Polymer-acid-metal quasi-ohmic contact for stable perovskite solar cells beyond a 20,000-hour extrapolated lifetime

The development of a robust quasi-ohmic contact with minimal resistance, good stability and cost-effectiveness is crucial for perovskite solar cells. We introduce a generic approach featuring a Lewis-acid layer sandwiched between dopant-free semicrystalline polymer and metal electrode in perovskite solar cells, resulting in an ideal quasi-ohmic contact even at elevated temperature up to 85 °C. The solubility of Lewis acid in alcohol facilitates nondestructive solution processing on top of polymer, which boosts hole injection from polymer into metal by two orders of magnitude. By integrating the polymer-acid-metal structure into solar cells, devices exhibit remarkable resilience, retaining 96% ± 3%, 96% ± 2% and 75% ± 7% of their initial efficiencies after continuous operation in nitrogen at 35 °C for 2212 h, 55 °C for 1650 h and 85 °C for 937 h, respectively. Leveraging the Arrhenius relation, we project an impressive T80 lifetime of 26,126 h at 30 °C.

Reviewer #1 (Remarks to the Author): Review of "Polymer-acid-metal ohmic contact for stable perovskite solar 2 cells beyond a 15-year operational lifetime" I believe this manuscript presents interesting results that warrant publication in a more energy focused journal, rather than a general journal like Nature Communications.
1.I do not believe it is suitable to put the term "…15-year operational lifetime" in a title.This is a bit subjective because it is determined by extrapolation and a solar simulator.Maybe "…15-year simulated operational lifetime" 2. From point 1, is the solar simulator in fact used for the stability testing?And if so, could the authors provide the spectra of the light source and indicate what is the amount of UV component?3. Line 75 -I do not think the word "engenders" applies in this context.4. Line 89 -should be accordingly.5.The authors do mention cost at a few points in the manuscript, but do not provide actual numbers.The ultimate polymer selected for the study, PDCBT, seems to have a cost of 1g/$3250 according to Ossila and 500mg/$1986 Thermo-Fisher, which are not low cost.If the authors have other values 1-Materials please indicate.6.The term crystalline polymers to describe the various polymers used in this study, but shouldn't a more appropriate term be semi-crystalline?These polymers are not fully crystalline.7. Why do the authors select these polymers over something like PTAA? Did the authors study devices with a structure …perovskite/PTAA/BCF derivative/gold?Thiophene containing polymers have a lower bandgaps than aromatic amine polymers like PTAA, thus can contribute to parasitic absorption of visible photons.8.With regards to BCF, please indicate the actual name of these materialstris(pentafluorophenyl)borane is BCF, lithium tetrakis(pentafluorophenyl)borate is Li-BCF, etc.Along these lines, BCF is clearly a Lewis acid, but I don't think the tetrakis(pentafluorophenyl)borate based molecules are considered Lewis acids since the empty boron orbital is no longer present.If the authors have another perspective on this, please indicate.9. Line 130 -should be "studied" Reviewer #2 (Remarks to the Author): This paper reports the use of a Lewis-acid layer (BCF) between the p-type dopant free polymer (PDCBT) and metal (Au) in n-i-p PSCs.This interlayer creates a good ohmic contact for extracting photogenerated holes.The ohmic contact is maintained over a wide range of temperatures.The resulting device showed good performance (about 21%) for dopant-free n-i-p PSCs.The stability of these devices is good from 35 to 85 degrees.An activation energy of about 0.55 eV was derived.An estimated T80 of about 23012 to 24390 hours at 30 degrees suggests a service life of more than 15 years.Overall, this paper provides interesting insights for dopant free HTL work in n-i-p PSCs.However, this paper is not suitable for publication in its current stage due to a few important points that need to be addressed.
1.The activation energy and stability results are based on three individual cells, one cell at one temperature.This is not acceptable.It is known in the PSC field that there is often a large variation for PSC efficiencies.The variation of stability can be much larger.Thus, to draw any conclusion, a statistically significant comparison based on at least 5-10 devices at each temperature should be conducted.The average should be used to determine the stability and the activation energy.The error bars for the analysis should also be given.
2. 23012 to 24390 hours operation at 30 degrees cannot be equated to 15 years of service lifetime.
For practical outdoor operation, the conditions (temperature, illumination, etc.) can change dramatically over days, weeks, and months.Thus, to draw any connection to outdoor operation, the authors should conduct outdoor ageing tests.
3. It is interesting that the degradation activation energy is about 0.55 eV for the n-i-p PSCs used in this study.This value is very similar to a recent Nature article on p-i-n PSC stability, which showed an activation energy of about 0.6 eV.This similarity is interesting given the two studies used different contact layers and perovskite composition.Can the authors comment on this, with respect to the potential origin of the activation energy for both n-i-p and p-i-n PSCs.
4. Why can MgF2 deposited on top of Au protect the device? 5. Is achieving good ohmic contact between the HTL and Au the most important factor for the improved stability?I think this is an overstatement about the importance of an ohmic contact.As for the state-of-the-art n-i-p PSCs, where spiro-OMeTAD is used, the stability is poor even at room temperature.However, I believe the contact is ohmic between spiro-OMeTAD and Au.This will contradict the conclusion from this study.
Reviewer #3 (Remarks to the Author): The authors have substantially improved the stability of perovskite solar cells through the use of a Lewis acid to dope the polymer hole transport material.The authors have measured the rate of degradation at three different temperatures and calculated the acceleration factor.They estimate that the cells would operate for over 15 years at 30 degrees C.They call that the operational lifetime.I do not think that name should be used.The cells would frequently reach 40-55 degrees C in operation.Cells are often 30 C hotter than the ambient temperature.I am still impressed that an acceleration factor was determined.I have rarely seen anyone do that for perovskite solar cells.I recommend published this excellent manuscript after some minor edits are made.It would be better to refer to the polymers as semicrystalline rather than crystalline.
I recommend telling the reader that the Lewis acids are "coupled with alcohol-soluble Lewis acids (BCF/Li-BCF/C-BCF/N-BCF/I-BCF)" earlier in the manuscript.I don't think there is a sentence early in the manuscript pointing people to look at Figure 1.Line 126: BCF should be spelled out.Line 113: "disovle" is misspelled.Line 181: "phonmena" is misspelled.Line 208: "yielding a negligible hysteresis PCE of around 21%" is slightly confusing.I prefer "yielding a PCE of 21 % with negligible hysteresis."

Point-by-point list of author actions in response to Reviewer comments
Manuscript #: NCOMMS-23-47889 Black color: Reviewer's comment Blue color: Author's response Red color: Revision in the revised manuscript

Reviewer #1
Review of "Polymer-acid-metal ohmic contact for stable perovskite solar cells beyond a 15-year operational lifetime".I believe this manuscript presents interesting results that warrant publication in a more energy focused journal, rather than a general journal like Nature Communications.
Response: We sincerely thank the reviewer for this comment.In our work, we underscore the pivotal role of a stable quasi-ohmic contact, a crucial element not only within the realm of energy conversion but also across various optoelectronic disciplines.By introducing a novel approach that combines a Lewis acid layer with a semi-crystalline conducting polymer, we showcase a new paradigm for achieving stable quasi-ohmic contacts.
We firmly believe that the implications of our research extend beyond the confines of a specific field, and will be of substantial interest to the broad readership of Nature Communications.Once again, we thank you for recognizing the importance of our work and for considering its relevance to a broader audience.
1.I do not believe it is suitable to put the term "…15-year operational lifetime" in a title.This is a bit subjective because it is determined by extrapolation and a solar simulator.Maybe "…15-year simulated operational lifetime".

Response:
We thank the reviewer for this professional suggestion.We have revised the title to "Polymer-acid-metal quasi-ohmic contact for stable perovskite solar cells beyond a 20,000-hour extrapolated lifetime".

Page 1, Line 1:
Polymer-acid-metal quasi-ohmic contact for stable perovskite solar cells beyond a 20,000-hour extrapolated lifetime 2. From point 1, is the solar simulator in fact used for the stability testing?And if so, could the authors provide the spectra of the light source and indicate what is the amount of UV component?

Response:
To clarify, the long-term stability assessment of our devices was conducted under metal halide lamps equipped with UV filters having a cutoff at 400 nm.This information is supplemented in the revised manuscript.

Page 23, Line 402:
under metal halide lamps equipped with UV filters having a cutoff at 400 nm.
3. Line 75 -I do not think the word "engenders" applies in this context.

Response:
We thank the reviewer for pointing out this issue.We have made the necessary correction by replacing the word "engenders" with a more fitting term, "gives rise to".

Page 4, Line 81:
the significant similarity in solubility of the two polymers gives rise to a processing challenge in which the upper polymer tends to dissolve the underlying polymer.

Response:
We feel sorry for this erratum.It is corrected in the revised manuscript.
5. The authors do mention cost at a few points in the manuscript, but do not provide actual numbers.The ultimate polymer selected for the study, PDCBT, seems to have a cost of 1g/$3250 according to Ossila and 500mg/$1986 Thermo-Fisher, which are not low cost.If the authors have other values 1-

Materials please indicate.
Response: Thank you for highlighting the importance of addressing cost considerations in our study.To address your inquiry, we have gathered the prices for the specific materials referenced in the manuscript.The reference to "lower cost" in our paper pertains to the comparison between PDCBT/BCF and PDCBT/PTAA.We are optimistic about achieving even more cost-effective outcomes by utilizing P3HT/BCF in future research, although the device performance is still low.We appreciate your attention to this matter and assure you that future iterations of our work will delve deeper into exploring more costefficient and high-performance alternatives.Thus, we successfully demonstrate an ideal quasi-ohmic contact with low cost (Supplementary Table 1) and high reproducibility by using a polymer-acid-metal architecture, providing an attractive alternative to commonly used metal oxides and doped conducting polymers.

Product
6.The term crystalline polymers to describe the various polymers used in this study, but shouldn't a more appropriate term be semi-crystalline?These polymers are not fully crystalline.

Response:
We fully agree with the reviewer.We have corrected the "crystalline polymer" to "semicrystalline polymer" throughout the whole manuscript.Response: In our investigation, we observed that semicrystalline PDCBT films exhibit efficient hole transportation, unlike amorphous PTAA films that necessitate bulk doping for adequate hole transport.As evidenced in Fig. R1, a device featuring the structure ITO/SnO2/PCBM/perovskite/PTAA/BCF/Au presents a low fill factor, highlighting the limitations associated with PTAA.
Regarding concerns about potential parasitic absorption due to PDCBT's light absorption around 600 nm, it's noteworthy that this property becomes less consequential when the perovskite layer exceeds 500 nm in thickness.This mitigation is attributed to the high absorption coefficient of the perovskite film at approximately 600 nm wavelengths.

Response:
We thank the reviewer for pointing out this issue.we have supplemented the revised manuscript with the full names of BCF and its derivatives for clarity.We redefine that Li-BCF, C-BCF, N-BCF and I-BCF are BCF's derivatives.
As you know, organic semiconductor doping is a complex issue because the doping effect is dependent on the structural and electrical properties of both the dopants and the organic semiconductors and on the interaction between them.9. Line 130 -should be "studied".

Response:
We feel sorry for this erratum.It is corrected in the revised manuscript.
Page 9, Line 148: Interface contact was further studied based on the vertical architecture of ITO/PDCBT/interlayer/Au.

Reviewer #2
This paper reports the use of a Lewis-acid layer (BCF) between the p-type dopant free polymer (PDCBT) and metal (Au) in n-i-p PSCs.This interlayer creates a good ohmic contact for extracting photogenerated holes.The ohmic contact is maintained over a wide range of temperatures.The resulting device showed good performance (about 21%) for dopant-free n-i-p PSCs.The stability of these devices is good from 35 to 85 degrees.An activation energy of about 0.55 eV was derived.An estimated T80 of about 23012 to 24390 hours at 30 degrees suggests a service life of more than 15 years.Overall, this paper provides interesting insights for dopant free HTL work in n-i-p PSCs.However, this paper is not suitable for publication in its current stage due to a few important points that need to be addressed.

Response:
We thank the reviewer for the appraisal of our work and appreciate his/her valuable time to give these helpful comments and suggestions.
1.The activation energy and stability results are based on three individual cells, one cell at one temperature.This is not acceptable.It is known in the PSC field that there is often a large variation for PSC efficiencies.The variation of stability can be much larger.Thus, to draw any conclusion, a statistically significant comparison based on at least 5-10 devices at each temperature should be conducted.The average should be used to determine the stability and the activation energy.The error bars for the analysis should also be given.

Response:
We sincerely appreciate your feedback and the valuable insights               The devices retain 96% ± 3%, 96% ± 2% and 75% ± 7% of their initial PCEs after continuous illumination at 35 °C for 2,212 hours, 55 °C for 1,650 hours and 85 °C for 937 hours with negligible J-V hysteresis.
Page 17, Line 297: the fitted activation energy is 0.61 ± 0.11 eV and 0.59 ± 0.10 eV for the data extracted from reverse and forward J-V scan, respectively.
Page 18, Line 313: we estimate the  2021,12,2191).We expect a higher activation energy for the perovskites with improved surface bonding, considering the initial degradation always starts with the surface layer.
Page 17, Line 298: The activation energy for our n-i-p structure is close to that reported in p-i-n structure with a similar perovskite composition, implying that perovskite decomposition determines the temperature-dependent degradation rate in

Why can MgF2 deposited on top of Au protect the device?
Response: There is a high risk of breakage for Au electrode during the ageing period at high temperatures, which is caused by trace vapors (e.g.I2/MF/FA) generated by perovskite film (Supplementary Fig. 30).This addition of MgF2 serves to suppress the breakage susceptibility of the Au electrode, thereby bolstering the overall stability of the PSCs under these conditions.
5. Is achieving good ohmic contact between the HTL and Au the most important factor for the improved stability?I think this is an overstatement about the importance of an ohmic contact.As for the state-of-the-art n-i-p PSCs, where spiro-OMeTAD is used, the stability is poor even at room temperature.However, I believe the contact is ohmic between spiro-OMeTAD and Au.This will contradict the conclusion from this study.

Response:
We acknowledge that while ohmic contact stands as a crucial factor, it is not the exclusive determinant of enhanced stability.Actually, the ohmic contact between spiro-OMeTAD and Au is found to be unstable during ageing process at high temperatures, which will be reported in our forthcoming work.
It's essential to note that the overnight oxidation process used in spiro-OMeTAD's fabrication primarily affects the interface rather than bulk doping, contributing to the observed instability.This aligns with our ongoing investigations, and we aim to present comprehensive insights into this matter in our subsequent research.
7. Why do the authors select these polymers over something like PTAA? Did the authors study devices with a structure …perovskite/PTAA/BCF derivative/gold?Thiophene containing polymers have lower bandgaps than aromatic amine polymers like PTAA, thus can contribute to parasitic absorption of visible photons.
Peter K.H.Ho et al. and Thuc-Quyen Nguyen et al. confirmed  the p-doping effect by organic salts by using FET device and EPR/IR spectra, and they further studied the mechanism of p-doping(Nat.Commun.2016, 7, 11948; ACS     Nano 2018, 12, 3938).The downshift of the Fermi level is attributed to the Hubbard/Coulomb interaction that attracts electrons from the conducting polymers, as shown in Fig.R2.We hope this additional information sheds light on the complexity of organic semiconductor doping and its intricacies related to BCF derivatives and their impact on the Fermi level.
regarding the statistical significance of our stability and activation energy analyses.We have revised the manuscript by including the trend of photovoltaic parameters obtained from stability tests conducted on five representative devices.These tests encompassed both reverse and forward J-V scans, providing a more comprehensive understanding of device performance.The error bar of photovoltaic parameters and activation energy are derived from the five individual devices at different temperatures (Figs.R3-R16).These additions enable a more robust analysis, resulting in a corrected activation energy of approximately 0.61 ± 0.11 eV for reverse J-V scans and 0.59 ± 0.10 eV for forward J-V scans.All the new data are supplemented in the revised manuscript and supplementary information.

Fig. R3
Fig. R3 Long-term stability of PSCs based on PDCBT/BCF/Au structure.a,b Data points originated from (a) reverse J-V scan and (b) forward J-V scan, respectively.c,d Normalized PCE originated from (c) reverse J-V scan and (d) forward J-V scan of PSCs plotted against the equivalent aging time at 30 °C by leveraging the Arrhenius relation.The error bars represent the standard deviations from 5 individual devices for each temperature.

Fig. R4
Fig. R4 Logarithm of degradation rate (k) versus 1/kBT.a,b The data of k originated from analyzing J-V curves with (a) reverse and (b) forward scans of PSCs based on ITO/SnO2/PCBM/perovskite/PDCBT/BCF/Au architecture.We used 5 individual devices for each stability test to obtain the average activation energy.
-p structure of glass/FTO/TiO2/Al2O3/CsPbI3/Cs2PbI2Cl2/CuSCN/Cr/Au, reporting activation energy of approximately 0.24 eV and 0.43 eV for the degradation rates with and without surface passivation, respectively.Given the intrinsic stability of transparent conducting oxides (TCO) and other transporting materials, our conclusions lean towards the influence of perovskite composition on the observed activation energy similarities across different PSC structures.Notably, our previous research on CsMAFA-perovskites also revealed an activation energy of approximately 0.71 eV for decomposition (Nat.Commun. of approximately 0.59 eV for the degradation rates in their latest Nature publication (https://www.nature.com/articles/s41586-023-06610-7).Additionally, we referenced Yueh-Lin Loo et al.'s work in Science, which explored a different n-i