NDUV, H2S Measurement, Frontal Elution Chromatography, Auto-carrier, online analysis, emissions monitoring, AFEC, AFECUV, Process UV Spectrometer.
Ultraviolet (UV) absorption spectroscopy provides rapid real-time analysis of many species of interest to industrial monitoring. In some applications, the presence of other interfering species obfuscates the measurement of interest and makes accurate and reliable analysis of trace components is difficult. The use of auto-carrier, frontal elution chromatography in conjunction with a process UV spectrometer allows for rapid online analysis of trace components even in complex streams containing interfering species. A unique implementation of this methodology is described and a thorough explanation of auto-carrier frontal elution chromatography is provided, along with data showing the
efficacy of the method combined with spectroscopic measurements in real-world applications.
Ultraviolet absorption spectroscopy is often used in process control and emissions measurement applications. Various compounds of interest have strong absorption bands in the UV, including sulfur compounds, oxides of nitrogen (NOx), aromatic hydrocarbons, ammonia, mercury, ozone, halogens, uranium hexafluoride, metal carbonyls, and many others. One of its major advantages over infrared techniques is the lack of interference from carbon dioxide (CO2), saturated hydrocarbons, and water vapor, which typically are present at much higher concentrations than the species of interest.
In certain applications, the only UV absorbing species is the analyte of interest. If this is the case, or its combination of concentration and molar absorbtivity is high enough relative to any trace interfering compounds, then an interference-free concentration measurement can be made. Many sample streams contain multiple species that have overlapping UV absorbance spectra. Analyzer spectrometer designs using only a single measuring wavelength are only accurate if no cross-interference is present. Although there are many single measuring wavelength UV analyzers still in use today, designs utilizing two or more measuring wavelengths are increasingly common1,2,3,4.
Designs of UV spectrometers with multiple measuring wavelengths will be either wavelength-dispersive or non-dispersive. Both designs have some advantages depending on what the requirements are for number and range of measuring wavelengths, size, cost, sensitivity, linearity etc. Figure 1 shows the design of a non-dispersive dual-beam UV spectrometer with exceptional stability, sensitivity and linearity. This spectrometer design is often used to measure multiple analytes in complex gas streams such as sulfur dioxide (SO2), nitric oxide (NO), and nitrogen dioxide (NO2) in stack gas or hydrogen sulfide (H2S), SO2, sulfur vapor (Sx), carbon disulfide (CS2) and carbonyl sulfide (COS) in Claus SRU tailgas1,3,4. Further details on the design of this spectrometer are available in Reference 1. This spectrometer design is capable of challenging and complex measurements but still has limitations in applications with a large number of UV absorbing components or where the high concentration of one component makes it difficult to measure the concentration of another with a much lower concentration.
Several applications exist where the standard measurement method has disadvantages such as complexity, high maintenance, inaccuracy, slow response time etc, but where the sample stream is unsuitable for using even a high accuracy multi-component UV analyzer such as the one described above. A good example of this type of application is the making this measurement, Gas Liquid Partition Chromatography (GLPC) or Lead Tape have various shortcomings (slow response time for both and high maintenance for Lead Tape), a UV spectrometer can’t make an accurate measurement because of interference from compounds with overlapping absorbance spectra. These interfering compounds may include benzene, toluene, ethylbenzene, xylenes (BTEX) and other aromatic hydrocarbons, and sulfur species that may include COS, mercaptans, thiophene, dimethyl sulfide etc., and various odorants that may be added. Figure 2 is a graph comparing the absorbance of 10 ppm of H2S to a typical natural gas sample that includes this same
concentration of H2S.
Chromatography is the term for various techniques used for the separation of mixtures. The components to be separated are distributed between two phases, one of the phases being stationary and with large surface area, and the other being a fluid (mobile phase) that flows through or past the stationary phase6. Chromatography may be either
preparative or analytical. Preparative chromatography is usually done when components of a sample are separated and collected for further use. When the separation is done for the purpose of measuring one or more of the components in the sample, then it is usually considered to be analytical chromatography6,7,8.
It is often assumed that gas chromatography was invented in the early 1950s but this is only true for gas-liquid partition chromatography (GLPC) which has a gas as the mobile phase and a stationary liquid phase coated on the inside of a column. This type of chromatography utilizes an elution mode of chromatographic separation, where a single precise dose of the sample mixture is injected into a flow of a carrier gas and the individual components in the sample are separated as they flow past the stationary phase and then are eluted as peaks. Retention time is the period of time after sample injection for a given component to elute from the column. Elution type absorption gas
chromatography using a solid phase in the columns was experimented with much earlier8.
In conventional gas chromatography, a small aliquot of sample gas is injected into a flowing stream of carrier gas, which then passes over a packed bed. As the gas flows through the bed, it diffuses to the surface of the particles, into pores, and is adsorbed and desorbed from the surface. As the carrier gas sweeps the analyte molecules through the column, this motion is inhibited by the adsorption of the analyte molecules either onto the column walls or onto packing materials in the column. The rate at which the molecules progress along the column depends on the strength of adsorption, which in turn depends on the type of molecule and on the stationary phase materials. Since each type of molecule has a different rate of progression, the various components of the analyte mixture are separated as they progress along the column and reach the end of the column at different times (retention time). A detector is used to monitor the outlet stream from the column; thus, the time at which each component reaches the outlet and the amount of that component can be determined. Generally, substances are identified (qualitatively) by the order in which they emerge (elute) from the column and by the retention time of the analyte in the column, and quantitatively by the size of the detector response, which may be characteristic to each species. Figure 3 portrays the transport of an injected peak through the column. The retention time and elution profile has been modeled extensively, in terms of theoretical chambers11, theoretical plates and film limited mass transfer12, While the peaks are often considered symmetric, peak tailing and asymmetry is noted and has been discussed extensively by Grubner13.
In frontal analysis chromatography (sometimes also called frontal elution chromatography), the sample flows continuously into the column with each component saturating the stationary phase before moving further down the column. The component with the least affinity for the column material is eluted from the column first in pure form. Then the component with the second lowest affinity elutes as a mixture with the first component. This process continues with each new component eluting as a mixture with all the previous components. While the sample mixture can be blended with a carrier gas, typically the pure mixture is used, in which case it is termed auto-carrier. Functionally, it is similar to some industrial processes using absorption beds or vessels. Because only the first component is eluted in pure form this method is not generally used in modern chromatography, since the non-specific detectors are incapable of speciating the eluting compounds. However, elution of mixtures as opposed to pure components is not necessarily a disadvantage in analytical chromatography, if a multi-component spectrometer is used to quantify concentrations of analytes that elute from the column instead of a simple detector (such as a thermal conductivity detector (TCD)). Only components in the mixture that absorb in the UV spectrometer wavelength operating range are important with the other components functioning as inert carrier gases. If a multi-component UV spectrometer is used, this allows the concentrations of several components to be determined before the column must be regenerated.
The same physical properties of mass transport apply to frontal elution chromatography as apply to conventional gas chromatography. The fundamental difference is the shape of the injection function and the shape of the elution profile. The pulse or delta function injection becomes a Heavy-side step function (the integral of the delta function over time), and the output function becomes the error function (the integral of the Gaussian over time). The input and output functions for frontal elution chromatography is shown in Figure 4. It should be noted that since the injection profile is no longer symmetrical (as in a delta function injection) the shape of the “peak” as it is transported through the bed
and the elution profile are no longer symmetrical either.
The retention of interfering compounds while allowing analytes of interest to elute quickly allows this method to be used in process analytical applications. The auto-carrier frontal elution method had been applied in the past using a non-specific detector (ultraviolet absorption at 202 nanometers) to measure only the first species to elute. This work, done by Western Research in the early 1980’s, resulted in a fast responding H2S analyzer for natural gas processing14.
Combining auto-carrier frontal elution chromatography (AFEC) with a multi-component UV spectrometer has advantages (other than a potentially large acronym) over other chromatographic methods while retaining the multi analyte measurement capability of GLPC. These advantages include:
The 933 analyzer uses the NDUV Spectrometer shown in Figure 1. Up to 6 bandpass optical filters can be used in this spectrometer to isolate wavelength emission lines from two hollow cathode (HC) lamps. For the natural gas application, measurements are made at four discrete wavelengths: 214 nm, 228 nm, and 326 nm from Source Lamp 1
(cadmium); 249 nm and 326 nm from Source Lamp 2 (copper). The 326 nm wavelength from both lamps is used to provide neutral drift compensation with the remaining wavelengths being used to measure H2S, COS and methyl mercaptan (MeSH). Analyzer operation is controlled by two microprocessors. One microprocessor controls the spectrometer, data processing, analytical calculations and spectrometer temperature, while the other handles I/O, calibration, solenoid controls, column temperature control, and digital Modbus communications.
The sample conditioning system uses auto-carrier frontal elution chromatography (AFEC) to remove any compounds from the sample that would interfere with the measurement of H2S, COS and MeSH. Figure 5 shows a flow schematic of the sample conditioning system for natural gas applications. If the natural gas pressure is higher than 400 psig (28 barg), it is reduced to 120–150 psig (8.3-10.3 barg) at the sample point using a pressure reducing probe or heated pressure regulator. Pressure reduction at the probe improves the analysis response time and lowers the sample dew point temperature. Heat traced sample lines are typically used to transport the sample from the probe to the analyzer where any particulates or liquid aerosols are removed with a three-stage filter that contains two membranes and one coalescing element.
After filtering, the pressure of the gas sample is reduced to 80 psig (5.5 barg) with a single-stage pressure regulator. An overpressure relief valve in the solenoid block limits the maximum pressure to 150 psig (10.3 barg) in the case of a failure in the upstream pressure regulation. Two three-way solenoid valves are used to control the flow of sample gas through the two columns. During normal operation, sample gas flows continuously through one three-way solenoid valve, through one of the columns in the column block, into the measuring cell, through a flow restrictor which reduces the pressure, backflushes the other column, and then is directed to vent by the other threeway solenoid valve. The columns have spiral paths in cylindrical cores located in the temperature controlled column block. Using a spiral column form allows different length columns to fit into the same space by adjusting the pitch on the helix. Appropriate
column length and porous polymer column material are selected so that H2S is the first UV absorbing component that elutes from the column. The next UV absorbing components that elute from the column are COS and MeSH.
The continuous sample flow acts as the carrier for the components to be measured and while one column is operating in absorbing mode, the other is being regenerated. Regeneration is achieved by reducing the pressure and backflushing with purified sample from the outlet of the measuring cell. This process is only effective when the columns are switched before the elution of the first compound that is not being measured. In a typical natural gas application the first eluting UV absorbing component that is not measured is ethyl mercaptan (EtSH). Since this regeneration process is similar to pressure swing absorption, it is important to maintain the 6 times pressure ratio between adsorption (94.7 psia) and desorption (close to 14.7 psia). The flow restrictors which achieve the pressure reduction for regeneration also serve to control the sample flow to approximately 2 LPM. After each column switch, a track and hold (T&H) function is used to hold the analyte concentrations at the value immediately before the column
switch, until each analyte has eluted from the new column. Typical hold and column switch times for H2S measurement in natural gas are 30 seconds and 2 minutes respectively. The sample gas composition does have some effect on the elution times with hold times and column switch times being about twice as long for a nitrogen sample as compared to natural gas. Response time is limited to 30 seconds by the hold time, although for the remaining 90 seconds flow time on a given column, the response time is much faster since any concentration change is relative to the absorption equilibrium with the column material rather than the regeneration equilibrium. For applications requiring COS and MeSH measurement, the column switch time is extended to about 3 minutes to allow the MeSH to completely elute to a stable plateau. The drawing in Figure 6 illustrates typical behavior of both the T&H and live concentrations.
Cell temperature and pressure measurements are used for live compensations. Measuring cell lengths between 0.5 mm and 81 cm can be used, with the most common lengths being 40 cm for 0-25 ppm H2S full-scale range and 81 cm for 0-5 ppm H2S ranges. With an 81 cm cell length, the minimum measuring ranges for COS and MeSH are 0-25 ppm and 0-15 ppm respectively. A two-way solenoid valve is used to automatically zero the analyzer every 24hrs using nitrogen or carbon dioxide. In the most common format, this analyzer is packaged in flameproof enclosures and has ATEX, CSA and UL certifications for Division 1 and Zone 1 hazardous areas. The flameproof packaged version of the analyzer is shown in Figure 7. Other available variations of this analyzer include a purged version for Division 2 locations with a sheet metal electronics enclosure, and also a general purpose rack-mount unit. A temperature controlled oven can be added to these versions along with an IR sensor to measure CO2 or a TC sensor to measure hydrogen.
Over 500 of these analyzers are measuring H2S in natural gas, biogas, CO2 and LPG applications around the world and have proven to be very reliable. Additional potential applications exist in catalytic reformer H2 recycle processes, tail-gas treating units (TGTU) and stack gas. This analyzer has several advantages compared to lead acetate tape analyzers including fast response time, long maintenance intervals of 6 to 12 months, COS and MeSH measurement capability, and self recovery after high concentration upsets. Because the wavelength of the hollow cathode lamp emission lines does not change, these analyzers have very low span drift, so span calibrations are not normally necessary. Although the initial analyzer cost is higher than a lead tape device, most users find that the higher reliability, lower maintenance and faster response time results in a much lower cost of ownership. Figure 8 shows the response time of the AFECUV analyzer and two different brands of lead tape analyzers connected to the same sampling system when the process gas is shut off and then 5.2 ppm H2S span gas is introduced. Although this shows that the response time for the AFEC/UV spectrometer based analyzer is faster than the lead tape devices, the difference would have been more dramatic if the span gas had a methane background instead of N2 which almost doubles the response time. It can also been seen that the lead tape analyzer that uses predictive modeling has a slower response time but then overshoots on the concentration reading and that both lead tape devices take longer to accurately read after switching back from span gas to sample. Since detection of an off specification H2S concentration calls for immediate block valve closure at many producing facilities, an accurate, fast responding H2S analyzer reduces the operating costs and improves the safety of the metering station and pipeline operations15.
Figure 9 shows some data from a lead tape analyzer and a AFECUV analyzer measuring H2S and COS during a natural gas amine contactor process upset, where both analyzers are measuring from the same sample system. Before the upset, there is very good agreement between the two analyzers, but the AFECUV analyzer H2S reading reaches 4 ppm 100 seconds before the lead tape device does. It is interesting that the COS reading starts to increase long before the H2S. This may be because COS is only physically absorbed by the amine solution instead of reacting chemically like H2S10. If the upset started out with foaming of the amine then this may limit the physical absorption of the COS before the absorption of the H2S is affected.
Most applications for the AFECUV analyzer in natural gas have had H2S ranges of 25 ppm or less, but ranges up to 1500 ppm are done routinely. Ranges above 1500 ppm aren’t usually required because the interference from other components in the natural gas becomes insignificant at these high concentrations. The exception to this is when an
application requires a low measuring range for monitoring normal in-specification operation, but also providing a higher for gauging the severity of process upsets. Using a multi-wavelength UV spectrometer with the AFEC technique allows for the provision of these dual ranges with only one measuring cell by selecting a measuring wavelength with
high molar absorbtivity for the low range and then one with much lower absorbtivity for the high range. Figure 10 shows some data from an analyzer with 0-5 ppm and 0-1.5% H2S ranges when the span gas concentration is changed from 3 ppm to 1% (at the 4 minute mark) and then back to 3 ppm (at the 10 minute mark). This data demonstrates
the self-recovery ability of this analyzer.
One application with a 15% H2S measuring range and 0.5 mm cell length has also been done, but it was an unusual gas stream from a SAGD (stream assisted gravity drainage) tar sand process with high levels of dozens of interfering compounds including aromatic hydrocarbons and various sulfur compounds including high molecular weight unknowns. While unusual, this application demonstrated that the AFEC technique achieved adequate separation with very high column loading of both interfering compounds and H2S.
Elution times for column materials with the AFEC technique are not readily available. The published relative retention indices for solid column materials used in elution chromatography can be used as a guide in choosing materials for AFEC, although it is not always accurate.
Since most of the analyzer sample system including the sample switching solenoids are at room temperature, some of the limitations of this analyzer include the requirement for a clean, gas phase sample at a temperature above the dew point and also adequate sample pressure of about 80 psig to achieve reliable column regeneration. If the column material gets coated with liquid, it will no longer remove interfering compounds effectively. In many applications these limitations can be overcome by using sample coolers and pumps to pressurize and dry the sample gas, but these methods are intended to remove trace quantities of liquids, not to work with a sample such as butane which is liquid at 80 psig and 20°C. In some applications, other than natural gas, there can be interfering components in the sample which cannot be blocked by the columns. An example of this is fuel gas streams with high levels of acetylene and butenes which can limit the accuracy of low H2S concentration measurements.
For natural gas applications, there is almost no effect of the background composition on the H2S concentration reading of the AFECUV analyzer. Some methods such as TDL may require the use of an H2S scrubber to reduce the effect of changes in the background composition on low range H2S measurements. In addition to being a consumable, using an H2S scrubber is undesirable because it does not account for changes in composition between adjustments and it also reduces the on-line time for the analyzer. The data in Table 1 compares the H2S reading from an AFECUV analyzer calibrated using a 90% methane/10% N2 background span gas compared to the same H2S concentration in very different background span gases supplied using mass flow controllers.
One natural gas application initially had issues with cross interference on the H2S measurement from an unknown material that was not being removed by the columns. Analysis of the ratio of the absorbance at 214 nm and 228 nm suggested that the interfering compound might be NO (nitric oxide). Subsequent analysis of the sample gas using GC-Mass Spectrometry verified that there was about 50 ppm of NO in the process gas. The natural gas was being sweetened using the Sulfa-Check process, which uses a sodium nitrite solution to remove H2S. Sodium nitrite reacts with H2S in the natural gas, forming elemental sulfur and other solid particulates plus small quantities of NOx A change in the measuring wavelength from 214 nm to 202 nm allowed the H2S to be measured without interference from NO. It would also be possible to measure both H2S and NO in natural gas, and there may be a need for these measurements in some applications such as production from tar sands using in-situ combustion.
For natural gas applications, increases in C2+ hydrocarbon content results in faster elution times and therefore faster response time. Some of this effect may be due to a reduction in column material pore volume due to the larger average molecular size of the background gas, but it is also partially because the sample flow rate increases. Although it is perhaps somewhat counter-intuitive, heavier aliphatic hydrocarbons have lower viscosities than methane. As the average molecular weight of the hydrocarbon mixture increases, the viscosity decreases, increasing the sample flow. Gas flow through sintered metal flow restrictors is approximately inversely proportional to viscosity.
Figure 11 shows some data from an analyzer that was configured to measure H2S in stack gas. It also has additional wavelengths to measure SO2, NO and NO2. In this test, the sample gas contains 9.5 ppm H2S, 150 ppm NO2, 700 ppm NO, 3000 ppm SO2, 3% O2, with the balance being nitrogen. The columns are being switched fast enough to block all
of the SO2 and the NO2, with most of the NO either being converted to NO2 or being blocked by the columns.
Ultraviolet absorption spectroscopy has long been known to provide accurate, sensitive and reliable measurements of trace process gases. However, in some applications, matrix effects have resulted in unacceptable levels of interference and made measurements difficult if not impossible. Combining separation sciences, in the form of auto-carrier frontal elution chromatography with a UV spectrometer allows the advantages of the spectroscopic method to be retained while reducing or eliminating the interference issues in several applications.
- Hamish Adam PRESENTED AT: L'Analyse Idustrielle, Paris, France Keywords Ultraviolet, Photometers, Absorption, Process Control Abstract Many different analytical techniques have made the transition from the laboratory to online process monitoring. UV photometry has made this transition more easily than most because it possesses some inherent advantages. Simple, robust optics and absence of interference from […]
- Hamish Adam PRESENTED AT: Annual ISA Analysis Division Symposium, Framingham, MA Keywords Ultraviolet, Photometers, Absorption, Process Control Abstract Many different analytical techniques have made the transition from the laboratory to online process monitoring. UV photometry has made this transition more easily than most because it possesses some inherent advantages. Simple, robust optics and absence […]
PRESENTED AT: ISA Analysis Division, Houston, TX Keywords ULTRAVIOLET ABSORPTION SPECTROSCOPY, APPLICATIONS, SULFUR RECOVERY, SCOT CONTACTOR, NATURAL GAS, HYDROGEN SULFIDE Abstract Ultraviolet absorption spectroscopy has been used for decades in the refining and petrochemical industry. Online spectroscopic methods allow for rapid compositional analysis of process streams and provide the capability for real-time monitoring and control. […]
