Download - sintering stainless steel

7/30/2019 sintering stainless steel

1/42

THE PRIMARY GOAL in stainless steel sinteringis to obtain good corrosion resistance along withgood mechanical properties and adequate dimen-sional tolerances. Most aspects of sintering havea bearing on corrosion resistance; therefore, inthe following, sintering is discussed with an

emphasis on its effect on corrosion resistance.In wrought stainless steels, superior corrosionresistance is of paramount importance, becausemechanical properties similar to and evensuperior to stainless steels can be obtainedmuch less expensively with conventional car-bon steels. However, over several decades,despite modest corrosion resistances, commercialsintered stainless steels found niche applications(for example, office machine parts, lock parts,mirror mounts, some appliance parts, etc.)where sintered stainless steels were able to com-pete with wrought or cast stainless steelsbecause their corrosion properties met the mod-

erate requirements. Also, powder metallurgy(PM) parts offered their typical advantages:good material utilization and low-cost netshape fabrication (no machining costs).

From the 1950s until the mid-1980s, beltfurnaces were the dominant method of industrialsintering of stainless steels in North America.Maximum sintering temperature was approxi-mately 1150 C (2100 F), and furnaceatmosphere was dissociated ammonia (DA).The lower-cost atmosphere and the higherstrength levels possible with sintering in DAwere attractive, but it was also more difficult toachieve good corrosion properties in DA than

in hydrogen or vacuum. Hence, the gradualshift to hydrogen and vacuum sintering, or theuse of a 90H2-10N2 atmosphere, during the past10 years. There was also a shift toward high-temperature (>1205 C, or >2200 F) sintering.

The majority of studies on the corrosion resist-ance of sintered stainless steels still lack a full

description of the experimental conditionsemployed. The most frequently omitted processparameters are the dewpoint of the sinteringatmosphere and the cooling rate after sintering.Because these and other parameters are of criti-cal importance to corrosion properties of

sintered stainless steels, only publications pro-viding critical processing data and/or permittingunambiguous conclusions are reviewed in thecontext of corrosion-resistance properties.

If sintering conditions are conducive to thedevelopment of good corrosion resistance, goodmechanical properties usually follow. Thereverse is not necessarily true. Each sinteringatmosphere has its own peculiarities with regardto stainless steels, mainly because each respondsdifferently to a number of chemical reactionsinvolving the interstitials carbon, nitrogen, andoxygen. The details of these reactions largelydetermine the corrosion and dynamic mechani-

cal properties of sintered parts. The manymisconceptions about sintering stainless steels(Table 1.1 in Chapter 1, Introduction) arise inlarge part from a lack of appreciation of theimportance of these chemical reactions and fromignoring their differences for the various sinter-ing atmospheres. Even though the extent of thesereactions typically varies from only several hun-dred to a few thousand parts per million, they areof critical importance. Viewing the sinteringatmosphere as mainly an inert cover to protectparts from oxidation, typical of the early years,grossly misjudges its importance.

In wrought stainless steel technology, oxygen,

carbon, and nitrogen are controlled at the refin-ing stage of the production process; in PM, theyare controlled during powder manufacture andsintering. Excessive amounts of carbon andnitrogen can give rise to the formation ofchromium carbides and chromium nitride, withnegative effects on corrosion resistance. These

CHAPTER 5

Sintering and Corrosion Resistance

2007 ASM International. All Rights Reserved.

owder Metallurgy Stainless Steels: Processing, Microstructures, and Properties (05200G)

www.asminternational.org


2/42

60 / Powder Metallurgy Stainless Steels

precipitates can be identified metallographicallyand through special corrosion tests. Furthermore,they resemble the corresponding phenomenain wrought stainless steels. However, precipi-tates of silicon dioxide that form during coolingafter sintering usually do not show up in a met-allographic cross section and are normallyabsent in properly finished wrought stainlesssteels.

5.1 Sintering Furnaces andAtmospheres

Notwithstanding the importance of powder selec-tion, the sintering process is of even greaterimportance for the successful processing of stain-less steels. It encompasses many more elements,from furnace type and atmosphere choice toprocess parameter choices. All of these influence

the quality of a sintered part. To the extent thatthese elements are common to general PMprocessing, their treatment in the following isonly cursory. For a detailed general treatment ofboth practice and fundamentals of sintering, thereader is advised to consult the literature sourcessuggested at the beginning of Chapter 1,Introduction. However, those elements andparameters that have a special bearing on sinteredstainless steels, both regarding their mechanicalproperties and, more so, their corrosion-resistanceproperties, are treated in detail.

Sintering Furnaces. Most commercial sinter-ing of stainless steel parts is performed in

continuous mesh belt conveyor furnaces at tem-peratures up to approximately 1150 C (2100 F).Pusher, walking beam, and vacuum furnaces areused for higher temperatures up to approximately1345 C (2450 F). In recent years, ceramic beltfurnaces have also been introduced for high-temperature sintering. The higher temperatures arefavored for improved mechanical and corrosion

properties. Vacuum and other high-temperaturefurnaces began to be more widely used in the1980s, as a result of increasing demands on mag-netic and corrosion-resistance properties.Although some sintering furnaces for carbonsteel parts now have gas quench capability intheir cooling zones, permitting so-called sinterhardening, most industrial furnaces for stainlesssteels presently lack this feature, despite its ben-efits in minimizing reoxidation in the coolingzone and reducing the risk of sensitization. In thisregard, vacuum furnaces, with their readily avail-able gas quench features, are advantageous.Among the belt furnace types, so-called hump-back furnaces (Fig. 5.1) (Ref 1), give lowerdewpoints. Their inclined entrance and exit zonesretain the lighter hydrogen better than the morecommon horizontal furnaces.

It is the inferior control of dewpoint and slowercooling after sintering in many commercial fur-

naces, compared to laboratory furnaces, that hasled to one of the half-truths (Table 1.1 in Chapter 1)about the corrosion resistance of stainless steelparts, namely, that PM parts possessing good cor-rosion resistance can be produced in laboratoryfurnaces but not in industrial furnaces.

Sintering Atmospheres. Typical sinteringatmospheres for stainless steels include hydro-gen, hydrogen-nitrogen mixtures, dissociatedammonia, and vacuum. Because a low-dewpointcapability is important for both hydrogen andhydrogen-nitrogen atmospheres, there is awidespread belief throughout the industry thatthe use of cryogenic nitrogen in hydrogen-

nitrogen mixtures makes it easier to attain therequired low dewpoints because of the drynessof cryogenic nitrogen. However, the reducibilitycriterion for nitrogen-containing hydrogendemands lower dewpoints than those for purehydrogen (Fig. 5.15).

Dissociated ammonia with dewpoints ofapproximately 45 C (50 F) was the most

Fig. 5.1 Schematic of a humpback mesh belt furnace. Source: Ref 1. Reprinted with permission from MPIF, Metal Powder IndustriesFederation, Princeton, NJ

Wire mesh belt

Driving

pulley

Loading

Entrance incline

Heating

chamber

Muffle

Cooling

chamberExit incline

Unloading

Idling pulley





3/42


4/42


acceptable mechanical properties (acceptable inthis case means that sintered parts will meet theproperties specified in Metal Powder IndustriesFederation, or MPIF, and ASTM standards) areobtainable with the common industrial sinteringatmospheres and with sintering times of only30 min, sintering under conditions that lower theamount of oxygen (oxides) clearly and signifi-cantly improves the dynamic mechanicalproperties of sintered stainless steels, that is,fatigue and impact strength.

For the same fundamental reason, propertiesof fully dense parts, made from gas-atomizedpowders with low oxygen contents and/or lowcontents of undesirable interstitials, can besuperior to their wrought counterparts becauseof their lower levels of interstitials and themore uniform (isotropic) distribution of theseinterstitials within the matrix. Superiormechanical and corrosion-resistance properties

have been documented for fully dense PMstainless steels, PM superalloys, and PM alu-minum alloys (Ref 5).

For vacuum sintering at high temperature(>1200 C, or >2192 F), the addition ofgraphite to the stainless steel powder prior tocompacting can produce oxygen contents of lessthan 300 ppm, with carbon contents of less than0.03% (section 5.2.5 in this chapter) and a 40%improvement in impact strength (Ref 6).

In hydrogen-nitrogen atmospheres, lowerdewpoint and higher hydrogen content give bet-ter reduction of oxides. Lower compact densitywill also produce lower oxygen contents

because of faster diffusion of the reducing gasand reaction products (H2O, CO, CO2). Longersintering time, of course, will also result in morereduction.

Differences in the mechanical properties ofsintered stainless steels as well as in their dimen-sional change (whether from lot to lot or fromproducer to producer) are mainly caused by dif-ferences in the amount and distributions of theinterstitials, oxygen, carbon, and nitrogen. Thesein turn arise from differences in processing. Withgood sintering practice, homogenization of themicrostructure takes place quite rapidly. This isdescribed in more detail in section 5.2.3 in this

chapter.While it is possible to obtain good corrosionresistance in any of the common sinteringatmospheres, each atmosphere demands its owncontrols. It is therefore convenient to discussthis subject individually for each sinteringatmosphere. However, the control of sintered

density and how it affects corrosion resistance iscommon to all types of sintering and is thereforeaddressed now.

5.2.2 Effect of Sintered Density onCorrosion Resistance

The corrosion resistance of stainless steels candiffer widely, depending on the testing environ-ment. Different mechanisms of corrosion havebeen correlated with certain environments.

Acidic Environment. Testing of sinteredstainless steels in acids, that is, H2SO4, HCl,and HNO3, shows that corrosion resistance,measured as weight loss, improves signifi-cantly with increasing density (Fig. 5.4) (Ref7). This relationship is confirmed elsewhere(Ref 810).

The detrimental effect of pores is attributed totwo factors: first, to the large internal surface

areas of sintered parts, which, at the typical den-sities of many structural parts (i.e., 80 to 85% oftheoretical), are still 2 orders of magnitude largerthan their exterior geometric surface areas andtherefore can be subject to increased general cor-rosion; second, to a lack of passivation withinthe pores of a sintered part. Open-circuit meas-urements (section 9.1.3 in Chapter 9, CorrosionTesting and Performance) of wrought stainlesssteels in an acidic environment show that thepotential typically increases with time (Ref 11).This can be interpreted as passivation and/orhealing of active areas. In contrast, sinteredstainless steels often exhibit decreasing poten-

tial, indicating activation of the surface. Itzhakand Aghion (Ref 12) and Raghu et al. (Ref 11)interpret the declining open-circuit potential ofsintered stainless steels as gradually increasing

Fig. 5.4 Relationship between sintered density and weightdecrease of three austenitic stainless steels in 40%

HNO3. Source: Ref 7. Reprinted with permission from MPIF,Metal Powder Industries Federation, Princeton, NJ

Weightloss,weight%





5/42

Chapter 5: Sintering and Corrosion Resistance / 63

activation as the acid penetrates the pores. Thisis accompanied by hydrogen evolution on thesurface of a part. The main reaction taking placeis 2H+ + 2e H2. Thus, corrosion in an acidicenvironment can be viewed as the operation of ahydrogen concentration cell between the exter-nal surface of a part and its internal pore surface.The surface of the pores acts as the anode andthe engineering surface as the cathode. Metaldissolution occurs primarily in the interior of thematerial. After 40 h, the activation processcomes to an end and the potential increases.

Neutral Chloride-Containing Environment.Corrosion resistance in neutral saline solutionshas been found to decline with increasing density(Fig. 5.5) (Ref 13).

The parts of Fig. 5.5 had been prepared fromtypical 100 mesh compacting-grade powders.The decline of corrosion resistance is moderateat low density but becomes very steep at a rela-

tive density of approximately 80 to 84% oftheoretical, depending on pore size, pore mor-phology, and possibly on residual oxygencontent. The fact that some specimens in Fig. 5.5are capable of bridging the low corrosion-resist-ance gap suggests that the effect is a borderlineone and that it may disappear by increasing theintrinsic crevice and pitting resistance of analloy. In fact, some of the specimens of Fig. 5.5,as a result of carbon-assisted vacuum sintering(section 5.2.5 in this chapter), had very lowoxygen contents, comparable to wrought stain-

less steels. Also, corrosion-resistance/densitycurves for 317L, a somewhat more corrosion-resistant material because of higher chromiumand molybdenum contents, appear to possess agreater number of specimens bridging thecorrosion-resistance gap.

Potential-time curves of sintered stainlesssteels in a neutral saline solution exhibit similarbehavior to those in an acidic environment; thatis, they also may be characterized by the poten-tial decreasing with time, indicating activationrather than passivation.

Raghu et al. (Ref 14) have performed cyclicpotentiodynamic polarization studies on sin-tered 316L prepared from narrow sievefractions. Densities varied from 37 to 71%, andtesting was performed in 3% NaCl. The differ-ence potential, E, which is a measure for amaterial susceptibility to crevice corrosion,increased with decreasing pore size (Fig. 5.6)

The detrimental effect of pores is very strongfor small pores up to approximately 20 m (asdetermined by the bubble point test method forfilters) and thereafter becomes much less pro-nounced. The important variable in this case ispore size rather than porosity.

These results are best explained by assumingthe operation of an oxygen concentration cell,which establishes itself in accordance with themechanism shown in Fig. 5.7 (Ref 15) andFig. 5.8 (Ref 16).

1,000

100

10Corrosionresistance(5%aq.NaClbyimmersio

n),

hA-rating

6.5 6.7 6.9 7.1 7.3 7.5

Density, g/cm3

Fig. 5.5 Effect of density on corrosion resistance of 316L parts. , pressed and sintered only; o, pressed, sintered, re-pressed, andannealed. Source: Ref 13. Reprinted with permission from MPIF, Metal Powder Industries Federation, Princeton, NJ





6/42

The overall reaction involves the dissolution ofmetal, M (immersed in aerated saline solution),and the reduction of oxygen to hydroxide inaccordance with:

Oxidation

M M+

+ e

Reduction

O2 + 2H2O + 4e 4OH

As a result of limited diffusion within thepore space of a part, oxygen within that spacebecomes depleted and oxygen reductionceases. However, as shown in Fig. 5.7, metaldissolution continues within the pore space.The latter creates a positive charge (M+)within the pore space, which is neutralized bythe migration of chloride ions into the porespace. The increased metal chloride concen-tration within the pore space undergoeshydrolyzation into insoluble hydroxide andfree acid according to:

M+Cl + H2O = MOH + H+Cl

The free acid increases metal dissolution,which in turn increases migration, representingan accelerating, autocatalytic process. As thecorrosion within the pore space increases, the


Fig. 5.6 Effect of pore size on size of hysteresis(E) for sintered316L in 3% NaCl (27 C, or 81 F ). Source: Ref 14.

NACE International 1989

Fig. 5.7 Crevice corrosion mechanisminitial stage.Source: Ref 15. Reprinted with permission, Fontana,

Corrosion Engineering, 2d ed. The McGraw-Hill Companies,Inc., 1978

Fig. 5.8 Crevice corrosion mechanismlater stage. Source:Ref 16. Reprinted with permission from MPIF, Metal

Powder Industries Federation, Princeton, NJ

330

300

270

240

210

180

150

120

90

60

30

0

Potential(E=EBEpp

),m

V

0 10 20 30 40 50 60 70 80 85

Pore size, m





7/42


rate of oxygen reduction on the internal poresurfaces also increases. This cathodically pro-tects the external surfaces, which explains whyduring crevice corrosion the attack is localizedwithin the porous or shielded areas, while theremainder suffers little or no damage.

Pore morphology of a sintered part is affectedby powder particle shape, particle size distribu-tion, compacting pressure, amount of shrinkageduring sintering, re-pressing, and so on. For astainless steel part made from a typical 100mesh compacting-grade stainless steel powder,minimum corrosion resistance, as measured byimmersion in 5% NaCl, appears at a relativedensity of approximately 87 to 90%, againdepending on pore morphology. Past the mini-mum, corrosion resistance increases again. Theincrease past the minimum is attributed to the dis-appearance of pores as sintered structural partsapproach the region of closed-off porosity at

approximately 92% of theoretical density. Amore uniform density distribution in a sinteredpart, such as is obtainable with isostatic pressing,or, more practical, with warm compaction, mayreduce the width of the crevice-corrosion densityregime.

It should be emphasized that corrosion resist-ance as shown in Fig. 5.5 was measured by thetime it took for the development of rust, basedon visual assessment. Weight change measure-ments are not reliable for assessing corrosionresistance of sintered parts that were tested in aneutral environment.

It is not clear why the corrosion resistance at

the higher densities past the corrosion-resist-ance minimum does not approach that of itswrought counterpart. Both pore morphologyand residual oxygen content may play a role. Infact, as is shown in section 5.3 on liquid-phasesintering of stainless steels, a high-density,boron-containing, liquid-phase-sintered 316L

had a chloride (immersion in 5% aqueous NaCl)corrosion resistance similar to wrought 316L,while other high-density-sintered stainlesssteels of the same composition but withoutboron had much lower corrosion resistances.The boron may have scavenged and redistrib-uted the residual oxygen of the sinteredmaterial, with the formation of less detrimentalborosilicates.

Conflicting with the aforementioned results,short-term potentiodynamic polarization testsby Lei et al. (Ref 17) pointed to a beneficialeffect of density in a saline environment. Thecontroversy was resolved when Maahn andMathiesen (Ref 18) observed that in short-termpolarization tests, there was not enough time forthe time-consuming buildup of localized attackwithin pores (Table 5.1) (Ref 19).

While the corrosion resistance related to theouter surfaces, given by ipeak and ipass, in gen-

eral improves with increasing density, with Epitremaining unchanged, more relevant long-termexposure techniques, such asEstp and salt spraytesting (NSS1 and NSS2) (Chapter 9, CorrosionTesting and Performance), show increasing sus-ceptibility to crevice corrosion with increasingdensity and increasing oxygen content. Thus, byusing slow, stepwise polarization (section 9.1.3in Chapter 9), the expected relationshipadecrease of the stepwise initiation potential,equivalent to deteriorating corrosion resist-ancewas observed.

In the past, many instances of corrosion ofsintered stainless steels were interpreted as

crevice corrosion because of porosity, when infact they were clearly the result of incorrect orsuboptimal sintering that produced metallurgi-cal defects that gave rise to intergranular orgalvanic corrosion or that, because of an exces-sive vacuum (section 5.2.5 in this chapter), ledto severe chromium depletion of the surfaces of

Table 5.1 Effect of density and oxygen content on corrosion resistance of hydrogen-sintered 316L

Specimens sintered at 1250 C (2282 F), 120 min in pure hydrogen

Compaction Sintered

pressure density, Open Ipeak(a), Ipass (a), Epit (a), Estp (b),MPa ksi g/cm3 pores, % O, ppm A/cm2 A/cm2 mV SCE mV SCE NSS1, h NSS2

295 43 6.34 19.4 340 31 20 475 0 >1500 9

390 57 6.62 15.5 1260 18 19 425 100 985 7490 71 6.86 12.3 970 25 15 475 75 36 4540 78 6.94 10.8 1900 18 15 500 200 60 3590 86 7.02 9.7 1410 21 14 450 125 28 2685 99 7.13 7.6 2150 9 7 500 225 48 2785 114 7.23 5.7 2040 7 7 475 200 24 2

(a) 0.1%CI, pH 5, 30 C (86 F), 5 mV/min. (b) 5% NaCI, 30 C (86 F), 25 mV/8 h





8/42


sintered parts. Mathiesen and Maahn (Ref 20)have used image analysis on 316L parts,sintered in hydrogen under various conditionsof time and temperature, to obtain a wide rangeof sintered densities. Considering the pores ascylindrical holes, they expressed the severity ofcorrosion in pores as:

S= ia 1/d

where dis the pore diameter, and ia is the corro-sion rate in the passive state. Figure 5.9 shows aplot of the visual rating of corrosion (10 = norust; 1 = 50% rust) after a 1500 h salt spray testversus the aforementioned severity value.

In the same investigation, Mathiesen andMaahn show that stepwise pitting potential(0.5% Cl) decreases with increasing sintereddensity of 316L (Table 5.2).

The numbers within the body of Fig. 5.9 referto the tables of the paper in which the variousexperiments are described. It is apparent thatcorrosion resistance begins to deteriorate rap-idly at a density of approximately 6.6 g/cm3

(82% of theoretical). The authors attribute thedeterioration of corrosion resistance withincreasing density to both a critical pore geom-etry and to impeded reduction of oxides.

Figure. 5.10 (Ref 15) shows the results of acrevice-corrosion test in accordance withASTM G 48 wrought 316L and sintered 316L.The density of the sintered 316L was 6.8 g/cm3

(85% of theoretical), that is, well within thesteep decay region for a compacting-gradematerial.

Interestingly, the sintered part showed only amild attack in comparison to the severelycorroded wrought stainless steel of the same

Fig. 5.9 Visual rating after 1500 h salt spray test versus severity value calculated as the reciprocal of average pore diameter. Reprintedwith permission from MPIF, Metal Powder Industries Federation, Princeton, NJ

Table 5.2 Effect of density for 316L cylindrical specimens sintered at 1250 C (2282 F), 120 minin pure hydrogen

Estp(c)

Green density, Density(a), Open Average pore Ferroxyl mV SCE, Cl:

g/cm3

g/cm3

pores(a),% diameter(b), m Roundness(b) NSS1, h NSS2 test, Cl

: 0.5% 0.1% 0.5%

5.80 6.40 19.3 9.5 0.73 1336 9 0 350 1505.91 6.51 17.8 8.8 0.74 >1500 10 0 250 1506.05 6.65 15.9 8.0 0.72 >1500 10 0 275 1006.19 6.75 14.4 8.7 0.79 >1500 10 0 250 1256.25 6.83 13.4 8.0 0.74 >1500 10 0 300 1006.38 6.93 11.8 7.3 0.75 1168 9 0 325 1006.44 7.01 10.7 6.1 0.71 192 5 0 300 50

(a) Measured by oil impregnation technique. (b) Measured by image analysis. (c) Stepwise polarization

10

9

8

7

6

5

4

3

2

1

0

NSS2

0.1 0.11 0.12 0.13 0.14 0.15 0.16 0.17

Severity, 1/eqv. diameter

55

5 55

54 4

4 44

4

4

3

2

2

2

3

3

3

3

3 3

22

2

2

5





9/42


designation. Evidently, because the entire PMpart, on account of its porosity at a relativedensity of 85%, already represents a system ofinterconnected crevices, any additional crevice

in accordance with ASTM G 48 test seems tohave only a minor effect. This relationship isanalogous to the lower notch sensitivity of sin-tered parts in comparison to wrought parts. Theauthors also found sintered type 304L and 316Lstainless steels to be less susceptible to crevicecorrosion than wrought 316L, on the basis of theareas of the hysteresis loops of their cyclicpolarization curves (Ref 15).

The reduced crevice sensitivity of sinteredstainless steels may be attributed to their inter-connected pores, which facilitate oxygendiffusion through and from neighboring pores.As such, it appears that the pore space surround-ing a crevice should be taken into account inassessing its susceptibility to crevice corrosion.Oxygen diffusivity within the pore space of asintered part, as a measure for its capability totransport oxygen to its internal surfaces in orderto maintain passivity, appears to be a better

characterization for its resistance to crevicecorrosion than an average pore diameter. A per-meability or diffusivity number takes intoaccount the entire pore space, including its tortu-osity. Characterization of the pore space, throughmercury porosimetry would also appear toprovide more relevant characterization than anaverage pore diameter. In mercury porosimetry(Fig. 5.11) (Ref 21), the measured pore sizesrepresent the bottlenecks between neighboringpores rather than pore diameters themselves. Itis the totality of these bottlenecks, rather than

Fig. 5.10 Comparison of wrought and sintered type 316Lstainless steels before and after testing in 10%

aqueous FeCl3. (a) Assembled crevice-corrosion test specimenof wrought type 316L (100% dense). (b) Assembled crevice-corrosion test specimen of sintered type 316L (85% dense). (c)Wrought specimen after test showing severe attack at fourcrevices under rubber bands and synthetic fluorine-containingresin ring. (d) Sintered specimen after test showing slight attackunder synthetic fluorine-containing resin ring. Source: Ref 15.Reprinted with permission from MPIF, Metal Powder IndustriesFederation, Princeton, NJ

Fig. 5.11 Mercury porosimetry curves of sintered steel parts of varying densities. Green skeletons were sintered at 1093 C(2000 F) for 20 min. Total porosity () is determined from sample weight and dimensions. Source: Ref 21

0.157





10/42


the actual pore diameters or pore volumes, thatexert the greater influence on a parts capabilityto facilitate gas diffusion through the porespace. According to Fig. 5.12, the majority ofthe pore bottlenecks of a sintered steel part(made from a compacting-grade, 100 mesh,water-atomized powder), of a relative density of80 to 84%, are 4 to 5 m in size, and at a rela-tive density of 87 to 90%, approximately 2 m.Assigning greater importance to the bottleneckswould also explain the shift of the crevice-corrosion minimum to a higher density as aresult of repressing (Fig. 5.6). In pressing orrepressing, densification comes about firstthrough the collapse of the larger pores (Ref22), whereas in sintering, it is the small poresand the connections between large pores, that is,the aforementioned bottlenecks, that, because oftheir small curvatures and greater surface ener-gies, become active. Thus, repressing simply

increases the density of a part without greatlyaffecting its bottleneck pores or its diffusivitycharacteristics, and hence, the shift of maximumcorrosion to a higher density.

According to Maahn et al. (Ref 19), the corro-sion attack in a ferric chloride test (ASTM G 48)may develop within the pores beneath the sur-face of a part. In this case, the superior surfaceappearance of a sintered part may be misleading,

and interior examination and testing for mechan-ical property degradation is appropriate.

For a better assessment of the effect of poremorphology on crevice-corrosion resistance inthe low-density range below approximately80% of theoretical, optimally sintered parts withoxygen contents below approximately 200 ppmshould be evaluated. Such parts could be pre-pared with carbon-assisted optimal vacuumsintering, or, easier, by optimal gravity sinteringof a low-oxygen-content, inert-gas-atomizedstainless steel powder, or by warm compactionand sintering of such a powder.

Molins et al. (Ref 23) have investigated theinfluence of several finishing operations on thecorrosion resistance of sintered 316L as meas-ured potentiodynamically according to ASTM B627 in a solution of 0.1 N NaCl and 0.4 NNaClO4 (Fig. 5.12).

Worsening passivation due to tumbling was

interpreted as due to smearing of pores andpotential contamination from additives. Thebest and most significant improvement resultedfrom operations that sealed surface porosity:grinding, turning, and shot blasting.

It should be stressed again that the effective-ness and/or ranking of such treatments willdepend on the quality of sintering. Thus, whileany such results may be relevant and practical

20

15

10

5

0

Currentdensity,mA/cm2

0.3 0.0 0.3 0.6 0.9 1.2 1.5

Turned

Shot blasted

GroundChemical passivated

As-sintered

Sized 4 Tn/cm2

Thermal passivated

Tumbled

Volts (SCE)

Fig. 5.12 Potentiodynamic curves of 316L stainless steels as a function of surface finishing treatment. Reprinted with permissionfrom MPIF, Metal Powder Industries Federation, Princeton, NJ





11/42


for an individual parts producer, they differ foreach parts producer. Only when sintering condi-tions approach optimal should results andranking be generalized.

There have been attempts to decrease or allevi-ate crevice corrosion in sintered stainless steelsby impregnating the pores with a resin, by metal-lurgical modification of the pore surfaces, or bythe use of higher-alloyed stainless steels, particu-larly those containing higher concentrations ofmolybdenum. The authors found resin impregna-tion beneficial only in cases where the stainlesssteel parts had been improperly sintered andtherefore had an initial low corrosion resistance.In optimally sintered parts, that is, parts thatendured a long exposure to the testing solution,the tested resins separated from the pore surfaces,and the testing liquid was able to seep into thespaces. In several instances, resin impregnationalso introduced ferrous contamination and unac-

ceptable galvanic corrosion. The approachesbased on surface modification and higher-alloy-ing additions (Chapter 6, Alloying Elements,Optimal Sintering, and Surface Modification inPM Stainless Steels) show promising results.Another promising approach to avoiding theproblem of long-term corrosion in a neutral saltsolution by the presence of crevice-sensitivepores is to make use of liquid-phase sintering andto achieve sintered densities greater than 7.4g/cm3 (section 5.3 in this chapter).

5.2.3 Sintering of Stainless Steelsin Hydrogen

Hydrogen has now become the most widelyused atmosphere for sintering stainless steels. Inthe interest of good corrosion resistance, the pri-mary goal in processing is to lower the oxygencontent of the green part as much as possible, toprevent reoxidation in the cooling zone of thefurnace, and to maintain a low carbon content ofapproximately 0.03% in austenitic stainlesssteels (0.02% for high-nickel contents), andpreferably, still lower for ferritic stainless steels.Apart from the sintering temperature, the twomost critical parameters are the dewpoint of thehydrogen atmosphere (a measure of the water

content of the atmosphere) and the cooling rateafter sintering.Oxygen Control during Sintering. For

elevated-temperature metallurgical reactions,equilibrium data are very informative because ofthe greater ease with which reactions take placeat high temperatures. For sintering of stainless

steels in hydrogen, the equilibria of interest areusually shown as so-called redox curves. Suchcurves show, as a function of temperature, atwhat water content of the sintering atmosphere ametal becomes oxidized. In scientific literature,the water content is usually shown in terms ofwater pressure, pH

2O; in technical engineering-

type literature, it is often shown in terms ofdewpoint of the atmosphere, because of the easyway to determine dewpoints. The two scales canbe converted into each other via temperature-pressure data for steam (Fig. 5.13) (Ref 24).

The dewpoint, , may also be calculated bythe following equation (Ref 25):

= 273 A/(lnpH2O B) [C]

where A= 6128 and B = 17.335 for 100 C 0 C and for 10 8pH2O 6 10

3

Figure 5.14 (Ref 19) shows such redox curves,

calculated from thermodynamic data, for severalof the pure, high-oxygen-affinity elements presentin stainless steels as well as for some of these ele-ments present as solid solutions in stainless steel.

Figure 5.15 (Ref 25) shows redox curves forpure metals and their oxides against both the par-tial pressure ratio pH2/pH2O and the dewpoint,as well as for H2-N2 mixtures.

105

104

103

102

10

1

H2/H2Oratio

100 50 0 50 100

Dewpoint, C

Fig. 5.13 Relationship between ratio of H2/H2O and dew-point. Source: Ref 24





12/42


Partial pressures of water and temperatures tothe left of a redox curve indicate that the oxideof that particular metal is stable under such con-ditions, whereas under conditions that lie to theright of that curve, the pure metal is stable.

Because reduction equilibria depend on theratio of the partial pressures of hydrogen andwater, that is,pH2/pH2O

, and not on the absolutewater pressure, as dewpoint does, the dew-points for oxide reduction of hydrogen-nitrogenmixtures differ from those of pure hydrogen. Theminimum temperature at which a metal oxidecan be reduced in a hydrogen-nitrogen mixtureof a given dewpoint can be derived from Fig.5.15 by drawing a horizontal line at the height ofthat dewpoint in the left part of the figure.

At the intersection of this horizontal line withthe gas mixture of the atmosphere, a perpendicu-lar line is drawn up to the curve for purehydrogen. From this intersection, a horizontal line

is extended into the right part of Fig. 5.15. Alloxides above this line can be reduced, whereas alloxides below this line are stable. The example inFig. 5.15 shows that with a dewpoint of 35 C(31 F), it is possible to reduce Cr2O3 at a tem-perature of approximately 1000 C (1830 F) orhigher in pure hydrogen (dashed line), whereas anH2-N2 atmosphere with 95% N2 requires a mini-mum temperature of almost 1600 C (2912 F)(solid line) or a dewpoint of almost 60 C (76F). For dissociated ammonia and atmospherescontaining lesser amounts of nitrogen, this dew-

point correction is relatively small. Figure 5.14shows that very low dewpoints are required, oronly very small amounts of water vapor can betolerated, if stainless steels are to be kept frombecoming oxidized during sintering. Also, thisrequirement is easier to fulfill as the sintering tem-perature increases. Furthermore, when an elementis present in the form of an alloy, its activity isdecreased, and it is easier to keep it from becom-ing oxidized than if the same element is present asa pure metal. Of the most oxidation-proneconstituents in conventional stainless steelsmanganese, chromium, and siliconsiliconexhibits the greatest affinity to oxygen. Also, forsilicon, the difference between pure and alloyedstates is particularly large and explains, accordingto Larsen, why silicon dioxide in sintered stainlesssteels can be reduced under commercial sinteringconditions, a fact that often had been doubted inearlier years. Another scenario for the successful

reduction of SiO2 at reasonable dewpoints isbased on the reduction of SiO2 to volatile SiO fol-lowed by the iron-catalyzed reduction of SiO tosilicon (Ref 16).

Figure 5.16 shows the Auger composition-depth profile of a 316L part sintered inhydrogen at 1260 C (2300 F).

A comparison with Fig. 3.10, which shows thesame profile for water-atomized 316L in thegreen condition, demonstrates the enormousdegree of reduction of SiO2. Furthermore, thewidth of the oxygen profile of Fig. 5.16 is narrow,approximately 30 atomic layers (~50 ), andmore akin to the thickness of a passive film.

This materials corrosion resistance was excel-lent. Figure 5.17 (Ref 26) shows a similarly

Fig. 5.14 Redox curves for oxides in equilibrium with 316Lin H2 at atmospheric pressure. Source: Ref 19.

Reprinted with permission from MPIF, Metal Powder IndustriesFederation, Princeton, NJ

100

50

0

50

100

Dewpoint,C

Nitrogen content0%50%80%95%

100 102 104 107

Partial pressure ratio,pH2

/pH2O

800 1600

Temperature, C

WO3

FeOMoO2

SnO2ZnO Cr2O3 SO2

MnO

SiO2

VOTiO

Al2O3MgO

Fig. 5.15 Dewpoint for various hydrogen-nitrogen mixtures inequilibrium with metal/metal oxide. Source: Ref 25

102

103

104

105

106

PartialpressureofH2O,bar

800 900 1000 1100 1200 1300 1400

SiO2

/Si

SiO 2

/316L

Cr2O3

/CrCr2

O3/316L

MnCr2

O4/31

6L

FeCr2

O4/31

6L

Temperature, C





13/42


excellent profile for a high-temperature (1295 C,or 2363 F) vacuum-sintered 316L part.

The absence of silicon in Fig. 5.17 appears tobe due to the authors use of the nonscanningmode of Auger analysis, because they were ableto confirm its presence when they were usingthe scanning mode.

Figure 5.18 shows the profiles of a 316L partsintered at 1120 C (2048 F) in hydrogen with adewpoint of 35 C (31 F).

Based on Fig. 5.18, these conditions are mar-ginal for SiO2 reduction, and it is therefore notsurprising that the corrosion resistance of this

material was quite inferior. Also, the width of theAuger oxygen profile of this material is widerthan that of 316L reduced at higher temperatures.

Table 5.3 (Ref 19) further illustrates how amarginal dewpoint (35 C, or 31 F) for 316Lparts sintered in hydrogen at 1120 and 1250 C

(2048 and 2282 F) affects the electrochemicalpassivation characteristics as well as the long-term-exposure corrosion resistance in 5%aqueous NaCl in comparison to a much lowerdewpoint of 70 C (94 F).

Microstructures. Thus, dewpoint value, togetherwith temperature, largely determines if adequateparticle bonding will take place during sintering.Interparticle bonding can readily be ascertainedthrough metallography. Figure 5.19 shows themicrostructure of a 316L stainless steel part sin-tered in hydrogen for 30 min at 1093 C (2000 F).Prior-particle boundaries and angular pores areevident as a result of insufficient sintering.

Figure 5.20 shows the polished cross sectionof undersintered 304L, with many oxides in thegrain boundaries.

In contrast, well-sintered 316L (Fig. 5.21),shows good interparticle bonding, well-roundedpores, and narrow and precipitate-free grain

boundaries in the austenitic structure.Figure 5.22 shows the same attributes for awell-sintered ferritic stainless steel, 434L,except for the absence of twin boundaries,which are characteristic of face-centeredaustenitic stainless steels.

Figure 5.23(a) (Ref 27) shows the surface andFig. 5.23(b) the cross section of 316L vacuumsintered for 1 h at 1150 C (2102 F). The oxideparticles seen in Fig. 5.23(a) are typically lessthan 1m in diameter.

After exposure to FeCl3, the oxide particle sitesdevelop corrosion pits (Fig. 5.24b) (Ref 27).

It was only recently recognized that, in order to

develop excellent corrosion properties in sinteredstainless steels, not only a thorough reduction ofoxides but also prevention of reoxidation aftersintering is required. The parts in Fig. 5.25 (Ref28), sintered under various conditions, had goodstatic mechanical properties, but they had a broadrange of corrosion resistances as measured bysubmersion in a saline solution. Corrosionincreased with increasing oxygen content of thesintered part.

The detrimental effect of oxygen on corrosionresistance was confirmed by Maahn andMathiesen (Ref 18). The pitting potential, ameasure of a steels resistance to pitting, declined

with increasing oxygen content in 316L.Kinetic Considerations. Using gas and massspectrometry analysis during sintering of stain-less steels in hydrogen and under vacuum, atvarious temperatures, and with additions ofgraphite to the powder, Tunberg et al. (Ref 6)

Fig. 5.16 Auger composition-depth profile of a 316L partsintered in hydrogen at 1260 C (2300 F)





14/42


and Larsen and Thorsen (Ref 29, 30) showedthat the various reduction reactions occurred infairly close agreement with the equilibrium gasconcentrations calculated from thermodynamicdata. More specifically, the dominant reactionsfor sintering in hydrogen are:

2H2 + SiO2 = 2H2O + Si316L

2C + SiO2 = 2CO + Si316L

C + 2H2 = CH4

Fig. 5.18 Auger composition-depth profile of 316L sintered for 30 min in hydrogen with a dewpoint of 35 C (31 F). Oxygencontent was 0.24%

Fig. 5.17 Auger composition-depth profile of a 316L part vacuum sintered for 30 min at 1295 C (2363 F). Oxygen content was0.20%. Source: Ref 26

60

40

20

0

Concentration,%

0 100 300200 400 500

Sputtering time, s

Fe

Cr

Ni

O

C

S

1000 s correspond to

about 300 atomic layers

60

40

20

0

Concentration,%

0 100 300200 400 500

Sputtering time, s

Fe

Cr

Ni

O

C

S

1000 s correspond to

about 300 atomic layers





15/42


and for sintering in vacuum:

2C + SiO2 = 2CO + Si316L

However, due to low equilibrium pressures of

the reaction products, it can take a long time toremove oxygen from the parts. Increased partdensity also slows down the reactions.

Larsen and Thorsen (Ref 30) have shown thatcarbon is a much more effective reducing agentthan H2 for the reduction of oxides, and carbonremoval is much faster in vacuum. Tunberg et al.

(Ref 6) were able to obtain a tenfold reduction

in oxygen content (from 0.31 to 0.03%) byadding 0.19% C to a 304L stainless steel pow-der and by vacuum sintering at 1200 C (2192F) for 1 h. Removal of surface oxides led toimproved interparticle bonding, as reflected inmarkedly improved dynamic mechanical prop-erties (elongation, impact strength). The

Table 5.3 Corrosion properties of 316L steel sintered in hydrogen with a dewpoint of 35 or 70 C(31 or 94 F) at different combinations of time and temperature

ipeak(a), ipass(a), Epit(a),

A/cm2 A/cm2 mV SCE NSS1, h NSS2

Dewpoint, C 35 70 35 70 35 70 35 70 35 70

1120 C/30 min 150 10 29 11 250 375 36 >1500 5 91250 C/30 min 105 7 20 12 325 325 288 1260 4 8

1120 C/120 min 120 10 25 10 325 375 48 1272 5 71250 C/120 min 83 4 19 9 325 500 24 96 1 7

(a) 0.1%C1, pH 5, 30 C (86 F), 5 mV/min. Source: Ref 19

Fig. 5.21 Well-sintered 316L (etched) revealing interparticlebonding, twin boundaries, rounded pores, and

precipitate-free grain boundaries

Fig. 5.22 Well-sintered 434L (etched) revealing interparticlebonding, rounded pores, and precipitate-freegrain boundaries

Fig. 5.19 Undersintered 316L (unetched) revealing prior-particle boundaries and angular pores

Fig. 5.20 Polished cross section of undersintered 304Lrevealing oxides in grain boundaries





16/42


investigators attributed the superior oxygen

reduction during vacuum sintering to the fasterremoval of carbon monoxide from the pores.The faster chemical reaction rates for vacuumsintering and the low oxygen levels achievablewith the addition of an appropriate amount ofcarbon should be of considerable commercialinterest (section 5.2.5 in this chapter).

Figure 5.26 (Ref 28) shows the very strong

effect of part size on weight loss during sinter-ing of 316L transverse-rupture specimens in dryhydrogen. The H2O and CO account for themajor portion of the total weight loss. Rate-controlled by the slow transportation of reactionproducts through the tortuous pores, oxides nearthe surface of a part are reduced first, followed

Fig. 5.24 (a) SEM and (b) light microscopy microstructures of vacuum-sintered 316L after exposure to 6% FeCl3. Reprinted withpermission from MPIF, Metal Powder Industries Federation, Princeton, NJ

Fig. 5.23 Microstructures of vacuum-sintered 316L. (a) SEM. (b) Light microscopy. After exposure to FeCl3, the oxide particle sitesdevelop corrosion pits (Fig. 5.24b). Source: Ref 27. Reprinted with permission from MPIF, Metal Powder Industries

Federation, Princeton, NJ





17/42


by the reduction of oxides farther inside a part.Again, this suggests that there exists an oxygencontent gradient in sintered stainless steels, withthe oxygen content increasing from the surfaceto the interior of a part. Part supports (ceramicbodies, metallic belts) and part spacing will alsoaffect the diffusion of gases into and out of thesintered bodies and therefore their reaction rateswith interior oxides.

In their tests with 316L parts, Samal and Klaret al. (Ref 31, 32) carefully eliminated allknown corrosion defects, such as contamina-tion, grain-boundary carbides and nitrides,reoxidation on cooling, and the crevice-sensitivedensity range around 6.9 g/cm3, except forresidual oxygen (oxides). A reduction of the

oxygen content of the sintered parts fromapproximately 1900 to approximately 1300 ppmthrough increasing the sintering temperaturefrom 1138 to 1316 C (2080 to 2400 F)improved the saline (5% NaCl) corrosion resist-ance by 400%, (Fig. 5.27). Vacuum-sintered316L parts with oxygen contents of approxi-mately 700 ppm showed a 700% improvementover the low-temperature hydrogen-sintered

parts. Sintered 316L parts with still lower oxygencontents (200 to 300 ppm) are expected to have ayet higher corrosion resistance, with reduced orno evidence of crevice corrosion, but with gen-eral corrosion, as measured by Ipass, reflectingthe larger effective surface areas of such parts.Such low-oxygen parts could be made from a

2100 F, 45 min

H2, 90 F dewpoint

316L

0.3

0.2

0.1Weightlossduringsintering,%

0 2 5 10 20

Transverse-rupture specimen size, g

Fig. 5.26 The effect of transverse-rupture specimen size onweight loss during sintering in hydrogen (densityof specimens: approximately 6 g/cm3). Source: Ref 28.Reprinted with permission from MPIF, Metal Powder IndustriesFederation, Princeton, NJ

0 1000 2000

3000

2000

1000

0

Corrosionresistance,h

(5%aq.NaCl,B-rating)

Oxygen content in sintered part, ppm

Fig. 5.27 Corrosion resistance of 316L stainless steel partssintered under various conditions under exclusion of

defects, except for residual oxides

3000

2500

2000

1500

1000

500

0

Oxygencontent,ppm

0.1 1 101 102 103 104

Corrosion time, h

(1120, 34, 188)

(1120, 34, 2914)

(1120, 34, 3210)

(1120, 51, 178)

(1260, 34, 2294)

(1260, 51, 2817)

(1260, 34, 22)(1260, 51, 42)

(1120, 42, 2050)(1120, 37, 78)

316L sintered in hydrogen

316L sintered in dissociated NH3316L-1.5Sn sintered in hydrogen

316L-1.5Sn sintered in dissociated NH3

Fig. 5.25 Effect of oxygen content on corrosion resistance of sintered 316L and tin-modified 316L (sintered density: 6.65 g/cm3;cooling rate: 75 C/min, or 135 F/min). Values in parentheses are sintering temperature (C), dewpoint, (C), and nitro-

gen content (ppm), respectively. Time indicates when 50% of specimens showed first sign of corrosion in 5% aqueous NaCl. Source:Ref 28. Reprinted with permission from MPIF, Metal Powder Industries Federation, Princeton, NJ





18/42


thermally agglomerated (Ref 33), low-oxygen-content, gas-atomized 316L powder or, morepractically, by warm compaction (Chapter 4,Compacting and Shaping) of a low-oxygen-content, gas-atomized powder in combinationwith, if necessary, an appropriate binder.

Oxygen Control during Cooling. Theimportance of a low dewpoint during the sinteringof stainless steels has been described previ-ously. However, it is also important to controlthe cooling conditions after sintering if maxi-mum corrosion resistance is to be achieved. Forillustration, Lei and German (Ref 34) subjecteda wrought 304L stainless steel specimen,together with PM parts, to sintering in dryhydrogen (dewpoint 35 C or 31 F) for60 min at 1250 C (2282 F). The corrosion rateof the wrought stainless steel (as measured bypotentiodynamic scanning in 3.5% saltwater)after exposure to sintering increased by a factor

of 100. Electrochemical testing of the exteriorsurfaces of the sintered PM stainless steelsshowed similar degradations. No cooling rateswere disclosed in these experiments, but theauthors had observed second-phase inclusionson the surfaces of the wrought stainless steelafter its simulated sintering cycle. They attrib-uted the large decrease in corrosion resistanceof both the sintered and wrought specimens tochromium losses from the surfaces due to

chromium evaporation. It is more likely, how-ever, that the culprit was reoxidation duringcooling, with the formation of spheroidaloxides on the exposed surfaces. Figure 5.28(Ref 13) shows a 316L stainless steel part thathad first been vacuum sintered to reduce theoxygen content to approximately 700 ppm (Ref27) and then allowed to cool in hydrogen (dew-point 40 C, or 40 F) from 1127 C (2061 F)at a cooling rate of 187 C/min (337 F/min). Theoxide particles formed during cooling werespherical and measured between 0.5 and 2.0 min diameter. To the naked eye, the surface bright-ness of such parts is not affected by this type ofreoxidation. Auger line analysis (Fig. 5.29)

Fig. 5.28 Spheroidal SiO2 particles formed on 316L part oncooling. Source: Ref 13

Fig. 5.29 SEM and Auger line analysis of 316L surfaces containing surface oxides formed during cooling. Reprinted withpermission from MPIF, Metal Powder Industries Federation, Princeton, NJ





19/42


identified the particles as consisting predomi-nantly of silicon oxide. Chromium was absent.

After exposure to aqueous FeCl3 for 7 h, theoxide particles had formed corrosion pits. Thechloride corrosion resistance had decreased to avery low value. As shown in Fig. 5.32, 316Lloses approximately 99% of its corrosion resist-ance in 5% saltwater when cooled under theaforementioned conditions. SS-100 parts,a higher-chromium, higher-nickel austeniticstainless steel, treated identically, exhibited verylittle pitting and had lost only approximately50% of their chloride corrosion resistance. Theoxide particles of the SS-100 material were ofspherical and triangular shape.

Redox curves and cooling-rate relationshipsprovide insight on why it is important to controlthe cooling process. The redox curves in Fig. 5.30(Ref 19) show two sintering scenarios.

In scenario 1, with a dewpoint of 40 C (40F), the sintered parts, as they enter the coolingzone and their temperature decreases, will beginto become oxidized at approximately 1070 C(1958 F), the temperature at which the dew-point of 40 C (40 F) crosses the redox curve.In scenario 2, with a lower dewpoint of 60 C(76 F), oxidation is delayed until the partshave cooled to a lower temperature of 960 C(1760 F). It is clear that the parts become moreoxidized under scenario 1 than under scenario 2.

The importance of a fast cooling rate for min-imizing reoxidation is self-evident. It is alsoobvious that a high dewpoint requires a fastercooling rate than a lower dewpoint, because ofthe higher concentration of water vapor in a

higher-dewpoint atmosphere and becausereoxidation starts at a higher temperature.Sands et al. (Ref 35) suggested a maximumwater content of 50 ppm (corresponding to adewpoint of 48 C, or 54 F) for slow coolingin hydrogen. Figure 5.31 illustrates these rela-tionships in a semiquantitative scheme.

Upper Critical Cooling Temperatures. Figure5.32 shows the upper critical cooling temperaturesfor 316L, that is, the lowest temperatures whererapid cooling must begin in order to avoid sensiti-zation. The required cooling rates, as a function ofdewpoint, are also shown in that figure.

The curves marked with percentage figures indi-

cate, semiquantitatively, how rapidly corrosionresistance (in 5% NaCl) deteriorates with decreas-ing cooling rate. Thus, a part cooled at 400 C/min(720 F/min) in a dewpoint atmosphere of

+20

0

20

40

60

80

100

Dewpoint,F

1200 1600 2000 2400

Temperature,F

Oxidation316

L/SiO 2

316L

/Si

1

2

Sintering

Fig. 5.30 Redox curves and sintering scenarios for 316L inH2 at atmospheric pressure (schematic). Source:Ref 19. Reprinted with permission from MPIF, Metal PowderIndustries Federation, Princeton, NJ

2400

2000

1800

1200

800

200

Temperature,

F

Temperature,

F

Oxidation

Oxidation

Slowcool

Slow

cool

Fastcool

Fastcool

Dewpoint 60 F Dewpoint 40 F

0 10 0 10

Cooling time, min Cooling time, min

Fig. 5.31 Temperature-time cooling profiles for two dewpoints,showing schematically approximate reoxidation

regimes for 316L

500

400

300

200

100

0

Coolingrate,

C/min

800

900

1000

1100

1200

1300

Temperature

C

30 40 50 70

Dewpoint of sintering atmosphere,C

1%

30%

100%

Upperc

riticalcoo

ling

Temper

ature

316L (


20/42


40 C (40 F) attains its full or optimal corro-sion resistance for the particular oxygen contentof the part, but only approximately 30% whenthe cooling rate is reduced to 200 C/min (360F/min). The data in Fig. 5.32 were obtainedfrom 316L parts sintered in hydrogen at varioustemperatures, dewpoints, and cooling rates.Their densities were under 6.6 g/cm3, and theiroxygen contents ranged from approximately1300 to 2200 ppm. It is not known how severelyoxygen content affects the position or shape ofthese curves. The upper critical cooling temper-ature in Fig. 5.32 is derived from the redoxcurve for 316L. It denotes the temperature atwhich cooling must start, at the latest, for avoid-ing reoxidation.

Thus, with rapid cooling, the extent of silicondioxide formation is minimized and corrosionresistance maximized. As the part cools, the oxi-dation rate of silicon decreases, and at

sufficiently low temperatures, the oxide layerformed contains less silicon and more chromium;in other words, it contains the elements that havea beneficial effect regarding the formation of thepassive layer.

Misconceptions regarding the correct dewpointfor sintering stainless steels still linger in theindustry (Ref 36). It is clear from the preceding

that the basis for sintering stainless steels is themore demanding redox equilibria for the siliconcontained in a stainless steel and not those forchromium (Fig. 5.14).

Figure 5.33 (Ref 13) shows cooling rate/dew-point curves for three austenitic stainless steels.The 316LSC is a tin-copper-modified 316L; SS-100 is a high-chromium, high-nickel stainlesssteel. As mentioned earlier, higher-alloyed steelsappear to be less sensitive, or more forgiving,because lower cooling rates are sufficient tokeep the surfaces free from reoxidation.

Lei et al. (Ref 38) confirm the important effectof cooling rate; 304LSC sintered at 1250 C(2282 F), for 45 min in H2 or 83%H2-17%N2(dewpoint 35 C, or 31 F) had passive cur-rent densities that increased by over 2 orders ofmagnitude when the cooling rate was changedfrom fast to slow. Such a big change cannot beattributed to a change in internal surface area as a

result of different cooling rates. The latter is onlysmall. A more probable interpretation is reoxida-tion of the outer surfaces as a result of slowcooling, in accordance with the examples shownin Fig. 5.32. The large increase of the passivecurrent density due to the formation of surfaceoxides illustrates the effect of metallurgicaldefects on electrochemical characteristics. As

500

400

300

200

100

0

Coolingrate,

C/min

30 40 50 70

Dewpoint of sintering atmosphere,C

(


21/42


mentioned earlier, many investigators in the pasthave failed to account for such defects and thenconcluded, mistakenly, that their low corrosion-resistance properties were due to the presence ofpores, crevice corrosion, or both.

It is appropriate to mention that the two auto-motive large-volume uses of sintered stainlesssteelsantilock brake sensor rings and exhaustflangesare based on high-temperature sinter-ing (>1200 C, or >2192 F) in a low-dewpointatmosphere of hydrogen but without acceler-ated cooling. Their oxygen contents aretypically between 1500 and 2000 ppm. Thisillustrates that maximum corrosion resistance isnot always necessary for successful use.Furthermore, in applications where some sur-face wear occurs readily, or in a strong enoughacidic environment, shallow surface defectssuch as oxides can disappear with time.

For wrought stainless steels, it is known that

rough surfaces from rolling and drawing opera-tions, or from pickling and passivation treatments,increase the tendency for pitting, as do surfaceoxides formed during annealing and weldingoperations. Only in the so-called bright annealingof wrought thin-gage stainless steel sheet are pro-cessing conditions similar to those employed inPM processing, namely, rapid cooling in low-dewpoint hydrogen with limited amounts ofnitrogen (Ref 39).

Little and conflicting data are available forthe chemical cleaning of sintered stainless steels(section 9.1 in Chapter 9, Corrosion Testingand Performance). Much of the published infor-

mation relates to stainless steel parts that hadrelatively low corrosion properties. Any benefitsfrom cleaning may not necessarily apply toproperly sintered parts. Also, the presence ofpores makes cleaning in solutions difficult,because of the capillary forces of the pores thattend to retain the cleaning liquid.

Carbon Control: Delubrication andSintering Conditions. A vast amount of litera-ture exists on the subject of chromium carbideprecipitation in wrought stainless steels. Most ofthese data are applicable to PM stainless steelsand are used here where relevant. For sinteredstainless steels, proper delubrication is impor-

tant for keeping carbon levels below where theycan cause sensitization.Thermodynamics and Kinetics Background.

Intergranular corrosion, one of the variousforms of corrosion in stainless steels, arisesfrom excessive amounts of carbon, which can

form chromium-rich carbide precipitates atgrain boundaries. Because these chromium-richcarbides (M23C6) have a higher chromium con-tent than the alloy, chromium in the surroundingmatrix, that is, next to the grain boundaries(Fig. 5.34) (Ref 40), is depleted to below the levelnecessary to maintain passivation. Chromium car-bide itself is not susceptible to rapid corrosion.

Figure 5.35 (Ref 41) illustrates chromium car-bide precipitates in sintered 316L for variouscarbon levels and typical (i.e., slow) commercialcooling conditions.

Austenitic Stainless Steels.At low carbon lev-els (Fig. 5.35a), the austenitic structure revealsdesirable clean and thin grain boundaries andample twinning; at intermediate carbon levels(Fig. 5.35b), so-called necklace-type chromium-rich carbide precipitates (if present, nitrogen canparticipate in the precipitation) are visible in thegrain boundaries; and at high carbon levels (Fig.

5.35c), the grain boundaries are heavily deco-rated with continuous precipitates. The lattertwo cases give rise to various degrees of inter-granular corrosion.

In austenitic stainless steels, chromium carbideprecipitation occurs in the temperature range of816 to 538 C (1500 to 1000 F). Carbon, due toits small atomic size, diffuses rapidly in the steelmatrix. Hence, during cooling from an elevatedtemperature, for example, after sintering in a typ-ical belt, pusher, or walking beam furnace, anycarbon present in excess of the limit of solubility

Chromium carbide

precipitate

Grain

boundaries

Chromium-depleted zone

Fig. 5.34 Schematic of sensitization. Source: Ref 40.Reprinted with permission from McGraw-Hill





22/42


can easily migrate out of the matrix to the grainboundaries, where it would combine withchromium to form chromium carbide. Figure5.36 (Ref 42) shows the limit of solubility ofcarbon in an austenitic stainless steel as a func-tion of temperature. Based on these data, themaximum amount of carbon in an austeniticstainless steel should be approximately 0.03%.

The diffusion rate of chromium in anaustenitic matrix is not rapid enough to make

up for the chromium lost due to chromiumcarbide precipitation. If, however, the sinteredpart is cooled very rapidly, carbon atoms willnot be able to diffuse out of the matrix andform chromium carbide precipitates butinstead will be held in solution. Hence, bycooling rapidly from an elevated temperature,sensitization can be either prevented or mini-mized in an alloy containing carbon in excessof its limit of solubility.

Critical cooling rates necessary for the preven-tion of sensitization in wrought stainless steels

are commonly depicted as time-temperature-sensitization (TTS) diagrams. Figure 5.37 (Ref42) is an example of a set of TTS diagrams forfive 18Cr-9Ni austenitic stainless steels withdifferent carbon contents.

According to Fig. 5.37, a steel containing0.08% C must be cooled through the sensitiza-tion range in less than approximately 30 s inorder to avoid chromium carbide precipitation.A steel containing only 0.03% C, however, maybe cooled through the same temperature rangein approximately 50 min without riskingsensitization. For constituents that eitherincrease or decrease the tendency for carbideprecipitation, see Ref 43 and 44.

Figure 5.38 (Ref 45) shows the effects of car-bon content and cooling rate on intergranularcorrosion for hydrogen-sintered 316 parts thathad been prepared with various amounts of lubri-cants and with various delubrication conditions

Fig. 5.35 Microstructures of type 316L stainless steel sinteredin hydrogen at 1150 C (2100 F) (glyceregia). (a)

Carbon is 0.015%; thin and clean grain boundaries. (b) Carbon is0.07%; necklace-type chromium-rich carbide precipitates in grainboundaries. (c) Carbon is 0.11%; continuous chromium-rich car-bide precipitates in grain boundaries. Source: Ref 41

1100

1000

900

800

700

600

500

400

300

Temperature,C

0 0.02 0 .04 0.06 0.08 0.10

2000

1800

1600

1400

1200

1000

800

600

Temperature,F

+ M23C6

Solubility limit of

carbon in

austenite

Carbon content, %

Fig. 5.36Solid solubility of carbon in austenitic stainlesssteel. Source: Ref 42





23/42


to produce parts that possessed a wide range ofcarbon contents, from 0.01 to 0.11%.

The curve separating the sensitized fromthe sensitization-free parts represents thecritical cooling rates necessary for variouscarbon contents to avoid intergranular corro-sion. The curve was derived from the TTScurves of Fig. 5.37 by drawing the tangentsfrom 1260 C (2300 F), the sintering temperatureused in this experiment, to the time minima ofthe various carbon-level curves. This allowsone to calculate an average cooling rate foreach carbon level. Good agreement betweenwrought and sintered stainless steel data confirmsthe applicability of wrought stainless steel datato sintered stainless steels.

The most reliable means to prevent sensitiza-tion in austenitic stainless steels is to restrict thecarbon content to 0.03% maximum. Alloys thusmodified are designated as L grades. Inwrought stainless steels, L grades are recom-mended for applications requiring welding and/orthermal cycling. For PM stainless steels, Lgrades are recommended if sintering is performedin typical commercial sintering furnaces withtheir slow cooling rates. All stainless steel pow-ders destined for conventional compaction andsintering are also of the L-grade designations,because of their superior compacting properties.However, as mentioned in section 3.1.3 inChapter 3, Manufacture and Characteristics ofStainless Steel Powders, for 304L and 316Lwithhigh nickel contents close to their upper limits,safe maximum carbon contents are only approxi-mately 0.02%.

Ferritic Stainless Steels. The phenomenon of

intergranular corrosion in ferritic stainless steelsdiffers somewhat from that in austenitic stain-less steels. The limit of solubility of carbon ismuch lower in ferritic stainless steels, and thediffusion rates of interstitials are much higherdue to their body-centered cubic (bcc) structure.As illustrated in Fig. 5.39 (Ref 46), these char-acteristics require very fast cooling rates inorder to prevent sensitization.

900

800

700

600Temperat

ure,

C

Temperature,

F

1652

1472

1292

1112

0.5 1 5 10 50 100

Minutes

C = 0.08%

C = 0.06%C = 0.05%

C = 0.03%

C = 0.02%

Fig. 5.37 Time-temperature-sensitization diagrams for five18Cr-9Ni austenitic stainless steels with different

carbon contents. Source: Ref 42. Reprinted with permission ofJohn Wiley & Sons, Inc.

Fig. 5.38 Effect of carbon content and cooling rate on intergranular corrosion of hydrogen-sintered 316. IG, intergranular. Source:Ref 45. Reprinted with permission from MPIF, Metal Powder Industries Federation, Princeton, NJ

0.12

0.1

0.08

0.06

0.04

0.02

0

Carbonlevel,%

10 102

102

103

103 104

Cooling rate, C/min

Sensitization

No sensitization

10

Critical cooling rate

Based on Fig. 2 (Ref 43)

No IG attack

Minimal IG attack

Widespread IG attack

Cooling rate, F/min





24/42


Ferritics with intermediate levels of interstitialsalso exhibit serious loss of ductility in addition toloss of corrosion resistance. As a result, itbecomes necessary to have either a very low car-bon plus nitrogen level of approximately 0.02%or to use a strong carbide former (stabilizer), suchas niobium, titanium, or tantalum, which forms amore stable carbide in preference to chromiumcarbide, thereby preventing sensitization. Withthe introduction of argon-oxygen decarburization(AOD) and with vacuum and electron beamrefining, it became possible to produce wroughtstainless steels possessing much lower levels ofinterstitials. Before the advent of AOD, the easewith which wrought ferritic stainless steels (i.e.,430 and 434) could be sensitized, together withductile-to-brittle transitions occurring aboveambient temperature, had limited their use. Theaddition of stabilizers has been a common prac-tice for wrought ferritic alloys intended for

applications requiring welding or exposure toelevated temperatures. The control of interstitialsto very low levels, with and without the use ofstabilizers, has led to the development of thehigh-chromium superferritics with good tough-ness, stress-corrosion resistance, and generalcorrosion resistance.

Among sintered PM stainless steels, only onestabilized grade is featured thus far in the MPIFand ASTM standards, namely, ferritic 409L, con-taining niobium as a stabilizer. However, Samalet al. (Ref 47) have shown that it is possible toobtain the equivalent of sintered superferritics byusing niobium as a stabilizer, by sintering (at

1148 C, or 2100 F) in a low-dewpoint atmos-phere of hydrogen, and by employing rapid

cooling. Although some of the niobium reactswith nitrogen during delubricating in nitrogenand with carbon from the lubricant, leading toincreased carbon and nitrogen levels, the amountof niobium was sufficient to precipitate theseinterstitials as carbides and nitrides and toachieve superior corrosion resistances.

Of the various stabilizers (titanium, tanta-lum, niobium) used in wrought stainless steelsto cope with higher carbon contents and com-bat sensitization, only niobium has been usedthus far with some success (409Nb, 434Nb ) insintered stainless steels. In wrought stainlesssteels, niobium carbide-stabilized steels havebeen found to be more resistant to intergranu-lar corrosion than titanium carbide-stabilizedsteels (Ref 48). Titanium and tantalum, proba-bly because of their higher oxygen affinities,form objectionable surface oxides duringwater atomization.

In contrast to the face-centered cubicaustenitic stainless steels, the diffusion rate ofchromium atoms in the bcc ferritic matrix isapproximately 100 times faster. Because of this,a ferritic stainless steel can be cured of its sensi-tized condition by a suitable annealing stepbetween approximately 704 and 954 C (1300and 1750 F). Replenishment of chromium-depleted regions can be satisfactorily achievedand corrosion resistance restored, despite thepresence of chromium carbides along the grainboundaries (Ref 49).

Delubrication. In wrought and cast stainlesssteels, carbon control is accomplished during

melting. In sintered PM stainless steels, the car-bon content is determined not only by the carboncontent of the powder but also by its lubricantand the delubrication and sintering conditions.Of these three, the lubricant contribution is themore complex. It is described in some detail,because it has been the cause for many under-performing sintered stainless steels in the past.

Ideally, for H2 and H2-N2 sintering atmos-pheres, a lubricant should be completelyremoved from the part by complete combustioninto volatile constituents during delubrication. Inpractice, however, combustion and volatilizationare incomplete, and at least a small portion of the

lubricant typically decomposes into carbon andother organic constituents. Lack of control caneasily increase the carbon content of a stainlesssteel to above the 0.03% limit and sometimes toas much as 0.1%. Moyer (Ref 50) has discussedthe delubrication and sintering conditions on the

900

800

700

600

500

400

300

Temperature,

C

102 101 1 10 102 103 104 105

Time, s

Austenitic stainless steel

Ferritic stainless steel

End sensitization

Begin sensitization

Fig. 5.39 Time-temperature-sensitization curves for austeniticand ferritic stainless steels of equivalent chromium

content. Source: Ref 46





25/42


efficiency of lubricant removal in 316L powdercompacts. Saha and Apelian (Ref 51) describedan empirical model and closed-loop controlsystem for delubrication. The degrees ofvolatilization and decomposition depend onmany factors, including part density, heat-up rate,dewpoint, furnace atmosphere, and gas flowrates. In the early years of stainless steel partsproduction, when much of the sintering was indissociated ammonia at 1120 to 1150 C (2050 to2100 F), lubricant removal was usually accom-plished in the preheat zone. Decomposition of thelubricant caused carbon contents to exceed0.03%. Because the low-temperature sinteringconditions were not conducive to lowering thecarbon content during sintering, part fabricatorsresorted to delubricating the parts separately inair, for approximately 30 min for small parts andlonger for larger parts. While this procedurereduced the carbon content to more acceptable

levels, it also caused oxidation (Fig. 5.40).This oxidation is lessened or avoided in dis-sociated ammonia. Figure 5.41 (Ref 41) shows

similar data for Acrawax as a lubricant. Incomparison to lithium stearate, Acrawax hascleaner burn-off characteristics, but, as mentionedin Chapter 3, Manufacture and Characteristics ofStainless Steel Powders, it does not impart thecompressibility advantage of lithium stearate.

Oxidation begins before complete lubricantremoval, even when delubrication is per-formed in dissociated ammonia. It appearsimpossible under these conditions to obtainmaximum carbon removal without additionaloxidation. Though oxides formed at low tem-peratures are more easily reduced duringsintering in a reducing atmosphere than thoseformed at very high temperatures during wateratomization, the goal is to keep oxidation aslow as possible.

Delubrication should always be viewed in thecontext of sintering. With higher sintering tem-peratures (>1205 C , or >2200 F), the reaction

between residual oxygen and carbon is morecomplete, and delubricating is therefore pre-ferably completed in a reducing atmosphere.

Fig. 5.40 Effect of delubrication temperature on oxygen, carbon, and weight loss of 316LSC parts of two densities (6.0 g/cm 3,dashed lines; 6.6 g/cm3, solid lines), lubricated with 1% lithium stearate and delubricated for 30 min in (a) air and

(b) dissociated ammonia (DA) (unpublished data)





26/42


As Fig. 5.41 shows for several stainless steelslubricated with 1% Acrawax and pressed to

green densities of 6.5 to 6.7 g/cm3

, delubricat-ing in dissociated ammonia prevents anysignificant oxidation up to 510 to 538 C (950 to1000 F). Although carbon removal under theseconditions is not yet at its maximum, and carboncontent is still >0.03%, sintering at a highertemperature will lower the carbon content tobelow 0.03%. It is in part because of these rela-tionships that parts sintered at high temperaturesexhibit better corrosion resistances.

Martensitic Stainless Steels. Of the variousfamilies of stainless steels, the martensiticstainless steels have the highest carbon contents,sometimes exceeding 1.2%. In this case, the

carbon function is to form martensite andprimary carbides that endow these steels withtheir hardness, strength, and abrasion-resistanceproperties for which they are known and used.Structural requirements limit the chromiumcontent of these steels; thus, their corrosionresistances are limited.

5.2.4 Sintering of Stainless Steelsin Hydrogen-Nitrogen Gas Mixtures

The primary goal in sintering stainless steels inH2-N2 mixtures is to achieve corrosion resist-ance equal or superior to sintering in hydrogen,in combination with markedly improvedstrength. In recent years, there has been a shiftfrom dissociated ammonia to hydrogen andvacuum sintering. This was clearly the result ofthe increasing emphasis on corrosion resistanceand the difficulty in achieving good corrosionresistance with dissociated ammonia. However,as is clear in the following, the shift to hydrogenand vacuum may be unfortunate in view of thelower cost of nitrogen-containing atmospheresand, more importantly, in view of the potency ofnitrogen to markedly increase the pitting resist-ance of a stainless steel at very low cost. In moststudies, this beneficial effect of dissolved nitro-gen has not been observed, because it wasovershadowed by the negative effect of Cr2Nprecipitation that causes sensitization andintergranular corrosion. However, as is seen, theuse of lower-nitrogen-content atmospheres,such as 90H2-10N2 instead of 75H2-25N2, cou-pled with appropriate cooling rates, readilyestablishes these benefits.

Like carbon, nitrogen has a strong affinity tochromium, and its absorption from the sinteringatmosphere can be exploited for increasingstrength and hardness of stainless steels.Nitrogen-containing sintering atmospheres aremainly used for austenitic stainless steels, where

nitrogen up to approximately 0.3% does notpromote sensitization with correct processing. Itis therefore superior to carbon as a means ofincreasing strength, particularly yield strength.Strengthening is caused by the lattice expansionof the phase (austenite) from the dissolvednitrogen, as well as the precipitation of finelydivided Cr2N. The latter is also beneficial to thefatigue properties but detrimental to corrosionresistance.

Sintering temperature and dewpoint require-ments for effective oxide reduction andinterparticle bonding are similar to those ofhydrogen sintering. Figures 5.42 (Ref 16) and

5.43 illustrate good and bad Auger composition-depth profiles of 316L parts, sintered indissociated ammonia under good and unaccept-able conditions.

While the part in Fig. 5.42 had excellent cor-rosion resistance, that in Fig. 5.43 was veryinferior. The temperature-dewpoint conditions

Fig. 5.41 Effect of delubricating temperature on (a) oxygencontent and (b) carbon content of stainless steel

parts (6.5 to 6.7 g/cm3), lubricated with 1% Acrawax and delubri-cated for 30 min in dissociated ammonia. Source: Ref 41





27/42


for Fig. 5.43 were to the left of the redox curvefor SiO2/316L in Fig. 5.14. As a result, the partpicked up oxygen during sintering. The oxygenprofile in Fig. 5.43 is extended to 2000 s of sput-tering, that is, approximately 30 times the valueof a properly sintered material. Under properreducing conditions, however, the challenge ofachieving good corrosion resistance in nitrogen-containing atmospheres is related to the controlof nitrogen.

Nitrogen Control during Sintering. For opti-mal sintering of stainless steels in H2-N2mixtures, that is, for exploiting the beneficialeffects of nitrogen for both strengthening andimproving corrosion resistance, it is importantto understand their equilibrium solubilities withnitrogen.

Thermodynamic Relationships: NitrogenSolubility of Stainless Steels.According to Zitterand Habel (Ref 52), the solubility of nitrogen in

austenitic chromium and chromium-nickel steelsis determined, on one hand, by the solution ofgaseous nitrogen in the matrix and, on the otherhand, by the precipitation of dissolved nitrogenas chromium nitride, Cr2N. The neglect of thefact that there are two equilibria to consider hasled to erroneous data in the literature. The inter-action of the two relationships leads to amaximum solubility for nitrogen, which dependsonly slightly on the chromium content but

Fig. 5.42 Auger composition-depth profile of 316L partsintered in dissociated ammonia at 1177 C (2151

F). Dewpoint 40 C (40 F). Source: Ref 16. Reprinted withpermission from MPIF, Metal Powder Industries Federation,Princeton, NJ

60

40

20

0

Concentration,%

0 100 200 300 400 500 2000 4000 6000 8000

Sputter time, s

O

Fe

Cr

Ni

C

Si

N

Mo

1000 s correspond to about

240 atomic layers

S or Cl observed in small

concentration on surface

Fig. 5.43 Auger composition-depth profile of 316L part sintered for 20 min at 1110 C (2030 F) in dissociated ammonia ofdewpoint 30 C (22 F). Oxygen content of sample was 0.39%





28/42


1

101

Nitrogen,wt%

10 9 8 7 6

Reciprocal absolute temperature, 104/K

800 900 1000 1100 1200 1300 1400

Cr in %

Cr in %

14

18

2214

14

18

18

22

22

Equilibrium with:

Cr2N

Dissociated ammonia (25% N2)

90% H2-10% N2

Temperature, C

Fig. 5.44 Solubility of nitrogen in chromium-nickel steels in equilibrium with gaseous nitrogen or nitrides, depending ontemperature and partial pressure of nitrogen. Source: Adapted from Ref 52

which shifts markedly to higher temperatureswith increasing chromium content and increas-ing partial pressure of nitrogen (Fig. 5.44)(Ref 52).

Nickel reduces the nitrogen solubility for theequilibrium with Cr2N (Ref 53).

For sintering in dissociated ammonia, withchromium contents of 14, 18, and 22%, the max-imum solubility temperatures are, according toFig. 5.44, 947, 1057, and 1163 C (1737, 1935,and 2125 F), and their nitrogen solubilities are0.39, 0.43, and 0.46%, respectively. For sinter-ing a 22% Cr austenitic stainless steel in90H2-10N2, the minimum sintering temperatureis approximately 1105 C (2021 F). Below thistemperature, chromium becomes fully nitridedto form insoluble Cr2N.

The negative temperature coefficient for sol-ubility is unusual for metals and accounts, inpart, as is seen subsequently, for the preferenceof sintering at higher temperatures in nitrogen-containing atmospheres. On cooling, Cr2Nbegins to precipitate at these temperatures ofmaximum solubility. Above these tempera-tures, only dissolved nitrogen exists in thesolid phase.

Dautzenberg (Ref 54) has shown the effect ofnitrogen content on ultimate tensile strength andelongation of 304L stainless steel. The amountof nitrogen absorbed follows known phase

equilibria in accordance with Sieverts law; thatis, nitrogen absorption is proportional to thesquare root of the partial pressure of nitrogen inthe sintering atmosphere (Ref 55).

The negative effect of nitrogen on ductilityand impact strength is also apparent from themechanical properties tables in Chapter 7,Mechanical Properties.

Miura and Ogawa (Ref 56) used mechanicalalloying of elemental powder mixtures withFe10N to produce high-nitrogen chromium-nickel and chromium-manganese stainlesssteel powders with nanostructures having awide composition range for austenite stability.The addition of AlN or NbN as dispersionagents allowed them to fully consolidate themechanically alloyed materials by hot rollingnear 1173 K (1652 F), while still retainingnanostructures.