Tribological Properties of Poly(methyl methacrylate) Brushes Prepared by Surface-Initiated Atom Transfer Radical Polymerization

Abstract

Surface-initiated living radical polymerization of methyl methacrylate (MMA) using copper/bipyridyl complex was carried out from silicon wafer immobilized with a 2-bromoisobutylate moiety, resulting in the assembly of polymer chains tethered by one end to a surface, a so-called “polymer brush”. The thickness of the PMMA brush, which was estimated by ellipsometry and atomic force microscopy, increased linearly with the molecular weight of the chains, and was controlled by the free initiator concentration. The graft density of the PMMA brushes was estimated to be as high as 0.56 chains/nm2. The frictional properties of the high density PMMA brush were characterized by sliding the stainless ball probe on its surface across a width of 20 mm at a rate of 90 mm/min under loading of 0.49 N in air at room temperature. The PMMA brush was found to have a lower frictional coefficient and much better wear resistance than the corresponding spin-coated PMMA film because of the anchoring of the chain ends in the brush. In addition, the friction coefficient of the polymer brush significantly decreased in response to soaking in acetone and toluene, and increased in response to immersion in hexane. The tribological properties depended on the solvent quality.

Full text

Tribological Properties of Poly(methyl methacrylate)
Brushes Prepared by Surface-Initiated Atom
Transfer Radical Polymerization
Hiroki SAKATA, Motoyasu KOBAYASHI, Hideyuki OTSUKA, and Atsushi TAKAHARA
y
Institute for Materials Chemistry and Engineering, Kyushu University,
6-10-1 Hakozaki Higashi-ku, Fukuoka 812-8581, Japan
(Received May 16, 2005; Accepted July 4, 2005; Published October 15, 2005)
ABSTRACT: Surface-initiated living radical polymerization of methyl methacrylate (MMA) using copper/bipyr-
idyl complex was carried out from silicon wafer immobilized with a 2-bromoisobutylate moiety, resulting in the assem-
bly of polymer chains tethered by one end to a surface, a so-called ‘‘polymer brush’’. The thickness of the PMMA brush,
which was estimated by ellipsometry and atomic force microscopy, increased linearly with the molecular weight of the
chains, and was controlled by the free initiator concentration. The graft density of the PMMA brushes was estimated to
be as high as 0.56 chains/nm2. The frictional properties of the high density PMMA brush were characterized by sliding
the stainless ball probe on its surface across a width of 20 mm at a rate of 90mm/min under loading of 0.49 N in air at
room temperature. The PMMA brush was found to have a lower frictional coefficient and much better wear resistance
than the corresponding spin-coated PMMA film because of the anchoring of the chain ends in the brush. In addition, the
friction coefficient of the polymer brush significantly decreased in response to soaking in acetone and toluene, and
increased in response to immersion in hexane. The tribological properties depended on the solvent quality.
[DOI 10.1295/polymj.37.767]
KEY WORDS Tribology / Friction Coefficient / ATRP / Surface-Initiated Polymerization /
Poly(methyl methacrylate) (PMMA) / Polymer Brush / Wear Resistance /
Wear-resistant materials with low friction are
required for various modern technologies ranging
from a hard disk drives to roller bearings in automo-
bile. In particular nonlubricanted moving devices will
be increasingly needed for nano-machine technologies
in future. To improve the physicochemical properties
of solid surfaces, we have been investigating the tribo-
logical properties of ultrathin films prepared by chem-
ical vapor adsorption of organosilane compounds with
long alkyl chains, which are immobilized on a silicon
substrate by means of silyloxy bonds, forming a high-
density monolayer structure.
1
This immobilized
monolayer has a low friction coefficient and great
wear resistance, probably because of the low interfa-
cial energy of monolayer surface and strong adhesion
between monolayer and substrate. Consequently, the
following study focused on a surface-grafted polymer,
referred to as a ‘‘polymer brush’’.
Over the last decade, high-density and well-defined
polymer brushes have been readily synthesized, since
controlled/‘‘living’’ polymerization techniques
2–4
have been applied to the surface-initiated polymeriza-
tion.
5–7
One of the advantages of living polymeriza-
tion is quantitative initiation, which means that poly-
mers are propagated efficiently from the initiators on
a substrate. Practically, only 1 of 10 initiators bound
to the surface is expected to initiate a polymerization
because of the steric hindrance of the growing poly-
mer, however, it is sufficient to afford a high-density
brush. Atom transfer radical polymerization (ATRP)
has been widely employed for the formation of poly-
mer brushes because ATRP is compatible with various
functionalized monomers,
8–12
and the living character
of the ATRP process yields polymers with a low poly-
dispersity. Many researchers are investigating the sur-
face properties of tailored polymer brushes,
13
includ-
ing adhesion,
14
chromatography,
15
biomaterials,
16
wettability,
17
and biocompatibility.
18
Although the molecular mechanism of tribology on
the brush is still difficult to understand experimentally
due to the complexity of friction, lubrication, and
wear phenomena, some research groups have tried
to analyze the frictional properties of the polymer
brush. Klein et al. have found a reduction in the fric-
tional forces between solid surfaces bearing polymer
brushes using a newly developed surface force bal-
ance,
19,20
and they have also reported that brushes of
a charged polymer (polyelectrolyte) can act as effi-
cient lubricants between mica surfaces in an aqueous
medium.
21,22
Similarly, Osada and his coworkers have
reported that the well-defined polyelectrolyte brushes,
prepared by controlled radical polymerization using
TEMPO, reduce the surface friction of hydrogels
in water.
23
They have also found that the friction
y
To whom correspondence should be addressed (Tel: +81-92-642-2721, Fax: +81-92-642-2715, E-mail: takahara@cstf.kyushu-u.ac.jp).
767
Polymer Journal, Vol. 37, No. 10, pp. 767–775 (2005)

depends on the brush length; i.e., gels with longer
polymer brushes show higher friction. The lubrication
of a polymer brush is interesting from both a scientific
and a technological perspective. Tsujii et al. have
measured topographic images and force-distance pro-
files of high-density PMMA brushes by scanning force
microscopy (SFM) using a micro silica sphere attached
to a cantilever head.
24,25
The highly anisotropic struc-
ture of the swollen brush in toluene revealed extremely
strong resistance against compression. These results
imply the potential of the polymer brush to serve as
a low-friction and a wear-resistant film, although its
macro tribology and mechanical properties have not
yet been well studied, despite their very importance
with regard to practical applications. Therefore we
started a project to synthesize tethered polymer brush-
es on flat silicon substrate by surface-initiated ATRP
of MMA, investigating the sliding friction with a stain-
less probe and the wear resistance of the brush surface
under a load of 108Pa pressure. The various solvent
effects on the tethered brush during sliding friction
tests are also described in this paper.
EXPERIMENTAL
Materials
Anisole was stirred with sodium tips at 383 K for
6 h, followed by distillation from sodium under
reduced pressure. Copper bromide (CuBr, Wako Pure
Chemicals Industries Ltd.) was purified by washing
with acetic acid and dried under vacuum.
26
Methyl
methacrylate (MMA) purchased from Wako Pure
Chemicals was distilled under reduced pressure over
CaH2before use. 4,40-Di-n-heptyl-2,20-bipyridine
(Hbpy) was prepared by the dilithiation of 4,40-dimeth-
yl-2,20-bipyridine followed by coupling with 1-bromo-
hexane according to the method of Matyjaszewski
et al.
27
Ethyl 2-bromoisobutylate (EB), purchased
from Tokyo Chemical Inc., was distilled before use.
Water for contact angle measurements and frictional
tests was purified with the NanoPure Water system
(Millipore, Inc.). All other reagents were purchased
from commercial sources and used as received.
Synthesis of 60-dimethoxymethyliylyhexyl 2-bromoiso-
butylate (DMSB)
Bromoisobutyl bromide (51.5 mmol) was added
dropwise to a stirred solution of 6-undecenyl-1-ol
(51.2 mmol) and triethylamine (71.9 mmol) in dry di-
chloromethane (85 mL) under a nitrogen atmosphere
with cooling in an ice bath, and the solution was then
stirred overnight at room temperature (Scheme 1).
7
After the filtration with suction to remove the hydro-
chloric salt, the reaction solution was washed with
0.5 N HCl and water, and dried over MgSO4. The
product was distilled under reduced pressure (321–
322 K/0.1 mmHg) to give 60-hexenyl 2-bromoisobu-
tylate as a colorless liquid (31.9 mmol, 76%). The ob-
tained 60-undecenyl 2-bromoisobutylate (16.2 mmol)
and chloroplatinic acid (0.019 mmol, 2-propanol solu-
tion) were placed in a three-necked flask, dimethoxy-
methylsilane (32.6 mmol) was slowly added to the
mixture over a period of 30 min. During the reaction
mixture was stirred at 323 K for 12 h, the conversion
was checked by NMR. After unreacted dimethoxy-
methylsilane was removed under a vacuum, the solu-
tion was passed through a short column of anhydrous
Na2SO4to remove the catalyst, and distilled to give
12.8 mmol (79%) of 60-dimethoxymethylslilylhexyl
2-bromoisobutylate (DMSB) as a colorless oily resi-
due. Further purification was not carried out before
use. 400 MHz 1H NMR (CDCl3): 0.1 (SiCH3), 0.6
(SiCH2), 1.4 (–CH2–), 1.7 (OCH2CH2–), 1.9 (CH3),
3.5 (SiOCH3), 4.2 (COOCH2). 100 MHz 13 C NMR
(CDCl3): 5:7(SiCH3), 13.1 (SiCH2), 22.7 (SiCH2-
CH2), 25.5 (SiCH2CH2CH2), 28.3 (OCH2CH2), 30.9
(CH3), 32.8 (OCH2CH2CH2), 50.2 (SiOCH3), 56.1
(CBr), 66.2 (OCH2), 171.8 (C=O).
OH Br Br Br
C
O
C
CH3
CH3
Et3N
C
O
C
CH3
CH3
O
HSiMe(OMe)2
H2PtCl6
MMA
CuBr / Hbpy
Si wafer
CVA
MMA
CuBr / Hbpy
EB
O
O
Si(CH2)6OOC
Me
CBr
CH3
CH3
EtOOC C Br
CH3
CH3
EtOOC C CH2
CH3
CH3
C
CH3
Br
COOCH3
free polymer
O
O
Si(CH2)6OOC
Me
CCH
2
CH3
CH3
C
CH3
Br
COOCH3
n
C
O
CBr
CH3
CH3
(CH2)6O(CH3O)2Si
CH3
DMSB
PMMA brush
n
+
Scheme 1.
H. SAKATA et al.
768 Polym. J., Vol. 37, No. 10, 2005

Preparation of Initiator-Immobilized Silicon Substrate
The silicon(111) wafers (40 mm 8mm) were
immersed in a mixture of conc. H2SO4and 30% H2O2
aqueous solution (70/30, v/v) at 373 K for 1 h to
remove the organic contaminant from their surface.
Successively, substrates were further cleaned by expo-
sure to vacuum ultraviolet-ray (VUV, ¼172 nm)
for 5 min under reduced pressure at 15 mmHg to result
in the hydrophilic surface. These silicon wafers and a
glass vessel filled with 10% toluene solution of DMSB
were packed in a Teflon container purged with N2gas,
and were allowed to stand in an autoclave at 373 K for
5 h. During the heating at this temperature, DMSB
vapor adsorbed on the surface of the wafers to make
an organosilane monolayer, which is known as the
chemical vapor adsorption (CVA) method.
28,29
After
these wafers were rinsed with toluene and ethanol,
they were dried in vacuo at 373 K for 10 min and
was stored in a dark place.
Surface-Initiated ATRP
Typical polymer brush growth was achieved by
placing the DMSB-immobilized substrates in a glass
tube equipped with a stop cock under argon gas and
adding a degassed anisole solution of CuBr (0.020
mmol), Hbpy (0.040 mmol), MMA (50.0 mmol), and
ethyl 2-bromoisobutylate (EB, 0.010 mmol) as a free
initiator. The total volume of the polymerization solu-
tion was approximately 10 mL. The polymerization
solution was degassed by repeating the freeze-and-
thaw process, and the glass tube was sealed off under
the vacuum condition. The polymerization was then
allowed to proceed for a set reaction time (2–24 h)
at 363 K, and was terminated by cooling the solution
to 273 K and the addition of a small amount of meth-
anol under ambient pressure. The conversion of MMA
was estimated by 1H NMR spectra of the polymeriza-
tion solution, comparing the relative intensities of the
signals due to the unreacted monomer and the pro-
duced polymer. The silicon substrates were washed
with toluene using a Soxhlet apparatus for 12 h to re-
move the free polymer absorbed on their surfaces, and
were dried under the reduced pressure at 373 K for 1 h.
The polymer solution was passed through the alumina
column using THF to remove catalyst, and was poured
into the methanol to precipitate the free polymer. Us-
ing the obtained free polymer, spin coat films on sili-
con wafers were also prepared from toluene solution.
The thickness of the spin coat film was adjusted to that
of the corresponding polymer brush by changing the
spinning rate and concentration of toluene solution.
Measurements
The number-average molecular weights (Mn) and
molecular weight distribution (MWD) of the free
polymer were determined by size exclusion chroma-
tography (SEC) recorded on a Tosoh GPC-8010 sys-
tem using polystyrene standards calibration, which
runs through two directly connected polystyrene gel
columns (Shodex GPC KF-804L, 1.0 mL/min) using
THF as an eluent at 313 K. The NMR spectra were
measured in CDCl3with a Jeol EX-400 (1H 400
MHz) system. IR spectra were obtained with a Per-
kin-Elmer Spectra-One KY type (Perkin-Elmer) sys-
tem coupled with a Mercury Chromium Tell detector.
The incident angle of the p-polarized infrared beam to
the silicon wafer was 73.7(Brewstar angle). The
thickness of the polymer brush and the spin coat film
on the silicon substrate were determined by an imag-
ing ellipsometer (Nippon Laser & Electronics Lab.)
equipped with a YAG laser (532.8 nm). The polarizer
angle was fixed at 50, and a refractive index of 1.49
was used for the calculations of the film thickness.
Atomic force microscopic (AFM) observation was
done with SPA 400 with an SPI 3800N controller
(Seiko Instruments Industry Co., Ltd.) in air at room
temperature, using a Si3N4integrated tip on a com-
mercial triangle 100 mmcantilever (Olympus Co.,
Ltd.) with a spring constant of 0.09 N/m. XPS meas-
urements were carried out on a PHI ESCA 5800 (PHI
Electronics Co., Ltd.) at 105Pa using a monochro-
matic Al-KX-ray source. The contact angles against
water were recorded with a drop shape analysis sys-
tem DSA10 Mk2 (KRU
¨SS Inc.) equipped with a video
camera. The frictional coefficient of the polymer
brushes and the cast films were recorded on a Tribos-
tation Type32 (Shinto Scientific Co., Ltd.) by sliding a
stainless ball (10 mm) on the substrates over a width
of 20 mm at a sliding velocity of 90 mm/min under
loading of 0.20–0.98 N at 298 K (Figure 1). The fric-
tion force on the ball probe was transmitted to a stress
gauge attached to a probe end, and was recorded auto-
matically. Friction tests in various solvents were also
carried out using the polymer brush substrates, which
were immersed in the corresponding solvents for
24 h, in advance. The morphologies of the wear trace
of the thin films were observed with an S-4300SE
20 ~ 100 g
Stainless ball
PMMA brush
Silicon wafer
Load
Sliding platform
Friction Probe
Figure 1. Schematic description of the friction tester setup.
Tribological Properties of PMMA Brushes
Polym. J., Vol. 37, No. 10, 2005 769

field-emission SEM (Hitachi Co., Ltd.) equipped with
an X-ray microanalysis system (Genesis 7000, EDAX
Co., Ltd.) in order to examine the elements on the
wear surface.
RESULTS AND DISCUSSION
Surface-Initiated ATRP of MMA
The CVA method is often used to prepare the
high-density monolayer films of organosilane com-
pounds.
30,31
Takai et al. have carried out AFM analy-
sis of the monolayer film prepared by CVA, and they
found that the monolayer surface is very smooth, with
no aggregates and low number of defects.
32
We pre-
pared here the silicon substrate immobilized with rad-
ical initiator, DMSB, by the CVA method. The typical
water contact angle of silicon wafer irradiated by
VUV is lower than 5, but water contact angle on
the substrate increased to 87after the immobilization
of DMSB by CVA. XPS spectra of initiator-immobi-
lized silicon wafer showed a carbon signal (C1s)at
286 eV associated with the organic portion of the
attachable initiator along with the bromide (Br3d) sig-
nal at 71 eV. The C1s signals attributed to C=O and
C–O bonds were also observed in the narrow scan
mode. These results are indicative of formation of a
DMSB thin layer on the silicon wafer.
Surface-initiated radical polymerizations of MMA
were carried out in the presence of EB as a free initia-
tor coupled with CuBr and Hbpy (mole ratio EB/
CuBr/Hbpy/MMA = 1/2/4/1000). Figure 2 shows
the plots of Mnand the Mw=Mnindex of free polymer
produced in the solution as a function of monomer
conversion, where the Mnand MWD values were
determined by polystyrene-calibrated SEC. The poly-
dispersities of the free polymers were relatively low,
and the Mnvalues were proportional to the monomer
conversion, with the slope being very close to the
theoretical value. These findings indicate that the con-
trolled polymerization proceeded with a restriction of
transfer and termination reactions. The Mnof surface-
grafted PMMA on a silicon wafer cannot be directly
determined yet, however, a polymer brush should
have the same molecular weight as the value of the
corresponding free polymer.
24
As shown in Figure 3,
the thickness of the polymer brushes increased linear-
ly with molecular weight. The thickness of the ob-
tained polymer brush was smaller than the theoretical
value of the all-trans conformation given by 0:254N
and was larger than that of the random coil conforma-
tion calculated by 2ðNb2=6Þ1=2, where Nand bare the
degree of polymerization and the statistical segment
length of 0.68 nm, respectively.
33
According to the
proportional relationship between the thickness Ld
(nm) and Mn, the graft density was estimated to
be ca. 0.56 (chains/nm2) by following equation,
¼dLdNA1021=Mn
where dand NAare the assumed density of bulk
PMMA at 293 K and Avogadro’s number, respective-
ly. This graft density is comparatively high, taking
into account the volume fraction of the polymer chain.
Hence, the tethered PMMA chains would have a rela-
tively extended conformation along the direction nor-
mal to the substrate surface. The AFM observation
revealed that a homogeneous polymer layer was
formed on the substrate. The root mean square
(RMS) of the surface roughness was found to be ap-
proximately 1.0 nm in a 10 10 mm2scanning area
at any location. The water contact angle of PMMA
brush was 78, which was very close to the value
(80) for spin-coated PMMA film. The thickness of
the brush did not change in response to repeated rins-
ing with toluene using a Soxhlet apparatus; therefore,
the polymer chains were not physically adsorbed but
were chemically anchored on the substrate. Formation
of the PMMA brush was also confirmed with XPS and
0
10
20
30
40
50
1.0
1.5
2.0
2.5
0102030405060
Mn x 10-3
Mw / Mn
Conversion/ %
Mn (theo)
Figure 2. Mnand MWD of free PMMA obtained with EB/
CuBr/Hbpy in anisole at 363 K: ( )Mn;( ) MWD. The dashed
line indicates the theoretical value of Mngiven by the ratio of
[MMA]/[EB] and conversion.
0
10
20
30
40
0 1020304050
Thickness of PMMA brush / nm
Mn x 10-3
Figure 3. Relationship between brush thickness and Mnof
free polymer.
H. SAKATA et al.
770 Polym. J., Vol. 37, No. 10, 2005

Brewstar FT-IR spectra, although the results are not
shown here.
Friction Behavior in Dried State
Dynamic friction tests were carried out by sliding a
stainless ball on the substrates at a rate of 90 mm/min
in air under the normal load ranging from 20–100 g at
room temperature. In the case of a non modified sili-
con wafer under a normal load of 50 g (0.49 N), the
theoretical contact area between a stainless probe and
substrate can be calculated to be 2:43 109(m2)by
Hertz’s theory,
34
and the average pressure on the con-
tact area was estimated to be 201 MPa. Although the
actual contact area on the PMMA brush substrate
might be larger than the theoretical value, the average
pressure supposed to be more than 102MPa. We
attempted to demonstrate here whether the polymer
brush could resist such a high pressure and friction
for industrial application. As shown in Figure 4(a),
the dynamic friction coefficient of the polymer brush
was found to be 0.20–0.22 under a normal load of
20 g, and 0.24–0.25 under a normal load of 50 and
100 g. However, the magnitude of the friction coeffi-
cient of the polymer brush was independent of the
brush thickness and sliding velocity,
35
and was almost
constant at all normal loads. Large error margins were
observed in short brush around 5–20 nm, while the
margins of the error bars decreased in size with the
thickness of the brush layer, and the friction coeffi-
cient converged to a constant value. These results
indicate that stable sliding was achieved on the sub-
strates, and wear resistance was improved due to the
tethered polymer chain end by covalent bonds on
the substrate. On the other hand, the dynamic friction
coefficient of the cast film increased with film thick-
ness. In addition, a larger friction force was detected
than that of brush; for example, the friction coeffi-
cients of cast film and brush under loading of 100 g
were 0.31 and 0.24, respectively. The error margins
of the cast films were also larger than those of brush
at any film thickness. The thicknesses of these cast
films and the molecular weights of spin-coated poly-
mer were almost the same as those of the correspond-
ing polymer brushes; therefore, the grafting structure
of polymer brush should contribute to lowering the
friction, as described later.
To evaluate the wear resistance of polymer brush, a
continuous friction test for 600 s were performed by
sliding a stainless ball on substrates covered with
20 nm of polymer brush or cast film under a load of
0.49 N at a rate of 90 mm/min (Figure 5). In the early
stage of friction test, the friction coefficient of the
brush surface increased to 0.5 from 0.25, but there
was no further increase during continuous friction,
as shown in Figure 5(a). On the other hand, the
friction coefficient of the cast film was gradually
0.10
0.15
0.20
0.25
0.30
0 5 10 15 20 25 30 35
Brush [100g]
Brush [50g]
Brush [20g]
Friction coefficient
Thickness of PMMA brush / nm
0.10
0.15
0.20
0.25
0.30
0 5 10 15 20 25 30 35
Cast [100g]
Cast [50g]
Cast [20g]
Friction coefficient
Thickness of cast film / nm
(a)
(b)
Polymer brush
Cast film
Figure 4. Friction coefficient under dry condition of PMMA
brush (a) and cast film (b).
0.0
0.50
1.0
1.5
2.0
0 100 200 300 400 500 600
Friction coefficient
Sliding time /s
0.0
0.50
1.0
1.5
2.0
0 100 200 300 400 500 600
Friction coefficient
Sliding time /s
(a) Polymer brush
(b) Cast film
Figure 5. Friction time dependence of friction coefficient on
PMMA brush (a) and cast film (b) under a load of 0.49 N at a slid-
ing velocity of 90 mm/min in air: Mnof PMMA ¼26000; Thick-
ness of brush and cast film = 20 nm.
Tribological Properties of PMMA Brushes
Polym. J., Vol. 37, No. 10, 2005 771

increasing with friction time [Figure 5(b)], and the
surface polymer gradually peeled away to form many
debris (‘‘wear elements’’), some of which were ad-
sorbed on the surface of the probe (‘‘transfer parti-
cles’’) or left on the wear track, interfering with
smooth sliding of the probe. A sudden increase of fric-
tion coefficient to 1.5 was observed after 400 s, prob-
ably because the polymer layer on the wear track had
completely peeled off.
The trace scratched by the sliding probe can be seen
in the SEM image of wear tracks after the 600-s fric-
tion test (Figure 6). No topographic difference in the
width and morphology of the scratched tracks could
be found between the polymer brush and cast film.
An accumulation of wear elements produced by
scratching can also be observed at both ends of the
sliding trace in both pictures. However, the elemental
analysis showed different results. Signals due to car-
bon Kand oxygen Kcould be clearly observed at
the point of wear track on polymer brush by energy-
dispersive X-ray (EDX) spectra, which indicated that
the PMMA component remained. The observed peak
at 1.8 keV attributed to silicon Kmust have originat-
ed from the silicon substrate. As the matter of course,
all peaks attributed with carbon, oxygen, and silicon
atoms were observed at the outside of wear tracks.
On the contrary, peaks due to neither carbon nor oxy-
gen were observed in the EDX spectrum from the
wear track of the spin-cast film. Accordingly, a sliding
probe readily scratched the polymer layer on cast film
because the polymer molecules on cast film were not
anchored on the substrates. In the case of polymer
brush, polymer chains were immobilized on the sub-
strates and were highly extended in a perpendicular
direction for steric reasons related to the high graft
density.
36
We suppose that such a bound and aniso-
tropic structure restricted the mobility of the end-
grafted chain to prevent stretching and scattering by
the sliding probe. Of course, some of the brushes were
worn away under the strong pressure, as shown in
SEM image, although the rest of the tethered compo-
nents remained during the friction test to result in the
low magnitude of friction coefficient.
Effects of Solvent Quality on Friction Behavior
Since the early stages of polymer brush synthesis,
many researchers have suggested that the thickness
of polymer brush changes in response to soaking in
various solvents; the polymer brush can be stretched
in good solvents, and is compressed in poor solvents
or when in a dry state on the substrate. Furthermore,
the surface morphology and roughness can also be
arranged using solvation of polymer brushes. For
example, Zhao et al. have reported observing that dif-
ferent surface morphologies of polystyrene-block-
PMMA brush on the substrates layer by AFM after
treatment with cyclohexane and dichloromethane.
37,38
The solvent quality affects whether the polymer brush
will be stretched or compressed. We next attempted to
investigate the influence of solvent quality on the fric-
tion of polymer brushes. Friction tests of cast films
were not performed because the cast films can be
dissolved in organic solvents and removed from the
substrates.
The friction coefficients of the PMMA brush with
various thickness were measured by sliding a stainless
probe under a load of 50 g at room temperature in
acetone as a good solvent, and in n-hexane as a poor
0.0 1.0 2.0
Intensity / cps
/keV
C
O
Si
0.0 1.0 2.0
Intensity / cps
/ keV
C
Κ
α
Κ
α
Κ
α
Κ
α
Κ
α
Κ
α
O
Si
100µm
250 µm
250 µm
50 µm
Cast film
Brush film
(a)
(b)
Figure 6. EDX spectra and SEM images of wear track on the PMMA brush (a) and the cast film (b) after the sliding friction test in
Figure 5.
H. SAKATA et al.
772 Polym. J., Vol. 37, No. 10, 2005

solvent [Figure 7(a)]. Compared with the friction
coefficients of polymer brush in the dry state, the val-
ue decreased in both organic solvents due to the fluid
lubrication effect. The friction coefficient in acetone
was lower than that in hexane, and it decreased with
increases in the brush thickness. The friction coeffi-
cient of the PMMA brush with a 30-nm thickness
was 0.05, which is a half magnitude of the non modi-
fied substrate, suggesting that the polymer in good
solvents performs as a good lubricant. In n-hexane,
however, the friction coefficient of brush with any
thickness was almost as the same as that of non modi-
fied silicon substrate. This finding suggests that the
polymer brushes in poor solvents did not perform well
as lubricants, in other words, only fluid lubrication oc-
curred. Similar results were observed in toluene and
cyclohexane solution, as shown in Figure 7(b). The
dynamic friction coefficient in toluene approached to
a value of 0.06 on a 25-nm thickness brush, while
the friction coefficient for sliding friction in a cyclo-
hexane system was higher than that in toluene, main-
taining a constant value around 0.14, regardless of the
brush thickness. These results suggest that the interac-
tion between the brush surface and the stainless ball
(friction probe) was moderated because acetone and
toluene were good solvents for PMMA. On the other
hand, the PMMA brush surface would be unwilling
to be in contact with poor solvent such as hexane and
cylcohexane, and prefers to interact with the stainless
probe, thus giving a higher friction coefficient.
The wear resistance in solvents was also measured
under a normal load of 0.49 N and a sliding velocity of
90 mm/min for 10 min in toluene and cyclohexane.
As we expected, better wear resistance of the polymer
brush was observed in toluene compared with that in
cyclohexane. Figure 8 shows the SEM image and
EDX spectra of the worn surface after the wear-resist-
ance test. The wear track in toluene seems to be slight
compared with the scratched surface under the non
solvent condition in Figure 6. Almost no accumula-
tion of wear elements was observed at the sides and
ends of the sliding trace. Strong peaks due to carbon
and oxygen Kwere observed in EDX spectra from
the worn track surface, indicating that PMMA brush
components still covered the substrate surface. We
believe that the polymer brush, together with toluene,
worked as a lubricant, and reduced the interaction
between the stainless probe and the brush. The molec-
ular motion and characteristic structure of polymer
brush might impact the wear resistance, especially in
a solution. As mentioned above, polymer brush chains
in an good solvents typically form an extended struc-
ture along the perpendicular direction and show high
repulsion against the compression, while the polymer
chains tend to be compressed or collapse in poor sol-
vents.
39
However, brush structures in the presence and
absence of a solvent would be squashed by a probe
under as high pressure as 102MPa. Therefore, we sup-
pose that the sliding friction and wear resistance
0.0
0.050
0.10
0.15
0 5 10 15 20 25 30 35
Thickness of PMMA brush / nm
In cyclohexane
In toluene
Friction coefficient
0.0
0.050
0.10
0.15
0 5 10 15 20 25 30 35
Friction coefficient
Thickness of PMMA brush / nm
In hexane
In acetone
(a)
(b)
Figure 7. Friction coefficient of the PMMA brush in hexane
and acetone (a), in cyclohexane and toluene (b) under a load of
0.49 N at room temperature.
0.0 1.0 2.0 / keV
C
Κ
α
Κ
α
Κ
α
O
Si
Intensity / cps
In toluene
100 µm
100 µm
Figure 8. EDX spectra and SEM photograph of wear track on PMMA brush after sliding friction test in toluene under a load of 0.49 N
at a rate of 90 mm/min for 600 s: Mnof PMMA ¼28000; Thickness of brush = 20 nm.
Tribological Properties of PMMA Brushes
Polym. J., Vol. 37, No. 10, 2005 773

depend more on the interaction of the brush and probe
than the extended structure of polymer brush.
CONCLUSIONS
The high-density PMMA brushes were obtained by
surface-initiated atom transfer radical polymerization
using a silicon wafer covered with a flat monolayer
of 2-bromoisobutylate derivatives prepared by the
CVA method. According to the proportional relation-
ship between the thickness and Mnof the polymer
brush, the graft density was estimated to be 0.56
chains/nm2, which is the so-called high-density brush.
The friction coefficient of the polymer brush under a
normal load of 0.49 N in air at room temperature
was a lower than that of spin-coated film having the
same Mnand thickness as the corresponding polymer
brush. The polymer brush revealed better wear resist-
ance than the spin cast film because of the end-grafted
structure of the polymer brush. The tribological prop-
erties of brush in solution were dramatically changed
by the solvent quality. The friction coefficient of
PMMA brush in a good solvent such as toluene,
decreased further compared with those in air and in
a poor solvent such as hexane. In good solvent, poly-
mer brush performed as a lubricant to reduced interac-
tion between the probe and brush surface; as such, the
wear resistance was also significantly improved. From
these results, it can be concluded that polymer brush
prepared by surface-initiated polymerization exhibits
excellent tribological properties compared with those
prepared by spin-coating.
Acknowledgment. This work was partially sup-
ported by a Grant-in-Aid for the 21st century COE
Program ‘‘Functional Innovation of Molecular Infor-
matics’’ from the Ministry of Education, Culture,
Sports, Science and Technology of Japan. The FE-
SEM observation was performed using S-4300SE
(Hitachi Co., Ltd.) at the Collabo-station II, Kyushu
University.
REFERENCES
1. H. Ishida, T. Koga, M. Morita, H. Otsuka, and A. Takahara,
Tribology Lett.,19, 523 (2005).
2. K. Matyjaszewski and J. Xia, Chem. Rev.,101, 2921 (2001).
3. C. J. Hawker, A. W. Bosman, and E. Harth, Chem. Rev.,
101, 3661 (2001).
4. M. Kamigaito, T. Ando, and M. Sawamoto, Chem. Rev.,
101, 3689 (2001).
5. M. Ejaz, S. Yamamoto, K. Ohno, Y. Tsujii, and T. Fukuda,
Macromolecules,31, 5934 (1998).
6. B. Zhao and W. J. Brittain, J. Am. Chem. Soc.,121, 3557
(1999).
7. M. Husseman, E. E. Malmstrom, M. McNamara, M. Mate,
D. Mecerreyes, D. G. Benoit, J. L. Hedrick, P. Mansky,
E. Huang, T. P. Russell, and C. J. Hawker, Macromolecules,
32, 1424 (1999).
8. T. von Werne and T. E. Patten, J. Am. Chem. Soc.,121,
7409 (1999).
9. M. Ejaz, K. Ohno, Y. Tsujii, and T. Fukuda, Macromole-
cules,33, 2870 (2000).
10. C. Perruchot, M. A. Khan, A. Kamitsi, S. P. Armes,
T. von Werne, and T. E. Patten, Langmuir,17, 4479 (2001).
11. W. Huang, J. B. Kim, M. L. Bruening, and G. L. Baker,
Macromolecules,35, 1175 (2002).
12. M. Ejaz, S. Yamamoto, Y. Tsujii, and T. Fukuda, Macromo-
lecules,35, 1412 (2002).
13. K. Kato, M. Uchida, E.-T. Kang, Y. Uyama, and Y. Ikada,
Prog. Polym. Sci.,28, 209 (2003).
14. J. J. Lin, J. A. Silas, H. Bermudez, V. T. Milam, F. S. Bates,
and D. A. Hammer, Langmuir,20, 5493 (2004).
15. X. Huang and M. J. Wirth, Anal. Chem.,69, 4577 (1997).
16. T. Matsuda, M. Kaneko, and S. Ge, Biomaterials,24, 4507
(2003).
17. A. M. Granville and W. J. Brittain, Macromol. Rapid
Commun.,25, 1298 (2004).
18. R. Iwata, P. Suk-In, V. P. Hoven, A. Takahara, K. Akiyoshi,
and Y. Iwasaki, Biomacromolecules,5, 2308 (2004).
19. J. Klein, E. Kumacheva, D. Mahalu, D. Perahia, and L.
Fetters, Nature,370, 634 (1994).
20. R. Tadmor, J. Janik, and J. Klein, Phys. Rev. Lett.,91,
115503 (2003).
21. U. Raviv, S. Giasson, N. Kamph, J.-F. Gohy, R. Je
´ro
ˆme, and
J. Klein, Nature,425, 163 (2003).
22. J. Klein, U. Raviv, S. Perkin, N. Kampf, L. Chai, and S.
Giasson, J. Phys.: Condens. Matter,16, S5437 (2004).
23. Y. Ohsedo, R. Takashina, J. P. Gong, and Y. Osada, Lang-
muir,20, 6549 (2004).
24. S. Yamamoto, M. Ejaz, Y. Tsujii, M. Matsumoto, and T.
Fukuda, Macromolecules,33, 5602 (2000).
25. S. Yamamoto, M. Ejaz, Y. Tsujii, and T. Fukuda, Macromo-
lecules,33, 5608 (2000).
26. T. Grimaud and K. Matyjaszewski, Macromolecules,30,
2216 (1997).
27. K. Matyjaszewski, T. E. Patten, and J. Xia, J. Am. Chem.
Soc.,119, 674 (1997).
28. T. Koga, H. Otsuka, and A. Takahara, Chem. Lett.,31, 1196
(2002).
29. A. Takahara, H. Sakata, M. Morita, T. Koga, and H. Otsuka,
Compos. Interfaces,10, 489 (2003).
30. H. Tada and H. Nagayama, Langmuir,10, 1472 (1994).
31. K. Hayashi, N. Saito, H. Sugimura, O. Takai, and N.
Nagagiri, Langmuir,18, 7469 (2002).
32. A. Hozumi, Y. Yokogawa, Y. Kameyama, H. Sugimura, K.
Hayashi, H. Shinohara, and O. Takai, J. Vac. Sci. Technol.,
A,19, 1812 (2001).
33. K. Tanaka, K. Kojio, R. Kimura, A. Takahara, and T.
Kajiyama, Polym. J.,35, 44 (2003).
34. If a circle with radius a(m) is regarded as a contact area
between a stainless probe and substrate under a normal load
P(0.49 N), Hertz’s theory afford a following relationship
using Young’s modulus of stainless and silicon wafer, EA
H. SAKATA et al.
774 Polym. J., Vol. 37, No. 10, 2005

(1:96 1011 Pa), EB(1:30 1011 Pa), and Poisson’s ratio vA
(0.30), vB(0.28);
2=E¼ð1vA
2Þ=EAþð1vB
2Þ=EB
a¼ð3=42=EPRAÞ3=2
where RAis curvature radius (5:00 103m) of stainless
ball.
35. Unpublished data. Even if the stainless probe slided on the
surface of polymer brush with 20–30 nm thickness at rate
of 90 and 900 mm/min, the friction coefficients were almost
same within error margin.
36. S. Yamamoto, T. Tsujii, and T. Fukuda, Macromolecules,
35, 6077 (2002).
37. B. Zhao, W. J. Brittain, W. Zhou, and S. Z. G. Cheng,
J. Am. Chem. Soc.,122, 2307 (2000).
38. S. G. Boyes, W. J. Brittain, X. Weng, and S. Z. G. Cheng,
Macromolecules,35, 4960 (2002).
39. A. Roters, M. Schimmel, J. Ruhe, and D. Johannsmann,
Langmuir,14, 3999 (1998).
Tribological Properties of PMMA Brushes
Polym. J., Vol. 37, No. 10, 2005 775