00:00:30 And I want to pay my respects to the Lubrizol Corporation for their continued support of

00:00:38 this award, making it possible for the Petroleum Division to select the candidates each year.

00:00:49 All the preceding speakers have talked about things that I have been very interested in

00:00:55 over the years.

00:00:56 Milt Orchard was not the first because he was the fellow who wanted to try things out.

00:01:02 Many years ago, Milt Orchard came over to me one day and said, over in Germany, they're

00:01:07 running some reaction, I don't exactly know what it is, but I think it makes aldehydes

00:01:13 from olefins and CO and hydrogen.

00:01:15 And out of that came that first 1948 paper that you referred to, and in it, by the way,

00:01:22 we mentioned that it is quite likely, or possible anyway, that free radical reactions

00:01:27 may be catalyzed by cobalt carbonyl, but it took a long time until you got to that point.

00:01:34 The subject of my talk is sort of pulling in, or trying to pull in, many of the talks

00:01:42 or the ideas that have been brought forth today.

00:01:47 It's an attempt, and don't be too put off by some of the things you may hear, but at

00:01:53 least it endeavors to pull things together.

00:01:57 What I'm going to do is try to relate and compare and perhaps contrast two reactions,

00:02:06 and two that you've just heard about.

00:02:09 The hydroformylation reaction, which is homogeneously catalyzed, and the oxo reaction, and the fissiotropes

00:02:15 reaction, which is heterogeneously catalyzed.

00:02:18 Now if I could have the first slide, please.

00:02:21 Now you may say that that is comparing apples and oranges, because a homogeneous and a heterogeneous

00:02:29 reaction are quite different.

00:02:32 This is true, but anyway, the heterogeneous people have really been using the homogeneous

00:02:46 catalyst people's results for many years.

00:02:50 They've been looking at the transition metal complexes that have been made, and seeing

00:02:54 the ligands that one is able to make, and then saying that these ligands, that these

00:03:01 moieties exist on the surface of the heterogeneous catalyst.

00:03:06 What has happened really is that so many ligands have been made by the homogeneous people,

00:03:12 that there aren't enough heterogeneous reactions around to encompass and use all the ones that

00:03:18 have been found by the homogeneous people.

00:03:21 Nevertheless, in talking about these things, I want to point out the relationship, at least

00:03:27 to us and to me, between them.

00:03:31 We at the Bureau had a large, I think barrel a day, fissiotrope plant some time ago, and

00:03:38 in fact the South Africans, who now run SESOL, used to visit Brewston regularly to see this

00:03:43 plant.

00:03:45 Bob Anderson labored on the second floor running the heterogeneous fissiotropes reaction, and

00:03:51 Milt and I decided that we would study this reaction by choosing a homogeneous reaction

00:03:57 that was similar to it.

00:03:59 That is how we came to study the oxo-hydroformylation reaction.

00:04:07 It is very interesting to talk about how the hydroformylation reaction was discovered.

00:04:17 Chuck Kibbe has just mentioned that in 1929, he had the right date, Smith, Hawke, and Golden

00:04:25 added ethylene to the fissiotropes reaction, and they noticed an increase in the oxygenates

00:04:30 that were produced.

00:04:31 Unfortunately, or it really doesn't matter, they didn't follow it up.

00:04:37 Years later, during the war, Otto Roland, who was a student of Franz Fischer, was studying

00:04:44 the fissiotropes mechanism at Rochemy, and he read the results of Smith, Hawke, and Golden,

00:04:51 and decided to repeat them and got exactly the same results.

00:04:55 He then added ethylene in reasoning that if I, from Le Chatelier's principle, if I raise

00:05:02 the pressure, I was adding the gas, so I would get the probability of ethylene, CO, and hydrogen

00:05:10 combining would be greater.

00:05:12 He raised the pressure, and lo and behold, he obtained, upon adding large amounts of

00:05:19 ethylene at 50 atmospheres, essentially only oxygenated products.

00:05:25 What he had done was take a heterogeneous reaction, a fissiotropes reaction, which produced

00:05:31 hydrocarbons plus some oxygenates, and turned it into a reaction which produced no hydrocarbons

00:05:39 and only oxygenates.

00:05:42 This was quite a relationship between the two reactions.

00:05:47 By the way, they were working with a cobalt catalyst, and that's a very happy circumstance.

00:05:53 I said to myself over the years, isn't it lucky that Otto Roland was studying the cobalt

00:06:00 catalyzed FT reaction?

00:06:02 It turns out that inevitably, the oxo reaction was going to be discovered, and as I will

00:06:08 show you in a little while, it need not have been only cobalt.

00:06:14 The catalyst that is used down at SESOL today, of course, is iron, and I doubt if you would

00:06:20 have discovered the hydroformylation reaction with iron.

00:06:24 However, when you run ruthenium, you can get essentially exactly the same results.

00:06:32 You have under certain conditions, we have a temperature and pressure, we have a state

00:06:40 where the fissiotropes reaction and the hydroformylation can both occur under the same conditions.

00:06:47 This is what happens with cobalt, and it turns out that as Pickler and Fernhaber were able

00:06:54 to show later, under almost identical conditions, you can run a fissiotropes reaction with ruthenium

00:07:03 and get an oxo reaction going on at the same time.

00:07:07 So it would have been discovered when you started playing around with ruthenium.

00:07:13 There are certain odd things about the oxo reaction that have been mentioned.

00:07:18 One thing, methylmethacrylate, as Dr. Orchin mentioned, is hydroformylated very well.

00:07:26 It doesn't polymerize, and so evidently CO inhibits the polymerization.

00:07:31 It's not polymerized also.

00:07:37 What about the effect of sulfur?

00:07:39 Sulfur is a poison for the fissiotropes reaction, at least in significant amounts.

00:07:45 Sulfur does not poison the oxo reaction to a significant amount.

00:07:50 In fact, Dr. Orchin and I have a patent on the hydrogenation of thiophene to tetrahydrothiophene

00:07:56 using the cobalt hydrocarbonyl system.

00:08:00 In the long run, if you run olefin feeds with large amounts of sulfur, you gradually form

00:08:09 cobalt sulfides which are not converted to the carbonyl, so the rate slows down.

00:08:14 But it leads one to the interesting thought of what would happen to a fissiotropes reaction

00:08:19 if you add a little bit of olefin and a lot of sulfur.

00:08:23 You then might poison the fissiotropes reaction under slightly milder conditions and have

00:08:30 an oxo-type reaction go on.

00:08:39 This is a reaction that's sort of interesting, and I just sort of had to show it.

00:08:44 We postulated a long time ago that cobalt hydrocarbonyl, if you can call it so, was

00:08:52 the catalyst, and that, of course, deserves quotation marks.

00:08:56 We were determined to make the substance, which we learned how to do and put it into

00:09:03 inorganic syntheses.

00:09:07 Its decomposition has been studied.

00:09:10 If you make pure hydrocarbonyl, which is 9 molar, it decomposes by a second molar reaction

00:09:17 and so it will decompose very rapidly.

00:09:20 But if you put it in dilute solution, it has to meet another hydrocarbonyl and so it

00:09:25 sticks around for quite a while.

00:09:27 We chose hexene, one hexene, because when you read C&E news, in the very last page they

00:09:34 always have a little bit of trivia.

00:09:37 I read that one time, just before we ran this reaction, and it said that the human nose

00:09:43 could detect heptaldehyde in the minutest amounts.

00:09:48 In fact, of all the substances known, that was detected most easily.

00:09:52 I imagine with the progress of science, we could detect it in the nose, they do a little

00:09:56 better these days.

00:09:58 But we were worried about getting a significant amount, if this reaction went, of aldehyde,

00:10:05 and I took a whiff of heptaldehyde, and really, it's not bad, it's just very distinctive.

00:10:13 We ran this reaction, both in the presence of nitrogen and in the presence of CO, and

00:10:18 I must tell you that when the reaction was over in a few seconds, we knew we had heptaldehyde.

00:10:24 It was the first time we were able to show that in one molecule, cobalt hydrocarbonyl,

00:10:33 regardless of that very complicated mechanism, shown by Dr. Orchard, shown by Dr. Oswald,

00:10:40 as time goes on, the mechanism of both the Fischer-Tropsch and the hydroformylation get

00:10:46 more complicated.

00:10:47 Thank goodness the plants were already built.

00:10:50 Next slide, please.

00:10:55 The hydroformylation is a commercial reaction.

00:10:59 There are plants all over the world.

00:11:01 There are plants in India, Iran, Bulgaria, Romania, Australia, every industrialized country,

00:11:08 and many of the third world countries.

00:11:11 It's a very useful one, and a new plant's going up.

00:11:14 Most of the new plants, I think, in the future will probably be based on rhodium, which I'm

00:11:19 based on Roy Pruitt's work.

00:11:21 It's an energy-saving device, and there are certain advantages to rhodium, which I guess

00:11:29 time does not permit me to go into, but nevertheless, and the Fischer-Tropsch, of course, is also

00:11:36 commercial in South Africa.

00:11:40 We are studying here, I think, in this country, so that one of these days we will get a variation

00:11:47 of the Fischer-Tropsch reaction, which we hope will be commercial.

00:11:51 I really have only two or three points to make in my talk, and this slide and one or

00:11:58 two others that follow are one of the main points, certainly.

00:12:02 That is that the fact that we discovered the hydrocarbonate, that Rowland actually

00:12:09 discovered the oxo reaction with cobalt, is sort of an accident.

00:12:15 In our very first paper that we referred to, we did say that it would be very interesting

00:12:22 to study rhodium, iridium, but I think unconsciously we felt, as Val Hansel must have felt once,

00:12:30 that this was a very precious metal and one doesn't fool around with rhodium, and so we

00:12:34 never tried it.

00:12:36 But naturally rhodium works, iridium works, and if you look on the board you see that

00:12:44 I have iron on there, and actually, essentially every element on that board works.

00:12:51 You can't get a type of hydroformylation with every element.

00:12:56 Now how is that possible?

00:13:01 The publications, for instance, there are some publications lately using platinum as

00:13:09 a catalyst.

00:13:11 Some weird ligands, in this case with platinum, people tend to use tin chloride, as Nifton

00:13:17 has and as Milk Orchard has, but nevertheless you get a very good hydroformylation.

00:13:22 There are reports in the patent literature that cobalt and even silver are catalysts,

00:13:28 and what you really have to do is get a hydrogen to get onto the metal, and if possible a CO.

00:13:37 If you can get a hydrocarbon yield out of these elements, you will get a type of hydroformylation

00:13:44 reaction.

00:13:45 So since you can, in this case, and molybdenum and tungsten also, as I will show you in a

00:13:53 moment I think, also can be catalysts, it turns out that what I'm saying for the oxo

00:14:01 reaction that all these elements can be hydroformylation catalysts more or less holds true for the

00:14:09 Fischer-Stokes reaction.

00:14:11 If I can have the next slide, please.

00:14:16 Now just switching over to the Fischer-Stokes reaction, this is a well-known table that

00:14:25 was put out by Brodin et al. some time ago, and he pointed out that these dashed lines

00:14:31 and these solid lines don't really mean as much as they look, as I will try to elaborate

00:14:37 in a little while.

00:14:39 What they were saying was that those elements to the left of that first dashed line dissociate

00:14:45 CO at room temperature.

00:14:49 One of the reasons that we never, we postulated, as Chuck Kibbe just said, that the oxo proceeded

00:14:54 by an insertion of CO is that CO is a diatomic molecule with the strongest bond energy, and

00:15:02 it seemed incomprehensible, frankly, to me at the time, that a reaction was going to

00:15:08 proceed by complete dissociation of the molecule.

00:15:12 But now, of course, when you realize the immense energy that is present in the surface of one

00:15:17 of these metals, it doesn't seem that incomprehensible at all.

00:15:23 But nevertheless, as you proceed to the right, you find that rhenium, osmium, ruthenium,

00:15:30 sodium, cobalt, all will dissociate CO to CH, to carbide, if you like it, CH, CH2, CH3

00:15:43 entities at synthesis temperatures, between 200 and 300 degrees, usually.

00:15:49 Now the elements to the right do not dissociate, iridium, platinum, palladium, and copper do

00:15:55 not dissociate CO at 300 degrees.

00:15:58 And as Rabot and Poutsman have shown, you can use this synthesis to these elements to

00:16:05 make methanol.

00:16:06 And I'm sure that most of you know they have published several papers in which they've

00:16:10 shown that the order is palladium is better than platinum, is better than iridium, and

00:16:16 under, call them FT conditions, these are methanol catalysts.

00:16:22 Of course, the methanol catalyst happens to be copper.

00:16:26 And copper and zinc oxide is the ICI methanol catalyst, and they sell about 80% of all

00:16:33 the methanol catalysts in the world.

00:16:35 And I just want to point out, it's a little bit amazing to me that the one thing you must

00:16:39 do to copper is never to add any alkali.

00:16:43 If you add a drop or a touch of alkali to the ICI or the methanol catalyst, you start

00:16:50 making ethanol and propanol.

00:16:53 And you have to think a little bit about that.

00:16:55 And we may or may not get back to that.

00:16:58 But this is sort of the way things are.

00:17:01 And what I really want to say that in spite of those lines, Dow, for instance, finds that

00:17:10 molybdenum have published papers showing that molybdenum is a rather good, in fact you can

00:17:17 use the word excellent, fissure charge catalyst.

00:17:19 In the presence of alkali, they've shown that you can only get, that you can restrict the

00:17:24 product mostly to C1, C5 hydrocarbons.

00:17:28 And manganese is a very good case in point.

00:17:32 I think it was Bob Anderson, who's sitting over there, who in his first book, which is

00:17:37 still the book, Fischer-Tropsch Reaction, and if Bob doesn't mind, I will tell you that

00:17:44 I think he's finishing another book at the moment, Bob in his first book I think said

00:17:50 that manganese was not a Fischer-Tropsch catalyst.

00:17:54 I happened to visit Colbell in Berlin, and he looked, that subject came up.

00:18:01 And he said, manganese is a Fischer-Tropsch catalyst.

00:18:04 And I can prove it to you.

00:18:07 You don't exactly prove things to people by yelling at them.

00:18:10 But nonetheless, he said that he had an iron manganese catalyst that he had used at Ryan

00:18:17 in person, and that he was using now in his laboratory, and that this particular catalyst

00:18:27 was an excellent one for producing C2, C4 olefins.

00:18:31 I since then conducted my own private canvass as to whether manganese is a catalyst.

00:18:36 That's about 50-50.

00:18:38 And most, with the ones that don't believe it is, believing that it may be, since manganese

00:18:44 oxidizes very easily, it may be a support.

00:18:48 Coming back again to the oxo reaction, while we were at the Bureau of Mines, we took manganese

00:18:56 hydrocarbonyl, which is a very stable substance, much more stable than cobalt hydrocarbonyl,

00:19:02 and tested this as an oxo catalyst, and it is a very good oxo catalyst.

00:19:07 So as long as we were able to make HMNCO5, we knew it was going to be a catalyst.

00:19:15 The only trouble with that catalyst was that after several turnovers, it suddenly slowed

00:19:22 down and ceased to be a catalyst.

00:19:24 It was catalytic to a certain extent, and we found that manganese, which is oxophilic,

00:19:30 likes to combine with oxygen, and you were making aldehydes, and so the manganese was

00:19:34 oxidizing, and you were destroying your catalyst.

00:19:38 So that was the story there, and there may be some relationship between that and what

00:19:43 is happening in the Fischer-Tropsch reaction.

00:19:48 May I have the next slide, please?

00:19:54 The hydroformylation reaction is one of those things that obviously has not remained still.

00:20:01 It was originally cobalt, and I might say, while rhodium is doing very well, Roy can

00:20:09 correct this, but probably 90% of the oxo capacity in the world is still based on cobalt.

00:20:16 By the way, some 10 billion pounds of products are made via the oxo reaction.

00:20:22 So you see what we're talking about.

00:20:27 I won't read all that to you.

00:20:29 I think you can see what is there.

00:20:33 The temperature with rhodium obviously is lower.

00:20:38 The pressure with rhodium obviously is much lower.

00:20:41 The volume of olefins per hour per reactor volume, LHSV, is higher with cobalt, meaning

00:20:51 you might have to use a larger reactor with rhodium.

00:20:55 The hydrocarbon formation is high with the shell catalyst.

00:20:59 Shell discovered that if you made the tributylphosphine ligand of hydrocarbonyl, you stabilized it.

00:21:10 When I say stabilized it, cobalt carbonyl decomposes at about 200 degrees.

00:21:15 You examine the hydroformylation literature, what happens in many of the papers is that

00:21:21 people have run hydroformylation reactions at 160 or 190 or even 200 degrees and not

00:21:30 realized that the CO reaction is so fast that the reactions become diffusion controlled

00:21:36 and you're using up your CO much more rapidly than it can be replaced.

00:21:44 But when you add the phosphine ligand, then you can lower the pressure because your carbonyl

00:21:50 is stable.

00:21:52 That HCOCO3PBO3 is much more stable so you're able to lower the temperature.

00:22:00 There is some hydrogenation to the hydrocarbons, which is not very great by the way.

00:22:07 There is also some hydrogenation to alcohols, which is greatest with the shell catalyst.

00:22:15 Rhodium is great if you want to make aldehydes.

00:22:17 Rhodium has some very marvelous properties in the sense that if you take a molecule like

00:22:23 butadiene with cobalt, you will hydrogenate one double bond and hydroformylate the other.

00:22:30 Rhodium tends to give you aldehydes without much hydrogenation.

00:22:34 If you take butadiene, you don't get a very good yield of a dipic aldehyde, straight chain,

00:22:42 but you can get a 50% yield where two CHOs add on.

00:22:47 Some of it will be straight and some of it will have one branched formal group on it.

00:22:54 Another case is styrene.

00:22:56 With cobalt, styrene hydrogenates to ethylbenzene to a large extent.

00:23:01 With rhodium it doesn't, you get aldehydes.

00:23:05 You have to look at this table and see that the hydroformylation reaction is one that

00:23:12 is constantly improving.

00:23:13 It will probably improve more.

00:23:15 I don't exactly know why you couldn't build another plant based on ruthenium actually

00:23:21 if the price of ruthenium came down, and I think it is down right now by the way.

00:23:25 There's an excess of ruthenium.

00:23:28 The next slide perhaps may show that.

00:23:33 You can run a hydroformylation reaction with ruthenium carbonyl, and it has been done by

00:23:42 Schultz.

00:23:43 That's the Schultz of Schultz and Pickler.

00:23:44 We'll come to Pittsburgh next week, we'll talk about that.

00:23:49 As the bottom line says that certain ruthenium catalysts are effective for both deoxo and

00:23:54 FT.

00:23:55 You can't read it.

00:23:57 Polymethylene synthesis in the same reaction conditions.

00:23:59 That's kind of an interesting point.

00:24:02 Pickler was making polyethylene with ruthenium.

00:24:06 It's a very active catalyst, and when you say polymethylene, you're really saying the

00:24:15 same as polyethylene, and I think there was some patent troubles, not troubles, but differences

00:24:26 between Pickler and some people who were making polyethylene in this country because he claimed

00:24:30 he made it this way.

00:24:34 Just as the hydroformylation reaction is changing, the Fischer-Tropsch reaction is changing because

00:24:39 you can take ruthenium as a carbonyl and lay it down on a support, and if you add a little

00:24:45 bit of alkali to that, and this has been published, you will make C2, C4 olefins in 70 to 80%

00:24:53 yield.

00:24:54 Here we have a reaction that if you read literature is excellent for making very long chain hydrocarbons,

00:25:01 and yet today you can, by playing around with this element, change its product distribution

00:25:08 completely.

00:25:09 It would be interesting.

00:25:11 People are interested in making lower alcohol, C2, C6 alcohols as fuels to use the Fischer-Tropsch

00:25:20 to make C2, C4 olefins, and then take those and put them through a hydroformylation reaction

00:25:27 and run and make alcohols that way.

00:25:30 You have to think about that a little bit, but it's not a bad idea.

00:25:34 Next slide, please.

00:25:36 I just wanted to show, if you could focus that again, that platinum complexes are also

00:25:45 hydroformylation catalysts, and in this particular one, I think a tin chloride ligand was also

00:25:54 used.

00:25:56 In this case, whether you have a neutral compound or you have PtSnCl3, in other words, an ionic

00:26:04 compound.

00:26:06 Next slide, please.

00:26:10 This is the place where you come to iron and chromium and tungsten and molybdenum and all

00:26:17 the things that were not hydroformylation catalysts, but which really are.

00:26:22 The reason that iron is not a catalyst is because if you treat iron carbonyl with hydrogen

00:26:29 at very high pressures and at very high temperatures, you'll get just a trace of a hydrocarbonyl,

00:26:35 if any at all.

00:26:38 There are reports that it is a catalyst in the literature, and that's usually due to

00:26:41 the fact that the fellow uses autoclave and once had some cobalt in it.

00:26:47 Reppy and Vetter discovered that if, instead of using hydrogen, they took an olefin and

00:26:52 CO in water, they were able to get very high yields of aldehydes and alcohols.

00:26:59 It may surprise you, but the Japanese, who were very inventive, looked at that reaction

00:27:08 and built a plant based on iron, not on cobalt, and ran that plant for some 20-odd years.

00:27:16 One of the reasons they picked that is because the normal people who run the hydroformylation

00:27:21 reactions are always interested in getting the normal product as against the isoproduct.

00:27:27 Since this reaction gives a very high ratio of normal iso, they like this, and I think

00:27:31 the plant's shut down now, but it did run for a long time.

00:27:35 Once you, and I'll let you read the slide, the essential point that it's easier to form

00:27:39 the hydrocarbonyl, once you form the hydrocarbonyl, the oxo reaction proceeds well.

00:27:45 Osmium does the same thing.

00:27:48 In fact, chromium does exactly the same thing, and molybdenum, and tungsten, and all of these

00:27:54 in the presence of a little base give you a hydrocarbonyl.

00:27:58 Once you get that, you get a hydroformylation reaction.

00:28:01 So we're not restricted.

00:28:02 I'm trying to say we're not restricted in either reaction to any particular element.

00:28:09 It is up to you to find out the best conditions in which you want to run it.

00:28:15 The Japanese is sort of a good example.

00:28:17 They did not use the cobalt at all.

00:28:21 With the advent of making fischer-tropsch catalysts based on carbonyls, which are deposited,

00:28:30 and by ion exchanging metals onto supports, one can now make finely divided catalysts,

00:28:38 and we may show one where you have a catalyst with no more than 10 atoms on a support where

00:28:46 you have totally different effects than you have with even a crystallite, which may have

00:28:50 as many as 500 atoms.

00:28:55 The next slide, please.

00:28:58 This very quickly, I just picked a few.

00:29:01 These are hydroformylation reactions with propylene, CO, and water, and that list could

00:29:07 easily be doubled, and I'll let you read it for yourself, but you see that you make

00:29:13 a good yield of C4 aldehyde next to the last column over C4 alcohol.

00:29:19 With the osmium compound, you essentially get all normal compounds, and since you get

00:29:25 H2FeCO4, and the reaction involves the formation, the reaction really involves, as I think you've

00:29:33 probably figured out, the attack by base on the carbonyl, giving you CO2.

00:29:40 So you get a molecule of CO2 and the H2FeCO4, and this essentially, if the H2CO4 then loses

00:29:48 this hydrogen, giving you FeCO4 minus minus, you will have a water-gas-shift reaction.

00:29:55 So all of these reactions are more or less also water-gas-shift reactions.

00:30:02 The next slide, please.

00:30:06 We studied the rates of hydroformylation of various olefins with cobalt, and it turns

00:30:12 out that the results that were later studied by Marko were essentially almost exactly the

00:30:18 same.

00:30:19 And rather than give you the table, I just want to show you that terminal olefins react

00:30:26 most rapidly, three times as fast as internal olefins.

00:30:30 That is, start putting a chain, you start putting a branch at the terminal olefin,

00:30:34 the rate goes down quite a bit, and if you start putting a branch and then making it

00:30:39 internal again, the rate will still go down.

00:30:42 And in the Fischer-Tropsch reaction, if you get any such molecule other than the terminal

00:30:47 olefin, the chances of it incorporating into the chain get lower and lower.

00:30:54 As to the cyclic olefins, it's easier to figure out, of course, why cyclopentene, it's more

00:31:01 strained, it doesn't have the tetrahedral angle, why cyclohexene is much closer to it,

00:31:06 and we could predict that.

00:31:08 And so I say this holds for both catalysts, and it will probably hold for all the oxo-catalysts

00:31:14 that you make.

00:31:15 Next slide, please.

00:31:19 This is, just dug this out because I wanted to show you that the last thing, group 1A

00:31:27 I think it was called, contains zirconium and is an unlikely oxo-catalyst.

00:31:33 Nevertheless, if you take the di-cyclopentadienyl zirconium, I'll call it hydrochloride, and

00:31:43 treat it with either, that's hexene, 1-hexene, or that looks like 3-hexene, cis and trans.

00:31:58 You get an insertion of CO, now there's no hydrocarbonyl involved here, and that means

00:32:05 that CO was picked up, but you never isolated the hydrocarbonyl, you don't have to isolate

00:32:10 it.

00:32:11 But CO was picked up, and all metals have this terrible desire when they have an R group

00:32:19 attached to it and a CO, for a migratory insertion reaction to occur, in other words, for CO

00:32:26 to insert between the metal and R. And we all know, of course, that it's not an insertion,

00:32:31 that the R group goes over with its pair of electrons to the CO, but nevertheless, migratory

00:32:37 insertion is the word that has come about.

00:32:39 So here we have a case of an element that should not, it's not catalytic by the way,

00:32:46 this is stoichiometric, you fool around with this, you make it catalytic, but the amazing

00:32:53 thing about this is that no matter where the double bond is, you only get the terminal

00:32:58 aldehyde.

00:32:59 So this is a case where it's an unusual hydroformylation, you get a result that people have been looking

00:33:07 for, you get the terminal, it's a straight chain aldehyde only, and if someone wants

00:33:13 to go back to the laboratory and play around with this, I think you might get some unusual

00:33:17 results.

00:33:18 The next slide, please.

00:33:22 This I, no matter what you say about the oxo reaction, and you can postulate all kinds

00:33:30 of mechanisms, what I've simply done is saying that they all have to sort of go through this.

00:33:38 I have it for cobalt, but it's really true for all of them.

00:33:40 You have to get a molecule that's coordinatively unsaturated, and in this case, we get HCOCO3,

00:33:48 and in the paper that's just been published about a year ago by Milt, by using an argon

00:33:53 matrix, he was able to pretty much spectroscopically identify the presence of that HCOCO3.

00:34:04 So it's calculated, by the way, that under oxo conditions with cobalt, you should have

00:34:10 about perhaps three-tenths of one percent of that molecule, which is perhaps a little

00:34:15 too much for an intermediate, it should be less.

00:34:19 Then the next thing that happens, as already, and I'm going to go over this quickly because

00:34:23 Dr. Orchard really showed the same sort of slide, you coordinate the olefin with the

00:34:30 coordinatively unsaturated carbonyl, and next slide, then you can get the formation

00:34:46 of either the straight chain, alkyl, cobalt carbonyl, or the branch chain, depending,

00:34:53 once you have the cobalt bonded to the double bond, hydrogen can add Markovnikov-wise or

00:35:00 anti-Markovnikov-wise, it essentially does both under most conditions, and you get the

00:35:06 normal RCH2, the top one over the bottom, and then in the presence of CO, you have again

00:35:11 the migratory insertion reaction, so you get the acyl, cobalt, carbonyl, and next slide,

00:35:21 we have, and then we come to a little difference of opinion, which is kind of important, with

00:35:28 the Heck and Breslow originally postulated this mechanism, I think a lot of other people

00:35:33 perhaps helped them along, but they really put it all together, and they postulated that

00:35:39 step four was the way hydrogen added to this, you had a coordinatively unsaturated molecule,

00:35:47 CO usually has nine electrons around it, and it only has seven in the first equation,

00:35:53 so hydrogen adds, and then by reductive elimination, you get the aldehyde, however, maybe the next

00:36:01 slide will show what happens, I think that what really happens is this, with cobalt at

00:36:08 least, you don't get hydrogen adding in the last step, I think you do get hydrogen, oxidative

00:36:13 addition of hydrogen with rhodium, but I do not think you get it with cobalt, and what

00:36:19 they did was show that by using a high pressure infrared cell, they took cobalt carbonyl and

00:36:26 hydrogen and studied that equilibrium, and they found that at equilibrium, you should

00:36:34 expect to have about 90% or 92% of hydrocarbonyl present, and a small amount of cobalt carbonyl

00:36:42 present, and this was reached after about an hour and a half, well, when that was reached,

00:36:47 looking at their cell under reaction conditions, they then injected a terminal olefin, and they

00:36:54 noticed that the hydrocarbonyl disappeared, and the cobalt carbonyl concentration went way up,

00:36:59 went up past its equilibrium concentration, which means that it was being used up, and the cobalt

00:37:07 carbonyl concentration, dicobalt octacarbonyl, was being formed in very large amounts, and the

00:37:13 only way to account for that is to say that the termination step is a bimolecular step,

00:37:18 where hydrocarbonyl reacts with the acyl cobalt carbonyl in this manner, and that is why you get

00:37:24 this large, suddenly great influx of, in the reaction vessel of cobalt carbonyl. Now, if we

00:37:31 go to the next slide, we want to get back to the homogeneous and heterogeneous comparison,

00:37:43 if you wish. In the hydroformylation type of reaction, you have a very reactive species

00:37:50 running around, which is the cobalt hydrocarbonyl or a rhodium hydrocarbonyl, and in this case,

00:37:59 it reacts with an acyl cobalt metal, cobalt carbonyl compound very easily. However, when

00:38:11 you form CO2-CO8, and that is not a compound M2-CO2, I mean M2 or CO2-CO8, CO2-CO8 is

00:38:19 coordinatively saturated. There's nothing it can do, and so what happens is the aldehyde comes off

00:38:26 the carbonyl, and that's your final product. In a heterogeneous system, you don't have a

00:38:34 mobile hydrocarbonyl species. You have hydrogen that is attached to a metal atom, and you have

00:38:42 an acyl complex that is attached to a metal atom, and it's likely then that the only way that this

00:38:49 can happen in a heterogeneous system, and it does happen in a heterogeneous system because it happens

00:38:53 in the Fischer-Tropsch reaction, is for the oxidation of addition of hydrogen to occur.

00:38:59 The next slide, please. There are some other implications of this binuclear hydroformylation

00:39:09 reaction, and that people, it's very interesting that in studying, in that very first slide where

00:39:15 I had the oxo reaction and the Fischer-Tropsch reaction, I find there are two classes of people.

00:39:21 Those who are taking the homogeneous hydroformylation reaction and trying to turn it

00:39:27 into a heterogeneous reaction, and on the other hand, there are the heterogeneous Fischer-Tropsch

00:39:31 reaction and trying their best to turn that into a homogeneous reaction, and they're both for

00:39:37 good reasons. And yet, if you really want to go back to it, the first group had their work

00:39:45 already done in a way, because the first reactor that was built by the Germans, who did not realize,

00:39:51 Roland at the very beginning did not realize that he had a homogeneous reaction, and so Roche

00:39:57 Rukmi built a Fischer-Tropsch type reactor. They used a cobalt, thorium, magnesium, and kiesel

00:40:05 reactor as their catalyst, and they ran this, and they were getting, under correct conditions,

00:40:12 oxo products only. And Natta and others studied the kinetics of this reaction, and they were

00:40:19 getting a homogeneous reaction, running it in a fixed-bit heterogeneous system. And so, the people

00:40:26 who were trying to do this, I want to say it's already been done, and the people have decided

00:40:31 that they better go over to the homogeneous way. The other second sentence is very

00:40:37 interesting, because if a bimolecular reaction is possible, then it might be possible to have

00:40:44 the two reactions carried out with two different metals. And I think in this very room, or the room

00:40:50 next door, I think the people of Delaware, Gates et al., reported that ruthenium and osmium give

00:40:58 different products. In other words, they were running Fischer-Tropsch reactions using

00:41:07 carbonyls of ruthenium and osmium, and they found that if they upped the osmium concentration, the

00:41:14 amount of oxygenates increased a great deal. And this is also a subject for further study. Next

00:41:21 slide, please. This is very quickly to show you that this is a bimolecular reaction where you get

00:41:31 essentially hydroformylation. The metal is molybdenum. This was done by Bergman at LBL. And the

00:41:38 other interesting part of it is that in this case, you don't even need an olefin. Here he had CH3,

00:41:46 which was attached. And this cannot be derived from an olefin. And you were able to get acid

00:41:52 aldehyde via this bimolecular reaction. And then the next slide, please, which shows, I'll let you

00:42:00 follow the right-hand side down, because it is, it really follows what we've said. But very

00:42:05 interestingly enough, the oxal reaction is pretty much called so, because in ethylene, it gives you

00:42:11 not only aldehydes, but some ketones. You don't get ketones in higher olefins. To any great extent,

00:42:18 you can't get them, certainly. And so the oxo came from that fact that ketones were formed. And here,

00:42:24 using this molybdenum compound, you see Bergman found that if he added ethylene to it, he inserted

00:42:31 ethylene into the molybdenum-ACO bond, giving the lower left-hand complex. And then this went over

00:42:39 to RCO, and then CH2 to molybdenum, which finally was decomposed to the ketone compound. So again,

00:42:49 it just shows you that this bimolecular reaction, by the way, is very common, many examples in

00:42:54 literature. And here's a case where you get ketones, if you like. And I assure you that you

00:43:00 could take molybdenum and another metal, make the complex that you want, and get the products

00:43:05 that you would like. Next slide, please. This is the, Chuck was just talking about the mechanism

00:43:14 of the Fischer-Tropsch reaction. And there's a great deal of evidence by very, very good people,

00:43:19 going from Somagi, to Bell, to Balloon, to Sackler, to Ponetz, to a great many Van Nuys, that the very

00:43:29 first thing that happens in the Fischer-Tropsch reaction is that you dissociate CO, and that you

00:43:34 get, and this is, I'll let you follow the mechanism on the board, I will not go over it, you don't

00:43:40 really get carbides out of it, you get MCHX, which I don't call a carbide, I'm sorry about that. And

00:43:46 then that, if you follow the work of Raleigh Pettit, who very sadly is no longer with us,

00:43:52 has done some marvelous work in this by just decomposing diazomethane on Fischer-Tropsch-type

00:44:00 catalysts, and showing that nothing much happened until he added a touch of hydrogen, and that

00:44:04 hydrogen was necessary to make that CH3 group. Once you got that CH3 group, he then got a

00:44:10 polymerization that gave him the Anderson type of product distribution in the Fischer-Tropsch.

00:44:20 And the funny part is, I believe that mechanism, although I was the one, along with others,

00:44:27 who postulated that CO insertion can happen. I believe it, next slide please, I'll just finish

00:44:35 this and then we'll talk about it. You terminate this reaction, the Fischer-Tropsch reaction,

00:44:39 in essentially three ways. You have a higher hydrogen concentration, or you don't even need

00:44:45 that, but anyway, you can get hydrocarbons simply by hydrogenating off that RCH2. You can get

00:44:50 olefin formation by simply beta hydride elimination, a very well-known reaction. And of course, here's

00:44:57 our oxo reaction, there's one bond missing between M and C. And here we find that CO insertion

00:45:03 occurs during the Fischer-Tropsch reaction, but people refer to it as a termination reaction.

00:45:07 Okay, and it probably reacts just that way, it is a termination reaction. But, next slide please,

00:45:15 maybe we'll come to that. I come back to this, and what I want to say is that these metals,

00:45:23 in spite of the lines that are drawn there, depending on conditions, can, we say, for

00:45:30 instance, that ruthenium and cobalt and rhodium dissociate CO at synthesis temperatures. Well,

00:45:37 they do dissociate CO. They also adsorb CO associatively, and there's a balance between

00:45:45 these two. And it depends on temperature and pressure and the dispersion of the metal on

00:45:52 the particular support that you have, and that you can get probably both mechanisms to operate.

00:46:00 In other words, you can get the, I'll call it the original carbide mechanism that is in vogue now,

00:46:07 and you can also get a mechanism whereby CO inserts and hydrogenates. And that what usually

00:46:14 happens is that both exist on the surface at the same time in varying amounts, and it depends on

00:46:20 you to determine how much of each will exist as such. Rhodium is particular of interest because

00:46:31 it can go either way. With rhodium, people in carbide have found that you can make C2 and

00:46:37 higher oxygenated compounds. Rhodium, of course, also is an excellent fisiotropic catalyst,

00:46:42 and you can make, if you wish, only hydrocarbons. If I can have the next slide.

00:46:51 I think I'll just let you read that because my little bell went off here. The next slide, please.

00:47:01 Well, what Katsura has done lately is by an ion exchange, an ion exchange technique,

00:47:08 he has put rhodium on alumina, silica, cerium oxide, magnesium, magnesium oxide,

00:47:16 magnesium oxide, and titanium oxide. And he has been able to show that he is getting very small

00:47:29 10-atom metal and with rhodium metal clusters, if you'd like to call it, dispersed. Now,

00:47:38 a crystallite normally is very big. It contains maybe 500 atoms, but when you get down to

00:47:45 something like 10 atoms, every atom is essentially an edge. It's even smaller than a single crystal.

00:47:52 So, in the next slide, I think, he was able to show with this finely dispersed supported rhodium,

00:48:01 less than 10 atoms per cluster, he was able to get two CO molecules to absorb a CO atom. Now,

00:48:07 that's amazing because on a single crystal, the most you can get is 0.7 CO atoms absorbed

00:48:14 on a metal. I don't know if heterogeneous people here are willing to fight about that.

00:48:18 Nevertheless, spectroscopically, he was able to show that he had essentially no linear

00:48:24 trace abridged and he had two COs. He had this dicarbonyl species. And if I can read that,

00:48:33 because it's worthwhile reading, rhodium, which lies between ruthenium, which produces only

00:48:38 hydrocarbons, and palladium, which produces only methanol, and even there you have to put in a

00:48:46 little catchphrase. If you raise the temperature even with palladium, platinum, and iridium,

00:48:52 you will get Fischer-Tropsch type of reactions. You still have to be careful even there.

00:48:57 Nevertheless, reading on, and palladium, which produces only methanol,

00:49:01 catalyzed formation of both alcohols and hydrocarbons. Small effects markedly alter

00:49:07 its selectivity. Thus, rhodium, as a bulk metal, yields mainly hydrocarbons. There's very small

00:49:13 clusters on acidic supports. Rhodium still only gives you hydrocarbons. But on basic supports,

00:49:20 and that means that you're pushing electrons into the rhodium. And if you remember that table I had

00:49:25 on the board, if you push electrons into rhodium, you're making palladium out of it because palladium

00:49:31 has more d-electrons than rhodium has. So if you add potassium, or if you have a basic support,

00:49:39 you're changing the nature of the rhodium and you're pushing it over to the right-hand side

00:49:43 of the table, and then it becomes its selectivity to alcohols. C2 alcohols and C3 alcohols is

00:49:50 enhanced. Now you make methanol this way too, and Katzer believes that methanol is not an

00:49:58 intermediate in the formation of higher alcohols. I don't believe it is either. One of the reasons

00:50:03 for this is because if you look at the higher alcohols that are formed, you get the regular

00:50:08 Anderson plot. In other words, CH2 polymerizes in a way in which you would not, it undergoes

00:50:18 a polymerization type of reaction in which you would not expect if methanol were the intermediate.

00:50:26 I think I'll let you read this. Understanding and control of CO insertion into the R-M bond

00:50:39 is essentially a key for determining the product distribution in the F-T reaction. Most of the CO

00:50:45 adsorbed in the catalyst. Well, I have to go on a little bit. I have to go on to the next slide

00:50:50 because that has to be the... What this and the next slide says is that if you run reactions

00:50:59 at high temperatures on both fish-soaked metals, you will get hydrocarbons. If you lower the

00:51:06 temperature, you'll start getting oxygenated products. What this means is that you have

00:51:12 CO dissociation and CO adsorbed associatively. On the very next slide, there are just simply

00:51:18 two more of the same. If you look through the literature, you'll find many others of the same.

00:51:23 If you lower the temperature, you have both dissociative and associative adsorption on F-T

00:51:29 catalysts. I think that's enough. I thank you very much for staying so late and missing that

00:51:34 lovely sunshine outside. Thank you very much.