1,3-Dimethoxy-5-methylene-1,3-cyclohexadiene Compounds with Leaving Groups at C6: Generation, Solvolytic Reactivity, and Their Importance in the Photochemistry of 3,5-Dimethoxybenzyl Derivatives

Article abstract of DOI:10.1021/jo000387o

The photochemistry of 3,5-dimethoxybenzyl compounds with the leaving groups acetate (1a), chloride (1b), bromide (1c), iodide (1d), diethyl phosphate (1e), and trimethylamine (1f), as the chloride, was examined by both product studies and flash photolysis. The isomeric triene, 5-methylene-1,3-cyclohexadiene derivative was observed for the acetate (2a), diethyl phosphate (2e) and trimethylammonium chloride (2f). The solvolysis of these derivatives, 2, was examined in alcohol solvents and the rate correlation with Y_OTS values gave m = 0.47 (2a) and 0.63 (2e), suggesting S_N1 reactivity but with an early transition state. Quantum yields for formation of 2a and 2e indicated that these trienes play only a minor role (～16%) in the overall photochemistry of the corresponding arylmethyl substrates.

Full text of DOI:10.1021/jo000387o

4698
J . Org. Chem. 2000, 65, 4698-4705
1,3-Dim eth oxy-5-m eth ylen e-1,3-cycloh exa d ien e Com p ou n d s w ith
Lea vin g Gr ou p s a t C6: Gen er a tion , Solvolytic Rea ctivity, a n d
Th eir Im p or ta n ce in th e P h otoch em istr y of 3,5-Dim eth oxyben zyl
Der iva tives
D. P. DeCosta,† N. Howell, A. L. Pincock, J . A. Pincock,* and S. Rifai
Department of Chemistry, Dalhousie University Halifax, Nova Scotia, Canada, B3H 4J 3, and
Department of Chemistry, Colombo University, Colombo, Sri Lanka
pincock@chem1.chem.dal.ca
Received March 17, 2000
The photochemistry of 3,5-dimethoxybenzyl compounds with the leaving groups acetate (1a ), chloride
(1b), bromide (1c), iodide (1d ), diethyl phosphate (1e), and trimethylamine (1f), as the chloride,
was examined by both product studies and flash photolysis. The isomeric triene, 5-methylene-1,3-
cyclohexadiene derivative was observed for the acetate (2a ), diethyl phosphate (2e) and trimethy-
lammonium chloride (2f). The solvolysis of these derivatives, 2, was examined in alcohol solvents
and the rate correlation with YOTS values gave m ) 0.47 (2a ) and 0.63 (2e), suggesting SN1 reactivity
but with an early transition state. Quantum yields for formation of 2a and 2e indicated that these
trienes play only a minor role (∼16%) in the overall photochemistry of the corresponding arylmethyl
substrates.
In tr od u ction
This observation that the ether 4, formed on photolysis
of 1a in methanol comes, at least in part, from ground
state solvolysis of a photochemically generated interme-
diate 2a , is relevant to the general question of consider-
able recent interest as to the pathway(s) for forma-
tion of ion-derived products in benzylic photosolvoly-
sis reactions.2 As outlined in eq 1, two different pro-
posals have been made: (1) competition between hetero-
lytic (khet) and homolytic (khom) cleavage occurs in the
originally formed excited singlet state and this competi-
tion is the major factor controlling the yields of ion- and
radical-derived products; (2) the main excited-state pro-
cess is homolytic cleavage (khom) and the competition is
then between electron transfer (ketri) converting the
radical pair to the ion pair and other chemistry of the
radical pair (kr). These concepts have been recently
reviewed.2
Recently we reported that the photolysis of 3,5-di-
methoxybenzyl acetate, 1a , in either methanol or hex-
anes, resulted in the formation of the isomeric triene 2a ,
Scheme 1.1 In hexanes, secondary photochemistry con-
verted 2a to another isomer, the bicyclic ester 3, which
was purified by silica gel chromatrography and charac-
terized by 1H and 13C NMR spectroscopy. The bicyclic
isomer 3 could then be converted by thermolysis in
hexane at 50 °C back to 2a which was also characterized
spectroscopically. In methanol, 2a was found to react by
ground-state solvolysis with a half-life of 2.7 min to give
the ether 4 (60%) and the original ester 1a (40%). Finally,
nanosecond laser flash photolysis (LFP) with a Nd:YAG
laser at 266 nm gave a transient with the same absorp-
tion spectrum as 2a and with a lifetime (>10ms) longer
than the instrument’s response time. However, multiple
flashes over about a minute duration allowed sufficient
quantity of this transient to be generated that its decay
could be monitored by conventional UV absorbance
spectra. That this transient was indeed 2a was confirmed
by measuring an identical solvolysis rate in methanol as
that obtained by generating 2a from 3, as described
above.
In the case of arylmethyl esters, kr is mainly the
•
decarboxylation (kCO ) of the acyloxy radical (R-CO2 ). On
2
the basis of the second mechanism, we have used product
yields and the radical clock method to obtain a large
number of values for both ketri and kCO as a function of
2
systematic variation in ester structure that seem to make
chemical sense;3 the ketri values can be fitted to Marcus
(1) Cozens, F. L.; Pincock, A. L.; Pincock, J . A.; Smith, R. J . Org.
Chem. 1998, 63, 434-435.
(2) Fleming, S. A.; Pincock, J . A. Organic Molecular Photochemistry;
M. Dekker: New York, 1999; Vol. 3, pp 211-281.
(3) Pincock, J . A. Acc. Chem. Res. 1997, 30, 43-49.
* To whom correspondence should be addressed. Phone: (902) 494-
3324. Fax: (902) 494-1310.
† Colombo University.
10.1021/jo000387o CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/06/2000

1,3-Dimethoxy-5-methylene-1,3-cyclohexadiene Compounds
J . Org. Chem., Vol. 65, No. 15, 2000 4699
Sch em e 1. P h otoch em ica l F or m a tion a n d Rea ctivity of 2a in Meth a n ol
Ta ble 1. P r od u ct Yield s for th e P h otolysis of 3,5-Dim eth oxyben zyl Aceta te in Alcoh ol Solven ts
a
solvent
ArCH3
ArCH2CH3
ArCH2)-2
isomer 9
ArCH2OS
M.B.b
% Cc
R/Id
YOTS
t-BuOH
i-PrOH
CH3CH2OH
MeOH
1.7
4.7
1.2
3.0
3.8
57.3
47.1
39.1
20.8
27.3
16.6
ND
ND
11.7
ND
1.3
6.0
ND
4.7
ND
23.0
42.4
40.1
59.8
68.8
97
86
88e
96
98
37
34
33
64
21
77/23
58/42
59/41
40/60
31/69
-3.74
-2.83
-1.96
-0.92
+1.77
TFE
a
b
d
Twice the molar yield. Mass balance based on amount of 1a reacted. c Percent conversion. Ratio of radical- to ion-derived product
normalized to 100%. e 9% of an unknown compound was also observed.
theory in both the normal and inverted region and the
photochemistry of 1a . For instance, if 2a is formed from
an in-cage singlet radical pair, as in the accepted mech-
anism for the photo-Fries reaction,11 and then it reacts
by rapid ground-state solvolysis, then the ion-derived
product 4 results from the initial photochemical genera-
tion of radicals, i.e., the homolytic pathway, khom. To
assess the importance and reactivity of the 5-methylene-
1,3-cyclohexadiene isomers 2a -f as a function of leaving
group in 3,5-dimethoxybenzyl chromophore chemistry, we
have now examined the derivatives 1a -f.
kCO values increase, as expected, with the stability of the
2
alkyl radical formed.4 Moreover, recent direct measure-
ment of rates of decarboxylation of acyloxy radicals by
LFP are in agreement with our indirectly determined
ones.5
However, mechanism (1) above, direct heterolytic
cleavage from S1, is certainly occurring for some sub-
strates. For instance, contact ion pairs have been ob-
served by LFP within the 20 ps instrument response time
for a set of substituted diphenylmethyl chlorides followed
by further growth of the ion pair by electron-transfer
conversion of the radical pair.6 Moreover, 3,5-dimethoxy-
benzyl substrates have been advocated as useful chro-
mophores for the development of photolabile protecting
groups on the basis of their ability to give high yields of
ion derived products with good efficiency.7 These studies
were founded on the pioneering study by Zimmerman
and Sandel8 that demonstrated enhanced yields of sol-
volysis products on irradiation of both 3-methoxy and 3,5-
dimethoxybenzyl acetate in aqueous dioxane. These
observations have been rationalized on the basis of MO
calculations, originally at the Hu¨ckel level8 but more
recently at much higher levels,9,10 that demonstrate that
resonance electron donating groups such as methoxy are
better donors, in the excited singlet state, from the meta
(and ortho) positions than from the para position. This
observation, which contrasts with ground state observa-
tions of ortho/para activation has been termed the
photochemical “meta effect”.8-10
Resu lts a n d Discu ssion
Solven t Effects on th e P h otoch em istr y of 1a (X
) OAc). We first examined the effect of solvent on the
product distribution for 1a . These results are shown in
(4) Hilborn, J . W.; Pincock, J . A. J . Am. Chem. Soc. 1991, 113, 2683-
2686.
(5) (a) Bockman, T. M.; Hubig, S. M.; Kochi, J . K. J . Am. Chem.
Soc. 1996, 118, 4502-4503. (b) Bockman, T. M.; Hubig, S. M.; Kochi,
J . K. J . Org. Chem. 1997, 62, 2210-2221.
(6) (a) Deniz, A. A.; Li, B.; Peters, K. S. J . Phys. Chem. 1995, 99,
12209-12213. (b) Lipson, M.; Deniz, A. A.; Peters, K. S. J . Am. Chem.
Soc. 1996, 118, 2992-2997. (c) Lipson, M.; Deniz, A. A.; Peters, K. S.
J . Phys. Chem. 1996, 100, 3580-3586.
(7) (a) Barltrop, J . A.; Schofield, P. J . Chem. Soc. 1965, 4758-4765.
(b) Chamberlin, J . W. J . Org. Chem. 1965, 31, 1658-1659. (c) Wieland,
T.; Birr, C. Peptides 1967, 103-106. (d) Birr, C.; Lochinger, W.;
Stahnke, G.; Lang, P. Liebigs Ann Chem. 1972, 763, 162-172. (e) Birr,
C. Peptides 1972, 72-77. (f) Pirrung, M. C.; Bradley, J .-C. J . Org.
Chem. 1995, 60, 1116-1117. (g) Cameron, J . F,; Wilson, C. G.; Frechet,
J . M. J . J . Am. Chem. Soc. 1996, 118, 12925-12937.
(8) Zimmerman, H. E.; Sandel, V. R. J . Am. Chem. Soc. 1963, 85,
915-922.
Explanation of product yields for ion- and radical-
derived products will be made more difficult to interpret
for both of these mechanistic proposals if the isomeric
triene 2a , is a major product resulting from primary
(9) Zimmerman, H. E. J . Am. Chem. Soc. 1995, 117, 8988-8991.
(10) Zimmerman, H. E. J . Phys. Chem. A 1998, 102, 5616-5621.
(11) Miranda, M. Organic Phtochemistry and Photobiology; CRC
Press: Boca Raton, 1995; p 570.

4700 J . Org. Chem., Vol. 65, No. 15, 2000
Sch em e 2. Mech a n ism for th e P h otolysis of 1a in Alcoh ol Solven ts
DeCosta et al.
Table 1 and eq 2 (Ar ) 3,5-dimethoxybenzyl). All reac-
Pa‚s)14 estimated from the diffusional escape equation
(k ) (RT/2πNr3η))15 assuming 6 Å spheres but faster than
in tert-butyl alcohol (kdif ) 5 × 108 s-1, η ) 5.9 × 10-3
Pa.s.14)
F la sh P h otolysis of 1a . As reported in our prelimi-
nary communication,1 nanosecond laser flash photolysis
at 266 nm of 1a in methanol, either with or without
nitrogen purging, gave a weak transient (λmax at 320 nm)
which did not decay at the long time limit of the
instrument response (10 ms). This transient was assigned
to the triene 2a . Higher concentrations of the transient
were generated by multiple pulses (∼50 at a repetition
rate of 1 Hz) and the decay of the transient monitored
by conventional UV spectra, Figure 1. The same transient
was also generated by classical microsecond flash lamp
photolysis16 and, again, its decay monitored by UV. This
transient was observed in all the solvents in Table 1,
except TFE. First-order rate constants for its decay are
given in Table 2.
The values in methanol with added salt (entries 2-4)
indicate a very modest salt effect, as expected for an SN1
ionization in a solvent as polar as methanol.17 There is
no indication of increased rate in the presence of the
added nucleophile (chloride ion), no internal return
(acetate ion) and no special salt effect (perchlorate ion).
A plot of log(k/T) versus 1/T according to the Eyring
equation (entries 1, 5, and 6) gives ∆Hq ) 16.3 kcal/mol
and ∆Sq ) -14 eu. The ∆Sq value is comparable to those
observed in the solvolysis of benzyl tosylates in 80%
aqueous acetone (YOTS ) -0.94)18 and in 30% ethanol:
70% water (YOTS ) 3.32).19
tions were done at 25 °C with nitrogen purging using all
16 lamps (254 nm) in a Rayonet Reactor. The percent
conversion and mass balance are given for samples taken
after 5 min of irradiation of 100 mg of substrate in 100
mL of solvent. The high percentage conversion after such
short times attests to the very high reactivity of this
particular substrate. In fact the excited-state lifetime of
1a is likely very short (<1 ns) but cannot be determined
because 1a does not fluoresce. In our previous publication
on the photochemistry of 1a in methanol,1 we had not
reported the yield of the isomer 9, because the yield is
low and we were not specifically looking for it. Its
retention time (GC) is nearly identical to that of the
starting material but our current high resolution column
gives baseline separation. It has been characterized
previously12 from irradiation of 1a in hexane and is
derived by tautomerization of the initially formed iso-
meric triene 2a , Scheme 2. It can, therefore, most likely
be regarded as a radical pair derived product as can 3,5-
dimethoxytoluene, 5, 3,5-dimethoxyethylbenzene, 6, and
the dimer 7. These products are rationalized by the
general mechanism shown in Scheme 2. The ratio of
radical-derived to ion-derived products, R/I ) yield[(5 +
6 + 7 + 9)/(8)], changes systematically in favor of the
ion-derived ether 8, as the ionizing power of the solvent
increases as measured by the YOTS values.13 There is also
a systematic increase in the yield of the ethylbenzene as
the viscosity increases and the ionizing power of the
solvent decreases over the series methanol, ethanol,
2-propanol, tert-butyl alcohol. This result is likely a
consequence of the increased lifetime of the radical pair
allowing decarboxylation of the acyloxy radical followed
by in-cage coupling. The rate constant for this decar-
An estimate of the enhanced reactivity of 2a relative
to 3,5-dimethoxybenzyl acetate 1a can be made as
(13) Bentley, T. W.; Llewellyn, G. Prog. Phys. Org. Chem. 1990, 17,
121-158.
(14) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photo-
chemistry; M. Dekker: New York, 1993; p 283.
(15) Benson, S. W. The Foundation of Chemical Kinetics; McGraw-
Hill: New York, 1960; p 495.
(16) Chiang, Y.; Hojatti, M.; Keeffe, J . R.; Kresge, A. J .; Schepp, N.
P.; Wirtz, J . J . Am. Chem. Soc. 1987, 109, 4000-4009.
(17) Raber, D. J .; Harris, M.; Schleyer, P. V. R. Ions and Ion Pairs
in Organic Reactions; J . Wiley and Sons: New York, 1974; Vol. 2, pp
247-374.
(18) Fujio, M.; Goto, M.; Susuki, T.; Akasaka, I.; Mishima, M.; Tsuno,
Y. Bull Chem. Soc. J pn. 1990, 63, 1146-1153.
(19) Hill, E. A.; Gross, M. L.; Stasiewicz, M.; Manion, M. J . Am.
Chem. Soc. 1969, 91, 7381-7391.
boxylation is ∼1 × 109 s-1 4
, slower than the rate of
diffusion (kdif ) 5 × 109 s-1) in methanol (η ) 0.59 × 10-3
(12) J aeger, D. A. J . Am. Chem. Soc. 1974, 96, 6i, 6216-6217.

1,3-Dimethoxy-5-methylene-1,3-cyclohexadiene Compounds
J . Org. Chem., Vol. 65, No. 15, 2000 4701
F igu r e 2. Grunwald-Winstein plots for the first-order decay
of 2a (X ) OAc), 2e (X ) O(PO)(OCH2CH3)2), and 2f (X )
N(CH3)3Cl).
nonaromatic isomer 2a . Because the entropies of activa-
tion are similar for solvolysis of both substrates, this very
large change in rate can be attributed almost entirely to
a lower enthalpy of activation as a consequence of the
decreased stability of the nonaromatic isomer 2a .
Finally, the data in Table 2 (entries 1, 7-9) allow the
construction of a Grunwald-Winstein13 plot. Because
acetate is a poor leaving group, only very limited data is
available for its SN1 reactivity. However, a kinetic study
of solvent effects for ferrocenylmethyl acetate has pro-
vided a YOAc scale21 but it only includes one of the solvents
(ethanol) used in the results in Table 2. We have
therefore used the YOTs scale which gives a fair correlation
(r ) 0.830) with the YOAc scale.21 The mYOTs plot gives m
) 0.47, Figure 2. There is some uncertainty in this value
because of the limited number of solvents examined but
it is certainly low for an SN1 solvolysis. For instance, for
benzyl tosylate solvolysis, using the modified Grunwald-
Winstein equation that includes a term for solvent
nucleophilicity, m ) 0.83 at 50 °C22 and m ) 0.80
(without including the term for solvent nucleophilicity)
for the solvolysis of 4-methoxybenzyl chloride at 25 °C.23
A reasonable explanation for the reduced value for 2a is
that the lower ∆Hq will make the transition state for ion-
pair formation earlier so that the charge development will
be less and the response to ionizing power of the solvent
lower.
F igu r e 1. (a) UV spectra of the decay of the intermediate 2a
in methanol at 25 °C. (b) First-order fit to the decay.
Ta ble 2. Ra te Con sta n ts for th e Disa p p ea r a n ce of
6-Acetoxy-3,5-d im eth oxy-5-m eth ylen e-1,3-cycloh exa d ien e,
2a
entry
k (s-1)/10-3 a
t (°C)
solvent
added
1
2
3
4
5
6
7
8
9
3.80 ( 0.06
4.27 ( 0.03
4.15 ( 0.02
4.22 ( 0.10
6.85 ( 0.09
2.25 ( 0.08
1.26 ( 0.05
0.500 ( 0.009
0.164 ( 0.008
25.0
25.0
25.0
25.0
30.5
18.9
25.0
25.0
25.0
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
EtOH
0.1 M NaCl
0.1 M NaOAc
0.1 M NaClO4
The products obtained in the solvolysis of 2a were
examined only in methanol. This was done by converting
3 (isolated by column chromatography) back to 2a by
thermolysis in hexane at 50 °C. This reaction could be
monitored by the appearance of 2a by UV spectra, Figure
i-PrOH
t-BuOH
a
Errors quoted are the average deviations for at least three
determinations of the rate constant.
3. The first-order rate constant is 1.3 × 10-4 s-1
.
follows. The rate constant for solvolysis of 3-methoxy-
benzyl tosylate in 80% acetone: 20% water (YOTS ) -0.94)
Evaporation of the hexane and addition of methanol
resulted in the formation of 3,5-dimethoxybenzyl acetate,
1a (X ) OAc), (40%) and 3,5-dimethoxybenzyl methyl
at 25 °C is 1.08 × 10-5 s-1 18
.
The additional methoxy
1
group in 1a and the change in solvent to methanol (YOTS
) -0.92) should have only a minor effect on the rate.
However, the change in leaving group from tosylate to
acetate gives an estimated change in rate by a factor 3.8
× 10-11, on the basis of leaving group abilities obtained
from the solvolysis of 1-phenylethyl derivatives.20 There-
fore, the estimated rate constant for solvolysis of 3,5-
dimethoxybenzyl acetate, 1a , in methanol at 25 °C is 4.2
× 10-16 s-1, approximately 1013 slower in rate than its
ether 4, (60%) as determined by H NMR integration.
Qu a n tu m Yield for F or m a tion of 2a (X ) OAc)
fr om 1a (X ) OAc). The quantum yield of formation (Φ
(20) Noyce, D. S.; Virgilio, J . A. J . Org. Chem. 1972, 37, 2643-2647.
(21) Kevill, D. N.; Pinhas, A. K. J . Org. Chem. 1993, 58, 197-201.
(22) Kevill, D. N.; D’Souza, M. J .; Ren, H. Can. J . Chem. 1998, 76,
751-757.
(23) Kevill, D. N.; D’Souza, M. J . J . Chem. Res., Synop. 1999, 336-
337.

4702 J . Org. Chem., Vol. 65, No. 15, 2000
DeCosta et al.
ally three dimers. They were shown to be the bibenzyl
derivative 7, eq 2, (65% of the total), and the two possible
aromatic coupled products 10 (25%) and 11 (10%). The
latter two products are formed by ortho and para coupling
of one arylmethyl radical to the other as has been
observed previously for unsubstituted benzyl radicals.27,28
The ratio of ortho to para coupling is almost statistical
(2:1) and quite different from that for benzyl radicals
where the ortho to para ratio in benzene at 28 °C is 1.04,
i.e., para is formed in a yield that is higher than
statistically predicted.28 The first of these dimers was
available from previous work in our laboratory29 and 10
and 11 were synthesized by Friedel-Crafts coupling of
3,5-dimethoxybenzyl chloride with 3,5-dimethoxytoluene.
These two dimers were not separated (despite attempts
on both normal and reverse phase silica gel) but could
be easily distinguished by 1H NMR at 400 MHz, 10
having three different methoxy groups with an intensity
ratio of 2:1:1 whereas 11 has only two different methoxy
groups in the ratio of 2:2.
F igu r e 3. (a) UV spectra for the formation of 2a (X ) OAc)
from 3 in hexane at 50 °C. (b) First-order fit to the growth.
) 0.061) of the triene 2a was determined by measuring
its absorbance (ꢀ ) 5.5 × 103 M-1 cm-1 at 320)1 after
generation by nanosecond LFP. The power of the laser
pulse was determined using the absorbance of the sol-
vated electron created from aqueous solutions of iodide
ion (ꢀ ) 1.33 × 104 M-1 cm-1 at 600 nm, Φ ) 0.29).24
Because only 60% of 2a proceeds to the methyl ether 4,
the quantum yield of formation of the ether by this
pathway is 0.036. To assess the importance of this
pathway in the overall photochemistry of 1a , we have
also measured the quantum yield (Φ ) 0.37) for its
disappearance in methanol. From the data in Table 1,
the yield of the methyl ether 5, is 60% and, therefore,
the quantum yield for ether appearance is 0.22. We
conclude that the yield for ether formation by the
pathway of excited-state conversion of 1a to 2a followed
by ground-state solvolysis of 2a accounts for only (0.036/
0.22) × 100 ) 16% of the total. Therefore, this pathway
does not significantly perturb the more normal excited-
state conversion of 1a to radical pairs and ion pairs.
P h ot olysis of 3,5-Dim et h oxyb en zyl Der iva t ives
(1b-f) in Meth a n ol. Results for the photolysis, again
at 254 nm in methanol at 25° C, are given in Table 3,
along with literature values where available.25,26
Nanosecond LFP of 1b in methanol gave the expected
transient at 320 nm that again did not decay on the
millisecond time scale. Because chloride is a considerably
better leaving group than acetate,20 we expected that the
(24) Siskos, M. G.; Zarkadis, A. K.; Steenken, S.; Karakostas, N. J .
Org. Chem. 1999, 64, 1925-1931.
(25) Appleton, D. C.; Brocklehurst, B.; McKenna, J .; McKenna, J .
M.; Thackeray, S.; Walley, A. R. J . Chem. Soc., Perkin Trans. 2 1980,
87-90.
(a ) 3,5-Dim eth oxyben zyl Ch lor id e, 1b (X ) Cl). As
shown in Table 3, the chloride is unusual for the high
yield of the dimer 7, of the arylmethyl radical. This
presumably reflects the formation of triplet radical pairs,
which will be longer lived and allow cage escape. A closer
examination, by GC/MS, indicated that there were actu-
(26) Appleton, D. C.; Bull, D. C.; Givens, R. S.; Lillis, V.; McKenna,
J .; McKenna, J . M.; Thackeray, S.; Walley, A. R. J . Chem. Soc., Perkin
Trans. 2 1980, 77-82.
(27) Langhals, H.; Fischer, H. Chem. Ber. 1978, 111, 543-553.
(28) La¨ufer, M.; Dreescamp, H. J . Magn. Reson. 1984, 60, 357-365.

1,3-Dimethoxy-5-methylene-1,3-cyclohexadiene Compounds
J . Org. Chem., Vol. 65, No. 15, 2000 4703
Ta ble 3. P r od u ct Yield s for th e P h otolysis of 3,5-Dim eth oxyben zyl Der iva tives 1b-f in Meth a n ol
X
ArCH3
ArCH2CH2OH
ArCH2)-2
ArCH2OCH3
R/Ia
1b, X ) Cl
3
14
6
48
5
34
66/34
14/86
6/94
6/94
77/23
1c, X ) Br
1 (trace)b
74 (95)b
90
1d , X ) I
1e, X ) O(PO)(OEt2)
1f, X ) N(CH3)3Cl
6
3
2
82
55 (46)c
7
7 (low)c
21 (12)c
a
b
Ratio of radical- to ion-derived product normalized to 100%. Value in brackets from ref 25. c Values in brackets from ref 26.
intermediate 2b (X ) Cl) would decay considerably faster
than the corresponding acetate 2a . To our surprise the
lifetime of the transient species was considerably longer,
with a half-life around 15 min, but the decay was unusual
in that there was considerable variation from experiment
to experiment. Our conclusion is that the main species
that we observed at 320 nm is the precursor to the dimer
10, i.e., 12, which has the same chromophore as 2b. We
also observed a band with λmax at 270 nm, consistent with
the expected precursor of the dimer 11, i.e., 13. For the
unsubstituted compound, 1-methylene-2,5-cyclohexadi-
ene, λmax ) 242 nm,30 and the added methoxy groups will
shift the band to longer wavelength by about 16 nm.31
This band also decays with a half-life of around 15 min.
The erratic decay of these transients is likely a result of
the formation of HCl as a byproduct in the photolysis of
1b (X ) Cl) because the conversion of 12 to 10 and 13 to
11, a tautomerization by hydrogen atom migration, will
be acid catalyzed. If any of the isomer 2c is formed in
the photolysis of 1c, its absorbance is probably hidden
under the large absorption band of 12.
(b) 3,5-Dim eth oxyben zyl Br om id e, 1c (X ) Br ) a n d
3,5-Dim eth oxyben zyl iod id e, 1d (X ) I). The higher
yields of the ion-derived product, the methyl ether 4,
seems surprising for these two compounds because the
heavy atoms, bromine and iodine, should induce inter-
system crossing to the triplet followed by homolytic
dissociation to triplet radical pairs. A similar high yield
of ether has been observed for 1-naphthylmethyl bromide
(88%,32 74%33) and iodide (>90%)32 for irradiation in
methanol. High yields of carbocation intermediates have
also been observed by LFP of diphenylmethyl bromide
(Φ ) 0.12) and chloride (Φ ) 0.13)34 in acetonitrile.
Moreover, sensitization experiments with acetophenone34
have shown that these cations are formed, along with
radicals, from the triplet state of a number of substituted
diphenylmethyl chlorides. This observation has been
explained by the suggestion that bond scission and ion
solvation are concerted in polar solvents so that triplet
state to ion pair conversion occurs by a solvent induced
intersystem crossing. Relevant to this question, very
recently the important observation has been made (LFP)
that the triplet state of benzoin diethyl phosphate 14,35
first forms the triplet state of the cation resulting from
heterolytic cleavage of the carbon-oxygen bond. As
expected these cations have very different reactivity from
their singlet state counterparts. This surprising process
may be specific to R-keto carbocations.
Nanosecond LFP experiments for 1c and 1d in metha-
nol gave no observable transient with λmax at 320 nm as
expected for the triene isomers 2c and 2d . For both
substrates, a broad transient from 330 to 450 nm with
λmax ) 380 nm was observed in both nitrogen and air sat-
urated samples. For the bromide 1c, the absorbance was
quite low. Similar transients have been observed in the
LFP study of 1-naphthylmethyl bromide and iodide32 and
were assigned to the dihalide radical ion, X2•-, resulting
from combination of halide anion with halogen atoms.
(c) Dieth yl (3,5-d im eth oxyben zyl) P h osp h a te, 1e
(X ) O(P O)(OCH2CH3)2). Again, nanosecond LFP in
methanol gave a transient at 320 nm which did not decay
on the long time scale of the instrument. The quantum
yield for its formation was 0.064. Because we were not
able to isolate this very reactive compound, we do not
know the yield of product(s) formed from its solvolysis
in methanol. As in the case of the acetate 2a , some
probably goes back to starting material 1e. The quantum
yield for disappearance of 1e in methanol was determined
to be 0.45 and because the phosphate gives a 94% yield
of the methyl ether, the quantum yield for ether forma-
tion is 0.42. Therefore, as in the case of the acetate
derivative 1a , the isomeric triene 2e, accounts for a
maximum of only about 16% of the excited-state chem-
istry. The very high yield of ion-derived products in
benzylic phosphate photochemistry is well documented.36
The transient 2e could also be generated by conven-
tional flash lamp photolysis in methanol at 25 °C and
its decay monitored in the usual alcohol solvents (k, s-1):
methanol (324 ( 10); ethanol (74.6 ( 2.8); 2-propanol
(22.7 ( 1.1); tert-butyl alcohol (5.11 ( 0.20). In methanol,
the rate enhancement for the better leaving group,
phosphate, relative to acetate, is ∼105. The mYOTs plot,
Figure 2, gives m ) 0.63, somewhat higher than for the
acetate 2a .
(d ) (3,5-Dim et h oxyb en zyl)t r im et h yla m m on iu m
Ch lor id e, 1f (X ) N(CH3)3Cl). As the results in Table
3 indicate, photolysis of the ammonium salt in methanol
is unusual in that a high yield of 3,5-dimethoxytoluene
is obtained. This is a consequence of the rapid dispro-
portionation possible in the (radical/radical ion) pair
generated from homolytic cleavage from S1, eq 3. This
observation has been made before for arylmethylammo-
nium salt photochemistry.26,37
(29) Pincock, J . A,; Wdege, P. J . J . Org. Chem. 1994, 59, 5587-5595.
(30) Plieninger, H.; Maier-Borst, W. Angew. Chem., Int. Ed. 1964,
3, 62.
(31) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric
Identification of Organic Compounds; J . Wiley: New York, 1991; p 298.
(32) Slocum, G. H.; Schuster, G. B. J . Org. Chem. 1984, 49, 2177-
2185.
(33) Arnold, B.; Donald, L.; J urgens, A.; Pincock, J . A. Can. J . Chem.
1985, 63, 3140-3146.
(34) Bartl, J .; Steenken, S.; Mayr, H.; McClelland, R. A. J . Am.
Chem. Soc. 1990, 112, 6918-6928.
(35) Rajesh, C. S.; Givens, R. S.; Wirz, J . J . Am Chem. Soc. 2000,
122, 611-618.
(36) Givens, R. S.; Kueper, L. W., III. Chem. Rev. 1993, 93, 55-66.
(37) Foster, B.; Gaillard, B.; Pincock, A. L.; Pincock. J . A.; Sehmbey,
C. Can. J . Chem. 1987, 65, 1599-1607.

4704 J . Org. Chem., Vol. 65, No. 15, 2000
DeCosta et al.
halides as solids. Both were chromatographed on silica gel with
5% ethyl acetate/hexanes and then recrystallized from hexanes
to give the pure compounds. The bromide, 1c(22%)25 and
iodide, 1d (30%)40 have been reported previously but without
spectral data (Table S1, Supporting Information).
Diethyl (3,5-dimethoxybenzyl) phosphate, 1e. To 2.52 g (15
mmol) of 3,5-dimethoxybenzyl alcohol and 2.01 g of pyridine
at 0 °C was added 2.59 g (16.5 mmol) of diethylchlorophosphate
(Aldrich) and the mixture was stirred for 6 h at 0 °C. Water
(25 mL) was then added and the mixture was extracted with
ether. The ether was washed with 0.1 M H2SO4, 5% NaHCO3,
and water, dried, and concentrated. The crude phosphate was
chromatographed on silica gel with 10% ethyl acetate/hexanes
to give the pure phosphate as a clear liquid (21% yield). 1H
(Figure S1) and 13C NMR (Figure S2) spectra are included in
the Supporting Information.
Nanosecond LFP in methanol again generated a tran-
sient at 320 nm assigned to the triene, 2f. Its decay could
be monitored by conventional flash photolysis (k ) 0.15
( 0.02 s-1), a value a factor of 39 faster than the acetate
2a , Figure 2. This rate ratio, determined for methanol,
will be very solvent dependent because the rate of
solvolysis for a substrate with a neutral leaving group
(triethylamine) will be only slightly affected by solvent
changes in contrast to the large effect for an anionic
leaving group (acetate). For reasons that we do not
understand, this transient was not observed in other
alcohol solvents.
Con clu sion s
(3,5-Dim eth oxyben zyl)tr im eth yla m m on iu m Ch lor id e,
1f. To 4.00 g (22.0 mmol) of 3,5-dimethoxybenzyl chloride, 1b,
in 100 mL of anhydrous ether was added, by slow bubbling,
gaseous trimethylamine (Matheson). The solution was then
left to stand for 16 h and the solid collected by vacuum
filtration to give 5.2 g (20.7 mmol) of crude salt. The salt was
washed with anhydrous ether and then dried under vacuum
at 100 °C and stored in a desiccator: mp 196-198 °C. This
compound has been prepared previously (mp 190-192 °C)41
but without 13C NMR spectral data (Table S1).
Syn th esis of 2- (10) a n d 4-(3,5-Dim eth oxyben zyl)-3,5-
d im eth oxytolu en e (11). Following the procedure of Miquel
et al.,42 1.21 g (7.2 mmol) of 3,5-dimethoxybenzyl chloride
(Aldrich), 2.00 g (13.1 mmol) of 3,5-dimethoxytoluene (Aldrich)
and 0.35 g of anhydrous ZnCl2 were stirred in 8.0 mL of
chloroform for 24 h. After adding 15 mL more chloroform, the
organic layer was washed with water, dried and concentrated
to give 2.5 g of crude material. GC/MS analysis indicated two
compounds of the appropriate m/z equal to 302. This material
was chromatographed on reversed-phase silica gel (EM Li-
Chroprep RP-18) using 80% methanol/20% water as eluent.
The dimers were removed from the eluate by adding 3× the
volume of water and extracting with 2× the volume of hexanes.
The hexanes were then dried (MgSO4) and concentrated to give
a mixture of 10:11 in a ratio of 65:35. The structures were
assigned by 1H NMR (CDCl3, 400 MHz) on the basis of the
higher symmetry for 11 (Supporting Information).
P h otolysis. In a typical reaction, 58 mg of 3,5-dimethoxy-
benzyl bromide, 1c, in 100 mL of HPLC grade methanol was
irradiated in a quartz tube in a Rayonet reactor with 254 nm
low-pressure mercury lamps while being purged with nitrogen.
The reaction was thermostated at 25 °C with an immersion
circulating water tube. The reactions were monitored by GC
on a 30 m by 0.25 mm J & W DB 200 column using helium as
carrier gas (split) and temperature programming (60 °C for 1
min, 15 °C/min to 180 °C, 180 °C for 7 min). Standard samples
of the products, except the dimers 10 and 11, (see above), were
available from previous work in our laboratory.29 Dark samples
were checked for ground-state reactions but none were ob-
served over the short times required for photolysis and
analysis.
La ser F la sh P h otolysis. The nanosecond laser flash
photolysis system at Dalhousie is of standard design using,
as excitation source, the fourth harmonic from a Continuum
Nd:Yag NY-61 laser (266 nm; e8 ns/pulse; e15 mJ /pulse). For
compounds 1a and 1b where the transient triene 2a and 2b
decays in the minutes time domain, samples of A = 0.5, in a
standard 1 cm cuvette, were subjected to repeated pulses at a
1 Hz repetition rate for about a minute and then transferred
to a thermostated HP-diode array spectrometer to measure
the decay kinetics. Light intensity for quantum yield measure-
ments was determined by KI actinometry as described previ-
ously.24
The most important conclusion reached form this study
is that the reactive trienes produced in the photochem-
istry of 3,5-dimethoxybenzyl derivatives are formed in
only low yield (∼16% for the acetate 2a , and the phos-
phate 2e) if at all (chloride, bromide and iodide). There-
fore, they do not play a significant role in the photochem-
istry of these substrates and their formation does not
complicate previous mechanistic arguments. However,
the solvolytic reactivity of these trienes is of interest.
They react by an SN1 mechanism via a transition state
that has less charge development than their benzylic
counterparts. Moreover, as expected, they have very large
rate enhancements: 2a (X ) OAc) reacts about 1013 times
faster than the benzylic acetate 1a . Finally, we obtained
an order of leaving group ability in methanol of acetate
< trimethylamine , phosphate (1:39:105). We have been
unable to find a literature value for trimethylamine as a
leaving group in SN1 reactions of benzylic compounds but
one is available for the reactions of tert-butyl substrates.
In that case, trimethylamine is more reactive than
acetate by a factor of about 65,38 again in methanol. The
rate ratio for phosphate to acetate in benzylic solvolysis
has been estimated20 at almost 109. Therefore, the order
we have obtained agrees with previous leaving group
studies, although the magnitude of the change in leaving
group ability is compressed, presumably because of the
very high reactivity of the compounds, 2.
Exp er im en ta l Section
Methanol (Caledon HPLC grade) was used without further
purification. Ethanol was dried by distillation from magnesium
ethoxide and 2-propanol and tert-butyl alcohol by distillation
from sodium.
Syn th esis of Su bstr a tes. The ester, 3,5-dimethoxybenzyl
acetate 1a , was prepared previously in our laboratory from
the alcohol.29 3,5-Dimethoxybenzyl chloride, 1b, was obtained
from Aldrich. Spectral data (1H and 13C NMR, GC/MS) are
given in Table S1 (Supporting Information) for the other
substrate, 1c-f.
3,5-Dimethoxybenzyl bromide, 1c, and 3,5-dimethoxybenzyl
iodide, 1d , were prepared in the same way using either NaBr
or NaI.39 A mixture of 3,5-dimethoxybenzyl alcohol (1.68 g, 10
mmol) and NaBr or NaI (15 mmol) and 15 mmol of freshly
distilled boron trifluoride etherate in 25 mL of acetonitrile was
stirred for 2 h at room temperature. Then, 10 mL of saturated
NaHCO3 was added and the mixture was extracted with
dichloromethane. The organic layer was washed (10% Na2S2O3
and sat. NaCl), dried and concentrated to give the crude
(40) Corey, E. J .; Floyd, D.; Lipshutz, B. H. J . Org. Chem. 1978, 43,
3418-3420.
(38) Thornton, E. R. Solvolysis Mechanisms; The Ronald Press
Company: New York, 1964; p 165.
(39) Mandal, A. K.; Mahajan, S. W. Tetrahedron Lett. 1985, 26,
3863-3866.
(41) Ridley, D. D.; Ritchie, E.; Taylor, W. C. Aust. J . Chem. 1970,
23, 147-183.
(42) Miquel, J . F.; Buu-Ho¨ı, N. P.; Royer, R. J . Chem. Soc. 1955,
3417-3420.

1,3-Dimethoxy-5-methylene-1,3-cyclohexadiene Compounds
J . Org. Chem., Vol. 65, No. 15, 2000 4705
Con ven tion a l F la sh P h otolysis. These experiments were
done in the laboratory of Professor A. J . Kresge in the
Department of Chemistry at the University of Toronto on an
apparatus previously described.16 Samples in 10 cm cylindrical
cuvettes were flashed perpendicularly to the long axis of the
cell by two Xenon Corp. FPA-5-100 C xenon flash lamps and
the absorbance monitored parallel to the same axis using an
Osram Model XBO 150-W xenon lamp. The transients gener-
ated from 1e and 1f were monitored on this instrument.
Qu a n tu m Yield Mea su r em en ts. Samples of 1a or 1e in
methanol were irradiated in a carousel apparatus at 25 °C
along with 4-methylbenzonitrile in acetonitrile. The quantum
yield for conversion of 4-methylbenzonitrile (Φ ) 0.025) to
3-methylbenzonitrile, used as an actinometer, can be measured
accurately at low conversions.43
Ack n ow led gm en t. We thank NSERC of Canada for
financial support. D.P.D. thanks the University of
Colombo for a sabbatical leave and a travel grant. N.H.
thanks NSERC of Canada for an Undergraduate Stu-
dent Research Award. We also thank F. Cozens and N.
Schepp of Dalhousie for help with the nanosecond LFP
results and Y. Kresge and J . Kresge for hospitality,
instructions, and experimental help with the microsec-
ond FP instrument in the Department of Chemistry,
University of Toronto.
Su p p or tin g In for m a tion Ava ila ble: Spectral data for
1
compounds 1c-f (Table S1), 10, and 11 and 250 MHz H and
13C NMR spectra for 1e. This material is available free of
charge via the Internet at http://pubs.acs.org.
(43) MacLeod, P. J .; Pincock, A. L.; Pincock, J . A.; Thompson, K. A.
J . Am. Chem. Soc. 1998, 120, 6443-6450.
J O000387O