Sorption of arsenic on sorghum biomass: a case study Nazmul Haque, Greg Morrison & Gustavo Perrusquía Water Environment Transport, Chalmers University of Technology, Göteborg, Sweden Irene Cano-Aguilera, Alberto F. Aguilera-Alvarado & Moisés Gutiérrez-Valtierra Facultad de Química, Universidad de Guanajuato, Guanajuato, Gto., Mexico ABSTRACT: Large scale, field experiments were conducted for the removal of arsenic from the groundwater of a well located in Guanajuato, Mexico region using non-immobilized sorghum bio- mass (NISB) as a sorbent, which was found highly efficient to adsorb As in previous laboratory experiments. The columns were run under gravity and pump flow conditions. Removal of arsenic under pump flow was slightly higher than the gravity flow due to the steady-state flow conditions. The maximum arsenic accumulation measured was 3.2 and 3.3mg of As/g of NISB for gravity and pump flow conditions, respectively. To determine the optimal hydraulic detention time, columns were operated under different flow rates and the maximum sorption occurred at a flow rate of 10 mL/min. Columns of different dimensions were run to obtain the optimal design parameter between surface loading and volumetric loading of the system. The optimal sorption condition can be achieved through the volumetric design of the system. 1 INTRODUCTION Mining and textile activities can greatly affect the quality of the local groundwater (Bowell et al. 1994, Gray 1997, Rösner 1998). It is more severe in semi-arid lands, where groundwater is the only source of drinking water. High arsenic concentrations in surface and groundwater are recognized to be a problem in mining and textile areas (Gough et al. 1979). It is also recognized that there are potential health consequences for the local population when exposed to drinking water with high arsenic content (Goldman & Darce 1991, Mass 1992, Qvarfort 1992). Uriangato (situated at 20.15°N latitude and 101.18°W longitude) is one of the important centers for textile activities in Guanajuato State of northeastern Mexico (Fig. 1) and, therefore, the city is highly affected by arsenic because these textile industries use a lot of chemicals for their products. One of them might be chromated copper arsenated for the colorization of the cloths. Then, these industries discharge their waste directly to the channel, which passes through the city and links between La Laguna de Cuitzeo and La Laguna de Yuriria as shown in Figure 1 (Virtual Maps & Photos, Mexico 2003). Thus the prevailing hypotheses for the arsenic contamination around this area came from textile industries. A new technology has been evaluated for the removal of arsenic using Sorghum Biomass (SB) as a sorbent (Haque 2003). The laboratory studies using SB both, immobilized and non-immobilized, suggested the following: • Arsenic can be removed effectively below the guideline value from aqueous solutions using SB. • The characterization of arsenic binding to the SB showed that the binding mechanism was pH dependent and the maximum percent of removal of arsenic on SB was at an initial pH of 5.0. • As far as the sorption mechanisms are concerned, it is assumed that arsenic is sorbed mainly by an outer surface mechanism. 247 Natural Arsenic in Groundwater: Occurrence, Remediation and Management – Bundschuh, Bhattacharya and Chandrasekharam (eds) © 2005, Taylor & Francis Group, London, ISBN 04 1536 700 X Copyright © 2005 Taylor & Francis Group plc, London, UK • SB that was saturated with arsenic showed the remarkable ability for arsenic recovery by treatment with 0.1 M HCl. This study presents a detailed investigation of arsenic sorption on SB under field conditions. A pilot scale studies were conducted in the field at an arsenic contaminated site in Mexico to elucidate the design factors using NISB as a sorbent in a column to remove arsenic from the aqueous solu- tions to get the levels consistent for drinking water standards. The columns were run under different flow conditions to determine the maximum efficiency. Optimizations of the columns were studied through the breakthrough curve analysis. 2 EXPERIMENTAL 2.1 Biosorbent The SB from Guanajuato, Mexico region was washed several times with distilled water to remove any particles, or soluble materials because it was waste from the farmers after harvesting their crops. Then, the biomass was dried at 60°C, ground in a bladed mixer to particles less than 0.1 mm in diameter, and sieved through different sieves according to ASTM-AASHO standard methods. These particles were washed with 0.01 M HCl until the supernatant was clear. The biomass was separated from the supernatant then dried in the oven at 100°C for 12 h. This biomass was used for all experiments and it was identified as non-immobilized sorghum biomass (NISB). 2.2 Field experiments The field experiments were conducted at a site called San Jose Curacurio (Fig. 1). The column was packed with NISB. To promote compaction, the column was filled continuously with the suspended biomass in deionized water while the sides of the column were lightly vibrated. The bottom of the 248 Figure 1. Study area in Uriangato, Guanajuato State in northeastern, Mexico. Copyright © 2005 Taylor & Francis Group plc, London, UK column was filled with clean sand (0.25–0.5 mm) to prevent clogging. The column was equipped with a pump and a tank for gravity flow. To estimate the hydraulic detention time of the column, pore volume was divided by the flow rate. The flow rate was observed and adjusted on an hourly basis. All columns had sampling ports that extended into the center of the bed to prevent sampling along the sidewall. Three different sets of experiments were conducted in the field. Firstly, the columns were run under gravity and pump flow conditions to compare the arsenic removal efficiency. Secondly, three columns were run under different flow rates to determine the hydraulic detention time for maximum As removal. Finally, two columns of different sizes but same bed volume were run to test the design parameters surface loading against volumetric loading. The various running condi- tions for each set of column experiments are shown in Tables 1, 2, and 3. The influent solution was taken directly from the extraction well and was never exposed to the atmosphere prior to passing through the columns. The columns were sampled every half-hour interval at the beginning of the experiments and eventually samples were taken less frequently. The samples were collected from the sample ports in 50mL glass vials containing few drops of con- centrated analytical grade nitric acid (Fisher). A plant influent sample was also collected to deter- mine influent arsenic concentration. On several occasions, samples were collected and analyzed for pH, conductivity, and temperature. Finally, the saturated columns were treated by low concentrated HCl acid to recover the sorbed arsenic from SB. 2.3 Arsenic content determination Total arsenic content in all experiments was determined by Hydride Generation, Perkin Elmer MHS 15 coupled to Atomic Absorption Spectrometry, Perkin Elmer Analyst 100, (HG-AAS). All 249 Table 1. Running conditions for arsenic sorption to non-immobilized sorghum biomass (NISB) column under gravity and pump flow. Column Active bed Pore volume Amount of Flow rate Flow condition diameter (cm) (cm) (mL) NISB (g) (mL/min) Gravity 5.5 40 535 150 10 Pump 5.5 40 535 150 10 Table 2. Running conditions for arsenic sorption to non-immobilized sorghum biomass (NISB) column under different flow rates. Flow rate Column Active bed Pore volume Amount of Flow (mL/min) diameter (cm) (cm) (mL) NISB (g) condition 10 5.5 40 535 150 Gravity 20 5.5 40 535 150 Gravity 30 5.5 40 535 150 Gravity Table 3. Running conditions for arsenic sorption to non-immobilized sorghum biomass (NISB) column having different sizes. Column Flow rate Active bed Pore volume Amount of Flow diameter (cm) (mL/min) (cm) (mL) NISB (g) condition 6.5 10 28 535 150 Gravity 5.5 10 40 535 150 Gravity 4.5 10 60 535 150 Gravity Copyright © 2005 Taylor & Francis Group plc, London, UK the samples were pre-reduced from As(V) to As(III) by adding 10% KI according to the pre- determined dilution factor (DF) to prevent the interferences between the two oxidation states of arsenic. Samples were stored at least 30 min in darkness and were analyzed within 3h of sampling. Each time a 1 mL sample was pipetted into a 50mL volumetric flask where 10 mL 1.5% of HCl was added for dilution. Subsequent inline hydride generation with argon and sodium borohydride reductants (3% NaBH 4 in 1% NaOH solution) and HG-AAS detection allowed detection of surely total arsenic species down to 10 g/L at a wavelength of 193.7 nm. The signals for each sample were read to provide a mean and relative standard deviation. Calibrations curves were performed for the removal range of analysis (10 to 5000g/L) and a correlation coefficient for the calibration curve of 0.986 or greater was obtained. The instrument response (50 L of the 1000 mg/L arsenic stock solution give an absorbance of approximately 0.2) (Perkin Elmer 2000) was periodically checked with known arsenic standards. 2.4 Error analyses The adsorption experiments were carried out in triplicate in order to evaluate the experimental reproducibility of the experimental results. The confidence of data generated in the present inves- tigations has been analyzed by standard statistical methods to determine the mean values and con- fidence intervals. Each data set was calculated at 95% confidence level (P Ͻ 0.05) to determine the error margin (William & Ricahard 1973). The correlation coefficient was computed as required to confirm the linear range for a minimum of 12 data points. 3 RESULTS AND DISCUSSION 3.1 Field experiments The influent arsenic concentrations for the entire field study are shown in Figure 2. The variability range from 471 to 541 g/L with an average concentration of 506g/L. An influent sample was taken during each sampling event. The first phase of the column experiments was conducted separately to compare the capacity under gravity and pump flow condition. Effluent arsenic concentrations for both flow conditions are shown in Figure 3. Arsenic removal below the guideline value (10 g/L), recommended by World Health Organization (WHO 2001), was observed for up to 475 and 500 pore volumes for the gravity and pump flow conditions, respectively. Breakthrough (Ͼ65 g/L and Ͼ60 g/L for gravity and pump flow, respectively) occurred after 333 and 400 h for gravity and pump flow, respectively. Then the columns slowly decreased in their ability to remove arsenic from the solu- tion and the column under gravity flow saturated after 1150 pore volumes and the column under pump flow after 1200 pore volumes. The maximum sorption capacity and the supply of fresh water (below the guide line value provided by WHO) for the column under gravity flow were 3.2 mg of As/g of NISB and 254 L and for the column under pump flow were 3.3 mg of As/g of NISB and 267 L. The slight difference of results between the two flow conditions is attributed due to the vari- ability in flow. A steady state flow was achieved under pump flow condition which might be obtained under gravity flow condition if the water level can be kept constant in the tank. It was necessary to measure the pH, redox potential, and the temperature of the influent solutions. Because all these parameters are important for sorption of arsenic to SB. From the pH profile experiments (Haque 2003), the results showed a strong influence of pH and the maximum removal of arsenic was observed at an initial pH range of 4.0–6.0. During the field experiments, the initial pH ranged from 4.6 to 6.4 which were favorable for the sorption of arsenic to SB. However, it might not be same for other conditions where the solutions pH changes a lot. The second set of the column experiments were carried out to determine the hydraulic detention time. The breakthrough curves for different flow rates are shown in Figure 4. It can be easily seen from the breakthrough curves that arsenic can be sorbed more with less flow rate. From the previous 250 Copyright © 2005 Taylor & Francis Group plc, London, UK 251 0 100 200 300 400 500 600 0 200 400 600 800 1000 1200 Pore Volumes µg/L Figure 2. Influent arsenic concentrations of the natural source. 0 1 0 200 400 600 800 1000 1200 Gravity Flow Pore Volumes 0,8 0,6 0,4 0,2 C/Co Pump Flow Figure 3. Breakthrough curves under pump and gravity flow conditions. The flow rate of the incoming As solution into the column was 10mL/min. 0,0 0,2 0,4 0,6 0,8 1,0 0 200 400 600 800 1000 1200 Pore Volumes C/Co 10 ml/min 20 ml/min 30 ml/min Figure 4. Breakthrough curves under different flow rates. The columns were run under gravity flow. Copyright © 2005 Taylor & Francis Group plc, London, UK sorption isotherm analysis (Haque 2003), it was noticed that the arsenic binding was rapid within 1 hour. The rapid binding of the metal ions by SB could indicate the metals are being adsorbed onto the surface of the biomass, instead of absorbed. It has been found that the adsorption processes in most of the biomasses are within 5 minutes and only the surface areas of the sorbent are highly active during this time period (Vaishya and Prasad 1991). High flow rate might break the binding surface between arsenic and SB and, therefore, the sorption was less. The final set of column experiments were carried out to determine the design parameter surface or volumetric loading. The size of the columns was different but the total volume of the sorbent was same. Figure 5 shows the breakthrough curve among the different sizes of column but same bed volumes. The breakthrough curves were at 200, 450, and 650 pore volumes for the column diameter of 6.5, 5.5, and 4.5 cm, respectively. The sorption capacity of the column increased with the lower diameter i.e. the higher length of the column. So, the efficiency of the column can be increased by volumetric design but the ratio of column length to column diameter should be ϳ5 to avoid plugging of the column and other undesirable effects. From the batch experiment (Haque 2003), it has been found that the recovery of arsenic from the saturated biomass is achievable. Thus, it was intended to recover sorbed arsenic from the saturated column by the addition 0.1M HCl. Approximately 95% of the arsenic bound was recovered from the column as the acid was passed through it. The little arsenic remaining may be responsible for the slight decrease on arsenic binding capacity, when the column will be used again. 4 CONCLUSIONS This research clearly demonstrated the effectiveness of SB to remove arsenic under natural conditions. More specific conclusion can be drawn from the results of each specific experimental case: • Due to the steady-state flow condition, arsenic removal under pump flow was slightly higher than the gravity flow. • The maximum arsenic accumulation measured was 3.2 and 3.3mg of As/g of NISB for gravity and pump flow conditions, respectively. • The sorption capacity of the column can be optimized through the lower flow rate (10 mL/min) without alteration on the efficiency of the column. • The system can be designed on a volumetric loading basis. • The saturated column can be reused after recovery of bound arsenic by using 0.1M HCl. 252 0 0,2 0,4 0,6 0,8 1 0 200 400 600 800 1000 1200 Pore Volumes C/Co Column Dia. 6.5 cm Column Dia. 5.5 cm Column Dia. 4.5 cm Figure 5. Breakthrough curves among the different sizes of column but same bed volumes. The flow rate of the incoming As solution was 10 mL/min under gravity flow. Copyright © 2005 Taylor & Francis Group plc, London, UK Finally, this case study showed that the groundwater of Guanajuato, Mexico region could be treated for drinking purposes by using SB. The adsorbents are not only inexpensive but also available. This innovative technology provides a reusable material (SB) which is not biodegradable, but environmentally friendly. ACKNOWLEDGEMENTS The Swedish Foundation for International Cooperation in Research and Higher Education (STINT) are greatly acknowledged for their financial support. REFERENCES Bowell, R.J., Morley, N.H. & Din, V.K. 1994. Arsenic speciation in soil pore waters from the Ashanti Mine, Ghana. Appl Geochem 9: 15–22. Fotos y Mapas Virtuales Departmento, Mazatlan, Mexico 2003. Ave. Cruz Lizarraga, No. 712–3, Col. Palos, Prietos, Sinaloa, Mexico 82100. Goldman, M. & Darce, J.C. 1991. Inorganic arsenic compounds: are they carcinogenic, mutagenic, teratogenic? Environ Geochem Health 13: 179–191. Gough, L.P., Shacklette, H.T. & Case, A.A. 1979. Element concentrations toxic to plants, animals, and man. USGS Bull 1466. Gray, N.F. 1997. Environmental impact and remediation of acid mine drainage: a management problem. Environ Geol 30: 62–71. Haque, N. 2003. Sorption of arsenic on Sorghum Biomass and Iron Fillings. Master Thesis. Department of Water Environment Transport, Chalmers University of Technology, Göteborg, Sweden. Mass, M.J. 1992. Human carcinogenesis by arsenic. Environ Geochem Health 14: 44–54. Perkin Elmer 2000. User’s Guide for Mercury Hydride System, USA. Qvarfort, U. 1992. The high occurrence of arsenic in Macadam products from an iron mine in central Sweden: significance for environmental contamination. Environ Geochem Health 14: 87–90. Rösner, U. 1998. Effects of historical mining activities on surface water and groundwater: an example from northwest Arizona. Environ Geol 33: 224–230. Vaishya, R.C. & Prasad, S.C. 1991. Adsorption of copper (II) on sawdust, Indian J. Environmental Protection. 11(4): 284–289. WHO 2001. Arsenic compounds: In Environmental Health Criteria 224, 2nd edition, World Health Organization, Geneva. William, M. & Ricahard, L.S. 1973. Mathematical Statistics with Applications. Duxbury Press, Massachusetts: pp. 274–290. 253 Copyright © 2005 Taylor & Francis Group plc, London, UK . assumed that arsenic is sorbed mainly by an outer surface mechanism. 247 Natural Arsenic in Groundwater: Occurrence, Remediation and Management – Bundschuh, Bhattacharya and Chandrasekharam. where groundwater is the only source of drinking water. High arsenic concentrations in surface and groundwater are recognized to be a problem in mining and textile areas (Gough et al. 1979). It. sample ports in 50mL glass vials containing few drops of con- centrated analytical grade nitric acid (Fisher). A plant influent sample was also collected to deter- mine influent arsenic concentration.