PROOFS

Aging of Copper-Titanium Dilute Alloys in Hydrogen Atmosphere:

Influence of Prior-Deformation on Strength and Electrical Conductivity*

Satoshi Semboshi1, Shin-ichi Orimo1, Hisashi Suda2, Weilin Gao2 and Akira Sugawara2

1Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan2DOWA METALTECH Co., Ltd., Iwata 438-0025, Japan

The influence of prior-deformation on the mechanical and electrical properties of Cu-4.2 mol% Ti alloys aged in a hydrogen pressure of0.8 MPa was examined. This follows from the results of aging solution-treated Cu-Ti alloys in a hydrogen atmosphere, which significantlyimproved their electrical conductivity over alloys conventionally aged in vacuum, without degradation of the mechanical strength. Themaximum-strength was enhanced in the prior-deformed specimen, and the strengthening and increase in electrical conductivity were acceleratedduring aging in a hydrogen atmosphere, compared to that for the non-deformed specimen. As a result, the balance between the strength and theconductivity is improved within shorter aging time for specimens that are more severely deformed and then aged in a hydrogen atmosphere. Thestrengthening is mainly due to age-hardening by the growth of finely dispersed precipitates of Cu4Ti and TiH2, which are preferentiallynucleated at lattice defects such as dislocations and nano-sized deformation twins. The improved conductivity is closely related to reduction ofthe solute Ti concentration in the copper matrix, which is attributed to the precipitation of TiH2 and Cu4Ti. Thus, prior-deformation assists torender a good combination of strength and electrical conductivity for Cu-Ti dilute alloys during aging in a hydrogen atmosphere.[doi:10.2320/matertrans.M2011173]

(Received August 1, 2011; Accepted September 27, 2011; Published xxxx yy, zzzz)

Keywords: copper-titanium alloy, hydrogen, aging, prior-deformation, strength, electrical conductivity, precipitation

1. Introduction

Age-hardenable copper based alloys are widely used inelectrical applications such as for connectors and lead-frames. The Cu-Be alloys exhibit by far an excellent balanceof mechanical strength (more than 1000 MPa) and electricalconductivity (30% IASC1)).2,3) However, the substitution ofCu-Be alloys is required, due to the relatively high costs andpotential health hazards of beryllium. Cu-Ti alloys contain-ing approximately 1 to 6 mol% Ti are an attractive substitutebecause their mechanical strength is comparable to that ofCu-Be alloys, and they exhibit good stress-relaxation behav-ior and higher thermostability.4–7) While, Cu-Ti alloys areinferior with respect to electrical conductivity, due to themuch larger contribution of the solute Ti to the resistivitythan that of beryllium in the Cu-Be alloys.8) To extendthe industrial applicability of Cu-Ti alloys, it is stronglydesirable to provide Cu-Ti alloys that exhibit both highstrength and high conductivity.

It has been reported that the electrical conductivity ofCu-Ti dilute alloys is significantly improved by aging ina hydrogen atmosphere rather than conventional aging invacuum.9–11) Improvement of the conductivity is due to asignificant reduction in the concentration of the solute Ti inthe matrix, which is caused by the formation of not onlyCu4Ti, but also the titanium hydride (TiH2) phase. Insubsequent research, it was demonstrated that the adjustmentof aging conditions such as temperature and hydrogenpressure was effective to improve the balance of conductivityand strength for these alloys;12,13) aging at a low temperatureprovided a reasonable balance of these properties, although along aging-time was required. Aging under high hydrogenpressure allowed for a rapid improvement in the conductivity,

although the use of high pressure hydrogen gas should beavoided from a safety perspective. Thereby, in order to applythese preliminary findings for the practical fabrication ofCu-Ti alloys with high-strength and high-conductivity, it isnecessary to understand the influence of the alloy composi-tion and pre-treatment, together with the aging conditions,on the mechanical and electrical properties.

Recently, we proposed that cold-rolling deformation couldimprove the balance of hardness and conductivity in Cu-Tialloys during aging in a hydrogen atmosphere within a shortaging time, compared to non-deformed specimens agedunder the same atmospheric conditions.14) However, thepublished data are confined to a mild deformation up to areduction in thickness of 15%. In this work, we firstexamined systematically the strength and conductivity ofCu-Ti dilute alloys that were processed using a solution-treatment and cold-working to reduce thickness by 0 to 60%,followed by aging in a hydrogen atmosphere, to clarifythe effect of prior-deformation on the resulting properties.A promising process-condition of the extent of thicknessreduction during prior-deformation and the aging temper-ature were then adopted and demonstrated for the fabricationof alloys with high-strength and high conductivity compara-ble to those of Cu-Be alloys. The microstructural evolution ofthe alloys prepared was also confirmed using transmissionelectron microscopy (TEM).

2. Experimental Procedure

The nominal composition of the alloy used in this workwas Cu-4.2 mol% Ti, which is the same as a commercialhigh-strength alloy. Sheets of the alloy were prepared bymelting pure copper (99.99%) and titanium (99.99%) as rawmaterials and then hot-rolling to 0.22 mm thickness. Thesheets were solution-treated at 1223 K for 15 min in air andimmediately quenched in water, which resulted in a recrys-

21 / M2011173 / Total page 6

*This Paper was Originally Published in Japanese in J. JRI Cu. 50 (2011)

185–189.

Materials Transactions#2011 Journal of Japan Research Institute for Advanced Copper-Base Materials and Technologies

PROOFS

tallized microstructure with a grain size of approximately10 mm. Some of the sheets were unidirectionally cold-rolleddown to 0.18, 0.15 and 0.08 mm thickness, which correspondto a reduction in thickness of 15, 30 and 60%, respectively.The sheets were cut into plates of 50 mm long and 6 mmwide, and tensile specimens of 20 mm long and 4 mm wide.The longitudinal direction of each tensile specimen wasparallel to the rolling direction. The specimens weremechanically polished with 2000 grade SiC paper to removethe surface oxide layer and were then aged at temperaturesranging from 653 and 623 K under a hydrogen pressure of0.8 MPa.

The electrical conductivity of the aged specimens wasmeasured at room temperature using the standard DC four-probe technique. The Vickers hardness was examined withan applied load of 500 g and a holding time of 10 s. Thehardness number was determined by averaging the results ofmore than ten tests, excluding the maximum and minimumvalues. Tensile tests were conducted using a universal tensiletesting machine (Shimadzu Autograph AG-Xplus) at roomtemperature with a strain rate of 1:67� 10�4 s�1. Themicrostructure of the specimens aged at 623 K under thehydrogen pressure of 0.8 MPa was examined using TEM(Jeol JEM-3010). Thin-foil samples for TEM observationswere first ground to less than 80 mm thick and then electro-polished in a solution of 10 vol% nitric acid in methanolat 243 K with a DC voltage of less than 5 V, followed bylow-angle ion milling using an argon ion beam acceleratedat 3 kV.

3. Results and Discussion

3.1 Electrical conductivity and strengthFigure 1 shows the variations of the electrical conductivity

(a) and Vickers hardness (b) for specimens aged at 653 Kunder a hydrogen pressure of 0.8 MPa after deformation by 0(as solid-solution), 15, 30 and 60% reduction of thickness.The conductivity for all of the deformed specimens beforeaging was approximately 3.5% IACS. This indicates noobvious influence of the deformation on the conductivitywithin the accuracy of this study. The conductivities of allspecimens increased steadily with time during aging in ahydrogen atmosphere (Fig. 1(a)), which exceeded the in-crease in conductivity of approximately 12% to 14% IACSfor the specimen aged in vacuum.8,15) The conductivity forthe specimen deformed to a greater extent increased morerapidly during aging in the hydrogen atmosphere.

Vickers hardness values for the specimens deformed by areduction of 0, 15, 30, and 60% were HV 127, 187, 209,and 215, respectively, which was due to working-hardeningeffects. For the specimen deformed by a reduction of 0%,the hardness kept increasing during aging in the hydrogenatmosphere, even after 144 h (Fig. 1(b)). The hardnessreached a maximum at 96 h for the specimens deformed bya reduction in thickness of 15%, at 48 h for 30% reductionand at 6 h for 60% after aging in the hydrogen atmosphere.Prior-deformation was therefore effective to accelerate thetime to reach maximum hardness during aging in the sameatmosphere. The maximum-hardness was obtained for thespecimen deformed to a greater extent and then aged.

Figure 2 shows the relationship between the conductivityand tensile strength of the specimens prior-deformed at 15,30, and 60% reduction in thickness and then aged at 653 Kunder a hydrogen pressure of 0.8 MPa. Both the strength andelectrical conductivity increased before reaching the max-imum strength, followed by a trade-off relationship betweenthe conductivity and strength. In the specimen deformed to a

21 / M2011173 / Total page 6

200

220

240

260

280

300

320

1 10 100

60% rolled

Aging time, t/ h

Vic

kers

hard

ness

, Hv

/-C

ondu

ctiv

ity,

σ (%

IAC

S)

15% rolled

30% rolled

Quenched

(a)

(b)

0

10

20

30

40

50

Fig. 1 (a) Variations of electrical conductivity and (b) Vickers hardness for

Cu-4.2 mol% Ti alloy cold-rolled to reduction in thickness of 0 to 60% and

then aged at 653 K in a hydrogen atmosphere of 0.8 MPa. The conductivity

is expressed in terms of % IACS, a percentage for the conductivity of

annealed pure copper at 298 K.

700

750

800

850

900

950

1000

1050

0 10 20 30 40 50

Tens

ilest

reng

th,σ

UT

S/M

Pa

Conductivity, σ (% IACS)

60% rolled

15% rolled

30% rolled

36

1224

48 723

6

1224 48

72

36

12

24

48

72

144

1

144

144

Fig. 2 Relationship between the electrical conductivity and tensile strength

of Cu-4.2 mol% Ti alloys cold-rolled to a reduction of 15, 30, and 60% and

then aged at 653 K in a hydrogen atmosphere of 0.8 MPa. The number

beside each plot indicates the aging time in hours.

2 S. Semboshi, S. Orimo, H. Suda, W. Gao and A. Sugawara

PROOFS

greater extent, the maximum value of the tensile strength wasenhanced during aging in the hydrogen atmosphere, whichwas consistent with the variation of the hardness (Fig. 1(b)).Furthermore, Fig. 2 shows that the plots in the over-agingstage were shifted to the upper-right area for the specimendeformed to a greater extent. Therefore, a reasonableimprovement of the balance between the conductivity andstrength of the specimen deformed to a greater extent can beachieved by aging in a hydrogen atmosphere.

Aging of the specimen deformed more severely aftersolution-treatment in a hydrogen atmosphere rendered a morefavorable balance of strength and conductivity in a shorteraging period. In addition, it has been reported that aging atlow temperature under a high hydrogen pressure can alsoimprove the balance of properties for solution-treated speci-mens.12,13) Based on these results, we attempted to fabricateCu-4.2 mol% Ti alloys with an excellent combination ofconductivity and strength using a synergic effect of prior-deformation and aging conditions. Prior-deformed speci-mens, reduced by 60% and aged at a low temperature of623 K under a high hydrogen pressure of 0.8 MPa, wereadopted. Here, the hydrogen pressure employed was possiblethat was still safe. Figure 3 shows the relationship betweenthe electrical conductivity and tensile strength of the speci-mens, in addition to that for Cu-Ti alloys conventionally agedin vacuum and commercial Cu-Be alloys.16) The specimenthat were severely prior-deformed and then aged at a lowtemperature of 623 K exhibited an improved balance ofconductivity and strength, compared to that for the conven-

tional alloys and the specimens aged at 653 K, even thoughit required a long aging period; the specimen exhibited amaximum strength of 1052 MPa with a conductivity of 14%IACS by aging for 24 h. After aging for 120 h, the specimenstill had a strength of more than 1000 MPa and highconductivity of 30% IACS. The combination of strengthand conductivity were comparable with that for somecommercial Cu-Be alloys. The reason why the specimenaged in a lower temperature improved in the balance of thestrength and conductivity is primary because finer dispersionof Cu4Ti precipitates and a larger number of TiH2 proceededin parallel with each other during the aging. The detail hasbeen discussed in the previous work.12)

3.2 Microstructural evolutionThe microstructural evolution of Cu-Ti dilute alloys

solution-treated and then aged at 673 K and 773 K in ahydrogen atmosphere has already been investigated;10,12,17) inthe early stage of aging, spinodal decomposition progressesin the solid solution phase, followed by the precipitation anddispersion of Cu4Ti (MoNi4 structure: space group I4=m,a ¼ 0:583 nm, c ¼ 0:362 nm).4,5) In the subsequent stage,particles of TiH2 (fcc: Fm�33m, a ¼ 0:444 nm18)), were alsoformed by the reaction of dissolved hydrogen atoms withTi atoms in the matrix or in Cu4Ti precipitates. The co-precipitation of Cu4Ti and TiH2 promotes the decrease in theconcentration of solute Ti in the solid solution phase. Onfurthermore aging, TiH2 particles continue to grow at theexpense of Cu4Ti particles, and the microstructure eventuallyconsists of two phase, of a much diluted solid solution andTiH2 particles. The microstructural evolution must beessentially the same as that for the specimen aged at atemperature lower than 673 K in a hydrogen atmosphere.

Figure 4 shows a bright field (BF) TEM image of thespecimen solution-treated and then aged at 653 K for 48 h inthe hydrogen atmosphere. This image shows the modulatedstructure attributed to spinodal decomposition,4,5) but noformation of Cu4Ti and TiH2 precipitates is evident. There-fore, in the case of the solution-treated specimen, aging for48 h is not sufficient, and aging for a longer period is requiredfor the Cu4Ti and TiH2 precipitates to appear.

21 / M2011173 / Total page 6

700

750

800

850

900

950

1000

1050

1100

1150

0 10 20 30 40 50

Tens

ilest

reng

th,σ

UT

S/M

Pa

Conductivity, σ (% IACS)

144

72

36

12

24

48

1224

4872 120

168

Aged at 653 KAged at 623 K

Fig. 3 Relationship between the electrical conductivity and tensile strength

of Cu-4.2 mol% Ti alloys cold-rolled to a reduction of 60% and then aged

at 623 K in a hydrogen atmosphere of 0.8 MPa. The number beside each

plot indicates the aging time in hours. For comparison, the relationships

of alloys aged at 653 K (shown in Fig. 2), representative Cu-Ti alloys

conventionally aged in vacuum, and commercial Cu-Be alloys are also

given (dotted lines).16)

10 nm

Fig. 4 BF TEM image of Cu-4.2 mol% Ti alloy solution-treated, and then

aged at 653 K for 48 h in a hydrogen atmosphere of 0.8 MPa.

Aging of Copper-Titanium Dilute Alloys in Hydrogen Atmosphere: Influence of Prior-Deformation 3

PROOFS

Figure 5 shows BF TEM images and a selected areadiffraction (SAD) pattern of a specimen deformed by areduction of 60%. The image from the transverse direction(Fig. 5(a)) shows the grains extend to the rolling direction.In the higher magnified image (Fig. 5(b)), a high density ofdislocations and deformation twins with an average widthof 15 nm are evident. According to the SAD pattern takenfrom the twin boundary, the twin planes were (111)Cu. Thespecimens deformed by a reduction of 15 and 30% alsocontained deformation twins with widths of approximately50 and 20 nm, respectively. Thus, prior-deformation to amore severe extent introduces a higher-density of latticedefects in the specimens.

Figure 6 shows BF TEM images and a SAD pattern of aspecimen deformed by a reduction of 60% and then aged at653 K for 3, 6, 48, and 144 h under a hydrogen pressure of0.8 MPa. The deformation twins generated by cold-rolling,were not recovered by aging for 3 to 144 h. In the BFTEM image of the specimen aged for 3 h (Fig. 6(a)), moirecontrasts of several nano-meters in size were observedespecially on the twin boundaries, although the onlydiffracted spots detected in the SAD pattern were from thematrix phase of copper solid solution (fcc, a ¼ 0:361 nm).Fine dispersion of the precipitation of Cu4Ti is evident inthe early stage of aging in a hydrogen atmosphere,10,12,17)

therefore, it is suggested that the moire contrasts in the BFTEM image should correspond to the precipitation of Cu4Ti.During the peak-hardened period of aging for the 6 h agedspecimen (Fig. 6(b)), such moire contrasts were formed inthe overall. In the specimen aged for 48 h (Fig. 6(c)), notonly moire contrasts, but also bright granular contrasts ofapproximately 5 nm were visible between the twin bounda-ries. In the corresponding SAD pattern, weak spots due toTiH2 (marked by solid circles) and those of its doublediffractions (by dotted circles), together with fundamentalspots from the matrix phase, were detected. Therefore, the

granular contrasts correspond to TiH2. TiH2 particles weregrown between the twin boundaries by aging for 144 h(Fig. 6(d)). The size of the TiH2 particles was approximately5 nm, which is similar to that for 48 h aging. The TiH2

particles were not significantly coarsened during the agingprocess, because the twin boundaries suppressed the growthof TiH2 particles.

3.3 Influence of prior-deformationStrain and lattice defects such as dislocations and

deformation twins are accumulated by cold-rolling thesolution-treated specimens. In particular, the width of thedeformation twins is less than 50 nm by a reduction inthickness of more than 15%, which is much smaller than thatfor pure copper.19) This suggests that the supersaturated solidsolution of Ti in Cu exhibits a lower stacking fault energythan pure copper. Figures 6(a) and 6(c) shows that duringaging in a hydrogen atmosphere, precipitates of Cu4Ti andTiH2 are preferentially nucleated and grown in dislocationsand deformation twins, because of their high fault energies;the lattice defects in the matrix phase behave as nucleationsites for the precipitation. In addition, a comparison ofFigs. 4 and 6(c) shows that the precipitates of Cu4Ti andTiH2 are developed more rapidly in the prior-deformedspecimen than in the non-rolled specimen after the sameperiod of aging in a hydrogen atmosphere. This is becausethe strain and lattice defects assist in nucleation of theprecipitates, similar to that for aging in vacuum.20,21) Thus,prior-deformation to a more severe extent causes more strainand a large number of lattice defects, which resulting in anincrease of the number of nucleation sites and the nucleationrate for precipitates of Cu4Ti and TiH2.

Strengthening of the Cu-Ti dilute alloys by aging in ahydrogen atmosphere is primarily due to the precipitation ofCu4Ti particles,10) and the improvement in conductivity iscontrolled by the concentration of Ti atoms dissolved in the

21 / M2011173 / Total page 6

500 nm

(a)

50 nm

111Cu

000

220Cu

(b)

Fig. 5 BF TEM images of Cu-4.2 mol% Ti alloys cold-rolled to a reduction of 60%. The rolled (RD) and normal (ND) directions are

indicated by the arrows in (a). The higher magnification image in (b) shows some twin boundaries in the matrix phase, which were

confirmed by the corresponding SAD pattern.

4 S. Semboshi, S. Orimo, H. Suda, W. Gao and A. Sugawara

PROOFS

matrix, which is efficiently reduced by the formation ofTiH2.9) In the severely deformed specimens, the numberdensity of Cu4Ti precipitates is effectively increased duringaging in a hydrogen atmosphere, which results in efficientdispersion-strengthening. Furthermore, the nucleation ofCu4Ti and TiH2 precipitates is accelerated in the deformedspecimen, so that the period for maximum-strengthening isshortened and the conductivity is increased more rapidlyduring aging in a hydrogen atmosphere.

4. Conclusion

The effect of the processing conditions on the strengthand electrical conductivity of Cu-4.2 mol% Ti alloys agedisothermally at 623 to 653 K under the hydrogen pressure of0.8 MPa was investigated, together with the microstructural

evolution of the alloys. The salient results obtained aresummarized as follows.(1) The conductivity increased more rapidly in those

specimens prior-deformed by a greater reduction inthickness. In addition, the maximum values of hardnessand tensile strength were enhanced within a shorteraging time. Controlling the conditions of not only theextent of prior-deformation, but also the aging temper-ature and hydrogen pressure, imparts an excellentbalance of strength and conductivity, of more than1000 MPa and 30% IACS, respectively.

(2) Dislocations and deformation twins with average sizesof several tens of nano-meters in width were generatedin specimens prior-deformed to a reduction in thicknessof more than 15%. These lattice defects were effectiveto increase the nucleation sites in the specimen and

21 / M2011173 / Total page 6

50 nm 50 nm

50 nm50 nm

TiH2

(a) (b)

(c) (d)

Fig. 6 BF TEM images of Cu-4.2 mol% Ti alloys cold-rolled to a reduction of 60%, and then aged at 653 K for (a) 3, (b) 6 (peak-aging

condition), (c) 48, and (d) 144 h in a hydrogen atmosphere of 0.8 MPa. The image in (a) shows moire patterns indicated by circles. The

image in (c) shows TiH2 particles between twin boundaries, which were confirmed by the corresponding SAD pattern (inset). The weak

spots marked by solid circles in the SAD pattern are from TiH2, and the other weak spots marked by dotted circles are due to double

diffraction.

Aging of Copper-Titanium Dilute Alloys in Hydrogen Atmosphere: Influence of Prior-Deformation 5

PROOFS

growth rates of Cu4Ti and TiH2 precipitates duringaging in a hydrogen atmosphere. Therefore, finedispersion of precipitates was formed more rapidly inthe specimens that were prior-deformed to a moresevere extent.

Acknowledgement

The authors thank Profs. S. Hanada and N. Masahashiof the Institute for Materials Research (IMR) of TohokuUniversity, and Prof. H. Numakura and Mr. T. Kondo ofOsaka Prefecture University for useful discussions andcomments, and Mr. E. Aoyagi and Y. Hayasaka of the IMRfor their technical supports. This work was partly performedunder the co-operative research program of AdvancedResearch Center of Metallic Glasses, IMR of TohokuUniversity. Financial support provided by the New Energyand Industrial Technology Development Organization(NEDO), and the Japan Science and Technology Agency(JST) is gratefully acknowledged.

REFERENCES

1) The percentage based on the electrical conductivity of an International

Annealed Copper Standard at 298 K, 5:80� 107 ��1m�1.

2) P. Wilkes: Acta Metall. 16 (1968) 153–158.

3) P. J. Rioja and D. E. Laughlin: Acta Metall. 28 (1980) 1301–1313.

4) A. Datta and W. A. Soffa: Acta Metall. 24 (1976) 987–1001.

5) W. A. Soffa and D. E. Laughlin: Prog. Mater. Sci. 49 (2004) 347–366.

6) S. Nagarjuna, M. Srinvivas, K. Balasubramanian and D. S. Sarma: Acta

Metall. 44 (1996) 2285–2293.

7) A. W. Thompson and J. C. Williams: Metal. Trans. A 15A (1984) 931–

937.

8) S. Nagarjuna, K. Balasubramanian and D. S. Sarma: Matar. Sci. Eng. A

225 (1997) 118–124.

9) S. Semboshi and T. J. Konno: J. Mater. Res. 23 (2008) 473–477.

10) S. Semboshi, T. Nishida and H. Numakura: Mater. Sci. Eng. A 517

(2009) 105–113.

11) A. Kamegawa, T. Iwata and M. Okada: Mater. Sci. Forum 654–656

(2010) 1319–1322.

12) S. Semboshi, T. Nishida, H. Numakura, T. Al-Kassaab and R.

Kirchheim: Metall. Mater. Trans. A, in press.

13) S. Semboshi, T. Nishida and H. Numakura: Mater. Trans. 52 (2011)

605–609.

14) S. Semboshi, H. Numakura, H. Suda, W. L. Gao and A. Sugawara:

Mater. Sci. Forum 645–656 (2010) 1315–1318.

15) S. Suzuki, K. Hirabayashi, H. Shibata, K. Mimura, M. Isshiki and Y.

Waseda: Scr. Mater. 48 (2003) 431–435.

16) S. Nagarjuna, M. Srinivas, K. Balasubramanian and D. S. Sarma:

Mater. Sci. Eng. A 259 (1999) 34–42.

17) S. Semboshi, T. Al-Kassaab, R. Gemma and R. Kirchheim: Ultra-

microscopy 109 (2009) 593–598.

18) S. B. Qadri, E. F. Skelton, M. Nagumo, A. W. Webb, F. E. Lynch and

R. W. Marmaro: Phys. Rev. B 49 (1992) 1193–1194.

19) L. Balogh, T. Ungar, Y. Zhao, Y. T. Zhu, Z. Horita, C. Xu and T. G.

Langdon: Acta Mater. 56 (2008) 809–820.

20) S. Nagarjuna and M. Srinivas: Mater. Sci. Eng. A 498 (2008) 468–474.

21) S. Nagarjuna, K. Balasubramanian and D. S. Sarma: J. Mater. Sci. 32

(1997) 3375–3385.

21 / M2011173 / Total page 6

6 S. Semboshi, S. Orimo, H. Suda, W. Gao and A. Sugawara