ma2014

Available online at www.sciencedirect.com

www.elsevier.com/locate/actamat

ScienceDirect

Acta Materialia xxx (2013) xxx–xxx

Mechanical behavior and strengthening mechanisms in ultrafinegrain precipitation-strengthened aluminum alloy

Kaka Ma a, Haiming Wen a,b, Tao Hu a, Troy D. Topping a, Dieter Isheim b,c,David N. Seidman b,c, Enrique J. Lavernia a, Julie M. Schoenung a,⇑

a Department of Chemical Engineering and Materials Science, University of California Davis, One Shields Avenue, Davis, CA 95616, USAb Department of Materials Science and Engineering, Northwestern University, 2220 Campus Drive, Evanston, IL 60208-3109, USAc Northwestern University Center for Atom Probe Tomography (NUCAPT), 2220 Campus Drive, Evanston, IL 60208-3109, USA

Received 23 July 2013; received in revised form 18 September 2013; accepted 23 September 2013

Abstract

To provide insight into the relationships between precipitation phenomena, grain size and mechanical behavior in a complex precip-itation-strengthened alloy system, Al 7075 alloy, a commonly used aluminum alloy, was selected as a model system in the present study.Ultrafine-grained (UFG) bulk materials were fabricated through cryomilling, degassing, hot isostatic pressing and extrusion, followed bya subsequent heat treatment. The mechanical behavior and microstructure of the materials were analyzed and compared directly to thecoarse-grained (CG) counterpart. Three-dimensional atom-probe tomography was utilized to investigate the intermetallic precipitatesand oxide dispersoids formed in the as-extruded UFG material. UFG 7075 exhibits higher strength than the CG 7075 alloy for eachequivalent condition. After a T6 temper, the yield strength (YS) and ultimate tensile strength (UTS) of UFG 7075 achieved 734 and774 MPa, respectively, which are �120 MPa higher than those of the CG equivalent. The strength of as-extruded UFG 7075 (YS:583 MPa, UTS: 631 MPa) is even higher than that of commercial 7075-T6. More importantly, the strengthening mechanisms in eachmaterial were established quantitatively for the first time for this complex precipitation-strengthened system, accounting for grain-boundary, dislocation, solid-solution, precipitation and oxide dispersoid strengthening contributions. Grain-boundary strengtheningwas the predominant mechanism in as-extruded UFG 7075, contributing a strength increment estimated to be 242 MPa, whereas Oro-wan precipitation strengthening was predominant in the as-extruded CG 7075 (�102 MPa) and in the T6-tempered materials, and wasestimated to contribute 472 and 414 MPa for CG-T6 and UFG-T6, respectively.� 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Al alloys; Precipitation; Strengthening mechanism; Ultrafine-grained materials; Atom-probe tomography

1. Introduction

Conventional coarse-grained (CG) precipitation-strengthened 7000 series aluminum (Al) alloys have beenwidely used for aerospace and transportation applicationsbecause of their high strength and heat treatability [1,2].In this class of precipitation-strengthened alloys, extremelysmall and uniformly dispersed precipitates, which act asobstacles to dislocation movement, form within the Al

1359-6454/$36.00 � 2013 Acta Materialia Inc. Published by Elsevier Ltd. All

http://dx.doi.org/10.1016/j.actamat.2013.09.042

⇑ Corresponding author. Tel.: +1 (530) 752 5840; fax: +1 (530) 752 9554.E-mail address: [email protected] (J.M. Schoenung).

Please cite this article in press as: Ma K et al. Mechanical behaviostrengthened aluminum alloy. Acta Mater (2013), http://dx.doi.org/1

matrix upon heat treatment and thus strengthen the mate-rials [3]. This phenomenon is generally referred to as pre-cipitation strengthening. Natural aging, or precipitateformation at room temperature, occurs in most 7000 seriesalloys [4,5]. It has been generally accepted that the precip-itation starts with the formation of Guinier–Preston (GP)zones, which may be regarded as coherent metastable pre-cipitates [3,6–8]. The typical diameter of GP zones is of theorder of a few nanometers [8]. Subsequent evolution of theprecipitates involves the replacement of the metastable GPzones with a metastable semicoherent phase, g0-MgZn2 [9].This occurs primarily because GP zones are isostructural

rights reserved.

r and strengthening mechanisms in ultrafine grain precipitation-0.1016/j.actamat.2013.09.042


mailto:[email protected]



2 K. Ma et al. / Acta Materialia xxx (2013) xxx–xxx

with the matrix and, therefore, have a lower interfacialenergy than intermediate or equilibrium precipitate phasesthat possess a different crystal structure. As a result, thenucleation barrier for GP zones is significantly smaller[1]. The incoherent equilibrium hexagonal phase, g-MgZn2, forms from g0-phase precipitates at higher agingtemperatures and longer aging times. These g-phase pre-cipitates are generally larger in size (diameter > 50 nm)and are preferentially located at grain boundaries. The pre-cipitation sequence can be summarized as follows: super-saturated solid solution! GP zones! g0 (MgZn2)! g(MgZn2) [3,8]. Basically, the presence of a high numberdensity of GP zones and fine g0-phase precipitates isresponsible for the strengthening of the material [6,10]. Inan effort to accelerate aging kinetics, artificial aging is per-formed at higher temperatures, during which the strengthachieves a maximum value. After long aging times orhigher aging temperatures, the strength begins to decreaseand the alloy becomes over-aged [3,11]. The commonlyused artificial aging treatment, called a T6 temper, thatresults in peak microhardness for these alloys with conven-tional coarse grains, is performed at 120 �C for 24 h [5].

Interest in nanocrystalline or nanostructured (NS; graindiameter < 100 nm) and ultrafine-grained (UFG; graindiameter > 100 nm, but less than 1000 nm) materials, orig-inally motivated by reports of novel deformation mecha-nisms as well as by the potential to attain notableenhancements in mechanical properties [12–16], has gradu-ally moved from pure metals and simple alloys to morecomplex precipitation-strengthened alloys with many alloy-ing components. Prior studies of Al 7075, a representativeprecipitation-hardenable aluminum alloy, have revealedthat these alloys can be further strengthened by incorporat-ing grain refinement. Zhao et al. [4,9] fabricated UFG 7075with a grain diameter of �400 nm utilizing commercial7075 rod (grain diameter �40 lm) through equal channelangular pressing (ECAP). They reported that the yieldstrength and the tensile strength of the UFG 7075 were650 and 720 MPa, respectively, with natural aging for amonth after ECAP, which represent strength increases of103% and 35%, respectively, over its commercial Al 7075counterpart. The improvement in the strength was ascribedto grain refinement and higher number densities of bothGP zones and dislocations in the UFG material. In arelated study, Zhao et al. also documented a simultaneousincrease in both ductility and strength for NS 7075 (aver-age grain diameter �100 nm, processed by cryorolling)with subsequent artificial aging compared with the unagedcondition [17]. In addition to the fine GP zones, g0- and g-phase precipitates were introduced in the nanograinsthrough aging, increasing the dislocation density. Theincreased dislocation density led to an improvement inthe work-hardening rate and consequently contributed tothe enhanced uniform elongation. It was concluded thatthe high dislocation density and fine grain size of the NSsample were primarily responsible for its improved strengthover the CG sample, while the high density of second-phase


precipitates was responsible for its improved ductility overas-processed NS 7075 without aging [17]. In a relatedstudy, Panigrahi and Jayaganthan [18] applied cryorollingto produce an UFG 7075 material with high-angle grainboundaries that exhibited improved strength due to theHall–Petch effect and a higher dislocation density. It wasdocumented that the microhardness and tensile strengthof the cryorolled UFG 7075 was reduced after annealingat temperatures of 150–250 �C and subsequently remainedconstant when the annealing temperature was increased[19]. More recently, a NS 7075 material exhibiting an extre-mely high yield strength of �1 GPa, combined with a uni-form elongation of �5%, was successfully produced byhigh-pressure torsion (HPT) [20]. It was suggested thatthe formation of a nanostructured architecture, whichcomprised a solid solution including a high number densityof dislocations, sub-nanometer intragranular solute clus-ters, nanometer-scale intergranular solute clusters andgrains of tens of nanometers in diameter, contributed tothe dramatic increase in strength.

Despite ample evidence that grain refinement furtherimproves the strength of precipitation-strengthened Alalloys, precise determination of the underlying strengthen-ing mechanisms has been hindered by the complexity of thepossible mechanisms, including: grain-boundary strength-ening (Hall–Petch effect), solid-solution strengthening, dis-location strengthening and precipitation strengthening.Accordingly, the goal of the present study is to formulatea quantitative insight into strengthening mechanisms byproviding a direct comparison of an UFG precipitation-strengthened material with an otherwise equivalent powdermetallurgy (PM)-derived CG material. It is noted thatunlike CG materials made by casting, the PM-derivedCG material is expected to exhibit a relatively fine graindiameter, in the range of 1–5 lm [21]. To the best of ourknowledge, this is the first time a direct comparisonbetween UFG and CG materials, consolidated and heat-treated using identical processing steps, has been docu-mented for such a complex precipitation-strengthened Alalloy. More importantly, fundamental insights into theinterrelationships between grain refinement, precipitationcharacteristics and mechanical behavior are provided.

2. Experimental procedure

Al 7075 was chosen as a representative alloy for thisinvestigation, partly because the precipitation sequenceand kinetics in this system have been extensively studied.Cryomilling, a mechanical attrition technique in a cryo-genic environment [14,15,22–24], was utilized in our studyto obtain NS 7075 powder. This technique takes advantageof the low boiling temperature, 77.2 K, of liquid nitrogen,which suppresses recovery and recrystallization in the pow-der, and leads to nanocrystalline grain structures and rapidgrain refinement. Cryomilling also results in a high numberdensity of dislocations in the material through severe plas-tic deformation (SPD) [14,15]. Additionally, fine oxide/



Table 2Parameters used in cryomilling.

Cryomilling parametersMilling media Slurry of stainless steel balls and LN2

Ball-to-powder weightratio

32:1

Impellor rotation speed 180 rpmMilling time 12 hMilling temperature ��183 to �190 �CProcess control agent

(PCA)Stearic acid �2 g (0.2 wt.% of the loadedpowder)

K. Ma et al. / Acta Materialia xxx (2013) xxx–xxx 3

nitride dispersoids are introduced into the material duringcryomilling [25,26]. UFG 7075 bulk materials were subse-quently fabricated by consolidating the nanocrystallinepowder through degassing, hot isostatic pressing, andextrusion [27]. Equivalent CG 7075 bulk materials werefabricated employing the same techniques by consolidatinggas atomized powder. The microstructures of the sampleswere characterized by transmission electron microscopy(TEM) and X-ray diffraction (XRD). Atom-probe tomog-raphy (APT) [28–32] was utilized to characterize the inter-metallic precipitates and oxide dispersoids in the UFG7075 on an atomic scale; this advanced technique deter-mines the elemental identity and position of individualatoms with sub-nanometer resolution in three dimensions(3-D), and therefore enables the quantification of the size,morphology, composition, number density and volumefraction of the nanoscale precipitates and dispersoids [28–30,33–35]. The mechanical behavior of the materials wasinvestigated through standard tensile testing. Details areprovided below:

2.1. Material processing

2.1.1. Powder modification

The starting material used in the present study is gas-atomized Al 7075 powder provided by Valimet, Inc.(Stockton, CA) with a particle diameter of 200 mesh(�74 lm). The nominal chemical composition of the pow-der is provided in Table 1 [36]. The feedstock powder wascryomilled in liquid nitrogen (LN2) for 12 h in a modifiedSzegvari attritor. The cryomilling parameters are summa-rized in Table 2. The cryomilled powder was subsequentlyfilled into a 1 inch diameter by 3 inch long can fabricatedfrom Al 6061 [36] and hot vacuum degassed at tempera-tures up to 500 �C for 12 h with a final pressure in the range10�6 Torr [37]. Degassing is necessary to remove hydratesand stearates attributable to powder handling, and the pro-cess control agent (PCA), stearic acid, utilized during cryo-milling [37].

2.1.2. Consolidation

After cryomilling and hot vacuum degassing, the canwas crimped, welded and hot isostatically pressed (HIPed)at 400 �C and 172 MPa (25 ksi) in argon to achieve fullconsolidation. The Al 6061 can material was removed bymachining and the consolidated Al 7075 was machined intoa workable billet, which subsequently underwent a slow-strain-rate (SSR) extrusion at 350 �C using a 3.56 MN(400 ton) Hypress Technologies Inc. model UGP400-6press equipped with a custom-built resistive band heatingsystem. Omega temperature controllers using K-type ther-

Table 1Nominal chemical composition of Al 7075 alloys [36].

Elements Al Zn Mg Cu C

wt.% Balance 5.1–6.1 2.1–2.9 1.2–2 0


mocouples measured and controlled the temperature of thedie. The extrusion die is a cylinder of �108 mm(4.25 inches) diameter fabricated from a H13 tool steel witha 25.4 mm (1 inch) hole bored through its center. Thereduction ratio was �10:1 in area ratio, which gave a finaldiameter of �7.9 mm. The strain rate, �3.7 � 10�2 s�1,was based on a nominal velocity of 100 mm min�1. Forthe CG 7075 counterpart, as-received gas-atomized Al7075 powder was degassed and consolidated followingthe same processing steps as used for the cryomilledpowder.

2.1.3. Heat treatmentSolution heat treatment was performed on some of the

tensile specimens, which were wrapped in Al foil, in an elec-tric resistance furnace (model C601K, Cress Mfg Co., ElMonte, CA) at 500 �C for 1 h. Subsequently, the specimenswere quenched in ice water. The artificial aging treatmentwas performed at 120 �C in a tube furnace (Carbolite,UK) for 24 h. The schedule for the artificial heat treatmentwas chosen following the T6 temper for conventional Al7075 materials [5]. Natural aging occurred after solution-ized samples were stored at room temperature for a week.A summary of the sample identification (ID) and the corre-sponding process/heat-treatment conditions is provided inTable 3. A pictorial description of the thermal history thematerials experienced is provided in Fig. 1 [38].

2.2. Characterization

2.2.1. Room temperature tensile testing

Cylindrical, threaded tensile specimens were machinedwith a gauge length of �12 mm (0.5 inch) � 3 mm(0.12 inch) diameter, with dimensions close to the specifica-tions for sub-size ASTM E 8M standard. The tensile testswere performed at room temperature using a universal test-ing machine (Instron 8801, Norwood, MA) with strainmeasured by a standard video extensometer. The strainrate utilized was 10�3 s�1. Two samples were tested foreach condition to confirm reproducibility. The solution

r Fe Mn Si Ti Other

.18–0.28 60.5 60.3 60.4 60.2 60.15



Table 3Sample identification (ID) and the corresponding processing conditions.

Sample ID Processing conditions

NC7075 As-cryomilled 7075 powderUFG7075-E Consolidated 7075 made from NC7075, which went

through degassing plus HIP plus extrusionUFG7075-E-sol UFG7075-E was solution heat treated and quenched

in ice waterUFG7075-E-Nag UFG7075-E-sol was naturally aged at ambient

temperature in air for 1 weekUFG7075-E-T6 UFG7075-E-sol was artificially aged at 120 �C in air

for 24 hCG7075-E Consolidated 7075 made from as-received gas

atomized (i.e. coarse grained, CG) powder, whichwent through degassing plus HIP plus extrusion

CG7075-E-sol CG7075-E was solution heat treated and quenched inice water

CG7075-E-Nag CG7075-E-sol was naturally aged at ambienttemperature in air for 1 week

CG7075-E-T6 CG7075-E-sol was artificially aged at 120 �C in air for24 h

C7075-T6 Commercial extruded 7075 bar in T6 temper


heat-treated samples and T6-tempered samples were alltested within 1 h after the heat treatment to exclude anynatural aging effect.

2.2.2. Phase identification and microstructure

XRD analyses were performed on both the powder andconsolidated bulk samples to investigate the phase consti-tution after each processing step, using a Scintag X-ray dif-fractometer equipped with a graphite monochromatorusing Cu Ka (k = 0.15406 nm) radiation. Additionally,quantitative analyses of grain diameter and microstrainwere performed on the NC7075 powder and UFG7075-Econsolidated sample according to the Williamson–Hallmethod [39]. As a measure of peak broadening, the full-width at half-maximum (FWHM) of the peaks wasobtained by fitting the XRD peak profiles using the Pear-son 7 function. The true peak broadening, B, was derivedfrom:

B ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiB2

obs � B2inst

q; ð1Þ

where Bobs is the observed peak broadening and Binst is theinstrumental broadening.

Fig. 1. Temperature vs. processing time for Al 7075 alloy from powder


The microstructures of the powder and consolidatedsamples were characterized employing a Phillips CM12transmission electron microscope and a JEOL 2500 high-resolution transmission electron microscope, operating at120 and 200 kV, respectively. The grain diameter was esti-mated by measuring and averaging the length and width ofthe strongly diffracting grains using ImageJ� image analy-sis software [40]. Truncation effects caused by the surface ofthe TEM foil were neglected. APT studies of the as-extruded UFG 7075 were performed using a Camecalocal-electrode atom-probe (LEAP) 4000X-Si tomograph[28–31]. Parallelepipeds of 400 lm � 400 lm � 8 mm werecut and subsequently electropolished at ambient tempera-ture to obtain a needle-shaped tip with a radius of curva-ture of �50 nm at the apex. APT was performed usingUV picosecond laser pulsing (355 nm wavelength), a pulserepetition rate of 500 kHz, a pulse energy of 200 pJ and adetection rate of 0.2–0.5%. The specimen base temperaturewas 60 K, and steady-state DC voltages between 2.0 and6.0 kV were applied for controlled field-evaporation inthe voltage pulsing mode. 3-D reconstructions and dataanalyses of the APT data were performed using Cameca’sIVAS� software, version 3.6.1. Compositions of precipi-tates and dispersoids were obtained employing the proxim-ity histogram concentration profile technique [41]. Moredetails regarding APT can be found in Refs. [25,28–31].

3. Results

3.1. Tensile behavior

The representative engineering tensile stress–straincurves for CG and UFG materials for different heat treat-ment conditions are presented in Fig. 2a and b, respec-tively. As a reference, a commercial extruded 7075 bar,T6 temper, was tested in the same condition as the othersamples; the result is presented as curve 5. The mechanicalproperty data are summarized in Table 4. It is clear that theUFG 7075 materials exhibit higher strengths than the CG7075 materials for each equivalent condition. In theextruded condition, the tensile yield strength, rys, and ulti-mate strength, ruts, of sample UFG7075-E are 583 and631 MPa, respectively, which are approximately 106%and 45% greater, respectively, than those of sampleCG7075-E (283 and 436 MPa). For the T6 temper, the

to consolidated UFG bulk samples and subsequent T6 temper [38].



Fig. 2. Comparison of tensile stress–strain curves for: (a) CG 7075materials and (b) UFG 7075 materials, after selected heat-treatmentconditions. Curve 5 is for a reference specimen of commercially extruded7075 bar in T6 temper, which was tested under the same conditions as theother samples. Symbols are added every 50 data points to differentiate thecurves from one another.


rys and ruts values of sample UFG7075-E-T6 increased to734 and 774 MPa, respectively, which are 20% and 17%,respectively, greater than those of sample CG7075-E-T6(613 and 659 MPa). Moreover, the rys and ruts values ofsample UFG7075-E-T6 are 36% and 31% greater, respec-tively, than those of the commercial Al 7075-T6 alloy(541 and 593 MPa). Note that the strength of the as-extruded UFG7075-E sample is also greater than that ofthe commercial Al 7075-T6 alloy and is close in value tothat of sample CG7075-E-T6. These results indicate thatthe cryomilling process might eliminate the need for thesolutionizing and aging treatments to achieve the desirablestrength, with acceptable ductility (elongation of 2.8%).Additionally, a decrease in stress, i.e. strain softening,occurred after yielding in all the curves for the UFG7075 materials, while this phenomenon was absent for theCG 7075 materials. Strain softening has been observed inother Al alloys, e.g. Al 5083, that were cryomilled and con-


solidated in a similar manner [24,27,42]. This phenomenonis, however, beyond the scope of the current study.

Interestingly, after the solutionizing treatment, the CG7075 material exhibited a decrease in rys from 283 MPa(CG7075-E, curve 1) to 236 MPa (CG7075-E-sol, curve2) with almost the same ductility; while sampleUFG7075-E-sol (curve 7) maintained both the same rys

and ruts as those of sample UFG7075-E (curve 6), butexhibited an increase in elongation from 1.9% to 5.3%.Additionally, both natural aging and artificial aging con-tributed more significantly to increasing the strength inthe CG 7075 materials than in the UFG 7075 materials.The increases in rys for samples CG7075-E-Nag andCG7075-E-T6 are 70% and 160%, respectively, comparedwith the baseline values for sample CG7075-E-sol. In con-trast, the increases in rys after natural aging and T6 temperof the UFG 7075 materials are only 13% and 26%,respectively.

3.2. Microstructure

3.2.1. Grain diameterFig. 3 displays a representative bright-field TEM image

of the as-cryomilled Al 7075 powder (NC7075) and a histo-gram of the grain diameter distribution that ranges from 20to 90 nm with a mean value of 46 nm, which demonstratesthat nanocrystalline Al 7075 powder was successfully fabri-cated during the cryomilling process. No contrast relatedto precipitates could be discerned in the TEM micrographsand we concluded that the alloy is largely precipitation-freeafter the cryomilling process. Due to the thermal exposureduring degassing, HIPing and extrusion, significant graingrowth occurred in the materials, as observed in the micro-graphs in Fig. 4. The length of the elongated grains in theUFG7075-E sample ranged from 95 to 730 nm along theextrusion direction (Fig. 4a) [38]. Averaging the meanlength (315 nm) and the mean width (175 nm) of the elon-gated grains yields an estimate for the mean grain diameterof 245 nm [38]. In contrast, sample CG7075-E exhibitsmuch coarser equiaxed grains ranging from 0.5 to 1 lm,with an average grain diameter of 894 nm (Fig. 4b). Afterthe T6 temper, the average grain diameter of the UFG7075 increased to 422 nm (Fig. 4c), while that of the CG7075 increased to about 1 lm (Fig. 4d). Additional TEMimages and grain diameter histograms appear elsewhere[38].

3.2.2. Precipitates

A detailed and robust investigation of the precipitates interms of type, morphology, size and density, as well as for-mation mechanisms, in both the CG and UFG 7075 mate-rials, in both the as-fabricated condition and after a T6temper, can be found elsewhere [38]. The morphology, sizeand number density of the three families of precipitates—GP zones, and g0- and g-phase precipitates—wereobserved to vary significantly between the CG and UFGsamples. The critical relevant results for explaining the



Table 4Mechanical data for the consolidated Al 7075 materials.

Sample ID 0.2%YS rys (MPa) UTS ruts (MPa) Strain at fracture (%) Elongation (%)

CG7075-E 283 436 11.9 11.4UFG7075-E 583 631 3.0 1.9CG7075-E-sol 236 441 11.9 11.3UFG7075-E-sol 584 635 6.4 5.3CG7075-E-Nag 402 564 13.6 11.5UFG7075-E-Nag 658 686 3.5 2.5CG7075-E-T6 613 659 5.0 3.9UFG7075-E-T6 734 774 4.0 2.8C7075-T6 541 593 14.1 13.2

Fig. 3. Representative TEM image of as-cryomilled Al 7075 powder(NC7075) with an embedded grain size histogram.


mechanical behavior are summarized in Table 5. Represen-tative TEM bright-field images of the consolidated CG andUFG 7075 materials are displayed in Fig. 4. AdditionalTEM images, found in a prior study [38], support thedescription of the precipitates provided herein. SampleUFG7075-E possessed a relatively low number density ofGP zones and g0-phase precipitates, both with averagediameters <5 nm, as observed within the grain interiors[38]. Irregularly shaped g-phase MgZn2 precipitates, whichcontain some Cu and Al in solution with the Zn, were occa-sionally observed, with an average diameter of 90 nm [38].In contrast, sample CG7075-E (Fig. 4b) exhibited a num-ber of plate-like g0-phase (diameter �55 nm) and lath-likeg-phase precipitates (length �174 nm) in grain interiors;these precipitates were located along dislocation lines[38]. After the T6 temper, the grains in the UFG 7075material (Fig. 4e) contained a large number of nanoscalespherical GP zones (average diameter �3 nm) and plate-let-like g0-phase precipitates (diameter 6 5 nm); in con-trast, although sample CG7075-E-T6 (Fig. 4f) alsoexhibited many spherical GP zones, with diameter <5 nm,


the g0-phase precipitates were coarser with an averagediameter of 61 nm [38].

To incorporate this descriptive information concerningthe precipitates into the strengthening mechanism estimates(Section 4.4) it was also necessary to estimate the meanradii and mean edge-to-edge interprecipitate spacing, kp.The mean radii, r, in Table 5, were estimated by averagingthe radii of all the precipitates regardless of precipitatemorphology. The values for kp, in Table 5, were estimatedfrom 60 measurements on several representative TEMimages.

3.2.3. X-ray diffraction

The XRD patterns for the UFG and CG 7075 materialsare displayed in Fig. 5a and b, respectively. The g0- and g-phases were evaluated together due to the overlapping oftheir XRD peaks. The XRD peaks for the g0- and g-phasesare present in the as-received gas-atomized 7075 powder,but disappear after cryomilling, indicating that cryomillingdissolved the precipitates and solutionized the powder towithin the limits of detection. The g0- and g-phases weredetected, however, in the consolidated samples and theintensity of these XRD peaks are stronger for sampleCG7075-E than for sample UFG7075-E, which is consis-tent with our TEM observations. During the solutionizingstep, the g-phase precipitates were dissolved in samplesCG7075-E-T6 and UFG7075-E-T6. The amount of theg0-phase was reduced, while the number density of GPzones increased after T6 temper in sample CG7075-E-T6.Therefore, the XRD peaks for the g0- and g-phases becameweaker for sample CG7075-E-T6 compared to those forsample CG7075-E. Similarly, the dissolution of theg-phase precipitates and the extremely small size of theg0-phase precipitates resulted in a diminution of theirXRD peaks in sample UFG7075-E-T6 compared to sampleUFG7075-E.

The grain diameter, d, and microstrain, e, of the bulksamples were calculated from the XRD peak broadening,B, using the Williamson–Hall method [39]. This approachassumes that B consists of grain refinement broadeningand strain broadening, which is given by:

B cos hB ¼Kkdþ e sin hB; ð2Þ



Fig. 4. Representative TEM images of samples: (a) UFG7075-E; (b) CG7075-E; (c) UFG7075-E-T6; (d) CG7075-E-T6; (e) a higher-magnification imageof UFG7075-E-T6; and (f) a higher-magnification image of CG7075-E-T6.

Table 5Summary of the type, average diameter, mean radius and mean edge-to-edge interprecipitate spacing of the precipitates present in select samples.

Sample ID Precipitate type[38]

Average diameter 2r (nm)[38]

Mean radius of all the precipitates r

(nm)Mean edge-to-edge interprecipitate spacingkp (nm)

UFG7075-E GP zone 2.5 1.5 173g0 3.3

UFG7075-E -T6

GP zone 3.1 1.5 19g0 4.4 (length)

1.5 (width)CG7075-E g0 55 40 194

g 174 (length)35 (width)

CG7075-E-T6 GP zone 2.9 16 35g0 61


Please cite this article in press as: Ma K et al. Mechanical behavior and strengthening mechanisms in ultrafine grain precipitation-strengthened aluminum alloy. Acta Mater (2013), http://dx.doi.org/10.1016/j.actamat.2013.09.042


Fig. 5. X-ray diffraction patterns for (a) UFG 7075 and (b) CG 7075materials.

Fig. 6. Atom-probe tomographic 3-D reconstructions of sampleUFG7075-E: (a) reconstruction with only Zn atoms displayed and asuperimposed 8 at.% Zn isoconcentration surface; (b) reconstruction withonly Mg atoms displayed and a superimposed 8 at.% Mg isoconcentrationsurface superimposed; (c) reconstruction with only O atoms displayed anda superimposed 5 at.% O isoconcentration surface; and (d) reconstructionwith only N atoms displayed and a superimposed 5 at.% Nisoconcentration.


where k is the wavelength of Cu Ka radiation, i.e. 1.54 A, K

is �0.9, e is the microstrain and hB is the Bragg angle [25].Plotting B cos hB vs. sin hB and performing a linear regres-sion analysis, the values of d and e were obtained from theslope and intercept of the fitted curve. For materials sub-jected to SPD, the dislocation density, q, in terms of dand e is given by [4]:

q ¼ 2ffiffiffi3p

edb

; ð3Þ

where b = 0.286 nm is the magnitude of the Burgers vectorfor Al [43]. Applying the values of d and e obtained fromEq. (2), the values of q for samples UFG7075-E,UFG7075-E-T6, CG7075-E and CG7075-E-T6 are4.5 � 1014, 4.1 � 1014, 1.7 � 1014 and 5.6 � 1013 m�2,respectively. These data are used to calculate the contribu-tion to strengthening from dislocations.

3.2.4. Atom-probe tomography

An APT tomographic 3-D reconstruction of a volume inthe UFG7075-E sample is presented in Fig. 6a–d. Imagesof each individual element (Zn, Mg, O or N) are displayed


in a separate box, with 8 at.% Zn, 8 at.% Mg, 5 at.% O and5 at.% N isoconcentration surfaces superimposed, respec-tively; this volume contains 23 million atoms. The precipi-tates are clearly delineated by the isoconcentrationsurfaces. The inhomogeneous distributions of the Zn,Mg, O and N atoms indicate the formation of precipitatesin the material. The locations of Mg-enriched regions coin-cide with Zn- or O-enriched region, indicating the coprecip-itation of Mg and Zn atoms, and the presence of MgOdipsersoids. A slight enrichment of N is observed whereO is enriched, suggesting that some MgO dispersoids con-tain a small concentration of N. The presence of N is adirect result of cryomilling, which occurs in liquid nitrogen[14,22,44]. Proximity histogram concentration profiles [41]were employed to quantify the specific chemical composi-tions of the precipitates and dispersoids. Results from anumber of proximity histograms reveal that three distincttypes of precipitates and dispersoids exist in sampleUFG7075-E: MgZn GP zones, MgZn2 g0-phase precipi-tates and MgO dispersoids. Fig. 7a displays isoconcentra-tion surfaces of 8 at.% Zn with a representative exampleof a GP zone indicated by an arrow; the corresponding rep-




resentative proximity histogram is in Fig. 7b. Examples ofg0-phase precipitates and MgO dispersoids and the corre-sponding proximity histograms are displayed in Fig. 7c–f,respectively. Although some of the GP zones and g0-phaseprecipitates contain up to �5 at.% Cu, and some of theMgO dispersoids contain up to 10 at.% N, these phasesare referred to below without reference to their complexchemistries. Quantitative information on precipitate diam-eter distribution, number density and volume fraction werealso obtained following the methodology detailed in Refs.[25,35]. The precipitate diameter distributions for the GPzones, g0-phase precipitates and MgO dispersoids are pre-sented in Fig. 8a and b, respectively. The standard devia-tion was used as the measure of uncertainty in theprecipitate diameters. The statistical precipitate and disper-soid characteristics are presented in Table 6. The averagediameter of GP zones/g0-phase precipitates is 3.5 nm,which agrees with the TEM observations. The volume frac-tion of these precipitates is about 0.07%, and their numberdensity is 1.8 � 1022 m�3. Although TEM was unable toreveal the presence of MgO dispersoids, the APT results

Fig. 7. (a) 8 at.% Zn isoconcentration surface for sample UFG7075-E, where tanalysis is highlighted and arrowed; (b) proximity histogram concentration parrowed in (a); (c) 8 at.% Zn isoconcentration surface, where the isoconcentrahighlighted and arrowed; (d) proximity histogram concentration profile based o(e) 5 at.% O isoconcentration surface, where the isoconcentration surface of a prand (f) proximity histogram concentration profile based on the 5 at.% O isoco


indicate an average diameter of 4 nm, with a volume frac-tion of �0.14% at a number density of 1.3 � 1022 m�3.

4. Discussion

Our recent investigation [38] revealed that grain sizeinfluences precipitation kinetics, resulting in significant dif-ferences in the size, composition and spatial distribution ofprecipitates between the CG and UFG 7075 materials. GPzones, as well as plate-like g0-phase precipitates, wereobserved to nucleate homogeneously in the grain interiorof sample UFG7075-E; in contrast, large numbers ofg0-phase precipitates formed either on dislocation lines orin the vicinity of the dislocations in sample CG7075-E.During artificial aging, a high number density of GP zoneswith an average diameter of �3 nm and platelet-shapedg0-phase precipitates formed via homogeneous nucleationand growth in the interior of the grains in the UFG mate-rial. Alternatively, the presence of dislocations in the grainsof the CG material assisted the heterogeneous nucleationand growth of plate-like g0-phase precipitates, whereas

he isoconcentration surface of a precipitate used for a proximity histogramrofile based on the 8 at.% Zn isoconcentration surface of the precipitatetion surface of a precipitate used for the proximity histogram analysis isn the 8 at.% Zn isoconcentration surface of the precipitate arrowed in (c);ecipitate used for proximity histogram analysis is highlighted and arrowed;ncentration surface of the precipitate arrowed in (e).



Fig. 8. Precipitate diameter distributions for: (a) GP zones/g0-phase and(b) MgO. The uncertainty in the average precipitate diameter correspondsto the standard deviation.

Table 6Summary of precipitate characteristics in sample UFG-7075-E from atom-probe tomography (APT).

Precipitates Average diameter(nm)

Volumefraction (%)

Number density(m�3)

GP zone/g0-MgZn2

3.5 ± 1.6 0.07 1.8 ± 0.5 � 1022

MgO 4.0 ± 3.1 0.14 1.3 ± 0.4 � 1022


GP zones formed homogeneously at a high number densityin regions without dislocations in sample CG7075-E-T6[38]. It was proposed that a reduction in grain size, withan increase in grain boundary area per unit volume,resulted in a small concentration of vacancies in theUFG 7075, and thus homogeneous nucleation of the pre-cipitates was inhibited during aging [38]. The dislocationsubstructure, which provided heterogeneous nucleationsites for g0-phase precipitates, governed the precipitationkinetics in CG 7075 [38]. The modification of the feedstockpowder by cryomilling reduced the length scale of the grainsize and subsequently changed the morphology, size and


nucleation mechanism of the precipitates in the final con-solidated UFG 7075 material, as well as introducing nano-scale oxide dispersoids. The current investigationdemonstrates that these microstructural changes conse-quently lead to an increase in tensile strength in the UFG7075 when compared with CG 7075. To provide insightinto the measured differences in the stress–strain behaviorbetween these materials, it is important to establish theactive strengthening mechanisms. The CG and UFG7075 materials have experienced several thermomechanicalprocessing (TMP) steps during the consolidation proce-dure, in which significant plastic deformation was alsointroduced. Considering the nature of this class of precipi-tation-strengthened Al alloys, the influence of cryomillingand the plastic deformation that occurred during TMP,the following strengthening mechanisms need to be evalu-ated to explain the higher strength in the UFG 7075 mate-rials: grain-boundary strengthening, solid-solutionstrengthening, dislocation strengthening and precipitate/dispersoid strengthening. Quantitative estimates of the con-tribution of each mechanism in the UFG and CG 7075materials are described below; the strength incrementsstemming from the four mechanisms are summarized inTable 7. The base strength of pure Al, the lattice frictionstress, r0, is not included in this table. Although there arevarious studies in the literature in which linear summationof strength increments have been applied [25,35,45], thiswas not the purpose of the current investigation. Rather,our goal was to highlight the differences in mechanismsthat dominate strengthening behavior between the UFGand CG materials. The physical meaning and values ofthe symbols in the equations used in the current studyare summarized in Table 8 [25,36,43,46–50].

4.1. Grain-boundary strengthening (Hall–Petch effect)

One of the most significant consequences of cryomillingis grain size refinement, creating a high volumetric densityof grain boundaries that impede dislocation movement anddislocation propagation to adjacent grains, therebystrengthening the materials [14,15,51]. The grain-boundarystrengthening mechanism is usually described by the Hall–Petch equation [47,52–54]:

ry ¼ r0 þkyffiffiffi

dp ; ð4Þ

where d is the average grain diameter, r0 is the frictionstress and ky is the Hall–Petch slope.

Several publications have reported a breakdown or devi-ation in the Hall–Petch relationship in UFG materials fab-ricated by SPD processes such as accumulative rollbonding (ARB) and HPT [55–57]. Two explanations havebeen proposed for this breakdown: (i) the effective grainsize for the Hall–Petch relationship is larger than the mea-sured value because the mobile lattice dislocations pass eas-ily through the non-equilibrium grain boundaries that weregenerated in the material during the SPD processing [55];



Table 7Estimated strength increment for different strengthening mechanisms.

UFG7075-E

UFG7075-E-T6

CG7075-E

CG7075-E-T6

Drgrain boundary

(MPa)242 185 127 120

Drsolid solution

(MPa)<82 �82 <82 �82

Drdislocation (MPa) 99 95 61 35Drorowan (MPa) 45 414 102 472

Table 8Physical meaning and values of different symbols used in the strengtheningmechanism calculations [25,36,43,46–50].

Symbol Meaning Values Unit

a Lattice constant =0.405 for fcc Al nmb Magnitude of the

Burgers vector=p

2/2a = 0.286 for fccmetals

nm

ky Hall–Petchcoefficient

=0.12 MPa=ffiffiffiffimp

M Mean orientationfactor

=3.06 for the fccpolycrystalline matrix

Dimensionless

G Shear modulus =26.9 for Al 7075 GPaa Constant =0.2 for fcc metals Dimensionlesst Poisson ratio =0.33 for Al 7075 Dimensionlessae Constant =2.6 for fcc metals Dimensionless


(ii) increased participation of extrinsic dislocations are ableto move in the non-equilibrium grain boundaries andthereby reduce the apparent strength [57]. Non-equilibriumgrain boundaries are responsible for the deviation from aHall–Petch behavior. In our work, the non-equilibriumgrain boundaries introduced during cryomilling are effec-tively eliminated due to the annealing that occurs duringthe subsequent degassing step and thermomechanical con-solidation steps, which involve prolonged thermal exposureat elevated temperatures (see Fig. 1). This is supported bypublished studies [14,15,58], which fail to reveal non-equi-librium grain boundaries in the microstructure of consoli-dated cryomilled materials. Also, several publishedstudies have applied Dr = kd�1/2 to estimate grain-bound-ary strengthening for consolidated cryomilled materialswith grain size ranges similar to those in the current work[45,58,59]. Therefore, it is reasonable to assume that aHall–Petch relationship with a power of d�1/2 is applicableto the UFG Al 7075 in the current study. Wert et al. stud-ied the effect of grain size on the yield strength of 7000 ser-ies Al alloys and revealed that the Hall–Petch coefficient,ky, for peak-aged Al 7075 was �0.12 MPa=

ffiffiffiffimp

[48,49].Assuming the values of r0 are the same for CG andUFG 7075, the strength increase from grain-boundarystrengthening, Drgb, is proportional to d�1/2. For UFG7075, the grain diameter was obtained by averaging thelength and the width of the elongated grains. Therefore,the increase in yield strength due to grain-boundarystrengthening is calculated to be �242, �185, �127 and


�120 MPa for samples UFG7075-E, UFG7075-E-T6,CG7075-E and CG7075-E-T6, respectively. It is evidentthat the strength increment from grain-boundary strength-ening in sample UFG7075-E is almost twice that in sampleCG7075-E, while the strength increment from grain-boundary strengthening in sample UFG7075-E-T6 is 54%higher than in sample CG7075-E-T6.

4.2. Solid-solution strengthening

Solid-solution strengthening occurs when other elementsare alloyed with a metal matrix as solute atoms that differfrom the matrix atoms in size and/or shear modulus, whichcan cause a variation of strain fields. Local strain fields arecreated that interact with dislocations and impede theirmotion, leading to an increase in the yield strength of thematerial. It has been generally accepted that solid-solutionstrengthening is governed by the Fleischer equation [60,61]:

Drss ¼ MGbe32ss

ffiffifficp: ð5Þ

The meaning and values of the symbols in this equationare listed in Table 8. Some studies have demonstrated a needto modify the power of c from ½ to 1 for nanostructuredmaterials with grain diameters <30 nm [62]. Nevertheless,this is not necessary for the present system because the grainsize in the UFG7075 is of the order 100 nm and larger. TheFleischer equation has been widely used in the literaturefor UFG materials with grain diameters in this range[25,35,45]. Therefore, it is reasonable to assume that Eq.(5) is applicable in the current investigation. From Eq. (5),the value of Drss depends on the difference in the shearmoduli between the solute and the matrix, the concentration,c, and the difference in size between the solute and solventatoms (causing lattice strain, e). Table 1 demonstrates thatAl 7075 primarily contains Zn, Mg and Cu solute atoms.The difference in radii and the theoretical contributions tothe yield strength from these elements are listed in Table 9,with data for high-purity binary solid-solution alloys[50,63]. Mg, Zn and Cu are not all in solid solution in theextruded condition nor in the T6-tempered conditionbecause they form second-phase precipitates or segregateto the grain boundaries [38]. As an upper bound, assumingthat all of the solute atoms are in solid solution and the effectsfrom the different solute atoms are additive, solid-solutionstrengthening in this alloy accounts for a strength increaseof �82 MPa. Because the actual contribution from solid-solution strengthening is <82 MPa in the T6-tempered sam-ples, where a significant fraction of the solute atoms haveprecipitated, we conclude that solid-solution strengtheningprovides a small contribution to the total strength of theas-extruded and T6-tempered materials.

4.3. Dislocation strengthening

Dislocations interact with themselves and impede theirown motion. Thus, increasing the dislocation density in a



Table 9Data on the primary solute atoms in Al 7075 and their contribution toyield strength.

Element Difference inatomicradii [50,63],(rx � rAl)/rAl

(%)

Yield strengthaddition[50,63](MPa wt.%�1)

Concentration(wt.%)

Contributionto yieldstrength(MPa)

Zn �6 2.9 5.4 16Mg 11.8 18.6 2.4 44Cu �10.7 13.8 1.6 22


metal increases the yield strength of the material [47]. Toevaluate and compare the role of the residual dislocationsto strengthening in the CG and UFG 7075 materials, theBailey–Hirsch relationship was applied in the current study[47,64]:

Drd ¼ MaGbq12: ð6Þ

The meaning and values of the symbols in Eq. (6) aregiven in Table 8. The applicability of the Bailey–Hirschrelationship to UFG materials is supported by previouswork by other researchers [25,35,45,65]. However, it isnoteworthy that Huang et al. [66] reported an interestingphenomenon: hardening by annealing and softening bydeforming for a nanostructured Al prepared by ARB,which is in contrast to the typical behavior of a metal. Theyproposed that many dislocation sinks available in the formof closely spaced high-angle boundaries reduce the numberof dislocation sources during annealing. Consequently, anincrease in the yield stress is expected during straining inorder to activate new dislocation sources. In the presentwork, however, we aimed to estimate the contribution fromthe residual dislocations to strengthening during tensiledeformation. The dislocation density values were deter-mined by XRD (Section 3.2.3). The strength incrementcaused by dislocation sources is calculated to be �99 and�95 MPa in samples UFG7075-E and UFG7075-E-T6,respectively. In contrast, dislocation strengthening contrib-uted an increase of �61 and �35 MPa for samplesCG7075-E and CG7075-E-T6, respectively. An additionalstrengthening mechanism related to dislocations, the so-called dislocation source-limited hardening, may also oper-ate for UFG, but not CG, materials, as a higher stress isrequired to activate alternative dislocation sources[56,66]. Combined, these results suggest that dislocationstrengthening plays a more significant role in the UFG7075 materials than in the CG 7075 materials.

4.4. Precipitate and dispersoid strengthening

Precipitates (GP zones, g0- and g-phases) are present inboth CG and UFG 7075 materials, in both the as-extrudedand the T6-tempered conditions. Nitrogen- or oxygen-richdispersoids are only present in the UFG 7075 materialsbecause they are a by-product of cryomilling [14,15,22].


Precipitation and dispersoid strengthening are governedby either the Orowan dislocation bypassing or dislocationshearing mechanisms. The one causing a smaller strengthincrement is the operative mechanism [25,47]. When pre-cipitates or dispersoids are bypassed by the Orowan dislo-cation bypassing mechanism, the yield strength increment,Drorowan, is [46,47,67]:

Drorowan ¼ M0:4Gb

pffiffiffiffiffiffiffiffiffiffiffi1� tp lnð2�r=bÞ

kp; ð7Þ

where M, G, b, t are defined in Table 8; �r is the mean radiusof a circular cross-section in a random plane for a sphericalprecipitate, �r ¼

ffiffiffiffiffiffiffiffi2=3

pr, where r is the mean radius of the

precipitates. In the shearing mechanism, three factors con-tribute to the increase in yield strength: coherency strength-ening (Drcs), modulus mismatch strengthening (Drms) andorder strengthening (Dros). The larger of Drcs + Drms orDros is the total strength increment from the dislocationshearing mechanism [25,46,47]. The values of Drcs, Drms

and Dros are calculated from Eqs. (8)–(10), respectively[25,43,68,69]:

Drcs ¼ MaeðGecÞ32

rf0:5Gb

� �12

; ð8Þ

Drms ¼ M0:0055ðDGÞ32

2fG

� �12 r

b

� �3m2 �1

; ð9Þ

Dros ¼ M0:81capb

2b3pf

8

� �12

; ð10Þ

where M, G, b and r are listed in Table 8; ae = 2.6 for face-centered cubic (fcc) metals; m = 0.85; DG is the modulusmismatch between the matrix and the precipitates; ec isthe constrained lattice parameter misfit; f is the volumefraction of the precipitates; and capb is the antiphaseboundary free energy of the precipitate phase.

To estimate the strength increment from the precipitatesand dispersoids in the CG and UFG 7075 materials, theoperative mechanism for each type of precipitate and dis-persoid must first be identified. For precipitates, e.g. theg-phase precipitates that are incoherent with the Al-matrix, the operative mechanism is Orowan dislocationbypassing [1,6,70]. If the precipitate is coherent or semico-herent with the matrix, e.g. GP zones and the g0-phase pre-cipitates, the strength increment resulting from dislocationshearing needs to be evaluated and compared with Drorowan

to determine the operative mechanism [25,47]. From Eq.(7), it is evident that Drorowan is only dependent on r andkp, and is independent of the intrinsic properties of the pre-cipitates or dispersoids, e.g. chemical composition andcrystal structure [47]. The strength increment from disloca-tion shearing (Eqs. (8)–(10)) is, however, dependent on theintrinsic material properties, which derive from chemicalcomposition and crystal structure [46,47,67]. Due to thechallenges of identifying the chemical composition andcrystal structure of each precipitate and dispersoid fromTEM images, the upper bound of Drorowan was estimated




by assuming that Orowan dislocation bypassing is opera-tive for all the precipitates and dispersoids (GP zones,g-phase precipitates, g0-phase precipitates and MgO dis-persoids). This is only the starting point since the actualvalue of Drorowan for individual precipitate/dispersoid typeswill be lower than this upper bound, and it is possible thatOrowan dislocation bypassing is not the operativemechanism for each precipitate or dispersoid type. Usingthe values of r and kp, Table 5, in Eq. (7), the upper boundfor the Drorowan for the aggregate effect of all the precipi-tates and dispersoids, as well as for each precipitate anddispersoid type, in samples UFG7075-E, UFG7075-E-T6,CG7075-E and CG7075-E-T6, are estimated to be 45,414, 102 and 472 MPa, respectively.

Next, for comparison and identification of the operablestrengthening mechanisms, the strength increments fromdislocation shearing for g0-phase precipitates, GP zonesand MgO dispersoids are estimated. This mechanism isnot a viable option for the g-phase precipitates becausethey are incoherent with the Al matrix. The relationshipbetween f, kp and �r is [71]:

kp ¼ 2�rffiffiffiffiffiffip

4f

r� 1

� �: ð11Þ

Substituting f for kp using Eq. (11), Eq. (8) becomes:

Drcs ¼3

2

� �14

ffiffiffiffiffiffi2pb

rMGae

ðec�rÞ32

kp þ 2�r: ð12Þ

The g0-MgZn2 phase has an hexagonal structure:a = 0.496 nm and c = 1.402 nm [72]. The orientation rela-tionship between the g0-phase and the Al matrix is(001)g0//{11 1}Al and [110]g0//h112iAl [72]. The interrela-tionship between the lattice parameters of the g0-phaseand the Al matrix are given by d100 (g0) = 3d220(Al) andd001(g0) = 6d111(Al), which makes the g0-semicoherent withthe Al fcc lattice [72]. To calculate the semicoherent latticeparameter misfit between precipitates (hexagonal structure)and matrix (cubic structure), the unique constrained effec-tive misfit strain, ec, is defined by [25,73]:

ec ¼ð1þ tÞ

3ð1� tÞeeff

ð1þ 4G=3BcÞ; ð13Þ

where Bc is the bulk modulus of the g0-MgZn2 precipitates,63.5 GPa [74], and eeff is:

eeff ¼ffiffiffi2p

3½ðe11 � e22Þ2 þ ðe22 � e33Þ2 þ ðe33 � e11Þ2�

12: ð14Þ

where

e11 ¼ap

am� 1; e22 ¼

ffiffiffi3p

apffiffiffi2p

am

� 1; e33 ¼cpffiffiffi2p

am

� 1; ð15Þ

where am is the lattice parameter of the Al matrix,0.405 nm; ap and cp are the lattice parameters of theg0-phase precipitate: ap = 0.496 nm and cp = 1.402 nm,respectively [72]. Utilizing Eqs. (11)–(13), the value of ec

is �0.3. Substituting ec and the values of �r for the g0-phase


and the values of kp into Eq. (12), the lower bound of thestrength increments due to coherency strengthening,Drcs, are approximately several GPa for all the materials.Thus, the upper bound of the strength increment from Oro-wan dislocation bypassing is smaller than the lower boundof strengthening from coherency for all cases, i.e.Drorowan < Drcs, implying that Orowan dislocation bypass-ing is the operative mechanism for the g0-phase because thesmaller of Drorowan and Drshearing is the operative mecha-nism [25,47].

In the case of GP zones, Berg0s study [8] revealed thatGP zones are Zn-rich layers on {111} planes, with internalorder in the form of elongated h110i domains and a spac-ing between rows of atoms �6–8% less than in the Almatrix. The reduced spacing, relative to the Al matrix lay-ers above or below the sheet, was proposed to be associatedwith the smaller radius of Zn atoms [8]. In Eq. (12), weassumed the ec came from this reduced spacing and is equalto 6%. Utilizing the values of �r for GP zones and kp intoEq. (12), the lower bound of Drcs for GP zones in samplesUFG7075-E, UFG7075-E-T6 and CG7075-E-T6 areapproximately 96, 1000 and 590 MPa, respectively. Thesevalues are greater than the upper bound of Drorowan, i.e.Drorowan < Drcs. Hence, similar to the g0-phase, Orowandislocation bypassing is determined to be the operativemechanism for GP zones.

For the UFG 7075 materials, oxide dispersoids mustalso be considered because cryomilling promotes their for-mation [14,15]. The APT analyses demonstrate that thevolume fraction of MgO dispersoids is �0.14% inUFG7075-E, with an average diameter of �4 nm. UsingEq. (11), kMgO is �73 nm, the bulk modulus of MgO is155 GPa [75] and MgO has a cubic structure with a latticeparameter of 4.21 A [76]. The lattice parameter misfitbetween MgO and Al, e, is 0.04. Applying Eq. (12), thecoherency strengthening, Drcs, from MgO, is 220 MPa.This value is larger than the upper bound of Drorowan inUFG7075-E (45 MPa). Therefore, Orowan dislocationbypassing is identified as the operative strengthening mech-anism for MgO dispersoids. Although sample UFG7075-E-T6 was not characterized by APT, it is reasonable toassume that strengthening from oxide dispersoids is similarto that in sample UFG7075-E, because the thermal treat-ments from solutionizing and T6 temper are not expectedto modify the oxide dispersoid distribution characteristics.

Consequently, Orowan dislocation bypassing has beendetermined to be the operative strengthening mechanismfor all of the second-phase particles (the three families ofprecipitates and the oxide dispersoids) in all the materials,which leads to strength increments of 45, 414, 102 and472 MPa for samples UFG7075-E, UFG7075-E-T6,CG7075-E and CG7075-E-T6, respectively. Comparingthese calculated values, it is clear that precipitationstrengthening plays a more significant role in the CG7075 materials than in UFG 7075 for equivalent condi-tions. The experimental results, Table 4, suggest that sam-ple UFG7075-E-sol maintained the same strength as that




of sample UFG7075-E. We speculate that the loss of pre-cipitation strengthening (45 MPa) after solution treatmentwas compensated for by the increase from solid-solutionstrengthening because 45 MPa is within the range of theideal solid-solution strengthening increment with an upperlimit of �82 MPa. In contrast, sample CG7075-E-solexhibited a strength decrease of 47 MPa, compared withthe strength of sample CG7075-E. Comparing the calcu-lated values, the loss of precipitation strengthening(�102 MPa) in sample CG7075-E after solution treatmentcannot be totally compensated for by solid-solutionstrengthening, less than �82 MPa, which is consistent withour experimental observations.

5. Conclusions

The mechanical properties of UFG Al 7075 materials,fabricated by cryomilling feedstock powder with subse-quent consolidation and various heat-treatment conditions,were investigated, with direct comparison to those of CG7075 counterpart materials that were consolidated and heattreated through identical processes using commercially gas-atomized powder as the feedstock. The microstructures ofsamples in the as-extruded and the T6 temper conditionwere characterized by XRD and TEM. Additionally,APT was utilized to study the UFG 7075 material to inves-tigate the nanoscale precipitates and oxide dispersoids thatare produced during cryomilling. The contributions fromdifferent strengthening mechanisms were quantitativelymeasured (including grain-boundary (Hall–Petch strength-ening), dislocation, solid-solution, precipitation and oxidedispersoid strengthening) in the UFG 7075 material inthe as-extruded condition and after the T6 treatment, incomparison with the CG 7075 counterparts. With multiplefamilies of precipitates and oxide dispersoids forming in thematerials, Orowan dislocation bypassing is the dominantstrengthening mechanism in the CG 7075 materials, espe-cially after the T6 temper. Grain-boundary strengtheningis dominant in the as-extruded UFG 7075, and it alsomakes a significant contribution in sample UFG7075-E-T6. In summary, we conclude the following:

(i) The UFG 7075 materials exhibit higher strength thanthe CG 7075 materials in each equivalent condition.The strength of the as-extruded UFG 7075 (YS:583 MPa, UTS: 631 MPa) is even higher than com-mercial Al 7075 with the T6 temper (YS: 541 MPa,UTS: 593 MPa) and is close in value to the CG7075 sample with the T6 temper (YS: 613 MPa,UTS: 659 MPa).

(ii) After a T6 temper, both the UFG and CG 7075 mate-rials exhibit a considerable increase in strength. TheYS and UTS of sample UFG7075-E-T6 are 734 and774 MPa, respectively, which are �120 MPa higherthan those of sample CG7075E-T6 (YS: 613 MPa,UTS: 659 MPa). The age-hardening effect, i.e. thestrength increment after natural aging or T6 temper,


in the UFG 7075 materials is not, however, as signif-icant as that in the CG 7075 materials. The increasesin yield strength of samples CG7075-E-Nag andCG7075-E-T6 are 70% and 160%, respectively, rela-tive to that of CG7075-E-sol. In contrast, theincreases in yield strength after natural aging andT6 temper for the UFG 7075 materials are only13% and 26%, respectively.

(iii) Sample UFG7075-E possesses a relatively small vol-ume fraction, �0.07% from APT, of GP zones, andg0-phase precipitates in grain interiors: the averagediameter of both precipitate types is <5 nm. APTresults demonstrate that this material also contains0.14% volume fraction of MgO dispersoids, with anaverage diameter of �4 nm. In contrast, sampleCG7075-E exhibits a number of plate-like g0-phaseprecipitates (�55 nm diameter) and lath-like g-phaseprecipitates (length �174 nm) in grain interiors: theseprecipitates pin dislocations. After the T6 temper, thegrains in the UFG 7075 material contain a volumefraction �1.5% of spherical GP zones and platelet-like g0-phase precipitates, both in nanoscale (diame-ter 6 5 nm), while sample CG7075-E-T6 exhibitsmany spherical GP zones, diameter < 5 nm, but coar-ser g0-phase precipitates whose average diameter is�61 nm.

(iv) The analyses of the contributions from differentstrengthening mechanisms indicate that grain-bound-ary strengthening is the predominant mechanism inthe as-extruded UFG 7075 material, contributing astrength increment of �242 MPa, and it is also animportant contribution (�185 MPa) in the T6-tem-pered UFG 7075; while precipitation Orowan strength-ening is predominant in the T6-tempered UFG 7075,contributing a strength increment of 414 MPa. In con-trast, Orowan strengthening is the primary strengthcontributor in the CG 7075 materials, both in the as-extruded and T6-tempered states, with values ofapproximately 102 and 472 MPa, respectively.

Acknowledgements

The authors would like to acknowledge financial sup-port provided by the Office of Naval Research (GrantNo. ONR N00014-12-1-0237), Dr. Lawrence Kabacoff asthe Program Officer. The authors are also grateful for tech-nical discussions with Dr. Ali Yousefiani from Boeing Re-search & Technology. Atom-probe tomography wasperformed at the Northwestern University Center forAtom-Probe Tomography (NUCAPT). The local-electrodeatom-probe (LEAP) tomograph was purchased and up-graded with funding from NSF-MRI (DMR-0420532)and ONR-DURIP (N00014-0400798, N00014-0610539,N00014-0910781) grants. Instrumentation at NUCAPTwas also supported by the Initiative for Sustainabilityand Energy at Northwestern (ISEN). This research also




made use of the Shared Facilities at the Materials ResearchCenter of Northwestern University, supported by theNational Science Foundation’s MRSEC program(DMR-1121262).

References

[1] Fine M. Metall Mater Trans A 1975;6:625.[2] Nakai M, Eto T. High strength heat treatable 7000 series aluminum

alloy of excellent corrosion resistance and a method of producingthereof. United States: Kabushiki Kaisha Kobe Seiko Sho (Kobe,JP), US Patent 6048415, 2000.

[3] Park JK, Ardell AJ. Mater Sci Eng A 1989;114:197.[4] Zhao YH, Liao XZ, Jin Z, Valiev RZ, Zhu YT. Acta Mater

2004;52:4589.[5] Viana F, Pinto AMP, Santos HMC, Lopes AB. J Mater Process

Technol 1999;92–93:54.[6] Richard D, Adler P. Metall Mater Trans A 1977;8:1177.[7] Adler P, DeIasi R. Metall Mater Trans A 1977;8:1185.[8] Berg LK, Gjonnes J, Hansen V, Li XZ, Knutson-Wedel M, Waterloo

G, et al. Acta Mater 2001;49:3443.[9] Zhao YH, Liao XZ, Zhu YT, Valiev aRZ. J Mater Res 2005;20:288.

[10] Chen H, Yang B. Mater Trans 2008;49:2912.[11] Donoso E. Mater Sci Eng 1985;74:39.[12] Gleiter H. Nanostruct Mater 1995;6:3.[13] Gleiter H. Acta Mater 2000;48:1.[14] Lavernia EJ, Han BQ, Schoenung JM. Mater Sci Eng A

2008;493:207.[15] Witkin DB, Lavernia EJ. Prog Mater Sci 2006;51:1.[16] Zhu Y, Langdon T. JOM J Miner Metal Mater Soc 2004;56:58.[17] Zhao YH, Liao XZ, Cheng S, Ma E, Zhu YT. Adv Mater

2006;18:2280.[18] Panigrahi SK, Jayaganthan R. Mater Des 2011;32:3150.[19] Panigrahi S, Jayaganthan R. Metall Mater Trans A 2011;42:3208.[20] Liddicoat PV, Liao X-Z, Zhao Y, Zhu Y, Murashkin MY, Lavernia

EJ, et al. Nat Commun 2010;1:63.[21] Pickens JR, Christodoulou L. Metall Trans A 1987;18:135.[22] Ma K, Schoenung JM. Philos Mag Lett 2010;90:739.[23] Tang F, Hagiwara M, Schoenung JM. Mater Sci Eng A 2005;407:306.[24] Zhang Z, Topping T, Li Y, Vogt R, Zhou Y, Haines C, et al. Scripta

Mater 2011;65:652.[25] Wen H, Topping TD, Isheim D, Seidman DN, Lavernia EJ. Acta

Mater 2013;61:2769.[26] Li Y, Zhang Z, Vogt R, Schoenung JM, Lavernia EJ. Acta Mater

2011;59:7206.[27] Topping T, Ahn B, Li Y, Nutt S, Lavernia E. Metall Mater Trans A

2012;43:505.[28] Miller MK, Cerezo A, Hetherington MG, Smith GDW. Atom-probe

field-ion microscopy. Oxford: Clarendon Press; 1996.[29] Seidman DN. Annu Rev Mater Res 2007;37:127.[30] Seidman DN, Stiller K. MRS Bull 2009;34:717.[31] Kelly TF, Miller MK. Rev Sci Instrum 2007;78:031101.[32] Blavette D, Deconihout B, Bostel A, Sarrau JM, Bouet M, Menand

A. Rev Sci Instrum 1993;64:2911.[33] Mulholland MD, Seidman DN. Acta Mater 2011;59:1881.[34] Ringer SP, Hono K. Mater Charact 2000;44:101.[35] Wang JS, Mulholland MD, Olson GB, Seidman DN. Acta Mater

2013;61:4939.[36] Metals handbook, properties and selection: nonferrous alloys and

special-purpose materials, 10th ed. Materials Park, OH: ASMInternational; 1990.

[37] Zhang Z, Dallek S, Vogt R, Li Y, Topping TD, Zhou Y, et al. MetallMater Trans A 2010;41:532.

[38] Hu T, Ma K, Topping TD, Schoenung JM, Lavernia EJ. Acta Mater2013;61:2163.

[39] Williamson GK, Hall WH. Acta Metall 1953;1:22.[40] Collins TJ. Biotechniques 2007;43:S25.


[41] Hellman OC, Vandenbroucke JA, Rusing J, Isheim D, Seidman DN.Microsc Microanal 2000;6:437.

[42] Topping TD, Lavernia EJ. Strain hardening, strain softening and thePortevin-Le Chatelier effect in cryomilled, ultrafine grained AA 5083.In: Weiland H, Rollett AD, Cassada WA, editors. ICAA13: 13thinternational conference on aluminum alloys. Hoboken, NJ: Cass;2012.

[43] Seidman DN, Marquis EA, Dunand DC. Acta Mater 2002;50:4021.[44] Li Y, Liu W, Ortalan V, Li WF, Zhang Z, Vogt R, et al. Acta Mater

2010;58:1732.[45] Cao B, Joshi SP, Ramesh KT. Scripta Mater 2009;60:619.[46] Kim YW, Griffith WM. In: Dispersion strengthened aluminum alloys:

proceedings of the six session symposium on “dispersion strengthenedaluminum alloys”. Phoenix, AZ: TMS; 1988. p. 157.

[47] Courtney TH. Mechanical behavior of materials. Long Grove: Wave-land; 2005.

[48] Kim YW, Griffith WM. PM aerospace materials. Shrewsbury: MPR;1984.

[49] Wert JA. Strength of metal alloys. In: Gifkins RC, editor. Oxford:Pergamon Press; 1980. p. 339.

[50] J.R. Davis & Associates, ASM international handbook committee.Aluminum and aluminum alloys. ASM specialty handbook. Materi-als Park, OH: ASM, International; 1993. p. 784.

[51] Malygin GA. Phys Usp 2011;54:1091.[52] Pande CS, Cooper KP. Prog Mater Sci 2009;54:689.[53] Hall EO. Proc Phys Soc Sec B 1951;64:747.[54] Petch NJ. J Iron Steel Inst Lond 1953;174:25.[55] Loucif A, Figueiredo RB, Baudin T, Brisset F, Chemam R, Langdon

TG. Mater Sci Eng A 2012;532:139.[56] Kamikawa N, Huang X, Tsuji N, Hansen N. Acta Mater

2009;57:4198.[57] Furukawa M, Horita Z, Nemoto M, Valiev RZ, Langdon TG. Acta

Mater 1996;44:4619.[58] Hayes RW, Witkin D, Zhou F, Lavernia EJ. Acta Mater

2004;52:4259.[59] Carsley JE, Fisher A, Milligan WW, Aifantis EC. Metall Mater Trans

A 1998;29:2261.[60] Fleischer RL. Acta Metall 1962;10:835.[61] Fleischer RL. Acta Metall 1963;11:203.[62] Rupert TJ, Trenkle JC, Schuh CA. Acta Mater 2011;59:1619.[63] Topping TD. PhD dissertation: Nanostructured Aluminum Alloys

and Their Composites, University of California Davis; 2012.[64] Bailey JE, Hirsch PB. Philos Mag 1960;5:485.[65] Kumar N, Mishra RS. Mater Sci Eng A 2013;580:175.[66] Huang X, Hansen N, Tsuji N. Science 2006;312:249.[67] Srivatsan TS. Mater Manuf Process 1991;6:565.[68] Ardell A. Metall Mater Trans A 1985;16:2131.[69] Booth-Morrison C, Dunand DC, Seidman DN. Acta Mater

2011;59:7029.[70] Fribourg G, Brechet Y, Deschamps A, Simar A. Acta Mater

2011;59:3621.[71] Brown L, Ham R. Strengthening methods in crystals. Amster-

dam: Elsevier; 1971. p. 9.[72] Li XZ, Hansen V, Gjonnes J, Wallenberg LR. Acta Mater

1999;47:2651.[73] Mulholland MD. PhD thesis, Materials Science and Engineering,

Northwestern University; 2011.[74] Wu M-M, Wen L, Tang B-Y, Peng L-M, Ding W-J. J Alloy Compd

2010;506:412.[75] Madelung O, Rossler U, Schulz M, editors. Magnesium oxide (MgO)

Young’s, shear and bulk moduli, Poisson’s ratio, vols. III/17B-22A-41B. Springer Materials—The Landolt-Bornstein, Database. doi:http://dx.doi.org/10.1007/1081719_210.

[76] Madelung O, Rossler U, Schulz M, editors. Magnesium oxide (MgO)crystal structure, lattice parameters, thermal expansion, vols. III/17B-22A-41B. SpringerMaterials—The Landolt-Bornstein, Database. doi:http://dx.doi.org/10.1007/10681719_206.


http://refhub.elsevier.com/S1359-6454(13)00727-1/h0005




































































































http://dx.doi.org/10.1007/1081719_210

http://dx.doi.org/10.1007/10681719_206


Date post:	20-Jul-2016
Category:	Documents
Upload:	quangsangbk
View:	7 times
Download:	5 times

ma2014

Documents