Solvolysis and ring closure of quinone methides photogenerated from biaryl systems

1306

Solvolysis and ring closure of quinone methides

photogenerated from biaryl systems¹

Yijian Shi and Peter Wan

Abstract: A variety of biaryl quinone methides have been photogenerated with a range of efficiencies from biaryl pre-

cursors 4–6 and 8, 10, and 11, all having hydroxyl and hydroxymethyl substituents on alternate rings. These novel

biaryl quinone methides, which cannot be readily generated via thermal chemistry, are trapped by added nucleophiles

such as MeOH and ethanolamine; two that cannot undergo electrocyclic ring closure (from 8 and 11) are readily ob-

servable by nanosecond laser photolysis, with long wavelength maxima (λ_max) of 600 and 520 nm, respectively.

Photogenerated o,o′-biaryl quinone methides undergo electrocyclic ring closure to give the corresponding chromene

(pyran) products in high yield. Since the precursor biaryl alcohols have highly twisted structures in the ground state

(dihedral angle of up to 90° by molecular mechanics calculations), a significant twisting motion to planarity is required

to achieve reaction. Using steady-state fluorescence studies, we present evidence to suggest that the mechanism of

quinone methide formation may occur via one of the following mechanisms: (i) dissociation of the proton from ArOH

that precedes twisting; or (ii) ArOH dissociation and twisting taking place either simultaneously or in quick succession.

Key words: biaryl quinone methide, photosolvolysis, photodeprotonation, photocyclization.

Résumé : Un certain nombre de biaryl quinone méthides diverses a été photogénéré avec des efficacités variables à

partir des précurseurs biarylés 4–6, 8, 10 et 11 qui portent tous des substituants hydroxyle et hydroxyméthyle sur les

autres cycles. Ces nouveaux biaryl quinone méthides, qui ne peuvent pas être générés facilement par le biais de la

chimie thermique, peuvent être piégés par l’addition de nucléophiles tel que le méthanol et l’éthanolamine; deux

d’entre eux qui ne peuvent pas subir des fermetures électrocycliques de cycles (obtenus à partir de 8 et 11) peuvent fa-

cilement être observés par photolyse laser nanoseconde avec des longueurs d’onde maximales (λ_max) de 600 et 520 nm

respectivement. Les o,o′-biaryl quinone méthides photogénérées subissent des fermetures électrocycliques de cycle

conduisant à la formation des chromènes (pyran) correspondants, avec des rendements élevés. Puisque les structures de

chaque alcool précurseur biarylé sont de nature fortement déformée à l’état fondamental (angle dièdre allant jusqu’à

90° sur la base de calculs de mécanique moléculaire), la réaction ne peut se faire qu’à la suite d’un mouvement de dé-

formation important pour atteindre un état planaire. Faisant appel à des études de fluorescence en état stationnaire, on

a obtenu des données qui permettent de suggérer que le mécanisme de formation de la quinoneméthide peut impliquer

l’une des voies suivantes: (i) une dissociation de proton à partir du ArOH qui précède la déformation ou (ii) une disso-

ciation du ArOH et une déformation qui se produisent soit de façon simultanée soit en succession rapide.

Mots clés : biaryl quinone méthide, photosolvolyse, photodéprotonation, photocyclisation.

[Traduit par la Rédaction] Shi and Wan 1323

Introduction

strength joining the two aryl groups, hence, a twisting mo-

tion to a more planar structure. This phenomenon has been

utilized only on a limited basis to design or induce new

photochemistry: most of the photochemical reactions reported

for biphenyls do not appear to explicitly require a twisting

motion to planarity (3). Interest in the photocontrol of supra-

molecular assemblies (4) and our wish to explore the mecha-

nistic details and scope of quinone methide photogeneration

from biaryl systems prompted us to investigate in detail the

generality of photocyclization reactions reported by us over

several years (5). For example, twisted 2-(2′-hydroxy-

phenyl)benzyl alcohol (1) (dihedral angle of 68°) is effi-

ciently converted (in aq. CH₃CN) to the essentially planar

The moderate to highly twisted ground-state structures of

biaryls (as measured by the dihedral angle between the two

phenyl rings) are known, in general, to become more planar

when electronically excited. This twisting motion towards

planarity inherent in many biaryls on the excited-state sur-

face has been generally rationalized by Zimmerman and

Crumrine (1) and Imamura and Hoffmann (2), by noting that

the coefficients of the two carbons joining the two aryl rings

in the LUMO have the same sign, whereas they have oppo-

site signs in the HOMO. Thus, population of the LUMO on

electronic excitation causes an increase in the π-bond

Received 25 January 2005. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 19 October 2005.

Y. Shi and P. Wan.²Department of Chemistry, Box 3065, University of Victoria, Victoria, BC V8W 3V6, Canada.

¹This article is part of a Special Issue dedicated to organic reaction mechanisms.

²Corresponding author (e-mail: pwan@uvic.ca).

Can. J. Chem. 83: 1306–1323 (2005)

doi: 10.1139/V05-138

Shi and Wan

1307

Scheme 1.

Electrocyclic

h␯

2

3

2

syn

2

anti

5H-dibenzo[b,d]pyran (3) (dihedral angle of 24°) via the

proposed o,o′-biphenyl quinone methide syn-2 with high

quantum efficiency (Φ up to 0.5 at pH > 11) (Scheme 1) (5).

The anti conformer (anti-2) is presumably also formed but is

converted back to substrate 1 by nucleophilic attack by H₂O.

The incorporation of a naphthalene ring will, in general,

make such biaryls more twisted than biphenyls owing to

higher steric congestion induced by the naphthalene ring.

For this reason, we decided to examine the photochemistry

of naphthalene incorporated biaryls 4–6, to test whether the

photocyclization can occur for these more sterically con-

gested systems. Biphenyl derivatives 7–11 were also studied

to examine the photochemistry of those systems that cannot

undergo photocyclization, but that contain the appropriate

hydroxy and hydroxymethyl substituents to enable it to po-

tentially generate the appropriate biphenyl quinone methide

intermediate. We show that photocyclization takes place even

for the highly sterically congested binaphthyl 6 and that

biphenyl quinone methide intermediates are detectable by

nanosecond laser flash photolysis only when the cyclization

pathway is not available. In addition, photolysis of m-

hydroxy substituted biphenyl 10 gave rise to a formally non-

Kekulé biphenyl quinone methide.

Results and discussion

Materials

Biaryl alcohols 4–6 were readily made via LiAlH₄reduc-

tion of the corresponding lactones 12–14 (e.g., eq. [1]), all

known except for lactone 13, which was made by adapting

the procedure reported for 12 and 14 involving use of palla-

dium(II) acetate mediated intramolecular coupling of aryl

bromides (6). Biphenyl alcohols 8 and 10 were made by

BBr₃demethylation of methoxybiphenyls 7 and 9, respec-

tively, which in turn were readily made using palladium(0)

mediated Suzuki coupling (7, 8) of the appropriate aryl bro-

mide and arylboronic acid, followed by functional group

modification, as shown in Scheme 2 for the synthesis of 7,

which starts from 2-bromotoluene (15). Biphenyl alcohol 11

was made by LiAlH₄reduction of commercially available 4-

hydroxy-4′-biphenylcarboxylic acid.

4

[1]

Many o,o′-disubstituted biphenyls are found in the syn ge-

ometry (as shown by syn-18) as opposed to the anti geome-

try (as shown by anti-18) with respect to the two o-

substituents in the solid state (even in the absence of any

1308

Can. J. Chem. Vol. 83, 2005

Scheme 2.

3

2

Ether

3 4

3

2

intramolecular hydrogen bonding), although the general rea-

son for this remains unclear (9). We have shown that the

crystal structure of 1 and its α,α-dimethyl derivative also

take on the syn conformation, although for these alcohols

intramolecular H bonding may play a role in favouring such

a geometry (5a). With respect to the other biaryls studied in

this work, we anticipate that their diaryl dihedral angles to

be quite large, with also a preference for syn geometry. Mo-

lecular mechanics calculations (Chem 3D) of 4–11 confirm

this; indeed the dihedral angles found for 4–6 were close to

90°.²Although these biaryls are expected to have highly

twisted geometries, their solution structures may be differ-

ent. Roberts (9) has shown there is close agreement between

the dihedral angles of biphenyls observed in X-ray crystal-

lography and those deduced from ¹³C NMR studies. Thus,

there is strong reason to believe that all biaryls 4–11 have

highly twisted structures in solution. Moreover, if the prefer-

ence for syn geometry observed and (or) predicted in the

crystal for o,o′-disubstituted biaryls is also applicable in so-

lution, then these molecules are also structurally optimized

(in solution) to produce syn QMs (e.g., syn-2), which have

the correct geometry for electrocyclic ring closure.

conveniently followed by UV–vis spectrophotometry as all

the chromene products show a red-shifted strong absorption

band. Figure 1 shows the UV–vis spectral changes observed

on photolysis of 4 in 100% CH₃CN, which gave chromene

19 quantitatively after sufficient photolysis. Note the pres-

ence of vibrational fine structure in the absorption of the

photoproduct 19, which is consistent with its more rigid

structure (less torsional flexibility). The presence of

isosbestic points throughout the conversion indicates a very

clean reaction. The spectral traces for the analogous photo-

chemical conversion of 6 to 21 have been published previ-

ously (5c) and show the same qualitative features.

The dihedral angle for chromene 21 is a point of interest

since this would be the most sterically hindered of all the

chromene products. A dihedral angle of 30° was calculated

using molecular mechanics (Chem 3D). Thus, the photo-

chemical conversion of 6 to 21 requires a substantial twist-

ing motion, from about 90° to about 30°. Similar twisting

motions are required for the conversions of 4 and 5 to the

corresponding chromenes.

Photolysis (λ_ex= 254 nm) of 4 in H₂O–CH₃CN (1:1, v/v)

results in a photostationary state containing 56% of the

chromene product 19 and 44% of alcohol 4 (Fig. 2). Pro-

longed photolysis does not change this product to starting

material ratio. The photostationary state composition was

confirmed by exhaustive photolysis of a solution of 19 under

the same conditions, which regenerated 4 and gave a

photostationary state containing 60% of chromene 19 and

40% of alcohol 4. Photolysis of 5 (or 6) under identical con-

ditions resulted in a photostationary state containing >95%

chromene product 20 (or 21) and <5% alcohol starting mate-

rial.

Photolysis of 4 in neat CH₃OH gave a photostationary

state containing approximately 24% of chromene 19 and

76% of the methyl ether product 22. Prolonged photolysis,

however, gave rise to a small amount (~10%) of unidentified

side products, owing to the secondary photodecomposition

of methyl ether 22. The reversibility of this reaction was

demonstrated by photolysis of chromene 19 in neat CH₃OH,

which gave a photostationary state containing 21% of 19 and

79% of 22. In H₂O–CH₃OH (1:1), photolysis of 19 resulted

in a photostationary state containing 19, 22, and 4 in a ratio

of 8:69:23. When the concentrations of CH₃OH and H₂O are

taken into account, the relative rates of CH₃OH and H₂O ad-

syn

anti

Product studies

Photolysis (254 or 300 nm) of 4–6 in 100% CH₃CN

cleanly gave the corresponding ring closure chromene

(pyran) products, as shown in Scheme 3 for 4 (correspond-

ingly, 5 gave 20 and 6 gave 21, Table 1), although with dif-

ferent quantum yields. These ring closure reactions can be

²Unpublished X-ray analyses (A. MacKinnon, J.A.K. Howard, Y. Shi, and P. Wan) of some single crystals of these compounds confirm the

expected large dihedral angles in the solid state. In addition, syn geometry was observed in all of 4–6.

Shi and Wan

1309

Scheme 3.

3

h␯

2

3

h␯

3

h␯

4

3

Table 1. Quantum yields for photocyclization of biaryl alcohols 1 and 4–6.

Φ^c(H₂O–CH₃CN)

Biaryl

alcohol

PS ratio^bin

H₂O–CH₃CN (1:1)

0%

1%

10%

H₂O

50%

H₂O

Φ^c

(CH₃OH)

Φ^c

(THF)

λ_max(nm) (⑀)^a

Product

H₂O

60:40

>95:5

≈98:2

4

5

19

20

21

3

337 (10 600)

350 (7 950)

350 (11 500)

308 (6 775)

0.29

0.079

0.11

0.12

0.30

0.090

0.21

0.10

0.25

0.080

0.34

0.21

0.12

0.047

0.18

0.21

0.081

0.039

0.16

0.39

0.11

0.15

0.11

6

1^d

^aAbsorption max and extinction coefficient (number in brackets) of the product.

^bProduct to biaryl alcohol ratio at the photostationary (PS) state (λ_ex= 254 nm).

^cProduct quantum yield (for photocyclization) measured using potassium ferrioxalate actinometry (error: 10%) (10). Water content quoted is by volume

in CH₃CN.

^dUsed for comparison. Some data taken from ref. 5a.

dition is approximately 7:1 in favor of CH₃OH. Photolysis

of 5 and 6 in the presence of CH₃OH also resulted in the for-

mation of the corresponding methyl ether products accompa-

nied with a reduced yield for the corresponding chromene

product. In the presence of NaBH₄(~0.5 mol/L), photolysis

of 4 (in 2-propanol or aq. CH₃CN) gave the reduction prod-

uct 23 and 19 (Scheme 3). Photolysis of 1 and 6 also gave

similar reduction products.

Photolysis of the biphenyl alcohols 7 and 8 in H₂O–

CH₃OH (1:1) gave the corresponding methyl ether substitu-

tion products 24 and 25 (Scheme 4). In the presence of ethyl

vinyl ether (~0.4 mol/L), photolysis (254 nm, 15 min) of 8

in aq. CH₃CN (H₂O–CH₃CN, 1:2) gave aldehyde 26 in high

yield (>60%). In this reaction, ethyl vinyl ether simply func-

tions as a nucleophile (β-vinyl carbon of ethyl vinyl ether)

followed by hydrolysis. No evidence for [4+2] cycloaddition

products was observed. Although less efficient, photolysis of

11 under similar conditions also resulted in the formation of

the corresponding methyl ether product from CH₃OH, but no

aldehyde product was observable in the presence of ethyl vi-

nyl ether.

It was previously observed (5a) that when the phenolic

OH group of the parent compound 1 was replaced with

a methoxy group, photolysis of the substrate in aqueous

methanol does not result in the corresponding methyl

ether substitution product. This is also true for 3-[2′-

1310

Can. J. Chem. Vol. 83, 2005

Fig. 1. UV–vis spectral traces showing the smooth conversion of

4 to chromene 19 in 100% CH₃CN (λ_ex= 254 nm; each trace

represents about 10–20 s photolysis, except for the last trace,

which was obtained after extended photolysis, >10 min). The fi-

nal spectrum obtained is essentially identical to that of pure

chromene 19.

Scheme 4.

3

h␯

2

3

h␯

mol/L

2

3

Interestingly, irradiation of 10 in CH₃CN–H₂O (1:1) re-

sulted in a major product 3-hydroxyfluorene (27) and a mi-

nor product in approximately a 10:1 ratio. The minor

1

product had the same molecular weight as 27 and its H

NMR spectrum showed only absorption at the aromatic re-

gion and a singlet at δ 3.82 ppm, corresponding to Ar-CH₂-

Ar. This minor product is consistent with 1-hydroxyfluorene

(28) (eq. [2]). In CH₃OH–H₂O (1:1), photolysis of 10 gave

Fig. 2. Plot of % 4 (triangles) and % 19 (squares) vs. photolysis

time in 100% CH₃CN (filled) and in CH₃CN–H₂O (1:1) (un-

filled).

1

an additional minor product. The H NMR spectrum of the

100

80

60

40

20

0

product mixture is consistent with the formation of the sub-

stitution product methyl ether 29 (eq. [2]). The ring closure

reaction (to give 27 and 28) can also be monitored by UV

spectrophotometry. Since the absorption spectrum of 27 is

red shifted by about 20 nm relative to the starting material

(10), as the reaction proceeds, a structured new band corre-

sponding to 27 in the 300–320 nm region grows in with

photolysis.

Product quantum yields

Product quantum yields for the photocyclization of biaryl

alcohols 4–6 and for the photosovolysis/photocyclization of

7–11 were measured using potassium ferrioxalate actino-

metry (10) (λ_ex= 280 nm) as the standard, and UV–vis and

1

(or) H NMR to monitor for reaction progress. Photocycli-

0

10

20

30

40

50

zation product quantum yields for 4–6 are summarized in

Table 1. Large solvent dependence was observed in the

photocyclization quantum yields. Although H₂O was previ-

ously observed to enhance the photocyclization of the parent

compound 1 (5a), the water effect on the photocyclization of

4 was the opposite. Added H₂O (when the H₂O content

≥10%) noticeably decreased the quantum yield. In contrast

Photolysis Time (min)

(hydroxymethyl)phenyl]anisole (9), where prolonged irradia-

tion of 9 in CH₃OH–H₂O (1:1) did not give any product, and

the starting material was completely recovered.

h␯

[2]

3

2

3

Shi and Wan

1311

Table 2. Fluorescence data of biaryl alcohols and model compounds.

k_f⁰(s^–1)^c

∆ν (w) (cm^–1)^d

Compound

τ_f^a(ns)

Φ_f^b

(CH₃CN)

CH₃CN

0.01 mol/L NaOH

Biphenyls

27

11

33

7

8

34

9

10

30

1

1705(1411)

2979(1574)

3216(1754)

4235(2063)

4968(2096)

5838(2506)

0.24(0.30)

0.16(0.19)

0.15(0.16)

0.13(.026)

0.12(.087)^e

0.12(0.12)

0.14(0.04)

3239(1921)

3368(2040)

6238(2402)

5590(2138)

3623(2017)

3003(1801)

3276(1913)

Naphthols and naphthylphenols

35

1676(1319)

1834(1525)

1689(1899)

1786(1430)

1737(1361)

1715(1450)

1680(1354)

3034(1583)

2958(1528)

5

4

5.8(10.8)

0.23

0.0059

0.15

0.26

0.050

0.18

4.0×10⁷

6

6.7

2.2×10⁷

3.8×10⁷

1.4×10⁷

2.6×10⁷

31

23

32

6.8(15.0)

3.6

6.9(3.0)

2975(1561)

~7800

6865(2264)

^aMeasured in neat CH₃CN. The lifetime of the corresponding phenolate or naphtholate (in 0.01 mol/L NaOH)

quoted in brackets.

^bDetermined in neat CH₃CN using naphthalene and 2-aminopyridine as the standard. The number in brackets was

determined in H₂O–CH₃CN (1:1).

^cNatural fluorescence rate constant as defined by k_f⁰= Φ_f/τ_f.

^d∆ν is the Stokes shift and w is the spectral width (at half height), as defined by Berlman (11, 13).

^eDue to stronger emission from the phenolate ion, Φ_fin aqueous solution was calculated from the Φ_fof 34 in neat

CH₃CN and the relative fluorescence intensity in the two solvents.

to 4, the quantum yield for 6 and 5 showed an initial in-

crease with added water and a decrease at higher water con-

tents. The quantum yield for 5 was also much smaller than

those of 4 and 6. Biaryl alcohol 6 is expected to be most

sterically congested system, whereas 4 and 5 are expected to

be similar in this respect; therefore, the explanation for the

observed quantum yield difference probably lies in an elec-

tronic factor that we do not currently understand.

Fluorescence measurements

It is well-known that biphenyl and many of its derivatives

have large geometry changes upon electronic excitation (11).

When compared with more rigidly planar (in both ground

and electronically excited states) molecules such as fluorene

and anthracene, these molecules have larger energy differ-

ences (thus, a larger Stokes shift) between the Frank–Con-

don and the fluorescent state. For example, biphenyl and

phenyl naphthalenes have significantly larger Stokes shifts

(3000–3200 cm^–1) than fluorene (1430 cm^–1) and 1,2-

benzofluorene (1400 cm^–1) (in cyclohexane) (11). Thus, ex-

cited-state twisting of biaryls can be qualitatively related to

the Stokes shift values determined from fluorescence emis-

sion spectra (11, 12). In addition, transitions from a more

planar initial state (S₀or S₁) tend to give narrower as op-

posed to broad spectra (12). Thus, both Stokes shift (∆ν) and

spectral width (w) of the fluorescence emission band provide

useful information with respect to geometry changes associ-

ated with the excited state.

Due to highly efficient photocyclization via S₁, some of

the biaryl alcohols studied were only weakly fluorescent.

Therefore, the fluorescence behavior of a series of model

compounds (7, 9, 23, 27, and 30–35) that are either non-

reactive or reacted only with low quantum efficiency were

also studied. The fluorescence spectral properties (∆ν, w),

fluorescence quantum yields, and the fluorescence lifetimes

for some of these compounds are also tabulated in Table 2.

As shown in Table 2, the Stokes shifts (∆ν) for all the

“loose” biphenyl systems in neat CH₃CN are approximately

3300 300 cm^–1, which is significantly larger than that of the

The quantum yield for the photocyclization (to the

fluorene ring system) of 10 in CH₃CN–H₂O (1:1) (eq. [2])

was 0.17 (for the formation of both 27 and 28). There was

significant variations in measured quantum yields for the

photosolvolysis of the remaining biphenyl alcohols 7, 8, and

11. Photolysis of 8 in H₂O–CH₃OH (1:1) gave 25 with Φ =

0.23, while photolysis of 11 under identical conditions gave

the corresponding methyl ether in much lower quantum

yield (Φ = 0.026). Substitution of the phenol OH with an

OCH₃group significantly lowers the reactivity of these sub-

strates. For example, the quantum yield (Φ = 0.075) for

photolysis of 7 to give 24 is only one-third of that for

photolysis of 8 to give 25. Indeed, there was no methyl ether

product (or ring closure product) observable in the

photolysis of 9. In general, the product quantum yields mea-

sured previously were unaffected in the pH 1–7 region

(slight increases of up to about 10% was observed on lower

acidity beyond pH 1). In the pH > 9 region, the compounds

were more reactive as indicated by enhanced loss of the sub-

strate, but quantification was problematic because of the for-

mation of more side products as a result of direct excitation

of the phenolate.

1312

Can. J. Chem. Vol. 83, 2005

Fig. 3. Transient spectra of biaryl quinone methides 36 and 37

obtained from nanosecond LFP of 8 and 11, respectively, in

100% H₂O.

3

0.08

BQM 36

BQM 37

0.06

0.04

0.02

0.00

3

“rigid” biphenyl molecule 27 (1705 cm^–1). The spectral width

(w) of all the o-substituted biphenyl derivatives (w ≈ 1750–

2000 cm^–1) is significantly larger than analogs not bearing a

substituent at the o-position (11, w = 1574 cm^–1) and the

rigid compound 27 (w = 1411 cm^–1). The spectral width of

11 is only slightly larger than that of 27. This phenomenon

is a result of the increased steric repulsion owing to o-substi-

tution that results in a less rigid and less planar S₁in these

molecules. A similar trend was also observed in 0.01 N

NaOH(aq) solutions, where the emitting species are the cor-

responding phenolate ions. For example, the Stokes shift and

spectral width of 34 (∆ν = 6238 cm^–1, w = 2402 cm^–1, model

of 10) and 30 (∆ν = 5590 cm^–1, w = 2138 cm^–1, model of 1)

are larger than those of the rigid molecule 27 (∆ν =

4235 cm^–1, w = 2063 cm^–1). Thus, it can be concluded that

both the neutral and the phenolate ion forms of these

biphenyl systems can twist to the more planar geometry

upon electronic excitation.

In contrast to the biphenyl systems, biaryls with one or

two naphthalene rings (4–6 and their model compounds, bot-

tom half of Table 2) do not show large Stokes shifts in neat

CH₃CN. The fluorescence spectra of these molecules in neat

CH₃CN resemble that of 2-naphthol (35) or 2-

(hydroxymethyl)naphthalene. The Stokes shifts (~1750

80 cm^–1, Table 2) of these compounds are also similar to (or

only slightly greater than) that of 35, but significantly

smaller than that of 1-phenylnaphthalene (3200 cm^–1) (11),

which suggests that these molecules are just as twisted in the

S₁state as in the ground state. The spectral width for these

molecules, except for 4, is also small (~1400–1500 cm^–1) in-

dicating rigid S₁states.

In aqueous solution, neither 4 nor 6 showed any detect-

able fluorescence emission owing to the very efficient photo-

chemical cyclization from S₁. Thus, model compounds 23

and 32 provided valuable information. In CH₃CN–H₂O

(1:1), both of these compounds only gave one markedly red-

shifted emission band centered at ~560 nm. These emission

bands are identical to the emission spectra of the corre-

sponding phenolate and (or) naphtholate ions obtained by

using 0.01 mol/L NaOH–CH₃CN (1:1) as the solvent. The

absence of the emission from the neutral species of 23 and

32 at pH 7 suggests rapid excited-state proton transfer

(ESPT) to the solvent. In 0.01 mol/L NaOH–CH₃CN (1:1),

the Stokes shift and spectral width (∆ν = 6865 cm^–1, w =

2264 cm^–1) of 32 and that of 23 (∆ν ≈ 7800 cm^–1, this emis-

sion is too weak for w to be measured accurately) are much

320

420

520

620

720

Wavelength (nm)

larger than the naphtholate ion of 35 (∆ν = 3034 cm^–1, w =

1583 cm^–1, Table 2). These extraordinary large Stokes shifts

are consistent with large geometry changes in S₁for the

deprotonated forms of 32 and 23. Thus, it seems reasonable

that the deprotonated forms of 4 and 6 also twist to the pla-

nar (or more planar) geometry in the excited state. The lack

of any observable emission from these compounds is due to

an inherently efficient cyclization reaction that takes places

from this more planar state.

In contrast, 5 and its model compound 31 show strong

emissions (at nearly equal intensity) from both the neutral 2-

naphthol moiety (λ_max≈ 360 nm) and the naphtholate ion

(λ_max≈ 425 nm) in H₂O. In 0.01 mol/L NaOH–CH₃CN

(1:1), only the emission of the naphtholate ion was observ-

able, but the spectral properties (Stokes shift, spectral width,

and λ_max) of this emission closely resemble those of 2-

naphthol 35 (the 2-naphtholate ion). It is concluded that 5,

even in its deprotonated form, can not twist to the planar ge-

ometry efficiently in S₁, which is required for the ring clo-

sure reaction. This is consistent with its low quantum

efficiency observed for the ring closure reaction (Table 1).

Laser flash photolysis

Nanosecond laser flash photolysis (LFP) (266 nm,

Nd:YAG laser, ≤20 mJ) of 8 in aq. CH₃CN gave a strong and

very broad transient absorption spectrum with the major

band at ~600 nm (Fig. 3). Although the transient spectrum

was broad (450–800 nm), the decay at all wavelengths (in-

cluding the weak band at 350 nm) was single exponential

and had identical rates. It was also observed that the water

content of the solvent affected the intensity as well as the

decay rate of this transient. In neat CH₃CN, this transient

was not observed; in neat H₂O, the intensity of the spectrum

was ca. threefold larger than that in CH₃CN–H₂O (1:1). This

indicated that H₂O was necessary for the formation of this

transient. The λ_maxof the transient was red shifted with in-

creasing solvent polarity (575 nm in CH₃CN–H₂O (1:1) to

600 nm in neat H₂O). The shape of the spectrum was inde-

pendent of pH; photolysis of 8 in pH 2.5, 7, and 12 gave es-

sentially the same spectrum, although the intensity varied.

The decay rate was not affected by O₂but showed strong

Shi and Wan

1313

Fig. 4. pH Dependence for photogeneration of biaryl quinone

methide 36 as observed by LFP of 8 (errors are 10% of quoted

value).

sponding to the pK_aof the ground state). This is in fact re-

vealed in Fig. 4. The first titration between pH 2–4 is clearly

evident while the second at pH 10–12 is less marked. The

modest increase in efficiency upon excitation of the pheno-

late might be due to the already high yield of the phenolate

ion on adiabatic deprotonation from the phenol.

Although product studies showed that the photolysis of 7

in H₂O–CH₃OH (1:1) also gave the corresponding methyl

ether product in reasonable yield, LFP of 7 did not produce

any observable transient species assignable to quinone methides

or carbocations. This seems reasonable since there is no sim-

ple mechanism for photogenerating quinone methides from

the methoxy compound. Moreover, the anticipated carbo-

cation intermediate responsible for the observed photosol-

volysis would be of the benzylic kind and would be too

short-lived in aqueous solution to be detectable by nanosec-

ond LFP.

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0

Laser flash photolysis of 11 in aq. CH₃CN also gave a

transient absorption spectrum with two major bands at 360

and 525 nm (Fig. 3). Its decay rate in 100% (pH 7) H₂O

(k_obs= 1.50 × 10⁴s^–1) is less than that observed for 36. The

transient is quenched by H⁺and added nucleophiles and thus

can be assigned to BQM 37. A Stern–Volmer plot using

H₂NCH₂CH₂OH as a quencher gave k_q= 2.61 × 10⁵s^–1

(mol/L)^–1, which is significantly smaller than that observed

for 36. The fact that the spectral width of 37 is narrower

than that of 36 is consistent with its more rigid ground state

(less steric repulsion at the o-position). Unlike 36, the

phenolate ion of 11 gives rise to BQM 37 much more effi-

ciently than the neutral molecule; the relative quantum yield

for the photogeneration of 37 is 3.5 times higher at pH 12

than at pH 7.

Burnham and Schuster (14) have studied the photochem-

istry of 6 and 21. They suggested that the photocyclization

of 6 to 21 proceeded via a BQM intermediate (viz., 38).

They also suggested that BQM 38 was observable by nano-

second laser flash photolysis in cyclohexane with a lifetime

of 5 µs. However, using nanosecond LFP, we have not been

able to detect any transient species in the 450–700 nm re-

gion assignable to the BQM for any of 1, 4–6, and 10 in aq.

CH₃CN. We conclude that in aq. CH₃CN any BQM interme-

diate from these substrates is too short-lived to be detectable

on the nanosecond timescale, i.e., electrocyclic ring closure

from intermediates such as 39, to give two new aromatic

rings in 19 (Scheme 5) would be expected to be very fast

(k >> 10⁹s^–1).

0

2

4

6

8

10

12

pH

solvent and pH dependence. In the CH₃CN–H₂O mixed sol-

vent system, the decay rate increased by ca. three orders of

magnitude when the water content changed from 10%–100%

H₂O. A plot of the logarithm of the decay rate vs. the

molarity of H₂O was essentially linear. The transient could

be efficiently quenched by H⁺. In pH 4–12, there was essen-

tially no pH effect on the decay rate constant (k_obs), although

it was slightly affected by added electrolytes (in neat H₂O

and 0.01 mol/L NaOH, k_obs= 2.0 × 10⁶s^–1; in pH 4 and 7

buffered solutions, k_obs= 2.5 × 10⁶s^–1). However, the decay

rate increased dramatically at lower pH (pH 2 to 3), suggest-

ing quenching by H⁺. In this pH region, the observable tran-

sient intensity also decreased significantly with pH while the

fluorescence intensity increased considerably. Due to inter-

ference from fluorescence emission and weaker transient ab-

sorption, it was not possible to measure the decay rate

accurately at low pH. It was estimated that k_obs~10⁷s^–1at

pH 3 and ~10⁸s^–1at pH 2, corresponding to a quenching

rate constant (by H⁺) of k_q≈ 10¹⁰s^–1(mol/L)^–1. The tran-

sient was also quenchable by added nucleophiles such as

ethanolamine (k_q= 3.6 × 10⁶s^–1(mol/L)^–1).

We thus assigned this transient spectrum to biaryl quinone

methide (BQM) 36. The broad and structureless absorption

spectrum is consistent with a ground state that is not rigidly

planar. The fact that irradiation of 8 in 0.01 mol/L NaOH

gave exactly the same transient spectrum as in pH 7, indi-

cates that this same transient can be generated by irradiation

of the corresponding phenolate ion of 8. The relative quan-

tum efficiency for generating this transient from the neutral

form of 8 vs. its deprotonated form (the phenolate ion) is

shown in Fig. 4 in which the ∆A at 600 nm at different pH

values was measured immediately after each laser pulse, and

the ratio ∆A/∆A_pH7was plotted against the pH of the sol-

vent. When the energy of the laser pulse and the absorbance

of the solution at the excitation wavelength were kept con-

stant, the ratio ∆A/∆A_pH7is a direct measure of the relative

quantum yield for formation of BQM 36 at this pH (relative

to pH 7). It is expected that such a plot would show two

transitions, one at pH ≈ 2–4 (corresponding to the pK_a* of

the excited-state phenol) and the other at pH ≈ 10 (corre-

Reaction mechanisms

We begin our mechanistic discussion for the most reactive

compound that undergoes photocyclization, viz., 4 (Scheme 5).

The quantum yield for the photocyclization of 4 is high in

aprotic solvents such as CH₃CN (Φ = 0.3) and THF (Φ =

0.4). Since a small amount of added H₂O does not change

the quantum yield dramatically (Table 1), the reaction can-

not be ascribed by a mechanism mediated by traces of resid-

ual water or other protic solvent impurities. Although the

molecules may be intermolecularly H bonded in the crystal-

line state, such intermolecular H bonding is not likely in di-

lute solution in a polar solvent such as CH₃CN. Thus, the

reaction in aprotic solvent is most likely via an excited-state

intramolecular proton transfer (ESIPT) mechanism, as

1314

Can. J. Chem. Vol. 83, 2005

Scheme 5.

h␯

-H

3

Twist

(Fast)

2

-OH

shown in Scheme 5, in which the ArOH proton is first trans-

ferred to the ArCH₂OH oxygen atom of the same molecule

via an intramolecular H bond to give 4a*. This is followed

by twisting to the planar geometry to give 4b*, followed by

the loss of a molecule of H₂O to give the BQM intermediate

39, which subsequently undergoes electrocyclization in a

fast reaction (k >> 10⁹s^–1) to give the chromene product 19.

This mechanism is supported by the Stokes shift data ob-

served in the fluorescence measurements, which indicated

that the neutral forms of 4 and 6 do not twist to the planar

geometry, but their deprotonated forms do.

Since the product quantum yield for 4 is also much higher

than its analogs (1, 5, and 6) in aprotic solvents, a faster

ESIPT rate could be responsible for this. More insights on

this aspect are obtained from the fluorescence quantum yield

and fluorescence lifetime studies (Table 2). The kinetic ex-

23, 31, and 32); (k_f⁰+ Σk_d) is the reciprocal of the fluores-

cence lifetime (τ_f, Table 2) of the model compound. The

ESIPT rate k_–Hestimated for 4 using this method is 2 ×

10⁹s^–1, which is nearly two orders of magnitude larger than

those for 5 (2 × 10⁷s^–1) and 6 (3 × 10⁷s^–1).

Addition of H₂O to CH₃CN can change the mechanism

for photocyclization. It was observed that upon the addition

of water (0–2.5 mol/L), the fluorescence intensity of 31

(model of 5) increased by 11% while both 32 and 23 showed

significant fluorescence quenching (fluorescence intensity

decreased by 41% and 53%, respectively), suggesting that

23 (or 4) and 32 (or 6) undergo efficient ESPT to water.

Thus, the increase in cyclization quantum yield with added

water (Table 2) can be rationalized as being due to the oper-

ation of a more efficient cyclization pathway in water, viz.,

ESPT to water to generate the phenolate or napththolate ion,

followed by twisting to a more planar state and the loss of

hydroxide ion. However, when the water content is suffi-

ciently high, the quantum yield for cyclization decreases ow-

ing to the competition of nucleophilic attack on the quinone

methide by water itself, which leads back to the starting ma-

terial (i.e., 39 to 4). In the case of 5, where the cyclization

quantum yield is intrinsically lower, other effects appear to

dominate when water is added, making the fluorescence

emission stronger at low water content than in pure CH₃CN.

In 0.01 mol/L NaOH, if we also assume that the reactive

species has similar k_f⁰and Σk_dvalues as its model com-

pound, k_–OHfor 5 (its naphtholate ion) can be estimated

from the fluorescence lifetime of 5 (τ_f= 10.8 ns) and that of

its model compound 31 (τ_f= 15.0 ns), i.e., k_–OH= (1/10.8 –

1/15) × 10⁹= 2.6 × 10⁷s^–1, which is in the same order of

magnitude as the natural fluorescence rate constant (k_f⁰) for

these types of molecules. This also explains the fact that 5

has a strong fluorescence emission in aq. 0.01 mol/L NaOH

solution and a lower reaction quantum yield. In contrast, 6

showed complete fluorescence quenching in this medium,

pression for fluorescence quantum yield (Φ) is the ratio of

f

the natural fluorescence rate constant (k_f⁰) to the sum of the

rates for all possible deactivation pathways, given by eq. [3],

[3]

Φ_f= k_f⁰/(k_f⁰+ Σk_d+ Σk_r)

where Σk_drepresents the sum of nonradiative deactivation

rates such as the rate of internal conversion (k_ic) and the rate

of intersystem crossing (k_isc); Σk_rrepresents the sum of all

possible chemical reactions of S₁. According to the proposed

mechanism (Scheme 5), Σk_rin this case is the ESIPT rate k_–H.

For the model compound 23, Σk_ris zero in neat CH₃CN. If

we assume that 4 and 23 have similar k_f⁰and Σk_dvalues,

then k_–Hcan be estimated by eq. [4]:

⎡

⎤

(Φ_f)_mdl

(Φ_f)_rec

0

[4]

k_−H= (k_f+ Σk_d) _⎢

−1

⎥

⎦

⎣

where the rec subscript refers to the reactive compound (i.e.,

4–6) and mdl to the corresponding model compound (i.e.,

Shi and Wan

1315

Scheme 6.

+

h␯

-

Scheme 7.

although its model compound 32 does show reasonably

strong fluorescence emission. This suggests that k_–OHfor 6

is much larger than k_f⁰, i.e., k_–OH>> 10⁷s^–1. It is therefore

clear that the low quantum efficiency observed for the

photocyclization of 5 is due to both a smaller k_–Hand a

h␯

smaller k_–OH

.

LFP and product studies indicate that the efficiencies for

the photogeneration of BQMs 36 and 37 are significantly

different, which suggests that 8 and 11 have significantly

different reactivities (different deprotonation rate k_–Hand

(or) dehydroxylation rate k_–OH). Fluorescence quenching ex-

periments on these systems reveal some interesting insights

on these processes. It can be easily seen from Table 2 that

significant fluorescence quenching effect is only observable

for the biphenyl system bearing both the phenolic and the

benzylic OH (and also at the right positions), i.e., 8 and 10.

2

anti

Φ of 8 decreased by a factor of five and that of 10 decreased

f

by a factor of 3.5 when the solvent was changed from neat

CH₃CN to H₂O–CH₃CN (1:1), suggesting very large k_–Hand

(or) k_–OHfor these molecules. When either the CH₂OH moi-

ety (of ArCH₂OH) is replaced by CH₃or the phenolic OH is

replaced by an OCH₃, there is either no fluorescence

quenching or only a minor quenching effect. In addition, no

fluorescence emission from the phenolate ions of 8 and 10

could be detected, although model compounds 33 and 34 did

show strong fluorescence emission from their respective

phenolate ions. These observations suggest that the depro-

tonation and dehydroxylation processes in 8 and 10 are un-

usually fast and may actually be concerted, although it is

shown as a step-wise process in Scheme 6. Note that in this

mechanism, the twisting about the biphenyl ring system

takes place without the need for deprotonation of the phenol,

consistent with the fluorescence results.

syn

intermediates, the working mechanism for the reaction is

shown in Scheme 7. The isolation of fluorene products 27

and 28 is consistent with the formation of the zwitterionic

BQMs anti-39 and syn-39. The quantum yield for formation

of these intermediates is quite high (Φ = 0.17, vide supra)

and indicates that there is substantial charge transfer (elec-

tronic communication) between the o and m positions on dif-

ferent rings in the biphenyl system. Note that there is little to

no electronic communication between such positions in the

ground-state molecule, so the previous finding could be used

to design systems where electronic communication can be

switched on by photoactivation.

Although results from product studies indicate that the

photolysis of 10 (Scheme 7) should proceed via BQM 39,

LFP studies showed no detectable transients longer than a

few ca. 10 ns. This is consistent with the expectation since

benzylic-type carbocations 39 should be shorter lived than a

few nanoseconds (15). Assuming the presence of BQM 39

Summary

The main conclusion from this work is that a variety of

1316

Can. J. Chem. Vol. 83, 2005

biaryl quinone methides (extended analogs of simple benze-

noid quinone methides) are readily photogenerable from the

corresponding hydroxymethyl substituted phenols (or equiv-

alent compounds) in aqueous solution. Indeed, it would ap-

pear that at least for the biphenyl quinone methide series, all

isomeric biphenyl quinone methides are photogenerable in

aqueous solution, albeit with differing quantum efficiencies.

The lifetimes of these quinone methides will range from

subnanoseconds to as long as a few milliseconds. Two driv-

ing forces have been identified as being responsible for this

kind of chemistry: (i) the tendency for twisted ground-state

biaryls to twist to a more planar geometry in the excited

state; and (ii) a large degree of charge transfer (to the other

ring) inherent in the excited-state phenolate ion.

Materials

The following compounds were purchased from Aldrich:

naphthalene (recrystallized from 95% CH₃CH₂OH), 2-

aminopyridine (99.9%, used as secondary standards for fluo-

rescence quantum yield experiments), 2-naphthol (35)

(recrystallized from 95% EtOH), BBr₃(1 mol/L in CH₂Cl₂),

2-bromobenzoic acid, 3-bromoanisole, 4-bromoanisole, 2-

bromotoluene (15), and 4-(4′-hydroxyphenyl)benzoic acid.

Biphenyl 1 was synthesized according to a previous method

(5a). Lactones 12 and 14 were synthesized from known pro-

cedures (6).

2′-Naphthyl-2-bromobenzoate

2-Bromobenzoic acid (20 g, 0.1 mol) was dissolved in a

mixture of SOCl₂(300 mL) and toluene (500 mL) and

refluxed under N₂for 3 h. The solvent and extra SOCl₂were

then removed under reduced pressure distillation to give the

acid chloride. The residue was then redissolved in CH₂Cl₂

(500 mL). To this solution, a solution of 2-naphthol (14.4 g)

and triethylamine (15 g) in 250 mL CH₂Cl₂was added

dropwise and stirred at room temperature (rt) for several

hours. After reaction, the mixture was first washed with wa-

ter, then dried over MgSO₄. The solvent was evaporated to

give a brown solid. Pure 2′-naphthyl-2-bromobenzoate was

obtained upon recrystallization from toluene (~80%); mp 79

Experimental section

Instrumentation

¹H NMR spectra were recorded on a Bruker WM-250

(250 MHz), AMX (360 MHz), or AC-300 (300 MHz) instru-

ment in CDCl₃or (CD₃)₂CO, unless otherwise noted. IR

spectra were recorded on a PerkinElmer 283 instrument.

UV–vis spectra were measured on a Cary 5 instrument.

Mass spectra were obtained on a Finnigan 3300 instrument

(CI). High-resolution mass spectra (HR-MS) were obtained

on a Kratos Concept H (EI) instrument. Melting points were

determined on a Koefler hot stage microscope (uncorrected).

Fluorescence spectra were obtained on a Photon Technology

International (PTI) A-1010 fluorimeter. Fluorescence life-

times were determined on a PTI LS-1 instrument in lifetime

mode. Gas chromatography was performed on a Varian 3700

instrument with a Hewlett Packard 3390A integrator, with a

column temperature of 100–250 °C (initial 100 °C, rate

10 °C/min, final 250 °C). Measurements of pH was carried

out on a Corning pH meter 140. Preparative photolyses were

carried out using a Rayonet RPR 100 photochemical reactor

using 254 or 300 nm lamps. An Oriel 200 W Hg light source

and an Applied Photophysis monochromator on an optical

bench were employed in small scale photolysis and in prod-

uct quantum yield determinations.

1

to 80 °C. H NMR (250 MHz, (CD₃)₂CO) δ: 7.45–7.66 (m,

5H), 7.78–7.88 (m, 2H), 7.90–8.00 (m, 2H), 8.02 (d, 1H, J =

8.8 Hz), 8.06–8.16 (m, 1H). MS (CI) m/z (relative intensity):

327/329 (M + 1) (14/13), 183/185 (M – C₁₀H₈O) (100/99).

5H-Dibenzo[c,f]chromen-5-one (13)

2′-Naphthyl-2-bromobenzoate (5.0 g, 15 mmol), sodium

acetate (2.5 g, 30 mmol), triphenylphosphane (2.4 g,

9 mmol), and palladium(II) acetate (0.10 g, 0.45 mmol) were

dissolved in dry N,N′-dimethyl acetamide (180 mL) and

heated to 120–130 °C for 24 h. After cooling to rt, CH₂Cl₂

(720 mL) was added. This solution was washed with 10%

HCl three times. The organic layer was dried over MgSO₄

then the solvent was removed using a rotary evaporator. A

mixture of 13 and its β-isomer (3:1 mixture, total 1.5 g,

6 mmol, 40%) was obtained after chromatography (silica

gel, CH₂Cl₂). This mixture was separated by slow

recrystallization from hexanes (the two isomers separated

from the mother liquor at different times). Pure 13 was ob-

tained after an additional recrystallization procedure (final

yield 25%); mp 160–162 °C. IR (KBr, cm^–1) ν: 3040 (w, C-

H aromatic), 1710 (s, C=O), 1605, 1590 (m, C=C), 1500

(m), 1458 (m), 1300 (s), 1270 (s), 1210 (s), 1110 (s), 1065

(s), 805 (s), 780 (m), 740 (s). ¹H NMR (250 MHz,

(CD₃)₂CO) δ: 7.50 (d, 1H, J = 7.9 Hz), 7.60 (ddd, 1H, J = 8,

8, 1 Hz), 7.73 (ddd, 2H, J = 8, 8, 1 Hz), 7.98–8.14 (m, 3H),

8.41 (dd, 1H, J = 8, 1 Hz), 8.75 (d, 1H, J = 8 Hz), 8.85 (d,

1H, J = 8 Hz). HR-MS (EI) calcd. for C₁₇H₁₀O₂: 246.0681;

found: 246.0688. Chemical anal. calcd. for C₁₇H₁₀O₂(%): C

82.91, H 4.09; found: C 82.80, H 4.08.

Common laboratory reagents

Dichloromethane (ACS reagent grade) was distilled be-

fore use. Anhydrous THF was obtained by distillation from

potassium metal. Anhydrous diethyl ether and dimethyl

acetamide were purchased from the Aldrich Chemical Co.

and used as received. Other solvents (ACS reagent grade)

used for synthetic purposes were used without further purifi-

cation. Deuterated solvents, D₂O, CDCl₃, (CD₃)₂CO, CD₃CN,

and DMSO-d₆, were purchased from the Cambridge Isotope

Laboratory containing 99.9% D. Standard buffer solutions

used for fluorescence and LFP studies were purchased from

the Fisher Scientific Co. Solvents used for fluorescence

studies, CH₃CN (HPLC grade), 95% CH₃CH₂OH, and C₆H₁₂

(spectral grade), were checked for spurious emission before

use. Anhydrous CH₃CN used for fluorescence studies was

distilled over CaH₂and used immediately. Preparative TLC

was carried out on 20 cm × 20 cm silica gel GF Uniplates

(Analtech).

General procedure for LiAlH₄reduction

A solution of the substrate in dry THF was added (under

stirring, rt) dropwise into a round-bottomed flask charged

with LiAlH₄(excess) and a small amount of dry THF. After

the addition was completed, the mixture was refluxed for

Shi and Wan

1317

1 h. Then the mixture was cooled to rt and poured into a

mixture of ice and 10% HCl. The corresponding biaryl

methanol was obtained after standard extraction (Et₂O–

CH₂Cl₂, 1:2), drying (MgSO₄), and evaporation (rotary

evaporator) of the solvent. Pure product (4–6) was obtained

after recrystallization from toluene (yield ≥90%).

Chemical anal. calcd. for C₁₃H₁₂O₂(%): C 77.98, H 6.04;

found: C 78.06, H 6.03.

2-Hydroxy-2′-methyl-1,1′-binaphthyl (32)

A sample of 10 mg chromene 21 (photocyclization prod-

uct of 6) was refluxed with LiAlH₄in dry THF (ca. 10 mL)

for 6 h. After reaction, the solution was then quenched with

5% HCl–ice and extracted with Et₂O (4 × 50 mL). The sol-

vent was evaporated and the residue was purified on prepa-

ratory TLC (CH₂Cl₂). Recrystallization from toluene gave a

white crystal, mp 163 to 164 °C. ¹H NMR (360 MHz,

(CD₃)₂CO) δ: 2.14 (s, 3H, CH3), 6.85 (d, 1H, J ≈ 8 Hz),

7.13 (d, 1H, J ≈ 8 Hz), 7.16–7.31 (m, 3H), 7.35–7.43 (m,

2H), 7.55 (d, 1H, J = 8.5 Hz), 7.88–7.97 (m, 4H), 7.98 (s,

1H, exchangeable with D₂O, ArOH). ¹³C NMR δ: 20.3

(CH₃), 119.2, 119.3, 123.7, 125.0, 125.7, 126.4, 126.7,

127.2, 128.6, 128.8, 129.0, 129.6, 129.8, 130.2, 132.4,

133.4, 134.3, 134.8, 136.6, 153.2. HR-MS (EI) calcd. for

C₂₁H₁₆O: 284.1201; found: 284.1214.

2-[2′-(Hydroxymethyl)-1′-naphthyl]phenol (4)

According to the general procedure for LiAlH₄reduction,

1

2.5 g of 12 gave 2.4 g of 4 (95%); mp 167 to 168 °C. H

NMR (360 MHz, CD₃CN) δ: 4.42 (s, 2H, CH₂), 6.99–7.11

(m, 3H), 7.30–7.42 (m, 3H), 7.47 (ddd, 1H, J = 7.4, 7.4,

1.5 Hz), 7.74 (d, 1H, J = 8.5 Hz), 7.91 (d, 1H, J = 8.1 Hz),

7.95 (d, 1H, J = 8.6 Hz); additional ¹H NMR peaks in

(CD₃)₂CO: 4.20 (s, broad, 1H, exchangeable with D₂O, OH),

8.03 (s, broad, 1H, exchangeable with D₂O, ArOH). MS (CI)

m/z (relative intensity): 233 (M – OH^–) (100). Chemical

anal. calcd. for C₁₇H₁₄O₂(%): C 81.58, H 5.64; found: C

81.54, H 5.54.

4-(2′-Methylphenyl)anisole (17) and 3-(2′-methyl-

phenyl)anisole

1-[2′-(Hydroxymethyl)phenyl]-2-naphthol (5)

According to the general procedure for LiAlH₄reduction,

3-Bromoanisole (80 mmol) was dissolved in 100 mL of

toluene followed by the addition of Pd(PPh₃)₄(0.70 g). After

all the catalyst was dissolved, a solution of boronic acid 16

(100 mmol in 200 mL of 95% ethanol) and a solution of

K₂CO₃(20 g in 100 mL of water) were subsequently added.

The mixture was refluxed for 24 h, then worked up by addi-

tion of NaOH (10%, ca. 400 mL) followed by extraction into

CH₂Cl₂(3 × 400 mL). The organic layer was dried over

MgSO₄and the solvent evaporated to give the crude product

as a colourless oil (~90%). This crude product was used di-

rectly in the next step without further purification. 3-(2′-

1

2.0 g of 13 gave 1.9 g of 5 (95%); mp 167 to 168 °C. H

NMR (360 MHz, CD₃CN) δ: 4.21 (AB quartet, 2H, J =

13.5 Hz, CH₂), 7.06–7.12 (m, 1H), 7.13–7.17 (dd, 1H, J = 9,

2 Hz), 7.22 (d, 1H, J = 8.8 Hz), 7.27–7.33 (m, 2H), 7.41

(ddd, 1H, J = 7.4, 7.4, 1.5 Hz), 7.49 (ddd, 1H, J = 7.4, 7.4,

1.5 Hz), 7.67 (d, 1H, J = 7.4 Hz), 7.80–7.86 (m, 2H); addi-

1

tional H NMR peaks in (CD₃)₂CO: 4.10 (s, broad, 1H, ex-

changeable with D₂O, OH), 8.05 (s, broad, 1H,

exchangeable with D₂O, ArOH). MS (CI) m/z (relative in-

tensity): 233 (M – OH^–) (100). Chemical anal. calcd. for

C₁₇H₁₄O₂(%): C 81.58, H 5.64; found: C 81.50, H 5.56.

1

Methylphenyl)anisole: H NMR (90 MHz, CDCl₃) δ: 2.20

(s, 3H, CH₃), 3.79 (s, 3H, OCH₃), 6.85–7.05 (m, 3H), 7.05–

7.30 (m, 5H). MS (CI) m/z (relative intensity): 199 (M + 1)

(100). In a similar manner, use of 4-bromoanisole gave 4-

2-Hydroxy-2′-hydroxymethyl-1,1′-binaphthyl (6)

According to the general procedure for LiAlH₄reduction,

1

3.0 g of 14 gave 2.8 g of 6 (95%); mp 172 to 173 °C. H

(2′-methylphenyl)anisole (17): H NMR (90 MHz, CDCl₃)

NMR (360 MHz, CD₃CN) δ: 4.30 (AB quartet, 2H, J =

13.3 Hz, CH₂), 6.82 (d, 1H, J = 8 Hz), 7.50 (d, 1H, J =

8 Hz), 7.19 (ddd, 1H, J = 7.6, 7.6, 1.4 Hz), 7.26 (ddd, 1H,

J = 7.6, 7.6, 1.3 Hz), 7.30 (ddd, 1H, J = 7.6, 7.6, 1.3 Hz),

7.31 (d, 1H, J = 8.9 Hz), 7.46 (ddd, 1H, J = 7.6, 7.6,

1.3 Hz), 7.86 (d, 1H, J = 8.6 Hz), 7.90 (d, 1H, J ≈ 9 Hz),

7.94 (d, 1H, J = 9.0 Hz), 7.98 (d, 1H, J ≈ 9 Hz), 8.06 (d, 1H,

J = 8.4 Hz); additional ¹H NMR peaks in (CD₃)₂CO: 4.10 (s,

broad, 1H, exchangeable with D₂O, OH), 8.20 (s, broad, 1H,

exchangeable with D₂O, ArOH). MS (CI) m/z (relative in-

tensity): 283 (M – OH^–) (100). Chemical anal. calcd. for

C₂₁H₁₆O₂(%): C 83.98, H 5.37; found: C 83.54, H 5.24.

δ: 2.20 (s, 3H, CH₃), 3.79 (s, 3H, OCH₃), 6.80–7.00 (m,

2H), 7.05–7.35 (m, 6H). MS (CI) m/z (relative intensity):

199 (M + 1) (100).

4-[2′-(Bromomethyl)phenyl]anisole

To a solution of 16 g of 4-(2′-methylphenyl) anisole 17

(80 mmol) in 250 mL CCl₄, a slurry of 14.2 g of NBS

(80 mmol) in 100 mL CCl₄was added. Then 0.3 g of

benzoperoxide was added. The mixture was refluxed for

10 h then allowed to cool to rt and filtered. Upon washing

with H₂O (3 × 100 mL), drying over MgSO₄, and evapora-

tion of the solvent (rotary evaporator), 20 g (~90%) of crude

product was obtained. ¹H NMR (90 MHz, CDCl₃) δ: 3.79 (s,

3H, OCH₃), 4.35 (s, 2H, CH₂), 6.75–7.05 (m, 2H), 7.05–

7.65 (m, 6H). The crude product was used in the next step

without further purification.

4-[4′-(Hydroxymethyl)phenyl]phenol (11)

According to the general procedure for LiAlH₄reduction,

2.0 g of 4-(4′-hydroxyphenyl)benzoic acid (in 200 mL dry

THF) gave 1.75 g (93%) of pure 11 after recrystallization

1

3-[2′-(Bromomethyl)phenyl]anisole

from a mixture of toluene and hexanes; mp 205–208 °C. H

NMR (300 MHz, (CD₃)₂CO) δ: 4.18 (t, 1H, J = 5.5 Hz, ex-

changeable with D₂O, OH), 4.64 (d, 2H, J = 5.5 Hz, CH₂),

6.90–6.99 (m, 2H), 7.35–7.45 (m, 2H), 7.45–7.60 (m, 4H),

8.43 (s, 1H, exchangeable with D₂O, ArOH). MS (CI) m/z

(relative intensity): 201 (M + 1) (61), 183 (M – OH) (100).

To a solution of 9.7 g of 3-(2′-methylphenyl) anisole

(49 mmol) in 800 mL CCl₄, a slurry of 9.0 g of NBS

(50 mmol) in 200 mL CCl₄was added. Then 0.5 g of

benzoperoxide was added. The mixture was refluxed for

30 h and allowed to cool to rt and filtered. Upon washing

1318

Can. J. Chem. Vol. 83, 2005

with H₂O (3 × 100 mL), drying over MgSO₄, and evapora-

tion of the solvent (rotary evaporator), 20 g (~90%) of crude

product was obtained. H NMR (300 MHz, CDCl₃) δ: 3.87

(s, 3H, OCH₃), 4.49 (s, 2H, CH₂), 6.85–7.10 (m, 3H), 7.20–

7.45 (m, 4H), 7.50–7.59 (m, 1H). MS (CI) m/z (relative in-

tensity): 277/279 (M + 1) (7/6), 197 (M – Br) (100). The

crude product was used in the next step without further puri-

fication.

4.10 (t, 1H, J = 5.5 Hz, exchangeable with D₂O, OH), 4.55

(d, 2H, J = 5.5 Hz, CH₂), 6.80–6.90 (m, 3H), 7.10–7.40 (m,

4H), 7.60 (d, 1H, J = 8 Hz), 8.40 (s, 1H, exchangeable with

D₂O, ArOH). MS (CI) m/z (relative intensity): 183 (M –

OH) (100). Chemical anal. calcd. for C₁₃H₁₂O₂(%): C

77.98, H 6.04; found: C 78.19, H 6.02.

1

2-(2′-Methyl-1′-naphthyl)phenol (23) and 2-(2′-

methylphenyl)phenol (30)

The corresponding biaryl alcohol (4 or 1, 50–100 mg) was

photolyzed (254 nm) in 100 mL of 2-propanol in the pres-

ence of excess (~0.5 g) NaBH₄for 2 h. Then 10% HCl (50–

100 mL) was added, and the mixture was extracted with

Et₂O–CH₂Cl₂(1:2, 3 × 80 mL). Pure 23 or 30 were isolated

by preparative TLC (CH₂Cl₂–hexanes, second component)

4-[2′-(Hydroxymethyl)phenyl]anisole (7) and 3-[2′-

(hydroxymethyl)phenyl]anisole (9)

4-[2′-(Bromomethyl)phenyl]anisole (15 g) or 3-[2′-

(bromomethyl)phenyl] anisole (15 g) was dissolved in

750 mL of acetone–H₂O (1:1, v/v) with 5.0 g of added

NaHCO₃. The mixture was refluxed overnight. The solution

was cooled to rt and extracted with CH₂Cl₂(twice) then with

diethyl ether (once). The extract was combined and dried

over MgSO₄. Upon evaporation of the solvent, 11 g (90%)

of crude product was obtained. Pure product was obtained

after recrystallization from toluene–hexanes. 4-[2′-

(Hydroxymethyl)phenyl]anisole (7): mp 77 to 78 °C. ¹H

NMR (360 MHz, (CD₃)₂CO) δ: 4.17 (t, 1H, J = 5.5 Hz, ex-

changeable with D₂O, OH), 4.55 (d, 2H, J = 5.5 Hz, CH₂),

6.96–7.01 (m, 2H), 7.18–7.23 (m, 1H), 7.26–7.37 (m, 4H),

7.58–7.63 (m, 1H). MS (CI) m/z (relative intensity): 215

(M + 1) (23), 197 (M – OH) (100). Chemical anal. calcd. for

C₁₄H₁₄O₂(%): C 78.48, H 6.59; found: C 78.56, H 6.59. 3-

1

as a colourless oil. 2-(2′-Methyl-1′-naphthyl)phenol (23): H

NMR (300 MHz, CD₃CN) δ: 2.18 (s, 3H, CH₃), 6.6 (s, 1H,

exchangeable with D₂O, ArOH), 6.95–7.09 (m, 3H), 7.25–

7.52 (m, 5H), 7.77–7.90 (m, 2H). MS (CI) m/z (relative in-

tensity): 235 (M + 1) (100). 2-(2′-Methylphenyl)phenol (30):

¹H NMR (360 MHz, (CD₃)₂CO) δ: 2.16 (s, 3H, CH₃), 6.89

(ddd, 1H, J = 7.4, 7.4, 1.2 Hz), 6.95 (dd, 1H, J = 8.0,

1.2 Hz), 7.05 (dd, 1H, J = 7.4, 1.7 Hz), 7.11–7.25 (m, 5H),

8.15 (s, 1H, exchangeable with D₂O, ArOH). ¹³C NMR δ:

20.1, 116.4, 120.3, 126.3, 128.0, 129.3, 129.8, 130.4, 130.9,

131.6, 137.6, 139.6, 155.1. HR-MS (EI) calcd. for C₁₃H₁₂O:

184.0888; found: 184.0893.

1

[2′-(Hydroxymethyl)phenyl]anisole (9): mp 63 to 64 °C. H

NMR (300 MHz, (CD₃)₂CO) δ: 3.80 (s, 1H, OCH₃), 4.25 (t,

1H, J = 5.5 Hz, exchangeable with D₂O, OH), 4.65 (d, 2H,

J = 5.5 Hz, CH₂), 6.85–7.00 (m, 3H), 7.15–7.45 (m, 4H),

7.55–7.66 (d, 1H, J = 8Hz). MS (CI) m/z (relative intensity):

197 (M – OH) (100). Chemical anal. calcd. for C₁₄H₁₄O₂

(%): C 78.48, H 6.59; found: C 78.31, H 6.46.

General procedure for the synthesis of 31, 33, and 34

A solution of the biaryl alcohol 5, 7, or 10 (1 equiv.) in

CH₂Cl₂(70 mL per 10 mmol) was added dropwise into a so-

lution of BBr₃(1.5 equiv., 0.15 mol/L in CH₂Cl₂) at rt over

1 h. The solution was stirred at rt for another 4 h. The reac-

tion was then quenched by the addition of water. The water

layer was extracted with diethyl ether (3 × 150 mL). The

combined organic extract was dried over MgSO₄and the sol-

vent evaporated on the rotary evaporator to give an oily resi-

due. This residue was dissolved in dry THF and the solution

was added (under stirring, rt) dropwise into a round-

bottomed flask charged with LiAlH₄(excess) and a small

amount of dry THF. After the addition was completed, the

mixture was refluxed for 1 h. Then the mixture was cooled

to rt and poured into a mixture of ice and 10% HCl. The

corresponding biaryl methanol (crude product) was obtained

after extraction (Et₂O–CH₂Cl₂, 1:2), drying (MgSO₄), and

evaporation (rotary evaporator) of the solvent.

4-[2′-(Hydroxymethyl)phenyl]phenol (8) and 3-[2′-

(hydroxymethyl)phenyl] phenol (10)

A solution of 10 g (36 mmol) 4-[2′-(hydroxymethyl)phe-

nyl]anisole (7) or 3-[2′-(hydroxymethyl)phenyl]anisole (9)

in 750 mL CH₂Cl₂was added dropwise into a solution of

BBr₃(0.1 mol/L in CH₂Cl₂, 180 mL) at rt over 1 h. The so-

lution was stirred for an additional 4 h. The reaction was

quenched by addition of ca. 200 mL of water. The water

layer was extracted with diethyl ether (3 × 150 mL). The

combined organic extract was dried over MgSO₄and the sol-

vent evaporated to give an oily residue. This residue was dis-

solved in ca. 500 mL of acetone–H₂O (1:1, v/v) and refluxed

overnight. The solution was cooled to rt and extracted with

CH₂Cl₂(twice) then with diethyl ether (once). The extract

was combined and dried over MgSO₄. The crude product

was obtained upon evaporation of the solvent. Pure product

(6.0 g, 80%) was obtained after recrystallization from tolu-

ene. 4-[2′-(Hydroxymethyl)phenyl]phenol (8): mp 121 to

1-(2′-Methylphenyl)-2-naphthol (31)

5 (10 mg) was converted to 31 using the preceding gen-

eral procedure. Pure 31 (8 mg) was obtained as a white solid

after purification by preparatory TLC (CH₂Cl₂–hexanes,

1

1:1), mp 91 to 92 °C. H NMR (300 MHz, (CD₃)₂CO) δ:

1.99 (s, 3H, CH₃), 7.10–7.20 (m, 2H), 7.25–7.40 (m, 6H),

7.80–7.87 (m, 2H), 7.95 (s, 1H, exchangeable with D₂O,

ArOH). MS (CI) m/z (relative intensity): 235 (M + 1) (100).

1

122 °C. H NMR (300 MHz, (CD₃)₂CO) δ: 4.10 (t, 1H, J =

5.5 Hz, exchangeable with D₂O, OH), 4.55 (d, 2H, J =

5.5 Hz, CH₂), 6.83–6.93 (m, 2H), 7.12–7.35 (m, 5H), 7.53–

7.62 (m, 1H), 8.45 (s, 1H, exchangeable with D₂O, ArOH).

MS (CI) m/z (relative intensity): 183 (M – OH) (100).

Chemical anal. calcd. for C₁₃H₁₂O₂(%): C 77.98, H 6.04;

found: C 78.03, H 6.15. 3-[2′-(Hydroxymethyl)phenyl]phe-

nol (10): mp 96 to 97 °C. ¹H NMR (300 MHz, (CD₃)₂CO) δ:

4-(2′-Methylphenyl)phenol (33)

According to the preceding general procedure, 1.07 g of 7

gave 0.85 g (90%) of pure 33 as a white solid after

1

recrystallization from toluene, mp 76 to 77 °C. H NMR

(360 MHz, DMSO-D₆) δ: 2.21 (s, 3H, CH₃), 6.78–6.84 (m,

Shi and Wan

1319

2H), 7.09–7.15 (m, 2H), 7.11–7.16 (m, 1H), 7.16–7.22 (m,

2H), 7.22–7.26 (m, 1H), 9.55 (s, 1H, exchangeable with

D₂O, ArOH). MS (CI) m/z (relative intensity): 185 (M + 1)

(100).

7.9, 7.9, 1.5 Hz), 7.42 (d, 1H, J = 8.3 Hz), 7.52 (ddd, 1H,

J = 7.4, 7.4, 1.4 Hz), 7.58 (ddd, 1H, J = 7.6, 7.6, 1.6 Hz),

7.87 (d, 1H, J = 8.2 Hz), 7.94–7.98 (m, 1H), 8.05–8.09 (m,

1H), 8.56–8.61 (m, 1H). HR-MS (EI) calcd. for C₁₇H₁₂O:

232.0888; found: 232.0892.

3-(2′-Methylphenyl)phenol (34)

For analytical runs (in CH₃CN), the preceding solution

was sampled at set irradiation times and injected into the GC

to determine the conversion. The GC retention times (t_r)

were 12.8 min for 19 and 13.1 min for 4.

According to the preceding general procedure, 0.20 g of

1

10 gave 0.17 g (90%) of pure 34 as a colourless oil. H

NMR (360 MHz, (CD₃)₂CO) δ: 2.24 (s, 3H, CH₃), 6.74–6.87

(m, 3H), 7.14–7.30 (m, 5H), 8.20 (s, 1H, exchangeable with

D₂O, ArOH). MS (CI) m/z (relative intensity): 185 (M + 1)

(100).

Photolysis of 2-[2′-(hydroxymethyl)-1′-naphthyl]phenol

(4) in H₂O–H₃CN (1:1)

A solutionof 4 (50 mg) in 100 mL of solvent was photo-

lyzed (254 nm) in a quartz tube for 45 min using the general

procedure. After photolysis, the solution was worked using

the general procedure (CH₂Cl₂–Et₂O (2:1), 3 × 150 mL). GC

General photolysis procedure

Preparative scale photolysis was carried out in a Rayonet

RPR 100 photochemical reactor using 254 or 300 nm lamps.

The reaction mixture was contained in a 100 or 200 mL

quartz tube, which was cooled to ≈15 °C using a cold finger

(tap water) and purged continuously during photolysis with a

stream of argon via a stainless steel syringe needle. Micro-

scale photolysis was carried out in a 3 mL quartz cuvette or

in an NMR tube, and irradiated either in a Rayonet RPR 100

photochemical reactor (254 or 300 nm) or on an optical

bench (280 nm) equipped with an Oriel 200 W Hg lamp and

an Applied Photophysics monochromator. Solutions were

purged with a stream of argon via a stainless steel syringe

needle for 5 min and sealed before photolysis.

1

and H NMR showed that this sample contained ca. 60% of

19 and ca. 40% of 4. No side products were detectable. Ana-

lytical runs were carried out as described previously, using

GC analysis.

Photolysis of 2-[2′-(hydroxymethyl)-1′-naphthyl]phenol

(4) in 100% CH₃OH

A solution of 4 (50 mg) in CH₃OH (100 mL) was

photolyzed (254 nm) in a quartz tube for 5–45 min using the

general procedure. The solvent was then removed by rotary

evaporation. The residue (maximum yield, 75% of methyl

1

ether 22 and 25% chromene 19 by H NMR) was chromato-

graphed on preparative TLC (silica, CH₂Cl₂). The second

band was collected to give pure 22 as a colourless oil. H

NMR (300 MHz, (CD₃)₂CO) δ: 3.24 (s, 3H, OCH₃), 4.35

(AB quartet, 2H, J = 12.5 Hz, CH₂), 6.95–7.15 (m, 3H),

7.30–7.55 (m, 4H), 6.69 (d, 1H, J = 8.4 Hz), 7.87–7.96 (m,

2H), 8.20 (s, 1H, exchangeable with D₂O, ArOH).

General workup procedure for photolysis in aqueous

solution

After photolysis, satd. NaCl solution was added and the

mixture was extracted three to five times with either CH₂Cl₂

or Et₂O. The combined organic extract was then dried over

MgSO₄and the solvent evaporated on the rotary evaporator.

1

Photolysis of 6H-naphtho[2,1-c]chromene (19) in 100%

CH₃CN or THF

Photolysis of 2-[2′-(hydroxymethyl)phenyl]phenol (1) in

CH₃CN

A solution of 20 mg of 19 in 60 mL of CH₃CN (or THF)

was photolyzed (254 nm) in a quartz tube for 30 min using

the general procedure. The solvent was then removed by ro-

tary evaporation. H NMR of the residue was identical to

that of the starting material 19.

A solution of 1 (50 mg) in CH₃CN (80 mL) was

photolyzed (254 nm) in a quartz tube for 30 min using the

general procedure. The solvent was then removed by rotary

evaporation. The residue (80% conversion to chromene 3 by

¹H NMR) was chromatographed on preparative TLC (silica,

hexanes–CH₂Cl₂, 2:1). The first band was collected to give

1

Photolysis of 6H-naphtho[2,1-c]chromene (19) in H₂O–

CH₃CN (1:1)

pure 6H-benzo[c]chromene (3) as a colourless oil. H NMR

(300 MHz, CD₃CN) δ: 5.12 (s, 2H, CH₂), 6.96 (dd, 1H, J =

8, 1 Hz), 7.06 (ddd, 1H, J = 8, 8, 1 Hz), 7.18–7.44 (m, 4H),

7.76 (d, 1H, J = 8 Hz), 7.80 (dd, 1H, J = 8, 1 Hz). MS (CI)

m/z (relative intensity): 183 (M + 1) (100).

A solution of 20 mg of 19 in 60 mL of H₂O–CH₃CN (1:1)

was photolyzed (254 nm) in a quartz tube for 30 min using

the general procedure then worked up using the general pro-

cedure (CH₂Cl₂, 3 × 100 mL). The residue contained 60% of

1

19 and 40% of 4 as determined by H NMR.

Photolysis of 2-[2′-(hydroxymethyl)-1′-naphthyl]phenol

(4) in aprotic solvent (100% CH₃CN or THF)

Photolysis of 6H-naphtho[2,1-c]chromene (19) in H₂O–

CH₃OH (1:1)

A solution of 4 (50 mg) in 100 mL of solvent was photo-

lyzed (254 nm) in a quartz tube for 45 min using the general

procedure. The solvent was then removed by rotary evapora-

According to the general procedure for microscale

photolysis, a solution (3.0 mL, 2.14 × 10^–4mol/L) of 19 in

H₂O–CH₃OH (1:1) was photolyzed in the Rayonet (254 nm)

in 30 s intervals. Between each irradiation period, the UV

spectrum of the solution was taken. When the UV spectrum

of the solution remained unchanged (after a total of five irra-

diations), the composition of this solution was determined

by UV photospectrophotometry by the absorbance of the so-

1

tion. The residue (>95% conversion to chromene 19 by H

NMR) was chromatographed on preparative TLC (silica,

hexanes–CH₂Cl₂, 2:1). The first band was collected to give

pure 6H-naphtho[2,1-c]chromene (19) as a colourless crystal

after recrfystallization from CH₃CN; mp 96 to 97 °C (lit.

1

value (16) mp 101 °C). H NMR (360 MHz, (CD₃)₂CO) δ:

5.13 (s, 2H, CH₂), 7.12–7.23 (m, 2H), 7.33 (ddd, 1H, J =

1320

Can. J. Chem. Vol. 83, 2005

lution at 337 nm (for 19, ε₃₃₇= 1.04 × 10⁴, ε₂₈₃= 3.57 ×

10³) and the absorbance increases at 283 nm (for 22 and 4,

ε₂₈₃= 9.23 × 10³), as 8% of 19 and 92% of the ring-opening

product 22 + 4. The solutions for several runs were collected

and extracted with CH₂Cl₂(5 × 5 mL). The combined ex-

Photolysis of 1-[2′-(hydroxymethyl)phenyl]-2-naphthol (5)

in 100% CH₃OH

Using the general procedure, a sample of ca. 3 mg of 5 in

3 mL of CH₃OH was irradiated in the Rayonet reactor for

15 min then solvent was evaporated. The ¹H NMR

(300 MHz, CDCl₃) of the residue showed a singlet at δ 5.07

corresponding to the formation of 20 (10%). In addition,

there was a weak, but sharp, singlet at δ 3.23 (OCH₃) and an

AB quartet centred at 4.12 ppm (J ≈ –11 Hz, CH₂), suggest-

ing the formation of traces of the corresponding methyl

ether product (<1%).

1

tracts were dried over MgSO₄and evaporated. H NMR of

the residue showed that alcohol 4 and methyl ether 22 were

formed in a 1:3 ratio, corresponding to a photostationary ra-

tio of 8:23:69 of 19:4:22.

Photolysis of 6H-naphtho[2,1-c]chromene (19) in 100%

CH₃OH

Photolysis of 5H-dibenzo[c,f]chromene (20) in H₂O–

CH₃OH (1:1)

According to the general procedure for microscale photo-

lysis, a solution (3.0 mL, 1.25 × 10^–4mol/L) of 19 in CH₃OH

was photolyzed in the Rayonet (254 nm) in 15 s intervals.

Between each irradiation, the UV spectrum of the solution

was taken. After 10 irradiations, the UV spectrum remained

unchanged. The composition of this solution contained 24%

of 19 and 76% of the ring-opening product 22 as determined

by the method described previously using UV spectrophoto-

metry. In another attempt, ca. 3 mg of 19 was dissolved in

3 mL of CH₃OH and photolyzed for 5 min. After the photo-

Using the general procedure for microscale photolysis, a

sample of ~2 mg of 20 was dissolved in 3 mL of H₂O–

CH₃OH (1:1) and photolyzed in the Rayonet reactor for

30 min and then worked up according to the general proce-

dure. The ¹H NMR (300 MHz, CDCl₃) of the residue showed

that the methyl ether product was obtained in ≈1% yield.

Photolysis of 2-hydroxy-2′-hydroxymethyl-1,1′-binaphthyl

(6) in H₂O–CH₃CN (1:1), 100% CH₃CN, or 100% THF

A solution of 6 (50 mg) in 100 mL of solvent was

photolyzed (254 nm) in a quartz tube for 30 min, then

worked up (CH₂Cl₂–Et₂O (2:1), 3 × 150 mL) using the gen-

eral procedure (aqueous) or solvent evaporated rotary

1

lysis, the solvent was evaporated. H NMR of the residue in-

dicated that methyl ether 22 (50%) was the only product.

Photolysis of 1-[2′-(hydroxymethyl)phenyl]-2-naphthol (5)

in H₂O–CH₃CN (1:1)

1

(nonaqueous). The residue (contained ~90% of 21 by H

NMR) was chromatographed on preparative TLC (silica,

hexanes–CH₂Cl₂, 2:1). The first band was collected to give a

white solid, which was recrystallized from CH₃CN to give

pure 4H-benzo[f]naphtho[2,1-c]chromene (21), mp 176–

A solution of 5 (50 mg) in 100 mL H₂O–CH₃CN (1:1)

was photolyzed (254 nm) in a quartz tube for 90 min and

then worked up (CH₂Cl₂–Et₂O (2:1), 3 × 150 mL) using the

1

general procedure. The H NMR of the residue showed that

1

178 °C. H NMR (360 MHz, (CD₃)₂CO) δ: 4.99 (d, 1H, J =

5H-dibenzo[c, f ]chromene (20) was obtained in 60% yield

(with 40% starting material left). Prolonged photolysis

(>10 h) could raise the conversion to ca. 90% with decom-

position of the product (ca. 10%). Pure 20 was isolated as a

white solid by preparative TLC (silica gel, CH₂Cl₂–hexanes

(1:1), first band collected). This solid was recrystallized

from CH₃CN, mp 55 to 56 °C. ¹H NMR (300 MHz,

(CD₃)₂CO) δ: 5.09 (s, 2H, CH₂), 7.25 (d, 1H, J = 8 Hz),

7.35–7.54 (m, 4H), 7.59 (dd, 1H, J = 8, 8 Hz), 7.85 (d, 1H,

J = 9 Hz), 7.93 (d, 1H, J = 8 Hz), 8.03 (d, 1H, J = 8 Hz),

8.55 (d, 1H, J = 9 Hz). HR-MS (EI) calcd. for C₁₇H₁₂O:

232.0888; found: 232.0893. Chemical anal. calcd. for

C₁₇H₁₂O (%): C 87.91, H 5.21; found: C 87.58, H 5.16.

–12.9 Hz, CH₂), 5.36 (d, 1H, J = –12.9 Hz, CH₂), 7.30–7.46

(m, 4H), 7.51 (ddd, 1H, J = 7.5, 7.5, 1.2 Hz), 7.60–7.67 (m,

3H), 7.93–8.03 (m, 4H). HR-MS (EI) calcd. for C₂₁H₁₄O:

282.1045; found: 282.1032. Chemical anal. calcd. for

C₂₁H₁₄O (%): C 89.34, H 5.00; found: C 88.97, H 4.84.

When photolyzed in neat THF or CH₃CN using the preced-

ing procedure, the reaction also gave the same product, but

the efficiency is significantly lower (by ca. 30%–50%).

For analytical runs, two methods were used: (i) a solution

of 6 was photolyzed in a quartz cuvette, and the formation of

21 was followed by UV spectrophotometry (absorbance at

360 nm, ε = 1.04 × 10⁴(mol/L)^–1cm^–1); (ii) ~5 mg of 6 was

dissolved in 0.4 mL of CD₃CN and purged with Ar for

10 min then sealed in an NMR tube. This sample was irradi-

ated in the Rayonet reactor (in the NMR tube, λ_ex= 300 nm)

for a series of set times and the conversions were determined

Analytical runs were carried out in a 3 mL quartz cuvette

using the general procedure for microscale photolysis, and

the yield of 20 was determined by the absorbance increases

of the solution at 350 nm (ε₃₅₀= 7.952).

1

by H NMR.

Photolysis of 1-[2′-(hydroxymethyl)phenyl]-2-naphthol (5)

in 100% CH₃CN or THF

Photolysis of 2-hydroxy-2′-hydroxymethyl-1,1′-binaphthyl

(6) in H₂O–CH₃OH (1:1) and 100% CH₃OH

Using the general procedure for microscale photolysis, 3

to approximately 4 mg of 5 was dissolved in 3 mL of solvent

and photolyzed in the Rayonet reactor for 10–15 min. After

A sample of 6 (20 mg) in 100 mL solvent was photolyzed

in the Rayonet reactor for 20 min, then worked up using the

1

method previously described. H NMR of the residue showed

1

photolysis, the solvent was evaporated. H NMR of the resi-

ca. 15% yield of the methyl ether product and ca. 85% of

21. The formation of the corresponding methyl ether product

was inferred from the observation of the following three sets

due indicated that 20 was obtained in 10%–20% yield. Due

to the low efficiency of this reaction, the conversion was

kept low in these experiments to avoid significant decompo-

sition of the reactants.

1

of characteristic H NMR (300 MHz, (CD₃)₂CO) peaks δ:

3.19 (s, 3H, OCH₃), 4.23 (AB quartet, 2H, J = –13 Hz,

Shi and Wan

1321

CH₂), and 6.88 (d, 1H, J = 8 Hz). No attempt was made to

isolate this product.

Photolysis of 4-[2′-(hydroxymethyl)phenyl]phenol (8) in

H₂O–CH₃OH (1:1)

A solution of 8 (30 mg) in CH₃OH–H₂O (1:1, 60 mL) was

irradiated in a quartz tube for 10 min, then worked up

(CH₂Cl₂, 3 × 150 mL) using the general procedure. The H

Photolysis of 4H-benzo[f]naphtho[2,1-c]chromene (21) in

CH₃CN–H₂O (1:1)

1

A solution of 21 (1.0 × 10^–4mol/L, in H₂O–CH₃CN (1:1))

was placed in a cuvette (3.0 mL) and irradiated at 254 or

300 nm. No changes were detectable by UV–vis spectro-

photometry even on prolonged irradiation, indicating the

lack of any conversion to the ring-opened product 6. When

this solution was photolyzed on the optical bench using

370 nm, the 360 nm band decreased and a new band at

280 nm grew in, suggesting the formation of 6.

NMR of the residue showed 60% yield of 4-[2′-(methoxy-

methyl)phenyl]phenol (25), which was isolated on prepara-

tive TLC (silica, CH₂Cl₂) as a colourless oil. ¹H NMR

(250 MHz, (CD₃)₂CO) δ: 3.28 (s, 3H, OCH₃), 4.31 (s, 2H,

CH₂), 6.87–6.95 (m, 2H), 7.15–7.35 (m, 5H), 7.45–7.51 (m,

1H), 8.60 (s, broad, exchangeable with D₂O, 1H, ArOH).

Photolysis of 4-[2′-(hydroxymethyl)phenyl]phenol (8) in

the presence of ethyl vinyl ether

A solution of 8 (30 mg) in CH₃CN–H₂O (2:1, 100 mL)

with 3.5 mL of ethyl vinyl ether added was irradiated in a

quartz tube for 15 min according to the general procedure.

After photolysis, the solution was worked up (CH₂Cl₂, 3 ×

Photolysis of 4H-benzo[f]naphtho[2,1-c]chromene (21) in

CH₃OH–H₂O (1:1)

A solution of 21 (1.0 × 10^–4mol/L in H₂O–CH₃OH (1:1),

saturated) was placed in a cuvette (3.0 mL) and irradiated at

254 nm. As the photolysis proceeded, the 360 nm band de-

creased and a band at 280 nm grew in, suggesting the forma-

tion of 6 and (or) the methyl ether substitution product. The

spectrum stopped changing at a state corresponding to 20%

of chromene 21 and 80% ring-opened products (6 + the

methyl ether product).

1

150 mL) using the preceding procedure. H NMR showed

that there was one major product, 3-[2′-(4′′-hydroxy-

phenyl)phenyl]propanal (26), which was purified on prepara-

1

tive TLC (silica, CH₂Cl₂). H NMR (300 MHz, (CD₃)₂CO)

δ: 2.55–2.61 (m, 2H, CH₂), 2.88–2.94 (m, 2H, CH₂), 6.87–

6.93 (m, 2H), 7.11–7.17 (m, 3H), 7.18–7.28 (m, 2H), 7.28–

7.33 (m, 1H), 8.50 (s, 1H, exchangeable with D₂O, ArOH),

9.64 (t, 1H, J = 1.4 Hz, RCHO).

Photolysis of 4H-benzo[f]naphtho[2,1-c]chromene (21) in

100% CH₃OH

Photolysis of 4-[2′-(hydroxymethyl)phenyl]phenol (8) at

different pH

A solution of 21 (ca. 5 mg in 3 mL CH₃OH) was photo-

lyzed (254 nm) in a quartz cuvette for 5 min, then the sol-

vent was removed by bubbling through the solution with a

stream of air. The cuvette was then placed in a round-

A solution of 8 (1.50 mL, 9 mg in 4.6 mL of CH₃OH) was

added to each of the cuvettes charged with either 1.50 mL of

H₂O or 0.05 mol/L H₂SO₄, respectively. These samples were

irradiated on the optical bench (280 nm) for 4.0 min, then

worked up using the general procedure. The residues of all

three samples contained (15 1)% of 25 as determined by

¹H NMR, suggesting no pH dependence in this region.

1

bottomed flask and left under vacuum for 1 h. H NMR of

the residue showed that the methyl ether substitution product

was obtained in about 30% yield.

Photolysis of 4-[2′-(hydroxymethyl)phenyl]anisole (7) in

CH₃OH–H₂O (1:1)

Photolysis of 3-[2′-(hydroxymethyl)phenyl]phenol (10) in

CH₃CN–H₂O (1:1)

A solution of 7 (90 mg) in CH₃OH–H₂O (1:1, 100 mL)

was photolyzed (254 nm) in a quartz tube for 10 min, then

worked up (CH₂Cl₂, 3 × 150 mL) using the general proce-

A solution of 10 (50 mg) in CH₃CN–H₂O (1:1, 100 mL)

was photolyzed (254 nm) in a quartz tube for 15 min, then

worked up (CH₂Cl₂, 3 × 150 mL) using the general proce-

dure. The ¹H NMR of the residue showed ~100% conversion

to 3-hydroxyfluorene (27) and 1-hydroxyfluorene (28) in a

ca. 10:1 ratio. The residue was chromatographed on prepara-

tive TLC to give a white solid (mixture of 27 and 28).

Recrystallization from toluene–hexanes gave pure 27, mp

1

dure. The H NMR of the residue showed two products, 4-

[2′-(methoxymethyl)phenyl]anisole (24) (23%) and 4-(2′-

methylphenyl)anisole (~3%). The residue was chromato-

graphed on preparative TLC (silica, hexanes–CH₂Cl₂, 1:1) to

1

give pure 24 as a colourless oil (second band collected). H

NMR (250 MHz, (CD₃)₂CO) δ: 3.28 (s, 3H, ROCH₃), 3.82

(s, 3H, ArOCH₃), 4.31 (s, 2H, CH₂), 6.95–7.03 (m, 2H),

7.18–7.39 (m, 5H), 7.47–7.53 (m, 1H).

1

138 to 139 °C. H NMR (360 MHz, (CD₃)₂CO) δ: 3.78 (s,

2H, CH₂), 6.80 (dd, 1H, J = 8.1, 2.4 Hz), 7.27 (ddd, 1H, J =

7.4, 7.4, 1.3 Hz), 7.30 (d, 1H, J = 2.4 Hz), 7.31–7.39 (m,

2H), 7.50–7.55 (m, 1H), 7.74–7.78 (m, 1H), 8.34 (s, 1H, ex-

changeable with D₂O, ArOH). MS (CI) m/z (relative inten-

Photolysis of 4-[2′-(hydroxymethyl)phenyl]anisole (7) in

different pH

1

sity): 183 (M + 1) (100). Pure 28 was not isolated. H NMR

A solution of 7 (1.50 mL, 10 mg in 4.6 mL of CH₃OH)

was added to each of the cuvettes charged with either

1.50 mL of H₂O or 0.05 mol/L H₂SO₄, respectively. These

samples were irradiated in the Rayonet (254 nm) for 12 min

and worked up according to the general procedure. The

yields of methyl ether 24 at different pH is in the ratio of

of 28 has CH₂at δ 3.82 (s).

Photolysis of 3-[2′-(hydroxymethyl)phenyl]phenol (10) in

CH₃OH–H₂O (1:1)

A solution of 10 (50 mg) in CH₃OH–H₂O (1:1, 100 mL)

was photolyzed (254 nm) in a quartz tube for 15 min, then

worked up (CH₂Cl₂, 3 × 150 mL) using the general

1

1:1.1:1.2 (pH 7 : pH 1 : H_o= +0.04) as determined by H

NMR, indicating that the quantum yield increased slightly at

low pH.

1

procedure. The H NMR of the residue showed two peaks at

1322

Can. J. Chem. Vol. 83, 2005

δ 3.27 (OCH₃) and 4.32 (ArCH₂OR) with integrations in a

3:2 ratio, in addition to those corresponding to 27 and 28

(≥90%), suggesting the formation (5%–10%) of methyl ether

29.

Φ = 3.00 × 10^–3× ∆A/(ε_λ× I × ∆t)

where ε_λis the extinction coefficient of the product at λ_mon.

,

I is the light intensity determined using potassium ferri-

oxalate actinometry (10), and t is the time of irradiation. The

initial concentration of the substrate was adjusted such that

at 1% conversion, ∆A > 0.3 and the absorbance of the solu-

tion at 280 nm >> 2 (typically, >10^–4mol/L). The conversion

was kept at ca. 1% for all runs. Finally, the sample was irra-

diated for a longer time to enhance the conversion to ~15%

Photolysis of 4-[4′-(hydroxymethyl)phenyl]phenol (11) in

CH₃OH–H₂O (1:1)

A solution of 11 (40 mg) in CH₃OH–H₂O (1:1, 100 mL)

was photolyzed (254 nm) in a quartz tube for 8 min using

the general procedure. The solution was then worked up

(CH₂Cl₂, 3 × 150 mL) using the general procedure. The resi-

due contained 16% of 4-[4′-(methoxymethyl)phenyl]phenol).

Pure 4-[4′-(methoxymethyl)phenyl]phenol was isolated by

preparative TLC (silica, hexanes–CH₂Cl₂(1:1), first band

collected) as a colourless oil. ¹H NMR (250 MHz,

(CD₃)₂CO) δ: 3.33 (s, 3H, OCH₃), 4.43 (s, 2H, CH₂), 6.87–

6.95 (m, 2H), 7.31–7.40 (m, 2H), 7.40–7.58 (m, 4H), 8.45

(s, 1H, exchangeable with D₂O, ArOH).

1

and the conversion rechecked by H NMR. The results from

¹H NMR were consistent with that measured by UV spectro-

photometry. The quantum yields of other solvents were mea-

sured relative to that obtained in CH₃CN–H₂O (1:1) using

this method. The ε_λat λ_mon.are essentially independent of

the solvents being studied. The λ_maxand ε_λdata for all

chromene products are summarized in Table 2.

General procedure for determination of photosolvolysis

quantum yields for 7–8 and 11

Laser flash photolysis

A

solution of the substrate (100 mL, typically

Experiments were carried out at the University of Victoria

LFP Facility at 20 2 °C. A Spectra-Physics GCR 12 YAG

laser at 266 nm (<30 mJ) was used for excitation and signals

were digitized with a Tektronix TDS 520 recorder. Samples

of OD ≈ 0.3 at 266 nm were prepared and irradiated in

quartz cells. Generally, flow solutions were used to avoid the

accumulation of any photoproduct. Samples were continu-

ously purged with a stream of N₂or O₂.

4.00 mmol/L in CH₃OH–H₂O (1:1)) and a sample of a refer-

ence compound were irradiated in the Rayonet reactor for a set

time and worked up using the general procedure. The forma-

tion of the methyl ether products (15%–20% conversion)

were determined by ¹H NMR and GC. The photometh-

anolysis of two reference compounds, 2,6-dimethoxybenzyl

alcohol (Φ = 0.31) and 2-hydroxybenxyl alcohol (Φ = 0.24),

were employed (17).

Fluorescence lifetime measurements

Acknowledgments

All fluorescence lifetimes were measured on a PTI LS-1

instrument using the time-correlated single photon counting

(SPC) technique (typically, using 10 000 counts). Samples

were saturated with Ar and the absorbance at the excitation

wavelength was adjusted to ~0.1. The lifetime was extracted

from the SPC decay trace using the iterative reconvolution

program that came with the instrument.

This research was supported by the Natural Sciences and

Engineering Research Council of Canada (NSERC). We

thank Dr. C.-G. Huang for initial studies of some of the

biaryl systems.

References

Steady-state fluorescence and fluorescence quantum

yield measurements

1. H.E. Zimmerman and D.S. Crumrine. J. Am. Chem. Soc. 94,

498 (1972).

All fluorescence spectra were recorded on a PTI A-1010

fluorimeter at rt. Samples were degassed by purging with ar-

gon gas prior to measurement. The concentrations of all

samples were kept at ≤10^–4mol/L (OD ≈ 0.05). In determin-

ing the absolute fluorescence quantum yields, two secondary

standards, naphthalene (in C₆H₁₂) and 2-aminopyridine (in

0.1 N H₂SO₄), were used.

2. A. Imamura and R. Hoffmann. J. Am. Chem. Soc. 90, 5379

(1968).

3. (a) A. Padwa, C. Doubleday, and A. Mazzu. J. Org. Chem. 42,

3271 (1977); (b) S. Lazare, R. Lapouyade, and R. Bonneau. J.

Am. Chem. Soc. 107, 6604 (1985); (c) E.C. Taylor, B. Furth,

and M. Pfau. J. Am. Chem. Soc. 87, 1400 (1965); (d) K.

Mislow and A.J. Gordon. J. Am. Chem. Soc. 85, 3521 (1963);

(e) B.S. Udayakumar and G.B. Schuster. J. Org. Chem. 57,

348 (1992); ( f ) J.S. Swenton, T.J. Ikeler, and G.L. Smyser. J.

Org. Chem. 38, 1157 (1973).

Photocyclization quantum yields for 1, 4–6, and 10 in

CH₃CN–H₂O (1:1)

4. V. Balzani and F. Scandola. Supramolecular photochemistry.

The quantum yield for the chromene (or hydroxyfluorene)

product formation was determined using UV spectrophoto-

metry. Generally, a quartz cuvette containing 3.00 mL solu-

tion of the biaryl alcohol substrate was purged with Ar for

~5 min. and then irradiated for a set time (1~10 min) on the

optical bench (λ_ex= 280 nm, slit 10 nm). The change of

absorbance of the solution before and after irradiation (∆A_λ)

at the monitoring wavelength (λ_mon.) was measured. The

quantum yield (Φ) was then calculated using the following

equation:

Ellis Horwood, New York. 1991. Chap. 7.

5. (a) C.-G. Huang, K.A. Beveridge, and P. Wan. J. Am. Chem.

Soc. 113, 7676 (1991); (b) Y. Shi and P. Wan. J. Chem. Soc.

Chem. Commun. 1217 (1995); (c) Y. Shi and P. Wan. J. Chem.

Soc. Chem. Commun. 273 (1997).

6. (a) G. Bringmann, T. Hartung, L. Göbel, O. Schupp, C.L.J.

Ewers, B. Schöner, R. Zagst, K. Peters, H.G. von Schnering,

and C. Burschka. Liebigs Ann. Chem. 225 (1992); (b) G.

Bringmann, B. Schöner, O. Schupp, K. Peters, E.-M. Peters,

and H.G. von Schnering. Liebigs Ann. Chem. 91 (1994).

Shi and Wan

1323

7. (a) A. Suzuki. Pure Appl. Chem. 63, 419 (1991); (b) A.R.

Martin and Y.H. Yang. Acta Chem. Scand. 47, 221 (1993).

8. C.M. Unrau, M.G. Campbell, and V. Snieckus. Tetrahedron

Lett. 2773 (1992).

9. R.M.G. Roberts. Magn. Reson. Chem. 23, 52 (1985).

10. C. Hatchard and C. Parker. Proc. R. Soc. London Ser. A, 235,

518 (1956).

13. I. Berlman. J. Chem. Phys. 52, 5616 (1970).

14. K.S. Burnham and G.B. Schuster. J. Am. Chem. Soc. 120,

12619 (1998).

15. R.A. McClelland. Tetrahedron, 52, 6823 (1996).

16. G. Bringmann, T. Hartung, L. Göbel, O. Schupp, K. Peters,

and H. Schering. Liebigs Ann. Chem. 769 (1992).

17. P. Wan and B. Chak. J. Chem. Soc. Perkin Trans. 2, 1751

(1986).

11. I. Berlman. Handbook of fluorescence spectra of aromatic

molecules. Academic Press, New York. 1971.

12. T.C. Werner. Modern fluorescence spectroscopy. Vol. 2. Edited

by E.L. Wehry. Plenum Press, New York. Chap. 7.

Article Doi

Solvolysis and ring closure of quinone methides photogenerated from biaryl systems

DOI: 10.1139/v05-138

Source and publish data:

Authors:

Article abstract of DOI:10.1139/v05-138

Full text of DOI:10.1139/v05-138

Products guided by the article

R&D Labs maybe for 195-25-5

Relevant to this article

Hot Product