Photochemical Strategies for the Seed-Mediated Growth of Gold and

Nov 6, 2012 - Gold nanoparticles (AuNP) can be used as seeds for the synthesis of larger ... Seed-Mediated Synthesis of Gold Nanoparticles of Controll...
0 downloads 0 Views 393KB Size
Download PDF
Article pubs.acs.org/Langmuir

Photochemical Strategies for the Seed-Mediated Growth of Gold and Gold−Silver Nanoparticles Katherine L. McGilvray,† Chiara Fasciani,† Carlos J. Bueno-Alejo, Rachel Schwartz-Narbonne, and Juan C. Scaiano* Department of Chemistry and Centre for Catalysis Research and Innovation, University of Ottawa, 10 Marie Curie, Ottawa, Ontario K1N 6N5, Canada S Supporting Information *

ABSTRACT: Gold nanoparticles (AuNP) can be used as seeds for the synthesis of larger AuNP of controllable size with narrow size distribution by photochemical reduction of additional Au(III) using water-soluble benzoins or H2O2 as sources of reducing radicals. Further, beyond simply enlarging the AuNP, it is possible to add a shell of another metal, such as silver, leading to Au/Ag core−shell structures with controllable dimensions for both core and shell. This strategy illustrates the fine spatial and temporal control achievable using clean photochemical techniques without the addition of hard surface ligands often necessary to control the size and structure of gold−silver nanostructures. The mild nature of the surface coverage makes these nanomaterials ideal for further surface modification.

■

INTRODUCTION When exploiting metal nanoparticles in a particular application, from catalysis to chemical sensing to drug delivery, a uniformity in nanoparticle surface reactivity is critical, underlining the demand for colloids composed of identical chemical composition, particle size, and morphology.1−3 One particular approach to ensure such a uniform distribution is referred to as seed-mediated growth, where small particles of a given composition (gold in our case) are placed in a solution of metal ions and a suitable mild reducing agent for particle enlargement.4,5 Seeding strategies, such as those devised by Natan et al.,6 take advantage of surface catalysis to generate larger particles. Small nanoparticles are added to a “growth solution” of metal salt and mild reducing agent such that the surface of the metal particle catalyzes the reduction of the metal salt occurring more readily at the surface than in the bulk solution. This stepwise strategy effectively separates the nucleation and growth period of nanoparticle formation, resulting in enhanced particle monodispersity relative to synthetic methods where concentrations of reagents and stabilizers must be adjusted to obtain a desired size.7 The choice of reducing agent is also essential for effective growth; for strong reducing agents can favor the additional generation of new nuclei in solution, a process termed secondary nucleation, and thus leads to an undesirable size distribution. The choices of reducing agent and reaction conditions are paramount to favor growth over the formation of new nucleation centers.4−6,8−11 On occasion, photochemical seeding has also been explored.12,13 Typical reducing agents employed in seeding studies include ascorbic acid, hydrazine, and hydroxylamine hydrochloride; results depend on a range of easily controllable parameters, © 2012 American Chemical Society

such as reagent concentration, the presence of additives, and the order and rate of addition of reagents, where some variability is common.5,9,14 In this paper we report two methodologies for the growth of gold nanoparticles from seeds with additional layers of either gold or silver, where a photoactivated reducing agent is employed for both the initial nanoparticle synthesis and their seeded growth. In both cases, surface protection is minimal and only involves mild stabilizers that can be easily substituted. The first technique involves the photogeneration of a key reducing agent in the presence of seed; 2-hydroxylpropyl radical, or the “ketyl” radical from acetone, produced by photolysis of the substituted benzoin commercialized as Irgacure-2959 (I-2959). The ketyl radical can be viewed as a caged electron, ready to be supplied when a suitable acceptor, such as our easily reduced metal ions, is available (Scheme 1). We have recently shown that this electron transfer in the formation of initial seed should be viewed as an example of multisite proton coupled electron transfer (PCeT), where the presence of a receptor for the proton is critical.15 This photochemical strategy has been successfully used in our laboratory to produce several types of metal nanoparticles.10,16−19 A second methodology takes advantage of the use of H2O2 as a source of a photoactivated reducing agent, extending a methodology that we recently reported.20 Note that the H2O2 method can only be used with AuNP, as AgNP is reactive toward H2O2.21 Traditional seeding methodologies require the presence of a weak reducing agent for the initial reduction of Au(III) to Received: July 12, 2012 Revised: October 20, 2012 Published: November 6, 2012 16148

dx.doi.org/10.1021/la302814v | Langmuir 2012, 28, 16148−16155

Langmuir

Article

a photoreactor to yield spherical particles of 10.3 ± 3.4 nm.16 SEM images of these nanoparticles are shown in Figure S2. Seed-Mediated Growth. Aliquots of AuNP were added to growth solutions containing metal salt and reducing agent following four different protocols (vide inf ra). One-pot seeding of either monometallic or bimetallic core/shell NP were prepared with addition of AuNP seed to growth solutions containing I-2959 and metal salt followed by UVA photolysis. A sequential seeded growth method was performed using either I-2959 or H2O2 as a reducing agent. AuNP seeds (prepared solely using the I-2959 method) were added to a growth solution with 15 min of UVA photolysis and 10 min of ripening, where the newly enlarged AuNP were used as an aliquot for the next generation of growth. Characterization. Absorbance spectra of samples at various stages of growth were recorded using a Cary 50 UV−vis spectrophotometer. Zeta potentials were measured using a Malvern Zetasizer Nano NS. Particle sizes were determined by analysis of representative electron microscopy images. SEM images were acquired using a JSM-7500F field emission scanning electron microscope from JEOL Ltd. TEM micrographs were collected with a high-resolution transmission electron microscope (HRTEM); TEM and HRTEM studies were performed using a JEOL JEM-2100F field emission transmission electron microscope equipped with an ultrahigh-resolution pole piece operating at 200 kV. In preparation of samples for electron microscopy, 10 μL aliquots of particle suspensions were deposited onto 400 mesh copper grids, followed by solvent evaporation.

Scheme 1. Photolysis of I-2959 and Oxidation of Its Radical Products in the Synthesis of Metal Nanoparticles

Au(I). While the reduction potential of the Auatom/Au(I)aq is −1.5 V vs NHE, the ability of the particle surface to be sacrificially oxidized to reduce an adsorbed Au(I) ion is in fact quite favorable, as indicated by the high reduction potential for the surface process Aumetal/Au(I)aq which is +1.68 vs NHE.22 While typical thermal methods for seeding require the use of a weak reducing agent for the prevention of secondary nucleation, reducing species with both strong and weak reduction potentials, i.e., the ketyl radical ((CH3)2CO, H+/ (CH3)2C•OH, E0 = −1.80 V vs NHE) or the hydroperoxyl radical (O2, H+/HOO•, E0 = −0.05 V vs NHE), both provide a system where surface-mediated growth is controlled by the reducing species, metal salt concentration, or irradiation intensity where secondary nucleation is essentially negligible. HAuCl4 is the most common source of Au(III), where reduction to Au(0) involves several reduction and disproportionation steps.16,17,23 Following the initial reduction, atoms nucleate into small clusters, followed by a growth period. This growth can either start new nucleation centers or add particles to an existing one if available. When the chemistry of Scheme 1 is employed to produce new particles, these particles are mildly stabilized by Cl− and 4-hydroxyethoxybenzoic acid (HEBA),24 a carboxylic acid product of I-2959 formed when the photochemistry takes place in aerated solutions (Scheme 1). In this work, we report strategies to steer the newly formed atoms toward seed-mediated growth rather than new nucleation; photochemistry can provide fine control of these processes.

■

■

RESULTS A number of studies were carried out using thermal reduction methods employing ascorbic acid and hydroxylamine hydrochloride as reducing agents, which were reported in part in an earlier contribution.14 The experiments provided herein describe different strategies for particle enlargement, highlighting the advantages in this photochemical method. For comparison to a thermal method, the one-pot seed-mediated growth employing NH2OH−HCl is provided in the Supporting Information (see Figure S1). i. One-Pot Seed-Mediated Growth of AuNP with Variation in Seed Concentration. Photochemically generated AuNP seeds, prepared as described above, were added to a growth solution containing HAuCl4 and a photoinitiator as a source of reducing species, followed by 30 min UVA irradiation at 36 W/m2. The reducing ketyl radicals are generated as indicated in Scheme 1 and reported in earlier contributions.16,17,25 The content of the growth solution was held constant with 0.25 mM HAuCl4 and 0.4 mM reducing agent, while the concentration of AuNP seed was varied. Control experiments were performed using NH2OH−HCl as a reducing agent to compare the current method to the thermal method.9,14 A variation in particle size was achieved by changing the concentration of seeds added, with increased seed concentrations leading to reduced growth of individual particles. Particle growth was easily identified with a red shift of the surface plasmon band maximum of the solution and later confirmed by SEM. The expected diameter of AuNP following growth can be calculated with eq 1. This equation assumes particles as perfect spheres and that all the Au(III) in the growth solution adds as Au(0) to the existing seeds.26

EXPERIMENTAL METHODS

Materials. HAuCl4·3H2O, AgNO3, sodium citrate, and H2O2 (30 wt %) were used as received (Sigma-Aldrich). The photoreducing ketone Irgacure-2959 (I-2959 from Ciba Specialty Chemicals) was recrystallized from ethanol. All aqueous samples were prepared with deionized water, obtained from an academic Millipore Milli-Q system (resistivity 18.2 MΩ at 25 °C). Irradiation. Solutions containing HAuCl4 and radical precursors were prepared in fused silica cuvettes or in polypropylene 24-well plates (BD Falcon, non-tissue-culture-treated) and irradiated in a Luzchem LZC-4 photoreactor with 14 UVA lamps (36−70 W/m2 for UVA, λ > 300 nm, centered at 350 nm). Synthesis of AuNP Seeds. AuNP was synthesized by a photochemical approach developed in our group.16,17 Briefly, aqueous solutions containing 0.33 mM HAuCl4 and 1.0 mM I-2959 are irradiated using 14 UVA lamps for 15 min under aerated conditions in

dfinal = dseed 3

[Au 0] + [Au 3 +] [Au 0]

(1)

Table 1 shows a direct comparison of the growth of AuNP seeds, initially 10.3 ± 3.4 nm in diameter, using NH2OH−HCl as a mild reducing agent and I-2959 as a source of reducing 16149

dx.doi.org/10.1021/la302814v | Langmuir 2012, 28, 16148−16155

Langmuir

Article

solution and photoirradiated to induce growth. The solutions were prepared in one of two ways as highlighted in Scheme 2,

Table 1. Comparison of Particle Diameter As Measured by SEM Analysis for the Seed-Mediated Growth of AuNPa particle size (dfinal, nm) seed concn as [Au0], μM 6.61 4.53 1.65 0.45 0.33

predicted 35 40 55 85 94

± ± ± ± ±

b

11 13 18 28 31

NH2OH−HCl method

Scheme 2. Sequential Photochemical Seeding Employing a Photoinitiator as a Source of Reducing Radicalsa

I-2959 method 48 ± 15

36 ± 16 68 ± 16 167 ± 45 98 ± 23

AuNP seeds (0.33−6.61 μM) added to growth solutions containing 0.25 mM HAuCl4 and 0.4 mM reducing agent (NH2OH−HCl or I2959). bPredicted diameter calculated from a seed particle diameter (dseed) of 10.3 ± 3.4 nm using eq 1. a

a

An aliquot of AuNP prepared by the I-2959 method was used as seeds for addition to a growth solution, followed by UVA photolysis. Enlarged particles were then used as seeds for subsequent photolysis and further seeding. X = A when 1/8th seed volume was added to 7/ 8th growth solution volume. X = B when 1/10th seed volume was added to 9/10th growth solution volume.

ketyl radicals. We note experimental and calculated diameters are in reasonable agreement. Further, the fact that the sizes obtained with the I-2959 method are slightly larger than calculated values suggests that the substoichiometric amounts of I-2959 did not prevent the reduction of AuCl4− to Au(0), although we note that relative errors in the AuNP volume are larger than diameter errors and allow for some gold (particularly as Au+) to remain in solution following incomplete reduction. Table 1 provides a comparison of a traditional thermal method with the photochemical approach employing I-2959. Almost all measured particle sizes fall within the range of the predicted particle size, except for the NH2OH−HCl method at higher seed concentration. The data for the I-2959 method indicate improved monodispersity among most samples. At high seed concentration, however, the I-2959 photoseeding and the NH2OH−HCl method showed similar results. Under low seed concentration, the I-2959 offered better monodispersity of larger particle size (low seed concentration). Importantly, while AuNP were formed in the absence of seed using 0.4 mM I-2959 in the growth solution (see Figure S3), high monodispersity of enlarged particles is achieved at low seed concentration. This preference for particle enlargement over the formation of new particles at low seed concentration supports the surfacecatalyzed growth of newly reduced metal atoms on the growing particle surface. Occasional prisms are present in the I-2959 AuNP samples, yet the thermal method displayed significantly greater populations of plates, prisms, and rods. This photochemical control over shape and size is further emphasized in UV−vis spectra and SEM images in Figure S3. The experimental method was later optimized by introducing two changes: (1) minimizing the concentration of reducing agent by using stoichiometric or substoichiometric concentration relative to Au3+ and (2) employing a sequential photoseeding method. Stoichiometric ratios would require three times as much I-2959 as Au(III); to our surprise, the best results were obtained with [I-2959] ≤ [Au(III)]. We will speculate on the chemistry underlying this behavior in the Discussion section. ii. Subsequent Photochemical Growth with Variation in Seed Concentration. Photochemically prepared seeds of 10.3 nm diameter were added to a growth solution containing 0.5 mM HAuCl4 and either 0.25 mM I-2959 or 0.5 mM H2O2. These two reducing radical sources provide a contrast of reagents with strong and weak reduction potentials, respectively. Following particle seeding and ripening, an aliquot of the enlarged particles was subsequently added to a replicate growth

building on a protocol established by Henglein.27 In one approach, 1/8th the volume of seed solution was added to 7/ 8th volume of growth solution for a total volume of 2.0 mL. Samples prepared with this protocol are noted as A and their subsequent seeding stages as A′, A″, etc. In a second approach, 1/10th seed volume was added to 9/10th growth solution volume for a total volume of 2.5 mL, and such samples and their seeds were noted B and B′, B″, etc., respectively. Variation in the seed:growth solution volume ratio allows further control over particle size. Samples were irradiated (36 mW/m2 of UVA light) for 15 min and ripened for 10 min prior to subsequent growth. The ripening period ensured that particle growth from the previous step in the synthesis was complete. The sequential photoseeding method employing I-2959 as a radical source was performed, and the resulting absorption spectra and representative SEM images for the enlarged particles following protocol A are shown in Figure 1. SEM images of resulting particles prepared following protocol B are available in the Supporting Information, along with their corresponding absorption spectra (Figures S4 and S6, right). A bar graph is also provided in Figure 3 (vide inf ra) combining the results for methods A and B to illustrate the ability to tune the solution volumes for desired particle size in the range 10− 150 nm. While seeding appeared to be effective for generation of particle diameters in the 20−90 nm range, seeding for larger particles tended to generate polymorphic samples including aggregated spherical particles. Further tuning of irradiation time and seed concentration along with in situ addition of stabilizer may enable to synthesis of particles >100 nm with reasonable polydispersity and consistent morphology. We have recently shown that H2O2 is responsible for reducing Au(III) to Au(0) in a process that is mediated by the HOO• radical generated upon abstraction of a hydrogen atom from hydrogen peroxide by chlorine atoms (from AuCl4−) under UV photolytic conditions.20 The hydroperoxyl radical offers the advantage of a weak reducing agent (−0.05 V vs NHE) for surface-mediated growth under photochemical conditions. The stepwise sequential seed-mediated growth approach as illustrated in Scheme 2 was employed with H2O2 as a reducing agent in the growth solution. A variety of concentration ratios were evaluated for the growth solution, and best results were obtained with 0.5 mM as the standard concentration for both 16150

dx.doi.org/10.1021/la302814v | Langmuir 2012, 28, 16148−16155

Langmuir

Article

Figure 1. UV−vis absorption spectra of AuNP prepared via seedmediated photochemical growth using I-2959 and protocol A with 1/ 8th volume seed addition to 7/8th volume growth solution containing 0.5 mM HAuCl4 and 0.25 mM I-2959 where the total sample volume was 2.0 mL. Top: representative SEM images a−d of samples prepared using this method: (a) original seed, (b) sample A, (c) sample A″, (d) sample A‴. Each sample was subjected to 15 min UVA followed by 10 min ripening.

Figure 2. UV−vis absorption spectra of AuNP prepared via seedmediated photochemical growth using H2O2 and protocol A with 1/ 8th volume seed addition to 7/8th volume growth solution containing 0.5 mM HAuCl4 and 0.5 mM H2O2 where the total sample volume was 2.0 mL. Top: representative SEM images a−d of samples prepared using this method: (a) original seed, (b) sample A, (c) sample A″, (d) sample A‴. Each sample was subjected to 15 min UVA followed by 10 min ripening. The scale bars for the SEM images each represent 100 nm.

HAuCl4 and H2O2 in the growth solution. Figure 2 shows the spectral and SEM changes that take place during the stepwise growth process via protocol A. SEM images of resulting particles prepared following protocol B are available in the Supporting Information along with their corresponding absorption spectra (Figures S5 and S6, left). A comparison of particle diameter and change in diameter upon growth from generation to generation using both photochemical radical sources with the same I-2959-generated AuNP seeds is shown in Table 2. The concentration of seeds in solution, reported in terms of Au0, is determined using the extinction coefficients for gold nanoparticles of various sizes in aqueous solution but remains an approximation at best for larger particles owing to the high degree of scattering.5 The particle size and size distribution obtained using the sequential growth method with protocols A and B are compared in Figure 3. It is apparent that the H2O2 method generally yields larger particles than the I-2959 method with improved monodispersity. We also note that using H2O2 leads to stronger SPB absorptions than in the case of I-2959, likely the result of a higher concentration of photoinitiator in the H2O2 method. Increased growth with H2O2 can also be attributed to the ease of surface-catalyzed growth20 with an absence of stabilizing molecules on the particle surface as compared to the I-2959 method with HEBA stabilization. In the case of H2O2, particle stability is limited to a few days; we have recently shown that laser ablation techniques can be used to reverse the aggregation or the growth or large plates.28 Nanoparticle stability can also be readily improved to several weeks by addition of HEBA. In fact, HEBA is a very mild stabilizer that can be easily removed or replaced if a different coverage is required.

The relationship between concentration of seed added and particle enlargement is interesting and tends to emphasize a difference between the one-pot seeded growth and subsequent seeding methods described herein. Using the one-pot seeded growth method, larger particles are generated with lower concentration of seeds, as evidenced in Table 1. With fewer seeds in solution, additional Au3+ deposit on each growing particle surface. In the subsequent growth method where identical seeds are subjected to protocol A or B, it is interesting that the size of B′ particles are smaller than A′ particles. This is observed with both transient reducing agents. Thermal methods use an excess of reducing agent for the effectively complete consumption of Au3+ in solution, while under our experimental conditions the sequential photochemical method allows the delivery of reducing agent at a rate of 0.0026 s−1 in the case of I-2959 following unimolecular decomposition of the photoinitiator (see Supporting Information Figure S7). Au3+ remains in solution after 15 min photolysis, as suggested by the absorbance of the LMCT AuCl4− band at ∼300 nm (see Supporting Information Figure S6). As a result of the remaining metal salt in solution, eq 1 for predicting sizes of seeded particles in Table 2 overestimates the size of the particles. The particles grown from sample A to form sample A′ using the I2959 method and the H2O2 method, for example, are estimated to be 91 ± 38 and 109 nm, respectively. The actual sizes of samples for A′, however, were 53 ± 12 and 62 ± 11 nm, respectively. iii. One-Pot Seed-Mediated Growth of Au/Ag Core− Shell Nanoparticles. These experiments are very similar to those described in part i, except that the metal ion added to the 16151

dx.doi.org/10.1021/la302814v | Langmuir 2012, 28, 16148−16155

Langmuir

Article

Table 2. Summary of Sequential Photochemical Seeding via I-2959 or H2O2 Reduction (See Scheme 2) particle diam (nm) I-2959 method a

Cseed [Au(0)], μM

1st gen (A) 2nd gen (A′) 3rd gen (A″) 4th gen (A‴) 1st gen (B) 2nd gen (B′) 3rd gen (B″)

17.0 16.8 44.2 16.8 13.6 13.5 35.5

seed volume fraction protocol A

protocol B

exptl 29 53 103 147 22 41 89

± ± ± ± ± ± ±

12 12 27 19 + plates 5 14 41

H2O2 method Δ diam

exptl

19 24 50 >44 12 19 48

35 62 90 156 29 53 84

± ± ± ± ± ± ±

10 11 34 35 + plates 10 15 16

Δ diam 25 27 28 >66 19 24 31

a

Samples prepared from 1/8th seed volume are labeled A, with subsequent samples A′, etc. Samples containing 1/10th seed are labeled B, with subsequent seeded samples B′, etc.

Figure 3. Summary of particle sizes obtained following a photochemical seeding approach employing I-2959 or H2O2 with protocols A (1/8th seed volume to 7/8th volume growth solution) and B (1/ 10th seed volume to 9/10th volume growth solution). Initial AuNP seed aliquots prepared from the I-2959/AuCl4− reduction are used in both approaches. The same seeds are used for both I-2959 and H2O2.

Figure 4. UV absorption spectra of the Au/Ag core−shell nanoparticles 50:50 right after the synthesis (10 min) and days after the reaction. For this experiment the concentration of HAuCl4 and AgNO3 used was 0.11 mM. Top: pictures of the sample before (A) and after 10 min irradiation (B).

AuNP seeds is silver rather than gold. Typically, 1.0 mL of AuNP seeds, 2.0 mL of a AgNO3 and a varied concentration of I-2959 and citrate were combined solution prior to irradiation. The mixture was purged under N2 for 15 min and irradiated for 10 min using a Luzchem photoreactor. Figure 4 shows the absorption spectra obtained after reduction of Ag(I) in the presence of AuNP; in this case the ratio between Au and Ag is 50:50. The images show the sample before and after generation of the bimetallic particle. The reduction of Ag(I) on the gold particle surface occurs within minutes, where the new silver surface is prone to oxidation, as evidenced by changes in the absorption spectra within just a few minutes. Addition of citrate to the growth solution, however, is sufficient to stabilize the particles for several months. The presence of an additional band around 400 nm, characteristic of AgNP plasmon absorption, and a blue shift in the maximum of the band corresponding to the AuNP plasmon absorption are indicative of a modification on the surface of these AuNP.18,29 In a different set of experiments we also observed that increasing the UVA dose leads to an approximately linear growth the SPB at 400 nm. Changes in the spectra after 2 min irradiation are provided in Figure S8. We further investigated the possible variation in the core− shell structure changing the ratio between Au and Ag. Three proportions were chosen: Au/Ag = 50/50 (0.11 mM HAuCl4

and 0.11 mM AgNO3); Au/Ag = 65/35 (0.11 mM HAuCl4 and 0.073 mM AgNO3), and Au/Ag = 35/65 (0.11 mM HAuCl4 and 0.15 mM AgNO3) (see Table 3). A different change in color of the solution was observed depending on the thickness of silver shell formed. The wine-red color of 100:0 AuNP Table 3. Summary of Physical and Chemical Properties of the Core−Shell Nanoparticles with Different Au/Ag Atomic Ratios (Citrate Present)a Au/Ag ratio

Au:Ag ratio (measd by EDS)

λmax

size (nm)d

zeta potential (mV)c

100:0b 65:35 50:50 35:65

100:0 65.8:34.2 52.8:47.2 36.8:63.2

522 520 400; 489 405

10 18.8 ± 6.9 13.8 ± 1.2 14.1 ± 3.5

−30 −35.1 −35.9 −41.0

a

Citrate concentration of 0.83, 0.69, and 1.08 mM were used for 65:35, 50:50, and 35:65 Au/Ag ratios, respectively. bGold nanoparticles prepared by the I-2959 method in water with pH 3.0 produced a zeta potentials of approximately −30 mV. cTypical errors for Z-potential measurements are ∼10%. dHydrodynamic radii were collected using dynamic light scattering from a large ensemble of particles.

16152

dx.doi.org/10.1021/la302814v | Langmuir 2012, 28, 16148−16155

Langmuir

Article

changed to a brighter orange color with increasing silver shell thickness, which is further demonstrated in the changes in UV− vis in Figure 5. The metal composition, absorption maxima, particle size, and zeta potential of the core−shell particles are provided in Table 3.

Figure 5. UV absorption spectra of the core−shell nanoparticles synthesized with different Au/Ag proportions. The ratios were obtained using a different proportion of HAuCl4 and AgNO3 (Au/ Ag 35:65−0.11 mM HAuCl4 and 0.15 mM AgNO3, Au/Ag 65:35− 0.11 mM HAuCl4 and 0.073 mM AgNO3, Au/Ag 50:50−0.11 mM HAuCl4 and 0.11 mM AgNO3).

For the Au/Ag = 65/35 solution, a small blue shift from 522 to 520 nm is observed for the AuNP maximum as well as an increase in absorption in the region between 350 and 500 nm. A stronger bathochromic shift from 522 to 489 nm is observed for the 50:50 Au/Ag solution, in addition to a maximum around 400 nm. The intensities of the two maxima are comparable, as expected from the initial ratio used between Au and Ag. Finally, the Au/Ag = 35/65 proportion displays a higher peak around 405 nm, attributed to the stronger Ag absorbance in the core−shell structure, compared to the other shoulder at 490 nm attributed to the Au contribution. The changes are in line with those observed in other Au/Ag bimetallic systems.18,30,31 In order to confirm the formation of core−shell nanostructures in each of the solutions, samples were submitted for TEM and EDS analysis. Figure 6 shows TEM images for the three different Au:Ag ratios. In all cases, structures with a darker region at the center (Au) and a lighter portion in the edges (Ag) are clearly identified. EDS measurements confirm that the ratio of metals (Ag and Au) reflects the initial proportions used and confirms the absence of newly formed monometallic AgNP. These results are provided in Table 3.

Figure 6. TEM images of core−shell nanoparticles solutions synthesized with different proportions Au/Ag: Au/Ag = 65/35 (top); Au/Ag = 50:50 (middle); Au/Ag = 35/65 (bottom). The size bar is 20 nm in all panels.

points summarize the paradigm for this convenient synthesis of AuNP. In the one-pot, one-exposure synthesis, the 3:1 stoichiometry (I-2959 to AuCl4−) works well to prepare ∼10 nm particles that can be used as seeds. The average size of AuNP can be controlled by adjusting the light intensity, with high UVA intensity providing smaller particles.16 Typical irradiances around 40 W/m2 yield particles around 10−12 nm. When AuNP seeds are placed in a growth solution of I-2959 and AuCl4− and irradiated, AuNP surface catalysis favors growth over new nucleation, and the particles can be enlarged in a predictable manner. In this case substoichiometric amounts of I-2959 seem to yield the best results. The stepwise growth approach illustrated in Scheme 2 improves the fine control of AuNP synthesis and leads to lower polydispersity. This approach is also very convenient when a library of AuNP of different sizes is needed or than a specific target size. Each step in this growth strategy offers a new particle size, creating seeds for further growth. We note that early radiolysis work from Heinglein’s group took advantage of the easy generation of •CH2OH in methanol solutions containing N2O and H2O.27 One of the advantages of the photochemical method with I-2959 or H2O2 is that the

■

DISCUSSION It is common to see the spatial and temporal control of photochemical reactions highlighted in studies of photochemistry. The first example in this paper, where Au(III) and the photodecomposition of I-2959 are used as the growth medium for AuNP seeds, provides an excellent example of temporal control. Thus, for the I-2959 system, combining our earlier results with data acquired here provides a basic paradigm for the controlled synthesis of AuNP with fine size control and with low polydispersity. While not strictly “naked”, these particles are largely unprotected, with just chloride ions and HEBA as mild coverage providing enough stabilization to ensure a long shelf life.24 There are no covalent bonds to the stabilizers, and these particles achieve stability in the absence of N, P, and S ligands on the surface.16,32 Thus, the following 16153

dx.doi.org/10.1021/la302814v | Langmuir 2012, 28, 16148−16155

Langmuir

Article

490−500 nm. The former correspond to Ag and the latter to Au, although the blue shift from the original 530 nm for AuNP seeds clearly indicates that this cannot be simple physical mixture of AuNP and AgNP. A graph comparing the absorption of the core−shell nanoparticles with a mixture of AuNP and AgNP is shown in Figure S9. It is interesting that this system does not show any complications or evidence for the precipitation of AgCl, perhaps due to stabilization as AgCl2−. Further, EDS at the surface particle level clearly shows the bimetallic composition of the particle (vide supra), and finally individual images of the particles (see Figure 6) show two domains of different contrast, consistent with a Au core and Ag shell.

synthesis is very fast; suitable particles are obtained in minutes, without the use of ionizing radiation and without reliance on bimolecular reactions of excited states for radical generation. While Henglein’s method required increasing irradiation time (13−23 h) and complete consumption of metal salt in the growth solution, our modification further emphasizes the custom growth of nanoparticles within minutes with residual Au ions remaining in solution and easily removed through centrifugation. In a related synthesis for Au−Ag core−shell nanoparticles,33 Mallik and co-workers suggested that chlorine atom generation led to hydrogen abstraction from the surfactant used (Triton100) and the formation of surfactant-derived reducing ketyl radicals. Our work has shown that these chlorine atoms can abstract from alcohols as well as ethers generating radicals capable of reducing gold ions to Au(0).20 It is very likely that both I-2959 and HEBA (Scheme 1), each containing alcohol and ether functionalities, may play a similarly reactive role producing reducing radicals. In this sense the extended functionality of I-2959 is not limited to the benzoin function but rather may involve the HOCH2CH2O− moiety present in I-2959 and preserved in its product HEBA. Multiple functionality of I-2959 and HEBA provides a simple explanation for the success of substoichiometric concentrations of initiator.22 Further, the fact that HEBA is a known surface stabilizer may cause nascent reducing radicals to be formed at the surface of AuNP seeds and thus favor growth over new nucleation. We note that some Au+ may remain in solution after synthesis when substoichiometric amounts of I-2959 are used. This is consistent with the smaller than predicted particle dimensions observed. The use of H2O2 in the growth medium also favors enlargement and growth on the particle surface, without adding any additional organic stabilizers. In general, the particles prepared in this manner are somewhat cleaner and more monodisperse that those prepared with I-2959 or from NH2OH−HCl. The fact that H2O2-derived particles are deficient in surface stabilizers is also evidenced by their poor stability, which is limited to just a few days but can be dramatically enhanced by addition of HEBA, thus illustrating how important this molecule is in giving stability to I-2959-derived particles.24 Clearly, other specific coating agents could be added, such as citrate, to provide custom stabilized particles for the application of choice. Photoseeding can also be used to produce core−shell particles. Given a AuNP seed, it should be possible to add shells of metals that reduce well with ketyl radicals. In our case, we utilize the example of silver to prepare Au/Ag core−shell nanoparticles. These have a cleaner structure than those prepared in surfactant media before.18,33,34 Throughout this process it is possible to control the overall chemical composition of the nanostructures as well as the dimension of both core and shell. Systems with bimetallic composition could in principle have three different structures at the nanometric level: (a) alloy, (b) core−shell, and (c) a physical mixture of both monometallic nanoparticles. Alloys can be readily ruled out, as they are known to present a single plasmon band in a region intermediate between the separate SPBs for the two metals;2,18,29,35,36 this is clearly not the case (see Figure 5). Our nanoparticles clearly show two distinct plasmon absorptions: one in the ∼400 nm region and the other around

■

CONCLUSION This study demonstrates how photoseeding can be used to control the size, surface coverage, and chemical composition of noble metal nanoparticles. The stepwise photoseeding strategies developed can be used to fine-tune particle size, and narrow size distributions have been achieved for colloidal particles