Article pubs.acs.org/ac
Temperature-Programmed Sputtered Micromachined Gas Chromatography Columns: An Approach to Fast Separations in Oilfield Applications R. Haudebourg,†,§ J. Vial,*,† D. Thiebaut,† K. Danaie,‡,§ J. Breviere,∥ P. Sassiat,† I. Azzouz,† and B. Bourlon§ †
Laboratoire Sciences Analytiques, Bioanalytiques et Miniaturisation, ESPCI Paristech, CNRS UMR PECSA 7195, 10 rue Vauquelin, 75005 Paris, France ‡ ESIEE Paris, 2 boulevard Blaise Pascal, 93162 Noisy le Grand, France § MEMS Technology Center, Schlumberger, 10b rue Blaise Pascal, 78990 Elancourt, France ∥ GeoServices, 127 Avenue du Bois de la Pie, Paris Nord II, BP 67049 95971, Roissy, France ABSTRACT: In a previous study, a new stationary phase deposition technique for micromachined gas chromatography columns was presented. The rerouting of the sputtering technique to this purpose enabled collective and reproducible fabrication of microcolumns in a silicon wafer. Silica-sputtered micromachined columns showed promising separations of light alkanes in isothermal conditions. In order to go beyond the limitations of isothermal separations, the columns were equipped with sputtered platinum filaments to enable highspeed and low-power temperature programming. The separation performances of temperature-programmed silica- or graphite-sputtered microcolumns were investigated: a separation of light alkanes (C1−C5) was completed in 9 s, and heavier alkanes (until C9), cyclic, isomeric, and unsaturated hydrocarbons were also successfully separated. Versatility of these microcolumns was demonstrated with a high-temperature C1−C2 separation and a C1−C5 separation with nitrogen as carrier gas instead of helium. By matching the requirements of a gas chromatography-based monitoring sensor, in terms of low-cost and industry-ready fabrication process, fast temperature programming and analysis, low power consumption, and good versatility (ambient temperature, carrier gas), these columns should be used in various applications related to oilfield gas analyses.
A
fabrication,14−22 but also microinjectors23 or microdetectors,9,24−26 still using clean room silicon micromachining processes. However, in microcolumn fabrication, the critical step of column coating is hardly compatible with extreme miniaturization. Indeed, most of the protocols resorted to polymer-based liquid stationary phases such as poly(dimethylsiloxane) (PDMS), or to molecular sieves-based solid stationary phases, which could not match the criteria of a collective, reproducible, and industry-ready fabrication process required for a low-cost sensor. In a previous publication,27 we presented how the rerouting of the sputtering technique to coat columns with a thin film of silica was able to address this issue (patent pending). With this method, the whole fabrication process of the microcolumns is collective, reproducible, and industry-ready (nine columns per wafer, accurately controlled thickness of stationary phase thin film). Promising isothermal separations of light linear alkanes
considerable part of the modernization efforts recently made in advanced gas chromatography deals with miniaturization, whose usually recognized advantages are portability and cost (fabrication, utilization, and sustaining). Indeed, with miniaturization of the different components of the chromatograph (mainly column, but also injection system and detector), the bulk of the apparatus can be significantly decreased. Thus, in situ analysis can easily be considered, instead of laboratory analysis implying tedious transport of the sample to analyze. But also with miniaturization, the batch fabrication of the microcomponents is made possible, and the power required to heat and cool the system is dramatically decreased. Numerous publications1−11 or product developments (C2 V, SLS, Torion, DefianTech, ...) over the past few years aimed at creating new generations of sensor-like portable gas chromatographs for various applications (general purposed,1−4 lab-on-chip, air monitoring,5,6 safety,7 light hydrocarbons8−11). In the late 1970s, Terry and co-workers designed the first air analyzer fabricated on a silicon wafer.12,13 Since, various approaches have been proposed regarding microcolumn batch © 2012 American Chemical Society
Received: August 2, 2012 Accepted: November 20, 2012 Published: November 20, 2012 114
dx.doi.org/10.1021/ac3022136 | Anal. Chem. 2013, 85, 114−120
Analytical Chemistry
Article
cm ×2 cm microcolumns (Figure 1). Two ∼5 cm long fusedsilica capillaries by Polymicro Technologies (i.d. 100 μm, o.d. 360 μm) were plugged and glued on the inlet and outlet vertical holes on the back of the microcolumn silicon chip using a twocomponent gastight temperature-resistive hysol epoxy resin by Loctite. Analytical GC Device. A conventional GC was used to evaluate the microcolumns. (Varian 3800). The chromatograph was equipped with a 1079 split−splitless injector set at 200 °C and a flame ionization detection (FID) system set at 300 °C. Helium or nitrogen were used as carrier gas (unless there is additional indication, helium was used as carrier gas). Helium was purchased from Air Liquide (grade Alphagaz 1). Nitrogen was purchased from Praxair (Ultraplus 6.0). Hydrogen was obtained from an F-DBS NMH2 250 hydrogen generator. A compressor with F-DBS GC 1500 air generator and Donaldson filters were used for the air gas source. Makeup gas flow rate was set at 30 mL·min−1, hydrogen flow rate at 30 mL·min−1, and air flow rate at 300 mL·min−1. The column was connected to the injector and the detector of the GC device using universal press-fit connectors (Restek). Gas Samples. Gas samples were prepared by filling a sampling bag (Calibond, Calibration Technologies Inc.) with a 25% methane, 25% ethane, 25% propane, and 25% butane mixture by Air Liquide (Crystal grade), or with a 1/3 methane, 1/3 ethane, 1/3 propane mixture by Praxair, or with GeoServices calibration mixtures, which consisted in artificially mixed hydrocarbons and makeup gas, species and proportions in these mixtures being chosen to mimic genuine oilfield samples. Mixture 1 was composed of methane (100 000 ppm), ethane (25 000 ppm), propane (25 000 ppm), methyl-propane and butane (10 000 ppm each), methyl-butane and pentane (2500 ppm each), and 2-methyl-pentane, 1-methyl-pentane, and hexane (300 ppm each). Mixture 2 was composed of ethene (5000 ppm), hexane, heptane and octane (250 ppm each), benzene and toluene (100 000 ppm each), and helium (100 000 ppm). For both calibration mixtures, nitrogen was the makeup gas. For species liquid at ambient temperature (pentane, hexane, heptane, octane, nonane, cyclopentane, cyclohexane, except in GeoServices calibration mixtures), gaseous samples consisted in vapors extracted from the headspace of the flasks (GPR Rectapur by VWR). Fillings were carried out so that the volume proportion of each compound was approximately the same (except in GeoServices calibration mixtures). A 10 μL glass syringe (Hamilton) was used to sample the gas mixture and inject it into the GC injector. Data Acquisition. A commercial acquisition board was used to interface the Varian GC with a computer. Galaxie 1.9 was used as acquisition software. For experiments involving thermal management, a National Instruments Data Aquisition PAD USB 6008 OEM was used and ran with a homemade Labview program. Column Temperature Calibration. The output analogical voltage recorded by the NI DAQ PAD was compared to the actual temperature of the chip measured with a Omega Datalogger thermometer, for temperatures between 10 and 190 °C, and the resulting calibration curve was found linear (R2 = 0.99). Thermal Block Building. The temperature sensing filament was connected to the NI DAQ PAD board, through a voltage divider bridge and a basic amplification circuit (external power supply +12 V/−12 V Tracopower). The
(C1−C4) in 50 s on silica-sputtered microcolumns were obtained. The efficiency of this type of microcolumn was proven to be as high as 2500 theoretical plates/m. In typical hydrocarbons monitoring applications such as mud gas real-time analysis, a C1−C4 separation time of 50 s is too long. Indeed, in routine analysis, the average analysis time of light alkanes C1−C5 is 40 s (or C1−C8 in 1 min 30 s).10 Moreover, the ability of sputtered columns to separate more various oilfield-related compounds such as isomeric alkanes, cyclic alkanes, and unsaturated hydrocarbons still remains to be evaluated. To shorten analysis time, for both punctual and cycled analyses, the microcolumn’s chip should be temperatureprogrammable (fast heating and cooling). Thanks to the small volume (0.4 cm3) and to the low thermal capacity of the chip, and to the high thermal diffusivity of silicon, temperature programming of the chip is expected to be fast, efficient, and economical in terms of power consumption. In this paper, it is shown how very fast separations of various oilfield-related compounds can be obtained on temperature-programmed sputtered micromachined gas chromatography columns, through metallic filaments deposition (also by sputtering) on the back of the chip, to allow resistive heating and temperature sensing.7,17,20 The versatility of these columns, toward harsh ambient temperature conditions and carrier gas changing, is also investigated.
■
MATERIALS AND METHODS Columns Fabrication. As described previously,27 2.2 m × 100 μm × 100 μm columns have been fabricated on a doublesided polished silicon wafer (nine columns per wafer). To enable temperature programming of the columns microchips, two platinum resistive filaments per chip (one 7.5 Ω filament for resistive heating, one 120 Ω filament for temperature sensing) were deposited by sputtering on the back side of the wafer (Figure 1). Channels were etched to a chosen depth by deep reactive-ion etching (DRIE) using an anisotropic standard Bosch process. The wafers were sputtered with a thin layer of silica or graphite using the appropriate target. The silicon wafer was anodically bonded to a Pyrex substrate. In a similar manner, inlet and outlet back access holes were etched on the back side of the wafer. After fabrication the wafer was diced to obtain ∼2
Figure 1. Picture of the back of the column’s microchip (2 cm × 2 cm × 1 mm): (1) heating filament; (2) temperature sensing filament. Filaments are platinum films, deposited by sputtering, and enable temperature programming of the chip. 115
dx.doi.org/10.1021/ac3022136 | Anal. Chem. 2013, 85, 114−120
Analytical Chemistry
Article
heating filament was connected to an external power supply (Multmetrix XA3052) set at 30 V. To enable fast cooling, the column’s chip was placed on a Peltier device (Muliticomp 30 W, and 3.3 V external power supply, Stontronics). The Peltier device was placed on a dissipative radiator equipped with minifans (Sunon, external power supply 5 V). Thermal contacts between the chip and the Peltier device, and between the Peltier device and the radiator, were realized using a thermal paste (Dow Corning 340). Solid-state relays (electronically controlled switches) were placed between the power supplies and the components, to enable real-time automate feedback control of the temperature of the chip by pulse width modulation28 (see also Figure 2): at each measure of the
Figure 3. Chip heating: (1) 12 °C/s; (2) 16 °C/s; (3) 20 °C/s. Initial temperature, 13 °C; final temperature, 220 °C; ambient temperature, 30 °C. Input voltage: 30 V.
studies should be soon led to confirm both heating and cooling to be homogeneous within the chip. Oilfield Gaseous Compounds Temperature-Programmed Separations on Silica-Sputtered Microcolumns. As expected, temperature programming of the chip shortened considerably the analysis time (Figure 4). Whereas
Figure 2. Controlled linear heating of the column’s microchip: (1) chip temperature; (2) theoretical ramp of 10 °C/s; (3) instantaneous electrical power consumption (width-modulated pulses); (4) average electrical power consumption.
temperature of the chip, the software compared this actual temperature to the instantaneous set point temperature in the ramp; if the actual temperature was lower, the switch was closed, and a pulse of power was delivered to the heating filament, and if the temperature was higher, the switch was open and no power was delivered to the heating filament; working frequency of the system was 27 Hz. Instantaneous electrical power consumption was calculated using Joule’s law P = V2/R, where P was the instantaneous electrical power delivered to the chip, V was the voltage applied to the heating filament, and R was the instantaneous resistance of the heating filament.
■
RESULTS AND DISCUSSION Temperature Programming. As shown in Figure 2, linear heating of the chip was enabled by pulse width modulation. Heating rates, linear in the range of 13−220 °C, and up to 20 °C/s, were achieved (Figure 3). Above this temperature, the chip could be damaged. Electrical power consumption to bring the chip’s temperature from 13 to 220 °C was as low as 13 W (at 5 °C/s), 32 W (at 10 °C/s), and 49 W (20 °C/s). As well as fast heating, fast cooling of the chip was also a crucial requirement for cyclic applications. The thermal block (Peltier device, radiator, and fans) allowed very fast cooling of the chip from 200 °C to below ambient (30 °C) in less than 60 s, with an average power consumption of 10 W. Thus, heating rates were more than 10 times faster than usual rates in conventional GC (GC Varian 3800 maximum heating rate was 1.7 °C/s), with a much lower power consumption than conventional GC ovens, and shorter cooling durations. Although thermal diffusivity of silicon is known to be high, complementary
Figure 4. Isothermal (top) and temperature-programmed (bottom) light alkanes separations on a silica-sputtered microcolumn: (1) methane, (2) ethane, (3) propane, (4) butane, (5) pentane; isothermal separation, T = 30 °C, d = 0.27 mL/min; temperature-programmed separation, Tambient = 30 °C, Tinitial = 15 °C, ramp = 15 °C/s, Tlimit = 200 °C, d = 1.2 mL/min; Vinj = 5 nL; column 2.2 m × 100 μm × 100 μm, sputtered silica 3 μm.
isothermal separation of light alkanes (from methane to pentane) took 6 min, temperature-programmed separation of the same light alkanes took only 9 s, using a heating rate of 15 °C/s. Starting the separation at a temperature slightly below ambient (Tinitial = 13 °C, instead of Tambient = 30 °C), thanks to a preliminary cooling, allowed the use a higher flow rate (1.2 mL/min instead of 0.27 mL/min) without losing the resolution between methane and ethane and, thus, to obtain an even shorter separation. Average routine methane to pentane 116
dx.doi.org/10.1021/ac3022136 | Anal. Chem. 2013, 85, 114−120
Analytical Chemistry
Article
did not allow a full resolution between peaks. For methane and ethane, significant differences in their respective quantities (4:1 ratio) damaged even more the quality of the separation. Regarding hexane, peak shape was also damaged due to the use of a limited temperature. The column dimension, the stationary phase film’s thickness, and/or the method should be optimized to improve the quality of the separation. Yet, it should be reminded here that in this targeted application, separation time was as crucial as resolution. The compromise between speed and accuracy should be the subject of a further publication. Light cycloalkanes (cyclopentane and cyclohexane) were successfully separated in 10 s (Figure 7); resolution was complete, but additional tests should be led with a mixture composed of both cyclic and noncyclic alkanes.
separation time was thus divided by a factor more than 4. However, fast cooling of the chip is still to be improved, for the whole thermal cycle (heating while separating and cooling back to start a new cycle) should remain shorter than at most 40 s. Figures 5−8 show successful separations of various oilfieldrelated compounds on temperature-programmed silica-sputtered microcolumns.
Figure 5. Temperature-programmed light linear alkanes separation on a silica-sputtered microcolumn: (1) methane, (2) ethane, (3) propane, (4) butane, (5) pentane, (6) hexane, (7) heptane, (8) octane, (9) nonane; Tambient = 30 °C, Tinitial = 15 °C, ramp = 15 °C/s, Tlimit = 200 °C, d = 1.2 mL/min, Vinj = 5 nL; column 2.2 m × 100 μm × 100 μm, sputtered silica 3 μm. Figure 7. Temperature-programmed light cyclic alkanes separation on a silica-sputtered microcolumn: (1) cyclopentane, (2) cyclohexane; Tambient = 30 °C, Tinitial = 50 °C, ramp = 10 °C/s, Tlimit = 200 °C, d = 1.2 mL/min, Vinj = 20 nL; column 2.2 m × 100 μm × 100 μm, sputtered silica 3 μm.
Linear alkanes (from methane to nonane) were separated in 15 s (Figure 5); resolutions between peaks were very satisfying (C7−C9) or complete (C1−C7). Because of the limit temperature of 200 °C, compounds that were not eluted after 12.3 s (octane and nonane) finished eluting under isothermal conditions (200 °C), which could explain their poorer resolution. The chip’s thermal robustness should be improved so that temperatures above 200 °C could be used without risks, allowing the elution of all compounds under gradient conditions. Calibration mixture 1 (isomeric alkanes from methane to hexane, by GeoServices) was as well analyzed in 15 s (Figure 6); although that each isomer could be clearly identified (even those in low quantities), the selectivity of the stationary phase
Eventually, calibration mixture 2 [unsaturated hydrocarbons (ethene, benzene, toluene) and linear alkanes (hexane, heptane, octane) by GeoServices] was analyzed in 23 s (Figure 8).
Figure 8. Temperature-programmed light unsaturated hydrocarbons and alkanes mixture separation on a silica-sputtered microcolumn: (1) ethene, (2) hexane, (3) heptane, (4) octane, (5) benzene, (6) toluene; Tambient = 30 °C, Tinitial = 40 °C, ramp = 10 °C/s, Tlimit = 190 °C, d = 1.1 mL/min, Vinj = 1 μL; column 2.2 m × 100 μm × 100 μm, sputtered silica 3 μm.
Observed ethene retention time was 3.2 s, which fell exactly within the retention times observed for ethane (2.7 s) and propane (3.7 s) on the chromatogram shown in Figure 8 (obtained in the same conditions). Both benzene and toluene could be successfully separated from hexane, heptane, and octane. Retention of xylene should be also tested as a complementary experiment.
Figure 6. Temperature-programmed light isomeric alkanes (mixture 1) separation on a silica-sputtered microcolumn: (1) methane, (2) ethane, (3) propane, (4) methyl-propane, (5) butane, (6) methylbutane, (7) pentane, (8) 2-methyl-pentane, (9) 1-methyl-pentane, (10) hexane; Tambient = 30 °C, Tinitial = 15 °C, ramp = 15 °C/s, Tlimit = 180 °C, d = 1.1 mL/min, Vinj = 50 nL; column 2.2 m × 100 μm × 100 μm, sputtered silica 3 μm. 117
dx.doi.org/10.1021/ac3022136 | Anal. Chem. 2013, 85, 114−120
Analytical Chemistry
Article
Graphite-Sputtered Microcolumns. Temperature-programmed separations on silica-sputtered microcolumns showed a fair methane−ethane separation but only thanks to the moderate ambient temperature (30 °C) and to a preliminary cooling (13 °C). As oilfield monitoring is a potential application for such columns, the possibility of working at high ambient temperatures (between 50 and 150 °C), where an initial temperature of 13 °C is hardly reachable, had to be considered. In this case, silica as a stationary phase was not retentive enough to provide a fully resolved methane−ethane separation (see Figures 10 and 12). As graphite is known to be a highly retentive stationary phase for alkanes,20,29 graphite-sputtered microcolumns were fabricated and tested. Figure 9 shows a scanning electron microscope (SEM) picture of the inner wall of a graphite-sputtered microcolumn.
Figure 10. Comparison of the retention of light alkanes on graphitesputtered (above) and silica-sputtered (below) microcolumns: (1) methane, (2) ethane, (3) propane; Tambient = 30 °C, d = 0.55 mL/min, Vinj = 5 nL; columns 2.2 m × 100 μm × 100 μm, sputtered graphite 0.78 μm (above), sputtered silica 3 μm (below). Retention factors k for ethane and propane: 0.67 and 12.4 (graphite-sputtered column), 0.28 and 1.80 (silica-sputtered microcolumn).
Figure 9. SEM picture of the graphite-sputtered layer deposited on the inner wall of a silicon micromachined gas chromatography column. The structure in sheets is due to the etching process; the structure in clusters is due to the deposition process.
In addition of a wall structure in sheets due to the silicon etching process, a structure of carbon graphite clusters due to the deposition process is clearly visible, resulting in a high surface-to-volume ratio layer. As expected, graphite-sputtered microcolumns showed very strong retention properties toward light hydrocarbons, compared to silica-sputtered microcolumns (Figures 10 and 11). Figure 10 shows a comparison of typical chromatograms obtained with these two stationary phases. Strong adsorption mechanism for graphite resulted in much higher retention factors k (0.67 and 12.4 for ethane and propane on graphite, whereas it was 0.28 and 1.80 on silica) and in dramatic tailing peak asymmetries (3.25 and 2.88 for ethane and propane on graphite, whereas it was 2.47 and 2.37 on silica), even with a 4 times thinner layer (0.78 μm for graphite, 3 μm for silica). Further results on retention properties (Van’t Hoff plots) are displayed on Figure 11. The relationship between log k and (1/ T) was confirmed to be linear in the temperature range explored, and the standard enthalpy for graphite was calculated to be 1.4 times higher than the standard enthalpy for silica. However, efficiency was dramatically decreased when sputtered graphite was used as a stationary phase, compared to silica (Figure 12). Indeed, minimal plate height obtained with graphite-sputtered microcolumns was only 16 mm, whereas it was as low as 0.40 mm with silica-sputtered
Figure 11. Van’t Hoff plot for propane for graphite-sputtered (2.5 μm) and silica-sputtered (3 μm) microcolumns (2.2 m × 100 μm × 100 μm). Graphite: R2 = 0.997 and ΔrH0 = 43 kJ. Silica: R2 = 0.993 and ΔrH0 = 31 kJ. Temperatures were set using the temperatureprogramming system presented here.
microcolumns, for identical columns dimensions and stationary phase film thicknesses.27 The strong retention properties of graphite as stationary phase were then utilized to enable methane−ethane separations at high temperature. As shown in Figure 13, where silica-sputtered microcolumns failed to separate methane and ethane with a sufficient resolution (0.81) at high temperature (100 °C), graphitesputtered microcolumns were able to separate methane and ethane with an excellent resolution (1.97). Changing the stationary phase to be sputtered inside the columns did not imply fundamental modifications in the fabrication process: indeed, it only consisted in choosing the 118
dx.doi.org/10.1021/ac3022136 | Anal. Chem. 2013, 85, 114−120
Analytical Chemistry
Article
Figure 12. Van Deemter plot for ethane for a graphite-sputtered microcolumn; Tambient = 30 °C, P from 5 to 85 psi, Vinj = 5 nL; column 2.2 m × 100 μm × 75 μm, sputtered graphite 0.78 μm; HETPmin = 16 mm (HETPmin for a 0.75 μm silica-sputtered column with the same dimensions: 0.4 mm).
Figure 14. Temperature-programmed light linear alkanes separation on silica-sputtered microcolumn with nitrogen instead of helium as carrier gas: (1) methane, (2) ethane, (3) propane, (4) butane, (5) pentane; Tambient = 30 °C, Tinitial = 30 °C, ramp = 16 °C/s, Tlimit = 195 °C, d = 0.64 mL/min, Vinj = 5 nL; column 2.2 m × 100 μm × 100 μm, sputtered silica 3 μm.
■
CONCLUSION Micro gas chromatography columns were fabricated through a fully collective, reproducible, and industry-ready fabrication process, including the steps of sputtering deposition of the stationary phase and of the metallic filaments for resistive heating and temperature sensing. These columns were coupled with a temperature-programming system, allowing fast linear heating of the column’s chip up to 20 °C/s and fast cooling (from 220 to 30 °C in less than 1 min). Temperatureprogrammed silica-sputtered micromachined columns enabled fast separation of light linear alkanes (C1−C5) in 9 s. Separation of heavier linear alkanes (C1−C9) was carried out in 15 s, and isomeric alkanes (C1−C6), cyclic alkanes (cyclopentane−cyclohexane), and unsaturated compounds (ethene, benzene, toluene) were also successfully separated. Average power consumption was lower than 50 W. Graphitesputtered microcolumns were used to enable full separation of methane and ethane at high temperature (100 °C). Light alkanes (C1−C5) were also successfully separated in 11 s with nitrogen as carrier gas instead of helium. Thus, such columns could easily match the requirements of an oilfield monitoring sensor based on a gas chromatography method (safety, mud logging), in terms of fabrication process, separated compounds, fast thermal management and analysis, low power consumption, and good versatility (ambient temperature, carrier gas). In mud gas analysis application, methane to pentane separation time was divided by 4, with a lower power consumption. However, cooling of the microchip still remains the limiting step in the routine analysis cycle. Enabling ultrafast cooling of the chip should be the next perspective toward the use of these columns in field applications. Moreover, the design and fabrication of the up- and downstream components of a micro gas chromatograph (i.e., a microinjector,23 and a micro thermal conductivity detector30), also through conventional silicon micromachining techniques, have already been reported. Although some critical steps remain to be done (air−methane separation, packaging, quantitative analyses), the framework and the proof of concept of the essential components of a GC-based oilfield monitoring sensor are henceforth present.
Figure 13. High-temperature (100 °C) methane−ethane separation on silica-sputtered (left) and on graphite-sputtered (right) microcolumn: (1) methane, (2) ethane; Tambient = 100 °C, d = 0.52 mL/min (left), d = 0.47 mL/min (right), Vinj = 5 nL; column 2.2 m × 100 μm × 100 μm, sputtered silica 3 μm (left), sputtered graphite 2.5 μm (right).
appropriate target in the sputtering machine, knowing that our sputtering machine hosted four different targets. These results opened numerous prospects: many target materials should thus be evaluated as stationary phases, in order to adapt to the various applications and requirements that will be met in the future. Separations with Nitrogen as Carrier Gas. Helium as carrier gas is generally not compatible with the low-bulk, lowcost, and low-pressure gas chromatography required in an oilfield environment. In most oilfield applications, in situ purified air as carrier gas is preferred to helium.10,11 To address this issue, preliminary separation tests on silica-sputtered temperature-programmed micromachined columns with nitrogen instead of helium as carrier gas were carried out. Figure 14 shows a successful temperature-programmed separation of light alkanes (from methane to pentane) with nitrogen as carrier gas in 11 s. These results reasonably suggested that all separations presented above with helium as carrier gas could be run with air as carrier gas without major loss of performances. Thus, additional tests should be made with air as carrier gas to comprehensively demonstrate the appropriateness of these columns to an oilfield environment. Such tests should include a kinetic comparison (Van Deemter plots) between both carrier gases for these columns. 119
dx.doi.org/10.1021/ac3022136 | Anal. Chem. 2013, 85, 114−120
Analytical Chemistry
■
Article
(25) Cruz, D.; Chang, J. P.; Showalter, S. K.; Gelbard, F.; Manginell, R. P.; Blain, M. G. Sens. Actuators, B 2007, 121 (2), 414−422. (26) Chen, K.; Wu, Y. E. Sens. Actuators, A 2000, 79 (3), 211−218. (27) Vial, J.; Thiébaut, D.; Marty, F.; Guibal, P.; Haudebourg, R.; Nachef, K.; Danaie, K.; Bourlon, B. J. Chromatogr., A 2011, 1218 (21), 3262−3266. (28) Ball, S. Analog Interfacing to Embedded Microprocessor Systems, 2nd ed.; Newnes: Burlington, MA, 2004; pp 287−297. (29) Bruner, F.; Lattanzi, L.; Mangani, F.; Attaran Rezaii, M. Chromatographia 1994, 38 (1−2), 98−108. (30) Kaanta, B. C.; Steinecker, W. H.; Zhdaneev, O.; Lambertus, G. R.; Chen, H.; Zhang, X.; Bourlon, B.; Donzier, E. P. (Schlumberger). U.S. Patent Application PCT/US2010/052147, 2011.
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
REFERENCES
(1) Bruns, M. W. In Industrial Electronics, Control, Instrumentation, and Automation, Proceedings of the 1992 International−Conference on Power Electronics and Motion Control; San Diego, CA, November 9−13, 1992, ISBN: 0-7803-0582-5. (2) Agah, M.; Lambertus, G. R.; Sacks, R.; Wise, K. J. Microelectromech. Syst. 2006, 15 (5), 1371−1378. (3) Lu, C. J.; Steinecker, W. H.; Tian, W. C.; Oborny, M. C.; Nichols, J. M.; Agah, M.; Potkay, J. A.; Chan, H. K. L.; Driscoll, J.; Sacks, R. D.; Wise, K. D.; Pang, S. W.; Zellers, E. T. Lab Chip 2005, 5 (10), 1123− 1131. (4) Nishino, M.; Takemori, Y.; Matsuoka, S.; Kanai, M.; Nishimoto, T.; Ueda, M.; Komori, K. IEEJ Trans. Electr. Electron. Eng. 2009, 4 (3), 358−364. (5) Wise, K. D. Sens. Actuators, A 2007, 136 (1), 39−50. (6) Zampolli, S.; Elmi, I.; Mancarella, F.; Betti, P.; Dalcanale, E.; Cardinali, G. C.; Severi, M. Sens. Actuators, B 2009, 141 (1), 322−328. (7) Potkay, J. A.; Lambertus, G. R.; Sacks, R. D.; Wise, K. D. J. Microelectromech. Syst. 2007, 16 (5), 1071−1079. (8) Yu, C. M.; Lucas, M.; Koo, J. C.; Stratton, P.; DeLima, T.; Behymer, E. J. Microelectromech. Syst. 1998, 66, 481. (9) Narayanan, S.; Alfeeli, B.; Agah, M. In Procedia Engineering, 24th Eurosensors Conference, Linz, Austria, September 5−8, 2010; Elsevier, Vol. 5, 2010, pp 29−32. (10) Breviere, J.; Herzaft, B.; Mueller, N. In Petrophysics, SPWLA 43rd Annual Logging Symposium, June 2−5, 2002, Oiso, Japan; Society of Petrophysicists and Well Log Analysts, 2002, p 424. (11) Blanc, P.; Brevière, J.; Laran, F.; Chauvin, H.; Boehm, C.; Fréchin, N.; Capot, M.; Benayoun, A. In Mile High Meeting of the Minds, Proceedings−SPE Annual Technical Conference and Exhibition 2003, Denver, CO, Oct 5−8, 2003; Society of Petroleum Engineers (SPE), USA, pp 2861−2879. (12) Terry, S. C.; Jerman, J. H.; Angell, J. B. IEEE Trans. Electron Devices 1979, 26 (12), 1880−1886. (13) Jerman, J. H.; Terry, S. C. Environ. Int. 1981, 5 (2), 77−83. (14) Hannoe, S.; Sugimoto, I.; Yanagisawa, K.; Kuwano, H. Proceedings of the 1997 International Conference on Solid-State Sensors and Actuators, Part 1 (of 2), Chicago, IL, USA, June 16−19, 1997, Code 47124; IEEE, Piscataway, NJ, USA, 1997, pp 515−518. (15) Wong, C. C.; Adkins, D. R.; Frye-Mason, G. C.; Hudson, M. L.; Kottenstette, R.; Matzke, C. M.; Shadid, J. N.; Salinger, A. G. Proc. SPIE 1999, 3487, 120. (16) Spangler, G. E. J. Microcolumn Sep. 2001, 13 (7), 285−292. (17) Agah, M.; Potkay, J. A.; Lambertus, G.; Sacks, R.; Wise, K. D. J. Microelectromech. Syst. 2005, 14 (5), 1039−1050. (18) Ali, S.; Ashraf-Khorassani, M.; Taylor, L. T.; Agah, M. Sens. Actuators, B 2009, 141 (1), 309−315. (19) Radadia, A. D.; Morgan, R. D.; Masel, R. I.; Shannon, M. A. Anal. Chem. 2009, 81 (9), 3471−3477. (20) Reid, V. R.; Stadermann, M.; Bakajin, O.; Synovec, R. E. Talanta 2009, 77 (4), 1420−1425. (21) Nishiyama, S.; Nakai, T.; Shuzo, M.; Delaunay, J. J.; Yamada, I. In Effect of micropillar density on separation efficiency of semi-packed micro gas chromatographic columns, IEEE Sensors Conference, 2009; Christchurch, October 25−28, 2009; IEEE, Piscataway, NJ, USA, 2009, Article no. 5398351, pp 1935−1938. (22) Vangelooven, J.; Desmet, G. J. Chromatogr., A 2010, 1217 (52), 8121−8126. (23) Nachef, K.; Bourouina, T.; Marty, F.; Danaie, K.; Bourlon, B.; Donzier, E. J. Microelectromech. Syst. 2010, 19 (4), 973−981. (24) Sorge, S.; Pechstein, T. Sens. Actuators, A 1997, 63 (3), 191− 195. 120
dx.doi.org/10.1021/ac3022136 | Anal. Chem. 2013, 85, 114−120
