AI‑Driven Materials for Climate Mitigation
Key Points
- Stacey Gifford, an IBM Research scientist, frames her work by asking how it impacts the world, leading her to explore how AI can address the urgent challenge of climate change.
- She emphasizes that climate change is fundamentally a chemistry problem driven by rising CO₂, and that mitigation—through new low‑carbon materials and chemistries—is the preferred strategy.
- AI‑enabled discovery can accelerate the development of green‑chemistry processes, energy‑efficient designs (e.g., aerodynamic structures), advanced batteries, and carbon‑capture/utilization/storage materials.
- By using AI to design and screen novel compounds and materials, researchers aim to create “sponge‑like” membranes for CO₂ capture and convert captured carbon into valuable products, directly supporting climate‑mitigation goals.
- This “chemistry edition” of AI for climate showcases how data‑driven material science can rapidly expand the toolbox needed to limit global temperature rise to 1.5 °C.
Sections
- AI‑Driven Chemistry for Climate Solutions - Stacey Gifford outlines how artificial intelligence can be harnessed to design new materials, positioning AI‑enhanced chemistry as a key tool in tackling climate change.
- Challenges in CO₂ Utilization Technologies - The speaker outlines how converting captured carbon dioxide into products or storing it underground faces hurdles such as high energy penalties, material toxicity, limited stability, and elevated costs that hinder large‑scale adoption.
- AI‑Driven Literature Mining for Hypothesis Generation - The speaker describes using natural‑language processing and knowledge‑graph technologies to ingest and interlink vast scientific publications, enabling rapid, broad‑scope hypothesis creation—such as discovering novel carbon‑capture solvents—far beyond what a single researcher could manually achieve.
- AI‑Accelerated Membrane Material Discovery - The passage explains how AI‑guided monomer prediction, high‑throughput simulation, and automated synthesis can rapidly design, fabricate, and evaluate polymer membranes for selective CO₂ separation, vastly increasing efficiency and reducing cost.
Full Transcript
# AI‑Driven Materials for Climate Mitigation **Source:** [https://www.youtube.com/watch?v=92yV9afmc38](https://www.youtube.com/watch?v=92yV9afmc38) **Duration:** 00:11:43 ## Summary - Stacey Gifford, an IBM Research scientist, frames her work by asking how it impacts the world, leading her to explore how AI can address the urgent challenge of climate change. - She emphasizes that climate change is fundamentally a chemistry problem driven by rising CO₂, and that mitigation—through new low‑carbon materials and chemistries—is the preferred strategy. - AI‑enabled discovery can accelerate the development of green‑chemistry processes, energy‑efficient designs (e.g., aerodynamic structures), advanced batteries, and carbon‑capture/utilization/storage materials. - By using AI to design and screen novel compounds and materials, researchers aim to create “sponge‑like” membranes for CO₂ capture and convert captured carbon into valuable products, directly supporting climate‑mitigation goals. - This “chemistry edition” of AI for climate showcases how data‑driven material science can rapidly expand the toolbox needed to limit global temperature rise to 1.5 °C. ## Sections - [00:00:00](https://www.youtube.com/watch?v=92yV9afmc38&t=0s) **AI‑Driven Chemistry for Climate Solutions** - Stacey Gifford outlines how artificial intelligence can be harnessed to design new materials, positioning AI‑enhanced chemistry as a key tool in tackling climate change. - [00:03:03](https://www.youtube.com/watch?v=92yV9afmc38&t=183s) **Challenges in CO₂ Utilization Technologies** - The speaker outlines how converting captured carbon dioxide into products or storing it underground faces hurdles such as high energy penalties, material toxicity, limited stability, and elevated costs that hinder large‑scale adoption. - [00:06:11](https://www.youtube.com/watch?v=92yV9afmc38&t=371s) **AI‑Driven Literature Mining for Hypothesis Generation** - The speaker describes using natural‑language processing and knowledge‑graph technologies to ingest and interlink vast scientific publications, enabling rapid, broad‑scope hypothesis creation—such as discovering novel carbon‑capture solvents—far beyond what a single researcher could manually achieve. - [00:09:18](https://www.youtube.com/watch?v=92yV9afmc38&t=558s) **AI‑Accelerated Membrane Material Discovery** - The passage explains how AI‑guided monomer prediction, high‑throughput simulation, and automated synthesis can rapidly design, fabricate, and evaluate polymer membranes for selective CO₂ separation, vastly increasing efficiency and reducing cost. ## Full Transcript
Hi, everyone. My name is Stacey Gifford,
and I'm a research staff member at IBM Research,
and as a scientist, I often ask myself the question of,
"how is the work that I'm doing impacting the world around me?"
And I think that this is an important question to ask.
So today we're going to try and answer it
for one of the fastest growing areas of research,
which is AI, or artificial intelligence, and
one of the biggest problems that we have to address today,
which is climate change.
So we're going to ask the question of,
"how can AI help solve climate change?"
And, full disclosure, I'm a chemist
so we're going to focus on how we can use AI
to develop new materials to help solve climate change.
So we're going to call this the chemistry edition.
OK, so before we dive in,
let's talk a moment about the problem at hand.
So most of you probably already know this,
but climate change is the observed increase
in global temperature that we've observed over time.
And if we look at the curve of the past hundred years or so,
it looks something like this.
And we know that we're on track today
to hit a 1.5°C increase in temperature by 2040.
And that's not very far away.
And while 1.5°C may not seem
like a whole lot of change, we know
that this is going to lead to more extreme events
like hurricanes,
tornadoes, floods, etc..
OK, so we need to address this problem urgently.
And at its core, this is really a chemistry problem.
So the reason that we're
seeing this increase in global
temperature is due to rising
greenhouse gases
and primarily carbon dioxide.
Now we can solve this problem in two ways:
we can either mitigate it,
which means that we stop it before it happens,
or we can adapt,
which means we deal with it once it's here.
And from a materials and chemistry
perspective, really what we want to
do is mitigate.
We want to develop new materials and
new chemistries that can help solve
climate change and stop it from happening.
And there's a couple of different ways we can do this.
So there are what we call low carbon technologies,
and this includes things like green
chemistry that can be synthesized
with lower carbon emissions
and produce things that are more
environmentally friendly.
It also includes things that operate
with less energy requirements.
So if you think of something that's
more aerodynamic, it's going to
require less fuel, which means fewer
carbon emissions.
It's also includes things like
energy storage.
And so this is batteries,
which are really critical for
enabling renewable energy,
which is an important part of how
we're going to address climate change.
And then finally it includes things
like carbon capture, utilization,
and storage technologies.
And these are materials that act a
bit like a sponge or membrane that
separate out CO2 from either the
atmosphere or point sources like
power plants.
And then the utilization and storage
part is, once we have that carbon
dioxide in hand, we can either
convert it into things that are useful,
like drugs and new materials,
or we can store it underground
where it turns into rock
and it's stable for millions of years.
So these technologies all exist
today in some capacity, but they've
been limited by a couple of
different challenges, and
they can kind of be broken down into
a couple of key areas.
So the first is performance.
And one example of this is if you
have a material like a solvent or
a solid absorber for carbon capture,
usually you have to heat it up
or pressurize it to release that CO2
back so you can go do something with it.
And that requires fuel
and it requires energy, and it means
that you're actually emitting some
carbon dioxide while you're capturing it.
And we want to minimize that energy
penalty as much as possible.
And there's a lot of room for improvement today.
We also have issues with toxicity.
So if you think about the materials
that go into batteries today,
many of these are very toxic to the environment.
So while we're solving one problem
and helping to address renewable energy,
we're creating other problems
elsewhere in the system.
Stability is another problem.
So we know that many of these materials
break down quickly over time
and ultimately this
all leads to an increase in cost.
And in order to really drive large
scale adoption of these
technologies, we need to bring the
cost down.
And at its core, all of these things
are can be addressed by
designing new materials and new
processes.
But if you think about doing that in
a traditional laboratory setting,
it becomes really complicated and
almost overwhelming.
So if we take one problem, right, so
let's talk about creating a new
solvent that's better than
any other solvent on the market
today for carbon capture.
The solvent space alone has
thousands of candidates that already
exist today, even if we don't
discover a completely new one.
And then if we start talking about
solvent blends where you're mixing
two or three of these together to
get better performance, now you're
talking about millions of different
combinations.
And that's not even considering the
processes and the operating
operating conditions that these work
under. So you can improve the pH,
the temperature, the pressure,
the concentration of CO2, all of
those things have an impact on
performance.
And so now you're starting to talk
about millions of different
experiments that have to be
performed, and this becomes
an incredibly hard problem to solve
on an experimental level.
And this is really where AI comes
into the picture.
So if we go back to our
high school science class and we
think about the sort of tried and
true scientific method, it looks
a little bit something like this,
right? So you have a question.
You have a hypothesis.
You test that hypothesis with
experiments.
You observe the results.
And then you draw some conclusions.
And then we iterate through this
many, many times
before we really decide
on some fundamental conclusions from
our experiments.
So this can take a long time to do,
especially if you have lots of
different conditions that you have
to try.
And we can use AI to speed this
process up and make it work better.
So the first place that we can use
AI is in the question and
hypothesis generation part.
So we can use techniques like
natural language processing to
ingest thousands of journal articles
that exist out there, which is
really sort of the foundation of our
scientific knowledge.
And start to link entities so
we can link materials and different
properties in their performance
metrics and other properties
and build knowledge graphs.
And these allow us to explore
the literature in a way that's much
more meaningful than if you were to
have a single scientist reading
through a journal article after
journal article.
Because if you think about it, the
scientists is going out to create a
new solvent for carbon capture,
they're going to read the carbon
capture literature first and maybe
the solvent literature, but they're
probably not going to explore a
whole lot beyond that because
there's just too much to look at.
But if you have a way of ingesting
all of that literature and linking
it up all together, then you can
really start to make insights in new
areas using other materials that
maybe nobody's looked at before.
And that's really where the
interesting and exciting hypotheses
come from.
We can also use query technologies
to start to ask interesting
questions, so we can say, "what is
the best performing solvent on the
market today?", and it's probably
going to tell us MEA because that's
what's most widely used, but there
could be other candidates that get
linked up through those entity
recognitions to help
bring to light new materials that
nobody's really looked at before.
So that's one area.
The second area is that we can start
to predict new materials.
And we can do this through
generative modeling.
And this allows
us to
set boundaries and say, "we want a
material that has these properties
or these performance metrics", and
we can start to predict new
chemistries and new materials that
meet those requirements.
We can also use data driven
techniques like machine learning to
bring in large datasets.
So say we have a dataset on the
performance of thousands
of solvents.
We can start to look at how we mix
this together to actually predict
new mixtures that can outperform
what's there today.
And then we can use other techniques
like quantum chemistry.
And this can allow us to predict
single properties of materials and
how they might perform and rank them
against each other.
Then we can start to look at their
performance so we can assess
them and
down select.
And this is really before we even
get into a chemistry lab so
we can use
AI-enhanced simulation
to predict how
different materials will perform
under certain conditions.
So, for example, if we're able
to predict new candidates
of monomers that can perform
that can create a polymer
that we can use as a membrane to
separate CO2 from nitrogen,
we can simulate that
membrane and then we can look at
its permeability and selectivity for
carbon dioxide versus nitrogen
versus NOx versus SOx versus other
contaminants.
And we can do this in a way that's
higher throughput than we could do
in a laboratory setting, and it's
much cheaper because we don't have
to go create all of these membranes
from scratch.
And this really drives up the
efficiency. So we know that this
increases efficiency several orders
of magnitude.
And it makes a big difference on our
ability to discover new materials.
And then finally,
when we actually do want to move
into the real world and start
thinking about creating things,
we can use AI-enhanced
synthesis
and automated labs
to speed this up.
So using enhanced synthesis
means that we can
predict retrosynthetic
pathways
to predict the best ways to make new
materials. And "best"
can mean a lot of different things,
right? So it could be that it's
the cheapest way to make it, or it's
the most environmentally friendly
way to make it, or it's the purest
way to make it.
And then once we have those
conditions, we can go into an
automated lab and synthesize them in
a high-throughput manner and then
also test their performance.
And really, what this means is that
we're able to both validate
and we can generate data.
And that data can
feed back into
both our knowledge graph, so that we
get a better understanding of the
full picture of the scientific
knowledge that's out there, and also
into our models so that we can be
better at predicting new materials
moving forward.
And so hopefully what you're seeing
here is that the same cycle that we
use for the scientific method
is replicated in the AI-enhanced
version of the scientific method.
And I think what we're going to see
over the next couple of years is
that this process is really going to
accelerate discovery and we're going
to see significant impact on how we
address climate change with new
materials.
Thanks.
Thank you. If you have any questions,
please drop us a line below.
And if you want to see more videos like this in the future,
please like and subscribe.
If you want to learn more about AI for climate change,
please visit the links in the description.