Physics of the Solid State, Vol. 43, No. 4, 2001, pp. 665–669. Translated from Fizika Tverdogo Tela, Vol. 43, No. 4, 2001, pp. 639–643. Original Russian Text Copyright © 2001 by Ivanov, Khazanov, Taranov, Mikhaœlova, Gropyanov, Abramovich.

DEFECTS, DISLOCATIONS, AND PHYSICS OF STRENGTH

Grain Boundaries and Elastic Properties of Aluminum-Oxide and Stainless-Steel-Based Cermets S. N. Ivanov*, E. N. Khazanov*, A. V. Taranov*, I. S. Mikhaœlova**, V. M. Gropyanov**, and A. A. Abramovich** *Institute of Radio Engineering and Electronics, Russian Academy of Sciences, Mokhovaya ul. 11, Moscow, 103907 Russia **St. Petersburg Technological University of Plant Polymers, ul. Ivana Chernykh 4, St. Petersburg, 198095 Russia e-mail: [email protected] e-mail: [email protected] Received July 25, 2000; in final form, September 11, 2000

Abstract—Elastic properties and thermal-phonon scattering are investigated in Al2O3 + 0.3% MgO ceramics and cermets of different porosities based on them. The cermets, reinforced with a metallic frame of the steel 12X18H9T, are obtained by dry compaction followed by sintering. It is shown that the elastic moduli of cermets are determined by their porosity and that the grain boundaries can be investigated in detail by a nonequilibriumphonon propagation method. © 2001 MAIK “Nauka/ Interperiodica”.

INTRODUCTION New composites based on ceramics and metals (cermets) have been being developed since the 1960s, and unique materials have been produced which combine the virtues of both components, such as heat resistance, temperature stability, wear resistance, service reliability, and chemical stability [1]. At the present time, cermets are already in service in the machine-building industry (high-temperature elements of gas turbines, rolling bearings, high-precision temperature-stable tools), the medical and food industries (parts of pumps and devices for pumping over corrosive liquids), and other fields [2]. However, the development of cermets of new types and the prediction and investigation of their properties involve difficulties. Among these are the sophisticated high-temperature synthesis technology (often requiring a vacuum and high pressures), not entirely known mechanisms for the grain formation and for the grain-boundary influence on the strength and other properties of cermets, and the lack of methods providing reliable information about grain boundaries. In this work, we investigate cermets based on Al2O3 and on the stainless steel 12X18H9T with the aim of (1) developing a comparatively simple and cheap technology for the fabrication of the cermets and (2) obtaining samples which have a high strength in combination with a low density. We synthesized a material which is characterized by large elastic moduli and high wear resistance and thermal stability and, therefore, has an application potential in machine building and other fields of engineering. It is known that the mechanical properties of cermets depend heavily on their porosity and the quality of grain boundaries [3], which are determined by the syn-

thesis technique of the original material, as well as by the fabrication technology of the samples and their finishing thermal treatment. To investigate the influence of these factors, we apply two different dynamic elasticwave methods in this work. At room temperature, lowfrequency ultrasound waves are used, while in the liquid-helium temperature range, we apply a nonequilibrium acoustic-phonon propagation method. The former method gives information about the elastic and, hence, strength properties of the materials. The data obtained by the latter method allowed us to construct a model for grain boundaries in the ceramic Al2O3 and cermets based on it. 1. EXPERIMENTAL TECHNIQUE We investigated samples of the original (“basic”) ceramic Al2O3, which contained a stabilizing addition MgO (0.3%), and cermets based on this ceramic in combination with 20% of the commercial-quality stainless steel 12X18H9T (18% Cr, 9% Ni, 1% Ti, 72% Fe). The basic ceramic samples were prepared by dry compaction of the mixture under a pressure of 80– 100 MPa followed by sintering in a vacuum at 1940°C. These samples and their characteristics will be called basic in what follows, because their preparation technology with adding small amounts of MgO is typical of the synthesis of Al2O3 ceramics. The samples thus obtained had a volume porosity lower than 1%, with the average grain diameter being 10–3 cm. To fabricate a cermet, the initial fine-grained mixture was prepared by milling Al2O3 powder in a ball mill in the presence of balls 1–2 cm in diameter made from the stainless steel 12X18H9T. The milling was

1063-7834/01/4304-0665$21.00 © 2001 MAIK “Nauka/Interperiodica”

IVANOV et al.

666

Measured parameters of the Al2O3 ceramic and cermets based on it Sample

Composition

Basic 1 2

Al2O3 + 0.3% MgO Al2O3 + 20% stainless steel Al2O3 + 0.5% Cr2O3 + 1% TiO2 + 0.5% MnO2 + 20% stainless steel Al2O3 + 20% stainless steel

3

Sintering Elastic moduli, GPa tempe- Porosity, % G E K rature, K 1920 1920 1640

≤ 1.0 9.1 19.0

1690

36.8

terminated when the steel content in the Al2O3 powder became equal to 20%. Then, the mixture obtained was doped with a plasticizer and subjected to dry compaction under a pressure of 80–100 MPa and then sintered in a vacuum at 1640–1920°C. Finally, the samples were cooled in the furnace at an average rate of 100°C/h and no further treatment was made. The cermets thus prepared (samples 1–3 in table) ranged in volume porosity from 9 to 37%, but had the same steel content (the composition of the basic ceramic in sample 2 differed insignificantly from that in the other samples). The cermet samples were then cut into smaller parts of the required dimensions, which were further ground and polished, depending on the measuring method. To investigate the elastic properties of the cermets, we measured their density by the hydrostatic method and the velocities of longitudinal and transverse ultrasound waves of a frequency of 1.7 MHz. With the data obtained, we calculated Young’s modulus E, the shear modulus G, the bulk modulus K, and Poisson’s ratio from the well-known formulas of the elasticity theory for an isotropic medium [4]. We used the pulsed phaseinterferometry method, which made it possible to determine the velocity of ultrasound within an accuracy of 1–2%, the elastic moduli within 5%, and Poisson’s ratio within 10–20%. The strength of samples was measured with an IP6011-500-1 hydraulic press. The ultimate compression strength was evaluated from the formula σ = 4P/πd 2, where P is the breaking load and d is the diameter of the sample; for each type of composite, the compression test was performed at a loading rate of 0.04– 0.05 kg/s for ten samples and the results were averaged. Examination of the surface structure (cleaved facet) of the Al2O3 ceramic samples was made with a JSM840 scanning electron microscope (the Jeol company) and a P4-SPM-MTD scanning probe microscope (atomic-force microscopy regime) operating in the contact mode. The kinetics of phonons at liquid-helium temperatures was investigated by the “heat pulse” method [5]. For this purpose, a gold film was sputtered on one face of a plate of the material under study; this film was heated by a short current pulse (≅10–7 s) and served as an injector of nonequilibrium phonons into the sample.

Strength Poisson’s Deff , cm2/s σ, MPa ratio, ν (T = 3.8 K)

144.0 358.0 234.5 900–1000 71.4 170.3 238.7 560 58.5 153.1 151.1 425 35.1

83.0

48.9

140

0.24 0.35 0.33

1.2 × 102 1.2 0.82

0.21

0.74

A meander-shaped bolometer with the dimensions 0.3 × 0.25 mm2 was deposited on the opposite plate face. A weak bias magnetic field (~2 × 102 Oe) was applied to the bolometer in order to measure the temperature dependence of the scattered intensity of nonequilibrium phonons in the sample over the temperature range 1.7–3.8 K. The power dissipated in the heater was so low that the injected phonons could be characterized by a temperature equal to that of the thermostat (bath), and their frequency distribution was close to the Planck distribution. 2. EXPERIMENTAL RESULTS AND DISCUSSION Micrographs of the surfaces of the Al2O3 ceramic and cermet samples obtained by scanning electron microscopy are presented in Fig. 1. It is seen that, for the most part, grains in the basic ceramic have isomeric hexagonal faces (Fig. 1a) and the grain faces meet at an angle close to 120° at interface junctions (see also [6]); the average grain diameter is about 10–3 cm. In the cermet samples (Fig. 1b), metal grains having a nearspherical shape are clearly visible against the background of faceted grains of the polycrystalline Al2O3. Examination of a fairly large area of cleaved facets of cermet samples revealed that the metal grains are distributed uniformly and do not form clusters or filamentlike structures. This is supported by resistance measurements, according to which the cermets remained insulators and had no “junctions.” A more detailed structure of the metal grains as obtained by atomicforce microscopy is presented in Fig. 1c; this allows one to estimate their minimum size with a reasonable accuracy. The dimensions of metal grains are strongly scattered and lie in a range of Rm ≈ 200–1000 nm (Figs. 1b and 1c). The practically important room-temperature strength characteristics of the Al2O3 composites, their initial compositions, and the sintering temperatures are listed in the table. Note that the mechanical characteristics (elastic moduli and strength) of the cermets differ from their respective values for the basic ceramic and vary rather smoothly with increasing overall volume porosity. The measured elastic moduli correlate well with the data presented in [7], and their dynamics is PHYSICS OF THE SOLID STATE

Vol. 43

No. 4

2001

GRAIN BOUNDARIES AND ELASTIC PROPERTIES OF ALUMINUM-OXIDE

controlled only by the porosity of the material; that is, low-frequency measurements and strength data carry no information about the properties of grain boundaries in the Al2O3 and metal–α-Al2O3 interfaces. The last column of the table lists the values of the effective phonon diffusion coefficient Deff at a liquidhelium temperature. These values of Deff are first invoked to characterize the ceramic materials; they allow one to infer the character of phonon scattering at grain boundaries in ceramics and to predict the mechanical characteristics of the material at room temperature. Let us discuss the data on the propagation of weakly nonequilibrium phonons in the basic Al2O3 ceramic and cermet samples. The curves in Fig. 2 describe the propagation of a heat pulse in the basic Al2O3 sample at different temperatures. These bell-shaped bolometer signal curves are typical of diffusive pulse propagation and have a well-defined maximum. For phonons traveling the thickness L of the ceramic sample, the signal maximum will be observed at a time tmax, which characterizes the properties of grain boundaries and is given by the expression [8] 2

667

1 µm

(‡)

Al2O3

Steel

1 µm

(b)

2

L L S t max ≅ -------- ≅ -------------------- . D eff v s RΣ f ω

(1) (c)

Here, vs is the velocity of sound in a ceramic grain, R is the average dimension of grains, S is the area of the grain surface, Σ is the total contacting area per grain, and fω is the probability that a phonon passes through the contact area. The temperature dependence of tmax is determined by the temperature dependence of fω, which, in the model used [8], is a function of the wavelength (energy) of the phonon (i.e., of the temperature of the sample in our experiments) and determined by the mechanism of phonon scattering at the grain boundaries, which depends on the ceramic sintering process, the composition of the initial powder, and the other specific features of the ceramic fabrication. For the basic Al2O3 samples, the temperature dependence of tmax is closely approximated by the expression tmax = A + BT 4 (see inset in Fig. 2), where the first term is determined by the acoustic matching of the contacting ceramic grains for nonequilibrium phonons and the second is due to phonon scattering at the interface of the grains. In the range of the plateau, one can estimate the coefficient fω by putting S/Σ ≅ 1, vs ≅ 7 × 105 cm/s, and the grain diameter to be ≅10–3 cm for the dense basic ceramic with the result that fω is equal to 0.5–0.8, which is evidence that the grains matched fairly well in this material. In the most likely case of phonon scattering at grain boundaries, the temperature dependence tmax ~ T 4 will take place if the condition qlb ! 1 is met [9], where q is the phonon wave vector and lb is the grain boundary thickness. In our experiments, q = (1–2) × 106 cm–1; therePHYSICS OF THE SOLID STATE

Vol. 43

No. 4

2001

nm

1000

500

0

Steel

500

1000

nm

Fig. 1. Electron micrographs of cleaved facets of (a) ceramic Al2O3–0.3 wt % MgO (basic sample in table) and (b, c) cermet Al2O3 + 20 wt % stainless steel (sample 1 in table) differing in image scale.

fore, lb is estimated to be lb ≅ 0.5–1.0 nm, which supports the conclusion that grain boundaries are perfect in the dense α-Al2O3 ceramic. In cermet samples, we have a completely different situation (Fig. 3). The time tmax is fully two orders of magnitude longer; that is, the effective phonon diffusion coefficient is smaller (see table). However, the most important result is that the temperature dependence of tmax is radically altered; namely, we have ∂tmax/∂T < 0 in cermet samples. The slower phonon diffusion and negative ∂tmax/∂T in cermets cannot be due to metal grains, whose acoustic characteristics are different from those of α-Al2O3. Estimations of the possi-

IVANOV et al.

668

tmax, µs

5

S, arb. units

4

10 1

10

T, K

3

2 1

0

50 t, µs

100

Fig. 2. Time dependence of the signal amplitude S of nonequilibrium phonons in the basic ceramic Al2O3 for L = 0.5 mm and different temperatures T (K): (1) 3.83, (2) 3.44, (3) 2.18, (4) 2.57, and (5) 2.28. The inset shows the temperature dependence of tmax, with the solid line being tmax = A + BT 4.

1 2

S, arb. units

3 4

5 6

0

1000

2000

3000

4000

t, µs

Fig. 3. Time dependence of the signal amplitude S of nonequilibrium phonons in a cermet sample with a porosity of 19% and L = 0.3 mm for different temperatures T (K): (1) 3.8, (2) 3.48, (3) 3.27, (4) 3.01, (5) 2.79, and (6) 2.61. PHYSICS OF THE SOLID STATE

Vol. 43

No. 4

2001

GRAIN BOUNDARIES AND ELASTIC PROPERTIES OF ALUMINUM-OXIDE

ble contribution from electron–phonon interaction to phonon scattering in the framework of classical papers [10] show that phonon scattering in the grains of a diameter less than 10–4 cm is insignificant and the direct contribution to the damping from the phonon– phonon interaction is also very small; therefore, it is most likely that phonon diffusion becomes slower because of phonon scattering by the metal–α-Al2O3 interfaces. According to [9], this can be true if there are many fine voids in these interfaces (for instance, because the metal–insulator wetting is poor [11, 12] and their contacts do not become close in the process of sintering). In this case, we have open interfaces with a noticeably lower density and elasticity. At lb ≥ 1 nm, such interfaces are significantly less transparent for phonons and can be the reason for the temperature dependence with ∂tmax/∂T < 0. Estimations give reasonable values lb ≥ 5.0–10.0 nm for this case. For the cermet samples under study, the variation in Deff correlates with the dynamics of the overall porosity P, which suggests that the porosity in the interfaces is proportional to P. In closing, we note that the data obtained by the method of nonequilibrium phonon propagation (heat pulse method), proposed in this paper for investigating ceramic samples, allow one to construct a model of grain boundaries in the ceramic Al2O3 and cermets based on it. Analysis of the influence of grain boundaries on the mechanical properties of cermets at room temperature will be the objective of further investigation. ACKNOWLEDGMENTS The authors are grateful to O.V. Karban’ for the examination of cleaved facets of ceramics by atomicforce microscopy.

PHYSICS OF THE SOLID STATE

Vol. 43

No. 4

2001

669

This work was supported by the Russian Foundation for Basic Research, project no. 00-02-17426. REFERENCES 1. S. M. Lang, NBS Monogr., No. 6, March 1, 45 (1960). 2. P. S. Kislykh, N. I. Bodnaruk, and M. S. Borovikova, Cermets (Naukova Dumka, Kiev, 1985). 3. W.-P. Tai and I. Watanache, J. Am. Ceram. Soc. 81 (6), 1673 (1998). 4. L. Bergmann, Der Ultraschall und seine Anwendung in Wissenschaft und Technik (S. Hirzel, Zürich, 1954; Inostrannaya Literatura, Moscow, 1957). 5. S. N. Ivanov, A. V. Taranov, and E. N. Khazanov, Zh. Éksp. Teor. Fiz. 99 (4), 1311 (1991) [Sov. Phys. JETP 72, 731 (1991)]. 6. S. N. Ivanov, E. N. Khazanov, and A. V. Taranov, Fiz. Tverd. Tela (St. Petersburg) 37 (10), 2902 (1995) [Phys. Solid State 37, 1601 (1995)]. 7. Physical Acoustics: Principles and Methods, Vol. 3, Part B: Lattice Dynamics, Ed. by W. P. Mason (Academic, New York, 1965; Mir, Moscow, 1968). 8. S. N. Ivanov, A. G. Kozorezov, A. V. Taranov, and E. N. Khazanov, Zh. Éksp. Teor. Fiz. 102 (2), 600 (1992) [Sov. Phys. JETP 75, 319 (1992)]. 9. L. M. Brekhovskikh and O. A. Godin, Acoustics of Layered Media (Nauka, Moscow, 1989; Springer-Verlag, New York, 1990). 10. Physical Acoustics: Principles and Methods, Ed. by W. P. Mason (Academic, New York, 1968–1970; Mir, Moscow, 1974), Vol. 7, Chap. 3. 11. V. N. Eremenko, Yu. V. Naœdich, and A. A. Nosovich, Élektronika, No. 5, 136 (1959). 12. Cermets, Ed. by J. R. Tinklepaugh and W. B. Crandall (Reinhold, New York, 1960; Inostrannaya Literatura, Moscow, 1962).

Translated by Yu. Epifanov