- Email: [email protected]
Friction measurements with yoghurt in a simulated tongue-palate contact
BIOTRI-00026; No of Pages 11 Biotribology xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Biotribology journal homepage: http://www.elsevier.com/locate/biotri
Friction measurements with yoghurt in a simulated tongue-palate contact S. Tsui a, J. Tandy a, C. Myant a,b, M. Masen a, P.M. Cann a,⁎ a b
Tribology Group, Department of Mechanical Engineering, Imperial College London, SW7 2AZ, United Kingdom Dyson School of Engineering Design, Imperial College London, United Kingdom
a r t i c l e
i n f o
Article history: Received 2 December 2015 Received in revised form 18 February 2016 Accepted 21 February 2016 Available online xxxx Keywords: Friction Oil/water emulsion Boundary lubrication Oral processing
a b s t r a c t The perception of many food attributes is related to mechanical stimulation and friction experienced in the tongue-palate contact during mastication. Friction in the tongue-palate is determined by the changing film properties (composition, component distribution, thickness) in the conjunction. We suggest this evolution is essentially determined by tongue-palate film loss rather than shear flow entrainment which predominates in conventional bearing lubrication. The paper reports friction measurements in a simulated tongue-palate contact for a range of high and low fat dairy foods. A reciprocating, sliding contact with restricted stroke length (bcontact width) was used; under these conditions there is negligible shear-entrainment of fluid from outside the contact area. The tongue-palate contact was simulated by a PDMS ball and glass surface. The effect of hydrophobic and hydrophilic surfaces on friction was investigated for different fat contents (0, 4.2, 9.5% wt fat). Friction was measured over 60 s of rubbing. Significant differences were observed in the friction change with time for different fat contents (μ 9.5 b μ 4.2 b μ 0 wt%) and for different surface energy conditions (μ hydrophilic b μ hydrophobic). Post-test visualisation of the rubbed films showed that low friction coefficient was associated with the formation of a thin oil film on deposited particulate solids. © 2016 Published by Elsevier Ltd.
1. Introduction Oral processing or mastication is a dynamic process whereby food structure is broken down and transported to the pharynx prior to swallowing. It is a complex series of processes which involve a number of surface interactions including tooth-tooth and tongue-palate mechanisms [1]. Sensory perception of food involves visual, physical and physiological elements in which the role of structure, rheology and the mechanisms of food breakdown are not fully understood. The perception of taste and texture which includes thickness, smoothness, slipperiness [1] and to some extent astringency are usually related to mechanical stimulation and hence friction experienced in the mouth. Semi-solid food, which includes cream, yoghurt, custard and mayonnaise, are oil-in-water (O/W) emulsions which also contain proteins, emulsifiers, carbohydrates (including sugar) and other soluble components. Most formulation work has centred on research by food scientists [2,3] and the development of complex emulsions and micro-phase fluids. The focus has been on controlling the chemical and physical characteristics to improve the stability and storage properties [2]. However in recent years the role of food composition and structure in oral processing has received increasing attention. As Dalgleish [2] states “These structures in turn give rise to the perception of texture as they are consumed”. The development of low and zero-fat food with the mouth⁎ Corresponding author. E-mail address: [email protected] (P.M. Cann).
feel of high fat content analogues remains an important objective of the food industry [2,3]. The emphasis has been the role of fat in the original product structure [2] and how this evolves during mastication. The low/very low fat yoghurt market in the UK is currently worth over £500 million/year [4] and is growing rapidly. However dietary focus has recently switched to the reduction of sugar in foods and unfortunately many low-fat formulations contain higher levels of sugar (see Table 1). During mastication the structure of these foods change and the contribution of all components, not just fats, to perceived taste and texture must be understood if we are to effectively develop new products that satisfy consumer expectations. The problem remains of how to quantify the mouth-feel of these products without excessive use of panel tests. Mouth-feel is the tactile sensation created when food is rubbed between the tongue and palate, up to the point of swallow. Many studies [5] have tried to link oral perception to bulk rheology of the foods however this method does not capture many important aspects of the problem. As Malone et al. [6] concluded “thin film tribological properties correlated more closely with the composition and texture of the samples than the bulk rheological properties”. When food is masticated the material in the tongue-palate conjunction is sheared and reduces so that the surface and thin film properties and the structural evolution, rather than the bulk fluid, determine oral perception. Selway and Stokes [7] note “The change in lengthscale (film thickness between oral surfaces) is identified as a key feature driving the transition from a regime where the bulk fluid properties govern the sensory response, to one where surface interactions dominate.”
http://dx.doi.org/10.1016/j.biotri.2016.02.001 2352-5738/© 2016 Published by Elsevier Ltd.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
2
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Table 1 Composition of test fluids. Composition/100 g
0% wt fat
4.2% wt fat
9.5% wt fat
Total fat of which saturates Carbohydrates Of which sugars Protein
0g 0g 8.5 g 8.5 g 5.9 g
4.2 g 2.5 g 4.2 g 2.7 g 4.6 g
9.5 g 5.9 g 6.4 g 6.4 g 4.58 g
In recent years there has been a growing number of studies [6–18] reporting the tribological evaluation of semi-solid foods (e.g. yoghurt, salad dressings, milk, chocolate) with the intention of correlating measured friction with oral perception and to provide more relevant assessment of new products. The approach has been very successful as Chojnicka-Paszu et al. [9] concluded from their study of milk “indicated a good correlation between creamy attributes and measured friction coefficient, a result that validates the use of tribology as an analytical technique to better programme specific sensory products in product development and reformulation”. Research has usually focussed on correlation of friction behaviour with fat content [6,9,18] of the sample however in future we must be able to relate measured friction to the wider chemical composition and component distribution properties of the lubricating film. A thorough review of the topic is provided by Stokes et al. [15] who consider more general aspects of the relationship between rheology, lubrication and oral processing. The focus of the current paper is to reexamine the tribology of the tongue-palate and development of improved test methodologies to better simulate oral processing.
Food tribology tests are usually run as “Stribeck” curves [6,7,12–14,] with average friction measured over a range of speeds. The technique has been successfully used to distinguish between different fat contents in semi-solid foods including yoghurt [7], milk [9] and O/W emulsion [6, 11]. Typical examples of MTM speed-sweep curves carried out in our laboratory are shown in Fig. 1a. The friction coefficient in the lowspeed region (usually b50 mm/s) decreases with increasing fat content as has been reported in other studies [6,7,9]. For the speed-sweep results friction coefficient at 20 mm/s is ranked 9.5% fat (μ = 0.025) b4.2% fat (μ = 0.04) b 0% fat (μ = 0.15). However measurements at constant speed (20 mm/s) gave different results for the 9.5% fat yoghurt (average μ = 0.09) as shown in Fig. 1b. This result was unexpected and possibly indicates that the shear history experienced by the food is important in determining friction response. The friction/speeds curves are usually interpreted in terms of classical fully-flooded lubrication regimes (boundary, mixed, full film) denoted by changes in friction coefficient [6,7,8,11–15]. For the O/W emulsions the reduction in friction at slow speeds is attributed to the preferential entrainment of oil [4,7] and the formation of an oil “boundary” film [8] on the contacting surfaces. For example Selway and Stokes [7] concluded “The medium- and high-fat products both generate friction curves identical to that of the pure oil phase over the entire range of speeds measured, indicating preferential entrainment of oil and exclusion of the thickened aqueous phase from the contact zone.” Thus in these tests the conditions experienced by the fluid as it is entrained through the inlet into the contact are critical in determining both the composition of the film formed (“preferential entrainment of oil”) and the resulting friction behaviour. Thus the film in the contact zone is not representative of multiphase composition of the food.
2. Research into food tribology 2.2. Tribology analysis of oral processing 2.1. Conventional tribology test methods A variety of test devices have been used to study friction properties of foods, Prakash et al. [8] provide a recent review. Most studies of fluid or semi-solid food tribology [6,7,9,12,13,14] have used the Mini Traction Machine (MTM) or Optical Tribological Configuration (OTC) [11,17]. The MTM uses a ball-on-disc contact immersed in excess of fluid and measures friction coefficient as a function of different test conditions including entrainment speed, slide-roll ratio and temperature. The OTC uses a pin (PDMS or other soft surface) reciprocating against a glass counterface [11,17]. Lee et al. [10] also used a pin-on-disc (unidirectional sliding) device (CSM Instruments) to study the tribology properties of molten chocolate. In tribology terms the tongue-palate contact is characterised as a soft, rough, viscoelastic specimen (tongue) rubbing against a harder, smooth counterface (palate). The contact pressure is generally low (~ MPa) and the sliding speeds in the range 10–30 mm/s to match tongue movement in eating [19]. The various test methods described in the literature use a wide range of materials (polymer, zirconia, glass, steel, porcine tongue), specimen surface properties (smooth/textured, hydrophobic/hydrophilic) contact conditions (pressures, temperature, speed range) and kinematics (sliding/rolling/reciprocation). The test specimen materials are chosen to simulate the tongue-palate contact and include “soft/soft” (polymer/polymer) [7] and “soft/hard” (polymer/steel or glass, porcine tongue/glass) [11,12,13,14,17] combinations. The kinematic condition is particularly important as this plays a significant role in determining lubricant film properties. The MTM test has usually been used in a rolling-sliding configuration (50% SRR) where both the ball and disc are driven at different speeds [6,12–15]. Other studies have used simple sliding [10] or reciprocating sliding [11,16,17]. In some cases very high speeds (~ 500 mm/s [12]) were used, however the usual range is 5–80 mm/s [10,11,17]. This is supported Hiiemae and Palmer [19] who quote an average measured speed range of 10–32 mm/s for tongue movement during eating
The preceding discussion has emphasised the role of the inlet in determining the film formation and friction response of two-phase (emulsions) fluids. Lubricant is entrained into the contact region by the relative motion of the surfaces and thus forms a film which supports the load and separates the surfaces. A fundamental principle of tribology is that under normally lubricated conditions i.e. where there is continued supply of lubricant it is the properties (viscosity, composition) of the fluid in the inlet region that determine the lubricating film properties (chemistry, thickness) in the contact region. The shear conditions in the inlet (due to narrowing gap height) are severe and even for modest speeds (mm/s) can reach 105 s−1 for the classical ball/disc arrangement. This can lead to shear thinning or degradation, phase and compositional changes, particularly for emulsions, prior to the contact (Fig. 2a left image) and is clearly not representative of the eating process. Semi-solid foods are placed in the oral cavity and broken down either in the tooth-tooth or tongue-palate rubbing contacts. There is no continued flow of fresh material into the contact and a representative distribution of components is present at the start of mastication. The role of the inlet in determining film properties in the contact has been extensively studied in classical (oil-based) lubrication [20–22]. O/ W fluids are used as metal-working lubricants where shear degradation and coalescence of oil droplets in the inlet occurs to provide an oil-rich reservoir which supplies the lubricating film in the contact zone [20,21]. At low speeds the film in the contact is predominately composed of oil giving a thicker film and lower friction than the continuous water phase [20]. At higher speeds replenishment of the oil reservoir is not maintained and entrainment of the bulk fluid, which is predominately water, occurs [20,21]. The film thickness of industrial O/W emulsions often drop with increasing speed because of this effect [20]. Thus the oil/water composition in the contact will change depending on the properties of the fluid and the ability of the contact to generate and retain the oil reservoir (materials, surface energy, entrainment speed, chemical composition) [20,21].
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
3
Fig. 1. MTM test results for 0, 4.2 and 9% w/w fat yoghurts, 50% SRR, 35 °C, excess fluid (a) speed sweep results (b) constant speed as a function of time at 20 mm/s.
In addition to the oil/water phase semi-fluid foods often contain a dispersed solid phase (e.g. proteins) where the nearest engineering equivalent is lubricating grease which is a two-phase system (base oil/ thickener). Tribology studies of grease has shown complex lubrication behaviour where the film properties depend on the entrainment speed and lubricant supply (starvation) to the inlet [22]. At slow speeds much thicker films than expected are formed due to preferential entrainment of thickener particles [22]. At higher speeds starvation of the lubricant supply may occur or shear-degradation of the thickener matrix [22]. Thus in the usual test configuration (ball/disc) where a complex (multi-phase) fluid is entrained by the relative movement of the surfaces the chemical composition of the film in the contact is likely to be different to the bulk fluid. As a result lubricating film properties (thickness, friction) may change in a complex fashion depending on the inlet flow condition and local fluid composition. As all foods contain a mixture of components and in some cases phases it is impossible to interpret friction data, unless we understand more about the composition of the film in the contact zone and the effect of inlet flow. Although correlations have been made between food composition and friction response [7,8] it is difficult to foresee how the traditional lubrication test
can be developed further to provide a more relevant research tool without this information. Although the foregoing argument has emphasised the role of the inlet in the conventional test configuration this is not necessarily relevant to the oral processing problem as the inlet to the contact is effectively absent. During eating food is captured in the tongue-palate under load and then repeatedly sheared (Fig. 2a right image), thus the mixture of food components initially present is that occurring in the bulk material. The distribution and condition of food components in the tongue-palate conjunction is not filtered or sheared by flow through the converging gap of the inlet. Under repeated loading and shear the two/three-phase structure will break down and the distribution of food components changes due to surface adsorption, agglomeration, degradation or selective loss [23,24,25]. The change in component distribution due to simple shear can be seen clearly in the phase contrast microscopy images shown in Fig. 2b. In these images a small sample of yoghurt was sheared under simple, reciprocating sliding between two glass slides. The formation of “roll” structures can clearly be seen transverse to the sliding direction. These images are discussed in more detail later in the paper. In conclusion we suggest it is the loss
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
4
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Fig. 2. Inlet-controlled and zero-entrainment lubrication conditions and phase contrast microscopy images of sheared emulsion films.
mechanisms occurring during mastication rather than shearentrainment and inlet “filtering” which determine film thickness and component distribution in the tongue-palate contact and hence the evolution of friction. A second consideration is the timescale of the measurements; food is introduced into the tongue-palate and sensory information is immediately recorded during mastication. The dwell time of food in the mouth is short and the friction experienced in the first few seconds as the bolus thins which will likely determine the sensory assessment. Both the MTM and OTC tests can be used to provide friction change with rubbing time (Fig. 1b) however the time scale is much longer than experienced in normal mastication [7,15,9]. When considering the design and applicability of tribology tests for assessing mouth-feel it is instructive to consider the oral mechano-stimulation process. Mechano-receptors at the oral surface detect local friction and pressures changes at the food-mucosa interface during mastication [26]. The timescale of receptor response will be important when assessing the influence of friction changes contributing to oral perception. The development of new test methods which capture the time dependency of textural changes has also been discussed in a recent review paper by Stokes et al. [13].
In conclusion friction experienced in the tongue-palate is determined by the composition, component distribution and film thickness in the conjunction, however in traditional tribology tests these properties are controlled by the inlet flow. During mastication fresh food is supplied by periodically separating the tongue-palate interface; effectively this is a transient loading condition. Thus film thickness is predominately determined by frequency of this event and the tonguepalate food loss mechanisms rather than shear flow entrainment. The intention in this study is to develop a test method which replicates the tongue-palate loss and resupply mechanisms. 2.3. New tribology test method In this paper a new approach has been developed where friction change for a food sample is measured over the first 60 s of rubbing. A sliding, reciprocating test device (HFRR High Frequency Reciprocating Rig PCS Instrument UK) is used. Ranc et al. [16] used a similar test method to study the effect of PDMS surface texture (simulating the tongue) on friction. In the standard configuration the HFRR utilises a ball sliding against a flat disc. The test specimens are held in a temperature controlled bath. The major advantage is that a small amount of sample is
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
required and this is subject to repeated shearing. Friction measurements were taken every second. Food held in the tongue-palate interface is sheared repeatedly during mastication, thus the structure and distribution of food components will change and this will result in varying friction. Time-dependent changes particularly in the initial stages of mastication are likely to contribute to the perception of texture. In the first instance a simplified form of the test is reported using PDMS/glass as a tongue/palate model [16] operating at room temperature [18] and in the absence of saliva mucins [12]. However as the test develops these factors will be changed to improve the simulation of the tongue-palate oral processing. The Hertzian contact diameter (1.6 mm for 2N) is larger than the stroke length (1 mm) and thus reciprocation does not entrain significant amounts of fresh fluid into the contact zone making it analogous to movement of the tongue against the palate. In initial tests a constant load condition was used, in later work the load was removed every 30 s which represents a more realistic tongue-palate re-supply condition. A series of commercial dairy foods (yoghurt) with different fat contents was studied. 3. Experimental 3.1. HFRR tests The HFRR uses a reciprocating, sliding contact where the upper specimen is loaded and rubs against a lower disc. The load (2N), temperature (25 °C), stroke length (1 mm) and reciprocation frequency (10 Hz) are controlled by a computer. Friction was measured by a force transducer mounted on the lower specimen support. The test device uses a 19.05 mm diameter PDMS ball (Duro 30 supplied by PCS
5
Instruments) and glass microscope slide (VWR Premium Grade). A schematic diagram and photograph of the test device are shown in Fig. 3. Test conditions are summarised in Table 2. Two series of tests were carried out with either “hydrophobic” or “hydrophilic” surfaces. For hydrophobic (HB-HB) tests both specimens were cleaned in detergent solution, distilled water and then isopropyl alcohol (Sigma Aldrich Analar Grade). For the hydrophilic (HL-HL) tests both specimens were cleaned in isopropyl alcohol followed by treatment in a plasma cleaner. This procedure is used to remove organic contaminants and form an oxidised, hydrophilic surface [11]. All plasma-modified specimens were tested within 5 min of treatment. Photographs of water droplets on glass surfaces before and after plasma treatment are shown in Fig. 4. In a number of tests the effect of repeated addition of fresh food was investigated using a “lift-load” procedure. The tests were run as usual except after 30 s the load was lifted so that the ball and glass slide were separated (zero load condition); this motion pulled fresh yoghurt into the contact region. The load was then replaced and the test continued. The reciprocating motion was maintained during this procedure although the friction coefficient fell to zero during the load release period. It was possible to repeat this process 3–4 times during the standard test time. This motion replicated the periodic addition of fresh food and was thus an improved representation of the eating process. 3.2. Examination of rubbed food surface films For a limited number of cases at the end of the test the remaining fluid in the rubbed zone on the lower specimen was examined under a low power optical microscope. These films were formed on the glass slide and thus it was possible to use phase contrast imaging to visualise
Fig. 3. HFRR test device and test specimens.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
6
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Table 2 HFRR test conditions. Test parameter
Value
Test condition Load Temperature Time Kinematics Sliding speed
2N (Pmax = 0.25 MPa) 23–25 °C 60 s Reciprocating sliding: 1 mm stroke 10 Hz Mid stroke 20 mm/s
Test specimens Upper Lower
19.8 mm diameter PDMS ball (Duro 300): Young's modulus: 0.0024 GPa Glass microscope slide: Young's modulus 70 GPa
Surface treatment Hydrophobic (HB) Hydrophilic (HL)
None: water contact angle N20° O2 plasma: water contact angle ~0°
different components (Fig. 2). Phase contrast imaging uses polarised light to distinguish between components with similar refractive index, for example emulsion droplets. Images were taken from the rubbed films immediately after the test was completed. These measurements were confined to the 9.5% wt fat yoghurt.
3.3. Test foods: yoghurt A range of commercial yoghurts (same manufacturer: Yeo Valley Farms, different fat content) was tested. Yoghurt is a protein stabilised oil-in-water emulsion, where emulsion droplets are embedded in protein gel matrix. The test samples and given composition are listed in Table 1. The yoghurts were purchased from a national supermarket and new batches bought weekly. All samples were used before their sell-by date expired. They were kept in a refrigerator until required and brought to room temperature before testing when the samples were gently stirred for a few seconds and a small amount taken from the centre of the pot.
4. Results and discussion
Fig. 5. Repeatability of friction results (a) water (solid line) and 4.2% fat yoghurt (dotted line) HB-HB specimens (b) HB-HB: 9.5% wt fat effect of solvent cleaning procedure.
4.1.1.1. 0% wt fat. The low-fat fluids gave rapid increases in friction coefficient with rubbing time. The HB-HB and HL-HL specimens reached their maximum friction after 6 s of rubbing (HB-HB: μ = 0.59, HL-HL μ = 0.52). Both of these values are close to the water friction coefficients (HB-HB: μ = 0.61, HL-HL μ = 0.55 decreasing to 0.47 at 60 s).
4.1. HFRR friction results 4.1.1. Constant load Friction coefficient is plotted against rubbing time in Figs. 5–8. Overall repeatability was very good, example plots are shown in Fig. 5a for HB-HB with water and 4.2% wt fat yoghurt. Problems were encountered in early work with the HB-HB 9.5% tests where two different friction curves were recorded, these are shown in Fig. 5b. The test was found to be very sensitive to surface contamination possibly by residual isopropanol. These results demonstrated the importance of the surface energy condition. Averaged friction coefficient results are plotted against rubbing time for 0, 4.2 and 9% wt fat yoghurt are shown in Figs. 6 (HB-HB) and 7 (HLHL). The HL-HL 4.2% wt fat yoghurt also gave bifurcated results (although this was less distinct than for HB-HB/9.5% wt fat) both types of behaviour are plotted in Fig. 7. The friction results are summarised below:
4.1.1.2. 4.2% wt fat. The 4.2% wt fat results were very different for the two types of surfaces (HB-HB and HL-HL), this can be seen clearly in Fig. 8 which compares the results. HB-HB friction increases rapidly in the first 5 s, drops and then rapidly climbs again to the “water” level of ~μ = 0.6 after 60 s. The HL-HL system also shows an initial increase to μ = 0.35 at 3 s and then progressively drops to give lower friction values of μ = 0.13–0.25. 4.1.1.3. 9.5% wt fat. For HL-HL tests; an initial increase to μ = 0.18 after 5 s followed by a slight decrease and then increase to a maximum at 20 s (μ = 0.22). The friction coefficient values slowly dropped to μ = 0.21 after 60 seconds rubbing. The HB-HB friction coefficient increased slowly from μ = 0.25 at 5 s to μ = 0.54 after 60 s. In some tests low stable friction curves (μ b 0.22) were obtained this appeared to be linked to residual solvent contamination of the test surfaces, this problem is discussed in more detail in Section 4.1.3.
Fig. 4. Photographs of water droplets on HB and HL glass surfaces.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
7
Fig. 6. Averaged friction results for yoghurt samples 0, 4.2, 9.5% wt fat for HB-HB specimens.
4.1.2. Lift load tests Averaged friction coefficient results for load-removal tests with HBHB surfaces are shown in Fig. 9. In these tests the load was lifted momentarily after 30 seconds rubbing and then replaced; this action pulled fresh fluid into the contact zone mimicking the action of eating. The load-lift procedure was repeated 3 times over 120 seconds test time. Two results are shown for 0 and 4.2% wt fat. The initial friction curves obtained over the first 30 s are similar to those shown in Fig. 6. Friction climbs more rapidly for the 0% wt fat sample reaching the “water” value of ~ 0.58–0.6 within 10 s. The 4.2% wt fat friction coefficient increases more slowly to reach μ = 0.57 at 30 s when the load is removed. After load replacement the friction coefficient drops again to a very low value which increases as rubbing progresses. Both samples now show delayed friction increase curves reaching μ N 0.57 after 25 s. The corresponding results for HL-HL surfaces are shown in Fig. 10 for 0 and 4.2% wt fat yoghurts. The 0% wt fat yoghurt did not show the delayed friction rise after load-lift which was observed for the HB-HB tests, if anything the final friction coefficient at 30 s appeared to increase. Two types of behaviour were again observed for the 4.2% wt fat sample and both curves are plotted in Fig. 10. It is interesting that the “high” friction behaviour tends towards the “low” friction curve as the test progresses. This would suggest that progressive deposition of particular food component occurs and in some cases a sufficient amount is not captured at the start of the test. The contact area is fairly small compared to the
Fig. 7. Averaged friction results for yoghurt samples 0, 4.2, 9.5% wt fat for HL-HL surfaces.
Fig. 8. Effect of surface wetting properties (HL-HL and HB-HB) on friction behaviour of 4.2% and 9.5% wt fat samples.
bulk distribution of components and thus in some tests a representative composition is not initially captured. 4.1.3. Effect of cleaning procedure Some test combinations appeared to be very sensitive to the cleaning procedure, and possibly trace levels of solvents, giving very different friction response curves. An example is shown in Fig. 5b for HBHB 9.5% wt fat samples which shows the two different friction curves. In tests were the full cleaning procedure of detergent/water rinse/ isopropanol was used the friction coefficient increased to μ N 0.4 at the end of the test. In tests where only isopropanol was used to clean the
Fig. 9. Average friction results for “lift load” tests; HB-HB for 0 and 4.2% wt fat yoghurt.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
8
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Fig. 10. Average friction results for “lift load” tests; HL-HL for 0 and 4.2% wt fat yoghurt.
surfaces the friction coefficients were lower μ ~ 0.2. Films remaining in the rubbed track were examined for both the “high” and “low” friction condition (Fig. 11). 4.2. Examination of rubbed surface films The phase contrast images shown in Fig. 2b were taken from yoghurt samples sheared between (hydrophobic) glass surfaces under low load, reciprocating sliding. The un-sheared condition is shown on the left hand side. Reciprocating shearing has the effect of forming rolls, presumably of mixed protein/carbohydrates, transverse to the sliding direction. Generally we see sparser “roll” formation for the 0% wt than the 4.2 or 9.5% wt fat yoghurts. The rolls are thought form due the dairy particulates not adhering to the HB glass surfaces. The contribution of these structures to perceived friction is not known. In the friction test images very limited “roll” formation was observed, these are described below. Phase contrast images of the film in the rubbed contact and in the surrounding region on the glass slide are shown in Fig. 11. These results are limited to the 9.5% wt fat yoghurt as the intention was to examine oil droplet release and coalescence which is usually identified as the mechanism contributing to friction reduction [9,11,24]. Images (bar: 100 μm). in Fig. 11 (all 9.5% wt fat yoghurt) include: (a) Fresh yoghurt: Oil droplets (outlined) embedded in protein/carbohydrate matrix (b) HB-HB high friction test (μ = 0.5 at 60 s): in-track region. “Roll” formation in track, very little deposited particulates, small, isolated oil droplets (outlined) (c) HB-HB high friction test (μ = 0.38 at 60 s): in-track region. Some “roll” formation in track, very little deposited particulates, oil droplets (outlined) still embedded in matrix (d) HB-HB low friction test (μ = 0.2 at 60 s): In track coalesced oil droplets spread on particulate layer. These results suggest that the “roll” condition is associated with nonadherence and high friction conditions. In low-friction tests oil droplets are observed to coalesce and spread on a deposited particulate layer. 4.3. Discussion The present paper considers the development of a new approach to understanding the tribology of oral processing where the focus is the on the film loss mechanisms and the way these control friction. Replenishment of food in the tongue-palate interface is considered to be through an unloading/loading cycle rather than shear flow. We consider tribology studies of oral processing must replicate the factors which control film loss and replenishment.
The results for the different yoghurts clearly show very different friction behaviour both in the final friction coefficient and time dependent changes; both of which correlate with the fat content. The initial sharp increase in friction over the first couple of seconds of rubbing is a typical “static friction” behaviour possibly due to adherence of the PDMS to the glass slide. In some cases this is followed by a decrease in friction but in other tests an increase occurs; we suggest this behaviour reflects the competing mechanisms of film loss (giving higher friction) and fat release and deposition (giving lower friction). The increasing friction with rubbing time measured with the 0% wt fat and HB surfaces is thought to be due to progressive loss of the fluid film from the contact and demonstrates the lack of entrainment replenishment. For the HL surfaces (4.2 and 9.5% wt fat) the friction coefficient dropped or stabilised at a low level (μ b 0.3) suggesting a thicker film is retained in the contact. This is expected as higher energy hydrophilic surfaces would be expected to retain both water and dairy protein components. It is also possible that shearing releases oil droplets from the protein matrix and they are deposited on the rubbing track contributing to increased film thickness and lower friction. This behaviour is also indicated by the images from the rubbing films shown in Fig. 11. The deposition and retention of dairy fat in the rubbing contact will presumably contribute to improved “mouthfeel” over a longer period. It is likely that the friction experienced in the initial stages of rubbing is the most representative of “mouth feel” and the results after 5 s of rubbing are summarised in Fig. 12a. Longer term results (60 s) are shown in Fig. 12b. Generally lower friction coefficient correlates with higher fat content as has been reported by a number of tribology studies [7,11]; for example Chojnicka-Paszun et al. [11] in tests on milk of varying fat content. They suggested this was due to shear, surface induced coalescence of the oil droplets, which created a lubricating film between the two surfaces in contact, thereby lowering the friction coefficient value. In tests where the load was lifted and then replaced fresh material was pulled periodically into the contact zone and a better mimic of the sequential eating process obtained. It was noted for the HB-HB 0% and 4.2% wt tests that the friction increase was delayed after the first load-lift step. These results suggest deposition of a residual surface layer in the rubbing contact during the first 30 s which affects subsequent friction behaviour. The deposition and survival of this layer could be controlled to manipulate “mouth feel” properties of foods. Two surface wetting conditions were studied: “hydrophobic” (HBHB) where the water contact angles were N20° and “hydrophilic” (HLHL) where the water contact angles were ~ 0°. Differences were observed in the friction/time behaviour of the HB-HB and HL-HL pairs depending on the fluid composition. In all cases the HL-HL samples gave lower friction coefficients (typically 15–20% less) at the start (5 s) of the test. At the 60 second stage the differences were much greater. The HB-HB samples friction coefficients were close to the water level (μ ~ 0.6). For HL-HL the 0% wt fat samples the final friction coefficient was again close to the water value, however the 4.2 and 9.5% wt fat samples friction coefficients stayed low (μ = 0.2 and 0.22 respectively at 60 s). Although overall the tests showed good repeatability in some cases bifurcated behaviour was observed (see for example Figs. 5b and 7). In some early HB-HB tests with the higher fat content fluids variability in the results appeared to be due to the level of surface cleaning. The differences observed in the HL-HL 4.2% wt fat results might be due to differences in component distribution between the tests. The test uses a very small contact size (1.6 mm) and it is possible it not capture a representative fat distribution for each tests. Work in the future will use a much larger contact area to avoid this problem. Previous studies with hydrophobic or hydrophilic test pairs [9] have generally shown that for O/W emulsions lower friction coefficient are obtained for hydrophobic surfaces. Results for the OTC test method [11] using vegetable oils and O/W emulsions stabilised with whey protein showed reduced friction (at the lowest speeds) for hydrophobic
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Fig. 11. Microscope images from fresh and rubbed yoghurt samples.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
9
10
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Fig. 12. Summary of average friction results yoghurt tests at 5 and 60 s.
(μ = 0.4) compared to hydrophilic surfaces (μ = 0.5). However these results are difficult to interpret as in both cases the friction coefficient for water was less than for the emulsions (water hydrophobic μ = 0.35, hydrophilic μ = 0.38). The results did show increased friction with increasing protein concentration (corresponding results for hydrophilic surfaces were not shown). Dresselhuis et al. [11] also concluded that “Sensorial analyses indicate that the small un-deformable droplets have hardly any influence on perception and the perceived friction is increased instead of decreased.” Friction reduction and thus fat perception is attributed to the ability of the oil droplets to coalesce and spread on the rubbing surfaces [9,24,25]. The results from this study would support these conclusions, in tests with low stable friction traces coalesced droplets are observed in the track (Fig. 11D). However in this case the extended oil phase appears to be spread over dispersed protein/carbohydrate particles, when this layer is absent there is very little oil retained or the droplets are “un-deformed” and pushed to the side of the track. In contrast to the tests reported in [9] lower friction was recorded for the hydrophilic surfaces. We speculate this is because the HL-HL rubbing interface is better at retaining the proteincarbohydrate matrix allowing it to be sheared down and released oil droplets to form. The degraded structure is deposited as a dispersed particulate film in the rubbed track which is overlaid by coalesced oil droplet (Fig. 11D). The particulate film appears to retain oil droplets and possibly aids coalescence; this might be due to increased surface roughness or chemical effects. The tongue is usually considered to be hydrophobic but rendered more hydrophilic by the presence of mucosa glycoproteins [9]. Clearly from our tests the effect of surface roughness and/or protein films play an important role in the deposition and retention of oil films and hence reducing friction and increasing perceived “creaminess”. The developed test has shown friction dependence with rubbing time, discrimination between foods of different fat content and the effect of specimen surface energy. The film in the interfacial region is now determined by depletion rather than entrainment of food components. Thus we must consider further the mechanisms of film loss and
friction change and whether they are relevant to the tongue-palate condition. Film depletion will be influenced by the kinematics (stroke length, reciprocation frequency), contact pressure and surface properties. The use of laboratory tests to simulate more complex systems can be used simply to discriminate between different formulations or more usefully provide insights into mechanisms occurring in real conditions. In conventional tribology tests it is the inlet flow condition which controls film formation and friction but in the mouth it is the interfacial shearing and loss mechanisms which predominate. We consider the new test offers an opportunity for studying in more detail mechanisms occurring in the tongue palate contact. However to do this further improvements are required in the simulation conditions. The friction response with time represents changes in film thickness and component distribution in the conjunction and the fluid composition. The film thickness is determined by the balance between fluid lost and entrained. However the test conditions of short stroke length (1 mm b the contact diameter 1.6 mm) results in reduced entrainment of fresh fluid. Thus the friction change is due to changing composition and preferential loss of some components from the contact as the fluid is shear degraded. This assumes that the initial film captured when the load is applied contains a representative range of constituents. Clearly all of these effects will be influenced by the system parameters; specimen materials, surface properties (hydrophilic/hydrophobic), sliding speed, stroke length, frequency and temperature. We consider this new test approach captures many of the important tribology features of mastication. However this will be improved as the method develop to include improved test specimen properties (surface chemistry and texture, Young's modulus), kinematics (lower frequency reciprocation) and the presence of saliva mucins. We are currently working with PCS Instruments to achieve this. 5. Conclusions Friction experienced in the tongue-palate interface during mastication contributes to our perceptions of taste and texture of food. In this paper the tribology conditions occurring during oral processing have been discussed and a new test method developed. The conclusions from this study are as follows: 1. In the absence of fluid replenishment to the contact the film thickness is dominated by the loss mechanism and friction tests should replicate the appropriate factors (surface properties, kinematics) involved. 2. We suggest food replenishment is controlled by periodic loading and unloading of the tongue palate rather than shear-flow through an inlet region. 3. Under these conditions semi-solid food is trapped within the tonguepalate contact and undergoes degradation and film-thinning due to shear and loading forces. Friction is now time/loss dependent rather than entrainment speed dependent. 4. The tribology test uses a simple, linear reciprocating motion where the stroke length is less than the contact width. Film thickness is determined by the initial capture of components and the subsequent loss from the contact zone. 5. Differences were observed in the friction change with time for different fat contents (μ 9.5 b 4.2 b 0 wt%) and for different surface energy conditions (μ hydrophilic b hydrophobic). Low friction coefficient was associated with the formation of a thin, coherent oil film spread on deposited dairy solids.”
Acknowledgements The authors wish to thank Dr. Clive Hamer of PCS Instruments PLC for helping with the development of this test method.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
References [1] J. Chen, L. Engelen (Eds.), Food Oral Processing: Fundamentals of Eating and Sensory Perception, Wiley-Blackwell, 2012. [2] D.G. Dalgleish, Food emulsions — their structure and structure-forming properties, Food Hydrocoll. 20 (2006) 415–422. [3] E. Dickinson, Interfacial structure and stability of food emulsions as affected by protein–polysaccharide interactions, Soft Matter 4 (2008) 932–942. [4] DairyCo, DairyCo Market Information, http://www.dairyco.org.uk/market-information/dairy-sales-consumption/yogurt-market/yogurt-market/#.U4mndvldWg5; May 30 2014. [5] G. Tabilo-Munizaga, G.V. Barbosa-Cánovas, Rheology for the food industry, J. Food Eng. 67 (2005) 147–156. [6] M.E. Malone, I.A.M. Appelqvist, I.T. Norton, Oral behaviour of food hydrocolloids and emulsions. Part 1. Lubrication and deposition considerations, Food Hydrocoll. 17 (2003) 763–773. [7] N. Selway, J.R. Stokes, Insights into the dynamics of oral lubrication and mouthfeel using soft tribology: differentiating semi-fluid foods with similar rheology, Food Res. Int. 54 (2013) 423–431. [8] S. Prakash, D.D.Y. Tan, Chen, Applications of tribology in studying food oral processing and texture perception, Food Res. Int. 54 (2013) 1627–1635. [9] A. Chojnicka-Paszun, H.H.J. de Jongh, C.G. de Kruif, Sensory perception and lubrication properties of milk: influence of fat content, Int. Dairy J. 26 (2012) 15–22. [10] S. Lee, M. Heuberger, P. Rousset, N.D. Spencer, A tribological model for chocolate in the mouth: general implications for slurry lubricated hard/soft sliding counterfaces, Tribol. Lett. 16 (2004) 239–249. [11] D.M. Dresselhuis, H.J. Klok, M.A. Cohen Stuart, R.J. De Vries, G.A. Van Aken, E.H.A. De Hoog, Tribology of O/W emulsions under mouth-like conditions: determinants of friction, Food Biophys. 2 (2007) 158–171. [12] T. Mills, I.T. Norton, S. Bakalis, Development of tribology equipment to study dynamic processes, J. Food Eng. 114 (2013) 384–390. [13] J.H.H. Bongaerts, K. Fourtouni, J.R. Stokes, Soft-tribology: lubrication in a compliant PDMS–PDMS contact, Tribol. Int. 40 (2007) 1531–1542.
11
[14] J. de Vicente, J.R. Stokes, H.A. Spikes, Soft lubrication of model hydrocolloids, Food Hydrocoll. 20 (2006) 483–491. [15] J.R. Stokes, M.W. Boehm, S.K. Baier, Oral processing, texture and mouthfeel: from Rheology to tribology and beyond, Curr. Opin. Colloid Interface Sci. 18 (2013) 340–359. [16] H. Ranc, C. Servais, P.-F. Chauvy, S. Debaud, S. Mischler, Effect of surface structure on frictional behaviour of a tongue/palate tribological system, Tribol. Int. 39 (2006) 1518–1526. [17] D. Dresselhuis, E. De Hoog, M. Cohen Stuart, G. Van Aken, Application of oral tissue in tribological measurements in an emulsion perception context, Food Hydrocoll. 22 (2008) 323–335. [18] R.A. de Wijk, J.F. Prinz, The role of friction in perceived oral texture, Food Qual. Prefer. 16 (2005) 121–129. [19] K.M. Hiiemae, J.B. Palmer, Tongue movements in feeding and speech, Crit. Rev. Oral Biol. Med. 14 (2003) 413–429. [20] M. Ratoi-Salagean, H.A. Spikes, Optimizing film formation by oil-in-water emulsions, Tribol. Trans. 40 (1997) 569–578. [21] A.A. Lubrecht, D. Mazuyer, P.M. Cann, Starved elastohydrodynamic lubrication theory: application to emulsions and greases, CR Acad. Sci. Paris t.2 (Série IV) (2001) 717–728. [22] S. Hurley, P.M. Cann, Grease composition and film thickness in rolling contacts, NLGI Spokesman 63 (4) (1999) 12–22. [23] S. Giasson, J. Israelachvili, H. Yoshizawa, Thin film morphology and tribology study of mayonnaise, 62 (1997) 640–647. [24] D.M. Dresselhuis, E.H.A. De Hoog, M.A. Cohen Stuart, M.H. Vingerhoeds, G.A. Van Aken, The occurrence of in-mouth coalescence of emulsion droplets in relation to perception of fat, Food Hydrocoll. 22 (2008) 1170–1183. [25] D.M. Dresselhuis, G.A. van Aken, E.H.A. de Hoogac, M.A. Cohen Stuart, Direct observation of adhesion and spreading of emulsion droplets at solid surfaces, Soft Matter 4 (2008) 1079–1085. [26] R.S. Johansson, M. Trulsson, K.A. Olsson, K.-G. Westberg, Mechanoreceptor activity from the human face and oral mucosa, Exp. Brain Res. 72 (1988) 204–208.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
Contents lists available at ScienceDirect
Biotribology journal homepage: http://www.elsevier.com/locate/biotri
Friction measurements with yoghurt in a simulated tongue-palate contact S. Tsui a, J. Tandy a, C. Myant a,b, M. Masen a, P.M. Cann a,⁎ a b
Tribology Group, Department of Mechanical Engineering, Imperial College London, SW7 2AZ, United Kingdom Dyson School of Engineering Design, Imperial College London, United Kingdom
a r t i c l e
i n f o
Article history: Received 2 December 2015 Received in revised form 18 February 2016 Accepted 21 February 2016 Available online xxxx Keywords: Friction Oil/water emulsion Boundary lubrication Oral processing
a b s t r a c t The perception of many food attributes is related to mechanical stimulation and friction experienced in the tongue-palate contact during mastication. Friction in the tongue-palate is determined by the changing film properties (composition, component distribution, thickness) in the conjunction. We suggest this evolution is essentially determined by tongue-palate film loss rather than shear flow entrainment which predominates in conventional bearing lubrication. The paper reports friction measurements in a simulated tongue-palate contact for a range of high and low fat dairy foods. A reciprocating, sliding contact with restricted stroke length (bcontact width) was used; under these conditions there is negligible shear-entrainment of fluid from outside the contact area. The tongue-palate contact was simulated by a PDMS ball and glass surface. The effect of hydrophobic and hydrophilic surfaces on friction was investigated for different fat contents (0, 4.2, 9.5% wt fat). Friction was measured over 60 s of rubbing. Significant differences were observed in the friction change with time for different fat contents (μ 9.5 b μ 4.2 b μ 0 wt%) and for different surface energy conditions (μ hydrophilic b μ hydrophobic). Post-test visualisation of the rubbed films showed that low friction coefficient was associated with the formation of a thin oil film on deposited particulate solids. © 2016 Published by Elsevier Ltd.
1. Introduction Oral processing or mastication is a dynamic process whereby food structure is broken down and transported to the pharynx prior to swallowing. It is a complex series of processes which involve a number of surface interactions including tooth-tooth and tongue-palate mechanisms [1]. Sensory perception of food involves visual, physical and physiological elements in which the role of structure, rheology and the mechanisms of food breakdown are not fully understood. The perception of taste and texture which includes thickness, smoothness, slipperiness [1] and to some extent astringency are usually related to mechanical stimulation and hence friction experienced in the mouth. Semi-solid food, which includes cream, yoghurt, custard and mayonnaise, are oil-in-water (O/W) emulsions which also contain proteins, emulsifiers, carbohydrates (including sugar) and other soluble components. Most formulation work has centred on research by food scientists [2,3] and the development of complex emulsions and micro-phase fluids. The focus has been on controlling the chemical and physical characteristics to improve the stability and storage properties [2]. However in recent years the role of food composition and structure in oral processing has received increasing attention. As Dalgleish [2] states “These structures in turn give rise to the perception of texture as they are consumed”. The development of low and zero-fat food with the mouth⁎ Corresponding author. E-mail address: [email protected] (P.M. Cann).
feel of high fat content analogues remains an important objective of the food industry [2,3]. The emphasis has been the role of fat in the original product structure [2] and how this evolves during mastication. The low/very low fat yoghurt market in the UK is currently worth over £500 million/year [4] and is growing rapidly. However dietary focus has recently switched to the reduction of sugar in foods and unfortunately many low-fat formulations contain higher levels of sugar (see Table 1). During mastication the structure of these foods change and the contribution of all components, not just fats, to perceived taste and texture must be understood if we are to effectively develop new products that satisfy consumer expectations. The problem remains of how to quantify the mouth-feel of these products without excessive use of panel tests. Mouth-feel is the tactile sensation created when food is rubbed between the tongue and palate, up to the point of swallow. Many studies [5] have tried to link oral perception to bulk rheology of the foods however this method does not capture many important aspects of the problem. As Malone et al. [6] concluded “thin film tribological properties correlated more closely with the composition and texture of the samples than the bulk rheological properties”. When food is masticated the material in the tongue-palate conjunction is sheared and reduces so that the surface and thin film properties and the structural evolution, rather than the bulk fluid, determine oral perception. Selway and Stokes [7] note “The change in lengthscale (film thickness between oral surfaces) is identified as a key feature driving the transition from a regime where the bulk fluid properties govern the sensory response, to one where surface interactions dominate.”
http://dx.doi.org/10.1016/j.biotri.2016.02.001 2352-5738/© 2016 Published by Elsevier Ltd.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
2
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Table 1 Composition of test fluids. Composition/100 g
0% wt fat
4.2% wt fat
9.5% wt fat
Total fat of which saturates Carbohydrates Of which sugars Protein
0g 0g 8.5 g 8.5 g 5.9 g
4.2 g 2.5 g 4.2 g 2.7 g 4.6 g
9.5 g 5.9 g 6.4 g 6.4 g 4.58 g
In recent years there has been a growing number of studies [6–18] reporting the tribological evaluation of semi-solid foods (e.g. yoghurt, salad dressings, milk, chocolate) with the intention of correlating measured friction with oral perception and to provide more relevant assessment of new products. The approach has been very successful as Chojnicka-Paszu et al. [9] concluded from their study of milk “indicated a good correlation between creamy attributes and measured friction coefficient, a result that validates the use of tribology as an analytical technique to better programme specific sensory products in product development and reformulation”. Research has usually focussed on correlation of friction behaviour with fat content [6,9,18] of the sample however in future we must be able to relate measured friction to the wider chemical composition and component distribution properties of the lubricating film. A thorough review of the topic is provided by Stokes et al. [15] who consider more general aspects of the relationship between rheology, lubrication and oral processing. The focus of the current paper is to reexamine the tribology of the tongue-palate and development of improved test methodologies to better simulate oral processing.
Food tribology tests are usually run as “Stribeck” curves [6,7,12–14,] with average friction measured over a range of speeds. The technique has been successfully used to distinguish between different fat contents in semi-solid foods including yoghurt [7], milk [9] and O/W emulsion [6, 11]. Typical examples of MTM speed-sweep curves carried out in our laboratory are shown in Fig. 1a. The friction coefficient in the lowspeed region (usually b50 mm/s) decreases with increasing fat content as has been reported in other studies [6,7,9]. For the speed-sweep results friction coefficient at 20 mm/s is ranked 9.5% fat (μ = 0.025) b4.2% fat (μ = 0.04) b 0% fat (μ = 0.15). However measurements at constant speed (20 mm/s) gave different results for the 9.5% fat yoghurt (average μ = 0.09) as shown in Fig. 1b. This result was unexpected and possibly indicates that the shear history experienced by the food is important in determining friction response. The friction/speeds curves are usually interpreted in terms of classical fully-flooded lubrication regimes (boundary, mixed, full film) denoted by changes in friction coefficient [6,7,8,11–15]. For the O/W emulsions the reduction in friction at slow speeds is attributed to the preferential entrainment of oil [4,7] and the formation of an oil “boundary” film [8] on the contacting surfaces. For example Selway and Stokes [7] concluded “The medium- and high-fat products both generate friction curves identical to that of the pure oil phase over the entire range of speeds measured, indicating preferential entrainment of oil and exclusion of the thickened aqueous phase from the contact zone.” Thus in these tests the conditions experienced by the fluid as it is entrained through the inlet into the contact are critical in determining both the composition of the film formed (“preferential entrainment of oil”) and the resulting friction behaviour. Thus the film in the contact zone is not representative of multiphase composition of the food.
2. Research into food tribology 2.2. Tribology analysis of oral processing 2.1. Conventional tribology test methods A variety of test devices have been used to study friction properties of foods, Prakash et al. [8] provide a recent review. Most studies of fluid or semi-solid food tribology [6,7,9,12,13,14] have used the Mini Traction Machine (MTM) or Optical Tribological Configuration (OTC) [11,17]. The MTM uses a ball-on-disc contact immersed in excess of fluid and measures friction coefficient as a function of different test conditions including entrainment speed, slide-roll ratio and temperature. The OTC uses a pin (PDMS or other soft surface) reciprocating against a glass counterface [11,17]. Lee et al. [10] also used a pin-on-disc (unidirectional sliding) device (CSM Instruments) to study the tribology properties of molten chocolate. In tribology terms the tongue-palate contact is characterised as a soft, rough, viscoelastic specimen (tongue) rubbing against a harder, smooth counterface (palate). The contact pressure is generally low (~ MPa) and the sliding speeds in the range 10–30 mm/s to match tongue movement in eating [19]. The various test methods described in the literature use a wide range of materials (polymer, zirconia, glass, steel, porcine tongue), specimen surface properties (smooth/textured, hydrophobic/hydrophilic) contact conditions (pressures, temperature, speed range) and kinematics (sliding/rolling/reciprocation). The test specimen materials are chosen to simulate the tongue-palate contact and include “soft/soft” (polymer/polymer) [7] and “soft/hard” (polymer/steel or glass, porcine tongue/glass) [11,12,13,14,17] combinations. The kinematic condition is particularly important as this plays a significant role in determining lubricant film properties. The MTM test has usually been used in a rolling-sliding configuration (50% SRR) where both the ball and disc are driven at different speeds [6,12–15]. Other studies have used simple sliding [10] or reciprocating sliding [11,16,17]. In some cases very high speeds (~ 500 mm/s [12]) were used, however the usual range is 5–80 mm/s [10,11,17]. This is supported Hiiemae and Palmer [19] who quote an average measured speed range of 10–32 mm/s for tongue movement during eating
The preceding discussion has emphasised the role of the inlet in determining the film formation and friction response of two-phase (emulsions) fluids. Lubricant is entrained into the contact region by the relative motion of the surfaces and thus forms a film which supports the load and separates the surfaces. A fundamental principle of tribology is that under normally lubricated conditions i.e. where there is continued supply of lubricant it is the properties (viscosity, composition) of the fluid in the inlet region that determine the lubricating film properties (chemistry, thickness) in the contact region. The shear conditions in the inlet (due to narrowing gap height) are severe and even for modest speeds (mm/s) can reach 105 s−1 for the classical ball/disc arrangement. This can lead to shear thinning or degradation, phase and compositional changes, particularly for emulsions, prior to the contact (Fig. 2a left image) and is clearly not representative of the eating process. Semi-solid foods are placed in the oral cavity and broken down either in the tooth-tooth or tongue-palate rubbing contacts. There is no continued flow of fresh material into the contact and a representative distribution of components is present at the start of mastication. The role of the inlet in determining film properties in the contact has been extensively studied in classical (oil-based) lubrication [20–22]. O/ W fluids are used as metal-working lubricants where shear degradation and coalescence of oil droplets in the inlet occurs to provide an oil-rich reservoir which supplies the lubricating film in the contact zone [20,21]. At low speeds the film in the contact is predominately composed of oil giving a thicker film and lower friction than the continuous water phase [20]. At higher speeds replenishment of the oil reservoir is not maintained and entrainment of the bulk fluid, which is predominately water, occurs [20,21]. The film thickness of industrial O/W emulsions often drop with increasing speed because of this effect [20]. Thus the oil/water composition in the contact will change depending on the properties of the fluid and the ability of the contact to generate and retain the oil reservoir (materials, surface energy, entrainment speed, chemical composition) [20,21].
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
3
Fig. 1. MTM test results for 0, 4.2 and 9% w/w fat yoghurts, 50% SRR, 35 °C, excess fluid (a) speed sweep results (b) constant speed as a function of time at 20 mm/s.
In addition to the oil/water phase semi-fluid foods often contain a dispersed solid phase (e.g. proteins) where the nearest engineering equivalent is lubricating grease which is a two-phase system (base oil/ thickener). Tribology studies of grease has shown complex lubrication behaviour where the film properties depend on the entrainment speed and lubricant supply (starvation) to the inlet [22]. At slow speeds much thicker films than expected are formed due to preferential entrainment of thickener particles [22]. At higher speeds starvation of the lubricant supply may occur or shear-degradation of the thickener matrix [22]. Thus in the usual test configuration (ball/disc) where a complex (multi-phase) fluid is entrained by the relative movement of the surfaces the chemical composition of the film in the contact is likely to be different to the bulk fluid. As a result lubricating film properties (thickness, friction) may change in a complex fashion depending on the inlet flow condition and local fluid composition. As all foods contain a mixture of components and in some cases phases it is impossible to interpret friction data, unless we understand more about the composition of the film in the contact zone and the effect of inlet flow. Although correlations have been made between food composition and friction response [7,8] it is difficult to foresee how the traditional lubrication test
can be developed further to provide a more relevant research tool without this information. Although the foregoing argument has emphasised the role of the inlet in the conventional test configuration this is not necessarily relevant to the oral processing problem as the inlet to the contact is effectively absent. During eating food is captured in the tongue-palate under load and then repeatedly sheared (Fig. 2a right image), thus the mixture of food components initially present is that occurring in the bulk material. The distribution and condition of food components in the tongue-palate conjunction is not filtered or sheared by flow through the converging gap of the inlet. Under repeated loading and shear the two/three-phase structure will break down and the distribution of food components changes due to surface adsorption, agglomeration, degradation or selective loss [23,24,25]. The change in component distribution due to simple shear can be seen clearly in the phase contrast microscopy images shown in Fig. 2b. In these images a small sample of yoghurt was sheared under simple, reciprocating sliding between two glass slides. The formation of “roll” structures can clearly be seen transverse to the sliding direction. These images are discussed in more detail later in the paper. In conclusion we suggest it is the loss
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
4
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Fig. 2. Inlet-controlled and zero-entrainment lubrication conditions and phase contrast microscopy images of sheared emulsion films.
mechanisms occurring during mastication rather than shearentrainment and inlet “filtering” which determine film thickness and component distribution in the tongue-palate contact and hence the evolution of friction. A second consideration is the timescale of the measurements; food is introduced into the tongue-palate and sensory information is immediately recorded during mastication. The dwell time of food in the mouth is short and the friction experienced in the first few seconds as the bolus thins which will likely determine the sensory assessment. Both the MTM and OTC tests can be used to provide friction change with rubbing time (Fig. 1b) however the time scale is much longer than experienced in normal mastication [7,15,9]. When considering the design and applicability of tribology tests for assessing mouth-feel it is instructive to consider the oral mechano-stimulation process. Mechano-receptors at the oral surface detect local friction and pressures changes at the food-mucosa interface during mastication [26]. The timescale of receptor response will be important when assessing the influence of friction changes contributing to oral perception. The development of new test methods which capture the time dependency of textural changes has also been discussed in a recent review paper by Stokes et al. [13].
In conclusion friction experienced in the tongue-palate is determined by the composition, component distribution and film thickness in the conjunction, however in traditional tribology tests these properties are controlled by the inlet flow. During mastication fresh food is supplied by periodically separating the tongue-palate interface; effectively this is a transient loading condition. Thus film thickness is predominately determined by frequency of this event and the tonguepalate food loss mechanisms rather than shear flow entrainment. The intention in this study is to develop a test method which replicates the tongue-palate loss and resupply mechanisms. 2.3. New tribology test method In this paper a new approach has been developed where friction change for a food sample is measured over the first 60 s of rubbing. A sliding, reciprocating test device (HFRR High Frequency Reciprocating Rig PCS Instrument UK) is used. Ranc et al. [16] used a similar test method to study the effect of PDMS surface texture (simulating the tongue) on friction. In the standard configuration the HFRR utilises a ball sliding against a flat disc. The test specimens are held in a temperature controlled bath. The major advantage is that a small amount of sample is
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
required and this is subject to repeated shearing. Friction measurements were taken every second. Food held in the tongue-palate interface is sheared repeatedly during mastication, thus the structure and distribution of food components will change and this will result in varying friction. Time-dependent changes particularly in the initial stages of mastication are likely to contribute to the perception of texture. In the first instance a simplified form of the test is reported using PDMS/glass as a tongue/palate model [16] operating at room temperature [18] and in the absence of saliva mucins [12]. However as the test develops these factors will be changed to improve the simulation of the tongue-palate oral processing. The Hertzian contact diameter (1.6 mm for 2N) is larger than the stroke length (1 mm) and thus reciprocation does not entrain significant amounts of fresh fluid into the contact zone making it analogous to movement of the tongue against the palate. In initial tests a constant load condition was used, in later work the load was removed every 30 s which represents a more realistic tongue-palate re-supply condition. A series of commercial dairy foods (yoghurt) with different fat contents was studied. 3. Experimental 3.1. HFRR tests The HFRR uses a reciprocating, sliding contact where the upper specimen is loaded and rubs against a lower disc. The load (2N), temperature (25 °C), stroke length (1 mm) and reciprocation frequency (10 Hz) are controlled by a computer. Friction was measured by a force transducer mounted on the lower specimen support. The test device uses a 19.05 mm diameter PDMS ball (Duro 30 supplied by PCS
5
Instruments) and glass microscope slide (VWR Premium Grade). A schematic diagram and photograph of the test device are shown in Fig. 3. Test conditions are summarised in Table 2. Two series of tests were carried out with either “hydrophobic” or “hydrophilic” surfaces. For hydrophobic (HB-HB) tests both specimens were cleaned in detergent solution, distilled water and then isopropyl alcohol (Sigma Aldrich Analar Grade). For the hydrophilic (HL-HL) tests both specimens were cleaned in isopropyl alcohol followed by treatment in a plasma cleaner. This procedure is used to remove organic contaminants and form an oxidised, hydrophilic surface [11]. All plasma-modified specimens were tested within 5 min of treatment. Photographs of water droplets on glass surfaces before and after plasma treatment are shown in Fig. 4. In a number of tests the effect of repeated addition of fresh food was investigated using a “lift-load” procedure. The tests were run as usual except after 30 s the load was lifted so that the ball and glass slide were separated (zero load condition); this motion pulled fresh yoghurt into the contact region. The load was then replaced and the test continued. The reciprocating motion was maintained during this procedure although the friction coefficient fell to zero during the load release period. It was possible to repeat this process 3–4 times during the standard test time. This motion replicated the periodic addition of fresh food and was thus an improved representation of the eating process. 3.2. Examination of rubbed food surface films For a limited number of cases at the end of the test the remaining fluid in the rubbed zone on the lower specimen was examined under a low power optical microscope. These films were formed on the glass slide and thus it was possible to use phase contrast imaging to visualise
Fig. 3. HFRR test device and test specimens.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
6
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Table 2 HFRR test conditions. Test parameter
Value
Test condition Load Temperature Time Kinematics Sliding speed
2N (Pmax = 0.25 MPa) 23–25 °C 60 s Reciprocating sliding: 1 mm stroke 10 Hz Mid stroke 20 mm/s
Test specimens Upper Lower
19.8 mm diameter PDMS ball (Duro 300): Young's modulus: 0.0024 GPa Glass microscope slide: Young's modulus 70 GPa
Surface treatment Hydrophobic (HB) Hydrophilic (HL)
None: water contact angle N20° O2 plasma: water contact angle ~0°
different components (Fig. 2). Phase contrast imaging uses polarised light to distinguish between components with similar refractive index, for example emulsion droplets. Images were taken from the rubbed films immediately after the test was completed. These measurements were confined to the 9.5% wt fat yoghurt.
3.3. Test foods: yoghurt A range of commercial yoghurts (same manufacturer: Yeo Valley Farms, different fat content) was tested. Yoghurt is a protein stabilised oil-in-water emulsion, where emulsion droplets are embedded in protein gel matrix. The test samples and given composition are listed in Table 1. The yoghurts were purchased from a national supermarket and new batches bought weekly. All samples were used before their sell-by date expired. They were kept in a refrigerator until required and brought to room temperature before testing when the samples were gently stirred for a few seconds and a small amount taken from the centre of the pot.
4. Results and discussion
Fig. 5. Repeatability of friction results (a) water (solid line) and 4.2% fat yoghurt (dotted line) HB-HB specimens (b) HB-HB: 9.5% wt fat effect of solvent cleaning procedure.
4.1.1.1. 0% wt fat. The low-fat fluids gave rapid increases in friction coefficient with rubbing time. The HB-HB and HL-HL specimens reached their maximum friction after 6 s of rubbing (HB-HB: μ = 0.59, HL-HL μ = 0.52). Both of these values are close to the water friction coefficients (HB-HB: μ = 0.61, HL-HL μ = 0.55 decreasing to 0.47 at 60 s).
4.1. HFRR friction results 4.1.1. Constant load Friction coefficient is plotted against rubbing time in Figs. 5–8. Overall repeatability was very good, example plots are shown in Fig. 5a for HB-HB with water and 4.2% wt fat yoghurt. Problems were encountered in early work with the HB-HB 9.5% tests where two different friction curves were recorded, these are shown in Fig. 5b. The test was found to be very sensitive to surface contamination possibly by residual isopropanol. These results demonstrated the importance of the surface energy condition. Averaged friction coefficient results are plotted against rubbing time for 0, 4.2 and 9% wt fat yoghurt are shown in Figs. 6 (HB-HB) and 7 (HLHL). The HL-HL 4.2% wt fat yoghurt also gave bifurcated results (although this was less distinct than for HB-HB/9.5% wt fat) both types of behaviour are plotted in Fig. 7. The friction results are summarised below:
4.1.1.2. 4.2% wt fat. The 4.2% wt fat results were very different for the two types of surfaces (HB-HB and HL-HL), this can be seen clearly in Fig. 8 which compares the results. HB-HB friction increases rapidly in the first 5 s, drops and then rapidly climbs again to the “water” level of ~μ = 0.6 after 60 s. The HL-HL system also shows an initial increase to μ = 0.35 at 3 s and then progressively drops to give lower friction values of μ = 0.13–0.25. 4.1.1.3. 9.5% wt fat. For HL-HL tests; an initial increase to μ = 0.18 after 5 s followed by a slight decrease and then increase to a maximum at 20 s (μ = 0.22). The friction coefficient values slowly dropped to μ = 0.21 after 60 seconds rubbing. The HB-HB friction coefficient increased slowly from μ = 0.25 at 5 s to μ = 0.54 after 60 s. In some tests low stable friction curves (μ b 0.22) were obtained this appeared to be linked to residual solvent contamination of the test surfaces, this problem is discussed in more detail in Section 4.1.3.
Fig. 4. Photographs of water droplets on HB and HL glass surfaces.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
7
Fig. 6. Averaged friction results for yoghurt samples 0, 4.2, 9.5% wt fat for HB-HB specimens.
4.1.2. Lift load tests Averaged friction coefficient results for load-removal tests with HBHB surfaces are shown in Fig. 9. In these tests the load was lifted momentarily after 30 seconds rubbing and then replaced; this action pulled fresh fluid into the contact zone mimicking the action of eating. The load-lift procedure was repeated 3 times over 120 seconds test time. Two results are shown for 0 and 4.2% wt fat. The initial friction curves obtained over the first 30 s are similar to those shown in Fig. 6. Friction climbs more rapidly for the 0% wt fat sample reaching the “water” value of ~ 0.58–0.6 within 10 s. The 4.2% wt fat friction coefficient increases more slowly to reach μ = 0.57 at 30 s when the load is removed. After load replacement the friction coefficient drops again to a very low value which increases as rubbing progresses. Both samples now show delayed friction increase curves reaching μ N 0.57 after 25 s. The corresponding results for HL-HL surfaces are shown in Fig. 10 for 0 and 4.2% wt fat yoghurts. The 0% wt fat yoghurt did not show the delayed friction rise after load-lift which was observed for the HB-HB tests, if anything the final friction coefficient at 30 s appeared to increase. Two types of behaviour were again observed for the 4.2% wt fat sample and both curves are plotted in Fig. 10. It is interesting that the “high” friction behaviour tends towards the “low” friction curve as the test progresses. This would suggest that progressive deposition of particular food component occurs and in some cases a sufficient amount is not captured at the start of the test. The contact area is fairly small compared to the
Fig. 7. Averaged friction results for yoghurt samples 0, 4.2, 9.5% wt fat for HL-HL surfaces.
Fig. 8. Effect of surface wetting properties (HL-HL and HB-HB) on friction behaviour of 4.2% and 9.5% wt fat samples.
bulk distribution of components and thus in some tests a representative composition is not initially captured. 4.1.3. Effect of cleaning procedure Some test combinations appeared to be very sensitive to the cleaning procedure, and possibly trace levels of solvents, giving very different friction response curves. An example is shown in Fig. 5b for HBHB 9.5% wt fat samples which shows the two different friction curves. In tests were the full cleaning procedure of detergent/water rinse/ isopropanol was used the friction coefficient increased to μ N 0.4 at the end of the test. In tests where only isopropanol was used to clean the
Fig. 9. Average friction results for “lift load” tests; HB-HB for 0 and 4.2% wt fat yoghurt.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
8
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Fig. 10. Average friction results for “lift load” tests; HL-HL for 0 and 4.2% wt fat yoghurt.
surfaces the friction coefficients were lower μ ~ 0.2. Films remaining in the rubbed track were examined for both the “high” and “low” friction condition (Fig. 11). 4.2. Examination of rubbed surface films The phase contrast images shown in Fig. 2b were taken from yoghurt samples sheared between (hydrophobic) glass surfaces under low load, reciprocating sliding. The un-sheared condition is shown on the left hand side. Reciprocating shearing has the effect of forming rolls, presumably of mixed protein/carbohydrates, transverse to the sliding direction. Generally we see sparser “roll” formation for the 0% wt than the 4.2 or 9.5% wt fat yoghurts. The rolls are thought form due the dairy particulates not adhering to the HB glass surfaces. The contribution of these structures to perceived friction is not known. In the friction test images very limited “roll” formation was observed, these are described below. Phase contrast images of the film in the rubbed contact and in the surrounding region on the glass slide are shown in Fig. 11. These results are limited to the 9.5% wt fat yoghurt as the intention was to examine oil droplet release and coalescence which is usually identified as the mechanism contributing to friction reduction [9,11,24]. Images (bar: 100 μm). in Fig. 11 (all 9.5% wt fat yoghurt) include: (a) Fresh yoghurt: Oil droplets (outlined) embedded in protein/carbohydrate matrix (b) HB-HB high friction test (μ = 0.5 at 60 s): in-track region. “Roll” formation in track, very little deposited particulates, small, isolated oil droplets (outlined) (c) HB-HB high friction test (μ = 0.38 at 60 s): in-track region. Some “roll” formation in track, very little deposited particulates, oil droplets (outlined) still embedded in matrix (d) HB-HB low friction test (μ = 0.2 at 60 s): In track coalesced oil droplets spread on particulate layer. These results suggest that the “roll” condition is associated with nonadherence and high friction conditions. In low-friction tests oil droplets are observed to coalesce and spread on a deposited particulate layer. 4.3. Discussion The present paper considers the development of a new approach to understanding the tribology of oral processing where the focus is the on the film loss mechanisms and the way these control friction. Replenishment of food in the tongue-palate interface is considered to be through an unloading/loading cycle rather than shear flow. We consider tribology studies of oral processing must replicate the factors which control film loss and replenishment.
The results for the different yoghurts clearly show very different friction behaviour both in the final friction coefficient and time dependent changes; both of which correlate with the fat content. The initial sharp increase in friction over the first couple of seconds of rubbing is a typical “static friction” behaviour possibly due to adherence of the PDMS to the glass slide. In some cases this is followed by a decrease in friction but in other tests an increase occurs; we suggest this behaviour reflects the competing mechanisms of film loss (giving higher friction) and fat release and deposition (giving lower friction). The increasing friction with rubbing time measured with the 0% wt fat and HB surfaces is thought to be due to progressive loss of the fluid film from the contact and demonstrates the lack of entrainment replenishment. For the HL surfaces (4.2 and 9.5% wt fat) the friction coefficient dropped or stabilised at a low level (μ b 0.3) suggesting a thicker film is retained in the contact. This is expected as higher energy hydrophilic surfaces would be expected to retain both water and dairy protein components. It is also possible that shearing releases oil droplets from the protein matrix and they are deposited on the rubbing track contributing to increased film thickness and lower friction. This behaviour is also indicated by the images from the rubbing films shown in Fig. 11. The deposition and retention of dairy fat in the rubbing contact will presumably contribute to improved “mouthfeel” over a longer period. It is likely that the friction experienced in the initial stages of rubbing is the most representative of “mouth feel” and the results after 5 s of rubbing are summarised in Fig. 12a. Longer term results (60 s) are shown in Fig. 12b. Generally lower friction coefficient correlates with higher fat content as has been reported by a number of tribology studies [7,11]; for example Chojnicka-Paszun et al. [11] in tests on milk of varying fat content. They suggested this was due to shear, surface induced coalescence of the oil droplets, which created a lubricating film between the two surfaces in contact, thereby lowering the friction coefficient value. In tests where the load was lifted and then replaced fresh material was pulled periodically into the contact zone and a better mimic of the sequential eating process obtained. It was noted for the HB-HB 0% and 4.2% wt tests that the friction increase was delayed after the first load-lift step. These results suggest deposition of a residual surface layer in the rubbing contact during the first 30 s which affects subsequent friction behaviour. The deposition and survival of this layer could be controlled to manipulate “mouth feel” properties of foods. Two surface wetting conditions were studied: “hydrophobic” (HBHB) where the water contact angles were N20° and “hydrophilic” (HLHL) where the water contact angles were ~ 0°. Differences were observed in the friction/time behaviour of the HB-HB and HL-HL pairs depending on the fluid composition. In all cases the HL-HL samples gave lower friction coefficients (typically 15–20% less) at the start (5 s) of the test. At the 60 second stage the differences were much greater. The HB-HB samples friction coefficients were close to the water level (μ ~ 0.6). For HL-HL the 0% wt fat samples the final friction coefficient was again close to the water value, however the 4.2 and 9.5% wt fat samples friction coefficients stayed low (μ = 0.2 and 0.22 respectively at 60 s). Although overall the tests showed good repeatability in some cases bifurcated behaviour was observed (see for example Figs. 5b and 7). In some early HB-HB tests with the higher fat content fluids variability in the results appeared to be due to the level of surface cleaning. The differences observed in the HL-HL 4.2% wt fat results might be due to differences in component distribution between the tests. The test uses a very small contact size (1.6 mm) and it is possible it not capture a representative fat distribution for each tests. Work in the future will use a much larger contact area to avoid this problem. Previous studies with hydrophobic or hydrophilic test pairs [9] have generally shown that for O/W emulsions lower friction coefficient are obtained for hydrophobic surfaces. Results for the OTC test method [11] using vegetable oils and O/W emulsions stabilised with whey protein showed reduced friction (at the lowest speeds) for hydrophobic
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Fig. 11. Microscope images from fresh and rubbed yoghurt samples.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
9
10
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
Fig. 12. Summary of average friction results yoghurt tests at 5 and 60 s.
(μ = 0.4) compared to hydrophilic surfaces (μ = 0.5). However these results are difficult to interpret as in both cases the friction coefficient for water was less than for the emulsions (water hydrophobic μ = 0.35, hydrophilic μ = 0.38). The results did show increased friction with increasing protein concentration (corresponding results for hydrophilic surfaces were not shown). Dresselhuis et al. [11] also concluded that “Sensorial analyses indicate that the small un-deformable droplets have hardly any influence on perception and the perceived friction is increased instead of decreased.” Friction reduction and thus fat perception is attributed to the ability of the oil droplets to coalesce and spread on the rubbing surfaces [9,24,25]. The results from this study would support these conclusions, in tests with low stable friction traces coalesced droplets are observed in the track (Fig. 11D). However in this case the extended oil phase appears to be spread over dispersed protein/carbohydrate particles, when this layer is absent there is very little oil retained or the droplets are “un-deformed” and pushed to the side of the track. In contrast to the tests reported in [9] lower friction was recorded for the hydrophilic surfaces. We speculate this is because the HL-HL rubbing interface is better at retaining the proteincarbohydrate matrix allowing it to be sheared down and released oil droplets to form. The degraded structure is deposited as a dispersed particulate film in the rubbed track which is overlaid by coalesced oil droplet (Fig. 11D). The particulate film appears to retain oil droplets and possibly aids coalescence; this might be due to increased surface roughness or chemical effects. The tongue is usually considered to be hydrophobic but rendered more hydrophilic by the presence of mucosa glycoproteins [9]. Clearly from our tests the effect of surface roughness and/or protein films play an important role in the deposition and retention of oil films and hence reducing friction and increasing perceived “creaminess”. The developed test has shown friction dependence with rubbing time, discrimination between foods of different fat content and the effect of specimen surface energy. The film in the interfacial region is now determined by depletion rather than entrainment of food components. Thus we must consider further the mechanisms of film loss and
friction change and whether they are relevant to the tongue-palate condition. Film depletion will be influenced by the kinematics (stroke length, reciprocation frequency), contact pressure and surface properties. The use of laboratory tests to simulate more complex systems can be used simply to discriminate between different formulations or more usefully provide insights into mechanisms occurring in real conditions. In conventional tribology tests it is the inlet flow condition which controls film formation and friction but in the mouth it is the interfacial shearing and loss mechanisms which predominate. We consider the new test offers an opportunity for studying in more detail mechanisms occurring in the tongue palate contact. However to do this further improvements are required in the simulation conditions. The friction response with time represents changes in film thickness and component distribution in the conjunction and the fluid composition. The film thickness is determined by the balance between fluid lost and entrained. However the test conditions of short stroke length (1 mm b the contact diameter 1.6 mm) results in reduced entrainment of fresh fluid. Thus the friction change is due to changing composition and preferential loss of some components from the contact as the fluid is shear degraded. This assumes that the initial film captured when the load is applied contains a representative range of constituents. Clearly all of these effects will be influenced by the system parameters; specimen materials, surface properties (hydrophilic/hydrophobic), sliding speed, stroke length, frequency and temperature. We consider this new test approach captures many of the important tribology features of mastication. However this will be improved as the method develop to include improved test specimen properties (surface chemistry and texture, Young's modulus), kinematics (lower frequency reciprocation) and the presence of saliva mucins. We are currently working with PCS Instruments to achieve this. 5. Conclusions Friction experienced in the tongue-palate interface during mastication contributes to our perceptions of taste and texture of food. In this paper the tribology conditions occurring during oral processing have been discussed and a new test method developed. The conclusions from this study are as follows: 1. In the absence of fluid replenishment to the contact the film thickness is dominated by the loss mechanism and friction tests should replicate the appropriate factors (surface properties, kinematics) involved. 2. We suggest food replenishment is controlled by periodic loading and unloading of the tongue palate rather than shear-flow through an inlet region. 3. Under these conditions semi-solid food is trapped within the tonguepalate contact and undergoes degradation and film-thinning due to shear and loading forces. Friction is now time/loss dependent rather than entrainment speed dependent. 4. The tribology test uses a simple, linear reciprocating motion where the stroke length is less than the contact width. Film thickness is determined by the initial capture of components and the subsequent loss from the contact zone. 5. Differences were observed in the friction change with time for different fat contents (μ 9.5 b 4.2 b 0 wt%) and for different surface energy conditions (μ hydrophilic b hydrophobic). Low friction coefficient was associated with the formation of a thin, coherent oil film spread on deposited dairy solids.”
Acknowledgements The authors wish to thank Dr. Clive Hamer of PCS Instruments PLC for helping with the development of this test method.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001
S. Tsui et al. / Biotribology xxx (2016) xxx–xxx
References [1] J. Chen, L. Engelen (Eds.), Food Oral Processing: Fundamentals of Eating and Sensory Perception, Wiley-Blackwell, 2012. [2] D.G. Dalgleish, Food emulsions — their structure and structure-forming properties, Food Hydrocoll. 20 (2006) 415–422. [3] E. Dickinson, Interfacial structure and stability of food emulsions as affected by protein–polysaccharide interactions, Soft Matter 4 (2008) 932–942. [4] DairyCo, DairyCo Market Information, http://www.dairyco.org.uk/market-information/dairy-sales-consumption/yogurt-market/yogurt-market/#.U4mndvldWg5; May 30 2014. [5] G. Tabilo-Munizaga, G.V. Barbosa-Cánovas, Rheology for the food industry, J. Food Eng. 67 (2005) 147–156. [6] M.E. Malone, I.A.M. Appelqvist, I.T. Norton, Oral behaviour of food hydrocolloids and emulsions. Part 1. Lubrication and deposition considerations, Food Hydrocoll. 17 (2003) 763–773. [7] N. Selway, J.R. Stokes, Insights into the dynamics of oral lubrication and mouthfeel using soft tribology: differentiating semi-fluid foods with similar rheology, Food Res. Int. 54 (2013) 423–431. [8] S. Prakash, D.D.Y. Tan, Chen, Applications of tribology in studying food oral processing and texture perception, Food Res. Int. 54 (2013) 1627–1635. [9] A. Chojnicka-Paszun, H.H.J. de Jongh, C.G. de Kruif, Sensory perception and lubrication properties of milk: influence of fat content, Int. Dairy J. 26 (2012) 15–22. [10] S. Lee, M. Heuberger, P. Rousset, N.D. Spencer, A tribological model for chocolate in the mouth: general implications for slurry lubricated hard/soft sliding counterfaces, Tribol. Lett. 16 (2004) 239–249. [11] D.M. Dresselhuis, H.J. Klok, M.A. Cohen Stuart, R.J. De Vries, G.A. Van Aken, E.H.A. De Hoog, Tribology of O/W emulsions under mouth-like conditions: determinants of friction, Food Biophys. 2 (2007) 158–171. [12] T. Mills, I.T. Norton, S. Bakalis, Development of tribology equipment to study dynamic processes, J. Food Eng. 114 (2013) 384–390. [13] J.H.H. Bongaerts, K. Fourtouni, J.R. Stokes, Soft-tribology: lubrication in a compliant PDMS–PDMS contact, Tribol. Int. 40 (2007) 1531–1542.
11
[14] J. de Vicente, J.R. Stokes, H.A. Spikes, Soft lubrication of model hydrocolloids, Food Hydrocoll. 20 (2006) 483–491. [15] J.R. Stokes, M.W. Boehm, S.K. Baier, Oral processing, texture and mouthfeel: from Rheology to tribology and beyond, Curr. Opin. Colloid Interface Sci. 18 (2013) 340–359. [16] H. Ranc, C. Servais, P.-F. Chauvy, S. Debaud, S. Mischler, Effect of surface structure on frictional behaviour of a tongue/palate tribological system, Tribol. Int. 39 (2006) 1518–1526. [17] D. Dresselhuis, E. De Hoog, M. Cohen Stuart, G. Van Aken, Application of oral tissue in tribological measurements in an emulsion perception context, Food Hydrocoll. 22 (2008) 323–335. [18] R.A. de Wijk, J.F. Prinz, The role of friction in perceived oral texture, Food Qual. Prefer. 16 (2005) 121–129. [19] K.M. Hiiemae, J.B. Palmer, Tongue movements in feeding and speech, Crit. Rev. Oral Biol. Med. 14 (2003) 413–429. [20] M. Ratoi-Salagean, H.A. Spikes, Optimizing film formation by oil-in-water emulsions, Tribol. Trans. 40 (1997) 569–578. [21] A.A. Lubrecht, D. Mazuyer, P.M. Cann, Starved elastohydrodynamic lubrication theory: application to emulsions and greases, CR Acad. Sci. Paris t.2 (Série IV) (2001) 717–728. [22] S. Hurley, P.M. Cann, Grease composition and film thickness in rolling contacts, NLGI Spokesman 63 (4) (1999) 12–22. [23] S. Giasson, J. Israelachvili, H. Yoshizawa, Thin film morphology and tribology study of mayonnaise, 62 (1997) 640–647. [24] D.M. Dresselhuis, E.H.A. De Hoog, M.A. Cohen Stuart, M.H. Vingerhoeds, G.A. Van Aken, The occurrence of in-mouth coalescence of emulsion droplets in relation to perception of fat, Food Hydrocoll. 22 (2008) 1170–1183. [25] D.M. Dresselhuis, G.A. van Aken, E.H.A. de Hoogac, M.A. Cohen Stuart, Direct observation of adhesion and spreading of emulsion droplets at solid surfaces, Soft Matter 4 (2008) 1079–1085. [26] R.S. Johansson, M. Trulsson, K.A. Olsson, K.-G. Westberg, Mechanoreceptor activity from the human face and oral mucosa, Exp. Brain Res. 72 (1988) 204–208.
Please cite this article as: S. Tsui, et al., Friction measurements with yoghurt in a simulated tongue-palate contact, (2016), http://dx.doi.org/ 10.1016/j.biotri.2016.02.001