Imaging of Anisotropic Cellulose Suspensions Using Environmental

Biomacromolecules 2003, 4, 510-517

510

Imaging of Anisotropic Cellulose Suspensions Using Environmental Scanning Electron Microscopy Aline F. Miller and Athene M. Donald* Polymers and Colloids Group, Cavendish Laboratory, Department of Physics, University of Cambridge, Cambridge CB3 0HE, U.K. Received July 15, 2002; Revised Manuscript Received September 27, 2002

The effect of concentration on anisotropic phase behavior of acid-hydrolyzed cellulose suspensions has been examined using conventional polarizing microscopy and the novel technique of environmental scanning electron microscopy (ESEM). Microcrystalline cellulose dispersed in water formed biphasic suspensions in a narrow concentration range, 4-12 wt % for a suspension pH of 4, where the upper and lower phases were isotropic and anisotropic (chiral nematic), respectively. It is known from previous work that within the biphasic regime total suspension concentration affects only the volume fractions of the two phases, not phase concentration or interfacial packing. As the total suspension concentration surpassed the upper critical limit (c**), however, a single anisotropic phase of increasing concentration was observed. It was evident from polarizing microscopy that the chiral nematic pitch of the anisotropic phase decreased with increasing concentration, which has been attributed to a reduction in the electrostatic double layer thickness of the individual rods, thus increasing intermolecular interactions. Chiral nematic textures were also visible using ESEM. This technique has the advantage of studying individual rod orientation within the liquid crystalline phase as it permits the high resolution of electron microscopy to be applied to hydrated samples in their natural state. To our knowledge this is the first time such lyotropic systems have been observed using electron microscopy. Introduction There has been great interest in recent years in applying soft condensed matter physics to living systems.1 The disciplines of both physics and biology require an understanding of structure-property relationships in complex fluids on the mesoscopic scale, for example the structure and dynamics of colloids, liquid crystals, or polymers. Interdisciplinary research has been driven by advances in material synthesis that have provided well-defined molecules together with the realization that many of the mathematical physical models can be applied to biological systems. Such biological systems are wet, and not suited to extensive sample preparation that more traditional materials undergo before characterization in an electron microscope. The need is great, therefore, for new techniques to provide insights into the morphology and properties of materials of interest in their natural state. In this article we show the development of one technique, namely, environmental scanning electron microscopy (ESEM), and apply it to study the most common biopolymer, cellulose. ESEM is a relatively new technique that has already expanded the range of samples examined in their original state with electron microscopy, from oil-water emulsions2 to aggregating amyloid fibrils3 and Langmuir films.4 Like conventional scanning electron microscopy (SEM), ESEM is a surface technique offering high spatial resolution. * To whom correspondence may be addressed. E-mail address: amd3@ phy.cam.ac.uk.

Conventional SEM has two main drawbacks: the sample chamber has to be kept under high vacuum conditions and the sample itself must be conducting, or be given an electronically grounded conductive coating. The ESEM, conversely, allows the presence of a pressure (up to 20 Torr) inside the sample chamber, and when water vapor is selected as the gas, hydrated and insulating samples can be analyzed. This allows the sample structure to be probed in its natural state with no sample preparation. Morphologies obtained for biofilms and pharmaceutical formulations using both environmental and conventional SEM have been compared, and such studies have demonstrated that standard sample preparation for conventional SEM has, at least sometimes, introduced artifacts.5-7 The powerful technique of ESEM is therefore ideal for the high-resolution study of biological samples in their natural environment. Many biological macromolecules that play a key role in self-assembly processes in nature display liquid crystalline characteristics: for example rod-shaped materials such as tobacco mosaic virus8 have displayed nematic order and DNA fragments,9 chitin,10 collagen,11 and κ-carageenan12 chiral nematic (cholesteric) order, all in aqueous suspensions. The terms cholesteric and chiral nematic are both used, but chiral nematic is more accepted today. Cellulose, the world’s most common biopolymer, forms chiral nematic phases when dispersed in water,13-18 where the source of the chiral arrangement arises from the tight packing of the screwlike rods.19-21 A schematic diagram of such organization is given in Figure 1, and when viewed in the polarizing microscope

10.1021/bm0200837 CCC: $25.00 © 2003 American Chemical Society Published on Web 03/01/2003

Imaging of Anisotropic Cellulose Suspensions

Biomacromolecules, Vol. 4, No. 3, 2003 511

Figure 1. Schematic representation of rod orientation in both the isotropic and chiral nematic (chiral nematic) phase.

consists of near-parallel sets of swirling lines, reminiscent of fingerprint patterns. The individual lines are associated with the layerlike optical periodicity along the helix axes, and the line spacing is half the helical pitch. Such supramolecular structure of cellulose has been found in cell walls of plants22 but is important commercially too as cellulose has an estimated production of approximately 1011 tons per year.23 Paper products constitute the largest use of cellulose, but other applications include fillers in tablets, fat replacement, texturizing agent, and stabilizer in foods. Commercial microcrystalline cellulose is mainly made from hydrolysis with HCl and formed in high shear rates. When sulfuric acid is used to hydrolyze cellulose, sulfate ester groups are introduced onto the surface of the cellulose rod and can electrostatically stabilize the suspensions.13 By careful control of the preparation method (hydrolysis time, temperature) the surface charge can be manipulated.17 It has been found that at sufficiently high suspension concentrations an ordered, birefringent phase was formed, and when the sample was purified it formed a chiral nematic phase above a critical concentration.14 The organization and phase formation of such particles have been the subject of many theoretical papers24-29 where behavior is postulated to depend on axial ratio, suspension concentration, and ionic strength. Experimentally, much work has involved characterization of the cellulose structure on the subnanometer level using for example NMR spectroscopy,30 neutron spectroscopy,16 and various microscopy techniques31-36 including transmission electron, atomic force, and optical microscopy. Atomic force microscopy (AFM) has been most successful,35,36 giving topographic profiles of individual cellulose chains down to the atomic level, but artifacts due to the convolution of the tip dimensions with sample dimension were introduced. Other studies have fallen foul either of extensive sample preparation, of limited resolution, or of expensive source of probing radiation. ESEM overcomes all of these problems. Phase behavior and organization of isotropic and anisotropic cellulose suspensions have been investigated here using both optical and environmental scanning electron microscopy to provide an understanding of cellulose rod self-assembly at the meso- and nanoscopic length scale, in their natural state. ESEM could have been used as a stand-alone technique, but since this is the first time liquid-crystalline patterns of hydrated samples have been observed using electron microscopy, we use the conventional polarizing microscope to validate our results.

Figure 2. Schematic representation of the column of the ESEM, showing how pressure varies down the column. DP and RP denote diffusion and rotary pump, respectively.

Background to ESEM The environmental scanning electron microscope is a development of conventional SEM that has the advantage of being able to image uncoated and hydrated samples. For a detailed description of the technique, the reader is referred to early reviews by Danilatos36-40 and previous experimental papers.41,42 We outline briefly here the two major differences between environmental and conventional SEM; these are the vacuum system operation and the mechanism of electron detection. It is possible to keep water in an electron microscope and maintain high resolution by using a differential pumping system to provide a pressure gradient down the column (depicted schematically in Figure 2). Such a system provides the high vacuum (10-7 Torr) required for the operation of the electron gun while simultaneously allowing the presence of water vapor in the sample chamber (varied between 0 and 20 Torr and maintained indefinitely). By careful control of this pressure and sample temperature, a 100% humidity in the chamber can be achieved, and manipulation of either parameter can induce evaporation or condensation of water on the sample. Typical conditions for a hydrated sample are obtained when the sample temperature is ∼275 K while the chamber pressure is maintained at ∼4 Torr. Under these lowpressure conditions, high-magnification images (up to ∼10 nm) can be obtained provided the working distance is ca. 5-10 mm to ensure minimal multiple scattering of electrons. It should be noted here that care must be taken during the initial pump-down procedure as any dehydration or “boiling” of water in the sample could introduce artifacts. To avoid this, the pump-down sequence, developed by Cameron and Donald,43 was employed in all experiments and drops of water were placed at numerous positions inside the chamber to ensure the sample environment was kept as close to saturation as possible. Second the nonvacuum conditions of the sample chamber required the development of a new detection system, as those used in conventional SEM involved high voltages that result in the breakdown of the gas. This led to the development of the gaseous secondary electron detector (GSED),44,45 whose operation capitalizes on the presence of the gas by using it

512

Biomacromolecules, Vol. 4, No. 3, 2003

to enhance the secondary electron (SE) signal via gaseous cascade amplification process: as SEs migrate through the sample chamber toward the positively biased detector, they undergo ionization collisions with gas molecules producing more free electrons and positive ions. The “daughter” electrons produced undergo further collisions, thus producing a cascade and an amplified signal, while the positive ions drift back to the sample surface where they dissipate any charge buildup, thus negating the need for a conducting surface. (Further details can be found in refs 37-42 and references therein.) Experimental Section (A) Sample Preparation. The microcrystalline cellulose suspensions were prepared by acid hydrolysis13,14 of Whatman No. 1 filter paper (Aldrich). The experimental procedure has been set out in an earlier paper,46 hence only a brief synopsis is given here. Hydrolysis conditions (temperature and duration of hydrolysis) were kept constant to ensure identical numbers of charged sulfate ester groups were introduced onto the cellulose chains, within experimental error. This was checked by conductometric titration with dilute sodium hydroxide solution (0.05 M). The resulting suspensions were all dialyzed against 0.01 M HCl in order to remove residual sulfuric acid and also to maintain a constant ionic strength. The polydispersity of the axial ratio was subsequently reduced by size fractionation by concentrating the suspension via evaporation until 20% of the solution was anisotropic. This lower anisotropic phase was discarded, and the remaining suspension was concentrated further until only ca. 20% remained isotropic. The upper phase was also discarded, and the process repeated to ensure the removal of extra long and short length rods. A series of concentrations were prepared by careful dilution of a concentrated stock sample with distilled water, and the desired concentration of HCl was added until all solutions were at a constant pH of 4. Isotropic/anisotropic volume fractions were measured by placing each homogeneous suspension in a sealed glass vial (5 cm high with a 2 cm diameter), exposing the solution to 1 min in the ultrasonicator to break up any coagulated flocs and allowing 2-3 days for equilibration. Phase separation was quantified simply by measuring the height fraction of each phase using a ruler. (B) Microcrystalline Characterization. (1) Atomic Force Microscopy. Rod dimensions have been determined by atomic force microscopy (AFM) (Digital Instruments Nanoscope II). An aliquot of the suspension was diluted (0.1 total weight (wt) %) and applied to a freshly cleaved mica surface. The grid was then washed with 3 × 10 mL of distilled water, and fibrils were observed using AFM in the scanning mode. (2) Polarizing Microscopy. The texture of the liquid crystalline phase was determined by pipetting an aliquot from each sample into a flat, glass tube, 0.4 mm wide, sealing the tube at both ends and viewing it through crossed polars (Zeiss universal microscope). The chiral nematic pitch was determined by measuring the distances between the fingerprints using a calibrated slide.

Miller and Donald

(3) Environmental Scanning Electron Microscopy. For all ESEM experiments reported here, we used a FEI XL-30 field emission gun instrument with a Peltier sample stage and gaseous secondary electron detector (GSED) to produce an image. All samples were contained within a circular brass stub with a diameter of 1 cm and a depth of 0.5 cm. The temperature of the stub, and hence sample, was controlled using the Peltier cooling stage with a water/propanol coolant maintained at 283 K. A few drops of the undiluted sample were placed on the stub and left for a few minutes to equilibrate, before initiating pump down of the chamber from ambient to a few Torr. Images from the CCD camera confirmed that the sample remained stable; i.e., there was no “boiling” of water during this pump-down procedure as the method adopted (pumping down in steps) ensured that a high percentage of water vapor remained in the sample chamber/environment.43 In all cases, the sample was cooled to ∼275 K (2 °C) and surrounded by water vapor at a pressure of ca. 4 Torr. Working distances were typically 8-10 mm, and images were collected using an accelerating voltage of 5 keV. Under such conditions signal-to-noise ratio was maximized, sample dehydration minimized, and detailed high-resolution images were obtained. Rod dimensions and their separation were subsequently determined directly from the images using computer software built into the ESEM. It should be noted that these samples were highly susceptible to beam damage, especially at high magnification. Beam damage is a ubiquitous problem for all electron microscopists working with organic materials but has been found to be more prevalent in the environmental SEM.47 When the electrons interact with water molecules, they can produce a small number of highly mobile free radicals (hydroxide and hydrogen peroxide) through the process of radiolysis48 which subsequently attack, propagate, and degrade the sample. Damage is clearly visible as a blistering effect on the micrometer length scale where smoothing and distortion of the cross-sectional sample surface and perimeter shape appears (for examples of beam damage the reader is referred to work by Royall et al.48 and references therein). Beam damage was minimized here, however, by optimizing chamber conditions and minimizing the time spent at high magnification. If damage did occur. then either a fresh area or a new sample was examined. Results and Discussion Rod Dimensions. Figure 3 shows a typical AFM micrograph of a thin film cast from a 0.1 wt % microcrystalline cellulose solution. Such dilute solutions were necessary to image individual rods. The rods observed are bundles of microcrystalline chains that self-associate along their long axes through electrostatic interactions. Despite attempts at size fractionation, the polydispersity of the rods remained relatively high: ∼150-210 nm in length and ∼5-11 nm wide. For the remainder of the discussion the dimensions were averaged and taken at fixed values of 180 nm long and 8 nm wide, giving an axial ratio of 22. Pre` cis of Phase Behavior. A complete description of the phase behavior of our cellulose suspensions has been presented,46 and discussed in relation to the extensive work



Figure 3. Atomic force micrograph of a dried 0.1 wt % sample of microcrystalline cellulose suspension.

by Gray and co-workers,13-17 in an earlier publication. Only an outline of the results will be presented here to aid interpretation of the liquid crystalline behavior. A summary of phase behavior for a series of suspension concentrations (1-15 wt %) of constant ionic strength (pH 4) is given in Figure 4a. Above the first critical concentration, c*, of 4 wt % a biphasic suspension formed with a sharp phase boundary, the lower and upper being anisotropic and isotropic, respectively. The volume fraction of the anisotropic phase was found to increase linearly with total suspension concentration, in agreement with previous experiments14,46 and theoretical predictions24-27 (and references therein). Above the second critical concentration, c**, only the anisotropic phase formed. Phase separation, interfacial tension, and hence packing of the lyotropic phase are predicted to be highly dependent on the concentration of the individual coexisting phases.24-29 Weight percent concentrations for each phase were determined, therefore, for all coexisting phases and results, given in Figure 4b, reveal no significant trend in either phase. This was expected from the linear increase in anisotropic volume fraction across the biphasic region in the phase diagram. Such behavior, and the difference between the coexisting phases, have been discussed previously.46 It is important to note, however, that the concentration, hence molecular packing, of the anisotropic phase increases linearly above the upper critical concentration. Microscopy. A range of concentrations of both isotropic and anisotropic phases were examined using both optical and environmental scanning electron microscopy. The advantage of both techniques is that samples do not require dilution or coating; hence the process of microcrystalline self-assembly can be investigated as a function of concentration on both the meso- and nanoscopic length scale. Optically, the isotropic phase displayed no birefringence, as expected. By use of ESEM, however, individual microcrystalline aggregates were visible. This is the first time, to our knowledge, that cellulose organization could be elucidated from isotropic phases in their natural state, i.e., without any sample dilution or preparation that is usually required for conventional SEM or TEM work. A typical ESEM image is given in Figure 5a, and rod dimensions were calculated

Figure 4. (a) Phase diagram of microcrystalline suspension where water medium is maintained at a constant pH of 4.0. The lower (c*) and upper (c**) phase boundaries are represented by the solid lines. (b) Individual concentrations of isotropic and anisotropic phases as a function of total suspension concentration for a constant pH of 4.0.

Figure 5. ESEM micrograph of isotropic phase of an 3 wt % aqueous cellulose suspension where the chamber pressure and sample temperature were (a) 4.2 Torr and 275 K and (b) 2.5 Torr and 275 K, respectively.

514


Miller and Donald

Figure 6. Optical micrograph of anisotropic phase (nematic domains) of a 16 total wt % aqueous cellulose suspension under crossed polars after 10 min. The scale bar represents 10 µm.

Figure 7. Optical micrograph of anisotropic phase (chiral nematic) of a 16 total wt % aqueous cellulose suspension under crossed polars after 1 day. The scale bar represents 100 µm.

using computer software. It is evident that the average aggregate size was ca. 0.6-1.0 µm in length and 0.1-0.2 µm wide and they have enough space and freedom to assume random orientation and, hence, are lying with their long axes parallel to the surface to minimize surface tension.46 These dimensions are significantly higher than those extracted from AFM studies, presumably due to rod self-association and/or swelling. It is generally accepted that although crystalline cellulose I does not swell in neutral water, it can in acidic (or basic) media. Our suspensions are kept constant at pH 4; hence fiber swelling is expected and, furthermore, will be encouraged by the added repulsion present due to the electrical double layer effect. To investigate the degree of swelling (and also any aggregation of rods upon drying), controlled dehydration experiments were performed where water was gradually pulled away from our sample by decreasing the pressure in the sample chamber stepwise from 4 down to 2.5 Torr (at constant temperature). The dry film imaged (Figure 5b) revealed that the average aggregate diameter decreased by ca. one-third indicating the individual rods were highly swollen in the presence of pH 4 water. The dimensions of the dry fibers from Figure 5b, ∼400600 nm in length and ∼60-100 nm wide, were still considerably higher than those dimensions obtained from AFM work (ca. 180 and 10 nm in length and width, respectively). Such results suggest that self-association of rods was either present already in the isotropic phase where rods self-assembled along a single axial direction (these aggregates will be referred to as fibers) or the drying process induced bonding and aggregation. The anisotropic phase was also characterized using both optical and electron microscopy. Initial polarizing microscopy studies revealed that all concentrations of anisotropic phase spontaneously formed birefringent domains with a distinct extinction direction. Typical domains, shown in Figure 6, were all rectangular but of varying size, ranging from 24 to 32 µm in length and 10-12 µm wide. A preferred directional orientation was visible within each domain, and as the polarizer and the analyzer were rotated by 45° some of the birefringence turned from bright to dark or dark to bright, indicating the molecular axes, and hence axes of the cellulose rods, were parallel to the preferred orientation. With time they reorganized and subsequently merged, to produce

continuous, fingerprint textures, indicative of a chiral nematic phase (see Figure 7). The time for the transition into fingerprint textures was concentration dependent and ranged from several hours to 1 day, for the lowest and highest concentrations, respectively. This equilibration time dependence is attributed to the variation of viscosity influencing rod diffusion and mobility: higher concentration, and hence higher viscosity, will reduce diffusion and mobility therefore longer equilibration times will be expected. Viewing the rectangular domains using ESEM was relatively straightforward as a few drops of sample were placed on the sample stub and imaged directly, after careful pump down of the chamber. Typical images are given in Figure 8 and at low magnification are comparable to those obtained from optical microscopy. The advantage of ESEM becomes apparent when we attempt to image inside the domains as we can see the alignment of the individual rods (An example is given in Figure 8c where the rods appear bright within the dark water media.) It was evident that all rods align with their long axes parallel to the water surface and pack with nematic order, i.e., they have long-range orientational order but only short-range positional order. The rods were highly swollen in the fully hydrated state (diameter was ca. 50 nm) and the inter-rod distances were uniform and estimated to be 55 nm. It was found that within the biphasic region the anisotropic concentration did not vary, and hence no influence on inter-rod distance was observed. This distance between the individual domains did, however, decrease in proportion to the increase in anisotropic phase concentration in the single-phase region (i.e., above c**). Imaging the chiral nematic textures in the ESEM proved more testing as the domains did not merge spontaneously in the sample chamber. Problems arose due to the difficulty of maintaining a fully hydrated sample in the chamber over a period of several hours, and furthermore the low temperature of 275 K possibly led to an increase in suspension viscosity and, hence, to a reduction in mobility of the rods and domains. Such difficulties were overcome by preparing a sample outside the ESEM chamber by storing it overnight (or longer if necessary) in a saturated water environment at room temperature, before inserting it carefully into the ESEM for subsequent analysis. With this procedure, characteristics


Figure 8. ESEM micrographs of anisotropic phase (nematic domains) of an 18 total wt % aqueous cellulose suspension placed directly into sample chamber where the scale bar represents (a) 50 µm, (b) 10 µm, and (c) 1 µm. The cellulose and water media appear bright and dark, respectively. The ESEM chamber pressure and sample temperature were 4.2 Torr and 275 K, respectively.

of helicoidally oriented chiral nematic phases, reminiscent of the fingerprint type patterns observed using optical microscopy, were observed. Typical ESEM micrographs obtained are given in Figure 9 for a total suspension concentration of 18 wt %. To confirm the ESEM images relate exactly to the optical micrographs, the chiral nematic pitch was calculated, and compared, as a function of concentration (in the single anisotropic phase region). The pitch is a measure of cellulose interactions (smaller pitch, stronger interaction) and is determined experimentally by measuring (and doubling) the distance between the fingerprints of the chiral nematic texture (see Figure 1 for a schematic illustration of organization and Figure 9 for an example of experimental measurement). Results are summarized in Figure 10 where each data point is an average of 10 separate measurements. It is evident that both imaging techniques gave comparable results, within experimental error, suggesting that the polarized light and ESEM micrographs are indeed equivalent. Moreover the chiral nematic pitch of the anisotropic phase decreased (phase becoming more highly twisted) from ∼70 to 5 µm, with increasing suspension concentration, from 12 to 18 total wt %. This is


Figure 9. ESEM micrographs of anisotropic phase (chiral nematic) of an 18 wt % aqueous cellulose suspension after sitting in saturated environment for 1 day where the scale bar represents (a) 50 µm, (b) 10 µm, and (c) 500 nm. The cellulose and water media appear bright and dark, respectively. The ESEM chamber pressure and sample temperature were 4.0 Torr (for a and b), 2.8 Torr for c, and 275 K, respectively.

usually the case for lyotropic systems49 and has in previous work been attributed to the compression of the packing of the molecules.14,46 We have been able to prove this postulate by studying high-resolution ESEM images and measuring any change in the distance between the individual rods. This proved extremely difficult when the rods were fully hydrated as samples were susceptible to beam damage. Dehydrating the sample reduces sample degradation (due to the formation of fewer radicals) however, and is simple to do in the ESEM by reducing chamber pressure, stepwise, from 4.2 to 2.5 Torr, in a controlled manner. Under such conditions a slight reduction in rod swelling was inevitable, but dimensions were used as a first approximation for comparative purposes. Results are plotted as a function of concentration in Figure 11 and indeed a decrease is observed, from 45 to 25 nm for a concentration increase of 12 to 18 wt %, confirming increasing packing density reduces the pitch of the anisotropic phase. Such a trend perhaps seems unsurprising to the reader, but not only have we proven it unequivocally as

516


Figure 10. Effect of suspension concentration on chiral nematic pitch of the anisotropic phase: a comparison of polarizing (0) and ESEM (×) results.

Figure 11. Effect of suspension concentration on interfibril distance in the chiral nematic phase.

opposed to indirect inferences, in doing so we have demonstrated the versatility and power of the ESEM technique. We still need to explain why we see liquid crystalline textures in the ESEM, i.e., what gives rise to the contrast? Contrast in the ESEM is an ongoing area of research but to date is still not completely understood.50-52 When the primary electron (PE) beam impinges on a sample surface, it can produce a variety of signals including secondary electrons (SEs), backscattered electrons (BSEs), characteristic X-rays, and cathodoluminescence which are used to characterize the sample. Using the GSED detector, we will have primarily collected SE signals as these are low in energy in comparison to the primary or backscattered electrons, and hence they are more readily amplified via the gaseous cascade amplification process and accelerated toward the positively biased detector.53 The amount of SE emission from the sample surface is influenced by several factors, including the surface potential barrier, SE escape depth, transport to the surface, and the number of SEs generated per unit volume.54,55 The generation rate is governed by the efficiency of primary beam interaction with the material; i.e., the higher the PE energy loss rate, the greater the number of SEs produced. After generation, excited electrons need to diffuse through the

Miller and Donald

Figure 12. (a) Representation of a chiral nematic texture where the length of the nails drawn represent the projected length of the molecule onto the page, and the head of the nail represents the end point coming out of the page. (b) Schematic of SE signal intensity and interaction volume from undulating surface.

sample, during which inelastic collisions occur, and still have enough energy left to overcome the surface barrier. Since SEs by definition have small energies (e∼10 eV), only the SEs generated at the near surface (tens of nanometers in dielectrics) can escape and contribute to the final image. It is easy to envisage, therefore, the important influence chemical composition, density, and electronic and dielectric properties of the phases present at the sample surface has on SE generation.54,55 In our isotropic and nematic cellulose suspensions, we speculate that variations in material density, and hence electronic structure, are the source of contrast causing cellulose rods to appear brighter than the surroundings. The crystalline cellulose rods will have a higher mass density than water, resulting in a higher secondary yield via two possible, synergistic mechanisms where the increased molecular density will (a) increase the PE loss of energy leading to higher generation rates of SE's and (b) increase the band gap of the material54 and the SE path length, thus facilitating SE escape. Such variations will also be present in the chiral nematic phase. The chiral nematic phase consists of numerous nematic planes, which have been twisted periodically about an axis perpendicular to the director (see Figure 1). If the planes are perpendicular to the surface, then their directors will rotate around an axes parallel to the surface (see Figure 12a for clarification). Such organization will lead to variation in mass and electronic density across the surface as the rods with their long axes parallel to the surface will have a higher packing density than those normal to the surface. These changes in director orientation could affect electron affinity, thus influencing the surface energy barrier; i.e., it is possible when the director is parallel with the liquid surface electron affinity will increase, thus reducing the surface energy, although at this stage this is only speculation. The opposite scenario would occur when the director is normal to the surface, inducing an increase in surface energy. It is conceivable that this will lead to contrast variation where domains with the director parallel and perpendicular to the surface will appear as bright and dark, respectively. This analogy will also account for the additional detail observed “between” the fingerprint lines, as there will be a gradual decrease in contrast as the director rotates through 180° (half pitch length) to become parallel again with the surface.


An alternative or concurrent source of contrast in the chiral nematic state is surface topography54,55 where the surface is not macroscopically flat but undulating sinusoidally due to the fluid nature of the phase. Such possible spatial variation is shown schematically in Figure 12b and could be induced by changes in molecular packing as the director rotates systematically at the surface in synchrony with the chiral nematic pitch. Such surface undulations would be expected54,55 to lead to reduced effective emission from the “troughs” in comparison with the “peaks”, as fewer electrons from the “troughs” will enter the cascade process in the gas. To our knowledge this is the first time that lyotropic liquid crystalline textures in their natural state have been observed using electron microscopy. The contrast in the images is obvious, although the details of the mechanisms giving rise to the contrast are not as yet unambiguously established. Understanding this in more detail will form the focus of future work. Nevertheless these results indicate the power of the ESEM for such liquid crystalline dispersions, and it would be expected that other similar systems would also prove amenable to such studies to explore local packing in both nematic and chiral nematic phases. Conclusions We have shown that ESEM can be used successfully to image the structure of both isotropic and anisotropic phases of microcrystalline cellulose suspensions. Results from ESEM have been verified by comparing with those obtained from the more conventional technique of polarizing microscopy. As expected microcrystalline rods adopt random orientation in the isotropic phase but form ordered phases in the more dense anisotropic phase. In this liquid crystalline phase nematic domains, over time, merged to form fingerprint patterns indicative of chiral nematic textures. Furthermore we have shown that the pitch of the liquid crystalline phase decreases due to close packing of the rods as the suspension concentration increases. We thus have demonstrated the ability of ESEM to observe lyotropic textures in their natural environment and the advantages offered by its operation at much higher resolution than optical methods. Furthermore ESEM offers the chance to observe directly structures on the nanometer scale that could previously be investigated only by nonimaging methods such as scattering or inferred from optical textures. Acknowledgment. We gratefully acknowledge stimulating discussions with Drs. John Craven and Milos Toth. This work was funded by the EPSRC (Grant Number GR/N14484). References and Notes (1) Poon, W.; McLeish, T. C. B.; Donald, A. M. Phys. World 2001, May, 33-38. (2) Stokes, D. J.; Thiel, B. L.; Donald, A. M. Langmuir 1998, 14, 44024408. (3) Miller, A. F.; Donald, A. M.; Dobson, C. M.; MacPhee, C. E. In preparation. (4) Miller, A. F.; Cooper, S. J. Langmuir 2002, 18, 1310-1317. (5) Uwins, P. J. R. Mater. Forum 1994, 18, 51-75. (6) Little, B.; Wagner, P.; Ray, R.; Pope, R.; Scheetz, R. J. Ind. Microbiol. 1991 8, 213-222. (7) D’Emanuele, A.; Gilpin, C. Scanning 1996, 18, 552-527.

Biomacromolecules, Vol. 4, No. 3, 2003 517 (8) Oster, G. J. Gen. Physiol. 1950, 33, 445-449. (9) Livolant, F.; Leforestier, A. Prog. Polym. Sci. 1996, 21, 1115-1164. (10) Revol, J. F.; Marchessault, R. H. Int. J. Biol. Macromol. 1993, 15, 329-335. (11) Giraud-Guille, M.-M Biol. Cell. 1981, 67, 97-99. (12) Borgstrom, J.; Quist, P.-O.; Piculell, L. Macromolecules 1996, 29, 5926-5933. (13) Revol, J. F.; Bradford, H.; Giasson, J.; Marchessault, R. H.; Gray, D. G. Int. J. Biol. Macromol. 1992, 14, 170-172. (14) Dong, X. M.; Kimura, T.; Revol, J. F.; Gray, D. G. Langmuir 1996, 12, 2076-2082. (15) Revol, J. F.; Godbout, L.; Gray, D. G. J. Pulp Pap. Sci. 1998, 24 (5), 146-149. (16) Revol, J. F.; Godbout, L.; Dong, X. M.; Gray, D. G. Liq. Cryst. 1994, 16, 127-134. (17) Dong, X. M.; Revol, J. F.; Gray, D. G. Cellulose 1998, 5, 19-32. (18) Araki, J.; Wada, M.; Kuga, S.; Okano, T. Langmuir 2000, 16, 24132415. (19) Furuta, T.; Yamahara, E.; Konishi, T.; Ise, N. Macromolecules 1996, 29, 8994-8995. (20) Revol, J.-F.; Orts, W. J.; Godbout, L.; Marchessault, R. H. Polym. Mater. Sci. Eng. 1994, 71, 334-339. (21) Orts, W. J.; Godbout, L.; Marchessault, R. H.; Revol, J.-F. Macromolecules 1998, 31, 5717-5725. (22) Roland, J. C.; Reis, D.; Vian, B. Tissue Cell 1992, 24, 335-345. (23) Lonnberg, B. Industrial Cellulose. In Regenerated Cellulose Fibres; Woodings, C. R., Ed.; Woodhead Publishing Ltd.: Cambridge, 2001; Chapter 2. Johnson, T. F. N. Current and Future Market Trends. In Regenerated Cellulose Fibres; Woodings, C. R., Ed.; Woodhead Publishing Ltd.: Cambridge, 2001; Chapter 10. (24) Onsager, L. Ann. N. Y. Acad. Sci. 1949, 51, 627-659. (25) Doi, M.; Kuzuu, K. J. Appl. Polym. Sci., Appl. Polym. Symp. 1984, 41, 65-68. (26) McMullen, W. E. Phys. ReV. A 1988, 38 (12), 6384-6395. (27) Chen, Z. Y.; Noolandi, J. Phys. ReV. A 1992, 45 (4), 2389-2392. (28) Chen, W.; Sato, T. Macromolecules 1998, 31, 6506-10. (29) Koch, D. L.; Harlen, O. G. Macromolecules 1999, 32, 219-226. (30) Peemoeller, H,; Sharp, A. R. Polymer 1985, 26, 859-864. (31) Pothan, L. A.; Bellman, C.; Kailas, L.; Thomas, S. J. Adhes. Sci. Technol. 2002, 16, 157-178. (32) Revol, J.-F. Carbohydr. Polym. 1982, 2, 123-127. (33) Sugiyama, J.; Harada, H.; Fujiyoshi, Y.; Uyeda, N. Mokuzai Gakkaishi 1984, 30, 98-101. (34) Revol, J.-F. J. Mater. Sci. Lett. 1985, 5, 1347-1349. (35) Hanley, S. J.; Giasson, J.; Revol, J.-F.; Gray, D. G. Polymer 1992, 33, 4640-4642. (36) Hanley, S. J.; Gray, D. G. J. Pulp Pap. Sci. 1999, 25, 196-200. (37) Danilatos, G. D. AdV. Electron. Electron Phys. 1988, 71, 109-250. (38) Danilatos, G. D. J. Microsc. 1990, 162, 391-402. (39) Danilatos, G. D. Scanning 1981, 4, 9-20 (40) Danilatos, G. D. Scanning 1985, 7, 26-42. (41) Donald, A. M. Curr. Opin. Colloid Interface Sci. 1998, 3, 143147. (42) Stokes, D. J. AdV. Eng. Mater. 2001, 3, 126-130. (43) Cameron, R. E.; Donald, A. M. J. Microsc. 1994, 174, 227-237. (44) Thiel, B. L.; Bache, I. C.; Fletcher, A. L.; Meredith, P.; Donald, A. M. J. Microsc. 1997, 187, 143-157. (45) Fletcher, A. L.; Thiel, B. L.; Donald, A. M. J. Phys. D: Appl. Phys. 1997, 30, 2249-2257. (46) Miller, A. F.; Donald, A. M. Langmuir 2002, 18, 10155-10162. (47) Kitching, S.; Donald, A. M. J. Microsc. 1998, 190, 357-365. (48) Royall, C. P.; Thiel, B. L.; Donald, A. M. J. Microsc. 2001, 204, 185-195. (49) Guo, J. X.; Gray, D. G. Cellulosic Polymers, Blends and Composites; Gibert, R. D., Ed.; Hanser: New York, 1994; pp 25-45. (50) Thiel, B. L.; Donald, A. M. J. Microsc. (Oxford) 1991, 164, 107126. (51) Toth, M.; Thiel, B. T.; Donald, A. M. J. Microsc. 2002, 205, 8695. (52) Toth, M.; Phillips, M. R.; Craven, J. P.; Thiel, B. T.; Donald, A. M. J. Appl. Phys. 2002, 91, 4492-4499. (53) Toth, M.; Phillips, M. R.; Thiel, B. L.; Donald, A. M. J. Appl. Phys. 2002, 91, 4479-4491. (54) Goldstein, J. I. Scanning Electron Microscopy and X-ray Microanalysis; New York: Plenum Press: 1992. (55) Reimer, L. Scanning Electron Microscopy; Berlin: Springer, 1985.

BM0200837

Imaging of Anisotropic Cellulose Suspensions Using Environmental

Recommend Documents