利用過硫酸鹽熱活化法整治受NAPL污染之含水層是一種具有潛力的源頭清除方法。本研究利用MODFLOW-PHT3D程式其具有使用者自訂化學反應模式之功能,將其改寫成可以描述過硫酸鹽與溶解態NAPL反應之動力學模式,同時並將溫度對反應速率常數之影響也一併考慮,利用此模式來模擬TCE以NAPL狀態出現於含水層中時以過硫酸鹽熱活化法整治時NAPL污染物之去除狀況。本研究考慮之各種模擬條件包括:(1)NAPL出現於均質含水層中未加熱與加熱處理之差異;(2)在異質含水層與在均質含水層中加熱處理時NAPL去除效率之差異。由模擬結果發現,加熱區域周圍NAPL污染物濃度下降之速率遠較未加熱區域快速。而當NAPL出現於異質含水層中時,因過硫酸鹽試劑之擴散範圍較廣,因此較易符合擬一階動力學模式之假設。最後,本研究證實MODFLOW-PHT3D程式經修改後可以用來模擬含水層受NAPL污染,而以過硫酸鹽熱活化法整治時污染物之去除狀況,其模擬結果以加藥中心點最為正確。
1. Introduction Remediation of NAPL-contaminated aquifers by thermally activated persulfate is a promising source removal method. In this study, the MODFLOW-PHT3D program is employed to simulate this remediation process. Due to the complexity of its reaction mechanism, a simplified pseudo-first-order kinetics is usually adopted for persulfate oxidation reactions. Since the effects of temperature on the observed first-order rate constant have been correlated via the Arrhenius equation, it is now possible to use such correlation to describe temperature effects on NAPL removal by thermally activated persulfate in MODFLOW-PHT3D. The objectives of this investigation are to simulate the effects of heating on the removal of NAPL and to determine the effects of aquifer heterogeneity on the removal of NAPL. 2. Materials and Methods The MODFLOW package used in this study is PM8 MODFLOW, and the Seer3D program is used for output visualization. Three different remediation scenarios are considered: (1) using persulfate with and without any thermal activation in treatment of TCE in homogeneous aquifers; (2) using thermally activated persulfate in treatment of TCE in both homogeneous and heterogeneous aquifers. The initial TCE liquid saturation is set to be high upstream and distributed in upper aquifers, and decreases gradually downstream and distributed into deeper aquifers, to mimic field conditions. The heating unit employed has a center temperature of 50℃. Its heat radiates gradually to the background temperature of 20℃ within three meters. The hydraulic conductivity in heterogeneous aquifers is obtained by the field generator program in PM8. The kinetic equations describing pseudo-first-order kinetics and its temperature dependence are written in the PHT3D package, which is essentially based on the PHREEQC-2 program using the BASIC programming language. 3. Results and Discussion From the simulation results in homogeneous aquifers, it is seen that persulfate is distributed rather rapidly across the contamination zone, and its concentration is much greater than that of dissolved TCE. The pseudo-first-order kinetics employed in our model can thus be considered as valid, especially in the vicinity of dosing wells. When heating zones are installed to initiate thermal activation of persulfate, removal efficiency increases drastically in critical areas. For example, the observation wells OBS1 and OBS2 need around 100 days to achieve complete removal of NAPL in treatment without heating, but need only 50 days when heating is applied. The simulation results in heterogeneous aquifers reveal that the overall effects of mechanical dispersion are higher than in homogeneous aquifer, and this can lead to faster persulfate distribution. By comparing the concentration distribution between aqueous and NAPL TCE in the same area, it is realized that the rate of NAPL removal around the heating center is close to the mass-transfer control regime. One major limitation of the program is that the simulation generally leads to correct predictions in the vicinity of the dosing wells, but may introduce some errors at larger distance away from the dosing wells, where second-order kinetic equations should be used. 4. Conclusions Using the MODFLOW-PHT3D program with its user-defined kinetic option, one can simulate the in-situ chemical oxidation (ISCO) process in the removal of NAPL contaminants by thermally activated persulfate without developing new codes. Despite the model's limitations outside the dosing well areas, it is still the most appropriate one that is currently available.