In dye-sensitized solar cells (DSSCs), light-absorbing dye molecules inject excited electrons into a nanoporous semiconductor, commonly titanium dioxide (TiO2). Those electrons then move through the semiconductor toward a collecting contact. Because the porous network can make transport irregular, a conventional diffusion equation may not capture every aspect of electron movement. A fractional diffusion model offers a way to represent this complexity and examine how transport behavior affects calculated cell performance.
The model discussed here applies fractional diffusion to electron transport in the TiO2 conduction band. Its purpose is to account for the random-walk behavior associated with the semiconductor’s porous structure, then relate the simulated electron distribution to measures such as short-circuit current density, open-circuit voltage, and efficiency. The results are numerical and depend on the model assumptions and parameter values; they are not a general performance guarantee for DSSCs.
Why model electron transport in porous TiO2?
DSSCs use a nanoporous semiconductor rather than relying on a high-purity semiconductor structure like that used in conventional silicon solar cells. The dye absorbs incoming light and injects electrons into the semiconductor. Those electrons travel through the TiO2 network, while some can recombine at interfaces involving the semiconductor and electrolyte.
Mathematical models help describe how these processes shape a cell’s current–voltage relationship. Earlier diffusion-reaction models represent electron density as it changes over position and time, accounting for light absorption as an electron source and recombination as a loss process. However, a standard diffusion description does not explicitly represent the effects of the porous network’s geometry on electron movement.
A fractional approach introduces a parameter, γ, that describes the transport behavior associated with a continuous-time random walk. In the model, γ ranges from values above zero to one. When γ equals one, the equation reduces to the standard diffusion model. Lower values represent slower transport and longer electron path lengths through the nanoporous material.
How the fractional diffusion model works
The model describes electron density, n, as a function of position within the TiO2 layer and time. Its equation combines three contributions:
- Fractional diffusion: Describes electron transport while accounting for the random-walk characteristics of the porous network.
- Light absorption: Adds electrons at a rate that varies with depth, based on the incident photon flux and absorption coefficient.
- Recombination: Represents electron loss relative to the equilibrium electron density.
A fractional derivative differs from an ordinary derivative because it incorporates the system’s history into the mathematical description. The study uses a Caputo fractional derivative in time as part of the diffusion term. Under this formulation, the parameter γ connects the model to the power-law behavior of mean-square displacement in the random-walk description.
Boundary and initial conditions specify electron density at the contact and at the start of the simulation, as well as a no-flux condition at the opposite edge of the TiO2 layer. Together, these assumptions define the problem the numerical method solves. The study uses a finite difference scheme, including an L1 approximation for the fractional derivative, to calculate electron density over the layer and through time.
Connecting transport to cell performance
Once the electron-density distribution is calculated, the model estimates current and voltage characteristics. Short-circuit current density, Jsc, is derived from the electron flux at the contact. For γ equal to one, the expression reduces to the conventional flux relationship; for fractional transport, the calculation incorporates the fractional time derivative.
The open-circuit voltage, Voc, is calculated from the diode equation using the short-circuit current density and dark saturation current density. The model then identifies a maximum-power operating point and calculates efficiency as the maximum electrical power divided by incident light power. This sequence provides a way to connect a transport assumption about the porous semiconductor with familiar photovoltaic performance measures.
What the numerical results indicate
For the parameter values and simulation conditions examined, the calculated short-circuit current density increased as γ increased. The open-circuit voltage changed less than the current density, while the estimated efficiency was lower at smaller values of γ. In the reported cases, efficiency rose from about 2.81% at γ = 0.25 to 7.03% at γ = 1. The latter is the standard-diffusion case within this model, not a universal benchmark for DSSCs.
The simulations also showed that lower γ values took longer to approach steady state. This is consistent with the model’s interpretation of lower γ as slower transport. The electron density was notably higher in some of the lower-γ simulations, a result the authors identify as a concern when interpreting the profiles. They note that flux estimation and numerical stability become challenging at lower γ, and that computations can become expensive. These issues matter when using the model to draw conclusions from simulated performance.
The study also considered γ = 0.612, using a finer spatial and temporal discretization than in the comparison cases. Under those particular conditions, the reported estimates were Jsc = 116.9059 A/m2, Voc = 0.6245 V, and efficiency of 6.0835%. These figures illustrate the model’s output for the chosen inputs; they should be understood in that context rather than treated as expected values for other cells.
Practical implications and next steps
For researchers and technical teams evaluating DSSC transport models, the key contribution is a framework for representing the influence of TiO2’s porous network through a fractional transport parameter. Comparing different γ values can reveal how assumptions about electron movement affect predicted current, voltage, and efficiency. It also highlights the need to assess the electron-density profiles and numerical behavior—not just the final efficiency estimate.
The model does not include a detailed treatment of the electrolyte couple, which the authors identify as an area for further development. Extending the framework to include electrolyte transport could address additional interactions at the semiconductor–electrolyte interface. More broadly, the fractional diffusion model provides a mathematical route for studying how irregular transport in nanoporous TiO2 may influence DSSC performance, while making clear that model assumptions and numerical limitations must guide interpretation.