| Abstract: | 流體化床結晶軟化反應槽常用的操作方式，系統中使擔體流體化之上升流速會將因過飽和產生的懸浮固體物帶出槽體，導致出流水濁度偏高，因此需在流體化床後面增設一過濾單元來防止出流水含有大量的懸浮固體物，此一設計不僅佔地面積大且會增加建設費用成本。有鑑於此，本研究設計一新反應槽「移動結晶床」並以鉛做為汙染物來探討反應槽之可行性，此反應槽不僅佔地面積小且兼有結晶與過濾效果兩種功能，可改善流體化床操作之問題。
綜合文獻與反應槽設計歸納出可能影響去除率與回收率以及濁度的因子為鉛濃度、pH值、曝氣量、CO32-/Pb2+之比例以及砂層高度。濁度在本研究中只在篩選因子時用來探討砂層的過濾效果。藉由25-1部分因子實驗設計篩選出顯著因子為鉛濃度與pH值後，以反應曲面法建立這兩個因子對去除率與回收率的迴歸模型。依反應曲面法所配適出的去除率與回收率之迴歸模型進行最佳解分析可得一組最佳操作條件。此條件是在pH值為8.6，鉛濃度為1.45e-4M，預測之去除率為100.00%、回收率為94.90%，經實驗後實際得到之去除率為98.75%、回收率為87.28%。模型預測與實驗所得之去除率與回收率的誤差分別為1.25%和8.00%，推測造成誤差的原因是模型缺適值顯著，此意味著模型的預測能力不是相當準確。因此，模型所預測之最佳操作條件在實驗中可能無法達到相同的效果。
將最佳操作條件下培養出來的擔體進行質量平衡分析，經實驗得到的回收率87.28%中有18.59%的鉛離子結晶在擔體上，剩餘的68.69%的鉛被砂層過濾在反應槽內。此結果說明移動結晶床在設計上仍有不足，僅可回收18.59%的鉛。
Fluidized-bed type reactors are frequently used in crystallization process. In this type of reactor, suspend solids (SS) produced by primary nucleation under supersaturated condition would flow out with upward flow, resulting in high SS in the effluent. Installation of filter unit is needed to prevent suspend solids in the discharge. In this study, a novel reactor named “moving-bed reactor” having two functionalities, namely filtration and crystallization, was studied for lead (Pb) removal from aqueous solutions.
According to literatures and design parameters of moving-bed reactor, five potential factors which may affect the process performance, are indentified, including lead concentration, pH, air-flow rate, the ratio of CO32-/Pb2+ and the height of sand. A 25-1 fractional factorial design is utilized to discuss the effect of factors on Pb removal efficiency, recovery efficiency and turbidity. Lead concentration and pH are the most significant factors affecting process performance. Subsequently, RSM with CCD design are used to build regression models for removal efficiency and recovery efficiency, respectively. The best operating condition was determined by uniting two regression models with pH 8.6 and 1.45×〖10〗^(-4)M of lead concentration obtained. Under the best operation condition, Pb removal efficiency and recovery efficiency are predicted to be 100.00% and 94.90%, respectively, by regression models. However, they are only 98.75% and 87.28%, respectively, experimentally. The differences are 1.25% and 8.00%, respectively, for Pb removal efficiency and recovery efficiency. The experiment can not reach the target that models predict because prediction of models is not precise.
Based on quality analysis, crystallized Pb on the sand surface and precipitated Pb filtered by sand are 18.59% and 68.69%, respectively. The result indicated design of moving-bed reactor is inadequate to recover Pb. |
