Properties of the front contact

A very successful approach was the introduction of a heterojunction where a semiconductor with higher bandgap is used as front contact. In most thin film solar cells a wide bandgap semiconductor is used as front contact. Usually it is one of so called transparent conducting oxides (TCOs), ZnO, In₂O₃/SnO₂ (ITO), or SnO₂:F (FTO).

In order to be transparent for visible light the front contact is required to have a bandgap of more than about 3.3 eV which is the case for the mentioned TCOs. Absorption in the infrared (IR) is an important issue for TCOs because of free carriers absorption just like in metals.

We need reasonable conductivity in the front contact which we can only obtain it by a high doping level. A general requirement for the front contact is a resistivity below 10 Ω_sq (specific area resistance). The specific resistivity ρ of the film is obtained by multiplying with the layer thickness. Assume a thickness of one μm and a typical mobility of about 20 cm²/Vs, then we need carrier densities between 10²⁰ and 10²¹ cm^-3.

The figure below shows the dielectric function as well as the refractive index of an aluminium doped ZnO layer [1]. In order to illustrate the power of the simple model, calculated curves are superimposed to the data and extended to energies below 0.6 eV (dashed lines). A high frequency dielectric constant of ε of 3.9, a plasma frequency ω_p of 1x10¹⁴ 1/s (corresponding to a carrier density of about 3x10²⁰ 1/cm^-3), and a relaxation time τ of 6x10^-15 s corresponding to a mobility of about 30 cm²/Vs were assumed. The relations between mobility and relaxation time is discussed in the chapter about semiconductor physics.

For photon energies of more than 3 eV the behaviour is dominated by absorption phenomena. In the figure below the simple oscillator model has been used which attributes for the dispersion between 2 and 3 eV, but close to the bandgap again the model is far too simple. A better description requires the inclusion of excitonic phenomena [2].

From the Drude part we expect metallic reflection and absorption phenomena in the near infrared, somewhere between wavelengths of 1.5 and 3 μm. The figure below presents transmission, reflection and absorption of a ZnO:Al layer with lower mobility. In this case the free carrier absorption is broad and can extend almost into the visible wavelength regions where we want light transmission into the absorber.

With respect to photon energy we can distinguish several regions:

• For infrared photons below 0.3 eV we observe reflection. Absorption is low because the radiation does not enter the film.
• In the near IR the spectrum is dominated by a broad absorption peak.
• At the transition from IR to visible light (at photon energies just above 1 eV) some free carrier absorption persists. For a better efficiency of the solar cell the absorption in this region should be decreased, either by a lower carrier density or by better mobility.
• In the visible range ZnO behaves like a thin dielectric layer and shows interference fringes as discussed in the previous section. The thickness of the shown film is about 1.5 μm. The absorption is negligible, but the reflection and transmission measurements were probably not performed at exactly the same positions, thus the maxima and minima of the interference fringes do not match exactly.
• Photons with energies of more than 3.5 eV are absorbed by excitation of electrons over the bandgap.