electron microscopy of nanoemulsions_an essential tool for characterisation and stability assessment

7/23/2019 Electron Microscopy of Nanoemulsions_An Essential Tool for Characterisation and Stability Assessment

1/19
http://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.micron.2011.07.014mailto:[email protected]:[email protected]://www.elsevier.com/locate/micronhttp://www.sciencedirect.com/science/journal/09684328http://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.micron.2011.07.014


2/19


3/19

V. Klang et al. / Micron 43 (2012) 85103 87

Fig. 1. (A) Visual appearance and cryo TEM image of a nanoemulsion. Images reprinted from Sonneville-Aubrun et al. (2004) with permission from Elsevier. (B) Cryo TEM

image of a nanoemulsion with nano-sized oil droplets (homogeneously filled circles) and vesicles (unfilled circles). The black scale bar represents 200nm. Image reprinted

from Norden et al. (2001) with permission from Elsevier. (C) TEM micrograph of an amphotericin B nanoemulsion with particle sizes between 100 and 400nm, prepared

by the freeze-fracturing technique. Image reprinted from Benita and Levy (1993) with permission from Wiley. (D) Visualisation of the size distribution and morphology of

nanoemulsions by atomic force microscopy, reprinted from Marxeret al. (2011) with permission from Elsevier.

(Klang et al., 2011a; Takegami et al., 2008; Yilmaz and Borchert,

2005). In addition, different methods of low-energy emulsification

have been described which employ the chemical properties of the

system to create nano-sized emulsion droplets from a microemul-

sion matrix (Anton et al., 2007; Bouchemal et al., 2004; Sole et al.,

2010; Tadros et al., 2004). In both cases, highly fluid emulsion

systems emerge. The optical appearance may range from trans-

parent systems with droplet sizes well below 100 nm to opaque ormilky systems with droplet sizes slightly above 100 nm. A bluish

tone points to the occurrence of Rayleigh Scattering for small

droplet sizes. Emulsions with larger droplet sizes appear as plain

white fluids due to multiple scattering of light (Klang and Valenta,

2011c; Mason et al., 2006). Representative examples of a translu-

cent nanoemulsion and a corresponding cryo TEM image are given

in Fig. 1A.

The characterisation of nano-sized colloidal systems requires

diverse techniques and a certain experience. Basic formulation

characteristics such as visual appearance, pH, mean particle size,

particle surface charge, chemical stability of employed excipients

and the localisation of incorporated drugs provide useful infor-

mation (Klang and Valenta, 2011c). For an exhaustive overview of

the different physicochemical techniques for nanoemulsion char-acterisation the reader is referred to recent reviews on this topic

(Benita and Levy, 1993; Klang and Valenta, 2011c; Solans et al.,

2005; Tadros et al., 2004).

Especially the mean particle size and the particle size dis-

tribution are frequently employed to characterise the long-term

stability of novel formulations. These parameters can be deter-

mined by optical light scattering techniques. More specifically,

dynamic light scattering (DLS, photon correlation spectroscopy) is

frequently employed for determination of nano-sized oil droplets

within emulsions. Among the obtained results, the mean particle

size as intensity-weighted mean of the hydrodynamic diameter

and the polydispersity index (PDI) are most frequently presented

(Hoeller et al., 2009; Klang et al., 2011a; Preetz et al., 2010; Yilmaz

and Borchert, 2005). The latter characterises the width of the

particle size distribution and thus the homogeneity of the for-

mulation. A small PDI below 0.2 indicates a narrow droplet size

distribution and thus better stability against destabilisation phe-

nomena such as Ostwald ripening (Klang and Valenta, 2011c).

The characterisation and stability assessment of nanoemulsions is

strongly associated with their droplet size and PDI. If both param-

eters remain largely unchanged during a prolonged observation

period, a formulation is usually considered physically stable. Ini-tial and regular subsequent measurements are required for an

exact stability assessment. Apart from light scattering techniques,

microscopic methods can be employed to monitor changes in

droplet size. Thus, electron microscopy is not only a valuable

tool for formulation characterisation, but also for the stability

assessment of nanoemulsions since certain changes in formula-

tion properties remain undetected during DLS analysis, as detailed

below.

2.2. Laser light scattering versus electron microscopy

In context with nanoemulsion characterisation and stability

assessment, DLS exhibits certain limitations and may provide

incomplete information. Firstly, it may fail to recognise thepresence of a small population of large droplets present in

nanoemulsions (Benita and Levy, 1993). Likewise, other surfac-

tant aggregates such as liposomal vesicles or lamellar structures

are not detected; the exact composition of the colloidal system

thus remains unknown. However, such structures are frequent

by-products of high-pressure homogenisation and should be

accounted for (Klang et al.,2011a). Fig.1B demonstratesthe compo-

sition of sucha mixedcolloidal dispersionof nano-sized oil droplets

and liposomes. Moreover, the shape of the analysed oil droplets is

usually assumed to be a perfect sphere for calculation of the DLS

results, which is not always the case. Thus, determined particle

sizes for droplets of variable shape may not be entirely repre-

sentative. Furthermore, most samples have to be diluted prior to

DLS measurements to ensure sufficient transparency for accurate


4/19

88 V. Klang et al. / Micron 43 (2012) 85103

droplet size determination (Klang and Valenta, 2011c). As a conse-

quence, reversible destabilisation phenomena such as flocculation

or the appearance of larger aggregates may remain unnoticed.

In order to account for these issues, additional techniques of

analysis are highly recommendable. Methods such as sedimen-

tation field flow fractionation (Benita and Levy, 1993), nuclear

magnetic resonance spectroscopy (Norden et al., 2001; Takegami

et al., 2008) or Fourier transform infrared spectroscopy (Scheuing,

1990; Whittinghill et al., 1999) have been proposed in this con-

text. However, the microscopic visualisation of the investigated

nanoemulsions might represent the most reliable and informative

method for formulation characterisation.

When employing microscopic techniques for nanoemulsion

characterisation, the presence of larger droplets is not an entirely

uncommon observation (Hatanaka et al., 2010; Preetz et al., 2010),

albeit a rarely reported one. Experience has shown that it is possi-

ble to obtain excellent DLS data for nanoemulsions over months of

stability monitoring while a microscopic analysis of the same sam-

ple reveals a definite change of the internal structure. Recently,

Preetz et al. (2010) demonstrated the importance of microscopic

analysis for the characterisation of nanoemulsions and nanocap-

sules. It was found that the mean droplet size determined by

DLS was around 150 nm for all investigated systems. In contrast,

freeze-fracture TEM revealed variable droplet sizes between 50and 500nm with the highest frequency around100 nm, which was

additionally confirmed by atomic force microscopy (Preetz et al.,

2010).

Thus, the importance of electron microscopic techniques for

the analysis of nanoemulsion droplet size and overall morphology

needs to be emphasized. Surprisingly, hardly any review articles

can be found on this specific topic. An excellent overview of cryo

TEM analysis of colloidalsystems in general was recently published

(Kuntscheet al.,2011). Cryo TEMis certainlyamong the most useful

techniques for the investigation of nanoemulsions since it delivers

detailed information about the internal structure of the observed

colloidal systems in their native state. The present review aims to

give an overview of the electron microscopy techniques that have

been employed for the analysis of nanoemulsions so far and possi-ble future developments. It is important to note that all mentioned

techniques have theirbenefits anddrawbacks. In anycase, thestud-

ied images have to be representative of the whole sample. Image

analysis software shouldbe employed only for systems with a suit-

ablecontrast andcomposition.Several rounds of analysis are highly

recommendable. A good overview of the investigated systems can

be obtained by combination of DLS or static laser diffraction and

cryo TEM (Kuntsche et al., 2009).

2.3. Development of microscopic techniques for visualisation of

nanoemulsions

Optical light microscopy can be considered the simplest method

to examine the microstructure of nanoemulsions. However, it wassoon realised that this technique is futile for the exact visualisa-

tion of nano-sized droplets below 500 nm (Benita and Levy, 1993;

Drzymala and Krajczyk, 1985). In context with nanoemulsions, the

applications of light microscopy are limited to the detection of

pronounced destabilisation phenomena such as coalescence and

Ostwald ripening (Jumaa and Mueller, 1998; Welin-Berger and

Bergenstahl, 2000) or monitoring of phase transitions (Jahanzad

et al., 2010; Rao and McClements, 2010). Likewise, the presence of

larger aggregates, droplets (Baker and Naguib, 2005; Burapapadh

et al., 2010; Corts-Munoz et al., 2009; Wang et al., 2008) or

undissolved drug crystals (Akkar and Mueller, 2003; Araujo et al.,

2011) can be determined and the background movement within

the images indicates the presence of nano-sized droplets in Brow-

nian motion (Schalbart et al., 2010). For an exact visualisation of

the nanoemulsion structure and determination of the particle size

a higher resolution is necessary. Thus, electron microscopy is an

essential tool for these tasks.

Early methods for visualisation of parenteral fat emulsionswere

presented by Du Plessis et al. (1986) who analysed thin sections of

emulsions enclosed in agar capsules after fixation with osmium

tetroxide. Well-defined droplets with particle sizes around 250nm

could be visualised. This methodproved to be time-consuming and

expensive; first results could only be obtained after 30h. Faster

and less expensive methods were subsequently developed which

involved fixation of the diluted sample and placing a thin layer on

a formvar-coated copper grid for TEM examination after nega-

tive staining (Benita and Levy, 1993; Du Plessis et al., 1987). As

for all hydrated systems investigated by conventional electron

microscopy, artefacts induced by fixation, inclusion in agar, nega-

tive staining and drying have to be expected. Structural alterations

such as distortion or collapse of the colloidal system caused by

investigation under vacuum and beam damage always need to be

taken into consideration wheninterpreting such images (Kuntsche

et al., 2011).

At roughly the same time, it wasrealised that SEM likewise rep-

resents a valuable tool to determine the droplet size of parenteral

fat emulsions (Benita and Levy, 1993; Bullock, 1984; Hamilton-

Attwell et al., 1987). A definite benefit of the SEM technique isthe topography dependentnature of the obtained images, i.e. addi-

tional topographical information at a considerable depth of focus

on a 2D-image can be obtained. However, the fixation of lipids is

a challenging task and specific fixation protocols for SEM are nec-

essary that preserve the shape and size of nanoemulsion droplets

(Benita and Levy, 1993). One of the earliest techniques in this

respect was developed by Hamilton et al. which combines fixation

of a nanoemulsionwith osmium trioxideand deposition of thefixed

droplets on a filter support (Benita et al., 1991; Hamilton-Attwell

et al., 1987).

In addition, freeze-fracturing techniques (Bachmann and

Schmitt-Fumian, 1973) emerged in context with electron

microscopy of nanoemulsions (Benita and Levy, 1993; Menold

et al., 1972; Sjoeblom and Friberg, 1978), thus increasing accuracyof the obtained images. The technique of rapid freezing and

fracturing of the sample to prepare carbon replicas was found

time-consuming since it involved a two-stage Triafol-carbon

replica technique and required the use of a special vacuum coater

(Drzymala and Krajczyk, 1985). However, it allowed for the

identification of finely structured molecules such as lipids and

various formations composed of surfactant. Early research already

produced images of high quality (Fig. 1C) (Benita and Levy, 1993).

Thus, the detection of vesicles, micelles, liquid crystals and other

structures together with fat droplets became feasible (Buchheim,

1982; Groves et al., 1985; Norden et al., 2001; Rotenberg et al.,

1991b). It was mainly by means of freeze-fracturing and cryofix-

ation techniques (Groves et al., 1985; Rotenberg et al., 1991b)

that the well-established presence of liposomes in nanoemulsionswas first observed. The development of cryo-preparation methods

for liquid specimens (Bachmann and Talmon, 1984) offered new

possibilities for nanoemulsion characterisation by cryo-TEM

(Rotenberget al., 1991a; Teixeira et al., 2000) andcryo-SEM (Saupe

et al., 2006).

Other microscopic techniques such as atomic force microscopy

(AFM) have been successfully employed for the investigation of

nanoemulsions (Fig. 1D) (Corts-Munoz et al., 2009; Fang et al.,

2004; Marxer et al., 2011; Preetz et al., 2010; T akegami et al.,

2008). Since the focus of this review lies with electron microscopy,

AFM will not be further discussed in this context. The electron

microscopic techniques that are most frequently employed for the

analysis of nanoemulsions today as well as recent methodological

advances will be elucidated in the following chapters.


5/19


3. Scanning electronmicroscopy

3.1. Experimental setup, sample preparation and potential

artefacts

InSEM(Ardenne,1938a,b; Knoll, 1935; Reimer, 1993), theimage

is formed step by step by scanning a focused electron beam across

the specimen. The primary electrons penetrate the solid specimen

and are deflected by a large number of elastic scattering processes.

Various signals are generated as a result of the impact of the inci-

dent electrons, which are collected to form an image or to analyse

the sample surface (Bogner et al., 2007). The electrons that are col-

lected by the detector system give specific information and types

of contrast as detailed below:

I. The surface topography of the sample is primarily registered by

secondary electrons, i.e. all emitted electrons with exit energies

below 50eV. Secondary electrons can emerge from the speci-

men only from within a thin surface layer of a few nanometers.

The image contrastdepends on theselectedangular range of the

electrons collected (Goldstein et al., 1981).

II. Material contrast can be obtained by back-scattered electrons

which possess energies between 50eV and the primary energy

at the point when they pass through the surface of the specimen.This contrast results in an increase in intensity with increasing

mean atomic number (Reimer, 1993).

The SEM technique has a few important characteristics which

make it a highly popular microscopic technique for the ultra-

structural investigation of different kinds of hydrated materials,

including nanoemulsions. Firstly, topography dependent informa-

tion of the sample surface can be obtained in this fashion, as

opposed to the two-dimensional projections of the section volume

which can be obtained by TEM. Secondly, a great depth of focus is

achieved by SEM. At low magnifications it can be in the range of

a few millimeters which is especially important for samples with

high surface corrugation. However, there are certain limitations.

These include a lack of internal details, a somewhat limited resolu-tion and the risk of electron beam damage (Reimer, 1993).

3.2. SEM-based techniques for characterisation of nanoemulsions

Calderillaet al. investigated differentcolloidal systems basedon

sucrose esters. While SEM was employed for the characterisation

of nanocapsules by a fixation procedure with osmium tetroxide

and covering of the dried samples with gold particles, no such

approach was reported for the nanoemulsions presented in this

study (Calderilla-Fajardo et al., 2006). Hatziantoniou et al. (2007)

investigated both nanoemulsions and solid lipid nanoparticles by

SEM. The samples were placed on a polycarbon substrate and

left to dry at room temperature, followed by drying in a criti-

cal point drier. The samples were then sputter coated with goldand examined. The authors stated that SEM was a suitable tech-

nique for the investigationof nanoemulsions and advantages when

compared to FF-TEM were found, such as lower costs and the

opportunity to examine large areas of dispersed particles in a short

time.

Another variation of a SEM analysis for visualisation of propo-

fol nanoemulsions was performed by Masaki et al. (2003) to

analyse notable changes in droplet size. A one-step fixation tech-

nique for the oil droplets on filter papers with a combination

of glutaraldehyde-malachite green and osmium tetroxide was

employed. These compounds are otherwise used to stabilise i.v.

fat emulsions. Malachite green is useful as a dye and at the same

time stabilises lipiddissolvedin aqueous glutaraldehyde. Thus,this

specific fixation technique permitted visualisation of flawless oil

droplets in the SEM. A filter paper immersed in the saline dilution

medium alone was used as control. The samples were dehydrated

bydifferent alcohols and after 2 h of freeze-drying they were coated

with gold vapour and observedundera SEMto detectthemaximum

size of the observed droplets. The resulting images clearly showed

the coalescence process of the emulsion droplets with propofol

after addition of lidocaine. Not surprisingly, the authors reported a

certain discrepancy between the observed maximum droplet sizes

andthose previously determined by DLS. As already discussed, DLS

only determines an average diameter of all droplets and thus may

fail to acknowledge the presence of a few large droplets among

a large population of small droplets. However, the presence of

larger droplets may be detrimental to thelong-term stabilityof the

nanoemulsion and may render it unsuitable for intravenous appli-

cation. This underlines the need for additional methods of analysis

such as electron microscopic techniques.

A SEM-based technique that can be used for dynamic experi-

ments of hydratednanocarriersis environmental scanningelectron

microscopy (ESEM) (do Amaral et al., 2005; Ruozi et al., 2011). This

technique is based on the use of a multiple aperture and a gradu-

ated vacuum system that allows the chamber to be maintained at

pressures around 5000Pa. Specimens can be viewed under water

vapour or other auxiliary gases. Moreover, dehydration of hydrated

samples can be partially inhibited by using a pump-down pro-cedure and by controlling the temperature using a Peltier stage

(Ruozi et al., 2011). However, the resolution is not sufficient to

obtain detailed information regarding the surface properties and

architecture of nanoscale structures (Mohammed et al., 2004). This

lack of resolution is likewise the current main limitation for other

SEM-based techniques such as wet scanning transmission elec-

tron microscopy (wet STEM in ESEM) or SEM of fluids in flow cells

(Bogneret al., 2005,2007). Nevertheless, accurate results regarding

droplet size and size distribution of nano-sized emulsions can be

obtained bywet STEM inESEM (Bogneret al., 2005; do Amaral et al.,

2005). Thus, further developments in this field may be anticipated

with interest.

4. Transmission electronmicroscopy

4.1. Experimental setup, sample preparation and potential

artefacts

Conventional transmission electron microscopes (Ruska and

Knoll, 1932) are electron optical instruments analogous to light

microscopes.However,the specimenis notilluminated by light, but

by an electron beam. This requires operation in a vacuum since gas

moleculeswoulddeflect theelectrons. At the topof the microscope

column is an electron gun and a system of electromagnetic lenses

focuses theelectron beam onthe sample. Theimage contrastin TEM

is obtained by the interactions of the electrons with the material,

i.e. electron scattering. The resolution in TEM is directly propor-tional to the acceleration voltage of the electrons. High resolution

is obtained due to the short wavelength of the electrons when the

voltage increases. However, increasing acceleration voltage leads

to poorer contrast since the scattering of the electrons is decreased

at higher velocity (Kuntsche et al., 2011). Typical instruments are

capable of voltages from 40 to 120kV andmicroscopes in the range

of 200400kV are becoming more common. In TEM investigations

of colloidal systems, voltages between 80 and 200 kV are usually

employed (Kuntsche et al., 2011).

The resolution in TEM is rather limited by the properties of the

specimens than by those of instrumentation. Its present level is at

0.31.0nm (Massover, 2011). The absorption of electrons into the

specimen is unusual, but electrons scattered to large angles do not

contribute to the image in the usual bright field mode and thus


6/19


appear to be absorbed. In the case of ordered or crystalline sam-

ple material, this results in diffraction contrast, which is strongly

dependent on the crystal orientation. In amorphous materials,

a mass thickness contrast is formed, where the image bright-

ness depends on the local mass thickness. Thus, darker regions

in the bright field image mode are regions of higher scattering

(Sawyer and Grubb, 1996). For more detailed information on elec-

tron microscopy, general literature is recommended (Bozzola and

Russell, 1999; Egerton, 2005).

In case of nanoemulsions, cryo preparation methods are of

course the most suitable methods of analysis. If cryo TEM is not

available, a conventional negative staining analysis with or with-

out dilution can be performed on nanoemulsions to obtain certain

basic informations. Staining techniques are frequently employed

for imaging of colloidal systems with TEM since they are easy, fast

and universally applicable (Kuntsche et al., 2011; Massover, 2008).

Themost commonstaining agents are salts of heavy metals such as

molybdenum, tungsten or uranium which possess atomic numbers

between42 and 92. These agents must be benign to the wet speci-

mens, form a thin glassy layer upon dryingand must resist electron

beam radiation damage to a satisfying extent (Massover, 2008).

In context with nanoemulsion analysis, phosphotungstic acid or

its salt solutions (Desai et al., 2008; Ganta and Amiji, 2009; Liu

and Yu, 2010; Pathan and Setty, 2011; Seki et al., 2004) or uranylacetate (Araujo et al., 2011; Hatanaka et al., 2010; Schalbart et al.,

2010) are most frequently employed. During sample preparation,

a droplet of the nanoemulsion is placed on a carbon coated grid

onto which it is rapidly adsorbed. Subsequently, an aqueous solu-

tion of a heavy metal salt is applied for staining. The sample is

then left to dry and finally observed by TEM at room tempera-

ture. This technique allows for the identification of the dehydrated

shells of the nanoemulsion droplets which are stabilised by sur-

factant. The strongly scattering metal ions form an amorphous

shieldenveloping the weakly scattering oil droplets to enhance the

electron microscopy contrast (Brenner and Horne, 1959). A high

reverse contrast is thus seen in bright field TEM images, with light

dropletsagainsta darker background.Negative stainingcan be used

to visualise the size, shape and internal structure of the sample.The negative stain not only provides contrast for weakly scattering

specimens, butalso physical support against collapse of the sample

structure during drying and protection against electron beam dam-

age (Massover, 2008). Thesample may also be coatedwith a carbon

film under ambient conditions before investigation (Burapapadh

et al., 2010).

However, it should be kept in mind that conventional electron

microscopy is prone to artefacts in case of surfactant solutions, i.e.

hydrated colloidal dispersions (Egelhaaf et al., 2000). For hydrated

samples such as nanoemulsions, the factors affecting the preserva-

tion of the structural integrity are identical in both TEM and SEM

(Mueller, 1991). Both drying and staining techniques canaffect the

structure and morphology of the sample; thus, great care should

be taken during interpretation of the obtained images (Friedrichet al., 2010). Severe shrinkage or even complete collapse, selec-

tive dimensional modification and aggregation of the constituents

of colloidal systems due to complete dehydration and drying usu-

ally cause strong structural changes. In addition, the use of heavy

metal salts leads to a selectivesample appearance since only struc-

tures that can be reached by or react with the staining agents can

be detected (Hayat, 2000). Consequently, the uppermost surface

of the sample as well as the compounds of the system that can

chemically reactwith thestainingagentsare clearly observedwhile

many compounds of the sample remain practically invisible. As a

result,thefinalEMimagesmay describecompletelymodifiedstruc-

tureswhich have nothing in common withthe original formulation

morphology. Moreover, conventional negative staining on contin-

uous carbon support films bears the risk of sample distortion due

to adsorption and flattening during the drying of the thin aque-

ous film of negative stain or evaporation in the TEM (Harris, 2008).

Apartfrom adsorptionartefacts, variable spreadingand incomplete

specimen coverage by the stain solution can lead to non-uniform

staining results. Specimen distortion from surface tension forces

duringevaporative dryingor the formationof a saturated salt solu-

tion before the final drying may occur as well (Massover, 2008).

Techniques basedon cryofixation and low temperature electron

microscopy help to overcome these major problems. However, not

all TEM laboratories are equipped with cryo electron microscopy

facilities. Therefore, the application of air-dried negative staining

techniques for biological specimens or other aqueous systems such

as colloidal dispersions remains justified (Harris, 2008).

4.2. Conventional TEM for characterisation of nanoemulsions

Conventional TEM is frequently employed for the analysis of

colloidal systems since it can visualise coexisting structures and

microstructure transitions (Bali et al., 2010; Silva et al., 2009).

Despite the obvious risk of artefacts, a large body of experi-

mental data can be found in the respective literature. Image 2

gives an overview of images obtained by TEM at room tempera-

ture after negative staining. Although these images might providesome basic information in regard to droplet stability or general

size range, they are of course far from a successful microscopic

analysis. Conventional TEM may have its place as a useful pre-

liminary technique to obtain a rapid impression of a colloidal

system. However, this approach should not be confused with

a substantiated structural analysis. We present some images of

variable quality to show the potential as well as the obvious lim-

itations of this technique. Seki et al. (2004) employed TEM for

visualisation of two different emulsion systems. The samples were

placed on a collodion-coated specimen mesh and negative staining

with sodium phosphotungstate was performed before observa-

tion. As can be seen in Fig. 2A, a pronounced difference in mean

droplet diameters was demonstrated for the two systems. Like-

wise, Kelmann et al. (2007) investigated the microstructure ofa carbamazepine-loaded nanoemulsion by TEM. A droplet of the

sample was placed on a carbon-coated copper grid, stained with

uranyl acetate and covered with formvar. As demonstrated in

Fig. 2B, a clear image of the undiluted nanoemulsion was obtained.

A similar staining technique using uranyl acetate was employed for

visualisation of a nanoemulsion for the administration of paclitaxel

and curcumin(Ganta and Amiji, 2009). Likewise, topical nanoemul-

sions with genistein were visualised in this fashion (Silva et al.,

2009). The increased contrast observed at the interface of the oil

droplets was assumed to be related to an affinity of the staining

agent to interfacial components.

In another approach, nanoemulsions stabilised by semisolid

polymer interphases were analysed by TEM without specified sam-

ple preparation (Namet al., 2010). Significant variations werefoundin size distribution and morphology of nanoemulsions in depen-

denceof differentoilphases (Fig.2C). The surface morphology of the

systems might havebeen affected by thedrying process. Gantaetal.

(2008) employedTEM afternegative stainingwithphosphotungstic

acid. The peculiar arrangement of the nano-sized droplets (Fig. 2D)

was most likely caused by the sample preparation. A similar tech-

nique was employed for visualisation of gelated nanoemulsions

(Mou et al., 2008); however, the contrast of the obtained images

was poorer, possibly due to the altered formulation properties. In

another work, the same negative staining method was employed

for ocular submicronemulsions(Ibrahim et al., 2009). The obtained

images revealed a crowded droplet structure and poor contrast

for emulsions with castor oil (Fig. 2E). Coalescence phenomena

were observedfor emulsionswith soybean oil(Fig.2F). It remained


7/19


Fig. 2. (A) Electron micrographs obtained after negative staining of a lipid nanoemulsion system (lipid nano-spheres, LNS,left-hand side) and a conventional fat emulsion

for parenteral nutrition (LM, right-hand side). Images reprinted from Seki etal. (2004) with permission from Elsevier. (B) TEM photomicrograph of a carbamazepine-loaded

nanoemulsion. Image reprinted from Kelmann et al. (2007) with permission from Elsevier. (C) TEM micrographs of nanoemulsions prepared with different types of oil:

(1) phenyl trimethicone, (2) poly-dimethylsiloxane, (3) cetyl ethylhexanoate, (4) dioctanoyl-decanoyl-glycerol, (5) isopropyl myristate and (6) liquid paraffin. Scale bars

represent 150nm. Images reprinted from Nam etal. (2010) with permission from theAmerican Chemical Society (ACS).(D) TEMphotomicrograph of a parenteral emulsion

after negative staining using phosphotungstic acid. The scale bar represents 500nm. Image reprinted from Ganta et al. (2008) with permission from Elsevier. (E), (F) TEM

photomicrographs of nano-sized emulsionswith castoroil (E)or soybean oil (F). Thescale bars represent500 nm. For the latter system, coalescence phenomena are visible.

Images reprintedfrom Ibrahim et al.(2009) with permission fromElsevier.(G) TEM photomicrographsof thalidomid-loaded and blanknanoemulsionswith polysorbate. The

scale bars represent100 nm.

Image reprinted from Araujo et al. (2011) with permission from Elsevier.

unclear whether the latter was inherent to the formulation ormerely an artefact caused by sample preparation.

Several authors employed specific diffraction modes during

conventional TEM bright field imaging to visualise the form and

size of nanoemulsions and to determine the amorphous or crys-

talline character of their components (Bouchemal et al., 2004;

Shafiq et al., 2007). However, the contrast of the obtained imagesof

dark droplets against a bright background was comparatively poor

in these as well as in similar cases (Shakeel et al., 2007; Singh and

Vingkar, 2008).

Overall, nanoemulsion images of good quality obtained by con-

ventional TEM are comparatively rare. An appropriate staining

technique appears to be essential to obtain sufficient contrast;

both type and amount of the staining solution should be optimised

before the respective task. Nevertheless, oil droplet aggregation

or destruction has to be expected under the dehydration condi-tions of electron microscopy (Liu and Yu, 2010). Even if images

of more or less intact oil droplets are obtained, the information

that is conveyed is frequently limited since a very restricted area of

the investigated specimen may remainintact and canbe presented

(Fig. 2G). In this case, no information is obtained about the polydis-

persity of the droplet size distributionor the presence of additional

structures within the sample.

Apparently, conventional TEM may be most successfully

employed for the real-space determination of nanoemulsion

droplets after continuous phase-evaporation if the employed oil

has a large molecular weight, as observed by Mason et al. (2006). A

silicon oil-in-water nanoemulsion was stained with uranyl acetate

and placed on a TEM grid coated with a monolayer polymer

film. After evaporation of the water the sample was observed at


8/19


Fig. 3. Comparison of cryo TEMand conventional TEM after negative staining with uranyl acetate: nanoemulsions stabilised by either5% (A and B) or 2.5% (w/w) of sucrose

stearate (C andD) were investigated with both methods. On theleft hand side (A,C), thecryo TEMimages aregiven.On therighthand side (B,D), thecorresponding images

obtained by conventional TEM at room temperature are given. For conventional TEM, the risk of beam damage is imminent as indicated in the lower right hand corner of

image D.

Image A and B are reprinted from Klang et al. (2011b), image C is reprinted from Klang et al. (2011a) with permission from Elsevier.

room temperature. At low oil volume fractions, coalescence of the

droplets during evaporation was apparently avoided since they

did not completely cover the grid. This effect was most likely

enhanced by the staining agent. However, even nanoemulsion

dropletsthat hadbeen concentratedand deformedduringthe evap-

oration process could be observed. Apparently, the silicone oil was

of sufficiently large molecular weight so as not to significantly

evaporate under the experimental conditions. The specific reasonswhy this simple method provided such comparatively good results

remained unclear. In recent experiments (Klang et al., 2011a,b),

conventional nano-sized oil-in-water emulsions were investigated

by TEM at room temperature after negative staining with uranyl

acetate (Fig.3B andD).A directcomparison with imagesof thesame

samples obtainedby cryo TEM(Fig.3A and C)suggeststhatthemor-

phology of thesystems is comparatively well preserveddespite the

unfavourable surroundings within the TEM. Surprisingly intact, if

deformed and aggregated droplet shells were obtained. Neverthe-

less, care has to be taken to avoid artefactscaused by electronbeam

damage (Fig. 3D, right-hand corner), which are more imminent in

TEM at room temperature.

Summarising our experiences with TEM analysis of nanoemul-

sions at room temperature, it may be assumed that it is not

primarily the nature of the oil which is decisive for the quality of

the obtained images, but the efficacy of the employed surfactant

to stabilise the oil droplets. In Fig. 3, a conventional cosmetic oil

of a molecular weight around 300 g/mol was emulsified using dif-

ferent amounts of a sucrose ester surfactant. Perfectly clear phase

boundaries of the droplet shells after evaporation in the TEM were

foundfor systems of an idealsurfactantconcentration of 2.5%(w/w)

and high physical stability (Fig. 3D). A surplus of surfactant did notimprove the formulations general properties (Fig. 3B), but rather

destabilised the system by introducing aggregates and leading to

droplet deformation.The obtainedimagesof bothformulations cor-

responded comparatively well to the native structures as observed

by cryo TEM. Although the obtained TEM images are of course far

from describing the real state of the emulsion, they provide cer-

tain information that might be useful in nanoemulsion studies, e.g.

about sample stability against external stress. The stability of the

emulsified oil droplets is after all an important quality parameter.

However, it is again emphasised that the potential of TEM at room

temperature is limited to such basic aspects; an exact visualisation

of the droplet shape remains confined to cryo TEM. Interpretation

of TEM images of nanoemulsions should therefore be performed

with caution.


9/19


5. Cryo-preparation techniques for SEM

5.1. FF-SEMand cryo SEM: experimental setup, sample

preparation and potential artefacts

Freeze-fracture techniques and direct imaging by cryo electron

microscopy are clearly more appropriate methods for analysis of

nanoemulsions than conventional electron microscopy. In freeze-

fracture electron microscopy the specimens are rapidly frozen,

fractured, shadowed with metal and replicated (Belkoura et al.,

2004). The metal replica of the fracture surface are then viewed

either by SEM or TEM, the former of which may include additional

steps sample preparation. Further details on the freeze-fracture

technique for analysis of nanoemulsions are given in context

with freeze-fracture TEM in section 6.1., which is more frequently

employed in this field.

In contrast to freeze-fracture electron microscopy, the frozen

specimens prepared for cryo electron microscopy are not repli-

cated, but immediately transferred to a low temperature stage

within the microscope and viewed directly (Belkoura et al., 2004).

In cryo SEM, additional steps such as coating with metal for

enhanced conductance of the electrons are usually performed

(Preetz et al., 2010). Details on cryo preparation methods are given

in context with cryo TEM in Section 6.3, which is by far the mostfrequently employed technique for nanoemulsion analysis.

Images of nanoemulsions in their natural hydrated state can be

obtained by these techniques of sample preparation before SEM

analysis. The benefits of cryo SEM are particularly useful for the

analysis of certain types of colloidal nanoemulsion systems which

are not entirely suitable for the commonly employed cryo TEM

investigation. For samples which are highly viscous, have a strong

tendency to aggregation or contain large amounts of oil, cryo SEM

of the freeze fracture-freeze dried samples is among the best solu-

tions (Severs and Robenek, 2008). On the one hand, an ultrathin

layer of such a solution for cryo TEM investigation can hardly be

obtained. Viscous nanoemulsions are adsorbed firmly onto the car-

bon coated grid and are not readily removed by filter paper. The

resulting layer may not be thin enough to be transparent for theelectron beam. On the other hand, the dimensions of the holes in

the carbon support film are usually much smaller than the size

of large aggregates or multilamellar structures potentially present

in nanoemulsions. Thus, such structures may remain undetected.

In cryo TEM micrographs only a small part of an entire aggregate

structure can be observed, which may be insufficient for a com-

prehensive understanding of the sample organisation. In contrast,

cryo SEM delivers information about the distribution of specific

structures such as aggregates on the surface of the sampleirrespec-

tive of their location on the grid. Thus, cryo SEM provides a more

adequate impression of the overall morphology of inhomogeneous

and/or viscous nanoemulsions. Likewise, comparatively large sam-

ple areas can be investigated by freeze-fracture SEM, which also

delivers a good overview of the sample morphology and homo-geneity.

When working with either freeze-fracture or cryo electron

microscopy, some general aspects on the benefits and limitations

of the different methods of sample preparation should be con-

sidered to identify the most appropriate technique for a specific

nanoemulsion sample. Vitrification of samples for TEM is typi-

cally achieved by plunging the thin liquid specimens into liquid

ethane. In case of oil-rich samples, apolarcompounds such as hydr-

carbons, alkales or aromatics will dissolve in the cryo-medium

ethane, thus rendering analysis of the intact system impossible

(Kesselman et al., 2005). Since oil-in-water nanoemulsions usu-

ally do notcontainlargeamounts of apolarcompounds or solvents,

plunge-freezing is usually a suitable approach. However, in case

of water-in-oil nanoemulsions or microemulsions dissolution of

sample compounds may occur. Therefore, high pressure-freezing

represents the optimal cryofixation methodin such cases. It allows

the vitrification of samples with a thickness of up to 200m.Moreover, during high pressure freezing the sample is encapsu-

lated in a metal shell, which helps to avoid dissolution problems.

Conventional freeze-fracture techniques allow for the direct visu-

alisation of oil-rich samples which would dissolve in the liquid

ethane used for vitrification of the samples during cryo-techniques

(Belkouraet al.,2004). In addition, non-blottingmethods for freeze-

fracture preparation andcryo-energyfilteredtransmissionelectron

microscopy are suitable alternatives (Belkoura et al., 2004; Burauer

et al., 2003; Schmidtgen et al., 1998).

5.2. FF-SEMand cryo SEM for characterisation of nanoemulsions

Cryogenic treatment of samples prior to SEM allows for the

investigation of soft carriers such as nanoemulsions in their native

state. Saupe et al. investigated different nano-structured lipid car-

riers such as nanoemulsions and solid lipid nanoparticles by cryo

SEM (Saupe et al., 2006). The liquid colloidal systems were plunge-

frozenand viewedin theirnatural state after fracturing andcoating

withplatinum. Deviating nano-sized structures which may emerge

during production could be detected by this technique since bothshape and size of the lipid structures were visualised (Fig. 4A).

The technique of freeze-fracture SEM is among the more rarely

employed techniques for the analysis of nanoemulsions (Mason

et al., 2006). Few reports can be found in the literature (Bilbao-

Sainz et al., 2010). A freeze-fracture SEM technique was employed

to characterise the interfacial and bulk structure of silica-coated

nano-sized emulsionsas wellas conventionalnanoemulsions with-

out additional coating (Eskandaret al. 2009). The coated emulsions

were stabilisedby mixedinterfacial layers of either lecithin or oley-

lamine and hydrophilic silica nanoparticles. The reported method

includedcryofixation of the emulsion,fracturing, etching, platinum

coating and subsequent imaging. The samples were injected into a

split brass tube and were cryofixed by plunging the latter into a

liquid nitrogen-solid nitrogen slush together with the cryo trans-fer specimen holder. The specimen was then transferred to the

exchange chamber under vacuumand the split brass tube was bro-

ken with a precooled scalpel. The surface ice was removed during

careful sublimation to avoid droplet disintegration. The fractured

and etched sample was then coated with platinum to provide con-

ductivity prior to SEM imaging. The obtained images revealed that

the adsorption of the silica nanoparticles to the oil droplet inter-

face was influenced by the nature of the surfactant and the charge

it conveyed to the interfacial film (Fig. 4B). The distribution of the

silicananoparticles between the interfacial filmand the bulk phase

could be determined in this fashion.

6. Cryo-preparation techniques for TEM

6.1. Freeze-etching and freeze-fracturing for TEM: experimental

setup, sample preparation andpotential artefacts

Freeze-etching, freeze-fracturing and cryo electron microscopy

of frozen fluids are complementary techniques mainly because the

respectively obtained information is based on different mecha-

nisms (Severs and Robenek, 2008; Steinbrecht and Zierold, 1987).

Freeze-fracture electron microscopy techniques emerged during

the 1950s to 1960s and have been successfully employed for the

analysis of hydrated specimens for well over 30 years. The freeze-

fracture technique physically fractures a frozen biological sample;

the structural details exposed by the fracture plane are visualised

by vacuum-deposition of platinum-carbon to make the replica for

examination in a TEM (Severs, 2007). Thus, no drying process is


10/19


Fig. 4. (A) Cryo SEM images of nanoemulsions. Image reprinted from Saupe et al. (2006) with permission from Elsevier. (B) Freeze-fracture SEM images of conventional

nanoemulsions stabilised by lecithin (L) or oleylamine (O) and corresponding silica-coated lecithin (LSA) or oleylamin (OSA) nanoemulsions.

Images reprinted from Eskandar et al. (2009) with kind permission from Springer Science and Business Media.

required and additional information aboutthe internal structure ofnano-sized colloidal carriers is obtained (Kuntsche et al., 2011).

The key steps involved in this procedure are rapid freezing of

thesample, fracturing,replicationand replica cleaning.A schematic

overview of the freeze-fracturing process for TEM as well as SEM

analysis is given in Fig. 5.

I. The rapid freezing, i.e. cryofixation, of the nano-sized suspen-

sion is usually performed by swiftly immersing the sample

into a liquid coolant such as subcooled liquid nitrogen. In this

context, pre-treatment with cryoprotectants such as glycerol

is sometimes necessary to avoid ice crystal damage. Chemical

fixation with glutaraldehyde beforehand serves to avoid arte-

facts induced by the cryoprotectant. In many cases, successful

freezing of hydrated samples requires ultrarapid freezing tech-

niques, such as optimised plunge freezing, jet freezing, sprayfreezing, high-pressure freezing or freezing by impact against a

cold metal block (Severs, 2007).

II. Subsequently, the fracturing of the sample is carried out under

vacuum at liquid nitrogen temperature by breaking the sample

in a hinged device or by using a liquid nitrogen-cooled micro-

tomeblade.If deemed necessary, an additional etchingstep may

be performedwhich consists of vacuumsublimationof iceafter

fracturing. In other words, theice can be removed from the sur-

face of the fractured specimen by freeze-drying by increasing

the temperature to about100 C for several minutes to let ice

sublime.

III. The replicas are then prepared by shadowing and backing of

the specimen. The surface of the sample is usually shadowed

with platinum to achieve a good topographic contrast and


11/19


Fig. 5. Schematic illustration of the sample preparation during the freeze-fracture

process followed by thefreeze-etching technique.The obtained specimenscan sub-

sequently be investigated by TEM or SEM.

then covered with a strengthening layer of electron-lucent car-

bon to stabilise the ultra-thin metal film (Preetz et al., 2010).

More specifically, the cold fractured surface, possibly etched,

is shadowed with evaporated platinum or gold at an average

angle of 45 in a high vacuum evaporator. A second coat of car-

bon, evaporated perpendicular to the average surface plane, is

often performed to improve stability of the replica coating. The

topographical features of the frozen, fractured surface are thus

transformed into variations in the thickness of the deposited

platinum layer of the replica (Severs, 2007).

IV. The specimen is then returned to ambient temperature and

pressure and the extremely fragile pre-shadowed metal

replica of the fracture surface is released from the underly-ing biological material by careful chemical digestion with acid

solutions or detergents. The still-floating replica is thoroughly

washed free from residual chemicals, carefully placed on fine

grids,dried andtheninvestigatedin the TEM(Preetzetal.,2010;

Severs and Robenek, 2008).

Further details and practical advice on freeze-fracture elec-

tron microscopy can be found in the literature (Gulik-Krzywicki,

1997; Severs, 2007). Overall, freeze-fracture electron microscopy

can be employed for the analysis of a large spectrum of differ-

ent materials, including liquids and dispersions, at intermediate

to low resolution. Freeze-fracture TEM (FF-TEM) is well adapted

to study lipid-containing colloidal suspensions, such as liposomes,

nanoemulsions and nanoparticles despite the relatively low signalto noise ratio of the replicas. Polymer solutions, microemulsions

and biological systems can be investigated as well (Brandl et al.,

1997; Gulik-Krzywicki, 1997; Yan et al., 2005; Zhou et al., 2010).

The most importantfeature of this technique is the tendency of the

fracture plane to follow a plane through the central hydrophobic

core of frozen membranes, thus splitting them in half. As a result,

planar views of the internal structure of the samples are obtained

(Severs, 2007).

The main drawback of FF-TEM is that the obtained information

strongly depends on the quality of the fracture (Belkoura et al.,

2004). Freeze-fracturing techniques are complex in nature and the

different steps of sample preparation, such as chemical fixation,

cryoprotective pre-treatment,cryofixation, freeze-fracturing, etch-

ing and replication, may significantly influence the appearance of

the investigated sample (Benita and Levy, 1993; Buchheim, 1982).

As for all microscopic techniques, care must be taken to avoid

misinterpretationdue to artefacts. Artefacts mayoccurduetoinsuf-

ficient freezing rates or re-deposition of solvent molecules ontothe

sample plane after fracturing (Kuntsche et al., 2011). Therefore,

considerate specimen preparation is essential to ensure repro-

ducible and reliable data.

6.2. Freeze-fracture TEM for characterisation of nanoemulsions

The morphology of a lecithin-based nanoemulsion for topical

application was investigated by FF-TEM following a standard pro-

tocol of sample cryofixation, freeze-fracturing, freeze-etching and

covering with platinum/carbon (Zhou et al., 2010). The influence

of increasing amounts of glycerol within the systems was clearly

shown (Fig. 6). The droplet size, shape and size distribution could

be monitoredvery well by thistechnique. Likewise, instability phe-

nomena suchas the agglomeration of dropletsas observedin Fig.6A

could be detected. In the same fashion, Zeevi et al. (1994) analysed

positively charged submicron emulsions using a freeze-fracture

and etching technique, thus confirming a spherical particle shape

and random particle distribution within the fracture plane.

6.3. Cryo TEM: experimental setup, sample preparation and

potential artefacts

Cryo TEM allows for the direct investigation of colloids in the

vitrified frozen-hydrated state. As with freeze-fracture TEM, infor-

mation about the internal structure of the colloidal system is

obtained (Kuntsche et al., 2011). By means of a complex sample

preparation the formulation microstructure is displayedin its origi-

nal stateand a clear differentiation between nano-sized oil droplets

and other structures can be obtained (Fox, 2009). It is therefore

the electron microscopic technique of choice for an artefact-free

visualisation of nanoemulsions.

Conventional TEM analysis of hydrated systems only provides a

limited amount of information since the aqueous compounds of

the system will evaporate rapidly under the vacuum within anelectron microscope. Therefore, the development of cryo-electron

microscopy of vitrified specimens represented a major progress in

this field (Bouchet-Marquis and Hoenger, 2011; Steinbrecht and

Zierold, 1987). During vitrification, a thin film containing the spec-

imen is plunged into a suitable cryogen such as liquidethane.Liquid

nitrogen is used to maintain the temperature of the ethane near its

melting point of183 C. Witha freezingrate around 1,000,000 K/s,

thefluidsurrounding the specimendoesnothavetimeto form crys-

talline ice, which would damage the fragile sample; it is vitrified

instead. The nature of the liquid is preserved since the first-order

exothermic phase transition between the liquid and the solid does

not take place (Bachmann and Mayer, 1987). Thus, the specimen

that is embedded within the layer of vitreous ice is essentially

preserved in its native state at high spatial resolution.Different methods for preparation of a thin vitrified layer of

aqueous specimens exist, such as plunge-, impact-, spray- or high

pressure-freezing. For investigation of nanoemulsions, plunge-

freezing and high pressure-freezing are the most commonly used

techniques. Detailed information on these cryofixation procedures

can be found in the literature (Dubochet et al., 1982; Lepault et al.,

1983; Sitte et al., 1987). Various kinds of cryo specimenholdersand

plunge-freezing devices are available.

One of the basic requirements for TEM is that the specimen

must be very thin. To produce films of adequate thickness, three

main techniques can be employed. Firstly, blotting of the sample

is performed followed by vitrification (Danino and Talmon, 2000).

This method is simple and rapid, but is limited to the analysis of

liquid suspensions of moderate viscosity with particle diameters


12/19


Fig. 6. Freeze-fracture TEM micrographs of different lecithin-based nanoemulsions. Images reprinted from Zhou et al. (2010) with kind permission from Springer Science

and Business Media. The influence of increasing contents of glycerol within the formulation is demonstrated by decreasing particle sizes and a more homogeneous droplet

size distribution. Coalescence phenomena as indicated by thearrows in image A can be detected.

of less than 200n m. Secondly, cryo-sectioning using high pres-

sure freezing devices after vitrification of the bulk sample can be

performed (Moor, 1987; Mueller and Moor, 1984). In this con-

text, different cutting techniques to obtain vitrified sections thin

enough for high resolution observation are available (Al-Amoudi

et al., 2003; Dubochet et al., 1983; McDowall et al., 1983). Thirdly,a film of the liquid sample can be prepared that is subsequently

vitrified as a thin layer. The thicknessof the films can be controlled

interferometrically before vitrification (Denkov et al., 1996). The

latter two methods can be applied for various tasks, but are rather

complex and demanding in nature. Thus, vitrification after blotting

of the sample is the most widely employed technique.

The sample preparation for cryo TEM involves three main

steps:

I. A small aliquot (approximately 3l) of a fluid suspension con-taining the sample is applied to the surface of a supporting

substrate such as a holey or continuous carbon film that is

attachedto the surface ofa standardTEM specimengrid(Fukamiand Adachi, 1965; Reichelt et al., 1977).

II. Subsequently, the droplet is carefully blotted with filter paper

until most of the supernatant liquid is removed and only a thin

layer of approximately 100 nm thickness is left on the support

substrate. On perforated carbon support films, the thin sample

film is left stretched over the holes (Fig. 7A).

III. The thin fluid layer is rapidly immersed into a suitable cryo-

gen of high heat capacity, which leads to instantaneous and

contaminant-free freezing. This vitrification or shock-freezing

by propelling the grids into a cryogen is ideally performed by

means of a plunging device and a humidity- and temperature-

controlled environmental vitrification system (Dubochet et al.,

1987;Norden et al., 2001). Thevitrifiedsamples then need to be

transferred into the TEM cryo holder of the microscope under

liquidnitrogen and areexaminedat temperatures around 100 K

(Marxer et al., 2011; Yilmaz and Borchert, 2005).

There are two key factors regarding specimen preparation that

are critical to obtaining high quality cryo TEM data, assuming thatbiochemical integrity of the specimen is given: the proper prepa-

ration of the support substrate and the considerate blotting of the

sample droplet to a thin fluid layer on the substrate prior to freez-

ing. The rapid drop in temperature during vitrification provides

the possibility to capture hydrated specimens in their momentary

movement without changing their structure and helps to reduce

the effect of electron beam damage.

Most colloidal systems such as nanoemulsions give compara-

tively poor contrast in cryo TEM. The inherently low contrast of

unstained vitrified specimens can be compensated to a large extent

by an optimal use of phase contrast (Adrian et al., 1984). A varia-

tion in focus canserve to revealdifferent parts of a structurewhile a

total defocus creates optical artefacts (Kuntsche et al., 2011). Addi-

tional staining techniques canbe employed to produce images withimproved contrast (Adrian et al., 1998).

As in allelectronmicroscopictechniques,different artefactsmay

hinder cryo TEM analysis. Adequate cooling of the sample at all

stages of preparation and analysis is essential. Otherwise, artefacts

may emerge during the freezing process. Ice crystal formation or

modifications due to humidity or temperature changes may hinder

the structural analysis. Inadequate cooling within the microscope

leads to damage of the vitrified sample within the vacuum of the

TEM by a kind of freeze-drying effect (Kuntsche et al., 2011).

Moreover, electron beam damage may occur during observa-

tion of frozen materials in cryo TEM although a certain protection

of the sample is obtained after the cryo-fixation procedure. Elec-

tron beam damage is especially severe in the presence of water

(Belkoura et al., 2004). Since most colloidalsystems are sensitiveto


13/19


Fig. 7. Images of a typical nanoemulsion analysed by cryo TEM after plunge freezing. (A) A conventional TEM grid coated with a holey carbon film is shown. The letters

indicate thecarbon layer (C), theholesin thecarbonfilm (H)and thefrozen hydrated sample (S). Thescheme on theright hand side depicts theformation of thethin liquid

sample layer which remains after blotting with a filter paper just before the freezing procedure. (B) Segregation of nanoemulsion droplets: smaller droplets remain in the

thinner parts of the sample filmwhile larger ones areonly found in thethicker parts of thesample layer.

radiation damages, the time period for sample viewing is limited

and low dose conditions are recommended. Such minimal dose

microscopy canbe performedon all modern electronmicroscopes;

thus, the strength of the electron beam can be handled (Belkoura

et al., 2004). Severe electron beamdamage may result in bubbling

of the sample and out-of-focus images (Kuntsche et al., 2011). The

modified techniqueof cryonegative staining,which should provide

both maintained sample hydration and protection against electron

beam damage, bears the risk of selective particle orientation due to

interfacial forces and flattening of fragile structures (Harris, 2008).

Thepotential presence of cryogen residues that remainfrom the

plunge-freezing process likewise needs to be taken into account.

An increased energy input may remove these remnants, but may

induce a phase transition of the vitrified ice into cubic or hexag-onal ice (Kuntsche et al., 2011). Likewise, ice contamination may

occur due to a high content of evaporated water in the TEM col-

umn (Friedrich et al., 2010). For examples of the potential artefacts

in cryo TEM analysis of colloidal systems the reader is referred to a

recent article by Kuntsche et al. (2011).

Apart from the risk of artefacts, there are certain general limita-

tions to the cryoTEM technique for investigationof nanoemulsions.

To obtain adequatefilm thicknessis currently themajor difficulty in

preparingsamplesfor direct imagingwith cryoTEM (Belkoura et al.,

2004). The maximum specimen thickness that can be observed is

limited to a few hundred nanometers, i.e. the film has to be thin

enough to allow investigation by cryo TEM. However, at the same

time the filmmust be thickenough notto influence themicrostruc-

ture of the sample (Belkoura et al., 2004).As already mentioned, blotting of the sample followed by vitri-

fication is the most commonly employed technique in cryo TEM of

nanoemulsions. However, the blotting procedure bears the risk of

preparation artefacts. Concentration changes of the thinned sam-

ple drop, size segregation and changes of internal structure due

to shearing effects may occur (Belkoura et al., 2004). Thus, the

blotting procedure has to be adjusted for each particular sample

(Talmon, 1999). Nevertheless, cryo electron micrographs of poly-

disperse systems may be biased towards small particles due to the

preparation technique. The specimen preparation involves appli-

cation of the liquid sample on the microscopic grid and removal

of the surplus liquid with filter paper until an ultra-thin sample

film remains in the holes of the grid, particularly in their cen-

tre (Fig. 7B) (Jores, 2004a; Jores et al., 2004b). Structures which

exceed the thickness of this film are either removed or relocated

to thicker areas of the film during this procedure. Unfortunately,

areas of increased thickness are often too sensitive towards the

electron beam to deliver reliable results upon investigation. As a

consequence, aggregates or droplets with large dimensions may

remain undetected. Thus, a direct comparison of droplet sizes as

observed in cryo TEM with the results of particle size measure-

ments by DLS or laser diffraction should be performed with great

caution (Jores, 2004a,b). The data obtained by cryo TEM should

be regarded as complementary qualitative information about the

shape and size of the observed particles. A quantitative evaluation

of cryo electron micrographs of a certain formulation aiming at an

accurate size distribution of the observed droplets would require

evaluation of large amounts of individual images and specific pro-grammes.

Nevertheless, cryo TEM of frozen-hydrated unstained spec-

imens is presently among the preferred approaches for high-

resolution studies because it provides data on the fully native

structure and some protection of the specimens against elec-

tron radiation damage (Massover, 2008). Technical improvements

of all steps of sample preparation, such as environmental con-

trol during vitrification and improved data processing through

CCD cameras and digitalisation, promote the further use of this

technique (Kuntsche et al., 2011). Three-dimensional reconstitu-

tion (Orlova et al., 1999) or cryoelectron tomography (Koning

and Koster, 2009) are interesting developments to obtain infor-

mation about the three-dimensional structure of the investigated

systems. Thus, the projection limitation of TEM can be par-tially solved. Three-dimensional tilt-series-based tomography or

even simple stereo pairviews obtained with tilting capable stages

and CCD camera image acquisition give a rapid insight into the

three-dimensional structure of nanocarrier systems not only from

vitrified, but also from negatively stained or freeze-fractured

samples (Baumeister, 2002; Frank, 1992; Hoppe, 1981). Further-

more,freeze-fracture direct imagingoffers interesting prospectsfor

investigation of hydrated colloidal samples (Belkoura et al., 2004).

This method combines elements of the freeze-fracture technique

with direct imaging by cryo TEM. Many experimental artefacts

caused by the blotting procedure in cryo TEM can be avoided;

thus, this technique, although not yet applied to nanoemulsion

samples, represents a promising approach to obtain artefact-free

images.


14/19


Cryo TEM is particularly useful to investigate structural details

of colloidal nano-sized systems, e.g. to detect the presence of vesi-

cles among nanoemulsion droplets (Fig. 1B) (Norden et al., 2001).

Overall, nanoemulsion droplets are comparatively simple to dis-

tinguish in cryo TEM images. They always appear as spherical dark

droplets while solid lipid nanoparticles, liposomes or other related

lipid structures may appear as needle- or rod-like structures when

viewed edge-on (Jores, 2004a,b). The systems can be investigated

with (Hatanaka et al., 2010) or without dilution (Marxer et al.,

2011). In general, dilution of the investigated nanoemulsions is

advisable since examination of undiluted samples frequently leads

to crowded images where individual structures cannot be clearly

characterised.

6.4. Cryo TEM for characterisation of nanoemulsions

A number of recent cryo-microscopic investigations of

nanoemulsions based on eudermic surfactants have allowed for

detailed insights into the versatile nature of such colloidal sys-

tems (Klang et al., 2010, 2011a,b; Teixeira et al., 2000). Teixeira

et al. reported unusual bicompartmental structures with associ-

ated oil and water phases among regular nanoemulsion droplets

(Teixeira et al., 2000). These handbag-shaped structures wereslightly deformed (Fig. 8A), which was ascribed to molecular

rearrangements at the interface due to an interaction between

the employed lecithin and the positively charged stearylamine.

Similar conclusions were reported for nanoemulsions with the

positively charged phytosphingosine (Yilmaz and Borchert, 2005)

although the latter image (Fig. 8B) is of questionable quality. This

emphasises the need for careful interpretation of the obtained

results so as to avoid misinterpretations which are still frequently

observed.

In more recent approaches, we have investigated the morphol-

ogy of lecithin-based nanoemulsions both in the presence (Klang

et al., 2010) and absence (Klang et al., 2011a) of additional emulsi-

fiers orstabilisers. Itwas foundthatthe lecithinmixture aloneled to

the formation of a random mnagerie of rather irregularly shapeddroplets and vesicles (Fig. 9A). This corresponds well with the

respective literature. Phosphatidylcholine alone is rather unsuit-

able forthe formationof curvedsurfaces(Shchipunov, 1997; Trotta

et al., 2002), and the performance of phospholipids mixtures which

largely consist of phosphatidylcholine can be improved by addi-

tional surfactants (Hoeller et al., 2009; Trotta et al., 2002; Yilmaz

andBorchert, 2005). Interestingly, a remarkable change in formula-

tionmicrostructurewas alsoachieved by additionof-cyclodextrinas stabilising agent (Fig. 9C). The number of additional vesicular

structures decreased notably while the droplet shape appeared

more homogeneous and spherical. These findings highlight the

importance of a considerate choice of excipients during formulation

development and the role of electron microscopy to monitor the

results. Comparative studies with a sucrose ester mixture, namelysucrose stearate S-970, revealed that this mixture of eudermic sur-

factantswas moresuitable for the formationof nano-sized droplets

than the lecithin mixture (Fig. 9B). In this case, further addition

of -cyclodextrin had no notable impact on the composition ofthe colloidal system or the droplet shape (Fig. 9D). In fact, any

surplus of surfactant might lead to the formation of additional

aggregates such as multilamellar structures (Fig. 10A) which may

or may not be favourable for the respective formulation proper-

ties (Klang et al., 2010). Similar observations were recently made

fora nanoemulsionstabilised by 5% of sucrose stearate (Klang et al.,

2011b). Ascanbeseenin Fig. 10B, a highly crowded microstructure

was observed despite dilution of the sample. The emulsion droplets

appeared slightly deformed, which may have been related eitherto

altered formulation viscosity and behaviour during high-pressure

homogenisation or to droplet repulsion due to the high negative

zeta potential values (Klang et al., 2011b).

7. Advanced techniques: cryo analytical TEM (cryoATEM)

The technique of electron energy-loss spectroscopy is of great

interest for the analysis of lipid nanocarriers, especially in com-

bination with cryo-preparation techniques (Cryo EELS). Cryo EELS

represents a valuable approach to analyse both the morphology

of a formulation and its chemical composition since the energy

spectrum of electrons passing through the specimens also contains

information about the element composition. When the electron

beam impacts the particle, some of the electrons are inelastically

scattered and lose a part of their energy. Every element possesses

its own specific energyloss.Thus, the elemental composition of the

particle or droplet can be determined by analysis of this specific

energy by means of a spectroscope attached to the electron micro-

scope (Egerton, 1986). Based on this phenomenon, EELS can be

used for chemical analysis of various liquid dispersions, including

systems containing polymers.

Another technique whichis widely used for the elemental anal-

ysis in combination with electron microscopy is energy dispersive

X-ray spectroscopy. When a high-energy electron beam impingesupon a specimen, X-ray photons are produced. Characteristic X-

rays have well defined energies which are characteristic of the

atoms in the specimen (Sawyer and Grubb, 1996; Sutton et al.,

2003). Thus, the elemental composition of the investigated sample

can be determined. In most cases EELS is expected to offer higher

spatial resolution than EDX because the effect of beam broaden-

ing and aberrations of the probe-forming lens can be controlled by

means of an angle-limiting collection aperture. The technique of

EELS as well as energy-filtered transmission electron microscopy

(EFTEM) imaging are usually applied to the detection of light ele-

ments suchas C,N orO, where EDX islesssensitive (Egerton, 1986).

Isaacson and Johnson (1975) provided the theoretical basis for the

prediction of detectionlimits in EELSby introducingthe concepts of

minimum detectable mass (MDM) and minimum detectable massfraction (MMF). The MDMdescribes the smallest amount of mate-

rial that can be detected in a given matrix. A small beam diameter is

desirable, thus the use of a scanning transmission electron micro-

scope equipped with a field emission sourceis preferable (Krivanek

et al., 1991; Leapman, 2003). Alternatively, the MFF represents the

smallest concentration of elements that canbe measured in a given

matrix. Thisparameter depends primarilyon thetotal currentavail-

able in the probe. Thermal emission tips that provide large beam

currents and conventional transmission electron microscopy mode

may thereforebe preferable(Kothleitner andHofer, 1993; Zhuet al.,

2001). In the ideal case of a sample that is not susceptible to beam

damage, traces of chemical elements as low as 0.030.01% can be

detected when using a 200kV instrument (Riegler and Kothleitner,

2010). However, the cryo EELS detection limit is much lower, thusrendering the analysis of nanoemulsion compositions a challeng-

ing task (Yakovlev et al., 2010). Frozen hydrated nanoemulsions

may contain only low amounts of specific atoms like N, P, Cl or F.

Thus, the statistical noise obscures the weak energy-loss near edge

structure (ELNES) signals leading to increased errors in background

extrapolation and further deterioration of the detection limits.

Moreover, cryo EELS usually requires the use of primary energies

of 200kV (in contrast to the conventionally used 80120kV for

cryo TEMimaging) andhighbeamcurrents, whichleadto complete

beam damage of the specimen structure before the adequate spec-

tracanbe recorded (Egerton,1986). Currentlythere are twogeneral

types of energy filters (post-column(GIF) and column-integrated

(omega)) which are used for EELS/EFTEM analysis (Egerton, 1986,

2005).


15/19


Fig. 8. (A) Cryo TEM micrograph of a positively charged nanoemulsion containing deformed bicompartmental structures. Image reprinted from Teixeira et al. (2000) with

kind permission from Springer Science and Business Media. (B) Cryo TEM image which presumably shows similar structures containing phytosphingosine and ceramides.

The latterimageappears to sufferfrom severeout-of-focus problems,perhapsdue to insufficient freezing of thesample. Thelargestructures indicatedby thearrows cannot

be clearly identified as emulsion droplets and might be ice crystals.

Image reprinted with permission from Elsevier (Yilmaz and Borchert, 2005).

Fig. 9. Visualisation of differences in formulation morphology: cryo TEM micrographs of diluted nanoemulsions (1:10, v/v) stabilised by (A) 2.5% of lecithin alone (mean

particle size: 186.4111.06, n=3) or(C) 2.5% oflecithinand-cyclodextrin (meanparticle size:175.8200.47, n =3), (B)2.5% of sucrose stearate alone (mean particle size:141.2108.73, n=3) or (D) 2.5% ofsucrose stearate and -cyclodextrin (meanparticle size:144.7710.61, n=3). Theaddition of thecyclodextrin apparently promotes the

formationof spherical droplets in case of lecithin-based systems (Fig.7A versus7C). Thearrowheads included in Image 7A indicate thepresence of vesicular structures, such

as liposomesand multilamellar layers. In case of sucrose-stearate based systems, hardly any differences were observed after addition of-CD.

All images reprinted from Klang et al. (2011a) with permission from Elsevier.


16/19


Fig.10. (A)Cryo TEMmicrograph of a nanoemulsionstabilised by lecithin,sucrose stearate and-cyclodextrin. The excess of surfactant induces the formationof additional

multilamellar structures. Image reprinted from Klang et al. (2010) with permission from Elsevier. (B) Cryo TEM micrograph of a nanoemulsion stabilised by comparatively

high amounts of sucrose stearate (5%, w/w) after dilution with distilled water (1:10, v/v). It remains to be investigated whether thedeformation of thedroplets was caused

during production or during sample preparation forcryo TEM.

Image reprinted from Klang et al. (2011b).

Cryo EELS has not yet found broader application for nanoemul-

sion characterisation. Firstly, frozen hydrated nanoemulsions

usually contain specific atoms like N, P, Cl or F in such low con-

centrations that the detection limit of cryo EELS does not allow to

obtain fine ELNES details of these elements in the spectra, which

is not the case for conventional biological samples or water map-

ping. Secondly, cryo EELS usually requires usage of an electron

beam voltage up to 200 kV, which leads to complete beam damage

Fig. 11. Application of cryo EELS for the analysis of frozen hydrated samples: differentiation of calcium carbonate particles (A) from ice crystal formations (B). The scale

bars represent 100nm. The corresponding chemical composition of the analysed compound is depicted on the right hand side (C, D). Vitrified specimens were examined in

a field emission Tecnay F20 TEM operating at an accelerating voltage of 200kV using an Oxford CT3500 cryo holder (Oxford Instruments, UK) that maintained the vitrified

specimens at 160

C during sample observation.Images were recorded digitally on a cooled UltraScan CCD (Gatan, USA).


17/19


18/19


19/19


Patravale,V.B., Mandawgade,S.D., 2008.Novel cosmetic delivery systems:an appli-cation update. Int. J. Cosmet. Sci. 30, 1933.

Piemi, M.P.,Korner, D.,Benita,S., Marty, J.-P., 1999.Positively andnegatively chargedsubmicron emulsions for enhanced topical delivery of antifungal drugs. J. Con-trol. Release 58, 177187.

Pinnamaneni, S., Das, N.G., Das, S.K., 2003. Comparison of oil-in-water emul-sions manufactured by microfluidization and homogenization. Pharmazie 58,554558.

Preetz, C., Hauser, A., Hause, G., Kramer, A., Mader, K., 2010. Application of atomicforcemicroscopy and ultrasonic resonatortechnology on nanoscale: distinctionofnanoemulsions from nanocapsules. Eur. J. Pharm. Sci. 39, 141151.

Rao, J., McClements, D.J., 2010. Stabilization of phase inversion temperaturenanoemulsionsby surfactantdisplacement.J. Agric.Food Chem.58, 70597066.

Reichelt, R., Koenig, T., Wangermann, G., 1977. Preparation of microgrids as speci-mensupports forhigh resolution electron microscopy. Micron8, 2931.

Reimer, L., 1993. Image Formation in Low-Voltage Scanning Electron Microscopy.SPIE Optical Engineering Press, Bellingham & Washington.

Riegler, K., Kothleitner, G., 2010.EELS detectionlimits revisited:Rubya casestudy.Ultramicroscopy 110, 10041013.

Rotenberg, M., Rubin, M., Bor, A., Meyuhas, D., Talmon, Y., Lichtenberg, D., 1991a.Physico-chemical characterization of Intralipid emulsions. Biochim. Biophys.Acta 1086, 265272.

Rotenberg, M., Rubin, M., Bor, A., Meyuhas, D., Talmon, Y., Lichtenberg, D., 1991b.Physico-chemical characterization of IntralipidTM emulsions. Biochim. Biophys.Acta 1086, 265272.

Ruozi, B.,Belletti, D.,Tombesi,A., Tosi, G.,Bondioli, L.,Forni, F.,Vandelli, M.A., 2011.AFM, ESEM TEM, andCLSM in liposomal characterization: a comparative study.Int.J. Nanomed. 6, 557563.

Ruska, E.,Knoll, M., 1932. Das Elektronenmikroskop. Z. Phys. 78, 318339.Saupe, A., Gordon, K.C.,Rades, T., 2006. Structuralinvestigationson nanoemulsions,

solid lipid nanoparticles and nanostructured lipid carriers by cryo-field emis-sion scanning electron microscopy and Raman spectroscopy. Int. J. Pharm.314,5662.

Sawyer, L.C., Grubb, D.T., 1996. Polymer Microscopy, 2nd ed. Alden Press, Oxford,England.

Schalbart, P.,Kawaji, M., Fumoto, K., 2010. Formation of tetradecane nanoemulsionby low-energy emulsification methods. Int. J. Refrigeration 33, 16121624.

Scheuing, D.R., 1990. Fourier TransformInfrared Spectroscopy in Colloid and Inter-face Science. American Chemical Society, Washington, DC, pp. 121.

Schmidtgen, M.C., Drechsler, M., Lasch, J., Schubert, R., 1998. Energy-filtered cry-otransmission electron microscopyof liposomespreparedfrom humanstratumcorneum lipids.J. Microsc. 191,177186.

Schwarz, J.S., Weisspapir, M.R., Friedman, D.I., 1995. Enhanced transdermal deliv-ery of diazepam by submicron emulsion (SME) creams. Pharm. Res. 12, 687692.

Seki, J., Sonoke, S., Saheki, A., Fukui, H., Sasaki, H., Mayumi, T., 2004. A nanometerlipid emulsion, lipid nano-sphere (LNS),as a parenteral drug carrier for passivedrug targeting. Int. J. Pharm. 273, 7583.

Severs,N., Robenek,H., 2008.Freeze-fracturecytochemistryin cellbiology.MethodsCell Biol. 88, 181204.Severs, N.J., 2007. Freeze-fracture electron microscopy. Nat. Protocols 2, 547576.Shafiq, S., Shakeel, F., Talegaonkar, S., Ahmad, F.J., Khar, R.K., Ali, M., 2007. Design

anddevelopmentof oraloil inwater ramiprilnanoemulsionformulation: invitroand in vivoassessment. J. Biomed. Nanotech.3, 2844.

Shakeel, F., Baboota, S., Ahuja, A., Ali, J., Aqil, M., Shafiq, S., 2007. Nanoemulsions asvehicles for transdermal delivery of aceclofenac. AAPSPharm. Sci.Tech. 8, E104.

Shchipunov, Y.A., 1997. Self-organising structures of lecithin. Russ. Chem. Rev. 66,301322.

Silva, A.P., Nunes, B.R., De Oliveira, M.C., Koester, L.S., Mayorga, P., Bassani, V.L.,Teixeira, H.F., 2009. Development of topical nanoemulsions containing theisoflavone genistein. Pharmazie 64, 3235.

Singh, K.K.,Vingkar,S.K., 2008.Formulation, antimalarial activityand biodistributionoforal lipid nanoemulsion of primaquine. Int. J. Pharm.347, 136143.

Sitte, H., Edelmann, L., Neumann, K., 1987. Cryofixation without pretreatment atambient pressure.In: Steinbrecht,R., Zierold,K. (Eds.),Cryotechniques in Biolog-ical ElectronMicroscopy. Springer,Berlin,Heidelberg, Germany,pp. p.87p.113.

Sjoeblom, E., Friberg, S., 1978. Light-scattering and electron microscopydetermina-tions of association structures in W/O microemulsions. J. Colloid Interface Sci.67, 1630.

Solans, C., Izquierdo, P., Nolla, J ., Azemar, N., Garcia-Celma, M.J., 2005. Nano-emulsions. Curr. Opin. Coll. Int. Sci. 10, 102110.

Sole, I., Pey, C.M., Maestro, A., Gonzalez, C., Porras, M., Solans, C., Gutierrez, J.M.,2010. Nano-emulsions prepared by the phase inversion composition method:preparation variablesand scale up. J. Colloid Interface Sci. 344, 417423.

Sonneville-Aubrun,O., Simonnet, J.T.,LAlloret,F., 2004.Nanoemulsions:a newvehi-cle forskincare products. Adv. Colloid Interface Sci. 108109, 145149.

S te inbrecht , R.A., Ziero ld, K. , 1987. Cry o Techniques in Biological Electro nMicroscopy. Springer, Berlin, Heidelberg, Germany.

Sutton, S.R.,Newville,M., Rivers, M.L.,2003. Synchrotron X-raymicroscopeanalysis.Geophys. Res. Abstracts 5, 13260.

Tadros, T., Izquierdo, P., Esquena,J., Solans, C., 2004.Formation andstabilityof nano-emulsions. Adv. Colloid Interface Sci. 108109, 303318.

Takegami, S.,Kitamura, K., Kawada, H., Matsumoto, Y., Kitade, T., Ishida, H.,Nagata,C., 2008.Preparationand characterization of a newlipid nano-emulsioncontain-ing two cosurfactants, sodium palmitate fordroplet size reductionand sucrosepalmitatefor stabilityenhancement.Chem.Pharm.Bull. (Tokyo) 56, 10971102.

Talmon, Y., 1999. Cryogenic temperature transmission electron microscopy in thestudy of surfactant systems. In: Binks, B.P.E. (Ed.), Modern CharacterizationMethods of Surfactant Systems. Marcel Dekker, New York, p. p147.

Tamilvanan, S., Benita, S., 2004. The potential of lipid emulsion for ocular deliveryoflipophilic drugs. Eur. J. Pharm. Biopharm. 58, 357368.

Teixeira, H., Dubernet, C., Rosilio, V.,Benita, S.,Lepault, J., Erk, I., Couvreur, P.,2000.Newbicompartmental structuresareobservedwhenstearylamineismixed with

trigly

electron microscopy of nanoemulsions_an essential tool for characterisation and stability assessment

Documents