Article pubs.acs.org/Langmuir
Effect of Ionic Liquid Impurities on the Synthesis of Silver Nanoparticles Laura L. Lazarus,† Carson T. Riche,‡ Noah Malmstadt,*,‡ and Richard L. Brutchey*,† †
Department of Chemistry and ‡Mork Family Department of Chemical Engineering and Materials Science, University of Southern California, Los Angeles, California 90089, United States S Supporting Information *
ABSTRACT: Imidazolium-based ionic liquids have been widely utilized as versatile solvents for metal nanoparticle synthesis; however, reactions to synthesize silver nanoparticles that are performed identically in different commercially obtained lots of 1-butyl-3-methylimidazolium tetrafluoroborate (BMIM-BF4) give divergent results. This suggests that impurities in these nominally identical solvents play an important role in the resulting silver nanoparticle quality. To test the effect that impurities have on the quality of silver nanoparticles synthesized in BMIM-BF4, silver nanoparticles were synthesized in carefully prepared and purified BMIM-BF4 and compared against silver nanoparticles that were synthesized in the purified BMIM-BF4 that had been spiked with trace amounts of water, chloride, and 1-methylimidazole. It was clearly demonstrated that trace amounts of these common ionic liquid impurities cause significant deviation in size and shape (creating polydisperse and irregularly shaped ensembles of both large and small particles), and also negatively impact the stabilization of the resulting silver nanoparticles.
■
INTRODUCTION Room-temperature ionic liquids have been touted as green solvents with many unique properties, making them potential alternatives to traditional organic solvents for a wide variety of applications.1 In particular, ionic liquids based on the alkylimidazolium cation have been widely studied, and many examples based on this motif have been synthesized in the past decade.2 This is perhaps due to the ease with which one can synthetically tune the properties of imidazolium-based ionic liquids by simply altering their constituents (e.g., type of anion, alkyl chain length, and addition of functional groups). These solvents can be recycled and reused; are generally thermally stable; and possess negligible vapor pressures and high dielectric constants, all of which give them the ability to solvate a wide variety of species.1,2 An ever-growing application for imidazolium-based ionic liquids is in metal nanoparticle synthesis, where they serve as interesting dual-function solvents and stabilizers.3−6 For example, silver nanoparticles (AgNPs) of varying sizes and shapes have been successfully synthesized in 1-butyl-3methylimidazolium tetrafluoroborate (BMIM-BF4) using several types of reducing agents.7−9 Like traditional surface ligands, ionic liquids are thought to stabilize metal nanoparticles by coordinating to their surfaces, although these interactions are generally considered to be somewhat weak compared to traditional ligands (e.g., thiols, carboxylates, and amines).5 Evidence also suggests that imidazolium-based ionic liquids are highly structured in the condensed phase, containing supramolecular networks held together by hydrogen bonds and π−π stacking between imidazolium rings.10 This extended ordering © 2012 American Chemical Society
of imidazolium-based ionic liquids results in local hydrophilic and hydrophobic regions whose respective volumes may affect the size and shape of the resulting metal nanoparticles.11,12 Collectively, these properties make imidazolium-based ionic liquids interesting media for synthesizing metal nanoparticles; however, real challenges exist for utilizing these ionic liquids that few reports have directly addressed.13 Impurities found in imidazolium-based ionic liquids can differ between lots of the same solvent and may affect the reproducibility and quality of reactions that are particularly sensitive to them. For example, Dash and Scott first reported that small differences in concentration of 1-methylimidazole affect the stabilization of gold nanoparticles synthesized in BMIM-PF6.14 Typical impurities in imidazolium-based ionic liquids are water, halides, and alkylimidazole starting reagents.13 Hydrophilic imidazolium-based ionic liquids, such as BMIM-BF4, are particularly difficult to purify because water-soluble impurities cannot be easily extracted without losing a considerable amount of the ionic liquid to the water layer. They also have a strong affinity for water, which makes them hygroscopic and therefore difficult to dry.15 Many of the reactions performed in these ionic liquids (e.g., catalysis, organic reactions, and nanoparticle syntheses) are known to be sensitive to moisture, halides, and other trace impurities but impurity concentrations are rarely reported. While several reports have systematically examined how ionic liquid impurities affect homogeneous catalysis and biocatalReceived: September 8, 2012 Revised: October 17, 2012 Published: October 23, 2012 15987
dx.doi.org/10.1021/la303617f | Langmuir 2012, 28, 15987−15993
Langmuir
Article
ysis,16−19 there have not heretofore been any systematic examinations of the effects of ionic liquid impurities on nanoparticle synthesis. It has been well documented that impurities in reagents and solvents can affect the quality of nanoparticles from batch to batch in some syntheses. For example, halide impurities in a surfactant critical for gold nanorod synthesis were shown to dramatically inhibit nanorod growth.20,21 Likewise, it was demonstrated that a popular solvent used in quantum dot synthesis contains impurities that are critical for producing high-quality CdSe nanocrystals.22 These studies prompted us to investigate whether trace impurities commonly found in ionic liquids have an effect on metal nanoparticle synthesis. For this purpose, we synthesized AgNPs in carefully purified BMIM-BF4 and compared the results against BMIM-BF4 that was spiked with various impurities (i.e., water, chloride, and 1-methylimidazole) to determine how each of these affects the quality of the resulting AgNPs, as probed by a combination of transmission electron microscopy (TEM) and UV−vis spectroscopic analysis.
■
isolated by vigorously shaking the oil with a 1:2 (v/v) mixture of acetonitrile and ethyl acetate (45 mL total volume), which resulted in the immediate precipitation of needle-like, off-white crystals of 1-butyl3-methylimidazolium chloride (BMIM-Cl). The solid was isolated and recrystallized several times from acetonitrile and ethyl acetate and dried under vacuum (7 h) to obtain an off-white crystalline solid (99.2 g, 0.568 mol, 90.6% recovered yield). BMIM-Cl (199 g, 1.14 mol) was dissolved in a mixture of distilled water (53 mL) and dichloromethane (CH2Cl2, 74 mL) and cooled in an ice bath before HBF4 [155 mL, 49.3% (actual) aqueous solution, 1.25 mol] was added in 15.5 mL portions over 30 min. The mixture was then allowed to come to room temperature and was stirred for 48 h under nitrogen, which resulted in the formation of two layers, an aqueous phase and a CH2Cl2 phase containing the product. The aqueous phase was removed and extracted with CH2Cl2 (5 × 100 mL) and the CH2Cl2 extracts were combined. The CH2Cl2/BMIM-BF4 phase was further extracted with distilled water (30 mL) until the water extracts were no longer acidic. The solvent was removed and the product dried for several more hours under vacuum (85 °C) to yield a light yellow, viscous liquid (171.8 g, 68.2% recovered yield). Attempts were made to decolorize the BMIM-BF4 by a reported procedure.26 A column was packed with activated charcoal on the top (ca. 14 cm), followed by silica gel (1 cm), then Celite 545 (1 cm). The column was primed with CH2Cl2 (200 mL) before a solution of BMIM-BF4 in CH2Cl2 (1:3 v/v) was passed through, followed by another portion of CH2Cl2 (200 mL) to elute the remaining BMIM-BF4. Nitrogen was used to help purge the column of the eluate. The solvent was removed under vacuum (85 °C). Using this procedure, a small amount of colored impurities were qualitatively removed, but a completely colorless product was not achieved. Less volatile impurities were removed using a cold finger inserted into a round-bottom flask containing the product. The BMIM-BF4 was heated under vacuum (85 °C) while dry ice−ethanol was used to maintain a temperature of −116 °C in the cold finger. After ca. 12 h, the cold finger residue was rinsed into a vial with acetonitrile and analyzed via GC. This procedure was followed for two consecutive days until impurity peaks were no longer observable. 1H NMR (400 MHz, DMSO-d6, TMS, δ): 9.084 (s, 1H), 7.757 (t, 1H, J = 1.8 Hz), 7.690 (t, 1H, J = 1.8 Hz), 4.156 (t, 2H, J = 7.2 Hz), 3.844 (s, 3H), 1.800−1.726 (m, 2H), 1.307−1.214 (m, 2H), 0.902 (t, 3H, J = 7.4 Hz). 13C NMR (100 MHz, DMSO-d6, TMS, δ): 136.46, 123.60, 122.25, 48.49, 35.72, 31.33, 18.76, 13.25. 19F NMR (376 MHz, DMSO-d6, TFT, δ): −148.30, −148.36. Characterization of Common Impurities in Purified BMIMBF4. Water and chloride content were determined by Karl Fischer titration and suppressed ion chromatography, respectively (Galbraith Laboratories Inc.; Knoxville, TN). The water content determination combined coulometry with Karl Fischer titration. In this analysis, the sample was mixed with an amine−methanol mixture containing predominantly iodide and sulfur dioxide, and the iodine produced at the anode through the electrolysis is allowed to react with water. The amount of water was measured directly from the number of coulombs required for electrolysis. Gas chromatography (GC) was used to determine 1-methylimidazole concentration by the standard addition method.27 Solutions of BMIM-BF4 (28 wt %) and 1-methylimidazole (45 wt %) were prepared in acetonitrile. To 1 mL aliquots of the BMIM-BF4 solution were added 0.1, 0.2, 0.3, and 0.4 mL aliquots of the 1-methylimidazole solution and diluted to 5 mL total volume with acetonitrile. A 1 μL sample was injected into the GC that was equipped with an ultrainert inlet liner with glass wool to prevent contamination of the column with the nonvolatile ionic liquid. The retention time of 1-methylimidazole using this method was ca. 4 min (inlet temperature 200 °C, initial column temperature 40 °C for 2 min, then ramp to 200 °C at 110 °C min−1, hold 5 min split ratio 20:1). Each measurement was performed in triplicate and resulted in R2 values of >0.995. Synthesis of 1-Butyl-3-methylimidazolium Borohydride (BMIM-BH4). BMIM-BH4 was synthesized following a literature procedure.28 Briefly, BMIM-Br (5.14 g, 23.5 mmol) and NaBH4 (1.07 g, 28.1 mmol) were stirred in dry acetonitrile for 24 h at 25 °C under a nitrogen atmosphere. The colorless solution was separated from the
EXPERIMENTAL SECTION
Materials and Methods. Tetrafluoroboric acid solution (HBF4, 48% min w/w aqueous solution), 1-chlorobutane (>99.5%), 1methylimidazole (1-MI, 99%), 1-butyl-3-methylimidazolium chloride (BMIM-Cl, 96%), and 1-butyl-3-methylimidazolium tetrafluoroborate (BMIM-BF4 >98%, lots 10153814 and 10163139) were purchased from Alfa-Aesar. Silver(I) tetrafluoroborate (AgBF4, 98%), silver(I) chloride (AgCl, 99.999%), dodecanethiol (≥98%), trioctylamine (98%), BMIM-BF4 (>98%, lots BCBG3452 V and BCBF5160 V), 1butyl-3-methylimidazolium bromide (BMIM-Br, ≥97%), sodium borohydride (NaBH4, 99%), and Celite 545 were purchased from Sigma-Aldrich. Activated carbon powder (Darco G-60, acid-washed, steam-activated) was obtained from JT Baker, Inc. All chemicals were used as received. 1 H, 13C, and 19F NMR spectra were collected on a Varian 400-MR or Varian VNMRS-500 2-channel NMR spectrometer at 25 °C. Tetramethylsilane (TMS, δ 0 ppm, 1H and 13C) and trifluorotoluene (TFT, δ −63.7 ppm, 19F) were used as internal standards. Gas chromatography (GC) data were collected using a 7890A GC System (Agilent Technologies) outfitted with an HP-5 column (J&W Scientific) and FID detector operated at 300 °C. Transmission electron microscopy (TEM) was performed on a JEOL JEM-2100 electron microscope at an operating voltage of 200 kV, equipped with a Gatan Orius CCD camera. Samples for TEM analysis were prepared by dropping hexane dispersions of AgNPs onto copper grids (Carbon type-B, 200 mesh; Ted Pella). TEM micrographs were processed in Matlab to analyze nanoparticle size statistics. Nanoparticle diameters were calculated based on the projected area, in a manner consistent with NIST protocol.23 Grayscale images were converted to binary images with discrete nanoparticles on a uniform background, using a consistent thresholding technique for an accurate comparison of separate samples. UV−vis absorption spectra of AgNP suspensions in hexanes were collected on a Shimadzu UV-1800 spectrophotometer, and spectra were normalized to the peak localized surface plasmon resonance band intensity because of variable yields after phase transfer and purification. Synthesis of 1-Butyl-3-methylimidazolium Tetrafluoroborate (BMIM-BF4). BMIM-BF4 was synthesized following a literature procedure found to produce an ionic liquid with minimal chloride impurities compared to other methods.24,25 1-Chlorobutane (72.0 mL, 0.689 mol), 1-methylimidazole (50.0 mL, 0.627 mol), and dry toluene (62.5 mL) were added to a round-bottom flask in an ice bath under a nitrogen atmosphere. The solution was allowed to warm to room temperature before refluxing for 48 h under nitrogen. This resulted in a two-phase mixture consisting of a viscous amber oil and a light yellow toluene phase that was removed by decanting. The product was 15988
dx.doi.org/10.1021/la303617f | Langmuir 2012, 28, 15987−15993
Langmuir
Article
NaBr byproduct via cannula filtration and then dried in vacuo to yield the BMIM-BH4 product (3.53 g, 97.4% isolated yield) as a low melting solid. The product appears to be stable for long periods of time when stored under nitrogen. 1H NMR (500 MHz, DMSO-d6, TMS, δ): 9.127 (s, 1H), 7.774 (s, 1H, J = 1.75 Hz), 7.706 (s, 1H, J = 1.75 Hz), 4.162 (t, 2H, J = 7.0 Hz), 3.849 (s, 3H), 1.793−1.734 (m, 2H), 1.298− 1.223 (m, 2H), 0.903 (t, 3H, J = 7.25 Hz), −0.068−0.558 (BH4). 13C NMR (125 MHz, DMSO-d6, TMS, δ): 137.45, 124.59, 123.24, 50.26, 36.89, 32.67, 20.01, 13.72. Synthesis of Silver Nanoparticles (AgNPs) in BMIM-BF4. All silver nanoparticle (AgNP) syntheses followed the same general procedure. Solutions of AgBF4 (1.95 mg mL−1, 10 mM) and BMIMBH4 (15.4 mg mL−1, 100 mM) were prepared in BMIM-BF4 with stirring overnight at room temperature. The silver precursor solution was protected from light until just before the reaction. Quickly, 0.5 mL AgBF4 solution was added to a vial followed by 0.5 mL of BMIM-BH4 solution with rapid stirring for 5 min under ambient conditions. The reaction was quenched with a prepared solvent mixture containing hexanes (3.5 mL), ethanol (2 mL), 1-dodecanethiol (0.15 mL), and trioctylamine (0.15 mL) to initiate a phase transfer of the AgNPs out of the ionic liquid and into the hexane phase for ease of analysis. The colored organic phase containing AgNPs was separated and the AgNPs precipitated by centrifugation with the addition of methanol (ca. 3.5 mL). The isolated AgNPs were washed twice by replacing the colorless supernatant with fresh ethanol (4 mL), bath sonication (10 min), and centrifugation (6000 rpm for 3 min) to precipitate. The AgNPs were redispersed in hexanes (2 mL) and sonicated once more (30 min) prior to UV−vis and TEM analyses. Preparation of Impurity-Spiked BMIM-BF4 Solutions and AgNP Synthesis. Purified BMIM-BF4 (vide supra) was used to make all of the following impurity-spiked solutions. A 0.1 wt % chloridespiked ionic liquid was prepared by adding BMIM-Cl to the purified BMIM-BF4 (6 mg mL−1) with stirring overnight at room temperature. Once dissolved, this solution was further diluted with BMIM-BF4 to produce 0.05 and 0.025 wt % chloride solutions. A 0.2 wt % waterspiked ionic liquid was prepared by adding distilled water to the purified BMIM-BF4 (2.4 mg mL−1) with stirring. This was diluted with pure BMIM-BF4 to produce 0.1, 0.05, and 0.025 wt % solutions. A 0.05 wt % solution of 1-methylimidazole-spiked ionic liquid was prepared by stirring 1-methylimidazole with the purified BMIM-BF4 (0.60 mg mL−1). This was diluted with pure BMIM-BF4 to prepare 0.025 and 0.01 wt % solutions. All solutions were prepared under nitrogen. AgNPs were synthesized and characterized as above using the impurity-spiked ionic liquids.
comparison (lots #10153814 and #10163139 from Alfa-Aesar and #BCBG3452 V and #BCBF5160 V from Sigma-Aldrich). Silver nanoparticles were synthesized by 1-butyl-3-methylimidazolium borohydride (BMIM-BH4) reduction in each of the BMIM-BF4 batches under the same reaction conditions. We previously reported that the ionic liquid reducing agent BMIMBH4 has improved solubility in BMIM-BF4 compared to NaBH4 without introducing sodium impurities.29 UV−vis spectra of AgNP suspensions resulting from the various lots of BMIM-BF4 and subsequently phase transferred into hexanes all contained an absorption band resulting from the localized surface plasmon resonance (LSPR) of the AgNPs (see Supporting Information, Figure S1).30 Silver nanoparticles synthesized in the purified BMIM-BF4 were fairly monodisperse (4.4 ± 0.8 nm), and very few agglomerates were observed by TEM (Figure 1). Correspondingly, a relatively
Figure 1. TEM image of AgNPs produced in purified and dried BMIM-BF4 (lot #LL08112).
■
RESULTS AND DISCUSSION Observed Lot-to-Lot Variations in AgNPs Synthesized in Commercial BMIM-BF4. In developing synthetic methods to fabricate AgNPs in BMIM-BF4, variations in the size and polydispersity (and, consequently, colloidal stability) of AgNPs prepared in different commercial lots of this ionic liquid were observed. As a reference point, a purified batch of BMIM-BF4 (lot #LL08112) was synthesized by the reaction of 1-butyl-3metylimidazolium chloride (BMIM-Cl) with HBF4, taking care to limit as many known impurities as possible. Attempts were made to decolorize the resulting ionic liquid and remove less volatile impurities (vide supra). The BMIM-BF4 was also thoroughly dried and care was taken to handle it under strictly anhydrous conditions. Water (
