Theoretical and Experimental Studies of the Optical Properties of

Nov 17, 2010 - Andrew J. Logsdail and Roy L. Johnston. The Journal of ... D T Tran , I P Jones , J A Preece , R L Johnston , K Deplanche , L E Macaski...
1 downloads 0 Views 3MB Size
J. Phys. Chem. C 2010, 114, 21247–21251

21247

Theoretical and Experimental Studies of the Optical Properties of Conjoined Gold-Palladium Nanospheres Andrew J. Logsdail,*,† Nikki J. Cookson,† Sarah L. Horswell,† Z. W. Wang,‡ Z. Y. Li,‡ and Roy L. Johnston*,† School of Chemistry and Nanoscale Physics Research Laboratory, School of Physics and Astronomy, UniVersity of Birmingham, Birmingham, B15 2TT, U.K. ReceiVed: September 6, 2010; ReVised Manuscript ReceiVed: October 25, 2010

We identify and report on UV-vis spectral features of AucorePdshell nanoparticles. Chains of conjoined nanospheres are formed with clear core/shell segregation and identified via experimental spectral red shifts of λmax and scanning transmission electron microscopy (STEM) imaging. Theoretical calculations performed using the discrete dipole approximation (DDA) validate our findings. The DDA simulation method is then used to delve further into possible spectroscopic implications of geometric and environmental changes. Structural configurations including extended chains of nanospheres, differing levels of nanosphere conjoinment, and increased Pd shell thickness are investigated, as well as variations of the surrounding dielectric medium. Introduction Plasmonics is the title given to optical phenomena related to plasma oscillations in metals. Original work on surface plasmons is attributed to Ritchie,1 who postulated that electrons traveling through thin films would experience less energy loss than those traveling through bulk, which was subsequently verified by Stern and Ferrell.2 The excitation of surface plasmons by light on metal planar surfaces is called surface-plasmon resonance (SPR) or localized SPR for metal clusters in the nanometre size range. Nanoplasmonics has received considerable attention in recent times because of technological advances which now allow for the manipulation and structural characterization of clusters on the nanometre scale and important applications of these features, for example in sensors. The optical properties of nanoparticles are size and shape dependent, allowing for tunability; in the case of heterogeneous nanoparticles, composition and chemical ordering allows for further tunability.3 Nanorod shapes are of particular current interest because they display two axisdependent SPRs.4 Exact solutions to Maxwell’s equations of classical electromagnetism were first presented by Mie5 and are only known for special geometries such as spheres,6 ellipsoids,7 and infinite cylinders.8 As a result, approximate methods are often required. An example is the discrete-dipole approximation (DDA), which is a flexible method for computing the absorption and scattering by nanoparticles with an arbitrary geometry.9 Recent theoretical work performed using the DDA, with comparisons to experimental results, has investigated Au nanospheres,10-12 rods,4,13-15 disks,16 and cages17 and the effect of coupling between some of these geometries.10,12,18,19 A recent study has also been conducted on AuPd rods by Xiang et al.20 Calculations have shown that the plasmonic responses of complex structures can be regarded as interactions of simpler geometries.21 AuPd nanoclusters are of considerable industrial interest because of their catalytic properties which are dependent on * To whom correspondence should be addressed. E-mail: andylogsdail@ stchem.bham.ac.uk and [email protected]. † School of Chemistry. ‡ Nanoscale Physics Research Laboratory, School of Physics and Astronomy.

structure and chemical ordering and are enhanced by their large surface-to-volume ratio.22 Although the PdcoreAushell configuration is thermodynamically favored at low temperatures,23,24 AucorePdshell and mixed configurations have also been characterized experimentally, generated as metastable structures.25-27 Here, simulations performed by using the DDA for linearly aggregated AucorePdshell nanospheres are presented, to aid the interpretation of experimental work. Our simulations are compared with previous theoretical work on Au and AucorePdshell nanoparticles. We hope to find agreement between theoretical and experimental results to help explain the structural origin of each resonance and thus to control the appearance of the extinction spectra. Experimental Method Synthesis of AucorePdshell Nanoparticles. Hydrogen tetrachloroaurate (III) (Premion, 99.999%), potassium tetrachloropalladate (II) (Premion, 99.99%), and trisodium citrate dihydrate (99%) were obtained from Alfa Aesar and used as received. Sodium borohydride (95%, Riedel de Hae¨n) and hydroxylamine (99%, Fluka) were used without further purification. Ultrapure water, obtained from a tandem Elix-Milli-Q Gradient A10 system (resistivity > 18.2 MΩ cm, TOC < 5 ppb, Millipore, France), was used throughout. All glassware was cleaned by heating in a 1:1 mixture of concentrated nitric and sulphuric acids for at least 1 h. The glassware was allowed to cool, rinsed with copious quantities of water, soaked for several hours in water, and rinsed again. The Au seed particles were prepared according to the method reported by Brown et al.26 A total of 1 ml 1% HAuCl4 was added to 90 mL water at room temperature. After 1 min stirring, 2 mL 38.8 mmol dm-3 sodium citrate was added. After 1 min, 1 mL fresh 0.075% NaBH4 in sodium citrate was added. The solution was stirred for an additional 5 min and stored in the dark at 4 °C. Au particles were coated with Pd by adapting a method reported by Cao et al. for Pt deposition.28 Appropriate quantities (according to the desired Au:Pd ratio) of K2 PdCl4 and 0.9 mL 1 wt% hydroxylamine were added to 50 mL of the seed solution,

10.1021/jp108486a  2010 American Chemical Society Published on Web 11/17/2010

21248

J. Phys. Chem. C, Vol. 114, No. 49, 2010

Logsdail et al.

and the solution was stirred at 60 °C for a range of times between 30 min and 20 h. Characterization Techniques. Absorption Spectra. The particles were diluted in H2O, and UV-vis absorption spectra were acquired using a Camspec M550 spectrometer with a resolution of 2 nm. STEM Measurements. The characterization of the nanoparticles was conducted employing a 200 kV Tecnai F20 scanning transmission electron microscopy (STEM) with a high-angle angular dark field (HAADF) detector. Use of this method allowed for the determination of particle size, along with the identification of chemical ordering by using Z-contrast imaging.29,30 The specimen was prepared by using the drop-cast method on a standard TEM copper grid, covered with an amorphous carbon film. DDA Simulations. In this work, the SPR spectra have been simulated within the DDA, using the program DDSCAT.9,31 DDSCAT 7.0 is a free Fortran software package which enables the simulation of a variety of nanoclusters via implementation of the DDA, allowing for the definition of shape, size, and composition. Relevant quantities calculated by DDSCAT are the absorption efficiency factor (Qabs), the scattering efficiency factor (Qsca), and the extinction efficiency factor (Qext), which are calculated, respectively, from 2 Qabs ≡ Cabs /πaeff

(1)

2 Qsca ≡ Csca /πaeff

(2)

Qext ≡ Qabs + Qsca

(3)

where Cabs and Csca are the absorption and scattering cross sections for a target, respectively, and aeff is characterized as the effective radius, calculated as aeff ≡ ((3V)/(4π))1/3, where V is the volume of the target. In our work, we have looked at the UV-visible region (300 < λ < 1200 nm); therefore, the maximum particle length we can study is ∼100 nm. Previous work has shown that the DDA is converged to a few percent when the number of dipoles N g 103.32 To ensure accuracy, for our work, N > 105 throughout, with 1000 dipoles nm-3. Complex dielectric constants are wavelength-dependent for metallic systems. Experimental values from thin-film SPR measurements for Au and Pd were used.33,34 Calculations were performed on the University of Birmingham BlueBEAR linux cluster, which has 384 dual-core worker nodes, each with two 2.6 GHz AMD processors and 8 Gb memory, and a pool of over 150 Tb of storage space.35 Results and Discussion Experimental. UV-vis spectra for isolated Au seeds (Figure 1) show strong absorption at 520 nm. This feature is characteristic of Au spheres on the nanometre scale and has been well documented previously.10,19 Addition of a Pd shell results in a red shift of the main feature (λmax) to 550 nm. HAADF-STEM imaging shows that AucorePdshell particles have formed. Because of the Z-contrast, the brighter Au core (Z ) 79) is distinguished from the darker Pd shell (Z ) 46) (Figure 2c); however, merging of these Pd shells is visible, corresponding to conjoined AucorePdshell particles. Further HAADF-STEM analysis shows that the AucorePdshell nanospheres have formed

Figure 1. Experimental UV-vis absorption spectra for Au seeds and AucorePdshell nanoparticles formed after 8 h heating, normalized to λmax (indicated by vertical lines).

chains, many particles in length, and that chain dispersion increases with heating time up to 8 h (Figure 2a,b). Simulation. DDSCAT allows for user-defined arbitrary shapes, which we utilize in this study to replicate the merging of spherical AucorePdshell nanoparticles witnessed experimentally. Geometries were created representing two and three colinear spherical particles with touching shells, overlapping shells, and touching cores (Figure 3). A radius of 2 nm is used for the Au core of the spherical particles throughout, with a Pd-shell thickness of 0.5 nm, giving an overall radius of 2.5 nm (as measured from the STEM images). For isolated spheres, this corresponds to a volume stoichiometry of ∼1:1. For a single AucorePdshell particle, the Au SPR features are quenched considerably by the addition of a Pd coating (Figure 4);25 however, a small ripple remains at ∼500 nm. Initially, calculations were carried out in vacuo (ε0 ) 1) for two shell-overlapping spheres with the interparticle axis parallel (0°) and perpendicular (90°) to the incident radiation (k) to identify the potential SPR contributions of longitudinal (LSPR) and transverse (TSPR) waves (Figure 4). The prominent contribution to the UV-vis absorption spectra is shown to be from the LSPR, but TSPR influences cannot be discounted in our calculations. Previously, red shifts in LSPRs have been witnessed with increasing aspect ratio (AR) of Au rods4,13-15 and also coupling of spheres,10 where AR is merely the length-to-width ratio of a particle. For the case of conjoined particles, the length is defined along the interparticle axis, and the width is equal to the diameter of a particle. In our calculations, the TSPR feature neither moves nor intensifies, implying the change is not associated with coupling effects. For coupled particles (i.e., not touching), the overall charge neutrality must be respected for each particle; in our case the overlapping dielectric Pd shells mean that the intraparticle charge neutrality does not have to be respected for the Au cores; thus, the particle overlap results in nanorod-like features (dependent on the AR) with longitudinal plasmon oscillations.12 Henceforth, the radiation source is set at θ ) 45° to the interparticle axis to account for contributions from both the LSPR and the TSPR. Increasing the number of conjoined particles in the chain results, intuitively, in increased ARs: for one, two, and three conjoined particles, with overlapping shells, the ARs are 1, 1.9, and 2.8, respectively. In turn, this increase results in a red shift of λmax. However, the increase in λmax is nonlinear with AR: between two and three particles, the red shift is 120 nm,

Studies of the Optical Properties of Conjoined AuPd Nanospheres

J. Phys. Chem. C, Vol. 114, No. 49, 2010 21249

Figure 2. Left to right: HAADF-STEM images of AucorePdshell particles after (a) 30 min and (b) 8 h heating time. (c) A higher-resolution image of part of a AucorePdshell particle chain formed after 8 h heating, illustrating the AucorePdshell segregation.

Figure 3. Left to right: schematics of the models used for (a) touching shells, (b) overlapping shells, and (c) touching cores. Black and white shadings represent the Au core and Pd shell, respectively.

Figure 5. DDA-calculated SPR spectra for increasing numbers of conjoined AucorePdshell particles. λmax is indicated by vertical lines of matching key. Au and Pd are represented in the schematics by black and white, respectively.

Figure 4. DDSCAT calculations for isolated Au and AucorePdshell spheres compared to AucorePdshell particles oriented with conjoining axis parallel and perpendicular to k. Au and Pd are represented in the schematics by black and white, respectively.

considerably greater than the red shift of 40 nm between one and two particles (Figure 5). This nonlinearity is due to the SPR dampening effect of the Pd shell between the Au cores, like previous reports for Au nanorods which give linear relationships13-15 up to a width of 30 nm, at which point the quasistatic approximation no longer remains valid.36 When varying the degree of overlap between two AucorePdshell particles, using the models previously described, we see changes in λmax (Figure 6, upper panel). Initially, for touching particles, the extinction spectrum is a superposition of the spectra for individual nanospheres; although the intensity has increased, the resonance shape and position is identical to an isolated AucorePdshell particle (λmax ) 500 nm). This can be explained by the quenching effect of Pd dominating the spectrum, because it shields the Au cores from each other. Overlapping of the particles causes a red shift of λmax to 528 nm. The peak is broad because of Pd shielding of the Au cores. Further overlap, to the extent that the Au cores are touching, results in a blue shift of λmax back to 516 nm because of the decreased AR. Here, the shift is accompanied by the peak becoming more pronounced. This increase in definition and intensity is attributable to the reduced levels of Pd, and thus shielding, between the Au cores.

Calculations for three conjoined AucorePdshell particles yield similar results (Figure 6, lower panel). Touching particles again represent a superposition, with λmax ≈ 500 nm. Shell overlapping results in a dramatic red shift of λmax to 658 nm, accompanied by peak broadening. Bringing the Au cores into contact with each other leads to λmax blue shifting back to 578 nm, with increased intensity and definition of the peak. The greater spread displayed by λmax for three conjoined particles can be explained by the increased chain length, and thus AR, leading to longerwavelength LSPR oscillations. Overall, for three particles, intensity is significantly increased across the spectrum compared to the case with two particles. Calculations were repeated by defining the surrounding medium as water (ε0 ) 1.333) and toluene (ε0 ) 1.497). Previously, the increasing dielectric of a surrounding medium has been found to cause red shifting of the spectra relative to vacuum,11 and our calculations are consistent with this (Figure 7). For two shell-overlapping AucorePdshell particles, we see a significant red shift, broadening, and quenching of the LSPR feature [Figure 7, upper panel]. However, in all calculations, a peak remains at ∼510 nm. This can be attributed to the increasing strength of the TSPR; calculations for Au cores on their own show a quasilinear growth in intensity with ε0 values. This is contrary to our previous observations, where the TSPR contributed little to the overall spectrum (Figure 4), but implies that the increasing ε0 value results in stronger dipolar interactions with the Pd shell, consequently decreasing its damping effect on the Au core.

21250

J. Phys. Chem. C, Vol. 114, No. 49, 2010

Logsdail et al.

Figure 8. DDA-calculated SPR extinction spectra for conjoined AucorePdshell particles with Au:Pd volume stoichiometry of ∼1:4. Au and Pd are represented in the schematics by black and white, respectively.

Figure 6. DDA-calculated SPR extinction spectra for differing levels of overlap between two conjoined AucorePdshell particles (top) compared to similar calculations for three conjoined AucorePdshell particles (bottom).

Figure 9. Comparison of the position of λmax in our results for AucorePdshell particles with 1:1 volume stoichiometry with previous work on Au nanorods by Kooij et al. on prolate ellipsoids and cylinders13 and by Yan et al. on hemispherically capped cylinders.37

Figure 7. DDA-calculated SPR extinction spectra for two AucorePdshell particles with shell overlapping (top) and core touching (bottom), with various surrounding medium dielectrics (vacuum, ε0 ) 1; water, ε0 ) 1.333; toluene, ε0 ) 1.497).

For two core-touching spheres, there is a well-defined peak, with λmax of 516, 578, and 624 nm for vacuum, water, and

toluene, respectively. Similarly, for three core-touching spheres, λmax is 578, 716, and 793 nm for vacuum, water, and toluene, respectively (these results are not shown). These shifts show a close-to-linear variation with increasing ε0. The effect of the Pd shell on the SPR spectra was tested with models of core size 2 nm and Pd shell of thickness 1.5 nm, hence, a combined radius of 3.5 nm and volume stoichiometry of ∼1:4 (Figure 8). Here, we found that the overweighting of Pd resulted in complete quenching of the Au features seen previously for thinner shell thickness and individual AucorePdshell particles (Figures 5 and 6). We conclude that the Pd shell, therefore, is thick enough for a volume stoichiometry of 1:4 so that only Pd SPR features are prominently visible. This effect has been observed, on occasions, experimentally in our work and is consistent with previous studies of isolated AucorePdshell nanospheres where a Pd coating was found to quench features in the UV-visible region.25 Figure 9 shows λmax versus AR as previously documented for Au rods13,37 compared with our results for AucorePdshell particles with 1:1 volume stoichiometry. For touching AucorePdshell particles, we have an almost flat line, with minimal coupling-based interactions. A completely horizontal line would be expected for noninteracting particles because there is no distinguishability between the TSPR and LSPR. Similarly, there is only a small λmax red shift for two core-touching and shelloverlapping particles (Figure 6, upper panel). Only for three

Studies of the Optical Properties of Conjoined AuPd Nanospheres conjoined AucorePdshell spheres do we see a significant increase of λmax (Figure 6, lower panel), which we can attribute to the LSPR. The red shift of λmax is consistent in part with previous work by Zhong et al. on colinear Au spheres: they found separation of the TSPR and LSPR and a red shift of the main SPR when Au spheres were brought together in the simulation.38 Conclusions We have identified previously unreported optical-spectra phenomena associated with conjoined AucorePdshell nanospheres by matching computational simulations with experimental results. Furthermore, we have shown that different extents of overlap between AucorePdshell nanospheres result in the evolution of the extinction spectral features from those characteristic of a sphere toward those of rod-shaped geometries. LSPR features of Au are not quenched by a Pd layer of 0.5 nm; however, a Pd shell of 1.5 nm quenches all Au features in the absorption spectrum. Increasing the refractive index of the surrounding medium results in strong red shifts and broadening of prominent peaks, with the appearance of a residual TSPR feature when Au cores are not in contact. Comparison of our results with previous work on Au nanorods illustrates the progression from AucorePdshell spheres toward more rod-like attributes in the extinction spectrum (Figure 9). Further work is needed to investigate the effects of Pd shells on Au rods to validate our conclusions with regards to the change in feature types, as well as further experimental studies to validate our predictions for differing surrounding media AucorePdshell. Acknowledgment. A.J.L. acknowledges financial support from the EPSRC, U.K. (DTA Award Reference: EP/P504678/ 1) and the University of Birmingham, U.K. R.L.J. and Z.Y.L. acknowledge financial support from the EPSRC, U.K. The authors acknowledge support from COST Action MP0903 “Nanoalloys as Advanced Materials: From Structure to Properties and Applications” and thank Ruth Chantry for a critical reading of the manuscript. References and Notes (1) Ritchie, R. H. Phys. ReV. 1957, 106, 874–881. (2) Stern, E. A.; Ferrell, R. A. Phys. ReV. 1960, 120, 130–136. (3) Ferrando, R.; Jellinek, J.; Johnston, R. L. Chem. ReV. 2008, 108, 845–910. (4) Kooij, E. S.; Poelsema, B. Phys. Chem. Chem. Phys. 2006, 8, 3349– 3357. (5) Mie, G. Annal. Phys. 1908, 25, 377. (6) Wait, J. R. J. Appl. Sci. Res. B 1963, 10, 441–450. (7) Gans, R. Annal. Phys. 1912, 342, 881.

J. Phys. Chem. C, Vol. 114, No. 49, 2010 21251 (8) Wait, J. R. Can. J. Phys. 1955, 33, 189–195. (9) Draine, B. T.; Flatau, P. J. J. Opt. Soc. Am. A 1994, 11, 1491– 1499. (10) Sandrock, M. L.; Foss, C. A. J. Phys. Chem. B 1999, 103, 11398– 11406. (11) Kelly, K. L.; Coronado, E.; Zhao, L. L.; Schatz, G. C. J. Phys. Chem. B 2003, 107, 668–677. (12) Romero, I.; Aizpurua, J.; Bryant, G. W.; Abajo, F. J. G. D. Opt. Express 2006, 14, 9988–9999. (13) Brioude, A.; Jiang, X. C.; Pileni, M. P. J. Phys. Chem. B 2005, 109, 13138–13142. (14) Prescott, S. W.; Mulvaney, P. J. Appl. Phys. 2006, 99, 123504. (15) Lee, K.-S.; El-Sayed, M. A. J. Phys. Chem. B 2005, 109, 20331– 20338. (16) Larsson, E. M.; Alegret, J.; Ka¨ll, M.; Sutherland, D. S. Nano Lett. 2007, 7, 1256–1263. (17) Chen, J.; Wiley, B.; Li, Z.-Y.; Campbell, D.; Saeki, F.; Cang, H.; Au, L.; Lee, J.; Li, X.; Xia, Y. AdV. Mater. 2005, 17, 2255–2261. (18) Jain, P. K.; Eustis, S.; El-Sayed, M. A. J. Phys. Chem. B 2006, 110, 18243–18253. (19) Ghosh, S. K.; Pal, T. Chem. ReV. 2007, 107, 4797–4862. (20) Xiang, Y.; Wu, X.; Liu, D.; Jiang, X.; Chu, W.; Li, Z.; Ma, Y.; Zhou, W.; Xie, S. Nano lett. 2006, 6, 2290–2294. (21) Prodan, E.; Nordlander, P. J. Chem. Phys. 2004, 120, 5444. (22) Liu, F.; Wechsler, D.; Zhang, P. Chem. Phys. Lett. 2008, 461, 254– 259. (23) Pittaway, F.; Paz-Borbo´n, L. O.; Johnston, R. L.; Arslan, H.; Ferrando, R.; Mottet, C.; Barcaro, G.; Fortunelli, A. J. Phys. Chem. C 2009, 113, 9141–9152. (24) Ismail, R.; Johnston, R. L. Phys. Chem. Chem. Phys. 2010, 12, 8607–8619. (25) Kan, C.; Cai, W.; Li, C.; Zhang, L.; Hofmeister, H. J. Phys. D: Appl. Phys. 2003, 36, 1609–1614. (26) Brown, K. R.; Walter, D. G.; Natan, M. J. Chem. Mater. 2000, 12, 306. (27) Ferrer, D.; Blom, D. A.; Allard, L. F.; Mejia, S.; Perez-Tijerina, E.; Jose´-Yacama´n, M. J. Mater. Chem. 2008, 18, 2442. (28) Cao, L.; Tong, L.; Diao, P.; Zhu, T.; Liu, Z. Chem. Mater. 2004, 16, 3239. (29) Li, Z. Y.; Yuan, J.; Chen, Y.; Palmer, R. E.; Wilcoxon, J. P. Appl. Phys. Lett. 2005, 87, 243103. (30) Li, Z. Y.; Wilcoxon, J. P.; Yin, F.; Chen, Y.; Palmer, R. E.; Johnston, R. L. Structures and optical properties of 4-5 nm bimetallic AgAu nanoparticles. Faraday Discuss. 2008, 138, 363. (31) Yurkin, M. A.; Hoekstra, A. G. J. Quant. Spectrosc. Radia. Transf. 2007, 106, 558–589. (32) Sosa, I. O.; Noguez, C.; Barrera, R. J. Phys. Chem. B 2003, 107, 6269–6275. (33) Johnson, P. B.; Christy, R. W. Phys. ReV. B 1972, 6, 4370–4379. (34) Johnson, P. B.; Christy, R. W. Phys. ReV. B 1974, 9, 5056–5070. (35) http://www.bear.bham.ac.uk/bluebear. (36) Slaughter, L. S.; Chang, W.-S.; Swanglap, P.; Tcherniak, A.; Khanal, B. P.; Zubarev, E. R.; Link, S. J. Phys. Chem. C 2010, 114, 4934– 4938. (37) Yan, B.; Yang, Y.; Wang, Y. J. Phys. Chem. B 2003, 107, 9159. (38) Zhong, Z.; Patskovskyy, S.; Bouvrette, P.; Luong, J. H. T.; Gedanken, A. Phys. ReV. B 2004, 108, 4046–4052.

JP108486A