A study of saliva lubrication using a compliant oral mimic
Due to ethical issues and the difficulty in obtaining biological tissues, it is important to find synthetic elastomers
that can be used as replacement test media for research purposes. An important example of this is friction testing
to understand the mechanisms behind mouthfeel attributes during food consumption (e.g. syrupy, body and
clean finish), which requires an oral mimic. In order to assess the suitability of possible materials to mimic oral
surfaces, a sliding contact is produced by loading and sliding a hemispherical silica pin against either a polydimethyl
siloxane (PDMS), agarose, or porcine tongue sample. Friction is measured and elastohydrodynamic
film thickness is calculated based on the elastic modulus of the samples, which is measured using an indentation
method. Tests were performed with both saliva and pure water as the lubricating fluid and results compared to
unlubricated conditions.
PDMS mimics the tongue well in terms of protein adhesion, with both samples showing significant reductions
in friction when lubricated with saliva versus water, whereas agarose showed no difference between saliva and
water lubricated conditions. This is attributed to PDMS's eOeSi(CH3)2- group which provides excellent adhesion
for the saliva protein molecules, in contrast with the hydrated agarose surface. The measured modulus of
the PDMS (2.2 MPa) is however significantly greater than that of tongue (3.5 kPa) and agarose (66–174 kPa).
This affects both the surface (boundary) friction, at low sliding speeds, and the entrained elastohydrodynamic
film thickness, at high speeds.
Utilising the transparent PDMS sample, we also use fluorescence microscopy to monitor the build-up and flow
of dyed-tagged saliva proteins within the contact during sliding. Results confirm the lubricous boundary film
forming nature of saliva proteins by showing a strong correlation between friction and average protein intensity
signals (cross correlation coefficient=0.87). This demonstrates a powerful method to study mouthfeel mechanisms.
1. Introduction
Due to ethical issues and the difficulty in obtaining biological tissues,
it is necessary to find synthetic elastomers that can be used as
replacement test media for research purposes. A key example of this is
friction testing to understand the mechanisms behind mouthfeel attributes
during food consumption (e.g. syrupy, body and clean finish),
which requires an oral mimic. This is important since the acceptability
of food and beverages depend critically on their mouthfeel, which results
from tribological and rheological processes (Stokes, Boehm, &
Baier, 2013). Moreover, a poor understanding of these processes
currently limits the development of healthy formulations that can replicate
foods while reducing ingredients such as fat (Dresselhuis, 2008),
(Drewnowski, 1997). When mimicking the oral mucosa for these in
vitro tribological studies of foods and beverages, consideration must be
made of the mucosal pellicle. Like the acquired enamel pellicle on teeth,
this is a subset of salivary proteins that specifically bind to oral epithelial
cells (Gibbins, Proctor, Yakubov, Wilson, & Carpenter, 2014).
Unlike the acquired enamel pellicle, the mucosal pellicle is mostly
composed of mucins and secretory IgA. This layer is driven by the interaction
of salivary mucins (muc5b and muc7) with membrane-bound
mucin (muc1) expressed on oral epithelial cells (Vijay, Inui, Dodds,
Proctor, & Carpenter, 2015). Mucins are large highly glycosylated
proteins which retain considerable amounts of water when initially
secreted (Corfield, 2015). Thus, in addition to saliva lubricating the
surface, there is also a hydrogel-like layer adjacent to the surface. All
too frequently however, saliva is omitted from in vitro tests as it was
cited as being too inconvenient to collect in sufficient quantities or
considered too complex to give consistent results.
Previously, the oral mucosa was mimicked using glass or other hard
substrates (Chen & Stokes, 2012). More recently elastic substrates have
been used which introduced soft-tribology with Hertzian mechanics. In
an important investigation by Dresselhuis et al. (Dresselhuis et al.,
2007), the surface characteristics of pig tongue were compared with
those of PDMS. Their investigation concluded that PDMS showed dissimilarities
in surface characteristics to those of a tongue surface, since
the oral mucosa and PDMS rubbers, even with a structured surface to
reproduce biological scenarios, were not interchangeable in tribological
experiments. However, this widely cited paper has a critical shortcoming
in that it used only emulsion as the lubricant and saliva interactions
were completely ignored. Other work carried out on biological
surfaces, but without the presence of saliva include studies by Adams
et al. (Adams, Briscoe, & Johnson, 2007) and Tang et al. (Tang &
Bhushan, 2010), (Tang, Bhushan, & Ge, 2010) into the lubricating
properties of human skin. Adams et al. used a smooth glass or polypropylene,
spherical tipped probe sliding against a human forearm,
while Tang et al. employed shaved porcine skin. Results were reported
for a range of lubricating conditions, but repeatability of testing was
difficult to achieve. Prinz et al. (Prinz, de Wijk, & Huntjens, 2007) did
investigate the frictional properties between two pig mucosal surfaces
lubricated with human saliva. However, scant data is presented and no
comparison is made between different component materials.
For the majority of research, crosslinked polydimethyl siloxane
(PDMS) has been chosen because of its elastic properties, easy handling
and relatively low stiffness, comparable to soft biological tissues (Cox,
Driessen, Boerboom, Bouten, & Baaijens, 2008; Khanafer, Duprey,
Schlicht, & Berguer, 2009). PDMS is utilized as one (de Vicente, Stokes,
& a Spikes, 2006), (Malone, Appelqvist, & Norton, 2003), (Tang &
Bhushan, 2010), (Tang et al., 2010) or both (Stokes, Bongaerts, &
Rossetti, 2007), (Lee & Spencer, 2005), (Bongaerts, Fourtouni, & Stokes,
2007) of the contacting surfaces in the tribological contact to maintain
low contact pressures and create the conditions for isoviscous-elastohydroynamic
lubrication (I-EHL) to occur. One key advantage of PDMS
which has contributed to its widespread use is its ease of fabrication.
Prior to crosslinking, PDMS can be cast into suitable moulds of almost
any desired shape. Other attractive features of PDMS include its physiological
inertness, availability, low unit cost, as well as its good
thermal and oxidative stability.
PDMS is a transparent silicon-based organic polymer, used to represent
biological materials in numerous tribological studies (e.g.
(Bongaerts et al., 2007) (Dresselhuis et al., 2007), (De Vicente, Spikes,
& Stokes, 2004)). It is highly compliant, with a Young's modulus
E≈0.57–3.7 MPa (depending on degree of crosslinking) (Wang,
Volinsky, & Gallant, 2014), due to its uniquely low glass transition
temperature (Tg≈−125 °C) (Lötters, Olthuis, Veltink, & Bergveld,
1999). The surface of PDMS is hydrophobic, due to its repeating eOeSi
(CH3)2- group (Adams et al., 2007) but can be made hydrophilic by
plasma cleaning. In addition to this, PDMS is being used extensively in
polymeric microfluidics (e.g. (Eddings, Johnson, & Gale, 2008)) research
and findings from this area may be usefully applied in this study.
The tribological properties of PDMS are now fairly well understood.
Vorvolaskos and Chaudhury (Vorvolakos & Chaudhury, 2003) investigated
the effect of molecular weight and test temperature on
friction in a pure sliding contact between a PDMS and metal surface.
Bongaerts et al. (Bongaerts et al., 2007) investigated the effect of surface
roughness of PDMS on the lubricating properties of biopolymers
and aqueous solutions. PDMS, like most elastomeric surfaces, is by
nature hydrophobic but an oxidation treatment can be employed to
create a hydrophilic surface. Hillborg et al. (Hillborg & Gedde, 1998),
(Hillborg, Sandelin, & Gedde, 2001) and Schneemilch et al.
(Schneemilch & Quirke, 2007) investigated the wettability of PDMS
before and after oxidisation by several techniques and studied the effect
of crosslink density on oxidation. de Vicente et al. (de Vicente, Stokes, &
Spikes, 2005) looked at the influence of surface modification of PDMS
on its aqueous lubrication properties. However, there remains some
debate over the suitability of PDMS as a model biosurface and instances
of PDMS being tested under saliva conditions are few in number.
The second soft matrix to be considered here as a potential substrate
to mimic the oral mucosa is agarose. Agarose, the agaropectin deficient
fraction of agar derived from seaweed and consisting of β-1,3 linked α-
galactose and α-1,4 linked 3,6-anhydro-αL-galactose residues
(Normand, Lootens, Amici, Plucknett, & Aymard, 2000), is used to
create a hydrogel-like matrix. The compliance of agarose varies enormously
depending on concentration, with Young's moduli ranging
from ∼1.5 kPa to ∼3 MPa (Benkherourou, Rochas, Tracqui, Tranqui, &
Guméry, 1999), (Normand et al., 2000), (Chen, Suki, & An, 2004). In
addition to this, agarose has the ability to grow cells in suspension and
has therefore been used in tissue culturing systems (Chen et al., 2004).
This combination of properties make agarose an attractive choice in
biomedical research, for example, as a cartilage mimic (Saris et al.,
2000), or as a phantom material for magnetic resonance elastography
(Muthupillai et al., 1995). It is therefore surprising that agarose has
been used in few tribological studies and seems to have been overlooked
completely as an oral mimic. Fernández Farrés studied its frictional
behaviour, but did so under glucose and glycerol lubrication
rather than saliva (Fernández Farrés & Norton, 2015). Shewan et al.
also recently studied the lubrication performance of agarose, but as
particles in suspensions rather than a substrate (Shewan & Stokes,
2015).
It can be concluded that it is important to be able to mimic the oral
mucosa surface and various materials have been studied for this purpose.
However, these have rarely been compared with actual biological
materials (probably because of the difficulty in source, preserving and
securing them) and almost never when lubricated by saliva. To address
this, the current study characterises the friction and film thickness
performance of polydimethyl siloxane (PDMS), agarose and porcine
tongue, with the aim of assessing their suitability as an oral mimic for
tribological testing. Particular attention is paid to the compliance and
protein binding behaviours of these substrates.
2. Test methods
2.1. Specimen preparation
PDMS specimens were moulded using a commercially available
Sylgard 184 kit from Dow-Corning, containing a base and curing agent
to produce a material with a Young's modulus 1.84 MPa at 23 °C.
Agarose gel was produced by dissolving powdered agarose (Sigma-
Aldrich, Poole, UK) into water at 1 or 2% w/v. To aid dissolution the
solution was heated to 90 °C then allowed to cool to a temperature
below the coil-helix transition at around 35 °C. At this point agarose
forms a gel, consisting of an infinite three-dimensional network of fibre
helices (Normand et al., 2000).
Prior to collection the subject refrained from food and drink for 1 h.
Resting whole mouth saliva (WMS) was collected from a single subject
by drooling into a pre-weighed tube, kept on ice. After collection saliva
was briefly centrifuged (3000 g for 3 min) to remove sloughed cells and
other debris.
Porcine tongue was procured and tested on the same day. Its underside
was removed to produce a parallel slab. This specimen was then
bonded onto a flat plate using cyanoacrylate adhesive and mounted in
the friction rig.
G. Carpenter et al. Food Hydrocolloids 92 (2019) 10–18
11
2.2. Indentation and surface roughness measurements
The elastic modulus of each sample material was measured using an
indention test performed on a Mach 1 rig (Biomomentum Inc., Laval,
Canada). This involved indenting the sample at 1 mm/s with a spherical
indenter with radius 3.175 mm, during three repeat tests, while measuring
the normal force and the vertical displacement. The normal force
was measured using 1.5 N single-axis load cell with a resolution of
75 μN and the vertical displacement was measured by the moving stage
of the rig with a resolution of 0.1 μm. A depth of penetration of 0.6mm
was used for agarose 1% w/v and 0.4mm for each of the other samples.
This was done in accordance with Van Dommelen et al.’s (van
Dommelen, van der Sande, Hrapko, & Peters, 2010) suggestion that the
sample thickness does not significantly affect the data if indentation
depths are restricted to less than 10% of the sample thickness. Nevertheless
a formulation that considers the finite thickness of the sample
was used (Hayes, Keer, Herrmann, & Mockros, 1972) to calculate
Young's moduli. Contact mechanics equations were fitted to the data to
give the Young's modulus, specifically,
χ = a
dR
2
(1)
=
− κ P ν
aGd
(1 )
4 (2)
Where d is the displacement of the indenter, R is the radius of the indenter,
a is the radius of the contact region, P the applied load, G the
shear modulus, and ν the Poisson's ratio. A schematic of the test is
provided in Fig. 1.
The reaction force and the indenter displacement are recorded by
the Mach-1 Motion Software and enter the above equations as P and d
respectively. The Poisson's ratio is assumed to be equal to 0.5 (incompressible
materials). The specimens' height h and the indenter radius
R are also known. The values of χ and κ are given in Table 2 in
(Hayes et al., 1972) for different values of a, h and ν. The radius of the
contact region a is estimated during the curve fit with equation (1).
Once the fitting algorithm converges, the Young's Modulus E is computed
from the Shear Modulus G using the equation
E = 2G(1 + ν) (3)
The roughness of each of the specimens was measured three times
(each at a different location) on the surface, using a Veeco optical
profilometer.
2.3. Protein staining measurements
SDS-PAGE (sodium dodecyl sulfate–polyacrylamide gel electrophoresis)
was used to assess the degree of binding of different proteins
to the surface of the oral mimics. This immunoblotting technique targets
proteins in a sample with specific dyes and measures their progression
through a gel, due to an applied electric field. In this way,
different proteins in a sample with different molecular weights are separated.
Staining involved incubating for an hour at room temperature
with whole mouth saliva from a single subject. Coomassie Brilliant Blue
(CBB) was used stain of all proteins. In addition to this, Periodic Acid
Schiff's (PAS) was used to stain for highly glycosylated proteins and
specific antibody with sensitive chemiluminescent detection was used
for the saliva protein muc7. Samples were removed from the surface of
each of the oral mimic surfaces and tested in this way to investigate
which proteins were present.
2.4. Friction measurements
A contact was produced by loading and sliding a 5mm radius silica
hemisphere against the compliant disc specimen, using a UMT2
(Universal Materials Tester), manufactured by CETR, (Campbell, USA).
This equipment was operated in pin-on-disc mode, so that the PDMS
specimen rotated, while the silica hemisphere was held stationary. The
lower specimen was located on a rotating table, capable (with certain
modifications), to run at speeds from 0.01 rpm up to∼4000 rpm.
Friction force (Fx) and normal load (Fz) were measured using strain
gauges, bonded to the housing above the stationary silica hemisphere
specimen. Sensitive, low-load sensors were chosen for this purpose,
with measurement ranges of±0.65 N and±6 N for Fx and Fz respectively.
This experimental setup is shown in Fig. 2 a. Friction data
was recorded over a speed range from 0.002 to 0.35 ms−1 with an
applied load of 0.2 N.
2.5. Laser induced fluorescence measurements
The custom-built Laser Induced Fluorescence (LIF) microscope is
shown by the photograph and schematic in Fig. 2b. It comprises an LED
light source, which produces a beam that is focussed through the
transparent PDMS specimen onto the contact interface. For certain
tests, the proteins in the lubricating saliva were tagged with a dye,
fluorescein isothiocyanate, (FITC) in order for them to fluoresce when
excited by the LED. The emitted light is filtered and collected by a highspeed
EMCCD camera. For film thicknesses greater than 200 nm, the
recorded intensity of the fluorescence light emitted from the contact is
proportional to the thickness of the liquid in the interface. This means
that the images acquired by the camera represent maps showing the
distribution of proteins in the contact. Further details of the fluorescence
technique can be found in (Myant, Reddyhoff, & Spikes, 2010),
(Reddyhoff, Choo, Spikes, & Glovnea, 2010).
3. Results
3.1. Indentation and roughness results
Fig. 3 shows the force displacement curves for the four materials
during the indentation tests using the Biomomentum Mach-1 rig.
Equations (1)–(3) were applied to this data giving the Young's Modulus
values shown in Table 1. As, expected the Young's Modulus of the
porcine tongue at 3.5 kPa is lower than other measurements of biological
tissue found in the literature – e.g. human skin: 25–101 kPa
(Akhtar, Sherratt, Cruickshank, & Derby, 2011), human muscle:∼7 kPa
(McKee et al., 2011). These values were most closely mimicked by the
agarose with a modulus 66 and 174 kPa for the 1 and 2% concentrations.
The modulus of the PDMS was nearly two orders of magnitude
Fig. 1. Schematic diagram of indentation setup. higher than the biological sample.
G. Carpenter et al. Food Hydrocolloids 92 (2019) 10–18
12
Table 2 shows the surface roughness measurements for each of the
specimens, which are separated by approximately an order of magnitude
(PDMS < Agar < Tongue). The effect this variation has on friction,
however, is counteracted by the different stiffness which increases
in the opposite sense (e.g. the asperities on the tongue surface are
readily flattened). The range of values displayed for each measurement
refers to the standard error, which is due to the variation over surface of
the specimens, rather than any error in the measurement.
3.2. Protein staining results
When incubated for an hour at room temperature with whole mouth
saliva from a single subject, neither agarose nor PDMS bound significant
amounts protein as shown by Coomassie Brilliant Blue (CBB)
staining of all proteins as shown in Fig. 4. Small amounts of amylase,
the single most abundant protein in saliva, is the only protein apparent
(identity based on apparent molecular weight). When the same gel was
stained with Periodic Acid Schiff's (PAS), a stain for highly glycosylated
proteins, small amounts of muc5b and muc7 were visible in the agarose
gel but nothing in the sample eluted from PDMS. Immunoblotting for
muc7 using a specific antibody with sensitive chemiluminescent detection
again suggested agarose gel bound some mucins whereas PDMS
did not. Incorporating potentially muco-adhesive agents such as chitosan
and the lectin WGA (AWGL) into the agarose appeared to enhance
protein, and mucin in particular, binding to the agarose.
3.3. Friction results
In this section, Stribeck curves have plotted the speed on the x-axis
rather than the product of speed×viscosity as is customary (de Vicente
et al., 2006). This is because the viscosity of saliva, being highly non-
Newtonian, varies strongly as a function of shear rate (Rantonen &
Meurman, 1998) and is therefore not constant throughout each test.
Another obstacle in assuming a single viscosity is the inhomogeneous
and surface active nature of saliva means that it is not possible to assume
whether it is the high viscosity proteins or just water molecules
are entrained between the surfaces.
Fig. 5 shows the variation in friction with sliding speed for the
agarose-glass contact. Under unlubricated and water lubricated conditions,
this substrate exhibits lower friction, due to the agarose being a
hydrogel which releases water when compressed. When agarose was
submerged in water it exhibits Stribeck curve behaviour with higher
friction at low speeds which decreases rapidly with speed due to the
formation of an elastohydrodynamic film. However, when lubricated
with saliva, the friction behaviour is completely unchanged compared
to pure water.
Fig. 6 shows the variation in friction with sliding speed for the
PDMS-glass contact under different conditions. When the contact is
unlubricated, the coefficient of friction remains between 3 and 4, due to
the strong adhesive interaction between the surfaces. The increase
followed by a decrease in friction with sliding speed may be attributed
to the viscoelastic properties of the elastomer (the friction that arises
Fig. 2. Laser Induced Fluorescence setup, a) Photograph, b) Schematic diagram of indentation setup.
Fig. 3. Force-displacement curves for each material obtained during indentation.
Table 1
Young's modulus results for each test material in kPa.
Porine tongue Agarose (1%) Agarose (2%) PDMS
3.46 66.4 174 2270
Table 2
Surface roughness results for each test material. Examples of the corresponding
surface topographies are shown in the appendix.
Roughness (nm)
Average (Ra) RMS (Rq)
Tongue 5480 ± 667 656 ± 403
PDMS 10.1 ± 0.16 13.1 ± 0.23
Agar 1% 399 ± 91 514 ± 109
Agar 2% 325 ± 14 420 ± 18
G. Carpenter et al. Food Hydrocolloids 92 (2019) 10–18
13
from the deformation of the PDMS varies as a function of speed due to
its viscoelastic response).
At the lowest speed, the dry and water submerged friction values are
similar, showing that no water is present between the surfaces even
when submerged (i.e. no boundary film is formed). This is because the
speed is insufficient to separate the surfaces hydrodynamically and also
water molecules are not attracted to either the PDMS or glass surface. In
contrast, when flooded with whole mouth saliva, very low friction
(more than two orders of magnitude less than the dry case) is observed.
These observations are in agreement with those of Stokes and coworkers
(Bongaerts et al., 2007).
Fig. 7 shows the friction versus speed behaviour for the porcine
tongue sample. When lubricated with pure water, this sample shows
high boundary friction which reduces with speed due to lubricant entrainment.
In addition to this, the low speed boundary friction is significantly
reduced when lubricated with saliva when compared to
water. The shape of the dry, water and saliva lubricated curves are similar
for PDMS and tongue, however there was a significant difference
in terms of the magnitude of the friction.
3.4. Laser induced fluorescence results
An advantage of PDMS over both the tongue and the agarose samples
is that it is transparent, which enables imaging of the contact. To
demonstrate this, the LIF microscopy results in Fig. 8 show the build-up
and flow of FITC-dyed saliva proteins within the contact during sliding.
Images a-d in this figure are intensity maps of the contact showing the
distribution of proteins (these are frames taken from the videos provided
as Supplementary Material). Here, bright colours represent high
concentrations of proteins in the contact and the dark blue circular
region is the pressurised contact area. Proteins agglomerations of
varying morphologies are evident as they are entrained due the sliding
motion from the inlet at the top of the figure to the outlet at the bottom.
The figure also plots the variation in friction over time alongside a
measure of the fluorescence intensity within the contact. The latter was
obtained by counting the number of pixels within the contact with an
intensity greater than the test average (using a Matlab program).
There is a clear correlation between the coefficient of friction and
the presence of proteins within the contact zone. This is highlighted by
Fig. 4. Staining showing binding of proteins to each mimic surface. (note: the whole mouth saliva sample is labelled WMS).
Fig. 5. Friction versus sliding speed for an agarose disc pressed against a stationary silica hemisphere with a force of 0.2 N. a) linear scale, b) log scale.
G. Carpenter et al. Food Hydrocolloids 92 (2019) 10–18
14
the calculated cross correlation coefficient of 0.872 and the visible
occurrence of peaks (shown by ˄) in one single coinciding with troughs
(shown by v) in the other signal, and vice versa.
4. Discussion
The stiffness of the tongue sample is far closer to that of the agarose
than the PDMS. This means that, for the agarose contact, the area and
pressure match more closely those found in the mouth. Moreover, if this
is considered in isolation, it suggests the boundary friction and hydrodynamic
film thickness separating the surfaces for the agarose are more
realistic. But it is important also to consider the mucosal pellicle for
lubrication of oral surfaces by saliva and to implement this we added
mucoadhesive components to agarose gels to enhance mucin binding.
In some ways this appeared successful with greater amounts of all
salivary proteins, including the two mucins (muc 5b and muc7),
binding in greater amounts to the chitosan and WGA lectin containing
agarose, shown by protein staining. However, there was little effect on
the tribology when the mucoadhesive agarose was compared to agarose
alone. Indeed, there was almost no difference between agarose lubricated
by water or saliva. This suggests that this substrate is already
being lubricated by the surface itself – probably water being expelled
from the hydrogel under the pressure of the tribo-pairing. Furthermore,
the interchangeability of the curves for water and saliva lubricated
contacts in the full film regime, where friction is dominated by viscous
drag, suggests high viscosity saliva proteins are not even being entrained
into the contact at entrainment high speed.
The behaviour of PDMS showed much stronger protein interactions.
When sliding at low speed (∼0.1 mm/s) in the boundary regime (i.e.
when there is insufficient hydrodynamic entrainment of liquid to separate
the surfaces), the coefficient of friction for PDMS when lubricated
by saliva is two orders of magnitude lower than when lubricated
with pure water (∼0.01 vs ∼2). Since saliva is made up of
99.5% water and<0.5% protein molecules, this shows the proteins are
highly effective surface active lubricating additives, which adhere to
PDMS and oral surfaces to produce a lubricous low shear strength interface.
More specifically, PDMS, like the tongue is hydrophobic
(Dresselhuis et al., 2007) and due to its charged eOeSi(CH3)2- group it
attracts proteins indiscriminately (Phillips & Cheng, 2005) (in fact the
adherence of biological proteins to PDMS is a problematic occurrence in
biological lab-on-chip systems (Phillips & Cheng, 2005)). The viscosity
difference between water and saliva (0.89 cP and ∼5 cP (Rantonen &
Meurman, 1998)) is insufficient to explain this difference.
It could also be hypothesised that the elasticity of the bulk saliva
may be responsible for the differences in the hydrodynamic/rheological
response of the PDMS compared to water. However, at such low speeds
elasticity should not play a role. Furthermore, as shown, the friction is
strongly affected by the chemistry of sample surface, which would not
be the case under full film hydrodynamic lubrication. Finally, as shown
by Davies et al., the elasticity of resting saliva, as tested here, is significantly
lower than that of acid stimulated saliva (Davies, Wantling, &
Stokes, 2009).
The shape of the dry, water and saliva lubricated curves for tongue
are most similar to those of PDMS, which supports the latter's use an
oral mimic. However, there was a significant difference in terms of the
magnitude of the friction. Under dry, unlubricated conditions, the
Fig. 6. Friction versus sliding speed for PDMS disc pressed against a stationary silica hemisphere with a force of 0.2 N. a) linear scale, b) log scale.
Fig. 7. Friction versus sliding speed for porcine tongue pressed against a stationary silica hemisphere with a force of 0.2 N. a) linear scale, b) log scale.
G. Carpenter et al. Food Hydrocolloids 92 (2019) 10–18
15
PDMS shows a friction coefficient of around 3.5 in contrast to 1.5 for
the tongue sample. When water is replaced with saliva, the PDMS
friction reduces to ∼0.02 while the tongue sample only falls to 0.25.
This difference in friction coefficient magnitude between PDMS and
tongue, under low speed conditions, when the surfaces are in contact,
can be analysed as follows. As predicted by Schallamach (Schallamach,
1958) and Roberts (Barquins & Roberts, 2000), using Hertz theory, the
coefficient of friction under dry/boundary lubrication conditions (i.e.
when not liquid is separating the surfaces) is given by:
μ = πS (9R/16E)2/3W−1/3 (4)
where R is the reduced radius, E is the elastic modulus, S is the interfacial
shear stress and W is the load. This shows that higher friction
coefficients arise in contacts between compliant materials, since these
deform and produce a larger contact area to be sheared. Equation (4)
can be used to calculate the shear stress within the contact, S, under
boundary lubrication conditions since all other quantities are known,
which gives values of 0.53 and 3.2 kPa for tongue and PDMS respectively.
This suggests that, when lubricated by saliva, the lower friction
of the PDMS surface arises due to its higher stiffness and smaller contact
area, but per unit area the protein covered tongue is in fact more easily
sheared. Another factor is the difference in roughness between the two
samples. Under dry conditions, the lower roughness of the PDMS increases
the real contact areas and hence adhesion, whereas under
protein lubrication lower roughness aids the formation of a complete
surface film.
The highly lubricious nature of the saliva proteins and their adherence
to the PDMS surface are confirmed by the in-contact LIF results.
In addition to demonstrating the effectiveness of this technique to
study saliva protein entrainment, these results shed light on the details
of this intermittent process. More specifically, the observed highly
transient nature of the protein entrainment is similar to that demonstrated
by Fan et al. (Fan, Myant, Underwood, Cann, & Hart, 2011) who
attributed the build-up and breakdown of proteins within the contact to
the following inlet aggregation mechanism. Due to the contact geometry
and flow path of the lubricant, proteins are transported into the
contact inlet. Some of these proteins attach to the converging surfaces.
Over time additional proteins become entangled with the surface protein
branches, forming a larger protein mass in the inlet zone. A critical
point is then reached where surface friction forces and lubricant hydrodynamic
forces cause this protein mass to breakdown, allowing
large agglomerate of proteins to be dragged into the contact zone. This
can be observed in Fig. 8, highlighted on the plot with a * symbol,
where peak protein presence occurs with minima in friction coefficient.
Fig. 8. Laser Induced Fluorescence results from a sliding test of silica hemisphere loaded against PDSM disc and lubricated with FITC dyed saliva. a) Intensity maps
for unloaded contact, b) to d) Intensity maps during sliding, e) Variation of friction coefficient (blue) and fluorescence signal (orange), obtained by counting number
of pixels with intensity greater than the test average. To highlight the correlation, example peaks are labelled with ˆ and example troughs are labelled with v. The
arrows around 400 s highlight symmetrical trends in the two signals. Note: the step changes in fluorescence observed at 5 and 440 s correspond to increase and
decrease in in-contact proteins during the loading and unloading of the contact. (For interpretation of the references to colour in this figure legend, the reader is
referred to the Web version of this article.)
G. Carpenter et al. Food Hydrocolloids 92 (2019) 10–18
16
The difference in lubricating properties of saliva compared to water
are assumed to relate to the salivary proteins such as mucins and statherin.
Mucins contribute to the viscosity of saliva which may aid the
hydrodynamic mode of lubrication (Bongaerts et al., 2007) whereas
statherin, a small surface active protein is regarded as a boundary lubricant
(Douglas et al) (Harvey, Carpenter, Proctor, & Klein, 2011),
although it is entirely possible that other proteins also contribute to the
lubricating properties.
5. Conclusions
From a surface chemistry point of view, PDMS is suitable at replicating
the oral mucosa, since, like the tongue, it is hydrophobic
(Dresselhuis, 2008) and its charged groups, which attract proteins
(Phillips & Cheng, 2005). This resulted in PMDS showing similar friction
versus speed trends to the biological sample. Agarose on the other
hand shows only a minor difference in friction when lubricated by
saliva versus water. This is attributed to the hydrated agarose surface
weakly adhering to the saliva proteins. The friction properties of
agarose did not improve even after the agarose was treated with mucoadhesive
components to enhance mucin binding.
Although PDMS rubbers have similar hydrophobic qualities to a
tongue, PDMS has an elastic modulus two orders of magnitude larger.
Furthermore, even if the degree of cross linking is limited the modulus
of PDMS reduces only to around 570 kPa (Wang et al., 2014) versus
3.4 kPa for tongue. This is significant shortcoming, since the stiffness of
the sample affects both the boundary friction (μ α E′−2/3 (Schallamach,
1958)) and the elastohydrodynamic film thickness (h α E′0.66 (de
Vicente et al., 2005)). There is also considerable variation in roughness
between the specimens tested, with agarose matching the tongue most
closely. However, the effect this has on friction is limited due to the incontact
flattening of the rougher materials, which have lower stiffness.
An advantage of PDMS is that being transparent it allows in-contact
imaging of saliva lubrication mechanisms. This was demonstrated using
laser induced fluorescence and the resulting strong correlation (0.87)
between friction and protein intensity signals confirms the lubricous
boundary film forming ability of saliva proteins. Protein aggregation
was shown to be highly transient in nature. The application of this
technique to study the tribological interactions between saliva and
foods and beverages in order to scientifically characterise mouthfeel
attributes is the subject of ongoing research.
Acknowledgements
S.K. Baier and R.V Potineni are employed by PepsiCo, Inc. The views
expressed in this research article are those of the authors and do not
necessarily reflect the position or policy of PepsiCo, Inc. The research
was funded by PepsiCo, Inc (grant number: P55310-1).
Appendix A. Supplementary data
Supplementary data to this article can be found online at 01.049.
Appendix. Surface topography measurements
Fig. A1. Surface topographies of the three materials, measured using a Veeco optical profilometer, a) porcine tongue, b) PDMS, c) agrose.
References
Adams, M. J., Briscoe, B. J., & Johnson, S. a. (2007). Friction and lubrication of human
skin. Tribology Letters, 26(3), 239–253.
Akhtar, R., Sherratt, M. J., Cruickshank, J. K., & Derby, B. (2011). Characterizing the
elastic properties of tissues. Materials Today, 14(3), 96–105.
Barquins, M., & Roberts, A. D. (2000). Rubber friction variation with rate and temperature:
Some new observations. Journal of Physics D Applied Physics, 19(4), 547–563.
Benkherourou, M., Rochas, C., Tracqui, P., Tranqui, L., & Guméry, P. Y. (1999).
Standardization of a method for characterizing low-concentration biogels: Elastic
properties of low-concentration agarose gels. Journal of Biomechanical Engineering,
121(2), 184–187.
Bongaerts, J. H. H., Fourtouni, K., & Stokes, J. R. (2007). Soft-tribology: Lubrication in a
compliant PDMS-PDMS contact. Tribology International, 40(10–12), 1531–1542 SPEC.
ISS.
Chen, J., & Stokes, J. R. (2012). Rheology and tribology: Two distinctive regimes of food
texture sensation. Trends in Food Science & Technology, 25(1), 4–12.
Chen, Q., Suki, B., & An, K.-N. (2004). Dynamic mechanical properties of agarose gels
modeled by a fractional derivative model. Journal of Biomechanical Engineering,
126(5), 666–671.
Corfield, A. P. (2015). Mucins: A biologically relevant glycan barrier in mucosal protection.
Biochimica et Biophysica Acta (BBA) - General Subjects, 1850(1), 236–252.
Cox, M. A. J., Driessen, N. J. B., Boerboom, R. A., Bouten, C. V. C., & Baaijens, F. P. T.
(2008). Mechanical characterization of anisotropic planar biological soft tissues using
finite indentation: Experimental feasibility. Journal of Biomechanics, 41(2), 422–429.
Davies, G. A., Wantling, E., & Stokes, J. R. (2009). The influence of beverages on the
stimulation and viscoelasticity of saliva: Relationship to mouthfeel? Food
Hydrocolloids, 23(8), 2261–2269.
De Vicente, J., Spikes, H. A., & Stokes, J. R. (2004). Complex fluids in thin lubricating films
between soft surfaces. 1–4.
van Dommelen, J. A. W., van der Sande, T. P. J., Hrapko, M., & Peters, G. W. M. (2010).
Mechanical properties of brain tissue by indentation: Interregional variation. Journal
of the Mechanical Behavior of Biomedical Materials, 3(2), 158–166.
W. H. Douglas, E. S. Reeh, N. Ramasubbu, P. A. Raj, K. K. Bhandary, and M. J. Levine,
"Statherin: A major boundary lubricant of human saliva," Biochemical and
Biophysical Research Communications, vol. 180, no. pp. 91-97.
Dresselhuis, D. M. (2008). The fate of fat: Tribology, adhesion and fat perception of food
emulsions. Wageningen Universiteit.
Dresselhuis, D. M., Klok, H. J., Stuart, M. A. C., De Vries, R. J., Van Aken, G. A., & De
Hoog, E. H. A. (2007). Tribology of o/w emulsions under mouth-like conditions:
Determinants of friction. Food Biophysics, 2(4), 158–171.
Drewnowski, A. (1997). Why do we like fat? Journal of the American Dietetic Association,
97(7), S58–S62.
Eddings, M. A., Johnson, M. A., & Gale, B. K. (2008). “Determining the optimal
PDMS–PDMS bonding technique for microfluidic devices. Journal of Micromechanics
and Microengineering, 18(6), 067001.
Fan, J., Myant, C. W., Underwood, R., Cann, P. M., & Hart, A. (2011). Inlet protein aggregation:
A new mechanism for lubricating film formation with model synovial
fluids. Proceedings of the Institution of Mechanical Engineers, Part H: Journal of
G. Carpenter et al. Food Hydrocolloids 92 (2019) 10–18
17
Engineering in Medicine, 225(7), 696–709.
Fernández Farrés, I., & Norton, I. T. (2015). The influence of co-solutes on tribology of
agar fluid gels. Food Hydrocolloids, 45, 186–195.
Gibbins, H. L., Proctor, G. B., Yakubov, G. E., Wilson, S., & Carpenter, G. H. (2014).
Concentration of salivary protective proteins within the bound oral mucosal pellicle.
Oral Diseases, 20(7), 707–713.
Harvey, N. M., Carpenter, G. H., Proctor, G. B., & Klein, J. (2011). Normal and frictional
interactions of purified human statherin adsorbed on molecularly-smooth solid substrata.
Biofouling, 27(8), 823–835.
Hayes, W. C., Keer, L. M., Herrmann, G., & Mockros, L. F. (1972). A mathematical analysis
for indentation tests of articular cartilage. Journal of Biomechanics, 5(5), 541–551.
Hillborg, H., & Gedde, U. W. (1998). Hydrophobicity recovery of polydimethylsiloxane
after exposure to corona discharges. Polymer, 39(10), 1991–1998.
Hillborg, H., Sandelin, M., & Gedde, U. W. (2001). Hydrophobic recovery of polydimethylsiloxane
after exposure to partial discharges as a function of crosslink density.
Polymer, 42(17), 7349–7362.
Khanafer, K., Duprey, A., Schlicht, M., & Berguer, R. (2009). “Effects of strain rate, mixing
ratio, and stress-strain definition on the mechanical behavior of the polydimethylsiloxane
(PDMS) material as related to its biological applications. Biomed.
Microdevices, 11(2), 503–508.
Lee, S., & Spencer, N. D. (2005). Aqueous lubrication of polymers: Influence of surface
modification. Tribology International, 38(11–12), 922–930 SPEC. ISS.
Lötters, J. C., Olthuis, W., Veltink, P. H., & Bergveld, P. (1999). The mechanical properties
of the rubber elastic polymer polydimethylsiloxane for sensor applications. Journal of
Micromechanics and Microengineering, 7(3), 145–147.
Malone, M. E., Appelqvist, I. A. M., & Norton, I. T. (2003). Oral behaviour of food hydrocolloids
and emulsions. Part 1. Lubrication and deposition considerations. Food
Hydrocolloids, 17(6), 763–773.
McKee, C. T., Last, J. A., Ph, D., Russell, P., Ph, D., Murphy, C. J., et al. (2011).
“Indentation versus tensile measurements of Young's modulus for soft biological
tissues. Tissue Engineering Part B Reviews, 17(3), 155–164.
Muthupillai, R., Lomas, D., Rossman, P., Greenleaf, J., Manduca, A., & Ehman, R. (1995).
Magnetic resonance elastography by direct visualization of propagating acoustic
strain waves. Science (80), 269(5232), 1854–1857.
Myant, C., Reddyhoff, T., & Spikes, H. A. (2010). Laser-induced fluorescence for film
thickness mapping in pure sliding lubricated, compliant, contacts. Tribology
International, 43(11), 1960–1969.
Normand, V., Lootens, D. L., Amici, E., Plucknett, K. P., & Aymard, P. (2000). New insight
into agarose gel mechanical properties. Biomacromolecules, 1, 730–738.
Phillips, K. S., & Cheng, Q. (2005). Microfluidic immunoassay for bacterial toxins with
supported phospholipid bilayer membranes on poly(dimethylsiloxane). Analytical
Chemistry, 77(1), 327–334.
Prinz, J. F., de Wijk, R. A., & Huntjens, L. (2007). Load dependency of the coefficient of
friction of oral mucosa. Food Hydrocolloids, 21(3), 402–408.
Rantonen, P. J. F., & Meurman, J. H. (1998). Viscosity of whole saliva. Acta Odontologica
Scandinavica, 56(4), 210–214.
Reddyhoff, T., Choo, J. H., Spikes, H. A., & Glovnea, R. P. (2010). Lubricant flow in an
elastohydrodynamic contact using fluorescence. Tribology Letters, 38(3), 207–215.
Saris, D. B., Mukherjee, N., Berglund, L. J., Schultz, F. M., An, K. N., & O'Driscoll, S. W.
(2000). Dynamic pressure transmission through agarose gels. Tissue Engineering, 6(5),
531–537.
Schallamach, A. (1958). Friction and abrasion of rubber. Wear, 1(5), 384–417.
Schneemilch, M., & Quirke, N. (2007). Effect of oxidation on the wettability of poly(dimethylsiloxane)
surfaces. The Journal of Chemical Physics, 127(11).
Shewan, H. M., & Stokes, J. R. (2015). Viscosity of soft spherical micro-hydrogel suspensions.
Journal of Colloid and Interface Science, 442, 75–81.
Stokes, J. R., Boehm, M. W., & Baier, S. K. (2013). Oral processing, texture and mouthfeel:
From rheology to tribology and beyond. Current Opinion in Colloid & Interface Science,
18(4), 349–359.
Stokes, J. R., Bongaerts, J. H. H., & Rossetti, D. (2007). The lubricating properties of
human whole saliva. Tribology Letters, 277–287.
Tang, W., & Bhushan, B. (2010). Adhesion, friction and wear characterization of skin and
skin cream using atomic force microscope. Colloids and Surfaces B: Biointerfaces,
76(1), 1–15.
Tang, W., Bhushan, B., & Ge, S. (2010). Friction, adhesion and durability and influence of
humidity on adhesion and surface charging of skin and various skin creams using
atomic force microscopy. Journal of Microscopy, 239(2), 99–116.
de Vicente, J., Stokes, J. R., & a Spikes, H. (2006). Rolling and sliding friction in compliant,
lubricated contact. Proceedings - Institution of Mechanical Engineers Part J J. Eng.
Tribol. 220(2), 55–63.
de Vicente, J., Stokes, J. R., & Spikes, H. A. (2005). Lubrication properties of non-adsorbing
polymer solutions in soft elastohydrodynamic (EHD) contacts. Tribology
International, 38(5), 515–526.
Vijay, A., Inui, T., Dodds, M., Proctor, G., & Carpenter, G. (2015). Factors that influence
the extensional rheological property of saliva. PLoS One, 10(8), 1–11.
Vorvolakos, K., & Chaudhury, M. K. (2003). The effects of molecular weight and temperature
on the kinetic friction of silicone rubbers. Langmuir, 19(17), 6778–6787.
Wang, Z., Volinsky, A. A., & Gallant, N. D. (2014). Crosslinking effect on polydimethylsiloxane
elastic modulus measured by custom-built compression instrument.
Journal of Applied Polymer Science, 131(22), 1–4.

顺应性口腔模拟唾液润滑的研究

由于伦理问题和获取生物组织的困难，寻找可作为研究用替代试验介质的合成弹性体非常重要。这方面的一个重要例子是摩擦测试，以了解在食用过程中（如糖浆、身体和清洁剂）口感属性背后的机制，这需要一个口头模拟。为了评估可能的材料对模拟口腔表面的适用性，通过将半球形硅胶针加载并滑动到聚二甲基硅氧烷（PDMS）、琼脂糖或猪舌样品上，产生滑动接触。测量摩擦，并根据试样的弹性模量计算弹流动力膜厚度，该弹性模量采用压痕法测量。以唾液和纯水为润滑液进行试验，并将结果与未润滑条件进行比较。PDMS在蛋白质粘附方面很好地模拟了舌头，当用唾液和水润滑时，两种样品的摩擦力都显著降低，而琼脂糖在唾液和水润滑条件下没有差异。这归因于PDMS的eOeSi（CH3）2-基团，与水合琼脂糖表面相比，它为唾液蛋白分子提供了良好的粘附性。然而，PDMS（2.2mpa）的测量模量明显大于舌（3.5kpa）和琼脂糖（66-174kpa）。这会影响低速滑动时的表面（边界）摩擦力，以及高速滑动时的夹带弹流膜厚度。利用透明的PDMS样品，我们还使用荧光显微镜来监测滑动过程中接触到的染色标记唾液蛋白的积累和流动。结果表明，唾液中蛋白质的摩擦强度信号与平均蛋白质强度信号之间存在很强的相关性（互相关系数=0.87），从而证实了唾液蛋白质的润滑边界膜形成性质。这是研究口感机制的有力方法。一。由于伦理问题和获取生物组织的困难，有必要寻找可作为研究用替代试验介质的合成弹性体。这方面的一个关键例子是摩擦测试，以了解在食用过程中（如糖浆、身体和清洁剂）口感属性背后的机制，这需要口头模拟。这一点很重要，因为食品和饮料的可接受性在很大程度上取决于其口感，而口感是摩擦学和流变学过程的结果（Stokes、Boehm和Baier，2013）。此外，对这些过程的不了解目前限制了健康配方的开发，这些配方可以复制食品，同时减少脂肪等成分（Dresselhuis，2008），（Drewnowski，1997）。在模拟口腔粘膜进行食品和饮料的体外摩擦学研究时，必须考虑粘膜膜。与牙齿上获得的珐琅质膜一样，这是唾液蛋白的一个子集，专门与口腔上皮细胞结合（Gibbins、Proctor、Yakubov、Wilson和Carpenter，2014）。与获得性釉质膜不同，粘膜膜主要由黏蛋白和分泌性IgA组成。这一层是由唾液粘蛋白（muc5b和muc7）与口腔上皮细胞表达的膜结合粘蛋白（muc1）相互作用驱动的（Vijay，Inui，Dodds，Proctor和Carpenter，2015）。粘蛋白是一种大的高度糖基化的蛋白质，在初分泌时能保留相当数量的水（Corfield，2015）。因此，除了唾液润滑表面外，在表面附近还有一个水凝胶状层。然而，唾液在体外试验中经常被忽略，因为它被认为不方便收集足够数量的唾液，或者被认为太复杂而无法给出一致的结果。以前，口腔粘膜是用玻璃或其他硬基质模拟的（Chen&Stokes，2012）。近年来，随着赫兹力学的引入，弹性基底被广泛应用于软摩擦学领域。在Dresselhuis等人的一项重要调查中。（Dresselhuis等人，2007），猪舌的表面特征与PDMS进行了比较。他们的研究结论是，由于口腔粘膜和PDMS橡胶，即使表面结构能够再现生物场景，在摩擦学实验中也不能互换，PDMS在表面特征上与舌表面的表现不同。然而，这篇被广泛引用的论文有一个严重的缺点，那就是它只使用乳状液作为润滑剂，而忽略了唾液的相互作用。在生物表面进行的其他工作，但是没有唾液的存在包括亚当斯等人的研究。（Adams、Briscoe和Johnson，2007）和Tang等人。（Tang&Bhushan，2010），（Tang，Bhushan，&Ge，2010）人类皮肤润滑特性研究。亚当斯等人。使用光滑的玻璃或聚丙烯，球形探针顶着人类前臂滑动，而Tang等人。用剃过的猪皮。结果报告了一系列润滑条件，但重复性的测试难以实现。Prinz等人。（Prinz，de Wijk和Huntjens，2007）研究了用人类唾液润滑的两个猪粘膜表面之间的摩擦特性。然而，缺乏数据，没有对不同的组分材料进行比较。在大多数研究中，选择交联聚二甲基硅氧烷（PDMS）是因为其弹性特性、易于处理和相对较低的硬度，可与软生物组织相比（Cox、Driessen、Boerboom、Bouten和Baaijens，2008；Khanafer、Duprey、Schlicht和Berguer，2009）。PDMS被用作一个（de Vicente，Stokes，&a Spikes，2006），（Malone，Appelqvist，&Norton，2003），（Tang&Bhushan，2010），（Tang et al.，2010）或两者（Stokes，Bongaerts，&Rossetti，2007），（Lee&Spencer，2005），（Bongaerts，Fourtouni，&Stokes，在摩擦学接触中保持低接触压力并为等粘弹性流体动力润滑（I-EHL）创造条件。PDMS的一个关键优点是易于制造，这也促进了PDMS的广泛应用。在交联之前，PDMS可以浇铸成几乎任何所需形状的合适模具。PDMS的其他优点包括其生理惰性、可用性、单位成本低以及良好的热稳定性和氧化稳定性。PDMS是一种透明的硅基有机聚合物，用于在许多摩擦学研究中代表生物材料（例如（Bongaerts等人，2007年）（Dresselhuis等人，2007年），（De Vicente，Spikes和Stokes，2004年）。由于其*的低玻璃化转变温度（Tg≈125°C），其弹性模量E≈0.57–3.7兆帕（取决于交联程度）（Wang、Volinsky和Gallant，2014）（Lótters、Olthuis、Veltink和Bergveld，1999）。PDMS的表面是疏水的，因为其重复的eOeSi（CH3）2-基团（Adams等人，2007），但可以通过等离子体清洗使其亲水。除此之外，PDMS在聚合物微流控领域（如Eddings，Johnson，&Gale，2008）得到了广泛的应用，该领域的研究成果可以在本研究中得到有益的应用。PDMS的摩擦学特性现在已经相当清楚了。Vorvolakos和Chaudhury（Vorvolakos和Chaudhury，2003）研究了分子量和试验温度对PDMS与金属表面纯滑动接触摩擦的影响。Bongaerts等人。（Bongaerts等人，2007）研究了PDMS表面粗糙度对生物聚合物和水溶液润滑性能的影响。PDMS，像大多数弹性体表面，本质上是疏水的，但是可以采用氧化处理来产生亲水表面。希尔堡等人。（Hillborg和Gedde，1998），（Hillborg、Sandelin和Gedde，2001）和Schneemilch等人。（SnEnMeLCH和奎克，2007）通过几种技术研究了PDMS在氧化前后的润湿性，研究了交联密度对氧化的影响。de Vicente等人。（de Vicente，斯托克斯，&Spikes，2005）研究了PDMS表面改性对其水润滑性能的影响。然而，对于PDMS作为生物表面模型的适用性仍存在一些争论，并且PDMS在唾液条件下被检测的例子很少。第二种软基质被认为是模拟口腔粘膜的潜在基质是琼脂糖。琼脂糖是从海藻中提取的琼脂中缺少琼脂蛋白的部分，由β-1,3-连接的α-半乳糖和α-1,4-连接的3,6-脱水-αL-半乳糖残基（Normand，Lootens，Amici，puckennet，&Aymard，2000）组成，用于制造水凝胶样基质。琼脂糖的顺应性随浓度变化很大，杨氏模量从1.5千帕到3千帕不等（Benkherourou、Rochas、Tracqui、Tranqui和Guméry，1999年），（Normand等人，2000年），（Chen、Suki和An，2004年）。此外，琼脂糖还具有悬浮培养细胞的能力，因此被用于组织培养系统（Chen等人，2004）。这种特性的结合使得琼脂糖在生物医学研究中成为一种有吸引力的选择，例如，作为软骨模拟物（Saris等人，2000），或作为磁共振弹性成像的模型材料（Muthupillai等人，1995）。因此，令人惊讶的是，琼脂糖在摩擦学研究中的应用很少，而且似乎*被忽略作为一种口服模拟物。费尔南德斯·法雷斯研究了其摩擦行为，但在葡萄糖和甘油润滑而非唾液润滑下进行了研究（费尔南德斯·法雷斯和诺顿，2015）。Shewan等人。近还研究了琼脂糖的润滑性能，但它是悬浮液中的颗粒，而不是基质（Shewan&Stokes，2015）。因此，能够模拟口腔粘膜表面是非常重要的，为此，人们对各种材料进行了研究。然而，它们很少与实际的生物材料进行比较（可能是因为它们在来源、保存和固定方面的困难），而且几乎从来没有被唾液润滑过。为了解决这一问题，本研究对聚二甲基硅氧烷（PDMS）、琼脂糖和猪舌的摩擦和膜厚性能进行了表征，目的是评估它们作为摩擦学试验的口腔模拟物的适用性。特别注意这些底物的顺应性和蛋白质结合行为。2。试验方法2.1。样品制备PDMS样品使用道康宁市市面上可买到的Sylgard 184试剂盒进行模压，该试剂盒含有一种碱和固化剂，以在23°C下产生杨氏模量为1.84兆帕的材料。琼脂糖凝胶是通过溶解粉状琼脂糖（Sigma-Aldrich，Poole，（UK）在1%或2%w/v的水中。为了帮助溶解，将溶液加热至90°C，然后在35°C左右冷却至低于螺旋转变的温度。此时，琼脂糖形成凝胶，由无限的三维纤维螺旋网络组成（诺曼德等人，2000）。在收集之前，受试者1小时内不进食和饮水。从单个受试者中收集静止的全口唾液（WMS），方法是将其流涎到预先称重的试管中，保存在冰上。收集后，对唾液进行短暂离心（3000 g，持续3分钟）以去除脱落的细胞和其他碎片。当天采集猪舌并进行检测。它的下侧被移除，形成一个平行的板。然后，使用氰基丙烯酸酯粘合剂将该试样粘合到平板上，并安装在摩擦装置中。G、 Carpenter等人。食品水胶体92（2019）10–18 11 2.2。压痕和表面粗糙度测量在Mach 1试验台（加拿大拉瓦尔市Biomomentum公司）上进行压痕试验，测量每个样品材料的弹性模量。这包括在三次重复试验期间，用半径为3.175 mm的球形压头以1 mm/s的速度压入样品，同时测量法向力和垂直位移。采用1.5 N单轴测压单元测量75μm的法向力，用0.1μm分辨率的钻台移动台测量垂直位移，每个样品的穿透深度为0.6mm，琼脂糖为1% W/V，0.4mm为0.4mm。这是根据Van Dommelen等人（Van Dommelen、Van der Sande、Hrapko和Peters，2010）的建议完成的，即如果压痕深度限制在样品厚度的10%以下，样品厚度不会对数据产生显著影响。然而，考虑到样本有限厚度的公式（Hayes、Keer、Herrmann和Mockros，1972）被用于计算杨氏模量。将接触力学方程拟合到数据中，得到杨氏模量，具体地说，χ=a d R 2（1）=−k P v a G d（1）4（2），其中d是压头的位移，R是压头的半径，a是接触区域的半径，P是施加的载荷，G是剪切模量，v是泊松比。试验示意图见图1。用Mach-1运动软件记录反作用力和压头位移，分别以P和d形式输入上述方程。假设泊松比等于0.5（不可压缩材料）。试样的高度h和压头半径R也已知。表2（Hayes et al.，1972）中给出了a、h和ν不同值的x和k值。在用方程（1）进行曲线拟合期间，估计接触区域a的半径。一旦拟合算法收敛，根据剪切模量G计算杨氏模量E，使用方程E=2G（1+v）（3）使用Veeco光学轮廓仪在表面上测量每个试样的粗糙度三次（每个试样在不同位置）。2.3条。蛋白质染色测量SDS-PAGE（十二烷基硫酸钠-聚丙烯酰胺凝胶电泳）用于评估不同蛋白质与口腔模拟物表面的结合程度。这种免疫印迹技术以带有特定染料的样品中的蛋白质为靶点，并通过凝胶测量它们在外加电场作用下的进展。就这样，在不同分子量的样品中分离出不同的蛋白质。染色包括在室温下用一名受试者的全口唾液孵育一小时。考马斯亮蓝（CBB）染色所有蛋白。此外，采用周期性酸Schiff's（PAS）对高糖基化蛋白进行染色，并用特异性抗体对唾液蛋白muc7进行敏感的化学发光检测。从每个口腔模拟表面的表面取下样本，并以这种方式测试哪些蛋白质存在。2.4条。摩擦测量使用CETR（美国坎贝尔）制造的UMT2（通用材料测试仪）将直径为5mm的二氧化硅半球加载并滑动到柔顺圆盘试样上，产生接触。该设备在销-盘模式下运行，使PDMS试样旋转，而二氧化硅半球保持静止。下试样位于旋转台上，能够（经过某些修改）以0.01转/分到4000转/分的速度运行。摩擦力（Fx）和法向载荷（Fz）是用应变计测量的，该应变计与固定硅半球试件上方的壳体连接。为此，选择了灵敏的低负载传感器，Fx和Fz的测量范围分别为±0.65n和±6n。该实验装置如图2a所示。在0.002至0.35 ms-1的速度范围内，在0.2 N.2.5的负载下记录摩擦数据。激光诱导荧光测量定制的激光诱导荧光（LIF）显微镜如图2b中的照片和示意图所示。它包括一个LED光源，该光源产生的光束通过透明PDMS样品聚焦到接触界面上。在某些测试中，润滑唾液中的蛋白质用染料异硫氰酸荧光素（FITC）标记，以便它们在被LED激发时发出荧光。发射的光被高速EMCCD摄像机过滤和收集。当薄膜厚度大于200nm时，从接触点发射的荧光光的记录强度与界面中液体的厚度成正比。这意味着摄像机采集的图像代表了显示接触中蛋白质分布的地图。荧光技术的更多细节可以在（Myant，Reddyhoff，&Spikes，2010），（Reddyhoff，Choo，Spikes，&Glovnea，2010）中找到。三。结果3.1。压痕和粗糙度结果图3显示了使用Biomomentum Mach-1试验机进行压痕试验期间四种材料的力-位移曲线。方程（1）–（3）应用于该数据，给出了表1所示的杨氏模量值。正如预期的那样，3.5千帕下猪舌的杨氏模量低于文献中发现的其他生物组织测量值，例如人类皮肤：25-101千帕（Akhtar、Sherratt、Cruickshank和Derby，2011年），人类肌肉：7千帕（McKee等人，2011年）。在1%和2%浓度下，这些数值 于模数为66和174kpa的琼脂糖。PDMS的模量几乎是图1的两个数量级。缩进设置示意图。高于生物样本。G、 Carpenter等人。食品水胶体92（2019）10 - 18 12表2示出了每一个样品的表面粗糙度测量，它们被分离大约一个数量级（PDMS<琼脂<舌）。然而，这种变化对摩擦力的影响被相反意义上增加的不同刚度抵消（例如，舌面上的小突起容易变平）。每次测量显示的数值范围是指标准误差，这是由于试样表面的变化而不是测量中的任何误差引起的。3.2条。蛋白质染色结果：当在室温下与来自单个受试者的全口唾液孵育1小时时，所有蛋白质的考马斯亮蓝（CBB）染色显示，琼脂糖和PDMS均未结合大量蛋白质，如图4所示。少量的淀粉酶是唾液中 的单一蛋白质，是 明显的蛋白质（根据表观分子量确定）。当用高糖基化蛋白的周期性酸Schiff's（PAS）染色同一凝胶时，琼脂糖凝胶中可见少量muc5b和muc7，但PDMS洗脱的样品中没有。用特异性抗体和敏感的化学发光检测对muc7进行免疫印迹，再次表明琼脂糖凝胶结合了一些粘蛋白，而PDMS没有。在琼脂糖中加入潜在的粘液粘合剂，如壳聚糖和凝集素WGA（AWGL），似乎可以增强蛋白质，特别是粘蛋白与琼脂糖的结合。3.3条。摩擦结果在这一段中，StbBek曲线绘制了X轴上的速度，而不是惯用的速度×粘度的乘积（de Vicente等人，2006）。这是因为唾液的粘度是高度非牛顿的，随着剪切速率的变化而变化很大（Rantonen&Meurman，1998），因此在每次测试中都不是恒定的。假设单一粘度的另一个障碍是唾液的非均质性和表面活性，这意味着无法假设唾液表面夹带的是高粘度蛋白质还是水分子。图5显示了琼脂糖玻璃接触的摩擦随滑动速度的变化。在无润滑和水润滑条件下，由于琼脂糖是一种水凝胶，压缩时会释放水分，因此这种基质表现出较低的摩擦力。当琼脂糖浸没在水中时，它表现出Stribeck曲线行为，低速时摩擦更大，由于形成弹性流体动力膜，摩擦随速度迅速减小。然而，当用唾液润滑时，与纯水相比，摩擦行为*没有变化。图6示出在不同条件下PDMS玻璃接触的摩擦随滑动速度的变化。当接触是无润滑的，摩擦系数保持在3和4之间，由于表面之间的强粘着相互作用。随着滑动速度的增加，摩擦随之减少，这可归因于弹性体的粘弹性特性（图2所示的摩擦）。激光诱导荧光装置，a）照片，b）压痕装置示意图。图3。压痕过程中获得的每种材料的力-位移曲线。表1各试验材料的杨氏模量结果（单位：kPa）。多孔舌琼脂糖（1%）琼脂糖（2%）PDMS 3.46 66.4 174 2270表2每种试验材料的表面粗糙度结果。附录中给出了相应的地表地形图示例。粗糙度（nm）平均值（Ra）RMS（Rq）舌片5480±667 656±403 PDMS 10.1±0.16 13.1±0.23琼脂1%399±91 514±109琼脂2%325±14 420±18 G。Carpenter等人。食品水胶体92（2019）10–18 13由于其粘弹性响应，PDMS的变形随速度的变化而变化）。在低速度下，干摩擦值和水下摩擦值相似，表明即使在水下，表面之间也不存在水（即没有形成边界膜）。这是因为速度不足以以流体动力学的方式分离表面，而且水分子不被PDMS或玻璃表面所吸引。相反，当全口唾液充满时，可以观察到非常低的摩擦力（比干燥情况下小两个数量级以上）。这些观察结果与斯托克斯和同事的观察结果一致（Bongaerts等人，2007）。图7显示了猪舌样品的摩擦与速度特性。当用纯水润滑时，该样品显示出高的边界摩擦，由于润滑剂夹带而随着速度降低。此外，与水相比，用唾液润滑时，低速边界摩擦显著减少。PDMS和舌头的干、水和唾液润滑曲线形状相似，但摩擦大小有显著差异。3.4条。激光诱导荧光结果PDMS比舌和琼脂糖样品的优点是它是透明的，这使得接触成像成为可能。为了证明这一点，图8中的LIF显微镜结果显示了FITC染色唾液蛋白在滑动过程中在接触面内的积聚和流动。图中a-d是显示蛋白质分布的接触强度图（这些是从作为补充材料提供的视频中获取的帧）。在这里，明亮的颜色代表了接触中高浓度的蛋白质，深蓝色的圆形区域是加压接触区域。由于从图形顶部的入口到底部的出口的滑动运动，不同形态的蛋白质聚集明显。该图还绘制了摩擦随时间的变化以及接触内荧光强度的测量值。后者是通过计算接触点内的像素数，其强度大于测试平均值（使用Matlab程序）。摩擦系数与接触区内蛋白质的存在有明显的相关性。图4突出显示了这一点。染色显示蛋白质与每个模拟表面结合。（注：全口唾液样本标记为WMS）。图5。琼脂糖圆盘在0.2N力作用下对固定二氧化硅半球的摩擦力与滑动速度a）线性标度，b）对数标度。G、 Carpenter等人。食品水胶体92（2019）10–18 14计算出的互相关系数为0.872，在一个信号中与另一个信号中的波谷（v）重合的可见波峰（以˄表示）出现，反之亦然。四。舌苔的硬度与琼脂糖的硬度比PDMS更接近。这意味着，对于琼脂糖接触，面积和压力与口腔中发现的更接近。此外，如果将其单独考虑，则表明分离琼脂糖表面的边界摩擦和流体动力膜厚度更为真实。但是，考虑唾液对口腔表面的润滑作用的粘膜膜也是很重要的，为了实现这一点，我们在琼脂糖凝胶中添加了粘着成分，以增强粘蛋白结合。在某些方面，这似乎是成功的与更多的所有唾液蛋白质，包括两个粘蛋白（muc 5b和muc7），结合在壳聚糖和WGA凝集素含有琼脂糖，显示出蛋白质染色。但粘着琼脂糖与单独琼脂糖相比，对摩擦学性能影响不大。事实上，水或唾液润滑的琼脂糖几乎没有区别。这表明，这种基底已经被表面本身润滑了——可能是在摩擦学配对的压力下，水从水凝胶中排出。此外，在摩擦主要由粘性阻力控制的全膜状态下，水和唾液润滑接触曲线的互换性表明，高粘度唾液蛋白甚至没有以高速夹带进入接触。PDMS的行为表现出更强的蛋白质相互作用。当在边界区域（即当液体的流体动力夹带不足以分离表面）以低速（∼0.1 mm/s）滑动时，唾液润滑时的摩擦系数比纯水润滑时的摩擦系数低两个数量级（∼0.01 vs∼2）。由于唾液是由99.5%的水和<0.5%的蛋白质分子组成的，这表明这些蛋白质是高效的表面活性润滑添加剂，它们粘附在PDMS和口腔表面上，形成lubricous低剪切强度界面。更具体地说，PDMS就像舌头是疏水的一样（Dresselhuis等人，2007），由于其带电的eOeSi（CH3）2-基团，它不分青红皂白地吸引蛋白质（Phillips&Cheng，2005）（事实上，生物蛋白质粘附到PDMS是生物实验室芯片系统中的一个问题（Phillips&Cheng，2005））。水和唾液之间的粘度差异（0.89 cP和∼5 cP（Rantonen和Meurman，1998））不足以解释这种差异。也可以假设，与水相比，PDMS的流体动力学/流变学响应的差异可能是由唾液的弹性引起的。然而，在如此低的速度下，弹性不应发挥作用。此外，如图所示，摩擦受到样品表面化学性质的强烈影响，而在全膜流体动力润滑条件下则不是这样。后，如Davies等人所示，此处测试的静止唾液的弹性明显低于酸刺激唾液的弹性（Davies，Wangling，&Stokes，2009）。舌头的干燥、水和唾液润滑曲线的形状与PDMS 相似，这支持后者使用口腔模拟。然而，在摩擦力的大小方面存在显著差异。在干燥、无润滑的条件下，如图6所示。在0.2N的力作用下，PDMS圆盘对固定二氧化硅半球的摩擦与滑动速度a）线性标度，b）对数标度。图7。猪舌在0.2n.a）线性标度，b）对数标度力作用下对静止二氧化硅半球的摩擦力与滑动速度的关系。G、 Carpenter等人。食品水胶体92（2019）10–18 15 PDMS显示摩擦系数约为3.5，而舌样为1.5。当用唾液代替水时，PDMS的摩擦力减小到0.02，而舌苔的摩擦力只有0.25。在低速条件下，当表面接触时，PDMS和舌片之间的摩擦系数大小差异可以分析如下。根据Schallamach（Schallamach，1958）和Roberts（Barquins&Roberts，2000）的预测，使用赫兹理论，干/边界润滑条件下的摩擦系数（即，当非液体分离表面时）由以下公式给出：μ=πS（9R/16E）2/3W−1/3（4），其中R是减小半径，e是弹性模量，S为界面剪应力，W为荷载。这表明，柔性材料之间的接触会产生更高的摩擦系数，因为这些材料会变形并产生更大的接触面积以进行剪切。方程（4）可用于计算边界润滑条件下接触点S内的剪应力，因为所有其他量都已知，其中舌和PDMS的值分别为0.53和3.2kpa。这表明，当唾液润滑时，PDMS表面的摩擦力较低，这是因为它的硬度较高，接触面积较小，但每单位面积的蛋白质覆盖舌头实际上更容易剪切。另一个因素是两个样品之间粗糙度的差异。在干燥条件下，较低粗糙度的PDMS增加了实际接触面积，从而增加了粘附力，而在蛋白质润滑下，较低粗糙度有助于形成完整的表面膜。唾液蛋白的高润滑性及其与PDMS表面的粘附性已被接触LIF结果证实。除了证明这项技术对研究唾液蛋白夹带的有效性外，这些结果还揭示了这一间歇过程的细节。更具体地说，观察到的蛋白质夹带的高度瞬态性质与Fan等人所证明的类似。（Fan，Myant，Underwood，Cann，&Hart，2011）他将接触中蛋白质的积累和分解归因于以下的入口聚集机制。由于润滑剂的接触几何形状和流动路径，蛋白质被输送到接触入口。其中一些蛋白质附着在会聚表面上。随着时间的推移，更多的蛋白质与表面的蛋白质分支纠缠在一起，在入口区形成一个更大的蛋白质团。然后到达一个临界点，在这个临界点上，表面摩擦力和润滑剂流体动力会导致蛋白质团分解，从而使蛋白质的大团块被拖进接触区。这可以在图8中观察到，在带有*符号的曲线图中突出显示，其中蛋白质峰值出现在摩擦系数小的情况下。图8。激光诱导荧光是在PDSM盘上加载二氧化硅半球并用FITC染色唾液润滑的滑动试验的结果。a） 空载接触的强度图，b）至d）滑动过程中的强度图，e）摩擦系数（蓝色）和荧光信号（橙色）的变化，通过计算强度大于试验平均值的像素数获得。为了突出相关性，示例峰值用ˆ标记，示例波谷用v标记。400 s左右的箭头突出显示两个信号中的对称趋势。注：在5和440s观察到的荧光阶跃变化对应于在接触的加载和卸载过程中接触蛋白的增加和减少。（为了解释本图图例中对颜色的引用，请参阅本文的网络版本。）G.Carpenter等人。食品类亲水胶体92（2019）10–18 16唾液与水相比的润滑特性差异被认为与唾液蛋白质如粘蛋白和statherin有关。粘蛋白有助于唾液的粘度，这可能有助于流体动力润滑模式（Bongaerts等人，2007），而statherin，一种小的表面活性蛋白被视为边界润滑剂（Douglas等人）（Harvey，Carpenter，Proctor和Klein，2011），虽然*有可能其他蛋白质也有助于润滑性能。5个。结论从表面化学的角度来看，PDMS适合于复制口腔粘膜，因为它和舌头一样，是疏水的（Dresselhuis，2008）及其带电基团，吸引蛋白质（Phillips&Cheng，2005）。这导致PMDS显示出与生物样品相似的摩擦与速度趋势。另一方面，当唾液和水润滑时，琼脂糖在摩擦力上的差别很小。这是由于水合琼脂糖表面弱粘附唾液蛋白所致。琼脂糖经粘胶组分增强粘蛋白结合后，其摩擦性能没有改善。虽然PDMS橡胶具有与舌头相似的疏水性质，但PDMS的弹性模量要大两个数量级。此外，即使交联度受到限制，PDMS的模量也仅降至约570千帕（Wang等人，2014年），而舌头的模量为3.4千帕。这是一个明显的缺点，因为样品的刚度同时影响边界摩擦（μαE′-2/3（Schallamach，1958））和弹性流体动力膜厚度（hαE′0.66（de Vicente等人，2005））。被测样本之间的粗糙度也有相当大的差异，琼脂糖与舌头 匹配。然而，由于较粗糙材料的非接触压扁，这对摩擦的影响是有限的，因为较粗糙材料具有较低的刚度。PDMS的一个优点是它是透明的，可以对唾液润滑机制进行接触成像。利用激光诱导荧光证实了这一点，由此产生的摩擦和蛋白质强度信号之间的强相关性（0.87）证实了唾液蛋白质的lubricous边界膜形成能力。蛋白质聚集在本质上是高度短暂的。应用这项技术来研究唾液与食品和饮料之间的摩擦学相互作用，以科学地描述口感特性是正在进行的研究课题。致谢S.K.Baier和R.V.Potineni受雇于百事公司。本研究文章中表达的观点是作者的观点，并不一定反映百事公司的立场或政策。本研究由百事公司资助（批准号：P55310-1）。附录A.补充数据本文的补充数据可在01.049在线查询。附录。表面形貌测量图A1。三种材料的表面形貌，使用Veeco光学轮廓仪测量，a）猪舌，b）PDMS，c）农酶。参考文献：Adams，M.J.，Briscoe，B.J.，和Johnson，S.a.（2007）。人体皮肤的摩擦和润滑。摩擦学快报，26（3），239-253。Akhtar，R.，Sherratt，M.J.，Cruickshank，J.K.和Derby，B.（2011年）。描述组织弹性特性的。材料今天，14（3），96-105。Barquins，M.和Roberts，A.D.（2000年）。橡胶摩擦随速率和温度的变化：一些新的观察结果。物理与应用物理学杂志，19（4），547-563。Benkherourou，M.，Rochas，C.，Tracqui，P.，Tranqui，L.和Guméry，P.Y.（1999年）。低浓度生物制剂表征方法的标准化：低浓度琼脂糖凝胶的弹性特性。生物力学工程杂志，121（2），184-187。Bongaerts，J.H.H.，Fourtouni，K.，和Stokes，J.R.（2007年）。软摩擦学：在一个兼容的PDMS-PDMS接触润滑。摩擦学，40（10-12），1531-1542规范。Chen，J.和Stokes，J.R.（2012年）。流变学和摩擦学：食物质感的两种不同状态。食品科学与技术趋势，25（1），4-12。Chen，Q.，Suki，B.，&An，K.-N.（2004年）。用分数阶导数模型模拟琼脂糖凝胶的动态力学性质。生物力学工程杂志，126（5），666-671。Corfield，A.P.（2015年）。粘蛋白：粘膜保护中与生物学相关的聚糖屏障。生物化学与生物物理学学报（BBA）-普通学科，1850（1），236–252。Cox，M.A.J.，Driessen，N.J.B.，Boerboom，R.A.，Bouten，C.V.C.和Baaijens，F.P.T.（2008年）。有限压痕法研究各向异性平面生物软组织的力学特性：实验可行性。生物力学杂志，41（2），422-429。Davies，G.A.，Wannilling，E.，和Stokes，J.R.（2009年）。饮料对唾液刺激和粘弹性的影响：与口感的关系？食品水胶体，23（8），2261-2269。De Vicente，J.，Spikes，H.A.和Stokes，J.R.（2004年）。软表面间润滑薄膜中的复杂流体。1到4。van Dommelen，J.A.W.，van der Sande，T.P.J.，Hrapko，M.和Peters，G.W.M.（2010年）。脑组织压痕力学特性：区域间差异。生物医学材料力学行为杂志，3（2），158-166。W、 H.Douglas，E.S.Reeh，N.Ramasubbu，P.A.Raj，K.K.Bhandary和M.J.Levine，“Statherin：人类唾液的主要边界润滑剂”，生物化学和生物物理研究通讯，第180卷，第91-97页。Dresselhuis，D.M.（2008年）。脂肪的命运：食品乳剂的摩擦学、粘附性和脂肪感觉。瓦赫宁根大学。Dresselhuis，D.M.，Klok，H.J.，Stuart，M.A.C.，De Vries，R.J.，Van Aken，G.A.，和De Hoog，E.H.A.（2007年）。o/w乳液在类口条件下的摩擦学：摩擦的决定因素。食品生物物理学，2（4），158-171。Drewnowski，A.（1997年）。为什么我们喜欢脂肪？美国饮食协会杂志，97（7），S58–S62。Eddings，M.A.，Johnson，M.A.，和Gale，B.K.（2008年）。确定微流控器件的 PDMS-PDMS键合技术。微机械与微工程学报，18（6），067001。Fan，J.，Myant，C.W.，Underwood，R.，Cann，P.M.和Hart，A.（2011年）。入口蛋白聚集：模型滑液润滑成膜的新机制。机械工程师学会论文集，H部分：G.Carpenter等人期刊。食品水胶体92（2019）10–18 17医学工程，225（7），696–709。费尔南德斯·法雷斯，I.和诺顿，I.T.（2015）。共溶质对琼脂凝胶摩擦学性能的影响。食品水胶体，45186195。Gibbins，H.L.，Proctor，G.B.，Yakubov，G.E.，Wilson，S.，和Carpenter，G.H.（2014年）。结合口腔粘膜膜内唾液保护蛋白的浓度。口腔疾病，20（7），707–713。Harvey，N.M.，Carpenter，G.H.，Proctor，G.B.，和Klein，J.（2011）。纯化人statherin吸附在光滑固体基质上的正常和摩擦相互作用。生物污染，27（8），823-835。Hayes，W.C.，Keer，L.M.，Herrmann，G.，和Mockros，L.F.（1972）。关节软骨压痕试验的数学分析。生物力学杂志，5（5），541-551。Hillborg，H.和Gedde，U.W.（1998年）。聚二甲基硅氧烷在电晕放电后的疏水性恢复。聚合物，39（10），1991-1998。Hillborg，H.，Sandelin，M.，和Gedde，U.W.（2001）。聚二甲基硅氧烷局部放电后的疏水性恢复与交联密度的关系。聚合物，42（17），7349–7362。Khanafer，K.，Duprey，A.，Schlicht，M.，和Berguer，R.（2009年）。应变率、混合比和应力-应变定义对聚二甲基硅氧烷（PDMS）材料力学行为与其生物应用的影响。生物医学。微型设备，11（2），503–508。Lee，S.和Spencer，N.D.（2005年）。聚合物的水润滑：表面改性的影响。摩擦学，38（11-12），922-930规格。Lótters，J.C.，Olthuis，W.，Veltink，P.H.，和Bergveld，P.（1999）。传感器用橡胶弹性聚合物聚二甲基硅氧烷的机械性能。微机械与微工程学报，7（3），145–147。Malone，M.E.，Appelqvist，I.A.M.，和Norton，I.T.（2003）。食品水胶体和乳状液的口服行为。 部分。润滑和沉积注意事项。食品水胶体，17（6），763-773。McKee，C.T.，Last，J.A.，Ph，D.，Russell，P.，Ph，D.，Murphy，C.J.，等人。（2011年）。软生物组织杨氏模量的压痕与拉伸测量。组织工程B部分综述，17（3），155-164。Muthupillai，R.，Lomas，D.，Rossman，P.，Greenleaf，J.，Manduca，A.，和Ehman，R.（1995）。声应变波传播的直接可视化磁共振弹性成像。科学（80），269（5232），1854-1857。Myant，C.，Reddyhoff，T.，和Spikes，H.A.（2010年）。纯滑动润滑柔顺接触中薄膜厚度映射的激光诱导荧光法。摩擦学，43（11），1960-1969。诺曼，V.，洛滕斯，D.L.，阿米奇，E.，普拉克内特，K.P.和艾马尔德，P.（2000）。琼脂糖凝胶力学性质的新认识。生物大分子，1730-738。Phillips，K.S.和Cheng，Q.（2005）。聚二甲基硅氧烷负载磷脂双层膜细菌毒素的微流控免疫分析分析化学，77（1），327–334。Prinz，J.F.，de Wijk，R.A.，和Huntjens，L.（2007年）。口腔粘膜摩擦系数的负荷依赖性。食品亲水胶体，21（3），402–408。Rantonen，P.J.F.和Meurman，J.H.（1998年）。整个唾液的粘度。斯堪的纳维亚齿科学报，56（4），210-214。Reddyhoff，T.，Choo，J.H.，Spikes，H.A.，和Glovnea，R.P.（2010年）。用荧光法研究弹流接触中的润滑油流动。摩擦学快报，38（3），207-215。Saris，D.B.，Mukherjee，N.，Berglund，L.J.，Schultz，F.M.，An，K.N.，和O'Driscoll，S.W.（2000年）。琼脂糖凝胶的动态压力传递。组织工程，6（5），531-537。Schallamach，A.（1958年）。橡胶的摩擦和磨损。磨损，1（5），384–417。Schneemilch，M.和Quirke，N.（2007年）。氧化对聚二甲基硅氧烷表面润湿性的影响《化学物理学杂志》，127（11）。Shewan，H.M.和Stokes，J.R.（2015年）。软球形微水凝胶悬浮液的粘度。胶体与界面科学杂志，4427581。Stokes，J.R.，Boehm，M.W.和Baier，S.K.（2013年）。口腔加工、质地和口感：从流变学到摩擦学及其他。胶体与界面科学的 观点，18（4），349–359。Stokes，J.R.，Bongaerts，J.H.H.，和Rossetti，D.（2007年）。人全唾液的润滑特性。摩擦学快报，277-287。Tang，W.和Bhushan，B.（2010年）。原子力显微镜对皮肤和护肤霜的粘附、摩擦和磨损特性的研究。胶体和表面B：生物界面，76（1），1-15。Tang，W.，Bhushan，B.，和Ge，S.（2010年）。利用原子力显微镜研究摩擦、附着力和耐久性以及湿度对皮肤和各种护肤霜附着力和表面电荷的影响。显微镜杂志，239（2），99-116。de Vicente，J.，Stokes，J.R.，&a Spikes，H.（2006年）。在柔顺、润滑的接触中滚动和滑动摩擦。机械工程师学会论文集第J部分。220（2），55–63。de Vicente，J.，Stokes，J.R.和Spikes，H.A.（2005年）。非吸附聚合物溶液在软弹流接触中的润滑特性。摩擦学，38（5），515–526。Vijay，A.，Inui，T.，Dodds，M.，Proctor，G.，和Carpenter，G.（2015年）。影响唾液拉伸流变特性的因素。公共科学图书馆一号，10（8），1-11。Vorvolakos，K.和Chaudhury，M.K.（2003年）。分子量和温度对硅橡胶动摩擦性能的影响。兰缪尔，19（17），6778-6787。Wang，Z.，Volinsky，A.A.和Gallant，N.D.（2014年）。自制压缩仪测定交联对聚二甲基硅氧烷弹性模量的影响。应用高分子科学杂志，131（22），1-4。