A time-resolved spectroscopic study of the bichromophoric phototrigger 3',5'-dimethoxybenzoin diethyl phosphate: Interaction between the two chromophores determines the reaction pathway

DOI: 10.1002/chem.200902891

A Time-Resolved Spectroscopic Study of the Bichromophoric Phototrigger

3

’,5’-Dimethoxybenzoin Diethyl Phosphate: Interaction Between the Two

Chromophores Determines the AHCUTNGRNENUGR eaction Pathway

[

a]

[b]

[a]

Chensheng Ma, Wai Ming Kwok, Hui-Ying An, Xiangguo Guan,

[

a]

[a]

Michael Yunyi Fu, Patrick H. Toy, and David Lee Phillips*

Abstract:

3’,5’-Dimethoxybenzoin

out for the reference compounds dieth-

yl phosphate acetophenone (DPAP),

and 3’,5’-dimethoxybenzylic diethyl

phosphate (DMBnDP) in the same sol-

vent. Comparison of the fs-TA spectra

reveals that the photoexcited DMBDP

exhibits distinctly different spectral

character and dynamic evolution from

those of the reference compounds. This

fact, combined with the related steady-

state spectral and density functional

theoretical results, strongly suggests the

presence in DMBDP of a significant in-

teraction between the two sub-chromo-

phores, and that this interaction plays a

governing role in determining the

nature of the photoexcitation and the

reaction channel of the subsequent

cyclization mechanism for DMBDP in

(

DMB) is a bichromophoric system

CH CN. By monitoring the direct gen-

3

that has widespread application as a

highly efficient photoremovable pro-

tecting group (PRPG) for the release

of diverse functional groups. The pho-

todeprotection of DMB phototriggers

is remarkably clean, and is accompa-

nied by the formation of a biologically

benign cyclization product, 3’,5’-dime-

thoxybenzofuran (DMBF). The under-

lying mechanism of the DMB depro-

tection and cyclization has, however,

until now remained unclear. Femtosec-

ond transient absorption (fs-TA) spec-

troscopy and nanosecond time-resolved

eration of the transient DMBF prod-

uct, the cyclization time constant was

determined unequivocally to be ꢀ1 ns.

This indicates that there is little rele-

vance for the long-lived intermediates

(>10 ns) in giving the DMBF product,

and excludes the stepwise mechanism

proposed in the literature as the major

pathway for the DMB cyclization reac-

tion. This work provides important new

insights into the origin of the 3’,5’-di-

methoxy substitution effect for the

DMB photodeprotection. It also helps

to clarify the many different views pre-

sented in previous mechanistic studies

of the DMB PRPGs. In addition to

this, our fs-TA results on the reference

3

resonance Raman (ns-TR ) spectrosco-

py were employed to detect the transi-

ent species directly, and examine the

dynamic transformations involved in

the primary photoreactions for DMB

diethyl phosphate (DMBDP) in aceto-

photophysical

and

photochemical

3

transformations. The ns-TR results

and their correlation with the fs-TA

spectra and dynamics provide evidence

for a novel concerted deprotection–

compound DMBnDP in CH CN pro-

3

vide the first direct observation (to the

best of our knowledge) showing the

predominance of a prompt (ꢀ2 ps)

heterolytic bond cleavage after photo-

excitation of meta-methoxybenzylic

compounds. This provides insight into

the long-term controversies about the

photoinitiated dissociation mode of re-

lated substituted benzylic compounds.

nitrile (CH CN). To assess the elec-

3

tronic character and the role played by

the individual sub-chromophore, that

is, the benzoyl, and the di-meta-me-

thoxybenzylic

DMBDP deprotection, comparative fs-

TA measurements were also carried

Keywords: chromophores · dime-

thoxybenzoin

·

photochemistry

·

moieties,

for

the

Raman spectroscopy

solved spectroscopy

·

time-re-

[

a] Dr. C. Ma, H.-Y. An, Dr. X. Guan, M. Y. Fu, Dr. P. H. Toy,

Prof. D. L. Phillips

Department of Chemistry, The University of Hong Kong

Pokfulam Road, Hong Kong (P. R. China)

Fax : (+852)2857-1586

[b] Dr. W. M. Kwok

Department of Applied Biology and Chemical Technology

The Hong Kong Polytechnic University

Hung Hom, Kowloon, Hong Kong (P. R. China)

Supporting information for this article is available on the WWW

under http://dx.doi.org/10.1002/chem.200902891.

E-mail: phillips@hkucc.hku.hk

5102

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2010, 16, 5102 – 5118

FULL PAPER

Introduction

DMB systems. Direct time-resolved transient absorption

[2a,13a]

3 [13a]

(TA)

and time-resolved resonance Raman (TR )

Because of the very efficient and biologically benign photo-

liberation of protected functional groups, 3’,5’-dimethoxy-

benzoin (DMB) systems have been developed extensively as

a photoremovable protecting group (PRPG) for many func-

studies of a parent benzoin-based PRPG compound have

provided solid pieces of evidence for an apparently single-

step deprotection–cyclization reaction mediated by a benzo-

yl-localized np* triplet with ꢀ11 ns time constant in organic

[1,2,3a]

tional groups, such as inorganic phosphates,

nucleoti-

solvents like acetonitrile (CH CN) and benzene. This agrees

3

[

2b,4]

[5–7]

[3b,8]

[3c]

des,

carboxylates,

amines,

and alcohols.

DMB

with the indirect results of facile quenching of the benzoin

[9]

[2,3a,7]

has also found applications as a PRPG in drug delivery,

deprotection by conventional triplet quenchers.

In con-

[

10]

[3d]

muscle relaxation studies,

chip fabrication,

lithographic synthesis,

bio-

trast to this, the DMB deprotection is not quenched by trip-

let quenchers even in a neat quencher such as piperylene.

Based on this evidence, diverse singlet character pathways

proceeding via various reactive intermediates (Scheme 1)

have been proposed in the literature, and the triplet state of

such as the DMB ester, which was found to be long-lived

[3c,e]

[11]

protein folding and unfolding, and so

forth. These practical applications of the DMB PRPG fol-

lowed from pioneering studies by Sheehan and co-workers

on benzoin esters and substituted analogues, which demon-

strated the potential utilization of these compounds for the

[7]

ꢁ6

photorelease of carboxylic acid. These and many other

later studies on related benzoinyl derivatives have found a

complex dependence of the photoproduct formation on a

number of factors, including the character of the substitu-

ent(s) in the aromatic rings, the nature of the leaving group,

and the properties of the solvent in which the reaction takes

(>10 s lifetime) and having absorptions at ꢀ340 nm and

ꢀ420 nm, has been considered to be nonreactive towards

[5,12]

deprotection and cyclization.

In general, three types of intermediates, including two cat-

ions (the cyclic cation 1 and carbocation 2) and a biradical

species 3, have been considered to be important in previous

mechanistic studies of the DMB photodeprotection and cyc-

lization reactions. Cation 1 is speculated to be produced by

either 1) a CꢁX bond rupture accompanied by a ring-open-

[1–8,12,13]

place, and so forth.

Among the broad range of photo-

coupling and addition reactions observed in many studies,

the photoelimination along with cyclization of the benzoinyl

group into the corresponding benzo[b]furan (BF) as an inert

byproduct [Eq. (1)] constitutes one of the principal reactions

for the photodeprotection of benzoin-based PRPG com-

pounds. The efficiency of this reaction depends strongly on

the substitution pattern on the benzylic ring, and when a

twofold meta-methoxybenzylic substitution is employed, as

in the DMB system, the photodeprotection and cyclization

reactions become very clean and highly efficient. For in-

stance, the photocyclization yield of DMB acetate is report-

ing of a strained bicyclic oxetane intermediate, which results

from an intramolecular Paterno–Bꢁchi cycloaddition

[7]

(Scheme 1, pathway a), or 2) a CꢁX heterolytic cleavage

of an intramolecular exciplex formed by charge transfer

(CT) from the dimethoxybenzylic ring to the carbonyl np*

[5]

singlet state (pathway b). In this case, the cyclization prod-

uct, 2-phenyl-5,7-dimethoxybenzofuran (DMBF), was con-

+

sidered to be formed from a subsequent H elimination of

the cation 1 intermediate. Results from a nanosecond laser

flash photolysis (LFP) study by Wan and co-workers on a

series of DMB esters led them to propose intermediate 1 to

[7a]

ed to be ꢀ0.64,

(ꢀ0.26) estimated for the corresponding unsubstituted ben-

zoin acetate under similar deprotection conditions. This

which is more than twice the yield

ACHTUNGTRENNUNG

[2a]

be assigned to a transient species with a l

absorption at

max

[5]

remarkable substituent effect suggests that DMB systems

can be significantly superior compared to the nonsubstituted

benzoin systems, and this accounts for the widespread appli-

cation of DMB systems for use as PRPG compounds. How-

ever, the underlying reasons for this substantial substituent

effect remain elusive, and there have been differing views

on the reaction kinetics and mechanism(s) of the DMB pho-

todeprotection.

ꢀ480 nm.

From the observation of a time-resolution

(10 ns) limited formation and the ꢀ1 ms lifetime of this spe-

cies, it was suggested that a stepwise mechanism, with a

quick deprotection (rate >10 s ) followed by a relatively

slow DMBF formation (rate ꢀ10 s ) for DMB esters in

8

ꢁ1

6

ꢁ1

[5]

CH CN, may occur.

3

The cation 2 intermediate, formed either by 1) a direct

heterolytic cleavage of the CꢁX bond or 2) a homolytic

cleavage followed by an electron transfer in the radical pair,

has been taken as the most likely intermediate to mediate

not only the deprotection and cyclization processes in non-

polar or polar aprotic solvents, but also to account for the

additional solvolysis reaction that takes place in water or

[3a]

largely water-containing solvents.

The involvement of a

homolytic character for the CꢁX cleavage is disfavored by

the absence of decarboxylation products observed after pho-

[5]

tolysis of DMB pivalate. This lack of decarboxylation

products, together with results from recent fs-TA kinetic ex-

periments carried out by Pirrung, Simon, and co-workers,

led to the proposal of an alternative stepwise mechanism

(Scheme 1, pathway c) with an ultrafast excited-state hetero-

Previous mechanistic studies have revealed a clear de-

scription of the photodeprotection pathways for the parent

[2a,13a]

benzoin system,

but the picture is not clear for the

Chem. Eur. J. 2010, 16, 5102 – 5118

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

5103

D. L. Phillips et al.

Scheme 1. Deprotection and cyclization pathways proposed in previous mechanistic studies of DMB compounds.

1

3

ꢁ1

lytic deprotection cleavage (rate >10 s ), giving inter-

phore, Bz or DMBn, plays the predominate role in the excit-

ed-state processes. This is because the locally excited pro-

cess of both chromophores has been expected to display dis-

tinct excited states or transient species with different reac-

tivities towards deprotection and cyclization. For example,

the precursors of cation 1 (Scheme 1) suggested in path-

1

ꢁ1

mediate 2 as a precursor (with a rate of ꢀ10 s ) to the

cyclic 1 intermediate, which then undergoes deprotonation

[3f]

to yield DMBF.

Distinct from the above-mentioned pathways a, b, and c

in Scheme 1, Chan and co-workers suggested that a biradical

intermediate 3 is formed from a radical-based addition of

the excited carbonyl oxygen to the substituted benzylic ring,

to be the first reactive intermediate in the photodeprotec-

tion of a water soluble DMB acetate. Upon the formation

of intermediate 3, acetoxy migration, followed by re-aroma-

tization or nucleophilic attack by water at the benzylic

AHCTUNGTRENNUNGw ays a (oxetane), b (exciplex), and d/e (biradical) were all

based on a Bz-localized excitation and the consideration of

a mainly radical-associated reactivity of the singlet np* ex-

[6]

[5,6,7,12]

cited state.

On the other hand, the attribution of

cation 2 in pathway c rests on the primary role played by an

excited state of the DMBn moiety, in that the 3,5-dimethoxy

substitution on the benzylic ring can greatly enhance hetero-

lytic cleavage at the benzylic carbon (e.g., through Zimmer-

manꢂs meta effect), leading to rapid formation of the associ-

carbon accounts for the cyclization (pathway d) in CH CN

3

or the modest extent of extra solvolysis in water, respective-

ly. A more recent fs-TA study by Wirz and co-workers on

DMB acetate and fluoride led to another proposed mecha-

nism (pathway e), which suggests that intermediate 3 is a

precursor to cation 1, with a ꢀ2–3 ns deprotection time con-

[3a,f,8]

ated cation.

There are, however, difficulties for both kinds of basic

conjectures to fully justify the very high efficiency of the

DMB deprotection and reactivity of the involved excited

state. For instance, in the case of the Bz-dominated excited

reaction, it is known that electronic excitation of the Bz

moiety typically causes rapid intersystem crossing (ISC),

with time constants of ꢀ2–3 ps and giving a close to unit

[

12]

stant for the two compounds in CH CN.

Within this

3

model, similarly to pathways a, b, and c, the cyclization pro-

ceeds by the subsequent deprotonation of cation 1. In this

work, intermediate 3 was attributed to an fs-TA band that

has l_m_a_xat ꢀ360 nm and is formed with a time constant of

[15–20]

ꢀ

20 ps.

yield of a triplet state.

Therefore, it is unclear 1) how a

DMB is a bichromophoric system containing a benzoyl

Bz) chromophore and a 3’,5’-dimethoxybenzylic (DMBn)

ꢀ20 ps process, like that ascribed to give the biradical 3 in-

termediate, can compete efficiently with the ISC process to

account for the high yield of the DMBF product, and

2) why the derived triplet is nonreactive (triplet energy

transfer from Bz to DMBn is forbidden from an energy con-

sideration), whereas the Bz-localized np* triplet, as revealed

in the parent benzoin system, has been proven to be very re-

(

chromophore. Among the previously proposed pathways,

the cation 1 species appears to be a common intermediate

attributed to the transient product of the heterolytic depro-

tection, and at the same time the immediate precursor to

the formation of the cyclization product DMBF. The differ-

ences in the various proposed assignments of the penulti-

mate precursor that leads to intermediate 1 originate from

different views on the fundamental issue of which chromo-

8

ꢁ1

active in giving rapid (ꢀ10 s rate) deprotection–cycliza-

tion. As for the mechanisms associated with the DMBn

chromophore, considering that the DMB photodeprotection

[13a]

5104

www.chemeurj.org

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2010, 16, 5102 – 5118

Bichromophoric Phototrigger 3’,5’-Dimethoxybenzoin Diethyl Phosphate

depends only slightly on excitation wavelength (l_e_x_c) and

FULL PAPER

tained for DMBDP with those for DPAP and DMBnDP

indicates that electronic interaction between the Bz and

DMBn moieties plays a decisive role from the moment of

photoexcitation, leading to a peculiar series of transient

states and intermediates that constitute a distinct pathway

for DMBDP deprotection and cyclization compared to sys-

tems in which this interaction is not present, such as DPAP,

DMBnDP, and the parent benzoin-caged diethyl phosphate

(BDP).

[3,5,7,8]

shows a better efficiency with l_e_x_c>300 nm,

it is not

clear how the meta effect can operate efficiently: within the

picture of localized photoexcitation, apparently only the Bz

group could possibly be excited electronically at these wave-

[

5,7,8,12]

lengths,

chromophore to the DMBn chromphore is energetically for-

bidden.

and the singlet energy transfer from the Bz

[5,21]

Besides this, the previous mechanisms also have

difficulties in their kinetic aspects in accommodating the ob-

served subtle dependence of the DMB deprotection and

cyclization efficiency on the leaving group and solvent polar-

[

3a,f,5,8,12]

ity (except water).

For example, according to the

Results

ꢁ

latest suggested pathway e, the heterolytic X elimination

from radical 3 was assigned as a key reaction step in

Steady-state study: Figure 1 shows the change in the UV

[12]

CH CN. However, this would suggest that the deprotec-

spectra of DMBDP upon 266 nm photolysis in CH CN re-

3

tion rate, and hence the efficiency of the deprotection,

would be substantially affected by the leaving group lability

corded after a series of defined periods of irradiation. Clear-

[15a,b,22]

and solvent properties,

which is not confirmed by ex-

perimental data. Additionally, the unusually high deprotec-

13

ꢁ1

[3f]

tion rate (ꢀ10 s ) proposed for pathway c has had little

explicit evidence for similar extremely fast and efficient

analogous reactions in the literature for this kind of mecha-

nism. The preceding uncertainties suggest that more experi-

mental characterization of the reaction intermediates and

their properties is needed to understand the DMB deprotec-

tion and cyclization pathway more fully.

3

In this paper, a combined fs-TA and ns-TR study on a

typical DMB-caged compound, that is, DMB diethyl phos-

phate (DMBDP), in CH CN is reported. To elucidate the

3

origin of the substitution effect and to explore the electronic

character and reactivity of the DMBDP excited state (such

as how and at which stage this may be influenced by interac-

tion between the Bz and DMBz moieties), we have also per-

formed a parallel fs-TA study on two model compounds, di-

ethyl phosphate acetophenone (DPAP) and 3’,5’dimethoxy-

benzylic diethyl phosphate (DMBnDP).

Figure 1. Change in the UV spectrum of DMBDP in CH

creasing 266 nm photolysis recorded prior to irradiation (0 s), and after 2,

, 6, 8, 10, 12, and 16 s of exposure to laser irradiation.

3

CN with in-

4

ly, with an increase of the photolysis time, the DMBDP ab-

sorption (the 0 s time spectrum in Figure 1) transforms grad-

ually into a different spectrum (the 16 s time spectrum in

Figure 1), which is characteristic of an absorption spectrum

[3,7,8]

due to the DMBF photoproduct.

The smooth spectral

transformation is accompanied by the appearance of three

well-defined isosbestic points at ꢀ208, 234, and ꢀ262 nm,

indicating that there is direct photoinduced conversion of

DMBDP into DMBF, as described in Equation (1). Using

the absorption extinction coefficient reported for DMBF in

ꢁ

1

ꢁ1 [3a]

The DPAP and DMBnDP compounds were synthesized

to serve as reference systems to provide a spectral and dy-

namical characterization of the excited-state processes initi-

ated by the individual electronic excitation of the Bz and

DMBn chromophores, respectively. These time-resolved

studies, combined with steady-state photochemistry meas-

urements and related theoretical calculations, provide evi-

dence for a novel concerted deprotection–cyclization mecha-

nism, which bypasses the commonly referred to intermedi-

ate 1, and which has a time constant for the DMBDP cycli-

zation determined as ꢀ1 ns. Comparison of the results ob-

ethanol (e₃₁₀_n_m=27480m cm )

value in CH CN, the quantum yield of DMBF formation

and assuming the same

3

[13c]

was estimated to be ꢀ0.62ꢂ0.1,

close to the values re-

[7a]

ported in the literature for DMB acetate.

To better understand the influence of the dimethoxy sub-

stitution, and to assess the possible interaction between the

Bz and DMBn moieties in the Franck–Condon region of

photoexcitation, Figure 2 displays a comparison of the

steady-state UV absorption spectra of DMBDP, DMBnDP,

and DPAP in CH CN. As expected, the DPAP spectrum

closely resembles the reported absorption spectrum of

3

Chem. Eur. J. 2010, 16, 5102 – 5118

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

5105

D. L. Phillips et al.

possible that the absorptions may arise from electronic tran-

sitions with noticeably different character, and are likely to

be due to excitations that involve both the Bz and DMBn

moieties. Indeed, our TDDFT computational results ob-

tained from the optimized structure of the ground S state

0

of DMBDP (see Figure 1S and Table 1S in the Supporting

Information) indicate that the absorption of the two lower

energy bands are due to electronic transitions correlated

with both the Bz and DMBn moieties. For example, the

ꢀ

270–300 nm absorption arises from localized pp* transi-

tions at either the Bz or DMBn rings, with sizeable contri-

butions due to an interchromophore pp* transition from the

DMBn to Bz moiety. For the 300–350 nm band, besides the

Bz-localized np* transition, participation also comes from

extensive delocalized transitions from the bonding orbital of

DMBn to the antibonding orbital of Bz and the carbonyl

group. The involvement of these CT types of transitions in

the FC excitation, which is not seen clearly in many other a-

Figure 2. UV absorption spectra of DMBDP (solid line), DPAP (dashed

line), and DMBnDP (dotted line), in CH CN. The inset shows enlarged

spectra for the weak absorption in the ꢀ270 to 350 nm wavelength range.

3

[

3a,16,19,20,23]

AP,

indicating that light absorption is accounted

[1,2,13,15,24,26,27]

for by the common Bz chromophore, and that an analogous

pattern of electronic transitions is involved in the photoexci-

tation of DPAP and AP. From the known spectral attribu-

keto phototriggers,

may present an important

aspect for the ensuing unique photodeprotection properties

of the DMB PRPG. This is verified by the fs-TA and ns-TR

measurements described in the rest of the paper.

3

[16,19,20,23]

tion of AP and related aromatic carbonyls,

the

strong DPAP absorption band of l ꢀ240 nm (eꢀ1.3ꢃ

max

4

ꢁ1

1

0 m cm ) can be ascribed to the allowed pp* transition

Femtosecond transient absorption (fs-TA) spectroscopy: To

help unravel the excited-state processes for the DMBDP

photodeprotection and to elucidate the roles played by the

Bz and DMBn moieties in the DMBDP photodeprotection,

comparative fs-TA measurements were carried out for

(

L , S ) of the phenyl ring, the relatively weaker band at

a 3

3

ꢁ1

ꢀ

260–300 nm (e₂₈₀_n_mꢀ1.1ꢃ10 m cm ) to the less-allowed

ring pp* (L , S ) transition, and the very weak but character-

b

2

ꢁ1

istic absorption at ꢀ300–350 nm (e₃₁₀_n_mꢀ80m cm , see

Figure 1, inset) to the carbonyl np* (S ) transition. The

DMBDP, DPAP, and DMBnDP in CH CN. The results ob-

1

3

DMBnDP spectrum is typical of the absorption spectra of

tained for DPAP and DMBnDP are presented first, because

they provide necessary reference information for the inter-

pretation of the DMBDP excited states and for assessment

of the various earlier proposed deprotection pathways. In

addition, the significance of the excited-state properties, re-

activity, and dynamics of these two compounds by them-

selves renders these femtosecond time-resolved data, which

have not been available in the literature, interesting in their

own right. For example, DPAP can be regarded as a phenac-

[21,25]

benzene and related Bn compounds.

However, com-

pared to the corresponding bands of these Bn derivatives,

the lowest energy pp* absorption in DMBnDP (l

max

3

ꢁ1

ꢀ

280 nm, eꢀ1.4ꢃ10 m cm ) is redshifted significantly,

and has a marked increased in its oscillator strength. This

agrees with the general trend of the spectral changes in this

[21]

pp* absorption as induced by the ring substitution, and

reflects that the ring 3,5-substitution by the strongly elec-

tron-donating methoxy groups leads to a substantial energy

stabilization of the correlated excited state relative to the

[1,24]

yl-caged compound.

For this type of PRPG, a previous

LFP study on the ns–ms timescale revealed a triplet hydro-

ground state (S ). It is noted that the transition energy of

gen abstraction pathway for the photodeprotection of the

0

[24]

this DMBnDP absorption is comparable to the L pp* ab-

phenacyl acetate. However, the earlier stage excited states

that exist before several nanoseconds have remained un-

characterized. DMBnDP, on the other hand, provides a

b

sorption in DPAP. However, DMBnDP shows no discernible

absorption at wavelengths >300 nm. From the above results

and the fact that the diethyl phosphate (DP) leaving group

absorbs only in the far UV region, the spectra of DPAP and

DMBnDP can each be taken as benchmark absorption spec-

tra for the Bz and DMBn chromophores, respectively.

[

14,28,29]

good example of a DMBn-caged compound.

pioneering work of Zimmerman

Since the

[14,29]

in showing the meta-

effect and the remarkable facile photosolvolysis reactivity

for di-meta-methoxy-substituted benzylic systems, the

DMBn chromophore has been frequently employed in the

design of practical PRPGs, and this has attracted much in-

terest in the excited-state behavior of DMBn com-

The UV/Vis absorption of DMBDP at wavelengths

>

240 nm appears to be similar to that of DPAP. However,

close examination shows that, besides the strong ꢀ240 nm

band, which appears to be fairly comparable in the two com-

pounds, the two lower energy bands of DMBDP at the

[28,30]

pounds.

Although many studies have provided valuable

information for the excited-state dissociation of DMBn sys-

[14,28–30]

ꢀ

270–300 and 300–350 nm regions clearly display stronger

tems,

it has remained unclear as to the fundamental

oscillator strengths than their DPAP counterparts. As the

DMBDP spectrum cannot be decomposed into the individu-

al absorptions due to the Bz and DMBn chromophores, it is

mode, homolytic versus heterolytic breaking, of the bond

cleavage; the rate of the bond breaking has also remained

unknown. Our results here for DMBnDP provide, to the

5106

www.chemeurj.org

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2010, 16, 5102 – 5118

Bichromophoric Phototrigger 3’,5’-Dimethoxybenzoin Diethyl Phosphate

FULL PAPER

best of our knowledge, the first direct determination of

these parameters, which are of general interest for DMBn

systems.

in a ꢀ2.5 ps time constant for the rising component ob-

served at early picosecond times.

The ꢀ350 nm absorption spectrum closely resembles the

[17,18]

triplet absorption spectrum of AP in the same solvent,

Femtosecond TA measurements for DPAP: Figure 3 shows

representative fs-TA spectra recorded for DPAP at time

delays of up to 5000 ps after 267 nm excitation, which direct-

and is also consistent with the triplet absorption of phenacyl

group compounds, as reported in previous LFP studies on

[24]

phenacyl esters.

Therefore, this spectrum can be confi-

dently attributed to the Bz-localized triplet (T ) state of

1

DPAP. From this, it is straightforward to attribute the

ꢀ

2.5 ps process to an ISC that leads to rapid conversion of

the system from the initially photoexcited singlet to a triplet

state. The rapid ISC agrees with the high triplet yield gener-

[15–20]

ally observed for related aromatic carbonyls.

Given the

extremely rapid nature of the internal conversions (IC)

from S to S , we tentatively attribute the very early (0.2 ps)

3

1

spectrum to the S₁np* state. The persistence of the

350 nm triplet spectrum at late picosecond times is consis-

ꢀ

tent with the lower reactivity of the T state in the inert

1

CH CN solvent (in terms of the hydrogen abstraction reac-

3

tion), in which the triplet decay is determined mainly by a

[15,24,31]

ns–ms diffusion-controlled oxygen quenching process.

The TA results for DPAP reveal a photophysical transfor-

mation [Eq. (2)] to account for the early time deactivation

associated with the Bz-localized photoexcitation. The data

provide characteristic absorption spectra for the Bz-local-

ized S (the 0.2 ps spectrum) and T (spectrum of 10 ps or

Figure 3. Fs-transient absorption spectra of DPAP obtained by 267 nm

excitation in CH

3

CN at early picosecond times of 0.2, 0.3, 0.6, 0.8, 1, 1.25,

.75, 2.5, 4.5, 6, 10 ps, and later picosecond times (inset) of 10, 20, 100,

00, 1000, and 5000 ps.

1

5

1

later) states.

ly populates the strong pp* (L ) state of DPAP. The tempo-

a

ral evolution of the spectra displays rapid conversion within

IC;<200fs

ISC;2:5ps

S ! S ! T

ð2Þ

3

1

ꢀ

10 ps from the early spectrum (the 0.2 ps spectrum, ab-

sorbing at < ꢀ340 nm wavelength region) into a spectrum

the 10 ps spectrum) that shows a relatively strong absorp-

Fs-TA measurements for DMBnDP: Figure 5 displays the

temporal evolution of the fs-TA spectra obtained after

67 nm photoexcitation of DMBnDP in CH CN. The

67 nm excitation takes the population of the system to the

pp* excited state, and the initial 0.2 ps spectrum can be as-

cribed to the S pp* state, the spectrum of which is broad

(

tion with its l_m_a_xat ꢀ350 nm. This 350 nm absorption dis-

plays little decay at later times (see Figure 3, inset). The ki-

netics for the spectral changes of the TA spectra were fol-

lowed using the 345 nm wavelength, and Figure 4 displays

the TA time profile at ꢀ345 nm together with a single expo-

nential fitting convolved with the instrumental response

function (IRF) of the TA measurements. The fitting results

2

3

1

^aⁿ^d^a^p^p^a^r^eⁿ^t^l^y^h^a^s^t^w^o^lmax at ꢀ370 nm and 510 nm. It is

clear from Figure 5 that, immediately after photoexcitation,

Figure 4. Time dependence of the DPAP transient absorption spectra in

CH

3

CN (shown in Figure 3) at ꢀ345 nm at early and later (inset) pico-

Figure 5. Fs-transient absorption spectra of DMBnDP obtained after

267 nm excitation in CH CN at delay times of 0.2, 0.4, 0.6, 0.8, 1, 1.2, 1.4,

3

1.6, 2, 2.25, 3, 3.5, 4.5, 7, 8.5, 12, 18, 25, and 50 ps.

second times. The solid lines indicate the kinetics fitting to the experi-

mental data points using an exponential function convolved with the IRF.

Chem. Eur. J. 2010, 16, 5102 – 5118

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

5107

D. L. Phillips et al.

the S spectrum decays quickly into a distinctly different

as a plausible attribution for the observed process and the

associated intermediate, and is supported by several pieces

of evidence, such as the fact that: 1) many benzyl cations

have been found to absorb in the ꢀ410–450 nm wavelength

1

profile (the spectrum of ꢀ5 ps or later), which features a

single absorption with l of ꢀ425 nm; this ꢀ425 nm spec-

max

trum then decays and disappears on the ꢀ50 ps timescale.

[32,33,37]

Figure 6 shows the TA time profiles at the representative

range,

and 2) the importance of the direct excited-

state CꢁX (X=leaving group) heterolytic cleavage has been

implicated in many earlier studies to interpret the noticeably

enhanced yield of ion-derived products observed for di-

meta-methoxy/substituted

phates. Further relevance to this end is that

CASSCF theoretical studies by Zimmerman

benzyl

esters

and

phos-

[14,28,30,38]

[

14]

indicate an

evident energy preference for the S heterolysis relative to

1

homolysis, and suggest the presence of an easily accessible

conical intersection between the S and S potential surfaces

1

0

to offer an efficient avenue for the excited-state ion pair to

decay radiationlessly to the S surface. Based on the preced-

0

+

ing results, we ascribe the ꢀ425 nm signal to the DMBn

ion, which is formed by direct CꢁO heterolysis of the S

Figure 6. Normalized time dependences of the DMBnDP transient ab-

1

sorption spectra in CH

425 nm (~).The solid lines indicate kinetics fitting to the experimental

3

CN (shown in Figure 5) at ꢀ370 nm (*) and

state. The observed rapid decay (ꢀ7.7 ps) of the ꢀ425 nm

ꢀ

signal can be readily explained by the high reactivity of the

data points using exponential functions convolved with the IRF.

+

nascent DMBn cation. Upon formation, this kind of cation

is subject to extremely efficient reactions, including the in-

cage ion-pair recombination to return to the original

DMBnDP molecule, and the solvolytic addition reac-

wavelengths of ꢀ370 and ꢀ425 nm. Fitting of this profile by

using single (the ꢀ370 nm data) or double (the ꢀ425 nm

data) exponential function(s) convolved with the IRF gives

a common time constant of ꢀ2.1 ps from both the data sets

[28,30,32,33,38]

tion.

Although the specific partitioning between

these two channels may depend on various factors, the very

high efficiency of the recombination pathway has been dem-

1

8

for the early S decay, and a ꢀ7.7 ps time constant (from

onstrated by O scrambling experiments on closely related

benzylic compounds, as reported by Givens and co-work-

1

the ꢀ425 nm profile) for the later disappearance of the

[38]

ꢀ

425 nm spectrum.

Of the many kinds of substituted benzyl compounds, the

ers. From these results, and given the lower reactivity of

[

39]

the CH CN solvent towards solvolytic addition, the very

3

+

di-meta-methoxy-substituted systems are among those

fast decay of DMBn can be mainly attributed to the ion

ꢁ

3

having the lowest fluorescence quantum yield (f of ꢀ10

pair collapse leading to re-formation of the reactant sub-

strate.

It is worth noting that the results here for DMBnDP in

f

[28,30]

order of magnitude) in polar solvents.

The ꢀ2.1 ps S

1

decay time observed here provides a direct determination of

the S deactivation rate, and explains these low f values.

CH CN represent the first direct spectral and dynamical

1

f

3

This very short S lifetime indicates a very high efficiency

characterization of the singlet excited state and associated

heterolysis for benzylic and ring-substituted benzylic com-

pounds. This observation lends strong experimental support

1

for the nonradiative event(s) that depopulates the S state.

1

Our data above suggest that as the S decays, the system is

1

[14,29]

converted into a short-lived (ꢀ7.7 ps) intermediate that is

to Zimmermanꢂs meta effect,

in that the meta-methoxy

responsible for the ꢀ425 nm spectra.

substitution favors strongly the excited singlet state heterol-

ysis. This also offers an explicit example indicating that the

ionic products observed for DMBnDP, and probably for

To help identify this intermediate and the underlying

pathway leading to the ꢀ425 nm species, we consider three

possible S deactivation channels proposed previously in the

some DMBn esters as well, in solvents such as alcohols and

1

[14,28–30]

[28,30]

literature:

1) ISC to a triplet state, 2) CꢁO bond ho-

water

could also be due to ionic intermediates (the

molysis to give a radical pair DMBnCCDP, and 3) CꢁO bond

cation) that originates from the direct heterolysis pathway,

rather than the indirect pathway of charge transfer preceded

by homolysis. However, given that the mode of the excited-

state bond rupture (homolysis versus heterolysis) can be af-

fected sensitively by diverse factors, such as properties of

the leaving group, the polarity and solvation capability of

+

ꢁ

heterolysis to produce an ion pair DMBn DP. Process 1

appears rather unlikely, because although the triplet of

some benzyl esters have been reported to absorb at

[28c]

ꢀ

400 nm (close to the observed absorption),

the triplet is

expected to have a much longer lifetime in the ns–ms range

[28,30,32,33]

due to its known nonreactivity towards an excited-state

the solvent medium, ring substitution, and so on,

[28–30]

bond cleavage.

Process 2 also seems unlikely, due to the

further study is required to establish whether, or to what

extent, the direct excited-state heterolysis mechanism can be

applied to other related benzyl compounds. Nonetheless, for

fact that the radical species, which has DMBn as a chromo-

phore, is expected to display an absorption in the ꢀ320–

3

40 nm region, similar to absorption spectra observed previ-

the purpose of the work reported here, the spectral and ki-

[32,33,34–36]

+

ously for various benzyl radicals.

Process 3 emerges

netic information acquired for the S and DMBn ion from

1

5108

www.chemeurj.org

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2010, 16, 5102 – 5118

Bichromophoric Phototrigger 3’,5’-Dimethoxybenzoin Diethyl Phosphate

FULL PAPER

DMBnDP provides the needed reference benchmark for a

discussion of the nature and dynamical transformation of

the DMBDP excited state, which is examined in the next

section.

Fs-TA measurement for DM-BDP: The fs-TA spectra re-

corded for DMBDP over a range of time intervals following

excitation at 267 nm reveals the presence of a series of con-

secutive spectral changes in the timescale up to 6 ns. To

show the temporal evolution of the spectra clearly in differ-

ent time regimes, the spectra obtained at several typical

time intervals are displayed separately in Figure 7. A 3D

overview of all the spectra is given in the Supporting Infor-

mation (Figure 2S).

Immediately after photoexcitation (Figure 7a), the spec-

trum shows evolution from a profile featuring a broad ab-

sorption with apparently two l_m_a_xat ꢀ335 and 450 nm (the

0.6 ps spectrum) to a different profile also exhibiting a

broad absorption, but with apparently one l_m_a_xat ꢀ345 nm

(

the 3.5 ps spectrum). This spectral change is accompanied

by three isosbestic points at ꢀ350, 430, and 530 nm, suggest-

ing the occurrence of a state-to-state conversion which is de-

noted temporally as A!B, with A related to the very early

time (<0.6 ps) spectra and B to the somewhat later 3.4 or

4.0 ps spectra. From Figure 7b and c it can be seen that, at

the time interval from 4 to ꢀ50 ps, the B-related spectra

decay, and there is simultaneous growth of a new band with

l_m_a_xat ꢀ370 nm, assumed to arise from a species denoted as

C₃₇₀. From 50 to 1500 ps, the C (the 50 ps spectrum is char-

370

acteristic of the C₃₇₀species) species spectrum decays again,

and at the expense of this decay, two new bands with l_m_a_xat

ꢀ

320 and 480 nm appear due to another species. From their

simultaneous growth, the ꢀ320 and ꢀ480 nm bands can be

considered to arise either from a common species or from

two individual ones. These two bands are denoted as being

due to the D₃₂₀and D₄₈₀species, respectively. The observa-

tion of two sets of isosbestic points associated with the two

spectral transformations in the two different time regimes

(

Figure 7b and c) indicates there are direct dynamic conver-

sions taking place, and mother-to-daughter relationships be-

tween the correlated species. After 1500 ps (Figure 7d), the

C₃₇₀- and D₄₈₀-associated bands both exhibit an intensity de-

^c^r^e^a^s^e^,^b^u^t^wⁱ^t^h^dⁱ^f^f^e^r^eⁿ^t^d^yⁿ^a^mⁱ^c^s^,^w^h^e^r^e^a^s^t^h^e^D³20-related

band shows relatively little change in intensity.

Figure 7. Fs-transient absorption spectra of DMBDP obtained subsequent

to 267 nm excitation in CH CN at delay time intervals from a) 0.6 to

3

.5 ps, b) 4 to 40 ps, c) 50 to 1500 ps, and d) 1500 to 6000 ps.

The kinetics for the above-mentioned spectral transforma-

tions, and possibly other intermediates involved in these

transformations, can be traced by following the TA time

profiles at typical absorption wavelengths for the species of

interest. Figure 8 displays the temporal changes of the TA

spectra at the ꢀ375, ꢀ495, and ꢀ420 nm wavelengths, at

different ranges of time intervals. The two-phase growth ex-

hibited by the ꢀ375 nm profile at early times (ꢃ15 ps, inset

in Figure 8a) reflects the dynamics for the A!B and B!

time TA decay (Figure 8b) is wavelength dependent. The

TA time profile at some wavelengths, such as ꢀ420 and

ꢀ340 nm, and also those associated with the D species at

4

80

ꢀ495 nm, display much slower decays than that of the pro-

file at ꢀ375 nm.

Global multiexponential fittings of the TA profiles at vari-

ous wavelengths including those in Figure 8 (e.g., solid lines

in Figure 8) give time constants of ꢀ1.7, ꢀ14, and ꢀ1050 ps

for the first three transformation steps, respectively. As for

the ensuing later time dynamics, the least square fitting of

C

transformations. The subsequent intensity decay and

70

3

corresponding growth of the ꢀ375 and ꢀ495 nm profiles, re-

spectively, at timescales up to 700 ps (Figure 8a) contain the

dynamics for the C₃₇₀!D₄₈₀(+D₃₂₀) conversion. The later

Chem. Eur. J. 2010, 16, 5102 – 5118

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

5109

D. L. Phillips et al.

from the T species. These two species may display over-

lapped absorptions to a different extent at some other wave-

lengths. Similarly to the case here, the TA spectra at a com-

parable time regime (<50 ps) from the joint contributions

due to the T and a l_m_a_x360 nm species (assigned to the bi-

radical 3 in path e of Scheme 1) have also been reported in

[12]

a fs-TA study on DMB acetate in CH CN.

3

The preceding kinetic analysis can be used to develop a

scheme for the dynamical transformations observed in the

TA spectra, and to account for the excited-state events oc-

curring after photoexcitation of DMBDP [see Eqns. (3)–

(

5)]. Among the various properties of the transient species

included in the scheme, the TA character determined for

the very early stage A and B states and the explicit identifi-

cation of the D₃₂₀species provide important new insights for

evaluating the different deprotection and cyclization path-

ways (Scheme 1) proposed in previous studies.

Figure 8. Normalized time profiles of the DMBDP transient absorption

in CH

CN (Figure 7) at ꢀ375 nm (~), ꢀ495 nm (*), and 420 nm (&).

3

The solid lines are from fitting of the experimental data by multiexpo-

nential functions convolved with the IRF. The inset in a) shows the two-

phase grow-in kinetics as indicated by the 375 nm time profile at early

delay times (see text for details).

As mentioned in the introduction section, the DMBDP

reaction pathways suggested earlier are based on local exci-

tation of either the Bz or DMBn moieties, and governed by

the corresponding transient states/species, that is, mainly the

the ꢀ375 and ꢀ420 nm time profiles requires the inclusion

of an additional fourth component of 010 ns time constant,

with a relatively higher weighted contribution for the

carbonyl localized S state in the case of the Bz moiety pre-

1

dominant process versus the heterolytic cation 2 involved in

the DMBn sub-chromophore predominant process

(Scheme 1). These states or species were speculated in the

literature to be extremely short-lived, and present only at

ꢀ

420 nm wavelength than the ꢀ375 nm profile. In the same

way, the slow decay of the D related profiles can be de-

4

80

[3a,f,5–7,12]

scribed by a time constant of 010 ns.

very early times after photoexcitation.

The validity

It is noted that these long-lived spectral components have

also been observed on a microsecond timescale by previous

of these hypotheses would be demonstrated by the expecta-

tion that the initial and early-stage excited states of photo-

excited DMBDP would display, to a certain extent, a spec-

tral and dynamical character comparable to the excited

states or species associated with either the Bz or the DMBn

chromophore. Our aforementioned TA data on DPAP and

DMBnDP supplies benchmark results for making such an

assessment. Comparing the early time (<5 ps) TA spectra

and the temporal evolution of photoexcited DMBDP (Fig-

ure 7a) with those of DPAP (Figure 3) and DMBnDP

(Figure 5) reveals, however, a distinctly different character

of the DMBDP spectra from either chromophore bench-

mark (DPAP and DMBnDP) or a combination of them. For

[5]

LFP measurements on DMB esters in CH CN. The D₄₈₀

3

associated component evidently corresponds to the l_m_a_xof

4

80 nm (ꢀ1 ms lifetime) LFP absorption, corresponding to

the cation 1 of previous studies (see Scheme 1). The 010 ns

absorption at wavelengths around ꢀ420 and ꢀ340 nm is

consistent with the triplet-state absorption reported for

[5,12]

DMB acetate,

and thus can be associated with the triplet

state (T) of DMBDP. The broad nature of the TA spectra at

early times (Figure 7a–c) together with the rather slow

decay of the T state makes it difficult to track the precise

timing for its formation. However, the spectral evolution

and wavelength-dependent decay in Figure 8b indicates that

the T and C₃₇₀species coexist, and both contribute to the rel-

atively early spectra shown in Figure 7b and c, but with the

l_m_a_x370 nm absorption contributed predominantly from the

C₃₇₀species, and the absorption at ꢀ420 and ꢀ340 nm more

example, the initial Bz-localized S state has an absorption

1

l_m_a_xat <340 nm, and then transforms by ꢀ2.5 ps ISC into

the Bz-localized T featured by the l ꢀ350 nm band, with

1

max

little absorption at wavelengths >450 nm. This contrasts

with the DMBDP results in that the A and B states both ex-

5110

www.chemeurj.org

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2010, 16, 5102 – 5118

Bichromophoric Phototrigger 3’,5’-Dimethoxybenzoin Diethyl Phosphate

FULL PAPER

3

hibit broad absorptions throughout the visible region. The

lack of a general common feature in the spectral profiles is

ly, were carefully chosen to facilitate the ns-TR measure-

ments. The 240 nm pump was selected due to the relatively

strong absorption of the DMBDP reactant compared to the

DMBF product at this wavelength. The 299 nm probe is

close to the l_m_a_xof the DMBF lowest energy absorption,

and is also close to the D₃₂₀absorption observed in the TA

spectra.

The ns-TR spectra obtained with the 240 nm pump and

299 nm probe wavelengths at the indicated delay times are

given in Figure 9. A 299 nm resonance Raman (RR) spec-

also seen clearly when comparing the DMBDP A- and B-

+

state TA spectra with those of the S and DMBn species of

1

DMBnDP. It also appears unlikely that the DMBDP excited

states could be described as an overlapped contribution

from the excited states/species associated with the individual

Bz and DMBn locally excited processes. This is because it

turns out to be difficult to decompose the early DMBDP

spectra into a weighted combination of the corresponding

DPAP and DMBnDP spectra, despite our many attempts to

do so.

3

This observation of little similarity in the early TA spectra

of the DMBDP excited state with those expected by the lo-

cally excited Bz and/or DMBn corroborates the UV/Vis ab-

sorption (Figure 2) and TDDFT results on DMBDP. These

together suggest that, for DMBDP, the electronic interaction

between the two moieties leads not only to the delocalized

S to S absorption transition(s), but also to the involvement

0

n

of discrete excited states and ensuing species, which could

not be accounted for only on the basis of locally excited

chromophores.

As for the D₃₂₀species, we note that our observation of

the simultaneous dynamical growth of the correlated l₃₂₀ab-

sorption with the l₄₈₀absorption (Figure 7c) has not been

reported in the literature to the best of our knowledge. This

could be due to the different time and spectral windows

[3f,5,12]

used in this work from those of previous studies,

and

may also relate to the presence of the promptly formed, but

long-lived, triplet-state absorption at a similar wavelength

[5,12]

region.

Considering the fact that the absorption of the

DMBF final product [Eq. (1)] extends to this wavelength

region and is much stronger than the absorption of the

DMBDP reactant (Figure 1), it turns out to be quite possi-

ble that the l₃₂₀absorption (Figure 7c and d) may arise from

a transient DMBF product resulting from the deprotection

and cyclization of the photoexcited DMBDP. The drastic

TA drop at the blue edge of the l₃₂₀absorption (Figure 7) is

due to the limited spectral detection window of the present

TA measurements. Making use of the ability of resonance

Raman spectroscopy to enhance selectively the Raman

spectra of species that are in resonance with the electronic

Figure 9. Nanosecond time-resolved resonance Raman spectra of

DMBDP in CH

lengths at the indicated delay times. At the top is the resonance Raman

spectrum of DMBF acquired with 299 nm excitation in CH CN.

3

CN obtained with 240 nm/299 nm pump/probe wave-

3

[13,15]

absorption of that species,

our resonance Raman (RR)

trum from a synthesized authentic sample of DMBF ac-

quired under the same conditions is also displayed in

Figure 9, to allow comparison and to help identify the transi-

3

and ns-TR results described below provide unequivocal evi-

dence for the attribution of the D₃₂₀species to the transient

DMBF. This constitutes a key piece of information, which

leads us to a dramatically revised picture for the DMBDP

deprotection and cyclization reaction.

3

ent revealed in the ns-TR spectra. The ns-TR spectra of

DMBDP in Figure 9 show an instant appearance of a series

ꢁ

1

of features (mainly the strong ꢀ1610 cm feature and weak

ꢁ

1

features at ꢀ1570, 1180, and 1000 cm ) with their intensi-

3

Nanosecond time-resolved resonance Raman (ns-TR ) spec-

ties reaching their maxima within the approximate time res-

3

troscopy of DMBDP: The ns-TR measurements were de-

olution of our ns-TR measurement (ꢀ10 ns), and these in-

signed and performed to explore and ascertain the identity

of the D₃₂₀species observed in the TA spectra, as well as to

determine the formation dynamics of the DMBF cyclization

tensities nearly constant after 10 ns (Figure 3S, Supporting

Information). Comparison of these transient Raman spectra

with the DMBF RR spectrum shows clearly that the vibra-

tional features in the former are essentially identical to

those of the latter. This offers explicit evidence for the as-

product for photoexcited DMBDP in CH CN. Using the

3

UV/Vis absorptions of DMBF and DMBDP (Figure 1),

pump and probe wavelengths of 240 and 299 nm, respective-

3

signment of the ns-TR features to the DMBF product.

Chem. Eur. J. 2010, 16, 5102 – 5118

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

5111

D. L. Phillips et al.

Taken together, the data in Figure 9 provide unambiguous

evidence for a rapid yield of the DMBF product with a for-

mation time of <10 ns. The constant ns-TR signal level ob-

the operation and way of operation of this electronic inter-

action could depend critically on the pattern and composi-

tion of the substitution on both the Bz and benzylic phenyl

rings. Extensive previous work on this aspect, based mainly

on product analysis, found consistently that the di-meta-me-

thoxybenzylic substitution represents the optimal chromo-

phore design in offering a clean and highly efficient photo-

3

served on the ns to ms timescale indicates there is little rele-

vance of a long-lived species (>10 ns lifetime) in the DMBF

product formation. To the best of our knowledge, the ns-

3

TR results here provide the first direct time-resolved moni-

[3,7,8]

toring of the cyclization reaction for DMP PRPGs.

deprotection and cyclization reaction.

This may imply

Incorporating these results with the temporal evolution

scheme observed based on the fs-TA data leads us to associ-

ate the TA observed D₃₂₀species with the DMBF product

revealed in the resonance Raman measurements. With this

assignment and the dynamic result in Equation (5), it is con-

sidered that the DMBDP photocyclization takes place from

the C₃₇₀precursor and proceeds with a ꢀ1050 ps time con-

further that the incorporation of this particular pattern of

substitution can induce an appropriate electronic modula-

tion to promote the needed delocalized excitation, and at

the same time facilitate the particular set of excited-state

steps to bring about a highly efficient photoliberation reac-

tion. To substantiate this, it appears relevant to make a com-

parison of the excited-state behavior and photochemical

outcome for DMBDP and the parental benzoin-caged dieth-

yl phosphate (BDP).

9

ꢁ1

stant, corresponding to a rate of ꢀ10 s . The C species

3

70

also serves as a common precursor to give the D₄inter-

80

mediate, a process which occurs competitively with the cycli-

zation reaction. The cyclization pathway revealed here dif-

fers fundamentally from those proposed in the earlier stud-

ies not only in the ordering of reactive intermediates, but

also in the dynamics of the reaction. The previously pro-

posed pathways, although they have diverse descriptions for

the very early time excited state/species, have a general con-

Substitution effect of the benzoinyl chromophore: Previous

[2a,13a]

studies on BDP

show that photolysis of this compound

in CH CN leads to the formation of a Bz-localized S , which

3

1

converts by a ꢀ2–3 ps ISC to a Bz-localized T state, from

1

which then proceeds, with ꢀ11 ns time constant, an appa-

rently single-step phosphate liberation and cyclization to

yield the BF product. It is consistent to this end that the

6

ꢁ1

sensus that the cyclization occurs with a ꢀ10 s rate and

had the l₄species (corresponding to D here) as the pre-

spectral features and the ISC dynamics of the BDP S and

80

480

3f,5–7,12]

1

[

cursor for the cyclization (Scheme 1).

T states closely resemble the corresponding results shown

1

here for DPAP. Unlike the triplet-natured pathway of BDP,

our observation that photolysis of DMBDP leads to the

DMBF product being formed on a timescale (ꢀ1 ns) much

faster than the decay (@10 ns) of its T state provides direct

evidence to exclude the T state as a reactive precursor to

the DMBDP cyclization. This affords an explicit explanation

for the previously reported lack of triplet quenching effect,

and on the other hand implicates a singlet-character channel

for the DMBDP photodeprotection and cyclization, which is

in accordance with most of the earlier proposed mechan-

Discussion

Understanding the nature of the series of photoinitiated

transformations, as revealed in the TA data, and the make-

up of the transient states and intermediates involved, is es-

sential for elucidating the mechanistic details of the

DMBDP photodeprotection and cyclization, and the under-

lying cause of the unusual di-meta-methoxy substitution

effect. Comparison of the steady-state UV/Vis and time-re-

solved fs-TA results obtained for the target DMBDP and

the two reference compounds DPAP and DMBnDP in

[3a,f,5–7,12]

isms

with regard to the excited-state multiplicity of

the DMBDP photoliberation. It is interesting to note that

the determined timing of DMBDP cyclization is more than

one order of magnitude faster than the BDP cyclization.

This may contribute to the much higher efficiency of the

DMBDP compared to the BDP reaction. The nonreactivity

of the DMBDP T, and the low T yield as can be inferred

from the high efficiency of the competing singlet pathway,

contrast sharply with the primary role of the BDP T state in

the BDP photoliberation reaction. The origin of this could

also be the peculiar electronic interaction between the Bz

and the DMBn sub-chromophores in DMBDP, which ren-

ders its triplet having intrinsically different features from

those of the Bz-localized triplet of BDP. Considering the

modest role of the T state in the DMBDP excited-state de-

activation and its minor relevance to the deprotection reac-

tion, this state will not be included in the following sections.

CH CN have revealed spectral and dynamical features of

3

the target molecule that are distinctive from the reference

systems. This indicates that the nature of the photoexcitation

and subsequent excited-state events and dynamics for

DMBDP are intrinsically different from those expected

from local excitation of either the Bz or DMBn sub-chromo-

phore. From this and the related TDDFT absorption, elec-

tronic-transition results, it is reasonable to consider that, for

DMBDP, the electronic communication between the Bz and

DMBn moieties upon and after photoexcitation may play a

decisive role in accounting for the delocalized excitation,

the involvement of the exclusive series of transient states

and species, and, more importantly, the explicit outcome of

the DMBDP photolysis. As implied by the complex and rich

photochemistry exhibited by the diverse substituted ben-

zoinyl compounds in terms of giving substantially varied

Possible assignments and mechanistic pathway for the

DMBDP photodeprotection and cyclization: Although our

[1–3,5–8,12]

yields of the cyclization and some other products,

5112

www.chemeurj.org

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2010, 16, 5102 – 5118

Bichromophoric Phototrigger 3’,5’-Dimethoxybenzoin Diethyl Phosphate

FULL PAPER

fs-TA data obtained for DMBDP provide clear evidence for

the participation of a series of dynamically correlated transi-

ent species in the excited-state deactivation and cleavage re-

action, it appears difficult to make straightforward assign-

matic) ketones bearing an electron-rich a- or b-aryl group

have suggested involvement of an exciplex or a complex

formed by CT from the aryl to the carbonyl to account for

the markedly efficient excited-state quenching and reac-

[42–47]

ments to these intermediates based solely on the broad, and

in some cases overlapped, absorption features. However, re-

ferring to the valuable insight presented in previous DMB

tions.

The observed very rapid nature (ꢀ1.7 ps) of the

S A to exciplex B transformation implies the process is es-

1

sentially a barrierless excited-state conversion, likely to be

associated with small structural changes. It is worth noting

that the timescale for this conversion in DMBDP is close to

that of the DPAP S to T ISC (ꢀ2.5 ps) and the cleavage

[3a,f,5–8,12]

mechanistic studies,

and taking into account the

electronic and spectral features reported in the literature for

some related compounds and intermediates, we consider the

following attributions to each of the observed transient spe-

cies to be rather reasonable, and form a basis to suggest an

alternative mechanistic picture (Scheme 2) for DMBX (X=

1

+

of S to DMBn for DMBnDP (ꢀ2.1 ps). However, the dis-

1

tinctly different character of the spectral profile revealed by

our TA data (Figures 3, 5, and 7) indicates clearly the partic-

ipation of entirely different species for these dynamically

similar processes. This underlines the importance of the TA

broad-band capability in differentiating transient intermedi-

ates involved in excited-state events of comparable

dynamics.

leaving group) deprotection and cyclization in CH CN. It is

3

envisaged that this single mechanism may apply also to the

deprotection reaction taking place in other polar aprotic sol-

vents, and also nonpolar solvents.

For the C₃₇₀intermediate, the

formation time (ꢀ14 ps) deter-

mined here for DMBDP is

comparable to that of the

l_m_a_x360 nm species reported for

DMB acetate and fluoride

A

H

U

G

R

N

U

G

ers. Since this intermediate is

the immediate reactive precur-

sor giving the cyclization reac-

tion [Eq. (5)], the observations

that 1) its formation dynamics

show little dependence on the

basicity

group,

of

the

leaving

[5,12]

and 2) the cyclization

yield is affected only modestly

by solvent polarity and is fa-

vored slightly in nonpolar over

[3a,f,5,8,7,12]

polar solvents,

lend

strong support to its attribution

as the biradical-natured inter-

mediate suggested by Wirz and

Scheme 2. Deactivation and photodeprotection and cyclization pathway of DMBDP in CH

the time-resolved spectra presented here and other results in the literature.

3

CN suggested from

[12]

co-workers and Chan and co-

[6]

workers (species 3 in path-

The 267 nm photoexcitation of DMBDP takes the system

to an upper singlet (S_n_>₁) excited state (Figure 2). The gen-

erally prompt IC is expected to lead the S_n_>₁population to

relax promptly to the S state. We thus relate the S state to

ways d and e in Scheme 1). Corroborating further this as-

cription, as well as the B exciplex to the C₃₇₀biradical con-

version, there have been many precedents in the literature

involving a quick transformation from an early CT exciplex

to a biradical precursor to account for photocyclization reac-

1

the denoted A state of the fs-TA spectra. This state corre-

sponds to the directly accessed singlet excited state when

using a >340 nm excitation, the desirable and frequently

used excitation wavelength in many of the previous studie-

[40,43,46,48–50]

tions of aromatic ketones.

We note, however, that

the previous studies by Wirz and Chan have provided no

specification as to the nature of the electronic state for the

[

3a,f,5–8]

[12]

s.

pacity of the Bz and DMBn moieties,

a probably enhanced ability induced by the electronic exci-

In view of the electron-donating and accepting ca-

C₃₇₀biradical.

The ꢀ1 ns lifetime of this species for

[3f,5,7,8]

respectively, and

DMBDP implies that it would correspond to a fairly well-

defined local minimum of the related potential hypersurface.

Our DFT calculations, intending to locate such a configura-

tion in the singlet ground-state surface by considering either

a syn or anti arrangement of the phosphate group relative to

the liable ortho-hydrogen of the DMBn subgroup, lead in-

[5,12,40,41]

tation,

we tentatively associate the observed B state

to the CT exciplex suggested by Wan and co-workers (in

pathway b of Scheme 1). It is pertinent to this point that

many of the previous studies on the excited states of (aro-

Chem. Eur. J. 2010, 16, 5102 – 5118

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

5113

D. L. Phillips et al.

evitably to structures with fission of the CꢁO bond connect-

er, in contrast to most of the literature work in considering

ing the phosphate and the DMB cage (see Figures 4S and

the cation 1 as a key precursor to the subsequent rate-deter-

+

5S, and Tables 2S and 3S in the Supporting Information for

mining H elimination to give the DMBF cyclization prod-

3

details of the optimized structures). From this, we conjecture

that, instead of being a ground-state species, the biradical

may still reside on the singlet excited-state surface, and is

formed by addition of the carbonyl oxygen (O) to the ortho-

carbon (C) of the benzylic ring. For a singlet-natured spe-

uct, our ns-TR data here indicate explicitly that this cation

is largely irrelevant to the cyclization, and that formation of

this species could represent a competing but energy-wasting

channel, probably leading mainly to re-formation of the re-

actant substrate or some other minor byproduct(s) with low

efficiency towards the cyclization reaction. As for the previ-

[31]

cies, such a bonding process is spin-allowed and,

in the

1

3

ꢁ1

particular case of the CT exciplex of DMBDP, the bonding

could be further facilitated by the intramolecular CT inter-

action between the DMBn-associated donor and the Bz-as-

ously suggested ultrafast (ꢀ10

s

rate) heterolysis cleav-

age pathway, which gives the cation 2 as a reactive inter-

[3f]

mediate (path c in Scheme 1), the distinct difference in

the early time (<1 ps) TA spectra of DMBDP from those of

DMBnDP, and the apparent lack in the DMBDP spectra of

the characteristic l_m_a_xꢀ420 cation absorption as exhibited

[3a,f,5,7,8,12,14,29]

sociated acceptor.

The rapid dynamics of biradi-

cal formation (ꢀ14 ps) suggests the addition reaction is a

fairly feasible excited-state process. Given the largely non-

planar conformation (with a nearly orthogonal relative ori-

entation of the Bz and DMBn ring planes, and the far-

beyond-bonding separation of the associated C and O

atoms) of the DMBDP ground state (see Figure 1S and

Table 1S in the Supporting Information) and the probably

similar structure of the exciplex precursor, it is possible that

the OꢁC addition is largely an entropy-controlled process

+

by the closely related DMBn , suggests that such a process,

if present, could play only a minor role in DMBDP photoly-

sis in CH CN.

3

From the above description of the assignments and dy-

namical conversions of the various intermediates, a pathway

for DMBDP deprotection can be constructed (Scheme 2).

After photoexcitation, the S transforms rapidly (ꢀ1.7 ps)

1

that requires substantial conformational rearrangement,

such as alignment to a certain extent of the two ring systems

and synchronous motion of the connecting carbon atoms,

and so forth. This parallels the widely recognized aspect that

the kinetics of many intramolecular photocyclizations are

subject to strong conformational and stereoelectronic con-

into a CT exciplex (B), which then undergoes intramolecular

CꢁO addition with ꢀ14 ps time constant to yield the prod-

uct-determining intermediate, that is, the singlet excited bi-

AHCTUNGRTEUNNGNr adical C₃₇₀. Bifurcation of the biradical decay leads primari-

ly to a concerted HX elimination, and to a much lesser

ꢁ

extent to an X elimination, with the former pathway pro-

[

42,43,45,51]

trol.

In nonpolar and polar aprotic solvents, such as

ducing the DMBF product directly, while the latter results

in the mainly non-product-forming D₄₈₀cation.

CH CN, the flexible construction of the DMBDP molecule

3

renders the needed conformation easily achievable, and this

may contribute to the rapid dynamics and high efficiency of

the OꢁC addition.

It is interesting to note that, although our ascription of

the C₃₇₀biradical to a singlet excited species implies a singlet

excited-state calculation (which is very difficult to perform

for this large system with two chromophores) is needed to

elucidate the detailed conformation and energy properties

of this species in giving the deprotection reaction and any

accompanied byproduct, our preliminary DFT results on the

ground state of this species, which are computationally

easier and more tractable (see Figure 4S–6S and Tables 2S–

4S in the Supporting Information), appear to contain useful

information for understanding the mode of the proposed

concerted HX elimination and the formation of the D₄₈₀

cation. On the one hand, the observation that the syn-ar-

rangement fragments into phosphoric acid and a structure

analogous to the DMBF product (Figure 4Sa and Table 2

Sa) is consistent with an intramolecular character syn elimi-

nation of HX in accounting for the concerted deprotection–

cyclization process. It is likely that for the operation of this

elimination mode, the phosphate group may act as an intra-

molecular base to help in removing the liable hydrogen. On

the other hand, the result that the anti-arrangement frag-

3

Our fs-TA and ns-TR data on DMBDP show that, upon

decay of the C₃₇₀biradical (Scheme 2), the system may con-

vert either directly to the DMBF product or to the D₄₈₀in-

termediate. With the biradical attribution of the C₃₇₀species,

the former channel presents a concerted HX elimination–

cyclization pathway for DMBDP photodeprotection. As this

channel works as the major pathway leading to the DMBF

product, the efficiency of the cyclization reaction would be

determined largely by the branching ratio of this to the com-

peting D₄₈₀formation. The generally high DMBF yield, as

reported consistently in previous studies, indicates a modest

yield of D₄₈₀. From this and the rather strong l₄₈₀absorption,

as suggested in the fs-TA spectra (Figure 7), a D₄₈₀species

with fairly large molar absorptivity may be implied. Our

TDDFT calculations for the structurally optimized cation 1

(

Scheme 1) reveals a strong absorption with an oscillator

strength of ꢀ0.45 at a wavelength of ꢀ440 nm (see Fig-

ure 6S and Table 4S in the Supporting Information for de-

tails). This appears to support the attribution of D₄₈₀to the

cation 1, a connection originally proposed by Wan and co-

ꢁ

ments into the phosphate anion (X ) and a counterpart

cation, as indicated by the calculated charge distribution

(Figure 5Sa and Table 3S in the Supporting Information),

may suggest formation following such a conformation of a

contact ion pair. Given the same chemical constitution and

the general resemblance in the charge and structural ar-

[

5]

workers. With this assignment, the C₃₇₀biradical to D

480

ꢁ

cation conversion can be considered as an X elimination

channel, similar to that suggested by Wirz and co-workers in

terms of this particular step (path e in Scheme 1). Howev-

[12]

5114

www.chemeurj.org

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2010, 16, 5102 – 5118

Bichromophoric Phototrigger 3’,5’-Dimethoxybenzoin Diethyl Phosphate

FULL PAPER

rangement of the counterpart cation to that of the D₄₈₀

proposed in the preceding paragraph. It is possible that the

cation (Figure 6S and Table 4S in the Supporting Informa-

tion), it is conceivable that the D₄₈₀cation may arise from

lack of the D cation for the photolysis of DMB acetate in

4

12]

80

[

cyclohexane may indicate a complete suppression of the

ꢁ

the charge separation of the contact ion pair and correspond

to the counterpart cation involved in the ion pair. This anti-

structure-associated ion-pair formation and charge separa-

X

elimination channel, thus rendering the HX elimination

an overwhelmingly predominant excited-state reaction in

the nonpolar environment.

ꢁ

tion may constitute a pathway giving the competing X

The photodeprotection mechanism of DMB photolysis in

aqueous solution is of particular interest due to the biologi-

elimination of the C₃₇₀biradical. Because the efficiency of

the ion-pair formation and the charge separation depends

cal relevance of the aqueous medium. However, like many

[15a,b,22,31,39]

[1,2,13,15,26,27]

critically on the polarity property of the solvent,

a

other useful a-keto PRPGs,

DMB photolysis in

higher yield of the D₄₈₀cation in a polar solvent is expected

within this assignment than in a nonpolar solvent. In fact, an

absence of the absorption due to the D₄₈₀cation has been

reported for photolysis of DMB-caged acetate in cyclohex-

water or largely water-containing solvent gives different

products from those in polar aprotic or nonpolar sol-

[6,12]

[12]

vents.

Photolysis of DMB acetate, fluoride,

and a

[6]

water-soluble DMB acetate in largely or wholly water sol-

utions has revealed that, as for the liberation of the leaving

groups, the cyclization is prevented in a substantial fraction

(030%) of photoproduct and, at the expense of this, there

is formation of a second solvolytic water addition product of

the corresponding benzoin. A similar solvent dependence of

product distribution has also been reported for photolysis of

parent benzoin compounds in aqueous versus aprotic sol-

[12]

ane (a nonpolar solvent), in contrast to the clear involve-

ment of such an absorption as observed for the same system

[5,12]

in CH CN (a polar solvent).

3

The mechanistic scenario proposed here (Scheme 2) rec-

onciles nicely with several important experimental observa-

tions reported for DMB deprotection in aprotic polar and

nonpolar solvents. Within this model, the decay rate of the

[2a,13a]

biradical C₃₇₀, which equals the sum of the rates of the HX

vents.

The appearance of the extra benzoin product at

ꢁ

(

k_H_X) and the X (k_X

ꢁ

) eliminations, serves as a measure of

the expense of BF for BDP photolysis in aqueous solution

has been found to be caused by the additional involvement

of a kinetically competitive heterolytic dissociation to bifur-

cate the triplet deprotection–cyclization and to result in the

formation of a triplet a-ketocation from the common pre-

8

ꢁ1

the overall rate (ꢀ1ꢃ10 s ) for liberation of the leaving

group; the concerted HX elimination is the rate-determining

step for the deprotection and cyclization with its efficiency

determined by the ratio of the k_H_Xto the overall rate. As-

suming the excited-state nature of the radical C₃₇₀, HX elim-

ination from this species to generate the extensively conju-

gated ground-state DMBF would appear to be a highly exo-

thermic and kinetically feasible reaction, likely to be associ-

[2a,13a]

cursor of the Bz-localized triplet.

In spite of the intrins-

ically different features of the DMBDP photodeprotection

from that of BDP, as illustrated in the preceding sections,

the observation of an analogous photosolvolysis product in

DMBDP may imply that its benzylic carbon must be accessi-

ble to nucleophilic attachment by water at a certain stage

along the photolysis pathway. In parallel to the photosolvol-

ysis pathway of the parent BDP, and also to take into ac-

count the high dissociation and solvation capacity of solvent

[31]

ated with an early transition state. The proposed intramo-

lecular nature of the HX elimination implies the reaction

requires little assistance from the surrounding medium.

These attributes appear to corroborate the general high effi-

ciency of the deprotection and cyclization and the fact that

this reaction is largely independent of the polarity property

water molecules in promoting rapid heterolytic cleava-

[3a,f,5,7,8,12]

[2a,13a,15a,b,17,26,39]

of the solvent (except in the case of water).

The

ge,

it appears rather reasonable to consider

suggestion of the leaving group working as an intramolecu-

lar base to facilitate the concerted syn elimination accom-

modates nicely the fact that the kinetic decay of the biradi-

cal C₃₇₀depends modestly on the leaving group involved in

the reaction. With the increase in lability of the leaving

group from fluoride to acetate to diethyl phosphate, the bir-

adical decay rate increases, so the decay time constants

become smaller, to range from ꢀ3 ns for fluoride to ꢀ2 ns

the inclusion of a similar cation species, the cation 2 pro-

[3a,f]

posed by Pirrung and co-workers,

as the transient precur-

sor to the solvolysis product from photolysis of the DMB

compounds in the predominantly water environment. Fur-

ther study is underway to detect directly the relevant transi-

ent species, and to explore the specific photodeprotection

pathway of DMB compounds in aqueous solutions.

[

12]

for acetate and further to ꢀ1 ns, as revealed here for the

phosphate leaving group. In addition, previous studies have

found a slightly lower cyclization efficiency of DMBX in

aprotic polar than nonpolar solvents. For example, the rela-

tive cyclization efficiency of DMB acetate was reported to

Conclusion

3

A combined fs-TA and ns-TR study has been performed to

characterize the reactive intermediates and to detect directly

the dynamics and pathways for DMBDP photodeprotection

[12]

be ꢀ86% in CH CN and ꢀ95% in cyclohexane.

From

3

our perspective, this can be rationalized in terms of a lower

and cyclization in CH CN. To assess the molecular cause of

3

involvement in the nonpolar than the polar solvent of the

the unusual di-meta-methoxy substituent effect, parallel fs-

TA measurements have been carried out on two reference

compounds, DPAP and DMBnDP, in the same solvent.

Comparison of the UV/Vis and fs-TA data obtained for

ꢁ

X

elimination process relative to the HX elimination pro-

cess in accounting for the biradical decay. This is as expect-

ꢁ

ed according to the anti-associated X elimination pathway

Chem. Eur. J. 2010, 16, 5102 – 5118

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

5115

D. L. Phillips et al.

these compounds and relevant TDDFT results reveals dis-

tinctively different spectral and dynamical features for

DMBDP from those of the reference systems. The results

suggest that, for DMBDP, electronic interaction between

the two sub-chromophores plays a governing role in ac-

counting for the involvement of a peculiar set of transient

states and species and the discrete deactivation and depro-

tection pathways. By directly monitoring the temporal gen-

eration of the DMBF product, the DMBDP cyclization time

monic of 800 nm, the regenerative amplifier fundamental) and probed by

a white light continuum probe beam produced from a rotating CaF plate

2

pumped by the fundamental laser pulses (800 nm). The pump and probe

laser beam spot sizes at the sample were about 200 and 100 mm, respec-

tively. The signals were focused into a monochromator and detected by a

2

liquid-N -cooled CCD. The time delay between the pump and probe

pulse was controlled by an optical delay line, and the time resolution of

the measurement was ꢀ200 fs. The TA measurements were carried out,

respectively, for DMBDP, DPAP, and DMBnDP in CH

3

CN at a sample

concentration of ꢀ1 mm.

3

Ns-time-resolved resonance Raman (ns-TR ) experiments: The ns-TR

constant was determined to be ꢀ1 ns in CH CN. This indi-

spectra presented here for DMBDP in CH CN were obtained by using

3

[

13,15]

cates explicitly that the long-lived species of the T state and

the commonly discussed l_m_a_x480 nm cation play only a

minor role in the DMBDP photodeprotection and cycliza-

tion. In contrast to the previously suggested singlet stepwise

deprotection and cyclization pathway, a revised singlet con-

certed deprotection–cyclization mechanism was proposed.

This mechanism includes a rapid ꢀ1.7 ps conversion of the

initial excited state into a CT exciplex that transforms with

the experimental apparatus and methods described previously.

Brief-

ly, the sample solution was pumped by a 240 nm excitation wavelength

and probed by a 299 nm wavelength light. The pump and probe pulse

were provided, respectively, by the first anti-Stokes and Stokes output

2

from a H Raman shifter, each pumped by the fourth harmonic output

(266 nm) of two individual Nd:YAG lasers. The time delay between the

pump and probe pulses was controlled electronically by a pulse genera-

tor, and the time resolution of these measurements was ꢀ10 ns. The

pump and probe laser beams were lightly focused onto a flowing sample

solution and the Raman scattering was collected through a backscattering

ꢀ

14 ps time constant to a singlet excited l_m_a_x370 nm biradi-

configuration and detected by a liquid-N -cooled CCD. The resonance

Raman spectrum of DMBF in CH CN was acquired by using a 299 nm

3

2

cal, which acts as the key reactive precursor to the ensuing

leaving group liberation and cyclization. The deprotection

of the biradical proceeds with ꢀ1 ns time constant and is

caused by competition between two different reactions, of

which the concerted HDP elimination constitutes the pre-

dominant pathway leading directly to yield the DMBF final

excitation wavelength, with the signal collection configuration similar to

3

that of the ns-TR measurements. Sample concentrations for these

Raman experiments were ꢀ1 mm. The wavenumber shifts of the reso-

3

nance Raman spectra were calibrated using the known CH CN solvent

3

Raman bands, and the ns-TR spectra presented were obtained by sub-

tracting an appropriately scaled probe-before-pump spectrum from the

corresponding pump-probe spectrum.

ꢁ

product, whereas the heterolytic DP elimination that gives

Density functional theoretical (DFT) calculations: The TDDFT calcula-

^t^h^e^lmax 480 nm cation works only as a minor pathway, bear-

ing little relevance to the cyclization reaction. This work

contributes important new insights for an improved mecha-

nistic description of DMB photodeprotection, and helps

give a better understanding of the structure/reactivity rela-

tionship of the DMB PRPG.

[52]

tions were carried out employing the Gaussian 03 program suite. The

calculations were performed to obtain transition energies and oscillation

strengths of electronic transitions of the species of interest. For each of

the systems, the calculation was performed based on the optimized struc-

ture of the studied system obtained firstly by DFT geometry optimization

computation. To find out the optimized structure(s) for the singlet

ground state of the DMBDP biradical, the energy optimization calcula-

tion was first done for the triplet ground state of this species. This gave

two optimized conformations, corresponding to diastereomers with the

diethyl phosphate group being syn and anti relative to the liable ortho-

hydrogen in the DMBn sub-group (see Figures 4Sb and 5Sb and Ta-

bles 2Sb and 3Sb in the Supporting Information). These geometries were

then used as initial conformations to acquire the corresponding optimized

structures for the singlet ground state of the biradical, as shown in Figur-

es 4Sa, 5Sa, and Tables 2Sa and 3a in the Supporting Information. All of

the calculations were performed using the B3LYP method with a 6–31G*

basis set.

Experimental Section

Materials: DMBDP, DPAP, and DMBnDP were synthesized following

the methods described in the Supporting Information. The identity and

purity of the samples were confirmed by NMR and UV absorption spec-

troscopy, and mass spectrometry. The details of the NMR characteriza-

tion of the synthesized samples are given in the Supporting Information.

DMBF was synthesized according to the photolysis method described in

[

2c]

the literature, but replacing BDP with DMBDP. Spectroscopic grade

CH CN was used as the solvent for all the steady-state and time-resolved

3

experiments.

Acknowledgements

UV absorption measurement: The UV absorption spectra from the dilute

sample solution in CH

3

CN (ꢀ1 mm concentration) were measured using

This research was carried out in the Ultrafast Laser Facility at the Uni-

versity of Hong Kong and supported by grants from the Research Grants

Council of Hong Kong (HKU 1/01C and HKU 7039/07P) to D.L.P. and

(PolyU 7029/06P and PolyU 7029/07P) to W.M.K. D.L.P. thanks the

Croucher Foundation for the Award of a Senior Research Fellowship and

the University of Hong Kong for the Outstanding Researcher Award.

a Perkin–Elmer Lambda 19 UV/Vis spectrometer. The molar extinction

ꢁ

1

ꢁ1

coefficients (M cm ) were calculated based on the Beer–Lambert law.

Photochemistry experiments: The proceeding of the deprotection reac-

tion initiated by photolysis of DMBDP in CH CN was monitored by

3

measuring spectral change in UV absorption of a DMBDP sample solu-

tion (in a 1 cm UV cell) after various periods of exposure to an unfo-

cused 266 nm laser line (ꢀ1 mJ). The 266 nm light is from the fourth har-

monic of a nanosecond Nd:YAG Q-switch laser.

[

1] R. S. Givens, L. W. Kueper, Chem. Rev. 1993 , 93 , 55–66.

Femtosecond transient absorption (fs-TA) experiments: The fs-TA meas-

urements were performed by using a commercial femtosecond Ti:Sap-

phire regenerative amplifier laser system. Details of the TA experimental

setup are described in references [13,15]. For the present measurements,

sample solutions were excited by a 267 nm pump beam (the third har-

[2] a) C. S. Rajesh, R. S. Givens, J. Wirz, J. Am. Chem. Soc. 2000 , 122 ,

611–618; b) R. S. Givens, P. S. Athey, L. W. Kueper, B. Matuszewski,

J. Y. Xue, J. Am. Chem. Soc. 1992 , 114 , 8708–8710; c) R. S. Givens,

P. S. Athey, B. Matuszewski, L. W. Kueper, J. Y. Xue, T. Fister, J.

Am. Chem. Soc. 1993 , 115 , 6001–6012.

5116

www.chemeurj.org

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2010, 16, 5102 – 5118

Bichromophoric Phototrigger 3’,5’-Dimethoxybenzoin Diethyl Phosphate

FULL PAPER

[

3] a) M. C. Pirrung; L. Fallon; D. C. Lever; S. W. Shuey, J. Org. Chem.

[27] a) K. Zhang, J. E. T. Corrie, V. R. N. Munasinghe, P. Wan, J. Am.

Chem. Soc. 1999 , 121 , 5625–5632 ; b) D. W. Brousmiche, P. Wan, J.

Photochem. Photobiol. A 2000 , 130 , 113–118.

[28] a) D. P. DeCosta, N. Howell, A. L. Pincock, J. A. Pincock, S. Rifai, J.

Org. Chem. 2000 , 65 , 4698–4795; b) J. A. Pincock, I. S. Young, Can.

J. Chem. 2003 , 81 , 1083–1095 ; c) J. W. Hilborn, E. MacKnight, J. A.

Pincock, P. J. Wedge, J. Am. Chem. Soc. 1994 , 116 , 3337–3346;

d) J. A. Pincock, Acc. Chem. Res. 1997 , 30 , 43–49.

1996 , 61 , 2129–2136, S. W. Shuey, J. Org. Chem. 1994 , 59 , 3890–

3897; b) M. C. Pirrung, C. Y. Huang, Tetrahedron Lett. 1995 , 36 ,

5883–5884; c) M. C. Pirrung, J. C. Bradley, J. Org. Chem. 1995 , 60 ,

1116–1117; d) M. C. Pirrung, C. Bradley, J. Org. Chem. 1995 , 60 ,

6270–6276; e) M. C. Pirrung; L. Fallon; D. C. Lever; S. W. Shuey, J.

Org. Chem. 1996 , 61 , 2129–2136; f) M. C. Pirrung, T. Ye, Z. Zhuo;

J. D. Simon, Photochem. Photobiol. 2006 , 82 , 1258–1264; J. D.

Simon, Photochem. Photobiol. 2006, 82 , 1258–1264.

[

29] H. E. Zimmerman, V. R. Sandel, J. Am. Chem. Soc. 1963 , 85 , 915–

22.

9

[

4] J. M. Peach, A. J. Pratt, J. S. Snaith, Tetrahedron 1995 , 51 , 10013–

[

30] a) T. Goshima, Y. Itoh, H. Shirai, M. Kojima, J. Photochem. Photo-

biol. A 2001 , 141 , 139–145 ; b) T. Goshima, Y. Itoh, M. Kojima, T.

Karatsu, J. Photochem. Photobiol. A 1999 , 127 , 75–81.

1

0024.

[5] Y. Shi, J. E.T Corrie, P. Wan, J. Org. Chem. 1997 , 62 , 8278–8279.

[6] R. S. Rock, S. I. Chan, J. Am. Chem. Soc. 1998 , 120 , 10766–10767.

[7] a) J. C. Sheehan, R. M. Wilson, A. W. Oxford, J. Am. Chem. Soc.

[

31] N. J. Turro, Modern Molecular Photochemistry, University Science

Book, Sausalitos, 1991.

32] J. Bartl, S. Steenken, H. Mayr, R. A. McClelland, J. Am. Chem. Soc.

1

971 , 93 , 7222–7228 ; b) J. C. Sheehan, R. M. Wilson, J. Am. Chem.

Soc. 1964 , 86 , 5277–5281.

1

990 , 112 , 6918–6918.

33] a) M. Lipson, A. A. Deniz, K. S. Peters, J. Am. Chem. Soc. 1996 ,

18 , 2992–2997 ; b) M. Lipson, A. A. Deniz, K. S. Peters, J. Phys.

[

8] J. F. Cameron, C. G. Wilson, J. M. J. Frechet, J. Am. Chem. Soc.

1

996 , 118 , 12925–12937.

1

9] C. P. McCoy, C. Rooney, C. R. Edwards, D. S. Jones, S. P. Gorman, J.

Am. Chem. Soc. 2007 , 129 , 9572–9573.

Chem. 1996 , 100 , 3580–3586 ; c) M. Lipson, A. A. Deniz, K. S.

Peters, Chem. Phys. Lett. 1998 , 288 , 781–784.

[

10] J. E. T. Corrie, D. R. J. Trentham, Chem. Soc. Perkin Trans. 1 1992 ,

409–2417.

[

34] G. Porter, M. W. Windsor, Nature 1957 , 180 , 187–188.

35] H. C. Christensen, K. Sehested, E. J. Hart, J. Phys. Chem. 1973 , 77 ,

2

[

11] a) R. S. Rock, S. I. Chan, J. Org. Chem. 1996 , 61 , 1526–1529;

b) R. P.-Y. Chen, J. J.-T. Huang, H.-L. Chen, H. Jan, M. Velusamy,

C.-T. Lee, W. Fann, R. W. Larsen, S. I. Chan, Proc. Natl. Acad. Sci.

USA 2004, 101, 7203–7310; c) R. S. Rock, K. C. Hansen, R. W.

Larsen, S. I. Chan, Chem. Phys. 2004 , 307 , 201–208.

9

83–987.

[

36] H. G. Heine, W. Hartmann, D. R. Kory, J. G. Magyar, C. E. Hoyle,

J. K. McVey, F. D. Lewis, J. Org. Chem. 1974 , 39 , 691–698.

37] J. M. Masnovi, S. Sankararaman, J. K. Kochi, J. Am. Chem. Soc.

1

989 , 111 , 2263–2276.

[

12] H. Boudebous, B. Kosmrlj, B. Sket, J. Wirz, J. Phys. Chem. A 2007 ,

38] a) R. S. Givens, B. Matszewski, P. S. Athey, M. R. Stoner, J. Am.

Chem. Soc. 1990 , 112 , 6016–6021 ; b) R. S. Givens, B. Matuszewski,

J. Am. Chem. Soc. 1984 , 106 , 6860–6861.

1

11 , 2811–2813.

13] a) C. Ma, W. M. Kwok, Y. Du, D. L. Phillips, Chem. Eur. J. 2007 , 13 ,

290–2305 ; b) W. S. Chan, C. Ma, W. M. Kwok, P. Zuo, D. L. Phil-

2

[

39] C. Reichardt, Solvent and Solvent Effect in Organic Chemistry,

VCH, Weinheim, 1988.

lips, J. Phys. Chem. A 2004 , 108 , 4047–4058; c) Y. Du, PhD. Thesis,

University of Hong Kong 2007 .

[

40] a) I. E. Kochevar, P. J. Wagner, J. Am. Chem. Soc. 1972 , 94 , 3859–

[

14] a) H. E. Zimmerman, J. Phys. Chem. A 1998 , 102 , 5616–5621;

b) H. E. Zimmerman, J. Am. Chem. Soc. 1995 , 117 , 8988–8991.

15] a) C. Ma, W. M. Kwok, W. S. Chan, P. Zuo, J. T. W. Kan, P. H. Toy,

D. L . Phillips, J. Am. Chem. Soc. 2005 , 127 , 1463–1472; b) C. Ma,

W. M. Kwok, W. S. Chan, Y. Du, J. T. W. Kan, P. H. Toy, D. L . Phil-

lips, J. Am. Chem. Soc. 2006 , 128 , 2558–2570 ; c) W. S. Chan, C. Ma,

W. M. Kwok, D. L. Phillips, J. Phys. Chem. A 2005 , 109 , 3454–3469 ;

d) P. Zuo, C. Ma, W. M. Kwok, W. S. Chan, D. L. Phillips, J. Org.

Chem. 2005 , 70 , 8661–8675; e) C. Ma, P. Zuo, W. M. Kwok, W. S.

Chan, J. T. W. Kan, P. H. Toy, D. L. Phillips, J. Org. Chem. 2004 , 69 ,

3

5

865 ; b) P. J. Wagner, R. A. Leavitt, J. Am. Chem. Soc. 1970 , 92 ,

806–5808.

[

41] R. W. Ricci, J. M. Nesta, J. Chem. Phys. 1976, 64–65 , 974–980.

42] a) J. C. Netto-Ferreira, W. J. Leigh, J. C. Scaiano, J. Am. Chem. Soc.

1

985 , 107 , 2617–2622 ; b) J. C. Netto-Ferreira, I. G. J. Avellar, J. C.

Scaiano, J. Org. Chem. 1990 , 55 , 89–92.

[

43] a) P. J. Wagner, M. A. Meador, B. Zhuo, B.-S. Park, J. Am. Chem.

Soc. 1991 , 113 , 9630–9639 ; b) P. J. Wagner, B. Zhou, T. Hasegawa,

D. L. Ward, J. Am. Chem. Soc. 1991 , 113 , 9640–9654; c) P. J.

Wagner, B. P. Giri, J. C. Scaiano, D. L. Ward, E. Gabe, F. L. Lee, J.

Am. Chem. Soc. 1985 , 107 , 5483–5490.

6

641–6647; f) C. Ma, W. S. Chan, W. M. Kwok, P. Zuo, D. L. Phil-

lips, J. Phys. Chem. A 2004, 108 , 9264–9276.

16] a) D. R. Kearns, W. A. Case, J. Am. Chem. Soc. 1966 , 88 , 5087–

[44] T. Wilson, A. M. Halpern, J. Am. Chem. Soc. 1981 , 103 , 2412–2413.

[45] a) J. N. Moorthy, S. L. Monahan, R. B. Sunoj, J. Chandrasekhar, C.

Bohne, J. Am. Chem. Soc. 1999 , 121 , 3093–3103; b) S. Samanta,

B. K. Mishra, T. C. S. Pace, N. Sathyamurthy, C. Bohne, J. N. Moor-

thy, J. Org. Chem. 2006 , 71 , 4453–4459.

[46] L. Giering, M. Berger, C. Steel, J. Am. Chem. Soc. 1974 , 96 , 953–

958.

[47] P. K. Das, K. Bobrowski, J. Chem. Soc. Faraday Trans. 1981, 77,

1009–1027.

[

5

097; b) W. A. Case, D. R. Kearns, J. Chem. Phys. 1970 , 52 , 2175–

2

191.

[

17] M. B. Ledger, G. Porter, J. Chem. Soc. Farady Trans. 1972, 68, 539–

53.

18] H. Lutz, E. Breheret, L. Lindqvist, J. Phys. Chem. 1973 , 77 , 1758–

762.

5

[

1

[

19] P. M. Rentzepis, Science 1970 , 169 , 239–247.

20] P. J. Wagner, A. E. Kemppaninen, H. N. Schott, J. Am. Chem. Soc.

[48] F. D. Lewis, T. A. Hillard, J. Am. Chem. Soc. 1970 , 92 , 6672–6674.

[49] R. R. Hautala, N. J. Turro, J. Am. Chem. Soc. 1971 , 93 , 5595–5597.

1

973 , 95 , 5604–5614.

[

50] N. C. Yang, M. H. Hui, S. A. Bellard, J. Am. Chem. Soc. 1971 , 93 ,

056–4058.

51] M. Lipson, P. F. McGarry, I. V. Koptyug, H. A. Staab, N. J. Turro,

D. C. Doetschman, J. Chem. Phys. 1994, 100–101, 7504–7512.

52] Gaussian 98, Revision A.11.3, M. J. Frisch, G. W. Trucks, H. B.

Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakr-

[

21] M. Montalti, A. Credi, L. Prodi, M. T. Gandolfi, Handbook of Pho-

tochemistry, CRC Press, New York, 2006.

22] a) M. Newcomb, J. H. Horner, P. O. Whitted, D. Crich, X. Huang, Q.

Yao, H. Zipse, J. Am. Chem. Soc. 1999 , 121 , 10685–10694 ; b) J. H.

Horner, E. Taxil, M. Newcomb, J. Am. Chem. Soc. 2002 , 124 , 5402–

4

5

410; c) P. O. Whitted, J. H. Horner, M. Newcomb, X. Huang, D.

zewski, J. A. Montgom ACHUTNRGNENUGe ry, Jr., R. E. Stratmann, J. C. Burant, S. Dap-

Crich, Org. Lett. 1999 , 1 , 153–156.

prich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O.

Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.

Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y.

Ayala, Q. Cui, K. Morokuma, N. Rega, P. Salvador, J. J. Dannen-

berg, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman,

J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A.

[23] S. Dym, R. M. Hochstrasser, J. Chem. Phys. 1969 , 51 , 2458–2468.

[24] A. Banerjee, D. E. Falvey, J. Am. Chem. Soc. 1998 , 120 , 2965–2966.

[25] S. Tournon, M. A. El-Bayoumi, J. Chem. Phys. 1972 , 56 , 5128–5134.

[26] P. G. Conrad, R. S. Givens, B. Hellrung, C. S. Rajesh, M. Ramseier,

J. Wirz, J. Am. Chem. Soc. 2000 , 122 , 9346–9347.

Chem. Eur. J. 2010, 16, 5102 – 5118

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

5117

D. L. Phillips et al.

Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin,

D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara,

M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong,

J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle, J. A.

Pople, Gaussian, Inc., Pittsburgh PA, 2002.

Received: October 20, 2009

Published online: March 26, 2010

5118

www.chemeurj.org

ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2010, 16, 5102 – 5118

Article Doi

A time-resolved spectroscopic study of the bichromophoric phototrigger 3',5'-dimethoxybenzoin diethyl phosphate: Interaction between the two chromophores determines the reaction pathway

DOI: 10.1002/chem.200902891

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.200902891

Full text of DOI:10.1002/chem.200902891

Products guided by the article

R&D Labs maybe for 107933-78-8

Relevant to this article

Hot Product