Femtosecond transient absorption and nanosecond time-resolved resonance raman study of the solvent-dependent photo-deprotection reaction of benzoin diethyl phosphate

DOI: 10.1002/chem.200600893

Femtosecond Transient Absorption and Nanosecond Time-Resolved

Resonance Raman Study of the Solvent-Dependent Photo-Deprotection

Reaction of Benzoin Diethyl Phosphate

Chensheng Ma, Yong Du, Wai Ming Kwok, and David Lee Phillips*^[a]

Abstract:

transient absorption (fs-TA) and nano-

second time-resolved resonance

A

combined femtosecond

Raman (RR) measurements, the TA

results reveal that the phenacyl local-

ized BDP triplet state (that is mainly

np* nature) is the common and imme-

diate precursor to the photo-deprotec-

tion reaction in both solvents. Howev-

er, the triplet deprotection follows dif-

ferent pathways in neat MeCN versus

the largely water containing solvent.

The deprotection reaction in MeCN

was determined to occur with a ~11 ns

time constant and the reaction was

found to be an unimolecular process

leading to elimination of the diethyl

phosphoric acid apparently concurrent

with cyclization to yield the benzofuran

product. In the water mixed solvent,

the triplet reaction was observed to

proceed with a ~15 ns time constant

and the reaction leads to not only the

deprotection–cyclization but also a het-

erolytic dissociation to release the di-

Raman (ns-TR³) study was performed

to directly detect the dynamics and elu-

cidate the mechanism of the excited

state deactivation and solvent-depen-

dent photo-deprotection pathways for

benzoin diethyl phosphate (BDP) in

neat acetonitrile (MeCN) and 75%

H₂O/25% MeCN. Comparison of the

TA spectral evolution observed in the

two solvents provides explicit evidence

that the photophysical deactivation of

the BDP singlet excited state has little

solvent dependence. The TA spectra

also indicate the related internal con-

version (IC) and intersystem crossing

(ISC) processes occur rapidly on hun-

dreds of femtoseconds and ~2–3 ps

time scales, respectively. From this and

in conjunction with a photochemistry

study and ground state resonance

ethyl phosphate anion through

a

branching and competing mechanism.

The ns-TR³spectra combined with rel-

evant DFT calculations have been used

to characterize the dynamics, structure

and vibrational frequencies to help

identify the important intermediates as

well as to explore the reaction pathway

leading to formation of the solvolysis

product in the largely water solvent. A

consecutive mechanism has been re-

vealed for the heterolysis–solvolysis re-

action in the water mixed solvent. The

present work provides direct and irrev-

ocable evidence for the dynamics and

mechanistic description of the overall

photophysics and deprotection related

photochemistry for BDP.

Keywords: Raman spectroscopy

solvent effects time-resolved

spectroscopy · transient absorption

·

Introduction

use in physiology experiments. Among the various protect-

ing groups in the literature, some aromatic a-keto

Photochemically removable protecting groups (PRPGs)

have found increasing utility in many applications demand-

ing temporally and spatially controlled release of reactive

species in situ.^[1–8]Examples include a variety of phototrig-

ger compounds which have been developed for potential

groups^[1,3,9,10]

such

as

p-hydroxyphenacyl

(pHP)

cages^{[11,12,13,14]}have gained interest due to their potential use

as phototriggers for the fast and efficient liberation of vari-

ous biologically active stimulants. Benzoin and substituted

benzoin esters are another important class of aromatic a-

keto compounds that display attractive photo-deprotection

properties such as a rapid liberation rate, a relatively high

quantum yield for release of leaving group and the genera-

tion of a highly fluorescent but inert benzofuran by-prod-

uct.^{[1,10,15,16–21]}The practical features of these compounds

have led to their extensive development as PRPGs for inor-

[a]Dr. C. Ma, Y. Du, Dr. W. M. Kwok, Prof. Dr. D. L. Phillips

Department of Chemistry, The University of Hong Kong

Pokfulam Road, Hong Kong S.A.R. (PR China)

Fax : (+852)2857-1586

E-mail: phillips@hkucc.hku.hk

ganic

phosphates,^{[1,15a,b,20a]}

nucleotides,^[15c,16]

carboxy-

Supporting information for this article is available on the WWW

under http://www.chemeurj.org/ or from the author.

AHCTREUNG

2290

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2007, 13, 2290 – 2305

FULL PAPER

compounds have also been exploited for use in lithographic

synthesis^[20d]and as photoliable linkers.^[18b]

bond producing the benzylic a-ketocation as a key interme-

diate^[20a]and ii) the involvement of a intramolecular exciplex

followed by heterolytic cleavage giving a short-lived cationic

intermediate^[17]were also suggested for the deprotection–

cyclization of a series of 3’,5’-dimethoxybenzoin esters

(DMBEs). In addition, the formation of a biradical inter-

mediate followed by acetoxy migration was proposed to

result in both the cyclization and solvolysis products ob-

served for the 3’,5’-bis(carboxymethoxy)benzoin acetate

compound in a buffered aqueous solution.^[18a]To evaluate

the preceding proposed mechanisms and substantiate the

pertinent reaction pathways, real time kinetic information as

well as unambiguous identification of the reactive inter-

mediate(s) are needed. To our knowledge, direct and explic-

it evidence of this kind has been scarce for the desyl PRPGs

(such as benzoin^[15a]and substituted benzoin esters). To help

provide this required information, a comparative femtosec-

ond time-resolved transient absorption (fs-TA) and nanosec-

ond time-resolved resonance Raman (ns-TR³) study on the

benzoin diethyl phosphate (BDP) photo-deprotection reac-

tion in MeCN versus. largely water mixed MeCN solvent is

reported here.

The versatile applications and the biochemical and syn-

thetic relevance of the benzoinyl PRPGs have attracted con-

siderable interest in understanding the photo-deprotection

mechanism(s) of these compounds.^{[1,15,17,18a,19,20a,21,22]}Previ-

ous studies have shown the diversity and complexity of the

reaction pathways leading to deprotection. The photoprod-

uct composition and reaction efficiency have been found to

depend on many factors. These factors include the position

and type of substituent(s) on the benzyl ring, the nature of

the leaving group and the character of the solvent environ-

ment. Photolysis of desyl compounds can yield a complicat-

ed mixture of coupling and addition products. Two typical

reactions occurring in normal organic solvent and solvents

with high ionizing ability, respectively, form the groundwork

for the desyl or substituted desyl compounds being exploited

as PRPGs or cages in phototrigger compounds. In organic

solvents such as MeCN and benzene, the photo-liberation is

accompanied by cyclization of the benzoinyl cage into the

corresponding 2-phenylbenzofuran (PBF) product. The PBF

conversion efficiency was found to be highest for the desyl

esters with a meta-methoxy benzyl substitution but to be rel-

atively modest for unsubstituted desyl compound-

s.^{[17,18a,19,20a,21]}In (a largely) aqueous solution, the corre-

sponding PBF yield decreases remarkably compared with

that observed in MeCN; in addition to PBF, a second prod-

uct (which is the primary product in largely aqueous sol-

vents) has been observed and assigned to the solvolytic nu-

cleophilic addition product of the corresponding benzoin

(BZ) or substituted benzoin compound based on various

product analysis methods.^[15a,18a]This solvent-dependence im-

plies the solvent plays a crucial role in deciding the excited

state reactivity and in controlling the competition between

the reaction channels for the formation of the various prod-

ucts observed in different solvent environments.

BDP is a representative unsubstituted desyl caged com-

pound exhibiting the characteristic solvent-dependent de-

protection photochemistry of benzoin esters [see Eqs. (1)

and (2)].

While the product identities and the reaction conditions

appear to be well-established, the underlying kinetics and

details of the reaction mechanism(s) leading to the depro-

tection and the PBF and BZ by-product formation as well as

the associated solvent dependence have remained unclear

for the benzoin and substituted benzoin esters. Substantially

different views and various intermediates have been pro-

posed in previous studies to interpret the experimentally ob-

served solvent effects and the influence of the ring substitu-

tion and leaving group on the desyl deprotection photo-

chemistry. For example, triplet mechanisms including i) ho-

molytic cleavage followed by electron transfer^[1,10]and ii) an

apparently single-step transformation through a rate-limiting

formation of a cyclized biradical intermediate^[15a]were pro-

posed for the deprotection–cyclization of unsubstituted ben-

zoin esters. In addition, a heterolytic triplet cleavage of the

Givens and co-workers reported that photolysis of BDP

in MeCN results in conversion to diethyl phosphoric acid

and PBF with a nearly quantitative yield and a quantum ef-

ficiency of 0.28. The same product mixture and a similar

quantum yield were also observed in MeOH and ben-

zene.^[1,10,15]However, in a largely aqueous solution, the de-

protection reaction gives the solvolytic product BZ as the

major product and the PBF as only a minor product.^[15a]The

predominant formation of the corresponding solvent nucleo-

philic addition product was also observed in highly ionizing

solvents such as trifluoroethanol or hexafluoro-2-propano-

l.^[15a]This product distribution and this typical solvent de-

pendence make BDP very suitable example for a mechanis-

tic study of the deprotection reactions of the desyl PRPG

(or benzoin esters). Based on laser flash photolysis (LFP)

and excited-state Stern–Volmer quenching studies, Givens

ꢀ

keto a-C O bond yielding the triplet cation was also sug-

gested to be responsible for the deprotection and solvolysis

observed for a benzoin caged phosphate in a largely water

containing solvent.^[15a]Furthermore, singlet mechanisms in-

ꢀ

cluding i) direct heterolytic cleavage of the ketone a-C O

Chem. Eur. J. 2007, 13, 2290 – 2305

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

2291

D. L. Phillips et al.

and co-workers provided convincing evidence for a triplet

state mechanism and proposed two different pathways to be

responsible for the solvent dependent product composition-

s.^[15a]However, the predominant factor(s) controlling the

mechanism and competition between the different pathways

remains unclear and further direct kinetic and structural in-

formation are needed to explicitly identify the actual inter-

mediates involved in the deprotection and elucidate the de-

tails of the reaction mechanism. Structural information on

the related intermediates and their importance to the ob-

served photochemistry also appears absent. The fs-TA and

ns-TR³results presented here provide real-time monitoring

of the reaction dynamics and direct spectral and structural

characterization of the intermediates involved in the photo-

chemistry. The ns-TR³spectra combined with density func-

tional theoretical (DFT) calculations to determine the struc-

tures and vibrational frequencies provide solid evidence for

the assignment of the reactive intermediates. To better un-

derstand the role of the solvent in influencing the reaction

pathways, the effect of inter-solvent–solute hydrogen-bond-

ing interaction and its structural relevance in the deprotec-

tion, cyclization and solvolyisis reactions were also explored.

These results shed new light on the photochemistry and lead

to an improved understanding of the photo-deprotection

mechanism for the desyl related PRPGs.

Figure 1. UV/Vis absorption spectra formed upon 267 nm laser photolysis

of BDP in MeCN solvent, a) without, b) after 2 min, c) 4 min, d) 6 min,

e) 8 min, f) 12 min, g) 60 min exposure to the laser light.

bles the BDP reactant (Figure S1 in the Supporting Informa-

tion), it is expected that the UV/Vis spectral changes ob-

served in the photolysis experiment does not afford a clear

indication of the BZ generation. However, as illustrated

below, the ns-TR³measurements do provide clear evidence

for the dynamical formation of BZ in the same solvent. In

general, the photochemistry observed here for BDP in the

two solvents is consistent with previously reported re-

sults^[15a,18a]and the photo-induced reactions depicted in

Results and Discussion

Equatoins (1) and (2). More importantly, as demonstrated

later in the fs-TA spectral evolution analysis, the broad and

slightly structured ~300 nm PBF absorption band (with sub-

peaks at ~315 nm, ~300 nm, and 290 nm) can serve as a

characteristic indication to track the cyclization dynamics

leading to the PBF formation.

Steady-state photolysis

Figure 1 shows UV/Vis spectra obtained for the photolysis

of BDP in MeCN solvent at varying irradiation times. It can

be seen that, with an increase of the irradiation time, the

BDP absorption spectrum (maxima at ~250 nm, spectrum a)

in Figure 1) transforms gradually into an alternative spec-

trum with two absorption bands with maxima at ~300 and

230 nm (spectrum g) in Figure 1). The latter spectrum is

identical to the absorption spectrum of PBF.^{[10,15a,20a,21]}This

together with the observation of the two well-defined iso-

sbestic points at ~240 and ~265 nm clearly indicates a direct

photo-induced conversion from BDP into PBF. The corre-

sponding photochemistry measurements performed for BDP

in 75% H₂O/25% MeCN solvent gave a similar spectral

transformation but with the relative absorption ratio of the

PBF product to the BDP reactant being substantially small-

er than that observed in neat MeCN, which is consistent

with the much reduced yield of PBF in the water mixed sol-

vents relative to MeCN.^[15a]Using the absorption extinction

coefficient reported for PBF in methanol (e=32220m^ꢀ1cm^ꢀ1

at 302 nm)^[10,15a]and assuming the same value in neat MeCN

and the water mixed MeCN, the PBF quantum yield can be

estimated to be roughly ~0.35ꢁ0.1 in neat MeCN and

~0.19ꢁ0.1 in the water mixed solvent. These values are

comparable to those reported in literature.^[1,10,15]

Resonance Raman (RR) spectroscopy

The RR spectra obtained with 267 nm excitation for BDP in

neat MeCN and 75% H₂O/25% MeCN solvents and their

respective comparison with the corresponding DFT calculat-

ed normal Raman spectra are displayed in Figure 2. The cal-

culated spectrum for comparison with the MeCN spectrum

is from the optimized geometry of free BDP molecule;

while that for comparison with the mixed solvent spectrum

is from the computation of the carbonyl hydrogen-bonded

complex containing one water molecule. The optimized

structures of the free and H-bonded BDP are displayed in

Figure 3 and the corresponding geometric parameters are

listed in Table S1 in the Supporting Information. Compari-

son of the experimental spectra obtained in the two solvents

reveals essentially identical vibrational features with the

same intensity distribution and frequencies except for the

predominant carbonyl stretching mode showing a ~11 cm^ꢀ1

frequency downshift upon going from the spectrum in

MeCN (1708 cm^ꢀ1) to that in the mixed solvent (1697 cm^ꢀ1).

The experimental observations and the frequency downshift

of the carbonyl stretching vibration are well reproduced by

the calculated results. A similar carbonyl stretching frequen-

Since the absorption spectrum of the second expected

product BZ in the water containing solvent closely resem-

2292

www.chemeurj.org

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2007, 13, 2290 – 2305

Time-Resolved Resonance Raman Studies

FULL PAPER

tions for the free and H-bonded BDP. The H-bonding leads

ꢀ

to a slight increase in the carbonyl C O bond length

(~0.005 ) and a modest variation in the relative reorienta-

tion between the benzoyl subgroup and benzylic moiety but

with the other structural parameters kept nearly unchanged

to those in the free BDP molecule. Based on the calculated

total energies, the stabilization energy due to the H-bonding

can be estimated to be ~5.4 kcalmol^ꢀ1. The calculations also

find that, with a similar H-bonding strength, the carbonyl

oxygen can H-bonded with an additional water molecule

that locates at a position symmetric relative to the first one.

This is consistent generally with the H-bonding configura-

tions observed for other aromatic carbonyl compounds^[23]

and indicates that, different from the free BDP molecule in

MeCN solvent, the ground state BDP molecule is present in

the complex form with H-bonding to the surrounding water

molecules in the largely water containing solvent.

Figure 2. RR spectra of BDP in neat MeCN (b) and 75% H₂O/25%

MeCN (c) and their respective comparison with the corresponding DFT

calculated normal Raman spectra (a) and (d). The asterisks (*) mark sol-

vent subtraction artifacts.

In addition to the H-bonding effect, it is important to

mention that, as summarized in Table S2 (Supporting Infor-

mation), most vibrational features enhanced by the 267 nm

excitation are due to the various normal modes associated

with the phenacyl subgroup. This indicates that this sub-

group is the chromophore responsible for the photoexcita-

tion. This localized nature of the BDP photoexcitation is

consistent with our previous excitation-dependent RR anal-

ysis.^[22]From this as well as from knowledge of the nature of

the transition of photo-excited aromatic carbonyl com-

pounds,^{[11,19,22–26]}it is obvious that the observed BDP RR

spectra in the two solvents are both correlated with the

strong pp* electronic transition (e.g. the S₃state) of the phe-

nacyl subgroup.

Fs-transient absorption (fs-TA) spectroscopy

Fs-TA measurement for BDP in MeCN: Figure 4 shows fs-

TA spectra of BDP obtained in neat MeCN solvent with

time delays up to 6000 ps after photo-excitation. To clearly

display the spectral evolution in different time regimes, the

spectra at early and late picosecond time-scales (correspond-

ing to before and after 10 ps, respectively) are given sepa-

rately in Figure 4. The temporal evolution of the early spec-

tra (Figure 4a) shows a rapid conversion of the initial spec-

trum (the 0.6 ps spectrum) that absorbs strongly at 300 nm

into a spectrum (the 10 ps spectrum) having a broad absorp-

tion with l_maxat ~350 nm. From Figure 4b, it can be seen

that the ~350 nm spectrum decays gradually at later times

and, at the expense of this decay, an additional spectrum

with an absorption maxima at ~300 nm grows in. The obser-

vation of isosbestic points for both the early and late spec-

tral conversions (at ~330 and 320 nm in Figure 4a and b, re-

spectively) indicates there are direct dynamical conversions

between two distinct state/species in each the case. The ki-

netics of these conversion processes can be followed by the

time-dependence of the TA intensity at 350 and 310 nm, re-

spectively. Such results are displayed in Figure 5 together

with exponential fittings of the experimental data points.

Obviously, the temporal change of the 350 nm band (Fig-

Figure 3. DFT optimized structures for the ground and triplet states of

free BDP and the corresponding carbonyl H-bonded complexes.

cy was computed to shift down by ~18 cm^ꢀ1from the free

BDP molecule to the H-bonded BDP molecule while the

frequencies of all the other modes are nearly identical for

both systems. The experimental and calculated frequencies

and the corresponding normal mode descriptions for the ob-

served RR vibrational features (refer to ref. [22]for a de-

tailed list of the BDP vibrational analysis) are provided in

Table S2 in the Supporting Information. It is therefore clear

that the local inter-solute–solvent hydrogen-bonding interac-

tion at the carbonyl site is responsible for the spectral differ-

ence observed in the two solvent systems. This is consistent

with the localized nature of the H-bonding interaction and

the slight structural differences found in the DFT calcula-

Chem. Eur. J. 2007, 13, 2290 – 2305

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

2293

D. L. Phillips et al.

ure 4a) can be fitted by two exponentials with a ~2.3 ps time

constant for the early-time growth and a ~11 ns time con-

stant for the late-time decay. As seen in Figure 4b, the same

corresponding time constants can also fit the early-time

decay and late-time growth of the absorption at ~310 nm.

Figure 4. Fs-transient absorption spectra of BDP at a) early and b) later

picosecond times recorded with 267 nm excitation in neat MeCN solvent.

Figure 5. Time-dependence of the transient absorption spectra for BDP

in neat MeCN at a) ~350 and b) ~310 nm. The solid lines indicate the ki-

netics fitting to the experimental data points using exponential functions.

The inserts show the kinetics at early picosecond time delays.

The ~350 nm (l_max) absorption spectrum (Figure 4, the

spectrum observed at 10 to 50 ps) is clearly from the BDP

triplet (T₁) state. Wirz and co-workers have reported the

BDP triplet absorption spectrum within the 280–450 nm

spectral region.^[15a]The spectrum observed here is consistent

with the reported spectrum but is displayed in a much

broader spectral window. In addition, we note that the BDP

triplet spectrum also closely resembles the triplet absorption

spectrum of acetophenone^[27,28]and the triplet absorption of

the phenacyl group in phenacyl caged compounds.^[3a]This in-

dicates the triplet excitation is localized on the BDP phenac-

yl subgroup. From this assignment and the phenacyl local-

ized nature of the triplet state, the observed ~2.3 ps time-

constant spectral transformation (Figures 4 and 5) can be at-

tributed confidently to intersystem crossing (ISC) conver-

sion from the lowest singlet (S₁) state to the lowest triplet

(T₁) state. The ~2.3 ps ISC time constant matches well the

values of rapid ISC rate reported previously for a wide

range of aromatic carbonyl compounds^[11,22–29]and is consis-

tent with the nearly unit triplet quantum yield reported for

these compounds.^[11,29,30]Therefore the initial spectrum (see

the 0.6 ps spectrum in Figure 4a) can be assigned to the S₁

! S_nabsorption from the BDP phenacyl S₁(np*) state.^[22]

The assignment of the S₁singlet as the precursor state for

ISC conversion is in agreement with a number of studies on

acetophenone and other closely related compounds showing

there is efficient ISC from the S₁state into the triplet

state.^{[22–24,26,28–40]}The BDP S₁absorption spectrum observed

here has not been previously reported probably due to its

short lifetime (~2.3 ps). An alternative assignment of the in-

itial TA spectra to the 267 nm photo-populated S₃

state^[11,19,22]and that the conversion occurs with the IC step

being rate-determining (~2–3 ps) seems rather unlikely

since previous relevant studies on the related aromatic car-

bonyl compounds have reported consistently a <100 fs life-

time for the S₃state^[11,23,32]and up to hundreds of femtosec-

ond lifetime for the intervening S₂state (weakly allowed

pp* state).^[32–35]This implies a rapid IC from the S₃into S₁

state occurring with an upper-limit time constant of hun-

dreds of femtoseconds. Considering the ~200 fs time-resolu-

tion of our TA setup, it appears reasonable to infer that IC

process involved in the 267 nm excited BDP might be domi-

nated by direct conversion from the S₃into S₁state that hap-

pens within tens to hundreds of femtosecond time scale,^[11a]

too fast to be easily detected by our present TA measure-

ment. The S₁spectrum is similar to the singlet absorption of

the phenacyl subgroup of the pHP caged phototrigger com-

pounds^[11b]implying again the phenacyl localized excitation

in the S₁state. Consistent with the phenacyl localized ab-

sorption revealed by the RR spectrum, the observation of a

2294

www.chemeurj.org

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2007, 13, 2290 – 2305

Time-Resolved Resonance Raman Studies

FULL PAPER

similar characteristic feature of the phenacyl associated sin-

glet and triplet excited state spectra and the correlated ISC

rate indicates the predominance of the phenacyl subgroup

being responsible for the photo-excitation and early photo-

physical processes.

is featured by a ~570 nm absorption, rather than the

~300 nm absorption in Figure 4b. It is thus clear that cation

2 proposed previously to be an essential intermediate for

the BDP photocyclization reaction is not actually involved

in the cyclization pathway but rather is present as a key in-

termediate responsible for the BDP deprotection–solvolysis

process occurring exclusively in the largely water containing

solvent.

The late-time picosecond spectrum with the predominant

absorption at ~300 nm (see Figure 4b) displays the charac-

teristic structured UV absorption of the PBF cyclization

product (see Figure 1). As displayed in Figure S2 (Support-

ing Information), this is shown by a direct comparison be-

tween the ~300 nm transient absorption spectrum (with the

absorption due to the BDP triplet state subtracted) with the

PBF UV spectra (spectrum g in Figure 1). Therefore the ob-

served ~11 ns spectral conversion (Figures 4b and 5b) re-

flects a direct BDP triplet state reaction leading to the PBF

production. This rather fast and straightforward BDP triplet

to PBF cyclization is a remarkable reaction since it requires

not only the net elimination of the phosphate leaving group

and the 2’-benzylic proton but also a coupled spin flip. It is

therefore necessary to be cautious about the identity of our

late picosecond ~300 nm transient spectrum. In this context,

we note that the spectrum is consistent with the ~300 nmns

LFP spectra (at 30 ns time delay) reported for BDP in

MeCN by Givens and co-workers.^[15a]In their work, the

~300 nm absorption has been assigned specifically to the

PBF photoproduct based on a dynamic measurement for

both the BDP absorption and emission spectra. This togeth-

er with the close resemblance in the absorption spectral pro-

file makes it quite certain for the tentative assignment of

our late ~300 nm spectrum to the ground state of PBF. An

alternative assignment of this absorption to other intermedi-

ates that may possibly occur in the triplet cyclization path-

way appears implausible. Such transient species include the

triplet state of BDP, the radical 1 and cation 2 intermediates

possibly formed as

Fs-TA measurements for BDP in 75% H₂O/25% MeCN

solvent: Figures 6 and 7 show the fs-TA spectra and the cor-

responding kinetics observed for BDP in a 75% H₂O/25%

MeCN solvent. Obviously, the spectral features and the cor-

related conversion dynamics observed in the early-time pi-

cosecond spectra (10 ps and before in Figure 6a and insets in

Figure 7a and b) parallels those observed in neat MeCN sol-

vent. This indicates that the temporal evolution of the corre-

sponding TA spectrum reflects the very similar ISC conver-

sion (with a ~2.7 ps time constant as obtained by an expo-

nential fit shown in the inserts in Figure 7) from the S₁(np*)

G

singlet state (represented by the 0.35 ps spectrum in Fig-

ure 6a) to the T₁triplet (represented by 10–30 ps spectra in

Figure 6a and b) excited state. Considering the remarkable

differences in polarity and H-bonding ability between

MeCN and water, the spectral and dynamic resemblance in

the photophysics observed in the two solvent systems sug-

gests there is only a modest influence of solvent property on

these processes. This is consistent with the nearly solvent-in-

dependent ISC dynamics reported recently in our combined

fs-time-resolved fluorescence (fs-TRF) and ps-TR³study on

pHP caged phototriggers.^[11]

However, the late time spectral evolution recorded in the

largely water containing solvent displays distinctly different

features from that observed in MeCN (Figure 4b and 5b). It

a result of the homolysis and heterolysis, respectively, of

[15,17–21]

ꢀ

the keto a-C O bond.

DFT computation at the

random-phase approximation (RPA) with the B3LYP

method and 6-311G** basis set found that the triplet state

BDP has its strongest absorption at ~437 nm (with an oscil-

lation strength f of ~0.6) and a modest absorption at

~338 nm (with f of ~0.2) (the details of the calculated RPA

results are listed in Table S3 in the Supporting Information).

Obviously, these absorption positions are not in accordance

with the ~300 nm absorption observed here in the TA spec-

tra and this implies the BDP triplet does not account for the

experimental spectrum. For radical 1, from the lack of the

decarboxylation for the acyloxy radical, previous studies on

benzoin and a series of methoxy substituted benzoin esters

have suggested only a possible minor importance of the ho-

molysis in the deprotection pathway.^[17–20]As for cation 2,

our following result and discussion indicates evidently that it

Figure 6. Fs-transient absorption spectra of BDP at a) early and b) later

picosecond times recorded with 267 nm excitation in 75% H₂O/25%

MeCN.

Chem. Eur. J. 2007, 13, 2290 – 2305

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

2295

D. L. Phillips et al.

suggests that this band results from a species with the same

identity, that is, the PBF product. Obviously, the ~570 nm

band is from an alternative species, neither the BDP triplet

state nor the PBF product. This band correlates well with

the ~570 nm absorption band reported in a BDP nanosec-

ond LFP spectrum (in aqueous solution with 4% MeCN) by

Givens and co-workers.^[15a]Based on DFT energy calcula-

tions, Givens and co-workers have associated this band with

the triplet cation resulted from direct BDP triplet heterolyt-

ic cleavage of the phosphate leaving group.^[15a]To confirm

further this assignment and to track the associated reaction

pathway so as to unveil its relevance to the solvolytic BZ

product, ns-TR³experiments combined with DFT results for

this intermediate are presented in the next section.

We note that the ~11 and ~15 ns BDP triplet decay times

determined here in neat MeCN and H₂O/MeCN, respective-

ly, are very close to the 10–25 ns triplet lifetimes reported by

Givens and co-workers based on the indirect diffusion con-

trolled triplet quenching study in these two kinds of solvent-

s.^[15a]The fs-TA results presented here provide, however, ad-

ditional and unequivocal evidence to indicate that the triplet

is i) the reactive precursor to the deprotection–cyclization

reaction in MeCN and ii) the common precursor to the de-

protection–cyclization and heterolytic deprotection leading

to photosolvolysis (below) in the largely water containing

solution. In the case of the largely water containing solvent,

it is unlikely that the observed later time 300 nm band is

from the same transient species responsible for the ~570 nm

band even though both bands display similar developing dy-

namics (Figure 5b, 6b and c). This is because the species re-

sponsible for the ~570 nm band was found to lead eventual-

ly and solely to the solvolytic BZ product (see next section

for details). This in turn leads to the product distribution

shown in Equation (2). In addition, the photochemical re-

sults requires the ~300 nm band being assigned to the PBF

product.

Figure 7. Time-dependence of transient absorption spectra for BDP in

75% H₂O/25% MeCN at a) 370, b) 305 and c) 570 nm. The solid lines in-

dicate the kinetics fitting to the experimental data points using exponen-

tial functions. The inserts show the kinetics at early picosecond time

delays.

Ns-Time-resolved resonance Raman (ns-TR³) spectroscopy

and comparison with DFT calculation results

Ns-TR³study with a 532 nm probe wavelength: According

to the fs-TA spectrum recorded in H₂O/MeCN (Figure 6b),

a 532 nm probe wavelength was selected for the ns-TR³ex-

periments to detect the intermediate responsible for the

~570 nm transient absorption band. Figure 8 displays repre-

sentative ns-TR³spectra of BDP in 75% H₂O/25% MeCN

acquired with time delays between the pump and probe

pulses varying from 0 to 5000 ns. The temporal evolution of

the spectra reveals a dynamic transformation from an early

species (manifested by the 10 ns spectrum in Figure 8) into a

late one (represented by the 200 ns spectrum in Figure 8)

that decays on the microsecond time scale. By fitting a typi-

cal Raman band area with a Lorentzian function, Figure 9

displays the intensity time-dependence of the early species

(represented by the 1560 cm^ꢀ1band area) and the later spe-

cies (represented by the 1626 cm^ꢀ1band area). The kinetics

shown in Figure 9 can be fit well by exponential decay and

can be seen from Figure 6 b that, as the BDP triplet spec-

trum decays, there are the simultaneously development of

two absorption bands. Besides the common ~300 nm ab-

sorption band also observed in neat MeCN, an additional

band with a broad absorption and a l_maxat ~570 nm is seen

exclusively in this largely water containing solvent. The

growth of the two latter bands are accompanied by isosbes-

tic points at ~320 and 510 nm, respectively. The kinetics of

the triplet decay and the development of the two new bands

are followed by temporal variations in the TA intensity at

typical wavelengths of ~370 (Figure 7a), 305 (Figure 7b) and

570 nm (Figure 7c), respectively. Exponential fitting of these

series of experimental data points results in a ~15 ns conver-

sion time from the triplet spectrum into the ~300 and

~570 nm absorption bands. The analogy of the ~300 nm

band seen in the water mixed solvent to that in neat MeCN

2296

www.chemeurj.org

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2007, 13, 2290 – 2305

Time-Resolved Resonance Raman Studies

FULL PAPER

leaving group and the desyl cage are the most plausible in-

termediates for the early and late species, respectively, ob-

served in the ns-TR³spectra (Figure 7).^{[15a,17,20a,21]}To investi-

gate this probable assignment, (U)B3LYP/6-311G** based

DFT calculations were done for the triplet and singlet cat-

ions. These calculations located more than one minima for

both the triplet and singlet cations and found that the lowest

energy of the singlet conformation is ~22.1 kcalmol^ꢀ1e more

stable than the lowest triplet counterpart, in agreement with

the singlet state as being the ground state of this cation spe-

cies.^[15a]

The calculations for the triplet cation located two minima

with the carbonyl lying in the benzylic ring plane but being

s-cis and s-trans, respectively, to the ring moiety. The s-cis

conformation was calculated to be ~4.7 kcalmol^ꢀ1(including

ZPE) lower in energy than the s-trans one. This together

with the reported relatively large rotation barrier for conver-

sion of the s-cis to s-trans-conformation^[15a]suggests that the

s-cis isomer would be predominantly responsible for the

early time species seen in the ns-TR³spectra. The major

structural parameters of the optimized s-cis triplet cation

are listed in Table S4 in the Supporting Information and the

structure is shown in Figure 10. Figure 11 displays a compar-

ison and tentative correlation of the 10 ns TR³spectrum to

the calculated Raman spectrum given by the s-cis triplet

cation. It can be seen that the computed spectrum reprodu-

ces reasonably well the experimental spectrum. This helps

to confirm its assignment to the triplet cation. Provisional

vibrational assignments to the observed Raman features

based on a direct comparison between the two spectra are

listed in Table S5 in the Supporting Information.

Figure 8. ns-TR³spectra of BDP in 75% H₂O/25% MeCN obtained with

267 nm excitation and a 532 nm probe wavelength at various time delays.

In relation to the BDP ground state conformation (see

Figure 3 and the structural parameters in Table S1 in the

Supporting Information), the major structural difference dis-

played in the triplet cation is due to the conformation asso-

ciated with the benzyl and carbonyl subgroups. For example,

*

Figure 9. Time-dependence of the early (&) and the later species ( ) for

BDP in 75% H₂O/25% MeCN observed in the ns-TR³spectra displayed

~

in Figure 8, and the kinetics for formation of the BZ product ( ) as ob-

served in the ns-TR³spectra displayed in Figure 12. The solid lines indi-

cate the kinetics fitting based on Equations (3) and (4) in the text.

ꢀ

the benzyl C8 C9 bond and the bridging C7 C8 bond

become significantly shorter (by ~0.09 and ~0.13 , respec-

tively) suggesting that the two bonds gain noticeable

double-bond nature in the triplet cation. This is also consis-

tent with the lengthening (by ~0.03 ) and shortening (by

~0.02 ), respectively, of the benzyl ring shoulder and

growth components as displayed by the solid lines in the

figure. From the exponential fitting, a time constant of 102ꢁ

6 ns was estimated for decay of the early intermediate and a

growth component with a similar time constant combined

with a ~230 ns decay time was found for the late one. Fur-

ther ns-TR³measurements reveal that the lifetime of the

early intermediate becomes longer upon a nitrogen purge of

the sample system (the corresponding ns-TR³spectra are

given in Figure S3 in the Supporting Information). Under an

oxygen purge of the sample, the early time species lifetime

becomes shorter. In contrast, the lifetime of the late species

is hardly affected by nitrogen or oxygen purging. The sensi-

tivity of the early species lifetime to the oxygen concentra-

tion suggests a triplet nature for this species while the insen-

sitivity observed for the later species implies it probably has

a singlet character.

ꢀ

center C C bond lengths. The benzylic ring of the triplet

cation gains some quinoidal character in comparison to its

ground state counterpart. The carbonyl bond becomes

slightly longer by ~0.05 in the triplet cation and changes

its orientation from being coplanar with the adjacent ring in

the ground state to coplanar with the benzyl moiety in the

triplet cation. This structural variation in the triplet cation

suggests there is an increased extent of the benzyl conjuga-

tion in the triplet cation relative to the BDP ground state.

In other words, the delocalization of the p electrons occur at

the benzyl ring moiety in the ground state but extends to ad-

ꢀ

ditionally involves the connecting C7 C8 and C8 C9 bonds

as well as the carbonyl group in the triplet cation. The struc-

ture of the triplet cation and its differences from the BDP

ground state are consistent with the frequency changes of

The benzoin triplet and singlet cation produced by hetero-

lytic cleavage of the C O bond connecting the phosphate

ꢀ

Chem. Eur. J. 2007, 13, 2290 – 2305

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

2297

D. L. Phillips et al.

(1028 cm^ꢀ1).^[22]However, there

is no specific vibrational feature

with a predominant carbonyl

stretching motion in the triplet

cation. According to the calcu-

lations, this is due to mixing of

the carbonyl stretching motion

with the benzylic ring carbon–

carbon stretching vibrations.

We note that most features ob-

served in the ns-TR³spectra are

due to vibrational motions asso-

ciated with the benzylic ring

and the nearby C-C chain (see

Table S3, Supporting Informa-

tion). This suggests that the ex-

tensively conjugated benzyl

moiety included subgroups are

the major chromophore probed

by the 532 nm ns-TR³measure-

ments and mainly account for

the ~570 nm absorption band

seen in the fs-TA spectrum

(Figure 6b).

Figure 10. DFT optimized structures for the free triplet and the singlet cation and their respective carbonyl H-

bonded complexes.

Calculations for the singlet

cation located three minima with energy differences within

~0.002 kcalmol^ꢀ1. These minima have generally similar

structures that differ mainly by modest variations in the car-

ꢀ

bonyl connecting C6 C7 and C7 C8 bond lengths (varying

by ~0.04 ). The three isomers exhibit almost identical

structural parameters for the carbonyl connecting ring and

the benzyl moiety; they all have the carbonyl group lying

(nearly) at the connected ring plane and with a similar rela-

tive orientation of this plane with the benzyl plane. The iso-

mers also give analogous calculated Raman spectra since

the computed Raman features with noticeable intensities

are predominantly the vibrational motions associated with

the subgroups (mainly the two rings) with the common

structural parameters. As a representative structure, the con-

formation of one such isomer is displayed in Figure 10 and

the corresponding structural parameters are listed in

Table S4 in the Supporting Information. A comparison be-

tween the corresponding computed Raman spectrum and a

late-time ns-TR³spectrum (at 200 ns) is showed in Figure 11

together the spectra for the triplet cation. The calculated

frequencies and tentative assignments of the experimental

features made based on correlation of the experimental to

the calculated spectrum are presented in Table S6 in the

Supporting Information. It is obvious from Figure 11 that

the experimental features seen in the ns-TR³spectrum are

reproduced reasonably well by the calculations and this

strongly supports assigning the spectrum to the singlet

cation. The lack of the calculated carbonyl stretching feature

at ~1700 cm^ꢀ1in the experimental spectrum is within ex-

pectations since the experimental spectrum was obtained

with the resonance enhancement effect while the calculated

one corresponds to a normal Raman spectrum. However,

Figure 11. Comparison of the experimental and DFT calculated Raman

spectra for the singlet and triplet cation. See the text for details.

the relevant vibrational modes observed in the two species.

For instance, in agreement with the increased quinoidal

character, the ring center C–C stretching of the benzylic

ring displays a ~31 cm^ꢀ1frequency up-shift on going from

the ground state (1595 cm^ꢀ1)^[22]to triplet cation (1626 cm^ꢀ1).

ꢀ

In accordance to the noticeable shortening of the C7 C8

bond length, the vibrational feature contributed mainly by

the stretching of this bond shifts up by ~158 cm^ꢀ1in the trip-

let cation (1186 cm^ꢀ1) relative to the ground state

2298

www.chemeurj.org

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2007, 13, 2290 – 2305

Time-Resolved Resonance Raman Studies

FULL PAPER

considering the calculated result that the isomers have very

close energy and give very similar Raman spectra, it appears

that they may all contribute significantly to the experimental

spectrum. This may also be a reflection that the singlet

cation has a flexible structure in terms of the specific bond

identify the intermediate observed in the ns-TR³spectra.

The temporal evolution of the spectra in Figure 12 shows

the gradual growth of a series of features (mainly the strong

feature at 1605 cm^ꢀ1and some weaker ones at ~1692 and

~1181 cm^ꢀ1) at the late nanosecond times. This reflects the

formation of a new species in this time regime. Clearly, the

newly developed spectrum closely resembles the RR spec-

trum of the authentic BZ, which indicates that the corre-

sponding species can be attributed confidently to the BZ

product. Thus, the temporal change of the transient BZ

band area seen in the ns-TR³spectra represents directly the

formation dynamics of the BZ product. The time-depend-

ence of the Lorentzian fitted ~1605 cm^ꢀ1band area of the

transient BZ is displayed in Figure 9 together with the kinet-

ics obtained for the T⁺and S⁺species based on the ns-TR³

measurements using the 532 nm probe wavelength. The gen-

eral correlation of the kinetics obtained for the three species

ꢀ

length for the bridging C C bonds. It may be possible that a

further search of the potential surface may locate additional

minima with a generally similar geometry but slightly differ-

ꢀ

ent bridging C C bond lengths from those seen in the iso-

mers already located. With the carbonyl group being far out

of the benzylic ring plane and with the substantial single

ꢀ

bond feature of the C7 C8 bond (Table S3, Supporting

Information), the singlet cation has a lesser extent of the

benzyl related conjugation than that of the triplet cation.

This might account for the structural flexibility and could

also be relevant to the subsequent reaction (see below).

k₁

k₂

Ns-TR³study on the formation dynamics for the photo-sol-

volysis product: To further explore the solvolysis reaction

pathway and to determine the formation dynamics of the

solvolytic product BZ, ns-TR³measurements with 267 nm

pump and 267 nm probe wavelengths were done for BDP in

75% H₂O/25% MeCN. Figure 12 displays the acquired ns-

TR³spectra recorded with the selected delay time up to

~2 ms. A RR spectrum from an authentic sample of BZ ob-

tained under the same condition and by using 267 nm excita-

tion is also shown in the Figure to compare with and to help

suggest a consecutive dynamics of T⁺

S⁺

BZ with the

! !

k₁and k₂representing the rate constants for the cation ISC

and the solvolytic addition step, respectively. In this reaction

scheme, the temporal variation of the T⁺concentration [T⁺

]follows a single exponential decay with the time constant

equal to the reciprocal of k₁, while the time-dependence of

the S⁺and BZ concentrations ([S⁺]and [BZ], respectively)

are expressed, respectively, by Equations (3) and (4), where

[T⁺]_orefers to the initial concentration of the T⁺precursor.

k₁

½S^þ ¼ ½T^þ

½e^{ꢀk t}ꢀe^{ꢀk t}

1

2

ð3Þ

ð4Þ

⁰k₂ꢀk₁

½T^þ

0

½k₂ð1ꢀe^{ꢀk t}Þꢀk₁ð1ꢀe^{ꢀk t}Þ

1

2

½BZ ¼

k₂ꢀk₁

Provided that the ISC from T⁺to S⁺is the predominant

channel to deactivate the T⁺population, the k₁in Scheme 1

can be taken as the decay rate of T⁺determined by the

532 nm probed ns-TR³measurement (k₁=1/t with t=

102 ns). With this value for k₁and based on Equations (3)

and (4), the least squares simulation has been done to fit the

kinetics result observed for the [S⁺]and [BZ]concentra-

tions. As displayed in Figure 9, this gives a reasonable fit to

the experimental data and results in a rate of k₂correspond-

ing to a time constant of ~210 ns.

DFT calculations for the BDP triplet state and H-bonding

configurations

Our fs-TA results (Figures 3 and 5) indicate clearly that the

BDP triplet is the reactive precursor to the photo-deprotec-

tion reaction. It is therefore necessary to gain information

about the structural and electronic properties of the BDP

triplet. Our attempt to obtain a ns-TR³spectrum for the

triplet state was hindered by overwhelming fluorescence

from the PBF product. We have then performed open shell

DFT calculations for this state and the computed energy

and the optimized structural parameters for the triplet state

are listed in Table S1 in the Supporting Information and the

optimized triplet conformation is displayed in Figure 3.

Figure 12. ns-TR³spectra of BDP in 75% H₂O/25% MeCN after 267 nm

laser excitation and using a 267 nm probe wavelength with the time

delays varying from 0 ns to 2000 ns. A resonance Raman spectrum of an

authentic sample of BZ recorded with 267 nm excitation is displayed at

the top.

Chem. Eur. J. 2007, 13, 2290 – 2305

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

2299

D. L. Phillips et al.

above for the BDP triplet based on the fs-TA analysis. The

calculated spin densities together with the triplet structure

suggest a mainly np* nature triplet with an electron-defi-

cient carbonyl oxygen. This is consistent with the triplet fea-

ture reported for unsubstituted phenyl alky ketones.^{[27,28,30,31]}

To examine and evaluate the possible influence of the sol-

vent–solute H-bonding associated solvent effect, we have

done (in parallel with the BDP singlet ground state compu-

tations) DFT calculations considering the carbonyl H-

bonded configurations for the BDP triplet (T₁) as well as

the triplet cation (T⁺) and the singlet cation (S⁺) species. In

general, these calculations suggest that the presence of the

H-bonding interaction can lead to energy stabilization as

well as modest structural modifications. The optimized struc-

tural parameters and the stabilization energies obtained for

the H-bonded conformations are summarized in Tables S1

and S4 in the Supporting Information together with the cor-

responding results calculated for the free conformations.

The structures for the DFT optimized H-bonded complexes

are displayed in Figure 3 and 10 together with the free coun-

terparts. It can be seen that, like the case of BDP ground

state, whilst the H-bonding does not cause significant

changes in the bond lengths, it does introduce noticeable

variations in the relative orientations among the subgroups

of the benzyl moiety, the carbonyl group and the carbonyl

connected ring. This is particularly true for the BDP T₁state

and the S⁺species. From the calculated energy, the H-bond-

ing stabilization energies are estimated to be ~4.8 kcalmol^ꢀ1

for T₁and 6.9 and 7.2 kcalmol^ꢀ1for the T⁺and S⁺species,

respectively. The magnitudes of these energies are close to

or even stronger than the corresponding BDP ground state

(S₀) stabilization energy (5.4 kcalmol^ꢀ1). Since our RR ex-

periment reveal that the BDP S₀state is present in the com-

plex form with the carbonyl H-bonding to surrounding two

water molecules in 75% H₂O/25% MeCN. This implies

that, in the same solution, the BDP T₁state and the T⁺and

S⁺species should also be in the corresponding H-bonded

forms.

Scheme 1. Deactivation and deprotection reaction pathways of 267 nm

excited BDP in neat MeCN and 75% H₂O/25% MeCN.

From the computed total energy (including ZPE) for the

BDP triplet state and the ground state, a triplet state energy

of ~64.4 kcalmol^ꢀ1can be estimated. This value is compara-

ble to the ~73 kcalmol^ꢀ1triplet energy as given by a BDP

phosphorescence spectrum.^[1,10,15a,c]In terms of the structural

changes in the triplet state, the triplet and singlet structures

differ mainly by the conformation of the phenacyl subgroup.

For example, upon going from the ground state to the triplet

state, the carbonyl bond lengthened by ~0.1 and twists

out of the phenayl plane by ~128. The carbonyl connecting

Summary of the photodeprotection pathways of BDP in

MeCN and 75% H₂O/25% MeCN: Our fs-TA results ob-

tained after the 267 nm photolysis of BDP in neat MeCN

and in 75% H₂O/25% MeCN indicate that the BDP triplet

is formed, barely dependent on solvent property, with a ~2–

3 ps time constant by ISC from the S₁excited state. A com-

parison of this ISC time with the dynamics determined for

the photochemical reactions (represented initially by the

~10–15 ns triplet decay time constant, Figures 5 and 7) re-

veals that this state is the common precursor to the subse-

quent photochemical events in both of the two solvent sys-

tems. This is consistent with the corresponding results re-

ported by Givens and co-workers.^[15a]Corroborated with the

phenacyl located singlet and triplet excitation as indicated

by our RR and fs-TA results, the photophysics dynamics ob-

tained here for the deactivation of the 267 nm excited S₃-

ꢀ

C6 C7 and C7 C8 bond shorten by ~0.07 and 0.05 , re-

spectively, and the phenacyl ring gains a little quinoidal fea-

ture from slight increases and decreases of bond lengths for

ꢀ

the ring shoulder C C bonds and the center C C bonds, re-

spectively. On the other hand, the geometric parameters re-

lated to the benzyl group, the phosphate leaving group as

well as the relative orientation of the benzylic ring with the

carbonyl adjacent ring remain quite similar in the two states.

This phenacyl localized structural variation in the BDP trip-

let state versus the ground state coincides and is consistent

with the phenacyl localized excitation feature inferred

AHCTREUNG

2300

www.chemeurj.org

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2007, 13, 2290 – 2305

Time-Resolved Resonance Raman Studies

FULL PAPER

observed generally for most aromatic carbonyl alkyl ketones

in the solution phase.^[11,23–29]This combined with the very ef-

ficient IC (with a time constant within hundreds of femto-

seconds) from the S₃and S₂state (pp*, formed due to

~250–300 nm excitation)^[11,32–35]to the S₁state (np*, formed

due to >300 nm excitation) has been suggested to account

for the high and substantially excitation wavelength inde-

pendent triplet formation yield for these compounds.^[29,30]

Due to the BDP deprotection proceeding on the triplet

manifold, the insensitivity of the triplet yield to the excita-

tion wavelength suggests the deprotection efficiency to be

also independent of the excitation wavelength. Indeed, as

reported by Givens and co-workers, the desyl deprotection–

cyclization reaction actually occurs with a similar efficiency

over a broad range of excitation wavelength ranging from

the deep UV to the far UV region.^[15a]This indicates that

the desyl photo-liberation does not necessarily require

tures and the electronic character. This suggests they may

also display similar triplet reactivity. For the latter com-

pounds, either the intermolecular triplet carbonyl hydrogen

abstraction (from the solvent) or the intramolecular carbon-

yl hydrogen abstraction pathway has been identified as the

predominant initial photochemical step for the photochemis-

try. From this and the ~11 ns BDP triplet cyclization time

constant, it may be conceivable that the BDP reaction could

also originate from an intramolecular triplet carbonyl ab-

sorption of the benzylic 2’ hydrogen. Further theoretical cal-

culations could be helpful to help determine the conforma-

tional details of the transition state and provide a more in

depth molecule-level description to this intriguing reaction

pathway.

In the largely water solvent, in addition to the deprotec-

tion–cyclization reaction, the triplet deactivation leads also

to a heterolytic dissociation to liberate the phosphate leav-

ing group in a competing and branching manner. The occur-

rence of the heterolytic cleavage is indicated by and con-

firmed by our combined ns-TR³(with a 532 nm probe wave-

length) and DFT results that identify the triplet cation (T⁺)

as the initial heterolytic product. The temporal evolution of

the ns-TR³spectra in conjunction with the DFT calculations

also indicate that ISC of the T⁺species into the S⁺species

with a ~102 ns time constant is followed by decay of the S⁺

within the hundred nanoseconds time scale. Correlation of

the S⁺decay kinetics with the formation kinetics of the sol-

volytic product BZ as observed by the 267 nm probe ns-TR³

results indicates there is a sequential pathway for the solvo-

lytic addition reaction.

An overall reaction scheme consistent with our previous

result is displayed in Scheme 1 for the photodeprotection

and solvolysis pathway of BDP in the two solvent systems.

According to this scheme, the triplet decay rate in the

water mixed solvent is the sum of the reaction rate for the

cyclization (k_cyc) and heterolysis (k_het). Therefore, the slight-

ly slower triplet decay rate in the largely water solvent

(~15 ns time constant) than that in neat MeCN (~11 ns time

constant) implies a noticeably slower k_cycin the former than

later solvent. Considering the predominance of the solvoly-

sis BZ product and assuming a ~25% of the chemical yield

of PBF in the water mixed solvent, the k_cycin the mixed sol-

vent can be estimated roughly to be at least four times

smaller in magnitude than that in MeCN.

The BDP ground state RR results together with the DFT

results considering the carbonyl H-bonding interaction with

the solvent water molecule(s) shows there is noticeable

energy stabilization caused by the H-bonding. This suggests

that, in the highly aqueous solution, not only the BDP

ground state species, but also the other relevant intermedi-

ates involved in the excited state processes are present in

the carbonyl H-bonded forms. At least as part of the first

solvent shell, this H-bonding configuration may affect the

triplet state reaction and contributes to the difference in the

triplet deprotection dynamics and pathway observed in the

largely aqueous solution versus neat MeCN where the H-

bonding interaction is absent.

direct S₁(np*) excitation, a condition that has been speculat-

G

ed to be essential for the desyl deprotection and cyclizatio-

n.^{[17,18a,19,20a,21]}

The phenacyl localized nature of the BDP triplet state

combined with the triplet identity as the photoreaction pre-

cursor indicates that the BDP photo-deprotection is primari-

ly the photochemistry of the phenacyl triplet state rather

than the photodissociation chemistry associated with the sin-

glet state of the benzylic subgroup^[36–38]as proposed in some

of the previous studies.^{[17,19,20a,21]}

In neat MeCN, the decay of the triplet leads to an appa-

rently single-step phosphoric acid elimination and cycliza-

tion reaction yielding the PBF product with ~11 ns time

constant. Such a seemingly straightforward and efficient

BDP triplet to PBF ground state transformation is an amaz-

ing reaction. For the reaction to occur, it requires extensive

electronic reorganization and conformational rearrangement

such as achieving an overall planar arrangement (as in the

PBF product)^[39]from the highly non-planar triplet confor-

mation (see Figure 3 and Table S1, Supporting Information);

it also needs an ISC at some point along the reaction path-

way to bring the system from the triplet manifold to the sin-

glet ground state. Considering the dependence of the depro-

tection–cyclization yield on the leaving group ability,^[21]it is

reasonable to infer that the reaction could create a charge

in the transition state such that the departing leaving group

may act as (to a certain extent) the hydrogen-abstracting

base to facilitate the deprotonation of the 2’ carbon so as to

assist the cyclization reaction. However, the small kinetics

and yield dependence of the reaction to the change in the

solvent polarity (from e=2.3 for benzene to e=35.9 for

MeCN)^[1,10,15]suggests that the transition state would be

only modestly polarized during the liberation of diethyl

phosphoric acid and the apparently concurrent cyclization

process.

It is interesting to note at this point that the phenacyl lo-

cated np* nature triplet state of BDP is analogous to the

triplet of phenacyl caged phenylacetate (PPA),^[3a]ortho-

methyl substituted phenacyl esters,^[9]acetophenone and ace-

tophenone derivatives^[27,28]in terms of both the spectral fea-

Chem. Eur. J. 2007, 13, 2290 – 2305

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

2301

D. L. Phillips et al.

Heterolysis–solvolysis and deprotection–cyclization reac-

tions in a largely water containing solvent

BDP triplet heterolysis observed here may follow a similar

mechanism where a significant degree of polar solvent reor-

ientation is required to facilitate and accommodate the

charge transfer so that the appropriate solvation of the trip-

let by the solvent ionizing and polarity ability becomes one

of the most important factors in influencing the height of

the relevant reaction barrier. Moreover, the stability of the

heterolysis produced T⁺intermediate and the good nucleo-

fugal property of the diethyl phosphate leaving group^[52]are

also of equal importance to jointly account for the occur-

rence of the efficient heterolytic deprotection in the largely

water containing solvent. With a pK_avalue of 1.4 for the

conjugate acid,^[11,48]the diethyl phosphate leaving group is a

good nucleofuge; and photoheterolytic cleavage has been

reported for many systems bearing the phosphate moiety as

a leaving group.^[11,47–49]

According to the DFT calculated structure and Mulliken

charge distribution, the stablization of the incipient T⁺car-

bocation is due to the resonance conjugation with the adja-

cent p system. Thus, the positive charge formed in the heter-

olysis reaction is not localized at the benzylic carbon under-

going the ionization but is instead delocalized into the ad-

joining p-electron system. In largely polar environments

such as a largely water containing solvent, the high stability

of T⁺and the counterpart phosphate anion not only enable

the heterolysis to compete with the deprotection–cyclization

reaction but also facilitate the geminate ion pair separation

so as to compete with the reverse ion recombination pro-

cess. However, the conjugation-induced stabilization of T⁺

may at the same time increase the “intrinsic barrier” for the

T⁺to be trapped by the solvent for undergoing the solvolyt-

ic nucleophilic addition.^[53,54]This may, on the other hand,

constitute one of the important reasons for the observed

stepwise BDP solvolysis scheme and a rationale for the nec-

essary involvement of the S⁺species as the immediate nu-

cleophilic addition precursor in the solvolysis pathway. Our

calculations suggest a lesser extent of conjugation in S⁺

than in T⁺and a substantial positive charge distribution on

the S⁺benzylic carbon that makes this carbon well suited

for solvent water nucleophillic attack leading to the forma-

tion of the BZ product.

Extensive studies have shown that the kinetics and products

of aromatic ketone photoreaction may be altered significant-

ly in water compared with organic solvents.^{[11–13,15a,18a,23,40–44]}

This has been attributed generally to the solvent effect asso-

ciated with the high polarity, high ionizing and high H-bond-

ing ability of water. In the case of BDP, this solvent effect

may mainly manifest itself as follows: i) by stabilization of

the closely lying pp* state and destabilization of the lowest

np* state,^[23,40–45]the high polarity and high H-bonding ca-

pacity of water could introduce increased pp* character into

the original np* dominated BDP triplet state. This in turn

could lead to a decreased reactivity for the np* based reac-

tions (the deprotection–cyclization). ii) As indicated by our

DFT results on the carbonyl H-bonded complex, the incor-

poration of the H-bonding interaction can alter the triplet

conformation especially in terms of the relative orientation

among the two ring planes and the carbonyl group. This

could result in a conformational effect and influence the

triplet reaction. iii) The high ionizing ability combined with

the high polarity of water provides a favorable environment

to help drive and encourage heterolytic dissociation.^[11,46–49]

The three factors could all contribute to varying degree to

the different reaction pathways observed for BDP in the

largely water containing solvent compared to in the neat

MeCN solvent. However, the i) factor appears to be of

minor importance since our fs-TA result show that, in either

75% H₂O/25% MeCN or in pure MeCN, the BDP triplet

exhibits a similar absorption band with the l_maxat ~350 nm

which is characteristic of an np* character triplet.^[3,27,28]

A

typical pp* character triplet state for aromatic ketones

show an absorption with the l_maxat ~400 nm.^[11b]The quali-

tatively similar np* triplet character in the two solvent sys-

tems is also supported by the fact that the BDP triplet de-

protection–cyclization reaction also occurs to a significant

extent in the largely water containing solvent. The slower

k_cycin the water mixed solvent than in the neat MeCN sol-

vent may reflect the significant sensitivity of the deprotec-

tion–cyclization reaction to conformational factors. This ap-

pears fairy reasonable since this one-step reaction does ne-

cessitate a high degree of structural reorganization (see

above section). Therefore, the H-bonding induced triplet

geometric change combined with the bulky H-bonded triplet

conformation (involving at least two H-bonded water mole-

cules at the carbonyl site) may constitute a steric factor

retarding the required structural rearrangement so as to

increase the reaction barrier height leading to a decrease

For the branching mechanism of the BDP triplet depro-

tection–cyclization and heterolysis–solvolysis in the water

mixed solvent, the relative yield of the PBF cyclization

product and the BZ solvolysis product may reflect the rela-

tive efficiency of the two competing reaction channels. This

relative yield of PBF and BZ would be expected to be gov-

erned by the relative barrier heights correlated with the re-

spective reaction pathways. In this context, the predomi-

nance of the BZ product over PBF product observed in this

solvent may imply there is a substantially lower triplet reac-

tion barrier associated with the heterolysis–solvolysis reac-

tion than with the deprotection–cyclization reaction. This is

within expectations since the heterolysis reaction is general-

ly associated with a highly polar transition state whereas the

transition state correlated with the cyclization has been in-

ferred to be modestly polar, which allows the high solvent

in k_cyc

.

The additional opening of the heterolysis channel for

BDP in the largely water containing solvent is obviously rel-

evant to the solvent effect iii). Solvent-assisted intramolecu-

lar dissociative charge transfer from the phenacyl triplet p*

orbital to the bond-breaking s* orbital has been proposed

to account for b-heterolysis of pHP caged phototriggers^[11,12]

and of radical anions of a-aryloxyacetophenones.^[50,51]The

2302

www.chemeurj.org

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2007, 13, 2290 – 2305

Time-Resolved Resonance Raman Studies

FULL PAPER

water polarity to favor the former reaction more than the

latter one.

~100 ns time constant. The singlet cation then reacts further

with the solvent water to yield the solvolytic nucleophilc ad-

dition product benzoin with a time constant of ~210 ns. The

decrease of the cyclization rate in the water mixed solvent

compared to the neat MeCN solvent is attributed mainly to

a steric retarding effect associated with the carbonyl H-

bonding configuration. It is suggested that three factors are

important and jointly account for the occurrence of the het-

erolysis pathway in the largely water containing solvent: the

high ionizing and polar ability of solvent to drive the hetero-

lytic dissociation, the good nucleofugal property of the leav-

ing group to make the bond breaking more feasible and the

stability of the cation intermediate formed by the heterolytic

dissociation to assist the heterolysis cleavage.

In general, the comparison of the excited state photophys-

ical deactivation and photochemical reaction established

here for UV photolysis of BDP in MeCN and H₂O/MeCN

demonstrate the importance of the solvent environment in

controlling the excited state reaction pathway. These results

afford new examples of phenacyl triplet based b-elimination

pathway and contribute new insight into the versatile depro-

tection chemistry of the phenacyl caged compounds. Further

work on the photo-deprotection mechanism and its (proba-

ble) solvent dependence for DMBEs is planned in this lab

to explore the remarkable substitution effects caused by the

benzylic meta-methoxy substitution^[17–21]and its correlation

with the deprotection pathways revealed here for the non-

substituted BDP.

Experimental and Computational Section

Conclusion

Materials: BDP was synthesized following the method detailed in refer-

ence 10. The identity and purity of the BDP sample was confirmed using

NMR, UV absorption, and mass spectrometric data. Spectroscopic grade

MeCN and deionized water were used as solvents for samples used in the

experiments.

We report here a combined fs-TA and ns-TR³spectroscopic

study to detect the dynamics and mechanism of excited state

deactivation and deprotection for BDP in neat MeCN and

75% H₂O/25% MeCN. The ns-TR³spectra in conjunction

with DFT calculations have been used to characterize the

structure and vibrational frequencies so as to help identify

important intermediates involved in the excited state pro-

cesses. DFT calculations based on a solvent–solute intermo-

lecular H-bonded complex model have been adopted to ex-

plore the conformational changes induced by the H-bonding

interaction and its relevance to the excited state reaction

pathway in largely water containing solvents. Our study pro-

vides direct time-resolved evidence that the excited state

photophysical deactivation due to the IC and ISC conver-

sion occurs rapidly on the hundred of femtoseconds and sev-

eral picosecond time scales, respectively. It has been re-

vealed unequivocally that, independent of the solvent prop-

erty, the phenacyl localized np* triplet is the common and

immediate precursor for the photo-deprotection reaction.

However, the triplet deprotection reaction has been found

to follow different pathways in neat MeCN and in the large-

ly water containing solvent. In MeCN, the triplet deprotec-

tion is an apparently a single-step process leading to the

elimination of the diethyl phosphoric acid accompanied with

cyclization to form a benzofuran product and the reaction

has been found to occur with a ~11 ns time constant. In the

largely water containing solvent, the triplet reaction has

been determined to proceed with a ~15 ns time constant

and in addition to the deprotection–cyclization, the reaction

also results in a heterolytic dissociation leading to the libera-

tion of the diethyl phosphate anion through a branching and

competing mechanism. The heterolysis is demonstrated ex-

plicitly by direct observation and identification of the triplet

cation as the initial product. The correlation of the dynamics

and the structural characterization of the intermediates indi-

cates that the triplet cation transforms subsequently into the

corresponding singlet cation by ISC conversion with a

Photochemistry experiments: The photochemistry of the photo-induced

reaction for BDP in MeCN and H₂O/MeCN mixed solvent was moni-

tored by measuring the UV/Vis absorption spectrum of the sample solu-

tion (~0.06 mm concentration) after exposure to an unfocused 267 nm

laser line (~1 mJ) from the fourth harmonic of a nanosecond Nd/YAG

Q-switch laser. The sample solution was put in a 1 cm UV grade cell and

excited in the 267 nm laser beam. As the reaction progressed, the UV/

Vis absorption spectrum of the photolyzed sample was measured by

using a Perkin–Elmer Lambda 19 UV/Vis spectrometer.

Fs-transient absorption (fs-TA) experiments: The fs-TA measurements

were performed based on a commercial femtosecond Ti/sapphire regener-

ative amplifier laser system. Details of the TA experimental setup are de-

scribed in ref. [11b]. For the present experiments, the sample solution

was excited by a 267 nm pump beam (the third harmonic of 800 nm the

regenerative amplifier fundamental) and probed by a white light contin-

ue laser beam produced from a rotating CaF₂plate pumped by the funda-

mental laser pulses (800 nm). The pump and probe laser beam spot sizes

at the sample were about 200 and 100 mm, respectively. The signals were

focused into a monochromator and detected by a liquid-N₂cooled 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 experiments were done for BDP in neat MeCN and

75% H₂O/25% MeCN with a sample concentration of ~1 mm.

Ns-time-resolved resonance Raman (ns-TR³) experiments: The ns-TR³

spectra presented here were obtained by using the experimental appara-

tus and methods described previously.^[23,24]Briefly, the sample solution

was pump by a 267 nm excitation wavelength provided by the fourth har-

monic of a Nd/YAG laser and probed by a 532 nm wavelength provided

by the second harmonic of a second Nd/YAG laser. The time delay be-

tween the pump and probe laser beams was controlled electronically by a

pulse generator and the time resolution of these experiments was ~10 ns.

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

liquid sample and the Raman scattering was collected via a backscatter-

ing configuration and detected by a liquid N₂cooled CCD. The wave-

number shifts of the resonance Raman spectra were calibrated using the

known MeCN solvent Raman bands and the spectra presented were ob-

tained from subtracting of an appropriately scaled probe-before-pump

spectrum from the corresponding pump-probe spectrum. The ns-TR³

measurements were also done for BDP in 75% H₂O/25% MeCN.

Normal resonance Raman (RR) experiments using 267 nm excitation

were performed for BDP in neat MeCN and 75% H₂O/25% MeCN. The

Chem. Eur. J. 2007, 13, 2290 – 2305

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

2303

D. L. Phillips et al.

RR experimental setup is similar to that of the ns-TR³but with the

probe beam from the second laser blocked. Sample concentrations for

the Raman experiments were in the ~1–2 mm range.

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

Chem. Soc. 1999, 121, 5625–5632; b) M. Fischer, P. Wan, J. Am.

Chem. Soc. 1999, 121, 4555–4562.

[14]A. Specht, S. Loudwig, L. Peng, M. Goeldner, Tetrahedron Lett.

2002, 43, 8947–8950.

Density functional theoretical (DFT) calculations: The DFT calculations

reported here were done employing the Gaussian 98 program suite.^[55]

Complete calculations were done to obtain the geometry optimization,

frequencies, and Raman activities of the vibrational modes analytically

by using the (U)B3LYP method with a 6-311G** basis set. For the calcu-

lations considering the inter-solute-solvent H-bonding interaction, the H-

bonded complexes containing the benzoinyl molecule of interest and one

or two water molecule(s) H-bonded with the carbonyl oxygen of the mol-

ecule were computed for the full geometric optimization, energy and fre-

quency calculations. The H-bonding stabilization energy was estimated

based on the energies (including the zero point energy, ZPE) calculated

for dissociation of the complex into the free water molecule(s) and the

benzoinyl molecule.

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

611–618; b) R. S. Givens, B. Matuszewski, J. Am. Chem. Soc. 1984,

106, 6860–6861; c) R. S. Givens, P. S. Athey, L. W. Kueper, B. Ma-

tuszewski, J. Y. Xue, J. Am. Chem. Soc. 1992, 114, 8708–8710.

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

10024.

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

[18]a) R. S. Rock, S. I. Chan, J. Am. Chem. Soc. 1998, 120, 10766–

10767; b) R. S. Rock, S. I. Chan, J. Org. Chem. 1996, 61, 1526–1529.

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

1971, 93, 7222–7228; b) J. C. Sheehan, R. M. Wilson, J. Am. Chem.

Soc. 1964, 86, 5277–5281.

[20]a) M. C. Pirrung, 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.

Acknowledgements

This research was done in the Ultrafast Laser Facility at the University

of Hong Kong and supported by grants from the Research Grants Coun-

cil of Hong Kong (HKU/7108/02P) and (HKU 1/01C) to D.L.P. W.M.K.

thanks the University of Hong Kong for the award of a Research Assis-

tant Professorship.

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

1996, 118, 12925–12937.

[22]W. S. Chan, C. Ma, W. M. Kwok, P. Zuo, D. L. Phillips, J. Phys.

Chem. A 2004, 108, 4047–4058.

[23]a) W. S. Chan, C. Ma, W. M. Kwok, D. L. Phillips, J. Phys. Chem. A

2005, 109, 3454–3469; b) P. Zuo, C. Ma, W. M. Kwok, W. S. Chan,

D. L. Phillips, J. Org. Chem. 2005. 70, 8661–8675.

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

[2]a) Y. V. Il ’ichev, M. A. Schworer, J. Wirz, J. Am. Chem. Soc. 2004,

126, 4581–4595; b) C. S. Rajesh, R. S. Givens, J. Wirz, J. Am. Chem.

Soc. 2000, 122, 611–618; c) M. A. Hangarter, A. Hormann, Y.

Kamdzhilov, J. Wirz, Photochem. Photobiol. Sci. 2003, 2, 524–535.

[3]a) K. Lee, D. E. Falvey, J. Am. Chem. Soc. 2000, 122, 9361–9366;

b) A. Banerjee, K. Lee, Q. Yu, A. G. Fan, D. E. Falvey, Tetrahedron

lett. 1998, 39, 4635–4638; c) A. Banerjee, D. E. Falvey, J. Org.

Chem. 1997, 62, 6245–6251.

[4]S. Namiki, T. Arai, K. Rujimori, J. Am. Chem. Soc. 1997, 119, 3840–

3841.

[5]H. Sugi, H. Iwamoto, T. Akimoto, H. Ushitani, Proc. Natl. Acad.

Sci. USA 1998, 95, 2273–2278.

[24]a) 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, 6641–6647; b) C. Ma, W. S.

Chan, W. M. Kwok, P. Zuo, D. L. Phillips, J. Phys. Chem. A 2004,

108, 9264–9276.

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

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

5097; b) W. A. Case, D. R. Kearns, J. Chem. Phys. 1970, 52, 2175–

2191.

[27]M. B. Ledger, G. Porter, Trans. Faraday Soc. 1972, 68, 539–553.

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

1762.

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

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

1973, 95, 5604–5614.

[31]N. Ohmori, T. Suzuki, M. Ito, J. Phys. Chem. 1988, 92, 1086–1093.

[32]S.-H. Lee, K.-C. Tang, I.-C. Chen, M. Schmitt, J. P. Shaffer, T.

Schultz, J. G. Underwood, M. Z. Zgierski, A. Stolow, J. Phys. Chem.

A 2002, 106, 8979–8991.

[33]B. K. Shah, M. A. J. Rodgers, D. C. Neckers, J. Phys. Chem. A 2004,

108, 6087–6089.

[34]S. T. Park, J. S. Feenstra, A. H. Zewail, J. Chem. Phys. 2006, 124,

174707.

[35]J.A. Warren, E. R. Bernstein, J. Chem. Phys. 1986, 85, 2365–2367.

[6]M. Lukeman, J. C. Scaiano, J. Am. Chem. Soc. 2005, 127, 7698–

7699.

[7]a) D. Geissler, Y. N. Antonenko, R. Schmidt, S. Keller, O. O. Krylo-

va, B. Wiesner, J. Bendig, P. Pohl, V. Hagen, Angew. Chem. 2005,

117, 1219–1223; Angew. Chem. Int. Ed. 2005, 44, 1195–1198; b) R.

Schmidt, D. Geissler, V. Hagen, J. Bendig, J. Phys. Chem. A 2005,

109, 5000–5004; c) T. Eckardt, V. Hagen, B. Schade, R. Schmidt, C.

Schweitzer, J. Org. Chem. 2002, 67, 703–710; d) B. Schade, V.

Hagen, R. Schmidt, R. Herbrich, E. Krause, T. Eckardt, J. Bendig, J.

Org. Chem. 1999, 64, 9109–9117.

[8]Y. Zhu, C. M. Pavlos, J. P. Toscanto, T. M. Dore, J. Am. Chem. Soc.

2006, 128, 4267–4276.

[9]M. Zabadal, A. P. Pelliccioli, P. Klan, J. Wirz, J. Phys. Chem. A 2001,

105, 10329–10333.

[10]R. S. Givens, P. S. Athey, B. Matuszewski, L. W. Kueper, J. Y. Xue,

T. Fister, J. Am. Chem. Soc. 1993, 115, 6001–6012.

[11]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.

[36]a) H. E. Zimmerman, J. Phys. Chem.

A 1998, 102, 5616–5621;

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

[37]a) J. A. Pincock, Acc. Chem. Res. 1997, 30, 43–49; b) J. W. Hilborn,

E. MacKnight, J. A. Pincock, P. J. Wedge, J. Am. Chem. Soc. 1994,

116, 3337–3346.

[38]M. Lipson, A. A. Deniz, K. S. Peters, J. Am. Chem. Soc. 1996, 118,

2992–2997.

[39]J. Catalan, E. Ferbero, F. Amat-Guerri, J. Phys. Chem. 1992, 96,

2005–2016.

[40]R. G. Zepp, M. M. Gumz, W. L. Miller, H. Gao, J. Phys. Chem. A

1998, 102, 5716–5723.

[41]P. K. Das, M. V. Encinas, J. C. Scaiano, J. Am. Chem. Soc. 1981, 103,

4154–4162.

[42]P. F. McGarry, S. Jockusch, Y. Fujiwara, N. A. Kaprinidis, N. J. Turro,

J. Phys. Chem. A 1997, 101, 764–767.

[43]J. C. Scaiano, J. Am. Chem. Soc. 1980, 102, 7747–7753.

[12]a) R. S. Givens, J. F. Weber, P. G. Conrad, G. Orosz, S. L. Donahue,

S. A. Thayer, J. Am. Chem. Soc. 2000, 122, 2687–2697; b) P. G.

Conrad, R. S. Givens, J. F. Weber, K. Kandler, Org. Lett. 2000, 2,

1545–1547; c) C. Park, R. S. Givens, J. Am. Chem. Soc. 1997, 119,

2453–2463; d) R. S. Givens, A. Jung, C. Park, J. Weber, W. Bartlett,

J. Am. Chem. Soc. 1997, 119, 8369–8370; e) R. S. Givens, C. Park,

Tetrahedron Lett. 1996, 37, 6259–6262.

2304

www.chemeurj.org

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2007, 13, 2290 – 2305

Time-Resolved Resonance Raman Studies

FULL PAPER

[44]C. Chudoba, E. T. Nibbering, T. Elsaesser, J. Phys. Chem. A 1999,

103, 5625–5628.

[53]J. P. Richard, T. L. Amyes, L. Bei, V. Stubblefield, J. Am. Chem.

Soc. 1990, 112, 9513–9519.

[45]M. Berger, E. McAlpine, C. Steel, J. Am. Chem. Soc. 1978, 100,

5147–5151.

[54]Y. Chiang, A. J. Kresge, P. Salomaa, C. I. Young, J. Am. Chem. Soc.

1974, 96, 4494–4499.

[46]C. Reichardt, Solvent and solvent Effect in Organic Chemistry, VCH,

Weinheim, 1988.

[47]a) J. H. Horner, E. Taxil, M. Newcomb, J. Am. Chem. Soc. 2002,

124, 5402–5410; b) M. Newcomb, J. H. Horner, P. O. Whitted, D.

Crich, X. Huang, Q. Yao, H. Zipse, J. Am. Chem. Soc. 1999, 121,

10685–10694.

[48]B. Schade, V. Hagen, R. Schmidt, R. Herbrich, E. Krause, T. Eck-

ardt, J. Bendig, J. Org. Chem. 1999, 64, 9109–9117.

[49]F. L. Cozens, M. OꢁNeill, R. Bogdanova, N. Schepp, J. Am. Chem.

Soc. 1997, 119, 10652–10659.

[50]C. P. Andrieux, J. Saveant, A. Tallec, R. Tardivel, C. Tardy, J. Am.

Chem. Soc. 1996, 118, 9788–9789.

[51]a) M. L. Andersen, N. Mathivanan, D. M. Wayner, J. Am. Chem.

Soc. 1996, 118, 4871–4879; b) M. L. Andersen, W. Long, D. M.

Wayner, J. Am. Chem. Soc. 1997, 119, 6590–6595; c) N. Mathivanan,

L. J. Johnston, D. M. Wayner, J. Phys. Chem. 1995, 99, 8190–8195.

[52]R. S. Givens, B. Matuszewski, P. S. Athey, M. R. Stoner, J. Am.

Chem. Soc. 1990, 112, 6016–6021.

[55]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-

zewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dap-

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.

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: June 23, 2006

Published online: December 11, 2006

Chem. Eur. J. 2007, 13, 2290 – 2305

ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

2305

Article Doi

Femtosecond transient absorption and nanosecond time-resolved resonance raman study of the solvent-dependent photo-deprotection reaction of benzoin diethyl phosphate

DOI: 10.1002/chem.200600893

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.200600893

Full text of DOI:10.1002/chem.200600893

Products guided by the article

R&D Labs maybe for 119-53-9

Relevant to this article

Hot Product