《金属材料强韧化与组织调控》教学资源（参考文献）Science_Hardening by Annealing and Softening by Deformation in Nanostructured Metals

REPORTS Hardening by Annealing and almost brittle.This is in contrast to the expected behavior after annealing-a decrease in strength and an increase in elongation or ductility Softening by Deformation in An increase in flow stress during annealing of a deformed metal is typical if the metal con- Nanostructured Metals tains alloying elements in solid solution that precipitate when the metal is annealed.This so- Xiaoxu Huang,1*Niels Hansen,1 Nobuhiro Tsuji? called precipitation hardening was not expected in the aluminum.which had a purity of 99.2% We observe that a nanostructured metal can be hardened by annealing and softened when However,other stable impurities might have subsequently deformed,which is in contrast to the typical behavior of a metal.Microstructural dissolved during processing and reprecipitated investigation points to an effect of the structural scale on fundamental mechanisms of dislocation- during annealing.To prove that this was not dislocation and dislocation-interface reactions,such that heat treatment reduces the generation the case,we carried out an experiment using and interaction of dislocations,leading to an increase in strength and a reduction in ductility. 99.99%pure aluminum as the starting material. A subsequent deformation step may restore the dislocation structure and facilitate the yielding As with the 99.2%pure aluminum,specimens process when the metal is stressed.As a consequence,the strength decreases and the ductility were produced by a six-cycle ARB processing increases.These observations suggest that for materials such as the nanostructured aluminum to an equivalent strain of 4.8,followed by an- studied here,deformation should be used as an optimizing procedure instead of annealing. nealing at 150C for 30 min.The specimens were tested under the same conditions.The rom the beginning of our civilization,the changes in properties and structure when results showed that,as expected,both the yield metalworkers have known that when a a nanostructured metal is annealed,and to use stress and the UTS were substantially reduced metal becomes too hard-for example, such findings to inspire the development of relative to the 99.2%Al,but the key phenome- when forged-it can be softened by annealing. new optimization processes. na (ie.,the hardening by annealing and the By choosing the right combination of annealing We investigated annealing behavior in a decrease in ductility)were reproduced.For temperature and time,a desired combination of fully dense,nanostructured aluminum of com- example,an increase of about 9%in the yield strength and ductility can be achieved.The cur- mercial purity (99.2%)that was prepared by a stress was observed after annealing.Therefore, rent focus is on nanostructured metals that have high-strain rolling deformation known as ac- the dissolution and reprecipitation of impurities, extreme strength but limited ductility and form- cumulative roll bonding (ARB)(/2).Alumi- if it occurred,did not contribute to the annealing- ability,which reduces their applicability.The num sheets of a final thickness of I mm were induced property changes. extreme strength is obtained through a structural produced by a six-cycle ARB processing to We used transmission electron microscopy refinement of the grains down to nanometer an equivalent strain of 4.8(/3).The ARB- (TEM)and high-resolution TEM(HRTEM)to dimensions,and an optimization of ductility has processed state showed a weak crystallo- characterize the structural parameters of ARB been sought through annealing.It has been graphic texture and a lamellar microstructure samples before and after annealing.The initial shown (/that when annealing under conditions of dislocation boundaries and grain boundaries structure (Fig.2A)is delineated by lamellar that produce a structure with bimodal distribu- characteristic of high-strain rolling of metals boundaries parallel to the rolling direction(RD) tion of grain sizes from nanometer to microme- and alloys.The lamellar boundaries are parallel and interconnecting boundaries parallel to the ter scales,the strength of nanostructured metals to the rolling plane,with an average spacing of normal direction (ND).Table 1 shows the quan- decreases slightly but the deformation-induced 180 nm.This lamellar morphology and the rel- tified structural parameters that are thought to hardening (i.e.,work hardening)of the coarse atively coarse boundary spacing ensure the contribute to the mechanical properties.Slight grains in the structure gives ductility.It has been elimination of grain boundary sliding during coarsening occurred during annealing (Fig.2B) also discovered that during annealing at low tensile testing.Tensile specimens of gauge di- for both the lamellar boundary spacing D and temperatures,which does not cause excessive mensions 10 mm by 5 mm were machined from the interconnecting boundary spacing Dic and heterogeneous coarsening of the nano- the sheets and tested at room temperature.The which produced a reduction of boundary surface structure,nanostructured metals may harden engineering stress-strain curve of the ARB area per unit volume. rather than soften-as observed,for example. sample (Fig.1,curve 1)shows a very high yield Statistical measurements of misorientation in metals produced by inert gas condensation stress (259 MPa)and ultimate tensile stress angles across the lamellar boundaries and inter- (24),electrodeposition (5,6),and plastic de- (UTS,334 MPa)and a reasonably good tensile connecting boundaries were made by Kikuchi formation to very high strains (7-10).Associ- ductility,as expressed by total elongation (7%) diffraction.Misorientation angles in both sam- ated with the hardening,a decrease in the tensile and uniform elongation(1.8%).This yield stress ples show a bimodal distribution,with one peak ductility has been reported (5,9,10)where ten- is nearly 10 times that of a coarse-grained ma- located in the range below 3 and the other sile tests were carried out to evaluate the mechan- terial with a grain size of 50 um (~28 MPa). located between40°and55°.More than60%of ical properties.This unusual,annealing-induced When the ARB sample was annealed at 150C the boundaries were high-angle boundaries hardening has been related to changes in struc- for 30 min,the yield stress increased by 8.9%to (>15)in the ARB samples both before and tural characteristics [e.g.,the grain boundary 281 MPa (curve 2,Fig.1)and the total elonga- after annealing.This high density,in combina- structure (//)]but various hypotheses have tion decreased markedly,making the material tion with the small boundary spacing.results in not been verified.The present work has two objectives:to improve our understanding of Table 1.Structural parameters in samples of different conditions.f,fraction of boundaries with Center for Fundamental Research:Metal Structures in misorientation angles less than 3;s fraction of boundaries with misorientation angles Four Dimensions,Materials Research Department,Rise between 3 and 15;fs,fraction of high-angle boundaries (>15). National Laboratory,DK 4000 Roskilde,Denmark.De- partment of Adaptive Machine Systems,Graduate School Sample P。(m-2) DLB (nm))Dice(nm）fes(%方-1s(%) f15(%） of Engineering,Osaka University,2-1 Yamada-Oka,Suita, ARB 1.33×1014 180 600 17.5 16.2 66.3 0saka565-0871,]apan. ARB annealed 0.53×1014 225 640 12.8 23.3 63.9 *To whom correspondence should be addressed.E-mail: at150°cfor30min xiaoxu.huang@risoe.dk www.sciencemag.org SCIENCE VOL 312 14 APRIL 2006 249

Hardening by Annealing and Softening by Deformation in Nanostructured Metals Xiaoxu Huang,1 * Niels Hansen,1 Nobuhiro Tsuji2 We observe that a nanostructured metal can be hardened by annealing and softened when subsequently deformed, which is in contrast to the typical behavior of a metal. Microstructural investigation points to an effect of the structural scale on fundamental mechanisms of dislocation￾dislocation and dislocation-interface reactions, such that heat treatment reduces the generation and interaction of dislocations, leading to an increase in strength and a reduction in ductility. A subsequent deformation step may restore the dislocation structure and facilitate the yielding process when the metal is stressed. As a consequence, the strength decreases and the ductility increases. These observations suggest that for materials such as the nanostructured aluminum studied here, deformation should be used as an optimizing procedure instead of annealing. From the beginning of our civilization, metalworkers have known that when a metal becomes too hard—for example, when forged—it can be softened by annealing. By choosing the right combination of annealing temperature and time, a desired combination of strength and ductility can be achieved. The cur￾rent focus is on nanostructured metals that have extreme strength but limited ductility and form￾ability, which reduces their applicability. The extreme strength is obtained through a structural refinement of the grains down to nanometer dimensions, and an optimization of ductility has been sought through annealing. It has been shown (1) that when annealing under conditions that produce a structure with bimodal distribu￾tion of grain sizes from nanometer to microme￾ter scales, the strength of nanostructured metals decreases slightly but the deformation-induced hardening (i.e., work hardening) of the coarse grains in the structure gives ductility. It has been also discovered that during annealing at low temperatures, which does not cause excessive and heterogeneous coarsening of the nano￾structure, nanostructured metals may harden rather than soften—as observed, for example, in metals produced by inert gas condensation (2–4), electrodeposition (5, 6), and plastic de￾formation to very high strains (7–10). Associ￾ated with the hardening, a decrease in the tensile ductility has been reported (5, 9, 10) where ten￾sile tests were carried out to evaluate the mechan￾ical properties. This unusual, annealing-induced hardening has been related to changes in struc￾tural characteristics Ee.g., the grain boundary structure (11)^, but various hypotheses have not been verified. The present work has two objectives: to improve our understanding of the changes in properties and structure when a nanostructured metal is annealed, and to use such findings to inspire the development of new optimization processes. We investigated annealing behavior in a fully dense, nanostructured aluminum of com￾mercial purity (99.2%) that was prepared by a high-strain rolling deformation known as ac￾cumulative roll bonding (ARB) (12). Alumi￾num sheets of a final thickness of 1 mm were produced by a six-cycle ARB processing to an equivalent strain of 4.8 (13). The ARB￾processed state showed a weak crystallo￾graphic texture and a lamellar microstructure of dislocation boundaries and grain boundaries characteristic of high-strain rolling of metals and alloys. The lamellar boundaries are parallel to the rolling plane, with an average spacing of 180 nm. This lamellar morphology and the rel￾atively coarse boundary spacing ensure the elimination of grain boundary sliding during tensile testing. Tensile specimens of gauge di￾mensions 10 mm by 5 mm were machined from the sheets and tested at room temperature. The engineering stress-strain curve of the ARB sample (Fig. 1, curve 1) shows a very high yield stress (259 MPa) and ultimate tensile stress (UTS, 334 MPa) and a reasonably good tensile ductility, as expressed by total elongation (7%) and uniform elongation (1.8%). This yield stress is nearly 10 times that of a coarse-grained ma￾terial with a grain size of 50 mm (È28 MPa). When the ARB sample was annealed at 150-C for 30 min, the yield stress increased by 8.9% to 281 MPa (curve 2, Fig. 1) and the total elonga￾tion decreased markedly, making the material almost brittle. This is in contrast to the expected behavior after annealing—a decrease in strength and an increase in elongation or ductility. An increase in flow stress during annealing of a deformed metal is typical if the metal con￾tains alloying elements in solid solution that precipitate when the metal is annealed. This so￾called precipitation hardening was not expected in the aluminum, which had a purity of 99.2%. However, other stable impurities might have dissolved during processing and reprecipitated during annealing. To prove that this was not the case, we carried out an experiment using 99.99% pure aluminum as the starting material. As with the 99.2% pure aluminum, specimens were produced by a six-cycle ARB processing to an equivalent strain of 4.8, followed by an￾nealing at 150-C for 30 min. The specimens were tested under the same conditions. The results showed that, as expected, both the yield stress and the UTS were substantially reduced relative to the 99.2% Al, but the key phenome￾na (i.e., the hardening by annealing and the decrease in ductility) were reproduced. For example, an increase of about 9% in the yield stress was observed after annealing. Therefore, the dissolution and reprecipitation of impurities, if it occurred, did not contribute to the annealing￾induced property changes. We used transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) to characterize the structural parameters of ARB samples before and after annealing. The initial structure (Fig. 2A) is delineated by lamellar boundaries parallel to the rolling direction (RD) and interconnecting boundaries parallel to the normal direction (ND). Table 1 shows the quan￾tified structural parameters that are thought to contribute to the mechanical properties. Slight coarsening occurred during annealing (Fig. 2B) for both the lamellar boundary spacing DLB and the interconnecting boundary spacing DICB, which produced a reduction of boundary surface area per unit volume. Statistical measurements of misorientation angles across the lamellar boundaries and inter￾connecting boundaries were made by Kikuchi diffraction. Misorientation angles in both sam￾ples show a bimodal distribution, with one peak located in the range below 3- and the other located between 40- and 55-. More than 60% of the boundaries were high-angle boundaries (915-) in the ARB samples both before and after annealing. This high density, in combina￾tion with the small boundary spacing, results in 1 Center for Fundamental Research: Metal Structures in Four Dimensions, Materials Research Department, Risø National Laboratory, DK 4000 Roskilde, Denmark. 2 De￾partment of Adaptive Machine Systems, Graduate School of Engineering, Osaka University, 2-1 Yamada-Oka, Suita, Osaka 565-0871, Japan. *To whom correspondence should be addressed. E-mail: xiaoxu.huang@risoe.dk Table 1. Structural parameters in samples of different conditions. f G3- , fraction of boundaries with misorientation angles less than 3-; f 3–15- , fraction of boundaries with misorientation angles between 3- and 15-; f 915- , fraction of high-angle boundaries (915-). Sample r0 (mj2) DLB (nm) DICB (nm) f 15- (%) ARB 1.33
1014 180 600 17.5 16.2 66.3 ARB annealed at 150-C for 30 min 0.53
1014 225 640 12.8 23.3 63.9 REPORTS www.sciencemag.org SCIENCE VOL 312 14 APRIL 2006 249

REPORTS 400 400 (edW) 300 3 200 200 c=4.17x10s1 元=4.17x104s1 100 0 2 46 8 10 Engineering Strain(%) 0 2 4 6 8 10 Engineering Strain(%) Fig.1.Engineering stress-strain curves for 99.2% pure AL.Curve 1:processed by six ARB cycles to an 200nm Fig.3.Engineering stress-strain curves for 99.2% equivalent strain of 4.8.Curve 2:same material as pure AL.Curve 2:ARB annealed at 150C for 30 min 1,plus annealing at 150C for 30 min.The strain (the same curve 2 as shown in Fig.1).Curve 3:same rate used for the tensile test is indicated.Refer as 2 but deformed 15%by cold rolling.Curve 4: to table S2 for sample numbering. same as 3 but again annealed at 150C for 30 min. Curve 5:same as 4 but deformed 15%by cold rolling.Refer to table S2 for sample numbering. a very large area per unit volume (S)of high- angle boundaries,which can act as dislocation sinks.As an example,S,in the ARB sample in order to activate new dislocation sources before annealing is ~5.6 x 106 m-,which is during straining.Such a correlation between the about 100 times the for a typical polycrystal- density of dislocation sources and the strength is line material with a grain size of 50 um.In the typically observed in nanoscale metals,both bright-field images(Fig.2),it is seen that a series M experimentally (18)and by atomic-scale of alterating bright and dark thickness contours modeling (/9).Furthermore,the decrease in characterize individual high-angle boundaries. the density of interior dislocations that can carry most of which correspond to the lamellar bound- the strain may efficiently reduce the elongation. aries.HRTEM observations show that the lattice images on both sides of a lamellar boundary 200nm These different effects are reflected in Fig.1. RD A critical test for the above hypothesis is to extend all the way to the boundary:no special see whether a softening and an increase in elon- region with a disordered structure is detected Fig.2.TEM images showing the lamellar struc- gation occur if dislocations are generated in the along the boundary.The low-angle boundaries tural morphology and dislocation configuration in annealed sample.An annealed sample was de- exhibit a certain width in the deformed state. two ARB samples,(A)before annealing and (B) formed 15%by cold rolling and tested under the but they become sharper and better defined upon after annealing at 150C for 30 min.The micro- same conditions.The tensile curve for this test annealing,which suggests the occurrence of a graphs were recorded when at least one lamella is plotted as curve 3 in Fig.3.It is seen that the recovery process by rearrangement of disloca- (M)in the area was under multiple-beam con- stress-strain behavior of this redeformed sample tions in these boundaries.The most remark- dition with the beam direction parallel to the returns to that of the initial ARB sample(Fig able change observed was the decrease in the [001]zone axis of the lamella to reveal the 1),as do the values for the measured yield stress density of interior dislocations,Po that exist interior dislocations.In (A),dislocation tangles are (256 MPa),UTS (333 MPa),total elongation in the volume between the boundaries (Fig. seen within the lamella marked M.In (B),dis- (6.6%).and uniform elongation (2.0%).The 2).A determination of the interior disloca- location tangles are replaced by fewer disloca- similarity in mechanical behavior between this tion density showed that it decreased from tions,many of which are pinned by the lamellar 15%cold-rolled sample and the original ARB 1.33 x 1014 m-2 in the deformed state to boundaries. sample suggests that applying 15%cold rolling 0.53 x 1014 m-2 after annealing (Table 1). has modified the structure in the annealed sam In previous strength-structural analysis of When a deformed structure is annealed at a ple to be similar to the structure in the initial lamellar structures,two additive strengthening temperature that does not cause recrystalli- ARB sample.TEM characterization shows that mechanisms have been proposed:(i)forest hard- zation,typical effects include a coarsening of a large number of dislocations are indeed intro- ening caused by the dislocations in the low- boundary spacing,recovery of low-angle bound- duced again in the volume between the bound- angle boundaries and in the volume between aries.and reduction in the dislocation density aries (Fig.4)and at triple junctions and grain the boundaries (/4),and (ii)grain boundary in the grain interior,at grain boundaries and boundary regions.As a result,the dislocation hardening caused by the high-angle boundaries triple junctions.In conventional materials with configuration is very similar to that observed in (14.15)taken to be inversely proportional to the medium to large grain sizes,these changes will the original ARB sample (Fig.2A).We ob- square root of the boundary spacing [i.e.,a Hall- cause softening by a reduction in dislocation tained an interior dislocation density of 1.14 x Petch relationship (/6,/7)].Not included is a hardening and grain boundary strengthening. 1014 m-2 and boundary spacings of 200 nm contribution from texture strengthening,as the However,the changes in the dislocation struc- (DB)and 650 nm (Dic)in this sample.These texture change during annealing is negligible. ture occurring in a nanostructured metal may parameters are close to the values measured for Therefore,the observed coarsening and de- play a distinct and different role.As a the original ARB sample (Table 1). crease in the dislocation density in the volume hypothesis,it is suggested that the many To further verify this hypothesis,we car- between the boundaries and the decrease in the dislocation sinks available in the form of closely ried out further annealing and deformation low-angle boundary fraction suggest a decrease spaced high-angle boundaries will reduce the experiments as well as tensile tests.Repeated in strength,which is opposite to the hardening number of dislocation sources during annealing. hardening and decrease in the elongation by low- observed experimentally. This may lead to an increase in the yield stress temperature annealing,and softening and in- 250 14 APRIL 2006 VOL 312 SCIENCE www.sciencemag.org

a very large area per unit volume (SV) of high￾angle boundaries, which can act as dislocation sinks. As an example, SV in the ARB sample before annealing is È5.6
106 mj1 , which is about 100 times the SV for a typical polycrystal￾line material with a grain size of 50 mm. In the bright-field images (Fig. 2), it is seen that a series of alternating bright and dark thickness contours characterize individual high-angle boundaries, most of which correspond to the lamellar bound￾aries. HRTEM observations show that the lattice images on both sides of a lamellar boundary extend all the way to the boundary; no special region with a disordered structure is detected along the boundary. The low-angle boundaries exhibit a certain width in the deformed state, but they become sharper and better defined upon annealing, which suggests the occurrence of a recovery process by rearrangement of disloca￾tions in these boundaries. The most remark￾able change observed was the decrease in the density of interior dislocations, r0, that exist in the volume between the boundaries (Fig. 2). A determination of the interior disloca￾tion density showed that it decreased from 1.33
1014 mj2 in the deformed state to 0.53
1014 mj2 after annealing (Table 1). In previous strength-structural analysis of lamellar structures, two additive strengthening mechanisms have been proposed: (i) forest hard￾ening caused by the dislocations in the low￾angle boundaries and in the volume between the boundaries (14), and (ii) grain boundary hardening caused by the high-angle boundaries (14, 15) taken to be inversely proportional to the square root of the boundary spacing Ei.e., a Hall￾Petch relationship (16, 17)^. Not included is a contribution from texture strengthening, as the texture change during annealing is negligible. Therefore, the observed coarsening and de￾crease in the dislocation density in the volume between the boundaries and the decrease in the low-angle boundary fraction suggest a decrease in strength, which is opposite to the hardening observed experimentally. When a deformed structure is annealed at a temperature that does not cause recrystalli￾zation, typical effects include a coarsening of boundary spacing, recovery of low-angle bound￾aries, and reduction in the dislocation density in the grain interior, at grain boundaries and triple junctions. In conventional materials with medium to large grain sizes, these changes will cause softening by a reduction in dislocation hardening and grain boundary strengthening. However, the changes in the dislocation struc￾ture occurring in a nanostructured metal may play a distinct and different role. As a hypothesis, it is suggested that the many dislocation sinks available in the form of closely spaced high-angle boundaries will reduce the number of dislocation sources during annealing. This may lead to an increase in the yield stress in order to activate new dislocation sources during straining. Such a correlation between the density of dislocation sources and the strength is typically observed in nanoscale metals, both experimentally (18) and by atomic-scale modeling (19). Furthermore, the decrease in the density of interior dislocations that can carry the strain may efficiently reduce the elongation. These different effects are reflected in Fig. 1. A critical test for the above hypothesis is to see whether a softening and an increase in elon￾gation occur if dislocations are generated in the annealed sample. An annealed sample was de￾formed 15% by cold rolling and tested under the same conditions. The tensile curve for this test is plotted as curve 3 in Fig. 3. It is seen that the stress-strain behavior of this redeformed sample returns to that of the initial ARB sample (Fig. 1), as do the values for the measured yield stress (256 MPa), UTS (333 MPa), total elongation (6.6%), and uniform elongation (2.0%). The similarity in mechanical behavior between this 15% cold-rolled sample and the original ARB sample suggests that applying 15% cold rolling has modified the structure in the annealed sam￾ple to be similar to the structure in the initial ARB sample. TEM characterization shows that a large number of dislocations are indeed intro￾duced again in the volume between the bound￾aries (Fig. 4) and at triple junctions and grain boundary regions. As a result, the dislocation configuration is very similar to that observed in the original ARB sample (Fig. 2A). We ob￾tained an interior dislocation density of 1.14
1014 mj2 and boundary spacings of 200 nm (DLB) and 650 nm (DICB) in this sample. These parameters are close to the values measured for the original ARB sample (Table 1). To further verify this hypothesis, we car￾ried out further annealing and deformation experiments as well as tensile tests. Repeated hardening and decrease in the elongation by low￾temperature annealing, and softening and in￾Fig. 3. Engineering stress-strain curves for 99.2% pure Al. Curve 2: ARB annealed at 150-C for 30 min (the same curve 2 as shown in Fig. 1). Curve 3: same as 2 but deformed 15% by cold rolling. Curve 4: same as 3 but again annealed at 150-C for 30 min. Curve 5: same as 4 but deformed 15% by cold rolling. Refer to table S2 for sample numbering. Fig. 2. TEM images showing the lamellar struc￾tural morphology and dislocation configuration in two ARB samples, (A) before annealing and (B) after annealing at 150-C for 30 min. The micro￾graphs were recorded when at least one lamella (M) in the area was under multiple-beam con￾dition with the beam direction parallel to the [001] zone axis of the lamella to reveal the interior dislocations. In (A), dislocation tangles are seen within the lamella marked M. In (B), dis￾location tangles are replaced by fewer disloca￾tions, many of which are pinned by the lamellar boundaries. Fig. 1. Engineering stress-strain curves for 99.2% pure Al. Curve 1: processed by six ARB cycles to an equivalent strain of 4.8. Curve 2: same material as 1, plus annealing at 150-C for 30 min. The strain rate e˙ used for the tensile test is indicated. Refer to table S2 for sample numbering. REPORTS 250 14 APRIL 2006 VOL 312 SCIENCE www.sciencemag.org

REPORTS 400 2.G.E.Fougere,)R.Weertman,R.W.Siegel,S.Kim, Scr.Metall.Moter.26,1879 (1992). 3.)R.Weertman,P.G.Sanders,Solid Stote Phenom.35-36, 300 249(1994). 4.]R.Weertman,Mater.Sci.Eng.A 166,161 (1993). 5.Y.M.Wang et ol.,Scr.Mater.51,1023 (2004). 200 6.F.Ebrahimi,Q.Zhai,D.Kong,Scr.Moter.39,315 (1998) 7.R.Z.Valiev,F.Chmelik,F.Bordeaux,G.Kapelski. e=4.17x10s1 B.Baudelet,Scr.Metoll.Mater.27,855 (1992). 8.]Languillaume et aL,Acta Metoll Moter.41,2953(1993) 100 9.N.Kamikawa,thesis,Osaka University (2005). 10.]R.Bowen,P.B.Prangnell,D.Juul Jensen,N.Hansen, Mater..Sci.Eng..A387-389,235(2004). 2 4 6 8 10 11.A Hasnaoui,H.V.Swygenhoven,P.M.Derlet,Acta Mater. Engineering Strain(%) 50,3927(2002). 12.N.Tsuji,Y.Saito,S.H.Lee,Y.Minamino,Adv.Eng.Moter. Fig.5.Engineering stress-strain curves for 99.2% 5,338(2003). 200nm pure AL.Curve 1:same as curve 1 in Fig.1.Curve 13.See supporting material on Science Online. RD 14.D.A.Hughes,N.Hansen,Acta Moter.48,2985 (2000). 6:same as 1 but deformed 15%by cold rolling. Refer to table S2 for sample numbering. 15.Q.Liu,X Huang,D.]Lloyd,N.Hansen,Acto Mater.50, Fig.4.TEM image showing the lamellar structural 3789(2002). 16.E.O.Hall,Proc.Phys.Soc.London B64,747 (1951). morphology and dislocation configuration in the processed by rolling to a large strain per pass and 17.N.)Petch,J.Iron Steef Inst.London 174,25 (1953). ARB sample processed by annealing at 150C for 30 18.]R.Greer,C.O.Warren,W.D.Nix,Acta Mater.53,1821 min,then deformed 15%by cold rolling.A some adiabatic heating may have taken place (2005). dislocation structure similar to that in the original (i.e.,the material may be in a recovered state) 19.]Schiotz,K.W.Jacobsen,Science 301,1357 (2003). ARB sample(Fig.2A)is introduced in the lamellae. (20).Such conditions are also typical of 20.N.Tsuji et aL,Mater.Sci.Eng.A 350,108 (2003). industrial processing.In accordance with the 21.B.L Li,A.Godfrey,Q.C.Meng.Q.Liu,N.Hansen,Acto Mater.52,1069(2004). present hypothesis,it is assumed that a light 22.Supported by the Danish National Research Foundation crease in the elongation by a low level of deformation of an ARB sample in the as- through the Center for Fundamental Research:Metal deformation,are obtained,as shown by curves delivered state may induce a small decrease in Structures in Four Dimensions,within which this work 4 and 5 in Fig.3.This repeated mechanical strength followed by an increase in ductility. was performed,and by the 21st Century COE Program behavior,combined with the structural charac- Curves 1 and 6 in Fig.5 confimm this assumption (the Center of Excellence for Advanced Structural and Functional Materials Design)at Osaka University through terization,confirms that the removal of disloca- The present investigation has focused on MEXT Japan.We thank D.Juul Jensen,B.Ralph,and ]A tions by annealing and their introduction by aluminum.The strategy described above may Wert for critical reading of the manuscript and helpful slight deformation are the cause of the changes also apply to metals such as nickel and intersti- discussions;E.Johnson for help with HRTEM;and in the mechanical properties.The deformation tial free steels that develop deformation micro- N.Kamikawa for preparing the samples used in this study. induced relatively small decreases in yield stress structure similar to that of aluminum (/4,2/) and UTS,and a large increase in the elongation Therefore,this strategy opens up a research area Supporting Online Material www.sciencemag.org/cgi/content/full312/5771/249/DC1 greatly improves the applicability of the material. of both fundamental and applied importance. Materials and Methods A further test of the beneficial effect of Tables S1 and S2 deformation as a final processing step is to References References and Notes deform the initial ARB sample 15%by cold 1.Y.Wang,M.W.Chen,F.H.Zhou,E.Ma,Nature 419,912 23 December 2005;accepted 16 March2006 rolling.The reason is that this sample has been (2002). 10.1126/science.1124268 Diels-Alder in Aqueous Molecular have made substantial progress in building mo- lecular hosts that emulate these enzymatic Hosts:Unusual Regioselectivity and pockets (3,4).Self-assembly of carefully con- structed organic and/or metallic building blocks in solution produces hollow host structures that Efficient Catalysis can bind small molecule guests (5,6).Among the many potential advantages of this strategy is Michito Yoshizawa,Masazumi Tamura,Makoto Fujita* the creation of hydrophobic reaction environ- ments in aqueous solution,widening the scope Self-assembled,hollow molecular structures are appealing as synthetic hosts for mediating of accessible reactivity in ecologically friendly chemical reactions.However,product binding has inhibited catalytic turnover in such systems,and media.However,these synthetic hosts have rarely selectivity has rarely approached the levels observed in more structurally elaborate natural conferred the orientational precision necessary to enzymes.We found that an aqueous organopalladium cage induces highly unusual regioselectivity guide reactions along otherwise unfavored path- in the Diels-Alder coupling of anthracene and phthalimide guests,promoting reaction at a terminal ways.Moreover,catalytic tumover has been in- rather than central anthracene ring.Moreover,a similar bowl-shaped host attains efficient catalytic hibited because the hosts bind products as turnover in coupling the same substrates(although with the conventional regiochemistry),most effectively as reactants,if not more so.In earlier likely because the product geometry inhibits the aromatic stacking interactions that attract the reports by Rebek (7,8),Sanders (9),and our planar reagents to the host. Department of Applied Chemistry,School of Engineering ffective synthetic homogeneous cata- more complex and derive much of their selectiv- University of Tokyo,and Core Research for Evolutional lysts have generally been structurally ity by bonding substrates through multiple in- Science and Technology (CREST),Japan Science and Technology Agency (ST),7-3-1 Hongo,Bunkyo-ku,Tokyo simple small molecules,which act by teractions in elaborate pockets,thereby forcing 113-8656,]apan. binding to substrates at or near the reaction the substrates into orientations that favor specific *To whom correspondence should be addressed.E-mail: site.In contrast,enzymes are much larger and reaction paths(1,2).In the past decade,chemists mfujita@appchem.t.u-tokyo.ac.jp www.sciencemag.org SCIENCE VOL 312 14 APRIL 2006 251

crease in the elongation by a low level of deformation, are obtained, as shown by curves 4 and 5 in Fig. 3. This repeated mechanical behavior, combined with the structural charac￾terization, confirms that the removal of disloca￾tions by annealing and their introduction by slight deformation are the cause of the changes in the mechanical properties. The deformation induced relatively small decreases in yield stress and UTS, and a large increase in the elongation greatly improves the applicability of the material. A further test of the beneficial effect of deformation as a final processing step is to deform the initial ARB sample 15% by cold rolling. The reason is that this sample has been processed by rolling to a large strain per pass and some adiabatic heating may have taken place (i.e., the material may be in a recovered state) (20). Such conditions are also typical of industrial processing. In accordance with the present hypothesis, it is assumed that a light deformation of an ARB sample in the as￾delivered state may induce a small decrease in strength followed by an increase in ductility. Curves 1 and 6 in Fig. 5 confirm this assumption. The present investigation has focused on aluminum. The strategy described above may also apply to metals such as nickel and intersti￾tial free steels that develop deformation micro￾structure similar to that of aluminum (14, 21). Therefore, this strategy opens up a research area of both fundamental and applied importance. References and Notes 1. Y. Wang, M. W. Chen, F. H. Zhou, E. Ma, Nature 419, 912 (2002). 2. G. E. Fougere, J. R. Weertman, R. W. Siegel, S. Kim, Scr. Metall. Mater. 26, 1879 (1992). 3. J. R. Weertman, P. G. Sanders, Solid State Phenom. 35–36, 249 (1994). 4. J. R. Weertman, Mater. Sci. Eng. A 166, 161 (1993). 5. Y. M. Wang et al., Scr. Mater. 51, 1023 (2004). 6. F. Ebrahimi, Q. Zhai, D. Kong, Scr. Mater. 39, 315 (1998). 7. R. Z. Valiev, F. Chmelik, F. Bordeaux, G. Kapelski, B. Baudelet, Scr. Metall. Mater. 27, 855 (1992). 8. J. Languillaume et al., Acta Metall. Mater. 41, 2953 (1993). 9. N. Kamikawa, thesis, Osaka University (2005). 10. J. R. Bowen, P. B. Prangnell, D. Juul Jensen, N. Hansen, Mater. Sci. Eng. A 387–389, 235 (2004). 11. A. Hasnaoui, H. V. Swygenhoven, P. M. Derlet, Acta Mater. 50, 3927 (2002). 12. N. Tsuji, Y. Saito, S. H. Lee, Y. Minamino, Adv. Eng. Mater. 5, 338 (2003). 13. See supporting material on Science Online. 14. D. A. Hughes, N. Hansen, Acta Mater. 48, 2985 (2000). 15. Q. Liu, X. Huang, D. J. Lloyd, N. Hansen, Acta Mater. 50, 3789 (2002). 16. E. O. Hall, Proc. Phys. Soc. London B64, 747 (1951). 17. N. J. Petch, J. Iron Steel Inst. London 174, 25 (1953). 18. J. R. Greer, C. O. Warren, W. D. Nix, Acta Mater. 53, 1821 (2005). 19. J. Schiøtz, K. W. Jacobsen, Science 301, 1357 (2003). 20. N. Tsuji et al., Mater. Sci. Eng. A 350, 108 (2003). 21. B. L. Li, A. Godfrey, Q. C. Meng, Q. Liu, N. Hansen, Acta Mater. 52, 1069 (2004). 22. Supported by the Danish National Research Foundation through the Center for Fundamental Research: Metal Structures in Four Dimensions, within which this work was performed, and by the 21st Century COE Program (the Center of Excellence for Advanced Structural and Functional Materials Design) at Osaka University through MEXT Japan. We thank D. Juul Jensen, B. Ralph, and J. A. Wert for critical reading of the manuscript and helpful discussions; E. Johnson for help with HRTEM; and N. Kamikawa for preparing the samples used in this study. Supporting Online Material www.sciencemag.org/cgi/content/full/312/5771/249/DC1 Materials and Methods Tables S1 and S2 References 23 December 2005; accepted 16 March 2006 10.1126/science.1124268 Diels-Alder in Aqueous Molecular Hosts: Unusual Regioselectivity and Efficient Catalysis Michito Yoshizawa, Masazumi Tamura, Makoto Fujita* Self-assembled, hollow molecular structures are appealing as synthetic hosts for mediating chemical reactions. However, product binding has inhibited catalytic turnover in such systems, and selectivity has rarely approached the levels observed in more structurally elaborate natural enzymes. We found that an aqueous organopalladium cage induces highly unusual regioselectivity in the Diels-Alder coupling of anthracene and phthalimide guests, promoting reaction at a terminal rather than central anthracene ring. Moreover, a similar bowl-shaped host attains efficient catalytic turnover in coupling the same substrates (although with the conventional regiochemistry), most likely because the product geometry inhibits the aromatic stacking interactions that attract the planar reagents to the host. Effective synthetic homogeneous cata￾lysts have generally been structurally simple small molecules, which act by binding to substrates at or near the reaction site. In contrast, enzymes are much larger and more complex and derive much of their selectiv￾ity by bonding substrates through multiple in￾teractions in elaborate pockets, thereby forcing the substrates into orientations that favor specific reaction paths (1, 2). In the past decade, chemists have made substantial progress in building mo￾lecular hosts that emulate these enzymatic pockets (3, 4). Self-assembly of carefully con￾structed organic and/or metallic building blocks in solution produces hollow host structures that can bind small molecule guests (5, 6). Among the many potential advantages of this strategy is the creation of hydrophobic reaction environ￾ments in aqueous solution, widening the scope of accessible reactivity in ecologically friendly media. However, these synthetic hosts have rarely conferred the orientational precision necessary to guide reactions along otherwise unfavored path￾ways. Moreover, catalytic turnover has been in￾hibited because the hosts bind products as effectively as reactants, if not more so. In earlier reports by Rebek (7, 8), Sanders (9), and our Fig. 4. TEM image showing the lamellar structural morphology and dislocation configuration in the ARB sample processed by annealing at 150-C for 30 min, then deformed 15% by cold rolling. A dislocation structure similar to that in the original ARB sample (Fig. 2A) is introduced in the lamellae. Fig. 5. Engineering stress-strain curves for 99.2% pure Al. Curve 1: same as curve 1 in Fig. 1. Curve 6: same as 1 but deformed 15% by cold rolling. Refer to table S2 for sample numbering. Department of Applied Chemistry, School of Engineering, University of Tokyo, and Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency (JST), 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan. *To whom correspondence should be addressed. E-mail: mfujita@appchem.t.u-tokyo.ac.jp REPORTS www.sciencemag.org SCIENCE VOL 312 14 APRIL 2006 251