Implementation of a validation protocol to ensure a reliable assessment of the biomethane quality: practical demonstration of its fit-for-purpose using ISO 2620:2024 to comply with the requirements set in the EN 16723 standards

Vol.:(0123456789) Accreditation and Quality Assurance https://doi.org/10.1007/s00769-025-01673-7 RESEARCH Implementation ofavalidation protocol toensure areliable assessment ofthebiomethane quality: practical demonstration ofits fit‑for‑purpose using ISO 2620:2024 tocomply withtherequirements set intheEN 16723 standards SandraHultmark1· KarineArrhenius1· LucyCulleton2· ChristopherBamforth2· NurAinNazirahBintiNajurudeen2· OliverWilliams2· JianrongLi3· StefanPersijn3· IrisdeKrom3· KatarinaHafner‑Vuk4· JanBeranek5· PetraDědková5· JuditFűkő6· TamásBüki6· TanilTarhan7· AylinBoztepe7· DilaraKurt7 Received: 4 March 2025 / Accepted: 20 July 2025 © The Author(s) 2025 Abstract Biomethane is expected to be important in meeting Europe’s greenhouse gas reduction target. The composition of biomethane varies, therefore, depending on the feedstock and process parameters. The European standards EN 16723 specify requirements for injecting biomethane into natural gas networks and its use as vehicle fuel. Ensuring that biomethane composition complies with the requirements set in the standards requires rigorous quality control processes and validated analytical methods. Method validation for parties lacking practical experience or training can be a complex process involving numerous steps and requires adequate guidelines, which were lacking for biomethane. The BiometCAP project has, therefore, developed a standardized protocol for evaluating gas analyzers in biomethane applications with detailed procedures for assessing instrument performance and ensuring accurate and reliable measurements. This article describes how to use the validation protocol with practical applications for evaluating the limit of detection and limit of quantification, the working range and linearity, the precision, the bias and finally to calculate the measurement uncertainties using the analytical method described in ISO 2620:2024 as an example. This method based on TD-CG/MS-FID can therefore be considered fit-for-purpose, providing reliable, precise and sensitive measurements for the analysis of VOCs as demonstrated for 3-carene, dichloromethane and hexamethyldisiloxane. Finally, the article also summarizes the measurement uncertainties obtained during an extensive evaluation exercise organized between seven NMIs across Europe. For any future validation work, measurement uncertainties of 1 to 10% relative for any regulated compounds can be used as reference. Keywords Biomethane· Quality assessment· Protocol· Validation * Karine Arrhenius [email protected] 1 Research Institutes ofSweden (RISE), Frans Perssons Väg 6, 41276Göteborg, Sweden 2 National Physical Laboratory (NPL), Hampton Rd, TeddingtonTW110LW, UK 3 Van Swinden Laboratorium B.V. (VSL), Thijsseweg 11, 2629JADelft, Netherlands 4 Institute ofMetrology ofBosnia andHerzegovina (IMBiH), Branilaca Sarajeva 25, 71000Sarajevo, BosniaandHerzegovina 5 Czech Metrology Institute (CMI), Okruznı 31/772, 63800Brno, CzechRepublic 6 Government Office oftheCapital City Budapest, BKFH, Váci Út 172-176, 1138Budapest, Hungary 7 TUBITAK National Metrology Institute (UME), Barış Mah. Dr. Zeki Acar Cad. No:1, 41470GebzeKocaeli, Turkey Accreditation and Quality Assurance Introduction Biomethane is rapidly emerging as a key sustainable alternative to fossil fuels, with its use already widespread across Europe and expected to grow significantly in the coming years [19]. As a renewable energy source, biomethane plays a crucial role in helping the European Union (EU) achieve its ambitious climate goals, including a 55% net greenhouse gas emission reduction by 2030 compared to 1990 levels and the goal of reaching net-zero emissions by 2050 [38]. Biomethane is produced by upgrading biogas from the anaerobic digestion of organic waste, such as food waste, sewage sludge, manure, organic rest from different industries and agricultural by-products. As these feedstocks have rather different compositions, the resulting biogas logically exhibits composition profiles that depend upon the substrate(s). As a result, and also due to different production process parameters (for instance, the temperature), methane, carbon dioxide and significant impurities (hydrogen sulfide, ammonia, oxygen, nitrogen) concentrations vary in biogas. But even more pronounced, the type of substrates clearly determines the fingerprint of the volatile organic compounds (VOCs) in the biogas. Biogases [1, 33] from landfill or wastewater treatment plants (WWTPs) contain mostly hydrocarbons, mostly ramified or cyclic alkanes and siloxanes. Biogases from household food waste contain mostly terpenes (p-cymene and D-limonene are clearly dominant). Biogases produced from manure varies depending on the origin of manure (swine, cattle, etc.), but some of VOCs such as 2-propanethiol, α-pinene, benzene, toluene, 2-butanone and 2-methyl-furan are among the more commonly found in theses samples. Biogases from industrial wastes (example: digestion of sludge from paper mill or brewery, fish rests) contain toluene, methyl-furans, dimethylfurans and ethyl-furans. The complex and variable composition of biogas is mirrored in biomethane composition, where impurities and their concentrations depend on both the composition of the biogas and the upgrading techniques (the four main employed technologies being membrane separation, pressure swing adsorption, ammine scrubbing and water wash). Some of these impurities can affect the performance of gas vehicles and infrastructure, especially when biomethane is injected into existing natural gas networks. Ensuring the quality (and consistency) of biomethane is critical for its safe and efficient use in both natural gas grids and as vehicle fuel. Rigorous quality control processes are essential to meet the requirements set in standards developed for each potential application. For quality control, different analyzers, based on various measurement principles, are used to measure and quantify impurities in biogas and biomethane, but only instruments and associated analytical methods with proven performance and traceable validation parameters can guarantee accurate and reliable results as required for any testing and calibration laboratories following ISO 17025 standard [2]. In 2016–2017, the European Committee for Standardization (CEN) Technical Committee (TC) 408 developed and published EN 16723 [20] [30] specifications for the injection of biomethane into natural gas networks and its use as a vehicle fuel. Measurement uncertainties of 1%–10% are targeted for EN 16723 limit values. The analytical methods proposed in the EN 16723 series were, in most cases, originally developed for air or, in the best-case scenario, for natural gas. Therefore, the development and validation of specific methods were required to address impurities unique to biomethane (siloxanes, for instance) and/or to address the specificity of the biomethane matrix. A number of test methods have been developed and validated specifically in biomethane during the EMRP project ENG54 Metrology for biogas [34] and the EMPIR project Metrology for biomethane [13]. Most of these methods were then standardized in ISO/TC 193/SC1/WG25 biomethane. These methods intended to measure a given compound around and above the maximum levels stated in EN 16723 are presented in Table1. These methods were developed by National Metrology Institutes (NMIs), which have a long and vast experience in validating methods using adequate tools, including guides [5] (and other references [14]) and certified traceable reference materials. However, method validation for parties lacking practical experience or training can be a complex process involving numerous steps. The list of method performance characteristics, their definitions, and how to assess those vary slightly depending on the source. Therefore, more specific guidelines, together with practical applications of these guidelines, can be beneficial [31]. Validation protocols specific for performance evaluation of analytical methods exist for hydrogen used as vehicle fuels; ISO 21087:2019 [43] and for natural gas used in gas transmission and distribution systems; ISO 10723:2012 [44]. Both standards specify test methods for determining whether any chosen analytical system is fit-for-purpose and includes the method, equipment and sample handling. They require some predefined performance specifications. They also give recommendations or considerations on calculating an uncertainty budget for the amount fraction. However, a cost-effective, efficient and standardized protocol to evaluate methods for measuring impurity contents in biomethane is currently lacking, hampering such methods’ implementation in laboratories and the field [31]. While these analytical methods are available for testing biomethane Accreditation and Quality Assurance under EN 16723, their applicability still requires validation. The implementation of harmonized testing methods demands standardized protocols that define critical validation parameters and assess method performance accordingly. The BiometCAP project, which started in 2022, addresses the lack of a protocol specifically designed for evaluating gas analyzers in biomethane applications. The project has developed a comprehensive protocol supporting current and future measurement technologies, which will be submitted to ISO as a new international standard. The BiometCAP protocol provides detailed procedures for assessing the instruments’ performance to detect key impurities in biomethane. This protocol functions as a complete toolkit for accurately evaluating biomethane quality, incorporating standardized analytical methods for key parameters, reference gas materials for instrument calibration and method validation with step-by-step guidelines for the validation process. It outlines procedures for assessing critical parameters and offers best practices for sample preparation to ensure precise and reliable measurements. By implementing this approach, the BiometCAP protocol ensures high standards and fosters consistency across the industry. To ensure that the protocol was fit-for-purpose, demonstration of its applicability has been done by seven NMIs across Europe: NPL in the UK, RISE in Sweden, VSL in the Netherlands, IMBiH in Bosnia and Herzegovina, CMI in the Czech Republic, BFKH in Hungary and UME in Türkiye for different families of compounds using already validated methods (See Fig.1). Table 1 Methods developed for biomethane IC: Ion chromatography TD-GC–MS/FID: Thermal Desorption–Gas Chromatography–Mass Spectrometry–Flame Ionization / detector GC-IMS: Gas Chromatography–Ion Mobility Spectroscopy µ GC-TCD: Micro Gas Chromatography–Thermal Conductivity Detector AES: Atomic Emission Spectrometry TDLAS: Tunable Diode Laser Absorption Spectroscopy Impurities Measurement principle ISO standard Amines TD-GC–MS/FID (specifically alcohol-amines and piperazines) ISO/TS 2610:2023 [42] ISO 2620:2024 [12] Halogenated compounds IC (HF and HCl) ISO 2611–1:2024 [45] TD-GC–MS/FID (halogenated organic compounds) ISO 2620:2024 [12] Ammonia TDLAS ISO 2612:2023 [29] Total silicon content or Siloxanes GC-IMS (siloxanes) ISO 2613–2:2023 [28] TD-GC–MS-FID (Siloxanes) ISO 2620:2024 [12] AES (total silicon) ISO 2613–1:2023 [27] Terpenes µ GC-TCD ISO 2614:2023 [39] TD-GC–MS-FID ISO 2620:2024 [12] Compressor oil GC–MS (or GC-FID) ISO 2615:2024 [41] Other VOCs (ketones, hydrocarbons, furans…) TD-GC–MS-FID ISO 2620:2024 [12] Fig. 1 Evaluation exercise of the protocol for validation of methods or qualification of instruments used in biomethane quality assessment conducted by seven NMIs Accreditation and Quality Assurance This article describes this evaluation exercise with examples of practical applications for siloxanes, terpenes and halogenated organic compounds using analytical methods and instrumentations described in ISO 2620:2024. In the standard, a mass spectrometer (MS) and/or a flame ionization detector (FID) can be used to quantify the VOCs. This paper also evaluates both detectors and discusses their advantages and disadvantages for the selected compounds in biomethane quality assessment. Finally, the article also summarizes the relative measurement uncertainties obtained by all the NMIs involved in the evaluation exercise of the BiometCAP protocol. This exercise demonstrates clearly that targeted measurement uncertainties of 1 to 10% relative can be achieved and as reference values for any future validation (for new methods or new instruments). Material andmethods A total of three traceable certified gas standards have been produced. Two of these were produced by the National Physical Laboratory (NPL). The standard used for terpenes contained 3.32 ± 0.1µmol.mol−1 of alpha-pinene (amount fraction followed by expanded uncertainties based on standard uncertainties multiplied by a coverage factor k = 2), 3.00 ± 0.1µmol.mol−1 of 3-carene and 3.18 ± 0.1µmol. mol−1 of D-limonene (Sigma Aldrich, Saint-Louis, Missouri, USA) in methane grade 6.0 (Buse Scientific, Crewe, Great Britain). The standard used for siloxanes contained 0.0874 ± 0.0044µmol.mol−1 of hexamethyldisiloxane (L2), 0.0557 ± 0.0034µmol.mol−1 of octamethyltrisiloxane (L3), 0.0542 ± 0.0038µmol.mol−1 of hexamethylcyclotrisiloxane (D3), 0.0429 ± 0.0026µmol.mol−1 of octamethylcyclotetrasiloxane (D4) and 0.0326 ± 0.002µmol.mol−1 of decamethylcyclopentasiloxane (D5) (Sigma Aldrich, Saint-Louis, Missouri, USA) in methane grade 6.0. All materials used underwent purity analysis as described in ISO 19229:2019 [7]. NPL’s reference gases were prepared in 10-L aluminum Luxfer cylinders with SpectraSeal® internal passivation (BOC, Surrey, England). The cylinders were evacuated for a minimum of 12h to a pressure below 3 × 10–7mbar. Pure liquid components were transferred into stainless steel transfer vessels which were weighed against a tare vessel of nominally identical shape and volume. The transfer vessel was then attached to the evacuated cylinder via evacuated connections to allow the component to evaporate into the cylinder. Following addition, the transfer vessel underwent a second high accuracy weighing to quantify the total mass of component transferred to the cylinder. The methane was added directly to the cylinder, which was weighed accurately before and after the addition. For the siloxane mixture, a secondary dilution step took place to dilute the amount fractions of analytes to the required < 1µmol.mol−1 level. The reference gas used for halogenated VOCs was produced at National Metrology Institute of the Netherlands (VSL) and contained 0.0526 ± 0.0026µmol.mol−1 of 1,2-dichloropropane, 0.0695 ± 0.0035µmol/mol of 1,1,2-trichloroethane, 0.050 ± 0.002µmol.mol−1 of dichloromethane, 0.0524 ± 0.0026µmol.mol−1 of tetrachloroethylene and 0.0528 ± 0.0026µmol.mol−1 of trichloroethylene in methane (grade 6.0). The gas mixture was prepared starting from a gravimetrically prepared liquid mixture. This mixture was prepared using a liquid mixture robot (LMR) developed at VSL. The LMR is equipped with a multipurpose sampler (MPS, a multifunctional autosampler and sample preparation robot, from Gerstel, Mülheim an der Ruhr, Germany), an analytical balance (from Mettler-Toledo International Inc., Columbus, Ohio, USA) and a sample preparation system (from PAL System, Zwingen, Basel-Stadt, Switzerland) for controlling of the MPS and the balance. The LMR allows automatic preparation of liquid mixtures containing up to 20 individual components. In addition, it allows the manual introduction of a solid into the mixture (as long as the solid will dissolve in the liquid mixture). The individual components were added according to their boiling point in a decreasing order. The prepared liquid mixture contains specific amount-of-substance fractions of the different halogenated VOCs calculated according to ISO 6142–1:2015 [8]. All materials used underwent purity analysis as described in ISO 19229:2019 [7]. The obtained liquid mixture is injected with a syringe into the cylinder and then vaporized and diluted with methane, to obtain a stock mixture [8] [3]. The mother mixture was further diluted in two steps: First, a gas mixture with nominal amount-of-substance fractions of 0.75µmol mol−1 for all halogenated VOCs was obtained. This mixture was further diluted to obtain the final gas mixtures containing amountof-substance fractions of approximately 0.05µmol mol−1 for all halogenated VOCs. For this purpose, the gas mixtures were prepared in aluminum high-pressure cylinders treated with an Experis VOC coating. All the cylinders were rinsed using nitrogen (grade 6.0) (6 × 15 bars and 2h vacuum); then the cylinders were evacuated overnight using a high vacuum turbo pump (p ≈ 2 × 10–7mbar). The amount fraction for each compound was then converted to µg.m−3 (at atmospheric pressure and 15°C). Transfer gas standards were produced using a dilution rig developed at RISE to calibrate the instruments. The rig has three lines for the supply of gases. Transfer gas standards were generated by blending certified gas mixtures with pure methane. The third line, vent line, can be used to supply nitrogen to flush the rig and remove methane or any other gases that have been introduced. Each line is built with chemically inert tubing, pressure rating suitable for the Accreditation and Quality Assurance operating conditions and an on–off valve for safety reasons. The three lines are connected using tee connectors. After the tee connectors, the gases reach a mixing tube which ensures that the gases are well-mixed. Each line is equipped with a calibrated laminar flow element (LFE) flow controller (MC Series, Alicat, Tucson, USA), for which the volumetric flow rate is a function of the pressure difference over the LFE and the viscosity of the gas to be measured. Flow range: 0 – 2l/ min, 0500ml/min, 0–100ml/min. The flow controllers are connected with quick connectors so they can easily be moved from one line to another depending on the gases and the required gas composition. The analytical system here is compatible with the standard ISO 2620:2024. The traceable certified gas standards and the transfer gas standards were produced using a dilution rig and were first collected on sorbent tubes containing 200mg Tenax TA. Substances adsorbed on the sorbent tubes were released through a two-stage thermal desorption (TD) process. First, the tubes were heated to 275°C for seven minutes using a Markes TD100 desorber, initiating the release of the adsorbed compounds. The released substances were then concentrated in a cold trap containing graphitized carbon at -10°C. Subsequently, the trap was rapidly heated to 300°C, releasing the compounds for separation in the GC column. The analyses were performed with a 6890N GC-FID-MS system (5975C inert MSD) (Agilent Technologies, Santa Clara, CA, USA) in electron ionization mode under standard conditions (70eV electron energy, mass scan range of 29–300 amu). A BPX5 column (5% phenyl polysilphenylene-siloxane equivalent) with dimensions of 60m × 32mm ID × 1μm film thickness (Trajan, SGE, Melbourne, Australia) was used. The column effluent was split 1:3 between the flame ionization detector (FID) and the mass spectrometer (MS). Methodology Description oftheprotocol The protocol describes a process/plan for determining whether an analytical system for biomethane composition analysis is fit for a defined purpose. It is assumed that the analytical system is intended to be applied to compositions that vary over the typical ranges found within biomethane.To determine performance characteristics, traceable calibration gases with defined uncertainties are required.The protocol for determining the performance characteristics of the method consists of eight steps (four planning steps, one experimental step and, three calculation steps). 1) Specify the components to be measured and the measuring range (planning) 2) Establish the expected function of the response (for example, linear response) (planning) 3) Specify the set of reference gas mixtures needed (planning) 4) Choose a gas mixture to be used for routine calibration (planning) 5) Collect data to evaluate the performance characteristics (experimental) 6) Evaluate/calculate the performance characteristics related to the range (LOQ, LOD) (calculation) 7) Evaluate/calculate the performance characteristics precision and bias (calculation) 8) Calculate the measurement uncertainties The protocol then proposes definitions for several performance characteristics, along with guidelines and examples on evaluating those (selectivity, limit of detection and limit of quantification, working range, trueness, precision). The validation parameters evaluated in this study were selectivity, limit of detection and quantification, working range and linearity, trueness/bias and precision. Using these results, the expanded uncertainties (k = 2) were calculated for each component using the software MUKit. Evaluation oftheprotocol anddefinition ofperformance characteristics The protocol has been evaluated using the method described in ISO 2620:2024 for the limit of detection and limit of quantification, the working range and linearity, the precision, the bias and finally to calculate the measurement uncertainties. Three compounds representative of terpenes (3-carene), halogenated VOCs (dichloromethane) and siloxanes (L2) have been selected. These compounds are been found in real biogas samples, dichloromethane has been identified in biogas from municipal solid waste landfill [4], 3-carene is found in biogases produced from many feedstocks such as food wastes, WWTP sludges and industrial wastes while L2 is mostly found in samples of biogas from landfills [1]. Another rational behind the selection of these specific compounds is that they cover a wide range of boiling points from 40 to 170°C, covering more than half of VOC boiling range (up to 250°C), and encompass a broad spectrum of chemical properties, covering different functionalities. The definition of the performance characteristics evaluated in this study is briefly given below. Limit ofdetection (LOD) andlimit ofquantification (LOQ) LOD refers to the lowest concentration of an analyte that can be detected but not necessarily quantified, whereas LOQ indicates the lowest concentration at which the analyte can Accreditation and Quality Assurance be both detected and quantified with acceptable accuracy and precision [5]. Working range andlinearity The working range can be defined [5] as the interval over which the method provides results with an acceptable uncertainty. The lower end of the working range is bounded by the limit of quantification LOQ. The upper end of the working range is defined by concentrations at which significant anomalies in the analytical sensitivity are observed. A calibration functionestablishes a relationship between the measured instrument response and the analyte concentration [6]. For the instrument used in ISO 2620:2024, the function is expected to be linear within the working range. The linearity can be defined as the concentration range over which the intensity of the signal obtained is directly proportional to the concentration of the species producing the signal [6]. The next step is to confirm the relationship between concentration and instrument response by examining the regression statistics and residual plot for the chosen model [5] Precision The precision [9] is the closeness of agreement between indications or measured quantity values obtained by replicate measurements on the same or similar objects under specified conditions. Measurement precision is related to random measurement error and is a measure of how close results are to one another. Precision can be divided into three categories: repeatability, intermediate reproducibility and reproducibility. Repeatability refers to the consistency of measurements taken by the same instrument or individual under identical conditions. Intermediate reproducibility examines the consistency of measurements under varying conditions, such as different times, operators or equipment, but within the same laboratory. Reproducibility, however, evaluates the ability to replicate an entire study or experiment across different laboratories, with different operators, and using different equipment. Bias The bias [9] assesses the closeness of agreement between the average of an infinite number of replicates measured quantity values and a reference quantity values. Considerations aboutISO 2620:2024 ISO 2620:2024 implies analysis with a flame ionization detector (FID) and/or a mass spectrometer (MS). These detectors have both advantages and disadvantages when analyzing biomethane samples. The FID is based on the principle that organic compounds, particularly those containing carbon-hydrogen (C-H) bonds, ionize when exposed to a hydrogen flame. The ionization produces charged particles, generating an electrical current proportional to the quantity of carbon atoms in the sample. Due to its inherent properties, the FID is highly responsive to most organic compounds. Two of the key benefits of FID are its universal response to nearly all hydrocarbons and the possibility to even quantify analytes that are not readily available as calibration standards. In that case, calibration can be done by using a reference compound, for example, toluene and relative response factors (RRF) against the reference compound, which can be calculated using different approaches [35] [37] [38]. RRFs have been shown to be reproducible over time and from one instrument to another [36]. This RRF can be calculated using existing models or from the literature. However, the FID’s performance is notably influenced by the presence of heteroatoms such as oxygen, sulfur, chlorine, fluorine or silicon within the analyte. Compounds containing these atoms tend to produce weaker signals compared to hydrocarbons. This variability arises because the detector’s response is primarily proportional to the number of carbon atoms, with hydrocarbons yielding the strongest signals. For compounds containing heteroatoms, particularly halogenated compounds, the response diminishes significantly. In fact, halogenated VOCs often yield even lower signals than those containing oxygen, sulfur or silicon, due to the specific combustion characteristics of halogenated compounds. Therefore, if a halogenated compound coelutes with a hydrocarbon, the FID will detect both substances but interpret the signal predominantly based on the hydrocarbon’s stronger ionization response. This can result in an overestimation of the concentration of the target compound. For this reason, FID is most effective when applied to samples rich in hydrocarbons (for example, landfill biogases and biogases produced from the digestion of sludge from wastewater treatment plants, and for analyzing compounds present at prevailing concentrations such as terpenes in biogas and biomethane samples produced from household food wastes). On the other hand, MS offers a more sensitive and selective approach, making it highly suitable for detecting VOCs at trace concentrations. MS operates by ionizing compounds and then sorting the ions based on their mass-to-charge ratio. This allows the MS to precisely differentiate between compounds, even in complex mixtures. One of the primary advantages of MS is its ability to detect a wide range of compounds, including those that contain heteroatoms such as sulfur, oxygen and halogens, without the bias observed in FID toward hydrocarbons. MS is particularly useful insituations where the target VOCs are expected to be present in very low concentrations. This makes it particularly useful insituations where the compounds of interest have been Accreditation and Quality Assurance largely removed through purification processes, such as the case with siloxanes in upgrading techniques [20] or for compounds present in much lower concentrations than other compounds in a sample (for example, dichloromethane in most biogases and biomethanes). However, the effectiveness of MS depends heavily on the availability of reference mixtures for calibration. In contrast to FID, where a single reference compound can be used to calculate relative response factors, MS requires precise calibration for each analyte, which can limit its practical utility in some cases. Also, MS’s ability to distinguish between isotopes or different ion fragments provides an analytical depth that FID cannot match. With gas chromatography (GC), it is sometimes impossible to achieve perfect separation mostly for complex samples for which the risk of impurities coeluting increases. This is particularly the case for biomethanes produced in landfill or wastewater treatment plants where the identification of the VOCs is challenging due to the presence of a large number of hydrocarbons not well-separated chromatographically and which cannot be identified individually. In that case, coelution may result in wrong determination using the MS library due to the superposition of two or more mass spectra. However, MS can filter out many coeluting compounds that the FID may not by extracting out characteristic ion fragments which are present on the mass spectrum of the targeted compound but not in the coeluting one. This is the case for example, for siloxanes and terpenes even if present in a mixture of hydrocarbons. Biomethanes produced from food wastes, manures or crops usually contain relatively few compounds, and their identification is then possible to a large extent [1]. Results anddiscussions To demonstrate the method’s applicability for different families of VOCs (here specifically terpenes, halogenated compounds and siloxanes), one compound was selected from each family: dichloromethane, L2 and 3-carene. LOD andLOQ According to EN 16273 standards, the limit values for total volatile silicon (as Si) are specified as follows: 0.3 to 1 mg. Si m−3 (Part 1) and 0.3 mg. Si m−3 (Part 2). Among the nine common siloxanes (trimethylsilanol, L2, L3, L4, L5, D3, D4, D5, D6), we assume that the concentration of L2 should be below 0.033 mg. Si m−3. Given that L2 consists of 29% silicon, a reasonable detection limit for L2 is approximately 0.1 mg.m−3. Currently, there are no established limits for terpenes or halogenated VOCs. For simplicity, the same detection limit is applied to all compounds in this study. The protocol states that the LOD and LOQ should be calculated in one of two ways: via replicate measurements of blank samples or via replicate measurements of test samples with a suitably low amount fraction of the analyte. Here, the second approach was chosen as it is a more suitable approach for gas chromatographic methods, the method of the replicate measurements of blank samples being more adapted for spectroscopy methods. The LOD and the LOQ were determined using the signal-to-noise (S/N) ratio approach. This standard procedure involves comparing the signal intensity from samples with known low concentrations of analytes against those from blank samples. As per international guidelines [10], the S/N threshold values were set at 3 for LOD and 10 for LOQ. The results for the selected compounds detected using either FID or MS are summarized in Table2. The LOD values ranged from 0.002 (3-carene using FID) to 0.03 mg.m−3 (dichloromethane using MS) calculated with a typical volume of 100 ml gas sampled. All these values are much lower than the reasonable detection limit of 0.1 mg.m−3, and the relatively low LOD values highlight the high sensitivity of the TD-GC/MS/FID technique. In this study, the LOD and LOQ values using the MS detector were slightly higher compared to those obtained using the FID detector. In another study, LOQ and LOD determined with GC/FID and GC/MS were found to be similar for siloxanes [21]. The MS detector’s higher values are likely due to its increased noise levels and the need to tune ionization parameters carefully. The LOQ/LOD determined in this study with mass spectrometry are in the same order of magnitude (LOD of 0.01mg.m−3 and a LOQ of 0.05mg.m−3 for siloxanes [22] analyzed with solid-phase microextraction followed by GC/ MS) or lower (0.032–0.087mg.m−3 and LOQ in the range of 0.11–0.28mg.m−3 for siloxanes with using only GC/MS) than what was found in other studies [23]. In contrast, FID operates at a lower noise baseline, allowing for enhanced sensitivity at low concentrations. Another study comparing GC/FID and GC/MS found that the LOQ and LOD were Table 2 Calculated limits of detection and quantification for, 3-carene, L2 and dichloromethane using ISO 2620:2024 Compounds LOD, mg.m−3 (S/N = 3) LOQ, mg.m−3 (S/N = 10) MS detector 3-carene 0.008 0.03 L2 0.002 0.006 Dichloromethane 0.03 0.09 FID detector 3-carene 0.002 0.007 L2 0.007 0.02 Dichloromethane 0.01 0.04 Accreditation and Quality Assurance similar with both detectors. Nevertheless, both detectors yielded LOD and LOQ values that meet or exceed the initial requirements. Working range andlinearity Predefined required values: Present mass spectrometers offer approximately three orders of magnitude linear dynamic range due to the limitations in ion detection and the complexity of ionization processes, which can affect the linearity at very high or very low concentrations [24]. In comparison, a flame ionization detector (FID) typically reports a linear dynamic range of up to six orders of magnitude [25]. FID’s broader range is attributed to its more straightforward detection mechanism, which is less affected by variations in ionization efficiency. However, demonstrating that the instrument is linear in the region covering half of the EN 16723 standard’s maximum limit value up to 10 times the maximum limit value is enough to assess the conformity of biomethane against the requirements set in the standard. In practice, for siloxanes, the interval is 0.15 to 3mg. Si m−3, which corresponds to 0.015 to 0.3mg. Si m−3 for an individual siloxane and specifically, to 0.05 to 1mg.m−3 for L2. Dichloromethane is also expected to be present in low concentrations, so the same interval is chosen here. Even though terpenes are not considered to be harmful at low levels, they could mask odorization if present in the gas at a certain concentration. The reported odor threshold for terpenes varies from source to source (for example, 0.07mg. m−3 [26], or 0.1mg.m−3 [11]). Moreover, in high concentration, they can contribute toward saturation of filters used in biomethane plants (for example, activated carbon filters used to remove compounds such as hydrogen sulfide and VOCs). To assess the instrument’s working range, the protocol recommends performing four steps: 1. Identify the range of interest (recommended to span ± 10% of the expected calibration range) and measure blanks and calibration standards at 6–10 amount fractions spread evenly across this range. 2. Plot response against amount fraction and visually examine plot to identify the approximate linear range. 3. To quantify linearity over the identified linear range, measure blanks and calibration standards at 6–10 amount fractions spread evenly across the linear range. 4. Plot response (y-axis) against amount fraction (x-axis) and calculate appropriate regression statistics. Plot the residuals (difference in observed and predicted y value by the straight line fit). Random distribution around zero confirms linearity. The linearity of the method was evaluated by analyzing different amount fractions of each analyte in mg.m−3 calculated for a sampling volume of 100ml. Results for 3-carene, L2 and dichloromethane are shown in Fig.1. The correlation coefficients obtained with the two detectors were all close to 1, indicating that the linear regression equation accurately fits the data. This suggests that the method exhibits a linear working range across the tested concentrations. Additionally, the residuals—calculated as the difference between the predicted and observed areas—were plotted against the known amount fractions in mg.m−3. The random distribution of residuals further confirms the method’s linearity and its valid working range (Fig.2). Precision Predefined values: Precision of an analytical instrument using the method described in ISO 2620:2024 [12] is related to the ability of data processing equipment to precisely integrate the signal from the gas chromatographic detector, the ability of the detector to provide the same signal for the same amount fraction of an analyte and the ability of the thermal desorber to provide the same desorption characteristics for samples containing the same amount fraction of a given analyte. Previous evaluation exercises [13] have shown that a total precision of 3% relative is achievable. The protocol states that the evaluation of the repeatability is done by measuring the analyte(s) of interest within a reference material at various amount fractions for at least 10 replicate measurements spread across the working range using the same method and by calculating the standard deviation and relative standard deviation. In this study, the intermediate precision (same instrument, two analysts) was evaluated for 3-carene, L2 and dichloromethane using 12 replicates measured over a 1-month period. Duplicates of the analytes were measured at different concentrations across the working range for several days. These concentrations are more or less similar to the one used shown in Fig.2. Using this result, the standard deviation from control samples, SRW and the standard deviation from routine replicate samples, Sr, were calculated. The within-laboratory reproducibility, u(Rw) , was then calculated according to Eq.(1): The results are presented in Table3. The low degree of scatter observed between a series of measurements taken from replicates indicates that the method demonstrates high repeatability and consistency and in agreement with the predefined value of 3%. (1) u (Rw)= √ S2 RW +S 2 r Accreditation and Quality Assurance Bias Bias was determined by calculating the differences between the measured concentrations and the expected concentrations of 3-carene from GC-FID and L2, and dichloromethane from GC–MS. By incorporating the uncertainty of the reference standard into these calculations, the total bias, u(bias), was calculated according to Eq.(2): Fig. 2 Plots of peak areas versus amount fractions for a hexamethyldisiloxane (L2) b dichloromethane c 3-carene for MS (purple stars) and FID (green squares) and, the residuals, against amount fractions. The correlation coefficient is close to a value of 1.00 for both MS and FID, suggesting that the equation for the linear regression fits the data. This observation implies that the method encompasses a linear working range within the analyzed amount fractions. The distribution of residuals is random, confirming the linearity and working range