Article pubs.acs.org/JPCA
Investigating the Synthesis of Ligated Metal Clusters in Solution Using a Flow Reactor and Electrospray Ionization Mass Spectrometry Astrid Olivares,‡ Julia Laskin,† and Grant E. Johnson*,† †
Physical Sciences Division, Pacific Northwest National Laboratory, P.O. Box 999, MSIN K8-88, Richland, Washington, 99352 Department of Chemistry, California Lutheran University, Thousand Oaks, California 91360, United States
‡
ABSTRACT: The scalable synthesis of ligated subnanometer metal clusters containing an exact number of atoms is of interest due to the highly size-dependent catalytic, electronic, and optical properties of these species. While significant research has been conducted on the batch preparation of clusters through reduction synthesis in solution, the processes of metal complex reduction as well as cluster nucleation, growth, and postreduction etching are still not well understood. Herein, we demonstrate a prototype temperature-controlled flow reactor for qualitatively studying cluster formation in solution at steady-state conditions. Employing this technique, methanol solutions of a chloro(triphenylphosphine)gold precursor, 1,4-bis(diphenylphosphino)butane capping ligand, and borane-tert-butylamine reducing agent were combined in a mixing tee and introduced into a heated capillary with a known length. In this manner, the temperature dependence of the relative abundance of different ionic reactants, intermediates, and products synthesized in real time was characterized qualitatively using online mass spectrometry. A wide distribution of doubly and triply charged cationic gold clusters was observed as well as smaller singly charged organometallic complexes. The results demonstrate that temperature plays a crucial role in determining the relative population of cationic gold clusters and, in general, that higher temperature promotes the formation of doubly charged clusters and singly charged organometallic complexes while reducing the abundance of triply charged species. Moreover, the distribution of clusters observed at elevated temperatures is found to be consistent with that obtained at longer reaction times at room temperature, thereby demonstrating that heating may be used to access cluster distributions characteristic of different stages of batch reduction synthesis in solution.
■
INTRODUCTION Subnanometer clusters containing an exact number of atoms are of interest due to their potential biological and chemical applications.1−7 In particular, the scalable synthesis in solution of gold nanoclusters with size-dependent chemical and physical properties may enable widespread application of these materials in areas such as medicine,4,8 catalysis,5,9−11 optics,12 and electronics.13 Therefore, understanding the fundamental processes of how clusters nucleate from reduced organometallic complexes, grow, and are etched in solution in the presence of organic ligands is necessary in order to fully exploit the promise of the size-dependent properties of these ultrasmall species.14,15 In addition, experimental identification of the initial organometallic complexes, intermediates, and final cluster products formed during different stages of batch reduction synthesis in solution will aid theoretical efforts to calculate the structures of these species as well as the energetics associated with their formation through competing pathways.16−19 In recent years, the batch synthesis of subnanometer gold clusters using bidentate diphosphine ligands has been receiving increasing attention.20−29 Diphosphine ligands have the ability to stabilize solutions of relatively monodisperse cationic gold clusters. Recent studies employing (C6H5)2P(CH2)nP(C6H5)2, where n = 2−5, as the capping ligands23,24,27,29 indicate that the distribution of gold clusters formed in solution depends on the © 2014 American Chemical Society
length of the CH2 chain separating the two phosphorus centers of the ligand.20,30 In addition, work by Robinson and coworkers demonstrated that the rate of cluster formation when n = 3 is relatively slow, while cluster formation when n = 4 occurs at an intermediate rate, and cluster formation when n = 5 and 6 is comparatively fast.29 This observation suggests that the length of the CH2 spacer also plays a role in determining the rate of reaction as well as the composition of the final products. In another publication, the same group demonstrated that the length of the CH2 spacer influences the reactivity of ligated gold cluster ions toward the activation of C−I bonds in the gas phase.31 Density functional theory (DFT) calculations of gold clusters ligated with diphosphines have examined the selectivity of different length ligands toward certain size clusters. It was predicted that additional CH2 groups in the spacer may allow more geometrical flexibility of the ligands, thereby resulting in a wider distribution of cluster sizes synthesized in solution than that obtained with analogous shorter ligands.18 Experimental mass spectrometry and spectroscopy studies by Hudgens and co-workers investigated the gas-phase dissociaSpecial Issue: A. W. Castleman, Jr. Festschrift Received: February 20, 2014 Revised: April 1, 2014 Published: April 1, 2014 8464
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
function of temperature. We demonstrate that the temperature of the flow reactor has a pronounced effect on the relative abundance of the metal complexes and gold clusters formed in solution at otherwise identical conditions of reactant concentration and solution flow rate. The results indicate that varying the temperature of the flow reactor is an effective approach for producing steady-state distributions of gold clusters that are characteristic of different stages of batch reduction synthesis at room temperature. This opens up a much broader range of clusters for analysis. The insight obtained from this initial study will be used to direct the preparation of stable streams of size-selected gold clusters in the flow reactor for a variety of applications such as online analysis by mass spectrometry and ion spectroscopy as well as immobilization on surfaces through soft landing of massselected ions.36−38
tion pathways of Au8, Au10, and Au11 clusters synthesized in batches in methanol/chloroform solution using 1,n-bis(diphenylphosphino)n-alkane ligands, where n = 3 and 5.32 The authors observed that loss of a neutral ligand as well as dissociation of the gold cores into AuL+ and Au3L+ were characteristic fragmentation pathways for these clusters. In addition, it was shown that the Au10 and Au11 clusters are more resistant to dissociation in the gas phase than Au8.32 In a following study, the same group investigated the influence of swapping ethyl for phenyl groups at the two phosphorus centers in otherwise identical diphosphine ligands.21 It was demonstrated that the ethyl-substituted diphosphine ligands promote the batch synthesis of singly charged Au11L4Cl2+, while the phenyl-substituted ligands favor formation of triply charged Au 11 L 5 3+ . The difference in cluster charge state and composition was attributed to the distinct electron-accepting properties of the two differently substituted diphosphine ligands.21 The influence of the substitution of functional groups in diphosphine ligands on the size and composition of gold clusters was further demonstrated by Johnson and co-workers, where it was shown that swapping phenyl for cyclohexane groups at the phosphorus centers in diphosphine ligands results in the preferential formation of Au93+ rather than Au113+ clusters at otherwise identical batch synthesis conditions.28 Mass spectrometry has been applied to investigate the ligand exchange reactions occurring during the early stages of synthesis to identify the specific organometallic complexes that lead to the formation of larger clusters.22 It has been shown that the ratio of the concentration of the diphosphine ligand to the gold precursor may be used to control the formation of the initial gold complex distributions that subsequently evolve into clusters of a specific size following reduction.23 Reaction mechanisms responsible for cluster nucleation and growth as well as postreduction size-focusing through etching have been proposed for several different diphosphine ligands.24,25 In particular, Hudgens and co-workers have indicated that there are two classes of diphosphine ligands, (i) ones that form mainly (AuLn2+) (n = 2−3) or (ii) bridged (Au2Ln22+) after ligand exchange with the dissociated gold precursor AuPPh3+.26 The majority of these previous experimental studies were conducted on aliquots extracted at different times from larger batch syntheses performed at room temperature. Therefore, the influence of temperature applied over a defined reaction time at steady-state conditions of reactant concentration and solution flow rate on the distribution of gold clusters formed in solution has not been characterized. In an effort to better qualitatively understand cluster formation through metal complex reduction, as well as cluster nucleation, growth, and etching during synthesis in solution, we applied an approach involving a flow reactor coupled to an online electrospray ionization mass spectrometer. A related method, developed previously by Konermann and co-workers, has been used for investigating the kinetics of enzyme reactions where chemical reaction is initiated by the rapid mixing of two reactant solutions.33−35 This approach has been used to examine processes such as the hydrolysis of acetylcholine as a function of pH,35 the hydrolysis of nitrophenyl acetate by chymotypsin,33 and the folding of the protein ubiquitin.34 As a first demonstration of a prototype technique for the study of inorganic reactions that produce clusters, a model reaction involving the reduction of a gold precursor in the presence of the capping ligand, 1,4-bis(diphenylphosphino)butane, by a weak reducing agent, borane tert-butylamine, was examined as a
■
RESULTS AND DISCUSSION Cationic gold nanoclusters ligated with diphosphine ligands were synthesized in solution using a steady-state flow reactor and analyzed online with electrospray ionization mass spectrometry (ESI-MS) in the positive ion mode. A relatively weak reducing agent, borane tert-butylamine, was used in this study to ensure that reactions would be ongoing throughout the course of the analysis and at different temperatures. Two separate streams of solvent were introduced into the vortex mixing chamber of a high-pressure mixing union using two syringe pumps, as shown schematically in Figure 1. One syringe
Figure 1. A schematic representation of the prototype flow reactor used in this study. Reactants are aspirated from two separate syringes using two independent pumps into a static mixing tee. The mixed solution is reacted for a predetermined length of time at a selected temperature in a heated silica capillary and then aspirated into the inlet of an electrospray mass spectrometer for online analysis.
was used to supply a mixed solution of the gold precursor (AuPPh3Cl) (PPh3 = triphenylphosphine) and the diphosphine ligand (1,4-bis(diphenylphosphino)butane) in methanol. The other syringe was used to provide a solution of the reducing agent (borane tert-butyamine) in methanol. The combined solvent stream exiting the mixing tee was introduced into a coiled fused silica capillary of a predetermined length (6 in. or 5 8465
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
influence of temperature on the distribution of clusters. A representative mass spectrum of gold clusters observed at an intermediate temperature of 55 °C but at otherwise identical conditions of concentration and flow is presented in Figure 2b. In agreement with Figure 2a, the singly charged metal complex (1,2)+ is still found to be by far the most abundant species observed in solution. In contrast with Figure 2a, however, when moderate heat is applied to the flow reactor, the formation of triply charged clusters is observed to be less pronounced. Specifically, triply charged (9,4)3+ and (11,5)3+ clusters are easily observed at low temperature but with decreased abundance after moderate heat has been applied to the flow reactor. Concomitant with the reduction in abundance of these triply charged clusters, the singly charged (2,2)+ complex is also partially consumed and exhibits decreased abundance in the mass spectrum in Figure 2b. This species is, therefore, likely an intermediate that yields larger clusters through higher-order reactions or decomposes to form more reactive smaller metal− ligand complexes. In addition, at 55 °C, the relative abundance of doubly charged (6,4)2+ is observed to increase slightly. Comparison of Figure 2a and b demonstrates that increasing the temperature of the flow reactor yields a qualitatively different distribution of ligated gold clusters and metal complexes than that produced at identical conditions but at lower temperature. Therefore, by using the flow reactor method and changing the temperature, it is possible to investigate the processes occurring at different time periods of the formation of gold clusters through reduction synthesis in solution without perturbing other relevant experimental parameters such as solution concentration and flow rate. In a previous publication, Hudgens and co-workers proposed a detailed series of chemical equations that describe the formation of gold nanoclusters through charged and neutral intermediates during batch reduction synthesis in solution at room temperature.24,27 By varying the ratio of the concentration of diphosphine ligands to the gold precursor, they were able to show that different size clusters form through the reduction of specific metal−ligand complexes. The method presented herein allows the use of another experimental variable, temperature, to study the formation of gold clusters in real time and at steady-state reaction conditions. As the temperature of the flow reactor is increased further, interesting trends in the abundance of different cluster species are observed. The mass spectrum of the solution flowing through the reactor at 70 °C is presented in Figure 2c. The relative abundance of (6,3)2+ is observed to be fairly consistent in each spectrum, indicating that this cluster remains stable in solution at higher temperatures. Similar to the results obtained at low temperature, at 70 °C, the relative abundance of the doubly charged (6,4)2+ cluster is found to be significantly higher than that of the triply charged (11,5)3+ cluster, which has almost disappeared from the mass spectrum in Figure 2c. Indeed, the relative abundance of (6,4)2+ compared to (11,5)3+ is much higher at 70 °C than that at either 25 or 55 °C. In addition, a large decrease in the abundance of the triply charged (9,4)3+ cluster is observed in Figure 2c. These results suggest that the larger triply charged clusters, while more abundant at low temperature, are not stable at temperatures above 55 °C. To better illustrate the qualitative trends in relative abundance observed with increasing temperature of the flow reactor, the temperature dependence of the abundances of representative clusters and metal complexes normalized to the total ion abundance is presented in Figure 3. The results for
ft) that was placed in contact with the surface of a hot plate. The temperature of the hot plate was monitored with a thermocouple attached to a digital meter. In this manner, the temperature of the flow reactor was controlled while monitoring the composition of the solution exiting the heated capillary with online mass spectrometry. At the lowest temperature (25 °C), steady-state formation of gold clusters using 1,4-bis(diphenylphosphino)butane as the capping ligand (L) resulted in the distribution of positively charged ions presented in the mass spectrum in Figure 2a.
Figure 2. Representative positive mode electrospray mass spectra of gold cluster solutions obtained at selected temperatures of (a) 25, (b) 55, and (c) 70 °C. The x-axis represents the mass-to-charge ratio (m/z), and the y-axis represents the normalized abundance.
The singly charged positive ions observed in Figure 2a include AuL2+, (1,2)+, which is the most abundant ion found in solution at low temperature, and Au2L2+, (2,2)+, which is relatively less abundant. This bracket notation will be used throughout the remainder of the text and in the figures to denote the number of gold atoms and diphosphine ligands as well as the ionic charge state of the clusters and complexes. Less abundant doubly charged gold clusters were also observed at room temperature including (6,3)2+, (4,3)2+, (6,4)2+, and (8,4)2+. These doubly charged species are listed in order of decreasing abundance as observed in Figure 2a. Two triply charged cationic clusters, (9,4)3+ and (11,5)3+, were also observed in the mass spectrum at low temperature. Therefore, without the application of any heat to the flow reactor, a wide distribution of ligated cationic gold clusters of different sizes and ionic charge states was observed in a straightforward manner. In addition, several singly charged gold ligand complexes were also identified. Many of these species have been observed previously as products at different stages of batch reduction synthesis in solution using mass spectrometry.26,29 In an effort to better qualitatively understand the process of gold cluster formation in solution, the temperature of the flow reactor was used as an experimental variable to systematically investigate the influence of applying heat on the distribution of gold clusters while all other experimental parameters such as reactant concentrations and solution flow rates were kept the same. It should be noted that the batch synthesis of gold clusters using diphosphine ligands is typically carried out at room temperature. Therefore, this study represents, to the best of our knowledge, the first effort to qualitatively understand the 8466
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
abundant than (6,3)2+ but also shows a similar trend of first increasing in abundance up to 60 °C and then decreasing in abundance at higher temperature. In addition, the abundance of the (8,4)2+ cluster shows the same sort of dependence on temperature as (6,3)2+ and (6,4)2+. This species initially increases in abundance with higher temperature and then drops off at temperatures above 60 °C. The behavior of two representative triply charged gold clusters formed in the flow reactor is illustrated in Figure 3c. The (11,5)3+ cluster is observed with highest abundance at room temperature and stays consistent with a small application of heat. Above 50 °C, however, this cluster decays in abundance until it is barely visible in the spectrum at higher temperatures. The (9,4)3+ cluster is found to be less sensitive to temperature than (11,5)3+ and exhibits only a small change in abundance with temperatures up to 80 °C. Above this temperature, the (9,4)3+ cluster also shows reduced abundance. The results described above, which were obtained at different temperatures in a steady-state flow reactor, share interesting similarities with previous findings for diphosphine ligated gold clusters synthesized in batch reactions at room temperature. For example, Hudgens and co-workers proposed that triply charged (9,4)3+ may be produced from reaction between (1,2)+ and (8,4)2+ for clusters ligated with the longer pentane-based ligand.27 The reverse process, involving decomposition of (9,4)3+ to form (8,4)2+ and (1,2)+, is shown in eq 1. (9, 4)3 + + 2L ↔ (8, 4)2 + + (1, 2)+
(1)
The steady decrease of (9,4)3+ and increase of (8,4)2+ and (1,2)+ with increasing temperature observed in Figure 3 between 25 and 60 °C indicates that eq 1 may also describe the reaction responsible for the trends in abundance observed for the shorter butane-based ligand used in this study. Alternatively, the (9,4)3+ cluster may also decompose in a reaction that generates (8,4)2+ and (1,1)+ according to eq 2.
Figure 3. Plots of the normalized abundance of selected gold clusters and complexes as a function of the temperature (ºC) of the capillary flow reactor: (a) complexes increasing in abundance with higher temperature, (b) clusters that exhibit a maximum in abundance at an intermediate elevated temperature, and (c) species that decrease in abundance with an increase in temperature.
(9, 4)3 + + L ↔ (8, 4)2 + + (1, 1)+
two singly charged gold complexes, (1,1)+ and (1,2)+, that exhibit increasing abundance at higher temperature are presented in Figure 3a. The (1,1)+ complex is observed to initially increase in abundance between 25 and 50 °C and then decrease slightly up to a temperature of around 70 °C. At temperatures higher than 70 °C, this complex appears to increase in abundance again. The (1,2)+ complex, in contrast, is far more abundant than (1,1)+ and is observed to increase gradually in abundance with temperature up to around 80 °C. The larger (2,2)+ complex is observed to be far less abundant than either (1,1)+ and (1,2)+ and is, therefore, presented in Figure 3c. This singly charged species exhibits a gradual monotonic decrease in abundance with increasing temperature followed by a small spike to higher abundance at around 80 °C. The behavior of four doubly charged ligated gold clusters is illustrated in Figure 3b. The smaller (4,3)2+ cluster is observed to have high abundance at low temperature, which remains constant with heating up to around 55 °C and then decreases slightly with further application of heat to the flow reactor. At temperatures up to 55 °C, therefore, the abundance of this cluster is found to be similar to that observed at room temperature. The slightly larger (6,3)2+ cluster is far more abundant than (4,3)2+ and exhibits different behavior with a steady increase in abundance with increasing temperature up to around 60 °C followed by a rapid decline in abundance at higher temperature. The related (6,4)2+ cluster is much less
(2)
The increasing abundance of (1,1)+ with temperature observed in Figure 3a is consistent with this alternative reaction, which also reduces the charge of the original 3+ cluster to 2+. The decrease in abundance of (2,2)+ observed in Figure 3c with increasing temperature is consistent with the previous identification of this species as a highly reactive radical by Hudgens and co-workers.24 In order to demonstrate the ability of the temperaturecontrolled flow reactor to produce cluster distributions that are characteristic of different stages of batch reduction synthesis, mass spectra were obtained at 15, 80, and 160 min after batch reactions were initiated at room temperature. This experiment was conducted by mixing the reagents in a single microfuge tube and loading the reacting solution into a single syringe. The reacting contents of the syringe were then aspirated into the mass spectrometer. All other experimental conditions were the same as those used with the mixing tee. The mass spectrum obtained after 15 min of batch reduction of the gold precursor in the presence of the butane-based ligand is presented in Figure 4a. Several prominent singly charged peaks are observed for metal complexes, including (1,1)+, (1,2)+, and (2,2)+. In addition, a small peak is observed for doubly charged (6,4)2+. The mass spectrum obtained after 80 min of batch reaction at room temperature is presented in Figure 4b. At this intermediate time, it is clear that (2,2)+ has decreased in 8467
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
Figure 4. Representative positive mode electrospray mass spectra of gold cluster solutions obtained at selected times of (a) 15, (b) 80, and (c) 160 min following the initiation of the batch reduction synthesis reaction. The x-axis represents the mass-to-charge ratio (m/z), and the y-axis represents the normalized abundance.
Figure 5. Representative positive mode electrospray mass spectra of gold cluster solutions obtained using a capillary flow reactor (a) 6 in. in length and (b) 5 feet in length. The x-axis represents the mass-tocharge ratio (m/z), and the y-axis represents the normalized abundance.
abundance, while (6,4)2+ has become more pronounced in the spectrum. Concomitant with the decrease in (2,2)+ is a small increase in the abundance of (6,3)2+. The mass spectrum obtained at the longest batch reaction time of 160 min is presented in Figure 4c. The (2,2)+ peak is completely absent at this time, while both the (6,3)2+ and (6,4)2+ peaks are increased in abundance. A prominent peak for (8,4)2+ is also observed at 160 min. These room-temperature results of batch synthesis are consistent with the spectra obtained previously by Hudgens and co-workers, which showed (6,4)2+ to be a stable species for clusters synthesized with the shorter propane-based ligand.32 In addition, Robinson and co-workers demonstrated the formation of (6,3)2+, (8,3)2+, (8,4)2+, (10,4)2+, and (11,5)3+ using a stronger sodium borohydride reducing agent after 1 day of reaction time.29 The shorter reaction time, milder reducing agent, and different mass spectrometer used in the current study may explain the absence of (8,3)2+, (10,4)2+, and (11,5)3+ in Figure 4. Close inspection of Figures 2 and 4 illustrates that by increasing the temperature of the flow reactor, it is possible to create steady-state cluster distributions that are characteristic of different stages of batch reduction synthesis. For example, by heating the flow reactor to around 80 °C, it is possible to reduce the abundance of (2,2)+ to a level similar to that observed following 160 min of batch reaction at room temperature. In addition, by heating the flow reactor to an intermediate temperature of around 60 °C, it is possible to increase the abundance of (6,3)2+, (6,4)2+, and (8,4)2+ to a level similar to that obtained at 160 min of batch reaction time at room temperature. As a further demonstration of the capabilities of the flow reactor, we examined the distribution of clusters produced at room temperature using capillaries of different lengths. The mass spectrum obtained for a reacting solution exiting a short 6 in. capillary is presented in Figure 5a. Similar to the cluster distribution obtained at a short batch reaction time in Figure 4a, (1,2)+ and (2,2)+ are observed to be the predominant species in solution with the 6 in. capillary. The mass spectrum of a solution exiting a much longer 5 ft capillary is presented in Figure 5b. It is clear that increasing the reaction time through lengthening the capillary results in a very different distribution of clusters, including (6,3)2+, (11,5)3+, and (6,4)2+. The similarities between the spectra obtained from batch
reduction at different times and from flow reactors of different lengths demonstrate the versatility of this steady-state approach. Furthermore, the use of temperature control is shown to be an effective method to access cluster distributions characteristic of different stages of batch reduction. It follows that transient species that may only be present for a short period of time at room temperature during batch synthesis as well as species that require long periods of time to form in solution may be obtained by varying the length and temperature of the flow reactor.
■
CONCLUSIONS Employing electrospray ionization mass spectrometry combined with a temperature-controlled variable length flow reactor, a new approach has been demonstrated for investigating the reduction of metal complexes in the presence of organic ligands to form nanoclusters. For a model system consisting of diphosphine-ligated gold clusters, a key singly charged (2,2)+ intermediate was observed with high abundance at low and intermediate temperatures and reduced abundance at high temperature. The doubly charged (4,3)2+, (6,3)2+, (6,4)2+, and (8,4)2+ clusters initially increased in abundance up to 60 °C and then decreased in abundance at higher temperatures. A stable abundance of triply charged (11,5)3+ was observed up to 50 °C, at which point this species decreased in abundance and was ultimately completely consumed at higher temperatures. The triply charged (9,4)3+ cluster also exhibited a steady decrease in abundance with increasing temperature. Our results illustrate how temperature and reactor length may be used as parameters to obtain steady-state distributions of clusters that are characteristic of different stages of batch reduction in solution. We anticipate that a secondgeneration flow reactor will be used to produce nanoclusters across a wide range of sizes and compositions for quantitative analysis using mass spectrometry and ion spectroscopy as well as characterization on surfaces following soft landing of massselected ions. 8468
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
■
Article
EXPERIMENTAL METHODS Synthesis of Ligated Gold Clusters. All chemicals were used as purchased with no further purification. Chloro(triphenylphosphine)gold(I) (99.9%, Sigma-Aldrich) was dissolved in methanol (99.8%, Fisher) to generate 50 mL of a 0.1 mM solution. 1,4-Bis(diphenylphosphino)butane (98%, SigmaAldrich), a bidentate capping ligand, was added to the gold solution to generate 50 mL of a 0.1 mM solution. In a separate Pyrex jar, a solution of reducing agent, borane tert-butylamine (97%, Sigma-Aldrich), was added to generate 50 mL of 0.5 mM solution in methanol. Both solutions were mixed in a highpressure mixing tee (M-540, IDEX Corporation) using two separate syringe pumps (KD Scientific). The fused silica capillary (Polymicro Technologies) introducing the sample into the mass spectrometer was heated using a hot plate (Fisher) in approximately 10 °C increments. The temperature of the flow reactor was determined using a K-type thermocouple connected to a meter (Fluke). The heated solution was then introduced to an electrospray ionization mass spectrometer (Bruker HCT Ion Trap, ESI-MS). The evolution of the mass spectra with time was monitored at room temperature without a mixing tee by combining the reagents in a single syringe (batch synthesis) and aspirating the reacting solution into the mass spectrometer. The influence of the length of the capillary on the distribution of clusters was determined by aspirating an identical reacting solution into 6 in. and 5 ft capillaries. The retention times of the 6 in. and 5 ft capillaries were estimated to be approximately 2 and 25 min, respectively. Analysis of Cluster Solutions. An electrospray ionization mass spectrometer, Bruker HCT ion trap (m/z = 100−2500), was used for analysis of the cluster solutions exiting the flow reactor. In the positive ion mode, the source conditions employed were as follows: capillary 4000 V, nebulizer 20 psi, drying gas 5 L/min, and drying gas temperature 200 °C. The internal instrument settings were as follows: skimmer 25 V, capillary exit 75 V, Oct 1 DC 15 V, Oct 2 DC −5 V, trap drive 100, Oct RF 75 Vpp, lens 1−10 V, and lens 2−50 V. The stock gold precursor and ligand solutions were diluted by a factor of 10 in methanol and loaded into glass syringes (Hamilton). The reducing agent solution was diluted to a 0.20 mM concentration. Both solutions were introduced at a flow rate of 100 μL/h using two separate syringe pumps (KD Scientific). The molecular formulas of the clusters and complexes were assigned on the basis of their mass-to-charge ratios (m/z), the charge-state-dependent separation of the individual peaks, and the overall shape of the isotopic distributions. Additionally, experimental spectra were compared with spectra simulated using the molecular weight calculator program (http://omics. pnl.gov/software/-MWCalculator.php).
■
support from the Linus Pauling Fellowship and the Laboratory Directed Research and Development Program at the Pacific Northwest National Laboratory (PNNL). This work was performed using EMSL, a national scientific user facility sponsored by the DOE’s Office of Biological and Environmental Research and located at PNNL. PNNL is operated by Battelle for the U.S. DOE.
■
REFERENCES
(1) Jin, R. C. Quantum Sized, Thiolate-Protected Gold Nanoclusters. Nanoscale 2010, 2, 343−362. (2) Li, G.; Jin, R. C. Atomically Precise Gold Nanoclusters as New Model Catalysts. Acc. Chem. Res. 2013, 46, 1749−1758. (3) Nimmala, P. R.; Dass, A. Au36(SPh)23 Nanomolecules. J. Am. Chem. Soc. 2011, 133, 9175−9177. (4) Dreaden, E. C.; Mackey, M. A.; Huang, X. H.; Kang, B.; El-Sayed, M. A. Beating Cancer in Multiple Ways Using Nanogold. Chem. Soc. Rev. 2011, 40, 3391−3404. (5) Herzing, A. A.; Kiely, C. J.; Carley, A. F.; Landon, P.; Hutchings, G. J. Identification of Active Gold Nanoclusters on Iron Oxide Supports for CO Oxidation. Science 2008, 321, 1331−1335. (6) Wilcoxon, J. P.; Abrams, B. L. Synthesis, Structure and Properties of Metal Nanoclusters. Chem. Soc. Rev. 2006, 35, 1162−1194. (7) Yeh, Y. C.; Creran, B.; Rotello, V. M. Gold Nanoparticles: Preparation, Properties, and Applications in Bionanotechnology. Nanoscale 2012, 4, 1871−1880. (8) Kudgus, R. A.; Bhattacharya, R.; Mukherjee, P. Cancer Nanotechnology: Emerging Role of Gold Nanoconjugates. AntiCancer Agents Med. Chem. 2011, 11, 965−973. (9) Li, G.; Jiang, D. E.; Liu, C.; Yu, C. L.; Jin, R. C. Oxide-Supported Atomically Precise Gold Nanocluster for Catalyzing Sonogashira Cross-Coupling. J. Catal. 2013, 306, 177−183. (10) Tsunoyama, H.; Liu, Y. M.; Akita, T.; Ichikuni, N.; Sakurai, H.; Xie, S. H.; Tsukuda, T. Size-Controlled Synthesis of Gold Clusters as Efficient Catalysts for Aerobic Oxidation. Catal. Surv. Asia 2011, 15, 230−239. (11) Turner, M.; Golovko, V. B.; Vaughan, O. P. H.; Abdulkin, P.; Berenguer-Murcia, A.; Tikhov, M. S.; Johnson, B. F. G.; Lambert, R. M. Selective Oxidation with Dioxygen by Gold Nanoparticle Catalysts Derived from 55-Atom Clusters. Nature 2008, 454, 981−U931. (12) Yao, H. Optically Active Gold Nanoclusters. Curr. Nanosci. 2008, 4, 92−97. (13) Perez, A.; Melinon, P.; Dupuis, V.; Bardotti, L.; Masenelli, B.; Tournus, F.; Prevel, B.; Tuaillon-Combes, J.; Bernstein, E.; Tamion, A.; et al. Functional Nanostructures from Clusters. Int. J. Nanotechnol. 2010, 7, 523−574. (14) Liu, C.; Li, G.; Pang, G. S.; Jin, R. C. Toward Understanding the Growth Mechanism of AuN(Sr)M Nanoclusters: Effect of Solvent on Cluster Size. RSC Adv. 2013, 3, 9778−9784. (15) Jin, R. C.; Qian, H. F.; Wu, Z. K.; Zhu, Y.; Zhu, M. Z.; Mohanty, A.; Garg, N. Size Focusing: A Methodology for Synthesizing Atomically Precise Gold Nanoclusters. J. Phys. Chem. Lett. 2010, 1, 2903−2910. (16) Guidez, E. B.; Hadley, A.; Aikens, C. M. Initial Growth Mechanisms of Gold−Phosphine Clusters. J. Phys. Chem. C 2011, 115, 6305−6316. (17) Ivanov, S. A.; Arachchige, I.; Aikens, C. M. Density Functional Analysis of Geometries and Electronic Structures of Gold−Phosphine Clusters. The Case of Au4(PR3)42+ and Au4(μ2-I)2(PR3)4. J. Phys. Chem. A 2011, 115, 8017−8031. (18) Hong, S.; Shafai, G.; Bertino, M.; Rahman, T. S. Toward an Understanding of Ligand Selectivity in Nanocluster Synthesis. J. Phys. Chem. C 2011, 115, 14478−14487. (19) Shafai, G.; Hong, S. Y.; Bertino, M.; Rahman, T. S. Effect of Ligands on the Geometric and Electronic Structure of Au13 Clusters. J. Phys. Chem. C 2009, 113, 12072−12078.
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the U.S. Department of Energy (DOE), Office of Basic Energy Sciences, Division of Chemical Sciences, Geosciences and Biosciences. A.O. acknowledges support from the DOE Science Undergraduate Laboratory Internship (SULI) program. G.E.J. acknowledges partial 8469
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
(20) Bertino, M. F.; Sun, Z. M.; Zhang, R.; Wang, L. S. Facile Syntheses of Monodisperse Ultrasmall Au Clusters. J. Phys. Chem. B 2006, 110, 21416−21418. (21) Golightly, J. S.; Gao, L.; Castleman, A. W.; Bergeron, D. E.; Hudgens, J. W.; Magyar, R. J.; Gonzalez, C. A. Impact of Swapping Ethyl for Phenyl Groups on Diphosphine-Protected Undecagold. J. Phys. Chem. C 2007, 111, 14625−14627. (22) Bergeron, D. E.; Coskuner, O.; Hudgens, J. W.; Gonzalez, C. A. Ligand Exchange Reactions in the Formation of DiphosphineProtected Gold Clusters. J. Phys. Chem. C 2008, 112, 12808−12814. (23) Pettibone, J. M.; Hudgens, J. W. Synthetic Approach for Tunable, Size-Selective Formation of Monodisperse, DiphosphineProtected Gold Nanoclusters. J. Phys. Chem. Lett. 2010, 1, 2536−2540. (24) Hudgens, J. W.; Pettibone, J. M.; Senftle, T. P.; Bratton, R. N. Reaction Mechanism Governing Formation of 1,3-Bis(diphenylphosphino)propane-Protected Gold Nanoclusters. Inorg. Chem. 2011, 50, 10178−10189. (25) Pettibone, J. M.; Hudgens, J. W. Gold Cluster Formation with Phosphine Ligands: Etching as a Size-Selective Synthetic Pathway for Small Clusters? ACS Nano 2011, 5, 2989−3002. (26) Pettibone, J. M.; Hudgens, J. W. Predictive Gold Nanocluster Formation Controlled by Metal−Ligand Complexes. Small 2012, 8, 715−725. (27) Pettibone, J. M.; Hudgens, J. W. Reaction Network Governing Diphosphine-Protected Gold Nanocluster Formation from Nascent Cationic Platforms. Phys. Chem. Chem. Phys. 2012, 14, 4142−4154. (28) Johnson, G. E.; Priest, T.; Laskin, J. Synthesis and Characterization of Gold Clusters Ligated with 1,3-Bis(dicyclohexylphosphino)propane. ChemPlusChem 2013, 78, 1033−1039. (29) Robinson, P. S. D.; Nguyen, T. L.; Lioe, H.; O’Hair, R. A. J.; Khairallah, G. N. Synthesis and Gas-Phase Uni- and Bi-Molecular Reactivity of Bisphosphine Ligated Gold Clusters, [AuxLy]n+. Int. J. Mass Spectrom. 2012, 330, 109−117. (30) Shichibu, Y.; Suzuki, K.; Konishi, K. Facile Synthesis and Optical Properties of Magic-Number Au13 Clusters. Nanoscale 2012, 4, 4125− 4129. (31) Robinson, P. S. D.; Khairallah, G. N.; Silva, G.; Lioe, H.; O’Hair, R. A. J. Gold-Mediated C−I Bond Activation of Iodobenzene. Angew. Chem., Int. Ed. 2012, 51, 3812−3817. (32) Bergeron, D. E.; Hudgens, J. W. Ligand Dissociation and Core Fission from Diphosphine-Protected Gold Clusters. J. Phys. Chem. C 2007, 111, 8195−8201. (33) Wilson, D. J.; Konermann, L. Mechanistic Studies on Enzymatic Reactions by Electrospray Ionization MS Using a Capillary Mixer with Adjustable Reaction Chamber Volume for Time-Resolved Measurements. Anal. Chem. 2004, 76, 2537−2543. (34) Wilson, D. J.; Konermann, L. A Capillary Mixer with Adjustable Reaction Chamber Volume for Millisecond Time-Resolved Studies by Electrospray Mass Spectrometry. Anal. Chem. 2003, 75, 6408−6414. (35) Kolakowski, B. M.; Simmons, D. A.; Konermann, L. StoppedFlow Electrospray Ionization Mass Spectrometry: A New Method for Studying Chemical Reaction Kinetics in Solution. Rapid Commun. Mass Spectrom. 2000, 14, 772−776. (36) Johnson, G. E.; Wang, C.; Priest, T.; Laskin, J. Monodisperse Au11 Clusters Prepared by Soft Landing of Mass Selected Ions. Anal. Chem. 2011, 83, 8069−8072. (37) Johnson, G. E.; Priest, T.; Laskin, J. Charge Retention by Gold Clusters on Surfaces Prepared Using Soft Landing of Mass Selected Ions. ACS Nano 2012, 6, 573−582. (38) Johnson, G. E.; Priest, T.; Laskin, J. Coverage-Dependent Charge Reduction of Cationic Gold Clusters on Surfaces Prepared Using Soft Landing of Mass-Selected Ions. J. Phys. Chem. C 2012, 116, 24977−24986.
8470
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
Investigating the Synthesis of Ligated Metal Clusters in Solution Using a Flow Reactor and Electrospray Ionization Mass Spectrometry Astrid Olivares,‡ Julia Laskin,† and Grant E. Johnson*,† †
Physical Sciences Division, Pacific Northwest National Laboratory, P.O. Box 999, MSIN K8-88, Richland, Washington, 99352 Department of Chemistry, California Lutheran University, Thousand Oaks, California 91360, United States
‡
ABSTRACT: The scalable synthesis of ligated subnanometer metal clusters containing an exact number of atoms is of interest due to the highly size-dependent catalytic, electronic, and optical properties of these species. While significant research has been conducted on the batch preparation of clusters through reduction synthesis in solution, the processes of metal complex reduction as well as cluster nucleation, growth, and postreduction etching are still not well understood. Herein, we demonstrate a prototype temperature-controlled flow reactor for qualitatively studying cluster formation in solution at steady-state conditions. Employing this technique, methanol solutions of a chloro(triphenylphosphine)gold precursor, 1,4-bis(diphenylphosphino)butane capping ligand, and borane-tert-butylamine reducing agent were combined in a mixing tee and introduced into a heated capillary with a known length. In this manner, the temperature dependence of the relative abundance of different ionic reactants, intermediates, and products synthesized in real time was characterized qualitatively using online mass spectrometry. A wide distribution of doubly and triply charged cationic gold clusters was observed as well as smaller singly charged organometallic complexes. The results demonstrate that temperature plays a crucial role in determining the relative population of cationic gold clusters and, in general, that higher temperature promotes the formation of doubly charged clusters and singly charged organometallic complexes while reducing the abundance of triply charged species. Moreover, the distribution of clusters observed at elevated temperatures is found to be consistent with that obtained at longer reaction times at room temperature, thereby demonstrating that heating may be used to access cluster distributions characteristic of different stages of batch reduction synthesis in solution.
■
INTRODUCTION Subnanometer clusters containing an exact number of atoms are of interest due to their potential biological and chemical applications.1−7 In particular, the scalable synthesis in solution of gold nanoclusters with size-dependent chemical and physical properties may enable widespread application of these materials in areas such as medicine,4,8 catalysis,5,9−11 optics,12 and electronics.13 Therefore, understanding the fundamental processes of how clusters nucleate from reduced organometallic complexes, grow, and are etched in solution in the presence of organic ligands is necessary in order to fully exploit the promise of the size-dependent properties of these ultrasmall species.14,15 In addition, experimental identification of the initial organometallic complexes, intermediates, and final cluster products formed during different stages of batch reduction synthesis in solution will aid theoretical efforts to calculate the structures of these species as well as the energetics associated with their formation through competing pathways.16−19 In recent years, the batch synthesis of subnanometer gold clusters using bidentate diphosphine ligands has been receiving increasing attention.20−29 Diphosphine ligands have the ability to stabilize solutions of relatively monodisperse cationic gold clusters. Recent studies employing (C6H5)2P(CH2)nP(C6H5)2, where n = 2−5, as the capping ligands23,24,27,29 indicate that the distribution of gold clusters formed in solution depends on the © 2014 American Chemical Society
length of the CH2 chain separating the two phosphorus centers of the ligand.20,30 In addition, work by Robinson and coworkers demonstrated that the rate of cluster formation when n = 3 is relatively slow, while cluster formation when n = 4 occurs at an intermediate rate, and cluster formation when n = 5 and 6 is comparatively fast.29 This observation suggests that the length of the CH2 spacer also plays a role in determining the rate of reaction as well as the composition of the final products. In another publication, the same group demonstrated that the length of the CH2 spacer influences the reactivity of ligated gold cluster ions toward the activation of C−I bonds in the gas phase.31 Density functional theory (DFT) calculations of gold clusters ligated with diphosphines have examined the selectivity of different length ligands toward certain size clusters. It was predicted that additional CH2 groups in the spacer may allow more geometrical flexibility of the ligands, thereby resulting in a wider distribution of cluster sizes synthesized in solution than that obtained with analogous shorter ligands.18 Experimental mass spectrometry and spectroscopy studies by Hudgens and co-workers investigated the gas-phase dissociaSpecial Issue: A. W. Castleman, Jr. Festschrift Received: February 20, 2014 Revised: April 1, 2014 Published: April 1, 2014 8464
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
function of temperature. We demonstrate that the temperature of the flow reactor has a pronounced effect on the relative abundance of the metal complexes and gold clusters formed in solution at otherwise identical conditions of reactant concentration and solution flow rate. The results indicate that varying the temperature of the flow reactor is an effective approach for producing steady-state distributions of gold clusters that are characteristic of different stages of batch reduction synthesis at room temperature. This opens up a much broader range of clusters for analysis. The insight obtained from this initial study will be used to direct the preparation of stable streams of size-selected gold clusters in the flow reactor for a variety of applications such as online analysis by mass spectrometry and ion spectroscopy as well as immobilization on surfaces through soft landing of massselected ions.36−38
tion pathways of Au8, Au10, and Au11 clusters synthesized in batches in methanol/chloroform solution using 1,n-bis(diphenylphosphino)n-alkane ligands, where n = 3 and 5.32 The authors observed that loss of a neutral ligand as well as dissociation of the gold cores into AuL+ and Au3L+ were characteristic fragmentation pathways for these clusters. In addition, it was shown that the Au10 and Au11 clusters are more resistant to dissociation in the gas phase than Au8.32 In a following study, the same group investigated the influence of swapping ethyl for phenyl groups at the two phosphorus centers in otherwise identical diphosphine ligands.21 It was demonstrated that the ethyl-substituted diphosphine ligands promote the batch synthesis of singly charged Au11L4Cl2+, while the phenyl-substituted ligands favor formation of triply charged Au 11 L 5 3+ . The difference in cluster charge state and composition was attributed to the distinct electron-accepting properties of the two differently substituted diphosphine ligands.21 The influence of the substitution of functional groups in diphosphine ligands on the size and composition of gold clusters was further demonstrated by Johnson and co-workers, where it was shown that swapping phenyl for cyclohexane groups at the phosphorus centers in diphosphine ligands results in the preferential formation of Au93+ rather than Au113+ clusters at otherwise identical batch synthesis conditions.28 Mass spectrometry has been applied to investigate the ligand exchange reactions occurring during the early stages of synthesis to identify the specific organometallic complexes that lead to the formation of larger clusters.22 It has been shown that the ratio of the concentration of the diphosphine ligand to the gold precursor may be used to control the formation of the initial gold complex distributions that subsequently evolve into clusters of a specific size following reduction.23 Reaction mechanisms responsible for cluster nucleation and growth as well as postreduction size-focusing through etching have been proposed for several different diphosphine ligands.24,25 In particular, Hudgens and co-workers have indicated that there are two classes of diphosphine ligands, (i) ones that form mainly (AuLn2+) (n = 2−3) or (ii) bridged (Au2Ln22+) after ligand exchange with the dissociated gold precursor AuPPh3+.26 The majority of these previous experimental studies were conducted on aliquots extracted at different times from larger batch syntheses performed at room temperature. Therefore, the influence of temperature applied over a defined reaction time at steady-state conditions of reactant concentration and solution flow rate on the distribution of gold clusters formed in solution has not been characterized. In an effort to better qualitatively understand cluster formation through metal complex reduction, as well as cluster nucleation, growth, and etching during synthesis in solution, we applied an approach involving a flow reactor coupled to an online electrospray ionization mass spectrometer. A related method, developed previously by Konermann and co-workers, has been used for investigating the kinetics of enzyme reactions where chemical reaction is initiated by the rapid mixing of two reactant solutions.33−35 This approach has been used to examine processes such as the hydrolysis of acetylcholine as a function of pH,35 the hydrolysis of nitrophenyl acetate by chymotypsin,33 and the folding of the protein ubiquitin.34 As a first demonstration of a prototype technique for the study of inorganic reactions that produce clusters, a model reaction involving the reduction of a gold precursor in the presence of the capping ligand, 1,4-bis(diphenylphosphino)butane, by a weak reducing agent, borane tert-butylamine, was examined as a
■
RESULTS AND DISCUSSION Cationic gold nanoclusters ligated with diphosphine ligands were synthesized in solution using a steady-state flow reactor and analyzed online with electrospray ionization mass spectrometry (ESI-MS) in the positive ion mode. A relatively weak reducing agent, borane tert-butylamine, was used in this study to ensure that reactions would be ongoing throughout the course of the analysis and at different temperatures. Two separate streams of solvent were introduced into the vortex mixing chamber of a high-pressure mixing union using two syringe pumps, as shown schematically in Figure 1. One syringe
Figure 1. A schematic representation of the prototype flow reactor used in this study. Reactants are aspirated from two separate syringes using two independent pumps into a static mixing tee. The mixed solution is reacted for a predetermined length of time at a selected temperature in a heated silica capillary and then aspirated into the inlet of an electrospray mass spectrometer for online analysis.
was used to supply a mixed solution of the gold precursor (AuPPh3Cl) (PPh3 = triphenylphosphine) and the diphosphine ligand (1,4-bis(diphenylphosphino)butane) in methanol. The other syringe was used to provide a solution of the reducing agent (borane tert-butyamine) in methanol. The combined solvent stream exiting the mixing tee was introduced into a coiled fused silica capillary of a predetermined length (6 in. or 5 8465
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
influence of temperature on the distribution of clusters. A representative mass spectrum of gold clusters observed at an intermediate temperature of 55 °C but at otherwise identical conditions of concentration and flow is presented in Figure 2b. In agreement with Figure 2a, the singly charged metal complex (1,2)+ is still found to be by far the most abundant species observed in solution. In contrast with Figure 2a, however, when moderate heat is applied to the flow reactor, the formation of triply charged clusters is observed to be less pronounced. Specifically, triply charged (9,4)3+ and (11,5)3+ clusters are easily observed at low temperature but with decreased abundance after moderate heat has been applied to the flow reactor. Concomitant with the reduction in abundance of these triply charged clusters, the singly charged (2,2)+ complex is also partially consumed and exhibits decreased abundance in the mass spectrum in Figure 2b. This species is, therefore, likely an intermediate that yields larger clusters through higher-order reactions or decomposes to form more reactive smaller metal− ligand complexes. In addition, at 55 °C, the relative abundance of doubly charged (6,4)2+ is observed to increase slightly. Comparison of Figure 2a and b demonstrates that increasing the temperature of the flow reactor yields a qualitatively different distribution of ligated gold clusters and metal complexes than that produced at identical conditions but at lower temperature. Therefore, by using the flow reactor method and changing the temperature, it is possible to investigate the processes occurring at different time periods of the formation of gold clusters through reduction synthesis in solution without perturbing other relevant experimental parameters such as solution concentration and flow rate. In a previous publication, Hudgens and co-workers proposed a detailed series of chemical equations that describe the formation of gold nanoclusters through charged and neutral intermediates during batch reduction synthesis in solution at room temperature.24,27 By varying the ratio of the concentration of diphosphine ligands to the gold precursor, they were able to show that different size clusters form through the reduction of specific metal−ligand complexes. The method presented herein allows the use of another experimental variable, temperature, to study the formation of gold clusters in real time and at steady-state reaction conditions. As the temperature of the flow reactor is increased further, interesting trends in the abundance of different cluster species are observed. The mass spectrum of the solution flowing through the reactor at 70 °C is presented in Figure 2c. The relative abundance of (6,3)2+ is observed to be fairly consistent in each spectrum, indicating that this cluster remains stable in solution at higher temperatures. Similar to the results obtained at low temperature, at 70 °C, the relative abundance of the doubly charged (6,4)2+ cluster is found to be significantly higher than that of the triply charged (11,5)3+ cluster, which has almost disappeared from the mass spectrum in Figure 2c. Indeed, the relative abundance of (6,4)2+ compared to (11,5)3+ is much higher at 70 °C than that at either 25 or 55 °C. In addition, a large decrease in the abundance of the triply charged (9,4)3+ cluster is observed in Figure 2c. These results suggest that the larger triply charged clusters, while more abundant at low temperature, are not stable at temperatures above 55 °C. To better illustrate the qualitative trends in relative abundance observed with increasing temperature of the flow reactor, the temperature dependence of the abundances of representative clusters and metal complexes normalized to the total ion abundance is presented in Figure 3. The results for
ft) that was placed in contact with the surface of a hot plate. The temperature of the hot plate was monitored with a thermocouple attached to a digital meter. In this manner, the temperature of the flow reactor was controlled while monitoring the composition of the solution exiting the heated capillary with online mass spectrometry. At the lowest temperature (25 °C), steady-state formation of gold clusters using 1,4-bis(diphenylphosphino)butane as the capping ligand (L) resulted in the distribution of positively charged ions presented in the mass spectrum in Figure 2a.
Figure 2. Representative positive mode electrospray mass spectra of gold cluster solutions obtained at selected temperatures of (a) 25, (b) 55, and (c) 70 °C. The x-axis represents the mass-to-charge ratio (m/z), and the y-axis represents the normalized abundance.
The singly charged positive ions observed in Figure 2a include AuL2+, (1,2)+, which is the most abundant ion found in solution at low temperature, and Au2L2+, (2,2)+, which is relatively less abundant. This bracket notation will be used throughout the remainder of the text and in the figures to denote the number of gold atoms and diphosphine ligands as well as the ionic charge state of the clusters and complexes. Less abundant doubly charged gold clusters were also observed at room temperature including (6,3)2+, (4,3)2+, (6,4)2+, and (8,4)2+. These doubly charged species are listed in order of decreasing abundance as observed in Figure 2a. Two triply charged cationic clusters, (9,4)3+ and (11,5)3+, were also observed in the mass spectrum at low temperature. Therefore, without the application of any heat to the flow reactor, a wide distribution of ligated cationic gold clusters of different sizes and ionic charge states was observed in a straightforward manner. In addition, several singly charged gold ligand complexes were also identified. Many of these species have been observed previously as products at different stages of batch reduction synthesis in solution using mass spectrometry.26,29 In an effort to better qualitatively understand the process of gold cluster formation in solution, the temperature of the flow reactor was used as an experimental variable to systematically investigate the influence of applying heat on the distribution of gold clusters while all other experimental parameters such as reactant concentrations and solution flow rates were kept the same. It should be noted that the batch synthesis of gold clusters using diphosphine ligands is typically carried out at room temperature. Therefore, this study represents, to the best of our knowledge, the first effort to qualitatively understand the 8466
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
abundant than (6,3)2+ but also shows a similar trend of first increasing in abundance up to 60 °C and then decreasing in abundance at higher temperature. In addition, the abundance of the (8,4)2+ cluster shows the same sort of dependence on temperature as (6,3)2+ and (6,4)2+. This species initially increases in abundance with higher temperature and then drops off at temperatures above 60 °C. The behavior of two representative triply charged gold clusters formed in the flow reactor is illustrated in Figure 3c. The (11,5)3+ cluster is observed with highest abundance at room temperature and stays consistent with a small application of heat. Above 50 °C, however, this cluster decays in abundance until it is barely visible in the spectrum at higher temperatures. The (9,4)3+ cluster is found to be less sensitive to temperature than (11,5)3+ and exhibits only a small change in abundance with temperatures up to 80 °C. Above this temperature, the (9,4)3+ cluster also shows reduced abundance. The results described above, which were obtained at different temperatures in a steady-state flow reactor, share interesting similarities with previous findings for diphosphine ligated gold clusters synthesized in batch reactions at room temperature. For example, Hudgens and co-workers proposed that triply charged (9,4)3+ may be produced from reaction between (1,2)+ and (8,4)2+ for clusters ligated with the longer pentane-based ligand.27 The reverse process, involving decomposition of (9,4)3+ to form (8,4)2+ and (1,2)+, is shown in eq 1. (9, 4)3 + + 2L ↔ (8, 4)2 + + (1, 2)+
(1)
The steady decrease of (9,4)3+ and increase of (8,4)2+ and (1,2)+ with increasing temperature observed in Figure 3 between 25 and 60 °C indicates that eq 1 may also describe the reaction responsible for the trends in abundance observed for the shorter butane-based ligand used in this study. Alternatively, the (9,4)3+ cluster may also decompose in a reaction that generates (8,4)2+ and (1,1)+ according to eq 2.
Figure 3. Plots of the normalized abundance of selected gold clusters and complexes as a function of the temperature (ºC) of the capillary flow reactor: (a) complexes increasing in abundance with higher temperature, (b) clusters that exhibit a maximum in abundance at an intermediate elevated temperature, and (c) species that decrease in abundance with an increase in temperature.
(9, 4)3 + + L ↔ (8, 4)2 + + (1, 1)+
two singly charged gold complexes, (1,1)+ and (1,2)+, that exhibit increasing abundance at higher temperature are presented in Figure 3a. The (1,1)+ complex is observed to initially increase in abundance between 25 and 50 °C and then decrease slightly up to a temperature of around 70 °C. At temperatures higher than 70 °C, this complex appears to increase in abundance again. The (1,2)+ complex, in contrast, is far more abundant than (1,1)+ and is observed to increase gradually in abundance with temperature up to around 80 °C. The larger (2,2)+ complex is observed to be far less abundant than either (1,1)+ and (1,2)+ and is, therefore, presented in Figure 3c. This singly charged species exhibits a gradual monotonic decrease in abundance with increasing temperature followed by a small spike to higher abundance at around 80 °C. The behavior of four doubly charged ligated gold clusters is illustrated in Figure 3b. The smaller (4,3)2+ cluster is observed to have high abundance at low temperature, which remains constant with heating up to around 55 °C and then decreases slightly with further application of heat to the flow reactor. At temperatures up to 55 °C, therefore, the abundance of this cluster is found to be similar to that observed at room temperature. The slightly larger (6,3)2+ cluster is far more abundant than (4,3)2+ and exhibits different behavior with a steady increase in abundance with increasing temperature up to around 60 °C followed by a rapid decline in abundance at higher temperature. The related (6,4)2+ cluster is much less
(2)
The increasing abundance of (1,1)+ with temperature observed in Figure 3a is consistent with this alternative reaction, which also reduces the charge of the original 3+ cluster to 2+. The decrease in abundance of (2,2)+ observed in Figure 3c with increasing temperature is consistent with the previous identification of this species as a highly reactive radical by Hudgens and co-workers.24 In order to demonstrate the ability of the temperaturecontrolled flow reactor to produce cluster distributions that are characteristic of different stages of batch reduction synthesis, mass spectra were obtained at 15, 80, and 160 min after batch reactions were initiated at room temperature. This experiment was conducted by mixing the reagents in a single microfuge tube and loading the reacting solution into a single syringe. The reacting contents of the syringe were then aspirated into the mass spectrometer. All other experimental conditions were the same as those used with the mixing tee. The mass spectrum obtained after 15 min of batch reduction of the gold precursor in the presence of the butane-based ligand is presented in Figure 4a. Several prominent singly charged peaks are observed for metal complexes, including (1,1)+, (1,2)+, and (2,2)+. In addition, a small peak is observed for doubly charged (6,4)2+. The mass spectrum obtained after 80 min of batch reaction at room temperature is presented in Figure 4b. At this intermediate time, it is clear that (2,2)+ has decreased in 8467
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
Figure 4. Representative positive mode electrospray mass spectra of gold cluster solutions obtained at selected times of (a) 15, (b) 80, and (c) 160 min following the initiation of the batch reduction synthesis reaction. The x-axis represents the mass-to-charge ratio (m/z), and the y-axis represents the normalized abundance.
Figure 5. Representative positive mode electrospray mass spectra of gold cluster solutions obtained using a capillary flow reactor (a) 6 in. in length and (b) 5 feet in length. The x-axis represents the mass-tocharge ratio (m/z), and the y-axis represents the normalized abundance.
abundance, while (6,4)2+ has become more pronounced in the spectrum. Concomitant with the decrease in (2,2)+ is a small increase in the abundance of (6,3)2+. The mass spectrum obtained at the longest batch reaction time of 160 min is presented in Figure 4c. The (2,2)+ peak is completely absent at this time, while both the (6,3)2+ and (6,4)2+ peaks are increased in abundance. A prominent peak for (8,4)2+ is also observed at 160 min. These room-temperature results of batch synthesis are consistent with the spectra obtained previously by Hudgens and co-workers, which showed (6,4)2+ to be a stable species for clusters synthesized with the shorter propane-based ligand.32 In addition, Robinson and co-workers demonstrated the formation of (6,3)2+, (8,3)2+, (8,4)2+, (10,4)2+, and (11,5)3+ using a stronger sodium borohydride reducing agent after 1 day of reaction time.29 The shorter reaction time, milder reducing agent, and different mass spectrometer used in the current study may explain the absence of (8,3)2+, (10,4)2+, and (11,5)3+ in Figure 4. Close inspection of Figures 2 and 4 illustrates that by increasing the temperature of the flow reactor, it is possible to create steady-state cluster distributions that are characteristic of different stages of batch reduction synthesis. For example, by heating the flow reactor to around 80 °C, it is possible to reduce the abundance of (2,2)+ to a level similar to that observed following 160 min of batch reaction at room temperature. In addition, by heating the flow reactor to an intermediate temperature of around 60 °C, it is possible to increase the abundance of (6,3)2+, (6,4)2+, and (8,4)2+ to a level similar to that obtained at 160 min of batch reaction time at room temperature. As a further demonstration of the capabilities of the flow reactor, we examined the distribution of clusters produced at room temperature using capillaries of different lengths. The mass spectrum obtained for a reacting solution exiting a short 6 in. capillary is presented in Figure 5a. Similar to the cluster distribution obtained at a short batch reaction time in Figure 4a, (1,2)+ and (2,2)+ are observed to be the predominant species in solution with the 6 in. capillary. The mass spectrum of a solution exiting a much longer 5 ft capillary is presented in Figure 5b. It is clear that increasing the reaction time through lengthening the capillary results in a very different distribution of clusters, including (6,3)2+, (11,5)3+, and (6,4)2+. The similarities between the spectra obtained from batch
reduction at different times and from flow reactors of different lengths demonstrate the versatility of this steady-state approach. Furthermore, the use of temperature control is shown to be an effective method to access cluster distributions characteristic of different stages of batch reduction. It follows that transient species that may only be present for a short period of time at room temperature during batch synthesis as well as species that require long periods of time to form in solution may be obtained by varying the length and temperature of the flow reactor.
■
CONCLUSIONS Employing electrospray ionization mass spectrometry combined with a temperature-controlled variable length flow reactor, a new approach has been demonstrated for investigating the reduction of metal complexes in the presence of organic ligands to form nanoclusters. For a model system consisting of diphosphine-ligated gold clusters, a key singly charged (2,2)+ intermediate was observed with high abundance at low and intermediate temperatures and reduced abundance at high temperature. The doubly charged (4,3)2+, (6,3)2+, (6,4)2+, and (8,4)2+ clusters initially increased in abundance up to 60 °C and then decreased in abundance at higher temperatures. A stable abundance of triply charged (11,5)3+ was observed up to 50 °C, at which point this species decreased in abundance and was ultimately completely consumed at higher temperatures. The triply charged (9,4)3+ cluster also exhibited a steady decrease in abundance with increasing temperature. Our results illustrate how temperature and reactor length may be used as parameters to obtain steady-state distributions of clusters that are characteristic of different stages of batch reduction in solution. We anticipate that a secondgeneration flow reactor will be used to produce nanoclusters across a wide range of sizes and compositions for quantitative analysis using mass spectrometry and ion spectroscopy as well as characterization on surfaces following soft landing of massselected ions. 8468
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
■
Article
EXPERIMENTAL METHODS Synthesis of Ligated Gold Clusters. All chemicals were used as purchased with no further purification. Chloro(triphenylphosphine)gold(I) (99.9%, Sigma-Aldrich) was dissolved in methanol (99.8%, Fisher) to generate 50 mL of a 0.1 mM solution. 1,4-Bis(diphenylphosphino)butane (98%, SigmaAldrich), a bidentate capping ligand, was added to the gold solution to generate 50 mL of a 0.1 mM solution. In a separate Pyrex jar, a solution of reducing agent, borane tert-butylamine (97%, Sigma-Aldrich), was added to generate 50 mL of 0.5 mM solution in methanol. Both solutions were mixed in a highpressure mixing tee (M-540, IDEX Corporation) using two separate syringe pumps (KD Scientific). The fused silica capillary (Polymicro Technologies) introducing the sample into the mass spectrometer was heated using a hot plate (Fisher) in approximately 10 °C increments. The temperature of the flow reactor was determined using a K-type thermocouple connected to a meter (Fluke). The heated solution was then introduced to an electrospray ionization mass spectrometer (Bruker HCT Ion Trap, ESI-MS). The evolution of the mass spectra with time was monitored at room temperature without a mixing tee by combining the reagents in a single syringe (batch synthesis) and aspirating the reacting solution into the mass spectrometer. The influence of the length of the capillary on the distribution of clusters was determined by aspirating an identical reacting solution into 6 in. and 5 ft capillaries. The retention times of the 6 in. and 5 ft capillaries were estimated to be approximately 2 and 25 min, respectively. Analysis of Cluster Solutions. An electrospray ionization mass spectrometer, Bruker HCT ion trap (m/z = 100−2500), was used for analysis of the cluster solutions exiting the flow reactor. In the positive ion mode, the source conditions employed were as follows: capillary 4000 V, nebulizer 20 psi, drying gas 5 L/min, and drying gas temperature 200 °C. The internal instrument settings were as follows: skimmer 25 V, capillary exit 75 V, Oct 1 DC 15 V, Oct 2 DC −5 V, trap drive 100, Oct RF 75 Vpp, lens 1−10 V, and lens 2−50 V. The stock gold precursor and ligand solutions were diluted by a factor of 10 in methanol and loaded into glass syringes (Hamilton). The reducing agent solution was diluted to a 0.20 mM concentration. Both solutions were introduced at a flow rate of 100 μL/h using two separate syringe pumps (KD Scientific). The molecular formulas of the clusters and complexes were assigned on the basis of their mass-to-charge ratios (m/z), the charge-state-dependent separation of the individual peaks, and the overall shape of the isotopic distributions. Additionally, experimental spectra were compared with spectra simulated using the molecular weight calculator program (http://omics. pnl.gov/software/-MWCalculator.php).
■
support from the Linus Pauling Fellowship and the Laboratory Directed Research and Development Program at the Pacific Northwest National Laboratory (PNNL). This work was performed using EMSL, a national scientific user facility sponsored by the DOE’s Office of Biological and Environmental Research and located at PNNL. PNNL is operated by Battelle for the U.S. DOE.
■
REFERENCES
(1) Jin, R. C. Quantum Sized, Thiolate-Protected Gold Nanoclusters. Nanoscale 2010, 2, 343−362. (2) Li, G.; Jin, R. C. Atomically Precise Gold Nanoclusters as New Model Catalysts. Acc. Chem. Res. 2013, 46, 1749−1758. (3) Nimmala, P. R.; Dass, A. Au36(SPh)23 Nanomolecules. J. Am. Chem. Soc. 2011, 133, 9175−9177. (4) Dreaden, E. C.; Mackey, M. A.; Huang, X. H.; Kang, B.; El-Sayed, M. A. Beating Cancer in Multiple Ways Using Nanogold. Chem. Soc. Rev. 2011, 40, 3391−3404. (5) Herzing, A. A.; Kiely, C. J.; Carley, A. F.; Landon, P.; Hutchings, G. J. Identification of Active Gold Nanoclusters on Iron Oxide Supports for CO Oxidation. Science 2008, 321, 1331−1335. (6) Wilcoxon, J. P.; Abrams, B. L. Synthesis, Structure and Properties of Metal Nanoclusters. Chem. Soc. Rev. 2006, 35, 1162−1194. (7) Yeh, Y. C.; Creran, B.; Rotello, V. M. Gold Nanoparticles: Preparation, Properties, and Applications in Bionanotechnology. Nanoscale 2012, 4, 1871−1880. (8) Kudgus, R. A.; Bhattacharya, R.; Mukherjee, P. Cancer Nanotechnology: Emerging Role of Gold Nanoconjugates. AntiCancer Agents Med. Chem. 2011, 11, 965−973. (9) Li, G.; Jiang, D. E.; Liu, C.; Yu, C. L.; Jin, R. C. Oxide-Supported Atomically Precise Gold Nanocluster for Catalyzing Sonogashira Cross-Coupling. J. Catal. 2013, 306, 177−183. (10) Tsunoyama, H.; Liu, Y. M.; Akita, T.; Ichikuni, N.; Sakurai, H.; Xie, S. H.; Tsukuda, T. Size-Controlled Synthesis of Gold Clusters as Efficient Catalysts for Aerobic Oxidation. Catal. Surv. Asia 2011, 15, 230−239. (11) Turner, M.; Golovko, V. B.; Vaughan, O. P. H.; Abdulkin, P.; Berenguer-Murcia, A.; Tikhov, M. S.; Johnson, B. F. G.; Lambert, R. M. Selective Oxidation with Dioxygen by Gold Nanoparticle Catalysts Derived from 55-Atom Clusters. Nature 2008, 454, 981−U931. (12) Yao, H. Optically Active Gold Nanoclusters. Curr. Nanosci. 2008, 4, 92−97. (13) Perez, A.; Melinon, P.; Dupuis, V.; Bardotti, L.; Masenelli, B.; Tournus, F.; Prevel, B.; Tuaillon-Combes, J.; Bernstein, E.; Tamion, A.; et al. Functional Nanostructures from Clusters. Int. J. Nanotechnol. 2010, 7, 523−574. (14) Liu, C.; Li, G.; Pang, G. S.; Jin, R. C. Toward Understanding the Growth Mechanism of AuN(Sr)M Nanoclusters: Effect of Solvent on Cluster Size. RSC Adv. 2013, 3, 9778−9784. (15) Jin, R. C.; Qian, H. F.; Wu, Z. K.; Zhu, Y.; Zhu, M. Z.; Mohanty, A.; Garg, N. Size Focusing: A Methodology for Synthesizing Atomically Precise Gold Nanoclusters. J. Phys. Chem. Lett. 2010, 1, 2903−2910. (16) Guidez, E. B.; Hadley, A.; Aikens, C. M. Initial Growth Mechanisms of Gold−Phosphine Clusters. J. Phys. Chem. C 2011, 115, 6305−6316. (17) Ivanov, S. A.; Arachchige, I.; Aikens, C. M. Density Functional Analysis of Geometries and Electronic Structures of Gold−Phosphine Clusters. The Case of Au4(PR3)42+ and Au4(μ2-I)2(PR3)4. J. Phys. Chem. A 2011, 115, 8017−8031. (18) Hong, S.; Shafai, G.; Bertino, M.; Rahman, T. S. Toward an Understanding of Ligand Selectivity in Nanocluster Synthesis. J. Phys. Chem. C 2011, 115, 14478−14487. (19) Shafai, G.; Hong, S. Y.; Bertino, M.; Rahman, T. S. Effect of Ligands on the Geometric and Electronic Structure of Au13 Clusters. J. Phys. Chem. C 2009, 113, 12072−12078.
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the U.S. Department of Energy (DOE), Office of Basic Energy Sciences, Division of Chemical Sciences, Geosciences and Biosciences. A.O. acknowledges support from the DOE Science Undergraduate Laboratory Internship (SULI) program. G.E.J. acknowledges partial 8469
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
The Journal of Physical Chemistry A
Article
(20) Bertino, M. F.; Sun, Z. M.; Zhang, R.; Wang, L. S. Facile Syntheses of Monodisperse Ultrasmall Au Clusters. J. Phys. Chem. B 2006, 110, 21416−21418. (21) Golightly, J. S.; Gao, L.; Castleman, A. W.; Bergeron, D. E.; Hudgens, J. W.; Magyar, R. J.; Gonzalez, C. A. Impact of Swapping Ethyl for Phenyl Groups on Diphosphine-Protected Undecagold. J. Phys. Chem. C 2007, 111, 14625−14627. (22) Bergeron, D. E.; Coskuner, O.; Hudgens, J. W.; Gonzalez, C. A. Ligand Exchange Reactions in the Formation of DiphosphineProtected Gold Clusters. J. Phys. Chem. C 2008, 112, 12808−12814. (23) Pettibone, J. M.; Hudgens, J. W. Synthetic Approach for Tunable, Size-Selective Formation of Monodisperse, DiphosphineProtected Gold Nanoclusters. J. Phys. Chem. Lett. 2010, 1, 2536−2540. (24) Hudgens, J. W.; Pettibone, J. M.; Senftle, T. P.; Bratton, R. N. Reaction Mechanism Governing Formation of 1,3-Bis(diphenylphosphino)propane-Protected Gold Nanoclusters. Inorg. Chem. 2011, 50, 10178−10189. (25) Pettibone, J. M.; Hudgens, J. W. Gold Cluster Formation with Phosphine Ligands: Etching as a Size-Selective Synthetic Pathway for Small Clusters? ACS Nano 2011, 5, 2989−3002. (26) Pettibone, J. M.; Hudgens, J. W. Predictive Gold Nanocluster Formation Controlled by Metal−Ligand Complexes. Small 2012, 8, 715−725. (27) Pettibone, J. M.; Hudgens, J. W. Reaction Network Governing Diphosphine-Protected Gold Nanocluster Formation from Nascent Cationic Platforms. Phys. Chem. Chem. Phys. 2012, 14, 4142−4154. (28) Johnson, G. E.; Priest, T.; Laskin, J. Synthesis and Characterization of Gold Clusters Ligated with 1,3-Bis(dicyclohexylphosphino)propane. ChemPlusChem 2013, 78, 1033−1039. (29) Robinson, P. S. D.; Nguyen, T. L.; Lioe, H.; O’Hair, R. A. J.; Khairallah, G. N. Synthesis and Gas-Phase Uni- and Bi-Molecular Reactivity of Bisphosphine Ligated Gold Clusters, [AuxLy]n+. Int. J. Mass Spectrom. 2012, 330, 109−117. (30) Shichibu, Y.; Suzuki, K.; Konishi, K. Facile Synthesis and Optical Properties of Magic-Number Au13 Clusters. Nanoscale 2012, 4, 4125− 4129. (31) Robinson, P. S. D.; Khairallah, G. N.; Silva, G.; Lioe, H.; O’Hair, R. A. J. Gold-Mediated C−I Bond Activation of Iodobenzene. Angew. Chem., Int. Ed. 2012, 51, 3812−3817. (32) Bergeron, D. E.; Hudgens, J. W. Ligand Dissociation and Core Fission from Diphosphine-Protected Gold Clusters. J. Phys. Chem. C 2007, 111, 8195−8201. (33) Wilson, D. J.; Konermann, L. Mechanistic Studies on Enzymatic Reactions by Electrospray Ionization MS Using a Capillary Mixer with Adjustable Reaction Chamber Volume for Time-Resolved Measurements. Anal. Chem. 2004, 76, 2537−2543. (34) Wilson, D. J.; Konermann, L. A Capillary Mixer with Adjustable Reaction Chamber Volume for Millisecond Time-Resolved Studies by Electrospray Mass Spectrometry. Anal. Chem. 2003, 75, 6408−6414. (35) Kolakowski, B. M.; Simmons, D. A.; Konermann, L. StoppedFlow Electrospray Ionization Mass Spectrometry: A New Method for Studying Chemical Reaction Kinetics in Solution. Rapid Commun. Mass Spectrom. 2000, 14, 772−776. (36) Johnson, G. E.; Wang, C.; Priest, T.; Laskin, J. Monodisperse Au11 Clusters Prepared by Soft Landing of Mass Selected Ions. Anal. Chem. 2011, 83, 8069−8072. (37) Johnson, G. E.; Priest, T.; Laskin, J. Charge Retention by Gold Clusters on Surfaces Prepared Using Soft Landing of Mass Selected Ions. ACS Nano 2012, 6, 573−582. (38) Johnson, G. E.; Priest, T.; Laskin, J. Coverage-Dependent Charge Reduction of Cationic Gold Clusters on Surfaces Prepared Using Soft Landing of Mass-Selected Ions. J. Phys. Chem. C 2012, 116, 24977−24986.
8470
dx.doi.org/10.1021/jp501809r | J. Phys. Chem. A 2014, 118, 8464−8470
