The Role of Nonaromatic Isomers in the Photochemistry of 3,5-Dimethoxybenzyl Acetate

Full text of DOI:10.1021/jo9722616

434
J . Org. Chem. 1998, 63, 434-435
1
of the H NMR spectrum of the crude reaction mixture,
using the signals previously reported and assigned to 2,
the yield of the photoisomer was 9%. This compound was
then partially purified by preparative chromatography.11
The original assignment of 2 was made only on the basis
Th e Role of Non a r om a tic Isom er s in th e
P h otoch em istr y of 3,5-Dim eth oxyben zyl
Aceta te
Frances L. Cozens, Alexandra L. Pincock,
J ames A. Pincock,* and Richard Smith
1
of 100 MHz H NMR spectra, and two of the reported
features were puzzling. First, the signal for Ha was at δ
3.58, a very low value for a doubly allylic methine
hydrogen at an ester carbon (an approximate value of δ
6.6 can be predicted using the corresponding CH2 in
o-isotoluene (C7H8) at δ 3.17 (C6H6)12 and δ 3.32 (CCl4)13
and the expected substituent effect of an ester oxygen of
δ 3.414) and, second, the two methoxy groups, although
in very similar chemical environments, gave quite dif-
ferent chemical shifts (δ 3.24 and δ 3.62). When 1H-
13C-correlated spectra15 demonstrated that both Ha and
He were on sp3-hybridized carbons at δ 76.6 and δ 52.6,
respectively, assignment of the structure 2 to this com-
pound became untenable. In contrast, the NMR spectra
are consistent with structure 5, particularly when the
1H and 13C chemical shifts, after correcting for substitu-
ent effects, are compared with those of the previously
reported13 unsubstituted compound, C7H8, another isomer
of toluene. Moreover, the structure of 5 makes chemical
sense, because an obvious mechanism for its formation
is the allowed disrotatory butadiene to cyclobutene ring
closure by secondary photochemistry of 2.16
As expected, freshly isolated samples of 5 were trans-
parent at 320 nm; however, 5 slowly isomerized to 2, and
this first-order conversion could be monitored by either
1H NMR or the increasing absorbance at 320 nm (t1/2
approximately 10 h at room temperature).17 The conver-
sion of 5 to 2 was cleanly first order in hexanes at 50 °C
(t1/2 ) 89 min)18 and 2 could then be isolated, spectra19
taken, and its reactivity studied. NMR samples of 2 in
CDCl3 were relatively unstable as 2 rearranged to the
toluene derivative 6; this process was quantitative after
2 h at 50 °C.20 In acetonitrile, 2 rearranges back to
Department of Chemistry, Dalhousie University, Halifax, Nova
Scotia B3H 4J 3, Canada
Received December 15, 1997
More than 20 years ago, J aeger reported1,2 that pho-
tolysis of 3,5-dimethoxybenzyl acetate (DMBA), 1, in
hexane, gave a 17% yield of the isomer 2 along with
radical-derived coupling products, the in-cage one, 3
(47%), being the major one (Scheme 1). We were in-
trigued by the possibility that 2 might also be a primary
photoproduct in polar, nucleophilic solvents, like metha-
nol, where photolysis of DMBA gives, as the major
product, the ether 4 (56%),3 clearly derived from an
intermediate arylmethyl cation. If 2 is formed in metha-
nol, it might be expected to undergo rapid ground-state
solvolysis to 4 complicating mechanistic arguments4,5
centered on the yield of ion-derived products in this
“photosolvolysis” reaction, i.e., is 4 formed entirely in the
primary photochemistry of DMBA or, at least in part, in
a secondary ground-state process from 2? This is an
important question because DMBA, and other multiple
methoxy substituted benzyl acetates, has been shown3
to give a higher yield of ion-derived products than
expected on the basis of a mechanism that emphasizes
formation of ion pairs from radical pairs by electron
transfer. Moreover, on the basis of the pioneering study
by Zimmerman6 on the influence of meta methoxy sub-
stituents on benzene excited state reactivity, the 3,5-
dimethoxybenzyl chromophore in benzoin derivatives has
recently been advocated in the design of practical ex-
amples of photo labile protecting groups.7,8
We therefore decided to isolate and examine the
reactivity of 2 and now report that (1) the compound
previously reported as 2 is, in fact, the bicyclic isomer 5;
(2) 2 can be prepared from 5 by pyrolysis in hexane; (3)
2 does undergo rapid ground-state solvolysis in methanol;
and (4) 2 is formed as a primary photoproduct of DMBA
in methanol.
(11) Silica gel dry-flash chromatography using 5% ethyl acetate in
hexanes as eluant gave fractions with purity as high as 90% but always
contaminated with DMBA.
(12) Kopecky, K. R.; Lau, M.-P. J . Org. Chem. 1978, 43, 525-526.
(13) Hasselmann, D.; Loosen, K. Angew. Chem., Int. Ed. Engl. 1978, 17,
606-608.
(14) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric
Identification of Organic Compounds, 5th ed.; J ohn Wiley & Sons: New
York, 1991; p 213.
Photolysis9 in hexanes results, as previously reported,1
in the rapid10 disappearance of DMBA. By integration
(15) NMR spectra for 5: 1H NMR (CDCl3 at 250 mHz) δ 5.66 (Ha), 5.24
(Hc or Hd), 5.16 (Hc or Hd), 4.74 (Hb), 3.71 (3H, s), 3.66 (He), 3.33 (3H, s) and
2.09 (3H, s); the observed multiplets could be simulated using the program
NMRSIM (Dr. T. P. Forrest, Chemistry Department, Dalhousie) to give the
following coupling constants, J (Hz): eb (1.22), ec (0.91), ed (1.83), ca. (2.44),
* To whom correspondence should be addressed. Tel.: (902) 494-3324.
Fax: (902) 494-1310. E-mail: PINCOCK@CHEM1.CHEM.DAL.CA.
(1) J aeger, D. A. J . Am. Chem. Soc. 1974, 96, 6216-6217.
(2) We thank Professor J aeger for sending us his original laboratory
notebooks, including spectra. This material was of invaluable help. Although
we are reinterpreting his observations, the quality of his experiments and
the clarity of the notebooks would serve as models for anyone. Without the
visual incentive of his spectra, we would have had even more difficulty
completing these experiments.
(3) Pincock, J . A.; Wedge, P. J . J . Org. Chem. 1994, 59, 5587-5595.
(4) Pincock, J . A. Acc. Chem. Res 1997, 30, 43-49.
(5) Zimmerman, H. E. J . Am. Chem. Soc., 1995, 117, 8988-8991.
(6) Zimmerman, H. E.; Sandel, V. R. J . Am. Chem. Soc. 1963, 85, 915-
922.
(7) Pirrung, M. C.; Bradley, J .-C. J . Org. Chem. 1995, 60, 1116-1117.
(8) Cameron, J . F.; Wilson, C. G.; Frechet, J . M. J . J . Am. Chem. Soc.
1996, 118, 12925-12937.
(9) Vycor-filtered, 450 W, medium-pressure Hanovia mercury lamp.
(10) Greater than 90% conversion of 400 mg of 1 in 280 mL of hexanes
after 30 min as determined by HPLC on silica gel with 5% ethyl acetate in
hexanes as eluant.
and da (2.14); 13C NMR (CDCl3 at 250 mHz)
δ 170.5 (CdO), 158.3
(dCOCH3), 144.6 (CdCH2), 112.3 (CdCH2), 95.7 (CHb), 81.0 (COCH3), 76.6
(CHa), 56.9 (CH3O), 52.6 (CHe), 52.3 (CH3O), and 21.1 (CH3). Assignments
were made on the basis of 1H-1H and 13C-1H correlated spectra.
(16) Although we have no evidence to support the assignment, the acetate
functional group is assumed to be in the more stable exo position.
(17) This observation clarified several experimental observations. First,
the initial report1 on the isolation of 5 (thought to be 2) included a UV
absorption band at 315 nm with ꢀ ) 400 M-1 cm-1; the wavelength agrees
with expectations for 2 but the ꢀ value is too low for a conjugated triene.
Our estimate of ꢀ ) 5500 M-1 cm-1 at 320 nm is more reasonable,
particularly when compared to the value of 4400 reported12 for the
unsubstituted compound, o-isotoluene (C7H8). Second, in our attempts to
isolate the photoisomer, we monitored chromatography fractions by HPLC
with UV detection at 320 nm. Those fractions that contained 5, when first
obtained, were transparent but, after standing overnight at room temper-
ature, were strongly absorbant. Both of these observations result from
increasing amounts of 2 contaminating 5 with time.
(18) For the unsubstituted compound, a value of t1/2 ) 210 min in THF
can be calculated from the Arrhenius values reported.13
S0022-3263(97)02261-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/22/1998

Communications
J . Org. Chem., Vol. 63, No. 3, 1998 435
Sch em e 1
DMBA, presumably because this more polar solvent
induces contact ion pair formation. Of particular rel-
evance to the photochemistry of DMBA in polar, nucleo-
philic solvents, the half-life for first-order solvolysis of 2
in methanol at room temperature is only 2.7 min at 25
°C! The products result from partitioning of the ion pair
between nucleophilic capture by methanol to give the
methyl ether, 4 (60%) and internal return to DMBA
(40%).
Finally, to demonstrate that 2 is formed in methanol,
laser flash photolysis21 of DMBA at 266 nm gave an
absorbance spectrum with a maximum at 320 nm that
remained constant on the millisecond time scale. More-
over, samples subjected to multiple 266 nm laser pulses
could be transferred to a diode array spectrometer and
the decay of the photogenerated transient monitored. The
lifetime obtained was identical to that obtained from the
preparative-scale NMR samples, conclusively demon-
strating that 2 is a primary photoproduct in methanol.22
These observations have important implications in the
ongoing debate concerning the mechanism for formation
of ion-derived products in the photochemistry of benzyl
acetates.4,5 There are three possible mechanisms for the
formation of 2 from DMBA in methanol: (1) concerted
sigmatropic rearrangement; (2) recombination of an
initially formed radical pair; and (3) recombination of an
initially formed ion pair. If either mechanism (1) or (2)
is important in nucleophilic solvents such as methanol,
then the yield of ion-derived product, the methyl ether
4, will be increased by a ground-state pathway for its
formation that is independent of the photochemical
pathway. By analogy with the photo-Fries reaction
where solvent caged radical pairs generated from the
singlet excited state are proven intermediates,23 mech-
anism 2 would seem to be preferred.
Moreover, preliminary results for DMBP (the tert-
butyl, or pivalate, ester derivative of 1) also support
mechanism 2. Both DMBA and DMBP produce an
absorbance at 320 nm after multiple 266 nm laser pulses.
For DMBA, after the decay of 2 is complete, the residual
absorbance remaining is still 15% of the starting absor-
bance value. This residual absorbance is assigned to the
more stable alkyl-substituted triene product that would
result by combination of the radical pair after the
decarboxylation of the acyloxy radical.24 In contrast, for
DMBP, the residual absorbance is 77% of the original
value. For the radical pair in this case, decarboxylation
of the acyloxy radical is faster25 so that formation of the
solvolytically reactive trienyl ester26 is less efficient
relative to the formation of hydrocarbon products.
Future experiments will be directed toward establish-
ing the photochemical yield and reactivity of isomers
analogous to 2 as a function of substrate (substituents
and leaving group) and reaction conditions (solvent,
wavelength of irradiation, etc.).
Ack n ow led gm en t. We thank NSERC of Canada for
funding for this research and for a WF Award to F.L.C.
We also thank Dr. Norman Schepp for assistance with the
laser experiments.
(19) Samples of 2 were still contaminated with the DMBA that had been
in 5. The 1H NMR spectrum of 2 in CDCl3 could not be analyzed because
the signals for Hb, Hc, Hd, and He were almost coincident at 250 mHz, giving
an unresolved multiplet. In contrast, these signals were well dispersed in
acetonitrile: 1H NMR (CD3CN at 250 mHz with δ values reported relative
to a value of δ 5.04 for the methylene hydrogens of DMBA, their chemical
shift in CDCl3) δ 6.10 (Ha), 5.33 (Hb), 5.24 (He), 5.19 (Hc or Hd), 5.08 (Hc or
Hd), 3.67 (3H, s), 3.65 (3H, s), 2.04 (3H, s); 13C NMR (CD3CN) δ 158.4 (C),
142.8 (C), 115.9 (CH2)), 97.5 (CHe), 92.0 (CHb), 69.3 (CHa), 56.3 (CH3O),
55.7 (CH3O), 21.4 (CH3). Assignments were made on the basis of 1H-13C
correlated spectra.
J O9722616
(23) (a) J iminez, M. C.; Miranda, M. A.; Scaiano,. C.; Tormos, R. Chem.
Commun. 1997, 1487-1488. (b) Mirand, M. A. In Handbook of Organic
Photochemistry and Photobiology; Horspool, W. M., Song, P. S., Eds.; CRC
Press: Boca Raton, 1995; pp 570-578.
(24) We have not been able to isolate this compound because it is
contained in chromatograpy fractions that are mixtures of hydrocarbons
dominated by the aromatic isomer, 3,5-dimethoxyethylbenzene.
(25) Hilborn, J . W.; Pincock, J . A. J . Am. Chem. Soc. 1991, 113, 2683-
2686.
(26) The rate of solvolysis of the trienyl ester from DMBP is a factor of
2 slower than for 2. This is in agreement with the fact that acetic acid (pKa
) 4.75) is a slightly stronger acid than pivalic acid (pKa ) 5.03), and
therefore, acetate is the better leaving group.
(20) J aeger1 reported the same process in benzene at 50 °C, and this
observation was used as support for the assignment of the structure 2 to
the photoisomer isolated. In fact, the process observed was thermal
conversion of 5 to 2 to 6.
(21) Continuum NY-61 Nd:YAG laser at 266 nm, e15 mJ /pulse, e8 ns/
pulse.
(22) Similar laser experiments as a function of temperature in methanol
gave ∆Hq ) 16 kcal/mol and ∆Sq ) -14 eu and in other solvents (ethanol,
2-propanol) at 25 °C gave a linear mY correlation with Y ) 0.57.