Environmental Engineering Reference
In-Depth Information
exited state
Fermi level
ground state
Figure 6.17 Principle of operation [83] of dye-
sensitized solar cell. Note load resistor R at
bottom, and light excitation of dye by photon h
n
(step 1, center of figure). (Left) Schematic of
nanoporous deposit of TiO
2
(titania, anatase
form) coated with dye molecules (dots), and
immersed in electrolyte containing iodine ions.
Charge is carried to the dye by the
iodide
oxidation couple in
solution. Typically, the left side of the cell is
commercial SnO
2
-layered conductive glass,
which has been coated with a Ti-nanoxide paste
and fired [83].
triiodide reduction
-
-
6.5.1
Principle of Dye Sensitization to Extend Spectral Range to the Red
The operation of this device follows the steps indicated in Figure 6.17: step 1 is
absorption of a photon by a dye molecule. Typically, the dye is a metal-organic
ruthenium complex with response in the wavelength range 700
900 nm, light that
is not absorbed by the titania. This is the idea of dye sensitization, to extend the
absorption range of the photocell to utilize more of the solar spectrum. In step 2,
the excited state electron jumps from the dyemolecule into the conduction band of the
TiO
2
. The transfer is rapid from the dye only if the dye level lies
-
0.2 eV above the
conduction band, and reverse transfers from conduction band to dye are found to
be slow. In step 3 the electron diffuses through the porous nanoparticle layer, whichmay
extend 10
m, and enters the electrode, and in step 4 the electron
ows through the
external load resistor R. In step 5, the electron is transferred into the electrolyte, which is
typically aided by a very thin platinum catalyst layer. In step 6 the dye is regenerated into
its ground state. From the diagram, it appears that the open-circuit voltage of the cell is
approximately the photon energy divided by the electron charge, minus a voltage drop
that occurs in the last step (step 6) when the redox couple brings charge back to the dye
molecule. This may be oversimpli
ed. Other factors including a potential difference
across the back contact and a Fermi level shift of the anatase nanoparticles are discussed
in an excellent review by Graetzel [82]. It is clear that these devices cannot be modeled in
the conventional treatment outlined for crystalline semiconductor PN junction cells,
m
