Visualisation and Modelling Tools for Evaluating Remediation Performance

Maureen: Well, thank you, Dane, and welcome from beautiful Ottawa, Ontario. And it really is a delight to have Grant Carey as our webinar speaker today. And he will be presenting on the topic of visualizing and modeling tools for evaluating remediation performance. Now, I’ve had the privilege of attending several talks that Grant has made, and I believe I was first introduced to the concept of back-diffusion and matrix diffusion, and the impact and significant impact that can have on remediation. And I also believe he was the first person that explained to me the difference between back-diffusion and matrix diffusion, and as a microbiologist, I do appreciate it and I do my best to keep it straight. I’m looking forward to this presentation that you’ve prepared for us today, and I think we’re all really trying to better understand the datasets that are generated from remediation sites, how to interpret results, and ultimately to develop an appropriate path forward. But, you know, before, you know, I turn this over to Grant, I want to briefly describe the reagents that we will be referring to in today’s presentation.

So, you know, we’re with REGENESIS, and for over 20 years, REGENESIS has been involved in research, development, and commercialization of environmental technologies. And we really specialize in developing in-situ solutions for treatment of wide range of contaminants in wide range of environmental challenges.

So the products that we’ll be focused on today and will be covered in a couple of the scenarios that Grant is going to present include our bioremediation product 3-D Microemulsion and our bioaugmentation culture BDI. The 3-D Microemulsion is an electron donor that’s applied as a dilute suspension, and it has some exceptional distribution characteristics. So it has a unique molecular structure that’s going to allow the distribution in the subsurface, and once that’s in place then it’s going to have a controlled release of organic acids for about two to three years, ultimately to stimulate reductive dechlorination. And the BDI is our bioaugmentation culture.

The other product that we are going to have as part of the presentation is PlumeStop. So PlumeStop, what that is is a solvent that’s actually designed to remove contaminants and promote degradation of groundwater contaminants. Now, the PlumeStop is composed of fine carbon particles or activated carbon that’s suspended in water through the use of a unique polymer distribution chemistry. So the PlumeStop is going to behave like a colloid behind to the aquifer matrix, rapidly sorb the contaminant, and then expedite the contaminant biodegradation. So in essence, we’re gonna paint the subsurface with a thin film of activated carbon. So in this animation, you’ll be able to see the difference between a powdered activated carbon and the liquid activated carbon with a PlumeStop, and how that typically can be distributed. The animation [inaudible 00:06:28]. We’ll get this animation thing going. But really, one of the most important features is for us to have the ability to deliver this liquid activated carbon under low pressure and get the widespread distribution that’s necessary.

So how is PlumeStop actually helping? So we’re often challenged with sites where we have to achieve a low cleanup standard, look to achieve very rapid results, and we also have to have results that are gonna be maintained for the long-term. So contaminants are removed rapidly through sorption, and sorption onto the activated carbon. But the PlumeStop is actually the platform that’s designed to support a wide range of biologies and chemistries so we can enhance biological degradation as well as abiotic reactions to enhance degradation.

Now, as part of this presentation, PlumeStop was used to address PFOA and PFOS. And under this scenario, this is a sorption-only mechanism as there isn’t really any degradation that’s been established or any biological remediation. So anyway, with this, I wanna get you back to Grant because he’s who you really wanna hear about. So, Grant, I look forward to your presentation and thank you.

Grant: Thanks, Maureen. Hello, everyone. It’s a pleasure to meet all of you. So as I’m sure a lot of you are well aware, during the site remediation process, we collect a lot of different types of data to help us design and monitor the remedy. So modeling and visualization tools can really help with remediation, especially of complex sites, including a number of the uses that I’m listing on this slide. For example, one thing that I find models particularly useful for is using them as interpretive tools that really help us to make more informed decisions about how best to remediate a site. And given the complexity of remedies today, visualization and modeling tools can also be very valuable in helping us to communicate or demonstrate key design aspects or performance monitoring results. And given the improved understanding of the role that back-diffusion can play at some sites, models can also be instrumental in helping us to evaluate more realistic remediation timeframes.

Today, we actually have a lot to cover. We’ve got two sets of tools, modeling tools and visualization tools we’re gonna run through with different case studies. So the first section is gonna be on modeling tools. I’m gonna start off just talking generally about some of the tools that are available, some of the different metrics you might wanna look at if you’re modeling this site, and then I’ll walk through a specific case study, as Maureen mentioned, where PlumeStop was used to remediate a source of PFOS and PFOA at a site in Central Canada. And then we’ll talk, in the second part of the seminar, about modified radial diagram visualization tool that can actually be really helpful for looking at both natural and enhanced biodegradation. It really helps to show you where biodegradation is occurring in groundwater. So at the end of this, I’m gonna show my email address. I am going through some of the slides fairly quickly. I’d be happy to provide you with a presentation after I’m done. Just feel free to send me an email if you like a copy of it.

So one tool I’m gonna talk about today, I use this actually more to learn and gain more insights about how remediation works, what things to look for with different technologies. It’s a model called In Situ Remediation, or ISR-MT3DMS is the long name. This is actually based in part on a BioRedox model I developed back in 1997-’98 with funding from CRA, which is now GHD. And the difference between BioRedox and the In Situ Remediation model, and say RT3D, for example, with reactive transport, is it really breaks the redox zones into discrete zones where only one electron acceptor is gonna be used at any time. And it allows you to actually simulate different types of reactions, whether it’s oxidation, reduction, mineral precipitation, and dissolution. Those reactions can actually vary depending on what redox zone a particular part of the plume is in. It does have also biodegradation rate stimulation and emission options.

So just as an example, Dr. Brian Looney of the Savannah River National Laboratory, I was working with him on an ITRC team back in around 2005-2006, the Enhanced Attenuation of Chlorinated Organics team, and he’d asked me as part of a separate project to actually use the model…I had a model set up for Plattsburgh Air Force Base at the time, and he was interested in mass balance. And that’s one thing we can actually use these modeling tools for, is better understand where does most of the degradation occur? Which is the redox zone that really drives where the degradation is gonna occur, and how much more degradation do you get, say, in the methanogenic zone versus sulfate-reducing stone or iron-reducing zone?

And just looking at the slide, I won’t go into the details, but you can see the plume here. This is a redox zone. And I use a particular way to visualize electron acceptors and metabolic by-products in these redox zone colors. It makes it easier to see where the different redox zones are. And in this site, we actually simulated the redox zones first with BTEX degradation driving the formation of these redox zones over about a 40-year time period, and then with chlorinated organics, we simulated different types of reactions depending on which redox zone each chemical was in. So with TCE, for example, we looked at reductive dechlorination with different rates depending on the redox zone. And then there was methane transport at this site, potentially into part of the aerobic zone, so we looked at methanotrophic cometabolism for DCE.

So the nice thing about modeling these kind of sites, it actually doesn’t take a lot of time. I really typically try to simplify as much as I can in the model and maybe spend a little more time looking at more sophisticated reactions so we can understand these processes, but it shows most of the mass of these chlorinated solvents was degraded in that methanogenic zone. And then a lot less mass in these other zones further downgradient in the plume. So that gives us more relative insights about where degradation is focused.

Other things in this model, there’s a simple NAPL Depletion Model that I use from time to time. And I use that as a screening model. It’s actually a free tool that’s available on our website if you wanna download it. And I use that to really get an estimate of the relative remediation timeframe for both natural attenuation of NAPL source zones and enhanced dissolution of NAPL source zones involving, for example, in-situ bioremediation or even strategic pump-and-treat, which may be useful at some sites for trying to get extra mass out of a source zone. So that tool is actually being developed and incorporated into the numerical model to provide a source term-type approach.

As Maureen mentioned, I’ve done a lot of work with back-diffusion modeling. There’s a paper we published a couple of years ago. I worked with Steve Chapman and Beth Parker at the University of Guelph and Rick McGregor, looking at just some special modeling tools for back-diffusion given it does take a long time to simulate these, to do it properly, there are some tools we can use to try and reduce our computer or simulation times. Whoops, I just went right to the end. Sorry, I’m just gonna scroll back up. I pressed ‘end’ instead of ‘page down’ So you’re getting a real preview here of this presentation. Lots of graphics as you can see.

So the other…one thing we’re also looking at is colloid transport. And that comes in actually with PlumeStop and emulsified oil. These are really little particles, very small particles, traveling through porous media. So colloid transport is something we can use to actually simulate the migration of, say, the liquid activated carbon with PlumeStop or emulsified oil.

Another thing I’m working on, I’m partnering with Deacon Larry Deschaine, who has a Ph.D. Larry’s with HGL and he’s developed an award-winning software program called Physics Based Management Optimization or PBMO. And that’s really an optimization, a true optimization software that he can link up with any groundwater flow and reactive transport model. So I’m working with Larry to link his Optimizer Code [SP] PBMO with the In Situ Remediation model so we can actually do true optimization of different remedial technologies. And really that’s to get a better understanding of the things we can do with what injection rates we use, what the solution composition is that may be actually speed up or give more effective performance. So that’s the kind of thing that we’re also looking at with remediation.

And one thing I’ve actually found interesting as well, especially if you’re looking at oxidants or reductants, the efficiency of those types of technologies really depends on the contact time in different parts of the source zone between the oxidant or reductant and the actual contaminants. So I’ve got a metric built into this model, and you can actually do it with MT3DMS as well, where you map the contact time over a 30-day, 60-day period so you can actually look for gaps in your source zone where maybe there’s not enough contact. So that’s something I’ve found actually works really well for trying to get better balance, better distribution of substrates that we’re injecting into the subsurface.

Okay, so that’s the introduction to the model. Let’s actually look at an actual case study of how we can use reactive transport models in a very simple way to actually learn more about what’s happening at some of these sites that we’re remediating. So this is a site in Central Canada. I’d like to acknowledge first Rick McGregor, who’s President of InSitu Remediation Services Limited. Rick’s a great guy. He’s in my opinion probably one of the best remediation contractors in the business, extremely experienced, very knowledgeable. Rick is actually the one who implemented this remedy at the site. I got the data from Rick to model after it had been implemented, to better understand some of the dynamics with PlumeStop. And I’ve also been working with Dr. Jeremy Birnstingl, who’s the Vice President of REGENESIS to look at modeling approaches for evaluating PlumeStop and PlumeStop longevity.

So I’m sure a lot of people on the call are familiar with PFOS compounds. I’m gonna go through just a really quick overview. These chemicals, we typically find PFOS compounds, especially two chemicals called PFOS and PFOA at sites where fire trainings have been implemented in the past or are currently, like airports, military sites. But these chemicals, these PFOS chemicals are also used extensively through a number of different industries for manufacturing because they’re engineered to be oil, water, and stain resistant. So these are or at least have been widely used chemicals that are widespread in the environment. And the thing about these chemicals, they are fairly toxic. Right now we have an EPA health advisory level of 70 nanogram per liter, part per trillion. And I have to get used to saying nanogram per liter because until recently, I haven’t said that word very often. Normally it’s microgram per liter or milligram per liter.

But the cleanup criteria are very low with these PFOS compounds, and that creates a big challenge for everybody because we have a lot of fairly large plumes right now that are orders of magnitude above cleanup criteria. And a real challenge with these chemicals, especially PFOS and PFOA is they’re very resistant to degradation, which means that in terms of in-situ remediation, we have pump-and-treat as one option to contain a source zone of a plume, or we have these carbon sorption-based remedies that we can…where we can eject carbon into the ground. And that’s what Rick did at this site is he injected PlumeStop to try to control this source.

So at the site, we have a shallow pretty thin silty sand aquifer. And initially, and still, this is actually a petroleum hydrocarbon site. That was the main focus when the site characterization was done. When Rick got involved after the characterization was completed, Rick’s job was to basically design and implement the remedy, he heard that there was a history of fire training and furniture manufacturing at the site, so he sampled for and actually found PFOA and PFOS in a limited source zone, limited plume at the site up to concentrations of 3,000 nanogram per liter PFOA, and up to about 1,500 nanogram per liter PFOS. So at this site, there are no regulatory limits at this site right now in Canada, where it’s located, so the client set up voluntary cleanup goals of 700 and 300 nanogram per liter with PFOA and PFOS.

So PlumeStop was injected into temporary direct push wells in one injection event in March 2016. Rick’s actually giving a talk, I think it’s December 7th or sometime in early December with REGENESIS. He’s gonna provide a lot more details on the actual remedy. So I’m just gonna really talk today about some of the results from monitoring the remedy and the performance that you might find useful. But the very interesting thing about this site, it’s the only…so far, the only in-situ remediation that’s actually been successful, that I’m aware of anyway, for PFOS and PFOA. After the injection of PlumeStop in March 2016, PFOA and PFOS have been non-detect at every well on the site, so less than 20 nanogram per liter in every monitoring event since the injection was done. So so far it’s been very successful. And one thing we wanted to look at with the modeling is what’s the long-term? Is it gonna be successful for the next couple of decades, next 100 years? And that’s what we’ll look at now with the modeling results.

This is actually an image Jeremy Birnstingl had sent this to me a while back. I find it very helpful because it really shows you what PlumeStop also called…which is also called liquid activated carbon, looks like in the subsurface. So these are very, very, small particles of activated carbon that have been injected. And because these particles are only in terms of diameter, 1 or 2 microns, very small, much smaller than the actual pore spaces that we see in sand zones, these small little particles of liquid activated carbon move very freely through sand lenses, through sand layers, so they get distributed nicely and they actually attach onto the sand grains. And that’s what you’re seeing in this image.

The other thing that’s key with liquid activated carbon or any activated carbon with PFOS compounds is sorption follows what’s called a Freundlich isotherm. And to make a long story short, the very ideal thing with PFOS compounds is that the lower the concentration with this type of adsorption [SP] mechanism, the higher the retardation coefficient is gonna be. And because PFOS compounds are actually typically present at pretty low concentrations relative to other chemicals, they actually have very high retardation coefficients in-situ when these liquid activated carbon particles are injected.

So the first thing I did with a reactive transport model is actually, try to calibrate what is the actual mass discharge coming from the source zone for both PFOS and PFOA that’s been sustaining these plumes that were observed at the site in this area? So what you’re seeing are the monitoring well concentrations. The blue symbols are the monitoring wells at the site. There are some further offset wells that show we have plume delineation to the south here. And we’re seeing on site we’ve got concentrations that are on the order of thousands up to 3,000 nanogram per liter for PFOA. And what I did is, I just calibrated a mass discharge rate of about 1.8 grams per year as coming from the source zone, and the source zone at this site follows this general area where it’s the darkest red shading. That was the source zone that we turned on in this model. And that reproduced the observed plume contours fairly well.

Now, when you’re modeling PlumeStop, if you wanna model the effect it has on say a plume detachment, downgradient the PlumeStop zone, if you’re gonna pick up a code like MT3DMS or any other program, there’s a step in here actually right now that’s not implemented and available reactive transport models. And that’s what I call the mass redistribution step. So before we inject liquid activated carbon, we’ve got mass in the sand layers stored in really two compartments on the left. We’re seeing the blue bars where we’ve got a fair bit of mass in the aqueous dissolve phase in groundwater, and we’ve also got mass sorbed to the native organic matter before the injection of the liquid activated carbon.

Now, when the carbon is injected into the soil, into the sand layers, now we’ve got a pretty quick re-equilibration, redistribution of the mass. So now we actually have three compartments in the sand layers where that mass is gonna be stored. There’s the dissolved and organic matter sorbed compartments but now there’s a new third one where we’ve got most of the mass almost all like 99.999% of the mass initially is actually sorbed to the liquid activated carbon in the sand layer. The amount of the distribution between the liquid activated carbon sorbed mass and these other compartments really depends on the sorption properties, the sorption isotherms. And I’m using sorption isotherm parameters that have been measured by REGENESIS for PFOS and PFOA.

So there is a simple equation that I developed just based on a mass balance approach. So CR is really the redistributed or the groundwater concentration after the injection of PlumeStop. So in each model grid cell, the In Situ Remediation Model recalculates the groundwater concentration based on this formula. And that’s a new starting point basically once the PlumeStop has been injected. So with the model, I basically tell it, “Okay, where’s the PlumeStop zone gonna be?” And the light blue color here represents the part of the PlumeStop zone at this site after Rick’s injection event that’s relevant to the PFOA and PFOS source zone. There is another PlumeStop zone to the north here, but this is actually, as I mentioned before, a hydrocarbon site. It’s not relevant to the discussion today, so I’m just gonna focus on the PlumeStop area relevant to the PFOA and PFOS.

And we see 10 days in, we’ve got really non-detect concentrations for PFOA in the PlumeStop zone, and we see the plume has actually detached here. And if I just move this forward quickly one month, two months, you can see the plume continues to move downgradient, but the source zone has now been controlled, even though there’s actually an after flux of mass continuing to come in in this area, which might be from back-diffusion, it could be rate-limited desorption from the sand particles, it might be infiltration from the overlying vadose zone although, it’s fairly thin here, or it could even be maybe there was a little bit of NAPL where PFOS had become entrained. But in this side, it doesn’t sound like that’s the case.

So three months in, this is the first event where Rick had actually gone out and looked, measured for PFOA and PFOS, PFOS and PFOA, and you can see three months in after the injection of PlumeStop, all the wells are now non-detect, including the ones that had been up to about 3,000 nanogram per liter. So they’re all less than 20 now. And six months in same thing, another event. No detection of PFOS or PFOA. There was another event done earlier this year. I’m not gonna show it here on the model, but same thing, still non-detect this year. And Rick’s gonna talk in December about some additional monitoring he’s been doing to characterize how things are distributed. He’ll actually show probably photographs of how PlumeStop was visible in the cores after the injection event, things like that.

So we know from the observed data and even from the model data that this PlumeStop worked great for the first year and a half, now the question is, okay, we’ve got mass sorbing but because in the model I think there probably is still an active source, there’s mass. If it’s coming from back-diffusion it’s still gonna come into the sand layer in that source zone and it’s gonna sorb to the carbon, but how does things change over time? So I wanted to model, okay, with an active flux coming into the source zone, more masses sorb in onto the carbon particles, what’s this gonna look like in 20 years, 50 years, 100 years?

So the first thing I did is I actually set up a very simple spreadsheet model. And I’ll show you the equations in a minute, you can program this yourself. And it’s the conservative model because it’s looking at what are the groundwater concentrations right in the middle of the source zone where it’s gonna be the highest concentration in groundwater and how do those change over time? So it’s called…I say it’s a conservative model because in the spreadsheet approach I’m looking at one model grid cell, I’m looking at mass flux coming in at the calibrated rates from the back-diffusion or desorption sources or infiltration, but I’m not looking at mass leaving that grid cell. I’ve got no outgoing mass you do advection or dispersion, and that’s why this is conservative. The numerical model does incorporate those basically mass out terms, but the spreadsheet model just assumes all the mass goes in and gets distributed between those three compartments we talked about earlier.

So one thing I had to do was actually look at okay, over 100 years, we know what the mass discharge is today, but over 100 years I expect that mass discharge to come down over time. So I did some simple analytical back-diffusion modeling to try and look at how the back-diffusion mass discharge rate might decline and get a conservative estimate for that. So based on that simple modeling, I’m using a mass discharge decline half-life of 30 years, and then this is the spreadsheet model for those of you who are really keen and wanna program it in Excel, you can. And I’m happy to send this PowerPoint to you. I’ll talk about how the numerical model compare to this.

So this chart shows the PFOA concentrations in the middle of the source zone over time. So time is the bottom axis. Time zero is just at the injection of PlumeStop after that mass redistribution has taken place, and the Y-axis is our groundwater concentration. So you can see PFOA sorbs extremely strongly to the liquid activated carbon after the carbon has been injected. We’ve got very low concentrations. They do come up. And this is the same thing if we’re using granular activated carbon for above-ground wastewater treatment or water treatment. We do see concentrations tick up over time, but, you know, even after 100 years, what we’re seeing, because the sorbing here is so strong for PFOA, concentrations come up to about 10 to the -6 nanogram per liter and then they level off for PFOA. And then just looking at retardation, because initially, the groundwater concentration is extremely low, the retardation coefficient is extremely high. Even after 100 years, we have a very high retardation coefficient because of the strong sorption properties of PFOA with the PlumeStop.

Now, PFOS is different. PFOS also sorbs very strongly to the liquid activated carbon but not as strongly as the PFOA. And this is based on the REGENESIS work where they’ve actually measured the sorption isotherms. So it’s interesting to see that PFOS does start after PlumeStop injection, it drops from 1,500 nanogram per liter down to something that’s way below detection limits, but it also comes up over time because of this ongoing mass flux into the source zone area. So we do have concentrations in the middle of the source zone that actually get up to about in the model 24 nanogram per liter. After 100 years, we’ve still got a retardation coefficient in the source zone of about 100,000. And just to let you know, the spreadsheet model and the numerical model are almost identical. And that’s probably because in the middle of the source zone you’ve got very, very small concentration gradients, so there probably isn’t much mass actually leaving the…there’s not much net mass leaving the source zone from that grid cell.

But when you look at it in plain view you actually see a very different picture. So if you only use a spreadsheet model, again, you’re looking in the middle of the source zone. Now, at this site, because there was a larger petroleum hydrocarbon source, PlumeStop is injected over a zone that extends to about 30 feet downgradient. The groundwater flow is from right to left here. So the PlumeStop zone extends about 30 feet downgradient of the PFOS and PFOA source zone. So even though in the source zone we see PFOS getting up to about 24 nanogram per liter in the model, after 100 years, at the end of the PlumeStop zone we’re still at 0.0004 nanogram per liter. So what’s happening is, the source zone, there is some mass showing up but it’s not…it hasn’t broken through the PlumeStop zone because there’s a pretty small flux in terms of 1or 2 grams per year of PFOS and PFOA. So that’s good to know.

So even for PFOS, which does have maybe detectable level concentrations in the source zone, at least the modeling is telling us that for many decades, we don’t expect that breakthrough beyond the end of the PlumeStop zone. So that’s good news for this site, one injection event. This little area probably cost…Rick said about $40,000 to do the remedy, for this one-time injection. There were probably, in this small area, maybe 20 to 30 injection wells, low injection rates. So pretty small cost for a passive remedy that looks like it’s gonna last for a long time into the future.

So the next thing I was curious about, well, I know that there are other sites where PFOS and PFOA are 10 to 100 times higher in concentration than there were at this site. So I wanted to see what happens with PlumeStop if we have higher mass fluxes into the source zone. How long is it gonna last for in those cases? So what I did is I took that same model, just created a hypothetical scenario, not representing the site at all and just said, “Well, let’s look at maybe another site that has mass discharge that’s 100 times higher and PFOS concentrations up to 260 microgram per liter, PFOA 440 microgram per liter in the source zone before the PlumeStop injection.” And then in the model, I said, “Okay, let’s put PlumeStop in around the same time, March 2016, what’s it gonna look like 100 years from now?”

Well, PFOA, because it sorbs so strongly, still non-detect everywhere on the site. PFOS, with the single PlumeStop injection, was actually…that one injection, even with the higher mass flux in the PFOS source zone with this hypothetical scenario was actually sufficient to stop breakthrough for 75 to 80 years. And at that point then we did start to see concentrations downgradient, the PlumeStop zone started to increase. So that’s still actually good news. We’ve got a passive remedy, low-cost that works still for decades and would be much cheaper than pump-and-treat if we needed to do anything at this hypothetical site.

So I’ll show you the modeling results. I saw a couple of things that were interesting here, so I just wanted to share this with you. So what we’re looking at first is what the PFOS contours look like, the plume contours before the PlumeStop injection. So here’s our source zone with this mass flux rate. It’s a mass discharge of about 100 grams per year in this hypothetical scenario. Concentrations of PFOS up to 260 part per billion. They declined a little bit but still pretty high downgradient of our source zone. And with organic matter sorption here using the linear isotherm, we’ve got a retardation coefficient of about 10 with the fraction of organic carbon that we have at this site. This is before the PlumeStop injection.

So after PlumeStop injection, this is a time of 1 year after injection, 7 years, 18 years afterwards, we clearly see in the PlumeStop area, we’ve got low concentrations. We do see up to…after one year we do see a source…because it’s a high mass flux we do see in the source zone concentrations getting up to about I’d say 0.1 to 1 microgram per liter. By about 20 years, we’re higher. We’re now at about 30 microgram per liter in the source zone, but 30 feet downgradient of that source zone, in the PlumeStop area, we’re still at 0.006 microgram per liter, which is a reduction of about 4.5 orders of magnitude from the initial concentrations before the injection.

And then if we just go and look into the future again, nice to have models to be able to look into the future, we still see, at the downgradient edge of the PlumeStop zone, we still see low concentrations. But you can see because we have the high mass discharge at100 grams per year, we do see the advancing plume within the PlumeStop area. So eventually, at about 75 to 80 years, we do see a little bit of breakthrough, and then if I go on in time, we do see more concentrations coming down. So there’s a couple of options here in terms of how we could manage this site knowing that we still have passive treatment here with PlumeStop injection that will work for many decades.

So we know what would work for at least 75 years as a passive remedy even at a very high mass discharge, at concentrations of hundreds of microgram per liter PFOS and PFOA. It’s certainly cheaper than pump-and-treat, which right now at least anyway is one of the only other options we have for these PFOS and PFOA compounds. I know there are others working very hard on. I’m trying to develop more destructive technologies, but at this point in time, we’re just not at the full-scale implementation yet. So we do have the option knowing that this is gonna last for about 75 to 80 years. We could have actually put in a longer PlumeStop zone initially, maybe put in a 60-foot buffer zone downgradient of the source initially, and that’s…we’re gonna get twice as long a performance in that case. Or in 75 to 80 years or whenever modern [inaudible 00:35:45] we need to, we can re-inject PlumeStop in the future, either in the same source zone or downgradient, or maybe in the future, there will actually be a technology available for in-situ destruction as another option.

So the big lesson here for me was that PlumeStop certainly is long-lasting. Even with fairly high concentrations, high mass fluxes of PFOS and PFOA, and the modeling just can be used as a fairly simple tool to look at longevity and give us some thoughts on what the long-term scenarios might be.

Now, PFFOS compounds, there are hundreds of PFOS compounds. There are some great experts out there doing research on identifying what these compounds are. REGENESIS has done some work to look at PlumeStop performance with other compounds. We know that with granular activated carbon, shorter chain compounds may sorb less strongly to activated carbon. REGENESIS has found at least with some preliminary work that we do get good sorption even for some of the other shorter chain compounds, less effective than PFOS and PFOA but still actually very effective. In this experiment I’m showing you, they had 99.8% reduction in groundwater concentrations even for the shorter chain PFOS compounds in this case.

One thing I’m also doing is I’m working with a group. Basically, we call it the PFOS Remediation and Research Group. It’s got academic partners, including Dr. Brent Sleep at the University of Toronto. Dr. Anh Pham who’s at Carleton University Civil and Environmental Engineering, and we’ve got industrial partners, at least the initial partners. We’re hoping to expand the list. The industrial partners are putting in cash and in-kind services. The academic partners are actually doing most of the work. And what we’re looking at, we’re starting on a couple of things, one is we’re looking at PlumeStop with PFOS remediation, and we’re looking at a number of different scenarios so we can better understand how PlumeStop works for a wide range of different sites and different conditions. And we’re also looking at freight and transport models, taking the in-situ remediation model and developing that a little bit further so it could be used as a more specialized tool for PFOS freight and transport.

Okay, so we’ve got about 10, 15 minutes left. What I’m gonna do next is just talk about the next tool, which is a completely different tool. So we’re switching tracks here. We’re getting out of the modeling world and now we’re actually just gonna be using observed data and looking at different ways of presenting it, that help us better understand and better communicate what the data is telling us.

So I’m not gonna have a lot of time for background here, but very quickly, when biodegradation is occurring, I’m sure a lot of you know that we need electron donors. These are chemicals that give up the carbon that microbes use for cell growth. We also need electron acceptors. Those electron acceptors are basically transformed resulting in the release of energy that’s used by the microbes for cell metabolism. And typically, the key electron acceptors that we measure in groundwater to see if biodegradation is occurring are dissolved oxygen, nitrate, and sulfate. We also look for what we call metabolic byproducts, which are produced during biodegradation. That’s dissolved manganese, dissolved iron, and methane. So those are six redox indicators that we typically look at or measure in groundwater to evaluate natural attenuation, see if degradation is occurring, and especially to look at how enhanced biodegradation is working as well.

So you might be asking why do we look at all those chemicals. Well, one thing I look at, if I’m looking at natural attenuation or MNA as a remedy, is it viable? Or if I’m looking at enhanced biodegradation, one of the key things I try and understand is, where are the major redox zones in groundwater, before and after enhanced remediation, for example? And the reason we do that is because…I’m just gonna use TCE as an example here, trichloroethene is chlorinated solvent. This is data from a really groundbreaking paper, Suarez and Rifai, in 1999. They published a paper that showed how the degradation rates vary for different contaminants in different redox zones. And under strongly anaerobic conditions, where sulfate is being reduced or methane is being produced, TCA degrades most quickly. If it’s iron-reducing conditions, which are moderately anaerobic, TCE still degrades, but it’s a slower process. And you can see under nitrate-reducing conditions, aerobic conditions TCE is probably not gonna degrade unless you have something like methane present where TCE degrades very rapidly through a process called methanotrophic cometabolism. So being able to…if we can look at groundwater data and actually define where these different anaerobic and aerobic zones are, we can see where TCE is gonna biodegrade, which data products might be formed, and same thing with other chemicals.

So the visualization tool I’m gonna talk about, it’s a radial diagram tool, and the way it works is you plot radial diagrams at each moderating well location. And we’re looking at a site, this is actually a case study I did back in 1996. I was actually working with Conestoga-Rovers & Associates at the time, now GHD, and each radial diagram represents redox indicator concentrations for one monitoring well. And I’m not even gonna tell you which chemicals are shown on the different axes, I’m not even gonna show you the concentrations, because the nice thing about the way this modified radial diagram method works is you can tell just by looking at the size of those orange shapes, which represent redox concentrations at each monitoring well, the size of those shapes tell me whether we have an aerobic aquifer, like we see in the background part of this. This is a background well, and it’s a big orange shape, and that’s telling me, without even looking at the data, I have aerobic conditions there.

And I can see downgradient of a former drug disposal area, we have really smaller in shapes, and I know that those represent strongly anaerobic conditions where we’re probably getting a lot of biodegradation occurring along that flow path. And even downgradient of the municipal landfill, these are moderate small…they’re medium-sized shapes, so I know that we have slightly to moderately anaerobic conditions even downgradient of the municipal landfill telling me that there are some organic substrates from the landfill getting down into the underlying aquifer. So it’s a very simple way to look at the data without actually getting into the details of each individual chemical.

So back in 1999, I did develop a radial diagram tool. It was much better than what I have today. It was the Windows-driven program, data management system built in. That was funded by CRA, now GHD. They were very helpful with both the initial BioRedox tool and the SEQUENCE tool. I’d say true innovators to be able to recognize the value of these tools. Today that software, it’s out of date, it’s not updated. So at least to be able to make people…to have a tool available, I’ve developed a simple Visual Bio tool. I just made it available this year, very simple to use. It takes a little bit of time to get to learn, but at least you’ve got something you can use for this. So I’ll show you two ways to look at radial diagrams with all the data we collect. One is with…I call it CAH, chlorinated…it’s basically a chlorinated solvent radial diagram. That’s the first example. And then I’ll show you the redox diagrams. And what I’m gonna show you now is actually specific to a case study. This is a site in Michigan. Doug Davis from REGENESIS actually provided the data for this site. It was very helpful to get me started with being able to visualize the data. So this is a case study now that we’re gonna walk through.

The first diagram were looking at, it shows data for these chlorinated solvents. We see trichloroethane, TCE, on this west axis, if you wanna be simple, and that’s the parent species. And we know that TCE when it degrades, will go through the cis-1,2-dichloroethene, and if that degrades that may go to vinyl chloride. That may actually go to harmless end products through oxidation. And if vinyl chloride degrades, reductively, that goes to ethene. It can also degrade through oxidation where ethene is not produced. So here we see at this one well, we’re looking at concentrations for the four chemicals and we see a high parent concentration. That’s a high cis-DCE, but it’s showing that we’re getting degradation of TCE to cis-DCE. But we’re not getting vinyl chloride or ethene showing. We’re not getting a whole lot of…we’re probably not getting much complete transformation here at this well. And I’m saying that based on other data as well.

Another thing I do with these radial diagrams is I plot symbols to show where we have exceedances. At this site they’re not using MCLs, they’re using risk-based screening levels. So the symbols show where we have exceedances of groundwater criteria. And then another symbol to show where we have non-detect. In this case, ethene is non-detect. So now I just go and add another data series on the radial diagram for 2016 data. And you can see a dramatic reduction in all chlorinated solvent concentrations. Three and a half order of magnitude reduction for TCE from 2010 to 2016. And these are long scale, so you can easily read off the number of orders of magnitude and reduction. DCE is 2 order of magnitude reduction, vinyl chloride, a little bit of reduction, ethene still not detected today, which is good. It’s gone all the way through to ethene.

So now the beauty is, okay, that’s one well, let’s put them on a map. And you can actually see in 2010, we’re only looking at 2010 data now, the source area is below this building. Groundwater is flowing to the top of the page. So we can see fairly high TCE and cis-DCE concentrations all the way downgradient. So that’s how I know DCE is probably not being transformed through oxidation. That we’re not getting complete transformation here. Now, there was a very aggressive remediation and a very successful remediation through the different transects. Each line here is a transect of injection wells. So there were some transects. A lot of transects actually were emulsified oil. 3DMe was injected in 2013 and 2011. And then also the grey transects here, the lines represent injection wells where PlumeStop HRC and BDI that bacterial inoculum was injected through bioaugmentation. So a lot of work done between 2011 and 2015.

And when you look at the comparison, the orange data series represents 2010 data. And just looking at the inset map here, we’re talking about orders of magnitude, reductions, and concentrations at all of these wells. And the non-detect symbols are helpful because you can show that these chlorinated solvents are pretty much non-detect everywhere except for a few exceptions. Maybe not the best word in there but anyway, you see one small exceedance of cis-DCE, but certainly low concentrations. And it’s as you move downgradient cis-DCE is just not detected. So very successful remediation here. And this gives you a way not just to look at TVOC, which is another way to visualize the data before and after map, but now we’re actually looking at four chemicals and two different time periods on one map. So if we were doing this the normal way through contour maps, I’d have eight contour maps spread out across the table where I’m trying to compare trends over time, trends between parent species, and daughter products. And that can actually be a little challenging, especially trying to communicate the results of that kind of analysis. But now we’ve got a tool that shows all the data on a single map.

So we’ve talked about the benefits. I’ll just move on now to another advantage of the radial diagram tools is actually being able to look at the shift between parent species and as they transform through the daughter species. So I’ll just show you…I’ll back up. There’s one well it’s SB-230, but it is actually a permanent well now. I’m gonna look at data over time at that well and show you…so about this well, SB-230 is 50 feet downgradient of emulsified oil injection well transect. So right at the time of injection, you see high… and these are a slightly different scale. This is 1 milligram per liter…sorry, 10 milligram per liter at the extent of each axis. So TCE is about 1 milligram per liter, cis-DCE is 4 milligram per liter, BCE…sorry, vinyl chloride is about 0.1. And you can see even 90 days after the injection, still really no change.

And when we look at the redox we’ll see also no change. And that’s because it is 50 feet downgradient of the injection transact. But after 190 days you see a rapid transformation. All of the TCE is now non-detect, has been degraded to cis-DCE, and even that has been reduced. And we have an increase of vinyl chloride and a much bigger increase in ethene. So now we know we’re getting complete transformation because conditions are more anaerobic now and we have a lot more substrate on the ground to drive the biodegradation process. So just over time, you can see really a nice shift with the radial diagram. So that’s another advantage to use radial diagrams with chlorinated solvents.

The real key advantage of the radial diagram tool, though, is how we can actually visualize the redox indicators. There’s a unique little twist, pretty simple but it’s not available in commercial radial diagram software, which makes it so easy to show where we have strong anaerobic conditions or moderately anaerobic conditions relative to background. And this is the way it works. So each axis represents in this case electron acceptors shown in blue. And when we have electron acceptors axes, they increase from near the middle of the radial diagram to the outer extent of the radial diagram. So all these increase towards the outside of the radial diagram and concentration. The key thing here is metabolic by-products that are actually produced during degradation have the opposite trend. So where we have electron acceptors used during biodegradation, they decrease when things go from aerobic to anaerobic. We see reductions in DO concentrations and nitrate concentrations. When metabolic by-products are being produced, we see increasing concentrations. It’s the opposite trend. So we have the opposite direction for the metabolic by-product axis.

The key thing here is that in a background aerobic aquifer where we’ve got high electron acceptors and low metabolic by-products, all the concentrations at this well if it was aerobic would plot at the outer extent of the axes. So this is what a data series would look like at a well in anaerobic aquifer. In a strongly anaerobic aquifer, try and think about what that green shape would look like. So if you guessed that, good on you, you’re absolutely right. Everything moves in towards the origin because these axes have different directions depending on whether it’s an electron acceptor or a metabolic by-product.

So just looking if we had wells along a flow path, where we start with aerobic conditions in the MW-1 well, if things are going to be progressively more anaerobic, this is what you would see with the redox indicator concentrations. So normally, we’d be looking at tables where I’d have 4 wells, 6 indicators, 24 pieces of data. I’m trying to look at each well and look at what oxygen is compared to nitrate, how does that compare to iron? But visually, you can look at this and immediately see what kind of redox conditions we have. So looking at the site data now, we can see that we have in 2011, before the remediation was implemented, we do have anaerobic conditions, moderately anaerobic, probably iron and nitrate-reducing but not much methanogenesis or sulfate reduction going on.

And just to add in, 2016 now is the red data series. You can see a big transformation-ware near the source zone, much smaller shapes, meaning more strong anaerobic conditions because you’ve got more substrate present. And even as you move downgradient you see much more methane, less sulfate. Here you actually see decreasing iron and manganese concentrations probably indicating that iron and manganese are now depleted because there’s been so much degradation in this area. So we can look at this kind of thing and just looking at it, this is again, we looked at this well before, 50 feet downgradient of the emulsified oil injection transect, and we can see redox conditions. When they started to change more dramatically, that’s when we started to see the shift in the parent species.

There is a way, I’m running out of time so I won’t go into it in detail, but there is a way to look at the size of the green data series relative to the size of the outside data series, which is anaerobic reference data series. I just calculate a ratio based on the areas of those two data series, and I can actually calculate, where is the transition between say sulfate reduction and methanogenic as one zone? Where’s the transition in that relative redox area between that and manganese and iron reduction? So using that kind of simple ratio, semi-quantitative analysis, you can create a curve like this that shows how we had started with iron-reducing conditions and things got to sulfate reduction and methanogenesis and stayed that way for at least five years after the injection of the emulsified oil. So here we’ve got a long-lasting substrate with the emulsified oil and clearly sustained strong anaerobic conditions for a long time.

So that’s basically the conclusion here. I just wanna let you know there is this free tool available. I apologize, it’s pretty simple to use. It’s a simple tool. It’s a DOS-based tool. You really need Golden Surfer to actually do the graphics, to make the maps, to make the radial diagram maps, but you enter the data in DOS text files, you run the software and it creates files that you can import into Surfer quite easily to create these radial diagram maps. So if you just go to our website, shown up here, click on the Visual Bio, you can download a zip file with the user’s guide software, and an example for the Plattsburgh Air Force Base.

So on that note, I’d like to conclude the presentation and thank you all very much for your attention. I did wanna point out my email address is here. If you’d like a copy of the presentation, just send me an email and I’d be happy to send it to you. Thanks very much.

Dane: All right. Thank you, Grant. That does conclude the formal section of our presentation. And at this point, we’d like to shift into the question-and-answer portion of the webcast. Before we do this, just a couple of quick reminders. First, you’ll receive a follow-up email with a brief survey. We really appreciate your feedback, so please take a minute to let us know how we did. Also, after the webinar, you’ll receive a link to the recording as soon as it is available. All right, so let’s circle back to the questions. We do have a lot of questions today, so if we run out of time before we get to your question, we will make an effort to follow up with you after the webinar. All right, so here we go. The first question is how are co-contaminants accounted for in these models? And this person is talking specifically about hydrocarbons and other PFOS. And what effect would you expect them to have on the assumed life of PlumeStop remedies?

Grant: That’s an excellent question. So there’s a couple things I wanted…I’d like to say about that because I was expecting that question. I’ve had the same. First, there’s actually some really good papers out there that look at mixtures of PFOS compounds even with dissolved organic matter that’s come with groundwater samples collected where they’ve done sorption isotherms, not for liquid…well, actually, REGENESIS is doing testing with PFOS compounds and other mixtures today. There are some sites where you really might have a complex mixture and you might need to actually have someone like REGENESIS, and they offer this as a service, do sorption isotherms or testing to see how the liquid activated carbon performs with a site-specific mixture. So if you have fairly low concentrations of co-contaminants, the nice thing about liquid activated carbon, because they’re small particles, they have a very high surface area relative to, say, granular activated carbon. So because they have so many sorption sites, if you have low concentrations of all the co-contaminants, you’re probably not gonna see much effect.

And I’ve seen a paper, there’s a good paper by Hansen et al., they looked at lower concentration PFOS mixtures with DOC, dissolved organic carbon at the 1 to 3 milligram per liter range, and they actually, with particle activated carbon, which probably has low sorption potential than liquid activated carbon because particle size makes a difference, they saw similar types of sorption properties. So lower concentrations, I wouldn’t be too concerned about it. If you do have high concentrations and a mixture a fairly complex mixture then I suggest doing site-specific testing. And that’s similar to doing batch testing. Like, if you’re gonna do microcosms in the lab or other bench scale-type testing, I think with PlumeStop and complicated mixtures, you might wanna do site-specific sorption isotherms to really make sure you’ve got the right parameters when you’re modeling this.

Dane: Okay. All right, great, thank you. Next question here is please discuss the model parameter assumptions used to determine the retardation coefficients. Also, discuss the process used to calculate the expected desorption rates.

Grant: Okay. So first of all, the sorption and desorption, whether it’s Freundlich or Langmuir or linear isotherms, in this model, that’s an equilibrium process. So when you have a Freundlich isotherm, it’s more complicated than the linear isotherm. Linear isotherm…I don’t have the equation on a slide, in hindsight I should have had that, but it’s basically the sorb concentration is gonna be your KOC times your FOC times your water concentration, and your retardation coefficient would not change. Retardation is just 1 plus density divided by porosity times KD [SP], very simple. It ever changes.

But when it’s a Freundlich isotherm, it does change. There is an equation that we use. I’m not gonna verbally say it. I don’t think I have a slide that shows it, but it’s a widely used…basically, the retardation coefficient it’s equal to the slope of the sorption isotherm. And that’s really how it’s determined, or it’s related to this slope. So if there’s a simple equation, parameters would include the KF, which is your…really your sorption capacity. When you’re at unity concentration, there’s an exponent term in there, and there’s the water concentration. So as water concentration changes, with a Freundlich isotherm, your retardation coefficient changes as well. So in the reactive transport model, initially before a PlumeStop, there’s a switch that says, “There’s no Freundlich sorption it’s just linear isotherms, linear retardation coefficients.”

In the model, when PlumeStop is turned on, the model knows, “Okay, I’m in a grid cell where PlumeStop is present.” It’s got a certain…one of the parameters in the retardation coefficient as well is the fraction of liquid activated carbon per mass of soil similar to FOC. So the model will basically calculate the retardation coefficient based on the water concentration at each individual time step and each individual grid cell, and it’ll do the rest of the calculations accordingly.

Dane: Okay, all right. Great, Thanks. Let’s see here, next question is how was the extent of the PlumeStop confirmed in the case study.

Grant: Oh, good question. So Rick’s gonna talk more about this, but from talking to Rick, he used 10-foot spacing, 3-meter spacing for the injection wells. And what he found to confirm it, because I asked the same question, so that is a good one, he actually saw monitoring wells that were more than 10 feet away from the injection wells. They were actually measuring looking for PlumeStop. The nice thing about PlumeStop is when you first inject it is a black color liquid. It’s just basically aqueous water with black carbon in it, but you can clearly see when you have PlumeStop in a monitoring well at least after the initial injection. Over time, those particles will attach to the soil and you won’t see it in groundwater over weeks or maybe a month or two. I’ll let Rick talk more about that. But he clearly showed that there were monitoring wells more than 10 feet away from the injection wells breakthrough of PlumeStop, so he knows that he’s at least got the equivalent of a 3-meter radius of influence. So that’s what I use.

So basically I developed a PlumeStop zone, I traced where he had the injection wells. And these were very…they were offset a little bit from each other in terms of transects, but he had a lot of wells and I just traced around them with 1.5-meter radius outside of each injection well.

Maureen: Hi, Grant, just to add onto this. My understanding is they also collected soil borings as well to confirm. And so they were able to use, you know, both the visual from the groundwater, but there was confirmation where they looked at soil borings and could find the dark color. And so these are some of the standard methods that are used to make sure that we’re getting the distribution of the PlumeStop during an application.

Grant: That’s right. Thanks, Maureen. And Rick is also…they’re in the process of…they’ve sampled those cores [SP] and they’ve actually sent them to the lab to see if they can actually…they’re measuring the concentrations of the liquid activated carbon in the cores. But I’ll let Rick talk about that. It’s a fairly complex thing to do, apparently, so that’s a question for Rick in December.

Dane: Okay. All right, great. Well, thank you, Grant and Maureen. We are out of time, so that is gonna be the end of our chat questions. We did not get to all the questions, so if we didn’t get to your question, someone will be making an effort to follow up with you. If you would like more information about modeling and consulting services from Porewater Solutions, please visit porewater.com. And if you need immediate assistance with a remediation solution from REGENESIS, please visit regenesis.com to find your local technical representative, and they will be happy to speak with you. Thanks again to Grant Carey and to some Maureen Dooley, and thanks to everyone who could join us. Have a great day.

Grant: Thanks, everyone.