Thermal Independent Modulator for Comprehensive Two-Dimensional

Aug 18, 2016 - The 1D was connected to the inlet of modulation column of TiM by a capillary flow technology (CFT) purged union (Agilent Technologies, ...
2 downloads 38 Views 2MB Size
Download PDF
Technical Note pubs.acs.org/ac

Thermal Independent Modulator for Comprehensive TwoDimensional Gas Chromatography Jim Luong,*,†,‡ Xiaosheng Guan,§,∥ Shifen Xu,§ Ronda Gras,†,‡,⊥ and Robert A. Shellie‡,⊥,# †

Dow Chemical Canada ULC, Highway 15, Fort Saskatchewan, Alberta, T8L 2P4, Canada Australian Centre for Research on Separation Science (ACROSS) and ⊥ARC Training Centre for Portable Analytical Separation Technologies (ASTech), University of Tasmania, Private Bag 75, Hobart, Tasmania 7001, Australia § Agilent Technologies (Shanghai) Co., Ltd., 412 Ying Lun Road, Shanghai 200131, China # Trajan Scientific and Medical, 7 Argent Place Ringwood, Victoria 3154, Australia ‡

S Supporting Information *

ABSTRACT: We introduce a modulation strategy for comprehensive twodimensional gas chromatography (GC×GC) with complete thermal independence between the cooling and heating stages and without the need for GC oven heat for remobilization. Based on this approach, a compact thermal independent modulator (TiM) with thermoelectric cooling and micathermic heating has been successfully innovated for use in GC×GC. The device operates externally to a gas chromatograph, does not require liquid cryogen, and has minimal consumables requirements. The augmentation of an additional gas flow stream results in a number of critical chromatographic parameter improvements such as the decoupling of flows of first- and second-dimension columns to attain both efficiency and speed optimized flow in each dimension, the potential for independent retention time locking or scaling in either dimension, the improvement of modulator reinjection efficiency, as well as facilitating back-flushing for the first dimension to enhance system cleanliness and throughput. TiM was found to be useful for chromatographic applications over a volatility range equivalent to nC6 to nC24 under conditions used. Repeatability of retention time for model compounds such as benzene, toluene, ethyl benzene, and xylenes were found to be quite satisfactory with relative standard deviations of less than 0.009% in 1D and less than 0.008% in 2D (n = 10). Typical peak widths of 120 ms or less with a relative standard deviation of less than 4.7% were achieved for the aromatic model compounds. In this article, the performance of the modulator is demonstrated and a series of challenging chromatographic applications are presented to illustrate usefulness of the apparatus.

W

N2 (l)) with the latter being able to trap low boiling-point compounds down to n-propane (n-C3); meanwhile, remobilizing is achieved either by direct GC oven heating or by applying additional streams of hot gas. The amount of cryogen supplied to a modulator typically has a high operational cost5−7 and may need additional hardware and algorithms to control especially during a temperature-programmed GC×GC analysis, thereby adding another level of complexity to the approach. Efforts have been made to reduce or eliminate the need for cryogen via closed-loop chiller systems. For example, Muscalu et al. reported the use of a single stage thermal modulation for GC×GC.8 The device traps analytes using a proprietary stainless steel capillary trap compressed between two ceramic cooling pads. Analytes are thermally desorbed from the trap into the second column using resistive heating. Microfabrication technologies have also provided a route to eliminating cryogen. A low-thermal-mass, two-stage thermal modulator employing a multistage thermo-electric cooler

e introduce a strategy of complete thermal independence between the cooling and heating stages in a thermal modulator for comprehensive two-dimensional gas chromatography (GC×GC). Thermal modulators use a difference in temperature to achieve sampling and reinjection. When a thermal modulator is in sampling mode, the effluent from the first-dimension separation column (1D) is focused by strong retention or condensation in the column. When the modulator is switched to reinjection mode, the focused band is quickly desorbed or revaporized with thermal energy. Twostage thermal modulation is typically implemented to avoid the possibility of analyte breakthrough to the second-dimension separation column (2D) during switching between these two modes. The first dual-stage thermal modulator was invented by Liu and Phillips in 1991.1 Its trapping approach used air at ambient temperature and remobilized with resistively heated conductive paint applied on a fused silica column. Later in 1997, Marriott and Kinghorn pioneered a thermal modulation approach using a cryogenically cooled trap moving inside the GC oven.2−4 Trapped solutes were conveniently remobilized by direct GC oven heating. Contemporary thermal modulators use nonmoving jets; trapping uses CO2 (l) or dry N2 (g) (chilled using © XXXX American Chemical Society

Received: July 2, 2016 Accepted: August 18, 2016

A

DOI: 10.1021/acs.analchem.6b02525 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

zone, as illustrated in Figure 2. The trapping zone is a copper block sandwiched between cooling surfaces of a pair of threestage TEC elements (Marlow Industries, Texas, U.S.A.). A resistance temperature detector (RTD) is inserted in the copper block for temperature monitoring and control. The copper block and TECs are clamped together in an enclosure, which is formed by a 4-sided wall made of stainless steel and a pair of fan cooled heat-sinks for the TECs. The remobilizing zones are aluminum chambers with micathermic heaters attached at back and RTDs secured inside. On the side wall of each chamber facing the trapping zone, a copper tube is attached and, hence, is conductively heated by the chamber from this end. The other end of the copper tube, bearing a smaller outer diameter from the middle, protrudes into the enclosure through a silicone insulation seal embedded in the enclosure wall, but away from the copper block; therefore, thermal crosstalk between the three zones is kept to a minimum, allowing their temperatures to be independently controlled. The copper tubes are mechanically aligned with the throughhole of the copper piece, which permits a 0.36 mm o.d. fused silica modulation column to move freely inside. The modulation column is crimped by a Siltite metal ferrule, which in turn fits into the end hook of a gripper. The gripper is attached to a shaft extended from a rotary solenoid behind of the driving side chamber. The rotary solenoid turns by 45° between energized and de-energized positions and, through the gripper length, drives the modulation column back-and-forth by 30 mm, within 25 ms in each direction. The modulation column is bent 270° into a loop shape in each chamber and exits through a copper transfer line screwed onto the chamber wall. The transfer lines go through the GC oven wall into the oven so that they are heated from both ends by the chamber and the oven as well. The modulation column connects to 1D and 2D columns inside the GC oven at the end of the transfer lines on the driving side and the releasing side, respectively over the back inlet slot for ease of maintenance. Samples were injected by a 7693B automated liquid sampler (Agilent) into the GC×GC system. All the samples and standard solutions were analyzed by a GC×GC-FID setup, with helium as carrier gas. If preferred, hydrogen as a carrier gas can also be employed. The 1D was connected to the inlet of modulation column of TiM by a capillary flow technology (CFT) purged union (Agilent Technologies, Wilmington, Delaware, U.S.A.). The flow and pressure in the purged union was driven by the first channel of the PCM. The outlet of modulation column is connected to the 2D column by a quartz press-fit union. Unless otherwise stated, all chemicals used were obtained from Sigma-Aldrich (Oakville, Ontario, Canada). The hydrocarbon C18 to C24 test mixture was obtained from Restek (Bellefonte, PA, U.S.A.), while the hydrocarbon test mixture from C6 to C24 was blended in-house. Gasoline and diesel samples were obtained from fuel depots in Fort Saskatchewan, Canada. Table 1 lists the chromatographic conditions used. Data were collected using Agilent ChemStation B.04.03SP2 software. Control of TiM various heating and cooling zones was performed with Agilent TiM UI software version 1.0 (Agilent Technologies, Shanghai, China). Data processing was carried out using GC Image GC×GC software version 2.4B2 (GC Image, LLC, Lincoln, NE, U.S.A.) and was used for displaying separation results, providing peak statistics such as retention times and peak widths.

(TEC) offered a significant step toward a compact, coolant-free thermal modulation device.9−11 Due to thermal crosstalk between the two stages as well as problems of interfacing with conventional GC platforms, only a narrow volatility range from nC6 to nC10 has been attained with such a system.9−11 Here we utilize TEC and micathermic heating combined, but leverage conventional GC capillary columns and GC oven for GC×GC analysis of a series of complex samples.

■

EXPERIMENTAL SECTION This investigation used an Agilent 7890A gas chromatograph (Agilent Technologies, Wilmington, Delaware, U.S.A.) equipped with a split/splitless inlet operated in split mode, a flame ionization detector, an Agilent 7897B autosampler, and a two channel electronic pressure control module (PCM). A TiM, as described in U.S. patent #8277544, with a number of modifications made to it was constructed for this project. A simplified analytical system diagram is shown in Figure 1, while Figure 2 shows a diagram of key components of TiM.

Figure 1. Schematic diagram of the analytical system.

Figure 2. Diagram of key components of TiM.

The TiM was mounted on top of the GC oven above the vacant slot normally occupied by a back inlet. The TiM measures 190 mm × 120 mm × 70 mm. Installing the TiM over the back inlet does not interfere with the installation of the autosampler in the front inlet position. A standard 24 V DC and 180 W power supply supports full functioning of the TiM, which allows temperature programming from ambient to + 350 °C in the remobilizing zones, and from −50 °C to + 50 °C in the trapping zone. The TEC temperature above ambient is enabled by an implementation of automatic TEC polarity reversal. The device has a trapping zone in the middle (TEC) and two remobilizing zones on each side, namely the entry and exit B

DOI: 10.1021/acs.analchem.6b02525 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

annular space effectively eliminates ambient moisture from condensing in the trapping zone when the TECs operate below subambient temperature. A column guide is installed in each zone to prevent the polyimide coating of the fused silica modulation wearing as a result of contact with the side wall of the copper tubes. The column guide supports the column with an optimized micro bearing to keep the column straight, and maintains the column loop motion in the same plane. Mean time between modulation column failures with microbearings in place is on the order of one million modulations or equivalent to 2000 analyses for a typical diesel sample GC×GC analysis (500 modulations per analysis). To minimize thermal mass, 0.25 mm id fused silica tubing with 0.055 mm tubing wall thickness was used for the modulation column. Consequently, the most likely cause for system failure is the breakage of the modulation column. Failure is abrupt with loss of detector peak response and warning of loss of pressure and flow by the Pressure Control Module (PCM). Otherwise, TiM has been proven to be highly reliable with no unplanned event over the course of two years of continuous use. Preliminary evaluation of the modulator indicated much wider injection bands than those produced by commercial thermal modulators. This was attributed to two constraints of the design. First, modulation on a 0.25 mm i.d. column produces a much larger vapor volume upon remobilization than on narrower bore 2D columns as typically practiced; next, there is an extra length of modulation column past the second stage in the TiM before connection to 2D. This length may range on average 35 cm and contributes to additional band broadening. To achieve a better 2D injection performance, an auxiliary carrier gas flow was introduced to the inlet of the modulation column. When total flow through TiM is typically set to ∼3 mL/min, injection bands were significantly improved and measured at 20−25 ms at half height for n-alkanes up to C24 by direct connection of modulation column exit to FID. Figure 3 is a color plot of 1000 ppm (v/v) each of a mixture of nC8 to nC16 hydrocarbons, benzene, toluene, ethyl benzene,

Table 1. Gas Chromatographic Conditions for the Analysis of Hydrocarbons, Diesel, Dowtherm Q, and Pyrolysis Gasoline Gas Chromatographic Conditions columns injection parameters oven profile carrier gas detector modulation period driving zone (hot) cold zone releasing zone (hot)

30 m × 0.25 mm × 0.25 μm VF-1 ms (1D) 1.0 m × 0.18 mm × 0.18 μm HP-INNOWax (2D) 1.0 μL, split 200:1; 250 °C, liner (#5183−4647) 40 °C (3 min) to 250 °C (5 min) @ 5 °C/min He, 99,999%, 1 mL/min (1D), 3 mL/min (2D) FID TiM Conditions 4 s (release time of 2 s) 80 °C (3 min) to 280 °C (6 min) @ 5 °C/min −50 °C (15 min) to 30 °C (14 min) 160 °C (3 min) to 290 °C (13.5) @ 4 °C/min FID Conditions

H2 air N2 acquisition rate

35 mL/min 350 mL/min 25 mL/min 200 Hz

■

RESULTS AND DISCUSSION A two-stage thermal modulation is achieved, akin to Marriott and Kinghorn’s longitudinally modulated cryogenic system (LMCS), by a motion of the modulation column relative to the thermal zones [2−4]. Unlike the LMCS, TiM is unique in, first bringing the modulator outside of the oven so that the GC oven heating is absent from and will not interfere with cooling of either stages, and second, choosing a moving design to shuttle back and forth between thermally isolated cooling region (TEC) and heating region (heated copper tube) of TiM. The absence of thermal interferences, hence “thermal independence” between heating and cooling stage is a critical novelty of TiM. When the modulation column is driven toward the entry zone, a segment of column (first stage) previously inside the trapping zone is moved to the copper tube on the entry side where it is heated to remobilize trapped analytes; in the meantime another segment of column (second stage) previously inside of the copper tube on the exit zone side is moved into the trapping zone where it is cooled to retrap the analytes that have been remobilized from the first stage. When the modulation column is driven back toward the exit zone side, the first stage re-enters the trapping zone to capture newly incoming analytes, while the second stage enters the copper tube on the exit zone side to remobilize trapped analytes to 2D. Since the first stage remobilization and the second stage trapping are concurrent, there is a possibility for analyte breakthrough(s) to occur. To reduce/eliminate this possibility, the entry zone side temperature is set lower than the exit zone side and a slightly larger inner diameter copper tube (0.5 mm i.d.) was employed in the driving side to purposely cause a slower heat transfer rate than the release side (0.4 mm i.d.). Since the trapping zone periodically operates in subambient temperature, moisture condensed from ambient air can result in formation of ice in the trap. This can have negative impact on the performance of TiM and was alleviated by using 3−5 mL/ min of dry nitrogen as purge gas. The nitrogen purge gas is continuously introduced into the trapping zone through a line welded on the enclosure wall and is expelled through the narrow annular space between the copper tubes and the modulation column; a high linear gas velocity produced in the

Figure 3. Color plot of 1000 ppm (v/v) each of a mixture of nC8 to nC16 hydrocarbons, benzene, toluene, ethyl benzene, and xylenes, and 250 ppm (v/v) each of hydrocarbons from nC18 to nC22 in cyclohexane.

and xylenes, and 250 ppm (v/v) each of hydrocarbons from nC18 to nC22 in cyclohexane using conditions reported in Table 1. Excellent chromatographic performance was attained with peak width 1/2 heights for aromatic compounds of less than 120 ms under the conditions used. Repeatability of retention time for the target compounds were found to be highly C

DOI: 10.1021/acs.analchem.6b02525 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry satisfactory with relative standard deviations of less than 0.009% in 1D and less than 0.008% in 2D (n = 10). Typical peak widths of 120 ms or less with a relative standard deviation of less than 4.7% were achieved for the aromatic model compounds. The loadability of the modulator for light components such as hexane was studied using hexane standards in n-decane over a range from 100 ppm (w/w) to 1% (w/w). The results obtained show no significant loading issue was encountered up to 1% (w/w) of hexane at a flow of 1.85 mL/min through TiM (Supporting Information, Supporting material 1). Figure 4 shows a color plot of Canadian Winter Diesel Fuel using TiM under the conditions stated in Table 1. Applicability

inlet pressure to 2 psig. Back-flushing can substantially decrease analytical time as well as improves chromatographic system cleanliness.12−15 Retention time locking capability is demonstrated on 1D with an overlay of color plots of the same sample, with the carrier gas flow of 1D kept at 1.4 mL/min He and 2D flow varied from 1.4 to 5.0 mL/min He, as shown in Figure 6. With independent

Figure 4. Color plot of Canadian Winter Diesel Fuel.

Figure 6. Overlay of color plots of Canadian Winter Diesel with 1D flow at 1.4 m/min He and 2D flow at 1.4 and 5.0 mL/min He, respectively.

of the TiM approach for a volatility range spanning nC6 to nC24 is clearly demonstrated. Excellent chromatographic resolution was achieved in 2D with peak width at 1/2 height of less than 120 ms for the one-ring, two-ring, and three-ring aromatic compounds. With the advent of an auxiliary gas flow, independent flow control of 1D and 2D was achieved. The possibility of independent flow control offers additional options such enabling back-flushing as well as independent retention time locking. Figure 5 shows an overlay of color plots of the same Canadian Winter Diesel Fuel sample depicted in Figure 4, but back-flushed after 15 and 30 min respectively by dropping 1D

flow control and retention time locking (exact locking of 1D would require a slight change of 1D flow), resolution tuning or control of wrap around in 2D can be realized without impacting the already optimized 1D separation. It is a more versatile approach than a secondary oven solely to achieve similar effects. Once optimized and its reliability demonstrated, a number of industrial applications were successfully used to demonstrate the TiM performance. In the process of producing ethylene, pyrolysis gasoline is also generated.16 It is produced in the pyrolysis (high temperature) furnaces in steam cracker operations.17 Pyrolysis gasoline is a commodity of industrial significance with a very complex mixture of having more than 200 individual components and isomers of alkanes, branched alkanes, alkenes, conjugated alkenes, and aromatic compounds up with volatility up to nC12. With the high resolution achieved in the second dimension, aromatic and oxygenated compounds in pyrolysis gasoline produced with different feedstock, under different process conditions, and at different geographic locations can be accurately monitored and compared with TiM. A color plot of a commercial pyrolysis gasoline sample is shown in Supporting Information, Supporting material 2. Dowtherm Q is a highly effective commercially available aromatic based heat exchanger fluid. Dowtherm Q is comprised of a mixture of diphenyl ethanes and alkyl aromatics.18 A heat exchanger leak can lead to the presence of Dowtherm Q in the aliphatic solvent used in the reaction process with a potential of contaminating the final products produced. The high resolution attained by GC × GC using TiM allows Dowtherm Q to be accurately quantified. Supporting Information, Supporting material 3 shows a color plot of a standard of 50 ppm (v/v) of Dowtherm Q in an aliphatic recycled solvent.

Figure 5. Overlay of color plots of Canadian Winter Diesel Fuel backflushed after 15 and 30 min, respectively. D

DOI: 10.1021/acs.analchem.6b02525 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry Unlike in flow modulation, and with typical D2 column flow of no more than 3 mL/min, TiM is fully compatible with benchtop mass spectrometer without the need for effluent splitting.

(5) Haglund, P.; Harju, M.; Daneilsson, C.; Marriott, P. J. J. Chromatogr. A 2002, 962, 127−134. (6) Gaines, R. B.; Frysinger, G. S. J. Sep. Sci. 2004, 27, 380−388. (7) Rathbun, W. J. Chromatogr. Sci. 2007, 45, 636−642. (8) Muscalu, A. M.; Edwards, M.; Gorecki, T.; Reiner, E. J. Chromatogr. A 2015, 1391, 93−101. (9) Kim, S. J.; Reidy, S. M.; Block, B. P.; Wise, K. D.; Zellers, E. T.; Kurabayashi, K. Lab Chip 2010, 10, 1647−1654. (10) Serrano, G.; Paul, D.; Kim, S. J.; Kurabayashi, K.; Zellers, E. T. Anal. Chem. 2012, 84, 6973−6980. (11) Collin, W.; Nunovero, N.; Dibyadeep, P.; Kurabyashi, K.; Zellers, E. T. J.Chromatogr. A 2016, 1444, 114−122. (12) Deans, D. R. Chromatographia 1968, 1, 18−22. (13) Seeley, J.; Seeley, S. Anal. Chem. 2013, 85, 557−578. (14) Luong, J.; Gras, R.; Cortes, H. J.; Shellie, R. A. J. Chromatogr. A 2012, 1255, 216−220. (15) Luong, J.; Gras, R.; Cortes, H. J.; Shellie, R. A. J. Chromatogr. A 2013, 1271, 185−191. (16) Dow Chemical Material Safety Data Sheet, Pyrolysis Gasoline, publication number 50515/1001, Version 3.0, April 29, 2008. (17) Ali, J. The hydrogenation of pyrolysis gasoline over nickel and palladium catalysts. Ph.D. Thesis, University of Glasgow, Scotland, United Kingdom 2012. (18) Dow Chemical Material Safety Data Sheet, Dowtherm Q, publication number 00257, October, 2004.

■

CONCLUSIONS TiM represents a thermal independence strategy that effectively separates the heating of a GC oven from its modulator and eliminates thermal crosstalk between interior thermal zones. Minimal amount of consumables and no coolants are required to operate the TiM device. The device was found to be suitable over a relatively wide volatility range, it has a compact footprint and low power consumption. With the advent of an auxiliary gas flow, independent flow control of 1D and 2D was achieved. The possibility of independent flow control offers additional degrees of freedom such as improved band injection, more effective sweeping of downstream void volume, enabling backflushing as well as independent retention time locking or scaling in either dimension. Difficult industrial applications were successfully used to demonstrate the unique feature set of the device.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.6b02525. Supporting figures (PDF).

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: 780-998-8668. Fax: 780998-6709. Present Address

∥ J&X Technologies, 1599 Jungong Road, Yanpu District, 200438 Shanghai, China; E-mail: [email protected].

Author Contributions

The manuscript was written through contributions of all authors. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to express their deepest gratitude to Dr. Qiang Xu, Dr. Kai Chen, and Dr. Roger Firor of Agilent Technologies for their invaluable contributions. Dr. Monty Benefiel, Mrs. Shanya Kane, and Dr. Cynthia Cai of Agilent Technologies are acknowledged for their unconditional support and encouragement. Last but not least, we thank Dr. Martine Stolk, Dr. Judy Gunderson, and Mr. Mike de Poortere of the Dow Chemical Company for their support. This project was partially funded by Dow Chemical Analytical Technology Center 2014, 2015, and 2016 Technology Renewal and Development Program.

■

REFERENCES

(1) Liu, Z.; Phillips, J. J. Chromatogr. Sci. 1991, 29, 227−231. (2) Marriott, P. J.; Kinghorn, R. M. Anal. Chem. 1997, 69, 2582− 2588. (3) Marriott, P. J.; Shellie, R. A. TrAC, Trends Anal. Chem. 2002, 21, 573−583. (4) Marriott, P. J.; Ong, R.; Shellie, R. A. Am. Lab. 2001, 17, 44−46. E

DOI: 10.1021/acs.analchem.6b02525 Anal. Chem. XXXX, XXX, XXX−XXX