甲醛(HCHO)是大气中含量最为丰富的羰基化合物, 是非甲烷可挥发性有机化合物(NMVOCs)的最重要的中间产物之一, 广泛参与大气中的光化学反应, 同时也是气溶胶的重要前体物, 在大气化学中承担了非常重要的作用。 石油化工行业的VOCs类排放是城市大气中HCHO的重要来源, 而目前化工园区中的HCHO等NMVOCs类污染物主要通过点式设备获取近地面浓度, 缺乏立体监测数据。 差分光学吸收光谱(DOAS)技术已成功应用于SO2和NO2等污染气体监测, 甲醛由于其光学吸收强度相对较弱, 反演波段内其他气体交叉干扰强, 实际的监测应用相对较少。 选取某石化企业, 运用被动DOAS方法实现了甲醛柱浓度的精确反演。 研究通过建立甲醛吸收截面与其他参与拟合气体吸收截面的二维相关性矩阵, 选取甲醛吸收截面同其他气体吸收截面相关性最小的波段, 即实现其他气体对甲醛的DOAS反演交叉干扰最小的波段的获取。 同时选取外场实际采集的光谱, 选择不同起始波段和截止波段做迭代DOAS反演, 通过拟合残差来评估甲醛在不同波段的实际反演效果。 在截面间交叉干扰小, 拟合残差低的波段范围内, 选择尽量宽的波段作为最佳的拟合波段, 实现甲醛的精确DOAS反演。 由甲醛同其他气体吸收截面的二维相关性矩阵结果, 甲醛与NO2, SO2和O3和O4间在大部分波段内相关性均在0.5以下, 交叉干扰小; 甲醛同BrO在起始波长318～320 nm, 截止波长340～346 nm以及起始波长330～334 nm, 截止波长354～360 nm两个波段范围内截面间相关性小于0.5, 适合作为HCHO的反演波段。 通过选择不同起始波段和截止波段做甲醛的迭代DOAS反演, 结合拟合截面相关性分析结果综合考虑, 最终采用332.4～358.1 nm作为HCHO的反演波段, 拟合残差在10-4量级。 利用车载被动DOAS系统, 通过建立吸收截面间二维相关性矩阵并通过实测光谱的迭代反演, 获取了适用于该套系统的HCHO最佳拟合波段, 拟合残差降低至10-4量级, 在实现甲醛精确反演的基础上, 结合系统GPS信息, 获取了某化工企业甲醛柱浓度的空间分布, 整个外场观测期间, HCHO的反演误差低于6%。 结果表明, 车载被动DOAS系统在快速获取化工园区甲醛空间分布信息上可以发挥重要作用, 为城市大气中甲醛的立体监测提供了一种有效测量手段。

Formaldehyde (HCHO) is the most abundant carbonyl compound in the atmosphere. It is one of the most important intermediate products of non-methane volatile organic compounds. It is widely involved in photochemical reactions in the atmosphere and is also an important precursor of aerosols. Formaldehyde plays an important role in atmospheric chemistry. The emission of VOCs from the petrochemical industry is an important source of HCHO in the urban atmosphere. At present, the pollutants of NMVOCs such as HCHO in the chemical park are mainly near-ground concentration through point-type equipment and lack of stereoscopic monitoring data. The Differential Optical Absorption Spectroscopy technology has been successfully applied to the monitoring of pollutant gases such as SO2 and NO2. Due to the relatively weak optical absorption intensity of formaldehyde, the cross-interference of other gases in the inversion band is strong, and practical monitoring applications are relatively rare. The paper selects a petrochemical company and uses a passive DOAS method to accurately invert formaldehyde column concentrations. In this paper, the relationship between the absorption profile of formaldehyde and other two-dimensional correlation matrix that is used to fit the gas absorption cross-section is established, and the band with the smallest correlation between the absorption profile of formaldehyde and other gas absorption profiles is selected. The acquisition of the band that minimizes cross-interference from DOAS inversion of formaldehyde by other gases is achieved. At the same time, the spectrum actually collected in the field is selected, and different initial bands and cut-off bands are selected for iterative DOAS inversion. The residuals are used to evaluate the actual inversion effect of formaldehyde in different bands. In the region where the cross-interference between the cross-sections is small and the fitting residual is low, the widest band is selected as the best fitting band to achieve accurate DOAS inversion of formaldehyde. From the results of the two-dimensional correlation matrix of absorption profiles of formaldehyde and other gases, the correlation between formaldehyde and NO2, SO2, O3, and O4 is below 0.5 in most of the bands, and the cross-interference is small. The correlation between formaldehyde and BrO at the initial wavelength of 318～320 nm, cutoff wavelength of 340～346 nm and initial wavelength of 330～334 nm, and cutoff wavelength of 354～360 nm is less than 0.5 in these two wavebands, which is suitable as the inversion waveband of HCHO. Through the selection of different initial bands and cut-off bands for the iterative DOAS inversion of formaldehyde, combined with the correlation analysis results of fitting cross sections, 332.4～358.1 nm was finally used as the inversion band of HCHO, and the fitting residual was 10-4. In this paper, a passive vehicle-borne DOAS system is used to establish the HCHO best-fit band for the system by establishing a two-dimensional correlation matrix between absorption cross-sections and through iterative inversion of the measured spectrum. The fitting residual is reduced to 10-4. Based on the accurate inversion of formaldehyde and combined with GPS information of the system, the spatial distribution of formaldehyde concentration in a chemical company was acquired. During the entire field observation period, the error of HCHO inversion was less than 6%. The results show that the vehicle passive DOAS system can play an important role in quickly obtaining the spatial distribution information of formaldehyde in the chemical industry park, and provides an effective measurement method for the stereoscopic monitoring of formaldehyde in the urban atmosphere.