人工湿地具有生态、高效、节能的特点，受到世界各国推崇，广泛应用于分散或非点源污水处理。随着对污水处理要求的提高，人工湿地也从最初主要用于处理污水中悬浮物和有机物逐渐向除磷脱氮方向发展。人工湿地有着公认的脱氮潜能，但实际运行的人工湿地脱氮效率较低，一般不超过50%。人工湿地脱氮因涉及微生物、基质(包括土壤、填料)、植物等等因素，其复杂性决定了人工湿地中氮的转换和去除机理还远不清晰。这就制约了湿地实际脱氮性能的改进。因此研究湿地生物脱氮作用的机制，特别是阐明根际微生态环境中，根系分泌物、植物、脱氮微生物三者耦合关系及调控机制，对正确预测和评估湿地脱氮效率、改进湿地脱氮性能具有重要意义。 在各种人工湿地处理技术中，芦苇潜流人工湿地是应用最广的湿地类型。试验以芦苇潜流人工湿地为研究对象，人工配制以氨氮为唯一氮源的污水，探讨3种潜流人工湿地（空白湿地、芦苇湿地、根袋湿地）的运行效果，研究了湿地运行期各种污染物的沿程净化规律；采用根袋法区分湿地种根际与非根际环境，测定了芦苇根际土壤与非根际土体土壤ORP及土壤溶液中TOC的变化，检测了空白湿地与植物湿地中硝化细菌/反硝化细菌的数量与分布、湿地不同空间土壤硝化/反硝化强度；初步鉴定了芦苇根系分泌物中有机酸的类型及其对脱氮微生物性能的影响。实验结果表明： 空白湿地、芦苇湿地和根袋湿地3种小型潜流人工湿地对水中污染物均有着较好的去除效果，对NH4<下标!>+<上标!>-N去除率分别为54.4%、77.6%和51.9%；对TN的去除率分别为56.2%、76.5%和51.8%；对TP的去除率分别为60.0%、69.5%、59.1%；对COD的去除率分别为67.5%、79.3%、76.8%。 在HRT为5d，水位保持在35cm，进水C/N=5的运行条件下，空白湿地、芦苇湿地和根袋湿地对NH4<下标!>+<上标!>-N和TN的去除效果，分别达到了66.2％、94.2%、82.2%和67.2％、90.7%、76.1%。NH4<下标!>+<上标!>-N和TN去除具有沿程同步性，且在NH4+-N减少的同时，TN也有所降低，说明试验中存在同步硝化反硝化。从湿地沿程净化规律看，污染物的去除主要存在于湿地的前端，去除率达到总去除率的80～90%。 根袋湿地前端的根际土壤ORP明显高于空白湿地和湿地土体土壤，差值分别为11～311mV、62～261mV，根际附近较高的氧化还原电位有利于硝化反应的进行；根际土壤与非根际土体土壤中TOC含量分别为21.3～54.6mg•L-1<上标!>和6.65～12.0mg•L-1<上标!>，根际附近较高的TOC含量能促进反硝化反应的顺利进行，因此试验中有植物的湿地对氮的去除要好于空白湿地。 湿地中硝化强度存在明显的分层现象。芦苇湿地中上层土壤的硝化强度高于下层，沿程逐渐上升；空白湿地上层土壤硝化强度高于下层，沿程略有下降；植物根际土壤的硝化强度低于非根际土体土壤，芦苇湿地和根袋湿地硝化能力强于空白湿地，植物湿地中硝化强度与硝化细菌的空间分布之间没有明显的对应关系；反硝化强度在芦苇湿地中分层现象不明显，沿程基本不变；空白湿地中底层反硝化强度低于表层，沿程略有升高；根际土壤反硝化强度低于非根际土体土壤，3个湿地中，反硝化强度和反硝化细菌空间分布均存在较好的对应关系，其中以空白湿地的反硝化能力较高。 利用高效液相色谱初步鉴定出芦苇根系分泌物含有草酸，根据试验结果和文献报导，选用草酸和柠檬酸研究根系分泌物中有机酸对湿地土壤中分离出的硝化细菌和反硝化细菌活性的影响，结果表明，培养基中添加不同弄得的草酸和柠檬酸，对硝化细菌均有抑制作用。相同培养基条件下，添加含碳量均为0.1mg•L-1<上标!>草酸和柠檬酸，抑制率分别为37.5%和35.4%，且随着有机酸浓度增大，抑制率逐渐增大，当含碳量为3.0mg•L-1<上标!>，抑制率增加至80.7%、85.5%。；反硝化细菌可以利用草酸和柠檬酸作为反硝化所需的碳源，培养基中添加相同含碳量的草酸和柠檬酸，C/N达到12.3，脱氮率分别为16.5%和93.0%，倾向于利用柠檬酸。 根据试验结果，在芦苇湿地床人工湿地中，根际土壤ORP较高，TOC含量较多，但根际附近硝化强度较低，说明芦苇根际分泌物对硝化细菌存在一定的抑制作用，使硝化作用过程受阻；同时根际附近反硝化强度也较非根际低，不能缓解根系分泌物对硝化细菌的抑制作用，导致整个微生物脱氮途径不畅通。

Constructed wetlands, a kind of strengthen natural processing technique for water purification and pollution-control, has been extensive applied in dispersive sewage and micro-contaminated water around the world. With the high demand of sewage treatment in nitrogen and phosphorus removal, the research field has been changed into nitrogen and phosphorus removal. As everyone knows, constructed wetland has the protential in nitrogen removal, but the removal efficiency is always below 50%. Because nitrifitation is a complex progress involving wetland vegetation, substance (soil, padding) and associated microbial, the transformation of nitrogen and removal mechanism are still not very clear. So it is important to forecast and evaluate nitrogen removal efficiency and enhance nitrogen removal ability by illustrating the relationship between root exudate, wetland vegetation and microbial in rhizosphere. During different typies of constructed wetlands, subsurface flow wetland is the main type in engineering application. In this thesis, three subsurface flow constructed wetlands (CWs) with and without reed were investigated with artificial ammonium-rich wastewater. Root bag made of nylon sieve with 300 mesh was used to enwrap the reed root in one of reed CWs to distinguish reed rhizosphere from non-rhizosphere. The CWs with root bag enwrapped reed root hereinafter was called as Mesh CWs. Contaminants removal efficiencies were assessed during running period, also the removal of contaminants along the stream way, the change of the oxidation-reduction position (ORP) and concentration of total organic carbon (TOC) in rhizosphere and non-rhizosphere were detected. Furthermore, the relationship between the amount of microorganism and intensity of nitrification and denitrification in wetlands was researched. Finally, root exudate of reed was detected and the mechanism of nitrogen removal based on the different root exudate was discussed. The results were as follows: In running period, the removal efficiencies of NH4<下标!>+<上标!>-N in Control CWs, Reed CWs and Mesh CWs were 54.4%, 77.6% and 51.9%, respectively. TN removal efficiencies were 56.2%, 76.5% and 51.8% respectively.TP removal efficiencies were 60.0%, 69.5%, and 59.1%, COD removal efficiencies were 67.5%, 79.3%, and 76.8% respectively. When hydraulic rention time was 5 days, water level was 35 centimeters, C/N ratio was 5, removal efficiencies of NH4<下标!>+<上标!>-N in Control CWs, Reed CWs and Mesh CWs were 66.2%, 94.2% and 82.2%, respectively. TN removal efficiencies were 67.2%, 90.7% and 76.1% respectively. Simultaneous nitrification and denitrification phenomenon in this study was also observed. Most contaminants were degradated in the front of wetlands. The ORP in the front of Mesh CWs’s rhizosphere was much higher than that in control CWs and non-rhizosphere in Mesh CWs, which were 11～311mV and 62～261mV, respectively. Root exudate also showed the difference between rhizosphere and non-rhizosphere in Mesh CWs, the TOC of them were 21.3～54.63mg•L-1<上标!> and 6.65～12.0mg•L-1<上标!>. Due to the higher OPR and concentration of TOC, the nitrogen removal efficiency in Reed and Mesh CWs was much higher than that in Control CWs. The intensity of nitrification was different in two layers. Upper layer was higher than lower layer in Reed CWs, and the intensity increased along the bed. In Control CWs, also upper layer was higher than lower layer, but the intensity declined along the water stream. Rhizosphere had lower intensity than non-rhizosphere in Mesh CWs. The Reed and Mesh CWs’s nitrification ability was better than Control CWs, but the intensity of nitrification had no corresponding with the amount of nitrify bacteria in Reed and Mesh CWs. Denitrification intensity had no obvious difference in the two layers or along the length of Reed CWs, Contrary to Reed CWs, the upper layer was lower than lower layer in Control CWs, and the intensity was better along the steam in the CWs. Non-rhizosphere’s intensity was higher than that in rhizosphere in Mesh CWs. The amount of microorganism of denitrification had pretty good corresponding with intensity of denitrification, and Control CWs had the best ability of denitrification in the three CWs. Root exudate of reed was detected for oxalic acid by using HPLC. The influence of oxalic and citric on the microorganism of nitrification and denitrification was analyzed. The results showed that nitrification was inhibited by 0.1mg•L-1<上标!> of carbon in oxalic and citric. The inhibitation percentage was 37.5% for oxalic and 35.4% for citric. As the concentration of organic acid increased, the rate of nitrification would decrease. The inhibitation of oxalic and citric were 80.7% and 85.5% when carbon concentration was 3.0mg•L-1. Oxalic and citric can be used as carbon source for denitrification, TN removal efficiencieis was 16.5% and 93.0% when the C/N was 12.3 and citric was preferred to be utilized. According to the results, there was a contradiction between the lower nitification intensity and higher ORP and the concentration of TOC in rhizosphere in the reed constructed wetland, and it demonstrated that reed root exudate could inhibit nitrification. Meanwhile, lower denitrification intensity in rhizosphere could not eliminate the inhibitation of root exudate, so the nitrification-denitification was set back.