† Corresponding author. E-mail:
Project supported by the National Natural Science Foundation of China (Grant Nos. 61172047, 61502538, and 61501525).
The expressions of the second harmonic (2f) signal are derived on the basis of absorption spectral and lock-in theories. A parametric study indicates that the phase shift between the intensity and wavelength modulation makes a great contribution to the 2f signal. A self-calibration wavelength modulation spectroscopy (WMS) method based on tunable diode laser absorption spectroscopy (TDLAS) is applied, combining the advantages of ambient pressure, temperature suppression, and phase-shift influences elimination. Species concentration is retrieved simultaneously from selected 2f signal pairs of measured and reference WMS-2f spectra. The absorption line of acetylene (C2H2) at 1530.36 nm near-infrared is selected to detect C2H2 concentrations in the range of 0–400 ppmv. System sensitivity, detection precision and limit are markedly improved, demonstrating that the self-calibration method has better detecting performance than the conventional WMS.
C2H2, a species of toxic gases, has strong chemical activity and is easy to be decomposed, burn, and explode. In modern energy systems, C2H2 is one of the main components of the oil-decomposed fault gases in transformer.[1,2] During the research of plant growth, C2H2 can be used as a kind of plant hormone to ripen the fruits or wither the flowers. Therefore, there is a demand for the trace C2H2 detection in some practical applications.
Designing rapid and accurate trace gas measurements has become a hot topic. Numerous efforts have been contributed, such as photoacoustic spectrometry (PAS),[3] QEPAS (quartz-enhanced) with a new PAS technique,[4] metal oxide semiconductor (MOS),[5] and TDLAS,[6,7] among these methods, the TDLAS technique has been frequently used due to its high sensitivity and accuracy. In order to further improve the signal to noise rate (SNR), WMS with 2f signals is employed to TDLAS[8–11] technique. There is now a general consensus that the first harmonic (1f) signal can be used to normalize the 2f signal to improve accuracy.[12,13] In these studies, the phase shift (PS) between the intensity and the wavelength modulation has been demodulated to optimize the 2f/1f fitting.[12] Although this method can reduce the detection error, it is intricate to demodulate the phase shift and use the simulated line shape to satisfy the best fitting.
In this paper, we derive the expressions of the 2f signal according to the absorption spectral and lock-in theories. By analyzing the expression of 2f signal with a different way of Fourier expansion (by putting the PS to the wavelength modulation and separating it from wavelength modulation frequency without as a whole in the process of Fourier expansion), we will research the influence factors of a 2f signal. In particular, the influences of PS (varies from 0.5π to π) and pressure are analyzed and simulated. The scope of this research lies in ambient pressure, temperature suppression, and phase shift effects elimination. Therefore, we have utilized a self-calibration WMS measurement method based on TDLAS with an advanced gas flow cell and a reference cell to detect the gas concentration. The selected 2f signal pairs of measured and reference WMS-2f spectra are used for 2f peak calibration, combining the advantage of concentration least-squares algorithm optimization, leading to high accuracy and linearity concentration detection. In addition, the transition of C2H2 at 1530.36 nm is selected based on the Pacific Northwest National Laboratory (PNNL) spectra database and spectral data from experiment. The numerical results calculated by new method are compared with conventional results (extract concentration from 2f signal peak without reference 2f signal calibration). The accuracy and linearity, precision and stability, detection limit and sensitivity are evaluated to verify the performance of the self-calibration system for simultaneous measurements of C2H2.
Once the monochromatic radiation of the light source overlaps with a rotation transition of a gas species, absorption will occur, resulting in the attenuation of light intensity. The transmitted intensity It(v) at simultaneous time t associated with a rotation transition in a gas cell is given by the Beer–Lambert law, when the absorbance is lower than 5%,[13]
Compared with direct absorption measurements, WMS can improve the SNR and is less influenced by non-absorption transmission losses. In this paper, a saw tooth wave signal as well as a sine wave signal is injected into the distributed feedback laser (DFB) to tune the output wavelength. Getting the instantaneous center wavelength and laser intensity as follows:[10]
According to the statement of basic measurement principles, various influence factors of 2f signals including the pressure and PS ψ are investigated. This will be followed by descriptions of Simulink results and detailed presentations of how the 2f signals are affected.
Wavelength-scanned and wavelength-modulation spectroscopy with 2f signal detection is used to determine the absorption magnitude and gas concentration in actual measurements. To find out the influence factors of 2f signals’ peaks, the expressions of the 2f signals are further derived by Eq. (
Obviously, the absolute peak S2f is determined by the ambient pressure, temperature, laser intensity, H2a and H2b. As noted, H2a and H2b are much dependent on the PS between the intensity and the wavelength modulation. The PS ψ is determined by contributions from two effects, namely: slow thermal mechanisms and fast carrier injection effects.[16] (I) The thermal effect (the expansion of the optical path length in the DFB laser grating period) produces an increase in the wavelength with the current. At high modulation frequency, thermal delay causes that the wavelength modulation progressively lags the intensity modulation. (II) Carrier injection, through plasma and band filling effects, decreases the wavelength as the current increases. The wavelength modulation from this effect is π out of phase from intensity modulation. As illustrated in Fig.
Beyond that, the line-width of DFB laser also contributes a lot to the 2f signal peak. The peak expressions of 2f signal derived above have assumed that the line-width is zero, which cannot describe how DFB laser line-width affects the 2f signal peak. Here, we assume that the DFB line-width is v,[17] and then the output intensity of line-center v0 is
Moreover, temperature variation only affects the line-strength S(T), and line-strength has a negative linear correlation relationship with temperature. According to this negative linear correlation relationship and Eq. (
To verify the correctness of the proposed idea, a platform based on Matlab Simulink is established and the effectiveness of pressure and phase shift to peaks of 2f signals is analyzed. Here, the parameters of simulation conditions are similar to the parameters of the practical measurements. The gas temperature and absorption path length are 296 K and 40 cm, respectively. Meanwhile, the gas pressure varies from 0.8 atm to 1.2 atm,[19] and the PS ψ varies from 0.5π to π. The simulation results are shown in Figs.
The 2f signals under different pressures (assuming the PS ψ = 0, T = 296 K) are shown in Fig.
A self-calibration WMS experimental system is developed based on TDLAS, as shown in Fig.
A self-calibration scheme is introduced into the gas concentration experimental with a reference gas cell. The reference gas cell and the measurement gas cell share the same kind gas but with different concentrations. They are placed under the same ambient temperature and pressure. Referring to Eq. (
For an exact optical fiber gas concentration detection system, the absorption path length is known. The light beams in the two paths separated by the coupler are of the same intensity and PS ψ between the wavelength modulation and intensity modulation. Such an operation devotes the proportion of
The PS ψ has been measured to proof and express the correctness of the proposed theory in Section 2.2. A 0.016-cm−1 FSR etalon is utilized as the substitute of reference gas cell to record the relative wavelength variation accurately. The PS is extracted as shown in Fig.
The choice of C2H2 absorption line needs to fulfill the following requirements: (i) the absorption line is strong at the measurement temperature (near 300 K), providing good SNR and measurement precision; (ii) the transition is well isolated from neighbors and other gases; and (iii) the collisional width is small, providing narrow line-shape to enhance WMS signals. The DFB diode laser applied here possesses an operating center wavelength near 1530 nm. It means that a C2H2 absorption line near 1530 nm needs to be chosen. Figure
In the experiment, the reference gas cell (40-cm long with a fixed pressure of 0.8 atm) is contained with standard C2H2 gas of 100-ppmv concentration. A fiber-coupled gas flow cell (FC-16 FCM, 40-cm long) is employed as the measurement gas cell. The measurement gas is C2H2 with a concentration range of 0 ppmv–400 ppmv. The total pressure in the gas flow cell is also kept at ∼ 0.8 atm to minimize the line shape discrepancies between the measurement and reference 2f signals. The gas pressure of the measurement gas cell is measured and controlled by vacuum pressure gauge (BST-111) with an accuracy uncertainty of 0.01%. The laser wavelength is driven by a 30-Hz saw tooth wave summed in an adder with a 25-kHz sine wave to provide the wavelength modulation. Before each measurement, the gas sample was allowed to stabilize thermally with enough time. In this way, the temperatures of the two gas cells can be kept at room temperature and the influence of temperature can be minimized.
Examples of experiment measured 2f signals of detecting the 50-ppmv and 100-ppmv C2H2 standard gases under the pressure of 0.8 atm are shown in Figs.
To guarantee the extracted concentration accuracy of the measurements, only 2f signal pairs with highly consistent line shapes between measured and reference 2f signals are applied. The useful signal pairs are selected by comparing the HWHMs of measured and reference 2f absorbance spectra. The convenient and reliable data processing method is based on peak-trough seeking algorithm, and then extracting the relative HWHMs through the difference sampling points between peak and trough of measured and reference 2f signals. The selected signal pairs can be used to derive the C2H2 concentrations in the measurement gas cell from the normalized measured and reference 2f signal intensities’ peak heights according to Eq. (
Besides, 100-ppmv, 200-ppmv, and 400-ppmv C2H2 standard gases are detected to verify the accuracy and linearity of the self-calibration method. Figure
The stability of the instrument performance is analyzed based on the standard deviation. When the standard deviation is lower, instrument performance will be better. After over 3 h, 50-ppmv C2H2 standard gas is measured to evaluate the stability, as shown in Fig.
The detection limit of self-calibration detection system based on reference 2f signal calibration method is determined by Allan analysis.[20] Here, C2H2 measurements with 5-ppmv standard gas are performed during an estimated 1-h experiment. An Allan variance is utilized to analyze the stability and detection limit of the system. Figure
Here, the sensitivity of self-calibration detection system is determined by the ratio of variation in output light intensity and variation in C2H2 concentration. It can be described as
In this paper, self-calibration wavelength modulation spectroscopy for acetylene detection based on TDLAS has been presented. The expressions of 2f signal are derived according to the absorption spectral and lock-in theories, where the 2f signal peak can be recurrently influenced by phase shift between the wavelength modulation and intensity modulation. Moreover, a self-calibration experimental system with reference 2f signal calibration method is designed. In our experiment, the influences of pressure, temperature, and phase shift on the measurement concentrations are suppressed markedly. The numerical results indicate that the self-calibration system performs better accuracy and stability than the conventional one. The precision of the system can reach 1.6 ppmv through standard deviation analysis. Furthermore, by conducting Allan variance analysis and absorption principle discussion, the detection limit and sensitivity of 122 ppbv and 0.11 μW/ppmv have been achieved, respectively. This is significant for practical applications in many other gases.
1 | |
2 | |
3 | |
4 | |
5 | |
6 | |
7 | |
8 | |
9 | |
10 | |
11 | |
12 | |
13 | |
14 | |
15 | |
16 | |
17 | |
18 | |
19 | |
20 |