LETTER pubs.acs.org/NanoLett
Influence of Homogeneous Interfaces on the Strength of 500 nm Diameter Cu Nanopillars Dongchan Jang,*,† Can Cai,‡ and Julia R. Greer† †
Division of Engineering and Applied Science and ‡Division of Chemistry and Chemical Engineering, California Institute of Technology, 1200 East California Boulevard, Passadena, California 91125, United States ABSTRACT: Interfaces play an important role in crystalline plasticity as they affect strength and often serve as obstacles to dislocation motion. Here we investigate effects of grain and nanotwin boundaries on uniaxial strength of 500 nm diameter Cu nanopillars fabricated by e-beam lithography and electroplating. Uniaxial compression experiments reveal that strength is lowered by introducing grain boundaries and significantly rises when twin boundaries are present. Weakening is likely due to the activation of grain-boundary-mediated processes, while impeding dislocation glide can be responsible for strengthening by twin boundaries. KEYWORDS: Nanopillars, nanocrystalline, nanotwins, uniaxial compression, nanomechanics
U
nderstanding the influence of interfaces like grain and interphase boundaries on crystalline strength and mechanical properties is of paramount importance as they often serve as obstacles for dislocation motion, as well as dislocation sources and sinks. In conventional polycrystalline metals, for example, grain boundaries (GB) obscure the passage of the gliding dislocations, causing them to pile up against the boundaries,1 which leads to strengthening via a well-known Hall-Petch mechanism.2,3 These GB-dislocations interactions result in the scaling law, where the yield strength is inversely proportional to the square root of the grain size.2,3 However, when the grain size falls below ∼30 nm (nanocrystalline metals), the role of GBs shifts toward dislocation nucleation and absorption sites,4,5 and sometimes to the deformation path itself, such as GB sliding6-8 and Coble creep.9-11 This transition is manifested by a different scaling law, often referred to as “inverse Hall-Petch relation,” where the yield strength becomes lower with decreasing grain size (for a review on nanocrystalline materials, see ref 11). Strength-defining microstructural feature size, however, is not the only factor that may affect the deformation mechanisms in nanomaterials. It has now been ubiquitously demonstrated that the specimen size at the micrometer and submicrometer scales fundamentally changes the deformation processes, leading to size-dependent strength12-14 (for a review on this topic, see ref 15. For example, when reduced to the submicrometer scale, single-crystalline face-centered cubic (fcc) metals exhibit a “smaller is stronger” phenomenon12,13,16 due to the dislocation nucleation-governed plasticity with dislocations nucleating either from single-arm sources17 or from the surface sources,18 as opposed to through dislocation multiplication processes via the operation of Frank-Read sources.19 While significant effort has been dedicated to studying the mechanical behavior of single-crystalline metals at nanoscale, the effect of internal interfaces contained within nanometer-sized r 2011 American Chemical Society
samples on strength and mechanical deformation has not been studied to the same extent. Yet the combined effects of the characteristic microstructural dimensions and the external sample dimensions on strength are likely to interplay with one another and need to be thoroughly understood.15,20-22 Furthermore, most of studies have focused on the effect of the sample dimensions while maintaining the internal feature sizes constant, but no study has reported the influence of different interfaces with fixed specimen sizes. In this report, we aim at understanding the effect of homogeneous internal interfaces on the compressive strength of 500 nm diameter Cu nanopillars containing twins, grains, and a combination of both boundaries. We fabricated a large array of cylindrical Cu nanopillars with 500 nm diameter by electroplating Cu into the patterned polymethylmethacrylate (PMMA) templates.23 Precise control of the electroplating conditions enabled us to create the nanopillars with a rich variety of microstructures containing different interfaces: grain boundaries and coherent nanotwin boundaries (TBs). We then investigated the influence of these interfaces on the uniaxial compressive strengths in these 500 nm nanopillars, thereby shedding light onto a different type of “size effects”, where the submicrometer-sized external dimensions remain constant while microstructural size, twin boundary spacing and grain size, are varied. The 500 nm diameter Cu nanopillars were electroplated into a patterned PMMA template on top of a ∼100 nm seed layer of Au on Si substrate, a process described in detail in ref 23. The microstructure was controlled by varying the waveform, which has periodic rectangular signals, with the current density of Jon during the on-time (ton) and reduced to 0 A/cm2 for off-time, toff, as shown schematically in Figure 1A. The compositions of the Received: January 26, 2011 Revised: February 18, 2011 Published: March 09, 2011 1743
dx.doi.org/10.1021/nl2003076 | Nano Lett. 2011, 11, 1743–1746
Nano Letters
LETTER
electrolyte, Jon, ton, and toff used to create each microstructure are listed in Table 1. Scanning electron microscope (SEM) image of a typical as-fabricated pillar is displayed in Figure 1. To identify the internal microstructure, the nanopillars were examined using transmission electron microscopy (TEM) at 200 kV acceleration voltage (TF20, FEI Co.). The TEM specimens were prepared by lifting out underlying micrometer-sized Si substrate lamellae with the pillar of interest on top, using a nanomanipulator (AutoProbe 200, Omniprobe, Inc.) and transferring them onto Cu TEM grids. The pillar samples were then thinned to the electron-transparent thickness (∼100 nm) by using the focused ion beam (FIB) at the final condition of 30 kV and 10 pA. The uniaxial compression tests were carried out in the nanoindenter (G200, Agilent Technologies) equipped with custom-made flat punch tip at the nominal strain rate of 1 10-3 s-1. More than 10 nanopillars were tested for each microstructure. The effect of the nonrigid substrate was corrected using the methodology developed by Greer et al.13 Figure 2 shows bright- and dark-field TEM images, as well as diffraction patterns for each microstructure. The nanopillar in Figure 2A is pure nanocrystalline (NC) with the grain size of 160 nm. The sample shown in Figure 2B is nearly pure nanotwinned (NT), where the entire pillar consists of vertically stacked twin lamellae spaced at ∼8.8 nm with the exception of a few tiny grains on the right side. Crystallographic analysis using the electron diffraction pattern reveals that all twin lamellae share the same {111} planes aligned along the pillar axis direction. Finally, the nanopillar in Figure 2C is nanocrystalline nanotwinned (NCNT), where ∼6.7 nm-spaced nanotwins are embedded within ∼250 nmsized grains with random orientations. The lower right inset in Figure 2C is the zoomed-in image of region within the white square. For each TEM image, the corresponding electron diffraction pattern is shown in the inset at the top-right corner. Figure 3A-C shows several true stress-strain curves collected for each microstructure.
Figure 1. (A) The waveform of the pulsed electroplating current. (B) SEM image of a typical as-fabricated Cu nanopillar.
Interestingly, while the plastic flow of NC samples consists of many jerky strain bursts, the other two samples show a relatively smoother flow. One possible explanation can be that TBs may effectively impede the rapid deformation process. For example, TBs may act as barriers to dislocation glide for purely NT samples and slow down GB-mediated processes in NCNT samples.24 On the other hand, NT samples exhibit much more limited plasticity compared with the other two samples, as their failure is always accompanied by a large strain burst. The plot in Figure 3D shows the average yield strength of the pillars with each microstructure as well as that of 500 nmdiameter [111]-oriented single crystalline (SC) Cu nanopillars25 as a function of microstructure. Interestingly, there is a clear distinction of how these two different interfaces, GBs and TBs, appear to affect the strengths of Cu nanopillars. For example, the strength of the NC samples is lower than that of the SC samples by 20%, implying that the influence of the regular GBs is likely to lower the strength of the 500 nm nanopillars. On the other hand, the TBs seem to dramatically strengthen the nanopillars as the strengths of those nanopillars containing nanotwins are always stronger than their twin-less counterparts. Furthermore, the strength of NT samples is a factor of ∼3 higher than of all others, suggesting that the orientation of the interfaces may play an important role in strengthening mechanism. These findings may seem unusual, as the presence of GBs usually facilitates strength increase in conventional polycrystalline metals. The strengths of polycrystals are usually higher than those of singlecrystals mainly because (1) the compatibility condition at the GBs applies additional constraints to the deformation, and (2) the GBs obstruct dislocation motion.1 However, in nanocrystalline materials with grains smaller than some critical size (typically, on the order of 10-20 nm) it has been frequently reported that the strength decreases with the increasing volume fraction of GBs, that is, decreasing the grain size, due to the activation of GB-mediated plasticity.11 In this regime, the GBs may serve as a source of weakening due to the more energetically favored GB-mediated deformation mechanisms like GB sliding and grain rotation rather than dislocation-driven processes.8,11 Further, the authors’ previous work suggests that GB-mediated deformation processes can occur in the surface grains of nanocrystalline nanopillars even at much larger grain sizes than the critical grain size.22 The number of grains across the diameter of NC nanopillars in this study is ∼3, meaning that 2 out of 3 grains are surface grains. Therefore, it is conceivable that these GB-mediated deformation mechanisms contribute to the strength reduction in nanocrystalline nanopillars compared with the SC ones. As a result, two competing processes may emerge simultaneously upon the introduction of internal interfaces into these 500 nm nanopillars: Hall-Petch-like strengthening and GBmediated weakening. The strength of these nanopillars, therefore, is defined by the competition between these two processes.
Table 1. Electroplating Condition for the Fabrication of 500 nm Cu Nanopillars for Each Microstructure feature sizes
microstructure nanotwinned (NT) nanocrystalline (NC) nanocrystalline Nanotwinned (NCNT)
Jon
ton
toff
grain size
twin thickness
electrolyte composition
(A/cm2)
(ms)
(ms)
(nm)
(nm)
125 g/L of CuSO4 3 5H2O þ 50 g/L of H2SO4 100 g/L of C6H8O7 3 H2O þ 28 g/L of CuSO4 3 5H2O
0.4
1
100
n/a
8.8
1.2
1
100
160
n/a
0.72
1
100
250
6.7
þ 50 g/L of (NH4)2SO4 50 g/L of C6H8O7 3 H2O þ 28 g/L of CuSO4 3 5H2O þ 50 g/L of (NH4)2SO4
1744
dx.doi.org/10.1021/nl2003076 |Nano Lett. 2011, 11, 1743–1746
Nano Letters
LETTER
Figure 2. (A) TEM bright-field image and diffraction pattern (inset) of a typical nanocrystalline Cu nanopillar. TEM dark-field images and diffraction patterns (upper right insets) of (B) nanotwinned and (C) nanocrystalline nanotwinned Cu nanopillars. The lower right inset in (C) is the zoomed-up image of the region within the white box.
Figure 3. Selected true stress and strain curves for (A) NC, (B) NT, and (C) NCNT Cu nanopillars. (D) The average yield strength of each microstructure.
The results of this study suggest that the structure of the interfaces and their orientation to the loading axis are important factors when
considering the influence of interfaces on nanopillar strength. Given that the strength of NC pillars is ∼20% lower than that of SC ones, 1745
dx.doi.org/10.1021/nl2003076 |Nano Lett. 2011, 11, 1743–1746
Nano Letters
LETTER
authors also gratefully acknowledge critical support and infrastructure provided for this work by the Kavli Nanoscience Institute at Caltech.
’ REFERENCES
Figure 4. Illustration showing the effect of grain and nano twin boundaries on the strength of 500 nm Cu nanopillars.
there would be more weakening effect by the randomly oriented regular GBs. In contrast, strengthening appears to be a result of uniformly spaced nano-TBs as both NCNT and NT pillars are stronger than NC and SC, respectively. Additionally, the strength of NT samples is significantly higher than that of others by a factor of ∼3. The perpendicular orientation of twin boundaries with respect to the loading axes suppresses shear deformation parallel to the TB planes because there is no resolved shear stress in that direction. Therefore, the low-friction dislocation glide along the twin boundaries, a deformation mechanism recently revealed to be associated with nanotwinned microstructure by,26 is not accessible, resulting in a more pronounced strengthening. Strengthening by orthogonally oriented nano TBs is also consistent with the recent computational reports.27,28 A matrix displaying the influence of GBs and TBs on the strengths of 500 nm Cu nanopillars is summarized in Figure 4. It is important to note that here the attained mechanical strengths are a result of the collective effect of multiple interfaces simultaneously, rather than that of a well-defined single boundary. Therefore, individual GBs or TBs, may strengthen or weaken the sample depending on their structure and orientation. In summary, we fabricated 500 nm diameter Cu nanopillars with a broad range of possible microstructures and measured their compressive strengths. Unlike bulk-scale polycrystalline metals, introducing GBs into metallic nanostructures appears to weaken their strength. On the other hand, periodic nanotwin boundaries significantly increase the strength of such nanostructures. These findings provide insight into the possibility of tuning both microstructural critical size and sample dimensions to elicit desired mechanical properties.
(1) Courtney, T. H. Mechanical behavior of materials; McGraw-Hill: New York, 2000; pp xviii, 733. (2) Hall, E. O. Proc. Phys. Soc., London, Sect. B 1951, 64 (381), 747–753. (3) Petch, N. J. J. Iron Steel Inst. 1953, 174 (1), 25–28. (4) Van Swygenhoven, H.; Derlet, P. M.; Hasnaoui, A. Phys. Rev. B 2002, 66, 024101. (5) Chen, M. W.; Ma, E.; Hemker, K. J.; Sheng, H. W.; Wang, Y. M.; Cheng, X. M. Science 2003, 300 (5623), 1275–1277. (6) Ke, M.; Hackney, S. A.; Milligan, W. W.; Aifantis, E. C. Nanostruct. Mater. 1995, 5 (6), 689–697. (7) Kumar, K. S.; Suresh, S.; Chisholm, M. F.; Horton, J. A.; Wang, P. Acta Mater. 2003, 51 (2), 387–405. (8) Shan, Z. W.; Stach, E. A.; Wiezorek, J. M. K.; Knapp, J. A.; Follstaedt, D. M.; Mao, S. X. Science 2004, 305 (5684), 654–657. (9) Wang, D. L.; Kong, Q. P.; Shui, J. P. Scr. Metall. Mater. 1994, 31 (1), 47–51. (10) Masumura, R. A.; Hazzledine, P. M.; Pande, C. S. Acta Mater. 1998, 46 (13), 4527–4534. (11) Meyers, M. A.; Mishra, A.; Benson, D. J. Prog. Mater. Sci. 2006, 51 (4), 427–556. (12) Uchic, M. D.; Dimiduk, D. M.; Florando, J. N.; Nix, W. D. Science 2004, 305 (5686), 986–989. (13) Greer, J. R.; Oliver, W. C.; Nix, W. D. Acta Mater. 2005, 53 (6), 1821–1830. (14) Jang, D.; Greer, J. R. Nat. Mater. 2010, 9 (3), 215–219. (15) Greer, J. R.; De Hosson, J. T. M. Prog. Mater. Sci.; 2011; (article in press), doi:10.1016/J.pmatsci.2011.01.005. (16) Volkert, C. A.; Lilleodden, E. T. Philos. Mag. 2006, 86, 5567– 5579. (17) Tang, H.; Schwarz, K. W.; Espinosa, H. D. Phys. Rev. Lett. 2008, 100 (18), No. 185503. (18) Zhu, T.; Li, J.; Samanta, A.; Leach, A.; Gall, K. Phys. Rev. Lett. 2008, 1 (2), No. 025502. (19) Hirth, J. P.; Lothe, J. Theory of dislocations; John Wiley & Sons: New York, 1982; pp xii, 857. (20) Mara, N. A.; Bhattacharyya, D.; Dickerson, P.; Hoagland, R. G.; Misra, A. Appl. Phys. Lett. 2008, 92 (23), No. 231901. (21) Rinaldi, A.; Peralta, P.; Friesen, C.; Sieradzki, K. Acta Mater. 2008, 56 (3), 511–517. (22) Jang, D.; Greer, J. R. Scr. Mater. 2011, 64 (1), 77–80. (23) Burek, M. J.; Greer, J. R. Nano Lett. 2010, 10 (1), 69–76. (24) Chen, K. C.; Wu, W. W.; Liao, C. N.; Chen, L. J.; Tu, K. N. Science 2008, 321 (5892), 1066–1069. (25) Jennings, A. T.; Burek, M. J.; Greer, J. R. Phys. Rev. Lett. 2010, 104 (13), No. 135503. (26) Li, X.; Wei, Y.; Lu, L.; Lu, K.; Gao, H. Nature 2010, 464 (7290), 877–880. (27) Cao, A. J.; Wei, Y. G.; Mao, S. X. Appl. Phys. Lett. 2007, 90, No. 151909. (28) Deng, C.; Sansoz, F. Nano Lett. 2009, 9 (4), 1517–1522.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT The authors gratefully acknowledge the financial support of the NSF-sponsored MRSEC Center at Caltech (DMR-0520565). The 1746
dx.doi.org/10.1021/nl2003076 |Nano Lett. 2011, 11, 1743–1746