Hot Working Superalloys Research Summary

Simulating Microstructural Evolution

during the Hot Working of Alloy 718
Martin C. Mataya

INTRODUCTION
The simulation of microstructural evolu- The ability of secondary forgers to produce complex-shaped products with uniform
tion during the primary breakdown of pro- properties and microstructure hinges on the ability of primary processors to transform
duction-sized alloy 718 ingots and billets by as-cast ingots with highly nonuniform, coarse, as-cast grain structures into intermedi-
radial forging was accomplished in the labo- ate-sized product (billets, bar, slab, plate) with a uniform fine-grain structure. How-
ratory via multiple-stroke axial compression ever, hot working of even the simplest shapes, such as the reduction of ingot to bar or
testing of cylindrical specimens. The dwell plate, can produce widely differing microstructures from position to position. This
or hold time between strokes was varied to occurs because gradients in the critical intrinsic variables that control microstructural
simulate the deformation-time history for evolution (e.g., temperature, strain, and strain rate) are inevitably generated in the
three different locations along the radial- work piece during hot working.
forging work piece: lead-end, mid-length, The definition of the values of these critical parameters with regard to position in the
and tail-end positions. The microstructural work piece and time during hot working, coupled with a constitutive description of
evolution varied with simulated work piece their effect on recrystallization, recovery, and grain growth, is required to develop
position. Static, rather than dynamic, re- widely applicable models that can be applied in the design of primary breakdown
crystallization was responsible for the ob- schedules. The former can be provided via large plastic-strain finite-element analysis
served grain-size refinement, and its repeti- of the deformation and the latter by laboratory hot-deformation studies, such as those
tive occurrence during consecutive dwell conducted during the past 15 years at the Advanced Steel Processing and Products
periods resulted in the maintenance of a fine- Research Center at the Colorado School of Mines in Golden, Colorado.1–8
grain microstructure during multiple-stroke Typically, alloy 718 is produced by vacuum induction melting followed by vacuum
deformation sequences. For comparison, the arc remelting into a 406 mm diameter cylindrical ingot. Ingots and billets are subse-
total plastic strain was also applied in a quently reduced in diameter by working in a temperature range of 800°C to 1,150°C.
single-stroke test. The single- and multiple- A large finished size may be 225 mm in diameter. When processed via radial forging,
stroke techniques gave differing microstruc- work-piece speed through the forging dies may be varied from 2–5 m/min. The
tural results, indicating that multiple-stroke reduction in cross-sectional area can be varied 10–50 percent per pass (15–20% is
testing is necessary in modeling microstruc- common). The strain rate of deformation is approximately 1.0 s–1 .
tural evolution during primary breakdown. During radial forging, the work piece moves in one direction through the die set,
with both the lead and tail ends passing completely through the dies to receive the full
benefit of the applied reduction. The work piece is then reversed so that the tail end,
with respect to the first pass, is the first to be reduced during the second pass. A 2.5 m
long ingot traveling at 2.5 m/min. through the inlet side of a radial forging machine
would have a cycle time for the first complete pass of approximately one minute. A
dwell time of 10 s between material exit from the dies and reentry is common.
Figure 1 shows a schematic of the deformation dwell-time history during radial
forging for three different positions in a work piece (lead, mid-length, and tail, with
respect to the first pass) for the initial four passes. As shown, after the lead end is
deformed, one minute will elapse before the work piece exits the dies; 30 s will lapse
for the mid-length position before exit, and the tail end exits at zero time. Predictions
of the evolved microstructure after the first and possibly the second pass could be
made with existing data from single-stroke compression tests,3 however, the micro-
structural evolution during subsequent passes can not be predicted. Hence, multiple-
stroke compression testing was pursued, and the results are discussed here. For more
Figure 1. Simulated deformation-dwell time
detail on the experiments, refer to the sidebar “Experimental Methods.”
histories for three positions along the length SINGLE-STROKE VERSUS FOUR-STROKE TEST RESULTS
of a work piece processed by radial forging
during primary breakdown. Cycle time per Figure 2 shows typical four-stroke stress-strain (σ-ε) curves simulating deforma-
pass is one minute. tion-dwell time sequences for the mid-length position in the work piece. Also shown
are the results of the corresponding
Table I. A Comparison of Recrystallized Grain Size and Volume Percent Recrystallized
between Single-Pass and Four-Pass Compression Tests of Alloy 718* single-stroke tests (triangular symbols).
At 950°C, significant softening occurs
Recrystallized Volume Percent during the single-stroke test (beginning
Grain Size µ
(µ m) Recrystallized
Deformation at a ε of approximately 0.2) as compared
Temperature (°C) Strain Single-Pass Four-Pass Single-Pass Four-Pass to the four-stroke test. Such softening is
950 0.56 9 9 7 <1 often related to the progression of dy-
950 0.92 10 7 30 6 namic recrystallization; however, in this
1,050 0.56 41 22 100 89 case softening is due to some other source,
1,050 0.92 37 19 100 96 probably an increase in test-sample tem-
1,150 0.56 114 50 100 100
1,150 0.92 98 50 100 100
perature via deformational heating.
Heating in alloy 718 deformed under
* Initial 254 µm average grain diameter.
identical conditions after a ε of 1.0 was

18 JOM • January 1999

a b c
Figure 2. Flow curves of alloy 718 at (a)
found to be approximately 60°C.2 Assuming that the magnitude of deformational 950°C, (b) 1,050°C, and (c) 1,150°C deformed
heating is linearly proportional to ε, the amount of heating after a single-pass ε of 0.56 in four strokes (solid curves) and in a single
is estimated to be 34°C. The reduction in flow stress resulting from sample heating can stroke (discontinuous curve) to a total strain of
be approximated by the variation in flow stress with temperature obtained from the 0.56.
three four-pass σ-ε curves (Figure 2), assumed here to represent the isothermal
behavior because of the periodic 70 s dwell time, which facilitates dissipation of the
heat of deformation during the test. Comparing the four-pass curves, an increase in
temperature from 950°C to 1,050°C causes alloy 718 to soften by about 225 MPa, from
approximately 500 MPa at 950°C to 275 MPa at 1,050°C at a ε of 0.56. Thus, a 34°C
increase due to deformational heating during the single-pass test at 950°C would be
expected to result in a softening increment of about 76 MPa at ε = 0.56, assuming a linear

EXPERIMENTAL METHODS
The composition (in wt.%) of the alloy 718 used in lubricant to the sliding face during compression. Weis2 and 1,150°C. The strain applied to the specimen during
this investigation was, in addition to nickel, 18.99 Fe, verified the effectiveness of this technique and deter- each stroke of an individual four-stroke sequence was
18.32 Cr, 5.18 Cb, 3.05 Mo, 0.98 Ti, 0.55 Al, 0.35 Co, mined an optimum well geometry for the specimen used either 0.14 or 0.23. The total strain for each sequence
0.21 Si, <0.2 Ta, 0.12 Mn, 0.06 Cu, 0.044 C, 0.009 P, in this investigation. Two glass lubricants were em- was either 0.56 or 0.92 (corresponding to reductions in
<0.005 Mg, 0.0043 B, 0.004 Sn, <0.0005 Pb, <0.0005 ployed to accommodate the range of testing tempera- height of the compression sample of 43% and 60%).
Ag, and <0.5 ppm Bi. The starting ingot, provided by tures: Delta Glazes 349M for 950°C and 347M for The strain levels were chosen to simulate those typically
Carpenter Technology Corporation, was homogenized, 1,050°C and 1,150°C (Atcheson Chemical Company). applied during a four-pass radial forging sequence used
then radial-forged and rolled to 16.5 mm diameter The combination of hard-smooth dies, lubricant well to reduce 406 mm diameter ingot to either 305 mm or
bar. The bar was solution heat treated at 1,250°C for geometry, and glass lubricant provided excellent lubri- 254 mm diameter billet.
7.5 hours and water quenched to simulate the coarse cation for the duration of the compression test, demon- The hold or dwell time between strokes was varied to
grain microstructure typically found in billet material strated by the absence of barreling of the free surface on simulate the strain-dwell time history characteristic of
during primary breakdown. Figure A shows the solution- the compressed test sample (Figure B). Barreling oc- either the first end of the work piece to enter the radial
treated microstructure, with an average grain diameter curs when compression platens are colder than the forge machine on the first pass (hereafter referred to as
of about 250 µm (ASTM grain size number one). specimen; die chilling of the faces during compression the lead end, independent of the pass number), the mid-
Compression testing was accomplished in a 250 KN causes a restricted flow in the radial direction at the length position (mid-length), or the last end of the work
servohydraulic testing machine (manufactured by MTS faces. The utilization of dies at the same initial tempera- piece to enter on the first pass (tail end). Table 1 lists the
Corporation) outfitted with an electric resistance clam- ture as the work piece, as accomplished in this study, is strain dwell time history for the three positions. The final
shell furnace. The vertical load column was composed essential. Uniform deformation allows the assumption dwell time for each sequence was chosen to simulate
of two opposing Astroloy rams, which were hollowed to of a constant strain throughout the cross section of the that point at which the production work piece exits the
facilitate heating, and flat-smooth (lapped to an 8 RMS specimen; thus, the evolved microstructure can be radial forge machine for the fourth time. For example,
surface finish with opposite faces parallel to within associated with the strain value measured directly from the tail end exits the radial forge first on the fourth pass,
0.0127 mm over a 69.85 mm diameter) SiN compres- the reduction in height of the compression specimen. If before the lead end exits the machine. Therefore, the
sion dies fitted to the end of the rams. barreling occurs, strain varies widely across the com- test specimen simulating the tail end was held for 60 s
Cylindrical compression samples, 12.7 mm diameter pression specimen,8 and reliable correlations between (simulating one pass per minute) before quenching,
by 19.05 mm high, were machined from the 16.5 mm plastic strain and the resulting microstructure can not be while the specimen simulating the lead end was rapidly
rod. The end faces were recessed to form a lubricant obtained. A discussion of the assumptions used in quenched (within 2 s) immediately after the fourth strain
well, which is an effective technique for constraining compression testing and the effect of the deformation cycle was imposed.
heat (often referred to as adiabatic heating) on mea- Single-stroke compression tests were also carried
sured stress-strain properties is provided elsewhere.9 out to examine the difference in microstructural evolu-
Test specimens were loaded onto the bottom die in tion predicted by the two techniques. Upon deformation
the furnace in air, held for 10 min. at the test temperature cessation in the single-stroke tests, samples were held
(specimens required approximately 5 min. to equili- in the furnace at the deformation temperature for 240 s
brate), and compressed uniaxially. The velocity of the and then quenched in order to simulate the total dwell
moving die was varied in order to apply deformation at time for the mid-length position in the four-stroke test.
a constant strain rate. Die velocity, displacement, un-
loading, dwell time, and reloading were controlled by
computer software.
Four-stroke hot compression tests were performed
at a constant true-strain rate of 1.0 s–1 at 950°C, 1,050°C,

Table 1. Dwell-Time History Used to Simulate

Various Work-Piece Positions during Radial
Forging*
Dwell Time (s)
Position One Two Three Four
200 µm Lead End 130 10 130 2
Figure A. The starting microstructure of alloy Mid-Length 70 70 70 30
718 for radial-forging simulation study. Tail End 10 130 10 60 Figure B. Typical hot compression specimens
(Kalling’s etch.) * Dwell 1 is applied after Pass 1, etc. (left) prior to and (right) after deformation.

1999 January • JOM 19

 variation of flow stress with temperature. The calculated value is consistent with the
difference in flow stress (approximately 75 MPa) between the single-pass and four-
pass curves at the final strain.
At 1,050°C and 1,150°C, the single-pass and four-pass σ-ε curves are nearly identical
with respect to the flow-stress levels achieved, except for the transient behavior
encountered at low ε upon reloading during the four-pass tests. At 1,050°C, static
recrystallization causes a loss of strength during the dwell periods. At 1,150°C, aging
during each dwell period provides a strengthening effect characterized by the devel-
opment of an upper and lower yield point. Hardening during the 1,150°C dwell masks
the softening from static recrystallization.
Table I gives the measured values of the volume-percent recrystallized (Vrex) and
recrystallized grain size (Drex); Figure 3 shows the evolved microstructures of both the
single-pass and four-pass materials. At 950°C, the two test methods produced ap-
proximately the same Drex (7–10 µm); however, the single-pass method resulted in a
greater Vrex than the corresponding four-pass method—7% vs. less than 1% after 0.56
ε and 30% vs. 6% after 0.92 ε, respectively. Similarly, at 1,050°C, Drex and Vrex differ
between the two methods, with larger
values observed after single-stroke com-
pression. At 1,150°C, both materials are
fully recrystallized, with the single-
stroke material exhibiting twice the re-
crystallized grain size as the multiple-
stroke material.
Increased recrystallization in the
single-pass material is attributed at least
in part to a higher sample temperature
after deformation. This is due to
deformational heating, which favors
static recrystallization—the dominant
recrystallization mechanism for the de-
formation history applied here.3–5 In-
creased recrystallization in single-pass
material may also be due in part to the
a 25 µm b 25 µm extensive deformation that is applied
without much chance for recovery. This
recovery is afforded to the unre-
crystallized portion of the microstruc-
ture in the multiple-stroke test. Greater
retained work, in the form of an unre-
covered dense dislocation substructure
in the unrecrystallized matrix, would
favor static recrystallization during the
240 s hold after deformation. In com-
parison, static recovery of the unre-
crystallized matrix, assumed to occur to
some degree during the dwell periods of
four-stroke tests, would reduce the driv-
ing force for recrystallization in these
areas.10,11 With greater recovery and less
driving force for recrystallization in the
multiple-stroke test samples, one would
c 25 µm d 25 µm expect a larger Drex . On the contrary,
finer recrystallized grain sizes are ob-
served after multiple-stroke testing.
SOFTENING AND
RECRYSTALLIZATION DURING
DWELL PERIODS
Softening between passes during hot
working is usually due, in part, to the
progression of static recrystallization.
The amount of softening that occurs
during the dwell periods in the four-
stroke tests was calculated with
S = [(σf – σr)/(σf – σi)]100 (1)
where S is the percentage of the maxi-
mum softening that could be realized, σf
e 25 µm f 25 µm is the final flow stress prior to unload-
Figure 3. The microstructure of alloy 718 deformed to a total strain of 0.92 by (a) single stroke at ing, σr is the 0.2% offset yield stress upon
950°C, (b) four strokes at 950°C, (c) single stroke at 1,050°C, (d) four strokes at 1,050°C, (e) single reloading after a dwell period, and σi is
stroke at 1,150°C, and (f) four strokes at 1,150°C. the yield stress for the initial loading

20 JOM • January 1999

cycle. Table II gives the yield stress and softening values for the various deformation
temperatures, strain levels, and work piece positions. At 950°C, the yield stress
increases from strain cycle to subsequent strain cycle, indicating a general work-
hardening behavior rather than softening due to either static recrystallization or static
recovery. In fact, the calculated softening does not vary with strain cycle, strain level,
or dwell time, but remains constant at approximately 55%. Figure 3b shows that little
recrystallization is observed after a four-stroke compression test at 950°C. Most of the
observed softening must then be due to static recovery. In this case, the lack of
dependence of S on hold time indicates that the majority of the softening (55%) occurs
upon unloading, probably by back-stress-induced dislocation movement rather than
thermally activated dislocation annihilation and polygonization.
At 1,050°C, yield stress and softening vary dramatically (Table II). For example, a
four-stroke simulation of the work-piece lead end with an applied strain increment per
cycle of 0.14 shows that the yield stress increases from an initial value of 197 MPa to
242 MPa by the third strain cycle, then drops from 242 MPa to 217 MPa with the
application of the 130 s dwell prior to the fourth strain cycle. The softening remains
constant at about 50% for the first two dwell periods (a level not associated with the
occurrence of any static recrystallization), and then increases to 73% for the third dwell
period. If it is assumed that approximately 50% of the total softening is partitioned to
static recovery as described, then about half of the microstructure (23% as compared Figure 4. Yield stress vs. total accumulated
to the remaining 50% of available softening) recrystallizes statically during the third strain for alloy 718 obtained from four-stroke
dwell period. Softening values further indicate that about half of the microstructure for compression tests.
the 0.23 strain/cycle lead-end specimen statically recrystallizes during the first dwell
period. In fact, a microstructural analysis of two companion specimens (not shown),
one quenched within 2 s after the first 0.23 strain increment and the other after a
130 s dwell at temperature, showed that more than 30% of the structure recrystallized
during the dwell period. A similar test was performed for the second dwell period of
the tail end (0.23 strain/cycle). In this case, 70% of the structure recrystallizes statically
as compared to a predicted value of about 68%.
At 1,150°C, the amount of recovery and recrystallization that occurs is masked by
strengthening during the dwell periods (Figure 2c and Table II). As a result, a softening
analysis from the flow curves can not be used to estimate the volume fraction
recrystallized during each dwell period. Strengthening, which is manifested by the
establishment of a sharp yield point followed by a large stress relaxation, may be due
to short-range ordering and/or solute pinning of an existing dislocation substructure.
The data in Tables 1 and II show that strengthening and the re-establishment of the
yield drop progresses rapidly between 10 s and 70 s, and little further change is noted
for the 70–130 s interval.
Figure 4 shows the variation in yield stress (from Table II) with strain for the mid-
length 0.14 strain/cycle sequence at 950°C and 1,050°C. As discussed, the yield stress
of material deformed at 950°C increases monotonically with the application of succes-
sive strain cycles, indicating the lack of recrystallization during the dwell periods. At
1,050°C, however, a critical amount of strain is reached (approximately 0.15) and static
recrystallization proceeds during the dwell periods. A concomitant reduction in the
yield stress upon reloading occurs and is manifested in Figure 4 as a decrease in the
1,050°C curve. Also shown are the 950°C lead-end and tail-end 0.23 strain/cycle data.

Table II. Yield Stress (0.2% offset) for Alloy 718 After Each Simulated Radial Figure 5. The percent softening and percent-
Forging Pass* volume recrystallized for each individual dwell
Yield Stress (MPa) time vs. total accumulated strain obtained
Position in from four-stroke compression tests of alloy
Work Piece Strain Per Pass Pass 1 Pass 2 Pass 3 Pass 4 718. Note the mid-length 1,050°C, 0.14 strain
950°C per pass softening curve was calculated as-
Lead End 0.14 196 269 (60) 321 (54) 315 (60) suming an initial yield stress of 197 MPa, a
0.23 198 294 (65) 339 (51) 322 (54) value typical for lead- and tail-end samples,
Mid-Length 0.14 196 277 (60) 306 (62) 322 (58) because the measured value for the mid-
0.23 196 304 (62) 326 (56) 330 (55) length sample appeared to have an anoma-
Tail End 0.14 189 289 (51) 300 (62) 340 (52) lously high value (Table II).
0.23 196 313 (60) 317 (57) 347 (50)
1,050°C
Lead End 0.14 197 240 (49) 242 (47) 217 (73)
0.23 205 223 (74) 223 (72) 215 (77)
Mid-Length 0.14 208 244 (51) 228 (74) 221 (82)
0.23 201 227 (64) 216 (78) 215 (76)
Tail End 0.14 197 250 (36) 222 (69) 225 (64)
0.23 203 239 (51) 213 (84) 214 (80)
1,150°C
Lead End 0.14 185 175 (—) 136 (—) 170 (—)
0.23 174 173 (—) 151 (—) 162 (—)
Mid-Length 0.14 168 173 (—) 171 (—) 171 (—)
0.23 168 163 (—) 166 (—) 162 (—)
Tail End 0.14 174 152 (—) 170 (—) 154 (—)
0.23 176 148 (—) 167 (—) 153 (—)
* Percent softening shown in parentheses.

1999 January • JOM 21

 The apparent oscillation about the 950°C mid-length curves is attributed to the
alternating length of the dwell period, being either shorter or longer than the 70 s dwell,
which is characteristic of the mid-length position. For example, yield stress measured
after a 10 s hold is slightly higher than the associated yield stress after a 70 s hold and
likewise for the 70 s and 130 s holds. This behavior is consistent for all test conditions.
It appears that a small portion of the static recovery at 950°C is time-dependent. The
majority of the static restoration, however, occurs very rapidly upon unloading. At
1,050°C, the time-dependent phenomena that can account for the oscillating behavior
(not depicted in Figure 4) includes static recrystallization and static recovery.
Figure 5 shows the variation in softening with strain for 950°C and 1,050°C (upper
curves) and the percent volume recrystallized as measured metallographically on a
series of companion samples deformed at 1,050°C (lower curve). Softening for the
1,050°C mid-length material increases with accumulated strain up to a value of about
50%, then decreases slightly. The maxi-
mum softening recorded, approximately
75%, shows that as much as 50 vol.% of
the structure can recrystallize during a
70 s dwell time. With higher applied
strain (0.23 strain/cycle) as much as 70%
of the structure could recrystallize dur-
ing a single 130 s hold. Once again, at
950°C, softening is independent of the
test parameters and for the most part is
time-independent. At 1,050°C, the mag-
nitude of softening after the first pass at
a strain of 0.14 is unexpectedly below
that observed for 950°C. This may be
explained by the appearance of a mild
yield point on the 1,050°C flow curves in
Figure 2b. Such age hardening during
the dwell period would naturally reduce
a 25 µm b 25 µm the magnitude of the measured soften-
ing. This is consistent with the complete
elimination of measured softening at
1,150°C due to rapid age hardening.
MICROSTRUCTURAL
EVOLUTION
A detailed evolution of microstruc-
ture occurring during the multiple-stroke
test sequence corresponding to Figure
2b (1,050°C mid-length position with 0.14
strain/cycle) is presented in Figure 6.
The final microstructure is compared to
others that developed at higher and lower
temperatures and at different simulated
positions in the work piece in Figure 7
and Table III. The microstructure present
c 25 µm d 25 µm

e 25 µm f 25 µm g 25 µm
Figure 6. The microstructural evolution of alloy 718 at 1,050°C in the mid-length position. (a) After a strain of 0.14 (first stroke) plus a 70 s dwell. Note
the large recrystallized grain on the right, which nucleated at a high-angle grain boundary. The arrow on the right points to the “ghost” appearance
of a prior high-angle grain boundary and the arrow on the left indicates the location of a prior triple point. (b) After a second-strain increment of
0.14 (second stroke) followed by a rapid water quench. (c) Second stroke plus 70 s dwell. Note the large unrecrystallized grain (upper left) and the
prior “ghost” triple point (arrow). (d) Third stroke plus rapid quench. (e) Third stroke plus 70 s hold. (f) Fourth stroke plus rapid quench. (g) Fourth stroke
plus 30 s hold. Note in (d) and (f) that repeated recrystallization occurs within the previously recrystallized material. (Kalling’s etch.)

22 JOM • January 1999

 A COMPARISON OF ALLOY 718 BEHAVIOR TO TYPE 300
SERIES STAINLESS STEELS
Both single- and four-stroke tests were conducted after one increment of 0.14 strain). The recrystallization 718 is much larger (by a factor of ten), the recrystallized
on 304L with a starting grain size equivalent to alloy results in significant softening and a drop in the yield grain size of the two materials after one pass at 1,150°C
718 (250 µm, produced by a 1 h, 1,150°C heat stress from 123 MPa for the second pass to 95 MPa for and a dwell is nearly equivalent—200 µm and 160 µm,
treatment). Table 2 compares the microstructures of the third pass and 84 MPa for the fourth stroke. Upon respectively. Also plotted in Figure C is the final recrys-
alloy 304L after single-pass and four-pass compres- completion of the fourth pass, the material is 75% tallized grain size of the wrought 250 µm grain diameter
sion testing. Values of Drex in single-stroke 304L recrystallized (Table 4). By comparison, alloy 718 in the 718 and 304L alloys. Although the final grain size values
material are about twice those in four-stroke material. equivalent condition is only 6% recrystallized. after four passes at 1,150°C are nearly equivalent for
In addition, single-pass material exhibits a signifi- Table 4 also shows that the recrystallized grain size the four materials (cast 305, wrought 304L, cast 718,
cantly higher Vrex at the two lower temperatures. In in 304L is coarser than in alloy 718. The differences in and wrought 718), the recrystallized grain size is dra-
general, the observed differences between the val- behavior at T = 950°C may be due to the presence of matically finer for alloy 718 at 950°C. The grain refine-
ues of Drex and Vrex in 304L produced by the two test fine particles in alloy 718 that are not present in 304L. ment may be associated with the presence of delta-
methods are qualitatively similar to those observed It should be noted that regarding the magnitude of the phase Ni3Nb in alloy 718, which is stable to about
for alloy 718 (Table I). The similarity reinforces the measured softening due to recovery, Perttula and 1,040°C.1
proposition that microstructural evolution during con- Karjalainen14 report that by increasing the plastic-strain Roberts et al., 16 investigating two austenitic stainless
ventional hot-working processes is better simulated offset used in Equation 1 from 0.2% to 5%, softening in steels, showed that the rate of dynamic recrystallization
in the laboratory via multiple-stroke rather than single- alloy 304 due to recovery is effectively eliminated, varies inversely with initial grain size, Do, and that at a
stroke testing. providing a direct relationship of S to volume fraction given temperature and strain rate, the critical strain for
A comparison of the two alloys at T = 950°C shows recrystallized. dynamic recrystallization decreases significantly as Do
that for equivalent deformation conditions, recrystal- The distinct difference in behavior of the two materi- is decreased. The critical strain for static recrystalliza-
lization and grain growth are much more sluggish in als at 950°C is substantially moderated as the deforma- tion is likewise reduced. Thus, as coarse-ingot grain
alloy 718. For example, after a total strain of 0.92, both tion temperature is increased. At 1,150°C, values for structure is reduced via static recrystallization during
the single-pass and four-pass alloy 718 samples have Vrex and Drex for the two alloys are nearly equivalent. This the dwell times between the initial passes, the critical
much lower values of Vrex than the corresponding behavior is consistent with the apparent correlation of strain for dynamic recrystallization would also be re-
304L samples—30% vs. 99% (single-pass) and 6% deformation behavior between as-cast alloy 718 and duced. Guimaraes and Jonas17 apparently observed full
vs. 75% (four-pass), respectively. The corresponding as-cast 304L at high temperature.4 Thus, at high tem- dynamic recrystallization in alloy 718 during a strain of
recrystallized grain sizes are 10 µm vs. 34 µm (single- perature (T > 1,050°C) it appears that 304L could be 0.7 at 1,090°C.
pass) and 7 µm vs. 22 µm (four-pass). This difference employed as a surrogate material for modeling the Although not cited, the initial grain size was approxi-
originates from the fact that 304L begins to recrystal- microstructural behavior of alloy 718 during hot-work mately 150 µm, similar to that observed for alloy 718
lize at lower values of strain. For example, recrystal- processing. ingot after one pass (Figure C). If dynamic recrystalliza-
lization takes place after the second and third passes Figure C compares the predicted grain size of alloy tion occurs during latter passes in the multiple-stroke
(note the large amount of softening reported in Table 718 ingot material after an initial strain of 0.25 and a tests, it would be favored at the mid-length position by
3 prior to the start of the third and fourth passes, S = 70 s hold4 to that observed for as-cast type 305 stainless the finer grain size present at that location and at the
78% and 85%, respectively, compared to a value for steel for an equivalent strain and a 60 s hold.15 Although work-piece ends because of the shorter dwell time,
S of approximately 43% for unrecrystallized material the approximate starting grain size of the as-cast alloy which results in a greater cumulative strain that is
available to drive dynamic recrystallization during sub-
Table 2. A Comparison of Recrystallized Grain Size and Volume Percent Recrystallized of Single-Pass sequent deformation. However, from the extensive
and Four-Pass Compression Tests of Alloy 304L static recrystallization observed in the sequence of
µ m) microstructures presented in Figure 6 for 1,050°C mul-
Recrystallized Grain Size (µ Volume Percent Recrystallized
Deformation tiple-stroke compression, it appears that this is the
Temp. (°C) Strain Single-Pass Four-Pass Single-Pass Four-Pass dominant recrystallization mechanism controlling the
950 0.56 45 27 — 20 evolution of microstructure in alloy 718 as well as 304L.
950 0.92 34 22 99 75 In fact, significant softening of the flow curves, which
1,050 0.56 74 43 100 95 would be indicative of dynamic recrystallization, was
never observed.
1,050 0.92 68 39 100 100
1,150 0.56 130 66 100 100
1,150 0.92 116 52 100 100

Table 3. Yield Stress (0.2 % offset) for Type 304L Stainless Steel After Each Simulated Radial
Forging Pass*
Postion in Yield Stress (MPa)
Work Piece Strain Per Pass Pass 1 Pass 2 Pass 3 Pass 4
950°C
Mid-Length 0.14 64 125 (43) 126 (52) 115 (62)
0.23 64 123 (53) 95 (78) 84 (85)
1,050°C
Lead End 0.14 54 52 (102) 80 (78) 54 (100)
Mid-Length 0.14 54 66 (80) 54 (100) 54 (100)
Mid-Length 0.23 54 52 (103) 53 (101) 56 (97)
1,150°C
Mid-Length 0.14 46 42 (112) 42 (111) 44 (104)
0.23 48 36 (128) 44 (110) 43 (111)
* Percent softening shown in parentheses.

Table 4. The Microstructure of Type 304L Stainless Steel After Each Simulated Radial
Forging Four-Pass Sequence
0.14 Strain Per Pass 0.23 Strain Per Pass Figure C. Recrystallized grain size versus the
number of deformation steps in compression
Deformation Position in Volume Percent Average Grain Volume Percent Average Grain
µ m) µm) for as-cast alloy 718,4 as-cast alloy 305,15 and
Temperature (°C) Work Piece Recrystallized Diameter (µ Recrystallized Diameter (µ
wrought 718 and 304L. For alloys 718 and
950 Mid-Length 20 REX 27 75 REX 22 304L, the strain rate, dwell time, and strain per
1,050 Mid-Length 95 REX 43 100 REX 39 stroke are 1 s–1, 70 s, and 0.23. The strain rate,
1,150 Mid-Length 100 REX 66 100 REX 52 dwell time, and strain per stroke values for
REX—Recrystallized grains. alloy 305 are 10 s–1, 60 s, and 0.28.

1999 January • JOM 23

 in the sample after applying the first
stroke and quenching was deformed and
showed no evidence of recrystallization
(photograph not shown). During the
subsequent 70 s dwell time at 1,050°C
and prior to the imposition of the second
stroke, approximately 11% of the de-
formed microstructure recrystallized
(Figure 6a). The average recrystallized
grain diameter is 45 µm. During the fol-
lowing (second) stroke plus rapid
quench cycle, the volume fraction that
was recrystallized remains virtually un-
changed, however, about one-half of the
inherited recrystallized volume recrys-
tallizes again (Figure 6b). The new grains
a 100 µm b 100 µm
are approximately 11 µm in diameter.
The new grains nucleate preferentially
at the initial coarse-grain, high-angle
grain boundaries; at the previously re-
crystallized high-angle grain bound-
aries; at the grain boundaries separating
recrystallized from unrecrystallized ma-
terial; and at the surface of coarse inter-
metallic inclusions. The second wave of
recrystallization probably occurs stati-
cally during the period of time begin-
ning immediately after deformation and
ending at the quench, because there is
more time for recrystallization to take
place during this period (1–2 s) than
during deformation (0.15–0.25 s). This
c 100 µm d 50 µm assumption is consistent with the trans-
mission electron microscopy presented
Figure 7. The final evolved microstructures after four strokes, 0.14 strain/stroke of alloy 718. (a)
1,050°C, mid-length position corresponding to Figure 6g; (b) 1,150°C, mid-length position; (c–d) by Weis et al. 4 for the occurrence of static
1,150°C, lead-end position. Note the “ghost” triple point (center arrow) in (d), which was the recrystallization within 2 s after a strain
nucleation site for a coarse grain typically observed in (c), and the fine grain necklaces which of 0.25 at 1,150°C, but is contrary to the
nucleated on the “ghost” high-angle grain boundaries (upper arrow) in (d) of the coarse grain. findings of Camus et al. 12
Figure 6c shows the microstructure of
Figure 6b after a 70 s dwell time at 1,050°C. A comparison of the two micrographs
shows that the fine 11 µm diameter grains coarsen to approximately 33 µm. The
remaining recrystallized grains generated during the first dwell period coarsen from
45 µm to about 83 µm during the second dwell. Thus, the microstructure, at this point,
has a duplex grain size. The total volume fraction recrystallized increases from 11% to
43%, with 38% of the total volume being statically recrystallized during the second 70
s hold. The significant increase in volume fraction recrystallized (shown in the lower
curve of Figure 5) is consistent with the relatively large drop in yield stress observed
for the third strain cycle in Figure 2b and the large value of softening reported for the
start of Pass 3 in Table II and Figure 5 (upper solid curve).
Figure 6d shows the microstructure after the third strain cycle followed by a rapid
Table III. The Microstructure of Wrought 254 mm Grain Diameter Alloy 718 After Each Simulated Radial Forging Four-Pass Sequence
0.14 Strain Rate Per Pass 0.23 Strain Rate Per Pass
Deformation Position in Volume Percent Average Grain Volume Percent Average Grain
Temperature (°C) Work Piece Recrystallized µ m)
Diameter (µ Recrystallized µ m)
Diameter (µ
950 Lead End UNREX 3 REX 10
Mid-Length <1 REX 9 6 REX 7
Tail End 2 REX 5 17 REX 10
1,050 Lead End 84 REX Duplex (50, 8) 97 REX Duplex (40, 8)
Ave Duplex 26 Ave Duplex 16
UNREX 507 UNREX 224
(Max UNREX 1054) (Max UNREX 267)
Mid-Length 89 REX 22 96 REX 19
UNREX 372 UNREX 187
(Max UNREX 635) (Max UNREX 508)
Tail End 88 REX 27 99 REX 19
UNREX 440 UNREX 190
(Max UNREX 635) (Max UNREX 210)
1,150 Lead End 100 REX Duplex (120, 25) 100 REX Duplex (105, 25)
Ave Duplex 56 Ave Duplex 37
Mid-Length 100 REX 50 100 REX 50
Tail End 100 REX 57 100 REX 61
REX—recrystallized grains; UNREX—worked unrecrystallized grains.

24 JOM • January 1999

quench. Again, the presence of new fine recrystallized grains in coarser previously
recrystallized material shows that recrystallization occurred for a third time. In this
case, about 30% of the previously recrystallized volume recrystallized again, probably
during the 2 s time period after the third strain cycle and before the quench for reasons
previously discussed. The new fine grains are about 17 µm in diameter, while the
previously recrystallized grains from the prior two dwell periods have an average
diameter of 75 µm.
Figure 6e shows the microstructure of Figure 6b after a 70 s dwell period (the third
dwell applied in this sequence). During this period, the 17 µm grains grow to 48 µm,
and the total volume percent recrystallized increases from 43% to 78%. It is estimated
from microstructural observations that almost 50% of the total volume recrystallizes
statically during this hold time. The presence of new fine grains after the fourth strain
cycle demonstrates that previously recrystallized material recrystallizes again after
the applied strain cycle (Figure 6f). Figure 6g shows the microstructure of Figure 6f
after the final dwell period of 30 s. This is the simulated microstructure of the mid-
length position in a four-pass radially forged 305 mm diameter billet immediately after
the final pass. Subsequent cooling of the radially forged billet is not simulated here.
Figure 7 shows the final microstructure, the same as Figure 6f, at lower magnifica-
tion and with an etchant used to reveal grain boundaries. The structure is 89% References
recrystallized with an average grain diameter of 22 µm. Eleven percent of the 1. D.A.K.C. Chang, “The Influence of Ingot Heterogeneity
microstructure remains unrecrystallized, having an average grain diameter of 327 µm Upon the Hot Deformation of Alloy 718,” M.S. thesis T-3371,
Colorado School of Mines, Golden, CO (1987).
and an occasional large unrecrystallized grain on the order of 635 µm. 2. M.J. Weis, “The Hot Deformation Behavior of As-Cast
Alloy 718,” M.S. thesis T-3382, Colorado School of Mines,
The final microstructure produced by the various sequences studied in this inves- Golden, CO (1987).
tigation are provided in Table III. In general, deformation at 950°C is ineffective with 3. M.C. Mataya et al., “Grain Refinement during Primary
Breakdown of Alloy 718,” 29th Mechanical Working and Steel
respect to grain refinement, and the final structure obtained depends strongly on the Processing Conference Proceedings, vol. XXV (Warrendale, PA:
position in the work piece. For example, for strain cycles of 0.23, the tail end is 17% ISS, 1987), pp. 235–248.
4. M.J. Weis et al., “The Hot Deformation Behavior of an As-
recrystallized, while the lead end is only 3% recrystallized. Cast Alloy 718 Ingot,” Alloy 718—Metallurgy and Applica-
At 1,050°C, the mid-length 0.14 strain/cycle material has a finer grain size (22 µm vs. tions, ed. E.A. Loria (Warrendale, PA: TMS, 1989), pp. 135–
154.
26 µm) compared to the lead end, probably due to the shorter dwell times during which 5. M.C. Mataya and D.K. Matlock, “Effects of Multiple
grain growth is rapid (Figure 7a). The unrecrystallized grain size is also finer due to the Reductions on Grain Refinement During Hot Working of
Alloy 718,” Alloy 718—Metallurgy and Applications, ed. E.A.
greater extent of recrystallization (89% vs. 84%, respectively). Another difference Loria (Warrendale, PA: TMS, 1989), pp. 155–178.
6. D.R. Nielsen, “Grain Size Control in Ring-Rolled Alloy
associated with work-piece position is that the lead end has a distinct duplex morphol- 718,” M.S. thesis T-4450, Colorado School of Mines, Golden,
ogy with coarse grains surrounded by necklaces of fine grains. The application of a CO (1993).
7. D.R. Nielsen et al., “Grain Size Control in Ring-Rolled
greater stain/cycle results in finer recrystallized and unrecrystallized grain sizes and 718,” Proceedings of International Symposium on Superalloys
a greater volume fraction recrystallized (Table III). Similarly, Domblesky et al.,13 who 718, 625, 706 and Derivatives, ed. E.A. Loria (Warrendale, PA:
TMS, 1994), pp. 373–392.
also investigated simulated radial forging of alloy 718 via multiple-stroke compres- 8. A.C. Perry, “Hot Deformation Processing and Micro-
sion testing, found that either a greater strain per pass or an increased dwell time structural Evolution of Alloy 617,” M.S. thesis MT-SRC-095-
023, Colorado School of Mines, Golden, CO (1995).
between passes reduced the number of passes required for complete recrystallization. 9. M.C. Mataya and V.E. Sackschewsky, “Effect of Internal
At 1,150°C, both low-strain and high-strain four-stroke cycles result in a uniform, Heating during Hot Compression Testing on the Stress-
Strain Behavior and Hot Working Characteristics of Alloy
fully recrystallized, fine-grained microstructure at the work-piece mid-length posi- 304L,” Metall. Trans. A, 25A (1994), pp. 2737–2752.
tion (Figure 7b). The same is true for the tail-end position (Table III). Figure 7c shows 10. M.C. Mataya, E.L. Brown, and M.P. Riendeau, “Effect of
Hot Working on Structure and Strength of Type 304L Auste-
that, similar to the 1,050°C lead-end material described above, the 1,150°C lead-end nitic Stainless Steel,” Metall. Trans. A, 21A (July 1990), pp.
material also has a distinct duplex microstructure consisting of fine-grain necklaces 1969–1987.
11. M.C. Mataya, M.J. Carr, and G. Krauss, “The Effect of
surrounding coarse grains. The coarse recrystallized grains apparently nucleated and Hot Working on Structure and Strength of a Precipitation
Strengthened Austenitic Stainless Steel,” Metall. Trans. A,
grew at an earlier generation of high-angle grain boundaries and triple points (Figure 15A (February 1984), pp. 347–368.
7d, see arrow). The newer fine recrystallized grains nucleated at the high-angle grain 12. G. Camus, B. Pieraggi, and F. Chavet, “Hot Deformation
and Recrystallization of Inconel 718,” Formability and Metal-
boundaries and triple points of the current coarse recrystallized grains, forming the lurgical Structure, eds. A.K. Sachdev and J.D. Embury
observed necklaces. Thus, two generations of “ghost” prior grain boundaries are (Warrendale, PA: TMS, 1986), pp. 305–324.
13. J.P. Domblesky et al., “Prediction of Grain Size During
apparent in Figure 7d. The coarse grains could have originated during the first 130 s Multiple Pass Radial Forging of Alloy 718,” Proceedings of
hold or during the final 130 s hold. The fact that repeated recrystallization is so International Symposium on Superalloys 718,625, 706 and Vari-
ous Derivatives, ed. E.A. Loria (Warrendale, PA: TMS, 1994),
prevalent at 1,050°C, and that it occurs quite rapidly (within 2 s) gives credence for the pp. 263–272.
latter scenario—coarse grains nucleating and growing during the second 130 s hold, 14. J.S. Perttula and L.P. Karjalainen, “Recrystallisation Rates
in Austenite Measured by Double Compression and Stress
and the fine-grain necklaces nucleating and growing after cessation of the fourth strain Relaxation Methods,” Matls. Sci. and Tech., 14 (July 1998), pp.
cycle and before the quench. 2–6.
15. R. Staker and N.J. Grant, “The Effects of Strain, Strain
Rate, and Temperature on Grain Refinement and Hot Work-
OTHER CONSIDERATIONS WITH RESPECT TO ability of Type 305 Stainless Steel,” Mater. Sci. Engr., 75
MICROSTRUCTURAL GRADIENTS (1985), pp. 137–150.
16. W. Roberts, H. Boden, and B. Ahlblom, “Dynamic Recrys-
tallization Kinetics,” Met. Sci., 13 (Mar.-Apr. 1979), pp. 195–
In the modeling of microstructural evolution during breakdown, a number of other 205.
effects need to be considered. For example, during the forging of large cross sections, 17. A.A. Guimaraes and J.J. Jonas, “Recrystallization and
Aging Effects Associated with the High Temperature Defor-
strain does not penetrate efficiently from surface to center in the work piece. If the mation of Waspaloy and Inconel 718,” Metall. Trans. A, 12A
center of an ingot receives less strain than the surface during working, then grain (September 1981), pp. 1655–1666.
18. D.K. Matlock and D.A. Burford, “An Experimental Cor-
refinement at the center should lag the surface. Matlock and Burford18 showed that the relation of Plane Strain Deformation Zone Geometry and
gradient in strain from surface to centerline in deformed copper ingots, which resulted Forming Loads,” J. Applied Metalworking, 4 (4) (January 1987),
pp. 301–305.
in a distinct and steep gradient in microstructure during subsequent annealing, was 19. D. Zhao et al., “Three-Dimensional Computer Simulation
correlated to the deformation-zone geometry (∆). For example, material near the ingot of Alloy 718 Ingot Breakdown by Cogging,” Proceedings of
Int’l. Symposium on Superalloys 718,625, 706 and Various De-
surfaces was fully recrystallized while the center material was unrecrystallized for a rivatives, ed. E.A. Loria (Warrendale, PA: TMS, 1997), pp.
∆ greater than four, where ∆ is measured by dividing work-piece through-thickness by 163–172.
20. D. Zhao, S. Guillard, and A.T. Male, “High Temperature
die contact length. Therefore, steeper gradients in strain and, thus, in the evolved Deformation Behavior of Cast Alloy 718,” in Ref. 19, pp. 193–
microstructure from the surface to the center of alloy 718 are expected during the early 204.
21. R.A. Jaramillo et al., “Evaluation of an Inconel Alloy 718
stages of ingot breakdown, and milder gradients are expected during the latter stages. Microstructural Evolution Model,” in Ref. 19, pp. 257–266.
22. B. Marty et al., “Recrystallization and Work-Hardening
Because strain induces self-heating of the work piece, gradients in strain will result Prediction During Forging Process of Inconel 718,” in Ref. 19,
in gradients in temperature. Heating the work-piece surface relative to its center by pp. 331–342.

1999 January • JOM 25

 localized strain will be moderated by radiation, convection, and conduction. Whether
the surface is hotter or cooler than the center depends on the specifics of the process
and the rate of deformation. Zhao et al.19 recently observed that the center of alloy 718
ingot was more extensively recrystallized than the surface, and the observed gradient
could be explained on the basis of the results from finite-element modeling, which
predicted that the surface was chilled with respect to the center. In addition, high
friction at the tool work-piece interface in the model restricted flow at the surface. The
observed gradient in microstructure in this case is explained by the combination of
lower temperature and reduced strain at the surface. In the analysis, microstructural
results from single-stroke compression tests of alloy 718 were employed with appar-
ently reasonable success because the predicted microstructural gradient agreed fairly
well with the actual gradient. In a separate study, Zhao et al.20 also found that the
microstructure created in the single-stroke compression test samples of alloy 718 was
significantly different than that observed after a double-stroke test, an observation
consistent with the conclusions of this study.
NUMERICAL MODELING
The particular usefulness of finite-element modeling is that the evolution of the
critical intrinsic variables during hot working that control microstructural evolution
(e.g., strain, strain rate, and temperature) are predicted from point to point on a fine
scale. Simulations of microstructural evolution are, however, critically based on
ABOUT THE AUTHOR
developing and utilizing appropriate relationships that can trace the evolving micro-
structure with time as the hot-working process proceeds. To this end, Jaramillo et al.21
Martin C. Mataya earned his Ph.D. in metal- employed a numerical model for the recrystallization rate of alloy 718, based on the
lurgical engineering at Marquette University critical intrinsic variables and the measured activation energy for recrystallization in
in 1974. He is currently a research professor a finite-element model, to predict evolved microstructure in hot-rolled plate. Marty et
at the Advanced Steel Processing and Prod-
ucts Research Center at the Colorado School
al.22 proposed a different numerical model for the recrystallization rate in alloy 718
of Mines and a staff member in the Materials based on evolving dislocation density as modified by the various mechanisms of
Science and Technology Division at Los recovery. Of particular interest, the model was used to predict the progression of the
Alamos National Laboratory. Dr. Mataya is a volume fraction recrystallized in a multiple-stroke compression test, and the predic-
member of TMS. tions compared well with actual test results. Coupling the results of multiple-stroke
For more information regarding the Advanced
compression testing with appropriate numerical models of metallurgical behavior
Steel Processing and Products Research Cen- and subsequent employment of these models in finite-element simulations of complex
ter, contact D. Matlock at (303) 273-3775. processing schemes is now at hand.

News & Update (Continued from page 11.)

tus of the Society, concluding that although Societyline
the Society is not facing a cash flow
problem, operational revenue is be- Posthumous Honor ........................................................... Robert A. Laudise, a member of TMS since 1989, was
coming insufficient to cover all of the honored by the Federation of Materials Societies posthumously with its first Past Presidents’ Award for Excellence
activities the Society desires to do. in Leadership in the Materials Professional Community. The award was presented at a reception on December 9,
Hence, a plan must be developed to 1998, at the National Press Club in Washington, D.C. At the time of his death on August 20, 1998, in New York,
live within the Society’s means. Dr. Laudise was adjunct chemical director of Bell Labs, Lucent Technologies; adjunct professor of materials
science at Massachusetts Institute of Technology; and adjunct professor of ceramics at Rutgers University. He
• Public and Governmental Affairs Di-
joined Bell Labs in 1956 after receiving his Ph.D. in inorganic chemistry from the Massachusetts Institute of
rector Bhatka Rath announced that Technology. Dr. Laudise was president of the Federation of Materials Societies in 1995–1996.
the plans for establishing a congres-
sional fellow for TMS did not work St. Louis Local Section ..................................................... The TMS/Society for Mining, Metallurgy, and Explora-
out because no cosponsors could be tion St. Louis Local Section will hold meetings with the following speakers: January 15, Richard M. Whiting,
identified. president and chief officer of Peabody Holding; February 19, David Rice, president and chief executive officer of
• Student Affairs Director Tresa Pollock Savage Zinc; and March 19, Allan Weakly, Henry Krumb lecturer. On April 17, a dinner dance will be held. For more
reported on an exploratory meeting information on any activities, contact Walt J. Spitzmiller, Big River Zinc Corporation, Route 3 and Monsanto
with other key materials societies, in- Avenue, Sauget, Illinois 62201; (618) 274-5000.
cluding ASM, American Ceramics In Memory of ..................................................................... TMS has received notification of the deaths of the
Society, Materials Research Society, following members: Fred Demoney (Butte, Montana), year of election—1950; Earl Hayes (Laurel, Maryland), 1962;
the Iron and Steel Society, and the Orville Kimball (Oak Ridge, Tennessee), 1955; Robert Laudise (Berkeley Heights, New Jersey), 1989; Ian
Metallurgical Society of the Canadian MacGregor (New York), 1959; James McClain (Albany, Oregon), 1962; Niranjan Parikh (Highland Park, Illinois),
Institute of Mining, Metallurgy, and 1980; Jason Ryder (Huntingdon, Pennsylvania), 1998; Emery Valyi (Katonah, New York), 1941; Frank Wong
Petroleum to consider forming Mate- (Bronkville, New York), 1954.
rials Clubs at universities. The clubs Missing Members .............................................................. Due to address changes, TMS is unable to locate the
would replace the societies’ individual following members: I. Bernstein, Tufts University; Howard Chang; and Irene Purdy. If you know where these
student chapters. individuals are, contact TMS Customer Service Center, 420 Commonwealth Drive, Warrendale, Pennsylvania
• TMS Foundation President Robert 15086; (724) 776-9000, ext. 270; fax (724) 776-3770; e-mail csc@[Link].
Wagoner reported the approval of UMI InfoStore .................................................................... UMI InfoStore™, which has provided reprint versions
Lionel Kimerling as Foundation presi- (via microfilm reproduction) of out-of-print TMS publications, has closed all operations as of December 31, 1998.
dent, effective February 1999 for a Readers can still obtain copies of out-of-print TMS articles, issues, or publications through Linda Hall Library in
three-year term. Two new subcom- Kansas City, Missouri. The research library can be accessed via the World Wide Web at [Link].
mittees were formed: the Program
Committee and Fundraising Com- MetaLeaders ..................................................................... The following TMS members have received recogni-
mittee. tion in the form of awards and grants: Ali S. Argon (elected 1965), American Society of Mechanical Engineers’ Nadai
The next Board meeting will be held on Medal; Geoffrey Brooks (1962), Iron & Steel Society’s 1998 Ferrous Metallurgy Grant; John W. Cahn (1960),
General Electric Distinguished Lectureship in Materials Science and Engineering; Michael F.X. Gigliotti (1979), the
Wednesday, March 3, at the San Diego
GE Research and Development Center’s Coolidge Fellowship Award; and Joseph I. Goldstein (1959), honorary
Marriott Hotel and Marina in San Diego, member of Microbeam Analysis Society.
California.

26 JOM • January 1999