February 29, 2008

Formation of Fine Grained Globular Microstructures During Isothermal Annealing of PM T42 High Speed Steel: Role of Nanometric Proeutectoid Carbides

By Trabadelo, V Iturriza, I

The microstructure of powder metallurgy T42 high speed steel has been analysed after isothermal annealing. The heat treatment parameters (austenitising temperature, annealing time and annealing temperature) have been chosen aiming at obtaining a moderate hardness (50 HRC) adequate for structural applications. It has been demonstrated that it is possible to attain the targeted hardness through a globular pearlitic matrix microstructure at temperatures markedly lower than A^sub 1^ and by short time annealing cycles. The key factor for the generation of such microstructure is the precipitation of a homogeneous distribution of nanometric tungsten rich M^sub 6^C proeutectoid carbides in the austenite. The presence of these carbides avoids the growing of lamellar pearlite independently of the annealing time. Keywords: Isothermal annealing, Nanometric proeutectoid carbides, Undercooling, Globular pearlite, Microstructural evolution


The properties of a high speed steel (HSS) can be modified in a wide range by the application of different heat treatments.1-3 The main use of these steels is cutting tools, where a high hardness (60- 70 HRC) and wear resistance are needed. This imposes the use of hardening and tempering as the final heat treatment in order to obtain a final microstructure consisting of a tempered martensitic matrix with primary and secondary carbides embedded. Before this last processing step, annealing is conventionally used to soften the material in order to perform the required machining operation under the best conditions (hardness 15-25 HRC). From a microstructural point of view, annealing produces a microstructure consisting of uniformly dispersed spheroidised carbides in a ferritic matrix. The main spheroidising processes can be divided into four categories:4

(i) isothermal annealing at a temperature slightly below Ai (subcriticai annealing)

(ii) repeated heating and cooling cycles between temperatures just above and just below A^sub 1^

(iii) heating just above A^sub 1^ followed by slow cooling

(iv) heating just above A^sub cm^ and slow cooling or isothermal annealing at a temperature below A^sub 1^ for a long time.

The resulting spheroidised carbides ranged between 50 and 300 nm in diameter.5 In addition, the application of HSS as structural material makes the use of intermediate hardness between 35 and 55 HRC attractive. This range of hardness can be obtained by different heat treatments such as isothermal annealing at temperatures far below A^sub 1^, overtempering or transformation annealing.1,6,7

The main objective of the present work is to characterise the transformation of the austenite in a powder metallurgy T42 HSS isothermally annealed at temperatures far below A i and to propose a mechanism consistent with the observed microstructural features. The studied annealing treatments were developed by the authors in previous works6,7 in order to obtain a material with a hardness value of ~ 50 HRC.


T42 HSS powder was water atomised and subsequently hydrogen annealed by Hoganas Great Britain Ltd. The chemical composition of the powder given by the manufacturer is Fe-1.3C-10.4Co-4.2Cr-3.8Mo- 0.3Si-3.3V-10.0W (wt-%). An addition of 0.4 wt-% of elemental carbon in the form of graphite powder was made to the T42 powder in order to reduce its optimum sintering temperature and to widen its sintering window.8 The obtained powder mixture (T42 + C) was uniaxially cold pressed at 600 MPa to ~70% of the theoretical density. The green compacts were sintered to full density under vacuum atmosphere at 1220[degrees]C for 1 h and cooled down to room temperature inside the furnace.

The sintered specimens (cylinder of 16 mm diameter and 3 mm high) were austenitised for 15 min under an argon flowing atmosphere and subsequently rapidly immersed into a salt bath for the isothermal annealing. Once the isothermal soaking was completed, the samples were water quenched. The austenitising temperatures were varied between 950 and 1200[degrees]C and the annealing temperatures ranged from 600 to 675[degrees]C. The annealing time was varied between 5 min and 24 h. The hardness of the samples was measured with a Mitutoyo ATKF1000 hardness tester using the Rockwell C scale.

1 Images (SEM) of as sintered T42+C cooled down to room temperature inside sintering furnace

Cross-sections were mechanically polished, etched using a solution of 2% nital and observed by a Philips XL30 scanning electron microscope (SEM). Extraction replicas were obtained by etching in 2% nital or in diluted Villela's reagent and subsequently coated with a thin layer of carbon. The replicas were stripped from the specimens by electrolytic etching in 10% nital using a potential of ~ 30 V. The examination was carried out at 100 kV in a Philips CM- 12 transmission electron microscope (TEM) (LaB^sub 6^ filament) fitted with an energy dispersive X-ray spectroscopy (EDS) system. The chemical composition of the extracted precipitates was determined by analysing at least 30 particles by EDS.

Dilatometric tests under vacuum atmosphere were carried out in a Netzsch 402E/7 dilatometer to determine the A^sub 1^ temperature.


The microstructure of T42 + C after vacuum sintering and before the austenitising treatment consists of a lamellar pearlitic matrix where the well known primary blocky carbides MC and M^sub 6^C are embedded (Fig. 1).

Taking this microstructure as the starting point, it is possible to attain the desired hardness value of 50 HRC through several combinations of the austenitising temperature and the annealing temperature/time. The following cycles have been previously reported:6,7

2 Image (SEM) of T42+C annealed for 3 h (cycle 3) (secondary electrons)

(i) austenitising at 1000[degrees]C, isothermal annealing for 24 h at 625[degrees]C

(ii) austenitising at 975[degrees]C, isothermal annealing for 6 h at 650[degrees]C

(iii) austenitising at 950[degrees]C, isothermal annealing for 3 h at 660[degrees]C.

It has been demonstrated that the three cycles lead to qualitatively the same microstructure.6,7 The microstructure developed during the shortest annealing cycle is shown in Fig. 2. Apart from the primary blocky carbides, a fine distribution of equiaxial submicrometric precipitates can be observed.

A detailed TEM microstructural analysis has been carried out in order to determine accurately the size, composition and crystallography of those mentioned submicrometric precipitates. For the sake of clarity, only the results corresponding to the shortest cycle have been included but comments can be extended to the other annealing cycles. The size of the extracted precipitates varies from 200 to 25 nm (Fig. 3). Moreover, it has been determined that the chemical composition of the precipitates depends on the size of the analysed particle (Fig. 4): carbides between 100 and 200 nm are rich in Fe and Cr; while carbides ranging from 25 to 50 nm are rich in W and Fe.

The indexation of the diffraction patterns of the 100200 nm precipitates revealed that they possess an orthorhombic crystal structure. Taking into account their chemical composition and their crystalline lattice, these mentioned precipitates have been identified as chromium rich cementite. On the other hand, the composition of the 25-50 nm sized carbides is similar to that of the M^sub 6^C primary ones [3V-4Cr-28Fe-5Co-48W-12Mo (wt-%)]. Therefore it is reasonable to identify these nanometric carbides as M^sub 6^C. The intermediate size carbides, i.e. ranging from 50 to 100 nm, have been identified either as chromium rich cementite or M^sub 6^C. Consequently, it can be concluded that the austenite has transformed during the isothermal annealing cycle in equiaxial cementite and M^sub 6^C secondary carbides embedded in a ferrite matrix, i.e. M^sub 6^C+ globular pearlite.

In order to assess the development of the microstructure, the material austenitised at 950[degrees]C and annealed at 660[degrees]C for shorter periods of time (5, 30 min and 1 h) has been characterised. Figures 5-7 show the illustrative SEM images of the specimens heat treated under the above mentioned conditions.

3 Images (TEM) of carbon extraction replica taken at different magnifications (cycle 3)

4 Average chemical composition of extracted carbides as function of particle size: error bars indicate standard deviation

5 Image (SEM) of T42+C austenitised at 950[degrees]C and annealed at 660[degrees]C for 5 min (backscattered electrons)

6 Image (SEM) of T42+C austenitised at 950[degrees]C and annealed at 660[degrees]C for 30 min (backscattered electrons): M=martensite

7 Image (SEM) of T42+C austenitised at 950[degrees]C and annealed at 660[degrees]C for 1 h (secondary electrons): M=martensite

After annealing for 5 min (Fig. 5) the grain boundaries appear clearly visible, even though the specimen has not been etched. This means that in addition to the nanometric bright contrasted precipitation detected in Fig. 5 the eutectoid transformation of the austenite has already started (the nucleation of pearlite is a heterogeneous process taking place preferentially at high free energy sites, such as grain boundaries). During the subsequent water quenching the rest of the matrix transforms to martensite and consequently the hardness is 68 HRC. After annealing for 30 min (Fig. 6) the transformation is partially developed and the microstructure mainly consists of pearlite and martensite. It is worth remarking that the existing cementite is equiaxial. Again, the austenitic phase not transformed into pearlite during the isothermal annealing transformed into martensite during the subsequent cooling to room temperature. This explains the rather high hardness value of this specimen (57-5 HRC). After annealing for 1 h the transformation is almost completed as shown in Fig. 7. The microstructure mainly consists of globular pearlite while martensite is hardly detected. The small fraction of martensite is responsible for the increased hardness value (52 HRC) compared to that obtained after the 3 h cycle (50 HRC). So, although the initial, intermediate and final steps of the phase transformation have been characterised lamellar pearlite from the eutectoid transformation of the austenite, has not been detected in any case.


The eutectoid transformation temperature (A^sub 1^) of the T42 + C is 826[degrees]C as experimentally determined by dilatometry. This means that the austenite experiences a minimum undercooling of ~ 170[degrees]C for the highest of the annealing temperatures employed in the present work (660[degrees]C). The driving force for the transformation is large and consequently the eutectoid reaction tends to develop rapidly. Therefore, the system should adopt a fine lamellar configuration in order to minimise the carbon diffusion distances. However, from the above results it is demonstrated that these conditions lead to the formation of globular pearlite and even more, after Figs. 5-7, it can be concluded that the equiaxial cementite found after isothermal annealing does not originate from a lamellar pearlite through a spheroidisation process.

On the contrary, as shown in Fig. 1, the matrix of the as sintered specimens that suffered a continuous slow cooling inside the furnace consists of lamellar pearlite. Moreover, it has been observed that if these lamellar pearlitic specimens are held directly (without a prior austenitising treatment) for 3 h at 660[degrees]C there is no evidence of spheroidisation. This confirms the conclusion at the end of the former paragraph that cementite precipitates as spherical particles during the annealing cycles designed in the present work. The temperatures and times employed are not adequate to promote the spheroidisation of a lamellar configuration. In fact, previous works indicate that even a fine lamellar pearlitic microstructure needs a very long annealing cycle to spheroidise.9-11

Comparing the SEM images of Figs. 1b, 5 and 6, a homogeneous distribution of nanometre sized bright contrasted precipitates can be observed in the austenitised and isothermal annealed specimens. However, these precipitates are not detected in the continuously cooled specimens (as sintered samples), where pearlite adopts a lamellar configuration. Consequently, it is reasonable to assume a connection between the presence of those precipitates and the formation of globular pearlite.

8 Image (TEM) of carbon replica extracted from sample of T42+C austenitised at 950[degrees]C and annealed at 660[degrees]C for 5 min

Figure 8 shows a representative TEM image from a specimen annealed for 5 min (carbon replica). It can be seen that the size of the precipitates is in the range of 25100 nm.

Carbides ranging from 25 to 50 nm have a high W and Fe content and their composition is analogous to that of the same sized carbides detected in the specimen annealed during 3 h (Fig. 9). The largest fraction of the precipitates between 50 and 100 nm are also W and Fe rich carbides, while some of them have been identified as the first cementite particles nucleated during the 5 min isothermal annealing.

Consequently, the W rich carbides detected in the specimens where the pearlitic transformation has been completed (3 h) are analogous to those present at the beginning of the transformation (5 min). This indicates that the nanometric precipitation takes place before the eutectoid reaction begins, so it must occur during the fast cooling from the austenitising temperature to the annealing temperature. Since this type of precipitation precedes the eutectoid reaction, it has been termed proeutectoid precipitation.1,12 The indexation of the diffraction patterns obtained from these proeutectoid carbides (Fig. 10) revealed a cubic crystal structure. This fact together with the chemical composition similar to that of the primary M[degrees]C carbides confirms that proeutectoid carbides are M[degrees]C type.

9 Average chemical composition of carbides ranging from 25 to 50 nm as function of annealing time at 660[degrees]C: error bars indicate standard deviation

Assuming that proeutectoid carbides precipitate during the rapid cooling from the austenitising temperature to the annealing temperature, they will also precipitate during quenching to room temperature. Figure 11 shows the T42 + C microstructure after austenitising at 950[degrees]C and oil quenching. As expected, the matrix appears fully decorated by nanometric bright contrasted precipitates. It has been assessed by means of TEM observation and EDS analysis that those precipitates are M^sub 6^C. The presence of proeutectoid carbides in quenched HSSs has been reported elsewhere.12

Proposed mechanism

According to the previous experimental observations it is clear that:

(i) a rapid cooling promotes the precipitation of proeutectoid carbides

(ii) there exists a connection between the presence of proeutectoid carbides and the formation of globular pearlite. When pearlite adopts a lamellar configuration this type of carbides are not detected.

10 Diffraction patterns obtained from proeutectoid carbide and its corresponding indexation

11 Image (SEM) showing proeutectoid precipitation in T42+C austenitised at 950[degrees]C and oil quenched

12 Qualitative schematic representation on temperaturetime diagram for three types of thermal cycles discussed in present work

The different heat treatments carried out in the present work can be represented according to the schematic diagram shown in Fig. 12. The proeutectoid carbides precipitation line is plotted at a temperature below the minimum tested austenitising temperature (950[degrees]C) and above the maximum annealing temperature (660[degrees]C). On the other hand, since it occurs before the beginning of the pearlitic transformation, the line has been placed on the left of the C curves.

13 Mechanism of formation of globular pearlite when there is homogeneous distribution of intragranular nanometric proeutectoid M^sub 6^C carbides in austenite: proeutectoid carbides=black circles

At high cooling rates (during quenching or during the cooling step from the austenitising temperature in an isothermal annealing cycle) the proeutectoid carbides precipitate. However, a slow continuous cooling cycle can avoid that precipitation by only intersecting the C curves and not the proeutectoid carbides formation line. In that case, the eutectoid reaction takes place within an austenitic phase free from nanometric carbides, according to the well known mechanism of cooperative growth of ferrite and cementite leading to a lamellar configuration.13

On the other hand, if proeutectoid carbides precipitate, the austenite where the eutectoid transformation will take place has the aspect shown in Fig. 13a. As demonstrated in the present work, precipitation is profuse and particles are evenly distributed, so the mean free path between the nanometric carbides is nanometric as well.

The beginning of the transformation takes place according to the classical mechanism: first cementite (theta) nucleates at the austenitic grain boundaries (or at primary carbide/matrix interfaces). The austenite (gamma) beside the just formed nuclei is progressively depleted in carbon until the carbon content reaches a determined value and then ferrite (alpha) nucleates around cementite (Fig. 13b). Identically, the nucleation of ferrite causes carbon enrichment in the adjacent austenite promoting the formation of new cementite nuclei. However, the proeutectoid carbides embedded in the matrix limit the ferritic interface available for the nucleation process (Fig. 13c). In this sense, it is of great importance that the composition of the proeutectoid W rich carbides is very different from that of the cementite. If both compositions were similar the nanometric carbides would coalesce with the growing cementite and then for the undercooling of the present paper, the transformation of austenite into pearlite would lead to a lamellar configuration. Just below A^sub 1^ austenite would transform to cementite particles dispersed in a ferritic matrix by a divorced eutectoid transformation (DET).2,14,15

So, the high driving force involved in the isothermal annealing cycles developed at the present work promotes the formation of lamellar pearlite but the proeutectoid carbides prevent the cementite lamellae from growing. In other words, the available space for the growth of a cementite nucleus is limited by the adjacent proeutectoid carbides, which act as a physical barrier. The mean free path between the nanometric carbides is so small that cementite must adopt an equiaxial shape. The overall result is the cooperative growth of ferrite and equiaxial cementite. The reaction front advances from the grain boundary into the austenite (Fig. 13d), according to the observations derived from the micrographs shown in Fig. 6. Therefore, although globular pearlite is formed directly from austenite the process cannot be considered as a DET since the process is cooperative.


1. Rapid cooling from austenitising temperatures promotes the precipitation of W rich, nanometre sized M^sub 6^C proeutectoid carbides in the T42 HSS. The precipitation process takes place 660[degrees]C.

2. The proeutectoid carbides remain embedded in the steel matrix during the isothermal annealing treatment. They prevent the development of laminar pearlite because they act as a barrier for the growth of cementite lamellae by limiting the alpha/gamma interface available for the cementite nucleation. The overall result is the cooperative growth of globular pearlite directly from austenite at temperatures far below A^sub 1^ (i.e. in a highly undercooled austenite). 3. The precipitation of M^sub 6^C proeutectoid carbides (if abundant and evenly distributed in the austenite grains) allows to obtain fine grained globular microstructures by short duration annealing cycles.


The authors wish to thank Dr S. Gimenez for his inestimable collaboration. One of the authors (Dr V. Trabadelo) thanks the financial support provided by Torres Quevedo Programme.


1. G. Hoyle: 'High speed steels', 65; 1988, London, Butterworths.

2. G. Krauss: 'Steels: processing, structure and performance', 47; 2005, Materials Park, OH, ASM International.

3. In 'Metals handbook', 9th edn, Vol. 4, 'Heat treating'; 1981, Materials Park, OH, ASM International.

4. F. B. Pickering: in 'Materials science and technology', Vol. 7, 'Constitution and properties of steels', 152; 1992, Weinheim, VCH.

5. E. Pippel, J. Woltersdorf, G Pockl and G. Lichtenegger: Mater. Charact., 1999, 43, 41-55.

6. V. Trabadelo, C. Zubizarreta, S. Gimenez, I. Iturriza and F. Castro: Proc. PM2TEC 2002 World Cong., Orlando, FL, USA, June 2002, MPIF, 7-151.

7. V. Trabadelo, C. Zubizarreta, S. Gimenez, I. Iturriza, N. Khattab and E. Gordo: Proc. Eur. PM2003 Cong., Valencia, Spain, October 2003, EPMA, Vol. 1, 386.

8. S. Gimenez and I. Iturriza: Powder Metall, 2003, 46, 209-218.

9. Y. L. Tian and R. W. Kraft: Metall. Trans. A, 1987, 18A, 13591369.

10. V. V. Shkatov, A. P. Chemyshev and V. I. Lizunov: Phys. Met. Metallogr., 1990, 70, (4), 116-121.

11. O. E. Atasoy and S. Ozbilen: J. Mater. Sci., 1989, 24, 281- 287.

12. R. Wang, H. O. Andren, H. Wisell and G. L. Dunlop: Acta Metall. Mater., 1992, 40, (7), 1727-1738.

13. P. R. Howell: Mater. Charact., 1998, 40, 227-260.

14. T. Oyama, O. D. Sherby, J. Wadsworth and B. Walser: Scr. Metall., 1984, 18, 799-804.

15. J. D. Verhoeven and E. D. Gibson: Metall. Mater. Trans. A, 1998, 29A, 1181-1189.

V. Trabadelo1,2 and I. Iturriza*1

1 CEIT, P[degrees] Manuel Lardlzabal 15, 20018 San Sebastian, Spain

2 Now at Tekniker, Avda. Otaola 20. 20600 Eibar, Spain

* Corresponding author, email [email protected]

(c) 2008 Institute of Materials, Minerals and Mining

Published by Maney on behalf of the Institute

Received 20 February 2007; accepted 26 March 2007

DOI 10.1179/174328407X241061

Copyright Institute of Materials Jan 2008

(c) 2008 Materials Science and Technology; MST. Provided by ProQuest Information and Learning. All rights Reserved.