Cobalt Polypyridyl Mediators in Dye-Sensitized Solar Cells (DSCs)

Dye-sensitized solar cells (DSCs) have comprised a sizeable field of research over the past two decades due to their attractive blend of efficiency and low cost.  Until the past few years, the most successful of these cells were based on the original system of Ru(II/III) polypyridyl sensitizer and I^-/I₃^- mediator in systems capped at around 10% overall efficiency.¹  Efforts since have been focused on the engineering of novel dyes and mediators to combat the prototypical system’s issues.  First, these dyes are generally expensive due to the rarity of ruthenium.  Second, the I^-/I₃^- mediator has a variety of undesirable properties: it is corrosive, highly colored, volatile, and has a fixed E_1/2.  The last of these is perhaps the most problematic, as the maximum open-circuit voltage (V_OC) in a DSC is defined as the difference between the Fermi energy of the TiO₂ and the E_1/2 of the mediator.  Thus, improvements to V_OCmust be made elsewhere.

In 2011, Grätzel and coworkers reported a new record for a non-solid state cell of ~12% based on a Zn porphyrin D-π-A sensitizer and Co(bpy)₃ mediator (bpy = 2,2’-bipyridine).²  This type of mediator solves many of the issues associated with I^-/I₃^-, including a controlable E_1/2 that can be tuned depending on the substitution of the polypyridine rings.  However, these complexes suffer from poor diffusion through the TiO₂ framework arising from their much larger size.  This leads to mass transport-limited current when under full 1-sun illumination.  As the oxidation of I^-/I₃^- is a two-electron process and the Co complex oxidations are one-electron processes (i.e., from 2⁺ to 3⁺), Co(bpy)₃-type complexes are more prone to fast interfacial recombination from the TiO₂ to the oxidized form of the mediator.  In the Grätzel work, this undesirable process was mitigated by 1) the steric bulk of the dye preventing access to the TiO₂ and 2) a diarylamine donor which translated the hole further away from the surface of the TiO₂.

In principle, the problematic back-electron transport from TiO₂ to the oxidized mediator can be approached from several directions.  Designing the dye such that bulky groups block the surface as Grätzel did is one such option.  Past work in our group has shown that incorporating insulating alkyl groups at the 4,4’ positions on bipyridyl ligands reduces electronic coupling and therefore slows down interfacial recombination, resulting in a higher photocurrent.³  Hamann and coworkers have shown that coating the TiO₂ prior to sensitization with a monolayer of Al₂O₃ effectively blocks recombination while retaining rapid electron injection into the TiO₂, which results in an improvement in external quantum efficiency.⁴  Regardless of the method chosen to address this problem, the fact remains that Co(bpy)₃-type mediators are promising alternatives to the traditional I^-/I₃^-electrolyte.

Recent work in our group has revolved around the engineering of DSC systems that make use of Co(bpy)₃-type complexes  as mediators as well as surface chemistry that affects the reduction of such complexes at the cathode.  As it turns out, the incorporation of well-established additives to mediator solutions can have a profound effect on electron transfer kinetics in various parts of the cell. 

For instance, in an effort to address the high cost of the sensitizer, we developed a series of Cu(I) phenanthroline dyes.⁵  Reports by Constable and coworkers^6,7 showed the validity of Cu bipyridyl dyes, and we were encouraged that the increased rigidity of the phenanthroline ring structure would lead to improved cell performance.  In addition, the incorporation of phenothiazine electron donors (based on earlier work by Bignozzi and Meyer⁸) was thought to extend the excited state lifetime by translating the hole away from the surface analogous to the diarylamine in Grätzel’s work.  However, the regeneration of the Cu(I) ground state by the Co mediator was sufficiently slow that sensitizers which did not incorporate an electron donor underwent exciplex formation with 4-tert-butylpyridine (TBP), a common additive used to increase V_OC.  Ground state regeneration rates became much slower in the presence of TBP, and cyclic voltammetry revealed substantial peak-shifting and irreversibility of the Cu(I/II) redox couple. 

We have also investigated the redox behavior of these types of complexes on various electrode materials and have found a substantial dependence on both identity and history of such electrodes.^9,10  As mentioned earlier, steric bulk on the recesses of the complex can inhibit electronic coupling and therefore reduce back electron transfer.  However, back electron transfer does still occur.  As it turns out, the incorporation of Li⁺ into the mediator solution – another common additive that tends to increase photocurrent – results in increasing the over-potential needed for heterogeneous electron transfer to the mediator.  This is particularly interesting in the case of designing systems Co mediators with larger rates of interfacial recombination such as the root complex, Co(bpy)₃.

Our urrent work has focused on the continued use of electron donating moieties in sensitizers and further investigating the electron transfer kinetics of Co(bpy)₃-type mediators.  Among these priorities are the synthetic tuning of Co complexes with more positive E_1/2 values in an effort to increase V_OC, but in concert with designing the
system in such a way as to reduce unwanted recombination from the TiO₂.  The continued exploration of electrode surface chemistry is also teaching us new lessons on the electron transfer kinetics of these complexes at the cathode, which may be compounding the perceived mass transport limitation.

1. Grätzel et al. JACS 1993, 115, 6382-6390.

2. Grätzel et al. Science 2011, 334, 629-634.

3. Elliott et al. JACS 2002, 124, 11215-11222.

4. Klahr and Hamann J. Phys. Chem. C 2009, 113, 14040-

    14045.

5. Ashbrook and Elliott J. Phys. Chem. C 2013, 117,

    3853-3864.

6. Constable et al. Chem. Commun.  2008, 32, 3717–3719

7. Constable et al. Dalt. Trans. 2011, 40, 12584-12594.

7. Bignozzi et al. J. Phys. Chem. B 1997, 101, 2591-2597

8. Elliott et al. J. Electrochem. Soc. 2013, 160(6), H355-

    H359.

9. Elliott et al. Langmuir 2013, 29, 825-831.