00:00:00 Mark came to Caltech early in 1970, I think, and he didn't stay very long. He stayed a

00:00:28 couple of years, a little over two years, and we had to give him a Ph.D. and get him

00:00:34 on to our sister institution on the East Coast, which is well represented, I might say, Dietmar

00:00:42 Mark in these award addresses this morning. When Mark first came to Caltech, he rode a

00:00:54 motorcycle and he wore a giant helmet and he came in every morning. This helmet's about

00:01:03 as big as he is, and if you haven't seen him before, you'll see that you'll have to listen.

00:01:11 You may not be able to see him over the podium. I hope there's a special lowering device here

00:01:16 for the podium. Mark wore this giant helmet and would come buzzing in his motorcycle and

00:01:22 walk in the lab carrying a giant box of Twinkies, which is the food he ate, and so we gave him the

00:01:28 name Captain Twinkie. I understand MIT has since promoted him to General Twinkie. But I owe a lot

00:01:47 to Mark Wrighton and my own career has been dramatically influenced by him because many

00:01:57 years, a few years before he arrived, I had had many discussions with George Hammond about

00:02:02 starting some work in organic organometallic photochemistry, but George and I never seemed

00:02:08 to get together to start anything in this field until Mark Wrighton arrived. And Mark got started

00:02:17 in this field, he really started up on his own, and George Hammond and I had the pleasure of

00:02:23 working for him for a couple of years. He graduated Caltech after two and a half years,

00:02:32 1972, and went to MIT and became a full professor at MIT in 1977. We're very proud of him, I'm sure

00:02:42 MIT is, Caltech certainly is. He's done a tremendous number of interesting pieces of work in the ten

00:02:51 or so years he's been active in research. He's contributed to our understanding of the photochemistry

00:03:00 of metal clusters and metal-metal bonded systems. He really, really owns this field of cluster

00:03:07 photochemistry, I think. He's shown us the pathways that metal-metal bonded systems use

00:03:14 in photoreactions. He's really introduced the whole area of photogatalysis, organometallic

00:03:22 photogatalysis, where you start with a thermally inert species, radiate it, and generate a very

00:03:27 photochemically catalytically active species from an inert precursor. And quite importantly,

00:03:35 he's contributed significantly to photoelectrochemistry, particularly to the area of

00:03:43 stabilization of light-sensitive electrodes. The big problem there is stability when you

00:03:48 radiate them, and Mark Wrighton and his group have contributed very significantly to our

00:03:53 understanding of the factors that contribute to degradation and has gone even further than

00:03:58 that in showing how to overcome some of these degradative processes and induce tremendous

00:04:06 stability in some of these light-sensitive systems. His award address is entitled

00:04:12 Inorganic and Organometallic Photochemistry, and I ask you to join me in welcoming Mark Wrighton

00:04:20 and congratulating him for this award.

00:04:22 I'll move the mic down at least. The difficulty with being a captain at Caltech is that it's

00:04:41 difficult to get your lieutenants to move to the front. Unfortunately, I still have the task of

00:04:47 doing all the experiments, since George and Harry were always out of town. But it was a pleasure to

00:04:53 work with them, and it's a great pleasure to acknowledge their contributions to my career,

00:04:57 along with the contributions of Jack Saltile, the person with whom I did my undergraduate work. The

00:05:04 work that we've been doing at MIT has been in a large number of areas, but the general theme has

00:05:10 been optical excitation in inorganic substances. And what I'd like to do today is to give you an

00:05:15 overview of some of the concepts that we've been studying and some of the progress that we've been

00:05:20 making in the three areas of interest. On the first slide, I show you the three areas that we've been

00:05:28 working in. Just turn that on, please. The area of inorganic photochemistry, we've been concerned

00:05:41 with the study of primary photoprocesses, study of catalysis, and this third area of photoelectrochemistry.

00:05:47 There are a large number of reasons for studying each of these. Let me mention a couple. In the

00:05:53 area of primary photoprocesses, we're concerned with what happens after optical excitation of

00:05:58 substances. Generally speaking, everything that happens in the ground state can happen in the

00:06:04 excited state, plus a few other things. And each excited state can be regarded as a molecular

00:06:09 entity in its own right, with distinct chemical and physical properties, and we're interested in

00:06:14 understanding what these are and how these relate to our understanding of electronic structure. In

00:06:20 the area of catalysis, we believe that we're in a position to make new forms of catalytic materials,

00:06:25 and at the very least, we think we're able to study better the mechanisms of catalytic processes

00:06:31 in solution. In the third area of photoelectrochemistry, we believe that we have in

00:06:36 hand now the best systems for converting optical energy into electricity and into chemical fuel,

00:06:42 and we're interested in learning how to transduce excited states more efficiently so that we might

00:06:48 be able to use solar energy sometime in the distant future. There are areas of maturity in

00:06:56 photochemistry. Certainly, we can regard agriculture as a mature area of endeavor, and we can think of

00:07:04 this as a technology. Some of the things that we're working on, though, do have the possibility

00:07:10 in the future of becoming technologies such as energy conversion, synthesis, and catalysis. The

00:07:17 only other area that I think can be regarded as a mature technology is imaging. We're always

00:07:22 looking for better ways to image systems using optical excitation, and it turns out that many

00:07:28 of the important ones are based on inorganic substances. So, both from the fundamental side

00:07:33 and from the practical side, I think there are many reasons for being interested in the

00:07:38 photochemistry of inorganic substances. I'd like now to give you a feeling for the three areas of

00:07:44 research that we've been involved in. First, going over some of the primary photoprocesses that we've

00:07:49 uncovered in my laboratory and indicate that some of the products that we make in these primary

00:07:55 steps are themselves very labile. What we're interested in in optical excitation is generally

00:08:02 converting an inert substance into one which is very labile, the excited state. But sometimes the

00:08:07 excited state itself decays to something that's very labile, and an important area in the future,

00:08:13 I believe, will be the study of reactive intermediates that we generate with light.

00:08:17 Historically, this has been true, especially in the area of organic photochemistry. Secondly,

00:08:24 I'd like to show you what we've been able to do in the area of catalysis, and finally,

00:08:28 what we've been doing in the area of energy conversion. Much of our work on primary

00:08:33 photoprocesses has involved the study of rhenium complexes. It turns out that low valent organometallics

00:08:41 of rhenium exhibit a large number of different kinds of excited states which undergo optical

00:08:48 emission subsequent to excitation. Some of these are represented here. We see that emission from

00:08:55 mononuclear substances. Let me see if I can focus that a bit. We see a number of mononuclear

00:09:05 substances which are emissive, as well as some polynuclear substances containing transition

00:09:10 metal-metal bonds, and substances which contain transition metal, main group metal-metal-metal

00:09:15 bonds. Optical emission inherently is not very interesting in terms of chemical reactivity. But

00:09:23 as a property of excited states, it's extremely useful because it tells us something about how

00:09:27 long the excited state will live, and something about its orbital character in terms of electron

00:09:33 distribution. So we can use optical emission as a technique, or as a phenomenon really, that will

00:09:41 give us detailed information about the character of lowest excited states. Generally, the lowest

00:09:46 excited states in these substances are the emissive ones. For example, with this tetranuclear

00:09:52 polyhydride, we were able to investigate the effect of replacing light hydrogen with heavy

00:09:57 hydrogen on the character and the rates of non-radiative decay in these substances. Generally

00:10:04 speaking, changing the vibrational energies in such molecules without changing the geometrical

00:10:10 structure or the electronic structure gives us some important information in how these excited

00:10:16 states behave. In some of these other cases, you'll see subsequently that we've been able to

00:10:21 exploit the long-lived nature of these molecules in order to do bimolecular reactions, especially

00:10:28 electron transfer reactions. But the excited states do a number of different kinds of reactions,

00:10:35 including all of those that occur in the ground state. The difference in producing the excited

00:10:41 state is that sometimes these molecules are much more labile. We can think about increase in

00:10:47 lability by 20 orders of magnitude in some cases. Molecules which undergo chemistry with rate

00:10:53 constants in the range of 10 to the minus 10 per second in the excited state do chemistry with rate

00:10:57 constants greater than 10 to the 10th, and this with only a one-electron excitation. So we have

00:11:03 a very useful probe of electronic structure and a very useful way to create reactive species. We

00:11:11 can see electron transfer reactions by molecular reactions and unimolecular in some cases. We

00:11:17 can see proton transfer reactions. Proton transfer reactions and acid-base equilibria in the excited

00:11:23 state give us some direct information on the redistribution of electron density in these

00:11:28 excited states. The cornerstone reaction perhaps in inorganic and organometallic chemistry is

00:11:33 ligand substitution. Generally, ligand substitution occurs by a dissociative mechanism. In some cases,

00:11:39 associative components do exist, but generally a dissociative component dominates and systems

00:11:47 like these, these mononuclear systems involving the extrusion of a two-electron donor ligand,

00:11:52 undergo substitution from lowest-lying ligand field states. Metal-metal bonded compounds undergo

00:11:58 metal-metal bond cleavage associated with excited states involving redistribution of the electron

00:12:04 density mainly around the metal-metal bond. And in all of inorganic and organometallic chemistry,

00:12:10 we incorporate organic photochemistry of the ligands. And in some cases, the photochemistry

00:12:16 of the coordinated ligands can be significantly perturbed by its coordination. So this is just a

00:12:22 sampling of the kinds of systems and the kinds of reactions that we've been looking at. What I'd

00:12:27 like to do now is turn to some of the reactive intermediates that we've been generating and

00:12:31 indicate some of the chemistry that can be induced by photochemistry and its use in generating

00:12:38 reactive intermediates. Here, for example, is a situation where we've been able to take this

00:12:45 main group transition metal bonded system, excite it with visible light. This absorbs in the region

00:12:51 of 480 nanometers in the presence of an electron donor, such as an amine. The electron donation,

00:13:00 sorry, the electron accepting quencher here, such as methylbiologen, will generate the radical anion

00:13:07 or the one-electron reduced form of the quencher, creating this one-electron oxidized species. And

00:13:14 this is the labile species that we're interested in examining. This radical cation that we formed

00:13:20 from the metal-metal bonded system involves really a hole that's delocalized over these two metal

00:13:26 centers. In effect, we're withdrawing electron density from a bond that involves mainly, or an

00:13:33 orbital that involves mainly, the 10-medium bonding. And what we find is that the metal-metal

00:13:38 bond cleaves, creating this 18-electron compound after scavenging by the donor solvent and creating

00:13:46 a triphenyl-10 radical. The same kind of chemistry can be induced by conventional chemical oxidation

00:13:54 or by electrochemical oxidation, and we have independent routes then of generating the same

00:13:59 radical cation. By any of the routes that we use, we find that the rate constant for cleaving the

00:14:06 metal-metal bond in this one-electron oxidized state is something in the range of 10 to the

00:14:11 fifth per second. So the one-electron change does endow the compound with a fair amount of liability.

00:14:16 We use this kind of information to help us understand the character of the excited state

00:14:22 itself. The lowest excited state, which gives rise to an absorption at about 470, 480 nanometers,

00:14:29 involves removing electron density from the same orbital and putting it into the

00:14:35 phenanthroline ligand. This metal-metal to phenanthroline charge transfer gives us the

00:14:41 same kind of character in the sense that we've oxidized the metal-metal bond. And sure enough,

00:14:47 when we look at the excited state reactions themselves, we find that the chemistry in the

00:14:52 system is dominated by metal-metal bond cleavage with a rate constant that we've measured to be

00:14:57 very similar to that obtained with the radical cation in its ground state. So the excited state

00:15:03 can really be viewed as a situation where we shuffle electron density from the 10-renium bond

00:15:09 into the phenanthroline. The electron is stored there for a long enough period of time that we

00:15:14 can think of having literally oxidized the metal-metal bond. And in this charge transfer

00:15:19 excited state, we see dissociation of that metal-metal bonded system. Another kind of

00:15:26 chemistry that we've been able to see by electron transfer is a new mechanism for

00:15:31 photosubstitution in inorganic substances. Take this acetonitrile cation, for example. If we use

00:15:38 a quencher now that does donate an electron to this excited state, we find that we generate a

00:15:45 19 valence electron species. The 19 valence electron species that we make is substitution

00:15:51 labile. We've added an electron in the excited state to create a ground state 19 electron species

00:15:57 that undergoes rapid substitution in the presence of pyridine, triphenylphosphine, and other

00:16:03 entering groups. The species that we make by substitution is still a 19 electron species,

00:16:10 but it isn't the net product that we see. In fact, this 19 electron species can turn around and react

00:16:17 electron transfer with an original acetonitrile cation, generate another molecule of the 19

00:16:25 electron acetonitrile that goes back and substitutes. And we see as the final product

00:16:30 that substituted pyridine complex. What we have here is a quantum chain process for substitution

00:16:37 induced by excited state electron transfer. The chain comes about because the back reaction of

00:16:44 the quencher in its oxidized state and this reduced compound is slow relative to the substitution

00:16:50 at the 19 electron state and relative to this cross reaction here in terms of electron transfer

00:16:58 between these two species. We've been able to observe quantum yields for substitution which

00:17:02 exceed 50. This kind of photosubstitution where you have a large number of molecules reacting

00:17:09 per photon incident is a situation where you might be able to develop new imaging systems

00:17:15 since many inorganic and organometallic substances do undergo large color changes

00:17:19 upon ligand substitution. So this is a new mechanism for substitution unlike the dissociative

00:17:25 mechanisms and associative mechanisms that we've previously seen for the excited states

00:17:31 of organometallic substances. Another application of electron transfer is to do NIP chemistry on a

00:17:38 ligand. This chlorurhenium tricarbonyl bis ketone compound, ketone being coordinated by the pyridyl

00:17:46 group, undergoes or suffers electron transfer quenching. Triethylamine is a good donor. We

00:17:52 create the radical anion of the complex and we believe from electrochemical studies that the

00:17:58 electron is mainly localized on one of the ketone groups. With subsequent reactions involving

00:18:04 triethylamine and the triethylamine radical cation, we observe as the final rhenium containing

00:18:10 product a coordinated alcohol. If we substitute this coordinated alcohol with another molecule

00:18:18 of the pyridyl ketone, we have affected the net photo reduction of an organic ketone. This is a

00:18:24 reaction that's already known. You can shine light on ketones in the presence of triethylamine,

00:18:29 but the wavelengths of light that are required are those that are absorbed by the ketone. In this

00:18:35 experiment, the chemistry is initiated by absorption of this rhenium complex at wavelengths which are

00:18:43 much longer than the absorption of the free ketone. So in effect, we've brought about a new mechanism

00:18:49 for photo reducing ketones. We coordinate the ketone, induce a low energy charge transfer

00:18:55 absorption, and this allows us to initiate a series of electron transfer reactions leading to

00:19:01 the net reduction of the coordinated ligand accompanied by the usual oxidation products

00:19:06 from triethylamine. So this illustrates how we can exploit electron transfer that is a

00:19:12 bimolecular reaction of excited states. And of course, all of this is possible only because

00:19:19 the excited states are reasonably long-lived. And by reasonably long-lived, we're talking about

00:19:23 lifetimes in the microsecond range. Electron transfer is a reaction that can occur rapidly

00:19:29 in a bimolecular sense, and the quenchers that we've employed do quench these excited states

00:19:34 at nearly the diffusion controlled rate. So the chemistry can be very efficient.

00:19:41 Another kind of intermediate that we've been studying recently is the one that results from

00:19:47 the cleavage of direct metal-metal bonds. This is the original system that we investigated with

00:19:53 Dave Ginley and others in my laboratory where we showed that we could cleanly generate

00:19:58 17 valence electron species. Subsequently, these have been shown by Ted Brown and others to be very

00:20:04 substitution labile, and other workers have shown that in certain instances, redox products do

00:20:10 result from the 17 electron species. Recently in my group, we've been studying the electron transfer

00:20:16 behavior of these 17 valence electron radicals with what would appear to be outer sphere reagents

00:20:22 such as ferrocenium. And we see very clean oxidation of the 17 electron species in the

00:20:28 presence of a donor solvent such as acetonitrile. We observed the generation of this 18 electron

00:20:34 solvated manganese pentacarbonyl. A study of this reaction as a function of the substituents

00:20:40 in ferrocenium do allow us to conclude that this is an outer sphere reaction. For example, if we

00:20:47 look at a series of alkylated ferroceniums, dimethyl, pentamethyl, octamethyl, decamethyl

00:20:54 ferrocenium, we change the oxidizing power of these ferroceniums by about a half a volt. And we

00:21:02 see rate variation over three orders of magnitude with respect to this process in oxidizing these

00:21:10 radicals. On the other hand, if we look at a species like tetracyanoethylene, whose redox

00:21:17 potential in terms of oxidizing power is about the same as our pentamethyl ferrocene, we see that

00:21:23 tetracyanoethylene first of all ends up in the coordination sphere as manganese pentacarbonyl

00:21:29 tetracyanoethylene. The electron is localized on the tetracyanoethylene. We have a cationic anionic

00:21:37 system. And we find that it falls off of our plot of rate versus potential that we get from our

00:21:44 outer sphere reagents. So we're in a position then to look at the electron transfer behavior

00:21:49 of a number of photogenerated intermediates, this one being our prototype system.

00:21:56 Another class of very important intermediates that we can generate, of course, are those that

00:22:01 derive from the extrusion of ligands in the coordination sphere. This is a substitution

00:22:07 inert 18 valence electron alkyl complex. Illumination of this system in the near UV

00:22:14 results in the generation of a 16 electron compound. And the interest here is in studying

00:22:19 the reaction of the 16 electron compound. It turns out that we can examine the chemistry of this

00:22:25 system at very low temperatures. For example, we can generate this 16 electron compound

00:22:32 at temperatures as low as 40 degrees kelvin and then subsequently warm the system and watch the

00:22:39 transfer of the beta hydrogen onto the metal with a concomitant formation of the coordinated

00:22:44 ethylene. This reaction shown at the bottom, the transfer of the beta hydrogen, is one which occurs

00:22:49 at temperatures about minus 100 degrees centigrade. A reaction that has a very low activation energy.

00:22:55 And the point in doing this kind of chemistry is that we're in a position to study reactions

00:23:01 of this kind when we can enter into the system by a photochemical route that usually, or sometimes

00:23:08 at least, has a very low thermal activation. That is, we can do the generation of the reactive species

00:23:14 at 40 degrees kelvin and then this reaction becomes rate limiting. If you try to study the

00:23:19 same system thermally, you're in a situation where the rate limiting step is the extrusion of carbon

00:23:25 monoxide and the following step is very fast. So we're not in a position to study the same system

00:23:31 thermally, but photochemically we think that we can enter into a number of interesting situations

00:23:35 such as this. And this one is particularly important because it brings me to the systems

00:23:40 that we're interested in in terms of catalysis. You can appreciate that this system is one which

00:23:46 could isomerize simple alkenes. For example, if we extrude ethylene from the system and coordinate

00:23:53 cis-2-pentene, reinsert back to an alkyl, we could be at the 2-pentyl stage with 2-pentene.

00:24:02 Beta-hydride transfer again could lead us back to the trans-alkene extrusion replacement with cis,

00:24:08 and you can appreciate that we can be into a catalytic cycle. And we can be into that catalytic

00:24:13 cycle at a temperature where we're able to study something other than the rate limiting step

00:24:18 associated with the loss of carbon monoxide. In particular, we're at a position where we can

00:24:23 enter into the catalytic cycle at a point that matters, that is very close to the key steps in

00:24:29 product formation. So very generally, we're interested in catalytic cycles of this kind.

00:24:36 Generally, we have a series of intermediates involved in catalytic conversion of substrate

00:24:40 to product, and what we're interested in in looking at these kinds of mechanisms

00:24:45 is to enter this cycle at various points. For example, we would like photochemical precursors

00:24:50 to be which is the immediate precursor to product. Generally, we should be able to make a number of

00:24:56 precursors, but of course we're not going to be able to make all of them. In a large number of

00:25:01 instances, we've been able to enter catalytic cycles and see that we can affect the catalytic

00:25:06 conversion of substrate to product at low temperature compared to what is required

00:25:11 thermally, and in a few moments I'll show you some examples. An important thing that we can also

00:25:16 do photochemically, at least in principle, is recover dead catalysts. For example, in this scheme,

00:25:23 A to B to C to D to E to F, this would go on indefinitely except for the fact that I've indicated

00:25:28 an irreversible decomposition into G. G drops out of the catalytic cycle and is ultimately a drain

00:25:36 of the key intermediates. If somehow we could reactivate the catalysis, we could get back into

00:25:42 the scheme and sustain the conversion of substrate to product. The objective then is to photochemically

00:25:49 prepare or to photochemically rejuvenate the system by illuminating G and bringing us

00:25:55 back to D or to E or to any one of these other intermediates in the cycle. So the idea is that

00:26:00 we'll be able to make new kinds of catalysts. That kind of statement is justifiable since the excited

00:26:05 states will do unique chemistry, same kinds of reactions generally as ground state, but with

00:26:11 different specificity. So we can think of making new catalysts, and more importantly, perhaps at

00:26:16 this stage at least, we're able to study catalytic mechanisms and initiate catalysis at low

00:26:21 temperatures. One of the first examples in my group was that done by Mark Schroeder. He was able to

00:26:29 initiate hydrosylation and hydrogenation using chromium hexacarbonyl as a catalyst precursor.

00:26:36 With 1,3-butadiene, for example, we were able to affect the hydrosylation with trimethylsilane

00:26:41 and other trialkylsilanes as well as trialkoxysilanes to generate products that would

00:26:48 appear to arrive from 1,4 addition to the diene, and we only end up with the cis double bond in

00:26:55 the product. The product that you see is generated in essentially quantitative yield.

00:26:59 There is no further isomerization, no further hydrosylation, no further hydrogenation. A very

00:27:06 specific reaction carried out under relatively mild conditions. This kind of chemistry that we're

00:27:12 able to do at room temperature would require high temperature conditions if you insist on using

00:27:17 chromium hexacarbonyl, but as is usually the case when we're able to do this, the ability to do this

00:27:24 chemistry at room temperature indicates that we can enter into the catalytic cycle. Leads us to

00:27:29 conclude that there should be some good thermal route into the same catalyst, and sure enough,

00:27:34 we can take precursors such as the thermally labile trisacetonitrile chromium tricarbonyl

00:27:40 and enter into the same catalytic chemistry. So in these systems where the objective is to make

00:27:45 a catalyst, we haven't really been that successful in making catalysts that you can't make in other

00:27:50 ways, but it certainly is an easy way to do the groundbreaking research that will get you

00:27:57 into the kinds of systems that you should be able to make thermally. Moreover, of course, we will be

00:28:02 able to study mechanism in ways that aren't accessible by simple thermal activation. Another

00:28:08 important system for us has been that associated with the photochemistry of iron carbonyl in the

00:28:14 presence of olefins and oxidative addition substrates like silanes and hydrogen. In this

00:28:20 chemistry, again, that was pioneered by Mark Schroeder, we've been recently looking at the

00:28:25 rate of reaction associated with the intermediates that we've been making. Iron carbonyl is a system

00:28:31 that has been studied for a long time. The primary photo process is the extrusion of carbon monoxide,

00:28:36 and we're able to initiate this kind of chemistry by irradiating the iron carbonyl at liquid

00:28:42 nitrogen temperature and then warming it up. In those kinds of experiments, we find that in the

00:28:47 time it takes to get the sample to room temperature, we can essentially equilibrate

00:28:53 alkenes to the thermodynamic ratio of the internal isomers, and we can affect significant conversion

00:28:59 to these silated products. More recently, Jim Michener in my group has been investigating

00:29:05 this catalysis using intense pulsed laser light excitation. The objective of these experiments was

00:29:12 to irradiate with intense light for a short period of time, make a high steady state concentration

00:29:17 of catalyst, and monitor with molecular specificity the conversion of one pentene into some of these

00:29:23 products. What we were able to determine is that the catalysis is over after a 10 microsecond pulse

00:29:31 or a 10 nanosecond pulse. The chemistry is over on the time scale of seconds, but in that period

00:29:37 of time, activating something like 10 to the minus 3 molar iron pentacarbonyl, we don't know

00:29:43 actually the concentration of the catalyst, but starting with something like 10 to the minus 3

00:29:47 molar, we're able to see 50 percent conversion of neat solutions of substrate to product. So a

00:29:54 significant amount of chemistry can be done in very short order, establishing the intermediates

00:29:59 that we're making here as some of the most active catalysts that have been seen in homogeneous

00:30:04 solution. This kind of work, I think, will take us toward some understanding of just what kinds

00:30:10 of limits we will have in terms of activity. Light enables us to make extensively coordinatively

00:30:16 unsaturated intermediates, should be able to make very reactive species under conditions where we

00:30:23 can both spectroscopically characterize them and subsequently by warm-up under conditions where

00:30:29 we're able to study the interesting chemistry that they will do. Another application is that

00:30:36 in clusters, this first cluster should be the Tris triphenylphosphine complex, but Jim Graff

00:30:42 in my group compared the isomerization activity of this cluster and this mononuclear species,

00:30:48 looking at one pentene going to cis and trans 2 pentene. Not a very interesting reaction in

00:30:54 particular. The interest to us was that we knew that the primary photoprocess in this tetracarbonyl

00:31:01 is the extrusion of carbon monoxide to form ruthenium tricarbonyl triphenylphosphine.

00:31:07 This 16 electron species has the same empirical formula as this cluster. The cluster, we believe,

00:31:14 undergoes metal-metal bond cleavage upon excitation, and we felt that we were in a very

00:31:19 good position to establish whether the cluster remains intact when it's actually doing catalytic

00:31:25 chemistry. And the way that we determined this was to simply look at the cis to trans 2 pentene ratio,

00:31:31 assuming that if we get the same intermediates from the two species, we'd get the same ratio.

00:31:37 And in fact, we found different ratios. We find that the cis to trans ratio differs significantly

00:31:43 depending on whether you start with this mononuclear species, which gets us to the

00:31:47 same simplest formula as this cluster, and yet this persists in giving us a different kind of

00:31:52 chemistry. We invoke different intermediates, and the logical conclusion is that the metal-metal

00:31:57 bond does remain intact. At least one metal-metal bond is intact when this chemistry occurs.

00:32:05 A more interesting system, perhaps, is this one, the tetranuclear tetrahydride dodecacarbonyl.

00:32:12 In this system, we start out with a saturated cluster. In the presence of hydrogen and olefins,

00:32:20 we can affect reduction reactions. And in this case, we believe that the primary result,

00:32:26 primary chemical result of excitation, is the extrusion of carbon monoxide. We're able to do

00:32:31 photosubstitution on these clusters. The quantum yield is low, but it does allow us to bring all

00:32:36 of the elements of importance into the coordination sphere. You can lose a carbon monoxide, coordinate

00:32:42 the olefin, transfer the hydrogen, and then reductively eliminate in order to get to these

00:32:47 reduction products. We believe that the activation of clusters like this is important, perhaps because

00:32:53 we can relate it to surfaces, but at the least we think that we'll be able to make extensive

00:32:57 unsaturation in a situation where there are a large number of metals and all of the elements

00:33:03 needed to do catalysis of an interesting sort. So this is at a primitive stage really in development,

00:33:10 but we are able to affect some reactions that do not occur at room temperature. Similar chemistry

00:33:15 does occur when you heat these systems up, and we believe that similarities in the chemistry are

00:33:20 sufficiently profound that we conclude the same catalytic intermediates to be generated at room

00:33:25 temperature as generated at elevated temperatures thermally. So cluster activation is another area

00:33:31 where we feel we'll be able to contribute in terms of doing mechanistic studies of catalysis.

00:33:38 Finally, let me give you an example of the rejuvenation of a catalyst. It turns out that

00:33:44 dicobalt octacarbonyl is a fairly effective hydrocylation catalyst. However, after a few

00:33:50 hundred turnovers, you get a lot of the alkyl silane and you find that the cobalt is found

00:33:57 as this trialkyl silo cobalt tetracarbonyl. This trialkyl cobalt tetracarbonyl is thermally

00:34:04 inactive at room temperature, whereas the dicobalt octacarbonyl is very active.

00:34:11 Our initial idea was that we would be able to exploit what we've done in terms of metal-metal

00:34:15 bond cleavage to simply cleave the silicon-cobalt bond and regenerate the dicobalt octacarbonyl by

00:34:22 radical coupling, or maybe even generate the active species that derive from cobalt carbonyl

00:34:28 in the catalysis. It turns out that the primary photoprocess in the tetracarbonyl is the loss

00:34:34 of carbon monoxide. That's been confirmed by a number of different kinds of experiments,

00:34:39 but nonetheless, you do get back into the catalytic cycle and you can sustain the conversion

00:34:45 of alkene and silane to the hydrocylation products. The similarity in the product

00:34:49 distribution again indicates that we're back into the same catalytic cycle. So this is a

00:34:54 situation where we have a system that's dropping out of the catalytic cycle that's thermally inert

00:35:00 and we're able to get back in by extruding ligands from the coordination sphere,

00:35:04 suffering oxidative addition and coordination of the essential elements involved in forming

00:35:09 product, and getting back in and turning over a few hundred more times before activation is

00:35:14 needed again with the light. So this is an interesting illustration of the rejuvenation

00:35:19 of a catalyst. The final area that we've been involved in is photoelectrochemistry.

00:35:25 Got involved in photoelectrochemistry in 1975, and shortly thereafter, we came up with some

00:35:33 very interesting results. It turns out that systems that my group and others have worked on

00:35:40 are the best systems for converting optical energy into chemical and electrical energy,

00:35:44 and I think that we made some significant contributions to the state of progress in

00:35:48 this field. The first significant experiment that we did is represented by this illustration.

00:35:54 We have an electrochemical cell into which we've placed two electrodes,

00:35:58 and one of these electrodes is strontium titanate. As an n-type semiconductor,

00:36:04 the counter electrode is a good reversible electrode for hydrogen. Optical excitation

00:36:09 of strontium titanate evolves oxygen and hydrogen from alkaline solutions, alkaline aqueous solutions.

00:36:16 So this has obvious relevance then to converting light into chemical energy in the form of the

00:36:22 products that result from the decomposition of water. This is a very significant finding. It

00:36:29 remains as the most efficient system for sustaining the conversion of light to chemical energy,

00:36:35 the direct conversion, with no other energy inputs other than the light. The overall efficiency for

00:36:41 light in the ultraviolet is about 25 percent, a very significant efficiency. However, if you

00:36:49 would like to do solar energy conversion, the wavelengths of light that will promote this

00:36:53 reaction are those that will excite electrons in strontium titanate. The band gap of strontium

00:36:59 titanate dictates the wavelength response, and unfortunately, the wavelength response is such

00:37:05 that only ultraviolet photons are useful. Therefore, only three percent of the sun's

00:37:10 energy is actually absorbed by this system. Of the three percent absorbed, we can recover one

00:37:17 percent overall of the energy from the sun in the form of hydrogen and oxygen. So we do a good job

00:37:24 on the three percent. The efficiency would be about 25 or 30 percent, but overall we're only

00:37:29 doing conversion with sunlight at the level of about one percent. It's that 25 percent that we

00:37:35 like to emphasize, but of course the problem is that we have to get to significant efficiencies

00:37:40 in order to think about applications in solar conversion. But we do feel that there is

00:37:45 considerable promise in getting to visible light responsive systems. Another interesting aspect of

00:37:52 this particular system is that strontium titanate is an incredibly rugged material from which to

00:37:59 evolve oxygen. Oxygen is a notoriously difficult substance to generate with good kinetics.

00:38:06 We find, for example, here that using ultraviolet laser light, we're able to evolve oxygen with

00:38:12 current densities in the range of five amps per centimeter squared without suffering significant

00:38:18 losses in efficiency. That is, we're still operating at about 25 percent efficiency for

00:38:23 the ultraviolet light. In fact, the major drawback in terms of efficiency seems to be the fact that

00:38:29 we have so much bubble formation at the interface here that we have difficulty getting the light in.

00:38:34 It's scattered, literally frothing here at this electrode with oxygen evolution. The difficulty,

00:38:40 though, is that it doesn't respond to visible light. But it is an important prototype system,

00:38:46 and I underscore it because it is the existence proof that we may be able to do something

00:38:51 sustaining solar energy into other forms of energy. However, when we go to try to find

00:38:57 something that will be responsive to visible light, the problem that Harry Gray mentioned

00:39:02 becomes evident, and that problem is degradation. If we try to find a visible light responsive

00:39:08 electrode, for example, cadmium telluride would be nearly ideal, or cadmium selenide would be

00:39:14 nearly ideal. And in fact, these substances have been investigated, and from the point of view of

00:39:19 energetics, they are nearly ideal. We get, in theory, efficiencies for water splitting from

00:39:25 sunlight in the range of 20 percent, which would be an efficiency that would be high enough. However,

00:39:32 the decomposition dominates. In fact, if you try to use cadmium selenide or cadmium telluride in

00:39:38 this application, you certainly see oxidation chemistry, but it's the oxidation of the

00:39:43 electrode. The only elemental chalcogen that you form from cadmium selenide is selenium, and from

00:39:50 cadmium telluride, tellurium. Hydrogen is evolved, but only as long as the electrode lasts. So

00:39:57 degradation has been an important problem in terms of the photoanode in these devices. At this point,

00:40:04 we decided that there would be a way to do something, provided that we could make the electrode

00:40:10 durable. We have the energetics, but we don't have the kinetics. We have two reactions that are

00:40:16 possible. We have the decomposition, and we have oxygen evolution. Our approach was to introduce

00:40:21 a third possibility. We wanted to put something into the solution that would compete for the

00:40:26 oxidizing power, and here I've represented it very simply, A going to A+. What we needed was a

00:40:33 redox active species A that would completely successfully compete for the oxidizing equivalents.

00:40:40 If we could do that, at the very least, we'd be able to construct systems like that represented

00:40:45 here, where we oxidize A to A+, at the photoelectrode, and then reduce A plus back to A

00:40:51 at the counter electrode. At the same time, remember, the electrons are being pumped through

00:40:56 the external circuit. If we put nothing in the external circuit, just a wire, all we'd do is heat

00:41:01 up the solution. You just have a fancy way for doing self-exchange reactions. However, if we

00:41:08 introduce a load in series in the external circuit, we can recover electrical energy, and the efficiency

00:41:14 is represented here, or in terms of an equation, the efficiency is the voltage that we would

00:41:19 measure here times the current divided by the input optical power. It turns out that systems

00:41:25 like this, mainly through work at Bell Laboratories in the group of Adam Heller and Barry Miller,

00:41:31 systems like these have been the most efficient ways to directly convert sunlight into electrical

00:41:38 energy using chemical systems. The efficiencies that they've measured based on this concept

00:41:44 have been in the range of 14 percent, so very significant efficiencies can be derived.

00:41:49 The original work in my laboratory was done by Art Ellis and Steve Kaiser and Jeff Bolts using

00:41:55 systems like the polytelluride, the polysulfides. Here's a representation of what we did with the

00:42:01 cadmium telluride cell to make it durable. In this instance, we put in the telluride ion

00:42:07 in solution. It's an extremely air-sensitive system, so you have to be very careful in working

00:42:12 with it, and of course you want it to be strongly alkaline in order to avoid other problems. But

00:42:19 it turns out that this is a very good redox couple from the point of view of stabilizing

00:42:24 photoanodes. Not just cadmium telluride, but substances like gallium arsenide and other small

00:42:31 bandgap visible light responsive photoanodes. We can sustain the efficiencies essentially

00:42:36 indefinitely. Very durable systems, and for this particular one, we've seen conversion efficiencies

00:42:42 for monochromatic light, not sunlight, monochromatic light in the range of 10 percent.

00:42:47 This would be a solar efficiency somewhere in the range of 5 percent output voltage indicated here

00:42:54 at about a half a volt. So significant progress has been made in improving durability by selecting

00:43:01 the right redox couples, or you might say that we came up with them serendipitously. But it turns

00:43:07 out that a number of systems are now known for which stable photoanodes can be found.

00:43:15 One of the concepts that we've developed subsequently is the one represented here.

00:43:22 If we find the proper redox system A that will stabilize the anode, we thought that it would

00:43:28 be interesting to covalently anchor that redox system to the surface of an electrode. By doing

00:43:34 this, we confront the surface of the electrode with essentially pure redox reagent. Even when

00:43:40 there's only a monolayer of A on the surface, the electrode thinks it's only looking at A.

00:43:46 Photoexcitation of the electrode then will successfully oxidize A, and once oxidized,

00:43:53 A plus can then turn around and oxidize a species present in solution B. We'll be able to oxidize

00:44:00 anything that's oxidizable with A plus using the wavelengths of light that will excite the electrode.

00:44:06 So in this kind of system, we don't have to serendipitously uncover new redox systems B.

00:44:12 We simply find the one reagent that's completely successful, covalently anchor it to the surface,

00:44:18 and then we're in a position to do a large number of redox reactions.

00:44:22 And one system that we've worked on extensively is the one where the electrode is silicon as a

00:44:28 photoanode, and covalently anchored to it is a ferrocene derivative. Ferrocene is a good electron

00:44:34 transfer reagent, has fast outer sphere electron transfer kinetics, and it is durable in both

00:44:39 oxidation states, ferrocene and ferrocenium. Ferrocenium is not exactly a powerful oxidant,

00:44:46 but it will do a number of things, including the oxidation of iodide to iodine, the oxidation

00:44:51 of ferrocyanide to ferricyanide, and in short, we've been able to show that we can oxidize

00:44:56 anything that's oxidizable with ferrocenium using the wavelengths of light that will excite silicon.

00:45:02 Silicon responds in the region of 1100 nanometers and higher in terms of energy. It's a very

00:45:09 significant visible response. So the concept here is to put systems on the surface that will

00:45:14 stabilize the electrode and then subsequently allow us to do interesting chemistry. In the

00:45:19 meantime, we're always trying to find, by serendipity or by rational means, additives

00:45:28 that would be interesting in terms of doing sustained energy conversion at such interfaces.

00:45:35 And recently we had a very important discovery, this one. We've shown that a number of small

00:45:42 bandgap systems, photoanodes, such as molybdenum diselenide, molybdenum disulfide, and others of

00:45:48 a similar structure that respond to wavelengths of light in the range of a thousand nanometers,

00:45:54 are very durable as photoanodes for the evolution of chlorine, provided you confront the surface

00:46:00 with sufficient amounts of chloride. We can also do the same sort of experiment with bromine.

00:46:06 Efficiencies for these systems are in the range of five percent, that is for conversion of sunlight

00:46:12 to energy in the form of chlorine, and reduction products that would occur over here, such as

00:46:17 hydrogen. But the significance here goes beyond simply being able to find another stable system.

00:46:24 Chlorine is an extremely important material. It's energy intensive, and it's significant

00:46:29 that we're able to make something as potent as chlorine. Most electrodes that would be exposed

00:46:35 to chlorine would react spontaneously. Something like cadmium sulfide, for example, or cadmium

00:46:40 selenide would disappear essentially instantaneously upon exposure to chlorine, just

00:46:45 in the dark. These systems are incredibly rugged and allow us to sustain chlorine or bromine

00:46:51 generation at very high current densities using visible light and with sensible efficiencies.

00:46:56 Moreover, chlorine is a more potent oxidant thermodynamically and kinetically than molecular

00:47:02 oxygen, and it gives us some credibility, I think, in saying that we'll be able someday to affect

00:47:08 the efficient generation of oxygen using visible light, and I think that's an important thing to be

00:47:13 looking for in this area. On the other side of the cell, we're interested in doing reductions,

00:47:20 and it's possible to assist them with light. We're interested in making hydrogen, but like oxygen,

00:47:27 hydrogen evolution is something that has associated with it very poor kinetics,

00:47:33 and it turns out there's no good reason for the surfaces of semiconductors to be any different

00:47:38 than most other electrodes. That is, we expect the kinetics for hydrogen evolution to be poor,

00:47:43 and they are. Even for semiconductors which are durable and have the right wavelength response,

00:47:49 we're still unsuccessful in making hydrogen efficiently because the kinetics are so poor

00:47:53 that when we create our excited state in the semiconductor, the electron-hole pair,

00:47:57 the rate of formation of hydrogen with our excited electron is so slow that electron-hole

00:48:03 recombination dominates. So in this area, our approach has been to modify the surface,

00:48:09 modify the surface in such a way that we're able to accelerate the rate of the reaction of the

00:48:14 excited electron with water to form hydrogen. And our experiments which established that kinetics

00:48:19 were the problem here gave us the direction to take. We showed that we were able to reduce

00:48:25 certain substances that we put in the solution which are as difficult, if not more so,

00:48:30 than water reduction. That is, we put something in that's difficult to reduce,

00:48:33 as difficult as water reduction, and we showed that it was efficiently reduced.

00:48:38 At that point, we took that reagent and decided we would functionalize it in such a way that we

00:48:43 could covalently anchor it to the surface of the electrode. But the reduced substance that we made

00:48:48 was a derivative of bipyridinium, an NN-dialkyl bipyridinium. The reduced form of that substance

00:48:55 is not good for hydrogen evolution either. It merely takes the electron away from the

00:48:59 semiconductor and gives us a little bit more time to operate. We still had to incorporate the

00:49:04 catalyst. And the way that we did that was to do an ion exchange reaction on a polymer of the

00:49:11 bipyridinium. We covalently anchor the polymer, and then we exchange into this polycation

00:49:16 hexachloroplatinum 2-. We then go and use the electrode as a cathode. The platinum 4 is reduced

00:49:22 to elemental platinum. It's dispersed in the polymer, and optical excitation of that modified

00:49:27 surface now allows us to sustain the conversion of water to hydrogen in a thermodynamically uphill

00:49:33 sense, using light to operate on the system. We can sustain it for significant periods of time,

00:49:39 but not with constant output parameters, unfortunately. We do see some decay in terms

00:49:44 of efficiency as time goes on, but we've seen tens of thousands of turnovers. That is,

00:49:51 10,000 oxidation reduction cycles on the polymer can occur without significant decline in efficiency,

00:49:59 and far better efficiency always than for the unmodified surface.

00:50:04 Finally, like the others, I would like to acknowledge the extensive contributions of my

00:50:09 research group. On this slide, I list the graduate students that have been associated with me since

00:50:16 1972 at MIT, PhD recipients and present group members, master's recipients. On this slide,

00:50:26 postdoctoral associates and visiting scientists. I have a slightly smaller army than Dietmar Seifert,

00:50:33 but a very good one. Finally, I'd like to acknowledge my supporters. These agencies

00:50:40 and private institutions have contributed significantly to the progress that we've made

00:50:44 through their financial support. Finally, I'd like to thank the Alpha Chi Sigma Fraternity

00:50:49 for supporting this award, and thank you for your attention.

00:51:05 Thank you, Mark, for a really clear lecture on a illuminating subject.

00:51:09 Will you entertain some... Mark is willing to entertain a question or two or a comment. We have a few minutes.

00:51:21 Not from you, Harry. Yes, I'm sorry.

00:51:33 We expected this.

00:51:34 I'd just like to... That was a lovely talk. I'd like to point out that

00:51:45 the induction of rapid substitution in, say, a 19-electron system, which you talked about,

00:51:57 really has been known for a long time. And what you showed was really a lovely sort of general

00:52:04 bimolecular example of the old case where you you photolyze bromopentamine, cobalt-3

00:52:11 out in the UV. You pop an electron from bromide into cobalt-3 to make cobalt-2.

00:52:17 Bromine pops off. Things scramble in the 19-electron system. You make a very leap

00:52:22 and this induces substitution. So I think you should look at the bimolecular

00:52:29 quenching thing inducing rapid substitution as a special case of this old

00:52:35 intramolecular charge transfer thing.

00:52:41 I'd perfectly agree that there are a large number of situations where we can identify

00:52:46 unit changes in oxidation state with considerable changes in the liability, the classic cases of

00:52:51 cobalt-2 and cobalt-3, chromium-2 and chromium-3, and a number of others. The distinction in this

00:52:57 particular system is that the net substitution is what is our product. In cobalt-3 pentamine

00:53:04 halide complexes, the product is net reduction to form cobalt-2, although there is considerable

00:53:10 substitution liability at that state. The products that we end up with are the simple substitutions

00:53:16 of one ligand for another by way of the electron transfer mechanism,

00:53:20 just simply a quantum chain process.