Surface topography of heat-set whey protein gels by confocal laser scanning microscopy

Food Hydrocolloids 20 (2006) 468–474 www.elsevier.com/locate/foodhyd

Surface topography of heat-set whey protein gels by confocal laser scanning microscopy Jianshe Chen*, Thomas Moschakis, Luis A. Pugnaloni1 Procter Department of Food Science, University of Leeds, Leeds LS2 9JT, UK Received 15 October 2004; revised 23 February 2005; accepted 19 April 2005

Abstract Extensive research works have been carried out in investigating the microstructure of heat-set whey protein gels and their fractal nature, but little has been done on the surface studies of these systems due to the lack of suitable technique for surface characterization of wet and deformable food gels. This work intended to explore the possibility of applying confocal laser scanning microscope (CLSM) for surface characterization of such delicate systems. Surfaces of heat-set whey protein gels (14 wt%) have been investigated with and without the presence of salt (0 and 200 mM NaCl). High quality surface images across X–Z plane were obtained. These images were further analysed for their surface roughness, periodic length scale, and fractal nature. The roughness of these surfaces was quantified in terms of root-mean-square surface roughness (Rq) and arithmetic surface roughness (Ra). It was found that the protein gel without salt addition had a very smooth surface with Rq of 0.20 and 0.18 mm, respectively, but the gel containing 200 mM NaCl had a much rougher surface with large Rq and Ra (2.39 and 1.91 mm, respectively). The fractal nature of the surface was also revealed for this gel and a fractal dimension of 1.15 was obtained. q 2005 Elsevier Ltd. All rights reserved. Keywords: Surface texture; Surface roughness; Protein gels; Whey protein; Confocal microscope

1. Introduction Surface properties such as surface roughness and surface wetness strongly influence the visual and sensorial quality of food products (Malone, Appelqvist, & Norton, 2003). Consumers always have the preference of food products, which have appealing surface appearance and texture. Although the importance of surface properties of foodstuff has been well recognized, the creation and characterization of surface texture are still poorly understood. Little progress has been made so far on the surface quantification of food materials despite of various techniques that have been made available for surface researches of non-food materials. Very recently, surface friction measurement method has been * Corresponding author. Tel.: C44 113 343 2748; fax: C44 113 343 2982. E-mail address: [email protected] (J. Chen). 1 Current address: Instituto de Fı´sica de Lı´quidos y Sistemas Biolo´gicos, Calle 59 No. 789, 1900 La Plata, Argentina.

0268-005X/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodhyd.2005.04.002

successfully applied in authors’ group for the characterization of surface properties of heat-set whey protein gels (Chen, Moschakis, & Nelson, 2004). It was found that, by analysing the speed- and load-dependence of surface friction, the surface roughness and wetness of protein gels could be easily distinguished. This work will further explore the possibility of quantitative characterization of surface geometry of these protein gels based on the images from confocal laser scanning microscope. Surface geometry is by nature a three-dimensional feature. In theory, any measurement of two-dimensional profiles or sections cannot give a complete description of the real surface topography features. However, in practice, twodimensional measurements are still generally acceptable on the assumption that the surface is isotropic, that is, the surface has the same topography features across the X–Y plane. Based on the analysis of the profile of a surface, characteristic features of the surface can be quantitatively described using a number of statistical parameters, such as root-mean-square roughness (Rq) and average (or arithmetic) roughness (Ra). Root-mean-square roughness is defined as the square root of the mean of the height square

J. Chen et al. / Food Hydrocolloids 20 (2006) 468–474

deviations from the mean and is probably the most commonly used surface roughness parameter sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi PN 2 iZ1 ðzi K za Þ Rq Z ; (1) N where zi is the surface height at each measuring point, N is the total number of measuring points, and za is the mean height of the surface profile and is defined as za Z

N 1 X z: N iZ1 i

(2)

Average (or arithmetic) roughness is defined as the average of deviation of surface from its baseline and is also commonly used for surface geometry quantification: Ra Z

N 1 X jz K za j: N iZ1 i

(3)

Both root-mean-square roughness and average roughness gave a statistical value of surface roughness and were found effective in describing and differentiating surface topography of various materials (Maksumov, Vidu, Palazoglu, & Stroeve, 2004; Tay, Sikdar, & Mannan, 2002). Parameters such as peak-to-valley height, skewness of surface height distribution, and kurtosis of surface height distribution are also used for surface geometry characterization. There has been great achievement in developing techniques for surface characterization during last few decades (Myshkin, Grigoriev, Chizhik, Choi, & Petrokovets, 2003). Tribometry and contact profilometry are the classical techniques widely used for the surface characterization of solid materials (Chappard et al., 2003; Felder & Samper, 1994; Luengo, Tsuchiya, Heuberger, & Israelachvili, 1997; Tay et al., 2002). Optical Microscopy is a convenient technique for surface observation of almost any material, but application of optical lenses means that the resolution of normal optical microscope can only reach micrometer length scale. Surface scanning microscopy such as atomic force microscopy (AFM) is probably the most powerful technique so far for surface imaging and quantification. Since its introduction less than two decades ago (Binning, Quate, & Gerber, 1986), AFM has been widely used for high resolution profiling of surface morphology and nanostructure of various materials (Chakrapani, Mitchell, van Winkle, & Rikvold, 2003; Jalili & Lasminarayana, 2004; Jandt, 2001; Morris, Kirby, & Gunning, 1999; Pang, Baba-Kishi, & Patel, 2000). High resolution and capability of nanoscale surface profiling are unmatchable advantages of this technique. Scanning electron microscopy (SEM) is another powerful technique for high quality surface imaging and profiling of various materials (Brooks & McGill, 1994; Chappard et al., 2003; Huang, Li, Shen, Zhu, & Xu, 2002; Vansteenkiste, Davies, Roberts, Tendler, & Williams, 1998). With the help of ever advancing software, SEM can now produce high quality

469

three-dimensional surface images and perform complex surface quantification. Other established technique for surface characterization include methods based on the specularly reflected light of surfaces such as: Fibre Optic Reflectometer (Silvennoinen et al., 1993), Glossy Meter (Huang et al., 2002), surface glistening points method (Lu, Koenderink, & Kappers, 1999; Quevedo & Aguilera, 2004) and methods based on surface contact such as contact angel measurement (Meiron, Marmur, & Saguy, 2004). While the above mentioned techniques have their unique advantages for the characterization of certain type surfaces and materials, they all show great limitation for surface imaging and characterization of protein gels and other food materials. Surface deformability and the existence of surface moisture of such food materials are the main problems for the application of these techniques. For example, damage of a delicate surface could be easily caused by the tapping probe of an AFM and other techniques, which use a contacting surface-probe. Surface moisture of protein gels and other wet food material is the biggest problem for the application of SEM technique to wet surfaces. In these cases, surface dehydration would be essential, but surface distortion would almost be inevitable. Environmental scanning electron microscopy (ESEM) was further developed for surface imaging of wet materials without the need of a full removal of surface moisture. In reality, however, the quality of surface image will have to be compromised for high moisture containing systems such as protein gels whose water content could be as high as 99%. In addition to these limitation and drawbacks, there is a non-stated assumption for these surface characterization techniques: the surface asperities are cone-shaped or column-shaped, or that the tips of the asperities are at least no larger than their bases. However, if the bases of asperities are smaller than their tips or bodies, such as diamond-shaped, mushroom-shaped and upside down coneshaped asperities, surface profiles based on the above techniques could be seriously misleading. This is because the void space beneath the head of a mushroom-shaped asperity, for example, is simply undetectable by either a surface contacting probe or by surface reflected light. It is probably true that asperities of most engineered surfaces are cone-shaped or column-shaped. However, many naturally grown surfaces may not necessarily follow this rule. For example, particulate gels have networks of particle strands growing with a particular pattern in certain directions (Dickinson, 1994, 1995; Doi, 1993; Stading & Hermansson, 1991). Therefore, diamond-shaped, mushroomed-shaped, or upside down cone-shaped surface asperities could be highly possible in aggregated particle gels. Confocal laser scanning microscope (CLSM) could provide a solution to the above problems. As a nonconventional light microscope, CLSM technique has a number of advantages for microstructure observation of foodstuff and colloidal systems. With this technique, one can obtain images of the microstructure of the surface or an

470

J. Chen et al. / Food Hydrocolloids 20 (2006) 468–474

internal region of a sample without having to slice it. This capability has been widely used for the observation of the microstructure of particle gels across the X–Y plane either at the surface or deep inside the gel. This work intended to explore the possibility of using CLSM for microstructure observation and quantification across X–Z plane of a surface. No surface-dehydration and no surface-distortion were involved and, therefore, particle gels can retain their delicate surface texture and topographic features during the observation. Another unique advantage is that all possible forms of surface asperities (upside-down cone-shaped, diamond-shaped, mushroom-shaped, etc.) will be truly recorded without any distortion. In this work, images of heat-set whey protein gels were obtained using CLSM technique. These images were analysed to obtain quantitative description of surface geometry. The fractal nature of gel surface was also examined based on the CLSM images.

2. Materials and methods 2.1. Gel preparation Commercial whey protein isolate (Lacprodan DI-9224) was supplied by Arla Foods Ingredients amba (Videbaek, Denmark). The product has a high content of protein (O 93.5%), but a low content of fat (!0.2%) and carbohydrates (!0.2% lactose). The moisture content of the sample is less than 6.0%. AR grade NaCl (99.95) was purchased from Merck (Poole, UK). Distilled water was used for the preparation of protein solutions and the pH of the protein solutions was 6.8. Gels were prepared by heat treatment. Protein solutions (14 wt%) were carefully transferred into aluminium dishes (81 mm in diameter and 11 mm height) and foam bubbles created during the transfer were carefully removed. The dishes were then placed in a water bath at 90 8C. The levelling of sample dishes was carefully checked using a spirit levelling device. The sample dishes were removed after 1 h heat treatment and immediately cooled in a refrigerator at 5 8C and stored overnight. Protein gels containing 0 and 200 mM NaCl were investigated.

during the creation of each image. The image data files were transferred to a separate computer for further analysis. The CLSM was operated in fluorescence mode and Rhodamine B dye was used to stain protein. A staining solution of concentration 0.01% w/v was prepared by dissolving 1 mg of Rhodamine B in 10 ml of Milli-Q water (water purified by treatment with a Milli-Q apparatus, Millipore, Bedford). The solution was stored in a dark place. Fluorescence from the sample was excited with the 543 nm line of a green He–Ne laser. Samples of protein gels were transferred to a microscope slide. Rhodamine B was added on the top of protein gels. A round coverslip (0.17 mm thickness) was carefully placed on the top of the gels. An immersion oil with refractive index 1.518 was used. 2.3. Image analysis The CLSM images of the gels surface profile are shown in Fig. 1. The images give the cross-section images (X–Z plane) of the gels and the top boundary represents the topography profile of the gel. The colour was codified in 256 intensity levels (IL) ranging from high protein concentration to low protein concentration. Since, there is a high contrast between the gel and the air phase above the surface, the two phases could be separated by using a simple thresholding technique. Any pixel in the image with an intensity level below a certain threshold was considered as part of the air

2.2. Microstructure observation GLSM (Leica TCS SP2, Leica Microsystems Heidelberg GmbH, Mannheim, Germany) was used for the observation of surface geometry of protein gels. The microscope was mounted on a Leica Model DM RXE upright microscope base. A 40! oil-immersion objective with numerical aperture 1.25 was used in all of the experiments reported here. Each image contains 1024!1024 pixels, with a grey value ranging from 0 to 255. The voxel size, the volume (3d) resolution element was adjusted at 122 mm. The signal from the samples was collected and eight scans were averaged

Fig. 1. CLSM profile of the surface of a heat-set WPI gel (14 wt%). (A) No added salt, and (B) 200 mM NaCl. The inset in (B) is a 4! enhancement of part of the surface showing a mushroom-shaped tip.

J. Chen et al. / Food Hydrocolloids 20 (2006) 468–474

3. Results and discussion The applicability of CLSM for the investigation of surface topography was investigated for two heat-set whey protein gels with and without the presence of salt. CLSM images of the two gels are shown in Fig. 1. The two gels contain the same amount of protein (14 wt%) but different amount of salt (0 and 200 mM NaCl). We can see that the gel containing no salt has a very smooth and flat surface, almost featureless to the naked eyes, but the gel containing 200 mM NaCl has a rough surface, with numerous peaks and troughs. The surface images of whey protein gels are in general consistent with previous observation of the microstructure of these gels (Langton & Hermansson, 1996; Verheul & Roefs, 1998), where a fine stranded microstructure was observed for a pure whey protein gel but much coarser microstructure for the gels containing large amounts of salt. Surface asperities for the gel containing 200 mM NaCl correspond to the aggregates of protein particles. Their shape and size appear to be very irregular and variable. Although many of the asperities are indeed cone-shaped or column-shaped, there are cases, where the bases of the asperities are smaller than their tips or bodies. There are also cases where asperities show significant inward bending curvatures. This can be clearly seen from the insert in Fig. 1(B) for one particular asperity. The observation of such kinds of surface asperities could be highly unlikely by other surface characterization techniques using either tapping probe or light reflection. Surface profiles for the protein gels were extracted from their CLSM images based on the grey level intensity (see Fig. 2). For the convenience of analysis and

z (µm)

0.8

A

0.4 0.0 –0.4 –0.8 6

B

z (µm)

3 0 –3 –6 0

20

40

60 x (µm)

80

100

120

Fig. 2. Surface profile of heat-set WPI gels (14 wt%). (A) No added salt, and (B) 200 mM NaCl.

calculation, the average height of the surface was used as the baseline of a surface and was set as height zero. Fig. 3 shows the height distribution of surface asperities for the two gels. We can see almost symmetrical height distributions along the baseline, with almost identical numbers of peaks and troughs on either side of the baseline. Normal probability distribution function (Gaussian model) was applied to analyse the height distributions. The R-square values of Gaussian fitting for the two systems were 0.90 and 0.66, respectively. This means close to normal distribution of the surface height for the gel containing no salt. However, the surface height distribution for the gel containing 200 mM NaCl is hardly normal, but rather a multi-peaked distribution. A normal height distribution suggests random nature of the surface asperities such as most engineered surfaces. It is not yet clear what a multi-peaked height distribution means in terms of surface topography, but one may speculate that such a distribution may indicate the significance of certain height levels associated to the microstructure of the protein aggregates. B

A

60 50 Frequency

phase and set to be black (i.e. ILZ0). Any other pixel was set to be white (i.e. ILZ256). Then the surface profile was extracted from the image by locating the gel pixel (ILZ256) with the highest Z coordinate (Zmax) for each value of X in the image. Then the profile is defined the function f(X)Z Zmax. This function was discrete and consists of all data points over the width of the images in pixels. Although the CLSM technique is capable to observe a surface consisted of any shape of surface asperities, mathematical quantification of the surface profile could still be a problem. The above extraction technique of the surface profile f(X) may prevent us from studying the rich information provided by these CLSM images. Clearly, as the function f(X) is single-valued, any mushroom-shape asperity, for example, will be converted into a columnar shape by this procedure. However, having a single-valued function of X allows us to perform a variety of standard data analysis that is useful in describing general surface properties. A study of the full complex surface profile, which cannot be represented by a simple function of X, is still undergoing.

471

40 30 20 10 0

–0.6 –0.4 –0.2 0.0 0.2 0.4 0.6 –6

z (µm)

–4

–2

0

2

4

6

z (µm)

Fig. 3. Height distribution of surface asperities of heat-set WPI gels (14 wt%). (A) No added salt, and (B) 200 mM NaCl.

472

J. Chen et al. / Food Hydrocolloids 20 (2006) 468–474

However, we have to bear in mind that, like all imaging techniques, CLSM has its resolution limit. For the conditions applied in this work, the resolution limit on the Z-direction can be estimated using the Kinos formula (Pawley, 1995). For a wavelength of 543 nm and the numerical aperture of 1.25, the resolution limit in Z-direction was estimated to be around 0.7 mm. This value is in close range of the height changes for the surface profile of the gel containing no salt (see Figs. 1(A) and 2(A)). This shows that CLSM maybe less capable for surface differentiation of smooth protein gels due to its resolution limit. However, the CLSM technique is proved to be highly adequate in characterizing the surface profile for the saltcontaining gel whose surface height variation exceeds the resolution limit of the technique (Figs. 1(B) and 2(B)). Discrete Fourier Transform (FT) was applied to extract possible topography features from these digitalised surface profiles. FT is a powerful technique to represent a periodic function into a harmonic combination of periodic functions. A FT spectrum can be seen as a linear combination of individual profiles of particular wavelength. Each dominant frequency in the FT spectrum would, therefore, represent the characteristic wavelength of an individual profile (LZ 1/f). By analysing the characteristic frequencies of a surface profile in the Fourier space, we can then identify the characteristic wavelengths involved in the profile, which in turn are related to the size and shape of protein aggregates responsible for the surface asperities. Fig. 4 shows the FT of surface profiles of two protein gels. Although no dominant frequency could be identified in both gel systems, but one or two peaks can be seen standing out for each system. In graph A (the gel without salt addition), the peak at 0.61 mmK1 shows the largest amplitude, implying a periodic feature of 0.05

A

Table 1 Surface roughness of heat-set whey protein gels Gels

Rq (mm)

Ra (mm)

14 wt% WPI, no salt 14 wt% WPI, 200 mM NaCl

0.20 2.38

0.16 1.91

1.64 mm. In graph B, the gel containing 200 mM NaCl has two outstanding contributions at periodic length scales of 7.1 mm (0.14 mmK1) and 6.2 mm (0.16 mmK1). One may note that these length scales from FT analysis correlate to periodic width of asperities but have no direct links to the heights of surface asperities. Further investigation is needed for full understanding of the periodic feature of asperities both in width and in height. The quantification of surface geometry has also been carried out based on CLSM surface profiles (Fig. 2). Two statistical parameter, root-mean-square roughness Rq and arithmetic roughness Ra, are given in Table 1 for the two gels. The very smoothness of the gel containing no salt can be clearly seen by its small Rq (0.20 mm) and Ra (0.16 mm). And the drastic difference of the surface of the gel containing 200 mM NaCl can also be seen from its high Rq (2.38 mm) and Ra (1.91 mm). These surface quantification agrees well with the results from recent surface friction investigation of protein gels in this lab, where the surface friction force for the gel containing 200 mM NaCl showed much higher load-dependence than that for the gel containing no salt, an indication of much rougher surface of the former (Chen et al., 2004). The characteristic surface geometry of protein gels can be further analysed by the length of their surface profiles. Fig. 5 shows the integration length of surface profile against the profile width. We observe straight linear relationships for protein gels. This linear relationship indicates the reliability of CLSM surface profiling. For a drifted or 100

0.03 C

0.02

80

0.01 0.00 B

Amplitude

0.8 0.6 0.4

Total profile length (µm)

Amplitude

0.04

B 60 A 40

20

0.2 0 0

0.0 0.01

0.1 Frequency (µm–1)

1

Fig. 4. FT analysis of surface profiles of heat-set WPI gels (14 wt%). (A) No added salt, and (B) 200 mM NaCl.

10

20 30 Profile width (µm)

40

50

Fig. 5. Total profile length against the profile width for heat-set WPI gels (14 wt%). (A) A strictly flat surface, (B) no added salt, and (C) 200 mM NaCl.

J. Chen et al. / Food Hydrocolloids 20 (2006) 468–474

distorted surface profile, there will be unlikely a linear relationship. The dashed line in Fig. 5 has a slope of 1 representing the case of a strictly flat surface and can be used a reference. The protein gel without salt addition shows straight line with a slope of 1.42, but the gel containing 200 mM NaCl has a much higher slope of 1.98. Proportional increase of the length of surface profile against surface width suggests that each surface profile has a particular set length for each unit surface width and this set length is surface-dependent. High slope for the gel containing 200 mM NaCl (line C) suggests that this gel has the longest surface length (or the largest surface area if converted into two dimensions), while a low slope for the gel containing no salt (line B) indicate the smoothness and flatness of their surfaces. The slope in Fig. 5 is sometimes referred in literature as the ratio of the length of a surface profile to that of a flat surface (Rs). Heat-set whey protein gels have been well recognized for their fractal nature of three-dimension microstructure (Chen & Dickinson, 1998; Ikeda, Foegeding, & Hagiwara, 1999; Verheul, Roefs, Mellema, & de Kruif, 1998). The question has always been that whether the fractal nature of a network could be extended to its surface for these aggregated particle gel systems. For this purpose, a yardstick walk method was used to measure the length of a surface profile at different magnification levels as was explained by Mandelbrot (1983) L Z F31KD

(4)

or log L Z ð1 K DÞlog 3 C log F;

(5)

where L is the measured length of a curve with a yardstick of length 3, D is the fractal dimension of the curve and F is a constant. Fig. 6 shows the relationship between log (L) and log (3) for the gel containing 200 mM NaCl. Fractal nature of the surface of the gel was evidently revealed by the linear relationship between log (L) and log (3) and a fractal dimension of 1.15G0.01 was obtained. No data is yet available in the literature to prove this finding. However, it

log L (pixel)

3

2.9

2.8

0

0.2

0.4 logε (pixel)

0.6

0.8

Fig. 6. Fractal dimension analysis of a heat-set WPI gel (14 wt%) containing 200 mM NaCl. A linear log–log relationship was given between the total length of surface profile against the measuring unit length.

473

appears that the surface fractal dimension of this gel is smaller than that of UK’s coast line, where a fractal dimension of 1.3 was obtained (Maletsky, Perciante, & Yunker, 1992), but is in the range of that of stainless steel with various surface finishes, whose fractal dimensions were found to be between unit and 1.2 (Chesters, Wen, Lundin, & Kasper, 1989). Unfortunately, we were unable to calculate the fractal dimension of the smooth one due to not higher enough magnification for this smooth gel. Another observation worth of mentioning is the orientation of particle aggregates at the surface. This was particularly evident for the gel containing 200 mM NaCl, where all the aggregates appeared oriented toward the surface. This preference of aggregate growth in the Z-direction indicates an anisotropic nature of particle aggregates within the surface layer. For the observed thickness (around 20 mm), there was merely strands in Xand Y-directions. There is no question that this layer would have significantly different physicochemical properties (such as porosity, particle density, mass transfer rate, light reflectivity, etc.) compared to those in the bulk phase. We may call this anisotropic region the skin layer of the particle gel. So far, little research has been done on the surface properties of particle gel systems, even though there have been computer simulation studies on the particle aggregation within an adsorbed monolayer (Pugnaloni, Ettelaie, & Dickinson, 2003; Wijmans & Dickinson, 1999). Unfortunately, most of such simulation only considers the lateral interactions within a two-dimensional monolayer and does not take consideration of simultaneous aggregation that occurs in the bulk phase of the gel system. Results from such simulation could only provide limited information on the surfaces of aggregated particle gels. More sophisticated modelling would be needed in order to get more realistic picture of surface aggregation of particle gels.

4. Conclusions This work successfully applied CLSM technique for surface imaging of wet and deformable protein gels. The unique advantage of this technique is its no surface contacting, no surface dehydration, and, therefore, no surface distortion. The surface roughness/smoothness of heat-set whey protein gels by CLSM is consistent in general with their three-dimensional microstructure. The gel containing 200 mM NaCl has much rougher surface with a surface roughness Rq of 2.38 mm and Ra of 1.91 mm. The whey protein gel without salt addition has a Rq of 0.20 mm and a Ra 0f 0.16 mm. FT analysis revealed different periodic length scales for the two protein gels. Surface fractal nature was confirmed for the gel containing 200 mM NaCl, where a surface fractal dimension was found to be 1.15. This work demonstrated the feasibility of CLSM technique for surface characterization of aggregated particle gels both qualitatively and quantitatively, even though

474

J. Chen et al. / Food Hydrocolloids 20 (2006) 468–474

the technique has a limited capability for quantitative characterization of smooth surfaces due to its resolution limit. The finding of this work suggests that CLSM would be a suitable technique for surface topography investigation of wet and deformable surfaces, where other surface contacting and probing techniques have been proved to be less appropriate.

Acknowledgements L. A. Pugnaloni is a member of CONICET (Argentina). T. Moschakis wishes to acknowledge Greek State Scholarships Foundation (IKY) for his studentship. Authors would also like to acknowledge financial support from Food Processing Faraday for this project.

References Binning, G., Quate, C. F., & Gerber, Ch. (1986). Atomic force microscope. Physical Review Letters, 56, 930–933. Brooks, C. R., & McGill, B. L. (1994). The application of scanning electron microscopy to fractography. Material Characterization, 33, 195–243. Chakrapani, M., Mitchell, S. J., van Winkle, D. H., & Rikvold, P. A. (2003). Scaling analysis of polyacrylamide gel surfaces synthesized in the presence of surfactants. Journal of Colloid and Interface Science, 258, 186–197. Chappard, D., Degasne, I., Hure, G., Legrand, E., Audran, M., & Basle, M. F. (2003). Image analysis measurements of roughness by texture and fractal analysis correlate with contact profilometry. Biomaterials, 24, 1399–1407. Chen, J., & Dickinson, E. (1998). Viscoelastic properties of heat-set whey protein emulsion gels. Journal of Texture Studies, 29, 285–304. Chen, J., Moschakis, T., & Nelson, P. (2004). Application of surface friction measurements for surface characterization of heat-set whey protein gels. Journal of Texture Studies, 35, 493–510. Chesters, S., Wen, H. Y., Lundin, M., & Kasper, G. (1989). Fractal-based characterization of surface texture. Applied Surface Science, 40, 185–192. Dickinson, E. (1994). Computer simulation of particle gel formation. Journal of Chemical Society Faraday Transaction, 90, 173–180. Dickinson, E. (1995). Simulated structure of mixed particle gels. Journal of Chemical Society Faraday Transaction, 91, 51–58. Doi, E. (1993). Gels and gelling of globular proteins. Trends in Food Science and Technology, 4, 1–5. Felder, E., & Samper, V. (1994). Experimental study and theoretical interpretation of the frictional mechanisms in steel sheet forming. Wear, 178, 85–94. Huang, H., Li, Y., Shen, J. Y., Zhu, H. M., & Xu, X. P. (2002). Microstructure detection of a glossy granite surface machined by grinding process. Journal of Materials Processing Technology, 129, 403–407. Ikeda, S., Foegeding, E. A., & Hagiwara, T. (1999). Rheological study on the fractal nature of protein gel structure. Langmuir, 15, 8584–8589. Jalili, N., & Lasminarayana, K. (2004). A review of atomic force microscopy imaging systems: Application to molecular metrology and biological sciences. Mechatronics, 14, 907–945.

Jandt, K. D. (2001). Atomic force microscopy of biomaterials surfaces and interfaces. Surface Science, 491, 303–332. Langton, M., & Hermansson, A.-M. (1996). Texture as a reflection of microstructure. Food Quality and Preference, 7, 185–191. Lu, R., Koenderink, J. J., & Kappers, M. L. (1999). Surface roughness from highlight structure. Applied Optics, 38, 2886–2894. Luengo, G., Tsuchiya, M., Heuberger, M., & Israelachvili, J. (1997). Thin film rheology and tribology of chocolate. Journal of Food Science, 62, 767–772 (see also page 812). Maksumov, A., Vidu, R., Palazoglu, A., & Stroeve, P. (2004). Enhanced feature analysis using wavelets for scanning probe microscopy images of surfaces. Journal of Colloid and Interface Science, 272, 365–377. Maletsky, E., Perciante, T., & Yunker, L. (1992). Fractals for the classroom, part one, introduction to fractals and chaos. New York: Springer. Malone, M. E., Appelqvist, I. A. M., & Norton, I. T. (2003). Oral behaviour of food hydrocolloids and emulsions. Part 1. Lubrication and deposition considerations. Food Hydrocolloids, 17, 763–773. Mandelbrot, B. B. (1983). The fractal geometry of nature. New York: W.H. Freeman and Company. Meiron, T. S., Marmur, A., & Saguy, I. S. (2004). Contact angle measurement on rough surface. Journal of Colloid and Interface Science, 274, 637–644. Morris, V. J., Kirby, A. R., & Gunning, A. P. (1999). Atomic force microscopy for biologists. London: Imperial College Press. Myshkin, N. K., Grigoriev, A. Y., Chizhik, S. A., Choi, K. Y., & Petrokovets, M. I. (2003). Surface roughness and texture analysis in microscale. Wear, 254, 1001–1009. Pang, G. K. H., Baba-Kishi, K. Z., & Patel, A. (2000). Topographic and phase-contrast imaging in atomic force microscopy. Ultramicroscopy, 81, 35–40. Pawley, J. B. (1995). Handbook of biological confocal microscopy. New York: Plenum Press. Pugnaloni, L. A., Ettelaie, R., & Dickinson, E. (2003). Do mixture of proteins phase separate at interfaces? Langmuir, 19, 1923–1926. Quevedo, R., & Aguilera, J. M. (2004). Characterization of food surface roughness using the glistening points method. Journal of Food Engineering, 65, 1–7. Silvennoinen, R., Peiponen, K. E., Ikonen, K., Asakura, T., Luostarinen, J., & Gu, C. (1993). A fibre optic reflectometer. Optics and Lasers in Engineering, 18, 15–24. Stading, M., & Hermansson, A.-M. (1991). Large deformation properties of b-lactoglobulin gel structures. Food Hydrocolloids, 5, 339–352. Tay, F. E. H., Sikdar, S. K., & Mannan, M. A. (2002). Topography of the flank wear surface. Journal of Materials Processing Technology, 120, 243–248. Vansteenkiste, S. O., Davies, M. C., Roberts, C. J., Tendler, S. J. B., & Williams, P. M. (1998). Scanning probe microscopy of biomedical interfaces. Progress in Surface Science, 57, 95–136. Verheul, M., & Roefs, S. P. F. M. (1998). Structure of whey protein gels, studied by permeability, scanning electron microscopy and rheology. Food Hydrocolloids, 12, 17–24. Verheul, M., Roefs, S. P. F. M., Mellema, J., & de Kruif, C. G. (1998). Power law behaviour of structural properties of protein gels. Langmuir, 14, 2263–2268. Wijmans, C., & Dickinson, E. (1999). Brownian dynamics simulation of the displacement of a protein monolayer by competitive adsorption. Langmuir, 15, 8344–8348.