Photoexcited Phenyl Ring Twisting in Quinodimethane Dyes

Subscriber access provided by UNIVERSITY OF THE SUNSHINE COAST

Article

Photo-excited phenyl ring twisting in quinodimethane dyes enhances photovoltaic performance in dye-sensitized solar cells Yun Gong, Jacqueline Manina Cole, Hiroaki Ozoe, and Takeshi Kawase ACS Appl. Energy Mater., Just Accepted Manuscript • DOI: 10.1021/acsaem.7b00244 • Publication Date (Web): 06 Feb 2018 Downloaded from http://pubs.acs.org on February 17, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Energy Materials is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Energy Materials

Photo-excited Phenyl Ring Twisting in Quinodimethane Dyes Enhances Photovoltaic Performance in Dye-sensitized Solar Cells Yun Gong,a Jacqueline M. Cole,*abcd Hiroaki Ozoe,e Takeshi Kawasee

a

Cavendish Laboratory, University of Cambridge, J. J. Thomson Avenue,

Cambridge, CB3 0HE, UK. b

ISIS Neutron and Muon Source, STFC Rutherford Appleton Laboratory,

Harwell Science and Innovation Campus, Didcot, Oxfordshire, OX11 0QX, UK. c

Argonne National Laboratory, 9700 South Cass Avenue, Argonne, IL

60439, USA. d

Department of Chemical Engineering and Biotechnology, University of

Cambridge, West Cambridge Site, Philippa Fawcett Drive, Cambridge, CB3 0FS, UK. e

Graduate School of Engineering, University of Hyogo, 2167 Shosha,

Himeji, Japan

*Author for correspondence (J. M. Cole): [email protected]

Key words: Quinodimethane-based dyes  Molecular engineering  Dye sensitized solar cell  Photovoltaic performance  Phenyl ring twist  Diode1

ACS Paragon Plus Environment

ACS Applied Energy Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 50

like effect

Abstract Four metal-free organic quinodimethane-based dyes, which represent molecular building blocks for a new class of DSSC dyes, were studied with the objective to improve their photovoltaic performance via molecular engineering strategies. Such strategies can only be systematic and successful if they are derived from knowledge-based paradigms that relate individual aspects of the molecular structure for a given class of dyes to their photovoltaic properties. The optical and electrochemical properties of these dyes were investigated experimentally by UV-vis absorption and emission spectroscopy, as well as cyclic voltammetry and electrochemical impedance spectroscopy. Density functional theory (DFT) and timedependent DFT (TD-DFT) calculations on these dyes complement these experiments. Among other results, this concerted experimental and computational study revealed that the dialkylaminophenyl moiety in these dyes exhibits a large twist as a result of its ground-to-excited state optical transition. In particular, a near-perpendicular (88.98o) twist of the dimethylaminophenyl moiety relative to the -bridging unit that connects the three aryl rings in 2 was observed upon formation of its photo-excitedstate structure. Moreover, it was discovered that 2 affords the highest VOC and power-conversion efficiency (PCE) values among these dyes. The 2


Page 3 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


particularly high PCE for 2 was found to be due to this twisting of the phenyl ring, which blocks the pathway of electrons from TiO 2 to the donor, and hence suppresses undesirable electron recombination at the dye···TiO2 interface.

This diode-like effect minimizes

undesirable electron-

recombination effects and represents an unprecedented structure-property relationship that should be useful for the molecular engineering of larger chromophores in this class of dyes for DSSC applications.

1. INTRODUCTION The first dye-sensitized solar cells (DSSCs) with substantial photovoltaic output were reported by O’Regan and Grätzel in 1991,1 and subsequently these devices were considered to belong to the third generation of solarcell technology. Since then, DSSCs have become a promising alternative to fossil fuels, which is particularly interesting with respect to economic considerations given the excellent price-to-performance ratio of DSSCs. While metalorganic complexes were used initially as the original dyes for such DSSCs devices, organic dyes have become increasingly attractive as alternative chromophores, especially given environmental considerations. Moreover, metal-free dyes are usually cheaper, exhibit high molecular extinction coefficients, and are usually easier to synthesize than metalcontaining dyes. So far, the most efficient metal-free DSSCs have reached power conversion efficiency (PCE, ) values of up to 14.3%.2

3



Page 4 of 50

Mimicking the photosynthesis in plants, DSSCs separate light absorption and charge transportation. DSSCs thus usually consist of a photoanode (i.e., a nano-structured layer of a metal-oxide semiconductor) sensitized with dye molecules, a counter electrode, and a redox electrolyte. The dye is a crucial device component on two distinct accounts: on one hand, it harvests photons from sun light, while on the other hand, it injects electrons into the semiconductor layer of the working electrode, which is a regenerative process provided the provision of electrons from a redox shuttle.

Metal-free dyes often exhibit a donor-π-bridge-acceptor (D-π-A) structure. In order to improve the light-harvesting properties of DSSCs, and to better align the energy levels between the dye and those of the semiconductor or electrolyte, which affords the highest possible driving forces between device components, various organic moieties have been employed as electron

donors, including fluorine,

3,4

coumarins,5,6 indolines,7,8

cyanines,9,10 carbazoles,11,12 and triphenylamines.13,14,15 The electron acceptors are responsible for the generation of electron-hole pairs, and most commonly act simultaneously as anchors for the dye molecules to adsorb onto the surface of the semiconductor. On the basis of their anchoring strength, cyanoacrylic acid and carboxylic acid groups are, by far, the most commonly employed electron acceptors in D-π-A dyes.16

4


Page 5 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


To further improve the charge-separation properties of the electron acceptor, a phenyl ring can be introduced between the -bridge and the acceptor. The validity of this concept has been demonstrated for a DSSC device that contains a dye with a D--A-πphenyl-Anc (Anc = anchor ≠ acceptor) structural motif that features a benzothiadiazole donor and cyanoacrylic acid anchor, wherein electron back-transfer is apparently blocked and electron recombination delayed.15 The use of this D-π-Aπphenyl-Anc motif also increases the choice of acceptors, by isolating the function of the acceptor from that of the anchor. 15,16

We have recently developed a data-mining method that predicts new classes of DSSC dyes. The first-generation predictions exploited this expanded choice of electron acceptors, by identifying promising D-π-Aπphenyl-Anc motifs from a set of 118,465 organic molecules.17 The results of that data-mining study indicated that a quinodimethane-based motif showed particular promise as a molecular building block for DSSC dyes. Indeed, in D-π-A-πphenyl-Anc-based DSSC dyes, this moiety exhibited excellent intramolecular charge-transfer (ICT) properties for different electron donors (dimethylamino or diphenylamino groups) and anchors (cyanoacrylic or carboxylic acid groups).17 Figure 1 displays the corresponding chemical structures of (E)-2-cyano-3-(4-((3,5-di-tert-butyl4-oxocyclohexa2,5-dien-1-ylidene)(45



Page 6 of 50

(dimethylamino)phenyl)methyl)phenyl)acrylic acid (1), 4-((3,5-di-tertbutyl-4-oxocyclohexa-2,5-dien-1-ylidene)(4(dimethylamino)phenyl)methyl) benzoic acid (2), (E)-2-cyano-3-(4-((3,5di-tert-butyl-4-oxocyclohexa-2,5-dien-1-ylidene

(4-

(diphenylamino)phenyl)methyl)phenyl)acrylic acid (3), and 4-((3,5-ditert-butyl-4-oxocyclohexa-2,5-dien-1-ylidene)(4(diphenylamino)phenyl)methyl) benzoic acid (4).

Figure 1. Chemical structures of 1-4, whereby electron donors and anchoring groups are highlighted in green and blue, respectively.

The aforementioned data-mining study17 examined the influence of two 6


Page 7 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


different donor moieties on the optoelectronic, electrochemical, and photovoltaic properties of the resulting dyes, while the anchoring group (cyanoacrylate) remained constant. In order to test the potential effects of dye aggregation, which should be diminished by the presence of bulky groups, both small (1) and bulky donors (3) were selected.18

The study presented herein includes the corresponding properties of two other compounds that feature a carboxylic acid anchor (2, 4), using a concerted set of experimental methods and computational calculations. These methods are also used to further evaluate the properties of the originally studied compounds (1 and 3).17 Herein, we will demonstrate that 2 affords the best photovoltaic performance and overall conversion efficiency among this set of DSSC dyes (2 > 4 ~ 1 > 3). The results obtained from UV-vis absorption and emission spectroscopy, together with those of density functional theory (DFT) and time-dependent DFT (TD-DFT) calculations demonstrate that a twisting of the dialkylaminophenyl moiety in the excited-state geometry is responsible for the distinguishing photovoltaic properties that are observed by DSSC device testing. A rationale for this structure-property relationship is offered by considering the diode-like effect that this phenyl ring twist exerts upon the forward and backward electronic pathways that are required for electron injection and recombination, respectively. 7



Page 8 of 50

2. EXPERIMENTAL AND COMPUTATIONAL METHODS 2.1 Synthesis

The synthetic route to 1 and 3 is essentially the same, whereby 4-bromo2,6-di(tert-butyl)-1-(trimethylsiloxy)benzene

is treated with 4-N,N’-

dialkylamino-4’-formylbenzophenone ethylene acetal (alkyl = Me19 or Ph20) in the presence of n-butyl lithium in THF at -78C, followed by a Knoevenagel condensation with cyanoacetic or acetic acid in acetonitrile. 21 The synthesis of 1 and 3 has been described elsewhere.17 2 and 4 were synthesized in good yield from the corresponding 4-(diphenylmethylene)2,5-cyclohexadien-1-one derivatives that contain aldehyde instead of carboxylate groups by treatment with NaClO2 and NaH2PO4. Full details for the synthesis of 2 and 4 are given in the Supporting Information.

2.2 Optical and Electrochemical Measurements

UV/vis absorption spectra of 1-4 in solution and in thin dye…TiO 2 nanocomposite films were recorded on an Agilent8453 Diode Array spectrophotometer (resolution: 2 nm). Emission spectra of 1-4 in solution were recorded on a Cary Eclipse fluorescence spectrophotometer (excitation/emission slits: 20 nm). 8


Page 9 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Cyclic voltammograms were recorded using a computer-controlled Autolab PGSTAT101 potentiostat with a three-electrode setup (working electrode: glassy carbon; counter electrode: Pt wire; reference electrode: Ag/AgCl). The Ag/AgCl reference electrode was calibrated against the ferrocene/ferrocenium (Fc/Fc+) redox couple as the external standard. 1-4 were dissolved in acetonitrile (5

×

10-5 mol L -1), containing 0.1 M

tetrabutylammonium perchlorate (TBAP) as the supporting electrolyte. Measurements were carried out using a scan rate of 100 mV/s.

To estimate the extent of dye loading on the TiO 2 films, the sensitized electrodes were separately immersed in a NaOH solution (0.1 M) in water:ethanol (1:1, v/v) for the desorption of 1-4. The extent of dye loading was obtained by measuring the absorbance of the resulting solutions, based on the molar extinction coefficients of 1-4, which were obtained from the the UV-vis absorption spectra.

2.3 Fabrication and Testing of DSSC Devices

TiO2 (DSL 18NR-T, Dyesol) photoanodes were fabricated on pieces of well-cleaned FTO glass (Dyesol, TEC-15) using the doctor-blade technique, and subsequently dried. This doctor-blade/drying procedure 9



Page 10 of 50

was repeated twice to afford even layers (thickness: 4 μm). The resulting electrodes were subsequently sintered for 30 min at 500 ℃. Counter electrodes were fabricated using H2PtCl6·6H2O (Sigma) and a similar doctor-blade/drying procedure, followed by sintering for 30 min at 375 ℃. All photoanodes were then immersed overnight into methanolic solutions (5 × 10-5 mol L -1) of 1-4 and N719. The thus sensitized photoanodes and counter electrodes were assembled in a sandwich fashion including the redox electrolyte (5  10-2 M I-/I3-; HPE, Dyesol).

The current-voltage (J-V) curves and the overall DSSC photovoltaic performance of the thus obtained devices were measured using an ABET Sun 2000 solar simulator under AM1.5 illumination (100 mW cm -2 after spectral mismatch) with an active area of 0.5 cm 2.

2.4 Computational Details

Gas-phase and solution-based structures of 1-4 were optimized at the B3LYP22,23/6-31g(d)24 level of theory using the Gaussian 09 software. 25 Solution-based structures were optimized in methanol using the polarizable continuum model (PCM) with the dielectric constant of methanol (ε = 32.63).26,27 Single-point energy calculations were performed at the B3LYP/6-31g++(d,p) level of theory. 10


Page 11 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The interface between the dye molecules and the TiO2 nanoparticles was also modelled, under consideration of a feasible balance between scientific accuracy and computational costs. Accordingly, dye molecules with optimized gas-phase geometries were initially absorbed onto a (TiO2)9 cluster.28 For the cyanoacrylic and carboxylic anchoring groups, several possible adsorption modes exist for the attachment onto the (101) face of anatase TiO2, which is the most common type of TiO2 surface for dye adsorption.29-32 In general, bidentate coordination through the oxygen atoms of the carboxylate is considered the most stable anchoring mode for carboxylic acids.16,33,34 Checks for the most stable dye…TiO2 binding configuration were therefore based on the bidentate chelating and bridging anchoring modes in the protonated and deprotonated form. 2 was used as a representative example for modelling this dye…TiO2 interface, as 1 and 2 contain the same anchoring group. Bidentate coordination was also predicted as the most stable anchoring mode for cyanoacrylic acids, although this group presents more options, since the nitrogen of its nitrile group may also act as an anchoring point. 35-37 The COO/CN binding configuration was therefore studied as a possible anchoring mode together with the options that are available for a carboxylic acid anchoring group. In this case, 1 was used as a representative example for the dye…TiO 2 interface structural model for 1 and 3, since they exhibit the same 11



Page 12 of 50

anchoring group. The geometries of the adsorbed dye molecules were subsequently optimized at the same level of theory as the isolated dye molecules; associated vibrational-frequency calculations were carried out at the same level of theory on all geometry-optimized structures. Singlepoint energy calculations for all absorbed dyes were performed at the B3LYP/6-31g (d) level of theory. Theoretical UV/vis absorption spectra were calculated using timedependent density functional theory (TD-DFT) in combination with the M06-2X functional38 and the 6-31++g (d,p) basis set (isolated dyes) or the 6-31g (d) basis set (adsorbed dyes), based on the optimized ground-state structures. The geometries of all dyes in the excited state were optimized at the B3LYP/6-31g (d) level of theory.

3. RESULTS AND DISCUSSION The optical, electrochemical, and photovoltaic properties of 1-4 were evaluated via solution-state experiments. These were complemented by computational calculations for 1-4 in the gas phase and in solution. Moreover, computational calculations were extended to cases where 1 and 2 were adsorbed onto the (101) surface of anatase TiO2, which represents the dye…TiO2 working electrode.29-32 These calculations assessed the ground- and photo-excited state structures and optical properties of 1-4, the suitability of the frontier molecular orbital energies of 1-4 toward the 12


Page 13 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


alignment of the relevant quantum energies of the non-dye DSSC device components, the preferred dye adsorption mode on TiO2, and the associated energy modulation of these dyes upon forming the dye …TiO2 interface. 3.1 Experimental Characterization of Optical Properties.

(a)

13



Page 14 of 50

(b) Figure 2. Absorption spectra of (a) 1-4 in methanol ([dye] = 5 × 10-5 M) and (b) 1-4 on thin films of TiO2 (thickness: 4 μm).

The UV/vis absorption spectra of 1-4 in methanol ([dye] = 5 × 10-5 M) and of 1-4 on thin films of TiO2 are shown in Figure 2, and all associated data are summarized in Table 1. The absorption curves of the pairs with common anchoring groups, i.e., 1 / 3 and 2 / 4, feature similar shapes, which indicates that strong excitation processes (300-450 nm) occur at the sites of the anchoring group. For 1 and 3, which contain cyanoacrylic acid anchors, the absorption bands at shorter wavelength (1: 320 and 375 nm; 3: 301 and 372 nm) correspond to the aromatic π–π* transition of the conjugated backbone, while the longest-wavelength absorption band of 1 (λmax,

abs

= 503 nm) was assigned to an ICT17 that is bathochromically

shifted compared to that of 3 (λmax, abs = 482 nm). For 2 and 4, which contain carboxylic acid anchors, the longest-wavelength absorption bands show a similar trend, i.e., the ICT of 2 (λmax, abs = 490 nm) is bathochromically shifted compared to that of 4 (λmax, abs = 472 nm). The shorter-wavelength adsorptions of 2 and 4 exhibit shoulder peaks (310-400 nm) that may be caused by several π–π* transitions that are similar in energy. Moreover, the ICT band of 1 and 3 is more intense than that of 2 and 4, indicating stronger ICT interactions in the former compounds, possibly due to a more extended 14


Page 15 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


π-delocalization owing to the presence of the cyanoacrylic acid moiety.

The spectra of the adsorbed dyes (Figure 2b) revealed slight peak broadening and a hypsochromic shift of the ICT band for 1-4 compared to the spectra of 1-4 in solution. This shift to shorter wavelengths could be explained by the effect of the adsorption and the occurrence of dye aggregation, as both deprotonation and H-aggregation result in a hypsochromic shift.39 In this case, 2 (490437 nm,  = 53 nm) and 4 (473446 nm,  = 27 nm) exhibit more pronounced shifts relative to 1 (503464 nm,  = 39 nm) and 3 (482474 nm,  = 8 nm). Considering the identical backbone of 1-4, the differences between these pairs should be attributed predominantly to the anchor groups, which are deprotonated upon adsorption onto TiO2. The result that 3 and 4 exhibit less pronounced hypsochromic shifts than 1 and 2 indicates that dye aggregation also plays a role, since the steric bulk of the diphenylamino donor of 3 and 4 should prevent dye aggregation. Given the hypsochromic nature of the shift observed, the dye aggregation that manifests should involve face-to-face dye interactions.18 Considering the preponderance of phenyl rings in 1-4, … interactions are the most likely form of dye…dye interactions.

15



Page 16 of 50

Figure 3. Emission spectra of 1-4 in methanol ([dye] = 1 × 10-3 M,  ex = 400 nm).

The emission spectra of 1-4 (Figure 3) show emission peaks in the UV/vis range: 647 (1), 608 (2), 638 (3), and 616 nm (4). The Stokes shifts of 3 ( = 4907 cm-1) and 4 ( = 5073 cm-1) are substantially larger than those of 1 ( = 4405 cm-1) and 2 ( = 3961 cm-1), indicating that these shifts may be dominated by a structural reorganization of the donor group. A more detailed comparison of the molecules that contain the same donor (NMe 2) showed that the Stokes shift of 2 is slightly smaller than that of 1, which means that a small additional change in geometry between the ground state (GS) and the excited state (ES) must occur, whose nature affects the position of the anchoring group relative to that of the donor group. 16


Page 17 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Considering the individual structures, the most probable geometry reorganization from the GS to the ES is some form of twist of the aryl rings in the -bridge that connects the donor and anchoring groups. DFT and TD-DFT studies were employed to verify this chemical intuition regarding the nature of the geometry reorganization upon photo-excitation.

3.2 Computational Characterization of the Optical Properties of 1-4 Prior to performing any calculations on GS-to-ES optical transitions, DFT calculations on the GS of 1-4 were carried out. These DFT studies also served to corroborate the origins of the experimental UV/vis absorption spectra features described in Section 3.1.

3.2.1 Molecular Orbitals and First Excitation Energies of 1-4

Figure 4. Electron-density distributions for the frontier molecular orbitals of 1-4 calculated at the DFT/B3YLP/6-31++g (d, p)/PCM level of theory.

17



Page 18 of 50

Table 1 Summary of the optoelectronic parameters calculated using DFT and TD-DFT methods.

HOMO a)

EgTD b)

LUMO c)

[eV]

[eV]

[eV]

1

-5.408

2.499

-2.909

2

-5.385

2.661

-2.724

3

-5.447

2.590

-2.857

4

-5.434

2.718

-2.716

a) Calculated by DFT methods;

b)

calculated by TD-DFT methods;

c)

calculated from the sum of the HOMO energy and Eopg t.

The results in Figure 4 and Table 1 show that the highest occupied molecular orbitals (HOMOs) of 1-4 exhibit similar electronic distributions in the GS, whereby the electron density is predominantly located on the dimethylamino or diphenylamino donors and the quinone backbone. The electron distribution of the lowest unoccupied molecular orbitals (LUMOs) of 1-4 is localized on the carboxylic or cyanoacrylic acid anchoring groups and the quinone backbone, whereas small parts of the LUMOs of 2 and 4 are also situated on the phenyl ring that connects the quinone and donor moieties. In general, 1-4 show very good HOMO-to-LUMO charge 18


Page 19 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


redistribution from the donor to the anchoring groups, which renders the injection of excited electrons into the conduction band (CB) of TiO 2 possible. Moreover, it should be noticed that the quinone backbone between the donor and acceptor reveals sufficient HOMO and LUMO overlap to guarantee a fast charge-transfer transition.15 The first excitations of 1-4 in methanol were studied using TD-DFT methods. The  𝑝𝑒𝑎𝑘 𝑚𝑎𝑥 values predicted from the calculated first excitation energies (Figure 5) agree well with the results of the experimental UV/vis spectra for 1 and 3 in the case of the two most bathochromic  𝑝𝑒𝑎𝑘 𝑚𝑎𝑥 values, which are centered at 375 nm and 500 nm. Conversely, the  𝑝𝑒𝑎𝑘 𝑚𝑎𝑥 value centered at 300 nm is better simulated for 2 and 4. The different levels of agreement between experimental and computational values may be due to the suspected twist in the -bridge between the donor and anchor in their ES structures. The compositions of the excitations in terms of molecular orbital contributions were also calculated (Table S1), which show that the CT absorption bands of 1-4 consist predominantly of HOMO-LUMO transitions.

19



Page 20 of 50

Figure 5. Comparison of TD-DFT calculated oscillator strengths and frequencies (vertical lines) against the experimental UV/vis absorption spectra of 1-4 in methanol.

3.2.2 Optimized geometries for 1-4 in the Ground State (GS) and Excited State (ES).

To better understand the origins of the Stokes shifts of 1-4 and their ES properties, the geometries of these molecules in the ES were optimized by DFT and TD-DFT methods, using the same level of theory as per the GS. Notable differences in the molecular geometries of 1-4 between their GS and ES structures were observed, as enumerated in Table 2 and presented 20


Page 21 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


visually in Figure S3 by structural overlays of these GS and ES molecules.

Figure 6. Chemical structure of the core of 1-4 (R1 = –CH=C(CN)COO H or –COOH; R2 = –NMe2 or –NPh2). α, β, and γ represent the dihedral angles of the three aryl rings with respect to the C10-C20-C2 plane.

Table 2 Geometric parameters of the molecular structures of 1-4 that afford significant differences between the GS and ES. 21



a)

Page 22 of 50

α, β, and γ refer to the dihedral angles (o) C21-C20-C1-C2, C3-C2-C1-

C10, and C11-C10-C1-C20, respectively, while Δ refers to the difference GS α a)

β

γ

1

-46.84

-19.70

-35.87

2

-48.60

-19.72

3

-46.17

4

-48.14

C1-C2

b)

C1-C20

C1-C10

1.401

1.485

1.462

-34.92

1.400

1.487

1.461

-18.19

-40.32

1.396

1.483

1.472

-18.24

-38.46

1.395

1.486

1.470

ES 1

-23.74

-14.80

-86.81

1.422

1.433

1.499

2

-15.08

-23.08

-88.98

1.431

1.440

1.497

3

-17.95

-27.76

-61.59

1.418

1.437

1.499

4

-16.21

-23.98

-79.77

1.430

1.440

1.500

Δ(ES-GS) 1

23.10

4.90

-50.94

0.021

-0.052

0.037

2

33.52

-3.36

-54.06

0.031

-0.047

0.036

3

28.22

-9.57

-21.27

0.022

-0.046

0.027

4

31.93

-5.74

-41.31

0.035

-0.046

0.030

between GS and ES;

b)

C-C refers to the bond length (Å) between the two

specified carbon atoms. The ES molecular structures of 1-4 show consistent differences compared to those of the GS, particularly with regard to the relative orientation of the 22


Page 23 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


three aryl groups. The C1-C20 bond lengths are shorter in the ES, to the extent that its bonding manifests a more delocalized nature. The adjoining C1-C2 bond is longer in the ES, such that its distinctly double-bonded GS character (cf. 1.392 Å for typical C=C double bonds39) is better described as -delocalized in the ES. The -bridge that connects the quinone moiety with the phenyl ring that features the electronic acceptor would therefore appear to be more delocalized in the ES. The associated dihedral angle between the plane of the aryl ring and the reference plane,  shown in Figure 6, is smaller in the ES, indicating that -conjugation should extend more readily between these moieties, subject to a suitably delocalized bridging connection; as we have just seen, the nature of this is enhanced in the ES structure relative to that of the GS. The combined effects mean that for the ES structure, more ICT between these two planes will result, strengthening the role of the electron-acceptor group.

The C1-C10 bond involved in the -bridge that connects to the dialkylaminophenyl moiety becomes elongated in the ES, to the extent that it is best described as a single C-C bond (cf. 1.54 Å for typical C(sp3)-C(sp3) bonds39). The origin of this bond elongation can be understood by considering the substantial increase in dihedral angle between the reference plane and the involved aryl ring,  (Figure 6), upon realizing the ES structures of 1-4 that exhibit  values of 50.94, 54.06, 21.27, and 23



Page 24 of 50

41.31, respectively. The particularly large  value for 2 means that ICT will be hindered between the electron donor and the anchoring groups; as such, there will be little chance of electron back transfer for the D-to-Anc ICT that occurs. Moreover, the dimethylamino-phenyl moiety adopts an almost perpendicular (1: 86.81; 2: 88.98) arrangement relative to the C10-C20-C2 plane, which contains the -bridge that links all three aryl rings in 1-4. Accordingly, the ability of the -bridge to channel conjugation between rings in 1 and 2 should be severely compromised.

Considering the α and values together with the observed Stokes shifts for 1-4 from the UV/vis emission spectroscopy data (cf. Section 3.1), their ES geometries can be interpreted more effectively. The much larger values for 1 and 2 in combination with the distinctly smaller Stokes shifts compared to 3 and 4, indicate that the steric effect of the donor moiety may be one of the dominant factors, as the diphenylamino group is sterically more demanding than the dimethylamino group. However, the increased steric bulk of the anchoring group in 1 compared to 2 also needs to be considered. Despite 1 exhibiting a smaller than 2, 1 still shows a larger Stokes shift than 2 owing to its higher α value that is affected by the anchoring group, whereby the cyanoacrylate substituent in 1 is sterically more demanding and electron rich than the carboxylate group in 2. 24


Page 25 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In their entirety, these factors suggest that the π-conjugation in the acceptor group of 2 is particularly strengthened upon forming its ES structure, while it is attenuated in the donor; this effect should ultimately promote charge separation upon photoexcitation. In contrast, the large twist in the dimethylaminophenyl moiety of 2 hinders ICT, which almost eliminates a potential electron back transfer from TiO2 through to the donor group in 2 via its acceptor in the ES; this may effectively suppress electron recombination in DSSCs. Also, this large twist appears to raise the HOMO level, judging from the fact that the HOMO energy levels follow the same trend as , i.e., EHOMO 2 > EHOMO 1 > EHOMO 4 > EHOMO 3. Thus, 2 should offer the highest driving force for dye generation. In combination, these device-relevant factors should lead to a higher open voltage (VOC).

Accordingly, the suspected twisting of the phenyl group in the donor region of 1-4 is corroborated by these computational studies. Moreover, the calculations explain the origin of the distinct emission features that are experimentally observed for 2. Meanwhile, the photovoltaic prospects for 2 as a DSSC dye are attractive on account of the predicted diode-like action of the phenyl ring twisting upon formation of its ES, which will help to preclude undesirable electron-recombination effects.

25



Page 26 of 50

3.3 Electrochemical Characterization of 1-4 3.3.1 Cyclic voltammograms for 1-4

Figure 7. Cyclic voltammograms for 1-4 in acetonitrile ([dye] = 5 × 10-5 M, T = 295 K).

The photovoltaic prospects of 1-4 naturally depend on their electrochemical potential and optical properties. Accordingly, the electrochemical properties were assessed by cyclic voltammograms that were measured in acetonitrile. As shown in Figure 7, 1-4 exhibit reversible oxidation and reduction features. The oxidation potentials (Eoxo ) of 0.40 V (1), 0.32 V (2), 0.48 V (3), and 0.43 V (4) (all potentials vs Ag/AgCl) were assigned to the oxidation of the donor groups. Easier 26


Page 27 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


oxidation should be attributed to the lower electron-withdrawing properties of the dimethylaminophenyl donor group (1 and 2) compared to that of the diphenylaminophenyl donor group (3 and 4), which should lead to a weaker interaction between the acceptor and the donor groups. The energy levels of the HOMOs were determined by the onset of the oxidation potential, while the energy levels of the LUMOs were determined by the sum of the HOMO energy levels and the optical band gaps (estimated from the onset of the first absorption band in UV/vis spectra). A summary of all data is given in Table 3. Both HOMO and LUMO values agree well with those obtained from the theoretical calculations (cf. Table 1). Furthermore, the HOMO energy levels of 1-4 lie below that of the redox potential of the I-/I3- redox electrolyte, which should facilitate the regeneration of the dye molecules that is required for DSSC applications. This conclusion is of course based on the assumption that I-/I3- is the electrolyte, which seems reasonable given that it is, by far, the most commonly employed electrolyte in DSSC devices. Meanwhile, the LUMO energy levels of 1-4 are sufficiently high to provide the driving force for electron injection from the ES of the dyes to the CB of TiO2.

Table 3. Optical and electrochemical data for 1-4. λ max, abs [nm]

ε max

λ max, em [nm]

Stokes shift [cm-1]

o Eox

HOMO [eV]

a)

Egopt b) [eV]

LUMO [eV] 27


c)


[104 Lmol1cm-1]

1 2 3 4 a)

503 375 490 358 482 372 473 362

2.50 2.80 2.46 1.25 2.63 3.55 2.40 1.63

Page 28 of 50

[V]

647

4425

0.40

-5.15

2.06

-3.09

608

3961

0.32

-5.07

2.15

-2.92

638

5073

0.48

-5.23

2.16

-3.07

616

4907

0.43

-5.18

2.19

-2.97

EHOMO (Fc+/Fc) = -4.8 eV;

b)

estimated from the onset of the first

absorption band in the UV/vis spectra in methanol;

c)

calculated from the

sum of the HOMO value and Eopg t.

3.4 Adsorption of 1-4 onto TiO2 Surfaces to Form dye …TiO2 Working Electrodes

So far, the photovoltaic prospects of 1-4 have been considered in terms of their intrinsic molecular properties. However, these dyes must be able to adsorb onto TiO 2 surfaces and retain their attractive intrinsic molecular properties upon adsorption, if they are to act as viable dyes in working electrodes of DSSC devices. Accordingly, DFT methods were employed in order to: (i) assess possible dye…TiO2 adsorption modes for 1-4, and (ii) check that the extent by which the frontier molecular orbital energy levels of 1-4 may be modulated upon adsorption to TiO 2 does not deter a suitable alignment with the energy levels of other DSSC device components.

28


Page 29 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


3.4.1 Preferred Anchoring Configurations for 1-4 on TiO2 Surfaces

The possible adsorption modes of 1-4 on TiO2 were assessed using DFT calculations that employed a (TiO2)9 cluster model in order to simulate the interface between the dyes and the TiO2 films. For dyes that contain a carboxylic acid anchor (2 and 4), the four most common anchoring modes in protonated or deprotonated form are shown in Figure 8(a);34 these are: bidentate chelating (B and BH) and bidentate bridging (BB and BBH). These four possible binding modes were used as models to investigate the adsorption behavior of 2, taking this dye as representative for dye…TiO 2 adsorption of 2 and 4.

BBH

BB

BH

B

(a)

(b) 29



Page 30 of 50

Figure 8 (a) The four most common adsorption modes for the carboxylic acid anchoring group on TiO2: protonated (BBH) and deprotonated (BB) bidentate-bridging anchoring, as well as protonated (BH) and deprotonated bidentate-chelating anchoring (B). (b) The possible CN/COO adsorption mode for the cyanoacrylate acid anchoring group on TiO 2.

For dyes featuring a cyanoacrylate acid anchoring group (1 and 3), an additional adsorption mode is possible, given that the nitrogen atom of its nitrile group may also contribute to the adsorption via a COO/CN binding mode (Figure 8(b)).35-37 This COO/CN anchoring mode was therefore studied for 1, as a representative example for the dye…TiO2 adsorption of 1 and 3, in addition to the options that are available for the carboxylic acid anchoring group.

The dye…TiO2 adsorption energies were calculated according to: 𝐸𝑎𝑑𝑠 = 𝐸𝑑𝑦𝑒…𝑇𝑖𝑂2 − 𝐸𝑑𝑦𝑒 − 𝐸𝑇𝑖𝑂2 ,

(1)

wherein E refers to the DFT-calculated single-point energy values.

In the case of 1, the BBH mode showed a far higher adsorption energy than all the other possibilities, and was accordingly ruled out. The remaining anchoring options were considered in light of not only the adsorption energies but also the photoexcitation properties of 1. The TD-DFT results 30


Page 31 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


revealed that the B and BB modes show hypsochromic shifts of  = 45.51 and  = 29.49 nm, respectively for the first excitation energy upon adsorption, which is in good agreement with the experimentally observed hypsochromic shift by UV/vis spectroscopy ( = 39 nm; Figure 2(b); cf. Section 3.1). The DFT model containing the CN/COO anchoring mode shows the lowest adsorption energy, while its cognate TD-DFT results yielded a hypsochromic shift of  = 10.49 nm, which is much smaller than that obtained from experimental UV/vis spectroscopy. In contrast, the TD-DFT results for the BH mode suggest a bathochromic shift ( = ~16 nm) upon adsorption. Comparing the best overall merits of these modes in terms of adsorption and first excitation energies (Table 4), it would seem that the BB mode presents the most feasible adsorption mode for the dyeTiO2 interface. This notion should also strengthen the notion that deprotonation of the anchoring group upon dye adsorption may induce the hypsochromic shift observed in the UV/vis absorption spectra. That said, the possibility of a COO/CN binding configuration being the preferred adsorption mode cannot be ruled out given its adsorption energy is distinctly lower than all other feasible options. Thereupon, the difference between theoretical and experimental wavelength shifts in the optical absorption may conceivably be rationalized by considering possible variations between the solution-state UV/vis absorption spectral features and that of the dye once absorbed onto TiO 2. Indeed, this COO/CN binding 31



Page 32 of 50

configuration has recently been found in other dyes that contain cyanoacrylate anchoring groups.35-37

In the case of 2, the BBH mode exhibited a markedly higher adsorption energy than the other modes, similar to the case of 1, which precludes this option as a binding mode for this type of dye. In terms of the photoexcitation properties of 2, the TD-DFT results on the structures containing B, BH, or BB modes predict a hypsochromic shift of the first excitation energy upon adsorption onto TiO2. However, all these shifts (cf. Table 4) are noticeably smaller than that observed experimentally by UV/vis absorption spectroscopy ( = 53 nm), whereby those of the B and BB modes offer the closest agreement with the experimental values. Taking both adsorption mode and first excited state energies into account, the BB anchoring mode appears to be the best choice for 2 to anchor on the surface of TiO2; this conclusion also confers with those of the preferred anchoring mode for 1.

Table 4. Calculated first excitation and adsorption energies for the different adsorption modes of 1 and 2 on TiO2. Adsorption

First excitation

ΔFirst excitation

Adsorption

Mode

wavelength [nm]

energy a) [nm]

energy [eV]

32


Page 33 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1B

450.56

-45.51

-0.902

1BH

512.64

16.57

-1.099

1BB

466.58

-29.49

-0.957

1BBH

542.19

46.12

-0.197

1CN/COO

485.58

-10.49

-1.200

2B

425.90

-39.97

-0.892

2BH

459.09

-6.78

-0.993

2BB

429.82

-36.05

-0.985

2BBH

445.21

-20.66

-0.442

a)

ΔFirst excitation energy = ΔFirst excitation energy (dye…TiO2) - ΔFirst

excitation energy (dye in methanol). The sign “-” refers to a hypsochromic shift relative to the energy for the dye in the methanol solution prior to adsorption.

3.4.2 Extent of HOMO and LUMO Energy Modulations of 1 and 2 After Adsorption on TiO2

33



Page 34 of 50

Figure 9. Calculated electron-density distribution for the frontier molecular orbitals of 1 and 2 adsorbed on a (TiO2)9 cluster via a BB anchoring mode (calculated at the DFT/B3YLP/6-31g (d) level of theory).

To gain further insight into the nature of the electronic structures and the charge transfer of such dyes once they have been implemented in a DSSC device environment, the frontier molecular orbitals (MOs) of 1 and 2 were calculated after adsorption onto a (TiO2)9 cluster via the BB anchoring mode. The corresponding frontier MO isosurfaces (HOMO-1, HOMO, LUMO, LUMO+1) for the dye…TiO2 interface structure are shown in Figure 9. For 1 and 2, the HOMO is mainly located on the quinone backbone and donor group, while the HOMO-1 is situated exclusively on 34


Page 35 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the donor group. The LUMO and LUMO+1 are situated on the TiO2 cluster surface. These results reveal an obvious charge transfer between HOMO/HOMO+1 and LUMO/LUMO+1, indicating that the carboxylate and the cyanoacrylate acid group are appropriate anchoring groups for quinodimethane dyes that enable electron injection from the LUMO of the dyes into the CB of TiO2 via intramolecular charge transfer.

3.4.3 Extent of Dye Loading

To verify the effect of the anchor group on the extent of dye loading, which is a crucial parameter for the photovoltaic performance of a DSSC device, a 0.1 M NaOH solution in a water:ethanol (1:1, v/v) was used for the desorption of 1-4 on the TiO2 films (~ 4 μm).

Table 5. Extent of dye loading for 1-4 on TiO2

Dye loading Dye

[10-7mol cm-2]

1

0.89

2

0.82

3

0.63

4

0.76

35



Page 36 of 50

Table 5 summarizes the extent of dye loadings for 1-4. The fact that 1 and 2 show similar levels of adsorption on TiO 2 films, which are better than those of 3 and 4, indicates that the cyanoacrylic acid anchor may adopt the same adsorption mode as the carboxylic acid anchor. Moreover, the steric bulk of the NPh 2 group in 3 and 4 should require more surface area per molecule and thus reduce the total amount of adsorption sites available on the TiO2 surface.

3.5 The Photovoltaic Performance of 1-4.

DSSC device testing of 1-4 was undertaken for several reasons. On one hand, these tests naturally assess the feasibility of photovoltaic device applications for 1-4. It is important to remember that these four dyes are essentially the molecular building blocks of a new class of DSSC dyes, given the nature of the data-mining methods used for their materials discovery.17 Therefore, a large photovoltaic response should not be expected for 1-4; rather, a comparison of the results should help to discern the relative intrinsic photovoltaic merits of 1-4, the best of which could subsequently be adopted in the manufacturing of larger chromophores of this class of dyes that could function competitively in DSSC devices. For example, such a molecular engineering strategy might include the substitution of long alkyl chains (hydrophobic groups) to the core chemical 36


Page 37 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


motif of the best performing dye assessed herein; hydrophobic groups are well known to improve the photovoltaic performance of DSSC dyes as they suppress unfavorable dye aggregation.18 On the other hand, and perhaps more importantly, the results from DSSC device tests on 1-4 could be used to relate the structures of 1-4 to their photovoltaic properties. The generation of such structure-functionality relationships for this class of dyes presents a more informed opportunity to engineer larger chromophores for this class of dyes from these molecular building blocks, using systematic knowledge-based rules rather than empirical chemical intuition as a guide.

Figure 10. J-V curves for DSSCs containing 1, 2, 3, or 4.

Table 6 Photovoltaic performance parameters for DSSCs based on 1-4 37



Page 38 of 50

and N719. Dye

VOC [V]

JSC [mA/cm2]

FF [%]

η [%]

ηdye : ηN719 [%]

1

0.516±0.005

2.88±0.47

59.6±1.4

0.89±0.13

23.8

2

0.540±0.009

3.43±0.26

60.7±1.4

1.13±0.07

30.2

3

0.510±0.003

2.03±0.10

59.7±2.2

0.62±0.04

16.6

4

0.540±0.006

2.71±0.35

61.4±1.7

0.90±0.08

24.1

N719

0.593±0.013

9.68±0.30

65.3±3.7

3.74±0.04

100.0

A series of unsealed DSSCs was prepared using thin films of TiO2 (thickness: 4 µm) that should be compatible with the high extinction coefficients of 1-4 and minimal electron-recombination effects. Figure 10 shows the J-V curves for these DSSC devices, while the associated data are summarized in Table 6. Their power-conversion efficiencies, η, were normalized against that of N719, which is a Ru-based industrial standard DSSC dye that was fabricated and tested under the same conditions as 1-4. This normalization is important, as interpretations based on the ηdye: ηN719 ratio, rather than simply on ηdye, avoid the variability that is common in absolute DSSC device reporting. 17

Since 1-4 represent molecular building blocks of this new class of DSSC dyes, DSSC devices based on 1-4 cannot compete with those based on N719. This is also anticipated given that 1-4 are organic dyes and so they 38


Page 39 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


display lower current densities than Ru-based N719, owing to their narrow absorption bands and strong electron recombination. However, a comparison of the individual DSSC device photovoltaic properties of 1-4 can be readily made.

The dye with the best overall performance was 2 (η2 : ηN719 ~ 30%). The open-circuit voltage (VOC) values of 2 and 4 (with carboxylic group as anchor) are apparently higher than those of 1 and 3 (with cyanoacrylic group as anchor), indicating that the anchoring group in these dyes dominates the effect on VOC. The finding that the dyes with a cyanoacrylic acid group show lower VOC values may be due to the presence of cyano group which decreases the electron injection ability of the anchoring group. 40

The short-circuit current density, JSC, shows the same trend as Δγ, i.e.,

2 > 1 > 4 > 3, which corroborates the notion that the pronounced twist of the dialkylaminophenyl moiety blocks electron back transfer from the acceptor to the donor in the ES, and thus suppresses the recombination of electrons from the CB of TiO2 to the donor group of the dye. This diodelike effect represents a structure-functionality relationship that is characteristic for this class of dyes. Its encoding into a knowledge-based rule for the molecular engineering of derivatives of 1-4 should be highly advantageous in helping to produce superior photovoltaic properties in DSSC devices, while preventing undesirable electron-recombination 39



Page 40 of 50

effects.

4 CONCLUSIONS

In summary, we have investigated a new class of organic dyes for DSSC applications, which had previously been identified by a large-scale datamining approach.17 For that purpose, four subject dyes, 1-4, that possess a D-π-A-πphenyl-Aanc core structure and different anchor (cyanoacrylate, carboxylic acid) and donor (dimethylamino, diphenylamino) groups were characterized experimentally and computationally. All four dyes revealed a broad absorption in the visible region (473-503 nm). The energy levels of the HOMOs and LUMOs of 1-4 were obtained from UV/vis spectroscopy and cyclic voltammetry. These measurements revealed that the HOMO and LUMO energies of 1-4 are well aligned with the CB of TiO2 and the redox potential of the electrolyte (I-/I3-), which should enable electron injection and dye regeneration in DSSC applications. Moreover, the HOMO energies follow the trend 2 > 1 > 4 > 3, which suggests that 2 exhibits the largest driving force for dye regeneration. The ease of dye oxidation, determined by cyclic voltammetry, follows the same order. DFT and TD-DFT methods provided an insight into the electronic and geometric structures of 1-4 and their adsorption modes on the TiO2 surfaces. Theoretical calculations on the ES of 1-4 showed that a twist of the 40


Page 41 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


dialkylaminophenyl moiety explains the Stokes shifts observed in the emission

spectra

of

1-4.

Moreover,

we

discovered

that

the

dialkylaminophenyl moiety in 2 exhibits the largest torsional change upon photoexcitation, and that the extent of this twist in these dyes follows the previously established order (2 > 1 > 4 > 3). DSSC device photovoltaic performance tests supported this anticipated structural change and revealed a direct relationship between this photo-induced torsional change and the photovoltaic output. DSSC devices sensitized with 2 furnished the highest open-circuit voltage (VOC = 0.54 V) and power conversion efficiency (η2 : ηN719 ~ 30%) among 1-4, which is due to the diminished electron back transfer from the acceptor to the donor as a result of the large torsional change in its dialkylaminophenyl moiety. Moreover, the short-circuit current density, Jsc, values also follow the same order 2 > 1 > 4 > 3 as observed for . This newly discovered structure-property relationship should now be exploited to create larger chromophores for this class of dyes, which should exhibit a superior photovoltaic performance with good dye-regeneration characteristics, while precluding undesirable electron recombination via their intrinsic diode-like effect as a knowledge-based rule within the molecular engineering strategy for these dyes.

5 ACKNOWLEDGEMENTS J. M. C. is grateful to the 1851 Royal Commission for the 2014 Design 41



Page 42 of 50

Fellowship, and Argonne National Laboratory where work done was supported by DOE Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357. Y. G. thanks the Cambridge Overseas Trust for a PhD scholarship. The EPSRC UK National Service for Computational Chemistry Software (NSCCS), based at Imperial College London, and contributions from its staff, are also acknowledged for supporting this work.

ASSOCIATED CONTENT The Supporting Information is available free of charge on the ACS Publications website at DOI: XXXX. Synthetic details, (TD)DFT calculations for 1-4 (electron-density distributions of the HOMO-1 of 1-4, juxtaposition of the optimized structures of 1-4 in the ground and excited state), coordinates of the optimized structures of 1-4 in methanol, coordinates of the optimized structures of 1 or 3 adsorbed onto TiO 2, coordinates of the optimized structures of 1 adsorbed onto TiO 2, coordinates of the optimized structures of 1 adsorbed onto TiO 2.

6. REFERENCES

42


Page 43 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. O'Regan, B.; Grätzel, M., A Low-Cost, High-Efficiency Solar Cell Based on Dye-Sensitized Colloidal TiO2 Films. Nature 1991, 353, 737-740. 2. Kakiage, K.; Aoyama, Y.; Yano, T.; Oya, K.; Fujisawa, J.-i.; Hanaya, M., Highly-Efficient Dye-Sensitized Solar Cells with Collaborative Sensitization by Silyl-Anchor and Carboxy-Anchor Dyes. Chemical Communications 2015, 51, 15894-15897. 3. Chen, C.-H.; Hsu, Y.-C.; Chou, H.-H.; Thomas, K. R. J.; Lin, J. T.; Hsu, C.-P., Dipolar Compounds Containing Fluorene and a Heteroaromatic Ring as the Conjugating Bridge for High-Performance Dye-Sensitized Solar Cells. Chemistry – A European Journal 2010, 16, 3184-3193. 4. Yella, A.; Humphry-Baker, R.; Curchod, B. F. E.; Ashari Astani, N.; Teuscher, J.; Polander, L. E.; Mathew, S.; Moser, J.-E.; Tavernelli, I.; Rothlisberger, U.; Grätzel, M.; Nazeeruddin, M. K.; Frey, J., Molecular Engineering of a Fluorene Donor for Dye-Sensitized Solar Cells. Chemistry of Materials 2013, 25, 2733-2739. 5. Hara, K.; Kurashige, M.; Dan-oh, Y.; Kasada, C.; Shinpo, A.; Suga, S.; Sayama, K.; Arakawa, H., Design of New Coumarin Dyes Having Thiophene Moieties for Highly Efficient Organic-Dye-Sensitized Solar Cells. New Journal of Chemistry 2003, 27, 783-785. 6. Wang, Z. S.; Cui, Y.; Hara, K.; Dan-oh, Y.; Kasada, C.; Shinpo, A., A High-Light-Harvesting-Efficiency

Coumarin Dye

for Stable Dye-

Sensitized Solar Cells. Advanced Materials 2007, 19, 1138-1141. 43



Page 44 of 50

7. Horiuchi, T.; Miura, H.; Sumioka, K.; Uchida, S., High Efficiency of Dye-Sensitized Solar Cells Based on Metal-Free Indoline Dyes. Journal of the American Chemical Society 2004, 126, 12218-12219. 8. Kim, J. Y.; Kim, Y. H.; Kim, Y. S., Indoline Dyes with Various Acceptors for Dye-Sensitized Solar Cells. Current Applied Physics 2011, 11, S117-S121. 9. Sayama, K.; Tsukagoshi, S.; Mori, T.; Hara, K.; Ohga, Y.; Shinpou, A.; Abe, Y.; Suga, S.; Arakawa, H., Efficient Sensitization of Nanocrystalline TiO2 Films with Cyanine and Merocyanine Organic Dyes. Solar Energy Materials and Solar Cells 2003, 80, 47-71. 10. Ma, X.; Hua, J.; Wu, W.; Jin, Y.; Meng, F.; Zhan, W.; Tian, H., A HighEfficiency Cyanine Dye for Dye-Sensitized Solar Cells. Tetrahedron 2008, 64, 345-350. 11. Wang, Z.-S.; Koumura, N.; Cui, Y.; Takahashi, M.; Sekiguchi, H.; Mori, A.; Kubo, T.; Furube, A.; Hara, K., Hexylthiophene-Functionalized Carbazole Dyes for Efficient Molecular Photovoltaics: Tuning of SolarCell Performance by Structural Modification. Chemistry of Materials 2008, 20, 3993-4003. 12. Hara, K.; Wang, Z.-S.; Cui, Y.; Furube, A.; Koumura, N., Long-Term Stability of Organic-Dye-Sensitized Solar Cells Based on an AlkylFunctionalized Carbazole Dye. Energy & Environmental Science 2009, 2, 1109-1114. 44


Page 45 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


13. Kitamura, T.; Ikeda, M.; Shigaki, K.; Inoue, T.; Anderson, N. A.; Ai, X.; Lian, T.; Yanagida, S., Phenyl-Conjugated Oligoene Sensitizers for TiO2 Solar Cells. Chemistry of Materials 2004, 16, 1806-1812. 14. Chen, B.-S.; Chen, D.-Y.; Chen, C.-L.; Hsu, C.-W.; Hsu, H.-C.; Wu, K.-L.; Liu, S.-H.; Chou, P.-T.; Chi, Y., Donor-Acceptor Dyes with Fluorine Substituted Phenylene Spacer for Dye-Sensitized Solar Cells. Journal of Materials Chemistry 2011, 21, 1937-1945. 15. Haid, S.; Marszalek, M.; Mishra, A.; Wielopolski, M.; Teuscher, J.; Moser, J.-E.; Humphry-Baker, R.; Zakeeruddin, S. M.; Grätzel, M.; Bäuerle, P., Significant Improvement of Dye-Sensitized Solar Cell Performance by Small Structural Modification in π-Conjugated Donor– Acceptor Dyes. Advanced Functional Materials 2012, 22, 1291-1302. 16. Zhang, L.; Cole, J. M., Anchoring Groups for Dye-Sensitized Solar Cells. ACS Applied Materials & Interfaces 2015, 7, 3427-3455. 17. Cole, J. M.; Low, K. S.; Ozoe, H.; Stathi, P.; Kitamura, C.; Kurata, H.; Rudolf, P.; Kawase, T., Data Mining with Molecular Design Rules Identifies New Class of Dyes for Dye-Sensitised Solar Cells. Physical Chemistry Chemical Physics 2014, 16, 26684-26690. 18. Zhang, L.; Cole, J. M., Dye Aggregation in Dye-Sensitized Solar Cells. Journal of Materials Chemistry A 2017, 5, 19541-19559. 19. Kumagai, T.; Anki, T.; Ebi, T.; Konishi, A.; Matsumoto, K.; Kurata, H.; Kubo, T.; Katsumoto, K.; Kitamura, C.; Kawase, T., An Effective 45



Page 46 of 50

Synthesis of N,N-Dimethylamides From Carboxylic Acids and a New Route

from

N,N-Dimethylamides

to

1,2-Diaryl-1,2-Diketones.

Tetrahedron 2010, 66, 8968-8973. 20. Schiemenz, G. P.; Stein, G., Aromatische Phosphine Mit Substituenten Zweiter Ordnung—VII. Tetrahedron 1970, 26, 2007-2026. 21. Hagberg, D. P.; Marinado, T.; Karlsson, K. M.; Nonomura, K.; Qin, P.; Boschloo, G.; Brinck, T.; Hagfeldt, A.; Sun, L., Tuning the HOMO and LUMO Energy Levels of Organic Chromophores for Dye Sensitized Solar Cells. The Journal of Organic Chemistry 2007, 72, 9550-9556. 22. Lee, C.; Yang, W.; Parr, R. G., Development of the Colle-Salvetti Correlation-Energy Formula into a Functional of the Electron Density. Physical Review B 1988, 37, 785-789. 23. Becke, A. D., Density‐Functional Thermochemistry. III. The Role of Exact Exchange. The Journal of Chemical Physics 1993, 98, 5648-5652. 24. Hariharan, P. C.; Pople, J. A., The Influence of Polarization Functions on Molecular Orbital Hydrogenation Energies. Theoretica chimica acta 1973, 28, 213-222. 25. Frisch, M. J. T., G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; 46


Page 47 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö ; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian, Inc., Wallingford CT, Gaussian, version Gaussian.09W (accessed 2009). 26. Miertuš, S.; Scrocco, E.; Tomasi, J., Electrostatic interaction of a solute with a continuum. A direct utilizaion of AB initio molecular potentials for the prevision of solvent effects. Chemical Physics 1981, 55, 117-129. 27. Chipman, D. M., Reaction Field Treatment of Charge Penetration. The Journal of Chemical Physics 2000, 112, 5558-5565. 28. Sanchez-de-Armas, R.; San Miguel, M. A.; Oviedo, J.; Sanz, J. F., Coumarin Derivatives for Dye Sensitized Solar Cells: a TD-DFT Study. Physical Chemistry Chemical Physics 2012, 14, 225-233. 29. Pastore, M.; De Angelis, F., Aggregation of Organic Dyes on TiO 2 in Dye-Sensitized Solar Cells Models: An ab Initio Investigation. ACS Nano 2010, 4, 556-562. 30. Agrawal, S.; Pastore, M.; Marotta, G.; Reddy, M. A.; Chandrasekharam, 47



Page 48 of 50

M.; De Angelis, F., Optical Properties and Aggregation of PhenothiazineBased Dye-Sensitizers for Solar Cells Applications: A Combined Experimental and Computational Investigation. The Journal of Physical Chemistry C 2013, 117, 9613-9622. 31. Marotta, G.; Reddy, M. A.; Singh, S. P.; Islam, A.; Han, L.; De Angelis, F.; Pastore, M.; Chandrasekharam, M., Novel Carbazole-Phenothiazine Dyads for Dye-Sensitized Solar Cells: A Combined Experimental and Theoretical Study. ACS Applied Materials & Interfaces 2013, 5, 9635-9647. 32. Zhang, L.; Cole, J. M., Adsorption Properties of p-Methyl Red Monomeric-to-Pentameric Dye Aggregates on Anatase (101) Titania Surfaces: First-Principles Calculations of Dye/TiO 2 Photoanode Interfaces for Dye-Sensitized Solar Cells. ACS Applied Materials & Interfaces 2014, 6, 15760-15766. 33. Hagfeldt, A.; Boschloo, G.; Sun, L.; Kloo, L.; Pettersson, H., DyeSensitized Solar Cells. Chemical Reviews 2010, 110, 6595-6663. 34. Galoppini, E., Linkers for Anchoring Sensitizers to Semiconductor Nanoparticles. Coordination Chemistry Reviews 2004, 248, 1283-1297. 35. Zhang, F.; Ma, W.; Jiao, Y.; Wang, J.; Shan, X.; Li, H.; Lu, X.; Meng, S., Precise Identification and Manipulation of Adsorption Geometry of Donor−π–Acceptor Dye on Nanocrystalline TiO2 Films for Improved Photovoltaics. ACS Applied Materials & Interfaces 2014, 6, 22359-22369. 36. Jiao, Y.; Zhang, F.; Grätzel, M.; Meng, S., Structure–Property Relations 48


Page 49 of 50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


in All-Organic Dye-Sensitized Solar Cells. Advanced Functional Materials 2013, 23, 424-429. 37. Cole, J. M.; Blood-Forsythe, M. A.; Lin, T.-C.; Pattison, P.; Gong, Y.; Vázquez-Mayagoitia, Á.; Waddell, P. G.; Zhang, L.; Koumura, N.; Mori, S.,

Discovery

of

S···CN

Thiophenylcyanoacrylate-Based

Intramolecular Dye:

Realizing

Bonding Charge

in

a

Transfer

Pathways and Dye···TiO2 Anchoring Characteristics for Dye-Sensitized Solar Cells. ACS Applied Materials & Interfaces 2017, 9, 25952-25961. 38. Zhao, Y.; Truhlar, D. G., The M06 Suite of Density Functionals for Main Group Thermochemistry, Thermochemical Kinetics, Noncovalent Interactions, Excited

States, and Transition Elements: Two New

Functionals and Systematic Testing of Four M06-Class Functionals and 12 Other Functionals. Theoretical Chemistry Accounts 2008, 120, 215-241. 39. Allen, F. H.; Kennard, O.; Watson, D. G.; Brammer, L.; Orpen, A. G.; Taylor, R., Tables of Bond Lengths Determined by X-ray and Neutron Diffraction. Part 1. Bond Lengths in Organic Compounds. Journal of the Chemical Society, Perkin Transactions 2 1987, 0, S1-S19. 40. Chen, W.-C.; Nachimuthu, S.; Jiang, J.-C., Revealing the Influence of Cyano in Anchoring Groups of Organic Dyes on Adsorption Stability and Photovoltaic Properties for Dye-Sensitized Solar Cells. Scientific Reports 2017, 7, 4979.

49



Page 50 of 50

Graphical Abstract

50


Photoexcited Phenyl Ring Twisting in Quinodimethane Dyes

Recommend Documents