重金属作为PM2.5中重要的有毒有害污染物质，可对区域人群产生多种健康损害。同时，由于重金属环境来源广泛，既有机动车排放、工业排放、化石燃料燃烧等人为排放源，又受到扬尘及区域传输影响，如何针对多个排放源进行有效有序治理是当前亟需解决问题。因此，必须对区域重金属排放进行准确解析，从保护人群健康角度，评估重金属排放源的健康损害贡献，据此给出各排放源优先治理次序。但遗憾的是，当前大部分重金属健康风险评估研究仍然以常规致癌非致癌分别评价方法为主，风险结果不可统一，不利于管理决策。因此，开发可比较风险评估指标来对重金属排放源进行统一评估显得十分必要，为风险管理提供新的解决方案。 本文通过对北京城区两年逐日PM2.5采样（n=452），分析了13种金属元素，其中包括7种重金属元素（As、Cd、Co、Cr(VI)、Ni、Pb与V）。分析了重金属浓度水平的年际和季节变化；调查了区域重金属相关排放源的空间分布并通过ArcGIS展示，综合运用正定矩阵因子分解模型（Positive Matrix Factorization, PMF）、潜在源贡献因子模型（PSCF）和基于风速风向的条件概率模型（CPF），准确识别出重金属源的类别及潜在源区源向；基于收集的PM2.5滤膜样品，进行了肺溶出率实验，给出了重金属人体肺部生物可给性，并对Pb毒性参数进行本地化校正；运用美国环保署（US EPA）健康风险评估方法，采用基于关键效应（critical effect）的毒性参数，评估各排放源的致癌及非致癌风险贡献及其排序；通过检索分析大量原始文献，确定重金属的多种毒效应，建立暴露量-患病率（或癌发率）之间的数量关系；通过健康效用值及疾病持续时间参数的文献检索，疾病期望寿命损失年的计算，构建了重金属质量调整生命年评估方法（QALYs），包括7种重金属共11种疾病（肺癌、肺炎、COPD、上呼吸道疾病、哮喘、皮肤病、智力下降、鼻部炎症、慢性肾病、贫血及高血压）。基于QALYs，对不同排放源的疾病负担进行评价并依据其大小进行贡献排序，确定优先控制排放源；与常规健康风险评估结果对比显示，QALYs方法可以反映人群对疾病的健康偏好，区分不同疾病贡献，而且可将致癌非致癌健康效应统一评估。研究结果可为区域PM2.5载带重金属排放源的风险管理提供科学依据和指导方案。本论文的主要结论如下： （1） 北京市城区PM2.5浓度超过我国环境二级标准2-3倍。重金属中Cr(VI)超标，PM2.5及大部分金属元素从2016至2017年均具有下降变化趋势。季节变化特征为采暖季各元素浓度显著高于非采暖季。 （2） 北京及其周边地区重金属相关排放源分布较密集。北京市城区重金属主要来自4个排放源，贡献排序依次为扬尘与机动车排放源（81.0%）> 燃煤源（6.7%）> 燃料油燃烧源（2.7%）> Cr工业源（9.6%）。四个源的潜在源向与本地风频趋势不一致。扬尘与机动车排放源的贡献源区主要为蒙古及我国西北地区；燃煤源的贡献源区位于我国华北地区煤炭生产基地；燃料油燃烧源为北京市东南部及沿海地区；Cr工业源为河北及我国西北地区。 （3） 确定了肺溶出率实验中提取时间为48小时，固液比为1/1000 g/mL。不同重金属的生物可给性差异较大，大部分重金属巨噬细胞溶酶体酸性环境中生物可给性大于细胞间质液中性环境。其中，Pb酸性环境生物可给性为70%，而中性环境仅为0.5%。 （4） 各重金属非致癌风险之和（HI）为0.66，处于安全风险水平；致癌风险之和（CR）为3.31×10-5，超过了10-6风险水平。区域新生儿生活至1~80岁各年龄健康风险中低龄时具有更高的非致癌风险，而随着年龄增长，致癌风险逐渐增加。健康风险的概率估计中As可致区域人群致癌及非致癌风险的人群比例最高。不同排放源的致癌风险为：燃煤源（65.0%）＞扬尘与机动车排放源（18.5%）＞Cr工业源（14.7%）＞燃料油燃烧源（1.8%）；而非致癌风险为：燃煤源（63.0%）＞燃料油燃烧源（18.6%）＞扬尘与机动车排放源（15.8%）＞Cr工业源（2.6%）。 （5） 2017年新生儿人群因重金属暴露的QALYs损失为3665 QALYs，约12.5天/人。各排放源的QALYs损失贡献排序为：燃煤源（57.8%）> 燃料油燃烧源（20.0%）> Cr工业源（16.9%）> 扬尘与机动车排放源（5.3%）。燃煤源是最大的疾病负担贡献者，由于其排放更多具有较高毒性的As，Cd和Pb，建议进行优先管控。

Heavy metals were well-known toxic components on PM2.5, can cause a series of toxic effects, and originate from numerous sources, such as vehicle emissions, fossil fuel combustion, industries, resuspended dust, and long-range transportation, etc. How to effectively control the sources is urgently needed to be solved. Therefore, a thoroughly source apportionment should be conducted, based on which, their health risk and contribution were evaluated and ultimately, priority control sources were determined. But unfortunately, common health risk assessment remains at the stage of separately evaluating cancer and non-cancer health risk, namely, the health risk cannot be unified, which will hinder effective environmental management. Therefore, a comparable risk assessment index needs to be developed. In our study, 13 PM2.5-bound metals, including 7 heavy metals (As, Cd, Co, Cr(VI), Ni, Pb and V), were analyzed based on a two-years day-by-day sampling (n=452) in the urban area of Beijing. Levels and seasonal variation of heavy metals were clarified. Spatial distribution of regional heavy metal related sources was investigated and shown by ArcGIS, then, positive matrix factorization (PMF), potential source contribution function (PSCF) and conditional probability function (CPF) were comprehensively employed to identify the sources, their contribution regions and potential directions; An dissolution experiment in lung was carried out to investigate the bioaccessibility of heavy metals in human lungs by using filter collected PM2.5 samples. Health risk assessment was employed to give the risk contribution ranks of each source by the method provided by US Environmental Protection Agency (US EPA), which was based on toxicity parameters for critical effects; Multiple toxicity effects and dose-effect relationships of the heavy metals were retrieved from the toxicology database (ATSDR, IRIS, CalEPA, etc.) and original researches, and the toxicity parameters for each effect was further calculated. Quality adjustment life years (QALYs) of heavy metals was ultimately developed based on health utility value (HUV), duration of disease (DD) and expected years of life lost (EYLL), including 7 heavy metals, and 11 diseases (lung cancer, pneumonia, Chronic Obstructive Pulmonary Disease (COPD), upper respiratory diseases, asthma, skin diseases, mental retardation, nasal inflammation, chronic kidney disease, anemia and hypertension). Finally, QALYs lost attributed to different sources was estimated. Comparing with the common health risk assessment method, QALYs approach can reflect the preferences of population on different health states, can unify the assessment of cancer and non-cancer health effects. The research results can provide a scientific basis for environmental risk management of regional PM2.5-bound heavy metals. The main conclusions are as follows: (1) The annual mean PM2.5 concentration in Beijing urban area was 2-3 times above the national Ambient Air Quality Standard (AAQS) of China. The concentration of PM2.5-bound Cr(VI) exceeded the grade II value of this standard. Interannual variation showed that PM2.5 and most metals have decreased from 2016 to 2017. Seasonal characteristic was that the concentration in the heating season was higher than that in non-heating season for most metals. (2) Heavy metal related emission sources were intensively distributed in Beijing surrounding areas. Four sources of heavy metals in urban area of Beijing were identified, and their ranks and contribution were resuspended dust and vehicle emissions (81.0%), coal combustion (6.7%), fuel oil combustion (2.7%) and Cr-related industry (9.6%). The potential directions of the sources were inconsistent with the local wind frequency. The potential contribution region for re-suspended dust and vehicle emission was Mongolia and northwestern China; coal combustion was mainly from North China, the well-known coal production bases; fuel oil combustion was from southeastern Beijing and some coastal areas; and Cr-related industry was from Hebei and Northwest China. (3) The extract time of 48h and SLR of 1/1000 g/mL were determined for dissolution experiment in lung. Bioaccessibility of various heavy metals was quite different. And the bioaccessibility in acidic environment of phagocytosis of macrophages was greater than that of neutral environment of alveolar interstitial fluid for most metals. Therein, Pb was 70% and just 0.5% in acidic and neutral environment, respectively. (4) Non-cancer risk of total heavy metals (HI) was 0.66, which was within the safety level (HI<1); while, the cancer risk (CR) was 3.31×10-5, which has exceeded the risk level of 10-6. Health risk assessment for the newborns till their 1 to 80 years old showed that, larger non-cancer risk was observed at a lower-ages; while, higher-ages population had a higher cancer risk. From probability estimation by Monte-Carlo simulation, As had the highest probability of inducing cancer and non-cancer risk to local population. The cancer risk contribution for each source ranked as: coal combustion (65.0%) > resuspended dust and vehicle emission (18.5%) > Cr-related industry (14.7%) > fuel oil combustion (1.8%); while, non-cancer risks ranked as: coal combustion (63.0%) > fuel oil combustion (18.6%) > resuspended dust and vehicle emissions (15.8%) > Cr-related industry (2.6%). (5) The loss of QALYs is 3665 QALYs, around 12.5 days per person, if the newborns are exposed to PM2.5-bound heavy metals of current levels for 30 years. In terms of QALYs lost (disease burden), the sources contribution was ranked as: coal combustion (57.8%) > fuel oil combustion (20.0%) > Cr-related industry (16.9%) > resuspended dust and vehicle emission (5.3%). Coal combustion was the priority control source because it emitted more As, Cd and Pb with higher toxicity.