Partial Least-Squares Regression as a Tool to Retrieve Gas Concentrations in Mixtures Detected Using Quartz-Enhanced Photoacoustic Spectroscopy

Empirical Modal Decomposition Combined with Deep Learning for Photoacoustic Spectroscopy Detection of Mixture Gas Concentrations

In photoacoustic spectroscopy based multicomponent gas analysis, the overlap of the absorption spectra among different gases can affect the measurement accuracy of gas concentrations. We report a multicomponent gas analysis method based on empirical modal decomposition (EMD), convolutional neural networks (CNN), and long short-term memory (LSTM) networks that can extract the exact concentrations of mixed gases from the overlapping wavelength-modulated spectroscopy with second harmonic (WMS-2f) detection. The WMS-2f signals of 25 different concentration combinations of acetylene-ammonia mixtures are detected using a single distributed feedback laser (DFB) at 1531.5 nm. The acetylene concentrations range from 2.5 to 7.5 ppm and the ammonia concentrations from 12.5 to 37.5 ppm. The data set is enhanced by cyclic shifting and adding Gaussian noise. The classification accuracy of the test set reaches 99.89% after tuning. The mean absolute errors of the five additional sets of data measured under different conditions are 0.092 ppm for acetylene and 1.902 ppm for ammonia, within the above concentration ranges.

A Review on Photoacoustic Spectroscopy Techniques for Gas Sensing

The rapid growth of industry and the global drive for modernization have led to an increase in gas emissions, which present significant environmental and health risks. As a result, there is a growing need for precise and sensitive gas-monitoring technologies. This review delves into the progress made regarding photoacoustic gas sensors, with a specific focus on the vital components of acoustic cells and acoustic detectors. This review highlights photoacoustic spectroscopy (PAS) as an optical detection technique, lauding its high sensitivity, selectivity, and capability to detect a wide range of gaseous species. The principles of photoacoustic gas sensors are outlined, emphasizing the use of modulated light absorption to generate heat and subsequently detect gas pressure as acoustic pressure. Additionally, this review provides an overview of recent advancements in photoacoustic gas sensor components while also discussing the applications, challenges, and limitations of these sensors. It also includes a comparative analysis of photoacoustic gas sensors and other types of gas sensors, along with potential future research directions and opportunities. The main aim of this review is to advance the understanding and development of photoacoustic gas detection technology.

New silicon-based micro-electro-mechanical systems for photo-acoustic trace-gas detection

The achievable sensitivity level of photo-acoustic trace-gas sensors essentially depends on the performances of the acoustic transducer. In this work, the mechanical response of different silicon-based micro-electro-mechanical systems (MEMS) is characterized, aiming at investigating both their mechanical properties, namely the resonance frequency and the quality factor, and the minimum detection limit (MDL) achievable when they are exploited as an acoustic-to-voltage transducer in a trace-gas photoacoustic setup. For this purpose, a 4.56 µm Continuous-Wave (CW) quantum cascade laser (QCL) is used to excite a strong N2O roto-vibrational transition with a line strength of 2.14 × 10-19 cm/molecule, and the detection of MEMS oscillations is performed via an interferometric readout. As a general trend, the minimum detection limit decreases when the resonance frequency investigated increases, achieving a value of 15 parts per billion with a 3 dB cut-off lock-in bandwidth equal to 100 mHz, around 10 kHz.

Sensitive Detection of OCS Using Thermal Conversion Combined with Spectral Reconstruction Filtering Differential Optical Absorption Spectroscopy

Carbonyl sulfide (OCS) is a toxic gas produced during industrial processes that poses risks to both human health and industrial equipment. Therefore, detecting OCS concentrations plays a crucial role in early hazard warning. This paper presents an online system for detecting OCS at the ppb level using thermal conversion and spectral reconstruction filtering differential optical absorption spectroscopy (SRF-DOAS). First, OCS, which is not suitable for DOAS due to its weak absorption characteristics, is completely transformed into SO2 with strong absorption characteristics under high-temperature conditions. Then, the spectral reconstruction filtering method (SRF) is proposed to eliminate the noise and interference. The core idea of the method is to arrange the spectrum according to the spectral intensity from small to large rather than wavelength, reconstructing the spectrum into a new spectrum with linear characteristics. The reconstructed spectrum can remove noise and interference by linear fitting and retain the characteristic of SO2 oscillation absorption. Next, we demonstrate the ability of the reconstructed spectral method to remove noise and interference by comparing the spectra of the inverse-reconstructed gas mixture and SO2. The relative deviation of 0.88% at 100 ppb and detection limit of 7.26 ppb*m for OCS were obtained using the SRF-DOAS method. Finally, the reliability of the system was confirmed by measurements of OCS concentrations in mixture gas of OCS and air, as well as in human exhaled breath.

Multivariate analysis and digital twin modelling: Alternative approaches to evaluate molecular relaxation in photoacoustic spectroscopy

A comparative analysis of two different approaches developed to deal with molecular relaxation in photoacoustic spectroscopy is here reported. The first method employs a statistical analysis based on partial least squares regression, while the second method relies on the development of a digital twin of the photoacoustic sensor based on the theoretical modelling of the occurring relaxations. Methane detection within a gas matrix of synthetic air with variable humidity level is selected as case study. An interband cascade laser emitting at 3.345 µm is used to target methane absorption features. Two methane concentration ranges are explored targeting different absorptions, one in the order of part-per-million and one in the order of percent, while water vapor absolute concentration was varied from 0.3 % up to 2 %. The results achieved employing the detection techniques demonstrated the possibility to efficiently retrieve the target gas concentrations with accuracy > 95 % even in the case of strong influence of relaxation effects.

Quartz-Enhanced Photoacoustic Spectroscopy Assisted by Partial Least-Squares Regression for Multi-Gas Measurements

We report on the use of quartz-enhanced photoacoustic spectroscopy (QEPAS) for multi-gas detection. Photoacoustic (PA) spectra of mixtures of water (H2O), ammonia (NH3), and methane (CH4) were measured in the mid-infrared (MIR) wavelength range using a mid-infrared (MIR) optical parametric oscillator (OPO) light source. Highly overlapping absorption spectra are a common challenge for gas spectroscopy. To mitigate this, we used a partial least-squares regression (PLS) method to estimate the mixing ratio and concentrations of the individual gasses. The concentration range explored in the analysis varies from a few parts per million (ppm) to thousands of ppm. Spectra obtained from HITRAN and experimental single-molecule reference spectra of each of the molecular species were acquired and used as training data sets. These spectra were used to generate simulated spectra of the gas mixtures (linear combinations of the reference spectra). Here, in this proof-of-concept experiment, we demonstrate that after an absolute calibration of the QEPAS cell, the PLS analyses could be used to determine concentrations of single molecular species with a relative accuracy within a few % for mixtures of H2O, NH3, and CH4 and with an absolute sensitivity of approximately 300 (±50) ppm/V, 50 (±5) ppm/V, and 5 (±2) ppm/V for water, ammonia, and methane, respectively. This demonstrates that QEPAS assisted by PLS is a powerful approach to estimate concentrations of individual gas components with considerable spectral overlap, which is a typical scenario for real-life adoptions and applications.

A sub-ppbv-level Acetone and Ethanol Quantum Cascade Laser Based Photoacoustic Sensor - Characterization and Multi-Component Spectra Recording in Synthetic Breath

Trace gas analysis in breath is challenging due to the vast number of different components. We present a highly sensitive quantum cascade laser based photoacoustic setup for breath analysis. Scanning the range between 8263 and 8270 nm with a spectral resolution of 48 pm, we are able to quantify acetone and ethanol within a typical breath matrix containing water and CO2. We photoacoustically acquired spectra within this region of mid-infra-red light and prove that those spectra do not suffer from non-spectral interferences. The purely additive behavior of a breath sample spectrum was verified by comparing it with the independently acquired single component spectra using Pearson and Spearman correlation coefficients. A previously presented simulation approach is improved and an error attribution study is presented. With a 3σ detection limit of 6.5 ppbv in terms of ethanol and 250 pptv regarding acetone, our system is among the best performing presented so far.

Infrared dual-gas CH4/C2H2 sensor system based on dual-channel off-beam quartz-enhanced photoacoustic spectroscopy and time-division multiplexing technique

Highly sensitive and stable measurement of methane (CH₄) and acetylene (C₂H₂) based on a novel dual-channel off-beam quartz-enhanced photoacoustic spectroscopy and time-division multiplexing technique was realized by a compact 3D-printed gas cell with a size of 3 × 2 × 1 cm³. Two near-infrared distributed feedback diode lasers were employed to target the CH₄ absorption line at 6046.9 cm^-1 and the C₂H₂ absorption line at 6521.2 cm^-1, respectively. Second-harmonic wavelength modulation spectroscopy method was used for photoacoustic signal recovery. A minimum detection level of ∼ 7.63 parts-per-million in volume (ppmv) for CH₄ and a level of ∼ 17.47 ppmv for C₂H₂ were achieved with a 1 s lock-in integration time, leading to a normalized noise equivalent absorption (NNEA) coefficient of 7.24 × 10^-8 cm^-1·W·Hz^-1 and 3.73 × 10^-8 cm^-1·W·Hz^-1 for CH₄ and C₂H₂, respectively. Allan-Werle deviation analysis was employed to evaluate the stability and the minimum detection limit (MDL) of the developed photoacoustic CH₄/C₂H₂ dual-gas photoacoustic sensor. Owing to the high stability of the developed sensor system, an MDL of ∼ 0.73 ppmv and an MDL of ∼ 1.60 ppmv with a 100 s averaging time were achieved for CH₄ and C₂H₂, respectively.

Simultaneous Voltammetric Determination of Non-Steroidal Anti-Inflammatory Drugs (NSAIDs) Using a Modified Carbon Paste Electrode and Chemometrics

This work presents the simultaneous quantification of four non-steroidal anti-inflammatory drugs (NSAIDs), paracetamol, diclofenac, naproxen, and aspirin, in mixture solutions, by a laboratory-made working electrode based on carbon paste modified with multi-wall carbon nanotubes (MWCNT-CPE) and Differential Pulse Voltammetry (DPV). Preliminary electrochemical analysis was performed using cyclic voltammetry, and the sensor morphology was studied by scanning electronic microscopy and electrochemical impedance spectroscopy. The sample set ranging from 0.5 to 80 µmol L^-1 was prepared using a complete factorial design (3⁴) and considering some interferent species such as ascorbic acid, glucose, and sodium dodecyl sulfate to build the response model and an external randomly subset of samples within the experimental domain. A data compression strategy based on discrete wavelet transform was applied to handle voltammograms' complexity and high dimensionality. Afterward, Partial Least Square Regression (PLS) and Artificial Neural Networks (ANN) predicted the drug concentrations in the mixtures. PLS-adjusted models (n = 12) successfully predicted the concentration of paracetamol and diclofenac, achieving correlation values of R ≥ 0.9 (testing set). Meanwhile, the ANN model (four layers) obtained good prediction results, exhibiting R ≥ 0.968 for the four analyzed drugs (testing stage). Thus, an MWCNT-CPE electrode can be successfully used as a potential sensor for voltammetric determination and NSAID analysis.

Multi-gas quartz-enhanced photoacoustic sensor for environmental monitoring exploiting a Vernier effect-based quantum cascade laser

We report on a gas sensor based on quartz-enhanced photoacoustic spectroscopy (QEPAS) able to detect multiple gas species for environmental monitoring applications, by exploiting a Vernier effect-based quantum cascade laser as the excitation source. The device emission spectrum consists of ten separated emission clusters covering the range from 2100 up to 2250 cm-1. Four clusters were selected to detect the absorption features of carbon monoxide (CO), nitrous oxide (N2O), carbon dioxide (CO2), and water vapor (H2O), respectively. The sensor was calibrated with certified concentrations of CO, N2O and CO2 in a wet nitrogen matrix. The H2O absorption feature was used to monitor the water vapor within the gas line during the calibration. Minimum detection limits of 6 ppb, 7 ppb, and 70 ppm were achieved for CO, N2O and CO2, respectively, at 100 ms of integration time. As proof of concept, the QEPAS sensor was tested by continuously sampling indoor laboratory air and monitoring the analytes concentrations.

High-concentration methane and ethane QEPAS detection employing partial least squares regression to filter out energy relaxation dependence on gas matrix composition

A quartz enhanced photoacoustic spectroscopy (QEPAS) sensor capable to detect high concentrations of methane (C1) and ethane (C2) is here reported. The hydrocarbons fingerprint region around 3 µm was exploited using an interband cascade laser (ICL). A standard quartz tuning fork (QTF) coupled with two resonator tubes was used to detect the photoacoustic signal generated by the target molecules. Employing dedicated electronic boards to both control the laser source and collect the QTF signal, a shoe-box sized QEPAS sensor was realized. All the generated mixtures were downstream humidified to remove the influence of water vapor on the target gases. Several natural gas-like samples were generated and subsequently diluted 1:10 in N2. In the concentration ranges under investigation (1%-10% for C1 and 0.1%-1% for C2), both linear and nonlinear responses of the sensor were measured and signal variations due to matrix effects were observed. Partial least squares regression (PLSR) was employed as a multivariate statistical tool to accurately determine the concentrations of C1 and C2 in the mixtures, compensating the matrix relaxation effects. The achieved results extend the range of C1 and C2 concentrations detectable by QEPAS technique up to the percent scale.

Compact photoacoustic spectrophone for simultaneously monitoring the concentrations of dichloromethane and trichloromethane with a single acoustic resonator

Chlorinated hydrocarbons are frequently used as reagents and organic solvents in different industrial processes. Real-time detection of chlorinated hydrocarbons, as toxic air pollutants and carcinogenic species, is an important requirement for various environmental and industrial applications. In this study, a compact photoacoustic (PA) spectrophone based on a single acoustic resonator for simultaneous detection of trichloromethane (CHCl₃) and dichloromethane (CH₂Cl₂) is first reported by employing a low-cost distributed feedback (DFB) laser emitting at 1684 nm. In consideration of the significant overlapping of absorption spectral from trichloromethane and dichloromethane, the multi-linear regression method was used to calculate the concentrations of CHCl₃ and CH₂Cl₂ with special characterization of the absorption profile. The current modulation amplitude and detection phase in the developed PA spectrophone was optimized for high sensitivity of individual components. The measurement interference of CHCl₃ and CH₂Cl₂ on each other was investigated for accurate detection, respectively. For field measurements, all optical elements were integrated into a 40 cm × 40 cm × 20 cm chassis. This paper provides an experimental verification which strongly recommends this sensor as a compact photoacoustic field sensor system for chlorinated hydrocarbon detection in different applications.

Rapid Quantitative Analysis of IR Absorption Spectra for Trace Gas Detection by Artificial Neural Networks Trained with Synthetic Data

Infrared absorption spectroscopy is a widely used tool to quantify and monitor compositions of gases. The concentration information is often retrieved by fitting absorption profiles to the acquired spectra, utilizing spectroscopic databases. In complex gas matrices an expanded parameter space leads to long computation times of the fitting routines due to the increased number of spectral features that need to be computed for each iteration during the fit. This hinders the capability of real-time analysis of the gas matrix. Here, an artificial neural network (ANN) is employed for rapid prediction of gas concentrations in complex infrared absorption spectra composed of mixtures of CO and N₂O. Experimental data is acquired with a mid-infrared dual frequency comb spectrometer. To circumvent the experimental collection of huge amounts of training data, the network is trained on synthetically generated spectra. The spectra are based on simulated absorption profiles making use of the HITRAN database. In addition, the spectrometer’s influence on the measured spectra is characterized and included in the synthetic training data generation. The ANN was tested on measured spectra and compared to a non-linear least squares fitting algorithm. An average evaluation time of 303 µs for a single measured spectrum was achieved. Coefficients of determination were 0.99997 for the predictions of N₂O concentrations and 0.99987 for the predictions of CO concentrations, with uncertainties on the predicted concentrations between 0.04 and 0.18 ppm for 0 to 100 ppm N₂O and between 0.05 and 0.18 ppm for 0 to 60 ppm CO.

Adaptively Optimized Gas Analysis Model with Deep Learning for Near-Infrared Methane Sensors

Noise significantly limits the accuracy and stability of retrieving gas concentration with the traditional direct absorption spectroscopy (DAS). Here, we developed an adaptively optimized gas analysis model (AOGAM) composed of a neural sequence filter (NSF) and a neural concentration retriever (NCR) based on deep learning algorithms for extraction of methane absorption information from the noisy transmission spectra and obtaining the corresponding concentrations from the denoised spectra. The model was trained on two data sets, including a computationally generated one and the experimental one. We have applied this model for retrieving methane concentration from its transmission spectra in the near-infrared (NIR) region. The NSF was implemented through an encoder-decoder structure enhanced by the attention mechanism, improving robustness under noisy conditions. Further, the NCR was employed based on a combination of a principal component analysis (PCA) layer, which focuses the algorithm on the most significant spectral components, and a fully connected layer for solving the nonlinear inversion problem of the determination of methane concentration from the denoised spectra without manual computation. Evaluation results show that the proposed NSF outperforms widely used digital filters as well as the state-of-the-art filtering algorithms, improving the signal-to-noise ratio by 7.3 dB, and the concentrations determined with the NCR are more accurate than those determined with the traditional DAS method. With the AOGAM enhancement, the optimized methane sensor features precision and stability in real-time measurements and achieves the minimum detectable column density of 1.40 ppm·m (1σ). The promising results of the present study demonstrate that the combination of deep learning and absorption spectroscopy provides a more effective, accurate, and stable solution for a gas monitoring system.

Isotopic gas analysis by means of mid-infrared photoacoustic spectroscopy targeting human exhaled air

There is a great need for cost-efficient non-invasive medical diagnostic tools for analyzing humanly exhaled air. Compared to present day methods, photoacoustic spectroscopy (PAS) can provide a compact and portable (bedside), sensitive and inexpensive solution. We demonstrate a novel portable photoacoustic spectroscopic platform for isotopic measurements of methane (CH₄). We identify and discriminate the ¹²CH₄- and ¹³CH₄ isotopologues and determine their mixing ratio. An Allan deviation analysis shows that the noise equivalent concentration for CH₄ is 200 ppt (pmol/mol) at 100 s of integration time, corresponding to a normalized noise equivalent absorption coefficient of 5.1×10^-9Wcm^-1Hz^-1/2, potentially making the PAS sensor a truly disruptive instrument for bedside monitoring using isotope tracers by providing real-time metabolism data to clinical personnel.

Calibration of Quartz-Enhanced Photoacoustic Sensors for Real-Life Adaptation

We report on the use of quartz-enhanced photoacoustic spectroscopy for continuous carbon-dioxide measurements in humid air over a period of six days. The presence of water molecules alters the relaxation rate of the target molecules and thus the amplitude of the photoacoustic signal. Prior to the measurements, the photoacoustic sensor system was pre-calibrated using CO₂ mole fractions in the range of 0–10⁻³ (0–1000 ppm) and at different relative humidities between 0% and 45%, while assuming a model hypothesis that allowed the photoacoustic signal to be perturbed linearly by H₂O content. This calibration technique was compared against an alternative learning-based method, where sensor data from the first two days of the six-day period were used for self-calibration. A commercial non-dispersive infrared sensor was used as a CO₂ reference sensor and provided the benchmark for the two calibration procedures. In our case, the self-calibrated method proved to be both more accurate and precise.

Fiber-Coupled Quartz-Enhanced Photoacoustic Spectroscopy System for Methane and Ethane Monitoring in the Near-Infrared Spectral Range

We report on a fiber-coupled, quartz-enhanced photoacoustic spectroscopy (QEPAS) near-IR sensor for sequential detection of methane (CH₄ or C1) and ethane (C₂H₆ or C2) in air. With the aim of developing a lightweight, compact, low-power-consumption sensor suitable for unmanned aerial vehicles (UAVs)-empowered environmental monitoring, an all-fiber configuration was designed and realized. Two laser diodes emitting at 1653.7 nm and 1684 nm for CH₄ and C₂H₆ detection, respectively, were fiber-combined and fiber-coupled to the collimator port of the acoustic detection module. No cross talk between methane and ethane QEPAS signal was observed, and the related peak signals were well resolved. The QEPAS sensor was calibrated using gas samples generated from certified concentrations of 1% CH₄ in N₂ and 1% C₂H₆ in N₂. At a lock-in integration time of 100 ms, minimum detection limits of 0.76 ppm and 34 ppm for methane and ethane were achieved, respectively. The relaxation rate of CH₄ in standard air has been investigated considering the effects of H₂O, N₂ and O₂ molecules. No influence on the CH₄ QEPAS signal is expected when the water vapor concentration level present in air varies in the range 0.6–3%.

Intrinsic Spectral Resolution Limitations of QEPAS Sensors for Fast and Broad Wavelength Tuning

Quartz-enhanced photoacoustic sensing is a promising method for low-concentration trace-gas monitoring due to the resonant signal enhancement provided by a high-Q quartz tuning fork. However, quartz-enhanced photoacoustic spectroscopy (QEPAS) is associated with a relatively slow acoustic decay, which results in a reduced spectral resolution and signal-to-noise ratio as the wavelength tuning rate is increased. In this work, we investigate the influence of wavelength scan rate on the spectral resolution and signal-to-noise ratio of QEPAS sensors. We demonstrate the acquisition of photoacoustic spectra from 3.1 μm to 3.6 μm using a tunable mid-infrared optical parametric oscillator. The spectra are attained using wavelength scan rates differing by more than two orders of magnitude (from 0.3 nm s-1 to 96 nm s-1). With this variation in scan rate, the spectral resolution is found to change from 2.5 cm-1 to 9 cm-1. The investigated gas samples are methane (in nitrogen) and a gas mixture consisting of methane, water, and ethanol. For the gas mixture, the reduced spectral resolution at fast scan rates significantly complicates the quantification of constituent gas concentrations.