mixed amorphous P3HT:PCBM phase [93]. With increased PCBM density, the
PCBM also starts to form pure amorphous and crystalline domains [94, 95]. Solvent
annealing causes greater phase separation between P3HT and PCBM than thermal
annealing [96].
2.2.2 Thermal Annealing
A more commonly used method for causing phase separation between P3HT and
PCBM is thermal annealing. Thermal annealing of P3HT and mixtures of P3HT
with PCBM has been studied in great detail [42, 56, 59, 70, 76, 97–117]. Optimized
thermally annealed devices are spin-coated from a chlorobenzene solution in a ratio
that contains 35–45% PCBM by weight. After spin-coating, the metal electrode is
evaporated and then the device is heated for 5–30 min at ~150
C [42]. PCE as high
as 5% for this device type have been reported [42] but 3.5–4.5% PCE is more
common.
Moule ´ and Meerholz published two articles that described measurement of EQE
as a function of BHJ layer thickness. Several thousand individual P3HT:PCBM
OPV devices were fabricated to establish these data sets [70, 85]. We found that
OPV devices with PCE of ~4% could regularly and repeatably be fabricated using
the thermal annealing method with BHJ thicknesses of 80–350 nm. The solvent
annealing method occasionally brought “hero” devices, but in general was less easy
to control because the trace solvent atmosphere in the glove-box had a large effect
on the final morphology [118]. During the fabrication of so many devices we found
three issues that were likely to reduce device quality (without changing substrates
or metal type):
1. Occasionally, there is a PEDOT:PSS batch supplied that produces exclusively
S-shaped J/V curves with low FF [119, 120] This problem could only be solved
by ordering new PEDOT:PSS and is attributed to slight changes in the doping
level of the PEDOT:PSS itself.
2. The quality of P3HT varies widely between batches, which mostly has an effect
on the FF. Average FF > 0.6 is an indication of “good” P3HT. Typically, a
graduate student can achieve “good” results with a year of practice and with
~50% of the P3HT batches received. Cleaning the P3HT by dissolving and then
dripping into a solution of CH 3 OH:H 2 O improves the FF by an average of 0.05.
The improvement most probably comes from the removal of metals.
3. We have also informally found that evaporation of metal electrodes at a high rate
can reduce the FF because the hot metal damages the polymer. We use an initial
evaporation rate of 0.02 nm s
À1 .
Two different two-component nonequilibrium phase diagrams have been made
for P3HT:PCBM [121, 122]. Figure 9 shows the expected phase behavior of P3HT:
PCBM in a melt or solidifying melt [121]. Thermal analysis is able to show the
mixing ratio of P3HT:PCBM as a function of temperature. All investigations agree
that PCBM has ~30% miscibility with P3HT and that with higher PCBM content,
198
A.J. Moule ´ et al.
PCBM also starts to form pure amorphous and crystalline domains [94, 95]. Solvent
annealing causes greater phase separation between P3HT and PCBM than thermal
annealing [96].
2.2.2 Thermal Annealing
A more commonly used method for causing phase separation between P3HT and
PCBM is thermal annealing. Thermal annealing of P3HT and mixtures of P3HT
with PCBM has been studied in great detail [42, 56, 59, 70, 76, 97–117]. Optimized
thermally annealed devices are spin-coated from a chlorobenzene solution in a ratio
that contains 35–45% PCBM by weight. After spin-coating, the metal electrode is
evaporated and then the device is heated for 5–30 min at ~150
C [42]. PCE as high
as 5% for this device type have been reported [42] but 3.5–4.5% PCE is more
common.
Moule ´ and Meerholz published two articles that described measurement of EQE
as a function of BHJ layer thickness. Several thousand individual P3HT:PCBM
OPV devices were fabricated to establish these data sets [70, 85]. We found that
OPV devices with PCE of ~4% could regularly and repeatably be fabricated using
the thermal annealing method with BHJ thicknesses of 80–350 nm. The solvent
annealing method occasionally brought “hero” devices, but in general was less easy
to control because the trace solvent atmosphere in the glove-box had a large effect
on the final morphology [118]. During the fabrication of so many devices we found
three issues that were likely to reduce device quality (without changing substrates
or metal type):
1. Occasionally, there is a PEDOT:PSS batch supplied that produces exclusively
S-shaped J/V curves with low FF [119, 120] This problem could only be solved
by ordering new PEDOT:PSS and is attributed to slight changes in the doping
level of the PEDOT:PSS itself.
2. The quality of P3HT varies widely between batches, which mostly has an effect
on the FF. Average FF > 0.6 is an indication of “good” P3HT. Typically, a
graduate student can achieve “good” results with a year of practice and with
~50% of the P3HT batches received. Cleaning the P3HT by dissolving and then
dripping into a solution of CH 3 OH:H 2 O improves the FF by an average of 0.05.
The improvement most probably comes from the removal of metals.
3. We have also informally found that evaporation of metal electrodes at a high rate
can reduce the FF because the hot metal damages the polymer. We use an initial
evaporation rate of 0.02 nm s
À1 .
Two different two-component nonequilibrium phase diagrams have been made
for P3HT:PCBM [121, 122]. Figure 9 shows the expected phase behavior of P3HT:
PCBM in a melt or solidifying melt [121]. Thermal analysis is able to show the
mixing ratio of P3HT:PCBM as a function of temperature. All investigations agree
that PCBM has ~30% miscibility with P3HT and that with higher PCBM content,
198
A.J. Moule ´ et al.