Nobel Prize lecture: Louis Brus, Nobel Prize in Chemistry 2023

Scientists won the Nobel Prize
12 followers
1 Video View·Nov 19, 2025
The discovery that it is possible to create quantum dots using solution chemistry
was made by Professor Louis Bruce, who was born in Cleveland, Ohio,
received his PhD in 1969 from Columbia University, where he is now a professor.
Please welcome on stage for his Nobel lecture, Professor Louis Bruce.
Okay.
Well, I want to thank the Nobel Foundation for the great, great honor of this prize and
allowing me to talk this morning. A few mistakes creep into the slides, wrong date and so forth.
I got to go back. Okay. I'd like to talk about chemical quantum dots and our effort
to make, to create these things in solution that will typically be involving cadmium selenide.
I'm pushing the wrong button here. Oops, backwards one more time. Yeah, cadmium selenide, 2,6
semiconductor. The red balls are selenium atoms. The small ones, blue ones, are cadmium
atoms or ions. It's a tetrahedral material. Each blue ball is surrounded by four red ones
and each red one surrounded by four blue ones. I'll also talk some about short carbon
nanotubes in order to contrast their behavior to nanocrystals. Here's a slide showing nature's
on the left-hand side, biological nanoscience, starting with macroscopic objects coming down
in size all the way to proteins. Proteins are about the same size as nanocrystals here
or carbon nanotubes. This might be 35 angstroms in diameter. This one might be 10 angstroms
in diameter. This is carbon 60 or buckyball. It was awarded the Nobel Prize some years
ago, Rick Smalley and his collaborators. You can see the DNA over here, maybe 10 angstroms,
20 angstroms in diameter, roughly the same as a larger carbon nanotube. The quantum dots,
the colloidal quantum dots, began when I was a staff scientist in AT&T Bell Labs in the
late 1982 first observations. This Bell Laboratories was the research and development laboratory
of the telephone monopoly in the United States. They were mainly concerned with the science
that would support things that were critical in communications, new materials, mathematics
of the phone system, and computer chips. I want to thank the leadership of Bell Laboratories
for supporting this research in its earliest stages, long before it was recognized as important
by the academic community in the United States. Let's begin with what is a semiconductor nanocrystal.
As Alexei was describing, it's a small piece of rock. In this case, it might be 30 angstroms
in diameter, maybe 10 atoms across from one side to the other. Down here, you can see
that maybe 40% of the atoms are actually on the surface. Made by chemical synthesis,
I don't need to describe this right now, but creating a colloid that has this 30 angstrom
diameter particle with ligands on the surface, ligands to give stability and functionality,
as you'll see in a minute. I'd like to convince you this really is a branch of chemistry.
I show you the apparatus for making quantum dots of a certain size. This is a photograph
taken in my laboratory, a three-necked flask, thermocouple coming in one side, maybe 15
or 20 cc's of solvent on the bottom, a stirring bar, and septum for injection of chemicals,
heating mantle to heat up the solvent. In this situation, we can adjust the temperature
and we can initiate the reaction either by injection through the septum here or by having
one of the reagents already dissolved in the liquid on the bottom. This offers the opportunity
to control the composition as a function of size. You can begin to grow one nanocrystal,
one type of nanocrystal in the bottom, and then halfway through the growth, you can inject
reagents to grow another nanocrystal through the septum and get core shell nanocrystals.
Here's an example of what you can make with time. A cartoon. We start with the reagents
to make cadmium sulfide in the center. Nucleation and growth, you stop at this size, and then
you start injecting reagents to make cadmium sulfide. If the conditions are correct, you
will not nucleate any new cadmium sulfide, but it will grow on the surface of the preexisting
cadmium sulfide. Then you can continue by injecting at the final stage zinc sulfide
reagents, carboxylic acid capping on the surface at the end of the reaction. This is useful
because it's basically a piece of semiconductor surrounded by material with larger bandgaps.
Zinc sulfide is almost an insulator like that. Now let me talk in detail about how to go
through this and what we did. For the young students in the audience, let me say this
is, in fact, the way we gave talks in the 1980s before the invention of PowerPoint.
You see my handwriting cartoon effort to make sense of it all. Initially, we worked just
in solution. My first experiments were done in aqueous precipitation, in aqueous solution,
and sometimes in alcohols, and sometimes at low temperature. The problem with that is
that particles, you'll have two particles that are growing, and they'll diffuse and
touch each other, and they're reactive on their surfaces, so they'll fuse together to
make a dumbbell-shaped particle. Pretty soon you have an aggregation. You don't want that.
You want to grow individual particles. The first effort we made was to use inverse
micelle solutions. This is like in 1986 when Paul Alivisatos first came to be a postdoc.
Inverse micelle solution, maybe 99% heptane and 1% water with some soap. There's a spontaneous
phase separation into these little water droplets 1,500 angstroms across in the heptane solution.
The idea was to dissolve cadmium ions in the water, maybe five or ten, maybe just one.
This will be a tiny reagent vesicle that would allow you to grow one particle inside the
micelle, and it would be protected against aggregation by the others by the soap structure.
We're going to precipitate cadmium selenide, cadmium is dissolved in the water, bring in
an organometallic reagent, trimethylsiloselenide, that will hydrolyze and grow to make a small
particle. That works. As Alexei was describing, you can follow this process by the change
in the optical spectrum, because the color is proportional to diameter. I'll talk about
that in a minute. You can make a solution like this, and then we figured out you can
control the growth. You can stop adding reagents, cadmium all used up, everything precipitated,
let it sit for an hour, and then add more reagents. What you would find is if you added
the reagents slowly, the preexisting small particles would grow to be larger, but you
would not nucleate any new particles. We had a system by which we could grow various sizes.
You start with this and take the spectrum, add reagents, particles would grow larger.
You could tell it was growing larger by the change in the optical absorption and the change
in the fluorescence. Then we thought a little bit more about what to do with the surface.
The surface of the one particle remains very reactive. When Muji came as a postdoc after
Paul went to Berkeley as an assistant professor, we figured out, and it was mainly Mike Steigerwald
who was the driving force behind this, that we could change the nature of the surface
of the particle. If you have this growth process to make a particle, and then in the last stage
you eject phenyl trimethylsilyl selenide. That will react on the surface of the particle
and put phenyl radicals on the surface of the particle. In fact, that changes the particles
from hydrophobic to hydrophilic to hydrophobic. The particles, when they dissolve in the water,
they stay in the micellar solution. As soon as you change the surface like this, put phenyl
radicals on the surface, they come out of the water and go into the heptane. They're
not very soluble in the heptane either. They fall to the bottom of the beaker and you have
pale yellow powder on the bottom of this solution. That means you can extract the particles
out of the micellar solution and you have a pure sample, dry sample of this material.
Then we thought about this a little bit more. In principle, the selenium atoms on the surface
are bound to phenyl radicals. The cadmium atoms on the surfaces are bare in principle.
We wanted to do something with the surface, capacitate them even more. The thought was
to dissolve them in Lewis-based solvent. The lone pairs on the solvent would in fact coordinate,
made a dative chemical bond to the cadmium atoms on the surface. We did that. The first
experiment actually was in pyridine and then refluxing in pyridine. We realized that the
particles were getting better when we refluxed them above room temperature. The weakness
that we didn't understand in the beginning was all of this prior work was done at room
temperature, but the growth is better at higher temperature. Here's an experiment where we
refluxed in tributylphosphine, tributylphosphine oxide at 230 centigrade for three hours. It
made larger particles, beautiful red solution of the Wurzite structure. This was making
zinc blend. This was making Wurzite. You don't need to understand the difference between
these two structures. Anyway, there was a route found to better materials by working
in Lewis-based solvents at higher temperature. This was not understood. What was actually
going on here? It's a recipe for making larger particles, but it didn't work all the time.
It was intermittent. It was obviously being controlled by some impurities, something we
didn't have under control. It wasn't ready to publish by any means, but it was an existence
proof that you could. Nature had found some way to make better particles.
This is the problem that Mungi took to MIT when he went to be assistant professor. You'll
hear about all of this shortly. Bell Labs is a very collaborative institution. You see,
we're able to do a lot of characterization of these materials, this one and that one
as well. This synthesis, chemical synthesis in solution, can work spectacularly well in
when it's optimized. Many people over the last 30 or 40 years have contributed to this
improvement in the processing of the particles and invention of new routes to make different
materials. You see here some data from the laboratory of Chris Murray, who at that time
was at IBM and who's in the audience today. This is making lead cell nanocrystals. You
can see that perfect little cubes. Each one is a single nanocrystal. You see in this higher
resolution image the lattice planes of the individual atoms. Again, this is just an evidence
that you can make really good stuff if you optimize the synthesis. Synthesis is the critical
part of all of this. How was it that we could do this in Bell
Labs with enthusiasm? Bell Labs was concerned with the future of computing and the future
of microelectronics. The main thing that was going on at that time, 1985, was the miniaturization
of transistors, not made by chemical means, but by photolithography. This shows an old
slide, 45 nanometer size, year 2007, transistor on a chip like that, coming smaller and smaller
and smaller on down. The thing which drove the industry, made computers better and cheaper,
more powerful, was this miniaturization. That would have to change. Silicon has certain
properties as a bulk material. When you get it small enough, as Alexei has been describing,
its physical properties, the mobility of the electron, the optical properties would change
and the transistor design would have to change accordingly to reflect the fact that the properties
of the very small particles were different than the properties of the bulk material.
In our minds, we were doing long-range basic research, research of relevance to the future
of the industry, relevance maybe in 10 years, 20 years out, who knows. I looked up in preparation
for this talk the current status of these computer chips. Everyone has a cell phone.
Pull this thing out and in here is a chip designed by Apple, the M1 chip. This chip
has, it's hard to believe, but the truth is it has 110 billion transistors on one chip
in an organized fashion to do computation. Can you imagine, after all this work by so
many people over the decades, 110 billion transistors on one chip. This now has pushed
down into the regime that we were thinking about in the 1980s. The truth is that this
transistor business is one of the great achievements of mankind, in my mind. This has improvement
in communications, improvement in computers, has completely changed the modern world compared
with 50 years ago. A change as dramatic as the improvement of DNA technology for medicine
and for clinical work. We got into this accidentally, as many things
happened in research. It wasn't our intention to look for quantum size effects in the beginning.
I was doing experiments on the surface photochemistry of semiconductor colloids. A colloidal particle
can absorb one photon, create an excited state, which is an electron hole pair. If you have
an organic molecule on the surface, you can get oxidation or reduction reactions. That
experiment is best done in a colloid, because the surface area is so high between the solvent
and the surface of the solid material. We had these old recipes for making colloids.
Occasionally we'd make a colloid, and the band gap was a little bit larger than the
bulk band gap. I didn't know why. You can see that here. This is the optical absorption
spectrum of cadmium sulfide in a bulk material. Basically, it begins at 5,000 angstroms. It
absorbs light to the blue of 5,000. You'd make a small particle solution. The absorption
edge would shift, and a little bump would appear. The bump was, in fact, the beginning
of an exciton. This was evidence that we had, just as Alexei was describing, we were by
accident now getting small enough that we were intermediate or hybrid between molecular
properties and solid state properties. We were in this intermediate regime. Let's think
a little bit about why and how that might be the case. Alexei showed the equations for
the quantum size effect, which come from the band structure of the solid. I'd like to show
an equivalent way of understanding that quantum size effect using the language of chemistry.
The key thing is that in a large silicon crystal, you have tetrahedral coordination, sp3 hybridization,
silicon surrounded by four others in the bulk. That's the same as in a small molecule. Here's
a disilane molecule, sp3 hybridization, sigma-sigma bond between two silicons and then cathode
on the surface by hydrogen atoms. Local chemical bonding around each silicon atom is the same
in the molecule and the crystal. That's the key. Here's the way to understand how this
connects with the solid state. This little red dot is one silicon atom, and what energy
is plotted on this scale. Across here is the size of the nanocrystal, of the number of
silicon atoms in a nanocrystal. If you have two silicon atoms, disilane, sigma bond formed
between one silicon atom and the next like that, bonding orbital with two electrons in
it and the anti-bonding orbital with no electrons in it. This is what we teach in freshman chemistry.
This splitting, sigma to sigma star, is the strength, basically a measure of the strength
of the chemical bonding of the crystal. For a diamond crystal, this splitting is really,
really large. Bonding is very strong. The splitting is different for every semiconductor.
Suppose we get to a situation where we have a really small silicon cluster that has maybe
10 or 15 of these bonds in it with cathode on the surface. You have in fact 10 sigma
orbitals, 10 sigma star orbitals. These couple together to make a narrow band of molecular
orbitals like this, spreading out with the center of gravity with the original sigma
orbital. Many, many more states here and a width developing as a function of size. Mathematics
of this width is known. That mathematics is equivalent to the band structure equations
that Alexei showed. In molecular language, this is a molecule, big molecule. It's a HOMO
and a LUMO and optical transition from here to here. As it grows larger, you have more
and more states down here that get closer and closer together. Eventually they overlap
with respect to their natural widths and you form a band of occupied orbitals, which
in the solid state language is the valence band. There's a band up here of empty orbitals,
the conduction band. You can see that in the asymmetric limit of a larger crystal, there's
a band gap that controls the absorption. To the blue of this band gap, it's absorbing
at every wavelength and it's transparent to the red of this gap. This is the HOMO and
the LUMO of the bulk crystal. This is the HOMO and the LUMO of the nanocrystal. You
can see just by the way the problem is set up symmetrically, this HOMO LUMO transition
is larger in the crystal, the chemical quantum dot, than it is in the bulk material. This
energy difference between here and here is in fact the quantum size effect. It means
that the stability on a per atom basis in the cluster is less than in the bulk material.
The prediction is that in a quantum dot, you have discrete molecular orbitals, although
very close together. If the resolution is good enough, you can see these various transitions
close to each other. The continuous spectrum of the bulk material breaks up into a number
of closely lying optical transitions. It looks like a molecule, like an aromatic hydrocarbon.
An aromatic hydrocarbon solution might have a number of excited states that you can resolve.
Semiconductors luminesce. When they're made well, when they're crystalline and you control
the number of defects in them, they luminesce really well. The optical signals that are
carried in optical fibers are made by little semiconductor lasers. Christmas tree lights
are now made by light emitting diodes of semiconductors. The laser pointer in here is the laser pointer
that's semiconductor based. It's true as well that the quantum dots will emit light very
well if they're well made, crystalline, and the chemical valence on the surface is corrected.
There's another thing going on. It's a shame I don't have a good picture of that. You have
a quantum dot like this. It has an electron and a hole in it. The electric field lines
begin on the hole and they terminate on the electron. They pass through the semiconductor
mostly, but some of the electric field lines fringe out and then come back into the nanocrystal
to terminate on the opposite charge. That means the Coulomb attraction depends on the
dielectric constant of the solvent as well as the dielectric constant of the semiconductor.
This is an attraction and it's a little bit stronger. Let me go to the next slide. If
we drop back into the solid state language, this is the quantum of size effect due to
this incomplete band structure. This is the Coulomb attraction between the electron and
the hole. The HOMO to LUMO transition occurs at this energy, asymmetrically coming into
the bulk band gap but shifting blue as a function of smaller diameter. It's a very simple model.
It's approximately correct, but it has the virtue that there are no adjustable parameters
in this. It's based entirely on things which you can measure in the bulk and then in zero
order you can use it to predict what the band gap will be of a particle of a certain size
of that material. Effective masses and the dielectric constant are different for each
material. Putting this all together, we can recognize
spectroscopic regimes for quantum dot as it grows larger. If you have just maybe 10
atoms, it is a molecule in all respects and the chemical bonding is different than in
the bulk. It has no relation to the bulk. As it gets larger, the chemical bonding and
the structure will shift into being an excise little fragment of the bulk, especially if
you terminate the surface. In this size regime, you get these quantum dots that we're making
through solution chemistry. You get to even larger. When you get to the diameter equal
to the wavelength of light, then there's further changes in the spectra due to electromagnetic
effects. There are also kinetic size regimes that I don't have time to talk about.
I just have one minute. I'll give you what else is going on in nanoscience with respect
to the electrons. What are the electrons doing? When you think about this intermediate size
regime, there are a number of things which happen which are not characteristic either
of molecules or the bulk material. One is this quantum size effect that I've been talking
about. That's a single electron property shown by quantum dots. There's strong electron correlation.
This is single electron property. This is a multi-electron property. For example, if
you have electrons want to avoid colliding with each other. They all have the same charge.
In a bulk material, if one electron is coming like this and the other electron is coming
like this, they're heading towards each other. One can turn up and the other can turn down
and avoid each other. This process of avoiding each other is a very strong function of dimensionality
of the material. It's much harder to avoid the collision in two dimensions and especially
hard in one dimension. This can be shown in excitons and carbon.
Let me go back and get it to the right place. In a carbon nanotube, you can make an excited
state. An azorbic photon makes an excited state. Carbon nanotube is fundamentally different
than a nanocrystal because it has one dimension which is still long like this. An electron
in the carbon nanotube, if the electron is going around the circumference, it comes back
on itself and it has to show positive interference to make a stationary state. That's the same
thing as in a nanocrystal. You can think of it in a nanocrystal. Electron is going, bounces
off the wall, comes back into the interior of the nanocrystal, has to interfere with
itself constructively rather than destructively. This is a little bit like Bohr's original
model for the hydrogen atom. Bohr's model was that electrons are going around in orbits
and there has to be an integral number of wavelengths, de Broglie wavelengths going
around to make a stationary state. 1s, 2s, 1p, 2p, all these were stationary states in
Bohr's model. Since the time is up, let me just go ahead and try to bring this to a conclusion.
There's strong screening as well. If you think about, a carbon nanotube has electrons right
on the surface and so the field, if you create an electron hole pair, the electric field
will come out, pushing the wrong button, the electric field will come out and partially
go into the solvent and then come back again to the opposite charge in the nanotube. It
senses the properties of the solvent. The excited state senses the properties of the
solvent. That's basically a molecular property. A bulk solid state material, everything happens
inside the surface, it's not relevant for a big crystal. This is behaving more like
a molecule. We'll skip this one and skip this comparison. Conclusion, in nanocrystals,
the size dependence of the electrical and operative properties result from simple quantum
confinement with relatively little electron correlation. Nanocrystals are in fact excellent
classes of large molecules and excellent chromophores, if in fact you can make them well and the
structure is annealed. Carbon nanotubes are superb wires. In nanoscience, you can see
from what I've said and what Alexei has said, that synthesis is a critical technology. You
can theorize all day long about this, but what is important is to make them and actually
study what properties they have rather than just think about it abstractly. I'm profoundly
grateful to the Nobel Foundation for this honor, as I said. Also profoundly grateful
to my family for support over the decades, to my wife Marilyn, three children, and now
four grandchildren, all of whom are here in the meeting today. The essence of making progress
originally in Bell Labs was through collaboration, Tim Harris, Mike Steigerwald, Bill Wilson,
Munji, Paul, Jay Troutman, Sasha Efros, for all these different aspects of putting the
problem together. You can see this problem is partly theory, partly organic chemistry,
organometallic chemistry, material science. You need expertise in all these areas.
Going back to what I said about Bell Labs, I'm lucky to have been there when this project
started because it was the kind of institution that could support this strong collaboration
and which understood the importance of this on the 10 or 20 year scale. There was no work
of this sort in American universities at that time. Nobody was thinking about this problem.
If I had gone to the NSF and asked them for money, it wouldn't have worked. Nobody else
would think this was important. Last slide. I think I was in Hong Kong maybe 10 years ago
and had to overnight put together to give some thoughts to a room full of high school students
about this size about what they should be thinking about in doing research. This is
what I came up with. We're all trapped by our educational backgrounds. You come out of school
knowing a certain field. You don't know anything about other fields of science. That limits what
you can do for sure. The way to combat this is every day learn something new. I tell my
grad students that the greatest skill they have is to continue to learn by themselves after they
have left graduate school. Most of the things that I have used in my life were invented after I left
graduate school. I had to keep learning just to keep up with the field. What are your colleagues
working on and why do they think it's important? Search continuously for a better problem than the
you're working on at present. Choice of the problem is the most important scientific decision
you will ever make. If you're a competent scientist and someone gives you a well-defined problem,
you probably can work it out. The hard thing is to figure out what to do in the beginning
if someone doesn't give you the problem. That means to recognize where there's opportunity
that other people don't see yet. Problems are difficult in one field, can be easy in other
fields. The last thing is certainly true, you don't have to be a genius to make good progress.
You have to be dedicated, be intelligent, and have discipline to keep working but not be a true
genius. I tried to tell you a little bit of this and we only got through part of it. I thank you
for your attention today.
