J-Aggregates of Amphiphilic Cyanine Dyes for Dye-Sensitized Solar Cells: A Combination between Computational Chemistry and Experimental Device Physics

1 Nano-Photochemistry and Solarchemistry Labs, Department of Chemistry, Faculty of Science, Ain Shams University, Abbassia, Cairo 11566, Egypt 2 Center for Nanotechnology, School of Engineering and Applied Sciences, Nile University, Juhayna Square, Sheikh Zayed, 6th of October City, Giza 12588, Egypt 3 Nano-Photochemistry Laboratory, Environmental Studies and Research Institute, University of Sadat City (USC), Sadat City 32879, Egypt 4Department of Chemistry, Faculty of Education, Ain Shams University, Roxy, Cairo 11341, Egypt


Introduction
The design and synthesis of functional dyes have lately received much attention due to their potential applications as sensitizers in dye-sensitized solar cell (DSSC) [1].Dyes with wide absorption range, specifically in the visible region, and containing an anchoring group such as carboxylic acids, are ideal candidates [2].The DSSC process requires light-induced excitation of a HOMO electron.The anchoring group then acts as a bridge and injects this excited electron into the conduction band of TiO 2 and the dye gets oxidized.The oxidized dye is then neutralized to ground state by I 2 /I 3 − redox system [3].The efficiency of the cell is affected by three factors: the energy difference between the excited state of the dye and the conduction band of the semiconductor, the efficiency and nature of the dye-semiconductor binding, and the properties of the redox couple in the electrolyte [2].Additionally, the electron density should be localized near the injecting group in the excited state [4].It has also been shown that the redox potential of adsorbed dye depends on the pH of the electrolyte and on the potential applied to the semiconductor [5].By choosing an appropriate sensitizer, the tuning of the photoresponse of the semiconductor can be optimized [6].

International Journal of Photoenergy
The most successful charge transfer metal complexes of organic molecules employed so far in DSSC are cisdithiocyanatobis-(2,2  -bipyridyl-4,4  -dicarboxylato)ruthenium(II) (red dye) and trithiocyanato 4,4  ,4  -tricar-boxy2,2  ,6  ,2  -terpyridine ruthenium(II) (black dye), which yields overall conversion efficiencies of up to 11% under AM1.5 irradiation [7].However, ruthenium dyes are very expensive.To overcome the high cost, metal free sensitizers such as organic dyes and natural dyes are being investigated as alternative sensitizers for DSSC applications [8].Recent literature reports indicate achievement of efficiency of 9.8% using organic dyes [9].Several organic dyes have been reported in the literature [10,11].Although the main aim is to fabricate and optimize dyes of high efficiency, it is also of interest to investigate dyes from the viewpoint of fundamental research connected to issues like conjugation, charge transfer, and binding on TiO 2 surface.
The self-assembly of amphiphilic molecules into fibrous structures has been studied extensively due to their high potential for many profitable applications.Although very different in their head group chemistry, many natural as well as synthetic amphiphilic compounds derived from carbohydrates, carbocyanine dyes, or amino acids tend to form twisted fibrous structures by molecular self-assembly in water predominantly [12][13][14][15].For example, the spectral properties and photostability of the (Dye 1; Figure 1) J-aggregate have been investigated in solution and upon adsorption on TiO 2 nanoparticles [12].The self-assembly of Dye 1 has been also studied on a glass surface by noncontact atomic force microscopy (NCAFM).The dye molecules form a well-defined fiber-like structure that extends for tens of micrometers.The internal structure of the fibers has been clearly resolved and showed a number of small tubes wrapped around each other to form a helical structure [12].
As a new promising sensitizer for DSSC, here we are reporting studies on Dye 1, which exhibits interesting characteristics of self-organization and better charge transportation.It would be of interest to carry out a detailed DFT study of such a dye, to understand its donation capacity, and to shed light on the nature of interaction with TiO 2 anode.The light-to-electricity conversion efficiency () of DSSC based on that dye (Dye 1) as a metal-free dye sensitizer has been also tested.Reasonably good efficiency compared to the standard DSSC based on N719 dye sensitizer has been obtained with the solar device based on Dye 1 that indicated a good charge communication between the dye sensitizer and the TiO 2 conduction band electrode.

Materials and Device
Fabrication.The J-aggregating 5,5  -6,6  -tetrachloro-1,1  -dioctyl-3,3  -bis-(3-carboxypropyl)benzimidacarbocyanine (Dye 1) was obtained from FEW Chemicals GmbH, Wolfen, Germany, and was used as received.The absorption spectra of the free dye solution and the adsorbed dye on the TiO 2 film were already published in our previous work [12].Exactly similar findings were obtained in this work.The DSSCs were fabricated using the N719 dye and TiO 2 film electrodes included in a sample kit donated by Dyesol Company, Australia.The TiO 2 coated anodes with approximate thickness of 9 ± 1 m and effective area of = 0.16 cm 2 were sintered at 450 ∘ C for 4 hours and air-cooled before being immersed overnight in acetonitrile solutions of 2.8 × 10 −4 M of Dye 1 or N719.After the substrate was adequately washed with anhydrous alcohol and dried in moisture free air, the dye-sensitized TiO 2 electrode was obtained.A DSSC was assembled by filling an electrolyte solution (0.6 M tetrapropylammonium iodide, 0.1 M iodine, 0.1 M lithium iodide, and 0.5 M 4-tertbutylpyridine (TBP) in acetonitrile) between the dye-sensitized TiO 2 electrode and a platinized conducting glass electrode.The two electrodes were clipped together, and a sealing sheet was used as sealant to prevent the electrolyte solution from leaking.

Photocurrent-Voltage Curves Measurements.
The photovoltaic testing of Dye 1 and N719 DSSCs was carried out by measuring the - character curves using Keithley 2400 source meter and homemade software, under simulated AM 1.5 solar illumination at 100 mW cm −2 from a halogen lamp in ambient atmosphere.
The fill factor (FF) and overall light-to-electrical energy conversion efficiency () of DSSC were calculated according to the following well-known equations: where the  max is the maximum power output point in the --curve yielding maximum product of current  max (mA⋅cm −2 ) and voltage  max (V),  in is the radiation power incident on the cell (mW⋅cm −2 ),  SC is the short-circuit current density (mA⋅cm −2 ), and  OC is the open-circuit voltage (V), respectively.Profit 6 (Quantum Soft, Switzerland) was used for data analysis and handling.

Computational Details.
The geometry optimization of the positively charged molecule in its doublet ground state was carried out in vacuum as implemented in the Gaussian 09 package [16].The geometry of Dye 1 was optimized using the hybrid exchange-correlation B3LYP [16] functional with the 6-31G(d,p) basis set.The same computational method was applied to single TiO 2 molecule.Gausview 5, Gaussian Inc. (USA), was used for graphs visualization.

Molecular Interactions: Simplified View.
To investigate the structural and frontier orbitals of the interacting molecules included in the light to electricity energy conversion device, we performed DFT and time-dependent (TD) DFT calculations on simplified molecules.The results are depicted in Figures 3, 4, and 5. Optimized geometry of Dye 1 molecule studies shows that large intramolecular charge transfer takes place from the HOMO to LUMO, and the donor moiety is not coplanar with the other conjugated part of the molecule (twisted by (∼22 ∘ )) resulting in unsymmetrical orientation of the carboxylic groups, Figure 3. Furthermore, the anchoring carboxylic group is not contributing to frontier orbitals and International Journal of Photoenergy  thus seems to play a minor role in this type of molecules.This could give an idea about the correlation between the monomer molecular structure and performance in DSSC device, as discussed below.Additionally, based on its contribution to the frontier molecular orbitals (Figures 4 and 5) when interacting with TiO 2 , the carboxylic groups of Dye 1 should be excluded.

Surface Orientation of Dye 1: Topographic Imaging.
The self-assembly of amphiphilic molecules into fibrous structures has been the subject of numerous studies over past decades due to various current and promising technical applications [12][13][14].Although very different in their head group chemistry many natural as well as synthetic amphiphilic compounds derived from carbohydrates, carbocyanine dyes, or amino acids tend to form fibrous structures by molecular self-assembly in water predominantly twisted ribbons or tubes.Figure 2 shows the NCAFM image of Dye 1 aggregates adsorbed on the surface of the TiO 2 anode.One type of aggregates of well-defined fiber-like structure is clearly evident.It is horizontally aligned.This should result in better electronic charge transportation and thus a more efficient photovoltaic device.The following sections will shed more light on this expectation.

Performance of the Photovoltaic Device.
The photocurrent-voltage properties of DSSCs using either Dye 1 or N719 were measured and shown in Figure 6.The fill factors and the current-to-electricity photovoltaic efficiencies were FF = 0.69 and 0.79 and  = 3.75% and 7.6% for Dye 1 and N719 based devices, respectively.An illustrative energy level diagram (Figure 7) describes the electron charge donation from the sensitizer Dye 1 [16].Moreover, the experimental results could be explained on the basis of the DFT computations and AFM topographic image (Figure 2).The frontier orbitals of the self-assembled Dye 1 (depicted in Figure 3) do not spread over the carboxylic groups and, therefore, it is difficult to expect that these "anchoring" carboxylic groups play a significant role in the process of charge transportation to the TiO 2 conduction band.Moreover, Dye 1 molecules appear to be oriented in horizontal tube-like aggregates on the surface of the TiO 2 anode.The long-range electron transport to TiO 2 anode seems to be occurring through the terminals of the molecular helical wire cable.Although better charge transport within the long helical aggregate wire cable occurs, less effective TiO 2 anode molecules are involved in receiving electronic charge.Recent studies on similar J-aggregate molecule favor our explanation [15].Thus, it could be concluded that the orientation of J-aggregates  affects the electron transport capabilities in these artificial light-harvesting device systems.

Conclusion
Frontier orbitals of Dye 1-TiO 2 molecules have been calculated at DFT-B3LYP/6-31G+(d,p) level of theory, which exclude any significant role of the carboxylic groups in the dye anchor.The AFM topographic imaging of Dye 1 points to horizontal alignment of aggregates of Dye 1 assuring efficient charge transport through the helical wire's terminals to a limited number of receiving centers on the TiO 2 anode.Consequently, such a way of alignment of Dye 1 on the anode of DSSC leads to a lower electron relay.Contrary to our expectation, this is reflected on lower efficient light harvesting DSSC device.However, a relatively reasonable efficient DSSC device that is based on a metal-free dye sensitizer (Dye 1) has been obtained.The fill factors and the current-to-electricity photovoltaic efficiencies were FF = 0.69 and 0.79 and  = 3.75% and 7.6% for Dye 1 and N719 based devices, respectively.It is concluded that the nature of interaction with the anode and orientation of structural hierarchy dramatically affects the electron transport capabilities in these artificial light-harvesting device systems.Moreover, the anchoring carboxylic groups are not engaged in the electron transport process.

Figure 2 :
Figure 2: (a) NCAFM image of Dye 1 J-aggregates adsorbed on the TiO 2 anode.(b) The topographic profile along the line indicated in the topography image in (a).

Figure 3 :
Figure 3: DFT Optimized geometry of simplified Dye 1 (Cl atoms are not taken into account and where R group = CH 3 group).(a) Front view (180 ∘ on the plane of paper) and (b) 90 ∘ twisted view (perpendicular to the plane of the paper) reveal nonplanar molecule.

Figure 2 (
b) shows the topographic profile along the line indicated in the topography image in (a), length = 286 nm and height = 25.9 nm.Pico Image Basic software was used to build a basic surface analysis report on multilayer measurement data that is input from Agilent AFM 5500 Atomic Force Microscope (AFM) (N9410S), Agilent Technologies.Measurements were carried out via Dr. M. Ghobashy at National Center of Radiation Research, Atomic Energy Authority, Nasr City, Cairo, Egypt.

Figure 4 :
Figure 4: The (a) contour surface of HOMO and (b) presents contour surface of LUMO.Shown are the contour isosurfaces of orbital amplitude 0.02 (red) and −0.02 (green).

Figure 5 :
Figure 5: TiO 2 single molecule charge distributions in frontiers orbitals.The molecular orbitals were calculated at the TDDFT-B3LYP/6-31G+(d,p) level of theory.The upper plots present HOMO.(a) TiO 2 HOMO 3D surface representing charge density.(b) TiO 2 HOMO surface superimposed on the 2D contour of charge density and the lower plots present surface of LUMO.(c) TiO 2 LUMO 3D surface representing charge density.(d) TiO 2 LUMO surface superimposed on the 2D contour of charge density.Shown are the contour surfaces of orbital amplitude 0.02 (red) and −0.02 (green).