Modulation of carbene spin state population through precursor photophysics

Article abstract of DOI:10.1021/ol802934w

Dicarbomethoxycarbene can be generated by photolysis of S-C sulfonium ylides derived from thiophene. By manipulating the thiophene (leaving group) portion of the ylide, the initial spin distribution of the carbenes can be strongly influenced. With certain carbene traps, product distributions from dicarbomethoxycarbene depend on the initial spin state distribution in which the carbene is generated and this is used as a means to report on the initial spin state distributions. This approach should be general for other carbenes generated from analogous precursors.

Full text of DOI:10.1021/ol802934w

ORGANIC
LETTERS
2009
Vol. 11, No. 4
955-958
Modulation of Carbene Spin State
Population through Precursor
Photophysics
William S. Jenks,* Melanie J. Heying, and Erin M. Rockafellow
Department of Chemistry, Iowa State UniVersity, Ames, Iowa 50011
wsjenks@iastate.edu
Received December 19, 2008
ABSTRACT
Dicarbomethoxycarbene can be generated by photolysis of S-C sulfonium ylides derived from thiophene. By manipulating the thiophene (leaving
group) portion of the ylide, the initial spin distribution of the carbenes can be strongly influenced. With certain carbene traps, product distributions
from dicarbomethoxycarbene depend on the initial spin state distribution in which the carbene is generated and this is used as a means to report
on the initial spin state distributions. This approach should be general for other carbenes generated from analogous precursors.
We recently reported that photolysis of thiophene-based
sulfonium ylides TY, BTY, and DBTY results in formation
of dicarbomethoxycarbene, 1.1
It is notable that, for these carbene precursors, the
chromophore is centered on the thiophene, rather than on
the carbene. It was posited that the variation in 2:3 was due
to the intimate photophysics of each chromophore and the
relative initial populations of 1 and 1 that are generated.
Indeed, the generation of both singlet and triplet benzoylni-
trene from the closely related N-benzoyl dibenzothiophene
sulfilimine was recently documented.2
This supposition requires that trapping of the carbene is
faster than, or at least competitive with, intersystem crossing
by the carbene. Jones and co-workers showed that direct and
sensitized photolysis of methyl dizaomalonate gave differing
stereochemical outcomes for cyclopropanation,3 so the notion
Photoloysis of these compounds in methanol resulted in
different ratios of the formal OH insertion product 2,
interpreted as the product of the singlet carbene (11), and
1
3
3
dimethyl malonate (3), interpreted as the product of 1
(Scheme 1). The ratio of 2:3 varied from a high of about 14
for BTY to a low of about 1.4 for DBTY.
(1) Stoffregen, S. A.; Heying, M.; Jenks, W. S. J. Am. Chem. Soc. 2007,
129, 15746–15747.
(2) Desikan, V.; Liu, Y.; Toscano, J. P.; Jenks, W. S. J. Org. Chem.
2007, 72, 6848–6859.
10.1021/ol802934w CCC: $40.75
Published on Web 01/26/2009
2009 American Chemical Society

obtained at full conversion, as determined by NMR integra-
tion.9 Enough repetitions were done for the new compounds
to determine reasonable standard deviations of the yields
(Table 1). The fractions of products attributable to singlet
insertion, that is, 2, are also given.10
Scheme 1.
Photolysis of Thiophene Ylides in Methanola
Table 1. Yields of OH Insertion and Hydrogen Abstraction on
Photolysis of Sulfonium Ylides in Methanol
precursora
yield 2 (%)b
yield 3 (%)b
2/(2 + 3)
a A 22% yield of 4 was obtained from photolysis of TY.
TYc
52
5
0.91
0.95
0.94
0.90
0.88
0.64
0.58
0.53
0.52
0.14
0.30
0.044
2,5-dichloroTY
2-bromoTY
2,4-dibromoTY
3,4-dibromoTY
2-iodoTY
94.2 ( 1.3
94.7 ( 4.3
86.5 ( 2.5
86.2 ( 2.1
49.3 ( 1.4
54
50.3
51.5
9
28
4.8 ( 0.6
5.9 ( 1.1
9.4 ( 1.4
12.0 ( 2.1
27.6 ( 1.3
39
44.6
48.0
57
65
of intersystem crossing (ISC) being slower than or compa-
rable in rate to reactions of the carbene was not without
precedent. It also must be the case that the triplet energy of
the precursor is at least as high as the S-C BDE required to
produce the thiophene derivative and the triplet carbene.
Using data from a previous computational investigation4 the
S-C BDE in TY can be estimated at approximately 44 kcal/
mol, surely lower than the triplet energy of the ylide.
Similarly, an estimate of approximately 60 kcal/mol can be
obtained for DBTY, which is similar to or a few kcal/mol
lower than our best estimate of the triplet energy of that ylide,
based on dibenzothiophene sulfoxide.5
DBTYc
4-bromoDBTY
2,8-dibromoDBTY
3BP + DBTYd
6
7
3.7
80
a See illustration of parent compounds for ring numbering. b Yields at
full conversion. Error limits for yields of 2 and 3 are standard deviations.
c Reference 1. Compound 4 also formed from TY. d DBTY sensitized by
irradiation of benzophenone.
In this letter, we show that the initial spin distribution of
dicarbomethoxycarbene can be manipulated over a wide
range by adjusting the structure of the thiophene portion of
the sulfonium ylide precursor. This allows the decoupling
of the spin state of the carbene from its structure, instead
relying on the photophysics of the leaving group, where the
critical chromophore of the precursor resides. A heavy-atom
strategy seemed particularly plausible, given previous results
in which heavy atom substitution on dibenzothiophene-S-
oxide had modestly increased the quantum yield of photo-
chemical deoxygenation.6
Two compounds merit specific discussion. 2,4-Dibro-
moTY decomposed even in the dark over the course of tens
of hours. On standing in methanol in the dark for two days,
only 2,4-dibromothiophene and dimethyl methoxymalonate
(2) were observed.
Photolysis of 2-iodoTY produced not only the expected
products but also thiophene. However, closely monitored
photolyses showed that thiophene was a secondary photolysis
product of 2-iodothiophene. Thus, the distribution of 2 and
3 derives only from photolysis of 2-iodoTY and not from a
mixture of that and TY itself.
A series of substutited analogs of TY and DBTY were
prepared and photolyzed in methanol. The rate of reaction
It is apparent from the error limits that TY,10 2,5-
dichloroTY, and 2-bromoTY produce essentially the same
fraction of singlet product with methanol. The dibrominated
derivatives of TY did produce a larger fraction of triplet
product 3, as did 2-iodoTY. Similarly, the brominated
analogs of DBTY gave larger fractions of 3 than the
unsubstituted analog. Benzophenone-sensitized DBTY gave
more 3 than direct photolysis of any of the bromo- or iodo-
compounds (57%) and considerably less 2, but the mass
balance is not as good.
Photochemical cleavage of a selenophene ylide was also
investigated, based on the precedent of sulfoxide (or sele-
noxide) deoxygenation. Substitution of selenium for sulfur
enhances the efficiency of photochemical deoxygenation of
dibenzoselenophene-Se-oxide vs dibenzothiophene-S-oxide
7
1
between 1 and neat methanol should approach 1010 s-1.
1
3
While the rate of instersystem crossing from 1 to 1 is not
known, it is not likely to be competitive with such a reaction
velocity. However, the rate of reaction by hydrogen abstrac-
8
3
tion by 1 is likely to be ∼3 orders of magnitude slower.
The ratio of 2/3 was determined at low conversion (ca.
10%) by GC and was within experimental error of values
(3) (a) Jones, M., Jr.; Ando, W.; Hendrick, M. E.; Kulczycki, A., Jr.;
Howley, P. M.; Hummel, K. F.; Malament, D. S. J. Am. Chem. Soc. 1972,
94, 7469–79. (b) Jones, M., Jr.; Ando, W.; Kulczycki, A., Jr. Tetrahedron
Lett. 1967, 139, 1–6.
(4) Stoffregen, S. A.; McCulla, R. D.; Wilson, R.; Cercone, S.; Miller,
J.; Jenks, W. S. J. Org. Chem. 2007, 72, 8235–8242. A group additivity
approach must be used to approximate the BDE for TY or DBTY.
(5) Jenks, W. S.; Lee, W.; Shutters, D. J. Phys. Chem. 1994, 98, 2282–
2289.
(6) Nag, M.; Jenks, W. S. J. Org. Chem. 2004, 69, 8177–8182.
(7) Wang, J.-L.; Toscano, J. P.; Platz, M. S.; Nikolaev, V.; Popik, V.
J. Am. Chem. Soc. 1995, 117, 5477–5483.
(9) Overlap of 1H signals made integration difficult at low to moderate
conversion.
(10) The reaction of 1 with thiophene to form 4 is probably a singlet
reaction and therefore artificially lowers the yield of 2 (and its fraction)
when TY is the precursor. The analog of 4 was not found in any other
case.
(8) See, for example, Lusztyk, J.; Kanabus-Kaminska, J. M. Representa-
tive Kinetic Behavior of Selected Reaction Intermediates: Free Radicals.
In Handbook of Organic Photochemistry; Scaiano, J. C., Ed.; CRC Press:
Boca Raton, FL, 1989; Vol. 2, pp 177-209.
956
Org. Lett., Vol. 11, No. 4, 2009

by an order of magnitude,11 although it is not clear whether
the enhanced quantum yield is due to the weaker Se-O bond
or the heavy atom effect.
bypassing the carbene completely. We do not have any
evidence for similar reactivity of the sulfonium ylides.1
Here, thermolysis of TY and DBTY was carried out in
10% cis-4-octene in acetonitrile as a control for 100% singlet
generation,15 and nearly quantitative yields of cis-5 were
accompanied by only trace quantities of the trans adduct.
This established that the “singlet carbene limit” is nearly
quantitative cis-5, which can be compared to the results for
other conditions in Table 2.
The parent malonic selenophene ylide was known to be
insufficently stable at room temperature for these experi-
ments.12 However, malonic ylide of 2,5-dichloroselenophene
6 was straightforward to handle.13 Photolysis of 6 in
methanol led to a very high proportion of the triplet product,
greater than any other compound that was not sensitized
(Table 1).
Table 2. Yields of cis- and trans-Cyclopropanes from
Photolysis of Sulfonium Ylides in 10% cis-4-octene in
Acetonitrile
Control experiments, in which all of the substituted
thiophene ylides, dibenzothiophene ylides and 6 were heated
in a septum-sealed tube in methanol in an oven at 80 degrees
until all the material was decomposed, resulted in formation
of only 2, to the exclusion of 3. This implies that the
population of 11 is quantitatively trapped before ISC.
However the benzophenone-sensitized results at least suggest
that some 31 to 11 ISC occurs on the time scale of the slower
hydrogen abstraction reaction of the triplet carbene. Given
this result, and the low mass balance, compound 7 was also
prepared,14 with the idea that internal triplet sensitization
should be quantitative and that irradiation could be carried
out so that nearly all the light was absorbed by the
benzophenone moiety. The mass balance after photolysis was
considerably improved, compared to the bimolecular sensi-
tization case. As shown, the fraction of singlet product was
4.4%. We take this as an implication that there is a slight
leakage of 31 to 11 ISC that competes with hydrogen
abstraction.
A second classic probe for spin multiplicity is the
stereochemistry observed on cyclopropanation. Singlet car-
benes (and related species) are expected to react with alkenes
with retention of stereochemistry, while triplet carbenes give
cyclopropanes of mixed stereochemistry. Thus, starting from
a cis alkene, the appearance of the trans cyclopropane is
diagnostic of triplet reactivity.
Direct irradiation of methyl diazomalonate is known to
give approximately 90% retention of stereochemistry with
cis-4-methyl-2-pentene, whereas the opposite ratio was
obtained on sensitization with benzophenone.3 However, it
was concluded that, at high alkene concentrations, some of
the loss of stereochemistry on direct irradiation was due to
reaction between the excited diazo compound and the alkene,
precursor
yield cis-5a
yield trans-5
% singletb
TYd
95
5
3.7
3.2
8.3
8.6
20.4
40
35.4
36.8
41.7 ( 0.5
88
91
92
80
78
46
4.3
13
10
0c
2,5-dichloroTY
2-bromoTY
2,4-dibromoTY
3,4-dibromoTY
2-iodoTY
95.2
93.5
89.5
86.8
70.2
60
62.1
61.5
58.3 ( 0.3
DBTYd
4-bromoDBTY
2,8-dibromoDBTY
7
a Yields determined at complete conversion using GC. b Determined
from (cis-5 -1.40 • trans-5)/(cis-5 + trans-5), and assumes no intersystem
crossing. See text. c Used as the standard to establish a lower limit of singlet
chemistry. d Relative yields.
Using direct photolysis of methyldiazomalonate, Platz and
co-workers obtained rate constants near 3 × 108 M-1s-1 for
the reaction of cis-alkenes with 1 in concentration ranges of
up to a few M,7 similar to the concentration used here, near
1 M. They make no mention of curvature in their quenching
plots, which indicates a single kinetic species. This implies
that either intersystem crossing of 1 is very fast (subnano-
second) or that very little triplet carbene was generated in
their experiments. We find the latter interpretation to be more
consistent with current results, and believe that quantitative
trapping of 11 on the time scale of a few ns by ∼1 M alkene
to be very plausible.
In contrast, the all-31 product limit in 1 M cis-4-octene
should be a mixture of the cis and trans cyclopropanes. It
was established using compound 7. Multiple photolyses were
carried out in order to establish a standard deviation for
reproducibility, which was approximately 1% of the reported
yields. A ratio of cis to trans 5 of 1.40 was obtained.
Obviously, this is not the equilibrium ratio of the two
cyclopropanes. Two interpretations are possible: (1) ISC of
(11) McCulla, R. D.; Jenks, W. S. J. Am. Chem. Soc. 2004, 126, 16058–
16065.
(12) Bien, S.; Gronowitz, S.; Hoernfeldt, A. B. Chem. Scr. 1984, 24,
253–254.
(13) Compound 6 was prepared by coupling of dimethyl diazomalonate
in the usual fashion to 2,5-dichloroselenophene. Gronowitz, S.; Frejd, T.
Acta Chem. Scand. B 1976, B30, 439–449.
1
31 to 1 competes with its capture by the octene; (2) the
(14) Compound 7 was prepared by esterification of 3-benzoylbenzoic
acid with commercially available thienylmethanol, followed by coupling
of the diazomalonate in the usual fashion.
(15) Thermolysis was at 80 °C in a sealed tube, using the same initial
concentration and solvent mixture as for the photolyses.
Org. Lett., Vol. 11, No. 4, 2009
957

3
biradical formed on addition of 1 to the olefin does not
carbene would obviously be compound 7, whose preparation
equilibrate completely before its closure. We do not have
the data to distinguish between these possibilities with any
certainty, but we can recognize this as the maximum fraction
of trans-5 obtainable.
is very straightforward as well.
As a final remark, it is worth pointing out that the
photodissociation of these ylides to form carbenes has proven
to be considerably more flexible, with respect to the
thiophene derivative, than the corresponding generation of
O atoms from sulfoxides.6,11,16 Photochemical generation of
nitrenes from sulfilimines appears to be, reasonably, an
intermediate case.2,17 While a causal relationship has yet to
be established, a good correlation exists between this and
estimates of the relevant bond dissociation energies which
are weakest for the sulfonium ylides.18 We believe this
approach to carbene generation using one or more of the
sulfonium ylides should be general for preparation of any
carbene whose corresponding S,C ylide is synthetically
accessible.
Accordingly, the present ylides were photolyzed in the
same solvent mixture of 10% cis-4-octene in acetonitrile.
The results are given in Table 2. If it is assumed that only
31 is generated on photolysis of 7, and both the singlet and
triplet carbenes are quantitatively trapped by the olefin before
ISC, then the 58.3:41.7 (1.40) ratio of cyclopropane stere-
oisomers can be taken to reflect the limit achieved from
3
1
reaction of only 1. Given that 1 results only in the cis
isomer, then the fraction of singlet chemistry in cyclopro-
panation can be determined from the observed yields of cis
and trans cyclopropane 5 for a given precursor. These
numbers, given in Table 2, are only an upper limit on the
1
3
initial yield of 1, because we cannot establish that 1 is
quantitatively trapped.
Acknowledgment. We acknowledge the National Science
Foundation (CHE 0213375) for partial support of this
research.
The results in Tables 1 and 2 deserve comparison. The
parallelism between the methanol and olefin photolysis
results is striking, with 2,5-dichloroTY and 2,5-dibromoTY
giving nearly identical cis/trans ratios, followed descending
stereospecificity for the rest of the entries. However, in
contrast to the alkene results, the methanol chemistry clearly
shows that some singlet chemistry occurs even on quantita-
Supporting Information Available: Detailed procedures
and product characterization. This material is available free
of charge via the Internet at http://pubs.acs.org.
OL802934W
3
tive generation of 1, so the exact estimates of singlet
chemistry cannot align quantitatively.
(16) (a) Wan, Z.; Jenks, W. S. J. Am. Chem. Soc. 1995, 117, 2667–
2668. (b) Gregory, D. D.; Wan, Z.; Jenks, W. S. J. Am. Chem. Soc. 1997,
119, 94–102. (c) Nag, M.; Jenks, W. S. J. Org. Chem. 2005, 70, 3458–
3463. (d) Heying, M. J.; Nag, M.; Jenks, W. S. J. Phys. Org. Chem. 2008,
21, 915–924. (e) Arima, K.; Ohira, D.; Watanabe, M.; Miura, A.; Mataka,
S.; Thiemann, T.; Valcarcel, J. I.; Walton, D. J. Photochem. Photobiol.
Sci. 2005, 4, 808–816. (f) Kumazoe, K.; Arima, K.; Mataka, S.; Walton,
D. J.; Thiemann, T. J. Chem. Res. 2003, 60–61. (g) Thiemann, T.; Ohira,
D.; Arima, K.; Sawada, T.; Mataka, S.; Marken, F.; Compton, R. G.; Bull,
The obvious trend, regardless of the precision of the
estimate of the fraction of singlet carbene, is that heavy atom
substitution on the thiophene or dibenzothiophene chro-
mophore leads to greater fractions of triplet chemistry. This
confirms the original proposal1 that the variation from
thiophene to benzothiophene to dibenzothiophene was itself
due to variations in the photophysics of the precursor. As
expected, there is little effect with Cl substitution. It is clear
that the 3- and 4-positions are better than 2- and 5-positions
for increasing the triplet yield from the thiophene based
ylides from the bromo compounds, though we did not try to
duplicate this effect with iodo substituents. This is, in part,
because the most efficient manner for generating triplet
S.; Davies, S. G. J. Phys. Org. Chem. 2000, 13, 648–653
.
(17) (a) Fujita, T.; Kamiyama, H.; Osawa, Y.; Kawaguchi, H.; Kim,
B. J.; Tatami, A.; Kawashima, W.; Maeda, T.; Nakanishi, A.; Morita, H.
Tetrahedron 2007, 63, 7708–7716. (b) Fujita, T.; Maeda, T.; Kim, B. J.;
Tatami, A.; Miyamoto, D.; Kawaguchi, H.; Tsuchiya, N.; Yoshida, M.;
Kawashima, W.; Morita, H. J. Sulfur Chem. 2008, 29, 459–465. (c) Morita,
H.; Tatami, A.; Maeda, T.; Ju Kim, B.; Kawashima, W.; Yoshimura, T.;
Abe, H.; Akasaka, T. J. Org. Chem. 2008, 73, 7159–7163
.
(18) Stoffregen, S. A.; McCulla, R. D.; Wilson, R.; Cercone, S.; Miller,
J.; Jenks, W. S. J. Org. Chem. 2007, 72, 8235–8242.
958
Org. Lett., Vol. 11, No. 4, 2009