00:00:30 I'm very pleased to be able to talk to the inorganic section today about some of our

00:00:38 work in inorganic and organometallic chemistry. And because I, as well as my intellectual mentors,

00:00:51 have always been interested in the unity of chemistry rather than specialization,

00:01:01 I would like to take this opportunity to tell you a little bit about an analogy or to build up

00:01:07 a bridge between organic molecules and transition metal complexes. And that bridge will be formed

00:01:19 a concept called the isological analogy. But let me first begin by telling you a little bit about

00:01:27 the kinds of molecules that I would like to understand in some detail. And they are quite

00:01:35 typical. On one hand, there are quite typical organometallic molecules of the type that you see

00:01:40 here. For instance, the top line shows some complexes of iron tricarbonyl of molecules

00:01:48 of organic moieties, which by themselves would not be very stable. Cyclobutadiene, trimethylene,

00:01:55 methane, and enol form. The middle shows a fairly typical group of metallocenes and related

00:02:03 molecules, and the bottom some simpler clusters. I would like to see some, I would like to know

00:02:09 something about the electronic structure, so would we all, and we do know a great deal about

00:02:15 the electronic structure of some of these molecules. And I would also like to see something

00:02:21 about the electronic structure of somewhat more complicated things, not things that were made in

00:02:28 the classical antique days of organometallic chemistry, meaning 20 years ago, but maybe

00:02:35 things that are being made today. And here's a typical example of the kind of molecule which

00:02:42 I would like to get a simple picture of, the electronic structure. And this is something which

00:02:49 has a C3 bridge here with a paratallow group and two methyl groups over here. So there's a C3 bridge,

00:02:57 there's a Cp tungsten dicarbonyl, there's an iron tricarbonyl at right. Not too complicated,

00:03:04 but typical of the kind of thing which you'd find in your next issue of chemical communications,

00:03:10 and typical of the kind of thing past which you're, not to speak of a physical chemist,

00:03:16 but your organic chemist is likely to turn to the next page of the journal of chemical

00:03:22 communications. And I view one of my purposes in life is to stop organic chemists, and maybe

00:03:29 physical chemists, so they're almost hopeless, from turning past that page and to show them what

00:03:36 is interesting and exciting about this molecule. And the only way to show it to them is to make

00:03:41 them stop being afraid of the complexity of this molecule, to give them some unifying theme by

00:03:48 which they can tie this to something that they are familiar with, because that's how we could become

00:03:53 unafraid of complexity. And so to begin with, what I want to do is I want to show you some of the

00:04:02 what the particular approach to building a simplifying bridge to the organic chemist in

00:04:10 particular, concerning molecules which are obviously in some way organometallic. There's

00:04:18 obviously some relationship to organic molecules here. Let's see what that relationship is. It's a

00:04:23 pretty easy one. Now if we look at these things, there are many different ways to approach this.

00:04:28 You could do a calculation on the molecule as a whole, and you could do it on the next molecule

00:04:32 along the line, but I'd like to see in these molecules the fragments, the obvious building

00:04:40 blocks from which this is composed. Now this one has a C3 chain, it has a Cp tungsten dicarbonyl

00:04:47 group. I forgot to say, of course, that this molecule, as a number of others you'll see,

00:04:52 comes from the group of Gordon Stone at Bristol, and you'll see a few more of his molecules as we

00:04:58 go along. There's a transferable Cp tungsten dicarbonyl group, there's an iron tricarbonyl

00:05:04 group over here. This molecule comes from Stone's group, but there are other people who have

00:05:09 made this molecule as well. And if we look at some of the previous ones, there are also obviously

00:05:13 fragments in the metal with associated ligands, with 5, 4, 3 ligands, cyclopentadienyl groups,

00:05:20 and it is these fragments which I would like to take as the basis of a conceptual building up

00:05:28 of the electronic structure of these molecules. And in the process of doing that, I will also

00:05:35 get my bridge to organic fragments, because after all, organic chemistry can also be built up from

00:05:41 pieces. Organic molecules can be built up with CH3, CH2CH hydrocarbons. Of course, there are

00:05:50 heteroatoms to make those things important and biologically active, but we've got a plethora of

00:05:54 heteroatoms in transition metal chemistry, so that doesn't bother us. And these fragments of a metal

00:06:01 with associated ligands could be viewed, and in fact I will view them, as building blocks of

00:06:07 organometallic molecules, much as methyl, methylene, methine are the building blocks

00:06:12 of organic molecules. It is true that the logic of organic chemistry developed in the 19th century

00:06:17 in terms of functional groups and substituents, so people didn't bother to look at a benzene ring as

00:06:22 six CH groups. It was non-productive, it wasn't interesting, but there's nothing to stop us

00:06:27 from building up a benzene in that particular way. So we need the orbitals of the fragments,

00:06:33 and one particular, and this is a non-unique approach, there are many different ways to

00:06:37 the electronic structure of a molecule as complicated as this one. There are many different

00:06:42 ways, I'll just show you one. One particular way is to seek out the fragments, which are here a Cr,

00:06:48 a Cp tungsten dicarbonyl, and an iron tricarbonyl, and then after we get under control the orbitals

00:06:55 of the fragments, the frontier orbitals of the fragments, we can assemble the entire molecule

00:07:01 from them. Now just as there is an archetype in organic chemistry of methane, so we could view

00:07:09 as an archetypical form, an architectonic kind of principle, the octahedral principle which we have

00:07:17 inherited from Vernerian times, and we can see these fragments which we need, a metal with five

00:07:24 ligands, a metal with four ligands, metal with three ligands, even as you see a metal with a

00:07:30 cyclopentadienyl, which is really hiding there as a metal with three ligands electronically,

00:07:35 we can see these fragments in an octahedron. And so if we want to construct the electronic

00:07:42 structure of the fragments, these building blocks, one useful way is to use the octahedral starting

00:07:49 point. Now I don't have to go through with you how one would, what the orbitals are of an octahedron,

00:07:58 but I want to do it very quickly in here so that I can get at the ML5, ML4, ML3 orbitals.

00:08:05 And as usual in life, when you want to do something very quickly, you wind up doing it

00:08:09 in a valence bond way. Whenever you want to do anything but the simplest of all possible things,

00:08:15 you find it impossible to do it in a valence bond way. But here I am interested in just getting

00:08:20 very quickly at the orbitals of the fragments which are familiar to all of you, and so let's do it

00:08:25 in fact in a valence bond way. Now here is the way in a kind of mixed valence bond,

00:08:31 mixed valence bond MO picture that one might get at the electronic structure of an octahedron,

00:08:35 which is familiar to us from another way. One would take the valence orbitals of the metal,

00:08:40 the D, S, and P, just as if one had carbon, one would take the S and the three P's,

00:08:46 prepare the thing for octahedral bonding, and sure enough you can form six octahedral hybrids

00:08:53 using two of the D's and S's and three P's, a group theoretical exercise that was done by

00:08:58 Kimball and by Pauling a number of years ago, leaving three orbitals unhybridized, the X, Z,

00:09:03 Y, Z, X, Y, miracle of miracles, and no miracle at all, they turn out to be the T2G set of three

00:09:11 below two in a classical crystal field type analysis of an octahedral molecule. And once

00:09:17 this is prepared for bonding, just like a carbon atom is prepared for bonding, then if you have

00:09:22 six ligands, six Lewis bases approaching this metal atom that's been prepared for bonding,

00:09:28 they take care of these six orbitals, form six bonding, six anti-bonding combinations,

00:09:33 and leave in the frontier region, and that's what we're interested in, the T2G set, the three

00:09:41 orbitals which are not used in hybridization. Now that's the classical picture of an octahedral

00:09:46 complex. Where is the three below two, or rather where are the two? You see the three, the T2G set,

00:09:51 the two are up there among, lumped in this simplified picture among the metal to ligand

00:09:56 sigma anti-bonding orbitals. So that's five, that's six ligands, well what about five? Well

00:10:02 you bring in five, they take care of the six, five of the six hybrids, and one of the hybrids

00:10:07 here, pointing toward where the ligands ain't, is left untouched in the middle of the scheme.

00:10:15 So now it's three below the remnant of the octahedral set, one above. You bring in four,

00:10:21 and they take care of four of the hybrids, and leave two in the middle. You bring in three,

00:10:26 and they take care of three of the hybrids, and leave three in the middle. And here is then

00:10:31 the simplest of all possible pictures of the orbitals of an MLM fragment.

00:10:36 For ML5, three below one. For ML4, three below two. ML3, three below three. The three down below

00:10:45 are the memory of being an octahedron, and so to some extent are these up above. Now I assure you,

00:10:53 and that the details are different, that every fragment is different, whether the ligands are

00:11:00 carbon monoxides, or phosphines, or what you have on there. But still, I'm looking for a unity in

00:11:06 the thing. I can see the differences, and if you've read our papers, you've seen some of the

00:11:12 differences between the various fragments. But if I look for the similarities, that is the similarity

00:11:18 that I perceive if I want the simplest of all possible pictures of these orbitals. Three below

00:11:24 one, three below two, three below three. Now the only thing left is the number of electrons to put

00:11:30 in this. This is one of the greatest barriers that organic and physical chemists have to inorganic

00:11:35 chemistry. And with the help of this marvelous device, which you'll find front or back covers

00:11:43 of your freshman book or anywhere, you can tell your organic or inorganic colleagues

00:11:50 that in fact the iron atom in oxidation state zero has eight electrons to put into the pot.

00:11:56 And so the electrons are assigned to these fragments, and I have put in here, here's an

00:12:01 electron counter's periodic table where the anachronistic eight above the last three has

00:12:06 been replaced by an eight, nine, and ten to help you count electrons in the various elements. And

00:12:14 so we have ml5, ml4, ml3, and the varying numbers of electrons for the first transition series in

00:12:21 this fragments. And the analogy that I will seek out with you, the analogy to organic fragments,

00:12:28 is going to be formed, it's not the only way to do it, but it's going to be formed by my

00:12:33 concentrating on three fragments. The d7 ml5, manganese pentacarbonyl or anything like it,

00:12:40 the d8 ml4, and the d9 ml3. Because these three fragments, once I put in each of them

00:12:49 six electrons into the t2g set, I'm left with one, two, or three electrons to put into one, two,

00:12:56 or three hybrids. And that's going to form a similarity to methyl, methylene, and methine

00:13:02 in just a moment. Let's in fact do the first one right away. Here is manganese pentacarbonyl and

00:13:07 methyl with seven electrons in the valence orbitals of manganese pentacarbonyl, one in the

00:13:14 lone pair of methyl, and these two doublet molecules are alike by having one electron in

00:13:21 each of the, in these hybrids which are pointing away from the skeletal framework.

00:13:28 Now up to this point what I've done is I've told you that they're like, and I've used a

00:13:35 beautiful for log hemi template to draw these orbitals, that lovely yellow template which,

00:13:40 without which our papers couldn't be written, and that's very nice. But do they really look alike?

00:13:48 Do these orbitals, are they in fact similar to each other? Let me show you an attempt to approach

00:13:54 this question. They are similar and they're not similar. They could not be identical. After all,

00:13:58 these are different molecules as far apart almost as you could be from each other.

00:14:02 Here is a contour plot from an extended Hickel calculation of the orbital of the, of the,

00:14:08 of this one orbital in which the odd electron sits for methyl and for manganese with five

00:14:16 hydrides because I wanted to make it as similar to that, and then with the appropriate charge to

00:14:21 make this d7. And of course these orbitals look quite different. If you begin to look at them,

00:14:27 you see their nodal structure. In fact, you begin to see the d-like inside of this and a p-like

00:14:33 outside of that. So they don't look very different, but really what I'm after, you see, is I'm not

00:14:40 after these fragments, though a lot of people might be interested. And there is a lot of interest in

00:14:45 these fragments themselves, trapped in matrices in various ways. We have learned a good bit

00:14:51 through the work of people like Turner and Ozen about the fragments themselves,

00:14:55 their metastable molecules. I'm interested in using the fragments as the building blocks of

00:15:01 organic, metallic, and inorganic molecules. I'm interested in them only in their eventual

00:15:07 destruction, so to speak, while being incorporated in the molecule. So I'm not that interested in

00:15:13 the shape of this. I'm interested what this does when it interacts with something else

00:15:18 in its active principle in some way. And interacting with something else is going to

00:15:22 happen up there. And they look alike, and one way to convince you that they look alike

00:15:27 is in fact done in this plot over here, where I plotted on one axis the overlap between a probe

00:15:34 hydrogen here, a probe hydrogen which has been brought in a certain distance r, that's the other

00:15:40 axis here, to from the metal or from the methyl. And here is the overlap of that probe 1s orbital

00:15:48 with this particular hybrid that you saw in a previous slide. And the overlap is not identical,

00:15:53 of course, it can't be. They're different orbitals. But what's remarkable over how big

00:15:58 a range of distance that overlap parallels each other. And that's not true just for this one

00:16:04 orbital, but for many other orbitals. So in this sense, these are alike. To an incoming hydrogen,

00:16:10 they present a similar kind of overlap, a similar dependence with distance,

00:16:15 a similar shape of the interaction. And it's this which makes me think that, which makes me say that

00:16:21 methyl is like manganese pentacarbonyl. Well, if it's going to be like, if methyl is like manganese

00:16:27 pentacarbonyl, then it better do the same thing, the two things. And they do, to a certain extent,

00:16:33 that it's just like methyl radicals, if you have a lot of them, will dimerize to give ethane.

00:16:39 Well, if you generate manganese pentacarbonyl, and there is a painful history in the literature

00:16:44 that many of you can trace about processes of generating this and knowing what you've got,

00:16:49 but when you generate lots of it, you can get it to dimerize. And if you generate not so much of it

00:17:00 as you would have to have it dimerized, then just like methyl radicals,

00:17:05 these manganese pentacarbonyl or pentacyanocobaltate, another D7 system, they start

00:17:11 radical chains by abstracting atoms by reacting with anything in sight. And just as you can

00:17:16 dimerize the two organic pieces to give a dimer, a saturated molecule, and you dimerize the two

00:17:23 inorganic pieces, so you can co-dimerize the inorganic piece and the organic analog into

00:17:31 a molecule which is perfectly normal organometallic molecule, manganese pentacarbonyl methyl.

00:17:37 Now, no synthetic chemist in his right mind would try to make manganese pentacarbonyl methyl in this

00:17:44 way. You don't make molecules by generating two unstable species and hoping against all entropy

00:17:50 that they'll get together. You make molecules by acid-base reactions, if you can. If you can't do

00:17:55 it, then you may try something like this. But the similar, but that does not prevent us

00:18:03 from constructing a picture of this molecule over here at right, manganese pentacarbonyl methyl,

00:18:10 by dimerizing manganese pentacarbonyl and methyl on paper. I'm using here one of the very, very,

00:18:18 very, very few advantages that a theoretician has in an experimental science, and that is

00:18:25 molecules don't have to be made on paper in the same way that they're made in a laboratory.

00:18:29 And I'm making it on paper by this particular construction principle.

00:18:34 Now, there is a similarity there, and the fragments are similar to each other. They're

00:18:40 not isoelectronic. They're not isostructural. There's something similar about them, and I'm

00:18:44 afraid we've used a word, and we've called two such fragments isolobal, if the number

00:18:51 of symmetry properties, extent in space, and energy of their frontier orbitals are similar.

00:18:56 Not identical, but similar. This is one of those slightly vague, heuristically useful definitions

00:19:04 which make chemistry move ahead. Words like aromaticity come to mind for these things,

00:19:12 and it's not very precisely defined. It has to do with a similarity. It has to do, in fact,

00:19:19 with chemical logic, and that's what's most important, and what some people outside of

00:19:24 chemistry don't understand, that definitions don't have to be precise in order to be useful.

00:19:29 That's one of the great things about chemistry, is we all the time deal with definitions, and

00:19:34 we've learned how to do in our mature, psychologically mature science, to make do with

00:19:39 definitions that are not to make do, to extract maximum utility from definitions which are not

00:19:46 very precise. We've invented even a symbol for it, which you see down below here, which is a

00:19:55 two-headed arrow with half an orbital under it, to remind us of what isolobal means. And the reason I

00:20:02 have the names under there, of Mingos, Wade, Dahl, and Halpern, is to make sure that you don't think

00:20:08 that we've done something terribly original, because the ideas are all there before.

00:20:15 And in particular, Jack Halpern, especially in the middle 60s, in the context of looking at the

00:20:21 reactivity of both things like manganese pentacarbonyl, and of things like Vaska's complex,

00:20:29 looked very often at the similarity, or invoked very often, the similarity between a

00:20:35 D7-ML5 fragment and an alkyl radical, and between a D8-ML4 fragment and a carbene.

00:20:44 And Jack was fully aware of this analogy and has used it throughout. Larry Dahl, in his remarkable

00:20:51 series of structural investigations of cluster compounds, throughout that series, used an analogy

00:20:58 between the orbitals of an MLN fragment and a chalcogen, a sulfur, a selenium, a PR, something

00:21:04 like that. And he was very aware of this kind of analogy and its limits. And Mingos and Wade are

00:21:10 really perhaps the most important names on here. These people developed independently, in the

00:21:17 context of, independently, a theory of bonding and polyhedral transition metal complexes,

00:21:25 focusing on the similarity between the orbitals of such a polyhedral molecule, and another

00:21:34 aspect of chemistry where polyhedral molecules crop up, and that's polyhedral boron hydrides.

00:21:40 And so the analogy they drew was between MLN and BH. Now it's not very far from BH to CH.

00:21:46 And the only reason I go that step is because I want to connect up to a wider group of people

00:21:53 to talk to, and that's organic chemists, to whom BH is very far from CH. And that's why I moved that

00:22:00 one electron away to the parabens. So there's nothing special about this analogy. Let's just

00:22:06 extend it a little bit from here. If manganese pentacarbonyl is isolobal with methyl, well,

00:22:11 if you move one to the, if you move one to the left, that's like methyl cation, taking electron

00:22:16 from that. If you move one to the right, that's like methyl anion. The rules of the game are,

00:22:21 in this game, that you can move within the limits of this imprecise analogy, with its

00:22:29 not very careful definition, you can move up a column of the periodic table. And so if,

00:22:37 in ruthenium, the hybrids, which make for the isolobal character of, for the odd electron

00:22:46 in ruthenium pentacarbonyl, let's say, or the two electrons ruthenium pentacarbonyl,

00:22:50 if the orbitals involved are 4d and 5s and 5p, instead of 3d and 4s and 4p for iron,

00:22:58 it doesn't change the general appearance of the orbitals. And so you can just as well invoke that

00:23:04 for anything under a given element of the first transition series. There is nothing holy about

00:23:10 carbon monoxide. That's just a convenient handle. It could just as well put in phosphines or

00:23:14 chlorides, providing you keep the oxidation state the same. So by the time you put in chlorides,

00:23:19 you have ligands of a very different character, and some of your crystal field splittings will

00:23:23 be very different. Even things as complicated as Myron Rosenblum's favorite ligand, FIP here,

00:23:30 cyclopentadienyl iron dicarbonyl, of course, are also similar to these things. And the way to do

00:23:36 this is to remember that we've usually dealt with two electron ligands, but that cyclopentadienyl

00:23:42 minus. I tend to like to view the ligands for these purposes, for these electron counting

00:23:49 purposes. Of course, we all know for 18 electron rule counting purposes, it doesn't matter what

00:23:55 particular religion you belong to on counting electrons. But for this particular purpose,

00:24:01 for oxidation state formalism, I like to view things as even electron ligands. And so I view

00:24:06 methyl, for instance, as methyl anion, and cyclopentadienyl as cyclopentadienyl anion.

00:24:11 That makes it obvious that it's a six electron ligand. And so when you have FIP, cyclopentadienyl

00:24:17 iron dicarbonyl, the way to do this, I'm sorry for doing this out in public, cyclopentadienyl

00:24:24 minus is the equivalent of three carbonyls. Therefore, this is like iron pentacarbonyl plus,

00:24:30 and iron pentacarbonyl plus is like manganese pentacarbonyl, which is like methyl,

00:24:36 and so this is like methyl. Again, you can take that on faith or not, but if you've been looking

00:24:42 at the literature in the last few years, you've seen an excruciatingly detailed series of papers

00:24:47 by me and my co-workers in which we have built up slowly a library of these fragments. And

00:24:55 I assure you that the similarities are there. There are also differences, and there's no doubt

00:25:00 that if you insist on substituting that fragment by five different ligands with different

00:25:06 ligand field strengths, that you will ruin that analogy eventually. But that's just driving it to

00:25:12 an extreme. It's there approximately. Let's look very quickly at the fragment with eight electrons,

00:25:19 ML4, that you remember ML4 has three below, two above, and iron tetracarbonyl in this octahedron

00:25:26 minus two, or ruthenium ozone tetracarbonyl, is just like methylene in that it has two orbitals

00:25:32 and two electrons to put in them. It has all the problems of methylene for one fragment by itself.

00:25:40 That is, it has a competition between a triplet and a single state as a result.

00:25:45 There are differences in the level ordering between methylene and iron tetracarbonyl.

00:25:50 Of the two delocalized combinations which you'd form from the localized hybrids, which are so far

00:25:56 in this valence bond derivation. But that, those differences, the fact that the symmetric

00:26:00 combination of the orbitals is below the anti-symmetric one in one case, and the other

00:26:06 way around, those differences all wash out when this fragment approaches something else.

00:26:12 And that's what I'm really interested in doing. And so, therefore, you can do from these fragments,

00:26:20 you can put them together then in various ways. And so you can trimerize methylene to give

00:26:28 cyclopropane. Here is an extremely unpopular, or not so unpopular, here is one viewpoint

00:26:35 of iron tetracarbonyl ethylene, which is a well-known complex as a one-third inorganic

00:26:41 cyclopropane. Here is, we have a good number of these by now, perhaps not so many. We have a few

00:26:47 of two-thirds inorganic cyclopropanes, and here is an all-inorganic cyclopropane.

00:26:53 Ah, something has been pulled here. That is, I switched from iron to osmium. And what we have

00:26:59 run into is a limitation of the analogy. That is, that, in fact, osmium 3CO12 has this structure,

00:27:09 but iron 3CO12, as you know, has a structure with two bridging carbonyls.

00:27:15 And so what we have encountered is a limitation of the analogy. This mapping from the organic

00:27:22 to the inorganic side does not guarantee to get you into what is the most stable molecule

00:27:28 on the inorganic side. And in fact, we know that for some ligands, carbon monoxide, isocyanide,

00:27:36 but not others, not phosphenes, things like that, bridging is an everyday fact of life in

00:27:42 inorganic chemistry, especially in the first transition series. And so whereas the D3H

00:27:50 structure for Fe3CO12 is surely not very high in energy above the structure that's in the crystal

00:27:56 and in solution, the ground state structure, that structure is still not the stable point,

00:28:03 and we have to worry when we apply this that we don't necessarily reach the most stable

00:28:08 confirmation. Incidentally, organic chemistry is not all that clear. If you leave the parochial

00:28:17 field of molecules that you can have in bottles and enter the field of molecules flying down

00:28:23 a mass spectrometer tube, then even organic molecules often bridge, and there are carbonium

00:28:30 ions and radical cations which are often bridge structures in some ways analogous to these over

00:28:38 here. Here is the third of the fragments that I want to talk to because I want to get to some

00:28:42 applications. This is cobalt tricarbonyl, or rhodium, or rhidium, and that's analogous to

00:28:48 methyne CH. Why? Because each of these systems, methyne or this cobalt tricarbonyl, has

00:28:59 three hybrids which break up in certain different ways, but three hybrids basically into which to

00:29:05 put three electrons, and if I plotted for you the overlaps of those three hybrids of their symmetry

00:29:11 adapted functions with an approaching ligand, you would see that in fact there are very great

00:29:16 similarities in the overlaps, just as I showed you specifically for the simple case of manganese

00:29:21 pentacarbonyl and for a methyl group, and so these are similar to each other, and so you can make

00:29:28 tetrahedron and a one-quarter, one-half, three-quarters, and all inorganic tetrahedron.

00:29:35 Now I assure you that the makers of these lovely molecules did not call, let's see my light is

00:29:42 about to fade here, did not call this molecule over here a one-quarter inorganic tetrahedron

00:29:49 or this a one-half inorganic tetrahedron. They called this a cyclopropenium complex and they

00:29:54 called this an acetylene complex because that's how they were made, and there is substantial

00:30:00 utility in viewing those in that way, but there is also some, a little bit to be gained in looking

00:30:08 at these, and I think a lot to be gained for the unity of chemistry, in viewing this as a progression

00:30:14 in a series, which in which inorganic and organic pieces are being assembled in with each other

00:30:20 in every possible combination, and one of the interesting things is in fact that in terms of

00:30:26 that rough and word stability, whatever that means, a different thing to every person, that the

00:30:33 inorganic end of the series is more stable than the organic end of the series, and that hidden in

00:30:40 it has some commentary on the importance of strain in inorganic chemistry. And again I have cheated,

00:30:46 I've gone to iridium for the all-inorganic tetrahedron, not to cobalt, because as you know

00:30:54 cobalt has three bridging carbon monoxides, the same limitation of the model, and also we have

00:31:00 here another limitation of the model, and that is I have led you down a garden path by trimerizing

00:31:08 methylene and tetramerizing CH. Why didn't I dimerize the two to give ethylene and acetylene?

00:31:15 That's after all what CH2 would do if you generated lots of it. I didn't do that. The reason I didn't

00:31:21 do it is you know very well that I would have gotten to Fe2CO8, the dimer of the ethylene analog,

00:31:28 and cobalt 2CO6, the acetylene analog, and you know very well that the stable carbonyls are Fe2CO9,

00:31:35 not 8, and cobalt 2CO8, not 6, that these molecules are, if non-existent, then at least

00:31:45 terribly unstable, and so that they are coordinatively very unsaturated. Now that

00:31:53 worries one for a bit, but then I ask you to think in organic chemistry, I ask you to think about the

00:32:01 mapping in which you all you do is you keep carbon but you move down group 4 to silicon

00:32:09 germanium, tin, lead, and think about all the effort that it has taken to establish the

00:32:15 transitory existence and what difference in stabilities there are between carbon and silicon

00:32:21 and germanium and tin and lead in terms of a double bond. So once you view that in a context,

00:32:28 you aren't that surprised that this mapping over to the transition series is not necessarily going

00:32:33 to lead you to a position of great kinetic stability. It's mapping which says that the

00:32:38 orbitals are similar, but actually I'll tell you something about this molecule trapped in a complex

00:32:44 in just a while. In fact, now I have the isolobal analogy constructed and now I'm ready to use it.

00:32:53 I've gone through this very quickly because I think a number of these concepts, and especially

00:32:58 orbitals of fragments, are familiar to a primarily inorganic audience, but this is my analogy up

00:33:07 there and now I want to apply it. Now there are a number of ways in which I can apply it.

00:33:13 The particular first way that I'd like to show you is in a structural way. I'd like to use this

00:33:19 analogy to make sense of the geometrical structure of rather complicated inorganic complexes by

00:33:29 relating them to simpler, hopefully less complicated, at least to organic chemists, less complicated

00:33:40 molecules of one type or another. And let me show you some examples. This is easiest done

00:33:45 and the examples that I have have been, by chance, by the construction of this talk,

00:33:54 picked from what arrived in a Cornell library in July and August of last year.

00:33:59 And in so you'll see molecules which date from that period as far as the literature goes.

00:34:05 Here, for instance, it came in this structure from Herb Case, who is sitting here in the front row,

00:34:11 and we'll see a couple of his molecules, in an issue of Journal of Organometallic Chemistry,

00:34:17 which was rhenium-3, CO-12, dimethyl-10, H. Now the H is here because he was an honest

00:34:25 crystallographer. He didn't put it in. His students didn't put it in where they didn't see it.

00:34:31 In this case, probably bridges across that back edge over there. One idea we have found useful

00:34:38 in treating bridging hydrides is, for purposes of analyzing the structure, is to remove them as

00:34:43 protons and then to look at the electronic structure of the left-behind molecule and then

00:34:48 to reprotonate it. That's just the device. There's nothing wholly about that. And so when I saw this

00:34:54 molecule, I thought of it as rhenium-3, CO-12. You see very good these rhenium-CO-4 fragments.

00:35:02 Rhenium-3, CO-12, dimethyl-10 minus. Now the isological analogy that I want to use is dimethyl-10

00:35:11 with carbene, which most everyone will let me use, though there are very great differences between

00:35:17 these fragments, and of course between iron, between rhenium tetracarbonyl minus and iron.

00:35:25 Use the knight's move in this periodic table that you see over here. Now, given that these

00:35:32 are isolobal, these are isolobal with each other, and these are, they're isolobal, they're all isolobal

00:35:37 with each other, all of a sudden I remember that E.O. Fisher had made a number of years ago in his

00:35:44 group, let me see if I can focus this, this structure over here, which was rhenium-2, CO-8, with two

00:35:52 typical Fisher-type carbenes, sorry, two typical Fisher-type carbenes, paratalomatoxycarbene over

00:36:03 here. So this is related to the previous compound by a replacement of a rhenium tetracarbonyl minus

00:36:11 by the isolobal, by the isolobal carbene and the tin by a carbene. And then one can go in the other

00:36:18 direction. I also remember that Herb had made a while ago, though I'm not sure actually that he

00:36:25 did the structure, I think Mel Churchill did the structure of that one, of, he had made rhenium-4,

00:36:32 CO-16, 2-. This is now moving in the other direction. Now, I was very happy that Herb didn't

00:36:40 see the relationship between these compounds in his paper, but at least not obviously, but I think

00:36:48 it's sort of fun to see that these three molecules are related to each other. It makes you right away

00:36:53 think also, well, let's push more carbon into the thing. On what else can we do? And this is really

00:36:59 what's fun in chemistry for a theoretician is seeing relationships between things that

00:37:04 didn't seem related before, and in fact there are structural relationships of the inner rhenium-2,

00:37:10 CO-8 core which are quite similar to each other. Now, before I talk to you about this, this is

00:37:18 the Fe2CO8 analog of ethylene, which doesn't exist. Well, actually it does in the matrix.

00:37:24 There's good evidence for this in the matrix. That's a relatively unstable thing. We know

00:37:29 that it'll pick up a carbon monoxide if it's got one around to give Fe2CO9, and so here is an

00:37:35 unstable inorganic molecule. Now, the very first slide, how to stabilize it, the very first slide

00:37:41 of this talk tells you how to do it. The way to stabilize unstable organic molecules

00:37:50 is to stabilize them as transition metal complexes, so it's clear if you have an unstable

00:37:55 inorganic molecule, you should stabilize it as an organic complex. Just turn the tables around,

00:38:01 and with a little bit of alchemy, which relates tin to carbon, here it is. Here is two of these

00:38:08 Fe2CO8s stabilized by a tin atom in the middle. This is an old compound. I think this goes back

00:38:16 to the stone group as well, though someone should correct me on that. This, of course,

00:38:25 another way of looking at this is this is an all-inorganic spiropentane. Just replace each FeCO4

00:38:32 by a CH2 and the tin by a carbon atom in the middle, and there it is. Here is something that

00:38:39 came into the library in July from the factory that Jack Lewis and Brian Johnson have in Cambridge

00:38:45 for making metal carbonyl clusters. There came Osmium 5 CO19. The ORTEP drawing is terrible,

00:38:54 and this is a perfect example, this is taken directly from ChemComm, of the kind of thing

00:38:59 your organic friend would turn past. That is, it's a molecule which looks too complicated for

00:39:04 any human mind to think about, but really it's very simple. It's got five osmiums,

00:39:11 all in a plane, and all the carbon monoxides are roughly either perpendicular to that plane

00:39:16 or not, and really what it is, it's got a—they're either perpendicular to the plane or in it,

00:39:25 and what you see here is two of these ethylene-type Osmium 2 CO8 units connected to a

00:39:36 central Osmium CO3, which really is a trigonal bipyramid. This is the equator, and the one in

00:39:42 back and the one on top are the axes, and the other parts of the equator are formed by the midpoints.

00:39:48 It's really beautiful. These Osmium 2 CO8s are puckered back, just like you'd expect an

00:39:53 ethylene to be puckered back, except the puckering back is showing up in the disposition of the

00:39:59 equatorial carbonyls over here. So really what this is, it's really a derivative of

00:40:03 iron pentacarbonyl, where you've replaced the two equatorial carbonyls by ethylenes,

00:40:10 and then just to make you happy, because such molecules exist, done a bit of alchemy to replace

00:40:17 iron by Osmium, and then done a little bit more of the isolobo analogy to put two inorganic analogs

00:40:25 of ethylene into that. It's really a very simple molecule, and it leads you to think of some

00:40:29 chemistry. That is, if you think you can get off an ethylene unit out of this, maybe you can get off

00:40:35 an Osmium 2 CO8 out of this. Now let's do—let's push this a little bit further. Here is ethylene,

00:40:42 here is—I viewed the isolobo analogy of ethylene to Fe2CO8. Now here is the midway point, the half

00:40:50 inorganic, half organic ethylene, which is FeCO4 ethylene, alias a carbene complex, and there are a

00:40:58 few of those around. But now let's—you'll see in a moment my reason for wanting to do what I'm doing,

00:41:04 is let me further shift one electron from the iron to the carbon. That would make iron plus carbon

00:41:14 minus. Iron plus, with a little bit of alchemy, is manganese. Carbon minus is nitrogenous phosphorus,

00:41:21 and so a manganese tetracarbonyl phosphido entity is like iron tetracarbonyl carbene,

00:41:27 which is like ethylene. The reason I do this is because they came into the library,

00:41:33 this paper by—from the group of Pierre Braunstein of Strasbourg, with some of the

00:41:38 structures done in Strasbourg, some done in Grandjean's group in Rennes. This is a typical

00:41:43 reaction of a platinum 2 complex with two chlorides and a chlorodiphenylphosphine, two of those,

00:41:53 with manganese pentacarbonyl anion. In a typical inorganic reaction from which

00:41:59 at least five products were fished out in unspecified yields here, two of the products

00:42:07 were known. One was actually the coupling product of the oxidized anion here. The others were

00:42:14 complicated structures for which the crystal structures had to be done. They involve platinum,

00:42:22 they involve a diphenylphosphido group formed once the chloride was reduced here and popped off,

00:42:28 a carbon monoxide, manganese tetracarbonyl, one carbon monoxide gone, a carbon monoxide

00:42:33 sometimes jumps over to a platinum, a hydrogen comes from God knows where into this. Sometimes

00:42:40 there is a complicated—you could envisage several graduate student theses unraveling the mechanism

00:42:48 of this reaction. And let me show you the three structures which were not known. Incidentally,

00:42:56 this one over here, the one over here is also—the second one which was known—is, of course,

00:43:03 also isolobal to a typical platinum 2 complex. Replace manganese pentacarbonyl by methyl in that

00:43:10 and you have not a known molecule but an interesting one. The other three structures

00:43:15 are shown here. Here is one of them, and now remember what I'd like you to think—I have

00:43:20 to go back—every time that you see this unit, think ethylene. So here is one of these, a platinum

00:43:30 with a carbon monoxide, and here is a diphenylphosphido. The view is horrible because

00:43:35 one phenyl group is getting in front of the other one, but you see here a diphenylphosphido,

00:43:41 manganese tetracarbonyl. You're viewing it down the axis, the carbon monoxide axis here, and so

00:43:49 you can replace that unit that you see there on the side by an ethylene, and this is really

00:43:56 isolobal to what you see at the bottom, which is two ethylenes around a platinum and a carbon

00:44:06 monoxide, and that immediately connects up to a host of chemistry of such molecules made from

00:44:12 Stone's group, which involves several ethylenes and around the zero valent platinum or nickel.

00:44:19 Here is the second molecule. It's a hydride, which we don't see the hydrogen. Here is one

00:44:23 ethylene equivalent here. Here is diphenylphosphido, manganese tetracarbonyl, diphenylphosphido,

00:44:29 and that immediately connects up to a metallocycle and a host of compounds that

00:44:35 Padafat and Pipper and others have studied. Here is the most complicated of the highest

00:44:42 nuclearity complex that the Braunstein group isolated. It's two platinums, two phosphidos,

00:44:48 two ethylene equivalents, equivalent to another class of metallocycle. Now, the Braunstein group

00:44:54 was quite aware of the isolobal analogy in doing this. In fact, these collages that I have here

00:44:59 are taken directly from the French paper. The very same week that this paper came into the

00:45:06 library, the very same week that this came into the library, they came in a South African paper

00:45:13 which had—here is the Braunstein compound at left here, and over at right here you see

00:45:21 the compound—see my thing is fading here—you see here the compound which the South African

00:45:29 group made, which had an iron tetracarbonyl with one carbon monoxide desperately trying to bridge

00:45:35 to the rhodium here, two diphenylphosphidos, two diphenylphosphidos here. If you take one

00:45:44 one electron from platinum to manganese, then you have rhodium and iron. These are the same

00:45:50 molecule made by totally different methods, totally different chemistry, hung on totally

00:45:56 different pegs by their makers. The South African group advertised this, not quite correctly,

00:46:02 as the first molecule with four metal atoms in a row, and the French group had a different peg

00:46:08 to hang their molecules on, but they're really the same molecule. Now let's stretch within this

00:46:14 same group the isolobo analogy just a bit further. I've done this, and I've done this,

00:46:20 now let's go over to this. Now that's—if we go—the easiest way to go from this, of course, is

00:46:27 Cp minus is like three carbonyls, rhodium plus is like iron. Most of my slides have a little periodic

00:46:33 table just like the one that you have taped to your desk to help you along with this, and so

00:46:40 Cp rhodium carbonyl is like iron tetracarbonyl. It's an isolobal replacement, and the reason I

00:46:46 have two of them because, as you know, carbon monoxides move easily in and out of bridges

00:46:51 sometimes, and the structures that I'll show you, in fact, have both of these geometries.

00:46:58 One which has the unbridged geometry, you will see here, this is Hermann's complex of two Cp

00:47:06 rhodium carbonyls with a methylene. This is a two-thirds inorganic cyclopropane. The other one

00:47:13 that I want to show you, which has the bridging carbonyls, requires a little bit of preparation

00:47:20 because now I want to extend the isolobal analogy a little bit further. In particular, I want to

00:47:25 extend it from a d8 ml4 fragment, which is what I've had up to now, to an ml2 fragment, and the

00:47:34 easiest way, there are a number of ways of doing this, but the easiest way to do this extension

00:47:38 from the d8 ml4 to the d10 ml2 is to look at the orbitals, three below two, of the d8 ml4 and to

00:47:50 remove, conceptually, two axial ligands. Now, when you remove the two axial ligands, what you do, of

00:47:56 course, is you free, so to speak, the z-squared orbital. It's been pushed up by this axial crystal

00:48:02 field. z-squared orbital zooms down and rejoins the t2g block. Nothing much else changes, and so what

00:48:10 you have is that these two fragments are analogous if you watch your electron counts.

00:48:19 The electron count that you have to do is take d8 ml4 and relate it to d10 ml2, and you can do

00:48:26 the same thing for ml3 and the kind of fragment which appears in zysosalt, and this is the piece

00:48:34 of the isolobal analogy I want to use because instead of iron tetracarbonyl now, or methylene,

00:48:40 isolobal with it, I want to use a d10 ml2 fragment, in particular a rhodium dicarbonyl minus, to show

00:48:47 you a molecule that comes from Bergman's group in the first instance, though it's been made in a number

00:48:51 of groups elsewhere, and that has this beautiful Star Wars-like picture here, has a Cp rhodium

00:48:58 rhodium Cp, two bridging carbonyls in the background. That's that ethylene analog, and out in front,

00:49:06 at the command of this little module here, is a rhodium dicarbonyl anion. This is really a, this

00:49:12 whole molecule is an anion. This is another view of this. So this is now a three-thirds inorganic

00:49:20 cyclopropane, but with rather interesting mixture of two different kinds of isolobal fragments.

00:49:26 That same unit appears twice again in some of the structures that Gordon Stone has provided.

00:49:33 Gordon Stone, as you'll see in a moment, has been one of the strongest

00:49:37 utilizers and proponents of this isolobal analogy. He's made a fantastic array of molecules

00:49:44 in the next, in the last few years, and I just want to show you a couple more of them. Here is one he

00:49:49 made that you see over here, which has a platinum and two of these ethylene analogs, and earlier he

00:49:57 had made a platinum and two bulky acetylenes. These are analogous to each other, so if you

00:50:01 looked at the crystal structure of that, you wouldn't think that it looks anything like that.

00:50:06 Now here is another extension of the isolobal analogy, which I just want to do as the last

00:50:11 thing, because it's the, it's the, it's, it shows you how complicated you can get with this and how

00:50:18 much fun it can be. Here is another stone, typical stone compound. The kind of thing, again, similar

00:50:25 to the, what I showed in the second slide, which you turn away from, arene chromium dicarbonyl,

00:50:31 a Cr, and a Cp tungsten dicarbonyl. Now, you see a three-membered ring. You think cyclopropane.

00:50:38 This must be analogous to cyclopropane. Then you start applying the isolobal analogy in its

00:50:42 primitive form, and you get into trouble, because this is Cr. This is, once you replace the arene

00:50:50 by five, by three carbonyls, that's a chromium pentacarbonyl, which is CH3+, remember manganese

00:50:56 pentacarbonyl, or CH3. This thing here, Cp tungsten dicarbonyl, that's one electron less even

00:51:03 than this unit. That's like CH3 2+, so you come to a horrible thing. CH3 2+, CH, CH3+, that doesn't

00:51:12 look like anything like cyclopropane. It's got too many hydrogens. It's got pluses all over, but it's

00:51:18 not really as bad as we think. Let's look at that Cp tungsten dicarbonyl, which is one of Fisher's

00:51:24 and Stone's favorite fragments, and look at it again at D5, ML5. Here's ML5, the three T2g orbitals,

00:51:32 and the one hybrid, and when I first formed the isolobal analogy with you, I viewed that as a D7

00:51:40 CH3, and now I'm interested in D5 CH3, Cp tungsten dicarbonyl, or chromium pentacarbonyl plus,

00:51:48 and that looks like CH3++. It's bad. It dips into the T2g shell, and that's not any use to anyone.

00:51:55 Well, actually, let's dip into the T2g shell for the isolobal analogy. Let's identify the pi and

00:52:02 delta pseudosymmetry of these orbitals, so useful constructing metal-metal bonds, and draw a dotted

00:52:11 line around sigma and pi type orbitals, and leave two electrons in a delta orbital, which isn't going

00:52:18 to interact with anything organic, because nothing organic has any delta orbitals to interact with,

00:52:23 and all of a sudden we see three orbitals, one sigma and two pis, which are just what a CH group

00:52:30 has. A CH group, though I don't have it on here, has three orbitals, a sigma and two pis, and three

00:52:35 electrons to put into them. So this is analogous to that. If I drew my dashed line and used only

00:52:42 one of the two pi orbitals, and filled the other one, I'd have my five electrons so that I have one

00:52:50 in a pi and a sigma. One pi, one sigma is like methylene. That's what methylene has, a sigma and

00:52:55 a pi, and so this is like methylene plus. Curious. I'm getting that D5 ML5 is isolobal to CH, CH2+,

00:53:04 and CH3 2+, all three of them. This is a non-isomorphic mapping. This is a many into one mapping,

00:53:13 and this looks a little curious, but what it really is, it's an extension of the isolobal analogy,

00:53:18 which is easy to understand, because it's essentially counts just, the isolobal analogy just counts

00:53:25 the number of orbitals available for bonding, and there's a deprotonation analogy that's at work,

00:53:31 so that methane is as far as a ligand, similar to methyl anion. What happens when you remove

00:53:38 a hydrogen from methane? One of the anti, when you remove a proton from methane, one of the

00:53:43 anti-bonding orbitals goes down, goes away, one of the bonding orbitals goes up, comes up here.

00:53:48 What happens when you remove two protons from methane? Two orbitals come down, or three, or four.

00:53:55 All of these species have the same number of electrons. Now, they're not equal donors toward an

00:54:01 organic species, but they've all got four pairs of electrons, so that removing a proton from

00:54:07 something, and this is, I'm going rather quickly right now, the removing a proton, but you'll be

00:54:13 able to read about this all in science very soon, removing a proton makes no difference to the

00:54:20 isolobal analogy, and so in fact, all of these are in a way isolobal to each other. They're very

00:54:27 different bonding properties. I'm really stretching the analogy, and that's why I had what I had in

00:54:32 the previous slide, something like this CHCH2+, CH3 2+, and that allows me to see right away

00:54:41 the beautiful compounds that the stone group has made, all of them as being related to lots of

00:54:46 other molecules. Here is a selection of four compounds that the stone group has made in the

00:54:51 last year or so. This is a CH2, a CH2 analog, this is a CH3+, molybdenum pentacarbonyl,

00:55:00 which is really also CH2, that's cyclopropane. This is a CH analog, this is a CH2 analog, this

00:55:06 is cyclopropene. These lovely things that they, as well as a French group, as well as an Italian group,

00:55:14 as well as John Shapley have made, these are falling out of everybody's pots now,

00:55:21 these molecules over here. They're fantastic structures which have two carbons, three metals,

00:55:27 often a Cp tungsten dicarbonyl, and an osmium CO3 or a ruthenium CO3. These are all analogs

00:55:34 of cyclobutadiene iron tricarbonyl, easily seen to be and beautifully seen in the structures.

00:55:41 I urge you to look at some of these in the Shapley and Churchill structures or in the

00:55:47 stone group structures. And I think that is where I want to leave you. We began to work

00:55:56 in the inorganic field some eight years ago. We have done a lot of things. The isolobo analogy

00:56:05 that I showed to you is just one of them. I've had many able co-workers, some of whom of the

00:56:11 ones in the early days are listed here, and added to these should be in late days people such as

00:56:18 Cass Tatsumi and Odile Eisenstein. It's these people who've helped me to learn inorganic

00:56:25 chemistry as well as helped me to do a lot of this work. Let me just make two final comments. One is

00:56:31 that probably you find it hard to believe, but I'm prouder of the ACS award in inorganic chemistry

00:56:39 than I am of the Nobel Prize for a number of reasons. One is that it represents for me

00:56:47 life in a sense, that it represents an extension to another field of human endeavor, as little as

00:56:54 an extension as it is from organic chemistry to inorganic chemistry. It's what makes life

00:56:58 interesting. The other thing is it's possibly something that one could aspire to conceivably,

00:57:04 whereas the Nobel Prize is nothing of that sort, and for various political reasons.

00:57:12 And I am very happy to be able to tell you about this, and I'm also very happy about the story as a

00:57:21 whole, because it's basically a story of teaching as well as research, and I think it shows

00:57:30 here what I'm doing. If I haven't taught you anything, I think I teach the organic chemists

00:57:36 something. That is, I've taught them, I think, not to be afraid of molecules of this

00:57:43 complexity as you've seen over here, and I'm very proud of that, and thank you very much.