Carbon Nanotube/Polymer Composites as a Highly Stable Hole ...
Supplementary Information
Carbon Nanotube/Polymer Composites as a Highly Stable Hole Collection Layer in Perovskite Solar Cells
Severin N. Habisreutinger1, Tomas Leijtens1, Giles E. Eperon1, Samuel D. Stranks1, Robin J. Nicholas1 & Henry J. Snaith1*
1
Department of Physics, University of Oxford, Clarendon Laboratory, Parks Road, Oxford OX1 3PU, United Kingdom
*[email protected]
1
INDEX 1.
Methylammonium lead iodide/chloride degradation
2.
Degradation of devices with P3HT and PTAA
3.
Degradation of undoped spiro-OMeTAD
4.
Control devices
5.
Energy scheme
6.
SWNT-film deposition
7.
JV characteristics of the best-performing devices of all investigated HTM
8.
Thermal stability – 96 h study
9.
Thermal dependence of device performance
10. Water stability
2
1. Methylammonium lead iodide-chloride degradation Bare perovskite CH3NH3PbI3-xClx without any hole conducting top layer shows rapid degradation when heated (80°C) in air. Within 24 hours, its color changes from dark maroon to almost transparent yellow (figure S1 a)). The corresponding UV-Vis and XRD spectra demonstrate that this color change can be adjudicated to a change of the perovskite crystal structure. The diffraction pattern after heating does not exhibit the peaks characteristic for the CH3NH3PbI3-xClx perovskite. Instead there is only one dominating peak left at 12.7° which corresponds to a signature feature of a pure lead iodide film. This suggests that the degradation of the perovskite is correlated with the loss of methylammonium cations.
Figure S1: a) photograph of perovskite devices without hole transporting before and after 24 h exposed to 80°C. b) Absorption spectrum of the perovskite before and after heat exposure. The typical perovskite absorption across the visible range vanishes. c) XRD spectra of the perovskite before and after 24 h of heating, compared to a layer of pure lead iodide. The most pronounced features of the
3
perovskite crystal structure at 14.2° and 28.5° disappear fully and are replaced by a single peak at 12.7° which is associated with a film of pure lead iodide. That would also account for the yellow coloration of the degraded perovskite film.
4
2. Degradation of devices with P3HT and PTAA The absorption and XRD spectra for perovskite films cover with P3HT and PTAA, respectively, are shown in figure S2. Analogously to the observation for perovskite layers covered with Li-spiro-OMeTAD, samples with the polymeric hole-transporting material lose the absorption in the visible range, which is characteristic for CH3NH3PbI3-xClx. Correspondingly, the XRD measurements show that after 96 h of thermal stressing, the diffraction peaks associated with the perovskite have vanished. Instead, a new feature is observed at 12.80° which associated with PbI2 (figure S1).
Figure S2: a) Absorption spectrum of the perovskite samples covered with P3HT and PTAA taken in 24 h intervals. The typical perovskite absorption across the visible range slowly disappears. b) XRD spectra before and after 96 h of heating indicating degradation of the perovskite crystal structure.
5
3. Degradation of undoped spiro-OMeTAD The impact of lithium doping on the degradation process was evaluated by comparing doped and undoped spiro-OMeTAD. The heat induced degradation progressed noticeably faster for the spiro-OMeTAD HTL doped with Li-TFSI (figure S3 a)) which we attribute to the hygroscopic nature of lithium. In the absence of a dopant, the degradation proceeds more slowly, nevertheless already after 24 h, a reduction of absorption indicates that the perovskite structure begins to break down (figure S3 b)). The dffraction pattern after 96 h shows still some remaining features characteristic for the perovskite structure, however, the feature dominating the spectrum can be attributed to lead iodide which is indicative of the overall breakdown of the perovskite.
6
Figure S3: a) perovskite devices with undoped and doped spiro-OMeTAD as hole transporting layer exposed to 80°C. b) Absorption spectrum of the perovskite taken in 24 h intervals. The typical perovskite absorption across the visible range slowly disappears. c) XRD spectra of the perovskite before and after 96 h of heating.
7
4. Control devices The initial proof of principle was done by comparing devices without a dedicated hole-transporter layer, in which the electrode directly contacts the perovskite, with devices in which the HTL is comprised solely of a layer of P3HT/SWNTs. As we show in S4a the presence of SWNTs significantly improves charge extraction.
Figure S4a: Comparing the JV-characteristics of a device without hole-transporting layer and one with P3HT/SWNTs only. Despite being recombination limited the presence of SWNTs significantly improves hole extraction.
To make unambiguously sure that holes are exclusively transferred through the P3HT/SWNT nanohybrids in the stratified architecture, we also tested a PMMA layer without the P3HT/SWNT nanohybrids as “hole-transporting layer”. Expectedly, the insulating nature of the polymer does not support any current to flow as illustrated by the JV-curve of such a device in figure S4b.
8
Figure S4b: JV-characteristics of a device with neat PMMA as hole-transporting layer. Because of its insulating nature, there is no current flow.
9
5. Energy Scheme E(eV) -2.0 -2.8
-
-4.3
+
x -4.8
P3HT
Perovskite
TiO2
FTO
-5.3
-4.0 – -4.2
-4.0 -4.6
-4.8 – -5.0
-5.0 -6.0
Ag
-3.8
SWNTs
-4.0
-3.0
Figure S5: Energy levels of the various components of the device, showing energetically favorable pathways for holes and electron transfer to the respective electrodes. Because of its very high LUMO level, the transfer of photogenerated electrons from the perovskite layer to the nanotubes is blocked.
The position of the P3HT LUMO makes the transition of electrons from the perovskite to the nanohybrids highly unfavorable. P3HT may therefore act not only as a dispersing surfactant but also functionalize the SWNTs such that they become charge selective for photogenerated holes from the perovskite. Degrading the polymer chemically after the deposition leads to very poor device performance due to increased recombination losses.
10
6. SWNT-film deposition To obtain a homogenous SWNT film, we investigated several solvents and deposition methods. Dynamic drop-by-drop spin-coating from a chloroform solution was found to be the most effective and reproducible method for producing dense and resilient SWNT films. Other solvents such as chlorobenzene and toluene led to the formation of thick SWNT clusters and non-uniformity of the film. Films produced by non-dynamic spin-coating did not exhibit a comparable degree of uniformity and density as the dynamic drop-by-drop method.
Figure S6: Top view of a bare SWNT layer showing the thick mesh-like structure of the layer of P3HT-functionalized SWNTs.
11
7. JV characteristics of the best-performing devices of all investigated HTMs Before any thermal exposure devices with all investigated hole-transporting materials and structures were tested for their photovoltaic performance. The currentvoltage characteristics of the best-performing devices are shown in figure S8. The corresponding performance parameters are given in table S8.
Figure S7: JV-curves of the best-performing devices with various HTMs.
Table S1: performance parameters of the champion solar cells with various HTMs
2
HTL architecture
Jsc [mA/cm ]
Voc [V]
FF
PCE [%]
PTAA
15.3
0.94
0.51
7.3
P3HT
18.8
0.92
0.61
10.6
Li-spiro-OMeTAD
20.3
1.00
0.61
12.6
P3HT/SWNT-PC
21.4
1.00
0.69
14.8
P3HT/SWNT-PMMA
22.7
1.02
0.66
15.3
12
8. Thermal Stability – 96 h study After temperature exposure for 96 h photovoltaic performance of all devices was tested. Corresponding to the structural disintegration of the perovskite layer, the devices with Li-spiro-OMeTAD, PTAA and P3HT did not generate any power output. In contrast, the devices with P3HT/SWNT-PMMA and P3HT/SWNT-PC were still operational. Current-voltage characteristics of the best-performing devices are shown in figure S9.
Figure S8: JV-curves for devices with all investigated hole-transporting materials and structures after 96 h of thermal stressing at 80°C
Table S2: performance parameters of the best-‐performing devices after 96 h of heating
2
HTL architecture
Jsc [mA/cm ]
Voc [V]
FF
Max PCE [%]
Av PCE [%]
PTAA
0
x
x
0
0
P3HT
0
x
x
0
0
Li-spiro-OMeTAD
0
x
x
0
0
P3HT/SWNT-PC
19.8
1.03
0.70
14.3
10.7±3.7
P3HT/SWNT-PMMA
19.9
1.04
0.63
13.0
10.1±3.5
13
9. Thermal dependence of device performance Thermal dependence of device performance was tested with a custom-built sample chamber, which was externally heated. The temperature was controlled with an IR thermometer. Before measuring the photovoltaic performance of the devices, they were held at each temperature point for 20 minutes in order to let them equilibrate. The performance of the devices was measured in the temperature range between 25°C and 100°C. The performance parameters of all devices with different HTL structures measured at each temperature point are given in figure S10.1. Because the current drops to zero at 100°C for both P3HT and PTAA, the there are no corresponding values for Voc and FF.
Figure S9a: Performance parameters of devices with all investigated hole-transporting structures operating at temperatures between 25°C and 100°C
14
After measuring the photovoltaic performance at various temperatures between room temperature and 100°C, the devices were cooled down to room temperature and tested again. The standard HTLs did not recover from the thermally induced degradation accordingly they did not generate any power output under illumination (figure S10.2). The hole-transporting structures based on SWNTs were still operational. The performance of the PMMA HTL, however, significantly decreased. This is most likely called due to approaching the material’s glass transition, which impairs the integrity of the polymer matrix. As a consequence, direct contact between the perovskite layer and the electrode might become possible leading to increased shunting which would correspond to the observed JV-characteristic, namely the poor fill factor.
Figure S9b: current-voltage characteristic of devices measured at room temperature after having been tested at increased temperatures.
15
Table S3: performance parameters of the devices after having been tested at higher temperatures
2
HTL architecture
Jsc [mA/cm ]
Voc [V]
FF
PCE [%]
PTAA
0
x
x
0
P3HT
0
x
x
0
Li-spiro-OMeTAD
0
x
x
0
P3HT/SWNT-PC
23.0
0.97
0.60
13.5
P3HT/SWNT-PMMA
20.6
0.73
0.38
5.7
16
10. Water stability To demonstrate the ability of the P3HT/SWNT-PMMA composite structure to withstand moisture, we exposed devices with such a structure to a stream of water. As comparison, we did the same with devices with Li-spiro-OMeTAD as HTL. While devices with the nanotube composite exhibit no sign of degradation even after being exposed to the water stream for several minutes, the Li-spiro-OMeTAD devices start to degrade within a few seconds. After merely five seconds, the perovskite shows significant degradation (figure S11 a)). At this point the characteristic perovskite absorption onset has already almost fully disappeared from the absorption spectrum (figure S11 b)). In contrast, the absorption of the device covered with P3HT/SWNTPMMA remains unchanged.
17
Figure S10 a) photograph of devices with Li-spiro-OMeTAD and P3HT/SWNT-PMMA after having been exposed to a stream of running water. Degradation of the Li-spiro-OMeTAD devices is visible after merely 5 s of exposure. b) This degradation is also reflected by the change in the absorption of the Li-spiro-OMeTAD devices. In contrast, devices with the P3HT/SWNT-PMMA structure show no visible sign of degradation nor a change in their absorption characteristics (c)).
18
References 1.
Snaith, H. J. How should you measure your excitonic solar cells? Energy Environ. Sci. 5, 6513 (2012).
2.
Cambré, S., Wenseleers, W., Goovaerts, E. & Resasco, D. E. Determination of the metallic/semiconducting ratio in bulk single-wall carbon nanotube samples by cobalt porphyrin probe electron paramagnetic resonance spectroscopy. ACS Nano 4, 6717–24 (2010).
3.
Kim, S. S., Hisey, C. L., Kuang, Z., Comfort, D. a, Farmer, B. L. & Naik, R. R. The effect of single wall carbon nanotube metallicity on genomic DNAmediated chirality enrichment. Nanoscale 5, 4931–6 (2013).
19
Carbon Nanotube/Polymer Composites as a Highly Stable Hole Collection Layer in Perovskite Solar Cells
Severin N. Habisreutinger1, Tomas Leijtens1, Giles E. Eperon1, Samuel D. Stranks1, Robin J. Nicholas1 & Henry J. Snaith1*
1
Department of Physics, University of Oxford, Clarendon Laboratory, Parks Road, Oxford OX1 3PU, United Kingdom
*[email protected]
1
INDEX 1.
Methylammonium lead iodide/chloride degradation
2.
Degradation of devices with P3HT and PTAA
3.
Degradation of undoped spiro-OMeTAD
4.
Control devices
5.
Energy scheme
6.
SWNT-film deposition
7.
JV characteristics of the best-performing devices of all investigated HTM
8.
Thermal stability – 96 h study
9.
Thermal dependence of device performance
10. Water stability
2
1. Methylammonium lead iodide-chloride degradation Bare perovskite CH3NH3PbI3-xClx without any hole conducting top layer shows rapid degradation when heated (80°C) in air. Within 24 hours, its color changes from dark maroon to almost transparent yellow (figure S1 a)). The corresponding UV-Vis and XRD spectra demonstrate that this color change can be adjudicated to a change of the perovskite crystal structure. The diffraction pattern after heating does not exhibit the peaks characteristic for the CH3NH3PbI3-xClx perovskite. Instead there is only one dominating peak left at 12.7° which corresponds to a signature feature of a pure lead iodide film. This suggests that the degradation of the perovskite is correlated with the loss of methylammonium cations.
Figure S1: a) photograph of perovskite devices without hole transporting before and after 24 h exposed to 80°C. b) Absorption spectrum of the perovskite before and after heat exposure. The typical perovskite absorption across the visible range vanishes. c) XRD spectra of the perovskite before and after 24 h of heating, compared to a layer of pure lead iodide. The most pronounced features of the
3
perovskite crystal structure at 14.2° and 28.5° disappear fully and are replaced by a single peak at 12.7° which is associated with a film of pure lead iodide. That would also account for the yellow coloration of the degraded perovskite film.
4
2. Degradation of devices with P3HT and PTAA The absorption and XRD spectra for perovskite films cover with P3HT and PTAA, respectively, are shown in figure S2. Analogously to the observation for perovskite layers covered with Li-spiro-OMeTAD, samples with the polymeric hole-transporting material lose the absorption in the visible range, which is characteristic for CH3NH3PbI3-xClx. Correspondingly, the XRD measurements show that after 96 h of thermal stressing, the diffraction peaks associated with the perovskite have vanished. Instead, a new feature is observed at 12.80° which associated with PbI2 (figure S1).
Figure S2: a) Absorption spectrum of the perovskite samples covered with P3HT and PTAA taken in 24 h intervals. The typical perovskite absorption across the visible range slowly disappears. b) XRD spectra before and after 96 h of heating indicating degradation of the perovskite crystal structure.
5
3. Degradation of undoped spiro-OMeTAD The impact of lithium doping on the degradation process was evaluated by comparing doped and undoped spiro-OMeTAD. The heat induced degradation progressed noticeably faster for the spiro-OMeTAD HTL doped with Li-TFSI (figure S3 a)) which we attribute to the hygroscopic nature of lithium. In the absence of a dopant, the degradation proceeds more slowly, nevertheless already after 24 h, a reduction of absorption indicates that the perovskite structure begins to break down (figure S3 b)). The dffraction pattern after 96 h shows still some remaining features characteristic for the perovskite structure, however, the feature dominating the spectrum can be attributed to lead iodide which is indicative of the overall breakdown of the perovskite.
6
Figure S3: a) perovskite devices with undoped and doped spiro-OMeTAD as hole transporting layer exposed to 80°C. b) Absorption spectrum of the perovskite taken in 24 h intervals. The typical perovskite absorption across the visible range slowly disappears. c) XRD spectra of the perovskite before and after 96 h of heating.
7
4. Control devices The initial proof of principle was done by comparing devices without a dedicated hole-transporter layer, in which the electrode directly contacts the perovskite, with devices in which the HTL is comprised solely of a layer of P3HT/SWNTs. As we show in S4a the presence of SWNTs significantly improves charge extraction.
Figure S4a: Comparing the JV-characteristics of a device without hole-transporting layer and one with P3HT/SWNTs only. Despite being recombination limited the presence of SWNTs significantly improves hole extraction.
To make unambiguously sure that holes are exclusively transferred through the P3HT/SWNT nanohybrids in the stratified architecture, we also tested a PMMA layer without the P3HT/SWNT nanohybrids as “hole-transporting layer”. Expectedly, the insulating nature of the polymer does not support any current to flow as illustrated by the JV-curve of such a device in figure S4b.
8
Figure S4b: JV-characteristics of a device with neat PMMA as hole-transporting layer. Because of its insulating nature, there is no current flow.
9
5. Energy Scheme E(eV) -2.0 -2.8
-
-4.3
+
x -4.8
P3HT
Perovskite
TiO2
FTO
-5.3
-4.0 – -4.2
-4.0 -4.6
-4.8 – -5.0
-5.0 -6.0
Ag
-3.8
SWNTs
-4.0
-3.0
Figure S5: Energy levels of the various components of the device, showing energetically favorable pathways for holes and electron transfer to the respective electrodes. Because of its very high LUMO level, the transfer of photogenerated electrons from the perovskite layer to the nanotubes is blocked.
The position of the P3HT LUMO makes the transition of electrons from the perovskite to the nanohybrids highly unfavorable. P3HT may therefore act not only as a dispersing surfactant but also functionalize the SWNTs such that they become charge selective for photogenerated holes from the perovskite. Degrading the polymer chemically after the deposition leads to very poor device performance due to increased recombination losses.
10
6. SWNT-film deposition To obtain a homogenous SWNT film, we investigated several solvents and deposition methods. Dynamic drop-by-drop spin-coating from a chloroform solution was found to be the most effective and reproducible method for producing dense and resilient SWNT films. Other solvents such as chlorobenzene and toluene led to the formation of thick SWNT clusters and non-uniformity of the film. Films produced by non-dynamic spin-coating did not exhibit a comparable degree of uniformity and density as the dynamic drop-by-drop method.
Figure S6: Top view of a bare SWNT layer showing the thick mesh-like structure of the layer of P3HT-functionalized SWNTs.
11
7. JV characteristics of the best-performing devices of all investigated HTMs Before any thermal exposure devices with all investigated hole-transporting materials and structures were tested for their photovoltaic performance. The currentvoltage characteristics of the best-performing devices are shown in figure S8. The corresponding performance parameters are given in table S8.
Figure S7: JV-curves of the best-performing devices with various HTMs.
Table S1: performance parameters of the champion solar cells with various HTMs
2
HTL architecture
Jsc [mA/cm ]
Voc [V]
FF
PCE [%]
PTAA
15.3
0.94
0.51
7.3
P3HT
18.8
0.92
0.61
10.6
Li-spiro-OMeTAD
20.3
1.00
0.61
12.6
P3HT/SWNT-PC
21.4
1.00
0.69
14.8
P3HT/SWNT-PMMA
22.7
1.02
0.66
15.3
12
8. Thermal Stability – 96 h study After temperature exposure for 96 h photovoltaic performance of all devices was tested. Corresponding to the structural disintegration of the perovskite layer, the devices with Li-spiro-OMeTAD, PTAA and P3HT did not generate any power output. In contrast, the devices with P3HT/SWNT-PMMA and P3HT/SWNT-PC were still operational. Current-voltage characteristics of the best-performing devices are shown in figure S9.
Figure S8: JV-curves for devices with all investigated hole-transporting materials and structures after 96 h of thermal stressing at 80°C
Table S2: performance parameters of the best-‐performing devices after 96 h of heating
2
HTL architecture
Jsc [mA/cm ]
Voc [V]
FF
Max PCE [%]
Av PCE [%]
PTAA
0
x
x
0
0
P3HT
0
x
x
0
0
Li-spiro-OMeTAD
0
x
x
0
0
P3HT/SWNT-PC
19.8
1.03
0.70
14.3
10.7±3.7
P3HT/SWNT-PMMA
19.9
1.04
0.63
13.0
10.1±3.5
13
9. Thermal dependence of device performance Thermal dependence of device performance was tested with a custom-built sample chamber, which was externally heated. The temperature was controlled with an IR thermometer. Before measuring the photovoltaic performance of the devices, they were held at each temperature point for 20 minutes in order to let them equilibrate. The performance of the devices was measured in the temperature range between 25°C and 100°C. The performance parameters of all devices with different HTL structures measured at each temperature point are given in figure S10.1. Because the current drops to zero at 100°C for both P3HT and PTAA, the there are no corresponding values for Voc and FF.
Figure S9a: Performance parameters of devices with all investigated hole-transporting structures operating at temperatures between 25°C and 100°C
14
After measuring the photovoltaic performance at various temperatures between room temperature and 100°C, the devices were cooled down to room temperature and tested again. The standard HTLs did not recover from the thermally induced degradation accordingly they did not generate any power output under illumination (figure S10.2). The hole-transporting structures based on SWNTs were still operational. The performance of the PMMA HTL, however, significantly decreased. This is most likely called due to approaching the material’s glass transition, which impairs the integrity of the polymer matrix. As a consequence, direct contact between the perovskite layer and the electrode might become possible leading to increased shunting which would correspond to the observed JV-characteristic, namely the poor fill factor.
Figure S9b: current-voltage characteristic of devices measured at room temperature after having been tested at increased temperatures.
15
Table S3: performance parameters of the devices after having been tested at higher temperatures
2
HTL architecture
Jsc [mA/cm ]
Voc [V]
FF
PCE [%]
PTAA
0
x
x
0
P3HT
0
x
x
0
Li-spiro-OMeTAD
0
x
x
0
P3HT/SWNT-PC
23.0
0.97
0.60
13.5
P3HT/SWNT-PMMA
20.6
0.73
0.38
5.7
16
10. Water stability To demonstrate the ability of the P3HT/SWNT-PMMA composite structure to withstand moisture, we exposed devices with such a structure to a stream of water. As comparison, we did the same with devices with Li-spiro-OMeTAD as HTL. While devices with the nanotube composite exhibit no sign of degradation even after being exposed to the water stream for several minutes, the Li-spiro-OMeTAD devices start to degrade within a few seconds. After merely five seconds, the perovskite shows significant degradation (figure S11 a)). At this point the characteristic perovskite absorption onset has already almost fully disappeared from the absorption spectrum (figure S11 b)). In contrast, the absorption of the device covered with P3HT/SWNTPMMA remains unchanged.
17
Figure S10 a) photograph of devices with Li-spiro-OMeTAD and P3HT/SWNT-PMMA after having been exposed to a stream of running water. Degradation of the Li-spiro-OMeTAD devices is visible after merely 5 s of exposure. b) This degradation is also reflected by the change in the absorption of the Li-spiro-OMeTAD devices. In contrast, devices with the P3HT/SWNT-PMMA structure show no visible sign of degradation nor a change in their absorption characteristics (c)).
18
References 1.
Snaith, H. J. How should you measure your excitonic solar cells? Energy Environ. Sci. 5, 6513 (2012).
2.
Cambré, S., Wenseleers, W., Goovaerts, E. & Resasco, D. E. Determination of the metallic/semiconducting ratio in bulk single-wall carbon nanotube samples by cobalt porphyrin probe electron paramagnetic resonance spectroscopy. ACS Nano 4, 6717–24 (2010).
3.
Kim, S. S., Hisey, C. L., Kuang, Z., Comfort, D. a, Farmer, B. L. & Naik, R. R. The effect of single wall carbon nanotube metallicity on genomic DNAmediated chirality enrichment. Nanoscale 5, 4931–6 (2013).
19
