奈米薄膜對金門原水中天然有機物之去除研究

Abstract

目前台灣飲用水水質日趨嚴格，然部分淨水廠仍以傳統處理單元為主，尚難符合飲用水之水質標準；此外，水中的天然有機物更是造成後續消毒程序產生消毒副產物之主要前趨物質，故本研究以金門太湖淨水廠為例，利用奈米薄膜(NF270)對進行後續處理，期許達成有效處理水中天然有機物之目的。另外，以樹脂鑑定原水特性，並分析水中天然有機物之分子量分佈，結合薄膜特性分析，藉以進行奈米薄膜處理天然原水之效能評估，尋求最適操作條件。 金門太湖原水中DOC約6.54 mg/L、SUVA254值約1.70 L/mg m，主要以疏水性有機物為主(58.4%)，分子量分布範圍廣泛，小於1K Da之有機物約佔30%，1K - 5K Da約32%，大於5K Da 佔38%。本研究利用聚醯胺(polyamide)奈米複合(thin-film composite)薄膜(NF270)作後續處理，探討操作壓力(75~150 psi)、掃流速度(0.08~0.60 m/s)及前處理程序(SF及UF)對NF270薄膜程序之影響。研究結果顯示，濾液通量衰減隨操作壓力增加而增加，反之，提高掃流速度則可減緩通量之衰減。掃流速度對DOC處理成效影響不大，但UV254 值隨掃流速度增加而提升；另一方面，DOC及UV254處理成效隨壓力增加而提高，但當壓力超過100 psi時，去除率趨近定值。在操作過程中，薄膜表面會有溶質逐漸累積，進而造成阻力而導致濾液通量衰減，隨著操作壓力增加，薄膜結構更為緻密，溶質亦難通過薄膜而提升去除效率；當掃流速度增加時，薄膜表面易產生擾流現象，因在高掃流速度下會形成較少之質量傳送阻力，因此可減少濃度極化，提升濾液量，減緩濾液通量衰減。在不同前處理試驗中，以UF-NF程序可獲得較高之水通量，此外，UF較SF去除較多之疏水性物質，減緩通量衰減；然而，對於天然有機物之處理則以SF-NF較優。因此，以SF作前處理程序，結合掃流式NF薄膜在100 psi、0.30 m/s操作條件下將可有效處理天然有機物並達到省能源之目的。

The drinking water quality standards in Taiwan are becoming more stringent in the future. However, parts of water treatment plants in Taiwan still employ the conventional treatment processes which are hard to meet the standards. Besides, the natural organic matters (NOMs) are the main precursors of the disinfection by-products (DBPs) during the disinfection process. In this investigation, the Tai-Lake raw water in Kin-men Water Treatment Plant in Taiwan was selected to analyze the NOMs by resins and by molecular weight distribution. The NF270 membrane was characterized to evaluate its performance for reducing the NOMs and then to determine the optimum operating conditions. The DOC concentration in Kin-men raw water was found to be approximately 6.54 mg/L, SUVA254 value was about 1.70 L/mg m and the hydrophobic NOMs (58.4%) were the major components. The molecular weight distributed broadly, i.e., lower than 1K (30%), 1K to 5K (32%) and larger than 5K (38%). This study utilized the polyamide thin-film composite (TFC) nanofiltration membrane (NF270) to treat the water samples and investigated the effects of the transmembrane pressure (75 to 150 psi), the cross-flow velocity (0.08 to 0.60 m/s) and the pre-treatment process (SF and UF) on the rejection ratios of NOMs. The results displays that the permeate flux declines obviously with increasing the transmembrane pressure. On the contrary, increasing the cross-flow velocity could ease off the permeate flux declined. The DOC rejection ratio was not affected by changing the cross-flow velocity (0.08 to 0.60 m/s) but the reduction ratio of UV254 decreased with increasing the cross-flow velocity. On the other, DOC and UV254 reduced efficiently with increasing the applied pressure. However, while the pressure was over 100 psi, the reduction ratio was tended to be constant. During operation, the solute might accumulate on the membrane surface and resulted in the permeate flux declined causing by resistance. Increasing the pressure resulted in the membrane structure became more compact and the solute could hardly pass through the membrane. Hence, the rejection ratio increased. Further, with increasing the cross-flow velocity, the permeate flux increased but the flux declined decreased, due to the reduction of the concentration polarization effect. By comparing the pre-treatment process, the UF-NF processes could obtain the higher permeate flux. The UF membrane process rejected more hydrophobic NOMs than by the rapid sand filter (SF) process and decreased the permeate flux declined. But the SF-NF reduced the NOMs more efficiently than the UF-NF. The SF-NF was the proposed treatment process because it can remove the NOMs effectively with lower energy consumption.