Anal Bioanal Chem (2014) 406:1245–1247 DOI 10.1007/s00216-013-7289-z
NOTE
Sputtered alumina as a novel stationary phase for micro machined gas chromatography columns R. Haudebourg & Z. Matouk & E. Zoghlami & I. Azzouz & K. Danaie & P. Sassiat & D. Thiebaut & J. Vial
Received: 14 June 2013 / Revised: 23 July 2013 / Accepted: 1 August 2013 / Published online: 29 August 2013 # Springer-Verlag Berlin Heidelberg 2013
Abstract Silica and graphite sputtering have previously been reported as novel solid stationary phase deposition techniques for micro gas chromatography columns. As a conventional solid stationary phase in gas chromatography, compatible with sputtering yet so far unreported, alumina was evaluated in this study. Alumina sputtered semi-packed micro columns were fabricated (including an activation step) and proved able to separate a mixture of volatile alkanes (C1–C4 with isomers) in less than 1 min. Kinetic and a thermodynamic evaluation led to calculation of 4,500 theoretical plates for ethane in 1.1 m (HETPmin =250 μm) and a Gibbs free energy for propane of 30.2 kJ mol−1, making this stationary phase’s properties very close to those observed with silica-sputtered micro columns. Keywords Gas chromatography . Micro column . Semi-packed . Alumina . Sputtering
Introduction Advantages and potential of miniaturized chromatography have been reported several times. They mainly consist of the portability of the resulting apparatus, reduced production costs as a result of fabrication in silicon substrates, and the enabling
Published in the special issue Analytical Science in France with guest editors Christian Rolando and Philippe Garrigues. R. Haudebourg : Z. Matouk : I. Azzouz : P. Sassiat : D. Thiebaut : J. Vial Laboratoire Sciences Analytiques, Bioanalytiques et Miniaturisation, ESPCI Paristech, CNRS UMR PECSA, 7195, 10 rue Vauquelin, 75005 Paris, France R. Haudebourg (*) : E. Zoghlami : K. Danaie Schlumberger MEMS Technology Center, 10b rue Blaise Pascal, 78990 Elancourt, France e-mail:
[email protected]
of ultra-fast temperature programming for short and cycled separations. In oilfield applications, gas chromatographybased sensors for the real-time and continuous monitoring of C1–C4 (or C1–C6) alkane steams have been engineered. Reduced bulk, cost, and power consumption with increased analysis throughput is now conceivable because of progress in silicon micro machining techniques, and would be of the highest interest. To simultaneously reduce analysis time and column size, the length of the column must also be drastically decreased, making liquid stationary phases such as polydimethylsiloxane (PDMS) barely appropriate, because short lengths are not sufficiently retentive. However, since the first work of Terry et al. in the late 1970s [1] until recently, the vast majority of reported research on micro gas chromatography columns had resorted to PDMS as the stationary phase. This has been mainly because solid packings or coatings (which are still the key to fast separations of very volatile compounds) were hardly compatible with extreme miniaturization. Over the past few years, original solid stationary phase deposition processes compatible with extreme miniaturization for gas chromatographic applications have been reported: carbon nanotubes (CNTs) in-situ growth [2] and sputtering (silica and graphite) [3, 4]. Although carbon-based stationary phases (CNTs and sputtered graphite) were strongly retentive and rather appropriate for permanent gases or high temperature applications, sputtered silica was proved to be an adequate stationary phase for separation of C1–C4, but also C1– C9, isomers and unsaturated. Along with silica and graphite, alumina was compatible with sputtering and has always been used as a conventional stationary phase in gas–solid chromatography [5], but remained unreported. A preliminary chromatographic, kinetic, and thermodynamic study of alumina-sputtered semi-packed[3, 6] micro columns is presented in this note.
1246
R. Haudebourg et al.
Materials and methods Fabrication of micro columns As previously described [3, 4], semi-packed micro columns were fabricated on a double-sided polished silicon wafer. Column micro chips were equipped with a temperature programming system, including two platinum resistive filaments per chip (one 7.5-Ω filament for resistive heating, one 120-Ω filament for temperature sensing) deposited by sputtering on the rear of the wafer. A 200 μm wide, 100 μm deep, and 110 cm long column was etched by deep reactive-ion etching (DRIE) using an anisotropic standard Bosch process, with an inner structure filled with 10-μm square silicon-etched pillars. Alumina was then sputtered in the channels to form a thin layer of stationary phase (patent pending). The sputtering process is based on the acceleration of ionized argon (plasma) toward a chosen target by an electric field; the target, consisting in the material to be deposited on the substrate, is sputtered by the ions, and clusters of the target materials are deposited on the substrate (silicon wafer) as a thin and slightly porous layer with accurately controlled thickness. The silicon wafer was anodically bonded to a Pyrex substrate. In a similar manner, inlet and outlet rear access holes were etched on the rear of the wafer. After fabrication the wafer was diced to obtain ~2 cm×2 cm micro chips. Two ~5-cm-long fused silica 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 micro column silicon chip using a two-component gas-tight temperature-resistive Hysol epoxy resin by Loctite. Because of the strong adsorption of
Fig. 1 Temperature-programmed (10–120 °C at 6 ° s−1) separation of light alkanes mixture 1 (circled numbers: 1, methane; 2, ethane; 3, propane; 4 , butane; i4 , methylpropane). Separation at 6.75 bar (~0.03 mL min−1) on sputtered alumina (3 μm) semi-packed micro column (200 μm×100 μm×1.1 m)
Fig. 2 Van Deemter plot (2 to 6.75 bar) for ethane on sputtered alumina (3 μm) semi-packed micro column (200 μm×100 μm×1.1 m) at 30 °C
water by alumina, an activation treatment (1 h, 200 °C, under helium flow 5 bar) was applied to alumina-sputtered columns. Equipment Kinetic (Van Deemter plot, 2 to 6.75 bar) and thermodynamic (Van’t Hoff plot, 10 to 44 °C) evaluations of the column were performed with a conventional GC apparatus (Varian 3800
Fig. 3 Van’t Hoff plot (10 to 44 °C) for propane on sputtered alumina (3 μm) semi-packed micro column (200 μm×100 μm×1.1 m) at 6.75 bar
Sputtered alumina as a novel stationary phase
run with Galaxy software). Columns were connected to the GC by use of universal press-fit connectors (Restek). Helium, purchased from Air Liquide (grade Alphagaz 1), was used as carrier gas. Temperature programming through direct resistive heating, pulse-width modulation of a constant power source, and thermo-electric cooling, were used. 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. 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 (helium) 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. Gas samples were prepared by filling a sampling bag (Calibond, Calibration Technologies) with a 10 % methane, 2.5 % ethane and propane, and 1 % butane and isobutane in nitrogen mixture by GeoServices (mixture 1). For kinetic and thermodynamic evaluation, a 25 % methane, 25 % ethane, 25 % propane, and 25 % n-butane mixture, Crystal grade by Air Liquide (mixture 2), was used. A 10-μL glass syringe by Hamilton was used to sample the gas mixture and inject it into the GC injector. Each measurement was repeated three times and the results averaged.
Results and discussion A typical chromatogram obtained from temperatureprogrammed separation of mixture 1 on alumina-sputtered semi-packed columns is displayed in Fig. 1. The four alkanes were fully separated in 1 min with baseline resolution (ResC1– C2 =1.85 and ResiC4–nC4 =1.35). This relatively long separation time was mainly because of the semi-packed design and the low starting temperature (10 °C), both necessary to separate ethane from methane on sputtered alumina (open alumina-sputtered columns were fabricated but they could hardly separate propane from methane). A fast temperature ramp of 6 ° s−1 was used to elute butane in less than 1 min but, unfortunately, C4 isomers had to be eluted under isothermal conditions because of the very low and strict temperature limit (120 °C) chosen in this experiment. This results in a marked asymmetry for these peaks. Pressure was set to the highest level allowed by the GC apparatus (6.75 bar) to ensure best efficiency possible. Indeed, the Van Deemter plot for ethane displayed in Fig. 2 clearly indicates that the minimum theoretical plate height remained unreached at the highest pressure (6.75 bar). However, ethane separation efficiency at approximately 20 cm s−1
1247
was as high as 4,500 plates in 1.1 m (e.g. HETPmin =250 μm), which was quasi-identical to the results obtained with silicasputtered semi-packed micro columns of the same dimensions and film thickness (4,100 plates in 1.1 m, HETPmin =270 μm). A strong similarity with sputtered silica was also observed for thermodynamic properties. The Van’t Hoff plot for propane displayed in Fig. 3 enabled calculation of Gibbs free energy for propane on sputtered alumina (30.2 kJ mol−1), which was close to that calculated for sputtered silica under the same conditions (27.7 kJ mol−1). This similar behaviour reflects the similar retention mechanism on silica and alumina.
Conclusion In addition to previously reported sputtered silica and graphite micro columns, sputtered alumina semi-packed micro columns were fabricated and evaluated. They were proved able to separate very volatile alkane mixtures (C1–C4, with isomers), but only because of the choice of a very retentive column structure (semi-packed) and a low ramp initial temperature, otherwise ethane could not be separated from methane. However, maximum number of theoretical plates was measured to be as high as 4,500 plates in 1.1 m for ethane, which was slightly better than efficiencies measured on identical columns sputtered with the same thickness of silica instead of alumina. Thermodynamic similarities were also observed for both stationary phases. Contrary to silica, alumina required a tedious activation step, but was an interesting alternative. Several tests (various mixtures and species, thermal shocks, liquid water percolation, aging, repeatability, etc.) will be performed to seek potential specific advantages of sputtered alumina over sputtered silica.
References 1. Terry SC, Jerman JH, Angell JB (1979) IEEE Trans Electron Devices 26(12):1880–1886 2. Saridara C, Mitra S (2005) Anal Chem 77(21):7094–7097 3. Vial J, Thiebaut D, Marty F, Guibal P, Haudebourg R, Nachef K, Danaie K, Bourlon B (2011) J Chromatogr A 1218(21):3262–3266 4. Haudebourg R, Vial J, Thiebaut D, Danaie K, Breviere J, Sassiat P, Azzouz I, Bourlon B (2013) Anal Chem 85(1):114–120 5. Neuworth MB (1947) J Am Chem Soc 69(7):1653–1657 6. Ali S, Ashraf-Khorassani M, Taylor LT, Agah M (2009) Sensors Actuators B Chem 141(1):309–315
