Photochemistry of aryl tert-butyl ethers in methanol: The effect of substituents on an excited state cleavage reaction

Article abstract of DOI:10.1021/jo0002893

The photolysis of a set of 10 substituted aryl tert-butyl ethers, 8a-j, in methanol gave, as the major product, the corresponding phenol along with tert-butyl-substituted phenols resulting from photo-Fries reaction. The corresponding 4-cyanophenyl 1-adamantyl ether, 9, also gave 1-meth-oxyadamantane 16 (16%), indicating that, at least for this ether, some of the products were ion-derived. Quenching studies with 2,3-dimethylbutadiene for the tert-butyl ethers indicated that these reactions were occurring from the singlet excited state. Rate constants for the reaction, obtained from quantum yields and singlet lifetimes, were found to correlate reasonably well with σ^hv values, ρ = -0.77 (r = 0.975), a result that is unexpected for a reaction where the polarity of the bond breaking in the transition state is expected to be -O(δ-)···C(δ+).

Full text of DOI:10.1021/jo0002893

4162
J . Org. Chem. 2000, 65, 4162-4168
P h otoch em istr y of Ar yl ter t-Bu tyl Eth er s in Meth a n ol: Th e Effect
of Su bstitu en ts on a n Excited Sta te Clea va ge Rea ction
D. P. DeCosta,† A. Bennett, A. L. Pincock, J . A. Pincock,*,‡ and R. Stefanova
Departments of Chemistry, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4J 3, and
University of Colombo, Colombo, Sri Lanka
Received March 2, 2000
The photolysis of a set of 10 substituted aryl tert-butyl ethers, 8a -j, in methanol gave, as the
major product, the corresponding phenol along with tert-butyl-substituted phenols resulting from
photo-Fries reaction. The corresponding 4-cyanophenyl 1-adamantyl ether, 9, also gave 1-meth-
oxyadamantane 16 (16%), indicating that, at least for this ether, some of the products were ion-
derived. Quenching studies with 2,3-dimethylbutadiene for the tert-butyl ethers indicated that these
reactions were occurring from the singlet excited state. Rate constants for the reaction, obtained
from quantum yields and singlet lifetimes, were found to correlate reasonably well with σhν values,
F ) -0.77 (r ) 0.975), a result that is unexpected for a reaction where the polarity of the bond
breaking in the transition state is expected to be -O(δ-)‚‚‚C(δ+).
In tr od u ction
methoxy group is a better electron donor in the excited
state from the meta (and the ortho) position than it is
from the para position. Electron-withdrawing groups are
also predicted to be more effective electron-withdrawing
groups from the meta (and the ortho) position.12
To reliably correlate substitutent effects with excited
state rate constants, the ideal case would have an
unambiguous first (kr) or second (kr[Hx]) order rate
constant from a single excited state, S1, eq 1. The rate
The effect that substituents have on the excited state
reactivity of functional groups adjacent to benzene rings
continues to attract experimental and theoretical inter-
est. Recent examples, where quite a wide range of
electron-donating and -withdrawing substituents have
been examined include (1) the photolysis of benzylic
compounds with leaving groups (ArCH2-LG) where the
LG’s studied were trimethylamine (NR3),1 water (OH2),2
carboxylates (QO-CO-R),3 phosphates (QPO3R2),4 sulfides
(QSR),5 triphenylphosphine (PR3),6 and halides (QCl);7 (2)
the acid-catalyzed photohydration of styrenes8 and phen-
ylacetylenes;9 and (3) the excited state pKa of phenols.10,11
The interesting trend that has been observed in most (not
NR3 or QPO3R2) of these studies is that, contrary to
expectations based on ground state observations, meta
substitutents result in enhanced reactivity. This effect,
the photochemical meta effect, has been rationalized by
MO12-14 calculations which indicate that, for instance, a
constant can then be determined by the measurement
of the quantum efficiency of the reaction (Φr) and the
singlet lifetime, τs, eqs 2 and 3. However, few photore-
† University of Colombo.
‡ Phone: 902-494-3324. Fax: 902-494-1310. Email: pincock@
chem1.chem.dal.ca.
(1) Foster, B.; Gaillard, B.; Pincock, A. L.; Pincock, J . A.; Sehmby,
C. Can. J . Chem. 1987, 65, 1599-1607.
(2) (a) Turro, N. J .; Wan, P. J . Photochem. 1985, 28, 93-102. (b)
Wan, P. J . Org. Chem. 1985, 50, 2583-2586. (c) Wan, P.; Chak, B. J .
Chem. Soc., Perkin Trans. 2 1986, 1751-1756. (d) Wan, P.; Chak, B.;
Krogh, E. J . Photochem. Photobiol. A: Chem. 1989, 46, 49-57. (e) Hall,
B.; Wan, P. J . Photochem. Photobiol. A: Chem. 1991, 56, 35-42.
(3) Pincock, J . A. Acc. Chem. Res. 1997, 30, 43-49.
(4) Givens, R. S.; Kueper, L. W. Chem. Rev. 1993, 93, 55-66.
(5) Fleming, S. A.; J ensen, A. W. J . Org. Chem. 1996, 61, 7040-
7044.
(6) Imrie, C.; Modro, T. A.; Rohwer, E. R.; Wagener, C. C. P. J . Org.
Chem. 1993, 58, 5643-5649.
(7) (a) Cristol, S. J .; Bindel, T. H. J . Org. Chem. 1980, 45, 951-
957. (b) Cristol, S. J .; Bindel, T. H. J . Am. Chem. Soc. 1981, 103, 7287-
7293.
(8) (a) Wan, P.; Culshaw, C.; Yates, K. J . Am. Chem. Soc. 1982, 104,
2509-2515. (b) Wan, P.; Yates, K. Rev. Chem. Intermed. 1984, 5, 157-
181.
(9) McEwen, J .; Yates, K. J . Phys. Org. Chem. 1991, 4, 193-206.
(10) Wehry, E. L.; Rogers, L. B. J . Am. Chem. Soc. 1965, 87, 4234-
4238.
(11) Baldry, P. J . J . Chem. Soc., Perkin Trans. 2 1979, 951-953.
(12) (a) Zimmerman, H. E.; Sandel, V. R. J . Am. Chem. Soc. 1963,
85, 915-922. (b) Zimmerman, H. E.; Somasekhara, S. J . Am. Chem.
Soc. 1963, 85, 922-927.
actions satisfy these simple requirements. For instance,
in the examples (1) above, the photolysis of benzylic
compounds with leaving groups, both the excited singlet,
S1, and triplet, T1, states may react. Moreover, the excited
singlet may react by either homolytic (khom) or heterolytic
(khet) cleavage of the σ bond; both radical- and ion-derived
products are observed in most cases, eq 4. Because the
initially formed radical pair may be converted to the more
stable (in polar solvents) ion pair by electron transfer,
ketri in eq 4, the measured product ratio may not be a
reliable measure of khom vs khet. These concepts have been
recently reviewed.15
(13) Zimmerman, H. E. J . Am. Chem. Soc. 1995, 117, 9019-9025.
(14) Zimmerman, H. E. J . Phys. Chem. A 1998, 102, 5616-5621.
10.1021/jo0002893 CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/02/2000

Photochemistry of Aryl tert-Butyl Ethers in MeOH
J . Org. Chem., Vol. 65, No. 13, 2000 4163
Sch em e 1. P r od u cts Obta in ed fr om th e P h otolysis of th e Dia r yl Eth er s, 1, in Eth a n ol
We were intrigued by a recent observation of the
unusual substitution effects reported for the photochemi-
cal reactivity of diaryl ethers, 1 (Scheme 1), in ethanol.16
The major products, 2, 4 and 5, result from photo-Fries
type rearrangement with 2 derived from path a cleavage
and 4 and 5 from path b. Phenols 3 (path a) and 6 (path
b) were also formed. The ratio of (2 + 3):(4 + 5 + 6) )
path a:path b ) 42:0 (X ) CH3O), 69:0 (X ) OH), 64:0
(X ) NH2), 62:7 (X ) CH3), 25:3 (X ) F), 0:85 (X ) CO2-
CH3), 0:86 (X ) CN), varies considerably with substitu-
ents, electron-donating by their resonance effects (CH3O,
OH, NH2) having a marked preference for path a and, in
contrast, electron-withdrawing ones (CO2CH3, CN) se-
lecting path b. No explanation for this observation was
given. Moreover, a recent study,17 by MO calculations and
UV spectra, of the conformational surface for substituted
diaryl ethers gave results that supported the opposite of
this observation. For resonance electron-withdrawing
groups, such as CN, diaryl ethers adopt a conformation,
7 (Scheme 1), allowing intramolecular charge transfer
that shortens and presumably strengthens the C-O bond
for path b cleavage. In fact, path b cleavage is preferred
for electron-withdrawing groups.
We reasoned that substituent effects in aryl ether
photochemistry could be better studied using the set of
compounds, 8a -j, as well as the adamantyl derivative,
9, to compare with 8j. In contrast to the diaryl ethers 1,
these compounds have only one chromophore so there is
no question as to where the excited state energy will be
localized. As well, the bond cleavage should be highly
regioselective in the direction of the weaker tertiary
carbon-oxygen bond as opposed to the aryl carbon-
oxygen bond. Therefore, substituent effects on the ef-
ficiency and rate of the photochemical cleavage of a
simple single bond could be determined. Finally, the bond
cleavage would be either homolytic with a polarity such
that the substituents would interact with the oxygen
having a partial negative charge, -O(δ-), or heterolytic,
in which case the charge would be more fully developed.
Because this polarity is in the opposite direction to that
which has been studied in the past where the substituent
interaction was with a benzylic carbon and a developing
positive charge, -C(δ+), the manner in which the sub-
stitutents would influence the bond cleavage is obviously
of interest.
The photochemistry of arylalkyl ethers has been stud-
ied to some extent previously but usually with compounds
where the alkyl group was substituted in such a way as
to stabilize radical intermediates by delocalization. Alkyl
groups examined include triarylmethyl,13,18 diarylmeth-
yl,18 benzyl,19,20-22 and allyl.19,20,23,24 Again photo-Fries
compounds (photo-Claisen for the allyl compounds) and
phenols are the major products. In none of these cases
have substituent effects on the product distribution or
photochemical efficiency been examined.
Resu lts a n d Discu ssion
Syn th esis of Eth er s. The ethers 8a -i were synthe-
sized by the reaction of the Grignard reagent of the
(18) McClelland, R. A.; Kanagasabapathy, V. M.; Banait, N. S.;
Steenken, S. J . Am. Chem. Soc. 1989, 111, 3966-3972.
(19) Kelly, D. P.; Pinhey, J . T.; Rigby, R. D. G. Aust. J . Chem. 1969,
22, 977-991.
(20) Timpe, H.-J .; Friedrich, H.-J .; Dietrich, R.; Bo¨ckelmann, J .;
Friedel I. Z. Chem. 1978, 18, 25-26.
(21) Benn, R. Z. Naturforsch. 1979, 34b, 1002-1009.
(22) Pitchumani, K.; Devanathan, S.; Ramamurthy, V. J . Photochem.
Photobiol. A: Chem. 1992, 69, 201-208.
(15) Fleming, S. A.; Pincock, J . A. In Organic Molecular Photochem-
istry Vol. 3; Ramamurthy, V., Schanze, K. S.,, Eds.; New York, 1999;
pp 211-281.
(16) Haga, N.; Takayanagi, H. J . Org. Chem. 1996, 61, 735-745.
(17) Uno, B.; Iwamoto, T.; Okumura, N. J . Org. Chem. 1998, 63,
9794-9800.
(23) Pohlers, G.; Grimme, S.; Dreeskamp, H. J . Photochem. Photo-
biol. A: Chem. 1994, 79, 153-162.
(24) Pohlers, G.; Dreescamp, H.; Grimme, S. J . Photochem. Photo-
biol. A: Chem. 1996, 95, 41-49.

4164 J . Org. Chem., Vol. 65, No. 13, 2000
Ta ble 1. P h otop h ysica l P r op er ties of th e Eth er s 8a -j
DeCosta et al.
a
b
d
compd
λ
max, nm
ꢀ, M-1 cm-1
λ
max, nm
ꢀ, M-1 cm-1
λ0,0
,
nm
Es, kcal/mol
τs/10-9 c ns
Φf
8a , X ) H
271
280
272
278
272
269
263
260
1840
1730
1760
1910
1640
1200
947
220 (sh)
225
220
223
217
206 (sh)
211
224
211
245
7650
8890
7670
8660
6810
6930
7330
8190
6894
14400
14400
281
300
284
290
284
281
289
281
289
284
284
102
95
101
99
101
102
99
102
99
4.7
2.4
2.2
5.1
5.7
3.4
3.0
2.3
2.7
2.8
2.6
0.21
0.17
0.16
0.17
0.19
0.28
0.035
0.19
0.50
0.11
0.11
8b, X ) 4-OCH3
8c, X ) 3-OCH3
8d , X ) 4-CH3
8e, X ) 3-CH3
8f, X ) 4-F
8g, X ) 3-F
8h , X ) 4-CF3
8i, X ) 3-CF3
8j, X ) 4-CN
9, X ) Ad 4-CN
551
273
274 (sh)
274 (sh)
1163
1460
1220
101
101
244
a
b
Wavelength of the crossing point between the absorbance and emission spectra. Singlet excited state energy calculated from λ0,0
.
c Singlet lifetime by time-resolved fluorescence decay. Quantum yield of fluorescence.
d
corresponding bromobenzene with tert-butyl peroxyben-
zoate according to the method described previously, eq
5.25 Because we were also interested in studying cyano-
which should be somewhat lower than 64 kcal/mol, the
value for anisole.27 Therefore, homolytic cleavage from
S1 would be clearly exergonic for all substrates.
Although we have not obtained phosphorescence spec-
tra, triplet energies are available for some analogous
compounds: anisole (85 kcal/mol), 4-methoxyanisole (75
kcal/mol), and 4-methylanisole (78 kcal/mol).28 Therefore,
homolytic cleavage from T1 would also be exergonic.
Finally, on the basis of the reduction potential of the
phenoxy radical, 0.43 V vs NHE in DMSO,29 and the
oxidation potential of the tert-butyl radical, 0.33 V vs
NHE in CH3CN,30 conversion of the radical pair to the
ion pair by electron transfer, eq 7, should be only slightly
substituted substrates, which are incompatible with
Grignard reagents, the 4-cyano compounds 8j and 9 were
prepared by a different method. Reaction of the lithium
salt of the corresponding alcohol with 4-fluorobenzoni-
trile, eq 6, according to the method of Woiwode et al.,26
was successful. This nucleophilic aromatic substitution
only works if the fluorine is activated by the electron-
withdrawing group in the para position. No ether was
obtained when 3-fluorobenzonitrile was used, and there-
fore meta-cyano derivatives could not be prepared.
exergonic (2.3 kcal/mol). This estimate omits the Cou-
lombic attraction term which would contribute another
1.7 kcal/mol to stabilize the ion pair at a distance of 7 Å.
Moreover, for electron-donating groups (4-OCH3), the
reduction potential (0.134 V)29 of the phenoxy radical
becomes less favorable and the electron transfer becomes
endergonic by 0.2 V (4.6 kcal/mol). In contrast, for the
highly stabilized 4-cyano anion, the reduction potential
(0.90 V)29 makes the electron transfer favorable by 0.57
V (13 kcal/mol).
P h otoch em istr y of th e ter t-Bu tyl Eth er s 8a -j. The
ethers were irradiated in methanol at 25 °C using 254
nm lamps in a Rayonet reactor. The reactions were
31
P h otop h ysica l P r op er ties of Eth er s. UV absorption
spectra, quantum yields of fluorescence, singlet energies
(from the wavelength of the (0,0) band of the overlap
between the absorbance and fluorescence spectra), and
singlet state lifetimes (from time correlated single photon
counting) in methanol are given in Table 1. As expected,
there are two absorbance bands on the UV, the long
wavelength/low molar absorptivity, Lb, transition and the
shorter wavelength/higher molar absorptivity, 1La, band.
For the 4-cyano compounds 8j and 9, the conjugation
between the electron-donating oxygen of the ether and
electron-withdrawing cyano group shifts this 1La band
purged with N2 before and during photolysis. For all
the ethers, except 8j, this wavelength results in excitation
into an upper vibrational level of S1, the long wavelength
absorption band. For 8j, the excitation will be principally
into S2. For the tert-butyl ethers, 8a -j, the products, eq
8, obtained were simple, as the yield of the corresponding
phenol 10a -j, was always greater than 80% (GC). Minor
amounts of the photo-Fries products 11 were obtained
in all cases (GC/MS). These latter products were not
(27) McMillen, D. F.; Golden, D. M. Annu. Rev. Phys. Chem. 1982,
33, 493-532.
(28) Murov, S. L.; Carmichael, T.; Haig, G. L. Handbook of Photo-
chemistry, 2nd ed.; M. Dekker: New York, 1993; p 13.
(29) Bordwell, F. G.; Harrelson, J . A., J r. J . Org. Chem. 1989, 54,
4893-4898.
1
some 20 nm to longer wavelength so that the Lb band
appears only as a weak shoulder. The singlet energies
are all very similar (around 100 kcal/mol) with the
exception of the 4-methoxy compound 8b, which has a
somewhat lower energy at 95 kcal/mol. These values are
well above the carbon-oxygen bond dissociation energy
(30) Wayner, D. D. M.; McPhee, D. J .; Griller, D. J . Am. Chem. Soc.
1988, 110, 132-137.
(31) The more efficient removal of oxygen by freeze-pump-thaw
(FPT) cycles has been shown to be important in the photochemistry of
thioethers (ref 5), residual oxygen having an effect on quantum yields.
For the ethers studied here, the fact that the singlet lifetimes are short
(under 6 ns for all cases, Table 1) and also that the products are all
formed from an in-cage process, means that traces of oxygen will have
little effect. We have confirmed this in a few FPT experiments.
(25) Frisell, C.; Lawesson, S. O. Org. Synth. 1961, 41, 91-93.
(26) Woiwode, T. F.; Rose, C.; Wandless, T. J . J . Org. Chem. 1998,
63, 9594-9596.

Photochemistry of Aryl tert-Butyl Ethers in MeOH
J . Org. Chem., Vol. 65, No. 13, 2000 4165
Sch em e 2. Mech a n ism for th e P h otolysis of th e Eth er s, 8a -j fr om S1 in Meth a n ol
values equal to 33, 31, and 32 M-1, respectively. Com-
bined with the τs values in Table 1, the kq values for
quenching the excited singlet state of these ethers by the
diene are 7.0, 14, and 14 × 109 M-1 s-1, respectively, close
or equal to the diffusional value, 14 × 109 M-1 s-1, for
methanol.28 At 5 × 10-3 M diene, the ratio of quenching
of S1 (kQ[M] ) 7 × 109 M-1 s-1 × 5 × 10-3 M ) 3.5 × 107
s-1) to decay of S1 (1/τs ) 21 × 107 s-1) is 0.16. Therefore,
significant quenching of S1 occurs. In agreement with this
value, when the ether 8a (X ) H) was irradiated in the
presence of 5 × 10-3 M 2,3-dimethylbutadiene, products
(GC/MS) were observed that were photoadducts of the
two. The corresponding reaction of anisole with this diene
to give cycloadducts has been documented.34
characterized, except for 11a (using authentic samples
of 2- and 4-tert-butylphenol available from Aldrich), but
the GC peaks were counted. For instance, for 8a (X )
H), there were two such products (13%, ratio 1:1.4 of
2-tert-butylphenol:4-tert-butylphenol); for 8b (X ) 4-OCH3)
only one (4%) was observed and for 8c (X ) 3-OCH3)
three (18%, ratio 1:2:5) were observed. This is the number
expected for coupling ortho and para to the phenolic
oxygen as has been observed previously for benzyl phenyl
ethers.22 For the 4-methoxy compound 8b, we also
observed the formation of 8a (13%) as a result of
demethoxylation. These observations are consistent with
a mechanism of formation of radical pairs from the
excited state followed by disproportionation (path c) and
coupling (path d), Scheme 2. The intermediate 2,4-
cyclohexadienones shown in path d have recently been
observed by CIDNP,21 isolated32 in other photo-Fries type
reactions and also detected as transients in laser flash
photolysis experiments.33
Because both products in Scheme 1 could be derived
from in-cage radical pairs, the singlet excited state,
S1, would seem to be the likely precursor. However,
triplet yields for aromatic ethers are quite high (Φisc for
anisole ) 0.64),28 and, moreover, recent CIDNP experi-
ments with naphthyl ethers23,24 demonstrated that prod-
ucts are derived from both singlet and triplet radical
pairs. Earlier CIDNP results for phenyl ethers21 indicated
that only singlet radical pairs were involved. Additional
evidence that S1 was the reactive excited state for the
ethers studied here was provided by the following quench-
ing studies.
In contrast, irradiation at diene concentrations that
were a factor of 10 lower (5 × 10-4 M) gave only very low
yields (<1%) of cycloadducts. Therefore, as expected, S1
is essentially unquenched (kQ[M]/τs ) 0.016). Moreover,
simultaneous irradiation of 1a , in a carousel apparatus,
with (5 × 10-4 M) and without the quencher present,
resulted in an identical rate of disappearance of the
starting material. If triplet states were involved, they
would be expected to be longer lived and therefore still
quenched. Because the triplet lifetimes are not known
or easily determined because they are not detected by
triplet-triplet absorption in laser flash experiments, this
evidence for singlet reactivity is strong but not conclusive.
Irradiation of each of the ethers 8a -j in the presence
and absence of 2,3-dimethylbutadiene at 5 × 10-5 M gave
identical rates of disappearance in all cases. We conclude
that the reactions of these ethers occur from S1, the
excited singlet state.
The mechanism in Scheme 2 indicates radical pair
intermediates as in the accepted mechanism for photo-
Fries reactions. However, ion pairs cannot be easily
excluded because an intermediate ion-pair, 12 (Scheme
2), could give the same products. If the tert-butyl cation
were formed in methanol, it should perhaps give methyl
tert-butyl ether as a product, at least some of the time.
Because the yield of this volatile ether cannot be easily
determined by GC, we examined the adamantyl deriva-
tive 9, where the alkyl fragment is considerably larger.
Stern-Volmer fluorescence quenching plots for 8a
(X ) H), 8c (X ) 3-OCH3), and 8i (X ) 3-CF3) with 2,3-
dimethylbutadiene in methanol were linear and gave kQτs
(32) Yoshimi, Y.; Sugimoto, A.; Maeda, H.; Mizuno, K. Tetrahedron
Lett. 1998, 4683-4686.
(33) J ime´nez, M. C.; Miranda, M. A.; Scaiano, J . C.; Tormos, R.
Chem. Commun. 1997, 1487-1488.
(34) Gilbert, A.; Griffiths, O. J . Chem. Soc., Perkins Trans. 1 1993,
1379-1384.

4166 J . Org. Chem., Vol. 65, No. 13, 2000
DeCosta et al.
P h otoch em istr y of th e Ad a m a n tyl Eth er 9. The
adamantyl ether 9 provides a good comparison with tert-
butyl ether 8j for two main reasons: (1) the chemistry
of the alkyl fragment of the ether can be studied and (2)
the adamantyl radical, in contrast to the tert-butyl
radical, is known to be very unreactive toward dispro-
portionation35 by H atom donation as a consequence of
the low stability of 1-adamantene. Kropp and co-work-
ers36 have shown that the adamantyl group can be used
to study the partitioning between radical and ionic
behavior in the photochemistry of phenyl thio and seleno
ethers.
d4 had an (M/M + 1) ratio equal to 0.85, indicating that
it was less than 100% monodeuterated. However, the
adamantyl radical has a large isotope effect for H atom
abstraction with kH/kD values of 11 (toluene) and 26
(cyclohexane) at 65° C.35 Therefore, the adamantyl radical
must be abstracting hydrogen atoms from methanol-d4
in competition with some other nondeuterated source,
probably the adamantane group of the starting ether, 9.
The products 15 and 18 are best explained by photo-
chemical cleavage of the aromatic ether bond rather than
the adamantyl one. If, as seems likely, aryl-oxygen bond
cleavage is the pathway leading to these products, this
would be surprising because this bond will be consider-
ably stronger. Moreover, these products were not ob-
served for the tert-butyl ethers.
Photolysis of 9 in methanol as described previously
gave the products shown in eq 9. The yields given are
Finally, the methyl ether 16 is clearly derived from the
adamantyl cation and indicates that, at least to a small
extent, ionic intermediates are involved. In principle, the
ion pair could undergo an E1 like elimination as in 12
(Scheme 2) to form 1-adamantene which is known to
react very rapidly with methanol to give 16.37 This was
shown not to be the case, again by irradiation in
methanol-d4. The molecular ion of 16 had an m/z ) 169,
three more than the undeuterated material at m/z ) 166.
If 16 was formed from 1-adamantene, it would have
m/z ) 170, i.e., all four of the hydrogens from methanol-
d4 would be in the product.
based on reacted starting material (69% conversion), with
that for the dimer given as twice the molar yield. We see
no peaks (GC/MS) for photo-Fries products. The products
13, 14, and 17 are those expected from the radical pair
in methanol. The adamantyl radical has been shown to
be a very reactive H atom abstractor from studies on the
photochemistry of azoadamantane, 19.35 In fact, evidence
was presented that the dimer 14, formed from 19, was
These results clearly indicate that the 1-adamantyl
cation is an intermediate in the photochemistry of 9. We
have no evidence to indicate whether it is derived from
direct heterolytic cleavage of S1 or by electron transfer
in the initially formed radical pair. Also, we do not know
if the tert-butyl cation is an intermediate in the photo-
chemistry of the ethers 8, although it seems likely for
8j, the analogous 4-cyano compound to 9. In fact, as
discussed above, the ion pair for 8j and 9 will be more
stable than that for any of the other cases, but whether
this factor will make its yield higher or lower is difficult
to predict. Results for more adamantyl ethers would
clearly be of interest and we are in the process of trying
to synthesize some substituted derivatives.
only formed from in-cage recombination. Adamantyl
radicals that escaped the cage seemed to result only in
adamantane or aromatic addition products in solvents
such as benzene, toluene, cyclohexane, pentane, and tert-
butyl alcohol. For our results in methanol, the diada-
mantane must be formed from cage-escaped radicals so
there is competition between dimerization and H atom
abstraction. The same observation was made by Kropp
and co-workers36 for photolysis of adamantyl thio and
seleno ethers in methanol, although reliable yields were
not obtained because of the low solubility of the dimer
in methanol. We also noticed poor reproducibility for the
yield of the dimer but feel that the value given in eq 9 is
reliable because the substrate concentration was low
enough that solubility of the products should not be a
problem. The source of the H atom for formation of Ad-H
was probed by photolysis in methanol-d4. Authentic Ad-D
has an (M/M + 1) ratio in the mass spectrum of 0.281;35
the adamantane formed by photolysis of 9 in methanol-
Ra te Con sta n ts for th e P h otoch em ica l Clea va ge
of th e Eth er s 8a -j. Quantum yields (Table 2) for the
disappearance of the ethers in methanol at 25 °C were
measured in a carousel apparatus again using the 254
nm Rayonet reactor. These values, along with the singlet
lifetimes (Table 1), give the rate constants for photo-
chemical cleavage, kr, according to eq 2. Notice that for
the meta/para pairs the meta isomer is more reactive
than the para for all cases except for the strong, induc-
tively electron-withdrawing group, CF3. The photochemi-
cal meta effect seems to be operating but in a way which
suggests that electron donation from the meta position
is rate enhancing. This seems surprising for a reaction
that has an expected polarity with oxygen δ- and carbon
δ+. In fact, in a recent study of substituted benzyl phenyl
thioethers,5 the quantum yields reported for bond cleav-
age are in the ratio > 3.6 (4-OCH3:3-OCH3), 3.8 (4-CH3:
3-CH3), 0.11 (4-CN:3-CN), and 0.32 (4-NO2:3-NO2). There-
fore, electron-donating groups inhibit the reaction from
the meta position and electron-withdrawing groups en-
hance it from the meta position, again surprising for
(35) Engel, P. S.; Chae, W.-K.; Baughman, S. A.; Marschke, G. E.;
Lewis, E. S.; Timberlake, J . W.; Luedtke, A. E. J . Am. Chem. Soc. 1983,
105, 5030-5034.
(36) Kropp, P. J .; Fryxell, G. E.; Tubergen, M. W.; Hager, M. W.;
Harrs, G. D., J r.; McDermott, T. P.; Tornero-Velez, R. J . Am. Chem.
Soc. 1991, 113, 7300-7310.
(37) Gano, J . E.; Eizenberg, L. J . Am. Chem. Soc. 1973, 95, 972-
974.

Photochemistry of Aryl tert-Butyl Ethers in MeOH
J . Org. Chem., Vol. 65, No. 13, 2000 4167
Ta ble 2. Qu a n tu m Yield s of Rea ction a n d Der ived Ra te Con sta n ts for th e Rea ction s of Eth er s 8a -j
a
c
compd
Φr
kr/107 s-1 b
σex
σhν d
kf/107 s-1 e
kd/107 s-1 f
kdt/107 s-1 g
8a , X ) H
0.020
0.064
0.14
0.032
0.071
0.009
0.028
0.15
0.43
2.7
6.3
0.63
1.2
0.26
0.93
6.5
0
0
4.4
7.0
7.3
3.3
3.3
8.2
2.7
8.2
19
16
32
31
16
14
11
29
28
17
30
21
42
45
20
18
20
33
43
37
36
8b, X ) 4-OCH3
8c, X ) 3-OCH3
8d , X ) 4-CH3
8e, X ) 3-CH3
8f, X ) 4-F
-0.39
-0.10
-0.13
-0.10
0
-0.05
0.58
0.74
1.6
-1.17
-1.47
-0.53
-0.79
0.18
8g, X ) 3-F
-0.37
8h , X ) 4-CF3
8i, X ) 3-CF3
8j, X ) 4-CN
0.021
0.044
0.66
1.7
3.9
a
b
Quantum yield for disappearance of ether 8. Rate constant for disappearance of ether 8 from eq 2. The estimated error in these
values is (15% on the basis of estimated error ((10%) in the Φ values and (5% in the τs values. c From refs 10 and 11. From ref 9.
d
e Rate constant for fluorescence from kf ) Φfτs. f Rate constant for the decay of the excited singlet state by modes other than fluorescence
g
and reaction from kd ) kdt - kf - kr. kdt ) 1/τs.
F igu r e 2. Plot of log k versus σhν for the rates of reaction of
the aryl ethers, 8a -e (F ) -0.77).
F igu r e 1. Plot of log k versus σex for the rates of reaction of
the aryl ethers, 8a -j.
the rates of photochemical homolytic cleavage reactions.5
These values also give a very poor correlation with our
data, F ) -0.39 (r ) 0.24).
a reaction with the polarity of substituted benzylic carbon
δ+ and sulfur δ-.
We also wanted to correlate our rate constant data with
known scales of substituent effects for photochemical
reactions (Hammett plots). We thought that the data
would probably correlate with σex (Table 2) obtained from
the excited singlet state pKa values of substituted phe-
nols10,11 because the two reactions seemed likely to be
similar in electron demand, ARO-H (S1) f ArO(δ-)‚‚‚
H(δ+) and ArO-C(CH3)3 (S1) f ArO(δ-)‚‚‚C(CH3)(δ+).
In fact, the correlation was very poor (Figure 1) with F )
0.096 (r ) 0.12). A correlation with σ-, obtained from the
ground state pKa values for substituted phenols, was also
poor, F ) 0.27 (r ) 0.34). The data did however correlate
quite well with σhν values (Figure 2), F ) -0.77 (r )
0.975). These values are based on the excited state
protonation of arylalkenes and alkynes, a reaction that
generates δ+ on the benzylic carbon.9 Unfortunately that
photoprotonation reaction is too slow to be studied for
electron-withdrawing groups so that σhν values are
unavailable for CF3 and CN substituents. Finally, Flem-
ing and J ensen have recently proposed some σ‚hν values
on the basis of quantum yield ratios for the fragmentation
of substituted benzyl phenyl thioethers with the sugges-
tion that these values could be used to correlate data for
The negative F value obtained in the σhν plot confirms
that these reactions proceed in such a way that electron-
donating substituents enhance the reactions and meta
electron donation is preferred to para. In fact, the F value
of -0.77 is not much smaller than the value of -1.0 for
the defining reaction, the photoprotonation of S1 of
alkenes and alkynes, indicating that the response of the
reaction to substituents is comparable. Also noteworthy
is the fact that the compounds with strongly electron-
withdrawing groups in the para position are also quite
reactive, particularly 8h (X ) 4-CF3). In contrast, 8i
(X ) 3-CF3) is a factor of 10 less reactive than its para
analogue. We can assign σhν values to these substituents
by placing them on the correlation line in Figure 2,
σ
hν ) -1.62 (4-CF3), -0.42 (3-CF3), -0.87 (4-CN); but only
future studies with other substrates will confirm whether
they are reliable and useful values.
A possible reason for the unexpected correlation in
Figure 2 is that for these substrates the reactive center
is the aryl ether oxygen which will have its own effect
on electron distribution in the excited state, i.e., it is
electon donating itself. This situation is very different

4168 J . Org. Chem., Vol. 65, No. 13, 2000
DeCosta et al.
from the previous cases studied, where the substituent
was interacting with a benzylic carbon. How this oxygen
will interact with other substituents in the excited state
could be complicated and probably not obvious from
valence bond pictures. Perhaps high quality MO calcula-
tions will be of some help. The results may also be
affected by the intervention of reactive triplet states. The
rate constants for reaction (kr) and fluorescence (kf )
Φf/τs) account only for a relatively small fraction of the
total rate of decay (kdt ) 1/τs) of S1. These values are given
in Table 2. Quite likely, the major remaining path for
decay (kd) is intersystem crossing to T1. Our quenching
studies discussed above are not conclusive because we
do not know the triplet lifetimes. If both singlet and
triplet states are reacting to give products, then the
correlation with σhν values, defined from singlet state
reactions, will not likely be meaningful.
A referee has pointed out that determination of kr, as
in eq 2, may be an oversimplification because of internal
return of the radical pair by in-cage radical coupling to
the starting ether. This is a valid concern common to all
radical and ion pair chemistry. In fact, for photochemical
benzylic cleavage reactions, internal return has been
observed by 18O exchange in carboxylates, 20,38-40 and
phosphates, 21,41 and the magnitude is somewhat
in Figure 2 would be fortuitous. Results from other
substituted adamantyl ethers would help to clarify this
question.
Exp er im en ta l Section
P r ep a r a tion of th e Eth er s. The ethers 8a -i were pre-
pared by the reaction of tert-butyl peroxybenzoate (Aldrich)
with the Grignard reagent from the corresponding bromoben-
zene (Aldrich) using the procedure described by Frisell and
Lawesson.25 They were purified by vacuum distillation. All
have been reported previously but spectral data were not given
(Supporting Information). Photophysical data (UV absorbance
spectra, fluorescence quantum yields, excited singlet state
energies and excited singlet state lifetimes) are reported in
Table 1.
The ethers 8j and 9 were prepared by the reaction of the
lithium salt of the alcohols, tert-butyl alcohol and 1-adaman-
tanol (Aldrich), respectively, with 4-fluorobenzonitrile (Aldrich)
according to the method of Woiwode et al.26 Spectral data are
in the Supporting Information.
Ir r a d ia tion of Eth er s. The ethers were irradiated as
follows. A solution of 0.5-0.7 g in 100 mL of methanol was
purged with nitrogen and then irradiated in
a Rayonet
photochemical reactor using 16 lamps (75 W, 253.7 nm). The
progress of the reaction was monitored by GC, and the reaction
was stopped when the ether was >90% consumed. The
methanol was removed by rotary evaporation and the mixture
taken up in methylene chloride which was washed with water
and extracted into 5% NaOH. The basic layer was then
acidified and extracted with methylene chloride and analyzed
by GC/MS and NMR. The photoproduct phenols 10a -j and
11a (ortho and para) were characterized by comparison with
with authentic samples (Aldrich).
Quantitative photolyses were done in the same way as
described above except at lower concentrations, typically 50-
100 mg of the ether in 100 mL of methanol. Standard solutions
of the products 10a -j were prepared to determine their yields
in the photolysis reactions by comparing integrated areas by
GC/FID.
F lu or escen ce Mea su r em en ts. Fluorescence measure-
ments were carried out using a Shimadzu fluorescence spec-
trometer at 25 °C. Corrected spectra were obtained. Fluores-
cence quantum yields were determine by comparison with the
known fluorescence quantum yield of 0.13 for toluene42 in
methanol. Singlet lifetimes were measured using a PRA time
correlated single photon counting apparatus with a hydrogen
flash lamp of pulse width about 1 ns.
Qu a n t u m Yield Mea su r em en t s. The quantum yield of
reaction for the ethers in methanol were determined using
3-methoxybenzyl acetate (Φ ) 0.13) in aqueous dioxane as the
standard.12 The ethers, as well as the standard, were irradi-
ated in a thermostatted (25 °C), carousel apparatus using a
Rayonet reactor with four 75 W 254 nm lamps. Samples were
taken every 2 min for 14 min. Each sample was analyzed by
GC/FID in triplicate. A plot of percent conversion of the ether
or the ester divided by time versus time was extrapolated to
zero time. The ratio of the zero time values for the ether and
the standard multiplied by 0.13 provided the quantum yield
of reaction.
substituent dependent. In effect, kr(determined) )
kr(correct) × kp/(kp + ki), where kp/(kp + ki) gives the
fraction of the radical pair that proceeds to product (kp)
relative to return to starting ether (ki), i.e., kr(determined)
is less than kr(correct) by this factor. If this term is
constant for all substituents, then the rate constants
would be wrong by a constant factor but the correlation
in Figure 2 (which is logarithmic) would have the same
slope. If the effect is strongly substituent dependent, then
the correlation might be completely misleading. Experi-
mentally, for the radical pairs in the work reported here,
there is no convenient way to monitor internal return.
Fortunately, in the cases where internal return has been
measured, ki is always less than kr so that the error
created by internal return is always less than a factor of
2. Moreover, tert-butyl radicals do not couple very ef-
ficiently with other radicals, disproportionation normally
being preferred; for the ethers studied here, dispropor-
tionation (greater than 80%) of the radical pair to phenol
10 in eq 1 dominates coupling to photo-Fries products
11. Errors in kr of a factor of less than 2 will not
significantly effect the fundamental trend in Figure 2.
Finally, there is the possibility that both homolytic and
heterolytic cleavage are occurring from S1, particularly
in view of the formation of the methyl ether 16 from the
adamantyl substrate 9, eq 9. If kr is composed of two rate
constants (khet and khom, eq 4), and in a way which varies
from substitutent to substitutent, again the correlation
Ack n ow led gm en t. We thank NSERC of Canada for
financial support. D.P.D. thanks the University of
Colombo, Sri Lanka, for sabbatical leave and a travel
grant.
Su p p or tin g In for m a tion Ava ila ble: Spectral data for
compounds 8a -j and 9. This material is available free of
charge via the Internet at http://pubs.acs.org.
(38) Kim, J . M.; Pincock, J . A. Can. J . Chem. 1995, 73, 885-895.
(39) Givens, R. S.; Matuszewski, B.; Levi, N.; Leung, D. J . Am.
Chem. Soc. 1977, 99, 1896-1902.
(40) J aeger, D. A. J . Am. Chem. Soc. 1975, 97, 902-904.
(41) Givens, R. S.; Matuszewski, B.; Athey, P. S.; Stoner, R. M. J .
Am. Chem. Soc. 1990, 112, 6016-6021.
J O0002893
(42) Murov, S. L.; Carmichael, T.; Haig, G. L. Handbook of Photo-
chemistry, 2nd ed.; M. Dekker: New York, 1993; p 51.