Characterization and tribological investigation of Al2O3 and modified Al2O3 sol-gel films

Article abstract of DOI:10.1016/S0025-5408(01)00673-0

Thin films of Al₂O₃ and modified Al₂O₃ were prepared on a glass substrate by a dip-coating process from specially formulated aqueous sols. The tribological properties of the as-prepared thin films sliding against an Si₃N₄ ball were evaluated on a one-way reciprocating friction and wear tester. The worn surface morphologies and chemical compositions of the films were examined by means of atomic force microscopy (AFM), scanning electron microscopy (SEM), and X-ray photoelectron spectroscopy (XPS), respectively. Thermogravimetric analysis (TGA) and differential scanning calorimetric analysis (DSC) of dried sols were performed to explore the thermal events occurring during the annealing process of the sol-gel Al₂O₃ films. The wear mechanisms of the films are discussed based on SEM observation of the worn surface morphologies. As the results, modified sol-gel Al₂O₃ films are of higher wear resistance than the unmodified one. Meanwhile, the wear life of the modified films is correlated to the additive amount in the starting aqueous sol. In other words, the higher the additive amount in the starting aqueous gel, the longer the wear life of the sol-gel Al₂O₃ film. SEM observation of the morphologies of worn film surfaces indicates that wear of glass is characteristic of brittle fracture and severe abrasion. Wear of unmodified Al₂O₃ film is dominated by microfracture and abrasion, while that of modified films by plastic deformation and/or microcracking.

Full text of DOI:10.1016/S0025-5408(01)00673-0

Materials Research Bulletin 36 (2001) 1903–1914
Pergamon
Characterization and tribological investigation of Al2O3
and modified Al2O3 sol-gel films
Wenguang Zhang*, Weimin Liu, Qunji Xue
State Key Laboratory of Solid Lubrication, Lanzhou Institute of Chemical Physics,
Chinese Academy of Sciences, Lanzhou 730000, P.R. China
(Refereed)
Received 24 March 2000; accepted 17 August 2000
Abstract
Thin films of Al2O3 and modified Al2O3 were prepared on a glass substrate by a dip-coating
process from specially formulated aqueous sols. The tribological properties of the as-prepared thin
films sliding against an Si3N4 ball were evaluated on a one-way reciprocating friction and wear tester.
The worn surface morphologies and chemical compositions of the films were examined by means of
atomic force microscopy (AFM), scanning electron microscopy (SEM), and X-ray photoelectron
spectroscopy (XPS), respectively. Thermogravimetric analysis (TGA) and differential scanning cal-
orimetric analysis (DSC) of dried sols were performed to explore the thermal events occurring during
the annealing process of the sol-gel Al2O3 films. The wear mechanisms of the films are discussed
based on SEM observation of the worn surface morphologies. As the results, modified sol-gel Al2O3
films are of higher wear resistance than the unmodified one. Meanwhile, the wear life of the modified
films is correlated to the additive amount in the starting aqueous sol. In other words, the higher the
additive amount in the starting aqueous gel, the longer the wear life of the sol-gel Al2O3 film. SEM
observation of the morphologies of worn film surfaces indicates that wear of glass is characteristic of
brittle fracture and severe abrasion. Wear of unmodified Al2O3 film is dominated by microfracture and
abrasion, while that of modified films by plastic deformation and/or microcracking. © 2001 Elsevier
Science Ltd. All rights reserved.
Keywords: A. Thin films; A. Oxides; B. Sol-gel chemistry; D. Mechanical properties; D. Surface properties
1. Introduction
There have been many investigations of oxide films derived from the sol-gel method
because of several advantages, such as low processing temperature, homogeneity, the
* Corresponding author. Fax: ϩ86-931-8277088.
E-mail address: LSL@ns.LZb.ac.cn (W. Zhang).
0025-5408/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0025-5408(01)00673-0

1904
W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
possibility of coating on substrates with large areas, and low cost. Sol-gel oxide films have
found wide applications as functional materials in optical, microelectronics, and photoelec-
tronics industries, and for the purpose of protection as well [1–4]. It is known that the
tribological or mechanical properties of the films coated on the substrate will influence their
function, lifetime, and productivity [5]. Therefore the physical and mechanical properties of
the films have been largely focused on [6,7]. At the same time, the triblogical investigations
of ceramic films used as microdevice materials are drawing much attention of both tribolo-
gists and scientists engaged in material science [8]. Unfortunately, less is available about the
tribological properties of sol-gel films. Liu and co-workers found that sol-gel TiO2 thin films
had very good plasticity even at room temperature as well as good bonding ability with the
microscopic slide, which sets the prerequisite for the film to show excellent friction reduction
and wear resistant ability [9]. However, many basic problems, such as the relationship among
the process, structure, and the tribological properties of the film is still obscure.
Aluminum oxide thin films are widely used in many mechanical, optical, and microelectronic
applications because of the excellent properties such as chemical inertness, good mechanical
strength, and high hardness [10,11], transparency, high abrasive and corrosion resistance, as well
as insulating and optical property [12,13]. Usually, aluminum oxide film is prepared by tech-
niques such as chemical vapor deposition [14] and physical vapor deposition [15]. However, it
is difficult to prepare large, strong, and homogeneous film by these methods.
In this work we report on the fabrication, tribological behavior of Al2O3, modified Al2O3
thin films prepared by sol-gel, and dip-coating methods from inorganic salt. The relationship
among the process, structure, and the tribological properties of the sol-gel films is discussed
as well.
2. Experimental
2.1. Sample preparation
Crystal aluminum chloride (99.99%), ammonia, glacial acetic acid, and organic additive
polyethylene glycol (brand name P400, abridged as PEG, MW ϭ 400) were commercially
obtained and used without further purification as the raw materials for preparation of the aqueous
sol. The organic additive P400 was used as a kind of binder to improve the toughness of the
sol-gel ceramic film. In the preparation of the aqueous sol, excessive amount of ammonia solution
(5 M) was first dropped to an aqueous aluminum chloride solution (0.13 M) at room temperature
until a pH value of 9 was reached. The resultant hydrated precipitation was then filtered and
washed with distilled water and ethanol, respectively. Finally, the cleaned hydrated precipitation
was aged for 48 h and peptized with glacial acetic acid for 0.5 h at room temperature under
stirring to obtain the target translucent, homogeneous, and stable sols. The final pH value of the
aqueous sol was about 4. The aqueous sols (coded as A1, A2, and A3) for preparation of modified
ceramic coatings were prepared in the same way as above, except that the modifier P400 was
formulated in the starting solution at a volume fraction of 0.2, 0.4, and 0.6%, respectively. The
target aqueous sols were aged for 24 h and then served for preparation of the sol-gel Al2O3 films.
The glass substrate was first cleaned with ethanol–potassium hydroxide solution in an

W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
1905
ultrasonic bath for 20 min, then washed with distilled water and dried at ambient conditions.
Sol-gel Al2O3 films were prepared on the glass substrate by means of a dip-coating process.
This was realized by drawing and pulling of the substrate in the aqueous sols at a speed of
38 cm/min, accompanied by drying of the dipped substrate at room temperature for 15 min
and annealing at 500°C for 20 min in an oven. The target sol-gel Al2O3 films were finally
obtained by cooling of the sintered glass substrate to ambient temperature in the oven.
2.2. Experimental apparatus and measurements
The composition and structure of the films were characterized by means of TGA, DSC,
XPS, and AFM. DSC and TGA analyses were performed in nitrogen on a Perkin-Elmer
DSC-7 and TGA-7 system. The programmed heating rate was 10°C/min. The samples used
for thermal analysis were obtained by evaporation of the precursor solutions at room
temperature for 72 h. XPS analysis was conducted on a PHI-5702 XPS/AES system, using
Mg-K␣ radiation operating at 250 W and a pass energy of 29.35 eV. The binding energy of
C
1s (284.6 eV) was used as the reference. XPS depth profiling was performed with Arϩ ion
gun bombardment. The structure of the films was examined with an SPM-9500 atomic force
microscope, while the morphologies of the worn surfaces of the films observed with a
JSM-5600LV scanning electron microscope.
The tribological properties of the glass and coated glass sliding against a stationary ball
(diameter 3 mm, made of Si3N4) were evaluated on a Kyowa DF⅐PM model one-way
reciprocating friction tester at ambient conditions (RH: 40–44%). The sliding velocity and
stroke are 9 cm/min and 7 mm, respectively. The normal force is selected as 3, 2, and 1 N,
in anticipation that the sol-gel Al2O3 films would experience failure above a normal force of
3 N because of the excessive Hertzian contact stress. The physical and mechanical properties
of the counterpart Si3N4 ball are shown in Table 1. The coefficient of friction and sliding
passes were recorded automatically. A relatively low friction coefficient was recorded at the
early stage of sliding. It rose sharply to a higher stable value after sliding for a certain passes.
It is usually recognized that the film fails at this point. Therefore, the corresponding sliding
pass numbers were recorded as the wear life of the film. Three replicate tests were carried
out for each specimen, and the average friction coefficients and wear lives of the three
replicate tests were cited in this article. The relative error for the replicate tests was no more
than 5%. Prior to the friction and wear test, all the samples were cleaned in an ultrasonic bath
with ethanol and acetone for 10 min and then dried in hot air.
3. Results and discussion
3.1. Characterization of film precursor
Fig. 1a shows the DSC diagram of precursor A (without P400 as additive) and A3 (with
P400 additive at a volume fraction 0.6%). It is seen that precursor A shows two endothermic
peaks, while precursor A3 shows three endothermic peaks. A peak appears around 90°C for
both A and A3, indicating that the two precursors experience the same process at this

1906
W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
Fig. 1. DSC and TGA diagrams of the Al2O3 film precursor (A and A3).
temperature. The second process for A is delayed compared with A3, indicating that the
second process is easier to happen for A3. This is supposed to be related to the action of the
organic additive, which may act as a crosslinker. In other words, in the presence of the P400
additive, the molecules of sol are easier to crosslink with each other and enhance the
formation of Al2O3 particles in the subsequent heating process. Fig. 1b shows the TGA
diagram of precursor A and A3. In combination with the corresponding DSC analytical
results, it is inferred that the weight loss ratio for A and A3 is almost the same during the first

W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
1907
Fig. 2. XPS spectra of typical elements in as-prepared Al2O3 film (A) and modified Al2O3 film (A3).
phase, and then a larger weight loss is observed for A3, which is due to the loss of organic
compound or more coordinated water.
3.2. Characterization of the film
XPS is applied to detect the chemical states of some typical elements in as-prepared Al2O3
films (films A and A3). Fig. 2 shows the XPS spectra of Al2p and Si2p on the surface of A
and A3. The binding energy of Al2p at 74.20eV (Fig. 2a and c) is consistent with that of Al
in sapphire Al2O3 [16], indicating that target Al2O3 film has been successfully obtained in
our work. The signal of Si2p (Fig. 2b) is too weak to be detected, indicating that the surface
of film A is covered by Al2O3. The binding energy of Si2p at 102.5 eV (Fig. 2d) is assigned
to the glass substrate [16,17]. Fig. 3a shows the XPS depth profile of as-prepared Al2O3 film.
It is seen that the intensity of Al2p decreases gradually and that of Si2p and O1s increases with

1908
W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
Fig. 3. XPS depth profile of as-prepared Al2O3 film (A) and modified Al2O3 film (A3).
sputtering cycle increasing. Accordingly, it is concluded that diffusion took place between
the glass substrate and sol-gel Al2O3 film during the annealing process. This helps to increase
the bonding strength between the film and the substrate and, hence, improve the tribological
properties of the film. The peak of C1s disappears after two cycles of sputtering, illustrating
that contaminated carbon exists only in the very outer surface of the film, and the raw
materials were completely converted or removed during the heating process. Fig. 3b shows
the XPS depth profile of A3 film. Compared with film A, the most significant difference is
that the signal of C1s still exists even after five sputtering cycles, which is identified to the

W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
1909
Fig. 4. AMF images of Al2O3 films and corresponding friction force.
bulk film but not absorbed carbon. We suppose that such carbon originates from the additive
that was burned while heating and embedded in the film. This carbon may play a role to
modify the film structure. The second difference is that the intensity of Al2p in A3 keeps
stable and high with sputtering cycle extension, which indicates that the addition of P400
helps to generate thicker sol-gel Al2O3 film. The third difference is that the intensity of Si2p
in the first three cycles is higher compared with film A, indicating that the diffusion between
film A3 and the substrate is stronger. Thus, it can be concluded that P400 additive helps to
enhance the diffusion between the sol-gel film and the glass substrate as well.
Fig. 4 shows the AFM images and corresponding friction force pictures of the films. It can
be seen that the AFM images of A and A3 are different. Nano-size particles of film A arrange
in specific orientation, illustrating that the growth of unmodified Al2O3 nanoparticles is
accompanied by a preferential orientation. The surface of film A3 is relatively smooth and

1910
W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
Fig. 5. Friction coefficient of different couples as a function of wear passes near to failure.
some “island” distributes in the “marshland”, indicating that the addition of P400 has
changed the growing process of the film and led to totally different structure. From the
images of friction force of films A and A3 (Fig. 4b and d), it can be seen that the friction

W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
1911
Fig. 5 (continued)
force is correlated to the morphology to some extent. Namely, larger friction force is
recorded at the bulges of the film surfaces. Moreover, it is interesting to note that smoother
film A3 registers larger friction force than relatively rougher film A. This indicates that the
friction force of the sol-gel films is not only dependent on the morphological features but also
on other factors, especially the superficial and interfacial effects. As a matter of fact, the
friction at nanoscale or atomic level is largely dominated by surface forces including van der
Waals force, hydrogen bond force, and electrostatic and/or double layer forces [18]. Con-
sequently, the excitation of atomic lattice vibrations (phonons) and electronic excitations
contribute largely to the friction force at atomic level [19,20].

1912
W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
Fig. 6. SEM pictures of worn surfaces of substrate and Al2O3 and modified Al2O3 films.
3.3. Friction and wear
Fig. 5 shows the variation of friction coefficient as a function of sliding passes. It can be
seen that all the film samples record low initial friction coefficient about 0.1. It rises to a
stable and high value at a certain sliding pass, except that film A3 sliding against Si3N4 ball
registers a gradually decreased friction coefficient at extended sliding pass. All the sol-gel

W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
1913
Al2O3 films give considerably increased wear life compared with bare glass substrate, and
the longest wear life at load 1 N is recorded with film A3. However, the wear life of all the
film samples is short at load 3 N, indicating that the films have low load-carrying capacity.
The wear life of the samples under 1 N can be ranked as A3ϾA2ϾA1ϾA. In other words,
the higher the additive amount in the aqueous sol, the longer the wear life of the modified
sol-gel films. Besides, film A and the modified films show little difference in steady-state
friction coefficients, indicating that P400 as additive has little effect on the friction-reduction
behavior of the sol-gel Al2O3 films.
To gain more insights into the friction and wear mechanisms of the films, the worn
surfaces of the glass slide after 17 passes and of Al2O3 thin films sliding against Si3N4 after
failure have been observed by SEM. As shown in Fig.6, signs of severe brittle fracture are
visible on the worn surface of the glass slide even after 17 passes (Fig. 6a), accompanied by
the generation of large wear debris, which accounts for the abrasive wear of the slide and the
resultant high friction coefficient. Many microcracks are generated on the worn surface of
film A (Fig. 6b); the propagation of these cracks causes microfracture and leads to film
failure in a relatively short sliding duration. Contrary to the above, the worn surfaces of films
A1 and A3 show signs of slight plastic deformation (Fig. 6c and d), which is beneficial to
prevent the generation of microcracks. Besides, no sign of a microcrack is visible on the
worn surface of film A3 after sliding against Si3N4 at 3 N, while they are observed on the
worn surface of film A1 (Fig. 6c). This indicates that P400 additive in the aqueous sol helps
to improve the toughness of sol-gel Al2O3 film, which affects the spallation of film involving
the initiation and propagation of cracks in the film [21]. The above observation conforms
well to the friction and wear behaviors of the films. In other words, film A3 of higher
roughness shows better wear-resistance than film A1 of lower roughness. Moreover, com-
paring Fig. 6c with Fig. 6e and Fig. 6d with Fig. 6f, it is seen that film A1 or A3 shows
considerably different worn surface morphologies under different loads. Namely, the plastic
deformation and/or microcrack on the worn surfaces of films A1 and A3 are effectively
abated under relatively low loads (Fig. 6e and f). This agrees well with the extended wear
life of the two films at low load as well. In combination with the wear life of the films shown
in Fig. 5, it is rational to anticipate that film A3 would be a promising candidate for
protection and wear reduction of microdevices in relatively low-load applications.
4. Conclusions
Brittle Al2O3 ceramic having a high friction coefficient and severe wear under dry sliding
can be made into thin films of a low friction coefficient and long wear life by the sol-gel
process. Polyethylene glycol as an additive in the aqueous sol helps to improve the roughness
of the target sol-gel Al2O3 films, thereby to improve the tribological behavior of the films.
Wear of Al2O3 film is dominated by microcracking and fracture, that of modified Al2O3 films
characteristic of plastic deformation and/or microcracking. The plastic deformation is ben-
eficial to retard the propagation of fracture and peeling off of large wear debris; thus,
modified Al2O3 films show better wear resistance than unmodified Al2O3 film.

1914
W. Zhang et al. / Materials Research Bulletin 36 (2001) 1903–1914
Acknowledgments
The authors gratefully acknowledge the financial support from the National Natural
Science Foundation of China (Grant No.59825116) and the Ministry of Science and Tech-
nology of China (Grant No.95-Prereseach-07). They are also grateful to Engineer Jin-fang
Zhou, Ping-yu Zhang, and Shang-kui Qi for their assistance in TGA, DSC, AFM, and XPS
analyses.
References
[1] H. Dislich, J. Non-Cryst. Solids 57 (1983) 371.
[2] P.K. Nair, F. Mizukami, J. Nair, M. Salou, Y. Oosawa, H. Izutsu et al., Mater. Res. Bull. 33 (1998) 1495.
[3] G. Zhao, N. Tohge, Mater. Res. Bull. 33 (1998) 21.
[4] R. Pyper, Scho¨llhorn, J.J.T.M. Donkers, L.H.M. Krings, Mater. Res. Bull. 33 (1998) 1095.
[5] S. Hirai, K. Shimakage, M. Sekiguchi, K. Wada, A. Nukui, J. Am. Ceram. Soc. 82 (1999) 2011.
[6] M.J. Paterson, P.J.K. Paterson, B. Ben-Nissan, Tribology, 19 (1999) 166. (Chn).
[7] A.K. Jamting, J.M. Bell, M.V. Swain, L.S. Wielunski, R. Clissold, Thin Solid Films 332 (1998) 189.
[8] X.D. Li, B. Bhushan, Thin Solid Films 340 (1999) 210.
[9] W.M. Liu, W.G. Zhang, B. Li, G.X. Mai, J. Mater. Res., submitted.
[10] E. Do¨rre, H. Hu¨bner (Eds.), Alumina: Processings, Properties and Applications, Berlin: Springer-Verlag,
1984, p. 9.
[11] T.C. Chou, T.G. Nieh, S.D. McAdams, G.M. Pharr, Scripta Metal. Mater. 25 (1991) 2203.
[12] S.M. Edlou, A. Smajkiewicz, G.A. Al-Jumaily, Appl. Optics 32 (1993) 5601.
[13] K. Koski, J. Ho¨lsa¨, P. Juliet, Thin Solid Films 326 (1998) 189.
[14] J.H. Sun, E. Chang, B.C. Wu, Mater. Trans. JIM 34 (1993) 614.
[15] H. Bolt, F. Koch, J.L. Rodet, D. Karpov, S. Menzel, Surface Coat. Technol. 116–119 (1999) 956.
[16] C.D. Wagner, D.E. Passoja, H.F. Hillery, T.G. Kinisky, H.A. Six, W.T. Jansen et al., Sci. Technol. 21 (1982)
933.
[17] P.R. Anderson, W.E. Swartz, Inorg. Chem. 13 (1974) 2293.
[18] S.M. Hsu, Langmuir, 12, 4482 (1996)
[19] J. Krim, Langmuir 12 (1996) 4564.
[20] B.N. Persson, J. Phys, Rev. B 44, 3277 (1991). See also: L.S. Levitov, Europhys. Lett. 8 (1989) 499.
[21] Y.Q. Fu, X.D. Zhu, B. Tang, X.F. Hu, J.W. He, K.W. Xu et al., Wear 217 (1998) 159.