TY - JOUR
AU - Achyuthan, Komandoor E
AB - Abstract Despite promising advances with metal-organic frameworks (MOFs) as stationary phases for chromatography, the application of MOFs for one- and two-dimensional micro-gas chromatography (μGC and μGC × μGC) applications has yet to be shown. We demonstrate for the first time, μGC columns coated with two different MOFs, HKUST-1 and ZIF-8, for the rapid separation of high volatility light alkane hydrocarbons (natural gas) and determined the partition coefficients for toxic industrial chemicals, using μGC and μGC × μGC systems. Complete separation of natural gas components, methane through pentane, was completed within 1 min, with sufficient resolution to discriminate n-butane from i-butane. Layer-by-layer controlled deposition cycles of the MOFs were accomplished to establish the optimal film thickness, which was validated using GC (sorption thermodynamics), quartz-crystal microbalance gravimetric analysis and scanning electron microscopy. Complete surface coverage was not observed until after ~17 deposition cycles. Propane retention factors with HKUST-1-coated μGC and a state-of-the-art polar, porous-layer open-tubular (PLOT) stationary phase were approximately the same at ~7.5. However, with polar methanol, retention factors with these two stationary phases were 748 and 59, respectively, yielding methanol-to-propane selectivity factors of ~100 and ~8, respectively, a 13-fold increase in polarity with HKUST-1. These studies advance the applications of MOFs as μGC stationary phase. Introduction The largest petroleum fraction is high volatility light hydrocarbons (C1–C8) (1). Identification of these analytes in a gaseous mixture is widely employed in the petrochemical and energy industries during the production of natural gas (2). There is a great demand for gas chromatography (GC) systems that can efficiently separate and quantify proportions of light hydrocarbons. However, portable, high-speed analysis of natural gas analytes remains a largely unsolved challenge. Multidimensional microfabricated GC (μGC) (2–8) might be a solution to this challenge. μGCs have been mainly applied to the detection of medium- to semi-volatile compounds due to their typical use of polymer stationary phases. When dealing with natural gas analytes, polymer stationary phases do not have the retention power necessary for high-speed separations on μGC columns, and exploration of alternative stationary phases becomes necessary. Although high-speed, well-resolved natural gas separation in monolithic capillary columns has been demonstrated (9), portable, high-speed analysis of natural gas and other light hydrocarbon analytes remains challenging. Additionally, there is great interest in, and difficulty of, portable detection of toxic industrial chemicals (TICs). TICs represent a large class of ubiquitous chemicals, which are often highly volatile and polar, and which pose a threat to the environment and human life though accidental release or unlawful/nefarious use in warfare or terrorist attacks (10). Metal-organic frameworks (MOFs) are hybrid inorganic-organic porous materials possessing desirable properties such as large surface area, low density, high adsorption affinity, accessible cages and tunable pore dimensions along with chemical and thermal stabilities (11–14). These properties enable MOFs to accept analyte molecules by a variety of host-guest interactions such as molecular sieving, van der Waals forces, π-complexation and coordinative interactions with metal centers, leading to selective retention of analytes, and enabling chromatographic resolution of small molecules (11–14). Despite these advantages, the applications of MOFs in GC are still quite small (15, 16), and, to the best of our knowledge, the use of MOFs for one- and two-dimensional μGC and μGC × μGC has yet to be demonstrated. Here, we demonstrate μGC and μGC × μGC systems with Hong Kong University of Science and Technology-1 (HKUST-1; MOF-199, Cu-BTC, Basolite™ C300) and Zeolitic Imidazolate Framework-8 (ZIF-8) (11, 13, 14) MOF stationary phases, for the rapid separation of light alkanes, isomers and TICs. The selectivity of HKUST-1 coated μGC column was superior to porous layer open tubular (PLOT) column. We describe a method to determine MOF stationary phase film thickness for optimizing chromatographic performance. We also present thermodynamic data to demonstrate the separation characteristics of MOFs as μGC stationary phases. These studies expand the role of MOF stationary phases in GC applications. Experimental Analytes Natural gas samples were Scotty® 14 natural gas standard (Scotty Specialty Gases, now Air Liquide, Houston, TX) in the following composition (mol %): ethane, 12.5%; n-butane, 3.02%; i-butane, 3.05%; i-pentane, 0.503%; propane, 6.96%; balance methane, with other trace gases CO2 and N2. Where noted, additional propane, butane mixtures and liquid-headspace n-pentane and/or n-hexane were added to the GC vial with Scotty® gas standard, to increase the relative analyte proportions. All other analytes including ambient-liquid hydrocarbons and TICs were of analytical reagent quality from Sigma-Aldrich (St. Louis, MO). Chlorine gas was generated by reacting concentrated HCl with trichloroisocyanuric acid (CAS: 87-90-1, E-Z CLOR® pool water sanitizer from E-Z CLOR Chemicals). WARNING: chlorine gas is a strong oxidizer and is highly toxic. Great care should be observed to avoid inhalation and exposure (i.e., use of fume hood and personal protective equipment and carried out only by trained and knowledgeable personnel). Other chemicals and reagents Other chemicals and reagents from Sigma-Aldrich were methanol, catalog number 34860, for HPLC, >99.9%. Zinc nitrate hexahydrate, CAS: 101196-18-6, >99.0% pure. 2-Methylimidazole: CAS: 693-98-1, 99% pure, copper acetate (Cu(OAc)2), trimesic acid (benzene-1,3,5-tricarboxylic acid, H3BTC) and ethanol. Commercial capillary GC columns The commercial Rt-U-Bond® (Restek Corporation, Bellefonte, PA) fused silica PLOT capillary column coated with particles of divinylbenzene ethylene glycol/dimethylacrylate had an inner diameter of 320 μm and was cut to a length of 120 cm, with a film thickness, df, of 10 μm (β = 8). The commercial Rtx-Wax® polyethylene glycol (PEG) capillary column (Restek) also had an inner diameter of 320 μm, but with 0.25 μm df, and was cut to a length of 120 cm. μGC columns Three types of μGC columns were fabricated at Sandia’s Microsystems Engineering, Science and Applications Complex with dimensions as shown in Table I. These high-aspect-ratio μGC columns were fabricated by a Bosch-process, deep reactive-ion etching (DRIE) of <100> Si as described previously (17). A ~1 μm thermal oxide was grown on the channel walls using a standard semiconductor process, leaving a nominally hydroxylated SiO2 surface. Briefly, natural growth of silicon dioxide (SiO2) on Si surface at room temperature is very slow and stops at about 15 angstroms after 2 or 3 days. During microfabrication, SiO2 was grown in a high temperature furnace with an O2 source (“thermal oxidation”). Exposure time and temperature can be used to control the growth of the thermal oxide. In this way, SiO2 incorporates the Si consumed from the substrate and the supplied O2. Capillary ports were left sealed with a thin membrane of Si to prevent Si debris particles from entering and packing the column during the dicing stage (18). In the absence of hermetic seal, water-borne saw debris from the dicing process created a packed bed of Si microparticles that could not be removed. The membrane seal was easily punctured with a metal pin and breakage debris was minimal and was vacuumed out, leaving a pristine oxide channel. Table I μGC Column Details Channel dimensions . Column name/label . 70–120 . 30–90 . 30–30 . Depth (μm) 685 685 685 Width (μm) 70 30 30 Length (cm) 120 90 30 Channel dimensions . Column name/label . 70–120 . 30–90 . 30–30 . Depth (μm) 685 685 685 Width (μm) 70 30 30 Length (cm) 120 90 30 Open in new tab Table I μGC Column Details Channel dimensions . Column name/label . 70–120 . 30–90 . 30–30 . Depth (μm) 685 685 685 Width (μm) 70 30 30 Length (cm) 120 90 30 Channel dimensions . Column name/label . 70–120 . 30–90 . 30–30 . Depth (μm) 685 685 685 Width (μm) 70 30 30 Length (cm) 120 90 30 Open in new tab A 5 cm long (630 μm o.d. × 200 μm i.d.) polyimide-coated, fused silica capillary (Molex Inc. Polymicro-Technologies™, Phoenix, AZ, special order) was epoxied (Henkel, Inc., Dusseldorf, Germany, Loctite® 1CTM Hysol® two-part epoxy) at 80°C into both the inlet and outlet ports on the μGC for connection to either a commercial gas chromatograph or the surface-anchored/mounted MOF (SURMOF) deposition apparatus (described later). Photographic images and scanning electron microscope (SEM) micrographs of the μGC chip are shown in Figure 1. Figure 1 Open in new tabDownload slide μGC column fabrication. (Panels A and B) Photographs of top (back) and bottom (front) of μGC column with channel dimensions of 70 μm × 685 μm × 120 cm Bosch etched in Si and capped with an anodically bonded glass lid. Dimensions of μGC column are ~1 cm × 1 cm (B). “H” on the horizontal top and bottom bands of panel A refers to low power, high-efficiency integrated MEMS heaters of the chip for fast/efficient temperature programming. Capillary ports are sealed with a thin membrane of Si. Panel C shows SEM micrograph of a 30 μm × 685 μm × 90 cm μGC column and illustrates the high aspect ratio of the channels. Panel D shows the μGC × μGC setup of HKUST-1 (120 cm) and ZIF-8 (30 cm) columns. Figure 1 Open in new tabDownload slide μGC column fabrication. (Panels A and B) Photographs of top (back) and bottom (front) of μGC column with channel dimensions of 70 μm × 685 μm × 120 cm Bosch etched in Si and capped with an anodically bonded glass lid. Dimensions of μGC column are ~1 cm × 1 cm (B). “H” on the horizontal top and bottom bands of panel A refers to low power, high-efficiency integrated MEMS heaters of the chip for fast/efficient temperature programming. Capillary ports are sealed with a thin membrane of Si. Panel C shows SEM micrograph of a 30 μm × 685 μm × 90 cm μGC column and illustrates the high aspect ratio of the channels. Panel D shows the μGC × μGC setup of HKUST-1 (120 cm) and ZIF-8 (30 cm) columns. The comparable polydimethylsiloxane (PDMS)-coated μGC column was coated in-house with a 300 nm thin film of PDMS OV-1 from Ohio-Valley, Inc., Marietta, OH. MOF deposition ZIF-8 thin films were prepared in stepwise layer-by-layer (LBL) fashion (19) using the following two solutions in methanol: (a) 25 mM zinc nitrate hexahydrate and (b) 50 mM 2-methylimidazole. The column was manually filled with the first solution using a syringe and then capped with GC-inlet septa and allowed to sit for 30 min at room temperature (RT). The solution was then blown out of the column using dry N2 and then rinsed with methanol. This procedure was then used in alternating sequence with the second reactant solution for the desired number of layers; 5, 13, 40 and 60 cycles were investigated, and little correlation to specific chromatographic metrics was noted. There was some SEM evidence that ZIF-8 deposition was not forming conformal films, but instead, forming varying sizes of islands/clusters with disparate morphologies. The illustrative 2D μGC data shown in this paper were with a 60-cycle ZIF-8 LBL deposition in microfabricated column with channel dimensions of 685 μm deep, 30 μm wide and 30 cm long. The conformal MOF thin films of HKUST-1 were synthesized LBL in situ (20, 21). The deposition equipment was modified from peristaltic pumps (21) to syringe pumps (Harvard Apparatus, Holliston, MA; models PHD2000 and 33), to enable the high pressure necessary to pump HKUST-1 reactants through the μGC channel. The syringe pumps were controlled by a LabView® program to alternate between a flow of 0.2 mM Cu(OAc)2, in ethanol and 1.0 mM trimesic acid in ethanol (each reactant cycle was followed by a pure ethanol rinse, preventing the reactants from reacting in solution). Ethanolic solutions of the reactants were prepared freshly before each deposition cycle. The temperature of the reaction was constant at 50°C. Since the reaction of reagents in solution was undesirable, the flow system was designed such that the stock solutions did not come into direct contact with each other at any point during each deposition cycle. This was achieved by placing a solenoid valve (NResearch, Inc., West Caldwell, NJ; Teflon™ wetted part # HP161T032) in line after each syringe pump to ensure that the flow was completely stopped when the pump was not operating due to the continuation of flow resulting from built up pressure in the syringes. Additionally, polyether ether ketone tubing and fittings (VICI-Valco Instruments Company, Inc., Houston, TX) were used to prevent unwanted SURMOF growth on the supply tubing. The flow rate of each sub-cycle (CuOAc2/EtOH, trimesic acid/EtOH, and EtOH rinse) was set such that five times the volume of the microcolumn and supply tubing were pumped for each sub-cycle during a 10 min period. Temperature was controlled using hotplate or a laboratory oven, measured using a thermocouple and recorded throughout the deposition process. Such volume- and time-normalized flow rate allowed for equivalent deposition despite changing flow cell/column volume (such as those that might be encountered with a quartz crystal microbalance, QCM, flow cell). An example flow rate for the μGC columns is 150 μL/min. Scanning electron microscopy and energy dispersive X-ray spectroscopy A Zeiss Supra 55VP field emission gun scanning electron microscope (Carl Zeiss Microscopy, LLC, Thornwood, NY) was used for SEM micrograph determination of mean MOF film thicknesses and morphology. Energy dispersive X-ray spectroscopy (EDS) was used to verify the chemical composition of the deposited MOF films. SEM was used along with a Bruker quad silicon drift EDS detector with Bruker Esprit software for data collection and analysis (Bruker AXS, Inc., Madison, WI). The data illustrated the high levels of Si within the cracks in the film and high levels of Cu and C within the film itself, verifying the formation of MOF film in the μGC columns (data not shown). Quartz crystal microbalance The QCM system was RQCM PLO-10i phase lock oscillator with OEM crystal holder/flow cell, from Maxtek (now Inficon, Bad Ragaz, Switzerland). The 5 MHz resonant frequency quartz crystals were 25.4 mm in diameter and coated with an outer layer of SiO2 to best match the surface of the μGC channels (Inficon part # 149277-1). Reaction temperature was maintained at a constant 50°C using a laboratory oven, and the temperature was also monitored and recorded throughout the deposition cycle. Crystals were cleaned immediately prior to the deposition by UV-ozonization for 30 min. Frequency measurements were recorded in situ and dry, RT (~23°C) frequency measurements were recorded before and after MOF depositions. The oscillator was capacitance matched dry at 23°C (where the first frequency measurement was taken). It was then matched in ethanol at 50°C (for in situ measurements) and then dried in a stream of N2, cooled to 23°C and capacitance matched (which is the postdeposition dry measurement). Gas chromatography GC of light hydrocarbons and methanol was performed on an Agilent 7890N system (Agilent Technologies, Santa Clara, CA) with an auto injector and equipped with a flame ionization detector (FID). A 1 mm i.d. Siltek®-Dactivated Uniliner® inlet liner (Restek, part # 21053-214.5) was used with split-inlet flow (split ratios vary). The μGC flow connections were made from the GC inlet to FID using 250 μm i.d. deactivated fused silica Rxi® Guard capillary pigtails at 30 and 22 cm, respectively (Restek, part # 10029). Interconnect union fittings for this arrangement were CapTite™ unions made of ULTEM® (LabSmith, Inc., Livermore, CA). Comprehensive, stop-flow μGC × μGC chromatograms were generated using Agilent 7890N, but with the second pressure-controlled GC inlet connected to a miniaturized 24 VDC solenoid spider valve (Figure 1) for modulating flow between the two columns (Staiger GmbH & Co., Germany; part # 604 000 269, distributed by Humphrey, Inc., Kalamazoo, MI, part # VA 204-716 E 07 SAL 80 24/00). The valve seat for gas connections was custom designed by Sandia National Laboratories (Albuquerque, NM) and GenTech Concepts, LLC (Centennial, CO), and machined (TEAM Technologies, Inc., Albuquerque, NM). Valve modulation was accomplished with LabView® and μGC × μGC chromatograms were generated using an in-house MATLAB® program. Chlorine retention was measured with a LECO Corporation (St. Joseph, MI) Pegasus® 4D GC × GC time-of-flight (TOF) mass spectrometer in 1-D chromatography mode using the same configuration as the 7890N, except for a longer, 50 cm, pigtail from μGC to the detector. Thermodynamics Partition coefficients were calculated from the stationary phase film thickness, and the column’s channel dimensions by SEM, and from the retention time k, which was chromatographically determined as a function of temperature. Both measurements were used to calculate the partition coefficients for HKUST-1 and select analytes. To provide one example of the approach, the film thickness of HKUST-1 was determined after 60 cycles, using SEM. SEM was additionally used to measure the actual μGC column dimensions. Both measurements were then used to calculate the partition coefficients. These data are represented by the van‘t Hoff equation, ln(K) = ln(k × β) = A + B/T. The intercept “A” multiplied by the ideal gas constant “R” yields the entropy (ΔS) of sorption for an analyte/stationary phase pair. The slope “B” multiplied by “R” is the enthalpy (ΔH). Here, “K” is the partition coefficient, “k” the retention factor (time) and “T” the absolute temperature in Kelvin. The mean HKUST-1 film thickness and exact column dimensions were obtained by SEM using five measurements for each of three disparate number-of-cycle μGC columns, each of which had been previously characterized chromatographically (i.e., k(T) determined for several analytes). Results μGC column design Silicon substrate was used in the fabrication of μGC (Figure 1), which is also the dominant platform for μGCs (22). Channel dimensions of μGC could be designed and tailored to optimize chromatographic performance. Three major microchip column geometries have been reported: serpentine, circular spiral and square spiral (3, 22). Circular spiral geometry was adopted for μGC column fabrication (Figure 1). Circular spiral columns enable better separation of analytes, and simulation experiments have demonstrated that the carrier gas flow was more uniform on the circular spiral column. Flow rate of carrier gas and pressure within the μGC column critically affect performance, and pressure variation was reported to be greater with serpentine columns (23). Ordered, homogenous stationary phase coating improves chromatographic efficiency, which is enhanced in circular spiral column relative to serpentine column (23). This is because changes in the distribution of airflow rate in the corners of the serpentine column are large, whereas this rate is more uniform in the circular spiral (23). The difference of the airflow rate in the channel could change the stationary phase film thickness, leading to tailing (23). Pressure change rate in the serpentine is also larger than that of the circular spiral. The circular spiral column has a coil-like pattern with no sharp turns. Serpentine column has sharp 180° turns resulting in pooling of stationary phase in the curvatures, leading to chromatographic inefficiency (22). Such deleterious effects were all minimized in circular spiral columns. The circular spiral model (Figure 1) had a more aerodynamic geometry relative to the serpentine. For these reasons, circular spiral was chosen as the μGC column geometry (Figure 1B). The column design and microfabrication techniques allowed the incorporation of high-aspect ratio channels that maximized the surface area of stationary phase, while minimizing the diffusional distance an analyte molecule must travel to interact with this stationary phase (Figure 1C). Thus, optimization of analyte retention over very short column lengths was realized, which in turn enabled fast separation/elution times while maintaining high peak capacity (number of individual chemicals/peaks that could be resolved within a single analytical cycle). The microelectromechanical system (MEMS) heaters (Figure 1A, “H”) utilized less than 7.5 W and could complete a temperature ramp cycle from 25 to 300°C and back to 25°C in ~75 s. By contrast, μGCs fitted with Kapton® heaters only achieved a maximum temperature of ~200°C, at ~15 W (2× power) and took 300 s (4× longer) to complete the 25–200°C to 25°C cycle (unpublished data). Characterizing deposition and film thickness using QCM Conventional capillary columns can suffer from uncontrolled film formation on the inner capillary walls, affecting reproducibility (24). Therefore, in-situ monitoring of MOF deposition and the determination of film thickness were performed using QCM with a 5 MHz polished SiO2-coated crystal (Figure 2). Due to stoppage of the syringe pumps and closure of the solenoid valves, a perfectly quiescent stage arose where consistent frequency measurements could be recorded between deposition cycles. For thickness calculations, the Sauerbrey equation and HKUST-1 specific gravity of 0.76 were used (21), and the data shown in Figure 2 are for a 40-cycle LBL deposition of HKUST-1. These data yielded a final HKUST-1 film thickness of 264 nm with an average single deposition-layer thickness of 6.6 nm. These values were higher than those reported previously (21) and higher than the 168 nm calculated by measuring the crystal’s frequency in air at 23°C, before and after deposition. It seemed likely that the absorbed ethanol during the deposition, might explain the difference. It was also noted that the deposition rate (i.e., change in film thickness) decreased with increasing deposition cycle number (Figure 2C). It appeared that H3BTC did not react fully within the 10 min sub-cycle as the deposition progressed; in other words, the reaction rate of H3BTC was slowing with increased film thickness. Increasing the time for each deposition cycle will eliminate incomplete reaction. An alternate method for determining the film thickness using partition coefficients is described later. Figure 2 Open in new tabDownload slide HKUST-1 film thickness by QCM. (Panel a) Frequency data are split into two panels because of the need to run two back-to-back deposition of 20 cycles due to limited syringe volume. (Panel b) Details of QCM frequency change for one complete deposition cycle. For this cycle at ~200 min on the 21–40 cycle, the QCM did not reach an equilibrium frequency for H3BTC step. (Panel c). Film thickness showed a decline in the deposition rate with increasing deposition cycle number, perhaps due to increasingly slower reaction of H3BTC as deposition progressed. Figure 2 Open in new tabDownload slide HKUST-1 film thickness by QCM. (Panel a) Frequency data are split into two panels because of the need to run two back-to-back deposition of 20 cycles due to limited syringe volume. (Panel b) Details of QCM frequency change for one complete deposition cycle. For this cycle at ~200 min on the 21–40 cycle, the QCM did not reach an equilibrium frequency for H3BTC step. (Panel c). Film thickness showed a decline in the deposition rate with increasing deposition cycle number, perhaps due to increasingly slower reaction of H3BTC as deposition progressed. μGC separation of natural gas mixture A μGC column (630 μm × 70 μm × 120 cm, with 10 cm total 250 μm i.d. inlet- and outlet-port capillary) was coated with a 35-layer HKUST-1 film (~65 nm) at 50°C. The stability of the HKUST-1 stationary phase was excellent in the μGC columns and no adverse temperature (as high as 150°C) effects or analyte effects were observed. This was contrasted with the stationary phase degradation that was seen with several of the capillary columns, which we hypothesize was due to the MOF material fracturing and detaching from the capillary surface during handling or bending of the columns. HKUST-1 film thickness and therefore the chromatographic retention, was highly reproducible, as will be seen from the data presented elsewhere in this paper (for example, Figure 6). A temperature ramp (30°C/min) from 22 to 150°C was carried out using H2 carrier gas at 15 psi inlet pressure (flow rate and velocity). A small amount of neat n-pentane and n-hexane was added to the bottom of the GC vial for a total liquid depth of 2 mm. The vial was then filled with natural gas mixture, and the chromatogram is shown in Figure 3. Figure 3 Open in new tabDownload slide Natural gas mixture separation using HKUST-1-μGC (left panel) and PDMS-μGC (right panel). Left panel: the 35-layer HKUST-1-coated 70–120 μm μGC (Table I) column showing the complete separation of natural gas (C1–C7). Right panel: chromatogram for a 70–120 μm (Table I) μGC column coated with bonded 300 nm PDMS. Analytes were: 1, methane; 2, ethane; 3, propane; 4, i-butane; 5, n-butane; 6, n-pentane; 7, n-hexane. Figure 3 Open in new tabDownload slide Natural gas mixture separation using HKUST-1-μGC (left panel) and PDMS-μGC (right panel). Left panel: the 35-layer HKUST-1-coated 70–120 μm μGC (Table I) column showing the complete separation of natural gas (C1–C7). Right panel: chromatogram for a 70–120 μm (Table I) μGC column coated with bonded 300 nm PDMS. Analytes were: 1, methane; 2, ethane; 3, propane; 4, i-butane; 5, n-butane; 6, n-pentane; 7, n-hexane. The data demonstrated an exceptional ability of HKUST-1 film to resolve and baseline separate light hydrocarbon mixtures (C1–C5 high volatiles) within ~0.6 min, including discriminating i-butane and n-butane isomers (Figure 3). Complete separation of light hydrocarbons up to n-heptane was within 2 min, meeting the design goal of chromatographic separations in the shortest possible time (25). The elution sequence of analytes followed an increasing order of their boiling points, their respective kinetic diameters ranging between 0.38 and 0.5 nm, with no evidence of reverse shape selectivity or molecular sieving (26). Despite <12% difference in the kinetic diameters of n-butane (0.4687 nm) and i-butane (0.5278 nm) (27), the two isomers could be separated, albeit with some tailing (Figure 3), contrary to the notion that gases are difficult to separate when their molecular diameters differ slightly (28). Difference in diameters might be responsible for the peak broadening of i-butane relative to n-butane (Figure 3). This phenomenon was noted, but not investigated further. The selectivity of HKUST-1 could also be improved by increasing the temperature or adjusting the temperature ramp/rate (29), since the more retained, larger sized analytes are more susceptible to temperature effects and mobile phase influences, relative to smaller solutes. Although smaller molecules enter the pore preferentially and adsorption of smaller, slender molecules is preferred (26), hydrophobic interactions have an effect by increasing the migration time through the stationary phase. This is because, as the molecular critical diameter increases, it leads to decreasing distance between the atoms of the molecule and the pore wall, resulting in stronger van der Waals interactions. These results are in contrast with the performance of an identical μGC column coated with PDMS that was unable to retain and separate the same natural gas light hydrocarbon mixture (Figure 3). Unlike the present work, using PDMS (core) and MOF-5 (shell) as stationary phase, methane and ethane were ill-separated, and separation of light hydrocarbons up to n-hexane took 15 min (24). The HKUST-1-coated μGC described here has applications in portable natural gas monitoring or where constituent proportions of a gas mixture are of interest. Although μGC was used for separating light hydrocarbons (30–36), the combination of μGC with MOF stationary phase for separating light hydrocarbons is described here for the first time (Figure 3). Since in typical hydrocarbon monitoring applications, a C1–C4 separation time of 50 s was too long (32), it is important to note that a separation time of less than 30 s was achieved here (Figure 3). Indeed, the separation time demonstrated for light hydrocarbon alkanes in this work is comparable to performances reported previously (32–34), and 4- to 10-fold faster than before (31). Furthermore, methane, ethane and propane were poorly separated in previous studies (35, 36), in contrast to the present work (Figure 3). This was important, since C1–C3 light alkanes are important raw chemical feedstocks and energy resources (1, 2) and methane is both a greenhouse gas and a cleaner replacement for fossil fuels. Our work on light hydrocarbon separation using MOF-stationary phase-μGC system is also the first of its kind, since earlier work on using MOFs for this purpose employed either packed (37) or capillary GC columns (16, 25, 38, 39) that were not microfabricated. Furthermore, in earlier studies, separation times of light hydrocarbons were either comparable (16, 25) to our data (Figure 3), or were considerably longer, between 6 (39) and >12 min (37). HKUST-1 thermodynamics The partition coefficient, K, is a temperature-dependent property of a stationary phase/analyte pair, defining the amount of analyte a stationary phase can sorb at a particular concentration of the analyte vapor at equilibrium. It is an intensive property and is therefore independent of the stationary phase thickness. The partition coefficient is related to the chromatographic retention time of an analyte as follows: $$\begin{equation} k(T)={K}_{i,j}/\beta =\frac{t_{R,i}-{t}_M}{t_M} \end{equation}$$(1) $$\begin{equation} {\beta}_j\equiv \frac{V_{\mathrm{col}}}{V_j}=\frac{h\cdot w}{2\cdot{d}_{f,j}\left(w+h-2\cdot{d}_{f,j}\right)} \end{equation}$$(2) Retention time k was calculated using tR for the analyte i that is relative to the retention time of an unretained analyte such as methane, represented as tM (methane might have some retention time with MOF stationary phase, but still represented the most unretained analyte for the FID detector). β is the phase volume ratio, which is the ratio of free volume of the column to the volume of the stationary phase j. β is approximated by equation (2), where h and w are the rectangular micro-column channel height and width, and df the film thickness of the stationary phase j; this assumes a thin conformal film coating. Figure 4 shows van‘t Hoff plots for HKUST-1 and the polar Rt-U-Bond® PLOT stationary phases for various polar and nonpolar, highly volatile analytes. It was clear that HKUST-1 μGC was consistently more retentive than the PLOT column for all light hydrocarbons tested, and significantly more retentive than the PLOT phase for the highly polar methanol, a light alcohol, employed as a TIC surrogate. This was significant, because PLOT columns are well-established and commercially available for a range of applications involving different stationary phases (40). It was also significant because PLOT columns are considered the current state-of-the-art for low molecular mass analysis (e.g., natural gas separations) because their low flow impedance relative to packed columns enabled the use of longer columns resulting in better separations of small molecular weight nonpolar gases (25, 38, 40, 41). Indeed, Rt-U-BOND® was highlighted by the manufacturer as being their highest polarity porous polymer column (42), specifically targeted for the analysis of nonpolar and polar compounds (40, 42), including separation of deuterated isotopologues of methanol and acetaldehyde (42). Despite these attributes, quantitatively for methanol, HKUST-1 μGC had a partition coefficient of 1214 at 150°C, relative to a value of 3.6 for the PLOT stationary phase, representing an increase of 33,600%. Figure 4 Open in new tabDownload slide Thermodynamics of analyte sorption to HKUST-1 μGC and commercial PLOT column stationary phases. Note: abscissa values are ×103. HKUST-1 column was 70 μm × 685 μm × 120 cm (Table I) μGC column with 35 deposition cycles of MOF (df = 65 nm and β = 464). Commercial PLOT column had a diameter of 320 μm and length of 120 cm (df = 10 μm and β = 8). Figure 4 Open in new tabDownload slide Thermodynamics of analyte sorption to HKUST-1 μGC and commercial PLOT column stationary phases. Note: abscissa values are ×103. HKUST-1 column was 70 μm × 685 μm × 120 cm (Table I) μGC column with 35 deposition cycles of MOF (df = 65 nm and β = 464). Commercial PLOT column had a diameter of 320 μm and length of 120 cm (df = 10 μm and β = 8). Stationary phase polarities The choice of stationary phase is most important when selecting a column for a desired application, chiefly its polarity or selectivity. For the purpose of comparing the relative polarity of the HKUST-1 stationary phase to that of other materials, the selectivity factor α was employed, which is defined as the ratio of partition coefficients or retention factors for two analytes, in this case, of differing polarity. Such a comparison of HKUST-1 μGC and Rt-U-BOND® stationary phases is shown in Figure 5. Here, the retention factor for nonpolar propane is about the same for both MOF and PLOT stationary phases (~7.5). However, switching to the polar analyte methanol yielded retention factors of 748 and 59, respectively. This resulted in methanol-to-propane selectivity factors of ~100 and ~8, respectively, a 13-fold increase in HKUST-1 polarity by this metric. The entire concept of GC × GC (vide infra) relies on having two columns with disparate, preferably orthogonal retention mechanisms (e.g., polarity); the strong polarity of HKUST-1 stationary phase was crucial for GC × GC separation of light, polar molecules. Figure 5 Open in new tabDownload slide Selectivity factors. Comparison of α provides a useful metric for ranking relative polarities of stationary phases (α is the ratio of retention factors, k, for methanol-to-propane at 30°C). HKUST-1 had an α of >100, whereas α of Rt-U-BOND® PLOT stationary phase was 7.8, a 13-fold polarity increase and illustrates the exquisite selectivity of HKUST-1 for light polar molecules. Figure 5 Open in new tabDownload slide Selectivity factors. Comparison of α provides a useful metric for ranking relative polarities of stationary phases (α is the ratio of retention factors, k, for methanol-to-propane at 30°C). HKUST-1 had an α of >100, whereas α of Rt-U-BOND® PLOT stationary phase was 7.8, a 13-fold polarity increase and illustrates the exquisite selectivity of HKUST-1 for light polar molecules. MOF film thickness and growth It was important to understand the fundamentals of MOF film formation for SURMOFs and to determine MOF film thickness in the stationary phase for optimizing column performance (43). Since monitoring LBL growth of HKUST-1 thin films by SEM, QCM and X-ray diffraction (44) are destructive or ex situ techniques, a nondestructive approach was explored. Knowledge of the partition coefficient, K, combined with retention factor data could be used to back-calculate the HKUST-1 film thickness. Film thickness can then be correlated to the number of MOF LBL deposition cycles, enabling tailored chromatographic performance. Figure 6A presents HKUST-1 film thickness as it is related to the number of LBL deposition cycles and illustrated a range of film thicknesses (10 nm–100 μm) that could be obtained by this technique (43). The chart also includes two data points for the mean film thickness as determined directly using SEM. The mean thickness values for each column cycle were derived from the average retention factor values measured at four temperatures for the four analytes: ethane, propane, i-butane and n-butane. In other words, each df value charted in Figure 6 represented 16 individual measurements across four analytes and at four temperatures (total of 112 data points for the seven measurements of Figure 6A). Figure 6A demonstrates remarkable alignment between SEM-determined and chromatographically extrapolated df values above 17 deposition cycles (r2 = 0.98). Figure 6 Open in new tabDownload slide MOF film thickness by chromatography. (Panel A) HKUST-1 film thickness plotted as a function of deposition cycles. Squares along the profile represent thickness measurements by SEM; remainder of data on the tracing were from known partition coefficients for several analytes and retention time (k) to calculate film thickness. (Panel B) HKUST-1 film deposition rate consistent with nucleation-driven film growth. (Panel C) SEM of HKUST-1 film after 10 deposition cycles. (Panel D) SEM of HKUST = 1 film after 25 deposition cycles. Figure 6 Open in new tabDownload slide MOF film thickness by chromatography. (Panel A) HKUST-1 film thickness plotted as a function of deposition cycles. Squares along the profile represent thickness measurements by SEM; remainder of data on the tracing were from known partition coefficients for several analytes and retention time (k) to calculate film thickness. (Panel B) HKUST-1 film deposition rate consistent with nucleation-driven film growth. (Panel C) SEM of HKUST-1 film after 10 deposition cycles. (Panel D) SEM of HKUST = 1 film after 25 deposition cycles. The film thickness data were examined next in terms of deposition rate as it related to the number of deposition cycles (Figure 6B). It was clear that the deposition rate was low during the first five deposition cycles, after which the rate increased rapidly until about 17 cycles. At this stage, it reached a steady state of 1.8 nm/cycle. Although the transition was not captured, HKUST-1 film morphology by SEM showed conformal coverage at 25 (Figure 6D) and 60 deposition cycles, whereas after 10-cycles, there was only island-type film growth (Figure 6C). Nijem et al. (44) reported that HKUST-1 film growth proceeded by LBL deposition mechanism from the beginning and for at least 40 layers. These results support the hypothesis that the nonconstant deposition rate (for low numbers of cycles) was not due to another deposition mechanism, but instead was due to nonconformal LBL deposition only at nucleation sites until complete coverage was reached at ~17 cycles. This agreed with the growth mechanism for HKUST-1 SURMOF, wherein LBL deposition was of “Volmer-Weber” type, with “small crystallites nucleating and ripening on the substrate upon continued deposition cycles” (43). The SEM image for 10-cycle deposition corroborated the thickness data estimated by chromatography (Figures 6A and B). Furthermore, SEM imaging of the 10-cycle HKUST-1 μGC column channel wall at various magnifications showed that the film morphology and surface coverage varied from the top to the bottom of the channel wall. It is possible that small discrepancies in surface chemistry of the Bosch (DRIE)-etched, nominally hydroxylated SiO2 channel walls had a large effect on HKUST-1 film growth and highlighted the importance of surface preparation and temperature control prior to and during MOF deposition (43). μGC × μGC Comprehensive two-dimensional GC (GC × GC) is highly effective for separating complex mixtures of volatile organic compounds (VOCs), due to the orthogonal character of such systems (2–8), meeting the requirements of speed, resolution, and selectivity (25). To the best of our information, this is the first reported MOF-coated μGC × μGC system (Figure 1) for the analysis of natural gas mixture. A goal of this work was proof-of-concept demonstration of 2D separation using microfabricated components. In the first example, the μGC × μGC system comprised a 70 μm × 685 μm × 120 cm (Table I) μGC-1 column coated with HKUST-1 and a 30 μm × 685 μm × 30 cm (Table I) μGC-2 column coated with ZIF-8 (Figure 1D). This system represented MOFs of two distinctly different polarities being demonstrated for the first time in a gas chromatographic separation of light hydrocarbons. The data of Figures 7 and 8 are the first ever demonstration of μGC × μGC separation of model light alkane hydrocarbon mixture (C1–C4), with the use of MOF stationary phases of disparate polarity. Separation times for light hydrocarbons in our μGC × μGC system were an order of magnitude faster relative to a previously reported PLOT×Molecular Sieve nonmicrofabricated multidimensional GC data (2). Figure 7 Open in new tabDownload slide μGC × μGC (HKUST-1 × ZIF-8) separation of light alkane hydrocarbons (C1–C4). (Panel A) Chromatographic data without processing (i.e., detector amplitude as a function of time). Each individual slice represents a separate second column chromatogram (inset). (Panel B) Contour (3D) chromatogram showing the separation of C1–C4 analytes. The normalized amplitude contour scale ranges from 0 to 0.12, with the methane peak cut short for clarity. Figure 7 Open in new tabDownload slide μGC × μGC (HKUST-1 × ZIF-8) separation of light alkane hydrocarbons (C1–C4). (Panel A) Chromatographic data without processing (i.e., detector amplitude as a function of time). Each individual slice represents a separate second column chromatogram (inset). (Panel B) Contour (3D) chromatogram showing the separation of C1–C4 analytes. The normalized amplitude contour scale ranges from 0 to 0.12, with the methane peak cut short for clarity. Figure 8 Open in new tabDownload slide μGC × μGC (both HKUST-1) separation of light alkane hydrocarbons. (Panel A) 3-D surface plot. (Panel B) 3-D contour plot. Figure 8 Open in new tabDownload slide μGC × μGC (both HKUST-1) separation of light alkane hydrocarbons. (Panel A) 3-D surface plot. (Panel B) 3-D contour plot. Figure 8 shows the results from a μGC × μGC system, in which both the long and short GC-1 and GC-2 columns were coated with HKUST-1 film. Whilst this might not be as interesting from a separation science perspective, since the same stationary phase is used in both columns, it does demonstrate a higher resolution chromatogram with more symmetric peaks (Figure 8) than was possible with ZIF-8. This difference might be due to nonuniform ZIF-8 stationary phase films. That said, using the same stationary phase in GC analysis is not unique. For example, Chang et al. (45) used ZIF-8 for both solid phase microextraction (SPME) and as a GC stationary phase (SPME/GC) for the analysis of n-alkanes. Additionally, the figure demonstrates the reproducibility of the MOF film stationary phase, as the two microfabricated columns show identical retention characteristic, as seen by the linear response vector. If this response vector were to show any nonlinearity, that would indicate variation in the phase partition coefficients. The μGC × μGC demonstrated its value by the separating isomers. The μGC × μGC contour-plots (Figures 7 and 8) also demonstrated the benefits of multidimensional GC (2–8), due to clear depictions of the various analytes within a specific area of the 2D plane. The surface profile of a GC × GC chromatogram in the 1D presentation (Figure 3) when compared to the 2D version (Figures 7 and 8) demonstrated that comprehensive two-dimensional GC (2–8) had better separation/resolution. Visually too, 2D presentation of the data was both appealing and readily comprehensible. Toxic industrial chemicals The tremendous physical, structural and chemical diversity offered by MOFs make them suitable for applications relating to TICs (10). TICs are difficult to separate and detect with traditional polymer-based GC stationary phases. TICs typically have extremely high volatility and are often hydrophilic (polar), and as such, lack the necessary retention on polymer films to facilitate chromatographic separation. Even strongly polar polymers such as PEG and ionic liquids might lack sufficient retention for these high-volatility analytes on short columns. PEG also suffers from deterioration over time, especially in the presence of oxygen and heat. The next set of results showed the suitability of MOF-coated μGCs for the analysis of TICs and broadened the application space of MOFs in μGC. Due to safety rigor associated with TICs, methanol was used as a surrogate to demonstrate the separation behavior on the HKUST-1 μGC column. The partition coefficients of the remaining TICs shown in Figure 9 were analyzed on the HKUST-1 μGC column with an FID detector. Specifically, the CS2 used was HPLC grade (99.9% purity), and despite the relatively low FID signal, CS2 identification was achieved, consistent with earlier findings of FID sensitivity to CS2 (46). Retention times were thus determined for CS2, Cl2, CH3Cl and CH2Cl2, enabling the determination of partition coefficients at 50°C for these TICs, compared to light alkane hydrocarbons. The partition coefficient values for the TIC surrogates are intermediate to those of ethane and propane (Figure 9). Figure 9 Open in new tabDownload slide Partition coefficients at 50°C for HKUST-1 (20 deposition cycles) μGC column and selected TIC surrogates. Figure 9 Open in new tabDownload slide Partition coefficients at 50°C for HKUST-1 (20 deposition cycles) μGC column and selected TIC surrogates. Discussion Although promising applications of MOFs as stationary phase using packed or capillary GC columns were reported (11–14), packed column performance necessitates the particle size of the stationary phase to be large to avoid pressure drop during operations. Early work on MOF-GC was by using large crystals/pellets of MOF as stationary phase. Large-scale MOF synthesis yields mainly polycrystalline materials that are unsuited to packed columns due to pressure issues (14, 24). Packed columns require large-scale MOF synthesis adding to cost, and such columns could result in poor resolution of analytes with peak broadening (14). This is because, large-scale MOF synthesis yields mainly polycrystalline materials that are unfavorable to column performance. The polydispersity of such MOF preparations includes broad particle size distribution and nonspherical and irregular shape of the particles. These irregularities lead to high back-pressure, low column performance and peak broadening (14). To minimize these concerns, researchers switched to MOF-coated capillary columns (11–14). However, capillary columns can be fragile and might suffer from column bleeding due to spiking (unless a particle filter is used) (14). It is also difficult to coat capillary columns with MOF particles (11) and such columns require high power and longer time to both heat and cool during temperature-programmed chromatography. Incorporation of MOFs into monolithic columns can avoid these difficulties (14). However, to the best of our information, there have been no reports of MOFs as stationary phase in μGC. μGC involves columns microfabricated using planar substrates and MEMS for miniaturization (2–8, 22, 23). Current instruments for VOC detection are generally large, expensive, and power intensive. The advantages of μGC include lightweight, small footprint, low power requirement, rapid response, on-time/on-site analysis, favorable cost comparison to hybrid systems, low maintenance and portable instrumentation for field work relating to a wide range of civilian, military, homeland security, environmental, food and medical applications (2–8, 17, 22–24). The present work demonstrated for the first time, MOF-coated stationary phase in high-speed μGC applications, using natural gas and TICs as exemplars and broadened the scope of MOFs in GC. To illustrate size-scale of the μGC system, the HKUST-1 coated GC-1 is about the size of a U.S. penny without electronics, detector, battery, valve inlet/outlet plumbing, etc. Based on Sandia’s prior experience building portable systems, a portable version of this system would be expected to occupy <1 L, or roughly the size of a hardback book with flexibility to shrink, if necessary. This was ~230-fold smaller and ~100-fold lighter than commercial GC × GC-FID systems, not including the cryogenic source used for the commercial modulator or carrier gasses. As such, this μGC × μGC system represented a vast improvement for portable applications. The inextricable linkage between the df obtained from SEM and df the obtained from chromatography (Figure 6) raises the intriguing question whether the whole stationary phase thickness was used for separating the analytes. For the small molecular analytes examined in this study, it is speculated that the data indeed indicate that the entirety of the MOF film is absorbing the analyte and behaving as a polymer phase would (and that the estimation relying on the sorption equilibrium is still a good equilibrium). Phenomena, such as reverse (boiling point) chromatography and extreme tailing have been observed for other larger analytes. It would be interesting to see if the polymer absorption model still holds for these larger analytes as well; it is suspected that it would not. The MOF-μGC performance was superior to PLOT columns in the separation of light alkane hydrocarbons with respect to both polarity and absolute partition coefficients. We also demonstrated for the first time, μGC × μGC system coated with orthogonal MOF stationary phases for natural gas analysis. An LBL technique was used for MOF deposition upon μGC column surface with exquisite control over MOF film thickness and chromatographic performance. A nondestructive chromatographic technique was developed for estimating MOF film thickness upon μGC surface, that correlated well with thickness measurements from destructive or ex situ methods. The data suggested that MOFs behaved similarly in terms of their sorption thermodynamics relative to traditional polymer stationary phases, an unexpected result. The data led to the hypothesis that HKUST-1 film growth followed the Volmer-Weber model (43). These studies broaden the scope of MOFs as GC stationary phases, and for μGC applications. The typical U.S. natural gas supply would have no issues separating the light alkanes using existing 1D MEMS columns. However, the thrust of our paper was μGC and μGC × μGC proof-of-concept, using MOFs as stationary phases. The goal was also to demonstrate separation in the second dimension, due to the branched structure of butane isomers. Internationally, natural gas supply is adapting to increasing contributions from heavier supplies, with several countries routinely requiring not just C1–C6+, with all C6 and heavier fractions combined into a single measurement, but also including C1–C12+ to account for the greater energy contributions of the heavier fragments. With increasing carbon number, the number of naturally occurring isomers increases exponentially, and thus the separation becomes impractical on a 1D microfabricated column. Another key factor driving our development of polar MOF films and μGC × μGC was for field-portable separation of TICs in a battlefield or civilian emergency environment for rapid analysis of threat agents. It is unlikely that a 1D μGC column could achieve the required separation performance under such scenarios. Due to toxicity of TICs, it was neither safe nor practical to test the μGC columns with the TICs. Proof-of-concept was therefore achieved using natural gas as a surrogate. The tailing observed with n-butane (Figure 7) is not altogether an uncommon encounter in chromatography, since non-Gaussian shaped peaks do appear, and the shape of the peak can provide information regarding the variables present during GC. Tailing occurs when the analyte adsorption sites in the stationary phase strongly retain an analyte, due to concentration-dependent, nonlinear distribution coefficient. However, tailing peaks have not interfered with analyte quantitation (47), particularly when the tailing is reproducible and does not overlap with an adjacent peak. It is the overlap between n-butane and i-butane that makes quantitation challenging. The separation shown in Figure 7 illustrates yet another advantage of the μGC × μGC approach, by separating this partial coelution in the second dimension, thereby enabling more accurate quantitation. In conclusion, this work demonstrates for the first time the use of MOFs as stationary phases in μGC columns in one- and two-dimensional systems. The systems were used to demonstrate the separation of light hydrocarbons in natural gas and TICs. Such systems are envisioned for deployment in battlefield or civilian emergency situations. The studies described here advance the status of MOF in μGC and μGC × μGC applications. Acknowledgements This work/publication was funded by Sandia National Laboratories, Laboratory Directed Research and Development (LDRD) program, projects # 173133 and # 199974. We thank Dorina Sava-Gallis for guidance with MOF synthesis, Joseph Simonson for technical and programmatic advice, Matthew Moorman, John Anderson, and Ronald Manginell for μGC fabrication development, Daniel Porter for valve selection studies, Darren Graf for valve interface hardware design, Bonnie McKenzie for SEM imaging and Nathaniel Pfeifer for technical help. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525. This paper describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the United States Government. References 1. Mango , F.D. ; The light hydrocarbons in petroleum: a critical review ; Organic Geochemistry , ( 1997 ); 26 ( 7–8 ): 417 – 440 . Google Scholar Crossref Search ADS WorldCat 2. Luong , J. , Gras , R., Cortes , H.J., Shellie , R.A.; Multidimensional gas chromatography for the characterization of permanent gases and light hydrocarbons in catalytic cracking process ; Journal of Chromatography A , ( 2013 ); 1271 : 185 – 191 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Regmi , B.P. , Agah , M.; Micro gas chromatography: an overview of critical components and their integration ; Analytical Chemistry , ( 2018 ); 90 ( 22 ): 13133 – 13150 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Whiting , J.J. , Myers , E., Manginell , R.P., Moorman , M.W., Anderson , J., Fix , C.S. et al. ; A high-speed, high-performance, microfabricated comprehensive two-dimensional gas chromatograph ; Lab on a Chip , ( 2019 ); 19 ( 9 ): 1633 – 1643 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Prebihalo , S.E. , Berrier , K.L., Freye , C.E., Bahaghighat , H.D., Moore , N.R., Pinkerton , D.K. et al. ; Multidimensional gas chromatography: advances in instrumentation, chemometrics, and applications ; Analytical Chemistry , ( 2018 ); 90 ( 1 ): 505 – 532 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Collin , W.R. , Bondy , A., Paul , D., Kurabayashi , K., Zellers , E.T.; μGC × μGC: comprehensive two-dimensional gas chromatographic separations with microfabricated components ; Analytical Chemistry , ( 2015 ); 87 ( 3 ): 1630 – 1637 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Seeley , J.V. , Seeley , S.K.; Multidimensional gas chromatography: fundamental advances and new applications ; Analytical Chemistry , ( 2013 ); 85 ( 2 ): 557 – 578 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Krupcik , J. , Gorovenko , R., Spanik , I., Sandra , P., Armstrong , D.W.; Flow-modulated comprehensive two-dimensional gas chromatography with simultaneous flame ionization and quadrupole mass spectrometric detection ; Journal of Chromatography A , ( 2013 ); 1280 : 104 – 111 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Azzouz , I. , Essoussi , A., Fleury , J., Haudebourg , R., Thiebaut , D., Vial , J.; Feasibility of the preparation of silica monoliths for gas chromatography: fast separation of light hydrocarbons ; Journal of Chromatography A , ( 2015 ); 1383 : 127 – 133 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Bobbitt , N.S. , Mendonca , M.L., Howarth , A.J., Islamoglu , T., Hupp , J.T., Farha , O.K. et al. ; Metal-organic frameworks for the removal of toxic industrial chemicals and chemical warfare agents ; Chemical Society Reviews , ( 2017 ); 46 ( 11 ): 3357 – 3385 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Gu , Z.Y. , Yang , C.X., Chang , N., Yan , X.P.; Metal-organic frameworks for analytical chemistry: from sample collection to chromatographic separation ; Accounts of Chemical Research , ( 2012 ); 45 ( 5 ): 734 – 745 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Yu , Y. , Ren , Y., Shen , W., Deng , H., Gao , Z.; Applications of metal-organic frameworks as stationary phases in chromatography ; Trends in Analytical Chemistry , ( 2013 ); 50 : 33 – 41 . Google Scholar Crossref Search ADS WorldCat 13. Yusuf , K. , Aqel , A., ALOthman , Z.; Metal-organic frameworks in chromatography ; Journal of Chromatography A , ( 2014 ); 1348 : 1 – 16 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Zhang , J. , Chen , Z.; Metal-organic frameworks as stationary phase for application in chromatographic separation ; Journal of Chromatography A , ( 2017 ); 1530 : 1 – 18 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Borigin , T. , Sun , F., Zhang , J., Cai , K., Ren , H., Zhu , G.; A microporous metal-organic framework with high stability for GC separation of alcohols from water ; Chemical Communications , ( 2012 ); 48 ( 61 ): 7613 – 7615 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Munch , A.S. , Mertens , F.O.R.L.; HKUST-1 as an open metal site gas chromatographic stationary phase—capillary preparation, separation of small hydrocarbons and electron donating compounds, determination of thermodynamic data ; Journal of Material Chemistry , ( 2012 ); 22 ( 20 ): 10228 – 10234 . Google Scholar Crossref Search ADS WorldCat 17. Manginell , R.P. , Bauer , J.M., Moorman , M.W., Sanchez , L.J., Anderson , J.M., Whiting , J.J. et al. ; A monolithically-integrated μGC chemical sensor system ; Sensors , ( 2011 ); 11 ( 7 ): 6517 – 6532 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Moorman , M.W. , Manginell , R.P., Anderson , J.M., Simonson , R.J., Read , D.H.; Sealed micro-gas-chromatography columns and methods thereof . US Patent: 10,151,732 ; 2018 . 1 – 15 . 19. Eslava , S. , Zhang , Z., Esconjauregui , S., Yang , J., Vanstreels , K., Baklanov , M.R. et al. ; Metal-organic framework ZIF-8 films as low-κ dielectrics in microelectronics ; Chemistry of Materials , ( 2013 ); 25 ( 1 ): 27 – 33 . Google Scholar Crossref Search ADS WorldCat 20. Shekhah , O. , Wang , H., Strunskus , T., Cyganik , P., Zacher , D., Fischer , R. et al. ; Layer-by-layer growth of oriented metal organic polymers on a functionalized organic surface ; Langmuir , ( 2007 ); 23 ( 14 ): 7440 – 7442 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Stavila , V. , Volponi , J., Katzenmeyer , A.M., Dixon , M.C., Allendorf , M.D.; Kinetics and mechanisms of metal-organic framework thin film growth: systematic investigation of HKUST-1 deposition on QCM electrodes ; Chemical Science , ( 2012 ); 3 ( 5 ): 1531 – 1540 . Google Scholar Crossref Search ADS WorldCat 22. Ghosh , A. , Vibrio , C.R., Hawkins , A.R., Lee , M.L.; Microchip gas chromatography columns, interfacing and performance ; Talanta , ( 2018 ); 188 : 463 – 492 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Sun , J. , Cui , D.F., Cai , H.Y., Chen , X., Zhang , L.L., Li , H.; Design, modeling, microfabrication and characterization of the micro gas chromatography columns ; Advances in Gas Chromatography , ( 2012 ); 3 : 51 – 66 . Google Scholar OpenURL Placeholder Text WorldCat 24. Manju , Roy , P.K., Ramanan , A., Rajagopal , C.; Core-shell polysiloxane-MOF 5 microspheres as a stationary phase for gas-solid chromatographic separation ; RSC Advances , ( 2014 ); 4 ( 34 ): 17429 – 17433 . Google Scholar Crossref Search ADS WorldCat 25. Munch , A.S. , Seidel , J., Obst , A., Weber , E., Mertens , F.O.R.L.; High-separation performance of chromatographic capillaries coated with MOF-5 by the controlled SBU approach ; Chemistry A European Journal , ( 2011 ); 17 ( 39 ): 10958 – 10964 . Google Scholar Crossref Search ADS WorldCat 26. Chang , N. , Yan , X.P.; Exploring reverse shape selectivity and molecular sieving effect of metal-organic framework UIO-66 coated capillary column for gas chromatographic separation ; Journal of Chromatography A , ( 2012 ); 1257 : 116 – 124 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Gehre , M. , Guo , Z., Rothenberg , G., Tanase , S.; Sustainable separations of C4-hydrocarbons by using microporous materials ; ChemSusChem , ( 2017 ); 10 ( 20 ): 3947 – 3963 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Shimomura , S. , Higuchi , M., Matsuda , R., Yoneda , K., Hijikata , Y., Kubota , Y. et al. ; Selective sorption of oxygen and nitric oxide by an electron-donating flexible porous coordination polymer ; Nature Chemistry , ( 2010 ); 2 : 633 – 637 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Zhao , Z. , Wang , S., Yang , Y., Li , X., Li , J., Li , Z.; Competitive adsorption and selectivity of benzene and water vapor on the microporous metal organic frameworks (HKUST-1) ; Chemical Engineering Journal , ( 2015 ); 259 : 79 – 89 . Google Scholar Crossref Search ADS WorldCat 30. Lefebvre , D. , Ricoul , F., Charleux , B., Thieuleux , C.; A novel stationary phase for light alkanes separation in microfabricated silicon gas chromatography columns. In 17th International Conference on Miniaturized Systems for Chemistry and Columns, Freiburg, Germany . ( 27–31 October 2013 ), pp. 566 – 568 . 31. Sklorz , A. , Janssen , S., Lang , W.; Application of a miniaturized packed gas chromatography column and a SnO2 gas detector for analysis of low molecular weight hydrocarbons with focus on ethylene detection ; Sensors and Actuators B , ( 2013 ); 180 : 43 – 49 . Google Scholar Crossref Search ADS WorldCat 32. Haudebourg , R. , Vial , J., Thiebaut , D., Danaie , K., Breviere , J., Sassiat , P. et al. ; Temperature-programmed sputtered micromachined gas chromatography columns: an approach to fast separations in oil field applications ; Analytical Chemistry , ( 2013 ); 85 ( 1 ): 114 – 120 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Vial , J. , Thiebaut , D., Marty , F., Guibal , P., Haudebourg , R., Nachef , K. et al. ; Silica sputtering as a novel collective stationary phase deposition for microelectromechanical system gas chromatography columns: feasibility and first separations ; Journal of Chromatography A , ( 2011 ); 1218 : 3262 – 3266 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Lee , J. , Park , T.H., Kang , H.S., Lim , S.H.; Miniaturized gas chromatography module with micro posts embedded MEMS column for the separation of exhaled breath gas mixtures ; 2016 IEEE Sensors , ( 2016 ); ( October 30–November 3 ): 1 – 3 . Google Scholar OpenURL Placeholder Text WorldCat 35. Luo , F. , Feng , F., Hou , L., You , W., Xu , P., Li , X., et al. ; The exploration of mesoporous silica as a stationary phase support for semi-packed micro-fabricated gas chromatographic (GC) columns. In MEMS2017, 22–26 January 2017 . pp. 1071 - 1074 . 36. Hou , L. , Feng , F., You , W., Xu , P., Luo , F., Tian , B. et al. ; Pore size effect of mesoporous silica stationary phase on the performance of microfabricated gas chromatography columns ; Journal of Chromatography A , ( 2018 ); 1552 : 73 – 78 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Chen , B. , Liang , C., Yang , J., Contreras , D.S., Clancy , Y.L., Lobkovsky , E.B. et al. ; A microporous metal-organic framework for gas-chromatographic separation of alkanes ; Angewandte Chemie International Edition , ( 2006 ); 45 ( 9 ): 1418 – 1421 . Google Scholar Crossref Search ADS WorldCat 38. Fan , L. , Yan , X.P.; Evaluation of isostructural metal–organic frameworks coated capillary columns for the gas chromatographic separation of alkane isomers ; Talanta , ( 2012 ); 99 : 944 – 950 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Wharmby , M.T. , Pearce , G.M., Mowat , P.S., Griffin , J.M., Ashbrook , S.E., Wright , P.A. et al. ; Synthesis and crystal chemistry of the STA-12 family of metal N,N′-piperazinebis(methylenephosphonate)s and applications of STA-12(Ni) in the separation of gases ; Microporous Mesoporous Materials , ( 2012 ); 157 : 3 – 17 . Google Scholar Crossref Search ADS WorldCat 40. de Zeeuw , J. , Majors , R.E.; The development and applications of PLOT columns in gas-solid chromatography ; LCGC North America , ( 2010 ); 28 ( 10 ): 848 – 864 . Google Scholar OpenURL Placeholder Text WorldCat 41. Ji , Z. , Chang , I.L.; A new look at light hydrocarbon separations on commercial alumina PLOT columns: column selectivity and separation ; Journal of Separation Science , ( 1996 ); 19 ( 1 ): 32 – 36 . Google Scholar OpenURL Placeholder Text WorldCat 42. Schmarr , H.G. , Wacker , M., Mathes , M.; Isotopic separation of acetaldehyde and methanol from their deuterated isotopologues on a porous layer open tubular column allows quantification by stable isotope dilution without mass spectrometric detection ; Journal of Chromatography A , ( 2017 ); 1481 : 111 – 115 . Google Scholar Crossref Search ADS PubMed WorldCat 43. Brower , L.J. , Gentry , L.K., Napier , A.L., Anderson , M.E.; Tailoring the nanoscale morphology of HKUST-1 thin films via codeposition and seeded growth ; Beilstein Journal of Nanotechnology , ( 2017 ); 8 : 2307 – 2314 . Google Scholar Crossref Search ADS PubMed WorldCat 44. Nijem , N. , Fursich , K., Kelly , S.T., Swain , C., Leone , S.R., Giles , M.K.; HKUST-1 thin film layer-by-layer liquid phase epitaxial growth: film properties and stability dependence on layer number ; Crystal Growth & Design , ( 2015 ); 15 ( 6 ): 2948 – 2957 . Google Scholar Crossref Search ADS WorldCat 45. Chang , N. , Gu , Z.Y., Yan , X.P.; Zeolitic imidazolate framework-8 nanocrystal coated capillary for molecular sieving of branched alkanes from linear alkanes along with high resolution chromatographic separation of linear alkanes ; Journal of American Chemical Society , ( 2010 ); 132 ( 39 ): 13645 – 13647 . Google Scholar Crossref Search ADS WorldCat 46. Douglas , D.M. , Schaefer , B.A.; Response of the flame ionization detector to carbon disulfide ; Journal of Chromatographic Science , ( 1969 ); 7 ( 7 ): 433 – 438 . Google Scholar Crossref Search ADS WorldCat 47. Wahab , M.F. , Patel , D.C., Armstrong , D.W.; Total peak shape analysis: detection and quantitation of concurrent fronting, tailing, and their effect on asymmetry measurements ; Journal of Chromatography A , ( 2017 ); 1509 : 163 – 170 . Google Scholar Crossref Search ADS PubMed WorldCat © The Author(s) 2020. Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com This work is written by US Government employees and is in the public domain in the US. © The Author(s) 2020. Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com
TI - Metal-Organic Framework Stationary Phases for One- and Two-Dimensional Micro-Gas Chromatographic Separations of Light Alkanes and Polar Toxic Industrial Chemicals
JF - Journal of Chromatographic Science
DO - 10.1093/chromsci/bmaa005
DA - 2020-04-25
UR - https://www.deepdyve.com/lp/oxford-university-press/metal-organic-framework-stationary-phases-for-one-and-two-dimensional-BPmJEKklHo
SP - 389
EP - 400
VL - 58
IS - 5
DP - DeepDyve
ER -