REMOVAL OF SELENIUM FROM REFINERY WASTEWATER USING THE HONEYWELL UOP XCEED BIOREACTOR SYSTEM

Transcription

1 REMOVAL OF SELENIUM FROM REFINERY WASTEWATER USING THE HONEYWELL UOP XCEED BIOREACTOR SYSTEM Lori C. Donovan, R.E. Hanson, William Sheridan & F. Stephen Lupton Honeywell UOP UOP Industrial Wastewater Technologies Des Plaines, Illinois, INTRODUCTION Selenium is a very protean contaminant; speciation is complex with selenium existing in the oxidation states 2, 1, 0, +4, and +6 in nature, in both inorganic and organic forms [1]. Furthermore, selenium is both an essential trace nutrient and a highly toxic compound depending upon concentration [2]. As such, regulatory bodies are increasingly concerned about the anthropogenic release of selenium into the environment as a result of mining, power generation, and petroleum refining and are implementing strict limits for the release of selenium from these operations [3]. A number of physical/chemical and biological treatment technologies have been applied for remediation of selenium in wastewaters discharged from these industrial activities. These include chemical precipitation, adsorption, and biological treatments [4,5]. Multiple selenium species may be present in wastewaters such as selenite, selenate, and selenocyanate, which can be especially true of refinery waste streams. The varying forms of selenium that may be present pose a problem for many of the physical/chemical approaches which are often most effective for a single selenium species. Biological treatment has the potential to convert various forms of selenium to elemental selenium, which is relatively non-toxic, and separate the selenium from the wastewater streams as an inert solid associated with the biomass present in the solids recovered from the treatment process [6]. These solids typically pass the TCLP test for extractable selenium and are not considered hazardous waste, making it an economic route for remediation. For selenium reduction to occur, the treatment system must be anoxic. Nutrients required for bacterial growth, including carbon, nitrogen, and phosphorus, must also be present in the system. These can be naturally occurring or added during a biological remediation process to optimize growth. During treatment, bacteria couple the oxidation of organic carbon with the reduction of selenium to produce energy for growth and to generate more bacterial cells. If a more 2016 UOP LLC, A Honeywell Company. All rights reserved. Page 1

2 energetically favorable electron acceptor, such as oxygen or nitrate is present in the system, the bacteria will preferentially reduce these compounds prior to reducing selenium, increasing the consumption of organic carbon in the system. This paper highlights the evaluation of biological treatment using Honeywell UOP s XCeed bioreactor process technology to remove selenium from the waste water treatment plants end-ofpipe discharge location from a mature western U.S. refinery. The refinery is configured for production of gasoline and diesel transportation fuels, a crude oil slate, which is the source of the selenium. The refinery targeted to reduce selenium present in the final discharge from a level of approximately 200 µg/l to less than 50 µg/l in order to discharge to a local municipal treatment plant. At the point of discharge, the selenium is predominately in the selenite and selenate forms. The plant had previously run trials on various commercially available biological and chemical treatment technologies and subsequently evaluated the XCeed system to treat this effluent. The refinery needs for selenium treatment included: A robust treatment system that could handle influent water quality fluctuations while still achieving less than 50 µg/l total recoverable selenium in the treated effluent without post-treatment Cost effective treatment An easy to operate system with low labor and maintenance requirements Small footprint The ability to scale up treatment as refinery capacity grows Laboratory bench testing and a field demo pilot test were used to validate the XCeed treatment system performance and collect data for full-scale design. On-Site Pilot Demonstration of XCeed Technology for Selenium Removal from Refinery Wastewater Based upon the positive results from the bench scale treatability test, an on-site pilot study was conducted at the refinery site using a mobile pilot system. The XCeed pilot system was operated onsite, including inoculation, startup, and continuous forward flow. The pilot treated a slipstream of clarified water from the existing wastewater treatment plant (end of pipe) and evaluated different hydraulic test conditions. Major testing objectives for the selenium pilot included the following: Produce an effluent that contained less than 50 µg/l total recoverable selenium Optimize the design hydraulic retention time treatment performance for the XCeed system to minimize footprint and investment costs Develop engineering data to aid in the scale-up to full-scale treatment. Operate during typical refining operations, including upset conditions, to test system robustness Page 2

3 The XCeed field pilot was installed adjacent to the existing wastewater treatment plant. The system was designed to remove selenium using UOP s immobilized cell biological treatment technology, and consists of a series of bioreactors and UOP s proprietary biological growth media, along with all associated process equipment, instrumentation, and laboratory equipment necessary for operation. The system was inoculated using nitrate reducing cultures developed during the laboratory study. Organic carbon and phosphoric acid (an essential nutrient), were dosed inline prior to the first bioreactor to provide nutrients for biological growth. Nitrogen was not required due to nitrate present in the influent water. Influent Water Quality The pilot influent water quality for the testing period is summarized in table 1. The influent water quality demonstrated some variability, and included a period of dilute concentrations, at the beginning of the study, due to refinery startup as well as a period of high nitrate concentrations in the influent due to a process upset in the Sour Water Stripper. Pilot influent total recoverable selenium concentrations were variable, ranging from 0.05 to mg/l and averaging mg/l. The majority of the selenium present in the influent was dissolved, with selenite comprising approximately 80% of the dissolved selenium. The remainder of the selenium was primarily present as selenate. The influent water contained naturally occurring elevated dissolved oxygen (DO) and nitrate levels, which increased the organic carbon demand, since both DO and nitrate must be removed prior to selenium reduction. The DO averaged 3.7 mg/l and ranged from 1.7 to 5.9 mg/l. The nitrate averaged 16.6 mg/l NO3-N, and ranged from 4.3 to 41.9 mg/l NO3-N. Table 1. XCeed Selenium Pilot Test Influent Water Quality Summary Parameter Average ph (SU) ORP (mv)1 64 Dissolved Oxygen (mg/l) Nitrate (mg/l NO 3-N) 16.7 Selenium, total recoverable (mg/l) Selenium, filtered total recoverable (mg/l) Page 3

4 Effluent Water Quality The effluent water quality for the complete pilot test is summarized in table 2 below. The effluent water quality showed some variation, but on average, met the selenium treatment goal for the pilot study. Table 2. XCeed Selenium Pilot Test Effluent Water Quality Summary Parameter Average ph (SU) 7.47 ORP (mv) -335 Dissolved Oxygen (mg/l) Nitrate (mg/l NO 3-N) 1.0 Selenium, total recoverable (mg/l) Selenium, filtered total recoverable (mg/l) Selenium Removal The pilot test selenium profile is shown in figure 1. For all test conditions, effluent total recoverable selenium concentrations were typically less than the 0.05 mg/l total recoverable selenium target for discharge to the local sewer authority with two exceptions. The first excursion occurred during the 24-hr HRT after beginning forward flow. During this period, the microbial population was likely still adapting and selenium concentrations were below the limit for subsequent sampling events. The other excursion occurred after a power failure resulted in no nutrient addition to the bioreactor. After the power failure, influent nitrate concentrations increased. This combination of factors likely limited nitrate reduction in the bioreactors thereby limiting selenium reduction until complete nitrate reduction resumed. Despite still having a notfully developed microbial population, it only took approximately one week for the bioreactor to recover from the upset and stabilize performance similar to the bioreactor performance prior to the upset. Other key points include: During pilot testing, total recoverable influent selenium concentrations were never less than the discharge target. For approximately 30% of effluent sampling events, total recoverable selenium was not detected in the sample. Page 4

5 Nitrate (mg/l NO 3 -N) Selenium Figure 1. Removal of Selenium from Refinery Wastewater in the XCeed Bioreactor Influent Effluent Se Limit Nitrate Spike No Nutrients Days of Operation Nitrate Removal The pilot test nitrate profile is shown in figure 2. Nitrate concentrations varied in the influent, with a large influent nitrate concentration spike occurring on day 30 due to an upset in upstream operations. Although influent nitrate concentrations varied, the effluent nitrate concentration remained relatively stable averaging less than 1 mg/l NO3-N when upset conditions were excluded. Nitrate breakthrough was observed during the power loss, when sufficient nutrients were not present for nitrate reduction due to power interruption. Breakthrough was also observed during the high nitrate loading conditions until nutrient dosing rates were adjusted and the bioreactor adapted to the new loading conditions. Figure 2. Removal of Nitrate from Refinery Wastewater in the XCeed Bioreactor Influent Effluent No Nutrients Days of Operation Page 5

6 ORP Oxidation-Reduction Potential Oxidation-reduction potential (ORP) is the measurement of the potential to oxidize contaminants. Anoxic systems remove available dissolved oxygen, reducing ORP accordingly; additional reducing potential will further reduce ORP. Figure 3 shows the ORP during pilot testing. Figure 3. ORP Profile for the Treatment of Refinery Wastewater in the XCeed Bioreactor Influent Effluent No Nutrients Days of Operation Conclusions Laboratory and pilot testing successfully demonstrated that Honeywell UOP s XCeed bioreactor technology can meet the total recoverable selenium target of 50 µg/l without the need for post-treatment, allowing the refinery to comply with local regulations. The main testing objectives to achieve the local sewer total recoverable selenium limit, validate the design HRT treatment performance, and collect sufficient data to optimize the design of the full-scale system were met. Selenium removal. Effluent total recoverable selenium concentrations were typically less than half of the 50 µg/l total recoverable selenium target with two exceptions: the beginning of forward flow, where the biofilm was likely still adapting to conditions and during a process upset that was caused by a combination of loss of nutrient feeds and high nitrogen loading. This should not be a concern in the commercial system considering it will be fully developed and have a large biomass. Refinery wastewater quality. The wastewater was characterized during the laboratory and pilot studies to improve the understanding of influent water quality for the full scale treatment system. Pilot influent total recoverable selenium concentrations were variable. Approximately 80% of the influent was selenium and was present as selenite. The influent was also characterized by elevated DO and nitrate, which increase the organic Page 6

7 carbon demand as both DO and nitrate must be removed prior to selenium reduction. The influent temperature and ph were within the range necessary for selenium reduction. Dissolved oxygen and nitrate removal. Variable DO in the refinery wastewater was consistently reduced to less than 0.2 mg/l in the effluent during pilot testing. Likewise, although influent nitrate concentrations varied, the effluent nitrate concentration remained relatively stable and averaged less than 1 mg/l NO3-N when upset conditions were excluded for pilot testing. Nutrient addition. While the refinery wastewater contained some ammonia-nitrogen and phosphorus, nutrients were still required for microbial growth. Laboratory testing showed a carbohydrate based substrate provided the best selenium reduction. Organic carbon was added to provide an organic substrate for biological growth. Phosphoric acid was also added to provide phosphorus necessary for biological growth. A nitrogen based nutrient was not required as ammonia-nitrogen was present in the influent, and in the absence of ammonia-nitrogen bacteria can use nitrate-nitrogen for growth. Microbial inoculum. Laboratory testing showed nitrate reducing bacteria were more effective than sulfate reducing bacteria at reducing selenium in the refinery wastewater. The pilot was inoculated with nitrate reducing bacteria cultivated from the laboratory study, and successfully removed selenium after a comparatively limited inoculation and start-up period of less than seven days. References 1. P. H. Masscheleyn, R. D. Delaune and W. H. Patrick, Arsenic and Selenium Chemistry as Affected by Sediment Redox Potential and ph, Journal of Environmental Quality, Vol. 20 No. 3, p , M. Lenza, P. N.L. Lensa, The essential toxin: The changing perception of selenium in environmental sciences, Science of the Total Environment, Vol. 407, p , Selenium in Drinking-water Background document for development of WHO Guidelines for Drinking-water Quality, World Health Organization, WHO/HSE/WSH/10.01/14, 4. K. Smith, A. O. Lau, and F. W. Vance, Evaluation of Treatment Techniques for Selenium Removal Paper IWC 09-05, Proceedings the 70th Annual International Water Conference held 4-8 October 2009, Orlando, Florida, USA, Page 7

8 5. S.Santos, G. Ungureanu, R.Boaventura, C. Botelho, Selenium contaminated waters: An overview of analytical methods, treatment options and recent advances in sorption methods, Science of the Total Environment Vol , p , R. Ormeland, J. Hollibaugh, A. Maest, T. Presser, L. Miller, and C. Culberston, Selenate Reduction to Elemental Selenium by Anaerobic Bacteria in Sediments and Culture: Biogeochemical Significance of a Novel, Sulfate-Independent Respiration Applied and Environmental Microbiology, Vol. 55, p , 1989 Page 8