Arsenic (III) removal from water using natural sedimentary rocks

• Main Author: Yusof, Nur Zulaikha
• Format: Thesis
• Language: English
• Published: 2015
• Subjects: TA Engineering (General). Civil engineering (General)
• Online Access: http://eprints.utm.my/id/eprint/54708/1/NurZulaikhaYusofPFKA2015.pdf; http://eprints.utm.my/id/eprint/54708/; http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:94657

Arsenic (As) in groundwater is recognized as a threat to public health worldwide. Toxicity of As(III) is greater than As(V) on human health. Most of the arsenic removal techniques are effective only in removing As(V) and not As(III). In this study, ten locally available low cost adsorbents were screened for their capability to remove As(III). The shale sedimentary rock (SSR) and caustic sedimentary rock (CSR) were selected based on high As(III) removal. Based on the characterization of adsorbents using XRD, FESEM-EDX, BET and TGA analyses, it was found that raw SSR contains a mixture of goethite and hematite as its major composition whereas the major composition for CSR adsorbent was goethite. Upon calcination at 500oC, the composition of both adsorbents was completely changed to hematite. The activation of adsorbent was carried out by thermal treatment (250–600°C), acid treatment (0.1–1 M H2SO4) and metal impregnation (0.2-1 M of MnC12 and MgCl2) in order to choose the best treatment method for As(III) removal. Results showed that only by heating the adsorbents at 500°C for 1h, 0.2 g of each adsorbent was capable of reducing the residual As(III) concentration below 10 µg/L, for initial concentration from 100 to 700 µg/L and optimum pH ranges between 3 to 9 after 24 h of contact time. The experimental data were fitted to kinetic and diffusion models, such as pseudo-first order, pseudo-second order, Elovich and intra-particle diffusion models. The pseudo-second order model presented the best correlation (R2=0.999) for all adsorbents studied. The good fit of equilibrium data with the Langmuir isotherm indicated favourable As(III) sorption reaction with SSR-P (0.29 mg/g), SSR-G (0.36 mg/g) and CSR-G (0.65 mg/g), while the CSR-P (0.24 mg/g) was better fit with the Freundlich isotherm. The As(III) sorption occurred with catalytic oxidation of As(III) to As(V) of the surface oxide of adsorbents as evidenced from XPS investigation. Assessment of the breakthrough curve of granular SSR through a column study was examined for the effect of contact time based on operation parameters of bed depth and flow rate. The breakthrough times (10 µg/L) for contact time of 3.167, 4.75 and 6.33 min (10, 20, and 30 cm bed depth, flow rate of 3mL/min) were found to be 28, 90, and 150 h, with treated water of 5.04, 16.20 and 27 L respectively, while for the contact time of 2.85, 3.57 and 4.75 min (flow rates of 3, 4 and 5 mL/min, bed depth of 15 cm), 27, 21.6 and 15 L water can be treated at a breakthrough time of 150, 90 and 50 h respectively. Modeling of breakthrough point was carried out using bed depth/service time (BDST) model, Thomas model and Yoon-Nelson model. The BDST model gave results that were in very good agreement (R2=0.999) with the experimental results. The data obtained from the batch adsorption study was used to train back propagation learning algorithm having a 5-11-1 architecture. The model uses a tangent sigmoid transfer function and a linear transfer function. The network was found to be working satisfactorily since it gave a good degree of correlation (R2=0.919) indicating that the model is able to predict the percentage As (III) removal with reasonable accuracy. This adsorbent proved to be a promising method to meet the needs of rural populations of arsenic contaminated regions since it can effectively reduce arsenic concentration from water to environmentally acceptable levels using a simple method at affordable cost.