【摘要】 [背景]职业卫生场所复杂粉尘样品通常涉及面广、成分复杂、缺少附加信息，因而难以明确晶相构成，这对于晶体学筛查和进一步的准确定量构成了较大挑战。因此，针对无背景信息的职业卫生场所复杂粉尘样本，系统性构建非定向晶相筛查策略是其中的关键。[目的]立足于多种无损光谱学技术和计算机全谱拟合技术，构建一种适用于职业卫生场所复杂粉尘样品的非定向晶相筛查策略及对应的结果评判机制，为无背景信息的复杂样品的晶相筛选和无标样物相定量奠定技术基础。[方法]立足于职业卫生样本特性，建立了一套基于能量色散型X射线荧光光谱、红外光谱和拉曼光谱辅助的可拓展的非定向职业卫生场所相关晶相筛查策略以及相应的双重验证机制，实现了对样品晶相的准确筛查。以职业卫生环境粉尘常规监测样本中某一具有代表性的样品为例，进行了光谱学辅助的多晶相筛查策略的适用性验证，并从精修结果的准确性和各晶相含量分布在相关职业卫生场所环境中的合理性两方面进行双重验证。进一步地，为验证非定向晶相筛查策略的普适性，将其应用于耐火材料制造、建筑用石加工、钢压延加工等多个经典的易产生涉矽粉尘的职业卫生场所，同样以准确性和合理性的双重验证作为评价标准。[结果]基于某具有代表性的职业卫生场所复杂粉尘实际样品，从高精度X射线衍射法（XRD）扫描数据出发，以主成分定位、由主及次、先筛查关联性晶相再查漏补缺为整体思路，筛查得到了关联性较强的13种晶相，其中11种晶相具备强行业相关性。基于这一晶相筛查结果，进行了Rietveld全谱拟合计算，计算得出的谱图与实测谱几乎完全重合，误差线较为平直，表明晶相筛查的结果与样品中实际晶相分布较为契合。精修结果参数加权图形方差因子(R_wp)=5.121%、期望方差因子(R_exp)=2.955%、χ²=1.733，均表明拟合结果能较准确反映出样品中各晶相的实际情况，且精修拟合得出的各晶相的含量分布与相应的材料定位及相定位相符。在方法的普适性验证中，各样品的R_wp均<20%，且χ²均<2。此外，各样品的材料定位、相定位逻辑自洽且与其真实行业信息具有较高的匹配度。[结论]建立了一套结合多类型光谱分析技术和计算机全谱拟合技术的晶相筛查策略，能有效识别无信息样本中的多种结晶型物质，实现了对职业卫生场所复杂粉尘样本的非定向筛查，经由结果准确性和晶相分布合理性两方面实现了双重验证，并验证了其在多类职业卫生场所中的普适性，为无背景信息的职业卫生场所复杂粉尘样本分析工作提供了切实有效的方法示范。更多还原

【Abstract】 [Background] Occupational health places often involve complex dust samples characterized by diverse compositions and a lack of prior supplementary information. These factors pose challengs in ascertaining crystal phase composition, creating a hurdle for both crystallographic screening and subsequent precise quantification. Hence, establishing a systematic, non-targeted crystal phase screening strategy is essential for analyzing complex occupational dust samples lacking background data.[Objective] To develop a non-targeted crystal phase screening strategy and a corresponding evaluation mechanism for complex occupational dust, by leveraging various non-destructive spectroscopy techniques and automated full-spectrum fitting, aiming to establish a technical foundation for phases identification and standardless quantification in samples with unknown backgrounds.[Methods] An extensible and non-targeted screening strategy and a dual-verification mechanism were formulated specifically for occupational health samples. The approach integrated energy dispersive X-ray fluorescence spectroscopy（EDXRF）, Fourier-transform infrared（FTIR）, and Raman spectroscopy to ensure accurate phase identification. To demonstrate applicability, a representative routine monitoring sample was analyzed. The strategy was validated based on the precision of Rietveld refinement results and the logical consistency of phase distribution within specific occupational settings. Furthermore, universality was tested across multiple industrial sectors prone to silica-related dust, including refractory material manufacturing, construction stone processing, and steel rolling.[Results] Utilizing high-precision X-ray diffraction（XRD） data from a representative complex dust sample, a comprehensive workflow comprising primary component identification, sequential phase screening, and gap-filling analysis successfully identified 13 highly correlated crystal phases, 11 of which exhibited strong industry relevance. Subsequent to these screening results, Rietveld full-spectrum fitting yielded a calculated pattern closely matching the experimental data, with a relatively flat residual line and high statistical reliability(R_wp=5.121%, R_exp=2.955%, and χ²=1.733). In universality tests across various industrial sites, all samples achieved R_wp< 20% and χ² < 2. In addition, the quantitative phase distributions remained self-consistent and highly aligned with actual industrial processes and material positioning.[Conclusion] A crystal phase screening strategy combining multi-modal spectroscopic analysis with computer-aided full-spectrum fitting is successfully established. This method effectively identifies multiple crystalline substances in uninformed samples and achieves non-targeted screening of complex occupational dust samples. Validated through both numerical accuracy and phase distribution, the strategy provides a practical and robust methodological framework for analyzing complex occupational dust samples without background information.更多还原