A method for the highly accurate quantification of gas streams by on-line chromatography

Highlights

• A method for characterizing gas streams with on-line gas chromatography was designed.

• The on-line GC was provided with a TCD-methanizer-FID system of detection.

• Quantification methods were developed considering mass balances in the system.

• The internal and external standard methods for quantification were compared.

• The internal standard method produced highly accurate mass balances.

Abstract

Quantification of the gas streams from chemical systems such as catalytic reactors are routinely performed by on-line gas chromatography. Gas chromatographs used for this purpose are typically provided with a combination of thermal conductivity (TCD) and flame ionization (FID) detectors to be able to detect and quantify both permanent gases; CO_x, N₂, H₂, etc., and hydrocarbons. However, the accuracy of the quantification is hindered by the intrinsic limitations of each type of detector. Namely, TCD has low sensitivity and FID does not detect permanent gases. Therefore, modern gas chromatographs include methanizer units to partially overcome this shortcoming by converting CO_x to methane. However, as far as these authors know, the literature has not presented an analytical method to characterize gas streams with high accuracy by the simultaneous use of a combination of a TCD-FID detection system provided with a methanizer. This work is an attempt to solve this problematic; it consists of the formulation of a mathematical model for the well-known external and internal standard quantification methods in gas chromatography. The analysis of the gas stream from a catalytic reactor performing the combustion of methane was used to validate the developed method. The concentration ranges of the analysed gases were: 0.8–7.7 vol% of CH₄, CO₂, and CO, 7.7–38.5 vol.% of O₂, and 46.2–90.8 vol.% of N₂ at a total flow of 130 mL min⁻¹. It was found that the commonly applied external standard method leads not only to inaccurate quantification but also to physically meaningless carbon balances and conclusions on the behaviour of the selected model system. In contrast, the internal standard method led to a highly accurate quantification with a physically meaningful carbon balance. Considering these findings, this contribution also draws attention to the need for a thoughtful application of chromatographic methods when studying the reactivity of gas systems.

Introduction

On-line chromatography is the most used technique to quantitatively characterize the gaseous streams from chemical reactors and certain unit operations [1], [2], [3]. By “on-line,” it is implied that the gas chromatograph (GC) is connected to the outlet stream of the corresponding process unit. The gaseous samples must be introduced into the GC by intricate arrangements of valves. Fig. 1 sketches a basic six-way valve system that allows the automatic injection of the samples into the instrument. In a typical run, the outlet gas stream from the given process continuously passes through a loop, placed in the injection port of the chromatograph, and discharges in a vent (Fig. 1 top). When a sample is to be analysed, the valve is switched to a position in which the gas flowing by the loop is injected to the GC by passing the carrier gas of the instrument through the loop in order to carry the analyte into the chromatograph (Fig. 1 bottom). In most conventional on-line GC systems, the atmospheric conditions, namely, temperature and pressure, can affect the quantity of analyte collected by the loop to a degree where reproducibility between runs becomes a stringent constraint of the analysis [4]. To correct this issue, internal standards are added to the gaseous samples during each sample injection [2].; alternatively, some researchers have opted for implementing statistical methods for assessing the uncertainty in the quantity of analyte that is injected in the runs of their on-line GC systems [4,5]. After accomplishing the automatized injection of the sample, the carrier gas carries the analyte through the separation columns, where each gaseous compound is partially retained and separated and further eluted before reaching the detection system of the instrument.

Most on-line GC detection systems used for the analysis of gaseous streams from hydrocarbon or biomass related processes are typically configured with one or several detectors such as: i) one Thermal Conductivity Detector (TCD) able to detect gaseous compounds as a function of their differences in thermal conductivity and whose con-centration is higher than ~100 ppm; ii) one Flame Ionization Detector (FID) that detects hydrocarbons in concentrations higher than ~0.5 ppm, but cannot detect permanent gases like O₂, N₂, CO, CO₂, and H₂O [2]. Due to this limitation, most GC systems are provided by iii) a serial TCD-FID arrangement that allows detecting permanent gases with the TCD and hydrocarbons with the FID; and iv) TCD-methanizer-FID arrangements that allow detecting concentrations of CO₂ and CO below the 100 ppm limit of the TCD, by hydrogenating CO₂ and CO to methane after passing a fraction of the analyte through a small catalytic reactor; called the methanizer, and further submitting the outlet of the methanizer to the FID [6]. Despite the implementation of TCD-methanizer-FID arrangements, on-line GC systems still have other drawbacks when aiming to make a complete characterization of a gaseous stream containing both hydrocarbons and non-hydrocarbon gases.

On-line GC systems are fundamental for the study and monitoring of gaseous catalytic reaction units. Within this context, very accurate and reliable quantification of the concentration of each reactant is a must when assessing the performance of either the reaction unit or of the catalyst itself. Within this context, achieving closed mass balances from the input and output gaseous streams of the reaction systems is a necessary condition for process design and monitoring. Mass balances must be calculated from the relative concentrations quantified by the TCD and FID detectors of the on-line GC. However, such a task can be challenging up to a point where many process operators and even researchers either consider it helpless or tend to ignore it. The problem of achieving closed mass balances is associated with the inherent limitations of each type of detector as described before and with the fact that one needs to devise a reliable method for relating the concentrations of the different compounds from more than one detector at the same time.

As particular examples of the hurdles of measuring concentrations and closing mass balances from data read from more than one detector in an on-line GC, one can cite processes in which relatively simple reactions such as the total (Eq. (1)) or partial (Eq. (2)) combustion of methane, and the reforming of methane with carbon dioxide, a.k.a. dry re-forming, (Eq. (3)), must be followed by on-line GC.

$$\mathrm{C}{\mathrm{H}}_4+2{\mathrm{O}}_2\to \mathrm{C}{\mathrm{O}}_2+2{\mathrm{H}}_2\mathrm{O}$$

$$\mathrm{C}{\mathrm{H}}_4+3/2{\mathrm{O}}_2\to \text{CO}+2{\mathrm{H}}_2\mathrm{O}$$

$$\mathrm{C}{\mathrm{H}}_4+\mathrm{C}{\mathrm{O}}_2\to 2\text{CO}+2{\mathrm{H}}_2$$

To do so, the instrument must be at least be provided with both one TCD and one FID since the outlet gas stream is composed by CH₄, O₂, H₂, CO₂, CO, and H₂O [7], [8], [9], [10]. These reactions are traditionally studied as alternatives for the abatement or valorization of methane which is a 21 times more potent greenhouse gas than carbon dioxide [11], [12], [13]. At the research level, many scientists devote efforts towards the development of low-cost stable catalysts able to activate methane at the lowest possible temperature [11,12,[14], [15], [16], [17]]. For this purpose, the studies are most often carried out in fixed bed catalytic micro-reactors whose output is monitored by on-line GC with input streams whose composition is about 0.1–10 vol.% of methane, 2–50 vol.% oxygen diluted in nitrogen, argon or helium [11]. In this sense, one may consider the lower concentration of reactants and assume 80% conversion of methane to CO₂ (Eq. (1)), and 10% conversion to CO (Eq. (2)); the output stream would be composed for ~100 ppm CH₄, ~800 ppm CO₂, and ~100 ppm CO. If only a TCD is employed for quantification, the results would be 100% conversion of methane to CO₂, i.e., an overestimation of 10% in conversion and selectivity, which is due to the fact that both the hydrocarbon and CO are in concentrations hardly detected by the TCD.

For the above, a GC with at least a serial TCD-FID arrangement for detection is usually preferred. However, most investigators report quantification either by only the TCD or the FID or from an uncorrelated combination of both detectors. Furthermore, they seldom neither report statistics on the reliability of their measurements nor the mass balances for the analysed gaseous streams of their catalytic set-ups [18], [19], [20], [21], [22]. Besides the complications inherent to handle quantification with more than one detector, on-line GC analyses of the products from these reaction systems is complicated by the selection of an adequate internal standard. For example, the internal standard to be added must be more stable than the reactants under the usually harsh reaction conditions employed to test the catalysts [23,24]. Consequently, only gases such as N₂, He, and Ar are suitable as internal standards for methane combustion and dry reforming. However, none of these gases is detected by the FID hence hampering the analysis of the fluctuations in the amount of analyte injected after taken each sample. Under such circumstances, the analysis must rely on the response from the TCD even if the GC is also provided with a methanizer. As already mentioned, a TCD is two orders of magnitude less sensible as compared to an FID; hence relying solely on the former also harms product quantification.

This contribution presents an alternative method for quantifying the composition of gaseous streams by on-line GC. Particularly, a method that allows correlating the responses from a serial TCD-FID arrangement for compounds quantification, uses N₂ as an internal standard, and considers mass balance closure during the analysis of the gaseous effluents from a catalytic reactor was successfully implemented. As an example, the paper illustrates how this method was implemented to monitor the effluents from a lab-scale reaction set-up performing the catalytic combustion of methane. The mathematical deduction of a parameter called sensitivity factor allowed for the combined use of both the signals from the TCD and the FID for quantification despite their strong differences in sensibility. In general, the developed method can be applied to the on-line GC monitoring of any gaseous stream.