990 DOI: 10.1021/la902444x Langmuir 2010, 26(2), 990–1001Published on Web 09/18/2009

pubs.acs.org/Langmuir

© 2009 American Chemical Society

Development of Langmuir-Schaeffer Cellulose Nanocrystal Monolayers

and Their Interfacial Behaviors

Youssef Habibi,† Ingrid Hoeger,† Stephen S. Kelley,† and Orlando J. Rojas*,†,‡

†Department of Forest Biomaterials, North Carolina State University, Raleigh, North Carolina 27695-8005, and‡Department of Forest Products Technology, Faculty of Chemistry andMaterials Sciences, Helsinki University

of Technology, P.O. Box 3320, FIN-02015 TKK, Espoo, Finland

Received July 7, 2009. Revised Manuscript Received August 27, 2009

Model cellulose surfaces based on cellulose nanocrystals (CNs) were prepared by the Langmuir-Schaeffer technique.Cellulose nanocrystals were obtained by acid hydrolysis of different natural fibers, producing rodlike nanoparticles withdifferences in charge density, aspect ratio, and crystallinity. Dioctadecyldimethylammonium bromide (DODA-Br)cationic surfactant was used to create CN-DODA complexes that allowed transfer of the CNs from the air/liquidinterface in an aqueous suspension to hydrophobic solid substrates. Langmuir-Schaeffer horizontal deposition atvarious surface pressures was employed to carry out such particle transfer that resulted in CN monolayers coating thesubstrate. The morphology and chemical composition of the CN films were characterized by using atomic forcemicroscopy (AFM) and X-ray photoelectron spectroscopy (XPS). Also, their swelling behavior and stability aftertreatment with aqueous and alkaline solutions were studied using quartz crystal microgravimetry (QCM). Overall, it isconcluded that the Langmuir-Schaeffer method can be used to produce single coating layers of CNs that were shown tobe smooth, stable, and strongly attached to the solid support. The packing density of the films was controlled byselecting the right combination of surface pressure during transfer to the solid substrate and the amount of CNsavailable relative to the cationic charges at the interface.

Introduction

Cellulose model films have been extensively studied over thepast decade, especially in efforts to better understand a range ofinteractions and surface phenomena between this ubiquitousnatural polymer and solvents, enzymes, chemical additives, andother materials. Also, cellulose films have been used to developnovel suprastructures relevant to natural and industrial systems.Obtaining fundamental information from cellulose fibers ischallenging due to their chemically complex and heterogeneousfeatures, as well as their characteristic nanoscale surface topo-graphy. A review relevant to cellulose model layers has beenpublished recently by Kontturi et al.1 Regenerated cellulosedissolved in N-methylmorpholine oxide (NMMO) or lithiumchloride inN,N-dimethylacetamide (LiCl-DMA) as well as cellu-lose derivatives (mostly trimethylsilylcellulose, TMSC) have beenreported as useful in the preparation of thin films of cellulose.However, films obtained from such systems may not completelyrepresent native cellulose fibers because of differences inherent tochemical composition, morphology, and crystallinity. In fact, inthe plant cell wall, biosynthesis of cellulose fiber results from thecombined action of enzymatically controlled polymerization andsubsequent crystallization. These events are orchestrated byspecific enzymatic terminal complexes (TCs) that act as biologicalspinnerets producing crystalline cellulose nanofibrils.2 Fromstructural and morphological points of view, cellulose nanofibrilsare the smallest structural elements of the cellulose fibers.

By deconstruction of the cell wall of fibers, nanofibril elementscan be obtained which can be then employed in the development

of two- and three-dimensional structures that more accuratelyrepresent native cellulose. These substrates include nanofibrillaror microfibrillar cellulose (NFC or MFC, respectively)3-6 andcellulose nanocrystals.5,7,8 NFC films have been recently preparedby spin-coating aqueous suspensions of cellulose nanofibrilsobtained by combining enzymatic and mechanical treatments todisintegrate cellulose fiber pulp.3,4 The substructural nanoscaleunits or nano/microfibrils9,10 that are generated by mechanicalshearing processes differ from regenerated cellulose in theirinherent morphology, chemical composition, and crystallinity.For example, films of NFC typically contain both crystallinecellulose I and amorphous regions6,11,12 while spin-cast films areamorphous or cellulose allomorph type II.

It has been proposed that if the TCs are not perturbed duringcellulose biosynthesis, they can generate endless microfibrilshaving only a limited number of amorphous “defects” distributedalong the cellulose microfibrils.13 After acid hydrolysis of theamorphous defects, true cellulose nanorods or nanocrystals

*Corresponding author. E-mail: ojrojas@ncsu.edu. Telephone:+1-919-5137494. Fax +1-919-515 6302(1) Kontturi, E.; Tammelin, T.; €Osterberg, M. Chem. Soc. Rev. 2006, 35, 1287–

1304.(2) Saxena, I. M.; Brown, R. M. J. Ann. Bot. 2005, 96(1), 9–21.

(3) Ahola, S.; Oesterberg, M.; Laine, J. Cellulose 2008, 15(2), 303–314.(4) Ahola, S.; Salmi, J.; Johansson, L. S.; Laine, J.; Osterberg, M. Biomacro-

molecules 2008, 9(4), 1273–1282.(5) Ahola, S.; Turon, X.; Osterberg, M.; Laine, J.; Rojas, O. J. Langmuir 2008,

24(20), 11592–11599.(6) Aulin, C.; Ahola, S.; Josefsson, P.; Nishino, T.; Hirose, Y.; Oesterberg, M.;

Wagberg, L. Langmuir 2009, 25, 7675–7685.(7) Edgar, C. D.; Gray, D. G. Cellulose 2003, 10(4), 299–306.(8) Habibi, Y.; Foulon, L.; Aguie-Beghin, V.; Molinari, M.; Douillard, R.

J. Colloid Interface Sci. 2007, 316(2), 388–397.(9) Henriksson, M.; Henriksson, G.; Berglund, L. A.; Lindstroem, T. Eur.

Polym. J. 2007, 43(8), 3434–3441.(10) Turbak, A. F.; Snyder, F. W.; Sandberg, K. R. J. Appl. Polym. Sci.: Appl.

Polym. Symp. 1983, 37, 815–827.(11) Wagberg, L.; Decher, G.; Norgren, M.; Lindstrom, T.; Ankerfors, M.;

Axnas, K. Langmuir 2008, 24(3), 784–795.(12) Aulin, C.; Varga, I.; Claesson Per, M.; Wagberg, L.; Lindstrom, T.

Langmuir 2008, 24(6), 2509–2518.(13) Rowland, S. P.; Roberts, E. J. J. Polym. Sci., Part A: Polym. Chem. 1972, 10

(8), 2447–2461.

DOI: 10.1021/la902444x 991Langmuir 2010, 26(2), 990–1001

Habibi et al. Article

(CNs) can be obtained. Thus, CNs have the same cellulose Imorphology and crystallinity as that of the original cellulosefibers. In addition to the properties provided by NFC, it isexpected that films from CNs can be used to study the funda-mental interactions between cellulose surfaces and func-tionalization processes, surface polymer interactions, enzymereactions, and so on. Furthermore, recent interest in developingcoatings with ultrahigh surface strength and special optical,electrical, and mechanical properties are central interests in CNresearch.14

Smooth model films based on CNs have been prepared byLangmuir-Blodgett (LB)8 and spin-coating techniques.7 In filmsobtained by vertical deposition using the LB technique, the effectsof different CN aspect ratio, morphology, and crystallinity havebeen investigated. The resulting films, deposited on a solidsupport, were characterized as monolayers with thicknessesequivalent to the width of the precursor cellulose nanorods. Also,CNsdeposited usingLB techniques can be directed inmore or lessdefined orientation, usually in the dipping direction employed totransfer the monolayer to the solid support. In contrast to LBfilms, spin-coated films usually form multilayers and show apreferred radial microcrystal orientation.

LB deposition also allows for the manufacture of multilayeredstructures by successive transfers of monolayers from the air/water interface to the solid support.15 As a principal tool forfabrication of ultrathin and well-organized molecular films in thenanoscale, LB has been successfully used to produce cellulosefilms from other sources different than cellulose nanocrystals.However, typical LB deposition involves the use of subphases inwhich the target molecule is spread onto by using a suitable(spreading) solvent. The required solubilization of cellulose(in the spreading solvent) for such applications limits its use.Nevertheless, cellulose in such systems can be converted tocellulose after film deposition to produce amorphous celluloseor cellulose II. The spreading of cellulose derivatives at the air/water interface and respective surface isotherms were first inves-tigated more than 70 years ago.16,17 Monolayers and multilayersof cellulose esters with various alkyl chain lengths were firsttransferred onto solid supports by Kawaguchi et al. using thehorizontal and vertical lifting techniques.18,19 Since then, LB filmsof cellulose derivatives with different alkyl side chains, so-called‘‘hairy rod’’ ultrathin layers, have been extensively studied,mainly with respect to their spreading behavior and film proper-ties.20-26

Among the cellulose derivatives to prepare LBmodel cellulosefilms, TMSC is the most commonly used one,27-33 this is becauseof the readily and effective conversion of TMSC to regeneratedcellulose by desilylation. Desilylation can be performed directlyon the solid substrates by simply exposing the film toHCl vapors,29,30 thereby producing pure and ultrasmooth cellulose films.TMSC films have been conveniently deposited on hydrophobicsubstrates using either vertical (Langmuir-Blodgett) or horizontal(Langmuir-Schaeffer, LS) applications and then regenerated toproduce cellulose films with varying thickness depending on thenumber of dipping cycles.29-31

In this work, the Langmuir-Schaeffer technique was used todevelop novel CN films. Such technique was chosen over otherprocedures in order to facilitate the transfer of CNs from anaqueous subphase (where they were dispersed) to a solid support.An insoluble surfactant (dioctadecyldimethylammonium bro-mide, DODA), located at the air/water interface, was used as acarrier of the CNs. After horizontal LS deposition, the surfactantlocated between the CNs and the solid support acted as a binderlayer, thereby producing a top layer of surfactant-free CNs (theactive side of the solid). The effect of different morphologicalparameters of the precursor CN-DODA system (used to pro-duce highly packedCN films) is discussed. Finally, the stability ofthe obtained CN films to alkaline treatment is discussed.

Experimental Section

Materials. Three different fiber sources were used to producethe CNs including bleached cotton and sisal fibers (donated byDanforth International, NJ) and ramie fibers (Stucken MelchersGmbH & Co., Germany). Sulfuric acid, dioctadecyldimethylam-monium bromide (DODA-Br), 1-dodecanthiol, and all solventswere purchased fromSigma-Aldrich and used as received.Milli-Qwater with resistivity > 18 MΩ 3 cm was used in all experiments.Gold-coated quartz resonators used in the QCM-D experimentswere acquired from Q-Sense (V€astra Fr€olunda, Sweden).Methods. Preparation of Cellulose Nanocrystals. Acid

hydrolysis with 65% (w/w) H2SO4 solution was applied tohomogeneous suspensions of purified fibers from sisal, cotton,and ramie (see ref 8 for details about associated treatments). Thehydrolysis was carried out at 55 �C for 30 min under continuousstirring. The resulting suspensions were filtered through sinteredcrucibles (No. 1) to remove unhydrolyzed fibers, and the CNsuspensions were then washed with water and recovered bycentrifugationat 10000 rpm (10 �C) for 10min each.The resultingsuspensions were then dialyzed against deionized water and thenagainst Milli-Q water for a few weeks. The final CN stocksuspensions were stored at 4 �C after adding a few drops ofchloroform to prevent any possible biological growth. The sulfurcontent (fromsulfate groupspresent on theCNsafter sulfuric acidhydrolysis) was determined by elemental analysis.

Preparation of Gold Surfaces.Gold-coatedQCM resonatorswere cleaned with a H2SO4/H2O2 (7:3) solution for 10min, rinsedthoroughly with water, and blown dry with a nitrogen jet. Theywere hydrophobized by a monolayer of 1-dodecanthiol that wasself-assembled on gold after immersion overnight in a 10-3 Msolution in pure ethanol.34 After thiolation, the gold sensors werewashed thoroughly with pure ethanol to remove excess and

(14) Cranston, E. D.; Gray, D. G. Biomacromolecules 2006, 7(9), 2522–2530.(15) Petty, M. C. Langmuir-Blodgett films: An introduction; Cambridge University

Press: New York, 1996; p 245.(16) Adam, N. K. Trans. Faraday Soc. 1933, 29, 90–110.(17) Harding, J. B.; Adam, N. K. Trans. Faraday Soc. 1933, 29, 837–844.(18) Kawaguchi, T.; Nakahara, H.; Fukuda, K. Thin Solid Films 1985, 133,

29–38.(19) Kawaguchi, T.; Nakahara, H.; Fukuda, K. J. Colloid Interface Sci. 1985,

104(1), 290–293.(20) Basque, P.; de Gunzbourg, A.; Rondeau, P.; Ritcey, A. M. Langmuir 1996,

12(23), 5614–5619.(21) Fischer, P.; Brooks, C. F.; Fuller, G.G.; Ritcey, A.M.; Xiao, Y.; Rahem, T.

Langmuir 2000, 16(2), 726–734.(22) Kondo, T.; Kasai, W.; Kuga, S. Cell. Commun. 2002, 9(1), 18–22.(23) Maaroufi, A.; Mao, L.; Ritcey, A. M. Macromol. Symp. 1994, 87, 25–33

(Polymers: Progress in Chemistry and Physics) .(24) Schaub, M.; Fakirov, C.; Schmidt, A.; Lieser, G.; Wenz, G.; Wegner, G.;

Albouy, P. A.; Wu, H.; Foster, M. D.; Majrkzak, C.; Satija, S. Macromolecules1995, 28(4), 1221–1228.(25) Xiao, Y.; Ritcey, A. M. Langmuir 2000, 16(9), 4252–4258.(26) Zhang, Y.; Tun, Z.; Ritcey, A. M. Langmuir 2004, 20(15), 6187–6194.(27) Buchholz, V.; Adler, P.; Baecker, M.; Hoelle, W.; Simon, A.; Wegner, G.

Langmuir 1997, 13(12), 3206–3209.(28) Buchholz, V.; Wegner, G.; Stemme, S.; Odberg, L. Adv. Mater. 1996, 8(5),

399.

(29) Holmberg, M.; Berg, J.; Stemme, S.; Odberg, L.; Rasmusson, J.; Claesson,P. J. Colloid Interface Sci. 1997, 186(2), 369–381.

(30) Schaub,M.;Wenz, G.;Wegner, G.; Stein, A.; Klemm,D.Adv.Mater. 1993,5(12), 919–922.

(31) Tammelin, T.; Saarinen, T.; Osterberg, M.; Laine, J. Cellulose 2006, 13(5),519–535.

(32) Kontturi, K. S.; Tammelin, T.; Johansson, L.-S.; Stenius, P.Langmuir 2008,24(9), 4743–4749.

(33) Cromme, P.; Zollfrank, C.; Muller, L.; Muller, F. A.; Greil, P. Mater. Sci.Eng., C 2007, 27(1), 1–7.

(34) Stubenrauch, C.; Rojas, O. J.; Schlarmann, J.; Claesson, P. M. Langmuir2004, 20, 4977–4988.

992 DOI: 10.1021/la902444x Langmuir 2010, 26(2), 990–1001

Article Habibi et al.

unbound alkanethiol molecules. The thiolated, hydrophobicsurfaces were kept in a desiccator to avoid any contamination.The contact angles for the gold sensors before and after hydro-phobization were measured at room temperature (22 �C) with aRam�e-Hart contact angle goniometer model 100-00-115 (Ram�e-Hart InstrumentCo., NJ) using sessile drops ofMilli-Qwater andresulted to be 45 ( 2 and 90 ( 3, respectively.

CN Film Preparation. The formation of DODA monolayersat the air/water interface and their deposition on hydrophobizedgold substrates (thiolatedQCMsensors) wereperformedat20(1 �Cusing a Langmuir trough (KSV Technology, Minitrough modelwith 7.5�330 mm dimensions) equipped with a Wilhelmy-typefilm balance. Twentymicroliters of theDODAsolution (1mg/mL)was spread on the surfaces of freshly sonicated aqueous suspen-sions containing the respective concentration of CNs (sisal,cotton, or ramie CNs). The system was allowed to stabilize for15 min, and then the surface was compressed by reducing thedistance between the barriers on the through at a rate of 2 mm/min. At the appropriate surface pressure (or surface packing),material transfer on the thiolated gold sensors was performed byusing the horizontal deposition method, that is, by making thesolid substrate (positioned horizontally) to contact the com-pressed surface at a rate of 1 mm/min. The solid support wasallowed to make contact with the CN-DODA complex layer forfew seconds, and then the solid substrate was removed at aretraction rate of 1 mm/min. The substrate was then allowed todry for 10 min before removing it from the holder. The positivelycharged DODA molecules that formed an insoluble monolayeron the surface with a given packing density (surface pressure)acted as a carrier for the negatively charged CNs dispersed in theaqueous subphase. Hence, by using this approach, the solidsupport was coated with three layers consisting of alkanethiol(hydrophobic layer), DODA (binding layer), and the CNs (activelayer). These supported films are noted thereafter as “LS films”.

Quartz CrystalMicrogravimetry.Tests on swelling, stability,and enzymatic hydrolysis of the LS films prepared with CN fromdifferent sources were performed by using a quartz crystal micro-balance with dissipation monitoring (QCM-D) fromQ-Sense AB(model E-4, V€astra Fr€olunda, Sweden). The temperature in ourexperiments was varied between 20 and 40 ((0.02) �C using aPeltier element built in the QCM apparatus. For swelling andstability experiments, the LS films were thoroughly rinsed withwater and dried with nitrogen before testing. Water was injectedinto theQCMchamber at a flow rate 0.2mL/minwith aperistalticpump until the deposited films reached equilibrium. Afterward,freshly prepared sodium hydroxide solution (0.01 , 0.05, or 0.1MNaOH concentration) was injected at a flow rate of 0.2 mL/minfor 20min.Any shift inQCMfrequencyΔf or dissipationΔDwasrecorded as a function of time every 1.25 s.

Atomic Force Microscopy (AFM). AFM imaging was per-formed using a NanoScope III D3000 multimode scanning probemicroscope from Digital Instruments Inc. (Santa Barbara, CA).Surface morphologies of the LS films were analyzed in tappingmode, in air at room temperature, using a single crystal siliconSPM sensor tip (Pointprobe NCHR-50 from Nanoworld Inno-vative Technologies, Neuchatel, Switzerland) with a resonancefrequency of 320 kHz and force constant of 42 N/m. The drivefrequency of the cantilever was about 300 kHz, and the topo-graphic structure of the surface was recorded by maintaining theoscillation amplitude of the probe by using feedback to alter thetip-sample separation (defined as the z-axis). Several scan sizeswere performed, but the images reported here correspond to 5�5μm2

scans. Images were acquired on at least five different areas of thesample, and those representative of the respective surface topo-graphy are reported here.

Transmission ElectronMicroscopy (TEM). For TEM, a fewdrops of cellulose nanocrystal suspension in water (0.01%w/v)were deposited on carbon-coated electron microscope grids, andthen negatively stained with uranyl acetate and allowed to dry.

The grids were observed with a Hitachi HF2000 transmissionelectronmicroscope operated at an accelerating voltage of 80 kV.

Wide-Angle X-ray Diffraction (WAXS). WAXS experi-ments were carried out at ambient temperature on CN filmsobtained by casting-evaporation of the respective suspensions(the LS films were too thin to make practical WAXSmeasurements). A Philips (XLF, ATPS XRD 1000) diffract-ometer operated with a Cu KR anode (λ=0.15406 nm) was usedwith a 2θ range from 5� to 40� with steps of 0.05� and a countingtime of 60 s. The crystallinity index (CrI) was calculated by usingthe method of Segal et al.35 according to

CrI¼ 1-Iam

I002

where I002 is the maximum intensity (arbitrary units) of the 002lattice diffraction and Iam is the intensity of diffraction (in thesame units) at 2θ=18�.

The dimension of the crystal Dhkl was evaluated by usingScherrer’s formula:36

Dhkl ¼ 0:9λ

β1=2 cos θ

where θ is the diffraction angle, λ is theX-raywavelength, and β1/2is the peak width at half of maximum intensity.

X-ray Photoelectron Spectroscopy (XPS). XPS measure-ments were performed on dry Langmuir-Schaeffer films consist-ing of a singleDODA layer or after CNs-DODAwere depositedon the hydrophobized gold substrates (the surfaceswere dried andthen kept in a vacuum oven for a few hours at 40 �C prior toanalysis). A Kratos AXIS Ultra photoelectron spectrometer wasoperated at room temperature with a base pressure of 10-9 mbar.The monochromatic Al K X-ray source was operated at 300 W(15 kV, 20 mA) at a radiation angle of 15� for an estimatedpenetration depth of around 10 nm. Low-resolution survey scanswere taken with a 1 eV step and 80 eV analyzer pass energy andhigh-resolution spectra were taken with a 0.1 eV step and 20 eVanalyzer pass energy. Quantitative XPS analyses were performedwith the Kratos Vision software (version 2.1.2). The atomicconcentrationswere calculated from the photoelectron peak areasby usingGaussian-Lorentzian deconvolution. CarbonC1s spec-tra were resolved into different contributions of bonded carbon,namely, carbonwithout oxygen bonds (C-CandC-Hx), carbonwith one oxygen bond (C-O), carbon with two oxygen bonds(O-C-O), and carbon with three oxygen bonds (OdC;O).37

The chemical shifts were taken from the literature, and the spectrawere charge corrected by setting the carbon-without-oxygen-bond contribution in the C1s emission at 285.0 eV.37

Ellipsometry. A variable angle spectroscopic ellipsometer(VASE, J. A. Woollam Co., Inc.) with a wide spectral range of193-2500 nm was used to measure the film thickness. Data werecollected at different angles of incidence (60, 65, 70, 75, and 80�),and the spectral range probedwas 350-850 nm. ACauchymodelwas used to fit the ellipsometric angle (Δ,Ψ) data using the VASEsoftware (J. A. Woollam Co.) by minimizing the mean-squarederror. In the calculation, the effective indices of refraction of thethiol-DODA layer and cellulose nanocrystals layers were takento be equal to 1.5.

Results and Discussion

Characterization of Cellulose Nanocrystals. Acid hydro-lysis of the cellulose fibers (ramie, cotton, and sisal) produced

(35) Segal, L.; Creely, J. J.; Martin, A. E.; Conrad, C. M. Text. Res. J. 1959, 29,786–94.

(36) Klug, H. P.; Alexander, L. E.X-Ray Diffraction Procedures for Polycrystallineand Amorphous Materials; Wiley Interscience: New York, 1954; p 491.

(37) Watts, J. F.; Wolstenholme, J. An Introduction to Surface Analysis by XPSand AES; John Wiley & Sons Ltd: Chichester, 2003.

DOI: 10.1021/la902444x 993Langmuir 2010, 26(2), 990–1001

Habibi et al. Article

nanometer-scale rodlike crystalline residues. Typical TEMimagesare shown in Figure 1, and the corresponding dimensions can befound in Table 1. The CN length was generally of the order of afew hundred nanometers. However, the precise physical dimen-sions of the crystallites depended on several factors. The inherentnature of acid hydrolysis as a diffusion-controlled process and themechanical action involved in the CN recovery both prevented anarrow size distribution. Nanocrystals derived from sisal fiberspresented the smallest width (3-6 nm), while those from cotton

exhibited the largest width (10-15 nm). These values agreed wellwith those from the literature.14

WAXS diffraction patterns obtained for the CNs from differ-ent sources are also shown in Figure 1. They are typical of nativecrystalline cellulosic materials. Indeed, according to the mono-clinic indexation reported by Sugiyama et al., cellulose I ischaracterized by five well-defined diffraction peaks at 14.8�(d-spacing, 0.60 nm), 16.5� (0.54 nm), 21� (0.44), 22.6� (0.40 nm),and 34.5� (0.258 nm) that correspond to lattice planes (110), (110),

Figure 1. TEMmicrographs and X-ray diffraction patterns of (A) sisal, (B) ramie, and (C) cotton CNs.

Table 1. Characteristic Parameters of the Investigated Cellulose

Nanocrystalsa

CrI (%) length (nm) width (nm) charge (e/nm2)

sisal 81 115( 21 5.0( 1.5 0.37ramie 88 185( 25 6.5( 0.7 0.30cotton 88 140( 15 14.3( 2.0 0.32

aCrystallinity index of CNs were obtained fromWAXS, dimensions(length and width) from TEM images, and charge per unit area wascalculated using the procedure in ref 38.

Table 2. Molar Ratios of Sulfate Groups (From Elemental Analysis)to DODA Molecules Used in Langmuir-Schaeffer Experiments

n(S)/n(DODA)

wt(CNs)/wt(DODA) sisal cotton ramie

125 11.0 10.0 10.2250 21.6 19.3 20.3375 32.5 29.0 30.5500 43.2 38.6 40.7

994 DOI: 10.1021/la902444x Langmuir 2010, 26(2), 990–1001

Article Habibi et al.

(102)/(012), (200), and (040), respectively.39 These peaks areclearly shown in the case of highly crystalline cotton and ramieCNs. However, WAXS patterns obtained for sisal CNs arecharacterized by two peaks centered at 16� and 22�. One canobserve that in this case the peaks at 14.8� and 16.5� and those at21� and 22.6�, typical of cellulose I, tended to overlap. Thisobservation is believed to be a consequence of the small diameterof these cellulose crystals, as is apparent from the TEM images. Infact, native cellulose crystals are bundles of several cellulosechains stabilized by hydrogen bonds. Each cellulose cluster iscreated in the plant cell wall during the biosynthesis through theTCs. Depending on the origin of the cell wall, various arrange-ments ofTCs canoccur toproduce cellulose crystalswith differentshapes, geometries, and dimensions.14 Having diameters of only3-6 nm, sisal nanocrystals are expected to contain smalleramounts of cellulose chains. With such a small number ofcellulose chains, most of them are located at the surface of thecrystal, and therefore, only crystals with disorganized lateralcohesion, and no periodic lateral hydrogen bonds, are expectedto occur. The average cross-sectional dimensions of the differentCNswere also evaluated from the diffraction patterns by applyingScherrer’s formula. The estimated widths were 4.5, 5.3, and6.0 nm for CNs from sisal, ramie, and cotton, respectively. Thesevalues differ from the widths obtained by TEM, since in the lattercase the measurements are made in dry conditions which oftenleads to nanocrystal aggregation (a situation that is specially

observed in the case of cotton CNs; see Figure 1). Also, contrastartifacts could contribute to overestimation of the respectivedimensions by TEM. By using the dimensions obtained fromWAXS diffraction and assuming that CNs have a square crosssection geometry as considered byElazzouzi-Harfaoui et al.,40 therespective crystals contained a calculated number of cellulosechains of ca. 60 (sisal), 83 (ramie), and 110 (cotton), in agreementwith other reports.41 The size differences among the various CNsstudied here are shown in Table 1 along with the crystallinityindex and surface charge density. Cotton and ramie nanocrystalshad the highest degree of crystallinity (ca. 88%), while thecrystallinity of sisal nanocrystals was the lowest (ca. 81%).Langmuir Films of CN-DODA Complexes at the Air/

Water Interface. DODA is a cationic surfactant that has beenused in Langmuir films to investigate adsorption and bindingprocesses involving anionic polyelectrolytes at the air/water inter-face.42 Likewise, rodlike CNs, which are negatively charged dueto sulfate groups generated during sulfuric acid hydrolysis, areexpected to interact strongly with cationic DODA surfactants byway of electrostatic interactions.8 The concentration of surfacesulfate moieties as determined by sulfur elemental analyses of theCNs was 0.85, 0.76, and 0.80%on drymatter for sisal, ramie, andcotton CNs, respectively.

In typical experiments, a given amount of CNs was added tothe water subphase in a Langmuir trough and compressed to atarget surface pressure of 15, 30, 45, or 60 mN/m. For a constant

Figure 2. Compression isotherms of Langmuir films based on DODA-cellulose nanocrystals from sisal (4), cotton (O), and ramie (0) atdifferent n(S)/n(DODA) ratios: 10 (R=125) (A), 19.3 (R=250) (B), 29 (R=375) (C), and 38.6 (R=500) (D) (see alsoTable 2).DODA inpurewater (9) is given as reference.

(38) Araki, J.;Wada,M.; Kuga, S.; Okano, T.Colloids Surf., A 1998, 142, 75–82.(39) Sugiyama, J.; Vuong, R.; Chanzy, H. Macromolecules 1991, 24(14), 4168–

4175.

(40) Elazzouzi-Hafraoui, S.; Nishiyama, Y.; Putaux, J.-L.; Heux, L.; Dubreuil,F.; Rochas, C. Biomacromolecules 2008, 9(1), 57–65.

(41) Habibi, Y.; Chanzy, H.; Vignon, M. R. Cellulose 2006, 13, 679–687.(42) Engelking, J.; Menzel, H. Eur. Phys. J. E 2001, 5(1), 87–96.

DOI: 10.1021/la902444x 995Langmuir 2010, 26(2), 990–1001

Habibi et al. Article

amount of DODA (same in all experiments), a given concentra-tion of CN was used such that the weight ratio of CNs to DODAwas 0, 125, 250, 375, or 500. According to the measured sulfurconcentrations and the amount of CNs, the actual molar ratios ofsulfate groups toDODAmolecules (n(S)/n(DODA)) ranged from10 to 43 (seeTable 2). The fact that n(S)/n(DODA) in all caseswasthe highest for sisal CNs is not surprising, since among the threefiber sources sisal CNs are known to have the highest surfacecharge density. This high surface charge is related to the relativelysmall size of the nanocrystals which produce a large surface-to-volume ratio. Nevertheless, the data in Table 2 indicate that, forthe respective concentration of CN added to the subphase, theratio of sulfate to ammonium groups were relatively similar,regardless the type of cellulose nanocrystal. In our experiments,the amount of CN used in the subphase was such that the chargeratio relative to available cationic groups (DODA) was ca. 10-,20-, 30-, and 40-fold larger than the stoichiometric 1:1 ratio(i.e., an excess of negative charges from the sulfate groups onthe CNs with respect to the cationic DODA head groups wasalways used).

We turn our attention to the resulting Langmuir-Schaeffermonolayers produced with DODA alone or with the CN-DO-DA complex of varying n(S)/n(DODA) ratios. Figure 2 shows thecorresponding surface pressure (π) isotherms as a function ofmolecular packing (molecular surface area, A) resulting frommechanical compression of theDODAmonolayer in the presenceof various concentrations of CN (π-A isotherms). The pureDODA isotherm at the air/water interface is that of a typical 2-Dliquid-expanded layer. This is due to the large osmotic pressurethat exists between the polar head groups and also due to stericrepulsion forces between the head groups in the monolayer.DODA isotherms show three main regions within differentpacking densities: (1) an expanded monolayer at high area permolecule (over 80 A/molecule or surface pressures below 20mN/m),which turns into a (2) more condensed structure at a surfacepressure of approximately 20-40 mN/m (or area per molecule of

ca. 80 A/molecule) to finally reach a (3) plateau with a collapsepressure at 42.5 mN/m (corresponding to 52 A/molecule).

In the range of surface pressures corresponding to expandedmonolayers (surface pressure below 20 mN/m), the overall π-Aprofile did not changedwith addition ofCNs.Only a small shift atlow surface pressureswas observed.Although the shoulder for thetransition noted above at 80 A/molecule remained the sameat lowCN concentrations, the presence of CNs in the subphase signi-ficantly impacted the surface pressure profile above 80 A/molecule, at all concentrations tested. The most significant effectof adding CN to the subphase was that the collapse surfacepressure was noticeably higher (60mN/m) in the presence of CNs(compared to 42.5 mN/m for the DODA in pure water). It can besuggested that the addition of CNs in the subphase favored betterpacking of DODAmolecules at the interface. In other words, thepresence of CNs in the subphase screens the electrostatic repul-sions and results in a reduction in the lateral repulsive forcesbetween DODA head groups, thereby increasing their packingdensity (or surface pressure for a given compression).

The isotherms corresponding to cotton and ramie CNs at lowconcentrations showed a shoulder around 50 A/molecule. Thisshoulder can be attributed to the presence of “free” or uncom-plexed DODAmolecules. Moreover, at high CN concentrations,the surface pressure isotherm (above 20 mN/m) was shiftedtoward higher pressures and collapse occurred at a relativelylow area per amphiphilic molecule. This was clearly shown forthe CNs with the highest charge density, that is, sisal CNs(see Tables 1 and 2). It is worth noting that since n(S)/n(DODA)was close for all of the CNs (for any given wt(CNs)/wt(DODA)),it is the aspect ratio (length-to-width) which is believed to have themost significant effect on the packing of CNs at the air/waterinterface.

The rigid, rodlike features of CNs and the larger distancebetween the ionic sites could be also responsible for the increase inthe specific area of the CN-DODA complex seen at high CNsconcentration. Interestingly, similar behaviors at the air/water

Figure 3. AFM height images and the corresponding height profiles of gold substrates before any deposition (bare gold) (A), afteralkanethiol hydrophobization (B), and after alkanethiol hydrophobization followed by DODA deposition via LS transfer (C).

996 DOI: 10.1021/la902444x Langmuir 2010, 26(2), 990–1001

Article Habibi et al.

interface to those observed here for CN-DODA systems havebeen reported for the DODA Langmuir monolayers interactingwith anionic polyelectrolytes.42,43 By using the Fromherz techni-que and Avrami model, it was reported that upon compressionthe surface pressure showed a sigmoidal-shape isotherm; a func-tion of surface area was correlated with the formation of domainswith rigid, rodlike anionic poly(p-phenylene sulfonate) (PPSf). Inthe same report, the authors used X-ray reflectivity to relate theisotherm behavior to the liquid crystalline nature of the PPSf. Onthe basis of UV/vis spectroscopymeasurements, the formation ofhighly ordered monolayer structures with increasing surfacepressure was related to the electronic interaction and to changesin the balance of intramolecular and intermolecular forces.44 Thesame observations seem to apply to the present, stiff CNs.Transfer of CN-DODAComplexes ontoHydrophobized

Gold (LS Films). One of the main goals of this work was todevelop thin films of CNs and understand their transfer from theair/water interface to a solid substrate. Also, since the films wereintended for use in additional experiments with quartz crystalmicrogravimetry, QCM gold resonators were also employed assolid support (in addition to regular goldwafers). Such resonatorstypically have an active side and an electrode side. The use of theLangmuir-Schaeffer technique with horizontal deposition wastherefore useful in order to deposit the CNs only on the active sideof the sensor. Briefly, by using the LS deposition, the respectiveCN-DODA complex was transferred from the air/water inter-face to the solid support. Such transfer took advantage ofDODA

acting as carrier of the CN layer that was formed underneath theair/water interface. More specifically, adhesion between theDODA (carrying the CNs) and the alkyl chains coating the goldsurface (from the hydrophobic SAMof alkanethiol) is expected tohave been developed from strong, long-range hydrophobic inter-actions when they were in contact. Therefore, the depositionresulted in a three layer structure on the QCM gold sensor: analkanethiol base layer, a DODA binding layer, and a CN top(active) layer.

To fully develop the LS deposition, we investigated the topo-graphy and composition of DODA films transferred to the solidsupport in the absence ofCNbyusingAFMandXPS.The transferwas conducted at the collapse plateau, and the AFM images andXPS spectra are reported in Figure 3 and in Figure S1 of theSupporting Information, respectively. The results for bare andthiolated gold QCM sensors are also reported in the SupportingInformation. AFM images showed distinctive morphologicaldifferences between bare and thiolated gold sensors before andafter DODA deposition. Thiolated sensors with and withoutadsorbed DODA molecules both showed multiple circular, nano-scale domains compared to the featureless bare gold surface. Thesedomains were denser and larger after DODA deposition(Ø=75.7 ( 3.1 nm and 127.2 ( 2.2 before and after DODAdeposition, respectively). The surfaces remained smooth eventhough the root-mean-square roughness (rms) increased from0.99 nm for bare gold to 2.13 and 3.10 nm for thiolated andDODA coated gold, respectively. The difference in AFM height(from section analysis in the respective scan) changed from 4 nm toabout 6.8 nm after DODA deposition on the thiolated gold. Thisconfirmed the incorporation of DODA on the thiol layer thatcovered the gold surface. In good agreement with these values, the

Figure 4. AFMheight images ofLS films ofCNs (cotton) transferred at different surface pressureswithn(S)/n(DODA) of 38 (orweight ratioof 500). Surface pressures were 15 mN/m (A), 30 mN/m (B), 45 mN/m (C), and 60 mN/m (D).

(43) Engelking, J.; Ulbrich, D.; Menzel, H.Macromolecules 2000, 33(24), 9026–9033.(44) Engelking, J.; Ulbrich, D.; Meyer, W. H.; Schenk-Meuser, K.; Duschner,

H.; Menzel, H. Mater. Sci. Eng., C 1999, C8-C9, 29–34.

DOI: 10.1021/la902444x 997Langmuir 2010, 26(2), 990–1001

Habibi et al. Article

ellipsometric thickness of DODA film adsorbed on thiolated goldsensors was ca. 5.17 nm. From these figures, the calculated lengthsof 1-dodecanethiol andDODA thin filmswere, respectively, ca. 2.2and 3 nm, in close agreement with reported values.8,34

The obtained adsorbed layer thicknesses, taken together withthe observed uniformity of the AFM images, indicate the success-ful transfer ofDODA from the air/water interface to the thiolatedgold substrates by using the horizontal (LS) technique. XPSanalyses on these surfaces further confirmed the chemical natureof the layers (see survey spectra in the Supporting Information).We note that the C1s XPS signal in the case of bare gold mostlikely came fromever present organic contaminants that adsorbedon the high energy gold surface from air. Most importantly, XPSspectra indicated a significant increase of the carbon content afterthiol grafting and DODA deposition: the C/Au ratio increasedfrom 1.4 for bare gold sensors to 2 and 2.5 for thiolated goldsensors before and after DODA deposition, respectively.

In order to obtain homogeneous and densely packedCN films,their transfer from the air/water interface to the hydrophobizedgold substrate was optimized by adjusting the surface pressure ofthe DODA surfactant and the concentration of CNs in thesubphase. AFM images of the films produced after LS depositionat different surface pressures and CNs concentrations are shownin Figures 4 and 5. From Figure 4A, it is clear that no CNs weretransferred when the DODA monolayer was in the expandedstate at the air/water interface (surface pressure below 20mN/m),or when the surfactant was slightly less expanded (surfacepressure below 40 mN/m, Figure 4B). This is in agreement withBrewster angle microscopy analyses of DODAmonolayers at theair/water interface reported by Habibi et al.8 who indicated theformation of a uniform film or domain structures of DODA at

the air/water interface prior to the phase transition only at highsurface pressures. Therefore, it is suggested that at low surfacepressure minimum or no transfer of CN occurred due to a limitedformation of the CN-DODA complexes. CNs were effectivelytransferred by increasing the surface pressure to values above40 mN/m: at 45 mN/m, CNs were randomly but uniformlydistributed on the substrate (Figure 4C), and when the CNs weretransferred at a pressure close to collapse (ca. 60 mN/m) a highlypacked, well-organized CN film was obtained (Figure 4D).

It is worth noting that at highest deposition pressure orderedstructures are observed (Figure 4D) corresponding to a netorientation matching the principal axes of the transferred CNs.This ordered structure can be attributed to the self-assembly of therigid rodlike CNs. Indeed, CNs can display liquid crystallinebehaviors by varying their concentration in themedium.45-47 Forexample, they can form ordered chiral nematic structures from aseries of ordered nematic assemblies. Accordingly, at a low CNconcentration or below 45 mN/m surface pressures, the CNprobably existed as a randomly organized isotropic phase. At acritical local concentration brought about after compressing theCN-DODA film to higher surface pressures, the complexesformedanordered phase (at least locally, very close to the surface)and the resultingLangmuir-Schaeffermonolayer films displayedhigh anisotropy with uniaxial orientation. The same trends wereseen for the CNs from all three sources.

Figure 5. AFM height images of LS films of cotton CNs transferred to gold substrates at different n(S)/n(DODA) or weight ratiosR. n(S)/n(DODA) used were 10 (R=125) (A), 19.3 (R=250) (B), 29 (R=375) (C), and 38.6 (R=500) (D) (see also Table 2). Surface pressure duringdeposition was maintained at 60 mN/m.

(45) Dong, X. M.; Kimura, T.; Revol, J.-F.; Gray, D. G. Langmuir 1996, 12(8),2076–82.

(46) Giasson, J.; Revol, J. F.; Ritcey, A. M.; Gray, D. G. Biopolymers 1988, 27(12), 1999–2004.

(47) Revol, J. F.; Godbout, L.; Dong, X. M.; Gray, D. G.; Chanzy, H.; Maret,G. Liq. Cryst. 1994, 16(1), 127–34.

998 DOI: 10.1021/la902444x Langmuir 2010, 26(2), 990–1001

Article Habibi et al.

The effect of CN-DODA concentration on the morphology ofthe resulting LS filmswas also thoroughly investigated usingAFMimaging of films transferred to the solid substrate at different n(S)/n(DODA) (see Table 2). The case of cotton CNs is reported inFigure 5:At lowCNconcentrations (n(S)/n(DODAof 10orweightratio of 125), a patchy filmwas obtained, as seen in Figure 5A. Thefilms became more tightly packed at n(S)/n(DODA) of 19 (weighratio of about 250). For the next higher n(S)/n(DODA) of 29(weight ratio 375), the CNs were still randomly distributed.However, at the highest concentration studied (n(S)/n(DODA)of 38.6 (weight ratio of 500), the CNs were clearly organized. TheAFM results are consistent with the changing surface pressure,creating an orderedCN-DODAstructure that was then depositedon the substrate via the Langmuir-Schaeffer deposition. Byincreasing the surface concentration or packing of CN-DODAcomplexes, a transition from a random isotropic phase to anordered anisotropic phase was produced, and thus, ordered, two-dimensional CN films were obtained.Characterization of Langmuir-Schaeffer Cellulose Nano-

crystal Films. After optimization of the transfer conditions, thetopography, thickness, and chemical composition of the obtainedLS films (for CNs from the three different sources) were analyzedvia AFM, ellipsometry, and XPS. Here, we report the case ofmaximum n(S)/n(DODA) (weight ratio of 500) and maximumsurface pressure of 60mN/m.AFM images of theLS films ofCNsare shown in Figure 6, and the respective quantitative dataobtained from AFM analyses are summarized in Table 3. From

these images, it is clear that all films obtained with the differenttypes of CNs (sisal, cotton, and ramie) were highly ordered anduniformly packed. The height difference in section scans calcu-lated fromAFM images was between 3 and 5 nm (Table 3), whichis slightly higher than that of the DODA-coated substrate. Thethickness of the LS CN films was also estimated via AFM. AFMthicknesses of ca. 6 nm for sisal, 9 nm for cotton, and 10 nm forramie were obtained (see Table 3). There was close agreementbetween the AFM thicknesses and those obtained from ellipso-metry, as also reported in Table 3.

Comparing the ellipsometric and AFM thickness values to the(TEM) lateral dimensions of the respective single CN, it can beconcluded that the transferred LS films of cellulose nanocrystalswere equivalent to a layer with a single CN layer. In contrast tovertical dipping, the main advantage of the LS horizontaldeposition is that no surfactant molecules are adsorbed on theactive side of the transferred film (they are sandwiched betweentheCNand the alkanethiol layers on the substrate). Therefore, LSfilms of cellulose nanocrystals are a convenient and suitableplatform for studies of interfacial phenomena.

XPS elemental quantification confirmed the purity of the filmsof CNs after LS transfer. Low-resolution spectra of LS filmsprepared fromCNs from the different sources showed that carbonand oxygen atoms were the main chemical components (see theSupporting Information), which is expected for pure cellulose.Small amounts of sulfur (0.6-0.8 atom%) were also detected (S2pXPS signal at 168 eV). The sulfur was assumed to be from thethiolated gold support and sulfate groups on the CNs. High-resolution carbon spectra were acquired to study the local envir-onments of the carbon atoms (C-C or C-H; C-O; O-C-OandOdC;O) (see Figure 7). In all films, the component peaks can becategorized as C-C or C-H that can be assigned to thethiol-DODA support or C-O; O-C-O structures that canassigned to cellulose structures. The XPS spectrum of DODAfilms without any CN (Figure 7A) deposited on the thiolated goldsubstrate exhibited a peak assigned to C-Cor C-H,which can beassigned to alkane structures in the thiol and DODA. Spectra

Figure 6. AFMheight images and height profiles of LS films from cellulose nanocrystals from sisal (A), ramie (B), and cotton (C) depositedon thiolated gold sensors using DODA as carrier. Surface pressure during deposition was maintained at 60 mN/m.

Table 3. Thickness and rmsRoughness Obtained fromAFMAnalyses

for Langmuir-Schaeffer Films of Different Cellulose Nanocrystals

LSfilms

roughness(nm)

thickness(nm)a

ellipsometricthickness (nm)

sisal 3.7( 0.5 6.3( 1.1 6.3 ( 0.8ramie 5.2( 1.0 9.8( 1.3 10.8( 1.0cotton 4.3( 0.7 9.5( 0.9 10.7( 1.1

aThicknesses of alkanethiol and DODA layers were subtracted fromthe total thicknesses determined by AFM.

DOI: 10.1021/la902444x 999Langmuir 2010, 26(2), 990–1001

Habibi et al. Article

obtained for CN films (Figure 7B-D) exhibited contributionsfrom each type of carbon (C-CorC-HandC-OandO-C-O).Since pure cellulose does not contain any carbon atom withoutoxygen bonding, the presence of C-C or C-H bonds can be ex-plained by the contributions from the thiol-DODA sublayer. Thisis reasonable because the average escape depth of XPS photoelec-trons, about 10nm, is similar to the thickness of theCNmonolayer.These results also provide indirect confirmation of the thickness ofthe film and the likely occurrence of a single CN layer.

After subtraction of the XPS DODA C-C or C-H contribu-tion from that of the signal from the LS CNmonolayer, the ratioof carbon in O-C-O and in C-O was 0.19 (for all CN types).This ratio is very close to 0.2, the theoretical value for purecellulose (C6O5H9)n. It can be concluded from AFM, XPS, andellipsometry analyses that pure, ordered CN monolayers wereproduced by the Langmuir-Schaeffer deposition technique.48

Swelling and Stability of LS Films of Cellulose Nano-

crystals.Tobetter understand potential uses of LS filmswithCNmonolayers as a surrogate for “native” cellulose and in order toinvestigate their interactions with polymers, surfactants, andenzymes, we tested the stability of these CN films in a series ofdifferent challenging environments. For example, treatment withaqueous sodium hydroxide (NaOH) was expected to break downelectrostatic interactions between oppositely-charged CN filmand DODA carrier molecules. In addition, NaOH was alsoexpected to neutralize and saponify sulfate groups on the cellulosenanocrystals. Therefore, QCM, which has been used as a tool for

studying themass changes in thin films, was employed tomonitorthe effects of the treatments explained above.49,50 Briefly, for a flatand uniform film that is firmly attached to a piezoelectricresonator, the change in mass of the film is directly proportional

Figure 7. XPS high-resolution C1s spectra of LS films of (A) DODA and cellulose nanocrystals from (B) sisal, (C) ramie, and (D) cottondeposited on thiolated gold sensors. C1: C-C or C-Hx, 285.0 eV; C2: C-O, 286.5 eV; C3: O-C-O, 288.0 eV.

Figure 8. Shift in QCM frequency corresponding to the normal-ized third overtone of 15 MHz of gold sensors coated withLangmuir-Schaeffermonolayersof sisal nanocrystals treatedwithsodium hydroxide solutions of 0.01M (b), 0.05M (2), and 0.1 M(9) concentration.

(48) Dorris, G. M.; Gray, D. G. Cellulose Chem. Technol. 1978, 12(1), 9–23.

(49) Rodahl, M.; Hook, F.; Fredriksson, C.; Keller, C. A.; Krozer, A.;Brzezinski, P.; Voinova, M.; Kasemo, B. Rev. Sci. Instrum. 1995, 66, 3924–3930.

(50) Rodahl, M.; Kasemo, B. Sens. Actuators, A 1996, 54, 448–456.

1000 DOI: 10.1021/la902444x Langmuir 2010, 26(2), 990–1001

Article Habibi et al.

to the change in frequency, as shown in the Sauerbrey equation:51

Δmass=-C(Δfrequency/n), where C is a mass sensitivity con-stant (= 0.177 mg m-2 Hz-1 at f=5MHz) and n is the overtonenumber.

LS films of CNs (sisal, cotton, and ramie) deposited on aQCMgold sensor were immersed in water in the QCM chamber for 1 h;thereafter, NaOH aqueous solutions (of 0.01, 0.05, or 0.1 Mconcentrations or equivalent pH in excess of 11) were injected for20 min at a rate of 0.2 mL/min. After treatment with the highestNaOH concentration, the CN films were rinsed with water torestore the original aqueous environment so that the effectivechanges could be elucidated. Shifts in resonance frequency of (LS)CN-coatedQCM sensors as a function of alkaline treatment wererecorded (see Figure 8). Only the results from sisal CNs areshown, but similar behaviors were observed for films preparedfrom CNs from the other fiber sources. During the first hour,when the films were in contact with water, the frequency wasreduced only slightly, showing a very limited degree of wateruptake and indicating the physical and chemical stability of thecrystalline CN films in aqueous medium. This is in contrast to theextensive swelling or water absorption that has been observed inthe case of amorphous films of cellulose.52

An injection of NaOH solution across the CN film produced arapid reduction in resonance frequency (negative frequency shift)indicative of the combined effects of an increased density and/orviscosity of themedium and themass uptake by the sample (in theform of larger degree of solvation, coupling water or mass

adsorption). Rinsing with water produced an equilibrium plateauat a lower level than the initial frequency values recorded inwater,before alkali treatment. This was taken as an indication of anincreased CN film mass (see effects noted before) attributable toNaOH adsorption or water uptake.

Nevertheless, these changes were observed for the LS filmsprepared with CNs from different fiber sources. It can beconcluded that the CN monolayers were relatively stable toalkaline treatments and sulfate groups on the surface of theCNs could be eliminated by alkaline treatment up to pH 11,without damaging the integrity of the films.We note that strongeralkali treatment (NaOHconcentrations above 0.5MorpHhigherthan ca. 13) caused delamination or destruction of the CN films.

To further test the robustness of the CN monolayer films,AFM topographical analysis after alkaline treatment was carriedout. Representative AFM images of ramie CNs are shown inFigure 9. The AFM results indicate that, after water and alkalinetreatments, the CNs filmswere stable and strongly attached to thesolid support. No structural changes or damage in the nanocryst-als were observed after the alkaline treatments. The lack ofswelling or disintegration is ascribed to the highly crystallinenature of the CN monolayers.

Conclusions

After investigating the phase behavior of carrier moleculesconsisting of an insoluble surfactant (cationic DODA), transferof cellulose nanocrystal-DODA complexes from the air/liquidinterface to the surface of a hydrophobic solid substrate(alkanethiol-modified gold) was successfully accomplished. TheLangmuir-Schaeffer (LS) horizontal deposition technique was

Figure 9. One micrometer scan size AFM height (left) and phase (right) images of Langmuir-Schaeffer monlayers of ramie CNs afteralkaline treatments with 0.01 M NaOH (A) and 0.1 M NaOH (B).

(51) Sauerbrey, G. Z. Phys. A: Hadrons Nucl. 1959, 155, 206–222.(52) Turon, X.; Rojas, O. J.; Deinhammer, R. S. Langmuir 2008, 24(8), 3880–

3887.

DOI: 10.1021/la902444x 1001Langmuir 2010, 26(2), 990–1001

Habibi et al. Article

found suitable for producing stable, robustmonolayers of cellulosenanocrystals from several fiber sources (sisal, ramie, and cotton).The monolayers obtained from the different cellulose nanocrystalswere proved to be chemically pure, stable, and smooth. Overall,Langmuir-Schaeffer monolayers of cellulose nanocrystals areproposed to be suitable for studying interfacial phenomena rele-vant to the chemical and biological transformations of cellulose.Alternatively, these films can be used as a coating technology tomodify the surface of other materials to attain unique properties.

Acknowledgment. The authors would like to acknowledgefunding support from the NCSU Hofmann Fellowship (IH), theNational Research Initiative of the USDA Cooperative State

Research, Education and Extension Service, Grant No. 2007-35504-18290, and FiDiPro LignoCell project (O.J.R.). Also, theassistance of Drs. K. Efimenko and J. Genzer with ellipsometrymeasurements is gratefully acknowledged.

Supporting Information Available: Figure S1: XPS widescans of gold sensor, gold sensor after hydrophobizationwith alkanethiol, and gold sensor after hydrophobizationwith alkanethiol followed by DODA deposition. Figure S2:XPS wide scans of Langmuir-Schaeffer films of cellulosenanocrystals from sisal, ramie, and cotton deposited onthiolated gold sensors. This material is available free ofcharge via the Internet at http://pubs.acs.org.