JOMINY TEST APPLIED TO PM STEELS FOR HEAT TREATMENT

The hardenability of steels, as defined by ASM,1,2 is usually measured as the distance below a quenched surface at which the metal shows a specific hardness or a specific percentage of martensite in the microstructure. To rate the hardenability, two methods can be used: the Grossman test and the Jominy test. The latter3–5 is the simpler and the more commonly used one. While for fully dense steels the literature data cover a wide variety of grades, for PM steels the available database seems to be mainly limited to rather recent years.6–12 As everybody knows, the hardenability of PM steels depends on:
(i) chemical composition, that can be uneven, depending on powder nature and processing conditions (ii) austenitic grain size mainly depending on sintering condition
(iii) cooling rate inside parts within a given range. This rate depends on material (chemical compositions, density and sintering intensity) and on part geometry and size.

The main differences between PM and fully dense materials are the presence of diffused porosity and, in many cases, the typical heterogeneity of chemical compositions.12–16 Owing to the growing interest in sinter hardening, the knowledge of hardenability of PM steels is becoming more and more important to establish the maximum part weight and typical geometrical features of parts suitable for sinter hardening. In some previous papers,17–20 the hardenabilities of PM steels – based on some ferrous base powders – have been determined, through the Jominy end quench test and hardness measurements, as suggested by the standard. Also the possibility of obtaining ferrous materials characterised by hardness values (200–250 HV) suitable for specific mechanical applications by means of a sinter–hardening process applied to low or medium alloyed steels has been investigated.13 In this paper the Jominy test has been applied to PM ferrous materials produced from low or medium alloyed powders, suitable for quenching heat treatment in the case of demanding technical requirements.

Materials and methods Two low alloy powders, produced by Ho¨ gana¨ s AB (Sweden) have been considered; their chemical compositions are reported in Table 1. These grades, produced also by other manufacturers, are frequently selected for producing PM parts requiring heat treatment. Prisms with square cross-section (65665 mm2 ) and 63 mm height have been compacted at 7. 0 g cm23 nominal density and sintered under endogas atmosphere at 1393 K for 30 min. in belt type industrial equipment. After sintering and physical characterisation, each prism has been cut into four equal parts and from each quarter, by turning, a cylinder has been machined. The dimensions of the cylinders are: 25 mm in diameter and 63 mm high. Then, on each cylinder, at the centre, a threaded hole has been machined. This hole has been made to allow the insertion of a threaded extremity part,

1. Hardenability bands of investigated steels

in order to observe the requirement of the standard, which imposes a total length equal to 100 mm.4,5 Other three holes, equally spaced, have been machined on the specimens to make possible the insertion of thermocouples, with extremities placed at different, precise distances from the quenched end. The instrumented Jominy probes have been placed inside a muffle furnace, set at 900uC, with an hour time at temperature. Since no controlled atmosphere was available, to avoid carbon depletion on the surface zones, the specimens were placed inside small cylinders equipped with a graphite disc at the bottom. In order to follow the temperature course, three thermocouples have been inserted inside the Jominy probes, placed at half of the base radius and at distances of 3, 13 and 23 mm from the quenched surface. During the cooling after austenitising, at the three selected points, the temperatures have been recorded through a digital acquisition system. The present investigation includes: 

(i) measurement of Rockwell A hardness 

(ii) measurement of Vickers 0. 05 microhardness 

(iii) optical (LOM) and electronic (SEM) microscopy (iv) EDXS analysis.

Experimental results
Hardenability bands and microhardness 

After cooling in the Jominy test device, each probe has been ground along four generatrices, arranged at 90u each other, on which hardness measurements have been made. From the hardness measurements (Rockwell A) Jominy curves have been obtained for both materials. The results of these measurements are reported in Fig. 1. Along one generatrix, selected at random, also the HV0. 05 microhardness profile has been determined, with 1 mm pitch. The results are reported in Fig. 2, where each point of the profile is the average value of three measurements at that distance

The hardenability bands, depicted in Fig. 1, indicate some difference in hardenability between the investigated steels. Anyhow, assuming 70 HRA as the lower limit for a ‘hard’ structure, for both materials the ‘design’ hardening depth is y5 mm. Figure 2, for the steel obtained from AE base powder, shows that proceeding from the quenched face, on the first 4 mm, microhardness profile exhibits high values (800–860 HV0. 05), then the microhardness decreases very quickly, dropping to 400 HV0. 05 in the next 3 mm. From this point on, the profile continues to drop, but softly and slowly, till 350 HV0. 05. On the same figure, the hardness profile of the material obtained from Astaloy Mo shows microhardness values between 760 and 780 in the first eight measurements (always starting from the quenched end). Then, in the distance interval from 8 to 20 mm, the microhardness values are nearly halved, dropping to y400 HV0. 05. After this point, the microhardness decreases slowly, to reach 320 HV0. 05 in the last measurements. The different statistical populations of microhardness values are shown in Figs. 3 and 4, on normal probability graphs. Each plot represents the results of 450 measurements. Figure 3, relevant to AE coded PM steel, shows the overlapping of at least four different statistical populations. Less than 10% of the values exceed.

650 HV0. 05; they are concentrated within ,15 mm from the water quenched end. The median is y350 HV0. 05. Figure 4, relevant to AM coded PM steel, shows the overlapping of at least three statistical different populations. About 25% of the values exceed 650 HV0. 05; they are concentrated within ,30 mm from the water quenched end. The median is y370 HV0. 05. 

Microstructures 

Typical microstructures have been observed at increasing distances from the water quenched end of each Jominy specimen. No significant differences have been observed among probes of the same chemical family.

AE coded material 

The extreme part of the probe nearly completely presents a martensitic structure with some zones of transforming austenite (Figs. 5 and 6). The white areas observed at the optical microscope immediately below the quenched face are seen to be completely transformed when observed with the SEM (Fig. 7). Almost immediately, beyond this area, the microstructure suddenly changes and appears to be made up of very fine pearlite and austenite transforming in acicular structures (Figs. 8 and 9). Then, going on along the sample, very fine pearlite predominates with small quantity of upper bainite and austenite at different degrees of transformation (Figs. 10–13).

Then, at .5 mm distance from the quenched end, lower bainite becomes the predominant structure (Fig. 17). At y18 mm from the quenched face the presence of upper bainite becomes remarkable, intermixed with only some areas of lower bainite (Figs. 18–20). The areas white at LOM (Fig. 18), when observed at SEM, appear completely transformed in upper bainite (Fig. 21). Furthermore, at 20 mm distance, the structure is prevalently constituted by upper bainite, with the presence of pearlitic structures (Figs. 22–24). For the steels based on diffusion bonded powders (materials coded AE), EDXS analysis pointed out

uneven compositions as expected. The range of local Ni amount (wt-%) has been between 2 and 8% as usual at 1393 K sintering temperature with industrially acceptable times. Concluding remarks The results of this investigation confirm the good hardenability of both PM steels, but for limited distances from the quenched end. The hardenability bands indicate that the typical hardening depth is y5 mm for both materials, After this distance, the change towards non-martensitic structures is sudden for the steel from diffusion bonded powder and more soft for the steel based on Mo prealloyed powder. It should be observed that the hardenability bands are based on

HRA hardness instead of HRC hardness, as prescribed by the standard and customarily used for fully dense steels. At first sight, the use of a different scale could be criticised. It can mostly be observed, however, that when reducing the test load from 150 to 60 kp the probability of overtaking the border of reliable hardness measurements notably decreases. As we know, any Rockwell value is obtained by the difference between two positions of the indenter. The lower the difference between position, the higher the error of measurement. For this reason, any Rockwell value below 30/35 is affected by a measurement error that can alter the ‘true’ properties of the investigated materials, independently from any other variable. According to a well consolidated experience, 70 HRA corresponds to 39. 25 HRC.21 If the weakening effect of y11% porosity is duly considered, it should be unquestionable that 39 HRC on such porous steel can be achieved only if the microstructure is fully martensitic. So, the choice of the hardenability limit is based on these considerations. The data obtained from the experiments show that:

(i) as expected, the Jominy test is apt to rate the hardenability of PM steels

(ii) the hardenability bands and the microhardness profiles agree 

(iii) the microstructure features and the microhardness distributions agree 

(iv) the AE coded material shows high microhardness values (800–860 HV0. 05) only on the first 4 mm from the water quenched end. From this distance on, the microhardness decreases very quickly to 400 HV0. 05 at 7 mm distance. After this point the microhardness profile presents a slow drop to y300 HV0. 05 at .50 mm distance, i.e. at the centre of the standard Jominy specimen 

(v) near the water quenched end the AM coded material reaches lower microhardness values (760–780 HV0. 05) than the AE coded material, considering the same zone. However these values remain constant at y8 mm. Beyond this point the microhardness begins to decrease slowly, reaching 400 HV0. 05 at 20 mm from the water quenched end 

(vi) the AE coded material, at .5 mm distance from the quenched end, presents some austenite areas surrounding the pores. These areas are typically Ni rich, so that they prevent or obstruct the formation of martensite. The location of these

areas depends on the powder nature, where the diffusion–bonding process fixed very fine nickel particles on the surface of coarser iron at random. During sintering, the nickel facing the pores after compaction can diffuse only in one direction, while, again during sintering, the nickel situated between contacting particles after compaction can diffuse in two directions. In the first case, the concentration gradient will be inevitably more pronounced than in the second case. It is well known that the Ni rich austenite areas are definitely positive in the case of fatigue stresses acting in operation. On the contrary, their hardness and the corresponding wear resistance are rather poor. This latter remark enables to propose a distinction between the two materials. The distinction is based on the difference between the pattern of applied load. At equal hardness, in the case of prevailing and continuous wear stress, the PM steels based on diffusion bonded materials can be less resistant than the PM steels based on prealloyed powders. The contrary should be the operating behaviour in the case of combined stresses, including fatigue. This difference agrees with the common selection criteria followed by PM part manufacturers and users