-
HEAT TREATING PROGRESS NOVEMBER/DECEMBER 2009 25
Cooling time-temperaturedata are not routinelyshown, and
correlations ofcooling rate data, strength,and intergranular
corrosionwith either residual stress ordistortion are rarely
reported together. This article addresses this issue.
Patricia Mariane Kavalcoand Lauralice C. F. Canale*University of
So PauloSo Carlos, SP, Brazil
George E. Totten, FASM**Portland State UniversityPortland,
Oreg.
*Member of ASM International**Member of ASM International and
member,ASM Heat Treating Society
luminum is solution treatedat temperatures generally inthe range
of 400 to 540C (750to 1000F). During solutiontreatment, some
alloying ele-
ments are re-dissolved to produce asolute-rich solid solution.
The objec-tive of this process is to maximize theconcentration of
hardening elementsincluding copper, zinc, magnesium,and (or)
silicon in the solid solution.The concentration and rate of
dissolu-tion of these elements increases withtemperature.
Therefore, solutionizingtemperatures are usually near the liq-uidus
temperature of the alloy[1,2].
If an aluminum alloy is slowlycooled from an elevated
temperature,alloying elements precipitate and dif-fuse from solid
solution to concentrateat the grain boundaries, small voids,on
undissolved particles, at disloca-tions, and other imperfections in
thealuminum lattice as shown in Fig. 1[2].For optimal properties,
it is desirableto retard this diffusion process andmaintain the
alloying elements in solidsolution. This is done by quenchingfrom
the solution temperature. For
quench-hardenable wrought alloys(2xxx, 6xxx, and 7xxx) and
casting al-loys such as 356, this is accomplishedby the quenching
process. The objec-tive is to quench sufficiently fast toavoid
undesirable concentration of thealloying elements in the defect
andgrain boundary structure while at thesame time not quenching
faster thannecessary to minimize residualstresses, which may lead
to excessivedistortion or cracking. After quenching,aluminum alloys
are aged, and duringthis process, a fine dispersion of ele-ments
and compounds are precipi-tated that significantly increase
mate-rial strength. The diffusion process andprecipitation kinetics
vary with thealloy chemistry.
The cooling process of age-harden-able aluminum alloys not only
affectsproperties such as strength and duc-tility, but it also
affects thermal stresses.Thermal stresses are typically mini-mized
by reducing the cooling ratefrom the solutionizing
temperature.However, if the cooling rate is too slow,undesirable
grain boundary precipita-tion will result. If the cooling rate is
too
A
QUENCHING FUNDAMENTALS
QUENCHING OF ALUMINUM ALLOYS:COOLING RATE, STRENGTH, AND
INTERGRANULAR CORROSION
Fig. 1 Schematic illustration of the solid diffusion processes
that may occur during solution heat treatmentof aluminum.
Aged
Aged
Slow cc
ooled
Quenched
At ssolution hheat ttreating temperature
At rroom ttemperature After aaging
-
fast, there is an increased propensityfor distortion. Therefore,
one of the pri-mary challenges in quench-process de-sign is to
select quenching conditionsthat optimize strength while mini-mizing
distortion, and at the same timeensure that other undesirable
proper-ties are not obtained, such as intergran-ular corrosion,
which is also coolingrate dependent.
Bates and Totten have addressed theselection of quenchants and
quenchingconditions that will optimize materialstrength and
minimize the potentialfor distortion[3]. However, although it
iswell-known that properties such ascorrosion resistance are also
coolingrate dependent, the problem ofquenching system evaluation
with re-spect to strength, distortion, and cor-rosion are rarely
evaluated together.The objective of this article is to pro-vide an
overview of intergranular cor-rosion (IGC) of aluminum alloys andto
illustrate the effect of cooling rate onboth strength and IGC.
Pitting and Intergranular CorrosionPitting is the most common
corro-
sion process encountered with alu-minum alloys, and is a major
cause invariations in the grain structure be-tween adjacent areas
on the metal sur-faces in contact with a corrosive envi-ronment.
Pitting results in theformation of very small holes (pits) inthe
surface, which are covered bywhite or gray powder-like
depositsappearing as blotches on the surface[4].
Intergranular (intercrystalline) cor-rosion occurs most commonly
in thefollowing aluminum alloys: Al-Cu-Mg(2xxx); Al-Mg (5xxx),
which is similarto the Al-Cu-Mg alloys; Al-Mg-Si(6xxx); and
Al-Zn-Mg-Cu (7xxx)[5,6].
(The 2xxx, 6xxx, and 7xxx series areheat treatable). IGC refers
to a selectivedissolution of the surface grain-boundary zone, and
typically, thegrains below the surface zone are notattacked (Fig.
2). As discussed above,upon cooling from the
solutionizingtemperature, alloying elements mayconcentrate at the
grain boundaries toform intermetallic compounds thatdiffer
electrochemically from the adja-cent matrix and the metal adjacent
tothe grain boundaries[7,8].
The critical temperature range andtime (s) for the transition of
pitting tointergranular corrosion is shown bythe so-called TTP
(time-temperature-property) or C-curve illustrated in Fig.3[9]. The
C-curve for 2024-T4 shows thechange in corrosion behavior by
cor-relating the critical temperature rangewhere precipitation was
fastest. In-creased intergranular corrosion for2024-T4 will be
favored if cooling ratesare excessively slow during
quenching.Similar behavior can be shown forother aluminum alloys.
Therefore, it isimportant that cooling rates duringquenching be
sufficiently fast to avoidthis undesirable behavior.
As the IGC process continues, exfo-liation will result, which
refers to thelifting of the surface grains caused byexpansion due
to increasing volumeof the corrosion products accumulatingin the
subsurface grain boundaries[10].It has been reported that
exfoliation re-sults in severely reduced structuralstrength,
plasticity, and fatigue. Exfo-liation in aircraft aluminum alloy
struc-tural materials is most often observedwith extrusions, where
the grain thick-nesses are often less than the rolledforms[4].
Electrochemical Behavior of IGC Processes
IGC is caused by the formation of amicrogalvanic cell between
the inter-metallic compounds formed in thegrain boundary during
cooling andthe adjacent metal. These intermetalliccompounds may be
either anodic orcathodic, with respect to the adjacentmetal,
depending on their composi-tion. Figure 4 illustrates these two
situ-ations[11]. In one case, noble (inactive)alloying elements may
precipitate inthe grain boundaries leaving a de-pleted zone
adjacent to the grainboundary, which is electrochemicallyactive
(Fig. 4a). Conversely, electro-chemically active alloying
elementsmay precipitate at the grain boundary,and then the metal
adjacent to the
26 HEAT TREATING PROGRESS NOVEMBER/DECEMBER 2009
200 m
200 m
Fig. 2 Sample cross section after corrosion testillustrating
intermetallic deposition in the grain
boundaries in the surface region (a); Microstructure of sample
cross section (b).
Corrosion testing according to BS ISO 11846:1995method B. Sample
material is a model AlMgSi
aluminium alloy (6000 series) extruded using aresearch extrusion
press with a reduction ratio of34:1. Sample dimensions = 78 2.7 mm.
Source:G. Svenningsen, Norwegian University of Science
and Technology, Trondheim, Norway.
(a)
(b)
Fig. 3 Curve indicating cooling rate dependent mechanism of
corrosion attack on aluminum alloy 2024-T4sheet.
K
700
600
500
0.1 11 110 1100 11000Critical ttime, ss
Predominantly ppitting
Predominantly iintergranular
480
425
370
315
260
205
150
Tem
pera
ture
, C
-
grain boundary will be noble (Fig. 4b).Note that corrosion
behavior has alsobeen shown to be due to microstruc-tural changes
from the heat treatmentprocess, which will not be discussedhere.
The reader is referred to Refer-ence 12 for more detailed
discussion.
To illustrate this process, considerIGC occurring in 2xxx or
7xxx alloysthat would be caused by the loss ofcopper or sufficient
magnesium inareas near the grain boundaries tocreate an anodic
electrochemical po-tential. The electrochemical potential(referred
to as electromotive force, orEMF) for various aluminum
alloysprovided in Table 1 shows that thepresence of copper in solid
solutionwith aluminum makes it more ca-thodic[13]. The cathode is
the positiveelectrode in an electrochemical circuit,and, therefore,
it is the electrode thatgains electrons or electrons flow fromthe
anode (which possesses the morenegative potential) to the
cathode(more positive potential). An alu-minum alloy containing 4%
copper insolid solution has an EMF of -0.69 V.However, copper
concentrations in thegrain boundaries may reduce the EMFto -0.84 V,
making it more anodic.Grain boundary corrosion may alsooccur when
the grain boundary pre-cipitates are more anodic than the ad-jacent
solid solution. For example, Mg2,Al3, MgZn2 and Alx-ZnxMg are
morecathodic than CuAl2 and AlxCuxMg[14].When two dissimilar metal
com-pounds with different electron affini-ties (EMF values) are
connected, thereis a potential for electrons to pass fromthe
material with the smaller affinityfor electrons (anode the more
nega-tive pole) to the material with thegreater affinity for
electrons (cathode the more positive pole). A potentialdifference
between the materials willincrease until equilibrium is
achieved.This equilibrium potential is definedas the potential that
balances the dif-ference between the propensity of thetwo metals to
gain or lose electrons.
IGC ControlThe degree of IGC may be con-
trolled by the selection of the temperand maximizing the cooling
rate thatwill provide minimum distortion. Forexample, the T4 and T6
temper condi-tions are typically selected when op-timum resistance
to IGC is required[15].Schuler reported that the criticalcooling
rates (cooling rates between750 and 550F) for the 7xxx series to
be
-
that often in the industry, cooling be-havior of various quench
media is de-termined using an Inconel 600 probe
according to ISO 9950 or ASTM D6200.However, the thermal
conductivity ofInconel 600 is much less than that ofaluminum as
shown in Table 3[17].Clearly, the low thermal conductivityof
Inconel 600 versus aluminum ren-ders this probe to relatively
insensitiveto the cooling properties experiencedby an aluminum
alloy duringquenching.
Silver probes are also used to eval-uate quench severity
exhibited by dif-ferent quenchants. Because of the simi-larity of
the thermal characteristics ofsilver and aluminum, and because
ofthe significantly lower oxidation ten-dency for silver relative
to aluminum,the cooling behavior of AlMgSiCu anda silver (99.5 %)
probe was compared[17].
Thermal conductivity () and spe-cific heat capacity of various
materialsare provided in Table 3. Thermal con-ductivity is a
measure of the rate ofpropagation of temperature change ina body
and is related to the specificheat capacity by:
=
x Cp
where is thermal diffusivity, Cp is spe-cific heat capacity, is
thermal conduc-tivity, and is density of the material.
Tensi, et. al., compared coolingcurves recorded during quenching
ofan aluminum (AlMgSiCu) and a silverspecimen (Ag 99.5) in a Type I
water-soluble polymer quenchant solution.The Type I aqueous polymer
quen-chant concentration was 10% byvolume and the bath temperature
was25C. The temperature of both probematerials when quenched was
520C.Both probes were cleaned with 600 gritabrasive paper before
each test. Thecooling curves obtained are shown inFig. 5.
28 HEAT TREATING PROGRESS NOVEMBER/DECEMBER 2009
Table 2 Effect of cooling rate on maximum intergranular
penetration of 7075 as a function of cooling rate
Cooling rate, C/s Sample ID (50 mm diam bar) Location(a) Depth
of attack, mmA 53 Surface 0.46
Center 0.56B 50 Surface 0.30
Center 0.86C 30 Surface 0.46
Center 0.61D 17 Surface 0.74
Center 1.09(a) Surface: within 3.2 mm of cylinder surface.
Center: within 3.2 mm of centerline of the cylinder. Source: Ref
16.
Table 3 Thermal conductivity and specific heat capacity for
different materials
Material Thermal conductivity, m2s-1 Specific heat, kJ kg-1
K-1Aluminum 99.5 95 10-6 0.896Silver 99.5 174 10-6 0.235Nickel 14
10-6 0.448CrNi Steel(a) 4 10-6 0.477Inconel 600(b) 4 10-6 0.465(a)
Austenitic stainless steel SAE 30304. (b) Nickel based alloy.
Fig. 5 Comparison of the cooling processes of a cylindrical
AlMgSiCu probe (15 mm diam. 45 mm) with those of a silver probe;
cooled into a 10% solution of a water-soluble polymer at 25C
(temperatures recorded at the geometric center of the probe);
solution treating temperature is 520C for the AlMgSiCu probe;
annealingtemperature is 800C for the silver probe; (a) changes in
temperature and conductivity as a function of time; (b) cooling
rate as a function of temperature.
0 110 220 330
Immersion ttime, ss
100
0
Con
duct
ance
((G),
%%
800
600
400
200
0
Tem
pera
ture
((TZ)
, C
(a)
TeAg 999.5
AlMgSiCu
Te = TTsAg 999.5
AlMgSiCuTs
ts AlMgSiCu ts Ag 999.5
0 2200 4400 6600 8800Temperature ((TZ), C
300
250
200
150
100
50
0
Coo
ling
rrate
, KK/s
(b)
Ag 999.5 AAlMgSiCu
A
Grain ddirection
A DDisc ssectioned aand ppolished through ddiameter aat pplane A
aafter eexposure
-
The polymer film surrounding theprobe surface ruptured
simultane-ously around the entire surface, alsocalled explosive
rewetting, for bothprobes, and the rewetting times (tf tS) were
extremely short. How-ever, a stable film-boiling range lastingabout
4 s was observed for the silverprobe, which was not observed for
thealuminum probe. The centerlineprobe temperature was about
440Cfor silver and 500C for aluminum.The reason that the rewetting
of thesilver occurs about 4 s later for thesilver probe is the
greater oxidationresistance of silver. Initially, the pres-ence of
surface oxidation will facili-tate film rupture, thus enhancing
theboiling process. In this case, sufficientoxidation had not yet
occurered onthe surface of silver relative to alu-minum, thus there
was a 4s delay inthe rewetting process for silver. (Theratio of
heats of formation of Ag2O2and Al3O is 0.05.) When the quench-ing
temperature of the silver probe isincreased to 800C, considerable
stabi-lization of the film-boiling occurs asobserved in Fig. 5a.
Wetting nowstarts at about 260C after 24 s com-pared with the start
of wetting of theAlMgSiCu probe after 1 s at about500C.
The main reason for this stabiliza-tion of the film-boiling
phase of theentire surface of the silver probe is thereduction of
the silver oxide at thehigher temperature. The high-temper-ature
annealing of the silver probe re-moves the oxide and leaves a
baremetal surface, resulting in stabiliza-tion of film boiling
during quenching,especially in water-soluble polymers.Accordingly,
the maximum coolingrate of the silver probe is not reacheduntil a
centerline probe temperature
of 200C as shown in Fig. 5b.If distilled water at room
tempera-
ture is used as the quenchant insteadof an aqueous polymer, the
silver andthe AlMgSiCu probes show almostidentical cooling behavior
with coin-ciding rewetting kinematics as shownin Fig. 6. This means
that thequenching behavior determined inwater with silver probes
can be safelycompared with those obtained for alu-minum probes.
However, whenpolymer solutions are used, there areclear
differences, especially with re-spect to initial wetting.
It is important that whatever probematerial is used to
evaluatequenching behavior to be expectedwith the aluminum alloy of
interest,it should properly model the actualquenching process
occurring, espe-cially if cooling rate dependence ofthe quenching
media and process isbeing used to determine the potentialfor IGC.
For these reasons, it is recom-mended that heat treaters
requestcooling curve data obtained using ei-ther an aluminum or
silver probe.(The advantage of silver relative toaluminum is its
inertness and there-fore it may be reused, whereas alu-minum alloy
probes typically cannot.)
ConclusionsAn overview of the IGC process
was provided and contrasted to pit-ting corrosion. It was shown
that thepropensity for IGC of heat treatablealuminum alloys is
cooling rate de-pendent. It is recommended that inview of IGC
processes in many appli-cations such as aerospace and auto-motive,
that IGC should be evaluatedalong with strength and
residualstress/distortion as important andmore routine screening
parameters
HEAT TREATING PROGRESS NOVEMBER/DECEMBER 2009 29
Fig. 6 Comparison of cooling processes of a cylindrical AlMgSiCu
probe (15mm diam x 45mm) with those of a silver probe; probes
quenched into distilled water at25 (probe temperatures recorded at
the geometric center); solution treating temperature is 520C for
the AlMgSiCu probe; annealing temperature is 520C for thesilver
probe; (a) changes in temperature and conductivity as a function of
time; (b) cooling rate as a function of temperature.
(b)
Ag 999.5 AAlMgSiCu
The propensity for IGC ofheat treatable aluminum alloys is
cooling rate dependent. It is recommended that in viewof IGC
processes in manyapplications such as aerospace and automotive,that
IGC should be evaluated along withstrength and
residualstress/distortion as important and more routinescreening
parameters thanis now performed.
0 55 110 115Immersion ttime, ss
600
500
400
300
200
100
0
Tem
pera
ture
((TZ)
, C 100
0
Con
duct
ance
((G),
%%
(a)
TeAlMgSiCu
Ag 999.5
AlMgSiCu
Ag 999.5
0 2200 4400 6600Temperature ((TZ), C
300
250
200
150
100
50
0
Coo
ling
rrate
, KK/s
-
than is now performed. Furthermore,cooling time-temperature data
ispreferably obtained using the alu-minum alloy of interest.
However,silver probes may, under the appro-priate conditions,
reasonably modelthe cooling behavior of the aluminumalloy of
interest, but this needs to bedemonstrated experimentally. HTP
References1. G.E. Totten, G.M. Webster, C.E. Bates,Chapter 20
Quenching, Handbook of Alu-minum Volume 1: Physical Metallurgyand
Processes, Eds. G.E. Totten and D.S.MacKenzie, CRC Press, Boca
Raton, Fla., p971-1062, 2003.2. T. Croucher, Quenching of aluminum
al-loys: what this key step accomplishes, HeatTreating, Vol 14, No.
5, May, p 20-21, 1982.3. C.E. Bates and G.E. Totten, Procedurefor
Quenching Media Selection to Maxi-mize Tensile Properties and to
MinimizeDistortion in Aluminum-Alloy Parts, HeatTreatment of
Metals, No. 4, p 89-98, 1988.4. E.F. Rudolph, Corrosion on
Aircraft,Sport Aviation, April, p 7-8, 1961.5. P.R. Roberge,
Handbook of Corrosion En-gineering, McGraw-Hill, New York, N.Y.,p
505, 1999.6. D.G. Altenpohl, Aluminum: Technology,
Applications and Environment A Profile ofa Modern Metal, 6th
Ed., The AluminumAssociation, Washington, D.C., p 234,1998.7. F.
Andreatta, H. Terryn, and J.H.W. deWit, Effect of solution heat
treatment ongalvanic coupling between intermetallicsand matrix in
AA7075-T6, Corrosion Sci-ence, Vol 45, No. 8, p 1733-1746, 2003.8.
F. Andreatta, F.M.M. Lohrengel, H.Terryn, and J.H.W. de Wit,
Electrochem-ical characterisation of aluminiumAA7075-T6 and
solution heat treatedAA7075 using a micro-capillary cell,
Elec-trochimica Acta, Vol 48, p 3239-3247, 2003.9. J.W. Evancho and
J.T. Staley, Kinetics ofPrecipitation in Aluminum Alloys
DuringContinuous Cooling, Metallurgical Trans-actions, Vol 5,
January, p 43-47, 1974.10. F.-H. Cao, Z. Zhang, J.-F. Li,
Y.-L.Cheng, J.-Q. Zhang, and C.-N. Cao, Exfo-liation corrosion of
aluminum alloyAA7075 examined by electrochemical im-pedance
spectroscopy, Materials and Cor-rosion, 55, No. 1, p 18-23,
2004.11. G. Svenningsen, Corrosion of Alu-minium Alloys, Department
of MaterialsTechnology, 7491 Trondheim, Norway, In-ternet:
http://www.sintef.no/static/mt/norlight/seminars/norlight2003/Postere/Gaute%20Svenningsen.pdf12.
F. Andreatta, H. Terryn, and J.H.W. deWit, Corrosion behaviour of
different tem-
pers of AA7075 aluminium alloy, Elec-trochimica Acta, 49, p
2852-2862, 2004.13. E.H. Dix, The Resistance of AluminumAlloys to
Corrosion, Metals Handbook, 8thEdition, Vol. 1-Properties and
Selection ofMetals, Ed. T. Lyman, ASM International,Materials Park,
Ohio, p 916-935, 1961.14. J.E. Hatch, Aluminum: Properties
andPhysical Metallurgy, ASM International,Materials Park, Ohio, p
260-272, 1984.15. M.D. Schuler, Intergranular Corrosion,University
of California, May 18, unpub-lished report, 1998.16. UCON Quenchant
A QuenchingAluminum for Minimal Distortion andLow Residual
Stresses, published byTenaxol Inc., Milwaukee, Wis.17. H.M. Tensi,
P. Stitzelberger-Jacob, andG.E. Totten, Quenching
Fundamentals:Surface Rewetting of Aluminum, Ad-vanced. Materials
and Processes, Vol 156, No.5, p H15-H20, 1999.
For more information: George Totten,FASM, Portland State
University, Depart-ment of Mechanical and Materials Engi-neering,
Portland, Oreg.; tel: 206-788-0188;e-mail: [email protected]; or
Lauralice C.F.Canale, University of So Paulo, Depart-ment of
Materials, Aeronautical and Auto-mobile Engineering, School of
Engineering,So Carlos, SP, Brazil; e-mail: [email protected].
Donate by Phone.800-336-5152
Contact a staff member by phone.
Chuck HayesExecutive DirectorExt. 5506
Pergentina DeatherageProgram AdministratorExt. 5533
Ginny Shirk Foundation AssistantExt. 5538
Donate Today! Talk to a person Friendly voice Get answers
Donate Online.www.asmfoundation.org
Support Materials Education
Your donations help support the Foundations educational
programs.
Donate Today! Quick Immediate impact Secure payment Credit card
record
Donate by Mail.
ASM Materials Education Foundation9639 Kinsman RoadMaterials
Park, OH 44073-0002
Donate Today! Traditional Personal check for file
Three Convenient Ways toContribute to the ASM Foundation.
30 HEAT TREATING PROGRESS NOVEMBER/DECEMBER 2009
