Triplet species from dihydroipyrrolo[3,4-d]pyridazines, the diazene precursors of N-arenesulfonyl-3,4-dimethylenepyrroles

Article abstract of DOI:10.1021/ja963114c

In contrast to the photolyses of N-methyl- or N-pivaloylpyrrolo[3,4]dihydropyridazines in low-temperature matrices, which give the blue, ESR-inactive singlet 3,4-dimethylenepyrrole biradical at all wavelengths, the photolyses of the N-arenesulfonyl deriva

Full text of DOI:10.1021/ja963114c

1416
J. Am. Chem. Soc. 1997, 119, 1416-1427
Triplet Species from Dihydropyrrolo[3,4-d]pyridazines, the
Diazene Precursors of N-Arenesulfonyl-3,4-dimethylenepyrroles
Linda C. Bush,† Ljiljana Maksimovic, Xu Wu Feng, Helen S. M. Lu, and
Jerome A. Berson*
Contribution from the Department of Chemistry, Yale UniVersity,
New HaVen, Connecticut 06520-8107
ReceiVed September 4, 1996X
Abstract: In contrast to the photolyses of N-methyl- or N-pivaloylpyrrolo[3,4]dihydropyridazines in low-temperature
matrices, which give the blue, ESR-inactive singlet 3,4-dimethylenepyrrole biradical at all wavelengths, the photolyses
of the N-arenesulfonyl derivatives in this series are strongly wavelength-dependent. In three separate instances,
both singlet and triplet N-arenesulfonyl-3,4-dimethylenepyrrole biradicals have been observed. In each case, one or
two other triplets are seen by ESR spectroscopy after irradiation under special conditions. These are tentatively
assigned conformationally isomeric diazenyl biradical structures. The possible origins of the slow intersystem crossing
rates for singlet-triplet interconversions in N-arenesulfonyl-3,4-dimethylenepyrrole biradicals are examined. Near-
zero ionic character of the singlet wave function is considered to be a less likely cause than conformational control
of the spin state. The latter circumstance would require that intersystem crossing be coupled to conformational
change, which can be very slow at low temperatures. The zero-field splittings determined from the ESR spectra of
N-tosyl-3,4-dimethylenepyrrole biradical are found to be temperature-dependent, which could be caused by librations
and/or internal rotational (conformational) motions of the biradical in the matrix.
In recent work,1a,c-e we applied semiempirical quantum
593 nm generated by thermolysis or direct 370 nm photolysis
of diazene 5b is a singlet state of N-tosyl-3,4-dimethylenepyrrole
1b.1b,c This apparent failure of the computationally based
predictions of an accessible triplet of 1b was a sharp dis-
appointment.
chemical calculations, calibrated on experimental observations
and high-level ab initio computations, to develop guidelines for
the construction of non-Kekule´ molecules with near-zero
singlet-triplet (S-T) energy gap. These studies led to the
prediction that the preference for a singlet ground state
characteristic of the parent 3,4-dimethylenepyrrole biradical 1a
should decrease as the hydrogen of the N-H function is replaced
by increasingly electron-withdrawing (EW) groups. The N-
methyl substituent and even the moderately EW N-pivaloyl
substituent should still preserve the singlet ground state, in
accord with the observed chemistry of the corresponding
biradicals1 and with the absence of ESR signals from matrix-
immobilized samples generated at low temperatures. However,
with an N-substituent as strongly EW as arenesulfonyl (as in
1b,c), the T-S energy gap was anticipated to be very small1a,c
and about the same as that predicted (+2 to -1 kcal/mol) by
the most recent ab initio results2 for hydrocarbon tetramethyl-
eneethane (TME) derivatives. Three such hydrocarbon TMEs
(2-4) have been shown experimentally to have accessible triplet
states which have been characterized by electron spin resonance
(ESR) spectroscopy.3 Thus, there was legitimate reason to hope
that the heterocyclic biradicals 1b,c might have observable triplet
states also.
However, further study gradually revealed a complex set of
photochemical pathways of the diazene 5b. The photochemistry
is exquisitely sensitive to the temperature and irradiating
wavelength, and eventually, we were led to the discovery of
three new triplet species, one of which has properties consistent
with those expected of the long-sought triplet state of N-tosyl-
3,4-dimethylenepyrrole 1b. The nature of the other two triplets
in the b series is discussed below. In extensions of this work,
we also have prepared the corresponding persistent singlet and
triplet states from the N-brosyl precursor 5c,1a and we have
developed a new synthesis (Scheme 1) of 2-aryl-N-tosyl-3,4-
dimethylenepyrrolodiazenes, e.g., 5d. The latter compound
serves as a suitable precursor of the biradical N-p-toluene-
sulfonyl-2-phenyl-3,4-dimethylenepyrrole 1d and now also can
be used to generate both singlet and triplet species.
(3) (a) Dowd, P.; Chang, W.; Paik, Y. H. J. Am. Chem. Soc. 1986, 108,
7416. (b) Dowd, P.; Chang, W.; Paik, Y. H. J. Am. Chem. Soc. 1987, 109,
5284. (c) Roth, W. R.; Kowalczik, U.; Maier, G.; Reisenauer, H. P.;
Sustmann, R.; Mu¨ller, W. Angew. Chem., Int. Ed. Engl. 1987, 26, 1285.
(4) (a) Further evidence of the singlet nature of the blue transient 1b in
principle would be available from solid state 13C NMR spectroscopy, as is
reported5 for several other singlet biradicals. So far, our attempts to apply
these methods to the case of 1b have encountered insuperable experimental
obstacles and are described elsewhere.6
On the other hand, the experimental evidence1 so far leaves
little doubt that the ESR-silent blue transient species of λmax
)
† Current address: Department of Chemistry, Salisbury State University,
Salisbury, MD 21801.
X Abstract published in AdVance ACS Abstracts, February 1, 1997.
(1) (a) Preliminary communication: Bush, L. C.; Heath, R. B.; Berson,
J. A. J. Am. Chem. Soc. 1993, 117, 9830. (b) Heath, R. B.; Bush, L. C.;
Feng, X. W.; Berson, J. A.; Scaiano, J. C.; Berinstain, A. B. J. Phys. Chem.
1993, 97, 13355. (c) Accompanying paper: Bush, L. C.; Heath, R. B.; Feng,
X. W.; Wang, P. A.; Maksimovic, L.; Song, A. I. H.; Chung, W.-S.;
Berinstain, A. B.; Scaiano, J. C.; Berson, J. A. J. Am. Chem. Soc. 1997,
119, 1406. (d) Lu, H. S. M.; Berson J. A. J. Am. Chem. Soc. 1996, 118,
265. (e) Lu, H. S. M.; Berson, J. A. J. Am. Chem. Soc. 1997, 119, 1428
(accompanying paper).
(5) (a) Zilm, K. W.; Merrill, R. A.; Greenberg, M. M.; Berson, J. A. J.
Am. Chem. Soc. 1987, 109, 1567. (b) Zilm, K. W.; Merrill, R. A.; Webb,
G. G.; Greenberg, M. M.; Berson, J. A. J. Am. Chem. Soc. 1989, 111, 1523.
(c) Greenberg, M. M.; Blackstock, S. C.; Berson, J. A.; Merrill, R. A.;
Duchamp, J. C.; Zilm, K. W. J. Am. Chem. Soc. 1991, 113, 2318. (d)
Reynolds, J. H.; Berson, J. A.; Kumashiro, K. K.; Duchamp, J. C.; Zilm,
K. W.; Rubello, A.; Vogel, P. J. Am. Chem. Soc. 1992, 114, 763. (e)
Reynolds, J. H.; Berson, J. A.; Kumashiro, K. K.; Duchamp, J. C.; Zilm,
K. W.; Scaiano, J. C.; Berinstain, A. B.; Rubello, A.; Vogel, P. J. Am.
Chem. Soc. 1993, 115, 8073.
(2) (a) Du, P.; Borden, W. T. J. Am. Chem. Soc. 1987, 109, 930. (b)
Nachtigall, P.; Jordan, K. D. J. Am. Chem. Soc. 1992, 114, 4743. (c)
Nachtigall, P.; Jordan, K. D. J. Am. Chem. Soc. 1993, 115, 270.
(6) (a) Bush, L. C. Ph. D. Dissertation, Yale University, New Haven,
CT, 1994. (b) Bush, L. C.; Kumashiro, K. K.; Zilm, K. W.; Berson, J. A.
unpublished work at Yale University.
S0002-7863(96)03114-9 CCC: $14.00 © 1997 American Chemical Society

Triplet Species from Dihydropyrrolo[3,4-d]pyridazines
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1417
Scheme 1a
a Methods. (1) DIBAL, CH2Cl2, 81%. (2) MeI, Ag2O, CH3CN, 85%.
(3) NBS, THF, -72 °C, 83%. (4) BuLi, THF, -78 °C, Me3SnCl, 100%.
(5) PdCl2(CH3CN)2, PhI, 2-methyl-2-pyrrolidinone, 27%. (6) Ph3PBr2,
57%. (7) (t-BuO)2CN(K)N(K)CO2t-Bu, 59%. (8) HCl, Et2O,
MeO2CNdNCO2Me, 29%.
Synthesis of Diazene 5d (Scheme 1). The first four steps
of Scheme 1 leading to the trimethylstannyl derivative 8 are
identical to those1d in the synthesis of 1,3-di-2-pyrrylbenzenes.
The path (9 f 5d) to the phenyl derivatives in Scheme 1 used
a cross-coupling with iodobenzene (instead of 1,3-diiodo-
benzene, which was used1d to enter the 1,3-di-2-pyrrylbenzene
series). Conversion to the bis-carbamate 10 and the diazene
5d followed well precedented1 steps.
Scheme 2
ESR Spectroscopic Characterization of Triplet Species
Generated by Photolysis of Diazene Precursors 5b-d.1a For
ease of reference, the three types of triplet ESR spectra
encountered in this work are designated by the relative
magnitudes of their zero-field splitting (zfs) parameters |D/hc |
as small (S), medium (M), or large (L), respectively. The
substitution pattern of the biradical and its precursor is
designated as b, c, or d, as in 1b-d and 5b-d. Scheme 2
shows the assignments of structure to the carriers of the three
types of spectra in the b series. The assignment of the widest
spectrum, Lb, to the N-tosyl-3,4-dimethylenepyrrole biradical
31b rests upon the agreement between the observed and
calculated ESR zero-field splitting (zfs) parameter |D/hc | and
also upon chemical trapping experiments, which point to an
intact N-tosyl-3,4-dimethylenepyrrole structure. The other two
triplet species, the carriers of the narrower spectra Sb and Mb,
although less firmly identified, are consistent with diazenyl
Table 1. Generation of Triplet Species by Irradiation of Diazene
5b
matrix
hν, nma
T, K
5, 77
5, 85, or 90
5
5
5
5
80
5
obsd ESR spectrum
3
3
biradical structures 11b and 12b.
EAb,c
343-345
265
Sb
Lb + Sb
Generation of Diazenyl Biradical Triplet Species 311b and
propanolsd
3
312b and of N-Tosyl-3,4-dimethylenepyrrole Biradical 1b
230-325j
265
Lb + Mb + Sbg
Lb (+ Sb?)
Lb + Mb + Sbg
Lb + Mb + Sb
Lb + Sb
Freonse
in the Photolysis of Diazene 5b (Scheme 2). The azo
chromophore of N-tosyl diazene 5b shows a broad UV-vis band
with maxima at 370, 343, and ∼300 nm. As has been
described,1 370 nm irradiation in glassy matrices at 77 K
generated singlet 1b, an ESR-silent, blue transient, λmax 593
nm. Table 1 summarizes our studies of the effects of matrix,
wavelength, and temperature on the course of the photo-
deazetation. It is convenient to divide the data into three groups
according to irradiating wavelength:
1. Irradiations at 340-345, 366 (Band-Pass Filtered), or
370 nm. These generated spectra Sb and/or Mb, depending
on temperature. Thus, ESR examination of a matrix of 5b in
EA glass (diethyl ether-ethanol) that had been irradiated briefly
at 340 nm at 77 or at 5 K showed a strong four-line signal
assignable to the ∆ms ) 1 transitions of a randomly oriented
triplet species. This spectrum, Sb, also could be seen by
irradiation at 340-345 nm of diazene 5b in MTHF glasses at
77 K (Figure 1 Sb). Neither the EA nor the MTHF preparations
from the 340-345 nm photolyses showed the blue color or the
UV-vis maximum at 593 nm characteristic of the singlet
230-325j
265
MTHFf
265
230-325j
Lb + Mb + Sbg
Mb + Sbh
366k
5
370
5, 90i
Mb + Sbh
a Methods of irradiation are described in the Experimental Section.
b Diethyl ether-ethanol (1:1 v:v). c Reference 1. d PrOH:i-PrOH (2:3
v:v). e CCl3F:CF2BrCF2Br (1:1 v:v). We thank M. S. Platz for
f
suggesting this matrix. 2-Methyltetrahydrofuran. g This preparation was
i
visibly purple. h This preparation was visibly blue. Observed by signal
accumulation (see text). j 1 M NiSO4 solution filter, main transmission
230-325 nm. We thank P. Vaccaro for suggesting the use of a solution
filter. k Glass band-pass filter with main transmission 328-381 nm.
biradical 11b. The ESR signals persisted for many hours at 77
K in the dark. Irradiation of 5b in MTHF at 366 or 370 nm
gave the Sb and Mb spectra also (see below), but the samples
were visibly blue, because of the presence of the singlet biradical
11b.

1418 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Bush et al.
accumulation of signals generated in continuous photolyses (see
Table 1). We soon noticed a second triplet spectrum, Mb,
superimposed upon spectrum Sb in some of the samples. This
was especially clear in the spectra of samples prepared by
irradiation at 5 K (Figure 1). As judged by the zfs parameter
of Mb, |D/hc| ) 0.0198 cm-1 (2D′ ) 404 G), the two odd
electrons in the carrier of Mb are closer together than those in
the carrier of the narrow triplet signal Sb. We think it is likely
that the carrier of Mb is 312b, a conformational isomer of 11b
(Scheme 2).
The proximal and distal conformations assigned, respectively,
to biradicals 11b and 12b are only schematic. They symbolize
the inferences that the C-N-N angles, by analogy to theoreti-
cally predicted8-10 angles near 120° for simple diazenyl radicals,
are significantly bent, and that the larger D value for the
proximal biradical 12b (spectrum Mb) relative to the distal
biradical 11b (spectrum Sb) results from a closer average
propinquity of the two unpaired electrons.11a Occupation of a
σ-orbital by the N-centered electron8-10 and of a π-orbital by
the C-centered electron11b also seem likely additional features
of these diazenyl biradicals.
Thermal Interconversion of the Two Diazenyl Biradical
Conformers. We have observed a thermal transformation of
the Mb spectrum into the Sb spectrum. Figure 1 shows that
short irradiation at 366 nm of an MTHF glass of diazene 5b at
5 K generates an ESR trace in which Sb and Mb are
superimposed. Raising the temperature of this sample to 40-
50 K decreases the intensity of the Mb signals, but this is not
entirely a Curie Law effect: The intensity of the Sb signal
increases at the same time, and if the temperature is allowed to
rise briefly to 75 K, the only visible triplet spectrum is the
“narrow” signal Sb. Prompt cooling of the sample back to 5
K does not restore the Mb spectrum, but the Sb spectrum now
(Figure 1Sb) is more intense than it was originally. The Sb
spectrum itself slowly fades above 70 K (see below).
Figure 1. Mb + Sb. ESR spectrum obtained by band-pass filtered
(peak transmission 328-381 nm) or monochromatic (370 ( 5 nm)
irradiation of diazene 5b in MTHF glass at 5 K. The ∆ms ) 2 region
is not shown. Sb. ESR spectrum obtained by raising the temperature
of the preparation Mb + Sb to 75 K and re-cooling to 5 K. The ∆ms
) 2 region is not shown. The same Sb spectrum is obtained by 343
nm irradiation of 5b at 5 K.
2. Irradiations at 265 nm. These generated predominantly
spectrum Lb, but Sb and Mb could be observed under proper
conditions. The samples were colorless or yellow, and the 593
1
nm band of 1b was not seen in the UV-vis spectrum.
3. Irradiations with Broad-Band Light, NiSO4-Filtered,
Main Transmission at 230-325 nm, at 5 K. These generated
all three spectra Sb, Mb, and Lb. The preparations were
colorless or yellow and showed no absorption in the 590 nm
region of the UV-vis spectrum.
We interpret these changes as a thermal conformational
3
3
isomerization of 12b to 11b (Scheme 2). Since the Sb and
Mb spectra overlap appreciably (Figure 1), one cannot be sure
that the intensity gained by the Sb spectrum through the cycle
of temperature 5 K f 75 K f 5 K corresponds exactly to the
intensity lost from the Mb spectrum. It is therefore possible
Diazenyl Biradicals 11b and 12b, Carriers of the ESR
Spectra Sb and Mb. We consider first the spectrum Sb, which
as described above, was formed cleanly in the 340-345 nm
photolysis of diazene 5b at 77 and 5 K. The Sb spectrum
(Figure 1 Sb) is characterized by the zfs parameters |D/hc| )
0.0140 cm-1 (2D′ ) 302 G), |E/hc| ∼ 0. Although the
nominally “forbidden” ∆ms ) 2 transition was difficult to see
at 77 K, it was readily visible at 5 K. Even at 77 K, a weak
∆ms ) 2 resonance could be observed in matrices containing
ethyl iodide, a heavy-atom solvent, presumably because of spin-
orbit-mediated partial relaxation of the selection rule.
Although a triplet spectrum now was undoubtedly being
generated, the result again was frustrating, because we were
soon convinced that spectrum Sb was not associated with triplet
N-tosyl-3,4-dimethylenepyrrole 31b. That species should have
shown (see below) a D value near 0.025 cm-1, which would
correspond to a splitting between the outermost lines (2D′)
almost twice that observed in spectrum Sb. Although not
structurally diagnostic in direct terms, the zfs of Sb suggests
that the average separation of the unpaired electrons in its carrier
must be greater than that in 31b. The diazenyl biradical7
structure 311b, resulting from cleavage of only one of the C-N
bonds of the diazene 5b, would be consistent with this feature
of the Sb ESR spectrum and with the method of synthesis by
photoactivation of the diazene chromophore of 5b.
3
that a competitive thermal reaction of 12b, the carrier of the
Mb spectrum, could convert it in part to some other product
that is ESR-silent. Nevertheless, conversion to 311b, the carrier
of the Sb spectrum, must account for the major part, if not all,
3
of the thermal reaction that consumes 12b.
The kinetics of the Mb f Sb transformation are complex
and cannot be fitted well to either a first- or a second-order
rate law, probably because of a distribution of the immobilized
triplet molecules over nonequivalent sites in the matrix.11c For
this reason, a determination of true rate constants and an
(7) (a) Review: Engel, P. S. Chem. ReV. 1980, 80, 99. (b) Review: Du¨rr,
H.; Ruge, B. Top. Curr. Chem. (Triplet States III) 1976, 66, 53. (c)
Review: Suehiro, T. ReV. Chem. Intermed. 1988, 10, 101. (d) Engel, P. S.;
Gerth, D. B. J. Am. Chem. Soc. 1983, 105, 6849. (e) Adam, W.; Gillaspey,
W. D.; Peters, E.-M.; Peters, K; Rosenthal, R. J.; von Schnering, H. G. J.
Org. Chem. 1985, 50, 580. (f) Adam, W.; Do¨rr, M. J. Am. Chem. Soc.
1987, 109, 1240. (g) Reedich, D. E.; Sheridan, R. S. J. Am. Chem. Soc.
1988, 110, 3697.
(8) Adams, J. S.; Weisman, R. B.; Engel, P. S. J. Am. Chem. Soc. 1990,
112, 9115 and references cited therein.
(9) Bigot, B.; Sevin, A.; Devaquet, A. J. Am. Chem. Soc. 1978, 100,
2639.
(10) Baird, N. C.; Kathpal, H. B. Can. J. Chem. 1977, 55, 863.
(11) (a) Wertz, J. E.; Bolton, J. R. Electron Spin Resonance; McGraw-
Hill: New York, 1972; Chapter 10 and references cited therein. (b) Ibid.
Chapter 6. (c) For a review of such effects, cf.: Platz, M. S. In Kinetics
and Spectroscopy of Carbenes and Diradicals; Platz, M. S., Ed.; Plenum,
New York, 1990.
Spectrum Sb also could be discerned in 366-370 nm
photolyses of 5b at 5 K and, faintly, even at 90 K, by digital

Triplet Species from Dihydropyrrolo[3,4-d]pyridazines
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1419
responsible for the well-known12 photo deprotection of N-
arenesulfonamides (see below).
Under some conditions, the ∆ms ) 1 region of the Lb
spectrum shows six lines, of which two (3233 and 3392 G) flank
the impurity peak in the interior of the spectrum and are not
associated with the new triplet carrier of Lb, since their
intensities relative to those of the lines of Lb are variable. The
positions of the 3233 and 3392 G lines of Lb correspond exactly
to those of the strong inside lines (3227 and 3396 at 9.294 GHz)
of the previously described Sb triplet spectrum associated with
the distal diazenyl biradical 11b (see Figure 1). The weak outer
lines of the narrow Sb triplet spectrum, expected near 3164 and
3466 G (Figure 1), are not readily seen in the spectrum (Figure
2) of the 265 nm product because they are overlapped by
absorptions of Lb. Apparently, the 265 nm irradiation in MTHF
leads to two different triplets, the carriers of Lb and Sb.
In the photochemistry of diazene 5b that leads to the triplet
spectrum Lb, the diazene unit seems to be a necessary structural
feature of the precursor. For example, irradiation at 265 nm
and 77 K of the model compound N-tosylpyrrole generates
strong signals in the g ) 2 region which we assign to doublet
or radical pair species produced in the photodesulfonylation of
sulfonamides,12a-d but no trace is seen of a triplet signal of
sufficiently large zfs to account for the signal Lb observed in
the irradiation of 5b. This finding argues against several
possible assignments of the Lb triplet signal to species derived
from hypothetical processes localized to the pyrrole and
arenesulfonyl moieties.
Figure 2. ESR spectrum obtained by 265 nm irradiation of N-tosyl
diazene 5b in EIA glass at 77 K for 1.5 min. The spectrometer frequency
is 9.296 GHz. The upper trace shows the ∆ms ) 1 region; the peaks of
the Lb spectrum associated with 31b are marked with arrows and occur
at 3075, 3184, 3430, and 3558 G. The unlabeled peak in the middle of
the spectrum is associated with impurities formed in the photolysis.
The two peaks flanking it are at the same positions as the main peaks
of the Sb spectrum (see Figure 1). The lower trace is offset and shows
the ∆ms ) 2 region centered at 1662 G.
Arrhenius activation energy have not been feasible. By forcing
the data to first-order form (r ) 0.96), we derive a rate
“constant” for the 12b f 11b reaction at 57 K of 2.77 × 10-5
s-1, which corresponds to a free energy of activation of 4.4
kcal/mol. Anticipating criticism of this crude kinetic method,
we note that although site nonuniformity can lead to rate
“constants” that differ by a factor of 100 or more,11c the effect
on ∆G‡ at these temperatures is quite small. For example, rate
constants of 10-4 and 10-6 s-1 at 57 K correspond to ∆G‡ )
4.2 and 4.7 kcal/mol, respectively.
The assignment of the structure N-tosyl-3,4-dimethylene-
pyrrole 31b to the carrier of the triplet spectrum Lb is consistent
with its zero-field splitting. The D value matches closely the
value |D/hc| ) 0.0256 cm-1 calculated theoretically for N-ben-
zenesulfonyl-3,4-dimethylenepyrrole by a point-dipole ap-
proximation procedure (Table 2). Moreover, it is close to those
experimentally measured and theoretically calculated for other
known TME derivatives. Although this type of calculation is
not calibrated for σ-biradicals and hence is not reliable for
Similarly, the essentially unidirectional nature of the re-
arrangement of 312b to 311b should not cause surprise, since at
low temperatures, small energy differences, such as could be
associated with conformational isomerism, readily could account
for this result. For example, we estimate the ESR detection
3
3
prediction of the zfs of the diazenyl biradicals 11b and 12b,
the D values for these species, which are 1,5- or 1,6-biradicals,
would be expected to be smaller than that for 31b, a 1,4-
biradical, in agreement with observation.
Since the triplet 31b does not have structural 3-fold symmetry,
the near-zero value of |E/hc| observed suggests a near-equality
of two of the nominally anisotropic spin-dipolar interactions.
This presumably is the result of an accidental near-degeneracy
of two of the Cartesian directions in the molecular framework.11a
Similarly, the ESR spectrum of the triplet brosyl derivative 1c
(see below) shows only four lines, corresponding to |E/hc| ∼
0. This degeneracy is broken, however, in the case of the
N-tosyl-2-phenyl-3,4-dimethylenepyrrole biradical 31d (see be-
low), whose ESR spectrum shows six lines, corresponding to
|E/hc| > 0.
3
limit for 12b remaining after the rearrangement has reached
312b-311b equilibrium at 75 K to be about 5%, corresponding
to an equilibrium constant of about 20 in favor of 11b. ∆G°
for the conformational isomerization therefore could be as small
as 0.45 kcal/mol.
A Kinetically Stable Triplet State of N-Tosyl-3,4-dimeth-
ylenepyrrole 1b. Brief irradiation of the N-tosyl diazene 5b
in MTHF or other solvent glasses at wavelengths in the range
250-295 nm does not produce a detectable blue color or a UV-
vis maximum at 593 nm characteristic of the singlet 1b. Instead,
these preparations show a four-line ESR spectrum, Lb, char-
acteristic of a third triplet species. The zfs parameter, |D/hc|
) 0.0226 cm-1 (2D′ ) 483 G), is the largest of the three triplet
spectra. The spectrum can be fitted to a Hamiltonian with small
or zero values of the parameter |E/hc| (<0.0005 cm-1). At 265
nm in EIA glass, for example, 90 s of irradiation at 77 K
produces the spectrum shown in Figure 2, observed at a
microwave frequency of 9.296 GHz. A weak half-field transi-
tion (∆ms ) 2) associated with the Lb spectrum is seen centered
around 1662 G. This confirms the assignment of a triplet state
to the carrier of the Lb spectrum.11a
1
3
Like the blue singlet form of biradical 1b,1 the triplet 1b
was stable in the dark for weeks at 77 K but was photolabile.
Irradiation at 400 nm of a sample of 31b caused disappearance
of the Lb ESR signal. The impurity peaks in the center of the
spectrum remained after photobleaching of the Lb spectrum,
thereby demonstrating that they arise from a different carrier.
New ESR signals were not detected from the photobleaching
reaction, and it is tempting to conclude that the photoproducts
3
of the Lb carrier 1b are ESR-silent. However, since the Lb
(12) (a) Abad, A.; Mellier, D.; Pe`te, J. P.; Portella, C. Tetrahedron Lett.
1971, 4555. (b) Mellier, D.; Pe`te, J. P.; Portella, C. Tetrahedron Lett. 1971,
4559. (c) Umezawa, B.; Hoshino, C.; Sawaki, S. Chem. Pharm. Bull. 1969,
17, 1115, 1120; 1970, 18, 182. (d) Art, J. F.; Kestemont, J. P.; Soumillion,
J. Ph. Tetrahedron Lett. 1991, 32, 1425.
The peak near 3317 G we assign to a doublet or radical pair
impurity, which probably arises from photodissociation of the
S-N bond of the N-tosyl group. This type of reaction is

1420 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Bush et al.
Table 2. Values of the Zero-Field Splitting Parameter |D/hc|
Scheme 3
(cm-1) for TME Derivatives
compd
|D/hc|, obsd
|D/hc|,calcdh
0.025a
0.0236;d,f 0.025e
0.0259b
g
0.0204,b 0.024c
0.0226d
0.025e
0.0256d,f,h,i
0.0231d
0.0187d
g
g
(5) The D value of the Lb spectrum is in good accord with
that predicted by a point-dipole approximation that has given
agreement with a number of known π-conjugated triplet
molecules (Table 2).
a Reference 3a. b Reference 3c. c Reference 3b. d This work. e Rule,
M.; Matlin, A. R.; Seeger, D. A.; Hilinski, E. F.; Dougherty, D. A.;
Berson, J. A. Tetrahedron, 1982, 38, 787. A value of 0.0255 cm-1
has recently been calculated for this molecule by methods requiring
no empirical scaling factor: Prasad, B. L. V.; Radhakrishnan, T. P. J.
Mol. Struct. (Theochem) 1996, 361, 175. f The empirical divisor (see
ref 5e) used in this work was 2.11. g Not calculated. h Details of the
calculations are given in the Supplementary Material to ref 1a and in
ref 6b. i Calculated for the N-benzenesulfonyl-3,4-dimethylenepyrrole
biradical.
In the next section, we provide support for the assignment
of the Lb spectrum to 31b. For the present, no such additional
evidence is available for the diazenyl biradicals 11b and 12b,
the putative carriers of Sb and Mb. It will be obvious, for
example, that the observation of nitrogen hyperfine splitting in
the latter spectra would strengthen the assignments. However,
hyperfine splitting often is not seen under the broad lines of
such powder spectra, so that its absence is more a source of
regret than of surprise. As is described in the Supporting
Information, other confirmatory findings would be a demonstra-
tion that either or both of the diazenyl biradicals could be
thermally or photochemically converted to the N-tosyl-3,4-
dimethylenepyrrole biradical 1b. We note here that attempts
to demonstrate photoconversion of diazenyl biradicals 311b or
312b to 31b by irradiation at 265 nm, the wavelength needed to
generate 31b, are frustrated by the fact that, as we already know,
spectrum represents only about 1.6% conversion of diazene 5b
to biradical 31b (see below), it is possible that new triplet species
3
might be present in the photolysate of 1b but at too low a
concentration to be detected. The photobleaching of 31b does
not produce any visible blue color, such as would be associated
with the singlet state of 1b.
Structural Assignments to the Carriers of the ESR Spectra
Lb, Mb, and Sb. A. Spectroscopic Evidence. For review,
we list the principal spectroscopic bases for the assignments
3
the spectrum Lb associated with 1b appears upon 265 nm
3
3
3
given as, respectively, 1b, 12b, and 11b:
irradiation of diazene 5b, without prior generation of spectra
Sb or Mb. Therefore, the appearance of Lb upon 265 nm
irradiation of a preparation of Sb/Mb, which normally contains
unreacted 5b, would not be decisive. Further discussion of the
carriers of the Sb and Mb spectra, putatively 11b and 12b,
including the nature of the reaction that occurs above 70 K and
eventually causes the Sb ESR signal to vanish, is given in the
Supporting Information.
B. Chemical Evidence for the Structure of the Lb
Carrier. At the present stage of the study, the most significant
of the three triplets is the carrier of the Lb spectrum, which we
have assigned the N-tosyl-3,4-dimethylenepyrrole structure 31b.
We now describe experimental observations which exclude the
most plausible alternative and which demonstrate that it is
possible to involve the carrier of the Lb triplet in a cycloaddition
reaction that gives a product containing the intact N-tosyl-3,4-
dimethylenepyrrole structure.
(1) The diazene unit of the precursor 5b is a necessary
structural feature for the production of any of the three triplet
spectra. No such spectra are observed upon irradiation of model
compounds such as N-tosylpyrrole or the cycloadducts of
N-tosyl-3,4-dimethylenepyrrole with alkenes. This suggests that
the chemical change leading to any of the triplets is initiated
by a photoreaction of the diazene unit.
(2) The carriers of the Sb and Mb spectra are formed by
irradiation in the diazene chromophore with light of wavelength
340-370 nm. This again is consistent with a process involving
cleavage of a C-to-NdN bond. In the photodeazetation of
acyclic and cyclic 1,2-diazenes, a variety of mechanistic studies7
have provided persuasive indirect evidence implicating diazenyl
radicals and biradicals, respectively. A significant confirmatory
observation comes from time-resolved coherent anti-Stokes
Raman spectroscopy (CARS)8 of the evolved N2 in such
processes.
(3) The D values of the Sb and Mb spectra are smaller than
that of the Lb spectrum, in accord with a greater separation of
the unpaired electrons in the biradicals responsible for the former
two spectra.
(4) The occurrence and rate of the transformation of the Mb
carrier to the Sb carrier are consistent with a conformationally
isomeric relationship of the two.
Exclusion of a Rearranged Structure. Probably the most
vexatious of the alternatives to 31b would be a rearranged
structure, such as could be imagined to arise by the pathway
shown in Scheme 3. Hypothetically (but not necessarily), this
might occur through the pair of radicals 13-14. Recombination
in 13-14 could lead to an isopyrrole 15, which after tautomer-
ization to the rearranged pyrrole 16, could undergo deazetation

Triplet Species from Dihydropyrrolo[3,4-d]pyridazines
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1421
to the biradical 17. The overall structural result would be to
interchange the tosyl group at N with the hydrogen at C2 by
way of a nitrogen analog of a photo-Fries rearrangement.
Several examples of reactions analogous to 5b f 16b, especially
in N-tosylcarbazoles, have been observed.13a,b
Scheme 4
However, several aspects of the formulation of the carrier of
our ESR spectrum Lb as the rearranged structure 17 are
improbable. The reactions leading to the observed triplet species
are observed to occur at 5 K, so that either there would have to
be several independent photochemical steps in sequence, or any
thermal steps, for example, the 1,5-sigmatropic hydrogen shift
15 f 16, must have activation energies of <0.3 kcal/mol or
must occur by thermal tunneling processes. Also, there is no
obvious reason why biradical 17 should have an accessible triplet
state, since the electron-withdrawing Ts substituent is no longer
on the pyrrole nitrogen.1 Perhaps the most implausible require-
ment is that the rearranged diazene 16 must undergo deazetation
to biradical 17 but the original diazene 5b must not undergo
deazetation to biradical 1b. Despite these arguments, the
possibility of such a rearrangement was of sufficient concern
to motivate additional scrutiny.
By Analogy. In the first experiment of this series, we studied
the photolysis of the bis-carbamate 18 as a model for the diazene
5b, with the rationale that the carbamate units should have no
independent photochemistry under the conditions we would use.
This would provide a clear field for the observation of the
suspected photo-Fries rearrangement without the deazetation that
inevitably accompanies photolysis of the diazene. We found
that in any of several different solvents (benzene, CH3CN, THF,
MTHF, or MeOH), irradiation of the bis-carbamate 18 at room
temperature with light from a 1000 W Hg-Xe arc (unfiltered
or NiSO4-filtered) did indeed give the rearranged bis-carbamate
19 in about 20% yield after chromatographic isolation. Some
starting 18 was recovered and, in some of the solvents, a trace
of the detosylated bis-carbamate was formed. Significantly,
however, when irradiation of bis-carbamate 18 was carried out
either with 366 nm light (other conditions unchanged) or with
NiSO4-filtered light in MTHF matrix at 77 K, none of the
rearranged bis-carbamate was formed, and only the starting
material 18 was recovered. Thus, the photorearrangement of
the bis-carbamate 18 is effectively shut down in low-temperature
matrices. By analogy, the hypothetical rearrangement of the
diazene 5b (Scheme 3) under these conditions is unlikely.14a
the now known bis-carbamate 19. However, we found that
photolysis (at various wavelengths) of diazene 5b at 77 K in
MTHF matrix in the presence of 2.5 equiv of dimethyl
azodicarboxylate (DMAD) followed by warming to room
temperature gave as a trapping product only the unrearranged
bis-carbamate 18 and none of the rearranged one 19 (Scheme
4).
To demonstrate the connection betweeen the carrier of the
Lb ESR spectrum and the N-tosyl-3,4-dimethylenepyrrole
structure, we carried out a series of experiments (Scheme 4) in
which the triplet carrier 31b of the ESR spectrum was generated
as a reactive intermediate in the absence of the blue singlet 11b.
As we have previously shown, irradiation with monochromator-
filtered light of wavelength 265 nm of an MTHF matrix of
diazene 5b at 77 K (see Table 1) gave the Lb ESR spectrum of
31b, but no blue color or 590 nm UV-vis absorption of the
1
singlet 1b could be detected. Because the 265 nm emission
of the Hg-Xe photolamp is strongly reversed, and because the
transmission of the monochromator is only about 10% at this
wavelength,18 the irradiance is very low and the amount of triplet
actually generated under these conditions is only 1.6% by ESR
spin count. We have not been able to detect any trapping
product that can be ascribed to a photogenerated biradical under
(14) (a) To circumvent the assumptions associated with the use of a model
reaction, it would be desirable to know something of the chemistry of the
hypothetical rearranged diazene 16 itself. On a small scale, we have tried
to prepare this substance from the rearranged bis-carbamate 19, but our
usual procedures of hydrolysis and oxidation so far seem to give only the
corresponding hydrazone (NMR identification). We suspect that the presence
of the conjugated tosyl group at C2 renders the hydrogens adjacent to the
diazene linkage of 16 so acidic that rapid tautomerization to the more stable
hydrazone ensues. (b) A reviewer has asked whether the radical-pair 13-
14 (Scheme 3) might be the carrier of the Lb ESR spectrum. Substantial
evidence against this possibility is available. First, the D value of the Lb
spectrum, 0.0226 cm-1, is large for a radical pair and would require that
the radical partners remain close together. Since the same D value is
observed in several different matrices, one would have to postulate that,
fortuitously, the separation remains the same in all these media. Second, as
already has been stated, no such triplet spectra are seen in the photolyses
of model compounds such as N-tosylpyrrole or the cycloadducts of biradical
1b and alkenes. Third, to form, in the dark, the observed trapping product
18 from the Lb triplet preparations allegedly containing the 13-14 radical
pair would require some such mechanism as thermal deazetation of the
radical 14 at 193 K, followed by cycloaddition of the resulting triradical,
followed by recombination of the resulting adduct monoradical with radical
13. We consider these joint requirements sufficient to render the radical-
pair assignment highly improbable.
Directly. The Carrier of the Lb ESR Spectrum Contains
the Intact 3,4-Dimethylenepyrrole Unit. In addition to this
analogical argument, we have been able to provide direct
evidence that photolysis of the diazene 5b (Scheme 3) does not
produce appreciable quantities of rearranged diazene 16 or its
cognate biradical 17. Capture of 17 by an appropriate trapping
agent, dimethyl azodicarboxylate (DMAD), would have led to
(13) (a) Horspool, W. M. In The Chemistry of Sulphonic Acids, Esters
and their DeriVatiVes; Patai, S., Rappoport, Z., Eds.; John Wiley: New
York, 1991; Chapter 13. (b) Pillai, V. N. R. Org. Photochem. 1987, 9, 225.

1422 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Bush et al.
these monochromatically filtered conditions. However, with
NiSO4-filtered irradiation (230-325 nm), the light intensity was
much greater, and the ESR signal intensities of the resulting
Lb and Sb spectra were at least 5 times greater than those seen
in a comparable time of irradiation with monochromator-filtered
265 nm light. There is no doubt that much higher concentrations
of triplet species were being produced in the NiSO4-filtered
irradiation, but the characteristic Visible blue color and 590
Scheme 5. Irradiations in Glassy Matrices
1
nm UV-Vis band of the singlet 1b still did not appear.
The next task was to show that the bis-carbamate trapping
product 18 observed in the experiments of Scheme 4 arose from
photochemically generated intermediates and not just from
remaining unphotolyzed 5b in the warm-up step. Accordingly,
in the experiments of Scheme 4, the samples were warmed only
to 193 K, a temperature below the threshhold for thermal
decomposition of the diazene 5b. The remaining unphotolyzed
5b was then scavenged by adding a swamping concentration
(50 times the initial DMAD concentration) of a second trapping
agent, N-methylmaleimide (NMM), and allowing the temper-
ature to rise to 300 K. NMM is a faster diylophile than DMAD
and therefore should be essentially (>99%) the exclusive
trapping agent for any diyl liberated thermally at this point. If
the photolyses actually were generating diyls in interceptable
amounts, the ratio of photoproduct to thermal product (DMAD
adduct 18 to NMM adduct 20) should increase with increasing
photolysis time. Experimentally, this was what we observed.
The ratios were 22:78, 30:70, and 54:46, respectively, after 15,
30, and 60 m of photolysis. Again, the only DMAD product
observed was the unrearranged bis-carbamate 18. The HPLC
analytical conditions were such that a peak at the known
emergence time of the rearranged bis-carbamate 19 could have
been detected at the level of <0.1% of the intensity of the band
due to 18.
do not develop the blue color or UV-vis band (λmax 590 nm)
1
characteristic of the singlet form 1b after storage for days.
Similarly, the blue preparations1 of the singlet can be stored
for extended periods at 77 K without developing any of the
ESR spectrum of the triplet form (Scheme 5). If our structural
assignments are correct, these experiments haVe generated
independently the singlet and triplet forms of the same biradical,
and interconVersion of the spin isomers is slow on the time-
scale of days or weeks. We now have devoted some effort to
understanding the origins and implications of this long-liVed
spin isomerism.
One immediately obvious consequence of the slow intercon-
version is that serious doubt now is cast on attempts to assign
the spin of the molecular ground state of biradical 1b. For
example, the widely used technique16,17 of the Curie law plot
attributes a triplet ground state (or an exact degeneracy of triplet
and singlet) to a species whose ESR signal intensity is linear
with reciprocal temperature. Equilibration was assumed by
implication in the original application of ESR Curie plots,16 and
later, it was restated explicitly as a requirement.17d The present
observations call attention to the fact that an ESR-active triplet
not in rapid interconversion with another spin isomer can always
be expected to give a linear Curie plot, whether or not it is the
ground state. The time scale for “rapid interconversion” here
is the investigator’s pace of changing the temperatures and
making the intensity measurements. Under conditions of slow
interconversion, the ESR signal of such a triplet, like that of
any kinetically stable paramagnetic species, then simply will
respond linearly to reciprocal temperature.
3
If the ESR spectrum Lb, which we have assigned to 1b,
were instead due largely (say a 10:1 ratio 17:1b) to the
rearranged biradical 17 (Scheme 3), then in order for un-
rearranged adduct 18 to be formed in the ratio of >1000:1 over
rearranged adduct 19, the reactivity of 1b toward DMAD would
have to be >10 000 times that of 17. This seems unlikely. The
more reasonable conclusion is that, in competition with de-
azetation, not even a trace of photorearrangement of 5b to
diazene 16 occurs at 77 K, under the photolysis conditions that
produce the Lb ESR spectrum. When taken with the already
listed, inherently implausible subsidiary hypotheses needed to
sustain the proposal that the Lb ESR spectrum is associated
with rearranged biradical 17, the trapping results provide a
persuasive basis for the straightforward structural assignment
The Supporting Information describes details of Curie law
experiments on the N-tosyl-3,4-dimethylenepyrrole triplet 1b.
The slight downward curvature of the plot can be fitted formally
to an energy separation ET - ES ≈ 19 cal/mol, on the usual
(but now dubious) assumption of spin equilibrium between the
triplet and some singlet. If this singlet is the blue species of
λmax ) 593 nm, in a blue sample prepared by irradiation at 370
nm under specified conditions (see Supporting Information), an
energy gap of 19 cal/mol would imply the presence of an amount
3
of biradical 1b as the carrier.14b
These experiments strongly suggest that DMAD can trap
chemically either the triplet species responsible for the Lb ESR
spectrum, or a species readily derived from that triplet, to give
a product bis-carbamate 18 which contains an intact N-tosyl-
3,4-dimethylenepyrrole structure. With reference to other
attempts to trap 31b, which have uniformly led to products from
11b instead,1c it seems likely that the actual reactant in the
present DMAD trapping experiments is the singlet 11b formed
by intersystem crossing from 31b when the matrix is thawed to
effect the cycloadditon, as already described.1a-c In sum, the
observations link the carrier of the Lb ESR spectrum with the
N-tosyl-3,4-dimethylenepyrrole structural unit.
(15) For discussions of heavy-atom effects on intersystem crossing,
see: (a) McGlynn, S. P.; Azumi, T.; Kinoshita, M. Molecular Spectroscopy
of the Triplet State; Prentice-Hall: Englewood Cliffs, NJ, 1969; Chapter
5. (b) Michl, J.; Bonacic-Koutecky, V. Electronic Aspects of Organic
Photochemistry; Wiley: New York, 1990; pp 24, 42. (c) Fisher, J. J.; Michl,
J. J. Am. Chem. Soc. 1987, 109, 583. (d) Koziar, J. C.; Cowan, D. O. Acc.
Chem Res. 1978, 11, 334. (e) Turro, N. J. Modern Molecular Photochem-
istry; Benjamin-Cummings: New York, 1978; Chapter 6.
(16) Bijl, D.; Kainer, H.; Rose-Innes, A. C. J. Chem. Phys. 1959, 30,
765.
(17) Reviews: (a) McGlynn, S. P.; Azumi, T.; Kinoshita, M. Molecular
Spectroscopy of the Triplet State; Prentice-Hall: Englewood Cliffs, NJ,
1969. (b) Wasserman, E.; Hutton, R. S. Acc. Chem. Res. 1977, 10, 27. (c)
Reference 11a. (d) Berson, J. A. In The Chemistry of the Quinonoid
Compounds; Patai, S., Rappoport, Z., Eds.; Wiley: New York, 1988; Vol.
II, p 482 ff.
Long-Lived Spin Isomerism of the Singlet and Triplet
States of N-Tosyl-3,4-dimethylenepyrrole 1b. The Lb ESR
3
signal of 1b persists in the dark at 77 K for at least 1 month.
Even in EIA-heavy-atom15 solvent mixtures containing 1-bro-
mopropane, iodoethane, or xenon at 77 K, the triplet preparations
(18) Oriel Equipment Catalog, Vol. II, Oriel Corporation, Stratford, CT,
1994; p 1-43.

Triplet Species from Dihydropyrrolo[3,4-d]pyridazines
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1423
of the triplet easily sufficient to detect by the characteristic ESR
spectrum. Yet, such samples show no trace of an ESR triplet
signal. We conclude, therefore, that either the small curvature
in the Curie plot is artifactual (for example, a saturation effect)
and the triplet has not achieved equilibrium with any other
species, or if it has and the curvature is real, the species in
equilibrium with the triplet is not the blue singlet already
described.
rapidly disappeared when the matrix was thawed. UV-vis
examination of the green matrix showed only weak broad
absorption in the region 600-700 nm. By analogy to the cases
of the 370 nm photolyses of the N-tosyl- and N-brosyl diazenes
1
5b,c, it seems likely that the singlet biradical 1d is present.
The weak optical absorption perhaps may indicate that only a
low photostationary concentration of this species accumulates
under these conditions, plausibly because of a secondary
photoreaction analogous to that which photobleaches the singlet
11b. Further work would be needed to test this possibility. A
somewhat stronger and better defined UV-vis spectrum was
obtained when a glass of the dibromide 21 in MTHF containing
Incorporation of an Internal Heavy Atom. A Kinetically
Stable Triplet Form of N-p-Bromobenzenesulfonyl-3,4-
dimethylenepyrrole 31c. Since the rate of intersystem crossing
was not perceptibly increased by external heavy atoms in the
matrix (see above), we attempted to achieve this end by
incorporation of a heavy atom in the biradical itself. The N-p-
bromobenzenesulfonyl (brosyl)-3,4-dimethylenepyrrole biradical
1c was prepared from diazene 5c by methods similar to those
used for the preparation of the tosyl derivative 1b from its
precursor 5b.
tetrakis(dimethylamino)ethylene at 77 K was irradiated19 with
Pyrex-filtered light: λmax (nm) 400 (broad, strong), 640
(medium), and 795 (weak). It is possible that not all of these
absorptions are caused by one species. We have not yet studied
the chemistry of the singlet, but the UV-vis spectrum persists
upon storage at 77 K.
Irradiation of 5c at 340 and 370 nm. The photochemistry
of brosyl diazene 5c at 77 K was essentially identical to that of
the tosyl derivative 5b. Irradiation of 5c in MTHF at 77 K
with 370 nm light gave a pale blue preparation which showed
only a weak narrow ESR triplet spectrum, Sc. A much stronger
Sc spectrum, with resonances at the same positions as before,
was generated by irradiation at 340 nm. The ∆ms ) 2 transition
was weak but visible at 77 K. The zfs parameters of the Sc
ESR examination of the above sample obtained by irradiation
at 77 K of the N-tosyl-2-phenyl diazene 5d at 370 nm or of a
sample irradiated at 345 nm showed the presence of a 4-line
triplet spectrum Sd or Md with zfs parameters |D/hc| ) 0.0129
cm-1 and |E/hc| ∼ 0 cm-1, superimposed upon a weak narrow
impurity pattern. The separation of the outermost lines of the
∆ms ) 1 region of the Sd or Md triplet spectrum, 2D′ ) 275
G, was slightly smaller than the corresponding values of 302 G
observed for the narrow triplets Sb and Sc obtained by 345 nm
irradiation of the N-tosyl- and N-brosyl diazenes 5b and 5c,
but substantially smaller than the 404 G splitting in the Mb
spectrum, |D/hc| ) 0.0140 cm-1, |E/hc| e 0.0007 cm-1
,
matched those of the spectrum Sb of diazenyl biradical 11b
obtained by 340 nm irradiation of the N-tosyl diazene 5b.
Accordingly, we tentatively assign the diazenyl biradical
structure 11c to the carrier of the Sc ESR spectrum obtained
by 340 nm irradiation of 5c. This spectrum was still strong
after having been stored in the dark for 14 days at 77 K.
Irradiation of 5c at 265 nm. Irradiation at 265 nm of the
N-brosyl diazene 5c in MTHF matrix at 77 K gave rise to a
strong four-line ESR signal of a new triplet species Lc
superimposed on a weak central signal about 100 G wide
associated with an impurity. The new four-line triplet signal
was characterized by zfs parameters almost identical to those
of Lb, the spectrum of the N-tosyl triplet 31b: |D/hc| ) 0.0231
cm-1, |E/hc| e 0.0005 cm-1. A weak half-field transition (∆ms
) 2) also was observed. We assign the Lc spectrum to the
3
spectrum obtained from 5b. The phenyl substituent in 1d
should decrease the zfs (see below), but it is not easy to predict
just how much the decrease should be. Therefore, it is not
obvious from the splitting alone whether the 2D′ ) 275 G signal
3
from 5d should be called Sd and assigned to a carrier 11d,
structurally similar to the anti diazenyl biradical rotamer 311b,
or Md and assigned to 312d, corresponding to syn rotamer 312b.
At present, we are inclined to favor the first of these alternatives,
Sd, because the spectrum is stable at 77 K, as is the case with
Sb but not with Mb. The spectrum we now tentatively call Sd
was accompanied by a weak but observable ∆ms ) 2 signal at
half-field.
3
N-brosyl-3,4-dimethylenepyrrole biradical 1c. The spectrum
was observed at temperatures as low as 14 K and persisted with
undiminished intensity for 15 days at 77 K. As has been
reported in the accompanying article,1c the blue singlet form
11c does not give rise to any ESR resonances, even after long-
term storage at 77 K.
Irradiation of 5d at 265 nm. Irradiation of a 77 K glass of
diazene 5d in MTHF at 265 nm for 40 s caused the growth of
a strong six-line triplet ESR pattern Ld (Figure 3) in the ∆ms
) 1 region with transitions at 3101, 3192, 3207, 3404, 3423,
and 3502 G (microwave frequency ) 9.261 GHz) superimposed
upon weak impurity resonances in the center of the spectrum.
The zfs parameters were |D/hc| ) 0.0187 cm-1 and |E/hc| )
0.0005 cm-1. A half-field transition was observed at 1647 G.
We propose that the carrier of this spectrum is the triplet state
of N-tosyl-2-phenyl-3,4-dimethylenepyrrole biradical 31d. Note
the decrease in the value of the zfs parameter upon phenyl
substitution: 31b, |D/hc| ) 0.0226 cm-1; 31d, |D/hc| ) 0.0187
cm-1, corresponding to a narrowing of the separation of the
It is possible that the heavy atom effect of bromine in the
brosyl biradical 1c does in fact accelerate intersystem crossing,
but that even the increased rate remains too slow to detect. We
must point out that the location of the bromine in 31c probably
is not optimally conducive to the internal heavy atom effect,
which would have been expected15c to be stronger with the
bromine directly at a site of high spin density, namely C2 of
the pyrrole ring or the exocyclic methylene position. For the
present, we can only state that our attempts to induce a
discernible enhancement in the rate of intersystem crossing in
this series by external or internal heavy atoms have been
ineffectual.
A Triplet State of N-Tosyl-2-phenyl-3,4-dimethylene-
pyrrole. Irradiation of 5d at 345 and 370 nm. We also have
studied briefly the species obtained by photolysis of the N-tosyl-
2-phenyl diazene 5d. Irradiation at 370 nm in MTHF gave a
visibly green preparation whose color persisted at 77 K but
(19) (a) This is an application of a method for generating biradicals
developed by Haider, Platz, and co-workers19b and used in our laboratory
to prepare 3,4-dimethylenethiophene.19c (b) Haider, K.; Platz, M. S.; Despres,
A.; Lejeune, V. Migirdicyan, E.; Bally, T.; Hasselbach, E. J. Am. Chem.
Soc. 1988, 110, 2318. (c) Greenberg, M. M.; Blackstock, S. C.; Stone, K.
J.; Berson, J. A. Ibid. 1989, 111, 3671.

1424 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Bush et al.
interaction.21,22 In the theory of radiationless transitions,23 the
rate constant for isc is proportional (see eq 1) to a term Cvib
,
which incorporates constants and vibronic terms, and to the
square of the elements of the coupling matrix which expresses
the mixing of the singlet and triplet wave functions by the spin-
orbit coupling operator Hso.
2
kisc ) Cvib T|Hso|S
(1)
The non-zero matrix elements have the form of (2):
matrix element ) 4NZhabi (λ - 1 - λ )(úAB + úπAB) (2)
2
σ
x
As Salem and Rowland have emphasized,21a the components
of the matrix element have important implications for under-
standing the structural influences on the rate constant. Thus,
the term 4NZhabi expresses the requirement for proper orbital
orientation to accommodate the “torque” of the spin flip.
Ideally, the transition should occur between orbitals with axes
orthogonal to each other and to the axis about which orbital
angular momentum is created. The term λ - (1 - λ2)1/2 denotes
the difference between the coefficients of the two two-electron
wave functions21b which, in this approximation, are linearly
combined to constitute the singlet state; the term úσAB + úAπB is
related to distance/overlap properties of the orbitals involved
in the spin-flip.
Figure 3. ESR spectrum obtained by 265 nm irradiation of diazene
5d in MTHF at 77 K. The spectrometer microwave frequency is 9.28
GHz. The fields of the resonances are (from left to right) 3101, 3192,
3207, 3404, 3423, and 3502 G; they are superimposed upon weak
impurity resonances in the center of the spectrum. A weak ∆ms ) 2
transition (not shown here) occurs at approximately half-field.
outermost ∆ms ) 1 lines by about 90 G. This effect is consistent
3
with the structural assignment if the 2-phenyl ring of 1d can
acquire some spin density, thereby decreasing the average
propinquity of the unpaired electrons. If the mechanism of spin
delocalization involves only the π-electrons, the optimum
torsional geometry would require planarity of the phenyl and
pyrrole rings. However, it is not obvious that some delocal-
ization by spin polarization cannot also occur. In any case,
whatever the geometry, the empirical observations are remi-
niscent of other diminutions of the D value by phenyl substitu-
tion.20 Note also that the strong bands of the spectrum each
are comprised of two overlapping but discernible transitions.
Thus, the value of the zfs parameter |E/hc| in the spectrum Ld
of 31d is finite, which indicates that the accidental degeneracy
of two of the triplet sublevels seen in both 31b and 31c is lifted
The term λ - (1 - λ2)1/2 defines the ionic character of the
singlet state. Physically, the ionic character must be high to
facilitate isc, because only if the singlet wave function has
contributions from “ionic” configurations, with doubly occupied
orbitals, can the spin flip create orbital angular momentum. The
extreme cases are (i) a fully ionic state in which λ ) 1 and the
ionic character is unity and (ii) a fully covalent state in which
λ ) 1/x2 ) (1 - λ2)1/2 and the ionic character is zero. In the
latter case, it is immediately clear from eqs 1 and 2 that the
matrix element and the rate constant kisc will vanish.
3
by the unsymmetrical phenyl substitution in 1d.
Like the triplet biradicals 1b and 1c, the triplet 1d was
persistent for many hours in the dark at 77 K.
Borden and Davidson24 make the point that because of the
separability (near-zero overlap) of the NBMOs of disjoint
biradicals, one expects the singlet wave function of these species
to have low ionic character. For example, the ionic character
of the singlet state wave function of TME 1 is calculated to be
about 5%. The rate kisc for this molecule is at present
experimentally unknown, but it would be expected on the above
basis to be retarded relative to some ideal rate by a factor of
about (1/20)2 ) 1/400. When we found the extraordinarily slow
rates of isc in biradicals 1b-d, which are after all TME
derivatives, we wondered whether the ionic character of these
singlets might be low enough to depress the rate of isc to near
zero, and hence whether the long-lived spin isomerism observed
3
3
3
Is the Long-lived Spin Isomerism a Consequence of
Disjoint Character of the Biradicals 1b-d? The great
majority of rate constant measurements for singlet-triplet
intersystem crossings (isc) pertain to transitions from electroni-
cally excited states either to other electronically excited states
or to ground states that are much lower in energy. The known
values range over about 13 orders of magnitude.15 Carbonyl n,
π* S1-T1 crossing typically is characterized by fast rates near
1010-1011 s-1, whereas aromatic hydrocarbon S1-T1 rates are
near 108 s-1. Excited-to-ground state transitions can be much
slower: for benzene-d6, the T1-S0 rate is near 10-2 s-1 at 77
K. Little is known about the rates of intersystem crossing in
systems where the two multiplets are close competitors for the
designation of “ground state.” One example of the latter
category is the case of the spin isomers studied here, the
N-arenesulfonyl-3,4-dimethylenepyrrole biradicals 1b-d, which
persist for weeks. The rate constants must be <10-7 s-1, and
the rates are slower than that of the slowest known conventional
“excited state” case by at least 5 orders of magnitude.
(21) (a) Salem, L.; Rowland, C. Angew. Chem., Int. Ed. Engl. 1972, 11,
92. (b) See also: Michl, J. J. Am. Chem. Soc. 1996, 118, 3568. This paper
refines the earlier approximation21a with more accurate wave functions.
Qualitatively, the argument we give here persists in the later version of the
theory. We thank J. Michl for a discussion of this point. (c) Note that in
general hyperfine-induced S-T mixing will make a negligibly small
contribution to isc in π-conjugated non-Kekule´ biradicals, because it requires
that the singlet and triplet energies match to within the energy of the
hyperfine interaction (about 10-6 kcal/mol). This matching is easily achieved
during diffusive radical-pair encounters,22 but barring an accidentally exact
S-T degeneracy, will not occur in static systems such as those in the present
work.
(22) Reviewed by: Glarum, S. H. In Chemically Induced Nuclear
Polarization; Lepley, A. R., Closs, G. L., Eds.; Wiley-Interscience: New
York, NY, 1973; Chapter 1.
(23) (a) Robinson, G. W.; Frosch, R. P. J. Chem. Phys. 1962, 37, 1962.
(b) Robinson, G. W.; Frosch, R. P. J. Chem. Phys. 1963, 38, 1187. (c)
Jortner, J.; Rice, S. A.; Hochstrasser, R. M. AdV. Photochem. 1969, 7, 149.
(24) Borden, W. T.; Davidson, E. R. J. Am. Chem. Soc. 1977, 99, 4587.
We have considered two explanations for this additional
retardation. The first emerges from the nature of disjoint
biradicals. We assume that the only effective intramolecular
mechanism mediating isc in these systems is the spin-orbit
(20) Platz, M. S.; McBride, J. M.; Little, R. D.; Harrison, J. J.; Shaw,
A.; Potter, S. E.; Berson, J. A. J. Am. Chem. Soc. 1976, 98, 5725 and
references cited therein.

Triplet Species from Dihydropyrrolo[3,4-d]pyridazines
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1425
in these cases might be a dramatic and uniquely characteristic
manifestation of their disjoint nature.
with the H-S bond of the HSO2 group perpendicular to the
pyrrole ring plane than in the conformation with it in the ring
plane. This would be ample to support the proposed confor-
mational spin specificity. At the present stage of refinement,
however, the calculations on the model compound do not agree
particularly well with the deduction1 from semiempirical theory
that the gap should be very small in the N-arenesulfonyl
biradicals 1b-d. The ab initio calculations predict that the
absolute value of the smaller gap still should favor the singlet
of the HSO2 model compound by 5.3 (7.2) kcal/mol. Taken at
face value, this large a gap would have precluded detection of
an ESR triplet signal. That such a signal is nevertheless
observed could mean that the gap will narrow on further
computational refinement, or that the H-SO2 substituent is an
inadequate model for the ArSO2 groups we used experimentally,
Further examination,25 however, has cooled our enthusiasm
for this explanation. A rate constant of about 108 s-1 would be
a reasonable estimate of an average value for an idealized kisc
unretarded by any deficit of ionicity.15 The observed maximum
rates for 1b-d are slower than that by at least 15 orders of
magnitude. To ascribe the latter retardation entirely to low ionic
character, it would be necessary that the singlet wave function
15
x
for 1b-d be ionic to no more than one part in 10 , or
1/(107.5). Conceivably, perturbations by the nitrogen and its
attached substituent might reduce the ionic character of the TME
singlet from the calculated24 value of 5% to almost precisely
zero in one of the pyrrole derivatives. However, it seems
improbable that this accident could recur in four separate
molecules: 1b-d and the 1,3-di-2-pyrrylbenzene counterpart
described in an accompanying paper.1e Low ionic character of
the singlet wave function may well contribute to the rate
retardation, but it would seem that some other feature must be
responsible for a major portion of the effect. Accordingly, we
have proposed an alternative interpretation.
3
or that the ESR triplet signals we assign to 1b-d are caused
by other presently unknown species. For the moment, we press
ahead on the assumption that one (or both) of the first two of
these possibilities will prove to be correct.
Mechanism of the Photodeazetation of Diazene 5b to
Triplet Species. Important Common Features of the Mech-
anisms of Formation and the Properties of the Biradicals
1b-d. It should be clear that in order to observe the separate
kinetically stable singlet and triplet spin states of N-arene-
sulfonyl-3,4-dimethylenepyrrole biradicals, not only must the
interconversion of the multiplets be slow, but also, there must
be mechanisms for the photolysis of the diazene precursors that
give one or the other multiplet specifically in a wavelength-
dependent manner. Perhaps the most striking aspect of the
Is the Long-lived Spin Isomerism a Consequence of a
Conformation-Dependent State Ordering? The postulate that
underlies all of the present work is that a sufficiently strong
EWG as a substituent on nitrogen will produce a 3,4-dimeth-
ylenepyrrole biradical with a small energy separation between
the lowest singlet and triplet. In this narrow span of energies,
it seems possible that small structural or conformational changes
could alter the energy separation or even interchange the
energetic ordering of the multiplet states. Most plausibly,
conformational isomerism would be associated with internal
rotational orientation about the S-N and S-C bonds of the
arenesulfonyl group. Interconversion among conformations
could be slow, because of the high viscosity of the matrix and
because of one or more finite rotational barriers within the
biradical itself. Although no experimental determinations of
the rotational barriers in sulfonamides seem to be available,
Breneman29 has used ab initio theory to calculate that the barrier
in simple sulfonamides is in the range of 5-9 kcal/mol (gas
phase). Thus, even without considering a retardation by the
viscosity of the matrix, conformational lifetimes at 77 K or lower
temperatures in the present work might be expected to be of
the order of at least a few seconds to many centuries. By this
means, in a 1b biradical formed at low temperature in a
particular conformation, and hence in a particular spin state,
the multiplicity might be effectively frozen.
Under these circumstances, it becomes ambiguous to think
of which spin state is energetically preferred unless the
conformation is specified. For the present, there is no experi-
mental basis for assigning particular conformations to particular
spin states, but in the context of this work, the major idea is
that the conformation determines the favored spin. Note that
if the triplet-singlet energy gap is small, a conformationally
induced change in it of the order of a few tenths of a kcal/mol
would suffice to switch the obserVed multiplicity of the ground
state.
Recent high-level ab initio calculations26 on the hypothetical
model compound N-sulfonyl-3,4-dimethylenepyrrole confirm the
proposed dependence of the triplet-singlet energy separation
on conformation. At the CASSCF (CASPT2N)/6-31G* level,
the gap is smaller by 0.6 (0.8) kcal/mol in the conformation
1
observations is the switch from singlet product 1b-d formed
3
in 370 nm irradiation to triplet product 1b-d formed in 265
nm irradiation. One plausible explanation involves excitation
of the (localized) arenesulfonyl chromophore of diazenes 5b-
d, intersystem crossing to the excited triplet state, and intra-
molecular transfer of triplet energy to the (localized) diazene
chromophore. At present, we have no decisive experimental
evidence on these speculations, but the Supporting Information
discusses the issue further.
Search for Thermal Singlet-Triplet Interconversion in
Soft Matrices. To decrease the contribution of the matrix to
the overall conformational rotational barrier, thus facilitating
interconversion of the spin isomers of biradical 1b, one needs
a matrix significantly less viscous than those used in the studies
already described. However, the medium still must be suf-
ficiently viscous to prevent rapid motional averaging27 of the
ESR anisotropic interactions and permit observation of the triplet
transitions. We have photolyzed the N-tosyl diazene precursor
5b in a 2:3 (v/v) mixture of 1- and 2-propanols. This medium
previously has been shown in this laboratory27b to support the
characteristic ESR spectrum of the TMM biradical 2-isopro-
pylidenecyclopentane-1,3-diyl at temperatures as high as 143.6
K, where its viscosity η ∼ 2000 P.28 Under these conditions,
and even at temperatures down to 120.8 K, the biradical’s
mobility is sufficient to permit observation of its dimerization
kinetics and intermolecular capture by acrylonitrile on a
convenient time scale.
Although we have now observed what seem to be definite
motional effects on the zfs parameter |D/hc| in these and other
matrices containing 31b or other triplet species (see the
Supporting Information), we find no evidence for thermal
interconversion of the singlet and triplet states of 1b in propanol
(27) (a) Weissman, S. A. J. Chem. Phys. 1958, 29, 1189. (b) Platz, M.
S.; Berson, J. A. J. Am. Chem. Soc. 1980, 102, 2358.
(28) Greenspan, H.; Fischer, E. J. Phys. Chem. 1965, 69, 2466.
(29) Breneman, C. M.; Weber, L. W. Can. J. Chem. 1996, 74, 1271.
We thank C. M. Breneman for providing this information before publication.
(25) We are grateful to Professors Salem, Borden, and Davidson, whose
unanimous resistance to this idea forced us to examine its implications more
closely.
(26) Borden, W. T.; Hrovat, D. Personal communication, 1995.

1426 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Bush et al.
concentrated. Purification was achieved by column chromatography
(silica gel, 1:9 ethyl acetate/hexanes) gave a white solid (0.263 g,
57%): mp 120-122 °C; H NMR (250 MHz, CDCl3) δ 7.52 (s, 1H),
7.43-7.28 (m, 4H), 7.19-7.07 (m, 5H), 4.55 (s, 2H), 4.21 (s, 2H),
2.35 (s, 3H); 13C NMR (62.9 MHz, CDCl3) δ 145.22, 134.96, 134.31,
131.46, 129.49, 129.15, 128.34, 127.59, 127.38, 122.78, 122.31, 121.64,
23.85, 21.54; D/EI (20 eV) m/z 485 (M+ + 4, 15%), 483 (M+ + 2,
27%), 481 (M+, 14%), 404 (M+ + 2 - Br, 64%), 402 (M+ - Br,
59%), 323 (M+ - 2Br, 14%), 168 (M+ - Ts - 2Br, 100%).
matrices so far. The triplet signal persists up to about 90 K,
but no blue color is visible. Similarly, blue samples of 1b singlet
at temperatures up to 90 K do not show the characteristic ESR
resonances of triplet 1b. Since the blue samples generated in
propanols usually contain one or both of the the diazenyl triplets
11b and 12b, the latter experiment also serves to show that
thermal deazetation of the diazenyl biradicals, a reaction that
might have given the 1b triplet, does not occur in this medium.
This result resembles the findings already described in MTHF
glasses.
Conclusions. Both the singlet and the triplet states of the
three N-arenesulfonyl-3,4-dimethylenepyrrole biradicals 1b-d
can be prepared in glassy matrices. They are individually
kinetically stable for many days and thus constitute examples
of long-lived spin isomerism. The intersystem crossings are at
least 15 orders of magnitude slower than those of typical triplet-
singlet pairs. The most plausible current explanation for this
retardation is that the molecular conformation in 1b-d deter-
mines which spin state is energetically preferred. Intersystem
crossing then must be coupled to conformational isomerization
about the N-S and/or S-C bonds of the biradicals, which can
be a very slow process at low temperatures.
Conformational control of the ground state multiplicity is a
delicate indirect effect. It depends upon the circumstance that
conformational isomerization30 alters the singlet-triplet energy
gap by a small but finite amount (computationally26 less than 1
kcal/mol). In order for such control to operate as proposed for
the present examples, the energies of the triplet and singlet must
be nearly the same. This would be in accord with the
computationally based prediction1a,c that a sufficiently EW
N-substituent will damp the heteroatom lone-pair perturbation
in a 3,4-dimethylenepyrrole and restore the multiplets to the
near-degeneracy characteristic of the disjoint hydrocarbon TME.
1
1′,2′,3′,4′-Tetrahydro-2′,3′-di-tert-butoxycarbonyl-5′-phenyl-6′-p-
toluenesulfonylpyrrolo[3,4-d]pyridazine (10). A mixture of potas-
sium tert-butoxide (0.157 g, 1.33 mmol), 18-crown.-6 (0.064 g, 0.242
mmol), di-tert-butylhydrazine dicarboxylate (0.137 g, 0.590 mmol) in
16 mL of THF was heated to reflux for 1.5 h. To this was slowly
added the above dibromide (0.285 g, 0.590 mmol) in 8 mL of THF.
The reaction mixture was brought to reflux for another 9 h. After
workup, a brownish orange oil was obtained. The crude product was
purified by column chromatography (silica gel, 1:2 ether/hexanes) to
yield pale yellow solids (0.194 g, 59%). NMR showed that two
conformers were present: 1H NMR (300 MHz, CDCl3) δ 7.31-7.08
(m. 10H), 5.08 (bd. d, J ) 15 Hz, 1H), 4.88 (minor conformer, bd. d,
J ) 15 Hz), 4.72 (bd. d, J ) 15 Hz, 1H), 4.47 (minor conformer, bd.
d, J ) 15 Hz), 4.23 (minor conformer, bd. d, J ) 15 Hz), 4.11 (bd. d,
J ) 15 Hz, 1H), 3.97 (minor conformer, bd. d, J ) 15 Hz), 3.87 (bd.
d, J ) 15 Hz, 1H), 2.34 (s, 3H), 1.42 (s, 18H); 13C NMR (75.5 MHz,
CDCl3) δ 154.3, 144.84, 35.52, 131.11, 129.50, 128.58, 127.71, 127.27,
120.29, 118.65, 117.20, 81.27, 41.11, 42.70 (minor conformer), 41.26
(minor conformer), 40.92, 28.30, 28.20, 21.65; D/EI (20 eV) m/z 553
t
(M+, 3%), 397 (M+ - TsH, 28%), 242 (M+ - Ts - CO2 Bu - tBu +
H2, 84%), 57 (tBu, 100%); HRMS (FAB) calcd for C29H35N3O6S(Na+)
576.2144, found 576.2143.
1′,2′,3′,4′-Tetrahydro-5′-phenyl-6′-p-toluenesulfonylpyrrolo[3,4-
d]pyridazine. The hydrazine was generated from biscarbamate 10 in
the same manner as described for corresponding transformations in the
accompanying paper:1c 1H NMR (250 MHz, CDCl3) δ 7.32-7.10 (m,
10H), 3.98 (s, 2H), 3.69 (s, 2H), 2.34 (s, 3H).
1′,4′-Dihydro-5′-phenyl-6′-p-toluenesulfonylpyrrolo[3,4-d]-
pyridazine (5d). Diazene 5d was obtained by DMAD oxidation of
the hydrazine at -30 °C using the same procedure as described for
similar transformations.1c
Experimental Section
Photochemical, UV-vis spectroscopic, and solvent purification
techniques were similar to those described earlier.1a,5e,6a ESR spectro-
scopic procedures and methods for the AM1-CI calculations are given
in the Supporting Information to our preliminary communication.1a
Synthesis of Precursor 5d (Scheme 1). 2-Phenyl-3,4-bis(meth-
oxymethyl)-N-p-toluenesulfonylpyrrole (9). 2-(Trimethylstannyl)-3,4-
bis(methoxymethyl)-N-p-toluenesulfonylpyrrole (8)1d (1.96 g, 4.16
mmol) and bis(acetonitrile)-palladium dichloride (0.1 g, 0.386 mmol)
were added to a 25 mL round bottom flask, which was then evacuated
and flushed with argon three times. To this was added iodobenzene
(0.38 mL, 3.40 mmol), followed by 10 mL of anhydrous 2-methyl-1-
pyrrolidinone.
The mixture was stirred under argon for 16 h at room temperature.
After standard workup, the crude product was purified by column
chromatography (silica gel, 1:4 ethyl acetate/hexanes) to yield a clear
oil (0.384 g, 29%): 1H NMR (250 MHz, CDCl3) δ 7.41 (s, 1H), 7.34-
7.26 (m, 4H), 7.18-7.01 (m, 5H), 4.38 (s, 2H), 3.98 (s, 2H), 3.36 (s,
3H), 3.11 (s, 3H), 2.29 (s, 3H); 13C NMR (62.9 MHz, CDCl3) δ 144.56,
135.41, 133.93, 131.85, 129.32, 129.20, 128.40, 127.13, 127.08, 123.29,
121.09, 66.42, 64.95, 57.98, 57.50, 21.31; D/EI (20 eV) m/z 385 (M+,
28%), 354 (M+ - OMe, 6%), 323 (M+ - 2OMe, 4%), 198 (M+ - Ts
- MeOH, 100%).
3
Test for Equilibration of the Triplet and Singlet Biradicals 1b
1
and 1b as the Cause of the Slight Curvature in the Curie Law
Plots. We observed that a 0.005 M sample of tosyldiazene 5b, when
photolyzed at 370 nm for 20-30 min appears blue. At 77 K no hint
of ESR signal attributable to 31b was observed. Assuming, quite
conservatively, that only a volume of 50 µL of the sample is actually
“read” and that the total amount of diyl is only 5% of the initial diazene
concentration, as is suggested by the diminution of the UV-vis
absorption:
0.005 M (50 µL) ) 2.5 × 10-7 mol of diazene in cavity
2.5 × 10-7 mol of diazene (5% diazene to diyl) )
1.25 × 10-8 mol of diyl in cavity
1.25 × 10-8 mol of diyl (6.02 × 1023 molecules/mol) )
7.5 × 1015 molecules of diyl
The fraction of triplet is governed by the Boltzmann distribution. For
a gap of 19 cal/mol,
2-Phenyl-3,4-bis(bromomethyl)-N-p-toluenesulfonylpyrrole (21).
A suspension of the above crude compound 9 (0.366 g, 0.951 mmol),
dibromotriphenylphosphorane (1.07 g, 2.43 mmol), and acetonitrile (15
mL) was brought to reflux for 5 h. After cooling to room temperature,
the mixture was poured into saturated sodium bicarbonate solution (30
mL). This was extracted with ether (4 × 30 mL). The ether layers
were combined, washed with water (50 mL), dried over MgSO4, and
3e(-∆E/RT)
1 + 3e(-∆E/RT) 1 + 3e(-19/1.99 (77))
3e(-19/1.99 (77))
ftriplet
)
)
) 0.73
Now we calculate the number of spins which should be present:
7.5 × 1015 (73%) ) 5.5 × 1015
two spins per triplet biradical ) 1.1 × 1016 spins
(30) Although not mediated by internal rotational conformation, spin
isomerism has been observed in transition metal complexes. Review:
Gu¨tlich, P.; Hauser, A.; Spiering, H. Angew. Chem., Int. Ed. Engl. 1994,
33, 2024.
This extremely conservative value is still 1 order of magnitude greater

Triplet Species from Dihydropyrrolo[3,4-d]pyridazines
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1427
than the nanomolar detection limit of the instrument. We conclude,
Table 3. Product Ratios from the Experiments of Scheme 4
3
therefore, that either the triplet 1b is not in rapid equilibrium with
relative %
relative %
NMM Product 20
1
anything, or if it is then it is not with the blue species 1b.
time, min
DMAD Product 18
Synthesis of the Rearranged Bis-carbamate 1′,2′,3′,4′-Tetrahydro-
2′,3′-dicarbomethoxy-5′-p-toluenesulfonylpyrrolo[3,4-d]-
pyridazine 19 by Photo-Fries Rearrangement of 18. The N-tosyl
bis methyl carbamate 18 has been reported elsewhere6a and was prepared
in the present work by methods described1c for the synthesis of the
corresponding bis di-tert-butylcarbamate. The 1H NMR spectrum
matched that found previously.6a Irradiation of the N-tosyl bis-
carbamate 18 in solution in benzene, acetonitrile, THF, MTHF, or
methanol with the 1000 W Hg-Xe arc either unfiltered or NiSO4-
filtered gave the rearranged bis-carbamate 19 as product, together with
a small amount of starting material. In some of the solvents, traces of
the detosylated bis-carbamate were observed. In a specific run, a
solution of 30 mg of 18 in 2 mL of methanol was was irradiated without
filter for 50 m under a N2 atmosphere at room temperature. The solvent
was evaporated, and the dark brown residue was chromatographed on
silica gel with 50% EtOAc in pentane to give 6 mg (20%) of the
rearranged bis-carbamate 19. With NiSO4, the same product again was
observed, but the yield was lower. Under 366 nm irradiation, no
rearranged product was observed: 1H NMR (CDCl3) δ/ppm 9.9 (1H,
bs), 7.77 (2H, d J ) 8.3 Hz), 7.28 (2H, d J ) 8.3 Hz), 6.76 (1H, d J
) 2.5 Hz), 5.32-4.32 (4H, m), 3.75 (3H, s), 3.74 (3H, s), 2.40 (3H,
s); HRMS (CI) calcd for C17H20N3O6S (M + 1) 394.107282, found
394.108368.
Solution Phase Direct Photolysis of Diazene 5b by NiSO4-Filtered
Light in the Presence of Maleonitrile (MN). A sample of diazene
compound 5b1a-c (2 mL, 15 mM) and maleonitrile (0.032 mmol, 0.4
mL, 0.08 M) in 2-MTHF was degassed and sealed in a square quartz
tube equipped with a Pyrex extension. The sample was photolyzed
with NiSO4-filtered light for 30 min at -40 °C. It was then treated
with N-methylmaleimide (NMM) (0.875 mL, 1 M) and allowed to warm
to room temperature to scavenge any unreacted diazene.1c HPLC
analysis (20% EtOAc in hexane) showed no detectable product.
Apparently the products of the reaction are very unstable under these
reaction conditions. This was confirmed by separate control experi-
ments, which showed efficient photodecomposition of maleonitrilediyl
1b adducts under these conditions.
15
30
60
22
30
50-58
78
70
50-42
Irradiation with NiSO4-filtered light (1 M aqueous NiSO4, peak
transmission 230-325 nm) gave some change in the spectrum,
including new bands at ∼390 and ∼360 nm, which made difficult an
estimate of how much the 370 nm band decreased.
Direct Photolyses of N-Tosyl Diazene 5b with 265 nm, 366 nm,
and NiSO4-Filtered Light in MTHF Matrix Containing DMAD.
Three samples of diazene 5b (1 mL, 15 mM) in 2-MTHF and DMAD
(4 mg, 0.02 mmol) were sealed in the standard way in square quartz
tubes. One sample was photolyzed at 77 K at 265 nm for 20 min;
another at 366 nm for 15 min; the third with NiSO4-filtered light for
15 min. Each sample was then allowed to warm to room temperature
in the dark. The products were analyzed by 1H-NMR in CDCl3. The
only trapping product observed was the fused N-tosyl bis-carbamate
18. The spectrum also showed no growth of the rearranged 2-tosyl
bis-carbamate 19.
In another experiment (Scheme 4), a sample of diazene 5b (0.7 mL,
8.5 mM) in 2-MTHF and DMAD (20.43 mM) was placed in a capped
square quartz tube and purged with nitrogen. Three samples were
photolyzed by NiSO4-filtered light at 77 K for 15, 30, and 60 min,
respectively. After irradiation the tubes were stored at -80 °C, and
N-methylmaleimide (0.5 mL, 1 M) was added to scavenge unreacted
diazene. The samples then were allowed to warm to room temperature
in the dark. The products were analyzed by HPLC using 40% EtOAc/
hexane as a solvent system. The product peaks were assigned by
comparison to pure samples. The comparison of three runs (Table 3)
showed relative growth of trapping product N-tosyl bis-carbamate 18
with increasing reaction time, and relative decrease of the N-Me
maleimide trapped product 20. The HPLC chromatograms also showed
the absence (<0.1%) of the rearranged 2-tosyl bis-carbamate 19 in the
crude photolysis product.
Direct Photolyses of Diazene 5b. Extent of Deazetation by 265
nm, 343 nm, 366 nm, and NiSO4-Filtered Light by UV-Vis
Spectroscopy. These experiments were designed to monitor the extent
of deazetation in N-tosyl diazene 5b at 77 K using different filters.
The sample of azo compound 5b (15 mM solution in MTHF) was
irradiated at 265 or 343 nm monochromatic light at 77 K for 100 min
total time. The wavelengths were selected by means of an Oriel 7725
monochromator. Comparison of the UV spectra of the sample after
100 min total irradiation with the spectra prior to irradiation showed
no change.
Acknowledgment. This work was supported by the National
Science Foundation, by Dox Graduate Fellowships to L.C.B.
and H.S.M.L., and by an Arthur C. Cope Scholar Award of the
American Chemical Society. We thank W. T. Borden and D.
Hrovat (ref 26) and C. M. Breneman (ref 29) for information
in advance of publication.
Supporting Information Available: Curie Law, spin-count,
and temperature-dependent ESR zero-field splitting experiments
(37 pages). See any current masthead page for ordering and
Internet access instructions.
Irradiation by (mostly) 366 nm light (passed by Oriel colored glass
filter of peak transmission 328-381 nm) of N-Tos diazene 5b at 77 K
for 13 min total irradiation showed 28.8% decrease of the 370 nm band
in the spectrum.
JA963114C