Friction and wear of titanium alloys sliding against metal, polymer, and ceramic counterfaces

Friction and wear of titanium alloys sliding against metal, polymer, and ceramic counterfaces

Wear 258 (2005) 1348–1356 Friction and wear of titanium alloys sliding against metal, polymer, and ceramic counterfaces夽 Jun Qua,∗ , Peter J. Blaua ,...

619KB Sizes 0 Downloads 18 Views

Wear 258 (2005) 1348–1356

Friction and wear of titanium alloys sliding against metal, polymer, and ceramic counterfaces夽 Jun Qua,∗ , Peter J. Blaua , Thomas R. Watkinsa , Odis B. Cavinb , Nagraj S. Kulkarnia a

Metals and Ceramics Division, Oak Ridge National Laboratory, P. O. Box 2008, MS 6063, Oak Ridge, USA b University of Tennessee, Knoxville, USA Received 11 July 2003; received in revised form 21 September 2004; accepted 23 September 2004 Available online 11 November 2004

Abstract Recent advances in lower-cost processing of titanium, coupled with its potential use as a light weight material in engines and brakes has renewed interest in the tribological behavior of titanium alloys. To help establish a baseline for further studies on the tribology of titanium against various classes of counterface materials, pin-on-disk sliding friction and wear experiments were conducted on two different titanium alloys (Ti–6Al–4V and Ti–6Al–2Sn–4Zr–2Mo). Disks of these alloys were slid against fixed bearing balls composed of 440C stainless steel, silicon nitride, alumina, and polytetrafluoroethylene (PTFE) at two speeds: 0.3 and 1.0 m/s. The friction coefficient and wear rate were lower at the higher sliding speed. Ceramic sliders suffered unexpectedly higher wear than the steel slider. The wear rates, ranked from the highest to the lowest, were alumina, silicon nitride, and steel, respectively. This trend is inversely related to their hardness, but corresponds to their relative fracture toughness. Comparative tests on a Type 304 stainless steel disk supported the fracture toughness dependency. Energy dispersive spectroscopy (EDS) and X-ray diffraction (XRD) analyses confirmed the tendency of Ti alloys to transfer material to their counterfaces and suggested possible tribochemical reactions between the ceramic sliders and Ti alloy disks. These reaction products, which adhere to the ceramic sliders, may degrade the mechanical properties of the contact areas and result in high wear. The tribochemical reactions along with the fracture toughness dependency helped explain the high wear on the ceramic sliders. © 2004 Elsevier B.V. All rights reserved. Keywords: Titanium; Ceramics; Material transfer; Tribochemical reaction

1. Introduction In comparison to light weight alloys based on aluminum and magnesium, titanium alloys present interesting possibilities as tribomaterials, but they have not been widely investigated as bearing materials. They are harder and stiffer than Mg and Al alloys, and they resist exposure to heat and aqueous corrosion much better. Like Al and Mg, their high affinity 夽 Research sponsored by the U.S. Department of Energy, Assistant Secre-

tary for Energy Efficiency and Renewable Energy, Office of FreedomCAR and Vehicle Technologies, as part of the High Strength Weight Reduction Materials Program, under contract DE-AC05-00OR22725 with UT-Battelle, LLC. ∗ Corresponding author. Tel.: +1 865 574 4560; fax: +1 865 574 6918. E-mail address: [email protected] (J. Qu). 0043-1648/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2004.09.062

for oxygen results in the formation of an adherent surface oxide, but sub-stoichiometric TiO2 can act as a solid lubricant. A great deal is known about the physical metallurgy, heat treatment, and mechanical properties of titanium alloys, thanks to extensive aerospace-related research and development. Tribological concerns for Ti in aerospace components have focused mainly on their fretting behavior, leading to research on surface treatments like ion implantation and solid film lubrication [1,2]. Needs in the chemical process industry motivated a 1991 study of the galling and sliding wear behavior of commercial-purity Ti and alloy Ti–6Al–4V [3]. In that investigation, the best wear and friction results for Ti alloys were obtained for anodized counter-surfaces coated with MoS2 solid-film or with polytetrafluoroethylene (PTFE), but the abrasion resistance was poor. Relatively few additional

J. Qu et al. / Wear 258 (2005) 1348–1356

1349

Table 1 Compositions and characteristics of Ti–6Al–4V and Ti–6Al–2Sn–4Zr–2Mo Ti alloy

Compositions (wt%)* (balance Ti)

Microindentation hardness HV (GPa)

Tensile strength UTS (MPa)

Ti64 Ti6242

6.53 Al, 3.89 V, 0.13 Fe 5.85 Al, 1.98 Sn, 4.22 Zr, 1.95 Mo

3.36 ± 0.17 3.31 ± 0.10

954.9 957.0



Analyses supplied by Titanium Metals Corporation.

studies have been conducted on sliding wear mechanisms of Ti alloys. Molinari et al. highlighted the mechanisms responsible for the wear resistance under different load and sliding speed conditions in self-mated Ti–6Al–4V disk-on-disk sliding tests [4], and Dong and Bell [5] reported unexpectedly high wear rates for alumina sliding against Ti–6Al–4V (pinon-disk tests). Recent developments in Ti processing forecast the availability of lower-cost Ti and that has prompted further interest in exploring the tribological behavior of Ti alloys as bearing materials [6]. Focus by the U.S. Department of Energy on improved brake materials for fuel-efficient heavy trucks, has led to the consideration of Ti for disc brake rotors as well. In fact, coated Ti brake discs are already showing promise in auto racing [7]. This renewed interest in the friction and wear of Ti alloys has prompted the current laboratory study of the behavior of two commercially-available Ti alloys sliding against model metallic, ceramic, and polymeric counterfaces. One of the two alloys (Ti–6Al–4V) has had more tribological attention than the other, but the other (Ti–6Al–2Sn–4Zr–2Mo) has attractive elevated temperature properties and was felt to be of interest as well. There are very few studies on the tribological properties of Ti–6Al–2Sn–4Zr–2Mo in the literature. In this study, it is intended to establish baseline data for these alloys with which to compare the tribological behavior of new surface treatments or coatings in future work.

2. Materials and testing procedure Two titanium alloys, Ti–6Al–4V and Ti–6Al–2Sn– 4Zr–2Mo, were tested in this study. These alloys shall subsequently be referred to as Ti64 and Ti6242. The compositions

of these alloys, provided by the supplier for these heats of material, are listed in Table 1 and their microstructures are shown in Fig. 1(a) and (b), respectively. The typical ␣, ␣/␤-phase grain structure can readily be identified on the etched crosssection of Ti64. Ti6242 has finer grain size and is dominated by ␣-phase. The two alloys have similar microindentation hardness and tensile strength (see Table 1). Friction and wear tests were conducted using a pin-on-disk apparatus. The diameter of the fixed ball sliders was 9.53 mm. As shown in Table 2, 440C stainless steel, silicon nitride, alumina, and polytetrafluoroethylene (PTFE), were selected to represent metallic, ceramic, and polymeric bearing materials. The Ti alloy disks were 63.5 mm diameter and 12.7 mm thick. The disk surfaces were polished by 600 grit wet SiC abrasive paper and the pre-test arithmetic average surface roughness (Ra ), measured with a Taylor Hobson TalysurfTM 10 stylus profilometer with a 2.5 ␮m tip radius, was 0.11 ± 0.02 ␮m. Concentric wear tracks ranging from 18 to 52 mm in diameter were used, and the disk rotation rate was adjusted accordingly to provide either 0.3 or 1.0 m/s sliding speed. A 10 N normal load was applied and the test was run for 500 m sliding distance. In follow-on experiments, and to provide a comparison to the results for the Ti alloy disks, a 304 stainless steel disk (63.5 mm diameter and 6.35 mm thick) was tested against 440C stainless steel and ceramic (Si3 N4 and Al2 O3 ) sliders under similar sliding conditions. The microindentation Vickers hardness of the 304 stainless steel disk was about 3.16 GPa, slightly lower than that of the Ti alloy disks. The friction force was monitored by a load cell-based force measurement system. The wear volumes of the sliders and disks were determined by weight change measurements with an accuracy of 0.1 mg. The wear factor is defined as the wear volume normalized by the applied load and the sliding

Fig. 1. Microstructures of two Ti alloys.

1350

J. Qu et al. / Wear 258 (2005) 1348–1356

Table 2 Characteristics of slider materials Sliders

Supplier

Specification

Vickers hardness (GPa)

Fracture toughness (MPa m1/2 )

440C stainless steel Silicon nitride Alumina PTFE

McMaster-Carr Cerbec East Granby, CT Southern Bearing Service W.M. Berg, Inc. East Rockaway, NY

Grade 100 Hardened NBD200 Grade 5 AFBMA Grade 25 –

12.60a 19.37a 24.75a N/A

23.7b 5.2c 3–4c N/A

a

The Vickers hardness was measured using a 100 g load. The Izod impact strength of hardened 440C stainless steel is 4 ft lb [8]. The fracture toughness here was estimated based on Barsom–Rolfe’s empirical formula [9], with the assumption that the Izod impact strength is close to the Charpy V-notch impact strength. c The fracture toughness values were provided by the suppliers. b

distance of the pin. All the tests were conducted in ambient air conditions with temperature and humidity in the range of 18–22 ◦ C and 52–62%, respectively. At least two duplicates were run at each test condition. Good repeatability was obtained in both friction and wear results.

large fluctuation of the friction coefficient was thought to be caused by formation and periodic, localized fracture of a transfer layer. Titanium alloy commonly transfers to the counterface when rubbing against other metals or ceramics [3–5]. In this study, surface morphology examination and surface analysis confirmed this tendency. A transfer layer was easily identified on the wear scar of the 440C stainless steel ball (see Fig. 3(a)). The energy dispersive spectroscopy (EDS) analysis detected Ti and/or Al on the worn surfaces of the metal and ceramic balls, as shown in Fig. 4. More discussion on surface analysis can be found on Section 4.2. For metal and ceramic balls, lower friction coefficient and smaller instantaneous fluctuation were observed at 1.0 m/s compared to those at 0.3 m/s, as shown in Table 3. At higher sliding speed, the contact area had higher temperature, which generally reduced the shear strength and led to lower friction forces.

3. Results Results are summarized in Table 3(a) and (b) for Ti64 and Ti6242, respectively. Friction and wear results are presented separately. 3.1. Friction Table 3(a) and (b) present the average friction coefficient and its fluctuation at steady-state for each test condition. Selected friction traces of the four different sliders against Ti64 disks are shown in Fig. 2. The PTFE slider generated a fairly smooth friction trace (Fig. 2(d)) due to its self-lubricating nature. The metal and ceramic sliders produced friction coefficient in the range of 0.34–0.50 with relatively large fluctuation, as illustrated in Fig. 2(a)–(c). The

3.2. Wear As shown in Table 3(a) and (b), up to five times higher wear factors were obtained on both the slider and disk at 0.3 m/s

Table 3 Friction and wear results Slider material

Sliding speed 0.3 m/s Friction coefficient

1.0 m/s Wear factor

(mm3 /N m)

Ball

Disk

0.50 ± 0.05 0.47 ± 0.07 0.49 ± 0.07 0.28 ± 0.001

6.9 × 10−6 3.8 × 10−5 5.7 × 10−5 8.4 × 10−4

1.7 × 10−4 3.5 × 10−4 5.7 × 10−4 N/Ma

(b) Ti–6Al–2Sn–4Zr–2Mo disks 440C stainless steel 0.48 ± 0.05 Silicon nitride 0.47 ± 0.08 Alumina 0.49 ± 0.08 PTFE 0.27 ± 0.001

5.1 × 10−6 4.4 × 10-5 1.2 × 10−4 9.9 × 10−4

1.3 × 10−4 3.5 × 10−4 3.4 × 10−4 N/Ma

(a) Ti–6Al–4V disks 440C stainless steel Silicon Nitride Alumina PTFE

a

N/M, not measurable.

Friction coefficient

Wear factor (mm3 /N m) Ball

Disk

0.35 ± 0.05 0.36 ± 0.07 0.44 ± 0.07 0.29 ± 0.001

1.6 × 10−6 6.2 × 10−6 1.6 × 10−5 6.1 × 10−4

1.5 × 10−4 1.3 × 10−4 2.0 × 10−4 N/Ma

0.34 ± 0.04 0.37 ± 0.02 0.42 ± 0.04 0.29 ± 0.001

1.22 × 10−6 9.40 × 10-6 2.33 × 10−5 7.92 × 10−4

1.1 × 10−4 1.1 × 10−4 2.2 × 10−4 N/Ma

J. Qu et al. / Wear 258 (2005) 1348–1356

1351

Fig. 2. Frictional traces of different sliders against Ti64 disks.

than those at 1.0 m/s. Similar sliding speed dependency was also reported by other researchers [5]. The Ti disks suffered high wear rates, in the order of 10−4 mm3 /N m, against the metal and ceramic balls, and harder sliders generated relatively more (or at least comparable) wear on the Ti disks. Although harder balls were expected to have higher wear resistance, the results in Table 3(a) and (b) show a reverse order: the alumina ball wore more than the silicon nitride ball, which in turn wore more than the stainless steel ball. Remarkably, the wear factors of the ceramic balls were at least five times higher than those of the steel balls. Dong and Bell also reported a higher wear rate of an alumina ball than that of a steel ball when sliding against a Ti64 disk [5]. More analysis and discussion are presented in Section 4. Fig. 3 shows the wear scars on the metal and ceramic balls with features of abrasive wear, adhesive wear, and plastic deformation. Abrasive wear seemed to dominate the wear process at 0.3 m/s. Those wear scars were larger and flatter, corresponding to their higher wear factors. The wear scars generated at 1.0 m/s were smaller but much rougher with larger patches of transferred material implying more severe adhesive wear, possibly due to higher temperature at the contact area. EDS analysis also showed higher Ti and/or Al concentration on the wear scars at 1.0 m/s. The PTFE slider had the highest wear factor (10−3 mm3 /N m). Its counterface (Ti disk) was pro-

tected by the polymeric layer transferred from the PTFE ball and had almost no surface damage except a few shallow circular groves ground by third body particles, probably some metal debris, embedded in the PTFE ball. It has been seen that these two Ti alloys showed similar friction and wear behavior. Therefore, discussion will be focused on Ti64 only.

4. Discussion The most unusual finding of this study was the observation that the relatively hard ceramic sliders wore considerably more severely than the softer stainless steel slider. Mechanical and chemical analyses have been conducted to try to explain these results. 4.1. Fracture toughness The wear resistance of 440C stainless steel, silicon carbide, and alumina pins is in the reverse order as their relative Vickers hardness numbers, but in the same relative order as their fracture toughness. Recognizing that the current work was on sliding wear, the abrasive wear of ceramics has been proposed to be a function of both hardness and fracture toughness [10]. Further studies

1352

J. Qu et al. / Wear 258 (2005) 1348–1356

Fig. 3. SEM images of wear scars on the balls sliding against Ti64 disks.

on ceramic wear mechanisms showed that fracture toughness may play a dominating role in wear resistance. For example, Fischer [11] has demonstrated that, in the case of yttria stabilized zirconia ceramics, the wear resistance increases with the fourth power of fracture toughness. The pin-on-disk apparatus is mainly intended to evaluate sliding wear, but in practice, the load history can consist of

sliding and impact, since vibrations may occur if the disk surface is not perfectly normal to the axis of rotation. The tendency for impact to occur for small errors in alignment depends also on the sliding speed and how far the contact is from the center of rotation. Unlike sliding that usually causes plastic shearing in materials, impact may introduce catastrophic failures, such as cracking and crushing of the

J. Qu et al. / Wear 258 (2005) 1348–1356

1353

Table 4 Friction and wear results for 304 stainless steel disks against metal and ceramic sliders Slider material

Friction coefficient

Wear factor (mm3 /N m) Ball

Disk

440C stainless steel Silicon nitride Alumina

0.55 ± 0.05 0.68 ± 0.08 0.57 ± 0.03

<1 × 10−6 1.15 × 10−6 9.80 × 10−6

2.52 × 10−4 2.01 × 10−4 5.41 × 10−4

disk had much lower wear factors than those against the Ti disks at the same testing condition (see Tables 3 and 4), while the 304 stainless steel and Ti64 disks had similar hardness and comparable wear factors. This suggests that there might be other sources influencing the wear rate. 4.2. Tribochemical reactions

Fig. 4. EDS analysis of the wear scars on the sliders.

contact surfaces, leading to faster material removal and the production of sharp ceramic debris fragments that can in turn cause three-body abrasion. Brittle materials, like ceramics, are more sensitive to such repeated impact effects than are tougher metals. Rice et al. [12,13] have studied the wear rate and mechanism of compound impact (impact and sliding) on metals and superalloys. The material with lower fracture toughness had a higher wear rate and gave strong evidence for subsurface damage. To test the dependency of the wear rate on fracture toughness, comparative tests were conducted on a 304 stainless steel disk sliding against the 440C stainless steel, silicon nitride, and alumina sliders, under 10 N load and at 0.3 m/s speed for 500 m. Friction and wear results for the 304 stainless steel disk are shown in Table 4. The alumina, silicon nitride, and 440C stainless steel balls had the wear rate from high to low. This confirmed the fracture toughness effect. However, it has been noticed that the sliders against the steel

It is known that mechanically deformed surfaces usually have different chemical reactivity than purely thermally stressed solids [14]. Tribochemical reactions may significantly accelerate the wear process. Surface analyses (EDS and XRD) were conducted on the contact surfaces to explore the possibility of tribochemical reactions that may induce the unexpected high wear rates on the ceramic sliders. Fig. 4 shows the EDS spectra of the wear scars on the balls that slid against the Ti64 disks. The EDS analyses indicated Ti and Al on the worn surfaces of the steel and alumina balls (see Fig. 4(a) and (c)), which indicate material transfer from the Ti64 disk to the sliders. It is interesting to notice that only Al but no Ti was found on the wear scar of the silicon nitride ball (see Fig. 4(b)). No Al was observed on the unworn region of this ball. This may imply that the detected Al was probably not in metallic form (otherwise Ti should be present too), but had chemical compounds with other elements, such as Si, O, and/or N present in Fig. 4(b). X-ray diffraction was then used to further analyze the wear scars on the ceramic sliders. A four-axis goniometer [15] was employed for the grazing incidence (2◦ ) X-ray diffraction measurements using Cu K␣ radiation and parallel beam optics. That technique eliminates the sample surface displacement errors due to the spherical shape. Fig. 5 shows the XRD patterns of the wear scar and debris generated by silicon nitride against Ti64. Fig. 5(a) reveals that the wear scar on the silicon nitride ball contains silicon nitride (Si3 N4 ) in both ␣- and ␤-phase and silicon oxide nitride (Si2 N2 O). Due to the peak superposition, silicon aluminum oxide nitride (Si5 AlON7 ) cannot be distinguished from Si3 N4 . However, the presence of Al detected by EDS supports this possibility. As shown in Fig. 5(b), the XRD analysis on the wear debris found titanium, silicon nitride, and titanium nitride, but no indication of titanium oxides. This was a little surprising because titanium oxides have the lower Gibbs free energy of formation than the titanium nitride in the ambient environment. One possible explanation is that the titanium oxides were amorphous due to severe plastic deformation and were

1354

J. Qu et al. / Wear 258 (2005) 1348–1356

Fig. 5. X-ray diffraction analysis of the wear scar and debris for silicon nitride sliding against Ti64.

not detected by XRD. The XRD pattern of the worn surface on the alumina ball sliding against a Ti64 disk is shown in Fig. 6. The observed spinel (MgAl2 O4 ) was probably a sintering aid. The XRD pattern may suggest some possibility of forming titanium–aluminum intermetallic compounds (Al3 Ti, Al2 Ti), but there was no strong evidence. Titanium aluminides were also suspected by Dong and Bell [5] based

on the XRD analysis of the wear debris produced by alumina sliding against Ti64. The high wear rates of alumina and silicon nitride sliders may be attributed to the formation of chemical reaction products between them and the Ti and/or Al transferred from the Ti64 disks. Such tribochemical reactions were also aided by the lower thermal conductivity of Ti that promotes

J. Qu et al. / Wear 258 (2005) 1348–1356

1355

Fig. 6. X-ray diffraction pattern of the wear scar on the alumina slider against the Ti64 disk.

a higher temperature near the interface. These reaction products bonded to the ceramic contact surfaces may deteriorate their mechanical properties and result in micro fractures leading to high wear. The wear process continuously developed “fresh surfaces” and in turn accelerated the tribochemical reactions.

5. Summary The tribological behavior and responsible wear mechanisms for titanium alloys Ti–6Al–4V and Ti–6Al–2Sn– 4Zr–2Mo, sliding against 440C stainless steel, silicon nitride, alumina, and PTFE were investigated. The following observations and conclusions were obtained: (1) The two Ti alloys had similar friction and wear performance, although their grain structures and compositions are different. (2) Large frictional fluctuations occurred when metal and ceramic balls slid against Ti alloy disks, probably caused by formation and periodic, localized fracture of a transfer layer. (3) Higher friction coefficient with larger fluctuation and higher wear rate were observed at the lower sliding speed. (4) Despite their higher hardness, ceramic sliders experienced much higher wear and created more wear on the counterfaces than did the stainless steel sliders. (5) Fracture toughness and tribochemical reactions have been proposed to explain the unexpected high wear rates on the ceramic sliders. Comparative tests on a 304 stain-

less steel disk supported the fracture toughness dependency of the wear rate. (6) EDS and XRD analyses confirmed material transfer from the Ti alloy disks to their counterfaces and suggested possible tribochemical reactions.

Acknowledgements The authors with to acknowledge with appreciation Y. Kosaka of Titanium Metals Corporation, USA, for supplying alloy billets along with their chemical analyses. Support for this research was provided by the U.S. Department of Energy, Assistant Secretary for Energy Efficiency and Renewable Energy, Office of FreedomCAR and Vehicle Technologies, as part of the High Strength Weight Reduction Materials Program, under contract DE-AC05-00OR22725 with UT-Battelle, LLC. J. Qu and N. Kulkarni were supported in part by appointments to the ORNL Postdoctoral Research Associates Program administered jointly by ORNL and ORISE.

References [1] F.M. Kustas, M.S. Misra, Friction and wear of titanium alloys, in: P.J. Blau (Ed.), ASM Handbook, Friction, Lubrication, and Wear Technology, 18, ASM International, 1992, pp. 778–784. [2] R.B. Waterhouse, A. Iwabuchi, The effect of ion implantation on the fretting wear of four titanium alloys at temperatures up to 600 ◦ C, in: Proceedings of the International Conference on Wear of Materials, ASME, New York, 1985, pp. 471–484.

1356

J. Qu et al. / Wear 258 (2005) 1348–1356

[3] K.G. Budinski, Tribological properties of titanium alloys, Wear 151 (1991) 203–217. [4] A. Molinari, T.B. Straffelini, T. Bacci, Dry sliding wear mechanisms of the Ti6Al4V alloy, Wear 208 (1997) 105–112. [5] H. Dong, T. Bell, Tribological behavior of alumina sliding against Ti6Al4V in unlubricated contact, Wear 225–229 (1999) 874–884. [6] EHKT Technologies, Opportunities for low cost titanium in reduced fuel consumption, improved emissions, and enhanced durability heavy-duty vehicles, Oak Ridge National Laboratory Report, ORNL/Sub/4000013062/1, Oak Ridge, Tennessee, 2002, p. 59. [7] Ultra-Lite Brakes and Components, Literature, Red Devil Brakes, Inc., Mt. Pleasant, Pennsylvania, not dated. [8] S. Lampman, Fatigue and fracture properties of stainless steel, in: S.R. Lampman (Ed.), ASM Handbook, Fatigue and Fracture, 19, ASM International, 1996, pp. 712–732. [9] J.M. Barsom, S.T. Rolfe, Correlations between KIC and Charpy Vnotch test results in the transition temperature range, in: Impact

[10] [11] [12] [13]

[14] [15]

Testing of Materials STP 466, ASTM, Philadelphia, 1979, pp. 281– 302. A.G. Evans, D.B. Marshall, Fundamentals of Friction and Wear of Materials, ASM, 1980, p. 439. T.E. Fischer, Friction and wear of ceramics, Scripta Metall. Mater. 24 (1990) 833–838. S.L. Rice, The role of microstructure in the impact wear of two aluminum alloys, ASME Proc. Wear Mater. (1979) 27–34. S.L. Rice, H. Nowotny, S.F. Wayne, Characteristics of metallic subsurface zones in sliding and impact wear, ASME Proc. Wear Mater. (1981) 47–52. G. Heinicke, Tribochemistry, Carl Hanser Verlag Munchen Wien, Berlin, 1984. H. Krause, A. Haase, X-Ray Diffraction System PTS for Powder, Texture and Stress Analysis, in: H.J. Bunge (Ed.), Experimental Techniques of Texture Analysis, vol. 405–408, DGM Informationsgesellschaft, Verlag, 1986.