CHAPTER II SOLUBLE ARSENIC CHARACTERISTICS IN LAGOONS OF FERRIC CHLORIDE AND ALUM COAGULATION WATER TREATMENT PLANTS
The disposal of arsenic-containing wastes will receive more attention once the new arsenic MCL goes into effect. Arsenic that has been removed from water by sorption to ferric or aluminum hydroxides can accumulate in residuals to concentrations many times higher than in the source water. In many cases these residuals are stored in for months or even years. Many parameters may affect the soluble arsenic concentration and speciation in these lagoons. This study gives some baseline results for these lagoons, both with and without microbial activity and biological organic matter. It is also shown that lime can assist in keeping the arsenic sorbed to the solids and prevent release into the water.
Arsenic has long been recognized as a human health concern. It is known to cause skin cancer and has been linked to liver, lung, bladder, and kidney cancer (Smith et al, 1992). Currently, the maximum contaminant level (MCL) for arsenic in public drinking water is 50 µg/L. At this level as many as 13 in 1000 people could die from arsenic-induced cancer based on a lifetime consumption of 1 liter per day (Smith et al, 1992).
The 1996 Safe Drinking Water Act required the EPA to revise the 50 µg/L MCL to a value more protective of human health by January 2001. The purpose of the new standard was to maximize the benefits to human health (by reduction of cancer risks) at a cost that can be justified by the benefits. A new standard was released in January 2001 that requires public water supplies to meet an arsenic MCL of 10 µg/L by 2006, however the rule’s effective date has been delayed numerous times.
Once the revised arsenic MCL is in place it is also likely that stricter regulations will be implemented on the disposal of arsenic-containing residuals. Currently the limit to dispose of waste in landfills is a Toxicity Characteristic Leaching Procedure (TCLP) value of 5 mg arsenic/L. This is 100 times the value of the current drinking water standard.
A common practice among water utilities is to place these residuals in lagoons and eventually allow the lagoons to dry and then ship the waste to a landfill for disposal. Currently, there are no regulations regarding the concentration of arsenic allowed in these lagoons – only that the TCLP of the waste that is hauled away is less than 5 mg/L. The water from these lagoons is normally pumped back into the treatment plant but in the case of unlined lagoons some of this water will find its way into the groundwater.
Many parameters have the potential of affecting the soluble arsenic concentration in storage lagoons, including pH, dissolved oxygen concentration, microbial activity, age of solids, and other trace metal concentrations to name a few. Previous research has shown that the speciation of arsenic is extremely important. Arsenic in its reduced form (As3+) is much more toxic and much harder to remove from water than its oxidized form (As5+) (Edwards, 1994). A recent study by Langner and Inskeep (2000), however, suggests that the reduction of As5+ plays a minor role in the solubilization of As5+ sorbed to ferric hydroxide.
This study will measure changes in soluble arsenic in lagoons at ferric chloride and alum coagulation water treatment plants. The goal is to determine the effect of biological organic matter on arsenic leaching, and what effect lime (calcium) has on the soluble arsenic concentration in these lagoons. The mechanism by which arsenic is sorbed or released from sludge solids is also discussed in the context of the data from this study.
MATERIALS AND METHODS
Sludge Sources. Sludge for all experiments was obtained from six sources: the Los Angeles Department of Water and Power; the City of Great Falls, Montana; the Indiana-American Water Company; the Lockwood Water Users Association; the City of Billings, Montana; and the City of Helena, Montana. These treatment plants have relatively high levels of arsenic in their source waters and all use coagulation in their process trains. The City of Los Angeles, the Indiana-American Water Company, the Lockwood Water Users Association, and the City of Billings all use ferric chloride as their coagulant, while the City of Great Falls and the City of Helena both use alum.
Preliminary Tests. Each sludge was evaluated for solids content using standard method 2540G. Each sludge was also acid-digested per standard method 3030E so that the initial arsenic, iron, and aluminum present could be determined (Standard Methods, 1998).
Ferric Lagoon Simulations. Sludge from the four plants utilizing ferric chloride coagulant was received and divided into four 2-liter acid washed glass jars. The sludge in one jar was analyzed immediately for pH, dissolved oxygen, soluble arsenic, and soluble iron as follows: the sludge was separated into solid and liquid portions using a Whatman #41 filter (20-25 µm pore size). The liquid portion of the sludge was acid digested (EPA Method 3010) and analyzed with ICP-AES for total iron, aluminum, and arsenic. The other three jars were stored in the dark at a constant temperature of 20 °C to simulate lagoon storage. These jars were opened and analyzed at two, four, and six months. (Itle, 2001)
Ferric and Alum Lagoon Simulations with Biological Organic Matter.
Sludge from the City of Los Angeles was used for the ferric lagoon simulations and sludge from the City of Great Falls, Montana was used for the alum lagoon simulations. Three ferric lagoons and two alum lagoons were prepared in glass jars. Each lagoon held a total volume of 700 mL. Ferric sludge was added to three of these such that the solids concentration was 10%. Similarly, alum sludge was added to the other two such that the solids concentration was 6% (10% solids was extremely difficult to pour). Lime was added to two of the ferric lagoons such that the final concentration was 10% and 20% respectively (based on g lime/g dry solids). Lime was added to one of the alum lagoons for a final concentration of 10%. A third alum lagoon with 20% lime was not simulated due to the large amount of aluminum that would be dissolved at this pH. Activated sludge was obtained from the local wastewater treatment plant and used as a source of biological organic matter (BOM). Approximately 10 mL was added to each lagoon. Bactopeptone (500 mg/L), glucose (500 mg/L), and sodium sulfate (200 mg/L) were also added to provide a source of food for the microbes. The jars were stored in the dark at a constant temperature of 20 °C. Periodically, a 20 mL sample was withdrawn and analyzed for pH and soluble arsenic, iron, aluminum, and calcium. At the end of the test sulfide concentration was measured with a portable spectrophotometer. The speciation of the arsenic was also determined at the end of the test according to the procedure given in Edwards et al (1998).
Analysis. All samples were analyzed for arsenic, iron, aluminum, and calcium on a JY Ultima Inductive Coupled Plasma – Emission Spectroscopy (ICP-ES) according to standard method 3120B using continuous hydride generation (Standard Methods, 1998).
QA / QC. A blank sample was evaluated in conjunction with each test to determine if the reagents used affected the concentration of any element of concern. A matrix ‘spike and recovery’ was also performed at random on at least 10% of the samples.
RESULTS AND DISCUSSION
Results are organized into two sections. The first section is a discussion of the bench scale lagoons simulating lagoons at treatment plants that use ferric chloride as a coagulant; specifically, the mechanism(s) that influence the soluble arsenic concentration in these lagoons. In the second section the effect of lime addition to one of these lagoons, as well as an alum plant lagoon, in the presence of BOM is discussed.
Four treatment plants that use ferric chloride as a coagulant were evaluated during this study. Data from a companion thesis will be presented and modeled in an attempt to identify possible mechanisms of arsenic dissolution from these ferric hydroxide sludges (Itle, 2001)
Data and Observations. Dissolved oxygen in most cases was less than 1.0 mg/L and never exceeded 4.0 mg/L in any lagoon (Table 2-1). From these data it appears that an increase in the soluble iron concentration is associated with a corresponding increase in the soluble arsenic concentration (Figure 2-1). The change in soluble arsenic can also be plotted against the change in soluble iron (vs. initial concentration) (Figure 2-2) indicating that as more iron dissolves into the water, more arsenic is also made soluble. A strong correlation (r2 = 0.96) appears if the Los Angeles data are eliminated, reasons for which will be discussed later (Figure 2-3).
One hypothesis to explain why soluble arsenic increases as the soluble iron increases is that dissolution of the iron hydroxide sorbent causes release of arsenic to the water. A model based on Chen et al. (see Appendix) was used to evaluate the data (Chen, 2001). This model utilizes parameters such as ionic strength, pH, total arsenic, and total iron to estimate the percentage of arsenic that will be sorbed to a fresh iron surface. The modeling proceeded as follows. First, the total quantity of iron that was not dissolved was estimated from the size of the bench scale lagoon, the amount of dry solids present, and the amount of iron per weight of dry solid. The undissolved concentration of iron and solution pH was entered for each lagoon at each sampling interval. Total arsenic availability was chosen by matching the initial amount of arsenic sorbed (i.e., a total arsenic concentration was chosen so that the predicted soluble arsenic matched the observed value). All arsenic was assumed to be in the pentavalent (As5+) oxidation state at all times.
Table2-2 shows the results of this modeling exercise. While there are many reasons why the model predictions may vary from field or laboratory data, the results suggest that the dissolution of iron cannot account for the increase of soluble arsenic. It retrospect this seems logical because only 0.4% of the iron sorbent was dissolved.
A second hypothesis is that the arsenic is converted to the reduced trivalent (As3+) oxidation state in addition to dissolution of the iron hydroxide sorbent. Since As3+ sorbs less strongly to iron, this would release more arsenic to the water.
Modeling of this hypothesis was performed as before. Total arsenic was selected based on the soluble As5+ in the water. This time, however, the soluble arsenic was assumed to be completely reduced to As3+ and a new equilibrium established. The model results for both Indiana and Lockwood overpredict the soluble arsenic (Table 2-3). This means that the experimentally observed releases could be explained if a fraction of the arsenic is still in the oxidized pentavalent state. The Billings modeling results indicate that much more arsenic (2 – 5 times) is soluble than would be predicted; however, the trend is correct. Thus, for these waters it seems possible that conversion of As(V) to As(III) could explain the results.
The Los Angeles modeling results require further explanation. The time zero soluble arsenic is extremely high (358 ppb). If the sludge is fresh then the soluble arsenic concentration should be close to the concentration in the plant effluent (e.g. see time zero soluble arsenic concentrations from the other plants). Since Los Angeles obviously does not supply water with 358 ppb arsenic, some arsenic must have leached from the solids before our analysis during transport to our lab.
There are several explanations for the model’s inability to predict the elevated soluble arsenic concentrations at both Billings and Los Angeles. The model is only applicable to ‘fresh’ ferric hydroxide solids and ‘aged’ solids will have a lower surface area and a reduced number of surface sites (Dzombak and Morel, 1990). Less surface sites will mean fewer places for arsenic to sorb and therefore a higher arsenic concentration in solution. Another possible explanation is that the ferric iron may be transforming into ferrous iron and subsequently the ferrous iron combines with sulfide to produce a FeS solid. Although the iron is still in the solid phase FeS has a much lower sorption capacity as compared to ferric hydroxide and therefore more arsenic will be in solution (Davis, 2001). A third explanation may be that arsenic sulfide complexes could be forming as the lagoon environment becomes more reduced and complex formation would tend to increase arsenic leaching although other researchers did not find evidence that they formed (Davis, 2001).
Ferric and Alum Lagoon Simulations with BOM
It is widely known that the presence of bacterial activity and BOM lead to reducing conditions in a non-oxygenated environment. The effect of lime treatment on various lagoons seeded with BOM is presented so that its viability as a treatment option can be evaluated.
Data and Observations. Samples from each lagoon were taken periodically. Concentrations of soluble arsenic, iron, aluminum, and calcium, as well as pH were determined throughout the study. Sulfide concentration was determined at the end of the study. Results are shown in Table 2-4 and Figure 2-4.
Ferric Lagoons. Three ferric hydroxide lagoons with 0%, 10%, and 20% lime were studied. Various trends were noted and will be discussed in turn.
pH: The pH of each lagoon was relatively constant over the course of the study.
Arsenic concentration: The soluble arsenic concentration in the 0% lime lagoon continued to increase as the test proceeded, while the arsenic in the 10% and 20% lime lagoons decreased slightly and then leveled off. It can also be noted that the soluble arsenic concentration decreased as the lime dose was increased.
Iron concentration: The soluble iron concentration in the 0% lime lagoon also continued to increase over the course of the test. The soluble iron in the other lagoons was fairly constant at a value similar to the initial concentration in the 0% lime lagoon.
Calcium concentration: The soluble calcium concentration in the 0% lime lagoon was fairly constant throughout the test. Calcium in the other lagoons decreased during the test. This was probably due to lime or calcium carbonate precipitation.
Sulfide concentration: The sulfide concentration was fairly low for the 0% and 10% lime lagoons (0.066 mg/L and 0.025 mg/L respectively). The 20% lime lagoon, however, had a relatively high sulfide concentration of 0.22 mg/L and a noticeable rotten egg odor.
Additionally, tests were conducted to determine the speciation of the soluble arsenic in each lagoon at the end of the test. The arsenic in the ferric lagoon with 0% lime was 100% As3+ at the end of the test. The lagoon with 10% lime had 49% As3+ and 51% As5+. The lagoon with 20% lime had 64% As3+ and 36% As5+. These results are not surprising since it was expected that the reduction of As5+ would be less when lime was present since the high pH might reduce bacterial activity. It was unexpected, however, that the fraction of reduced arsenic would be slightly higher in the 20% lime lagoon versus the 10% lime lagoon and that sulfides were present in appreciable amount in the 20% lime lagoon. At this time we have no logical explanation for this occurrence. The presence of sulfide might suggest that the precipitation of orpiment (As2S3) is controlling arsenic solubility. At these levels of sulfide, however, there should be no arsenite in solution based on available constants.
As noted in the earlier bench scale lagoon tests, part of the reason arsenic is released as the soluble iron concentration is increased could be that arsenic is reduced to the trivalent oxidation state. This is exactly what was observed in this test as well. In the 0% lime lagoon where 100% of the arsenic is in the trivalent oxidation state, the soluble arsenic continued to increase as the test proceeded. On the other hand, in the 10% and 20% lime lagoons where the soluble iron concentration was fairly constant, the soluble arsenic concentration also remained fairly constant. We think that this is due to the lessened microbial activity in these lagoons; i.e. curtailed microbial activity means less arsenic is being reduced to the trivalent oxidation state.
CHAPTER I EFFECT OF CALCIUM ON ARSENIC RELEASE FROM FERRIC AND ALUM SLUDGES
Materials and Methods
Chemical Conditioning Test Protocol.
TCLP / CaWET Protocol.
Calcium Arsenate Solid Formation Test.
Divalent Cation Theory Test.
QA / QC.
Results and Discussion.
Calcium Arsenate Solid Formation..
CHAPTER II SOLUBLE ARSENIC CHARACTERISTICS IN LAGOONS OF FERRIC CHLORIDE AND ALUM COAGULATION WATER TREATMENT PLANTS
Materials and Methods
Ferric and Alum Lagoon Simulations with Biological Organic Matter
QA / QC
Results and DiscussionSimulations
GET THE COMPLETE PROJECT
EFFECT OF CALCIUM ON ARSENIC RELEASE FROM FERRIC AND ALUM SLUDGES AND LAGOONS