A Critical Review on Soil Gas Analysis: Modern Technologies and Problems

Abstract

In this review article, the main techniques for spectroscopic studies of gases in field conditions are considered. The issues related to the study of gas emissions from soils and the determination of their concentrations are analysed. The main types of spectroscopy used in portable devices for soil gas analysis, along with their design features and sampling approaches, are provided. Various studies aimed at optimising the operation of devices for analysing gases emitted from the soil, taking into account agronomic, agrochemical, and ecological specifics, are also presented. The effect of using different types of lasers and reflecting elements on the accuracy of optical measurements and the sensitivity to various substances in the gases is analysed.

1. Introduction

Soil gas analysis is very important in many fields, including engineering, environmental science, geochemistry, and agriculture, etc. [1,2] This analysis makes it possible to detect contaminants in soil [3], including greenhouse gases [4] and volatile organic compounds [5]. Another approach is to analyse the gases produced as a result of the presence of microorganisms in the soil. The analysis allows one to estimate the efficiency of the introduction of certain additives to the soil. In this way, the concentration of gases and the dynamics of their emission serve as indicators of microbiological processes in the soil, specifically microbial activity.

The soil consists of solid, liquid, and gas phases. Generally, the ratio of phases are 50%, 25%, and 25%, respectively [6]. Determining gas exchange and its rate is a complex task, as it depends on many factors. There are two types of gas transport processes in the soil: diffusion (according to Fick’s law) and advective transport (caused by the pressure drop between soil air and atmospheric air, as described by Darcy’s law) [7]. The majority of gases are nitrogen containing compounds [8,9,10,11] and carbon containing compounds [12,13,14]. The presence of plants induces an increase in gas fluxes from the soil. Soil with plants can produce various gases, such as CO₂ [15], N₂O [16,17], CH₄ [18,19], NO [20], NO₂ [21], etc. Some research focuses on the determination of greenhouse gases only [22,23,24]. This latter issue is actively studied, as the increase in temperature, along with the number of summer days, results in higher gas emissions, e.g., CO₂ emissions [15,25]. Typically, the emission of gases from soils is influenced by several factors, such as soil moisture, soil organic carbon, the presence of microorganisms, root processes, bulk density, and temperature [19,26,27]. The soil microbiome participates in biogeochemical processes, which should be further investigated [28].

The emission of gases from soil is a result of different processes, e.g., consumption and production, such as H₂ oxidation by soil enzymes; CH₄ oxidation by methanotrophic bacteria; CO cooxidation by the ammonium monooxygenase of nitrifying bacteria; N₂O reduction to N₂ by denitrifying bacteria, etc. [29]. All these processes differ depending on their occurrence in lowland and upland soils [29].

The formation of the chemical composition of the gas environment is influenced not only by plant life activity and anthropogenic factor in the form of agricultural activities but also by soil microflora. The monitoring of changes in the chemical composition of the gas and solid phase of soils is important for more accurate ecological and agricultural forecasting and planning of activities [30]. The completeness of information on the dynamics of soil formation is determined, among other things, by mass spectra obtained by spectrometry and chromatography. There are investigations devoted to field [31,32] and laboratory measurements of gas flux from soil [33,34]. However, each spectrometry imposes a number of requirements on the process of detection and further analysis of signals, which makes researchers and engineers look for ways of optimisation, especially for field studies where not only accuracy and resistance of measurements to noise but also the possibility of long-term studies, are important. One of the tasks that are solved by modern spectrometry methods is to provide non-destructive soil control in the fields of agriculture, ecology, and geology [35]. However, depending on the required quantity and quality of spectral data, the requirements for emitters, detection methods, and signal processing can vary. This is especially true for studies of microdoses of substances whose peaks in the spectrogram are negligible, but whose presence can provide certain information, e.g., data on the dynamics of microflora activity in the soil. Typically, investigating solid soil is a long-term and expensive process; however, gas analysis is faster and can be a non-destructive technique for this purpose [36].

The scheme of main methods for the investigation of gas emission from soils is shown in Figure 1.

There is no extensive classification of methods for soil gas analysis; all of them are based on different physical principles [37]. As can be seen, the majority of methods utilise optical techniques, such as Raman spectroscopy, infrared spectroscopy, and others. For some gases, e.g., CO₂, a traditional method such as titration with NaOH is used [38]. However, this method is not suitable for long-term measurements (months, weeks) and is time-consuming. Therefore, the application of instrumental techniques is the most effective approach for soil respiration analysis, as it requires frequent measurements and allows for automation.

This paper mainly reviews methods for determining gases in response to changes in organic amendments in the soil (soil respiration). This paper primarily reviews methods for determining gases in response to changes in organic amendments in the soil (soil respiration). This includes a review of spectroscopy techniques, detection principles, and the advantages and disadvantages of their use and optimisation for field studies. It also discusses examples of designs employed by other researchers to obtain mass spectra and how these designs affect the quality of the analysis. Section 2 provides an overview of the most common optical spectroscopy methods, as well as promising ones that can be used for non-destructive control in the field. It describes the pathways of gas-phase passage through radiation and detection elements in analysers, along with various design solutions utilised by researchers for specific methods. Section 3 delivers information on the mass spectrometry and gas chromatography for soil gas analysis. Section 4 deals with the gas analysis using various sensors. Section 5 is devoted to gas-phase sampling methods, particularly the use of various modifications of sampling chambers, which are employed for both soil analysis in the field and for pre-formed soil samples and slices. The influence of design features on the accuracy of measurements over time, gas losses during accumulation and pumping, and the occurrence of noise during analysis is also addressed. Section 6 outlines the challenges and prospects for spectroscopic techniques that researchers are currently exploring, the aspects of gas analysis that are developing most rapidly, and the techniques not previously used for soil gas analysis that may be considered for optimisation and improvement in agricultural applications.

2. Optical Spectroscopy in Soil Gas Analysis

Spectroscopy is based on the study of spectra of various types, as well as the detection of electron excitation in the molecules that constitute the substance and their transition to a higher energy level. The methods of spectral analysis are determined by the type of spectrum, the source of excitation of the molecules, and the type of interaction between the molecules of the substance and the sensor. Spectra differ in how electromagnetic waves interact with the molecules of matter, affecting several parameters used in subsequent mathematical calculations. Spectroscopy can include emission spectra, which detect the intensity of radiation; scattering spectra, which reflect information about the structure of matter in terms of the distribution, frequency, and polarisation of waves; and absorption spectra, which record the effect of wavelengths on electron excitation.

There are some limitations associated with each spectroscopic technique, which determine the ability to detect certain gases in soils. For example, infrared spectroscopy is limited in its ability to detect gases with non-polarizable chemical bonds (such as N₂, H₂, O₂, etc.), while gases with polarised bonds can be detected (e.g., CO₂). Some optical methods (e.g., Raman spectroscopy) do not have this limitation. Optical methods must have their own specific optical path length, which can affect the dimensions of the analysers, potentially making them relatively large. The issue of the size of analysers can be addressed by utilising a set of gas sensors that are small enough, such as chemiresistive sensors, optodes, non-dispersive infrared sensors, etc. Some methods enable the detection of gases produced by various isotopes (e.g., mass spectrometry) instead of optical spectroscopy, which is useful not only for soil analysis but also for geophysics and environmental studies. The distinctive feature of optical methods is the analysis of gases above the soil, since light does not pass through the soil sample.

2.1. Infrared Spectroscopy

Types of spectrometry differ in the use of different radiation sources, fixation methods, as well as optimisation of structural elements to improve the passage cycle of radiation through the analyte or excitation of atoms in the substance by other means. The main driving force behind the development of spectrometric methods towards universality and the possibility of express research is the advancement of laser technologies and the application of absorption spectroscopy. Absorption spectroscopy is currently the most common method for building setups of spectroscopic and hybrid (chromato–mass spectrometry) type [39]. Infrared spectroscopy is widely used in soil analysis [40,41], but the majority of papers are devoted to analysis of solid soil [42,43,44,45,46], whereas the number of papers devoted to gas analysis [47] is relatively low.

The principle of operation of infrared spectroscopy involves the passage of a beam from a radiation source through the gas phase, which results in the excitation of vibrational movements of molecules or their fragments and a change in the intensity of radiation. The structure of molecules in the sample affects the change in intensity, allowing, after passing the radiation through an optical filter, an infrared spectrum corresponding to the vibration of functional groups in the analyte to be obtained (Figure 2).

The spectrum is defined as a function of the intensity of transmitted radiation versus frequency (wavenumber). The spectrogram reflects absorption bands, the position and intensity of which provide information about the composition of the gas phase (Figure 3).

Regardless of the type of radiation, measurements are usually taken at all wavelengths within a certain range. They are divided into the near, middle, and far regions [50]. However, depending on the methodology, sources with narrower wavelength coverage may be used. The spectrum reflects the matching of the frequency of the wavelength with the frequency of the molecule’s vibration in the substance [50,51]. When absorbed, radiation causes a change in the dipole moment of the molecule [51], but some molecules, such as O₂, H₂, N₂, and others do not have it, so, for example, infrared spectroscopy is not used for their analysis [52], although these substances themselves can act as markers [53]. This is a main drawback of this technique. Therefore, additional different methods based on absorption, transmission of radiation, its scattering, and reflection are used for the detection of different substances [54].

For studying the emission gases of soil, infrared spectroscopy is an effective method of analysis because it allows automation of the analysis process, detecting from withdrawn samples, or conducting real-time research in field conditions. This facilitates environmental monitoring and agricultural planning of agrotechnical measures, allowing for precise local data on the gas composition in the soil and the dynamics of its changes to be obtained. The efficiency of infrared spectroscopy in soil analysis is described in [55] as this method covers a wide range of molecular vibration frequencies, allowing for non-destructive testing, and further cataloguing of spectra enables storage of data on the condition of soil masses for a long time and forecasting the dynamics of changes in physicochemical characteristics. New approach is the application of machine learning for gas sensing via infrared spectroscopy. For example, in [56] the one-dimensional convolutional deep learning neural network model has been demonstrated for detection of gases, showing the accuracy of 99% in a region of 400–4000 cm⁻¹.

Fourier transform infrared (FTIR) spectroscopy is an effective method for measuring the composition of trace gases, which is usually used in industrial soil gas analysers [57]. The use of different optical scheme in infrared spectrometers made it possible to create the faster and compact analyser for FTIR. This method is applicable for simultaneous measurement of multiple gases. The frequency and intensity of these individual absorption bands affect the overall spectrum, making it unique to the molecule of a specific substance. FTIR spectroscopy differs from the classical infrared method in that the spectrum is obtained not by measuring signal intensity, but by the response in spatial domain. For these purposes, optical spectroscopy methods use interferometers. The development of open-path gas analysers allowed for measuring the average mole fractions over large path lengths [58]. The open path FTIR spectrometers can be applied not only for agricultural purposes, but also for monitoring of greenhouse gases in geology [59].

The study in [60] describes the effectiveness of mid-infrared (MIR) spectroscopy for digital soil mapping. This method makes it possible to operate with solid and gaseous fractions in the soil. This technique is most commonly used because the vibrations of the majority of studied functional groups are comparable to the frequencies of mid-infrared wavelengths of 3–8 μm. Usually, a sieved sample or suspension of small particles is used as a sample for soil analysis. The application of diffuse reflection spectroscopy (DRS) for soil analysis is demonstrated in [61]. However, for the gases, other methods based on detecting the scattering spectrum are used because the requirements for sensor sensitivity and noise levels sharply increase in a pure gas environment without the use of an auxiliary eluent such as spraying or vapour creation.

Griffith et al. [62] used the FTIR for detection of trace gases, such as CO₂, CH₄, N₂O, and CO, reaching the precision of 0.1–0.5% and 1–2 min resolution. For some gases, the precision was high enough, e.g., N₂O measurements showed the precision of 0.2 ppbv with the variation within 0.5 ppbv. In [63], real-time monitoring of N₂O and its isotopologues was carried out. There was good agreement between the N₂O amount and consumptions of NO₂⁻ and NO₃⁻ when the atmosphere contained C₂H₂.

The other simple method to detect the gases is non-dispersive infrared (NDIR) spectroscopy, which is based on the measurement of characteristic infrared absorption. NDIR gas sensors are widely used for detection of various gases (CO_x, NO, SO₂, etc.); however, the increase in the number of gases to detect makes the equipment expensive and large in size, due to the need for many detectors and filters [64]. In [65], the NDIR device (Figure 4) for N₂O detection based on 59 cm path-length gas cell with microelectromechanical systems-based infrared emitter, pyroelectric detector, two anti-reflective coated optical windows, and convex lens has been created, covering the range of 1–2000 ppm without the impact of humidity. Laboratory tests showed good sensitivity (11.87 RMS/ppm at 3.87 ppm N₂O) to the soil atmosphere using the submerged diffusion cell.

Some approaches allow the creation of devices with multiplexed NDIR gas sensors by means of integrating the plasmonic metamaterial absorbers with pyroelectric detectors showing the enhanced response for eight gases (H₂S, CO₂, CH₄, CO, NO, CH₂O, SO₂, and NO₂; the LoD for methane, carbon dioxide, and carbon monoxide were 63 ppm, 2 ppm, and 11 ppm).

Overall, there is a lot of research devoted to infrared spectroscopy for the determination of gas species, whereas there have been few papers on the application of UV-visible molecular spectroscopy for this purpose. For example, the UV-visible molecular absorption spectrophotometer (with a path length of 1 cm) and diode array detector were used for the detection of VOCs in [52].

2.2. Raman Spectroscopy

One of the features of infrared spectroscopy is its ability to detect the molecules only with a dipole moment. However, some optical methods make it possible to overcome this obstacle, such as Raman spectroscopy [66,67,68]. This type of spectroscopy is also used for analysis of gas mixtures, e.g., analysis of natural gas [69,70,71], hydrocarbons [72], carbon dioxide [73], gases of fermentation (O₂, N₂, H₂, CH₄, and CO₂) [74], and fire gases [75]. Therefore, the task of analysing gases in soil using Raman spectroscopy seems to be similar to problems of such detection in chemical engineering, food engineering, metallurgy, and other fields.

Guo et al. [55] describe the method of cavity enhanced Raman spectroscopy (CERS) for analysing a multi-component gas medium. This method allows for effective detection of gas and enables work with compounds that do not have a dipole moment (N₂, O₂, H₂) and hydrocarbons (CH₄ and C₂H₄) in situ compared to infrared absorption spectroscopy. An example of the setup for Raman spectroscopy, which consists of a sampling cuvette whose structure is shown in Figure 5, a 532 nm laser emitting radiation through a 50 μm optical fibre into a probe via an adapter, and a 600 μm fibre for connecting the probe to the GRS-1000-532 spectrometer, has been shown. The cuvette in this method has a collimation node of excitation light, a resonator, and a receiving node. The Raman signal is formed by collimating the radiation from the source through lenses followed by reflection with concave lenses and signal accumulation by the lenses of the receiving device.

It can be noted that the number of papers devoted to the application of cavity-enhanced Raman spectroscopy in gas analysis increases significantly compared to those focusing on gas spectroscopy [76]. This suggests a growing application of cavity-enhanced Raman spectroscopy even in soil gas analysis.

The main difference in Raman spectroscopy compared to the infrared and its modifications lies in the fact that there is no change in the dipole moment, but rather a deformation of the electronic clouds. This effect is called the change in polarizability [48], and Raman spectra are obtained when analysing different media. As an example, the Raman spectrum of natural gas is shown in Figure 6.

The work [53], describes the design of a compact Raman spectrometer used for the analysis of greenhouse gases with a resolution of 1 cm⁻¹ by the Raman shift and detection limits of CO₂ isotopologues below 100 ppm, for CH₄ isotopologues < 25 ppm. It is assumed that the drawbacks of Raman spectroscopy in terms of low accuracy can be offset by using intense excitation beams in a gas cell with high pressure, so in this study, a gas chamber for high pressure made of titanium with anti-reflective enamel on the inner surface and windows made of optical glass K8 with antireflection coating was used. As for the analysis of soil emissions in situ in real time, a Raman spectroscopy system under pressure can be applied in a tunnel sample system, where the cuvette has a cylindrical shape and is placed at a specific depth in the soil.

Raman/FTIR spectroscopy is also applied for determination of concentration of gases [78]. In [79], the coupled Raman spectroscopy/FTIR analysis was used for monitoring of N₂, O₂, H₂O, and CO₂. It has been showed that Raman spectroscopy allows one to carry out measurements in wet conditions where FTIR spectroscopy does not show efficiency, even though the latter possesses higher sensitivity for CO₂ detection. The difference in signal to noise ratio (SNR) between these two methods, confirming the higher SNR of FTIR spectroscopy compared to Raman spectroscopy for CO₂ and water vapour, is shown in

Table 1. SNR of Raman and FTIR spectroscopy for various substances [79] (reproduced with the permission of Elsevier).

Raman spectroscopy	N2 250–400	O2 25–45	H2Ovapor 8–23	-	CO2 (1285 cm−1) 4–14	CO2 (1388 cm−1) 3–9	-
FTIR spectroscopy	-	-	H2Ovapor 500–710	H2Oliquid 500–710	CO2 (3609 cm−1) 135–300	CO2 (4983 cm−1) 7–20	H2Oliquid 500–710

.

Some modifications of Raman spectroscopy make it possible to carry out the in-field measurements. In [80] the shifted excitation Raman difference spectroscopy (SERDS) portable device, containing the sensor dual-wavelength diode laser (785 nm, 36 mW) with the SNR of 146 has been created. Surface enhanced Raman spectroscopy was also used for detection of nitrogen in there different types of soils by an ultra-portable Raman spectrometer and machine learning [81].

Even though Raman spectroscopy makes it possible to detect non-polar molecules (as mentioned above), the challenge of achieving a high signal-to-noise ratio (SNR) in certain environments remains critical. One of the most available and long-established methods of SNR improvement is the use of matched filters, but their use requires knowledge of the shape of the substance peak in the spectrum to match the filter’s impulse response. However, there are other techniques for mitigating Raman spectrometry. Fan et al. [82] used an automatic denoising method of convolutional denoising autoencoder for advancement of the SNR in Raman spectra without manual intervention (Raman spectrometer with 785 nm laser and Ne/Ar source). For example, the increase in SNR from 2.2 to 4.4 for a classical spectrometer, and from 3.2 to 7.2 for an optimised spectrometer was achieved. There are other approaches, such as spatially compressed illumination reported in [83]; deep learning-based approach in [84]; feature extraction method [85], etc. However, the majority of these methods were not tested for Raman spectra of gases, even in the field of soil science.

2.3. Cavity Ring-Down Spectroscopy

Cavity ring-down (CRD) spectroscopy is a direct absorption method that can be performed with pulsed or continuous light sources and has significantly higher sensitivity than achievable in conventional absorption spectroscopy [86,87,88,89,90,91,92,93]. The calculation of concentration is based on Beer’s Law [94]. The key feature of this method is measuring the rate of light absorption rather than the amount of light absorbed in a closed optical loop.

The scheme of CRD spectroscopy is shown in Figure 7 [95].

This method has several advantages over regular absorption spectroscopy [96]. Firstly, it is insensitive to fluctuations in the light source intensity, which enhances measurement stability and accuracy. Secondly, CRD spectroscopy allows for the creation of optical resonators with a long effective path length (it can reach the value of several kilometres [97]), which is particularly useful for measuring substances present in small concentration. In recent times, CRD spectroscopy has proven to be especially effective in spectroscopy by enabling the detection of even weak absorbing components. The scheme of analysis by CRD spectroscopy is shown in Figure 8.

The application of CRD spectroscopy makes it possible to detect the gases in quantities of several hundred ppb or even lower [99,100,101], which is an advantage of such approach for the analysis of soil gases. The second advantage is the possibility to create the device with low weight and low power [101]. The third advantage is the sustainability to noise of the laser, since the absorption path can be tens of kilometres [102,103,104]. Some research highlight the suitability of the method for environmental monitoring (e.g., detection of atmospheric δ¹³ CH₄ reported in [105]; detection of NO₃ radicals in atmospheric air [106]), monitoring of industrial gaseous mixtures [107], δ¹³C–CO₂ analysis in soil respiration [108] or even breath analysis [109].

The resonator is tuned to different wavelengths, and the change in ring-down time is measured. This change provides spectral information about the absorbing substances in the resonator and allows for determining their concentration. However, when applying this method, several considerations need to be taken into account. The reflective mirrors must have a very high reflectance coefficient [110]. For different types of gases, adjustments to laser parameters are necessary, which can be complicated by their technical characteristics. The requirement to use high-reflectivity mirrors (in order to achieve longer path lengths) is a feature of CRD spectroscopy and cavity-enhanced absorption spectroscopy, compared to chemical ionisation mass spectrometry (CIMS) [111]. For example, the mirrors with reflectivity of R ≥ 99.9985% with an effective path length of 20 km were used in [112]. The design of high-reflectivity mirrors represents a challenge for the implementation of CRD spectroscopy in soil gas analysis.

2.4. Off-Axis Integrated Cavity Output Spectroscopy

Off-axis integrated cavity output spectroscopy (OA-ICOS) [113,114,115,116,117,118,119] improves the accuracy by introducing laser light into the cavity at an angle relative to the main axis to avoid interacting with the high density of transverse modes. Although the intensity fluctuations are lower than those in direct axial ICOS, the method is still limited by low transmission and intensity fluctuations due to partial excitation of high-order transverse modes and typically can achieve sensitivity of ~10. OA-ICOS is used for detection of various gases in the nuclear industry (¹⁴CO₂ in CO₂ [113]), in breath analyses, as a non-invasive diagnostic (C₂H₆, CH₄, H₂O, acetone [120]; water isotopes [121]; C₂H₆ [122]), and leak monitoring of NH₃ [123], CH₄/C₂H₆ [124]).

In [125], a setup design for OA-ICOS that has been enhanced with radiofrequency white noise to improve the sensitivity of multicomplex gas-phase element detection has been shown (the LoD of CH₄ was 7.6 ppbv, corresponding to a minimum detectable fractional absorption scaled to the path length of 7.3 × 10⁻¹⁰ cm⁻¹). The design of the spectrometer is shown in Figure 9.

Wang et al. [126] presented an improved version of the OA-ICOS spectrometer using a time-sharing multiplexing method (Figure 10). A distinctive feature of the presented prototype is a dual-channel optical structure using two DFB lasers with ranges of 1603 and 1651 nm. In this system, input separators inject radiation from two sources into the optical cavity. In addition, the lasers are located off the optical axis, and the collimator system can be located on either side of the separators. Authors analysed the concentrations of CO₂ and CH₄ simultaneously (Figure 10b) with limit of detection (LoD) at the level of 0.271 and 1.743 ppb, respectively. This was achieved not only through a two-channel, two-source system, but also through the use of time division multiplexing.

Some papers related to detection of gases in other fields compared to agriculture, confirmed the prospects of OA-ICOS for soil gas analysis since the same gases are formed in soils. For example, in [113], it was shown that it is possible to carry out the ¹⁴CO₂ monitoring in the presence of N₂O and H₂O. He et al. [127] reported the application of wavelength modulated OA-ICOS based on a 50.8 mm long assembly Fabry–Perot cavity for CH₄ detection showing the 10.5 ppm LoD (averaging time was 30 s). The greenhouse gas analyser based on OA-ICOS technology for detection of CO₂ and CH₄ was created showing the accuracy 92.52–116.36 ppbv and 0.45–0.55 ppbv, respectively [128]. However, the detection of multiple gases is still an important problem to be solved in OA-ICOS [115].

2.5. Tunable Diode Laser Absorption Spectroscopy (TDLAS)

The basic setup for TDLAS consists of a tuneable diode laser light source, transmitting optics, an optically accessible absorbing medium, receiving optics, and a detector [129,130,131,132,133]. The wavelength of the emission from the tuneable diode laser is tuned to the characteristic absorption lines of the substance in the gas along the path of the laser beam. This tuning leads to a reduction in the intensity of the measured signal due to absorption, which can be detected by a photodiode and subsequently used to determine the gas concentration and other properties. The schematic of the TDLAS device for H₂ detection is shown in Figure 11.

There are two main approaches used in TDLAS: the application of multiple lasers for detection and a single laser approach [134]. Various diode lasers are employed depending on the application area and the range in which tuning needs to be performed. Typical examples include InGaAsP/InP (0.9–1.6 µm), InGaAsP/InAsP (1.6–2.2 µm), etc. These lasers can be tuned by adjusting their temperature or varying the injection current density in the amplification medium. While temperature changes allow tuning over more than 100 cm⁻¹, it is limited by the slow tuning speeds (several Hz) due to the thermal inertia of the system. On the other hand, current modulation can provide tuning at frequencies up to ~10 GHz, but it is limited to a smaller range (approximately from 1 to 2 cm⁻¹) in which tuning can be performed. The typical linewidth of the laser line is about 10⁻³ cm⁻¹ or less. Additional methods for tuning and narrowing the linewidth include the use of intracavity dispersive optics.

The main disadvantage of absorption spectrometry (AS), as well as laser absorption spectrometry in general, is that it is based on measuring a small signal change over a large background. Any noise generated by the light source or optical system degrades the detectability of the technique. Therefore, the sensitivity of direct absorption methods is often limited by absorption to ~10⁻³, which is far from the shot noise level, which for single-pass direct AS is in the range of 10⁻⁷–10⁻⁸. Since this is insufficient for many types of applications, AS is rarely used in its simplest operating mode. There are two main ways to improve this. One is to reduce noise in the signal; the other is to increase absorption. The first can be achieved using modulation methods, while the second can be obtained by placing the gas in a cavity where light passes through the sample several times, thereby increasing the interaction length. If this method is applied to detect trace particles, the signal can also be enhanced by detecting at wavelengths where transitions have greater line strength, for example, using fundamental vibrational bands or electronic transitions.

Most of the papers devoted to TDLAS applications for air analysis have considered the detection of gases, i.e., pollutants coming from industrial plants; however, some are related to soil gas analysis. Some research has demonstrated the efficiency of TDLAS compared to other methods. For example, this method was used to determine the molar fractions of ¹²CO₂ and ¹³CO₂ from grassland, along with field tests, independently, rather than using their ratio from mass spectrometry. The fractions of gases were extremely low (350–700 μmol mol⁻¹) [135]. Famulari et al. [136] presented data on field measurements of ammonia fluxes taken over 60 days in grassland in Southern Scotland. The research showed a small net emission of ammonia during the growing season (3.78 ng m⁻² s⁻¹). In [137], the TDLAS spectrometer for field detection of nitrous oxide and methane was developed, showing the precision of 0.1% with 1 s averaging.

Comparing TDLAS and CRD spectroscopy technologies, it is important to note that CRD spectroscopy allows for an increase in the optical path of absorption to thousands of metres using an optical mirror, which ensures extremely high detection accuracy. However, this method imposes quite high requirements on the experimental setup and complicates the conduct of online measurements on-site. TDLAS, on the other hand, is characterised by high adaptability to challenging conditions, excellent selectivity, affordability, and its measurement results meet most practical needs. The applied significance of this method has always attracted the attention of researchers. With the advancement of technologies, scientists are increasingly focusing on improving the detection sensitivity of the TDLAS system using various approaches [138].

The optical methods considered above are well-studied, whereas the application of some optical sensors (optodes) makes it possible to substitute large devices [37,139]. Currently, these sensors are used for the detection of ppbv levels of certain gases, such as ammonia [140,141,142]. However, their application in soil gas analysis could yield excellent results compared to the devices listed above.

3. Mass Spectrometry and Gas Chromatography

Mass spectrometry (MS) is one of the most rapidly developing areas of substance analysis due to its wide range of applications and the development of new methods to optimise the analysis process [143,144,145]. Measuring the spectra of substances in gaseous, liquid, or solid media and their subsequent systematisation allow for the extrapolation of obtained data to other measurements and the conducting of comparative analyses to determine the molecular composition of the medium. Despite the use of auxiliary components in most cases, such as liquid or gaseous eluents and various coatings of different compositions for ionisation, spectrometric methods are generally more suitable as a field research method, in contrast to chromatography. However, the combination of gas chromatography (GC) and mass spectrometry is frequently used [146,147] along with the separate research on the use of MS for soil gas analysis [148]. The main value to be analysed for carrying out the interpretation is the m/z ratio of ions analysed by MS, and the range of m/z limits the chemical compounds that can be determined. Taking into account that the majority of gases emitted from soils possess low molecular weights [27], the spectrometers and their design makes it possible to identify all of them. The significant difference and advantage of MS is its ability to recognise the gaseous compounds formed by isotopes (e.g., 12C and (e.g., ¹²C and ¹³C, ¹⁴N and ¹⁵N, ¹⁶O and ¹⁸O, as well as the compounds formed by these isotopes) [149], which is useful for the purposes of ecology, climate change, and geochemistry [150,151].

A typical graph of gas concentration determined using MS is depicted below (Figure 12).

Nakayama et al. [153] developed the time-of-flight portable mass spectrometer for soil gas analysis, making it possible to determine NO₂ and O₂ with the precision of ±34 ppbv and ±0.60 vol.%, respectively, operating with a resolution of 1 h within 5 days of field investigations.

Considering the advantages of MS application in soil gas analysis, the drawbacks can also be listed. Mass spectrometers are expensive, and the interpretation of analytical data is more complex. While the sensitivity of the technique is an advantage, it can also be a drawback, as all contaminants will be detected in the spectra. This sometimes results in data abundance, as the majority of the information pertains to contaminants that have no relation to the actual emissions from the soil.

MS has relatively high sensitivity and makes it possible to detect multiple gases, which is mainly a problem for gas chromatography [154]. It has almost no problems of interference sensing of gases, which is a feature of some optical methods, e.g., cavity ring-down spectroscopy [155]. Usually, the GC-MS devices are large and expensive enough; however, there are some data on the application of portable devices for analysis of gases, including field tests [156]. GC-MS technique makes it possible to detect even the hydrocarbon gases in soils, taking into account their interference. For example, the ion extraction method [157] was applied in to separate peaks from the aliphatic fractions, BTEX (benzene, toluene, ethylbenzene, xylenes), MTBE (methyl-tert-butyl-ether), ETBE (ethyl-tert-butyl-ether), halogenated VOCs, which made it possible to exclude the time-consuming process of treatment [143].

There are only a few papers on the application of gas chromatography specifically for soil gas analysis. Burford et al. [158] reported the use of gas chromatography for measuring CO₂ evolved from soils, using a He ionisation detector on a Porapak Q column. The treatment applied to the soil induced changes in CO₂ concentration. The results were obtained in only 6–7 min, which is significantly faster compared to the titration technique for CO₂ determination. Modini et al. [159] investigated the application of GC for measurement of CO₂, CH₄, and N₂O in soils with the detection limits of 2, 1, and 4 ppmv, respectively (the accuracy was –0.88, –0.94, and –3.17%, respectively).

One of the main disadvantages of gas chromatography in soil gas analysis is its complex design, the need for frequent calibration, and the requirement to maintain the stability of column operation. The majority of research is based on laboratory experiments [160]. There are almost no data on the field applications of portable GC devices for soil gas analysis.

4. Gas Analysis Using Chemiresistive Sensors

From some point, the method of analysing gas using chemiresistive sensors can be considered one of the types of adsorption spectroscopy, but it is not entirely accurate to call it spectroscopy [161]. The method is based on the application of a set of sensors calibrated for certain gases [162]. The mechanism of compound detection operates by utilising the effect of adsorption of gas molecules on the active sites. Adsorption can induce either an increase or decrease in the concentration of charge carriers in the sensing material, leading to a change in the sensor’s resistance [163,164]. The main component of the sensor is the material, which determines the sensor’s response, sensitivity, and selectivity. The sensors are prepared using various materials, such as metal oxides [165], dichalcogenides of transition metals [166,167,168], carbon nanomaterials [169,170], various composites [171,172], MXenes [173], etc.

The size of such sensors can be scaled depending on the material restrictions for the film and can take on various shapes, not just flat forms. Novel sensors are based on thin films that make it possible to create lightweight, flexible sensors [174].

The limits of sensor detection have become lower, reaching the ppb levels [175]. These sensors have high accuracy and selectivity, with the material for the film selected for specific types of gas environments. They are inexpensive to manufacture; however, drawbacks include their short lifespan, sensitivity to humidity, and reduced accuracy at excessive concentrations of particular molecules in the gas environment (due to excessive deposition of the molecules on the surface of the adsorbing layer).

There is not a great deal of research dedicated to the analysis of gases in soils; however, investigations of exhaled breath air for humans (low concentrations of gases such as CH₄ [176], CO₂ [176], NH₃ [177,178], NO₂ [179], VOCs [180,181], etc.) suggest that these sensors could be applied in this field. However, there are specific narrow areas where soil analysis has been conducted using chemiresistive gas sensors. For example, in [182] the sensors based on polypyrrole were used for detection of ammonium nitrate/fuel oil (ANFO) system used in improvised explosive devices showing the limit of detection of 73 ± 11 ppbv NH₃ under different soil types. The change in sensor resistance under contact with ammonia in soil is shown in Figure 13.

In [183], a chemiresistive sensor package for the detection of VOCs in soil was developed. The experiments demonstrated the efficiency of iso-octane detection in soil and field tests were successfully conducted in the Hazmat Spill Centre at the Nevada Test Site and at the Chemical Waste Landfill at Sandia National Laboratories. The sensor consisted of a polymer dissolved in a solvent and mixed with conductive carbon particles, which are then deposited on a solid substrate. Some articles have reported on advanced approaches in the treatment of data from chemiresistive gas sensors using calculation algorithms, which could benefit soil gas analysis. For example, Shooshtari [184] successfully used an electronic nose based on carbon nanotube-TiO₂ nanostructures for the detection of VOCs, along with support vector machine algorithms and principal component analysis (PCA) for data reduction. There are a lot of cases of PCA application for treatment the data of gas sensors [177]. It makes it possible to create the score plot of various characteristics of sensors (response, response time, recovery time, etc.) to various gases and find the distinguishable areas by various types of sensors for searching the most appropriate one [185]. The application of PCA-based approaches is also useful to improve the SNR values in other fields [186,187]. It can be suggested that such techniques may find its application in soil gas analysis.

Getino et al. reported one of the first papers dedicated to the application of chemiresistive gas sensors for soil analysis [188]. Sensor arrays were created for detection of combustion gases (NO_x, SO₂ and benzene) in soils.

The main problem is the necessity to use the defined type of chemiresistive gas sensor for analysis of certain gas. When the number of gases is high enough, the gas analyser becomes a complex device. However, the application of machine learning already helps to overcome this problem, making one sensing material appropriate for the analysis of multiple numbers of gases [189,190]. It can be noted that the machine learning is becoming applicable not only in chemiresistive gas analysis but also in soil gas analysis in general [2,191,192,193]. Of course, other types of sensors could be used for creating the arrays for soil gas analysis, e.g., quartz crystal microbalance (QCM) sensors [194] (there is an interesting paper since 1972 on detection of mercury vapours in soils using the QCM sensor [195]) and field-effect transistors (FETs) [196]. However, their price and availability are not the factor which could accelerate the application in this field. The price is one of the determining factors in the application of such sensors in soil gas emission analysis. However, it has already been shown in [197] that it is possible to create low-cost soil gas analysers using conventional gas sensors, e.g., MQ sensors, with a cost of around 50 USD. Additionally, there are more limitations of the application of chemosensors for soil gas analyzers. For example, the challenges include their recovery at room temperature, the effect of gas accumulation during adsorption, the stability of sensor response, and the modification of sensor materials for various fields of application.

It can be suggested that the soil gas analyser will be designed using various types of sensors that differ in their operating principles. For example, Romanak et al. [198] reported on the application of several commercial sensors used together for the analysis of N₂, CH₄, CO₂, O₂, and H₂O. Two infrared sensors were employed for the measurement of carbon dioxide and methane, while a galvanic cell was used for oxygen analysis. However, all-optical gas analysers using the same type of sensors (e.g., enhanced Raman gas sensors) are also found [199], but probably the combination of different types of sensors could bring higher limit of detections and responses for wide range of gases.

5. Methods of Sampling of Gas Emission

Gas exchange between the soil and the atmosphere is one of the key processes of the ecosystem, involving numerous processes occurring during the life activities of soil organisms and plants, as well as the migration of gases from anthropogenic interactions. Multiple studies of soil gas emission sampling methods show that the biological factor has a significant influence on the quality of the formed flux in the chamber and its concentration and dynamics of change over time. Moreover, these changes can be caused not only by the factor of the presence of a biological component in the study area but also by the design features of the flow chamber itself. It is mentioned in [200] that the activity of the root system of plants at the place of adjoining chambers of different manufacturers may differ, which inevitably leads to distortions of emission measurements. The control of microorganisms in the soil layer and their influence on the emission of greenhouse gases and other substances has received much attention due to the desire for environmentally friendly production of agricultural products. However, emissions from the greenhouse effect created by microorganisms and plants, especially when the task is to analyse the emission of soil gases from individual factors, increase the requirements for greater stability and accuracy of measurements from analytical equipment and auxiliary devices (flux chambers).

The design of the chamber for gas analysis plays an important role in obtaining correct results. There are two types of chamber systems: open systems (which allow ambient air to pass through the soil) and static systems (using non-flow-through-non-steady-state chambers [201]). The results of the first type strongly depend on the condition of the air (atmospheric pressure, relative humidity) [202]. Unfortunately, it is difficult to operate such systems, which require highly skilled staff. Static chambers are significantly cheaper and simpler compared to open systems [203]. Usually, the latter are applied for detection of greenhouse gases and non-reacting gases (CH₄, N₂O) [204]. Even so, the design of the non-steady state chamber is a critical problem (their seals, vents, dimensions of pipe, texture of soils, and a lot of other factors) affecting the quality of data obtained and errors occurring as a result of design [205,206]. It can be noted that these factors not only affect the gas exchange between soil and air but also there are many additional ones that is hard to account for them together: wind [207], rainfall/precipitation [208,209], soil water regime [210], groundwater level [211], etc. Therefore, the use of machine learning, artificial intelligence, and other information technologies is the critical need in this field.

The chambers are manufactured from plastics, metals, and glass. Usually, the materials should be inert chemically and do not react with gases to be measured. The use of some plastics (e.g., polyethylene, polyvinyl chloride, polycarbonate, etc.) should be checked before real field application [7]. From this point, metals (i.e., stainless steels) can be considered as the most appropriate materials; however, the problem of corrosion arises compared to plastics, which has no chemical and electrochemical corrosion under contact with soil (there are many other advantages of plastics for chambers for soil gas analysis, such as low density, ability to glue the elements easily). All chambers have thermal insulation for the protection of inner space from heating (polystyrene foam with foil for reflection of radiation heat transfer) [1]. Sometimes, the chamber for field measurements included the round shield on the top, protecting from the infrared radiation (e.g., CFlux-1 device, by Hansatech Co., Pentney, UK). The typical shape of such chambers is hemispherical; this shape also provides uniform mixing. The impact of soil temperature and moisture is considered by means of corresponding sensors installed in the chamber.

Glass-based chamber are not so frequently used, since it is fragile (for field tests, but appropriate for laboratory tasks), but it almost the most inert and convenient material for this purpose [7].

Studies show that gas emissions, especially carbon dioxide, depend on multiple factors influencing the measurement accuracy. In particular, the purity of CO₂ emission measurements is affected by the contact method of the sampling chamber with the soil surface, especially the depth of the sleeve installation, as indicated in [212]. In conditions of very loose desert soils, the location of the sleeve affects the surface temperature of the study area, indirectly influencing gas exchange under the temperature fluctuations in the desert throughout the day. The sampling process may vary not only due to the soil particle size distribution and climatic conditions but also because of the gases being studied and introductory conditions, such as emissions from geological activities or under the influence of agricultural work. It is mentioned in [213] that one of the industry standards in several countries, particularly in Canada, is the use of a tube-shaped rod chamber approximately 50 cm long (Figure 14). This sampling method is inherent in researching emissions in mining areas or areas where gas migration is presumed to occur during extraction processes. However, it helps better identify the correlation between gas flow and soil particle size distribution, as well as biological factors, as it encompasses multiple soil layers. An important element in studying soil gas emissions is whether the chamber is stationary or dynamic. A stationary chamber implies measuring the concentration of elements in an accumulating enclosed space without the release of the accumulated gas. However, this method may alter the diffusion gradient, affecting measurement accuracy. In dynamic chambers, gas is purged outside to an analyser in a volume equal to the amount of gas in the chamber. This method allows for connecting multiple flow chambers to a single analysing system.

In addition to the pillar type of chambers, a dome system, which is a container that establishes a tight seal with the soil surface, is used. It also features a channel for directing emission gases to the analyser. One of the advantages of this system is the ability to conduct analysis over a larger area and cover the most biologically active soil horizon, which is heavily influenced by anthropogenic factors and, thus, enhancing gas exchange more intensively. For example, in the work of [215], a non-flow-through (NFT) soil gas sampling system was used, adapted for assessing the emission of greenhouse gases such as N₂O and CO₂, which make a significant contribution to the chemical composition of the gas phase in pastures, areas with large herds of cattle.

In this experiment, the method of off-axis diode-laser spectroscopy with a declared dynamic range of the instrument of 0–10 ppm is used. To prevent a decrease in accuracy due to an increase in pressure inside the chamber, an oscillating fan is used, operating periodically, and the ventilation outlet is designed as a dish-type structure to create a variable gap. The study also assessed the types of sealing between the container and the lid, using both standard seals and filling with water, ballistic gel, Newtonian fluid, etc. Manufacturing rectangular metal lids for containers is a more accessible option for creating chambers with a large area. However, in certain situations, indicators such as air flow recirculation in a chamber with a specific geometry, as well as the degree of heating of the gas environment under a particular material, are crucial. In some cases, this can significantly affect measurement accuracy during prolonged real-time measurements.

The problem of pressure control during long-term measurements is important, and in some cases, it is recommended to install the pressure-sensitive transducers in dynamic chambers to somehow reduce the pressure deficit that may occur [216]. Each type of soil possesses the specific soil depth affected by the change in pressure, including the negative or positive pressure drop on the surface. This change in pressure should be compensated in the experimental apparatus [217].

In [28], the application of diffuse exchange system was reported using tuneable infrared laser direct absorption spectroscopy. The technique is based on the application of diffuse exchange membranes made of polymers (e.g., PTFE). These porous material-based techniques are frequently used [218]. However, the measurement system was more convenient for laboratory tests compared to field ones.

In many cases, this can significantly affect the accuracy of measurements during prolonged real-time monitoring. In their work, Maier et al. [219] presented recommendations for designing flow chambers to minimise noise during long-term measurements, taking into account factors such as photosynthetic activity, where the quantum yield of photosynthesis changes under weak illumination, or root exudation, which affects the activity of microflora and, thus, the dynamics of the soil gas composition. Groundwater pressure and precipitation also influence soil–atmosphere gas exchange. All gas flows, dependent on various factors, are depicted in Figure 15.

There are many features of experimental techniques linked with chambers. However, there are no strong recommendations. For example, the minimum depth of chamber insertion inside the soil is hard to unify. Usually it is within 5–20 cm [220,221].

Modern commercial flux chambers are made to ensure automation of the pumping process and minimise gas losses and the influence of external factors such as vibration, temperature, and infrared radiation. Such chambers include design solutions from Li-Cor, ABB, Picarro, PP Systems, and others. However, the systems have their drawbacks, which manufacturers try to emphasise in the design of the dome and gas exchange system. In [200], authors note that with some automated flux chambers, there may be problems with gas phase mixing under the dome as well as CO₂ diffusion under the chamber collars. However, improvements in the control of emission fluxes depending on the time of day, as well as paying more attention to the shape of the dome and the place where the chamber adjoins the ground, are made by the developers, which makes their design solutions more adaptive for long-term studies in different weather conditions and soil structure [214]. Researchers are trying to accumulate enough data on the use of flux chambers for specific climate zones. Bekin et al. [212] conducted experiments with automated gas chambers in desert conditions. They found that the emission of gases from soil in desert regions is not tied to periods of short-term rains, but depends much more on the time of day, as CO₂ and other emission components of the soil gas phase are absorbed at night. This fact, as well as the fact that in dry regions soil microflora and soil in general are more sensitive to temperature and other factors, increases the requirements to the design features of flux chambers. In particular, to ensure greater accuracy of measurements, it is necessary to minimise shading in the chamber, including in the area of contact with the surface. Artificial pressurisation and mixing of the gas phase can also be a serious factor, which also excludes the possibility of sequential connection of flux chambers. The authors’ experiment also showed the importance of the location of the collar in the soil. They measured the values of CO₂ emission and absorption, temperature fluctuations of the gas phase itself, and the ground surface at different types of collar placement 7.5 cm depth with a rim above the surface of 3.5 cm; 11 cm depth without a rim; 2.5 cm depth (Figure 16). A significant difference in CO₂ flux and temperature variations at specific time intervals between conventional and other types of collars was found. This is partly due to uneven heating of the surface due to shading, the influence of which was mentioned above.

Tropical and boreal zones also have their own specificity of flow measurements not only due to the influence of climatic factor, but also due to the peculiarities of gas exchange. In particular, tropical areas are a source of N₂O and a sink of CH₄. At the same time, very little information on the emission features has been collected, especially in certain regions such as the Congo Basin [222]. For a long time, there was a problem of adapting automatic chamber systems for different types of tropical forests (mountain, lowland, swamp). It is connected not only with places difficult to access for experiments, but also because of the specificity of gas emission, as well as the great influence of fauna activity on emission processes. In particular, researchers in tropical forests of Congo noted sharp increases in peaks in the spectra with increasing daily activity of termite colonies.

The other feature is the application of thermal desorption for components present in soil, which are volatile (VOCs, aromatic hydrocarbons, etc.). However, this approach is more directed toward environmental research and is not appropriate for gases released as a result of the action of microorganisms [223]. Some approaches are based on the desorption of VOCs with low boiling points, while others use high-temperature thermal desorption [224]. Nevertheless, the use of high temperatures may harm microorganisms; therefore, this method is mainly applied to contaminated soils.

Overall, a comparison of various characteristics of typical soil gas analysers is given in

Table 2. Comparison of various characteristics of conventional soil gas analysers.

Analyzer (Company)	Type of Measurement	Gases and Their Concentrations	Weight, g	Pressure of Pump, kPa	Country
Picarro G2508	CRD spectroscopy	N2O: 0.3–200 ppm, CH4: 1.5 –12 ppm, CO2: 380–5000 ppm, NH3: 0–300 ppb, H2O: 0–3%	22,600	40–133	USA
Gasmet GT5000	FTIR spectroscopy	N2O: oт 7 ppb, CH4: oт 40 ppb, CO2: oт 5 ppm, H2O: oт 0.01%, CO: oт 70 ppb, NH3: oт 70 ppb	9400	60–110	Finland
Li-Cor LI78xx	n/a	N2O: 0.4–100 ppm, H2O: 0–60,000 ppm, CO2: 50–2000 ppm, NH3: 0–30,000 ppb	10,500	70–110	USA
ABB GLA131/132/151	OA-ICOS	N2O: 0.5–40 ppm, CH4: 0.9–100 ppm, CO2: 0.35–20,000 ppm, H2O: 0–30,000 ppm	6400	70–110	Sweden-Switzeland
Hansatech Instuments CFLUX-1	IR spectroscopy	CO2: 1–30,000 ppm, H2O: 0.1–75%	8500	80–115	UK
Echo Instruments ECHO	Infrared, paramagnetic and electrochemical gas sensors	CO2: 0.35–5000 ppm, CH4: 0–10,000 ppm, O2: 0–25%, H2: 0–1000 ppm	7500	200–340	Slovenia

.

As seen, the weight of the analyser decreases when moving towards devices based on gas sensors. At the same time, the majority of analysis methods are based on optical methods with concentration limits as low as tens or hundreds of ppb. The trend towards developing low-LoD sensors will enable lower detection limits and significantly reduce the weight of analysers, which is beneficial for field applications.

6. Challenges

The main problem is caused by the features of soil gas analysis using optical spectroscopy analysers, particularly due to the design of the setups and the optical system, which makes them relatively large. However, the creation of small devices consisting of a multitude of gas sensors operating at room temperature could help overcome this issue, as the energy consumption, limits of detection (LoD), and relative response of chemiresistive gas sensors are low [164,225,226], making it possible to detect the concentration levels typically used in soil gas analysis. Solving such problem accompanied with the algorithms for enhancement of SNR and other sensor characteristics (as well as PCA or other approaches to eliminate the cross-sensitivity) will make it possible to approach the creation of the small gas analyser.

The second challenge is the sampling technique, which aims to create representative data from different points in the soil. It may be beneficial to treat the sampling points using AI approaches based on additional agronomic data. This means that the data from gas analysis should be considered alongside additional information on soils, such as pH, humidity, type, and other factors, since all of these can affect the pattern of gas emissions. The application of mathematical statistics and regression analysis to analyse gas data could enhance data quality. It is probable that the active involvement of machine, which is actively used for prediction of soil gas emission [227], will bring new outstanding results. Creation of maps of soil pollution with time and geographical points along with algorithms for their plotting will bring new data not only for agronomy, but for geochemistry, geophysics, environmental protection, etc. At the same time, the issues regarding the treatment of big data obtained from soils and their comparison with soil gas analysis data are important and need clarification. Even if the data could be obtained from different soils, the behaviour of these soils in various environments should be taken into account, representing an additional problem.

The issue of sample contamination and the presence of obstructive factors must also be taken into account. However, it is difficult to account it for all pollutants and biological agents in the soil that may affect gas emissions. Creation of recommendations and systematisation of research results could solve such a problem.

Interference from the signals of various gases could introduce errors in their detection. Therefore, applying principal component analysis and other methods should increase the precision of the acquired data.

7. Conclusions

In the present work, a review of modern instrumental techniques for soil gas analysis and the problems related to the determination of gases by each method has been conducted. The features of each technique, such as optical spectroscopy, mass spectrometry, etc., were analysed. The problem on the application of gas sensors for soil gas analysis was first considered. The studies demonstrated the potential of novel approaches for soil gas emission. Important problems that need to be addressed to improve the quality of gas analysis have been presented.