Removal of mercury from surface waters using coagulation

sherman island DWR sliderboxDr. Tamara Kraus with the USGS presents the latest laboratory and field test results using metal-based salts as a coagulant for mercury removal

Mercury contamination is a widespread problem, not just locally but also globally, and its serious effects on health are well known. Mercury in the environment comes from many sources: it is transported by wind and rain from local and global emission sources, it can be present in urban and industrial wastewater, and it can be naturally occurring in soils and springs, particularly in the Coast Range. Here in California, many of the state’s streams, rivers, and reservoirs are impaired for mercury, much of it a legacy from the Gold Rush days when mercury was mined in the Coast Range and transported to the Sierras for use in gold mining.

Open hydraulic mining pit, 1890-1900
Open hydraulic mining pit, 1890-1900

Gold-rush miners sought gold by eroding entire hillsides with high-pressure water cannons in a practice known as hydraulic mining; the sediment was then run through sluice boxes, where mercury was added to bind to gold.  Through this process, large quantities of the heavy metal made their way into sediment downstream, filling the rivers and streams with contaminated sediment and causing flooding in the Central Valley.  It is estimated that between 10% and 30% of the 26 million pounds of mercury brought into the gold mines for the gold recovery process was lost to the environment.  In 1884, the courts banned hydraulic mining, but the damage had already been done.

Even today, more than 150 years later, mercury is still found throughout the San Francisco Bay-Delta estuary and its watershed at elevated concentrations in the sediments and biota.  This mercury is of particular concern as significant acreages of tidal marsh, floodplain, and other wetland habitats are planned for the Bay Delta.

Suisun Marsh wetlands Photo courtesy of DWRMercury in its elemental form is non-toxic, but when the it reaches a wetted environment, such as a wetland or reservoir, it settles to the bottom where bacteria in the sediments convert it to the more toxic, organic form called methylmercury.  Methylmercury is bioaccumulative pollutant which concentrates as it moves up the food chain, from algae to zooplankton to prey fish and the predators that eat them, such as trout and bass.  Humans are exposed to harmful amounts of mercury through eating contaminated fish and shellfish.

As large as the problem is, there are significant efforts here in California and elsewhere to address the problem.  Some of the techniques include dredging and pumping of contaminated soils, chemical treatments with other compounds to make it biologically unavailable, physical barriers to contain it so it no longer spreads, using microbes to demethylate the methylmercury to a less toxic form, using plants to remove or immobilize it, and manipulating water quality parameters such as oxygen or pH to ensure that methylmercury production will not occur.

In this presentation, Dr. Tamara Kraus with the USGS Water Science Center summarizes the results of field and laboratory studies involving the use of coagulants to remove the methylmercury from the water column.  A number of researchers and agencies have been participating in the study which is known by the acronym LICD, or Low Intensity Chemical Dosing, which began in the early 2000s.

Kraus Mercury_Page_03Dr. Tamara Kraus began with the basics of coagulation.  “Coagulants are commonly used to remove dissolved organic carbon from surface water,” she said.  “It’s used by drinking water treatment plants, wastewater treatment plants, and in stormwater treatment; it’s also used to remove phosphate from lakes.  Basically a coagulant is anything that clarifies the water.  Romans used to do it by adding all kinds of stuff to pull particles just to make water look clearer so you wanted to drink it.  It’s been used for humans for centuries.”

There are metal-based coagulants; charcoal can also be considered a coagulant – anything that pulls things out of the water column, she said.  “The processes are charge neutralization, complexation, absorption and entrapment,” she said.  “There are a lot of different things happening under coagulation, both the removal of the dissolved organic carbon which you can think of that as binding, but then there’s also the physical properties that are happening as the particles form larger particles and then settle out of solution.”

Kraus Mercury_Page_05The original LICD study was designed to look at the use of coagulation on the central islands of the Delta that were at that time considered a major contributor of dissolved organic carbon to the waterways and that was bad for drinking water quality, she said.  “It was designed originally to look if we could remove dissolved organic carbon in situ rather than waiting for the water to reach the drinking water treatment plants and have it coagulate there.  Also can we have a secondary benefit to have constructed wetlands that have been shown to help with subsidence mitigation, actually build up some of these areas next to the levees and provide levee stability.”

It was in 2009 when the stop work order came so they weren’t able to build the field site, so they took the opportunity to add mercury to the study, so there were some benefits to the stop work order, she said.  In addition to looking at total and methyl mercury and how it’s impacted by coagulation, we also wanted to look at ecosystem effects, and bioavailability to organisms in these wetlands, she said.

There were three studies that were done.  The first one was a classic study where we looked at what happens when coagulants are added to water that has dissolved organic carbon and mercury from agricultural drainage waters.  Basically what they found was that mercury follows the dissolved organic carbon.  “The mercury is bound with the DOC; if you pull DOC out of the water, you’re pulling the mercury out of the water,” said Dr. Kraus.

Kraus Mercury_Page_06All of the lab studies were done using filtered water from Twitchell Island agricultural drains.  “About 70% of the total mercury is in the dissolved phase in this system, but we can assume that the particles will get pulled out too, because if you’re pulling out the dissolved stuff, and you can see that you are clarifying the water, you’re probably pulling out most of the particles,” said Dr. Kraus.  “Again, very effective removal; we found that at the highest coagulant doses, we could remove 80% of the DOC and the methyl mercury and we had even more effective removal of the inorganic mercury.”

Kraus Mercury_Page_07The second study utilized incubations where they put the floc (or precipitate that is formed in the coagulation process) and tried to hammer it with some reducing agents to see if the material was stable  in order to try to predict what the stability of this material would be in a wetland environment; it was stable in the bottle that was in the incubation, she said.  “There was no microbial activity involved here or evidence of DOC release; there was not a mercury component to that.”

The third study used isotopically-labelled mercury to determine how quickly does this interaction occurs between mercury entering the system, binding with DOC, and then getting pulled out, she said.  “Really it happened immediately, so there is really rapid association with the mercury.”

Kraus Mercury_Page_08Dr. Kraus then turned to the field studies which occurred on Twitchell Island.  There were nine cells 36×12; they are replicated cells of three treatments.  Three cells were uncoagulated drainage water that was flowed through the wetlands cells (shown in blue on the diagram); the other cells were treated with coagulants, three with iron sulphate (shown in green) and three with polyaluminum chloride (shown in orange).

Kraus Mercury_Page_10The mercury sampling was focused on the inflows and the outflows.  “There is the untreated water that’s entering, the water receives coagulant and it next goes through these pipes that mixes it and then it gets released into the wetlands,” she said.  “We’re sampling it in the pipe, so it’s already had the coagulant for a couple minutes, and then it flows through the wetlands.  We had an average residence time of about 3 days, but it ranged from about 2 to 7 days.”

We did have the wetlands fully instrumentated with flows at the inflows and the outflows so we have the concentration and the flow data, and then at the outflows, we also collected at the pipe where water was spilling out over a weir out into the drainage ditches,” she added.  “We have some data that we collected within the wetlands cells and we also did another study where we looked at bioaccumulation on fish living in these different wetlands.”

Kraus Mercury_Page_11Dr. Kraus noted that most of the data she is talking about is water that’s coming out over the weir and exiting the cells, but the water samples collected from within the water column looked quite different.  “There were chunks of algae floating around in there; there were plants, and the floc itself was not really settling out as maybe we had thought it would …  so just keep that in mind, sometimes what you sample flowing out of a wetland isn’t exactly what an organism living in the wetland would see, it’s often very different.

They didn’t really start collecting data until the fall of 2011; there is about 16 months where DOC concentration data was collected every week; there is also a one-year period where mercury samples were collected approximately monthly.  “The way we collected and analyzed samples is that we collected a whole water sample, we filtered it, we’d send the filtered water and the filters to the Wisconsin Lab, and then we would sum those together total, so we had filtered, particulate, and then the sum of those two,” she explained.

Dr. Kraus then presented some graphs showing the results of the study for dissolved organic carbon.  “Even within this coagulation study, the data tells us something about any wetland in general,” Dr. Kraus said.  “There’s a wealth of information about what does a constructed wetland do, with a high density of cattails in it, with about a three-day hydrologic residency time so we could look at the data that way.

We could also look at just what does coagulation do alone, so if we compare all the three inflows to each other, we can say, what happens if you’re just going to coagulate, you don’t necessarily plan to pass your water through a wetland, maybe you’re going to put in a retention basin like they do at water treatment plants and dispose of that offsite.  We can also compare what happens when you don’t flow that water through a field or a wetland,” she said.

Kraus Mercury_Page_13Dr. Kraus then presented a chart showing DOC concentration data, which ranged from 10 to 45 milligrams per liter.  “That is lower than we liked,” she said.  “We would have liked to situate our wetland right next to the main drain that pumps water off of the islands where concentrations are generally higher, but there are logistical reasons and regulatory reasons why we couldn’t do that, but a DOC of 10 is still pretty high on these peat soils.”

“I just want to show you the DOC concentrations in this system are higher in the winter because that’s when the rains come and the whole island floods up and there’s more drainage off of the peat material into these drainage ditches,” she said.  “If you compare the filled lines (or the inflow) to the outflow lines (the unfilled symbols), you can see that there was DOC production due to passage through the wetlands, except in the colder winter months.”

Kraus Mercury_Page_14She then presented a graph showing inflows, and noting that they were ratcheting up and down the coagulant dose on a daily basis as they were trying to target about a 60 – 85% DOC removal rate.  “We did a good job,” she said.  “We had in situ sensors … so we were able to get a really good control; so if the DOC concentration was higher, we would add more coagulant per liter of water.”

Kraus Mercury_Page_15Passage through the wetlands added dissolved organic carbon also in the coagulation treatment cells, but in different amounts.  So there’s a story there.  Something in these coagulation treatments were affecting the export of DOC from these wetlands.”

Dr. Kraus then turned to the mercury data.  “About 70% of the dissolved organic carbon was in the dissolved phase,” she said.  “About 70% is filtered and the 30% is particulate, and then after you add the aluminum coagulant, theoretically there should not be additional mercury because all you’ve done is add coagulant, but at the end that it did look like in general that maybe we were getting higher concentrations of mercury after adding the iron coagulant, which brings up the question of whether there was some mercury contamination in some of the coagulants we used.  But the error bars, in this case it wasn’t significantly different.  The coagulant immediately pulls the mercury out of solution and puts it into the particulate phase instead of the dissolved phase.”

Kraus Mercury_Page_18After the sample passes through the wetland, the particulate falls out of solution and settles to the bottom of the wetland.  “In this case, in January, we did not see much production of total mercury due to passage through the wetlands, so we saw this large benefit to the coagulation wetland system.  We transferred the mercury from the dissolved phase to the particulate phase and then those particles were retained in the wetlands.”

Kraus Mercury_Page_19However, in other seasons, it wasn’t so simple, she said.  “In some other months, even though you might have seen a loss of the particles, you also saw an increased in the filtered mercury fraction, so there was some production going on due to passage through the wetlands.  But .. however, you still saw an overall benefit, so you took a lot out but then you added some more in.  But not more to make it worse than not doing anything.”

Kraus Mercury_Page_20The methylmercury had a similar story to the total mercury, Dr. Kraus said.  “We had very high production of methylmercury during some periods of time due to passage through the control wetlands, less in the iron and aluminum treatments, but still some production, so there are seasonal effects.

Kraus Mercury_Page_21Dr. Kraus then presented another graph, noting that this was from the period of time when they collected samples for mercury on a monthly basis.  She reminded that they were ratcheting up and down the coagulant dose to get 60-75% removal of filtered total mercury.

Kraus Mercury_Page_23What happened due to passage through the wetlands?  We saw that there was some increase in filtered total mercury, though in some times of the year in the control wetland, there was a decrease and sometimes we saw an increase; however, we were always below as far as the filtered total mercury after the coagulation wetland system,” she pointed out.

Kraus Mercury_Page_25With the filtered mercury again, there was a little less, 40-70% removal of the methylmercury, she said.  “We had a lot of methylmercury formed in the control wetland, but less formed in the iron and the aluminum treatments.  But still, there was some methylmercury formed in the iron and aluminum treatments, and the iron seemed to be the best at reducing concentrations of mercury exported from these wetland cells.”

Kraus Mercury_Page_26So this is all the data together, just with the total mercury,” Dr. Kraus said.  “Again, the coagulation wetland systems generally lowered both total and methylmercury concentrations relative to the untreated water, but the untreated wetlands did contribute methylmercury during the summer months.”

Kraus Mercury_Page_27So what about the fish?  A truckload of mosquito fish were put into the wetlands, Dr. Kraus said.  “There was a large increase of methylmercury in the control wetlands, and less in the aluminum and even less in the iron,” she said.  “What was found in the fish was that there was no significant difference in the biotic uptake of mercury in the control and the aluminum, but we did see a 35% reduction in the fish living in the iron treated wetland.”

We talk about loads, and we talk about concentrations, now we have to talk about biotic uptake because we would have predicted, based on the fact that there were lower methylmercury concentrations, and lower mercury concentrations in the treated wetlands, that there would be lower uptake in both the aluminum and the iron treated cells, but that’s not what we saw,” she pointed out.

Kraus Mercury_Page_28Dr. Kraus pointed out that fish were collected from a lot of other locations on Twitchell Island at the same time.  “I think that’s really important, when we plot up what the fish are doing in our different wetland treatments, to put it in perspective,” she said, presenting a slide showing the bioaccumulation from the fish within the coagulant treatments relative to the fish pulled from other locations on Twitchell Island.

So this is again the data that we showed in the control, in the two coagulation treatments relative to the fish he pulled from the rice fields to the north, the main drainage canal, and the inflow and the outflow canals to the system we were studying.  Most of these were below the concentrations of concern,” she noted.

She then turned to the question of what happens to the material that is being accreted in the wetlands.   “There was very little material accreted in the controls over almost a two year period, whereas in the iron sulfate, sometimes we had six inches of material which was kind of like a moist chocolate cake texture, and quite a bit of material was deposited in the polyalunimum chloride wetlands.”

Dr. Kraus said there is currently a study underway to try to determine how long the mercury might be expected to stay in that material.  “Obviously if it’s going to come right back out, then maybe we don’t want to leave it in the wetland, but again, I think part of it will be what your objectives are – whether you are trying to reduce mercury export from a system and you don’t really care about what’s happening in your little wetland … if you are mostly just concerned with export, then what’s cycling in the sediment might not be as important, but if you think the mercury is going to be released under wet/dry oxidation conditions, that it will end up getting released back into the environment, so I think that is an important question and hopefully this study will shed some light on that.”

Kraus Mercury_Page_31So in summary, there are two approaches with coagulation, Dr. Kraus said.  “One is to think about it as just a treatment coagulation alone, what does that do,” she said.  “Maybe you would collect the material in a concrete basin and dispose of that offsite, or maybe you do want to collect that material in a natural environment, in which case you have to think about the cycling that happens in that wetland itself.”

Dr. Kraus noted that if you overdose with iron in particular, you can really lower the pH, but that was fairly easily controlled by monitoring pH.  “I definitely think that people should be aware that coagulants are going to have an impact on pH, she said.  “A lot of people were concerned about adding aluminum; I see no evidence of aluminum toxicity.  We did collect water samples for aluminum and the concentrations were very low in the water column.  Again we are interested in what adding iron sulfate and other kinds of compounds and the effect that removing DOC will have on methylation rates.  There is no evidence that we had higher methylation rates, but we have to look more closely at that.

There are concerns about the coagulants themselves possibly having some mercury contamination in there, which is a concern for regulators, and the cost and feasibility is an issue, she said.   Then the longer-term effects to see how stable this material is with respect to pulling mercury and DOC out of solution; there are also some concerns over the long-term about pulling too much phosphate out of a system, she said.

So how transferable is this to other sites?  Dr. Kraus said there are some project in the works to do some further coagulation tests up in the Cache Creek settling basin which are still in the formulation stages, but that is a very different system, she said.  “In that system, a lot of the mercury is in the particulate form, so they are looking more at trying to pull particles as well as the mercury associated with DOC out of solution.  The particles might be different and act differently to coagulants.”

You also have to think about different wetland properties,” she said.  “A lot of people think this is great, we can use this to reduce methylmercury in tidal wetlands, but that would probably not work too well, unless you want to flyover a bunch of coagulants on every tide.”

And that’s it …


Daily emailsSign up for daily email service and you’ll never miss a post!

Sign up for daily emails and get all the Notebook’s aggregated and original water news content delivered to your email box by 9AM. Breaking news alerts, too. Sign me up!

Print Friendly, PDF & Email
%d bloggers like this: