3The effect of nickel on the mechanical behavior of
4molybdenum P/M steels
5
B.A. Gething a , D.F. Heaney a ,∗ , D.A. Koss a ,b , T.J. Mueller a
6a The Center for Innovative Sintered Products, 118 Research Building West, 7The Pennsylvania State University, University Park, PA 16802, USA 8b Department of Materials Science and Engineering, The Pennsylvania State University, University Park, PA, USA
Received 15 May 2004; received in revised form 23 May 2004
9
10Abstract
This study has examined the effects of nickel alloying additions on the microstructural characteristics and mechanical properties of Fe–xNi–0.85Mo–0.4C-base steels that were powder processed using double-press double-sinter processing to maximize density. The steels were examined in the as-processed condition as well as in a quench-and-temper heat treated condition. Tensile behavior indicates that while nickel content (at levels of 2,4, and 6%) increased tensile strength in the as-sintered condition, it did not significantly affect tensile strength in the quenched and tempered condition. In both conditions increasing Ni content decreased elongation to fracture. The 4% Ni steel, which tended to have the smallest maximum pore size, also exhibited the greatest fatigue strength. © 2004 Published by Elsevier B.V.
4200, 4400 and 4600 series steels, which rely on nickel, car-
For a powder metallurgy (P/M) component to be consid-bon, and molybdenum additions. In most cases, the nickel and ered for high performance applications, porosity must be re-the carbon are added as elemental powder additions, whereas duced to a minimum; i.e. density must be maximized. Many the molybdenum is pre-alloyed with the base iron powder. methods to reduce porosity have been explored, such as the Due to the industrial importance of these PM steels, many hydropulsor technique [1], warm compaction [2], and high studies have addressed aspects of their processing and behavtemperature sintering [3]. One of the more common meth- ior [2–13]. However, each study has used a slightly different ods to reduce porosity is known as “double-press double-alloy or processing method, and comparisons as well as stan-sintering” (DPDS), which decreases bulk porosity of conven-dardized properties are difficult to ascertain. For example, tional ferrous-based components from approximately 10% while it is known that strength increases with an increase in after the conventional single-press single-sinter (SPSS) to nickel content [14], limited elongation data exists. Engstrom 5% porosity after DPDS [4,5]. and Allroth reported that elongation moderately increases
Another method to improve the mechanical properties of with nickel content, but at nickel contents greater than 6%, P/M materials is through the use of alloy additions. These ad-which are of lesser interest today [15]. Similarly, Morioka ditions can be pre-alloyed with the iron prior to powder pro- [7,8] e xamined fatigue strength for 2 and 4% nickel alloys, duction or added as an elemental powder addition to the base but each of these alloys had different molybdenum contents iron powder. For example, a “workhorse” series of alloys for and the 4% nickel alloy also contained copper additions. No
study has determined tensile and fatigue behavior as a func-
Corresponding author. Tel.: +1 814 865 7346; fax: +1 814 863 8211 tion of nickel content while maintaining constant process and E-mail address: dfh100@psu.edu (D.F. Heaney). alloy parameters.
0921-5093/$ – see front matter © 2004 Published by Elsevier B.V. doi:10.1016/j.msea.2004.05.087
UNCORRECTED PROOF
18Keywords: Molybdenum/nickel steels; Pressed and sintered steels; High strength; Double-press double-sintered steels; Tensile strength; Fatigue strength
19
1.Introductionmoderate-to high-stress P/M applications are the P/M grades
∗
MSA 18426 1–8
2 B.A. Gething et al. / Materials Science and Engineering A xxx (2004) xxx–xxx The objective of this work is to examine the effect of nickel Standard 40 geometry, with a thickness of approximately
58111
additions on the tensile and fatigue behavior of molybdenum 0.45 cm.
59112
P/M steels. High density processing was applied through All samples were initially pressed at 550 MPa and subse
60113
DPDS to reduce the effects of porosity, thereby increasing quently pre-sintered in an Inconel Lindberg retort furnace by
61114
performance and making the effects of the alloy additions being heated to 450 ◦Cat5 ◦C/min and held for 180 min for
62115
more prominent. This study examines the influences of the lubricant burnout, then being heated to 750 ◦Cat10 ◦C/min
63116
nickel additions on porosity, microstructure, hardness, tensile and held for 30 min in flowing 10% hydrogen/90% nitrogen
64117
strength, and elongation to fracture in both the as-sintered atmosphere to pre-sinter and anneal the samples, and then fur
65118
condition and a quenched-and-tempered heat-treated condi-nace cooled at a rate of approximately 15 ◦C/min. All samples
66119
tion; fatigue strength at 106 cycles was also determined for were then pressed a second time at 825 MPa. A resizing die
6712068the “heat-treated” condition. was used for the second pressing of the tensile specimens. The
121
opening of the die was larger than the initial die by 0.127 mm
122
in all directions, allowing for easier part entry to the die.
123
2.ExperimentalprocedureThe walls were tapered by the 0.127 mm over a distance of
69124
1.27 cm so that the inside of the die was the same size as the 125Samples were fabricated from water-atomized alloy pow-initial die. Before the samples were pressed a second time, 7012671der and elemental additions of nickel and carbon. If nickel they were lightly sprayed with zinc stearate to reduce friction
127
and carbon are pre-alloyed with the base iron material, the effects. High temperature sintering was performed in a CM
72128
compressibility is reduced resulting in lower overall density, 333 pusher furnace at a temperature of 1288 ◦C and a time of while the molybdenum has little effect on the compressibility 60 min. A sintering atmosphere of 50% H2 and 50% N2 was of the base iron powder. The base iron powder was Atomet used. Specimens were air cooled in the furnace cooling zone 4401, supplied by Quebec Metal Powders Limited. Atomet at an undetermined cooling rate. 4401 is water-atomized, pre-alloyed iron powder consisting After the double-press double-sinter processing, half of of 0.8% Mo and 0.16% Mn on a weight basis. The iron pow-the samples from the study were heat treated by austenitizing der had a rounded but irregular shape with a mean powder at 870 ◦C (1600 F) at a 0.7% C potential in an endothermic at-particle size of 37 �m. The elemental nickel powder, No-mosphere for 60 min and then quenched in oil. These samples vamet Type 123, had a mean particle size of 8 �m and was were subsequently tempered for 60 min at 390 ◦CinanN2 spherical and sponge-like in appearance. atmosphere. This temperature represents the upper range of
Carbon from Asbury Graphite Mills Inc., (PM 5) was temperatures used to temper high-strength powder-processed added in the form of graphite. The initial admixed carbon steels [3–12,16] inorder to ensure that the specimens regained amount was 0.46 wt.% which resulted in a nominal post-the maximum amount of tensile ductility. sintered composition of 0.40%. This reduction of 0.06% dur-The Archimedes technique was used to measure the dening sintering is believed to result from the carbon reduction sity of all samples, while the percent porosity was measured of the oxide layer on the Atomet 4401 powder, and thus, the using an AccuPyc 1330 pycnometer. Using this technique, carbon content identified for the alloys in this study refers samples were compacted and pre-sintered, then crushed into to post-sintered carbon content. Typically, prior studies often small fragments. These fragments were then measured for do not differentiate from pre-and post-sintered carbon con-density using the pycnometer; this density was utilized as the tent. It will be assumed that a pre-sintered carbon content of theoretical density. Comparing this value to the Archimedes 0.5% in prior studies is comparable to the post-sintered 0.4% values yielded the percent porosity. carbon content in the present study, as some decarburization Hardness measurements were taken using an LR-Series during the sintering process is likely. Rockwell-Type hardness indenter. Both Rockwell “B” and
A lubricant, Acrawax “C,” in the amount of 0.5 wt.% was “C” scale measurements were taken. Microhardness of indialso added to the alloy mixtures. A total of 100 g was mixed vidual steel phases was measured using a Leco M-400H-0 for each alloy in a Glen Mills Inc., turbula mixer for 30 min, testing apparatus using a Vickers diamond indenter. The ten-with the rotation speed set at 50 rpm. sile strengths were measured using an Instron 4100 tensile
Two different geometries of samples for mechanical tester, at a strain rate of 10−2 testing were used for the study. Standard P/M dog-bone Fatigue strength at 106 cycles was measured using the
were powder processed to the MPIF Stan-four-point-bend test fixture described above [17]. The fixture geometry, with a thickness of approximately was equipped with pins that allow for two directions of ar-After sintering, the section for quenched and ticulation to minimize stress concentrations during testing.
tempered specimens was reduced by machining to ap-Fatigue testing was performed using a MTS 1036 Materials
proximately half the original width. This modification in-Testing machine under load control with a stress ratio (R) stress in the gage section to ensure frac-of 0.1. All samples were tested at a frequency of 30 Hz at gauge section and to eliminate grip slip-ambient conditions. All specimens were surface conditioned was a problem in the unmodified samples. to remove sharpness and imperfections along edges, mini-
Four-point-bend fatigue bars were fabricated to the MPIF
UNCORRECTED PROOF
s.
tensile bars
dard 10
0.65 cm.
creased the
ture in the
page, which mizing stress risers that might cause inconsistent mechanical MSA 18426 1–8
B.A. Gething et al. / Materials Science and Engineering A xxx (2004) xxx–xxx
Fig. 1. Optical micrographs of as-sintered Fe–xNi–0.85Mo–0.4C steel containing (a) 2% Ni, (b) 4% Ni, and (c) 6% Ni. The microstructure designations are as follows: UB—upper bainite, F—pro-eutectoid ferrite, M—martensite, LB—lower bainite, and NR—Ni-rich retained austenite.
performance. Samples were placed in a vibratory polisher for B ni − A2 10 h. The polisher operates by vibrating samples and milling s = 1.62d � (ni)2 media together in a bell-shaped bowl. As these objects vibrate, they tumble onto one another, smoothing the surface In the above equations, S0 is the lowest stress level during a
particular staircase test, and d the interval spacing or differ-
In order to determine the fatigue strength of materials at ence of stress between two steps of the staircase. According 106 cycles, a staircase method was employed [18]. For this to Lipson and Sheth, the value for d should not exceed 10% procedure, the fatigue strength was first estimated based on of S0. A and B are parameters based on specimen responses. prior results by Morioka [8], and the first test was run at this The stress levels are given coded scores i, where i= 0 for stress amplitude. If the specimen did not fail at this stress the lowest stress level, and ni, is the number of failures (or level, the next test was run at a higher level, and the converse survivals) for a given staircase test. In Eq. (1) “ +1/2”is used was performed if the specimen failed. The test procedure con-if survivals are less frequent and “−1/2” if survivals are more tinued in this fashion with stress levels being raised and low-frequent. ered for sequential test specimens. Increments of 20 MPa, approximately 5% of the estimated fatigue strength, were used for all staircase tests, and approximately 15 specimens were 3.Resultsanddiscussiontested for each condition to determine the average fatigue strength at 106 cycles in a statistically reliable manner. The 3.1. Microstructural characteristics average fatigue strength (S) was determined by using only the failure or survival specimens, depending on which had The microstructures of several alloys were examined to the smaller total. The mean fatigue strength and standard de-determine the phases present in both the as-sintered and viation (s) were calculated by using the following equations heat-treated conditions. In the as-sintered condition shown from Lipson and Sheth [18]. in Fig. 1a .the microstructure of the 2% Ni steel is primar-
� ily upper bainite (UB) (Vickers hardness value, HVN = 209) ± 1 (1) with small, discontinuous regions of randomly dispersed, soft 2
UNCORRECTED PROOF
+ 0.029 (2)
and edges.
¯
A
190S = S0 + �
(HVN = 139) regions believed to be pro-eutectoid ferrite (F). MSA 18426 1–8
ni
B.A. Gething et al. / Materials Science and Engineering A xxx (2004) xxx–xxx
Fig. 2. Optical micrographs of quenched and tempered (a) 2% Ni alloy, (b) 4% Ni alloy, and (c) 6% Ni alloy. NR denotes regions that are primarily Ni-rich
The upper bainite regions are comparatively soft when com-hard martensite and bainite are more prevalent in the 6% Ni pared to previous results [19]; this softening could be a result alloy. of tempering of the upper bainite region during the slow cool-After the quench-and-temper heat treatment, the miing process after sintering. crostructures of all
The microstructures of the 4% Ni and 6% Ni alloys in tempered martensite the as-sintered conditions, Fig. 1b and c, are similar to those tained austenite; see observed by Kosko in a PM steel with similar composition austenite phase increases with increasing Ni content, [5], although no lamellar pearlite regions were observed in expected. the present case. The microstructure also includes regions The average porosity values were determined from five of upper bainite (UB) (HVN = 248) as well as lower bai-replicates of each alloy, and nite (LB) (HVN = 370) and areas of a needle-like, very hard Tables 1 and 2. Both the green and DPDS porosity levels (HVN = 575) martensite (M) phase. All three of these con-were the highest for the 6% Ni alloy, lowest for the 4% Ni, stituents appear to form in random locations. In addition, a and intermediate for the 2% Ni alloy. One would expect a soft (HVN = 180), nickel-rich austenite (NR) is also present; decrease in porosity of the alloy as the nickel content is in-it contains many secondary pores that were formed when the creased since the nickel powder is finer and would be more
prone to sinter densification. On the other hand, densification
Green and sintered densities of (Fe–xNi) alloys investigated in this study (error is standard deviation)
| Alloy Green porosity (%) | DPDS porosity (%) Densification | |
|---|---|---|
| (Fe–2Ni)–0.85Mo–0.4C | 10.62 ± 0.05 | 3.57 ± 0.01 7.05 ± 0.06 |
| (Fe–4Ni)–0.85Mo–0.4C | 10.93 ± 0.06 | 3.24 ± 0.02 7.69 ± 0.08 |
| (Fe–6Ni)–0.85Mo–0.4C | 11.62 ± 0.05 | 3.79 ± 0.02 7.83 ± 0.07 |
retained austenite.
three steels consist primarily of
with minor levels of Ni-rich re-
Fig. 2. The amount of retained
as
the results are shown in
nickel particles dissolved in the iron matrix. The regions of
Table 1
MSA 18426 1–8
B.A. Gething et al. / Materials Science and Engineering A xxx (2004) xxx–xxx
Table 2 Mechanical properties of Fe–xNi steels investigated in this study (error is standard deviation)
| Alloy | Hardness | Tensile strength (MPa) | Elongation (%) | Fatigue strength at |
| 106 cycles | ||||
| As-sintered properties | ||||
| (Fe–2Ni)–0.85Mo–0.4C | 81±2 (HRB) | 642 ± 8 | 8.7 ± 0.5 | |
| (Fe–4Ni)–0.85Mo–0.4C | 99±3 (HRB) | 760 ± 24 | 6.2 ± 0.7 | |
| (Fe–6Ni)–0.85Mo–0.4C | 108±3 (HRB) | 943 ± 21 | 5.0 ± 0.9 | |
| Quenched-and-tempered properties | ||||
| (Fe–2Ni)–0.85Mo–0.4C | 48 ± 1 (HRC) | 1530 ± 14 | 1.9 ±0.1 | 373 ± 6 |
| (Fe–4Ni)–0.85Mo–0.4C | 48 ±1 (HRC) | 1551 ± 27 | 0.9 ± 0.2 | 396 ± 26 |
| (Fe–4Ni)–0.85Mo–0.4C | 49 ± 1 (HRC) | 1579 ± 25 | 0.8 ± 0.2 | 361 ± 19 |
| 242 | of the cold-pressed compact during DPDS decreases with in- | Ni content results in a subsequent decrease in elongation for | 256 |
| 243244 | creasing amounts of Ni, as also observed elsewhere [20]. Thus, steels with low Ni additions have high compressibility | these alloys; tensile elongation is highest at 2% Ni (8.7%), and decreases to 5.0% at 6% Ni content. | 257258 |
| 245 | but low densification, and high-Ni-content steels have low | In contrast to the behavior in the as-sintered condition, | 259 |
| 246247 | compressibility with high densification, as shown in Table 2. The result is the 4% Ni alloy exhibiting the lowest amount of | Ni content has only a minor effect on the hardness and tensile strength in the quenched and tempered condition with | 260261 |
| 248 | final porosity, possessing both adequate compressibility and | all three alloys having tensile strength of ≈1550 MPa; see | 262 |
| 249 | densification. | Table 2. The 2% Ni alloy exhibits the highest elongation at 1.9%, while the 4 and 6% Ni alloys exhibit elongations of 0.9 | 263264 |
| 250 | 3.2. Mechanical properties: tensile behavior | and 0.8%, respectively, for the heat-treated condition. A comparison of these results with those previously ob | 265266 |
| 251252253254 | As indicated in Table 2, increasing Ni content causes an increase in both the hardness and tensile strength in the as-sintered condition. For example, tensile strength increases from 642 to 943 MPa as Ni content increases from 2% to 6%. | served for similar PM steels in the quenched and tempered condition is shown in Table 3. Prior studies often do not differentiate between pre-and post-sintered carbon content, and it will be assumed that a pre-sintered carbon content of 0.5% in | 267268269270 |
| 255 | The increasing strength and hardness of alloys with higher | prior studies is comparable to the post-sintered 0.4% carbon | |
| A comparison of the present results with those obtained in prior studies Table 3 | |||
| Density (g/cm3) Tensile strength (MPa) Study (alloy) | Elongation (%) Bending fatigue strength at 106 cycles (MPa) Processing and chemistry comments | ||
| Present (Fe–2Ni)–0.85Mo–0.4C 7.50 1530 (Fe–1.8Ni)–0.5Mo–0.5C 7.20 1300 (Fe–0.5Ni)–0.5Mo–0.5C 7.47 1711 (Fe–2Ni)–1.0Mo–0.5C 7.50 1946 (Fe–2Ni)–1.0Mo–0.6C 7.48 2000 approximately (Fe–2Ni)–1.0Mo–0.6C 7.43 1920 Narasimhan: [2] Hanada: [7] Hanada: [7] Morioka: [7] Furukimi: [6] UNCORRECTED PROOF | 1.9 373 N/A N/A N/A N/A N/A N/A N/A 420 approximately 106 cycles N/A N/A DPDS Lower density Pre-alloyed powder Diffusion bonded powder Diffusion bonded powder–higher C content Diffusion bonded powder–higher C content | ||
| Present (Fe–4Ni)–0.85Mo–0.4C 7.54 1551 (Fe–4Ni)–1.5Cu–0.5Mo–0.5C 7.50 1203 (Fe–4Ni–1.5Cu–0.5Mo–O.6C 7.41 1700 approximately Hanada: [7] Morioka: [8] | 0.9 395 N/A N/A N/A 350 approximately 106 cycles DPDS Cu addition Cu addition–higher C content | ||
| Present (Fe–6Ni)–0.85Mo–0.4C 7.55 1579 (Fe–6Ni)–0.85Mo–0.5C 7.42 1345 (Fe–6Ni)–1.5Mo–0.5C 7.42 1400 Narasimhan: [2] Narasimhan: [2] | 0.8 361 N/A N/A N/A N/A DPDS Warm compaction–lower density: Warm compaction–lower density: | ||
| All results are in the quench-and-tempered condition, however, temper temperatcompared to the present study that could significantly affect the results. | uresmay vary. The last column in the table identifies a difference in processing | ||
| MSA 18426 1–8 |
B.A. Gething et al. / Materials Science and Engineering A xxx (2004) xxx–xxx
Fig. 3. SEM micrographs of fatigue specimen fracture surface, showing (a) ductile tensile fracture and (b) fatigue fracture regions. Porosity is indicated with
content in the present study, as some decarburization during terms of alloying, because admixed Ni additions were made the sintering process is likely. Compared to previous studies to the alloy, as was the case for the present study. In general, at similar density levels [2,8,13], the 2% Ni alloy tested in the results of this study are similar to previous results of simi-this study had a lower strength. The primary reason for the lar studies; therefore, the fatigue data of the following section lower strength was the use of elemental, admixed Ni as ap-should also be representative of this class of P/M steels. posed to pre-alloyed or diffusion bonded Ni, resulting in soft (and larger) Ni-rich regions. 3.3. Mechanical properties: fatigue behavior
Another possible reason for the lower strength of this study is that the temper temperature of this study was 390 ◦C, which Fatigue tests were performed on the three alloys in the is the highest accepted tempering temperature for these alloys quenched and tempered conditions, and the results of the and is higher than most of the other studies. The 4% Ni alloy staircase tests appear in Table 2. exhibits a strength that is mid-range in value when compared of fatigue strength to tensile strength was in the range of to previous studies. For example, the strength results of this 0.23–0.25. The 4% Ni alloy shows the highest average fa-study were higher in comparison to Hanada et al. [7] b ecausetigue strength of approximately 395 MPa, while the 2% Ni Cu additions tend to decrease the “heat-treated” strength [6]. alloy exhibits intermediate fatigue strength of approximately On the other hand, tensile strengths attained in this study 373 MPa. The 6% Ni alloy, which had the highest tensile were lower than those of Furukimi et al. at 2% Ni [6]. Cu strength and porosity level ( Table 2), exhibited the lowest fa- additions were also present in the study of Furukimi and co-tigue strength of 361 MPa. This behavior is consistent with workers, but the carbon content was higher and the temper other studies that show porosity in P/M alloys decreases fa-temperature was lower, which can result in higher strengths. tigue strength significantly [21–23]. It should be noted that, For the 6% Ni alloy, no direct comparisons for DPDS are given the scale of the standard deviations of the fatigue available, but other high density processing (i.e. warm com-strengths, the differences between the fatigue strengths of paction) was performed on 6% nickel steels [2]. The 6% Ni the alloys are small. alloy of this study displayed a higher strength in comparison The present fatigue results may be compared with those of to the Narasimhan study [2], mainly due to higher density. Morioka, who examined a 2% Ni alloy with 0.6% C (as op-The Narasimhan study offers a direct basis of comparison in
UNCORRECTED PROOF
dotted arrows.
In all cases, the ratio
posed to 0.4% C here) and processed using diffusion bonded MSA 18426 1–8
B.A. Gething et al. / Materials Science and Engineering A xxx (2004) xxx–xxx
Table 4 Analysis of the projected pore content on the fracture surface of (Fe–2Ni) P/M steel, where both fatigue and ductile fracture modes were observed
Fracture surface Approximate crack length (mm) Porosity on fracture surface (%) Ratio to bulk porosity
Fatigue crack region 0.5 11.4 ± 1.2 2.4 ± 0.3 Ductile fracture region 3.0 33.7 ± 0.6 7.2 ± 0.2
321
powders [8], rather than elemental Ni in our case. As shown
322
in Table 3, while the 2% Ni alloy of this study has a lower
323
fatigue strength, it exhibits a slightly higher fatigue strength
324
to tensile strength ratio, 0.24, than that of Morioka who ob
325
tained a ratio of 0.21. For the 4% Ni alloy, the fatigue strength
326
was higher than that found by Morioka. As with the 2% Ni
327
alloy, the tensile strength to fatigue strength ratio was also
328
higher, 0.25 compared to 0.21, respectively.
329
3.4. On the effects of porosity on tensile and fatigue
330
properties
pore microstructure. In a tensile test, plasticity within the en
368
tire specimen gauge volume is the driving force; the eventual
369
fracture path is one of high pore content and is tortuous and
370
zigzagged in nature [26]. For crack growth processes, plastic
371
ity and damage is confined to the crack-tip plastic zone. For
372
crack lengths in Table 4 and correcting for the differences
373
between cyclic and monotonic plastic zone sizes, the plastic
374
zone, rp, will be approximately 24 times larger for the tensile
375
fracture stage of crack growth than for the fatigue fracture
376
portion [27]. Since damage is confined to a plane ± rp of the
377
main crack, only a comparatively small number of pores can
378
participate in the fatigue crack growth process. In contrast, In this study, porosity has shown to decrease both ten-the tensile overload region, rp, is large and pore participation
sile and fatigue strengths of the heat-treated P/M steels, as in the tensile crack growth process is about 3 times greater expected. To explore the influence of porosity on these dif-( Table 4). As a result, the fatigue fracture surface is comparferent modes of fracture, the fracture surfaces were examined atively flat, as fewer pores cause crack deviation, and it also
to determine the relative levels of porosity on the fatigue and contains a lower fraction of porosity. tensile fracture surfaces. As shown in Fig. 3, the fracture sur- The effect of pore size on the fatigue strength was also faces of the quenched and tempered fatigue specimens were investigated in this study. An analysis of the distribution of analyzed for projected area fraction of porosity, and the re-pore sizes shows that the 4% Ni steel, which had the highsults are shown in Table 4. est fatigue strength ( Table 2), also had the smallest percent- Quantification of porosity was performed via image anal-age of pores larger than either 10 �mor20 �m. Thus, as
ysis by means similar to Danninger et al. [24], although frac- has been observed in the influence of porosity on fatigue of ture was not performed at low temperatures which would have castings, large pores can significantly reduce high cycle fa-prevented ductile deformation. Despite this fact, the sintered tigue strength [28]. With the smallest fraction of large pores,
compacts showed little ductility and any miscalculation due the 4% Ni alloy will have the smallest likelihood to initito deformation can be reflected in the error reported with ate a long fatigue crack, consistent with its highest fatigue each value. The Dannigner study focused on the effective strength. load bearing cross-section (Ac) of the fracture surface, which is a parameter used to predict tensile and fatigue strength of sintered compacts. The present study focuses on the pore 4.Conclusionscontent of the fracture surface and how it changes with crack growth phenomena. Surface porosity and Ac are direct oppo-The influence of Ni content (2,4, and 6 wt.% levels) on the microstructure and mechanical properties has been ex-The amount of porosity on the ductile tensile fracture sur-amined in a series of Fe–xNi–0.85Mo–0.4C steels fabricated face is ≈7.2 times that of the bulk porosity or roughly one-by powder metallurgy using elemental Ni additions, rely-third of the fracture surface as viewed in Fig. 3. In contrast, the ing on a double-press double-sinter processing route to in-area fraction of porosity on the fatigue fracture surface was crease density. The results indicate that in the as-sintered much lower at ≈2.4 times the bulk porosity. The implication condition increasing nickel content increases tensile strength is that crack growth associated with tensile fracture follows and decreases elongation to fracture as a result of increased a rough, irregular plane of high pore content. In contrast, fa-levels of hard constituents (martensite and lower bainite) tigue crack growth is much less dependent on porosity, which within the microstructure. In contrast, after the quench-is consistent with earlier observations that porosity has only a and-temper heat treatment, Ni additions have no significant small effect on fatigue crack propagation rates [25]. Thus, the effect on tensile strength but decrease elongation slightly fracture surface of the fatigue crack region is less dependent for the tempered martensitic structures. Fatigue results for on porosity and contains a smaller number of pores. quenched and tempered conditions indicate that the 4% Ni The higher degree of pore participation in tensile crack steel has the highest fatigue strength at 106 cycles, consisgrowth can be understood in terms of the relative scales of tent with the lowest level of large (>20 �m) pores in this the crack-tip plastic zones when compared to the scale of the case.
UNCORRECTED PROOF
sites of each other.
MSA 18426 1–8
B.A. Gething et al. / Materials Science and Engineering A xxx (2004) xxx–xxx
415The authors wish to thank The Center for Innovative Sin-
[12] W.B. James, M.C. Baran, F.J. Semel, R.J. Causton, K.S. Narasimhan, 442
[13] A. Graham, T. Cimino, A. Rawlings, H. Rutz, Adv. Powder Metall. 444 416 tered Products for supporting this effort under BFTDA Grant Part Mater. 13 (1997) 75. 445
420 Metal Powders Limited are also acknowledged. [16] W. Jamesin, W.B. Eisen, B.L. Ferguson, R.M. German, R. Iacocca, 449P.W. Lee, D. Madan, K. Moyer, H. Sanderow, Y. Trudel (Eds.), 450 ASM Handbook: Powder Metal Technologies and Applications, vol. 451 07, Materials Park, OH, 1998. p. 947. 452
Park, PA, 2003. 454
[5] J. Kosko, in: W.B. Eisen, B.L. Ferguson, R.M. German, R. Iacocca, FL, June, 2002.
Lee, D. Madan, K. Moyer, H. Sanderow, Y. Trudel (Eds.), [22] T. Cimino, A. Graham, T. Murphy, Adv. Powder Metall. Part Mater.
ASM Handbook: Powder Metal Technologies and Applications, vol. 13 (1998) 33. 07, Materials Park, OH, 1998. p. 751. [23] T. Cimino, A. Graham, T. Murphy, A. Lawley, Adv. Powder Metall.
[6] O. Furukimi, K. Yano, S. Takajo, Adv. Powder Metall. 5 (1991) Part Mater. 7 (1999) 65.
[24] H. Danninger, U. Sonntag, B. Kuhnert, R. Ratzi, Prakt. Metallogr.
[7] M. Hanada, N. Motooka, T. Honda, Adv. Powder Metall. Part Mater. 39 (2002) 8.
[25] R.A. Queeney, P.S. Dasgupta, Int. J. Fatigue 2 (1980) 113.
[8] Y. Morioka, Met. Powder Rep. 45 (1990) 181. [26] H.J. Niu, I.T. Chang, Scr. Metall. 41 (1999) 481.
[9] F. Hanejko, H. Rutz, U. Engstrom, B. Johansson, Adv. Powder Met- [27] S. Suresh, Fatigue of Materials, Cambridge Press, Cambridge, UK, all. Part Mater. 10 (1995) 77. 1991.
[10] H. Khorsand, H. Yoozhashizade, S.M. Habibi, K. Janghorban, A. [28] Q.G. Wang, P.N. Crepeau, D. Gloria, S. Valtierra, Adv. Aluminum Nangir, S. Reihani, Met. Powder Rep. 57 (2002) 32. Cast. Tech. II (2002) 209.
UNCORRECTED PROOF
43459.
5 (1992) 215.
MSA 18426 1–8