3649
J. Sep. Sci. 2014, 37, 3649–3655
Nicole E. Heshka ∗ Darcy B. Hager CanmetENERGY, Natural Resources Canada, Devon, Alberta, Canada Received July 7, 2014 Revised August 20, 2014 Accepted October 3, 2014
Research Article
A multidimensional gas chromatography method for the analysis of hydrogen sulfide in crude oil and crude oil headspace Two-dimensional heart-cutting gas chromatography is used to analyze dissolved hydrogen sulfide in crude samples. Liquid samples are separated first on an HP-PONA column, and the light sulfur gases are heart-cut to a GasPro column, where hydrogen sulfide is separated from other light sulfur gases and detected with a sulfur chemiluminescence detector. Heartcutting is accomplished with the use of a Deans switch. Backflushing the columns after hydrogen sulfide detection eliminates any problems caused by high-boiling hydrocarbons in the samples. Dissolved hydrogen sulfide is quantified in 14 crude oil samples, and the results are shown in this work. The method is also applicable to the analysis of headspace hydrogen sulfide over crude oil samples. Gas hydrogen sulfide measurements are compared to liquid hydrogen sulfide measurements for the same sample set. The chromatographic system design is discussed, and chromatograms of representative gas and liquid measurements are shown. Keywords: Crude oil / Deans switch / GasPro / Hydrogen sulfide DOI 10.1002/jssc.201400727
1 Introduction GC has long been a staple in the field of petroleum analysis. Separations generally occur by boiling point, allowing users to perform a range of experiments from standard analyses, such as simulated distillation to complex multidimensional experiments that can fingerprint samples, making it possible to visualize trends in structure, heteroatom content, and carbon number. GC is also ideal for the analysis of dissolved gases present in crude oil, as gases are liberated from the liquid phase in the hot inlet and analyzed on the column set by the user desires. There are several excellent reviews that describe recent advances in GC applied to crude oil and petroleum-based samples [1–4]. The present paper focuses on the analysis of dissolved hydrogen sulfide (H2 S) in crude oil. H2 S is an important gas for the oil and gas industry, as it causes problems at many stages of refining and upgrading. H2 S gas is highly toxic, and can be fatal at concentrations as low as 100 ppm (http://www.cdc.gov/niosh/ idlh/7783064.html) [5, 6]. During upgrading, H2 S can preferentially react with catalysts, inhibiting the removal of sulfur compounds and the hydrogenation of others [7–9]. H2 S can also act as a corrosive agent, reacting with metals in storage containers or transport vessels [10, 11]. A common method Correspondence: Dr. Nicole E. Heshka, CanmetENERGY, Natural Resources Canada, 1 Oil Patch Drive, Devon, Alberta T9G 1A8, Canada E-mail: [email protected] Fax: +1 780 987 8676
Abbreviations: DEA, diethanolamine; FID, flame ionization detection; SCD, sulfur chemiluminescence detection C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
of dealing with H2 S is to add scavengers to crude or gas streams. These substances act to prevent H2 S gas phase release, but can cause additional problems with fouling and corrosion [12–17]. For these reasons, it is desirable to quantify dissolved H2 S in crude samples to assess potential for gas release, as well as to evaluate the success of potential removal methods. There are many standard methods for measurement of H2 S in petroleum or fuel-based samples, but none are designed for use with the heavy crudes commonly extracted from the Athabasca oil sands. Universal Oil Products method 163 utilizes titration for H2 S and mercaptan sulfur determination, with fuels and light distillates included in its scope [18]. The method is simple, but is not designed for heavy samples; moreover, the results can be user-dependent, since interpretation of titration curves is necessary. The standard method IP (Institute of Petroleum) 570 was developed using Seta Analytics’ H2 S analyzer (http://www.seta-analytics.com/documents/H2SCrude-Oil.pdf), and measures the H2 S released from a fuel oil sample with heating [19]. The H2 S analyzer is portable, and produces results that are free from userinterpretation bias. However, work with the analyzer on heavier crude samples has shown that it fails to consistently produce accurate results [20]. The addition of the vaporphase processor to this unit was designed to remedy this problem (http://www.seta-analytics.com/documents/Seta_ H2S_analyser_VPP.pdf), but it is not yet shown that the method can be successfully adapted to heavy crude samples. ∗ The author’s former name was Nicole E. Oro. Colour Online: See the article online to view Figs. 1 and 3 in colour.
www.jss-journal.com
3650
J. Sep. Sci. 2014, 37, 3649–3655
N. E. Heshka and D. B. Hager
The standard method most closely related to the work presented here is ASTM D5623 [21, 22]. This method, which is designed for use on light petroleum liquids, utilizes GC with sulfur selective detection. Cryogenic cooling is required, with a total run time of nearly 30 min. Clearly, an ambient separation with shorter analysis time would be preferable to this method. Di Sanzo et al. demonstrated separation and quantification of H2 S from other sulfur compounds in gasoline. However, this method is similar to ASTM D5623 in that it requires subambient temperatures and has not been shown to work with heavier crudes [23]. A recent communication by Luong et al. describes a configuration similar to this work, where Deans switching and DB-1HT and VF-1 GC columns are used to characterize multiple sulfur compounds in light and middle distillates [24]. This method has the advantage of characterizing an extended range of sulfur components, including H2 S, but is somewhat more complicated, with two planar microfluidic devices to control and program rather than the single device used in the present work. In this work, a 2D GC method with Deans switching was developed to selectively analyze H2 S in a heavy crude sample. The Deans switch, which has been in use since 1968, has one inlet port and two outlet ports. The switching mechanism operates by using an external valve to direct gas flow to one of two outlets attached to the switch [25–27]. In this way, effluent coming from the inlet port of the Deans switch can be pushed in the desired direction. Deans switches have been used in both heart-cutting and comprehensive 2D GC separations, with applications in petroleum analysis, food chemistry, environmental samples, fragrances, and medicine [28–32]. The benefit of using a Deans switch is the ability to selectively transfer compounds to a second dimension column that has a stationary phase suitable for precise separation. The GasPro GC column is a proprietary phase that effectively separates H2 S from other light gases (http://www.chem.agilent.com/cag/cabu/pdf/gaspro.pdf) [33–35], therefore, this stationary phase was selected as our second dimension column. To alleviate concerns about directly injecting long-chain hydrocarbons onto the GasPro column [35], an HP-PONA column was used to separate light sulfur gases from the hydrocarbons in the sample, and the Deans switch was used to transfer these gases to the GasPro column. Backflushing has been successfully used to minimize column contamination in complex sample matrices [32, 36], and was used in this work to protect the GasPro column from heavy hydrocarbons. The 2D method, with backflushing, allows for the analysis of H2 S-containing heavy crude with minimal effect on the analytical columns used. This paper describes the two-column GC method developed for the analysis of H2 S in both liquid crude oil samples and crude oil headspace. Results are presented for the analysis of 14 crude oil samples, with discussion focused on the configuration of the gas chromatograph and the separation capabilities of the method. To the best of our knowledge, this is the first report of a GC method capable of measuring H2 S in heavy crude oil samples at above-ambient temperatures. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2 Materials and methods 2.1 GC An Agilent 7890 gas chromatograph, equipped with a split– splitless inlet, a Deans switch with pressure control module (Agilent part number G2855A), a Sievers 355 sulfur chemiluminescence detector (SCD) and a flame ionization detector (FID; Agilent, Mississauga, ON, Canada) was utilized in this work. ChemStation version B.04.01 was used to control the GC system and analyze chromatograms.
2.1.1 Liquid measurements A two-column GC configuration with post run backflushing was used. The first column was an HP-PONA column (Agilent; 50 m × 0.200 mm id, 0.50 m film). This column was connected to the split-splitless inlet, and to the Deans switch on the outlet of the column. The second column was a GasPro PLOT column (Agilent; 30 m × 0.320 mm id). The inlet of this column was connected to the Deans switch, and the outlet was connected to the SCD. A fused-silica transfer line (1.73 m × 180 m, Agilent) connected the third Deans switch outlet to the FID. Hydrogen was employed as the carrier gas. Liquid measurements on the SCD were calibrated using a 100 ppm CS2 calibration standard. CS2 (ACS grade, Fisher Chemicals, Fairlawn, NJ, USA) was diluted with toluene (HPLC grade, Fisher Chemicals) to prepare the calibration solutions. These solutions were made on the day of use in four separate vials. Calibration was performed using four data points, with one injection taken from each vial. This procedure was used to prevent concentration changes due to evaporative losses between successive injections from the same vial. H2 S peak identity was confirmed by spiking liquid samples with H2 S gas (2.5% in He, Air Liquide, Edmonton, AB, Canada) and observing retention time of the resulting peak. Injections were performed using an Agilent 7683B liquid automatic sampler, with three replicates of each sample. Liquid samples analyzed were neat, except in the case of samples C, E, and F, which were too viscous to be pipetted quantitatively. These three samples were diluted 1:1 with toluene by weight to facilitate successful injection by the liquid autosampler. For analysis, the GC oven was held at 50⬚C for 2 min, then ramped at 100⬚C/min to 250⬚C, and held at 250⬚C for 1 min. The inlet was operated in split mode at 250⬚C, 40 psi, and a split ratio of 10:1. Flow through the HP-PONA column was configured at 2.7 mL/min with an average velocity of 57.0 cm/s. The GasPro column was configured with a flow of 2.99 mL/min and an average velocity of 53.7 cm/s. The pressure control module controlling the Deans switch was held at 6.89 psi during analysis, and the switch was “off” from 0.7 to 2.3 min. The liquid automatic sampler was used to inject 1 L of sample with a 6 s viscosity delay, with toluene serving as a wash solvent to clean the syringe. www.jss-journal.com
J. Sep. Sci. 2014, 37, 3649–3655
Following analysis of H2 S, a 16 min post-run backflush was performed. The oven was held at 250⬚C during this process. During backflush, the inlet pressure was reduced to 1 psi and the pressure control module attached to the Deans switch was increased to 40 psi. 2.1.2 Gas measurements The GC configuration for gas measurements was the same as for the liquid measurements. The split ratio for injection was changed to 40:1 for gases. Manual injections were performed with a 250 L gas-tight syringe (Hamilton Company, Reno, NV, USA), with 250 L of gas injected for both calibration and sample analysis. The GC was calibrated on a daily basis using a certified gas standard (Air Liquide, Edmonton, AB, Canada) containing 206 ppm H2 S, 99.8 ppm methyl mercaptan, and 498 ppm dimethyl sulfide. The SCD response was calibrated against the H2 S peak that was identified using retention time. Gas measurements were taken from the headspace of 500 mL glass bottles with septum tops (Chromatographic Specialties, Brockville, ON, Canada) after 24 h at a set temperature. Bottles were agitated hourly to ensure full equilibrium and H2 S release to gas phase. Backflushing was omitted for gas analysis.
2.2 Samples Fourteen different crude oil samples from a variety of sources were used to test this method. The crude samples were obtained from four different sources and span a range of densities and sulfur contents. The samples are labeled alphabetically, with ascending letters corresponding to increasing liquid phase H2 S content. Samples are not identified by location or supplier. Total sulfur content was determined by ASTM method D4294, while density was determined using ASTM method D4052.
3 Results and discussion 3.1 Chromatographic system design To analyze H2 S in crude, a 2D method was used. The use of the Deans switch allows for the diversion of flow between two different outlets, and the direction of flow is changed by switching an externally controlled valve. Samples are injected into the hot inlet, and in the case of liquid samples, vaporized and carried to the HP-PONA column. The initial separation occurring on the HP-PONA column allows light gases, including H2 S, to pass through the column essentially unretained. At the point when the light gases exit the column, the Deans switch is moved to the “off” position for a particular period of time, transferring a sample plug to the GasPro column for separation; this is commonly referred to as heart-cutting [25]. A wide heart cut was selected as smaller C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Gas Chromatography
3651
cut windows did not improve resolution, and a longer window allowed for additional H2 S identity confirmation based on retention of other sulfur gases found in the calibration mixture. Hydrocarbons remain on the HP-PONA column, with lower-boiling components eluting during the H2 S separation, detected by the FID. The oven temperature was held at 50⬚C at the onset of analysis to retard the movement of hydrocarbons through the HP-PONA column. Following the transfer of H2 S onto the GasPro column, the oven temperature was increased to 250⬚C, resulting in shorter separation times and a narrower H2 S peak shape. Once on the GasPro column, H2 S is well separated from other light sulfur-containing gases. As the SCD detects only compounds that contain sulfur, the chromatogram from the SCD does not experience any interference from hydrocarbons that may be present in the sample band transferred from the HP-PONA column [37,38]. The light sulfur gases elute off the GasPro and are detected by the SCD while the Deans switch is returned to the “on” position. This step protects the GasPro column from the introduction of any heavy hydrocarbons found in the crude samples. After H2 S separation and detection has been achieved, the pressure on the inlet is dropped and the carrier gas flow through the Deans switch is increased. This creates a backflush of any compounds still on the HP-PONA column out through the inlet. When coupled with an elevated oven temperature, this effectively burns off higher boiling crude components. No baseline problems were encountered after batches of repeat analyses, indicating that removal of uneluted components was complete. The backflushing process requires that the inlet liner be cleaned and/or replaced after approximately 50 injections, a procedure that is preferable to cutting or replacing either GC column. Analysis was performed in this way over five months, and daily blanks confirmed that no carryover or column bleed was occurring. Over 125 liquid samples were analyzed with no observable degradation or change in chromatographic performance or retention, demonstrating that heavy sample build-up is not a significant problem for this method. The use of two columns and the Deans switch requires more method optimization than a one-column GC method. However, the benefits of using this configuration outweigh the extra time and equipment investment. Measurement of H2 S that is free from both sulfur and hydrocarbon interferences is possible over a wide range of concentrations, in both liquid (1–500 ppm H2 S) and gas samples (0.7–9700 ppm H2 S). Analysis time is short (
J. Sep. Sci. 2014, 37, 3649–3655
Nicole E. Heshka ∗ Darcy B. Hager CanmetENERGY, Natural Resources Canada, Devon, Alberta, Canada Received July 7, 2014 Revised August 20, 2014 Accepted October 3, 2014
Research Article
A multidimensional gas chromatography method for the analysis of hydrogen sulfide in crude oil and crude oil headspace Two-dimensional heart-cutting gas chromatography is used to analyze dissolved hydrogen sulfide in crude samples. Liquid samples are separated first on an HP-PONA column, and the light sulfur gases are heart-cut to a GasPro column, where hydrogen sulfide is separated from other light sulfur gases and detected with a sulfur chemiluminescence detector. Heartcutting is accomplished with the use of a Deans switch. Backflushing the columns after hydrogen sulfide detection eliminates any problems caused by high-boiling hydrocarbons in the samples. Dissolved hydrogen sulfide is quantified in 14 crude oil samples, and the results are shown in this work. The method is also applicable to the analysis of headspace hydrogen sulfide over crude oil samples. Gas hydrogen sulfide measurements are compared to liquid hydrogen sulfide measurements for the same sample set. The chromatographic system design is discussed, and chromatograms of representative gas and liquid measurements are shown. Keywords: Crude oil / Deans switch / GasPro / Hydrogen sulfide DOI 10.1002/jssc.201400727
1 Introduction GC has long been a staple in the field of petroleum analysis. Separations generally occur by boiling point, allowing users to perform a range of experiments from standard analyses, such as simulated distillation to complex multidimensional experiments that can fingerprint samples, making it possible to visualize trends in structure, heteroatom content, and carbon number. GC is also ideal for the analysis of dissolved gases present in crude oil, as gases are liberated from the liquid phase in the hot inlet and analyzed on the column set by the user desires. There are several excellent reviews that describe recent advances in GC applied to crude oil and petroleum-based samples [1–4]. The present paper focuses on the analysis of dissolved hydrogen sulfide (H2 S) in crude oil. H2 S is an important gas for the oil and gas industry, as it causes problems at many stages of refining and upgrading. H2 S gas is highly toxic, and can be fatal at concentrations as low as 100 ppm (http://www.cdc.gov/niosh/ idlh/7783064.html) [5, 6]. During upgrading, H2 S can preferentially react with catalysts, inhibiting the removal of sulfur compounds and the hydrogenation of others [7–9]. H2 S can also act as a corrosive agent, reacting with metals in storage containers or transport vessels [10, 11]. A common method Correspondence: Dr. Nicole E. Heshka, CanmetENERGY, Natural Resources Canada, 1 Oil Patch Drive, Devon, Alberta T9G 1A8, Canada E-mail: [email protected] Fax: +1 780 987 8676
Abbreviations: DEA, diethanolamine; FID, flame ionization detection; SCD, sulfur chemiluminescence detection C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
of dealing with H2 S is to add scavengers to crude or gas streams. These substances act to prevent H2 S gas phase release, but can cause additional problems with fouling and corrosion [12–17]. For these reasons, it is desirable to quantify dissolved H2 S in crude samples to assess potential for gas release, as well as to evaluate the success of potential removal methods. There are many standard methods for measurement of H2 S in petroleum or fuel-based samples, but none are designed for use with the heavy crudes commonly extracted from the Athabasca oil sands. Universal Oil Products method 163 utilizes titration for H2 S and mercaptan sulfur determination, with fuels and light distillates included in its scope [18]. The method is simple, but is not designed for heavy samples; moreover, the results can be user-dependent, since interpretation of titration curves is necessary. The standard method IP (Institute of Petroleum) 570 was developed using Seta Analytics’ H2 S analyzer (http://www.seta-analytics.com/documents/H2SCrude-Oil.pdf), and measures the H2 S released from a fuel oil sample with heating [19]. The H2 S analyzer is portable, and produces results that are free from userinterpretation bias. However, work with the analyzer on heavier crude samples has shown that it fails to consistently produce accurate results [20]. The addition of the vaporphase processor to this unit was designed to remedy this problem (http://www.seta-analytics.com/documents/Seta_ H2S_analyser_VPP.pdf), but it is not yet shown that the method can be successfully adapted to heavy crude samples. ∗ The author’s former name was Nicole E. Oro. Colour Online: See the article online to view Figs. 1 and 3 in colour.
www.jss-journal.com
3650
J. Sep. Sci. 2014, 37, 3649–3655
N. E. Heshka and D. B. Hager
The standard method most closely related to the work presented here is ASTM D5623 [21, 22]. This method, which is designed for use on light petroleum liquids, utilizes GC with sulfur selective detection. Cryogenic cooling is required, with a total run time of nearly 30 min. Clearly, an ambient separation with shorter analysis time would be preferable to this method. Di Sanzo et al. demonstrated separation and quantification of H2 S from other sulfur compounds in gasoline. However, this method is similar to ASTM D5623 in that it requires subambient temperatures and has not been shown to work with heavier crudes [23]. A recent communication by Luong et al. describes a configuration similar to this work, where Deans switching and DB-1HT and VF-1 GC columns are used to characterize multiple sulfur compounds in light and middle distillates [24]. This method has the advantage of characterizing an extended range of sulfur components, including H2 S, but is somewhat more complicated, with two planar microfluidic devices to control and program rather than the single device used in the present work. In this work, a 2D GC method with Deans switching was developed to selectively analyze H2 S in a heavy crude sample. The Deans switch, which has been in use since 1968, has one inlet port and two outlet ports. The switching mechanism operates by using an external valve to direct gas flow to one of two outlets attached to the switch [25–27]. In this way, effluent coming from the inlet port of the Deans switch can be pushed in the desired direction. Deans switches have been used in both heart-cutting and comprehensive 2D GC separations, with applications in petroleum analysis, food chemistry, environmental samples, fragrances, and medicine [28–32]. The benefit of using a Deans switch is the ability to selectively transfer compounds to a second dimension column that has a stationary phase suitable for precise separation. The GasPro GC column is a proprietary phase that effectively separates H2 S from other light gases (http://www.chem.agilent.com/cag/cabu/pdf/gaspro.pdf) [33–35], therefore, this stationary phase was selected as our second dimension column. To alleviate concerns about directly injecting long-chain hydrocarbons onto the GasPro column [35], an HP-PONA column was used to separate light sulfur gases from the hydrocarbons in the sample, and the Deans switch was used to transfer these gases to the GasPro column. Backflushing has been successfully used to minimize column contamination in complex sample matrices [32, 36], and was used in this work to protect the GasPro column from heavy hydrocarbons. The 2D method, with backflushing, allows for the analysis of H2 S-containing heavy crude with minimal effect on the analytical columns used. This paper describes the two-column GC method developed for the analysis of H2 S in both liquid crude oil samples and crude oil headspace. Results are presented for the analysis of 14 crude oil samples, with discussion focused on the configuration of the gas chromatograph and the separation capabilities of the method. To the best of our knowledge, this is the first report of a GC method capable of measuring H2 S in heavy crude oil samples at above-ambient temperatures. C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2 Materials and methods 2.1 GC An Agilent 7890 gas chromatograph, equipped with a split– splitless inlet, a Deans switch with pressure control module (Agilent part number G2855A), a Sievers 355 sulfur chemiluminescence detector (SCD) and a flame ionization detector (FID; Agilent, Mississauga, ON, Canada) was utilized in this work. ChemStation version B.04.01 was used to control the GC system and analyze chromatograms.
2.1.1 Liquid measurements A two-column GC configuration with post run backflushing was used. The first column was an HP-PONA column (Agilent; 50 m × 0.200 mm id, 0.50 m film). This column was connected to the split-splitless inlet, and to the Deans switch on the outlet of the column. The second column was a GasPro PLOT column (Agilent; 30 m × 0.320 mm id). The inlet of this column was connected to the Deans switch, and the outlet was connected to the SCD. A fused-silica transfer line (1.73 m × 180 m, Agilent) connected the third Deans switch outlet to the FID. Hydrogen was employed as the carrier gas. Liquid measurements on the SCD were calibrated using a 100 ppm CS2 calibration standard. CS2 (ACS grade, Fisher Chemicals, Fairlawn, NJ, USA) was diluted with toluene (HPLC grade, Fisher Chemicals) to prepare the calibration solutions. These solutions were made on the day of use in four separate vials. Calibration was performed using four data points, with one injection taken from each vial. This procedure was used to prevent concentration changes due to evaporative losses between successive injections from the same vial. H2 S peak identity was confirmed by spiking liquid samples with H2 S gas (2.5% in He, Air Liquide, Edmonton, AB, Canada) and observing retention time of the resulting peak. Injections were performed using an Agilent 7683B liquid automatic sampler, with three replicates of each sample. Liquid samples analyzed were neat, except in the case of samples C, E, and F, which were too viscous to be pipetted quantitatively. These three samples were diluted 1:1 with toluene by weight to facilitate successful injection by the liquid autosampler. For analysis, the GC oven was held at 50⬚C for 2 min, then ramped at 100⬚C/min to 250⬚C, and held at 250⬚C for 1 min. The inlet was operated in split mode at 250⬚C, 40 psi, and a split ratio of 10:1. Flow through the HP-PONA column was configured at 2.7 mL/min with an average velocity of 57.0 cm/s. The GasPro column was configured with a flow of 2.99 mL/min and an average velocity of 53.7 cm/s. The pressure control module controlling the Deans switch was held at 6.89 psi during analysis, and the switch was “off” from 0.7 to 2.3 min. The liquid automatic sampler was used to inject 1 L of sample with a 6 s viscosity delay, with toluene serving as a wash solvent to clean the syringe. www.jss-journal.com
J. Sep. Sci. 2014, 37, 3649–3655
Following analysis of H2 S, a 16 min post-run backflush was performed. The oven was held at 250⬚C during this process. During backflush, the inlet pressure was reduced to 1 psi and the pressure control module attached to the Deans switch was increased to 40 psi. 2.1.2 Gas measurements The GC configuration for gas measurements was the same as for the liquid measurements. The split ratio for injection was changed to 40:1 for gases. Manual injections were performed with a 250 L gas-tight syringe (Hamilton Company, Reno, NV, USA), with 250 L of gas injected for both calibration and sample analysis. The GC was calibrated on a daily basis using a certified gas standard (Air Liquide, Edmonton, AB, Canada) containing 206 ppm H2 S, 99.8 ppm methyl mercaptan, and 498 ppm dimethyl sulfide. The SCD response was calibrated against the H2 S peak that was identified using retention time. Gas measurements were taken from the headspace of 500 mL glass bottles with septum tops (Chromatographic Specialties, Brockville, ON, Canada) after 24 h at a set temperature. Bottles were agitated hourly to ensure full equilibrium and H2 S release to gas phase. Backflushing was omitted for gas analysis.
2.2 Samples Fourteen different crude oil samples from a variety of sources were used to test this method. The crude samples were obtained from four different sources and span a range of densities and sulfur contents. The samples are labeled alphabetically, with ascending letters corresponding to increasing liquid phase H2 S content. Samples are not identified by location or supplier. Total sulfur content was determined by ASTM method D4294, while density was determined using ASTM method D4052.
3 Results and discussion 3.1 Chromatographic system design To analyze H2 S in crude, a 2D method was used. The use of the Deans switch allows for the diversion of flow between two different outlets, and the direction of flow is changed by switching an externally controlled valve. Samples are injected into the hot inlet, and in the case of liquid samples, vaporized and carried to the HP-PONA column. The initial separation occurring on the HP-PONA column allows light gases, including H2 S, to pass through the column essentially unretained. At the point when the light gases exit the column, the Deans switch is moved to the “off” position for a particular period of time, transferring a sample plug to the GasPro column for separation; this is commonly referred to as heart-cutting [25]. A wide heart cut was selected as smaller C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Gas Chromatography
3651
cut windows did not improve resolution, and a longer window allowed for additional H2 S identity confirmation based on retention of other sulfur gases found in the calibration mixture. Hydrocarbons remain on the HP-PONA column, with lower-boiling components eluting during the H2 S separation, detected by the FID. The oven temperature was held at 50⬚C at the onset of analysis to retard the movement of hydrocarbons through the HP-PONA column. Following the transfer of H2 S onto the GasPro column, the oven temperature was increased to 250⬚C, resulting in shorter separation times and a narrower H2 S peak shape. Once on the GasPro column, H2 S is well separated from other light sulfur-containing gases. As the SCD detects only compounds that contain sulfur, the chromatogram from the SCD does not experience any interference from hydrocarbons that may be present in the sample band transferred from the HP-PONA column [37,38]. The light sulfur gases elute off the GasPro and are detected by the SCD while the Deans switch is returned to the “on” position. This step protects the GasPro column from the introduction of any heavy hydrocarbons found in the crude samples. After H2 S separation and detection has been achieved, the pressure on the inlet is dropped and the carrier gas flow through the Deans switch is increased. This creates a backflush of any compounds still on the HP-PONA column out through the inlet. When coupled with an elevated oven temperature, this effectively burns off higher boiling crude components. No baseline problems were encountered after batches of repeat analyses, indicating that removal of uneluted components was complete. The backflushing process requires that the inlet liner be cleaned and/or replaced after approximately 50 injections, a procedure that is preferable to cutting or replacing either GC column. Analysis was performed in this way over five months, and daily blanks confirmed that no carryover or column bleed was occurring. Over 125 liquid samples were analyzed with no observable degradation or change in chromatographic performance or retention, demonstrating that heavy sample build-up is not a significant problem for this method. The use of two columns and the Deans switch requires more method optimization than a one-column GC method. However, the benefits of using this configuration outweigh the extra time and equipment investment. Measurement of H2 S that is free from both sulfur and hydrocarbon interferences is possible over a wide range of concentrations, in both liquid (1–500 ppm H2 S) and gas samples (0.7–9700 ppm H2 S). Analysis time is short (
