4. 5. 6. 7. 8.

S. F. Kondakov and O. A. Sarkisyan, "Effect of temperature on the perforation resistance of metallic objects," Probl. Prochn., No. 9, 69-71 (1980). M. L. Wilkins, Penetration Mechanics, UCRL-72111, University of California. Berkley Radiation Laboratory (1969), p. 40. J. Awerbuch and S. R. Bodner, "Experimental investigation of normal perforation of projectiles in metallic plates," Int. J. Solids Struct., iO, No. 6, 685-699 (1974). S. T. Mileiko, S. F. Kondakov, and E. G. Golofast, "On one case of perforation," Probl. Prochn., No. 12, 69-71 (1979). M. L. Wilkins, "Mechanics of penetration and perforation," Int. J. Eng. Sci., 16, No.

11, 793-807 (1978). 9. i0.

ii.

12.

J. Gering, "High-Speed Impact from the Engineering Standpoint," High-Speed Impact Phenomena [Russian translation], V. N. Nikolaevskii (ed.), Mir, Moscow (1973), pp. 468-517. S. T. Mileiko, S. F. Kondakov, and O. A. Sarkisyan, "On the perforation of metallic objects by a spherical projectile," Continuum Mechanics: Materials Presented at the AllUnion Conference on Continuum Mechanics, May, 1979, Tashkent (1982), pp. 42-46. G. Dorey and G. R. Sidey, "Residual strength of CFRP laminates after ballistic impact," Mechanical Properties High Rates Strain. Proceedings of the Conference, Oxford et al., (1974), pp. 344-351. V. P. Muzychenko and V. I. Postnov, Impact Mechanics of Systems Deformed through Heredity (Phenomenological Mechanics of Composite Perforation) [in Russian], Ya. Fabritsius D W A I U , Daugavpils (1985), p. 206.

EFFECT OF OXIDATION ON THE STRENGTH AND THERMAL STABILITY OF A MATERIAL BASED ON SILICON NITRIDE V. S. A. A.

A. I. F. B.

Lavrenko, A. A. Chernovolenko, Sopenko, V. I. Zubov, Alekseev, Yu. G. Gogotsi, Goncharuk, and V. V. Shvaiko

UDC 666.76.01

A structural ceramic based on silicon nitride and carbide is promising for the fabrication of various high-temperature mechanical devices [i, 2]. One of the most important factors determining the suitability of materials for use at high temperatures is their stability toward oxidation under operating conditions. This is because oxidation affects the mechanical properties of materials substantially. The general characteristics of the effect of oxidation on the strength properties of the given materials have not been demonstrated hitherto. For example, according to data in the literature the reaction with oxygen has a different effect on the strength of hot-pressed [3, 4] and reaction-bound [5, 6] materials based on Si~N~. There is almost no information on the corrosion resistance of materials in the Si~N~-SiC system at high temperatures. In the present work, which is a continuation of the investigation [7], where the material NKKKM-79 was considered (for materials of this type see [8]), we have studied the effect of high-temperature oxidation on the strength and thermal stability of the material NKKKM-83a. The specimens of N - ~ - 8 3 a were of a reaction-sintered material of composition Si3N4-SiC (70:30) in which 2 mass % of magnesium oxide was introduced as an activating additive. According to data from chemical analysis the nitrogen content in the specimens was 27.0-27.6 mass %. Spectral analysis showed that impurities of iron, boron, aluminum, calcium, titanium, and copper are present in the material. The specimens for testing were cut with a diamond tool from small bars having the dimensions 5.9 x 5.9 x 54 mm. The effect of oxidation on strength was studied by a procedure described previously [7]. The tensile strength was determined under conditions of bending by a concentrated force on an MIK-9 unit on specimens of height 2.3 mm, width 5 mm, and length 25 mm. The specimens were Kiev Polytechnic Institute. Institute of Strength Problems, Academy of Sciences of the Ukrainian SSR. Translated from Problemy Prochnosti, No. 8, pp. 67-70, August, 1986. Original article submitted August 14, 1984.

1070

0039-2316/86/1808-1070512.50

~ 1987 Plenum Publishing Corporation

20

800

~000

~200

f400

T,~

Fig. i. Dependence of the tensile strength of NKKKM-83a on the preoxidation temperature for Ttest = 20~ (i) and Ttest = 1200~ (2): a) treated surface; b) surface in the as-supplied state. cut off such that the surface occurring go mechanical treatment (Fig. i).

in the test under elongation

conditions did not under-

A boat containing the specimens was heated at a rate of -5~ in a tube furnace to a given temperature and maintained there for 3 h, after which the specimens were cooled rapidly (quenched) in air to room temperature. The dependence of the tensile strength of the material on the oxidation temperature over range 800-1500~ is presented in Fig. I. According to the data obtained the oxidation of the specimens at temperatures up to 1200~ leads to a certain reduction in the tensile strength measured at room temperature. In addition, oxidation over the temperature range indicated has virtually no effect on the strength of the material at 1200~ An increase in oxidation temperature to 1300~ causes an increase in the tensile strength, both at room temperature and at high temperature. Preliminary oxidation at 1400~ and above leads to a decrease in the tensile strength. It should be noted that the effect of oxidation on the strength of NKKKM-83a is expressed to a lesser extent than on the strength of NKKKM-79 [7]. For specimens preoxidized for 3 h at 1350~ we studied the temperature dependence the tensile strength. To eliminate the effect of additional oxidation during testing investigations were conducted with an argon atmosphere on a TsD-4 unit [9]. The tests conducted by the method of bending with a concentrated force on specimens 2.3 mm high, wide, and 54 mm long.

of the were 5.8 mm

The relationship obtained is presented in Fig. 2. As can be seen, the material NKKKM83a has its greatest strength at 1000~ the effect of preoxidation at this temperature appearing to be very appreciable. At higher temperatures there is a decrease in the tensile strength of the material. To examine the reasons for the appreciable difference in the behavior of the material being studied and NKKKM-79 and also explain the relationships obtained we studied the oxidation kinetics of these materials (Fig. 3). It was established that NKKKM-83a has a higher corrosion resistance than NKKKM-79, particularly over the range 800-II00~ This arises from the smaller average pore radius, as a result of which they are rapidly filled and the internal oxidation is lowered. It is also necessary to take into account that, according to spectral analysis data, the impurity content of NKKKM-83a is lower than in NKKKM-79. As is well known, a decrease in the impurity content of a material leads to an enhanced resistance to oxidation [i0]. At high temperatures the oxidation of these materials is characterized by approximately the same increase in mass with time. For all the NKKKM type of materials a large retardation of the oxidation process after a 30-mm exposure is characteristic in the temperature region above IIO0~ This can be explained by the formation of a liquid silicate phase on the surface of the specimens which seals all the open pores and prevents their further oxidation [2]. On specimens of NKKKM-79 heated at temperatures above II00~ a smooth glaze type transparent film is plainly seen. However, on NKKKM-83a, in which the impurity content is low, this film is appreciably thinner and is observed visually only on polished specimens oxidized at temperatures not lower than 1300~ X-ray phase analysis of the specimens, conducted with a DRON-2.0 diffractometer in copper radiation, showed the presence of a-cristobalite in the oxidized layer (Fig. 2b). A compari1071

a, ~ a

,

250

200

~

~

II

o --

I

9

[

135o~

l~

' fO0 Fig.

2.

JI q'O

301

20b[

-~

10 ie,deg

\i .Li,

20 400 500 a 800 ~000 ;200 1~.00T,oC Dependence of the t e n s i l e S t r e n g t h - o f NYdLH-8-3a on the t e s t

temperature in argon (a) (i represents the average values for specimens preoxidized at 1350~ for 3 h; 2 is the average value for the initial specimens) and typical x-ray diffraction diagrams from the surface of the specimen (b) [• line for Si~N~; A) line for SiC; 9 ) line for a-cristobalite]. son of the diffraction patterns recorded from the surfaces of the initial and oxidized specimens indicates that Si3N ~ is preferentially oxidized during heating; the intensity of the lines from Si3N 4 is lowered appreciably after oxidation relative to the lines from SiC. The smaller effect of oxidation on the strength of NKKKM-83a compared with NKKKM-79 is caused by the higher resistance of the former towards oxidation. The increase in the tensile strength of NKKKM-83a, which c o ~ e n c e s at temperatures above II00~ would seem to be caused by the healing of the surface and subsurface defects in the specimens. Tests of specimens 1.5 mm thick were conducted to confirm this hypothesis. With a decrease in the thickness of the specimen the effect of surface defects on strength should have more effect, which actually took place (Fig. 4). The 30% decrease in tensile strength at room temperature indicates that the thickness of the defect surface layer became commensurate with the thickness of the specimen. However, the tensile strength at 1200~ scarcely differs at all from that obtained for specimens 2.3 mm thick. This is caused by the healing of defects as a result of the oxidation which takes place while the specimens are being heated to the test temperature and during isothermal exposure at the given temperature over the course of 15 min. To confirm that the increase in tensile strength at II00-1200~ is, in fact, caused by oxidation and not by the effect of any other factors, a series of specimens were heated by the normal regime to 1200~ but were not tested at this temperature but cooled and tested at room temperature. The results obtained (Fig. 4) indicate conclusively the correctness of the conclusions drawn. Oxidation at 1400-1500~ leads to a decrease in the strength of the material (Fig. i). This is probably caused by the grain boundaries being etched during the oxidation, which leads to the formation of grooves [3], which are stress concentrators and grow into cracks with the application of a load. When considering the effect of oxidation on the stability of the material under steep thermal cycling conditions we studied its thermal stability by the method described in [ii]. This method provides for the determination of the weakening of the material (the change in the value of the tensile strength) after the action of thermal shock. A batch of previously oxidized specimens was heated in a stainless steel cassette placed in a vertical tube furnace. It is known from the results of a previous investigation [9] that for NKKKM-type materials the critical temperature drop &Tcr is 300-500~ After quenching with a temperature drop &T > &Tcr a large decrease occurs in the strength of the specimens, associated with the thermal damage to the structure of the material. The specimens were maintained at a temperature chosen from the range indicated over the course of 15-20 min and then dropped into a water bath having a temperature of 25~ After cooling they were dried in a drying cabinet at II0120~ for 6-8 h. To determine the residual strength the dried specimens were tested on an RM101M unit with pure bending on a support having a distance of 20 mm between the inner points of applying the load and 40 mm between the outer points. 1072

9 _2 Am/S, kg m "10 J

_jr ,~

3

/I 9

2

..---5

/

3

/Z
0

80

t20

160 ~ rain

Fig. 3. Kinetic curves for the oxidation of NI~-79 at 850, 1050, and 1200~ (curves 1-3) and NKKKM-83a at 850, 1050, 1200, 1300, and 1450~ (curves 4-8).

o,MPa

200 ~ f50

I tl,--1

o--2 lO0

[

I

20

200

#80

600

600

.

tO00 1200

T.~

Fig. 4. Dependence of the tensile strength of specimens of NKKKM-83a 1.5 mm thick on the test temperature in air: i) state as supplied; 2) after preliminary heating in air to 1200~ To determine the critical temperature drop from the experimental data we plotted a diagram for the resistance to thermal damage (an RTD diagram) which represents the dependence of the residual strength on temperature drop (Fig. 5). As may be seen, a decrease in the tensile strength of the material is observed after quenching with a temperature drop of more than 400~ For unoxidized material a gradual decrease in strength with an increase in temperature drop is characteristic (curve i in Fig. 5), while for material oxidized at 900 and I050~ the decrease is sharper. Let us also note that preoxidation at the temperatures indicated leads to a sharp drop in the residual strength for temperture drops exceeding ATcr. The residual strength of specimens oxidized at these temperatures is 20-30% lower than fo~ the unoxidized specimens. Such a decrease in strength is possibly caused by an increase in the degree of brittleness due to the increase in density which arises from the internal oxidation of the material which takes place over the temperature range 800-II00~ We may also note that the cristobalite filling the pores of the specimens at approximately 250~ undergoes a phase transformation which is accompanied by a 5% change in volume [ 5 ] . The stresses arising from this may lead to the formation of cracks and, consequently, to a decrease in the tensile strength of the material. For oxidation at temperatures above II00~ a liquid silicate phase appears on the surface of the specimens, which causes a rapid closing of the pores and a healing of the surface defects of the specimens. This leads to an approximately 10% increase in the tensile strength of the material for temperature drops less than ATcr. The RTD diagram of material oxidized at 12000C is similar in nature to that obtained for unoxidized specimens. As may be seen from Fig. 5 an increase in oxidation temperature to 1350~ causes an increase of 25-30~ in ATcr. A decrease in the strength of the material does not take place even after quenching with a temperature drop of 425~ However, even with a slight excess over ATcr a jumplike drop in the strength of the material is observed. Thus, it can be concluded that the oxidation of NKKKM-83a over the temperature range 800-II00~ leads to a decrease in its strength at room temperature and does not affect the strength at 1200~ Oxidation over the temperature range indicated also causes a reduction

1073

O~ MPa

22O

180

140

100 9

~5

400

T,~

GO

20

300

Fig. 5. RTD diagrams of the initial specimens (i) and specimens oxidized at temperatures of 900, 1050, 1250, 1350, and 1450~ (2-6). in the residual strength of specimens of material subjected to thermal shock with a temperature drop greater than the critical drop. Oxidation under the temperature conditions 13001350~ leads to an increase in the strength of the materials at temperatures of 20 and 1200~ and also to a certain increase in the critical temperature drop. With a further increase in the oxidation temperature a decrease in the strength of the material again takes place. LITERATURE CITED i. 2. 3.

4. 5. 6. 7. 8. 9.

i0. ii.

1074

G. S. Pisarenko and G. A. Gogotsi, "A study of the structural strength of ceramics as applied to components of the flow section of GTE," Probl. Prochn., No. 4, 3-10 (1980). N. R. Katz, "Some aspects of materials and structures engineering with ceramics for engine applications," ICM-3, Cambridge, 1979, August, Vol. i, pp. 257-278. S. M. Wiederhorn and N. J. Tighe, "Application of proof testing to silicon nitride," Proc. Workshop on Ceramics for Advanced Heat Engines, Orlando, Jan. 24-27, 1977, pp. 247-258. G. Ziegler and D. Munz, "Mechanical properties of precracked Si3N 4 after different annealing treatments," Sci. Ceram., 9, No. i, 502-509 (1977). A. G. Evans and R. W. Davidge, "The strength and oxidation of reaction sintered silicon nitride," J. Mater. Sci., 5, No. 4, 314-325 (1970). J. E. Easler, R. C. Bradt, and R. E. Tressler, "Effects of oxidation under load on strength distribution of Si3N~," J. Am. Ceram. Soc., 65, No. 6, 317-320 (1982). Yu. G. Gogotsi, S. I. Sopenko, and G. V. Trunov, "The effect of oxidation on the strength of silicon nitride ceramics," Probl. Prochn., No. i, 69-72 (1985). G. A. Gogotsi, "The strength of machine-structural nitride ceramics," Preprint, Inst. Probl. Prochn., Akad. Nauk UkrSSR, Kiev (1982). G. A. Gogotsi, "Some results from studying the mechanical properties of structural ceramics, as applied to engine components," Preprint, Inst. Probl. Prochn., Akad. Nauk UkrSSR (1983). R. Kossowsky, "Cyclic fatigue of hot-pressed Si3N~," J. Am. Ceram. Soc., 56, No. i0, 531-535 (1973). G. A. Gogotsi, A. A. Kushnirenko, and O. M. Kryukova, "A study of refractory ceramics for temperature loadings," Probl. Prochn., No. 6, 69-73 (1977).