Monolayer-Protected Gold Nanoparticles as a Stationary Phase for

Aug 5, 2003 - Gwen M. Gross,† David A. Nelson,‡ Jay W. Grate,‡ and Robert E. Synovec*,†. Center for Process Analytical Chemistry, Department o...
0 downloads 0 Views 169KB Size
Download PDF
Anal. Chem. 2003, 75, 4558-4564

Monolayer-Protected Gold Nanoparticles as a Stationary Phase for Open Tubular Gas Chromatography Gwen M. Gross,† David A. Nelson,‡ Jay W. Grate,‡ and Robert E. Synovec*,†

Center for Process Analytical Chemistry, Department of Chemistry, Box 351700, University of Washington, Seattle, Washington 98195, and Pacific Northwest National Laboratory, Richland, Washington 99352

The use of a thin film of monolayer-protected gold nanoparticles (MPNs) as a stationary phase for gas chromatography (GC) is reported. Deposition of a MPN film was obtained in a 2-m, 530-µm-i.d. deactivated silica capillary using gravity to force the solution containing the MPN material through the capillary. By SEM analysis, the average film thickness was determined to be 60.7 nm. The retention behavior for the dodecanethiol MPN column was studied using four compound classes (alkanes, alcohols, aromatics, ketones), and retention orders were objectively compared to a commercially available column (AT-1, 100-nm film thickness). Separation of an eightcomponent mixture was performed using both isothermal and temperature-programming methods with the dodecanethiol MPN phase and compared to an isothermal separation with the AT-1 phase. The AT-1 phase separation had an efficiency, N, of 6200 (k′ ) 0.33) while the dodecanethiol MPN phase separation had an efficiency, N, of 5700 (k′ ) 0.21) for the same analyte, octane. The reduced plate height, h, for octane was found to be less than 1 at the optimum linear flow velocity, indicating the MPN column operated near the optimum possible performance level. Robustness of the MPN phase is also discussed with consistent performance observed over several months. Overall, MPNs appear promising as a stationary-phase material for GC and as an experimental platform to study their thermodynamic and mass-transfer properties. Nanoparticles and nanoparticle-based materials are attracting great interest for their unique properties and potential for application in diverse areas.1-3 Use of nanoparticles for chromatographic applications has proven to be advantageous. For example, capillary electrophoresis has found applications of nanoparticles as a pseudostationary phase, enhancing selectivity, as well as a * Corresponding author. E-mail: [email protected]. † University of Washington. ‡ Pacific Northwest National Laboratory. (1) Brust, M.; Fink, J.; Bethell, D.; Schiffrin, D.; Kiely, C. J. Chem. Soc., Chem. Commun. 1995, 1655-1656. (2) Brust, M.; Bethell, D.; Kiely, C.; Schiffrin, D. Langmuir 1998, 14, 54255429. (3) Templeton, A.; Wuelfing, M.; Murray, R. Acc. Chem. Res. 2000, 33, 27-36.

4558 Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

production method for nanoparticles.4-6 Monolayer-protected nanoparticles (MPNs) are of particular interest because the surface monolayer stabilizes them relative to aggregation and their properties are influenced by the structure of the monolayerforming molecules. More recently, the use of high-performance liquid chromatography for the separation of different MPNs based upon nanoparticle size and the chosen monolayer was reported.7 An ordered arrangement of the thiol linkages with uniform, monolayer coverage can be obtained on flat gold surfaces.8 However, with small-diameter MPNs, the curvature of the gold core leads to a larger population of core surface gold atoms, which are defect sites in 2D self-assembled monolayers.9 This disorder is a desired effect, promoting a noncrystalline interparticle structure, facilitating fast mass transfer via diffusion. Studies of dodecanethiol MPNs have shown the closest core-core spacing is equal to the extended chain length of a single dodecanethiol monolayer, demonstrating the chains interpenetrate to form a disordered packing structure.3 Thus, a thin MPN film should function as an efficient stationary phase for gas chromatography (GC). It is envisaged that similar chromatographic efficiencies relative to commercial stationary phases can be realized. Since a wide range of thiol-based organic monolayers is available, ample chemical selectivity could be achieved for different MPN-based GC stationary phases. This report initiates a GC-based platform for thermodynamic and mass-transfer studies of MPNs. With a firm understanding of the chemical selectivity and mass-transfer properties, one could potentially devise MPNs that are tuned for a particular GC application of interest or for other chemical analyzers and sensors as well. The characterization of the MPNs is of considerable interest. In particular, Au-based MPNs have been utilized in a sensor format,10 but a greater understanding of their properties as a (4) Neiman, B.; Grushka, E.; Ovadia, L. Anal. Chem. 2001, 73, 5220-5227. (5) Sunderhaus, J.; Steinbock, B.; Steinbock, O. Abstr. Pap. Am. Chem. Soc. 2002, 223, 360-PHYS Part 2. (6) Viberg, P.; Jornten-Karlsson, M.; Petersson, P.; Spegel, P.; Nilsson, S. Anal. Chem. 2002, 74, 4595-4601. (7) Jimenez, V. L.; Leopold, M. C.; Mazzitelli, C.; Jorgenson, J. W.; Murray, R. W. Anal. Chem. 2003, 75, 199-206. (8) Sabapathy, R.; Bhattacharyya, S.; Leavy, M.; Cleland, W.; Hussey, C. Langmuir 1998, 14, 124-136. (9) Hostetler, M.; Wingate, J.; Zhong, C.; Harris, J.; Vachet, R.; Clark, M.; Londono, J.; Green, S.; Stokes, J.; Wignall, G.; Glish, G.; Porter, M.; Evans, N.; Murray, R. Langmuir 1998, 14, 17-30. (10) Grate, J. W.; Nelson, D. A.; Skaggs, R. Anal. Chem. 2003, 75, 1868-1879. 10.1021/ac030112j CCC: $25.00

© 2003 American Chemical Society Published on Web 08/05/2003

sensing medium was desired, serving as impetus for this work. Previous characterization has been done using techniques such as electron microscopies and thermal gravimetric analysis as well as other spectroscopic and thermochemical techniques.9-11 The GC platform reported herein can be used to further characterize the gold-based MPNs in a way that is different from the studies done on nanoparticles to date. Using techniques previously established for materials characterization, the sorptive properties of the nanoparticles as well as their chemical retention behavior can be explored.12-14 Gas chromatographic measurements provide a reliable, reproducible, and efficient platform for measuring retention behavior of many compounds on a given MPN phase and, hence, for evaluating chemical selectivity. Reported herein is the development of a stationary phase consisting of dodecanethiol-protected gold nanoparticles and initial characterization of this phase by and for capillary GC. Successful deposition of a thin film of MPNs within a capillary for GC was explored for the first time as well as the methodology for that deposition. Dodecanethiol MPN was coated in a thin film, nominally 60 nm thick, in a 530-µm-i.d. capillary with 2-m length. Separations of different compound classes and a test mixture, as well as characterization studies of the nanoparticle stationary phase, were completed. The characterization studies included exploration of the retention behavior of the stationary phase and chromatographic efficiency. An objective comparison between the novel dodecanethiol MPN phase and a commercial AT-1 GC stationary phase was completed. Mass-transfer behavior of the dodecanethiol MPN phase was studied by observing the band broadening of several analytes as a function of linear flow velocity. The retention and mass-transfer behavior provided by the dodecanethiol MPN phase will serve as a benchmark for comparison with future MPN-based stationary phases. EXPERIMENTAL SECTION Reagents and Chemicals. All chemicals were reagent grade or a higher grade. Many of the bulk reagents were purchased from Fisher (Fisher Scientific, Fairlawn, NJ). Chromatographic standards from PolyScience (AccuStandard, Inc., New Haven, CT) were used as analytes for the GC separations with the exception of samples containing benzene and hexane (both Fisher), anisole, octane, decane (all Aldrich, Milwaukee, WI), and chlorobenzene (J. T. Baker, Phillipsburg, NJ) (Table 1). Water used in the production of the column was filtered using a NANOpure II filter system (Barnstead/Thermolyne Corp., Dubuque, IA). The GC carrier gas was hydrogen and was supplied by a hydrogen generator (Whatman Inc., Kent, U.K.). Nanoparticle Synthesis. The method used for the synthesis of the gold-based MPNs was taken from Wolhtjen and Snow15 but was initially developed by Brust.16 The basic formula for the synthesis is (C8H17)4NBr

HAuCl4‚3H2O + RSH 9 8 Au:SR NaBH 4

Table 1. Twelve Components Chosen for Evaluation of the Dodecanethiol Monolayer-Protected Gold Nanoparticle Stationary Phasea compound

bp (°C)

k′ (MPN)

log L16

k′ (AT-1)

hexane ethanol benzene 3-pentanone 1-butanol octane chlorobenzene 1-pentanol 3-heptanone anisole 3-octanone decane

69 78 80 102 118 126 132 138 148 154 168 174

0.01 0.02 0.06 0.06 0.16 0.21 0.65 0.43 0.42 0.82 0.71 1.47

2.668 1.485 2.786

0.04 0.01 0.08 0.12 0.10 0.34 0.45 0.27 0.66 0.80 1.59 1.99

2.601 3.677 3.657 3.106 3.890 4.686

a The retention factors, k′, were determined using methanol as the dead time marker. For comparison, the elution data for two other nonpolar stationary phases are included: log L16 and AT-1. The elution order for hexadecane as a stationary phase is given by the log L16 values.

monolayer. Tetraoctylammonium bromide (4.5 g) was dissolved in 17 mL of toluene. A solution consisting of 65 mL of water with 1.03 g of hydrogen tetrachloroaurate was added to the toluene solution in a three-neck round-bottom flask attached to a nitrogen bubbler system. An additional 3 mL of water was used to completely transfer the gold-containing solution to the flask. Dodecanethiol (0.50 g) was dissolved in 2 mL of toluene and added to the flask. Sodium borohydride (0.79 g) was taken from under vacuum and dissolved in 50 mL of water while stirring. This solution was then added dropwise to the flask over a 5-min period. Once the addition was complete, the flask was capped and allowed to stir for 3.5 h, after which time the solution was transferred to a separation funnel. The toluene layer was washed twice with 150 mL of water. The product was placed under a nitrogen stream overnight prior to purification. Purification was done using 350 mL of ethanol, allowing the gold MPNs to precipitate from solution. The solution was centrifuged and the solvent decanted. Two additional 30-mL ethanol washes were completed in the same way. The product was dissolved in 5 mL of dichloromethane and transferred to a tared vial. An additional 3 mL of dichloromethane was used to completely transfer the dodecanethiol MPN product to the tared vial. The solvent was evaporated off under a stream of nitrogen, and a product yield of 0.78 g was achieved. The purity of the product was tested using thin-layer chromatography (TLC) and Fourier transform infrared spectroscopy (FT-IR) (Vector 33, Bruker Optics, Billerica, MA). Of particular interest was the presence of residual thiol or disulfide present in the product. Using a 9:1 dichloromethane/methanol solution, the TLC plates were run and then stained with iodine. No additional spots were detected with the iodine stain, indicating that neither residual thiol nor disulfide was present at detectable levels in the final product. The FT-IR spectra (KBr pellet) of the MPN product were consistent with the literature and confirmed the absence of any

A 1:1 molar ratio of gold to thiol was used, with a dodecanethiol (11) Warner, M.; Reed, S.; Hutchison, J. Chem. Mater. 2000, 12, 3316-3320. (12) Abraham, M.; Poole, C.; Poole, S. J. Chromatogr., A 1999, 842, 79-114. (13) Abraham, M. H.; Andonian-Haftvan, J.; Du, C. M.; Diart, V.; Whiting, G.; Grate, J. W.; McGill, R. A. J. Chem. Soc., Perkin Trans. 2 1995, 369-378.

(14) Abraham, M.; Ballantine, D.; Callihan, B. J. Chromatogr., A 2000, 878, 115124. (15) Wohltjen, H.; Snow, A. W. Anal. Chem. 1998, 70, 2856-2859. (16) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D.; Whyman, R. J. Chem. Soc., Chem. Commun. 1994, 801-802.

Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

4559

Figure 1. Characterization of the synthesized dodecanethiol monolayer protected gold nanoparticles. (A) Transmission electron micrograph of a thin film of the dodecanethiol MPNs. (B) Histogram of the measured core size distribution from the TEM data. (C) Thermal gravimetric analysis plot indicating the loss of the dodecanethiol monolayer from the gold core.

significant amount of phase-transfer agent in the final product.17,18 However, additional purification steps may be warranted to fully eliminate the phase-transfer agent.19 The final dodecanethiol MPN product was black and stored dry in a sealed vial. Characterization of the nanoparticles was done using thermal gravimetric analysis (TGA; model STA409A, Netzsch, Selb, Germany) and transmission electron microscopy (TEM; model JEM 2010, JEOL, Tokyo, Japan) with TEM software (Digital Micrograph, Gatan Inc., Pleasanton, CA). Characterization results are presented in Figure 1. Using TGA, it was determined that 75.09% by mass of the nanoparticles was due to the gold core with most of the monolayer loss occurring between 200 and 300 °C (Figure 1C). While the synthesis mole ratio for gold to thiol was 1:1, the final product mole ratio calculated from the TGA was ∼3: 1. The presence of nanoparticles was confirmed by TEM (Figure 1A) using a copper grid airbrushed with a dilute solution (0.2% by mass) of the nanoparticles dissolved in dichloromethane. The TEM image indicates the nanoparticle core size was polydisperse with diameters ranging from about 1.5 to 5 nm (Figure 1B) with an average of 3.2 nm. Column Preparation. Deactivated silica capillary, 530-µm i.d. with a 2-m length, was used for the production of the open tubular MPN column (Supelco, Bellefonte, PA). The capillary was washed with three 50-µL volumes of methylene chloride and with three volumes of water. The methylene chloride washes were evaporated from the capillary using a nitrogen stream while the water washes were evaporated from the capillary by baking in an oven at 300 °C for 1 h. This same treatment was used for a “blank” column (a column without MPN stationary phase to compare with the column with the dodecanethiol MPN phase, i.e., the MPN column). Sufficient dodecanethiol MPN material was placed in a 100µL conical insert inside a standard injection vial (Agilent Technologies, Palo Alto, CA), and 50 µL of methylene chloride was added. The resulting nanoparticle solution was introduced to prepare the MPN column via capillary action, resulting in a plug (17) Hostetler, M. J.; Stokes, J. J.; Murray, R. W. Langmuir 1996, 12, 36043612. (18) Cai, Q. Y.; Zellers, E. T. Anal. Chem. 2002, 74, 3533-3539. (19) Waters, C. A.; A. J., M.; Johnson, K. A.; Schiffrin, D. J. Chem. Commun. 2003, 540-541.

4560

Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

length of ∼4 cm. With the column in a vertical position, gravitational force was utilized to move the solution plug and the nanoparticles were deposited in the capillary via evaporation. When the solution plug reached the far end of the capillary, the column was inverted and the process repeated. Additional solution volumes were added until, via visual inspection, a uniform brown color was achieved. To help avoid MPN phase clumping a lowintensity hot air gun (Conair ProStylus 1500, Conair Corp., Stamford, CT) was used as a low heat source for faster evaporation of the solvent. Several MPN columns were produced with varying shades from light tan to black. The uniform brown for the thinfilm MPN column displayed minimal clumping with no observable bare spots and resulted in very reproducible GC separations. Determination of stationary-phase thickness, coverage, and confirmation of stationary-phase presence were done using SEM as will be described later. SEM Imaging of the MPN Stationary Phase within the Capillary Column. All images were obtained using a LEO 982 field emission SEM (LEO Electron Microscopy Ltd., Cambridge, England). Prior to analysis, four randomly selected pieces of capillary column, three containing MPNs and one blank, were mounted and then coated with a thin platinum layer using a Ladd electronic sputter coating machine (Ladd Research, Williston, VT). Images were obtained in both the normal and backscatter modes. Elemental analysis of the visible stationary phase was performed in the backscattering mode. Chromatographic Instrumentation. All chromatograms were obtained with an Agilent 6890 gas chromatograph using a standard FID detector and injector with ChemStation computer control (Agilent Technologies, Palo Alto, CA). Data were imported into Matlab (MathWorks, Natick, MA) for subsequent analysis. Chromatographic Experiments. All chromatograms were obtained with an injection source and FID temperature of 200 °C and a hydrogen carrier gas. Samples were introduced by headspace injection. A 2-mm depth of sample containing the components of interest was placed in the bottom of a standard injection vial, below the injection depth of the syringe into the vial. An injection size of 0.5 µL was used with a split injection of 50:1 or 1:1 as specified. The injection volume and split procedure did not introduce significant chromatographic band broadening. The vial

Figure 2. SEM image of the dodecanethiol MPN stationary phase within the capillary. Area A is the silica capillary. Area B is the goldbased dodecanethiol MPN stationary phase, 60.1 nm thick at this location. Using a random piece of column, this image was obtained using an end-on cross-sectional view.

was not heated, so the amount of each component present in the injected sample was determined by component volatility at room temperature. Using samples prepared in this manner did result in sample-to-sample concentration variation for each component. However, this sample preparation and injection procedure was satisfactory for chromatographic evaluation of retention time, efficiency, N, and reduced plate height, h, for the components of interest. Due to the relatively short length and wide diameter of the MPN column, a short restrictor column (100 µm i.d., 0.5 m in length) consisting of a methyl-deactivated silica capillary followed the MPN column to avoid FID blowout (Supelco). The oven temperature was 50 °C unless otherwise noted. Additional retention due to the restrictor column was found to be insignificant. The MPN stationary phase was compared to a commercial nonpolar AT-1 stationary phase (Alltech Associates, Inc., Deerfield, IL). All parameters were identical for the comparison except stationary-phase thickness. The AT-1 phase thickness of 0.10 µm was as near as possible to the MPN phase thickness, nominally 0.060 µm. The commercial column dimensions and chromatographic conditions were the same as those used for the MPN column separation: 530-µm i.d. and 2-m length, temperature of 50 °C, and linear flow velocity of 22 cm/s. RESULTS AND DISCUSSION Using SEM analysis, the presence of the gold-containing dodecanethiol MPN stationary phase within the capillary GC column was confirmed. Stationary-phase thickness was measured at random spots along the column. Visually, the majority of the surface area of the inside wall was a uniform brown (∼97%), while a minor fraction of the surface area contained small black clumps (∼3%). Both uniform and nonuniform areas were imaged using the SEM. For a typical location on the capillary wall appearing uniform, the stationary-phase thickness was 60.1 nm (Figure 2). There was very little variation ((2 nm) around the capillary circumference in these predominating uniform regions. For a rare area of nonuniformity, the average thickness was 55.7 nm, ranging from 12.7 to 115.2 nm. No bare areas of silica capillary were observed during any of the SEM measurements. The average film thickness for the full length of column was 60.7 nm.

The retention characteristics of the dodecanethiol MPN stationary phase were studied using four different compound classes (Figure 3). The analytes studied had a boiling point range from 69 to 174 °C (Table 1). It should be noted that the analytes are not retained strictly upon the basis of boiling point. 3-Octanone has a higher boiling point (168 °C), but the less polar anisole, with a boiling point of 154 °C, is more highly retained. Retention order is based upon boiling point only within a homologous series of compounds. This can be seen within the three homologous series of alkanes, ketones, and alcohols. Additional support for the separation due to the MPN stationary phase was obtained using the “blank” column. This experiment addressed the possibility of additional retention if exposed silica areas were present within the MPN column. Each of the four compound classes (Table 1) were analyzed with the blank column under identical conditions as the MPN column. The bare silica capillary did not contribute significantly to the net analyte retention. This experiment confirmed the separations achieved with the MPN column were not due to bare silica areas, consistent with the SEM results. An expanded study of the retention behavior for the alkane homologous series was also done (C10-C16). The log of the retention factor (k′) was plotted against log L,16 the Oswald solubility coefficient (gas-hexadecane partition coefficient at 25 °C). For straight-chain alkanes, log of k′ is directly related to the log L16 (Figure 4).14 The slope of the line is the l value of the stationary phase, used to describe the stationary phase in the solvation parameter model.12,14,20 The l term is a measure of the hydrophobicity of the stationary phase and is an indicator of the ability of a stationary phase to separate members of a homologous series.12 The solvation model is temperature dependent and so a change in the slope of the line was seen for the three temperatures examined (100, 75, 50 °C). The k′ values at 25 °C were extrapolated from the other three temperature series and plotted against the log L16 values. The linearity of the data indicates the dodecanethiol MPN stationary phase is well correlated with the nonpolar n-alkane series used (R2 of ∼0.99 for all temperature series) and behaves as expected by the model. The increased slope (l value) indicated the stationary phase separates n-alkanes better with a decrease in temperature. Lower temperatures are not always conducive to GC analysis, however, so it is useful to know that extrapolation of retention values can be done for the MPN stationary phase for use in future studies using the solvation model. From the original 12 analytes (Table 1), 8 analytes were selected to demonstrate a more complex mixture separation. Each of the four compound classes were represented in the mixture. An acceptable isothermal separation was obtained with the dodecanethiol MPN column in a little over 30 s as shown in Figure 5A, comparing reasonably with the commercial AT-1 column separation in Figure 5C. Using temperature programming, the same mixture was separated with the MPN column in just over 25 s with improved resolution (Figure 5B). The isothermal separation was done at 50 °C, far below the boiling point of the higher boiling analytes in the mixture. It is not surprising that an improvement in the separation was observed for the later-eluting (20) Du, C. M. Thesis. The Application of Physiochemical Desciptors to the Characterisation of Liquid and Solid Phases. University College, London, 1995.

Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

4561

Figure 3. Chromatographic separations using the dodecanethiol MPN column of four different compound classes: (A) alkanes, (B) aromatics, (C) alcohols, and (D) ketones. The extra peak eluting before 3-pentanone in (D) is an impurity peak. The separations were obtained using a 2-m, 530-µm-i.d. column with dodecanethiol MPN phase nominally 60 nm thick, at 50 °C with a 0.5-µL headspace injection using a 50:1 split on the inlet, operated under constant pressure conditions at 25 psi (∼22 cm/s).

Figure 4. The log of the retention factor (k′) versus log L16, the Oswald solubility coefficient (gas-hexadecane partition coefficient at 25 °C), for n-alkanes C10-C16. The k′ data were experimentally obtained at 50, 75, and 100 °C. The k′ values at 25 °C were extrapolated from the other three temperature data sets and plotted against the log L16 values.

peaks (Figure 5B), even for the short temperature program (4080 °C at 70 °C/min). Band broadening and separation efficiency were examined for the dodecanethiol MPN column. The reduced plate height, h, was plotted versus the linear flow velocity, u, for two aromatic analytes, chlorobenzene (k′ ) 0.65) and anisole (k′ ) 0.82). The van Deemter plots obtained for the two analytes were satisfactory (Figures 6 and 7).21 The reduced plate height, h, is the experi(21) Giddings, J. C. Unified Separation Science; Wiley: New York, 1991.

4562 Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

mentally determined plate height, H, divided by the inside diameter of the capillary. Using reduced plate height plots, future comparisons between different systems with varying column diameters will be straightforward. The minimum reduced plate height, hmin, is obtained at the optimum linear flow velocity. A hmin less than or equal to 1 is indicative of a high-performance open tubular GC system.22 The theoretically obtainable hmin for open tubular chromatography under ideal conditions is 0.8.23,24 Thus, the hmin values of 0.90 (chlorobenzene) and 0.88 (anisole) obtained from Figures 6 and 7, respectively, are indicative of an efficient open tubular system. The hmin near the optimum and the relatively shallow slope of h at high linear flow velocity both indicate the MPN column performance was consistent with the thin-film stationary phase being used (∼60 nm). The efficiency of the MPN column separations was compared to commercial AT-1 column separations. To compare the novel nanoparticle stationary phase with a commercial stationary phase it was decided that the two columns should be as dimensionally identical as possible. The AT-1 phase was chosen since it is a nonpolar phase based on 100% poly(dimethylsiloxane). The AT-1 phase thickness was 100 nm compared to 60 nm for the MPN column. Both columns were 2 m long with 530-µm internal diameters. The slight difference in film thickness was not anticipated to hamper the comparison. Using the same experi(22) Grant, D. Capillary Gas Chromatography; John Wiley & Sons: New York, 1996. (23) Knox, J. J. Chromatogr. Sci. 1980, 18, 453-461. (24) Novotny, M. Anal. Chem. 1988, 60, A500.

Figure 5. Eight-component mixture (subset of Table 1) separations with the dodecanethiol MPN column: (A) isothermal at 50 °C and (B) temperature programmed, 40-80 °C at 70 °C/min. The boiling point range was 78-174 °C with the following elution order: ethanol, benzene, 1-butanol, 3-heptanone, chlorobenzene, 3-octanone, anisole, and decane. (C) Isothermal separation at 50 °C on the AT-1 commercial column. Elution order: ethanol, benzene, 1-butanol, chlorobenzene, 3-heptanone, anisole, 3-octanone, and decane. All other parameters were the same as those in Figure 3.

Figure 6. A van Deemter plot for chlorobenzene (k′ ) 0.65) on the dodecanethiol MPN column. Reduced plate height, h, versus linear flow velocity, u, is plotted. The standard deviation error bars were calculated from sets of five runs. Column parameters: 1.7 m, 530 µm i.d. with MPN phase, operated at 50 °C. Sample was introduced via a 200 °C inlet under constant-pressure conditions with a sample size of 0.5 µL with a 1:1 split on the injected vapor.

mental parameters as initially used for the MPN column characterization, the AT-1 column was used to separate an alkane mixture (hexane, octane, decane) and the eight-component mixture previously described (Figure 5). For octane as the representative analyte, the AT-1 column had an efficiency, N, of 6200 (k′ ) 0.33) while the MPN column had an efficiency of 5700 (k′ ) 0.21), indicating the dodecanethiol MPN column at its optimum is operating essentially the same as the commercial AT-1 stationary phase. There were slight retention order differences obtained with the two phases, however, as indicated by the eight-component separations (Figure 5A and C). Note that the relative peak heights for each analyte in the two separations vary slightly due to the injection procedure employed. The AT-1 column separation occurred in the same time frame as the MPN column separation. However, two of the components not resolved for the AT-1 column separation, benzene (k′ ) 0.08) and 1-butanol (k′ ) 0.10), were resolved by the MPN column. In addition, the analyte peaks are distributed differently in the two chromatograms, indicating there is some difference in the chemical retention behavior despite both stationary phases being primarily nonpolar. The retention differ-

Figure 7. A van Deemter plot for anisole (k′ ) 0.82) on the dodecanethiol MPN column. All parameters are the same as those stated in Figure 6.

ences are summarized in Table 1 and compared to the log L16 values that indicate the elution order when hexadecane is applied as the stationary phase. Retention order variation is observed when all three stationary phases are compared. The variation in retention order for the AT-1 and MPN phases is not surprising, since it was not anticipated that they be identical. However, a more indepth analysis of the intermolecular interactions that contribute to the observed retention with the dodecanethiol MPN phase is warranted. Of particular interest in the MPN stationary-phase development is not only how it compares thermodynamically and kinetically with commercially available stationary phases but also the robustness of the phase. Little to no degradation in the MPN-based separations was observed even after a large number of injections, ∼650 sample runs to date. The hexane, octane, and decane separation was used as a “benchmark” for the MPN column performance at least once every 4 weeks. The benchmark runs indicated little to no degradation occurred over time. It is known from thermal gravimetric analysis of the MPNs that, at temperatures greater than ∼175 °C, the monolayer begins to rapidly strip off the gold cluster (Figure 1C). Since we desired to conclude all initial characterization studies on the same MPN column, this temperature was used as the upper limit for separation method Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

4563

evaluation. The actual temperature range studied was 30-150 °C, but compounds with boiling points up to 287 °C have been successfully studied. A more thorough study of the long-term impact of elevated temperature on retention time stability would be essential in order to determine more quantitatively the performance of MPNs as a GC stationary phase over extended periods. CONCLUSIONS This is the first report describing dodecanethiol MPNs as a thin-film stationary phase for open tubular capillary GC. The MPN materials can be successfully coated as a thin film, as demonstrated by SEM imaging, and perform well chromatographically, as indicated by evaluation of separation efficiency, N, and reduced plate height, h. An N of 5700 (k′ ) 0.21) was achieved for octane on a 2-m column, which is similar to the N achieved for octane on a comparable commercial AT-1 stationary-phase column. The results suggest the dodecanethiol MPN stationary phase can be applied in different methods of GC analysis, as well as continuing studies of the thermodynamic and mass-transfer characteristics of the nanoparticles themselves. Development of other MPN materials as GC stationary phases is also a logical step.

4564

Analytical Chemistry, Vol. 75, No. 17, September 1, 2003

ACKNOWLEDGMENT Special thanks to Jim Young at Pacific Northwest National Laboratory for assistance with the SEM images, to Todd Hart for assistance with the TGA analysis, to Chingmon Wang for assistance with the TEM images, and to Tim Huebler for insightful discussion on FT-IR and material purity. G.M.G. acknowledges partial support from the Joint Institute for Nanoscience funded by the Pacific Northwest National Laboratory (operated by Battelle for the U.S. Department of Energy) and the University of Washington. We also acknowledge the Center for Process Analytical Chemistry for partial support. A portion of the research described in this article was performed at the W. R. Wiley Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the U.S. Department of Energy’s Office of Biological and Environmental Research and located at Pacific Northwest National Laboratory. The Pacific Northwest National Laboratory is a multiprogram national laboratory that is operated for the U.S. Department of Energy by Battelle Memorial Institute. Received for review March 19, 2003. Accepted June 13, 2003. AC030112J