Combating Our Global Water Crisis Using AI

Published on by in Academic

Combating Our Global Water Crisis Using AI

Combating Our Global Water Crisis Using AI, with Junhong Chen (Ep. 78)

UChicago-Argonne scientist explores more sustainable ways to make use of water

There are a lot of problems in our world today, but if our water systems aren’t working, everything else takes a backseat. From a lack of freshwater to droughts on the West Coast to contaminants like PFAS and lead in many of our homes, our water systems are in trouble. But one scientist sees a solution to our making our water system sustainable by using artificial intelligence and machine learning.

Junhong Chen is a professor of molecular engineering at the University of Chicago and the lead water strategist at UChicago-affiliated Argonne National Laboratory. He’s using AI to address many of our global water crises in some surprising ways.

Transcript:

Paul Rand: Water. It’s all around us, but we often take it for granted. We don’t think much about it until suddenly, it’s all we can think about.

Tape: You may remember the pictures from the water crisis six years ago in Flint, Michigan. Hundreds of angry residents, holding up bottles of rust colored water and demanding answers.

Tape: New investigation by the USA Today Network is raising questions this morning about tap water nationwide. The report identified nearly 2000 water systems used by six million people where excessive levels of lead were detected in the past four years.

Tape: The drought that has hit the western United States for years is getting worse. Last week, the federal government declared what’s called an exceptional drought for much of the region.

Paul Rand: There are plenty of problems in our world, but if are water systems aren’t working. Everything else takes a back seat.

Junhong Chen: Yes, absolutely. Water is needed everywhere beyond just living our life.

Paul Rand: This is Junhong Chen. He’s a professor of molecular engineering at the Pritzker School of Molecular Engineering at the University of Chicago and has one of the most interesting job titles I’ve heard. Lead water strategist at Argonne National Laboratory

Junhong Chen: Without water. It cannot drive our economy and national security as well because if there’s no economy, no resource, no energy, our national security is endangered.

Paul Rand: Chen is working on the cutting edge of water system engineering. Trying to address some of the fast approaching water problems in the world, but he says “We’re not moving fast enough.” 

Junhong Chen: Nationally, compared to energy or climate change for example, we don’t have a department of water that drives all the water activities and that flows. So I think that there are several reasons that’s making this problem the less obvious and less urgent. But in fact. It’s not true.

Paul Rand: Luckily, Chen is on the front lines trying to stave off a water catastrophe, and he thinks the biggest solutions will actually come from AI and machine learning.

Junhong Chen: This is more like a powerful tool for us to enable some of the things that are impossible currently, right? So, that’s actually the most exciting role that AI could potentially play.

Paul Rand: From the University of Chicago podcast network. This is Big Brains, a podcast about the pioneering research and pivotal breakthroughs that are reshaping our world. On this episode, how artificial intelligence can change water forever? I’m your host, Paul Rand. You are overwhelmingly focused on water. Did that come to you one day? You’re taking a shower and you’re thinking this water stuff’s kind of what... Kind of interesting.

Junhong Chen: Yeah, that’s actually a very good question. So, obviously water is a eternal topic of interest because water is essential to our life. Without water. We can not live more than three or four days, right? According to the United nations, access to water is a basic human right.

Paul Rand: Okay. Well, I guess this question is what the average person thinks about water and maybe that’s changing as we have shortages. But, are we as concerned about water as we should be?

Junhong Chen: Yeah. That’s kind a interesting though, depending on who you are asked, this question. So most of us would take it for granted, water flows automatically from our tap and for us, for living close to the great lakes region. Obviously we even thinking water is plenty full, right? It’s everywhere.

Paul Rand: And fresh water.

Junhong Chen: Yeah. Fresh water, especially. So how nice, but for people who are living in California and Colorado, Arizona, they’re appreciating the water valley because they don’t have enough fresh water there. Actually, according to the national public radio, over the next 10 years, 40 out of 50 United States will experience water scarcity.

Paul Rand: My goodness. Okay.

Junhong Chen: Yeah. That’s a major issue.

Paul Rand: Indeed, but water scarcity doesn’t just mean a lack of drinking water.

Junhong Chen: Right? So what is needed to generate energy, to power our house, our offices, our universities, and our schools, hospitals, etc.

Paul Rand: When water fails, everything fails.

Junhong Chen: And we also need water to manufacture almost everything from our food to clothes to cars and electronics. I would just give you two examples, right? So to produce a handbook, it takes 24 liters of water, which is about 600 gallons of water. We call it virtual water footprint, right? To manufacturer a lithium-ion battery pack in an electrical vehicle that takes 7,400 gallons of water, huge amount of water.

Paul Rand: The way Chen sees it, our world faces a number of challenges with water.

Junhong Chen: First of all, there’s a limited supply of the renewable freshwater, right?

Paul Rand: Right.

Junhong Chen: Especially weighing in where it is needed. That really limits our economic development. And, but frankly we see future or population and economic grows. And also we talk of climate change and over migration, all of these factors were fully exacerbated. The growing water stress, globally and nationally here as well.

Paul Rand: Okay, what’s problem number two?

Junhong Chen: The second major issue is for our nation is the drinking water quality issue. We have seen the Flint water crisis, right? So that’s my example and many others. And that also brings about environmental justice. Underserved communities typically have more water quality violations, based on the U.S. Environmental protection agency standards, right? So, we need better infrastructure and more accessible treatment technologies.

Paul Rand: Our investment in water infrastructure has not kept up with the growing needs to replace aging systems. In it’s 2021 report, The American Society of Civil Engineers gave our drinking water infrastructure, a C minus grade. And if that wasn’t bad enough, our wastewater infrastructure got a D plus speaking of-

Junhong Chen: And finally it’s the energy consumption needed to treat this water. In the United States about three percent of the electricity load is consumed by municipal wastewater treatment. So we move all the organic matters and nutrients essential, right? So it is a very energy intensive process. With the current administrations, our goal to decarbonize our energy supply. We really need to do more to reduce energy consumption and reduce the emission of greenhouse gases related to the water treatment. So those are the major challenges.

Paul Rand: Chen is engaged on each of these problems and we’ll get to specifics soon, but there’s an overarching resource that Chen sees as crucial to address these issues, artificial intelligence and machine learning.

Junhong Chen: There are problems with rich data and complex interactions that are difficult for humans to fully comprehend, right? For water research, there are a number of areas that are falling into this potential category.

Paul Rand: One of those areas is contamination in our water supply.

Junhong Chen: So for example PFAS, poly fluoroalkyl substances is a very complicated contaminant with many of... Thousands of compounds

Speaker 6: PFAS, are toxic chemicals linked to serious diseases like oh, cancer, liver damage, and thyroid disease. And guess what, they never, ever degrade. In fact, they’re known as forever chemical.

Junhong Chen: They are widely used as firefighting foams, nonstick cookwares, in fast food wrappers, stained resistant copies and fabrics, cleaning products, even personal care products. So it’s everywhere.

Tape: They’re so prevalent, CDC scientists believe PFAS chemicals are in the bloodstreams of nearly all Americans.

Junhong Chen: In order to manage this type of compounds, it’s time to build a physics-based model to come up with a material that can effectively separate all these different types of PFAS, right? Because water contamination can occur between the water treatment plant and our kitchen.

Paul Rand: Because it’s coming through pipes.

Junhong Chen: Pipes. Yes. They are led in our pipelines and they are proxying our pipelines, bacteria can get in, right?

Paul Rand: Mm-hmm (affirmative).

Junhong Chen: So, all kinds of contamination events can occur. So how can we selectively separate the different contaminants from water? Let’s say, I want to take out PFAS. Can I do that selectively, because it’s not a trivial task, there are many contaminants co-existing with PFAS. So if you wanted to take out PFAS, most likely if others are present at a higher concentration, that’s posing a big problem. That’s where AI and machine learning can potentially come in to view different models based on what we have already learned about this chemicals and try to learn smartly, to develop new materials and sensors and devices, to be able to better manage this type of water contaminant.

Paul Rand: Chen is developing sensors that could work with AI and machine learning to model and detect contaminants in our drinking water, faster than any human ever could. It’s work he started specifically to target lead.

Junhong Chen: Currently to measure lead in the drinking water. You have to take water samples and send it to the analytical laboratories and it’s slow and it’s relatively expensive. It costs, somewhere between 30 to 50 dollars per sample. So average family will not be able to afford it. When you try to figure out the water quality urgently, you can not have the data until a few days later, right? So our vision was really... Can we develop sensors that can enable or empower the average person to measure water quality directly from their tap. Right?

Paul Rand: Right. So it could be... And the intention is that it’s a consumer product it sounds like.

Junhong Chen: Exactly. So while the benefit for this type of products is, people can measure their water quality. So they are ensured, their water quality is good. If it’s not good, they have then to find relevant measures to manage it. Right? So, this will help us restore the public trust of drinking water that has resulted from earlier water crisis, right? And also water utilities, they can use this type of low cost, portable technology to check our water quality rapidly. For example, after a severe storm they wanted to see if there are bacterial in the water, right? So, it’s going to be useful for them as well.

Paul Rand: Imagine, an affordable device, you could pick up at any hardware store to quickly and easily test for lead at home. No complicated phone calls with labs and city officials, no scheduling contractors or tearing up pipes.

Junhong Chen: The art is we are putting nano materials in action in these type of devices to make it highly sensitive to water contaminants. And they are also low cost because we can manufacture this chips and do this transistors at large scale, at a relatively low cost through my startup company called Nano Effects. We are trying to validate this prototype with our potential customers now, as we speak. This past couple of weeks, we just started it. So we’re hoping to be able to introduce this product into people’s hand, to be able to measure water quality, just like glucose testing at home.

Paul Rand: If you listen to our last episode with U Chicago President Paul Alivisatos. And I recommend you do, if you haven’t already. Something Chen said, probably perked up your ears, the sensors he’s designed use nano materials specifically nano crystals.

Junhong Chen: Yes. Obviously that’s very exciting. We are very happy to see President Alivisatos joining us because he is the true pioneer of nano crystals, right? For optoelectronic applications and for our applications, we’re using those nano crystals for sensors, for energy devices. For example, we use a lot of T-oxide nano crystals for detaching of toxic gases. We are using a lot of golden nano crystals to detect water contaminants and the biologic species, because gold is very stable, right? And also gold conjugation chemistry is well cited. That means you can attach different functioning that moves onto gold really reliably. That enables us to be able to detect different sense selectively.

Paul Rand: But when it comes to reducing contamination in waste, there’s an obvious problem here. Those electric sensors with themselves create waste and contamination in the supply.

Junhong Chen: All these sensors are part of the electronic devices. And while the challenge is with the electronic devices is the electronic waste. We’re generating a lot of this electronic waste.

Paul Rand: How big, of a problem is this?

Junhong Chen: It is a huge problem. And think about our personal electronic devices, right?

Paul Rand: Right. Yes.

Junhong Chen: Many generations of computers and cell phones, these electronics can have, all of this plastics, components and metals, and some of them are toxic. Plastics are not easily degradable in the environment. So they’re creating a lot of environmental issues.

Paul Rand: So, what to do? Chen has a solution, a really fascinating solution, but it’s kind of... Well-

Junhong Chen: Yeah, so it’s... Right now, it’s a science fiction.

Paul Rand: Okay. So this project may be futuristic, but it’s not quite fictional. It’s being funded by the National Science Foundation. And it’s very much based in real science.

Junhong Chen: Yes, absolutely. We are very excited about this party. We started in early January. It’s really to realize a vision of converting plants into electronics.

Paul Rand: Yes, you heard that right. Chen is working on making electronics out of plants.

Junhong Chen: Right? So we’re growing plants from water actually in this case. So hydroponic grows off plants. And then we extract components from this plants. For example, a cellulose and lignin, and converts this components into bio-based yanks that can be used to print electronic devices, such as the sensors we just talked about, right? And also leasing my own batteries, so that... This a battery power sensors can be further used to monitor the growth conditions of the plants, so that you can precisely goal the plants to enable the best composation to drive the printing passes and to achieve better performance for this devices.

Junhong Chen: So we are producing biodegradable electronic devices from plants, and we can do this better with a AI machine learning, because we are going to be able to learn throughout this process to make this growth process better and better. To achieve better device performance. And another important aspect of this project is in the end we’re hoping to be able to demonstrate a cyber manufacturing platform, it’s a virtual platform for everybody to access so that you can print your own devices from your home or from a nearby library using bio-based genes. Imagine, in the future you want it to have a sensor to monitor carbon monoxide, just go there and get onto the online platform and print it out.

Paul Rand: How far off is this, this type of goal in your mind?

Junhong Chen: I think in five years, this is a five-year project, but we’re hoping to be able to demonstrate such a concept-

Paul Rand: After the break, the connection between our water infrastructure and climate change and what Chen wants to do about it.

Paul Rand: If you’re getting a lot out of the important research shared on Big Brains, there’s another University of Chicago podcast network show you should check out, it’s called Not Another Politics Podcast. Not Another Politics Podcast provides a fresh perspective on the biggest political stories, not through opinions and anecdotes, but through rigorous scholarship, massive data sets and a deep knowledge of theory. If you want to understand the political science behind the political headlines, then listen to Not Another Politics Podcast. Part of the University of Chicago podcast network.

Paul Rand: Going back to our water challenge from before we not only need to make water cleaner, but we also need to find a way to optimize how much energy our systems use. Specifically wastewater systems as with our first problem, Chen sees a solution in AI.

Junhong Chen: In waste water utilities, they’re trying to treat the water to the right quality, right? So, we treat the water to the right level of purity for the right type of applications, right? So thinking about drinking water versus taking a shower versus irrigation or toilet flushing. They all need different quality of water, right? Obviously for drinking, we need the highest quality water to start with. And for taking a shower, maybe not so much, because the more purity you treat the water to, the more energy it takes, right?

Paul Rand: I see, yes.

Junhong Chen: So by treating the water to the right level of purity, for the right applications, you can save energy, AI and machine learning can really help accelerate that process.

Paul Rand: And that only becomes a big deal for all of us that have pets that think of the toilet bowl as a second water bowl source?

Junhong Chen: Exactly.

Paul Rand: Okay. So we’ll have to be careful, at some point leaving the lids down for the animals.

Junhong Chen: And frankly, if you look at the water challenge. Yeah. Overall, the only way to get out of this is to recycle out wastewater. To enable water reuse. That’s the only way to really meet this limited water supply to address growing need for water. So think about this. We need to be able to take up any contaminants as needed to produce precisely the type of water we need. We also need to think about the type of energy source we are using to treat water, right? Even better, we want it to use renewable energy. To treat the water so that it can reduce the greenhouse gas emission, or we can recover energy already containing the wastewater.

Junhong Chen: For example, municipal wastewater, there are lots of organic compounds, right? And also nutrients. And there are critical materials such as, cobalt. And you see what obviously this is even truer, lithium as well. And it is a critical energy and materials that could be recovered to enable better energy footprint and environmental footprint of our wastewater treatment.

Junhong Chen: And that actually... I wanted to finally point out, there’s a exciting opportunity there. To recover critical materials that are important for clean energy applications. I talk about, leasing from seawater and to some of the noble matters such as cobalt, that are important for energy storage applications. We need to be able to ensure adequate supply for this type of elements, but they are present. So we need to be able to recover those greater materials, to enable a sustainable energy future. And also water quality is directly contributing to our global house. But the other way around wastewater based epidemiology could help monitor community level infection of diseases such as COVID-19, right? So the virus shows up in the wastewater, much [inaudible 00:20:24] than clinical confirmation. So that can be used as a way to monitor the potential pandemic outbreak.

Paul Rand: And this doesn’t just stop with wastewater.

Junhong Chen: So more electric power plants, major water users, because they draw really large quantities of water to cool down the power plants. And they also dissipate a lot of heat, right? It’s through that process, right? So, we talked about AI machine learning that can help us optimize the use of the quoting model, depending on operating conditions. You can precisely dose the amount of cooling water needed so that you don’t waste the water to be used for the cooling, not too much. Right? So, that’s one major aspect. And also the AI machine learning models can help improve the energy efficiency or conversion efficiency from the similar energy to electricity, so that you don’t have a lot of heat to be wasted into the wastewater loss after the cooling. So those are some of the major opportunities there.

Paul Rand: All of these ideas are exciting, but it doesn’t mean we’re off the hook. Chen emphasizes that even with these technologies, we all play a role in fixing these problems.

Junhong Chen: And obviously, all this we need public educations, right? And educating everybody to consume water and mitigate water challenges, for example, producing less and waste. For example, plastics into our water, right? That can be very important.

Paul Rand: Like I said at the beginning, “Chen has one of the best job titles, lead water strategist at Argonne National Laboratory.” And if you haven’t heard of Argonne before you probably should.

Junhong Chen: Yeah. Argonne National Laboratory is a national laboratory funded by the department of energy to research energy related challenges from fundamental science to translational efforts so that we are on a trajectory to a safe and secure energy supply formation.

Paul Rand:

Argonne is celebrating its 75th anniversary, this year. It’s rich history of scientific discovery includes all sorts of research on energy infrastructure.

Junhong Chen: And we’re talking about clean energy supply, energy storage, energy management, etc. So it is a comprehensive research and development laboratory with a lot of cutting-edge facilities, world-class facility actually, including some of the user facilities that researchers across the nation, around the world, can come to do experiments and do computational offers. And we also have world-class expos ranging from engineering to science, to social science and life science, etc. A very wide range of expertise that are needed to address the energy challenge.

Paul Rand: One of the consistent themes that I hear on the Big Brains podcast is... Especially when we’re thinking about science at scale. And I think you’re touching on this. It’s not only the science that has to be looked at, you’re talking about public policy issues. You’re talking about economics issues, you’re talking about general society issues. So to solve even the water challenges that you’re talking about, it’s not just what you do in a lab. It has to be looked at holistically, doesn’t it?

Junhong Chen: It’s exactly true. It’s a holistic solution. It’s a convergent solution.

Paul Rand: Mm-hmm (affirmative). And being in a place... I imagine not only at Argonne, but also I imagine it here at the University of Chicago, is the ability to draw on expertise. Looking at all these areas, hopefully is one of the advantages of being at one of the world’s leading research universities that, we can impact.

Junhong Chen: Yes, that’s our advantage. We have all this expertise. Yes. I know across U Chicago and Argonne National Lab. And especially for you U Chicago, we are not only having a new engineering program here, a very frontier engineering program, but also we have traditionally been very strong in social economic sciences and political sciences, etc. We could potentially integrate those strengths into holistic solutions for water and energy problems for our society.

SOURCE  University of Chicago News

Taxonomy