NEWS VIEWS NANOSTRUCTURED METALS ess is more Annealing out dislocations in deformed metals usually leads to reduced strength and increased ductility.Exactly the opposite has been observed in bulk nanostructured aluminium. E.MA',T.D.SHEN2AND X.L.WUP a Figure 1 High-resolution are at 'Department of Materials Science and Engineering, transmission electron Johns Hopkins University,Baltimore,Maryland 21218,USA; micrographs showing Materials Science and Technology Division,Los Alamos dislocations in a National Laboratory,Los Alamos,New Mexico 87545,USA nanostructured aluminium State Key Laboratory of Nonlinear Mechanics,Institute of grain(120 nm in diameter) Mechanics,Chinese Academy of Sciences, produced by a large Beijing 100080,China deformation.a,Contrast due email:ema@jhu.edu to the presence of strain caused mainly by dislocations b,High-magnification view, etals deform under applied stresses because of with dislocations marked (T) the movement of dislocations-line defects b Note the circled dislocation in the crystal lattice.Many dislocations are dipoles.c.Fourier-filtered produced during plastic flow.The accumulation image of another grain.The and interaction of these crowded dislocations ellipses show examples of creates obstacles that make the propagation of dislocation dipoles. dislocations difficult.Further deformation therefore requires higher stresses,leading to work hardening. Conversely,annealing,a heat treatment often given to deformed metals to rearrange dislocations and relieve stresses,reduces strength and improves ductility as the dislocation roadblocks annihilate.This well- established picture for conventional metals apparently no longer holds for nanostructured aluminium,as reported recently by Huang et al.'.In fact,what they found was just the opposite,"hardening by annealing and softening by deformation". This result is intriguing because Huang et al. worked with 99.99%pure aluminium that presumably contained no alloying element.The bulk metal remains polycrystalline after processing,except that the crystallites inside are nanostructured by a severe plastic deformation process that has created a high density of grain boundaries(GBs)and dislocations After annealing,the density of dislocations in the grain interior decreased by 60%.This alone would reduce the strength by about 37%,on the conventional assumption that the strength scales with the square in nanostructured titanium.At even smaller grain root of the dislocation density.Yet the removal of sizes,electroplated nickel exhibited elevated strength dislocations actually rendered the metal 10%stronger without losing ductility after moderate annealing, rather than weaker'. and ball-milled nanocrystalline iron-based and Before discussing our views regarding the nickel-based alloys showed no decrease in hardness mechanisms that could make this happen,two after 90%of the total dislocations were removed questions immediately come to mind.First,does this during annealing(T.D.S.and S.H.Feng,unpublished happen in other bulk nanostructured metals?The observations).In all cases,however,there remains answer is probably yes.In addition to the examples the possibility that the segregation of small amounts cited by Huang and colleagues,low-temperature of impurities to the dislocations and GBs,much as in annealing also led to a 30%increase in the vield stress the well-known strain ageing scenario',contributed nature materials VOL 5|JULY 2006|www.nature.com/naturematerials 515 @2006 Nature Publishing Group
NEWS & VIEWS nature materials | VOL 5 | JULY 2006 | www.nature.com/naturematerials 515 E. MA1 , T. D. SHEN2 AND X. L. WU3 are at 1 Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, Maryland 21218, USA; 2 Materials Science and Technology Division, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA; 3 State Key Laboratory of Nonlinear Mechanics, Institute of Mechanics, Chinese Academy of Sciences, Beijing 100080, China email: ema@jhu.edu Metals deform under applied stresses because of the movement of dislocations — line defects in the crystal lattice. Many dislocations are produced during plastic fl ow. Th e accumulation and interaction of these crowded dislocations creates obstacles that make the propagation of dislocations diffi cult. Further deformation therefore requires higher stresses, leading to work hardening. Conversely, annealing, a heat treatment oft en given to deformed metals to rearrange dislocations and relieve stresses, reduces strength and improves ductility as the dislocation roadblocks annihilate. Th is wellestablished picture for conventional metals apparently no longer holds for nanostructured aluminium, as reported recently by Huang et al. 1 . In fact, what they found was just the opposite, “hardening by annealing and soft ening by deformation”. Th is result is intriguing because Huang et al. worked with 99.99% pure aluminium that presumably contained no alloying element. Th e bulk metal remains polycrystalline aft er processing, except that the crystallites inside are nanostructured by a severe plastic deformation process that has created a high density of grain boundaries (GBs) and dislocations. Aft er annealing, the density of dislocations in the grain interior decreased by 60%. Th is alone would reduce the strength by about 37%, on the conventional assumption that the strength scales with the square root of the dislocation density. Yet the removal of dislocations actually rendered the metal 10% stronger rather than weaker1 . Before discussing our views regarding the mechanisms that could make this happen, two questions immediately come to mind. First, does this happen in other bulk nanostructured metals? Th e answer is probably yes. In addition to the examples cited by Huang and colleagues, low-temperature annealing also led to a 30% increase in the yield stress in nanostructured titanium2 . At even smaller grain sizes, electroplated nickel exhibited elevated strength without losing ductility aft er moderate annealing3 , and ball-milled nanocrystalline iron-based and nickel-based alloys showed no decrease in hardness aft er 90% of the total dislocations were removed during annealing (T.D.S. and S.H. Feng, unpublished observations). In all cases, however, there remains the possibility that the segregation of small amounts of impurities to the dislocations and GBs, much as in the well-known strain ageing scenario4 , contributed Annealing out dislocations in deformed metals usually leads to reduced strength and increased ductility. Exactly the opposite has been observed in bulk nanostructured aluminium. NANOSTRUCTURED METALS Less is more 4 nm 1 nm 4 nm a b c Figure 1 High-resolution transmission electron micrographs showing dislocations in a nanostructured aluminium grain (120 nm in diameter) produced by a large deformation. a, Contrast due to the presence of strain caused mainly by dislocations. b, High-magnifi cation view, with dislocations marked (T). Note the circled dislocation dipoles. c, Fourier-fi ltered image of another grain. The ellipses show examples of dislocation dipoles. nmatnv0706-print.indd 515 matnv0706-print.indd 515 13/6/06 4:00:31 pm 3/6/06 4:00:31 pm ©2006 NaturePublishingGroup
NEWS VIEWS significantly to the hardening observed,even though for large grains,annealing readily sweeps the small the impurity level might have been below the grain volume clean.This dislocation exhaustion leaves detection limit'. the nucleation of dislocations at GBs as the main The second question is whether the removal of supply of mobile dislocations,which requires higher dislocations leading to higher strength has ever been stresses.In fact,the annealed GBs are more relaxed observed in metals such as aluminium.The answer and are also less likely to emit dislocations2. is yes,but only in carefully prepared single crystals Thus,in nanograins the strength is sensitive with dimensions on the micrometre scale.Obviously, to the atomic processes at the GBs,which are also the probability of finding defects is low when where impurities tend to segregate to,especially on the sample volume is small,such that defect-free annealing'.Even when the metal incorporates only a whiskers can exhibit an extremely high yield strength trace amount of impurities during its processing,the approaching the theoretical limits.Furthermore, segregation to the GBs could be sufficient to suppress dislocations existing in a small volume can all run out the dislocations from emerging,or to pin them down of the sample body on straining,causing dislocation as they propagate.The dislocation-GB interactions starvation,so that continued deformation demands are known to be thermally activated processess.. higher stresses to nucleate new dislocations"? One can view the high strength induced by These cases would be more in line with an intuitive nanostructuring as a result of severe constraints on expectation:the fewer the defects in a material,the dislocation activities.In this context the moderate higher its strength. annealing or brief deformation serves to exhaust But bulk engineering metals do not behave in the available dislocations further.To facilitate plastic this way,because many dislocations and their sources deformation,the idea advocated by Huang et al.is to are inevitably present inside the polycrystalline impose an intra-grain dislocation structure through grains.Without the possibility to rid the material of externally forced large deformation. dislocations completely,the practical strengthening Although it is clear that fewer dislocations strategy is rather to put in more dislocations so that do not necessarily mean softening and could in they get in each other's way. some cases even provide strengthening,a full Now,what is the difference if these polycrystals understanding of hardening by annealing and have tiny grains?The lattice dislocations no longer softening by deformation in nanostructured matter as much,because the strength becomes metals requires more in-depth studies.The dominated largely by how dislocations originate range of possible observations and mechanisms from and interact with GBs,as revealed by previous is fascinating and provides plenty of new computer simulations*and experiments.The stored opportunities for future research. dislocations are closely spaced,shown in Fig.1,and often take the form of dislocation dipoles'-a pair REFERENCES of dislocations close by but with opposite signs.A 1.Huang.X..Hansen,N.Tsuji.N.Science 312,249-251 (2006). 2.Valiev,R.Z..Sergueeva.A.V.Mukherjee,A.K.Ser.Mater.49,669-674 (2003). dislocation dipole generates only short-range forces 3.Wang.Y.M.et al.Ser.Mater.51,1023-1028 (2004). because of the screening of their stress fields".This is 4.Reed-Hill,R.E.&Abbaschian,R.Plrysical Metullurgy Principles 3rd edn. likely to be true for the dislocations stored inside the 288-294(PWS Publishing,Boston,1994). 5.Kellv.A St tangles in the nanostructured grains in ref.1.Thus, g Solids 2nd edn (Clarendon,Oxford,1973). 6.Uchic,M.D..Dimiduk,D.M.,Florando.I.N.Nix,W.D.Science 305, many of the dislocations contribute little to the long- 986-9892001}. range internal stress field and the overall strength.In 7.Greer,J.R,Oliver,W.C.Nix,W.D.Acta Mater.53,1821-1830 (2005). fact,their movement is easier under applied stresses 8.Van Swygenhoven,H..Detlet.P.M.Froseth,A.G.Acta Mater.54, 1975-1983(2006. than the generation of new dislocations from GBs". 9.Wang.Y.M.Hamza,A.V.Ma.E.Appl Phys.Lett 86,241917 (2005). On annealing,therefore,the elimination of the 10.Laird,C.in Plrysical Metallurgy Vol.3 (eds Cahn.R.W.Haasen,P.)2304- dislocations renders the material less prone to plastic 2313(Elsevier.Amsterdam,1996). flow.Indeed,annealing drives the dislocations to 11.Chen,H.S.Gilman,I.I.Head,A.K.L AppL Phrys.35,2502-2514 (1964). disappear into the abundant GB sinks near by',and to 12.Hasnaoui,A.Van Swygenhoven.H.Derlet.P.M.Acta Mater.50,3927 3939(2002). annihilate through recombination and climb.Unlike 13.Wang.Y.M.Ma,E.Appl Phrys Lett.85,2750-2752(2004). 516 nature materials|VOL 5|JULY 2006 www.nature.com/naturematerials @2006 Nature Publishing Group
NEWS & VIEWS 516 nature materials | VOL 5 | JULY 2006 | www.nature.com/naturematerials signifi cantly to the hardening observed, even though the impurity level might have been below the detection limit3 . Th e second question is whether the removal of dislocations leading to higher strength has ever been observed in metals such as aluminium. Th e answer is yes, but only in carefully prepared single crystals with dimensions on the micrometre scale. Obviously, the probability of fi nding defects is low when the sample volume is small, such that defect-free whiskers can exhibit an extremely high yield strength approaching the theoretical limit5 . Furthermore, dislocations existing in a small volume can all run out of the sample body on straining, causing dislocation starvation, so that continued deformation demands higher stresses to nucleate new dislocations6,7. Th ese cases would be more in line with an intuitive expectation: the fewer the defects in a material, the higher its strength. But bulk engineering metals do not behave in this way, because many dislocations and their sources are inevitably present inside the polycrystalline grains. Without the possibility to rid the material of dislocations completely, the practical strengthening strategy is rather to put in more dislocations so that they get in each other’s way. Now, what is the diff erence if these polycrystals have tiny grains? Th e lattice dislocations no longer matter as much, because the strength becomes dominated largely by how dislocations originate from and interact with GBs, as revealed by previous computer simulations8 and experiments9 . Th e stored dislocations are closely spaced, shown in Fig. 1, and oft en take the form of dislocation dipoles10 — a pair of dislocations close by but with opposite signs. A dislocation dipole generates only short-range forces because of the screening of their stress fi elds11. Th is is likely to be true for the dislocations stored inside the tangles in the nanostructured grains in ref. 1. Th us, many of the dislocations contribute little to the longrange internal stress fi eld and the overall strength. In fact, their movement is easier under applied stresses than the generation of new dislocations from GBs8 . On annealing, therefore, the elimination of the dislocations renders the material less prone to plastic fl ow. Indeed, annealing drives the dislocations to disappear into the abundant GB sinks near by1 , and to annihilate through recombination and climb. Unlike for large grains, annealing readily sweeps the small grain volume clean. Th is dislocation exhaustion leaves the nucleation of dislocations at GBs as the main supply of mobile dislocations, which requires higher stresses. In fact, the annealed GBs are more relaxed and are also less likely to emit dislocations3,12. Th us, in nanograins the strength is sensitive to the atomic processes at the GBs, which are also where impurities tend to segregate to, especially on annealing3 . Even when the metal incorporates only a trace amount of impurities during its processing, the segregation to the GBs could be suffi cient to suppress the dislocations from emerging, or to pin them down as they propagate3,8. Th e dislocation–GB interactions are known to be thermally activated processes8,9,13. One can view the high strength induced by nanostructuring as a result of severe constraints on dislocation activities. In this context the moderate annealing or brief deformation9 serves to exhaust the available dislocations further. To facilitate plastic deformation, the idea advocated by Huang et al. is to impose an intra-grain dislocation structure through externally forced large deformation. Although it is clear that fewer dislocations do not necessarily mean soft ening and could in some cases even provide strengthening, a full understanding of hardening by annealing and soft ening by deformation in nanostructured metals requires more in-depth studies. Th e range of possible observations and mechanisms is fascinating and provides plenty of new opportunities for future research. REFERENCES 1. Huang, X., Hansen, N. & Tsuji, N. Science 312, 249–251 (2006). 2. Valiev, R. Z., Sergueeva, A. V. & Mukherjee, A. K. Scr. Mater. 49, 669–674 (2003). 3. Wang, Y. M. et al. Scr. Mater. 51, 1023–1028 (2004). 4. Reed-Hill, R. E. & Abbaschian, R. Physical Metallurgy Principles 3rd edn, 288–294 (PWS Publishing, Boston, 1994). 5. Kelly, A. Strong Solids 2nd edn (Clarendon, Oxford, 1973). 6. Uchic, M. D., Dimiduk, D. M., Florando, J. N. & Nix, W. D. Science 305, 986–989 (2004). 7. Greer, J. R., Oliver, W. C. & Nix, W. D. Acta Mater. 53, 1821–1830 (2005). 8. Van Swygenhoven, H., Derlet, P. M. & Frøseth, A. G. Acta Mater. 54, 1975–1983 (2006). 9. Wang, Y. M., Hamza, A. V. & Ma, E. Appl. Phys. Lett. 86, 241917 (2005). 10. Laird, C. in Physical Metallurgy Vol. 3 (eds Cahn, R. W. & Haasen, P.) 2304– 2313 (Elsevier, Amsterdam, 1996). 11. Chen, H. S., Gilman, J. J. & Head, A. K. J. Appl. Phys. 35, 2502–2514 (1964). 12. Hasnaoui, A., Van Swygenhoven, H. & Derlet, P. M. Acta Mater. 50, 3927– 3939 (2002). 13. Wang, Y. M. & Ma, E. Appl. Phys. Lett. 85, 2750–2752 (2004). nmatnv0706-print.indd 516 matnv0706-print.indd 516 13/6/06 4:00:33 pm 3/6/06 4:00:33 pm ©2006 NaturePublishingGroup