Embed Size (px)
Citation preview
ORIGINAL PAPER
Improvement of photovoltaic performance by substituent effectof donor and acceptor structure of TPA-based dye-sensitizedsolar cells
Natalia Inostroza1 & Fernando Mendizabal2,5 & Ramiro Arratia-Pérez3,5 &
Carlos Orellana2,4 & Cristian Linares-Flores1
Received: 28 July 2015 /Accepted: 18 December 2015 /Published online: 7 January 2016# Springer-Verlag Berlin Heidelberg 2016
Abstract We report a computational study of a series of or-ganic dyes built with triphenylamine (TPA) as an electrondonor group. We designed a set of six dyes called (TPA-n,where n=0–5). In order to enhance the electron-injection pro-cess, the electron-donor effect of some specific substituentwas studied. Thus, we gave insights into the rational designof organic TPA-based chromophores for use in dye-sensitizedsolar cells (DSSCs). In addition, we report the HOMO,LUMO, the calculated excited state oxidized potentialEdye*(eV) and the free energy change for electron-injectionΔGinject(eV), and the UV-visible absorption bands for TPA-ndyes by a time-dependent density functional theory (TDDFT)procedure at the B3LYP and CAM-B3LYP levels with solventeffect. The results demonstrate that the introduction of theelectron-acceptor groups produces an intramolecular charge
transfer showing a shift of the absorption wavelengths ofTPA-n under studies.
Keywords Density functional theory . Electronic absorptionspectra . Molecular design . Organic dye-sensitized solar cells
Introduction
Solar energy will be an important sustainable energy source inthe near future, even though a variety of energy sources haveto be explored at least until the technique of solar-energy cap-turing devices are mature enough to replace the older energysources [1]. One of the more successful and promising ap-proaches is dye-sensitized solar cells (DSSCs), proposed byGrätzel [2, 3]. One of the most common problems regardingDSSCs is related with the efficiency of the dye-sensitizedsolar cells. To improve this latter, it is necessary to designsuitable molecular chromophores for the light capturing pro-cess. These systems are based on the adsorption of inorganicand organic molecules, usually called dyes, on nanocrystallinetitanium dioxide films [2–6]. In the last decades, rutheniumcomplexes [7], triarylamines [8], squarines [9], thiophenes[10 , 11 ] , po rphyr in s [12 , 13 ] and π - ex t endedtetrathiafulvalenes (exTTFs) [14] have been studied. On theother hand, inorganic compounds based on ruthenium com-plexes have a high efficiency, but a high monetary cost [15].
An important branch recently developed, which hasattracted much interest in recent years, because it offers sev-eral advantages such as a lower cost large-scale production,environmentally friendly, and high molar extinction coeffi-cients [16–18], are the organic DSSCs. An increment on theconversion efficiency is also expected to compete with thecurrent silicon based photovoltaic cells. As we mentionedabove, the general DSSCs consists of semiconductor material
* Fernando [email protected]
1 Inorganic Chemistry and Molecular Material center and Theoreticaland computational chemistry center, Facultad de IngenierÃa,Universidad Autonoma de Chile, El Llano Subercaseaux 2801, SanMiguel, Santiago, Chile
2 Departmento de QuÃmica, Facultad de Ciencias, Universidad deChile, P.O. Box 653, Las Palmeras 3425, Ñuñoa, Santiago, Chile
3 Doctorado en FisicoquÃmica Molecular, Relativistic MolecularPhysics (ReMoPh) Group, Universidad Andrés Bello, República275, Santiago, Chile
4 Departamento de QuÃmica, Facultad de Ciencias Básicas,Universidad Metropolitana de Ciencias de la Educación, Casilla 147,Santiago, Chile
5 Núcleo Milenio de IngenierÃa Molecular para Catálisis yBiosensores, ICM, Santiago, Chile
J Mol Model (2016) 22: 25DOI 10.1007/s00894-015-2893-9
as photoelectrode, a sensitizer, an electrolyte and catalyst be-tween two conductive transparent electrodes. The power con-version efficiency and stability of the cells are directly corre-lated with the nature of the organic photosensitizer, playing akey role in DSSC. Titanium dioxide (TiO2) is the most com-mon semiconductor used as activematerial due to high surfacearea, large band gap, good optics and electrical properties,nontoxic, and low cost [19–23]. The dyes have to absorbsunlight, to inject its electron toward the conductive bandsof TiO2 and receive an electron from the electrolyte. Thus,the optimization of dyes has resulted in the development ofnew compounds to improve the absorption, morphology, andstability.
In general, these systems feature a donor (D) andacceptor (A) systems separated by a pi-conjugatedbridge (D-π-A). For example, triphenylamine have beencommonly utilized as the electron donor (D) [24, 25].The carboxylic acid used to be an electron acceptor (A).Under this kind of systems, carboxilic acid will act asanchoring unit in the (D-π-A) structure. Various pi-conjugated bridges can be employed, however, we focuson one optimized π-brigde. Considering this fact, in thiswork, we focus on a series of organic donor–spacer–acceptor molecules involving triphenylamine (TPA).They are called TPA-n, where n is indicated 0, 1, 2,3, 4, and 5 new DSSCs (see Fig. 1). We designed aset of new organic dye sensitizer where the TPA is adonor group, phenylene as bridge (typically a π spacer),and cyanoacetic acid as acceptor/anchoring end group[26–31]. The expecting process involving organicDSSCs, have to guide light absorption regions of theDSSCs and subsequently the scale of the electron injec-tion from the excited state of the dyes to the semicon-ductor surface. Under this context, theoretical chemistryprovides a deep understanding about the behavior ofthese compounds. Thus, one aim of this work is tomade models that can optimize the DSSCs performancesthrough a rational molecular design [32].
Preat and co-worker in theoretical work [24, 25] calculated20 TPA-derivatives. They pointed out the relevance of thebridge part testing different types of it. Furthermore, aftersome modification to the usual structures, they proposed anoptimal structure which showed a better photovoltaic perfor-mance. Considering the main results of this above mentionedresearch, we selected their best structure as the starting pointof this current work. We fixed the donor group. The bridgeand acceptor group as well and considering the Hammentsubstituents effect we improve the performance inside a newset of TPA-n derivatives.
To improve the photovoltaic performance of somenew structures, a fundamental aspect has to be consid-ered: it is expected that the lowest unoccupied molecu-lar orbital (LUMO) has to be higher in energy than theconduction band (CB) edge of the semiconductor(TiO2). To obtain a better charge transfer (CT), the elec-tron mobility (conjugation) through the donor and an-choring group will be a key aspect. Due to this, weadded into the donor TPA-ring and acceptor-ring a setof specific substituents (-OCH3 and -NH2) to analyzethe effect of this modification into the electron-injection process for every TPA-n included in this work.
Models and computational details
To analyze the electronic structure and optical propertiesof our organic dyes sensitizers, density funtional theory(DFT) and time-dependent perturbation theory approach(TD-DFT) calculations were conducted. The TPA-nmodel, (where n is indicating 0–5 new DSSCs) usedin this study is shown in Fig. 1. In Table 1 we showthe dyes structures in detail included in this work. TheB3LYP and CAM-B3LYP functionals have providedbetter results than other functionals and have been wide-ly used in similar works [5, 6, 24]. Thus, geometry ofTPA-n dyes in the ground state were fully optimized at
R5
R4
R3
R1
R2
Fig . 1 Genera l scheme of TPA-n (n = 0-5) dyes . n = 0,R1 = R2 = R3 = R4 = R5 = -H; n = 1, R1 = R2 =R5 = -CN, R3 = R4 = -OCH3; n = 2, R1 = R2 = -H,R3 = R4 = -OCH3,R5 = -CN; n = 3,R1 = R2 = R3 = R4 = -NH2, R5 = -CN; n = 4, R1 = R2 = -NH2,R3=R4= -OCH3,R5 = -CN; n= 5, R1=R2=R5= -CN, R3=R4= -NH2
Table 1 Organic dyes TPA-n (n= 0–5) and electronic properties (ineV) estimated at the B3LYP/aug-cc-pVTZ with solvent effect
B3LYP TPA-0 TPA-1 TPA-2 TPA-3 TPA-4 TPA-5
R1 -H -CN -H -NH2 -NH2 -CN
R2 -H -CN -H -NH2 -NH2 -CN
R3 -H -OCH3 -OCH3 -NH2 -OCH3 -NH2
R4 -H -OCH3 -OCH3 -NH2 -OCH3 -NH2
R5 -CN -CN -CN -CN -CN -CN
HOMO −4.93 −5.26 −4.92 −4.63 −4.67 −5.21LUMO −2.69 −2.45 −2.39 −2.48 −2.37 −2.53Gap 2.24 2.81 2.53 2.15 2.30 2.68
25 Page 2 of 7 J Mol Model (2016) 22: 25
the B3LYP level in the gas phase. The excitation ener-gies were obtained by means of the TD-DFT [33, 34].The excitation spectra have been simulated from theoptimized geometry of the theoretical models. All UV/Vis spectra showed in the current discussed organicdyes were reflected with Gaussian curves of full-widthat half-maximun (FWHM) of 1500 cm−1 [5, 6].Literature reports that TD-DFT excitation energies arein general within a 0.4 eV deviation from experiment,the average errors are frequently smaller than this upperlimit.
All calculations have been carried out with the Turbomole(version 6.5) and Gaussian09 programs [35, 36]. The aug-mented correlation-consistent valence-triple-zeta (aug-cc-pVTZ) basis sets were used for C, N, O, and H. For compar-ison purpose, we employed the polarizable continuum model
(PCM) to include the dichloromethane solvent over TPA-n.These solvent calculations were carried out with the Cosmosprogram [37].
Results and discussion
Structures of TPA-n dyes and the effect of chemicalmodifications
We propose structural modifications that may improve theelectron injection efficiency of the TPA-based DSSCs. In thiswork, all dyes possess a terminal -COOH group on the accep-tor unit. The carboxylic group is the link between the dye andthe semiconductor (TiO2). As was emphasized earlier, TPA isthe donor part of our systems and we selected the better-
Tpa-1 Solution phase
HOMO
LUMO
Gas phaseFig. 2 Schematic representationof the frontier molecular orbitalenergies of the TPA-1 in solutionand gas phases at the B3LYP/aug-cc-pVTZ
TPA-2
TPA-3
TPA-4
TPA-1
Optimized structure
TPA-5
HOMO LUMO
TPA-0
Fig. 3 Optimized structure andfrontier molecular orbitals(HOMO and LUMO) of TPA-norganic dyes at the B3LYP/aug-cc-pVTZ with solvent effect
J Mol Model (2016) 22: 25 Page 3 of 7 25
reported bridge [24, 25]. Table 1 shows the results for a set ofsix compounds for studies in a solvent phase at the B3LYPlevel. Optimization geometries were done at the B3LYP andCAM-B3LYP levels. We defined a general skeleton accordingto the Fig. 1. There are no differences in the geometric param-eters to both levels of theory (B3LYP and CAM-B3LYP). Weincorporated a couple of substituents at specific positioncalled R1, R2, R3, R4 distributed over the three ring (referto Table 1). The R5 position in all structures was fixed.
The HOMO-LUMO gap between these structures changedwhen the substituents over donor/acceptor groups are modi-fied. It is observed that TPA-0 which have -H in all R1-R4position may show a good performance when we take as ref-erence electrode of experimental TiO2 (anatase) [38]. This isdue to if the lowest unoccupied molecular orbital (LUMO) ishigher in energy than the conduction band (BC) edge of thesemiconductor, since it will allow the charge injection fromthe excited state of the dye to the BC of the TiO2. BC of theTiO2 is between −4.0 and −3.5 eV [38]. It is easily noted thatthe substitution effect is the main cause of the HOMO-LUMOgap displacement observed for these six new DSSCs.
In order to demonstrate that the solvent effect produces thesame effect as in vacuo, we have used the TPA-1 dye as amodel at the B3LYP level. The same trend can be seen when
Optimized Structure HOMO LUMO
TPA-0
TPA-1
TPA-2
TPA-3
TPA-4
TPa-5
Fig. 4 Optimized structure andfrontier molecular orbitals(HOMO and LUMO) of TPA-norganic dyes at the CAM-B3LYP/aug-cc-pVTZ with solvent effect
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
eV
TiO2 Conduction Band
TPA-3 TPA-4 TPA-2 TPA-0 TPA-5 TPA-1
HOMOs
LUMOs
Fig. 5 Frontier molecular orbital energy levels (HOMO and LUMO) ofTPA-n at the B3LYP/aug-cc-pVTZ with solvent effect
Table 2 Principalelectronic transitions ofTPA-n dyes A, B, and Cbands maxima (λmax) innm at the B3LYP/aug-cc-pVTZ level withsolvent effect
TPA-n A B C
TPA-0 610 434 325
TPA-1 497 402 311
TPA-2 550 485 331
TPA-3 655 485 350
TPA-4 602 448 345
TPA-5 534 412 359
25 Page 4 of 7 J Mol Model (2016) 22: 25
we use CAM-B3LYP. For TPA-1, it is observed possible thatthe highest occupied molecular orbital (HOMO) of that sensi-tizer, both in vacuo (gas phase) and in solution (included sol-vent), is delocalized over the TPA-ring with the phenylenebrigde group (see Fig. 2). The lowest unoccupied molecularorbital (LUMO) of both sensitizers is orbital delocalizedacross the thiophene and cyanoacrylic groups, with sizablecomponents from the cyano- and carboxylic moieties.However, the LUMO in vacuo it is still over the phenylenebrigde.
Along the general structure, we identified four posi-tions that will be the subject of study. We selected R1and R2 position along TPA-rings, also the R3 and R4position over phenylene-ring to analyze the substituenteffects. The HOMO-LUMO gap between these struc-tures changed when the substituents along donor/acceptor groups are modified. To get the TPA-2, the -H atoms of R3 and R4 positions are replaced by twomethoxy groups. It is observed when those -H atomsare changed by those methoxy groups in meta positionsalong the phenyl ring given better results. It is shown inFig. 3 at the B3LYP/aug-cc-pVTZ with solvent effect.Figure 4 shows the frontier orbitals at the CAM-B3LYP/aug-cc-pVTZ. It is possible to see that the same trendsremain at both levels of calculation. The variation of theenergy gaps of these TPA-n can be observed in Table 1.
We included all TPA-n and their frontiers molecular orbitalat the B3LYP/aug-cc-pVTZ with solvent effect, and the ten-dencies are shown in Fig. 5. The data for the energy gaps aremarked at the appropriate orbital energy levels, and show therelative positions of the conduction band of TiO2. Figure 5shows the effect of substituent groups that directly connectto the TPA rings in the variation of the gaps.
Absorption properties
TD-DFT has proved to be an important tool for studying theoptical properties of the dyes in this work. For TPA-n dyes, theUV–Vis absorption spectra in the visible region, usually showsA, B, and C bands at higher energy. The calculated absorptionexcitation energies for TPA-n are reported in Tables 2 and 3,moreover spectra are reported in the Figs. 6 and 7 at the B3LYPand CAM-B3LYP levels with solvent effect. The studied sys-tems show the A and B bands in the region between 650 and400 nm, and the C band at around 300 nm. In this work, whendyes TPA-n (n=0–5) are excited at their wavelength (λ) max-imum absorptions, they exhibit absorption maxima at 610(0),497(1), 550(2), 655(3), 602(4), and 534(5) nm at the B3LYPlevel. The same tendency is obtained for CAM-B3LYP. Thissuggests that the excited states of dyes TPA-1 and TPA-2 needmore energy for the excitation than for TPA-4, TPA-5, andTPA-0. The absorption intensity reveals a more effectivequenching process in dye TPA-3, which means that there is abetter coupling between the donor and acceptor moieties in thisdye. This related a better light harvesting ability of the solarspectrum for dye TPA-3. We have made the same calculationsin gas phase, noting a shift to red when the solvent model isused. The spectra maintain the same band. These results are notshown here.
As described in some previously published papers [39, 40],the fundamental parameters are considered in the theoreticalcalculations, which are related to the experimental controls
Table 3 Principalelectronic transitions ofTPA-n dyes A, B, and Cbands maxima (λmax) innm at the CAM-B3LYP/aug-cc-pVTZ level withsolvent effect
TPA-n A B C
TPA-0 600 442 309
TPA-1 503 398 308
TPA-2 540 398 333
TPA-3 645 431 373
TPA-4 600 423 310
TPA-5 530 417 358
Fig. 6 The UV–Vis absorptionspectra of TPA-n dyes at theB3LYP/aug-cc-pVTZ withsolvent effect
J Mol Model (2016) 22: 25 Page 5 of 7 25
such as open-circuit (Voc) and short-circuit current density (Jsc)[39]. The photo-induced electron injection in DSSCs can beviewed as a charge transfer (CT) process. If we use the Marcustheory for electron transfer [41], the CT can be associated withthe free energy change for electron injection (ΔGinject) [42]. Ingeneral, the greater the ΔGinject, the greater the electron-injection efficiency (Φinject) and it can be calculated asΔGinject =Edye* - ECB [42]. ECB is the reduction potential ofthe CB of the TiO2. E
dye* is the excited state oxidation potential(ionization potential) of the dye, and it is determined by theredox potential of the ground state of the dye (Edye), and thevertical transition energy (λmax): E
dye*=Edye - λmax. The Edye is
approximated applying molecular orbitals eigenvalues on thebasis of Koopmans’ theorem as IP≈ -EHOMO [43]. This is ac-cepted in the literature within the framework of the conceptualdensity functional theory (CDFT) [44, 45].
In Tables 4 and 5, we show the Edye* andΔGinject values forA, B, and C bands of the TPA-n systems. We calculated the
ΔGinject value of every TPA-system with dichloromethane assolvent. We can see that there are no major changes betweenB3LYP and CAM-B3LYP, maintaining the same trend. TheTPA-2 and TPA-4 represented the best electron-injection per-formance. The ΔGinject value is even greater than TPA-0[24, 25]. Strong evidence exists that the inductive effectsin the -meta and -para positions are very similar for thissubstituent (-OCH3). By replacing –H atoms at R1 and R2positions of TPA-2 for two amino groups to design theTPA-4, the effect of the electron-donating power is evengreater for the latter, producing a better performance onthe ΔGinject value. Remarkably, the relationship betweensubstituent effect and ΔGinject is clear. The p-amino is anelectron donating group (groups which tend to increase theelectron density near the reaction site) and disfavors theionization to a negatively charged ion. TPA-3 which con-tains only amino groups over R1 to R4 position shows anegative ΔGinject as well.
Fig. 7 The UV–Vis absorptionspectra of TPA-n dyes at theCAM-B3LYP/aug-cc-pVTZ withsolvent effect
Table 4 The calculated excited state oxidized potential (Edye*/eV) andfree energy change for electron injection (ΔGinject/eV) of the A, B, and Cabsorption bands for TPA-n dyes at the B3LYP/aug-cc-pVTZ level oftheory including solvent effect. ECB is used with an experimental valueof −4.00 eV for the TiO2
TPA-n Edye* ΔGinject
A B C A B C
TPA-0 2.90 2.05 1.12 −1.10 −1.95 −2.88TPA-1 2.77 2.18 2.07 −1.23 −1.82 −1.93TPA-2 2.67 1.97 1.89 −1.33 −2.03 −2.11TPA-3 2.74 1.91 1.09 −1.26 −2.09 −2.91TPA-4 2.61 1.90 1.08 −1.39 −2.10 −2.92TPA-5 2.89 2.20 1.76 −1.11 −1.80 −2.24
Table 5 The calculated excited state oxidized potential (Edye*/eV) andfree energy change for electron injection (ΔGinject/eV) of the A, B, and Cabsorption bands for TPA-n dyes at the CAM-B3LYP/aug-cc-pVTZ levelof theory including solvent effect. ECB is used with an experimental valueof -4.00 eV for the TiO2
TPA-n Edye* ΔGinject
A B C A B C
TPA-0 2.12 0.92 0.40 −1.88 −3.08 −3.60TPA-1 2.14 1.54 1.24 −1.86 −2.46 −2.76TPA-2 1.93 0.72 0.38 −2.07 −3.28 −3.62TPA-3 1.75 1.54 1.31 −2.25 −2.46 −2.69TPA-4 1.74 0.68 0.10 −2–26 −3–32 −3.90TPA-5 2.23 2.08 1.74 −1.77 −1.92 −2.26
25 Page 6 of 7 J Mol Model (2016) 22: 25
Conclusions
We obtained a better theoretical performance in all organicdyes studied here. Since our starting the TPA-derivate wasalready showing higher photovoltaic properties than Preatand co-worker [24, 25]. We found that the TPA-3 and TPA-4new dyes showed even better photovoltaic properties.Furthermore, we found that through structural modificationsthe values of ΔGinject were greater, and gave us a better per-formance for organic dyes. The ΔGinject value is negative inall cases, which means that TPA-n dyes included in this studyinjected an electron to the hyphothetical semiconductor TiO2.These results postulate a potential group of future organic dyesensitizers which is improved by a rational modification ofspecific substitutions over donor acceptor groups. There is agood correlation between B3LYP and CAM-B3LYP.
Acknowledgments This work has been funded by Grants Conicyt-Aka-ERNC-001, Fondecyt 1140503 and 1150629, and ProjectRC120001 of the Iniciativa CientÃfica Milenio (ICM) del Ministerio deEconomÃa, Fomento y Turismo del Gobierno de Chile. N.I. wants toacknowledge the Fondecyt grant N° 11140770.
References
1. Singh G (2013) Energy 53:1–3502. Grätzel M (2009) Acc Chem Res 42:1788–17983. Urbani M, Grätzel M, Nazeeruddin MK, Torres T (2014) Chem
Rev 114:12330–123964. Pastore M, Fantacci S, De Angelis F (2013) J Phys Chem C 117:
3685–37005. Casanova D, Rotzinger FP, Grätzel M (2010) J Chem Theory
Comput 6:1219–12276. De Angelis F, Fantacci S, Mosconi E, Nazeeruddin MK, Grätzel M
(2011) J Phys Chem C 115:8825–88317. Funaki T, Funakoshi H, Kitao O, Onozawa-Komatsuzaki N,
Kasuga K, Sayama K, Sugihara H (2012) Angew Chem Int Ed51:7528–7531
8. Ning Z, Tian H (2009) Chem Commun 45:5483–54959. Yum J-H, Walter P, Huber S, Rentsch D, Geiger T, Nuesch F, De
Angelis F, Grätzel M, Nazeeruddin MK (2007) J Am Chem Soc129:10320–10321
10. Kwon T-H, Armel V, Nattestad A, MacFarlane DR, Bach U, LindSJ, Gordon KC, TangW, Jones DJ, Holmes AB (2011) J Org Chem76:4088–4093
11. Guo K, Yan K, Lu X, Qiu Y, Liu Z, Sun J, Yan F, Guo W, Yang S(2012) Org Lett 14:2214–2217
12. Imahori H, Umeyama T, Ito S (2009) Acc ChemRes 42:1809–181813. Yella A, Lee HW, Tsea HN, Yi CY, Chandirian AK, Nazeeruddin
MK, Diau EWG, Yeh CY, Zakeeruddin SM, Grätzet M (2011)Science 334:629–634
14. Wenger S, Bouit P-A, Chen Q, Teuscher JI, Censo DD, Humphry-Baker R, Moser J-E, Delgado JL, MarÃn N, Zakeeruddin SM,Grätzel M (2010) J Am Chem Soc 132:5164–5169
15. Zhang J, Li H-B, Sun S-L, Geng Y, Wu Y, Su Z-M (2012) J MaterChem 22:568–576
16. Mishra A, Fischer MKR, Bäuerle P (2009) Angew Chem Int Ed 48:2474–2499
17. Ooyama Y, Harima Y (2012) Chem Phys Chem 13:4032–408018. Kanaparthi RK, Kandhadi J, Giribabu L (2012) Tetrahedron 68:
8383–839319. Diebolt U (2003) Surf Sci Rep 48:53–22920. Macwan DP, Dave PN, Chaturvedi S (2011) J Mater Sci 46:3669–
368621. Kullgren J, Huy HA, Aradi B, Frauenhein T, Deák P (2014) Status
Solidi RRL 8:566–57022. Valencia S, MarÃn JM (2010) Restrepogo. OpenMater Sci J 4:9–1423. Buraidah MH, Teo LP, Yusuf SNF, Noor MM, Kufian MZ, Careem
MA,Majid SR, Taha RM, Arof AK (2011) Int J Photoenergy 2011:273683–27694
24. Preat J, Michaux C, Jacquemin D, Perpete EA (2009) J Phys ChemC 113:16821–16833
25. Preat J (2010) J Phys C 114:16716–1672526. Pastore M, Mosconi E, De Angelis F, Grätzel M (2010) J Phys
Chem C 114:7205–721227. Liu D, Fessenden RW, Hug GL, Kamat PV (1997) J Phys Chem B
101:2583–259028. Burfeindt B, Hannappel T, Storck W, Willig F (1996) J Phys Chem
100:16463–1646529. SayamaK, Tsukagochi S, Hara K, OhgaY, ShinpouA,AbeY, Suga
S, Arakawa H (2002) J Phys Chem B 106:1363–137130. Ning Z, Zhang Q, Wu W, Bo L, Tian H (2008) J Org Chem 73:
3791–379731. Jamorski C, Lüthi HP (2002) J Chem Phys 117:4146–415632. Preat J, Jacquemin D, Wathelet V, André JM, Perpéte EA (2006) J
Phys Chem A 110:8144–815033. Casida ME, Jamorski C, Casida KC, Salahub DR (1998) J Chem
Phys 108:4439–444934. Bauernschmitt R, Ahlrichs R (1996) Chem Phys Lett 256:454–46435. Turbomole: ab initio program, Ahlrichs R et al (1989) Chem Phys
Lett 162:165–16936. Frisch MJ, Trucks GW, Schlegel HB et al. (2003) Gaussian 09.
Gaussian Inc, Pittsburgh37. Klamt A, Schüürman G (1993) J Chem Soc Perkin Trans 2(5):799–
80538. Scanlon DO, Dunnill CW, Buckeridge J, Shevlin SA, Logsdail AJ,
Woodley SM, Catlon CRA, Powell MJ, Palgrave RG, Parkin IP,Watson GW, Keal TW, Sherwood P, Walsh A, Sokol AA (2013)Nat Mater 12:798–801
39. Chen J, Bai F-Q, Wang J, Hao L, Xie Z-F, Pan Q-J, Zhang H-X(2012) Dyes Pigments 94:459–468
40. Han L-H, Zhang C-R, Zhe JW, Jin N-Z, Shen Y-L, Wang W,Gong JJ, Chen Y-H, Liu Z-J (2013) Int J Mol Sci 14:20171–20188
41. Marcus RA (1993) Rev Mod Phys 65:599–61042. Fan W, Tan D, Deng W-Q (2012) ChemPhysChem 13:2051–206043. Lee MJ, Balanay M, Kim D (2012) Theor Chem Acc 131:1269–
127844. Geerlings P, De Proft F, Langenaeker W (2003) Chem Rev 103:
1793–187445. Torrent-Sucarrat M, De Proft F, Geerlings P, Ayers PW (2008)
Chem Eur J 14:8652–8660
J Mol Model (2016) 22: 25 Page 7 of 7 25
