Open Access
Add to BibSonomy
Abstract
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.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Chlorinated hydrocarbons, including trichloromethane (CHCl3, known as Chloroform) and dichloromethane (CH2Cl2), are commonly used as reagent, anesthetic, refrigerant and organic solvent in the industrial activities of foam plastic products, metal cleaning and pharmaceuticals production [1–4]. CHCl3 and CH2Cl2 are of concern as toxic air pollutants and carcinogen regulated by the international agency for research on cancer as 2A and 2B [5]. 30-300 ppm CHCl3 can lead to miscarriage in the pregnant mice [5] and CHCl3 can react with oxygen to produce highly toxic phosgene [6]. Breathing in CH2Cl2 vapour can produce narcotic effects including drowsiness, headache or unconsciousness and death at high concentrations [7]. Hence, real-time detection of chlorinated hydrocarbons vapors is very important for various environmental and industrial applications, particularly in confined spaces or where ventilation is inadequate.
Chloroform and dichloromethane as volatile organic compounds are usually analyzed by gas chromatograph–mass spectrometer (GC-MS) [1], electrochemical sensor [8,9] and optical absorption spectroscopy [4,10]. However, GC-MS technique requires time-consuming gas collection and preconcentration, extraction and separation of the samples prior to gas chromatography and mass spectrometry analysis, which leads to its application mostly applied in the lab [1]. Besides, GC-MS system is bulky and expensive for real time analysis. Electrochemical sensor with advantages of low-cost and portability are widely used in different industrial applications, while its detection is commonly interfered by the pollutants with same physic or chemical property [11]. Optical absorption spectroscopy is commonly used for substance analysis since it is a straight, noninvasive, in situ approach with advantages of fast response time, high selectivity and sensitivity compared with electrochemical sensor and GC-MS technique [12–14].
In last decades, high sensitivity optical absorption spectroscopies have been achieved by implementing multi-pass cells (White or Herriot cell) [12], modulation wavelength method [13] and high-finesse optical cavities [15]. However, each technique has its own limitations for multi-gas analysis or field measurement simultaneously guaranteed with high sensitivity and small size. The modulation wavelength method based on multi-pass cells highly depends on the background signal (not a zero-background approach) [16], and its measurement sensitivity is limited by the optical length in the optical cell. The high-finesse optical cavities composed of highly reflective dielectric mirrors is limited by effective wavelength band of dielectric mirrors and the use of relatively expensive high reflectivity mirrors requires careful protection [17]. Photoacoustic spectroscopy, with advantages of broadband operation, zero-background property and high sensitivity, is recognized as one of the best optical methods for gas analysis [18–20].
Photoacoustic spectroscopy is based on the detection of acoustic signals resulting from the light absorption of a modulated laser radiation by the target species (gas or particles) [21,22]. For chlorinated hydrocarbons detection, Mohebbifar et.al has used a CO2 laser in a wavelength range of 877 to 1086 cm−1 to realize a single species of CHCl3 measurement based a single acoustic resonator [4]. However, the light absorption of chlorinated hydrocarbons is not a Lorentzian profile that is the most suitable condition for optical gas spectroscopy [23]. For the complex molecules, the light absorption is companied with the broadening of absorption lines resulting in the structured absorption features over a wide spectral range, which increases the difficulty in the multi-gas analysis [23,24]. In the case of that each gas component does not spectrally interfered, multi-gas analysis is easily implemented. For the case of overlapping absorption spectra resulting from different gases, new advanced analysis tools approaches, such as multilinear regression (MLR) and partial least squares regression (PLSR), are needed to eliminate the measurement interference from other components [24–26]. MLR approach is reliable when dealing with uncorrelated variables, however in the prediction of correlated variables MLR can result in a lack of precision and accuracy [25]. Zifarefi et.al experimentally demonstrated that PLSR approach predicted gas concentrations with a calibration error up to 5 times better than MLR method in the case of overlapping absorption features of C2H2−CH4−N2O by exploiting quartz-enhanced photoacoustic spectroscopy (QEPAS) [25]. Among different advanced analysis tools, a linear combination approach of reference spectra still presents as a straightforward strategy and experimentally demonstrated by Sampaolo et.al through simultaneous detection methane, ethane and propane based on a QPEAPS sensor system in the spectral range 3.342–3.349 µm [26].
To realize multi-gas analysis, photoacoustic spectroscopy has exploited different approaches such as multi-acoustic resonators [27] or multi-laser coupled together [25,26,28] or a single laser covering several absorption lines of multi-gas species [29]. The methods of multi-acoustic resonators promote the complexity of sensing architecture and improve the sensor cost by requiring several sets of laser controller, signal process module such as lock-in amplifier [23]. In the approach of multi-laser coupled, the lasers can be operated in sequence, or simultaneously in a case of a two-component mixture with isolated absorbing features spectrally far from each other [24,28]. For multi-gas detection, single light source operating in the near or mid-infrared spectrum must be broadband for covering different absorption lines. Although trace gases in the mid-infrared spectrum have stronger absorption than those in the near-infrared band, near-infrared laser sources are more suitable in gas sensing due to their advantages of low cost and operation at room temperature [24,30]. In this work, a low-cost DFB laser that cover the light absorption of CHCl3 and CH2Cl2 in the near-infrared band was used for simultaneous measurement of CHCl3 and CH2Cl2.
In this paper, we report a compact photoacoustic spectrophone for simultaneous measurement of CHCl3 and CH2Cl2 based on a single acoustic resonator using a low-cost DFB laser emitting at 1684 nm. The feasibility of simultaneous measurement of CHCl3 and CH2Cl2 was investigated with strongly overlapping absorption band of CHCl3 and CH2Cl2. The detection inference from each component was investigated for accurately retrieving the concentration. To make the system compact, all analog signals for controlling diode laser and reference signal for lock-in amplifier were generated in a board-type data acquisition (DAQ) card. All key elements were integrated into a 40 cm × 40 cm × 20 cm chassis, which is well suitable for field campaign.
2. Experimental setup
2.1 Photoacoustic spectroscopy
Photoacoustic spectroscopy (PAS), as one of versatile photonic technology with advantages of its ease of use and relatively low cost, has the capability of trace gas measurement at the sub-ppb level [31]. This technique is based on photoacoustic effect with the generation of sound waves due to the absorption of modulated light in all materials [19,31]. The photoacoustic signal depends on laser power, microphone sensitivity and PA cell response constant described as following equation [32]:
where Mmic is the microphone sensitivity in mV/Pa, P is the modulated optical power of light radiation source, in watts (W). PA cell response constant, Ccell, has units of Pascal per inverse centimeters per watt (Pa/cm−1 W), Ntot is the total number density of molecules (molecule/cm3), c and σ are the concentration and absorption cross section of the analyte, respectively.In the wavelength modulation photoacoustic spectroscopy (WM-PAS), for achieving the harmonic detection, a sinusoidal waveform is usually used to modulate the injection current of light source at a resonant frequency of PA cell. The amplitude of sinusoidal waveform is proportional to the modulated exciting optical power (P). In WM-PAS, the laser frequency is scanned using a slow current ramp that adds to the sinusoidal signal. For multi-component detection using a single laser source, the slow current ramp should cover several absorption lines of multi-gas and the modulated amplitude of sinusoidal signal is used to excite photoacoustic signal for high sensitivity detection. Therefore, the amplitude of the current ramp and sinusoidal should be optimized for realizing multi-gas analysis and highly sensitive measurement.
2.2 Selection of absorption lines and characterization of the excitation source
HITRAN database was used to simulate the absorbance of chloroform and dichloromethane in the range of 1682 to 1685 nm for precise detection [33]. Figure 1 shows the simulation absorbance of 100 ppm CHCl3 and CH2Cl2, 20000 ppm H2O, 10 ppm CH4 with temperature of 300 K, air pressure of 1 atm and absorption path length of 1000 cm. 20000 ppm H2O at temperature of 298 K and air pressure of 1 atm corresponds to the relative humidity of 62% [34]. The regular CH4 concentration in the atmosphere is around 2 ppm. As shown in Fig. 1, CHCl3 presents a strong absorption peak with a 10−19 cm2/molecule at 1683.1 nm, which is beneficial to highly sensitive measurement of chloroform. The absorption lines of CHCl3 and CH2Cl2 around the wavelength of 1684.9 nm can be used for simultaneous measurement of CHCl3 and CH2Cl2 using two partially merged absorption peaks. The potential interference of atmospheric species (H2O and CH4) at 1684.9 nm will not affect the measurement of CHCl3 and CH2Cl2.
Fig. 1. Simulated absorbance spectra of 100 ppm CHCl3, 100 ppm CH2Cl2, 20000 ppm H2O, and 10 ppm CH4 with temperature of 300 K, air pressure of 1 atm and absorption path length of 1000 cm. The HITRAN database is used.
The wavelength space between the two absorption peaks shown in Fig. 1 is 0.07 nm, which is less than the spectral range of a single distribute feedback (DFB) laser through one periodical frequency scanning. A DFB laser (NEL, Tokyo, Japan) emitting at the wavelength of 1684 nm was used as the light source. Figure 2 presents the wavelength of the DFB laser at different operating temperatures from 25 to 32 °C with injection currents from 60 to 145 mA. The output light power at a temperature of 30°C was gradually increased from 7 mW to 20 mW with the injection current from 60 to 150 mA as shown in Fig. 2. The emitting wavelength of light source at the temperature of 30 °C is suitable for the target absorption spectral region for simultaneous measurement of CHCl3 and CH2Cl2. Therefore, an injection current of 60 to 145 mA combined with a 30 °C operation temperature was used for the simultaneous measurement of CHCl3 and CH2Cl2.
Fig. 2. The lasing wavelength and light power as the injection current varies at the operation temperature of 30 °C.
2.3 Compact photoacoustic spectrophone
Figure 3(a) describes the architecture of the developed compact PA spectrophone. A DFB (NEL, Tokyo, Japan) laser operating at 1684 nm was selected as radiation source for simultaneous detection of dichloromethane and trichloromethane. The temperature and current of light source were precisely controlled by a compact laser diode controller (ITC102, Thorlabs) integrated on a single circuit board with size of 100 × 160 mm for covering the absorption lines of CHCl3 and CH2Cl2. The temperature of light source was set to 30 °C around room temperature. The injection current of light source was composed of a sawtooth wave and a sinusoidal waveform, which were generated by a LabVIEW-controlled data acquisition (DAQ) card (USB 6211, NI) with a dimension of 90 × 60 mm. The sawtooth wave signal was utilized to scan the wavelength of light source for covering the absorption lines of CHCl3 and CH2Cl2. The sinusoidal waveform was used to modulate the light intensity of light source at a fundamental frequency of acoustic resonator. A single-board lock-in amplifier (LIA-BVD-150-L, Femto) with dimension size of 100 × 50 mm was deployed to demodulate the PA signal at the fundamental frequency of acoustic resonator. A homemade noise-reduction band-pass filter with effective operational frequency range from 3000 to 5000 Hz was used to amplify PA signal before connecting to the lock-in amplifier. An industrial laptop was used to perform data processing and result display via a Labview interface. A fan installed at the back of PA sensor was used to dissipate heat released from electronic elements. Figure 3(b) presents the developed compact photoacoustic sensing system with dimension of 40 cm × 40 cm × 20 cm and the internal construction with all optical and electronic elements.
Fig. 3. (a) Schematic view of the PA spectrophone. (b) The developed compact photoacoustic sensing system and its internal construction. DAQ: data acquisition card, MFC: mass flow controller, LIA: lock-in amplifier, LaCon: laser controller, f1: focus lens.
For investigating the detection inference from each component, a gas path shown in the blue square in Fig. 3(a) was designed with a total flow rate of 0.4 L/min. The gas line was separated into three channels: one channel for CHCl3 gas with a constant flow rate of 0.2 L/min controlled by MFC1 (Mass flow controller), the last two channels for CH2Cl2 gas and N2 with a total flow rate of 0.2 L/min controlled by MFC2 and MFC3. Therefore, the CHCl3 concentration in the mixture was kept constant and the CH2Cl2 concentration can be changed and controlled by the MFC2, for investigating the detection inference of CH2Cl2 on the measurement of CHCL3. The position of channels for CHCl3 and CH2Cl2 gas were exchanged for investigating the detection inference of CHCl3 on the measurement of CH2Cl2.
A cylindrical acoustic resonator with length of 23 mm and diameter of 6 mm was used to amplify the acoustic signal that can be measured using microphone in the PA sensing system. Four electret microphones (EK-23329-P07, Knowles) located at the middle of cylindrical resonator were used for detecting PA signal with a sensitivity of 22.4 mV/Pa. Figure 4 shows the experimental photoacoustic signals due to chloroform absorption at different modulation frequencies. A 1st longitudinal resonant frequency of 4580 Hz was experimentally observed. The frequency response was fitted by a Lorentz profile associated with a full width half maximum (FWHM) of 327 Hz, which leads to a quality factor Q of 14 (=4580/327).
3. Individual measurement of dichloromethane and trichloromethane
The performance of the developed PA spectrophone was firstly evaluated by individual measurements of CHCl3 and CH2Cl2 for optimizing the system settings including the selection of PA signal component, detection phase, current modulation amplitude before analyzing the gas mixture.
3.1 Characterization of the photoacoustic signal profile
Lock-in amplifier is a key device capable of recovering harmonic signal in the photoacoustic spectrophone system. However, the demodulated signal like quadrature components (X and Y) of PA signal in the lock-in amplifier are easily influenced by the signal phase shift [35]. The phase shift of PA signal is usually introduced by electronics, laser, microphone and the relaxation rates of different target molecules [36]. For eliminating the influence of phase shift, the quadrature components (X and Y) can be squared and summed to provide an estimated signal amplitude R = sqrt(X2 + Y2). In this section, the quadrature components (X, Y) and R value of PA signals at different concentration of CHCl3 and CH2Cl2 are characterized for accurate detection.
In the detection of harmonic waveforms, the theoretical amplitude of harmonic signal normally decreases with the increase of harmonic order [37]. The odd harmonic signal at the center of gas absorption line is close to zero and the two peaks of the odd harmonic signals is symmetry [37]. Figure 5(a), (b) show the recorded quadrature components X, Y at different concentrations of CHCl3 from 200 to 800 ppm and CH2Cl2 from 200 to 400 ppm with an optimized detection phase of 5° for maximizing the quadrature components X signal. The black line in Fig. 5(a) and (b) is the background signal. Figure 5(c) presents the corresponding R value of PA signal with different concentrations of CHCl3 and CH2Cl2.
Fig. 5. Quadrature components X (a), Y (b) and R (c) of photoacoustic 1-f signal with different CHCl3 and CH2Cl2 concentrations with a lock-in amplifier integration time of 500 ms. The sampling points correspond to different tuning points of wavelength.
Figure 5(c) indicates that the center point of 1st harmonic signal (1f) caused by CHCl3 light absorption (red line) is located at the sampling point 4200 (corresponding to 1684.85 nm) close to zero background. The signal amplitude of two peaks at the sampling point 3000 and 4800 are not symmetry, which is due to that the gas absorption of CHCl3 is not a convention voigt profile usually used to describe the broadening of gas absorption under atmosphere pressure. The different offset of R signals varying with CHCl3 concentration is caused by the broad light absorption at the wavelength far away from the peak. The photoacoustic signal caused by CH2Cl2 light absorption has the same features in the asymmetric peaks and the varied offset as shown in Fig. 5(c). Although R signal has advantage of removing the effect of phase shift, the recorded R signals of CHCl3 and CH2Cl2 are seriously emerged, which affects the accuracy of the measurement of CHCl3 and CH2Cl2. The profile of quadrature components X signal is less twisted than the profile of R signal, which is beneficial to calculate the concentrations of CHCl3 and CH2Cl. As shown in Fig. 5(a), the quadrature component X signal of different CH2Cl2 concentrations were intersected at the sampling points 570, 7500 and 9660, where we can retrieve CHCl3 concentrations without the interference of CH2Cl2 using this special characterization of the absorption profile.
3.2 Modulation amplitude optimization
For scanning the spectral range of CHCl3 and CH2Cl2 shown in Fig. 1, the operating temperature of DFB light source was set at 30 °C. The injection current was tuned from 65 mA to 145 mA, which was determined by a current summation between the current setpoint of laser controller and the injection signal amplitude (the ramp wave and sinusoidal waveform). The current setpoint in the laser controller was set to be 105 mA. The amplitude of ramp wave was 750 mV, corresponding to a current of 30 mA (=750 mV*40 mA/V) after multiplying with a transfer function coefficient of 40 mA/V in the laser controller. The frequency of the ramp wave was 50 mHz representing a period of 20 s for scanning the laser wavelength. The amplitude of sinusoidal waveform at a frequency of 4600 Hz was optimized with a variation from 40 mV to 220 mV corresponding to an injection current range from 1.6 to 8.8 mA. The recorded PA signal was the quadrature components X-1f signal demodulated by lock-in amplifier with an integration time constant of 500 ms.
To determine the optimal modulation amplitude for CHCl3 detection, the compact sensor was evaluated by a measurement of CHCl3 with a constant concentration of 620 ppm. Figure 6(a) shows the recorded PA signal (X-1f signal) for CHCl3 detection with different current modulation amplitudes from 40 to 220 mV. Figure 6(b) plots the relationships of PA signal peak values with 3100 sampling points on the different modulation amplitude. A highest peak value of 2.4 V with a modulation amplitude of 220 mV was obtained. As shown in Fig. 6(b), the peak values were increased linearly with the modulation amplitude from 40 to 180 mV and it wasn’t improved largely with the modulation amplitude more than 180 mV. In consideration of that the low modulation amplitude of 160 mV is beneficial to laser lifetime and also generate high PA signal, the modulation amplitude of 160 mV was used for CHCl3 detection.
Fig. 6. (a) Recorded PA spectral signals of CHCl3 with different current modulation amplitudes. (b) Peak values of the PA signals as the current modulation amplitude increases.
Figure 7 shows the measured spectrum of CH2Cl2 in a fixed concentration of 1030 ppm with different current modulation amplitude from 40 to 220 mV. The peak of CH2Cl2 spectrum is located at the sampling point 4480, which presents a distinguish absorption feature compared with the spectrum of CHCl3. The peak value of CH2Cl2 spectrum was linearly improved with the amplitude of current modulation as shown in Fig. 7(b). Due to the limit of maximum operation current for the light source, the photoacoustic signal with high amplitude of current modulation was not investigated. In order to simplify the PA sensing system, the system parameters including the amplitude and frequency of the sinusoidal waveform for the measurement of CH2Cl2 were kept the same as the setting in the detection of CHCl3.
Fig. 7. (a) Recorded PA spectral signals of CH2Cl2 with different current modulation amplitudes; (b) Peak values of the PA signals as the current modulation amplitude increases.
3.3 PA signal linearity
The instrument response to different CHCl3 and CH2Cl2 concentrations was evaluated by operating the light source at a temperature of 30 °C with a current modulation amplitude of 160 mV. The quadrature components X-1f signals from lock-in amplifier with a detection phase of 5° and an integration time of 500 ms were recorded. Two standard reference concentrations of CHCl3 in 620 ppm and CH2Cl2 in 1030 ppm were diluted by pure nitrogen (N2) to generate different concentrations from 100 to 1000 ppm using two mass flow controllers.
Figure 8(a) plots the recorded CHCl3 and CH2Cl2 spectrum related to different concentrations, ranging from 103 to 620 ppmv and 206 to 1030 ppmv, respectively. The relationship of each peak values of the CHCl3 and CH2Cl2 spectrum are plotted in Fig. 8(b) and (c). A saturation effect of PA signal reported by previous investigations [38,39] due to high concentration of the measured analytes was observed. In the case of CHCl3, the peak value of PA spectrum started to decrease at a concentration of 620 ppmv and deviated to the linearity: SPA,3100 = 3.15· CCHCl3+36 with a regression coefficient of 0.999 fitted by the peak values at the sampling point 3100 (SPA,3100) as a function of different CHCl3 concentration (CCHCl3) from 206 to 413 ppmv. The linearity of photoacoustic response from CHCl3 light absorption enabled an effective detection range from 0 to 400 ppmv in the compact PA spectrophone. For the detection of CH2Cl2, a linear fit of peak values of the CH2Cl2 spectrum at the sampling point 4480 (SPA,4480) versus on different concentration from 0 to 824 ppmv: SPA,4480 =2.61·CCH2Cl2+ 8 was achieved with a regression coefficient of 0.999, which determined an effective measurement range of CH2Cl2 from 0 to 800 ppmv. The saturation effect of PA signal due to high concentration of CH2Cl2 was also observed as shown in Fig. 8(c).
Fig. 8. (a) Photoacoustic signals for different CHCl3 concentrations (from 103 to 620 ppmv) and different CH2Cl2 concentrations (from 206 to 1030 ppmv) in pure nitrogen. Linear fit of the PA signal peak values for CHCl3 (b) and CH2Cl2 (c) detection as the gas concentration increases.
4. Simultaneous measurements of CHCl3 and CH2Cl2 in a gas mixture
The compact PA spectrophone has a good ability in the measurement of single component of CHCl3 or CH2Cl2. However, a gas mixture of the two components has a strong spectral overlap that seriously affects the measurement accuracy. As observed in Fig. 8(a), the quadrature component X signal of different CHCl3 concentrations were intersected at sampling point 4330, which is constant and located at the background line. This phenomenon indicates that the variation of PA signal at sampling point 4330 can be used for retrieving CH2Cl2 concentration without the interference from CHCl3. For the estimation of CHCl3 concentration, the quadrature component X signal of different CH2Cl2 concentrations were intersected at sampling points 570, 7500 and 9660, where we can retrieve CHCl3 concentrations. To improve the instrument signal-to-noise ratio (SNR), the peak values of the CHCl3 PA signal at the sampling point 3100 were also used for retrieving CHCl3 concentrations after removing the interference from CH2Cl2. Therefore, multilinear regression method was used for simultaneous measurements of CHCl3 and CH2Cl2 with the special characterization of absorption profile.
Figure 9(a) and (b) plot the linear relationships of PA signals at sampling points 4330 and 3100 as CH2Cl2 concentration varies, which can be fitted with the following equations both with a regression coefficient of r2 = 0.999:
where SPA,4330, SPA,3100 are the PA signals at sampling points 4330 and 3100 due to CH2Cl2 light absorption. CCH2Cl2 are the estimated CH2Cl2 concentration. Equation (2) was used to estimate the CH2Cl2 concentration in the developed PA sensor. With the determined CH2Cl2 concentration (CCH2Cl2), the PA signal at sampling point 3100 (SPA,3100) caused by CH2Cl2 light absorption could be removed for accurate estimation of CHCl3 concentration. After removing the influence of CH2Cl2, the CHCl3 concentration was determined by the following equation using the linear fit of the peak values on the CHCl3 concentration (SPA,3100 = 3.15*CCHCl3+36) shown in Fig. 8(b)):Fig. 9. Lear relationships of PA signal at the sampling point of 4330 (a) and 3100 (b) on different CH2Cl2 concentrations.
Table 1 summaries other several linear relationships of PA signals (SCHCl3,570, SCHCl3,7500, SCHCl3,9660) at sampling points 570, 7500 and 9660 on different CHCl3 concentrations with regression coefficient of 0.999,0.998,0.996, respectively. The CHCl3 concentration was determined by averaging the estimated CHCl3 concentrations using the linear fits of PA signals at sampling points 570, 3100, 7500 and 9660 as the following equations:
where CCHCl3,570, CCHCl3,7500, CCHCl3,9660 are the estimated CHCl3 concentration using the linear fit of PA signal at sampling points 570, 7500 and 9660 shown in Table 1.
Table 1. Linear Relationships of PA Signals (SCHCl3,570, SCHCl3,7500, SCHCl3,9660) at the Sampling Points of 570, 7500 and 9660 on Different CHCl3 Concentrations.
The validation test of this multi-linear regression method was carried out through simultaneous measurement of CHCl3 and CH2Cl2. The potential interference from each component was investigated. To produce different concentrations of CHCl3 and CH2Cl2, a reference CHCl3 gas (1030 ppmv) and a CH2Cl2 gas (620 ppmv) were diluted by nitrogen. For investigating the influence of CH2Cl2 on CHCl3, a flow gas mixture with constant CHCl3 concentration of 310 ppmv and varying CH2Cl2 concentration from 75 to 515 ppmv was generated and sampled in the PA cell. A constant CH2Cl2 concentration companied with different CHCl3 concentrations from 75 to 300 ppmv was also produced for establishing the influence of CHCl3 on CH2Cl2.
Figure 10(a) shows time series measurement of constant CHCl3 concentrations and varying CH2Cl2 concentration from 75 to 515 ppmv with an acquisition time of 20 s. The red line in Fig. 10(a) presents the retrieved CHCl3 concentrations, which were constant without fluctuation with different CH2Cl2 concentrations. Figure 10(b) presents the time series measurement of constant CHCl2 concentration and varying CHCl3 concentration. The determined CH2Cl2 concentration was kept constant in 350 ppmv and 515 ppmv, which indicates that different CHCl3 concentrations would not affect the measurement of CH2Cl2. The actual concentrations shown in Fig. 10 were determined by the ratio between the flow rate of each component to the total flow rate. A slight shift of the estimated constant CHCl3 and CH2Cl2 concentration (Grey line in Fig. 10(a) and (b)) might be caused by the sample preparation system.
Fig. 10. (a) Time series measurements results with constant CHCl3 concentration of 310 ppmv and CH2Cl2 concentration varying from 75 to 500 ppmv. (b) Time series measurements results with constant CH2Cl2 concentrations of 350 ppmv and 515 ppmv, and varying CHCl3 concentration from 75 to 300 ppmv.
The instrument parameters, including the detection limit, measurement accuracy and precision were evaluated by continuous measurement of CHCl3 in a fixed concentration of 310 ppmv and CH2Cl2 in 515 ppmv. Allan deviation of analysis was performed to evaluate the stability of the developed PA spectrophone. Figures 11(a) and (b) shows the retrieved concentrations of CHCl3 using the Eq. (5) and CH2Cl2 in Eq. (2) with an acquisition time of 20 s. The detection limit (1σ) was determined to be 6.3 ppmv for CHCl3 and 6.1 ppmv for CH2Cl2, deduced from the standard deviation of the data shown in Figs. 11(a) and (b). The corresponding normalized noise equivalent absorption (NNEA) coefficient of PA spectrophone was found as 3.20×10−8 W·cm−1·Hz−1/2 for CHCl3 and 3.58×10−8 W·cm−1·Hz−1/2 for CH2Cl2.
Fig. 11. Time series measurements results with constant CHCl3 concentration of 515 ppmv (a) and CH2Cl2 concentration of 310 ppmv (b). Allan deviation analysis for CHCl3 (c) and CH2Cl2 (d) plot as a function of averaging (integration) time. Red line represents white noise regime. Histogram showing the distribution of the estimated CHCl3 concentration (e) and CH2Cl2 concentration (f) with Gaussian profile fits.
Figure 11(c) and (d) plot the Allan deviation of the estimated concentration shown in Fig. 11(a) and (b) versus the average time for CHCl3 and CH2Cl2, respectively. The Allan deviation curves in Fig. 11(c) and (d) indicates an optimum sensing performance of the PA instrument: minimum Allan deviation (2 ppmv) for CHCl3 were obtained with optimum averaging (integration) time of 300 s and minimum Allan deviation (1 ppmv) for CH2Cl2 with optimum averaging (integration) time of 900 s, respectively. Figures 11(e) and (f) show the distribution histogram of the retrieved concentrations of CHCl3 and CH2Cl2 in the PA cell, accompanied with Gaussian profile fits. The value of the mean concentration of CHCl3 (Fig. 11(c)) was obtained to be 304 ppmv from the measurement of 310 ppmv CHCl3 inside the PA cell, which results in a measurement accuracy of 6 ppmv with a measurement uncertain of 1.9% and the measurement precision was deduced to be 9 ppmv (with an average time of 20 s) from the FWHM of the fitted Gaussian profile. Based on Fig. 11(d) with the measurement of 515 ppmv CH2Cl2, a measurement accuracy of 31 ppmv with an uncertain of 6.0% was obtained with a precision of 13 ppmv according to the mean concentration of 484 ppmv and the fit FWHM of 13 ppmv.
In the photoacoustic sensing system, the measurement uncertainty can be introduced by the signal phase shift due to the electronics elements and the relaxation rates of different target molecules, the fluctuation in laser power intensity, environmental acoustic noise and stray-light in the PA cell [35,36,40–42]. Another important factor for deteriorating the measurement accuracy is the hardly recognizable absorption features of different components as reported previous investigations [24,25]. Although several new advanced analysis tools such as PLSR approach had been applied in the overlapping absorption spectra for multi-gas analysis, the measurement accuracy was not highly improved [24,25]. In the future work, the instrument performance can be further improved by using higher light power of laser source, optimized PA cell immune from environmental acoustic noise and accurate phase information originating from different gas species with overlapping absorption spectra.
5 Conclusion
In this work, a compact photoacoustic spectrophone for simultaneous measurement of dichloromethane and trichloromethane gas was developed and validated by real-time detection of different concentrations. The compact photoacoustic sensor was realized using a DFB laser emitting at 1684 nm for scanning the absorption of two-gas mixture. Due to the strongly overlapping absorption of each component, multi-linear regression method was used for fitting the experimental photoacoustic spectra signals in the gas mixture. The performance of PA sensor was evaluated for individual component detection with an optimized current modulation amplitude and detection phase. The detection limit, measurement accuracy and precision were experimentally determined to be 6.3 ppmv, 1.9%, 9 ppmv for CHCl3 and 6.1 ppmv, 6.0%, 13 ppmv for CH2Cl2, respectively, which are sufficient for different industrial applications. The interference of CHCl3 and CH2Cl2 from each component was investigated through changing the concentration of each species. This sensor system shows the advantages of simultaneous multiple chlorinated hydrocarbons gas species with effective low-cost and reduced size.
Funding
Natural Science Foundation of Zhejiang Province (LQ22F050014); Key Research and Development Program of Zhejiang Province (2021C03178); NingboTech University (20201203Z0196).
Acknowledgments
G.W. and T.Z. designed/modified the experimental setup, G.W. performed the experiments and analyzed the data. G.W. wrote the manuscript. S.H. supervised the work and finalized the manuscript. All authors have read and agreed to the published version of the manuscript. We acknowledge the financial support from Zhejiang Provincial Natural Science Foundation of China (Grant No. LQ22F050014), Key Research and Development Program of Zhejiang Province (Grant No. 2021C03178), NingboTech University (Grant No. 20201203Z0196).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. Y. Wang, H. Han, T. Chang, X. Liu, Q. Zhu, F. Liu, Y. Yan, C. Shen, and Y. Chu, “A novel method for measuring the concentration of chloroform based on kinetic parameters at atmosphere,” Chem. Phys. Lett. 653, 5–10 (2016). [CrossRef]
2. R. Hossaini, M. P. Chipperfield, S. A. Montzka, A. A. Leeson, S. S. Dhomse, and J. A. Pyle, “The increasing threat to stratospheric ozone from dichloromethane,” Nat. Commun. 8(1), 15962 (2017). [CrossRef]
3. P. G. Simmonds, A. J. Manning, D. M. Cunnold, A. McCulloch, S. O’Doherty, R. G. Derwent, P. B. Krummel, P. J. Fraser, B. Dunse, L. W. Porter, R. H. J. Wang, B. R. Greally, B. R. Miller, P. Salameh, R. F. Weiss, and R. G. Prinn, “Global trends, seasonal cycles, and European emissions of dichloromethane, trichloroethene, and tetrachloroethene from the AGAGE observations at Mace Head, Ireland, and Cape Grim, Tasmania,” J. Geophys. Res. 111(D18), D18304 (2006). [CrossRef]
4. M. R. Mohebbifar, “High-sensitivity detection and quantification of CHCl3 vapors in various gas environments based on the photoacoustic spectroscopy,” Microw. Opt. Technol. Lett. 61(9), 2234–2241 (2019). [CrossRef]
5. List of Classifications – IARC Monographs on the Identification of Carcinogenic Hazards to Humans. Monographs.iarc.who.int. (2020). Retrieved 4 February 2021, from https://monographs.iarc.who.int/agents-classified-by-the-iarc/
6. T. Zhang, Y. Xing, G. Wang, and S. He, “High Sensitivity Continuous Monitoring of Chloroform Gas by Using Wavelength Modulation Photoacoustic Spectroscopy in the Near-Infrared Range,” Appl. Sci. 11(15), 6992 (2021). [CrossRef]
7. K. Dowker, “Real time measurement of dichloromethane containing mixtures,” Research report by Health and Safety Excutive 901-1000, RR982 (2013).
8. P. Kar and A. Choudhury, “Carboxylic acid functionalized multi-walled carbon nanotube doped polyanniline for chloroform sensor,” Sens. Actuators, B 183, 25–33 (2013). [CrossRef]
9. D. S. Chandrasakaran, I. Nainggolan, N. Derman, and T. Ikhsan, “Chloroform gas sensor based on chitosan biopolymer,” Appl. Mech. Mater. 679, 45–49 (2014). [CrossRef]
10. B. Jeffries, G. A. Raiche, and L. E. Jusinski, “Detection of chlorinated hydrocarbons via laser-atomization laser-induced fluorescence,” Appl. Phys. B 55(1), 76–83 (1992). [CrossRef]
11. D. W. Hatchett and M. Josowicz, “Composites of intrinsically conducting polymers as sensing nanomaterials,” Chem. Rev. 108(2), 746–769 (2008). [CrossRef]
12. Z. Du, S. Zhang, J. Li, N. Gao, and K. Tong, “Mid-Infrared Tunable Laser-Based Broadband Fingerprint Absorption Spectroscopy for Trace Gas Sensing: A Review,” Appl. Sci. 9(2), 338 (2019). [CrossRef]
13. J. Hodgkinson and R. P. Tatam, “Optical gas sensing: a review,” Meas. Sci. Technol. 24(1), 012004 (2013). [CrossRef]
14. W. Chen, H. Yi, T. Wu, W. Zhao, C. Lengignon, G. Wang, E. Fertein, and C. Coeur, “Photonic sensing of reactive atmospheric species,” Encyclopedia of Analytical Chemistry, 2017.
15. G. Gagliardi and H. P. Loock, Cavity-Enhanced Spectroscopy and Sensing, 2014, Publisher Name Springer, Berlin, Heidelberg,Heidelberg.
16. J. Li, B. Yu, W. Zhao, and W. Chen, “A review of signal enhancement and noise reduction techniques for tunable diode laser absorption spectroscopy,” Appl. Spectrosc. Rev. 49(8), 666–691 (2014). [CrossRef]
17. H. Ma, M. Sun, S. Zhan, Q. Zhang, S. Zha, G. Wang, K. Liu, M. Yi, and Z. Cao, “Compact dual-gas sensor for simultaneous measurement of atmospheric methane, and water vapor using a 3.38 mantimonide-distributed feedback laser diode,” Spectrochim. Acta, Part A 226, 117605 (2020). [CrossRef]
18. S. Manohar and D. Razansky, “Photoacoustics: a historical review,” Adv. Opt. Photonics 8(4), 586 (2016). [CrossRef]
19. L. Liu, A. Mandelis, H. Huan, and K. H. Michaelian, “Step-scan differential Fourier transform infrared photoacoustic spectroscopy (DFTIR-PAS): a spectral deconvolution method for weak absorber detection in the presence of strongly overlapping background absorptions,” Opt. Lett. 42(7), 1424 (2017). [CrossRef]
20. L. Liu, H. Huan, W. Li, A. Mandelis, Y. Wang, L. Zhang, X. Zhang, X. Yin, Y. Wu, and X. Shao, “Highly sensitive broadband differential infrared photoacoustic spectroscopy with wavelet denoising algorithm for trace gas detection,” Photoacoustics 21, 100228 (2021). [CrossRef]
21. A. Miklós and P. Hess, “Application of acoustic resonators in photoacoustic trace gas analysis and metrology,” Rev. Sci. Instrum. 72(4), 1937–1955 (2001). [CrossRef]
22. L. Liu, H. Huan, A. Mandelis, L. Zhang, C. Guo, W. Li, X. Zhang, X. Yin, X. Shao, and D. Wang, “Design and structural optimization of T-resonators for highly sensitive photoacoustic trace gas detection,” Opt. Laser Technol. 148, 107695 (2022). [CrossRef]
23. F. Merkt and M. Quack, “Molecular Quantum Mechanics and Molecular Spectra, Molecular Symmetry, and Interaction of Matter with Radiation,” 2011. John Wiley & Sons, Ltd.
24. A. Sampaolo, P. Patimisco, M. Giglio, A. Zifarelli, H. Wu, L. Dong, and V. Spagnolo. , “Quartz-enhanced photoacoustic spectroscopy for multi-gas detection: A review,” Analytica Chimica Acta 338894 (2021).
25. A. Zifarelli, M. Giglio, G. Menduni, A. Sampaolo, P. Patimisco, V. M. N. Passaro, H. Wu, L. Dong, and V. Spagnolo, “Partial least-squares regression as a tool to retrieve gas concentrations in mixtures detected using quartz-enhanced photoacoustic spectroscopy,” Anal. Chem. 92(16), 11035–11043 (2020). [CrossRef]
26. A. Sampaolo, S. Csutak, P. Patimisco, M. Giglio, G. Menduni, V. Passaro, F. K. Tittel, M. Deffenbaugh, and V. Spagnolo, “Methane, ethane and propane detection using a compact quartz enhanced photoacoustic sensor and a single interband cascade laser,” Sens. Actuators, B 282, 952–960 (2019). [CrossRef]
27. K. Liu, J. Mei, W. Zhang, W. Chen, and X. Gao, “Multi-resonator photoacoustic spectroscopy,” Sens. Actuators, B 251, 632–636 (2017). [CrossRef]
28. Y. Jiang, T. Zhang, G. Wang, and S. He, “A Dual-Gas Sensor Using Photoacoustic Spectroscopy Based on a Single Acoustic Resonator,” Appl. Sci. 11(11), 5224 (2021). [CrossRef]
29. Y. Ma, R. Lewicki, M. Razeghi, and F. Tittel, “QEPAS based ppb-level detection of CO and N2O using a high power CW DFB-QCL,” Opt. Express 21(1), 1008–1019 (2013). [CrossRef]
30. W. Zeller, L. Naehle, P. Fuchs, F. Gerschuetz, L. Hildebrandt, and J. Koeth, “DFB lasers between 760 nm and 16 µm for sensing applications,” Sensors 10(4), 2492–2510 (2010). [CrossRef]
31. M. W. Sigrist, Air Monitoring by Spectroscopic Techniques, JohnWiley and Sons, New York, USA, 1994.
32. T. Yang, W. Chen, and P. Wang, “A review of all-optical photoacoustic spectroscopy as a gas sensing method,” Appl. Spectrosc. Rev. 56(2), 143–170 (2021). [CrossRef]
33. I. E. Gordon, L. S. Rothman, C. Hill, R. V. Kochanov, Y. Tan, P. F. Bernath, M. Birk, V. Boudon, A. Campargue, K. V. Chance, B. J. Drouin, J. M. Flaud, R. R. Gamache, J. T. Hodges, D. Jacquemart, V. I. Perevalov, A. Perrin, K. P. Shine, M. A. H. Smith, J. Tennyson, G. C. Toon, H. Tran, V. G. Tyuterev, A. Barbe, A. G. Császár, V. M. Devi, T. Furtenbacher, J. J. Harrison, J. M. Hartmann, A. Jolly, T. J. Johnson, T. Karman, I. Kleiner, A. A. Kyuberis, J. Loos, O. M. Lyulin, S. T. Massie, S. N. Mikhailenko, N. Moazzen-Ahmadi, H. S. P. Müller, O. V. Naumenko, A. V. Nikitin, O. L. Polyansky, M. Rey, M. Rotger, S. W. Sharpe, K. Sung, E. Starikova, S. A. Tashkun, J. Vander Auwera, G. Wagner, J. Wilzewski, P. Wcisło, S. Yu, and E. J. Zak, “The Hitran2016 Molecular Spectroscopic Database,” J. Quant. Spectrosc. Radiat. Transfer 203, 3–69 (2017). [CrossRef]
34. S. A. Bell, “A guide to the Measurement of Humidity,” Technical Report. NPL, Teddington, 1996.
35. G. Wang, S. Reboul, J. B. Choquel, E. Fertein, and W. Chen, “Circular Regression in a Dual-Phase Lock-In Amplifier for Coherent Detection of Weak Signal,” Sensors 17(11), 2615 (2017). [CrossRef]
36. A. A. Kosterev, Y. A. Bakhirkin, F. K. Tittel, S. Blaser, Y. Bonetti, and L. Hvozdara, “Photoacoustic phase shift as a chemically selective spectroscopic parameter,” Appl. Phys. B 78(6), 673–676 (2004). [CrossRef]
37. Z. L. Wang, C. W. Tian, Q. Liu, J. Chang, Q. D. Zhang, and C. G. Zhu, “Wavelength modulation technique-based photoacoustic spectroscopy for multipoint gas sensing,” Appl. Opt. 57(11), 2909 (2018). [CrossRef]
38. D. C. Ddumitras, D. C. Dutu, C. Matei, A. M. Magureanu, M. Petrus, and C. Popa, “Laser photoacoustic spectroscopy: principles, instrumentation, and characterization,” J. Optoelectron. Adv. M. 9(12), 3655–3701 (2007).
39. F. Yehya and A. K. Chaudhary, “Designing and Modeling of Efficient Resonant Photoacoustic sensors for spectroscopic applications,” J. Mod. Phys. 02(04), 200–209 (2011). [CrossRef]
40. F. J. M. Harren and S. M. C. Cristescu, “Photoacoustic spectroscopy in trace gas monitoring, Encyclopedia of Analytical Chemistry,” John Wiley & Sons, Ltd, 2019.
41. R. Kordecki, J. Pelzl, and B. Bein, “Stray-Light Corrections in Frequency-Dependent Photoacoustics with Transparent Cells,” Proceed of Proceedings of the 5th International Topical Meeting: Photoacoustic and Photothermal Phenomena, Series in Optical Sciences Vol. 58, P. Hess and J. Pelzl, eds., Springer, 1988, 490–491.
42. L. Liu, H. Huan, X. Zhang, L. Zhang, X. Shao, A. Mandelis, and L. Dong, “Laser induced thermoelastic contributions from windows to signal background in a photoacoustic cell,” Photoacoustics 22, 100257 (2021). [CrossRef]
