FINAL REPORT

Aprü 1,1997 - December 31, 2000

Grant No. F49620-97-1-0234

PROCESSING AND PROPERTIES OF Al-Li-Cu-X ALLOYS

Submitted to:

Dr. Craig S. Hartley, Program Manager

Metallic Materials Program
Directorate of Aerospace and Mechanical Sciences
Air Force Office of Scientific Research
801 N. Randolph Street
Arlington, VA 22203-1977

Submitted by:

Edgar A. Starke, Jr.

Earnest Oglesby Professor

SEAS Report No. UVA/525869/MSE01/101

DISTRIBUTION STATEMENT A
Approved for Public Release
Distribution Unlimited

DEPARTMENT OF MATERIALS SCIENCE AND ENGINEERING

SCHOOL OF
ENGINEERING Ü
& APPLIED SCIENCE
University of Virginia
Thornton Hall
Charlottesville, VA 22903
20010316 073
 UNIVERSITY OF VIRGINIA
School of Engineering and Applied Science
The University of Virginia's School of Engineering and Applied Science has an undergraduate
enrollment of approximately 1,500 students with a graduate enrollment of approximately 600. There are
160 faculty members, a majority of whom conduct research in addition to teaching.
Research is a vital part of the educational program and interests parallel academic specialties.
These range from the classical engineering disciplines of Chemical, Civil, Electrical, and Mechanical and
Aerospace to newer, more specialized fields of Applied Mechanics, Biomedical Engineering, Systems
Engineering, Materials Science, Nuclear Engineering and Engineering Physics, Applied Mathematics
and Computer Science. Within these disciplines there are well equipped laboratories for conducting
highly specialized research. All departments offer the doctorate; Biomedical and Materials Science grant
only graduate degrees. In addition, courses in the humanities are offered within the School.
The University of Virginia (which includes approximately 2,000 faculty and a total of full-time student
enrollment of about 17,000), also offers professional degrees under the schools of Architecture, Law,
Medicine, Nursing, Commerce, Business Administration, and Education. In addition, the College of Arts
and Sciences houses departments of Mathematics, Phyics, Chemistry and others relevant to the
engineering research program. The School of Engineering and Applied Science is an integral part of this
University community which provides opportunities for interdisciplinary work in pursuit of the basic goals
of education, research, and public service.
 REPORT DOCUMENTATION PAGE
AFRL-SR-BL-TR-01-
1 hour par ratpor,*«. md

•ubiact to «n'y pmrtv taf Wine » eon^ty wl* • edl»etlon of kifoinwtlon if h dot» not (litplty ■ airt«wlyv»Hd
PLEASE DO *ÜOT RETURN YOUR FORM TO THE ABOVE ADDRESS. .-'
1. REPORT DATE (DD-MM-YYYYJ 2. REPORT TYPE DATES rnx/PRPn «Ssn TBT
2 Feb 2001 Final 4/1/97-12/31/00
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
Processing and Properties of Al-Li-Cu-X Alloys
Sb. GRANT NUMBER
AFOSR Grar" No. F49620-97-1-0234
6c PROGRAM ELEMENT NUMBER

AUTHOR(S) Sd. PROJECT NUMBER

Edgar A. Starke, Jr.

6t. TASK NUMBER

SI. WORK UNIT NUMBER

7. PERFORMING ORGANIZATION NAME«) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION

Department of Materials Science and Engineering REPORT NUMBER
University of Virginia
116 Engineer's Way, P.O. Box 400745
Charlottesville, VA 22904-4745
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESSES) 10. SPONSOR/MONITOR'S ACRONYM(S)
Air Force Office of Scientific Research
801 North Randolf Street
11. SPONSOR/MONITOR'S REPORT
Room 732 NUMBER(S)
Arlington, VA 22203-1977
c
12. DISTRIBUTION/AVAILABILITY STATEMENT AIR FORCE Of •FlCEOF3CIEMT!ri0nCSCAnCH(A
.. O^)
NOTICE OFTRA^ITTALßTiCTHISTeCHNlCALhHr^
HAS BEEN REVIEWED AND IS APPROVED FOR PUBUC RcU
LÄWAFR100-1^ DiSTRlBUT^ilSU^UMiTFn
13. SUPPLEMENTARY NOTES

14. ABSTRACT
The objectives of this research were to identify the mechanisms for the large difference in ductility
between the near peak aged AF/C-489 and AF/C 458 and then develop an aging schedule to
optimize the microstructure for high strength and ductility. Duplex aging treatments produced
significant increases in ductility for AF/C-489, approaching that of AF/C-458. Although strain
localization occurred in both alloys it was more severe in AF/C-489 due to the higher volume
fraction of shearable delta prime and a much larger grain size. Our results show that a
significant increase in ductility of the higher lithium AF/C-489 can be obtained by double aging
and reducing the grain size. The report describes the details of the study and the optimum
iraraicrÄRN^^1"^180091

16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF 18. NUMBER 19a. NAME OF RESPONSIBLE PERSON
a. REPORT b. ABSTRACT e. THIS PAGE ABSTRACT OF
PAGES
19b. TELEPHONE NUMBER Ondud» MM ceo»)

Standard Fom 298 (Rav. 8/98)

Piwcrfetd by AN« M. Z».1S
 Final Report

PROCESSING & PROPERTIES OF AL-LI-CU-X ALLOYS

AFOSR Grant No. F49620-97-1-0234

Dr. Edgar A. Starke, Jr.

Department of Materials Science and Engineering
University of Virginia
116 Engineer's Way
P.O. Box 400745
Charlottesville, VA 22904-4745

INTRODUCTION
The optimization of thermomechanical processing is a paramount consideration for Al-Li-Cu-X
alloys. The microstructural evolution, and therefore the mechanical properties, of these alloys
may be dramatically altered through precise changes in various processing stages. Primarily, the
imposed cooling rate from the solution heat treatment temperature and the amount of introduced
plastic deformation prior to aging have marked effects on both the competitive microstructural
evolution and subsequent mechanical properties. Other important metallurgical factors, such as
texture and grain size, have been shown to dictate the ductility of this class of alloys. It is
necessary to understand each of these thermomechanical-processing steps to truly optimize the
performance of this alloy system.

Although Al-Li-Cu alloys showed initial promise as lightweight structural materials,

implementation into primary aerospace applications has been hindered due in part to their
characteristic anisotropic mechanical and fracture behaviors. The Air Force recently developed
two isotropic Al-Li-Cu-X alloys with 2.1 wt pet Li and 1.8 wt pet Li designated AF/C-489 and
AF/C-458, respectively. The elongation at peak strength was less than the required 5 pet for the
2.1 wt pet Li variant but greater than 10 pet for the 1.8 wt pet Li alloy. The objectives of our
investigations were to first identify the mechanisms for the large difference in ductility between
the AF/C-489 and AF/C-458 alloys and then to develop an aging schedule to optimize the
microstructure for high ductility and strength levels.

MATERIALS
Our study focused on the two Air Force alloys;
AF/C-458: Al - 2.7 Cu - 1.8 Li - 0.6 Zn - 0.3 Mg - 0.3 Mn - 0.08 Zr
AF/C-489 : Al - 2.7 Cu - 2.1 Li - 0.6 Zn - 0.3 Mg - 0.3 Mn - 0.05 Zr
The AF/C-458 alloy was fabricated at Alcoa commercial facilities, while AF/C-489 was a
laboratory alloy from Wright Patterson Air Force Base. These alloys were received as plate of
3/4" and 0.5" thick, respectively. Both alloys were received in the unrecrystallized condition
consisting of 5-10um sub-grains. However, the grain size for the AF/C-489 was considerably
larger than the grain size for the AF/C-458.
EFFECT OF PRECIPITATE AND GRAIN STRUCTURE ON DUCTILITY

The precipitation response for both the AF/C-489 and AF/C-458 alloys was determine by TEM
and DSC, and used to design an improved micro structure through duplex aging for AF/C-489.
Duplex aging of the AF/C-489 alloy resulted in significant increases in ductility by as much as
85% with a small decrease of only 6.5% and 2.5% in yield and ultimate tensile strength,
respectively. No significant improvements were found with either of the duplex or triple aging
practices for the AF/C-458 commercial alloy indicating a large processing window. Quantitative
TEM demonstrated that the precipitation response significantly differed between the duplex
aging for the AF/C-489 and AF/C-458 alloys. The duplex aged AF/C-489 alloy competitively
precipitated more 9" at the expense of Ti whereas no significant changes in precipitation
response occurred for the AF/C-458 alloy. Furthermore, shearable 8' volume fraction was
calculated to be on average -3.0% greater in the AF/C-489 than the AF/C-458 alloy.
Conventional TEM also demonstrated that grain boundary Ti precipitation was unaffected by
duplex aging designed to competitively precipitate more matrix Ti than grain boundary Ti. In
fact, all investigated single and double aged samples contained grain boundary T\.

Conventional TEM, HREM, and SEM fractography demonstrated that formation of shear bands
occurred as a result of the localized shearing of both the 8' and T, strengthening precipitates in
both the AF/C-458 and AF/C-489 alloys. Our results indicated that the low ductility of the AF/C-
489 alloy in comparison to the AF/C-458 alloy is attributed to the increased volume fraction of
shearable 8' coupled with a larger grain size, both of which significantly increase the amount of
strain localization and stress concentrations at grain boundaries. This results in more low energy
intergranular fracture for the AF/C-489 alloy. The details of this phase of our research can be
found in the publication: "The Effect of Processing and Microstructure Development on the Slip
and Fracture Behavior of the 2.1 Wt Pet Li AF/C-489 and 1.8 Wt Pet Li AF/C-458 Al-Li-Cu-X
Alloys," A.A. Csontos and E.A. Starke, Jr., Metallurgical and Materials Transactions A.. Vol.
31A, August 2000, pp. 1965-1976, attached as Appendix I.

THE EFFECT OF PLASTIC DEFORMATION ON COMPETITIVE PRECIPITATION

The role of plastic deformation prior to artificial aging on the microstructural evolution and
mechanical properties of the AF/C 458 was investigated. Induced plastic deformation ranged
from a non-stretched or 0% stretch condition to an 8% stretch, with intermediate stretches of 2%,
4% and 6%. Tensile properties, fractography and quantitative precipitate analysis were acquired
from specimens that were water quenched from a solution heat treatment, immediately stretched
and artificially aged at 150°C. Fractography was investigated through scanning electron
microscopy (SEM). Quantitative transmission electron microscopy (TEM) determined the
variation in precipitate type, number density, size and volume fraction of the major strengthening
precipitates Al2CuLi (T,), Al2Cu (e'/8') and Al3Li (8). Age hardening curves for each level of
mechanical stretch illustrated the enhanced aging kinetics of plastically deformed material.
Quantitative TEM indicated that increasing amounts of pre-age stretch were found to greatly
affect the competitive precipitation kinetics of Ti and 076' in AF/C 458 augmenting the volume
fraction of fine matrix Ti plates and dramatically decreasing the volume fraction of 079' for
isochronal treatments. A quantitative microstructural comparison of specimens exhibiting a
given strength demonstrated that the imposed level of cold work dictated the density, size and
volume fraction of the competing precipitates. The tensile data indicated a trend of increasing
ductility for equivalent yield strengths with the increasing amount of pre-age mechanical stretch
and therefore shorter artificial aging times. The quantitative precipitate data was used with a
computer simulation for yield strength determination. The theoretical simulation reported
calculated yield strengths in good accord with experimental results and can thus be used to
predict the optimum microstructural configuration for high strength. The details of this phase of
our research can be found in the publication: "The Role of Plastic Deformation on the
Competitive Microstructural Evolution and Mechanical Properties of a Novel Al-Li-Cu-X
Alloy," B.M. Gable, A.W. Zhu, A.A. Csontos and E.A. Starke, Jr., Journal of Light Metals, Vol.
1, No. 1, January 2001, pp. 1-14, attached as Appendix II.

EFFECT OF QUENCH RATE ON MICROSTRUCTURE AND PROPERTIES

The objective of this phase of our research was to determine the effect of quench rate on the
microstructure and mechanical properties of the AF/C-458 and AF/C-489 alloys. TEM, SEM,
microhardness and tensile testing were utilized to ascertain these microstructure/property
relationships for both alloys in the T4, T6 and T86 tempers as a function of quench rate.
Subsequent losses in ductility for both alloys in all tempers with decreasing quench rate were
determined to be due to the precipitation of the equilibrium Al2CuLi (Ti) phase along subgrain
and grain boundaries which promoted intergranular fracture. Furthermore, yield and tensile
strengths increased for both alloys in the T4 temper but decreased in the T6 and T86 tempers
with decreasing quench rate. The increased strengths for the T4 condition resulted from the
heterogeneous precipitation of coarse Ti and naturally aged 8' phases. The decrease in yield and
tensile strength for the T6 and T86 tempers were also due to the coarse heterogeneous
precipitation of Ti which denuded regions of Cu thereby reducing the number density of fine
matrix 0" (T6) and Ti (86). Finally, a comparison of the quench sensitivity for both the AF/C-
458 and AF/C-489 alloys indicates that the mechanical properties for both alloys were less
quench rate sensitive than other typical Al-Li-Cu-X alloys. The details of this phase of our
research can be found in the publication: "The Effect of Quench Rate on the Microstructure and
Properties of AF/C-458 and AF/C-489 Al-Li-Cu-X Alloys," A.A. Csontos, B.M. Gable, A.
Gaber and E.A. Starke, Jr., Materials Science Forum. Vols. 331-337, July 2000, pp. 1333-1340,
attach as Appendix III.

PERSONAL SUPPORTED
Aladar A. Csontos Graduate Student, University of Virginia
Brian M. Gable Graduate Student, University of Virginia
Dr. Aiwu Zhu Postdoctoral Scientist, University of Virginia
Dr. Edgar A. Starke, Jr. University and Ogelsby Professor, University of Virginia

DEGREES GRANTED
Brian M. Gable, "The Role of Mechanical Stretch on the Microstructural Evolution of AF/C-
458," Master of Science Degree, University of Virginia, May, 2000.

Aladar A. Csontos, "Microstructural Effects on the Slip & Fracture Behavior of Isotropie Al-Li-
Cu-X Alloys," Doctor of Philosophy Degree in Materials Science and Engineering, University of
Virginia, February 2001.
PUBLICATIONS

"Computer Experiment on Superposition of Strengthening Effects of Different Particles," A.W.

Zhu, A. Csontos and E.A. Starke, Jr., Acta Materialia, vol. 47, June 1999, pp. 1713-1721.

"Need for New Materials in Aging Aircraft Structures," R.J. Bucci, C.J. Warren and E.A. Starke,
Jr., Journal of Aircraft, vol. 37, January-Feburary 2000, pp. 122-129.

"The Effect of Quench Rate on the Microstructure and Properties of AF/C 458 and AF/C 489
Alloys", A.A. Csontos, B.M. Gable, A.F. Gaber and E.A. Starke, Jr., Materials Science Forum,
vol. 331-337, April 2000, pp. 1333-1340.

"The Role of Mechanical Stretch on the Processing-Microstructure-Property Relationships of

AF/C 458", B.M. Gable, A.A. Csontos and E.A. Starke, Jr., Materials Science Forum, vol. 331-
337, April 2000, pp. 1341-1346.

"The Effect of Processing and Microstructure Development on the Slip and Fracture Behavior of
the 2.1 wt pet Li AF/C 489 and 1.8wt pet Li AF/C 458 Al-Li-Cu-X Alloys", A.A. Csontos and
E.A. Starke, Jr., Metallurgical and Materials Transactions A, vol. 31A, August 2000, pp.
1965-1976.
"The Role of Plastic Deformation on the Competitive Microstructural Evolution and Mechanical
Properties of a Novel Al-Li-Cu-X Alloy," Journal of Light Metals, vol. 1, January 2001, pp. 1-
14.
 APPENDIX I
The Effect of Processing and Microstructure Development
on the Slip and Fracture Behavior of the 2.1 Wt Pet Li
AF/C-489 and 1.8 Wt Pet Li AF/C-458 Al-Li-Cu-X Alloys
A.A. CSONTOS and E.A. STARKE, Jr.

Although Al-Li-Cu alloys showed initial promise as lightweight structural materials, implementation
into primary aerospace applications has been hindered due in part to their characteristic anisotropic
mechanical and fracture behaviors. The Air Force recently developed two isotropic Al-Li-Cu-X alloys
with 2.1 wt pet Li and 1.8 wt pet Li designated AF/C-489 and AF/C-458, respectively. The elongation
at peak strength was less than the required 5 pet for the 2.1 wt pet Li variant but greater than 10 pet
for the 1.8 wt pet Li alloy. The objectives of our investigations were to first identify the mechanisms
for the large difference in ductility between the AF/C-489 and AF/C-458 alloys and then to develop
an aging schedule to optimize the microstructure for high ductility and strength levels". Duplex and
triple aging practices were designed to minimize grain boundary precipitation while encouraging
matrix precipitation of the Tt (Al2CuLi) strengthening phase. Certain duplex aged conditions for the
AF/C-489 alloy showed significant increases in ductility by as much as 85 pet with a small decrease
of only 6.5 and 2.5 pet in yield and ultimate tensile strength, respectively. However, no significant
variations were found through either duplex or triple aging practices for the AF/C-458 alloy, thus,
indicating a very large processing window. Grain size and S (Al3Li) volume fraction were determined
to be the major cause for the differences in the mechanical properties of the two alloys.

I. INTRODUCTION pet Cu-2.1 wt pet Li-0.6 wt pet Zn-0.3 wt pet Mn-0.3 wt

IN this era of streamlined economics for the aerospace pet Mg-0.05 wt pet Zr. Unfortunately, the elongation at peak
industry, weight reduction of aerospace vehicles is best strength was lower than the acceptable level of 5 pet for
achieved through decreasing the density of the aircraft's aerospace applications..
structural materials. For aluminum alloys, lithium is added As a result of this research, the Air Force and ALCOA
to reduce the alloy's density with the added benefit of (Alcoa Center, PA) developed a derivative of AF/C-489 with
increasing the alloy's modulus. In fact, the addition of 2 wt 1.8 wt pet Li and 0.09 wt pet Zn, designated AF/C-458,
pet lithium can reduce the density of aluminum by 6 pet with the nominal composition of Al-2.7 wt pet Cu-1.8 wt
and increase the modulus by 12 pet,111 thus, decreasing the pet Li-0.6 wt pet Zn-0.3 wt pet Mn-0.3 wt pet Mg-0.09
system weight and operational costs while increasing the wt pet Zr. This AF/C-458 variant possessed much higher
intrinsic alloy strength. Nevertheless, extensive implementa- elongations, more than double the acceptable limit, while
tion of aluminum-lithium alloys into primary aerospace also maintaining similar isotropic mechanical properties.
structures has been hindered, in part, due to their characteris- Therefore, the objectives of our investigations were to first
tic anisotropic mechanical behavior,'21 unusual fracture identify the mechanisms for the large difference in ductility
behavior,01 poor short transverse properties,1*1 and higher between the AF/C-489 and AF/C-458 alloys and then to
production costs. develop an aging schedule to optimize the microstructure
A recent Air Force sponsored program focused on an for high ductility and strength levels.
effort to develop an aluminum-lithium alloy containing
greater than 2 wt pet lithium that possessed significantly
H. BACKGROUND
reduced anisotropic properties compared to those currently
available.131 Starke suggested that an intermediate «crystalli- Low ductility resulting from low energy intergranular
zation anneal be introduced between rolling stages to reduce fracture in Al-Li-Cu-X alloys has been attributed to a number
the sharp textures caused by the rolling and thereby decrease of factors that include tramp elements,l7,8,91 constituent parti-
the alloy's anisotropy.'61 This added step proved highly suc- cles,110,11,121 strain localization in the matrix,113-161 grain
cessful as the anisotropy of both the modulus and yield structure,1161 crystallographic texture,1171 and grain boundary
strength were reduced significantly from 20 to 25 pet for precipitates'181 with their corresponding precipitate-free
current aluminum-lithium alloys to less than 10 pet for the zones'191 (PFZs). Tramp impurity elements such as Na, K, Ca,
Air Force alloy. The subsequent Air Force alloy was desig- H, and S have been shown to segregate to grain boundaries,
nated. AF/C-489 with a nominal composition of Al-2.7 wt thereby weakening the boundary and leading to low energy
intergranular fracture.'7,8,91 The Fe and Si impurities, on the
other hand, lead to the formation of coarse constituent parti-
A.A. CSONTOS, Graduate Student, and E.A. STARKE. Jr., University cles during the casting process, which have been shown to
Professor and Oglesby Professor of Materials Science and Engineering,
are with the Department of Materials Science and Engineering, University
drastically degrade the fracture toughness of various alumi-
of Virginia, Charlottesville. VA 22903. num alloys.'101 Accordingly, both AF/C-489 and AF/C-458
Manuscript submitted September 17,1999. alloys were designed to minimize the concentrations of both

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 31A. AUGUST 2000—1965

the tramp elements (Na, K, Ca, H, and S) and constituent Mn. Their void nucleation and growth data coupled with
particle forming elements (Si, Fe) to well below those that their microstructural analysis suggests that the 0.31 wt pet
have been shown to cause low ductility. addition of Mn produced sufficient submicrometer sized
Apart from chemical inhomogeneities, strain localization dispersoids to homogenize slip without forming the large,
in the matrix has been shown to be a major cause for low coarse, primary Mn-rich particles, which were shown to
ductility in high-purity Al-Li-Cu-X alloys, e.g., AF/C-489 decrease the ductility.
and AF/C-458. This strain localization occurs by two dis- The Zr additions are designed to form ß (Al3Zr) disper-
tinctly different mechanisms: (1) cutting of shearable precip- soids, which act to greatly inhibit recrystallization during
itates such as 8' (Al3Li) and (2) the preferential deformation ingot breakdown and hot rolling. Nevertheless, the Zr addi-
of the softer PFZs. Binary Al-Li alloys with greater than tions strongly aid the formation of a pancake grain structure
5.8 at. pet Li or 1.6 wt pet Li and complex Al-Li-Cu-X and an intense crystallographic texture. Consequently, this
alloys, e.g., AF/C-489 and AF/C-458, with sufficient Li strong texture can produce adjacent grains and subgrains
concentrations homogeneously precipitate the ordered, with low misorientations, thereby, allowing glissile disloca-
coherent, spherical, and shearable strengthening precipitate, tions to penetrate their boundaries, wljich effectively
S'[m The concomitant shearing of these 8' precipitates with increases the grain size and, thus, the slip length. To reduce
each passing dislocation effectively softens the active slip the intense crystallographic texture associated with Zr addi-
plane, resulting in coarse planar slip and the formation of tions, methods such as a cross-rolling procedure as well as
intense slip bands, which impinge on grain boundaries, thus an intermediate recrystallization anneal between rolling
causing premature intergranular fracture. steps have been shown to significantly reduce the texture
Softer PFZs form in the matrix adjacent to grain bound- and, more interestingly, the anisotropy of certain Al-Li-Cu-
aries by two different mechanisms: (1) solute depletion and X alloys.'31-32-331
(2) vacancy depletion. In either case, the PFZ is significantly Nevertheless, the most effective method to improve the
softer than the precipitate hardened matrix, resulting in slip ductility of Al-Li alloys is to develop a microstructure that
localization in the PFZ and at the grain boundary triple contains various nonshearable precipitates in the matrix,
junctions, which also leads to premature intergranular frac- which act to homogenize slip.'15-30-331 Maximizing matrix
ture. Furthermore, the heterogeneous precipitation of coarse precipitation and minimizing coarse grain boundary forma-
grain boundary particles such as T2 (Al6Li3Cu)'201 and Tx tions of nonshearable metastable and equilibrium strengthen-
(AlxLiYCuz)'211 has been shown to lead to intergranular frac- ing phases is critical in order to reduce the slip length and
ture due to microvoid nucleation and coalescence. increase ductility. For Al-Li-Cu alloys, a preage stretch cou-
Minimizing strain localization in the matrix during defor- pled with a low-/high-temperature duplex aging schedule
mation is critical to increasing the ductility and improving has been shown to substantially increase the ductility and
the fracture behavior of Al-Li-X alloys. Slip length or the fracture toughness of various Al-Li alloys.134-38' The low-
measured distance between concurrent nonsheared micro- temperature aging maximizes the driving force for nucle-
structural features, e.g., incoherent precipitates, dispersoids, ation, while the preage stretch introduces dislocations as
competitive nucleation sites to grain boundaries. In complex
grain boundaries, etc., in the slip plane is a critical micro-
Al-Li-Cu-X alloys, a more substantial stretch is normally
structural parameter, which many fracture models'22-271 uti-
employed, since Tx nucleation is greatly influenced by dislo-
lize to accurately describe the degree of strain localization
cation jogs.'39,401 However, in order to determine the optimal
in an Al-Li-X alloy. In other words, microstructures with aging practices for both AF/C-489 and AF/C-458 alloys,
large slip lengths in Al-Li alloys encourage planar slip, while a detailed knowledge of the precipitation reaction of the
short slip lengths change the slip characteristics from planar metastable strengthening phases, i.e., 0', Tt, and 81, is
to wavy glide, thus, homogenizing deformation and reducing essential.
stress concentrations, which lead to premature fracture.
Microstructural features such as the grain size, dispersoid
size, and morphology, crystallographic texture, and hetero- ffl. EXPERIMENTAL PROCEDURES
geneous precipitation of strengthening particles all signifi- Air Force Research Labs and ALCOA provided both the
cantly affect the slip length and, thus, the fracture behavior 0.5-in.-thick 2.1 wt pet Li AF/C-489 and 0.75-in.-thick 1.8
of these Al-Li-Cu-X alloys. For example, grain size has been wt pet Li AF/C-458 alloys that were utilized for this investi-
shown to vary inversely to the true fracture strain (L-1 ~ gation. Both alloy ingots were direct chill cast, stress
ef) of aTi-Al binary alloy with shearable, coherent, spherical relieved, homogenized, hot rolled, annealed at 540 °C for 4
Ti3Al precipitates very similar to 8' in Al-Li alloys.116' Con- hours, and then slow cooled before being hot rolled again.
sequently, dispersoid-forming elements such as Mn and Zr This was followed by solution annealing at 540 °C for 1
are added to Al-Li-Cu-X alloys, i.e., AF/C-489 and AF/C- hour and an immediate water quench. Finally, a 6 pet preage
458, to control the grain size and structure. stretch was conducted within 8 hours of quenching to avoid
The Mn additions form the incoherent, nonshearable, natural aging effects, which can influence the homogeneity
cigar-shaped Al2oMn3Cu2[71 and Al6Mn'281 dispersoid parti- of the dislocation substructure and, thus, the heterogeneous
cles, which generally have dimensions less than 0.5 /tm in precipitation of the metastable strengthening phases. For
diameter and 1.0 /zm in length for high-purity Al-Li-Cu-X both AF/C-489 and AF/C-458 alloys, a 6 pet preage stretch
alloys. Starke and Lint29! and Walsh1301 et al. showed that both was determined to provide the greatest matrix Tt precipita-
coarse primary and submicrometer Mn dispersoids nucleate tion with the least amount of stretch.'411 Therefore, the T36
voids and thereby encourage the fracture process; however, temper for both the AF/C-489 and AF/C-458 alloys serves
their results also demonstrate that the maximum ductility of as the "as-received" material from which all aging studies
an Al-Cu-Mg-Zr alloy contained a 0.31 wt pet addition of were carried out for this investigation.

1966—VOLUME 3IA. AUGUST 2000 METALLURGICAL AND MATERIALS TRANSACTIONS A

 All aging studies were conducted in forced air furnaces aspect ratio roughly twice that of the AF/C-489 plate, while
with temperature fluctuations less than ±2 °C and immedi- the number of grains counted over a 500-mm circumference
ately water quenched. Continuous differential scanning calo- were from ~2.5 to 3 times greater in the AF/C-458 alloy
rimetry studies of the as-received T36 samples were than in the AF/C-489 plate.
accomplished with a PERKIN-ELMER* Series 7 model Low-magnification BF TEM was conducted for both AF/
C-489 and AF/C-458 to ensure that both alloys were received
*PERKIN-ELMER is a trademark of Perkin-EImer Physical Electronics,
Eden Prairie, MN.
in an unrecrystallized condition. The TEM micrographs in
Figures 1(c) and (d) clearly demonstrate the unrecrystallized
differential scanning calorimeter with a heating rate of 3 °C/ microstructures for both as-received T36 alloys as seen by
min. Tensile samples were machined into cylindrical and the approximate 5- to 10-/im subgrain size. The mechanical
dog-bone tensile bars conforming to ASTM E-8 substandard properties of both T36 alloys were quite similar except for
size regulations and tested utilizing the screw driven Sintech the 40 MPa variation in yield strength, which may be associ-
10/GL MTS system with a strain rate of 3 mm per minute. ated with the grain size difference or other microstructural
Optical samples were cut from the longitudinal (L), long features such as natural aging.
transverse (LT), and short transverse (ST) directions and Microstructural analysis of both T36 alloys, as seen in
then polished down to 0.05-/im colloidal silica solution and Figures 2(a) through (d), demonstrates the natural aging of
then etched with Keller's agent. Digital scanning electron the matrix 8' with the characteristic dislocation structure
microscopy fractographs were acquired through the use of resulting from the 6 pet preage stretch. Furthermore, Figure
a DSM982 Gemini field-emission scanning electron micro- 3 clearly shows the large strengthening effect that 8' natural
scope (SEM) operating at 10 kV equipped with dual back- aging has on the AF/C-458 T4 alloy. Due to this natural
scatter electron detectors, one perpendicular and the other aging of the shearable 8\ Alcoa stretched both alloy plates
at 45 deg to the surface. within the first 8 hours after quenching,'61 albeit, still too late
Transmission electron microscopy (TEM) samples were as a substantial 8' volume fraction had already precipitated.
cut from the grip section for microstructural analysis and Ideally, the preage stretch should be. accomplished within
the gage section for slip behavior investigations. The cut the first 3 hours of quenching to overcome the associated
wafers were then mechanically thinned to ~150-/im thick- problems with the naturally aged 8', i.e., nonuniform defor-
ness, punched into 3-mm disks, and electropolished in a mation due to planar slip. The resulting nonuniform disloca-
Tenupol twin-jet electropolisher at 11 V for the AF/C-458 tion structure could then affect the competitive precipitation
alloy and 8 V for the AF/C-489 alloy at a flow rate of 3 kinetics of the metastable Tt strengthening phase and, thus,
and photosensitivity of 9. A 3.5:1 methanol: nitric acid solu- the mechanical properties of the alloy.
tion was utilized as the electrolyte after being cooled to —25
°C. Conventional TEM was accomplished using a JEOL* B. Microstructural Evolution and Mechanical Properties
*JEOL is a trademark of Japan Electron Optics Ltd., Tokyo. of Both Single Aged T86 Alloys

2000FX transmission electron microscope at 200 kV, while Designing an artificial aging schedule subsequent to the
high resolution TEM was completed with the JEOL 4000 6 pet preage stretch required knowledge of the precipitation
high resolution electron microscope (HREM) operating at and strengthening response of both the AF/C-458 and AF/C-
400 kV. Centered dark-field (CDF) and bright-field (BF) 489 alloys as a function of time and temperature. Therefore,
techniques were used to observe the precipitate size and differential scanning calorimetry (DSC) and TEM studies
distribution. Stereological quantification of the T\ lath and were conducted to fully characterize the precipitation
6" plate shaped precipitates in thin TEM foils was completed response from the T36 temper for both alloys. Bright-field
through a statistical analysis'421 to overcome thin foil effects, TEM and diffraction pattern analysis of 6" streaking along
which underestimates the diameters and volume fractions g = (200)Jb = [001L and T, streaking-along g = (111)«/
of the 7\ and 6" precipitates from two-dimensional TEM b = [01 l]a indicated that the onset of 0" and T] precipitation
micrographs.1431 Furthermore, foil thickness was determined occurred between 90 °C to 95 °C and 120 °C to 125 °C,
by the well-established convergent-beam technique'441 and respectively, for both AF/C-458 and AF/C-489 alloys.
weak-beam dark-field (WBDF) was utilized with a large, In order to evaluate the strengthening effectiveness of
positive deviation parameter to examine the dislocation par- various aging routes, baseline hardness data were acquired
ticle interactions. to determine the various single-aged schedules. Figure 4
demonstrates the variation of the Rockwell B hardness val-
ues for both the AF/C-458 and AF/C-489 alloys as a function
IV. RESULTS AND DISCUSSION of single aging time at 100 °C, 125 °C, 130 °C, 135 °C, and
150 °C. Note that the slopes increase slowly and relatively
A. Grain Structure, Microstructure, and Properties of linearly for the 100 °C and 125 °C ages, where 6" is the
Both T36 and T4 Alloys main strengthening precipitate. However, the strengthening
The grain and subgrain structures of both T36 as-received rates for the 130 °C and 135 °C treatments increase more
AF/C-489 and AF/C-458 plates are demonstrated qualita- rapidly through the first day of aging and then decrease
tively by the optical micrographs in Figures 1(a) through to similar rates as the 125 °C curves. This initial rapid
(d) and quantitatively in Table I. Clearly, the 0.5-in.-thick strengthening may be due to initial Ti nucleation and early
AF/C-489 plate contains a much larger average grain size growth, which then slows at such low temperatures, i.e.,
in all three planes, i.e., rolling (LT), longitudinal (LS), and 10 °C, above the T, onset temperature. The 150 °C hardness
transverse (ST), as compared to the 0.75-in.-thick AF/C- curves are quite different, however, when compared to the
458 plate. Quantitatively, the AF/C-458 grains possessed an other lower temperature plots. The slopes of these curves are

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 31A. AUGUST 2000—1967

 El^^^^l

^K^SB

■ •* '-^^

H.-*:i"
Ftg. 1—(a) and (b) Optical and (c) (d) BF TEM micrographs illustrating the pancake grain structure and unrecrystallized subgrain structure for (a) and (c)
AF/C-489 and (b) and (d) AF/C-458 alloys.

Table I. Mechanical Properties, Grain Size, and Aspect Ratio Data for Both Alloys
AF/C-489 0.50-in. -Thick Plate
Yield Ultimate LT Number of LS Number of ST Number of
Strength Strength Fracture Aspect Grains/ Aspect Grains/ Aspect Grains/
(MPa) (MPa) Strain Ratio 500 mm Ratio 500 mm Ratio 500 mm
254 342 17.3 pet 3.4 18 19.7 75 10.1 83
AF/C-458 0.75-in. -Thick Plate
291 351 17.7 pet 6.7 42.2 34.2 228.5 24.6 261.3

substantially higher than the lower temperature counterparts, Also, note that the anisotropy between directions was deter-
which suggests that the T! nucleation and growth rates are mined to be, on average, less than 7 and 5 pet for the
much higher than at the lower temperatures. These results yield and ultimate tensile strengths, respectively, while the
indicate that the 150 °C single age practice provides a good elongations were on average greater than twice the accept-
combination of both high matrix Ti nucleation and growth able level of 5 pet. Therefore, this single age of 150 °C for
rates, i.e., near the nose of the time, temperature, and trans- 24 hours was utilized for all initial baseline mechanical
formation curve. property examinations.
The next step in determining a baseline 150 °C single Microstructural features of the baseline 150 °C 24 hours
aging schedule was to develop the best combination of single age for both the AF/C-458 and AF/C-489 alloys are
strength to ductility as a function of the aging time at 150 °C. shown in Figures 7 and 8. Qualitatively, the BF and CDF
Figures 5 and 6 demonstrate the tensile properties as a func- TEM micrographs clearly show that Ti, 6" engloved with
tion of aging time at 150 °C for the AF/C-458 alloy in both 8', and 8' are the major strengthening precipitates in both
the longitudinal and transverse directions, respectively. Note alloys at the 150 °C 24 hour single age. Matrix T, precipita-
that the best combination of strength to ductility from both tion is high in both alloys and, as seen in Figures 7(c) and
. the longitudinal and transverse data was in the slightly under- 8(c), two typical variants of grain boundary Tt are apparent.
aged condition, which was expected, at 150 °C for 24 hours. Grain boundary puckering and oriented T[ are two distinct

1968—VOLUME 31A. AUGUST 2000 METALLURGICAL AND MATERIALS TRANSACTIONS A

 since most grain boundaries were found not to be oriented
along a specific (111) habit plane.
/ ISO'CAilai
8S -

C. Microstructure and Mechanical Properties of

§ 80 - Multiple Aged T86 Alloys
c Since grain boundary Tj was observed in both 150 CC 24
1 75- jrtcTcMiÄ^ hour aged samples, duplex and triple aging studies were
conducted in an effort to minimize grain boundary Ti and
1 maximize matrix T] precipitation, thereby, increasing the
1 70-
ductility of these alloys. A preage stretch coupled with a
low-/high-temperature duplex aging schedule has been
65 -
—»- AF/C-489 Alloy shown to substantially increase the ductility and fracture
—*— AF/C-458 Alloy
toughness of various Al-Li alloys.'34-381 The low-tempera-
ture age above 120 CC was designed to maximize the driving
60 - 1 1 1 1
50 75
force for matrix T^ nucleation on dislocations and jogs fol-
Aging Time (Hours)
lowed by a higher temperature age to promote subsequent
matrix Ti growth. Furthermore, other duplex and triple aging
Fig. 4—Rockwell B hardness vs aging time. studies were designed to tie up the Cu atoms in the matrix
through the precipitation of a fine 6" dispersion at tempera-
tures between 95 °C to 120 °C. The higher temperatures of
580
the subsequent steps would then precipitate the matrix TIt
560 and since the Cu atoms were tied up in the matrix, the grain
boundary Tj variant would be minimized. The hope of these
' 540
various aging routes was to increase the matrix Tj density,
S 520
5> which would, in turn, reduce the slip length and strain local-
£
ization and promote greater ductility.
■3
H
Figures 9(a) and (b) show the results of the various duplex
3 480 M S
a and triple aging practices that were designed with the DSC
and TEM data. The duplex aging of the AF/C-489 alloy
resulted in promising increases in strain to fracture above
*" 440
the required 5 pet with only small losses in strength. In fact,
■ Yield Strength (L) •
420
- Ultimate Yield Strength (L)
as seen in Figures 10(a) and (b), the best duplex aging results
■ Elongation (L) of the AF/C-489 alloy reported significant increases in duc-
30 40 SO 60 70 80 tility by as much as 85 pet with a small decrease of only
Time at 150"C Age (Hours)
6.5 and 2.5 pet in yield and ultimate tensile strength, respec-
Fig. 5—Longitudinal tensile mechanical properties data for the AF/C- tively. Nevertheless, neither the double or triple aging prac-
458 alloy. tices designed for the AF/C-458 alloy proved effective in
substantially altering the yield strength, ultimate tensile
strength, and ductility. This point suggests that a very large
580 processing window exists for the AF/C-458 alloy even with
560
the wide variations in aging temperatures and time to achieve
comparable strength levels to the 150 °C single age. Figures
S 540
a. 10(a) and (b) also demonstrate the isotropic nature of the
2
? 520 mechanical properties between the longitudinal and trans-
BO

E 12 C.
verse directions for the AF/C-458 and, to a lesser degree, the
m 500
5
■o
AF/C-489 alloys. In fact, our results indicate the anisotropy
1 480
3> difference to be less than 5 pet for the AF/C-458 alloy, which
3 II g
is quite unique for Al-Li-Cu-X alloys.
| 460-
To examine the effect of the precipitate type, density, and
*" 440 size on the mechanical properties, quantitative microstruc-
420
Yield Slrenglh (T) tural analysis of the major precipitates is shown in Table II
Ultimale Yield Strength (T)
Elongation (T)
for both the single and duplex aged conditions corresponding
400
30 40 50 60 70 80
to Figures 10(a) and (b). The major difference between the
Time at 150-C Age (Hours) two alloys in the single aged 150 °C 24 hour condition was
the ~3 pet higher 8' volume percent for the AF/C-489 alloy,
Fig. 6—Transverse tensile mechanical properties data for the AF/C-458 while the duplex aged results showed that variations in the
alloy.
competitive precipitation of Ti and 0" were also evident for
the AF/C-489 alloy. For the AF/C-458 alloy, the T, and
6" volume percents and, correspondingly, the mechanical
variants of typical grain boundary precipitates found in both properties remain very similar for the single and duplex aged
alloys. Grain boundary puckered Ti was found to be more data, while only the number densities increase and mean
prevalent, as oriented grain boundary Tj was more unusual diameter decrease as anticipated by the duplex aging theory.

1970—VOLUME 31 A, AUGUST 2000 METALLURGICAL AND MATERIALS TRANSACTIONS A

 C)
*
1

- t

- »

\ 1

/ 1
/ l '"'' 1

»
Fig. 2—CDF micrographs of both T36(a) AF/C-489 and (b) AF/C-458 alloys after ~2 years of S natural aging after solutionizing at 540 °C for 1 h, b =
[0H]„, g = (100)*, WBDF TEM of both T36 (c) AF/C-489 and (d) AF/C-458 alloys indicating the dislocation structure in the matrix, b = [110]«,, g,
3g= (111).

60

-500 Days

24 Hours

-1— —i—
10 100 1000 10000 100000 Ioooooo
log Natural Aging Time (Minutes)
Fig. 3—Rockwell B hardness vs room temperature aging time for the AF/C-458 T4 alloy following a solution treatment at 540 °C for 1 h and water
quenching, demonstrating the strengthening effect of S1 precipitation.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 31A, AUGUST 2000—1969

Fig. 7—AF/C-458 150 °C 24 h micrographs: (a) CDF image showing 5' and 8" engloved with 5' (ff'18'), b = [Oil]«-, g = (lOO)j.; (b) BF indicating 0"
and matrix and grain boundary puckering T|, b = [110]„; (c) CDF micrograph showing matrix and grain boundary puckering Tj, b = [110]„, g = y(l 11)„.

Fig. 8—AF/C-489 150 °C 24 h micrographs: (a) CDF image showing S and 9" engloved with 5* (075*), b = [011]„ g = (100)«.; (6) BF indicating 0"
and matrix and grain boundary puckering T|, b = [110]^ (e) CDF micrograph showing matrix and grain boundary puckering Tlt b = [110],„g =i(lll)„.

As with the mechanical properties, the AF/C-489 precipi- tensile samples confirmed the formation of intense shear
tation behavior varied significantly with the added preage bands. The TEM analysis further showed that the cigar-
step in the duplex anneal. The Tj volume percents decreased shaped Al2oCu2Mn3 dispersoids were not sheared, but that
by nearly half, while the 9" volume percents and number the Ti adjacent to these dispersoids was still sheared.
densities increased significantly with the added low-temper- The SEM fractography results of the fractured tensile
ature age. These data suggest that the variation in tensile samples, as seen in Figures 14(a) through (d) and 15(a) and
properties between the single and duplex aged AF/C-489 (b), clearly demonstrate the flat and intergranular nature
alloy depends upon which strengthening precipitate, i.e., of the AF/C-489 single aged longitudinal and transverse
weaker 8" vs stronger Tb is dominant in the final fractured tensile samples as well as the duplex aged trans-
microstructure. verse fractured tensile samples. Conversely, the longitudinal
Nevertheless, as seen in Figures 11(a) through (f), dark- 100 °C (not pictured), 120 °C, and 130 °C duplex aged
field TEM indicates the presence of fine grain boundary Tt tensile samples fractured close to 45 deg, indicating more
along subgrains with less than 5 deg misorientation in both high energy transgranular shearing, which is favorable for
the AF/C-489 and AF/C-458 single and duplex aged sam- higher ductility. However, large regions of low energy inter-
ples. Coarse grain boundary T! and its associated PFZs were granular fractured grains were still evident along with
not observed in any of the single, duplex, and triple aged smaller areas of transgranular shearing, dimpled rupture, and
samples at or below 150 °C. Therefore, the tensile data microvoid coalescence (MVC) for these duplex aged AF/
coupled with the TEM analysis seem to indicate that the C-489 fractured tensile samples.
fine grain boundary Ti precipitation is not the limiting In contrast, as seen in Figures 14(a) and (c), all of the
parameter that affects the ductility of these alloys. longitudinal and transverse AF/C-458 tensile samples frac-
tured either with the classic ductile cup and cone or slant
45 deg fracture, thus, possibly indicating more of a transgran-
D. Analysis of Slip Behaviors for 8', 0", and T\ and
ular shear component and less intergranular fracture. Figure
SEM Fractography
15(a) confirms this point as the fractured grains were much
Whether S, 6", and/or Ti shear with mobile dislocations cleaner, indicating more high energy transgranular shearing
is a major factor that leads to strain localization and, thus, and grain boundary sliding with regions of high energy
premature intergranular fracture. Clear evidence of sheared dimpled rupture and MVC. Also note Figure 15(b) for a
8' and Ti and the subsequent strain localization in the form similarly aged AF/C-489 alloy, which shows large slip off-
of shear bands are shown in Figures 12(a) through (d) and sets characteristic of strain localization or slip concentrations
13(a) through (d) for both alloys aged at 150 °C 24 hours impinging a grain boundary. These large slip offsets are
and plastically strained to 2 pet. The TEM investigations of clearly absent in Figure 15(a) for the similarly aged AF/C-
the strained gage sections from all single and duplex aged 458 alloy, which suggests that strain localization played a

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 31A. AUGUST 2000—1971

 ' Y.S.
550 ' UTS

500

450

400

S > m
0 JT 300 o
I50CC I8H
IS<fC24H
*^^ a 'I
500
s. I50"C30H I 250
-«f
s IJO"C36H
IWC50H T® o 200 -5
IS<rC72H <s>
I 475- I SOT 2 Weeks
14O"C30H
^ 150

160*C20H 100
10tf'C24B/lSff'C24H
I20"CI8H/150"C24H 50
125°CI8H/IS0°C22H
I30°C24H/I5(T'C23H
I35"CI8H/I50°C2IH I T L T I- T L T
13rC24H/U(rC2IH 150gC24hra l00"C24hrs I20*C 18hn 130°C24hn
3> 135"C72H/I55',C22H IS0°C24hrs 150°C24hrs 150"C23hrs
<> I35*C72H/I65°C17H
Longitudnal Yield Strength Siegle & Daplex A(jn( Practices
V I45°C24H/16S°C7H
▼ !5ff'C2H/l70'C5H Loiiatiiidnal UTS
V I60°C3H/l3rC26H Transverse Ywkl Strength (a)
V I50*C2H/II5'CJ5H/170*C4H Transverse UTS

10 II 12 13 14 IS — Y.S.
— UTS
Strain (*• Elongation)

(a)

A ■ n

♦ sOn**
w
a
4 s*
o0$>V X 150'C24H
D 10O'C24H/l50TC16.5H
a B ■ 10O'C24H/15O,C18H
G3 100,C24H/15B'C24H
B 100,C24H/UOfC30H
El 100'C48H/150rci8H
t
Ä 475
A ♦
♦0 P 100*C72H/I50*C18H
O I20*C*.JH/lSirC21H
t T
150*C24taj
h T
100*C24hrs
L T
120*C48hn
L T
130*C24hrs
150"C24hrs U0*C22hrs lS0"C23hrs
•© 4> I20'C12H/150TC21H
CEB* fl20"C24H/15l)fC22H Stogie & Duplex Afhf Practices
T
12O'C24H/150rC24H
12O'C4»H/l50rC22H
X12S*CI8H/t50rC22H
130*C24H/l5<rC23H Fig. 10- —(a) AF/C-458 and (b) AF/C-489 mechanical properties for the
O 135'C18H/lS(rC21H
• l35*C72!i7MtfC17H 150 °C 24 h single age and duplex aged results.
0 135*C72H/l3(rC22H
▲ B i
■i Langiradnal Yield Strength
■i Langitudnel UTS
IB Transverse Yield Strength initial dislocations shear the precipitate followed by slip
JM TrmsvenseUTS
plane hardening due to the back stresses associated with the
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 IS
pileup at a major deformation barrier. Hence, Duva et al.[22]
Strain (% Elongation)
proposed a method to quantify strain localization as the
a» number of dislocations expected in a pileup since this is
Fig. 9—(a) AF/C-458 and (6) AF/C-489 strength vs strain for various measure of a microstructure's ability to deform by planar
aging treatments. slip:
r
N=f I/2 rl/2£ ^P
[1]
CBb
key role in the fracture process and, thus, the low ductility
of the AF/C-489 alloy in comparison to the AF/C-458 alloy. where N is the number of dislocations,/the volume fraction
Finally, these SEM fractographs also corroborate the optical of 8', r the 8' radius, L the slip length, b the burgers vector,
microscopy results in Figure 1, which showed the substantial Cp the constant that is related to the antiphase boundary
grain size variation between the AF/C-458 and AF/C-489 energy of 8', and CB the constant that depends on the matrix
alloys. shear modulus. Blankenship et al.{23] combined these slip
intensity calculations with estimates of the critical particle
size for the shearing-to-looping transition of S' phase and
E. Deformation, Fracture, and Slip Distribution predicted the slip distribution in complex Al-Li-Cu-X alloys.
By utilizing these relationships, Cassada et a/."5' and
Mechanistically, strain localization in Al-Li alloys with Blankenship and Starke'34' were able to develop microstruc-
5' has been shown to relate to slip plane softening after the tures that reduced the slip length and strain localization by

1972—VOLUME 31A, AUGUST 2000 METALLURGICAL AND MATERIALS TRANSACTIONS A

 Table II. Quantitative Precipitate Data for Both Single and Duplex Aged AF/C-4S8 and AF/C-489 Alloys Corresponding
to the Data Shown in Figures 10(a) and (b)
Alloy/ T, T| Number T( Mean 0" 6" Number 6" Mean 8' 8' Mean
Aging Vol Density Diameter Vol Density Diameter Vol Diameter
Practice Pet (nm) Pet (/"IT3)-1 (nm) Pet (nm)
AF/C-458
150 °C 24 hour 3.2 pet 7,690 58 0.23 pet 1,411 45 6.3 pet 7.8
100 °C 24 hour
150 °C 24 hour 3.0 pet 8,860 52 0.31 pet 10,330 19 5.7 pet 7.7
120 °C 18 hour
150 °C 24 hour 3.1 pet 10,920 48 0.24 pet 11,120 17 5.2 pet 7.8
130 °C 24 hour
150 °C 23 hour 3.0 pet 11,000 44 0.49 pet 12,900 22 5.8 pet 7.7
AF/C-489
150 °C 24 hour 2.9 pet 7,440 57 0.26 pet 4,200 29 9.1 pet 7.8
100 °C 24 hour
150 °C 24 hour 1.2 pet 5,550 43 1.38 pet 65,430 16 8.8 pet 7.6
120 °C 48 hour
150 °C 22 hour 1.7 pet 7,700 41 1.03 pet 28,060 22 8.9 pet 7.7
130 °C 24 hour
150 °C 23 hour 1.7 pet 7,660 42 1.06 pet 28,400 22 9.3 pet 7.7

Fig. 11—CDF TEM micrographs demonstrating the fine subgrain boundary T,, b = [110]„ g = UlU)a for (a) through (c) AF/C-458 and (d) through
(/) AF/C-489 alloys with the respective aging schedules denoted on the micrographs.

homogenizing slip. Cassada added Ge to an Al-Li binary Our investigation indicates that duplex aging had little
alloy to precipitate out Ge rods ~20 nm in length, which effect on the AF/C-458 ductility, while having nominal suc-
effectively homogenized slip. The Ge particles, thus, reduced cess for the AF/C-489 alloy. To understand the differences
the slip length from the grain size to the interparticle spacings between the duplex age results, we must consider the govern-
of the Ge precipitates, which changed the slip distribution ing parameters in Eq. [1]. The quantitative TEM data showed
from coarse planar to homogeneous slip. On the other hand, that the 5' volume fraction and diameter remained relatively
Blankenship developed a duplex age for AA8090, which constant between the single and duplex aging practices, i.e.,
reverted most of the 8 and continued the S' growth. This ~6 pet and ~8 nm for AF/C-458 and ~9 pet and -8 nm
age proved highly successful as the fracture toughness of for AF/C-489, respectively. However, ff volume fraction
the alloy was dramatically increased since planar slip was was on average 3 vol pet higher in the AF/C-489 alloy as
reduced due to the reduction of 8' volume fraction and the compared to the AF/C-458 alloy with similar heat treatments.
larger S' particles. Our results also indicate that the duplex aging of the AF/

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 31 A, AUGUST 2000—1973

 the number of dislocations impinging on a grain boundary
as compared to the AF/C-458 alloy, as seen directly in Fig-
ures 15(a) and (b). These major differences between alloys
seem to account for the much lower fracture strain of the
AF/C-489 alloy when compared to the AF/C-458 alloy with
similar aging.
The increased ductility of the duplex aged AF/C-489 alloy
with no change in fracture mode suggests that the superposit-
ion of the smaller, weaker, and denser 6" with the T i strength-
ening precipitates had a positive effect on the fracture
process. Quantitative TEM clearly showed that the duplex
aged AF/C-489 precipitated 6" with a (001) habit plane at
the expense of T! with a (111) habit plane, i.e., 0' volume
and number densities increased upward of 5- and 16-fold,
respectively, as compared to the single 150 °C 24 hour age.
This suggests increased ductility of the double aged AF/C-
489 alloy may be due in part to the superposition of the
0" and Tt strengthening effects. However, more work and
modeling needs to be conducted to clarify this point.

Fig. 12—AF/C-458 150 °C 24 h 2 pet plastic strained micrographs: (a)

V. CONCLUSIONS
CDF image showing sheared 5* (arrows), b = [llOJa, g = (lOO)j.; (b)
CDF indicating sheared T|, b = [110]„ g = £(111)„; (c) HREM micrograph
1 The precipitation response for both the AF/C-489 and AF/
showing sheared TV b = [110]„; and (d) BF image showing shear bands C-458 alloys was determined by TEM and DSC and was
along (lll)a trace, b = [110]«. used to design an improved microstructure through duplex
aging for AF/C-489.
2 Duplex aging of the AF/C-489 alloy resulted in significant
increases in ductility by as much as 85 pet, with a small
decrease of only 6.5 and 2.5 pet in yield and ultimate
tensile strength, respectively. No significant improve-
ments were found with either the duplex or triple aging
practices for the AF/C-458 commercial alloy, indicating
a large processing window.
3. Quantitative TEM demonstrated that the precipitation
response significantly differed between the duplex aging
for the AF/C-489 and AF/C-458 alloys. The duplex aged
AF/C-489 alloy competitively precipitated more 6" at
the expense of Tj, whereas no significant changes in
precipitation response occurred for the AF/C-458 alloy.
Furthermore, shearable S1 volume fraction was calculated
to be, on average, ~3 pet greater in the AF/C-489 than
the AF/C458 alloy.
4. Conventional TEM also demonstrated that grain bound-
ary T! precipitation was unaffected by duplex aging
designed to competitively precipitate more matrix Ti than
grain boundary Ti. In fact, all investigated single and
double aged samples contained grain boundary Tt.
5. Conventional TEM, HREM, and SEM fractography dem-
Fig. 13—AF/C-489 150 °C 24 h 2 pet plastic strained micrographs: (a) onstrated that formation of shear bands occurred as a
CDF image showing sheared 5* (arrows), b = [101]<„ g = (100)^ (b)
WBDF indicating sheared and bowed T,, b = [110]„ g, 3g = {(Ul)„;
result of the localized shearing of both the 8' and T,
(c) HREM micrograph showing sheared T,, b = [110]o; (d) BF image strengthening precipitates in both the AF/C-458 and AF/
showing shear bands along (111)„ trace, b = [110]„. C-489 alloys.
6. In summary, the low ductility of the AF/C-489 alloy in
comparison to the AF/C458 alloy is attributed to the
C-489 alloy promoted the formation of the weaker 0" at increased volume fraction of shearable & coupled with
the expense of the stronger Ti precipitate, while fine grain a larger grain size, both of which significantly increase
boundary T( precipitation seemed unaffected by the single the amount of strain localization and stress concentrations
and duplex ages for both AF/C-458 and AF/C-489 alloys. at grain boundaries. This results in more low energy
Since Ti was shown to shear, the distance between high intergranular fracture for the AF/C-489 alloy.
angle grain boundaries, i.e., grain size, determines the slip
length for these alloy microstructures. Accordingly, the much ACKNOWLEDGMENTS
larger grain size only exacerbates the higher 8' volume frac- The authors acknowledge helpful discussions with Dr.
tion for the AF/C-489 alloy, which significantly increases Kumar Jata during the course of this research. We also

1974—VOLUME 31A, AUGUST 2000 METALLURGICAL AND MATERIALS TRANSACTIONS A

i
 / Zr=)(\iVr :,= ii/i%'

fff: FJW;

/
%~L<i% f
u ^ | •

i
' f-<

\:

WC UH

= ri.A''/r,

Fig. 14—(a) through (d) SEM fractography of single and duplex aged AF/C-458 and AF/C-489 alloys from longitudinal and transverse fractured tensile
bars, as indicated.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 31 A, AUGUST 2000—1975

 !a)\ Ü A.K. Vasudevan, A.C. Miller, and M.M. Kersker: in Aluminum-Lithium
Alloys II, E.A. Starke, Jr. and T.H. Sanders, Jr., eds., TMS-AIME,
Warrendale, PA, 1983, p. 181.
10. A.K. Vasudevan, R.D. Doherty, and S. Suresh: in Aluminum Alloys-
t Contemporary Research and Applications, 1989, vol. 31, p. 446.
11. E.A. Starke, Jr. and J.T. Staley: Progr. Aerospace Sei., 1996, vol. 32,
p. 131.
12. J.G. Kaufman and J.S. Sanmer: in Applications of Fracture Mechanics

6'- A /
for Selection of Metallic Structural Materials, J.E. Campbell, W.W.
Gerberich, and J.H. Underwood eds., ASM, Metals Park, OH, 1982,
p. 169.
13. T.H. Sanders, Jr. and E. A. Starke, Jr.: Acta Metall., 1982, vol. 30, p. 927.
14. M. Furukawa, Y. Miura, and M. Nemoto: Trans. Jpn. Inst. Met., 1985,
/• vol. 26, p. 230.
j 15. W.A. Cassada, GJ. Shiflet, and E.A. Starke, Jr.: Acta Metall, 1986,
! •
'• M ^Ci'i.'
vol. 34, p. 367.
16. G. Terlinde and G. Luetjering: Metall. Trans. A, 1982, vol. 13A, pp.
U !
. *
M"
I >
i
i -i 1283-92.
17. S. Fox, D.S. McDarmaid, and H.M. Flower: Proc. Aluminum Technol-
ogy '86, Institue of Metals, London, 1986, Book 3.
18. A.K. Vasudevan and R.D. Doherty: Acta Metall., 1987, vol. 35, p. 1193.
19. R.E. Crooks, E.A. Kenik, and E.A. Starke, Jr.: Scripta Metall, 1983,
vol. 17, p. 643.
20. A.K. Vasudevan and S. Suresh: Mater. Sei. Eng., 1985, vol. 72, p. 37.
21. S. Suresh, A.K. Vasudevan, M. Tosten, and P.R. Howell: Acta Metall,
1987, vol. 35, p. 25.
22. J.M. Duva, M.A. Daeubler, E.A. Starke, Jr., and G. Luetjering: Acta
Metall, 1988, vol. 36, p. 585.
23. C.P. Blankenship, Jr., E. Hornbogen, and E.A. Starke, Jr.: Mater. Sei.
Eng., 1993, vol. A169, p. 33.
24. M. Sugamata, C.P. Blankenship, Jr., and E.A. Starke, Jr.: Mater. Scl
Eng., 1993, vol A163, p. 1
25. K.V. Jata and E.A. Starke, Jr.: Metall. Trans. A, 1986, vol. 17A, pp.
1011-26.
26. HJ. Roven: Scripta Metall. Mater., 1992, vol. 26, p. 1383.
27. E. Hornbogen: Z Metallkd., 1975, vol. 66, p. 511.
28. T.S. Srivatsan: J. Mater. Sei. Lett., 1988, vol. 7, p. 940.
29. E.A. Starke, Jr. and F.S. Lin: Metall. Trans. A, 1982, vol. 13A, pp.
2259-69.
Fig. 15—SEM fractography of (a) longitudinal AF/C-458 aged 150 °C 24 30. J.A Walsh, K.V. Jata, and E.A. Starke, Jr.: Acta Metall, 1989, vol.
h, indicating clean fractured grains, MVC, dimpled rupture, and no slip 37, p. 2861.
offsets; and (b) longitudinal AF/C-489 aged 150 °C 24 h, indicating slip 31. V.K. Jain, K.V. Jata, R.J. Rioja, J.T. Morgan, and A.K. Hopkins: J.
offsets and intergranular fracture with possible grain boundary Tt pullout. Mater. Processing Technol, 1998, vol. 73, p. 108.
32. K.V. Jata, S. Panchanadeeswaran, and A.K. Vasudevan: Mater. Sei.
Eng., 1998, vol. A257, p. 37.
33. K.V. Jata and A.K. Vasudevan: Mater. Sei Eng., 1998, vol. A241,
gratefully acknowledge the support of the Air Force Office p. 104.
of Scientific Research under Grant No. F49620-97-1-1034, 34.. C.P. Blankenship and E.A. Starke, Jr.: Metall. Trans. A, 1993, vol.
24A, p. 833.
Dr. Spenser Wu, Program Monitor. 35. R.F. Ashton, D.S. Thompson, E.A. Starke, Jr., and F.S. Lin: in Alumi-
num-Lithium Alloys III, C. Baker, PJ. Gregson, S J. Harris, and C.H.
Peel, eds., The Institute of Metals, London, 1986, p. 66.
REFERENCES 36. T. Ahrens and E.A. Starke, Jr.: in Aluminum-Lithium Alloys V, T.H.
Sanders, Jr. and E.A. Starke, Jr., eds., MCE Publications Ltd., Bir-
1. E.A. Starke, Jr.: Alloying, ASM, Metals Park, OH, 1988, p. 175. mingham, United Kingdom, 1989, p. 385.
2. A.K. Vasudevan, W.G. Fricke, Jr., M.A. Przystupa, and S. 37. C.P. Blankenship, Jr. and E.A. Starke, Jr.: Scripta Metall, 1992, vol.
Panchanadeeswaran: ICOTOM 8, TMS, Warrendale, PA, 1988, p. 26, p. 1719.
1071. 38. C.P. Blankenship, Jr. and E.A. Starke, Jr.: Acta Metall, 1994, vol. 42,
3. C.P. Blankenship and E.A. Starke, Jr.: in Aluminum-Lithium Alloys p. 845.
VI, M. Peters and PJ. Winkler, eds., DGM Informationsgesellschaft, 39. W.A. Cassada, G.J. Shiflet, and E.A. Starke, Jr.: Metall. Trans. A,
Oberursel, Germany, 1992, p. 187. 1991, vol. 22A, pp. 287-97.
4. K.T.V. Rao and R.O. Ritchie: Mater. Sei. Technol. 1989, vol. 5, 40. W.A. Cassada, G.J. Shiflet, and E.A. Starke, Jr.: Metall Trans. A,
p. 882. 1991, vol. 22A, pp. 299-306
5. Materials Directorate, Wright Laboratory, Air Force Material Com- 41. A.A. Csontos, B.M. Gable, A. Gaber, and E.A. Starke, Jr.: "The Effect
mand, Wright Patterson AFB, OH, under Contract No. F33615-92- of Quench Rate on the Microstrucrure and Properties of AF/C-458
C-5914. and AF/C-489 AI-Li-Cu-X Alloys," 7th International Conference on
6. A.K. Hopkins, K.V. Jata, and R.J. Rioja: in Aluminum Alloys, Their Aluminium Alloys: Their Physical and Mechanical Properties, Eds.
Physical and Mechanical Properties—Proceedings ICAA5, J.H. E.A. Starke, Jr., T.H. Sanders, Jr., and WA. Cassada, Trans Tech
Driver, B. Dubost, F. Durand, R. Fougeres, P. Guyot, P. Sainfort, and Publications, Zurich, Switzerland, pp. 1333-1340, 2000.
M. Suery, eds.. Materials Science Forum, Vols. 217-222, Transtec 42. W.A. Cassada: Ph.D. Dissertation, University of Virginia, Charlottes-
Publications, Zurich, Switzerland, 1996, p. 421. ville, VA, 1987.
7. E.A. Starke, Jr., T.H. Sanders, Jr., and I.G. Palmer: J. Met., 1981, vol. 43. J.E. Hilliard: TMS-AIME, 1962, vol. 224, p. 906.
33, p. 24. 44. D.B. Williams and C.B. Carter: Transmission Electron Microscopy,
8. D. Webster: Metall. Trans. A, 1987, vol. 18A, pp. 2181-93. New York, 1996, Plenum Press, New York.

1976—VOLUME 31A, AUGUST 2000 METALLURGICAL AND MATERIALS TRANSACTIONS A

 APPENDIX II JOURNAL OF
LIGHT METALS
ELSEVIER Journal of Light Metals I (2001) 1-14
www.elsevier.com/locate/Iigandmet

The role of plastic deformation on the competitive microstructural

evolution and mechanical properties of a novel Al-Li-Cu-X alloy
B.M. Gable, A.W. Zhu, A.A. Csontos, E.A. Starke Jr. *
Department of Materials Science and Engineering, University of Virginia. 116 Engineer's Way. P.O. Box 400745. Charlottesülle. VA 22904-4745. USA

Abstract
The role of plastic deformation prior to artificial aging on the microstructural evolution and mechanical properties of a novel Al-
Li-Cu-X alloy designated AF/C 458 was investigated. Induced plastic deformation ranged from a non-stretched or 0% stretch
condition to an 8% stretch, with intermediate stretches of 2%, 4% and 6%. Tensile properties, fractography and quantitative pre-
cipitate analysis were acquired from specimens that were water quenched from a solution heat treatment, immediately stretched and
artificially aged at 150°C. Fractography was investigated through scanning electron microscopy (SEM). Quantitative transmission
electron microscopy (TEM) determined the variation in precipitate type, number density, size and volume fraction of the major
strengthening precipitates Al2CuLi (r,), Al2Cu (ff'/ff) and Al3Li {5').
Age hardening curves for each level of mechanical stretch illustrated the enhanced aging kinetics of plastically deformed material
Quantitative TEM indicated that increasing amounts of pre-age stretch were found to greatly affect the competitive precipitation
kinetics of f, and ff /ff m AF/C 458 augmenting the volume fraction of fine matrix T, plates and dramatically decreasing the volume
fraction of ff/ff for isochronal treatments. A quantitative microstructural comparison of specimens exhibiting a given strength
demonstrated that the imposed level of cold work dictated the density, size and volume fraction of the competing precipitates The
tensile data indicated a trend of increasing ductility for equivalent yield strengths with the increasing amount of pre-age mechanical
stretch and therefore shorter artificial aging times. The quantitative precipitate data were used with a computer simulation for yield
strength determination. The theoretical simulation reported calculated yield strengths in good accord with experimental results and
can thus be used to predict the optimum microstructural configuration for high strength. © 2001 Elsevier Science Ltd All rights
reserved. °^

Keywords: Mechanical stretch; Competitive precipitation; Computer simulation; Aerospace materials; AF/C 458

1. Introduction and background hanced aging kinetics due to the increased number
density of strengthening precipitates and the shorter
Plastic deformation prior to artificial aging in Al-Li- time required to deplete the contiguous matrix of solute
Cu-X alloys has been found to enhance the strength, to obtain a specific strength. The increase in the volume
ductility and aging kinetics over non-deformed material fraction of matrix 7i combined with the enhanced aging
through the introduction of dislocations, which act as kinetics has been shown to lower the volume fraction of
preferential matrix nucleation sites for the primary grain boundary T, for the peak-aged condition [3-
strengthening phase 71 [1-15]. Increasing the amount of 6,8,10] and significantly lower the number density and
introduced plastic deformation leads to a greater num- volume fraction offf'/ff. Augmented 7"i volume fraction
ber of matrix dislocations. An increase in the number has been shown to affect Al3Li (<5') precipitation due to
density of nucleation sites [11] creates overlapping dif- limiting the available lithium for further nucleation and
fusion fields as precipitates grow, causing the average growth [5,6,13].
size of the precipitate to decrease. The refinement of the Cassada et al. [11,12] demonstrated that increasing
precipitate microstrücture corresponds with the en- high levels of plastic deformation prior to artificial aging
for a near-2090 alloy aged at 190°C resulted in marked
gains in yield strength for the under- and peak-aged
Corresponding author. Tel.: +1-804-924-6335; fax: + 1-804-924- conditions due to the enhanced volume fraction of fine
1353. matrix Tl. Matrix precipitation of this more abundant
E-mail address: [email protected] (E.A. Starke Jr.). and finer strengthening phase directly correlates to the
1471-5317/01/S - see front matter © 2001 Elsevier Science Ltd. All rights reserved
PlI: SI 4 7 I -53 1 7(00)00002 -X
 B.M. Gable et al. I Journal <•/ Light Metah I (2001) /-/</

increased strength and ductility of the material versus a initial gauge of 0.025 m. The extent of imposed plastic
non-stretched condition that exhibits only grain deformation ranged from a non-stretched or 0% stretch
boundary T{ precipitation [4-6,13]. Recent computer condition to an 8% stretch, with intermediate stretches
models by Zhu et al. [16] suggest that the strengthening of 2%, 4% and 6%. The tensile samples were aged in a
effect of matrix precipitation is optimized through the Stabil-Therm box furnace at 150°C for various times to
presence of two unshearable plate precipitates with investigate the mechanical properties and microstruc-
{111} and {100} habit planes, i.e., Tx and ff'/ff, re- tural evolution of the alloy as a function of prior plastic
spectively. deformation. The reported tensile data are an average of
Although it has been known for many years that a at least duplex specimens.
stretch prior to aging introduces dislocations that sub- Age hardening curves based upon the obtained ten-
sequently act as nucleation sites for precipitates with sile data for each stretch condition were obtained. Each
large interfacial strains, no systematic study has been tensile specimen was subjected to tensile loading until
conducted on the effect of stretch when different types of fracture on the same MTS Sintech 10/GL frame at a
precipitates are in competition for preferential matrix crosshead displacement of 0.003 m/min to obtain
nucleation sites and solute atoms. strength and ductility data. An isochronal investigation
This study investigated the role of varying prior was conducted to quantify the role of stretch on
plastic deformations on the competitive precipitation competitive matrix precipitation. The age hardening
kinetics of the primary strengthening phases Al2CuLi curves allowed for the comparison of the microstruc-
(T\) and AliCu {&'/&). Quantified transmission elec- tural evolution of specimens of a given yield strength
tron microscopy (TEM) data were related to the sub- but varying levels of plastic deformation and artificial
sequent tensile properties of AF/C 458, namely aging time.
strength and ductility. Number density, size and vol- Representative fractography specimens were sec-
ume fraction comparisons of the T\,&'/& and 5' tioned from fractured tensile samples and mounted for
strengthening precipitates were performed in order to scanning electron microscopy (SEM) investigation. A
quantify the competitive precipitation kinetics as a Zeiss DSM field emission gun (FEG) SEM was used for
function of thermomechanical processing for the alloy. the fractographic investigation. Specimens were sec-
These quantitative data were then used as a means of tioned from untested tensile specimens for TEM inves-
challenging the validity of recent computer simulations tigation on a JEOL 2000FX. Once sectioned, the
[16]. specimens were ground to approximately 150 urn, pun-
ched to into 3 mm disks, and chemically thinned using a
7:2 methanol-nitric solution cooled to -24°C in a
2. Experimental procedure Struers Tenupol-3 electropolishing unit that produced
electron transparent foils for TEM observation.
Alcoa fabricated the Al-2.7Cu-l.8Li-O.6Zn-0.3Mg- Through traditional point counting methods the
0.3Mn-0.08Zr (w/o) alloy designated AF/C 458 used in number density, size and volume fraction of the primary
this study. The alloy was ingot cast, homogenized and strengthening precipitates were quantified from the
subsequently formed through a two-stage rolling pro- transmission electron micrographs taken near
cess. The unrecrystallized plate underwent a 6% tensile B = [110]. Quantitative analysis required the foil thick-
stretch in the longitudinal direction resulting in a ness of each captured region, which was obtained
1.905 x 10"2 m through-thickness plate. This is consid- through convergent beam electron diffraction (CBED)
ered the "as-received" material. patterns [19,20], and a stereological method of correc-
Tensile samples were machined in accordance with tion based upon the size of the precipitates relative to
ASTM E8- 93 [17] parallel to the rolling direction of the the thickness of the TEM foil [21,22].
as-received plate. These specimens were then solution The quantified precipitate data for each thermo-me-
heat treated in a Lindberg three-zone furnace at 540°C chanical condition were then implemented in a com-
for 1 h and subsequently water quenched (290°C/s) and puter simulation that calculated the theoretical yield
immediately stretched. The pre-age stretch was per- strength. The precipitate type, volume fraction and di-
formed immediately upon quenching [13] from the so- mensions were required for simulation, which accounted
lution heat treatment to minimize the effects of an for disk-like precipitation on {11 U and {100} habit
inhomogeneous dislocation structure arising from the planes, T\ and ff'/ff, respectively, and homogeneous
presence of shearable naturally aged 8' precipitates [18]. shearable point-like particles, S'. The simulation was
The tensile specimens were stretched on an MTS Sintech based upon the stress required to pass a dislocation
10/GL frame employing a crosshead displacement of through a matrix consisting of the complex four-phase
1 x 10~4 m/min. The real-time elongation of each spec- microstructure. Additional terms encompassing the
imen was observed through a digital display via a matrix and solid solution strengthening effects were also
computer interface with a MTS extensometer across an taken into account.
 B.M. Gable et al. I Journal nf Light Metals I (2001) 1-14

3. Results 550

3.1. Artificial aging characteristics

Age hardening curves in terms of yield strength were

determined for each level of mechanical stretch, as
shown in Fig. 1. This plot illustrates the enhanced aging
kinetics and yield strength with greater amounts of pre-
age stretch.
The first aspect of this study was to examine the aging
characteristics of AF/C 458 as a function of the imposed
level of mechanical stretch prior to artificial aging for an 12 18 24 30 36
isochronal treatment. The second aspect of this study Aging Time @I50*C(H)
was to investigate the microstructural evolution of ma- Fig. l. Age hardening curve for various levels of pre-age stretch arti-
terial that experienced varying thermo-mechanical pro- ficially aged at 150°C.
cessings, namely different levels of pre-age stretch and
artificial aging times, to obtain similar yield strengths.

3.1.1. Tensile data - 150° C for 24 h 3.1.2. Tensile data - isostrength specimens
The tensile data from the specimens that were aged at The yield strength of 450 MPa was chosen as the
150°C for 24 h are given in Table 1. There is a marked basis for comparison between the various stretch con-
trend of increasing strength with pre-age stretch. It is ditions. Material in the under-aged condition was de-
important to note the marked gains in strength between sirable to avoid possible difficulties with quantifying an
the 0% and 2% stretch conditions. There were continued over-aged structure. The comparison of the tensile data
gains with increasing levels of stretch, but to a lesser and artificial aging times for the five stretched condi-
extent. The increased strength for each successive stretch tions may be seen in Table 2. Note that a direct rela-
condition was not accompanied by concomitant losses tionship exists between the level of imposed pre-age
in ductility. The strength x ductility factor increased mechanical stretch and the required artificial aging time,
with increasing levels of stretch, resulting in over a 25% with the most dramatic difference again being between
improvement between the extreme pre-age stretch con- the non-stretched and 2% stretch specimens. Further-
ditions. For brevity the 0%, 4% and 8% specimens were more, note that a possible trend of increasing ductility
used for the remaining fractographic and microstruc- with the amount of mechanical stretch between heat
tural comparisons for the isochronal treatment. treatments occurred.

Table 1
Tensile data after 150°C for 24 h

Quench media Pre-age stretch Yield strength S.D. Tensile strength S.D. Elongation S.D.
(cooling rate) (%) (MPa) (MPa) (%)
Water quench 0 324 12.6 434 12.8 12.9 1.2
(290°C/s) 2 415 9.8 473 15.3 10.3 0.9
4 471 7.0 513 1.5 10.4 0.6
6 488 15.5 519 12.8 9.9 0.6
8 506 19.6 531 21.1 10.6 1.5

Table 2
Tensile data for T8x specimens with similar strengths

Quench media Pre-age stretch Aging time (h) Yield strength Tensile strength Elongation
(cooling rate) (%) (MPa) (MPa) (%)
Water quench 20 454 515 9.1
(290°C/s) 35 452 503 9.1
20 451 500 9.9
15 452 496 10.9
10 452 497 10.5
 B.M. Gable et at. I Journal of Light Metals 1 (2001) 1-14

3.2. Scanning electron microscopy was approximately the same. The latter result is most
likely related to similar grain boundary precipitation.
3.2.1. Fractography for 24 h at 150°C
The micrographs in Fig. 2 indicate predominant 45° 3.2.2. Fractography for 450 MPa specimens
slant fracture for each tensile specimen. Each of the The fractography of the isostrength specimens may be
fracture surfaces has regions of transgranular fracture, found in Fig. 3. Each tensile specimen exhibits 45° slant
dimple rupture and grain boundary delamination. It is fracture with some degree of transgranular fracture and
important to note that each of these conditions should grain boundary delamination. There appears to be a
have approximately the same amount of grain boundary trend of increasing transgranular fracture with increas-
precipitation. For the isochronal aging treatment the ing levels of stretch, which may be related to decreasing
higher the imposed level of deformation prior to aging, artificial aging times and subsequent grain boundary Tt
the stronger the specimen although the strain to fracture growth. Those specimens aged at longer times appear to
have more intergranular failure coupled with grain
boundary delamination, which correlates with more
0% 24H pronounced grain boundary precipitation.

3.3. Transmission electron microscopy

As stated previously, the 0%, 4% and 8% stretch

conditions were utilized for quantitative microstructural
comparison. Matrix precipitation of Tu 6"/ff and 8
strengthening phases were quantitatively measured.

3.3.1. Microstructure for 24 h at 150°C

The TEM micrographs in Fig. 4 illustrate the Bf/0
r 50nm and T\ matrix precipitation as a function of the me-
chanical stretch. The most significant microstructural
4% 24H change is evident between the unstretched and stretched
conditions. The precipitation of the 0% stretch samples
is predominantly comprised of grain boundary rt along
with matrix ff'/tf and <5\ with the occasional matrix Tt
plate that precipitated on residual matrix dislocations. It
is important to note the (f'/0 precipitate-free zone that
is contiguous to the grain boundary Tt. Both the 4% and
8% stretch conditions indicate significant matrix Tt
precipitation relative to the unstretched specimen.

3.3.2. Microstructure for 450 MPa specimens

Figs. 5 and 6 illustrate the respective matrix precipi-
tation for those specimens with yield strengths of ap-
8% 24H proximately 450 MPa. The centered dark field images
highlighting 8' precipitation, Fig. 5, make it possible to
visualize the relative number density and size of the
matrix & due to the wetted exterior. There appears to be
a trend of decreasing ff diameter with increasing pre-age
stretch. Similarly, in Fig. 6, the shorter aging times ap-
pear to correlate with a finer 71 plate structure.

3.4. Quantitative microstructural analysis

Optical point counting and stereological correction

methods were utilized to obtain the number density, size
and volume fraction of the primary strengthening
Fig. 2. SEM fractography of specimens aged for 24 h at I50°C; the phases 7*,, ff'/& and 5' for each condition. The stere-
low-magnification composites illustrate the macroscopic fracture angle ological correction method is necessary to account for
of a tensile sample and a representative region of the fracture surface. particle truncation when the precipitates are largo rcla-
 B.M. Gable el al. I Journal of Light Metals 1 (2001) 1-14

0% 120H

2% 35H 4% 20H

6% 15H 8% 10H

50um|
Fig. 3. SEM fractography specimens stretched different percentages and aged for various times at 150°C to obtain a yield strength of 450 MPa.

tive to the thickness of the examined foil [12]. The 0%, precipitation in the unstretched condition relative to T\.
4% and 8% stretch specimens were used for micro- Conversely, the imposed deformation prior to aging
structural comparison, as in the fractography study. dramatically increased the number density and volume
fraction of matrix 7|, which continued with increasing
3.4.1. Strengthening precipitate data for 24 h at 150°C levels of deformation. These quantified microstructural
Table 3 conveys the quantified number density, di- data demonstrate that the enhanced precipitation of
ameter and volume fraction data for T\, tf'/ff and S' as matrix T\ dominates that of tf'/ff for a pre-age stretched
a function of prior deformation that was aged for 24 h at condition. Table 3 also illustrates a non-linear trend for
150°C. Table 3 indicates that &'/& dominates the matrix the corrected plate diameter for both T\ and ff as a
 B.M. Gable el al. I Journal of Light Metals I (2001) 1-14

0% 24H

4% 24H

8% 24H

Fig. 4. TEM micrograph illustrating the matrix precipitation of specimens after aging at 150»C for 24 h. Each condition has a centered dark field
,mage conveymg matnx fl»/?and 5^=1/2(200); andacentered dark field inuge for matrix rt Predpitation.?= 1/2(1 i l).LTfeaffS0fS
gram boundary on 0" precipitation in the 0% stretch condition. w i »;■ porcineanectottne

function of pre-age stretch, which will be discussed 3.4.2. Strengthening precipitate data for 450 MPa spec-
hereafter. It is important to note that the thickness of imens
the plates was not found to change for any artificial Table 4 indicates the quantified number density, di-
aging or stretch condition. The measured thickness of ameter and volume fraction data of Tu ff'/ff and S' for
the T\ plates was found to be on average 1.2 nm. The the various stretched and aged specimens that exhibited
only difference in thickness was between the 0" (0.8 nm) yield strengths of approximately 450 MPa.
and & (1.2 nm) from the unstretched condition. The quantified number densities for the precipitates
There is an apparent trend of decreasing volume were found to increase with the level of imposed me-
fraction of 5' with increasing levels of stretch. This re- chanical stretch as illustrated in Figs. 5 and 6. Con-
lationship would be expected due to the enhancement of versely, there is an apparent trend of decreasing
T] precipitation, however, it is important to note the precipitate diameter and volume fraction with increasing
significant error associated with the calculated values. level of pre-age stretch for each precipitate. This rela-
There does not appear to be a trend between the mea- tionship will be developed further, specifically consid-
sured d" diameter and the imposed level of predefor- ering the role of the stereological correction method on
mation. plates of very different sizes. As for the isochronal heat
 B.M. Gable et ai I Journal of Light Metals 1 (2001 j 1-14

0% 120H

2% 35H 4% 20H

6% 15H 8% 10H

Fig. 5. TEM centered dark field micrographs illustrating $ and wetted 8/ matrix precipitation of the specimens that experienced various levels of pre-
age stretch and were aged to obtain yield strength of 450 MPa. For each condition g = 1/2(2 0 0).

treatment the thickness of the plates was determined to 4. Discussion

remain constant.
The results from the theoretical simulation on the 4.1. The role of mechanical stretch
quantitative data for the isochronal and isostrength
conditions may be found in Tables 5 and 6, respec- 4.1.1. Specimens aged for 24 h at 150°C
tively. There are dramatic similarities between the ex- Initially, the relationship between the role of stretch
perimental and calculated yield strengths for the and the microstructural evolution of AF/C 458 was in-
isochronal specimens. The theoretical values for each vestigated by obtaining tensile data for specimens that
stretch condition are within several percent of those had undergone 0%, 2%, 4%, 6% and 8% mechanical
obtained through laboratory tensile testing. Similarly, stretches prior to artificial aging at 150°C for 24 h. These
the calculated yield strengths for the isostrength speci- specimens conveyed a direct trend of increasing aging
mens were quite good. There is a noticeable decreasing kinetics and strength with increasing levels of pre-age
trend between the level of imposed stretch and the stretch. It is important to note that there was greater than
simulated yield strengths. This phenomenon will be a 50% gain in yield strength for the isochronal condition
discussed later. between the non-stretch and 8% stretch specimens, with a
 B.M. Cable et al. I Journal of Light Metals I (2001) 1-14

0% 120H

2% 35H 4% 20H

6% 15H 8% 10H

Fig. 6. TEM centered dark field micrographs illustrating matrix 7", precipitation of specimens that experienced various levels of pre-age stretch ai
were aged to obtain yield strength of 450 MPa. For each condition g = 1/2(111).

Table 3
Quantitative precipitate data after 150°C for 24 h

Quantitative analysis Stretch (%) r, (Al2CuLi) S.D. er/er (AI,CU) S.D. S (AljCu) S.D.
3
Number density (#/um ) 0 286.0 50.0 6277.0 1023.0 15030.0 5677.5
4 916.2 280.2 493.8 382.8 15691.6 865.8
8 3035.0 183.9 176.4 106.0 12585.4 628.5
Precipitate diameter (nm) 0 64.7 2.2 33.9 0.9 12.1 2.0
4 80.1 5.0 63.2 8.4 11.8 1.7
8 49.8 4.6 38.4 3.9 12.2 1.9
Volume fraction (%) 0 0.52 0.17 1.14 0.14 11.10 2.25
4 2.23 0.23 0.54 . 0.27 10.70 1.63
8 2.84 1.00 0.07 0.04 9.50 1.50

minor loss in ductility. There was only a slight decreasing strength x ductility factor. The subsequent quantitatn
trend between elongation and the level of stretch, indi- microstructural analysis of various stretched conditior
cating the positive effect that the stretch has on the demonstrated a direct trend between the imposed level i
 B.M. Gable et al. I Journal nf Light Meiah I (2001) 1-14

Table 4
Quantitative precipitate data for specimens with a yield strength of 450 MPa

Quantitative analysis Pre-age Aging time r, S.D. ff (Al:Cu) S.D. S (Al, Li) S.D.
stretch (%) (h) (AljCuLi)

Number density (#/um3) 0 120 272.6 73.0 359.4 96.9 7778.0 1490.3
2 35 754.7 250.2 387.9 139.8 11356.5 1276.9
4 20 963.1 69.4 512.4 115.6 17359.6 2798.9
6 15 1283.1 337.7 548.6 180.1 22885.4 6332.6
8 10 1478.3 582.9 835.7 434.4 23675.4 11117.4
Precipitate diameter (nm) 0 120 158.7 12.5 88.2 30.3 16.3 3.7
2 35 ■ 86.8 8.8 81.3 9.0 13.2 2.2
4 20 55.6 6.0 40.1 3.6 11.9 2.0
6 15 53.6 7.7 42.8 7.1 10.7 1.8
8 10 40.1 5.7 28.5 5.0 10.6 1.6
Volume fraction (%) 0 120 2.53 0.47 0.95 0.73 14.00 5.00
2 35 2.12 0.51 0.65 0.39 10.90 2.19
4 20 1.64 0.82 0.23 0.05 12.30 3.21
6 15 1.42 0.37 0.30 0.09 11.70 5.25
8 10 0.94 0.58 0.20 0.15 11.70 8.27

Table 5 Table 6
Computer simulation results for isochronal condition Computer simulation results for isostrength specimens

Pre-age Aging Measured Calculated Devia- Pre-age Aging Measured Calculated Devia-
stretch (%) time (h) yield strength yield strength tion stretch time(h) yield strength yield strength tion (%)
(MPa) (MPa) (%) (%) (MPa)

0 24 324 302 -6.8 0 120 454 494 8.8

4 ' 24 471 490 4.0 2 35 452 479 6.0
8 24 506 488 -3.6 4 20 451 420 -6.9
6 15 452 438 -3.1
8 10 452 396 -12.4

plastic deformation, yield strength and the number

density and volume fraction of matrix T\ precipitates. between precipitates. This interparticle spacing is the
Conversely, the pre-age stretch dramatically lowered the limiting factor in the soft impingement of neighboring
number density and volume fraction of ff'/ff. This may precipitates. The greater the number of nucleation sites,
indicate the predominance of T\ in the competition for the finer the distribution of plates, the smaller the av-
heterogeneous matrix nucleation sites. This illustrates a erage diffusion field, and therefore the shorter the time
direct correlation between the imposed level of stretch, required to deplete the surrounding matrix. The en-
i.e., the matrix dislocation density, and the number of hanced aging kinetics of increasing levels of plastic de-
quantified T\ and &'/ff precipitation events. formation was illustrated in the age hardening curves
The obtained tensile data for the various pre-age seen in Fig. 1 and the time required for specimens to
stretched conditions, found in Table 1, indicate an initial obtain a yield strength of 450 MPa in Table 2.
"jump" of 28% (91 MPa) in yield strength associated In a similar fashion, there is a direct correlation be-
with a 2% stretch. This relationship is directly related to tween the quantified number density of 7i and &'/& and
the augmented matrix precipitation of T\ relative to that the respective measured volume fraction, as shown in
of 0". It is important to remember that the dense pop- Table 3. The matrix precipitation for the unstretched
ulation of 8" in the unstretched condition is adversely condition is dominated by 0", due to the few residual
affected through the addition of matrix dislocations. dislocations allowing for preferential T\ and & nucle-
These findings are in good agreement with the earlier ation. Furthermore, increasing the level of pre-age
studies that focused on the number density and precip- stretch dramatically increased the number of heteroge-
itate diameter of competing phases in Al-Cu-X systems neous matrix nucleation sites, which affected the com-
[11-14]. petitive precipitation kinetics between T\ and &'/&,
The enhanced aging kinetics of a given stretch con- resulting in a marked shift in the relative volume frac-
dition directly correlates with the marked increase in the tion.
number density of matrix T\. The increase in the number The quantified precipitate diameters also lend them-
density of nucleation sites decreases the average distance selves to discussion. The data in Table 3 do not convey
 10
B.M. Cable et at. I Journal of Light Metals I (2001) 1-14

any apparent linear trends between the size of either T\

Several pertinent studies on Al-Li-Cu-X alloys have
or ff'/ff as a function of imposed mechanical stretch.
been summarized in terms of their Li:Cu ratios in Table
However, the quantified relationships may be explained.
7. All of these ratios were calculated assuming a con-
For both precipitates there was a measured increase in
stant lithium solubility of 2.0 at.% and negligible copper
precipitate size between the 0% and 4% stretch condi-
solubility. In the case of AF/C 458, the copper content
tions followed by a refinement with increasing stretch to
was adjusted due to the presence of manganese (AF/C
8%. The increase in average size for the &'/& between
458 utilizes two types of dispersoids, one of which is an
the 0% and 4% stretch conditions is due to the intro-
AlioMnjCu: type). The ratios reported are in terms of
duction of heterogeneous matrix nucleation sites.
super-saturated atoms, i.e., those atoms available to
Although not supported through high-resolution mi-
form secondary phases. It is important to note that the
croscopy, it is believed that the matrix precipitation in
Li:Cu atomic ratio for AF/C 458 is approximately 9:2.
the unstretched condition is comprised solely of 9" for
Note that the ratio for the copper-rich alloy from the
early artificial aging times, rather than (?. This is taken
Silcock study, in which ff dominated, was less than 1.0
from the literature [1,23-25] in which case &' is the
whereas the remaining alloys exhibited dominance by 7/,
preferred matrix precipitate when there is an insufficient
and had a Li:Cu greater than 1.0. Stoichiometrically, the
number of matrix heterogeneous nucleation sites. This Li:Cu ratio must be greater than 1.0 for Tx (Al2LiCu) to
increase in quantified diameter correlates with the
possibly control the competitive co-precipitation with ff
transition from 0" to 0* with the significant loss in (Al,Cu).
number density due to the presence of both ff and Tu
A study by Noble and Thompson [2] investigated two
The refinement of both 0 and T\ with stretch to in-
Al-Li-Cu alloys, one copper-rich with moderate lithium
creasing 8% was due to an increase in the number den-
(alloy A), and the other with low copper and higher
sity of nucleation sites which directly relates to
lithium (alloy B). These alloys are also included in Table
overlapping diffusion fields, leading to finer precipitates.
7. The investigators found that alloy A, the copper-rich
It is important to recall the early work of Silcock [1]
alloy with moderate lithium, had homogeneous grain
that investigated the role of introducing heterogeneous
boundary and matrix precipitation of Ts without the
matrix nucleation sites, in the form of cadmium addi-
addition of matrix heterogeneous sites. This indicates
tions and matrix dislocations through a pre-age stretch.
that the relative super-saturation of the matrix overcame
Focusing on the mechanical stretch aspect, the author
the barrier for homogeneous matrix nucleation. Con-
found that for cold worked material the precipitate that
versely, the copper-poor alloy, alloy B, only exhibited
preferentially nucleated and grew in an Al-Li-Cu ter-
matrix Tt precipitation along grain boundaries and
nary was dictated by the alloy's chemistry. In a high
matrix dislocations. Due to the decrease in chemical
copper containing alloy the dominant precipitate that
driving force, from the less-saturated matrix, T, precip-
preferentially nucleated and grew was ff. An alloy with
itation required further reduction of the interfacial and
an additional 0.5 wt%Li exhibited an increase in the
strain energy barriers for nucleation in the form of
intensity of 7} diffraction spots relative to those of ff.
matrix dislocations. Noble and Thompson noted that
Similarly, 7/, was more dominant with respect to-fl' in a
lithium-rich alloy. This indicates that the driving force the change in precipitation indicated a loss in the driving
force for 7\ precipitation.
for the preferential precipitation is a function of the
relative saturation of the matrix, or the Li:Cu ratio. This The kinetics of co-precipitation are based upon the
ratio may indicate a significant driving force for the combinations between lattice defects, i.e., heterogeneous
nucleation sites, that minimize the energy barrier for
precipitation of either 7*, or ff, given heterogeneous
matrix sites. nucleation and the rate of formation for the competing .
phases [24]. The phases in competition with one another
Table 7
Atomic Li:Cu ratios for various AI-Li-Cu-X alloys of interest

Investigation [reference] Alloy Lithium (wt%) Copper rreaominant pha se Li:Cu ratio (atomic)
(wt%) in T8.v

ff A
Current study AF/C 458 1.80 2.70 X
Noble & Thompson P] Alloy A 4.43
3.50 3.50
Alloy B X 2.19
1.50 2.50
Silcock [I] #1 X 4.90
1.00 3.50 Very strong
#2 Weak 0.88
1.50 3.50 Medium
#3 Medium 2.19
2.00 1.00 Very weak
Cassada [11,12] Weak 12.24
Near-2090 2.45 2.45 X 6.68
 B.M. Gable et al. I Journal of Light Metals I (2001) 1-14

then form simultaneously at their respective preferred metastable phase. Calculated interfacial energies for the
sites. The less stable phase dissolves sacrificially during <5', 0", & and T\ precipitates may be found in Table 8.
growth, with the more stable phase becoming more These data illustrate the relative energy barriers for the
predominant, ruling the competitive precipitation. competing phases for various precipitate-defect combi-
When in competition the dominant precipitate is the one nations. It is important to recall that fundamental data,
with the lowest energy barrier for a given lattice defect. such as interfacial energies, are a strong function of
The energy barrier for a given site in a certain alloy is various metallurgical factors, e.g., alloy composition,
dependent upon the misfit strain energy, interfacial en- and should only be taken as relative values. Nicholson
ergy, and the driving force associated with local super- [32] has pointed out that although the present theory of
saturation, for example. heterogeneous nucleation is qualitatively satisfactory, it
Additional driving force for the preferential nucle- cannot be applied in a quantitative form in real systems
ation of T\ and & near a matrix dislocation comes from because of the lack of accurate interfacial energy data
the reduction of the energy barrier for that particular relating to a defect and a nucleus. However, if we as-
precipitate. In the case of AliCu, 0" is the preferred sume that the values in Table 6 are correct relative to
matrix precipitate when heterogeneous sites are absent one another we can discuss the competitive precipitation
because of the high coherency between 0" and the ma- kinetics for this alloy, which is in strong agreement with
trix. Despite the lowest relative driving force due to the quantified data from this investigation. As in Table
thermodynamic stability, 0" nucleates due to the mini- 3, 0" nucleation dominates the matrix for short times in
mal interfacial energy barrier [23]. Once heterogeneous the absence of heterogeneous sites, namely dislocations.
nucleation sites with sufficient strain fields, e.g., dislo- In the presence of defects that lower the energy barrier
cations, are introduced into the matrix the associated for nucleation, both Of and T\ plates nucleate and grow
strain lowers the misfit for the thermodynamically more due to the enhanced thermodynamic driving force.
stable phase, &, allowing it to preferentially nucleate and Although the interfacial energies of & and T\ are
grow [23,24]. This explains the shift from 0" to & be- similar, their corresponding volumetric strains differ by
tween the unstretched and stretched conditions, respec- an order of magnitude. In addition, their shear strains,
tively. although very similar, lie on different planes and along
Cahn [25] has derived equations for the rates of het- different directions, Table 9 [33,34]. The shear strain
erogeneous nucleation at various defects in a solid. For associated with ff has a magnitude of 0.333 on the
nucleation on dislocations, Cahn assumed elastic and (0 0 l)a plane in the [1 0 0]a direction, while the shear
surface isotropy and an incoherent nucleus and ex- strain associated with T\ has a magnitude of 0.354 on the
presses his results in terms of a dimensionless parameter (1 1 l)a plane in the [1 1 2]a direction [33,34]. Since the
[26] volume strain energy is substantially smaller than the
a = äGvfjJr/lTca1,
shear strain energy, the manner in which the shear
(1) component of the strain associated with the formation
where AGV is the change in volume-free energy, \i the of the precipitate is accommodated may be the key
shear modulus, b the Burgers vector and a is the inter- factor in controlling nucleation behavior [33]. Since the
facial energy required to form a new surface. The ef- shear strain for T\ is in the plane of the Burgers vector of
fectiveness of a dislocation in catalyzing nucleation dislocations in the aluminum matrix, i.e., {111} planes,
increases as a increases and thus as AGy and the mag- dislocations may thus have a greater effect on its nu-
nitude of the Burgers vector increases as the surface cleation than that of ff. It thus appears that a combi-
energy decreases. nation of the larger AGV and the {1 1 1} shear strain for
The change in volume-free energy associated with the T\ aids in its preferential nucleation on dislocations over
precipitation of T\ may be larger than that associated that of 0 given an appropriate alloy composition. This
with Of since T\ is an equilibrium phase and ff is a is supported by our quantified data that the relative

Table 8
Calculated interfacial energies for precipitates

Type of precipitation [reference] S (AljLi) Interfacial energy (J/m:) r, (Al:LiCu)

<7fi zone 0" (AI:Cu) ff (Al;Cu)

In the presence of dislocations [27] 0.13-0.23

Matrix precipitation [28] 0.003
Matrix precipitation [29] 0.05
Matrix precipitation [30] 0.179
Matrix precipitation [31] 0.014
 B.M. Cable et al. I Journal of Light Metals 1 12001) 1-14
Table 9
Precipitate structural data [33,34]

Phases (f (AI:Cu) T, (Al3CuLi)

Crystal structure Tetragonal Hexagonal
Lattice parameter a = 0.404 nm a = 0.496 nm
c = 0.580 nm c = 0.935 nm
Orientation (0 0 1)„//(OOI), (0001)r,//(lll),
Relationship [0 0 1]„//[10 0], [1010]r,//(U0),
Habit plane {100}, {HI},
Volumetric strain -0.05698 -0.00723
Shear strain 0.33333 0.35355
Es/Ev 40 10 000

driving force for T\ nucleation and growth is greater fine microstructure resulting from high levels of me-
than that for & in AF/C 458 when matrix dislocations chanical stretch. Conversely,, the coarse precipitate
are present. structure of large 7", and 0 plates and S' spheres is due to
prolonged exposure at artificial temperatures for those
4.1.2. Isostrength specimens specimens experiencing a lower amount of plastic de-
The second aspect of this investigation was to formation prior to aging.
compare the microstructural evolution for specimens The quantitative precipitate data suggest that a large
with similar yield strengths as a function of the thermo- number of fine matrix plates and spheres, as in the high-
mechanical processing, namely the imposed pre-age stretch conditions, strengthen on the same order as a
mechanical stretch and the artificial aging time. much coarser microstructure with more significant grain
The tensile data in Table 2 convey a trend of in- boundary precipitation. This demonstrates the role of
creasing ductility for equivalent yield strength with the the dislocations as heterogeneous matrix nucleation sites
increasing amount of pre-age mechanical stretch and for both 7i and 0 precipitates.
therefore shorter artificial aging times. This structure- The trend of decreasing volume fraction for each
property relationship should be directly related to the precipitate with increasing pre-age stretch introduced
aspects of the gram boundary precipitation, namely the some controversy. It is paramount to consider the error
average dimension, volume fraction and grain boundary associated with the stereological correction method that
area covered. Each of these aspects is a function of the was used for our analysis. The current method em-
artificial aging time. With the longer the artificial aging ployed-may have significant error associated with cor-
time, the greater the grain boundary precipitation and recting an inhomogeneous microstructure comprised of
the lower the ductility of the specimen. Nevertheless, truncated and very large plates. The measured diameter
note that each specimen exhibited at least 9% elongation for any truncated plate was significantly increased by
even after artificial aging for 120 h, which is nearly twice the correction program in order to normalize the aver-
the acceptable value for aerospace applications [18]. age plate diameter, bringing truncated and finer plates
As shown in Table 4, the quantified precipitate mi- closer to the mean length. This method enhances the
crostructure of the various conditions illustrates three corrected diameter and volume fraction values for the
linear trends with the increasing mechanical stretch. coarse microstructures of the non- and low-stretch
First, the number densities of both T\ and 0 were found specimens. The decreasing calculated volume fraction
to increase with the level of pre-age stretch, which relates for r, and 0 may simply be accentuated by the correc-
to the enhanced matrix dislocation density and aging tion method. As in the analysis for the isochronal
kinetics. The number density of 5' was also found to treatment there is significant error associated with the
increase with the level of imposed stretch. The disloca- calculated volume fraction of 5', which prohibits quan-
tion density directly relates to the number of precipita- titative discussion.
tion events and, therefore, required aging time to There may also be thermodynamic suppressive forces
achieve a given strength. The fewer the matrix disloca- that explain the dramatic volume fraction trend in the
tions the longer the required aging time and the coarser quantified microstructural analysis. The high number
the microstructure. density of fine matrix plates may introduce significant
Similarly, the corrected plate diameter for each pre- capillary effects that essentially shift the phase bound-
cipitate was found to decrease with increasing pre-age aries for a given dislocation density, with greater the
stretch as illustrated in Figs. 5 and 6. This illustrates that number density the more significant the effect [35].
the plate length, or sphere diameter in the case of <5', is a Therefore, the greater the matrix precipitation of fine
direct function of the artificial aging time with a very plates, the greater the pressure effects which shift the
 B.M. Gable et al. I Journal of Light Metals I (2001) 1-14 13
phase boundaries essentially driving less of the equilib- case of only two kinds (A and B) of plate-like precipi-
rium phases to precipitate. tates (for instance ff and 7"( ). as shown in [16] the
Ultimately, this second investigation illustrated that computer simulations may offer an approach to find the
material of a given strength may have vastly different optimum competitive precipitate structure for the asso-
microstructures. The quantitative results indicate that a ciated strengthening stresses using a characteristic pa-
large number of fine precipitates strengthen just as well rameter KAB introduced as
as a fewer number of large precipitates with more sig-
nificant grain precipitation. Future studies should focus
on other related properties between materials with a
^.c(*)'/(*!£*%),
\D J I \y y .a sin0 y
B B m B
(3)
given strength. Investigations into the fatigue properties,
creep response and environmental resistance should be where the coefficient £ accounts for the effect of the
of particular interest for AF/C 458. orientation of the plates (£ = 3/4 for ff- and T\ -precip-
itates), D.\ and yA are the mean diameter and the aspect
4.2. Theoretical modeling and simulation ratio of the A-plates, respectively, and Z)B and yB are
those associated with the B-plates. The terms Vm.A and
Using the quantified data of T\, ff and 5' precipitate Vm#, 0A and 0B are the molar volumes and dihedral
structures for each specimen, we conducted dislocation angles of the A and B plate-like particles with reference
slip modeling (for details of the method, c.f. [16]) and to the slip plane. For more complex structures consisting
evaluated the associated yield strength. In the simula- of three or more kinds of precipitates, as in AF/C 458,
tion, we assumed that T\ and ff with {111} and {10 0} more structural parameters are necessary, which may be
habit planes, respectively, to be unshearable during obtained through further experiments and computer
small plastic deformation until yielding. The 5' precipi- simulations, to determine the optimum structure. Based
tate required an order-strengthening mechanism with on the simulations, the practical approach of imple-
strength menting a pre-age stretch and aging treatment may be
developed to acquire the optimum structures for the
F = it/lrgy^ - ■ (2)'" given mechanical properties "of the alloy.
where rg is the radius of the 5' particle and the anti-
phase boundary energy yapb = 0.07 J/m2. Approxi-
mately 8000-15 000 obstacles formed by the intersection 5. Conclusions
of the T\, & and 5' with a {1 1 1} plane were taken into
account in a square of certain area (according to the 1. Pre-age plastic deformation enhances the aging kinet-
volume fraction of the particles) in the slip plane. An ics, strength and number density of fine strengthening
applied stress under which a dislocation line slips out of precipitates through the introduction of heteroge-
the area, that is, a micro-slip (plastic deformation) is neous matrix nucleation sites.
generated, is assumed to correspond to the contribution 2. The pre-age deformation significantly affects the com-
o-p of the precipitates to the total yield stress. Strength- petitive matrix precipitation between T\ and ff.
ening from grain boundaries and dispersoids (A^Zr and 3. The relative volume fraction of T\ to ff is greatly in-
Al2oMn3Cu2) are considered relatively small and con- creased with increasing amounts of deformation prior
stant throughout aging. Solid solution strengthening to artificial aging. The larger AGV and the accomoda-
from Li and Cu in the matrix was considered; every 1 tion of the {111} shear strain for T\ versus the
at.% Li gave rise to a 7 MPa increase [36] and 1 at.% Cu {10 0} shear strain for ff aids in its preferential nu-
gave rise to 4 MPa [37], while the matrix contribution cleation on dislocations over that of ff.
was 45 MPa [37]. 4. Specimens with a given strength may have vastly dif-
Table 5 summarizes the results for the specimens aged ferent microstructures without changing chemistry;
at 150°C for 24 h after different stretches and Table 6 for only a single thermo-mechanical processing step is
the specimens aged to the yield strength of 450 MPa necessary.
under various thermomechanical conditions. These ta- 5. Combining the dislocation-slip modeling with exper-
bles illustrate that the computer simulation furnished an imental tests, a practical approach to the optimiza-
excellent method to evaluate and predict the yield tion of mechanical properties for AF/C 458 may be
strength associated with the various complex precipitate developed.
structures generated by the pre-age stretch and artificial
aging treatments.
As previously described, the competitive precipitation Acknowledgements
between T\, ff and 8 and the associated mechanical
properties are extremely sensitive to pre-age deforma- The authors would like to acknowledge the helpful
tion as well as the artificial aging times. For a simple discussions with Professors Gary J. Shiflet and James M.
 14 B.M. Gable et al I Journal of Li^ht Metals 1 (2001) 1-14

Howe and financial support of the Air Force Office of [18] A.A. Csontos, E.A. Starke Jr.. Metall. Mater. Trans. A 31A
Scientific Research under Grant No. F49620-97-1-1034, (2000) 1965-1976.
Drs. Spenser Wu and Craig S. Hartley, Program Mon- [19] P.M. Kelly, A. Jostens, R.G. Blake, J.G. Napier. Phys. Stat. Sol.
A 31 (1975)771-780.
itors. [20] M. Otten, Philips Electron Optics - Applications Laboratory,
Building AAE, The Netherlands.
[21] S.A. Saltykov, Stereometrische Metallographie, Deutcher Verlag
fur Grundstoffindustrie, Leipzig, Federal Republic of Germany,
References
1974.
[22] E.E. Underwood, Quantitative Stereography, Addison-Wesley.
[1] J.M. Silcock, J. Inst. Met. 88 (1959-60) 357-364. Reading, MA, 1970.
[2] B. Noble, G.E. Thompson, Met. Sei. J. 6 (1972) 167-174. [23] D.A. Porter, K.E. Easterling, Phase Transformations in Metals
[3] F.S. Lin, S.B. Chakrobortty, E.A. Starke Jr., Metall. Trans. A and Alloys, second ed.. Chapman & Hall, London, 1993
13A (1982) 401-410. ' (reprinted).
[4] R.E. Crooks, E.A. Starke Jr., Metall. Trans. A 15A (1984) 1367- [24] E. Hornbogen. in: E.A. Starke Jr., T.H. Sanders Jr., W.A.
1377. Cassada (Eds.), Seventh International Conference on Aluminum
[5] R.F. Ashton, D.S. Thompson, E.A. Starke Jr., F.S. Lin. in: C. Alloys: Their Physical and Mechanical Properties, Trans Tech
Baker, P. Gregson, S. Harris, C. Peel (Eds.), Aluminum-Lithium Publications Ltd, Switzerland, 2000, pp. 879-888.
III, vol. Ill, The Institute of Metals, London, 1986, pp: 66-77. [25] J.W. Cahn, Acta Met. 5 (1957) 169-172.
[6] R.J. Rioja, P.E. Bretz, R.R. Sawtell, W.H. Hunt, E.A. Ludwic- [26] J.W. Martin, R.D. Doherty, B. Cantor, Stability of Microstruc-
zak, in: Proceedings of the International Conference on Alumi- ture in Metallic Systems, second ed., Cambridge University Press,
num Alloys - Physical and Mechanical Properties, vol. Ill, Cambridge, MA, 1997, p. 58.
Chameleon Press, London, 1986, pp. 1781-1797. [27] J.C. Huang, A.J. Ardell, in: G. Champier, B. Dubost, D.
[7] J. White, W.S. Miller, I.G. Palmer, R. Davis, T.S. Sani, in: C. Miannay, L. Sabety, (Eds.). Aluminum-Lithium IV, vol. IV, les
Baker, P. Gregson, S. Harris, C. Peel (Eds.), Aluminum-Lithium editions de physique, Les Ulix Cedex, France, 1987, pp. 373-
III, vol. Ill, The Institute of Metals, London, 1986, pp. 375-383. 383.
[8] J. White, W.S. Miller, in: G. Champier, B. Dubost, D. Miannay, [28] L. Lochte, A. Gitt, G. Gottstein, I. Hurtado, Acta Mater. 48
L. Sabety, (Eds.), Aluminum-Lithium IV, vol. IV, les editions de (2000) 2969-2984.
physique, Les Ulix Cedex, France, 1987, pp. 425-431. [29] E. Hornbogen,-Aluminium 43 (1967) 41, 115, 163.
[9] R.J. Sinko, T. Ahrens, G.J. Shiflet, E.A. Starke Jr., in: T.H. [30] D.Y. Li, L.Q. Chen, Acta Mater. 46 (1998) 2573-2585.
Sanders Jr., E.A. Starke Jr., (Eds.), Aluminum-Lithium Alloys V, PI] S.F. Baumann, D.B. Williams, Scripta Metall. 18 (1984) 611-616.
vol. V, 1989, pp. 375-383. [32] R.B. Nicholson, Phase Transformations, Chapman & Hall,
[10] T. Ahrens, E.A. Starke Jr., in: T.H. Sanders Jr., E.A. Starke Jr., London, 1970, p. 269.
(Eds.), Aluminum-Lithium Alloys V, vol. V, 1989, pp. 385-396. [33] J.F. Nie, H.I. Aaronson, B.C. Muddle, in: M. Koiwa, K. Otsuka,
[11] W.A. Cassada, G. J. Shiflet, E.A. Starke Jr., Metall. Trans. A 22A T. Miyazaki (Eds.), The Japan Institute of Metals Proceedings,
(1991) 287-297. Volume 12 (JIMIC-3), Solid-^Solid Phase Transformations, Part
[12] W.A. Cassada, G.J. Shiflet, E.A. Starke Jr., Metall. Trans. A 22A I, 1999, pp. 1128-1135.
(1991) 299-306. [34] B.C. Muddle, J.F. Nie, in: M. Koiwa, K. Otsuka, T. Miyazaki
[13] J.D. Kim, J.K. Park, Metall. Trans. A 24A (1993) 2613. (Eds.), The Japan Institute of Metals Proceedings, Volume 12
[14] S.P. Ringer, B.C. Muddle, IJ. Polmear, Metall. Mater. Trans. A (JIMIC-3), Solid-Solid Phase Transformations, Part II, 1999, pp.
26A (1995) 1659-1671. 157-160.
[15] A.K. Hopkins, K.V. Jata, R.J. Rioja, Mater. Sei. Forum (1996) [35] G.J. Shiflet, Department of Materials Science and Engineering,
421^*26. University of Virginia, Personal communication 4/26/2000.
[16] A.W. Zhu, A.A. Csontos, E.A. Starke Jr., Acta. Mater. 47 (6) [36] B. Noble, S.J. Harris, K. Dinsdale, Met. Sei. 16 (1982) 425.
(1999) 1713-1721. [37] J.R. Davis, Aluminum and Aluminum Alloys - ASM Special
[17] Annual Book of ASTM Standards, Metals Test Methods and Handbook, The Materials Information Society, Materials Park,
Analytical Procedures Section 3, ASTM, Philadelphia, PA, 1993. OH, 1993, pp. 49.
Materials Science Forum Vols. 331-337 (2Ü00) pp. 1333-1341) APPlTlVmTY TTT
© :00U Trans Tech Publications. Switzerland ArrufNLHA ill

The Effect of Quench Rate on the Microstructure and Properties of

AF/C-458 and AF/C-489 Al-Li-Cu-X Alloys

A.A. Csontos1, B.M. Gable1, A. Gaber2 and E.A. Starke, Jr.1

' Department of Materials Science and Engineering. University of Virginia,
Charlottesville, VA 22901, USA
2
Physics Department, Assiut University, Assiut, 71516. Egypt

Keywords: Aerospace Materials, Aluminium-Lithium Alloys, Pre-Age Stretch, Quench Sensitivity

ABSTRACT
The Air Force recently developed two isotropic Al-Li-Cü-X allovs with l.ST,<.Li and
2.r/.»Li designated AF/C-458 and AF C-489, respectively. The objective of this investigation was
to determine the effect ofquench rale on the microstructure and mechanical properties o'fthe AF C-
458 and AF/C-489 alloys. Tli.VI. SEM. microhardness, and tensile testing were utilized to ascertain
these microstructure/property relationships for both alloys in the T4, T6. and T86 tempers as a
function ofquench rate. Subsequent losses in ductility for both alloys in all tempers with
decreasing quench rate were determined to be due to the precipitation of the equilibrium AlCuLi
(Ti) phase along subgrain and grain boundaries which promoted intergranular fracture
Furthermore, yield and tensile strengths increased for both alloys in the T4 temper but decreased in
the T6 and T86 tempers with decreasing quench rate. The friereased strengths for the T4 condition
resulted from the heterogeneous precipitation of coarse T, and naturally aged Ö' phases The
decrease in yield and tensile strengths for the Tft and T86 tempers were also due to the coarse
heterogeneous precipitation ofT, which denuded regions of Cu thereby reducing the number
ienSA,tryv0t'!i,!,C matriX °"(T6) ;UUl T| ,TS(,)- ';inally' * comPari.son of the quench sensitivitv for both
he AF/C-458 and AF/C-489 alloys indicates that the mechanical properties for both allovs' were
less quench rate sensitive than other typical Al-Li-Cu-X alloys.

I. lNTiu)»r<TioN& BACK<;ROI;ND
In this era of streamlined economics for the aerospace industry, weight reduction of
aerospace vehicles is best achieved through reducing the density of the aircraft's structural
materials. For aluminum alloys, lithium is added to reduce the alloy's density with the added
oenet.t of increasing the strength and elastic modulus of the alloy. Most Al-Li-X allovs have
achieved approximately 10% reduction in density and approximately 10% increase in elastic
modulus [ i -4]. The major strengthening component of binary Al-Li alloys arises from the
precipitation of the metastable spherical AhLi (5') phase. The 5' possesses a LN ordered structure
which nucleates more or less homogeneously in Al-Li-Cu-X alloys upon quenching or natural adnu
«"th hth.um contents greater than 5.8a/«Li (1.6w/„Li). Due to their ordered structure and high " "
degree of coherency with the matrix. 5* precipitates can lead to undesirable planar slip [5] by the
concomitant shearing of these 5' with each passing dislocation which effectively softens the active
s«'P plane, thus, resulting in coarse planar slip. The addition of Cu and Mg to Al-Li alloys lowers
<ne solubility of lithium which drives the 5' precipitation and may lead to precipitation of other
strcngthcmng phases that may homogenize deformation, namely AI,Cu (81 & 9") AKCuMn (S1) and
AI2CUÜ (T,). In addition, phases such as AILi (8). Al6CuLi, (T2) and AbCtuLi (TB)"may "
precipitate through artificial aging or slow cooling from the solution heat treatment in
supersaturated alloys at intermediate temperatures [6.7].
Quench sensitivity is a measure of the precipitation of these metastable and equilibrium
pnases during an imposed quench from the solutionizing anneal. In Al-Li-Cu-X alloys, the
1334 Aluminium Alloys - Their Physical and Mechanical Properties

dissolved elements have a decreasing solubility with decreasing temperature, thus, when the alloy ^
cooled from the solutionizing anneal, the quench rate determines whether a supersaturated solid
solution forms or rather solute rich precipitates form. If the quenching rate is sufficiently rapid
there is insufficient time for atomic migration to form precipitates, thus, forming a supersaturated
soiid solution. However, at slow quenching rates, there may be sufficient time for solute atomic
migration, which promotes precipitation at high-energy location's, e.g. grain boundaries, of the
equilibrium phases. The type, size, and location of these precipitates strongly affect the subsequent
microstructural evolution, fracture behavior, and mechanical properties of these alloys.
Typically, the more solute present in the alloy, the higher the quench sensitivity of the alloy.
Alloys which fail to achieve their optimum properties due to the implementation of slower quench
techniques are considered to be quench sensitive. Limited information regarding the effect of
quench sensitivity on the microstructure and mechanical properties of Al-Li-Cu-X alloys has been
established [8-15]. A quantitative study has been performed elsewhere [8] which relates the
quenching rate to the microstructure and toughness of both lean and rich AA8090 alloys. The
obtained results of this study conveyed a good correspondence between the loss of fracture
resistance with the increase in volume fraction of T: and AhCuMg (S) phases. The most rapid loss
in fracture resistance for both alloys occurred as a result of isothermal aging at relatively high
temperatures (400 - 450°C) where T; precipitates were dominant. These results were supported by
another study [9] that showed that the influence of quenching rate on the mechanical properties of
AA8090. AA2091, and two CP276 alloys were temper dependent.

II. EXPERIMENTAL PROCEDURE

The unrccrystalli/ed 'A" thick AF/C-458 (Al-2.7Cu-l.8Li-0.6Zn-O.3Mg-O.3Mn-0.U8Zr
(wt%)) and 'A" thick AF/C-489 (AI-2.7Cu-2.1 Li-0.6Zn-O.3Mg-O.3Mn-O.O5Zr (wt%)) plates used in
this study were fabricated by Alcoa and the Air Force Research Labs, respectively. Both alloys
were direct chill cast, stress relieved, homogenized, hot rolled, annealed at 540°C for 4 hrs. and then
slow cooled before hot rolling again. This was followed by a solution anneal at 54()°C for 1 hour
and an immediate water quench. A 6% pre-age stretch was provided, therefore, the T36 temper for
both AF/C-458 and AF/C-489 alloys served as the "as-reccived" material from which all
solutionizing and quenching operations occurred.
Tensile specimens were machined in accordance with ASTM E8-93 from the received plates
so that the gauge lengths were parallel to the rolling direction. The tensile tests and pre-age
stretching were performed on a MTS 10/GL Sintcch test frame at a constant crosshead displacement
of 3mm/min. Vickers microhardness specimens were machined from their respective received plate
with dimensions of 4x5x10 mm. The Vickers microhardness measurements were performed using
the Vickers indenter on polished surfaces under a load of 4.9 Newtons for a duration of 10 seconds.
Each hardness value was obtained from the average of at least ten random measurements. Both
tensile and Vickers microhardness specimens were solution heat treated in a Lindberg three-zone
furnace at 550°C for 30 minutes and subsequently quenched in water, glycerol, or.still air and
immediately stretched and/or aged. For the T86 temper, the 6% pre-age stretch was performed
immediately upon quenching from the solution heat treatment to minimize the effects of an
inhomogeneous planar dislocation structure due to the natural aging of the fine shearable 5'
precipitates. The tensile and Vickers microhardness samples were then aged in a Stabil-Therm
forced air box furnace at 150°C for 32 hours and 24 hours for the T6 and T86 tempers, respectively.
To quantitatively measure the quench rate of both the tensile and Vickers microhardness samples, a
hole approximately 1.5mm in diameter was drilled on one side of dummy specimens to a depth of
approximately 2mm. A K-type thermocouple was inserted tightly in the hole to monitor the
temperature change with time during the quench. The resulting data yielded quenching rates for the
tensile samples to be 290°C/sec in water, 68°C/sec in glycerol, and 1.8°C/sec in still air while the
 Materials Science Forum Vols. 331-337 1335

smaller Vickers specimens were determined to be 500uC/sec in water, 130°C/sec in glycerol. and
3°C/sec in still air. Digital SEM fractographs were acquired through the use of a DSM982 Gemini
field-emission SEM operating at lOkV equipped with dual backscatter electron detectors, one
perpendicular and the other at 45° to the surface. A JEOL 2000FX conventional microscope was
utilized forTEM after samples were mechanically thinned, punched into 3mm disks, and
electropolished in a Tenupol twin-jet electropolisher with a 3.5:1 methanol: nitric acid solution.

III. RESULTS & DISCUSSION

As-Received T36 Condition
Prior research [16] quantitatively demonstrated that the smaller grained AF/C-458 alloy
possess an aspect ratio roughly twice that of the larger grained AF/C-489 plate where the number of
grains were 2.5 to 3 times greater in the AF/C-458 alloy than in the AF/C-489 plate. Furthermore,
the unrecrystallized grain structure of both alloys was confirmed by low magnification TEM which
demonstrated an approximate 5-10um sub-grain size.
As-Quenched and Naturally Aged T4 Condition
Due to 8' precipitation during the quench and resulting natural aging, analysis of the T4
temper is separated into the as-quenched T4 and naturally aged T4 conditions, which represents
those microstructures and mechanical properties acquired within the first 2 hours after quenching
and after 48 hours at room temperature, respectively. Previous investigations [16] indicate that 6'
natural aging occurred rapidly after quenching and nearly reached peak hardness after just 8 hours
at room temperature. The as-quenched microhardness values for both AF/C-458 and AF/C-489 are
illustrated in Figure I. The higher Li content AFC-489 alloy exhibits greater hardness values over
the leaner AF/C-458 at the slower quench rates, with only a slight increase in hardness for the water
quench. As seen in Figures 2u&b. the difference in the hardness with decreasing cooling rate from
the solution heat treatment may be directly attributed to o' and T( precipitation during the quench.
The as-quenched T4 anil naturaily aged T4 tensile data for the water, glycerol and air
quenched specimens are illustrated in Table I. The water and glycerol quenched samples, for both
AF/C-458 and AF/C-489, exhibited similar properties, however, the air cooled specimens displayed
strengths 30-40% higher and elongations f)2% and 82% lower, respectively. As seen in Figures
2a&b, the significantly increased strengths and reduced ductility for the air cooled T4 AF/C-458
and AF/C-489 alloys is associated with coarse T, precipitation during the quench. Furthermore, the
slightly higher strengths in the water quenched over the glycerol quenched specimens was most
likely due to the initial stages of o' natural aging where the increased number of quenched-in
vacancies from the rapid water quench enhanced Li dil'fusivity thereby o' precipitation and the
strengthening effect. Also, note the mechanical properly differences between the two alloys where
no -i —— . , the AF/C-458 exhibits higher
& VF/C458 AF/C489 strengths and elongations than
the AF/C-489 alloy for similar
conditions. As previously
reported [ 16], the differences in
strength and ductility arise from
the much finer grain structure of
the AF/C-458 alloy in
comparison to the AF/C-489
60-
plate, which acts to reduce the
so j | slip length, subsequent strain
100 200 :oo 400 500 localization, and premature
Quenching rate <C tec) intergranular fracture. In other
Figure I. As-quenched Vickers Microhardness vs. Quenching Rate words, the increased strengths
1336 Aluminium Alloys - Their Physical and Mechanical Properties

Water Quenched Glycerol Quenched Air Cooled

Jif3*-

Figure 2. a-b) As-quenched T4 SEM Iractographs & TEM micrographs B=[011). g=-' 4200 ]o\ g-'/i[l 11]T,.

and elongations for the finer-grained AF/C 458 alloys with qualitatively similar low precipitate
volume fractions may be attributed to Hall-Petch strengthening and reduced .-.train localization,
respectively.
As shown in Table 1 for the AF/C-458 alloy, the water and glycerol quenched specimens
exhibit higher strengths and lower ductility due to the natural aging of the o precipitates, i.e. T4 2 H
vs. T4 48 H. Vickers microhardncss data also demonstrates this trend for the water and glycerol
quenched samples, however, no natural aging trend was found for the air cooled condition. As seen
in Figure 2a-air cooled, the lack of natural aging can be attributed to the air cooled microstructure
where fine o' particles have already precipitated to a large extent during the quench. Finally, both
the AF/C-458 and AF/C-489 alloys were determined to be less quench sensitive when compared to
both T4 AA81W0 and AA2090 mechanical properties as a function of quench rate.
AI-/C-458 Allov \K r-489 Allov
YS UTS Elongation YS I;TS Elongation
Quench Media Aging
Temper (MPa) (MPa) (%) (MPa) iMPal (%)
T4 2H 161 287 25.3 142 268 23.2
Water T4 48H 205 352 19.4
T6 323 446 12.5 288 427 11.6
Quench
TX6 501 537 10.3 500 554 4.8
(290"C'sec)
T4 2H 150 275 26.1 129 255 23.3
Glycerol T4 48H 146 303 21.3
Quench T6 296 416 13.2 231 363 8.5
T86 360 413 6.6 334 392 3.9
(68"C/sec)
T4 2H 225 369 15.6 189 337 12.8
Air
Cooled T6 299 421 9.8 214 362 8.2
(l.8"C/sec) T86 1 411 455 69 412 433 3.0
, "
Table 1. Tensile Data for the Various Quenching Conditions
 Materials Science Forum Vols. 331-337 1337

Tft Condition M50°r for 32 hrsl

After the solutionizing anneal and respective quenches, the tensile samples were then peak
aged at 150°C for 32 hours and tested to obtain the mechanical property data in Table 1. The
artificial age greatly increased the strength and reduced the ductility for both alloys in each
quenching condition respective to the T4 temper. For both alloys, the water quenched specimens
exhibited the best combination of high strengths and ductility while the air cooled samples showed
the least ductility with strengths comparable to the glycerol quenched conditions. As seen in
Figures 3a&b. these trends can be directly related to the fracture behavior and microstrucrure of
both alloys as a function of quench rate.
The SEM fractographs clearly show the grain size variation between the alloys and that both
water quenched and AF/C-458 glycerol and air cooled samples fractured at 45° and contained
mostly transgranular shear, whereas the glycerol and air cooled AF/C-489 contained mostly
intergranular fracture. Furthermore, the TEM investigation of both alloys generally found larger
and higher volume fraction 6' and sub-grain and grain boundary T| precipitates coupled with a line
and highly dense matrix dispersion of Ö" with respect to the T4 temper. In particular, the water
quenched specimens for both alloys exhibited similarly sized Tt plates along sub-grain and grain
boundaries as well as residual matrix dislocations and incoherent dispersoid/matrix interfaces while
the glycerol and air cooled samples showed a bimodal distribution of similarly located T,. This T:
bimodal distribution suggests that the larger T| plates nucleated during the slower quench and grew..
with the artificial age while the smaller T| nucleated and grew only during the age. These large T
plates also effect the competing 0" precipitation as seen by the glycerol quenched AF/C-489 TliM
micrograph. The large grain boundary penetrating T| plate denuded the adjacent area ofCu. which
prevented the line 0" precipitation and produced a precipitate free zone (IM'Z) .surrounding the T|
plate. As seen by the slower quenched mechanical property data, these PFZs reduce the strength
and ductility of these alloys by preventing matrix precipitation of the major strengthening 0" phase
for this T6 temper and by promoting strain localization, respectively.
Also note that the AF/C-458 strength and ductility levels were 5-40% and 8-55% greater,
respectively, when compared to similar AF/C-489 samples with the least difference from the water
quenched condition. This disparity between the alloys indicates that along with the differences in
Li content and, thus, o' volume fraction, grain size may also have a large effect on the mechanical
properties which may not be due to the quenching conditions specifically. Initial observations, as
seen in Figures 3a&b. clearly suggest that the differences between the alloys may be due in part to
the grain size variation and correspondingly to the increased propensity for intergranular fracture
with decreasing quenching rate. e.g. AF/C-489. As mentioned previously, earlier work [ 1 <■>]
quantitatively showed that the grain size and o' volume fraction was ~3 times larger and -3%
greater, respectively, in the 2.1*'..Li AF/C-489 alloy when compared to the l.8w/oLi AF/C-458 alloy
in the T36 and T86 (150°C 24 hour) tempers. We may assume for the T6 condition that the -3% <>'
volume fraction difference between the two alloys should not change dramatically since the (>% pre-
age stretch should only affect the amount of the spherical and coherent ö'(AKLi) indirectly through
the preferred precipitation of T,(Al;CuLi) at the expense of G" (AI2Cu). Thus, the lower strength
and ductility levels of the T6 AF C-489 alloy may be attributed to the significantly larger grain size
and increased volume fraction of shearable 5' which both substantially increase the amount of strain
localization at grain boundaries, resulting in more low energy intergranular fracture.
186 condition (150°C for 24hrs)
A pre-age 6% stretch was incorporated for both alloys following the solution anneal and
corresponding quenches to increase the competitive precipitation of Ti during the 150°C 24 hour
age. The 6% pre-age stretch was utilized because Ti nucleation has been shown to be greatly
 1338 Aluminium Alloys - Their Physical and Mechanical Properties

Water Quenched Glycerol Quenched Air Cooled

speeimere exhibited fte hid« sSJa,5 ÄL L", J'°°°&Am- Xh° nm ^^

 Materials Science Forum Vols. 331-337
1339

Water Quenched Glycerol Quenched Air Cooled

Figure 4. a&b) T86(I50"C 24H) AFC-45X * AF/r_j«o ecu r

and varying amounts of intergranular SrVwiSw S^ u*" •AF/C-458 Water Wh

6
Although not presented here for bn^Ä^W "Ä* SEM
*"" C°°,ed AF/CJW9 all«.v-
•ntergranular fracture did play a role nth !wZfh A ^^tography indicated that
conditions. The significant l^^tXl^ ^«f*** and * cooled
'"crease in matrix T, volume fnSon andtmh/r 7 T'"" be d'reCtly rclated t0 the
me tracnon and „umber dens.ty at the expense of the 9" precipitate. Also.
1340 Aluminium Alloys - Their Physical and Mechanical Properties

note again the bimodal distribution of the T, plates that formed during the slower quenches for both
AF/C-458 and AF/C-489 samples. Furthermore, the only noticeable difference to account for the
10-20% lower strengths of the glycerol quenched as compared to the air cooled samples was the
apparent lower 5' volume fractions for the glycerol quenched specimens, although more work on
this point needs to be addressed. Once again, the low ductility of the AF/C-489 alloy in comparison
to the AF/C-458 alloy of similar quenching conditions was attributed to the -3% higher volume
fraction of 5' coupled with the larger grain size. Thus, both factors significantly increase the
amount of strain localization and stress concentrations at grain boundaries, which resulted in more
low energy intergranular fracture for the AF/C-489 alloy.

IV. CONCLUSIONS
This research has shown the effect of quench rate on the microstructural evolution and
overall mechanical behavior of the T4. T6. and T86 AF/C-458 and AF/C-489 alloys. A more rapid
quench was determined to be more desirable for optimizing the strengths and ductility for both
alloys. The AF/C-489 alloy appears to be more quench sensitive than AF/C-458 due to the higher
solute content and larger grain size. Nevertheless, both AF/C-458 and AF/C-489 alloys were shown
to maintain lower quench sensitivity values than their T4 Al-Li-Cu-X counterparts AA8O90 &
AA2090.

ACKNOWLEDGEMENTS
This research was funded under Air Force Office of Scientific Research Grant F49620-97-1-1034.

VI. RKKKRENCES
[ 11 D.B. Williams. R. Levi-setti. J.M. Chabala. Y.L. Wang, D.E., Newbury. Appl. Surface Sei. 37,
78(1989).
[2| A. Yamamato, H. Tsubakino, R. Nozato. 4lh Inter. Conf. On Al alloys, cd. by E.A. Starke. Jr.
T.I I. Sanders, Jr, Vol. II, Atlanta. Georgia. USA (1994) P. 191.
[31 A.F. Smith, J. de Physique. 48C3-49 & Cj-629 (1987).
[4| A.F. Smith. Mater. Sei. Technology, 5. 533 (1989).
|6| U.A. Ludwiczak and R.J. Rioja. Scripta Metall. Mat., 25, 1415(1991).
|7| K.S. Prasad. A.K. Mukhopadhyay, A.A. Gokhale, D. Benerjee and D.B. Goel. Scripta Metall.
Mat.. 30. 1299(1994).
[X| G.N. Coivin. E.A. Starke.Jr, SAMPE Quarterly. 19. 10 (1988).
[91 G.Lapasset, C.Damerval, M.Doudeau, 5,h Inter. Conf. on Al-Li alloys. Williamsbur« VA.
March 27-31, 365 (1989).
[10] S.Hirano, H.Yoshida, T.Uno. ibid. 335 (1989).
[11] J.T. Stalcy, R.D. Doherty, ibid. 345 (1989).
[12) M.E. Donnellan, W.E. Frazier. ibid. 355 (1989).
[13] J.T. Staley, Mater. Sei. Technol.. 3.923 (1987).
[14] J.T. Staley, PhD Thesis." Modeling quenching of precipitation strengthened alloys -
Application to an Al-Cu-Li Alloy", Drexel University (USA), 1989.
[15] P.A. Rometsch, G.B. Schaffer. Ji-Yong Yao, M.J. Couper, Proceed. 6* Int. Conf. on Al-AIloys,
Toyhashi-Japan. July 5-10,1998. Ed. T. Sato, S. Kumai, T. Kobayashi, Y. Murakami.
[16] A.A. Csontos. E.A. Starke, Jr., Met. Trans, In Press (2000)
[17] W.A. Cassada, G.J. Shiflet, and E.A. Starke, Jr., Metall. Trans. A., Vol. 22A, 1991, p. 287.
[18] W.A. Cassada, G.J. Shiflet, and E.A. Starke, Jr., Metall. Trans. A., Vol. 22A, 1991. p. 299.
[19] B. Gable, A.A. Csontos, and E.A. Starke, Jr., paper in progress
 DISTRIBUTION LIST

1-3 Dr. Craig S. Hartley

Program Manager, Metallic Materials
Directorate of Aerospace and Mechanical Sciences
Air Force Office of Scientific Research
801 N. Randolph St.
Arlington, VA 22203-1977

4 SEAS Postaward Administration

5-6 E. A. Starke, Jr.

7 W. A. Jesser

8-9 M. Rodeffer, Engineering Library, Clark Hall

10 SEAS Preaward Administration Files

JO#0840:ph