3 Plasmonically Enhanced Dye-Sensitized Solar Cells
137
necessary TiO 2 thickness to ∪5 nm [16], and allowing for higher coupling, and
absorption, by the dyes.
A comparison of the position of unprotected gold particles (type C, D and E)
showed that cells of type C exhibit twice the photocurrent of a metal-free cell, and
types D and E have reduced photocurrent compared to metal-free cells [44].
Du et al. [45] note that the size of the TiO 2 particles affects the electron injection
dynamics in a Au sphere/TiO 2 type D system, with larger TiO 2 particles reducing
recombination due to a longer diffusion length. Additionally, charging of the Ag in
a Ag/TiO 2 cell was shown to prevent recombination of the charge carriers due to a
large screening effect by the metal [46]. But the same effect also doubles the time
for the redox mediator to regenerate the dye [46].
Recent advances in growing the TiO 2 directly on the metal nanoparticles using
wet chemistry has allowed for the creation of type F cells [47]. It was found that as
little as 0.1 wt. % of core-shell particles improved efficiency from 7.8 to 9 % while
decreasing the thickness of the cell by 25 %. This decreased carrier distance is one
of the main appeals of incorporating plasmonic nanoparticles into photovoltaics.
Choi et al. performed measurements on DSSCs including both TiO 2 and SiO 2
coated Au nanospheres (type D) and noted that the TiO 2 coated particles increased
the overall cell efficiency from 9.29 % without Au to 9.78 % with Au, whilst the SiO 2
coated Au increased the cell efficiency to 10.21 % [17]. The difference was caused by
charge accumulation on the Au cores in the Au/TiO 2 case, which modifies the Fermi
level of the composite film [48]. This did not, however, explain the observation that
the Au/SiO 2 only increases the cell efficiency for particle loadings of less than 0.70
wt. %, above which the cell efficiency decreases.
The optimum particle loading that has been reported by reference [17] can be
explained by noting that a majority of light absorption in a metal loaded cell occurs
at the top of the cell, even for modest particle loadings. Take for example a 1 µm
thick cell loaded with 20 nm radius silver spheres. The cross-section of such spheres
is 0.013 µm 2 . Beer’s law relates the absorption cross-section, optical path length (l)
and number density of absorbers (N) to the transmission through the cell via T =
exp(−C abs l N ). For a 1 µm thick cell, a particle loading of N = 400 µm −3 is required
to achieve an optical density ( 1-T ) of 99 %. There are therefore 7.4 layers of silver
spheres in the cell with an areal density of 55 particles per square micron. The top
layer of spheres absorbs 49 % of the initial intensity of the incident light, which
is approximately 5 × 10
5 photons µm −2 s −1 , with each subsequent layer absorbing
49 % of the remaining light, i.e. the first layer of spheres absorbs 49 % of the photons,
and the remaining 6.4 layers absorb the remaining 50 %. It is easy to imagine that
the top layer of particles excites many more dye molecules than the redox couple can
reduce per second, resulting in lost photon absorption opportunities. This highlights
the importance of varying the density of the particle loading through the thickness
of the cell. It is vital that no single layer absorbs dramatically more photons than
the other layers, and therefore the particle positions in the cell must be optimized to
follow an exponential loading density that minimizes bleaching.
137
necessary TiO 2 thickness to ∪5 nm [16], and allowing for higher coupling, and
absorption, by the dyes.
A comparison of the position of unprotected gold particles (type C, D and E)
showed that cells of type C exhibit twice the photocurrent of a metal-free cell, and
types D and E have reduced photocurrent compared to metal-free cells [44].
Du et al. [45] note that the size of the TiO 2 particles affects the electron injection
dynamics in a Au sphere/TiO 2 type D system, with larger TiO 2 particles reducing
recombination due to a longer diffusion length. Additionally, charging of the Ag in
a Ag/TiO 2 cell was shown to prevent recombination of the charge carriers due to a
large screening effect by the metal [46]. But the same effect also doubles the time
for the redox mediator to regenerate the dye [46].
Recent advances in growing the TiO 2 directly on the metal nanoparticles using
wet chemistry has allowed for the creation of type F cells [47]. It was found that as
little as 0.1 wt. % of core-shell particles improved efficiency from 7.8 to 9 % while
decreasing the thickness of the cell by 25 %. This decreased carrier distance is one
of the main appeals of incorporating plasmonic nanoparticles into photovoltaics.
Choi et al. performed measurements on DSSCs including both TiO 2 and SiO 2
coated Au nanospheres (type D) and noted that the TiO 2 coated particles increased
the overall cell efficiency from 9.29 % without Au to 9.78 % with Au, whilst the SiO 2
coated Au increased the cell efficiency to 10.21 % [17]. The difference was caused by
charge accumulation on the Au cores in the Au/TiO 2 case, which modifies the Fermi
level of the composite film [48]. This did not, however, explain the observation that
the Au/SiO 2 only increases the cell efficiency for particle loadings of less than 0.70
wt. %, above which the cell efficiency decreases.
The optimum particle loading that has been reported by reference [17] can be
explained by noting that a majority of light absorption in a metal loaded cell occurs
at the top of the cell, even for modest particle loadings. Take for example a 1 µm
thick cell loaded with 20 nm radius silver spheres. The cross-section of such spheres
is 0.013 µm 2 . Beer’s law relates the absorption cross-section, optical path length (l)
and number density of absorbers (N) to the transmission through the cell via T =
exp(−C abs l N ). For a 1 µm thick cell, a particle loading of N = 400 µm −3 is required
to achieve an optical density ( 1-T ) of 99 %. There are therefore 7.4 layers of silver
spheres in the cell with an areal density of 55 particles per square micron. The top
layer of spheres absorbs 49 % of the initial intensity of the incident light, which
is approximately 5 × 10
5 photons µm −2 s −1 , with each subsequent layer absorbing
49 % of the remaining light, i.e. the first layer of spheres absorbs 49 % of the photons,
and the remaining 6.4 layers absorb the remaining 50 %. It is easy to imagine that
the top layer of particles excites many more dye molecules than the redox couple can
reduce per second, resulting in lost photon absorption opportunities. This highlights
the importance of varying the density of the particle loading through the thickness
of the cell. It is vital that no single layer absorbs dramatically more photons than
the other layers, and therefore the particle positions in the cell must be optimized to
follow an exponential loading density that minimizes bleaching.