Sulfur dioxide (SO₂) is a toxic gas which poses a great threat on environmental protection and human health even at low concentration. Hence, considerable attention has been devoted to achieve accurate SO₂ monitoring at ppb level. In this work, we report on a ppb-level SO₂ detection system via deep ultraviolet differential optical absorption spectroscopy (DUV-DOAS). A novel concentration-retrieval algorithm and spectral data-processing method were proposed to overcome the ubiquitous cross-sensitivity. The differential optical density (DOD) signal of SO₂ component was identified and decomposed by FFT (fast Fourier transform) from overall DOD signal of gas mixture, followed by a reconstruction by IFFT (Inverse fast Fourier transform) after frequency domain filtering. Via this method, SO₂ DOD signal was accurately extracted regardless of strong interference from NO, NO₂ and NH₃, indicating a remarkably enhanced SO₂ selectivity. Besides, a detection limit (DL) of 4 ppb and a quantitation limit (QL) of 20 ppb were achieved, which are among the best cases of related reports thus far. Moreover, a compensation model was offered to compensate for deviations caused by gas temperature variation and thus enhance the system accuracy and adaptability.

二氧化硫（SO ₂）是一种有毒气体，即使浓度低也对环境保护和人类健康构成巨大威胁。因此，已经投入了大量精力来实现ppb级的精确SO ₂监测。在这项工作中，我们通过深紫外差分吸收光谱法（DUV-DOAS）报告了ppb级SO ₂检测系统。提出了一种新颖的浓度检索算法和光谱数据处理方法，以克服普遍存在的交叉敏感性。SO ₂的差分光密度（DOD）信号从混合气体的总DOD信号中通过FFT（快速傅里叶变换）识别出分量并分解，然后在频域滤波之后通过IFFT（逆快速傅里叶变换）进行重构。通过这种方法，无论NO，NO ₂和NH ₃的强烈干扰如何，都能准确提取SO ₂ DOD信号，表明SO ₂选择性显着提高。此外，达到了4 ppb的检测限（DL）和20 ppb的定量限（QL），这是迄今为止相关报告的最佳案例。此外，提供了补偿模型来补偿由气体温度变化引起的偏差，从而提高了系统的准确性和适应性。