PERGAMON

Micron 31 (2000) 659667 www.elsevier.com/locate/micron

Atomic force microscopy analysis of wool bre surfaces in air and under water
J.A.A. Crossley a, C.T. Gibson b, L.D. Mapledoram a, M.G. Huson c, S. Myhra a,b,*, D.K. Pham c, C.J. Soeld a, P.S. Turner c, G.S. Watson b
AEA Technology plc., Analytical Services, 551 Harwell, Didcot, Oxon OX11 0RA, UK b School of Science, Grifth University, Nathan, Queensland 4111, Australia c CSIRODivision of Wool Technology, P.O. Box 21, Geelong, Victoria 3216, Australia Received 10 February 1998; accepted 2 November 1998
a

Abstract Wool bre surfaces have been treated by solvent cleaning which leaves the native covalently bound surface lipid layer intact, and by alcoholic alkali which removes the lipid layer. The resultant surfaces have been analysed by atomic force microscopy (AFM), with particular emphasis on forcedistance (Fd) methods. Methodologies were developed for investigation in situ in water of both the surface topography and the characteristics of the lipid layer. Longitudinal surface texturing was resolved in images of wool bre surfaces in air; the texturing remained prominent after exposure to water. High resolution Fd curves revealed features associated with the lipid layer. A simple formalism was used to show that the layer had a thickness of a few nm, and an effective stiffness of some 0:12 ^ 0:01 N=m: Strong adhesive interactions, equivalent to a pressure of 0.1 MPa, acted on the tip at the tip-to-substrate interface. The methodology and formalism are likely to be relevant in the broad eld of thin-lm analysis and for bre technology. 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Wool bre; surface structure; Lipid layer; Atomic force microscopy; Forcedistance analysis; Thin-lm analysis

1. Introduction Wool bre consists of elongated cortical cells surrounded by overlapping cuticle cells. The outer layer of the cuticle cells is a surface membrane 57 nm thick, commonly referred to as the epicuticle (Bradbury, 1973). The wool bre surface remains hydrophobic even after repeated solvent extraction. Lindberg (1953) reported that the hydrophobic surface was modied under alcoholic alkali conditions. Kopke and Nilssen (1960) suggested that the effect of alkali on wool could be explained if long chain fatty acids were attached by ester linkages to the wool bre surface. Leeder and Rippon (1985) treated wool with anhydrous potassium tertiary butoxide in the non-swelling solvent tertiary butanol, which was assumed to conne reaction to the bre surface. They observed a dramatic reduction in the hydrophobicity of the bre, which was attributed to the removal of the postulated lipid layer from the bre surface. Evans et al. (1985) showed that the major lipid component removed by alcoholic alkali treatment of wool was a
* Corresponding author. Tel.: 617-3875-7546; fax: 7656. E-mail address: s.myhra@sct.gu.edu.au (S. Myhra) 0968-4328/00/$ - see front matter PII: S0968-432 8(99)00076-1 617-3875-

methyl-branched 21-carbon fatty acid and suggested that it was bound to the bre surface by an ester or thioester bond, as originally proposed by Kopke and Nilssen (1960). Subsequent studies (Wertz and Downing, 1988; Negri et al., 1991) identied the fatty acid conclusively as 18-methyleicosanoic acid. The model for the wool bre surface proposed by Negri et al. (1993) consists of a protein matrix heavily acylated with 18-methyleicosanoic acid, subsequently described as the lipid layer, which orients away from the bre and forms a hydrophobic surface. The covalently bound lipid layer has a critical role in determining surface properties of the wool bre, which in turn are important in many aspects of wool processing. It is often necessary to modify the bre surface, chemically or physically, to achieve the required performance (see e.g. Millard et al., 1972; Makinson, 1979; Leeder and Rippon, 1985; Willis, 1993). Monitoring of alterations of bre surface as a result of a particular treatment or process is important for the understanding and improvement of the relevant technology. Atomic force microscopy (AFM) as a technique for characterisation of surfaces and interfaces has now reached a state of relative maturity. The underlying phenomenology of the technique and many of the methodologies which have

2000 Elsevier Science Ltd. All rights reserved.

660 Table 1 Characteristics of probes Supplier Designation

J.A.A. Crossley et al. / Micron 31 (2000) 659667

Tip parameters RTip (nm) a 40 40 1020 50 50 50

Ar b 0.7 0.7 3 0.7 0.7 0.7

L (mm) c 200 200 180 220 140 85

Lever parameters kN (N/m) d 0.03 0.06 0.3 0.03 0.1 0.5

TopoMetrix TopoMetrix Park Park Park Park

a b

pyramidal pyramidal Ultralever Microlever Microlever Microlever

Rtip species the nominal radius of curvature of the apex of the tip. Ar species the nominal aspect ratio of the tip. c L is the length of the lever from the base line of the V-shape to the tip. d kN is the nominal spring constant for normal deection of the lever.

been developed for dealing with particular systems have been described in the literature (e.g. see Guntherodt et al. (1995), and references therein). The impact of AFM in the eld of wool technology, or indeed in the broader area dealing with textile bres, has so far been relatively modest (Phillips, 1995; Phillips et al., 1995; Klausen et al., 1995). However, a great deal of work has been carried out on systems that are related, e.g. (Tao et al., 1992; Xu et al., 1994; Hoh and Schoenenberger, 1994). AFM has a number of attractive attributes relevant to characterisation of wool bre surfaces. While its lateral spatial resolution is comparable to that of eld-emission scanning electron microscopy (FeSEM), the z-resolution is greatly superior (sub-nm), and topographical excursions in the z-direction can be quantied over a depth of eld which is equal to the dynamic range of the instrumental z-range (usually 510 mm). Moreover, this performance is obtainable in situ in a uid environment so that the topographical evolution of a bre/uid interface can be tracked continuously in real time and space. The AFM can be operated in the so-called force-versus-distance (Fd) mode whereby the approaching/retracting tip senses interactions at the interface. Thus the nano-chemical (adhesive interactions) and nano-mechanical properties of surfaces and interfaces can be mapped as functions of environmental variables, duration and as functions of surface treatment. The objectives of the present project were to develop appropriate methodologies and to apply these to topical questions concerned with treated wool bre. In particular, in this paper are described methodological advances for investigation of the bre surface in situ in water and the outcomes of Fd analysis of the characteristics of the covalently bound lipid layer.

2. Materials and methods 2.1. Specimens Fine 19.8 mm Merino wool bres were used. The outer one-third of the raw bre was removed to avoid effects of

weathering. One type of characteristic bre surface was obtained from cleaning by Soxhlet extraction in a 2:1 mixture of chloroform/methanol for 5 h to remove wool wax, dried and then rinsed thoroughly in distilled water to remove reagents, inorganic and proteinaceous materials. This cleaning method ensured the removal of all unbound lipids and other contaminants from the bre surface (Ward et al., 1993; Brack et al., 1996). Another type of characteristic bre surface was prepared by further treating the above bre with 0.1 M KOH in ethanol for 30 min at room temperature. Previous study (Brack et al., 1996) has shown that this treatment removes the bound lipid layer from the bre surface. The samples were then rinsed in distilled water, dried and a second Soxhlet solvent extraction was performed in order to remove residual alkali and any other by-products of the treatment. The bres, being relatively fragile and crimped, are difcult to mount on a suitable substrate for AFM analysis. The following procedure was adopted: a half-round crosssection, 3 mm diameter and of length 6 mm, was prepared from mild steel (for easy anchorage to the magnetic scanner/ stage stub). The metal slug was then covered with nail varnish, so as to avoid corrosion, and wrapped with 23 layers of paralm a laboratory sealing lm (obtained from the Whatman Co.). The lm provided a sufciently compliant surface for partial embedding of the wool bre and allowed some accommodation for the crimped sections. A single wool bre was then wound under suitable tension around the cross-section. The slug was mounted so that the AFM probe was oriented perpendicularly to the long axis of the slug, in order to facilitate easy access to any of the bre sections. Optical magnication by 100 or 200 was adequate to locate a suitable section and to position the end of the probe in the general vicinity of a section at, or close to, the highest point of the rounded section. The procedure was adopted in order to prevent the lever (rather than the tip) interacting directly with the surface, and to preclude the tip landing on a steep slope where unwanted bending modes might be generated in the probe.

J.A.A. Crossley et al. / Micron 31 (2000) 659667

661

Fig. 1. FeSEM image of a scoured wool bre showing the cuticle structure.

2.2. AFM instrumentation and methodologies A TopoMetrix TMX-2000 Discoverer series instrument was employed at Grifth University, while an East Coast Scientic (ECS) instrument provided results at AEA Technology. The respective capabilities relevant to the present study were broadly comparable, both being based on the optical lever detection system, while some of the specic facilities were complementary. The analyses carried out on the TMX instrument utilised a scanner with a 70 70 mm2 eld of view, and the in situ analysis was performed in an open cell dened by a transparent window and the specimen substrate. The ECS instrument was tted with a 60 60 mm2 scanner, and the in situ analysis was facilitated by a transparent glass thimble specimen stage for easier access and more convenient control of ambient conditions. Most of the analyses were carried out in distilled and deionized water (DDW). In some cases the ionic conductivity of the aqueous environment was increased by the addition of NaCl to solution in order to suppress possible effects of surface charging (Emerson and Cox, 1994). Some 100 images were obtained, and several hundred Fd analyses were carried out during the conduct of the experiments. The runs were carried out in two separate laboratories over a period of more than six months. Several different V-shaped probes were used during the project. The relevant characteristics are summarized in Table 1. The two most signicant parameters are the normal

spring constant, kN, which determines the perpendicular force applied by the lever per unit deection, and the radius of curvature of the tip, Rtip, which affects the extent of indentation and thus the force per unit area at the point of contact. Cantilevers from the same wafer were calibrated by a method described in the literature (Gibson et al., 1996), and their spring constants were found to be in good agreement with the generic values reported in Table 1. All AFM images were acquired in the constant net repulsive force mode. The resolution was typically 400 400 points per image 512 512 for the ECS instrument), the offset forces corresponded to some 10 nm deection and were in the range 0.35 nN depending on the force constant of the lever. The fast scan frequency was 23 Hz. Images have been attened by tting a curved (polynomial) plane to the background. Most of the Fd curves were obtained using standard parameters, but the high-resolution curves obtained during thin-lm analysis of the lipid layer used a z-resolution of less than 0.1 nm between sampling points. The position-sensitive diode detector was calibrated by acquiring Fd curves on a non-compliant surface (i.e. glass). Operational and calibrative methods have been described in a recent review (Gibson et al., 1997).

3. Results and discussion 3.1. Surface structure The general structural features of a solvent cleaned wool bre surface are illustrated by the FeSEM image in Fig. 1. The characteristic cuticles are resolved clearly. An AFM contact mode image of a 20 20 mm2 eld of view (Fig. 2) shows a cuticle and apparently featureless surfaces in its vicinity. A number of higher resolution images were obtained in air and in water. The examples in Fig. 3 illustrate longitudinal texturing of a solvent cleaned bre surface (a) in air and (b) in water. A higher resolution image obtained in situ after brief exposure to water is shown in Fig. 3c. The extent of corrugation is shown in the line prole (c) drawn perpendicularly to the direction of texturing. Thus it can be seen that the transverse lateral spacing of the features is some 500 nm while the vertical amplitude of corrugation is ca 20 nm; other images revealed lateral spacings down to less than 200 nm. Similar structures were observed on alkali treated bre surfaces. It was observed reproducibly that the image quality deteriorated with extended exposure to water, as shown in Fig. 4. The features were smeared out and there was streaking and instabilities in the image. The observation of longitudinal texturing on wool bre surface in air and in water is in general agreement with a previous AFM study (Phillips, 1995). The longitudinal striations were originally observed by SEM, and have subsequently been investigated extensively (Watson, 1997). The AFM analysis shows that the striations remain prominent after exposure to water, rather than being due to dehydration

Fig. 2. An AFM contact mode image showing a cuticle and an apparent featureless surface. The image was obtained in air with a TopoMetrix probe having kN 0:06 N=m:

662

J.A.A. Crossley et al. / Micron 31 (2000) 659667

Fig. 4. AFM contact mode, kN 0:03 N=m; image of a solvent cleaned surface showing deterioration in image quality (compared to Fig. 3) due to softening after 90 h exposure to DDW.

Fig. 3. AFM contact mode, kN 0:03 N=m; images for solvent cleaned wool surfaces: (a) in air and (b) in water after brief exposure. The higher resolution image in (c) shows details of the surface texturing after a brief exposure to water. The image was levelled by a second order polynomial t, and shading has been applied from the left in order to enhance the contrast. The spacings and heights of surface texturing are shown by the line prole.

the lever (zero force) when the tip is far away from the surface. When the tip is close to the surface it feels strong short-range surface adhesive interactions, generally causing the approach curve to exhibit a small excursion into the net attractive force regime. When the effective force constant of the interaction is greater than that of the lever, the tip snaps into contact with the surface (not readily apparent in Fig. 5 but see for example Fig. 7a). Further travel by the specimen stage in the z-direction has the effect of bringing the tip-tosurface interaction back into the net repulsive force regime; the distance travelled by the stage to obtain zero deection represents compression of the surface due to the adhesive interaction. The steep rise in the approach curve, which follows, reects the increasingly large relative deection of the lever. In the retract curve the adhesive interaction is larger due to the much larger contact area arising from the maximum loading of the tip. The asymmetry in the snap-on and lift-off forces, and the extent to which deformations do not recover elastically, are the major causes of hysteresis in the approachretract cycle. 3.2.2. Curves in air Fd curves of solvent cleaned or alkali treated surfaces in air are dominated by a strong attractive meniscus force of some 2030 nN, ie. one to three orders of magnitude greater than the force applied by the lever (see for example Fig. 5a). Thus compliant surface structures with effective spring constants less than a few N/m will be substantially compressed during analysis in air. The meniscus force scales as the perimeter of wetting and is thus proportional to Rtip, to a rst approximation. However, the applied force 2 per unit area of contact will scale approximately as Rtip; in the case of a parabolically shaped tip the radius of the contact area is given by Rpara Rtip z1=2 where z is the distance of indentation. Accordingly, a sharper tip will not necessarily allow for greater sensitivity to deformable surface structures. Hence, the impetus for investigating compliant structures under water, where the water-to-air meniscus force will be absent. 3.2.3. Curves in water The results of Fd analyses in situ of solvent cleaned and

in the SEM vacuum. The deterioration in image quality with duration of exposure to water reects surface softening; this aspect, as well as the origin of the surface texturing, are currently the subjects of on-going studies by AFM analysis. Present indication is that transient contamination of the tip, by transfer of bre material to, and from, the tip is a contributing cause of the observed imaging instabilities. That hypothesis correlates well with the observation that the image quality improved with increasing force loading (greater offset and/or stiffer lever), which would have the effect of the tip tracing out contours over stiffer, and thus more stable, sub-surface structures. 3.2. Forcedistance analysis 3.2.1. General The general features of the approach and retract Fd curves (see Fig. 5) are as follows: moving from right to left along the x-axis as the sample surface approaches the tip, the horizontal section represents the undeected state of

J.A.A. Crossley et al. / Micron 31 (2000) 659667

663

Fig. 6. Typical high-resolution approach Fd curve, kN 0:03 N=m; for solvent cleaned wool in DDW. Fig. 5. Fd, kN 0:03 N=m; approach and retract curves for (a) solvent cleaned wool in air, (b) solvent cleaned wool in DDW and (c) alkali treated wool in DDW. Note the difference in vertical scales. The curves with the large lift-off feature refer to the retract cycle.

alkali treated wool obtained immediately after exposure to water are shown in Fig. 5b and c. It is apparent that the attractive forces are now much smaller and represent genuine tip-to-surface adhesive interactions. During the early stages of exposure an adhesive snap-on interaction is readily measurable (typically 0.5 nN for a TopoMetrix pyramidal tip). For example, the approach curve in Fig. 5c exhibits a small snap-on feature. It has been found that the effects of increasing duration of exposure to water are to soften the surface and to decrease the adhesive interactions. (Alterations in surface chemistry and surface-mechanical properties with duration of exposure will be discussed in a subsequent report.) 3.2.4. High resolution curves High resolution Fd analyses were carried out in situ on solvent cleaned wool bre surfaces during the early stages of aqueous exposure. A typical outcome is illustrated in Fig. 6. The signicant feature is the rst excursion during approach in the direction of net repulsive interaction, which was followed by an adhesive instability, and then by a normal excursion into the repulsive force regime. Analyses of the alkali treated bre surface showed no such adhesive instability. The absence of this effect in normal resolution curves (Fig. 5b) is most likely due to the decreased sensitivity of the analysis. These results can be explained by considering the system as consisting of three compressible elementsthe AFM lever, the lipid layer and the wool substrateseparated by two interfaces between the tip and lipid layer and between the lipid layer and the substrate. In the context of the present measurements both lever and substrate can accommodate innite deection/compression, while the lipid layer can be thought of as a linearly compressible element of nite thickness. Since the lateral spacing between lipid chains is

likely to be small in comparison with Rtip, it is plausible to treat the lipid layer as a continuous compressible structure, rather than as discrete linear molecular chains subject to bending as well as compression. In a serial multi-element system with each element having a characteristic stiffness, the compression of each element will be inversely proportional to the respective stiffness, for a given total force. During approach the tip may initially sense a small adhesive interaction at the tip-to-lipid interface. Then the three-element system will be compressed with the total combined dimensional change being measured along the horizontal axis and the displacement of the lever being measured along the vertical axis. The slope of the curve will reect the effective stiffness of the combined system. When the lipid layer is near its full compression, the tip senses the adhesive interaction at the lipid-to-substrate interface and is pulled into contact with the substrate. From this point onwards the lipid layer is passive, having effectively been compressed to a rigid layer of negligible thickness, and the system reverts to having only two compressible elements. Similar instabilities have been reported in systems containing hydrated layers (Senden et al., 1994) and CTAB surfactant monolayers (Drummond and Senden, 1994) when the applied pressure was high enough (small tip radius) to disrupt the layers. The behaviour of the alkali treated bre can be described by a system consisting of only two compressible elements, consistent with the expected absence of a lipid layer. The next section will provide a more formal description of the system. 3.2.5. Modelling The present system of three interacting compressible elements can be modelled by three linear springs and two interface interactions. A generic Fd curve is shown in Fig. 7a, while in Fig. 7b is sketched the physical conguration. For simplicity it has been assumed that the tip has a rectangular shape and that the compressed surface will adapt to the shape of the tip. A at tip is used for convenience as the

664

J.A.A. Crossley et al. / Micron 31 (2000) 659667

Fig. 7. Generic approach Fd curve (a) for a model with the physical conguration shown in (b) and represented as a series of springs in (c). The system has three compressible elements in seriesa substrate (ksub), a thin layer/lm (klayer) and a lever (kN). Two attractive interactions are sensed by the tip when it touches the outside surface of the layer/lm (F0), and when the layer is compressed to negligible thickness and the tip is at the layer-to-substrate interface (F1).

model becomes signicantly more complex and unwieldy for any other tip shape, effectively a non-linear spring. However, it is not a bad assumption for a layer of thickness of a few nanometers given that the radius of curvature of the tip is 50 nm. In Fig. 7c are sketched the three equivalent springs, and the points at which the interfaces reside and where interface interactions will come into play. It is assumed that the substrate spring is terminated by a xed point, and also that the thin-lm spring is linearly compliant from its original length, T, until full compression (negligible thickness) and then becomes a rigid element. Using symbols as dened in Fig. 7, and stage displacement, zd, taken from the original point of contact. Other displacements, zi, refer to compression of the relevant spring, and the force applied by the lever, kNzL, is referred to the point at which the lever is not deected. Then zd zsub ksub zsub zlayer zL F0 F1 dztip T 0 1 2

where dztip T is a notation taken to mean that the tip is within the reach of the layer-to-substrate interface. The rst expression, Eq. (1), states that the sum of the displacements of the compressible elements must equal the distance travelled by the sample stage, from the initial point of contact. The second, Eq. (2), reects the fact that the sum total of forces is zero. The two interface interaction forces F0 and F1 will come into play at the original point of contact and when the layer is compressed to zero thickness, respectively. For large displacements of the stage we have that klayer T p ksub zsub ; and the contributions of F1 and F0 will be negligible. Then ksub Dzsub kN DzL 3

where Dzi are relative displacements. It is now straightforward to show, in terms of the slope, Ssub DzL =Dzd ; of the curve when the substrate and the lever are the only two compressible elements, that zsub zd 1 Ssub 4

klayer zlayer

J.A.A. Crossley et al. / Micron 31 (2000) 659667

665

Fig. 8. A typical approach Fd curve for parafn coated sealing lm.

It is more convenient to dene the slope in terms of deections/compressions and stage travel as in Figs. 6 and 8, rather than to absorb a particular lever spring constant into the y-axis variable, as in Figs. 5 and 7 ksub kN Ssub 1 Ssub 5

For smaller total compressions, when the layer is compliant, it is convenient to treat the layer and substrate as a single spring with zeff zlayer zsub and keff klayer ksub =ksub klayer ; and the following relationships are obtained: zeff zd 1 DzL =Dzd zd 1 Slayer Slayer 6 7

analysed. The spring constants of the substrate and the layer were found to be 0:12 ^ 0:01 and 0:13 ^ 0:02 N=m; respectively. The inferred stiffness of the substrate should not be taken too literally; during the present series of experiments it was observed that the wool bre surface (substrate) became increasingly softer with duration of aqueous exposure. Most of the Fd results were obtained after an hour or more of exposure, arising from inevitable delays due to nding a suitable location and then establishing optimum conditions for imaging and Fd analysis. The thickness of the lipid layer, calculated from Eq. (8) and using high-resolution approach curves from the database, ranged from 2.5 to 7.7 nm. The lower value is likely to represent the better estimate; any point of contact on a sloping substrate or in a depression would result in a greater effective thickness. Also, transfer of material from the lipid layer to the tip cannot be ruled out, and would result in the compression of a thicker layer, and thus gives an anomalously high value for T. It has been shown (e.g. Weisenhorn et al., 1993; Radmacher, 1997) that on the assumption of either a parabolical or a cylindrical tip indenting a compliant surface, and assuming that the shape of the indentation conforms to that of the tip, Youngs Modulus, E, can be inferred from, respectively, Fload 4Rpara 1=2 =3E=1 Fload 2Rcyl E=1

m2 Dz1:5

10 11

keff kN Slayer =1

m2 Dz

where Slayer is the slope of the Fd curve when both layer and substrate are compressible. Then, in terms of the two slopes and the stage position, zd(1), at the point when F1 is just turned on, the layer thickness and effective spring constant can be written as T zd 1Ssub Slayer Slayer 8 9

klayer kN Slayer Ssub =Ssub

The extent of validity of thin-lm analysis based on the framework above will clearly depend on the linearity of the compliant elements, the range and strength of the adhesive interactions relative to thickness and stiffness of the lm, and the shape of the tip. Also, if the effective stiffness of the substrate is substantially less than that of the thin lm, then the Fd curve will be more difcult to interpret. Finally, the description above will need to be extended in order to describe multilayer structures. 3.2.6. Physical characteristics of bre surface Subject to uncertainties associated with the simplifying assumptions about the system, and errors associated with measurements of slopes, non-linearities in the z-translation piezo-stage and in the position-sensitive detector, the curves (eight separate analyses) obtained from high-resolution Fd analysis kN 0:03 N=m of solvent cleaned wool were

The load is now the sum of the restoring force of the compressed layer and the adhesive interaction, Fload klayer T F0 ; m is Poissons ratio (typically 0.30.5) and Dz T is the extent of indentation/compression. Most of the high-resolution curves were obtained after aqueous exposures in excess of 2 h, and F0 was generally below the limit of resolution ( ^ 0.02 nN). Using results from the Fd data base, and taking Rpara Rtip ; it was found that E for the lipid layer is in the range 815 MPa, if Eq. (10) is used. A similar estimate, taking the radius of the cylindrical section as the cross-sectional radius when the tip is indented to a depth of T, is arrived at by using Eq. (11). The slight curvature apparent in Fig. 6 suggests that the parabolic approximation is more appropriate, however. The result compares to 1320 MPa for polyurethane (Weisenhorn, 1993) and quoted values of 1.4 and 1 MPa for rubber and gelatin, respectively (Radmacher, 1997). The adhesive interaction with the tip at the layer-tosubstrate interface was found to be 0:3 ^ 0:1 nN for tips with Rtip 50 nm: Given that the lever spring constant was 0.03 N/m, the adhesive interaction gives rise to an indentation of some 10 nm. The contact area at the interface is therefore the spherical section of radius 50 nm and with a height of 10 nm. This converts into an estimated equivalent pressure of some 1 10 4 nN=nm2 (0.1 MPa). Caution must be exercised when using this result since the adhesive

666

J.A.A. Crossley et al. / Micron 31 (2000) 659667

interaction is likely to be dependent on the length of the aqueous exposure. Negri et al. (1993) estimated the thickness of the surface lipid layer as about 2.83.0 nm based on the length of 18methyleicosanoic acid. Ward et al. (1993) used measured atomic ratios from XPS and suggested an estimated thickness of about 0.9 nm for the lipid layer. Subsequently Peet et al. (1995) suggested that the discrepancy could be due to the anhydrous, high vacuum conditions in the XPS. Horr (1996) raised some concerns about the validity of the calculation method adopted by Ward et al. (1993) Another possibility is that the wool fabric used by Ward et al. (1993) had a relatively low surface carbon level compared to that on virgin bre surface (e.g. base section of the bre) (Brack et al. 1996). The low concentration of surface carbon suggests that some bound surface lipid on the fabric may have been removed during processing, and therefore may affect the estimation of the lipid layer thickness. The present method is more direct and has the additional merit that it has been obtained under more realistic ambient conditions. Thus it is pleasing that it lends support to the original estimate of layer thickness by Negri et al. (1993). 3.2.7. Thin lm analysis In principle the analysis above should be relevant for any multilayer material. Somewhat fortuitously the laboratory sealing lm provided a useful check on the methodology. The lm, with a nominal thickness of 110 mm, consists of polyethylene sandwiched between two layers of poly(isobutylene) with short parafn chains attached to the two external surfaces. Fd analyses were carried out under identical conditions to those for the wool surfaces. A typical high-resolution approach curve is shown in Fig. 8. The features are consistent with initial compression of the parafn surface layer followed by an adhesive attractive interaction when the surface layer is fully compressed and the tip is in reach of the interface. Finally there is compression of the stiffer substrate. Using the formalism above it was found that the stiffness of the substrate and parafn layer were 0:065 ^ 0:02 and 0:02 ^ 0:01 N=m; respectively, while the layer thickness was estimated to be 3 ^ 1 nm:

ness. Strong adhesive interactions, equivalent to a pressure of 0.1 MPa, acted on the tip at the tip-to-substrate interface. Organic coatings are ubiquitous for numerous technologies, and are at the core of many industrial products and processes. The present study demonstrates that AFM analysis can offer in addition to excellent spatial resolution for difcult materials, methodologies and protocols for evaluating in situ a number of relevant characteristics. Acknowledgements The work was supported in parts by Australian Woolgrowers through a grant from the International Wool Secretariat (Australia), by a Grifth University/CSIRO Collaborative grant and by the AEA Technology plc. Corporate Research Programme. The FeSEM analysis was ably carried out by John Fairchild. References
Brack, N., Lamb, R., Pham, D., Turner, P., 1996. XPS and SIMS investigation of covalently bound lipid on wool bre surface. Surf. Interface Anal. 24, 704710. Bradbury, J.H., 1973. The structure and chemistry of keratin bres. Adv. Protein Chem. 27, 111211. Drummond, C.J., Senden, T.J., 1994. Examination of the Geometry of longrange tip-sample interaction in atomic force microscopy. Colloids Surf., A 87, 217234. Emerson, L., Cox, G., 1994. Charging artifacts in atomic-force microscopy. Micron 25, 267269. Evans, D.J., Leeder, J.D., Rippon, J.A., Rivett, D.E., 1985. Separation and analysis of surface lipids of the wool cuticle. Proceedings of the Seventh International Wool Textile Research Conference, Tokyo 1, 135142. Gibson, C.T., Watson, G.S., Myhra, S., 1996. Determination of the spring constants of probes for force microscopy/spectroscopy. Nanotechnology 7, 259262. Gibson, C.T., Watson, G.S., Myhra, S., 1997. Scanning force microscopycalibrative procedures for best practice. Scanning 19, 564581. Guntherodt, H.-J., Anselmetti, D., Meyer, E. (Eds.), 1995. Forces in Scanning Probe Methods, NATO-ASI Series E: Applied Sciences, 286. Kluwer Academics, Dordrecht. Hoh, J.H., Schoenenberger, C.-A., 1994. Surface morphology and mechanical properties of MDCK monolayers by atomic force microscopy. J. Cell. Sci. 107, 11051114. Horr, T.J., 1996. Determining the lipid layer thickness on wool bre surfaces using XPS. Textile Res. J. 66, 122. Klausen, T., Thomas, H., Hocker, H., 1995. Inuence of oxygen plasma treatment of the chemical and morphological changes of the wool bre surface, Proceedings of the Ninth International Wool Textile Research Conference, 2 Biella, Italy, pp. 241248. Kopke, V., Nilssen, B., 1960. Wool surface properties and their inuence on dye uptakea microscopical study. J. Textile Inst. 51, 13981413. Leeder, J.D., Rippon, J.A., 1985. Changes induced in the properties of wool by specic epicuticle modication. J. Soc. Dyers Colour 101, 1116. Lindberg, J., 1953. Relationship between various properties of wool bres. Textile Res. J. 23, 6776 see also pp. 225236, 573588. Makinson, K.R., 1979. Shrinkproong of Wool, Marcel Dekker, New York. Millard, M.M., Pavlath, A.E., Lee, K.S., 1972. Continuous low-temperature discharge treatment of wool yarn. A process for making wool yarn oil repellent and shrink resistant. Textile Res. J. 42, 307313. Negri, A.P., Cornell, H.J., Rivett, D.E., 1991. The Nature of Covalently Bound Fatty Acids in Wool Fibres. Aust. J. Agric. Res. 42, 12851292.

4. Conclusions In situ application by AFMhigh spatial resolution imaging in combination with Fd analysison wool bre surfaces can provide information not readily obtainable by other techniques and methodologies. In particular, the native lipid layer under water was found to have a thickness of a few nm, in good agreement with an earlier estimate based on the length of the molecular carbon back bone. An effective stiffness for the layer of some 0:12 ^ 0:01 N=m was inferred from the relevant slopes of the Fd curves. Adoption of a simple model allowed calculation of Youngs modulus for the lipid layer from the effective stiff-

J.A.A. Crossley et al. / Micron 31 (2000) 659667 Negri, A.P., Cornell, H.J., Rivett, D.E., 1993. A model for the surface of keratin bres. Textile Res. J. 63, 109115. Peet, D.J., Wettenhall, R.E.H., Rivett, D.E., 1995. The chemistry of the cuticle surface of keratin bres. Textile Res. J. 65, 5859. Phillips, T.L. Scanning force microscopy of the surface of wool bres, Master of Applied Chemistry thesis, Royal Melbourne Institute of Technology, 1995. Phillips, T.L., Horr, T.J., Huson, M.G., Turner, P.S., 1995. Imaging wool bre surfaces with a scanning force microscope. Textile Res. J. 65, 445453. Radmacher, M., 1997. Measuring the elastic properties of biological samples with the AFM. IEEE Engng. Med. Biol. March/April, 4757. Senden, T.J., Drummond, C.J., Kekicheff, P., 1994. Atomic force microscopy: imaging with electrical double layer interactions. Langmuir 10, 358362. Tao, N.J., Lindsay, S.M., Lees, S., 1992. Measuring the microelastic properties of biological material. Biophys. J. 63, 11651169.

667

Ward, R.J., Willis, H.A., George, G.A., Guise, G.B., Denning, R.J., Evans, D.J., 1993. Surface analysis of wool by X-ray photoelectron spectroscopy and static secondary ion mass spectrometry. Textile Res. J. 63, 362368. Watson, D. CSIRO-Division of Wool Technology, 1997, unpublished data. Weisenhorn, A.L., Khorsandi, M., Kasas, S., Gotzos, V., Butt, H.-J., 1993. Deformation and height anomaly of soft surfaces studied with an AFM. Nanotechnology 4, 106113. Wertz, P.W., Downing, D.T., 1988. Integral lipids of human hair. Lipids 23, 878881. Willis, H.A. Surface analytical studies of shrink-resist treatments of wool, PhD thesis, University of Queensland, 1993. Xu, W., Mulhern, P.J., Blackford, B.L., Jericho, M.H., Templeton, I., 1994. A new atomic force microscopy technique for the measurement of the elastic properties of biological materials. Scanning Microscopy 8, 499 506.