Embed Size (px)
Citation preview
Ro
Fa
b
a
ARRAA
KMPSCPS
1
mMmmaaswvis
h1
Journal of Photochemistry and Photobiology A: Chemistry 284 (2014) 8–17
Contents lists available at ScienceDirect
Journal of Photochemistry and Photobiology A:Chemistry
jo u r n al homep age: www.elsev ier .com/ locate / jphotochem
evisiting the photophysics and photochemistryf methylene violet (MV)
. Ronzania, A. Trivellab, P. Bordata, S. Blanca, S. Lacombea,∗
IPREM UMR CNRS 5254, Université de Pau et Pays de l’Adour, Hélioparc, 2 avenue du Président Angot, 64053 Pau cedex 9, FranceICCF, UMR CNRS 6296, Université Blaise Pascal, Campus des Cézeaux, 24 avenue des Landais BP 80026, 63171 Aubière Cedex, France
r t i c l e i n f o
rticle history:eceived 12 February 2014eceived in revised form 25 March 2014ccepted 29 March 2014vailable online 8 April 2014
eywords:ethylene violet
henothiazinepectroscopyomputational analysishotochemistryinglet oxygen
a b s t r a c t
Methylene violet (MV) is known for its photosensitizing properties for singlet oxygen (1O2) generationupon visible-light irradiation and various examples of its potential use in photodynamic inactivation ofmicroorganisms and for photomedicinal purposes were reported. Notwithstanding its good photosensiti-zation properties, there is a loss of clinical tests in the recent literature, probably related to the incompleteinformation concerning the photophysics of this dye and the effects of the medium on its properties. Wethus studied both experimentally and numerically the solvatochromic effects and the presence of acidson the absorption and fluorescence spectra of MV. In chloroform, the highly favoured and reversibleformation of a mono-protonated derivative of MV, MVH+, was clearly demonstrated experimentally andsupported by ab-initio calculations. In acetonitrile, the more complex experimental absorption bandscan be explained by several assumptions including a thermodynamic equilibrium between MVH+ andMVH2
2+ or the formation of a possibly oblique dimer (MVH+)2 induced by the formation of intermolecularhydrogen bonds. All these data point out to the high sensitivity of MV to intermolecular hydrogen bondingand to protonation, inducing a significant influence of the chemical environment on the photosensiti-zation mechanisms of MV: from transient spectroscopy and photochemical experiments, an electron
transfer side-mechanisms was shown to occur only in aprotic solvents, together with a very general andefficient singlet oxygen production whatever the solvent. MV thus represents a very good singlet oxygensensitizer even though the influence of the chemical environment should be carefully considered for anyapplication. Moreover, MV might be a sensitive probe for the detection of acids in organic non-proticsolvents.© 2014 Elsevier B.V. All rights reserved.
. Introduction
Phenothiazines (PTZ) represent a very interesting class ofolecules, with suitable properties for widespread applications.ethylene blue (MB+) is probably the most well-known com-ercially available PTZ photosensitizer (PS) [1,2]. MB+ is used asedicinal and textile stain, antioxidant and antiseptic, diagnostic
gent, antidote for cyanide and nitrate, drug for malaria, Alzheimernd other neurodegenerative diseases [2]. MB+ efficiently producesinglet oxygen [3], it is non-toxic and it absorbs light at 650 nm,hich makes it suitable for photodynamic inactivation (PDI) of
iruses and bacteria in blood fractions, notably for plasma steril-zation [4–12], and for photodynamic therapy (PDT). Among theeveral MB+ derivatives, which can be prepared for tuning its
∗ Corresponding author. Tel.: +335 59407579.E-mail address: [email protected] (S. Lacombe).
ttp://dx.doi.org/10.1016/j.jphotochem.2014.03.019010-6030/© 2014 Elsevier B.V. All rights reserved.
photochemical/photophysical properties, the neutral dye methy-lene violet (MV, Chart 1), which has been studied for potentialapplications in biology, showed interesting preliminary results.MV presents a higher activity than MB+ for killing intracellularviruses thanks to a much easier penetration in cells due to theabsence of the positive charge. Nonetheless, MV is more easilytrapped by lipoproteins than MB+, which can limit its activity inbiological media [8,13]. MV can form singlet oxygen by energytransfer with relatively high efficiency and has a particular affinitytowards DNA which makes it a good PS for photodynamic ther-apy [14,15]. Some graftable MV derivatives have been prepared,and enhanced singlet oxygen production was achieved by addingiodine atoms on its structure (heavy atom effect) [8,16]. Previ-ous spectroscopic and photophysical data [17] on MV showed a
strong dependence on the solvent. For a better understanding ofMV properties, the solvatochromic behaviour of MV, its singletoxygen quantum yield, transient absorption spectra and photo-chemical properties were investigated in some selected solvents.
F. Ronzani et al. / Journal of Photochemistry and P
Oceo
2
lcdagite(d(aop
b1sFrnccsg1o2stms(t(dett
Aaortet(Nt
Chart 1. Chemical structure of methylene violet (MV).
n the basis of our experimental results and calculations on thisompound, we could highlight an impressive sensitivity to the pres-nce of H+, which makes of MV a very good photosensitizer andptical probe for the detection of H+ ions.
. Experimental part and calculation methods
Methylene violet (Bernthsen, MV, dye content > 72%), methy-ene blue (certified by Biological Stain Commission, MB+, dyeontent > 80%), rubrene (5,6,11,12-tetraphenylnaphtacene, 99%),ibutylsulfide (Bu2S, 97%) and methanesulfonic acid (CH3SO3H,nhydrous) were purchased from Sigma–Aldrich. All spectro-rade solvents, except for triethylamine (Et3N, Acros Organ-cs), were purchased from Sigma–Aldrich: n-heptane, carbonetrachloride (CCl4), diethyl ether (Et2O), toluene, dioxane,thyl acetate (AcEt), tetrahydrofuran (THF), acetone, acetonitrileACN), dichloromethane (CH2Cl2), chloroform (CHCl3), pyridine,imethylformamide (DMF), dimethylsulfoxide (DMSO), ethanolEtOH) and methanol (MeOH). All chemicals and solvents were useds received. The purity of MV was declared by the suppliers with-ut further specifications. It was not purified to be consistent withrevious published works [2,18].
UV–visible absorption spectra were recorded with a doubleeam Cary 5000 spectrophotometer in steps of 0.5 nm using a
cm quartz optical cell (Hellma). Corrected steady-state emis-ion spectra were measured using a photon counting EdinburghLS920 fluorescence spectrometer equipped with a Xe lamp. Time-esolved fluorescence experiments were carried out by using aano-second flash-lamp (nF900) and time-correlated single photonounting (TCSPC). The lifetime data were analyzed with the re-onvolution fit (including instrument response) of the Edinburghoftware. The concentrations of all compounds were adjusted toive an absorbance around 0.1 at the absorption maximum in a
cm fluorescence quartz optical cell (Hellma). For the detectionf transient species a Nd:YAG laser (GCR 130-1, pulse width 9 ns,66 nm or 355 nm) was used for sample irradiation. The monitoringystem consisted of a 150 W pulsed xenon arc lamp, a R928 pho-omultiplier and a 05-109 Spectra Kinetics Applied Photophysics
onochromator. Signals were digitized by a HP54522A oscillo-cope. The samples were irradiated in fluorescence quartz cells1 cm, Hellma); no influence of the excitation wavelength on theransient signals and spectra was noticed. The laser pulse energyP) was measured using a joulemeter Ophir Optronics Ltd. Eachata was the average of 5 measurements. The absorbance at pulsend was measured for various P values to check the linearity ofhe dependence of A on P and the monophotonic formation of theransient.
The spectrophotometric data recorded at various [H+] values inCN or CHCl3 were processed with Specfit program [19], whichdjusts the stability constants and the corresponding absorptivitiesf the species formed at equilibrium. Specfit uses factor analysis toeduce the absorbance matrix and to extract the eigenvalues prioro the multiwavelength least squares fitting of non-linear param-ters (equilibrium constants) of the reduced data set according
o a modified Marquardt-Levenberg algorithm. Linear parametersmolar absorptivities) are computed internally using Advancedewton–Raphson algorithm for equilibrium speciation calcula-ions. MV was considered to form a protonated species, MVH+ and
hotobiology A: Chemistry 284 (2014) 8–17 9
the protonation constant (Eq. (1)), ˇ1, may then be expressed asfollows:
MV + H+ � MVH+ ˇ1 = [MVH+]/[MV][H+] (1)
As each species is characterized by its own spectrum for UV/visexperiments, the Beer–Lambert law relates the absorbance A� forone cm path length as:
A� = [MV] × ε�,MV + [MVH+] × ε�,MVH+ (2)
where ε� is the molar absorptivity of the species (M−1 cm−1) mea-sured at the wavelength �.
The procedures for the detection of the transient species and thedetermination of the quantum yield of singlet oxygen production(˚�), via both direct and indirect detection, have been describedelsewhere [18].
Preparative batch reactions were carried out to identify theproducts of Bu2S photooxidation. Air-equilibrated ACN solutions(50 mL) containing MV (1.0 × 10−4 M) and Bu2S (5.0 × 10−3 M) wereirradiated in a Rayonet® reactor equipped with 16 × 8 W lampswith �max = 575 nm (RPR-5750 A). The reaction was checked bymeasuring the solution absorbance with a quartz cell with opti-cal path of 0.01 cm. The reactions were performed until no morevariations in the UV–vis spectrum of MV were noticed (approx.20 min). The reaction mixture was analyzed by 1H NMR on a BrukerAdvance 400 NMR spectrometer in CDCl3 at 25 ◦C using a 5 mminverse broadband probe, at 400.13 Hz, and calibrated with respectto the solvent signal.
The photobleaching and reactivity of MV in solution were stud-ied with the following set-up. 3 mL solutions (0.8 to 1.2 × 10−5 M,fluorescence quartz cells) containing the PS, magnetically stirred at25 ◦C, were irradiated with a 200 W Xe-Hg Lamp; a Cornerston 260motorized monochromator, inserted between the light source andthe analytical equipment, was set to � = 0 (white light). A PerkinElmer double beam, double monochromator Lambda850 UV-visspectrophotometer was used for the analysis. The quartz cells werepositioned and irradiated directly on the support of the spectropho-tometer (light source perpendicular to the analytical beam). Theamount of bleached PS was evaluated from the decrease of theabsorbance maximum, in terms of MMV s−1. From the irradiancespectrum of the light source and the absorption spectra of MV,the photon flux absorbed by the PS (Pa, Einsteins L−1 s−1) could beobtained by means of Eq. (3) [20]:∑
�
Pa,� =∑
�
P0,�(1 − 10−A� ) (3)
P0 (emitted photon flux) was measured with an InternationalLight ILT900 spectroradiometer. By dividing the decrease of PS con-centration by Pa, the quantum yield of photobleaching (�bleaching)was obtained (molMV Einsteins−1). In the presence of Bu2S (15 �L ofsulfide were added to the solution ([Bu2S] = 0.028 M, large excess)),the photobleaching could not be estimated owing to the spectralchanges induced by the formation of acids (see Section 5).
Quantum chemical calculations have been carried out usingGaussian 09 [21]. Standard calculations for structural optimizationhave been performed at the B3LYP/cc-pVTZ DFT level and electronictransitions have been calculated by TDDFT (time-dependent den-sity functional theory) at the same level for consistency. The solventis taken into account with the PCM model implemented in Gaussian09 [22]. Vibrational frequencies have been checked to ensure thatoptimized conformations correspond to minima. Moreover, extra-calculations have been performed with CAM-B3LYP and WB97XD
DFT functionals and extended bases as 6–31++G(3df,3dp) and AUG-cc-pVTZ. These extra-calculations confirm the results found withthe previous modest strategy (B3LYP/cc-pVTZ). So, in the followingdiscussion, interpretation is based on the results determined at the
10 F. Ronzani et al. / Journal of Photochemistry and Photobiology A: Chemistry 284 (2014) 8–17
Fig. 1. Normalized absorption spectra of MV in various solvents: n-heptane (greyd(A
Bt
3
3
iptt
ecsawtmeK
E
Table 1Wavelengths of absorption maximum (�max, first allowed transitions), wavelengthsof fluorescence emission maximum (�max,em), fluorescence quantum yields (˚F) andfluorescence lifetimes (�F) in ACN, acidified ACN and MeOH. Air-equilibrated solu-tions at room temperature. Reference for determining ˚F: methylene blue in MeOH(0.03) [29].
ACN ACN + CH3SO3H MeOH
�max/nm 549 620 600�max,em/nm 628 675 642
Ffas
ashed line), dioxane (black dashed line), acetonitrile (black dotted line), chloroformblack solid line), methanol (grey solid line) and water (grey dotted-dashed line).ir-equilibrated solutions at room temperature.
3LYP/cc-pVTZ level. Charge distribution was also investigated byhe ESP fit method.
. Experimental results
.1. Solvatochromism
The absorption spectra of MV in various solvents are reportedn Fig. 1: the differences induced by the solvent were significant,articularly in the visible range (Fig. SI 1, ESI). A shift in the absorp-ion maximum wavelength of more than 100 nm from n-heptaneo water was thus observed.
On the many expressions proposed to relate the chemical prop-rties of solvents to the wavelengths of light absorption by ahromophore, the Kamlet-Taft expression was found to be veryuccessful: we addressed the dependence of the position of thebsorption maximum considering the empirical parameters �*hich combines the polarity and polarizability of the solvents and
he Dimroth-Reichardt’s polarity factor ET(30) (kcal mol−1) whicheasures the ionizing power of a solvent [23,24]. This latter param-
ter was normalized (Eq. (4)) as usually to be included in the
amlet-Taft model for solvatochromism analysis [25–28]:NT = ET (30) − 30.7
32.4(4)
ig. 2. Variation of the absorption maximum wavelength of MV as a function of the polaor aprotic (grey solid line, grey empty circles) and protic (black dashed line, black diam
function of the normalized Dimroth-Reichardt’s polarity (ENT
) of the solvent; a unique colvents.
˚F 0.09 ± 0.01 0.01 0.03 ± 0.01�F/ns 1.49 ± 0.10 0.35 ± 0.05 0.50 ± 0.05
where 30.7 is the ET(30) value of tetramethylsilane (TMS) and 32.4represents the difference between water and TMS (63.1–30.7).
The absorption maxima recorded in several solvents, togetherwith the solvents’ properties considered for the Kamlet-Taft model,are reported in Table SI 1, ESI. The effect of hydrogen bond dona-tion ability (˛) and hydrogen bond acceptance (ˇ) on the absorptionspectrum of MV was analyzed. A linear correlation between and�max was observed (Fig. SI 2, ESI). However, since the absorptionwavelength also varied for solvents with = 0, it was concluded thatthis parameter cannot account for all the results. Considering �*(Fig. 2) two different linear correlations were found for non-proticand protic solvents (ethanol, methanol and water) respectively.This result suggested the presence of two forms of MV in the twokinds of solvents, as already outlined by Otsuki and Taguchi [17]. Agood linear fitting was also obtained between �max and the solventpolarity Dimroth-Reichardt’s scale, (Fig. 2 inset, Eq. (4)) and lead tothe additional conclusion that both species in each type of solventswere sensitive to its ionizing power.
The solvent also influenced the fluorescence emission spectra ofMV, as well as its fluorescence lifetime and fluorescence quantumyield (Table 1).
3.2. Effect of acidity
In protic solvents such as methanol and water, the absorptionspectra were close to those recorded for positively charged PTZdyes like MB+. For investigating in more detail the possible pro-tonation reactions occurring on MV in the presence of H+ ions[17], the variations in the absorption spectrum of MV was mon-itored by adding methanesulfonic acid (CH SO H) to MV diluted in
3 3chloroform (2 × 10−5 M, Fig. 3) or acetonitrile (3.2 × 10−6 M, Fig. 4).The addition of acid (2 × 10−6 to 2 × 10−4 M) to CHCl3 induced adecrease of the main band (552 nm) and the formation of newrity/polarizability parameter (�*) of the solvent; different correlations were foundonds) solvents. Inset: variation of the absorption maximum wavelength of MV asorrelation was found for aprotic (grey empty circles) and protic (black diamonds)
F. Ronzani et al. / Journal of Photochemistry and Photobiology A: Chemistry 284 (2014) 8–17 11
Fig. 3. Evolution of the absorption spectrum of MV in CHCl3 for subsequent addi-tions of CH SO H (2 × 10−6 to 2 × 10−4 M). Inset: Variation of MV absorbancea0
ad
1tn
eia[
•
•
ete
Fta0
Fig. 5. Evolution of the fluorescence emission spectrum (�ex = 550 nm) of MV in ACNfor subsequent additions of CH3SO3H (1.3 × 10−8 to 1.4 × 10−5 M). The decrease ofthe signal intensity is shown by the arrow (a); black dotted line: spectrum after theaddition of Et3N. (b) normalized fluorescence emission spectra at different CH3SO3H
3 3t 467, 552 and 611 nm with increasing relative concentrations of H+ for < [CH3SO3H]tot/[MV]tot ≤ 5 with [MV]tot = 2 × 10−5 M), l = 1 cm.
bsorption bands at 575 nm and 611 nm, giving rise to a well-efined intense band at higher H+ concentrations.
In ACN (Fig. 4), upon CH3SO3H addition (1.3 × 10−8 to.4 × 10−5 M), the intensity of the main band decreased giving riseo a maximum at 578 nm with a shoulder at 552 nm, with simulta-eous increase of two well defined new maxima at 611 and 467 nm.
In both solvents, the presence of isosbestic points suggestedquilibrium between two absorbing species. When comparing thensets of Figs. 3 and 4, where the decrease of the band of MVt 552 nm and the increase of the new bands are given vs theH+]tot/[MV]tot ratio, two main differences were noticed:
In CHCl3 for both 552 and 611 nm bands, the absorbancevaried until the ratio [H+]tot/[MV]tot ≈ 1 while in ACN the inten-sity of the three bands at 467, 552 and 611 nm varied up to[H+]tot/[MV]tot ≈ 3.The intensity of the 467 nm band in ACN was correlated with thatof the 611 nm band and was not observed to increase in CHCl3.
Another important feature was the total reversibility of this
ffect: after the addition of enough triethylamine (Et3N) to neu-ralize the amount of acid in the cell, the absorption spectra werexactly the same as in pure ACN or CHCl3, indicating that theig. 4. Evolution of the absorption spectrum of MV in ACN for subsequent addi-ions of CH3SO3H (1.3 × 10−8 to 1.4 × 10−5 M). Inset: Variation of MV absorbancet 467, 552 and 611 nm with increasing relative concentrations of H+ for
< [CH3SO3H]tot/[MV]tot ≤ 5 with [MV]tot = 3.2 × 10−6 M), l = 1 cm.
concentrations: the arrow points out the red-shift induced by the addition of acidand the black dashed line the reversibility (addition of Et3N) of the process.
variation of the MV absorbance was due to the presence of H+ ionsin solution.
Numerical fitting of the absorption spectra in ACN and CHCl3in the presence of acid was carried out (see experimental part). InCHCl3, for low relative acid concentrations ([H+]tot/[MV]tot < 5), wecould outline the formation of a protonated form of MV (MVH+,Eq. (1), Fig. SI 5, ESI), with absorption maxima at 611 and 575 nm,consistent with the spectrum at [H+]tot/[MV]tot ≈ 1 (Fig. 3, inset).In ACN, the interpretation of the four main transitions (467, 552,578 and 611 nm, Fig. 4) was more difficult. The data did not fitwith a simple model with one or two equilibria involving mono orbi-protonated species MVH+ or MVH2
2+. It was supposed that theparameters were correlated (similar absorptivity or concentrationratio) and could not afford reliable results. A quantitative analy-sis of the acid influence on the spectral changes would require theknowledge of the dissociation constants of the considered speciesin all the investigated solvents.
In CHCl3 (Fig. SI 4, ESI) and in ACN (Fig. 5), increasing amountsof methanesulfonic acid induced a strong decrease of the fluo-rescence emission intensity and a 20–30 nm bathochromic shiftof the band. Consistently to the absorption spectra, the additionof a base (Et3N) for H+ neutralization led to a final spectrumcompletely superimposable to the one before any acid addition.The presence of different species in ACN and acidified ACN wasalso suggested by the differences in fluorescence lifetimes andquantum yields in the two media (Table 1): �F in acidified ACNwas 0.35 ± 0.05 ns (against 1.49 ns in ACN) and ˚F was 0.01 (vs0.09 for MV in ACN). Noticeably, the �F and ˚F values in MeOHwere closer to the ones in acidified ACN. With such small fluo-rescence quantum yields and lifetimes, the new species formedupon acid addition would be hardly detectable by fluorescenceemission.
From laser flash photolysis experiments, a different behaviourof the transient at 410 nm, assigned to the triplet state, was noticedin MeOH and in ACN [18]. In MeOH, the triplet lifetime (5.9 �s)was approx. 30% shorter than in ACN (8.3 �s), but did not give riseto another transient (Fig. 6b). The triplet state was thus identifiedas the main transient species, efficiently quenched by oxygen, inboth solvents. A minor contribution of the semi-oxidized (MV•+,470 nm) and semi-reduced (MV•−, 380 nm) forms arising from thetriplet state was only noticed in ACN (Fig. 6a). Consistently, the
singlet oxygen quantum yield was found by both direct (emissionat 1270 nm) and indirect (chemical probes) methods to be as highas that of Methylene Blue (0.60 and 0.58 relative to phenalenone for
12 F. Ronzani et al. / Journal of Photochemistry and Photobiology A: Chemistry 284 (2014) 8–17
Fig. 6. Comparison between the transient absorption spectra obtained at the laser pulse end (diamonds, solid lines) and 1 �s after the pulse end (empty triangles, dashedlines, secondary axis) in (a) air-equilibrated ACN (5 × 10−6 M) and (b) air-equilibrated MeOH (2.5 × 10−5 M).
Fig. 7. Variation of the absorption spectra of MV in various solvents as a function of irradiation time, in the absence (left) and in the presence (right) of Bu2S. (a and b)ACN (irradiation: white light, 90 min (left) and 60 min (right), Pa = 5.1 × 10−6 Einsteins L−1 s−1); (c and d) dioxane (irradiation: white light, 40 min (left) and 20 min (right),Pa = 4.6 × 10−6 Einsteins L−1 s−1; (e and f) MeOH (irradiation: white light, 200 min, Pa = 4.2 × 10−6 Einsteins L−1 s−1).
and Photobiology A: Chemistry 284 (2014) 8–17 13
M(
3
ias5msiMaima(5S6Mwa(tATpeo
tpna
4
4
sMa(vdvMm
ab
•
•
Fig. 8. Relative energy of MV and various MVH+ and MVH22+ isomers in vacuum,
heptane, chloroform, methanol and acetonitrile, taking the most stable OH formof MVH+ and OH&N1H form of MVH2
2+ as zero in vacuum. The energy decreases
F. Ronzani et al. / Journal of Photochemistry
V and MB+ respectively) in ACN and higher in deuterated MeOH0.73 vs 0.52 respectively).
.3. Photoreactivity
The photosensitizing properties of MV in solution were testedn several solvents using Bu2S (Fig. 7) and rubrene (Fig. SI 6, ESI)s model reactants. As expected from the well-known reactivity ofulfides with 1O2, the sulfoxide (Bu2SO) and sulfone (Bu2SO2) in a0/50 ratio were only observed by 1H NMR in the crude reactionixture. No other products could be detected, meaning that pos-
ible side-products would be <2% in concentration. Surprisingly,n the presence of Bu2S, the absorption spectra of the irradiated
V solution significantly changed only in ACN and dioxane (Fig. 7bnd d respectively), but remained stable in MeOH (Fig. 7f), whilen the absence of Bu2S, only a slight or negligible decrease of the
ain absorption band was noticed in all the solvents (Fig. 7a, cnd e). The quantum yield of bleaching in the absence of sulfide�bleaching/10−5 molMV Einsteins−1) was estimated to 0.26 ± 0.03,.5 ± 0.8 and 17.0 ± 2.0 for MeOH, ACN and dioxane, respectively.uch values indicate that the photodegradation of MV was 20 and5 times more efficient in ACN and dioxane respectively than ineOH. On the other hand, the spectra observed after irradiationith Bu2S in ACN and dioxane were similar to the spectrum of
cidified MV in ACN (Fig. 4). By adding a suitable amount of a baseEt3N) to the reaction mixture, the reversibility of these latter spec-ral variations was observed, suggesting that by irradiating MV inCN or dioxane in the presence of Bu2S, some acid was formed.his peculiar reactivity, yielding some minor acid amounts in theresence of sulfides, was shown to be strictly photochemical (novolution of the mixture in the dark) and related to the presence ofxygen in solution (no evolution in N2-saturated solutions).
It should be noticed that during the MV-photosensitized oxida-ion of rubrene under the same experimental conditions as for thehotooxidation of Bu2S and even after its complete consumption,o formation of acid could be detected since no evolution of thebsorption spectrum of MV (in ACN) was observed (Fig. SI 6, ESI).
. Computational results
.1. Determination and optimization of the stable MV derivatives
In the following, quantum chemical calculations were used toupport and possibly account for the different protonated forms ofV suggested by the experimental spectroscopic results. First, MV
nd the different possible protonated forms of MVH+ and MVH22+
Scheme 1) were analyzed and optimized in vacuum and other sol-ents. Fig. 8 collects the energy of the optimized MV and protonatederivatives MVH+ and MVH2
2+ in vacuum and in solvents (energyalues in Table SI 2, ESI). The main geometrical features of MV,VH+ and MVH2
2+, as well as the charge distribution and dipoleoments are summarized in Table 2.Concerning MV, we first ruled out the Otsuki and Taguchi’s
ssumption [17] on a twisted isomer (90◦ rotation around the C Nond) of MV since (Fig. SI 7, ESI):
The C N2 bond length in MV was 1.37 A, similar to values typ-ical of sp2 C N bonds (1.34 A in pyridine, 1.37 A in pyrrole): ithad thus a partial sp2 character and it participated to the overallconjugation of MV,
The energetic barrier between the 0◦ and the 90◦ conformers wasestimated at approx. 50.2 kJ mol−1 when taking into account thenuclei’s relaxation and 71.1 kJ mol−1 for MV in vacuum and inMeOH, respectively,from vacuum to heptane (ε = 1.94), then chloroform (ε = 4.81) followed by methanol(ε = 37.2) and acetonitrile (ε = 37.5) whatever the derivatives.
• The 90◦ conformation represented an energetic maximum, inMeOH, relative to vacuum, the 0◦ conformer was more stabilizedthan the 90◦ one,
• The 90◦ conformer in vacuum had no significant electronic tran-sitions (absorption) at wavelengths higher than 406 nm.
For MVH+, whatever the solvent, the OH minimum was ener-getically favoured relative to the N1H and N2H minima (Scheme 1,Fig. 8). This most stable OH form of MVH+ was completely planar,with a double C N2 bond and a C O single bond shorter than inphenol (1.362 A) highlighting a strong conjugation on the wholemolecule. The charge distribution appeared surprisingly similar inboth MV and MVH+: the O atoms and the N1 atoms of the PTZskeleton carried important partial negative charges while the N2
atoms carried partial positive charges. The H atom on the hydroxylgroup carried about 45% of the positive charge of MVH+ while theremaining 55% was equally distributed on the whole molecule.
For the bi-protonated MVH22+, the OH&N1H form was ener-
getically more stable in vacuum and in all solvents, indicatingthis conformer as the most likely in case of a double-protonationreaction. The MVH2
2+ (OH&N1H form) was still planar with ashorter C N2 bond than in MV and MVH+ and a shorter C O bondthan in MVH+, indicating a stronger conjugation on the whole bi-protonated molecule. Two local O H and N1 H dipoles appearedwith two positively charged hydrogen atoms and strongly nega-
1
tively charged O and N .It is interesting to notice that the stabilization of MV, as well asMVH+ and MVH22+, increased with the dielectric constant of the
solvent, consistently with the polarity of the three forms, whose
14 F. Ronzani et al. / Journal of Photochemistry and Photobiology A: Chemistry 284 (2014) 8–17
S
N1
N2 O S
N1
N2 OH S
N1
N2 OH
H
MV MVH+ (OH form) MVH22+ (OH&N1H form)
S
N1
N2 O
H
S
N1
N2 OH
H+ H+
MVH+ (N1H form)
MVH+ (N2H form)
Scheme 1. Protonation of MV by acids and chemical structures of the three possible protonated MVH+ forms. The first protonation reaction occurs on the hydroxyl group(MVH+) and the second one on the N1 atom in the phenothiazine skeleton (MVH2
2+) preferably to the N2 of the dimethylamino group. See text.
Table 2Main features of the optimized geometries, charge distribution and dipole moments of MV and MVH+ and MVH2
2+ .
MV MVH+ MVH22+ MV MVH+ MVH2
2+
Bond Bond length/Å Atom Charge/eC N2 1.3744 1.3490 1.3364 N1< −0.665 −0.633 −0.308C O 1.2345 1.3397 1.3220 N2 +0.119 +0.159 +0.140C N1 1.3394 1.3355 1.3548 O −0.504 −0.494 −0.449N2 CH3 1.4682 1.4552 1.4777 S −0.064 +0.060 +0.135O H – 0.9723 0.9754 C (CH3) −0.285 −0.320 −0.329N1 H – – 1.0213 H (CH3) +0.110 +0.140 +0.16
H (OH) – +0.445 +0.466H (N1H) – – +0.354
Angle Angle size/◦ �/Debye2
d4
(l
4d
wMIdpof
TGt
H3C N C C 0.04 0.01 0.00
H O C C – 0.06 0.01H N1 C C – – 0.00
ipolar moments in vacuum were calculated to be 9.95, 2.43 and.44 D, respectively.
From now and in the following, the neutral, the monoprotonatedOH) and the biprotonated (OH&N1H) forms of MV are respectivelyabelled MV, MVH+ and MVH2
2+.
.2. Determination of the protonation affinity of MV and itserivatives
Very high Gibbs free energy of protonation of MV and MVH+
ere calculated (Table 3) demonstrating that both MVH+ andVH2
2+ could easily be formed in acidic medium or protic solvent.ndeed, for MV, it was found between −1124 and −1146 kJ mol−1
epending on the solvent, which was higher than the energy ofrotonation of aniline (−882.8 kJ mol−1 experimentally) [30,31]r of adenosine (calculated at −950 kJ mol−1) [32]. The Gibbsree energy of protonation of MVH+ to MVH2
2+ was still high,
able 3ibbs free energy (kJ mol−1) of the mono- and bi-protonation of MV according to
he following reactions: MV + H+ → MVH+ and MVH+ + H+ → MVH22+.
Vacuum CHCl3 MeOH ACN Water
�G0m,298(MV) −1022.4 −1124.3 −1145.8 −1146.1 −1147.9
�G0m,298(MVH+) −622.2 −981.2 −1065.1 −1066.4 −1073.8
9.95 2.43 4.44
although 60–80 kJ mol−1 lower. Therefore, the formation of MVH+
and MVH22+ may be assumed.
4.3. Computed electronic transitions of MV derivatives
In the next stage, we applied TDDFT to determine the singletand triplet electronic states of the three most stable MV derivatives(MV, MVH+ and MVH2
2+). The allowed singlet transitions beyond400 nm of MV derivatives in vacuum and in solvents (heptane,chloroform, acetonitrile and methanol) are collected in Table 4.Whatever the conditions, MV, MVH+ and MVH2
2+ had only twosinglet absorption bands, with oscillator strengths ranging from0.43 to 0.58. For these three species, quite similar amplitude of theabsorption bands was thus expected in the experimental spectra.Taking into account the coarse model used to mimic the sol-vent (PCM model), which can reasonably induce errors of at most40–50 nm, these calculations therefore suggested that the broadband observed in the absorption spectra of MV in chloroform or inacetonitrile could be assigned to the two electronic transitions ofMV presented in Table 4. In ACN for instance, the calculated MVS0–S2 and S0–S3 transitions at 546 nm and 515 nm respectively
could be assigned to the shoulder and the maximum observedexperimentally at 585 nm and 552 nm. Fig. 9 shows singlet andtriplet electronic states of MV, MVH+ and MVH22+ in ACN fromTDDFT calculations with the characteristic electronic transitions
F. Ronzani et al. / Journal of Photochemistry and Photobiology A: Chemistry 284 (2014) 8–17 15
Table 4Singlet electronic transitions: wavelength (nm) and oscillator strength (between brackets) of MV derivatives in vacuum, n-heptane, chloroform, acetonitrile and methanol.
Vacuum n-Heptane CHCl3 ACN MeOH
MV 508 (0.12) 524 (0.58) 538 (0.55) 546 (0.47) 545 (0.46)492 (0.32) 512 (0.01) 516 (0.06) 515 (0.13) 515 (0.14)
MVH+ 542 (0.09) 555 (0.19) 561 (0.22) 562 (0.20) 561 (0.20)
bi(r
5
snsgymettbmo
5c
fmdabiapMtb
Fami
486 (0.41) 500 (0.46)
MVH22+ 578 (0.41) 598 (0.53)
421 (0.04) 421 (0.05)
etween the ground state and the bold singlet excited states. Fornstance, in acetonitrile, MVH+ gave allowed transitions at 562 nmS0–S1) and 498 nm (S0–S2), while MVH2
2+ gave two more sepa-ated transitions at 583 nm (S0–S2) and 419 nm (S0–S3).
. Discussion
It emerged from our experimental results that the electronicpectra and reactivity of MV were very sensitive to the protic oron-protic character of the solvent (Fig. 2) and showed a reversiblepectral modification upon successive addition of acid and base. Inood agreement with the high measured singlet oxygen quantumield, irradiation of MV in the presence of Bu2S led to the main for-ation of sulfoxide and sulfone. However, in aprotic solvents, the
volution of the absorption spectrum during irradiation showedhat a side photochemical mechanism induced the formation ofrace amounts of acid. This result was correlated with the possi-le side-formation from the triplet state of MV, identified as theain transient species, of the semi-oxidized form of MV, MV•+,
nly observed in ACN but not in MeOH.
.1. Attribution of the absorption spectra of MV under variousonditions and proposition of assumed mechanisms
We first address the attribution of the absorption spectra withour well defined bands at 467, 552, 576 and 611 nm in the experi-
ental absorption spectra of MV in acidified ACN (Fig. 4) or two wellefined bands at 575 and 611 nm, with less defined shoulders at 545nd 480 nm in acidified CHCl3 (Fig. 3) for the same amount of acid. Inoth cases, the presence of several isosbestic points observed with
ncreasing [H+] suggested the presence of only two independentbsorbing species in solution. The calculation results strongly sup-
+
orted the formation of mono- or bi-protonated species, MVH andVH22+ on the basis of the calculated free Gibbs energy, althoughhe second protonation step was less favoured than the first oney 60–80 kJ mol−1 (Table 3). However, whatever the solvent, no
ig. 9. Schematic representation of singlet and triplet electronic states of MV, MVH+
nd MVH22+ in ACN calculated by TDDFT at the B3LYP/cc-pVTZ level with the PCM
odel, using Gaussian 09. Bold states are relevant singlet states which could benvolved in the absorption spectra.
501 (0.45) 498 (0.43) 497 (0.43)594 (0.54) 583 (0.51) 582 (0.51)420 (0.05) 419 (0.05) 419 (0.05)
more than two allowed transitions were calculated for each sep-arated species, MV, MVH+ and MVH2
2+ (Table 4) and the splittingbetween the two main allowed transitions increased from MV toMVH+ and MVH2
2+ (Fig. 9). The spectrum observed in acidifiedCHCl3 may be consistently assigned to MVH+. On the contrary, noneof these species alone could account for the four bands experimen-tally observed for the solute in acidic ACN. Tentatively, the twosplit transitions of MVH2
2+ could be respectively assigned to theextreme experimental bands (467 and 611 nm) and the two closertransitions of MVH+ to the intermediate bands (552 and 576 nm)of the absorption spectrum. Under our conditions, to account forthe presence of isosbectic points in the absorbance spectra withincreasing amounts of acid, a thermodynamic equilibrium betweentwo species (MVH+ and MVH2
2+) in ratio independent of [H+] mightbe proposed. Due to the absence of any relative variation of the con-centrations of the absorbing species, this hypothesis could not beverified by spectral analysis.
Alternatively, the formation of a different species characterizedby four absorption bands has to be assumed. Actually, methy-lene blue (MB+) and many other PTZ dyes are known to formdimers relatively easily: many examples have been reported inthe literature and the effect of several factors such as the ionicstrength, the counter-ions, surfactants and other additives has beenstudied by optical and NMR spectroscopy. In aqueous solutions,thionine and phenothiazines are often described to form dimersin H-configuration, even at low concentrations, giving rise to twoseparate blue- (H-band) and red-shifted (J-band) absorption bands[33–38].
It was therefore tempting to assume the formation of a dimerof MV derivatives. Since the peculiar spectroscopic behaviour inACN was limited to the presence of free H+ ions, we focusedon the possibility of dimers of MVH+, for which a very diffusespatial charge of +0.65 e on the PTZ skeleton and +0.45 e welllocalized on the hydroxyl hydrogen was calculated. It was thusreasonable to consider the formation of MVH+· · ·MVH+ dimers inacidified ACN: would the presence of such a dimer explain thefour bands absorption spectra recorded in this solvent? Accord-ing to the exciton theory [39], either for H- or J-dimers, the signalof each separated monomer would not split into bands, since onlyone transition is allowed. Another possibility would be that thedimer adopts a rather uncommon oblique configuration, charac-terized by a band-splitting inducing the presence of two allowedtransitions corresponding to every monomer transition. Accord-ing to this possibility, we could easily explain the fact that twoof the four observed bands were in the same position as in themonomer spectrum, whereas the others were bathochromicallyand hypsochromically shifted relative to the starting spectrum. Foroblique configuration to occur, hydrogen-bonding dimers must beformed, in order to fix the transition moment axes in such an ori-entation. Based on this assumption, such a dimer is proposed anddiscussed in Fig. SI 9, ESI.
To summarize, in pure CHCl3 and ACN, the neutral form of MV
had identical absorption spectra but depending on the solvent, thedye presented different sensitivity towards acid addition. In CHCl3,the formation of MVH+ remains consistent with all the availableexperimental and numerical data. In protic and acidified solvents,
1 and P
tabbbhoc
5
tgqmct
mofbttar(a(ttsareabm(c
M
[
[
M
M
M
O
B
B
rtub(tw
6 F. Ronzani et al. / Journal of Photochemistry
he formation of a dimer ((MVH+)2) was tentatively proposed toccount for the complex behaviour of MV. It may neverthelesse concluded that MV is highly susceptible to inter-molecular H-onding and to protonation. It has been reported that MV peculiarlyinds to DNA [14,15]: the formation of strong H-bonds between theydroxyl substituent of MVH+ (which should be formed in physi-logical media) and the several available lone pairs of nucleobasesould be a possible explanation of such a reactivity.
.2. Photosensitization mechanisms by MV
As already highlighted, irradiation of MV solutions of Bu2S, ledo the main formation of sulfoxide and sulfone after singlet oxy-en addition on sulfide, in agreement with the high singlet oxygenuantum yield found for MV. However, in dioxane and ACN, the for-ation of trace amounts of acidic products was deduced from the
hange in the MV absorption spectra. What mechanism, in additiono singlet oxygen production, could yield an acid?
In agreement with recent results from Manju et al. [40] on 10-ethylphenothiazine, and with the detection of MV•+ and MV•−
nly in non-protic solvents like ACN, in addition to singlet oxygenormation, the charge transfer (CT) complex [MV•+· · ·O2
•−] formedetween excited MV and O2 (Eq. (5)) could evolve mainly by energyransfer (ET) to 1O2 generation (Eq. (6)), and by a secondary electronransfer (ELT) reaction, yielding the formation of MV radical cationnd superoxide radical anion (Eq. (7)). Once formed, MV•+ couldeact with Bu2S to yield Bu2S•+ and MV in its ground state (Reaction8)). Alternatively, a first electron transfer reaction between MVnd the substrate could lead to the formation of Bu2S•+ and MV•−
Reaction (9)). MV•− could then produce superoxide radical anionhrough another ELT reaction (Reaction (10)). A radical chain reac-ion between O2
•− and Bu2S•+, involving H-abstraction from theulfide, could thus lead to the formation of hydroperoxyl radicalsnd H+ ions (Reactions (11) and (12)). The semi-dehydrogenatedadical of Bu2S (Bu2( H)S•), in the presence of oxygen, could alsovolve towards the formation of the corresponding sulfonic acidnd of sulfuric acid (Eq. (13)). These hypothetical pathways coulde responsible of the acidification of the medium. Considered theuch higher concentration of Bu2S (5 × 10−3 M) relative to MV
10−4 M), even at the side-process scale, these radical reactionsould produce enough H+ to justify the spectral changes observed.
V(S0)h�+ISC−→ MV∗(T1)
O2−→[MV∗(T1)· · ·O2] ↔ [MVı+· · ·Oı−2 ]CT (5)
CT]ET−→MV + 1O2
Bu2S−→Bu2SO + Bu2SO2 (6)
CT]ELT−→[MV•+· · ·O2
•−] → MV•+ + O2•− (7)
V•+ + Bu2S → MV + Bu2S•+ (8)
V∗(T1) + Bu2SELT−→MV•− + Bu2S•+ (9)
V•− + O2 → MV + O2•− (10)
2•− + Bu2S•+ → HOO• + Bu2( H)S• (11)
u2S•+ → H+ + Bu2( H)S• (12)
u2( H)S• → O2−→ → BuSO3H/H2SO4 (13)
Sulfides represent particular reactants, since they can easilyeact via both singlet oxygen addition and electron transfer reac-ions [41,42]. During the MV-photosensitized oxidation of rubrene,nder the same experimental conditions, no formation of acid could
e detected since no evolution of the absorption spectrum of MVin ACN) was noticed (Fig. SI 6, ESI). This result is consistent withhe known reactivity of rubrene, which selectively reacts with 1O2ithout undergoing any ELT reaction [43].hotobiology A: Chemistry 284 (2014) 8–17
In MeOH, since MV is most probably protonated in MVH+,the ELT mechanism proposed for ACN and dioxane cannot occur,accounting for the sole detection of triplet MV without any evi-dence for the radical-cation MV•+ or the radical-anion MV•−.Accordingly, higher quantum yield of singlet oxygen productionwas measured in MeOH than in ACN (0.73 vs 0.60) [18]. We canthus conclude that in protic solvents such as MeOH, i.e. in the pres-ence of MVH+, MV cannot undergo any electron transfer reactions,and that its photoreactivity in such media is limited to singlet oxy-gen production. Possibly, the positive charges on the protonatedform of MV, as well as the presence of H-bonds between the pho-tosensitizer and the solvent, prevent the evolution of the chargetransfer complex towards the charge separation and the formationof the two radical ions.
6. Conclusions
The photophysical and photochemical properties of the poorlyinvestigated methylene violet (MV), a possible candidate for appli-cations in PDT and PDI due to its very good photosensitizingproperties, were analyzed. A shift of the absorption maximum bymore than 100 nm from non-polar to polar and protic solvents,as well as a strong influence of acid addition on MV absorptionspectra and fluorescence properties were confirmed. Significantsolvent effects were also evidenced on the transient species pro-duced upon irradiation, as well as on the rather high quantumyields of singlet oxygen production (0.60 in ACN and 0.73 in deuter-ated MeOH). Computational data on the different neutral andprotonated derivatives of MV supported the analysis of the spec-troscopic properties of MV in different media: in aprotic (polarand non-polar) solvents, MV is the dominant species and its elec-tronic transitions depend on the solvent polarity. In protic media(alcohols and water for instance) and in acidified chloroform, themono-protonated form of the dye, MVH+, is formed. The forma-tion of a second species in acidified ACN, tentatively assigned to adimer of MVH+, also appeared from the analysis of the absorptionspectra.
Clearly, the presence of such protonated species influences thephotochemical properties of this dye. Whatever the nature of theprotonated species formed in the presence of organic acids, we con-clude that MV, independently on the solvent, efficiently undergoesintersystem crossing to its triplet excited state. From the tripletexcited state, energy transfer (ET) to molecular oxygen for singletoxygen generation efficiently takes place. The triplet transient statecan nevertheless evolve by different pathways, depending on thesolvent. In aprotic solvents, where only MV is present, electrontransfer (ELT) can occur, inducing a photoinduced radical chainreaction as a side-process. When such an ELT process from thereactant is not possible, as with rubrene for instance, ET to sin-glet oxygen production is only observed. In protic solvents, on thecontrary, singlet oxygen formation via energy transfer seems tobe the only possible photochemical pathway, probably due to thepresence of the protonated form of the PS issued from the strongH-bonding interaction with the solvent, which obviously preventsELT. It was also observed that in MeOH, MV photobleaching isreduced due to the presence of MVH+. Since, in this solvent, thetriplet excited state of MV was more efficiently quenched by O2leading to a more efficient singlet oxygen production than in ACN,MeOH as well as other protic solvents may be more suitable for anefficient use of this dye. In any case, for any application, the effectof the environment and the substrate on the complex photophys-
ical/photochemical properties of MV has to be carefully analyzed.Due to the strongly conjugated keto group on the aromatic skele-ton, MV displays a specific and complex behaviour among thePTZ dye family. Thanks to the high sensitivity of its absorption
and P
sa
A
M(rPa
A
i2
R
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
F. Ronzani et al. / Journal of Photochemistry
pectrum to the addition of acids, MV could act both as a PS ands a sensitive acid probe under carefully controlled conditions.
cknowledgments
This work has been supported by ANR (10-BLANC-0803), byESR (PhD grants for FR) and by Conseil Régional d’Aquitaine
funding). Simulations in this work were performed at the MCIAegional supercomputing facility, Bordeaux 1 University, and theCSTD, Université de Pau et Pays de l’Adour. They are gratefullycknowledged.
ppendix A. Supplementary data
Supplementary data associated with this article can be found,n the online version, at http://dx.doi.org/10.1016/j.jphotochem.014.03.019.
eferences
[1] A. Mills, D. Hazafy, J.A. Parkinson, T. Tuttle, M.G. Hutchings, Comment on solventeffect on the electronic spectra of azine dyes under alkaline condition, J. Phys.Chem. A 113 (2009) 9575–9576.
[2] A. Mills, D. Hazafy, J. Parkinson, T. Tuttle, M.G. Hutchings, Effect of alkali onmethylene blue (C.I. Basic Blue 9) and other thiazine dyes, Dyes Pigm. 88 (2011)149–155.
[3] F. Wilkinson, W.P. Helman, A.B. Ross, Quantum yields for the photosensitizedformation of the lowest electronically excited singlet state of molecular oxygenin solution, J. Phys. Chem. Ref. Data 22 (1993) 113–262.
[4] W.M. Sharman, C.M. Allen, J.E. van Lier, Photodynamic therapeutics: basic prin-ciples and clinical applications, Drug Discov. Today 4 (1999) 507–517.
[5] M. Wainwright, Methylene blue derivatives–suitable photoantimicrobials forblood product disinfection? Int. J. Antimicrob. Agents 16 (2000) 381–394.
[6] M. Wainwright, The emerging chemistry of blood product disinfection, Chem.Soc. Rev. 31 (2002) 128–136.
[7] M. Wainwright, R.M. Giddens, Phenothiazinium photosensitisers: choices insynthesis and application, Dyes Pigm. 57 (2003) 245–257.
[8] M. Wainwright, Photoinactivation of viruses, Photochem. Photobiol. Sci. 3(2004) 406.
[9] M. Wainwright, M.N. Byrne, M.A. Gattrell, Phenothiazinium-based photobac-tericidal materials, J. Photochem. Photobiol. B 84 (2006) 227–230.
10] M. Wainwright, H. Mohr, W.H. Walker, Phenothiazinium derivatives forpathogen inactivation in blood products, J. Photochem. Photobiol. B 86 (2007)45–58.
11] J.P.M.L. Rolim, M.A.S. de-Melo, S.F. Guedes, F.B. Albuquerque-Filho, J.R. de Souza,N.A.P. Nogueira, et al., The antimicrobial activity of photodynamic therapyagainst Streptococcus mutans using different photosensitizers, J. Photochem.Photobiol. B 106 (2012) 40–46.
12] T. Ben Amor, G. Jori, Sunlight-activated insecticides: historical backgroundand mechanisms of phototoxic activity, Insect Biochem. Mol. Biol. 30 (2000)915–925.
13] A. Skripchenko, D. Robinette, S.J. Wagner, Comparison of methylene blue andmethylene violet for photoinactivation of intracellular and extracellular virusin red cell suspensions, Photochem. Photobiol. 65 (1997) 451–455.
14] H. Morrison, T. Mohammad, R. Kurukulasuriya, Photobiological properties ofmethylene violet, Photochem. Photobiol. 66 (1997) 245–252.
15] T. Mohammad, H. Morrison, Simultaneous determination of methylene vio-let, halogenated methylene violet and their photoproducts in the presenceof DNA by high-performance liquid chromatography using an internal sur-face reversed-phase column, J. Chromatogr. B: Biomed. Sci. Appl. 704 (1997)265–275.
16] M.A. Houghtaling, R. Perera, K.E. Owen, S. Wagner, R.J. Kuhn, H. Morrison,Photobiological properties of positively charged methylene violet analogs, Pho-tochem. Photobiol. 71 (2000) 20–28.
17] S. Otsuki, T. Taguchi, Solvent-dependent photophysical properties of a phe-nothiazone dye as an optical probe, Bull. Chem. Soc. Jpn. 69 (1996) 2525–2531.
[
hotobiology A: Chemistry 284 (2014) 8–17 17
18] F. Ronzani, A. Trivella, E. Arzoumanian, S. Blanc, M. Sarakha, C. Richard,et al., Comparison of the photophysical properties of three phenothiazinederivatives: transient detection and singlet oxygen production, Photochem.Photobiol. Sci. 12 (2013) 2160–2169.
19] H. Gampp, M. Maeder, C.J. Meyer, A.D. Zuberbuehler, Evolving factor analysis,Comments Inorg. Chem. 6 (1987) 41–60.
20] P.A. Hoijemberg, A. Chemtob, C. Croutxé-Barghorn, J. Poly, A.M. Braun, Rad-ical photopolymerization in miniemulsions. Fundamental investigations andtechnical development, Macromolecules 44 (2011) 8727–8738.
21] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheese-man, et al., Gaussian 09, Revision A.1, Gaussian, Inc., Wallingford, CT,2009.
22] S. Miertus, E. Scrocco, J. Tomasi, Electrostatic interaction of a solute with a con-tinuum – a direct utilization of abinitio molecular potentials for the previsionof solvent effects, Chem. Phys. 55 (1981) 117–129.
23] K. Dimroth, C. Reichardt, T. Siepmann, F. Bohlmann, Über Pyridinium-N-phenol-betaine und ihre Verwendung zur Charakterisierung der Polarität vonLösungsmitteln, Justus Liebigs Ann. Chem. 661 (1963) 1–37.
24] C. Reichardt, Empirical parameters of the polarity of solvents, Angew. Chem.Int. Ed. Engl. 4 (1965) 29–40.
25] Y. Marcus, The properties of organic liquids that are relevant to their use assolvating solvents, Chem. Soc. Rev. 22 (1993) 409–416.
26] R.W. Taft, M.J. Kamlet, The solvatochromic comparison method. 2. The alpha-scale of solvent hydrogen-bond donor (HBD) acidities, J. Am. Chem. Soc. 98(1976) 2886–2894.
27] M.J. Kamlet, R.W. Taft, The solvatochromic comparison method. I. The beta-scale of solvent hydrogen-bond acceptor (HBA) basicities, J. Am. Chem. Soc. 98(1976) 377–383.
28] M.J. Kamlet, J.L. Abboud, R.W. Taft, The solvatochromic comparison method. 6.The pi* scale of solvent polarities, J. Am. Chem. Soc. 99 (1977) 6027–6038.
29] J. Olmsted, Calorimetric determinations of absolute fluorescence quantumyields, J. Phys. Chem. 83 (1979) 2581–2584.
30] Y.K. Lau, K. Nishizawa, A. Tse, R.S. Brown, P. Kebarle, Protonation and site ofprotonation of anilines. Hydration and site of protonation after hydration, J.Am. Chem. Soc. 103 (1981) 6291–6295.
31] E.P.L. Hunter, S.G. Lias, Evaluated gas phase basicities and proton affinities ofmolecules: an update, J. Phys. Chem. Ref. Data 27 (1998) 413–656.
32] D. Touboul, G. Bouchoux, R. Zenobi, Gas-phase protonation thermochemistryof adenosine, J. Phys. Chem. B 112 (2008) 11716–11725.
33] D. Hazafy, M.-V. Salvia, A. Mills, M.G. Hutchings, M.P. Evstigneev, J.A. Parkinson,NMR analysis of Nile Blue (C. I. Basic Blue 12) and Thionine (C. I. 52000) insolution, Dyes Pigm. 88 (2011) 315–325.
34] W.C. Lai, N.S. Dixit, R.A. Mackay, Formation of H aggregates of thionine dye inwater, J. Phys. Chem. 88 (1984) 5364–5368.
35] K. Patil, R. Pawar, P. Talap, Self-aggregation of Methylene Blue in aqueousmedium and aqueous solutions of Bu4NBr and urea, Phys. Chem. Chem. Phys.2 (2000) 4313–4317.
36] O. Yazdani, M. Irandoust, J.B. Ghasemi, S. Hooshmand, Thermodynamic study ofthe dimerization equilibrium of methylene blue, methylene green and thiazoleorange at various surfactant concentrations and different ionic strengths andin mixed solvents by spectral titration and chemometric analysis, Dyes Pigm.92 (2012) 1031–1041.
37] D. Heger, J. Jirkovsky, P. Klan, Aggregation of methylene blue in frozen aque-ous solutions studied by absorption spectroscopy, J. Phys. Chem. A 109 (2005)6702–6709.
38] A. Ghanadzadeh Gilani, T. Ghorbanpour, M. Salmanpour, Additive effect on thedimer formation of thiazine dyes, J. Mol. Liq. 177 (2013) 273–282.
39] M. Kasha, H.R. Rawls, M. Ashraf El-Bayoumi, The exciton model in molecularspectroscopy, Pure Appl. Chem. 11 (1965) 371–392.
40] T. Manju, N. Manoj, A.M. Braun, E. Oliveros, Self sensitized photooxidation ofN-methyl phenothiazine: acidity control of the competition between electronand energy transfer mechanisms, Photochem. Photobiol. Sci. 11 (2012) 1744–1755.
41] E.L. Clennan, New mechanistic and synthetic aspects of singlet oxygen chem-istry, Tetrahedron 56 (2000) 9151–9179.
42] I.N. Lykakis, S. Lestakis, M. Orfanopoulos, 9,10-Dicyanoanthracene photosensi-tized oxidation of aryl alkanols: evidence for an electron transfer mechanism,
Tetrahedron Lett. 44 (2003) 6247–6251.43] F. Ronzani, E. Arzoumanian, S. Blanc, P. Bordat, T. Pigot, C. Cugnet, et al., Efficientcyanoaromatic photosensitizers for singlet oxygen production: synthesis andcharacterization of the transient reactive species, Phys. Chem. Chem. Phys. 15(2013) 17219–17232.
