1 Introduction

High-efficiency dye sensitised solar cells popularly known as Gratzel cells (O’Regan 1991) were the first example of an economically competitive alternative to silicon-based solar cells. A DSSC is a photo electrochemical device which depends on the sensitisation of a wide-band semiconductor by a dye sensitizer to generate electricity. The sensitizer is one of its critical components that absorbs light energy and injects the photo-excited electron into the conduction band of the semiconductor (Zhang et al. 2014). The electronic structure of the sensitizer molecule, nature of anchoring and ancillary group and its HOMO and LUMO energy levels affect the light absorbance and electron injection, and hence the composite performance of the solar cell. As such development of metal complex dyes for use as sensitizers is an active field of research. Ruthenium polypyridyl complex dyes have been investigated extensively and dyes like N3 dye (Chiba et al. 2006; Gratzel 2004) and N719 (Nazeeruddin et al. 1993) have been reported to have reached 11% efficiency. However, their large-scale application is restricted due to high cost, limited availability and toxicity of ruthenium metal. This has motivated researchers to explore the possibility of using complexes of the first-row transition metals as inexpensive, abundant and non-toxic sensitizers in DSSCs.

2 Design Criteria for a Dye Sensitizer Molecule

An efficient dye sensitizer should have high absorbance in the visible region. Its HOMO–LUMO levels and redox potentials should appropriately match with the energy levels of the semiconductor and electrolyte for electron injection and regeneration respectively and should be able to sustain numerous redox turnovers under light irradiation (Gratzel 2001). Further, the sensitizer should be photo stable and chemically bound to the semiconductor. The MLCT transitions observed in these dye sensitizers play an important role in charge injection to TiO2 surface. Despite first-row transition metals fulfilling most of the above criteria and not presenting abundance or toxicity problems, research into first-row transition metal sensitizers has been comparatively limited. This is due to the fact that their excited states have an extremely short lifetime, probably due to crossing from charge-transfer state to closely spaced d-d excited states followed by non-radiative thermal decay (Linfoot et al. 2011). This factor limits electron injection into TiO2 leading to low efficiency.

The sensitizer design strategies implemented to enhance sensitising efficiency should, therefore, focus on extending the lifetime of the charge transfer states most probably by preventing crossing to d-d excited states. This has been reported to be achieved by either having complexes with completely filled d shells or by having a strong crystal field so as to raise the energy of d-d excited states. Based on these two strategies, several copper(I), Iron(II) and nickel(II)-based dyes have been recently investigated as sensitizers with great potential (Bozic-Weber et al. 2011). This review highlights the various strategies adopted by researchers in recent years in molecular designing of sensitizers based on these first-row transition metals so as to develop efficient, cost-effective and environment-friendly DSSCs.

3 Cu Sensitizers

Cu(I) bis (diimine) complexes have been reported to possess photo-physical properties similar to that of Ru(II) diimine complexes (Armaroli 2001; Armaroli et al. 2007). The closed-shell d10 Cu(I) dyes have MLCT excited state with considerable lifetime (Armaroli et al. 2006, 2007) resulting due to the excitation of a d electron from metal bonding orbital to ligand π* antibonding orbital. However, the MLCT excited state in ruthenium(II) and copper(I) sensitizers has rather different characteristics. In ruthenium compounds, the equilibrium geometry between the ground and excited states do not show much variation. However, in Cu(I) complexes the conformational geometry changes from distorted tetrahedral in the ground state to tetragonally flattened geometry in an excited state (Robertson 2008). Such geometric distortions slow down the electron injection into the semiconductor and allow a non-radiative relaxation of the excited state (Armaroli 2001; Armaroli et al. 2007). Due to this reason, the early investigations on homoleptic Cu(I)-based dye for DSSC by Sauvage and his co-workers (Alonso-Vante et al. 1994) in 1994 were reported to have much lower efficiency.

Investigations on a number of phenanthroline- and bipyridine-based copper(I) complexes have led researchers to the conclusion that efficient sensitizers can be developed by structural modifications of ligands to tune their photo-physical and electrochemical properties. In 2002, Sakaki and co-workers (Sakaki et al. 2002) reported a new Cu(I) dye with 5,5’ carboxy substituted bipyridyl ligands to ensure rapid electron injection with a conversion efficiency of 2.5%. This led to the conclusion that Cu(I) dyes have great potential as sensitizers for solar cells and provided evidence that molecular designing through structural changes in ligand and optimisation could improve efficiency. Despite these reports, little progress was made in Cu-based DSSCs and this remained an inactive field until 2008 when Constable, Housecroft and co-workers (Bessho et al. 2008) synthesised Cu(I) sensitizers with 6,6’-dimethyl 2,2’-bipyridyl ligands and found these dyes to be effective sensitizers for TiO2. They emphasised that the substituents at the 6 and 6’ positions in 2,2’ bipyridyl ligands are a basic structural requirement to enhance the stability of Cu(I) dyes. The stearic bulk of 6,6’ substituents effectively stabilises Cu(I) with respect to oxidation to Cu(II); thus avoids flattening distortion and produces long-lived excited states. Subsequently, many homoleptic Cu(I) bipyridyl dyes with various substituents (Constable et al. 2009; Bozic-Weber et al. 2011; Yuan et al. 2012; Colombo et al. 2013; Wills et al. 2013) were investigated and the one reported to have the highest performance up to now is the dye with 6,6’dimethyl-2,2’bipyridine-4,4’-dibenzoic acid ligands (Yuan et al. 2012). Lu et al. (2010) and Lopez et al. (2010) theoretically studied several homoleptic Cu(I)-based dyes by DFT and TDDFT approach to describe their geometries and spectral properties. They emphasised the importance of theoretical calculations in the design of new copper sensitizers with improved performances.

In homoleptic systems, the directional movement of photo-generated electrons across the dye is hampered due to high level of symmetry. This led to an emphasis on using heteroleptic Cu(I) dye sensitizers having one ligand with a group suitable for anchoring the TiO2 surface and the second ligand tuned structurally for light absorption. Thus, there is a push–pull system that enables the charge separation and provides directionality for electron injection to the TiO2 conduction band. In 2010, C. L. Linfoot et al. reported the first heteroleptic Cu(I) sensitising dye [Cu(POP)dcdpy][BF4] but with very low efficiency (Linfoot et al. 2010). The subsequent researches focused on the methodology of developing heteroleptic Cu(I) dyes and to synthesise D-π-A type sensitizers. Constable and Housecroft (2015) group developed a ‘Surfaces-as ligands, surfaces-as complexes’ strategy while Odobel et al. (Sandroni et al. 2013a, b) used ‘HETPHEN concept’ to develop stable heteroleptic Cu(I) complexes. Odobel et al. applied computational techniques such as DFT and TDDFT to elucidate the electronic structure and spectral properties of newly designed and synthesised Cu(I)-based sensitizers. In the past few years, effects of different ancillary groups (Hewat et al. 2014) and anchoring groups (Hewat et al. 2014), introduction of various substituents (Malzner et al. 2014; Brunner et al. 2015), introduction of co-adsorbents (Brauchli et al. 2014) and of incorporating hole transport motifs (Bozic–Weber et al. 2013b) on the efficiency of copper(I) DSSCs have been intensively investigated and DFT and TDDFT have been increasingly used to support the results. Detailed reviews by Housecroft (2015) and Magani (2016) provide a wealth of information on the strategy and design of heteroleptic Cu(I) sensitizers. These investigations have led to the conclusion that the development of efficient sensitizers is possible by accelerating electron injection either by using electron-releasing substituents on the ancillary ligand or by using a more rigid anchoring ligand to extend the lifetime of the excited state (Sandroni et al. 2013b).

4 Iron Sensitizers

Fe(II) polypyridyl complexes such as [Fe(dcbpy)2(CN)2] as sensitizers were first investigated by Ferrere and Gregg (1998) in 1998. Later, Meyer (Yang et al. 2002) and co-workers studied Na2[Fe(bpy)(CN)4] complexes. They observed two types of charge transfer mechanisms—(i) Direct mechanism in which an electron from the sensitising dye is directly injected in conduction band of TiO2 and (ii) Indirect mechanism involving interfacial electron transfer (IET) from initially excited charge-transfer state of dye into the TiO2 (Ferrere and Gregg 1998; Yang et al. 2002). Ghosh et al. (1998) observed a direct mechanism of charge transfer in [Fe(CN)6]4–. Iron complexes are considered to be a possible alternative to ruthenium sensitizers as it belongs to the same group as ruthenium. However, iron complexes unlike their ruthenium counterparts sensitise TiO2 very inefficiently inspite of an intense MLCT absorption. This is due to the fact that the photoactive MLCT state is extremely short-lived and gets deactivated into photo inactive metal centred (MC) states due to intersystem crossing (ISC) in sub-picoseconds (Monat and McCusker 2000; Juban et al. 2006; Smeigh et al. 2008). In short, the ultrafast ISC phenomenon triumphs over the relatively slow interfacial electron transfer leading to inefficient sensitization.

Hence, the development of Fe(II) complexes capable of IET at a faster rate as compared to ISC is an important aspect of improving sensitising efficiency. An understanding of different structural aspects that influence the rates of the ISC and IET, therefore, is the key to the development of more efficient iron-based sensitising dyes. Computational methods such as DFT and TDDFT are extremely helpful for such investigations because they enable evaluation of the effect of systematic structural modifications on light harvesting and other photo-physical properties. The efficiency of Fe(II) sensitizers can be improved by adopting basically two strategies, both of which have been explored in recent years by different research groups with considerable success. The first approach focuses on making IET faster than ISC process by increasing the rate of IET through structural modification in the ligands of sensitising dyes (Bowman et al. 2013, 2015a; Jakubikova and Bowman 2015; Nance et al. 2015). Bowman et al. (2013) studied three different dye-TiO2 systems having polypyridine scaffolds functionalised with three different electron-donating groups and concluded that electron-donating substituents may improve electron injection and hence the efficiency of Fe(II) polypyridine-based dyes (Bowman et al. 2013). Later, the same research group computationally investigated the effect of various linker groups in [Fe(bpy-L)2(CN)2] dye and determined the rate of IET in each case (Bowman et al. 2013). The molecular simulations predicted that the hydroxamate linker instead of carboxylic acid linker will have a high rate of IET and hence Fe(II) polypyridines with hydroxamate will be more efficient sensitising dyes.

The second approach towards enhancing the efficiency of these sensitising dyes focuses on slowing down the ISC process and thus increasing 3 MLCT lifetime. This was reported to be achieved by modifying the ligands in Fe(II) polypyridines to increase the ligand field strength. These modifications included changing the donor atom in ligands (Dixon et al. 2015; Mukherjee et al. 2015) or altering the ligand framework geometry (Mengel et al. 2015; Liu et al. 2015). Bowman et al. systematically modified donor atom and ligand framework to compare the relative ligand field strengths in various iron (II) dyes and concluded that formation of Fe–C bonds leads to the strongest ligand field. Liu and co-workers studied [Fe(CNC)2]2+ dye [CNC = 2,6-bis(imidazol-2-ylidene)pyridine] with strongly σ donating N-heterocyclic carbene ligands with an aim to raise the, eg orbital energy to slow down ISC. This study reported a record 3MLCT lifetime of t = 9 picoseconds (Liu et al. 2013), which was improved further by the same group to t = 26 picoseconds recently (Liu et al. 2016). Thus extending the lifetime of 3MLCT state in iron complexes is an  important strategy for improving efficiency of Fe(II) sensitizers.

5 Nickel Sensitizers

The literature on the use of nickel as sensitizers for DSSCs is scanty as some of the initially investigated Ni complex dyes reported insignificant photovoltaic activity. K. Neuthe used heteroleptic Ni(II) diimine dithiolenes as co-sensitizers (Neuthe et al. 2014) to known visible light-harvesting dyes for extending the spectral window to infra-red region. In 2011, Linfoot et al. reported [Ni(dcdpy)(qdt)] and [Ni(decbpy)(qdt)] as possible sensitizers and studied their electronic properties using DFT and TDDFT (Bozic–Weber et al. 2013a). These complexes exhibit intense visible transition between 500 and 600 nm that can be assigned to charge transfer from qdt to diimine ligand. However, the conversion efficiency is observed to be quite low in spite of satisfactory light harvesting probably due to short-lived excited state. This suggests that highly efficient Ni-based sensitizers can be designed if excited state lifetime is enhanced by suitable molecular design.

6 Conclusion

The research in first-row transition metal dye sensitizers for solar cell applications has tremendously progressed over the past decade. This review systematically summarises the innovative strategies employed by various groups for enhancing the efficiency of the dye sensitizers based on more available, cost-effective and non-toxic metals. Some complexes of Cu(I) and Fe(II) have led to promising results. However, the device performances with these sensitizers are still less than those of Ru(II) sensitizers. The development of first-row metal complexes as efficient and inexpensive sensitizers at commercial scale in the future requires important challenges to be tackled. Optimising the chemical properties to broaden the spectral coverage and enhancing photo-physical properties to enhance excited state lifetime and increasing the stability are some research areas that need continuous efforts.