Interparticle Spacing Control in the Superlattices of Carboxylic Acid

Sonny S. Mark, Magnus Bergkvist, Xin Yang, Leonardo M. Teixeira, Parijat Bhatnagar, Esther R. Angert, and Carl A. Batt. Langmuir 2006 22 (8), 3763-377...
0 downloads 0 Views 855KB Size
Download PDF
Langmuir 2004, 20, 10317-10323

10317

Interparticle Spacing Control in the Superlattices of Carboxylic Acid-Capped Gold Nanoparticles by Hydrogen-Bonding Mediation Hiroshi Yao,* Hiroyuki Kojima, Seiichi Sato, and Keisaku Kimura Graduate School of Material Science, University of Hyogo, 3-2-1 Koto, Kamigori-cho, Ako-gun, Hyogo 678-1297, Japan Received July 22, 2004. In Final Form: September 2, 2004 We have demonstrated that carboxylic acid-capped gold nanoparticles were self-assembled to form two-dimensional (2D) and/or three-dimensional (3D) superlattices at an air/water interface in the presence of a bifunctional hydrogen-bonding mediator such as 4-pyridinecarboxylic acid (PyC) or trans-3-(3-pyridyl)acrylic acid (PyA). Transmission electron microscopy revealed a hexagonal close-packed arrangement of nanoparticles in the superlattice with an extension of interparticle spacing. In the 2D superlattices, larger particles produced a higher-quality assembly having long-range translational ordering. Attenuated total reflectance IR (ATR-IR) spectroscopy revealed the presence of hydrogen bonds between the mediator used and the capping agents of carboxylic acid on nanoparticle surfaces. Since the experimentally obtained interparticle separation distance agreed approximately with that obtained by the geometrical model calculations, we conclude that the hydrogen-bonding mediation controlled the interparticle spacing or structure by monomolecular incorporation between adjacent nanoparticles in the superlattices.

Introduction Monolayer-capped nanoparticles of gold and other noble metals have attracted significant attention not only for their established preparation but also for their potential application in nanoelectronics.1 Recent interest has focused on the exploration of the collective properties of assembled arrays of these nanoparticles, since the assembling of nanoparticles into two-dimensional (2D) and/or three-dimensional (3D) ordered structures (nanoparticle superlattices) leads to the formation of new and fascinating materials with tunable, designer-specified optical, electronic, and catalytic properties.2 To construct a macroscopic assembly or superlattice, (i) availability of stable building blocks of nanoparticles with well-characterized uniform particle sizes and shapes and (ii) the suitable surface capping material that allows interparticle assembly are required. Numerous methods for the synthesis of such nanoparticle superlattices have been reported.3 Most of the reported superlattices are made of nanoparticles passivated with organic ligands having a long-chain alkane end, and these constituent nanoparticles are commonly hydrophobic. It has been shown that nanoparticles form spontaneous self-assembly structures when deposited from a solution onto a solid substrate by carefully adjusting the rate of solvent evaporation. The

interactions within the assembly include hydrophobic interaction between long alkyl chains and inherent van der Waals forces.4 The interparticle structure could be controllable by changing the molecular length or structure of the capping alkyl chains, although nanoparticles having a relatively short chain are hard to regularly assemble.3,4 Hence, this type of superlattice formed through slow solvent evaporation might limit our ability to multiply control interparticle structure. On the other hand, we have developed the syntheses of uniform-sized, carboxylic acid-capped, water-soluble gold nanoparticles.5 Application of such hydrophilic nanoparticles is one of the new fields for the construction of 2D or 3D superlattices not by weak van der Waals interactions but by strong hydrogen-bonding and/or electrostatic interactions.6 In particular, hydrogen-bonding interaction is known to be attractive and direction-specific,7 and it offers a flexible and controllable pathway for structural manipulation of assembly,8,9 as is recognized for highly regular structures of polypeptides consisting of amino acids that are building blocks of all living species. In the present study, we then focus the functions of the hydrogen-bonding network in the interparticle structure of nanoparticle assembly. In a superlattice, each nanoparticle is separated from every other one by an insulating barrier of the surface capping agent. Since the barrier thickness directly affects

* Corresponding author. E-mail: [email protected]. (1) (a) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. J. Chem. Soc., Chem. Commun. 1994, 801. (b) Templeton, A. C.; Wuelfing, W. P.; Murray, R. W. Acc. Chem. Res. 2000, 33, 27. (c) Whetten, R. L.; Khoury, J. T.; Alvarez, M. M.; Murthy, S.; Vezmar, I.; Wang, Z. L.; Stephans, P. W.; Cleveland, C. L.; Leudtke, W. D.; Landman, U. Adv. Mater. 1996, 8, 428. (2) (a) Brust, M.; Kiely, C. J. Colloids Surf., A 2002, 202, 175. (b) Sampaio, J. F.; Beverly, K. C.; Heath, J. R. J. Phys. Chem. B 2001, 105, 8797. (c) Zhong, C. J.; Maye, M. M. Adv. Mater. 2001, 13, 1507. (d) Storhoff, J. J.; Mirkin, C. A. Chem. Rev. 1999, 99, 1849. (e) Burghard, M.; Philipp, G.; Roth, S.; von Klitzing, K.; Pugin, R.; Schmid, G. Adv. Mater. 1998, 10, 842. (3) (a) Collier, C. P.; Vossmeyer, T.; Heath, J. R. Annu. Rev. Phys. Chem. 1998, 49, 371. (b) Collier, C. P.; Henrichs, S. E.; Shiang, J. J.; Saykally, R. J.; Heath, J. R. Science 1997, 277, 1978. (c) 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.

(4) Luedtke, W. D.; Landman, U. J. Phys. Chem. 1996, 100, 13323. (5) Chen, S.; Kimura, K. Langmuir 1999, 15, 1075. (6) (a) Yao, H.; Momozawa, O.; Hamatani, T.; Kimura, K. Chem. Mater. 2001, 13, 4692. (b) Yao, H.; Momozawa, O.; Hamatani, T.; Kimura, K. Bull. Chem. Soc. Jpn. 2000, 73, 2675. (c) Kimura, K.; Sato, S.; Yao, H. Chem. Lett. 2001, 372. (d) Sato, S.; Yao, H.; Kimura, K. Physica E 2003, 17, 521. (e) Wang, S. H.; Sato, S.; Kimura, K. Chem. Mater. 2003, 15, 2445. (7) (a) Yaghi, O. M.; Li, H.; Groy, T. L. J. Am. Chem. Soc. 1996, 118, 9096. (b) Jeffrey, G. A. An Introduction to Hydrogen Bonding; Oxford University Press: Oxford, U.K., 1997. (8) (a) Hao, E.; Lian, T. Chem. Mater. 2000, 12, 3392. (b) Hao, E.; Lian, T. Langmuir 2000, 16, 7879. (9) (a) Mayya, K. S.; Patil, V.; Sastry. M. Langmuir 1997, 13, 3994. (b) Kim, Y. J.; Johnson, R. C.; Hupp, J. T. Nano Lett. 2001, 1, 165. (c) Han, L.; Luo, J.; Kariuki, N. N.; Maye, M. M.; Jones, V. W.; Zhong, C. J. Chem. Mater. 2003, 15, 29.

10.1021/la0481501 CCC: $27.50 © 2004 American Chemical Society Published on Web 10/16/2004

10318

Langmuir, Vol. 20, No. 23, 2004

Yao et al.

the exchange coupling or dipole interactions between metal nanoparticles, fine-tuning of the interparticle spacing will be of crucial importance for the electronic structure design of the superlattice.3 Hydrogen-bonding mediation or linking allows us to tailor the interparticle spacing within the superlattices of hydrophilic nanoparticles. In this article, we report an interparticle spacing extension in the hydrophilic gold nanoparticle superlattices by intermolecular mediation using hydrogen-bonding bifunctional agents such as pyridinecarboxylic acids. Due to the fact that hydrogen-bonding interaction effectively works at an air/water (hydrophobic) interface,10 the specific binding mediation through hydrogen bonding will give durable and interparticle-spacing-controlled superlattices of gold nanoparticles in a free-standing form.10,11

Instrumentation. Transmission electron microscopy (TEM) studies were conducted on a Hitachi-8100 electron microscope operating at 200 kV. The superlattice films were scooped on a carbon-coated Cu grid and examined by using TEM. Attenuated total reflectance IR (ATR-IR) spectra were measured by placing the superlattice film on a ZnSe internal reflection trapezoid prism. The film was washed with methanol before the IR measurements to remove undesired impurities. Molecular Modeling. Calculations of molecular geometry were carried out using a MOPAC software of ChemBats3D Pro (CambridgeSoft Corp.). The structures of the “MSA-hydrogenbonding linker-MSA” triad were first drawn and optimized to obtain the minimum energy configurations under a vacuum. The calculated hydrogen-bonding distance was judged by comparing the literature values.7b

Experimental Section

Superlattice Formation from MSA-Capped Gold Nanoparticles. We have already found that MSA-capped gold nanoparticles self-assemble into superlattices or crystals by adding a pure concentrated hydrochloric acid (HCl).6c-e Figure 1a shows a typical TEM image of a 2D superlattice monolayer composed of the gold nanoparticles prepared in the absence of hydrogen-bonding mediators, that is, a TEM micrograph of pure-SL. From the image, the average diameter of the gold core (dcore) and the standard deviation were determined to be 3.3 and 0.50 nm, respectively (Figure 1b). Note that 3D crystals of nanoparticles are the major product in the assemblies.6c-e In the 2D superlattice shown in Figure 1a, a perfectly arranged domain size was < ∼30 nm. The interparticle spacing distance in the superlattice can be obtained by analyzing the Fourier transform (FT) 2D power spectrum of the image. Figure 1c shows the FT pattern of Figure 1a, exhibiting distinct rings. The ring patterns represent the ordering of nanoparticles and the polycrystallinity in the superlattice film. Although the observed ring patterns are somewhat distorted, brighter positions in the ring reflect the most probable spacing periodicity of the nanoparticle arrays. From the image analysis, therefore, the center-to-center distance between the nearestneighbor particles (superlattice constant (aSL)) was calculated to be ∼4.5 nm. Since the 2D superlattice was composed of 3.3-nm nanoparticles in core diameter, we could estimate an interparticle gap (Dgap ) aSL - dcore) of ∼1.2 nm. This separation distance is approximately twice longer than that of the MSA molecule (∼0.6 nm), consistent with the previous results of the superlattice constant of the 3D nanoparticle crystals.6d,e In Figure 2a or 3a is a typical TEM image of the large area of gold nanoparticle 2D superlattices prepared in the presence of PyC or PyA (namely, PyC-SL or PyA-SL), respectively, representing a long-range translational ordering in the superlattices. Figure 2b or 3b shows the magnified image of the superlattices surrounded by the square in Figure 2a or 3a, respectively. In both cases, it can be clearly seen that hexagonal close-packed structures are formed. Moreover, the interparticle spacing seems to be wider compared to that in the pure-SL. The mean particle gold core diameter and the standard deviation were determined to be 5.6 and 0.73 nm for the PyC-SL (Figure 2c), respectively, and 4.3 and 0.58 nm for the PyA-SL (Figure 3c), respectively. Note that no superlattices were generated in the absence of PyC or PyA under the present conditions. Considering that the car-

Chemicals. Hydrogen tetrachloroaurate tetrahydrate (HAuCl4‚ 4H2O, 99%), sodium borohydride (NaBH4, >90%), mercaptosuccinic acid (HS-CH(COOH)-CH2(COOH), abbreviated as MSA, 97%), 4-pyridinecarboxylic acid (GR grade, abbreviated as PyC), methanol (GR grade), and ethanol (GR grade) were received from Wako Pure Chemicals and used without further purification. trans-3-(3-Pyridyl)acrylic acid (abbreviated as PyA, 99%) was purchased from Aldrich and used as received. Either PyC or PyA is used as a bifunctional hydrogen-bonding mediator.7b Pure water was obtained by an Advantec GS-200 automatic water-distillation supplier. Synthesis of MSA-Capped Gold Nanoparticles. Watersoluble, mercaptosuccinic acid (MSA)-capped gold nanoparticles were prepared by a similar method to the one reported earlier.5 Briefly, 0.5 mmol of HAuCl4 dissolved in water (0.121 M) and 1.5 mmol of MSA were at first mixed in methanol (100 mL), followed by the addition of a freshly prepared 0.2 M aqueous NaBH4 solution (25 mL) at a rate of 2.5 mL/min under vigorous stirring. A dark-brown crude precipitate was produced. The precipitate was then thoroughly washed with water/ethanol (1/4) and methanol. Finally, after dissolving the precipitate into water (10 mL), dialysis was continued for 24 h, followed by freeze-drying to obtain a nanoparticle powder. The mean particle core diameter and the standard deviation of the as-prepared gold nanoparticles are 3.4 and 0.75 nm, respectively. Synthesis of Gold Nanoparticle Superlattices: (i) In the Absence of Hydrogen-Bonding Mediators. The as-prepared MSA-capped nanoparticle powder (4.0 mg) was dissolved in distilled water (2.0 mL), then HCl was added into the nanoparticle solution, and then this solution was stored in a closed vessel. The concentration of HCl in the solution was ∼0.4 M. After 3-4 days of storage under room temperature, crystallization took place at the air/water interface, showing mirrorlike light reflection due to the formation of gold nanoparticle superlattices. We abbreviate this superlattice as “pure-SL”. (ii) In the Presence of Hydrogen-Bonding Mediators. We selected different pyridinecarboxylic acids as hydrogenbonding mediators that might be incorporated between the MSA moieties on gold nanoparticles. To construct gold nanoparticle superlattices or assemblies in the presence of a hydrogen-bonding mediator, the aqueous solution of MSA-capped gold nanoparticles containing one of the hydrogen-bonding mediators (PyC or PyA) and HCl was stored in a sealed vessel for about 3-4 days to allow self-assembling of gold nanoparticles at the air/water interface. “PyC-SL” or “PyA-SL” is the abbreviation for the superlattices that were formed in the presence of PyC or PyA, respectively. The total amounts of the contained MSA-capped gold nanoparticle powder and the hydrogen-bonding mediator were 4.0 mg and 7.2 × 10-5 mol, respectively.12 Note that the HCl concentration was ∼0.04 M, ∼1/10 lower than that prepared in the absence of hydrogen-bonding agents. The final pH of the solutions was in the range 0.7-1.8. (10) Hoshi, H.; Sakurai, M.; Inoue, Y.; Chujo, Y. R. J. Chem. Phys. 1987, 87, 1107. (11) Yao, H.; Kojima, H.; Sato, S.; Kimura, K. Chem. Lett. 2003, 32, 698.

Results and Discussion

(12) On the basis of the elemental analyses of the MSA-capped gold nanoparticle powder (ref 6a), 4.0 mg of the powder contains -COOH groups of ∼7.2 × 10-6 mol. Therefore, the added hydrogen-bonding mediators are sufficient to link the surface functional groups of nanoparticles.

Carboxylic Acid-Capped Gold Nanoparticles

Langmuir, Vol. 20, No. 23, 2004 10319

Figure 1. (a) TEM micrograph of a 2D superlattice monolayer of MSA-capped gold nanoparticles prepared in the absence of hydrogen-bonding mediators (pure-SL). (b) Particle size distribution of the constituent gold nanoparticles. The average diameter of the gold core is 3.3 nm. (c) Fourier transform (FT) 2D power spectrum of the image from part a.

boxylic acid and pyridine groups are good hydrogenbonding partners,8 the formation of 2D superlattices at the air/water interface suggests that PyC or PyA would mediate the assembling through hydrogen-bonding interaction between MSA-capped gold nanoparticles.11 Interparticle attraction due to van der Waals dispersional interactions is also considered to contribute to the selfassembling.13 In the 2D PyC-SL and PyA-SL films, an increase in particle core size was observed compared to that of the pure-SL or original nanoparticles. Furthermore, the larger the particle size is, the higher the array quality becomes. At present, two possible reasons are considered: (i) The observed regions do not reflect ensemble average properties of original nanoparticles (namely, statistical uncertainty).14 Actually, we could see the variation in the core size of nanoparticles that produced 2D superlattices when a different area from that in Figure 3a was observed for the PyA-SL film (see the Supporting Information). It has been reported that larger gold nanoparticles readily organize into superstructures due to size-dependent van der Waals dispersional attractive forces (size segregation effect).13 When it works effectively together with the strong hydrogen-bonding interaction during the superlattice formation, superlattices composed of large-sized nanoparticles should be easily formed. Although the fraction (13) Ohara, P. C.; Leff, D. V.; Heath. J. R.; Gelbart, W. M. Phys. Rev. Lett. 1995, 75, 3466. (14) Sato, S.; Yao, H.; Kimura, K. Jpn. J. Appl. Phys. 2004, 43, L927.

of particles with a core diameter of ∼5.6 or 4.3 nm was relatively small within the original nanoparticles, the size increase would be explainable because the observed area shown in Figures 2 and 3 comes from the specific interfacial regions where long-range, well-ordered arrays of nanoparticles have emerged.15 (ii) Growth in the particle size would occur during the superlattice formation. However, since the pure-SL did not exhibit the increase in the mean particle size of the constituent nanoparticles, this situation would be less plausible. Figure 2d or 3d shows the FT patterns of the magnified TEM image, which also demonstrates the long-range nanoparticle ordering and its hexagonal symmetry. According to the spot positions, the superlattice constant (aSL) is calculated to be ∼7.5 or ∼6.1 nm for the 2D PyC-SL or PyA-SL, respectively. Considering that the respective superlattice was composed of 5.6- or 4.3-nm nanoparticles in core diameter, Dgap ()aSL - dcore) was determined to be ∼1.9 nm (PyC-SL) or ∼1.8 nm (PyA-SL). These data related to the interparticle structure are summarized in Table 1. Because the obtained Dgap value is about 3 times longer than the length of the MSA modifier, it is expected that hydrogen-bonding mediators (molecular length of ∼0.6 nm in both cases) are incorpo(15) During the preparation, distribution of the nanoparticles was observed: Superlattices were formed at the air/water interface, whereas some portions of the nanoparticles were still dissolved in water and some were precipitated. It is reasonable that particle size differences might be present in each of the portions.

10320

Langmuir, Vol. 20, No. 23, 2004

Yao et al.

Figure 2. (a) TEM micrograph of the large area of a 2D superlattice prepared in the presence of PyC (PyC-SL). (b) Magnified TEM image of the superlattice surrounded by the square shown in part a. (c) Particle size distribution of the constituent gold nanoparticles. The average diameter of the gold core is 5.6 nm. (d) FT 2D power spectrum of the image from part b.

rated in the nanoparticle assembly. The results also indicate that the interparticle spacing is extendable with hydrogen-bonding molecules. It is worth noting that 3D nanoparticle crystals (or superlattices) were also formed together with the 2D superlattices, although the number of the 3D crystals is relatively small. In Figure 4 are low-magnification images of MSA-capped gold nanoparticle 3D crystals assembled in the presence of PyC and PyA, respectively. The 3D crystals are well-faceted and hexagonally shaped. Figure 5a represents a magnified image of a superlattice edge of the PyA-SL. This superlattice shows remarkable in the perfect arrangement of gold nanoparticles. A closer look at Figure 5a reveals a triangular particle pattern (see the ellipsoidal region in the image). Such a triangular shape contrast is only typical for a hexagonal close-packed (hcp) ordering, as theoretically proven by Bentzon et al.16 Similar patterns have also been observed by Harfenist et al.17 and Stoeva et al.3c in a hcp structure of nanoparticles. When the hcp structure of the 3D superlattice is assumed here, the schematic representation of the particle ordering viewed along the [0001] direction of the superlattice can be shown in Figure 5b, which corresponds to the contrast sequence observed in Figure 5a. According to the image analyses, the superlattice constant (aSL) was found to be ∼5.9 nm ()5.1 × 2/x3 nm). To precisely determine the mean core diameter of the constituent gold nanoparticles, (16) Bentzon, M. D.; Tholen, A. R. Ultramicroscopy 1991, 38, 105. (17) Harfenist, S. A.; Wang, Z. L.; Whetten, R. L.; Vezmar, I.; Alvarez, M. M. Adv. Mater. 1997, 9, 817.

we counted the particles distributed in and by the 3D particle crystal edges and then obtained it to be ∼3.9 nm. Therefore, the interparticle separation gap (Dgap) is ∼2.0 nm, similar to that in the 2D superlattices even though we take some errors into consideration. We believe that the hydrogen-bonding mediator is also incorporated between the adjacent nanoparticles in the 3D crystals. Characterizations of the Superlattice Films by ATR-IR Spectroscopy. Hydrogen-bonding mediation in the superlattice film was identified by ATR-IR spectroscopy. Figure 6a shows the IR absorption spectrum of a superlattice film of PyC-SL along with that of pure PyC and MSA. The spectrum of the superlattice film has characteristic features of both PyC and MSA. A broad band at around 2800-3500 cm-1 is assigned to the carboxylic acid OsH stretching originated from MSA molecules.5 The peaks at 1603 and 1406 cm-1 are assigned to ring vibrations of pyridine groups.8,11 The broad absorption bands around 1900-2000 and 2400-2500 cm-1 are attributed to the OsH stretching bands based on hydrogen bonding between pyridine and carboxylic acid groups in a splitting pattern.18 Note that pure PyC possesses similar splitting bands. Therefore, in addition to the detection of hydrogen-bonded OsH groups, observation of the peak shift (1723 cm-1) of the CdO stretching vibration was decisive evidence for the interaction between MSA and PyC moieties; the peak position was different from that of pure PyC (1710 cm-1). We conclude that strong (18) Lee, J. Y.; Painter, P. C.; Coleman, M. M. Macromolecules 1988, 21, 954.

Carboxylic Acid-Capped Gold Nanoparticles

Langmuir, Vol. 20, No. 23, 2004 10321

Figure 3. (a) TEM micrograph of the large area of a 2D superlattice prepared in the presence of PyA (PyA-SL). (b) Magnified TEM image of the superlattice surrounded by the square shown in part a. (c) Particle size distribution of the constituent gold nanoparticles. The average diameter of the gold core is 4.3 nm. (d) FT 2D power spectrum of the image from part b. Table 1. Comparison of the mean core diameter (dcore), superlattice constant (aSL), and interparticle separation gap (Dgap) for the 2D superlattice films of pure-SL, PyC-SL, and PyA-SL superlattice

dcore

aSL

Dgap (nm)

pure-SL PyC-SL PyA-SL

3.3 5.6 4.3

4.5 7.5 6.1

1.2 1.9 1.8

hydrogen bonding between MSA on the surface of gold nanoparticles and PyC is present in the superlattice films. It should be noted that a distinctive new band at 1505 cm-1 that comes from the protonation of the pyridine unit could be detected,8,19 implying that the pyridine units are incorporated as a pyridinium form in the superlattice. A similar spectroscopic assignment can be made for the PyA-SL film. Figure 6b shows the IR spectrum of the PyA-SL film as well as pure PyA. The spectrum of PyA-SL exhibits characteristics of both PyA and MSA accompanied by a slight shift of the typical bands. The several peaks observed in the region of 1397-1612 cm-1 are ascribed to ring vibrations of pyridine units in PyA.8 Particularly, the band at 1612 cm-1 that is specific to the PyA-SL film can be assigned to the protonation of the pyridine unit in this case. The peak at 1644 cm-1 is due to the CdC stretching mode of the acrylic acid unit. The (19) Antonietti, M.; Wenz, E.; Bronstein, L.; Seregina, M. Adv. Mater. 1995, 7, 1000.

OsH stretching bands originated from hydrogen bonding between pyridine and carboxylic acid groups could be observed at around 1900-2000 and 2400-2500 cm-1 in the splitting pattern.18 The CdO stretching vibration in the PyA-SL film exhibited a peak at 1698 cm-1, different from that of PyA (1705 cm-1). The phenomenon is similar to that for the PyC-SL film, as described above, indicating that the PyA molecule also mediates the interparticle assembling through hydrogen-bonding interactions. A pure-SL exhibited the CdO stretching band at a different position (1620 cm-1) from that in the PyC-SL and PyA-SL films.20 Such a low-energy peak position suggests the existence of a polymeric hydrogen-bonded sCOOH group probably caused by water clusters in the superlattices.21 For PyC-SL and PyA-SL films, the CdO stretching modes appeared at higher energies. The result indicates that a hydrogen-bonding mediator directly linked to the MSA carboxylic moieties in the superlattices.9c Therefore, we conclude that the hydrogen-bonding mediators interconnect with the adjacent gold nanoparticles to form high-quality 2D and/or 3D nanoparticle superlattices. Model of the Interparticle Structure in the Hydrogen-Bonding Mediated Nanoparticle Superlattices. Hydrogen-bonding interaction between pyridine and (20) Wang, S.; Yao, H.; Sato, S.; Kimura, K. J. Am. Chem. Soc. 2004, 126, 7438. (21) Zheng, W. X.; Maye, M. M.; Leibowitz, F. L.; Zhong, C. J. Anal. Chem. 2000, 72, 2190.

10322

Langmuir, Vol. 20, No. 23, 2004

Yao et al.

Figure 4. Low-magnified TEM image of a 3D superlattice crystal of MSA-capped gold nanoparticles prepared in the presence of (a) PyC or (b) PyA. Hexagonal shapes with crystal habit can be seen in these images.

Figure 6. (a) IR absorption spectrum of PyC-SL along with that of pure PyC and MSA. (b) IR absorption spectrum of PyA-SL along with that of pure PyA.

Figure 5. (a) Magnified TEM image of a 3D superlattice crystal edge. The triangular shape contrast typical for hcp ordering can be seen in the ellipsoid shown in the image. (b) Schematic representation of the hcp nanoparticle ordering derived from the image from part a.

carboxylic acid groups is influenced by the degree of dissociation of the acids used, which directly correlates to the solution pH. The pH of the solutions in which the nanoparticle superlattices were formed was measured to be 0.7-1.8. According to the literature,22 the pKa of succinic acid in aqueous solution is 4.00 and 5.24, whereas that of (22) Martell, A. E.; Smith, R. M. Critical Stability Constants; Plenum Press: New York, 1974; Vol. 1-3.

PyC is 1.79 and 4.78 (these acids are dibasic acids). For PyA, we could not obtain the pKa value; however, considering that the pKa of benzoic acid or acrylic acid, which possesses structural similarity to PyC or PyA, respectively, shows a similar value (∼4.2),22 PyA can be assumed to possess similar pKa values to PyC. At pH 0.7-1.8, therefore, the pyridine group in PyC or PyA should be protonated and other carboxylic acid groups in PyC, PyA, or MSA are scarcely dissociated. The protonation of the pyridine units has been supported by the IR absorption results described in the preceding section. By taking these facts into account, we calculated an optimized stable geometry of “triad”, which shows that one hydrogen-bonding mediator (PyC or PyA) is incorporated between two MSA molecules through hydrogen bonding. Note that the capping MSA molecules are densely attached onto the gold nanoparticle surface (molecular area of 0.152 nm2/molecule),5 so that the intrusion of the mediator molecules inside the MSA corona region would be ruled out. Therefore, we employed a triad model based on the experimentally obtained interparticle distances (approximately 3 times longer than the length of MSA). The results are shown in Figure 7. For comparison, the geometry of the “MSA-MSA” diad is also shown. These molecular geometries describe that the sulfur-sulfur distance which corresponds to Dgap (arrow line in Figure 7b or c) is 1.93 or 1.86 nm for PyC-SL or PyA-SL,

Carboxylic Acid-Capped Gold Nanoparticles

Langmuir, Vol. 20, No. 23, 2004 10323

Figure 8. Schematic diagram of two adjacent MSA-capped gold nanoparticles linked by one PyC molecule via a hydrogenbonding interaction. A stiff capping MSA region is represented as a gray corona.

bonding and attractive dispersional interactions between constituent nanoparticles, well-ordered superlattices are formed at the air/water interface. Our finding shows that interparticle spacing or morphology can be controlled by manipulating interactions through hydrogen-bonding mediators. Future research in our group includes the arbitrary controls of the core size of constituent nanoparticles and the interparticle spacing in the nanoparticle superlattices by incorporating various-sized hydrogenbonding mediators.

Figure 7. Optimized molecular geometry of (a) MSA-MSA, (b) MSA-PyC-MSA, and (c) MSA-PyA-MSA. The arrow lines in parts b and c represent the distance between sulfur-sulfur atoms.

respectively. (The hydrogen-bond distance, H‚ ‚ ‚X, where X denotes N or O, is reasonably obtained to be ∼0.16 nm, consistent with values listed in the literature.7b) The distance agrees well with Dgap obtained from Figure 2 or 3, implying that the bridging through hydrogen bonding between the mediator and MSA in the 2D superlattices is reasonable. The incorporation of one mediator molecule between two MSA moieties is probably due to the energy differences in hydrogen-bonding interactions; the hydrogenbonding interaction between the mediator and MSA might be stronger than that between two mediator molecules. In the assembling of alkanethiolate-capped gold nanoparticles in the presence of R,ω-mercapto-alkanoic acids, Zhong and co-workers proposed a squeezed interparticle spatial model involving hydrogen bonding at the carboxylic acid groups and cohesive attractive van der Waals interactions through interdigitation of the capping alkanethiolate molecules.9c Compared to the similarity in this squeezed packing model, we believe that hydrogenbonding mediators would be statistically incorporated into the energetically preferable positions between the MSA moieties on the gold nanoparticles. A schematic diagram is shown in Figure 8, which explains the extension of interparticle spacing. A stiff capping MSA region is represented as a gray corona. Due to both hydrogen-

Conclusion In conclusion, mercaptosuccinic acid-capped gold nanoparticles were self-assembled to form 2D and/or 3D superlattices at an air/water interface in the presence of a bifunctional hydrogen-bonding mediator such as 4-pyridinecarboxylic acid (PyC) or trans-3-(3-pyridyl)acrylic acid (PyA). TEM measurements revealed that the particles were regularly assembled in a hexagonally close-packed style and the extension of the interparticle spacing was recognized. Larger particles produced higher-quality 2D superlattices having long-range translational ordering. IR spectroscopy revealed the presence of the hydrogen bonds between the mediator used and the functional capping agents on nanoparticle surfaces. These results suggest that both hydrogen-bonding and van der Waals dispersional interactions are responsible for the assembling forces at the air/water interface. The observed interparticle separation distances were consistent with those found by geometrical calculations; therefore, the hydrogen-bonding mediators expanded the interparticle spacing by monomolecular incorporation between adjacent nanoparticles in the superlattices. The fabrication strategy would be significant because the controllable interparticle morphology with a regulation in the superlattices might be utilized for manipulating electron transport based on the designed band structure. Acknowledgment. We are indebted to support in part by Grants-in-Aid for Scientific Research (B:13440212) and for Scientific Research in Priority Areas: Application of Molecular Spins (15087210) from MEXT. Supporting Information Available: TEM image of a 2D superlattice of gold nanoparticles (PyA-SL film) whose observation area was different from that shown in Figure 3a. This material is available free of charge via the Internet at http://pubs.acs.org. LA0481501