9th International Bologna Conference

Magnetic Resonance in Porous Media (MRPM9)

July 13 - 17, 2008

Schlumberger – Doll Research

Cambridge MA, USA

Book of Abstracts

Tutorials 2

Invited Talks 5

Contributed Talks 25

Posters 53

Author Index 169

TUTORIALS

2

Magnet Systems for Unilateral Magnetic Resonance

B. Balcom

MRI Centre, Department of Physics, University of New Brunswick, Fredericton, Canada

Magnetic Resonance in its various implementations has proven to be among the most

powerful, and certainly most general, analytical techniques available to modern science.

Following several decades of development, the state of the art for clinical MRI and

liquid/solid state spectroscopy instruments is very highly evolved, generally featuring

expensive high field superconducting magnets.

While low field permanent magnet based MR instruments feature inherently low

SNR, with thermal polarization, the ideas of unilateral magnetic resonance (also termed ex

situ NMR or single sided NMR) have captured the imagination of many researchers world-

wide. These instruments are attractive and exhilarating because the development work is very

hands-on and still very early stage - which means that new, simple, physical ideas are readily

translated to new designs.

Low field portable magnets (which includes luggable magnets), instruments, and

associated measurement developments, over the last decade, will be reviewed in this tutorial

lecture.

Our emphasis will remain true single-sided instruments rather than the more general

topic of low field magnetic resonance. Instrument developments over the last decade already

demonstrate a number of trends, which permit reasonable extrapolation to future instruments

and measurements. The tutorial will therefore conclude with judicious speculation on future

unilateral MR instruments and measurements.

3

Diffusion in porous media

Denis S. Grebenkov

LPMC, CNRS – Ecole Polytechnique, F-91128 Palaiseau, France

A geometrical confinement considerably affects the diffusive motion of the nuclei and the

consequent signal attenuation under inhomogeneous magnetic fields. In the tutorial lecture,

we focus on theoretical and numerical aspects of restricted diffusion in NMR. Starting from

the classical Bloch-Torrey equation, we obtain the free induction decay (FID) and the spin-

echo or gradient-echo signal in a compact matrix form. Each attenuation mechanism

(restricted diffusion, gradient dephasing, surface or bulk relaxation) is represented by a

matrix which is constructed from the Laplace operator eigenbasis and thus depending only on

the geometry of the confinement. In turn, the physical parameters (free diffusion coefficient,

gradient intensity, surface or bulk relaxivity) characterize the "strengths" of the underlying

attenuation mechanisms and naturally appear as coefficients in front of these matrices. Once

the Laplacian eigenfunctions for a given confinement are found (analytically or numerically),

further computation of the macroscopic signal is more accurate and much faster than by using

conventional simulation methods. The matrix technique is actually a simple numerical tool

to deal with arbitrary gradient waveforms, including simple or stimulated, single or multiple

spin echoes. We illustrate its efficiency by considering restricted diffusion in simple

domains: a slab, a cylinder, and a sphere. Classical and recent theoretical advances achieved

by using Laplacian eigenfunctions are overviewed.

4

INVITED

TALKS

5

Moving NMR

Bernhard Blümich and the Aachen team

Institute of Technical and Macromolecular Chemistry, RWTH Aachen University, Germany

A unique aspect of Nuclear Magnetic Resonance is that it is in uninterrupted

evolution towards new methodologies, instrumentation and applications since its first

experimental verification in condensed matter in 1945. The evolution of conventional NMR

seems to follow a path to stronger and homogeneous magnetic fields to make use of higher

sensitivity and wider spectral dispersion. At the same time, electronic components become

smaller and smaller and software becomes more powerful, leaving the magnet and possibly

the rf power amplifier as the volume determining quantities of NMR instruments. Mobile

NMR is moving that frontier by introducing small NMR sensors and with it methods capable

of dealing with the challenges associated with small magnets.1 Prominent challenges are low

field strength, low field homogeneity, temperature drift of permanent magnets, and low

electric power consumption for battery operated measurements.

The evolution of NMR is reviewed and the prominent challenges of mobile NMR are

addressed. General solution strategies are outlined and particular solution concepts of the

Aachen team are reported together with an overview of novel applications of mobile NMR.

References:

1. B. Blümich, F. Casanova, J. Perlo, Mobile Single-Sided NMR, Prog. Nucl. Magn. Reson.

(2008) in press, doi:10.1016/j.pnmrs.2007.10.002.

6

Probing Multi-Component Transport in Porous Media over a

Hierarchy of Length-Scales

L. F. Gladden a, T.C. Chandrasekera

a, Y.Y. Chen

a, J.H.P. Collins

a, C.P. Dunckley

a,

E.J. Fordham b, M.L. Johns

a, M.D. Mantle

a, J. Mitchell

a, M.H. Sankey

a,

and A.J. Sederman a

a University of Cambridge, Department of Chemical Engineering, Pembroke Street,

Cambridge CB2 3RA, U.K. b Schlumberger Cambridge Research, High Cross, Madingley Road,

Cambridge CB3 0HG, U.K.

Understanding transport processes in porous media is central to the design and operation of

many chemical and physical processes. Ongoing work in the group aims to develop a suite

of techniques which can probe transport processes in porous media over a hierarchy of

length-scales from 10-8

-10-2

m. This talk will summarise the motivation for applying

magnetic resonance techniques to study transport phenomena in a number of different

systems, and then present illustrative results. Three areas of application will be considered:

Oil recovery – identifying oil and water fractions in a permeable rock

Recently we have implemented a pulse sequence in which T1 relaxation times have been

encoded in the second dimension of two-dimensional relaxation correlation and exchange

experiments using a rapid “double-shot” T1 pulse sequence. The technique retains chemical

shift information (δ) for short systems. Thus, a spectral dimension is incorporated into a

T2-T1-δ correlation without an increase in experimental time compared to the conventional,

chemically insensitive T1-T2 correlation. This approach enables the unambiguous

identification of oil and water fractions in a permeable rock.

*

2T

Pharmaceutical delivery systems – quantifying rapidly evolving pore-size distributions

For reasons of both ease-of-use for the consumer and more effective treatment, there is

increasing motivation to understand and control the release of drug into the body from a

‘delivery system’; e.g. tablet, polymer extrudate. Magnetic resonance is well-established for

use in characterising pore-size distributions in, for example, rocks – in such systems the

pore size is usually constant with time. When considering pharmaceutical delivery systems,

the pore size will evolve with time. Recent data will be presented which demonstrate that

during release, both the pore structure and the dissolution of the pharmaceutical active can

be followed using combined 1H T2 relaxometry/ q-space molecular displacement imaging

and 19

F NMR, respectively.

Chemical reaction engineering – velocity mapping of gas and liquid in two-phase flows

The fixed-bed reactor is widely used throughout the chemical industry. The process unit

comprises a cylindrical column packed with porous catalyst particles. In the context of this

meeting, the fixed-bed reactor can therefore be considered as a model hierarchical porous

structure comprising the macroscopic pore space of the inter-particle space within the

cylinder, and the micropore space within the catalyst pellets. Here we will focus on recent

results in which we have imaged both gas and liquid velocities within the inter-particle

space. These data allow us to compare the characteristics of two-phase (gas-liquid) flow

with that of a single-phase flow within the same porous structure. Further, these data

provide us with first measurements of gas, liquid flow velocity and particle wetting upon

which predictive models of reactor performance can be developed.

7

Observation of Fragile-to-Strong Dynamic Crossover in Confined

and Hydration Water and Its Relation to the Liquid-Liquid Critical

Point in Supercooled Water

Sow-Hsin Chen

Department of Nuclear Science and Engineering, MIT, Cambridge, MA

We have observed a fragile-to-strong dynamic crossover phenomenon of alpha relaxation time

[1,4] and self diffusion constant [4,5] in 1-d and 2-d confined deeply supercooled water. The

alpha relaxation time is measured by Quasi-Elastic Neutron Scattering (QENS) experiments and

the self-diffusion constant by Nuclear Magnetic Resonance (NMR) experiments. Water is

confined in 1-d geometry in cylindrical pores of porous silica materials, MCM-41 and in Double-

Wall Carbon Nanotubes (DWNT) [6]. It is in a 2-d geometry as the hydration water on surfaces

of biopolymer, protein, DNA, and RNA.

The crossover phenomena can also be observed by measuring the Mean Square Hydrogen Atom

Displacement derived from an Incoherent Elastic Neutron Scattering experiment and from

appearance of a Boson peak in an Incoherent Inelastic Neutron Scattering experiment. We

observe a pronounced violation of the Stokes-Einstein relation at and below the crossover

temperature at ambient pressure [2]. Upon applying pressure to the confined water, the

crossover temperature is shown to track closely the Widom line emanating from the existence of

a liquid-liquid critical point buried in an unattainable deeply supercooled state of bulk water [1].

Relations of the dynamic crossover phenomena to the existence of a density minimum in

supercooled confined water [3] and to conformational flexibility of the hydrated biopolymers [4]

will be discussed. The crossover temperature is shown to be sensitively dependent on the degree

of hydrophilicity of the confining substrate [6]. The crossover phenomenon have also been

confirmed by an MD simulation of a hydrated lysozyme powder model [7].

1. Li Liu, Sow-Hsin Chen et al, Pressure dependence of fragile-to-strong transition and a possible second critical

point in supercooled confined water, PRL 95, 117802 (2005).

2. Sow-Hsin Chen, Francesco Mallamace et al, The violation of the Stokes-Einstein relation in supercooled water,

PNAS 103, 12974 (2006).

3. Dazhi Liu, Sow-Hsin Chen et al, Observation of the density minimum in deeply supercooled confined water,

PNAS, 104, 9570 (2007).

4. Sow-Hsin Chen, Li Liu, et. al., Observation of fragile-to-strong dynamic crossover in protein hydration water,

PNAS 103, 9012 (2006).

8

5. Franceso Mallamace and Sow-Hsin Chen et al, Role of the solvent in the dynamical transitions of proteins:the

case of the lysozyme-water system, JCP 127, 045104 (2007)

6. Xiang-Qiang Chu, Alexander I. Kolesnikov, A. P. Moravsky, V. Garcia-Sakai, and Sow-Hsin Chen,

Observation of a Dynamic Crossover in Water Confined in Double-wall Carbon Nanotubes, Physical Review

E 76, 021505 (2007).

7. Marco Lagi and Sow-Hsin Chen et al, The Low Temperature Dynamic Crossover Phenomenon in Protein

Hydration Water: Simulations vs Experiments, to appear in J. Phys. Chem. B (Letter), 2008.

9

Some new developments in multidimensional Fourier- and

Laplace-inversion NMR

Paul T. Callaghan, Lauren M. Burcaw, Petrik Galvosas, Daniel Polders

and Kate E. Washburn.

MacDiarmid Institute for Advanced Materials and Nanotechnology, School of Chemical and

Physical Sciences, Victoria University of Wellington, New Zealand.

In the last five years the use of two-dimensional inverse Laplace Transformation (ILT)

methods [1] has become well established in NMR studies of porous media. Indeed, the most

powerful applications of multidimensional inverse Laplace spectroscopy may well be in the

investigation of fluid displacement in porous media. A particular virtue is the ability to work

in low field, as well as high, making these approaches applicable in the widest possible range

of NMR from laboratory to portable NMR to bore-hole NMR. Separation, correlation and

exchange experiments have been demonstrated, with the Laplace domain traversing T1, T2

relaxation, diffusion and internal magnetic gradients, the latter manifest in the decay of

echoes due to diffusion in the presence of inhomogeneous local fields. Of course, the ILT

does somewhat distort the representation of distributions due to pearling effects, but in a

sense, the “binning” of data in this manner can prove quite useful. At the very least, the 2D

ILT gives us an excellent means of separation of relaxation and diffusion behavior in

complex materials, thus enhancing interpretation and resolution. But it has also given us

correlation and exchange analogues of considerable power. All three variants will be

traversed here [2-5]. The exchange experiment is particular effective in revealing dynamics

through the mixing time dependence of peak intensities [6].

Recently, it has become apparent that the addition of a Fourier dimension can be of particular

value. One example is where a 3rd

propagator dimension is added. For example, in the case

of relaxation exchange a third Fourier dimension relating to molecular displacements can be

added [7], allowing spatio-temporal insight regarding spin relaxation for fluid state molecules

in porous media [8]. Another example is in the correlation of local fields with relaxation,

diffusion and local gradients [9].

An overview of these methods will be presented along with some speculation as to their

future potential.

1. Y.Q. Song and L. Venkataramanan, J. Magn Reson, (2002) 154 261

2. P.T. Callaghan and I Fúro, , J. Chem. Physics, (2004) 120, 4032

3. Y. Qiao, P. Galvosas, and P. T. Callaghan, Biophys. J.” 89, 2899-2905 (2005)

4. Y. Qiao, P. Galvosas, T. Adalsteinsson, M. Schönhoff and P. T. Callaghan J. Chem. Phys

(2005) 122, 214912-1

5. K.E. Washburn, C.D. Eccles and P. T. Callaghan, J. Magn. Reson. (submitted) (2008)

6. K.E. Washburn and P. T. Callaghan, Physical Review Letters 97, 175502 (2006),

7. K.E. Washburn and P. T. Callaghan, J. Magn. Reson. 186, 337-340 (2007)

8. K.E. Washburn, C.H. Arns and P. T. Callaghan, Phys. Rev. E (in press) (2008)

9. L. M. Burcaw and P.T Callaghan, J. Magn. Reson. (submitted) (2008)

10

DNP-enhanced NMR analysis of water of soft material assembly

S. Han, E. McCarney, R. Kausik, B. Armstrong

Department of Chemistry and Biochemistry, University of California Santa Barbara, CA

A unique analysis tool for the selective detection of local water inside soft molecular

assemblies—hydrophobic cores, amyloid fibers, vesicle bilayers, micelles—contained in bulk

water is presented [1]. This was made possible through the use of the Overhauser effect for

dynamic nuclear polarization to amplify 1H NMR signal of water through its interaction with

stable radical probes that possess ~660 times higher spin polarization compared to 1H nuclei

[2-5]. We developed 1H-Overhauser spectroscopy to provide unique and complementary

information to cw electron spin resonance analysis of spin labeled molecular assemblies

through the characterization of water interacting with the site-specifically localized spin label

[1]. We demonstrate that (1) hydration and water diffusion versus chain dynamics inside

bilayer systems can be measured and (2) tau protein aggregation to bona fide fiber versus

non-specific tau agglomeration can be differentiated and dynamically monitored, as only the

former involves water exclusion through the formation of hydrophobic regions (Fig.1).

Fig. 1: Tau-187 aggregation monitored by 1H Overhauser spectroscopy (dotted line), turbidity

measurements (straight line) and electron microscopy (pictures) for spin labeled tau-187 mutant SL322 that

forms bona fide fibers versus mutant 413 that form non specific agglomerates.

Detection of site-specific water exclusion (Ç) distinguishes fiber vs. gel unlike light scattering (–)

hydrated gel

amyloid fiber

100nm

100nm

hydrated gel

amyloid fiber

References:

1. E.R. McCarney, B. Armstrong, S. Han, Dynamic nuclear polarization enhanced nuclear

magnetic resonance and electron spin resonance studies of hydration and local water

dynamics in micelle and vesicle assemblies, Langmuir (2008) in press.

3. K.H. Hausser, D. Stehlik, “Dynamic nuclear polarization in liquids”, Advances in

Magnetic Resonance 3 (1968), 79-139.

4. I. Nicholson, D.J. Lurie, F.J.L. Robb, “The Application of Proton-Electron Double-

Resonance Imaging Techniques to Proton Mobility Studies”, J. Magn. Reson. B 104

(1994), 250-255.

5. W. Barros, M. Engelsberg, “Enhanced Overhauser contrast in proton-electron double-

resonance imaging of the formation of an alginate hydrogel”, J. Magn. Reson. 184 (2007),

101-107.

11

Microimaging of Catalytic Hydrogenation using Parahydrogen

L.-S. Bouchard,1 S. R. Burt,1 M. S. Anwar,1 K. V. Kovtunov,2 I. V. Koptyug,2 T. Theiss,1 and A. Pines1

1Materials Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 947202International Tomography Center, Novosibirsk, Russian Federation.

We present a MRI method for the study of gas-phase reactions in microfluidic devices. With the useof parahydrogen induced polarization (PHIP) [1] and a novel heterogenized catalyst [2], we are able toprovide detailed flow map and visualization of the active regions in an operating catalytic microreactor.This provides a new platform that could be used for the evaluation of novel catalysts and reactors, wheretheir production in very small quantities could result in cost savings [3].

The hydrogenation of propylene gas in a microreactor (continuous flow mode) containing a solid cata-lyst is visualized (Fig. 1A). A mixture of 20% propylene, 40% parahydrogen and 40% orthohydrogen waspassed through the catalyst bed. The polarized product, propane, was imaged at 7.1–T. The thermally-polarized protons produced insufficient signal to generate high resolution (sub-mm) images, whereas thePHIP-polarized protons yielded a substantial sensitivity enhancement (Fig. 1A). Figure 1A is a map ofactive regions within the catalyst bed. The flow-compensated imaging sequence consisted of a pure phaseencoding protocol with zero velocity and acceleration moments. A high resolution flow map (Fig. 1B)was produced by velocity encoding along three axes.

Propane - PASADENA polarized

(A) (B) (C)Velocity map v = xx + zz

x

z

1 m/s scale

Controlled delivery of polarizationPropylene - thermally polarized

x

z

x

z

FIG. 1: MRI images of the microreactor during the hydrogenation of propylene into propane by PHIP. (A) 1H image

of the catalyst bed using thermal propene and polarized propane, (B) Flow map (2D projection of the velocity field)

of the polarized product, arrows correspond to the direction and magnitude of the velocity at each point, (C) Packet

of polarized propane released downstream of the catalyst bed, Td =0, 10 and 40 ms. The field of view is 2.3 by 7mm

in Figs. A and B and 6 by 23mm in Fig. C.

Figure 1C demonstrates the controlled delivery of polarized product downstream from the catalyst bedwith the use of isotropic mixing to prevent the otherwise rapid decay of the singlet states. Mixing allowsthe polarized product to escape the catalyst bed, which can be imaged after a delay Td , to allow the productto travel a well-defined distance. These results in gas-phase microflows during a chemical reaction shouldhave applications in chemical engineering, to correlate reactor morphology to active regions within thebed and to probe mass transport within the reactor. The timed release of polarization could also be used toextend the use of PHIP beyond hydrogenation reactions.

[1] C. R. Bowers and D. P. Weitekamp, Phys. Rev. Lett. 57, 2645 (1986).

[2] I. V. Koptyug et al., J. Am. Chem. Soc. 129, 5580 (2007).

[3] L.-S. Bouchard et al., Science 319, 442 (2008).

12

Remotely Detected NMR in Microfluidic Devices with High

Spatiotemporal Resolution

Elad Harel, Vik Bajaj, Monica Smith, Alex Pines

Materials Sciences Division, Lawrence Berkeley National Laboratory, and Department of

Chemistry, University of California at Berkeley, Berkeley, Californian 94720-1460, USA,

Microfluidics, a technology in which fluid flows are manipulated on short length scales, has

delivered integrated “lab on a chip” analytical platforms with increasingly routine applications in

molecular biology, synthetic

chemistry, and in fundamental

studies of fluid dynamics. While

analytes are typically detected

optically, magnetic resonance (MR)

is emerging as a versatile tool in

microfluidic applications because

spectra rich in chemical information

can be acquired without the need for

labeling of the analytes. Its principal

disadvantages as compared to optical

detection are low sensitivity and low

time resolution, which limits the time

scale of chemical dynamics that can

be probed by this technique. The

latter is fundamentally set by the time

required to observe the NMR

spectrum: long observation times are

needed to resolve closely spaced

resonances, complicating the direct

monitoring of fast dynamics.

Practically, it is also limited by the

geometry of the microfluidic

experiment, in which the ratio of

detectable spins to the sensitive volume of the detector is ~10-4

or less, precluding the use of fast

imaging sequences. Further, magnetic susceptibility gradients imposed by the geometry of the

chip and fluid channels dramatically limit the resolution of directly acquired spectra (~10 ppm

linewidths) and thus their chemical information content. Here, we demonstrate a variant of

remotely detected NMR which overcomes these limitations. By employing spatial encoding of

the flowing fluid together with conjugate imaging in the detector coil, we show how high

resolution spectra can be acquired with arbitrarily high temporal resolution and several orders of

magnitude greater sensitivity than is possible by direct MR measurements. The enhancement is

evidenced by recording spectrally resolved fluid mixing through an arbitrarily complex lab-on-a-

chip device at 500 frames per second - one of the fasted frame rates recorded in an MRI

experiment. This is achieved by combining remote detection NMR with a time ‘slicing’ of the

time-of-flight (TOF) dimension which eliminates the constraints of the limited observation time

by converting the time variable into a spatial variable through the use of magnetic field gradients.

This method has implications for observing fast processes, such as fluid mixing, with sub

millisecond time resolution at micron spatial resolution and as a new modality for on-chip

chromatography.

Figure 1. TOF partial images of fluid flow through a

lab-on-a-chip device. Information about the spatial

origin of each fluid species is encoded and stored

along the B0 axis before stroboscopic acquisition in

the detector by a microsolenoid detection coil.

13

Paramagnetic and Super-paramagnetic Nanoparticles: Magnetic

Beacons and Magnetic Bullets for Lesion Detection and Therapy

Yung-Ya Lin

Department of Chemistry and Biochemistry, UCLA, CA, 90095, USA

The strong fluctuating magnetic fields from paramagnetic or super-paramagnetic materials

enhance the longitudinal (1/T1) and transverse (1/T2) relaxation rates of surrounding water

protons. For examples, in porous media, the additional relaxation dynamics of protons with

paramagnetic impurities located on the grain surface provides information about transport

processes and characteristic length scales of the pore space. Nanoscale paramagnetic or super-

paramagnetic iron oxide (SPIO) cores of magnetite and/or maghemite with appropriate surface

chemistry have recently demonstrated their utility in molecular imaging for enhancing MR

contrast, allowing researchers to monitor not only anatomical changes, but physiological and

molecular changes as well. Applications have ranged from detecting inflammatory diseases via

the accumulation of non-targeted SPIO in infiltrating macrophages to the specific identification

of cell surface markers expressed on tumors.

In this presentation, our recent methodological developments in applying SPIO as "magnet

beacons" for lesion detection and as "magnet bullets" for lesion therapy will be highlighted.

SPIO as "magnet beacons"

(1) Sensitive detection of SPIO by feedback-enhanced magnetic resonance.

Through the use of feedback magnetic fields, active feedback electronic devices, and SPIO's

strong microscopic susceptibility effect, we are able to improve the detection sensitivity of early

tumor by ~5 times. This is demonstrated by using biomarkers consisting of 50nm SPIO

conjugated to mouse anti-carcinoembryonic antigen (CEA) which were injected intravenously

through the tail vein of an immune-gene knockout mouse with a human lung cancer (cell line

Colo 205) grown in the thigh.

(2) Detection of H5N2 virus by SPIO aggregation.

SPIO as "magnet bullets"

(3) Heating of SPIO in high-frequency magnetic field for electromagnetic hyperthermia.

Hyperthermia is a promising approach to cancer therapy. It is a minimal invasive method for

regional selective heat treatment, which is based on heating the target tissue to temperatures

between 42 to 46ºC. By this technique, the viability of cancer cells is reduced because they are

more sensitive to temperatures above 41ºC than are the normal cells. The magnetic materials

generate heat in an alternating magnetic field, which enables the induction of hyperthermia. The

sub-domain SPIO particles produce substantially more heat per unit mass than the much larger

multi-domain particles of similar composition. The mechanism of heating is based on the Brown

effect (rotation of the particle as a whole according to external magnetic field) and the Neél

effect (reorientation of the magnetic moment across an effective anisotropy barrier within each

particle).

14

Dispersion in packed beds

U.M. Scheven a , R. Harrisb and M.L. Johnsb,

aREQUIMTE/CQFB, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, Portugal, bMagnetic Resonance Research Laboratory, Department

of Chemical Engineering, University of Cambridge, UK

The experimental characterization of voidspaces and flow within natural rocks, packed bed reactors, chromatography columns, or in simple packs of mono-disperse solid spheres generally includes measurements of volume averaged properties such as permeability, porosity, dispersivity, and sometimes the hydrodynamic radius rh=V/S, where V and S are the volume and surface area of the pore space respectively. Displacement encoding NMR experiments have made significant contributions to this area of research, with measurements of short time restricted diffusion coefficients yielding the hydrodynamic radius of a pore space1, and with APGSTE flow propagator2 and dispersion3 experiments in packed beds determining pore space dispersivities or effective diffusion coefficients. It is clear, however, that NMR derived dispersivities in packed beds - the one random porous system for which there exist canonical but incompatible theoretical predictions with few or no adjustable parameters4,5 - can be affected by the same experimental complications which have substantially contributed to the puzzling scatter in published dispersion results based on elution experiments6. Notable among these are fast flow near walls, and inhomogeneous flow injection. We will discuss how, with data analysis accounting for these macroscopic flow heterogeneities, an APGSTE-NMR dispersion measurement on flow through a tube of Diameter D packed with monodisperse spheres of diameter d can yield the dispersivities of the infinite pack of spheres, provided d<<D.

References:

1. P.P. Mitra, P.N. Sen and L.M. Schwartz, Phys. Rev. B 47 (1993) 8565. 2. J. Kärger and W. Heink, J. Magn. Res. 51 (1983) 1.3. A. Ding and D. Candela, Phys. Rev. E 54 (1996) 656. 4. P.G. Saffman, J. Fluid Mech. 7 (1960) 194.5. D. Koch and J. F. Brady, J. Fluid Mech. 154 (1985) 399.6. N.W. Han, J. Bhakta and R.G. Carbonell, AIChE J. 31 (1985) 277.

15

Using Magnetic Resonance to Measure the Interplay of Structure

and Transport in Porous Media

Joseph D. Seymoura,c

, Sarah L. Coddb,c

,

aDepartment of Chemical and Biological Engineering, bDepartment of Mechanical and Industrial Engineering, and cCenter for Biofilm Engineering, Montana State University,

Bozeman, Montana U.S.A.

Determination of the structure and transport behavior of porous media is important in fields as diverse as geophysics and biomedicine. Whether modeling the fate and transport of contaminants and oil in the earth’s subsurface, or designing ceramics and porous filtration systems for biomedical, energy and environmental applications, an increased understanding of the correlation between porous media structure and transport would allow control of transport by structural modification. Biofouling of porous media due to attached microbial biofilm communities is an important feature of failure in biomedical filtration devices and in controlling subsurface transport with permeability altering biobarriers. Pulsed field gradient magnetic resonance (MR) measurements of the propagator of the motion and corresponding q-space diffraction in bead packs of varying particle to column diameter ratio indicate the impact of ordered packing on transport. Scaling of the effective longitudinal dispersion with Pe

2, the Taylor dispersion dominated regime, is observed in the most ordered bead pack in agreement with theory [1]. Power law scaling of the dispersion coefficient with Peclet number Pe

α decreases with increasing bead pack disorder. Porous media biofouling generates

Fig. 1: MR images taken with spatial resolution of 27.3 μm/pixel over an 8 mm thick slice for dp=100 μm particles in varying diameter columns. The impact of structural ordering on dispersion dynamics as a function of Peclet number is measured. The lines indicate the Taylor Dispersion Pe2 and power law fit to data Peα regimes

a transition in porous media from normal Gaussian transport associated with homogeneous structure to anomalous transport dynamics typical of heterogeneous media [2]. The potential to design porous media with structures in which transport dynamics respond in a controlled fashion to biofouling is considered.

References:

1. D.L. Koch et al., The effect of order on dispersion in porous media, J. Fluid Mech. 200 (1989) 173.

2. J.D. Seymour et al., Anomalous fluid transport in porous media induced by biofilm growth, Phys. Rev. Lett. 93 (2004) 198103.

16

NMR Quantification of Trabecular and Cortical Bone

Microstructure and Function

Felix W. Wehrli

Laboratory for Structural NMR Imaging

University of Pennsylvania School of Medicine

Bone is a complex composite material whose strength is determined by a combination of the

material’s intrinsic mechanical properties, its overall volume fraction and architecture at the

macro-, micro- and ultrastructural level. Structure at all scales is governed by the nature of the

remodeling process and any imbalance in homeostasis adversely affects the bone’s mechanical

competence. Trabecular bone, dominant in the vertebrae and ends of long bones, consist of a

network of interconnected plates and struts of about 100 m thickness immersed in a matrix of

marrow. The encasing shell of cortical or compact bone is characterized by a system of

interconnected channels and pores ranging in width from 100 nm to about 50 m.

NMR is uniquely suited to probe both trabecular and cortical bone either by direct

observation using high-resolution k-space sampling or indirectly by relaxometric techniques and

measurement of apparent diffusion. Bone is more diamagnetic than the surrounding marrow (by

about 2.5-3ppm S.I.) and the resulting induced magnetic fields in the marrow spaces cause a

structure-dependent modulation of the FID from which trabecular density and structural

anisotropy can be derived without the requirement for resolving the structures. Direct detection

requires resolution sufficient to at least partially resolve individual trabeculae, complicated

further by physiologic motion and image noise. It is shown that in combination with image

processing it is possible to resolve the 3D architecture of trabecular bone in vivo and characterize

the bone in terms of parameters of scale and topology, and further, that the method is able to

quantify structural changes in response to drug intervention in subjects with osteoporosis.

Alternatively, the image data can be fed into a finite element model from which the anisotropic

stiffness tensor can be retrieved.

The cortical bone’s pore structure is of a scale too small to resolve by NMR. A

complicating circumstance is that the protons of water occupying the pores, have extremely short

lifetime (<<1ms) caused by both, induced inhomogeneous fields and surface relaxation. We

show that by means of solid-state imaging techniques the water content and thus pore volume

fraction can be measured, providing a parameter relevant to bone strength. Further insight into

the properties of pore water and possibly pore architecture are provided by relaxometric

approaches and deuterium NMR, the latter making use of quadrupole splittings resulting from

molecular order of collagen-associated water. Lastly, diffusion across the haversian and

canalicular system, the sole process available for small-molecule transport between osteocytes

and the vascular system, is shown to be accessible through study of H2O - D2O exchange

kinetics.

In summary, NMR imaging has been shown to provide a wealth of information on bone

micro- and nano architecture yielding quantitative insight into its mechanical behavior in health

and disease.

17

Neural Tissue as Porous Media

P.J. Basser

Section on Tissue Biophysics and Biomimetics, NICHD, NIH

MR measurements of molecular displacements and porous media models provide useful information not only about transport properties and microstructure of inanimate porous materials, but also about living tissue, particularly neural tissue.

Hansen’s early ex vivo measurements of ADCs in nerves were followed by in vivo diffusion MRI studies demonstrating diffusion anisotropy in cat and in human white matter (WM). Diffusion tensor NMR and MRI (DTI), based on a Gaussian displacement PDF, characterized anisotropic water diffusion in brain WM using maps of material parameters derived from the apparent (or effective) diffusion tensor (ADT). Tuch used cross-property relations to derive an electrical conductivity tensor from this ADT.

Model independent approaches are also useful in characterizing neural tissue structure and organization. Callaghan and Xia’s "k- and q-space imaging" method was adapted to clinical scanners by Wedeen (DSI) to estimate a 3-D average propagator in each voxel, providing information about restriction, distinct compartments, etc. The burden of sampling E(q) uniformly throughout 3-D q-space led Tuch to consider functions of the average propagator, e.g., its orientation distribution function (ODF), obtained by collecting E(q) data only over a spherical shell in q-space. Alternatively, Pikalov et al. estimated the 3-D average propagator with less E(q) data by using CT reconstruction methods and a priori information.

Several model-based displacement MR approaches have also been used to estimate distinct features of neural tissue. One method treats the extra-axonal space in WM as hindered, described by a DTI model, and the intra-axonal space as restricted. This composite hindered and restricted model of diffusion (CHARMED) MRI framework has been extended to measure the diameter distribution in a pack of axons (AxCaliber), extending and adapting the approach Packer et al. used to estimate the diameter distribution in droplets.

Various porous media approaches have been developed to study gray matter (GM) microstructure. Treating GM as hierarchically organized, exhibiting fractal diffusion, Özarslan et al. proposed scaling relationships for time-dependent displacement distributions; this approach provides a means to estimate several parameters characterizing anomalous diffusion. Viewing dendrites in GM as a network of restricted tubes, Komlosh et al. have applied multiple PFG-MR sequences to measure and characterize microscopic diffusion anisotropy in GM, which appears using DTI to be macroscopically isotropic.

Useful analytical and simulation environments have been developed to aid in understanding water transport in neural tissue and its effect on the MR signal. Szafer and Stanisz used Monte Carlo methods in idealized WM to model MR signal attenuation. Wehrli’s group used digitized histological WM cross-sections to build more realistic micromodels to simulate diffusive transport and the MR signal attenuation. Frank has developed a computational environment to simulate diffusion in complex media that can treat microstructural heterogeneity over a large range of length scales. Sen et al. developed a model of diffusion within a pack of cylindrical tubes to assess how different microstructural parameters affect the ADCs parallel and perpendicular to the WM fiber axis.

In summary, the use of porous media concepts in conjunction with MR displacement measurements has resulted in the development of novel models and experimental methods to characterize biologically relevant transport properties in neural tissue, and measure microstructural features that otherwise could only be obtained by invasive and tedious histological analysis.

18

Diffusion MRI of complex order in tissue

Van Wedeen

The Athinoula A. Martinos Center for Functional and Structural Biomedical Imaging,

Massachusetts General Hospital, Charlestown, MA, USA

In the physics of continuous media including condensed physics, it is natural to begin the

description with consideration of order parameters; point-symmetry groups in crystals,

continuous groups in soft condensed matter, higher dimensional and outright unknown stuff in

the vacuum. Living tissues, as examples of ordered matter, present unique examples of complex

order. This talk will consider three cases - skeletal muscle, heart and brain - and discuss our

attempts to capture their order and disorder with diffusion MRI.

Local anisotropy of tissue is conveniently mapped by q-space diffusion MRI, in which images

are acquired of the mean time-dependent propagator of water diffusion at each spatial location x

in an NMR image p(x,r) where r denotes spatial displacement and q its Fourier-conjugate.

Methods have been proposed in which one either computes a digital image of the entire 3D

propagator at each location, effectively a 6-dimensional data set - 3 spatial and 3 diffusion - or

which approximate the propagator to second order, by a symmetric diffusion tensor D.

Historically, diffusion tensor methods preceded the more general q-space imaging methods by

nearly a decade: 1992 vs 2000, resp., causing much difficulty. Following Cory and Garroway,

this propagator approximates the spatial autocorrelation of the tissue at each MRI voxel.

Beginning with the simplest case, skeletal muscle locally specified by a direction vector at each

point. Since these fibers are inversion-symmetric, the set of directions comprise S2/±1 = RP

2.

Globally, most skeletal muscles are trivial as fiber bundles - smooth arrays of similar parallel 1D

fibers over a 2D base - but there are interesting exceptions. Crossings of fiber populations occur

in core of the tongue, where they enable the tongue to elongate as a “muscular hydrostat, and in

the wall of the esophagus, where fibers interweave in crossed helices to enable dynamic

narrowing of the lumen, in structure and function like Chinese finger cuffs.

Unlike skeletal muscle, the fibers of the myocardium are not discrete elements but a single

branching syncytium: the largest single cell. Just like a river that begins in and ends in un-

numbered streams and estuaries, myofibers have no defined beginnings or ends, but fan out in a

bi-cone of branching though which they couple to their myocardial environment - electrically,

chemically and mechanically. This cone is not circular but highly flattened. This flattening

separates the myocardium into ribbons of muscle 5-10 cardiac myocytes thick (50-100 ), in

humans many mm wide and of unknown length: “myocardial sheets”. Globally, myofiber

orientations encircle the ventricle, spiraling in toroidal paths concentric to the ventricular mid-

wall, a projection to 3D of the canonical action of exp(it) in SU(2) Q* the unimodular

quarternions. This special arrangement uniquely allows myocytes throughout the thick

ventricular wall to contract equally in ventricular ejection, shortening 11-13%, while generating

far larger spatially varying radial strains - thickening - of the wall of the heart that increase

smoothly from 20% to 50% from outer to inner wall. Combining MRI of fiber orientation with

velocity-sensitive MRI of myocardial 3D strain yields detailed maps of cardiac mechanics for the

first time noninvasively, as images, or in humans.

19

The global structure of the myocardial sheet system is not understood. They comprise non-

integrable plane field, for otherwise the heart presumably would disassemble, dilate and fail.

However, these planes sit asymmetrically in the ventricle, with alternating left and right pitch

over the free and septal walls. The sheets tend to sit at 45° to the radial direction of maximum

thickening and form a dense slip-system for cellular rearrangement, to augment myocardial

compliance, yet their motion has never been directly observed. It is known that sheet structure

may be compromised or lost entirely following cardiac damage, associated with increased

myocardial stiffness and reduced function.

While structure of cardiac and skeletal muscle is well represented by a second-order diffusion

tensor - fiber orientations by leading eigenvectors of diffusion tensors, sheet orientations by

second eigenvectors – such tensors fail to capture the complex fiber architecture of neural tissue,

particularly the brain. In particular, at scales accessible to MRI crossing fibers are indispensable

to neural tissue. Fortunately such fiber structure can be accurately captured by q-space diffusion

MRI. Numerical integration of fiber orientation fields in the brain, “tractography”, has produced

the first maps of fiber pathways in the human brain. This method is now becoming a clinical

standard in pre-surgical planning. The fiber pathways of the brain display striking and beautiful

3D order and symmetry, reflecting developmental history as well as their organizing principles

as at once a logical, topological and physical network. While multi-fiber order seems appropriate

and possibly sufficient for mapping white matter, the geometric order of gray matter including

the cerebral cortex is not yet known. To conclude, diffusion MRI of fiber structure in tissue

recasts familiar anatomy as the shadow order that freely extends into more than 3 dimensions.

20

Models and Applications of In Vivo Lung Morphometry with

Hyperpolarized 3He MRI in a Mild COPD Population

J. Quirka, A. Sukstanskii

a, B. Lutey

b, D. Gierada

a, J. Woods

a,c, M. Conradi

a,c, D. Yablonskiy

a,c

aMallinckrodt Institute of Radiology,

bDepartment of Pulmonology,

cDepartment of Physics,

Washington University, St. Louis, MO

Hyperpolarized 3He diffusion MRI is increasingly used to non-invasively quantify

local alveolar structure changes, such as those from Chronic Obstructive Pulmonary Disease

(COPD). Previously [1], we described an in vivo lung morphometry technique that decouples

the helium apparent diffusion coefficient (ADC) into components oriented along the

longitudinal (DL) and transverse (DT) axes of the acinar airways. The measured MRI signal is

modeled using Eq. (1), where is the error function.

1/2

1/2

0( ) exp ;

4T AN AN L T

AN

S b S bD bD D D DbD

(1)

More recently [2], we expanded this theory to express helium diffusion in terms of the

airway radii. In Weibel’s model of the lung [3], acinar airways are treated as cylindrical

objects covered by an alveolar sleeve, defined by its inner (r) and outer

(R) radii. In histological measurements of healthy human lungs at 90%

of TLC, the mean r ≈ 160 m and the mean R ≈ 350 m [3]. Following

this model, we simulated airways as a periodic structure of cylindrical

symmetry (shown right with one of the four alveoli removed). These

computer simulations revealed remarkable scaling relationships and

allowed us to represent the anisotropic apparent diffusion coefficients (DL and DT) in a rather

compact form. They also demonstrated the dependence of these ADCs on the b-value:

0 0 0 0(1 ), (1 )

L L L L T T T TD D bD D D bD (2)

where DL0, L , DT0 and T are functions of the model parameters, the diffusion-sensitizing

gradient waveform, and D0, the free diffusion coefficient of helium-3 gas in air. To gain

insight into the mechanisms of early COPD, we applied this theory to helium diffusion data

acquired from thirty “asymptomatic” subjects with significant smoking histories (50 ± 20

pack years, average age 62 ± 3 years) recruited from the National Lung Screening Trial, and

five healthy non-smokers (average age 34 ± 4 years). All procedures were performed with

IRB approval and 3He MRI was performed under a

3He IND FDA exemption.

In normal subjects, the mean airway inner and outer radii determined from the MRI

measurements were 170 ± 30 and 310 ± 40 m. Given that our measurements are performed

at about 60% of TLC, they are in a very good agreement with histology measurements [3].

The helium diffusion measurement showed increased mean internal radii in twenty five

smokers (mean 240 ± 40 m) and increased mean external radii in six (mean 325 ± 30 m).

Significant heterogeneity in these parameters was seen across the lungs in the majority of

smokers. Regions of increased internal radii were typically more extensive than regions of

increased external radii, suggesting that the inter-alveolar walls were damaged in this

population prior to gross airway dilation.

References:

[1] D. A. Yablonskiy, Proc Natl Acad Sci U S A (2002), 99, 3111-6

[2] A. L. Sukstanskii and D. A. Yablonskiy, J Magn Reson (2008), 190, 200-10

[3] B. Haefeli-Bleuer and E. R. Weibel, Anat Rec (1988), 220, 401-14

21

Nuclear Magnetic Resonance Applications to Unconventional Fossil

Fuel Resources

R.L. Kleinberg and G. Leu

Schlumberger-Doll Research, Cambridge, Massachusetts

Technical and economic projections strongly suggest that fossil fuels will continue to

play a dominant role in the global energy market through at least the mid twenty-first century.

However, low-cost “conventional” oil and gas will be depleted in that time frame. Therefore

new sources of energy will need to be developed. We discuss three relatively untapped

unconventional fossil fuels: heavy oil, oil shale, and gas hydrate. In each case, nuclear magnetic

resonance technology can play a key role in appraising the resource and/or providing information

needed for selecting production processes.

Heavy oil is a degraded form of petroleum found in extensive earth formations relatively

near the surface. One of the principal physical properties controlling its exploitation is its

viscosity. Viscosities of exploitable oils range from less than 1 mPa-s to more than 1,000,000

mPa-s (1 mPa-s = 1 centipoise). As long ago as 1948, Bloembergen, Purcell, and Pound showed

that NMR relaxation times are sensitive to the viscosity of fluids. In recent years it has become

possible to measure NMR relaxation times in earth formations in situ. However, heavy oils are

complex fluids with magnetization decays that are grossly nonexponential. Moreover, as found

in the earth, they are frequently mixed with water, which also relaxes nonexponentially due to

interactions with minerals that form the rock matrix. We have developed laboratory and

borehole methods that allow us to use NMR measurements to estimate the viscosity of oil, even

when the oil is mixed with water in the pores of natural rock.

Oil shale is a mixture of fine sediments and kerogen, which is organic matter that has not

been subjected to temperatures and pressures sufficient to convert it to petroleum. An immense

quantity of oil shale accumulated in a lake covering several counties in Colorado, Utah, and

Wyoming in the Eocene epoch, approximately 55 million years before present. Kerogen must be

heated to high temperature in order to convert it to oil and gas at economic rates. An NMR assay

of water content is potentially an important input into production process selection.

Gas hydrate is an ice clathrate that traps methane and other natural gases at elevated

pressures and temperatures above 0°C. Substantial quantities have been found in and below

permafrost, and under the seafloor at all latitudes. Building on novel experiments performed on

a submarine, and on a drilling rig on the North Slope of Alaska, NMR has been used to

determine how much gas hydrate is present in arctic and deep water accumulations.

22

Portable Quadrupole Resonance Systems

Hector Robert

Product Development Manager - Quantum Magnetics, Inc.

15175 Innovation Dr, San Diego, CA 92128 USA

Quadrupole Resonance (QR) sensors have the unique capability of detecting explosives with

remarkably high detection rates and low numbers of false alarm. This presentation describes

some of the technical and scientific accomplishments at Quantum Magnetics, Inc., a wholly-

owned subsidiary of GE Homeland Protection, that resulted in the fabrication of handheld,

portable QR systems for effective detection of energetic materials under field conditions. A

number of difficulties that impact operation in the field—such as external radio-frequency

interference (RFI), inhomogeneity of the excitation fields, and line broadening due to

temperature effects—have been mitigated to achieve the highest possible detection performance

with the lowest number of nuisance alarms. This presentation describes some of these

developments and presents results that demonstrate their usefulness for improved QR system

operation in the field.

23

CMOS mini-NMR biomolecular sensor

Donhee Ham

Division of Engineering & Applied Science

Harvard University, Cambridge, MA

My research group, in collaboration with Ralph Weissleder’s group at Massachusetts General

Hospital, has recently developed the smallest complete NMR relaxometry system ever built, at

the heart of which lies a sophisticated silicon RF chip that can elaborately manipulate water

proton spins and monitor their relaxation dynamics. It can detect biomolecules (e.g., cancer

marker proteins) in a sample (e.g. blood), by monitoring water (most bio samples contain a large

number of water molecules) proton spin dynamics affected by the biomolecules. This work is

marked with drastic size reduction and sensitivity enhancement. A state-of-the-art benchtop

NMR relaxometer is heavy (120 kg) and bulky (100-liter), due to a large magnet and discrete RF

electronics used. Our system weights only 2 kg (60 times lighter) and occupies 2.5 liter (40 times

smaller), while achieving 60 times better mass sensitivity. This was made possible by using a

small, fist-size magnet, but more crucially, by integrating RF electronics onto the silicon CMOS

chip. The design challenge we overcame was attaining RF performance far beyond what would

be typically sufficient for larger NMR systems, in order to detect and analyze NMR signals

severely degraded by the small magnet. This system can be used as a general NMR relaxometry

system, or for diagnostic analysis, in the small, portable platform at low cost, brought closer to

patients and doctors.

24

CONTRIBUTED

TALKS

25

Microscopic Anisotropy Revealed by Double-PFG NMR

E. Özarslana, P. J. Basser

a

aSection on Tissue Biophysics and Biomimetics, NICHD, NIH.

The multiple scattering generalizations [1] of the pulsed field gradient (PFG)

experiments have been predicted to be sensitive to restricted diffusion even at long wavelengths [2], i.e., when γδGa«1. Here, γ is the gyromagnetic ratio of the spins, G and δ are, respectively, the strength and duration of the applied magnetic field gradients, and a is a characteristic pore size. Such a pulse sequence is illustrated in Fig. 1.

We calculate the exact form of the quadratic term of the NMR signal attenuation,

obtained via NMR double-PFG experiments with arbitrary timing parameters (tm, Δ, and δ), from spins diffusing between two parallel plates, as well as in cylindrical and spherical pores. As shown in Fig. 2a, when tm is short, there is significant variation of the signal intensity with varying values of ψ; this is a manifestation of the sensitivity of the double-PFG experiment to microscopic anisotropy induced by the restricting walls of the spheres.

Fig. 1: The double-PFG experiment considered in this work. Two consecutive PFG blocks are separated by a mixing time tm. The separation between the two gradient pulses of each PFG block is denoted by Δ. The angle between the two gradients G1 and G2 is denoted by ψ.

In Fig. 2b, we consider a pack of coherently oriented infinitely long cylinders. The φ dependence of the signal when θ=900 is due to microscopic anisotropy whereas its θ dependence when φ=900 is indicative of ensemble anisotropy due to the coherence in the cylinders’ orientations. When the cylinders are isotropically distributed, similar to the case of spherical pores, only microscopic anisotropy can be observed. However, unlike in the case of spheres, signal attenuation can be enhanced significantly by increasing the diffusion time (Δ), eventually leading to a qualitatively different angular pattern (see Fig. 2c).

(c) randomly distributed cylinders

x

y

z

G1

G2

(b) coherently oriented cylinders(a) spherical pores

Fig. 2: NMR signal attenuation (E) vs. angle curves from (a) spherical pores of radius a; (b) cylindrical tubes of radius a coherently oriented along the z-axis where the orientations of the gradients are shown on the left; (c) isotropically distributed cylinders of radius a. D0 denotes the bulk diffusivity. Our results can be used to design experiments with many degrees of freedom and to obtain accurate information on pore microstructure using double-PFG acquisitions, including compartment size and fiber orientation distributions—all from the long wavelength regime of the NMR signal attenuation, i.e., using small gradient strengths. Therefore, the double-PFG technique, along with our quantitative findings, are expected to be useful in characterizing geometric features of small pores with possible biological and clinical applications.

References:

1. D. G. Cory, A. N. Garroway and J. B. Miller, Polym. Preprints. 31 (1990) 149. 2. P. P. Mitra, Phys. Rev. B. 51 (1995) 15074. This research was supported by the intramural research program of NICHD.

26

Propagator Resolved Transverse Relaxation Spectroscopy

K.E. Washburna,c

, C.H. Arnsb, P.T. Callaghan

a

aVictoria University of Wellington,

bAustralian National University,

cResLab, Reservoir

Laboratories AS

We present here a novel technique for characterization of porous media, propagator

resolved transverse relaxation exchange. This technique is the first experiment to combine

two inverse Laplace dimensions with a Fourier dimension. Previously, we had used a

transverse relaxation exchange experiment to determine the rate and extent of fluid exchange

between pores of differing sizes in Castlegate sandstone1. This 2D inverse Laplace

experiment is of similar form to Fourier exchange experiments, correlating the T2 values of

spin-bearing molecules at two different points in time. While T2 exchange is a useful

technique, it has the major limitation that it cannot differentiate between molecules that have

remained in their original environment and those that have moved to a new environment with

a similar T2 value. By adding the propagator dimension to the transverse relaxation

exchange experiment, we can not only determine where molecules are moving and how

quickly they get there, but how far they have moved as well.

We performed the propagator resolved T2 exchange experiments on tight-packed

quartz sand for a range of mixing times. The data are then Fourier transformed along the

diffusion axis to produce a propagator. Slices of T2-T2 exchange data are extracted from

along the propagator and inverted using the 2D inverse Laplace transform2. The difference in

the resulting spectra depending on displacement is marked. The signal from low molecular

displacements lies almost exclusively along the diagonal, which indicates no exchange. For

high displacement, multiple cross peaks appear, which, as with Fourier exchange

experiments, indicates movement between environments during the mixing time. We also

performed simulations of fluid movement within a pore glass system. By minimising the

difference between the measured and simulated results, we are able to estimate pore

characteristics such as pore size, inter-pore spacing, pore exchange times, and tortuosity.

Comparing our results to estimates from X-Ray CT, we see good agreement in pore size and

tortuosity values, but our method consistently underestimates inter-pore spacing. This most

likely stems from the limited distances that can be probed using diffusion, and we anticipate

work with flow will produce better agreement in the pore spacing estimates.

References:

1. K. E. Washburn and P.T Callaghan, Phys. Rev. Let 97 (2006) 175502.

2. L. Venkataramanan, Y.Q. Song, M.D. Hurlimann, IEEE Transactions on Signal

Processing 50 (2002) 1017.

27

Using heterogeneity spectra to describe porous media presenting

spatial heterogeneity on multiple length scales

Andrew E. Pomerantz, Peter Tilke, and Yi-Qiao Song

Schlumberger-Doll Research, 1 Hampshire St., Cambridge, MA 02139, USA

Naturally occurring porous media often present spatial heterogeneity on a wide range

of length scales. Carbonate rocks, which represent an important challenge to the oil industry

because they contain over half of proven oil reserves, are often considered to be

heterogeneous on length scales ranging orders of magnitude. MRI represents a powerful

experimental probe of these heterogeneous media because it can measure relevant quantities

with good spatial resolution over potentially large samples. In traditional quantitative

analysis of spatial heterogeneity, data from spatially resolved measurements such as MRI are

used to construct variograms or correlation functions, which are then fit to simple model

functions. This technique can determine the total heterogeneity (variance) and a single length

scale of heterogeneity (or a small number of discrete length scales when the data support a

nested structure). However, the interpretation becomes ambiguous for materials

simultaneously displaying many length scales of heterogeneity.

Here we present a statistical technique for inverting MRI measurements into a

heterogeneity spectrum, resolving that ambiguity by quantifying the extent of heterogeneity

as a function of length scale. A single experiment is sensitive to a range of heterogeneity

length scales determined by the experimental resolution and the sample size. The influence

of spatial averaging over the finite volume of a voxel (regularization) is modeled

mathematically, yielding the sensitivity of experimental data to heterogeneity at different

length scales. Those sensitivity functions are then used as basis functions for inverting the

data to a heterogeneity spectrum, showing heterogeneity as a function of length scale for all

length scales to which the measurement is sensitive. This transform presents the

experimental variogram in an easily interpretable form, just as an inverse Laplace transform

presents an echo train as an easily interpretable relaxation spectrum or a Fourier transform

presents an FID as an easily interpretable chemical shift spectrum.

We will present heterogeneity spectra derived from MRI images for a series of

sandstone and carbonate rocks. This measurement is sensitive to heterogeneity over the

range 10-0.5

– 102 mm. Over this range of length scales, the heterogeneity spectra of different

carbonates is shown to vary greatly: some carbonates present most of their heterogeneity on

relatively short length scales, some carbonates present most of their heterogeneity on

relatively long length scales, and some carbonates are as homogeneous as typical sandstones.

Examples are shown in the figure: rock A is quite heterogeneous, and most of that

heterogeneity occurs on relatively short length scales; rock B is somewhat less

heterogeneous, and heterogeneity at longer length scales dominates in this sample.

Figure 1: Example heterogeneity spectra

28

Dynamics at Surfaces : Probing the dynamics of polar and a-polar liquids

at silica and vapour surfaces.

J. Beau W. Webbera,b,d, John H. Strangeb, Philip Blandc, Ross Andersona, Bahman Tohidia aInstitute of Petroleum Engineering, Heriot-Watt University, Edinburgh, EH14 4AS, UK;

bSchool of Physical Sciences, University of Kent, Canterbury, Kent, CT2 7NH, UK; cImperial College London, South Kensington Campus, London SW7 2AZ, UK;

dLab-Tools Ltd., G19 Canterbury Enterprise Hub, University of Kent, CT2 7NJ, UK.

Recent studies by NMR and NS of the dynamics and phase-fractions of water/ice systems in templated porous silicas (SBA-15) show that what was believed to be a non-frozen surface water layer is actually plastic ice (Fig1a, b), the thickness varying with temperature.

Fig. 1a Variation with temperature of the long T2 relaxation time component for plastic ice in SBA-15 porous silica - Arrhenius plot. The lower activation enthalpy is due to the onset of rotational motion. Fig. 1b Variation with temperature of the amplitude of the long T2 relaxation time component for plastic ice in SBA-15. As the temperature is lowered the plastic ice converts reversibly to brittle ice.

More recent research has studied the dynamics of polar water/ice and a-polar organics

at both silica and vapour interfaces. The polar results are significant for water/ice systems in the environment. This research also points the way forward for wide-range cryoporometric metrology in ‘difficult’ systems such as high iron content clays and rocks, as well as aged concrete. Results will be presented for cryoporometric measurements on meteorite samples with a significant metallic content, exhibiting T2* relaxation times down to 2.5µs.

References:

1. Structural and Dynamic Studies of Water in Mesoporous Silicas using Neutron Scattering and Nuclear Magnetic Resonance. Beau Webber and John Dore. Invited article, IoP: Journal of Physics: Condensed Matter - Special Issue: Water in Confined Geometry. : 16, S5449-S5470, 2004. PII: S0953-8984(04)78970-5 .

2. Plastic ice in confined geometry: The evidence from neutron diffraction and NMR relaxation. J. Beau W. Webber, John C Dore, John H. Strange, Ross Anderson, Bahman Tohidi. J. Phys.: Condens. Matter 19, 415117, (12pp), 2007, Special Issue: Proceedings of The International Workshop On Current Challenges in Liquid and Glass Science .

3. Nuclear Magnetic Resonance Cryoporometry J. Mitchell, J. Beau W. Webber and J.H. Strange. Physics Reports, 461, 1-36, 2008. doi:10.1016/j.physrep.2008.02.001 .

6543

0.01

0.02

0.1

0.2

1

2

10 0C -50C -100C

1000/T K -1

T2

ms

0C

Cooling

Warming

Activation Enthalpy :810 kJ.mole -1

Activation Enthalpy :4.6 kJ.mole -1

T2 relaxation time for H 2Oin SBA-15 F5 silica.

x CPMG data

o FID data ( T2* deconvolved)

250200150

1

0.5

0

Temperature K

Rel

ativ

e A

mpl

itude

Amplitude for plastic icein SBA-15 F5 silica.

x CPMG data o FID data Cooling Warming

Max. plastic ice fraction = 0.52

Then, for Cylindrical geometry,max fractional thickness ofplastic layer = 0.31

Which for 86Å pore diameter gives maxthickness of layer of plastic ice = 13Å. Cylindrical pore,

diameter 86Å,surface layerof plastic iceof thickness13Å - 0.14Å.K -1 .

29

Direct probing of the wettability of plaster pastes at the nanoscale by proton field cycling relaxometry

J.-P. Korb and P. Levitz

Laboratoire de Physique de la Matière Condensée, Ecole Polytechnique, CNRS, 91128 Palaiseau, France.

How is it possible to probe non destructively the water wettability of a reactive porous plaster paste and how is it related to its macroscopic properties? Answering these two questions is very important in civil engineering to control the mechanical properties of such a widely used material. Here, we propose the proton nuclear magnetic relaxation dispersion (NMRD) to estimate the average lifetime of a water molecule on the pore surface of gypsum, a material of great interest in the building industry. Our first result at room temperature is the frequency dependence of 1/T1 that is constant below a cross-over frequency 22 kHz over which it behaves as a power law, 1/T1~ω-0.83 over more than three orders of magnitude either for light or heavy water. This is in favour of an intramolecular diffusive process on and at proximity of the entangled needle-shape gypsum flat surface [1, 2]. The second result for light water is that such a power law is preserved when varying the water-to-plaster weight ratio between 0.4 and 1 and the temperature between 25 and 45°C. However, one observes an anomalous behaviour of 1/T1 that increases with the temperature. This reveals an interaction with the solid surface. Last, one observes a net decrease of the exponent of the power law up to 1/T1~ω-0.50 in presence of adsorbed Sodium trimetaphosphate adjuvant. We propose an analytical model of the NMR relaxometry involving elementary time steps near the interface (e.g. bulk bridges, adsorption trails and escaping tails. This close-form model is supported by our experimental data. The model introduces a characteristic frequency )2( 22

0 ADτδω = that is related to the water size δ=0.3 nm, the translational bulk diffusion coefficient D of water and the average adsorption correlation time τA. The comparison with our data shows that this characteristic frequency is about ω0~0.1 MHz in normal hydration and increases to 875 MHz in presence of 6.5% of adjuvant. This gives an adsorption time scale that is τA=5.7 ns without adjuvant and decreases to 64 ps with 6.5 % of adjuvant.

In summary, the direct probing of water adsorption time on solid surface gives access to an original characterization of the surface nano-wettability of porous material. Possible extension of this method to various other materials is actually under investigation..

References: [1] P.E. Levitz, J. of Phys. Cond. Matter 17 (49), 54059-54075 (2005). [2] P.E. Levitz and J.-P. Korb, Europhys. Lett. 70, 684-689 (2005).

30

Fire spalling of concrete as studied by NMR

L. Pel, G.H.A. van der Heijden

Group Transport in Permeable Media,

Department of Applied Physics, Eindhoven University of Technology,

Eindhoven 5600 MB, The Nederlands

ABSTRACT

During the past thirty years concrete has developed enormously in both strength and

durability. A downside of these improvements is the increased risk of explosive spalling in

case of fire. The moisture inside the concrete plays an important role in this spalling

mechanism. As the temperature is increasing during a fire the water will start to boil in the

concrete. However due to its low permeability the vapour transport is hindered and a large

pressure will built up in the material. This pressure can ultimately lead to the explosion of the

concrete during fire. This phenomenon has been observed in the various tunnel fires during

the last years.

In order to study the moisture transport inside concrete during heating a special

Nuclear Magnetic Resonance setup was built. This setup can be placed inside a 1.5 T MRI

scanner. With this setup one dimensional moisture profiles can be measured non-

destructively while the concrete sample is heated up to 400 °C with heating rates of up to 10

°C/min. The setup can handle representative concrete samples up to 8 cm in diameter and 10

cm in length. Using this setup the moisture transport has been studied for various types of

concrete.

31

Identification of Endohedral Water in Single-Walled Carbon

Nanotubes by 1H NMR

Qiang Chen1, Julie L. Herberg

2, Gregory Mogilevsky

1, Hai-Jing Wang

1, Michael Stadermann

2,

Jason K. Holt2, Yue Wu

1

1Department of Physics and Astronomy and Curriculum in Applied and Materials Sciences,

University of North Carolina, Chapel Hill, NC 27599-3255 2Chemical Sciences Division, Lawrence Livermore National Laboratory,

Livermore, CA 94550

Water confinement within single-walled carbon nanotubes (SWCNTs) has been a topic of

current interest, due in part to their potential applications in novel molecular transport.

Experiments have recently validated molecular dynamics predictions of flow enhancement

within these channels, although few studies have probed the detailed structure and dynamics of

water in these systems. 1H Nuclear Magnetic Resonance (NMR) is a technique capable of

providing some of these details, although care must be exercised in separating confined water of

interest from exterior water. Using controlled experiments with both sealed and opened

SWCNTs, and providing a quantitative measure of water content through desorption

experiments, a signature for confined water in SWCNTs has been positively identified. This

interior water is characterized by a relatively broad feature located at 0.0 ppm, shifted upfield

relative to bulk water. With the identification of a signature for water inside SWCNTs, further

studies aimed at probing water dynamics including diffusion by NMR in SWCNTs will be

enabled.

Prepared by LLNL under Contract DE-AC52-07NA27344.

32

High resolution q-space imaging studies of water in elastin

G.S. Boutis*, C. Renner, T. Isahkarov, T. Islam, L. Kannangara, P. Kauer, E. Mananga, A.

Ntekim, Y. S. Rumala, D. Wei

York College of The City University of New York

Q space NMR imaging is a well-known non-invasive experimental technique for structural

investigations of a variety of complex systems relevant to problems in industry, material science

and biology. The technique allows one to accurately measure the morphology of a confining

pore, and molecular diffusion rate of mobile molecules within interstices of a structurally

complex system. In our laboratory we have recently designed a variable temperature NMR

microscope capable of delivering gradient pulses on the order of 20,000 G/cm for high resolution

scattering studies.

Elastin is an insoluble and highly cross-linked protein in the extra cellular matrix responsible for

the elastic properties of vertebrate tissues. Its soluble precursor, tropoelastin, has two main

domains. The hydrophilic cross-linked domains are rich in lysine and alanine and the

hydrophobic domain, which is responsible for elasticity, is rich in valine, proline and glycine.

Several models have been proposed to account for the heterogeneous and complex nature of

elastin at microscopic level. These models can be divided into two major categories, single and

two-phase models .The single phase model is also known as the random chain model. The two-

phase model can further divide into liquid–drop, oiled-coil and the β-spiral models. NMR

relaxation studies performed by Ellis and Packer showed the existence of water in two distinct

environments within the elastin fiber. All models for elasticity require water as a plasticizer yet

provide little insight into it’s dynamics in elastin. In this presentation, we report on the strong

gradient, variable temperature NMR microscope we designed and the measurements of the water

diffusion coefficients as a function of temperature. The instrument was successfully implemented

to measure the surface to volume ratio of pores within elastin fibers and associated changes as a

function of temperature. The impact of these findings on the structure of elastin and its

dependence on hydration are discussed.

References 1. W. Zhang and D. G. Cory (1998) Journal of Magnetic Resonance 132: 144-149

2. S. M. Partridge (1962) Advances in Protein Chemistry 17: 227-302

3. B. B. Aaron and J. M. Gosline (1980) Nature 287: 865 – 867

4. C. A. Hoeve and P. J. Flory (1974) Biopolymers 13: 677 – 686

5. J. M. Gosline (1978) Biopolymers 17: 677-695

6. T. Weis-Fogh, and S. O. Andersen (1970) Nature 227: 718 – 721

7. R.W. Gray, L.B. Sandberg and J.A. Foster (1973) Nature 246: 21 – 28

8. D. W. Urry, B. Starcher and S. M . Partridge (1969) Nature 222: 795-796

9. G.E. Ellis and K.J. Packer (1976) Biopolymers 13: 813 – 833

10. G.S. Boutis*, C. Renner, T. Isahkarov, T. Islam, L. Kannangara, P. Kauer, E. Mananga,

A. Ntekim, Y. S. Rumala, D. Wei (2007). 'High resolution q-space imaging studies of

water in elastin' Biopolymers 87, 5-6, 352- 359.

33

Sensitivity Enhancement of Multi-Acquisition/Multi-Dimensional Earth’s Field NMR

Meghan E. Halse and Paul T. Callaghan

The MacDiarmid Institute, Victoria University of Wellington, Wellington, New Zealand

Earth’s field nuclear magnetic resonance (EFNMR), which is nearly as old as NMR itself,

has long been used in the field of magnetometry and for teaching the principles of magnetic resonance to students. More recently, however, EFNMR has been shown to have promise for a much broader range of applications including multi-dimensional imaging, high resolution spectroscopy and measurements of molecular diffusion.

While EFNMR enjoys the advantage of sub-hertz spectral resolution due to the natural homogeneity of the Earth’s field, it suffers from low sensitivity, a consequence of the approximate B0

2 dependence of signal-to-noise ratio (SNR) on field strength, which for the Earth’s field is on the order of 50 µT. This limitation can be partially overcome through the use of pre-polarization techniques, whereby the sample is allowed to come to thermal equilibrium in a strong pre-polarizing field prior to excitation and detection in the Earth’s magnetic field. This technique has enabled high SNR to be achieved in single-shot NMR experiments where a 1 T Halbach magnet is used for pre-polarization and the sample is manually transported to a B1 coil far removed from this magnet for detection [1]. However, this method is not readily adapted to multi-acquisition or multi-dimensional NMR experiments.

The goal of the current research is to explore several methods for enhancing the sensitivity of multi-acquisition and multi-dimensional Earth’s field NMR experiments. In addition to Halbach pre-polarization, the advantages and disadvantages of a simple pulsed electromagnet for pre-polarization will be considered. Polarization enhancement via dynamic nuclear polarization (DNP), a method which has been shown to be highly effective for improving the sensitivity of continuous wave NMR magnetometers [2], will also be explored. Preliminary experiments and comparisons will be presented and discussed. References: 1. Appelt, Stephan, Haesing, F. Wolfgang, Kuehn, Holger, Sieling, Ulrich, Bluemich, Bernhard,

Chemical Physics Letters, 440 (2007) 308-312. 2. Kernevez, N and Glenat, H, IEEE Transactions on Magnetics, 27 (1991) 5402-5404.

34

Flow of concentrated suspensions in asymmetric bifurcations

Nina C. Shapley a and Chunguang Xi a

aDepartment of Chemical Engineering, Columbia University 500 W. 120th Street, MC 4721, New York, NY 10027

The bifurcation geometry is a basic component of porous media, where it is interesting to consider how dispersed particles in a suspension flowing through the bifurcation partition between downstream branches. Previous work on the behavior of dispersed particles in branching flows has generally emphasized dilute suspensions where the particle diameter is similar to the channel width. Meanwhile, a high loading of small particles, where the suspension can be compared to a continuum material, and the resulting impact on the concentration and flow fields have not received as much attention.

0 25 50 75 100 125 150 175 2000.0

0.1

0.2

0.3

0.4

0.5

Sid

e br

anch

frac

tion

Suspension flow rate (ml/min)

Flow rate(a) Flow rate(b) Particle flux(a) Particle flux(b) Area(a) Area(b)

Fig. 1: Side branch fraction of particle flux, flow rate and inlet area at five different flow rates (bulk particle volume fraction= 0.4) for the equal branch flow cell (a) and unequal branch flow cell (b).

In our study, suspensions of neutrally buoyant, noncolloidal spheres in Newtonian liquids undergo steady, pressure-driven flow in a rectangular channel (4:1 aspect ratio) that divides into two branches at an asymmetric T-junction. We examine two cases, where the downstream branches either have equal width or unequal widths in a ratio of 3:2. Particle concentration and velocity profiles are obtained by nuclear magnetic resonance imaging (NMRI). We aim to determine the effect of the branching ratio and geometry on the observed concentration and flow fields, for particle volume fractions of 0.4-0.5 and low flow and particle Reynolds numbers. We find that the particles follow flow streamlines fairly closely for the unequal branch flow cell, while in the case of equal branches, the particles are more evenly distributed between the downstream branches than expected.[1,2] Recent results from bifurcation flow experiments will be presented, comparing the two bifurcation geometries in terms of dividing streamlines, concentration inhomogeneities, and particle fluxes.

References: 1. C. Xi and N. C. Shapley, Flows of concentrated suspensions through an asymmetric

bifurcation, J. Rheology, 52 (2008), in press. 2. C. Xi and N. C. Shapley, Comparison of concentrated suspension flows through

bifurcations with equal and unequal branches, Phys. Fluids, submitted, March, 2008.

35

MRI Pressure Measurement in Novel Homogeneous Soft Solids

R. Morrisa, M. Bencsik

a, N. Nestle

b, R. Kaur

c, P. Galvosas

d, Y. Perrie

c

aNottingham Trent University,UK;

b BASF SE Ludwigshafen, Germany;

cAston University, UK;

dUniversität Leipzig, Germany

The study of pressure using MRI has been previously limited to fluid contrast agents

comprising of gases1 or liquids

2-5. The most recent of these studies uses a polysaccharide gel

as the base of the contrast agent. Here we extend this work by producing a soft-solid contrast

agent by cross linking the polymer chains with salt. This could find applications in many

disciplines, particularly that of chemical engineering.

We have shown that the diffusion of water in polysaccharide gels remains relatively

unhindered5, 6

for a wide range of very high mesoscopic viscosities. In these gels, the addition

of salt produces a soft-solid in which water is nearly free to diffuse. By adding gas filled

microbubbles during production of these soft solids, MR pressure contrast is achievable. In

this work, a preliminary study is presented in which two soft-solids are tested, one degassed

control and one containing microbubbles to demonstrate the source of contrast.

Fig 1 demonstrates the sensitivity of this technique in a cylindrical soft solid

(Ø=25mm h=30mm) inside a capped syringe experiencing changes in pressure up to 1.4 bar

during a RARE7 sequence in which the RARE factor is set to the number of lines so as to

acquire a single shot image.

Fig. 1: a-d) Raw RARE data. The regions of high intensity around the samples are bulk water. The top row is a

control sample and the lower contains microbubbles. a) and b) are collected with no external pressure whilst c)

and d) are collected for an applied pressure of 1.4 bar. e-f) are the differences of the RARE images. The right

hand plot is the average signal intensity over three slices for the control and microbubble samples (top and

bottom respectively).

Contrast is shown to require the presence of microbubbles and excellent MR

sensitivity to pressure changes (160% signal change per bar) can be seen promising a range

of exciting new MR experiments in the near future.

We gratefully acknowledge E. Fukushima for providing some of the drive behind this

work, P. Morris for facilitating access to an MRI scanner and P. Gardner for supplying

vacuum equipment

References: 1. M. Bencsik and C. Ramanathan, MRI, 2001, 19, 379-383.

2. A. Alexander, T. Mccreery et al, Mag. Res. Med., 1996, 35, 801-806.

3. R. Dharmakumar, D. Plewes et al Phys Med Biol, 2005, 50, 4745-4762.

4. R. Morris, M. Bencsik, A. Vangala and Y. Perrie, MRI, 2007, 25, 509-512.

5. R. Morris, M. Bencsik et al. Under Review: J. Mag. Res., 2008.

6. N. Nestle, P. Galvosas et al, EENC 2000, University of Leipzig, 2000.

7. J. Hennig, A. Nauerth et al, Mag Reson Med, 1986, 3, 823-833.

1 1.05 1.1 1.15 1.24

5

6

7

8

9

10

11x 10

4

Pressure, bar

Me

an

Sig

na

l In

ten

sity

Solid Gel

Control

3% Microbubbles

3% Microbubbles

4

No Microbubbles

-2 0 2 4

x 104

0 1 2

x 105

2 -

a) c) e)

b) d) f)

36

Investigation of H2O2 decomposition in heterogeneous catalysts L. Buljubasich1 , Thomas Oehmichen2 , Leonid B. Datsevich2 , A. Jess2 , B. Blümich1 and S. Stapf 3

1Dept. of Macromolecular Chemistry, ITMC, RWTH Aachen, 52074 Aachen, Germany 2Dept. of Chemical Engineering, University of Bayreuth, 95447 Bayreuth, Germany 3Dept. of Technical Physics II, TU Ilmenau, PO Box 100565, 98684 Ilmenau, Germany

Heterogeneously catalysed reactions mostly take place in the presence of finely dispersed catalysts (i.e. metals such as Ni, Pt, Pd, …), these in turn are localized in materials of large internal surfaces, i.e. porous media. The reaction efficiency depends on parameters such as internal surface area; homogeneity of metal distribution; porosity and tortuosity of the pellet; transport of the reactants and products between the pellets (flow, diffusion) and inside the pellets (diffusion).

In most technically interesting reactions, gas occurs as one of the involved components. For not too small pores, this leads to the generation of gas or steam bubbles, which grow and eventually leave the pore. The additional dynamics introduced by bubbles can greatly enhance mass transport and reactor efficiency.

The decomposition of hydrogen peroxide represents a simple gas-forming reaction, H2O2 (liquid) → H2O (liquid) + ½ O2 (gas). While a pure hydrogen peroxide solution is rather stable, decomposition is catalyzed by the presence of heavy metals, and oxygen gas and decomposition heat are produced.

In this contribution we present results obtained monitoring this reaction in commercial Al2O3 porous catalyst particles, with different metals as catalytically active sites. The motivation in this study is twofold: to follow the decomposition by means of properties easily accessible with NMR methods, e.g. by monitoring the oxygen concentration and the effective diffusion coefficient of the liquid in the vicinity of the pellet or in a defined closed volume; and to visualize the heterogeneity of reaction inside the pellet. In both cases, the change of relaxation times and diffusion coefficients were observed in real time and were combined with imaging of the spatial distribution of these parameters. In the fluid phase surrounding the pellet, the dependence of the liquid’s T2 on parameters such as the oxygen concentration allows the monitoring of the state of the system during the reaction. At the same time, the production and motion of gas bubbles produces a random change in the velocities of the liquid around the pellet. The increase of the effective diffusion coefficient (Deff) therefore is indicative for the reaction rate in the system, as is shown in the figure. The dependence of T2 and Deff inside the pellet is shown to correlate with the localized distribution of metal and the distance from the outer pellet surface.

Figure: Deff measured with the gradients aligned in z-direction during the reaction, in the presence of an Al2O3 pellet doped with Cu. The gray continuous line represents the diffusion coefficient of bulk water.

1. L. Datsevich, Appl. Cat. A 247 (2003) 101. 2. L. Datsevich, Appl. Cat. A 262 (2004) 149.

37

Advances in the NMR/MRI studies of hydrogenation processes

Igor V. Koptyug

International Tomography Center, SB RAS, Novosibirsk, Russia

NMR/MRI studies of porous media are often sensitivity-limited due to a relatively low spin density of the nuclei detected (especially when gases are involved) and the short nuclear spin relaxation times. Therefore, such applications could significantly benefit from the implementation of signal enhancement techniques. One of the promising directions is the use of parahydrogen in hydrogenation of substrates with double or triple bonds. This approach can provide signal enhancement of several orders of magnitude, but until recently was possible only with homogeneous hydrogenations in solution. We have demonstrated recently [1] that parahydrogen-induced polarization (PHIP) can be generated in heterogeneous hydrogenation reactions as well. This significantly expands the area of possible applications of the PHIP technology and allows one to produce polarized hydrocarbon gases as well as catalyst-free polarized liquids. These polarized fluids can then be used for NMR/MRI studies in materials research, chemical engineering and catalysis [2,3], biomedical research in vivo, etc. At the same time, the observation of PHIP in heterogeneous catalytic reactions can be used to develop a highly sensitive tool for the detailed studies of the mechanisms and kinetics of heterogeneous catalytic processes which take place in porous supported catalysts. We have demonstrated successfully that both metal complexes immobilized on porous supports and supported metal clusters are able to produce PHIP effects, providing some new insight into the mechanisms of these reactions [1,4].

The PHIP approach can be integrated with the more conventional ongoing studies of the processes in porous catalysts and catalyst beds. The character of mass transport in a reactor can change dramatically under reactive conditions, making the studies of operating catalytic reactors imperative. We have performed dynamic MRI studies of catalytic hydrogenation of α-methylstyrene or 1-octene in a packed bed reactor with continuous or modulated reactant feed, demonstrating a complex dynamic behavior caused by the non-linear coupling of mass transport with the chemical reaction. Another issue of paramount importance in exothermic processes such as hydrogenation is that of heat transport. Based on the direct imaging of the solid phase (27Al MRI of the Pd/Al2O3 catalyst bed), we are developing an approach for the spatially resolved NMR thermometry of operating catalytic reactors. A pronounced temperature dependence of the 27Al NMR signal was used to reveal the variations of the temperature along the catalyst bed in the course of hydrogenation of propylene into propane.

The grants from RFBR (08-03-00661, 07-03-12147), RAS (5.1.1, 5.2.3), SB RAS (11), support of leading scientific schools (NSh-3604.2008.3), CRDF (RUC1-2581-NO04) and Russian Science Support Foundation are acknowledged.

References:

1. I.V. Koptyug, K.V. Kovtunov, S.R. Burt, M.S. Anwar, C. Hilty, S. Han, A. Pines, R.Z. Sagdeev, J. Amer. Chem. Soc. 129 (2007) 5580.

2. L.-S. Bouchard, K.V. Kovtunov, S.R. Burt, M.S. Anwar, I.V. Koptyug, R.Z. Sagdeev, A. Pines, Angew. Chem. Int. Ed. 46 (2007) 4064.

3. L.-S. Bouchard, S.R. Burt, M.S. Anwar, K.V. Kovtunov, I.V. Koptyug, A. Pines, Science 319 (2008) 442.

4. K.V. Kovtunov, I.E. Beck, V.I. Bukhtiyarov, I.V. Koptyug, Angew. Chem. Int. Ed. 47 (2008) 1492.

38

A Hybrid Spin Echo SPI method for Density Imaging in Porous Media

L. Li,a H. Hui,a B. Balcoma

aMRI Centre, Department of Physics, University of New Brunswick

A simple density image of fluid distribution in porous media, with no relaxation time

contrast, is remarkably difficult to achieve with conventional methods. Short-lived transverse relaxation times (T2) yield signal loss while multi exponential T1 and T2 yield variable signal attenuation in a simple spin echo image. Density imaging is however highly desirable in a wide variety of porous media imaging experiments.

The SPRITE class of MRI methods have proven to be robust and general in their ability to generate pure density images in porous media however the short encoding times required, with correspondingly high gradients and filter widths, and low flip angle RF pulses, yield sub-optimal SNR images.

90° 180° 180° 180° 180°

31 Echoes

ACQ

ACQACQ

ACQ

T2

*

T2

TE

Time

M0

90° 180° 180° 180° 180°

31 Echoes

Gradient

Gradient Fig. 1: The k space origin data point is acquired on the FID with a minimal evolution time. Subsequent k space data points are acquired from individually phase encoded spin echoes. Positive and negative k space data sets are combined for FT image reconstruction.

We have recently explored pure phase encode imaging with spin echo methods for high resolution thin film imaging. These methods have T2 contrast since the k space origin point is acquired from the first echo.

In the present work we modify the spin echo SPI acquisition to acquire the k space origin data point with a near zero evolution time from the FID following the 90 deg excitation pulse. Since k=0, no gradient is applied and the k=0 data point has a pure density weighting. Subsequent data points do suffer from T2 attenuation but the echo time for pure phase encoding reasonable size core plugs may be reduced to less than 500 usec with rapid gradient switching. T2 attenuation of the pure phase encoded echoes introduces a convolution to the subsequent density weighted image. While in principle deconvolution of the profiles generated is possible, in practice it is not required for most samples since the resolution loss which results is acceptable. T2 is no longer an uncontrolled contrast parameter, it is a blurring parameter.

The hybrid spin echo SPI method outlined features high sensitivity since it permits digital filtering set to the natural line width, with 90 and 180 degree pulses generating maximum transverse magnetization. Short echo times are possible with the gradient waveform mapping and correction methods outlined in a separate abstract.

Our principal goal is one dimensional profiling of fluids in porous media however these ideas may be extended to higher dimensionality, and will also permit magnetization preparation for variable contrast imaging.

39

Velocity maps measured in a single-shot using multi-echoes with

independent encoding

Andrea Amar, Federico Casanova, and Bernhard Blümich

Institut für Makromolekulare Chemie, RWTH-Aachen University, Germany

NMR velocimetry has proven to be a remarkably useful tool for non-invasive investigations of

fluid dynamics. Among the different approaches developed to spatially resolve velocity

distributions, one of the most powerful ones combines displacement encoding in the phase of the

NMR signal with the successive application of fast read-out imaging sequences like RARE or

EPI. Although such sequences have succeeded to produce reasonable velocity maps, they are

limited to the study of systems evolving sufficiently slowly so that displacement during the

acquisition of the complete k-space can be neglected. This particular limitation comes from the

fact that the velocity is encoded in a period prior to the k-space sampling. In this work we present

a new multi-echo sequence where the velocity is encoded and decoded before and after the

formation of each single echo. As the velocity information is refreshed for each echo, velocities

almost two orders of magnitude higher than accessible with conventional sequences can be

measured. Besides tolerating much higher velocities, the fact that the echoes are independently

encoded along the train, allows one to measure velocity maps along all three directions in a single

shot. The performance of the sequence was extensively studied by means of numerical

simulations, which allowed us to evaluate the efficiency of velocity compensation along read and

phase directions, as well as that of velocity encoding. Moreover, the simulations clearly show

that, as no phase needs to be preserved during the echo train, the sequence is highly immune to B1

inhomogeneities, an extremely critical source of distortions even in conventional birdcage

resonators.

The new sequence was successfully implemented to resolve the flow pattern in toluene

drops levitating in a water counter-flow. The good time resolution allowed us to study the effect

of a third component (acetone) diffusing from the continuous phase (water) into the drop. The

results provide also important evidence of the influence of acetone on the surfactants present in

the system.

Fig. 1. Pulse sequence proposed to

independently encode velocity for each

echo. It uses two pairs of velocity

encoding gradients: one before and one

after the echo (to rewind the phase shift

proportional to the velocity introduced

by the first one). Extra bipolar pulses

with opposite polarity are applied along

the phase direction to cancel the phase

proportional to displacement accumula-

ted after the first pair.

40

Improving Accuracy and Speed of NMR Flow Propagators

J. Mitchella, E. J. Fordham

b, D. A. Graf von der Schulenburg

a, D. J. Holland

a, M. L. Johns

a,

and L. F. Gladdena

aDepartment of Chemical Engineering, University of Cambridge, Pembroke Street, Cambridge,

CB2 3RA, UK. b

Schlumberger Cambridge Research, High Cross, Madingley Road, Cambridge, CB3 0HG, UK.

NMR flow propagators are an important method for analyzing fluid transport in porous

media. The flow propagator measurement provides a probability distribution of displacement ζ

as a function of observation time ∆ [1]. There are two distinct drawbacks to these measurements:

(a) spin relaxation impairs the quantitative nature of the data, and (b) the experiments can be

considerably time consuming since both the PFG gradient strength g and ∆ need to be varied

independently. We address these two limitations with separate solutions.

(a) We improve the accuracy of the measurement by characterizing the relaxation

properties of the system under flow using ζ-T2 [2] and T1-T2 [3] correlations. The ζ-T2

correlation provides T2 distributions as a function of displacement, and combined with the ratio

T1/T2 as a function of T2, the T1 relaxation weighting can be removed from the probability

distribution. The corrected probability distributions are seen to provide observed displacements

equal to the expected displacements for brine solution flowing through typical permeable

reservoir rocks [4], see Fig. 1. Results from sedimentary rocks at low magnetic fields will be

shown.

Fig. 1. Measured / expected mean displacement of brine solution flowing through porous

rocks as a function of observation time. <ζ > / <ζ >0 tends to 1 for large displacements.

(b) The speed of the propagator measurements can be increased using the rapid Difftrain

pulse sequence [5] that allows propagators with a range of observation times to be acquired

simultaneously. This technique has been shown to yield accurate displacement probability

distributions – equivalent to those from the conventional APGSTE sequence – for fluid flow in

permeable rocks [6]. This should allow variations in flow to be observed in time critical

experiments such as dissolution or deposition within porous media.

References:

1. J. Karger and W. Heink, J. Magn. Reson. 51 (1983) 1.

2. M. M. Britton, et al., J. Magn. Reson. 169 (2004) 203.

3. Y. Q. Song, et al., J. Magn. Reson. 154 (2002) 261.

4. J. Mitchell, et al., J. Magn. Reson. submitted.

5. J. P. Stamps, et al., J. Magn. Reson. 151 (2001) 28.

6. J. Mitchell, et al., J. Magn. Reson. doi:10.1016/j.jmr.2007.12.014.

41

Characterization of internal magnetic field in porous media H.Cho

1,2, S.Ryu

2, J.L.Ackerman

3, and Y.-Q.Song

2,3

1Memorial Sloan Kettering Cancer Center, New York, NY 10069

2Schlumberger Doll Research, Cambridge, MA, 02139

3The Athinoula A. Martinos Center for Functional and Structural Biomedical Imaging,

Massachusetts General Hospital, Charlestown, MA 02129

When a porous material is inserted in a uniform magnetic field, spatially varying fields

typically arise inside the pore space due to susceptibility contrasts between the solid matrix

and the surrounding fluid. Susceptibility contrast is present in many porous media of interest,

such as fluid filled rocks, cements, granular media, colloids and trabecular bones. The

presence of non-uniform internal fields often interferes with NMR relaxation and diffusion

measurements. Recently, the idea of utilizing internal gradients as a finger-print of pore

geometry for detailed pore structure information has been implemented and gave a new

insight upon quantifying the structure related properties of internal gradients in porous

media.[1,2] In this presentation, we further investigate and develop new NMR methods

utilizing susceptibility induced internal gradients.

Direct visualization of internal gradients in a 2D media : We utilize the decay of the

magnetization due to the spins diffusing in the internal magnetic field (DDIF) for the direct

experimental visualization of the internal gradients with microimaging in a 2D system

composed of uniform cylindrical capillary tubes. Interference patterns of the cross-terms

between the internal and applied gradient are also resolved along different pulsed field

gradient (PFG) orientations to extract the vector components of the internal gradient.

Excellent agreement with theoretical calculations is observed and shown in Figure 1.[3]

Measurement of magnetic field correlation in a 3D media : We demonstrate that the

structure factor of the granular and porous media can be approximated by the pair correlation

function of inhomogeneous internal magnetic field. Experiments and simulations are

performed on randomly packed bead system and the results are shown in Figure 2.[4]

Figure 1 (Left) : (A) shows DDIF decay rates obtained from imaging experiment and (B) shows the strength of

local gradient from theoretical calculations. (C) and (D) compare decay rates with corresponding gradient

strengths for various packing configurations along the labeled lines in (A) and (B)

Figure 2 (Right) : (A) shows numerical calculation of field correlation function (K) with different porosities of

randomly packed bead system. (B) shows initial slope of K as a function of surface to volume ratio obtained by

varying the porosity of the pack. (C) shows measured field correlation function along different applied external

gradient. (D) compares experimental measurement and calculations.

References 1.Song YQ et al., Nature 406,178 (2000). 2. Lisitza N et al., Phys. Rev. B, 65, 172406 (2002). 3. H.Cho et al., to

be submitted. 4. H.Cho et al., Phys.Rev.Lett. 100, 025501 (2008)

42

Intermolecular interactions reveal molecular compartmentation and dynamics in tissues and cells as studied in controlled porous glass using 1H MAS NMR Jin-Hong Chen, Rachael O’Connor, Penelope DeCarolis, Elliott R. Brill, Bernadette U. Laxa, Samuel Singer Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10065

Controlled porous glass (CPG) share many common physical features with tissue and are regularly used as a model system for tissue. Small molecules within CPG and tissues exhibit different dynamics than they do in water. However, the susceptibility effects in these systems lead to large NMR linewidth and low detection sensitivity, especially in a proton spectrum. Magic angle spinning (MAS) averages out the isotropic susceptibility contribution and has become a valuable tool to study the molecular compartmentation and dynamics in these systems.

Strong negative nuclear Overhauser effect (NOE) is observed under MAS between water and some but not all small molecules in tissues or cells, as opposed to the weak positive intermolecular NOE between small molecules in liquid state1. This selectivity effect may be a consequence of a co-existence of the involved molecule and water in a confined compartment within the cells. The confinement results in a much slower molecular dynamics. A CPG with a pore size of 75 Å (CPG-75) is used to mimic the possible small compartments in cells. 2D NOESY MAS NMR of CPG-75 shows that creatine has strong negative intermolecular NOEs (positive cross peaks relative to diagonal peaks) with the residual water in D2O and negative intramolecular NOEs between its own proton groups (Fig. 1a). This is in sharp contrast to the regular positive intramolecular NOEs and intermolecular NOEs with water when creatine is in the same solution but without CPG (Fig. 1b). The negative intra/intermolecular NOEs suggest the dynamics of creatine and water in the pores of CPG is significantly slower than that in a free liquid state. Fig. 1a is consistent with the spectrum obtained in skeletal muscle2. Therefore, the observation of negative intermolecular NOEs between water and small metabolites in tissue indicates that these small metabolites may exist in micro-compartments within tissue and cells.

This method of studying the intermolecular interaction between small molecules in porous medium may be generalized as an alternative approach for analysis of molecular dynamics and compartmentation of a wide variety of porous materials.

Fig. 1. 2D NOESY acquired with creatine dissolved in 99.9% D2O using MAS rate of 5000 Hz and mixing time 400 ms: (a) 13 µl solution in 26.3 mg CPG-75; (b) creatine solution without CPG. The 1D spectra are the slices taken at the dashed line in the 2D. Reference: 1. Macura, S, Ernst, R.R., Elucidation of cross relaxation in liquids by two-dimensional N.M.R. spectroscopy. Mol. Phys. 100, 135-147 (2002). 2. Chen J.H, et. al., Resolution of Creatine and Phosphocreatine 1H Signals in Isolated Human Skeletal Muscle using HR-MAS 1H NMR. Magn. Reson. Med. In print.

ppm

4567 ppm

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.5

7.0 a)

ppm

4567 ppm

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.5

7.0 b)

4.0 3.9

x64

43

Self-assembly and molecular diffusion of amphiphilic block

copolymers in mesopores

F. Grinberga, K. Ulrich

a, J. Kaerger

a, W. J. J. Stevens

b, V. Meynen

b, and P. Cool

b

aUniversity of Leipzig, Department of Physics, Linnéstr. 5, D-04103 Leipzig, Germany

bUniversity of Antwerpen, Department of Chemistry, CDE, Universiteitsplein 1, B-2610

Wilrijk, Belgium

A renewed explosion of interest to block copolymers (BCP) was caused by their

fascinating properties of self-assembly and spontaneous structure formation on the

mesoscopic length scale. During the last two decades, these properties were successfully

exploited in nanotechnology and molecular engineering. In particular, the BCPs found a

spreading use as templates in manufacturing of novel mesoporous materials with tailored

functionalities. However, the relationship of complex dynamic and structural properties of the

BCPs as well as the effects of confinements on molecular self-assembly remain only poorly

understood. One of the purposes of this work was a detailed study of the influence of super-

molecular organization on molecular transport properties of the BCPs. A specific focus was

on getting a deeper insight into effects of the nanoscaled confinements on molecular self-

assembly and self-diffusion as well.

Diffusion of the BCPs based on poly(ethyleneoxide) (PEO) and poly(propyleneoxide)

(PPO) in water was studied with the help of the Pulsed Field Gradient NMR. In mixtures with

water, these BCPs form four super-molecular structures: micellar, cubic, hexagonal, and

lamellar. For the first time the diffusivities in all these formations could be measured with the

same block copolymer and at the same temperature, Fig. 1. Crucial effects of self-assembly

on molecular transport properties were elucidated. Diffusion in the hexagonal and lamellar

mesophases was shown to be anisotropic. First (unique) diffusion study was performed also

with the PEO-PPO-PEO+water system confined in the mesoporous material (SBA-15).

Diffusion attenuation curves at ambient temperatures exhibited strong deviations from the

exponential behaviour, Fig. 2, and could be well described by the model of anisotropic

diffusion (solid curve). The results obtained were discussed in terms of the constraints

imposed by specific molecular ordering and pore walls on the diffusion mechanisms.

Acknowledgements. We thank Dr. P. Galvosas for development of the unique PFG-unit with

high-intensity magnetic field gradients enabling the measurements of ultra-slow diffusion.

V. M. acknowledges the FWO flanders for the research grant.

Fig. 2: Diffusion attenuation curves of the systems

PEO-PPO-PEO (P123) + water confined in the

mesoporous SBA-15.

Fig. 1: The diffusivities of PEO-PPO-PEO (P123

and P104) and PEO molecules in water as a

function of polymer concentration at 23 oC.

44

Molecular exchange between intra- and extracellular

compartments in a cell suspension

Ingrid Åslund, Agnieszka Nowacka, Markus Nilsson and Daniel Topgaard,

Physical Chemistry 1, Lund University, P.O.B. 124, SE-221 00 Lund, Sweden

Exchange time between different domains is important when doing NMR experiments

because depending on the time scale of the exchange the end results can look very different.

The exchange time can also say something about the state of the system of interest.

To measure the exchange time a new NMR-method has been tested on a yeast suspension.

The NMR-method is based on two ordinary Pulsed-Field-Gradient Spin-Echos with a mixing

time inbetween. It is a new protocol which is more sensitive and more time-efficient

compared to other methods of measuring exchange.

Two domains, with different diffusion abilities, were obtained by preparing a suspension

of baker's yeast. Here the intracellular water was used as the domain with slow diffusing

water and the extracellular water was regarded as the domain with fast diffusing water. A

verification of the possibility to detect differences in exchange time was done by changing

the measuring temperature.

The results show a clear exchange between the two domains. They also show that the

exchange time is dependent on temperature.

45

0 25 50 75 100 125 150 175

-2.0

-1.5

-1.0

-0.5

0.0

ln(I/I

0

)

k x 10-10

/m-2

s

= 200 ms

= 100 ms

= 50 ms

Diffusion-exchange studies of probe molecules through the

nanoporous walls of hollow polyelectrolyte capsules

R. P. Choudhurya,b

, D. Chakrabortya,b

, P. Galvosasc, M. Schönhoff

a

aInstitute of Physical Chemistry, University of Münster,

bInternational NRW Graduate School

of Chemistry, Münster, cFaculty of Physics and Earth Science, University of Leipzig

Polyelectrolyte multilayer (PEM) capsules formed by layer-by-layer assembly have

attracted interest as materials for controlled release and delivery applications. The time scale

of permeation of active molecules through the nanoporous wall, or their incorporation into it

is crucial for these applications. We study different types of probe molecules by Pulsed Field

Gradient (PFG)-NMR diffusion experiments at different diffusion times, , and demonstrate

here two different cases of polymeric and small probe molecules, respectively.

For the permeation of polyethylene oxide (PEO), PFG-NMR and diffusion-relaxation

correlation spectroscopy (DRCOSY) yield two fractions of polymer and imply the presence

of mobile PEO chains in the capsule interior. For the permeation rate, a scaling law of the

molecular weight dependence is extracted, which suggests a transition between two different

mechanisms of permeation as the molecular weight is increased. In both regimes, the

permeation time can be described by a scaling law ~ Nb, with b = 4/3 for short chains and b

= 1/3 for long chains. We discuss the mechanisms controlling these exponents, which clearly

differ from theoretical predictions for single chain translocation through a nanopore [1].

Fig. 1: PFG-NMR echo decays for different diffusion times for

polyethyleneoxide in capsule dispersion. The lines are fits by the

exchange model. The inset shows the encapsulated and the free site.

Small hydrophobic molecules such as phenol show an interesting regime with

monoexponential echo decays, but -dependent diffusion coefficients, which is identified as

a regime with T2bound << exch [2]. The probe molecules are partly adsorbed in the capsule

wall and thus are practically not observed. However, via their indirect influence on the

apparent diffusion coefficient of the free site, an analysis becomes feasible[3]: From diffusion

and additional relaxation data, the adsorbed fraction is extracted, and the exchange between

the bound and free site is analysed, resulting in exchange times of ~ 30 ms.

References:

1. see for example: J. L. A. Dubbeldam, A. Milchev, V.G. , T.A. Vilgis, arXiv:cond-mat/

0701664v2 (2007).

2. D. E. Woessner, Concepts Magn. Reson. 8 (1996) 397.

3. R. P. Choudhury and M. Schönhoff, J. Chem. Phys. 127 (2007) 234702.

46

2D NMR approaches to characterize water dynamics, exchange and

membrane permeability in biological tissue: towards imaging?

Henk Van As, Natalia Homan, Carel W. Windt, Harke Pera, Frank J. Vergeldt

Lab of Biophysics and Wageningen NMR Centre, Wageningen University, Wageningen, The Netherlands

A number of 2D (or even 3D) correlation spectroscopy methods are available to

characterize porous (bio)materials in terms of water dynamics in different compartments and exchange between these compartments. Correlated D-T2, flow-T2, T2-T2 exchange and even propagator resolved T2-T2 exchange experiments have been demonstrated (1). Each of these approaches has different limitations and advantages when applied to biological tissues and results in different information. This will be illustrated by results obtained in plant material.

T2-T2 exchange is very attractive. The intensity of cross peaks as a function of mixing time can be used to quantify exchange rates. Only exchange between water fractions with different T2s (different subcellular compartments) can be studied. However, the apparent exchange times are T1 weighted. The observed T1 values will be affected by exchange rates as well and actual exchange times are hard to obtain. The same problem is encountered if exchange between flowing water and surrounding non-flowing (cell) water is studied by measuring the amount of flowing (and stationary) water in a propagator as a function of Δ.

Exchange times are also dependent on the size of the compartments. In biological tissue membrane permeability is the parameter to characterize membrane transport. Membrane permeability can only be obtained from the exchange times if information on the surface-to-volume ratio (S/V) of the compartment and geometry of the tissue is available. To extract the information on S/V and membrane permeability D-T2 measurements are in use. S/V can be obtained from the slope of D as a function of Δ in the short diffusion time limit.

In the limit of long diffusion time (Δ>>R2/D0) D becomes dependent on cell-to-cell transport or the tissue permeability, P. P can be calculated from these results, but again assumptions on tissue geometry are needed (2).

The results of the 2D measurements can be analyzed in different ways: e.g. using 2D least square fitting methods using a discrete sum of a limited number of exponentials and ILT, an efficient 2D fitting procedure based on a fast Inverse Laplace Transformation (3). The latter is quite sensitive to signal-to-noise ratio, and can easily result in non-realistic cross peaks in e.g. a T2-T2 exchange experiment. The first method may be to simplistic for complex tissue material. An acceptable compromise may be to limit the kernel used in ILT by using the result of a high quality T2 measurement. Doing so, the 2D data set can be reduced and lower S/N is acceptable for the analysis. This can be used in both D-T2 and T2-T2 exchange experiments. Doing so, the 2D measurements can even be combined with fast imaging (e.g. RARE) with acceptable measurement times.

References:

1. K.E. Washburn, P.T. Callaghan, J. Magn. Reson. 186 (2007) 337. 2. H. Van As, J. Exp. Botany 58 (2007) 743. 3. L. Venkataramanan, Y.Q. Song, M.D. Hurlimann, IEEE Trans. Signal Process. 50 (2002)

1017.

47

R2-dispersion simulation of foam microstructure

S. Baetea,b and Y. De Deenea,b,

a Laboratory for Quantitative Nuclear Magnetic Resonance in Medicine and Biology, ECNURAD, Ghent University, Belgium, bMedisip-IBBT, Ghent University, Belgium

The spin-spin relaxation rate R2 (=1/T2) in hydrogel foams measured by use of a multi spin echo sequence is found to be dependent on the echo time spacing. This property, referred to as R2-dispersion, has several contributions, of which two are important in this study. In hydrogel foams, magnetic susceptibility differences between the gel and air phase give rise to an inhomogeneous magnetic field BIG(r) (fig. 1b). A contribution to R2-dispersion originates from molecular self-diffusion of water within these inhomogeneous magnetic fields. Surface relaxation of the diffusing water molecules at the gel-air interface is the second contribution. As a result R2-dispersion can be used to probe microstructures with surface relaxation and/or different magnetic susceptibilities. In hydrogel foams, correlations between the average air bubble size and R2-values are found [1]. Random walk diffusion is simulated to correlate the R2-dispersion with the foam microstructure (i.e. the mean air bubble radius and standard deviation of the air bubble radius) and foam composition properties (i.e. magnetic susceptibilities, diffusion coefficient and surface relaxivity). Simulations of R2-dispersion are in agreement with NMR measurements of a hydrogel foam. R2-dispersion curves are well parametrised by an arctangent function [2]. A set of syntheticic foams, with various foam microstructure and composition properties, is generated. Through the correlations between R2-dispersion parameters and microstructure properties the mean air bubble size can be derived from measured R2-dispersion curves. The R2-derived mean air bubble size of a hydrogel foam is in agreement with the bubble size measured with X-ray micro-CT. This illustrates the applicability of 1H R2-dispersion measurements for the macroscopic determination of the size of air bubbles in hydrogel foams and alveoli in lung tissue.

References: 1. S.H. Baete and Y. De Deene. Microstructural Analysis of Foam-Like Structures by Use of R2 Dispersion. Proc. Intl. Soc. Magn. Reson. Med. 15:37, 2007. 2. G. Borgia, R. Brown, P. Fantazzinni, Scaling of spin-echo amplitudes with frequency, diffusion coefficient, pore size, and susceptibility difference for the NMR of fluids in porous media and biological tissues, Phys. Rev. E. 51 (1995) 2104–2114.

Fig 1: (a) The porous structure of a

hydrogel foam is build up of thin liquid films and Plateau borders. Pressure differences caused by surface tension drive gas from smaller to larger bubbles. (b) Diffusion of water molecules in the gel phase of the foam is responsible for R2 dispersion. The transverse relaxation rate R2 is increased by interactions at the gel-air interface (the surface relaxation R2S) and by diffusion in the inhomogeneous magnetic field BIG(r) caused by susceptibility differences between gel and air (R2IG).

48

Water-Oil Capillary Pressure Measurement in Petroleum

Reservoir Core Plugs by MRI

Prisciliano F. de J. Canoa,b

, Bruce. J. Balcoma, Michael J. McAloon

a, Derrick P. Green

c,

Josh Dickc

aMRI Centre, Physics Department, University of New Brunswick, Canada

bCIIDIR IPN

Oaxaca, Mexico, cGreen Imaging Technologies, Inc., Canada

Capillary pressure (Pc) results from the interaction between a wetting fluid and a

nonwetting fluid in porous media. Capillary pressure curves are used in the petroleum

industry to predict the potential hydrocarbon recovery from a reservoir and to determine

many other basic petrophysical properties. The centrifuge Pc method, traditionally used in the

industry, involves rotating a core plug at different speeds and measuring the amount of water

expelled at equilibrium1. The traditional approach requires the use of several assumptions

related to the unknown fluid distribution in the core and gravity effects. It is a time

consuming measurement since only one point in the curve is obtained at each centrifuge

speed, making the overall test on the order of weeks.

By changing the centrifugation speed one changes the distribution of fluids in the core

plug. Spatially resolving the fluids distribution through imaging in the new MRI

measurement means we can analyze the force balance within a core plug as a function of

position and saturation (the Pc curve). Equilibrium at only one, or a few, rotational speeds

permits determination of the entire curve an order of magnitude faster than traditional tests.

The Double Half K-Space SPRITE method2 yields a direct saturation measurement

with facile calibration. It is ideally suited for Pc determination with our new method3. The

Centrifuge-Magnetic Resonance Imaging method requires only the use of a moderate

rotational speed centrifuge and an inexpensive low field permanent magnet. In this

presentation we emphasize two phase water-oil capillary pressure measurement. Capillary

pressure curves for primary drainage, imbibition, and secondary drainage were obtained in

core plugs with a range of permeabilities. By establishing an appropriate free water level,

both negative and positive capillary pressure data were obtained for imbibition and secondary

drainage. The imbibition curve is compared to the data obtained with the traditional

centrifuge method.

The tests described are being commercialized by Green Imaging Technologies with

the assistance of major oil companies, Chevron, Conoco Philips, Exxon Mobil and others.

The speed improvement of the new test, combined with greater precision of the results, given

the value of the measurement to the oil industry, may mean that the capillary pressure test

will become a routine core analysis measurement.

References:

1. Hassler G, and Brunner, Measurement of capillary pressures in small core samples, Trans.,

AIME 160 (1945) 114-123.

2. Mastikhin, I., Balcom, B., Prado, P., Kennedy C., SPRITE MRI with Prepared

Magnetization and Centric k-Space Sampling, J. Magn. Res. 136 (1999) 159–168.

3. Chen, Q. and Balcom, B. J., Measurement of Rock Core Capillary Pressure Curves Using a

Single-Speed Centrifuge and One Dimensional MRI, J. Chem. Phys. 122 (2005) 214720.

49

NMR T2 Inversion Method for a Block of CPMG Echo Trains

Boqin Sun, Mark Skalinski, and Keh-Jim Dunn1

Chevron Energy Technology Company

Abstract

Nuclear magnetic resonance logging has been routinely used to measure mineralogy

independent porosity, irreducible water saturation, and permeability of earth formation.

The acquired CPMG echo train data in NMR logging are first inverted by Laplace

transformation to yield NMR T2 distribution. Further interpretation based on T2

distribution allows us to derive additional important petrophysical properties of earth

formation such as pore size distribution, fluid types, and oil viscosity. The quality and

speed of the T2 inversion depend on the inversion methods. We have recently developed

fluid component decomposition method for fast NMR data inversion. This method

assumes that fluids, either in bulk forms or saturated in porous media, have certain

predetermined functional shapes, which can be predetermined T2 distributions. These

predetermined shapes can be Gaussian, B-spline, or any functions predetermined

experimentally or empirically. This approach significantly reduces the computation time

for NMR data inversion especially for multi-dimensional data sets from oil well

measurements, without sacrificing the smoothness and accuracy of the inverted

distributions. Such method has a new application as a solution constraint for a group of

NMR data with sequential well depths where spurious signals frequently result in

dissimilar T2 distributions for the same rock type. The successful implementation of this

method as a constraint of T2 components inverted from T2 echo trains at different depths

involves nontrivial matrix manipulations and allows us to use T2 distribution as a rock

type indicator.

1 Retired, now works as a consultant.

50

High performance shimming with permanent magnets

Ernesto Danieli, Juan Perlo, Federico Casanova, and Bernhard Blümich

Institut für Makromolekulare Chemie, RWTH-Aachen University, Germany

Portable and open NMR probes built from permanent magnets like the NMR-MOUSE offer

several advantages over conventional NMR systems. They can be used to study arbitrarily large

samples, are small, inexpensive, and robust [1]. However, the magnetic field generated by open

magnets is believed to be inherently inhomogeneous, precluding the acquisition of chemical-shift

resolved NMR spectra. In this work we break with this assumption, and demonstrate

experimentally that the field of an open magnet can be shimmed to high homogeneity in a large

volume external to the sensor [2].

To shim the highly inhomogeneous stray field of conventional open magnets, standard

shim coils must be discarded simply because of excessive requirements for the shim currents.

However, we demonstrate that the required shim fields can be generated by means of a single-

sided shim unit built from small permanent magnet blocks. The final shim unit geometry

designed to homogenize the field of a conventional U-shaped magnet uses two pairs of movable

magnet blocks with opposite polarizations placed in the gap of the main magnet. By suitable

displacement of the magnet block pairs with respect to the main magnet a total of eight shim

components are generated (x, y, z, x2, z

2, xy, xz, and yz). Using this approach, high-resolution

proton spectra were measured outside a portable magnet with a spectral resolution of 0.25 parts

per million (Fig. 1).

We also exploit this concept to shim the field of closed magnets, for example, that of

Halbach arrays. Two prototypes were optimized, a wide-bore magnet with an extended working

volume suitable for MRI and a narrow-bore one for high-resolution 1H spectroscopy.

Fig. 1. a) Solid line: spectrum of a water sample much larger than the sensitive volume. The FWMH of the line is

0.25 ppm. Dashed line: best spectrum obtained by ex situ spectroscopy using the nutation-echo method. The line

width is 8 ppm [3]. b-c) 1H NMR spectra of different liquid samples obtained within a measuring time of 1 min.

References:

1. B. Blümich, J. Perlo, and F. Casanova, Mobile Single-Sided NMR, Prog. Nucl. Magn. Reson.

Spectrosc. (2008) in press.

2. J. Perlo, F. Casanova, B. Blümich, Science 315 (2007) 1110.

3. J. Perlo, V. Demas, F. Casanova, C.A. Meriles, J. Reimer, A. Pines, B. Blümich, High-

resolution spectroscopy with a portable single-sided sensor, Science 308 (2005) 1279.

51

Investigation of molecular exchange using DEXSY with ultra-high pulsed field gradients

M. Gratz, P. Galvosas

Universität Leipzig, Fakultät für Physik und Geowissenschaften, Germany

With the present work we successfully combined the diffusion exchange spectroscopy (DESXY) as suggested in [1] with methods for the application of ultra-high pulsed field gradients as introduced in [2]. The DEXSY pulse sequence traces the molecular displacements in two subsequent intervals, separated by a mixing time τm. After the data processing with the inverse Laplace transformation several peaks may appear in a two-dimensional map characterizing the motional behavior of molecules in different sites. This method is applicable for several systems, e.g. for the exchange of dextran through polyelectrolyte multilayer hollow capsules, as used in [3].

Recently, we investigated a water-oil-water emulsion (moisturizer), which consists of three different phases with distinct diffusional characteristics. Small droplets of water are dispersed in bigger droplets of oil, which again are embedded in a water phase. Water molecules are able to move through the oil membrane from the inner to the outer phase and vice-verse, resulting in a change of their dynamic state.

Using a mixing time of only 10 ms, thus clearly demanding ultra-high pulsed field gradients, most of the signal is to be found in the diagonal peaks (see fig. 1), representing molecules with unchanged diffusion coefficient during τm. On the contrary, the off-diagonal peaks, which only exist for the two phases with the highest diffusion coefficients, represent molecules changing their diffusional properties, thus revealing exchange between two different sites of the sample. So far, we associate this exchange process to the water in the emulsion. The slowest phase, which is therefore assign to the only existing oil phase in the sample shows, as expected, no such behavior.

Fig. 1: Laplace transformed NMR signal map of moisturizer with a mixing time τm=10 ms and its projections showing the three components

References:

1. P. T. Callaghan, S. Godefroy and B. N. Ryland, Magn. Reson. Imaging, 21 (2003) 243–248.2. P. Galvosas, F. Stallmach, G. Seiffert, J. Kärger, U. Kaess and G. Majer, J. Magn. Reson.,

151 (2001) 260–268.3. Y. Qiao, P. Galvosas, T. Adalsteinsson, M. Schönhoff and P. T. Callaghan, J. Chem. Phys.,

122 (2005) 214912.52

POSTERS

53

Nuclear magnetic relaxation and diffusion of gas in partially

water saturated natural soil

R. P. Choudhurya and S. Stapf

b

aInstitute for Technical and Macromolecular Chemistry, RWTH Aachen, Germany, bDept. of

Technical Physics II, Institute of Physics, TU Ilmenau, Germany

Nuclear magnetic resonance (NMR) of fluids has proven to be a very sensitive tool for the characterization of porous systems. In general, nuclear magnetic relaxation and the restricted diffusion of water molecules obtained from water saturated porous systems can provide a wealth of information including the pore surface-area-to-volume ratio, the average pore size in systems and the visualization of fluid transport under flow. However this approach is usually limited to probing length scales ∼20 µm, because spin relaxation quenches the NMR signal before molecules can diffuse across longer distances. For this reason, liquid phase NMR has been unable to give information about the long-range properties of porous media, e.g. interconnectivity and tortuosity of the pore space1. In contrast, NMR techniques based on an imbibed inert gas has some advantages, e.g., gas diffusion coefficients are typically orders of magnitude larger than those of water, and the weak interaction of inert gas molecules with pore surface leads to long NMR lifetime. In addition, the diffusion coefficients of the gases can be altered by varying pressure, which significantly expands the range of distances that can be probed in comparison to liquid NMR.

NMR spectroscopy of noble gases has been used for years to probe physical properties of porous media. Particularly 129Xe NMR has become popular because of its exceptional sensitivity to local environmental effects. The inherent disadvantage of low spin density of Xe is overcome by high-pressure set-ups and hyperpolarizing techniques . Despite this, it has been observed that F NMR of the multinuclear fluorinated gases possess great potential for transport and material studies since they are relatively inert, of high density, and have short relaxation times that can be used in fast imaging and lead to image qualities not much inferior to those obtained with liquids .

129 2

19

3

Here we have studied the nuclear magnetic relaxation times and the diffusion properties of the fluorinated gases (SF6, CF4 and C4F8) in dry and partially water saturated natural soils to characterise the soil pore space as well as water distribution and transport in soils, which depends on a variety of parameters. The most important ones are 1) the pore structure, i.e., the pore size distribution and pore connectivity, 2) the wetting properties, and 3) the directional water supply by unsaturated water fluxes. The spin-spin relaxation times and the diffusion coefficients of those gases are changed in presence of the soils. This indicates that the dynamics of the gas molecules are affected by the pores of soils. To get more detailed information about the pore space and the long-range connectivity of soil, spin lattice relaxation and time dependent diffusion experiments at different gas pressures have been done and the observed results are compared to known soil properties4.

References:

1. M. D. Hurlimann, K. G. Helmer, L. L. Latour and C. H. Sotak, J. Magn. Reson. A 111 (1994) 169.

2. R. P Wang, M. S. Rosen, D. Candela and R. W. Mair, Magn. Reson. Imaging. 23 (2005) 203 3. D. O. Kuethe, R. Montaño and T. Pietraβ, J. Magn. Reson. 186 (2007) 243. 4. A. Pohlmeier. S. Haber-Pohlmeier, and S. Stapf, Vadouse Zones. J. submitted.

54

Time-dependent diffusion coefficient of fluids in partially filled porous media

Carlos Mattea, and Siegfried Stapf.

Department of Technical Physics II, TU Ilmenau, Germany.

Diffusion of water and organic solvents partially filling porous media is measured as a function of time in the stray magnetic field of a single side NMR_MOUSE device. The stimulated spin echo sequence [1] with constant magnetic field gradient (in the order of 10 T/m) is used. The liquids are confined in porous silica glasses with mean nominal pore sizes of 1 µm and 10 µm, and porosity of 43% and 36%, respectively.

In completely filled pore spaces, the transport properties are predominantly affected by the geometrical restrictions, i.e., the tortuosity. The time-dependent diffusion coefficient D(t) then exhibits a reduction from the molecular diffusion coefficient D0 in bulk when increasing the diffusion time. When the porous media are non-saturated, the effective diffusion coefficient is modified due to the presence of the vapor phase and the reduction in the volume of the liquid fraction. Exchange between liquid and its vapor phase can produce an enhancement in the diffusion process [2]. But in the general case, D(t) has opposite tendencies upon reducing the filling fraction (enhancement or reduction) depending of several parameters like nature of the solvent, pore size or pore structure [3,4]. A general two-phase exchange model recently published (see Ref. [3]) accounts for these features. At the diffusion times used in those studies (200 and 300 milliseconds), fast molecular exchange between the liquid and vapor phase was observed in the case of nanometer length-scale confinement. However, the fast exchange regime is not valid anymore when the pore size is in the order of micrometers. This is evidenced in the attenuation of the stimulated echo versus the square of the wave number which is not exponential. That deviation depends on the mean residence time of molecules in the liquid phase.

In the present contribution the diffusion studies are performed in partially filled porous media for different diffusion times. The length scale of the pore diameter is extended to the order of tenths of micrometers. The molecular exchange model of reference [3] for the vapor and liquid phases is applied. It is found that the average residence time in the liquid phase at low filling factors is in the order of milliseconds for porous samples with ten micrometers pore size. The influence of spin-lattice relaxation, T1, at long diffusion times is also discussed in particular at low filling factors.

References:1. M. Hürlimann, J. Mag. Res. 148 (2001) 367. 2. F. D’Orazio, S. Bhattacharja and W. Halperin, Phys. Rev. Lett. 63 (1989) 43. 3. I. Ardelean, C. Mattea, G. Farrher, S. Wonorahardjo and R. Kimmich, J. Chem. Phys. 119

(2003) 10358, I. Ardelean, G. Farrher, C. Mattea and R. Kimmich, J. Chem. Phys. 120(2004) 9809.

4. G. Farrher, I. Ardelean and R. Kimmich, Mag. Res. Imag. 25 (2007) 453.

55

Poly(dimethyl siloxane) films adsorbed in porous media

S. Ayalur-Karunakarana, S. Stapf

b

aInstitute for Technical and Macromolecular Chemistry, RWTH Aachen, Germany,

bDept. of

Technical Physics II, Institute of Physics, TU Ilmenau, Germany

The dynamics of polymer thin (i.e. nm-scale) films is becoming of growing interest

both from an experimental as well as theoretical point of view; its investigation with NMR

methods is still a rather new field and usually requires the existence of a large surface area as

is given in porous media. Under these circumstances, the actual spatial distribution and

molecular conformation of polymers at very small amounts remains an open question; the

assumption of equally thick films must break down at a particular amount of polymer that

depends on the interaction parameters with the surface, but certainly below the value required

for generating a complete coverage (monolayer). In the present study, this aspect is addressed

by investigating the molecular dynamics of a linear flexible polymer, poly(dimethyl siloxane)

(PDMS), confined at amounts corresponding to a few to less than one monolayer inside

microporous alumina membranes of well-defined pore shapes using NMR techniques.

Employing Fast Field Cycling (FFC) NMR provides information about motion in the kHz to

MHz range of frequencies which can be compared with the established models observed for

bulk1 and completely confined polymer chains

2 while transverse relaxation and Double

Quantum (DQ) NMR have been used to study slow dynamic processes and ordering.

Using FFC, is can be shown that the dynamics gradually slows down in films of all

thicknesses, while there is a qualitative deviation from bulk behavior for the lowest polymer

coverages, equivalent to one monolayer or less. The effect of temperature on relaxation is

likewise less pronounced for decreasing polymer amounts. Transverse relaxation decays of

these systems were found to be multiexponential in all cases, unlike the much simpler

behaviour of the bulk melts; data are discussed in terms of Inverse Laplace Transforms of the

decay curves and were compared to a simplified biexponential fit.

103

104

105

106

107

108

10-2

10-1

100

298 K

PDMS (Mw=30000 g/mol) bulkThin films in 200 nm pores

4.50 nm layer 2.10 nm layer 1.25 nm layer

ν [Hz]

T1

/m

s

1 2 3 4 5

1

10

100

T2

Long

T2

Short

T2,e

ff / m

s

Layer Thickness / nm

PDMS Thin layers

1 2 3 4 5200

300

400

500

600

700

800

900

PDMS Thin layers Residual Dipolar

coupling constants

RD

C / H

z

Layer Thickness / nm

Fig. 1: Relaxation behavior of PDMS (Mw=30000 g/mol) films inside alumina pores of size 200 nm. (Left)

FFC relaxation profile, (center) T2,eff times and (right) RDC constant of films with varying nominal thickness.

By applying DQ NMR, the existence of dynamic heterogeneities in these films is

confirmed, and the values of the Residual Dipolar Coupling (RDC) constants indicate an

increase of order with decreasing nominal film thickness. PDMS of different molecular

weights has been studied in two different geometries and was compared to other types of

polymers, allowing to draw conclusions leading to a tentative pictorial model of the actual

spatial distribution of polymer molecules on internal surfaces of porous media.

References:

1. P.J. Rouse, J. Chem. Phys. 21 (1972) 1272.

2. P.G. de Gennes, J. Chem. Phys. 55, (1971) 572.

56

Fast Field Cycling Relaxometry: From Saturated to Unsaturated

Natural Soils

S. Haber-Pohlmeier

a, S. Stapf

b

aDepartment of Macromolecular Chemistry, RWTH-Aachen University, Germany,

bDepartment of Technical Physics II, TU Ilmenau, Germany

Soils as natural porous media exhibit the important ability of water retention and

transport. Both properties are mainly controlled by the pore size distribution, which is

furthermore related to NMR relaxation times of water molecules. In this study the

longitudinal relaxation time T1 is determined by fast field

cycling relaxometry (FFC) at low fields corresponding to

Larmor frequencies between 5 kHz and 20 MHz.

In this work1 we investigated three natural soils (Sand FH31,

Kaldenkirchen and Merzenhausen) which cover a broad range

with respect to their texture, i.e. from fine sand to silt loam. In

addition we decreased the water contents from saturation of

100% down to 6% corresponding to naturally occurring

conditions.

The results of T1- relaxation time distributions are obtained

by the analysis of the T1 relaxation curves by means of the

inverse Laplace transformation2,3

. For example Kaldenkirchen

soil shows relatively broad bimodal distribution functions D(T1)

in contrast to sand FH31 with an narrower unimodal behavior

(Fig.1,see lower case). Using the Brownstein-Tarr model4 for

NMR relaxation in porous media we get first information about

the pore size distribution which reflects the T1 relaxation

behavior scaled by the surface relaxivity. In fine sand nearly all

water is present in macropores with characteristic T1 of about

1000 ms. The opposite is found for Merzenhausen silt-loam, i.e. T1 is about 2 ms

corresponding to the major water fraction present in micropores. Kaldenkirchen with more

than 10 % of silt and clay represents an intermediate case.

With decreasing water content we observed a considerable decrease of T1 relaxation by

factors of five to ten (e.g. fine sand in Fig. 1).

Independent of the exact texture of the soils a

general trend is conceivable, where first macro-

pores desiccate leading to thin water films and

residual water at grain contact points (Fig. 2).

References: 1. A.Pohlmeier, S. Haber-Pohlmeier and S. Stapf, Vadose Zone Journal (2008), submitted.

2. S.W. Provencher, Comp. Phys. Comm. 27 (1982) 213.

3. Y.-Q. Song, L. Venkataramanan, M.D. Hürlimann, M.Flaum. P.Frulla and C. Straley,

J. Magn. Res. 154 (2002) 261.

4. K.R. Brownstein and C.E. Tarr, J. Magn. Res. 26 (1977) 17.

10 100 1000

0

1

2

3

4

5

rel. water content

100 %

50 %

23 %

12 %

6 %

D(T

1),

a.u

. sh

ifte

d

T1 (ms)

Fig. 1: T1 distribution func-

tions of fine sand at a Larmor

frequency of 20 MHz.

Fig. 2: Model for desiccation of fine sand.

57

Low Field NMR of Water in Soils. A Case Study.

Oscar Elías Sucre1, Federico Casanova

1, Andreas Pohlmeier

2, and Bernhard Blümich

1

1ITMC, RWTH Aachen University, Aachen, Germany

2Forschungszentrum Jülich, Jülich, Germany

The evolution of water in soils is a physical phenomenum of importance in soil science and

climatology. While most investigations of soil moisture are performed inside large

superconducting magnets in the laboratory, this eventually needs to be studied by mobile

NMR equipment in the field. As a part of a DFG-funded interdisciplinary project (TR32), this

work establishes preliminary results from the use of mobile NMR to measure moisture in

soils. To demonstrate the ability of the NMR technique to follow the drying process of water

in soils, daily moisture measurements were performed with a mobile NMR endoscope on

three different types of soil (silt, sand, and a natural soil sample). The soils were packed in

columns approximately 1 meter high. The NMR measurements were cross-validated by

repetitive measurements of the mass drift due to the drying process. The NMR-endoscope

exhibits a cylindrical geometry incorporating the principle of the u-shaped NMR-MOUSE. It

could be raised and lowered inside a plastic tube in the soil column similar to a wire-line

logging tool. For the purpose of modelling, a simulation of the water evolution (described by

the Darcy Equation) was executed with the Hydrus1D Program for comparison with the

experimental results. The optimization of the endoscope, the measurements, and the

simulations are discussed as well as, the statistical uncertainty and the noise shielding.

58

Monitoring Water Fluxes in Porous Media by Magnetic Reso-

nance Imaging using D2O as a tracer

A. Pohlmeier, D. van Dusschoten, L. Weihermüller, H. Vereecken, and U. Schurr,

ICG, Forschungszentrum Jülich, 52425 Jülich, Germany

Soil functions like water and nutrient supply for food production are strongly con-trolled by water fluxes in these unsaturated natural porous media. As well known, MRI is a very powerful method for non invasive monitoring of such flow processes. But for that pur-pose the usage of tracer substances is essential since flow velocities are relatively slow (i.e << mm/s). In the past most prominent tracers were paramagnetic ions (1,2). However, their major drawback is possible sorption at the solid-liquid interface which may cause considera-bly slower, apparent transport velocities than the liquid flow that is of interest. From this per-spective D2O seems to be an ideal tracer due to its high chemical similarity to H2O. In this study (3) we first investigated the behaviour of D2O as tracer in an unsaturated, heterogeneous model soil composed of fine sand and a built-in cylindrical obstacle in order to force the flux around it. Water was purged in steady state through the column at a constant water potential of 39 cm (corresponding to a negative pressure of 3.8 kPa) applied at the lower boundary. Tracer pulses of 55% (w/w) D2O and for comparison of 0.01 M NiCl2 were repeatedly applied with infiltration water. The motion of the tracer plumes were monitored by MRI using a fast spin echo sequence over a period of 20 minutes with mean pore water ve-locities between 0.21 and 0.82 cm min-1. The most important result is that the motion of the D2O-pulse is not retarded and moves with the water flux combined with smearing by disper-sion. Furthermore, the motion of the Ni2+ tracer as reference under known, non-adsorbing conditions (pH = 5) corresponds to that of D2O. In order to validate the experimental results we compared the observed motion with an inde-pendent simulation based on the Richards and van-Genuchten equations for water flux in unsaturated porous media and the convection-dispersion equa-tion for tracer transport as implemented in the model HYDRUS2D (4). Fig. 1 compares the posi-tion of the tracer cloud 7 minutes after injection with the corresponding simulated scenario. The agreement proves that D2O really behaves as con-servative tracer and thus opens up a new field of applications.

References:

1. K.-H. Herrmann; A. Pohlmeier, S. Wiese, N.J. Shah, O. Nitzsche, and H. Vereecken, J. Env. Qual. 31, 506-514 (2002)

2 H. van As and D. van Dusschoten, Geoderma 80: 389-403 3. A. Pohlmeier, D. van Dusschoten, L. Weihermüller, H. Vereecken, and U. Schurr, Magn.

Res. Imag. (2008), submitted. 4. J. Šimunek J, M. Sejna, and M.T. van Genuchten MT. The HYDRUS-2D software pack-

age, Version 2.0, IGWMC - 1999

Fig. 1. Left: D2O tracer cloud 7 minutes after injection, right: simulated tracer cloud for the same time point

59

Molecular motion and glass transition for ethylene glycol

adsorbed in zeolites

Oezlen F. Erdema,b

, Pavel Sedykha, Dieter Michel

a,

aFakultät für Physik und Geowissenschaften, Universität Leipzig, Linnéstrasse 5,

04103 Leipzig, Germany, michel@physik.uni-leipzig.de, bMax-Planck-Institut für Bioanorganische Chemie, Stiftstraße 34-36, 45470 Mülheim an der

Ruhr, Germany

Structure and mobility of molecules adsorbed in restricted geometry like zeolites has

been widely investigated by means of NMR spectroscopy [1-3] by using rare nuclei (like e.g. 13

C and 15

N), where a sufficient spectral resolution could be achieved. In case of proton NMR

spectroscopy, however, the applicability of NMR spectroscopy to molecules adsorbed on

surfaces is often limited due to the poor resolution. A notable enhancement in resolution of

the spectra may be achieved when the susceptibility broadening is averaged out by the

application of MAS techniques. In all these studies the definite preparation of the MAS

samples in vacuo plays a decisive role. It will be shown how these special measurements

could successfully be realized.

Furthermore, we will discuss how the information on thermal mobility derived from

the side-band analysis in 1H MAS NMR experiments can be combined with the study of

proton spin-lattice relaxation over a wider temperature range. The results obtained will be

also compared with of broadband dielectric measurements. The results will be compared with

extensive dielectric studies by Kremer et al. [4-5] on comparable systems. The main question

is whether the adsorbed species show a so called single-molecule behavior characterized by

an Arrhenius type temperature dependence of the respective dielectric relaxation times. In

contrast, a Vogel-Fulcher-Tammann (VFT) type temperature dependence of the (mean)

dielectric relaxation rate points to collective motions and is typical for the appearance of a

glass-transition. We will discuss how this transition depends on the filling factor of the

zeolites, on the geometry of the internal voids in our porous materials and the competition

between molecule-to-molecule and molecule-to-surface-interactions. The importance of the

choice of the matrix system will be shown for the case of the NaX zeolites with well-defined

dimensions of the supercages. To get further insights in the complex dynamics of confined

molecules, 2H NMR line shape and spin-lattice relaxation measurements over a wide range of

temperatures were also included.

References:

1. G. Engelhardt, D. Michel, High-Resolution Solid-States NMR of Silicates and Zeolites,

John Wiley & Sons: Chichester, 1987.

2. U. Schwerk, D. Michel, M. Pruski, J. Magn. Reson. 119 (1996) 157.

3. J. Roland, D. Michel, Magn. Reson. Chem. 38 (2000) 587.

4. F. Kremer, A. Huwe, M. Arndt, P. Behrens, W. Schwieger, J. Phys. Cond. Mat. 11 (1999)

A175.

5. A. Huwe, F. Kremer, J. Kärger, P. Behrens, W. Schwieger, G. Ihlein, Ö. Weiss, F.

Schueth, J. Mol. Liquids 86 (2000) 173.

60

Studies of barium titanate embedded into mesoporous sieves and

isolated barium titanate fine particles

P. Sedykha, J. Haase

a, D. Michel

a, E. V. Charnaya

b

aUniversity of Leipzig, Faculty of Physics and Geosciences, Linnéstrasse 5, 04103 Leipzig,

Germany bSt. Petersburg State University, Faculty of Physics, Petrodvorets, St. Petersburg 198504,

Russia

Phase transition temperatures, electrical polarization, coercive field, and other

properties potentially depend on the particle size. In order to study size effects in barium

titanate (BaTiO3) it was embedded into MCM-41 mesoporous materials and the behaviour of

the samples was studied in comparison to the bulk barium titanate and very fine BaTiO3

powders. The MCM materials used possess effective diameters of channels of about 3.7 nm.

The filling of the small pores was controlled by nitrogen adsorption isotherm measurements

on empty mesoporous sieves and those filled with BaTiO3. The ultrafine BaTiO3 powders

were prepared from a monomeric metallo-organic precursor through combined-solid state

polymerization and pyrolysis (CPP method) [1]. The particle diameter varied in the range

between 15 and 155 nm and the powders were carefully characterized by FT-Raman, XRD,

and other methods.

Dielectric measurements were carried out for the MCM-41 molecular sieves filled

with barium titanate. In contrast to the large dielectric constants (d. c.) measured for the bulk

BaTiO3 in the tetragonal phase, the d. c. values of the BaTiO3/MCM-41 materials practically

did not differ from those measured for the empty MCM-41 rods. The behavior shows that the

materials do not have the ferroelectric properties. To prove this conclusion further

measurements on SBA 15 samples are in progress.

The conclusion derived from the MCM-41 samples is consistent with the critical

diameters estimated for very fine BaTiO3 powders. The influence of the grain size on

properties of the ultrafine barium titanate powders was also investigated by 137

Ba NMR.

NMR spectra were measured for central transitions. The line shape was calculated assuming a

dominant influence of the quadrupole coupling and the validity of the second order

perturbation theory. These suggestions were verified by measuring the spectra at different

magnetic fields (9 and 17.6 T). Temperature dependencies were measured within a large

temperature interval in the ferroelectric tetragonal phase including also the 1st order phase

transition from the tetragonal to the cubic phase. The complex NMR lines were deconvoluted

into two lines [2] which revealed two different structures for the fine particles. The line shape

parameters estimated for both of the contributions will be measured over a broad range of

temperatures including the phase transition from the cubic to the tetragonal phase for the bulk

material. This analysis allowed us to derive detailed conclusions about confinement effects.

The results will be also analyzed in comparison to the results obtained in previous ESR

measurements on nanoparticles [3].

References:

1. H.-J. Gläsel, E. Hartmann, D. Hirsch, R. Böttcher, C. Klimm, D. Michel, H.-C.

Semmelhack, J. Hormes, H. Rumpf, J.Mater.Sci. 34 (1999) 2319.

2. P. Sedykh, J. Haase, D. Michel, E. V. Charnaya, Ferroelectrics in press, May 26, 2008.

3. R. Boettcher, C. Klimm, D. Michel, H.-C. Semmelhack, G. Völkel, H.-J. Gläsel, E.

Hartmann, Phys. Rev. B. 62 (2000) 2085.

61

Mixture diffusion in porous media studied by MAS PFG NMR

M. Fernández , J. Kärger, D. Freude

Abteilung Grenzflächenphysik, Universität Leipzig, Germany

One novel combination of MAS NMR and PFG NMR techniques was applied to

study the diffusion properties of molecular mixtures absorbed in nanoporous materials. MAS

PFG NMR provides high resolved spectra and allows the performance of selective diffusion

measurements in mixtures [1].

Diffusivity in binary mixtures of n-butane and isobutane adsorbed in zeolite MFI

silicalite-1 were studied in dependence of temperature and partial concentration [2].

The diffusivity of both n-butane and isobutane absorbed in MFI differ in more than one order

of magnitude. The diffusion coefficient of n-butane in mixture presents a strong dependence

with the partial concentration, see Fig. 1. It was found that n-butane diffusivity decreases

with increasing amount of isobutane, and presents one strong drop or discontinuity at about

0.5 n-butane molecules per channel intersection. Further on, with decreasing content of n-

butane (increasing content of isobutane) the diffusivity of n-butane drops exponentially, with

the value extrapolated to zero loading of n-butane coinciding with the pure isobutane

diffusivity. Molecular simulations are in good agreement with these experimental results.

CBMC simulations show that n-butane can locate along either straight or zig-zag channels of

MFI and isobutane locates preferentially at the intersections. Molecular simulation has shown

that at 0.5 isobutane molecule per intersection, the molecular traffic of n-butane is brought to

a virtual stand-still because of blocking of the intersection sites by isobutane molecules.

0.0 0.2 0.4 0.6 0.8 1.0 10

−12

10−11

10−10

10−9

10−8

r / n-butane molecules per intersction

T = 363 K

n-butane

isobutane

D /

m2 s

−1

Diffusion in liquid mixtures of acetone with different alkanes in porous glasses (4 nm

and 10 nm pore diameter) was also studied [3]. In the narrow pores one dependence of

acetone diffusivity with the odd or even number of carbons in the alkane chain was found [4].

This oscillation effect in the acetone diffusion coefficient by increasing n-alkane length

depends on the intermolecular interaction between the different alkanes and suggests the

formation of complexes under strong confinement.

References:

1. A. Pampel, M. Fernandez, D. Freude, J. Kärger, Chem. Phys. Lett. 407 (2005) 53.

2. M. Fernandez, J. Kärger, D. Freude, A. Pampel, J.M. van Baten, R. Krishna, Microporous

Mesoporous Mater. 105 (2007) 124.

3. M. Fernandez, A. Pampel, R. Takahashi, S. Sato, D. Freude, J. Kärger, Phys. Chem. Chem. Phys.

submitted.

4. R. Takahashi, S. Sato, T. Sodesawa, T. Ikeda, Phys. Chem. Chem. Phys. 5 (2003) 2476.

Fig. 1 Experimental data on self-diffusion

coefficients of nC4 in nC4/iC4 mixtures in

MFI at 363 K as a function of the loading of

nC4 in the mixture. Also shown is the self-

diffusivity of pure iC4.

62

MAS PFG NMR diffusion studies of liquid crystals and liquid

mixtures confined in porous glasses

E. E. Romanova, F. Grinberg, J. Kärger, D. Freude

Abteilung Grenzflächenphysik, Universität Leipzig, Germany

The pulsed-field gradient nuclear magnetic resonance (PFG NMR) technique is a direct and

non-invasive method for measuring self-diffusion coefficients. It has proven to be an indispensable

tool of studying isotropic liquids [1] and molecules in confined geometries [2]. However, if dipole-

dipole interactions are not completely averaged out by molecular motions, the NMR signals are

very broad, and the free induction signals decays is very fast [3]. This limits the time available for

the application of gradient pulses [4] and makes impossible measuring diffusion in liquid crystals

without special line-narrowing techniques [5]. In the case of liquid mixtures confined in nanoporous

materials [6] dipolar broadening deteriorates the resolution in the chemical shift scale necessary for

selective diffusion measurements of mixture components. This kind of problems can be overcome

by combining magic-angle spinning (MAS) with pulsed field gradients. This relatively new

technique is referred to as MAS PFG NMR, see e. g. [7]. MAS implies the orientation of the

spinning axes with respect to the external magnetic field at the angle of θm ≈ 54.7°. In comparison

with other techniques [5] this type of measurements has considerable advantages. First, the

increased resolution in the ppm scale permits one to observe separately each individual group with

identical electronic surroundings. Second, the longer transverse relaxation time upon MAS allows

for sufficient time for the application of pulsed magnetic field gradients.

Dynamic properties of the confined nematic liquid crystal 5CB has been studied by several

techniques including the Dipolar Correlation Effect [8] and Field Cycling Relaxometry [9, 10].

However, no NMR diffusion measurements were performed so far. In this work we apply for the

first time the MAS PFG NMR technique to liquid crystals. The nematic liquid crystal 5CB confined

in Bioran glass was studied. It was shown, that the diffusion coefficient of 5CB confined in Bioran

glasses exhibit an Arrhenius-like temperature dependence with an apparent activation energy Ea

iso =

26.7±2.7 kJ mol−1

close to that in the isotropic phase. A minor discontinuity at the phase transition

was observed in agreement with the results of other techniques [8-10]. The measured diffusivities

however do not differ significantly from the values in bulk 5CB. It is shown that diffusion studies

of confined liquid crystals permit one to get a deeper insight into aspects of anisotropic molecular

interactions and the effects of confinements.

Measurements of methanol / cyclohexane-mixtures in porous glass with 7.5 nm diameter

(CPG7.5) yield the self-diffusion coefficients for both compounds. A comparison with the values

obtained for single-component adsorption shows that the diffusivity of methanol is by factor of 2

smaller in the confined binary mixture than in the system methanol + CPG7.5. At the same time

cyclohexane has the same self-diffusion coefficient in the confined mixture and in the single

component system cyclohexane + CPG7.5.

References:

1. Price, W.S., Concepts in Magnetic Resonance, 1997. 9(5): p. 299-336.

2. Kärger, J. and D.M. Ruthven, 1992, New York: Wiley & Sons.

3. Romanova, E.E., et al., Solid State Nucl. Magn. Reson., 2008. in print.

4. Kruger, G.J., Physics Reports-Review Section of Physics Letters, 1982. 82(4): p. 229-269.

5. Dvinskikh, S.V. and I. Furo,. Journal of Chemical Physics, 2001. 115(4): p. 1946-1950.

6. Fernandez, M., et al., Phys. Chem. Chem. Phys., 2008. submitted.

7. Pampel, A., et al., Microporous and Mesoporous Materials, 2006. 90(1-3): p. 271-277.

8. Grinberg, F., et al., Journal of Chemical Physics, 1996. 105(21): p. 9657-9665.

9. Grinberg, F.,. Magnetic Resonance Imaging, 2007. 25(4): p. 485-488.

10. Grinberg, F., in Diffusion Fundamentals ΙΙ,. 2007.

63

Diffusion in mesopores during melting and freezing processes

R. Valiullin, M. Dvoyashkin, A. Khokhlov, J. Kärger

Department of Interface Physics, University of Leipzig, Leipzig, Germany

Phase transitions of fluids confined to mesopores often exhibit pore size-dependent

shifts in the transitions. Moreover, transition pairs, like freezing and melting considered in

this work, in most cases do not coincide with each other giving thus rise to a hysteresis

phenomenon. The microscopic mechanisms leading to such behavior are still under

discussion, and additional information about the growth process of a new phase along both

transitions may appreciably contribute to a further understanding of this phenomenon.

Molecular diffusivity is a sensitive probe of the geometrical characteristics of the space in

which it takes place. Melting and freezing processes, occurring in pores, modify the pore-

space characteristics, in turn modifying the diffusion behavior of the liquid molecules. Hence,

it may be used to monitor the character of the structural changes during these transitions.

In this work, we present the study of freezing and melting behavior of nitrobenzene in

mesoporous silicon (PS) with one-dimensional channels with an artificially superimposed

variation of the pore size along the pore axis. The combined experiments using NMR

cryoporometry [1] and pulsed field gradient NMR allowed us to correlate the amount of

frozen nitrobenzene in the pores and the molecular diffusivities. In mesoporous silicon

prepared to provide structural correlations [2], the diffusivities measured as a function of

temperature clearly shown the effect of the correlated distribution of pores with different pore

sizes. These findings have been further used to understand and analyze the data for material

with a random distribution of the pores. Most remarkably, there is a well-pronounced

difference in the diffusivities measured along the heating and cooling branches, revealing

different distributions of the frozen domains during melting and freezing. It is also discussed

that the used experimental procedure may reveal very important information on the

interconnectivity of the pore space, alternatively to the approach suggested in [3].

References:

1. J. H. Strange, M. Rahman, E. G. Smith, Phys. Rev. Lett. 71 (1993) 3589.

2. A. Khokhlov, R. Valiullin, J. Kärger, F. Steinbach, A. Feldhoff, New J. Phys. 9 (2007)

272.

3. K. E. Washburn, P. T. Callaghan, Phys. Rev. Lett. 97 (2006) 175502.

64

Probing the meso- and microstructural heterogeneity of low dense

networks with dendrimer probes and NMR/MRI diffusometry

Gert-Jan W. Goudappel

a, Magnus Nydén

b, John P.M. van Duynhoven

a

aUnilever Food and Health Research Centre, Unilever R&D, PO Box 114, 3130 AC

Vlaardingen, The Netherlands, b

Applied Surface Chemistry, Chalmers University of

Technology, SE-412 96 Göteborg, Sweden

The functionality of most food materials is determined by their meso- and microstructure.

Structural characterisation at these length scales is challenging and requires the deployment of

complementary microscopic and spectroscopic techniques. For porous materials with dense

networks, probing the diffusion behaviour of water by NMR and MRI is an established approach

for obtaining descriptive meso- and microstructural parameters.

For low dense networks, however, water is not a suitable probe, hence recently the use of

macromolecular probes has been proposed. So far this approach was only demonstrated for

homogeneous low dense networks. Thus we embarked on a feasibility study to explore whether

this approach could also be applied to systems which display meso- and microstructural

heterogeneity. As a model system we used alginate networks, which can be prepared as

homogeneous gels, as well as beads which display a radial crosslink density gradient. First, a

homogeneous alginate gel was prepared and characterised by probing the diffusion behaviour of

dendrimer probes by NMR. By monitoring the diffusion behaviour of the dendrimer probe with a

range of sizes, the mesostructural heterogeneity of alginate networks could be established.

Subsequently, the diffusion behaviour of dendrimers in heterogeneous alginate beads was

probed.

65

Quantitative imaging of moisture ingress in cereal crackers: an

SPI study

C. Windta, M. Witek

a, F. Vergeldt

a, E. Gerkema

a, J. van Duynhoven

b, H. Van As

a

aLaboratory of Biophysics, Wageningen University, Wageningen, the Netherlands.;

b

Unilever Food and Health Research Institute, Unilever R&D, Vlaardingen, the Netherlands.

Shelf life in multi-component foods with texture contrast (e.g., crispy crackers with

moist filling) depends largely on texture stability. Because of differences in water activity,

Aw, migration of moisture from wet to dry components will occur. This is undesirable. A

number of processes can play a role in the transmission of moisture through cereal

materials: gas phase wetting (moist air); capillary transport of liquid water over the surface;

and diffusion of liquid water through the cereal matrix. It is not well understood which of

these mechanisms dominate under what conditions.

To image water ingress in materials with low water contents, conventional imaging

methods do not suffice: the T2’s are too short. In crackers the transition from crispy to soggy

occurs at an Aw of ~0.5. 1H relaxometry revealed that in this transition the largest changes

are visible at a T2 of 125 µs. The only readily available methods, suitable to image very

short T2 components, are based on Single Point Imaging (1). SPI was sensitive enough to

observe changes of water content: Crackers at Aw’s from 0.15 to 1 were successfully

imaged at a TD of 125 µs; contrast was sufficient to follow the transition from crisp to

soggy. SPI thus provides a tool to observe changes in texture stability.

We carried out moisture ingress experiments on cube-shaped pieces of dry crackers

which were exposed to humid air (Aw = 0.85). Quantitative moisture content profiles could

be obtained with adequate time-resolution, and this allowed monitoring of the transition

from the crispy to the soggy state in a real time manner. The quantitative profiles, which

were obtained for crackers with different porosity and geometries, matched with our

theoretical expectations.

References:

1. B.J. Balcom et al, J. Magn. Reson. A 123 (1996) 131

66

Anomalous diffusion expressed through fractional order

differential operators in the Bloch-Torrey equation

B.S. Akpaa, O. Abdullah

b, R.L. Magin

b

aDepartment of Chemical Engineering, University of Illinois at Chicago,

bDepartment of

Bioengineering, University of Illinois at Chicago

We report the use of PGSE NMR to characterize heterogeneity in the structure of

Sephadex gels and an oil-in-water emulsion. We demonstrate the application of fractional

order models to describe the restricted diffusion characteristic of such systems. Several

recent studies1,2 have investigated the so-called anomalous diffusion stretched exponential

model – exp[-(bD)α], where α is a measure of structural complexity that can be derived from

fractal models. In this paper we propose an alternative derivation for the stretched

exponential model using fractional order space derivatives. We consider the case where the

spatial Laplacian in the Bloch–Torrey equation is generalized to incorporate a fractional order

Brownian model of diffusivity3. This treatment reverts to the classical result for integer order

operations. The fractional order dynamics derived were observed to fit the signal attenuation

in diffusion-weighted images obtained from Sephadex gels. We have further investigated the

extent to which the fractional order model can describe restricted diffusion in an emulsion.

0 1 2 3 4

x 1011

10-2

10-1

100

b (s/m2)

S/S

0

data

Stejskal Tanner

biexponential

Karger

Magin

stretched

b (s/m2)b (s/mm2)

S/S0

Fig. 1: (A) Normalized signal intensity plotted versus b (b = (γδg)

2∆ ) for selected regions of interest in samples

of distilled water and Sephadex G-25, G-50, and G-100. Experimental data were fit to the fractional order

model. The inset shows a T2-weighted spin echo image of the sample. (B) Normalized signal intensity plotted

versus b for an oil-in-water emulsion. Five different models were used to fit the data. For this sample and

observation time (∆=15 s), the fractional order model yields the best fit.

References:

(1) K.M. Bennett et al., Magn. Res. Med. 50 (2003) 727.

(2) K.M. Bennett et al., Magn. Res. Med. 56 (2006) 235.

(3) R.L. Magin et al., J. Magn. Reson. 190 (2008) 255.

(A) (B)

67

Pore-Water Interactions in Hydrated Cement Pastes by NMR

V. Rodina, A. Valori

a, P.J. McDonald

a,

aDepartment of Physics, University of Surrey, GU2 7XH, UK.

Although cement based materials are ages old, and although they are now ubiquitous across

the world, a detailed understanding of their CSH microstructure and of pore water interactions

within this structure still remains elusive. To obtain this understanding is important because,

without it, predictive refinement and improvement of cement properties is not possible.

In recent years, NMR relaxometry has begun to challenge the long established picture of a

network of interconnected gel (∼nm) and capillary (µm) pores. NMR 2D relaxometry has shown

exchange of water between two reservoirs of different gel pores, has allowed measurement of the

water exchange rate between them (∼ms-1

) and has revealed a surprising lack of capillary pore

water reservoir1. Increasingly the NMR picture is supported by numerical modelling and

statistical analysis of high resolution micrographs2.

Notwithstanding, however, the NMR results have raised a series of further questions

concerning interpretation and detailed analysis of “NMR mass balance”:

• Current opinion suggests that the two dominant diagonal peaks seen in 2D exchange

relaxometry are associated with different size gel pores between CSH bricks of different packing

density in different environments. However an alternative explanation is that rather than different

pore sizes, the reservoirs can be associated with pores of similar size in regions with different

Fe3+

impurity densities leading to different surface relaxivities. Can this alternative picture be

proved or discounted?

• If the reservoirs are associated with gel porosity, then can experiments be performed to

quantify the intra-CSH layer water and the bound water fraction so as to complete a detailed

“NMR mass balance” and confirm the “two types of gel pore hypothesis”?

• Can the slow exchange rate (∼ms-1

) be rationally explained if it is recognized that this

corresponds to a distance of the order of microns across a nano-scale structure?

• Why can exchange not be observed in old ( >months) samples?

It is the purpose of this paper to address the first two of these questions and speculate on the

others. We describe a series of relaxation studies of “tailored” grey and D2O exchanged cements

to address the first question and a variable temperature double quantum filter NMR study of

hydrated pastes to address the second.

References:

1. L. Monteilhet, J.-P. Korb et al., Phys. Rev. E 74 (2006) 061404-9.

2. NANOCEM Workshop, CEA, Paris, 11th

March 2008. (www.nanocem.org)

68

Why you can’t use water to make Cryoporometric measurements of the pore size distributions in meteorites

- or in high iron content clays, rocks or concrete.

J. Beau W. Webbera,c,d, Philip Blandb, John H. Strangec, Ross Andersona, Bahman Tohidia

aInstitute of Petroleum Engineering, Heriot-Watt University, Edinburgh, EH14 4AS, UK; bImperial College London, South Kensington Campus, London SW7 2AZ, UK;

cSchool of Physical Sciences, University of Kent, Canterbury, Kent, CT2 7NH, UK; dLab-Tools Ltd., G19 Canterbury Enterprise Hub, University of Kent, CT2 7NJ, UK.

Many porous materials have high susceptibility magnetic gradients in the pores, due to the presence of iron or other magnetic materials. Thus if probe liquids are placed in the pores they exhibit fast decaying signals with a short T2

*. Usually the actual T2 of the liquids is also reduced, due the presence of paramagnetic ions in the pore walls. The usual solution in NMR is to measure an echo (or echo train) at short times. However recent work (Ref. 1.) has shown that water/ice systems near a pore wall form rotator phase plastic ice, with T2 relaxation times in the region of 100 to 200 ms. Thus if a NMR Cryoporometric measurement is attempted with a measurement time significantly less than 1 or 2 milli-seconds, the result is make a measurement based on the phase properties of the brittle to plastic ice phase transition, not that of the brittle ice to water phase transition. This gives rise to artefacts of small pore sizes that may not actually be present. This work successfully uses a-polar liquids instead.

0.01

0.1

1

Por

osity

µl

.Å-1

.g-1

Pore Diameter100Å 1000Å 1µm

Meteorites:AllendeACFER094ALH77307

Fig. 1: NMR Spin-Echo from an a-polar liquid in a Meteorite sample ALH77307. Due to susceptibility

gradients, the echo has a decay time HWHM of 2.5 µs.

Fig. 2: Pore size distributions for 3 Meteorite samples, as measured by NMR Cryoporometry,

using an a-polar liquid.

References:

1. Plastic ice in confined geometry: The evidence from neutron diffraction and NMR relaxation. J. Beau W. Webber, John C Dore, John H. Strange, Ross Anderson, Bahman Tohidi. J. Phys.: Condens. Matter 19, 415117, (12pp), 2007, Special Issue: Proceedings of The International Workshop On Current Challenges in Liquid and Glass Science .

2001000

0.3

0.2

0.1

0

Time µs

Sig

nal A

mpl

itude

V

High metal content in meteoritesgives rise to high suseptibilitymagnetic gradients, and henceshort T2* times.

This sample exhibits an echo witha HWHM decay time of 2.5 µs,with B 0 = 0.5 T, yet good NMRcryoporometry results are obtainableusing an a-polar liquid, but not water.

69

1H and

23Na Imaging of Sodium Sulfate Crystallization in a Drying

Porous Core

F. Maricaa, B. MacMillan

a, A. Hamilton

b, C. Hall

b and B. J. Balcom

a

aMRI Centre, Dept. of Physics, University of New Brunswick, Fredericton, Canada

bCentre for Materials Science and Engineering, The University of Edinburgh, Scotland, UK

Salt crystallization in the pore space of building materials and public works of art, such as

statues and monuments, is a common precursor to cracking and subsequent decay. We have chosen sodium sulfate as our model salt to study this phenomenon because of debate in the literature about the salt form [1] which precipitates, and the practical importance of this salt in the decay of porous materials.

Sodium sulfate is an attractive salt for MRI studies because of the possibility of undertaking both 23Na and 1H studies to identify and discriminate the various forms of crystalline sodium sulfate, from solution. In the case of 1H, we rely on the large volume fraction of water incorporated into the hydrated salts at room temperature solubilities, yielding solid like 1H signals. The three crystalline forms of sodium sulfate are Na2SO4.10H2O (mirabilite), anhydrous Na2SO4 (thenardite), and Na2SO4.7H2O (heptahydrate). The hydrated salts yield transverse 23Na lifetimes, which are less than 75 μsec at 7 Tesla, short lived, but still possible to visualize in a SPRITE experiment.

The hydrated salts yield 1H transverse lifetimes, which are similarly short lived at 2.4 Tesla, but with a Sinc-Gaussian lineshape for the mirabilite. Quantification of the solution and solid component signals relies on fitting the solid and liquid decay terms, spatially resolved, for both the 23Na and 1H experiments with a variable encoding time in a centric scan SPRITE profile experiment. The minimum encoding time was 11 μsec.

One-dimensional experiments on a partially sealed core of calcium silicate, saturated with a solution of sodium sulfate, were undertaken with controlled drying in a temperature and humidity controlled drying chamber. The 1H and 23Na experiments show increased crystal deposition near the drying face. The imaging results are temperature dependent, since changes in solubility with temperature are readily revealed in the local solid/liquid ratio. References: 1. A. Hamilton and C. Hall, Sodium sulfate heptahydrate: a synchrotron energy-dispersive diffraction study of an elusive metastable hydrated salt, J. Anal. At. Spectrom. (2008). 2. L. A. Rijniers, H. P. Huinink, L. Pel and K. Kopinga, Experimental evidence of crystallization pressure inside porous media, Physical Review Letters 94 (2005) 075503.

70

Time dependent diffusion studies in partially filled nanometric

and micrometric pores

Germán Farrhera, Ioan Ardelean

b, Rainer Kimmich

a

a University of Ulm, Sektion Kernresonanzspektroskopie, 89069 Ulm, Germany;

b Technical

University of Cluj Napoca, Physics Department, 400020 Cluj-Napoca, Romania

Provided that some conditions are fulfilled, diffusion measurements of liquids partially

filling porous media have indicated an enhanced self-diffusion coefficient relative to the bulk

phase [1, 2]. The reason for such observations is

the molecular exchange process between two

phases: liquid and saturated vapor. Molecular

exchange means that the solvent molecules are

intermittently subject to diffusion features in

either phase. Translational displacements in the

vapor phase are much faster than in the liquid

phase and contribute to the enhancement of the

effective diffusion coefficient.

In our previous studies [2] the

contribution of the vapor phase to molecular

diffusion in silica glasses with nanometer and

micrometer pores partially filled with

cyclohexane (non-polar) or water (polar) was

investigated for a given diffusion time (300 ms)

with the aid of pulsed field-gradient NMR

diffusometry. In the present investigations we

are focused on time dependent diffusion

coefficients. Experimental diffusion times

ranging between 100 sμ and 1 were probed using two unconventional NMR diffusometry

techniques [3]. The experimental data were compared with Monte Carlo simulations (see

Fig.1) on model structures showing a qualitatively equivalent behavior in the common time

window. On this basis we could conclude that the vapor phase contribution to the effective

diffusivity is particularly efficient on a diffusion time scale corresponding to root mean

squared displacements of the order of pores dimension.

s

10-5

10-4

10-3

10-2

10-1

100

101

102

103

104

105

106

0

10

20

f=

0.04

0.10

1.00

0.1

0.2

De

ff

(10

-9 m

2/s

)

time (µs)

10-5

10-4

10-3

10-2

10-1

100

101

102

103

104

105

106

0

10

20

f=

0.04

0.10

1.00

0.1

0.2

De

ff

(10

-9 m

2/s

)

time (µs)

Fig.1: Measured (dots) and simulated (lines)

effective (= overall averaged) diffusion

coefficient versus diffusion time for different

filling factors.

This work was financed by the Alexander von Humboldt Foundation, the Deutsche

Forschungsgemeinschaft, and the Romanian MEC (CNCSIS 1292/2006).

References:

[1]. F.D’Orazio, S.Bhattacharja, P.Halperin, and R.Gerhardt, Phys.Rev.Lett. 63, 43(1989).

[2]. I. Ardelean, G. Farrher, C. Mattea, R. Kimmich, Magn. Reson. Imag. 23, 285(2005);

[3]. G. Farrher, I. Ardelean, and R. Kimmich, J. Magn. Reson. 182, 215 (2006).

71

Magnetic field strength dependence of correlated Internal

Gradient-relaxation time distributions in Heterogeneous

Materials K.E. Washburn

a,c, C.D. Eccles

b, P.T. Callaghan

a

aVictoria University of Wellington,

bMagritek Limited,

cReservoir Laboratories AS

When there exists a magnetic susceptibility difference between two materials placed

in an applied magnetic field, local field gradients will develop at their interfaces. These

inhomogeneities in the magnetic field are referred to as “internal gradients”. The internal

gradients can have detrimental effects upon experiments, leading to distorted images and

inaccurate diffusion and T2 measurements. At the same time, the dependence of internal

gradients upon pore characteristics means they can provide useful information about the pore

space.

Previous research has correlated T2 with internal gradients1. To better understand the

behaviour of internal gradients, we present here a novel NMR technique to correlate T1

relation with internal gradients. In contrast with T2 or the restricted diffusion coefficient,

measurement of T1 relaxation is not susceptible to the presence of internal gradients. This

makes it ideal for use at high and ultra-high fields where internal gradients could potentially

be significant. We observe how the distribution of T1-internal gradient correlations in tight-

packed quartz sand and Mt. Gambier limestone change at 12 MHz, 200 MHz, 400 MHz, and

900 MHz. Surprisingly, even at the ultra-high field of 900MHz there exists signal at

relatively low internal gradients.

The results of the correlations provide experimental evidence to support the theory2

that the effective internal gradients present in a sample can scale as B0 while maximum

observable gradients can scale at up to B0

3/2. Our results also show that it is possible to

reliably perform experiments on even highly heterogeneous samples at ultra high fields and

that advantages come at these high fields.

References:

1. B. Sun, K.J. Dunn Phys. Rev. E. 65 (2002) 051309

2. M.D. Hurlimann, J. Magn. Reson. 131 (1998) 232

72

Monitoring of hydrocarbon uptake in highly conductive porous

media using a unilateral NMR instrument

H. Adriaensena, M. Bencsik

a, S. Brewer

b

aSchool of Science and Technology, Nottingham Trent University, Clifton Lane, Nottingham,

NG11 8NS, UK, bDSTL, Porton Down, Salisbury, Wiltshire, SP4 0JQ, UK

Studies of adsorption on porous media are of a great interest to the industry. Some of its characteristics can be determined by several methods such as volumetric and gravimetric measurements. A more sophisticated technique like calorimetry can be used[1] and gives information about chemical or physical adsorption. Furthermore, Magic Angle Spinning (MAS) Nuclear Magnetic Resonance (NMR) spectroscopy is one of the best techniques for studying adsorption on porous material at a molecular level[2].

However, using MAS spectroscopy requires that the porous solid has already experienced adsorption prior to the real experiment. Here we show that we can monitor the time course of the uptake of an adsorbant, using a unilateral NMR instrument.

Heptane (1.5 mL) was placed in a small (i.d. 1.8 cm, L 0.9 cm) cylindrical PTFE well, on top of which a monolithic cylinder of activated carbon (see photo) was placed. The profile

NMR MOUSE®[3] and its lift were flipped upside-down so as to monitor the NMR signal at the top of the monolith, in a slice (~400 µm x 20 mm x 20mm) 2 mm away from the instrument, overlapping with the edge of the sample.

The heptane pool evaporates within approximately two hours, whilst the NMR was monitored over more than 7 hours, allowing both adsorption and desorption to be detected.

The spin echoes were processed so as to show the time course of the effective relaxation rate (R2

eff) and the NMR amplitude (see figure). A similar experiment was repeated with a less dense ‘composite’ of activated carbon, and with an activated carbon cloth. We only show here the results for the monolithic activated carbon.

This work demonstrates that the uptake process of an organic compound (heptane) in an activated carbon can be

successfully monitored with the MR MOUSE®, in spite of the low resistivity (565 Ω.m-1) of the porous material. Different NMR signatures of the uptake and desorption are seen in the three different samples, and these will be discussed.

We are currently actively focusing in trying to quantitate the signal in terms of adsorbed heptane.

References:

1. T. A. Centeno, F. Stoeckli, recent advances in adsorption processes for environmental protection and security, pp: 9-18, Sep 09-12, 2006 Kiev ukraine

2. R. K. Harris, T.V. Thompson, P. Forshaw, N. Foley, K.M. Thomas, P.R. Norman and C.Pottage, Carbon Vol. 34, No. 10, pp. 1275-1279, (1996)

3. J.Perlo, journal of magnetic resonance 176 (1): pp. 64-70 Sep (2005)

73

MICROSCOPIC WETTABILITY OF CARBONATE ROCKS :A PROTON FIELD CYCLING NMR APPROACH

G. Freimana, J.-P. Korba, B. Nicotb, P. Ligneulb

aLaboratoire de Physique de la Matière Condensée, Ecole Polytechnique, CNRS, 91128 Palaiseau, France bSchlumberger SDCR, PO Box 2836, 31952 Al Khobar, Saudi Arabia

Nuclear Magnetic relaxation dispersion (NMRD) is strongly sensitive to the microscopic wettability of oil and brine bearing carbonate rocks. Exploring a very large range of low frequency enables isolating the typical NMRD dispersion features, 1/T1Surf, associated to the different processes of molecular surface dynamics. This allows an experimental separation of the surface and bulk microdynamics of oil and water even for a diphasic saturation of petroleum rocks. Several surface dynamical parameters such as translational correlation time, time of residence, surface affinity were determined and related to the concept of microscopic wettability of oil and water in porous media. We show our last results for diphasic mixtures (dodecane/water-brine) saturating carbonate rocks of different porosities and permeabilities when irreducible saturation of water is reached.

74

BLOCH-SIEGERT EFFECT IN MAGNETIC-RESONANCE

SOUNDING OF AQUIFERS

Oleg A. Shushakov1, 2, 3

1Baker Atlas Russian Science Center, Novosibirsk, Russia,

2Institute of Chemical Kinetics and

Combustion SB RAS, Novosibirsk, Russia. 3Novosibirsk State University

The magnetic resonance basic idea consists in reorientation of spin magnetization caused

by a weak-amplitude radiofrequency field when its frequency is in resonance with the frequency

with which the magnetization precesses in a strong static field. The special case of a weak field

rotating around the strong field is usually used in the magnetic resonance applications while it is

easily treated. F. Bloch and A. Siegert [1] studied more general case of the magnetic resonance

with elliptic polarization of the radiofrequency field in particular the commonly used case of

simple linear oscillation.

The magnetic resonance sounding (MRS) is a particular case of the magnetic resonance

in the Earth’s magnetic field as a static field BBo that is quite weak (of the order of 5•10 T).

Whereas the radiofrequency-field B1

-5

B produced by the surface antenna is linearly polarized and

can be compatible with the geomagnetic field amplitude, therefore the effect of strong

nonrotating radiofrequency field BB1 can take place in the MRS. Fig. 1 exemplify a good

agreement between measured data and calculated taking into account the Bloch-Siegert effect at

the Ob reservoir near Novosibirsk used as a model of aquifer with known free water content

(100%), depth (from 1 to 11 m under ice), and some other parameters.

0 4000 8000 12000 16000 20000

Pulse moment (A*ms)

0

1000

2000

3000

4000

MR

S a

mpli

tud

e (

nV

)

Ob reservoir

0-1m - ice

1-11m - water

40 ms pulse duration measured

40 ms pulse duration calculated

80 ms pulse duration measured

80 ms pulse duration calculated

No Bloch-Siegert effect

Fig. 1: The MRS amplitude vs pulse moment at 40ms and 80ms pulse durations measured

(circles) and calculated (lines). Dashed line – calculated without the Bloch-Siegert effect.

Conclusion

New physical model of MRS has been developed taking into account non secular effects

of spin Hamiltonian with strong and linearly polarized radiofrequency field, its experimental

proof has been carried out for subice water at the Ob reservoir (modeling an aquifer).

Reference

1. F. Bloch and A. Siegert. Magnetic resonance for nonrotating fields. Phys.Rev. 57 (1940)

522

75

Internal Magnetic Field Gradients: Experimental Study

A.R. Mutina and V.D. Skirda.

Kazan state university.

Russia, 420008, Kazan, Kremlevskaya str., 18.

Nuclear magnetic resonance is one of the most powerful techniques for the investigation of

porous media. When the sample is placed in a uniform external field, the magnetic

susceptibility difference between the porous material and the diffusing fluid leads to the

appearance of non-uniform internal magnetic fields in the porous media. The

inhomogeneities depend on many factors, including H0, value, pore morphology and

geometry. On one hand, these internal fields introduce ambiguity in the classical

interpretation of NMR experimental data. On the other hand, the internal magnetic fields and

their gradient distributions are determined by the porous media properties and can be used to

obtain information about the porous media, the fluid condition and the localization in it [1,2].

In the present work, internal magnetic fields and their gradient distribution functions were

studied by different NMR techniques, including the “tau-scanning” experiment [2]. Model

(glass beads, Vycor) and natural (sand, quartz sand) porous media, fully and partially filled

by different hydrocarbons, were studied. The dependences of internal field properties and

their gradient distributions on different factors (pore size, diffusant molecular mobility, pore

filling) were examined.

References

[1] Song, Y.-Q. Using internal magnetic fields to obtain the pore size distributions of porous

media, Concepts in Magnetic Resonance, 2003, V.18A, No2, P. 97 – 110.

[2] A.R Mutina, V.D. Skirda Porous media characterization by PFG and IMFG NMR, J.

Magn. Reson., 2007, V.188, No1, P.122 – 128.

Acknowledgements

The work was supported by RFBR No 07-03-01004 and Schlumberger Moscow.

76

ν=2.5 KHz (BEarth)

0.1

1

10

100

10 100 1000 10000d (µm)

R1 (

s-1)

Glass SpheresSea SandOregon, IL SandOttawa, IL SandWater

The 1H NMR R1 of Some Hydrated Synthetic and Natural Sands

C.L. Bray,a R.G. Bryant,

b M.J. Cox,

a G. Ferrante,

c Y. Goddard,

b S. Sur,

d and J.P. Hornak

a

aRIT, Rochester, NY;

bUniversity of Virginia, Charlottesville, VA;

cStelar s.r.l., 27035 Meade

(PV), Italy; dUniversity of Rochester, Rochester, NY

The 1H NMR R1 of hydrated sands is important in determining the poroscity of aquifers

using magnetic resonance sounding. Large R1 variations have been reported in laboratory

measurements of hydrated sands.[1]

We believe these variations are attributable to the

differences in the resonance frequency (ν) and mineralogy. To test this hypothesis, we

measured the R1 of fully-hydrated quartz (Sea, Ottawa, and Oregon) and synthetic

(borosilicate glass spheres) sands as a function of grain diameter (d) and ν using a field

cycling (Stelar FFC-2000) and 7 T high resolution (Bruker DRX-300) NMRs. Samples were

cleaned, dried at 200 ºC, sieved, placed in NMR tubes, and hydrated with 18 MΩ·cm water.

R1 data for all 28 samples displayed monoexponential behavior. R1 values increased

with both decreasing d and ν. Between 0.01 < ν < 30 MHz, samples displayed a linear

relationship between R1 and log(ν), and a power law relationship between R1 and d. The two

smallest diameter sieved Ottawa sands (84 & 117 µm) displayed an order of magnitude faster

R1. EDXM (JEOL JSM-6400V) analysis revealed the presence of pyrite, probably dislodged

by mechanical abrasion during the sieving and stopped by the smallest sieves. In the glass

spheres and the Ottawa 84 and 117 µm samples there is a dispersion between 30 and 300

MHz caused by electron spin flips associated with the larger amount of paramagnetic

material in glass and the Ottawa 84 and 117 µm samples.

Extrapolation of the R1 vs. log(ν) relationship allows prediction of R1 at ν = 2.5 kHz or

BEarth. (Fig. 1.) Measurable R1 differences were observed between the three natural quartz

sands and glass beads. These differences can not be attributed to the geometric

characteristics of the grains and porosity differences, as these were identical within

experimental uncertainty. The three quartz sands had different quantities of trace

paramagnetic impurities in the grains as determined by ESR (Bruker ESP-300,), which may

explain the different slopes in Fig. 1.

We conclude that literature variations in R1 are attributable to the differences in the ν of

the measurements and paramagnetic impurities in the sand grains.

References:

1. M. Müller, S. Kooman, U. Yaramanci, Near Surf Geophys, 3:275-285 (2005).

2. C.L. Bray, et al., J. Env. & Eng. Geophys. 11:1-8 (2006).

Figure 1. Predicted 1H R1 as a function of

particle diameter at ν=2.5 kHz for fully

hydrated, random-packed natural and

synthetic sands. Solid lines are a power law

fit to the data. The dashed line represents the

R1 value for bulk water. Error bars in d

represent ±1σ in the range of d values for the

sample. The flattening of the glass sphere

data at large d values is attributed to a beak

down in the packing efficiency when d

approaches the diameter of the sample tube

(dTube). Both the Bray[2]

data at dTube = 4.5

mm and our data at dTube = 7.7 mm display

this at d > 0.22dTube.

77

129Xe NMR of Xenon Trapped in Fully Dehydrated

Mesoporous Silica referred to Molecular Sieves 5A and 13X Mineyuki Hattori

a*, Kikuko Hayamizu

a, and Nobuhiro Hata

b

aPhotonics Research Institute,

bAdvanced Semiconductor Research Center,

National Institute of Advanced Industrial Science and Technology.

129Xe NMR techniques have been applied to probe porosity of mesoporous silica and

the pore size is known to relate with the chemical shift. Since the Van der Waals radius of Xe

is known to be 0.216 nm, the possible pore size to adsorb xenon should be larger than 0.4 nm

in diameter. Then the mean pore diameters ranging from 0.4 to 300 nm are the possible target

to show the relationship experimentally. We have developed an apparatus to produce the laser

induced hyperpolarized (HP) Xe gas[1] and tried to apply it to a self-assembled porous silica

sample (Lowk1), which is known as a candidate of low-dielectric constant materials for

interconnects in future ultra-large scale integrated circuits (ULSIs)[2].

The temperature dependent 129

Xe NMR spectra were measured at the frequency of

74.7 MHz from 168 to 373 K. Since the pore size is small and the diameter is about 2 nm, the

comparison was focused on materials having small pore sizes of commercial available

molecular sieves 5A (0.5 nm) and 13X (1 nm). When the HP xenon gas was introduced into

the samples at room temperature, the 129

Xe NMR signal always moved a little to the lower

field side to reach an equilibrium state and the life-time of the HP xenon was shorter in 5A

and 13X compared with Lowk1. The broad signal in the initial stage moved to the lower field

with narrowing to approach equilibrium states, accompanied by the gradual decrease of

signal intensity and the HP 129

Xe signal disappeared. After a little while, at the same position

the 129

Xe signal came out and gradually the intensity increased in the opposite phase to

approach to equilibrium states in intensity. Waiting about 10 min, the 129

Xe shift values were

obtained as shown in Fig. 1. The changes in the line widths are plotted versus temperature in

Fig. 2. The 129

Xe NMR spectra for mesoporous materials measure the averages of the

residence times or numbers of xenon atoms in the mesopores under adsorbed and gas phase

equilibrium at each temperature. The residence times or numbers in the mesopores vary

depending on temperature and pressure, which influences the 129

Xe chemical shift.

Fig. 1: 129

Xe shifts of 5A, 13X and Lowk1. Fig. 2: 129

Xe full line width at half height.

A 129

Xe NMR study on mesoporous materials under atmospheric pressure has

established. Combination of UHV treatment of the sample and the hyperpolarized 129

Xe

NMR techniques at atmospheric pressure indicate that estimation of the size and distribution

of pore by measurements of both chemical shift and line width.

References:

1. M. Hattori, Engineering Materials 52 (2004) 86.; N. Ohtake, M. Murayama, T. Hiraga, M.

Hattori, K. Homma, Japanese Patent 2003-4304.

2. N. Hata, C. Negoro, K. Yamada, and T. Kikkawa, Jpn. J. Appl. Phys. 43 (2004) 1323.

78

Spin-lattice relaxation of SF6 in heterogeneous porous materials

Juhani Lounila, Henri Tervonen, Jukka Jokisaari

NMR Research Group, Department of Physical Sciences, University of Oulu, P.O. Box 3000, FIN-90014 University of Oulu, Finland

The spin-lattice relaxation times (T1) of fluorine nuclei in gases such as CF4 and C2F6

are known to increase considerably when the gases are confined to small pores [1]. Such behavior is expected for collisionally interrupted intramolecular interactions in the limit of extreme narrowing. As a matter of fact, the dominant relaxation mechanism for these gases is the modulation of the spin-rotation interaction by molecular collisions. In the bulk gas, T1 is proportional to the density, as the correlation time of molecular angular momentum is determined only by the molecule-molecule collisions. In porous materials, the collision frequency is increased by the presence of molecule-wall collisions. Hence, the increase of the relaxation time may be attributed to the increase of the collision frequency [1].

A method for measuring the surface/volume ratio of porous materials by measuring how T1 of CF4 gas changes with confinement has recently been introduced [2]. The applicability of the method was demonstrated by analyzing samples of fumed silica (SiO2). This material is an ultrafine powder with a high air content and very little particle-particle contact, and its surface/volume ratio can be changed by compression.

However, in most practical applications the studied material is granular. Then the analysis is complicated by the presence of voids where the gas molecules can reside. In the present work, we look for a generalization of the method to macroscopically inhomogeneous materials. To this end, we have studied the 19F spin-lattice relaxation time of sulfur hexafluoride gas (SF6) adsorbed in different types of granular porous materials.

References: 1. M. J. Lizak, M. S. Conradi and C. G. Fry, J. Magn. Reson. 95 (1991) 548. 2. D. O. Kuethe, R. Montano and T. Pietrass, J. Magn. Reson. 186 (2007) 243.

79

Imbibition delay in porous media owing to thin low permeability surface layers

V. Bortolottia, M. Gombiab, A. Campagnolia, M. Camaitic, R.J.S. Brownd, P. Fantazzinib

aDICMA, University of Bologna, Via Terracini 28, 40131 Bologna, Italy

bDepartment of Physics, University of Bologna, Viale Berti Pichat 6/2, 40127 Bologna, Italy cICVBC, CNR, Via Madonna del Piano 10, 50019 Sesto Fiorentino (FI), Italy

d953 W Bonita Ave, Claremont CA 91711-4193, USA

Coatings are gaining an ever-increasing importance in many technological materials, as well as in the study of the mechanisms for the ingress of water inside these materials. In a previous paper1, fluid imbibition in rocks has been studied by Magnetic Resonance Imaging (MRI), before and after treatments to artificially change capillary and surface properties.

Using MRI, the presence of water inside each sample could be visualized and the height z(t) reached by the wetting front as a function of time during experiments of capillary absorption before and after treatment could be quantified. Very good fits to the data were obtained with theoretical and empirical models of absorption kinetics, starting from the Washburn model for capillary rise, adapted to porous media instead of capillary tubes. In some cases, the model had to be modified for application to a sample having a thin low-permeability layer as a result of a treatment process. In those cases we observed some delay in the absorption for both z(t) and mass absorbed, as shown in Fig. 1a and 1b, respectively. A mechanism to explain this behaviour may be the presence of a thin layer of very low permeability at the absorbing surface, and good fits to these data were obtained by introducing a delay parameter into the absorption equation. In this paper, that model has been checked on model porous media made of two rock samples of different permeabilities.

Fig. 1: Height reached by the wetting front during imbibition process (a) and total mass of absorbed water (b) as functions of square root of time for two samples. Circles and solid lines refer to an about 100 mm high sandstone sample; squares and dotted lines refer to a model porous medium constituted by the previous sample and a small, about 1 mm high, low permeability layer made of calcareous rock.

References: 1. M. Gombia, V. Bortolotti, R.J.S. Brown, M. Camaiti, P. Fantazzini, Models of water imbibition in untreated and treated porous media validated by quantitative Magnetic Resonance Imaging, J. Appl. Phys., in press 2008.

a b

80

Desalination of historical objects as studied by NMR

L. Pel, V. Voronina, K. Kopinga

Group Transport in Permeable Media,

Department of Applied Physics, Eindhoven University of Technology,

Eindhoven 5600 MB, The Nederlands

ABSTRACT

Salt crystallization is one of the most important reasons for the decay of historical

objects. A well known example is the historical city of Venice in which there is a lot salt

decay. Reducing salt contamination is important for conservation of old historic buildings.

Therefore often desalination treatments are done. The desalination of non-movable objects,

like buildings, is usually made through application of so-called poultices. That is a moistened

absorbent material is put on the surface of masonry. Water will than penetrate and desolve the

soluble salt present. By subsequent drying the ions will be transport fror the substrate into the

tpoultice. The removal of water-soluble salts sounds easy, but it can prove difficult in

practice. To achieve a better poultice performance, knowledge about salt transport in the

combination poultice/substrates is needed. We have studied the moisture and ion transport for

various desalination systems by NMR. Based on these measurements one can categorize

poultices according to their working principle of salt extraction and indicate the effectiveness

for various desalination methods.

81

Spontaneous crystallization of meta stable sodium sulfate as observed by NMR

T. Saidov , L. Pel

Group Transport in Permeable Media,

Department of Applied Physics, Eindhoven University of Technology, Eindhoven 5600 MB, The Nederlands

Salt crystallized in pores can give rise to damage of porous materials due to

existence of surface tension between pores and formed crystals. Sodium sulfate is one of the most damaging salts, but there are still many questions about its crystallization and damaging mechanisms. Especially which form of crystal will crystallize out firstly in porous media. In our experiments we have investigated the crystallization of sodium sulfate in porous building materials like fired-clay brick and in sandstone. In these experiments the crystallization was introduced by changing the temperature as the solubility of sodium sulfate is strongly temperature dependent.

Using NMR we have measured non destructively the Na concentration in these materials during crystallization. These experiments show the formation of meta stable heptahydrate form of sodium sulfate. Moreover these experiments indicate the existence of a spontaneous crystallization line. That is during cooling there is a maximum super saturation which can be reached. It was found that the type of porous material has no influence on the spontaneous crystallization line. This proves that crystallization process in these cases is caused only by the internal properties of solution. This behavior of sodium sulfate opens new questions in the understanding of crystallization damage.

82

High resolution MRI to probe

drying and barrier properties of coatings

S.J.F. Erich*+, L. Pel*, H.P. Huinink*, V. Baukh* and O.C.G. Adan

+

* Department of Applied Physics, Eindhoven University of Technology,

P.O. Box 513, 5600 MB Eindhoven, The Netherlands +

Built environment and geosciences, TNO, P.O. Box 49, 2600 AA Delft

In the last twenty years the environmental concerns regarding chemicals used by the

coating industry has been growing rapidly. More and more coatings are reformulated and

turned into waterborne systems. Still many problems concerning waterborne systems remain

and new investigations and measurements techniques are required. Since the use of

waterborne coatings is significantly increasing and waterborne coatings are sensitive to water

it is important to investigate the connection between water transport and degradation of such

coatings. In addition recently, the carcinogenicity of cobalt based catalyst was reported.

Therefore alternative catalyst systems are studied for application in alkyd coatings, based on

Mn or Fe. So techniques able to monitor drying of coatings and moisture transport in coatings

are needed.

At present, spatial resolution nuclear magnetic (NMR) imaging is the only technique

that can measure moisture transport and drying processes in coatings in-situ non-destructively

with sufficient high resolution (~ 5 μm) [1]. The principle of NMR is based on the resonance

of nuclei at a specific frequency that is proportional to the magnitude of the magnetic field. If

this field is inhomogeneous, nuclei in different positions will resonate with different

frequencies enabling imaging. By following the so-called GARfield approach (a specific

design of the magnet poles) [2], the desired field gradient (in our case 36.4 T/m) and spatial

resolution is obtained.

MRI probes the cross linking of the coating film by changes in polymer mobility,

which decreases with increased cross linking. The results of studies on alkyd coatings indicate

that the drying dynamics with cobalt as a catalyst is different from those with manganese

based catalyst. In case of cobalt a cross linking front movement proportional to √t was

observed [3], whereas this was not observed in case of the manganese based catalysts, not

even at high concentrations.

In addition studies have been made in which high resolution MRI shows the

absorption of a solvent borne coating by an underlying porous substrate. For solvent borne

coating the absorption is higher than in case of a waterborne coating, in which only the water

is absorbed. Also, NMR imaging is able to distinguish different layers, measure moisture

transport, measure swelling and/or shrinkage, and fluorine distribution [4].

References:

1. S.J.F. Erich, J. Laven, L. Pel, H.P. Huinink and K. Kopinga, Comparison of NMR and confocal Raman

microscopy as coatings research tools, Prog Org Coat 52 (2005), p. 210.

2. P.M. Glover, P.S. Aptaker, J.R. Bowler, E. Ciampi and P.J. McDonald, A novel high-gradient permanent

magnet for the profiling of planar films and coatings, J. Mag. Res. 139 (1999) (1), pp. 90–97.

3. S.J.F. Erich, J. Laven, L. Pel, H.P. Huinink, K. Kopinga, Dynamics of cross-linking fronts in alkyd coatings,

Appl. Phys. Letters 86 (2005) 134105

4. Diki, T.; Erich, S.J.F.; Ming, W.; Huinink, H.P.; Thune ,P.C.; van Benthem, R.A.T.M.; de With, G,

Fluorine depth profiling by high-resolution 1D magnetic resonance imaging, Polymer 48 (14) (2007), 4063

83

The formation of meta-stable sodium sulfate heptahydrate during drying in porous media as studied by NMR

T. Saidov , L. Pel

Group Transport in Permeable Media,

Department of Applied Physics, Eindhoven University of Technology, Eindhoven 5600 MB, The Nederlands

Salt weathering is a major cause of deterioration of porous building materials. Of

the salts responsible in practice, especially sodium sulfate is seen as very damaging. However many questions arise concerning which sodium sulfate phase will crystallize out during salt weathering. In this study we focused on the crystallization during isothermal drying of a sample saturated with a sodium sulfate solution. As the material is drying moisture will leave and hence the salt concentration will rise until the maximum solubility is reached. From that point on crystals will be formed. Using NMR we have measured quasi simultaneously both the moisture and Na profiles during drying. These experiments have been performed at various temperatures and concentrations. In our NMR experiments we observe the formation of a metastable phase of sodium sulfate, the heptahydrate crystals.

84

Impact of multi-scale moisture transport

on durability of hardened cement pastes

H. Chemmia,b

, D. Petita, P. Levitz

a, J.-P. Korb

a. C. Tourné-Péteilh

c, J-M Devoisselle

c.

aLaboratoire de Physique de la Matière Condensée, Ecole Polytechnique, CNRS, Palaiseau

91128, France. bATILH, 7 Place de La Défense - 92974 Paris La Défense Cedex, France.

cLaboratoire de Matériaux Catalytiques et Catalyse en Chimie Organique UMR

5618/ENSCM/Université Montpellier 1 8 rue de l’Ecole Normale, 34296 Montpellier Cedex

5, France.

Improving sustainibility and performance of cements is a key point to limit CO2 emission.

Long term durability is closely related to unsaturated moisture transport at different length

scales of these materials. In our study, hardened cement pastes exhibit threefold hierarchy:

intra-CSH (hydraulic binder) nanopore, mesopore structure at a scale ranging from 3nm to 20

nm and capillary pore network above 100 nm. The moisture transport at different water

filling controlled by temperature and relative humidity ratio is followed by a proton NMR

multi-scale approach using PFG diffusometry, field cycled relaxometry and spectroscopy

correlated with T1, T2 and T1ρ measurements. In parallel, a similar study is conducted in three

types of reference pore network, having nano, meso and macro pore network MCM41, vycor,

and controlled pore glass CPG respectively. The role and the efficiency of the three scales on

the moisture transport are reported.

85

Molecular dynamics of ionic liquids confined in solid silica matrix for lithium batteries

D. Petita, J.-P. Korba, P. Levitza, J. Le Bideaub, A. Vioux b, D. Brevetb

aLaboratoire de Physique de la Matière Condensée, Ecole Polytechnique, CNRS, Palaiseau 91128, France bInstitut Charles Gerhardt, Université Montpellier 2, CNRS, 34095 Montpellier, France

Ionic liquids are known for their high ionic conductivity and their wide electrochemical potentialities. They have recently been used as electrolytes in solar and fuel cells [1, 2] and lithium batteries [3]. For such applications, these ionic liquids have been immobilized in a solid matrix [4, 5]. However, the molecular dynamics of these liquid-like ions within a disordered solid matrix is still unknown. Here, we choose the (1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide) [BMI][TFSI] as an anion-cation pair of ionic liquid confined within a silica-like mesoporous matrices made by a sol-gel route from hydrophobic methyl groups precursors (ionogels made from tetramethoxysilane, methyltrimethoxysilane ; lithium salt Li TFSI was added). In a first step, we have measured the proton nuclear magnetic relaxation dispersion (NMRD) of the confined proton-bearing cation [BMI]. The frequency dependence of 1/T1 behaves as a power law, 1/T1~ω-1/2, over more than three orders of magnitude. This suggests a very slow decay of the intramolecular dipolar fluctuations of this confined cation at proximity of the pore surface. Such a power law remains over a very large range of temperature (10°C-70°C). This suggests a translational diffusion process at proximity of the pore surface. Several dynamical parameters have been determined from these proton NMRD such as: translational correlation time, activation energy as well as a surface diffusion coefficient that is similar to the one determined by quasi-elastic neutron scattering [6]. Moreover, we have observed a modification of the diffusive regime above 300K in conformity to recent conductivity measurements [5]. An estimation of the length of persistence associated to an average radius of curvature of the pores has been reached from the cross-over to a frequency independence of 1/T1 observed at low frequency. Last, we show the 19F NMRD of the proton-free anion [TFSI] and obtained a power-law behaviour almost similar to the protons. This is in favour of a very-correlated dynamical motion of the anion-cation pair at room temperature within the solid and disordered silica matrix. Both the methods and the theories presented here can be applied more widely to other conducting porous media.

This work is supported by the Agence National de la Recherche ANR-06-NANO-003

[1] B. O’Reagan and M. Graetzel, Nature 353, 737-740 (1991). [2] H. Nakamoto and M. Watanabe, Chem. Commun. 2339-2541 (2007). [3] M. Diaw, A. Chagnes, B. Carre, P. Wilmann and D. Lemordant, J. Power sources 146,

682-684 (2005). [4] J. Le Bideau, P. Gaveau, S. Bellayer, M.A. Néouze and A. Vioux, Phys. Chem. Chem.

Phys. 9, 5419-5422 (2007). [5] M.-A. Neouze, J. Le Bideau, P. Gaveau, S. Bellayer and A. Vioux, Chem. Mater., 2006,

18, 3931–3936. [6] S. Mitra, J.-M. Zanotti, M.-A. Néouze, S. Bellayer, A. Vioux and J. Le Bideau,

submitted. 86

Surface Diffusion in Catalysts Probed by APGSTE NMR

D. Weber, M.D. Mantle, A.J. Sederman and L.F. Gladden

Department of Chemical Engineering, University of Cambridge, Cambridge CB2 3RA, UK

The performance of heterogeneous catalytic processes is closely linked to the access

of reactant molecules to active catalytic centres on the surface. It therefore follows that the

successful operation of catalytic processes relies on a fundamental understanding of

molecular diffusion in heterogeneous catalysts. We have recently developed the experimental

protocol to measure chemically-specific surface diffusion coefficients using Pulsed Field

Gradient (PFG) Nuclear Magnetic Resonance (NMR) techniques. While the study of

diffusion phenomena with PFG NMR techniques evolved in the 1960s and is nowadays part

of the standard NMR toolbox, our approach allows us to quantify both the self-diffusion

coefficients of molecules in a surface layer and the bulk pore liquid of a porous medium

using the 13-interval APGSTE pulse sequence as outlined by Cotts [1]. For this it is

necessary to acquire signal with a large number of scans at high gradient strengths of up to

1200 G cm-1

to assess the surface layer diffusion coefficient which is typically one to two

orders of magnitude slower than the corresponding bulk liquid diffusion coefficient.

We have used this method to study solvent effects in the hydrogenation of methyl ethyl

ketone (MEK) in porous 1 wt% Pd/Al2O3 catalyst trilobes and 1 wt% Ru/SiO2 catalyst

pellets. Water has been reported to have a significant impact on reaction rates during

carbonyl hydrogenations [2-4]. Consequently we have assessed surface layer and bulk pore

liquid diffusion coefficients of MEK in mixtures with two solvents – water and isopropyl

alcohol (IPA) – across the entire concentration range (0-100 mol% water) under non-reacting

conditions. The comparison of these results with the reaction rates of MEK at these ternary

mixture compositions will be presented. The data will be used to suggest hypotheses for the

role of the solvents in this hydrogenation reaction.

References:

1. R. M. Cotts, M. J. R. Hoch, T. Sun, and J. T. Markert, J. Magn. Reson. 82 (1989) 252.

2. P. Kluson, L. Cerveny, Appl. Cat. A 128 (1995) 13.

3. A. Drelinkiewicz, R. Laitinen, R. Kangas, J. Pursiainen, Appl. Cat. A 284 (2005) 59.

4. B. S. Akpa, L.F. Gladden, K. Hindle, M. Neurock, D.W. Rooney, N. Sinha, E.H. Stitt,

D. Weber, ISCRE19, Potsdam, Germany (2006).

87

Kinetics and Microstructure of Hydrating Plasters

K. M. Song and L. F. Gladden

Department of Chemical Engineering, Pembroke Street, University of Cambridge,

Cambridge CB2 3RA, UK

Gypsum (CaSO4·2H2O) is widely used in the construction industry; it can be pre-

fabricated into wallboards for interior wall and ceiling applications, and is sold as

“plasterboard.” There are two varieties of plaster, α and β, produced by wet or dry methods

respectively, which can be hydrated to form solid gypsum products. The two forms of plaster

differ in their reactivity with water and in the mechanical properties of the hydrated products

[1]. 1H NMR provides non-destructive methods of probing the microstructure of construction

materials (e.g., cement [2]) continuously during hydration. This work is part of a broader

study investigating improvements to the mechanical properties of set plaster under conditions

of high relative humidity.

We present a comparative study of the two different types of plaster using various

NMR techniques. Only the β form has been investigated in depth previously [3, 4]. NMR T2

relaxation and imaging are used to compare the hydration kinetics and microstructural

changes that occur during the hydration of the two different types of plasters. We observe a

distinct difference in the hydration kinetics, as illustrated in Fig. 1, for the α- and β-plasters

with a water-to-plaster ratio of w/p = 1. The α-plaster starts hydrating immediately following

the addition of water, whereas the hydration of β-plaster is characterised by a long induction

period. The β-plaster hydrates much faster, although the time required for both plasters to

reach the same residual water content is similar.

We have also examined the hydration kinetics, microstructure and morphology of the

gypsum crystals as a function of w/p ratio and inclusion of additives. The NMR study is

supported by X-ray Microtomography data, which will also be presented.

0.7

0.8

0.9

1

0 10 20 30 40 50Hydration time (min)

Sig

na

l in

ten

sit

y (

a.u

.)

α-plaster (w/p=1.0)

β-plaster (w/p=1.0)

Hydration period for β

Hydration period for α

Fig. 1: Variation of average signal intensity from 1D profiles of two

different plasters during the hydration.

References: 1. N. B. Singh and B. Middendorf, Prog. Cryst. Growth. Character. Mater. 53 (2007) 57.

2. J. -P. Korb, Magn. Reson. Imaging. 25 (2007) 466.

3. H. Jaffel et al., J. Phys, Chem. B 110 (2006) 7385.

4. E. Badens et al., J. Cryst. Growth. 198/199 (1999) 704.

88

Understanding T2 relaxation times in hardening cement pastes using

the NMR exchange and the Powers’ hydration models

F. Stallmach, K. Friedemann

Faculty of Physics and Earth Sciences, University of Leipzig, Leipzig, Germany

Transverse relaxation times (T2) of physically bound water in cement pastes decrease

monotonously from values of about 10 ms to 20 ms observed in freshly mixed glutinous cement

pastes down to less than 1 ms in the hardened cements [1-3]. These T2 changes occur during the

first day(s) of cement hydration and are qualitatively rationalized by the decreasing pore sizes

due to the growth of solid calcium silicate hydrate (CSH) phases.

The Powers’ model of cement hydration [4] assumes that changes of macroscopic

properties of the cement pastes due to hydration may be described by interrelated changes of the

volume fractions of the unreacted cement, the capillary water, the solid hydration products and

the gel water. Among these volume fractions, the capillary and gel water can be observed by

low-field 1H NMR relaxometry. However, CPMG NMR studies with ordinary white and

Portland cement pastes show only one peak in the T2 relaxation time distribution, which means

that it is not possible to distinguish between the gel and capillary bound water fractions using

expected differences in their T2 relaxation times. Thus, there must be a fast exchange between

both water fractions on a microscopic scale, which allows one to observe only an effective,

averaged relaxation time in CPMG NMR.

The NMR two-site fast exchange model allowed us to predict the effective relaxation rate

(1/T2) from the relaxation rates of the gel (1/T2,gw) and the capillary (1/T2,cw) bound water

fractions and their relative amounts in the cement paste. Using the Powers’ model, the required

relative amounts can be replaced by the initial water-to-cement (wc) ratio and the degree of

cement hydration (α) yielding:

This equation is found to be in excellent agreement with experimentally determined

relaxation times measured in hydrating white and Portland cement pastes of different wc ratios

over a period of two days after cement paste preparation. Since the NMR measurements are also

used to determine the degree of hydration (α), the relaxation rates of the gel (1/T2,gw) and the

capillary (1/T2,cw) water fractions are the only adjustable parameter in this approach. Thus, they

may be used in future NMR studies to characterize different cement mixtures and the cement

hydration under the influence of additives changing, e.g., the water available during the

hydration reaction.

( )

1

23.0

19.023.01

23.0

19.01

,2,22 cwgw Twc

wc

TwcT αα

αα

⋅−⋅+−

+⋅−

⋅=

References:

1. K. Friedemann, F. Stallmach, J. Kärger, Cement and Concrete Research 36 (2006) 817.

2. N. Nestle, C. Zimmermann, M. Dakkouri and J. Kärger, J. Phys. D: Appl. Phys. 35 (2002)

166.

3. A. Plassais, M.P. Pomies, N. Lequeux, P. Boch; J.P. Korb, D. Petit, F. Barberon, Magnetic

Resonance Imaging 21 (2003) 369.

4. O. M. Jensen, P. F. Hansen, Cement and Concrete Research 31 (2001) 647.

89

Analysis of a Novel Ceramic Pore Structure using PGSE NMR

TR Brostena, KV Romanenko

c, SL Codd

a, JD Seymour

b and SW Sophie

a

a Department of Mechanical and Industrial Engineering, Montana State Universityb Department of Chemical and Biological Engineering, Montana State University

cMRI Centre, Department of Physics, University of New Brunswick

Ceramic pore structures produced by a novel tape casting method suggest the opportunity to

optimize porous transport properties for application in a variety of industrial sectors; e.g. fuel

cells, filtration, etc. [1]. Magnetic Resonance Microscopy techniques have been used to

probe the nature of these ceramic pore structures. Spatially resolved pulsed gradient spin

echo (PGSE) techniques were used to obtain displacement propagators of pressure driven

octane flowing through a tape cast ceramic sample. The displacement distributions suggest

anomalous (non-Gaussian) transport of octane through the ceramic structure. In addition,

spatially resolved PGSE techniques were used to observe the pore structure induced restricted

diffusion of no-flow octane. The impact of restricted diffusion on the signal attenuation

clearly demonstrates both the one dimensional affine scaling of the average pore dimensions

and the quasi-elliptical transverse pore shape, see Figs. 1 & 2.

C

A

B

A

C B

x

y

z

C

A

B

C

A

B

A

C B

x

y

z

x

y

z

z

Fig. 1: SEM image of tape cast ceramic, bar=400 m

E( )E( )

Fig. 2: PGSE signal attenuation of stationary octane saturating a

tape cast ceramic as a function of longitudinal position and gradient

vector orientation, constant q

References:

1. S.W. Sofie, Fabrication of functionally graded and aligned porosity in thin ceramic

substrates with the novel freeze-tape casting process, (2007) J. Amer. Ceram. Soc., 90

[7]

2.25mm

90

Ion and polymer mobility in swelling hydrogels

J. Martinsa, T. Mang

b, S. Stapf

c

aInstitute for Technical and Macromolecular Chemistry, RWTH Aachen University,

bInstitute for Applied Polymer Science, Aachen University for Applied Science,

cDepartment of Technical Physics II, TU Ilmenau

The possibility for applying hydrogels as sensors and actuators is to a large degree

dependent on the feasibility to generate fast responses to a range of different external stimuli;

these can be temperature, pH value, or ion concentration in a solvent. The response is

commonly expressed in swelling or shrinking of the gel, where volume changes of two or

three orders of magnitude have been observed. Such hydrogels can then be used e.g. as drug

release carriers, valves, filters or mechanical switches. In most cases, hydrogels represent a

porous mesh constructed of chemically cross-linked polymer chains, but additional

macroscopic porosity can be superimposed into the gel matrix.

The fast swelling response in ionic

hydrogels, of which poly(Na acrylate) is a

prominent example, is due to the osmotic

pressure difference caused by the concentration

of ions inside and outside the hydrogel matrix.

In the presence of water the carboxylic groups

of the polymer backbone will dissociate, adding

a supplementary swelling force by repelling

each other. The swelling parameters can be

controlled by the crosslinker concentration and

the charge density which both influence the

mechanical properties. Hydrogels with a small

amount of crosslinker possess an increased

mobility of the polymer backbone and are able to imbibe large volumes of water.

Furthermore, foreign ions like calcium, magnesium and aluminum will replace the sodium

ions present near the carboxylic groups. Because of the bivalency character of these ions,

they restrict the mobility of the polymer backbone, while the osmotic equilibrium is partially

destroyed affecting the swelling properties. The dynamics and concurrence reactions between

the sodium and the foreign ions inside and outside the hydrogel matrix are not sufficiently

understood. This understanding is fundamental to control the swelling parameters.

In order to investigate the swelling response and the polymeric and ionic dynamics, a

series of polyacrylate hydrogels partially neutralized with sodium hydroxide and dried over a

lyophilized process were prepared. The concentrations of the crosslinker as well as of the

ions in solution were varied systematically, and 1H and

23Na relaxation, diffusion and

imaging experiments were carried out. In this work we report on the dependence of the

sodium dynamics within the pore space of the hydrogel, and provide real-time 23

Na

microscopic images during the swelling and drying processes that allow quantification of the

spatially dependent backbone and ion mobility as a function of external stimuli.

23Na Hydrogel

Fig. 1: 23

Na Spin-density MRI of Poly-

(Na acrylate) Hydrogel

91

Time evolution of the “Solid” and “Liquid” 1H signals duri ng cement paste hydration

V. Bortolottia, R.J.S. Brownb, A. Campagnolia, P. Fantazzinic, M. Gombiac, F. Peddisa

aDICMA, University of Bologna, Via Terracini 28, 40131 Bologna, Italy

b953 W Bonita Ave, Claremont CA 91711-4193, USA cDepartment of Physics, University of Bologna, Viale Berti Pichat 6/2, 40127 Bologna, Italy

NMR Relaxometry of 1H nuclei has been widely applied to cement pastes during hydration, and models have been presented to interpret the experimental data. In this study we used a particular kind of data computation, to follow the changes of freshly prepared cement pastes, starting from the beginning up to many hours and days after the preparation. The method consists in the quasi-continuous T1 distribution analysis of proton signals, after “Solid” and ”Li quid” signal separation, as described below.

As well known, many sample materials give both “Solid” signal components, from protons of low mobility, and also “Liquid” components, from protons with higher molecular mobility. The “Solid” gives a very fast FID decay that in a log-signal plot is initially quadratic in time (initially approximately Gaussian), while the “Liquid” gives initially exponential decay. In the course of time we settled on an algorithm (T1-Filter) to filter the FIDs obtained at different Inversion Recovery times in IR experiments able to give estimates of four parameters: Tg, the time constant of a Gaussian with the same initial log-FID curvature, X2, its corresponding extrapolated amplitude, T2_FID, the T2 corresponding to the initial slope of the “Liquid” component, and X1, its corresponding extrapolated amplitude. The Solid/Liquid amplitude ratio, X2/X1, is also calculated. Special attention has been devoted to avoiding the introduction of excessive scatter in the computed results, by forming stacks of FIDs on which computation were made, and to take into account the eventual changes in the FID slopes of the “Liquid” component with Recovery time, due to possible T2 multi-exponential decays. In such a way we were able to get separate files for input to our multi-exponential inversion software (UPENWin)1 for both “Solid” and “Liquid”, with an accurate determination of the two signal amplitudes at each Inversion Recovery time. We have observed the two components (“Solid” and “Liquid”) in many biological samples, bone, wood, cartilage, collagen. Also cement pastes during hydration show the two components, that we have studied by using the four parameters described above and by quasi-continuous T1 distributions of both components.

The distributions vary very quickly and are clear markers of the rapid changes of the cement pastes in the first few hours after preparation. As a new result, we found that the parameter X2/X1 changes drastically in the first hours and takes days to reach a plateau.

References: 1. V. Bortolotti, R. J .S. Brown, P. Fantazzini, UPENWin, a software for multi-exponential inversion decay, exploited by University of Bologna, 2008.

92

Time-dependent diffusion coefficient of proton in perfluorosulfonated membrane

T. Ohkubo, K. Kidena, A. Ohira,

National Institute of Advanced Industrial Science and Technology, Polymer Electrolyte Fuel

Cell Cutting-Edge Research Center, Japan

The perfluorosulfonated membranes have attracted much attention in the past decade because of their applications in polymer electrolyte fuel cells (PEFC). However, little is known about pore structure in long-range order (submicron) when resolving the detailed transport mechanism relevant to the proton conductivity. In this study, we focused on probing the pore structure of submicron size of three-dimensional structure of the polymer membrane by means of observation of the time-dependent diffusion coefficient, D(∆eff), of proton. Diffusion measurements were carried out by two kinds of pulse sequences, oscillating gradient spin echo (OGSE)1 and bipolar gradients with a longitudinal eddy delay (BPPled)2. Nafion 117 was used as a perfluorosulfonated membrane. The water content of the sample was controlled in a humidity chamber (r.h.=50%) or with saturated aqueous solution of K2SO4 (r.h.=97%). The resulted water content was estimated dry and wet membrane weight.

Fig. 1 shows the D(∆eff) in the perfluorosulfonated membrane with 6 and 24% of water contents at various temperature from 233 to 308 K. The diffusion coefficients in ∆eff>30 ms indicated a constant value regardless of water content and temperature. On the other hand, the diffusion coefficients in ∆eff<2 ms were remarkably changed at low temperature. This tendency was more pronounced with 6% of water content. It was suggested that the pore structure with lower water content have pathways with restricted geometry for proton transport. To evaluate the surface-to-volume ratio of pore, we applied Mitra equation3 to D(∆eff) of the perfluorosulfonated membrane with 6% of water content at 233K. A magnitude of restricted space for proton diffusion could be evaluated as hundreds nanometers, which was clearly larger than well-known water domain size, about several nanometers. We think that it is likely to reflect superstructures originated from clustering of semicrystals observed via ultra small angle neutron scattering4.

0.01

0.1

1

10 100

1H Self-Diffusion coefficients [10-9 m

2/s]

1H Self-Diffusion coefficients [10-9 m

2/s]

∆eff [m s]

0.01

0.1

1

10 100∆eff [m s]

278K293K308K

233K248K263K

278K293K308K

233K248K263K6% w ater content 24% w ater content

Fig. 1: Time-dependent diffusion coefficients of proton in N117 with 6 and 24% of water content at various temperatures.

References: 1. B. Gross and R. Kosfeld, Messtechnik 77 (1969) 171. 2. D. Wu, A. Chen, and C. S. Johnson, J. Magn. Reson. A 115 (1995) 260. 3. P. P. Mitra, P. N. Sen, L. M. Schwartz and P. Le Doussal, Phys. Rev. Lett. 68 (1992) 3555. 4. M. Kim, C. J. Glinka, S. A. Grot and W. G. Grot, Macromolecules 39 (2006) 4775.

93

Liquid Crystals Constrained in Thin Film Capillary Investigated by DHK Spin Echo SPI Measurements

Jing Zhang, Rod MacGregor, Bruce J. Balcom

Physics Department, University of New Brunswick, Canada

Liquid crystals are anisotropic fluids and it is well known that self-diffusion behavior can

provide valuable information on the molecular assemblage in liquid crystal phases [1]. Nuclear magnetic resonance is a well-established technique to study orientational order and dynamics of liquid crystals in porous media [2]. In this work we present a study of liquid crystal properties in constrained thin film capillary (thickness ~ 200 µm) at different temperatures. The liquid crystal under study is the common N-(p-methoxybenzylidene)–p –n –butylaniline (MBBA). Its nematic phase extends from 20°C to 43°C in bulk. Translational self-diffusion measurements have been performed in the mesophases of MBBA liquid crystal by means of pulsed field gradient approach as a function of temperature in the nematic and isotropic phases. We found that in the isotropic phase the diffusion coefficient is not a simple scalar quantity and it has orientational dependence due to the constraint of the capillary wall (Dzz/Dyy≈7.3). To characterize the diffusion distribution of liquid crystals within the thin film, a double half k (DHK) spin echo single point mapping (SPI) diffusion mapping experiment was performed using a parallel plate RF probe. The method provides accurate diffusion tensor measurements with pixel resolution of 4µm.

References: 1. G. J. Kruger, Phys. Rep.-Rev. Sec. Phys. Lett., 82 (1982), 229. 2. M. Vilfan, T. Apih, A. Gregorovic, B. Zalar, G. Lahajnar, S. Zumer, G. Hinze, R. Bohmer,

G. Alhofff, Magn. Reson. Imag., 19 (2001) 433.

94

Application of low field and solid-state NMR spectroscopy to

study the liquid/liquid interface in porous space of clay minerals

and shales

A. Borysenkoa, B. Clennell

b, I. Burgar

c, D. Dewhurst

b, R. Sedev

a, J. Ralston

a

aIan Wark Research Institute UniSA - Adelaide Australia,

bCSIRO Petroleum - Perth

Australia, cCSIRO Materials Science and Engineering - Melbourne Australia

The study of complex multi-component systems such as liquid/liquid/mineral

interactions in porous space of clays and shales is a challenging task. In petroleum research

understanding displacement, redistribution, adsorption of oil and water plays an important

role in considering: (i) oil production (ii) the problem of reservoir top seal, and (iii)

environmental considerations.

The samples, minerals, clays and shales, were prepared in the form of powder to

simplify the process of liquid saturation and crude oil treatment. The preliminary screening of

the samples using optical/fluorescence and scanning electron microscopy and XPS revealed

differences in wetting behaviour of the samples. Several NMR methods were applied to

characterise these systems: oil and water phases competing for mineral surfaces. The

liquid/liquid displacements and distribution within porous medium as studied with the

“Maran Ultra-2”, T2 were up to 10000 ms. The oil interaction with the mineral surfaces does

affect the oil mobility, T2, values only up to 100 ms. Using a combination of Bruker NMR

Surface Analyser MOUSE (15 MHz proton resonance), NMR Minispec (20 MHz), and the

solid-state NMR (300MHz) we were able to investigate surface-adsorbed hydrocarbons.

The detailed NMR analysis shows that the process of liquid/liquid displacement can

be characterised and analysed considering the relaxation times and signal amplitudes. The

results from low field NMR measurements are in good correlation with the solid-state data,

where the changes of oil dynamics in the solid matrix were correlated to the different crude

oils and different shale composition – variable hydrophobicity. All in all the NMR results

were found to correlate well with other data and they have been consistent so we could

consider a combination of low and high-field NMR spectroscopy as a tool of choice for

investigating shale and clay wettability.

95

PFG-NMR and Goldman-Shen study of water diffusion and cross-

relaxation in Polyelectrolyte Multilayers

C. Wende, M. Schönhoff

Institute of Physical Chemistry, University of Münster, Germany

The self-assembly of polyelectrolytes of alternating charge from aqueous solutions onto charged

surfaces leads to the formation of multilayered films. Their permeation properties for different

types of molecules and their porosity are of general interest[1-3], since the multilayers can act as

separation membranes or colloidal hollow carriers. A direct determination of diffusion

coefficients of small molecules within and through the multilayers is attractive, but hard to

achieve for nanometer thin films.

Here, we study free-standing films, which are densely stacked to achive a high filling factor, and

subsequently equilibrated at different relative humidities. We investigate the diffusion of water

molecules in a film of poly(styrene sulfonate)/poly(diallyldimethyl ammoniumchoride) by means

of Pulsed-Field-Gradient-NMR. A non-Gaussian diffusion behaviour is found, which strongly

depends on the relative humidity. Furthermore, a pronounced dependence of the PFG echo decay

on the diffusion time is observed, which might suggest restricted diffusion in a porous structure.

However, an evaluation in a model of restricted diffusion does not lead to conclusive results.

Therefore, a potential influence of cross-relaxation between water and polymer 1H spins is

investigated by a Goldman-Shen pulse sequence. The data are analysed in a model established

for cross-relaxation in polymer hydrogels[4,5], which yields cross relaxation rates. These are of an

order of magnitude relevant in the interpretation of diffusion echo decays, which leads to a re-

interpretation of the diffusion data.

From both the diffusion and cross relaxation experiments it can be concluded that the dynamic

behavior of water in polyelectrolyte multilayers is highly dependent on the water content.

References:

[1] A. Jin, A. Toutianoush, and B. Tieke, Appl. Surf. Sci. 246 (2005) 444.

[2] X. Liu and M. L. Bruening, Chem. Mater. 16 (2004) 351.

[3] F. Vaca Chávez and M. Schönhoff, J. Chem. Phys. 126 (2007) 104705.

[4] L.J.C. Peschier, J. A. Bouwstra, J. de Bleyzer, H. E. Junginger, J. C. Leyte, J. Magn. Reson.

B 110 (1996) 150.

[5] D. Topgaard, O. Söderman, Langmuir 17 (2001) 2694.

96

Time dependent NMR spectroscopy on ionic ferrofluids

D. Heinricha, A.R. Goñi

b, L. Cerioni

c,d, T. Osán

d, D.J. Pusiol

c,d, C. Thomsen

a

a Institut für Festkörperphysik, Technische Universität Berlin, Hardenbergstraße 36,

10623Berlin, Germany, bICREA, Institut de Ciència de Materials de Barcelona, Esfera UAB,

08193 Bellaterra, Spain, cCONICET, Argentina,

dSpinlock S.R.L, Av. Padre Viera 255, X5186KSE, Alta Gracia, Córdoba, Argentina

Magnetic nanoparticles colloidally suspended in a ferrofluid exhibit a tendency to form

clusters and chain-like structures under the influence of an external magnetic field [1-7]. Its

behavior resembles a liquid in a porous media. Recently we used Raman spectroscopy to monitor

the metastable cluster formation and its dynamics in surfacted and ionic ferrofluids [5-7]. In this

work we present results of a complementary study of the magnetic-field induced behavior of a

water-based ionic ferrofluid (IFF) with a concentration of 1 vol.% using nuclear magnetic

resonance (NMR) spectroscopy. For the measurements we used a low-resolution NMR

spectrometer working at room temperature with a homogenous magnetic field of 400 mT.

In the experiments, an electrostatically stabilized ferrofluid which has not been exposed

previously to any magnetic field is placed at 300 K in the bore of the NMR spectrometer. The

NMR spectrum displays an asymmetric feature which is composed by two peaks of different

intensity (Fig.1). The main peak is blue-shifted by approximately 17 kHz from the resonance

frequency of “pure and free” water molecules [8]. The less intense peak appears to be centered at

around 3 kHz above the pure water frequency. Both peaks are attributed to two different dynamic

environments of water molecules in the ferrofluid [8].

Figure 2 shows the time evolution of the peak amplitude of both peaks in a time scale of

more than one hour. The amplitude of the NMR signal of the 17 kHz feature exhibits a slight

increase in the first 300 s followed by a strong reduction in intensity, reaching its minimum after

approximately 15 min. Simultaneously, the amplitude of the 3 kHz peak corresponding to the

NMR signal stemming from water molecules far from the magnetic nanoparticles increases

monotonically, saturating at times longer than one hour. This contrasting behavior of both NMR

signals is readily understood by considering the dynamical processes within the ferrofluid

triggered by an external magnetic field, as revealed by Raman spectroscopy [7].

Fig.1 NMR spectra of an ionic ferrofluid at

different times in a magnetic field of 400 mT.

Fig.2 Time evolution of the amplitude of the

17 kHz (solid points) and the 3 kHz (open

symbols) NMR peaks. Lines are fits using

simple exponential functions.

97

TD-NMR investigation of the effect of curing-temperature on the

hydration and porous microstructure development in cement

S. Ghosh, Z.Harry Xie

The minispec Division, Bruker Optics, Inc. The Woodlands, TX – 77381, USA

Time-domain NMR experiments can be implemented using an instrument with low-

field magnet. This makes implementation of fast quantitative material characterization

combined with portability, possible. The hydration and porous microstructure development in

cement during curing is key to its structural attributes. How the ambient temperature during

the curing process of cement, of various types, affects these processes, is investigated in this

work through time-domain NMR. Variable temperature NMR relaxation and diffusion

measurements were utilized to look into the hydration kinetics and the evolution of

microstructure during the curing process. All these measurements were done with probes

tuned to the Hydrogen nucleus, and thus looking at the change in mobility and diffusion

characteristics of water molecules present in the curing cement system. The NMR data will

also be supported by IR-imaging results. This research work will be helpful in non-invasive,

fast, and quantitative assessment of proper curing conditions (e.g. temperature) for a specific

type of cement, as well as predictive mathematical model development of the process.

98

withdrawn

99

withdrawn

100

Methodical aspects of 2D NMR spectroscopy under conditions of ultra-high pulsed field gradients

M. Gratz, C. Horch, S. Schlayer, P. Galvosas

Universität Leipzig, Fakultät für Physik und Geowissenschaften, Germany

Multidimensional NMR correlation experiments based on an inverse Laplace transformation (see e.g. [1]) have been established as a powerful tool for the investigation of material structure and properties within the last few years. We successfully combined these recent methods with ultra-high pulsed magnetic field gradients of up to 35 T/m [2], which allows us to correlate molecular displacements in the order of 100 nm with NMR parameters, such as relaxation time T2.

Basic concepts for the handling of ultra-high pulsed field gradients were adapted accordingly. This concerns in particular the introduction of a read gradient at suitable positions of the pulse sequences as well as the development of a script automating the steps taken during the experiment. The software interface controlling the NMR experiment ensures the detection of mismatched pulsed field gradients and their subsequent correction as proposed in [2].

However, an additional requirement for experiments involving the inverse Laplace transformation for data processing is to attenuate the NMR signal down to a signal-to-noise ratio of about one. This demand tends to let the aforementioned detection of gradient mismatches fail for evanescent signals. We overcame this challenge by the introduction of a termination condition. Hence, the NMR measurement is terminated, if the signal is attenuated below a predefined signal-to-noise ratio and/or if the pulsed field gradient matching procedure returns unreasonable values.

First experiments were successfully applied to several lubrication oil samples with distributions of diffusion coefficients and NMR relaxation times (fig. 1). Please note the small diffusivities and short T2 which indeed require short and ultra-high pulsed field gradients.

Fig. 1: Diffusion-Relaxation correlation experiment with lubricating oil at T=340 K

References:

1. M. D. Hürlimann and L. Venkataramanan, J. Magn. Reson., 157 (2002) 31–42.2. P. Galvosas, F. Stallmach, G. Seiffert, J. Kärger, U. Kaess and G. Majer, J. Magn. Reson.,

151 (2001) 260–268.

(a) NMR signal map (b) Inverse Laplace transformation of (a)

101

Two–dimensional Laplace inversions applied to multi–component T2–T2 exchange experiments

R. Fechetea, D. Moldovana, D. E. Demcoa,b and B. Blümichc,

a Departments of Physics, Technical University of Cluj-Napoca, Romania

bDWI an der RWTH-Aachen, Germany c Institute of Technical and Macromolecular Chemistry, RWTH-Aachen University, Germany

The two-dimensional (2D) T2-T2 molecular exchange NMR experiments with a period of

magnetization storage between the two T2 relaxation encoding periods [1] are presented. A CPMG pulse sequence with variable echo number is used to encode the signal amplitude in the indirect dimension and the magnetization decay is recorded by a CPMG sequence in the directly detected dimension. The two-dimensional time map was inverted using a fast Laplace algorithm [2, 3] to obtain the T2–T2 exchange map. The relaxation exchange from the fast relaxing nuclei in molecules associated with pore surfaces or superficial liquid shell and the more slowly relaxing nuclei associated with molecules in the bulk liquid filling the materials pores can be observed as cross–peaks in the T2–T2 exchange maps. T2–MZ(store)–T2 2D 1H NMR spectra, recorded at high and low homogeneous magnetic fields of water and oil in sand, air bubbles in water and foams, exchange of liquid–foam and liquid–saturated vapors of chloroform, are presented. Systematic studies as a function of the mixing time were performed. Interesting and at the same time challenging results, from the interpretation point of view, are reported for a mixing time comparable to or smaller than the maximum duration of the CPMG pulse sequence. In the majority of cases only unidirectional exchanges, from low to high or from high to low relaxation times are observed. A clear bidirectional exchange was observed for water, the superficial liquid shell, and foam. The 2D Laplace inversion method of T2–T2 exchange NMR is a valuable tool to observe the motion of molecules in heterogeneous environments. Nevertheless, a quantitative interpretation of the T2–T2 exchange maps in the fast exchange limit is possible only using numerical simulations. The preliminary T2–T2 exchange NMR experiments, shows that the investigations of heterogeneous processes, like catalysis in small chemical reactors, can successfully be investigated.

References: 1. L. Monteilhet, J.-P. Korb, J. Mitchell, and P. J. McDonald, Observation of exchange of

micropore water in cement pastes by two-dimensional T2-T2 nuclear magnetic resonance relaxometry, Phys. Rev. E 74, 061404 (2006).

2. L. Venkataramanan, Y. Q. Song, M. D. Hürlimann, Solving Fredholm Integrals of the First Kind With Tensor Product Structure in 2 and 2.5 Dimensions, IEEE Trans. Sig. Process. 50, 1017-1026 (2002).

3. Y. Q. Song, L. Venkataramanan, M. D. Hürlimann, M. Flaum, P. Frulla, and C. Straley, T1-T2 Correlation Spectra Obtained Using a Fast Two-Dimensional Laplace Inversion, J. Magn. Reson., 154, 261-268, (2002).

102

Monte Carlo simulations of the two-dimensional NMR T2-T2 exchange of fluids in porous media

D. Moldovana, R. Fechetea, D. E. Demcoa,b, E. Culeaa, and B. Blümichc,

a Departments of Physics, Technical University of Cluj-Napoca, Romania

bDWI an der RWTH-Aachen, Germany c Institute of Technical and Macromolecular Chemistry, RWTH-Aachen University, Germany

Classical two-dimensional (2D) NMR T2-T2 exchange experiments with a period of

magnetization storage between the two T2 relaxation encoding periods T2–MZ(store)–T2 [1] assumes that the molecular exchange processes take place during the store period. Fast NMR experiments use the CPMG pulse sequence to stroboscopically encode the T2 relaxation time in the indirect and the directly observed dimensions. There are numerous samples, characterized by long T2 relaxation times, for which the relevant exchange time, associated with the storage period, is comparable to or less than the duration of CPMG echo train decay. In these cases, dynamic processes during the preparation and exchange periods can have a direct influence on the decoding of T2 relaxation via a second CPMG sequence. In the fast exchange limit, the simplest CPMG decay can be affected by molecular exchange. The interpretation of results in the presence of rapidly exchanging relaxation is not trivial and can be done correctly only by numerical simulation of molecular dynamics. The effect of molecular exchange processes on the one-dimensional (1D) T2 and two-dimensional (2D) NMR T2-T2 Laplace distributions [2, 3] were studied by Monte–Carlo simulations. A homogeneous static magnetic field is assumed, and no disturbing effects like the pore surface magnetic susceptibility is considered. A systematic study is conducted as a function of NMR pulse sequences parameters, transverse relaxation rates, time, and distribution of exchange rates. A system from free random walk molecules and excluded volume bulk molecules was considered. The difference in the dynamic behavior inside saturated and unsaturated pore is highlighted. Systems in dynamical equilibrium and non-equilibrium were also under investigation. The geometrical factor is studied for unidirectional flow, single pore bulk-to-surface shell exchange, and systems with connected pores. Particular molecular dynamics including exchange phenomena lead to particular features in 1D and 2D data. Matching the experimental data with simulated ones can be developed into a useful tool for the characterization of complex materials. For example the knowledge of fast exchange rate can provide a powerful means of characterizing the pore surface-to-volume ratio.

References: 1. L. Monteilhet, J.-P. Korb, J. Mitchell, and P. J. McDonald, Observation of exchange of

micropore water in cement pastes by two-dimensional T2-T2 nuclear magnetic resonance relaxometry, Phys. Rev. E 74, 061404 (2006).

2. Y. Q. Song, L. Venkataramanan, M. D. Hürlimann, M. Flaum, P. Frulla, and C. Straley, T1-T2 Correlation Spectra Obtained Using a Fast Two-Dimensional Laplace Inversion, J. Magn. Reson. 154, 261-268, (2002).

3. L. Venkataramanan, Y. Q. Song, M. D. Hürlimann, Solving Fredholm Integrals of the First Kind With Tensor Product Structure in 2 and 2.5 Dimensions, IEEE Trans. Sig. Process. 50, 1017-1026 (2002).

103

Using 2D Inverse Fourier Transformation to Observe Internal

Magnetic Fields and Internal Magnetic Field Gradients

Lauren Burcaw and Paul T. Callaghan

MacDiarmid Institute for Advanced Materials and Nanotechnology, School of Chemical and

Physical Sciences, Victoria University of Wellington, New Zealand.

Multidimensional NMR measurements, including inverse Fourier transformation (IFT), and

more recently, inverse Laplace transformation (ILT), are frequently used in separation,

correlation, and exchange experiments [1]. Often these experiments allow us to obtain

knowledge about a system that is otherwise difficult to attain with a one dimensional

experiment.

It is well known that porous media in an applied magnetic field will experience internal

magnetic field inhomogeneities due to susceptibility differences from its various components.

This was first discovered by Brown [2] using ferromagnetic grains suspended in water with

5% carboxy-methylcellulose, and expanded upon by Drain [3]who examined the broadening

of magnetic resonance lines in powdered samples. More recently, the internal field has been

studied in more depth by Audoly [4] using a finney pack of non-penetrating spheres.

We study these internal magnetic field inhomogeneities using exchange spectroscopy with

various sized soda glass bead packs in distilled water. Our experiment involves an

acquisition and evolution domain each consisting of a free induction decay so that the two

Fourier domains correspond to the spectral distribution of local fields. By performing the

exchange experiment over several different mixing times, twe obtain 2D data from which we

study the effects of changes in location of a diffusing water molecule in the presence of these

internal magnetic field, using 2D IFT. We show that exchange in the internal magnetic field

can be seen in a 2D IFT by peak broadening. This is potentially useful for studying the

internal magnetic field gradients in more complicated systems such as sandstones or other

porous media. We anticipate the extension to 2D correlation experiments involving the

Fourier domain of local field and an Inverse Laplace domain [5] of internal magnetic field

gradient.

References:

[1] P.T. Callaghan, C.H. Arns, P. Galvosas, M.W. Hunter, Y. Qiao, and K.E. Washburn,

Magn. Reson. Imaging 25, 441-444 (2007)

[2] R.J. Brown, Phys. Rev. 121, 1379-1382 (1961)

[3] L.E. Drain, Proc. Phys .Soc. 80, 1380-1382 (1962)

[4] B. Audoly, P.N. Sen, S. Ryu, and Y.-Q. Song, J. Magn. Reson. 164, 154-159 (2003)

[5] J.G. Seland, K.E. Washburn, H.W. Anthonsen, J. Krane, Phys. Rev. E., 70, 051305

(2004).

104

Two-dimensional NMR of Diffusion Systems Yi-Qiao Song, Lukasz Zielinski, and S. Ryu

Schlumberger-Doll Research, Cambridge MA 02139

2D NMR of relaxation and diffusion is a significant advance in the study of porous materials. There have been many applications of 2D NMR as a way to distinguish signals from different phases e.g. oil, water and gases, and different geometries. 2D NMR has also been used to probe water diffusion from one pore to another. The analogy to such experiments is the conventional NOESY, the 2D NMR in frequency domain. Such experiments offer the potential to probe the connectivity of pore structures and permeability of membranes.

We show that the diffusion system itself, even within a single, shows a whole range of behaviors in 2D NMR, including the appearance of off-diagonal peaks in T1-T2 experiments (see top panels in the Figure) which have been interpreted as arising from exchange. We confirm, however, that under certain circumstances, the off-diagonal peaks become significantly enhanced by diffusive coupling between pores (see bottom panels in the Figure) and could, in principle, be used as a signature of connectivity. We study these effects within the general framework of the Torrey-Bloch formulation, which becomes the diffusion equation,

, and ,

for the evolution of longitudinal magnetization. Here m is the magnetization deviation from its equilibrium, and and are the eigenfunctions and eigenvalues, respectively.

The spectroscopy of such an eigen system characterizes the pore space. In fact, with different surface relaxivity for longitudinal and transverse magnetization decays, two sets of eigenfunctions and eigenvalues are relevant, the longitudinal L-modes and the transverse T-modes. Each is a complete set of functions, mutually orthogonal within the set. The overlap between the L-modes and T-modes depends on the surface relaxivity. This overlap is critical for the variety of behavior of the 2D NMR results, including the appearance and intensity of the off-diagonal peaks. Though our present results are limited to one-dimensional coupled pores, our approach based on the Torrey-Bloch equation allows numerical simulations of 2D NMR for arbitrary 3D pore geometry.

Single pore, T1-T2 and T2-T2 experiments

Two coupled pores T1-T2 experiments, from strongest coupling (left panel) to weakest (right panel)

105

Spatially Resolved Two-Dimensional T2-D MRI in Porous Media

L. Li,a O. Petrov,a B. Balcoma

aMRI Centre, Department of Physics, University of New Brunswick

Discrimination of water and oil in petroleum reservoir exploration is a critical measurement for down-hole NMR. The most robust and successful measurements have been developed by Schlumberger and feature a 2D inverse Laplace transform to resolve the T2 distribution, which largely depends on the pore structure and pore size distribution, from the diffusion coefficient distribution which largely depends on molecular mobility. Neither suffices in isolation to fully characterize the fluids or the medium.

In our laboratory studies of fluids in porous media, frequently reservoir core plugs, we seek a similar ability to characterize the fluids and matrix - but spatially resolved. Spatial resolution is critical for heterogeneous samples and for examination of processes that introduce a heterogeneous fluid distribution – such as core flooding.

In this work we explore a new version of the stimulated echo PFG T2-diffusion experiment. Diffusive attenuation is modified by changing the PFG gradient amplitudes, not the stimulated echo evolution time. We add pure phase encode spatial encoding gradients in linear combination with the PFG diffusion gradients for spatial resolution of the experiment.

This simple modification of the experiment is remarkably successful, and robust, yet seems not previously to have been reported in the literature. We limit spatial resolution in core plug samples to 1 dimension in order to ensure reasonable acquisition times for the overall experiment.

Fig. 1: Single pixel T2-Diffusion spectrum from a one dimensional phantom. Inverse Laplace transformation was undertaken with Magritek software.

We ensure short echo times and accurate gradient switching through the use of a magnetic field gradient measurement and adjustment method outlined in a separate abstract.

References: 1. M. Hürlimann and L. Venkataramanan, J. Magn. Reson., 157, 2002, 31-42.

Fitting value: T2 = 91 ms Bulk measurement, T2 = 95 ms

Fitting value: D= 1.71E-009 m2/s Bulk value: D= 1.9E-009 m2/s

106

Diffusional coupling from T2-store-T2 NMR experiments in bimodal pore systems.

Marc Fleury and Jawed Soualem

Institut Français du Pétrole, Rueil Malmaison, France

Diffusional pore coupling is an important issue in the NMR characterization of multimodal porous systems such as complex carbonates and shaley sandstones. Essentially, the pore size distribution deduced from conventional T2 distributions may give wrong estimations of porosity partitioning, especially at high temperatures typical of oil reservoir at which wide unimodal instead of bimodal distributions may be observed. Recently, Montheillet et al.1 proposed a two dimensional T2-store-T2 experiment to study proton exchanges in cement pastes. We applied this sequence to study and quantify the pore coupling in carbonates and clay-sand mixtures. The T2-store-T2 experiment is an elegant method to demonstrate and quantify pore coupling. Depending on the exchange rate between two pore populations, off-diagonal peaks may appear in a T2-T2 correlation plot. Using an analytical solution, a parametric study indicates that the amplitude of the off-diagonal peaks is usually only a few percent of the largest on-diagonal peaks. Using our 2D Laplace inversion software, we show that the detection of small amplitude peaks is possible but their location may not be accurate. We show different examples (Fig. 1) of T2-store-T2 experiments. For each case, the existence of pore coupling is also verified qualitatively by observing the shift of the micro/macro porosity peaks in the T2 distribution at different temperatures.

T2 (ms)

T 2 (ms)

100

101

102

103

100

101

102

103

Fig. 1: Result of a T2-store-T2 experiment on a bimodal carbonate sample (Lavoux). Mercury porosimetry and MEB observations clearly indicate two pore populations, whereas the 1D T2 distribution is wide and unimodal. The strong coupling occurring on that sample is expressed as a distribution of off-diagonal peaks. (Storage time 90 ms, Larmor frequency: 23.7 MHz).

References: 1. L. Monteilhet, J.-P. Korb, J. Mitchell, P.J. McDonald, Physical Review E74, 061404 (2006)

107

Method and Experimental Study of 2-D NMR Logging

Guangzhi LIAO, Lizhi XIAO, Ranhong XIE, Shaoqing FU, Huijun Yu

State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum, Beijing, 102249

Abstract As the signal of reservoir fluids, such as oil, gas and water, always overlap in the T2

distribution of nuclear magnetic resonance (NMR). It is difficult to identify fluid types effectively

by 1-D NMR logging. Considering the characteristic of transverse relaxation (T2), longitudinal

relaxation (T1), diffusion coefficient (D) synthetically, we can perform the measurement of

two-dimension NMR (2D-NMR), which has provided the foundation for qualitative identifying

and quantitative evaluation of the reservoir-fluids. Three methods of 2D-NMR logging were

proposed and several forward models were designed accordingly. And 2D information of (T2, D),

(T2, G) and (T2, T1) was acquired by measuring multiple CPMG pulse sequence with variable

echo spacing (TE) and waiting times (TW) respectively. Then we had got the 2D proton density

distribution through multi-exponential inversion method. At last, in order to prove the measuring

and data processing methods, the experiments of 2D-NMR with some man-made rocks and natural

core samples are performed. It had identified fluid types clearly, calculated fluid-saturation exactly,

and detected the distribution of internal field gradients of core samples which contain

paramagnetic minerals.

Keywords 2D-NMR logging, Transverse-Relaxation, Longitudinal-Relaxation,

Diffusion-Coefficient, Gradient Fields

108

Methodology on Searching for Similarities between the NMR

Relaxation Spectra and Reference Logs

P. Romero

Baker Hughes, division Baker Atlas

This paper presents a technique on searching for similarities between Nuclear

Magnetic Resonance (NMR) logging spectra and reference logs or associated information.

The fundamental approach for implementing this technique relies on understanding each of

the bins, or discrete values, of the NMR spectra as individual logs or a time series (depth).

These NMR bin logs can be easily set in relationship to additional logs, e.g. resistivity,

saturation or gamma ray, without restricting their general applications.

The Pearson correlation and the entropy-based mutual information have been chosen

as two basic tools for searching for log-log similarities. Figure 1 shows an averaged T2

distribution over a certain depth interval, and a color-coded degree of correlation with the

resistivity log. When using the Pearson correlation approach, the p-value provides an

additional method to validate the degree of correlation on a statistical basis.

Figure 1 shows an averaged T2 distribution for a certain depth interval and color-

coded information regarding the degree of correlation of the T2 spectra with a resistivity log.

Dark blue color can be interpreted as predominantly water, and dark red can be interpreted as

hydrocarbon.

Fig. 1 T2 distribution and color-coded correlation with resistivity log

This method can be applied widely. The information it generates can contribute to

reduce uncertainties in the formation-evaluation process.

References:

1. A. L. Edwards, The Correlation Coefficient, Ch. 4 in An Introduction to Linear Regression

and Correlation. San Francisco, CA: W. H. Freeman, pp. 33-46, 1976

2. http://homepages.inf.ed.ac.uk/rbf/CVonline/LOCAL_COPIES/GILLES1/entropy.html

109

On the use of two dimensional inverse Laplace routines on NMR data

Geir Humborstad Sørlanda, Åsmund Ukkelbergb, Hege C. Widerøec

aAnvendt Teknologi AS, Norway, bUniversity of Oslo, Norway, cStatoilHydro Research Centre, Norway Keywords. Inverse Laplace, relaxation, diffusion, porous media When applied to NMR data arising from relaxation time and diffusion experiments, either in combination (two dimensional) or as single experiments, the use of Inverse Laplace routines has become a crucial tool in NMR data analysis.

After working with numerous rock core plugs at different saturation states over a period of five years we have concluded that it is impossible to draw good conclusions from the NMR data when applying well known algorithm which estimates continuous distributions of the inverse Laplace transform. A typical problem with the routines currently in use, is the fundamental error that is produced when defining the grid as an input to the numerical routines. As the responses from any fluid confined a porous rock may locate peaks off from the grid points, the resulting distributions will be wrong. In particular we have found that the current routines fail to fit the estimates of the total NMR signal. One may of course adjust the grid so that one achieves as good a fit as possible, but that relies on the brave assumption that any multimodal distribution may be fitted to that set of grid points. In practical life this is the exception not the rule, i.e. it is not possible to fit all peaks to the grid points available once having defined the grid. The underlying problem is that estimating an inverse Laplace transform with continuous distributions is an attempt at forcing too much information out of experimental data. When using the ANAHESS algorithm, the situation is different. No grid is defined, and the numerical routine fits the NMR data to the peak values that yield the best fit. When comparing the residuals from the two routines we find the ANAHESS overwhelmingly superior.

PDF laget med uregistrert versjon av pdfFactory www.westsoft.no

110

A Fast Monte Carlo Sampler for NMR T2 Inversion

M. Prange

Schlumberger-Doll Research

The inversion of noisy NMR T2 echo data a T2 spectrum is widely recognized as an

inherently non-unique process [1]. One approach to quantifying this uncertainty is to use

Monte Carlo sampling. Measurement noise is well described by an uncorrelated normal

distribution. When combined with the non-negativity constraint on T2 spectral values, this leads to spectral values following a non-negative normal distribution. There are published

samplers for truncated normal distributions [2], of which non-negative normal samples are a

subset, but I show that these converge very slowly for the highly ill-conditioned covariance

matrices that are present in NMR T2 spectral inversion. The reason for this is that they are based on Gibbs’ samplers that update the spectral estimate just one T2 component at a time.

When all of the spectral elements are fixed but one, that one has little room for change

without violating the noise constraints on the data. This means that each spectral sample can

only be slightly different from the preceding sample, indicating a high degree of statistical correlation and thus slow convergence. I resolve this by simultaneously updating two

neighboring spectral components at a time, allowing changes due to one spectral component

to be offset by changes in its neighbor. Core to this improvement is a fast 2D sampler for

non-negative normal distributions. I demonstrate that this improves convergence by more than two orders of magnitude. Such speedup allows routine Monte Carlo inversion of 1D

NMR spectra, and opens the door for the inversion of 2D NMR spectra.

References:

1. R. Parker, Y-Q Song, Assigning uncertainties in the inversion of NMR relaxation data, J. Mag. Res. 174 (2005) 314-324.

2. G. Rodriguez-Yam et al., Efficient Gibbs sampling of truncated multivariate normal

with application to constrained linear regression, Technical Report, Colorado State

University (2004).

111

Signal Optimization in Inhomogeneous Fields: Application of

Optimal Control Theory to NMR Oil-Well Logging

T. Bornemana, D. Cory

a, M. Hürlimann

b

aDepartment of Nuclear Science and Engineering, Massachusetts Institute of Technology

bSchlumberger-Doll Research

Nuclear Magnetic Resonance (NMR) techniques are especially useful in well logging to

characterize hydrocarbon bearing geological formations surrounding a bore hole. The Carr-

Purcell-Meiboom-Gill (CPMG) sequence forms a key element is all measurements that

determine the relaxation and diffusion properties (see Ref. 1). Our ability to accurately perform

these measurements, however, is severely limited by environmental constraints. The necessity of

designing a magnetic resonance tool to be placed inside a drilled well leads to gross in-

homogeneity of the applied static and RF fields. Being a resonance phenomenon, NMR requires

the Larmor precession frequencies of nuclear spins in a static field to be close to the frequency of

the applied RF field. Field in-homogeneity leads to a spatial variation of Larmor frequencies and

the effective RF field strength a spin experiences. Thus, the obtainable signal is limited by the

number of spins excited and refocused uniformly by the pulses in the CPMG sequence. We

demonstrate the use of Optimal Control Theory (OCT) techniques to design complex amplitude

and frequency modulated pulses that extend the effective bandwidth of uniform control to a

range well beyond that achievable via conventional ‘hard’ pulse techniques. OCT techniques

have previously been used in the design of NMR pulses, (see e.g. ref. 2), but this is the first time

that it has been used to the study of Carr-Purcell dynamics in inhomogeneous fields. We find that

these advanced techniques allow us to maximize the signal over a significantly wider dispersion

of Larmor frequencies. Compared to standard hard pulses, the excitation/refocusing bandwidth

for OCT pulses can be increased at least four-fold. The pulses also have the convenient property

that they are phase-distortionless. They can therefore be directly substituted into any sequence to

enhance performance without further need for modification, making the pulses appropriate for a

wealth of applications beyond what we have considered in this study. Additionally, analysis of

the asymptotic echo amplitude and phase created by directly substituting our pulses into the

Carr-Purcell sequence demonstrates stability in the absence of T2-relaxation and diffusion

consistent with the Meiboom-Gill correction. This indicates that the CPMG sequence using

directly substituted OCT pulses will significantly enhance the signal-to-noise ratio of diffusion

and relaxation measurements in oil-well logging.

References: 1. M. D. Hürlimann and D. D. Griffin, Spin Dynamics of Carr-Purcell-Meiboom-Gill-like

Sequences in Grossly Inhomogeneous B0 and B1 Fields and Application to NMR Well

Logging, J. Magn. Res. 143 (2000) 120-135.

2. T. E. Skinner, et al. Application of optimal control theory to the design of broadband

excitation pulses for high-resolution NMR, J. Magn. Res. 163 (2003) 8-15.

112

Split-180o Sequences

D. E. Freed a, U. M Scheven

b, and M. D. Hürlimann

a

aSchlumberger-Doll Research,

bDepartment of Chemistry, FCT-Unversidade Nova de Lisboa

In applications of NMR in inhomogeneous fields, sequences based on the Carr-

Purcell-Meiboom-Gill (CPMG) sequence play a central role. The standard CPMG sequence

consists of an initial 90o excitation pulse, followed by a long string of 180

o refocusing pulses.

This creates a series of echoes that decay with a characteristic relaxation time T2eff.

Here we present a modified sequence, the so-called Split-180o sequence that

specifically takes advantage of grossly inhomogeneous fields. In its simplest implementation,

the 180o refocusing pulse of the CPMG sequence is split into two separate pulses. This

sequence, which can be viewed as a modification of the CPMG sequence, simultaneously

generates two types of signal that can be separately detected. One is a CPMG-like signal that

decays with the expected relaxation time T2eff. In addition, a second type of signal builds up

and approaches a steady-state. The amplitude of this dynamic equilibrium depends on the

ratio of the longitudinal to the transverse relaxation times, T1/T2.

We present experimental results and a new theory that describes both signals in a

unified manner.

Fig. 1: Experimental results of the split-180o sequence on a

sample of doped water with a relaxation time T1 = T2 of 100

ms. The left panel shows the amplitudes of the CPMG-like

signal that decay towards zero with T2, whereas the right panel

displays the amplitudes of the dynamic equilibrium signal that

rise towards a constant amplitude.

Fig. 2. Echo shape of the CPMG-like signal (left) and the

dynamic equilibrium signal (right). The CPMG-like and the

dynamic equilibrium signals form out of phase to each other.

113

Measurements of diffusion, T1 and T2 in one-shot by MMME

X.-H.Rena, H. Cho

b, and Y. -Q.Song

c,

a Schlumberger Reservoir Completions, Rosharon, TX 77583

b Memorial Sloan-Kettering Cancer Center, New York, NY 10065

c Schlumberger-Doll Research, Cambridge, MA 02139

In the presence of a constant gradient field, a sequence of n hard RF pulses will allow

multiple coherence pathways (q) to produce multiple signals. This class of sequences is called

multiple modulation multiple echoes (MMME)[1-4]. Different q creates different spatial

phase modulation and yields echo signal at different time. Consequently, in one scan of the

sequence, different modulations can be acquired. By setting the time period appropriately, the

4-pulse MMME sequence can give well separated 13 echoes for acquisition. The

magnetization of each echo coherence pathway (i.e. each echo q) is a product of three factors:

the RF pulses ( ), the diffusion (D), and the relaxations (T1, T2). It can be written as:

=2

2

1

1

0exp)exp()(

T

c

T

cDbfMM

qq

qqq,

where M0 is the initial total magnetization, the coherence pathway q. The symbols

21, qqq ccb are sequence dependent diffusion, T1, and T2 weighting factors, respectively.

In practice, for two measurements with same RF flip angles, but different time delays, the

ratio of the corresponding echo magnitudes will eliminate fq( ). The amplitude ratios S1/S2

and weighting factors of the corresponding 13 echoes give sufficient data to uniquely

determine D, T1 and T2 simultaneously.

Different echo spacing and gradient strength will result in different diffusion and

relaxation weightings for each coherence pathway, thus offer a mechanism for optimization.

For the sequence with the optimized parameters, synthetic data were generated by giving a

random error of 3% on each echo signal, and the standard deviation of T1, T2 and D were

calculated for different T1, T2 in the range of 20 - 2000 ms (T1 T2) with a constant D value.

It can be seen, that using giving sequences, the error propagators in D, T1 and T2 are

controlled within 10% (Fig.1). Experimentally, the MMME obtained values showed a good

agreement with those from standard Pulsed Filed Gradient (PFG), Inversion Recovery (IR),

and CPMG measurements.

Fig. 1: Calculated standard deviation of D, T1, T2 for optimized MMME sequences with a random

error of 3% on each echo signal.

References:

1 Y. -Q. Song and X. Tang., J. Magn. Reson. 170 (2004) 136.

2 Y. -Q.Song et al., J. Magn. Reson. 172 (2005) 31.

3 H. Cho, L. Chavez, E.E.Sigmund, D.P.Madio, Y.-Q.Song, J. Magn. Reson. 180 (2006) 18.

4 H. Cho, X.-H.Ren, E.E.Sigmund, Y.-Q.Song, J. Chem. Phys. 126 (2007) 154501

114

Double-PFG Diffusion-Diffraction in Ellipsoidal Pores

E. Özarslana, C. G. Koay

a, P. J. Basser

a

aSection on Tissue Biophysics and Biomimetics, NICHD, NIH.

Repeated application of diffusion gradient pulse pairs [1] in a pulsed field gradient

(PFG) experiment provides important insights into pore microstructure. For example, the dimensions and eccentricity of yeast cells were measured from the fourth order term of the double-PFG signal attenuation when the mixing time between the two encoding blocks is long [2]. In this abstract, we propose an alternative double-PFG technique to address the same problem, which exploits the diffusion-diffraction phenomenon [3] in double-PFG experiments [4]. In our approach, all diffusion gradients in a single acquisition are applied along the same direction with a mixing time of 0. The experiment is subsequently repeated with diffusion gradients applied along different directions.

Fig. 1: The double-PFG pulse sequence considered in this work. The duration of the diffusion gradients (δ) are assumed to be very short. q=γδG/2π, where γ is the gyromagnetic ratio and G is the gradient strength. The separation between subsequent pulses (Δ) is assumed to be long enough so that a spin can probe the largest dimension within the pore.

As demonstrated in [4], the expected signal for this pulse sequence is given by ( ) ( )∗qq 2~~ 2 ρρ , where ( )qρ~ is the Fourier transform of the pore shape function. This

expression for the NMR signal attenuation leads to two interesting observations: (i) the first signal minimum occurs at exactly half the q-value necessary to observe nonmonotonicity in a single-PFG experiment; (ii) the diffraction minima are replaced by zero-crossings making the diffraction pattern robust to the heterogeneity of the specimen. We calculated the NMR signal attenuation from ellipsoidal pores with different eccentricities using the expressions in [5].

=00

=900

=600

=900

Fig. 2: The continuous and the dashed portions of each curve correspond to the positive and negative signal values, respectively. Black and red curves are computed for coherently oriented, axially symmetric ellipsoids whose ratio of major to minor axes was 4. The black curve is expected when the gradients are applied along the major axis, whereas the red curve is expected when the gradients are in the transverse plane. The blue curve shows the signal from ellipsoids whose axes are uniformly distributed over a spherical cap between the polar angles θ=00 and θ=600. The green curve is predicted forisotropically distributed pores.

The proposed method makes it possible to observe diffusion-diffraction phenomenon in anisotropic pores even when the pore orientations are randomly distributed. The first zero-crossings of the diffraction patterns with gradients applied along different directions can be used to quantify the compartment size as well as the pore shape anisotropy using double-PFG acquisitions with 0 mixing times, hence mitigating the effect of relaxation-related signal loss.

References:

1. D. G. Cory, A. N. Garroway and J. B. Miller, Polym. Preprints. 31 (1990) 149. 2. Y. Cheng and D. G. Cory, J. Am. Chem. Soc. 121 (1999) 7935. 3. P. T. Callaghan, A. Coy, D. MacGowan, et al., Nature 351(1991) 467. 4. E. Özarslan and P.J. Basser, J. Magn. Reson. 188 (2007) 285. 5. C. G. Koay, J. E. Sarlls and E. Özarslan, Magn. Reson. Med. 58 (2007) 430. This research was supported by the intramural research program of NICHD.

115

Accurate RTOP Estimation from PFG-NMR Signal

E. Özarslana, C. G. Koay

a, P. J. Basser

a

aSection on Tissue Biophysics and Biomimetics, NICHD, NIH.

The return to the origin probability (RTOP) for diffusing molecules is a valuable

indicator of porous media microstructure [1-2]. For example, in isolated pores with non-

relaxing walls, the pore volume is related to the RTOP at long diffusion times. Similarly, in

disordered media, the temporal scaling of the RTOP is necessary in the estimation of the

fractal dimension of the medium [3]. However, the RTOP is related to the pulsed field

gradient (PFG) NMR signal via an integration over the entire q-space. The unavailibility of

data at large wavevectors is a serious problem particularly in restricted domains where the

NMR signal does not attenuate significantly even at relatively large wavenumbers.

In principle, the extrapolation of the signal values can be performed by model fitting

to data. However, very different signal profiles are possible depending on the particular

specimen under examination whose structure is not known a priori. Another alternative is the

cumulant expansion, which may fail to converge to the true signal attenuation. Fig. 1a shows

the failure of both the cumulant expansion and sometimes used biexponential fitting in

describing the theoretical signal attenuation from rectangular pores at long diffusion times.

In this work, we propose to represent the PFG-NMR signal in terms of a complete set

of Hermite functions. The basis possesses many interesting properties relevant to q-space

NMR, such as the ability to represent both the signal and its Fourier transform. Unlike the

previously employed methods, this approach is linear and capable of reproducing

complicated signal profiles, e.g., those exhibiting diffraction peaks. The estimation of the

coefficients is fast and accurate while the representation lends itself to a direct reconstruction

of ensemble average propagators as well as calculation of useful descriptors of it, such as the

RTOP and its moments.

extrapolationregion

10 orderth

30 orderth

60 orderth

Fig. 2: PFG-NMR signal attenuation curves from rectangular pores at long diffusion times. (a) Cumulant

expansion of and biexponential fit to simulated data. (b) The performance of the proposed approach with the

same data. (c) Propagator reconstructed from the basis functions whose value at x=0 is the RTOP.

We performed simulations to assess the performance of the RTOP estimations for

one-, two- and three-dimensional, isotropic, restricted media as well as mono- and bi-

exponential decays and even the presence of flow was considered. RTOP values in 1-, 2- and

3- dimensions all yielded very accurate results (<6% error in restricted geometries, <1% in

others). The proposed approach is expected to improve the accuracy of RTOP estimates and

parameters deduced from it.

References:

1. M. D. Hürlimann, L. M. Schwartz, P. N. Sen, Phys. Rev. B. 51 (1995) 14936.

2. L. M. Schwartz, M. D. Hürlimann, K. Dunn, et al., Phys. Rev. E. 55 (1997) 4225.

3. E. Özarslan, P. J. Basser, T. M. Shepherd et al., J. Magn. Reson. 188 (2007) 285.

This research was supported by the intramural research program of NICHD.

116

Measurement and simulation of the non-local dispersion

tensor in porous media

M.W. Huntera, A.N. Jackson

b, P.T. Callaghan

a

a MacDiarmid Institute for Advanced Materials and Nanotechnology, Victoria University of

Wellington, Wellington, New Zealand., b SUPA School of Physics, University of Edinburgh, Edinburgh EH9 3JZ, Scotland, United Kingdom

Dispersion describes the phenomenon whereby particles on the same streamline separate during flow. The physics of dispersion is governed by stochastic processes arising from the interplay between advective velocity gradients, molecular diffusion and boundary layer effects1. The dispersion tensor, D*, is a local measurement in the sense that it does not depend on positional relationships and is measured as time asymptotes2. For situations where the length- and time-scales on which transport occurs are not much larger than the scale of the fluctuations in the velocity field, a non-local description is required3. The tensor is written as

. ( ) ( ) ( ) ( )*NL , ,0 ,P ,τ τ τ= +rD R u r R u r R

Pulsed Gradient Spin Echo (PGSE)-NMR provides a wealth of information about the velocity correlations in porous media. Presented here is a set of NMR pulse sequences and a superposition designed to extract the velocity correlations necessary to calculate the dispersion as a function of displacement and hence the non-local dispersion. The experimental method will be tested against the calculable non-local dispersion in a Couette cell4. Further experiments are performed on porous media over a range of Peclet number and further tensors will be discussed. The Lattice-Boltzmann algorithm has been shown to successfully predict the flow field in porous media5, and has been used to model the flow field through our model porous medium. This flow field is used to simulate a large ensemble of virtual tracer particles, from which numerical estimates for both the local and non-local dispersion can be determined. Our implementation of this approach is presented here along with a comparison between the experimental and computational results.

0 1 2 3 4

0

50

100

150

Simulated D*

NL

Z (mm)

D* zZ

z NL (

mm

/s2)

0 1 2 3 4

0

50

100

150

Measured D*

NL

Z (mm)

D* zZ

z NL (

mm

/s2)

Fig. 1: The non-local dispersion of flow through a beadpack, Pe = 2000, Vt = 8mm/s, τ = τv, dbead = .5mm.

1. A. A. Khrapitchev and P. T. Callaghan, Phys. Fluids, 15, 9 (2003) 2. J. D. Seymour and P. T. Callaghan, AIChE J. 43, 8 (1997). 3. D. L. Koch and J. F. Brady, J. Fluid Mech. 180 (1987) 4. M.W. Hunter and P. T. Callaghan PRL 99, 210602 (2007) 5. B. Manz et al., AIChE J. 45, 9 (1999).

117

Restricted self-diffusion in pure water measured by NMR

J. Stepišnika,b, I. Seršab, A. Mohoriča

aUniversity of Ljubljana, FMF, Physics Dept., bInstitute J. Stefan, Ljubljana Slovenia

The current thinking, influenced greatly by molecular modeling simulations1, is that on a very short pico-second time scale, water is more like a "gel" consisting of a single, huge hydrogen-bonded cluster. Rotations and other thermal motions cause individual hydrogen bonds to break and re-form in new configurations, inducing ever-changing local discontinuities whose extent and influence depends on the temperature and pressure. Over very small volumes, localized (H2O)n polymeric clusters may have a fleeting existence, and many theoretical calculations have been made showing that some combinations are stable enough to prolong their lifetimes. But it is considered that they does not remain intact long enough to be observed in ordinary bulk water at normal pressures.

Fig. 1: Molecular velocity autocorrelation spectra of pure water at different temperatures as measured by modulated gradient spin echo method. The constant spectrum value of non polar ethanol emphasizes the role of hydrogen bonding at the diffusion of water molecules

Thus, the measurements of the pure water self-diffusion by a new modulated gradient

spin echo method2 came as surprise giving the frequency dependence of the water velocity autocorrelation spectrum. The spectrum decrease in the low frequency range with almost linear dependence on ( )νD versus

ν1 plot resembles the restricted molecular motion.

Constant spectrum value of nonpolar ethanol in the low frequency range, as shown in Fig. 1., gives an indication about the water diffusion constrained by the hydrogen bond interaction.

.

References:

1. B. Chen, I. Ivanov, M.L. Klein, M. Parrinello, Hydrogen Bonding in Water, Phys. Rev, Letters 91.215503 (2003).

2. J. Stepišnik, S. Lasič, A. Mohorič, I. Serša, A. Sepe. Spectral characterization of diffusion in porous media by the modulated gradient spin echo with CPMG sequence. J. Magn. Reson., 182, 195-199 (2006)6.

118

0

3

6

9

12

15

0 15 30 45 60 75 90angle between fibers and B0 [°]

∆R

2 [

1/s

]

glass fiber

Dyneema

Nylon

0

0.1

0.2

0.3

0.4

0.5

0.6

0 0.2 0.4 0.6 0.8Fiber Density (FD)

fracti

on

al

an

iso

tro

py

(F

A)

.

simulation glass fibersimulation Dyneemasimulation Nylonglass fiberDyneemaNylon

0

1

2

3

4

5

6

7

8

0 0.2 0.4 0.6 0.8Fiber Density (FD)

R2

[s]

glass fiber

Dyneema

Nylon

Fig. 1: Fiber phantom: longitudinal

view (a) ,transversal view (b), fiber

tracking results (c), color-encoded

FA-map (d).

a

c d

b

Design of anisotropic fiber phantoms for the validation of diffusion

weighted magnetic resonance imaging

E.Fieremansa, Y.De Deene

a,b, S.Baete

a,b, I.Lemahieu

a

aMEDISIP, Ghent University-IBBT-IBiTech,

bQMRI, Ghent University Hospital, Belgium.

A gold standard for the validation of diffusion weighted

magnetic resonance imaging (DW-MRI) is essential for clinical

purposes but still not available. Anisotropic diffusion fiber

phantoms (Fig. 1) have been proposed recently for this purpose

[1]. A crucial question is how the different material properties

and size of the fiber phantoms influence the outcome of the DW-

MRI experiment.

Three potential fiber materials for constructing fiber phantom

were studied: Dyneema (∅ 16 µm), Nylon (∅ 64 µm) and glass

fiber (∅ 7 µm). Besides the fractional ansitropy (FA), the T2-

relaxation time and proton density of the fiber phantoms were

evaluated since they define the signal-to-noise ratio (SNR).

The fiber phantoms can be considered as two dimensional porous media with a matrix consisting

of parallel randomly packed cylinders and the theory for diffusion [2] and R2-relaxation [3] in

porous media was used as a model. In addition, the diffusion inside the fiber phantoms and the

effect of surface relaxation has been simulated using Monte Carlo simulations of random walk

and verified against experimental FA and R2-values. The fiber density and fiber diameter turned

out to be the major factors determining the FA (Fig. 2), while for the R2, the surface relaxation

(Fig. 3) and the magnetic susceptibility of the fiber (Fig. 4) are the most determinant. The most

appropriate fiber phantoms for the quantitative validation of DW-MRI are densely packed fiber

phantoms made from a hydrophobic material with a magnetic susceptibility close to water.

References:

1. E. Fieremans et Al., J. Magn. Reson. 190(2008) 189.

2. P.P. Mitra et Al., Phys. Rev. B Cond.Matt. 47(1993) 8565.

3. W.F. Slijkerman and J.P Hofman, Mag. Reson. Imag. 16 (1998) 541.

Fig. 2: Comparison between the

measured and simulated FA-values.

FA increases with increasing fiber

density and decreasing diameter.

Glass fiber and Dyneema have the

highest FA.

Fig. 3: R2 rates depend differently

on the fiber density due to the

different fiber surface relaxivities.

Dyneema has the lowest surface

relaxation (0.13 µm/s).

Fig. 4: The increase in R2 when

changing the angle between the

fibers and B0 depends on the

susceptibility of each fiber

materials. Nylon has the closest

susceptibility to water.

119

A numerical method to optimize presaturation sequences on multiexponential samples

M. Gombiaa, S. Sykorab, V. Bortolottic, E. Vacchellid, P. Fantazzinia

aDepartment of Physics, University of Bologna, Viale Berti Pichat 6/2, 40127, Bologna, Italy

bExtrabyte, Via Raffaello Sanzio 22c, 20022 Castano Primo (MI), Italy cDICMA, University of Bologna, Via Terracini 28, 40131, Bologna, Italy

dScuola di Specializzazione in Fisica Sanitaria, University of Bologna, Italy

We have developed a mathematical model to simulate the effects of pulse sequences on composite samples and coded it in Matlab. The object-oriented code accommodates pulse sequences and samples of any complexity, including ones with any distribution of relaxation rates and offsets. Of equal importance is the fact that the most common instrumental artifacts (B0 and B1 inhomogeneity) can be easily taken into account. The package permits to simulate the magnetization of a composite sample during the whole sequence by using Bloch equations to track the magnetization of each sample component. The software allows one to obtain three quality factors related to: the highest residual magnetization among all components (Q1); the square mean of all the residual magnetization components (Q2); the modulus of the total magnetization vector of the sample (Q3). Low values of Qi (i=1,2,3) indicate good zeroings of the sample residual magnetization. In particular, a low value of Q1 indicates good zeroing of all the sample residual magnetization components.

This approach has been applied to the problem of fast and efficient presaturation by a suitable Sample Magnetization Suppression pulse sequence (SMS) in the presence of a wide spread of offsets, relaxation rates, and magnetic field imperfections. This led us to the Logarithmically-distributed A-Periodic Saturation Recovery sequence (LAPSR) which comes as close as possible to suppressing the absolute magnetization of all sample components and is, in this respect, much better than the classical sequences Saturation Recovery (SR) and A-Periodic Saturation Recovery (APSR). LAPSR is characterized by the following pulse sequence:

P(αααα) – D – P(αααα) – D·q – P(αααα) – D·q2 – … – P(αααα) – D·qn-2 – P(αααα) – D·qn-1 – P(αααα) – τ – 90°, where D is the delay between the first two presaturating pulses and α the nutation angle of the magnetization vector for each pulse. We notice that delays between presaturating pulses decrease logarithmically by a factor q<1 to reach the value D·qn-1 between the last two pulses of the presaturating sequence. The code was also used both to optimize LAPSR parameters, i.e. the delay D, the angle α, the number n of presaturating pulses, the q-value, and to show the effects of field inhomogeneities.

The performance of LAPSR has been verified on a composite large-volume phantom.

120

Study of water permeability and dynamics in micelle, vesicle and

lipid bilayer systems using Dynamic Nuclear Polarization

Ravinath Kausik, Hongjun liang, Evan R. McCarney, Galen Stucky, Songi Han

Department of Chemistry and Biochemistry, University of California Santa Barbara, CA 93106

Materials Research Laboratory, University of California Santa Barbara, CA 93106

In nature water dynamics and content play a key role in the characteristics and functions

of micelle-vesicle and lipid bilayer systems. Experimentally these are difficult parameters to

access because the internal or bound water is difficult to discern from bulk water by conventional

spectroscopic methods. We use dynamic nuclear polarization (DNP) to amplify the 1H NMR

signal via electron spins that are in dynamic dipolar interaction with the 1H nuclei of water. Thus,

DNP detects amplified 1H NMR signal through polarization transfer from the spin label’s

unpaired electron giving us information about local water dynamics, while electron spin

resonance (ESR) directly detects the spin label’s signature and reports on the rotational diffusion

of the spin labeled molecular segment. Here we show how 1H DNP and ESR can give

complimentary information on water permeability, dynamics and chain mobility in vesicle

structures formed by Fatty acids (oleate) or cationic and anionic lipids respectively. Lipid bilayer

model systems with and without membrane protein Proteorhodopsin incorporated, or with DNA

spacers have also been investigated, by means spin labeles at head (16-PC Tempo) and tail (16-

PC Doxyl) groups.

Fig 1. The DNP (maximum enhancement) values show us how hydrated cationic and anionic vesicle systems are in

comparison with oleate vesicles. The rotational correlation times obtained by fitting the Cw X-band ESR spectra

give us an idea of the mobility of the chain segment in the vesicles.

References 1. Evan R. McCarney, Songi Han “Dynamic Nuclear Polarization Enhanced magnetic resonance

analysis as a unique tool to study local water dynamics in Micelle and Vesicle systems”

Langmuir submitted.

2. Hongjun Liang, Gregg Whited, Chi Nguyen, and Galen D. Stucky, Proc. Natl Acad. Sci. USA 104

(20), 8212-8217 (2007).

121

Enhanced Overhauser contrast in low field images of a gelling

process

Wilson Barros1

1 Harvard-Smithsonian Center for Astrophysics, Cambridge, MA, USA.

Magnetic Resonance Imaging (MRI) is a useful tool in material science applications. By

looking at the image profiles it is possible to assess aspects of the structure, the underlying

mechanisms of transport and chemical kinetics of processes. In MRI the signal amplitude

detected from the magnetized sample is known to be proportional to the polarizing magnetic

field. However, the MR signal can be enhanced by methods other than by increasing the field.

One example is dynamic nuclear polarization (DNP), in particular the electronic Overhauser

effect, whereby the polarization of a pool of unpaired electrons (a nontoxic free radical

dissolved in water) is transferred, via saturation of the electronic transition, to the water

molecules in the sample. It turns out that, in addition to the improved signal-to-noise ratio, for

systems exhibiting changes in water mobility there is a remarkable enhancement in the

contrast compared to that obtained by conventional T2-weighted imaging [1]. This

methodology offers accessibility to potential new effects [2] while using magnet design and

cost much more affordable.

T = 120 min. T = 128 min.

T = 15 min. T = 23 min.

Aqueous solution of sodium alginate

(3 % w/v )

Aqueous solution of CaCl2

(60 mM)

Semipermeable membrane

Fig.2 Gelling process of a sodium alginate

solution in the presence of calcium ions crossing

a semi-permeable membrane. Left column shows

images obtained via PEDRI for two different

gelling periods. The right column shows similar

intervals where the contrast originates only from

T2- weighting.

Fig.1 Sketch of the experiment: A cylindrical vial

containing a semi-permeable (permeable only to small

ions) dialysis membrane. The membrane contains a

dilute solution of sodium alginate that is immersed

into the vial containing CaCl2. When the CaCl2

crossed the membrane it reacts with the alginate

forming the gel.

Although gels exhibit solid-like characteristics on a macroscopic level, microscopically they

behave in a liquid-like fashion suitable for MRI investigation. Indeed, a considerable

variation in water mobility is expected during the formation of gel systems. In order to

demonstrate the enhanced mobility-contrast sensitivity generated via the Overhauser effect,

we compare images of the formation process of alginate hydrogels obtained under the same

experimental conditions by a technique named Proton-Electron Double-Resonance Imaging

(PEDRI) [3] and by conventional T2-weighted imaging.

1. W. Barros and M. Engelsberg, J. Magn. Reson 184, 101 (2006).

2. W. Barros and M. Engelsberg, Phys. Rev. E 67, 021905 (2003).

3 D. J. Lurie et al., J. Magn. Reson. 76, 336 (1988).

122

MRI with Hyperpolarized 3He and

129Xe at Very-Low-Field

Ross W. Mair1, Wilson Barros

1, Rachel Scheidegger

1,2, Daniel Chonde

1,2,

Matthew S. Rosen1,3

, Ronald L. Walsworth1, 3

1 Harvard-Smithsonian Center for Astrophysics, Cambridge, MA, USA.

2 Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA, USA.

3 Department of Physics, Harvard University, Cambridge, MA, USA.

Studies employing hyperpolarized spin samples offer the possibility of high-sensitivity NMR

and MRI at very low fields. Benefits include significantly reduced effects from

susceptibility-induced background field gradients, often resulting in significantly longer T2

*;

and a greater skin depth that allows imaging inside conductors [1]. Without hyperpolarized

samples, studies have been performed in the earth’s field, partly for some of the above

reasons [2,3]. For lossy samples, such as human subjects, additional benefits can be

recognized. We have designed and built an open-access, very-low-field human MRI system

that allows the study of pulmonary function with subjects in a variety of postures, an

important advance in lung physiology. Despite the lower B0 of our system, certain

measurements that rely on accurate knowledge of B1 homogeneity and RF flip-angle

uniformity are simplified by operation at very-low-field.

T2

* Contribution B0 = 1.5 T B0 = 6.5 mT

T2 (s) 10 10

T2, inhomogeneity (s) 0.143 0.016

T2, susceptibility (s) 0.0046 1.10

T2, diffusion (s) 0.031 1672

T2, gradients (s) 0.512 3.6

T2

* (ms) 3.8 15.6

Fig 1: Measured SWR as a function of

frequency for our human chest coil built for 3He

imaging at 210 kHz, when unloaded, and loaded

by two different human subjects. The coil

resonance shifts slightly, but coil Q is unaffected

by subject loading.

Table 1: A comparison of T2

* contributions at 1.5 T and

6.5 mT, along with the typical total value of T2

*. A value

of T2 = 10 s is used in both cases [4], while T2,inh 0.143

s is used for 1.5 T [5]. All other terms are calculated

from known values. 3He T2

* is based on our measured

B0 0.4 G, and 0.1 G for a 1.5 T scanner [6].

Operation below the “sample-noise dominance” regime means that our RF coil loads

negligibly and reproducibly, regardless of subject. Coil Q is unaffected (Fig. 1). Therefore,

the excitation flip-angle can be calibrated in advance, rather than being incorporated into

every functional human lung study, as is done at high fields. Additionally, our calculations

show that T2

* contributions from susceptibility differences and diffusion through

inhomogeneous background fields are negligible, and these effects combine to produce a T2

*

significantly longer than measured at high field (Table 1). As hyperpolarized liquid species

gain popularity, similar benefits could be realized in many biomedical and materials studies.

1. C.-H. Tseng et al, Low-field MRI of laser polarized noble gas. Phys Rev Lett, 81, 3785 (1998).

2. P. T. Callaghan et al, An earth’s field nuclear magnetic resonance apparatus suitable for pulsed gradient spin

echo measurements of self-diffusion under Antarctic conditions, Rev Sci Instrum, 68, 4263 (1997).

3. A. Mohoric et al, Self-Diffusion Imaging by Spin Echo in Earth’s Magnetic Field, J Mag Res, 136, 22 (1999).

4. L. Darrasse et al, Low-field He-3 nuclear magnetic resonance in human lungs. CR Acad Sci, 324, 691 (1997).

5. J. Parra-Robles et al, Theoretical signal-to-noise ratio and spatial resolution dependence on the magnetic field

strength for hyperpolarized noble gas magnetic resonance imaging of human lungs. Med Phys, 32, 221 (2005).

6. L. Tsai et al, Posture-Dependent Human 3He Lung Imaging in an Open Access MRI System: Initial Results,

Acad Radiol, in press (2008).

123

Accumulation of NMR Data in Polar Coordinates and Numerical Methods to minimize Systematic Errors

Gianni Ferrante1, Paolo Golzi2, Davide Canina1

Giuseppe Martini2, Carla Vacchi2

1 Stelar s.r.l., Via Enrico Fermi 4, Mede (PV), Italy 2 Department of Electronics, University of Pavia, Via Ferrata 1, 27100 Pavia, Italy

Abstract: In this work we present new instrumental concepts and new NMR methods based on the acquisition and the accumulation of the signal magnitude. The method has been developed to overcome field stability limits and allow the acquisition of Hahn echo decays in Fast Field Cycling Technique

Nevertheless the accumulation of signal magnitudes of NMR data in low signal-to-noise ratio (SNR) regimes produces signal-dependent noise bias that reduces the accuracy of the measurements. This is particularly important for T1 and T2 measurements when the relaxation curve are fitted by a multi-exponential acquisition from polar coordinates. In fact the white noise, with a Gaussian distribution, that affect the NMR signal, become a Rician noise, characterized from the Rayleight distribution, on the accumulated signal magnitudes resulting in a systematic error.

We have developed the correction scheme on a reconfigurable digital hardware architecture, to apply the factor directly over the data-flow on our acquisition channel

Further developments are focused on the improved hardware implementation of numerical methods used to calculate the correction factor. We have also proposed a different technique in order to better evaluate SNR dependencies and extract the real signal intensity from noisy magnitude MNR signals. This alternative technique presumes the possibility of measuring the power of the noisy signals.

Several numerical approaches have been proposed in literature to de-noising NMR signals including a non-linear and wavelet transform. These methods are suitable to be use in a post-processing data evaluation. Used in conjunction with our hardware correction technique, become a complete tool to compensate and/or remove noise from digitalized NMR signals.

1. Cheng Guan Koay, Peter J. Basser, Journal of Magnetic Resonance 179 (2006) 317-322. 2. S.O. Rice, Bell System Technical Journal - Mathematical Analysis of Random Noise 23-24 (1944).

124

Delayed Focus-Pulses in Projection Reconstruction

Joshua Varner, Benedict Newling

MRI Center, Departments of Physics, University of New Brunswick, Fredericton, NB, Canada

With the advent of newer and more robust MRI techniques, Projection

Reconstruction (PR) has become a somewhat retired method due to the probe dead time plaguing the gross structure of its images. In fact, virtually all FID experiments suffer from the same dead time, which effectively ignores the first few microseconds of any signal, the first few crucial k-space data points likewise in MRI. Dead time aside, the ability to rapidly sample an FID signal on a polar k-space raster while monitoring its progress makes Projection Reconstruction in principle a very worthy contender for imaging short T2* species, including porous media and fluid flow measurements. In this talk, the Delayed-Focus Pulse [1] (DFP) will be introduced as a potential cure to the dead time plague. The DFP is so-called due to its phase-refocusing property that causes the signal to achieve phase coherence at a time tf after the initial pulse based entirely on the customizable shape of the pulse itself. Combining the concepts of PR and the DFP, a new imaging sequence will be discussed and compared with other well-practiced techniques such as the Double-Half k, RARE and SPRITE. Understanding where this new method stands will be a step forward in hopes that its uniqueness has a place in materials MRI.

References: [1] X.-L. Wu, P. Xu, R. Freeman, Delayed-Focus Pulses for Magnetic Resonance Imaging: An Evolutionary Approach, M. R. Med 20, (1991) 165-170

Fig: A collection of four Delayed-Focus Pulses whose delay

times tf are listed in the legend below.

125

Optimal k-space Sampling for Single Point Imaging of Transient

Systems

P. Parasoglou1, A.J. Sederman

1, J. Rasburn

2 and H. Powell

2 and M.L. Johns

1

1MRRC, Department of Chemical Engineering, University of Cambridge, Cambridge, UK

2Nestlé Product Technology Centre, York, UK

Magnetic Resonance Imaging has been used to visualise moisture absorption from humidified

air, and subsequent diffusion of this absorbed water, in initially dry samples of food wafer.

Due to the high porosity of the wafer (typically 80 – 90 %) and the typically low moisture

content (of the order of 2-7 wt %), this is particularly challenging. The wafer also exhibits

very short spin-spin relaxation constants, T2* (typically of the order of 50 µs) and T2.

Consequently conventional frequency encode techniques are precluded due to the

unfavourable signal relaxation properties. Thus Single Point Imaging (SPI) [1; 2] is the main

MRI technique used in the current study.

To overcome the main drawback of SPI, its relatively long acquisition time, a novel approach

or trajectory for sampling k-space is presented (the relatively long T1 of the absorbed moisture

excludes the use of SPRITE [3]). The sequence is optimised with respect to the achievable

Signal-to-Noise ratio (SNR) over a given time interval via selective sparse k-space sampling,

the trajectory of which is dictated by prior knowledge of the overall object shape. Further

modest improvements in image quality are also shown via the use of complete sampling of k-

space at the start or end of the series of imaging experiments (when imaging time is not

limited by system dynamics), followed by subsequent use of this data for un-sampled k-space

points for intermediate images, as opposed to zero filling. This technique was successfully

applied to follow the moisture absorption process by the wafer. Good image reproduction

was achieved with significant under-sampling of k-space. Error is shown to be significantly

reduced relative to alternative sampling trajectories.

References:

[1] S. Emid, Ultra high-resolution multiple quantum spectroscopy in solids. Physica B & C

128 (1985) 79-80.

[2] S. Emid, and J.H.N. Creyghton, High-resolution imaging in solids. Physica B & C 128

(1985) 81-83.

[3] B.J. Balcom, R.P. MacGregor, S.D. Beyea, D.P. Green, R.L. Armstrong, and T.W.

Bremner, Single-point ramped imaging with T-1 enhancement (SPRITE). Journal of

Magnetic Resonance Series A 123 (1996) 131-134.

126

A new theoretical insight on time-dependent diffusion coefficient

D. S. Grebenkov

LPMC, CNRS – Ecole Polytechnique, F-91128 Palaiseau, France

In NMR, measuring the time-dependent diffusion coefficient D(t) is an efficient tool to probe the geometry of porous media1,2,3. Although the diffusive motion is well understood in single-scale domains (slab, cylinder, and sphere)4, many issues remain unclear for multi-scale porous structures like sedimentary rocks, cements, or biological tissues. To get a better theoretical insight onto restricted diffusion in multi-scale geometries, we study the spin-echo signal attenuation due to diffusion in circular and spherical layers, x∈Rd : L–l<|x|<L, presenting two geometrical lengths, the radius L and the thickness l. Since the Laplace operator eigenbasis is known explicitly, many analytical results can be derived, in particular, the closed form for the time-dependent diffusion coefficient,

D(t)/D = ∑k=0∞ λ1kB

200,1k w(D t λ1k/L

2), where D is the free diffusion coefficient, λnk the Laplace operator eigenvalues (here, only λnk with n=1 are involved), λ1kB

200,1k the explicit weighting coefficients5. The function w(p) is

determined by the temporal profile (or waveform) of the applied magnetic field gradient, e.g., w(p) = (1/p2 – (exp(-p) – 4exp(-p/2) + 3)/p3)/12

for the simple bipolar gradient (two rectangular pulses of duration δ=t/2). For thin layers (l << L), a perturbative analysis gives surprisingly accurate results, e.g. λ10 ≈ 1 and λ1k ≈ π2k2(L/l)2 for circular layers. The “gap” between λ10 and λ11 is bigger for larger separation between two geometrical lengths L and l. A new, intermediate diffusion regime emerges for l << (2Dt)1/2 << L, when the echo time t is long enough for particles to travel between two boundaries, but still insufficient for exploring the whole domain. This diffusion time t appears as an experimental “yardstick” for probing geometrical lengths of the confinement. This intermediate regime resembles the tortuosity regime in porous media.

. In conclusion, we considered restricted diffusion in model two-scale geometries. In a

single mathematical frame, we observed the transition from the slow diffusion to a new intermediate regime (analogous to the tortuosity regime in porous media), and then to the fast diffusion. These features should appear and be relevant for natural multi-scale structures.

References: 1. P. N. Sen, Concepts Magn. Reson. 23 A, 1-21 (2004). 2. P. P. Mitra, P. N. Sen, L. M. Schwartz, P. Le Doussal, Phys. Rev. Lett. 68, 3555 (1992). 3. T. M. de Swiet, P. N. Sen, J. Chem. Phys. 104, 206 (1996). 4. D. S. Grebenkov, Rev. Mod. Phys. 79, 1077-1137 (2007). 5. D. S. Grebenkov, J. Chem. Phys. (2008, in print).

Fig. 1: Time-dependent diffusion coefficient D(t) for thick (l/L = 0.5, dashed blue line) and thin (l/L=0.1, solid red line) circular layers. For thick layer, the scale window 0.25 << Dt/L2 << 1 (shown by vertical dotted lines) is too narrow, so that a mere transition between slow and fast diffusion is observed, as for the disk. For thin layer, the scale window 0.01 << Dt/L2 << 1 is large enough to reveal the new intermediate regime. At this time and length scales, restricted diffusion in a two-dimensional thin layer is effectively one-dimensional so that the apparent diffusion coefficient is reduced by factor 2, a kind of “tortuosity” of the thin layer.

127

Quantitative NMR microscopy of iron transport in methanogenic

aggregates

F. Vergeldta, J. Bartacek

b, E. Gerkema

a, M. Osuna

c, J. Philippi

a, P. Lens

b,d and H. Van As

a

aLaboratory of Biophysics and Wageningen NMR Centre, and

bSub-department of

Environmental Technology, Wageningen University, The Netherlands, cChemical and

Environmental Engineering Department, University of Basque, Bilbao, Spain, dPollution

Prevention and Control core, UNESCO-IHE, Delft, The Netherlands

Up-flow Anaerobic Sludge Bed (UASB) reactors are the most widely applied type of

anaerobic waste water treatment reactors. The actual purification process takes place in

methanogenic granular sludge that consist of rigid, well settling microbial aggregates that

developed by the mutual attachment of bacterial cells in the absence of a carrier material. Heavy

metal transport into these inhomogeneous, spherical biofilms is an important process both from

the point of view of toxicity and bioavailability. Detailed experimental data on metal transport

and biosorption within biofilm matrices is lacking due to the absence of adequate, non-invasive

analytical tools, hindering refined modelling of these complex processes.

In this study 3D spatially resolved concentration mapping of iron as an environmental

relevant, paramagnetic model metal is aimed for under strict requirements for spatial and

temporal resolution. Although 3D T1 mapping would be the most direct method, this approach

does not fulfil these requirements. Mohoric et al. demonstrated the application of centred-out 3D

RARE imaging to follow the highly dynamic process of cooking of a single rice kernel in a

spatially and temporal well resolved manner (1). This technique was combined with the approach

demonstrated by Nestle et al. to use image contrast to map metal concentration in alginate gels

(2). Due to the short effective echo-time contrast is based on changes in T1 as a function of metal

concentration. Recently Graf von der Schulenburg et al. presented a less sensitive 2D approach

using T2 contrast to map cobalt concentrations in a model biofilm reactor (3).

The reaction-diffusion process of iron in two types of methanogenic granules from

different UASB reactors was investigated. The inner structure of the different granule types

revealed by 3D T2 maps varied from a homogeneous matrix to a heterogeneous structure with

voids and cracks. Detailed analysis of the 3D RARE datasets resulted in quantitative

concentration profiles of iron inside the granules with a radial resolution of 200 µm and a time

resolution of 11 minutes. Although iron transport into the granules is facilitated by presence of

channels, the actual iron penetration into the matrix is mainly determined by the overall water

content and diffusivity of the matrix.

References:

1. Mohoric, A., Vergeldt, F., Gerkema, E., de Jager, A., van Duynhoven, J., van Dalen, G. and

Van As, H., Magnetic resonance imaging of single rice kernels during cooking. J. Magn.

Reson. 171 (2004) 157.

2. Nestle, N. and Kimmich, R., NMR microscopy of heavy metal absorption in calcium alginate

beads. Appl. Biochem. Biotechnol. 56 (1996) 9.

3. Graf von der Schulenburg, D., Holland, D., Paterson-Beedle, M., Macaskie, L., Gladden, L.

and Johns M., Spatially resolved quantification of metal ion concentration in a biofilm-

mediated ion exchanger. Biotechnol. Bioeng. 99 (2008) 821.

128

Asphaltene aggregation in crude oil studied by low field

relaxation and diffusion measurements L. Zielinski

a, D. Freed

a, M. D. Hürlimann

a, I. Saha

a,b

aSchlumberger-Doll Research, 1 Hampshire Street, Cambridge, MA 01239

bSouthern Illinois University, Department of Chemistry and Biochemistry,

Carbondale, IL 62901

We used low-field NMR bulk relaxation and diffusion experiments to study

aggregation properties of asphaltenes in crude oils. Asphaltenes are of particular interest in

the oil industry, where their precipitation may clog up permeable pores and pipelines, causing

huge losses at various production stages. Though many methods exist to study asphaltene

properties, NMR has the potential to measure the asphaltenes in situ, downhole in their native

environment, at the earliest stages of oil exploration.

We studied three sets of samples with varying asphaltene concentration at

temperatures ranging from 45ºC to 92ºC and at Larmor frequencies of 1.7 MHz and 5 MHz.

Asphaltene was first removed from the crude oil by heptane precipitation and then mixed

back in different proportions, resulting in concentration range from 0% to 11%. Increasing

asphaltene contents enhanced the relaxation, Fig.1(a), while leaving diffusion properties

largely unchanged. T2 relaxation was affected more strongly than T1, consistent with the

presence of large asphaltene clusters and resulting slowed correlation times. We found that

only maltene (deasphalted crude) protons were observed, as asphaltene relaxed faster than

our detection window. All the extra relaxation seen in Fig. 1(a) comes from maltene

interactions with asphaltene clusters. Asphaltene, therefore, acts as a contrast agent with

respect to the solvent maltene, and its effects are related to its aggregation state.

Fig. 1: (a) CPMG for seven samples with the same base oil and

asphaltene concentrations from 0% (sample #1) to 11% (sample #7).

(b) The same samples with time axis rescaled by constant ξ2’s (see

legend). Inset shows the range from panel (a) for easy comparison.

Remarkably, the data for different asphaltene concentrations, both for T2 (see Fig. 1(b)) and

for T1 relaxation (not shown), collapses onto a single decay curve upon rescaling the time

domain. This is a highly nontrivial phenomenon, given the multi exponential nature of the

relaxation. We have developed a simple model, based on interactions between asphaltene

clusters and maltene chains, that relates the scaling coefficients ξ2 and ξ1 (corresponding to

T2 and T1 relaxation, respectively) to the size of asphaltene clusters. Applying our formulas

to the present data yielded cluster sizes of the order of 4 nm, consistent with other methods

for assessing asphaltene aggregate sizes.

129

Chemical Resolution in T1-T2 correlations

T. C. Chandrasekeraa, J. Mitchell

a, E. J. Fordham

b, L. F. Gladden

a, and M. L. Johns

a.

a Department of Chemical Engineering, University of Cambridge, Pembroke Street, Cambridge,

CB2 3RA, UK. b

Schlumberger Cambridge Research, High Cross, Madingley Road, Cambridge, CB3 0HG, UK.

T1-T2 correlations are useful for determining the ratio T1/T2 as a characteristic parameter

for permeable rocks [1]. However, the presence of both oil and water can complicate the

analysis. Sedimentary rocks are therefore analyzed with D-T2 correlations at low magnetic fields

because of the distinct difference in diffusion coefficient D of the two liquids [2]. Here we

present a novel T2-T1-δ correlation [3] where the chemical shifts δ allow the water and oil

fractions to be identified; see Fig. 1.

Fig. 1. Chemically resolved T2-T1 correlations reveal the water and oil fractions within

a sandstone rock core.

The T2-T1-δ correlation is acquired in the same experimental time as a conventional T1-T2

correlation. This is achieved through the use of a “double-shot” measurement of T1 [4], so-called

because it requires the spin ensemble to be stored “up” and “down” in successive scans [5]. In

this way the longitudinal relaxation information is encoded in an exponential decay (rather than a

recovery). Since this T1 measurement is based on the acquisition of multiple FIDs, chemical shift

information can be extracted.

Furthermore, we propose a novel T1-T1 exchange measurement and demonstrate this by

determining exchange between cellulose fibers at high magnetic fields where the susceptibility

induced field gradients distort the traditional T2-T2 exchange measurements.

References:

1. Y. Q. Song, et al., J. Magn. Reson. 154 (2002) 261.

2. M. D. Hürlimann, et al., J. Magn. Reson. 157 (2002) 31.

3. T. C. Chandrasekera, et al., J. Magn. Reson. To be submitted.

4. J. J. Hsu and I. J. Lowe, J. Magn. Reson. 169 (2004) 270.

5. E. E. Sigmund, et al., Solid State Nucl. Magn. Reson. 29 (2006) 232.

130

withdrawn

131

NMR diffusion and relaxation studies of molecules confined

inside core-shell polymeric capsules

I. Ardeleana, M. Bogdan

b, C. Badea

a, A. Parnau

b

aTechnical University from Cluj-Napoca, Department of Physics,

400020 Cluj–Napoca, Romania; bNational R&D Institute for Isotopic and Molecular

Technologies, 400293 Cluj–Napoca, Romania

Understanding the surface effects on dynamics of molecules confined in nanocapsules

is important both for obtaining of theoretical scientific knowledge and for designing of new

nanocapsules to be used in controlled drug delivery [1]. Obtaining of information about the

surface effects on confined molecules and the molecular exchange through capsules

membrane can be favorably accomplished using nuclear magnetic resonance diffusometry

and relaxometry techniques [2]. Surface interactions give rise to orientational order near the

surface with the molecules diffusing from the isotropic bulk region to the ordered surface

layer and vice versa. At each return the molecules orientation is controlled by the interaction

with the surface. Consequently, de dipolar correlation function is characterized by a slowly

decaying part owing to the repeated returns of molecules to the surface. This mechanism of

relaxation that is based on “reorientations mediated by translational displacements” (RMTD)

[3] is the subject of our investigations.

In our contribution we will report diffusion and relaxation studies performed on

miglyol and hexadecane molecules encapsulated in polymeric nanocapsules of different

dimension. The core-shell polymeric capsules were prepared by two different methods based

on emulsification – diffusion technique [4]. The diffusion experimental data obtained with a

high field instrument (9.4T) using the fringe field stimulated echo (FFStE) technique [5] were

compared with those obtained with a low field instrument (0.47 T) by using the well known

pulsed field gradient stimulated echo (PFGStE) technique. The proton relaxation data were

obtained at 20 MHz using the CPMG technique. The apparent diffusion coefficient extracted

from the first part of echoes decay allowed us to evaluate the capsules dimension. The

coincidence with the dimension observed from electron micrograph is remarkable. In order to

obtain a correct interpretation of the experimental data these were also compared with Monte

Carlo simulations.

This work was financed by the CEEX – MATNANTECH 58/2006 project. We thank Prof.

Rainer Kimmich from Ulm University for allowing us to perform the FFStE measurements.

References:

[1]. V.R. Kumar, J.Pharm. Pharmaceut Sci. 3, 234(2000)

[2]. C. Mayer, Prog. Nucl. Magn. Reson. Spectroscopy, 40, 307(2002)

[3]. T. Zavada and R. Kimmich, J. Chem. Phys. 109, 6929(1998)

[4]. D.Quintanar, H.Fessi, E.Allemann, and E.Doelker, Int.J.Pharm.143, 133(1996)

[5]. G. Farrher, I. Ardelean, and R. Kimmich, J. Magn. Reson. 182, 215 (2006)

132

MRI Evaluation of the Effectiveness of Colloidal Dispersion Gels for Enhanced Oil Recovery

B. Babalola,a L. Li,b L. Romero-Zeron,a B. Balcom, b

aDepartment of Chemical Engineering, University of New Brunswick, bDepartment of

Physics, University of New Brunswick

Sweep efficiency in mature or marginal conventional oil fields is significantly affected by fluid channeling through rock cracks, fractures, and high permeability “thief zones”. Colloidal dispersion gels (CDG) have been proposed as an enhanced oil recovery (EOR) method to block deleterious fluid channeling in petroleum reservoir formations. Controversy exists in the literature however about the effectiveness of CDG.

Centric-scan SPRITE was employed to visualize the behavior of CDG in controlling fluid channeling in laboratory studies with core plug samples. The Centric-scan SPRITE method provides quantitative fluid density images in porous media that are difficult to image by conventional MRI techniques due to short transverse signal lifetimes and multi- exponential T1, T2 behavior.

The experimental results showed CDG is an effective technique for controlling conformance problems in porous media. Centric-scan SPRITE MRI permitted accurate monitoring of displacement fronts between displaced and displacing fluids through the porous matrix and high permeability channel.

133

Characterisation of pulsing flow in trickle-bed reactors

using ultra-fast MRI

T. T. M. Nguyen, A. J. Sederman, L. F. Gladden,

Departments of Chemical Engineering, University of Cambridge, Pembroke Street,

Cambridge CB2 3RA, UK

Trickle-bed reactors (TBRs) are widely used in petroleum refining and throughout the

chemical industry. TBRs take the form of columns packed with solid porous catalyst pellets.

The most common mode of operation is with concurrent down-flow of liquid and gas

reactants. TBRs can therefore be thought of as porous systems on two different length-

scales; a macro-pore structure formed by the void space between the catalyst particles, which

interacts with the micro-pore system of the catalyst pore structure. It has been suggested that

operation of TBRs at high liquid and gas flow rates in the so-called pulsing regime, can

improve the reactor performance in terms of mass, heat transfer rates and better catalyst

utilisation [1, 2]. However, an understanding of the hydrodynamics within the reactor

operating under these conditions has not yet been achieved. In this work, we use ultra-fast

MRI techniques to characterise hydrodynamics during pulsing flow and compare these results

with those obtained using conductance measurements – a technique well established for this

application within the chemical engineering community [3].

This paper presents the results of ultra-fast one-dimensional (1-D) Fast Low Angle

SHot (FLASH) MRI to characterise the liquid pulses in TBRs. 1H MRI experiments were

performed at 200 MHz on a 4.7 T magnet. 1-D liquid holdup profiles along the flow

direction were acquired at spatial and temporal resolutions of 352 µm pixel-1

and 3.3 ms,

respectively. It is demonstrated that properties such as pulse velocity, duration and frequency,

obtained with MRI are in good agreement with the conductance data. MRI is shown to have

three advantages compared to the traditional conductance measurement approach:

• Individual pulses are tracked and therefore the pulse velocity measured with MRI is

more accurate when only part of TBR is pulsing.

• The distribution of pulse velocities is obtained as opposed to a time-averaged

measurement.

• Since individual pulses are tracked, splitting of a single stream into two, and merging

of two streams into one are imaged directly. Such information cannot be obtained

using the conductance approach.

References:

1. Y. Lemay, G. Pineault, J.A. Ruether. Ind. Eng. Chem. Proc. Des. Dev. 14 (1975) 280.

2. N. A. Tsochatzidis, A. J. Karabelas. J. Appl. Electrochem. 24 (1994) 670.

3. J. G. Boelhouwer, H.W. Piepers, A. A. H. Drinkenburg. Chem. Eng. Sci. 57 (2002) 4865.

134

Study of Miscible and Immiscible Flows in a Microchannel Using

Magnetic Resonance Imaging

B.S. Akpaa, S.M. Matthews

b, A.J. Sederman

b, K. Yunus

b, A.C. Fisher

b,

M.L. Johnsb, and Lynn F. Gladden

b

aDepartment of Chemical Engineering, University of Illinois at Chicago,

bDepartment of

Chemical Engineering, University of Cambridge

Microfluidic devices have attracted much interest in the fields of biology,

biotechnology, and analytical and synthetic chemistry with applications as varied as protein

crystallization, analyte diagnostics, cytometry, and combinatorial chemistry. These

miniaturized fluidic systems have many advantages over their macroscale equivalents and

have made feasible the integration of multiple processes on one device – the so-called lab on

a chip or micro total analysis system. Many attempts have been made to characterize mixing

performance in microfluidic systems with a view to optimizing their design and operation.

Both flow and concentration mapping have typically been achieved by using optical methods.

Despite the achievements of workers using optical techniques, there remain some inherent

limitations of these methods. For example, the applicability of optical methods is limited to

systems that have been fabricated and sealed with optically transparent materials. Optical

methods are also often limited with respect to the type of device geometry that can be

studied. More recently, researchers have begun to explore MRI as a tool for the study of

microscale systems1,2,3

.

In this work, MRI has been used for the first time to obtain both cross-sectional

velocity and concentration maps of flow through an optically opaque Y-shaped microfluidic

sensor. Images of 23 µm × 23 µm resolution were obtained for a channel of rectangular cross

section (250 µm × 500 µm) fed by two square inlets (250 µm × 250 µm). Both miscible and

immiscible liquid systems have been studied. These include a system in which the coupling

of flow and mass transfer has been observed, as the diffusion of the analyte perturbs local

hydrodynamics. Our motivation for

developing MRI tools for the study of

microchannels extends beyond

experimental visualization of mixing

performance. The data thereby obtained

can be used in the validation of

numerical codes for simulating

micromixing processes. Such codes

will play a key role in the design and

optimization of microfluidic systems.

Images presented in this work have

been used by Sullivan et al.4 to validate

a lattice-Boltzmann code capable of

simulating coupled diffusive mass

transfer and hydrodynamics in 3D.

References:

(1) C. Hilty et al. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 14960. (2) E. Harel et al. Phys.

Rev. Lett. 2007, 98, 017601. (3) S. Ahola et al. Lab Chip 2006, 6, 90. (4) S.P. Sullivan et al.

Sens. Actuators, B: Chem. 2007, 123, 1142.

(A)

(B)

Low

Intensity

High

Intensity 0 43 mm s-10 43 mm s-1

Velocity

(C)

(D)

Fig. 1: Images of multiphase, immiscible flow

through the microchannel. Selective saturation was

used to image either the aqueous or organic phase

independently. (A) Organic phase. (B) Aqueous

phase. (C), (D) Velocity maps of the organic and

aqueous phases respectively.

135

Velocity and diffusivity mapping of water flow in a bead

pack by a purely phase-encoded MRI method Zhi Yang, Olusegun Adegbite, Joshua Varner, Ben Newling

MRI Center, Departments of Physics, University of New Brunswick,

Fredericton, NB, Canada

In the past years, magnetic resonance imaging (MRI) has been proved to be a valuable method for measuring flow properties. In this talk, we present a motion sensitized version of the MRI method SPRITE (Single Point Imaging with T1 Enhancement [1]) measuring fast, mechanically dispersed water flow in a pack of large glass beads (Velocity 20 ~40cm/s, Reynolds number up to 2000, Average bead size 1.6cm). This method is quantitative, three dimensional and has potential time saving over traditional pointwise techniques for measuring fluid flow. Use of a purely phase-encoded method means that the encoding interval can be kept very short, allowing greater control of intravoxel phase dispersion than many frequency-encoded techniques. The SPRITE family of techniques is essentially immune to artifacts caused by material boundaries in magnetic susceptibility. The mean velocity and turbulent diffusivity of water flow were mapped. SPRITE has great potential for future study in dispersed fluid flow and CFD (Computational Fluid Dynamics) software validation.

References:

[1] B. Newling, C.C. Poirier, Y. Zhi, J.A. Rioux, A.J. Coristine, D. Roach, B.J. Balcom, Velocity imaging of highly turbulent gas flow, Phys. Rev. Lett. 93 (2004) 154503.

Fig: 2D slice of 3D spin density image of the bead pack, spatial resolution 64 by 64 pixels

136

Touching-Vug velocity and exchange with matrix pores in models and samples from the Biscayne aquifer

L. Floreaa, S. Altobellib, K. Cunninghama

aU.S. Geological Survey, Ft. Lauderdale FL, bNew Mexico Resonance, Albuquerque, NM

We used stimulated-echo based velocimetry to measure three-dimensional steady velocity fields in the centimeter scale vugs and short echo-time spin warp imaging to study the washout of D2O from the porous matrix. Washout experiments were done under static and flowing conditions.

The Biscayne aquifer in southeast Florida is one of the most permeable aquifers in the world.1 The large scale macroporosity is due in part to shrimp burrows.2 As part of a combined experimental and computational effort aimed at modeling the aquifer on many scales, we investigated the feasibility of some simple MR techniques. Figure 1 is an example that shows the almost immediate washout of D2O from the macropores and the exchange that occurs with the water in the rock matrix over a longer time scale.

The experiments were done in a 1.89T horizontal bore system controlled by a Tecmag (Houston, TX) Libra. The 3d model used in the velocity measurements was constructed by a rapid-protoyping 3d epoxy printer. The evolution time in the stimulated echo experiments was approximately 1s, T1 was 2.7s, flow rates were set to match the superficial velocity in the aquifer, resulting in average velocities of 0.045 cm/s and 0.025 cm/s. The higher flow rate was used in the washout experiment shown in Figure 1.

Our results show that MR imaging is a promising technique for studying flow fields and transport properties pertinent to the Biscayne aquifer.

Fig. 1: MR images, FOV 8x8 cm, taken periodically show the incursion of water into an initially D2O saturated carbonate rock (ML-3) from an outcrop near Biscayne Bay.

References:

1. Fish, J. E. and Stewart, M., USGS Water-Resources Report 90-4108, 19912. Cunningham, K. J. and Curran, H. A. Int. Ichnofabrics Workshop IX, 2007

137

withdrawn

138

Viscosity Prediction in Polymer Melts using low field NMR Relaxation

Benjamin Nicot1,2, Marc Fleury1, Jacques Leblond2, Madeleine Djabourov2

1 Institut Français du Pétrole, Rueil Malmaison, France 2 Laboratoire de Physique Thermique, ESPCI, Paris, France

For many viscous liquids, NMR relaxation techniques can be used to estimate viscosity. However, the usual semi-empirical correlations fails in polymer melts due to their specific molecular dynamic. We present a combined viscosity-NMR study on a model system composed of linear polymers of increasing molecular weight (Polyethylene glycol, PEG). We show that the system studied behaves as expected for cross-linked polymers1. Viscosity, longitudinal and transversal relaxation times T1 and T2, and diffusivity clearly show the entanglement transition at Mc∼3000 g/mol and the expected power law dependences are verified for viscosity (η∝Mw

3) and diffusivity (D∝Mw

-2). However, for polymer melts, the standard viscosity prediction from T2 relaxation data (T2∝(η/T)-1) is not valid. In addition, T1 does not vary with viscosity when varying the molecular weight although it varies with temperature. We interpret these experimental data in the framework of a two correlation time model and we used the spectral density function proposed by Lipari and Szabo2. The T1 and T2 viscosity and temperature dependence is due to a different sensitivity to fast (local) and slow (global) motions of the molecules. The different time scales of the system can be deduced from the measured quantities D, T1, T1ρ, T2. In the present case, the order parameter S of the Lipari-Szabo model is linked to the proton fraction involved in the entanglements. We propose a new predictive model for viscosity combining T1 and T2 (Fig. 1).

y = 36.7x-1.05

R2 = 0.995

1.E-02

1.E-01

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03

ηηηη/T (mPa.s/K)

T 2.(T

2/T 1

)2 (ms)

Fig. 1: Viscosity prediction valid for entangled polymers (full circles); the prediction can also be used for non entangled polymers (triangles, T2 /T1≅1) but with a different pre-factor. This relation has been established on PEG samples of molecular mass Mw varying from 200 up to 35 000. For each molecular mass, the temperature was varied from 60 up to 80°C.

References: 1. R. Kimmich and N. Fatkullin, Polymer Chain Dynamics and NMR. Springer. 2004. 2. G. Lipari and A. Szabo, Journal of American Chemical Society, 104: 4546, 1982.

139

Diffusion tensor micro-imaging of articular cartilage as a probe of

tissue microstructure

S.K. de Vissera, J. C. Bowden

a, R.M. Wellard

a, K. Momot

a,b and J.M. Pope

a,b

aSchool of Physical & Chemical Sciences and

bInstitute of Health and Biomedical Innovation

Queensland University of Technology, Brisbane, Australia

Osteoarthritis is a disease of articular cartilage that affects an estimated 20% of the

population, but the early stages of osteoarthritis are difficult to diagnose and the progression of

the disease is poorly understood. Articular cartilage covers the ends of the bones in joints,

providing a smooth low friction surface that absorbs shocks and distributes load to the

subchondral bone. Healthy adult human cartilage is 2-4 mm thick and has a complex

composition and structure. It contains collagen (15−20%), proteoglycans (PG, 3−10%), lipids

(1−5%), chondrocyte cells, and water (65-80%). The collagen forms a cross-linked network, in

which three zones of alignment are usually distinguished: the fibres at the articular surface are

aligned parallel to the surface and become increasingly aligned normal to the surface at greater

depth. Collagen is primarily responsible for the compressive stiffness of cartilage, while the PG's

are responsible for maintaining osmotic equilibrium.

Cartilage is invisible on plane x-rays and its thickness can only be inferred from the

separation of the subchondral bone surfaces. The surface of the cartilage can be viewed using

arthroscopy, but this is invasive and provides no information on changes in structure and

morphology of the cartilage beneath the articular surface. In conventional clinical MRI the

appearance of cartilage is complicated by the presence of bands of high and low signal intensity

that depend on the orientation of the normal to the articular surface with respect to the applied

static magnetic field. These bright and dark bands reflect the orientation dependence of the spin-

spin relaxation time T2, which is determined by constraints on the mobility of water protons

imposed by the underlying structural anisotropy of the collagen fiber matrix. Thus, while the

complex nature of contrast observed in clinical MRI of cartilage has to date been seen as a

complicating factor for diagnosis, MRI has the potential to provide information on changes in

molecular structure of cartilage that could assist in the early diagnosis of disease and in

monitoring response to therapy.

In studies of bovine cartilage in vitro1,2

we have shown that diffusion tensor imaging

(DTI) can be used to monitor the orientation of the collagen fibers as a function of depth from

the articular surface and to measure changes in fiber orientation produced by both enzymatic

degradation and mechanical compression. Results were correlated with data from polarized light

microscopy (PLM)3 and polarized Fourier transform infra-red spectroscopy (FTIR). We are

currently modeling the orientation dependent relaxation of protons on water molecules bound to

collagen fibers with a view to correlating T2 anisotropy with molecular sub-structure in cartilage.

References:

1. R. Meder, S.K. de Visser and J. M. Pope, Osteoarthritis and Cartilage 14, 875-881 (2006).

2. S. K. de Visser, R. W. Crawford and J. M. Pope, Osteoarthritis and Cartilage 16, 83-89 (2008).

3. S. K. de Visser, J. C. Bowden, L. Rintoul, E. Wentrup-Byrne, T. Bostrom, J. M. Pope and K. I.

Momot, Osteoarthritis and Cartilage (in press).

140

Role inversion in magnetization exchange between water and macromolecular protons in articular cartilage components

at different hydration levels

P. Fantazzinia, M. Gombiaa, R.J.S. Brownb, F. Vitturc

aDipartimento di Fisica, University of Bologna, Viale Berti Pichat 6/2, 40127, Bologna, Italy

b953 W Bonita Ave, Claremont CA 91711-4193, USA cDipartimento di Scienze della Vita, University of Trieste, Via Giorgieri 1, 34127 Trieste, Italy

NMR techniques are acquiring an ever-increasing importance both in basic science and

clinical studies of articular cartilage, a connective tissue that can be considered, for many aspects, as a porous medium. NMR is unique as it furnishes non-destructive tools to evaluate the changes occurring in the cartilage due to pathologies and ageing, both in vitro and in vivo. In particular, MRI can be repeatedly used to monitor changes of articular cartilage of the human joints, without radiation damages. One of the most relevant MRI parameters is the magnetization transfer rate, that has been related to the concentration of the collagen fibres.

We studied the exchange of magnetization between water and macromolecular protons both ex vivo in native cartilage, and in its purified components, collagen and proteoglycans, at different water content. All the free induction decays could be fit by the sum of an exponential (liquid-like) and a Gaussian or modified Gaussian (solid-like) signal. The “solid-like” component is due to macromolecular 1H of restricted mobility, the “liquid-like” to water. When quasi-continuous distributions of longitudinal relaxation time were computed without imposing the usual non negativity constraint, the effect of the magnetization exchange between macromolecular and water 1H was evident. Data could be formally fit by a simple two-site exchange model, already used by us for a study of hydration in seasoned woods1, where the Solid-to-Liquid exchange times were found to be of the order of few tens of ms. An inversion of the Solid-Liquid roles in the magnetization exchange was observed with changing the hydration level of the macromolecules. Such an inversion has been seen to occur at the hydration level 20-26% (w/w) in a model system (Sephadex G10)2.

References:

1. P. Fantazzini, A. Maccotta, M. Gombia, C. Garavaglia, R.J.S. Brown, M. Brai, Solid-Liquid NMR relaxation and signal amplitude relationships with ranking of seasoned softwoods and hardwoods, J Appl Phys 100: 0749071-7 (2006). 2. P. Fantazzini, L. Lendinara, 1H-NMR relaxation in hydrated Sephadex system, in Physics in environmental and Biomedical Research, S.Onori, E. Tabet Editors, pp. 263-266, World Scientific Publishing Co., 1986.

141

GARField NMR profiling studies of

human stratum corneum and viable epidermis in vivo

P. J. McDonalda, S. Pitts

a, E. Ciampi

b, M. van Ginkel

b and S. Singleton

b

aDepartment of Physics, University of Surrey, Guildford, Surrey, GU2 7XH, UK

bMeasurement Science, Unilever, Colworth Science Park, Sharnbrook, Bedford, MK44 1LQ, UK

In a series of recent papers1 describing a limited number of trials, we have shown how

GARField magnetic resonance profiling is an ideal tool to study small molecule transport in

human skin stratum corneum and viable epidermis both in-vitro and in-vivo. This is possible

because, with GARField, the resolution (pixel size) is better than 10 μm with a first echo time as

short as 150 μs. In this previous work, there was indication that profiling of magnetic relaxation

parameters can access the qualitative parameter that is often dubbed “skin-moisturisation”: this

being a reflection of skin suppleness rather than true hydration. The current work described here

goes forward with two objectives.

In one set of experiments, a 5 pulse sequence (P90 - (τ-δ)/2 - P180 - (τ+δ)/2 – P90 - Δ -

(τ+δ)/2 - P90 - (τ-δ)/2 – P180 - echo) has been used to measure self-diffusivities in stratum

corneum and viable epidermis in vivo under the constant gradient conditions of GARField2. This

sequence allows simple diffusion of a single species to be measured independent of relaxation

weighting using constant τ and Δ intervals. By sampling the full 3 dimensional parameter space,

δ, τ, and Δ measurements of more complex diffusion scenarios with different characteristic

length scales as a function of time are possible. On human stratum corneum and viable epidermis

these experiments reveal (for the first time and in vivo) diffusion of water in a spatially and

temporarily defined manner showing different characteristic behaviours.

In other experiments, we quantitatively compare changes in stratum corneum and viable

epidermis relaxation rates occurring as a result of the application to the skin of materials typical

of model skin care formulation systems, with changes in the more standard measurement of

trans-epidermal water loss and with corneometer testing. Excellent correlation between the

results of the different methods is found. In one particular study, the volunteer ate a light meal

and took moderate exercise during the trial. These activities are known to affect core body

temperature and skin hydration and are faithfully echoed in the data sets. Single-point signal-to-

noise ratio and bio-reproducibility are comparable between the different methods. However, the

GARField MR alone provides depth resolution within the skin layers and has allowed a simple

qualitative model of skin swelling and moisturisation to be inferred.

References: 1 P. J. McDonald, A. Akhmerov, L. J. Backhouse, S. Pitts, J. Pharm Sci. 94 (2005)1850.

2 R. Kimmich, NMR - Tomography, Diffusometry, Relaxometry, Springer-Verlag, Berlin, 1997.

142

MRI VISUALISATION OF THE MOISTURE INGRESS INTO

POROUS TISSUE OF THE DECAYED TEETH

Władysław P. WĘGLARZ*, Marta M. TANASIEWICZ**, Tomasz W. KUPKA**,

Marco L.H. GRUWEL***, Boguslaw TOMANEK***

*Department of Magnetic Resonance Imaging, H. Niewodniczański Institute of Nuclear Physics Polish

Academy of Sciences, Kraków, Poland

**Department of Propaedeutics in Dentistry, Medical University of Silesia, Bytom, Poland

***NRC Institute for Biodiagnostics, Winnipeg, Canada

Introduction: During last decade MRI has been used in the research of the healthy and decayed teeth for

visualization of the dental surface geometry as well as for distinction between soft tissue (pulp) and

mineralized tissue (enamel, dentine and root cement) in the extracted teeth [1-3].

The aim of this work was to asses applicability of different MRI techniques for detection and

visualisation of the water ingress into porous structure of the decayed teeth in vitro.

Materials and Methods: Five extracted human teeth with different level of decay were used for 3D

Spin Echo (SE) MR Imaging done in MR Tomography Lab INP, Krakow, Poland, using 4.7 T research

MRI system. 3D (512x128x128) images of the teeth with resolution of 30x120x180 µm3 were obtained.

Single Point Imaging (SPI) [5] were performed using 11.7 T, vertical bore Bruker MRI system,

in IBD, Winnipeg, Canada. In this case resolution of 315x315x940 µm3 was achieved.

MR data were visualized and analyzed using home developed IDL based software and ImageJ

software.

Results

a b c d

Fig.1 The photo image (a), corresponding cross-sections through 3D SE MR image (b) and the pseudo

3D reconstruction (c) of the water in porous tissue of decayed tooth, obtained from 3D SE MR data. In

fig (d) 2D cross-section from 3D SPI data of tooth with decay comparable to that for SE images is

shown.

Discussion & Conclusions: Spin-echo based MR imaging methods allow for high resolution

visualization of the moisture entering into heavily decayed tooth, which are paths for microorganisms

causing further damage of the tissue. With this methods it is not possible to see any details in mineralized

tooth’s tissue, especially gradual decrease in tissue density in early stages of the caries development.

With this respect, the SPI method which allow for visualization of the mineralized tissue is more

promising.

References: 1. Lloyd CH, Scrimgeour SN, Chudek JA, HunterG, MacKay RL, Quintessence Int. 1997, 28(5), 349.

2. Lloyd CH. et.al. Caries Res 2000, 34, 53.

3. Węglarz WP, Tanasiewicz M, Kupka T, Skórka T, Sułek Z, Jasiński A, Solid State NMR, 2004,25, 84.

4. Latta P, Gruwel MLH, Edie E, Sramek M, Tomanek B, J. Magn. Reson. 2004, 170, 177.

143

Correlated analysis of the diffusion and T2 relaxation of water in multi-compartmental tissue of the spinal cord of a

rat

W.P. Węglarz, P. Rosicka, T. Banasik, U. Tyrankiewicz,

Dept. of Magnetic Resonance Imaging, Institute of Nuclear Physics, Polish Academy of Sciences, Kraków, Poland, Wladyslaw.Weglarz@ifj.edu.pl

Introduction: Present understanding of water diffusion in nervous tissue is based on assumption of three water compartment: intra-cellular, myelin and extra-cellular, with different diffusion and relaxation characteristics. There is still lot of debating concerning details of the diffusion model which includes all three compartments. To clarify this, proper characterization of all compartments is necessary. Especially myelin water is difficult to detect due to short T2. The aim of this work was to set up protocol for combined diffusion/relaxation characterization of all three compartments. Subject and Method: The male Wister rats used for measurements were anaesthetised using halothane in oxygene. MR measurements were performed on the 4.7 T research tomograph equipped with Maran DRX (Resonance Instruments Ltd.) console and the home built gradient coils and the surface coils. ECG and breath triggered 3D spin echo (SE) sequence without slice selection was used to measure diffusion for b-values up to 30000 s/mm2 and for different echo times (TE) in range from 10 ms to 400 ms. Results: The diffusion weighted MR images (128x128) of the spinal cord as a function of b-values were obtained for range of TE values. Dependence of signal amplitudes on TE and b-values for directions perpendicular and parallel to the spinal cord axis for white matter and gray matter were assessed and described in terms of multi-exponential as well as continuous 2D distribution of the T2/diffusion decays.

Fig.1. (a) MR images of the spinal cord of a rat in vivo with diffusion gradients oriented in direction perpendicular (top) and parallel (bottom) to spinal cord axis, for b equal to: 650, 1300, 1800, 2300, 2800, 3300 s/mm2; (b) T2 decays in white matter for the diffusion gradients perpendicular to the axis of the spinal cord. Diffusion decays in direction along spinal cord are close to mono-exponential, especially in WM. Diffusion in direction perpendicular to spinal cord axis deviates substantially from mono-exponential (Fig 1b). Discussion/Conclusions: 3D Spin Echo based sequence was found useful in characterisation of diffusion in different water compartments in nervous tissue, due to possibility of measurements with short TE. Observed dependence of diffusion/relaxation decays on TE was attributed to compartmentalisation of the water within nervous tissue. Acknowledgments: This work was partially supported by research grant no. 1 P03B 009 28 from Ministry of Science and Informatisation of Poland and European Reintegration Grant no. 46386. 144

In vivo imaging of DDIF contrast in the Human Knee

E. E Sigmunda, R. Regatte

a,M. Schweitzer

a, H. Cho

b, Y.-Q. Song

c

aNew York University Radiology Department, New York, NY

bMemorial-Sloan Kettering

Cancer Center, New York, NY cSchlumberger-Doll Research, Cambridge, MA

Trabecular bone (TB)

is a porous structure found in

load-bearing skeletal regions

and joints (vertebrae, radius,

femur, calcaneus), and is

monitored to determine risk

of fracture or osteoporosis.

However, the clinical

densitometry standard fails to

account for half of fragility

variations, and diagnostics

that incorporate architectural

“bone quality” information

are in high demand(1).

Recent in vitro work

has applied an MR structural contrast—decay from diffusion in the internal field (DDIF)—to

trabecular bone, including both bulk measurements(2) and microimaging(3) over a wide

range of structure and strength. This work showed that DDIF probes the projected surface-

to-volume ratio along the applied field. The present study extended this technique to the first

in vivo DDIF contrast results in healthy human volunteer knee joints.

Imaging was performed in a full body 3 T Siemens Trio scanner with a single channel

CP knee coil. Four healthy volunteers were scanned with an imaging protocol including a

FLASH-3D scan for high resolution anatomical reference (0.3 x 0.3 x 1.5 mm), gradient echo

imaging at variable echo time for T2* mapping, inversion recovery prepared turbo spin echo

imaging at variable inversion time TI for T1 mapping, and a stimulated echo prepared

BURST imaging at variable mixing time tD for DDIF mapping. The parametric maps were

acquired at 1.2 x 1.2 x 3 mm resolution. T2* and TDDIF maps were generated by single

exponential fitting, and T1 maps by single exponential recovery fitting.

Figure 1 shows a comparison of the short decay time limits of T2* and TDDIF maps in

an axial slice through the condyles of the distal femur. Both show high contrast in the

intercondylar notch, where trabecular density and mechanical stresses are high. Beyond this

commonality the two spatial patterns are similar but not identical. Higher precision (e.g.

higher SNR sequence) will be required to explore the unique sensitivity of DDIF. However,

this pilot study confirms that DDIF probes trabecular structure in vivo.

References:

1. Wehrli FW, et. al. Quantitative MRI for the assessment of bone structure and function.

Nmr in Biomedicine 2006;19(7):731-764.

2. Sigmund EE,et.al.Diffusion-based MR methods for bone structure and evolution. Magnetic

Resonance in Medicine 2008;59(1):28-39.

3. Sigmund EE, et.al. High resolution MRI of internal field diffusion-weighting in trabecular

bone. Nmr in Biomedicine 2008;submitted.

Figure 1: T2* and TDDIF (te = 40 ms) contrast in an axial slice of a

healthy volunteer distal femur at 3 T. Windowed parameter maps

are overlaid on a high resolution FLASH image.

86420

T2* (ms)

340320300280260

T_DD

IF (ms)

T* DDIFT2

145

Helium Diffusion Kurtosis MRI of the Lungs

G. Johnson, R. Trampel, J. Jensen,

Department of Radiology, New York University School of Medicine

Diffusion MRI of hyperpolarized 3He is a promising method of characterizing the

small airspaces (1). Diffusion is restricted by lung structures and hence measured diffusion

coefficients reflect airspace size. However, with typical diffusion times of a few milliseconds

(limited by the short T2* of helium) diffusion measurements largely reflect alveolar size and

are insensitive to changes in the bronchioles. Diffusion kurtosis imaging (DKI) measures the

degree to which diffusion is non-Gaussian and is particularly sensitive to diffusion over

larger distance scales (2). Including the kurtosis term MR signal intensity is

⎟⎠⎞

⎜⎝⎛ +−= KDbDbSS 22

0

6

1exp

where S0 is the signal intensity without diffusion weighting, b is the diffusion weighting

constant, D is the diffusion coefficient and K is kurtosis (2). Kurtosis can thus be measured

with conventional diffusion imaging sequences with at least three b-values.

DKI was performed on five normal controls and three firefighters involved in the

rescue attempts at the World Trade Center. The firefighters suffered from symptoms

including cough, wheezing, shortness of breath, chest pain, and dyspnea. Gradient-echo

diffusion images were acquired on a 1.5T scanner (flip-angle 4°; matrix 80×128; slice

thickness 15mm; FOV 330×380mm; TE 11ms; TR 14ms; b= 0,3,6,9,12 and 15s/cm2).

Fig. 1. Coronal images of a normal control (top) and

firefighter (bottom) with evidence of small airway

disease. Left: ventilation images. Middle: Diffusion

coefficient maps (range 0.0 – 0.6). Right: Diffusion

kurtosis maps (range 0.0 – 1.0). Diffusion is similar in

both subjects but kurtosis is reduced and less

heterogeneous in the firefighter.

Measured parameters were: firefighters D=0.263±0.049, K= 0.213±0.022; controls

D=0.337±0.038, K=0.212±0.024. Although there was no statistically significant difference in

the D (p=0.95), the firefighters had a substantially reduced K with a p value approaching

significance (p=0.098).

Decreased kurtosis in combination with normal diffusion coefficients suggests

bronchiolar constriction in the firefighters, an interpretation supported by the finding of air

trapping demonstrated by CT.

DKI may provide information that is complementary to that provided by conventional

diffusion imaging.

References:

1. X.J. Chen et al. Magn. Reson. Med. 42 (1999) 721; B. Saam et al. Magn. Reson. Med. 44

(2000) 174; D.A. Yablonskiy et al. Proc. Natl. Acad. Sci. USA 99 (2002) 3111.

2. J.H. Jensen et al. Magn. Reson. Med. 53 (2005) 1432; R. Trampel et al. Magn. Reson.

Med. 56 (2006) 733.

146

Toward accurate and precise S/V measurements of lung airspaces using noble-gas diffusion NMR

G. Wilson Millera, Michael Carlb, Gordon D. Cates Jr.ba, and John P. Mugler IIIa,

aDepartment of Radiology, bDepartment of Physics, University of Virginia, Charlottesville, Virginia USA

In-vivo diffusion measurements can be made in lung airspaces using gaseous NMR

signal sources such as hyperpolarized 3He or 129Xe. Whereas the relatively large self-diffusivity of these noble gases is an advantage for accessing long length scales corresponding to the tortuosity limit in porous solids, it a disadvantage for accessing short length scales corresponding to the surface-to-volume ratio (S/V) of lung airspaces [1]. For noble-gas diffusion in normal human lung, extracting S/V values from the short-time-scale limit of the time-dependent diffusivity curve requires that the curve be sampled at diffusion times much less than a millisecond [2]. This requirement generally precludes the use of refocusing RF pulses between individual lobes of the diffusion-sensitizing gradient, and also makes it impractical to generate measurable NMR signal attenuation using narrow gradient pulses. Although quantitative corrections can be applied that compensate for systematic errors resulting from the use of wide gradient pulses [3,2], it is nonetheless difficult to make clean diffusivity measurements subject to these constraints.

We recently developed a new gradient-echo-based pulse sequence for measuring noble-gas diffusion, which provides improved measurement accuracy and precision at sub-millisecond diffusion times [4]. These advantages were demonstrated using both Monte-Carlo simulations and phantom measurements in simple pore geometries. We are currently exploring the ability of our technique to quantify S/V in the pore space between loosely packed microspheres, using 129Xe gas at atmospheric pressure. A significant component of this effort is to characterize the effects of susceptibility-induced background gradients on the measured diffusivities. Taken as a whole, these developments represent a significant step toward establishing the feasibility of making reliable in-vivo lung S/V measurements using NMR of noble-gas diffusion.

This research was supported by NIH grants R01-HL079077 and R21-HL089525, and by Siemens Medical Solutions.

References: 1. R. W. Mair, G. P. Wong, D. Hoffmann, M. D. Hürlimann, S. Patz, L. M. Schwartz, and R.

L. Walsworth, Probing Porous Media with Gas Diffusion NMR, Phys. Rev. Lett. 83 (1999) 3324.

2. G. W. Miller, M. Carl, J. F. Mata, G. D. Cates Jr., and J. P. Mugler III, Simulations of short-time diffusivity in lung airspaces and implications for S/V measurements using hyperpolarized-gas MRI, IEEE Trans. Med. Imaging 26 (2007) 1456.

3. L. J. Zielinski and P. N. Sen, Effects of finite-width pulses in the pulsed-field gradient measurement of the diffusion coefficient in connected porous media, J. Magn. Reson. 165 (2003) 153.

4. M. Carl, G. W. Miller, J. P. Mugler III, S. Rohrbaugh, W. A. Tobias, and G. D. Cates Jr., Measurement of hyperpolarized gas diffusion at very short time scales, J. Magn. Reson. 189 (2007) 228.

147

Human Pulmonary Function with Hyperpolarized 129

Xe

S. Patza, I. Muradyan

a, M.I. Hrovat

b, F.W. Hersman

c, J.P. Butler

d

aBrigham and Women’s Hospital & Harvard Medical School, Boston, MA, bMirtech Inc, Brockton, MA, cUniversity of New Hampshire and Xemed, LLC, Durham, NH, dHarvard

School of Public Health, Boston, MA The function of the lung is gas exchange; avelolar oxygen diffuses into blood and CO2 diffuses in the opposite direction. This is accomplished by the lung’s large surface area (~100m2) due to its fine porous structure together with a thin septal barrier separating alveolar gas and capillary blood. Here, we describe a noninvasive method to determine three functional components of the lung: alveolar surface area per unit volume of gas (S/V), which decreases in emphysematous disease; septal thickness h, which increases in interstitial disease; and capillary transit time τ, important for vascular diseases.

S/V regime

Saturation

Figure 1. Example of human CSSR data.

√t (s0.5)

Blood Flow

The spectrum of 129Xe in the lung exhibits a gas peak at 0ppm and dissolved state peaks at ~+200ppm. A chemical shift saturation recovery (CSSR) pulse technique is used to first saturate the dissolved phase 129Xe magnetization and then observe its recovery1 due to diffusion of magnetized 129Xe from the gas spaces to the septa. The fraction F of the dissolved state signal at time t after the saturation pulse relative to the initial gas phase signal is measured. The data is fit to an analytical form obtained from the solution of a 1D diffusion geometry and includes a constant velocity blood flow term. Input parameters to the model are the diffusivity of 129Xe in tissue and the Ostwald solubility. Fitted parameters are pulmonary S/V, h, and τ. Figure 1 shows sample data from a healthy human subject after inhalation of 1L of 86% enriched 129Xe polarized to ~50%. The field strength is 0.2T. Data (green points) are acquired in a 10s breath-hold. The fitted curve is shown in blue. A red curve, calculated by setting τ→∞, shows septal saturation in the absence of blood flow, where the saturation value depends on h. The early time behavior is ~√t and allows determination of S/V

2.

(a) (b) (c) (d)

Figure 2. Example of single breath XTC for t=62ms. (a-c) serially acquired gradient echo images (d) calculated F(t) map.

To obtain this type of information regionally, we modified the XTC method3 to a single breath-method2; each experiment produces a map of F at a single t. Figure 2 shows 3 gradient echo images serially acquired (a-c). Signal attenuation between images is due to (r1) RF depletion of the nonrenewable hyperpolarization, (r2) T1 decay due to oxygen, and (r3) interphase diffusion. Attenuation between (a) and (b) allows determination of r1r2. Between (b) and (c), attenuation = r1r2r3 where r3 is enhanced due to multiple applications of [180o

dissolved – t] between the two images. Determination of r3 allows calculation of F (Fig 2d). A separate breath of ~1L of hyperpolarized 129Xe is required for each time point to acquire data similar to Figure 1 for each voxel. This work was supported by NIH RO1 HL073632.

References:

1. Butler et al, J Phys: Cond

Matter, 14, L297-L304 (2002). 2. Patz et al, Eur J Rad, 2007. 3. Ruppert et al., Magn Reson Med. 44, 349–357 (2000).

148

Magnetic Resonance Microscopy application to Biofouling in

Porous Media

Jennifer A. Hornemanna,b

, Sarah L. Coddb,c

, Joseph D. Seymoura,b

, and Konstantin

Romanenkoc

aDepartment of Chemical and Biological Engineering, Montana State University

bCenter for Biofilm Engineering, Montana State University

cDepartment of Mechanical and Industrial Engineering, Montana State University

From water service utilities to pharmaceutical processing, biofilms either create

havoc or are crucial to the system function in almost every water-based industrial

process. In subsurface bioremediation, biofilm establishment and maintenance is crucial

to many processes intended to degrade or contain contaminants. Many of these situations

in which biofilm growth needs to be understood involve water transport in a growth

medium that can be interpreted as a porous media. Thus, there is an urgent need to better

understand the structure and transport changes that occur in biofouled porous media.

Magnetic Resonance Microscopy (MRM) techniques have been used to study the

effective change in pore structure and pore connectivity due to biofilm growth and the

impact this growth has on transport phenomena. The porous media used in this study are

model beadpacks constructed from various glass and plastic spheres with diameters

between 100-300 m.

149

MR spectral characterization of diffusion with chemical shift resolution: study of highly concentrated emulsions

S. Lasica,b, I. Åslunda, N. Arteagaa, D. Topgaarda

aLund University, Physical Chemistry 1, Sweden, bUniversity of Ljubljana, Faculty for Mathematics and Physics, Slovenia

Emulsions are important for many technological applications for example in

pharmacy as drug delivery systems and in food or cosmetic products. Diffusiometry is a well-established approach to study morphology of porous materials and emulsions. Self-diffusion of different substances in highly concentrated water-in-oil emulsions have been studied by the pulsed gradient technique [1]. In the case of fairly mono-dispersed emulsions the diffraction pattern of signal intensity vs. phase factor q can reveal the information about droplet size. The effects of finite length gradient pulses in characterising emulsions have been thoroughly examined [2].

The finite gradient pulse effects are not relevant to the modulated gradient spin-echo method (MGSE) [3,4]. It provides information about molecular diffusion in frequency domain. Instead of using gradient pulses it employs harmonically oscillating gradients. This results in spin-echo attenuation, which is proportional to the spectrum of velocity autocorrelation function of the spin-bearing particles. The effect of spin de-phasing, due to diffusion, is accumulated over many gradient oscillation cycles, giving rise to adequate diffusive attenuation on a shorter time scale compared to the pulsed field gradient method.

We present an application of the MGSE method, which uses sinusoidaly shaped gradient pulses separated by 180-degree RF pulses. This produces an apodized cosine gradient modulation waveform, similar to the one employed by Parsons et al. in combination with imaging [5]. The RF pulses used in the presented technique efficiently refocus chemical shifts and de-phasing due to susceptibility differences, resulting in undistorted, high-resolution diffusion weighed spectra. This allows for simultaneous study of several molecular components with different chemical shifts.

The method was used to study highly concentrated water-in-oil emulsion, which consists of 95 wt% 0.2 M tetramethyl ammonium chloride (TMACl) in Millipore water as aqueous phase, 3.5 wt% heptane as oil phase and 1.5 wt% Brij 92 as surfactant. The droplet size is fairly mono-dispersed around 3.3 µm. In the range between 1-20 ms we observe the transition from restricted to free diffusive behavior of water and TMACl, while oil diffusion is occluded but Gaussian.

References: 1. C. Malmborg, D. Topgaard, O. Söderman, J. Colloid Interface Sci. 263 (2003) 270. 2. C. Malmborg, D. Topgaard, O. Söderman, J. Magn. Reson. 169 (2004) 85. 3. J. Stepisnik, Prog. Nucl. Magn. Reson. Spectrosc. 17 (1985) 187. 4. P.T. Callaghan, J. Stepisnik, J. Magn. Reson. A 117 (1995) 118. 5. E. C. Parsons, M. D. Does, J. C. Gore, Magn. Reson. Imaging 21 (2003) 279.

150

Pore surface of cellulose as studied by Low-Resolution NMR

E. Nikolskayaa, Y. Grunin

a, D. Karasev

a, John C. Edwards

b, L. Grunin

a

aMary State Technical University, Lenin sq. 3, Yoshkar-Ola, Mari El, 424000, Russia,

bProcess NMR Associates LLC, 87A Sand Pit Rd, Danbury, CT 06810 USA

Polysaccharides are frequently considered as sorbents since they possess good absorbent ability due to branchy porous structure. The question of saccharides supermolecular organization is controversial so far, so the conception of its porous structure is not absolutely clear. Supramolecular organizations of polysaccharides is normally considered as a combination of crystalline part with well-ordered, highly oriented and densely packed macromolecules and amorphous part which is presented as a less ordered and the most reactive region. Further on we suggest that so-called “surface” region of the sample consists of atoms of glucopyranose ring which are situated on the surface of crystallites and they form nearly all the amorphous part. “Internal” regions of polysaccharides are densely packed areas with crystal-like behaviour. Surface between crystallites constitutes pores which have mainly fissured form. So it could be supposed that amorphous part organizes porous structure of polysaccharide.

Due to last tendencies NMR relaxation is very convenient method for investigation of polysaccharides sorption ability and our study basically oriented to exploit this method.

Acquired for cellulose decay contains quickly falling component, which corresponds to crystallites and slowly falling component which is referred to amorphous phase of polysaccharide. FID data potentially could be used for determination of cellulose crystallinity extent and for its porosity evaluation. But the difficult task of the FID data interpretation is caused by impossibility of satisfactory fitting of the decay.

The Fourier transformation of FID signal measured on the shift from resonance frequency offers more informative data. Doublet line width ∆ωdoubl is referred to rigid lattice protons of ordered regions with low mobility. Central peak, which is narrower (∆ωCP) comprises contributions of reorienting parts of macromolecules in amorphous phase and adsorped water signal.

The central peak of cellulose spectrum has a complicated form. There are two wide lines, which are observed in condition of high magnetic field stability and large number of scans. The explanation of this fact requires additional scientific investigations. In our work different types of cellulose were treated with heating processing at the range of temperatures from 106ºC to 136ºC. Crystallinity extent, which is the most prevalent and the most studied characteristic of polysaccharides structure, finds good correlation with the height of central peak. While the temperature of treatment increasing the amplitude of central peak is decreasing.

Time of spin-lattice relaxation is the characteristic of structural order. It flows from the crystallites, where it goes slowly, to flow-out centers of magnetization located on the surface. So the longitudinal relaxation T1 is bound up with crystallinity extent. Our investigations show obvious dependence the relaxation time T1 on crystallinity extent. The higher crystallinity extent the greater T1 value.

Crystallinity extent and the dimensions of crystallites determine the porosity of cellulose. The overall amount of pores is reduced with crystallites dimensions rise. There is a possibility to predict both the sorption ability and pore volume of cellulose and its crystallinity extent.

Thus proton magnetic longitudinal and transverse relaxation measurements turn to be a promising tool to determination of porous structure of natural sorbents like polysaccharides.

151

Spatial Microheterogeneity of Polyelectrolyte-Protein Coacervates

revealed by 1H Pulsed Field Gradient NMR Diffusion Study

Amrish R. Menjoge1, A. Basak Kayitmazer

2#, Paul L. Dubin

2 and Sergey Vasenkov

1

1 Department of Chemical Engineering, University of Florida, Gainesville, FL 32611

2 Department of Chemistry, University of Massachusetts Amherst, MA 01003

# Present address: Biomedical Engineering, Northwestern University, Evanston, IL 60208

Polyelectrolyte-protein coacervates are dense macromolecular solutions composed of a protein

and an oppositely charged polyelectrolyte. Coacervates can be used in such applications as

micro-encapsulation of active food components, active drug ingredients or enzymes, since all

essential properties of proteins remain preserved in the coacervate environment. Understanding

the structure of coacervates on the submicrometer length scale is of importance for using

coacervation to its fullest potential. Here, we report the results of pulsed field gradient (PFG)

NMR studies of diffusion in coacervates formed from a globular protein (Bovine Serum

Albumin) and poly(diallyldimethlyammoniumchloride), a narrow weight distribution cationic

polymer. High magnetic field gradients (up to 30 T/m) were used for the observation of

polyelectrolyte diffusion in coacervates on the length scale as small as 100 nm. Diffusion

properties of polyelectrolyte in coacervates formed at different pH and ionic strengths have been

investigated. Our results indicate the presence of dense domains with sizes in the range of 100

nm and larger formed by the protein and the polyelectrolyte. These domains were observed to be

in a process of constant formation and breakup on a time scale of milliseconds. The domains are

surrounded by a continuous dilute phase, which is characterized by a lower concentration of both

species. These results represent a significant step towards understanding the structural and

dynamic properties of coacervates on the nanoscale.

152

Pore Opening and Closing in Biodegradable Polymers and its

Effect on Mass Transport Studied by NMR Cryodiffusometry

E .L. Perkinsa, S. P. Rigby

a,*, K. J. Edler

b, J. P. Lowe

b

Departments of aChemical Engineering and

bChemistry, University of Bath, Claverton

Down, Bath, BA2 7AY, UK.

NMR cryodiffusometry has been used to study the structural evolution of poly(lactic-

co-glycolic acid) (PLGA) microspheres prepared by a double emulsion (w/o/w) technique.

The mass transport of water within the polymer matrix was studied as a proxy for the

mobility of small drug molecules entrapped for controlled release. Over time, it was found

that the matrix ‘pores’ open and close as a result of water ingress between the polymer chains

constricting pores and forcing water into new spaces. In cryoporometry, the freezing curve is

often neglected due to a lack of clear understanding of hysteresis mechanisms. However, in

our previous work [1], we have shown that this curve can provide information on pore neck

sizes. Figure 1 shows the temporal evolution of the bimodal neck size distribution following

immersion of microspheres. The neck size for the major peak first shrinks and then grows,

while the minor peak grows in intensity but remains at a constant neck size. It is likely that

the evolution of the neck sizes for the matrix pores controls the mobility of drug molecules

within the matrix, and therefore influences the drug release profile [2]. The melting curve

pore size distribution remains unchanged over the same period (not shown).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1 1.5 2 2.5

Nominal pore neck diameter/nm

dI/dD /arbitrary units

Day 1

Day 10

Day 20

Day 27

Day 34

Day 42

Fig. 1: Temporal evolution of the nominal diameter (D) pore necks of the PLGA

matrix.

Results from cryodiffusometry experiments for molten aqueous phase solely within

the polymer matrix gave rise to curved log-attenuation plots, suggesting a heterogeneous

system, which fitted well to a two-component model. Hence, two distinct diffusive

environments exist for water within the polymer matrix pores. Currently the drug release

mechanism from PLGA microsphere systems is not fully understood, but with the additional

use of cryodiffusometry, a rich data set can be found to enhance the understanding of these

complex systems to improve the design of controlled drug delivery systems.

References:

1. E.L Perkins, J.P. Lowe, K. J. Edler, N. Tanko, S. P. Rigby, Chem. Eng. Sci. 2008,

doi:10.1016/j.ces.2007.12.022.

2. Kang, J.C. and S.P. Schwendeman, Molecular Pharmaceutics, 2007. 4(1) 104-118.

153

Temporal correlation function around spheres and cylinders

C. H. Ziener1, T. Kampf1, V. Herold1, P. M. Jakob1, W. R. Bauer2, W. Nadler3

1Experimentelle Physik 5, Universitat Wurzburg, Wurzburg, Germany2Medizinische Klinik and Poliklinik I, Universitat Wurzburg, Wurzburg, Germany3John von Neumann Institute for Computing, Forschungszentrum Julich, Germany

Introduction: Signal attenuation in NMR is partly due to incoherent dephasing of nuclearspins. This could be caused by independent stochastic motion of nuclear spins in inhomoge-neous local magnetic fields, e.g. created by magnetized objects or voids inside magnetizedbulk materials. If the magnetization time decay is not monoexponential the description withthe transverse relaxation time T ∗

2as single parameter will poorly describe the signal and the

underlying geometry of the magnetic field.Another way to describe the effects of diffusion is in terms of a correlation function [1].

Since it was demonstrated recently that the frequency correlation function of nuclear spinscan be measured directly in NMR experiments [2], a theoretical analysis of such functions isof paramount interest.Methods and Results: In this work we provide a numerically exact analysis of that cor-relation function K(t) for restricted diffusion in the inhomogeneous field around cylindersand spheres (cf e.g. Fig. a). With the local larmor frequency ω(r) and the probabilitydensity p(r, r0, t) to find a spin after time t at position r if it started at r0 K(t) is given as

K(t) =1

V

∫V

d3r

∫V

d3r0 ω(r) p(r, r0, t) ω(r0) .

V is the diffusion volume. K(t) is solved in terms of a spectral expansion of p(r, r0, t).The functional form may exhibit 3 regimes: after an initial transient an algebraic regime

with a t−d/2 time dependence - d being the space dimension - may form followed by an expo-nential cutoff due to finite system size effects. The main parameter controlling the formationand range of the regimes is the volume fraction of the magnetic inhomogeneities. Comparingthe algebraic regime with the case of long time behavior in unrestricted diffusion equal timebehavior is found and can be used at this time scale. The results obtained in this work mayhave implications on data analysis and interpreting results and could help choosing a properfit function for better parameter estimation.

RS

ϕ

θR

r

B0

a) b)

Fig. a) Scheme of a spherical inho-mogeneity inside a spherical voxel b)Scheme of a double logarithmic plot ofa correlation function for this geometry(K exact, Ka short time limit from un-restricted diffusion, Kb long time limitfrom unrestricted diffusion, Kc expo-nential cut off from restricted diffusion,τS correlation time, a, b and c denotesthe different regimes)

References[1] J. H. Jensen, R. Chandra, Magn. Reson. Med. 44, 144 (2000).

[2] J. H. Jensen et al., Magn. Reson. Med. 55, 1350 (2006).

154

ANAHESS, a new second order inverse

Laplace transform algorithm

Åsmund Ukkelberga, Geir Humborstad Sørland

b, Eddy Walter Hansen

a and Hege C.

Widerøec

aThe University of Oslo (UiO), Norway,

bAnvendt Teknologi AS, Harstad, Norway,

cStatoilHydro Research Centre, Trondheim, Norway

Keywords: inverse Laplace, Hessian, discrete solution, analytical expressions

The operation known as the Inverse Laplace Transform (ILT) is frequently applied in

the numerical analysis of nuclear magnetic resonance experiments. Estimates of properties

such as pore size distribution are frequently made on the basis of numerical ILT operations.

Good ILT algorithms are therefore essential when extracting physical information from NMR

experiments.

The ILT is the decomposition of a response curve in the first order case (1D-ILT) or a

response surface in the second order case (2D-ILT) into a sum of exponentials. It is well

known that this problem is ill conditioned, and that interpretation of the results should be

considered with caution.

The aim of our research was to develop an algorithm which performs 2D-ILT with

minimal assumptions, and produces models that are as simple as possible. We have

constructed an inverse Hessian (or Gauss-Newton) algorithm with analytical expressions for

the gradient and the Hessian matrix. Non-negative constraints are implemented by defining

ILT parameters as squares of the working variables in the minimization problem.

We propose a new algorithm called ANAHESS which computes 2D-ILT with a

specified number of discrete exponential components. Comparative studies show that the

sum of squared errors gets lower with a fewer number of components when using ANAHESS

than with a state of the art continuous distributions algorithm.

Given the premise that the ILT is ill conditioned, our conclusion is that numerical

solutions to ILT problems should be as simple as possible, with discrete components rather

than estimates of continuous distributions.

155

Uncertainties in the Assignments of Rock Porosities Obtained from

Laplace Inversion of NMR Relaxation Logs

Indrajit Saha1, 2, John Franck1, 3, Tim Hopper1, Boqin Sun4, Yi-Qiao Song 1

1Schlumberger-Doll Research, Cambridge, MA 02139

2Department of Chemistry and Biochemistry, Southern Illinois University Carbondale, IL 62901 3Department of Chemistry, University of California Berkeley, CA 94720 4Chevron Texaco Energy Technology Company, San Ramon, CA 94583

Abstract: Calculation of uncertainties from the Laplace inversion of NMR data used in the determination of oil field rock porosity is an important and under-valued area. An improved technique for calculating the uncertainty of NMR data has been developed. This method, denoted as A10 uses a constraint based on the means of the first 10 echoes in addition to the χ2 constraint used in the Parker et. al. method [4,5]. The Parker method is sensitive to the characterization of noise in the raw data and as a result the bounds can vary significantly. This new method produces much tighter, more realistic values for the porosity bounds. Using advanced processing algorithms and multi level processing platforms, we were able to reduce the calculation time down to minutes.

Figure-1: Left panel: The upper and lower bounds of rock porosities using A10 (green) and original Parker (blue) method. Right Panel: The relative widths of the ranges as determined by both methods. The inset highlights the differences between the methods. [1] Kleinberg et al.: The Log Analyst: 37(6), 20 (1996) [2] Song et al.: J. Chem. Phys.: 122, 104104 (2005). [3] Skilling et al.: Kluwer Academic: 45–52 (1989) [4] Parker et al.: J. Magn. Reson.: 174(2), 314 (2005) [5] Song et al.: Magn. Reson. Imaging: 25; 445 (2007)

156

Numerical modeling of the carbonate and the sandstone formations

Seungoh Ryu

Schlumberger Doll Research, Cambridge, MA, USA

It is of critical interest in the well-logging to make a reliable estimation of permeability of the formation via more accessible properties such as the length scale derived from NMR response. When the pore geometry is simple (quasi-periodic glass bead packs or clean sandstones), there are working empirical models that correlate the porosity and the length scale to the permeability. For heterogeneous carbonate rocks, such laws, based essentially on the notion of tortuous pipes of an effective hydraulic diameter and simplified interpretation of the NMR data, fail to give a consistent prediction and require an ad-hoc modification specific to the formation. The presence of ramified channels of varying capacity in these systems brings in a transport behavior qualitatively distinct from that of a quasi-periodic system. Heterogeneity and diffusive coupling further undermine the validity of the MR interpretation in extracting the controlling length scale parameter. Using the parallel Lattice Boltzmann and random walk simulations, we study transport and diffusion properties in various types of pore geometry, from simple 2D micro-fluidic mazes, 3D glass-bead packs and sandstones to more complex carbonates. In this work, we focus on the distinction between the sandstones and the carbonates as manifested in the numerically observed flow profile and diffusive/relaxation properties and try to develop a sound interpretational framework for complex porous media beyond the glass bead packs and sandstones. The author would like to thank colleagues at SDR and SCR for providing me with tomograms of carbonate rocks.

157

Spin Dynamics Simulations of Multiple Echo Spacing Pulse

Sequences in Grossly Inhomogeneous Fields

R. Heidlera, H. N. Bachman

a, Y. Johansen

a

aSchlumberger Oilfield Services, Sugar Land, TX 77478

Pulse sequences with multiple lengths of echo spacings are used in oilfield NMR

logging for diffusion-based NMR applications such as rock and fluid characterization. One

specific implementation is the so-called diffusion editing sequence comprising two long echo

spacings followed by a standard CPMG at a shorter echo spacing1. The echoes in the CPMG

portion contain signal from both the direct and stimulated echoes.

Modern oilfield NMR logging tools are designed for continuous depth logging of

earth formations by projecting both the static (B0) and dynamic (B1) fields into the formation.

Both B0 and B1 profiles are grossly inhomogeneous which results in non-steady-state

behavior in the early echoes. The spin dynamics effects present a challenge for processing the

echo amplitudes to measure porosity (amplitude extrapolated to zero time) and attenuations

for fluid or pore size characterization.

In this work we describe a calculation of the spin dynamics of the diffusion editing

sequence with two long echo spacings. Quantitative understanding of the early echoes allows

for improved characterization of heavy oil, gas, and other applications. The calculation takes

into account full B1 and B0 field maps, and comparisons will be made for sensors and

parameters typical of oilfield logging tools and environments.

References:

1. Hürlimann et al., Journal of Magnetic Resonance 157, 31–42 (2002).

158

A numerical analysis of NMR pore-pore exchange measurements using Xray-μCT techniques

C. H. Arnsa, Y. Meleana, K.E. Washburnb, and P.T. Callaghanb

aAustralian National University, bVictoria University of Wellington

Pore-pore exchange experiments are a recent development1 and hold great promise to spectrally derive length scales relevant for transport in porous media. In particular, NMR Laplace techniques like relaxation or exchange spectroscopy reach a spectral length scale resolution much higher than available from μCT images, but provide no spatial resolution. A combination of both techniques can lead to better models for regions of unresolved porosity in μCT images, implying increased accuracy for transport calculations based on such images.

In this study we carry out numerical NMR pore-pore exchange experiments on selected Xray-CT images of sandstones and carbonate rock, while at the same time tracking information about the geometry and topology of the pore space. In particular, we use random walk techniques and observe the radius of gyration for random walks in each encoding or diffusion interval as well as measuring local diffusion averaged relaxation rates in a forward modeling approach, including the effect of internal fields. Diffusion averaged internal field distributions and relaxation time distributions are in good agreement with experiment. We use pore partitioning techniques and geometric distance fields to relate the pore-pore exchange and relaxation spectra to underlying structural quantities, namely pore size, initial distance to a relaxing surface, and travelled distance. In addition to the concept of ‘pores’ and ‘throats’, e.g. convex/simple objects within the pore space connected by narrow constrictions, we use a diffusion averaged local surface to volume ratio to compare simulation data with experimental measurements.

References: 1. K. E. Washburn and P.T. Callaghan, Phys. Rev. Lett. 97 (2006) 175502.

159

Relaxation-relaxation exchange experiments with portable

Halbach-Magnets

A. Haber,a R. Fechete,

b J. Perlo,

a F. Casanova

,a E. Danieli,

a D. E. Demco,

a B. Blümich

a

aInstitute of Technical and Macromolecular Chemistry, RWTH Aachen University, Germany, bDepartment of Physics, Faculty of Material Science and Engineering, Technical University

Cluj-Napoca, Romania.

Relaxation-relaxation exchange experiments were performed on two-phase gas-liquid

systems. To conduct such experiments at low field with mobile NMR under the extraction

hood, new Halbach-Magnets were designed and built. One is characterized by maximum

simplicity, the other by an improved homogeneous magnetic field and a maximum accessible

sensitive volume. The construction of the magnets and their performance are described and

the results of first experiments are reported. We anticipate that this and related 2D NMR

experiments will be of great use to monitor multi-phase reactions on line.

References:

1. H. Raich, P. Blümler, Design and construction of a dipolar Halbach array with a

homogeneous field from identical bar magnets: NMR Mandhalas, Concepts in Magnetic

Resonance 23B (2004) 16-25.

2. S. Anferova, V. Anferova, J. Arnold, E. Talnishnikha, M. A. Voda, K. Kupferschläger, P.

Blümler, C. Clauser, B. Blümich, Improved Halbach sensor for NMR scanning of drill

cores, Magnetic Resonance Imaging 25 (2007) 474-480.

160

Non-invasive NMR profiling of painting layers

Federica Presciutti,1 Juan Perlo,2 Federico Casanova,2 Stefan Glöggler,2 Costanza Miliani,1 Marika

Spring,3 Bernhard Blümich,2 Brunetto G. Brunetti,1 Antonio Sgamellotti1

1 Centre of Excellence SMAArt and ISTM-CNR c/o Chemistry Department of University of Perugia,

Via Elce di Sotto 8, Perugia 06123 Italy.

2 Institute of Technical and Macromolecular Chemistry, RWTH Aachen University, D-52056 Aachen, Germany.3 Science Department of National Gallery of London, Trafalgar Square London, WC2, UK.

Restoration and conservation of paintings requires in-depth knowledge of the materials used, their properties as function of time, the environmental exposure and the distribution of materials in artwork. As moving paintings for analysis is connected with great effort or not possible at all, it is advantageous to move non-invasive diagnostic equipment to the paintings. Common non-invasive methods of analysis are restricted to surface layers, and lower layers cannot be accessed non-destructively. This is why we investigated the use of single-sided NMR sensors as an alternative method. With the Profile NMR-MOUSE, deeper layers can be accessed non-invasively to collect information about the stratigraphy of paintings and their support.This work reports results obtained by the Profile NMR-MOUSE. Binder aging and the thickness of proton-rich layers in paintings were characterized in situ without contact to the surface. The suitability and accuracy of this method were tested on two easel painting models consisting of copper acetate and cobalt blue in comparison to optical and electron microscopy. Depth profiles were measured that revealed amplitude variations consistent with the different layers of the paintings. The accuracy of thickness values determined non-destructively by NMR was conformed by destructive optical microscopy of painting models to be of the order of 20 µm. Different Italian Renaissance paintings were analyzed by the NMR-MOUSE in the National Gallery of Umbria in Perugia. My measuring NMR depth profiles at contingent spots of a painting, it is possible to build a 3D map of the statigraphy. Furthermore, the nature of the binder and its aging characteristics can be studied by their T1 and T2 relaxation signature. Artificially aged and naturally aged tempera binders can at least in the cases studied be differentiated. This makes it possible to identify recently restored sections in a painting and raises the hope to potentially identify forged tempera paintings.

161

Single-shot depth profiling with a single-sided NMR sensor:

application to skin measurements

Ernesto Danieli, Juan Perlo, Federico Casanova, and Bernhard Blümich

Institut für Makromolekulare Chemie, RWTH-Aachen University, Germany

Portable open NMR probes built from permanent magnets like the NMR-MOUSE offer several

advantages over conventional NMR systems. They are small in size, low in cost and robust.

However, associated with the open geometry of the magnets are a number of challenging

limitations. The strong stray field gradient of such sensors limits the thickness of the sensitive

slice to be much below one millimeter, imposing an important restriction to the experimental

time required to profile large depth ranges into the sample [1]. In this work we exploit the

concept of field shimming by means of a shim unit based on permanent magnet blocks [2] to

strongly reduce the magnetic field gradient along the depth direction while keeping the lateral

components under control. Under these conditions depth profiles over a range of some

millimeters can be measured with a resolution of 50 µm in a single-shot as the Fourier transform

of the echo signal acquired in the presence of a static gradient of about 1 T/m. The increment in

the thickness of the sensitive volume eliminates the need for repositioning the sensor with respect

to the object allowing us to cover the whole depth range in a single measurement. An important

advantage of this new magnet design is that by proper design of the shim unit, the reduction in

the static gradient is not achieved at expenses of a strong reduction in the magnitude of the

average field, a price usually paid in conventional designs. The field of the first prototype is

about 0.3 T, which is comparable to the field generated by a previous sensor of the same size [3].

References 1. J. Perlo, F. Casanova, and B. Blümich, Profiles with microscopic resolution by single-sided

NMR, J. Magn. Reson. 176, 64 (2005).

2. J. Perlo, F. Casanova, and B. Blümich, Ex Situ NMR in Highly Homogeneous Fields: 1H

Spectroscopy, Science 315, 1110 (2007).

3. J. Perlo, F. Casanova, and B. Blümich, 3D imaging with a single-sided sensor: an open

tomograph, J. Magn. Reson. 166, 228 (2004).

162

Magnetic Field Gradient Monitor (MFGM)

Hui Han, Bruce J. Balcom

MRI Centre, Department of Physics, University of New Brunswick, Fredericton, N.B, E3B

5A3 Canada

Porous media are of considerable interest in a very diverse range of activities. Pore

diameters range from subnanometer to macroscopic dimensions. Frequently their use or

importance is intimately connected with their transport properties. Diffusion and flow

measurements in porous systems are therefore of practical relevance.

Pulsed-field-gradient NMR (PFG NMR) has proven valuable for the study of both

diffusive and coherent molecular motions. Accurate PFG NMR studies of motions of large

molecules are, however, at present still difficult. For large, slowly moving species the

required gradient pulses must have large amplitudes, and the resulting eddy current fields can

cause errors in the phase encoding [1-2]. Restricted diffusion of smaller molecules in porous

media encounters similar difficulties.

Despite continued advances in NMR hardware, imperfections in the magnetic field

evolution during measurements still hamper numerous MR procedures. Large and fast

gradient switching provokes undesired dynamic field perturbations by various mechanisms

such as eddy currents, limited gradient bandwidth or heating effects. Reproducible field

perturbations can be determined approximately by existing measurements [3-5].

In this work we will present a novel method, related to but significantly different from

our previous methods [6]. We will compare our new method to other recent advances in

magnetic field gradient waveform/k-space trajectory measurement. Our new method

significantly improves applications, which either require high gradient waveform fidelity

and/or suffer from eddy current distortions (i.e. diffusion, flow, and phase contrast imaging).

References:

1. E.V. Meerwall, M. Kamat, J. Magn. Reson. 83 (1989) 309-323

2. S.J. Gibbs, C.S.Johnson, , J. Magn. Reson. 93 (1991) 395-402

3. G. Mason et al. Magn Reson Med. 38 (1997) 492-496

4. Duyn et al. J. Magn. Reson. 132 (1998) 150-153

5. Pruessmann et al., ISMRM 2005, p.681

6. D.J. Goodyear, M. Shea, B.J. Balcom, J. Magn. Reson. 163 (2003) 1-7

163

Compositional analysis of materials with unilateral NMR

Oleg V. Petrova , Igor V. Mastikhin

a, Bruce J. Balcom

a

aMRI Centre, Department of Physics, University of New Brunswick, Fredericton, Canada

We were interested in developing experimental protocols that enable the

discrimination between different sorts of liquid (e.g. oil and water) on the basis of their

molecular mobility, using unilateral (open-access) magnets. Because of the field

inhomogeneity inherent to unilateral magnets, our study was bounded by spin echo

experiments, particularly by those based on the Carr-Purcell-Meiboom-Gill (CPMG)

sequence. Here we present two ways to utilize CPMG for our purpose. First is to employ a

difference in relaxation times (T2) of the constituents, which is manifested in a multi-

exponential decay of a CPMG signal. To know how much of each T2 component is present in

the signal, we performed a multivariate regression of the decays, using a classical least-

square algorithm. A performance of the method is demonstrated on model food samples,

made of water and cod oil mixture stabilized by agar, as well as on a series of bitumen sands.

The prediction accuracy is strongly dependent of whether the calibration samples had

characteristic T2 similar to those in the test samples. Thereby, a proper choice of calibration

samples is the main factor of accuracy.

The second way is to perform a diffusion-weighted CPMG experiment, where

constituents are discriminated through a difference in their diffusivity rather than T2. The

diffusion delay that is needed for efficient suppression of a ‘faster’ constituent (water)

depends on the magnetic field gradient, hence the method includes a respective calibration

procedure. Once an appropriate diffusion delay is found, the amount of a ‘slow’ constituent is

proportional to the CPMG amplitude. This method is insensible to the T2 variation over the

samples, as we demonstrate here on example of commercial beef samples, but requires for

slower and faster diffusing components to have comparable T2.

164

Design methodology of a unilateral NMR apparatus

S. Utsuzawaa, E. Fukushimab, Y. Nakashimac

aNew Mexico Resonance, Albuquerque, NM, bABQMR, Albuquerque, NM, cAIST, JAPAN

Unilateral NMR apparatus using a barrel magnet, which was originally introduced by

Fukushima and Jackson [1], was found to be useful for investigating away from the apparatus

due to the large sample volume [2]. However, it was pointed out that there is a non-negligible

contribution from the regions with large magnetic field gradients. This problem is similar to

the one already known for the CMR (Combinable Magnetic Resonance) by Schlumberger [3].

In the presence of magnetic field gradients, diffusive motion leads to a signal decay that will

augment the normal T2 decay. Due to the complex pattern of the field gradients of the barrel

magnet, it is not straightforward to extract the intrinsic T2 value. Therefore, it is desired to

reduce the contribution from such regions.

Local sensitivity is defined by B0 and B1 distributions, both of which show complex

patterns on a unilateral NMR apparatus. Thus, it is not straightforward to obtain the optimum

setup just by improving the homogeneity of one of the magnetic fields. Therefore, we try to

accomplish our goal by using a sensitivity map for evaluation. The voltage induced in the coil

from a uniform sample in grossly inhomogeneous fields can be written as [3]

drtrmrFI

rrBtV yxyx ,

,0

12

0,. (1)

Here, F(Δω0) is the frequency response of the detection system, including the coil response

and any hardware or software filters. The quantity mx,y(r, t) is the transverse magnetization at

point r and time t, which depends on both hardware properties and pulse sequence. Figure 1

shows a sensitivity map calculated for the current apparatus with a barrel magnet and double-

D shaped rf coil. Radially extending lines correspond to the regions with large field gradients.

In order to obtain a coil design that only excites the region with a small field gradient,

Δω0 distribution weighted by sensitivity at each point was calculated from Eq. (1) with

changing the position and length of the straight wires composing a double-D coil. Figure 2

shows a sensitivity map with improved Δω0 distribution, which was obtained as a result of

brute-force procedure in a relatively small search space. By using an appropriate optimization

algorithm, it is expected to extend the search space that may lead to a better solution.

Reference 1. E. Fukushima and J. A. Jackson, U.S.Patent No. 6,489,872.

2. S. Utsuzawa and E. Fukushima, 9th

ICMRM, Aachen, Germany (2007).

3. M.D. Hülimann and D.D. Griffin, J. Magn. Reson. 143 (2000) 120-135.

Fig. 1. Sensitivity map for the current apparatus.

Surface of the apparatus is located at the bottom of

the figure. Radially extending legs correspond to the

regions with large field gradients.

Fig. 2. Sensitivity map for the new coil design.

Contribution from the radially extending legs is

reduced from Fig. 1.

165

Strategies and New Technologies for Improved Production of

Hyperpolarized 3He and

129Xe

Daniel Chonde1, Wilson Barros

1, Chih-Hao Li

1,2; Matthew S. Rosen

1,2; Michael Barlow

3;

Ross W. Mair1; Ronald L. Walsworth

1, 2

1 Harvard-Smithsonian Center for Astrophysics, Cambridge, MA, USA. 2 Department of Physics, Harvard University, Cambridge, MA, USA.

3 Sir Peter Mansfield Magnetic Resonance Centre, Univ. of Nottingham, Nottingham, UK.

NMR and MRI with hyperpolarized noble gases have a myriad of applications in medicine

and material science, including porous and granular media. In our experiments, nuclear-spin

polarized 129

Xe or 3He gases are produced by optically pumped spin-exchange with Rb

atoms. Signal-to-noise efficiency is generally limited by either low polarization or slow

production rates. One key technical problem in optimizing 129

Xe or 3He polarization levels is

the amount of obtainable resonant laser light from the current generation of Laser Diode

Arrays (LDA), which have broad optical spectra (see Fig.1). A recent solution has been the

development of Volume Holographic Gratings (VHG) that self-seed the laser to provide a

tunable narrow resonant line output with only minimal laser power reduction [1]. We have

recently employed novel, prototype-commercial VHG-narrowed lasers for spin-exchange

optical pumping of noble gases for the first time [2].

Fig 1: Laser line-shape from a

typical LDA. The pink line is

the measured LDA output; the

green line is the laser spectrum

after traversing the optical

pumping cell. The blue line is

the calculated absorption by Rb

during 129

Xe spin exchange.

Fig 2: Laser line-shape from a

commercial VHG-line-narrowed 30W

laser. The spectra were taken after the

light has traversed the 3He optical

pumping cell. Notice that the line-

width of the laser is ~ 0.2 nm and all

the laser power is absorbed when the

cell temperature is at 190°C.

Fig 3: Simulated magnetic field

plots, and contours showing the

calculated 3He T1 due to diffusion

in field gradients. The Helmholtz

pair previously used to provide

the holding field for optical

pumping (left), and the modified

three-coil design (right).

For hyperpolarized 3He in particular, the time required to polarize a suitable volume of gas (~

15-20 hours for ~ 0.5 L) has also been a major obstacle. To address this problem, we have

designed four storage cells to hold the hyperpolarized 3He gas in our system for 30-40 hours

with minimal polarization loss. This design required a much more homogeneous magnetic

field region than had been used for optical pumping alone. Adding another coil at the center

of the polarizer, in addition to the previous Helmholtz pair, has increased the working volume

by a factor of 6 as shown in Fig. 3 [3]. The storage cells will then be coupled with a switch to

Rb/K hybrid spin-exchange [4], which can enhance 3He production rates by a factor of 4.

1. F. Havermeyer et al., Volume holographic grating-based continuously tunable optical filter, Opt. Eng., 43,

2017 (2004).

2. R. Mair et al., Towards Posture-Dependent Human Pulmonary O2 Mapping using Hyperpolarized 3He and an

Open-Access MRI System, in Magnetic Resonance Microscopy, J. Seymour, S. Codd eds., in press (2008).

3. J. Wang et al., An improved Helmholtz coil and analysis of its magnetic field homogeneity, Rev. Sci.

Instrum., 73, 2175 (2002).

4. E. Babcock et al., Hybrid Spin-Exchange Optical Pumping of 3He, Phys. Rev. Lett., 91, 123003 (2003).

166

Low cost CE-NMR with microcoils for chemical detection Julie L. Herberg

1, Kristl Adams

1, Greg Klunder

1, Vasiliki Demas

1, Vince Malba

1,

Anthony Bernhardt1, Lee Evan

1, Chris Harvey

1, and Robert Maxwell

1,

1Lawrence Livermore National Laboratory, 7000 East Ave, Livermore, CA 94550

Understanding speciation in solids and solutions is important for environmental and

toxicological purposes. Capillary electrophoresis (CE) is a simple rapid separation

method that can be used to identify species in solution. One of the challenges with CE is

obtaining a method of direct detection that can provide speciation information. Nuclear

magnetic resonance (NMR) has been used to identify chemical species in aqueous

solutions and has also been coupled to CE. We are developing separation protocols to

determine the speciation of chemical complexes in solutions with minimal perturbation to

the original sample equilibrium. On-line NMR measurements will be made downstream

of the UV detector. We will discuss our development of a low-cost microcoil CE-NMR

system for in situ characterization of samples of interest.

Prepared by LLNL under Contract DE-AC52-07NA27344.

167

A new spin exchange optical pumping (SEOP) modality for

hyperpolarized 129

Xe production and use in porous media

M. J. Barlowa, N. Whiting

b, P. Nikolaou

b, N. Eschmann

b, R. Mair

c & B. M. Goodson

b

aSir Peter Mansfield Magnetic Resonance Centre, Univ. of Nottingham, Nottingham, UK

bDept of Chemistry and Biochemistry, Southern Illinois University, Carbondale, IL, USA

c Harvard-Smithsonian Center for Astrophysics, Cambridge, MA, USA

Recent studies using hyperpolarized 129

Xe as the fluidizing gas to probe exchange

measurements in this system were subject to extreme temporal averaging due to the very low 129

Xe polarization – see Fig. 1 [1]. The requirements of a continuous flow of hyperpolarized 129

Xe and that it not be diluted with another gas of high diffusivity severely limits the

polarization that can be achieved by conventional continuous flow spin exchange optical

pumping (SEOP). The low polarization is a combination of short residence time of the xenon

in the optical pumping cell and the high partial fraction of xenon leading to higher electronic

Rubidium spin destruction. In an attempt to improve SEOP efficiency and thus xenon

polarization we used a frequency narrowed laser based on Volume Bragg Grating (VBG)

techniques developed by our team [2]. Preliminary results in a continuous flow cell using a

90% Xe, 10% Nitrogen mix showed a tripling of polarization levels compared to pumping

with a conventional broad-band laser diode source [3]. Best 129

Xe polarization was achieved

with the narrowed laser pumping on the wings of the Rb absorption profile [3]. Further

investigation of pumping with a VBG laser both in continuous flow and stop-flow modalities

has lead to the discovery that the achievable 129

Xe polarization is sensitive to the

interdependence of the optimal cell temperature and the Xe density [4]. We report here for

the first time record high levels of 129

Xe polarization at high partial fraction of xenon within

the SEOP cell – see Fig. 2. The levels achievable will have profound impact on experiments

in porous and granular media allowing non-averaged single shot NMR spectra to be recorded.

This method will also enable stop-flow production for 129

Xe clinical lung MRI [4].

References:

1. T. Pavlin et al., Appl. Magn. Res. 32 (2007) 93.

2. M.J. Barlow et al., Proc. 46th ENC Providence RI USA (2005).

3. R.W. Mair et al., Magn. Reson. Imaging. 25 (2007) 549

4. N. Whiting et al., Proc. 49th

ENC Pacific Grove CA USA (2008).

−10 −5 0 5 10 15 20 25 30 350

5

10

15

20

25

Frequency (kHz)

Arb

itrar

y U

nit

1525

3545

5565

7585

95105

125 SCCM

12−8 −6 −4 −2 0 2 4 6 8 10

00.10.20.30.40.50.60.70.80.9

45 SCCM

Fig. 1: Plot of Xe spectra for range of

flow rates through the bed. Inset is for

case of 45 SCCM

Fig. 2: Xenon polarization as a function of cell

temperature and partial fraction of xenon within a

2000 Torr mix. Balance of mix is Nitrogen. All

results were taken with a 30W VBG laser module.

168

MRPM9 Author Index Abdullah O. P-014 Ackerman J.L. O-17 Adams K. P-114 Adan O.C.G. P-030 Adegbite O. P-083 Adriaensen H. P-020 Akpa B.S. P-014 Akpa B.S. P-082 Altobelli S. P-084 Amar A. O-15 Anderson R. O-04 Anderson R. P-016 Anwar M.S. I-06 Ardelean I. P-018 Ardelean I. P-079 Armstrong B. I-05 Arns C.H. O-02 Arns C.H. P-106 Arteaga N. P-097 Åslund I. O-20 Åslund I. P-097 Ayalur- Karunakaran S. P-003 Babalola B. P-080 Bachman H.N. P-105 Badea C. P-079 Baete S. O-23 Baete S. P-066 Bajaj V. I-07 Balcolm B.J. O-14 Balcolm B.J. O-24 Balcolm B.J. P-017 Balcolm B.J. P-041 Balcolm B.J. P-053 Balcolm B.J. P-080 Balcolm B.J. P-110 Balcolm B.J. P-111 Balcolm B.J. T-01 Banasik T. P-091 Barlow M.J. P-113 Barlow M.J. P-115 Barros W. P-069 Barros W. P-070 Barros W. P-113 Bartacek J. P-075 Basser P.J. I-12 Basser P.J. O-01 Basser P.J. P-062

Basser P.J. P-063 Bauer W.R. P-101 Baukh V. P-030 Bencsik M. O-11 Bencsik M. P-020 Bernhardt A. P-114 Bland P. O-04 Bland P. P-016 Blümich B. I-01 Blümich B. O-12 Blümich B. O-15 Blümich B. O-26 Blümich B. P-005 Blümich B. P-049 Blümich B. P-050 Blümich B. P-107 Blümich B. P-108 Blümich B. P-109 Bogdan M. P-079 Borneman T. P-059 Bortolotti V. P-027 Bortolotti V. P-039 Bortolotti V. P-067 Borysenko A. P-042 Bouchard L.S. I-06 Boutis G.S. O-08 Bowden J.C. P-087 Bray C.L. P-024 Brevet D. P-033 Brewer S. P-020 Brill E.R. O-18 Brosten T.R. P-037 Brown R.J.S. P-027 Brown R.J.S. P-039 Brown R.J.S. P-088 Brunetti B.G. P-108 Bryant R.G. P-024 Buljubasich L. O-12 Burcaw L. P-051 Burcaw L.M. I-04 Burgar I. P-042 Burt S.R. I-06 Butler J.P. P-095 Callaghan P.T. I-04 Callaghan P.T. O-02 Callaghan P.T. O-09 Callaghan P.T. P-019 Callaghan P.T. P-051

Callaghan P.T. P-064 Callaghan P.T. P-106 Camaiti M. P-027 Campagnoli A. P-027 Campagnoli A. P-039 Canina D. P-071 Cano P.F.J. O-24 Carl M. P-094 Casanova F. O-15 Casanova F. O-26 Casanova F. P-005 Casanova F. P-107 Casanova F. P-108 Casanova F. P-109 Cates G.D. P-094 Cerioni L. P-044 Chakraborty D. O-21 Chandrasekera T.C. I-02 Chandrasekera T.C. P-077 Charnaya E.V. P-008 Chemmi H. P-032 Chen J.H. O-18 Chen Q. O-07 Chen S.H. I-03 Chen Y.Y. I-02 Cho H.J. O-17 Cho H.J. P-061 Cho H.J. P-092 Chonde D. P-070 Chonde D. P-113 Choudhury R.P O-21 Choudhury R.P P-001 Ciampi E. P-089 Clennell B. P-042 Codd S.L. I-10 Codd S.L. P-037 Codd S.L. P-096 Collins J.H.P. I-02 Conradi M. I-14 Cool P. O-19 Cory D.G. P-059 Cox M.J. P-024 Culea E. P-050 Cunningham K. P-084 Danieli E. O-26 Danieli E. P-107 Danieli E. P-109 Datsevich L.B O-12

169

De Deene Y. P-066 De Deene Y.D. O-23 de Visser S.K. P-087 DeCarolis P. O-18 Demas V. P-114 Demco D.E. P-049 Demco D.E. P-050 Demco D.E. P-107 Devoisselle J.M. P-032 Dewhurst D. P-042 Dick J. O-24 Djabourov M. P-086 Dubin P.L. P-099 Dunckley C.P. I-02 Dunn K.J. O-25 Dvoyashkin M. P-011 Eccles C.D. P-019 Edler K.J. P-100 Edwards J.C. P-098 Erdem O.F. P-007 Erich S.J.F. P-030 Eschmann N. P-115 Evan L. P-114 Fantazzini P. P-027 Fantazzini P. P-039 Fantazzini P. P-067 Fantazzini P. P-088 Farrher G. P-018 Fechete R. P-049 Fechete R. P-050 Fechete R. P-107 Fernandez M. P-009 Ferrante G. P-024 Ferrante G. P-071 Fieremans E. P-066 Fisher A. P-082 Fleury M. P-054 Fleury M. P-086 Florea L. P-084 Fordham E.J. I-02 Fordham E.J. O-16 Fordham E.J. P-077 Franck J. P-103 Freed D.E. P-060 Freed D.E. P-076 Freiman G. P-021 Freude D. P-009 Freude D. P-010 Friedeman K. P-036 Fu S. P-055 Fukushima E. P-112

Galvosas P. I-04 Galvosas P. O-11 Galvosas P. O-21 Galvosas P. O-27 Galvosas P. P-048 Gerkema E. P-013 Gerkema E. P-075 Ghosh S. P-045 Gierada D. I-14 Gladden L.F. I-02 Gladden L.F. O-16 Gladden L.F. P-034 Gladden L.F. P-035 Gladden L.F. P-077 Gladden L.F. P-081 Gladden L.F. P-082 Glögler S. P-108 Goddard Y. P-024 Golzi P. P-071 Gombia M. P-027 Gombia M. P-039 Gombia M. P-067 Gombia M. P-088 Goñi A.R. P-044 Goodson B.M. P-115 Goudappel G.J.W. P-012 Graf von der Schulenburg D.A O-16 Gratz M. O-27 Gratz M. P-048 Grebenkov D.S. P-074 Grebenkov D.S. T-02 Green D.P. O-24 Grinberg F. O-19 Grinberg F. P-010 Grunin L. P-098 Grunin Y. P-098 Gruwel M.L.H. P-090 Haase J. P-008 Haber A. P-107 Haber- Pohlmeier S. P-004 Hall C. P-017 Halse M.E. O-09 Ham D. I-17 Hamilton A. P-017 Han H. P-110 Han S. I-05 Han S. P-068 Hansen E.W. P-102 Harel E. I-07

Harris R. I-09 Harvey C. P-114 Hata N. P-025 Hattori M. P-025 Hayamizu K. P-025 Heidler R. P-105 Heinrich D. P-044 Herberg J.L O-07 Herberg J.L P-114 Herold V. P-101 Hersman F.W. P-095 Holland D.J. O-16 Holt J.K. O-07 Homan N. O-22 Hopper T. P-103 Horch C. P-048 Hornak J.P. P-024 Hornemann J.A. P-096 Hrovat M.I. P-095 Hui H. O-14 Huinink H.P. P-030 Hunter M.W. P-064 Hürlimann M.D. P-059 Hürlimann M.D. P-060 Hürlimann M.D. P-076 Isahkarov T. O-08 Islam T. O-08 Jackson A.N. P-064 Jakob P.M. P-101 Jensen J. P-093 Jess A. O-12 Johansen Y. P-105 Johns M.L. I-02 Johns M.L. I-09 Johns M.L. O-16 Johns M.L. P-073 Johns M.L. P-077 Johns M.L. P-082 Johnson G. P-093 Jokisaari J. P-026 Kampf T. P-101 Kannangara L. O-08 Karasev D. P-098 Kärger J. O-19 Kärger J. P-009 Kärger J. P-010 Kärger J. P-011 Kauer P. O-08 Kaur R. O-11 Kausik R. I-05 Kausik R. P-068

170

Kayitmazer A.B. P-099 Khokhlov A. P-011 Kidena K. P-040 Kimmich R. P-018 Kleinberg R.L. I-15 Klunder G. P-114 Koay C.G. P-062 Koay C.G. P-063 Kopinga K. P-028 Koptyug I.V. I-06 Koptyug I.V. O-13 Korb J.P. O-05 Korb J.P. P-021 Korb J.P. P-032 Korb J.P. P-033 Kovtunov K.V. I-06 Kupka T.W. P-090 Lasic S. P-097 Laxa B.U. O-18 Le Bideau J. P-033 Leblond J. P-086 Lemahieu I. P-066 Lens P. P-075 Leu G. I-15 Levitz P. P-032 Levitz P. P-033 Levitz P.T. O-05 Li C.H. P-113 Li L. P-053 Li L. P-080 Li M. O-14 Liang H. P-068 Liao G. P-055 Ligneul P. P-021 Lin Y.Y. I-08 Lounila J. P-026 Lowe J.P. P-100 Lutey B. I-14 MacGregor R. P-041 MacMillan B. P-017 Magin R.L. P-014 Mair R.W. P-070 Mair R.W. P-113 Mair R.W. P-115 Malba V. P-114 Mananga E. O-08 Mang T. P-038 Mantle M.D. I-02 Mantle M.D. P-034 Marica F. P-017 Martini G. P-071

Martins J. P-038 Mastikhin I.V. P-111 Mattea C. P-002 Matthews S. P-082 Maxwell R. P-114 McAloon M.J. O-24 McCarney E.R. I-05 McCarney E.R. P-068 McDonald P.J. P-015 McDonald P.J. P-089 Melean Y. P-106 Menjoge A.R. P-099 Meynen V. O-19 Michel D. P-007 Michel D. P-008 Miliani C. P-108 Miller G.W. P-094 Mitchell J. I-02 Mitchell J. O-16 Mitchell J. P-077 Mogilevsky G. O-07 Mohoric A. P-065 Moldovan D. P-049 Moldovan D. P-050 Momot K. P-087 Morris R. O-11 Mugler III J.P. P-094 Muradyan I. P-095 Mutina A.R. P-023 Nadler W. P-101 Nakashima Y. P-112 Nestle N. O-11 Newling B. P-072 Newling B. P-083 Nguyen T.T.M. P-081 Nicot B. P-021 Nicot B. P-086 Nikolaou P. P-115 Nikolskaya E. P-098 Nilsson M. O-20 Nowacka A. O-20 Ntekim A. O-08 Nydén M. P-012 O’Connor R. O-18 Oehmichen T. O-12 Ohira A. P-040 Ohkubo T. P-040 Osán T. P-044 Osuna M. P-075 Ozarslan E. O-01 Ozarslan E. P-062

Ozarslan E. P-063 Parasoglou P. P-073 Parnau A. P-079 Patz S. P-095 Peddis F. P-039 Pel L. O-06 Pel L. P-028 Pel L. P-029 Pel L. P-030 Pel L. P-031 Pera H. O-22 Perkins E.L. P-100 Perlo J. O-26 Perlo J. P-107 Perlo J. P-108 Perlo J. P-109 Perrie Y. O-11 Petit D. P-032 Petit D. P-033 Petrov O. P-053 Petrov O.V. P-111 Philippi J. P-075 Pines A. I-06 Pines A. I-07 Pitts S. P-089 Pohlmeier A. P-005 Pohlmeier A. P-006 Polders D. I-04 Pomerantz A.E. O-03 Pope J.M. P-087 Powell H. P-073 Prange M. P-058 Presciutti F. P-108 Pusiol D.J. P-044 Quirk J. I-14 Ralston J. P-042 Rasburn J. P-073 Regatte R. P-092 Ren X.H. P-061 Renner C. O-08 Rigby S.P. P-100 Robert H. I-16 Rodin V. P-015 Romanenko K. P-096 Romanenko K.V. P-037 Romanova E.E. P-010 Romero P. P-056 Romero-Zeron L. P-080 Rosen M.S. P-070 Rosen M.S. P-113 Rosicka P. P-091

171

Rumala Y.S. O-08 Ryu S. O-17 Ryu S. P-052 Ryu S. P-104 Saha I. P-076 Saha I. P-103 Saidov T. P-029 Saidov T. P-031 Sankey M.H. I-02 Scheidegger R. P-070 Scheven U.M. I-09 Scheven U.M. P-060 Schlayer S. P-048 Schönhoff M. O-21 Schönhoff M. P-043 Schurr U. P-006 Schweitzer M. P-092 Sederman A.J. I-02 Sederman A.J. P-034 Sederman A.J. P-073 Sederman A.J. P-081 Sederman A.J. P-082 Sedev R. P-042 Sedykh P. P-007 Sedykh P. P-008 Sersa I. P-065 Seymour J.D. I-10 Seymour J.D. P-037 Seymour J.D. P-096 Sgamellotti A. P-108 Shapley N.C. O-10 Shushakov O.A. P-022 Sigmund E.E. P-092 Singer S. O-18 Singleton S. P-089 Skalinski M. O-25 Skirda V.D. P-023 Smith M. I-07 Song K.M. P-035 Song Y.Q. O-17 Song Y.Q. O-03 Song Y.Q. P-052 Song Y.Q. P-061 Song Y.Q. P-092 Song Y.Q. P-103 Sophie S.W. P-037 Sørland G.H. P-057 Sørland G.H. P-102 Soualem J. P-054 Spring M. P-108 Stadermann M. O-07

Stallmach F. P-036 Stapf S. O-12 Stapf S. P-001 Stapf S. P-002 Stapf S. P-003 Stapf S. P-004 Stapf S. P-038 Stepisnik J. P-065 Stevens W.J.J. O-19 Strange J.H. O-04 Strange J.H. P-016 Stucky G. P-068 Sucre O.E. P-005 Sukstanskii A. I-14 Sun B. O-25 Sun B. P-103 Sur S. P-024 Sykora S. P-067 Tanasiewicz M.M. P-090 Tervonen H. P-026 Theiss T. I-06 Thomsen C. P-044 Tilke P. O-03 Tohidi B. O-04 Tohidi B. P-016 Tomanek B. P-090 Topgaard D. O-20 Topgaard D. P-097 Tourné-Péteilh C. P-032 Trampel R. P-093 Tyrankiewicz U. P-091 Ukkelberg Å. P-057 Ukkelberg Å. P-102 Ulrich K. O-19 Utsuzawa S. P-112 Vacchelli E. P-067 Vacchi C. P-071 Valiullin R. P-011 Valori A. P-015 Van As H. O-22 Van As H. P-013 Van As H. P-075 van der Heijden G.H.A. O-06 van Dusschoten D. P-006 van Duynhoven J.P.M. P-012 van Duynhoven J.P.M. P-013 van Ginkel M. P-089 Varner J. P-072 Varner J. P-083 Vasenkov S. P-099 Vereecken H. P-006

Vergeldt F. P-013 Vergeldt F. P-075 Vergeldt F.J. O-22 Vioux A. P-033 Vittur F. P-088 Voronina V. P-028 Walsworth R.L P-070 Walsworth R.L P-113 Wang H. O-07 Washburn K.E. I-04 Washburn K.E. O-02 Washburn K.E. P-019 Washburn K.E. P-106 Webber J.B.W. O-04 Webber J.B.W. P-016 Weber D. P-034 Wedeen V. I-13 Weglarz W.P. P-090 Weglarz W.P. P-091 Wehrli F.W. I-11 Wei D. O-08 Weihermüller L. P-006 Wellard R.M. P-087 Wende C. P-043 Whiting N. P-115 Widerøe H.C. P-057 Widerøe H.C. P-102 Windt C. O-22 Windt C. P-013 Witek M. P-013 Woods J. I-14 Wu Y. O-07 Xi C. O-10 Xiao L. P-055 Xie R. P-055 Xie Z.H. P-045 Yablonskiy D. I-14 Yang Z. P-083 Yu H. P-055 Yunus K. P-082 Zhang J. P-041 Zielinski L. P-052 Zielinski L. P-076 Ziener C.H. P-101

172