Tribology International 69 (2014) 3945

Contents lists available at ScienceDirect

Tribology International
journal homepage: www.elsevier.com/locate/triboint

Boundary and elastohydrodynamic lubrication studies of glycerol

aqueous solutions as green lubricants
Yijun Shi a,c,n, Ichiro Minami a, Mattias Grahn b, Marcus Bjrling a, Roland Larsson a
a

Division of Machine Elements, Lule University of Technology, SE97187 Lule, Sweden

Division of Sustainable Process Engineering, Lule University of Technology, SE97187 Lule, Sweden
c
Department of Chemistry and Materials Sciences, Institute of Science, Nanjing Forestry University, Nanjing 210037, China
b

art ic l e i nf o

a b s t r a c t

Article history:
Received 4 June 2013
Received in revised form
25 August 2013
Accepted 28 August 2013
Available online 4 September 2013

In this paper, the boundary and elastohydrodynamic lubricating behaviour of glycerol and its aqueous
solutions are discussed in both rolling and sliding contacts with a view on assessing the use of glycerol as
a green lubricant. To understand the lubricating mechanism, the lm thickness of glycerol and its
aqueous solutions were studied at different velocities. The results show that the viscosity of glycerol can
be controlled for a wide range by adding different amounts of water. The lubricating behaviour of
glycerol in all lubricating regimes can be improved by adding water. The results suggest that glycerol
aqueous solutions have great potential to replace rapeseed oils as environmentally friendly base oils in
several applications.
& 2013 Elsevier Ltd. All rights reserved.

Keywords:
Glycerol
Boundary lubrication
Elastohydrodynamic lubrication
Green lubricants

1. Introduction
It is estimated that every year more than 10 million tons of mineral
oil-based lubricants and hydraulic uids leak into the ground or
waterways, or are disposed of in the environment [1]. This is
problematic from an environmental point of view since mineralbased oils may inhibit the growth of plants, are toxic to aquatic life,
and can contaminate groundwater for up to 100 years [2]. One liter
of oil released into the environment can cover an area the size of a
football pitch and contaminate as much as one million liters of
water [3].
There is growing interest in the use of green lubricants. Vegetable
oils are potential replacements for mineral oils as base oils to solve the
pollution problem [4,5]. However, the low thermal and oxidation
stabilities, narrow viscosity range and poor ow properties at low
temperatures are major problems for vegetable oils when used as
lubricants [6,7]. On the other hand, glycerol aqueous solutions have
good biocompatibility, low cost, and better low temperature properties
than vegetable oils [8]. Glycerol may thus be a better green lubricant
than vegetable oils in some applications.
During the past decades there has been a tremendous increase
in the production of bio-diesel as a substitute for the traditional
mineral oil-based diesel. Conventional production of bio-diesel, of
the fatty acid methyl ester (FAME) type, is a chemical process

n
Corresponding author at: Division of Machine Elements, Lule University of
Technology, SE97187 Lule, Sweden. Tel.: 46 920 492064; fax: 46 920 491399.
E-mail address: yijun.shi@ltu.se (Y. Shi).

0301-679X/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.triboint.2013.08.013

based on transesterication of triglycerides such as vegetable oils

or animal tallow with methanol, which results in glycerol being
formed as an inevitable by-product [9]. The bio-diesel production
process generates approximately 1 kg of glycerol for every 10 kg of
bio-diesel produced [10]. The rapid growing production of biodiesel has generated an oversupply of glycerol in the market, and
novel applications of glycerol are therefore sought for.
Today, the main areas of glycerol usage are the production of
food additives, pharmaceuticals, cosmetics, personal care products,
and industrial applications [11,12]. Nevertheless, there is still a drive
to nd new application areas for glycerol to make use of the surplus
glycerol produced and thus increases the protability of bio-diesel
production.
Recently, Martin et al. found that glycerol can generate a very low
friction coefcient for boundary and elastohydrodynamic lubrication
[1315]. The friction coefcient could hardly be measured when
glycerol was used to lubricate the contacts involving two surfaces
coated with a hydrogen-free Diamond-Like Carbon coating [16]. X-ray
Photoelectron Spectroscopy (XPS), Time of Flight Secondary Ion Mass
Spectrometry (ToFSIMS), Karl Fischer titrations, NMR spectroscopy and
computer simulations were employed to show that the sp2-carbon
atoms on the uppermost surfaces of the coating react with glycerol
and that a tribo-degradation of glycerol occurs. This leads to the
formation of water. Martin et al. believe that the water lm is easily
sheared and that is the reason for the super low friction.
Martin et al. were focusing on the investigation of the friction
behaviour of pure glycerol. However, the freezing point of pure
glycerol is about 18 1C [8], which means that pure glycerol cannot be
used as a lubricant at low temperatures. The viscosity of pure glycerol

40

Y. Shi et al. / Tribology International 69 (2014) 3945

at room temperature is about 890 Pa s [17], which is about 20 times

higher than that of the most of mineral base oil. The high viscosity at
low temperatures is unwanted in most applications since too high
viscosity means that energy is wasted in shearing the thick lubricant.
The excess energy is converted into heat which increases both the
local temperature and the bulk lubricant temperature. This accelerates
lubricant decomposition reactions and mechanical failure of the
surfaces, leading to reduction in the lifetime of the lubricant and
equipment. Our tests showed that the viscosity of glycerol can be
controlled by blending different amounts of water with a quantity of
glycerol. The freezing point of glycerol can thus be lowered to about
50 1C [8]. Interestingly, Martin et al. also reported a simulation of the
tribological properties of glycerol using a thermal elastohydrodynamic
model, which showed that lm thickness and friction coefcient of
glycerol decreased with increasing water content in the elastohydrodynamic lubrication regime [15].
Lubrication may be divided into three regimes: (1) full lm or
elastohydrodynamic lubrication, (2) mixed lubrication and (3)
boundary lubrication [18]. The lubricating behaviour of these three
lubrication regimes differs greatly. In full lm lubrication, there is
enough hydrodynamic lift from the lubricant to keep the surfaces
separated, thus avoiding metal to metal contact. In this regime, the
load is carried by the lubricant, and friction and wear is low. In
boundary lubrication there is not enough hydrodynamic lift to
keep the surfaces separated: this may for instance occur when the
load is high or when the speed of the surfaces is low. The load is
consequently carried by the asperities (small summits) on the
contact surfaces. These contacts lead to high friction coefcients
(typically around 0.1) and high wear rates. In mixed lubrication the
load is carried both by the uid and the asperities. Most lubricants
are required to cover all lubrication regimes.
The boundary and elastohydrodynamic lubricating behaviour of
pure glycerol and glycerol aqueous solutions were studied to investigate the effect of water on the lubricating properties of glycerol. The
lm thickness of the glycerol aqueous solutions in the contact region
was measured in-situ to study the lubricating mechanism of glycerol
aqueous solutions. The performance of glycerol and its aqueous
solutions was also benchmarked against a representative for environmentally friendly base oils of today such as rapeseed oil.

prepared by mixing the glycerol with de-ionised water. The water

content in glycerol was in the range 550 wt%. The commercially
available rapeseed oil was supplied by Zeta Fernando di Luca,
Sweden, and was used as received.
2.2. Rheology study
The viscosities of glycerol and its aqueous solutions at different
shear rate were investigated using a Bohlin CVO 100 rheometer.
A concentric cylinder geometry was used with a 25 mm diameter
inner cylinder and a 27 mm diameter outer cylinder. During the
experiments, the temperature of the lubricant was maintained at
25 1C throughout the measurement. In each measurement, the
shear rate was increased logarithmically from 0.1 to 300 s  1.
2.3. Elastohydrodynamic lubrication study
A WAM (model 11) ball-on-disc test rig (Wedeven Associates,
Inc., Edgmont, PA, USA) was used for evaluating the elastohydrodynamic lubrication behaviour of glycerol and its aqueous solutions
at room temperature (ca 25 1C), at a load of 100 N (1.35 GPa
maxium Hertzian pressure), a speed of 1 m/s and a Slide-Roll Ratio
(SRR) between 0 and 30%. A schematic gure of this tribometer is
shown in Fig. 1. At slide-roll ratios larger than zero, the ball rotates
faster than the disc. Before each test, the device and specimens
were thoroughly cleaned with acetone and ethanol. Thereafter the
specimens, the ball and the disc, were assembled, and the entrainment speed was set to 1 m/s. Then relative position of the ball and
disc was corrected to pure rolling and the desired slide-roll ratio
was set. The disc was constantly lubricated using a recirculation
pumping system attached to the test rig. A full description of this
equipment can be found in paper [19].
All specimens used in the tests, both balls and discs were made
from AISI 52100 bearing steel. The balls are taken directly from the
factory and the discs are processed in the same way as bearing
raceway material. The balls are grade 20 with a surface roughness
(Ra) of 30 nm, 20.637 mm as outer diameter, and a hardness of
about 60 HRC. The discs have a surface roughness (Ra) of 90 nm,
101.6 mm as outer diameter, a circumferential grind and are through
hardened to about 60 HRC.

2. Experiment

2.4. Boundary lubrication study

2.1. Material

An Optimol SRV-III oscillating friction and wear tester was used

to evaluate friction-reducing and anti-wear properties of the
lubricants under boundary lubrication conditions, in accordance
with the ASTM D 6425 protocol. During the test, the upper steel

Glycerol (purity Z99%) used in this study was reagent grade

purchased from Sigma-Aldrich. The glycerol aqueous solutions were

CCD
Camera

Microscope

Steel ball

Steel ball
Steel disc
Lubricant

Pump

Steel disc
Lubricant

Pump

Fig. 1. Diagram of the ball on disc tribometer (WAM) for: (a) friction test; and (b) lm thickness test.

Y. Shi et al. / Tribology International 69 (2014) 3945

ball (100Cr6 steel, diameter 10 mm, surface roughness (Ra) 20 nm)

slides under reciprocating motion against a stationary steel disc
(100CR6 ESU hardened, 24 mm  7.9 mm, and surface roughness
(Ra) 120 nm). Both the ball and disc were supplied by Optimol
Instruments Prftechnik GmbH, Germany. Before each test the
device and specimens were cleaned with acetone and ethanol.
All tests were conducted under a load of 24 N (1.35 GPa Maxium
Hertzian pressure) at room temperature (ca 25 1C), a sliding frequency
of 50 Hz, and an amplitude of 1 mm. The friction coefcient curves
were recorded automatically with a data acquiring system linked to
the SRV-III tester. After the friction tests the wear volumes of the
lower discs were determined using an optical proling system
(Wyko NT1100, Veeco).

41

24 N (1.35 GPa Maxium Hertzian pressure) at room temperature

(ca 25 1C), a sliding frequency of 50 Hz, and an amplitude of 1 mm.
The friction coefcient curves were recorded automatically with a data
acquiring system linked to the SRV-III tester. After the friction tests the
wear volumes of the lower discs were determined using an optical
proling system (Wyko NT1100, Veeco).
Three duplicate rheological and tribological tests were conducted to minimise the data scattering. The typical value and the
average wear volume were reported in this paper.

3. Results and discussion

3.1. Rheological properties

2.6. Boundary lubrication study

An Optimol SRV-III oscillating friction and wear tester were used to
evaluate friction-reducing and anti-wear properties of the lubricants
under boundary lubrication conditions, in accordance with the ASTM
D 6425 protocol. During the test, the upper steel ball (100Cr6 steel,
diameter 10 mm, surface roughness (Ra) 20 nm) slides in reciprocating
motion against a stationary steel disc (100CR6 ESU hardened,
24 mm  7.9 mm, surface roughness (Ra) 120 nm). Both the ball
and disc were supplied by Optimol Instruments Prftechnik GmbH,
Germany. Before each test the device and specimens were cleaned
with acetone and ethanol. All tests were conducted under a load of

The viscosities of the glycerol solutions and the rapeseed oil as a

function of shear rate are shown in Fig. 2. The salient feature is that
the viscosity of all lubricants tested is constant with increasing shear
rate. This means that both the glycerol solutions and the rapeseed oil
1

Viscosity (Pa s)

The lm-forming characteristics of pure glycerol, its aqueous

solutions, and rapeseed oil were also investigated using the WAM-11
(Fig. 1 (b)). A super-polished steel ball with a roughness of 10 nm
(Ra) is brought into contact and loaded against a glass disc coated
with a chromium semi-reecting layer with a roughness of 1 nm
(Ra). The disc and the ball are driven by separate motors. During
the tests the steel ball is partially submerged in the lubricant.
The lubricant is dragged into the contact by the ball and forms a
lubricating lm [20].
A microscope and a CCD camera are used to capture the
chromatic interference pattern which occurs when the contact is
illuminated by white light. The thickness of the lubricant lm
formed in the contact area is calculated by the interference colour
evaluation technique [21]. Discs coated with Cr lm were used to
measure the lm thicknesses between 80 nm and 800 nm. To
obtain a high resolution at lm thicknesses below 80 nm discs
with an additional SiO2 spacer layer deposited on the Cr lm were
used. The elastic modulus and Poisson's ratio are 64 GPa and 0.2 for
the glass disc, and 210 GPa and 0.3 for the AISI 52100 steel ball. The
outer diameter of the steel ball was 20.625 mm and of the glass disc
was 114.3 mm. A load of 20 N was applied to generate a maximum
Hertzian contact pressure of 0.46 GPa and a mean contact pressure
of 0.31 GPa. The disc and the ball were run in the pure rolling mode,
and the entrainment speed was in the range 0.024 m/s. All tests
were carried out at room temperature (ca 25 1C).
Before each test, the refractive index (n) of the lubricant was
tested with a Zeiss Abbe Refractometer. The refractive index values
of the lubricants are reported in Table 1.

Pure
5wt% water
10wt% water
15wt% water
20wt% water
30wt% water
40wt% water
50wt% water
Rapeseed
oil

0.1

0.01

100

10

Shear rate (1/s)

Fig. 2. Rheological behaviour of glycerol, its aqueous solutions and rapeseed oil at 25 1C.

0.1

Friction coefficient

2.5. Film thickness investigation

0.01
pure
5 wt% water
10 wt% water
15 wt% water
20 wt% water

1E-3

10

15
20
SRR (%)

30 wt% water
40 wt% water
50 wt% water
Rapeseed
oil

25

30

Fig. 3. Friction coefcients as a function of SRR for different lubricants.

Table 1
The refractive index of pure glycerol, its aqueous solutions and rapeseed oil.
Lubricant

Pure glycerol

5 wt% Water

10 wt% Water

15 wt% Water

20 wt% Water

Refractive index
Lubricant
Refractive index

1.466
30 wt% Water
1.427

1.464
40 wt% Water
1.412

1.457
50 wt% Water
1.400

1.449
Rapeseed oil
1.447

1.443

42

Y. Shi et al. / Tribology International 69 (2014) 3945

behave as Newtonian uids. It can also be seen that the viscosity of

pure glycerol is very high: about 17 times higher than that of
rapeseed oil, which is unwanted as discussed previously. However,
the viscosity of the glycerol solutions decreases greatly with increasing water content. The viscosity of glycerol with a water content of
10 wt% is about a fth of that of pure glycerol. The viscosity of a
solution with 50 wt% water in glycerol is about 200 times lower than
that of pure glycerol. The rheological results show that the viscosity
of glycerol can be adjusted by varying water content.

It can also be seen from Fig. 3 that the friction coefcients of

glycerol solutions at different SRR decrease with increasing water
content, up to a content of 15 wt%. At higher water content, the
friction coefcient increases with increasing water content, probably
due to an increasing amount of mixed lubrication. This means that,
for a specied condition, there is an optimum amount of water which
is benecial for reducing the friction coefcient of the glycerol.

3.3. Boundary lubrication

0.26

Pure
5 wt% water
10 wt% water
15 wt% water
20 wt% water
30 wt% water
40 wt% water
50 wt% water
Rapeseed oil

0.24
0.22

Friction coefficient

Friction coefcients for glycerol aqueous solutions and rapeseed

oil under different SRR are shown in Fig. 3. Full lm lubrication can
be identied as those cases with constantly increasing friction as
SRR increases. Full lm lubrication is obtained at all SRR for
rapeseed oil and glycerol solutions when the water content is
20 wt% or less. At higher water content the viscosity of the glycerol
solution is not high enough to provide full separation of the surfaces
at all SRR. This is seen in Fig. 3 as the glycerol solutions with higher
water content than 20 wt% have much higher friction coefcients at
low SRR than the glycerol solutions with a water content above
20 wt%. For the uids with high (420 wt%) water content it was,
however, unexpected that the friction decreases at higher SRR.
A combination of the benecial boundary lubrication properties of
glycerol may lead to this effect. It may be explained by a polishing
effect obtained by improved running-in of the surfaces since the
test was run at increasingly higher and higher SRR. Evidence of
polishing was found by examining the surface roughness of the
specimen after the tests (Table 2). The surface roughness of the
track decreased for the specimen when the water content was above
20 wt%, while the specimens tested with water content below 20 wt%
showed almost the same roughness as the original surface.
The friction coefcient of rapeseed oil is about three times
higher than that of pure glycerol. Based on the viscosities of pure
glycerol and rapeseed oil presented in Fig. 2, where the viscosity of
pure glycerol is signicantly higher than that of the rapeseed oil, it
might be expected that the glycerol should provide higher friction.
The friction in partly sliding elastohydrodynamic lubrication is, in
fact, only marginally controlled by the viscosity at atmospheric
pressure and instead is strongly inuenced by the rheological
behaviour of the uid at the contact pressure [22]. For pure glycerol
and the glycerol aqueous solutions, the viscosity at the contact
centre will not increase signicantly since both glycerol and water
have low pressureviscosity coefcients of  5.4  10  9 Pa  1 for
glycerol [23] and about 8.0  10  10 Pa  1 for water (calculated
from the viscosity data in [24]), while the pressureviscosity
coefcient of rapeseed oil is 18.9  10  9 Pa  1 [25]. Pressure on
rapeseed oil therefore has a much greater inuence on the viscosity
of rapeseed oil so that there is a risk of an almost solid lm forming
at the contact centre under these conditions, which will greatly
increase the friction coefcient. Secondly, the limiting shear stress
of glycerol at GPa-level pressure is low [14], which will also lead to a
low friction coefcient. Martin et al. [13] explained the low limiting
shear stress of glycerol by the formation of an easily sheared water
layer formed in the contact region during sliding, which can greatly
reduce the friction.

The friction coefcient and wear volume loss generated under

reciprocating sliding motion are shown in Fig. 4 and Fig. 5,
respectively. It can be seen from Fig. 4 that the friction coefcient
of the glycerol solutions reaches a steady state after a running-in

0.20
0.18
0.16
0.14
0.12
0.10
0.08
0

30

60

90

120

Time (min)
Fig. 4. Typical friction coefcient as a function of test time for glycerol solutions
and rapeseed oil.
14

Wear volume loss ( 10-5 mm3)

3.2. Elastohydrodynamic lubrication

12
10
8
6
4
2

1.29

1.15

0
Pure

5wt%
water

10wt% 15wt% 20wt% 30wt% 40wt% 50wt% Rapeseed

water water water water
water water
Oil

Sample
Fig. 5. The wear volume loss of the disc lubricated by glycerol solutions and
rapeseed oil.

Table 2
Surface roughness (Ra-nm) of the track after WAM tests.
Lubricant

Native surface

Pure glycerol

5 wt% water

10 w.% water

15 w.% water

Ra (nm)
Lubricant
Ra (nm)

92
20 wt% water
80

90
30 wt% water
75

88
40 wt% water
65

86
50 wt% water
50

88
Rapeseed oil
87

Y. Shi et al. / Tribology International 69 (2014) 3945

3.4. Film thickness investigation

In order to understand more about the lubricating mechanism
of glycerol aqueous solutions and the differences in behaviour as
compared to rapeseed oil, the central lm thickness of the
lubricants as a function of the entrainment speed was studied.
According to previous studies, SRR has little effect on the central
lm thickness [26,27]. The lm thickness test here was therefore
run under pure rolling conditions.
The central lm thickness for the lubricants is shown in Fig. 6.
It can be seen that with a water content below 30 wt%, the central
lm thickness shows a linear relationship with the entraining
speed when plotted on a logarithmic scale from 2 mm/s to
3000 mm/s, which is similar to the behaviour of traditional
mineral oils. For 40 wt% and 50 wt% water content, the central
lm thickness shows a linear relationship from 40 mm/s to
3000 mm/s and from 100 mm/s to 3000 mm/s, respectively. The
deviation of central lm thickness at lower speeds may be due to
the SiO2 spacer layer on the glass disc being easily and quickly
scratched at lower entrainment speeds for glycerol at a water
content of 40 wt% and 50 wt%.

1000
800

Central film thickness (nm)

time of about 15 min when the water content is below 20 wt% and
also the stable friction coefcient obtained after the running in
period decreased with increasing water content. Martin et al. [13]
came up with a hypothesis that water produced inside of the
contact region was the reason for the low friction coefcient when
the friction surface is lubricated by pure glycerol. According to this
hypothesis, water is benecial for glycerol to decrease its friction
coefcient.
Fluctuation of the friction coefcient is observed when the
water content is higher than 30 wt%. This indicates that a stable
boundary lm cannot be formed when the water content is higher
than 20 wt%. This may be due to the decreasing viscosity of
glycerol aqueous solutions with the increase of water content,
which leads to the decrease of load carrying capacity. Glycerol
aqueous solutions cannot develop effective uid lm between
friction surfaces when the water content is above 20%. This results
in more solid-to solid contact during the test, which increases the
friction coefcient. At the same time, the wear of the friction
surface also becomes severe (Fig. 5). Wear particles generated may
then lead to a high and unstable friction coefcient.
Moreover, the friction coefcient at the steady state of rapeseed
oil is about 3050% higher than that of the glycerol aqueous solutions
when the water content is below 20%. Combined with the elastohydrodynamic lubricating results, it can be concluded that glycerol and
glycerol aqueous solutions with low water content have lower
friction behaviour than traditional environmental friendly lubricants
in all lubricating regimes under the conditions studied.
As shown in Fig. 5, the wear volume loss of the disc increased
with increasing water content for the glycerol aqueous solutions. It
may thus be concluded that water increases wear of material
though a small amount of water decreases the friction coefcient
effectively. This may be attributed to the corrosive behaviour of
water [15] and the decreasing load carrying capacity of glycerol
aqueous solutions with increasing water content.
The wear volume loss for glycerol with 10 wt% water is about
twice as high as that for pure glycerol, while for glycerol with
50 wt% water the wear volume loss is about 12 times higher than
for that lubricated by pure glycerol. The wear volume loss of the
disc lubricated by rapeseed oil is a little higher than that for pure
glycerol, which shows that the anti-wear property of glycerol is
better than that of rapeseed oil. Considering that the viscosity of
pure glycerol is much higher than that of pure rapeseed oil, the
better anti-wear property of pure glycerol should be attributed to
the viscosity effect.

43

100

pure
5%water
10%water
15%water
20%water
30%water
40%water
50%water
Rapeseed oil

10

1
1

10

100

1000

Entraining speed (mm/s)

Fig. 6. Central lm thickness in elastohydrodynamic lubrication contact for pure
glycerol, its aqueous solutions and rapeseed oil as a function of entraining speed.

The experimental data were tted using the following power

function which is derived from the HamrockDowson equation [28]:
H c a  U be

where Hc is the central lm thickness (nm), Ue is the entrainment

speed (mm/s), a and b are the dimensionless parameters. The tting
results are shown in Fig. 6 (in solid lines). All the tted lines show the
same slope using logarithmic scales. For both pure glycerol and
glycerol aqueous solutions, the lm thickness increased with the
entrainment speed at a rate of around U0.59, while the value is U0.66
for rapeseed oil. The lubricating lm formation of glycerol aqueous
solutions is therefore less sensitive to the entrainment speed than that
of rapeseed oil.
Considering the rheological results of the lubricant in Fig. 2, it
can also be seen from Fig. 6 that the central lm thickness has a
strong relationship with the viscosity of the lubricant. The central
lm thickness for pure glycerol at an entraining speed of 1 m/s is
higher than 800 nm (out of the operating range of the set-up) and
the value is about 310 nm and 119 nm for glycerol with 10 and
20 wt% water, respectively. The lm thickness for glycerol with 30,
40, and 50 wt% water was 94 nm, 50 nm and 27 nm, respectively
at an entrainment speed of 1 m/s. According to previous studies on
lm thickness it is evident that viscosity and entrainment speed
have a much greater effect on lm thickness than the contact
pressure [2628]. It is therefore concluded that the solutions
giving the lowest lm thicknesses in the interferometry test would
also give the lowest lm thicknesses in the higher contact pressure
steel on steel friction test. This would give weight to the hypothesis that the irregular behaviour for the glycerol solutions with
high water content in the friction test presented in Fig. 3 are due to
mixed lubrication.
It can also be seen that the lm thickness determined for the
20 wt% water lled glycerol is about two times lower than that of
rapeseed oil, while the viscosities of 20 wt% water lled glycerol
and rapeseed oil are similar to each other. Similar results were also
previously reported by Wan and Spikes [29,30], who attributed the
lower lm thickness to a reduction of pressureviscosity coefcient
(, will be discussed later) as water was added to polyglycol.
The pressureviscosity coefcient is one of the most important
factors which will affect the lm thickness in the elastohydrodynamic lubrication regime. Based on Van Leeuwen's work [31], the
HamrockDowson equation can be used to calculate the pressure
viscosity coefcient from lm thickness data and this equation is
considered to be superior to other equations used. Therefore, the

44

Y. Shi et al. / Tribology International 69 (2014) 3945

Table 3
Calculated pressureviscosity coefcients of glycerol aqueous solutions and rapeseed oil.
Lubricant

Pure glycerol

5 wt% Water

10 wt% Water

15 wt% Water

20 wt% Water

30 wt% Water

Rapeseed oil

(  10  9 Pa  1)

4.9

4.3

4.2

4.2

4.0

4.0

14.1

HamrockDowson equation [28] is employed here to calculate the

pressureviscosity coefcient of the glycerol aqueous solutions
and rapeseed oil. The HamrockDowson centre lm thickness
equations are described as follows:
H c 1:345Rx U 0:67 G0:53 W 0:067 C o
U

0 U e
0
Rx E

3
4. Conclusion

2 U FN

R2x U E0

G 2E0

5
0:64

C o 10:61e0:752Rx =Ry
Ue

U1 U2
2

that of rapeseed oil, which may be one of the reasons for the low
friction coefcient of glycerol aqueous solutions. As discussed in
Section 3.2, the friction coefcient for the lubricant in the elastohydrodynamic lubrication regime is governed by the viscosity of the
lubricant in the high pressure contact zone. The high of the
rapeseed oil causes much higher viscosity under these conditions,
which leads to higher friction coefcient [18].

6
7

where Hc is the central lm thickness (nm), Rx is the radius of

curvature of the ball in x-direction (mm), Ry is the radius of
curvature of the ball in y-direction (mm), U is the dimensionless
speed parameter, Ue is the entrainment speed (mm/s), U1 is the
ball speed (mm/s), U2 is the disc speed (mm/s), G is the geometry
parameter, W is the load parameter, Co is the ellipticity inuence,
0 is the viscosity at ambient pressure (Pa s), E is the equivalent
Young modulus (Pa), FN is the normal force (N) and is the
pressureviscosity coefcient (Pa  1).
The calculated pressureviscosity coefcients can be found in
Table 3. The calculated pressureviscosity coefcients for pure
glycerol is about 4.9  10  9 Pa  1 which is a little lower than the
value of 5.4  10  9 Pa  1 reported by Cook et al. [22]. This difference
may be due to the deviation of the relationship between lm
thickness and the entrainment speed for glycerol aqueous solutions.
The calculated for rapeseed oil is about 14.1  10  9 Pa  1,
obviously lower than the 18.9  10  9 Pa  1 reported by Hglund
[25]. However, this can be attributed to test conditions being
different in the work by Hglund (20 1C and 150 MPa) from those
of this work (25 1C and 460 MPa). The pressureviscosity coefcient
decreases with increasing temperature [25], and the pressure
viscosity coefcient also decreases with increasing pressure when
the pressure is below about 500 MPa [32]. The pressureviscosity
coefcient of glycerol is much lower than that of rapeseed oil. This
may be due to shorter molecular chains and shorter side chains of
glycerol than those of rapeseed oil. Glycerol thus does not have
many entanglements which can increase the pressureviscosity
coefcient obviously [33,34].
As can be seen from Table 3, the pressureviscosity coefcient of
glycerol is lower when water is added to the glycerol. Similar results
were seen for water/polyglycol mixture by Wan and Spike [29,30].
They suggested this effect arose from an intermolecular bonding
and that water formed adducts with polyglycol to give a structured
uid. However, different from their work, the pressureviscosity
coefcient for the glycerol aqueous solutions do not vary greatly
with water content. It may be due to the different size and shape of
molecular between glycerol and polyglycol resulting in less intermolecular bonding behaviour of glycerol. The pressureviscosity
coefcient of glycerol and its aqueous solutions is much lower than

In this work the lubricating properties of glycerol aqueous solutions

and the effect of water were studied to assess glycerol aqueous
solutions as candidates for green lubricants. The results were compared with those for rapeseed oil serving as a reference lubricant.
In general, both glycerol solutions and rapeseed oil behave as Newtonian uids under the shear rate conditions used here. The viscosity
of the glycerol can be controlled by adding water, and viscosity
decreases with increasing water content. The viscosity of glycerol with
a water content of 50 wt% is more than two orders of magnitude lower
than that of pure glycerol. For both elastohydrodynamic- and boundary lubrication, there is an optimum amount of water which is
benecial for reducing the friction coefcient. The friction coefcient
of the glycerol solutions is substantially lower than that of rapeseed oil
for elastohydrodynamic lubrication (more than three times lower) and
for boundary lubrication (3050% lower) under the conditions studied
though the viscosity of glycerol solutions is higher than that of
rapeseed oil. Water in the glycerol promoted the wear of steel. The
wear volume loss of the disc for glycerol solutions increased with
increasing water content. The wear volume loss of the disc for pure
glycerol is lower than that of rapeseed oil. The central lm thickness of
the glycerol aqueous solutions shows a linear relationship with
entrainment speed in logarithmic coordinates when the water content
is below 30 wt%. At 1 m/s entrainment speed, full lm elastohydrodynamic lubrication can be achieved when the water content in
glycerol is lower than 30 wt%. With higher water content, the lm
thickness will be decreased until mixed lubrication or boundary
lubrication occurs. Glycerol aqueous solutions have great potential as
green lubricants and their lubricating properties are much better than
those of rapeseed oil especially when the water content is below
20 wt%.

Acknowledgements
The authors are grateful for the nancial support of FORMAS
(Project no. 2009-925), the National Natural Science Foundation of
China (Project no. 51005123), the Priority Academic Programme
Development of Jiangsu Higher Education Institutions and the
fruitful discussions with Dr. Mikael strand at Lantmnnen, AB,
Sweden.
References
[1] Wilson B. Lubricants and functional uids from renewable sources. Industrial
Lubrication Tribology 1998;50:615.
[2] Bartz WJ. Lubricants and the environment. Tribology International 1998;31:3547.
[3] Spilsbury R, Spilsbury L. The oil industry. The Rosen Publishing Group, Inc.;
2012.

Y. Shi et al. / Tribology International 69 (2014) 3945

[4] Gong QY, He WR, Liu WM. The tribological behaviour of thiophosphates as
additives in rapeseed oil. Tribology International 2003;36:7338.
[5] Zhu FK, Fan WX, Wang AR, Zhu Y. Tribological study of novel SN style 1,3,4thiadiazole-2-thione derivatives in rapeseed oil. Wear 2009;266:2338.
[6] Quinchiac LA, Delgadoa MA, Francoa JM, Spikes HA, Gallegos C. Lowtemperature ow behaviour of vegetable oil-based lubricants. Industrial Crops
and Products 2012;37:3838.
[7] Fox NJ, Stachowiak GW. Vegetable oil-based lubricantsA review of oxidation.
Tribology International 2007;40:103546.
[8] Trejo JA, Longinotti MP, Corti HR. The viscosity of glycerol-water mixtures
including the supercooled region. Journal of Chemical & Engineering Data
2011;56:1397406.
[9] Katryniok B, Kimura H, Skrzynsk E, Girardon JS, Fongarland P, Capron M,
Ducoulombier R, Mimura N, Paul S, Dumeignil F. Selective catalytic oxidation
of glycerol: perspectives for high value chemicals. Green Chemistry
2011;13:196079.
[10] Chi Z, Pyle D, Wen Z, Frear C, Chen S. A laboratory study of producing
docosahexaenoic acid from biodieselwaste glycerol by microalgae fermentation. Process Biochemistry 2007;42:153745.
[11] Pagliaro M and Rossi M. The Future of Glycerol. 2nd edn (RSC Green Chemistry
Book Series): 2010.
[12] Chheda JN, Huber GW, Dumesic JA. Liquid-phase catalytic processing of
biomass-derived oxygenated hydrocarbons to fuels and chemicals. Angewandte Chemie International Edition 2007;46:716483.
[13] Matta C, Joly-Pottuz L, De Barros MI, Martin JM. Superlubricity and tribochemistry of polyhydric alcohols. Physical Review B 2008;78:08543643.
[14] Joly-Pottuz L, Martin JM, De Barros MI, Belin M. Anomalous low friction under
boundary lubrication of steel surfaces by polyols. Tribology Letters 2009;34:219.
[15] Habchi W, Matta C, Joly-Pottuz L, De Barros MI, Martin JM, Vergne P. Boundary
lubrication and tribochemistry in steel circular contacts lubricated with
glycerol. Tribology Letters 2011;42:3518.
[16] De Barros MI, Matta C, Th. Le-Mogne, Michel JM, Sagawa T, Okuda S, Kano M.
Improved mixed and boundary lubrication with glyceroldiamond technology.
TribologyMaterials, Surfaces & Interfaces 2007;1:2832.
[17] Nalam PC, Clasohm JN, Mashaghi A, Spencer ND. Macrotribological studies of
poly(L-lysine)-graft-poly(ethylene glycol) in aqueous glycerol mixtures. Tribology Letters 2010;37:54152.
[18] Jones MH, Scott D. Industrial Tribology, the Practical Aspects of Friction,
Lubrication and Wear. Amsterdam: Elsevier; 1983.

45

[19] Bjorling M, Larsson R, Marklund P, Kassfeldt E. Ehl friction mapping: the inuence
of lubricant, roughness, speed and slide to roll ratio. Proceedings of the Institution
of Mechanical Engineers, Part J: Journal of Engineering Tribology 2011
;225:67181.
[20] Lord J, Jolkin A, Larsson R, Marklund O. A hybrid lm thickness evaluation
scheme based on multi-channel interferometry and contact mechanics.
Journal of Tribology 2000;122:1622.
[21] Hartl M, Krupka I, Poliscuk R, Liska M. Computer-aided chromatic interferometry. Computer & Graphics 1998;22:2038.
[22] Bair S, Liu Y, Wang QJ. The pressureviscosity coefcient for newtonian EHL
lm thickness with general piezoviscous response. Journal of Tribology
2006;128:62431.
[23] Cook RL, King HE, Herbst CA, Herschbach DR. Pressure and temperature
dependent viscosity of two glass forming liquids: glycerol and dibutyl
phthalate. Journal of Chemical Physics 1994;100:517889.
[24] Bridgman PW. The viscosity of liquids under pressure. Proceedings of the
National Academy of Sciences USA 1925;11:6036.
[25] Hglund E. Inuence of lubricant properties on elastohydrodynamic lubrication. Wear 1999;232:17684.
[26] Gold PW, Schmidt A, Dicke H, Loos J, Assmann C. Viscositypressuretemperature behaviour of mineral and synthetic oils. Journal of Synthetic
Lubrication 2001;18:5179.
[27] Nijenbanning G, Venner CH, Mose H. Film thickness in elastohydrodynamically lubricated elliptic contacts. Wear 1994;176:21729.
[28] Hamrock BJ, Dowson D. Ball Bearing Lubrication: The Elastohydrodynamics
Elliptical Contacts. New York: Wiley; 1981.
[29] Wan GTY, Spikes HA. The elastohydrodynamic lubricating properties of waterpolyglycol re-resistant uids. ASLE Transactions 1984;27:36672.
[30] Wan GTY, Kenny P, Spikes HA. Elastohydrodynamic properties of water-based
re-resistant hydraulic uids. Tribology International 1984;17:30915.
[31] Van LH. The determination of the pressureviscosity coefcient of a lubricant
through an accurate lm thickness formula and accurate lm thickness
measurements. In: Proceedings of the institution of mechanical engineers,
Part J. Journal of Engineering Tribology 2009;223:114363.
[32] Bair S, Khonsari M, Winer WO. High-pressure rheology of lubricants and
limitations of the Reynolds equation. Tribology International 1998;31:57386.
[33] Biresaw G, Bantchev GB. Pressure viscosity coefcient of vegetable oils.
Tribology Letters 2013;49:50112.
[34] Pensado AS, Comuas MJP, Fernndez J. The pressureviscosity coefcient of
several ionic liquids. Tribology Letters 2008;31:10718.