Rearrangement of alkylchlorocarbenes: 1,2-H shift in free carbene, carbene-olefin complex, and excited states of carbene precursors

Article abstract of DOI:10.1021/ja952700n

Photolysis of alkylchlorodiazirines (1) in the presence of olefins gives a cyclopropane (3) by addition of the generated carbene to the olefin and a vinyl chloride derivative (2) resulting from a 1,2-H shift rearrangement. This rearrangement may occur either in the carbene or in some excited state, precursor of the carbene (RIES mechanism), or in a 'carbene + olefin complex' on the way to the formation of 3 (COC mechanism). Results obtained by time-resolved photoacoustic calorimetry as well as by thermolysis and photolysis of ClCH₂C(N₂)Cl and CH₃(CH₂)₂C(N₂)Cl in the presence of tetramethylethylene clearly indicate that both the RIES and COC mechanisms play a role but with efficiencies which greatly depend on the nature of the diazirine. Reexamination of the results previously obtained with benzylchlorodiazirines indicates that, for this class of diazirines, the RIES mechanism is temperature dependent and has a very low efficiency at room temperature and below, whereas the nonlinearity of the plots [3]/[2] vs [olefin] is mainly due to the COC mechanism.

Full text of DOI:10.1021/ja952700n

J. Am. Chem. Soc. 1996, 118, 3829-3837
3829
Rearrangement of Alkylchlorocarbenes: 1,2-H Shift in Free
Carbene, Carbene-Olefin Complex, and Excited States of
Carbene Precursors
,†
‡
‡
§
Roland Bonneau,* Michael T. H. Liu, Kyu Chul Kim, and Joshua L. Goodman
Contribution from the URA 348 du CNRS, UniVersit e´ de Bordeaux I,
3
3405 Talence Cedex, France, Department of Chemistry, UniVersity of Prince Edward Island,
Charlottetown, Prince Edward Island, Canada, and Department of Chemistry, UniVersity of
Rochester, Rochester, New York 14627
X
ReceiVed August 9, 1995
Abstract: Photolysis of alkylchlorodiazirines (1) in the presence of olefins gives a cyclopropane (3) by addition of
the generated carbene to the olefin and a vinyl chloride derivative (2) resulting from a 1,2-H shift rearrangement.
This rearrangement may occur either in the carbene or in some excited state, precursor of the carbene (RIES
mechanism), or in a “carbene + olefin complex” on the way to the formation of 3 (COC mechanism). Results
obtained by time-resolved photoacoustic calorimetry as well as by thermolysis and photolysis of ClCH2C(N2)Cl and
CH3(CH2)2C(N2)Cl in the presence of tetramethylethylene clearly indicate that both the RIES and COC mechanisms
play a role but with efficiencies which greatly depend on the nature of the diazirine. Reexamination of the results
previously obtained with benzylchlorodiazirines indicates that, for this class of diazirines, the RIES mechanism is
temperature dependent and has a very low efficiency at room temperature and below, whereas the nonlinearity of
the plots [3]/[2] Vs [olefin] is mainly due to the COC mechanism.
Introduction
Scheme 1
Alkylchlorocarbenes with a C-H bond in the R position of
the carbene center, 4, which may be produced by photolysis or
thermolysis of diazirines, 1, generally undergo a fast 1,2-H
migration (rate constant ki) to give a vinyl chloride derivative,
2
. In recent years, much information has been obtained on the
1
kinetics of this process. This rearrangement is in competition
with intermolecular reactions such as addition to olefins (rate
constant kr) to give a cyclopropane, 3. The simplest mechanism
to describe the system, given in Scheme 1, predicts a linear
relation between the ratio [3]/[2] and the concentration of the
olefinic reactant, [olefin].
Scheme 2
But the distribution of the products obtained under either
photolysis or thermolysis do not fit this simple mechanism. Plots
of [3]/[2] Vs [olefin] are curved, indicating that, at high
concentrations of olefin reactant, the amount of [2] is larger
than expected (Figure 1). This has been explained by either a
rearrangement in excited states (RIES) or by a 1,2-H migration
in a diazo intermediate or by the formation of a carbene-olefin
complex (COC).
made by several research groups3,4 also lend support to the RIES
mechanism shown in Scheme 2.
With R ) kC/(k*i + kC) being the yield of formation of the
carbene, this mechanism which may be invoked only in the case
of photolysis gives
The RIES mechanism is supported by the fact that, in the
absence of olefin, the rearrangement products obtained by
thermolysis and by photolysis of some diazirines show a quite
different distribution of isomers.2 The more recent observations
[2]/[3] ) k* /k + [k /(Rk )]/[olefin]
i
C
i
1
†
Universit e´ de Bordeaux I.
University of Prince Edward Island.
University of Rochester.
Abstract published in AdVance ACS Abstracts, March 15, 1996.
The nature of the excited state(s) involved in this process is
uncertain. It could be the following: (i) The excited state of
diazirines, which may have a nonnegligible lifetime since the
fluorescence of several akyldiazirines has been reported.4
Theoretical studies indicate that several reaction pathways are
‡
§
X
(
1) (a) Moss, R. A. AdVances in Carbene Chemistry; Brinker, U., Ed.;
JAI: Greenwich, CT, 1994; Vol. 1. (b) Jackson, J. E.; Platz, M. S. AdVances
in Carbene Chemistry; Brinker, U., Ed.; JAI: Greenwich, CT, 1994; Vol.
1
. (c) Liu, M. T. H. Acc. Chem. Res. 1994, 27, 287. (d) Nickon, A. Acc.
Chem. Res. 1993, 26, 84.
2) (a) Chang, K. T.; Shechter, H. J. Am. Chem. Soc. 1979, 101, 5082.
b) Frey, H. M. AdVances in Photochemistry; Noyes, W. A., Hammond, G.
S., Pitts, J. N., Eds.; Wiley & Sons: New York, 1966; Vol. 4, p 225.
(3) (a) LaVilla, J. A.; Goodman, J. L. Tetrahedron Lett. 1990, 31, 5109.
(b) Moss, R. A.; Liu, W. J. J. Chem. Soc., Chem. Commun. 1993, 1597.
(4) (a) Modarelli, D. A.; Morgan, S.; Platz, M. S. J. Am. Chem. Soc.
1992, 114, 7034. (b) Modarelli, D. A.; Platz, M. S. J. Am. Chem. Soc. 1991,
113, 8985.
(
(
0
002-7863/96/1518-3829$12.00/0 © 1996 American Chemical Society

3
830 J. Am. Chem. Soc., Vol. 118, No. 16, 1996
Bonneau et al.
Figure 1. (a) Plots of the ratio [3]/[2] Vs [TME] for n-propylchlorodiazirine + TME in isooctane, under photolysis at 10 °C (*), 20 °C (O), 30 °C
(
]), 40 °C (4) and 55 °C (0) and under thermolysis at 79 °C (b), 84 °C ([) 90 °C (2), and 97 °C (9). (b) Plots of the ratio [2]/[3] Vs 1/[TME]
for the same system, with same conditions and symbols.
possible from the excited state of a diazirine.5 Although the
only paths considered were those leading to (carbene + N2)
and to a diazo compound as products, the complexity of the
system of energy surfaces makes quite plausible an H migration,
more or less concerted with extrusion of N2. (ii) An excited
state of the carbene may also, from an energetic point of view,
be produced from the excited state of the diazirine and has been
proposed as a possible intermediate for this kind of
Scheme 3
8
Theoretical calculations give a well-defined orientation between
dihalocarbenes and olefin, even at long distance, but predict no
minimum on the free energy profile connecting the separated
species and cyclopropane. However these profiles are strongly
dependent on the nature of the carbene: at the distance
corresponding to the postulated complex, the ∆G curve present
a sharp maximum for CCl2 but is nearly flat for CBr2. It seems
possible that the same type of calculations would yield a
mimimum for alkylchlorocarbenes.
6
a
rearrangement. (iii) An excited state of a diazo intermediate,
the adiabatic formation of which is thermodynamically feasible.
The photolysis of diazirines gives, with variable efficiencies,
diazo compounds. In the case of chlorodiazirines, this efficiency
is low and the chlorodiazo derivatives are unstable, even at low
temperature. For instance, photolysis of a benzylchlorodiazirine
in 3-methylpentane at 100 K6b or in argon matrix at 10 K
gives “traces of benzylchlorodiazomethane”, observed by IR
spectroscopy, but this diazo compound rapidly disappears at
6c
9
6
b
In 1987, M. T. H. Liu used the COC mechanism to explain
2
00 K. Similar results were reported for the photolysis of
6d
the curvature of the plots of [3]/[2] Vs [olefin] obtained under
thermolysis and photolysis of benzylchlorodiazirines in the
presence of various olefins. In this mechanism, the COC could
be a “π-complex”, with bonding interactions between the empty
orbital of the carbene and the filled olefinic π-orbital and/or
between the filled orbital of the carbene and the empty olefinic
π*-orbital. Furthermore, because the temperature dependence
of the ratio (ki/k1) measured under photolysis and thermolysis
gives a single Arrhenius line, Liu concluded that the RIES
mechanism was inefficient in the case of benzylchlorocarbenes.
phenylchlorodiazirine in argon matrix at 10 K. One of the
referees suggested that this decomposition of the diazo produces
2
: this pathway would bypass the carbene in a similar manner
as the RIES process, but it could still be operating in thermolysis
if the thermolysis of diazirines produces diazo ground state).
(
However, the global yield of formation of 2 Via the diazo ground
state pathway is probably negligible for alkylchlorodiazirines
since (i) the reported “traces of benzylchlorodiazomethane”
probably represent a yield of formation lower than 5 or 10%
and (ii) the thermal decomposition of a diazo (possibly acid
catalyzed) would give not only the rearranged product (2) but
also the carbene (4) which could even be the major product.
With the olefin being tetramethylethylene (TME) and â )
k2/(k′i + k2) being the yield of formation of 3 from the COC,
this mechanism gives
The COC mechanism, represented in Scheme 3, has been
7a
7b
7c
suggested by Skell et al., Turro et al., and Tomioka et al.
[
2]/[3] ) k′ /k + [k /(âk )]/[TME] )
i 2 i 1
(5) (a) Yamamoto, N.; Bernardi, F.; Bottoni, A.; Olivucci, M.; Robb,
M. A.; Wilsey, S. J. Am. Chem. Soc. 1994, 116, 2064. (b) M u¨ ller-Rammers,
P. L.; Jug, K. J. Am. Chem. Soc. 1985, 107, 7275. (c) Bigot, B.; Ponec, R.;
Sevin, A.; Devaquet, A. J. Am. Chem. Soc. 1978, 100, 8575.
(
1 - â)/â + [k /(âk )]/[TME]
i 1
(
6) (a) Warner, P. M. Tetrahedron Lett. 1984, 25, 4211. (b) White, W.
With both the RIES and COC mechanisms being efficient,
R.; Platz, M. S. J. Org. Chem. 1992, 57, 2841. (c) Wierlacher, S.; Sander,
W.; Liu, M. T. H. J. Am. Chem. Soc. 1993, 115, 8943. (d) Ganzer, G. A.;
Sheridan, R. S.; Liu, M. T. H. J. Am. Chem. Soc. 1986, 108, 1517.
(8) Houk, K. N.; Rondan, N. G.; Mareda, J. J. Am. Chem. Soc. 1984,
106, 4291 and 4293.
(7) (a) Skell, P. S.; Cholod, M. S. J. Am. Chem. Soc. 1969, 91, 7131.
(
b) Turro, N. J.; Lehr, G. F.; Butcher, J. A., Jr.; Moss, R. A.; Guo, W. J.
(9) (a) Liu, M. T. H.; Soundararajan, N.; Paike, N.; Subramanian, R. J.
Org. Chem. 1987, 52, 4223. (b) Liu, M. T. H.; Suresh, R. V.; Soundararajan,
N.; Vessey, E. G. J. Chem. Soc., Chem. Commun. 1989, 12.
Am. Chem. Soc. 1982, 104, 1754. (c) Tomioka, H.; Hayashi, N.; Izawa, Y.;
Liu, M. T. H. J. Am. Chem. Soc. 1984, 106, 454.

Rearrangement of Alkylchlorocarbenes
J. Am. Chem. Soc., Vol. 118, No. 16, 1996 3831
Scheme 4
Scheme 4 yields the relation
[
2]/[3] ) (1 - â)/â + k* /(âk ) + [k /(Râk )]/[TME]
i C i 1
Thus, in all cases, the plot of [2]/[3] Vs 1/[TME] must be linear
but (i) the slopes of the plots obtained under thermolysis
(ki/(âk1)) and photolysis (ki/(Râk1)) must differ by a factor R
and (ii) the intercepts, ([2]/[3])∞ (value of [2]/[3] when [TME]
f ∞), must be larger under photolysis than under thermolysis,
by an amount equal to (k*i/(âkC)) ) (1 - R)/Râ.
This paper presents the results obtained by time-resolved
photoacoustic calorimetry as well as by thermolysis and
photolysis of ClCH2C(N2)Cl and CH3(CH2)2C(N2)Cl in the
presence of TME. These results clearly indicate that both the
RIES and COC mechanisms play a role, but with efficiencies
which depend on the nature of the considered diazirine.
Reexamination of the results previously obtained with benzyl-
chlorodiazirines confirms that, for this class of diazirines, the
nonlinearity of the plots [3]/[2] Vs [TME] is mainly due to the
COC mechanism. But it also reveals that a RIES mechanism,
temperature-dependent and poorly efficient (especially at room
temperature and below), cannot be neglected.
Experimental Section
Diazirines XCH
(1d), p-ClC
2
C(N
2
)Cl with X ) H (1a), Cl (1b), CH
3 2
CH (1c),
(1f) were prepared by the
C H
6 5
6
H
4
(1e), and p-MeC H
6 4
1
0
hypochlorite oxidation of the corresponding amidines. Tetrameth-
ylethylene was purchased from Aldrich and used without purification.
Irradiation was carried out with 350 nm UV lamps in a Rayonet photo-
reactor until all the diazirine was destroyed. The photoacoustic
1
1
apparatus has been previously described.
Details on the analysis and identification of the products and on the
photoacoustic measurements are given as Supporting Information.
Results
Photolysis and Thermolysis. The relative amounts of
products obtained under photolysis and thermolysis of diazirines
1
b and 1c in the presence of various concentrations of TME at
several temperature are given in Tables 1 and 2 as ratios of
9
[
3]/[2] and [2-E]/[2-Z]. Previously published data were used
for comparing diazirines 1d-f with diazirines 1b and 1c.
Photoacoustic. Photoacoustic calorimetry (PAC) monitors
the amplitude and time evolution of heat depositions in a system
following photoexcitation. This information can provide kinetic
(
10) Graham, W. H. J. Am. Chem. Soc. 1965, 87, 4396.
(11) (a) LaVilla, J. A.; Goodman, J. L. J. Am. Chem. Soc. 1989, 111,
6
7
877. (b) LaVilla, J. A.; Goodman, J. L. J. Am. Chem. Soc. 1989, 111,
12.

3
832 J. Am. Chem. Soc., Vol. 118, No. 16, 1996
Bonneau et al.
Table 3. Experimental Enthalpies and Rate Constants Determined
by PAC
a,b
k (× 10 s-1)
7
diazirine
∆H a
1
∆H a
2
∆H
T
1a
1b
1d
-26.1
-39.9
+3.4
-39.9
-28.8
-55.1
-66.0
-68.7
-55.1
0.14
2.0
9.1
a
Values are in kcal/mol. b Total enthalpy of reaction, ∆H
T
) ∆H
1
+
∆H
2
.
Figure 2. Energetic diagram explaining the principle of the measure-
ments of the RIES efficiency by time resolved-photocacoustic calo-
rimetry.
and enthalpic information about transient intermediates.1
1-14
The
details of this method have been previously reported.1
1,12
Photoexcitation of diazirine (1) initially yields the excited
state (1*), which either directly or via another intermediate
rapidly forms (<10 ns) the singlet carbene (4) by loss of nitrogen
and vinyl halide (2) via the RIES process (with 1,2 hydrogen
rearrangement and loss of nitrogen). Experimentally, photo-
excitation of 1a, 1b, and 1d in heptane results in two heat
depositions: the first one, ∆H1, is fast (τ < 5 ns) and reflects
carbene and vinyl halide formation from the diazirine; the
second, ∆H2, is slower (τ ) 1/k) and is due to vinyl halide
formation by rearrangement of the carbene. The PAC reaction
enthalpies can be determined from ∆H1 ) (1 - a1) Ehν and
∆
H2 ) -a2Ehν, where Ehν is the incident photon energy, 84.8
kcal/mol, and an is the fraction of incident photon energy lost
to the solution in the given heat deposition.
Deconvolution of the PAC waveforms yields the values of
∆H1, ∆H2, and k which are given in Table 3. The rate constants
for the 1,2-H rearrangement of 4a and 4d obtained by PAC are
similar to those determined by nanosecond absorption spec-
troscopy.
Assuming that the overall quantum efficiency for the forma-
tion of 2 is unity, then 1* can form the carbene 4 and the vinyl
halide 2 with efficiencies R and (1 - R) respectively (Figure
2
∆
). The value of R can be determined using the equation R )
H2/∆Hr (4 f 2), which is simply the ratio of the obtained
(
12) (a) Rudzki, J. E.; Goodman, J. L.; Peters, K. S. J. Am. Chem. Soc.
985, 107, 7849. (b) Westrick, J. A.; Goodman, J. L.; Peters, K. S.
Biochemistry 1987, 22, 8313.
13) (a) Heihoff, K.; Braslavsky, S. E.; Schaffner, K. Biochemistry 1987,
2, 1422. (b) Kanabus-Kaminska, J. M.; Hawari, J. A.; Griller, D.;
1
(
2
Chatgilialoglu, C. J. Am. Chem. Soc. 1987, 109, 5267. (c) Mulder, P.;
Saastad, O. W.; Griller, D. J. Am. Chem. Soc. 1988, 110, 4090. (d) Ni, T.;
Caldwell, R. A.; Melton, L. A. J. Am. Chem. Soc. 1989, 111, 457.
(14) PAC measures the total volume change of the associated reaction.
This includes thermal, ∆Vth, and reaction, ∆Vrx, volume changes. In the
reaction (1 f 4), ∆Vrx should be posittive which would give an erroneously
large value for ∆H1. However, ∆Vrx should be small for the reaction (4 f
2) so that ∆H2 should accurately reflects the thermal volume change.

Rearrangement of Alkylchlorocarbenes
J. Am. Chem. Soc., Vol. 118, No. 16, 1996 3833
r
Table 4. Calculated ∆H (4 f 2) and Efficiency of Carbene
Formation
diazirine
∆H
f
(4)a,b
∆H
f
(2)b
∆H
r
(4 f 2)b
R
c
1a
1b
1d
61.2
59.6
87.2
8.6
1.1
-52.6
-58.5
-58.8
0.76
0.49
0.94
c
a
28.4
a
PM3 values obtained using Spartan, ver. 3.1, Wavefunction, Inc.,
b
c
1
994. Values are in kcal/mol. See: Cox, J. D.; Pilcher, G. Thermo-
chemistry of Organic and Organometallic Compounds; Academic
Press: London, 1970.
PAC enthalpy to that of the known heat of reaction. Unfortu-
nately, ∆Hr (4 f 2) has not been determined for these carbenes
but can be calculated by using the values of ∆Hf (2) and ∆Hf
(4) given in Table 4. When not available from the literature,
the values of ∆Hf were calculated using the PM3 method which
usually gives results in agreement with experimental values
within a few kcal/mol. The value of R for 1a, 0.67, is similar
to that obtained previously3a using TME as trapping agent for
∞
Figure 3. Values of ([2]/[3]) Vs temperature for the diazirines 1b (4
and 2), 1c (O and b), and 1f (0 and 9) under thermolysis (open
the carbene and suggest that the value used for ∆Hr (4 f 2) is
reasonable.
symbols) and photolysis (closed symbols). These values of the ratio
[2]/[3] when [TME] f ∞ are obtained from the intercepts of plots
such as shown in Figure 1b.
Interpretation and Discussion
Determination of the RIES Efficiency. The PAC method
gives R ) 0.49, 0.76, and 0.94 for ClCH2C(N2)Cl, CH3C(N2)Cl,
and C6H5CH2C(N2)Cl, respectively (see Table 4), indicating that
the efficiency of the RIES process, ΦRIES ) (1 - R), greatly
depends on the structure of the diazirine. The exact value of R
depends on the value of ∆Hr (4 f 2) which can be calculated
from the values of ∆Hf (2) and ∆Hf (3), obtained by theoretical
chemistry methods.
The value ΦRIES ) 0.24 thus obtained for 1a is similar to
the 0.33 yield of vinyl chloride found3 under photolysis of 1a
a
+
TME when [TME] f ∞. The difference may be due to some
vinyl chloride produced, in the later case, via the COC
mechanism, but this clearly indicates the limits of the PAC
method when the value of ∆Hr is not experimentally well
determined by some reliable experimental method. It is
therefore of interest to compare the values of R deduced from
PAC measurements with those which may be obtained from
product analysis.
Figure 4. Arrhenius plots for the diazirines 1b (4 and 2), 1c (O and
b) and 1e (0 and 9, upshifted by 0.3 unit) under thermolysis (open
symbols) and photolysis (closed symbols). The “slope” used on the
Y-axis is the slope of the plots of [2]/[3] Vs 1/[TME], equal to
It has been shown above that the slopes of the plots of [2]/[3]
Vs 1/[TME] obtained under photolysis (ki/(Râk1)) and thermoly-
sis (ki/(âk1)) must differ by a factor R and that the intercepts of
these plots, ([2]/[3])∞, must be larger under photolysis than under
thermolysis, by an amount equal to (k*i/(âkC)) ) (1 - R)/(Râ).
This provides two methods to estimate the value of R, but since
photolysis and thermolysis are usually not performed to the same
temperature, measurements must be made at various tempera-
tures for both methods and the results should then be extrapo-
lated to a common temperature.
The value of ([2]/[3])∞ obtained by photolysis and thermolysis
of diazirines 1b, 1c, and 1f in the presence of TME are plotted
Vs the temperature in Figure 3. There is no measurable
difference between the values under photolysis and thermolysis
in the case of 1f, but differences around 0.96 and 0.50 are
obtained for 1b and 1c, respectively. With the values of the
intercept under thermolysis, ([2]/[3])∞ ) (1 - â)/â, this method
gives
(k
i
/(âk
1
)) under thermolysis and to (k
i
1
/(Râk )) under photolysis (see
equations related to Scheme 4).
difference between the slopes of the plots [2]/[3] Vs 1/[TME]
measured under photolysis (ki/(Râk1)) and thermolysis (ki/(âk1)).
But since the processes related to ki and k1 have quite different
activation energies, these slopes are strongly temperature
dependent. The shift between the corresponding Arrhenius lines
must be equal to log R. This is shown in Figure 4 where the
log of the slope of the plots [2]/[3] Vs 1/[TME] is plotted Vs
1
/T, for 1b, 1c, and 1e.
For ClCH2C(N2)Cl and CH3(CH2)2C(N2)Cl, the Arrhenius
line obtained under photolysis is shifted with respect to the line
obtained under thermolysis ; the amplitude of this shift gives R
≈
0.42 and 0.62, respectively. On the contrary, the efficiency
of the RIES process must be very small for ClC6H4CH2C(N2)-
Cl (and other benzylchlorodiazirines) because data obtained
under photolysis and thermolysis fall on the same line as
â ≈ 0.85
â ≈ 0.62
â ≈ 0.77
and
and
and
R ≈ 0.54 for ClCH C(N )Cl
2
2
9
a
previously noticed.
R ) 0.72 for CH (CH ) C(N )Cl
3
2 2
2
On the Existence and the Properties of a COC. The RIES
process can explain, at least in part, the non linear plots of [3]/[2]
Vs [olefin] obtained under photolysis of some diazirines in the
presence of olefins, but this process cannot be invoked for data
R ≈ 1.00 for ClC H CH C(N )Cl
6
4
2
2
Similar values can be obtained for R by considering the

3
834 J. Am. Chem. Soc., Vol. 118, No. 16, 1996
Bonneau et al.
obtained by thermolysis. In fact, under thermolysis, the amount
of 2 produced from the COC and responsible for the curvature
is usually very small, compared to the total amount of 2, because
the 1,2 rearrangement of the free carbene is very fast at high
temperature. Consequently, in many cases, the curvature of the
plots of [3]/[2] Vs [olefin] is rather faint and the accuracy of
the value of ([2]/[3])∞ is rather poor. But the thermolysis of
the system CH3(CH2)2C(N2)Cl + TME gives plots of [3]/[2]
Vs [TME] which are clearly curved (see Figure 1) and the
intercept of the plots [2]/[3] Vs 1/[TME], ∼0.618 ( 0.010,
differs from zero by an amount which is far beyond all possible
experimental incertitudes.
For benzylchlorocarbenes, it seems well established that the
efficiency of the RIES process is either zero or, with a
reasonable limit for the incertitude on experimental measure-
ments, less than 10-15%.
Under photolysis of benzylchlorodiazirines in the presence
of some olefins, the curvature of the plots of [3]/[2] Vs [olefin]
is very clear, far beyond possible experimental errors, yielding
for instance values of ([2]/[3])∞ which are around 0.3, 0.6, 2.0,
and 3.0 for the photolysis at 24 °C of 1d with TME,
R-chloroacrylonitrile (Cl-ACN), diethyl fumarate (DEF), and
diethyl mesaconate (MES), respectively.15 Such large values
of ([2]/[3])∞ corresponding to 23, 37, 66, and 75% of 2 cannot
be explained by a RIES process with an efficiency < 10-15%.
Furthermore, the RIES process should give the same amount
of 2 whatever the olefin is so that ([2]/[3])∞ should be
independent of the nature of the olefin unless the RIES
efficiency is dependent on the solvent polarity. This does not
seem to be the case since there is no apparent correlation
between the values of ([2]/[3])∞ and the polariries of TME, Cl-
ACN, DEF, and MES.
Figure 5. Relations between the values of the ratios [2-E]/[2-Z] and
[3]/[2] for the system diazirine 1b + TME under photolysis at 18 °C
(O), 30 °C (]), 45 °C (4), and 52 °C (0) and under thermolysis at 75
°C (9), 95 °C ([), 102 °C (2), and 110 °C(b).
Scheme 5
This effect may be analyzed quantitatively by using Scheme
5
and expressing the yield of formation of the cyclopropane,
Φ(3) ) R{k1[TME]/(ki + k1[TME])}â, and the yield of formation
of 2-Z, Φ(2-Z) ) (1 - R)z* + R{ki/(ki + k1[TME])}z + R (1-
â){k1[TME]/(ki + k1[TME])}z′ where the three terms correspond
respectively to 2 produced by RIES, from the free carbene and
from the COC. These two expressions yield Φ(2-Z) ) Rz + (1
Another demonstration of the existence of a COC (as well
as another confirmation of the efficiency of the RIES process)
may be obtained from the changes in the distribution of the E
(trans) and Z (cis) isomers of the rearrangement product, 2-E
and 2-Z, as a function of the concentration of the olefin reactant.
These changes are indeed quite large for the system TME +
ClCH2C(N2)Cl.
-
R)z* - {z′ + (z - z′)/â}Φ(3) under photolysis and Φ(2-Z) )
z - {z′ + (z - z′)/â}Φ(3) under thermolysis since then R ) 1.
The plots of [2-Z] Vs [3] must therefore be straight lines for
thermolysis and photolysis, and these lines should have the same
slope. The values of [2-Z] and [3], obtained from the (3/2)
and (2-E/2-Z) ratios by assuming that Φ(2) + Φ(3) ) 1 and Φ(2)
In the absence of olefin reactant, the thermolysis of ClCH2C-
(N2)Cl gives a mixture of (Z)- and (E)-dichloroethylene with
an E/Z ratio ≈ 0.065, (nearly) independent of the temperature
on the range (75-110 °C). Therefore the rearrangement of the
free carbene ClCH2CCl gives ∼6% of 2-E and ∼94% of 2-Z.
)
Φ(2-Z) + Φ(2-E), are plotted on Figure 6: the slopes are nearly
identical (-0.954 (phot) and -0.990 (therm) and the very small
difference, if significant, may be due to a small change of one
(or several) of the parameters z, z′, and â with the temperature.
The slope of the line obtained from thermolysis, -0.99 )
-{z′ + (z - z′)/â} and the value â ) 0.85 obtained from ([2]/
Photolysis of the same diazirine in the absence of olefin
reactant also gives dichloroethylene with an E/Z ratio (nearly)
independent of the temperature on the range (0-52 °C), but
this ratio is now equal to 0.18, i.e. ∼15% of 2-E and ∼85% of
2
-Z. This is explained by the RIES process which produces 2
[3]) ) 0.175 yield z′ ) 0.64. Alternatively, the value of the
∞
with a much larger E/Z ratio than the free carbene. If the
efficiency of RIES is ∼52%, then the RIES process gives ∼25%
of 2-E and ∼75% of 2-Z.
E/Z ratio when [TME] f ∞, (E/Z)∞ ≈ 0.70 (obtained by
nonlinear regression of E/Z Vs [TME]), gives z′ ≈ 0.59 which,
with the above expression of the slope, yields â ) 0.875.
The behavior of the system ClCH2C(N2)C + TME under
photolysis and/or thermolysis on the temperature range 10-
110 °C can be fully described by the set of parameters given in
Table 5. For CH3(CH2)2C(N2)Cl, a similar analysis yields the
data shown on the line 2 of the same table.
Upon addition of TME, both the ratios 2-E/2-Z and 3/2
increase, and as shown in Figure 5, there is a clear correlation
between these changes, independent of the temperature but
strongly dependent on the method, photolysis or thermolysis,
used to drive the reaction.
The E/Z ratio appears to be very sensitive to the nature of
the species undergoing the rearrangement. We can use this
sensitivity to determine the efficiency of the RIES process in
benzylchlorodiazirines.This efficiency, found as very small from
the analysis of the values of 2/3 or equal to 6% ((6%) from
the PAC measurements, is in fact nonnegligible.
A common intermediate must be responsible for the produc-
tion of the cyclopropane 3 and for the formation of dichloro-
ethylenes 2 with a high percentage of E isomer. In this common
intermediate, the “carbene-olefin complex”, the perturbations
of the properties of the carbene moiety by the interactions with
the olefin change the ratio E/Z drastically.
The plot of (Z/E)0, the value of the ratio [2-Z]/[2-E] in the
absence of olefin, as a function of the temperature for 3-
(15) Liu, M. T. H.; Bonneau, R. J. Am. Chem. Soc. 1990, 112, 3915.

Rearrangement of Alkylchlorocarbenes
J. Am. Chem. Soc., Vol. 118, No. 16, 1996 3835
Figure 6. Relations between the relative amount of 2-Z, i.e [2-Z]/([2] + [3]), and the relative amount of 3, i.e [3]/([2] + [3]), produced by
photolysis of the system diazirine 1b + TME at 10 °C (O), 18 °C (4), 30 °C (0), 45 °C (]) and 52 °C (X) and by thermolysis at 75 °C (b), 95
°
C (9), 102 °C ([), and 110 °C (2). The relative amount of 2-E is also shown with common symbols for the four thermolysis (O) and the five
photolysis (0) data sets.
Table 5. Values of Φ(RIES) ) (1 - R), Φ(COCf3) ) â, and of the 2E:2Z distributions for 2 Generated from RIES, Carbene, and COC (from
data obtained by photolysis of several diazirines + TME)
TME + diazirine:
Φ
(RIES) ) (1 - R)
0.52 ( 0.06
2-E:2-Z (RIES)
2-E:2-Z (carbene)
Φ
(COCf3) ) (â)
2-E:2-Z (COC)
∼40:60
1b
1c
1e
1f
∼25:75
∼44:56
∼50:50
∼50:50
6:94
28:72
0.85 ( 0.03
0.62 ( 0.03
a
0.28 ( 0.05
∼20:80
b
c
∼0.05
∼88:12 (90 °C)
∼0.80
∼78:22
b
c
∼0.05
∼89:11 (90 °C)
∼0.80
∼83:17
a
From photolytic data. b At 20 °C. At 90 °C.
c
(
R) and (ii) by RIES, for a fraction 1 - R, with a much larger
Z/E ratio.
The value R ) 0.94 given by the PAC measurements yield
a 92:8 distribution of Z and E isomers for 2 produced by RIES,
which does not seem quite reasonable; RIES is expected to give
a distribution of isomers which is less selective than the
rearrangement of the free carbene (as in the case of 1b and 1c);
there seems to be no reason for RIES to produce the isomers of
2
with a large selectivity reversed with respect to that given by
the free carbene! Assuming a distribution around 50:50 for the
Z and E isomers of 2 produced by RIES, one gets R ) 0.88.
Most probably, the real value of R at 60 °C is in the range of
0
.80-0.95, i.e. R ) 0.87 ( 0.07. With such a value, the shift
between the lines “log(ki/kt) Vs 1/T” obtained under photolysis
and thermolysis, as shown in Figure 4, would be only 0.06 which
can easily escape observation. But a 10% efficiency for RIES
should be detected on the plots of ([2]/[3])∞ Vs T such as those
shown on Figure 8.
Figure 7. Variations, with the temperature, of the ratio [2-Z]/[2-E]
for the decomposition of diazirines 1d (2), 1e (0), and 1f (b), by
photolysis (below 60 °C) and by thermolysis (above 60 °C), in the
absence of TME.
For the systems 1e or 1f + TME, when the values of ([2]/
[
3])∞ are plotted Vs T on an enlarged scale, the values obtained
under thermolysis show a large scatter, but they are more or
less independent of the temperature whereas those obtained
under photolysis show a clear tendency to increase with
temperature and are significantly larger than the former when
extrapolated to the range of temperatures used for thermolysis.
In the case of “1e + TME” for instance, the ratio ([2]/[3])∞ is
around 0.25 under thermolysis (76-100 °C) whereas it is ∼0.28
at 13.5 °C, 0.33 at 30 °C, and 0.375 at 60 °C under photolysis.
With [2]∞ ) (1 - R) + R(1 - â) and â ) 0.8, these values of
([2]/[3])∞ yield R ) 0.97 at 13 °C, 0.94 at 30 °C, and 0.90 at
60 °C, in good agreement with R ) 0.94 at 30 °C, obtained
from PAC measurements, and R ) 0.88 at 60 °C found above
from the ratio Z/E by assuming that RIES gives a 50:50
distribution for 2-Z:2-E. Therefore it seems that, for benzyl-
benzylchlorodiazirines, XC6H4CH2C(N2)Cl with X ) H, Cl, and
Me, is given in Figure 7. The ratio (Z/E)0 slightly increases
with temperature as well under photolysis as under thermolysis,
and for a given temperature, it is larger under photolysis than
under thermolysis.
Under thermolysis at 60 °C for instance, (Z/E)0 ≈ 0.10,
corresponding to a 91:9 distribution of the E and Z isomers,
characteristic of the free carbene. Under photolysis, (Z/E)0 ≈
0.16 (when extrapolated to the same temperature) corresponding
to a 86:14 distribution; the relative amount of the Z isomer is
thus nearly twice larger than under thermolysis. This can only
be explained by assuming that, under photolysis, 2 is produced
(i) from the free carbene with a 9:91 distribution, for a fraction

3
836 J. Am. Chem. Soc., Vol. 118, No. 16, 1996
Bonneau et al.
Figure 9. Relations between the relative amounts of 2-Z (b and 9)
and 2-E (O and 0) and the relative amount of 3 produced by photolysis
Figure 8. Temperature dependence of the ratio [2]/[3] when [TME]
f ∞ for diazirine 1f under photolysis (T < 60 °C) and thermolysis
(0 and 9, 4 temperatures) and by thermolysis (O and b, three
temperatures) of the system diazirine 1e + TME.
(T > 60 °C): (b) from intercepts of linear plots of [2]/[3] Vs 1/[TME]
and (0) from nonlinear regression of [3]/[2] Vs [TME].
olefins to yield ultimately 3 after N2 loss. However, as stated
in the Introduction, the yield of formation of such diazo
intermediate from chlorodiazirines at ambient temperatures is
very small and the yield of 2 produced in this manner must be
even smaller. We are confident that this pathway alone cannot
account for the amount of 2 predicted to be produced in
thermolysis of solutions where the concentration of olefin would
be infinite, i.e. ca. 15% for 1b, ca. 20% for 1d-f or nearly
chlorodiazirines, the efficiency of the RIES process increases
with temperature.
It may be noted that this temperature dependence of the
efficiency of the RIES process is not observed with ClCH2C-
(
N2)Cl: the values of ([2]/[3])∞ obtained under photolysis are
remarkably constant over the whole temperature range 0-50
C. This may be explained by the fact that, in the case of 1b,
the rates for RIES and for carbene formation are similar (R ≈
.5), so that these two processes which most probably originate
°
4
0% for 1c. Our results certainly do not prove the existence of
COC, but this elusive species can best explain the data.
0
1a
Recently, Moss stated “... there are two presently unsolVed
from the same excited state of the diazirine must have similar
temperature dependence.
questions that temper our conclusions. These are (1) the
possible interVention of carbene/alkene complexes, and (2) the
extent of carbene-mimetic reactions that occur Via the excited
states of the carbene precursors...”. The choice between these
two alternatives has been a matter of much controversy. Our
present experimental results provide evidence that both mech-
anisms, COC and RIES, are operating, but with relative
efficiencies depending on the system studied.
For the system “1f + TME” under thermolysis, the plots of
[
2-E] and [2-Z] Vs [3] give straight lines with the following
equations: [2-E] ) 0.886 - 0.899[3] and [2-Z] ) 0.114 -
.101[3]. The values of [2-E] and [2-Z], calculated with [3] )
and [3] ) [3]∞ ) 0.8, yield respectively (E/Z)carb ) 7.77 and
E/Z)COC ) 5.0. Similar values are obtained for the system 1e
TME, as shown in Table 5.
0
0
(
+
The efficiency of the RIES process, ΦRIES, has been previ-
The overall E/Z ratio measured under thermolysis changes
ously examined and correlated to the R-CH bond dissociation
little when [TME] increases, because the E:Z distribution for 2
produced from the carbene and from the COC are similar,
whereas the small amount of RIES, increasing with temperature,
is sufficient to modify noticeably the E/Z ratio measured under
photolysis because the E:Z distribution for 2 produced by RIES
is quite different. This clearly appears on the plots of [2-E]
and [2-Z] Vs [3], in Figure 9, where the amount of [2-Z]
obtained under photolysis is always ∼4% higher than under
thermolysis, whatever the amount of 3 is, i.e. whatever the
concentration of TME is. These 4% represent in a first
approximation the amount of 2-Z produced by RIES with an
efficiency ∼50% whereas both the free carbene and the COC
produce mainly the 2-E isomer.
1b,6b
energy;
the stronger the R-CH bond, the greater yield of
carbene, and presumably the lower the RIES efficiency. The
relative yields for carbene formation from a series of alkyl-
chlorodiazirines, estimated from the amplitude of the pyridinium
ylide absorption, were found to be 1, 0.52, 0.38, 0.29, and 0.02
when the alkyl group is cyclopropyl or tert-butyl, methyl,
benzyl, ethyl, and isopropyl, respectively. Since the quantum
efficiency for carbene formation from tert-butylchlorodiazirine
3b
was measured to be ∼0.90, these yields correspond to ΦRIES
values of 0.10, 0.53, 0.66, 0.74, and 0.98. The general trend
shown by these numbers does correlate with the R-CH bond
dissociation energy, but the absolute values are not in good
agreement with those given herein. For methyl- and benzyl-
chlorodiazirines, PAC measurements and product analysis gives
ΦRIES values less than 0.33 and 0.05, respectively, significantly
less than those observed by the ylide trapping method. In
addition, experiments on chloromethyl- and propylchlorodiaz-
irines suggest lower values for ΦRIES than obtained by the ylide
method. The reason for these discrepancies is not immediately
Conclusion
The features explained by a COC such as the branching ratios
([2]/[3]) and (2-E/2-Z) could also be rationalized by assuming
that there are two types of “encounter complexes” between a
carbene and an olefin, one giving only (2 + olefin) and the
other producing only the cyclopropane 3. However it seems
reasonable and economical to assume that there is only one type
of encounter complex.
1
6,17
apparent
but it seems that the relation between ΦRIES and
the dissociation energy of the R-CH bond should be considered
with care.
A diazo compound could also explain the data if it yields 2
directly (i.e. without formation of carbene) and can react with
The efficiency of the carbene rearrangement, ΦCOCf2 ) 1 -
â, can be obtained by substracting ΦRIES from the yield of

Rearrangement of Alkylchlorocarbenes
J. Am. Chem. Soc., Vol. 118, No. 16, 1996 3837
Table 6.a Yields of 2 Generated from the COC, ΦCOCf2
Calculated from the Ratio [2]/[3] , in the Photolysis of Diazirine 1d
in the Presence of Various Olefins
,
rearranged product, Φ2∞, which is obtained from the ratio
∞
([2]/[3])∞. It seems difficult to propose a relation between this
efficiency and the nature of both the carbene and the olefin since
there are only a few relevant available data. For TME and
methylchlorocarbene, ΦCOCf2 ) 0.09 using ΦRIES ) 0.24 and
olefin
TME
BVE
HEX
MES
DEF
Cl-ACN
[2]/[3]∞
[2]
0.30
23%
0.18
0.44
30%
0.25
0.52
34%
0.29
3.15
76%
0.71
2.0
67%
0.62
0.64
39%
0.34
(
[2]:[3])∞ ) 67:33.3a This value is similar to those found for
Φ
COCf2
alkylchlorocarbenes 1b, 1d, 1e, and 1f with the same olefin
which have ΦCOCf2 values around 0.2.
The ΦCOCf2 values for benzylchlorocarbene and various
a
Φ2-COC calculated with ΦRIES ) 0.05 at 25 °C: BVE, n-butyl vinyl
ether; HEX, 1-hexene; MES, diethyl mesaconate; DEF, diethyl fuma-
rate; Cl-ACN, R-chloroacrylonitrile.
9
a
olefins, calculated by using previously published data, are
given in Table 6. No apparent correlation is found between
ΦCOCf2 and the rate constant of trapping of the carbene by the
olefin or between ΦCOCf2 and the “π-molecular electronega-
tivity” of the olefin. Most of the olefins, including very reactive
TME, unreactive 1-hexene, electron deficient chloroacrylonitrile,
and electron rich TME give ΦCOCf2 around 0.2 or 0.3.
However, olefins diethyl fumarate and diethyl mesaconate have
high ΦCOCf2 values: 0.62 and 0.71, respectively. The reason
for these large values is not obvious, but might be due to the
interaction of the electrophilic carbene with the ester groups
which does not lead to product but rather to rearrangement or,
more probably, to a steric effect of the two ethyl ester
substituents in a trans position on the double bond. This steric
factor would not affect the rate constant for trapping of the
carbene by the olefin (i.e. the rate constant of formation of the
COC) if the carbene-olefin distance is relatiVely long in the
COC, but it might be quite decisive when the carbene-olefin
distance must be strongly reduced to form the cyclopropane
adduct.
(16) In fact, in the Figure 3 of ref 6b, White and Platz use a “normalized”
value of the ylide absorption which greatly improves the correlation between
the yield of carbene formation and the dissociation energy of the R-CH
bond. Normalization is “justified” by expected change in the value of the
absorption coefficient between pyridinium ylides of benzylchlorocarbene
and other alkylchlorocarbenes, but this yields an even more unrealistic value
of the quantum yield of carbene formation, ∼4%!
Supporting Information Available: Details of the synthesis
of diazirines, irradiation and product analysis, and product
studies (2 pages). This material is contained in libraries on
microfiche, immediately follows this article in the microfilm
version of the journal, can be ordered from the ACS, and can
be downloaded from the Internet; see any current masthead page
for ordering information and Internet access instructions.
(
17) The pyridine ylide method relies on the assumption that, when
[
pyridine] is large enough, i.e. ∼1-5 M, every carbene produced is trapped
by the pyridine. This is justified when the carbene lifetime is longer than
a few nanoseconds since the rate constant for ylide formation, kyl, is usually
9
-1 -1
in the range (1-5) × 10 M
s . This may be unjustified for carbenes
with a lifetime around or shorter than a nanosecond, especially if substituent-
(s) increasing the rate of the 1,2-H shift also decrease the value of kyl by
increasing the electron density on the carbene center.
JA952700N