Theoretical study on Ru2+, Cu+, and Fe2+ complexes toward the application in dye sensitized solar cell

Nguyen Ngoc Ha Department of Chemistry, Hanoi National University of Education, 136 XuanThuy Street, Hanoi 10000, Vietnam Mai Anh Tuan International Training Institute for Materials Science, Hanoi University of Science and Technology, No. 1, Dai Co Viet Road, Hanoi 10000, Vietnam Dang Xuan Thu Department of Chemistry, Hanoi National University of Education, 136 XuanThuy Street, Hanoi 10000, Vietnam Luong T. Thu Thuy Department of Chemistry, Hanoi National University of Education, 136 XuanThuy Street, Hanoi 10000, Vietnam e-mail: thuyltt@hnue.edu.vn Theoretical Study on Ru21, Cu1, and Fe21 Complexes Toward the Application in Dye Sensitized Solar Cell This paper reports the application of the Ru2þ, Cuþ, and Fe2þ complexes in form of RuL2(SCN)2, CuL2(SCN)2  for dye-sensitized solar cell (DSSC) development. The calcula- tion results, given by quantum chemistry, demonstrated that the complex containing cop- per is more suitable than the one containing iron. The modification of Cu(I) complex by using various numbers of ligands enhanced photon absorption capacity as well as the absorption range. The addition of an organic ligand such as an electron attraction group to the benzene ring gave a better result as compared to the inorganic ones. Based on the analysis conducted, CuM2(SCN)2  is considered as potential material for N3 replacement. [DOI: 10.1115/1.4028582] Keywords: dye-sensitized solar cells, ruthenium(II) complex, photon absorption 1 Introduction DSSCs, since the first proposed by O’Regan and Gr€atzel [1], have been widely investigated owing to their advantages of low cost, and ease of fabrication from stable and abundant resource materials. This is owing to the durability of the dye-sensitized cell together with the ability to absorb more radiation per area as com- pared to those obtained by conventional silicon-based cells, as well as the possibility of flexible cells, which are able to be imple- mented for various applications that are not applicable in rigid silicon cells [2–7]. A typical DSSC consists of a nanocrystalline TiO2 electrode modified with a dye prepared on a transparent conducting oxide, a metal counter electrode, and an electrolyte solution redox couple between the electrodes, Fig. 1. The dye works as a photon adsorp- tion and, in corporation with the TiO2 layer, converts the solar radiation into electron current. Scientists have made great efforts of investigation of the photon dye sensitized materials for better photon absorption, ease of fabrication, and lowering the cost. The mechanism of a DSSC is described in four main periods: (1) Dyeþ h! Dye* (excited state), in this phase, electron from highest occupied molecular orbital (HOMO) is excited into lowest unoccupied molecular orbital (LUMO) of the dye. (2) Dye*! Dyeþþ e (TiO2), this is a stage that electron is transfer from Dye* to TiO2 layer. (3) Dye þþ 3/2 I! Dyeþ 1/2 I3 is a stage that electrons jump from I to Dyeþ and (4) 1/2 I3 þ e (Pt) ! 3/2 I is a stage where electron is transported to I3 from plati- num cathode. In DSSC, the ruthenium based complexes have been widely used as photon sensitized dye because of its high photon energy conversion rate and stability. The RuL2(SCN)2 complexes (where L stands for bipyridine), cis-di(thiocyanato)-bis[2,20-bipyridyl- 4,40-dicarboxylic acid] ruthenium(II) denoted as N3, operates by the conversion process Ru2þ to Ru3þ (Ru2þ e $ Ru3þ). Thus, one of the conditions that makes a complex become photon sensi- tive agent is that its HOMO must be basically built from the atomic orbital (AO) of Ru2þ. In addition, the complex should absorb the solar radiation from near violet and visible range. Ru2þ is one of a typical photon sensitized material (with the conversion efficiency great than 10%, [8]) and popular but the disadvantage of the N3 material is that it is highly expensive. Replacing Ru2þ by alternative and low cost metal ions, Mnþ is one of the solutions that brought the DSSC to mainstream use. The following require- ments required for the replacing ion metals: (i) convertible Mnþ e$M(nþ1)þ (ii) hOMO of the complex is mainly formed from the AO of the Mnþ This work aims at replacing the rare-earth Ru2þ in N3 by alter- native Cuþ and Fe2þions. Other organic ligands slightly changed but SCN remained. 2 Models and Computational Methods All calculation (energy, structure optimization in ground state) has been calculated by density functional theory (DFT) in the GGA (Generalized Gradient Approximation)/PBE (Perdew- Burke-Ernzerhof) form; the DZP (Double Zeta Polarization) basis set and the norm-conserving pseudopotentials [9] were used for all the atoms. The plane wave cutoff of 150 Ryd for the grid has been chosen for all calculations. Fig. 1 Structure and operation mechanism of a DSSC Contributed by the Solar Energy Division of ASME for publication in the JOURNAL OF SOLAR ENERGY ENGINEERING: INCLUDING WIND ENERGY AND BUILDING ENERGY CONSERVATION. Manuscript received February 20, 2014; final manuscript received August 24, 2014; published online September 30, 2014. Assoc. Editor: Santiago Silvestre. Journal of Solar Energy Engineering APRIL 2015, Vol. 137 / 021006-1CopyrightVC 2015 by ASME Downloaded From: on 11/27/2014 Terms of Use: From optimized geometries, the time-dependent density func- tional theory (TDDFT) method using a frequency-domain CASIDA model was employed to calculate UV–Vis spectra (in vacuum), and the CASIDA formalism is one of the most popular frequency-domain DFT methods. For a finite system, the polariz- ability must be given by aij xð Þ ¼ X I fI;ij x2I  x2 (1) where the I index labels different excited states; x2I and fI;ij are the corresponding excitation energy and oscillator strength. For adia- batic XC kernels, Casida deduced that the square of the excitation energies were all eigenvalues of a specific matrix equation XFI ¼ x2I FI (2) While the oscillator strengths could be deduced from the eigen- vectors FI of the same equation. This matrix X is defined in the space of electron-hole excita- tions, namely, combinations of one occupied and one unoccupied wave function, for one spin channel. The TDDFT simulations were run using the OCTOPUS code [10] with the adiabatic local density approximation. For these simula- tions, the geometrical functions were represented on a grid with a spacing parameter of 0.175 A˚ and a radius of 4 A˚ for each atom. The unified pseudopotential format pseudo-potentials were used for all elements. The numbers of occupied and virtual orbitals used in our CASIDA calculations are 20 and 100, respectively. Mayer bond order was calculated as BAB ¼ X a2A X b2B ½ðPSÞabðPSÞba þ ðPsSÞabðPsSÞba (3) Here, P and Ps are total density and spin-density matrices, respec- tively; S is overlap integral. By using these tools, the calculation would predict the compatibility of the alternative materials to replace the Ru2þ in dye complex. 3 Results and Discussion 3.1 RuL2(SCN)2 Complex. As N3 was confirmed by the manufacturers, the investigation of electron structure of N3 is helpful for further study using different ion metal complexes. Figure 2 shows that the main peaks at 410 and 510 nm of N3 given by computational calculation are close to those that were obtained by experimental measurement (at 405 and 550 nm, respectively), [11]. The main peak at 410 nm is caused by electron transfer from HOMO ! LUMO when N3 absorbs the high inten- sity photons in near UV region (Fig. 2(a)). As seen in Fig. 2(b), the HOMO was mainly formed from AO of Ru2þ and SCN. The HOMO analysis proves that this wave function was built from 4dxy and 4dyz AOs of Ru 2þ with corresponding ratio of 0.18 and 0.53. The total square of these two values makes 0.3133, meaning that 4dxy and 4dyz AOs of Ru 2þ contribute to 31.33% probability of the HOMO. The rest belongs to contribution of other AOs in N3 (191 AO). This calculation led to a confirmation that the HOMO was mainly formed from Ru2þ. Using the Mayer method [12], the analysis of the bond order found that the Ru-N bond order of SCN was 0.78, and that of pyridine was 0.72. This result confirms that SCN ligand strongly links with Ru2þ ion by means of chemical bond. The Ru-N bond length of SCN (1.99 A˚) is comparable with that of pyridine (2.01 A˚) matching with the bond order given by Mayer calculation (the longer bond, the smaller bond order). The above calculations indicate that all single Ru-N bonds were formed from the donor–acceptor bonds, which were created by the combination between the lone (electron) pair of N and empty, hybrid AO d2sp3 of Ru2þ. Figure 3 represents the density of state (DOS) of AOs dxy, dyz, and other AOs of Ru2þ. Below the Fermi level, the DOS of dyz is greater than that of dxy. This means that the probability of the electron’s presence in dyz AO of Ru 2þ dominates (note that, because of the known problem of LDA and GGA, the underesti- mation of GGA functional used, Fig. 3 appears an artificial gap- less electronic states). Thus, in N3, the electron transfer from HOMO ! LUMO occurred mainly by the contribution of dyz electron of Ru2þ. 3.2 CuL2(SCN)2 2 and FeL2(SCN)2—the Complex Modification Investigation. Replacing Ru2þ in RuL2(SCN)2 by Cuþ and Fe2þ creates a new—CuL2(SCN)2  and FeL2(SCN)2. The Fig. 2 (a) The computational UV–VIS spectra; (b) HOMO; and (c) LUMO of N3 Fig. 3 DOS of AO d (of Ru21) in N3 021006-2 / Vol. 137, APRIL 2015 Transactions of the ASME Downloaded From: on 11/27/2014 Terms of Use: corresponding results on UV–VIS, bond order, and bond length were obtained when using the similar investigation tools and methods as implemented in RuL2(SCN)2. The comparison of the results achieved in RuL2(SCN)2, CuL2(SCN)2 , and FeL2(SCN)2 may give a general conclusion of compatibility of the alternative metal ion Feþ2, Cuþ2 in new complexes for DSSC application. In general, the conversion efficiency of solar cell is proportional with the photon absorption intensity of the dye. In Fig. 4, the pho- ton absorption intensity of CuL2(SCN)2  is much better than that of FeL2(SCN)2 but smaller than that of RuL2(SCN)2. The photon absorption of FeL2(SCN)2 complex is effective in near UV region while that of CuL2(SCN)2  represents the absorption in both UV and visible range (at 550 nm). This result has significant meaning in improving the conversion efficiency of the cell compared to FeL2(SCN)2 complex. The results obtained lead to the conclusion that the CuL2(SCN)2  complex is much more effective than of FeL2(SCN)2 for application in DSSC. Thus, the paper will focus on the results of CuL2(SCN)2 . Figure 5 shows that the HOMO of CuL2(SCN)2  was mainly formed from the AOs of Cuþ (and SCN) is similar to N3. How- ever, the LUMO is found in one benzene ring of pyridine. Thus, the geometry arrangement in complex-TiO2 binding is also essen- tial for the electron transfer from LUMO of the complex into TiO2 of the DSSC. The effectiveness of an attachment between benzene ring and TiO2 complex is always below 100% which reduces the conversion efficiency of CuL2(SCN)2  based DSSC. The (calculated) bond order for the two bonds Cu-N of bipyri- dine was 0.001 and 0.255 with corresponding Cu and N length at 2.38 and 2.13 A˚, respectively. The Cu-N bond order of SCN was 0.469 and the bond length was 2.05 A˚. So, Cuþ in CuL2(SCN)2  formed coordinated number 4. Different from Ru2þ with 4d65s05p0 configuration (forming the low spin complex N3, d2sp3 hybridization, and coordinated number 6), Cuþ with 3d104s04p0 configuration forming sp3 hybridization, creating coordinated number 4 in the complex. The Cu-N bond order is smaller, in particular, the Cu-N bond of pyridine. Comparing with Ru-N bond order in N3, it is found that the CuL2(SCN)2  is not thermodynamically stable as RuL2(SCN)2. This can be considered as a drawback of Cu þ com- paring with Ru2þ. The analysis results of HOMO and bond order also revealed the important role of the ligands (SCN in N3) play in the photon sen- sitized complex and has led to new approach of improving the photon sensitive efficiency by modification of the inorganic ligands. In this work, several elements were studied to investigate the SCN replacing possibility by F, Cl, Br, CN, and H2O. Table 1 shows that the bond order decreases for the halogen series (F, Cl, and Br). This can be interpreted by the increase in radius of X (X¼ F, Cl, Br) in Cu-X bond and consequently led to decrease in the effective interaction to form the Cuþ complex. The photon absorption intensity, from F to Br, was observed to have drop and the maximum wavelength too shrunk (approaching the near UV). Concerning the CN ligand, the bond order of Cu and C is con- siderable (0.63), and the photon absorption intensity of the form- ing complex is higher than that of halide complex but less than the photon absorption intensity of SCN complex (0.17). Replac- ing SCN by water will create a complex that is not sensitive to the radiation in near UV and visible range. This means that mate- rial is not suitable for dye synthesis in DSSC. From the Table 1, it is found that SCN is most suitable candidate for photon sensi- tized complex development as compared to F, Cl, Br, CN, and H2O ligands. Implementing the similar experiments, the functional group– COOH of pyridine was replaced by Cl, OH, CH3, and CF3. The results were listed in Table 2. In Table 2, the COOH replacement by other groups did not change the bond order of Cu-N (pyridine). The reason is that the Cu-N bond length is almost unchanged (2.13 A˚). However, the electron structure was varied by different electron repulsive/ attractive capacity of different groups versus the benzene ring, especially in the two following cases: þ -OH groups strongly push electrons into benzene ring (by þC effect). þ -CF3 groups pull electrons out of benzene ring (I, H effect), similar to COOH group (electron attraction by C effect). For Cl (I, þC effect) and CH3 (þI, þH effect), the elec- tron repulsive/attractive capacity versus the benzene ring is fairly weak. Thus, the electron attraction groups like CF3 and COOH, when immobilized to the benzene ring of pyridine will enhance the photon absorption intensity and the best absorption was found in near UV range. Fig. 4 The computational UV–VIS spectra of FeL2(SCN)2 and CuL2(SCN)2 2 Fig. 5 The computational HOMO and LUMO of CuL2(SCN)22 Journal of Solar Energy Engineering APRIL 2015, Vol. 137 / 021006-3 Downloaded From: on 11/27/2014 Terms of Use: The replacement of COOH by Cl, OH, CH3, and CF3 did not significantly improve the photon absorption intensity and maximum wavelength (0.17 and 373 nm for COOH), but the result suggests that binding the electron attraction groups to the benzene ring will enhance the photon absorption intensity, and hence improve the performance of DSSC [13,14]. Moreover, the photon absorption capacity of CF3 based complex is comparable to that of COOH with the wavelength shifted to the red (383 nm as compared to 373 nm in COOH). This finding is also essential because the IR and visible range cover the majority of the solar energy. Following the above discussion, an electron attraction group, CH¼O (with C effect), was added at the –o position of the benzene ring in CuL2(SCN)2  complex with the hope that will improve the photon absorption capacity with the wavelength shifted to the red band (Fig. 6). The modified complex is denoted as CuM2(SCN)2 . An amazing result was obtained. The photon absorption capacity is considerably increased (0.19; higher than 0.17 of CuL2(SCN)2  and comparable to 0.20 of N3), and espe- cially, the maximum wavelength was found at 460 nm (373 nm for CuL2(SCN)2  and 410 for N3). It can also be seen, in Fig. 6(b), that the HOMO of CuM2(SCN)2  complex almost formed from the AOs of Cuþ and SCN ligand. 4 Conclusion This work investigated the electron structure detail of N3, the possible factors influenced the conversion efficiency of solar radi- ation into electricity in N3. The obtained results are fundamentals for further research on different complexes. The calculation proved that CuL2(SCN)2  is more suitable than FeL2(SCN)2 complex in development of dye sensitized solar cell due to the photon absorption capacity as well as wider absorption range. According to results obtained, although the photon absorp- tion capacity of CuL2(SCN)2  is lower than that of N3, but as it is much more cheaper than N3 and therefore considered as a poten- tial candidate for future research. The modification of Cu(I) complex by using various numbers of ligands are to enhance the photon absorption capacity as well as the absorption range can be predicted by the quantum chemis- try calculation. By altering the organic ligand, adding an electron attraction group to the benzene ring, for instant, bring a better result as compared to the inorganic ligand. According to achieved analysis, CuM2(SCN)2  is considered as potential material for N3 replacement. Further empirical research will be conducted to verify this conclusion. Acknowledgment This work was financially supported by the Vietnamese National Foundation for Science and Technology Development (NAFOSTED) for a basic research Project No. (104.99-2011.44). Reference [1] O’Regan, B., and Gr€atzel, M., 1991, A low-cost, “High-Efficiency Solar Cell Based on Dye-Sensitized Colloidal TiO2 Films,” Nature, 353(6346), pp. 737–740. [2] Robertson, N., 2006, “Optimizing Dyes for Dye-Sensitized Solar Cells,” Angew. Chem. Int. Ed. Engl., 45(15), pp. 2338–2345. [3] Hamann, T. W., Jensen, R. A., Martinson, A. B. F., van Ryswyk, H., and Hupp, J. T., 2008, “Advancing Beyond Current Generation Dye-Sensitized Solar Cells,” Energy Environ. Sci., 1(1), pp. 66–78. [4] Gr€atzel, M., 2009, “Recent Advances in Sensitized Mesoscopic Solar Cells,” Acc. Chem. Res., 42(11), pp. 1788–1798. [5] Wu, J. H., Lan, Z., Lin, J. M., Huang, M. L., Hao, S. C., Sato, T., and Yin, S., 2007, “A Novel Thermosetting Gel Electrolyte for Stable Quasi-Solid-State Dye-Sensitized Solar Cells,” Adv. Mater., 19(22), pp. 4006–4011. Table 1 Intensity and absorption band of Cu(I) complex when SCN2 was replaced F Cl Br CNa H2O SCN  Maximum absorption intensity peak (arb. unit) 0.13 0.12 0.08 0.13 1.47 0.17 Maximum wavelength (nm) 551; 722 496; 672 465 477 285 550 Bond order with Cu 0.47 0.40 0.30 0.63 0.10 0.47 aCu bonded to C (instead of N) in CN to form a more stable complex (less than 76 kJ/mol). Table 2 Intensity and absorption band of Cu(I) complex when2COOH was replaced Cl OH CH3 CF3 COOH Maximum absorption intensity peak (arb. unit) 0.14 0.05 0.10 0.16 0.17 Maximum wavelength (nm) 374 504 364 383 373 Bond order of N (in pyridine) with Cu 0.40 0.40 0.40 0.40 0.47 Fig. 6 (a) The computational UV–VIS spectra and (b) HOMO of CuM2(SCN)22 021006-4 / Vol. 137, APRIL 2015 Transactions of the ASME Downloaded From: on 11/27/2014 Terms of Use: [6] Wu, J., Hao, S., Lan, Z., Lin, J., Huang, M., Huang, Y., Li, P., Yin, S., and Sato, T., 2008, “An All-Solid-State Dye-Sensitized Solar Cell-Based Poly (N-Alkyl-4-Vinyl-Pyridine Iodide) Electrolyte With Efficiency of 5.64%,” J. Am. Chem. Soc., 130(35), pp. 11568–11569. [7] Dloczik, L., Ileperuma, O., Lauermann, I., Peter, L. M., Ponomarev, E. A., Redmond, G., Shaw, N. J, and Uhlendorf, I., 1997, “Dynamic Response of Dye- Sensitized Nanocrystalline Solar Cells: Characterization by Intensity- Modulated Photocurrent Spectroscopy,” J. Phys. Chem. B, 101(49), pp. 10281–10289. [8] Jitchati, R., Thathong, Y., and Wongkhan, K., 2012, “Three Synthetic