质子荧光分析——PIXE（Proton-induced X-ray emission），是一种重要的离子束分析方法。这种分析方法的分析灵敏度和技术方法不断地被改善，应用广度、深度及与其它技术的结合也不断被发展。本研究基于实验室原有的外束PIXE分析系统，建立了外束PIXE双探测器快速分析系统。（1）设计和制作了新型的外束引出管道，利用7.5 μm Kapton薄膜作为外束引出窗口将真空和大气隔开，外束管道内可以在5分钟内从大气压达到4×10-4 Pa真空度。通过引2.5 MeV质子束流，该外束窗口能有效地从真空中引出直径最大为φ 6.8 mm的质子束流，质子能量损失约为179 KeV。（2）在新型管道内安装了金硅面垒型半导体探测器。通过金硅面垒探测器探测Kapton膜质子背散射能谱，并利用SIMNRA软件模拟，研究了不同样品种类、He气中残余空气杂质及Kapton膜厚度变化对质子背散射能谱的影响。通过选取适当的金硅面垒探测器几何位置和合适的背散射能谱范围进行计数积分，消除了各种因素对能谱及其“计数积分-束流积分比例系数”的影响。在不同样品条件和束流大小下，得出了背散射能谱“计数积分-束流积分比例系数”为(28.95±0.59) Kc /μC，相对标准误差RSD为2.0 %。利用Kapton膜质子背散射能谱实现了束流积分监测，为厚样品定量分析打下了基础。（3）利用新增SDD探测器探测较轻元素的特征射线（Mg镁~V钒），加240 μm Mylar膜吸收片的Si(Li)探测器探测中重元素的特征射线（Cr镉~Pb铅），这两个探测器匹配成为双探测器外束PIXE系统。解决了原单Si(Li)探测器计数率低，无法进行大束流快速分析问题。（4）利用标准样品，对系统在4种探测条件下进行H因子刻度，并应用实际PM2.5采样样品进行验证。得到了有效和准确的H因子值。（5）在不同探测条件下，对同一批PM2.5采样样品进行了实际分析，得到了各个情形下的各元素LOD平均水平。新系统分析时间为原系统3分之1时，单SDD探测器条件下元素LOD水平比原有外束系统平均降低43 %，双探测器条件下，LOD水平平均降低57 %。并探讨了不同探测器、不同束流积分值和不同吸收片条件对元素LOD的影响。

PIXE（Proton-induced X-ray emission）is one of important ion beam analysis methods. It has been applied for elements analysis in many fields. The application range of PIXE has been widening, sensitivity and accuracy has been improving. Based on the pre-existing external beam proton induced X–ray emission analysis facility, fast external beam PIXE analysis system was established. (1) A 7.5 μm Kapton foil was used to isolate vacuum from atmosphere. Vacuum can reach 4×10-4 Pa in the external beam line. 2.5 MeV, Φ 6.8 mm proton beam was extracted from the window and the energy lose was about 179 KeV. (2) An Au-Si surface barrier detector was installed in vacuum to detect the protons backscattered by the Kapton extraction window. By detection and SIMNRA simulation, proton backscattering spectra were studied under several sample situations. It’s found that the Kapton film thickness, He atoms density and air impurity in the He flow had influence on spectra. Through appropriate detector’s geometric arrangement, influences to the proton backscattering spectra was eliminated. CQ (ratio between integral backscattering protons counts and integral beam charge) under different sample situations show high consistency. In beam current dependence study, the CQ statistic result was (28.95±0.59) Kc/ μC with RSD of 2.0 %. CQ is applicative for beam charge integration in external beam PIXE-PIGE analysis. Nondestructive quantitative PIXE-PIGE analysis in thick sample situations is achieved. (3) A Si(Li) detector was attached 240 μm Mylar foil to measure medium-high Z (Cr~Pb) elements’ X-ray. A Silicon drift detector was employed to detect low-Z (Mg~V) elements’ X-ray. Above two detectors were combined as the new X-ray measurement system in fast PIXE analysis. (4) H factors of the new fast PIXE system were scaled by Micromatter standard samples in four different detect situations. They were also verified effective by PM2.5 samples. (5) Limit of detection (LOD) of the new fast PIXE system was studied under 4 kinds of detect situations. Compared with the pre-existing system, elements’ LOD were reduced by 43% in average in single SDD situation and 57 % in double-detectors situation. At the same time, only one third time was needed in analysis. LOD level was also compared in different beam charge integration, detector situation and X-ray filters.