ELECTROCHEMICAL REPLICATION OF SELF-ASSEMBLED BLOCK COPOLYMER NANOSTRUCTURES - Electrochemical Nanofabrication: Principles and Applications - page 103

Table 2.3 Calculated and measured roughness factors for each array

morphology and photovoltaic performances in 1.2 µm liquid electrolyte
cells under AM1.5 illumination

Morphology RF calc a RF meas Porosity (%) J SC (mAcm -2 ) V OC (mV) FF η (%)

Wires 100 60 65 1.79 845 0.71 1.07

Gyroid 125 130 60 5.83 713 0.71 2.97

Particles 120 116

100

60 5.42 785 0.63 2.67

a

Model calculations using 'smooth' geometric constructions. Wires: 12 nm diameter,
21 nm centre-to-centre hexagonal standing rods; cubic gyroid: 50 nm unit cell,
consisting of cylindrical rods of 12 nm diameter; nanoparticles: 20 nm diameter
spheres, 60% porosity with point-like inter-particle contact.

The absorption spectra of the dye-sensitized arrays before
cell construction is shown in Fig. 2.27a. Both the gyroid and wire
spectra show vertical offsets caused by scattering. Comparing the
(offset corrected) peak optical density of the arrays at 800 nm, the
nanoparticle and gyroid have comparable degrees of dye-loading
while the nanowire array has only half the peak absorbance,
indicating a considerably lower accessible area in the device for dye
uptake. The spectral response, that is, the wavelength-dependent
conversion efficiency of incident photons to extracted charge, of
cells based on the three arrays are shown in Fig. 2.27b. At these low
light intensities (∼0.1 mW cm ), the nanoparticle film shows the

highest peak external quantum efficiency (EQE) of 34% compared
with 28% for the gyroid array, while the nanowire cell produces a
peak EQE of only 9%.
Solar cell current-voltage curves under simulated AM1.5 solar
illumination at 100 mW cm

−2

are shown in Fig. 2.27c and summarized
in Table 2.2. The gyroid, nanoparticle and nanowire cells have
power conversion efficiencies of 3.0, 2.7 and 1.1%, respectively.
The relatively weak adsorption, low EQE and power conversion
efficiencies of the nanowire devices are consistent with low dye
loading. Indeed, the electrochemically determined surface area of
arrays replicated in Pt (Table 2.3) is only 60% of the value expected
for a perfect standing array. Low power conversion efficiency could
also arise from physical damage to a wire during processing which
would, for instance, prevent collection of charge above the level of a
fracture. This is not generally the case for network structures in which

−2