pubs.acs.org/Langmuir © 2009 American Chemical Society

A Simple Method of Superlattice Formation: Step-by-Step Evaluation of Crystal Growth of Gold Nanoparticles through Seed-Growth Method Mandeep Singh Bakshi* Department of Chemistry, Acadia University, 6 University Avenue, Elliot Hall, Wolfville, NS B4P 2R6, Canada

Downloaded by MT ST VINCENT UNIV on October 28, 2009 | http://pubs.acs.org Publication Date (Web): July 20, 2009 | doi: 10.1021/la901767c

Received May 18, 2009. Revised Manuscript Received June 26, 2009 Self-assembled arrangement of monodisperse nanoparticles (NPs) forms one-dimensional (1D) to three-dimensional (3D) superlattice (SL) with many useful applications. A simple seed growth (S-G) method is presented to achieve monodisperse gold (Au) NPs and simultaneously arrange them into SL formation. It can simply be done by controlling the hydrophobicity of a capping surfactant and can very well be extended to semiconductor NPs such as PbS as well. It is demonstrated by step-by-step evaluation of a three-step S-G method in the presence of a series of strongly hydrophobic Gemini surfactants. NPs of each step are analyzed by transmission electron microscopy (TEM) and UV-visible measurements to evaluate the mode of aggregation in dried and colloidal bulk phases, respectively. Both studies show complementary results. Crystal growth of NPs is followed through different steps by measuring the X-ray diffraction (XRD) patterns. It allows one to identify different reaction conditions such as the number of nucleating centers (seeds) and concentration of the surfactant to achieve monodisperse morphologies of NPs. All studies pertaining to different steps of the S-G method under different reaction conditions collectively lead to a single conclusion that better capping ability of strongly hydrophobic surfactants allows NPs to achieve both monodisperse morphologies as well as SL formation simultaneously.

Introduction Ordered assemblies of monodisperse nanoparticles (NPs) form a superlattice (SL) under appropriate conditions.1 SL formation has received much attention recently due to their useful and interesting applications.2 Different methods can be implemented to design SL structures. Solution phase synthesis of SL formation proves to be an easy and straightforward method. The seed growth (S-G) method3 is a very convenient way to produce small monodisperse NPs. NPs of Au and Ag possess the ability to arrange themselves in long-range hexagonal close pack (hcp) arrangement,1 which basically arises *E mail: [email protected]. (1) (a) Zhang, Q. B.; Xie, J. P.; Yang, J. H.; Lee, J. Y. ACS Nano 2009, 3, 139– 148. (b) Chattopadhyay, S.; Mukherjee, R.; Datta, A.; Saha, A.; Sharma, A.; Kulkarni, G. U. J. Nanosci. Nanotechnol. 2009, 9, 190–194. (c) Abecassis, B.; Testard, F.; Spalla, O. Phys. Rev. Lett. 2008, 100, 115504-1–115504-4. (d) Xu, J.; Zhou, L. H.; Liu, H. L.; Hu, Y. J. Exp. Nanosci. 2006, 1, 103–111. (e) Pyrpassopoulos, S.; Niarchos, D.; Nounesis, G.; Boukos, N.; Zafiropoulou, I.; Tzitzios, V. Nanotechnology 2007, 18, 485604. (f) Didiot, C.; Pons, S.; Kierren, B.; Fagot-Revurat, Y.; Malterre, D. Nat. Nanotechnol. 2007, 2, 617–621. (g) Kim, Y. H.; Lee, D. K.; Cha, H. G.; Kim, C. W.; Kang, Y. S. Chem. Mater. 2007, 19, 5049–5051. (h) Yang, Y.; Kimura, K. J. Phys. Chem. B 2006, 110, 24442–24449. (i) Gutierrez-Wing, C.; Santiago, P.; Ascencio, J. A.; Camacho, A.; JoseYacaman, M. Appl. Phys. A: Mater. Sci. Process. 2000, 71, 237–243. (j) Lin, X. M.; Sorensen, C. M.; Klabunde, K. J. Chem. Mater. 1999, 11, 198–202. (k) Stoeva, S. I.; Prasad, B. L. V.; Uma, S.; Stoimenov, P. K.; Zaikovski, V.; Sorensen, C. M.; Klabunde, K. J. J. Phys. Chem. B 2003, 107, 7441–7448. (2) (a) Stapleton, J. J.; Harder, P.; Daniel, T. A.; Reinard, M. D.; Yao, Y. X.; Price, D. W.; Tour, J. M.; Allara, D. L. Langmuir 2003, 19, 8245–8255. (b) Dieluweit, S.; Pum, D.; Sleytr, U. B.; Kautek, W. Mater. Sci. Eng., C: Biomimetic Supramol. Syst. 2005, 25, 727–732. (c) Martin, J. E.; Wilcoxon, J. P.; Odinek, J.; Provencio, P. J. Phys. Chem. B 2000, 104, 9475–9486. (3) (a) Murphy, C. J.; San, T. K.; Gole, A. M.; Orendorff, C. J.; Gao, J. X.; Gou, L.; Hunyadi, S. E.; Li, T. J. Phys. Chem. B 2005, 109, 13857. (b) Jana, N. R.; Gearheart, L.; Murphy, C. J. Langmuir 2001, 17, 6782. (c) Nikoobakht, B.; El Sayed, M. A. Chem. Mater. 2003, 15, 1957. (d) Gao, J.; Bender, C. M.; Murphy, C. J. Langmuir 2003, 19, 9065. (e) Jana, N. R.; Gearheart, L.; Murphy, C. J. Chem. Mater. 2001, 13, 2313. (f) Srnov
a-loufov
a, I.; Vlckov
a, B.; Bastl, Z.; Hasslett, T. L. Langmuir 2004, 20, 3407. (g) Bakshi, M. S.; Possmayer, F.; Petersen, N. O. J. Phys. Chem. C 2008, 112, 8259. (h) Bakshi, M. S.; Possmayer, F.; Petersen, N. O. Chem. Mater. 2007, 19, 1257. (i) Bakshi, M. S.; Possmayer, F.; Petersen, N. O. J. Phys. Chem. C 2007, 111, 14113. (j) Bakshi, M. S.; Sachar, S.; Kaur, G.; Bhandari, P.; Kaur, G.; Biesinger, M. C.; Possmayer, F.; Petersen, N. O. Cryst. Growth Des. 2008, 8, 1713.

Langmuir 2009, 25(21), 12697–12705

from the core NP facet-driven phenomena. SL can be easily separated from the solution phase and arranged on a solid support.1a,g,h,k Fascinating morphologies can be generated by simply arranging NPs on templated structures such as polymer strands4 driven by specific interactions as they exist between thiol groups and Au NPs.5 Alkanethiols5 direct monodisperse Au or Ag NPs to self-assemble in a manner usually different from their natural hcp arrangement. Thiolcapped NPs, when subjected to electrostatic or nonelectrostatic intermolecular interactions between the capping layers of adjoining NPs, constitute diverse SL morphologies. Stronger interactions can even drive large nanorods (NRs) into threedimensional (3D) assemblies and are best known for doubletail conventional surfactants or phospholipids.6 Two methods can be adopted to generate SL structures. First, presynthesized NPs are treated with capping molecules such as surfactants or lipids to develop a capping layer, which then acts as a glue for the self-assembled 3D structure.5,6 Second, monodisperse NPs are directly synthesized in the presence of strong surface active agents, and NPs then acquire a uniform selfassembled state upon interacting with each other through capping surfactant layers.1a,1k The second method is much easier and convenient to follow if an appropriate capping agent is used. The S-G method can be easily implemented for the second method because it is a well-studied method.3 Although several aspects of the S-G method are related to both thermodynamical as well as kinetic controlled factors, the choice of an appropriate surfactant is highly important to obtain monodisperse NPs.3i,3j Strongly hydrophobic surfactants such as double-tail Gemini surfactants and lipids are the best shape-directing agents.3 Later, property is (4) Chen, C.-L.; Zhang, P.; Rosi, N. L. J. Am. Chem. Soc. 2008, 130, 13555. (5) Chen, C.-F.; Tzeng, S.-D.; Chen, H.-Y.; Lin, K.-J.; Gwo, S. J. Am. Chem. Soc. 2008, 130, 824. (6) Nakashima, H.; Furukawa, K.; Kashimura, Y.; Torimitsu, K. Langmuir 2008, 24, 5654.

Published on Web 07/20/2009

DOI: 10.1021/la901767c

12697

Article

Bakshi

Downloaded by MT ST VINCENT UNIV on October 28, 2009 | http://pubs.acs.org Publication Date (Web): July 20, 2009 | doi: 10.1021/la901767c

Scheme 1

mainly related to an effective liquid-solid interfacial adsorption, which originates primarily from its stronger amphiphatic nature. In the present study, the three-step S-G method has been analyzed to understand the different factors responsible for SL formation. After analyzing the shape and structure of NPs produced at each step, it has been concluded that an effective capping ability of a capping agent along with its strong hydrophobicity play an important role in gathering NPs in selfassembled SL formation.

Experimental Section Materials. Tetrachloroauric acid (HAuCl4), sodium borohydride (NaBH4), and trisodium citrate (Na3Cit) were obtained from Aldrich. Dimethylene bis(decyldimethylammonium bromide) (10-2-10), dimethylene bis(dodecyldimethylammonium bromide) (12-2-12), and dimethylene bis-tetradecyldimethylammonium bromide (14-2-14), Scheme 1, were synthesized as reported in the literature7 and used after repeated crystallization from ethanol. Ultra pure water (18 MΩ cm) was used for all aqueous preparations. Synthesis of 10-2-10/12-2-12/14-2-14 capped Au NPs was followed by S-G method essentially similar to that reported by Murphy et al.3a Briefly, 25 mL of seed solution was prepared by mixing [HAuCl4]=0.5 mM and [Na3Cit]=0.5 mM, and followed by the addition of 0.6 mL of aqueous NaBH4 ([NaBH4] = 0.1 mol dm-3) solution under constant stirring. Growth solution was prepared by dissolving [10-2-10]/[12-212]/[14-2-14]=1 mM in total 5 mL water along with [HAuCl4]= 0.5 mM, and was followed by the addition of 0.2 mL of freshly prepared ascorbic acid (AA) aqueous solution ([AA]=0.1 M). Different quantities of previously prepared seed solution were added at the end to start the growth process. Two methods were followed. In Method I, the three steps of each S-G reaction sequence were carried out at an interval of 1 min in which the entire reaction mixture of Step 1 was used as the seed for the next subsequent step. A 1 min time interval seems to be the best option3a because, within this period, the rate of the reaction in which nucleating centers undergo a growth process is maximum and provides better results rather than any extended period of time interval. In Method II, a specific amount of seed solution (i.e., 0.5, 0.25, and 0.125 mL) from Step 1 was used for the next subsequent step. Schematic flow diagrams of both methods have been shown and explained in the Discussion section. All samples were kept undisturbed overnight and then purified by repeated washing and centrifugation (at least three times) to obtain surfactant-free colloidal suspensions. Gemini surfactants are known7 for their micellar structure transitions such as thread-like micelles and vesicle formation just above the critical micelle concentration (cmc). In order to avoid the soft (7) (a) Bai, G.; Wang, J.; Yan, H.; Li, Z.; Thomas, R. K. J. Phys. Chem. B. 2001, 105, 3105. (b) Zana, R.; Benrraou, M.; Rueff, R. Langmuir 1991, 7, 1072. (c) Wettig, S. D.; Verrall, R. E. J. Colloid Interface Sci. 2001, 244, 377.

12698 DOI: 10.1021/la901767c

template effect of such structure transitions on the SL formation, thorough purification was necessary. Methods. UV-visible spectra of as prepared NP solutions were taken by a UV spectrophotometer (Multiskan Spectrum, model no. 1500) in the wavelength range of 200-900 nm to determine the absorbance due to surface plasmon resonance (SPR). Shape and size of NPs were characterized by transmission electron microscopy (TEM). Samples were prepared by mounting a drop of a sample on a carbon-coated Cu grid and were allowed to dry in air. They were observed with the help of a Philips CM10 Transmission Electron Microscope operating at 100 kV. The X-ray diffraction (XRD) patterns were characterized by using a Bruker-AXS D8-GADDS with Tsec = 480. Samples were prepared on glass slides by spotting a concentrated drop of aqueous suspension and were then dried in a vacuum desiccator.

Results and Discussion Method I (Total Addition). As mentioned in the Experimental Section, the total amount of the previous step is used as a seed for the subsequent steps. The left panels of Figure 1 show the reaction sequences of each sample. For example, sample Au1 is prepared by taking 0.5 mL of the previously made seed solution (see Experimental Section) in first step, then the entire first step is used as a seed for the second step within 1 min, and then the second step is used for the third step. TEM images of various samples shown in Figure 1 are of the third step in each case. Samples Au1, Au2, and Au3 are prepared in the presence of 1 mM of 10-2-10, while samples Au4, Au5, and Au6 are prepared in the presence of 1 mM of 12-2-12, by starting corresponding reactions with 0.5, 0.25, and 0.125 mL of seed solutions, respectively. A marked difference is observed between the morphologies of both series of samples.3j Au1 shows a predominant presence of NRs with an aspect ratio of 3.0 ( 0.5 (see corresponding histogram, Au1h) which are fused together in an end-to-end manner. Nothing like this can be seen in Au4, where large NPs of 55.3 ( 18.8 nm (Au4h) with different shapes are present. Likewise, sample Au2 shows a predominant amount of NRs with an increased aspect ratio of 4.5 ( 0.9 (Au2h). They are less fused and more prominent than those observed in Au1. Here, side-by-side attachment of NRs is also evident (see dotted circles). NPs of Au5 have morphologies almost similar to those of Au4, with slightly bigger size, 57.8 ( 10.8 nm (Au5h). Au3 now shows the presence of well-defined NRs with even higher aspect ratio, 5.8 ( 1.1 (Au3h), which still prefer to assemble in an end-to-end fashion (dotted circles). Similarly, NPs of Au6 have now acquired more clear morphologies with average size of 60.4 ( 18.7 nm (Au6h) relative to the corresponding previous samples (i.e., Au4, and Au5). Collectively, samples Au1-Au3 produce self-assembled NRs, whereas samples Au4-Au6 produce multifaceted icosahedrons (see particles in dotted circles). Well-defined NR formation has a direct relation with a decrease in the amount of seed3a-e,j from 0.5 to 0.125 mL. Use of 0.5 mL of seed (Au1) in comparison to 0.125 mL (Au3) generates a 4 times greater number of nucleating centers in sample Au1 than Au3 for equal amounts of gold ions at both places. As gold ions are reduced into atoms, they find a relatively much less number of nucleating centers in Au3. They will accommodate a large number of freshly prepared gold atoms at the {111} lattice planes of face-centered cubic (fcc) Au resulting in a regular increase in the aspect ratio from Au1 to Au3 due to a surface selective passivation of low-density {100}/{110} planes by surfactant monomers (see Figure 2a for a schematic representation). As both 10-2-10 and 12-2-12 have identical cationic dimeric head groups, the strength of interfacial Langmuir 2009, 25(21), 12697–12705

Downloaded by MT ST VINCENT UNIV on October 28, 2009 | http://pubs.acs.org Publication Date (Web): July 20, 2009 | doi: 10.1021/la901767c

Bakshi

Article

Figure 1. Left panels of this figure illustrate the flow diagram of three-step S-G reaction sequences for different samples. Each reaction starts with fixed amount of previously prepared seed. For instance, the first beaker refers to the first step of each reaction sequence and shows the amount of seed added to start the reaction. NPs were collected from the third beaker for each sample, and their TEM images are shown in the next two vertical rows. Samples Au1, Au2, Au3, and Au3, Au4, Au5 were prepared in the presence of 1 mM of 10-2-10 and 12-2-12, respectively. Dotted circles highlight the areas of interest (see text). The next two vertical rows show the size distribution histograms of NRs and NPs of corresponding samples. For example, Au1h belongs to Au1 and Au4h belongs to Au4.

adsorption lies with the overall magnitude of the hydrophobic component.8 A longer hydrophobic tail will face stronger expulsion from aqueous phase than a shorter one would, and hence would have greater potential for interfacial adsorption.9 Even the bulk properties show that one additional methylene group actually reduces the cmc to half its value.8,9 Thus, 12-2-12 monomers would possess greater ability to tightly and compactly cap not only low-density {100}/{110} planes but, to some extent, {111} planes as well, which would largely restrict any anisotropic growth and eventually lead to icosahedral shape bound with {111} facets1a (see Figure 2b for schematic representation). UV-visible measurements confirm these morphologies in the bulk phase. Figure 3 shows the UV-visible spectra of above six samples. Samples Au1, Au2, and Au3 show prominent absor(8) (a) Li, F.; Rosen, M. J.; Sulthana, S. B. Langmuir 2001, 17, 1037–1042. (b) Bai, G.; Yan, H.; Thomas, R. K. Langmuir 2001, 17, 4501–4504. (c) Yoshimura, T.; Ohno, A.; Esumi, K. Langmuir 2006, 22, 4643–4648. (d) Matsuoka, K.; Yoshimura, T.; Shikimoto, T.; Hamada, J.; Yamawaki, M.; Honda, C.; Endo, K. Langmuir 2007, 23, 10990–10994. (e) Bakshi, M. S.; Singh, J.; Kaur, G. J. Colloid Interface Sci. 2004, 285, 403. (f) Bakshi, M. S.; Sachar, S. Colloid Polym. Sci. 2005, 224, 671. (g) Bakshi, M. S.; Singh, J.; Kaur, G. J. Photochem. Photobiol. 2005, 173, 202. (h) Bakshi, M. S.; Kaur, G. J. Colloid Interface Sci. 2005, 289, 551. (9) (a) Okano, T.; Tamura, T.; Abe, Y.; Tsuchida, T.; Lee, S.; Sugihara, G. Langmuir 2000, 16, 1508–1514. (b) Okano, T.; Tamura, T.; Nakano, T. Y.; Ueda, S. I.; Lee, S.; Sugihara, G. Langmuir 2000, 16, 3777–3783. (c) Hisatomi, M.; Abe, M.; Yoshino, N.; Lee, S.; Nagadome, S.; Sugihara, G. Langmuir 2000, 16, 1515–1521. (d) Sulthana, S. B.; Rao, P. V. C.; Bhat, S. G. T.; Nakano, T. Y.; Sugihara, G.; Rakshit, A. K. Langmuir 2000, 16, 980–987. (e) Honda, C.; Itagaki, M.; Takeda, R.; Endo, K. Langmuir 2002, 18, 1999–2003.

Langmuir 2009, 25(21), 12697–12705

bances at higher wavelengths due to longitudinal surface plasmon resonance (LSPR)10 of NRs. Significant red shift with linear variation versus aspect ratio is evident, i.e., Au1(3.0 ( 0.5) < Au2(4.5 ( 0.9) < Au3(5.8 ( 1.1) (inset, Figure 3). No appreciable shift is observed in transverse surface plasmon resonance (TSPR) located close to 520 nm. Both LSPR as well as TSPR produce comparable intensities, because the absorbance at 520 nm also includes SPR contribution of small polyhedral NPs in all cases. Kamat and co-workers11 have observed a similar red shift experimentally, whereas El-Sayed et al.10b have demonstrated this from discrete dipole approximation simulation. LSPR arises from the collective resonance of conduction band electrons of surface atoms with light photons along the long axis. As the length of the long axis (i.e., aspect ratio) increases, the magnitude of weakly bound electrons increases along this axis, and hence the absorbance due to LSPR red-shifted. Red shift is also induced by an end-to-end linkage, and its magnitude increases from dimer to higher order.10b The present trend in LSPR is the sum of both factors, i.e., increase in the aspect ratio as well as end-to-end linkage. Although, side-by-side plasmon coupling of NRs induces blue shift in LSPR, its magnitude is considered to be much weaker (10) (a) Liz-Marzan, L. M. Langmuir 2006, 22, 32. (b) Jain, P. K.; Eustis, S.; ElSayed, M. A. J. Phys. Chem. B 2006, 110, 18253. (c) El-Sayed, M. A. Acc. Chem. Res. 2001, 34, 257. (d) Eustis, S.; El-Sayed, M. A. Chem. Soc. Rev. 2006, 35, 209. (11) Thomas, K. G.; Barazzouk, S.; Ipe, B. L.; Joseph, S. T. S.; Kamat, P. V. J. Phys. Chem. B 2004, 108, 13066.

DOI: 10.1021/la901767c

12699

Downloaded by MT ST VINCENT UNIV on October 28, 2009 | http://pubs.acs.org Publication Date (Web): July 20, 2009 | doi: 10.1021/la901767c

Article

Bakshi

Figure 2. Schematic representations of NR and icosahedral NP formation in the presence of 10-2-10 (a) and 12-2-12 (b). (c,d) The respective self-assembled states driven by the hydrophobic interactions between surfactant tails of capping layers.

Figure 3. UV-visible absorption spectra of various samples Au1-Au6. Note the absorbances at higher wavelengths due to LSPR of NRs. See text regarding the dependence of LSPR on the aspect ratio and self-assembled behavior of NRs. Absorbances located close to 530 nm are due to icosahedral NPs as well as TSPR (see text for details). Inset shows a linear variation of absorption wavelength with respect to aspect ratio.

with few such assemblies (see TEM image of Au2). On the other hand, samples Au4, Au5, and Au6 produce broad and prominent SPR located close to 540 nm. They are the consequence of large polyhedral NPs and become more prominent with particles bearing twin boundaries12 (Figure 3). Small spherical NPs produce equally (12) (a) Schwartzberg, A. M.; Zhang, J. Z. J. Phys. Chem. C 2008, 112, 10323– 10337. (b) Schwartzberg, A. M.; Olson, T. Y.; Talley, C. E.; Zhang, J. Z. J. Phys. Chem. B 2006, 110, 19935–19944. (c) Roma€un-Vela€uzquez, C. E.; Noguez, C.; Zhang, J. Z. J. Phys. Chem. A 2009, 113, 4068–4074.

12700 DOI: 10.1021/la901767c

distributed degenerate oscillations along all three axes and generate sharp absorbances. However, nondegenerate oscillations arising from twin boundaries with several different modes produce broad absorbances. Several NPs of Au6 (Figure 1) show twin boundaries (see NPs in dotted circles). Apart from this, weak LSPRs are also evident at higher wavelengths (Figure 3) as a result of the presence of few NRs in each sample (Figure 1). Method II (Partial Addition), Reaction Sequence in the Presence of 10-2-10. A second set of experiments was carried Langmuir 2009, 25(21), 12697–12705

Downloaded by MT ST VINCENT UNIV on October 28, 2009 | http://pubs.acs.org Publication Date (Web): July 20, 2009 | doi: 10.1021/la901767c

Bakshi

Article

Figure 4. Left panels show a three-step reaction sequence from top to bottom in the presence of 10-2-10. The top beaker represents the first step with a fixed amount of seed (i.e., 0.5/0.25/0.125 mL). The next three vertical rows show TEM micrographs of various samples obtained from each corresponding step. For example, Au7, Au10, and Au13 belong to the first step of three different reactions carried out with 0.5, 0.25, and 0.125 mL of seed solutions (see the amount of seed used at the top of each vertical row). Likewise, Au8, Au11, Au14, and Au9, Au12, Au15 belong to second and third steps, respectively. Right panels show the UV-visible spectra of corresponding samples. Panel a shows the spectra of all samples of the first step, i.e., Au7, Au10, and Au13; likewise panels b and c show that of the second and third steps, respectively.

Figure 5. XRD patterns of Au13, Au14, and Au15. Inset shows a variation in the intensity ratio. Langmuir 2009, 25(21), 12697–12705

DOI: 10.1021/la901767c

12701

Downloaded by MT ST VINCENT UNIV on October 28, 2009 | http://pubs.acs.org Publication Date (Web): July 20, 2009 | doi: 10.1021/la901767c

Article

Bakshi

Figure 6. Left panels show a three-step reaction sequence from top to bottom in the presence of 1 mM 12-2-12. The top beaker represents the first step with a fixed amount of seed (i.e., 0.5/0.25/0.125 mL). The next three vertical rows show TEM micrographs of various samples obtained from each corresponding step. For example, Au16, Au19, and Au22 belong to the first step of three different reactions carried out with 0.5, 0.25, and 0.125 mL of seed solutions (see the amount of seed used at the top of each vertical row). Likewise, Au17, Au20, Au23, and Au18, Au21, Au24 belong to the second and third steps, respectively. Right panels show the UV-visible spectra of corresponding samples. Panel a shows the spectra of all samples of the first step, i.e., Au16, Au19, and Au22; likewise panels b and c show that of the second and third steps, respectively.

out by choosing partial addition of the seed solution rather than the whole amount. The left panel of Figure 4 shows a reaction sequence of three steps. For example, the first vertical row of TEM micrographs belong to three samples of a reaction sequence produced by using 0.50 mL of the seed solution; likewise, the second and third rows represent those produced with 0.25 and 0.125 mL, respectively. Step 1 of all three reactions (top panels of all three rows) produces mostly small spherical NPs arranged in chains. Their UV-visible spectra show clear sharp absorbances due to SPR around 520 nm (Figure 4a). As the amount of seed decreases from Au7 f Au10 f Au13, NRs (aspect ratio=2.6 ( 0.5) appear in Au13 exactly following the route of Figure 2a, and show a weak LSPR located at 650 nm (Figure 4a). Step 2 produces a large amount of NRs in Au8 (3.9 ( 0.5), Au11 (4.7 ( 0.3), and Au14 (3.6 ( 1.5). NRs of these samples prefer to exist in side-by-side assembly (see NRs in dotted rectangles) driven by strong hydrophobic interactions between the double hydrocarbon tails of adjoining monolayers of different NRs, as shown schematically in Figure 2c. It is not only restricted to NRs; even polyhedral NPs show similar association in Au8 (schematic representation in Figure 2d). Double-chain-double-chain hydrophobic interactions are expected to be much intense than single-chain-single-chain 12702 DOI: 10.1021/la901767c

interactions. Geometric calculations suggest that an extended bilayer of 10-2-10 should be ∼1.25 nm. UV-visible behavior of Au8, Au11, and Au14 (Figure 4b) requires more careful analysis on the basis of symmetry-breaking plasmon coupling as demonstrated by El-Sayed et al.10b Linear red shift in LSPR is observed for Au8
Downloaded by MT ST VINCENT UNIV on October 28, 2009 | http://pubs.acs.org Publication Date (Web): July 20, 2009 | doi: 10.1021/la901767c

Bakshi

Article

Figure 7. XRD patterns of Au22, Au23, and Au24. Inset shows a variation in the intensity ratio.

are responsible for a red shift in LSPR and produce lower net dipole moments leading to lower intensity.10b TEM micrographs of Step 3 samples (Au9, Au12, and Au15) are much different from those of Step 2 samples. Here, the shape deformation3j sets in for both polyhedral NPs as well as NRs, which is a clear indication of a surfactant capping limit. A comparison between the equilibrium concentrations of both Au NPs and 10-2-10 shows that the molar concentration of NPs, ∼0.5 mM (assuming all Au3þ ions are reduced to Au0), is half the amount of 10-2-10=1 mM. Although double the amount of 10-2-10 is available for complete capping of all NPs, 1 mM is about 1/6.5 the cmc of 10-2-10.13 Complete capping is only expected with surfactant concentration greater than the cmc. Hence, as size grows beyond a certain limit, anisotropic crystal growth starts even on previously passivated {100}/{110} crystal planes. Actually, the roughly star shaped morphologies shown in the TEM micrograph of sample Au15 are the consequence of such a growth process.3j Recently, Liz-Marz
an et al.14 reported a systematic change in NR morphology to different geometries simply by varying the HAuCl4/NRs molar ratio. More pronounced changes are observed at high ratios as a result of a faster growth along the {100}>{110}>{111} planes, which is altogether related to the depleting capping layer. NRs of Au9 and Au12 are predominantly end-to-end attached with >90° angle between the NR axes. A shift from the side-by-side arrangement (Step 2) to end-to-end is obviously related to the shortage of capping surfactant on the {100}/{110} planes, which cannot produce sufficient hydrophobic interactions necessary for side-by-side attachment. UV-visible spectra (Figure 4c) authenticate this behavior very well in the bulk phase. Sample Au9 shows two absorbances located close to 540 and 960 nm, which belong to large polyhedral NPs (48.3 ( 3.5 nm) and long NRs (4.1 ( 0.7), (13) Bakshi, M. S.; Kaur, G. J. Colloid Interface Sci. 2005, 289, 551. Bakshi, M. S.; Singh, K. J. Colloid Interface Sci. 2005, 287, 288. Bakshi, M. S.; Sachar, S.; Singh, K.; Shaheen, A. J. Colloid Interface Sci. 2005, 286, 369. (14) (a) Carbo-Argibay, E.; Rodriguez-Gonzalez, B.; Pacifio, J.; PastorizaSantos, I.; Perez-Juste, J.; Liz-Marzan, L. M. Angew. Chem., Int. Ed. 2007, 46, 1. (b) Myroshnychenko, V.; Rodriguez-Gonzalez, B.; Pastoriza-Santos, I.; Fuston, A. M.; Novo, C.; Mulvaney, P.; Liz-Marzan, L. M. Chem. Soc. Rev. 2008, 37, 1792.

Langmuir 2009, 25(21), 12697–12705

respectively. Notably, sample Au8 (see TEM image) also contains NRs with an aspect ratio of 4, but its LSPR is located at 750 nm, which is much lower than 960 nm. This is due to a blue shift induced by symmetric coupling in comparison to a red shift induced by end-to-end attachment in Au9. An inter-NR angle of >90° further contributes toward the red shift due to antisymmetric coupling10b arising from the longitudinal plasmon of one NR to the transverse of the other. Likewise, sample Au12 shows only one very broad absorbance close to 600 nm due to large polyhedral NPs (97.5 ( 8.2 nm), while that of NRs (4.6 ( 0.6 nm) further shifts to the near-infrared (NIR) region of the spectrum due to antisymmetric coupling. Shape transformation from Step 1 to Step 3 has been further supported by XRD studies. Figure 5 shows the XRD patterns of Au13, Au14, and Au15 samples. All diffraction peaks can be indexed to Au fcc geometry. In order to follow the course of crystal growth from Au13 to Au15, plots of the intensity ratios of (111)/(200) and (111)/(220) (Figure 5, inset) show a steep increase and instantaneous fall, respectively. Both parameters indicate that more and more NPs bound with {111} and {110} facets15 are produced as the reaction progresses from Au13 through Au15. That promotes anisotropic crystal growth from sample Au13 to Au15, as evident from Figure 4. Method II (Partial Addition): Reaction Sequence in the Presence of 12-2-12. Similar reaction sequence was carried out in the presence of 1 mM of 12-2-12, and TEM micrographs are shown in Figure 6. Step 1 (Au16) with seed=0.5 mL produces fine monodisperse NPs (6.9 ( 0.6 nm) arranged in a self-assembled SL. They turn into an SL of one-dimensional (1D) chains (10.1 ( 0.8 nm) with seed=0.25 mL (Au19). Note this was not the case in any of the previous samples synthesized in the presence of 10-2-10, which clearly shows enhanced influence of hydrophobicity.5 Further half (seed=0.125 mL, Au22) leads to interparticle fusion and eventually produces different shapes (17.0 ( 2.8 nm) such as triangles and hexagons. UV-visible spectra (Figure 6a) of these samples produce significant sharp peaks around 520 nm due to (15) Pastoriza-Santos, I.; Liz-Marzan, L. M. Adv. Func. Mater. 2009, 19, 679.

DOI: 10.1021/la901767c

12703

Downloaded by MT ST VINCENT UNIV on October 28, 2009 | http://pubs.acs.org Publication Date (Web): July 20, 2009 | doi: 10.1021/la901767c

Article

Bakshi

Figure 8. The first horizontal row shows TEM micrographs of Au25, Au26, and Au27 samples synthesized by following a three-step reaction sequence in the presence of 1 mM 14-2-14 at 70 °C. Note the long order of SL formation by NPs and NRs (see details in the text). Panel a shows the corresponding UV-visible spectra of these samples. The second horizontal row shows TEM micrographs of PbS1, PbS2, and PbS3 samples synthesized by using 12-2-12, 14-2-14, and 12-0-12 (see details in the text). (b) Schematic representation of self-assembled behavior of different morphologies driven by surfactant capping layers. (c) Flow diagram showing a systematic dependence of crystal growth on the hydrophobicity of capping surfactant (see details in the text).

SPR. In Step 2, small monodisperse NPs of Au16 undergo clear oriented attachment in Au17 and form 1D nanochains.3i This is, however, not so clear in Au20, but roughly hexagonal shapes (shown in dotted circles) and NRs are considered to be the outcome of oriented attached NPs shown in Au19. Likewise, a trigonal attachment of NPs in Au22 (shown in dotted circle) eventually leads to a triangular shape in Au23. UV-visible spectra (Figure 6b) show a clear red shift in SPR around 520 nm in the order of Au17 (18.7 ( 3.2 nm) < Au20 (55.6 ( 5.1 nm) < Au23 (66.0 ( 12.3 nm) as a result of an increase in the size of polyhedral NPs and their interparticle fusion.12,15 Step 3 produces clear large morphologies for Au18 (59.1 ( 12.8 nm), Au21 (102 ( 21 nm), and Au24 (143 ( 47 nm), which are fully supported by their broad SPR15 in UV-visible spectra (Figure 6c). XRD analysis of samples Au22, Au23, and Au24 (Figure 7) helps to understand the overall growth process. Plots of (111/200) and (111/220) (inset, Figure 7) show a regular fall from Step 1 to 3, suggesting more and more NPs are bound with {100}/{110} facets. A careful analysis of the TEM micrographs of these samples indeed shows this trend from Au22 to Au24, and demonstrates that the oriented attachment begins via {111} planes of adjoining NPs (Au22) and 12704 DOI: 10.1021/la901767c

ultimately leads to the formation of single-crystal (see SEAD, inset Au24) plate-like morphologies16 such as triangles and hexagons (Au24). Comparative Analysis and SL Formation. A comparison among the NPs shown in Figures 1, 4, and 6 leads to a conclusion that both 10-2-10 and 12-2-12 can induce self-assembly behavior that is very much related to their capping efficiencies. Effective capping generates small monodisperse NPs with better chances of SL formation, which is true for 12-2-12 in Au16. It is possible only when several nucleating centers (seed=0.5 mL) are present, which grow uniformly into large surface area, i.e., small monodisperse NPs (Au16). It works through a layer-by-layer deposition of freshly generated atoms on already thermodynamically stable centers1a,17 (morphologies). Relatively weaker capping efficiency3j of 10-2-10 in Au7 triggers nonuniform deposition of atoms and results in interparticle fusion, and, consequently, no SL formation occurs. Two distinct long-order arrangements of NPs are visible in Au16. One is triangular shaped (see dotted triangle), (16) Pastoriza-Santos, I.; Liz-Marzan, L. M. Nano Lett. 2002, 2, 903. (17) (a) Baletto, F.; Mottet, C.; Ferrando, R. Reentrant Phys. Rev. Lett. 2000, 84, 5544. (b) Baletto, F.; Mottet, C.; Ferrando, R. Phys. Rev. B 2001, 6315.

Langmuir 2009, 25(21), 12697–12705

Downloaded by MT ST VINCENT UNIV on October 28, 2009 | http://pubs.acs.org Publication Date (Web): July 20, 2009 | doi: 10.1021/la901767c

Bakshi

and the other is stacks of chains (see dotted rectangle). Triangular arrangement is common with noble metal NPs possessing fcc geometry and hcp ordering, while stacking of 1D chains is typically driven by intermolecular interactions among the capping layers5 (Figure 2c,d). As the S-G reaction proceeds from Step 1 to 2 (i.e., from Au16 to Au17), stacking of 1D chains survives while triangular arrangement diminishes. The same trend is followed from Au16 to Au19 if the number of nucleating centers is reduced (i.e., from 0.5 to 0.25 mL of seed). The absence of triangular arrangement in Au17 and Au19 is mainly attributed to a significant reduction in the monodispersity and promotion of oriented attachment. Further enhancing capping efficiency (as in the case of 14-2-14) helps us to understand this behavior even more precisely. Figure 8 shows TEM micrographs of Au NPs (Au25 (7.1 ( 1.3 nm), Au26 (10.1 ( 1.9 nm), and Au27 (13.1 ( 2.3 nm)) synthesized under exactly identical reaction sequences as shown in Figure 6 by choosing 1 mM 14-2-14 at 70 °C to avoid soft template effect.3g A long-order 1D SL is quite evident in sample Au25 when 0.5 mL of seed solution is used. Both linear and curved arrangements with less extended triangular patterns are visible. 1D side-by-side arrangement of NRs (2.7 ( 0.3) also appears as the amount of seed reduces to 0.25 mL (Au26). NPs grow in size while retaining long order with overlapped 1D layers, even with seed=0.125 mL (Au27). It demonstrates that 14-2-14 possesses greater ability to produce monodisperse NPs at all seed concentrations (i.e., 0.1, 0.05, and 0.025 mL) because of its stronger capping efficiency relative to that of 12-2-12 and 10-2-10. Absorbances due to SPR (Figure 8a) located at 530 further confirm the presence of monodisperse NPs in all cases. No LSPR is observed for 1D stacked NRs of Au26 presumably due to a blue shift caused by side-by-side attachment.10b Intensity difference in the order of Au25 < Au26 < Au27 is the only difference one can see, which means that the NPs are not fused, but they grow in size while retaining SL formation. Thus, 14-2-14 not only helps in controlling the monodisperse behavior of NPs over a wide range of seeding, but also helps them to arrange in SL formation as a result of stronger hydrophobic interactions between the capping layers. It is true not only for Au NPs but for PbS NPs18 as well. Interestingly, similar long-order 1D stacking of SL formation is visible in different monodisperse morphologies such as cubes, spheres, and hexagons. Low hydrophobicity (as in 12-2-12) generates icosahedral SLs of NPs bound with {111} planes, which turn into SLs of cubes bound with {100} facets with increased hydrophobicity (as in 14-2-14). However, further increase in the hydrophobicity (as in 12-0-12) produces SLs of spheres with (18) Bakshi, M. S.; Thakur, P.; Sachar, S.; Kaur, G.; Banipal, T. S.; Possmayer, F.; Petersen, N. O. J. Phys. Chem. 2007, 111, 18087.

Langmuir 2009, 25(21), 12697–12705

Article

almost no facets. The flow diagram in Figure 8c demonstrates how surface passivation determines the overall shape. Relatively weaker surface passivation by 12-2-12 allows {111} planes to be eliminated faster than {100}, whereas the much stronger surface passivation by 12-0-12 even further decreases the facet formation. Thus, high hydrophobicity not only produces monodisperse spherical NPs but also arranges them in a compact SL formation. In addition, symmetry aspects related to tessellations (tilings) also play a significant role in the SL formation. Generally, hexagons, squares, or equilateral triangles lead to regular tessellations with greater possibility of compact SL formation than any other shape. Thus, a compact SL formation is primarily related to the monodisperse nature of such morphologies. Comparing Au25, Au26, and Au27 with PbS1, PbS2, and PbS3 suggests that tessellation is more prevalent in the latter case than in the former because PbS NPs possess a better degree of monodispersity than Au NPs.

Conclusions A simple S-G method has been presented to demonstrate SL formation by choosing two different ways of adding seeds into growth solutions. One is the complete addition of the entire volume of the previous step as a seed into the next step, while, in the other, only a partial amount of the previous solution was used as seed. A systematic evaluation of the shape and size of NPs through different steps of both S-G methods allows one to conclude that SL formation can only be achieved with monodisperse NPs stabilized with a capping surfactant. Additionally, the second method of partial addition provides better probability of SL formation. SL formation is induced by strong hydrophobic interactions among surfactant monolayers, and longer chains with greater hydrophobic strength provide greater ability to gather NPs in a SL structure. Although polydisperse NPs do possess some affinity of ordered arrangement, they lack SL formation because of lattice selective surfactant adsorption. Thus, well-defined NRs and icosahedral NPs easily arrange themselves in a self-assembled short-order 1D stacking through oriented attachment. A large number of seeds and high surfactant concentration (>cmc) favor the formation of small monodisperse NPs arranged in SL, but stronger hydrophibicity (longer tail surfactants) further preserves SL arrangement even over a wide number density of seeds. It is not only restricted to Au NPs; even monodisperse PbS NPs can easily be arranged in an SL formation by choosing strongly hydrophobic surfactants. Supporting Information Available: Size distribution histograms and other information. This material is available free of charge via the Internet at http://pubs.acs.org.

DOI: 10.1021/la901767c

12705