Carboranylcarbenes

Russian Chemical Bulletin, Vol. 42, No. 4, April, 1993 593

Reviews

Carboranylcarbenes*

M.Jones, Jr.

Department of Chemistry, Princeton University, Princeton, New Jersey, 08544, USA. Fax: (609) 258 2383

The restflts of studies of carboranylcarbenes and boron-substituted carboranylcarbenes are reviewed. Their structure and reactivity in addition to alkenes, insertion into carbon- hydrogen bonds, and intramolecular reactions are compared.

Key words: carboranes, carboranylcarbenes, singlet and triplet states, photolysis, diazo compounds, cycloaddition, insertion, stereochemistry.

It is a great pleasure for a newcomer to the field of carborane chemis t ry to contr ibute to this volume! Much of the past and present work in carborane chemistry is Russian, and m y group at Pr ince ton has benef i ted very greatly from the p ioneer ing and current research of Russian chemists . I very much appreciate the chance to describe here our work in this area, and, I hope, to return some smal l por t ion of what we have learned from work in Russia.

I will descr ibe here react ions and some spectroscopy of carboranylcarbenes . In te rmolecula r chemistry of three differently subst i tuted carboranylcarbenes will be the pr imary focus of the discussion of reactions, al though some in t ramolecu la r chemis t ry will be ment ioned as well.**

Introduction to carboranes. Three-center , two-e lec - t ron bonding can be a fine way to hold a toms together . l The icosahedra l carboranes, C2BIoH~2 , come in three patterns: ortho, in which the two carbons are adjacent,

*This paper is based on a talk given at the Fifth Conference on Carbene Chemistry, Moscow, September, 1992. **Our research at Princeton has been supported by the Na- tional Science Foundation and the Petroleum Research Fund.

meta, in which the two carbons are in a 1,3 relat ionship, and para, in which the two carbons lie at opposi te vertices of the icosahedron. In these molecules a set of thir teen bonding molecu la r orbitals accomoda tes a total of 26 f ramework electrons.***

ortho-carborane meta-carborane para-oarborane

This bonding si tuat ion should be reminiscent of that in benzene in which six electrons fully occupy a set of three bonding ~ molecu la r orbitals. In both cases excep- t ional t he rmodynamic stabili ty is a ccompan ied by an imperat ive to preserve the s tab le system. In benzene chemist ry this t ranslates into <<Preserve the aromat ic sextet!,> and in carborane chemist ry to, <<Preserve the 12-vertex cage!,>. The reactivi ty of the two species is

***Hereinafter full dots refer to carbon atoms, other vertices are occupied by boron atoms.

Translated from Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 4, pp. 637--645, April, 1993.

1066-5285/93/4204-0593 $12.50 �9 1994 Plenum Publishing Corporation

594 Russ. Chem.Bull., VoL 42, No. 4, April, 1993 Jones

similar in some aspects, and is dominated by classical aromatic substitution reactions in both cases. 1,2

Br 2 P

FeBr 3

Br

AICI 3

The two species do differ in reactivity at a carbon adjacent to the ring or cage. In the two-dimensional, classically aromatic molecules, the reactivity of the benzyl position is enhanced, and is dependent upon interaction of an orbital at the benzyl position with ring molecular orbitals. In the cage molecules, overlap between an exo- cage orbital and the orbitals of the cage is very weak, and there is no special reactivity associated with the "benzyl" position. As we shall see, some of the best evidence for the lack of orbital connectivity comes from the EPR spectra of triplet carboranylcarbenes.

1-o-Carboranylcarbene. Phenylcarbene might well serve as the prototypal carbene, less reactive than the voraciously electrophilic methylene, but still reactive enough to exemplify the properties that have made carbenes so attractive to some of us. Much of carbene chemistry can be encapsulated in the single diagram:

k~

Sin g let .~-..-.---~ Trip let

,Lks k_, ~k t

Products Different Products

Like many carbenes, phenylcarbene has two reactive spin states fairly close in energy. The triplet is the ground state 3 but is essentially undetectable in chemical reactions in fluid solution. Even if one enters the equilibrium from the triplet side through photosensitized decomposition of phenyl diazomethane, it is the singlet carbene that is the reactive species. The problem is that even though the equilibrium between the two spin states must favor the ground state triplet, the rates of reaction of the singlet are many orders of magnitude greater ichan those of the triplet.

The triplet carbene can be detected when it is immo- bilized at low temperature. The triplet is bent, with an HCPh angle of 140--155 ~ The zero field splitting parameter [D/hc[ (0.5098 cm -~) shows that one electron is substantially delocalized over the benzene ring, as would be expected for an electron in what is essentially a benzyl position. 3

Our first work in carborane chemistry was to generate a carbon-substituted carboranylcarbene and make the obvious comparison to the classical phenylcarbene intermediate. 4 One problem is how to make the precur- sor carboranyl diazo compound. The CH bonds of carboranes are reasonably acidic (pK a -21), and this offers an opportunity for manipulation. Our route to t -d iazomethyl -o-carborane (1) and 1-methyt-2- (diazomethyl)-o-carborane (2) takes advantage of the acidity and availability of 1-vinyl-o-carborane. Ozonolysis to the aldehyde is followed by tosylhydrazone formation, generation of the sodium salt, and heating at about 80~ to generate the diazo compounds. 4

~ BuLi = Li __ H31 ~ 3 ~

2

/ ~ I'TsNHN H 2 , / ~ 1"0 3 ~ . ~ 2.Nail ~ ; ~

R = H , CH 3 R = H , CH 3 1 : R = H 2 : R = CH 3

R. S. Hutton and H. D. Roth of Bell Laboratories, together with S. Chaff of our group at Princeton used irradiation of the diazo compounds 1 and 2 in frozen solutions of methyltetrahydrofuran or propylene carbon- ate at 5 K to generate triplet EPR spectra for carbenes 3 and 4. 5 It was also determined that 3 was very likely the ground state species. The zero-field splitting parameters [D/hc[ and IE/hc] could be extracted from the spectra. In contrast to phenylcarbene, the [D/hc] values (3: 0.6860, 4: 0.6920) showed that little delocalization into the cage was possible. This is strong evidence that the frontier orbitals of the cage are poorly connected to external free valence. Moreover, no evidence for hydrogen abstraction by 4 to give 5 could be seen. However, just because the triplet is the ground state of 1-o-carboranylcarbenes does not mean that the triplet will be the more reactive state. It remains for an examination of chemical reactivity to tell us whether the singlet or triplet is the more reactive species.

! : R = H 3 : R = H 2 : R =CH 3 4 : R = C H 3 5

Carboranylcarbenes Russ.Chem.Bull., Vol. 42, No. 4, April, 1993 595

The traditional test for the reacting spin state, which, as we have seen, is not necessarily the ground state, is the stereochemistry of addition to alkenes. The presumption is that the singlet state will form both new bonds in a single step, and will therefore be unable to disturb the stereochemical arrangement of any groups on the alkene. The triplet, by contrast, first forms a triplet 1,3-diradical in which rotation can scramble the stereochemistry originally present in the alkene. Irradia- tion of 1 in cis-2-butene gives only a mixture of the cis, anti and cis,syn diastereomers 6 and 7. Ten years after the original report, 4 a reinvestigation of the reaction with trans-2-butene revealed 4% of the cis,anti cyclopropane 6 along with 96% of trans cyclopropane 8. 6

R 6 8 R = o-carboranyl or Ph

congested cis,syn compound, and this preference is not present in the less sterically demanding addition of phenylcarbene. Addition must be from above the plane of the alkene, and the carborane can point away from the methyls and towards the hydrogens in the addition to cis-2-butene. Explanations for the preference of phenylcarbene for cis,syn addition over cis, anti all de- pend upon orbital connection between the phenyl ring and orbitals on the extra-ring carbon. As the EPR spectra of 3 and 4 show, there is not such a connection in the carboranylcarbenes. So, it is not surprising to see the strong preference of 3 for the cis, anti cyclopropane, 6.

H

hv 1 ~ 6 + 7

\ . / 84% ~6%

hv 1 ~ 6

4%

(R = o-earboranyl)

+ 8

96%

CHN2

h _ ~ = 6 + 7 + 8

- - 46% 51% 2 .5%

CHN2

~ 7 + 8 hv

3% 97%

(R = Ph)

The overall picture rather closely resembles phenylcarbene. Cycloaddition must be largely through the singlet state, even though the triplet is the ground state. It is always dangerous to make much of the small amounts of nonstereospecific addition that sometimes crop up in singlet carbene additions: Still, it may be that singlet addition to the trans alkene, in which both sides of the alkene are guarded by a methyl group, is more difficult than addition to the cis alkene, in which one side is open. For the carboranylcarbene, there is a strong preference for the cis,anti cyclopropane over the more

If, as seems likely, the small amount of the cis cyclopropane formed in the reaction with trans-2-butene is the result of triplet carbene addition, it should be possible to estimate the mix of reacting spin states. However, in this analysis it is necessary to augment the amount of cis, anti compound 6 with however much trans cyclopropane 8 derives from the triplet. In the absence of data on the stereochemical preferences of pure triplet, one can only estimate. This is especially hard in this case because the relative energies of the cyclopropane isomers are not known. Nonetheless, if the triplet is responsible for the foiTnation of 6 from trans-2-butene, some 8 must be formed as well. We take, as a guess for the thermodynamic preference, a 60/40 ratio of trans/cis. If this is correct, the 4% of observed 6 translates into 10% triplet reaction in the addition of 3 to trans-2-butene. There is no sign of triplet reactivity in the addition to cis-2-butene.

Other reactions of 3 can be compared to those of phenylcarbene. Carbenes not only add to rc systems, but insert into carbon-hydrogen bonds as well. The ability of a carbene to select among various kinds of carbon- hydrogen bonds, generally primary, secondary, and ter- tiary, has been used as a measure of reactivity. For example, phenylcarbene, presumably singlet, generated from several sources shows the same selectivity when inserting into the different carbon-hydrogen bonds of pentane. This has been taken as strong evidence that the free carbene is involved in all the reactions. For the reaction of singlet phenylcarbene with pentane, a secondary carbon-hydrogen bond is about 8.4 times as reactive as a primary carbon-hydrogen bond. 7 In the transition state for insertion, developing free valence is stabilized both by the benzene ring and the substituents on the carbon from which the hydrogen is being plucked.

596 Russ. Chem.Bull., VoL 42, No. 4, April, 1993 Jones

Whether the transition state is polar, or more likely, radical in nature, the more substituted the carbon, the better.

:CH H .......... cl L , .... ,

+ H--C--i I

t ransi t ion state

Carbene 3 shows similar, but slightly reduced selectivity. 4 In butane, secondary carbon-hydrogen bonds are more reactive than primary by a factor of 3. The carboranylcarbene is less selective than phenylcarbene. It is risky to rationalize small energy differences, but this difference is nonetheless consistent with a pair of transition states, one of which is stabilized by delocalization into the ring and another in which resonance stabilization is greatly attenuated.

" " " : C - -

. . . . . I .... I

delocalized not delocalized

1-m-Carboranylcarbenes. Synthesis of the diazo compound 9 follows the model of the ortho species, 1 and 2. A different protocol is used to make the aldehyde, but transformation into the tosylhydrazone salt, and diazo compound 9 are patterned exactly on the reactions used earlier. 4

@ HCi 1.BuLi ~ CH(OEt)2 HOAc

2.PhOC H(OEt)2

@ C H O ~ ~ - - - - ~ ~ C H N 2

Irradiation of the diazo compound 9 in cis- and trans-2-butene led to the products shown in Scheme 1. Cyclopropane formation is the major process, and the addition is largely stereospecific. The singlet carbene 10 must be responsible for most of the reactions. If, once again, we assume that triplet carbene is the sole source of the "wrong" cyclopropane, and if we assume that the thermodynamic ratio of cyclopropanes is cis/trans = 40/60, then there is approximately 8--12% triplet reactivity expressed in these additions.

Intramolecular chemistry of carbon-substituted earboranylearbenes. The most elaborate chemistry of phenylcarbenes appears in the gas phase, in what has come to be known as the phenylcarbene rearrangement. s The tolylcarbenes provide a nice example. Gen- eration of any tolylcarbene in the gas phase leads to benzocyclobutene and styrene. The mechanism involves a series of ring expansions and contractions resulting in an interconversion of the set of phenylcarbenes (Scheme 2).

When 4 is generated in the gas phase by thermal decomposition of diazo compound 2, the products are 1-methyl-o-carborane (11), 1,2-dimethyl-o-carborane (12), 1,2-ethano-o-carborane (13), and, most interest- ing, 1-vinyl-o-carborane (14), the analogue of styrene in the reactions of phenylcarbene. Of course, insertion into the adjacent methyl group must be the source of 13, but the origins of the other products are less clear (Scheme 3).

Scheme 1

10 R 6

5%

R = m-carboranyl

H 7 23%

+

R 8 5%

R

75%

+ insertion p rod ucts

14%

+ insertion products

20%

Carboranylcarbenes Russ. Chem.Bull., Vol. 42, No. 4, April, 1993 597

S c h e m e 2 Scheme 3

:CH 3 . H

: C ~

HaC 0[-t HaC H3C CH3

oas .has~ ~ t /

4 11 12

H

+

13 14

Generation of the methyl-labeled carbene 15 elimi- nates 13 as a source of 14. Further mechanistic specula- tion must continue to await our long-overdue labeling experiments. In particular, we do not yet know if ring- expanded, 13-vertex intermediates are permissable intermediates. Calculations by McKee at the MP2/6- 31G*/ /3-21G+ZPC level suggest that they are. 9

H 3 H *

~450~

gas phase

15 13 14

= 13 C

. / H

No similar reactions of m-carboranylcarbenes are known. Generation of 16 in the gas phase at -450~ does not lead to a vinyl carborane?

_ 450oc

gas phase

16

3-o-Carboranylcarbene, a boron-substituted carbene. An important theme in carbene chemistry has been the effect of substituents on the mix of singlet and triplet chemistry. Most attention has been paid to the rich chemistry of the narrowly separated singlet and triplet arylcarbenes and to the effects of halogen substitution. In the chemistry of arylcarbenes, reactions of both sin-

F_

8 .....

glets and triplets have been uncov- ered. Generally, halogen, oxygen, or nitrogen substitution preferentially stabilizes singlets as electrons are shared between a filled p orbital on the heteroatom and the empty 2p

orbital on carbon. For example, the chemistry of the halocarbenes is completely dominated by the singlet state, le

Almost no experimental attention has been paid to the other side of the periodic table where empty orbitals

X

- 9

tend to be more prominent than filled. However, theoretical work gives us some idea of what to ex- pect. For an atom such as boron, the presence of an empty p orbital should stabilize the singlet state

through overlap with the filled orbital of the divalent carbon. 1

In this case, however, there is a compensating effect stabilizing the triplet state. This operates through the c system and depends upon the electronegativity difference between the two atoms, lla The effect is to donate electrons through the ~r system and, for atoms less electronegative than carbon, favors the triplet state. Thus, for a simple boron-substituted earbene such as H2BCH , there should be compensating effects stabilizing both the singlet and triplet states. Calculations bear this out, with the most recent efforts predicting the two spin states to be close in energy. In the most simple example, triplet H2BCH is predicted to be the ground state by 4-6 kcal/mol. 12

598 Russ. Chem.BulL, Vol. 42, No. 4, April, 1993 Jones

In 1990, Ji Li at Princeton developed a route to the first boron-substituted carbenes (17,18, and 19) through manipulations of boron-substituted o-carboranes. 643

H H HaC OH 3 H3C" H

17 18 19

Although Li's method does yield carbenes adjacent to a boron atom, it must be emphasized that this is no ordinary boron. In particular, the empty 2p orbital, so influencial in the chemistry of "normal" trisubstituted boranes, is occupied in the web of three-center, two- electron bonding making up the icosahedral cage. Al- though the singlet-stabilizing ~ effect seems likely to be largely absent in a carboranylcarbene, the triplet-stabilizing ~ effect remains. So, our preliminary expectation was that a boron-substituted carboranylcarbene might show enhanced triplet reactivity, as the triplet should be especially favored at equilibrium.

The new carbenes are generated by the traditional route, photolysis of a diazo compound. The diazo compounds are made from the tosylhydrazone salts, them- selves available through the route shown. The tosylhydrazones were made by insertion of a vinyl- substituted boron into Li2B9C2Hu, 14 methylation of one or two cage carbons for 18 and 19, ozonolysis, and reaction with tosylhydrazine. Heating the sodium salts under vacuum (120~ Torr) led to the diazo compounds in about 80% yield (Scheme 4).

Table 1. EPR spectra of carboranylcarbenes, zero-field splitting parameters (cm -1)

Carbene ID/hc[ IE/hcl

1 -o-carboranylcarbene (3) 0.686 0.0302 1-methyl-2-o-carboranylcarbene (4) 0.6920 0.0293 3- o-carboranylcarbene (17) 0.657 <0.002

1,2-dimethyl- 3- o-carboranyl- 0.661 <0.002 carbene (18)

In collaboration with Linda Bush of the Berson group at Yale University, Rodney Blanch has determined that photolysis of these diazo compounds at ~ 11 K leads to persistent triplet EPR signals. At about 32 K, the signals irreversibly decay. Analysis of the zero-field splitting parameters showed that both triplet 17 and 18 are linear ([E/hcl ~0). Carbene 17 was shown to be a ground state triplet as the Curie law was obeyed. Table 1 summarizes the EPR spectra of carboranyIcarbenes. 15

Table 2 compares the stereochemistry of addition of 3, 17, and 18 to cis- and trans-2-butene and estimates the amount of triplet addition in these reactions.

It is the results of addition to cis-2-butene that are most revealing, trans-2-Butene is likely to give over- whelming amounts of trans adduct no matter what the mechanism, but a stepwise addition to the cis isomer should lead to substantial amounts of the more stable trans adduct. The last column of Table 2 is calculated on the assumption that the thermodynamic preference for trans adduct over cis is about 60/40. Although only small amounts of triplet are involved in any of the cycloaddition reactions, there seems to be more reactive triplet 18

Scheme 4

H H

/~8.'C',.~/c(c H s)3

1 .Bu Li2B 9 C 2 H 11 z,. Li

Br 2B""~ ~ jC(cH3)3 2.CH3I

R R R R

, , . , _-

N 2NNTs 2.120~ 0.02 Torr

R R

~ ~ B ~ " ~ J 'c(cH3)3

1"031 2"(CH3)2S

R R

R = CH 3 or H

Carboranylcarbenes Russ. Chem.Bull., VoL 42, No. 4, April, 1993 599

Table 2. Stereochemistry of addition reactions carboranylcarbenes.

CH3 H CH3

R A B R C CH3

R = o-carboranyl

of

Carbene Alkene

Products, %

A B C

Products of triplet reaction*

17 " ~ - - - [00 0 \

17 ~ 74 24 2 3

18 ~ 9 91 15 ---h

18 ~ / 63 25 12 20

3 ~ 4 96 10 ---h

3 ~ 84 16 0 0

*Calculated in an assumption of a -60/40 ratio of trans- to cis-2-butene.

than 17. This is surely reasonable given the additional interference to two-bond-forming reactions (singlets) of the pair of methyl groups on 18. One-bond forming reactions (triplets), taking place at the periphery of the molecule, will have an advantage for sterically congested carbenes, and this seems to be expressed here.

The reactions of carbenes 17, 18, and 19 with alkenes lead to products other than cyclopropanes. Both inter- and intramolecular carbon-hydrogen insertion reactions

occur with alkenes, and substantial amounts of 3-methyl-o-carboranes are also formed (Scheme 5).

Table 3 gives the relative amounts of cyclopropanes, total intermolecular insertion products, products of intramolecnlar insertion (the "cyclobutanes" 20 and 21), and B-methyl compounds 22, 23, and 24 for B-substituted carbenes 17, 18, and 19 and the C-substituted carbene 3.

It appears as though cyclopropane formation is favored somewhat by easy access to the alkene. Thus, cis-2- butene, in which one side of the alkene is unguarded by methyl groups gives the most cyclopropane, and tetramethylethylene, in which the maximum shielding of the double bond exists, gives the least.

As the products of allylic rearrangement were not stable over time, they could not be used as a diagnostic for triplet carbene activity. Therefore, a labelling experiment was undertaken in the style of Doering and Prinzbach, who, in 1959 had the daring to irradiate a sample of 14C-labeled isobutylene in what was essentially neat diazomethane.16 They determined that ailylic insertion took place without rearrangement, and attributed this result to the singlet nature of reacting methylene. Deuterium-labelled isobutylene was synthesized by Ji Li and used as a substrate for carbene 17. As the figure shows, the triplet should reveal itself through the appearance of protons in the terminal methylene position of the insertion products 25 and 26 (Scheme 6).

Analysis of the crude products by IH N M R spectroscopy showed only small amounts of protium in the position of absorption of the terminal methylenes of 25 (5 = 4.74, 5--10%). Although it was not possible to make an accurate quantitative determination of the amount of triplet participation in allylic insertion, clearly the insertion product was overwhelmingly 25, the isomer derived from the singlet, not 26, the isomer diagnostic for the triplet. The reason may well be that carboranyl radicals prefer abstraction of a second hydrogen to recombination. Methyl compounds 22, 23, and 24 are substantial products of the react ions of carboranylcarbenes with alkenes. Similarly, 22 appears

Scheme 5

R R"

1 7 : R = R ' = H 1 8 : R = H , R ' = C H 3 1 9 : R = R ' = C H 3

2 0 : R = OH 3 21 : R = H

R R"

~ - ~ C H s

2 2 : R = R ' = CH 3

2 3 : R = H, R ' = OH 3

2 4 : R = R ' = H

600 Russ.C hem.Bull., Vol. 42, No. 4, April, 1993 Jones

Table 3. Products from carbenes 3, 17, 18, and 19.

*R = H, CH 3

A B

Carbene Total Total

Alkene allylic cyclo- A B insertion propanes

I8

18

18

17

!7

17

19

3

3

3

36 44 4 15

40 33 4 23 --h

38 25 8 29

"~......_/ 28 50 19

41 36 23

48 16 36

49 18 3 31

8 84 8

~ - - 7 87 6

~ : ~ 27 67 - 6

in 15% yield in the reaction of isobutylene with 17. It seems safe to assign most insertion products to the singtet.

An analysis of the reaction of 17 with cyclohexane and dodecadeuterio-cyclohexane 17 was used to verify the notion that most insertion product comes from singlet carbene, and to provide an estimate for the isotope effect for the singlet insertion reaction. Carbene t7 was generated from irradiation of the diazo compound in both pure cyclohexane and dodecadeuteriocyclohexane to produce insertion products 27 and 28. In a third experiment, 17 was allowed to react with a 1:1 mixture of cyclohexane and dodecadeuteriocyclohexane. The presumption in this experiment was that triplet 17 would be revealed through formation of cross products 29 and 30 formed from recombition of radicals produced by hydrogen and deuterium abstraction (Scheme 7).

Adducts 27 and 28 were welt resolved by gas chro- matography, but the cross products 29 and 30 differ by only a single deuterium from 27 and 28 and could not be separated. Analysis of mass spectra showed that within the peak for 27 there was a small amount of a product with an extra deuterium (presumably 38), and that within the peak for 28 was a compound with one less deuterium, (29). This analysis indicates that triplet abstraction could account for only about 5N of the overall insertion reaction.

When the reaction with the mixture of cyclohexanes was run in the presence of oxygen, a species likely to scavenge the radicals formed by triplet abstraction, t8 the ratio of 27 and Z8 changed from 1:1 to 1.2:1. This provides a measure of the isotope effect, kH/k D for the singlet insertion reaction of 17 with cyclohexane.

So, both cyclopropanation and insertion reactions of the boron-substituted carbenes seem to be largely the result of singlet reactivity. There is no great difference

S c h e m e 6

H H H H �9 .

triplet F HaC ~ C H H3C ::CD2abstractio n ~ H2+ L H2C~:::OD2

+ H3 C 17

singlet l direct in sertion H 3CN, ~

CD 2 H H H2C

H3%CD2 H H H2C

25 25

HaC~_~D 2 H2C

recombine

H3C~cD2 H H2C H

26

Carboranylcarbenes Russ. Chem.Bull., Vol. 42, No. 4, April, 1993 601

Scheme 7

H H

k__._/

27

Direct, singlet insertion

H H

17

Direct, sincjlet I insertion

H \_ /H , / - - -~ D11 /~----'~... CHD--( " ,~

2 8

~ H1

-...,. H .

~ N iTl~Plrtif,ot . . . . tep

D12 H H ~ / bHO

H H

V "cross products"

H H

31)

between these intermediates and 3 or 4, the carbon- attached carbenes. What, then, of our expectation of increased triplet reaction? Where is the triplet? The answer seems to be, in the methyl groups of compounds 22, 23, and 24. These compounds must arise by a double abstraction reaction and are most reasonably attributed to the triplet. In order to test this notion and to verify our mental separation of products into those formed f rom singlet (cyclopropanes of retained

stereochemistry and most insertion products) and triplet products (wrong cyclopropanes plus an estimated por- tion of the other cyclopropanes, and methyl compounds) we have repeated the reaction of 18 with cis-2-butene in an atmosphere of oxygen. Oxygen is a known trap for triplets, is and should reduce or eliminate compounds that owe their origin to a triplet reaction. Scheme 8 shows the products of the react ion of 18 with cis-2-butene in the absence and presence of oxygen.

Scheme 8

1 8

R

11%

R

7%

R CH3 : ~ 3 C \__./ c"3 - - + + H3

R 25% 28% 11%

R CH3

R

32% 37% 21%

CH3

R CH3 5%

CH3 + H~ +

R CH3 <1%

+ 2 0 + 2 2

4% 15%

2 0 + 2 2

2% trace

R = o-carboranyl

602 Russ. Chem.Bull., Vol. 42, No. 4, April, 1993 Jones

Table 4. Singlet and triplet reactivity for some carbenes.

Carbene Alkene Triplet Singlet carbene ~ (%) carbene b (%)

3 cis-2-Butene 9 9I trans- 2- Butene 16 84 2, 3- Dimethyl - 2-butene >6 <94

10 cis-2-Butene 12 88 trans- 2- Butene 8 92

17 cis-2-Butene 22 78 trans- 2- Butene 23 77 2,3-Dimethyl- 2-butene >36 <64

19 cis-2-Butene 46 54 trans-2-Butene 38 62 2,3-Dimethyl- 2-butene >29 <71

18 2,3-Dimethyl- 2-butene >31 <69

a"Wrong" cyclopropane plus some "right" cyclopropane, plus methyl compound. b"Right" cyclopropane plus allylic insertion, plus self-insertion.

As predicted, the products attr ibuted to triplet carbene, the trans cyclopropane and methyl compound 22, vanish. In their place are found o-carborane-3- carboxaldehyde, o-carborane-3-carboxylic acid, and 3- o-carboranyl formate (not shown in scheme 6). The 3- aldehyde is the expected product of oxidation of the triplet carbene, and the other two compounds are known to derive from overoxidation of the aldehyde. 19 We now feel it justified to provide a new estimate of the amount of triplet reactivity present in the reactions of carbenes 3, 1 17, 18, and 19 with cis- and trans-2-butenes (Table 4).

The remaining uncertainties are the amount of triplet reactivity to assign to the allylic insertion reaction, and the extent (if any) to which triplet participates in the self-insertion reaction to give 20 and 21. These will add only very small corrections to the numbers of Table 4, which assumes that the insertion products, like the cyclopropanes, are mainly derived from the singlet carbene. Boron-substi tuted carbenes 17, 18, and 19 show roughly twice the triplet reactivity as do 3 and 10.

Intramolecular Chemistry of Boron-substituted Carbenes. Although the four-membered ring compounds 20 and 21 make up only a small amount of the products when carbenes 18 and 19 are generated in solution, the situation charges markedly when these carbenes are produced in the gas phase. Here, of course, interrno- lecular chemistry is minimized in favor of intramolecular reactions, and 20 and 21 become major products. No evidence o f phenylcarbene rearrangements can be found, as vinyl carboranes are not present in more than trivial amounts.

450~ •••B CH3

-~ 20

" ,

450~ 21

References

i. (a) G. A. Olah, G. K. Surya Prakash, R. E. Williams, L. D. Field, and K. Wade, Hypercarbon Chemistry, Wiley, New York, 1987; (b) E. L. Muetterties, Boron Hydride Chemis- try, Academic, New York, 1975; (c) R. N. Grimes, Carboranes, Academic, New York, 1970; (d) T. Onak, Organoborane Chemistry, Academic, New York, 1975. (e) V. I. Bregadze, Chem. Rev., 1992, 92, 209.

2. J. S. Andrews, J. Zayas, and M. Jones Jr., lnorg. Chem., 1985, 24, 3715.

3. A. M. Trozzolo, R. W. Murray, and E. Wasserman, J. Am. Chem. Soc., 1962, 64, 4990.

4. S. L. Chari, S.-H. Chiang, and M. Jones Jr., J. Am. Chem. Soc., 1982, 104, 3138.

5. R. S. Hutton, H. D. Roth, and S. Chari, J. Phys~ Chem., 1981, 85, 753.

6. J. Li, D. J. Caparrelli and M. Jones Jr., J. Am. Chem. Soc., 1993, 115, 408.

7. (a) H. Dietrich, G. W. Griffin, and R. C. Pettersen, Tetrahedron Lett. 1968, 153; (b) G. L. Closs and R. A. Moss, J. Am. Chem. Soc., 1973, 86, 4042.

8. P. P. Gaspar, J.-P. Hsu, S. Chari, and M. Jones Jr., Tetrahedron, 1985, 41, 1479.

9. (a) M. L McKee, J. Am. Chem. Soc., 1991, 113, 9448; (b) M. L. McKee, J. Am. Chem. Soc., 1992, 114, 5856.

10. M. Regitz Ed., Carbene(oide), Georg Thieme, Stuttgart, 1989.

l 1. (a) E. R. Davidson, in Diradicals, Ch 2, W. T. Borden Ed., Wiley, New York, 1982; (b) D. Feller, W. T. Borden, and E. R. Davidson, J. Chem. Phys., 1979, 71, 4987.

12. (a) B. T. Luke, J. A. Pople, M.-B. Krough-Jespersen, Y. Apeloig, M. Karni, J. Chandrasekhar, and P. v. R. Schleyer, J. Am. Chem. Soc., 1986, 108, 270; (b)P. v. R. Schleyer, B. T. Luke, and J. A. Pople, Organometallics, 1987, 6, 1997.

13. J. Li and M. Jones, Jr . . , J. Am. Chem. Soc., 1991, 114, 1094.

14. M. F. Hawthorne, D. C. Young, P. M. Garrett, D. A. Owen, S. G. Schwerin, F. N. Tebbe, and P. A. Wegner, J. Am. Chem. Soc. 1968, 90, 862.

15. R. J. Blanch, J. Li, L. C. Bush, and M. Jones, Jr., J. Am. Chem. Soc. 1992, 114, 9236.

16. W. von E. Doering and H. Prinzbach, Tetrahedron, 1959, 6, 24.

17. R. L. Barcus, L. M. Hadel, L. J. Johnston, M. S. Platz, T. G. Savino, and J. C. Scaiano, J. Am. Chem. Soc., 1986, 108, 3928.

18. S. Morgan, M. S. Platz, M. Jones, Jr., and D. R. Myers, J. Org. Chem., 1991, 56, 1351.

Received February 5, 1993

Documents

Carboranylcarbenes