ARTICLE pubs.acs.org/JPCA
Fluorescence Solvatochromism in Lumichrome and Excited-State Tautomerization: A Combined Experimental and DFT Study N. Shaemningwar Moyon and Sivaprasad Mitra* Department of Chemistry, North-Eastern Hill University, Shillong 793 022, India
bS Supporting Information ABSTRACT: Fluorescence solvatochromism of lumichrome (LC) was studied by steady-state and time-resolved fluorescence spectroscopy. The excited-state properties of LC do not show any correlation with solvent polarity, however, reasonably good correlation with solvent ET(30) parameter was observed. A quantitative estimation of contribution from different solvatochromic parameters, like solvent polarizability (π*), hydrogen bond donor (R), and hydrogen bond acceptor (β) ability of the solvent, was made using linear free energy relationship on the basis of Kamlet-Taft equation. The analysis reveals that hydrogen bond donating ability (acidity) of the solvent is the most important parameter that characterizes the excited-state behavior of lumichrome. Quantum mechanical calculations using density functional theory (DFT) were done to study the most stable structure and excited-state tautomerization process of LC toward the formation of isoalloxazines. Charge localization in the excited state and formation of hydrogen-bonded cluster through solvent hydrogen bond donation on the N10 atom of alloxazine moiety were predicted to be the key step toward this water-catalyzed tautomerization process.
1. INTRODUCTION Lumichrome (7,8-dimethylalloxazine, LC, structure I in Scheme 1) represents a group of heterocyclic compounds related to lumazines and biologically important flavins. Although isoalloxazines (structure II in Scheme 1), especially flavins, are very closely related compounds to alloxazines, the spectral and photophysical properties of these two groups of compounds are distinctly different. Particularly, the fluorescence intensity and the excited-state lifetime for flavins are substantially higher than the corresponding alloxazines because of the presence of yellow chromophore characteristics of flavoproteins, enzymes occurring widely in animals and plants. The ground- and excitedstate properties of flavins and its representatives like riboflavin, flavin mononucleotide (FMN), and flavin adenine dinucleotide (FAD) are well characterized;1-5 however, alloxazines have received relatively little attention. The interest in LC and other substituted alloxazines has intensified recently because of their important role in different biological systems.6-10 For example, LC is known as a triplet photosensitizer to generate singlet oxygen (1O2), which initiates the oxidation of many biological substrates like enzymes, proteins, nucleic acids, hormones, and so forth.11,12 Alloxazine group of ligands are also found to be very effective in binding to several proteins and nucleic acids.13-16 Furthermore, LC is known to inhibit the flavin reductase in living Escherichia coli cells, the riboflavin uptake by human-derived liver cells Hep G2, colonic epithelial NCM460 cells, and also Caco-2 human intestinal epithelial cells.7,8 The application of LC has r 2011 American Chemical Society
been reported in photodegradation of polyamidehydroxyurethane polymers in aqueous solution,12 polymerization of 2-hydroxyehtyl methacrylate,17,18 as an optical transistor device,19 and so forth. Alloxazine nucleosides, for which the hydrogen-bonding characteristics resemble thymidine, can further be used as a fluorescence probe.20 It has already been reported that hydrogen bonding with acetic acid or pyridine derivatives promotes the tautomerization in alloxazine scaffold to produce isoalloxazine type of structure.21,22 In an earlier report, it was also suggested that solvent hydrogen bonding plays an important role on the photophysics of LC.23 More particularly, both the absorption and the emission maxima show red shift, emission quantum yield increases, and also the fluorescence state becomes more stable in polar protic environments when compared with the aprotic medium. However, to the best of our knowledge, so far, no systematic study of LC solvatochromism with solvent hydrogen bonding ability is available in the literature. The nature of the interaction of a solute molecule with amphiprotic solvents like water, alcohol, and so forth can have far-reaching consequences in terms of solubility as well as in chemical properties of the solute. In some cases, the solvents can function as mere supporting matrixes with their physical Received: October 27, 2010 Revised: February 7, 2011 Published: March 09, 2011 2456
dx.doi.org/10.1021/jp1102687 | J. Phys. Chem. A 2011, 115, 2456–2464
The Journal of Physical Chemistry A Scheme 1. Tautomerization of Lumichrome (I) and Isoalloxazine (II)
parameters like dielectric constant, viscosity, refractive index, polarizability, and so forth. However, in many cases, the interactions of heterocyclic organic compounds with amphiprotic solvents involve creation of new molecular complexes through hydrogen-bond formation. The chemical nature of the solute is mostly controlled by the properties of the hydrogen-bonded complex. Hydrogen bonding can occur in different modes depending on the structure of the solute and solvent. The situation becomes more complicated when a solute molecule, like LC, possesses multiple hydrogen bonding sites, and the solvent molecule, like water or acetic acid, can act both as a proton donor as well as a proton acceptor. Under this condition, the competition among different molecular species resulting from hydrogen-bonding interaction between the solute and the solvent molecules becomes inevitable. In a recent publication, we have discussed water-assisted hydrogen-bonding phenomenon in luminol with steady-state fluorescence spectroscopy as well as high-level density functional theory (DFT) calculation.24 LC also provides an equally important example to study the hydrogenbonding effect because the molecule itself can exist in more than one prototropic species having multiple hydrogen-bonding sites (chart 1). The most stable structure and associated spectroscopic properties will depend strongly on the relative abundance of several species as well as on their hydrogen-bonding mode with the solvent. Furthermore, the efficacy of hydrogen-bond formation in the excited state and consequently the tautomerization process may change because of charge redistribution after excitation. In this paper, we use steady-state and picosecond timeresolved fluorescence spectral properties in a series of pure solvents with varying polarity as well as hydrogen-bond donor and acceptor abilities to find quantitative information about their relative contribution on LC solvatochromism. Furthermore, solution-phase spectral properties of LC were theoretically modeled by using density functional approach to gather the information on the molecular origin of specific solvent effect and on the driving force toward the excited-state tautomerization process.
2. MATERIALS AND METHODS 2.1. Chemicals. Lumichrome (LC) was received from SigmaAldrich Chemical Pvt. Ltd. (product no. 103217) and was used without any further purification. The organic solvents used were of spectroscopic grade (>99.5%) as received from Alfa Aesar and, in some cases, from Aldrich Chemical Co. The analytical grade type II water, also used as solvent, was obtained from Elix 10 water purification system (Millipore India Pvt. Ltd.). The chromophore concentration (6-8 μM) was very low to avoid any aggregation and was kept constant during spectral measurements in different solvents. 2.2. Experimental Procedure. Steady-state absorption spectra were recorded on a Perkin-Elmer model Lambda25
ARTICLE
Chart 1. Possible Isomeric Structures of Lumichrome (I)
absorption spectrophotometer. Fluorescence spectra were taken in a Hitachi model FL4500 spectrofluorimeter, and all the spectra were corrected for the instrument response function. Quartz cuvettes of 10 mm optical path length received from PerkinElmer, United States, (part no. B0831009) and Hellma, Germany, (type 111-QS) were used for measuring absorption and fluorescence spectra, respectively. Fluorescence quantum yields (φf) were calculated by comparing the total fluorescence intensity under the whole fluorescence spectral range with that of a standard (quinine bisulfate in 0.5 M H2SO4 solution, φsf = 0.54625) with the following equation using adequate correction for solvent refractive index (n).26 !2 i -As ni i s F 1 - 10 φf ¼ φf 3 s 3 ð1Þ i F 1 - 10-A 3 ns where Ai and As are the optical density of the sample and standard, respectively, and ni is the refractive index of solvent at 293 K. The relative experimental error of the measured quantum yield was estimated within (10%. Fluorescence decay analysis was performed using time-correlated single-photon-counting (TCSPC) technique as implemented in time-resolved spectrofluorimeter FL-920 (Edinburg Instruments, United Kingdom), details of which have been described elsewhere.27 2.3. Theoretical Calculations. Density functional theory (DFT) has successfully been applied to study the hydrogen bonding in various systems, and the results are found to be in good agreement even with the results of several other computationally costlier methods like Møller-Plesset (MP2) or coupledcluster (CC) calculations.28-31 Myshakina et al. used B3LYP functional along with 6-311þþG(d,p) basis set to estimate the impact of hydrogen bonding on amide frequency for model peptides,32 whereas in a recent paper, the effectiveness of DFT method using a similar level of calculations to predict the dative bond energy of Lewis adducts was checked to be within ∼1 kcal/ mol of the benchmark results.33 Several recent reports are also available in literature dealing with ab initio or DFT calculation on isolated alloxazines and related compounds23,34,35 along with some earlier quantum-mechanical (QM)/molecular-mechanics (MM) results.36-39 The success of the DFT method in elucidating the complex phenomenon like hydrogen bonding prompts us to use it further in studying the energetic parameters of isolated LC and its complex with water molecules. Several conformers of LC are possible, and the structures are given in Chart 1. To elucidate the lowest energy conformer, full geometry optimization was performed for all the structures of LC both in isolated 2457
dx.doi.org/10.1021/jp1102687 |J. Phys. Chem. A 2011, 115, 2456–2464
The Journal of Physical Chemistry A
ARTICLE
Figure 1. Steady-state absorption (left panel) and fluorescence emission (right panel) spectra of ∼6.0 10-6 mol dm-3 LC solution in 1,4-dioxane (a), acetonitrile (b), and water (c). The excitation wavelengths are 330 nm (solid line) and 380 nm (open circles) for both a and b, whereas for water, the high-energy excitation was done at 350 nm.
and in their hydrated complexes using the B3LYP method and the 6-311þþG(d,p) basis set as implemented in Gaussian03 program package.40 Frequency calculations were done in each stationary point to characterize the minimum energy equilibrium structure. Vertical transition energies up to the first 10 singlet excited states and corresponding oscillator strengths for all the structures were estimated using time-dependent DFT method (TD-DFT) at the same level of calculation. The effect of solvent on the energy parameters was incorporated by self-consistent reaction field calculation using polarizable continuum model (SCRF-PCM) as implemented in Gaussian03 software. Similar methodology was successfully applied recently for characterizing hydrogen-bonded structure and the spectroscopic properties of large organic heterocyclic systems.24,41-43
3. RESULTS AND DISCUSSION 3.1. Steady-State Spectral Properties in Pure Solvents. Figure 1 shows some representative absorption and emission spectra of LC in different solvents. It is observed that LC shows two absorption peaks; the first one is within the spectral range of 280∼350 nm with a maxima at ∼330 nm, whereas the second one is relatively structured and within the spectral range of 350∼420 nm with a peak position around ∼375 nm. Fluorescence emission spectra obtained by exciting at 330 nm show good mirror image relationship with the absorption profile having a structured emission in the 350-400 nm range with a clean broad emission peak at 450 nm. Interestingly, in aqueous media, the structured profile is lost in both the absorption and emission spectra. Excitation at the absorption peak (∼340 nm) as well as at the shoulder (∼380 nm) gives similar structureless and broad emission at 475 nm (Figure 1, top panel). The
excitation spectra obtained by monitoring the broad low-energy emission peak nicely resemble the absorption spectra. Multiple peaks in the absorption spectra and structured fluorescence indicate that the excited states of LC are vibronically coupled even in polar solvents. However, in aqueous media, there exists a strong specific interaction of the excited states with the solvent resulting in broad and structureless profile both in absorption and emission spectra. From the relatively large absorption coefficient (εmax ∼ 5 104 dm3 mol-1 cm-1) of the low-energy absorption band, it can be concluded that this transition is S1(π) r S0(π) in nature. The assignment of the absorption band as well as the origin of the structured absorption/fluorescence band is further confirmed from theoretical calculations described in the following sections. 3.2. Solvatochromism of LC Photophysics: Importance of Hydrogen-Bond Acidity of the Solvent. The spectroscopic behavior of LC was studied in a series of solvents with varying polarity and hydrogen-bonding parameters given in Table 1. Table 2 summarizes the steady-state spectral behavior of LC in these solvents. Although the results do not show any regular variation of steady-state spectral properties, careful observation reveals several interesting trends. For example, fluorescence maxima (λem) show appreciable shift in protic solvents along with almost 2-fold increase in fluorescence quantum yield (φf) when compared with their aprotic counterpart. Similar observations were also reported in an earlier report, although the number of solvents taken was relatively few.23 To verify the effect of solvent polarity, several steady-state spectral parameters (P) like emission maxima (νem) and Stokes shift (Δνss) of LC in a variety of solvents mentioned in Table 1 were plotted against the solvent polarity parameter Δf (ε, n) with the general form of LippertMataga (LM) equation given below.44,45 2458
dx.doi.org/10.1021/jp1102687 |J. Phys. Chem. A 2011, 115, 2456–2464
The Journal of Physical Chemistry A
ARTICLE
Table 1. Solvent Parameters Kamlet-Taft solvent parameters no.
Δf(ε, n)
a
solvents
ET(30)
b
R
β
π*
1
benzene
0
34.3
0.0
0.10
0.59
2
toluene
0.02
33.9
0.0
0.11
0.54
3
1,4-dioxane
0.03
36
0.0
0.37
0.55
4
ethyl acetate
0.19
38.1
0.0
0.45
0.55
5
tetrahydrofuran
0.21
37.4
0.0
0.55
0.58
6
dichloromethane
0.22
40.7
0.0
0.1
0.81
7
1-pentyl alcohol
0.25
49.1
8 9
1-butanol DMSO
0.26 0.26
49.7 45.1
0.84 0.0
0.84 0.76
0.47 0.1
10
DMF
0.27
43.2
0.0
0.69
0.88
11
1-propanol
0.27
50.7
0.84
0.90
0.52
12
isopropanol
0.27
48.4
0.76
0.95
0.48
13
acetone
0.28
42.2
0.08
0.48
0.71
14
acetonitrile
0.3
45.6
0.19
0.40
0.75
15
methanol
0.31
55.4
0.98
0.66
0.60
16
water
0.32
63.1
1.17
0.47
1.09
a Polarity parameter (=[(ε - 1)/(2ε þ 1)] - [(n2 - 1)/(2n þ 1)]) where solvent dielectric constant and refractive indices are represented by ε and n, respectively. b Reichardt solvent parameter.
Table 2. Steady-State Spectral Properties LC in Homogeneous Solventsa solventsb
νabs/cm-1
νem/cm-1
Δυss/cm-1
φf/10-2
1
26 316
23 419
2897
3.1
2 3
26 316 26 316
23 474 23 255
2842 3060
2.3 1.9
4
26 316
23 474
2842
5.9
5
26 247
23 474
2773
4.4
6
26 247
23 529
2717
11.3
7
25 974
22 471
3502
8.2
8
25 974
22 421
3553
9.3
9
26 042
22 779
3263
1.5
10 11
26 110 25 974
22 988 22 522
3121 3452
4.8 8.8
12
26 042
22 573
3468
5.9
13
26 316
23 310
3006
3.8
14
26 247
22 779
3468
3.5
15
26 042
22 371
3670
6.8
16
25 974
21 551
4422
9.3
Abbreviations used: ν = absorption and emission energy; Δνss = Stokes shift; φf = fluorescence quantum yield.; b The name of the solvents are listed in Table 1. a
P ¼ P0 þ a Δf ðε, nÞ
ð2Þ
where the orientation polarizability, Δf(ε, n), is related to the solvent dielectric constant (ε) and the refractive index (n) with the relation Δf ðε, nÞ ¼
ε-1 n2 - 1 - 2 2ε þ 1 2n þ 1
ð3Þ
P0 is the measured property in gas phase or in noninteracting solvents like hexane, and the slope of the equation a represents a
term containing the difference in dipole moment between the ground and the excited states. From the results given in Figure 2 (lower panel), it is clear that the spectroscopic properties of LC do not show any regular solvatochromism behavior on the solvent polarity parameter. This observation points to the existence of specific solute-solvent interactions. As a first trial, the empirical solvent polarity scale, ET(30), built with a betaine dye, was used to correlate the solvent dependence of the steady-state spectral properties of LC. The uniparametric scale depends on both the solvent dielectric properties and the hydrogen-bonding acidity, but it does not take care of solvent hydrogen-bonding acceptor basicity.46 The specificity of Lewis acid base interactions in ET(30) parameter arise from the negative charge localized on the phenolic oxygen of the betaine molecule. As it is seen in Figure 2 (middle panel) again, a relatively better correlation (with correlation coefficient R ≈ 0.95) is obtained for all the spectroscopic parameters like fluorescence maxima (νem) and Stokes shift (Δνss). This clearly indicates that apart from solvent polarity, LC solvatochromism is mostly modulated by solvent hydrogen bond donor acidity (R), whereas solvent hydrogen bond acceptor basicity (β) is relatively less important. To confirm this prediction further, the steadystate spectral properties of LC are interpreted by means of the linear solvation energy relationship (LSER) concept using Kamlet-Taft eq 447 P ¼ P0 þ sπ þ aR þ bβ ð4Þ where P is the value of the solvent-dependent property to be modeled and P0, s, a, and b are the coefficients determined from the LSER analysis. The term π* indicates the measure of solvent dipolarity/polarizability,48 whereas R and β are the scale of hydrogen bond donation acidity and acceptance basicity of the solvent, respectively.49 The corresponding parameters for 14 solvents are taken from literature50,51 and are given in Table 1. Regression analysis of fluorescence maxima (νem) and Stokes shift (Δνss) with solvent properties results in the following 2459
dx.doi.org/10.1021/jp1102687 |J. Phys. Chem. A 2011, 115, 2456–2464
The Journal of Physical Chemistry A
ARTICLE
Figure 2. Plot of fluorescence maxima (left panel, open squares) and Stokes shift (right panel, shaded squares) against polarity parameter, Δf(ε, n), using Lipper-Mataga plot (a), solvent ET(30) parameter (b), and theoretically calculated values using Kamlet-Taft equation (c) of LC in different solvents.
correlation equations: νem ðexp t, cm-1 Þ ¼ 24161:66 - 1090:3π - 1048:96R - 303:3β
ð5Þ
ΔνSS ðexp t, cm-1 Þ ¼ 2337:47 þ 880:78π þ 874:12R þ 42:16β
ð6Þ
The correlation diagrams of the experimental values with those calculated from the above equations are shown in the uppermost panel of Figure 2. Reasonably acceptable correlation coefficient (R) and the value of the slope of the linear plot close to unity indicate a very good correlation of the experimentally observed and theoretically calculated parameters. A close look into the above equations reveals several interesting features for LC solvatochromism: (1) in general, the contributions from a parameter is always more significant than b beside the s parameter indicating the importance of solvent hydrogen-bonding acidity in LC spectroscopy as indeed pointed out in the discussion earlier; (2) the importance of hydrogen-bonding acidity in LC spectroscopy points toward an efficient charge localization in LC upon excitation (see DFT calculation results in section 3.4.2); finally, (3) the large emission spectral shift in water is due to the negative value of a parameter and its corresponding larger contribution. In summary, the solvatochromic analysis reveals that in polar protic solvents like water, for example, several spectroscopic species may be present because of the hydrogen-bonded donor and acceptor properties of the solvent in the ground state; however, in the excited sate, the main fluorescing species is originated because of the hydrogen-bonded complex formation of LC through the solvent hydrogen-bonding acidity behavior. Theoretical calculation results (see below) indeed confirm that the extensive charge localization on N10 (see Scheme 1 for atom numbering) in the excited state plays a
Figure 3. Time-resolved fluorescence decay profile (open circles) and simulated data (solid line) of ∼5.0 10-6 mol dm-3 LC solution in different solvents. IRF indicates the instrument response function. The magnitude of τf in different solvents is tabulated in the inset. Solvent numbers are the same as mentioned in Table 1.
key role toward the formation of hydrogen bond in protic medium like water or acetic acid with LC, which subsequently initiates the tautomerization leading to the formation of isoalloxazinic structure through a six-membered cyclic transition state. This prediction is supported by a two-step mechanism for acidcatalyzed proton transfer in LC proposed earlier.52 3.3. Time-Resolved Fluorescence Measurements in Homogeneous Environments. The fluorescence decay times (τf) of LC are measured in different solvents at room 2460
dx.doi.org/10.1021/jp1102687 |J. Phys. Chem. A 2011, 115, 2456–2464
The Journal of Physical Chemistry A
ARTICLE
Table 3. Ground-State Energy Parameters (in Hartree) of Different Possible Conformers of LC Obtained by Full Geometry Optimization at the B3LYP/6-311þþG(d,p) Level of Calculation in the Gas Phase as well as in Some Representative Solvents conformera
a
gas phase
1,4-dioxane
acetonitrile
water
I
-833.0499109
-833.076708227
-833.082797005
-833.084065647
Ia Ib
-833.0132127 -833.0275746
-833.040351540 -833.055739194
-833.046950983 -833.062265591
-833.048484606 -833.06809155
Ic
-833.0249842
-833.054456108
-833.061430244
-833.063097388
Structures are given in Chart 1.
Table 4. Energy of Absorption (nm), Its Oscillator Strength (f), and Contribution of Different Orbitals in the Transition to the Corresponding Excited States (ES) of LC (Structure I) in Gas Phase as well as in Some Representative Solvents Obtained by TDDFT Calculation at B3LYP/6-311þþG(d,p) Level of Calculationa ES 1
2
gas phase 365.8 nm (f = 0.0626)
acetonitrile
381 nm (f = 0.0626)
water
383.2 nm (f = 0.0626)
383.5 nm (f = 0.0626)
62 f 64
0.1279
62 f 64
0.1231
62 f 64
0.1303
62 f 64
0.1319
62 f 65
0.1315
62 f 65
0.1095
62 f 65
0.1084
62 f 65
0.1083
63 f 64
0.6475
63 f 64
0.6538
63 f 64
0.6521
63 f 64
0.6517
362.3 nm (f = 0.0014)
353.6 nm (f = 0.0018)
351.6 nm (f = 0.0016)
351 nm (f = 0.0019)
60 f 64
0.1169
60 f 64
0.1034
60 f 64
0.1026
60 f 64
0.018
61 f 64
0.6748
61 f 64
0.6799
61 f 64
0.6807
61 f 64
0.6809
3
321.4 nm (f = 0.2173) 62 f 64 0.1169
4
310.2 nm (f = 0.0002)
63 f 65
a
1,4-dioxane
-0.2230
333.7 nm (f = 0.3472) 62 f 64 0.6316 63 f 65
334.8 nm (f = 0.3407) 62 f 64 0.6290
-0.1852
63 f 65
299.4 nm (f = 0.0001)
-0.1878
292.2 nm (f = 0.0001)
335.0 nm (f = 0.3398) 62 f 64 0.6287 63 f 65
-0.1878
296.7 nm (f = 0.0001)
57 f 64
-0.1364
57 f 64
-0.1383
57 f 64
-0.1451
57 f 64
-0.1469
59 f 64
0.6218
60 f 64
0.6643
60 f 64
0.6640
60 f 64
0.6638
63 f 65
0.2259
The highest occupied and lowest unoccupied MOs are represented by numbers 63 and 64, respectively.
temperature using time-correlated single-photon-counting method. Some of the representative decay traces are shown in Figure 3, and the results are provided in the inset table. All decays could be reproduced with single exponential decay function. Fluorescence decay times measured in this study are in very good agreement with some of the data available in literature.23 In general, the fluorescence decay time decreases with decrease in solvent hydrogen-bonding acidity (e.g., 0.24 ns in DMSO compared to 2.69 ns in aqueous buffer medium). This indicates that specific interaction through solvent hydrogen bond donation stabilizes the fluorescing states of LC. The increase in τf values, that is, the stabilization of the fluorescing state in the presence of hydrogen bond donor, is quite obvious from the discussion made above as well as from the theoretical calculation discussed below. However, the reasons of almost 3-fold decrease of τf in DMSO (0.24 ns) when compared with noninteracting solvent, for example, dichloroethane (0.61 ns23), are not clear. 3.4. Theoretical Calculation Using Density Functional Theory. 3.4.1. Energetics of Different Conformers in the Ground State. Full geometry optimization of different conformers of LC (structures are given in Chart 1) in isolated condition as well as in different solvents with SCRF/PCM model was done using B3LYP/6-311þþG(d,p) methodology. The fully optimized structures along with all the geometrical parameters are shown in Figure 1S of the Supporting Information. The corresponding energies are listed in Table 3. It is clear that structure I is the most stable conformer in the isolated condition as well as in different solvent systems.
Therefore, the ground-state population of LC is mainly contributed by structure I, and we will confine our discussion of spectral properties on the basis of this structure only. To understand the absorption spectral behavior, we have carried out TDDFT calculation of conformer I in different solvents for the first 10 excited singlet states. The contributions of different orbitals in the corresponding excited states within the experimentally observed spectral range (280-420 nm) are given in Table 4. The calculated excitation wavelengths are in very good agreement with the absorption spectra shown in Figure 1. From the table, it is clear that the excited state of LC is mainly contributed from the HOMO (H) f LUMO (L) transition with a little contribution from H - 1 and L þ 1 orbitals also. Careful analysis of electrondensity distribution in all these orbitals (see below) confirms the π* r π nature of transition involved in optical absorption spectroscopy of LC. 3.4.2. Analysis of Frontier Orbitals: Charge Localization in the Excited State. All the four MOs involved in electronic transition of LC are depicted in Figure 4. From the discussion made in the previous section and also from the results given in Table 4, it is clear that the major contribution for the electronic absorption arises from the transition to the lowest unoccupied molecular orbital (LUMO). Careful analysis of this orbital and comparing with the occupied molecular orbitals reveals that the charge redistribution of LC occurs in the excited state. It is indeed observed that the charge localization to a greater extent occurs on the N10 atom, whereas N1 hydrogen becomes strongly acidic on excitation. Therefore, it is highly probable that excited-state 2461
dx.doi.org/10.1021/jp1102687 |J. Phys. Chem. A 2011, 115, 2456–2464
The Journal of Physical Chemistry A
ARTICLE
Figure 4. Frontier orbital diagram of different MOs for LC (I). The block arrows in LUMO indicate charge accumulation (filled arrow) removal (empty arrow) on N10 and N1, respectively, due to electronic excitation.
photophysics of LC will be strongly modulated by hydrogen bond donation of the solvent with N10 atom; on the other hand, hydrogen bond accepting solvents like DMSO mainly acts through proton abstraction from the N1 atom. Analysis of solvent dependence of LC fluorescence discussed earlier has shown that the fluorescence properties show good correlation even with solvent ET(30) parameter that mainly takes care of solvent hydrogen bond donation ability (acidity). Consequently, it is reasonable to believe that in an amphoteric medium like water or acetic acid, the key step involves the formation of hydrogen bond through solvent hydrogen bond acidity at the N10 atom. Once this hydrogen bond formation occurs, geometric rearrangement of the solvent occurs to form a cyclic sixmembered structure through solvent hydrogen bond basicity with N1 proton of LC. A similar mechanism was proposed earlier by Sikorska et al. for proton transfer in LC-acetic acid complex.52 The authors termed the second step as the formation of an appropriate structure to account for the slow proton transfer following the very fast initial hydrogen bond donated complex at the N10 position of LC with acetic acid. However, contrary to the hypothesis proposed by the authors, we believe that the first hydrogen-bonded complex is not available in the ground state. Charge localization in the excited state is the origin of this specific hydrogen bond to occur at the N10 of LC which serves as the catalytic pathway for the excited-state proton transfer toward the formation of isoalloxazine (II). 3.4.3. Water-Assisted Proton Transfer in LC: Analysis of Transition State. The ground-state energy parameters of the alloxazine (I) and isoalloxazine (II) structures in gas phase as well as in different solvents calculated using SCRF/PCM model at B3LYP/6-311þþG(d,p) level are listed in Table 1S of the Supporting Information. It is found that although structure II is less stable by 54.06 kJ mol-1 energy relative to that of structure I in the gas phase, the energy difference decreases substantially in the presence of solvents (for example, ∼26.27 kJ mol-1 in the presence of water and ∼26.30 kJ mol-1 in the presence of acetonitrile). In contrast, full geometry optimization with specific
Figure 5. Ground- and excited-state potential energy surface for watercatalyzed conversion of I to II.
hydrogen-bonded structures of I and II with water molecules predicts an energy difference of ∼34.0 kJ mol-1 between them. These results indicate the importance of specific hydrogen bond formation in the proton-transfer process and confirm that only solvent physical parameters like polarity, as imposed by SCRF/ PCM model, are not sufficient to catalyze the process. Rather than acting merely as a matrix with dielectric continuum, the solvent directly takes part in the chemical transformation through hydrogen-bonding interaction. The optimized structures along with the geometrical and energy parameters for these two conformers, specifically hydrogen bonded with water molecule, are given in Figure 2S of the Supporting Information. Only primary hydrogen-bonded complex of I and II with water was considered to give different complexes. It is obvious that additional solvent molecules will combine to give secondary water cluster and that the actual hydrated structure is far more complex than what is considered for these calculations. However, we 2462
dx.doi.org/10.1021/jp1102687 |J. Phys. Chem. A 2011, 115, 2456–2464
The Journal of Physical Chemistry A
ARTICLE
Figure 6. Ground-state structure of the transition state (TS) for water-catalyzed I f II conversion obtained by IRC calculation using B3LYP/6311þþG(d,p) method. The bond lengths are given in angstroms, and the angles are in degrees.
restrict ourselves only in the first hydration layer as the effect of internal hydrogen bonding (IHB) is expected to diminish very fast from the center of origin. Consequently, it is reasonable to believe that any further addition of water structure will have insignificant contribution toward the relative energy parameters. The calculated energy difference between I and II in the ground state is ∼34.0 kJ mol-1. The potential energy surface (PES) (Figure 5), constructed by intrinsic reaction coordinate (IRC) calculation from the transition state (TS), indicates that the water-assisted conversion of I to II is associated with a large activation barrier of ∼89.88 kJ mol-1. Thus, in ground state, the possibility of proton transfer is very less, both from kinetic and thermodynamic points of view. However, TD-DFT calculation results show that the relative energy difference between I and II is very small in the first excited state (∼1.8 kJ mol-1) and that the barrier height for the proton transfer also reduces substantially to ∼68.5 kJ mol-1. As a result, the water-catalyzed protontransfer process becomes highly feasible in the excited state. The TS structure (Figure 6) is confirmed by the presence of an imaginary frequency (∼1590.35 cm-1) which is active toward the proton translocation within N1 and N10 atoms of alloxazine moiety.
’ CONCLUSIONS The fluorescence behavior and the water-catalyzed excitedstate tautomerization process of lumichrome have been studied by steady-state and time-resolved fluorescence spectroscopy as well as by DFT calculations. The observed solvatochromism of lumichrome fluorescence is rationalized by multiparametric approach using Kamlet-Taft equation. It has been found that the photoluminescence behavior of lumichrome is strongly modulated by specific solute-solvent interaction through solvent hydrogen bond donation. These observations have also been supported by DFT calculations on lumichrome in isolated condition as well as in the presence of some representative solvents. The results predict that extensive charge localization on the N10 atom and subsequent hydrogen bond formation with
the solvent are the key processes in the excited state, which essentially explains the dominance of solvent hydrogen bond acidity toward the lumichrome fluorescence behavior.
’ ASSOCIATED CONTENT
bS
Supporting Information. Energy parameters of I and II in gas phase as well as in some representative solvents (Table 1S), full optimized ground-state geometrical parameters in gas phase of I, Ia, Ib, and Ic (Figure 1S), and specific hydrogen-bonded water complex of I and II (Figure 2S). This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*Phone: (91)-364-2722634. Fax: (91)-364-2550076. E-mail:
[email protected],
[email protected].
’ ACKNOWLEDGMENT Financial support through research project 2009/37/26/ BRNS from Board of Research in Nuclear Sciences (BRNS), Government of India, is gratefully acknowledged. The authors thank Dr. A. Bhasikuttan of Bhaba Atomic Research Center (BARC) for helpful discussion. Thanks are also due to AIRF, JNU for their help in TCSPC measurement. ’ REFERENCES (1) Dutta Choudhury, S.; Mohanty, J.; Bhasikuttan, A. C.; Pal, H. J. Phys. Chem. B 2010, 114, 10717–10727. (2) Chosrowjan, H.; Taniguchi, S.; Mataga, N.; Nakanishi, T.; Haruyama, Y.; Sato., S.; Kitamura, M.; Tanaka, F. J. Phys. Chem. B 2010, 114, 6175–6182. (3) Murakami, M.; Ohkubo, K.; Fukuzumi, S. Chem.—Eur. J. 2010, 16, 7820–7832. 2463
dx.doi.org/10.1021/jp1102687 |J. Phys. Chem. A 2011, 115, 2456–2464
The Journal of Physical Chemistry A (4) Radoszkowicz, L.; Huppert, D.; Nachliel, E.; Gutman, M. J. Phys. Chem. A 2010, 114, 1017–1022. (5) Kao, Y. T.; Saxena, C.; He, T. F.; Guo, L.; Wang, L.; Sancar, A.; Zhong, D. J. Am. Chem. Soc. 2008, 130, 13132–13139. (6) Chastain, J.; McCormick, D. B. In Chemistry and Biochemistry of Flavoenzymes, 1; Muller, F., Ed.; CRC Press: Boston, MA, 1991; pp 196-200. (7) Said, H. M.; Ortiz, A.; Moyer, M. P.; Yanagawa, N. Am. J. Physiol. Cell Physiol. 2000, 278, C270–C276. (8) Cunningham, O.; Gore, M. G.; Mantle, T. J. Biochem. J. 2000, 345, 393–399. (9) Tyagi, A.; Penzkofer, A.; Mathes, T.; Hegemann, P. J. Photochem. Photobiol., B: Biol. 2010, 101, 76–88. (10) van den Berg, P. A.; van Hoek, A.; Walentas, C. D.; Perham, R. N.; Visser, A. J. Biophys. J. 1998, 74, 2046–2058. (11) Sikorska, E.; Sikorski, M.; Steer, R. P.; Wilkinson, F.; Worall, D. R. J. Chem. Soc., Faraday Trans 1998, 94, 2347–2353. (12) Onu, A.; Palamaru, M.; Tutovan, E.; Ciobanu, C. Polym. Degrad. Stab. 1998, 60, 465–470. (13) Li, H.; Jiang, Z. Q.; Zhang, R. H. Chin. Sci. Bull. 2010, 55, 2829–2834. (14) Rajendar, B.; Nishizawa, S.; Teramae, N. Org. Biomol. Chem. 2008, 21, 670–673. (15) Grininger, M.; Zeth, K.; Oesterhelt, D. J. Mol. Biol. 2006, 357, 842–857. (16) Rhee, H. W.; Choi, S. J.; Yoo, S. H.; Jang, Y. O.; Park, H. H.; Pinto, R. M.; Cameselle, J. C.; Sandoval, F. J.; Roje, S.; Han, K.; Chung, D. S.; Suh, J.; Hong, J. I. J. Am. Chem. Soc. 2009, 131, 10107–10112. (17) Bertolotti, S. G.; Previtali, C. M.; Rufs, A. M.; Encinus, M. V. Macromolecules 1999, 32, 2920–2924. (18) Bairi, P.; Roy, B.; Nandi, A. K. J. Phys. Chem. B 2010, 114, 11454–11461. (19) Zen, Y. H.; Wang, C. M. J. Chem. Soc., Chem. Commun. 1994, 2625–2626. (20) Wang, Z. W.; Rizzo, C. J. Org. Lett. 2000, 2, 227–230. (21) Song, P.-S.; Sun, M.; Koziolowa, A.; Koziol, J. J. Am. Chem. Soc. 1974, 96, 4319–4323. (22) Koziol, J. Photochem. Photobiol. 1966, 5, 41–54. (23) Sikorska, E.; Khmelinskii, I. V.; Prukala, W.; Wiliams, S. L.; Patel, M.; Worall, D. R.; Bourdelande, J. L.; Koput, J.; Sikorski, M. J. Phys. Chem. A 2004, 108, 1501–1508. (24) Moyon, N. S.; Chandra, A. K.; Mitra, S. J. Phys. Chem. A 2010, 114, 60–67. (25) Meech, S. R.; Phillips, D. J. Photochem. 1983, 23, 193–217. (26) Mataga, N.; Kubota, T. In Molecular Interactions and Electronic Spectra; Marcel Dekker Inc.: New York, 1970. (27) Moyon, N. S.; Mitra, S. Chem. Phys. Lett. 2010, 498, 178–183. (28) Ireta, J.; Neugebauer, J.; Scheffler, M. J. Phys. Chem. A 2004, 108, 5692–5698. (29) Ni, W.; Mosquera, R. A.; Perez-Juste, J.; Liz-Marzan, L. M. J. Phys. Chem. Lett. 2010, 1, 1181–1185. (30) Hao, M.-H. J. Chem. Theory Comput. 2006, 2, 863–872. (31) Chen, X.; Brauman, J. I. J. Am. Chem. Soc. 2008, 130, 15038–15046. (32) Myshakina, N. S.; Ahmed, Z.; Asher, S. A. J. Phys. Chem. B 2008, 112, 11873–11877. (33) Potter, R. G.; Camaioni, D. M.; Vasiliu, M.; Dixon, D. A. Inorg. Chem. 2010, 49, 10512–10521. (34) Klaum€unzer, B.; Kr€oner, D.; Saalfrank, P. J. Phys. Chem. B 2010, 114, 10826–10834. (35) Salzmann, S.; Marian, C. M. Photochem. Photobiol. Sci. 2009, 8, 1655–1666. (36) Salzmann, S.; Silva-Junior, M. R.; Thiel, W.; Marian, C. M. J. Phys. Chem. B 2009, 113, 15610–15618. (37) Szymusiak, H.; Konarshi, J.; Kozioz, J. J. Chem. Soc., Perkin Trans. 1990, 2, 229–236. (38) Komasa, J.; Rychlewski, J.; Kozioz, J. J. Mol. Struc. (THEOCHEM) 1988, 170, 205–212.
ARTICLE
(39) Sun, M.; Moore, T. A.; Song, P. S. J. Am. Chem. Soc. 1972, 94, 1730–1740. (40) Frisch, M. J. et al. Gaussian 03, revision D.01; Gaussian, Inc.: Wallingford, CT, 2004. (41) Mitra, S.; Singh, T. S.; Mandal, A.; Mukherjee, S. Chem. Phys. 2007, 342, 309–317. (42) Singh, T. S.; Mitra, S.; Chandra, A. K.; Tamai, N.; Kar, S. J. Photochem. Photobiol., A: Chem. 2008, 197, 295–305. (43) Moyon, N. S.; Singh, T. S.; Mitra, S. Biophys. Chem. 2008, 138, 55–62. (44) Mataga, N.; Kaifu, Y.; Koizumi, M. Bull. Chem. Soc. Jpn. 1956, 29, 465–470. (45) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3rd Edition; Springer: New York, 2006; pp 208-210. (46) Reichardt, C. Chem. Rev. 1994, 94, 2319–2358. (47) Kamlet, M. J.; Abboud, J. M.; Taft, R. W. Prog. Phys. Org. Chem. 1981, 13, 485–630. (48) Abboud, J. M.; Kamlet, M. J.; Taft, T. W. J. Am. Chem. Soc. 1977, 99, 8325–8327. (49) Miller, J. V.; Bartick, E. G. Appl. Spectrosc. 2001, 55, 1729–1732. (50) Marcus, Y. Chem. Soc. Rev. 1993, 22, 409–416. (51) Kamlet, M. J.; Abboud, J. M.; Abraham, M. H.; Taft, R. W. J. Org. Chem. 1983, 48, 2877–2887. (52) Sikorska, E.; Koziozowa, A.; Sikorski, M.; Siemiarczuk, A. J. Photochem. Photobiol., A: Chem. 2003, 157, 5–14.
2464
dx.doi.org/10.1021/jp1102687 |J. Phys. Chem. A 2011, 115, 2456–2464
