基于可调谐激光吸收光谱技术, 主要采用数据处理的方法如互相关、 小波分析等降低光谱吸收技术的测量下限。 采用可调谐激光吸收光谱技术中的波长调制技术, 对较低浓度下的NH3进行了降低测量下限的试验研究。 采用近红外波段2.25 μm附近的一组ν2+ν3 NH3吸收谱线, 其强度远高于1.5 μm处谱线, 结合波长调制技术在最优信噪比的条件下对常温常压下不同浓度的NH3进行了测量。 为了保证实验条件下测量得到的二次谐波信号的峰值高度达到最大且信噪比最优, 实验在调制系数为2.2左右的情况下, 通过加载最优的高频调制信号来保证信噪比(SNR)。 在10 m长的Herriott池中探测到了浓度为0.6×10-6的二次谐波信号, 其中信号处理部分主要采用相关分析、 多次平均以及小波变换分析来控制中心波长的移动和降低噪音的干扰。 结果表明, 经过数据处理以后的谐波信号, 其检测下限降低到处理前的约百分之一, 且不需要增加任何实验设备就可以很好的抑制噪音的影响, 将相关分析和小波分析与波长调制技术相结合, 这种数据处理方法对于在线检测技术具有很好的实用价值。

To obtain the weaker second harmonic signal of low concentration, reduce the minimum measurable limit and improve the sensitivity and accuracy of absorption measurement, a serious of data processing methods are proposed based on tunable diode laser wavelength modulation spectroscopy. The experiment on lower NH3 concentration at 2.25 μm was carried out in a 10.13 m absorption cell with different concentration. The peak height of the second harmonic signal is maximum at m=2.2, which optimizes the signal-to-noise ratio. In order to guarantee the optimal signal-to-noise ratio, the experiment was carried out by loading the optimal high frequency modulation signal. WMS-2f was performed at a repetitive scan rate of 200 Hz and a current-modulation rate of 15 kHz, wavelength modulation spectroscopy with the optimal signal-to-noise ratio was adopted for its better noise immunity to measure different lower NH3 concentration in the Herriott cell. This survey is focused on the ν２＋ν３ bands of absorption spectra near 2.25 μm in near-infrared region at ambient temperature and pressure, the line strengths of 2.25 μm are much larger than the absorption lines in the telecommunication bands, using stronger NH3 absorption lines can offer the potential of lower detection limits. During the data processing, the background signal of the original harmonic should be deducted at first, the second harmonic signal of 0.6×10-6 was obtained in a 10 m long-path Herriott cell after data processing, these signal processing mainly consist of cross-correlation analysis, multiple averages and wavelet transform analysis, the cross-correlation analysis was used to control the shift of center wavelength, the multiple averages and wavelet transform analysis were used to reduce influences of the environment noise, after that we get the revised second harmonic signal and improve the accuracy of the measurement results. The experimental results show that these data processing methods can obviously improve the signal quality and reduce the minimum measurable limit about 100 times lower than before. The experiment doesn’t need to add any laboratory equipment and can well restrain the influence of the environmental noise and other disturbance, so these signal process combined with wavelength modulation technique will be more useful for on-line gas detection technology.