Presence and absence of excited state intramolecular charge transfer with the six isomers of dicyano- N, N -dimethylaniline and dicyano-(N -methyl- N -isopropyl)aniline

ARTICLE

pubs.acs.org/JPCA

Presence and Absence of Excited State Intramolecular Charge

Transfer with the Six Isomers of Dicyano-N,N-dimethylaniline and

Dicyano-(N-methyl-N-isopropyl)aniline

Victor A. Galievsky,*^,†,‡Sergey I. Druzhinin,*^,†Attila Demeter,^†,§Sergey A. Kovalenko,*^,||

Tamara Senyushkina,^†Peter Mayer,^{^}and Klaas A. Zachariasse*^,†

^†Max-Planck-Institut f€ur biophysikalische Chemie, Spektroskopie und Photochemische Kinetik, 37070 G€ottingen, Germany

^‡B.I. Stepanov Institute of Physics, National Academy of Sciences of Belarus, pr. Nezavisimosti 68, 220072 Minsk, Belarus

^§Institute of Materials and Environmental Chemistry, Chemical Research Center, Hungarian Academy of Sciences, P.O. Box 17,

1525 Budapest, Hungary

Institut f€ur Chemie, Humboldt Universit€at zu Berlin, Brook-Taylor Strasse 2, 12489 Berlin, Germany

^{^}Department Chemie und Biochemie, Ludwig-Maximilians-Universit€at, Butenandtstrasse 5-13, Haus F, 81377 M€unchen, Germany

S Supporting Information

b

ABSTRACT:

The excited state behavior of the six m,n-dicyano-N,N-dimethylanilines (mnDCDMA) and m,n-dicyano-(N-methyl-N-isopropyl)-

anilines (mnDCMIA) is discussed as a function of solvent polarity and temperature. The dicyano moiety in these electron donor

(D)/acceptor (A) molecules has a considerably larger electron aﬃnity than the benzonitrile subgroup in 4-(dimethylamino)benzonitrile

(DMABN). Nevertheless, the ﬂuorescence spectra of the mnDCDMAs and mnDCMIAs in n-hexane all consist of a single emission

originating from the locally excited (LE) state, indicating that a reaction from LE to an intramolecular charge transfer (ICT) state

does not take place. The calculated energies E(ICT), obtained by employing the reduction potential of the dicyanobenzene

subgroups and the oxidation potential of the amino substituents trimethylamine (N(Me)₃) and isopropyldimethylamine

(iPrNMe₂), are lower than E(LE). The absence of an LE f ICT reaction therefore makes clear that the D and A units in

the dicyanoanilines are not electronically decoupled. In the polar solvent acetonitrile (MeCN), dual (LE + ICT) ﬂuorescence is

found with 24DCDMA and 34DCDMA, as well as with 24DCMIA, 25DCMIA, and 34DCMIA. For all other mnDCDMAs and

mnDCMIAs, only LE emission is observed in MeCN. The ICT/LE ﬂuorescence quantum yield ratio Φ⁰(ICT)/Φ(LE) in MeCN at

25 ꢀC is larger for 24DCDMA (1.2) than for 34DCDMA (0.35). The replacement of methyl by isopropyl in the amino substituent

leads to a considerable increase of Φ⁰(ICT)/Φ(LE), 8.8 for 24DCMIA and 1.4 for 34DCMIA, showing that the LE a ICT

equilibrium has shifted further toward ICT. The appearance of an ICT reaction with the 2,4- and 3,4-dicyanoanilines is caused by a

relatively small energy gap ΔE(S₁,S₂) between the two lowest excited singlet states as compared with the other m,n-dicyanoanilines,

in accordance with the PICT model. The observation that the ICT reaction is more eﬃcient for 24DCMIA and 34DCMIA than for

their mnDCDMA counterparts is mainly caused by the fact that iPrNMe₂is a better electron donor than N(Me)₃: E(D/D⁺) =

0.84 against 1.05 V vs SCE. That ICT also occurs with 25DCMIA, notwithstanding its large ΔE(S₁,S₂), is due to the substantial amino

twist angle θ = 42.6ꢀ, which leads to partial electronic decoupling of the D and A subgroups. The dipole momentsμ_e(ICT) range between

Received: May 16, 2011

Revised:

July 28, 2011

Published: July 29, 2011

r

2011 American Chemical Society

10823

dx.doi.org/10.1021/jp2045614 J. Phys. Chem. A 2011, 115, 10823–10845

|

The Journal of Physical Chemistry A

ARTICLE

18 D for 34DCMIA and 12 D for 25DCMIA, larger than the corresponding μ_e(LE) of 16 and 11 D. The diﬀerence between μ_e(ICT)

and μ_e(LE) is smaller than with DMABN (17 and 10 D) because of the noncollinear arrangement of the amino and cyano

substituents (diﬀerent dipole moment directions). The dicyanoanilines that do not undergo ICT, have LE dipole moments between

9 and 16 D. From plots of ln(Φ⁰(ICT)/Φ(LE)) vs 1000/T, the (rather small) ICT reaction enthalpies ΔH could be measured in

MeCN: 5.4 kJ/mol (24DCDMA), 4.7 kJ/mol (24DCMIA), and 3.9 kJ/mol (34DCMIA). With the mnDCDMAs and mnDCMIAs

only showing LE emission, the ﬂuorescence decays are single exponential, whereas for those undergoing an LE f ICT reaction the

LE and ICT picosecond ﬂuorescence decays are double exponential. In MeCN at 25 ꢀC, the decay times τ₂have values between

1.8 ps for 24DCMIA and 4.6 ps for 34DCMIA at 25 ꢀC. Longer times are observed at lower temperatures. Arrhenius plots of the forward

and backward ICT rate constants k_aand k_dof 25DCMIA in tetrahydrofuran, obtained from the LE and ICT ﬂuorescence decays, give

the activation energies E_a= 4.5 kJ/mol and E_d= 11.9 kJ/mol, i.e., ΔH = ꢀ7.4 kJ/mol. From femtosecond transient absorption

spectra of 24DCDMA and 34DCDMA at 22 ꢀC, ICT reaction times τ₂= 1/(k_a+ k_d) of 1.8 and 3.1 ps are determined. By combining

these results with the data for the ﬂuorescence decays and Φ⁰(ICT)/Φ(LE), the values k_a= 49 ꢁ 10¹⁰s^ꢀ1(24DCDMA) and k_a=

23 ꢁ 10¹⁰s^ꢀ1(34DCDMA) are calculated. An LE and ICT excited state absorption is present even at a pump/probe delay time of

100 ps, showing that an LE a ICT equilibrium is established.

Chart 1

’ INTRODUCTION

Of the 19 diﬀerent cyano(CN)-substituted N,N-dimethylanilines

(DMAs), seven molecules have been investigated in connection with

the occurrence of a reaction from a locally excited (LE) precursor to

an intramolecular charge transfer (ICT) state. A deﬁnition of the terms

LE and ICT is given in ref 1. For the three DMAs substituted with

one CN group (Chart 1), only the para-isomer 4-(dimethyl-

amino)benzonitrile (DMABN) undergoes ICT and emits dual

ﬂuorescence from an LE and an ICT state.^2ꢀ7With the ortho-

and meta-substituted derivatives 2-(dimethylamino)benzonitrile

(oDMABN) and 3-(dimethylamino)benzonitrile (mDMABN),

such an ICT reaction accompanied by LE + ICT dual ﬂuores-

cence has not been observed under any condition of solvent

polarity or temperature.^8,9This inability to undergo an ICT reac-

tion has been attributed to the fact that the energy gap ΔE(S₁,S₂)

between the two lowest excited singlet states is much larger than

for DMABN: 7710 cm^ꢀ1(oDMABN), 7940 cm^ꢀ1(mDMABN),

and 2700 cm^ꢀ1(DMABN), in n-hexane, for example.¹⁰

Of the six possible isomers of DMA with two CN-substituents,

only for 3,4-dicyano-N,N-dimethylaniline (34DCDMA) and

3,5-dicyano-N,N-dimethylaniline (35DCDMA) a possible ICT reac-

tion has been discussed.^9ꢀ18LE ﬂuorescence, but no ICT emis-

sion, was then reported with 34DCDMA and 35DCDMA, even

in the strongly polar solvent acetonitrile (MeCN, ε²⁵= 36.7),

although ﬂuorescence spectra were not shown.^9,11ꢀ15In accor-

dance with this observation, the LE ﬂuorescence decays, at a time

resolution of ∼10 ps, are single exponential with times in the

nanosecond range and no indication of a second, shorter

picosecond decay time.^9,11The absence of ICT emission was,

similar to what is described above for oDMABN and mDMABN,

explained by the magnitude of ΔE(S₁,S₂), which, in n-hexane, is

again much larger than in the case of DMABN: 4360 cm^ꢀ1for

34DCDMA and 7200 cm^ꢀ1for 35DCDMA, as compared with

2700 cm^ꢀ1for DMABN.¹⁰

computations.¹⁵For pentacyano-N,N-dimethylaniline (PCDMA)

in MeCN and tetrahydrofuran (THF), the ﬂuorescence spectra

likewise consist of a single LE emission, and the femtosecond

transient absorption spectra only show the presence of a single

LE state.¹⁰In conclusion, similar to the DMABNs and DCDMAs

treated above, the absence of an ICT reaction for TCDMA and

PCDMA has also been attributed to their large energy gaps

ΔE(S₁,S₂), 5370 cm^ꢀ1(TCDMA,¹⁵MeCN) and 7170 cm^ꢀ1

(PCDMA,¹⁰MeCN), in accordance with the planar ICT (PICT)

model.^{7,9ꢀ11,13ꢀ15}

As has been discussed before,^9ꢀ15the absence of an ICT reaction

with the electron donor (D)/acceptor (A) molecules 34DCDMA,

35DCDMA, TCDMA, and PCDMA is at ﬁrst sight unexpected,

because the dicyanobenzene, tricyanobenzene, and pentacyanoben-

zene A moieties have a considerably larger electron aﬃnity than

benzonitrile (the A subunit of DMABN). This can be seen from their

reduction potentials E(A^ꢀ/A).^10,19For each of these D/A molecules,

under the assumption of weakly interacting D and A subgroups

(the twisted ICT (TICT) model),⁶the calculated energy E(ICT) of

the ICT state is lower than the energy E(S₁,LE) of the excited state

precursor, in n-hexane as well as in MeCN. Under these conditions, an

LE f ICT reaction should be found: ΔH < 0.¹⁰The fact that such a

reaction is not observed^9ꢀ13,15makes clear that the D and A groups in

a hypothetical ICT state of 34DCDMA, 35DCDMA, TCDMA, and

PCDMA are not electronically decoupled, as would be required in the

TICT approach.⁶

34DCDMA and 35DCDMA Computations. From computa-

tions on 35DCDMA, it has been concluded that its meta-CN

substituents influence the S₁as well as the S₂state, in contrast to

DMABN, for which the interaction with CN mainly takes place

for S₂.¹⁶The S₁and S₂states of 35DCDMA, with similar dipole

moments, are therefore equally stabilized when the solvent

polarity increases. This means that the energy gap ΔE(S₁,S₂),¹⁶

which in the gas phase (calculations) is already much larger

than for DMABN,^20,21does not become smaller with increasing

solvent polarity, explaining why dual emission comparable to that

of DMABN is not observed with 35DCDMA. It is then concluded

The photophysical behavior of anilines with three CN-sub-

stituents, 2,4,6-tricyano-N,N-dimethylaniline (TCDMA), 2,4,

6-tricyano-N-methylaniline (TCMA), and 2,4,6-tricyanoaniline

(TCA), has been investigated by employing photostationary

absorption and emission spectra, picosecond ﬂuorescence de-

cays, and femtosecond transient absorption.¹⁵With these mea-

surements, only a single LE state could be detected for these

molecules, without any evidence for the occurrence of an LE f

ICT reaction, even in a strongly polar solvent such as MeCN.¹⁵

This experimental ﬁnding was supported by the results of

10824

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Chart 2

Chart 3

that its fluorescence spectrum is that of a single LE S₁emission band

and that even for 35DCDMA in very polar solvents a S₁(LE) state

with planar conformation is favored.¹⁶In another computation on

34DCDMA and 35DCDMA, it was found that their S₁states have

CT character and that these molecules have a significantly larger

energy gap ΔE(S₁,S₂) than DMABN, explaining the absence of

dual fluorescence, again in accordance with the PICT model.¹⁷

In a time-dependent density-functional calculation, without

geometry optimization for the excited states, the absence of dual

ﬂuorescence of 34DCDMA and 35DCDMA in polar solvents is

discussed.¹⁸These molecules have again a large ΔE(S₁,S₂) gap,

which was taken as an explanation that an ICT reaction does not

take place (PICT model). The single S₁ﬂuorescence band of

34DCDMA is identiﬁed as coming from a TICT state, with a

dipole moment μ_e(S₁) of 24 D. The energy E(S₁) of 35DCDMA

is almost independent of the amino twist angle. This S₁state has

important CT character, as seen from its computed μ_e(S₁) of

19 D. The FranckꢀCondon state S₁(FC), reached by excitation

of the equilibrated S₀ground state, is an ICT state with a

calculated dipole moment μ_e(S₁,abs) of 17 D for 34DCDMA

and 15 D for 35DCDMA. The S₂(FC) state is claimed to be of LE

nature, although μ_e(S₂,abs) of both molecules is with 16 D,

similar to μ_e(S₁,abs). ΔE(S₁,S₂) is hence not reduced by increas-

ing the solvent polarity, an explanation for the absence of an ICT

reaction, as mentioned above.

This compound was reacted with dimethylsulfate, producing

23DCDMA (mp 104.4ꢀ105.6 ꢀC).^22a23DCMIA was obtained by

adding methyldiisopropylamine to a solution of 3-nitro-1,2-dicyano-

benzene and triethylamine in N,N-dimethylacetamide at 100 ꢀC,

with a ﬁnal treatment (10 min) under microwave irradiation

(CEM, Discover). 24DCDMA (mp 109.0ꢀ109.4 ꢀC), 25DCDMA,

and 35DCDMA (mp 179.8ꢀ180.2 ꢀC) were made in a reaction

between the corresponding m,n-dibromoaniline and methyl

iodide (MeI) followed by a treatment²³with CuCN in quinoline

or dimethylformamide. 24DCMIA, 25DCMIA, and 35DCMIA

(mp 107.1ꢀ107.6 ꢀC) were likewise synthesized by adding

MeI to a solution of the m,n-dibromoaniline and isopropyl iodide

(iPrI). To obtain 26DCDMA (mp 145.0ꢀ145.4 ꢀC) and

34DCDMA, the appropriate m,n-dibromoaniline was reacted

with dimethysulfate, whereas 26DCMIA resulted from 2,6-di-

bromoaniline and iPrI in methanol. 34DCMIA was synthesized

from 4-ﬂuoro-1,2-dicyanobenzene and methylisopropylamine by

microwave irradiation for 3 min at 100 ꢀC. 2,4-Dicyano-

N-isopropylaniline (24DCIA, mp 98.8ꢀ99.1 ꢀC) was made from

2,4-dibromoaniline and iPrI, giving 2,4-dibromo-N-isopropylani-

line and a subsequent reaction with CuCN. Finally, 24DCP4C

(mp 191.3ꢀ191.4 ꢀC) was synthesized from 2,4-dicyano-

1-ﬂuorobenzene (24DCF) and azetidine in dimethylsulfoxide.

24DCF was obtained in a reaction between 2,4-dibromo-

1-ﬂuorobenzene and CuCN. For all molecules investigated,

HPLC was the last puriﬁcation step. All solvents were chromato-

graphed over Al₂O₃just prior to use. The solutions, with an optical

density between 0.4 and 0.6 for the maximum of the ﬁrst band in the

absorption spectrum, were deaerated by bubbling with nitrogen for

15 min. The measurement and treatment of the ﬂuorescence

spectra, quantum yields, nanosecond and picosecond single photon

counting decays, and femtosecond transient absorption spectra have

been described elsewhere.^7,24ꢀ28The error limits (data repro-

ducibility) of the total ﬂuorescence quantum yields, extinction

coeﬃcients, and ﬂuorescence decay times are estimated at 2ꢀ3%.

The accuracy of the maxima of the ﬂuorescence and absorp-

tion spectra is around 50 cm^ꢀ1(better for narrow bands). For

the separated LE and ICT ﬂuorescence bands, the uncertainty

of the ﬂuorescence maxima and quantum yields increases by a

factor of 3, as a result of unavoidable errors and intrinsic

ambiguity during the band separation procedure.

In all three calculations,^16ꢀ18only the S₀structure was opti-

mized, and this structure was subsequently used for the excited

states. For the TICT states, a rotation around the N-phenyl bond

was introduced, leaving the rest of the molecular structure

unchanged. Therefore, whereas absorption spectra can be ob-

tained from the calculations, the computed states are not the

relaxed S₁and ICT states from which the ﬂuorescence is emitted.

In the present Article, the photophysics of all six mnDCDMA

isomers and N-(2,4-dicyanophenyl)azetidine (24DCP4C) is

discussed (Chart 2). In addition, the six m,n-dicyano-(N-methyl-

N-isopropyl)anilines (mnDCMIA) and 2,4-dicyano-N-isopropyl-

aniline (24DCIA) are studied (Chart 3).

’ EXPERIMENTAL SECTION

3-Nitro-1,2-dicyanobenzene was treated with HCl in methanol

under addition of iron powder, resulting in 2,3-dicyanoaniline.

10825

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Figure 1. Crystal structures of 24DCDMA (a,b), 24DCMIA (c,d), 25DCMIA (e,f), and 34DCMIA (g,h). For each molecule, a view from above (a,c,e,g)

and one along the axis from the amino group to the dicyanobenzene moiety (b,d,f,h) is presented. The amino twist angle θ is deﬁned as

(C(2)C(1)N(7)C(8) + C(6)C(1)N(7)C(9))/2. The pyramidal angle j is the angle between the vector N(7)C(1) and the plane C(8)N(7)C(9). See

Chart 4 and Table 1.

’ RESULTS AND DISCUSSION

caused by a larger amino twist angle θ and pyramidality of N(7),

and consequently a relatively small value of ΣN, the sum of the

angles around N(7), of 344.6ꢀ for 25DCMIA, as compared with

359.8ꢀ for 24DCDMA, 358.6ꢀ for 24DCMIA, and 356.5ꢀ for

34DCMIA (Table 1).

Electronic Decoupling of the Amino Group in 25 DCMIA. The

exceptionally large amino twist angle θ = 42.6ꢀ and pyramidality

angle j = 38.4ꢀ of N(7) in 25DCMIA will lead to an electronic

decoupling of the amino and dicyanobenzene subgroups, as

shown by the increase of the N(7)ꢀC(1) bond length to 140.6

pm (see above). Such a decoupling can cause a change in the

mechanism of the ICT reaction, as encountered with 4-(di-tert-

butylamino)benzonitrile³⁰(DTABN) and 3,5-dimethyl-4-(dimethyl-

amino)benzonitrile^{4,9,11,13,21,31}(MMD). This will be further

discussed in a subsequent section.

Phenyl Ring. In 24DCDMA, 24DCMIA, and 25DCMIA, the

bonds C(1)ꢀC(2) and C(2)ꢀC(3) adjacent to the ortho CN-

group are clearly lengthened as compared with DMABN. For

example, for 24DCDMA C(1)ꢀC(2) = 142.9 pm and C(2)ꢀ

C(3) = 139.4 pm, larger than for DMABN with 140.8 and

137.0 pm.

Absorption and Fluorescence of the Six mnDCDMAs and

24DCP4C at 25 ꢀC. mnDCDMAs in n-Hexane. The absorption

and fluorescence spectra of the six mnDCDMAs in n-hexane at

25 ꢀC are depicted in Figure 2. Two bands appear in the

absorption spectra, a lower-energy band of smaller absorbance

than the one at higher energy. These bands are attributed to the

transition from S₀to the singlet excited states S₁and S₂, respec-

tively.^9ꢀ18The order of the energies of the first absorption band

~v^max(S₁,abs) for the mnDCDMAs in n-hexane (Table 2) is: 34 >

24 > 35 > 26 > 23 > 25. The same order has been found for the

m,n-dicyanoanilines in n-heptane.³²

X-ray Crystal Analysis. Results of an X-ray crystal analysis of

24DCDMA (Figure 1a,b), 24DCMIA (Figure 1c,d), 25DCMIA

(Figure 1e,f), and 34DCMIA (Figure 1g,h) are listed in Table 1,

together with data for DMABN,²⁹4-(diisopropylamino)benzo-

nitrile (DIABN,²¹Figure S1 in Supporting Information (SI)),

and 4-(azetidinyl)benzonitrile (P4C),²⁶included for comparison

(Chart 4).

Amino Twist Angle θ. For 34DCMIA a small amino twist

angle θ of 2.9ꢀ is observed (Figure 1). The presence of an ortho

CN-substituent leads to larger angles θ in the dicyanoanilines:

13.9ꢀ (24DCDMA), 16.7ꢀ (24DCMIA), and 42.6ꢀ (25DCMIA).

The azetidinyl group in 24DCP4C does not cause such steric

hindrance: θ = 3.6ꢀ. Note that DIABN, with two isopropyl

substituents, has a larger twist angle (14.1ꢀ and 14.4ꢀ) than

DMABN (0.0ꢀ) and P4C (1.0ꢀ). The exceptionally large angle

θ = 42.6ꢀ of 25DCMIA is a result of the somewhat surprising

location of isopropylamino in proximity of the ortho cyano

group, different from 24DCDMA, 24DCMIA, and 34DCMIA,

in which isopropyl has a position opposite of the ortho- or meta-

cyano substituent (see Figure 1).

Amino Pyramidal Angle j. For the dicyanoanilines with an

ortho cyano substituent, the amino pyramidal angle j (Figure 1)

is related with the twist angle θ: 25DCMIA has the largest value

for j (38.4ꢀ) as well as for θ (42.6ꢀ). For 34DCMIA with θ =

2.9ꢀ, j is nevertheless substantial (18.4ꢀ). Also for DMABN and

P4C a zero or small twist angle is accompanied by a relatively

large j, 12.1ꢀ (DMABN) and 25.8ꢀ (P4C) (see Table 1).

AminoꢀPhenyl Bond N(7)ꢀC(1). The length of the bond

N(7)ꢀC(1) between the amino nitrogen and the phenyl ring for

24DCDMA, 24DCMIA, and 34DCMIA (around 136 pm) is

similar to that of DMABN (136.5 pm) and somewhat longer than

the 133.9 pm for 24DCP4C. With 25DCMIA, a clearly longer

bond length of 140.6 pm is found. This increase in N(7)ꢀC(1) is

The ﬂuorescence spectra consist of a single relatively narrow

emission band (half-width Δ(1/2) = 3000ꢀ3650 cm^ꢀ1) with some

vibrational structure, typical for an LE emission. The ﬂuorescence

10826

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Table 1. Data on the Ground State Structure of 24DCDMA, 24DCMIA, 25DCMIA, 34DCMIA, and 24DCP4C from X-ray Crystal

Analysis^a

24DCDMA

24DCMIA

25DCMIA

34DCMIA

24DCP4C

DMABN^b

DIABN^cA

DIABN^cB

P4C^d

N(7)ꢀC(1)

135.4

146.1

145.4

142.9

141.4

139.4

143.3

137.5

136.0

146.1

148.2

142.9

142.1

139.5

143.8

138.0

140.6

146.9

149.0

141.6

139.7

139.3

143.9

137.7

136.9

146.0

147.4

141.1

141.0

138.3

133.9

146.1

146.6

142.0

141.7

139.8

143.3

138.4

136.5

144.8

140.8

137.0

137.6

148.3

147.0

142.0

141.2

137.4

148.7

147.9

141.9

141.4

137.1

136.4

146.5

146.6

140.4

139.8

137.7

N(7)ꢀC(8)

N(7)ꢀC(9)

C(1)ꢀC(2)

C(1)ꢀC(6)

C(2)ꢀC(3)

C(2)ꢀC(12)

C(3)ꢀC(4)

140.5

143.9

143.3

138.7

137.4

151.6

151.2

114.6

114.8

118.5

120.7

117.3

ꢀ8.8

14.6

138.8

139.1

139.3

139.5

C(3)ꢀC(12)

C(4)ꢀC(14)

143.6

140.5

136.2

143.5

139.8

136.6

152.3

152.0

114.9

121.9

121.0

115.7

23.9

143.4

140.9

136.6

153.3

142.7

138.8

137.0

143.7

139.7

137.5

152.7

152.3

143.7

139.1

137.6

152.5

151.5

143.6

139.2

137.9

154.2

C(4)ꢀC(5)

139.2

139.0

152.1

151.5

114.7

114.5

115.0

115.2

114.4

19.6^e

65.8^f

C(5)ꢀC(6)

C(9)ꢀC(10)

C(9)ꢀC(11)

C(12)ꢀN(13)

C(14)ꢀN(15)

C(1)ꢀN(7)ꢀC(8)

C(1)ꢀN(7)ꢀC(9)

C(8)ꢀN(7)ꢀC(9)

C(8)ꢀN(7)ꢀC(1)ꢀC(2)

C(9)ꢀN(7)ꢀC(1)ꢀC(6)

twist angle θ^g

ΣN^h

114.7

114.6

120.4

124.2

115.4

11.4

115.2

115.3

135.1

130.3

94.5

114.5

120.6

121.5

116.4

ꢀ7.7

7.7

115.0

122.3

120.0

115.2

23.8

114.7

122.5

120.0

114.9

ꢀ24.9

ꢀ3.9

14.4

114.4

127.4

128.3

94.1

ꢀ5.6

ꢀ1.7

3.6

ꢀ22.0

24.0

16.5

9.6

4.4

13.9

16.7

42.6

2.9

0.0

14.1

1.0

359.8

2.4

358.6

11.3

344.6

38.4

356.5

18.4

359.9

2.3

358.5

357.5

15.1

357.5

15.2

349.8

25.8

pyramidal angle jⁱ

12.1

c

^aThe structural data for DMABN, DIABN, and P4C are included for comparison. ^bRef 29 (253 K). Ref 21 (293 K). ^dCrystal structure data of

P4C have appeared in ref 26, under the name AZABN. ^eC(8)N(7)C(1)C(6), see Chart 4. C(9)N(7)C(1)C(2), see Chart 4. ^gTwist angle θ:

f

(C(2)C(1)N(7)C(8) + C(6)C(1)N(7)C(9))/2 (Chart 4 and Figure 1). ^hSum of the angles around the amino nitrogen. ⁱPyramidal angle j: angle

between the vector N(7)C(1) and the plane C(8)N(7)C(9) (Chart 4 and Figure 1).

Chart 4

remains single down to ꢀ95 ꢀC. It originates from an LE state, as

there is no indication of an additional red-shifted ICT ﬂuorescence.

This is clear evidence that an LE f ICT reaction does not take place

with the six DCDMAs in n-hexane at 25 ꢀC.

The ﬂuorescence quantum yields Φ(LE) range between

0.07 (24DCDMA and 26DCDMA) and 0.31 (25DCDMA),

with large intersystem crossing yields Φ(ISC): 0.78, 0.81,

and 0.54 (Table 2). The process of internal conversion (IC)

from S₁(LE) directly to S₀is relatively important, with yields

Φ(IC) between 0.12 (26DCDMA) and 0.46 (35DCDMA),

see Table 2.

mnDCDMAs in MeCN. The absorption and fluorescence spec-

tra of the six mnDCDMAs in MeCN at 25 ꢀC are shown in

Figure 3 (Table 3). The absorption spectra in this polar solvent are

red-shiftedrelative tothoseinn-hexane (cf. Figure2, Table 2). The

order of the energies of the first absorption band ~v^max(S₁,abs) for

the mnDCDMAs in MeCN (Table 3) is: 24 > 34 > 26 = 35 > 23 >

25. A similar order has been reported for the m,n-dicyanoanilines

in MeCN and H₂O, as well as in 95% C₂H₅OH.^32,22bThis means

that there is in this respect no substantial influence of steric

hindrance between the dimethylamino group and an ortho-cyano

substituent in the mnDCDMAs with mn = 23, 24, 25, and 26. It

was stated in ref 22b that this order cannot be explained in terms of

resonance theory, as discussed previously.³⁵

24DCP4C in n-Hexane. Thefluorescencespectrumof24DCP4C

in n-hexane at 25 ꢀC (Figure S2a in SI) consists of a single LE

emission, similar to what has been observed^{11ꢀ13,26,33}with its parent

compound P4C (Table 2).

The ﬂuorescence spectra of four of the mnDCDMAs (23-, 25-,

26-, and 35DCDMA) consist of a single unstructured emission

10827

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Figure 2. Fluorescence (LE) and absorption (Abs) spectra of the six m,n-dicyano-N,N-dimethylanilines mnDCDMA (Chart 2) in n-hexane at 25 ꢀC.

The ﬂuorescence spectra are attributed to the single emission from a locally excited (LE) state.

Table 2. Data Obtained from the Fluorescence and Absorption Spectra of the Six mnDCDMAs, 24DCP4C, and P4C in n-Hexane

at 25 ꢀC

23DCDMA

24DCDMA

25DCDMA

26DCDMA

34DCDMA

35DCDMA

24DCP4C

P4C^a

~v^max(LE) [cm^ꢀ1

Δ(1/2)(LE)^b[cm^ꢀ1

]

24410

3400

26370

3410

24340

3120

25180

3030

26470

3650

25910

3000

27100, 26040

3220

29330

3470

4710

0.26

]

Δ(1/4)(LE)^c[cm^ꢀ1

]

4610

4650

4300

4110

4940

4150

4370

Φ(LE)

0.13

0.065

0.31

0.066

0.17

0.095

Φ(ISC)^d

Φ(IC)^e

0.53 ( 0.08

0.34

0.78 ( 0.04

0.16

0.54 ( 0.03

0.15

0.81 ( 0.03

0.12

0.57 ( 0.06

0.26

0.45 ( 0.07

0.46

0.67

0.07

~v^max(S₁,abs) [cm^ꢀ1

^max(S₁,abs) [M^ꢀ1cm^ꢀ1

~v^max(S₂,abs) [cm^ꢀ1

^max(S₂,abs) [M^ꢀ1cm^ꢀ1

E(S₁)^h[cm^ꢀ1

ΔE(S₁,S₂)ⁱ[cm^ꢀ1

]

27570

29540

5020

27420

4410

27830

29670

28450

29330

34600

31110

f

g

f

ε

]

36100

34390

26220

27880

4850

36740

9800

34450

26280

34030

35760

35670

29950

31060

4560

ε

]

25850

8530

25590

28000

4360

26900

7310

27720

5270

]

9320

6620

c

^aSee ref 26. ^bFull width at half-maximum of the total LE ﬂuorescence band. Full width at quarter-maximum of the total LE ﬂuorescence band.

^dMeasurements (at 23 ꢀC) as in refs 26 and 34. ^eΦ(IC) = 1 ꢀ Φ(LE) ꢀ Φ⁰(ICT) ꢀ Φ(ISC). ^fNot suﬃciently soluble in n-hexane. ^gNot enough

substance available. ^hCrossing point of the normalized ﬂuorescence and absorption spectra. ⁱEnergy diﬀerence ~v^max(S₂,abs) ꢀ ~v^max(S₁,abs).

band (Figure 3a,c,d,f), with an asymmetric shape similar to that

of the LE ﬂuorescence of these molecules in n-hexane (Figure 2a,

c,d,f) and that of, e.g., TCDMA¹⁵in MeCN. The bands are

therefore likewise assigned to the LE state. It should be noted,

however, that the ﬂuorescence band of 25DCDMA (Figure 3c) is

somewhat broader than that of the other three isomers, as seen

from its half-width Δ(1/2) = 4580 cm^ꢀ1and quarter-width

Δ(1/4) = 6550 cm^ꢀ1(Table 3). The overall ﬂuorescence quantum

yield of 25DCDMA (0.046) is lower than that of 23DCDMA

(0.283), 26DCDMA (0.096), and 35DCDMA (0.162) (Table 3),

an additional indication that with 25DCDMA in MeCN an LE f

ICT reaction is starting to occur. This interpretation is strengthened

by the appearance of an ICT emission band in the spectrum of the

related compound 25DCMIA in MeCN (Figure 5c, below).

With 24DCDMA and 34DCDMA in MeCN, in contrast, an

additional ﬂuorescence band is present at the low-energy side of

the LE emission (Figure 3b,e), and the band is attributed to the

ICT state. The half-width Δ(1/2) of the overall ﬂuorescence

10828

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Figure 3. Fluorescence (LE or LE + ICT) and absorption (Abs) spectra of the six m,n-dicyano-N,N-dimethylanilines mnDCDMA (Chart 2) in

acetonitrile (MeCN) at 25 ꢀC. The ﬂuorescence spectra are attributed to the emission from a locally excited (LE) state, except for those of 24DCDMA

and 34DCDMA, which show dual ﬂuorescence from an LE and an intramolecular charge transfer (ICT) state (see text). The ICT emission band has

been obtained by subtraction with the shifted LE ﬂuorescence spectrum in n-hexane (Figure 2).

Table 3. Data Obtained from the Fluorescence and Absorption Spectra of the Six mnDCDMAs, 24DCP4C, and P4C in MeCN at

25 ꢀC

23DCDMA

24DCDMA

25DCDMA

26DCDMA

34DCDMA

35DCDMA

24DCP4C

P4C^a

~v^max(LE) [cm^ꢀ1

Δ(1/2)(ﬂu)^b[cm^ꢀ1

Δ(1/2)(LE)^c[cm^ꢀ1

Δ(1/2)(ICT)^d[cm^ꢀ1

Δ(1/4)(ﬂu)^e[cm^ꢀ1

Δ(1/4)(LE)^g[cm^ꢀ1

]

21180

-

23880

19400

8670

21070

-

23170

-

23040

19160

4940

22510

-

24580

27610

19940

3680

3640

4980

12530^f

5340

7130

0.034

0.017

0.70

-

]

4270

-

4580

-

3420

-

3930

-

3720

-

]

4200

4300

]

4960

6800

]

5680

-

10700

5750

6550

-

4830

-

7450

5530

-

5100

-

]

5970

Δ(1/4)(ICT)^h[cm^ꢀ1

]

7010

9630

Φ(LE)

Φ⁰(ICT)

Φ⁰(ICT)/Φ(LE)

Φ(ISC)ⁱ

Φ(IC)^j

0.28

0

0.013

0.016

1.2

0.046

0

0.096

0

0.022

0.0077

0.35

0.16

0

0.28

0

0.08 ( 0.02

0.64

26770

4050

35730

9470

24020

8960

0.69 ( 0.02

0.28

0.08 ( 0.01

0.87

26490

4100

36600

8420

23900

10110

0.90 ( 0.01

0.0

0.09 ( 0.01

0.88

0.57 ( 0.02

0.27

27350

2900

34850

12890

24910

7500

0.90

0.05

~v^max(S₁,abs) [cm^ꢀ1

^max(S₁,abs) [M^ꢀ1cm^ꢀ1

~v^max(S₂,abs) [cm^ꢀ1

^max(S₂,abs) [M^ꢀ1cm^ꢀ1

E(S₁)^k[cm^ꢀ1

ΔE(S₁,S₂)^l[cm^ꢀ1

]

29000

5230

27440

3520

34250

8320

25130

6810

28500

6500

28900

33900

ε

]

33830

25930

26680

4870

33100

23320

25930

4600

34290

27700

30600

ε

]

26690

5000

]

^aSee ref 26. There, Φ⁰(ICT) = 0.034 and Φ(LE) = 0.017, at 20 ꢀC. ^bFull width at half-maximum (fwhm) of the total ﬂuorescence band. ^cFwhm for the

separated LE band. ^dFwhm for the separated ICT band. ^eFull width at quarter-maximum (fwqm) of the total ﬂuorescence band. ^fSeparated LE and

ICT bands. ^gFwqm for the separated LE band. ^hFwqm for the separated ICT band. ⁱMeasurements (at 23 ꢀC) as in refs 26 and 34. ^jΦ(IC) = 1 ꢀ

Φ(LE) ꢀ Φ⁰(ICT) ꢀ Φ(ISC). ^kCrossing point of the normalized ﬂuorescence and absorption spectra. ^lEnergy diﬀerence ~v^max(S₂,abs) ꢀ ~v^max(S₁,abs).

10829

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Figure 4. Fluorescence (LE) and absorption (Abs) spectra of the six m,n-dicyano-(N-methyl-N-isopropyl)anilines mnDCMIA (Chart 3) in n-hexane at

25 ꢀC. The ﬂuorescence spectra are attributed to the single emission from a locally excited (LE) state.

band is very large for 24DCDMA (8670 cm^ꢀ1), larger than for

34DCDMA (4940 cm^ꢀ1), with a corresponding diﬀerence for

Δ(1/4): 10 700 and 7450 cm^ꢀ1(Table 3). These spectra have an

ICT/LE ﬂuorescence quantum yield ratio Φ⁰(ICT)/Φ(LE) of 1.2

(24DCDMA) and 0.35 (34DCDMA), see Table 3. The LE

ﬂuorescence quantum yield Φ(LE) of 24DCDMA (0.013) and

34DCDMA (0.022) is relatively low (Table 3), due to the quench-

ing by the ICT reaction. The yield Φ⁰(ICT) is larger for 24DCDMA

(0.016) than for 34DCDMA (0.0077), a further indication, besides

Φ⁰(ICT)/Φ(LE), that for the former molecule the LE a ICT

equilibrium has shifted toward the ICT state. As the ﬂuorescence

band of 24DCDMA fully covers the visible spectrum (half-width

Δ(1/2) = 8670 cm^ꢀ1, between 16 970 and 25 640 cm^ꢀ1), this

molecule in MeCN is a so-called white light emitter, similar to

crystal violet lactone (CVL) in an ethyl acetate/MeCN mixture.³⁶

The intersystem crossing yield Φ(ISC) of the mnDCDMAs is

small for mn = 23 (0.08), mn = 25 (0.08), and mn = 34 (0.09) but

large for mn = 24 (0.69, via ICT), mn = 26 (0.90), and mn = 35

(0.57), see Table 3, without any obvious correlation with the

absence or presence of an ICT emission. The same is the case for

the internal conversion yield Φ(IC), being zero for mn = 26

(0.00), but clearly larger for the other mnDCDMAs.

mnDCDMAs (Table 3) reveals that the two mnDCDMAs

clearly showing ICT fluorescence in MeCN at 25 ꢀC (Figure 3b,e)

have the smallest gaps (4870 cm^ꢀ1for 24DCDMA and

4600 cm^ꢀ1for 34DCDMA), as compared with the molecules

only showing LE emission, with much larger gaps ΔE(S₁,S₂):

8960 cm^ꢀ1(23DCDMA), 10110 cm^ꢀ1(25DCDMA), 6810 cm^ꢀ1

(26DCDMA), and 7500 cm^ꢀ1(35DCDMA). A similar correlation

between the magnitude of ΔE(S₁,S₂) and the occurrence (DMABN,

2700 cm^ꢀ1, in n-hexane)²¹or absence (oDMABN, mDMABN,

TCDMA, and PCDMA) of an LE f ICT reaction has been found

before, as mentioned in the Introduction.^{7,9ꢀ11,13ꢀ15,21}This correla-

tion is an essential part of the PICT model.^{10,12ꢀ14,21,24,37}

24DCP4C and P4C. An Argument in Favor of PICT. The

absence of the ICT fluorescence with 24DCP4C in contrast to

its presence with 24DCDMA in MeCN at 25 ꢀC (Table 3),

notwithstanding the practically identical ΔE(S₁,S₂), is remi-

niscent of the strong reduction of the efficiency of the LE f

ICT reaction for P4C in MeCN as compared with

DMABN.^{7,11ꢀ13,33,38,39}The very low ICT efficiency of P4C

has been attributed to a larger amino nitrogen inversion

barrier than that of DMABN, an important argument against

the validity of the TICT model.^{11ꢀ13,26,38,39}This argument is

also valid here for 24DCP4C as compared with 24DCDMA.

Absorption and Fluorescence of the Six mnDCMIAs and

24DCIA at 25 ꢀC. mnDCMIAs and 24DCIA in n-Hexane. The

absorption and fluorescence spectra of the six mnDCMIAs in

n-hexane at 25 ꢀC are presented in Figure 4 (Table 4). The

absorption spectra consist of two bands, again attributed to S₀f S₁

and S₀fS₂transitions, as in the case of the mnDCDMAs (Table 2).

The energy gap ΔE(S₁,S₂) of the mnDCMIAs (Table 4) is some-

what (60ꢀ740 cm^ꢀ1) smaller than that of the corresponding

24DCP4C in MeCN. The fluorescence spectrum of 24DCP4C

in MeCN at 25 ꢀC (Figure S2b in SI) consists of a single LE

emission, clearly different from the dual LE + ICT fluorescence

of 24DCDMA in this solvent (Figure 3b). This is at first sight

surprising, in view of the similar spectral properties, such as the

energy gap ΔE(S₁,S₂) of the two molecules (Table 3), as will be

discussed in the next section.

Energy Gap ΔE(S₁,S₂) and ICT Reactions with mnDCDMA

in MeCN. An inspection of the energy gaps ΔE(S₁,S₂) of the

10830

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

Table 4. Data Obtained from the Fluorescence and Absorption Spectra of the Six mnDCMIAs and 24DCIA in n-Hexane at 25 ꢀC

ARTICLE

23DCMIA

24DCMIA

25DCMIA

26DCMIA

34DCMIA

35DCMIA

24DCIA

~v^max(LE) [cm^ꢀ1

Δ(1/2)(LE)^a

Δ(1/4)(LE)^b

Φ(LE)

Φ(ISC)^c

Φ(IC)^d

]

23910

3470

25630

3920

23730

3510

24640

3200

25940

3580

25300

3160

4300

0.082

27100

3130

4320

4780

5450

4850

4430

4890

0.13

0.061

0.27

0.032

0.78 ( 0.03

0.19

0.16

0.59 ( 0.06

0.28

0.85 ( 0.02

0.09

0.56 ( 0.05

0.17

0.70 ( 0.04

0.14

~v^max(S₁,abs) [cm^ꢀ1

]

27270

3510

29230

5130

27060

3970

27710

2170

29280

4410

28050

2810

29260

35580

ε

^max(S₁,abs) [M^ꢀ1cm^ꢀ1

]

~v^max(S₂,abs) [cm^ꢀ1

]

35560

10920

25540

8290

34020

24470

27460

4790

36300

9290

33590

5410

33680

22750

27590

4400

35280

15400

26480

7230

ε

^max(S₂,abs) [M^ꢀ1cm^ꢀ1

E(S₁)^e[cm^ꢀ1

ΔE(S₁,S₂)^f[cm^ꢀ1

]

25290

9240

25990

5880

28430

6320

]

^aFull width at half-maximum of the total LE ﬂuorescence band. ^bFull width at quarter-maximum of the total LE ﬂuorescence band. ^cMeasurements (at

23 ꢀC) as in refs 26 and 34. ^dΦ(IC) = 1 ꢀ Φ(LE) ꢀ Φ⁰(ICT) ꢀ Φ(ISC). ^eCrossing point of the normalized ﬂuorescence and absorption spectra.

^fEnergy diﬀerence ~v^max(S₂,abs) ꢀ ~v^max(S₁,abs).

mnDCDMAs (Table 2), except for 34DCMIA with a slightly large

gap (+40 cm^ꢀ1). A similar observation that the introduction of an

isopropyl substituent causes a decrease of ΔE(S₁,S₂) has previously

been made for the series DMABN, 4-(isopropylmethylamino)-

benzonitrile (IMABN), and 4-(diisopropylamino)benzonitrile

(DIABN), leading to the appearance of an ICT reaction with

dual LE + ICT fluorescence for IMABN and DIABN, even in

n-hexane.⁴⁰

The ﬂuorescence spectra of all six mnDCMIAs in n-hexane at

25 ꢀC (Figure 4) and down to ꢀ95 ꢀC consist of a single band,

with some vibrational structure, as also observed with the

mnDCDMAs (Figure 2). This band is again attributed to LE,

as an additional ICT emission is not found. The same is the case

for 24DCIA in n-hexane (Figure S3a in SI). The ﬂuorescence

quantum yields Φ(LE) have values between 0.032 (26DCMIA)

and 0.27 (25DCMIA). The yields Φ(ISC) range between 0.56

(25DCMIA) and 0.85 (24DCMIA), from which Φ(IC) is

calculated: from 0.09 (24DCMIA) to 0.28 (23DCMIA)

(Table 4). These results are comparable to those obtained with

the mnDCDMAs (Table 2).

molecules extends over the entire visible spectrum, from 14 700

to 24 180 cm^ꢀ1, making them potential white-light emitters,

similar to what has been reported for CVL.³⁶

25DCMIA (LE + ICT). MeCN. With 25DCMIA in MeCN at

25 ꢀC, dual LE + ICT fluorescence is observed (Figure 5c), but

with a smaller ratio Φ⁰(ICT)/Φ(LE) (2.7) than for 24DCMIA

(8.8) (Figure 5b) under these conditions (Table 5). The emis-

sion spectrum of 25DCMIA is broader (Δ(1/2) = 6040 cm^ꢀ1

,

Δ(1/4) = 8630 cm^ꢀ1) than the LE bands of 26DCMIA

(3950 cm^ꢀ1, 5620 cm^ꢀ1) and 35DCMIA (4110 cm^ꢀ1

,

5740 cm^ꢀ1), see Figure 5c,d and Table 5. Also, the spectral

shape is different from an asymmetric LE emission, more

resembling a typical symmetric ICT fluorescence band, such as

for DMABN in MeCN.⁷The overall fluorescence spectrum of

25DCMIA in MeCN can therefore be separated into its LE +

ICT contributions, resulting in a ratio Φ⁰(ICT)/Φ(LE) = 2.7

(Figure 5c, Table 5). The presence of dual LE + ICT emission is

supported by the increase of Φ⁰(ICT)/Φ(LE) to 4.6, upon

cooling to ꢀ45 ꢀC (Figure 6b). In Figure 6a,c it is shown that for

24DCMIA and 34DCMIA in MeCN at ꢀ45 ꢀC Φ⁰(ICT)/

Φ(LE) has likewise become larger (21.7 and 2.1) than at 25 ꢀC

(8.8 and 1.4), see Table 5.

mnDCMIAs and 24DCIA in MeCN. The absorption and

fluorescence spectra of the six mnDCMIAs in MeCN at 25 ꢀC

are presented in Figure 5 (Table 5).

23DCMIA (LE + ICT?). The emission band of 23DCMIA is also

26DCMIA, 35DCMIA, and 24DCIA (LE). With 26DCMIA,

35DCMIA (Figure 5d,f), and 24DCIA (Figure S3b in SI), a

single fluorescence band is observed, with the characteristic

asymmetric shape of an LE emission, comparable to the LE

spectra of 26DCDMA and 35DCDMA in MeCN (Figure 3d,f).

24DCMIA and 34DCMIA (LE + ICT). 24DCMIA and

34DCMIA in MeCN (Figure 5b,e), however, show dual LE +

ICT fluorescence, with a ratio Φ⁰(ICT)/Φ(LE) of 8.8 and 1.4,

respectively (Table 5). Note that these ratios are much larger

than those of the corresponding mnDCDMAs: 1.2 (24DCDMA)

and 0.35 (34DCDMA) in MeCN (Figure 3b,e and Table 3). This

enhancement in ICT efficiency, caused by the introduction of an

isopropyl substituent into the amino group, is similar to that

found for IMABN and DIABN as compared with DMABN,⁴⁰as

discussed above.

broader (Δ(1/2) = 5110 cm^ꢀ1; Δ(1/4) = 7230 cm^ꢀ1) and less

asymmetric than the LE spectrum of 26DCMIA (3950 cm^ꢀ1

;

5620 cm^ꢀ1) and 35DCMIA (4110 cm^ꢀ1; 5740 cm^ꢀ1), although

smaller than for 25DCMIA (6040 cm^ꢀ1; 8630 cm^ꢀ1). Never-

theless, a spectral separation was not attempted. It should,

however, be concluded that most likely a small amount of

additional ICT fluorescence is present with 23DCMIA in MeCN

at 25 ꢀC, pointing to an incipient ICT reaction.

ICT and Energy Gap ΔE(S₁,S₂) in MeCN. 24DCMIA and

34DCMIA. The appearance of an ICT emission with

24DCDMA and 34DCDMA in MeCN, different from the

other four DCDMAs, was attributed above to their relatively

small energy gap ΔE(S₁,S₂). With the mnDCMIAs in MeCN,

24DCMIA and 34DCMIA likewise have the smallest

ΔE(S₁,S₂) gap: 4690 and 4550 cm^ꢀ1. The decrease of ΔE(S₁,S₂)

in MeCN of the mnDCMIAs as compared with the mnDCDMAs

(Tables 3 and 5) is, however, quite small: only 180 cm^ꢀ1

for 24DCMIA and 50 cm^ꢀ1for 34DCMIA. An influence of

With 34DCMIA, the emission half-width is exceptionally

large, Δ(1/2) = 9480 cm^ꢀ1(Figure 5e and Table 5), as with

24DCDMA in MeCN (Figure 3b). The ﬂuorescence of these

10831

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Figure 5. Fluorescence (LE or LE + ICT) and absorption (Abs) spectra of the six m,n-dicyano-(N-methyl-N-isopropyl)anilines mnDCMIA (Chart 3)

in acetonitrile (MeCN) at 25 ꢀC. The ﬂuorescence spectra are attributed to the emission from a locally excited (LE) state, except for those of

24DCDMA, 25DCDMA, and 34DCDMA, which show dual ﬂuorescence from an LE and an intramolecular charge transfer (ICT) state (see text). The

ICT emission band has been obtained by subtraction with the shifted LE ﬂuorescence spectrum in n-hexane (Figure 4).

Table 5. Data Obtained from the Fluorescence and Absorption Spectra of the Six mnDCMIAs and 24DCIA in MeCN at 25 ꢀC

23DCMIA

24DCMIA

25DCMIA

26DCMIA

34DCMIA

35DCMIA

24DCIA

~v^max(LE) [cm^ꢀ1

~v^max(ICT) [cm^ꢀ1

Δ(1/2)(ﬂu)^a[cm^ꢀ1

Δ(1/2)(LE)^b[cm^ꢀ1

Δ(1/2)(ICT)^c[cm^ꢀ1

]

21030

-

23620

19430

4840

22060

19220

6040

22520

-

22510

17280

9480

22320

25960

-

]

-

]

5110

-

3950

-

4110

-

3430

-

]

4180

3960

4760

]

4680

5150

5320

Δ(1/4)(ﬂu)^d[cm^ꢀ1

Δ(1/4)(LE)^e[cm^ꢀ1

Δ(1/4)(ICT)^f[cm^ꢀ1

]

7230

-

7300

8630

5620

-

11570

6610

5740

-

4750

-

]

5940

5560

]

6740

7280

7420

Φ(LE)

Φ⁰(ICT)

Φ⁰(ICT)/Φ(LE)

Φ(ISC)^g

Φ(IC)^h

0.17

0

0.0025

0.022

8.8

0.0033

0.025

2.7

0.061

0

0.0054

0.0077

1.4

0.20

0

0.19 ( 0.01

0.65

26560

3430

35210

9570

24260

8650

0.83 ( 0.02

0.15

0.12 ( 0.01

0.85

0.58 ( 0.02

0.36

27930

3270

33490

4690

25080

5560

0.22 ( 0.02

0.77

~v^max(S₁,abs) [cm^ꢀ1

^max(S₁,abs) [M^ꢀ1cm^ꢀ1

~v^max(S₂,abs) [cm^ꢀ1

^max(S₂,abs) [M^ꢀ1cm^ꢀ1

E(S₁)ⁱ[cm^ꢀ1

ΔE(S₁,S₂)^j[cm^ꢀ1

]

28830

5300

26330

3790

28270

7140

27090

2830

28840

5300

ε

]

33520

24210

26480

4690

36170

8170

32830

25380

25770

4550

34470

14310

24730

7380

33540

24220

27630

5780

ε

]

24580

]

9840

^aFull width at half-maximum (fwhm) of the total ﬂuorescence band. ^bFwhm for the separated LE band. ^cFwhm for the separated ICT band. ^dFull width

at quarter-maximum (fwqm) of the total ﬂuorescence band. ^eFwqm for the separated LE band. ^fFwqm for the separated ICT band. ^gMeasurements

(at 23 ꢀC) as in refs 26 and 34. ^hΦ(IC) = 1 ꢀ Φ(LE) ꢀ Φ⁰(ICT) ꢀ Φ(ISC). ⁱCrossing point of the normalized ﬂuorescence and absorption spectra.

^jEnergy diﬀerence ~v^max(S₂,abs) ꢀ ~v^max(S₁,abs).

10832

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Scheme 1

(5560 cm^ꢀ1) and 35DCMIA (7380 cm^ꢀ1), which only show LE

emission (Figure 5). Nevertheless, efficient ICT emission with

a large Φ⁰(ICT)/Φ(LE) of 2.7 is observed, as mentioned above

(Table 5). Similarly, ΔE(S₁,S₂) is only 270 cm^ꢀ1smaller

for 25DCMIA than for 25DCDMA (Tables 3 and 5) although

an efficient LE f ICT reaction is found with 25DCMIA

(Φ⁰(ICT)/Φ(LE) = 2.7), whereas such a reaction only just

starts to appear with 25DCDMA (Figures 5c and 3c).

This observation is a clear indication that the magnitude of

ΔE(S₁,S₂) is not the only criterion which determines whether an

LE f ICT reaction will take place or not with D/A-substituted

benzenes. The appearance of an ICT reaction with 25DCMIA in

MeCN will in fact be due to the large amino twist angle θ = 42.6ꢀ

of 25DCMIA (Table 1). Such a large twist angle leads to an

eﬃcient electronic decoupling between the amino and benzoni-

trile moieties. Under these conditions, the ICT state becomes

similar to a weakly coupled intermolecular exciplex ¹(A^ꢀD⁺), as

with DTABN,³⁰mDTABN,³⁰MMD,³¹6-cyanobenzoquinucli-

dine (CBQ),^11,13,39and also CVL.³⁶Evidently, the magnitude of

ΔE(S₁,S₂) does not govern the occurrence of an ICT reaction in

these D/A molecules with electronically decoupled D and A subunits,

diﬀerent from DMABN and its strongly coupled derivatives.⁷

Intersystem Crossing and Internal Conversion. With the

mnDCDMAs as well as the mnDCMIAs, ISC and IC are important

deactivation channels at 25 ꢀC (Tables 2ꢀ5).^15,34,37In n-hexane,

Φ(ISC) of the mnDCDMAs ranges between 0.45 and 0.81 and

between 0.56 and 0.85 for the mnDCMIAs, with Φ(IC) between

0.12 and 0.46 (mnDCDMA) and between 0.09 and 0.28

(mnDCMIAs). In MeCN, Φ(ISC) of the mnDCDMAs lies

between 0.08 and 0.90 and between 0.12 and 0.83 for the

mnDCMIAs, with Φ(IC) between 0 and 0.88 (mnDCDMA) and

between 0.15 and 0.85 (mnDCMIAs). Note that in n-hexane ISC and

IC originate from S₁(LE), whereas with 24DCDMA, 34DCDMA,

24DCMIA, 34DCMIA, and 25DCMIA in MeCN these deactivation

processes start from the LE as well as the ICT state.

Figure 6. Fluorescence (LE+ICT) of (a) 24DCMIA, (b) 25DCMIA,

and (c) 34DCMIA in MeCN at ꢀ45 ꢀC.

ΔE(S₁,S₂) can hence not be reponsible for the increase in

Φ⁰(ICT)/Φ(LE).

A second consequence of the presence of an isopropyl group

in 24- and 34DCMIA, however, is the fact that iPrNMe₂is a

better electron donor than trimethylamine N(Me)₃, a model for

the donor substituent in the mnDCDMAs. This can be deduced

from their vertical ionization energies (vIE): 8.20 eV (iPrN(Me)₂)

and 8.53 eV (N(Me)₃).⁴¹Other values of vIE in the literature are

8.00 eV^42afor iPrN(Me)₂and 8.44 eV^42bfor N(Me)₃, leading to

the same conclusion. An oxidation potential of iPrN(Me)₂is not

available. Via extrapolation in a plot of vIE vs E(D/D⁺) for

aliphatic amines (Table S1 and Figure S4 in SI), an oxidation

potential of 0.84 V vs SCE could be determined, showing that

iPrN(Me)₂is indeed a better donor than N(Me)₃(1.05 V vs

SCE).¹⁰As the calculated energy E(ICT) of the ICT state of the

six mnDCDMAs in n-hexane as well as MeCN is already lower

than E(S₁),¹⁰the same will certainly be the case for the

mnDCMIAs: E(ICT) is 0.21 eV = 20 kJ/mol lower (diﬀerence

in E(D/D⁺)),whereas E(S₁) is only slightly smaller (at most 0.08 eV,

Tables 2ꢀ5). The absence of an LE f ICT reaction with the

mnDCDMAs and mnDCMIAs in n-hexane therefore makes clear

that E(ICT) can not be calculated employing the redox potentials of

the D and A subgroups,¹⁰which means that the D and A units in

the dicyanoanilines are not electronically decoupled.

Φ⁰(ICT)/Φ(LE) as a Function of Temperature. StevensꢀBan

Plots. For systems with two kinetically interconverting states, such

as LE and ICT in Scheme 1, the ratio Φ⁰(ICT)/Φ(LE) can be

expressed by the following equation.^{7,24,34,40,43,44}

Φ⁰ðICTÞ=ΦðLEÞ ¼ k_f⁰ðICTÞ=k_fðLEÞfk_a=ðk_dþ 1=τ⁰_oðICTÞÞg

ð1Þ

When k_d. 1/τ⁰_o(ICT), with a small ICT reaction enthalpy

difference ꢀΔH,⁴⁴eq 1 simplifies the high-temperature limit

(HTL) to eq 2.

Φ⁰ðICTÞ=ΦðLEÞ ¼ k_f⁰ðICTÞ=k_fðLEÞk_a=k_d

ð2Þ

Alternatively, when k_d, 1/τ⁰_o(ICT), the low-temperature limit

(LTL) (large ꢀΔH),⁷eq 3, holds.

25DCMIA. Especially for 25DCMIA (9840 cm^ꢀ1), but also for

23DCMIA (8650 cm^ꢀ1), the energy gap ΔE(S₁,S₂) in MeCN is

considerably larger than that of 24DCMIA (4690 cm^ꢀ1) and

34DCMIA (4550 cm^ꢀ1) but also larger than for 26DCMIA

Φ⁰ðICTÞ=ΦðLEÞ ¼ ðk_f⁰ðICTÞ=k_fðLEÞÞk_aτ⁰_oðICTÞ

ð3Þ

In eqs 1ꢀ3 and Scheme 1, k_aand k_dare the rate constants of the

forward and backward ICT reaction; τ₀(LE) is the fluorescence

10833

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Figure 7. StevensꢀBan (SB) plots of Φ⁰(ICT)/Φ(LE) vs 1000/T in

MeCN of (a) 24DCDMA, (b) 24DCMIA, and (c) 34DCMIA. The ICT

reaction enthalpies ΔH of the LE f ICT reaction are indicated in the

ﬁgures.

Table 6. ICT Stabilization Enthalpy Diﬀerence ꢀΔH(SB)

and ICT/LE Fluorescence Quantum Yield Ratio

Φ⁰(ICT)/Φ(LE) at 25 ꢀC for 24DCDMA, 24DCMIA, and

34DCMIA in MeCN

D/A molecule

ꢀΔH(SB) (kJ/mol)

Φ⁰(ICT)/Φ(LE)

24DCDMA

24DCMIA

5.4

4.7

3.9

1.2^a

8.8^b

1.4^b

34DCMIA

^aTable 3. ^bTable 5.

lifetime of the model compound (no ICT); τ⁰₀(ICT) is the

fluorescence lifetime of the ICT state; and k_f(LE) and k⁰_f(ICT)

10834

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Table 8. Data from the Solvatochromic Analysis of the LE or ICT Fluorescence Maxima ~v^maxof the mnDCDMAs

D/A molecule

F^a[Å]

μ_g^a[D]

slope (LE)

μ_e(LE)^a[D]

slope (ICT)

μ_e(ICT)^a[D]

23DCDMA (eq 4)

4.43

7.5

ꢀ10480 ( 800

1.13 ( 0.03

1.73 ( 0.11

ꢀ7760 ( 650

0.44 ( 0.04

1.20 ( 0.08

ꢀ10600 ( 950

0.61 ( 0.07

1.76 ( 0.10

ꢀ6350 ( 850

0.37 ( 0.04

1.08 ( 0.10

ꢀ11060 ( 750

0.63 ( 0.04

1.82 ( 0.10

ꢀ11080 ( 950

0.63 ( 0.06

1.83 ( 0.10

ꢀ5840 ( 650

0.94 ( 0.03

0.34 ( 0.04

14.0 ( 0.4

23DCDMA (vs DIABN)

23DCDMA (vs DMABN)

24DCDMA (eq 4)

4.43 (4.68)

4.43 (4.20)

4.43

7.5 (6.8)

7.5 (6.6)

5.1

14.5 ( 0.4 (18)

12.8 ( 0.3 (10)

11.1 ( 0.4

ꢀ12360 ( 5120

0.66 ( 0.30

13.2 ( 2.1

24DCDMA (vs DIABN)

24DCDMA (vs DMABN)

25DCDMA (eq 4)

4.43 (4.68)

4.43 (4.20)

4.43

5.1 (6.8)

5.1 (6.6)

1.4

11.6 ( 0.4 (18)

10.2 ( 0.4 (10)

10.3 ( 0.5

13.5 ( 2.4 (18)

13.7 ( 0.9 (17)

0.56 ( 0.10

25DCDMA (vs DIABN)

25DCDMA (vs DMABN)

26DCDMA (eq 4)

4.43 (4.68)

4.43 (4.20)

4.43

1.4 (6.8)

1.4 (6.6)

3.3

10.9 ( 0.6 (18)

9.1 ( 0.3 (10)

9.3 ( 0.5

26DCDMA (vs DIABN)

26DCDMA (vs DMABN)

34DCDMA (eq 4)

4.43 (4.68)

4.43 (4.20)

4.43

3.3 (6.8)

3.3 (6.6)

7.8

9.8 ( 0.5 (18)

8.4 ( 0.3 (10)

14.4 ( 0.4

ꢀ15350 ( 2500

0.79 ( 0.20

16.1 ( 0.9

34DCDMA (vs DIABN)

34DCDMA (vs DMABN)

35DCDMA (eq 4)

4.43 (4.68)

4.43 (4.20)

4.43

7.8 (6.8)

7.8 (6.6)

5.2

15.0 ( 0.4 (18)

13.3 ( 0.3 (10)

12.7 ( 0.4

16.2 ( 0.9 (18)

16.8 ( 0.4 (17)

0.73 ( 0.10

35DCDMA (vs DIABN)

35DCDMA (vs DMABN)

24DCP4C (eq 4)

4.43 (4.68)

4.43 (4.20)

4.53

5.2 (6.8)

5.2 (6.6)

5.9

13.3 ( 0.5 (18)

11.6 ( 0.3 (10)

10.9 ( 0.4

24DCP4C (vs DIABN)

24DCP4C (vs DMABN)

4.53 (4.68)

5.9 (6.8)

5.9 (6.6)

9.9 ( 0.1 (18)

11.3 ( 0.5 (10)

4.53 (4.20)

^aValues in parentheses are for DIABN or DMABN. Ground state dipole moment μ_gcalculated by PM3; results obtained by scaling via a comparison of

the calculated μ_g(DMABN) = 5.94 D (PM3) and its experimental value of 6.6 D (refs 4 and 52).

n²ꢀ 1

are the radiative rate constants. The reciprocal of these lifetimes is

fðn²Þ ¼

ð6Þ

equal to the sum of their radiative and nonradiative rate constants.

In Figure 7, StevensꢀBan (SB)⁷plots, ln(Φ⁰(ICT)/Φ(LE))

vs 1000/T, are presented for (a) 24DCDMA, (b) 24DCMIA,

and (c) 34DCMIA in MeCN. Φ⁰(ICT)/Φ(LE) increases upon

cooling, which means that the systems are in the HTL condition

(eq 2), with k_d. 1/τ⁰₀(ICT), which appears with relatively

small ꢀΔH values,⁴such as the ꢀH(SB)s determined here

(Table 6): 5.4 kJ/mol (24DCDMA), 4.7 kJ/mol (24DCMIA), and

3.9 kJ/mol (34DCMIA). For DMABN in MeCN, by comparison,

ꢀΔH = 27.0 kJ/mol,⁷whereas for DMABN in toluene

(HTL),^43,4ꢀΔH = 11.6 kJ/mol.

2n²þ 1

In eqs 4ꢀ6, μ_gand μ_eare the ground and excited state dipole

moments; ε is the dielectric constant; n is the refractive index; h is

Planck’s constant; c is the speed of light; and F is the Onsager

radius^46,52of the solute.

Six mnDCDMAs. The fluorescence maxima ~v^max(LE) and

~v^max(ICT) of the six mnDCDMAs in a series of solvents at

25 ꢀC (Table 7) are plotted vs the~v^max(ICT) of DIABN in Figure

S5a,c,e,g,i,h (SI) and also against the solvent polarity parameter

g(ε,n) (eqs 4ꢀ6) in Figure S5b,d,f,h,j,l (SI). The LE and ICT

Solvatochromic Measurements. Excited State Dipole Mo-

ments μ_e(LE) and μ_e(ICT). For the six mnDCDMAs and

mnDCMIAs, the maxima ~v^max(LE) and ~v^max(ICT) of the LE

and ICT fluorescence bands of these compounds were measured

in a series of solvents, from the nonpolar n-hexane (ε²⁵= 1.88)

to the strongly polar MeCN (ε²⁵= 36.7), see Table 7. From

these data, μ_e(ICT) can be determined from the slope of a plot

of ~v^max(ICT) vs the solvent polarity parameter g(ε,n) = f(ε) ꢀ

1/2f(n²), whereas μ_e(LE) can be obtained from the slope of

the plot of ~v^max(LE) vs the solvent polarity parameter g(ε,n) =

f(ε) ꢀ f(n²) (eqs 4ꢀ6).^{26,40,44ꢀ49}This difference in g(ε,n)

appears while the ICT state is not directly populated by absorption

in S₀but is formed from the initially excited LE precursor (see

below).^26,40,46

emission maxima are, as usual,^15,44,50also plotted against the ~v^max

-

(ICT) of DIABN(μ_e(ICT) = 18 D).⁵¹With this procedure, the

scatter in the data points is generally reduced by mutually compensat-

ing the specific solute/solvent interactions.^26,40,48,51

For the four mnDCDMAs, 23DCDMA, 25DCDMA,

26DCDMA, and 35DCDMA, which only show a single LE

ﬂuorescence in n-hexane and MeCN (Figures 2a,c,d,f and 3a,c,

d,f), the following dipole moments μ_e(LE) are calculated, taking

the mean value of the results from the plots of ~v^max(LE) vs

~v^max(ICT,DIABN) and g(ε,n) = f(ε) ꢀ f(n²) (eqs 4ꢀ6): 14 D

(23DCDMA), 11 D (25DCDMA), 10 D (26DCDMA), and 13

D (35DCDMA) (see Table 8).

For the two DCDMAs showing dual LE + ICT ﬂuores-

cence in MeCN (Figure 3b,e), the averaged μ_e(LE) values are

11 D (24DCDMA) and 15 D (34DCDMA), whereas from

the plots of ~v^max(ICT) vs g(ε,n) = f(ε) ꢀ 1/2f(n²) somewhat

larger dipole moments μ_e(ICT) are found: 13 D for

24DCDMA and 16 D for 34DCDMA (Table 8). The increase

in dipole moment when going from LE to ICT is for

24DCDMA and 34DCDMA smaller than that observed for

1

max

ν~ ðfluÞ ¼ ꢀ

μ_eðμ_eꢀ μ_gÞgðε, nÞ þ const:

ð4Þ

ð5Þ

2hcF³

ε ꢀ 1

fðεÞ ¼

2ε þ 1

10835

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

DMABN, for which μ_e(LE) = 10 D and μ_e(ICT) = 17 D.⁵²

The reason for this perhaps unexpected diﬀerence may

ﬁnd its origin in possibly diﬀerent dipole moment directions

of the LE and ICT states in the dicyano-substituted

compounds.⁵³This small diﬀerence between μ_e(LE) and μ_e(ICT)

for 24DCDMA and 34DCDMA also shows that not the absolute

magnitude of a dipole moment determines whether its origin is

an LE or an ICT state, as stressed in our deﬁnition of an LE and

ICT states in ref 1. For DMABN in n-hexane (no ICT emission),

μ_e(LE) is already relatively large (10 D), which does not mean

that it is the dipole moment of an ICT state.

Comparison with Calculations. From computations¹⁸with

34DCDMA (gas phase), the following results are obtained: μ_g(S₀)

= 10.7 D, μ_e(S₁,TICT) = 23.7 D, and μ_e(S₂,LE) = 21.2 D. It should

be noted, as mentioned in the Introduction, that the lowest excited

and hence emitting state is considered to be a TICT state, without

the occurrence of dual fluorescence,¹⁸different from our observa-

tion. These calculated dipole moment data are substantially larger

than our results (Table 8): μ_g(S₀) = 7.8 D (PM3), μ_e(S₁,LE) = 15 D

(vs DIABN), and μ_e(ICT) = 16 D. The computations¹⁸with

35DCDMA (gas phase, only LE) also result in larger dipole moments

than those obtained from our investigations: μ_g(S₀) = 7.5, μ_e(S₁,CT)

= 19.4 D, and μ_e(S₂,LE) = 19.4 D (ref 18) as compared with μ_g(S₀) =

5.2 D (PM3) and μ_e(LE) = 13 D (vs DIABN) (see Table 8). The

difference between the calculated and experimental dipole moments

μ_e(ICT) and μ_e(LE) could indicate that the most stable molecular

structure obtained in the computations is not the one encountered in

our experiments. For a similar conclusion in the case of DMABN, see

ref 21.

Six mnDCMIAs. For the six mnDCMIAs(Table 9), thesolvato-

chromic plots vs ~v^max(ICT,DIABN) and g(ε,n) are presented in

Figures S6a,c,e,g,i,k and S61b,d,f,h,j,l (SI) (eqs 4ꢀ6). The aver-

aged dipole moments μ_e(LE) (Table 10) are: 15 D (23DCMIA),

12 D (24DCMIA), 9 D (25DCMIA), 10 D (26DCMIA), 16 D

(34DCMIA), and 13 D (35DCMIA). These LE dipole moments

are comparable with those obtained for the corresponding

mnDCDMAs (Table 8). The ICT dipole moments μ_e(ICT) of

the three mnDCMIAs for which a separation of the overall

fluorescence spectrum in an LE and ICT contribution could be

made (Figure 5b,c,e), are again somewhat larger than their LE

dipole moments: 15 D for 24DCMIA, 11 D for 25DCMIA, and

18 D for 34DCMIA (Table 10).

24DCP4C. The dipole moment of 24DCP4C has an average

value of 11 D (μ_e(LE), Table 8), supporting the identification of

its fluorescence as a single LE emission, as discussed above.

LE Fluorescence Decays with the mnDCDMAs and

mnDCMIAs. Literature Data. Picosecond fluorescence decays,

with a time resolution of 5 ps, have appeared in the literature for

35DCDMA in MeCN and diethyl ether (DEE) as well as for

34DCDMA in DEE. These decays are single exponential, with

times τ₁in the nanosecond domain, without any indication of a

second, shorter picosecond decay time.^9,11For 35DCDMA, τ₁

equals 18.8 ns in MeCN⁹at 25 ꢀC and 16.9 ns in DEE¹¹at

ꢀ89 ꢀC, whereas for 34DCDMA in DEE¹¹τ₁is 6.88 ns at 25 ꢀC

and 10.2 ns at ꢀ100 ꢀC. In view of their single exponential

character, these decays were attributed to LE. It was then concluded

that an ICT reaction does not take place under the experimental

conditions, caused by the relatively large energy gap ΔE(S₁,S₂) of

35DCDMA and 34DCDMA, as discussed above.^9,11ꢀ14

Nanosecond Fluorescence Decays of mnDCDMAs and

mnDCMIAs Without ICT Emission. n-Hexane and MeCN.

The LE fluorescence decays of all mnDCDMAs and mnDCMIAs

10836

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Table 10. Data from the Solvatochromic Analysis of the LE or ICT Fluorescence Maxima ~v^maxof the mnDCMIAs

D/A molecule

F^a[Å]

μ_g^a[D]

slope (LE)

μ_e(LE)^a[D]

slope (ICT)

μ_e(ICT)^a[D]

23DCMIA (eq 4)

4.66

7.5

ꢀ9920 ( 950

0.57 ( 0.04

1.56 ( 0.12

ꢀ6700 ( 150

0.378 ( 0.02

1.06 ( 0.08

ꢀ5390 ( 1020

0.31 ( 0.07

0.91 ( 0.10

ꢀ6800 ( 900

0.38 ( 0.06

1.07 ( 0.13

ꢀ11410 ( 700

0.65 ( 0.02

1.79 ( 0.16

ꢀ9980 ( 900

0.57 ( 0.04

1.57 ( 0.12

14.4 ( 0.5

23DCMIA (vs DIABN)

23DCMIA (vs DMABN)

24DCMIA (eq 4)

4.66 (4.68)

4.66 (4.20)

4.66

7.5 (6.8)

7.5 (6.6)

6.1

15.0 ( 0.4 (18)

13.0 ( 0.4 (10)

11.8 ( 0.1

ꢀ13520 ( 2500

0.69 ( 0.12

15.1 ( 1.0

24DCMIA (vs DIABN)

24DCMIA (vs DMABN)

25DCMIA (eq 4)

4.66 (4.68)

4.66 (4.20)

4.66

6.1 (6.8)

6.1 (6.6)

2.5

12.2 ( 0.3 (18)

10.7 ( 0.3 (10)

8.7 ( 0.7

15.2 ( 1.0 (18)

15.4 ( 0.4 (17)

11.5 ( 0.9

0.60 ( 0.04

ꢀ10230 ( 1740

0.51 ( 0.11

25DCMIA (vs DIABN)

25DCMIA (vs DMABN)

26DCMIA (eq 4)

4.66 (4.68)

4.43 (4.20)

4.66

2.5 (6.8)

2.5 (6.6)

2.8

9.2 ( 0.8 (18)

7.9 ( 0.4 (10)

9.8 ( 0.5

11.4 ( 1.0 (18)

11.5 ( 0.9 (17)

0.43 ( 0.08

26DCMIA (vs DIABN)

26DCMIA (vs DMABN)

34DCDMA (eq 4)

4.66 (4.68)

4.66 (4.20)

4.66

2.8 (6.8)

2.8 (6.6)

8.2

10.2 ( 0.7 (18)

8.6 ( 0.4 (10)

15.5 ( 0.3

ꢀ18010 ( 2130

0.89 ( 0.15

18.2 ( 0.8

34DCMIA (vs DIABN)

34DCMIA (vs DMABN)

35DCMIA (eq 4)

4.66 (4.68)

4.66 (4.20)

4.66

8.2 (6.8)

8.2 (6.6)

5.5

16.2 ( 0.2 (18)

14.1 ( 0.4 (10)

13.1 ( 0.5

18.0 ( 1.0 (18)

18.5 ( 0.8 (17)

0.79 ( 0.10

35DCMIA (vs DIABN)

4.66 (4.68)

5.5 (6.8)

13.7 ( 0.4 (18)

11.7 ( 0.3 (10)

35DCMIA (vs DMABN)

4.66 (4.20)

5.5 (6.6)

^aValues in parentheses are for DIABN or DMABN. Ground state dipole moment μ_gcalculated by PM3; results obtained by scaling via a comparison of

the calculated μ_g(DMABN) = 5.94 D (PM3) and its experimental value of 6.6 D (refs 4 and 52).

in n-hexane at 25 ꢀC are single exponential, with times τ₁in the

nanosecond time range, between 0.82 ns (26DCDMA) and

9.77 ns (25DCMIA) (Tables 11 and 12). In MeCN at 25 ꢀC,

single exponential nanosecond decays are likewise observed with the

dicyanoanilines for which dual LE + ICT fluorescence could not be

detected with certainty: 23DCDMA, 25DCDMA, 26DCDMA, and

35DCDMA (Figure 3a,c,d,f) as well as 23DCMIA, 26DCMIA, and

35DCMIA (Figure 5a,d,f) (see Tables 11 and 12). These results

support our conclusion based on the absence of an ICT emission

(see Figures 2ꢀ5) that an LE f ICT reaction does not take place

with these molecules. k⁰_f(ICT) is smaller than k_f(LE) (Tables 11

and 12), as generally observed,^{4,6,7,30,34,43}because of the larger CT

character of the ICT as compared with the LE state.¹⁴

shows a growing-in (Figure 8b, A₂₂= ꢀ1.8, eq 8). For 34DCDMA

in DEE at ꢀ110 ꢀC (Table 13), the LE and ICT decays are

double exponential. The ICT reaction is made possible by the

strong increase in solvent polarity upon cooling, from ε²⁵= 4.24 to

ε

^ꢀ110= 9.79.

i_fðLEÞ ¼ A₁₁expð ꢀ t=τ₁Þ þ A₁₂expð ꢀ t=τ₂Þ

ð7Þ

i_fðICTÞ ¼ A₂₁expð ꢀ t=τ₁Þ þ A₂₂expð ꢀ t=τ₂Þ

with A₂₂¼ ꢀ A₂₁

ð8Þ

ð9Þ

A ¼ A₁₂=A₁₁

Double Exponential LE and ICT Picosecond Decays.

24DCDMA, 24DCMIA, 34DCMIA, and 25DCMIA. The picose-

cond fluorescence decays of 24DCDMA, 24DCMIA, and

34DCMIA in MeCN at 25 ꢀC (Figures 8ꢀ10), which undergo

an ICT reaction and show dual LE + ICT fluorescence (Figures 3b

and 5b,c,e), can not adequately be fitted with a single exponential.

The LE and ICT fluorescence decays in MeCN of 24DCDMA are

shown in Figure 8 (25 ꢀC), of 24DCMIA in Figure 9 (25 ꢀC), and

of 34DCMIA in Figure 10 (25 and ꢀ45 ꢀC). Also for 25DCMIA

in THF, n-propylcyanide (PrCN), and MeCN at lower tempera-

tures (Figure 11), double exponential LE and ICT fluorescence

decays are observed, supporting our conclusion discussed above

that its fluorescence spectrum consists of overlapping LE and ICT

bands (Figure 5c).

24DCDMA in MeCN at 25 ꢀC. The LE and ICT picosecond

fluorescence decays of 24DCDMA in MeCN at 25 ꢀC are shown

in Figure 8. From a global analysis, two times are obtained: τ₂=

1.8 ps and τ₁= 3800 ps (eqs 7ꢀ9). The shortest time τ₂(taken

from femtosecond transient spectra, see Figures 14 and S7 (SI)

below) is kept fixed. The minor additional decay time of 153 ps is

attributed to photoproducts, other impurities, or experimental

problems and is hence not taken into account in the data analysis

(as done before).^40,55The ICT fluorescence response function

As τ₂, τ₁, the amplitude ratio A = A₁₂/A₁₁(eq 9) of the LE decay

(Figure 8a) is approximately equal to the ratio of the forward and

backward ICT rate constants k_a/k_d(Scheme 1).^7,34In such a case,

also 1/τ₂≈ (k_a+ k_d).^7,34From the data in Figure 8a, τ₂= 1.8 ps

and A = 7.58, the rate constants k_a= 49 ꢁ 10¹⁰

s

^ꢀ1and k_d= 6.5 ꢁ

10¹⁰s^ꢀ1are then calculated (Table 13).

These results show that not only the LE f ICT forward reaction

of 24DCDMA in MeCN at 25 ꢀC is very fast, but that this also is the

case for the ICT f LE back reaction, a clear indication that ꢀΔH is

relatively small, in accordance with ꢀΔH(SB) = 5.4 kJ/mol

(Table 6). The connection between k_dand ꢀΔH is illustrated by

the following data. For NTC6 in n-hexane at 25 ꢀC, with a very small

ꢀΔH of 2.4 kJ/mol, k_d= 32.2 ꢁ 10¹⁰s^{ꢀ1 24}

140 times smaller than for 24DCDMA (6.5 ꢁ 10¹⁰s^ꢀ1), due to a

much a larger ꢀΔH = 27.0 kJ/mol than the 5.4 kJ/mol for

24DCDMA (Table 6).⁷

. With DMABN in

MeCN at 25 ꢀC, for comparison, k_d(0.047 ꢁ 10¹⁰

s

^ꢀ1) is nearly

24DCMIA in MeCN at 25 ꢀC. The LE and ICT picosecond

fluorescence decays of 24DCMIA in MeCN at 25 ꢀC are shown in

Figure 9. From the LE decay (Figure 9a), the times τ₂= 3.0 ps and

τ₁= 5460 ps are obtained, with an amplitude ratio A₁₂/A₁₁= 24.4

(Table 13). From these data (see preceding section), k_a= 32 ꢁ

10¹⁰s^ꢀ1and k_d= 1.3 ꢁ 10¹⁰

s

^ꢀ1are calculated. For 24DCMIA, k_d

10837

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Table 11. Fluorescence Quantum Yields Φ(ﬂu), Nanose-

cond Decay Times τ₁, and Radiative Rate Constants k_ffor the

Six mnDCDMAs in n-Hexane, DEE, and MeCN at 25 ꢀC

(ꢀ45 ꢀC), to 21.8 ps (ꢀ115 ꢀC), whereas the corresponding

amplitude ratio A₁₂/A₁₁increases: 5.9, 24.9, and 75.2 (Table 13).

This means that the ratio k_a/k_d(≈A₁₂/A₁₁)⁷also increases,

because k_ddecreases stronger than k_a, as the activation barrier E_d

is larger than E_a, which is always the case when a molecule

undergoes an ICT reaction, i.e., when ΔH (= E_aꢀ E_d) is

negative.^{4,7,34,43,44,56ꢀ58}Note that ΔH for DEE is not a constant,

as its polarity and hence ꢀΔH increase upon cooling, from

k_f(LE)^cor k⁰_f(ICT)^c

mnDCDMA

solvent

Φ(ﬂu)^a

τ₁^b[ns]

[10⁷s^ꢀ1

]

23DCDMA n-hexane 0.13

3.27

4.0 (LE)

DEE

0.33

0.28

ε

²⁵= 4.24 to ε^ꢀ115= 11.55.

MeCN

13.3

2.1 (LE)

4.5 (LE)

2.4 (LE)

0.34 (LE)

34DCMIA in MeCN at 25 and ꢀ45 ꢀC. A global analysis of the

24DCDMA n-hexane 0.065

1.45

LE and ICT picosecond fluorescence decays of 34DCMIA in

MeCN at 25 ꢀC is presented in Figure 10a, with a growing-in

(negative A₂₂, eq 8) for the ICT fluorescence response function.

From the times τ₂= 4.6 ps and τ₁= 1360 ps, together with A₁₂/A₁₁

DEE

0.077

3.15

MeCN

0.013 (LE)

0.016 (ICT)

3.80^d,e

8.31

0.42^c,^f(ICT)

= 5.5, k_a= 18 ꢁ 10¹⁰

s

^ꢀ1and k_d= 3.3 ꢁ 10¹⁰

s

^ꢀ1are obtained

25DCDMA n-hexane 0.31

3.7 (LE)

DEE

0.45

(Table 13). Both rate constants become smaller upon cooling, to be

expected for activated processes. At ꢀ45 ꢀC, k_a= 14 ꢁ 10¹⁰s^ꢀ1and

MeCN

0.046

2.90

0.82

1.6 (LE)

8.0 (LE)

k_d= 0.76 ꢁ 10¹⁰

s

^ꢀ1(Table 13), calculated from τ₂= 7.0 ps, τ₁=

890 ps, and A₁₂/A₁₁= 17.6 (see Figure 10b).

26DCDMA n-hexane 0.066

DEE

0.055

0.096

25DCMIA in MeCN, PrCN, and THF at Different Tempera-

tures. MeCN. Picosecond LE and ICT fluorescence decays of

25DCMIA in MeCN were measured from ꢀ44 to ꢀ20 ꢀC (Table

S2, SI). The LE decays have a predominant ultrashort time τ₂.

Even at ꢀ44 ꢀC, τ₂= 1.4 ps (Figure 11a), a time definitely below

the experimentaltime resolution (≈3 ps).^7,10,15,44Thatthese short

decay times can nevertheless be readily observed is caused by the

very high amplitude ratio A₁₂/A₁₁(Table S2, SI). As the ICT

reaction time generally becomes longer and hence better acces-

sible in solvents less polar than MeCN (ε²⁵= 36.7, ε^ꢀ44= 49.9),

25DCMIA was investigated in PrCN (ε²⁵= 24.2, ε^ꢀ40= 34.0) and

THF (ε²⁵= 7.39, ε^ꢀ60= 11.0).⁵⁹In n-hexane (ε²⁵= 1.88), an ICT

reaction does not take place (Figure 4c).

PrCN. LE and ICT fluorescence decays in PrCN were mea-

sured at temperatures between ꢀ40 and ꢀ110 ꢀC. The short

decay time τ₂increases from 2.5 ps at ꢀ40 ꢀC (Figure 11b) to

5.1 ps at ꢀ70 ꢀC (Table S2, SI). In PrCN at lower temperatures

(below ꢀ75 ꢀC for N-phenylpyrrole (PP)),⁵⁷the solvent di-

electric relaxation has to be taken into account, as it starts to

interfere with the LE f ICT reaction.

MeCN

3.07

3.1 (LE)

5.4 (LE)

4.9 (LE)

2.0 (LE)

0.70^c(ICT)

3.1 (LE)

34DCDMA n-hexane 0.17

3.15

DEE

0.34

6.84^d

1.10^d

3.08

MeCN

0.022 (LE)

0.0077 (ICT)

35DCDMA n-hexane 0.095

DEE

0.21

MeCN

0.16

18.6

0.86 (LE)

^aLE ﬂuorescence quantum yield Φ(LE), unless otherwise indicated.

Data for n-hexane are from Table 2 and those for MeCN from Table 3.

^bFrom nanosecond ﬂuorescence decays, with a time resolution better

than 0.5 ns (ref 54). For the picosecond decays of 24DCDMA and

c

34DCDMA in MeCN, see footnote d. Radiative rate constant: k_f=

Φ(ﬂu)/τ₁. In the case of the dual ﬂuorescent molecules 24DCDMA and

34DCDMA in MeCN, only approximate values are listed for k_f(LE) and

k⁰_f(ICT), as the expressions for these rate constants are more complex

(ref 34). See footnote e. ^dTime resolution ≈3 ps (ref 7). ^eSee Figure 8.

^fThe ICT lifetime τ⁰_o(ICT) = 3.44 ns (Table 13), hence k⁰_f(ICT) =

Φ⁰(ICT)/τ⁰_o(ICT) = 0.47 ꢁ 10⁷s^ꢀ1

.

THF. The LE and ICT fluorescence decays at ꢀ60 ꢀC are

shown in Figure 11c: τ₂= 7.4 ps, τ₁= 5160 ps, A = 9.0 (Table S2,

SI). The temperature dependence between ꢀ60 and ꢀ105 ꢀC is

presented in Figure 12a. From these data k_a, k_d, and τ⁰_o(ICT) are

calculated (Table S2, SI). Arrhenius plots of these data are

depicted in Figure 12b: E_a= 4.5 kJ/mol, E_d= 11.9 kJ/mol, i.e.,

ꢀΔH = 7.4 kJ/mol (Table 14). From the preexponential factors

is smaller than for 24DCDMA (k_d= 6.5 ꢁ 10¹⁰

s

^ꢀ1, Table 13).

Replacing a dimethylamino substituent by isopropylmethylamino

results in a larger ICT efficiency, i.e., in a larger ꢀΔH, suchaswiththe

pair DMABN and IMABN (see above).⁴⁰This is, however, not

the case for 24DCDMA and 24DCMIA, having about the same

ꢀΔH(SB): 5.4 and 4.7 kJ/mol (Table 6). Nevertheless, a larger

Φ⁰(ICT)/Φ(LE) of 8.8 is observed for 24DCMIA than for

24DCDMA (1.2); see Tables 3 and 5. The ICT decay of 24DCMIA

can in fact be fitted with a single exponential: τ₁= 5120 ps

(Figure 9b,c). The short time τ₂= 3.0 ps only has a fractional

contribution of 3/5460 = 0.055% (for equal amplitudes ꢀA₂₂

and A₂₁, eq 8) to the overall ICT fluorescence response function.

24DCMIA in DEE at 25, ꢀ45, and ꢀ115 ꢀC. From the LE and

ICT fluorescence decays of 24DCMIA in DEE at 25 ꢀC (Table 13),

the decay times τ₂= 3.4 ps and τ₁= 5250 ps, with A₁₂/A₁₁= 5.9

(eqs 7ꢀ9), are determined by global analysis. These data result in

k_a=25ꢁ 10¹⁰s^ꢀ1and k_d=4.3ꢁ 10¹⁰s^ꢀ1. For 24DCMIA in MeCN

k_a⁰= 1.6 ꢁ 10¹²s^ꢀ1and k_d= 1.2 ꢁ 10¹³s^ꢀ1, ΔS = R ln(k_a⁰/k_d) =

ꢀ17 J K^ꢀ1mol^ꢀ1is calculated. Comparable ΔS results have been

found for other D/A molecules: ꢀ38 J K^ꢀ1mol^ꢀ1(DMABN in

MeCN),⁷ꢀ28 J K^ꢀ1mol^ꢀ1(PP in MeCN),⁵⁷and ꢀ23 J K^ꢀ1

mol^ꢀ1(PP in PrCN).⁵⁷

0

ICT Lifetimes τ⁰_o(ICT) of 25DCMIA in THF and PrCN. The

ICT lifetime τ⁰_o(ICT) of 25DCMIA in THF decreases when the

temperature is lowered: from 4.81 ns at ꢀ60 ꢀC to 3.93 ns at

ꢀ105 ꢀC (Table S2, SI). In PrCN, τ⁰_o(ICT) similarly decreases

from 4.13 ns at ꢀ40 ꢀC to 3.78 ns at ꢀ101 ꢀC. As 1/τ⁰_o(ICT) =

k⁰_f(ICT) + k⁰(ISC) + k⁰(IC), this could be caused by k⁰_f(ICT),

which becomes larger upon cooling (k⁰_f(ICT) ꢁ n²),^58,60ꢀ62

whereas k⁰(ISC) and k⁰(IC) will decrease or remain unchanged.

Shape of the LE and ICT Picosecond Fluorescence Decay

Curves. The LE fluorescence decay curves of the m,n-dicyanoa-

nilines in Figures 8ꢀ11 have a clearly different shape than the

at 25 ꢀC (Table 13), k_dis smaller (1.3 ꢁ 10¹⁰

s

^ꢀ1), due to the fact

that the stabilization of the ICT state (ꢀΔH) will be larger in the

polar MeCN (ε²⁵= 36.7) than in DEE (ε²⁵= 4.24).^7,44,56

The LE decay time τ₂of 24DCMIA in DEE becomes longer

when the temperature is lowered, from 3.4 ps (25 ꢀC), via 4.8 ps

10838

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Table 12. Fluorescence Quantum Yields Φ(ﬂu), Nanoseond Decay times τ₁, and Radiative Rate Constants k_ffor the Six

mnDCMIAs in n-Hexane, DEE, and MeCN at 25 ꢀC

mnDCMIA

solvent

Φ(ﬂu)^a

τ₁^b[ns]

k_f(LE)^cor k⁰_f(ICT)^c[10⁷s^ꢀ1

]

23DCMIA

n-hexane

DEE

0.13

4.08

3.2 (LE)

2.3 (LE)

1.7 (LE)

2.7 (LE)

0.25

10.65

9.77

2.27

MeCN

n-hexane

DEE

0.17

24DCMIA

25DCMIA

0.061

0.051

MeCN

0.0025 (LE)

0.022 (ICT)

0.27

5.46^d,e(ICT)

0.43^f(ICT)

2.8 (LE)

n-hexane

DEE

9.77

15.8

7.9

0.29

1.8 (LE)

THF

EtCN

MeCN

0.028

0.0033 (LE)

0.025 (ICT)

0.032

2.70 (ICT)

1.12^d

0.93 (ICT)

2.9 (LE)

26DCMIA

34DCMIA

n-hexane

DEE

1.90

3.74^d

3.08^d

MeCN

n-hexane

DEE

0.061

1.6 (LE)

5.2 (LE)

4.3 (LE)

0.16

0.29

6.75

MeCN

0.0054 (LE)

0.0077 (ICT)

0.082

1.36^d,g(ICT)

3.50

0.57^h(ICT)

2.3 LE)

35DCMIA

n-hexane

DEE

0.16

MeCN

0.20

15.9

1.3 (LE)

^aLE ﬂuorescence quantum yield Φ(LE), unless otherwise indicated. Data for n-hexane are from Table 4 and those for MeCN from Table 5. ^bFrom

nanosecond ﬂuorescence decays, with a time resolution better than 0.5 ns (ref 54). For the picosecond decays of 24DCMIA and 34DCMIA in MeCN,

see footnote d. ^cRadiative rate constant: k_f= Φ(ﬂu)/τ₁. In the case of the dual ﬂuorescent molecules 24DCMIA, 25DCMIA, and 35DCMIA in MeCN,

only approximate values are listed for k_f(LE) and k⁰_f(ICT), as the expressions for these rate constants are more complex (ref 34). See footnote e.

e

f

^dTime resolution ≈3 ps (ref 7). From Figure 9a. For 24DCMIA in MeCN, the ICT lifetime τ⁰_o(ICT) = 5.31 ns (Table 13), hence k⁰_f(ICT) =

Φ⁰(ICT)/τ⁰_o(ICT) = 0.41 ꢁ 10⁷s^ꢀ1. ^gFrom Figure 10a. ^hFor 34DCMIA in MeCN, the ICT lifetime τ⁰_o(ICT) = 1.17 ns (Table 13), and k⁰_f(ICT) =

Φ⁰(ICT)/τ⁰_o(ICT) = 0.66 ꢁ 10⁷s^ꢀ1

.

ICT response functions, with a dominant fast decay (τ₂) for LE

and a corresponding fast rise for ICT. Some of the decay times τ₂

(Figures 8, 11a, 11b; Tables 13 and S2 (SI)) are shorter than the

time resolution⁷of around 3 ps. Nevertheless, even these decay

curves reveal that an ICT reaction takes place, with a reaction

time τ₂below 3 ps. Femtosecond transient absorption experi-

ments with 24DCDMA and 34DCDMA in MeCN, presented in

the next section, will allow an accurate view of the temporal

evolution of these ICT reactions.

Femtosecond Transient Absorption Spectra in n-Hexane

and MeCN. Femtosecond transient absorption spectra (at

22 ꢀC) of 24DCDMA in n-hexane and MeCN and of 34DCDMA

in MeCN were measured to obtain kinetic information on their

ICT reactions.

24DCDMA in n-Hexane. The transient absorption spectra of

24DCDMA in n-hexane are shown in Figure 13a. After correction

for bleaching (BL) and stimulated emission (SE), the excited state

absorption (ESA) spectra are obtained (Figure 13b), with maxima

at 315, 440, 580, and 670 nm (Table 15). A decay is not observed

for the band integral BI(300,650) of the ESA spectra between 300

and 650 nm over the range of pumpꢀprobe delay times between

0.2 and 100 ps. This means that an LE f ICT does not occur.

Therefore, the ESA spectrum is attributed to the LE state. The LE

spectrum of 24DCDMA is similar to that of DMABN in n-hexane,

with maxima at 300, 320, 445, and 745 nm (Table 15).⁷The ESA

maximum of 24DCDMA at 780 nm also appears in the spectrum

of TCDMA (Table 15).¹⁵

24DCDMA in MeCN (0.1ꢀ9.0 ps). The transient absorption

spectra of 24DCDMA in MeCN for pumpꢀprobe delay times

between 0.1 and 9.0 ps are presented in Figure 14a. After

correction for bleaching (BL) and stimulated LE and ICT emis-

sion (SE(LE) and SE(ICT)), the excited state absorption (ESA)

spectra are obtained (Figure 14b). The ESA spectrum consists of

three bands with maxima at 320, 438, and 630 nm (Table 15).

Note the appearance of an isosbestic point at around 460 nm: an

indication of the presence of two kinetically interacting excited

states, LE and ICT. In the ESA spectra, a decay is observed for the

absorption maximum at 630 nm, which is hence attributed to LE.

A comparable growing-in occurs for the bands with maxima at 320

and 438 nm, which are therefore assigned to the ICT state. A

global analysis of the band integrals BI(310,350), BI(410,460),

and BI(600,680) in Figure 14c results in a decay time τ₂of 1.77 ps,

with a growing-in (negative A₂, eq 10) for BI(410,460), in the

spectral range of the ICT absorption.

BI ¼ A₂expð ꢀ t=τ₂Þ þ A₀

ð10Þ

24DCDMA in MeCN (0.2ꢀ100 ps). The transient absorption

spectra of 24DCDMA in MeCN for pumpꢀprobe delay times

10839

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Figure 8. Picosecond ﬂuorescence decays of 24DCDMA in MeCN at

25 ꢀC, (a) at 400 nm and (b) at 550 nm, with a time resolution of 1.025 ps/

channel and a time window of 1500 eﬀective channels. The 400 nm

decay is attributed to LE and that at 550 nm to ICT (see Figure 3b). The

LE and ICT decays are analyzed simultaneously (global analysis),

resulting in decay times τ₂and τ₁with amplitudes A_1i(LE) and A_2i(ICT)

(eqs 7 and 8). The shortest time is listed ﬁrst. The time τ₂= 1.8 ps comes

from femtosecond transient spectra (Figures 14 and S7 (SI), below) and

is kept ﬁxed in the analysis. Excitation wavelength: 298 nm. The time in

parentheses is attributed to photoproducts, intrinsic impurities, or

experimental problems (see the text). The weighted deviations sigma,

the autocorrelation functions AꢀC, and the values for χ²are also indicated.

Figure 9. Picosecond ﬂuorescence decays of 24DCMIA in MeCN at

25 ꢀC, with times τ_iand amplitude A_1i(LE) (eq 7): (a) at 390 nm and

(b) at 550 nm, both with a time resolution of 0.497 ps/channel and a

time window of 1400 eﬀective channels. The 390 nm decay is attributed

to LE and that at 550 nm to ICT (see Figure 5b). In (c), the ICT decay

with τ₁= 5120 ps is presented at 10.04 ps/channel. Excitation

wavelength: 272 nm. See the caption of Figure 8.

between 0.2 and 100 ps are shown in Figure S7a in the SI. The

ESA spectra, obtained after correction for bleaching (BL) and LE

and ICT stimulated emission (SE(LE) and SE(ICT)), are

presented in Figure S7b (SI), again with an isosbestic point at

around 460 nm. The bands with maxima at 320 and 438 nm show

a growing-in, whereas that at 630 nm presents a decay. The first

two bands hence belong to the LE ESA spectrum, and the third is

part of the ICT ESA spectrum, similar to what was concluded

above for 24DCDMA over the time range from 0.1 to 9.0 ps

(Figure 14). A global analysis of BI(310,350), BI(410,460), and

BI(600,680) in Figure S7c (SI) results in a decay time τ₂of 1.83

ps, identical to the 1.77 ps from Figure 14c (Table 15).

34DCDMA in MeCN. The transient absorption spectra of

34DCDMA in MeCN for pumpꢀprobe delay times between

0.08 and 9.0 ps are shown in Figure 15a. After correction for

bleaching (BL) and LE and ICT stimulated emission (SE(LE) and

SE(ICT)), the excited state absorption (ESA) spectraare obtained

(Figure 15b). The ESA spectra have a main maximum at 330 nm

and two smaller maxima at 570 and 635 nm. From a global analysis

of BI(315,350), BI(380,430), and BI(610,680) in Figure 15c,

a decay time τ₂= 3.1 ps is determined. The BIs around the

maxima at 330 and 635 nm show a decay with a positive

amplitude A₂(eq 10): they belong to the LE state. In the case

of BI(380,430) around 385 nm, a growing-in (negative

amplitude A₂) is observed. This BI therefore covers the

spectral range of the ICT ESA spectrum. Similar ESA spectra

are measured with 34DCDMA in MeCN for pumpꢀprobe

delay times between 0.2 and 100 ps (Figure S8 in SI). From a

10840

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Figure 10. Picosecond LE and ICT ﬂuorescence decays of 34DCMIA

in MeCN at (a) 25 ꢀC and (b) ꢀ45 ꢀC. The decays were measured at

420 and 580 nm, with a time resolution of 0.497 ps/channel and a time

window of 1400 eﬀective channels. The 420 nm decay is attributed to LE

and that at 580 nm to ICT (see Figure 5e). The LE and ICT decays are

analyzed simultaneously (global analysis), resulting in decay times τ₂

and τ₁with amplitudes A_1i(LE) and A_2i(ICT) (eqs 7 and 8). Excitation

wavelength: 272 nm. See the caption of Figure 8.

global analysis of three BIs over the same spectral ranges as in

Figure 15c, a somewhat longer decay time τ₂= 4.0 ps is found,

likewise with a growing-in for BI(380,430).

Kinetic Analysis of the ESA Decay Times τ₂of 24DCDMA

and 34DCDMA in MeCN at 22 ꢀC. From the ESA spectrum of

34DCDMA in MeCN, it is seen that a substantial LE absorption

has remained alongside the ICT band at a pumpꢀprobe delay of

100 ps (Figure S8c in SI). This is also the case, although to a

somewhat smaller extent, for the ESA spectrum of 24DCDMA in

MeCN at this long delay time (Figure S7c in SI). The observations

show that an LE a ICT equilibrium is established. The ICT

reaction times τ₂of 1.8 ps (24DCDMA) and 3.1 ps (34DCDMA)

are therefore in fact equal to the reciprocal of the sum of the forward

and backward ICT rate constants 1/(k_a+ k_d), see Scheme 1.

A similar equilibrium situation has been encountered with

N-(4-cyanophenyl)carbazole (NP4CN).⁵⁰For 24DCDMA in

MeCN at 25 ꢀC (Table 13), τ⁰_o(ICT) = 3.44 ns, hence

Figure 11. Picosecond LE and ICT ﬂuorescence decays of

25DCMIA in (a) MeCN at ꢀ44 ꢀC, (b) PrCN at ꢀ40 ꢀC, and (c)

THF at ꢀ60 ꢀC. The decays were measured at 420 nm (a,b,c), 590 nm

(a,b), and 600 nm (c), with a time resolution of 1.978 ps/channel and

a time window of 1400 eﬀective channels. The 420 nm decays are

attributed to LE and those at 590 and 600 nm to ICT (see Figure 5e).

The LE and ICT decays are analyzed simultaneously (global ana-

lysis), resulting in times τ₂and τ₁with amplitudes A_1i(LE) and

A_2i(ICT) (eqs 7 and 8). Excitation wavelength: 276 nm. See the

caption of Figure 8.

1/τ⁰_o(ICT) = 0.029 ꢁ 10¹⁰s^ꢀ1. As k_d= 6.5 ꢁ 10¹⁰

s

^ꢀ1, k_d.

Separate Rate Constants k_aand k_d. The separate forward

and backward ICT rate constants k_aand k_d(Scheme 1) of

24DCDMA and 34DCDMA in MeCN can be determined by

employing the data for the quantum yield ratios Φ⁰(ICT)/Φ(LE),

τ⁰_o(ICT). Within this so-called high-temperature limit (HTL),⁷

the general expression for Φ⁰(ICT)/Φ(LE) (eq 1) is trans-

formed to eq 2.

10841

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

Table 13. Fluorescence Decay Times and Kinetic Parameters (Equations 7ꢀ9, Scheme 1) for the ICT Reaction of 24DCDMA,

24DCMIA, 34DCMIA, and 25DCMIA

mnDCXYA

solvent

T [ꢀC]

τ₂[ps]

τ₁[ps]

A₁₂/A₁₁(eq 7)

k_a[10¹⁰s^ꢀ1

]

k_d[10¹⁰s^ꢀ1

]

τ⁰_o(ICT) [ns]

24DCDMA

34DCDMA

24DCMIA

MeCN^a

DEE

MeCN^e

25

1.8^b

18

3800^c

9250

5460

5250

5860

5470

1360

890

7.6

19.9

24.4

5.9

49.1

5.28

32.0

25.2

20.0

4.5

6.5

3.44

9.08

5.31

4.73

5.72

5.42

1.17

0.85

2.32

ꢀ101^d

25

0.26

1.3

3.0

DEE

25

3.4

4.3

DEE

ꢀ45

ꢀ115

25

4.8

24.9

75.2

5.5

0.80

0.060

3.3

DEE

21.8

4.6

34DCMIA

25DCMIA

MeCN^f

MeCN^g

MeCN^h

18.4

13.5

70.8

ꢀ45

ꢀ44

7.0

17.6

0.76

1.4

2340

114

0.62

^aFigure 8. ^bFrom femtosecond ESA spectra (Figures 14 and S7 (SI)). ^cSee Table 11. ^dICT reaction possible because of the increase in solvent polarity

upon cooling: ε^ꢀ110= 9.79, ε²⁵= 4.24. ^eFigure 9a. ^fFigure 10a. ^gFigure 10b. ^hFigure 11a.

Figure 12. 25DCMIA in THF. (a) Fluorescence decay times (τ₂, τ₁)

and the LE amplitude ratio A = A₁₂/A₁₁(eqs 7ꢀ9) as a function of

temperature and (b) Arrhenius plots of the ICT rate constants (k_a, k_d)

and τ⁰₀(ICT), see Scheme 1. ΔH = (E_aꢀ E_d) = ꢀ7.4 kJ/mol, k_a⁰

mol^ꢀ1(Table 14).

=

1.6 ꢁ 10¹²

s

^ꢀ1, k_d⁰= 1.2 ꢁ 10¹³

s

^ꢀ1, and ΔS = R ln(k_a/k_d) = ꢀ17 J K^ꢀ1

0

Table 14. Thermodynamic Parameters of the ICT Reaction

of 25DCMIA in THF

0

E_a

E_d

ꢀΔH

k_a⁰

k_d

ꢀΔS

[kJ/mol] [kJ/mol] [kJ/mol] [10¹²s^ꢀ1

]

[10¹²s^ꢀ1

]

[JK^ꢀ1mol^ꢀ1

]

4.5

11.9

7.4

1.55

11.7

16.8

Figure 13. 24DCDMA in n-hexane at 361 nm excitation (spectral

range: 295ꢀ683). (a) Transient absorption spectra and (b) excited

state absorption (ESA) spectra, at pumpꢀprobe delay times be-

tween 0.2 and 100 ps, after correction for bleaching (BL) and

stimulated emission (SE). The BL and SE (LE, cf Figure 2b) spectra

are also depicted. (c) For the band integral BI(300,650), between

300 and 650 nm in the ESA spectrum, a decay is not found between

0.2 and 100 ps: an ICT reaction does not take place. mΔOD is the

optical density/1000.

under HTL conditions (eq 2). In this procedure, k⁰_f(ICT)/k_f(LE) is

needed. For DMABN in MeCN at 25 ꢀC, k⁰_f(ICT)/k_f(LE) = 0.12,⁷

whereas for DIABN in n-hexane at 25 ꢀC,⁵¹k⁰_f(ICT)/k_f(LE) = 0.17.

Also for PP in MeCN at 25 ꢀC, k⁰_f(ICT)/k_f(LE) is much smaller

(0.19) than unity.³⁴

The mean value of k⁰_f(ICT)/k_f(LE) = 0.15 from the data

for DMABN and DIABN is adopted here for 24DCDMA

and 34DCDMA. For 24DCDMA, with Φ⁰(ICT)/Φ(LE) =

k⁰_f(ICT)/k_f(LE), k_a/k_d= 1.2 (Table 3), then k_a/k_d= 8.0.

Together with (k_a+ k_d) = 1/τ₂= 1/(1.8 ps) = 56 ꢁ 10¹⁰s^ꢀ1

(Figures 14c and S7c (SI)), one obtains: k_a= 49 ꢁ 10¹⁰s^ꢀ1

the same as those (49 and 6.5 ꢁ 10¹⁰s^ꢀ1) derived from the

LE and ICT ﬂuorescence decay parameters τ₂, τ₁, and

A₁₂/A₁₁(Table 13). Similarly, for 34DCDMA with

and k_d= 6.2 ꢁ 10¹⁰

s

^ꢀ1(Table 15). These data are practically

10842

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

Table 15. Excited State Absorption (ESA) Maxima (in nm) and ESA Decay Times (in ps) of 24DCDMA and 34DCDMA at 22 ꢀC^a

ARTICLE

solvent

ESA maxima (nm)

τ₂[ps]

(k_a+ k_d)^b[10¹⁰s^ꢀ1

]

k_a[10¹⁰s^ꢀ1

]

k_d[10¹⁰s^ꢀ1

]

24DCDMA

n-hexane^c

MeCN^d

MeCN^e

n-hexane

MeCN

315, 440, 670

-

320, 438, 630

1.8

3.1

55.6

32.3

49.4

22.5

6.2

9.8

34DCDMA

DMABN ^f

330, 385, 635

300, 320, 445, 745 (LE)

320, 355, 440, 710 (LE) 315, 425 (ICT)

310, 335, 510, 580 (LE)

290, 380, 520 (LE)

4.07

24.1

24.0

0.047

TCDMA^g

PCDMA^h

MeCN

^aSee Figures 13ꢀ15 and S7 and S8 (SI). Data for TCDMA and DMABN, TCDMA, and PCDMA are included for comparison. ^b(k_a+ k_d) = 1/τ₂

(see text). ^cFigure 13. ^dFigures 14 and S7 (SI). ^eFigures 15 and S8 (SI). ^fRef 7. ^gRef 15. ^hRef 10.

Figure 14. 24DCDMA in MeCN at 361 nm excitation (305ꢀ683 nm).

Figure 15. 34DCDMA in MeCN at 361 nm excitation (305ꢀ683 nm).

(a) Transient absorption spectra and (b) excited state absorption (ESA)

spectra, at pumpꢀprobe delay times between 0.1 and 9.0 ps. A decay is

spectra, at pumpꢀprobe delay times between 0.08 and 9.0 ps. A decay is

observed for the absorption maximum at 630 nm, attributed to LE. A

observed for the absorption maxima at 330 and 630 nm, attributed to LE.

growing-in occurs for the maxima at 320 and 438 nm, which are assigned

A growing-in occurs around 385 nm, which is assigned to the ICT state.

(c) A global analysis of the band integrals BI(315,350), BI(380,430), and

BI(610,680) results in a decay time τ₂of 3.1 ps, with a growing-in

(negative A₂, eq 10) for BI(380,430), in the spectral range of the ICT

to the ICT state. (c) A global analysis of the band integrals BI(310,350),

BI(410,460), and BI(600,680) results in a decay time τ₂of 1.77 ps, with a

growing-in (negative A₂, eq 10) for BI(410,460), in the spectral range of

the ICT absorption. See the caption of Figure 13.

absorption. See the caption of Figure 13.

Φ⁰(ICT)/Φ(LE) = 0.35 (Table 3), k_a/k_dbecomes equal to 2.3.

from Φ⁰(ICT)/Φ(LE): 0.35 for 34DCDMA and 1.2 for

24DCDMA (Table 3).

From k_a+ k_d= 1/τ₂= 1/(3.1 ps) = 32 ꢁ 10¹⁰

s

^ꢀ1(Figure 15c and

S8c (SI)), one so ﬁnds k_a= 23 ꢁ 10¹⁰

s

^ꢀ1and k_d= 9.5 ꢁ 10¹⁰s^ꢀ1

(Table 15).

’ CONCLUSIONS

The lower k_aand larger k_dvalues of 34DCDMA as compared

with 24DCDMA point to a smaller ICT reaction enthalpy

diﬀerence ꢀΔH for 34DCDMA, which causes the LE a ICT

equilibrium to be shifted further toward LE. This also follows

The ﬂuorescence spectra of the six isomers of mnDCDMA

and mnDCMIA in n-hexane consist of a single LE ﬂuorescence

band. This means that an LE f ICT reaction does not take place

10843

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

in this solvent, although the dicyanobenzene subgroups in the

dicyanoanilines are considerably better electron acceptors than

the benzonitrile moiety in DMABN, for which ICT just starts

to occur in n-hexane. The larger electron aﬃnity can be deduced

from the reduction potentials E(A^ꢀ/A) vs SCE: ꢀ1.68 V

(1,2-dicyanobenzene), ꢀ1.95 V (1,3-dicyanobenzene), and

ꢀ1.63 V (1,4-dicyanobenzene) as compared with ꢀ2.36 V for

benzonitrile. In MeCN, four of the mnDCDMAs (mn = 23, 25,

26, 35) and three of the mnDCMIAs (mn = 23, 26, 35) likewise

only show an LE emission, whereas dual LE + ICT emission is

observed with 24DCDMA and 34DCDMA, as well as with

24DCMIA and 34DCMIA and to a smaller extent also with

25DCMIA. The occurrence of an LE f ICT reaction with the

2,4- and 3,4-dicyanoanilines is connected with the magnitude

of the energy gap ΔE(S₁,S₂), which is much smaller for these

molecules than for the other isomers: 4870 and 4690 cm^ꢀ1for

24DCDMA and 24DCMIA against 8650 cm^ꢀ1for 23DCMIA in

MeCN, as an example. By comparison, for DMABN, the proto-

type of a D/A molecule undergoes an ICT reaction, ΔE(S₁,S₂) =

2700 cm^ꢀ1, in n-hexane. The observation that ICT is more

eﬃcient (larger Φ⁰(ICT)/Φ(LE)) for 24DCMIA and 34DCMIA

than with 24DCDMA and 34DCDMA is due to the stronger

electron donor character (smaller (E(D/D⁺) vs SCE) of

iPrN(Me)₂(0.84 V) than that of N(Me)₃(1.05 V). With

25DCMIA in MeCN, ICT ﬂuorescence is observed, notwith-

standing its relatively large ΔE(S₁,S₂) of 9240 cm^ꢀ1. This is caused

by the fact that 25DCMIA has a strongly twisted amino group

(θ = 42.6ꢀ), which leads to electronic decoupling of the D and A

subunits, making the molecule react as an intermolecular exciplex.

Under such conditions, ΔE(S₁,S₂) is no longer a criterion for the

appearance of an LE f ICT reaction. For the other dicyanoanilines

showing dual (LE + ICT) ﬂuorescence, the angle θis much smaller,

13.9ꢀ (24DCDMA), 16.7ꢀ (24DCMIA), and 2.9ꢀ (34DCMIA),

and ΔE(S₁,S₂) remains an essential factor for the appearance of

dual ﬂuorescence in these strongly coupled D/A systems.

1.8 ps for 24DCDMA and 3.1 ps for 34DCDMA. The reciprocal of

this time τ₂is equal to (k_a+ k_d). By combining these results with

the data for the ﬂuorescence decays and Φ⁰(ICT)/Φ(LE), the

values k_a= 49 ꢁ 10¹⁰

s

^ꢀ1(24DCDMA) and k_a= 22.6 ꢁ 10¹⁰s^ꢀ1

(34DCDMA) are determined.

’ ASSOCIATED CONTENT

S

Supporting Information. Crystal structure; ﬂuorescence

b

and absorption spectra; oxidation potentials; solvatochromic plots;

ESA spectra. This material is available free of charge via the Internet

at http://pubs.acs.org.

’ AUTHOR INFORMATION

Corresponding Authors

*E-mail: galievsky@imaph.bas-net.by (V.A.G.); druzhinin@chemie.

uni-siegen.de (S.I.D.); skovale@chemie.hu-berlin.de (S.O.K.);

kzachar@gwdg.de (K.A.Z.).

’ ACKNOWLEDGMENT

ManythanksareduetoProf.N.P.Ernsting(HumboldtUniversity

Berlin) for the use of the femtosecond equipment in the investiga-

tions reported here. We are grateful to Mr. W. Bosch for the synthesis

of the dicyanoanilines. We also thank Mr. J. Bienert for carrying

out HPLC puriﬁcations and Mr. H. Lesche for technical support.

We further express our gratitude to Mr. C.-P. Adam for making the

ﬁgures.

’ REFERENCES

(1) Deﬁnition of LE and ICT. The term LE (locally excited) may

lead to confusion in the discussion of ICT reactions. It originated from

intermolecular exciplex formation, where the excitation in the ¹A* + D f

¹(A^ꢀD⁺) reaction is initially localized on the monomer ¹A* and becomes

more delocalized over the molecules A and D upon exciplex formation.

When a donor (D)/acceptor (A) compound only emits a single ﬂuorescence

band, we call this an emission from an LE state, irrespective of the dipole

moment μ_e(LE) of this state. An example is DMABN in n-hexane, with

μ_e(LE) = 10 D as compared with μ_g(S₀) = 6.6 D (refs 7 and 21). In such a

case, we say that an ICT reaction did not take place upon excitation, although a

substantial increase in charge separation in fact occurs: Δμ ≈ 3.4 D. We do

speak of an ICT reaction only when a new excited ICT state is produced from

the relaxed initially excited S₁(LE) state as the precursor. For this ICT state,

the condition μ_e(ICT) > μ_e(LE) generally holds, such as for DMABN

in polar solvents with dipole moments of ≈17 D (ICT) and ≈9 D (LE)

(ref 52). In the present article, the increase in dipole moment during the

LE f ICT reaction is rather small, around 2 D for 24DCDMA and

34DCDMA (Table 8) and 3 D for 24DCMIA, 25DCMIA, and 34DCMIA

(Table 10). For these molecules, an ICT reaction nevertheless occurs: two

excited states, LE and ICT, are observed.

(2) (a) Lippert, E.; L€uder, W.; Boos, H. In Advances in Molecular

Spectroscopy; European Conference on Molecular Spectroscopy, Bologna,

Italy, 1959; Mangini, A., Ed.; Pergamon Press: Oxford, U.K., 1962; p 443.

(b) Lippert, E.; L€uder, W.; Moll, F.; N€agele, W.; Boos, H.; Prigge, H.;

Seibold-Blankenstein, I. Angew. Chem. 1961, 73, 695.

(3) Rettig, W. Angew. Chem., Int. Ed. Engl. 1986, 25, 971.

(4) Leinhos, U.; K€uhnle, W.; Zachariasse, K. A. J. Phys. Chem. 1991,

95, 2013.

(5) Ma, C.; Kwok, W. M.; Matousek, P.; Parker, A. W.; Phillips, D.;

Toner, W. T.; Towrie, M. J. Phys. Chem. A 2002, 106, 3294.

(6) Grabowski, Z. R.; Rotkiewicz, K.; Rettig, W. Chem. Rev. 2003,

103, 3899.

(7) Druzhinin, S. I.; Ernsting, N. P.; Kovalenko, S. A.; Pꢀerez Lustres,

L.; Senyushkina, T.; Zachariasse, K. A. J. Phys. Chem. A 2006, 110, 2955.

The dipole moments μ_e(ICT) are 13 D for 24DCDMA and

16 D for 34DCDMA, whereas for 24DCMIA and 34DCMIA

slightly larger values are determined: 15 and 18 D. The corre-

sponding μ_e(LE)'s are somewhat smaller, with 11 and 15 D for 24-

and 34DCDMA and 12 and 16 D for 24- and 34DCMIA. With

25DCMIA, smaller dipole moments are measured: μ_e(ICT) = 11 D

and μ_e(LE) = 9 D. The diﬀerences between μ_e(ICT) and μ_e(LE) of

the dicyanoanilines are smaller than for DMABN, with μ_e(ICT) =

17 D and μ_e(LE) = 10 D. This is probably caused by the diﬀerent

dipole moment directions in the ICT and LE states of the former

molecules. The change in enthalpy diﬀerence ΔH(SB) of the ICT

reaction in MeCN, obtained from plots of ln(Φ⁰(ICT)/Φ(LE)) vs

1000/T, is relatively small: ꢀ5.4 kJ/mol (24DCDMA), ꢀ4.7 kJ/mol

(24DCMIA), and ꢀ3.9 kJ/mol (34DCMIA).

Picosecond LE and ICT ﬂuorescence decays are determined

for 24DCDMA, 24DCMIA, 34DCMIA, and 25DCMIA in a

number of solvents at various temperatures. From these decays,

the forward and backward ICT rate constants k_aand k_dand also

the ICT lifetime τ⁰_o(ICT) are calculated. From the temperature

dependence of k_aand k_dfor 25DCMIA in THF, the activation

energies E_a= 4.5 kJ/mol and E_d= 11.9 kJ/mol are obtained,

giving ΔH = ꢀ7.4 kJ/mol.

The femtosecond ESA spectra of 24DCDMA and 34DCDMA

in MeCN at 22 ꢀC consist of a superposition of LE and ICT

absorptions, showing that an LE a ICT equilibrium is estab-

lished in around 4 ps. The ICT reaction time, derived from the

decay timeτ₂of the LE and the rise of the ICT absorption bands, is

10844

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

The Journal of Physical Chemistry A

ARTICLE

(8) Grabowski, Z. R.; Rotkiewicz, K.; Siemiarczuk, A.; Cowley, D. J.

Nouv. J. Chim. 1979, 3, 443.

(9) Zachariasse, K. A.; von der Haar, Th.; Hebecker, A.; Leinhos, U.;

K€uhnle, W. Pure Appl. Chem. 1993, 65, 1745.

(37) Druzhinin, S. I.; Kovalenko, S. A.; Senyushkina, T. A.; Demeter,

A.; Machinek, R.; Noltemeyer, M.; Zachariasse, K. A. J. Phys. Chem. A

2008, 112, 8238.

(38) Zachariasse, K. A.; von der Haar, Th.; Leinhos, U.; K€uhnle, W.

(10) Zachariasse, K. A.; Druzhinin, S. I.; Galievsky, V. A.; Demeter, A.;

Allonas, X.; Kovalenko, S. A.; Senyushkina, T. A. J. Phys. Chem. A 2010,

114, 13031.

(11) von der Haar, Th.; Hebecker, A.; Il’ichev, Yu. V.; Jiang, Y.-B.;

K€uhnle, W.; Zachariasse, K. A. Recl. Trav. Chim. Pays-Bas 1995, 114, 430.

(12) von der Haar, Th.; Hebecker, A.; Il’ichev, Yu.; K€uhnle, W.;

Zachariasse, K. A. In Fast Elementary Processes in Chemical and Biological

Systems, Lille, France, 1995, AIP Conference Proceedings 1996; Vol.

364, p 295.

(13) Zachariasse, K. A.; Grobys, M.; von der Haar, Th.; Hebecker,

A.; Il’ichev, Yu. V.; Jiang, Y.-B.; Morawski, O.; K€uhnle, W. J. Photochem.

Photobiol. A: Chem. 1996, 102, 59. Erratum: J. Photochem. Photobiol. A:

Chem. 1998, 115, 259.

J. Inf. Rec. Mater. 1994, 21, 501.

(39) Zachariasse, K. A.; Grobys, M.; von der Haar, Th.; Hebecker,

A.; Il’ichev, Yu. V.; Morawski, O.; R€uckert, I.; K€uhnle, W. J. Photochem.

Photobiol. A: Chem. 1997, 105, 373.

(40) Zachariasse, K. A.; Druzhinin, S. I.; Bosch, W.; Machinek, R.

J. Am. Chem. Soc. 2004, 126, 1705.

(41) Nelsen, S. F. J. Org. Chem. 1984, 49, 1891.

(42) (a) Solling, T. I.; K€otting, C.; Zewail, A. H. J. Phys. Chem. A

2003, 107, 10872.(b) Kimura, K.; Katsumata, S.; Achiba, Y.; Yamazaki,

T.; Iwata, S. Handbook of He(I) Photoelectron Spectra of Fundamental

Organic Molecules; Japan Scientiﬁc Societies Press: Tokyo, 1981.

(43) Il’ichev, Yu. V.; K€uhnle, W.; Zachariasse, K. A. J. Phys. Chem. A

1998, 102, 5670.

(14) Zachariasse, K. A. Chem. Phys. Lett. 2000, 320, 8.

(15) Zachariasse, K. A.; Druzhinin, S. I.; Galievsky, V. A.; Kovalenko,

S. A.; Senyushkina, T. A.; Mayer, P.; Noltemeyer, M.; Boggio-Pasqua,

M.; Robb, M. A. J. Phys. Chem. A 2009, 113, 2693.

(16) Gedeck, P.; Schneider, S. J. Photochem. Photobiol. A: Chem.

1999, 121, 7.

(44) Druzhinin, S. I.; Kovalenko, S. A.; Senyushkina, T. A.; Demeter,

A.; Januskevicius, R.; Mayer, P.; Stalke, D.; Machinek, R.; Zachariasse,

K. A. J. Phys. Chem. A 2009, 113, 9304.

(45) Baumann, W.; Bischof, H.; Brittinger, J.-C.; Rettig, W.; Rotkie-

wicz, K. J. Photochem. Photobiol. A: Chem. 1992, 64, 49.

(46) Yoshihara, T.; Galievsky, V. A.; Druzhinin, S. I.; Saha, S.;

Zachariasse, K. A. Photochem. Photobiol. Sci. 2003, 2, 342.

(47) Il’ichev, Yu. V.; K€uhnle, W.; Zachariasse, K. A. Chem. Phys. 1996,

211, 441.

(48) Suppan, P.; Ghoneim, N. Solvatochromism; The Royal Society

of Chemistry: Cambridge, U.K., 1997.

(49) Liptay, W. In Excited States; Lim, E. C., Ed.; Academic Press:

New York, USA, 1974; Vol. 1, p 129.

(17) Parusel, A. B. J.; Schamschule, R.; Kohler, G. Z. Phys. Chem. 2002,

216, 361.

(18) Jamorski, C. J.; L€uthi, H.-P. Chem. Phys. Lett. 2003, 368, 561.

(19) The following reduction potentials E(A^ꢀ/A) (vs SCE) are

taken from ref 10 and publications cited therein: ꢀ2.36 V (benzonitrile),

ꢀ1.68 V (1,2-dicyanobenzene), ꢀ1.95 V (1,3-dicyanobenzene), ꢀ1.63

V (1,4-dicyanobenzene), ꢀ1.36 V (1,3,5-tricyanobenzene), and ꢀ0.29

V (pentacyanobenzene).

(50) Galievsky, V. A.; Druzhinin, S. I.; Demeter, A.; Mayer, P.;

Kovalenko, S. A.; Senyushkina, T. A.; Zachariasse, K. A. J. Phys. Chem. A

2010, 114, 12622.

(20) Gꢀomez, I.; Reguero, M.; Boggio-Pasqua, M.; Robb, M. A. J. Am.

Chem. Soc. 2005, 127, 7119.

(21) Druzhinin, S. I.; Mayer, P.; Stalke, D.; von B€ulow, R.; Noltemeyer,

M.; Zachariasse, K. A. J. Am. Chem. Soc. 2010, 132, 7730.

(22) (a) Griﬃths, J.; Roozpeikar, B. J. Chem. Soc., Perkin Trans. 1

1976, 42. (b) Griﬃths, J.; Lockwood, M.; Roozpeikar, B. J. Chem. Soc.,

Perkin Trans. 2 1977, 1608.

(51) Demeter, A.; Druzhinin, S.; George, M.; Haselbach, E.; Roulin,

J.-L.; Zachariasse, K. A. Chem. Phys. Lett. 2000, 323, 351.

(52) Schuddeboom, W.; Jonker, S. A.; Warman, J. M.; Leinhos, U.;

K€uhnle, W.; Zachariasse, K. A. J. Phys. Chem. 1992, 96, 10809.

(53) Herbich, J.; Rotkiewicz, K.; Waluk, J.; Andresen, B.; Thulstrup,

E. W. Chem. Phys. 1989, 138, 105.

(23) Adcock, A. D.; Dewar, M. J. S. J. Am. Chem. Soc. 1967,

89, 386.

(54) Zachariasse, K. A.; Yoshihara, T.; Druzhinin, S. I. J. Phys. Chem.

(24) Druzhinin, S. I.; Kovalenko, S. A.; Senyushkina, T.; Zachariasse,

K. A. J. Phys. Chem. A 2007, 111, 12878.

A 2002, 106, 6325.

(55) Yoshihara, T.; Druzhinin, S. I.; Zachariasse, K. A. J. Am. Chem.

Soc. 2004, 126, 8535.

(56) Zachariasse, K. A.; Druzhinin, S. I.; Kovalenko, S. A.; Senyushkina,

T. A. J. Phys. Chem. A 2010, 114, 1621.

(25) Demas, J. N.; Crosby, G. A. J. Phys. Chem. 1971, 75, 991.

(26) Galievsky, V. A.; Druzhinin, S. I.; Demeter, A.; Jiang, Y.-B.;

Kovalenko, S. A.; Pꢀerez Lustres, L.; Venugopal, K.; Ernsting, N. P.;

Allonas, X.; Noltemeyer, M.; Machinek, R.; Zachariasse, K. A. Chem.

Phys. Chem. 2005, 6, 2307.

(57) Druzhinin, S. I.; Galievsky, V. A.; Yoshihara, T.; Zachariasse,

K. A. J. Phys. Chem. A 2006, 110, 12760.

(27) Kovalenko, S. A.; Dobryakov, A. L.; Ruthmann, J.; Ernsting,

N. P. Phys. Rev. A 1999, 59, 2369.

(58) Mac-anita, A. L.; Zachariasse, K. A. J. Phys. Chem. A 2011, 115,

3183.

(28) Ernsting, N. P.; Kovalenko, S. A.; Senyushkina, T. A.; Saam, J.;

Farztdinov, V. J. Phys. Chem. A 2001, 105, 3443.

(29) Heine, A.; Herbst-Irmer, R.; Stalke, D.; K€uhnle, W.; Zachariasse,

K. A. Acta Crystallogr. 1994, B50, 363.

(30) Druzhinin, S. I.; Dubbaka, S. R.; Knochel, P.; Kovalenko, S. A.;

Mayer, P.; Senyushkina, T.; Zachariasse, K. A. J. Phys. Chem. A 2008,112, 2749.

(31) (a) Zachariasse, K. A.; Druzhinin, S. I.; Kovalenko, S. A.;

Senyushkina, T. A. J. Chem. Phys. 2009, 131, 224313. (b) Zachariasse,

K. A.; Druzhinin, S. I.; Kovalenko, S. A.; Senyushkina, T. A. Chem. Phys.

Lett. 2009, 484, 28.

(32) Oshima, J.;Yoshihara, T.;Tobita, S.Chem. Phys. Lett. 2006, 423, 306.

(33) Dahl, K.; Biswas, R.; Ito, N.; Maroncelli, M. J. Phys. Chem. B

2005, 109, 1563.

(59) Landolt-Bo€rnstein, Numerical Data and Functional Relationships

in Science and Technology, New Series; Madelung, O., Ed.; Group IV;

Springer: Berlin, 1991; Vol. 6.

(60) (a) Birks, J. B. Photophysics of Aromatic Molecules; Wiley:

London, 1970. (b) Hirayama, S.; Phillips, D. J. Photochem. 1980,

12, 139. (c) Morris, J. V.; Mahaney, M. A.; Huber, J. R. J. Phys. Chem.

1976, 80, 969. (d) Saltiel, J.; Waller, A. S.; Sears, D. F.; Garrett, C. Z.

J. Phys. Chem. 1993, 97, 2516.

(61) A detailed treatment of the dependence of the radiative rate

constant k_fon the solvent refractive index n has been presented in ref 62.

(62) Toptygin, D. J. Fluoresc. 2003, 13, 201.

(34) Yoshihara, T.; Druzhinin, S. I.; Demeter, A.; Kocher, N.; Stalke,

D.; Zachariasse, K. A. J. Phys. Chem. A 2005, 109, 1497.

(35) Murrell, J. N. The Theory of Electronic Spectra of Organic

Molecules; Methuen: London, 1963.

(36) Karpiuk, J.; Karolak, E.; Nowacki, J. Phys. Chem. Chem. Phys.

2010, 12, 8804.

10845

dx.doi.org/10.1021/jp2045614 |J. Phys. Chem. A 2011, 115, 10823–10845

Article Doi

Presence and absence of excited state intramolecular charge transfer with the six isomers of dicyano- N, N -dimethylaniline and dicyano-(N -methyl- N -isopropyl)aniline

DOI: 10.1021/jp2045614

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jp2045614

Full text of DOI:10.1021/jp2045614

Products guided by the article

R&D Labs maybe for 1334204-25-9

Relevant to this article

Hot Product