[00:00:36] Camille Morhardt: Hi, and welcome to today’s InTechnology podcast. We’re gonna talk What That Means: carbon footprinting. And I’ve got with me today MIT professor and Edgerton Chair of the Material Science and Engineering Department, Elsa Olivetti, welcome to the show.
[00:00:53] Elsa Olivetti: Thank you so much, Camille.
[00:00:55] Camille Morhardt: Elsa, I am looking forward to getting into some level of detail about carbon footprint and lifecycle tracking. But can we start with you defining carbon footprint for us?
[00:01:11] Elsa Olivetti: Sure. So carbon footprint is actually grounded in an overall tool to assess impacts of products and services. It’s called lifecycle assessment. So lifecycle assessment is a very systematic, analytical approach by which we can estimate the environmental impacts, all the way from extracting materials to manufacturing those materials, to making them into some sort of product use of product and disposal at end of life.
So the lifecycle assessment methodology basically is an accounting for all of those lifecycle phases and the inputs and outputs of energy and materials and emissions associated with that. And carbon footprinting is one metric that we could pull from that accounting. We could talk about what are the CO2 equivalent emissions from that lifecycle. And then our kind of nickname for that is carbon footprinting.
But because lifecycle assessment is much more comprehensive, you could also talk about damage to resources or ecosystems or human health, those would be broader impacts associated with a particular product. And so carbon footprinting is just one metric associated.
[00:02:17] Camille Morhardt: Is carbon footprinting something we can measure? Whereas when you talk about sort of social impact, it’s like very subjective and it’s harder to set metrics? or are people setting metrics across the board. It doesn’t matter?
[00:02:29] Elsa Olivetti: So there’s kind of a couple ways of thinking about that. You know, there’s no kind of meter that measures carbon, right? We don’t put it on the edge of a material process and measure the CO2 emissions in the comprehensive way that that lifecycle assessment accounts for that. But we can derive models and analyses that say, “okay, based on the sets of processes we’re using, here’s the energy associated with that. And then the carbon intensity of the electricity grid associated with that energy.” You know, whatever’s being plugged in or whatever’s being extracted.
So it’s probably, you know, somewhere in between putting something on a scale and weighing it and something that, you know, what you were sort of referring to is the social context of it, which has a lot more challenges associated with the accounting, but it’s highly complex to distill down all the different lifecycle stages of a product into one number that’s about CO2 equivalent emissions. And depending on where our sources of data are in terms of time and geography, you know that then we’re gonna have the ability to do that with more and less precision and accuracy.
[00:03:30] Camille Morhardt: So you work on an algorithm that attempts to achieve this or achieve a metric or a quantitative figure, and it’s called PAIA and it stands for Product Attribute to Impact Algorithm. Can you tell us a little bit more about that and what its goal is?
[00:03:46] Elsa Olivetti: Yeah, so the motivation for the development of PAIA and the consortium that we’ve been working with is based on the way I just described carbon footprinting; it’s a very resource intensive activity. And so for products like are in the ICT space or you know, electronics, they’re highly complex. The supply chains are very complicated and very dynamic. And so for putting in tremendous amount of resources to account for every little screw, housing material or every step in a manufacturing process associated with integrated circuits, we’re putting in a lot of effort. And by the time we’re done with the measurement, so to speak, the product will have already evolved.
So our motivation in developing PAIA was could we have a streamlined approach to carbon footprinting where we have a statistical understanding of how good our number is, relative to whatever decision we’re trying to make. So if it’s a product change or if it’s some working with the supply chains to mitigate some emissions, we have some confidence that what we’re going after is important relative to other impacts. But we don’t necessarily have the number that knowing the number for carbon footprint is very resource intensive.
And so the motivation for PAIA was twofold: one, can we focus the data collection efforts where it matters most based on a collaboration with industry based on some statistics of understanding that. But can we also be then realizing the number for a particular product based on the attributes of the product? So that’s the, the reason for the name. So can we map things that we would think of as in when we’re like buying a laptop–screen size, what’s the configurations? Where are we located? where do we expect the manufacturing to take place? You know, those are sort of the attributes of the product and also the process. And then build statistical and, and regression based relationships to how that, what that means in terms of what the footprint activities are.
And another motivation for this product attribute. Framing is that maybe then it also, if we were successful, could be a tool that designers could use to help think about changes that they’re contemplating in a product and how are those impacted by the various associated manufacturing and, and materials choices.
[00:05:57] Camille Morhardt: So if we break it down, when you say where it matters in data collection, you’re kind of looking at where in the lifecycle is the biggest carbon footprint generated and you look across the whole lifecycle; so that could be in the extraction of the raw materials that go into the making of the product or to like the ongoing electric consumption of the product as it’s running in the case of electronics. Am I on the right track with that?
[00:06:26] Elsa Olivetti: Yeah, exactly. So which phase in the lifecycle, but also what part of the product, right? What component within the product? So if we’re talking about a notebook, is it the display? Is it the motherboard? You know, all of these things do contribute to the overall footprint, but what matters most, and also what are we most uncertain about? Right? Because that’s where we would then focus data collection efforts. But yes, you’re very much on the right track, what lifecycle stage, but then also what component within the product.
[00:06:55] Camille Morhardt: Okay. And then mapping what matters. When you say like what’s important, the input, so how is that different than the aspects of the product in data collection?
[00:07:05] Elsa Olivetti: We can think about it in terms of a laptop itself. So in our data collection efforts, in terms of understanding what matters, we would say, the critical lifecycle stages would be the manufacturing of the display within the notebook and also the integrated circuits that are present within in the processor for, just as an example. So, and then there will be different parts of each of those components–the manufacturing burden or the materials that are then also driving impact. So that would be where we focus our data collection efforts.
I was differentiating that from once I built the tool and I as a user want to say, “okay, I wanna know what the carbon footprint of my notebook is; can I ask the person who wants to get the footprint for the attributes of the product, like screen size, resolution of the screen, for example.” And have that be the kind of input that the user needs to, to give in order to get the carbon footprint rather than having them talk about the bill of materials, for example, in a lot of detail. Right. So I, I like to sort of, uh, have that in contrast to just, if you were going to buy a laptop, what are the products that you would need in order to customize the product that you wanted? That’s what I mean by the, the product attribute side, if that is helpful.
[00:08:21] Camille Morhardt: Okay. Who’s using the tool? You said it’s a consortium. It’s, it’s coming back under the governance of MIT I think right now. Is it not?
[00:08:27] Elsa Olivetti: MIT is back to managing it, as well. And, but it’s a consortium of about 15 electronics companies right now. And so those are the users and they work with their supply chains, like their customers who are asking for carbon footprints. That’s another motivation, right? So much of the industry is interested in this right now. And so any effort we can try to do to have these on a level playing field, so we know that the assumptions are the same, um, to try to help them be useful in a, in a broad way and that we’re both, we’re harmonizing and streamlining also the requests for data that are happening. That’s another motivation.
And all this gets back to where are we putting resource, right? We gotta make sure we still have resource left to try to do some mitigation based on the footprint result. Right.
[00:09:11] Camille Morhardt: Right. So, well, I guess one question is, what have you found, I mean, you specifically focus on electronics. You know, I assume that that would be different than if you were looking at boat manufacturing or textile manufacturing.
What have you found to be some of the bigger contributors to the carbon footprint? I know electronics is incredibly broad, but maybe even, what are the different categories that people should even consider when it comes to electronics?
[00:09:41] Elsa Olivetti: You, you said this earlier, but the idea of is the lifecycle stage where the footprint is higher, is that in the use phase or in the manufacturing phase? That’s kind of like the first order question. And so that’s one important opportunity for us is where is that crossover point between materials manufacturing and use phase. And so a way to think about that is where are efforts best spent in terms of improving the energy efficiency, um, of the use phase? And as designers and, and you know, electronics manufacturers are driving towards energy efficiency and use, are there trade-offs they’re making in the manufacturing complexity, right? And so that’s, that would be an important finding is kind of where those tipping points are as a function of ways in which emerging technologies would be maybe more specialized and therefore have a higher manufacturing footprint.
So that’s one kind of opportunity that we find in these approaches. Um, and then within the manufacturing, across the whole product, mostly what we find is this balance between what’s being driven by electricity use in a manufacturing facility versus things that are direct emissions. If folks are familiar with the Carbon Disclosure Projects, um, you know, Scope 1, Scope 2, Scope 3 framing.
So it’s basically is it the burden being driven by electricity in the fab or some kinds of processed chemicals or fuel that’s combusted on the facility? And then that, that again, starts to point to what are the ways to prioritize mitigation strategies or decarbonizing strategies? Is it about running things on a renewables heavy grid and or, uh, trying to address some of those specific process level emissions? So those are the kinds of things that we’re working on.
[00:11:27] Camille Morhardt: Are you looking holistically? It sounds like you are across the entire lifecycle of a product such that if you found a product to be maybe higher resource intensity during the manufacturing process, but you could counter that by extending the lifecycle like eightfold, by doing certain other things–like maybe making things like batteries modular so that you could like increase the lifecycle, the rest–does that kind of balance or offset or create, you know, a better reflection on the product overall?
[00:12:01] Elsa Olivetti: Yeah, exactly. That’s exactly the sorts of opportunities that we wanna identify. So in your particular example, we could try to look at what are trade-offs in extending the lifetime–so therefore lowering the manufacturing burden–but you wanna make sure you’re not doing that at the expense of efficiency in the use phase, as new technologies would be adopted.
So, yeah, exactly. And then, you know, there are other kinds of trade-offs in other lifecycle stages. I focused on use phase in manufacturing cause that is our finding of where the biggest burdens are. But you could also think about tradeoffs within the distribution, the transportation associated with the product, even how it’s dealt with that end of life and the, the recycled content of some of the materials. So exactly those sorts of trade offs are part of why we take that lifecycle perspective.
But again, as I said back in the beginning, because it’s focused on carbon footprint. We’re very focused on CO2 emissions where when you start to think about other life cycle stages, end of life in particular, there are other kinds of environmental impacts that we need to be thinking about. Or also extraction processes. Right. Things that, you know, sort of what’s the land use impact or what’s the sorts of human health implications across the lifecycle. So that wouldn’t be under the, the purview of a carbon footprint, but is very relevant when you start thinking about the environmental impacts more broadly of the products.
And, and I think, you know, water use is another one that I should be sure to mention. Mm-hmm. And I think that this, the method that we’ve looked at and within PAIA could be extended to any of these other categories of environmental impact and can also be extended to other products per your earlier comment about textiles, and we’ve done some work in the building space, for example.
[00:13:43] Camille Morhardt: So it’s an algorithm. How does it work? Uh, I mean, am I like filling out a form and it’s basically saying how much?
[00:13:49] Elsa Olivetti: Yeah. So it’s a set of inputs. Mm-hmm. What are the configurations of whatever product you’re looking at? Where are you using it? Um, what assumptions do you want to be inputting based on your particular context? And then the algorithm is, effectively taking what that implies for materials use and what the emissions factors are associated with those materials, but also the manufacturing burden associated with that. And it points to one of the really important efforts within PAIA is trying to make sure we’re using best available data for that, that calculation, like you’re saying. And that’s a significant challenge. And I think I couldn’t emphasize that enough of kind of, you know, a collaborative industry level data collection and transparency. There’s still a tremendous way to go in improving that.
[00:14:39] Camille Morhardt: You mean currently you have companies, let’s say, in even the same industry who may be using different assumptions about how much, emissions there is in extraction of certain raw materials. That may not even be standardized. So when they report their carbon footprint according to their own goals, we don’t even know if, if the measure is the same?
[00:15:01] Elsa Olivetti: That’s at varying levels of being standardized, I guess I’ll say. You know, so a really important part of this exercise, you know, globally is developing product category rules. So what are the sets of assumptions that a sector–so whether it’s cement or heavy duty equipment or electronics or textiles–you know, what are the grounding, level-setting assumptions that that sector wants to assume or, or take on.
The trick with that is there isn’t a global governing board of product category rule development. So often those are done at a national level. In some cases, or they’re done kind of by sector, depending on the sort of convening body there. So, The answer to your question is no, there is not enough specificity in the standards such that, that it’s very difficult to do comparative assessments.
The field is kind of working towards that, but it’s very dependent on the sets of data being more standardized and then the sets of assumptions being more standardized.
[00:16:05] Camille Morhardt: So I’m interested in something I was reading about on your website; basically coming up with novel materials, like replacement materials for things, or substitution or even recycling and making sure the quality is the same or they’re adequately robust for whatever the end purpose is.
I think we’ve all heard of sort of like more eco-friendly cement, things like that, that everybody’s like, “well, you know, we hope it works, but, (laughs) You know, only time will tell some of these things.” Your algorithms are using natural language processing now to sort of comb through everything we know about physics and chemistry that’s been written down. You know, I, I guess I picture it like working away in the background going, well, how about this? Have you thought of this? How about this thing?
[00:16:51] Elsa Olivetti: Yeah. So there are a couple of different applications of that approach. One way to think about it in the materials development side is that the materials community has really done a tremendous job leveraging computational capabilities as those have expanded, right? And so there’s lots of work to do simulations of what kinds of properties would new materials have, and as you said, trying to be predictive about that across a set of properties. Um, but then also, our confidence in, in how that would evolve. So, so that’s one thing. This is sort of like simulation that’s used by the materials community.
And so what we’ve tried to do with the natural language processing tools is have those simulations and even experimental work be grounded in kind of the history of what has been done by the community. So, I sometimes call it like a literature search on steroids, right? So there’s, you know, if you really are trying to effectively source a new material, a new alloy, that you’re saying, “okay, well who else has tried this and what did they find?” That kind of thing. It’s basically how do we expand the knowledge base from which we’re innovating?
And my interest in developing that capability was could we do something predictive to say, well, what would the environmental impact of that be? And maybe we could map that to things like, what’s the complexity of the process to make the material? Or what are our expected byproducts, or what’s the ability to make use of scrap materials within those?
So that’s how this new materials design could potentially leverage the kind of set of knowledge from the material science archive and that’s found in the literature.
[00:18:37] Camille Morhardt: Everything that you’re saying makes it sound like it gets more and more complicated the more you know, the more you know, the more you know you don’t know. Right? And it makes decisions more nuanced and more, more difficult. And people really have to understand, like you’re saying, “well, what weights do I have? What levers do I have even to play with? Which one should I focus on?” And that’s a lot to think about for people who are not, uh, well, we’ll just say for consumers of the products as opposed to maybe the manufacturers.
So do we have this problem where all of the decisions are, are being made realistically by the manufacturers and consumers are not ultimately going to be able to understand really what they should even be asking for? Or is there a way to keep this information or framing of things in the hands of the many, or people can understand what it is they’re asking for and why, and have a clear answer to things and make clear, crisp demands, do you know what I mean?, of the manufacturers?
[00:19:38] Elsa Olivetti: Yeah. There is this ultimate dilemma that we as consumers need to wrestle with and continue to talk to each other about. The most important thing for us to be able to do is decouple growth from resource use; so improvements in quality of life and having that be something we can be working towards that is not always directly coupled to needing more stuff, needing more things, needing more energy. And so that’s a very clear thing we need to be working on and demanding really that each other are very creative about. How can we improve quality of life that is not directly coupled to use of resources.
And so one clear thing to demand for manufacturers is improved service per product, if that makes sense. So the product as service and as a customer and as a consumer, testing and buying into, and being open to understanding the opportunity that is business models that are embracing that. That’s a sort of like very fundamental challenge that we face as a society is when we have efficiency gains, whether it’s energy or manufacturing, we use those cost savings to acquire more. If we want to drive towards the societal change that we need to have, that, we need to learn what it is to push ourselves towards that.
[00:21:22] Camille Morhardt: Well, uh, Elsa Olivetti, thank you so very much. Joining us from the Material Science and Engineering Department at MIT and also a contributor to PAIA, the Product Attribute to Impact Algorithm that MIT puts together along with the consortia to help designers, manufacturers, researchers understand how they can appreciate the carbon footprint throughout a product lifecycle, and then look at what levers can be adjusted to improve it. Thanks, Elsa.
[00:21:43] Elsa Olivetti: Thank you.
