ISSN 0031-918X, The Physics of Metals and Metallography, 2008, Vol. 105, No. 1, pp. 45–55. © Pleiades Publishing, Ltd., 2008. Original Russian Text © A.I. Uvarov, V.A. Sandovskii, V.A. Kazantsev, E.I. Anufrieva, N.F. Vil’danova, Yu.I. Filippov, 2008, published in Fizika Metallov i Metallovedenie, 2008, Vol. 105, No. 1, pp. 50–61.

STRUCTURE, PHASE TRANSFORMATIONS, AND DIFFUSION

Effect of Heat and Thermomechanical Treatments on the Structure and Physical and Mechanical Properties of the N30K10T3 Invar A. I. Uvarov, V. A. Sandovskii, V. A. Kazantsev, E. I. Anufrieva, N. F. Vil’danova, and Yu. I. Filippov Institute of Metal Physics, Ural Division, Russian Academy of Sciences, ul. S. Kovalevskoi 18, Ekaterinburg, 620041 Russia Received December 11, 2006; in final form, May 28, 2007

Abstract—Invar alloy N30K10T3, whose austenite is metastable with respect to the martensitic γ α transformation that occurs upon cooling below the martensitic point (Ms = –80°C), has been studied. The following six ways of the alloy strengthening have been tested: (1) aging (a) in a temperature range of ∆Ta = 20–700°C; (2) liquid-nitrogen cooling (lnc) of the material preliminarily hardened by aging under the aforementioned conditions (route 1) (a + lnc); (3) preliminary phase-transformation-induced hardening (ph) (γ α γph) and aging in the temperature range of ∆Ta (ph + a); (4) liquid-nitrogen cooling of the material preliminary hardened via route 3 (ph + a + lnc); (5) preliminary cold deformation (to 30%) at room temperature and aging in a temperature range of ∆Ta (cd + a); and (6) liquid-nitrogen cooling of the material preliminary hardened via route 5 (cd + a + lnc). The six ways of hardening were found to affect the hardness, electrical conductivity, magnetic permeability, and temperature dependence of the thermal expansion coefficient. PACS numbers: 81.40.Gh, 81.30.Kf DOI: 10.1134/S0031918X08010055

INTRODUCTION

in the structure upon deformation and decomposition of the supersaturated solid solution, (6) (liquid-nitrogen cooling) of the material preliminary deformed and aged (cd + a + lnc) to transform part of deformed austenite into martensite, which additionally hardens the invar. The first five routes were studied earlier [1–4]; the procedure comprising cold deformation, aging, and cooling in liquid nitrogen (cd + a + lnc) is considered in this work. The aim of this work is to compare changes in the structure and physical and mechanical properties produced by these six ways of hardening. This will allow researchers and design engineers to choose optimum conditions of hardening upon the production of concrete parts made of an aging invar, which will ensure the part a preset combination of physical and mechanical properties.

The invar alloy N30K10T3 subjected to water quenching from 1150°ë has a ferromagnetic structure. The Curie temperature of the austenite (γ) phase in the alloy is TC ≈ 200°ë; the phase is metastable upon cooling in liquid nitrogen (lnc), since its martensitic point is Ms ≈ –80°ë [1, 2]. Moreover, the γ phase is a supersaturated solid solution capable of undergoing decomposition upon subsequent heating. This material can be hardened using the following six ways: (1) aging (a) in a temperature range of 550–750°ë; the hardening takes place at the expense of precipitation of an intermetallic phase in the austenite [1, 2]; (2) liquid-nitrogen cooling (lnc) of the material preliminarily hardened via route 1 (a + lnc); the cooling-induced martensite which is formed in the quenched and aged samples upon the γ α transformation additionally hardens the alloy [1, 2]; (3) aging of samples preliminary subjected to phase hardening (ph + a) [3]; the decomposition of the supersaturated solid solution in the phase-hardened samples is realized in the same temperatures range as in route 1; (4) (liquid-nitrogen cooling) of the material after phase hardening and aging (ph + a + lnc); the arising cooling-induced martensite additionally hardens the alloy; (5) after quenching cold deformation to 30% at 20°ë is performed followed by aging in the same temperatures range as in route 1 (cd + a) [4]; the hardening occurs at the expense of both dislocations formed

EXPERIMENTAL A model N30K10T3 alloy 10 kg in weight, containing 0.01 wt % C, 30.0 wt % Ni, 10.0 wt % Co, 3.2 wt % Ti, and Fe (for balance) was melted in an induction vacuum furnace using pure components and forged at 1100°ë to rods of different sections. The preliminary treatment consists in an isothermal holding at 1150°ë for 2 h and subsequent water-quenching. The quenched rods were hardened using the six aforementioned ways. 45

46

UVAROV et al. T, °C 4

As–Af temperature range (Fig. 1, point 4), where As and Af are the temperatures of the start and finish of the reverse martensitic transformation, respectively. The heating results in the reverse α γph transformation and formation of the so-called phase-hardened austenite γph. Subsequently, the sample is cooled to room temperature (Fig. 1, point 5). In this case, the alloy structure is characterized by the presence of the hardened γph phase and retained austenite γret (which did not participate in the γ α γph transformations). At room temperature, the phase-hardened austenite is stable, since the Ms temperature of the alloy is negative [6].

α → γph

Af

As

20

γ

γret + α 1

3

–100 –200

γph + γret

RESULTS AND DISCUSSION Structure of the Alloy

5

γ→α 2 Time

Fig. 1. Diagram of heat treatment of metastable austenitic alloys upon phase-transformeation-induced hardening [6].

The following characteristics were measured using the hardened rods: (1) Vickers hardness (measured by the standard procedure); (2) initial magnetic permeability (measured by the procedure described in [7]); (3) electrical conductivity (measured using a Kelvin bridge [8]); and thermal expansion (measured using a DL-1500 PHP (ULVAK-RIKO) dilatometer) and thermal expansion coefficient β (calculated based on the data obtained). The quenched rods were subjected to deformation by cold rolling (at 20°ë) to 30% using a rolling mill equipped with grooved rolls. The structure of the samples was studied using a NEOPHOT optical microscope and a JEM-200 CX electron microscope. The polished sections for the metallographic studies were etched using an electrolyte consisting of 90 and 10 parts of acetic and perchloric acids, respectively. The finishing etching was performed using a 4% solution of nitric acid in ethanol. The foils for the electronmicroscopic studies were prepared using an electrolyte consisting of 800 ml orthophosphoric acid and 170 g of chromic anhydride. The phase-hardening method is described in [6]. Figure 1 shows a schematic diagram of the procedure. Let us consider this diagram in more detail. At room temperature, the initial nonhardened austenite is present (Fig. 1, point 1). The forward martensitic γ α transformation (Fig. 1, point 2) is realized upon cooling in liquid nitrogen to below the martensitic point. Subsequently, the sample is heated to room temperature (Fig. 1, point 3) In this case, the alloy becomes twophase (γ + α) and is heated to a temperature above the

The average grain size of the quenched N30K10T3 invar alloy is 300–350 µm. In spite of the fact that the austenite of the alloy under study is a supersaturated solid solution, no aging of the alloy takes place up to a temperature Ta = 500°ë. As is known [5], the decomposition of the supersaturated solid solution in Fe–Ni–Ti austenitic alloys occurs via two—continuous and discontinuous—mechanisms. In the case of the continuous mechanism, a γ' phase with a chemical composition Ni3Ti precipitates. Similar to the austenite, this intermetallic compound has an fcc lattice. The γ'-phase particles precipitate simultaneously over the entire volume of the sample; however, the rate of precipitation is different in different regions. In the case of discontinuous mechanism, particles of an η phase precipitate but with an hcp lattice. The structure resulting from the discontinuous decomposition consists of plate-type colonies (cells) resembling pearlite colonies in structural low-alloy steels subjected to quenching and tempering. The cells consist of alternating straight and parallel plates of the η and γ (austenite) phases. The γ phase is depleted of nickel and titanium as compared to the quenched alloy. In the temperature range of 550–650°ë, the decomposition of the supersaturated solid solution occurs only via the discontinuous mechanism [1, 2]. After aging at Ta = 700°ë, beginning stages of continuous decomposition are observed along with the discontinuous decomposition. This manifests itself in a thickening of grain boundaries observed in an optical microscope. The temperature range of discontinuous decomposition (DD) of the solid solution is 200 K. It starts at Ta = 700°ë and finishes at Ta = 900°ë. The maximum content of phases formed in the course of DD is observed after aging at Ta = 800°ë. Electron-microscopic studies of quenched invar show the presence of rare martensite plates inside austenite grains (Fig. 2a). The presence of the α phase in the alloy at room temperature indicates the fact that a slightly pronounced isothermal γ α transformation occurs in the quenched N30K10T3 alloy [2]. Upon

THE PHYSICS OF METALS AND METALLOGRAPHY

Vol. 105

No. 1

2008

EFFECT OF HEAT AND THERMOMECHANICAL TREATMENTS

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

47

Fig. 2. Structure of the N30K10T3 invar alloy subjected to different hardening treatments: (a) martensitic crystals in a quenched sample, dark-field image taken in the (10 1 )α reflection, ×73000; (b) quenched and subsequently cooled (in liquid nitrogen) alloy, ×600; (c) crystals of phase-hardened austenite γph; ×200; (d) crystals of athermal α martensite in phase-hardened invar alloy, darkfield image taken in an α-phase reflection, ×59000; (e) strain-induced martensite, dark-field image taken in the (112)α reflection, ×140000; (f) banded structure formed by dislocations in austenitic crystals, ×30000; (g) slip bands decorated with an intermetallic phase; ×200; and (h) aging-induced martensite in deformed and aged alloy, ×100.

cooling of the invar to Tcool ≤ –80°ë, a clearly pronounced athermal martensitic γ α transformation begins. The martensitic point of aging austenite alloys heated to different temperatures after quenching changes [5]. This statement is demonstrated schematically in Fig. 3. Let us discuss this diagram. No aging of the alloy takes place in a temperature range from room temperature T0 to Ta = T1; Therefore, the martensitic point Ms is unchanged is this range. The T1 temperature depends on the titanium content in the quenched austenite. As is shown in [5], the higher the titanium content, the lower the T1 temperature [5]. When the aging temperature increases from T1 to T2, Ms decreases from (Ms)1 to a minimum value (Ms)2 (Fig. 3). As the aging temperature increases from T2 to T3, the martensitic THE PHYSICS OF METALS AND METALLOGRAPHY

point increases to (Ms)3 = (Ms)1. A further increase in the aging temperature from T3 to T4 leads to an increase in the martensitic point to (Ms)4 > (Ms)1. The process of a decrease in the Ms temperature is usually called the austenite stabilization; correspondingly, the temperatures in the T1–T3 range are called stabilizing temperatures [5]. As the aging temperature increases to above T3, an austenite destabilization occurs; the Ms temperature reaches a maximum value and begins decreasing. This can be explained by the fact that, at Ta > T4, the degree of decomposition of the supersaturated solid solution approaches the equilibrium boundary of the γ'-phase solubility in austenite [6]. At Ta = 700°ë, the maximum decomposition of the supersaturated solid solution is observed [6]. Vol. 105

No. 1

2008

48

UVAROV et al. Ms (Ms)4

(Ms)1 = (Ms)3

(Ms)2 T0

T1

T2

T3

T4 Ta

Fig. 3. Dependence (schematic diagram) of the martensitic point Ms on the aging temperature Ta for aging austenitic steels with metastable austenite; T0 = 20°ë [5].

The austenite stabilization can be explained by different causes. We support the conception discussed in [5, 6]. When cooling the aged metastable austenite to T < Ms, the following transformations can occur: (1) γ'-phase particles, remaining coherent, transform simultaneously with the γ phase (austenite), i.e., gain a martensitic (bcc) structure; and (2) γ'-phase particles retain the fcc structure but an incoherent interphase boundary is formed between the martensitic crystals and untransformed particles. The martensitic point decreases owing to the fact that an additional energy is consumed for either fcc–bcc transformation (first variant) or the formation of the interphase boundary between the intermetallic particles and martensite crystals (second variant). Thus, two factors—stabilizing and destabilizing— work upon the decomposition of the supersaturated solid solution. The former factor is due to the energy consumption because of the inheritance of the γ' phase by martensite crystals. The effect of the latter factor is caused by the depletion of the solid solution of nickel and titanium. At aging temperatures of T1–T3, the Ms(Ta) dependence is affected by the first factor; at aging temperatures above T3, the second factor is dominating. After aging at temperatures Ta > T3, the martensitic point increases to above room temperatures. Upon cooling of the alloy with Ms > 20°ë from the aging temperature to room temperature, the γ αa transformation occurs; the αa phase is called the aging-induced martensite [6]. This term is used to distinguish the aging-induced martensite from the cooling-induced martensite (α phase). In the case of cooling-induced martensite, the martensite crystals are formed in alloys with Ms below 0°ë upon cooling in liquid nitrogen (the γ α transformation). After aging of the invar alloy under study at Ta = 600°ë, no αa phase is formed. The

absence of the γ αa transformation is due to the fact that the aforementioned aging temperature corresponds to the martensite-stabilization region. The aging at Ta = 700°ë and cooling to room temperature determine the formation of athermal aging-induced martensite crystals in the alloy structure [2]. The cooling of a quenched sample in liquid nitrogen leads to the formation of athermal α-phase crystals inside the γ phase (Fig. 2b) [1]. Analogous martensitic crystals also are formed upon cooling of aged samples in liquid nitrogen. When the quenched invar is subjected to phase hardening, the isothermal martensite formed upon quenching disappears. This occurs since upon heating to 800°ë the martensitic crystals undergo the reverse α γph transformation. An analogous process is realized in the case of heating of the athermal martensite that is formed upon cooling of a quenched material in liquid nitrogen. In optical micrographs the crystals of the phase-hardened austenite (γph phase) are seen as black needles against the background of the white retained austenite (which did not participate in the γ α γph transformation) (Fig. 2c) [3]. Electron microscopic studies of the phase-hardened invar show the presence of individual thin crystals of the athermal martensite in the alloy structure (Fig. 2d). Their appearance is likely to be due to the following causes [3]. In the course of the reverse martensitic α γph transformation, an austenite aging occurs and the γ' phase precipitates. The precipitation of intemetallics leads to a depletion of the matrix of nickel and titanium; the martensitic point increases to above room temperature. Upon cooling from Ta = 800°ë to room temperature, the phase-hardened material undergoes a martensitic γ αa transformation. The αa crystals are not seen in the optical microscope because of their dispersivity. A study of aged phase-hardened samples using an optical microscope allows us to infer that (1) the higher the aging temperature, the less clear the outlines of phase-hardened austenite crystals because of the precipitation of intermetallic particles inside the γph particles; and (2) an athermal αa phase is formed in the retained austenite after aging at temperatures above 600°ë. The higher the aging temperature, the less distinct the martensite needles become. The structure of the deformed alloy was studied using optical and electron microscopes. The optical microscope did not allow us to reveal the presence of strain-induced martensite. In the electron microscope, individual disperse strain-induced martensite crystals were found (Fig. 2e) [4]. In the γ phase, dislocations forming a banded structure we observed (Fig. 2f) [4]. Upon aging, these dislocations are decorated with the intermetallic phase; as a result, this banded structure can be observed not only in the electron microscope but also in the optical microscope (Fig. 2g) [4]. Thus, the

THE PHYSICS OF METALS AND METALLOGRAPHY

Vol. 105

No. 1

2008

EFFECT OF HEAT AND THERMOMECHANICAL TREATMENTS σ, MS/m

HV 510 480 450 420 390 360 330 300 270 240 210 180

49

(a)

4 5 6 3 2 1 0

200

400

4 2.4 (b) 3 2.3 2.2 2.1 2.0 1.9 4 1.8 5 1.7 1.6 1.5 6 1.4 2 1.3 3, 5 1.2 1 1.1 1.0 600 800 0 200 400 600 800 Ta, °C Ta, °C

(c)

µ 50 40

2 1 3

30 20 10

5 4 6 0

200

400

600

800 Ta, °C

Fig. 4. Dependences of (a) the hardness HV, (b) electrical conductivity σ, and (c) magnetic permeability µ on the aging temperature Ta for samples subjected to different treatments: (1) quenching (q) + aging (a); (2) q + a + liquid-nitrogen cooling (lnc); (3) phase hardening (ph) + a; (4) ph + a + lnc; (5) q + cold deformation (cd) + a; and (6) q + cd + a + lnc (τa = 1 h).

structure of the deformed invar consists of the austenite with a high density of dislocations, which are decorated with intermetallic particles, particles of strain-induced martensite αd, and deformed isothermal marteniste αis that precipitates during the quenching of the material. The decomposition of the supersaturated solid solution at Ta ≥ 650°ë destabilizes the austenite; this results in the martensitic γ αa transformation upon cooling of the alloy from the aging temperatures to room temperature. Inside austenitic grains, crystals of athermal αa phase appear (Fig. 2h) [4]. As the aging temperature increases to Ta = 700°ë, the boundaries of austenite grains become thicker. This indicates that the discontinuous mechanism of the decomposition of the supersaturated solid solution manifests itself at this heating temperature. No cells of discontinuous decomposition are seen in an optical microscope. They can be resolved only in an electron microscope. The cooling of the deformed and aged samples in liquid nitrogen causes a partial transformation of the γ phase into the cooling-induced athermal martensite (γ α). The structure of the deformed alloys aged at Ta ≥ 650°ë is a mixture of two martensites, i.e., cooling-induced (α) and aging-induced (αa). Hardness of the Alloy After quenching, the invar alloy has a hardness HV = 190, which is unchanged upon aging at Ta = 500°ë (Fig. 4a, curve 1) since, at these temperatures, no decomposition of the supersaturated solid solution occurs. In the case of treatment via route 1, the heating in a temperature range of 600–750°ë increases the THE PHYSICS OF METALS AND METALLOGRAPHY

hardness of the material due to both the precipitation of intermetallic particles in the austenite and formation of aging-indiced martensite crystals (αa) in the matrix. The cooling of a quenched sample in liquid nitrogen leads to the occurrence of the γ α transformation. The treatment via route 2 (γ α) determines an additional increase in the hardness of all aged samples (Fig. 4a, curves 1, 2). However, for both routes of treatment, the HV(Ta) dependences show the same behavior (compare curves 1 and 2 in Fig. 4a). After strengthening of the quenched invar via phase hardening (γ α γph), its hardness increases (Fig. 4a, curve 3). Upon heating from room temperature to 400°ë, the level of strength properties of the phasehardened alloy remains unchanged, since no decomposition of the supersaturated solid solution occurs at these aging temperatures. An increase in the hardness is observed only at Ta ≥ 500°ë. A comparison of curves 1 and 3 (Fig. 4a) indicates that the cycle of γ α γph transformations accelerates aging. Upon aging at Ta = 700°ë, an overaging takes place in both quenched and phase-hardened alloy; therefore, a comparison of the hardnesses of the invar alloys after treatments at Ta = 700°ë is not demonstrative. The cooling of phase-hardened and aged samples in liquid nitrogen increases their hardness (compare curves 3 and 4 in Fig. 4a); the hardness remains virtually unchanged at temperatures from room temperature to Ta = 500°ë after which it increases and reaches a maximum value at Ta = 700°ë. The plastic deformation (to 30%) of the quenched invar alloy increases its hardness (compare curves 1 and 5 in Fig. 4a). As is seen from Fig. 4a (curve 5), the heatVol. 105

No. 1

2008

50

UVAROV et al. HV 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 180

(a)

4

6 2 3 5

1

0

1

2

3 4 τa, h

5

6

σ, MS/m 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1.0

2

(b)

4

6 3

µ 50

(c)

40

1

30

2 5 3 4 6

20 1 5 0

1

2

3 4 τa, h

5

6

10 0

0

1

2

3 4 τa, h

5

6

Fig. 5. Dependences of (a) the hardness HV, (b) electrical conductivity σ, and (c) magnetic permeability µ on the aging time τa at Ta = 600°ë for samples subjected to different treatments: (1) q + a; (2) q + a + lnc; (3) ph + a; (4) ph + a + lnc; (5) q + cd + a; and (6) q + cd + a + lnc.

ing of the deformed γ phase to Ta = 400°ë does not change its hardness; it increases upon aging at Ta ≥ 550°ë. A comparison of the HV(Ta) dependences for the undeformed and deformed material (Fig. 4a, curves 1, 5) allows us to infer the intensification of the decomposition of the saturated solid solution in the deformed material. The acceleration of aging decreases to zero in the case of heating of the alloy to Ta = 700°ë. This is due to the overaging of the solid solution. The sixth way of strengthening (cd + a + lnc) does not change the behavior of the HV(Ta) dependence as compared to that is route 5 (compare curves 5 and 6 in Fig. 4a) but increases the hardness of the deformed and aged material. The increase in the hardness is due to the transformation of part of austenite into martensite (γ α) upon cooling. The hardness of the alloy changes with changing not only aging temperature but also aging time. After aging at Ta = 600°ë, the hardness increases monotonically with increasing time (Fig. 5a, curve 1). However, after isothermal holding for τa = 2 h, virtually no increase in the hardness is observed. The monotonic increase in the hardness at Ta = 700°ë is violated already after aging for τa > 1 h. The decrease in HV is related to overaging processes that structurally manifest themselves in the fact that a discontinuous decomposition begins to develop at grain boundaries; in this case, γ'-phase particles become larger.

The quenched alloy subjected to treatments via routes 1 and 2 exhibit analogous HV (τa) dependences (compare curves 1 and 2 in Fig. 5a). The difference consists in the fact that after route 2 the hardness is higher owing to an additional formation of coolinginduced martensite in the structure. The hardness of the phase-hardened invar aged at Ta = 600°ë increases monotonically with increasing time of isothermal holding (Fig. 5a, curve 3). In this case, the hardness is higher than that of quenched alloy (that was not subjected to phase hardening) (compare curves 1 and 3 in Fig. 5a). The comparison of these HV(τa) dependences allows us to affirm that the decomposition of the supersaturated solid solution in the phase-hardened invar is more intense as compared to the quenched alloy. As the aging temperature increases from 600 to 700°ë, the HV(τa) dependence exhibits nonmonotonic changes. In this case, the maximum hardness is reached at τa = 1 h and, owing to overaging processes, the hardness decreases with increasing aging time [3]. The alloys aged at Ta = 700°ë and 600°ë exhibit virtually the same maximum hardness. Let us consider the HV(τa dependences for the phase-hardened and aged alloys after cooling in liquid nitrogen. At Ta = 600°ë, the hardness increases monotonically (Fig. 5a, curve 4). The cooling of the alloy in liquid nitrogen results in an increase in the hardness (compare curves 3 and 4 in Fig. 5a). At Ta = 700°ë, the

THE PHYSICS OF METALS AND METALLOGRAPHY

Vol. 105

No. 1

2008

EFFECT OF HEAT AND THERMOMECHANICAL TREATMENTS

dependence exhibits nonmonotonic variations; the hardness reaches the maximum value at τa = 1 h [3]. In this case, the hardnesses of the material cooled in liquid nitrogen differs from that of the uncooled material only slightly. The HV(τa) dependence of the invars deformed to 30% at room temperature decreases monotonically after aging at Ta = 600°ë; after τa = 6 h, the increase in the hardness virtually ends. However, when the aging temperature increases to Ta = 700°ë, the HV(τa) dependence becomes nonmonotonic and the hardness reaches a maximum magnitude at τa = 1 h. As the time of isothermal holding increases (τa > 1 h), the hardness decreases owing to the development of overaging processes [4]. The HV(τa) dependences of the alloys subjected to treatments via routes 5 and 6 are analogous (compare curves 5 and 6 in Fig. 5a). The difference consists in the higher hardness of the alloy after treatment via route 6. This is due to the fact that a cooling-induced martensite is formed additionally in the structure of the latter alloy. Electrical Conductivity of the Alloy Upon aging, the electrical conductivity of the quenched alloy remains unchanged to Ta = 600°ë; at Ta = 700°ë, the electrical conductivity is maximum (Fig. 4b, curve 1). Upon cooling of the quenched alloy in liquid nitrogen, its electrical conductivity increases, i.e., the formation of martensite crystals in the austenitic structure leads to an increase in σ (compare curves 1 and 2 in Fig. 4b). In the case of these two treatments, the σ(Ta) dependences exhibit the same behavior. The phase hardening virtually does not change the conductivity of the alloy. Upon aging of the phasehardened alloy, its conductivity remains unchanged to Ta = 600°ë and then increases abruptly with increasing aging temperature (Fig. 4b, curve 3). It should be noted that the hardness increases (Fig. 4a, curve 3) and the conductivity is unchanged (Fig. 4b, curve 3) after aging in a temperature range of 400–600°C. This is likely to indicate that the effect of dispersivity of intermetallic particles on these parameters is different. Disperse particles virtually do not change σ and markedly increase HV. A comparison of curves 3 and 4 (Fig. 4b) shows that, upon aging in a temperature range of 20– 600°ë, the electrical conductivity of the alloys cooled in liquid nitrogen is higher than that of the non-cold-treated alloy. In both cases, the maximum σ magnitudes (at Ta = 700°ë) virtually coincide. The conductivity of the quenched alloy deformed to 30% decreases only slightly (compare curves 1 and 5 in Fig. 4b); σ remains unchanged upon aging in a temperature range of 20–600°ë and increases substantially after heating to Ta = 700°ë (Fig. 4b, curve 5). The increase in the electrical conductivity is due to the THE PHYSICS OF METALS AND METALLOGRAPHY

51

appearance of the aging-induced martensite in the structure. The cooling of deformed (to 30%) alloy in liquid nitrogen increases σ. After treatments via routes 5 and 6, the σ(Ta) dependences are analogous (compare curves 5 and 6 in Fig. 4b). The difference consists in that after treatment via route 6 the electrical conductivity is higher for all aging temperatures owing to the formation of cooling-induced martensite. The decomposition of the supersaturated solid solution at Ta = 600°ë affects only slightly the electrical conductivity. The virtually initial level of σ begins to change after isothermal holding for τa ≥ 4 h (Fig. 5b, curve 1). However, at the aging temperature Ta = 700°ë, the σ(τa) dependence is a monotonically increasing function [1]. Such a difference in the σ(τa) dependences obtained upon aging at 600 and 700°ë can be explained by the fact that upon aging at 600°ë there operates a single factor (the precipitation of intermetallic particles), whereas upon aging at 700°ë, the formation of aging-induced martensite crystals (resulting from the γ αa transformation) additionally affects the electrical conductivity. The liquid-nitrogen cooling of the samples aged at Ta = 600°ë substantially increases their conductivity (Fig. 5b, curves 1, 2). After treatment via route 2, the σ(τa) increases monotonically. With increasing time of isothermal holding at Ta = 600°ë, the conductivity of the phase-hardened material increases monotonically (Fig. 5b, curve 3). At Ta = 700°ë, the σ(τa) dependence also is a steadily increasing function [3]. Thus, at all aging temperatures, the conductivity of the phase-hardened alloy is higher substantially as compared to that of the quenched invar (compare curves 1 and 3 in Fig. 5b). After liquid-nitrogen cooling of the samples aged at Ta = 600°ë, their conductivity changes only slightly (Fig. 5b, curve 4). At Ta = 700°ë, the conductivity is maximum at τa = 1 h; with increasing time of isothermal holding, the magnitude remains unchanged [3]. At Ta = 600°ë, the increase in τa virtually does not affect the electrical conductivity of deformed (to 30%) alloy (Fig. 4b, curve 5); at Ta = 650°ë, the deformation leads to a monotonic increase in σ [4]. After aging at Ta = 700°ë for 6 h, the conductivity increases substantially [4]. The high σ value after aging at Ta = 700°ë is likely to be due to both the presence of aging-induced martensite crystals in the structure and increase in the amount of precipitating intermetallics. The liquidnitrogen cooling of the deformed samples aged at Ta = 600°ë increases σ (compare curves 5 and 6 in Fig. 5b). In this case, the σ(τa) dependences obtained for the samples treated in accordance with routes 5 and 6 ways differ slightly (Fig. 5b, curves 5, 6). Vol. 105

No. 1

2008

52

UVAROV et al.

Magnetic Permeability of the Alloy Upon aging of the quenched alloy, the magnetic permeability remains unchanged to Ta = 500°ë and subsequently decreases as the aging temperature increases to Ta = 700°ë (Fig. 4c, curve 1). The decrease in the magnetic permeability is due to the following two factors: (1) decomposition of the supersaturated solid solution and (2) transformation of part of austenite into aginginduced martensite (γ αa) whose µ magnitude is lower than that of the austenite. Upon cooling of the quenched sample in liquid nitrogen, the magnetic permeability is virtually unchanged. The tendency to a decrease in µ upon treatment via route 2 is observed for all aged samples. After treatment via routes 1 and 2, the µ(Ta) dependences are analogous (Fig. 4c, curves 1 and 2). After the cycle of γ α γph transformations, the magnetic permeability of the invar changes insignificantly. The aging of the phase-hardened austenite at temperatures below Ta = 500°ë does not change µ. The magnetic permeability of the phase-hardened alloy decreases at Ta ≥ 600°ë and reaches a maximum value µ = 23 at Ta = 700°ë (Fig. 4c, curve 3). The decrease in the magnetic permeability of the phase-hardened alloy after aging is due to the same causes as in the case of the quenched alloy. After cooling in liquid nitrogen, the magnetic permeability of the phase-hardened and aged alloys decreases (compare curves 3 and 4 in Fig. 4c). Upon aging in a temperature range of 20–500°ë, the magnetic permeability of the cooled alloy remains unchanged; the value of the permeability is lower than that for the non-cold-treated alloy. After aging at Ta = 700°ë, the µ value of the cold-treated alloy is minimum. The deformation to 30% at 20°ë results in a decrease in the magnetic permeability (compare curves 1 and 5 in Fig. 4c). As the aging temperature increases, the level of µ of the deformed material remains virtually unchanged to Ta = 600°ë; after aging at Ta = 700°ë, the magnetic permeability decreases (Fig. 4c, curve 3). The decrease in µ is likely to be due to two causes; these are (1) the appearance of the aginginduced martensite in the alloy structure and (2) intense decomposition of the supersaturated solid solution. Upon cooling of the deformed and aged samples in liquid nitrogen, the magnetic permeability decreases (compare curves 5 and 6 in Fig. 4c). The µ(Ta) dependence in the case of treatment routes 5 and 6 are analogous; the treatment via route 6 leads to the lower values of µ. The decrease in µ is due to the transformation of part of austenite into the cooling-induced martensite (γ α). As the time of aging increases, the magnetic permeability of the quenched alloy decreases (Fig. 5c, curve 1). After treatment via routes 1 and 2, the µ(τa) dependences are analogous (compare curves 1 and 2 in

Fig. 5c). The difference consists in the fact after route 2, the µ values are lower. The increase in the aging temperature from 600 to 700°ë leads to an additional decrease in µ; the general behavior of the µ(τa) curves remains the same [1]. After treatments via routes 3 and 4 (Fig. 5c, curves 3, 4), the regularities of the behavior of the µ(τa) curves are analogous to those observed for routes 1 and 2 (Fig. 5c, curves 1, 2). Curves 3 and 4 lie below curves 1 and 2 (Fig. 5c). An increase in τa at Ta = 600°ë does not affect the magnetic permeability of the deformed (to 30%) alloy (Fig. 5c, curve 5). The isothermal holding of the deformed sample at Ta = 700°ë for τa = 6 h leads to a monotonic decrease in µ [4]. The cooling of deformed and aged samples in liquid nitrogen decreases the magnetic permeability (compare curves 5 and 6 in Fig. 5c). The magnetic permeability of the sample aged at Ta = 600°ë for 6 h is µ = 5, i.e., it is equal to the magnetic permeability of the alloy aged at Ta = 700°ë (τa = 6 h). Variations of the Thermal Expansion Coefficient Figure 6 shows a schematic diagram of variations of the thermal expansion coefficient β (TEC) of samples that were hardened using different treatment routes (curves 1–6). In general, the temperature dependence of the TEC β(T) can be represented by curve 1 (Fig. 6). This dependence consists of three parts. In the temperature range of invar behavior ∆T = T1 – T0 (where T0 is room temperature and T1 is the maximum temperature of alloy heating at which no abrupt changes in β occurs), the austenite is ferromagnetic. In the temperature range of T1–T2 (Fig. 6, curve 1), substantial changes in β are observed. In this case, the increase in β is related to the fact that the ferromagnetic austenite transforms into the paramagnetic state. In the temperature range of T2–T3 (Fig. 6, curve 1), a gradual quite insignificant increase in β of the paramagnetic austenite is observed. In the case of the single-phase (γ phase) state of the invar, the subsequent gradual slight increase in β is observed in a temperature range of T3–T4 (Fig. 6, curve 1). However, if the alloy is two-phase (γ + α), a reverse transformation of martensite crystals (α phase) into the phase-hardened austenite (α γph) occurs upon heating to a temperature TB (Fig. 6, curve 1). As a result of the reverse martensitic transformation, the β(T) dependence exhibits a minimum, which is shown in Fig. 6 (curve 1) as ABC. Let us consider the ABC minimum for the N30K10T3 alloy in more detail [1, 2]. Since Fig. 6 shows a schematic diagram of variations of the β(T) dependence, it should be kept in mind that the position of point B in curve 1 is arbitrary, i.e., the β value determined in the β1 axis is not related to the β value at point β1. The real values corresponding to these points will be given below. Curve ABC only indi-

THE PHYSICS OF METALS AND METALLOGRAPHY

Vol. 105

No. 1

2008

EFFECT OF HEAT AND THERMOMECHANICAL TREATMENTS

cates the decrease in β in the course of the reverse martensitic transformation upon heating. All what was said above about the magnitudes corresponding to point B (curve ABC) is true for points M (DMH), P(NPQ), and K (LKS). The TEC magnitude of the quenched N30K10T3 alloy at room temperature is 2.5 × 10–6 K–1 (point β1 in curve 1, Fig. 6). The β value gradually increases in the temperature range of the invar behavior of 20–175°C [1, 2]. Upon heating of the material at T > T1, β increases, which is due to the transition of ferromagnetic austenite into the paramagnetic state. At T2 ≈ 200°ë, the transformation finishes. The smooth behavior of the β(T) dependence is violated near 560°ë and β decreases (Fig. 6, curves 1, ABC). The minimum TEC value at point B in curve ABC is βmin = –1 × 10–6 K–1 [2]; at points A and C, the β value is about 16 × 10–6 K–1. However, it should be understood that at TB ≈ 560°ë, β ≠ βmin. At T = 560°ë, crystals of isothermal martensite undergo a reverse transformation into the phase-hardγph) [3]; i.e., the ABC minimum ened austenite (αis (Fig. 6, curve 1) is due to a dynamic effect. After the aforementioned reverse transformation is completed γph), the invar has a value of β ≈ 16 × 10–6 K–1. (αis Thus, near the heating temperature of 560°ë, the TEC magnitudes virtually fall in the AC line (Fig. 6, curve 1). The AC line is drawn as a dashed line in order to emphasize the dynamic character of the β(T) variation in the temperature range under consideration. The dynamic effects in curves 3 (DMH), 4 (NPQ), 5 (WGF), and 6 (LKS) (Fig. 6) should be considered analogously. As a result of aging at Ta = 600°ë (τa = 1 h), the TEC at T0 = 20°ë increases from β1 = 2.5 × 10–6 K–1 to β2 = 5 × 10–6 K–1 (β2 > β1, compare curves 1 and 2 in Fig. 6), whereas the T1 temperature decreases to 115 °ë [2]. The ABC minimum observed for the quenched state is absent (compare curves 1 and 2 in Fig. 6). This is related to the fact that the heating of the sample to 600°ë causes the transformation of isothermal martensite crystals formed upon quenching into the γph phase (α γph). As the aging temperature increases from 600°C to 700°ë (τa = 1 h), the TEC magnitude at T0 increases to β3 = 8 × 10–6 K–1. Moreover, the β(T) dependence demonstrates an abrupt decrease in β in the temperature range of 350– 600°ë and a minimum at Tm = 500°ë. The appearance of the DMH minimum (Fig. 6, curve 3) is due to the fact that, upon heating of the sample (which was preliminary aged at Ta = 700°ë) in the dilatometer, a reverse transformation of the γph) aging-induced martensite into austenite (αa γph transformation occurs. The wide range of the αa is likely to be due to the inhomogeneity of the martensite crystals (αa), which inherit intermetallic particles precipitated in the course of the preliminary decomposition of the supersaturated solid solution [5, 6]. THE PHYSICS OF METALS AND METALLOGRAPHY

β

53 A

C

B β1

1

β2

2

β3

3

H

D

N

M

β4

4

β5

5

β6

6

Q

P F

W G

L

S

K 20 T0

TK

T1 T2

T3

TM TB TG TP

T4

Temperature, °C Fig. 6. A schematic diagram of the temperature dependence of the thermal expansion coefficient β for the N30K10T3 alloy subjected to different treatments: (1) q; (2) q + stabilizing aging (Ta = 600°ë); (3) q + destabilizing aging (Ta = 700°ë); (4) ph; (5) q + cd; and (6) q + cd + a + lnc.

Figure 6 (curve 4) shows the β(T) dependence for the phase-hardened alloy. A comparison of these dependences for the quenched and phase-hardened alloys (Fig. 6, curves 1, 4) gives grounds for the following conclusions: (1) the disappearance of the ABC minimum (Fig. 6, curve 1) is due to the transformation of isothermal-martensite crystals into the phase-hardened γph) during the heating of the alloy to austenite (αis 800°ë; (2) the phase-hardened alloy exhibits a slight decrease in the TEC measured at room temperature from β1 = 2.5 × 10–6 K–1 to β4 = 2 × 10–6 K–1; (3) the cycle of γ α γph transformations results in a decrease in the T1 temperature from 175 to 140°ë, i.e., the range of the invar behavior becomes narrower; (4) in the temperature range of heating of the phasehardened alloy of 575– 675°ë, there is observed a decrease in the TEC. The NPQ minimum (Fig. 6, curve 4) in the β(T) dependence at TP ≈ 640°ë can be explained as follows. To realize the cycle of γ α γph transformations, some excess holding at 800°ë should be used. Vol. 105

No. 1

2008

54

UVAROV et al.

Such a heating process of the alloy leads to the increase in the martensitic point of quenched austenite from the range of negative temperatures to the range of positive temperatures. So upon cooling of the invar from 800°ë to room temperature, aging-induced martensite (αa phase) is formed in it. Upon subsequent heating, the γph αa phase becomes unstable and the reverse αa transformation occurs near TP (Fig. 6, curve 4). For the quenched and aged (Ta = 700°ë, τa = 1 h) alloy, β = 8 × 10–6 K–1 at T0 = 20°ë; the TEC of the phase-hardened alloy subjected to the same heat treatment increases and reaches a magnitude β = 10 × 10–6 K–1. After such a heat treatment, the DMH minimum (Fig. 6, curve 3) is observed for both quenched and phase-hardened alloys, since the causes of its appearance are the same in both cases. If the quenched N30K10T3 Invar is deformed to 30% at room temperature, its TEC decreases from β1 = 2.5 × 10–6 K–1 to β5 = 1.5 × 10–6 K–1 [4]. Such a plastic deformation does not change the range of invar behavior. For the deformed alloy, an WGF minimum has been observed in the β(T) dependence (Fig. 6, curve 5) at T0 = 620°ë [4]. The formation of this minimum can be explained as follows. The structure of the deformed alloy consists of three phases; these are the austenite, crystals of isothermal matrensite αis, and strain-induced martensite αd. Upon heating of this three-phase structure in the dilatometer in the temperature range of 550– 675°ë, the martensitic phase becomes unstable and a reverse (αis + αd) γph transformation occurs. The decomposition of the supersaturated solid solution in the deformed alloy at Ta = 600°ë (τa = 1 h) does not lead to the formation of αa crystals inside austenitic grains. No aging-induced martensite is formed after the aforementioned treatment, since the temperature Ta = 600°ë corresponds to the austenite-stabilization region. The aging results in an increase in the TEC from 1.5 × 10–6 to 2.5 × 10–6 K–1. In this case, the range of invar behavior becomes narrower [4]. As the time of isothermal holding increases from 1 to 6 h, the TEC increases and reaches 3.5 × 10–6 K–1; the invar range continues the narrowing [4]. If the decomposition of the supersaturated solid solution is realized at 700°ë (τa = 1 h) rather than at 600°C, the TEC measured at T0 = 20°ë increases to β ≈ 6 × 10–6 K–1. Such an aging leads to the formation of crystals of aging-induced martensite (γ αa) within austenitic grains. The formation of the αa phase is due to the fact that the temperature Ta = 700°ë corresponds to the destabilization temperature range. In this case, upon heating of the invar in the temperature range of 350– 670°ë, the aging-induced martensite is unstaγph transformation occurs. ble and the reverse αa This transformation determines the appearance of the DMH minimum in the case of both quenched and aged phase-hardened samples.

After cooling of a quenched sample in liquid nitrogen, its TEC at T0 = 20°C increases to β6 = 5 × 10–6 K–1. This increase is due to the formation of coolinginduced martensite in the alloy. The α phase is unstable near TK ≈ 150°ë (in a temperature range of 100–200°ë) and undergoes the reverse α γph transformation [1], which results in the LKS minimum in the β(T) dependence (Fig. 6, curve 6). An analogous β(T) dependence is also observed in the case of deformed and liquidnitrogen-cooled austenite. If a sample whose structure is characterized by the presence of aging-induced martensite is cooled in liquid nitrogen, the β(T) dependence exhibits the behavior shown in Fig. 6, curve 3. Upon the formation of cooling-induced martensite in an aged sample, the α phase is likely to inherit intermetallic particles and becomes equivalent (in its structural state) to the aging-induced martensite. CONCLUSIONS (1) The decomposition of the supersaturated solid solution in the quenched, phase-hardened, and deformed N30K10T3 invar alloy increases its hardness HV and electrical conductivity σ and decreases its magnetic permeability (µ). Maximum HV and σ and minimum µ values are observed after aging at Ta = 700°ë. The changes in HV, σ, and µ are due to the formation of both intermetallic particles and aging-induced martensite, which is formed in the alloy owing to the increase in the martensitic point to above room temperature. (2) The isothermal holding upon aging leads to a monotonic increase in the hardness and electrical conductivity and a decrease in the magnetic permeability. (3) Upon cooling in liquid nitrogen, both quenched, phase-hardened, and deformed samples after aging undergo a martensitic γ α transformation, which determines an increase in HV and σ and a decrease in µ. In this case, the behavior of the HV(Ta), σ(Ta), and µ(Ta) dependences are the same for both cold-treated and non-cold-treated samples. (4) At room temperature, the quenched N30K10T3 invar alloy exhibits a low thermal expansion coefficient β = 2.5 × 10–6 K–1. The range of invar behavior of the quenched alloy corresponds to a temperature range of 20–175°ë. The phase hardening decreases β and narrows the range of invar behavior. Cold (at 20°C) plastic deformation (to 30%) decreases β; in this case, the range of invar behavior remains unchanged and corresponds to the range of quenched undeformed material. (5) The decomposition of the supersaturated solid solution increases the initial TEC of quenched, phasehardened, and deformed samples. In this case, the higher the temperature and the time of aging, the higher the increase in the β magnitude; the range of invar behavior becomes narrower. (6) In quenched, phase-hardened, and deformed samples, different martensitic crystals (isothermal,

THE PHYSICS OF METALS AND METALLOGRAPHY

Vol. 105

No. 1

2008

EFFECT OF HEAT AND THERMOMECHANICAL TREATMENTS

aging-induced, and deformation-induced martensite, respectively) are formed. Upon aging at Ta = 700°ë, in all the cases crystals of aging-induced martensite are formed additionally. All these martensitic phases are unstable upon heating and transform into the phasehardened austenite. This transformation is accompanied by the formation of corresponding minima in the β(T) dependences. (7) Upon cooling of the quenched and aged material in liquid nitrogen, cooling-induced martensite α is formed. Upon heating, the α phase transforms into phase-hardened austenite, which determines the formation of a minimum in the β(T) dependence. REFERENCES 1. V. A. Sandovskii, A. I. Uvarov, V. A. Kazantsev, et al., “Study of Decomposition of Supersaturated Solid Solution and Martensitic Transformation in N30K10T3 Invar,” Fiz. Met. Metalloved. 96 (5), 67–73 (2003) [Phys. Met. Metallogr. 96 (5), 502–508 (2003)]. 2. A. I. Uvarov, V. A. Kazantsev, E. I. Anufrieva, and T. P. Vasechkina, “Effect of Heat Treatments on the Thermal Expansion Coefficient and Magnetic Moment of Samples of an Invar Alloy N30K10T3,” Fiz. Met.

THE PHYSICS OF METALS AND METALLOGRAPHY

3.

4.

5. 6.

7. 8.

55

Metalloved. 98 (4), 35–43 (2004) [Phys. Met. Metallogr. 98 (4), 368–376 (2004)]. A. I. Uvarov, V. A. Sandovskii, V. A. Kazantsev, et al., “Effect of Heat Treatments on the Physicomechanical Properties of a Transformation-Hardened Invar Alloy N30K10T3,” Fiz. Met. Metalloved. 99 (6), 94–102 (2005) [Phys. Met. Metallogr. 99 (6), 640–648 (2005)]. A. I. Uvarov, V. A. Sandovskii, V. A. Kazantsev, et al., “Effect of Preliminary Plastic Deformation on the Structure and Physicomechanical Properties of Aging Invar Alloy N30K10T3,” Fiz. Met. Metalloved. 101 (4), 392– 399 (2006) [Phys. Met. Metallogr. 101 (4), 362–368 (2006)]. V. V. Sagaradze and A. I. Uvarov, Strenghtening Austenitic Steels (Nauka, Moscow, 1989) [in Russian]. K. A. Malyshev, V. V. Sagaradze, I. P. Sorokin, et al., Phase-Transformation-Induced Hardening of Fe–NiBased Austenitic Alloys (Nauka, Moscow, 1982) [in Russian]. V. V. Dyakin and V. A. Sandovskii, Theory and Calculation of Laid-on Eddy-Current Transducers (Nauka, Moscow, 1981) [in Russian]. B. G. Livshits, V. S. Kraposhin, and Ya. L. Linetskii, Physical Properties of Metals and Alloys (Moscow, Metallurgiya, 1980) [in Russian].

Vol. 105

No. 1

2008