The Effect of Molasses Concentration on Bacterial Treatment of Selenium in Agriculture Waste Water in the San Joaquin Valley

In the San Joaquin Valley, selenium concentrations in agricultural waste water have become a serious concern as an environmental pollutant causing birth defects and death in birds, small mammals, and fish. One method of lowering the concentration of selenium in the agriculture discharge is to use biological treatment. A pilot treatment system was developed at Panoche Drainage District using local bacterial strains in reduction ponds to minimize the amount of selenium being discharged into the San Joaquin River. In order to keep operation costs down and determine optimal growth for the bacteria, this project examines how molasses substrate concentration and trace nutrient additions for bacterial growth change the total selenium reduction. Three different concentrations of molasses (0.1g/L, 0.2g/L and 0.4g/L) were prepared in triplicate using influent collected at Panoche and processed at Lawrence Berkeley Laboratory. These samples were analyzed periodically (4 or 7 days) in a controlled environment, using atomic absorption spectrometry to determine selenium concentrations. Final results for varying concentration of molasses between 0.1g/L, 0.2g/L and 0.4g/L did not differ for total selenium concentration. Nor did the total selenium concentration of 0.4g/L with trace nutrient buffer differ from the other concentrations of molasses. Total selenium reduction averaged ten percent from the initial concentration. These results did not meet expectations. Previous experiments had a reduction of eighty percent under similar conditions. This implies that at low concentrations of molasses inadequate carbon is available for maximum bacterial growth and therefore selenium reduction. Introduction Sources of selenium pollution are various and include industrial effluents from thermal power plants, oil refineries, smelting plants, and in the production of semiconductors, pigments, and solar batteries (Kashiwa et al, 2000). Finding a reliable treatment may be applicable to a broad range of industries. The focus of this project is the analysis of the current selenium laden agricultural waste water problem in California. In California’s San Joaquin Valley, an area of extensive agriculture, high levels of selenium have been found naturally occurring in soils (Oswald et al, 2000). For California to maintain its high level of food production, adequate water supplies must be available. Extensive canal systems were built to bring water to the fertile valley, but no canals were built to remove the waste water. Unfortunately, while irrigating fields with selenium rich soils it has been found that selenium leaches into the water. Concentrations between 75 μg/L and 1400 μg/L are measured in the subsurface drainage water (Fan et al, 2001). This excess drainage is then pumped up to the surface and sent to lakes or discharged into the San Joaquin River (Quinn et al, 2000). Some of the effects on organisms when selenium is present in aquatic environments are reproductive dysfunction, deformities, anemia, and death in many species of birds, fish and mammals (Amweg et al, 2003). Since the discovery of selenium accumulation in vertebrates, law makers have tried to establish safe levels of selenium in discharge waters (Amweg et al, 2003). Because much of the selenium cycle is not clearly understood, the establishment of safe levels of selenium in water has been difficult to determine (Fan et al, 2001). The result of this situation is that the Environmental Protection Agency (EPA) has tried to reduce the amount of total selenium entering the watershed as a means to reduce the risk to the environment (Quinn et al, 2000). In 1987, the EPA set a chronic exposure level for freshwater aquatic life at 5 μg/L of total selenium (Fan et al, 2001). Selenium can be found in four different oxidation states (-II, 0, IV, VI). The chemical form of selenium will determine its solubility and availability to organisms (Zhang, 1999). Selenate (selenium VI), selenite (selenium IV), and selinde (selenium – II) are all water soluble and therefore considered to be the most important sources of selenium in water (Amweg et al, 2003). Though its solubility is agreed upon, there are differing opinions about which forms are most toxic. Zhang, Moore, and Frankenberger cite Mikkelsen, Bingham, and Page (1999) to assert that selenate is generally considered to be the most toxic. Whereas Amweg, Stuart, and Weston (2003) assert that organic forms of selenium are thousands of times more bioavailable than selenate and therefore pose the most important risk to the environment. Since most of the selenium from agriculture runoff in the San Joaquin Valley is primarily in the form of selenate, a problem arises as to how to appropriately manage selenium discharge without impacting agriculture production. Methods such as chemical precipitation, catalytic reduction, and ion exchange are effective for the removal of selenite but are not effective in removing selenate (Kashiwa et al, 2000). These methods are also costly (Kashiwa et al, 2000). Due to a lack of affordable treatment of selenium to meet concentration objectives there has been a regulatory shift to reducing the selenium load (Quinn et al, 2000). It may prove to be that bioremediation of selenium by bacteria into less toxic and more a stable form (elemental selenium) is the most cost effective method of reducing the selenium load (Quinn et al, 2000). In Panoche Drainage District near Firebaugh in the San Joaquin Valley an algalbacterial selenium removal system was created to treat drainage water (Oswald et al, 2000). The waste water is not only high in selenium but also in nitrate. Algae were originally used to remove the nitrate from the influent before the reduction pond where bacteria reduce the selenium. The old algae could then act as a carbon source for the bacteria and minimize external inputs into the system. However the algae component has since been discontinued due to experiments that showed better selenium reduction with out the drainage first passing though the algae system (T. Lindqust, 2003). The carbon source for the bacteria could be replaced by many sources found from byproducts of food production, and in the San Joaquin Valley molasses is readily available at the price of $60 to $90 per ton (Quinn et al, 2000). Once in the reduction pond local bacteria strains first remove the nitrate. After the nitrate is removed the bacteria reduce selenate to selenite then to elemental selenium in anoxic conditions (Oswald et al, 2000). Elemental selenium is non-toxic and insoluble (Kashiwa et al, 2000). Removal of elemental selenium from the effluent can then be accomplished by a physical method such as settling ponds or by filtration. The treatment pond in Panoche Drainage District has effectively removed up to eighty percent of the total selenium by the reduction of selenate by bacteria (Oswald et al, 2000). These levels meet regulation needs to reduce the selenium load discharged into surface waters. This project examines the how bacterial treatment is affected by molasses substrate concentration and trace nutrient additions on the total soluble selenium reduction by using biodigestion with anaerobic bacteria. Does more molasses substrate lead to higher reduction of selenium? Are trace nutrients a limiting factor for bacterial reduction? Does more molasses cause an increase in bacterial growth leading to more organic selenium? These questions will help to define the optimum concentration of molasses. This will save money in treatment costs and may give insight to minimize organic selenium discharged. The results from these experiments have direct implications on California’s water shed. The California Water Quality Board and Bureau of Land Management are desperately looking for reasonable solutions to the selenium problem. If a solution can be found that effectively meets standards at an affordable price then these treatment ponds could save California millions of dollars not only in treating the wastewater but by having clean water and a healthy ecosystem. Methods Collection Water samples were collected from the influent to the treatment facility at the Panoche Drainage District located near Los Banos, California. Bacteria were collected from the reduction pond at Panoche. Both water and bacteria samples are kept in one liter plastic bottles and kept cool in a portable cooler for transportation. They are then brought to the Berkeley Lawrence Laboratory and prepared for biological treatment. In the laboratory setting, the reduction pond at Panoche was mimicked by creating an anaerobic environment. Incubation of samples Two tests were run consecutively with the water and bacteria collected from Panoche Drainage District on November 11 2003. The first test had two treatments, addition of 0.1g of molasses per liter of drainage water and 0.2g of molasses per liter of drainage water. The purpose was to test how molasses concentration as a source of carbon for the bacterial growth affected the rate of selenium reduction. In addition, controls were run with this experiment that included a treatment of drainage water with bacteria but no molasses, and plain drainage water called Panoche Influent (PI). The second test also had two treatments, 0.4g of molasses per liter of drainage water and 0.4g of molasses per liter of drainage water with micronutrients in a phosphate buffer. The second test examined if the lack of additional micronutrients were acting as a limiting factor needed to promote bacteria growth and reduction. Controls ran with this experiment were plain PI and drainage water (PI) with bacteria and micronutrients. To accomplish these tests the appropriate concentration of molasses was first added to 2000mL of drainage water for each treatment of molasses. In the case of the micronutrient treatment 1mL of phosphate buffer was added to each liter. The phosphate buffer solution was made by combining 2.0g KH2PO4, 2.1g K2HPO4, and 2.0g NH4Cl brought up to 500mL with deionized water. Then the samples were thoroughly shaken and 80mL of each treatment was added into nine 100mL glass bottles to allow for three different testing dates in triplicate. Next, the bacteria were mixed and added to the samples. This was done in an anaerobic environment to keep the local bacteria from being exposed to oxygen, which slows down the reduction process. One gram of concentrated bacteria flocks were broken up and mixed with 100ml of drainage water. This mixture was then decanted to remove large flocks in order to keep the bacterial solution homogeneous. Two milliliters of bacterial solution was then added to each 80mL sample in a 100mL glass bottle. This was done in an anaerobic hood, (5% H2, 5% CO2, 90% N2 that has an autovacuum, triple sealed and pressurized), and then sealed with rubber tops. The bacteria will only reduce the selenium after the oxygen and nitrate is removed from the water. At that point the samples are placed into a temperature controlled environment, of 28 degrees Celsius, for 4, 7, 14, or 21 days. Everyday the bottles were vigorously shaken for 30sec to prevent bacteria from settling to the bottom of the bottle or sticking to the sides which would reduce the bacteria’s ability to reduce the selenium. The samples are then tested for selenium content by hydride generation atomic absorption spectrometry (AA), using two different processes to determine soluble selenium and total selenium concentrations. Preparation of Samples The AA uses two types of preparation of the samples before analyses, Alkaline Digest or Acid digest. Since the AA can only read selenite it is necessary to convert other forms of selenium into selenite by one of these methods. The Acid Digest uses the filtered sample treated with hydrogen chloride and persulfate to determine the total soluble selenium. The soluble selenium is what is currently the focus of reduction in the watershed as set by the EPA (Oswald, 2000). To filter the samples a 0.22 micron glass filter was used. This is small enough to remove most of the particulates from the sample. Two and a half milliliters of the sample is added into two test tubes. Two and a half milliliters of 12M HCl was added to each test tube. Two percent ammonium persulfate was added, 0.1mL or 0.2mL, into one of the two test tubes for each sample. Ammonium persulfate levels can have an effect on the selenium readings therefore the highest reading of multiple concentrations is accepted. Next the test tubes are set on the heating block for 30 minutes at 98 degrees Celsius. The Alkaline Digest uses an unfiltered sample treated with sodium hydroxide and hydrogen peroxide. The method is similar as above; hydrogen peroxide is added at 1.5mL or 2.0mL to 2.5mL of sample. The test tubes are then set on the heating block for an hour and a half. Next the samples are followed by a hydrogen chloride digest to determine the total selenium concentration in the water. To assure that the AA is running with precision and accuracy the last four test tubes are for quality assurance and quality control (QA/QC). Two split test tubes are prepared from a randomly selected sample and are prepared in the same manner as the rest of the samples. Two spike test tubes are prepared with 2.25mL of randomly selected sample with the addition of 0.25mL of 1000ppb standard selenium stock solution then the process is the same as the rest of the samples. Analysis The AA is turned on and optimized for an hour before samples are ran. Before the samples can be run a new selenium concentration curve must be established. Standards are run at 1, 5, and 10ppb prepared from 1000ppb stock solution. In addition to accurately test the concentration of selenium the samples may have to be diluted to read inside of this curve. Two common dilution factors are 1:7 and 1:40. 1 in 7 Dilution 0.50mL sample and 3.0mL of DDI 1 in 40 Dilution 0.25mL sample and 9.75mL of DDI The results from the two experiments were entered into Microsoft Excel spreadsheet and the progression of selenium was graphed over time. A t-test was run to determine if there was any significant difference between the initial concentration and the final values. Results Due to limited time and money, this experiment was missing some of the components that I had hoped that it would have. Namely, the selenite and organic selenium concentrations would give a fuller understanding of the species in the drainage water over time of treatment. I was only able to test for soluble and total selenium. The first experiment which examined the molasses substrate concentrations had showed no difference in the final concentration of total soluble selenium and total selenium. See figure 1. Initial concentration of selenium was 466ppb ± 60ppb. Final results for varying concentration of molasses between 0.1g/L, and 0.2g/L and did not differ for total selenium concentration (426ppb ± 15ppb, and 405ppb ± 20ppb respectively). The final concentration of total soluble selenium in varying concentration of molasses between 0.1g/L, and 0.2g/L did not show any difference (initial concentration of 466ppb ± 60ppb, to 405ppb ±8, and 418ppb ±13ppb respectively). Varying Molasses Concentrations