betaFeedback

WorksPhysics and Astronomy

Compositional and Interface Engineering of Hybrid Metal Halide Perovskite Thin Films for Solar Cells

Abstract

Perovskite solar cells (PSCs) are the fastest-growing photovoltaic (PV) technology in the solar cell community and have reached an efficiency close to that of commercial silicon (Si) solar cells. The organic-inorganic halide perovskite solar cell is an emerging PV technology and grabbed much attention due to its low cost, high efficiency, and ease of fabrication at lower temperatures 100-200 C by solution-processed spin coating or thermal evaporation techniques. Further, we introduce a self-assembled monolayer (SAM) based hole transport layer (HTL) in the p-i-n device architecture PSC. In this work, we used the mixing engineering strategy of SAM with a conjugated polyelectrolyte. We dealt with the hydrophobicity and tailored the work function of the mixed SAM based HTL. Therefore, the HTL/perovskite interface is engineered, and associated device physics is discussed. In addition, we observed the lowest dark current for specific mixed (9:1) HTL, which is a prerequisite in photodetector applications. Therefore, a detailed photodetection analysis is discussed to study the scalable photodetector device. This thesis thoroughly studies perovskite compositional and interface engineering via various optoelectronic measurements. An in-depth device physics is discussed to study the interfacial defects between the charge transport and the perovskite layers. This thesis will be helpful in exploring a new class of perovskite materials and interface modification engineering for fabricating reproducible, stable, and highly efficient hybrid organic-inorganic metal halide perovskite solar cells.

Submitted in partial fulfillment of the requirements

of the degree of

Doctor of Philosophy

by

(Roll No. 174120006)

Supervisor

Prof. Dinesh Kabra

Department of Physics

INDIAN INSTITUTE OF TECHNOLOGY BOMBAY

Approval Sheet

This thesis/dissertation/report entitled “Compositional and Interface Engineering of Hybrid Metal Halide Perovskite Thin Films for Solar Cells” by Kashimul Hossain (174120006) is approved for the degree of Doctor of Philosophy.

Examiners




Supervisor (s)


Chairman


Date: 17th May 2024

Place: IIT Bombay, Powai-400076, Mumbai, Maharashtra, India.

Declaration

I declare that this written submission represents my ideas in my own words and where others' ideas or words have been included, I have adequately cited and referenced the original sources.

I also declare that I have adhered to all principles of academic honesty and integrity and have not misrepresented or fabricated or falsified any idea/data/fact/source in my submission. I understand that any violation of the above will be cause for disciplinary action by the Institute and can also evoke penal action from the sources which have thus not been properly cited or from whom proper permission has not been taken when needed.


(Signature)

Kashimul Hossain

(Roll No. 174120006)

Date: 17th May 2024 we replace the opaque (Ag) electrode with a semi-transparent electrode (IZO) of the best performing opaque PSC to fabricate the semi-transparent solar cells. We demonstrate an efficient four-terminal tandem solar cell architecture of PSCs with silicon (Si) and cadmium telluride (CdTe) solar cells. In addition, we observed the lowest dark current for specific mixed (9:1) HTL, which is a prerequisite in photodetector applications. Therefore, a detailed photodetection analysis is discussed to study the scalable photodetector device. This thesis thoroughly studies perovskite compositional and interface engineering via various optoelectronic measurements. An in-depth device physics is discussed to study the interfacial defects between the charge transport and the perovskite layers. This thesis will be helpful in exploring a new class of perovskite materials and interface modification engineering for fabricating reproducible, stable, and highly efficient hybrid organic-inorganic metal halide perovskite solar cells.

Publications included in this thesis:

  1. K. Hossain, S. Singh, and D. Kabra “Role of Monovalent Cations in the Dielectric Relaxation Processes in Hybrid Metal Halide Perovskite Solar Cells” ACS Appl. Energy Mater. 2022, 5 (3), 3689–3697.

  1. K. Hossain, A. Kulkarni, U. Bothra, B. Klingebiel, T. Kirchartz, M. Saliba, and D. Kabra “Resolving the Hydrophobicity of the Me-4PACz Hole Transport Layer for Inverted Perovskite Solar Cells with Efficiency >20%>20\%” ACS Energy Lett. 2023, 8 (9), 3860–3867.

  1. K. Hossain, D. Sivadas, D. Kabra, and P. R. Nair “Perovskite Solar Cells Dominated by Bimolecular Recombination - How Far Is the Radiative Limit?” ACS Energy Lett. 2024, 9, 2310–23127.

  1. A. Paul{sup}†, A. Singha{sup}†, K. Hossain{sup}†`, S. Gupta, M. Misra, S. Mallick, K. R. Balasubramaniam, A. Munshi, and D. Kabra “4-T CdTe/Perovskite Thin Film Tandem Solar Cells with Power Conversion Efficiency >24%>24\%” ACS Energy Lett. 2024, 9, 3019–3026. († indicates equal contribution to the work).

  1. K. Hossain, B. Bhardwaj, and D. Kabra “Low Dark Current with High Speed Detection in a Scalable Perovskite Photodetector” Device. 2024, 2, 100513.

  1. K. Hossain, S. Nayak, and D. Kabra “Challenges and Opportunities in High Efficiency Scalable and Stable Perovskite Solar Cells” Appl. Phys. Lett. 2024, 125, 170501–170510.

  1. K. Hossain, A. Paul, S. Gupta, S. Koul, V. Chityala, and D. Kabra “Highly Efficient Four Terminal Tandem Solar Cells and Optical Loss Analysis” (Due for submission).

Miscellaneous Publications:

  1. B. Sharma, S. Singh, K. Hossain, Laxmi, S. Mallick, P. Bhargava, and Kabra, D. Additive Engineering of 4, 4′-Bis (N-Carbazolyl)-1, 1′-Biphenyl (CBP) Molecules for Defects Passivation and Moisture Stability of Hybrid Perovskite Layer. Sol. Energy 2020, 211, 1084–1091.

  1. T. S. Chandra, A. Singareddy, K. Hossain, D. Sivadas, S. Bhatia, S. Singh, D. Kabra, and P. R. Nair “Ion Mobility Independent Large Signal Switching of Perovskite Devices” Appl. Phys. Lett. 2021, 119 (2).

  1. N. Gaur, M. Misra, K. Hossain, D. Kabra “Improved Thermally Activated Delayed Fluorescence based Electroluminescent Devices using Vacuum Processed Carbazol based Self-Assembled Monolayer” ACS Energy Lett. 2024, 9, 1056-1062.

Patents:

  1. D. Kabra, K. Hossain, and N. Gaur “Tailored Work Function Based Hole Transport Layer in Optoelectronic Devices”. (Application Number: 202321054565).

  1. A. Paul, A. Singha, K. Hossain, S. Gupta, and D. Kabra “Perovskite/CdTe Thin Film Solar Cells Fabrication Methods and Thereof”. (IITB Ref. No. PAT/PH/I12115-2/23-24)

Presentations made at the Conference and symposium:

  1. Resolving the Hydrophobicity of the Me-4PACz Hole Transport Layer for Inverted Perovskite Solar Cells with Efficiency >20%>20\%. SymPhy, Organised by IIT Bombay, India, 2024. (Oral presentation).

  1. Role of the Tailored Hole Transport Layer in the Inverted Architecture Based Perovskite Solar Cells. NaMoSBio. Organized by the IISER Kolkata, India. 2024. (Poster presentation- won the best poster award)

  1. Role of Monovalent Cation in the Dielectric Relaxation Process and Correlation of Defects with the Thermal Stability of Hybrid Metal Halide Perovskite Solar Cells. 2nd International Conference on Materials for Humanity (MH 22), Organized by MRS, Singapore, 2022 (Poster presentation).

  1. Mixing Engineering of SAM, Me-4PACz with Ionic Polymer to Eliminate the Hydrophobic Issue of Me-4PACz/Perovskite Interface for Highly Efficient Reproducible PSCs. IUMRS – ICA 2022. Organized by the IIT-Jodhpur, India. 2022, (Poster presentation).

  1. Perovskite –Organic Monolithic Tandem Solar Cells Optimization and Progress. SUNRISE Symposium, 2021. (Online-Oral presentation).

Acknowledgments

I am profoundly grateful to my supervisor, Prof. Dinesh Kabra, whose unwavering support and guidance have been pivotal to the successful completion of my PhD. His constant motivation and encouragement for research in the area of perovskite photovoltaics always helped me in the ups and downs of my research work. I feel incredibly fortunate to have had such a dedicated mentor during my PhD career, and I sincerely appreciate his constant support and mentorship during those years. He always gave me the freedom to choose the research problem and encouraged independent thinking. His broad perspective on research problems during our discussions continually enhanced my understanding of advanced optoelectronic measurements. His extensive knowledge of optoelectronic devices greatly enriched my expertise during our research conversations. His spirit and passion for research have always motivated me and will carry with me in my future life regardless of the field I pursue. Besides the research guide, he is a lovely person and a source of positive energy with huge motivation. I am always grateful to him for giving me confidence and motivating me to make the correct decisions.

I am thankful to Prof. Pradeep R. Nair from the Department of Electrical Engineering, IIT Bombay, for the fruitful discussions, collaborative work, and scientific advice during my PhD candidature. I am thankful to Prof. K. L. Narasimhan from the Department of Electrical Engineering, IIT Bombay, for his constant helpful research feedback and scientific advice during my PhD candidature. I am thankful to my research progress committee members, Prof. Aftab Alam and Prof. Md. Aslam, for their valuable feedback and suggestions during my annual progress seminar. I am thankful to Prof. Aldrin Antony from Cochin University of Science and Technology for the fruitful discussions and suggestions during his stay at the Department of Electrical Engineering, IIT Bombay. I thank Prof. Michael Saliba, Prof. Thomas Kirtchaz, and Dr. Ashish Kulkarni for the collaborative work.

I would like to pay my deep gratitude to the IIT Bombay, National Centre for Photovoltaic Research and Education (NCPRE), Industrial Research and Consultancy Centre (IRCC), Sophisticated Analytical Instrument Facility (SAIF), UKRI Global Challenge Research Fund project, SUNRISE (EP/P032591/1), Indo-Swedish joint project funded by DST-India (DST/INT/SWD/VR/P-20/2019).

I am thankful to Dr. Shivam Singh for his immense help in the lab work, discussions, and collaboration. I am thankful to Dr. Amrita Dey for helping me learn spectroscopic measurements. I thank Dr. Ashish Kulkarni for discussing the research problems and collaborative work on the self-assembled monolayer-based hole transport layer. I thank Dr. Dhanashree Moghe for the fruitful discussions on spectroscopic measurements. I thank Dr. Bosky Sharma for the collaborative work and for discussing research problems. I thank Mr. Saurabh Gupta and Mr. Ananta Paul for the tin oxide layer deposition using the ALD process and transparent electrode IZO deposition using the sputtering technique, respectively, for the semi-transparent perovskite solar cell fabrication. I thank Dr. Sakshi Koul for the optical loss analysis calculation in the tandem solar cells architecture. I thank Prof. Amit H. Munshi from Colorado State University for providing the CdTe solar cells to study the four terminal CdTe – Perovskite tandem solar cells.

I am grateful to have some fabulous colleagues, juniors, and friends in and out of the lab: Laxmi, Venkatesh, Shiwani, Gopa, Urvashi, Gangadhar, Kalyani, Sumukh, Neha, Phalguna, Surya, Chinmay, Dr. Sakshi, Dr. Mohit, Nrita, Manas, Saurabh, Preetam, Ananta, Abhijit, Shreyasi, Bhupesh, Vijay, Dr. Deeksha, Johns, Subir, and Karthik for research work discussions and wonderful memorable moments. Apart from the lab, life would have been incoherent in the IIT Bombay campus without my exuberant friends Dibarkar, Monish, Bikash, Dr. Poulab, Manisha, Sucharita, Pritam, Jadupati, Gurudayal, and Abhimanyu. I am also thankful to my close friends Nur, Oyaresh, Javed, Baban, Abhisek, Akashneel, Subhojit, Sushmita, Sujit, Moon, Isha, Ansar, Muktar, Selim, Samrat, Bapi, Nadia, Praveen, Kamlesh, Bhanu, Ankit, Manendra and Pintu.

I am deeply grateful to my favourite teacher, Tr. Paritosh Chandra Roy, who first ignited my passion for mathematics and consistently motivated me to pursue higher education. I extend my heartfelt thanks to my favourite dedicated mentors, Tr. Koushik Chakraborty, Prof. Ashok Ghosh, Prof. Chinmoy Basu, and Prof. Rabiul Islam for their exceptional teaching, immense support, and unwavering encouragement.

I am eternally grateful to my grandmother, Fatima Bewa, for her unwavering support in every circumstance, which has been truly invaluable. I am profoundly grateful to my beloved wife, Tania, for her steadfast support and for standing by me through every situation. I am deeply grateful to my parents (Mr. Jiaur Rahaman & Mrs. Kashima Bibi), Cousins (Sonali, Sanjib, Monisha, Samrat, Neha, Martuj, Nilufa, Nisha, Josim), relatives (Mr. Ajijul Ali, Mr. Aynat Ali) and in-laws (Mr. Tafiqueuddin Ahmed, Mrs. Maiful Khatun, Mrs. Mim, Mr. Alam, Mihan, Arman, Jewel) for their unconditional love, immense support and care without which my life would not have been flourished.

I dedicate this thesis to my late grandparents.

Introduction to the perovskite solar cells

Introduction to the perovskite solar cells

Background

Energy plays a crucial role in sustaining life and supporting various activities in our daily lives. Energy in the form of heat, light, chemical, mechanical, electrical, etc., is essential to meet our diverse demands in daily life needs. Still, we depend on non-renewable energy sources such as coal, fossil fuels, oil and natural gas, etc. But these are limited and will end very soon. Besides the limited resources, consuming these energies leads the world to dangerous situations like climate change and global warming, which are further uninhabitable for humans and other creatures.[1] Keeping this in mind, using renewable energy sources is being started with wind, biomass, geothermal, solar light energy, etc. The worldwide energy generation in 2020 data indicates that non-renewable energy sources are still dominant over renewable sources Figure 1a.[2] Among all the 11.73% renewable energy applications, the worldwide 36% contribution comes from the solar PV industry in 2020 Figure 1b.[2] Recently, the International Energy Agency published data on the share of power capacity from 2010 to 2027, and it is shown that solar power will lead to energy production in the near future because of the continued presence of sunlight and the growing PV industry Figure 1c.[3] The dependency on energy will increase daily as the population is increasing linearly with time.[4] Hence, the security of the energy requirement is essential to live a hassle-free, sustainable life.

Figure 1. (a) Pie chart of all mixed energy generation worldwide.[2] (b) Pie chart of likely worldwide renewable energy generation.[2] (c) Share of power capacity in the upcoming future.[3]

As per the report proposed by the Ministry of Power, Government of India, in February 2019, the projected gross electricity generation during the year 2029-2030 will be 23% from solar photovoltaic renewable energy Figure 2a.[5] Long-term studies observed that coal-based plants capacity expansion is insignificant. In the near future, renewable energy production will lead to green energy production Figure 2b.[5] India projects to reduce the use of coal by 50% in 2030 by promoting renewable green energy production. As per the recent report on 31.10.2023 by the Ministry of New and Renewable Energy, Government of India, the total solar power installed in India is 72.02 GW.[6] The annual solar power installation is increasing rapidly in India to avoid an energy crisis from the limited conventional resources. Currently, the Indian government is highly enthused by the green energy revolution, especially in the solar cell energy sector Figure 2b.

Pie chart of likely gross generation capacity and long term generation planning

Figure 2. (a) Pie chart of likely gross generation (831 GW) capacity in 2029-2030 (Source: Central Electricity Authority, Government of India).[3] (b) Long term generation planning of the mixed energy sources.

Figure 3. Best research solar cell efficiency chart from NREL. [7]

This thesis will focus on solar energy production considering the geographical locations. There are diverse ways to harvest solar energy, such as solar water heater[8], solar cooker[9], concentrator solar power, i.e., steam engine[10], photovoltaic (PV) or solar cell[11]. Among all, solar cell technology is efficient and popular in renewable green energy production. Even in solar cell technology, numerous photovoltaic technologies evolved, such as crystalline silicon (Si) solar cells[12], single junction GaAs[13], multi-junction cells[14], CIGS[15], CdTe[16], Amorphous Si:H[17], thin film solar cells[18], etc. Si-based solar cells are commercialized for terrestrial applications, and continuous efforts have been made to reduce manufacturing costs. However, there is not enough space to reduce the manufacturing cost, and the market is almost saturated.[19] Therefore, a new photovoltaic technology must be adopted to harvest cost-effective solar energy where the material and manufacturing costs are cheaper and easy to fabricate. The emerging photovoltaic technology stands as one of the promising candidates to harvest cost-effective solar energy. Currently, the emerging PV technologies are dye-sensitized solar cells[20], organic cells[21], quantum dot cells[22], inorganic (CZTSSe) cells[23], perovskite cell[24], etc. Among all emerging thin-film PV, the perovskite solar cells (PSCs) are the fastest-growing and show a huge jump in efficiency from 3.8% to ≥25%\ge25\% in just one decade, which is close to the market-leading single junction crystalline Si solar cell Figure 3.[25, 26, 27] This motivates people to engage more in the perovskite photovoltaic devices. Recently, perovskite solar cells become very popular due to their application in tandem solar cell architecture. A tandem solar cell, also known as a multi-junction solar cell, consists of multiple semiconductor layers with different bandgaps stacked on top of each other. Each layer is designed to absorb a specific portion of the solar spectrum, allowing the cell to capture a broader range of sunlight and convert it into electricity more efficiently.[28] Currently, the Si perovskite tandem solar cell has reached an efficiency of >33%, which promises the future of more efficient light-to-electrical energy conversion.[29, 30, 31]

Levelized cost of energy (LCOE)

Apart from the efficiency, it is essential for the investor/industry to understand the lifespan and production cost of the cells before starting the business. Levelized cost of energy (LCOE) is a techno-economical term that is being used to evaluate the cost of energy production rate considering the material cost, fabrication cost, efficiency of the cell, lifetime of the cell, etc. The levelized cost of energy is defined as

LCOE=sum of costs over lifetimesum of the energy produced over lifetime\mathit{LCOE} = \frac{\textit{sum of costs over lifetime}}{\textit{sum of the energy produced over lifetime}}
LCOE prediction and cost analysis of tandem solar cells

Figure 4. (a) LCOE prediction of the perovskite photovoltaics over the last seven years. (b) Cost analysis of the perovskite-silicon and perovskite-perovskite tandem solar cells.

Recently, it is shown that the LCOE of the perovskite solar cells varies from 3 to 6 US cents per kWh−1\mathrm{kWh}^{-1} Figure 4a.[32] Whereas, for the Si solar cell, the LCOE decreased from $76 to 30 US cents in 2015 and is almost saturated.[33] Another report by Zongqi et. al. in 2018 shows that the LCOE of the Si–perovskite tandem solar cell comes down to 5.22 US cents per kWh−1\mathrm{kWh}^{-1} and further decreased to 4.22 US cents kWh−1\mathrm{kWh}^{-1} for the perovskite–perovskite tandem cell Figure 4b. [34] Interestingly, the LCOE of perovskite is much lower than that of commercialized Si solar cells and will be further decreased by increasing stability, i.e., the lifespan of perovskite solar cells. The cost-effective production of energy for the perovskite solar cells grabbed much attention in the solar cells community, and companies like Microquanta Semiconductor (China), RenShine Solar (China), Kunshan GCL (China), Hanwha Q CELLS (South Korea), Oxford PV (UK & Germany), and so on involved into R&D for aiming commercialization of perovskite solar cells in single junction and tandem device architecture. Microquanta demonstrated 20.2% efficiency on 20 cm2^2 solar cells area and 14.24% efficiency for large area (200 cm ×\times 800 cm) solar modules. [35] RenShine Solar achieved a certified conversion efficiency of 19.42% on a 30 cm ×\times 40 cm perovskite solar module and passed IEC61215 reliability tests. [36] Kunshan GCL showed an efficiency of 19.04% on the single junction perovskite module of area 100 cm ×\times 200 cm, and aiming for the world's first large-scale 2GW perovskite production line. [37] However, these recipes are not available in the public domain. Evidently, further progress is required to minimize the LCOE for commercialization of perovskite solar cells. [34]

Figure 5a shows the perovskite panel of commercial components α\alpha, size 1245 ×\times 635 ×\times 6.4$ mm by Microquanta, with a 12-year linear power output warranty along with a 25-year product material and workmanship warranty. [38] As per release time (03/12/2023) by Microquanta, they demonstrated the 1 MW perovskite ground–mounted photovoltaic power station. [39] They showed the world’s largest perovskite grid-connected ground power station, truly realizing the large-scale manufacturing and application of photovoltaic technology Figure 5b. [39]

Perovskite commercial component and ground-mounted photovoltaic power station

Figure 5. (a) Microquanta Semiconductor demonstrated perovskite commercial components α\alpha, size 1245 ×\times 635 ×\times 6.4$ mm: 25-year product material and workmanship warranty, 12-year linear power output warranty. [38] (b) World’s largest perovskite grid connected ground power station on 03/12/2023, demonstrated by Microquanta Semiconductor. [39]

Importance of perovskite solar cells

Perovskite is a class of semiconducting material with the crystal structure of ABX3\mathrm{ABX}_{3}, similar to the calcium titanium oxide (CaTiO3\mathrm{CaTiO}_{3}) mineral. The mineral was discovered by Prof. Gustav Rose in 1839 and is named after the Russian mineralogist Lev Perovski. The crystal structure was discovered by Victor Goldschmidt in 1926 in his tolerance factor work[40] and later published in 1945 from X-ray diffraction data on barium titanate (BaTiO3\mathrm{BaTiO}_{3}) by crystallographer Helen Dick Megaw.[41] In the ABX3\mathrm{ABX}_{3} crystal structure, A and B are the positively charged atoms/molecules of different sizes, and X is the halide. Figure 6 is the schematic of the typical crystal structure where the black ball at the center of the crystal represents the monovalent A cations (Cs+\mathrm{Cs}^{+}, FA+\mathrm{FA}^{+}, MA+\mathrm{MA}^{+}), the blue balls at the center of each octahedral represent the divalent B cations (Pb2+\mathrm{Pb}^{2+}, Sn2+\mathrm{Sn}^{2+}) and the red balls at the corner of the octahedral represents X halides (I−\mathrm{I}^{-}, Br−\mathrm{Br}^{-}, Cl−\mathrm{Cl}^{-}).[42]

Figure 6. Schematic of the ABX3\mathrm{ABX}_{3} perovskite crystal structure where the black ball at the center of the crystal represents the monovalent A cations (Cs+\mathrm{Cs}^{+}, FA+\mathrm{FA}^{+}, MA+\mathrm{MA}^{+}), the blue balls at the center of each octahedral represent the divalent B cations (Pb2+\mathrm{Pb}^{2+}, Sn2+\mathrm{Sn}^{2+}), and the red balls at the corner of the octahedral represents X halides (I−\mathrm{I}^{-}, Br−\mathrm{Br}^{-}, Cl−\mathrm{Cl}^{-}).

A class of perovskite materials can be developed by choosing various combinations of the different compositions of A, B cations, and X halides. To get a stable perovskite crystal structure, the choice of the A, B, and X is such that the crystal structure must follow the Goldschmidt tolerance factor (tt) and octahedral factor (μ\mu) rules.[43, 44]

The tolerance factor (tt) limit

0.81<t=RA+RX2(RB+RX)<1.110.81 < t = \frac{R_A + R_X}{\sqrt{2}(R_B + R_X)} < 1.11

Octahedral factor limit

0.44<μ=RBRX<0.90.44 < \mu= \frac{R_B}{R_X} < 0.9

Where RAR_A, RBR_B and RXR_X are the radii of the corresponding ions.

In an ideal case, the cubic perovskite crystal structure can be obtained when the tolerance factor t≈1t \approx1.[45] In this case, the size of the A cation must be larger than the B cation. Generally, the B site is occupied with large Pb or Sn atoms for the halide perovskite, so A cation must be much larger than B.[46] However, Cs is almost the largest group-I element in the periodic table but still not large enough to hold the cubic perovskite crystal structure, so it must be replaced with a larger atom/molecule. However, various additive engineering chemicals are developed to maintain the stable cubic crystal structure of the CsPbI3_3 perovskite.[47] Therefore, the A cation is chosen as an organic molecule, such as CH6_6N+^+, CH5_5N2+_2^+, CH6_6N3+_3^+, etc., to obtain a stable organic-inorganic halide perovskite crystal structure without additives.

Perovskite is a direct bandgap semiconductor with a high optical absorption coefficient, making it superior to the other commercialized silicon (Si) and gallium arsenide (GaAs) solar cells. The absorption coefficient of the perovskite material is 104−10610^4 - 10^6 cm−1^{-1}, which is higher than the commercialized Si, Ge, etc., semiconductor.[48] This results in a 300-600 nm thick perovskite layer that is sufficient to absorb almost all the 1- Sun light; hence, it becomes a potential candidate for solar cell application.[49, 50] Whereas the thickness of the first generation (silicon) solar cell absorber layer is 100-300 μ\mum[51, 52, 53] and the thickness of the second generation (GaAs) solar cell is 1-3 μ\mum[54, 55, 56]. The spin-orbit coupling plays an essential role in the band structure, hence the optoelectronic properties.

Figure 7. The schematic of the optical absorption of the (a) first generation, (b) second generation, and (c) halide perovskite solar cell absorber.[46] The GaAs absorber has been chosen as a second generation, and MAPbI3_3 as a halide perovskite solar cell absorber.

The optical absorption process for the first generation, second generation, and the perovskite solar cell absorber is shown in Figure 7. Figure 7a For the first generation (Si) solar cells, the optical absorption takes place between the Si pp orbital and the Si pp and ss orbital. However, silicon is an indirect bandgap semiconductor whose transition probability between the band edges is two orders of magnitude lower than that of the direct bandgap semiconductor. This results in the absorber layer being two orders of magnitude thicker and increasing the cost of solar cell fabrication. Despite being the indirect band gap of the Si absorber, the most commercialized solar cell in the PV community is the Si solar cell. The reason is that the Si absorber is fabricated from the sands by Czochralski (CZ) crystal growth or float zone(FZ) growth at temperature ∼1400∘C\sim1400^\circ\mathrm{C}.[57, 58] At such a high temperature, the Si absorber crystal growth is so pure that the charge carrier lifetime can be high as ∼1\sim1 ms, diffusion length 100100-300300 μm\mu\mathrm{m}, results efficient reproducible Si solar cells fabrication.[59] However, all these fabrication processes are expensive and require alternative approaches such as low cost and easy fabrication of perovskite solar cells. The bandgaps of the second generation (GaAs) and perovskite absorber are direct and hence possess stronger optical absorption than that of Si solar cells. The absorption coefficient of the perovskite absorber is one order of magnitude higher in the visible range than GaAs due to strong spin-orbit coupling.[60] The electronic structure of GaAs and perovskite absorbers are different. The valence band of GaAs is made up of p orbitals of As, and the conduction band is composed of the s orbital of Ga and the s orbital of As Figure 7b. However, the valence band of the perovskite is formed of the s orbital of Pb and p orbital of halides (Cl−\mathrm{Cl}^{-}, Br−\mathrm{Br}^{-}, I−\mathrm{I}^{-}), whereas the conduction band is made up of p orbital of Pb Figure 7c. The conduction band of GaAs is primarily made up of dispersive s orbital, whereas for the perovskite, it is made up of degenerate p orbitals. The atomic p orbitals are more localized than the s orbitals. As a result, the density of states (DoS) at the bottom of the conduction band is significantly higher for the perovskite than for the GaAs absorber.[46] The intra-atomic transition occurs between the valence band (Pb s) and conduction band (Pb p), facilitating a higher transition probability in the perovskite absorber. Hence, the optical absorption strength is higher in the perovskite than in GaAs. Apart from the high absorption coefficient, the other optoelectronic properties such as easy bandgap tunability (1.2 eV–3.2 eV), low binding energy (∼kBT\sim k_{B}T), long diffusion length (∼μm\sim\mu\mathrm{m}), higher charge carrier mobility (10110^{1}–10210^{2} cm2V−1s−1\mathrm{cm}^{2}\mathrm{V}^{-1}\mathrm{s}^{-1}), low temperature (100–200°C) solution and thermal evaporation processable technique promise the future of perovskite solar cells in the PV community.[61, 62]

Motivation

To commercialize any PV technology, the essential requirements are (i) reproducible high efficiency, (ii) long-term stability, and (iii) scalability. The perovskite solar cells have already passed the efficiency number compared to the commercialized silicon solar cells, but long-term stability and scalability are still significant challenges. Most of the reported perovskite solar cells have an active area of ∼0.1 cm2\sim0.1\ \mathrm{cm}^{2} and the anti-solvent approach is applied to control the crystallization to get the pinhole free and smooth perovskite film by solution processable technique. However, for scalable perovskite solar cells, the pinhole free uniform deposition of the perovskite layer over a large area is challenging using anti-solvent treatment. Hence, additive engineering, spray coating, thermal evaporation, screen printing, doctor blades, and carbon-based devices can be alternative approaches. However, the thermal evaporation-based perovskite layer deposition technique is promising for efficient and reproducible device fabrication.[63] Apart from efficiency, stability is essential to commercializing solar cells. Various engineering techniques are adopted to increase the solar cell devices moisture/thermal stability, such as additive engineering using ionic liquids or small molecules, interface passivation using 2D materials, encapsulation of the device using epoxy, etc.[64, 65] There are various ways to measure the stability of the solar cells, such as maximum power point tracking (MPPT) under continuous illumination, moisture stability at relative humidity (RH) of 50% or 85%, thermal stability at 65°C or 85°C and these are classified using standard ISOS protocols.[66] The interface between the perovskite and charge transfer layers plays a crucial role in the devices performance and stability operation. The charge transfer layers should be chosen to match the energy band well with the perovskite layer. However, during the stability operation study under continuous illumination, the halide ions come at the interface, which impedes the charge collection and deteriorates the photovoltaic performance. So, proper selection of the charge transfer layers and additive engineering solve such problems. Furthermore, the ITO/FTO substrate sheet resistance is also problematic in large-area devices, and hence, a mini-module kind of architecture is adopted in large-area devices.[67, 68] However, apart from the device engineering, the photophysics of the perovskite materials surprises every day. A few of them are that the perovskite semiconductor has a positive temperature coefficient[69, 70], double emission of PL peak at low temperature,71 contrast behaviour of enhanced PL intensity vs. decreased charge carrier lifetime at low temperature with respect to room temperature,72 photon recycling vs. defect-mediated recombination,73 the origin of non-radiative recombination74, the role of ions vs. ferroelectricity,75, etc. Perovskite semiconductors are full of defects, unlike silicon, and possess disorders (static or dynamic), dislocation, distortion, ion migration, and phase segregation (mixed halide perovskites), resulting in the unique physics of this novel class of semiconductor.76,77 Hence, it is necessary to understand these imperfections (defects/disorders) by various optoelectronic measurements to improve the quality of the perovskite material further and understand the device physics.

Outline of the thesis

This thesis focuses on the compositional and interface engineering of hybrid metal halide perovskite thin film solar cells. Chapter 2 demonstrates the fundamentals of solar cells with a literature survey. This chapter covers the challenges and future perspectives of fabricating scalable, efficient perovskite solar cells. Chapter 3 describes the perovskite thin film/solar cell fabrication procedure and the characterization techniques to support our experimental results and outcome. Chapter 4 describes the role of monovalent (A) cation in the ABX3ABX_3 perovskite crystal structure in terms of the dielectric relaxation process and correlates with the thermal stability of the perovskite solar cells. This chapter will introduce the study of dielectric relaxation and defects using frequency-dependent photocurrent measurement of a complete solar cell device and eventually correlate with the thermal stability of perovskite solar cells. Chapter 5 deals with the hydrophobic self-assembled monolayer (SAM) hole transport layer in the inverted (pp-ii-nn) architecture-based device. In this chapter, we studied the mixing engineering strategy of SAM with polyelectrolyte polymer and tailored the work function of the mixed HTL for 1.6 eV perovskite solar cells. The interface of the device is modified for different mixing ratios of HTL, and associated device physics is discussed. However, the device shows excellent efficiency performance, the lowest dark current, and the ideality factor close to the unity. The lowest dark current is a prerequisite for a photodetector application. Hence, we carried out a photodetector study of the best-performing HTL-based device in Appendix A. However, the unity ideality factor is a primary signature of solar cells working near the radiative limit. Hence, Chapter 6 deals with the dominant recombination mechanism on a complete solar cell using steady-state and transient measurements. We developed a consistent characterization scheme for studying a solar cell operating near the radiative limit. Chapter 7 describes the application of perovskite solar cells in the four-terminal tandem solar cell structure. In this case, we replaced the opaque (Ag) electrode with a semi-transparent electrode (IZO) to fabricate semi-transparent solar cells and use it in the four-terminal tandem architecture with Si and CdTe solar cells. Finally, chapter 8 concludes all the research work in the thesis and demonstrates the future outlooks related to this thesis.

References

369.

Fundamentals of solar cells and literature survey

Fundamentals of solar cells and literature survey

Introduction

Chapter 1 describes the motivation and outlines of the thesis to understand the compositional and interface engineering of hybrid organic-inorganic metal halide perovskite thin film solar cells. The fundamentals of solar cells and the literature survey will be discussed in Chapter 2. This chapter will cover the literature review on the fundamental aspects of perovskite solar cells and their progress in terms of efficiency, stability, and scalability. This chapter also justifies the need for work on the research problems addressed in this thesis.

Working principle of solar cells

A solar cell is an optoelectronic device that transduces light energy to electrical energy through the photovoltaic effect. The photovoltaic (PV) effect was first discovered by the French physicist Alexandre Edmond Becquerel in 1839 while investigating the light interaction with electrochemical cells.[1] After that, it took 115 years to demonstrate the 6% efficient PV solar cell using silicon (Si) absorber by C.S. Fuller and G. Pearson at Bell Laboratories in 1954.[2] After that, tremendous progress can be seen in the PV solar cell field as a consequence of the first-generation, second generation and today emerging PV technology came in the solar cell community.[3] The performance of the solar cells depends on the semiconductor used in the device and the corresponding fundamental mechanisms involved in it. The fundamentals of the solar cell are discussed below.

When photons are absorbed by a semiconductor absorber of energy equal to or more than the energy bandgap (EgE_g), the electron and hole pairs are generated at the conduction and valence bands respectively. The incident sunlight consists of photons with energies ranging from infrared to visible to ultra-violet region Figure 1a.[4] If the energy of the incident photon is less than the bandgap of the absorber, the photon will not be absorbed, leading to below bandgap Loss Figure 1b. If the energy of the incident photon is more than or equal to the bandgap of the semiconductor, it will be absorbed and create electron and hole pairs. For the photons of energy more than the bandgap, the excited electrons will go to the higher energy states of the conduction band, undergo various non-radiative fast relaxation processes, and eventually arrive at the bottom of the conduction band. These fast relaxation processes happen via phonon interaction and release heat to the semiconductor, referred to as thermalization loss.[5] However, the free electrons at the conduction band and the holes at the valence band take part in charge transport, i.e., current flow in the device. The interaction of light with the photovoltaic semiconductor absorber leads to (i) absorption/generation and (ii) recombination process.

Figure 1 (PDF p. 55). (a) 1-Sun solar spectrum (AM 1.5G).[4] (b) Schematic of absorption and generation of the charge carriers by optical excitation.

Absorption and Generation

For an efficient solar cell, the rate of generation of the electrons and holes should be as high as possible. The generation of electrons and holes in the solar cell occurs when the incident sunlight has energy higher than the bandgap of the absorber. However, the generation of the charge carriers in the semiconductor happens via direct band-to-band transition or through sub-gap states.[6] If the sub-gap states are close to the conduction or valence band, they are referred to as shallow trap states or tail states. However, if the sub-gap states lie in the middle of the energy band, called the mid-gap states.[7] The optical absorption process through direct band-to-band and sub-bandgap states is shown in Figure 2. The optical transitions through the direct band are called intrinsic transitions, whereas the transitions through the sub-gap states (due to impurity/defects) are referred to as extrinsic optical transitions.[8] The extrinsic optical transitions are not desirable for efficient solar cell application. Apart from the extrinsic transitions, the semiconductor materials can also be direct and indirect bandgap. In the direct bandgap semiconductor, the minimum of the conduction band and the maximum of the valence band are in the same crystal momentum space (k\mathbf{k}). In contrast, in the indirect bandgap, the momentum space is different.[9] The absorption coefficient α(λ)\alpha(\lambda) of the direct bandgap semiconductor is sharp at the band edge (Perovskite, CdTe, CuInSe2_2)[10, 11, 12] whereas for the indirect bandgap, it slowly increases (Si)[13] Figure 3.

Figure 2 (PDF p. 56). Optical absorption process through direct band-to-band and sub-gap states transition.[7]

The bandgap of a semiconductor as a function of the absorption coefficient is defined by the Tauc method as[14]

(αhν)1/γ=B(hν−Eg)(1)(\alpha h\nu)^{1/\gamma}=B(h\nu-E_g) \tag*{(1)}

Where hh is the Planck constant, ν\nu is the frequency of the photon, EgE_g is the energy bandgap of the semiconductor, and B is a constant. The factor γ\gamma depends on the nature of electron transfer and is equal to 2 for indirect and 1/2 for direct transition bandgaps respectively. The bandgap of a semiconductor is calculated using the above eq. (1).[9, 15]

Absorption coefficient of direct and indirect bandgap semiconductors

Figure 3 (PDF p. 57). Absorption coefficient of direct and indirect bandgap semiconductors.[13]

The sub-bandgap states do not contribute to the absorption coefficient significantly, but they provide information about the distribution of the band structure.

The Urbach energy (EuE_{\mathrm{u}}) is the measure of disorder in the bulk of the perovskite semiconductor and can be estimated from the absorption spectra using the following equation,[16, 17]

α(E)=α0exp⁡(E−EgEu)(2)\alpha(E) = \alpha_{0} \exp\left(\frac{E-E_{\mathrm{g}}}{E_{\mathrm{u}}}\right) \tag*{(2)}

Where EE denotes the photon energy, and α0\alpha_{0} is the material-dependent absorption coefficient.

Recombination

Upon optical excitation of a semiconductor, electrons and holes are generated in the bulk of the semiconductor. However, the generated electrons and holes recombine through various relaxation channels and decrease the charge carrier density. At steady state condition, the carrier dynamics under illumination) for an undoped semiconductor is,

∂n∂t=G−k1n−k2n2−k3n3(3)\frac{\partial n}{\partial t} = G-k_{1}n-k_{2}n^{2}-k_{3}n^{3} \tag*{(3)}

where ∂n∂t\frac{\partial n}{\partial t} is the rate of change of charge carrier density at steady state condition G is the generation of the carriers upon optical excitation, and the parameters k1k_{1}, k2k_{2}, and k3k_{3} denote the monomolecular, bi-molecular recombination, and Auger recombination rates, respectively. The recombination parameter k1k_{1} is contributed by trap-assisted recombination centers, also called Shockley Read Hall (SRH) recombination, k2k_2 is corresponds to the bimolecular recombination and k3k_3 is due to the Auger recombination process.

Figure 4 (PDF p. 58). Schematic representation of band-to-band, trap-assisted, and Auger recombination process.

The band-to-band recombination process merely involves the direct annihilation of the conduction band electron to the valence band hole and is significant in the direct bandgap semiconductor only. The electron jumps from the allowed conduction band minimum to the vacate valence band maximum, and the excess energy is released in the form of light of energy equal to the bandgap of the semiconductor. This recombination process is also referred to as bi-molecular recombination.[18] The trap-assisted recombination, also called the Shockley Read Hall (SRH) recombination, is a non-radiative process through the localized trap states. Since the traps can absorb the difference in momentum between the carriers, SRH is dominant in Si or other indirect bandgap semiconductor. However, such SRH can also be dominant in direct bandgap semiconductors under very low carrier densities (low-level injection) or materials with high trap density, such as perovskites. The Auger recombination is a three-particle interaction. The energy released by an excited state electron during band-to-band transition or trap-assisted transition can transfer to an electron in the conduction band and move to a higher energy state. However, this process is significant only when the charge carrier density is very high.

Shockley – Quisser (SQ) limit

In 1961, William Shockley and Hans-Joachim Queisser estimated the efficiency of 30% for a single junction solar cell of absorber material bandgap 1.1 eV using a 6000K black-body spectrum as an approximation of the solar spectrum.[19] Further, the theoretical maximum efficiency of 33.7% was achieved at a bandgap of 1.34 eV considering the AM 1.5G (1000 mW/cm2^{2}) solar spectrum.[20] The model is done considering the cell temperature at 300K and the sun temperature at 6000K. The device performance of the solar cell is modeled as a function of the bandgap of the absorber, as shown in Figure 5. According to the estimates, if the bandgap of the absorber is below 1.34 eV, then the energy loss is due to the thermalization relaxation process, and for a bandgap higher than 1.34 eV, the energy loss is due to the absorption loss or below bandgap loss Figure 5a. Each photon in the solar spectrum produces electron – hole pairs and results current flow in the solar cell. At higher bandgap, there are fewer photons in the solar spectrum, results in lower current which is attributed to spectrum loss Figure 5b. Figure 5c the red dotted line represents the VOCV_{\mathrm{OC}} of the solar cell and is equal to the bandgap of the absorber when there is no recombination, but due to radiative recombination it is lower represented by the black solid line.

SQ limit efficiency, short circuit current density, and open circuit voltage versus bandgap

Figure 5 (PDF p. 59). The plot between the (a) SQ limit efficiency, (b) SQ limit short circuit current density, and (c) SQ limit open circuit voltage vs. bandgap of the semiconductor.[19]

The SQ theoretical efficiency limit is based on the five assumptions as discussed below,

SQ limit assumptions:

(i) If the energy of the incident photon, hν>Egh\nu> E_g, the photon will be absorbed, whereas for photons of energy hν<Egh\nu< E_g, do not interact with the cell, i.e., the absorptivity A(hν)A(h\nu) of the cell is a step function, 0 for hν<Egh\nu< E_g, and 1 for hν>Egh\nu> E_g. Where, EgE_g represents the bandgap of the absorber.

  • (ii) The absorption of one photon generates exactly one electron-hole pair that contributes to the photocurrent at short-circuit conditions (JSCJ_{\mathrm{SC}}).

  • (iii) For the high-energy photons, the excess energy hν−Egh\nu- E_g, releases in the form of heat in femtosecond (fs), and the cell is always in equilibrium with the environment at 300K.

  • (iv) At this point, the generated electron–hole pair either get collected at their respective electrodes or recombine within the cell. The only allowed recombination in the cell is the radiative with emission of photon. The emitted photon further reabsorbs, creating a new electron-hole pair (photon recycling).[21] The relation between absorption and emission of photons in a semiconductor results from the principle of detailed balance limit. This can be described as every microscopic process must have the same rate as its inverse process in thermal equilibrium, or else thermal equilibrium could not be reached.[22]

  • (v) Each contact that exchanges only one carrier type (electrons or holes) with the absorber layer is ideal and has negligible resistance.

Violation of SQ limits in practical solar cells

In reality, no practical solar cell holds the assumption 1, i.e., the step function of the absorptivity, because, in semiconductor absorbers, there are always sub-gap states resulting from the impurity/defects. Further, the parasitic absorption of photons in contact layers or by free carriers (electrons) in the bulk of the absorber reduces the average number of photogenerated electron-hole pairs and violates assumption 2. In assumption 3, the thermalization losses release heat from the cell but remain in equilibrium with the surroundings, which is not possible in real life. In assumption 4, the allowed recombination is radiative, but the violation of assumptions 1 and 2 indicates the presence of sub-gap states and leads to non-radiative recombination.

Illustration of SQ-model and energy losses

Figure 6 (PDF p. 61). (a) Illustration of SQ–model as a function of bandgap (always at maximum power point) using AM 1.5G solar spectrum. (b) Energy losses for a given bandgap energy are depicted as a function of photon flux vs. photon energy (adapted from ref.[23]). By dividing the flux with elementary charge qq and multiplying the energy with qq, the axes can be current density vs. voltage (JJ vs. VV).

In assumption 5, the contact electrodes are assumed to be ideal with negligible resistance, but in reality, it is not possible due to the finite mobility or carrier concentration of the selective layer.

However, in the practical device single junction solar cells, the total current under illumination is due to a short circuit current (JSCJ_{SC}) along with diode current (JdJ_d) eq. 2.4. The diode current is dependent on the reverse saturation current density J0J_0 which is due to the recombination mechanism. Thus, the violation of assumptions 1 and 2 results in a lower short circuit current than the SQ limit. 0

J=JSC−J0[exp⁡(qVnkT)−1](4)J = J_{SC} - J_0 \left[\exp\left(\frac{qV}{nkT}\right)-1\right] \tag*{(4)}

Violation of assumptions 2, and 3 increases cell temperature as the recombination current is thermally activated, increasing exponentially with temperature. This increase in recombination current reduces the VOCV_{OC} of the device compared to the SQ limit. [23]

VOC=nkTqln⁡(JSCJ0+1)(5)V_{OC} = \frac{nkT}{q}\ln\left(\frac{J_{SC}}{J_0}+1\right) \tag*{(5)}

Further, violation of assumption 5 increases the series resistance of the device and reduces the fill factor (FFFF). Figure 6a shows the fraction of power as a function of bandgap considering SQ limit assumptions. fig:6b shows the SQ limit J−VJ-V plot considering SQ limit assumptions. Thus, in a single junction practical solar cell, the efficiency is always lower than the SQ limit. However, the efficiency can be improved in the practical solar cells by minimizing the defect states and choosing the right bandgap absorber. However, the model shows a multi-junction cell with an infinite number of absorber layers, the efficiency limit is 68.7% using AM 1.5 G solar spectrum and 86.8% using concentrated sunlight.[24, 25] The multi-junction technique has been applied in the various tandem solar cells and surpasses the single junction SQ limit efficiency.[3, 26]

I-V characteristics and PV parameters of solar cells

The fundamental way to test a solar cell is by measuring the current density vs. voltage (J−VJ-V) characteristics under dark and illumination conditions with AM 1.5G spectrum. The current flow in a solar cell under illumination (considering the single-diode model) can be expressed as[27]

I=−IL+I0[exp⁡(qVnkT)−1](6)I=-I_L+I_0\left[\exp\left(\frac{qV}{nkT}\right)-1\right] \tag*{(6)}

Where ILI_L is the illumination current, I0I_0 is the reverse saturation or leakage current, which flows in the opposite direction to the illumination current, and II is the total current. The nn represents the ideality factor, which lies between 1 and 2 (depending on the bulk and interface quality)[27] qq is the electronic charge, and kk is the Boltzmann constant. The numerical factor 1 can be avoided for a forward bias of 100 mV (or higher) as the diode current increase is exponential. On the other hand, at lower applied bias or in reverse bias, the illumination current ILI_L dominates over I0I_0. However, the device's total current depends on the active area of the solar cells. Hence, the current density (JJ) is used to estimate the device's efficiency. Thus the above expression can be written in terms of JJ as

J=−JL+J0exp⁡(qVnkT)(7)J=-J_L+J_0\exp\left(\frac{qV}{nkT}\right) \tag*{(7)}

The typical illumination J−VJ-V characteristic is shown in fig:7a, and the power conversion efficiency as a function of the applied bias voltage is shown in fig:7b.

The power conversion efficiency (PCE) of solar cells can be expressed as

PCE(%)=Pmax⁡Pin(100 mW/cm2)×100%(8)PCE(\%) = \frac{P_{\max}}{P_{\mathrm{in}}(100\ \mathrm{mW}/\mathrm{cm}^{2})} \times100\% \tag*{(8)}

Which can be written in terms of the PV parameters as

PCE(%)=JSC×VOC×FF (%)=Jmax⁡×Vmax⁡(9)PCE(\%) = J_{SC} \times V_{OC} \times FF\ (\%) = J_{\max} \times V_{\max} \tag*{(9)}

Where, JSCJ_{SC} is the short circuit photocurrent density, VOCV_{OC} is the open circuit voltage, and FFFF is the fill factor (Figure 7a). The FF is defined by the maximum square of the J−VJ - V curve, and it can be represented as,

FF=Jmax⁡×Vmax⁡JSC×VOC(10)FF = \frac{J_{\max} \times V_{\max}}{J_{SC} \times V_{OC}} \tag*{(10)}
Schematic representation of current density versus voltage characteristics and power plot for a typical perovskite solar cell

Figure 7 (PDF p. 63). (a) Schematic representation of current density vs. voltage (J−VJ - V) characteristics with the PV parameters for a typical perovskite solar cell. (b) The power plot as a function of applied bias voltage.

In practical solar cells, series (RsR_s) and shunt (RshR_{sh}) resistances are the two important parameters that affect the device's performance. The series resistance of a solar cell originates from the contact resistance between the absorber layer and the charge transport layer, along with the sheet resistance of the bottom and top contacts. However, the shunt resistance originates from the physical and electronic defects. [28, 29] The specific origin of electronic defects is yet to be known.82 However, the physical defects (point defects, grain boundaries, pinholes, dangling bonds at the surface of the film, etc.)[30, 31] can be prominent considering low temperature (~100º C) solution processable techniques widely used for perovskite layer deposition and chances of non-uniformity over the whole substrate.[29, 32, 33]

Therefore, upon applying the 1-Sun (100 mW/cm2^2) light, the photogenerated charges take place, and the total current in terms of RsR_s and RshR_{sh} can be written as [34]

J=−JL+J0[exp⁡(V−JRsηkT/q)−1]+V−JRsRsh(11)J=-J_L+J_0\left[\exp\left(\frac{V-JR_s}{\eta kT/q}\right)-1\right]+\frac{V-JR_s}{R_{sh}} \tag*{(11)}

These RsR_s and RshR_{sh} resistances can be evaluated from the illuminated J−VJ-V as follows

Rs=dVdJ at V=VOC(12)R_s=\frac{dV}{dJ}\ \text{at}\ V=V_{OC} \tag*{(12)}
Rsh=dVdJ at V=0(13)R_{sh}=\frac{dV}{dJ}\ \text{at}\ V=0 \tag*{(13)}

However, these RsR_s and RshR_{sh} parameters can be extracted more accurately from the dark current as the illuminated curve can be affected by slight fluctuations in light intensity, low diffusion length, poor collection properties, etc.

Under dark conditions, the eq. 2.11 becomes1

Jd=[exp⁡(V−JdRsηkT/q)−1]+V−JdRsRsh(14)J_d=\left[\exp\left(\frac{V-J_dR_s}{\eta kT/q}\right)-1\right]+\frac{V-J_dR_s}{R_{sh}} \tag*{(14)}

Figure 8. The typical dark current of an inverted (pp-ii-nn) architecture based perovskite solar cell of device active area 1 cm2^2.

At the low voltage, the dark current flows merely through the shunt path and is dominated by RshR_{sh} (region 1 Figure 8). The 2{sup}nd` term of eq. 2.14 represents the shunt current, and region 1 is used to measure the shunt resistance from the inverse of the slope. When the applied bias increases the dark current increases exponentially in region 2. This portion represents the recombination current, and the fitting of region 2 estimates the ideality factor (nn). Further increase of the applied bias the current is limited by RsR_{\mathrm{s}}, and it saturates at higher voltage regions, as represented by region 3. The inverse slope of region 3 estimates the value of RsR_{\mathrm{s}}.

Suns - VOCV_{\mathrm{OC}} characterization and pseudo JJ−VV estimation

The device performance depends on the series resistance of the solar cell device for instance larger areas of solar cells, as the FF gets affected the most. The Suns-VOCV_{\mathrm{OC}} measurement is a characterization technique of solar cells where the effect of series resistance is avoided, and the pseudo JJ−VV characteristic can be generated. The Suns-VOCV_{\mathrm{OC}} technique is widely used for the characterization of silicon (Si) solar cells.[36] This technique more properly can be termed as illumination VOCV_{\mathrm{OC}}.[37] The basic concepts of Suns-VOCV_{\mathrm{OC}} technique was outlined in 1963 but was not popular until Sinton et.al. in the year 2000 and the subsequent availability of the commercial tool.[38, 39] Generally, in the Suns-VOCV_{\mathrm{OC}} technique, the VOCV_{\mathrm{OC}} of a solar cells is measured with varying illumination intensity (1-Sun to 0.01 Sun.) Table 1. In the solar cells device, the short circuit current density (JSCJ_{\mathrm{SC}}) related to the intensity (PP) as JSC=PαJ_{\mathrm{SC}} = P^{\alpha}, where the exponent α≈1\alpha\approx1. Hence the intensity (PP) can be replaced by JSCJ_{\mathrm{SC}} at each intensity and a pseudo JJ−VV i.e (JSCJ_{\mathrm{SC}}−VOCV_{\mathrm{OC}}) can be constructed, where the voltage (VV) is the VOCV_{\mathrm{OC}} at each intensity. The detail application of the Suns-VOCV_{\mathrm{OC}} technique will be shown in Chapter 6. It is worth noting that, the JJ−VV plot (Figure 9a) is shown in the first quadrant for convenience (as Si solar cell community do) and the pseudo JSCJ_{\mathrm{SC}} estimated as per column D in Table 1. For better understanding we have shown the 1-Sun JJ−VV and the pseudo JJ−VV in the same figure (Figure 9a) in the first quadrant. The green plot represents the pseudo JJ−VV, where the voltage axis corresponding to VOCV_{\mathrm{OC}} and the current density axis is JSCJ_{\mathrm{SC}} at each intensity (Table 1). The reason for higher FF of the pseudo JJ−VV is due to avoiding the series resistance at VOCV_{\mathrm{OC}} point where the current flow in the circuit is zero (Figure 9b).

Col(A)Col(B)Col(C)Col(D)
IntensityJSCJ_{SC}VOCV_{OC}Pseudo JSC=Jsc,0−Col(B)J_{SC} = J_{sc,0} - \text{Col(B)}
I0I_0Jsc,0J_{sc,0}VOC,0V_{OC,0}Jsc,0−Jsc,0J_{sc,0} - J_{sc,0}
I1I_1Jsc,1J_{sc,1}VOC,1V_{OC,1}Jsc,0−Jsc,1J_{sc,0} - J_{sc,1}
I2I_2Jsc,2J_{sc,2}VOC,2V_{OC,2}Jsc,0−Jsc,2J_{sc,0} - J_{sc,2}
I3I_3Jsc,3J_{sc,3}VOC,3V_{OC,3}Jsc,0−Jsc,3J_{sc,0} - J_{sc,3}
I4I_4Jsc,4J_{sc,4}VOC,4V_{OC,4}Jsc,0−Jsc,4J_{sc,0} - J_{sc,4}
I5I_5Jsc,5J_{sc,5}VOC,5V_{OC,5}Jsc,0−Jsc,5J_{sc,0} - J_{sc,5}
I6I_6Jsc,6J_{sc,6}VOC,6V_{OC,6}Jsc,0−Jsc,6J_{sc,0} - J_{sc,6}
I7I_7Jsc,7J_{sc,7}VOC,7V_{OC,7}Jsc,0−Jsc,7J_{sc,0} - J_{sc,7}
I8I_8Jsc,8J_{sc,8}VOC,8V_{OC,8}Jsc,0−Jsc,8J_{sc,0} - J_{sc,8}
I9I_9Jsc,9J_{sc,9}VOC,9V_{OC,9}Jsc,0−Jsc,9J_{sc,0} - J_{sc,9}
I10I_{10}Jsc,10J_{sc,10}VOC,10V_{OC,10}Jsc,0−Jsc,10J_{sc,0} - J_{sc,10}

Table 1. Example of the pseudo J−VJ - V estimation from the intensity dependent J−VJ - V measurement.

Figure 9. (a) Comparison of the 1-Sun J−VJ - V and the pseudo J−VJ - V estimated from Suns–VOCV_{OC} measurement. The intensities varied for the Suns–VOCV_{OC} measurement is 1-Sun to 0.01 Sun. (b) Represents the zoomed X-axis for a better understanding of the FF.

Introduction to perovskite solar cells

Progress in perovskite solar cells

The journey of perovskite solar cells began with the concept of dye-sensitized solar cells (DSSCs). The historical evolution of hybrid organic-inorganic perovskite solar cells, starting from the DSSC is shown in Figure 10[40] In 2006 and 2009, Miyasaka et. al. used hybrid organometallic halide perovskite MAPbBr3\mathrm{MAPbBr_3} and MAPbI3\mathrm{MAPbI_3} in the DSSCs and achieved an efficiency of 2.6% and 3.8% respectively.[41, 42] In 2011, Park et. al. used MAPbI3\mathrm{MAPbI_3} of 2-3 nm sized nanocrystal and surface-treated TiO2\mathrm{TiO_2} and achieved an efficiency of 6.5%.[43] The electrolytes used in the DSSC are detrimental to the device as it has corrosion problems. In 2012, M. Gratzel and N.G. Park replaced the electrolyte with solid-state Spiro-MeOTAD and immersed perovskite into the TiO2\mathrm{TiO_2} scaffold.[44] Interestingly, the efficiency of the cell boosted to 9.7%.[44] In the same year, Snaith et. al. replaced the TiO2\mathrm{TiO_2} with electrically inert Al2O3\mathrm{Al_2O_3}, and the efficiency was boosted to 10.9%.[45] In the DSSC, the TiO2\mathrm{TiO_2} acts as an electron acceptor to assist electron transport. Snaith et. al. showed that the perovskite could transport electrons itself without scaffold TiO2\mathrm{TiO_2}.[45, 46] However, these findings indicate that the perovskite itself has good electrical transport properties and is promising for PV applications. Further, Snaith et. al. replaced the solution-processed perovskite deposition with thermal evaporation-based perovskite in the same device architecture and showed an efficiency of 15.4%. This indicates that the efficiency can be improved by improving the perovskite quality. Later, Gratzel and Kim et. al. used quantum dot (QD) SnO2\mathrm{SnO_2} as an electron transfer layer (ETL) and FAPbI3\mathrm{FAPbI_3} perovskite as an active layer, showing an efficiency 25.4%, which is comparable to the commercialized Si solar cells.[47] The above device structure originated from liquid-based DSSC, and it is a regular (nn-ii-pp) device architecture. In parallel, an inverted architecture (pp-ii-nn) was used by Guo et. al. in 2013 using PEDOT:PSS as HTL and PCBM as ETL, which showed an efficiency of 3.9%.[48] The pp-ii-nn architecture is beneficial as it deals with the low-temperature processed charge transport layers (CTLs) such as poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT: PSS), poly (triaryl amine) (PTAA), self-assembled monolayers (SAM), PCBM, ICBA, fullerene (C60\mathrm{C_{60}}), etc.[48, 49, 50, 51, 52, 53] Currently, the efficiency achieved on the pp-ii-nn architecture based PSC is 25%, similar to the nn-ii-pp architecture.[54] In addition, the pp-ii-nn architecture-based device is crucial in monolithic tandem solar cells such as perovskite-perovskite, perovskite–organic, and Si-perovskite tandem solar cells. Such a speedy journey of efficiency improvement of the perovskite solar cells (PSCs) (3.9% to >25%)[42, 55, 56] can be credited to the unique optoelectronic features of perovskites such as a high absorption coefficient, low exciton binding energy, and a tuneable bandgap.[57, 58, 59, 60, 61] Unprecedentedly, Etgar et. al. found that the MAPbI3\mathrm{MAPbI}_{3} perovskite could be pp-type[62] whereas You et. al. showed nn-type behavior[63]. Thus, MAPbI3\mathrm{MAPbI}_{3} is bipolarly dopable, and different processes lead to different doping behaviors (pp-type or nn-type). However, various perovskites can be prepared by choosing different mixing compositions of the A, B, and X ions in the ABX3\mathrm{ABX}_{3} perovskite. The underlying mechanism of perovskite is still unclear due to the low temperature processed deposition, which leads to point defects, making perovskite unintentionally doped. Perovskite is full of defects and disorder, but the unique properties of perovskite are that the dominating defects are not harmful as they do not create detrimental deep-level defect states within the bandgap.[64]

Historical evolution of hybrid organic-inorganic perovskite solar cells

Figure 10. Historical evolution of hybrid organic-inorganic perovskite solar cells, starting from the DSSC. (adapted from ref.[40])

Defects in perovskite semiconductor

Ideally, each atom in a crystal lattice should be located at its designated position and any deviation from this results in a defect state Figure 11a. The structural non-periodicity or any foreign atom in the periodic lattice introduces disorder into the crystal lattice bonding, and impacting the physical properties and density of states. Such disorders are not desirable for PV application as they produce localized sub-gap electronic states. When the charge carrier transitions happen, these disorders play a significant role as non-radiative centers Figure 11b. The trapping and detrapping of the charge carriers depend on the position of the trap states. If the trap states are located in the mid of the bandgap, they are called mid-gap states. If the trap states are located close to the conduction or valence band, they are called shallow traps. If the charge carriers trap at the mid-gap states, then it will be difficult to detrapped, however, if the charges trap at the shallow traps, then they can be detrapped by additional energy (activation energy), via further optical excitation or thermal energy (kT=25.7kT = 25.7 meV at room temperature) Figure 11c. A brief of the origin of various defects and defect passivation engineering will be discussed below.

Schematic diagrams of charge localization, charge trapping kinetics, and density of states in a disordered semiconductor

Figure 11. (a) Schematic of a charge localization caused by non-periodic disorder and defect in the lattice. (b) Schematic of charge trapping kinetics in a semiconductor, which inhabits charge transport by trapping and detrapping events. (c) Schematic representation of the density of states (DoS) is a disordered semiconductor whereby both the radiative (band-to-band) and non-radiative (trap-assisted) recombination can happen. (adapted from ref.65).

The efficiency of the PSCs can be limited due to the presence of non-radiative defect centres and uncontrolled uniform deposition of perovskite films over the large area substrates. The low temperature (~100° C) processed perovskite films are generally polycrystalline films with a thickness of ≈500\approx500 nm and create impurities in the bulk and surface of the film. The imperfections in the perovskite are classified into two categories: (i) intrinsic and (ii) extrinsic. The extrinsic imperfections are caused by environmental surroundings and unsaturated bonds, which result in polycrystalline networks, grain boundaries, line defects, surface defects (due to finite thickness), etc. However, the intrinsic defects are also known as point defects, which depend on the growth of the crystal structure under the chemical environment and the used composition of the perovskite system. These defects play the role of trap centers of the charge carriers, as discussed above, and result in the lower performance of the solar cell. Hence, it is essential to understand the origin of such defects and passivation mechanisms to obtain improved performance in perovskite solar cells.

Schematic of intrinsic and extrinsic defects in perovskite layer compared to the ideal crystal lattice

Figure 12. Schematic of intrinsic (int.) and extrinsic (ext.) defects in perovskite layer compared to the ideal crystal lattice. For all the crystal structures, green, black, and blue spheres represent A, B, and X elements in the ABX3\mathrm{ABX}_{3} perovskite crystal structure, while grey and red represent different impurities. The defects which are identified experimentally are marked by ‘*’. [66, 67] (adapted from [65]).

Figure 12 shows the schematic representation of intrinsic and extrinsic defects. For instance, the intrinsic defects are vacancy (missing atoms)[68], interstitial (same or different atom occupies interstitial site)[69], antisite (atoms of different type exchanges positions)[70], impurity(foreign atoms that replace some atom)71, etc. In contrast, the extrinsic defects are line defects (whole atoms in a row arranged anomalously)72, grain boundaries (separation line between polycrystalline grains)73, unsaturated surface bonds (boundary dangling atoms)74, etc. In figure 12, “int.” refers to intrinsic defects, and “ext.” corresponds to extrinsic defects. Some of the defects that can be directly observed experimentally are marked with ‘*’. However, there are various device engineering techniques by which these intrinsic and extrinsic defects can be minimized.

Defects passivation engineering of perovskites solar cells

Even though the first experimental demonstration of perovskite solar cells was done by Miyasaka at. el. in 2006 using MAPbBr3\mathrm{MAPbBr}_{3} perovskite, but the theoretical investigation using the first principle theory studied before that. In 2004, Change et. al. used local density approximation (LDA) theory to study the cubic phase of inorganic perovskite CsPbI3\mathrm{CsPbI}_{3} and hybrid organic–inorganic halide perovskite MAPbX3\mathrm{MAPbX}_{3} (where X= Cl, Br, I).75 They observed that the effective mass of holes was higher than the electron mass, which is not the case with conventional semiconductors.75 However, various perovskites are later being used to study the intrinsic defects in the perovskite crystal structure system using the first principle calculation.76–79 There are various kinds of point defects present in the perovskite crystal structure, and the main such point defects are (i) interstitial, (ii) vacancy, (iii) antisites or substitutional (iv) Frankel, (v) Schottky, (vi) impurity, etc. However, the MAPbI3\mathrm{MAPbI}_{3} is the widely studied perovskite, and it observed that there are mainly 12 dominating point defects present in the crystal structure e.g., three vacancies (VMAV_{\mathrm{MA}}, VPbV_{\mathrm{Pb}}, VIV_{\mathrm{I}}), three interstitials (MAi\mathrm{MA}_{i}, Pbi\mathrm{Pb}_{i}, Ii\mathrm{I}_{i}) and six antisites (MAPb\mathrm{MA}_{\mathrm{Pb}}, PbMA\mathrm{Pb}_{\mathrm{MA}}, MAI\mathrm{MA}_{\mathrm{I}}, PbI\mathrm{Pb}_{\mathrm{I}}, IMA\mathrm{I}_{\mathrm{MA}}, IPb\mathrm{I}_{\mathrm{Pb}}) defects, etc.80 These defects act as non-radiative recombination centres and are always detrimental to the device performance.

In 2018, Lu and Yam studied the FAPbI3_3-based perovskite crystal structure under different conditions (I-rich, Pb-rich, and moderate).81 It is observed that the VFA_{\mathrm{FA}}, FAI_{\mathrm{I}}, and IFA_{\mathrm{FA}} have much lower formation energies. Hence, antisites FAI_{\mathrm{I}} and IFA_{\mathrm{FA}} create deep levels in the band gap. These defects can act as non-radiative recombination centres and reduce the charge carrier lifetime, resulting in lower open circuit voltage. To avoid the formation of FAI_{\mathrm{I}} defects, the crystal should be grown in the I-rich environment. In addition, it is observed that the mixing of organic cations increases the formation energy of deep-level defects, which is expected to reduce the non-radiative recombination Figure 13b. Thus, the compositional engineering of organic cations is helpful to fabricate deep-level defect free perovskite and offer improved device performance.

The compositional engineering of hybrid organic-inorganic perovskite will be discussed in Chapter § 14. The mixing of the cations must follow the Goldschmidt tolerance factor (tt) and octahedral factor (μ\mu) rules as discussed in Chapter § 8.82,83 Apart from the organic cation mixing, in 2018, Ghosh et. al. showed that the addition of a small amount of Cs (≤25%\le25\%) in the FAPbI3_3 perovskite distorted the crystal structure (octahedral tilting) due to mismatch in the size of the cations.84 The slight octahedral tilting helps to form a strong hydrogen bond (N-H….I) between the FA cation and the halide ions.84 This eventually helps in the stabilization of crystal structure and promotes greater thermal stability with improved performance. In 2016, Saliba et. al. showed that the addition of 5% of CsI in the MAFAPb(IBr)3_3 perovskite improves the perovskite quality and results in an improvement in stability, reproducibility, and high efficiency.85 However, various reports show that the mix cation, i.e., compositional engineering, is a promising technique to improve the perovskite quality, which results in stable, reproducible, and efficient perovskite solar cells.85,86

There are various extrinsic defects present in the solution-processed perovskite film, such as grain boundaries, line defects, pinholes, unsaturated surface bonds, etc. These physical defects can be seen experimentally by scanning electron microscope (SEM) or tunneling electron microscope (TEM) images.[66, 67] However, passivation of such defects is essential to obtain improved performance of perovskite solar cells. Various additive engineering and surface treatments are adopted to passivate the extrinsic defects. Figure 14 shows the schematic of the grains and pinhole defects. For instance, in passivating defects, several large molecules in the form of Lewis acids and base are used to improve the crystal quality and improve charge carrier lifetime.87–89 Poly(methyl methacrylate) (PMMA) is such a Lewis base and is used at the interface between HTL-perovskite90 and ETL-Perovskite91. The PMMA used here by the spin coating technique suppresses the grain boundary and pinhole defects.88

Schematic of extrinsic defects and passivation of defects

Figure 14. Schematic of extrinsic defects grain boundaries, pinhole. (b) Passivation of defects using PMAA. (adapted from ref.88).

Further, additive engineering is a promising way to enhance device efficiency and lifetime. The additives applied to the perovskite bulk or at the interface interact with the perovskite absorber and the transport layers, resulting in defect passivation and ion immobilization, improving device performance and stability Figure 15a. The additives form strong ion bonds with uncoordinated Pb2+\mathrm{Pb}^{2+} or organic cations such as MA, FA, with the halide ions (Cl−\mathrm{Cl}^{-}, Br−\mathrm{Br}^{-}, I−\mathrm{I}^{-}) to enhance the Pb–I bonding and MA–I, FA-I bonding etc. Hence, the MA, FA, iodine vacancies, and other defects can be reduced. Son et. al. showed that adding a small amount of potassium iodide (KI) in halide perovskite is a universal approach to fabricating hysteresis-free PSCs. Wu et al. showed that the additive engineering of the ionic-liquid additive of methylammonium acetate (MAAc) modulates the thin film's crystalline quality and overall morphology.92

Figure 15. (a) Additive engineering for the passivation of defects to enhance efficiency and stability in perovskite solar cells. (adapted from ref.93) (b) Schematic illustration for understanding the stability improvement of the 3D/LD structure PSCs. (adapted from ref.94)

The MACl additive into the perovskite solution before deposition shows an increase in grain size from 200 to over 1000 nm, where the MACl has no effect on the absorption or photoluminescence (PL) spectra of the perovskite films.95 Li et al. reported a one-step solution-processing strategy using phosphonic acid ammonium additives, resulting in efficient PSCs and enhanced stability.96

Apart from the defects passivation, the stability of the perovskite solar cell is essential. The low dimensional (LD) capping layer plays a crucial role in surface passivation, heterojunction formation, and environmental stress resistance.97,98 Nazeeruddin et.al. reported that employing a thin 2D perovskite of phenethylammonium iodide (PEAI) layer on top of a 3D perovskite film shows improved efficiency and stability. The devices retained 85% of the initial efficiency even stressed under 1-Sun illumination for 800 hours at 50 °C.99 Liu et al. introduced a 2D capping layer of A2PbI4A_{2}PbI_{4} over the 3D perovskite layer and showed devices retained 90% of their initial efficiency for 1000 hours in humid air and 1-sun light.100 Wang et al. reported the incorporating n-butyl-ammonium cations of 2D perovskite into a 3D perovskite FA0.83Cs0.17Pb(IyBr1−y)3FA_{0.83}Cs_{0.17}Pb(I_{y}Br_{1-y})_{3} shows an efficiency of 19.5% which maintains the 80% of the initial efficiency of its post-burn after 1,000 hours in air, and close to 4,000 hours when encapsulated.101

Charge transport layers and interfacial engineering in perovskite solar cells

The journey of perovskite solar cells (PSCs) began with the concept of dye-sensitized solar cells (DSSCs) in the n−i−pn-i-p (aka “regular”) device architecture.[42] In 2009, Miyasaka et al. used hybrid organometallic metal halide perovskite MAPbI3MAPbI_{3} in the DSSCs and achieved an efficiency of 3.81%.[42] Currently, the reported efficiency of the perovskite solar cells is ~26% in both n−i−pn-i-p102 and p−i−np-i-n (aka “inverted”)103 device architecture, where the practical achievable limit104 could be ~30%. This thesis deals with the p−i−np-i-n architecture-based solar cells. The p−i−np-i-n device architecture has several advantages, including low-temperature processability of the charge transport layers (CTLs) and long-term operational stability derived from non-doped hole transport layers.105 In 2013, the p−i−np-i-n architecture perovskite solar cell started with PEDOT:PSS as the hole transport layer and PCBM or ICBA as the electron transport layer.[48]

However, in the p−i−np-i-n architecture, although C60 and PCBM are widely used as electron transport layers (ETL), numerous hole transport layers (HTLs) have evolved. The widely used hole transport layers are, inorganic (NiOx, Cu2O)106,107, polymeric (PEDOT:PSS, PTAA, PolyTPD)108–110, small molecules111, and self-assembled monolayer (Me-2PACz, MeO-2PACz, Me-4PACz,).112–114 The PEDOT:PSS is hygroscopic and acidic in nature, which impedes the efficient performance and stability of perovskite solar cells.115 The chemical stability of NiOx is challenging when connected with the perovskite, and hence, an additional interface passivation layer is used on the top of the NiOx layer.116,117 The PTAA-based HTL shows a offset of 200-300 meV energy band alignment mismatch for perovskite of bandgap 1.6 eV.113,118 Recently, various phosphonic acid group-anchored carbazole-based SAM HTLs have grabbed significant attention due to the high performance and excellent stability of perovskite solar cells. Figure 16a shows the molecular structure of 2PACz SAM deposited on ITO-coated glass substrates.119 The phosphonic acid group strongly binds with ITO and shows a strong monolayer fingerprint.113 In 2021, Levine et. al showed that, in the p-i-n architecture solar cells, the dominating loss mechanisms are associated with hole-selective buried interfaces. Compared to other SAMs, the Me-4PACz-based device showed suppressed non-radiative recombination at the perovskite/Me-4PACz interface, resulting in a faster hole transfer rate Figure 16b.112 In 2020, Al-Ashouri et al., used various phosphonic acid group SAMs such as 2PACz, MeO-2PACz,

Chemical structure of 2PACz SAM and different SAMs used at the ITO-perovskite interface

Figure 16. (a) The chemical structure of 2PACz SAM, and it is deposited on the ITO-coated glass substrate. (adapted from ref.119) (b) Different SAMs used at the ITO-perovskite interface and surface photovoltage transients with a minimalistic kinetic model applied to extract interfacial electron trap densities and hole transfer rates and their correlation with open-circuit voltages and fill factors. (adapted from ref.112)

and Me-4PACz as HTL, demonstrating superior device performance than the widely employed PTAA as HTL.113,120,121 In particular, by using Me-4PACz SAM as HTL, Al-Ashouri et al. showed that, a certified PCE of ~29% in silicon-perovskite two-terminal tandem solar cells with improved stability.121 Recently, Li et al., employed a mixture of MeO-2PACz and 2PACz to improve the charge extraction and reported a PCE of 25% for all-perovskite tandem solar cells.122 Deng et. al. showed that co-assembled monolayers help in suppressing the non-radiative recombination and the resultant device showed a PCE of 23.59%.123 Due to the exceptional performance of the SAM-based HTLs, they can promise to deliver commercialized p-i-n architecture-based PSCs.

Future perspective of perovskite solar cells

Perovskite solar cells (PSCs) are the fastest-growing PV technology and hold great promise for the photovoltaic industry due to their low-cost fabrication and excellent efficiency performance. To get it for the commercial readiness level (CRL), most important factor would be yield beyond 95% at PSC module levels. The current essential requirements for PSC are the reproducibility of high efficiency devices, scalability and stability. The reported certified high-efficiency (24-26%) papers are based on FAPbI3_3 perovskites with bandgap of Eg≈1.5E_g \approx1.5 eV and the typical device’s active area of ≈0.1\approx0.1 cm2^2 to maximum of 1 cm2^2. Scaling up the device's active area without compromising fill factor (FF), hence efficiency is non-trivial. Therefore, understanding the loss mechanism in large area devices is crucial. The stability analysis reported in the literature is inconsistent, preventing data comparison and identifying various degradation factors or failure mechanisms. Moreover, how the accelerated tests would be useful in predicting the real lifetime of the solar cells is yet to be developed. So, the knowledge and the technological gaps between laboratory and industry-scale production are crucial to develop. Therefore, in this section we have discussed the future perspective of perovskite solar cells in terms of the scalable and stable high efficiency PSCs.

Scalable perovskite solar cells and efficiency limitation

Solution-processed photovoltaic technologies grabbed much attention in the solar cell community and reported certified high-efficiency (>25%) using FAPbI3_3 (or ~5% Cs added) perovskite absorber of bandgap ~1.5 eV over the device active area of ≤0.1\leq0.1 cm2^2. [56] It is evident that, in terms of lab-scale device efficiency, this value is competitive with established commercialized technologies. However, the perovskite community is still far from commercializing the perovskite photovoltaic technology as the efficiency is not similar in the slightly large area (1 cm2^2) devices yet, regardless of the deposition techniques. In academic labs, people have demonstrated significant progress in PSCs over device active areas of up to 1 cm2^2 and mini-modules. However, a few industries e.g., Kunshan GCL, showed efficiency >19% on the single junction perovskite module of area 2 m2^2, but the recipe is not available in the public domain.124 Therefore, researchers must focus on developing strategies for high performance scalable perovskite solar cells and perovskite modules fabrication. Since this is a thin-film device technology with many layers being deposited via the solution route, it seems to be one of the major challenges in scaling up. Typical thicknesses of absorber layers in other established thin-film PV technologies like; GaAs and CdTe are typical of 10-12 μ\mum and 4-6 μ\mum respectively, however in the case of PSCs, it is hardly 600-800 nm. Further, both established thin-film PV technologies use typically vacuum-based deposition techniques (Vacuum-transfer deposition for CdTe and MBE for GaAs PVs), whereas halide perovskite PV performance is typically processed via solution route. Vacuum-processed perovskite is lagging behind in terms of efficiency. At least in the academic laboratory, scaling up via the solution route is challenging. Recently, we have demonstrated that the fabrication of large area PSCs by spin coating is challenging as the leakage current in the device increases exponentially, which may be due to non-uniform deposition of the perovskite layer in a multi-layered stack device.125 However, different deposition techniques have been developed, such as vacuum processing (thermal evaporation, pulsed laser deposition), solution processing (blade coating, spray coating, slot dye coating, inject printing, screen printing) apart from spin coating technique to make uniform perovskite layer over large area substrates. Even though the blade coating and evaporation based technique showed an efficiency ~25% over a small area (≤0.1\leq0.1 cm2^2) devices, however, the up-scaled perovskite solar modules (PSMs) of area ~100 cm2^2 showed better efficiency (~10%) in evaporated devices due to conformal deposition over large area substrates. However, the solution based techniques such as spray coating, slot die coating, and inject printing-based devices have lower cell efficiency in small area cells (0.1 cm2^2) but relatively improved efficiency performance in PSMs. Therefore, contrasting the photovoltaic performance with cell and module sizes of different deposition techniques makes it difficult to compare their performance disparities fairly and consistently. Chalkias et. al. develop a mathematical equation to ascertain the scaling-up factor (fscaling_upf_{scaling\_up})126,127 as

f=1−ηmoduleηcelllog⁡(AmoduleAcell)×100%f = \frac{1-\frac{\eta_{module}}{\eta_{cell}}}{\log\left(\frac{A_{module}}{A_{cell}}\right)} \times100\%

Where, η\eta is the PCE and A is the active area of the cell or module. The lower value of the ff indicates less scaling-up loss from PSCs to PSM, which is essential for commercialization. Interestingly, it is observed that f=3.48%f = 3.48\% for the inject printing process (Acell=0.105A_{cell} = 0.105 cm2^2, Amodule=804A_{module} = 804 cm2^2, ηcell=20.7%\eta_{cell} = 20.7\%, ηmodule=17.9%\eta_{module} = 17.9\%) and f=8.57%f = 8.57\% for evaporation based process (Acell=0.16A_{cell} = 0.16 cm2^2, Amodule=228A_{module} = 228 cm2^2, ηcell=24.8%\eta_{cell} = 24.8\%, ηmodule=18.1%\eta_{module} = 18.1\%), indicating the possible direction of scaling up the process in the future.128 People have reported in various review papers that scaling up of the PSCs active area results in poor efficiency, but which PV parameter affects the most is not well elaborated. Additive and interface engineering were adopted to improve the efficiency over large area PSC devices but the efficiency and reproducibility both decrease on increased device's active area, which restricts the PSCs toward commercialization.129 Even though the PSCs research started with the millimeter square device area, people have recently focused on taking it to the centimeter square device area. Deplorably, the literature study shows that there is a lack of representation of the performance of PSC in large areas. Many of the reported papers have demonstrated high efficiency PSC’s performance but don’t report the device area, we deliberately avoid to refer them here, however, would like to pass this is an important parameter to be checked at all levels, i.e., authors, editors, and reviewers. Even some groups report the device's active area in the SI file, which is a bad practice, and the reader needs to open another file to get the information on the device area. It would be good if the authors report the device's active area with dimension (length×width\mathrm{length} \times\mathrm{width}) in the main text along with the device image. Also, the sheet resistance of the transparent conductive oxide (TCO) substrate on which the device is fabricated could influence the device's performance, hence it is important to mention the sheet resistance (Ω/Sq\Omega/\mathrm{Sq}) of the substrates.130 It is also important to include the list of materials, device characterization tools, and detailed information on characterization techniques. Saliba et. al. studied 16000 papers and showed that the short circuit current density (JSCJ_{SC}) measured from J−VJ - V scans is 4-5% higher than the JSCJ_{SC} measured from integrating the EQE spectrum.131 This mismatch could be attributed to either edge effect132 from the active area of the device or pre-bias measurement condition 133. Some groups use a non-reflective metal mask aperture on the top of the active area during J−VJ - V scans, which could avoid edge effect.[56] Since the dimension of the aperture masks influences the PV parameters, such as VOCV_{OC} and FFFF, so it should be ≥80%\ge80\% of the device's active area.134

Photovoltaic parameters as a function of device active area, electrode designs, simulated series resistance distributions, and $J$–$V$ characteristics

Figure 17. Photovoltaic (PV) parameters as a function of device active area for the highest efficiency reported perovskite solar cells (PSCs).102,103,135–137 (b) 1-cm2^2 active area n-i-p architecture based PSC with the conventional electrode (left) and U-shaped new electrode design (right). (adapted from ref. 130). (c) Left images indicate the simulated series resistance distribution for 10 × 10 mm2^2 active area device (red dashed lines define active area) and right panel JJ − VV characteristics shows the performance comparison of the conventional and new electrode designed PSCs. (adapted from ref. 130).

Apart from that, recently, people reported excellent efficiency on the small area (<0.1 cm2^2) and relatively larger area (1 cm2^2) devices. However, they certify the efficiency on small area devices only or use an aperture mask of small area (<0.1 cm2^2), and that results in lack of confidence in large area PSCs.136 Figure. 17a represents the reported high efficiency PV parameters of PSC (Eg=1.52±0.3 eVE_g = 1.52 \pm0.3\ \mathrm{eV}) over device active area of ~0.1 cm20.1\ \mathrm{cm}^{2} and 1 cm21\ \mathrm{cm}^{2}. 102,103,135–137 It is evident that the JSCJ_{SC} and the VOCV_{OC} are not affected significantly by increasing the active area from 0.1 cm20.1\ \mathrm{cm}^{2} to 1 cm21\ \mathrm{cm}^{2}, whereas the FFFF does, hence the device efficiency. This effect could be severe in further large area devices. It is of utmost importance to understand the loss in the FFFF due to increasing the device's active area to look forward the scalability of PSCs. Ho-Baillie and co-workers showed that the TCO substrate sheet resistance could influence the FFFF of large-area PSC devices apart from the non-uniform deposition of the device stack.130 They fabricated n-i-p mesoscopic PSC device architecture on a conventional and U-shaped designed bottom gold electrode (150 nm thick) and showed certified efficiency of 19.63% with FFFF of 75.8% over an active area of 1.02 cm21.02\ \mathrm{cm}^{2} (see Figure. 17b). Depositing U-shaped electrodes on the TCO substrates enhances the electron collection efficiency in n-i-p device architecture. Figure. 17c shows the simulated series resistance distribution for 10×10 mm210 \times10\ \mathrm{mm}^{2} active area device, and it is lower for the U-shaped modified electrode, and the corresponding fill factor (FFFF) increased from 50.0% to 72.2%. The FF of the device also depends on the device geometry, as shown in Figure. 18a. The series resistance is higher for the square-shaped electrode because the electrons travel a longer distance from the center, whereas it is lower for the strip-shaped electrode and shows an improved fill factor. However, the bottom electrode modification strategy is possible for the n-i-p device where the ETL (c-TiO2\mathrm{TiO}_{2} 50 nm/mp-TiO2\mathrm{TiO}_{2} 200 nm) thickness ~250 nm could be deposited by spray pyrolysis/spin coating technique over 100-150 nm thick bottom U-shaped electrode. But similar strategy is difficult to adopt in solution-processed p-i-n PSCs, as the HTL layer thickness is usually 2-3 nm of recently developed self-assembled monolayers (SAMs).112,114,138 Such a 100-150 nm bottom U-shaped gold electrode would act as a barrier at the edges of the substrates and make it difficult to deposit uniform HTL using the solution-processed spin coating technique. Therefore, for large area (≥1 cm2\geq1\ \mathrm{cm}^{2}) p-i-n device architecture, the U-shape electrode deposition is done (thermal evaporation or silver paste) after scratching the spin-coated thin film layers. Figure. 18b, c, d are the photographic images of large area (1 cm21\ \mathrm{cm}^{2}) PSCs using different electrode shapes.

Simulated series resistance distributions, simulated and measured current–voltage curves, and rectangular device active areas

Figure 18. (a) Left panel represents the simulated series resistance distributions for 10×1010 \times10 mm2^{2}, 6×176 \times17 mm2^{2}, and 4×254 \times25 mm2^{2} devices using electrode designs with different geometries. The right panel shows the simulated (solid line) and measured (dashed line) J−VJ - V curves of devices using new electrode designs with different geometries (adapted from ref.130). (b) Rectangular device active area of 0.9480.948 cm2^{2} with one side TCO connection.139 (c) Rectangular device active area 1.0081.008 cm2^{2} with all side TCO connection.140 and (d) Rectangular device active area 1.0001.000 cm2^{2} with U-shaped (three side TCO connection) bottom electrode deposited after scribing the spin-coated layers.141

Unfortunately, most of the reports do not elaborate on the electrode design for large-area devices, which is a bad practice and lacks progress toward the scalability of PSCs. We believe reporting the device engineering, including electrode design, is required for successful scalable PSC fabrication apart from uniform perovskite layer deposition. In addition, a systematic simulation or experimental study needs to be carried out to understand the series and shunt resistance losses for large-area devices to better understand the FF loss issues.142

Stability analysis of perovskite solar cells

To ensure the economic viability of perovskite technology, long-term stability is crucial apart from high efficiency and scalability. To compete with the market-leading commercialized PV technologies, the expected lifetime of the PV devices should be 20-25 years. It is impractical to use decade-long protocols for stability testing. Hence, accelerated ageing tests are required to understand the lifetime span of PSCs. The stability of the PSCs depends on extrinsic (oxygen, moisture, UV light exposure) and intrinsic (thermal and structural degradation, ion migration, electrical bias) stresses. Extrinsic stress, such as oxygen-induced degradation, moisture degradation, and UV-light degradation, can be avoided by advanced encapsulation techniques. However, the intrinsic stability of the PSCs is a crucial factor, and understanding them and corresponding solutions need to be discussed for commercializing perovskite PV technology. Considering the material stability, the mixed organic-inorganic halide perovskites represented the better choice for high efficiency and stable performance.85,143 However, the existing accelerated aging tests described in an International Electrotechnical Commission (IEC) standards are used for testing the Si solar panel performance.144 Unfortunately, these tests cannot be carried out in DSSC, OPV and PSCs as they are fundamentally different materials and PSCs also show ion migration, performance recovery in dark conditions, etc., which are unknown in Si photovoltaics. In 2011, a board consortium of researchers at the International Summit of Organic PV Stability developed standardized ageing tests called ISOS protocols for testing the lab-scale OPV devices to ensure the comparability of PV testing performed at different laboratories.145 Those protocols include the following stability testing standards (i) dark storage (ISOS-D), (ii) light soaking (ISOS-L), (iii) outdoor stability (ISOS-O), (iv) thermal cycling (ISOS-T), (v) solar-thermal cycling (ISOS-LT), and each of which has three different levels determined by thermal stress, relative humidity stress, environment setup, and light source & circuit bias (maximum power point tracking: MPPT or open circuit: OC). People reported different stability conditions and data formats in PSCs, preventing data comparison and identifying various degradation factors or failure mechanisms. The consensus statement in 2020 by Khenkin et. al. proposes that the ISOS stability could be carried out in PSCs along with additional ageing experiments such as (i) bias stability (ISOS-V) (ii) light cycling (ISOS-LC), and also reported a checklist that should be carried out for reporting the stability of PSCs. The reason for adding bias stability and light cycling for PSC is discussed below.

The ISOS-O-1 protocol suggests the periodic measurements of JJ–VV curves under solar simulator light illumination in MPPT or OC bias conditions, whereas ISOS-O-2 requires natural sunlight. ISOS-O-3 requires both in situ MPP tracking under natural sunlight and periodic performance measurements under a solar simulator. Figure 19a shows the stability performance (ISOS-O) obtained by JJ–VV measurements and MPP tracking of PSCs do not coincide, although they generally have similar trends. Therefore, it is crucial when characterizing PSCs to describe the load and recovery time before each JJ–VV measurement. The perovskite solar cell degradation mode is known to be reversible partly or entirely in dark conditions, which is often referred to as metastability (see Figure. 19b). Therefore, cycling through light-dark periods to mimic day-night cycles results in a significantly different stress test than applying constant illumination (ISOS-L). The improvement in the power conversion efficiency (PCE) under illumination after storage in the dark condition could be attributed to the passivation of interfacial defects by photogenerated charge carriers or ion migration induced modification of the built-in electric field.146 However, the PCE dynamics during a cycle depend on the present status of cell degradation condition. Therefore, the ISOS protocols were revised for PSCs and included light-dark cycling protocols to account for the recovery phenomena (ISOS-LC). In addition, the electrical bias causes PSC degradation by stimulating the ion migration or charge carrier accumulation, resulting in thermally activated trap formation.147 Therefore, ISOS protocols revisited for PSCs to include the ISOS-V, in which the behaviour of the cell is analysed when exposed to a certain electrical forward bias (maximum power point voltage: VMPPV_{MPP} or open circuit voltage: VOCV_{OC} as measured in AM 1.5G) in the dark (see Figure. 19c). Interestingly, people also show additional degradation studies beyond ISOS protocols, such as exposure of PSCs using 10 Suns illumination at the MPPT condition of the encapsulated devices (See Figure. 19d) and shows excellent thermal stability of FA0.83Cs0.17PbI2.7Br0.3\mathrm{FA}_{0.83}\mathrm{Cs}_{0.17}\mathrm{PbI}_{2.7}\mathrm{Br}_{0.3} composition based PSC.143 When the 3D perovskites come in contact with the environment (moisture, oxygen, UV-light), they are subjected to severe degradation and, limiting the solar cell's lifetime. Different 2D and quasi-2D perovskite materials demonstrated significant improvement in the efficiency and stability of 3D PSCs. The 2D perovskites exhibit multiple-quantum-well structures and offer tunable optoelectronic properties, defects passivation in the bulk and interface, superior thermal and light stability, hydrophobicity due to large organic cation, suppressing ion migration etc.148 To achieve simultaneously high efficiency and high stability for 3D PSCs using 2D perovskites, numerous device engineering strategies studies have been conducted, including compositional or additive engineering, dimensional engineering, interfacial engineering, etc. Sum et. al. used different 2D (and low dimensional) perovskite materials to passivate 3D perovskite surface defects and showed an efficiency 24.1% with T95T_{95} >1000 hours at MPPT.94 Wolf el. al. tailored the number of octahedral inorganic sheets in 2D perovskite passivation layers for 3D PSCs, and showed efficiency 24.3% with T95T_{95} >1000 hours under damp heat testings (85º C and RH 85%).152 Mohite et. al. carried out dimensional tunability and grew pure 2D phase perovskites of controlled thickness on the 3D perovskite layer using solvent engineering strategy and showed efficiency 24.5% with T99T_{99} >2000 hours under continuous 1-sun light at 55º C and RH 65% (see Figure. 19e).150 However, for the practical application of perovskite solar cells, the encapsulation of the PSC device is mandatory to prevent the penetration of oxygen, moisture, and lead leakage from the lead-based PSCs.

Power conversion efficiency and stability measurements of perovskite solar cells

Figure 19. (a) Power conversion efficiency (PCE) extracted from current density vs. voltage (J−VJ-V) scans (black dots) and continuous MPP tracking (red curve) for the same PSC. (adapted from ref. 149). (b) PCE variation of PSCs upon continuous (blue curve) or cycled (6/6 h, red curves) illumination by white light emitting diodes. (adapted from ref. 149). (c) Normalized PCE of PSCs changes with forward bias in the dark condition. (adapted from ref. 149). (d) Stabilized power output (SPO) of the perovskite devices, continuously recorded under 10 Suns concentration illumination. (adapted from ref. 143) (e) ISOS-L-1 stability measured at MPPT in ambient condition under continuous 1-sun illumination (55°C) for epoxy encapsulated control-3D and 2D passivated 3D PSC. (adapted from ref. 150) (f) Indoor cycled and outdoor stability of encapsulated PSCs with triple cation perovskite and different transport layers. The outdoor power output of the representative cells is depicted at the corresponding irradiance and temperature. (adapted from ref. 151).

Emery et. al. recently performed the outdoor stability (ISOS-O3) of FA0.85Cs0.15PbI2.55Br0.45\mathrm{FA}_{0.85}\mathrm{Cs}_{0.15}\mathrm{PbI}_{2.55}\mathrm{Br}_{0.45} composition based PSC after encapsulating using glue-based epoxy and Polyolefin Elastomer (POE) + butyl lamination in a glass–glass stack. 153 They observed that the glue-based encapsulation lasts for only 3 months whereas the POE-based encapsulation lasts for ten months. However, the POE-based encapsulation is technically challenging to implement, but it provides excellent stability even under damp heat and outdoor test conditions. The POE + butyl lamination in a glass-glass stack of PSC restricts its use as it requires an annealing temperature of 150 °C for 20 minutes for the lamination process, which could damage the methylammonium iodide (MAI) containing perovskites. Therefore, developing encapsulation technology for perovskite solar cells is of utmost importance for commercialization of it. Nevertheless, it is crucial to understand that the indoor stability test results as per ISOS protocols do not match the outdoor stability results due to the transient behaviour of PSCs not being accounted for in the standard ISOS constant illumination testing conditions. Abate et. al. showed that light cycling shows much better agreement with outdoor observations, at least qualitative (and even quantitative in some cases) as shown in Figure. 19f. Apart from all these different accelerated ageing tests, the important question is how can we predict the practical lifetime of the PSCs from these ageing tests? As far as we know, no methods have been developed yet to estimate the practical lifetime from these accelerated tests for PSCs. So, to speed up the commercialization of PSCs, it is of utmost importance to develop a simulation model to predict the real field lifetime of the PSCs using the accelerated ageing test results. However, Solar Energy Technologies Office (SETO) in DOE, USA and Perovskite PV Accelerator for Commercializing Technology (PACT), have made a joint effort to perform the field deployment testing. PACT has planned to develop a standardized testing protocols to establish outdoor testing to ensure their accelerated ageing tests accurately reproduce the real field degradation mechanisms and allow researchers to corroborate their prototypes in the real field. 154

References

(2022).

Methodology

Methodology

This chapter will introduce all the experimental techniques employed to fabricate thin films and solar cell devices, as well as the characterization method used to study both the perovskite thin films and solar cell devices. To understand the photophysics and device physics of perovskite semiconductors, we have used different experimental techniques such as absorption spectra, steady-state and time-resolved photoluminescence spectra, photoluminescence quantum yield, field emission scanning electron microscopy, dektak profilometer, X-ray diffraction, atomic force microscopy, kelvin probe force microscopy, contact angle measurement, nuclear magnetic resonance, dark and illuminated current density vs. voltage (J−VJ - V) characteristics, intensity-dependent (J−VJ - V), Suns–VOCV_{\mathrm{OC}}, incident photon to current efficiency (IPCE), current density - voltage - light (J−V−LJ - V - L) measurement, steady-state electroluminescence, transient photovoltage/photocurrent measurement, and frequency-dependent photocurrent measurement.

Thin films and device fabrication

Substrates preparation

The indium tin oxide glass substrates (10-15 Ω/sq\Omega/\mathrm{sq}) of dimension 60 mm ×\times 60 mm are purchased from Luminescence Technology Corp. The dimensions of the patterned ITO substrates used for device fabrication are 15 mm ×\times 15 mm. The following steps are involved in preparing the substrates (see Figure 1)

  1. Cut the substrates in 60 mm ×\times 15 mm (i.e. in four strips) and then place a 12 mm width thermal tape on one side of the ITO substrate.

  2. In the open area where the ITO etching will be done, deposit zinc power paste (in water) and dry it for 5 minutes.

  3. Slowly drop 37% HCl (purchased from Sigma Aldrich) solvent on the zinc-coated area to etch the ITO.

  4. Dip the substrates in DI water to remove HCl completely.

  5. Remove the thermal tape, wipe it with isopropanol, and dry it in the air for 5 minutes.

  • (vi) Cut the patterned 60 mm × 15 mm substrates into four pieces i.e. 15 mm × 15 mm.

  • (vii) For a 1 cm2^{2} device, the substrate’s size is 20 mm × 20 mm. In this case, the 60 mm × 60 mm substrates are cut in 60 mm × 20 mm i.e. in three pieces, and use a thermal tap of 15 mm width. Repeat the above ITO etching process with zinc powder.

  • (viii) The substrates were sequentially cleaned with soap solution, deionized (DI) water, acetone, and isopropanol for 10 minutes each in an ultra-sonicator.

  • (ix) The cleaned substrates are dried with a nitrogen gun and kept on the hot plate at 100∘C100^{\circ}\mathrm{C} for 10 minutes to remove the residual solvents.

  • (x) Take the substrates inside the oxygen plasma ashing chamber for 15 minutes. The plasma ashing is done at an RF power of 18 watts.

  • (xi) After plasma ashing, take the substrates immediately inside the N2_{2} environment glove box (O2_{2}<0.1 ppm, H2_{2}O<0.1 ppm) for the spin coating process.

Figure 1 (PDF p. 106). Steps involved in pattering the ITO substrates. The 60 mm × 60 mm ITO-coated glass substrate was cut in 60 mm × 15 mm i.e. 4 pieces (step -1). Then, place a 12 mm width thermal tap on one side (step-2). After that, put zinc paste on the open area (step-3). Drop HCl to remove the ITO (step-4). Wipe the substrate with IPA (step -5). Cut the substrate into four pieces i.e. 15 mm × 15 mm (step -5). Finally, clean the substrates to fabricate solar cell devices.

Perovskite solar cells fabrication

In this thesis, we have used the pp-ii-nn (inverted) architecture-based perovskite solar cell (PCS) devices. The schematic diagram of the device architecture is shown in Figure 2. The typical device structure used in this thesis is indium tin oxide (ITO)/HTL /perovskite (CsFAMA) /phenyl-C61_{61}-butyric acid methyl ester (PC61_{61}BM)/ bathocuproine(BCP)/ silver (Ag). We have used two hole transport layers (HTL): PTAA or Me-4PACz. In the device structure (Figure 2), the ITO acts as a transparent electrode, PTAA or Me-4PACz acts as HTL, perovskite acts as an absorber layer, PCBM and BCP act as an electron transport layer (ETL) and buffer layer respectively, and finally silver acts as a back metal electrode. The hole transporting layer PTAA or Me-4PACz is spin-coated at 4000 rpm for 30 seconds over the plasmas-ashed ITO substrates and annealed at 100∘100^\circ C for 10 minutes. The perovskite layer was deposited on the hole-transporting layer at 5000 rpm for 30 seconds, and in the last 7 seconds, anti-solvent chlorobenzene dropped for faster crystallization. The perovskite films were immediately annealed at 100∘100^\circ C for 30 minutes. After cooling down the perovskite layer for 5 minutes at room temperature, the PC61_{61}BM layer spin coated at 2000 rpm for 30 seconds and air dried for 5 minutes. After that, the BCP layer spin coated at 5000 rpm for 20 seconds. Finally, the substrates are taken in an N2N_2-filled thermal evaporator for opaque back electrode deposition. 150 nm silver (Ag) was deposited using a metal shadow mask under a vacuum of 2×10−62 \times10^{-6} mbar pressure at a rate of 0.1 A˚\text{\AA}/second for the fast 10 nm and 1 A˚\text{\AA}/second for the rest 140 nm. The specific details of the solution preparation and spin coating are discussed in each chapter.

Schematic diagram of the p-i-n (inverted) architecture-based perovskite solar cells

Figure 2 (PDF p. 107). The schematic diagram of the p-i-n (inverted) architecture-based perovskite solar cells.

Device contact with metal pins

After depositing the back metal electrode, we fix the customized metal pins to the device using a contact pin fitter zig. The metal pins are made of alloy Copper Tin (Phosphor Bronze) and purchased from Batten & Allen Ltd. The different active area-based solar cells are made using different metal masks but in all the devices we used the same kind of metal pin as shown in Figure 3.

Digital image of perovskite solar cells with metal pins and different active areas: Area 0.175 cm², Area = 0.350 cm², Area 0.805 cm², and Area = 1.00 cm²

Figure 3 (PDF p. 108). Digital image of the perovskite solar cells having different active areas. The devices are connected with metal pins to contact the ITO and Ag electrode.

Thin film characterizations

Steady-state absorption spectroscopy

The steady-state absorption spectra of all the solution-processed thin films have been taken using the Perkin Elmer Lambda 950 spectrometer with the integrated sphere. The Perkin Elmer Lambda 950 spectrometer consists of two light sources (i) deuterium and (ii) tungsten halogen lamp. These two lamps provide a full range of UV-vis-NIR spectrum (175 nm – 3300 nm). The samples are illuminated from the front side for the absorption/transmission spectra and placed at position P1. In this case, the reflection port at place B is closed by the reflectance standard plate (coated with Barium Sulfate). For the reflection measurement, the samples are placed at position P2. The samples are illuminated from the glass side for absorption/transmission measurement of the thin films deposited on the glass substrates. For accurate optical density (O.D.) measurement of the thin films deposited on the glass substrates, first, the calibration is done using a glass substrate on which the films are deposited. The absorption measurement is carried out using glass as a baseline correction sample. The photodetectors used in the Lambda 950 spectrometer are (D1) PbS & (D2) InGaAs and are placed at the bottom of the integrating sphere as shown in Figure 4.

Figure 4 (PDF p. 109). Schematic diagram of the absorption spectroscopy measurement using Perkin Elmer Lambda 950 spectrometer.

The absorbance of the thin film can be expressed by Beer Lambert’s law as,

A(λ)=−log⁡(II0)(15)A(\lambda) = -\log\left(\frac{I}{I_0}\right) \tag*{(15)}

Where I0I_0 and II indicate the intensities of the incident and the transmitted light respectively. The absorption coefficient can be expressed as

α(λ)=−2.303×A(λ)t(16)\alpha(\lambda) = -\frac{2.303 \times A(\lambda)}{t} \tag*{(16)}

Where tt is the thickness of the perovskite absorber layer.

The bandgap of a semiconductor as a function of the absorption coefficient is defined by the Tauc method as[1]

(αhν)1/γ=B(hν−Eg)(17)(\alpha h\nu)^{1/\gamma} = B(h\nu- E_g) \tag*{(17)}

Where hh is the Planck constant, ν\nu is the frequency of the photon, EgE_g is the energy bandgap of the semiconductor and B is a constant. The factor γ\gamma depends on the nature of electron transfer and is equal to 2 for indirect and 1/21/2 for direct transition bandgaps respectively. The bandgap of a semiconductor is calculated using the above relation.[2, 3]

The Urbach energy (EuE_u) is the measure of disorder in the bulk of the perovskite semiconductor and can be estimated from the absorption spectra using the following equation,[4, 5]

α(E)=α0exp⁡(E−EgEu)(18)\alpha(E) = \alpha_0 \exp\left(\frac{E-E_g}{E_u}\right) \tag*{(18)}

Where EE denotes the photon energy, and α0\alpha_0 is the material-dependent absorption coefficient.

Steady state and time resolved photoluminescence spectroscopy

Figure 5 (PDF p. 110). Schematic diagram of the steady state and time resolved photoluminescence spectroscopy measurement using a pulsed laser and gated i-CCD (Andor iStar) coupled with spectrometer.

When photons incident on a semiconductor of energy equal to or more than the bandgap, they will be absorbed and create electron and hole pairs at the conduction and valence bands respectively.[6] The absorption of the photons by the semiconductor takes place on the time scale of femtosecond (fs).[7] After the absorption process, the excited electrons undergo various fast transitions within the conduction band and relax to the minimum of the conduction band, called thermal relaxation. Further, the electrons spontaneously come down from the conduction band minimum to the valence band maxima and recombine with the holes, resulting in the re-emission of photons.[6, 8] This process is called photoluminescence (PL), and the emitted photons have energy equal to the bandgap of the semiconductor material. The time scale of the photoluminescence process happens in the perovskite material in the order of a few ns to μ\mus.[9, 10] The steady state photoluminescence (SSPL) measurement of the thin films performed in the custom made chamber under a vacuum of 10−510^{-5}mbar. The steady state PL spectra give information about the PL peak position and peak intensity. To get information about the charge carrier lifetime, time resolved (TR) PL measurement is useful.[11] For the TRPL measurement, we used a highly sensitive gated intensified charge coupled device (i-CCD, Andor iStar). The perovskite film is kept inside the sample holder at a pressure of 10−510^{-5}mbar using a vacuum pump station as shown in Figure 5. To excite the film, we used a 3rd harmonies at 355 nm of Nd:YAD solid state laser (from Innolas) with fundamental wavelength 1064 nm, operating at 1 kHz with pulse width 840 ps. The pulse width of the laser must be lower than the expected lifetime of the charge carriers. The number of the generated electron-hole pairs depends on the power used to excite the sample. The power calibration was done using a Si-based power meter. The maximum energy limit of the Nd:YAG laser is 22 μ\muJ and it can be varied at the sample by using a varying neutral density filter wheel having an optical density of 0 to 4. The excited electrons from the conduction band decay and recombine with the holes via a radiative or non-radiate recombination process. The decay process of the charge carriers via radiative recombination can be characterized by the lifetime of the charge carriers. The lifetime of the charge carrier is defined as the time taken by the PL intensity decay by 1/e1/e times of initial intensity. To get the PL decay time, the gate of the i-CCD triggered with the external circuit of the laser so that excitation and decay time difference can be accurately estimated. There is a time delay of ~75 ns between the gate opening and the optical signal reaching the gate. Therefore, we put a delay of ~75 ns to consider it as the zero time. The PL emission dispersed through a grating spectrometer with 500 nm blaze and 600 lines/mm ruling (placed inside of the Shamrock 303i spectrometer). The spectrometer is integrated with the gated i-CCD (Andor iStar). The minimum optical gating time of the i-CCD is ~780 ps with a minimum resolution of 10 ps. To increase the signal to noise (SNR) ratio, 70 accumulations per acquisition have been taken.

Generally, for the perovskite thin films, the PL decay contribution comes from fast and slow decay.[11, 12, 13] So, we fit the PL decay with a bi-exponential decay function to estimate the average lifetime of the charge carriers. The average lifetime (τavg\tau_{avg}) is calculated using the following equation.

τavg=(A1t12+A2t22A1t1+A2t2)(19)\tau_{avg} = \left(\frac{A_1t_1^2 + A_2t_2^2}{A_1t_1 + A_2t_2}\right) \tag*{(19)}

Where A1,A2A_1, A_2 are the weightage of the fast and slow decay time t1t_1 and t2t_2 respectively.

Photoluminescence quantum yield

Figure 6 (PDF p. 112). Schematic of the absolute photoluminescence quantum yield (PLQY) measurement using the integrating sphere in the following steps, (a) No sample in the integrating sphere, i.e. measuring the laser intensity. (b) Thin film placed off the axis of the laser beam to measure the self-absorption of the scattered light. (c) Thin film placed on the axis of the laser beam to measure the total PL.

In the steady state photoluminescence (SSPL) measurement, we just know the PL peak position and intensity of the PL spectra (depends on the calibration) but absolute PL intensity can’t be measured because of not considering all emitted photons. To do so, we need an integrating sphere where the inner surface is highly reflective (~100%, Barium Sulfate coated) and results in isotropic redistribution of the light.2 The photoluminescence quantum efficiency indicates the radiative efficiency of any sample and is defined as3

PLQY=number of emitted photonsnumber of absorbed photonsPLQY = \frac{\textit{number of emitted photons}}{\textit{number of absorbed photons}}

Higher PLQY is required for better performance of solar cells. The higher PLQY indicates the suppression of non-radiative channels. The PLQY can be increased by improving the material quality.4 The absolute PL measurement can be done in three steps as shown in Figure 6. In the first step, there is no sample in the integrating sphere and only the laser intensity is measured Figure 6a. In the second step, a thin film is placed off the axis of the laser beam such that the laser beam directly doesn’t hit the sample. This process accounts for the emission due to the reabsorption of the scattered light by the thin film Figure 6b. In the final step, the sample is placed on the axis of the laser beam. This process accounts for the emission due to the direct absorption of the laser beam along with the reabsorption of the scattered light Figure 6c.

PLQY Calculation:

Schematic spectral profile showing laser spectra L_a, L_b, and L_c and emission spectra E_b and E_c

Figure 7 (PDF p. 113). Schematic of the PLQY measurement in three steps. The term ‘L’ represents laser spectra and ‘E’ for emission spectra.

Figure 7 represents the schematic spectral profile of the excitation laser beam and the emission spectra inside the integrating sphere, as discussed above. LaL_a, LbL_b, and LcL_c are the areas under the curves of the laser beam spectra, and EbE_b and EcE_c represent the areas under the curves of the emission spectra in the three steps, as depicted in Figure 6. Hence, we can write,

Lb=La×(1−μ)(20)L_b = L_a \times(1-\mu) \tag*{(20)}
Lc={La×(1−A)}×(1−μ)(21)L_c = \{L_a \times(1-A)\} \times(1-\mu) \tag*{(21)}

Where AA and μ\mu represent the fraction of the direction absorption of the laser beam and reabsorption of the scattered light.

Simplifying the above equations, the absorption coefficient of the thin film can be expressed as,

A=(1−LcLb)(22)A = \left(1-\frac{L_c}{L_b}\right) \tag*{(22)}

The component of the scattered light which contributes to the total emission is

(1−A)×(Eb+Lb)(1-A) \times(E_b+L_b)

The total emission output in the 3rd step is

Lc+EcL_c+E_c

The emission due to the laser beam can be written as

ηPL×A×La\eta_{PL} \times A \times L_a

Where ηPL\eta_{PL} represents the PL quantum yield of the thin film.

Thus, considering no loss of light inside the integrating sphere, we can write

Lc+Ec=(1−A)×(Lb+Eb)+ηPL×A×La(23)L_c+E_c = (1-A) \times(L_b+E_b)+\eta_{PL} \times A \times L_a \tag*{(23)}

Simplifying the above equation, the PLQY is given by

ηPL=Ec−Eb(1−A)La×A(24)\eta_{PL} = \frac{E_c-E_b(1-A)}{L_a \times A} \tag*{(24)}

Field emission scanning electron microscopy

Field emission scanning electron microscopy (FESEM) is a highly advanced technology used to study the topography of nano/micro-structured samples.[17] The FESEM can be used to study the various topographic images such as surface morphology (crystalline size, grain boundaries), the cross sectional view (thickness of the individual layer of a multi-layered thin film device). Using the energy dispersive X-ray spectroscopy (EDS) facility in the FESEM system we can study the composition of the material i.e. each constituting atom in percentage. We have used Carl Zeiss Ultra 55 FESEM instrument to study the thickness, cross section, and compositional analysis of atoms/molecules of the perovskite films. The FESEM imaging is superior over the optical microscope because of the small wavelength of the electron in the range of 0.859 Å to 0.037 Å. It facilitates the high resolution images of magnification up to 300,000×300{,}000 \times.[18] The high energy electron beam (1-30 keV) generated from the electron gun in ultrahigh vacuum (10−910^{-9} mbar) and used as a source. The primary electrons are accelerated using the anode electrode and then directed using the electromagnetic lens (condenser lens, magnetic lens, and scanning coils) Figure 8. The scanning coil deflects the electron beam over the sample according to the zig-zag pattern. Due to the high-energy electrons hitting the sample’s surface, the secondary electrons are emitted from the sample and collected by the secondary electron detector. These secondary electrons create the electrical signal, which is further converted to the image/videos in the monitor. For good quality images, the samples should be conductive to reduce the charging effect.[17] Therefore, we prepared the perovskite films on the ITO-coated glass substrates and before measurement, the films were grounded on the sample holder using silver paste.

Figure 8 (PDF p. 115). Schematic diagram of field emission scanning electron microscopy

Dektak profilometer

The thickness of the solution processed perovskite thin film was measured using a Bruker dektak profilometer (DektakXT). The system takes measurements over the surface electrochemically by moving a diamond-tipped stylus. In addition to the thickness measurement, the system measures the surface roughness in the nanometre range. The scan range of the system is 6.5 μm\mu\mathrm{m} to 524 μm\mu\mathrm{m} with resolution 0.1 nm to 8 nm respectively. To measure the thickness, the applied stylus force can be 1 mg to 15 mg but we used 3 mg only so that the film does not get scratched during measurement which can result in inaccurate thickness. To measure the thickness of the perovskite film deposited on the glass substrates, we scratched the film at different places using a sharp needle. The difference between the top surface and scratched bottom surface (i.e. glass) gives the value of the thickness Figure 9.

Thickness measurement of the perovskite thin film using Dektak profilometer

Figure 9 (PDF p. 116). Thickness measurement of the perovskite thin film using Dektak profilometer. The thickness of the perovskite thin film ∼420 nm\sim420\ \mathrm{nm} marked with a red closed loop.

X-ray diffraction

X-ray diffraction is a non-destructive phenomenon where the high energy X-ray interacts with a crystalline material elastically and provides information about the crystallographic structure, chemical composition, physical properties, etc. of the material. A crystalline material is a regular periodic arrangement of atoms, ions, or molecules in a solid material. The X-ray light diffracted from the periodic lattice can interfere constructively or destructively. However, in the X-ray diffraction system, the detector detects only constructive interference. The schematic of the X-ray diffraction analysis is shown in Figure 10.

The constructive interference in the X-ray diffraction can be explained by Bragg’s law[19]

2dsin⁡θ=nλ(25)2d\sin\theta= n\lambda \tag*{(25)}

Where dd is the spacing between the diffracting plans, θ\theta is the angle of incidence, nn is the integer number, and λ\lambda is the wavelength of the X-ray.

Figure 10 (PDF p. 117). Schematic representation of the X-ray diffraction by the atoms of a periodic crystal lattice.

The perovskite films used for making solar cell devices are polycrystalline. The polycrystalline perovskite film’s grain size was observed from FESEM imaging with an average grain size of ∼200−300\sim200 - 300 nm. We used a Smartlab Rigaku high-resolution X-ray diffractometer (HR-XRD) to measure the X-ray diffractograms for the perovskite thin films. The wavelength of the diffractometer is Cu Kα\alpha radiation λ=1.54\lambda= 1.54 Å.

The XRD pattern used to calculate the crystalline size (D) of the spherical domains of the perovskite films using Debye Scherrer formula,

D=kλβcos⁡θ(26)D = \frac{k\lambda}{\beta\cos\theta} \tag*{(26)}

Where kk is the shape factor (~0.9)[20], λ\lambda is the wavelength of the X-ray, β\beta is the full width at half maximum (FWHM), and θ\theta is the angle of incidence.

Atomic force microscopy

An atomic force microscopy (AFM) instrument images the surface topography of a sample surface by scanning it using a highly sensitive cantilever tip.[21] The AFM provides information regarding the particle size, morphology, surface roughness, grain size, etc.[22] in the nanometre scale. The sharp tip scanning probe is attached to a flexible cantilever and moves along the surface for a very short distance, 0.2 nm to 10 nm. The AFM measures the intermolecular force between the sharp tip and the atom/molecule on the surface of the sample and results in deflection. The laser/photodiode system measures this deflection, capturing the sample’s topography. We used the Bruker Multimode Nanoscope IV instrument to capture the surface image in tapping mode. The cantilever tip has a tip radius of 8-10 nm (RTESP-300) and can measure an average roughness of ~1 nm. The minimum scanning size is 0.4 μm\mu\mathrm{m} × 0.4 μm\mu\mathrm{m} and the maximum size is 125 μm\mu\mathrm{m} × 125 μm\mu\mathrm{m}. Higher scanning size has lower resolution hence we used in general 5 μm\mu\mathrm{m} × 5 μm\mu\mathrm{m} or is 10 μm\mu\mathrm{m} × 10 μm\mu\mathrm{m} at the different locations of the surface.

Figure 11 (PDF p. 118). Schematic diagram of the atomic force microscopy.

Kelvin probe force microscopy

The kelvin probe force microscopy (KPFM) is an electrostatic force microscopy technique that is extensively used to study the surface potential of various conductive or semi-conductive samples.[23] The KPFM allows us to measure the quantitative local surface potential distribution and estimate the work function. The KPFM working principle is based on the Fermi-level alignment of the tip and the sample when they are in contact during measurement. In this case, the electrons flow from the tip to the sample or vice versa. This Fermi level alignment results in an offset in the vacuum level which in turn introduces the contact potential difference (VCPDV_{CPD}) between the tip and the sample.

The work function of the sample is calculated using the following formula

VCPD=WFtip−WFsample−e(27)V_{CPD} = \frac{WF_{\mathrm{tip}} - WF_{\mathrm{sample}}}{-e} \tag*{(27)}

Where WFtipWF_{\mathrm{tip}} is the work function of the tip and WFsampleWF_{\mathrm{sample}} is the work function of the sample.[24, 25]

To accurately measure the work function of the perovskite layer and the charge transfer layer, the work function of the known sample (gold W.F. = 5.1 eV) is used to calibrate the tip work function.[25] Later, the estimated tip work function and measured VCPDV_{CPD} used to calculate the sample work function.

Figure 12 (PDF p. 119). Schematic representation of the Kelvin probe force microscopy.

Contact angle measurement.

The contact angle is the measure of hydrophilicity/hydrophobicity of any surface. The contact angle is defined as the angle between the solid surface and the liquid where they meet.[26] A higher contact angle (i.e. hydrophobicity) is desirable where degradation due to moisture is prone to happen. However, the lower contact angle (i.e. hydrophilicity) is necessary where a layer is depositing on a surface.[27] A surface will be hydrophobic if it contains methyl group ($-\mathrm{CH}_{3}$) or long alkyl chains which will be discussed in more detail in chapter 5. We have used the GBX Digidrop instrument to measure the contact angle, which also measures the surface energy, surface tension, absorption, etc. The schematic diagram of contact angle measurement is shown in Figure 13. A water droplet of 11-2 μL2\,\mu\mathrm{L} is dropped from the syringe to the surface of the sample. The volume of the water droplet is controlled by the software Windrop in the GBX Digidrop system. Once the water drop touches the substrate surface, it slowly spreads with time and after a few seconds, it becomes stable. The water drop spreading is captured by the camera. The angle between the surface and the liquid at stable conditions results in the contact angle.

Schematic diagram of the water contact angle measurement showing a syringe, water drop, sample stage, camera, light source, and computer

Figure 13 (PDF p. 120). The schematic diagram of the water contact angle measurement between the surface and the water droplet.

Nuclear magnetic resonance spectroscopy

Nuclear magnetic resonance (NMR) spectroscopy is a highly advanced technology based on the orientation of the atomic nuclei with non-zero spins placed in a strong external magnetic field.[28] A spinning charge generates a magnetic field and that results in a magnetic moment of strength proportional to the spin of the charge. Thus, in the presence of the external magnetic field, there are two spin states exist i.e. one in the direction of the magnetic field and another opposite to the direction of the field. The reorientation of the magnetic moments occurs with the absorption of electromagnetic radiation in the RF frequency range 4 – 900 MHz, and changes with the external magnetic field (6 - 24 T) and results in chemical shift Figure 14.[29]

The NMR spectroscopy is widely used to study the molecular dynamics, week intermolecular interactions, ionization energy, chemical analysis, etc. There are various NMR active nuclei used to study the NMR spectroscopy such as 1H^{1}\mathrm{H}, 2D^{2}\mathrm{D}, 13C^{13}\mathrm{C}, 15N^{15}\mathrm{N}, 19F^{19}\mathrm{F}, etc. We used JEOL, ECZR series 600 MHz NMR spectrometer for analyzing the solution state proton NMR, which will be discussed in Chapter 5.

Figure 14 (PDF p. 121). The schematic diagram of the nuclear magnetic resonance spectroscopy.

Device characterization

Current density vs. voltage (J−VJ - V) characterization

The fundamental way to test a solar cell's performance is by measuring the current vs. voltage (I−VI - V) characteristics. The current (II) depends on the active area (AA) of the solar cells; hence, the current density (J=I/AJ = I/A) is an important factor to study the efficiency performance of the device. All the photovoltaic measurements were carried out under ambient conditions. Photocurrent density vs. applied voltage (JJ-VV) measurement has been carried out using Keithley 4200 SCS and LED solar simulator (ORIEL: LSH-7320) after calibrating through a standard Si solar cell (RERA SYSTEMS-860 reference cell) Figure 15. The JJ-VV measurement was performed with a scan rate of 100 mV/s and a hold time of 10 ms. For the intensity-dependent J−VJ - V measurement, we used a set of ND filters with different optical densities (this will be explained in detail in Chapter 6).

The power conversion efficiency (PCE) of a device is calculated by using the following formula[30]

PCE=JSC×VOC×FF(28)PCE = J_{SC} \times V_{OC} \times FF \tag*{(28)}

Where, JSCJ_{SC} is the short circuit photocurrent density, VOCV_{OC} is the open circuit voltage, and FFFF is the fill factor Figure 15. The FF is defined by the maximum squareness of the J−VJ - V curve, and it can be represented as,

FF=Jmax⁡×Vmax⁡JSC×VOC(29)FF = \frac{J_{\max} \times V_{\max}}{J_{SC} \times V_{OC}} \tag*{(29)}

The solar cell is a diode and the dark current is similar to the p-n junction diode which increases exponentially with the applied forward voltage as

J=J0exp⁡(qV/nkT)(30)J = J_0 \exp(qV/nkT) \tag*{(30)}

Where J0J_0 is the reverse saturation current density, qq is the elementary charge, nn is the ideality factor, kk is the Boltzmann constant, and TT is temperature.

Upon applying the 1-Sun light, the photogenerated charges take place, and the total current in the device can be written as [31]

J=−JL+J0[exp⁡(V−JRsnkT/q)−1]+V−JRsRSh(31)J = -J_L + J_0 \left[\exp\left(\frac{V - JR_s}{nkT/q}\right) - 1\right] + \frac{V - JR_s}{R_{Sh}} \tag*{(31)}

Where JLJ_L is the illuminated current, RSR_S and RShR_{Sh} are the series and shunt resistance shown in the equivalent circuit diagram in Figure 16.

Figure 15 (PDF p. 123). Schematic representation of I-V measurement and the current density vs. voltage (J−VJ - V) characteristics with the PV parameters for a typical perovskite solar cell.

Figure 16 (PDF p. 123). Schematic representation of the equivalent circuit diagram under illumination conditions including the series and shunt resistance with a single diode model.

Incident photon to current efficiency

The incident photon to current efficiency (IPCE) is a powerful technique to characterize a solar cell and helps in the optical engineering of the solar cell device. The IPCE technique is used to estimate the spectral responsivity (SR), external quantum efficiency (EQE), and internal quantum efficiency (IQE). The IPCE spectrum of all the perovskite solar cells was measured using the Bentham PVE 300 system. The schematic diagram of the IPCE spectrum measuring set-up is shown in Figure 17. In the Bentham PVE 300 set-up, there are two lamps: (i) 75W Xenon lamp and (ii) 100W Quartz halogen lamp (QTH), which are used to make a white light source to produce photons. The intensity of the lamp is less than 1 mW/cm21\ \mathrm{mW}/\mathrm{cm}^{2}. The light passes through a monochromator to get the desired wavelength. The wavelength selection and IPCE measurement are done with BenWin+ software in the Bentham PVE 300 system. The measurement can be done in three modes (DC, AC, and transformer), and for the AC or transformer mode, the DC light is chopped with a chopper at a frequency of 390 Hz and is considered as a reference signal for the lock-in amplifier. The chopped light passes through optical lenses and focuses on the device. The light spot should be lower than the active area of the device (should fall inside the active area) and for that different slits are used for different active area devices. The Bentham PVE 300 set-up is calibrated every time before starting the measurement using a standard reference Si diode (300-1100 nm)

Figure 17 (PDF p. 124). Schematic diagram of the incident photon to current efficiency (IPCE) measurement set-up.

The EQE of a solar cell is defined as the ratio of the number of charge carriers collected by the solar cell to the number of incident photons (of a given energy) on the device.[32]

EQE(λ)=number of charge carriers collected (λ)number of incident photons (λ)EQE(\lambda)=\frac{\textit{number of charge carriers collected }(\lambda)}{\textit{number of incident photons }(\lambda)}

The Bentham PVE 300 set-up has an integrating sphere to measure the reflection R (%) and transmission T (%) spectra. The IQE spectra are obtained from the EQE spectra by using the reflection and transmission spectra as follows[33]

IQE(%)=EQE(%)1−0.01×R(%)−0.01×T(%)(32)IQE(\%)=\frac{EQE(\%)}{1-0.01\times R(\%)-0.01\times T(\%)} \tag*{(32)}

Electroluminescence quantum efficiency measurement

The steady state electroluminescence quantum efficiency (QEEL_{\mathrm{EL}}) is measured in an open atmosphere and in the dark condition to avoid any randomly scattered light. To do the QEEL_{\mathrm{EL}} measurement, we used a 100 mm2^2 Si-photodetector which is connected with a 12V DC supply, and a Keithley 2400 multimeter (for photovoltage measurement) as shown in Figure 18. The device was injected a current with a Keithley 2000 source meter. Both the Keithley 2400 and Keithley 2000 are connected in series using National Instrument (NI) GPIB to GPIB cable. Both the Keithleys are interfaced using the LabView program to record the real-time current-voltage-luminescence (J−V−LJ-V-L). The EL spectra are collected using an optical fiber and Ocean optics spectrometer (FLAME-T-XR1-ES).

Figure 18 (PDF p. 125). Schematic diagram of the electroluminescence quantum efficiency (QEEL_{\mathrm{EL}}) measurement set-up.

Transient photovoltage/photocurrent measurement

Transient photovoltage (TPV) is an optoelectronic measurement technique widely used in the organic solar cell community, where the steady-state device is kept in an open circuit condition (along with background 1-Sun white light bias) using a high external resistance across the device and a short-lived laser perturbation is applied to the device.[34] The schematic of the TPV measurement is shown in Figure 19. For the TPV measurement, the termination resistance used is 1MΩ. The solar cell is illuminated by 1-sun DC white light using a THORLABS white lamp S/N M00304198. A 490 nm TOPTICA diode laser is used to perturb the device. The diode laser is modulated with a frequency of 1 kHz and duty cycle of 0.05% (for 500 ns pulse) using an ArbStudio 1104 function generator. The perturbed laser intensity is chosen such that the perturbed voltage ΔVOC≈10 mV\Delta V_{OC} \approx10\ \mathrm{mV} while the background DC light intensity was 1-sun.[35] The perturbed signal is measured using a digital oscilloscope Tektronix DPO 4104B. The intensity of the background DC white light was changed using a set of neutral density (ND) filters for the intensity dependent TPV measurement. The intensity-dependent TPV analysis will be discussed in detail in Chapter 6. For the TPC measurement, the termination resistance used is 50Ω. In this case, the generated perturbed charge carriers leave the device due to short circuit conditions.[36, 37] The TPC analysis will be discussed in good detail in appendix A.

Figure 19 (PDF p. 126). Schematic diagram of the transient photovoltage/photocurrent measurement set-up.

Frequency dependent photocurrent measurement

Frequency-dependent photocurrent measurement is a highly promising technique to understand the defect states and dielectric relaxation process in solar cell devices.[38] A frequency modulated laser is applied to the solar cell device and the corresponding photocurrent is recorded in a lock-in amplifier. A 490 nm TOPTICA diode laser is used whose ON/OFF frequency (duty cycle 50%) is modulated using an ArbStudio 1104 function generator. We used an SR830 DSP lock-in amplifier whose reference channel frequency range is 1 mHz to 102 kHz hence we modulated the laser ON/OFF frequency from 200 Hz to 102 kHz. The photogenerated current is frequency dependent and has real and imaginary parts. By analyzing the imaginary part, the dielectric relaxation is determined for perovskite solar cells.[39] This technique will be discussed in good detail in chapter 4.

Figure 20. Frequency-dependent photocurrent measurement set-up.

References

Role of Monovalent Cation on the Dielectric Relaxation Processes in Hybrid Metal Halide Perovskite Solar Cells and Correlation with Thermal Stability

Role of Monovalent Cation on the Dielectric Relaxation Processes in Hybrid Metal Halide Perovskite Solar Cells and Correlation with Thermal Stability

Perovskite solar cells (PSCs) are the fastest growing photovoltaic devices in the solar cells community and offer a bright future for cheap solar electricity. In the last few years, it has been observed that the efficiency and stability of the PSCs can be enhanced by introducing multications into the perovskite crystal structure. Herein, we have examined the triple monovalent cation based PSC (FA0.83MA0.17)0.95Cs0.05Pb(I0.90Br0.10)3(\mathrm{FA}_{0.83}\mathrm{MA}_{0.17})_{0.95}\mathrm{Cs}_{0.05}\mathrm{Pb}(\mathrm{I}_{0.90}\mathrm{Br}_{0.10})_3 (abbreviated as CsFAMA) over single monovalent cation based PSC MAPbI3\mathrm{MAPbI}_3 (MAPI) through frequency dependent photocurrent and dielectric measurements in terms of the dielectric relaxation process. The dielectric relaxation time constant (τd\tau_d) is lower for the CsFAMA based PSC as compared to the MAPI based PSC. The lower τd\tau_d is attributed to a lower dielectric constant using classical semiconductor physics. Unprecedentedly, the relaxation process is correlated with the presence of a monovalent cation of CsFAMA vs MAPI as an absorber in PSCs, which is well correlated with the presence of relative defect density. We note that defect formation under illumination conditions reduces JSCJ_{\mathrm{SC}} as a major component in PV parameters, which is significant in MAPI PSCs. This can be hypothesized based on relative ion density accumulation at the interface of charge extracting contact as per available literature. This study provides a unique, in-depth knowledge towards the role of monovalent cations in the dielectric relaxation process and their connection with the defects and thermal stability of halide perovskite semiconductors based solar cells.

Introduction

Organic-inorganic halide perovskite solar cells (PSCs) have enticed much interest in the semiconductor community and promise to deliver the next generation of highly cost-effective photovoltaic devices. The PSCs are the fastest growing photovoltaic devices in the solar cell community and reached the certified efficiency from 3.8% in 2009[1] to >25.6%.[2, 3] There are various engineering techniques reported in the perovskite research community to optimize the active and charge transport layers, such as additive engineering, anti-solvent treatment, interface, and surface modification, etc. to improve the power conversion efficiency (PCE) and the stability of the PSCs.[4, 5, 6] In general, the enhanced performance of the PSC is attributed to the improved morphology and suppressed non-radiative recombination. However, in-depth insight into the stability of the PSC is still being explored in the community, which involves both intrinsic and extrinsic factors. One needs to start at the intrinsic properties of the absorbing semiconductor as a pre-requisite and then at solar cell device physics.[7, 8, 9]

It is widely known that PCE and stability are two important factors that are essential for the commercialization of PSCs. The PCE of PSCs has reached close to the conventional commercialized single junction Si solar cells[2], but the stability is still under optimization.[10, 11, 12] The stability of the PSCs depends on various factors, for example, moisture resistivity, oxygen induced degradation, thermal degradation or ion migration, etc.[13, 14, 15] The moisture and oxygen induced degradations can be reduced by advanced encapsulation techniques.[16, 17] However, thermal degradation is crucial and needs to be addressed for future stable PSCs. MAPbI3\mathrm{MAPbI_3} (MAPI) perovskite is the most widely explored material in the perovskite community in terms of photophysics and device applications. However, single monovalent cation based PSCs possess numerous defects and lead to lower PCE and stability.[18, 19, 20] These defects act as recombination centers for the charge carriers and act as migration channels for the ions under illumination.[21] The defects in the perovskite can be reduced by changing the cation and anion or both simultaneously. Hence, instead of MA+\mathrm{MA^+}, thermally stable FA+\mathrm{FA^+} ion-based perovskites are introduced.[22] It is found that FAPbI3\mathrm{FAPbI_3} (FAPI) is potentially more suitable than the conventional MAPI perovskite, but it has also been reported that the FAPI trigonal α\alpha phase is quite sensitive to humidity and easily transforms into a non-photo active hexagonal δ\delta-phase at room temperature.[23] It is observed that the partial introduction of Cs cation in the FAPI perovskite further increases the structural and thermal stability of the photoactive cubic α\alpha-phase.[24, 25] From the first principle calculation, it is reported that the formation energy of the point defects can be maximized by mixing the different A cations in the ABX3\mathrm{ABX_3} perovskite, and correspondingly the point defects reduce, as we discussed in Chapter 2.[26] Later on, it is found that the triple cation (CsFAMA) and mixed halide (IBr) based perovskite thin films are used to fabricate highly efficient single junction PSCs.[7, 27, 28] It is reported that the mixing of different A cations leads to a higher tolerance factor (tt) value through lattice strain relaxation and thus it minimizes the defects to achieve a relatively stable crystal structure.[29, 30] In order to have more clarity behind the stability issue of these ionic perovskite semiconductors is still being explored in the community as an important research problem. Perovskite is being debated to be a ferroelectric material, however, it is proven to have a frequency dependent dielectric polarization as expected.[31, 32, 33] A distinct change in the performance of halide perovskites as a function of A-site is evident and being explained using first principle calculations, however, there is no experimental macroscopic study which connects this stability issue to the intrinsic semiconductor property of absorber. Semiconducting absorbers are known to have their dielectric relaxation time (τd\tau_d) depends on dielectric constant and conductivity.[34] These two macroscopic quantities have been reported however, a direct link of them to τd\tau_d is missing. Recently, Singh et. al. reported a correlation between the τd\tau_d to charge transport properties in MAPI films.[35] Whether τd\tau_d can be linked to the defects in the perovskite and connects to the device stability for these ionic character semiconductors has not explored so far as per our knowledge.

In this work, we fabricated conventional pp-ii-nn architecture based monovalent single cation MAPI and triple cation (FA0.83MA0.17)0.95Cs0.05Pb(IBr)3\mathrm{(FA_{0.83}MA_{0.17})_{0.95}Cs_{0.05}Pb(IBr)_3} (CsFAMA) based PSCs to study the dielectric relaxation process and correlate with the defects studies. A frequency dependent photocurrent (Iph(f)I_{ph}(f)) measurement has been carried out to study the dielectric relaxation time constant (τd\tau_d). We observed that the τd\tau_d for CsFAMA perovskite is lower than that for MAPI. To validate the Iph(f)I_{ph}(f) measurement, we carried out conductivity studies and frequency dependent capacitance measurements for CsFAMA and MAPI based perovskites. We carried out various optoelectronic measurements to compare the defects present in the MAPI and CsFAMA perovskite and found that the CsFAMA perovskite has lower defects than MAPI perovskite in confirmation with literature.[28, 29] Interestingly, the relatively lower defect states for CsFAMA perovskite can be attributed to measured lower capacitance. We also measured the stability of MAPI and CsFAMA based PSCs and found the stability of CsFAMA is much higher than that of MAPI based PSCs. We correlated the Iph(f)I_{ph}(f) with the polarization of A-cation in terms of defect mediated dielectric relaxation process to validate the enhanced stability of CsFAMA-based PSCs.

Experimental Section

Materials: Methylammonium iodide (MAI), lead iodide(PbI2\mathrm{PbI}_{2}), formamidinium iodide (FAI), cesium iodide (CsI), lead bromide (PbBr2\mathrm{PbBr}_{2}) all were purchased from TCI chemical and used as received. Phenyl-C61-butyric acid methyl ester (PC61BM\mathrm{PC}_{61}\mathrm{BM}) were purchased from Solenne BV and used as received. The poly(9,9-bis(3′-(N,N-dimethyl)-N-ethylammoinium-propyl-2,7-fluorene)-alt-2,7-(9,9-dioctylfluorene))dibromide (PFN-Br) ordered from 1-Material and used as received. Poly (triaryl amine) (PTAA) and bathocuproine (BCP) were purchased from Sigma Aldrich and used as received

Solution preparation: (FA1−XMAX)0.95Cs0.05PbI3(\mathrm{FA}_{1-X}\mathrm{MA}_{X})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} solution prepared by mixing of FA0.95Cs0.05PbI3\mathrm{FA}_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} and MA0.95Cs0.05PbI3\mathrm{MA}_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} in different volume ratios. FA0.95Cs0.05PbI3\mathrm{FA}_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite precursor prepared by mixing of 212 mg of FAI, 17 mg of CsI, and 599 mg of PbI2\mathrm{PbI}_{2} in 1 ml of a mixed solvent of DMF and DMSO in a 4:1 volume ratio. MA0.95Cs0.05PbI3\mathrm{MA}_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite precursor prepared by mixing of 196 mg of MAI, 17 mg of CsI, and 599 mg of PbI2\mathrm{PbI}_{2} in 1 ml of a mixed solvent of DMF and DMSO in a 4:1 volume ratio. We added FA0.95Cs0.05PbI3\mathrm{FA}_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} and MA0.95Cs0.05PbI3\mathrm{MA}_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} in the volumetric ratio to obtain X = 0.00, 0.25, 0.50, 0.75, 1.00. MAPI perovskite precursor solution was prepared by mixing 581 mg PbI2\mathrm{PbI}_{2} and 209 mg MAI in 1 ml of a mixed solvent of DMF and DMSO in a 4:1 volume ratio and stirred at room temp for 4 hours before spin coating. The CsFAMA perovskite precursor was prepared by mixing of 16.9 mg of CsI, 33.1 mg of MAI, 71.6 mg of PbBr2\mathrm{PbBr}_{2}, 176.6 mg of FAI, and 509.4 mg of PbI2\mathrm{PbI}_{2} in 1 ml of a mixed solvent of DMF and DMSO in 4:1 volume ratio and stirred overnight at room temperature before spin coating. 0.5 mg of PFN-Br is dissolved in 1 ml of methanol and stirred overnight at room temperature. 1.5 mg of PTTA is dissolved in 1 ml of toluene and stirred overnight at room temperature. 20 mg PC61BM\mathrm{PC}_{61}\mathrm{BM} is dissolved in 1 ml of 1,2-dichlorobenzene and stirred for overnight at room temperature. 0.5 mg BCP is dissolved in 1 ml of anhydrous isopropanol and stirred overnight at room temperature and 10 minutes at 70° C before spin coating.

Perovskite solar cells fabrication and characterization: ITO-coated glass (10-15 Ω\Omega/square) substrate was sequentially cleaned with hellmanex soap solution, deionized (DI) water, acetone, and isopropanol for 10 minutes each. After drying the substrate with a nitrogen gun, we kept them inside the oxygen plasma ashing chamber for 10 minutes. After plasma ashing, we immediately spin-coat PTAA at 2000 rpm for 30 seconds and then annealed at 100° C for 10 minutes under a nitrogen environment. After that, we cooled down the substrates for 5 minutes, and then PFN-Br solution was spin-coated at 5000 rpm for 30 seconds, and immediately the perovskite solution spin-coated on the ITO/PTAA/PFN-Br substrates. All the perovskite precursors were filtered using 0.45 μ\mum PTFE filter before spin coating. The (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite spin-coated at 4000 rpm for 30 seconds, and at the last 10 seconds, we used 150 μ\mul of chlorobenzene anti-solvent treatment on a 1.5 cm by 1.5 cm substrate and then annealed at 100°C for 30 minutes. The MAPbI3\mathrm{MAPbI}_{3} (abbreviated as MAPI) perovskite spin-coated at 4000 rpm for 30 seconds and at the last 10 seconds, we used 150 μ\mul of chlorobenzene anti-solvent treatment on a 1.5 cm by 1.5 cm substrate and then we did 5 minutes of room temperature dry and then annealed at 100°C for 10 minutes. The (FA0.83MA0.17)0.95Cs0.05Pb(I0.90Br0.10)3(\mathrm{FA}_{0.83}\mathrm{MA}_{0.17})_{0.95}\mathrm{Cs}_{0.05}\mathrm{Pb}(\mathrm{I}_{0.90}\mathrm{Br}_{0.10})_{3} (abbreviated as CsFAMA) perovskite spin-coated at 5000 rpm for 30 seconds and at the last 10 seconds, we used 150 μ\mul of chlorobenzene anti-solvent treatment on a 1.5 cm by 1.5 cm substrate and then annealed at 100°C for 30 minutes. After that, PC61BM\mathrm{PC}_{61}\mathrm{BM} was spin-coated at 2000 rpm for 30 seconds and BCP was spin-coated at 5000 rpm for 20 seconds. Finally, 120 nm of Ag was deposited under a vacuum of 2×10−62 \times10^{-6} mbar using a metal shadow mask. For all studied devices, the active area of the device is 17.5 mm2\mathrm{mm}^{2}.

All the photovoltaic measurements were carried out under ambient conditions. Photocurrent density versus applied voltage (JJ-VV) measurement was carried out using Keithley 4200 SCS and an LED solar simulator (LSH-7320) after calibrating through standard Si solar cells provided by ABET, IIT Bombay. The JJ-VV measurement was performed with a scan rate of 40 mV/s. EQE measurement has been carried out to measure the photoresponse as a function of wavelength using the Bentham quantum efficiency measurement system (Bentham PVE 300). The XRD measurements were carried out in a Rigaku smart lab diffractometer with Cu Kα\alpha radiation (λ\lambda=1.54Å). Θ\Theta-2Θ\Theta scan has been carried out from 10°-45° with step size 0.001°. Morphological analysis was done using field emission scanning electron microscopy (FESEM). Optical absorption spectra were carried out using a spectrometer (PerkinElmer LAMBDA 950). Steady-state PL measurement was done on thin films in a vacuum at a pressure of 10−310^{-3} mbar in a custom-made chamber. Excitation energy and wavelength were 70 nJ and 355 nm respectively. For time-resolved PL, a gated i-CCD (ICCD, Andor iStar) was used for detection with excitation laser (3rd harmonic Nd:YAG with emission wavelength of 355 nm) pulse width of 840 ps pulse with 1 kHz repetition rate. The PLQY measurement was done using a 4-inch customized integrating sphere and Nd3+\mathrm{Nd}^{3+}:YAG 355 nm laser and ICCD, Andor iStar-based spectrometer using a method developed earlier by de Mello et al. [36] Steady-state current-voltage-light characteristics were measured using a Keithley 2400 source meter, Keithley 2000 multimeter, and calibrated Si photodiode (RS components). The capacitance of the devices was measured by using Novocontrol Technology, concept 80. TPV was measured by using a 490 nm TOPTICA diode laser, THORLABS white lamp S/N M00304198, ArbStudio 1104, and digital oscilloscope Tektronix DPO 4104B.

Results

Dielectric relaxation process in perovskites

We fabricated the inverted (pp-ii-nn) heterojunction PSCs with device structure of indium tin oxide (ITO)/ poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine(PTAA)/poly(9,9-bis(3′-(N,N-dimethyl)-N-ethylammonium-propyl-2,7-fluorene)-alt-2,7-(9,9 dioctylfluorene)) dibromide (PFN-Br)/ perovskite / phenyl-C61_{61}-butyric acid methyl ester(PC61_{61}BM)/ bathocuproine(BCP)/ silver (Ag) to understand the role of monovalent cations in dielectric relaxation process. Figure 1a represents the crystal structure of ABX3_3 perovskite where A is a monovalent cation (MA+^+/FA+^+/Cs+^+) represented by grey balls; B is a divalent cation (Pb2+^{2+}) represented by blue balls at the center of the octahedral cage, and X is the halide (I−^-/Br−^-) represented by pink balls at the corner of octahedral. In this work, we focus on the thermal stability of the PSCs as a function of ‘A’ cation and correlate it with the dielectric relaxation process. Figure 1b depicts the device

Figure 1 (PDF p. 139). (a) Schematic diagram of ABX3_3 perovskite crystal structure. (b) Energy level diagram of each layer used in the pp-ii-nn architecture based PSC, values are taken from literature.[37, 38] (c) Typical frequency dependent photocurrent measurement (Iph(f)I_{\mathrm{ph}}(f)) set-up used

for this study and an inset image of actual device picture with two dots (blue on device & red off device) to show sites at which light is irradiated for Iph(f)I_{\mathrm{ph}}(f) measurements. (d) Quadrature part of the frequency dependent normalized photocurrent for (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3}, where X=0.00X = 0.00, 0.25, 0.50, 0.75, and 1.00. (e) The frequency dependent device capacitance was measured for (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite absorbers. (f) The variation in capacitance and corresponding photocurrent resonance frequency with respect to XX in (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} composition.

configuration of the pp-ii-nn architecture based PSCs with an energy level diagram of each layer.[37, 38] Detailed information about the perovskite precursor preparation, device fabrication, and characterization are given in the experimental section above.

The Iph(f)I_{\mathrm{ph}}(f) measurement has been carried out to understand the dielectric relaxation process in PSCs.[39, 40] Figure 1c represents the schematic diagram of the experimental set-up used to determine the τd\tau_{\mathrm{d}}. The frequency of the input signal (490 nm diode laser) is varied from 200 Hz to 100 kHz with a duty cycle of 50% using the Arb Studio 1104 function generator. The photogenerated current is measured through the SRS830 DSP lock-in amplifier, which also provides the phase delay (ϕ\phi) between the device signal and the input signal. The inset figure represents the actual experimental device having 3 pixels (with Ag electrode) with an active area of each pixel is 17.5 mm2^{2}. The rest of the area that is not covered by the Ag electrode is termed as outside the active area of the device. In order to study the effect of monovalent ‘A’ cation on the dielectric relaxation process in the ABX3\mathrm{ABX}_{3} perovskite, we choose a series of perovskite systems denoted by (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3}, where XX varies from 0.00 to 1.00 with a step of 0.25 and the atomic concentration of Cs, Pb, & I remains constant. The crystallographic XRD study shows incorporating of smaller size MA cation by increasing XX in the (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite system, the (110) peaks shift towards higher angle from 13.94° to 14.11° Figure 2.[41] Optical studies UV-vis-NIR absorption spectra and steady-state photoluminescence (SSPL) suggest that incorporating of smaller size MA cation by increasing XX in the (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite system increases the bandgap Figure 3. We fabricated a series of pp-ii-nn architecture based (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} PSCs (Figure 1b). The illuminated JJ-VV characteristics and corresponding PV parameters of (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} based PSCs are shown in Figure 4 and Table 1 respectively. In this chapter, the Iph(f)I_{\mathrm{ph}}(f) is used to determine the relaxation time of the photogenerated charge carriers in the PSC by analyzing the quadrature (imaginary) component of the photocurrent. Figure 1d shows a typical quadrature part of the normalized photocurrent (IPh,qdI_{\mathrm{Ph},qd}) as a function of frequency for (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite systems. The in-phase (IPh,in)(I_{\mathrm{Ph,in}}) part of the normalized photocurrent (IPh,qd)(I_{\mathrm{Ph,qd}}) as a function of frequency is shown in Figure 4b. It is observed that Iph,qd(f)I_{\mathrm{ph,qd}}(f) slowly increases with an increase in frequency and then attains a maximum

Normalized XRD patterns, peak positions, and peak shift

Figure 2 (PDF p. 142). (a) Normalized XRD pattern of (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite films where X = 0.00, 0.25, 0.50, 0.75, and 1.00. (b) Peak position of (110) plane of (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite films. (c) the shift of (110) peak in 2θ2\theta degree with respect to X composition.

Absorption, steady-state normalized PL spectra, and PL peak position and bandgap

Figure 3 (PDF p. 143). (a) Absorption, (b) steady-state normalized PL spectra for the (FA1−XMAX)0.95Cs0.05PbI3(\mathrm{FA}_{1-X}\mathrm{MA}_{X})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite films where X=0.00X=0.00, 0.250.25, 0.500.50, 0.750.75, and 1.001.00. (c) PL peak position and bandgap for the (FA1−XMAX)0.95Cs0.05PbI3(\mathrm{FA}_{1-X}\mathrm{MA}_{X})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} perovskite films.

Illuminated J-V characteristics and in-phase part of the frequency dependent photocurrent

Figure 4 (PDF p. 144). (a) Illuminated JJ-VV characteristics and (b) in-phase part of the frequency dependent photocurrent of the (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} PSCs, where X=0.00X = 0.00, 0.250.25, 0.500.50, 0.750.75, and 1.001.00.

DeviceJSCJ_{SC} (mA/cm2^2)VOCV_{OC} (V)FFFFPCE (%)
X = 0.0020.560.980.6012.12
X = 0.2520.251.050.6313.47
X = 0.5019.120.940.6912.44
X = 0.7517.681.010.6611.74
X = 1.0017.531.040.7413.56

Table 1 (PDF p. 144). The PV parameters of the (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} PSCs, where X=0.00X = 0.00, 0.250.25, 0.500.50, 0.750.75, and 1.001.00.

photocurrent value at a specific frequency (termed as resonance frequency). After the resonance frequency, the photocurrent starts decreasing and attains zero value at a higher frequency. It is observed that the resonance frequency decreases with an increase in XX in composition (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3}. Further, to underpin the trend in resonance frequency for different compositions of XX in the Iph(f)I_{\mathrm{ph}}(f) measurement, a frequency dependent capacitance measurement is carried out for the (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} based PSCs, and shown in Figure 1e. The frequency dependent capacitance increases with an increase in XX. It suggests that the capacitance is also a function of the A-site cation.[42] The variation in the resonance frequency and the capacitance with respect to the X (%) for (FA1−x_{1-\mathrm{x}}MAx_{\mathrm{x}})0.95_{0.95}Cs0.05_{0.05}PbI3_3 based PSCs is shown in Figure 1f. The effect of A-site cation on the resonance frequency will be explained in the discussion part in good detail.

Image not extracted

Figure 5 (PDF p. 145). Top-view FESEM images of (a) MAPI and (b) CsFAMA perovskite films. The cross-section FESEM images of the complete device of (c) MAPI and (d) CsFAMA PSCs.

Image not extracted

Figure 24 (PDF p. 145).

Image not extracted

Figure 6 (PDF p. 146). (a) X-ray diffraction (XRD) pattern of MAPI and CsFAMA perovskite films. (b) The energy bandgap is calculated from the absorption spectra.

Image not extracted

Figure 7 (PDF p. 146). (a) Dark and illuminated JJ-VV characteristics of MAPI and CsFAMA based PSCs in forward and reverse scan. (b) EQE spectra of MAPI and CsFAMA based PSCs and the integrated JSCJ_{\mathrm{SC}} according to the corresponding EQE spectra (Dip in EQE is due to reflection losses, see Figure 9). (c) Quadrature part of the frequency dependent normalized photocurrent for MAPI and CsFAMA based PSCs fitted with Lorentzian peak function. (d) Frequency dependent capacitance of MAPI and CsFAMA based PSCs.

DeviceJSCJ_{\mathrm{SC}} (mA/cm2^{2})VOCV_{\mathrm{OC}} (V)FFFFPCE (%)Int. JSCJ_{\mathrm{SC}} (mA/cm2^{2})HI index (%)
MAPI (F)20.301.060.7215.49
18.36 ± 1.631.02 ± 0.030.71 ± 0.0413.98 ± 0.8119.344.32
MAPI (R)20.921.060.7316.19
19.59±1.5719.59 \pm1.571.03±0.031.03 \pm0.030.72±0.030.72 \pm0.0314.50±0.7914.50 \pm0.79
CsFAMA (F)23.0223.021.121.120.780.7820.1120.1121.8821.881.671.67
22.04±0.8122.04 \pm0.811.10±0.021.10 \pm0.020.75±0.030.75 \pm0.0318.15±0.8918.15 \pm0.8921.8821.881.671.67
CsFAMA (R)23.1123.111.121.120.790.7920.4520.4521.8821.881.671.67
22.53±0.8822.53 \pm0.881.10±0.021.10 \pm0.020.76±0.030.76 \pm0.0318.76±0.8418.76 \pm0.8421.8821.881.671.67

Table 2. PV parameters of the best performed and averaged over 24 PSCs under 1-sun illumination conditions in forward (F) and reverse (R) scan directions.

Further, we extended the frequency dependent photo current study to the widely explored MAPI and CsFAMA based PSCs to understand the dielectric relaxation process in these two wonderful and interesting perovskite systems.[43, 44] These devices were fabricated in pp-ii-nn configuration (Figure 1b). It is found that both the films are compact and grain size is relatively bigger in CsFAMA film in comparison to MAPI film (Figure 5). Figure 7a represents the dark and illuminated current density vsvs voltage (JJ-VV) characteristics in forward and reverse scan direction. The illuminated JJ-VV characteristics are carried out under 1-Sun (100 mW/cm2^2) condition, and all the photovoltaic parameters for the champion MAPI and CsFAMA based PSCs are listed in Table 2. The illuminated PV parameters for the MAPI based PSCs are, short circuit current density (JSCJ_{SC}) = 20.92 mA/cm2^2, open-circuit voltage (VOCV_{OC}) = 1.06 V, fill factor (FFFF) = 0.73 and PCE = 16.19 %. The illuminated PV parameters for the CsFAMA based PSCs are, JSCJ_{SC} = 23.11 mA/cm2^2, VOCV_{OC} = 1.12 V, FFFF = 0.79 and PCE = 20.45 %. The dark JJ-VV characteristics in the semi-log scale for both devices suggest three times lower leakage current density (at the bias of - 0.5V) in CsFAMA than MAPI based PSC (Figure 8b (PDF p. 181)).[45] The higher value of JSCJ_{SC} for CsFAMA based PSCs may be attributed to the higher built-in potential VbiV_{bi} = 0.89 V than MAPI VbiV_{bi} = 0.84 V measured from dark JJ-VV characteristics (Figure 8c (PDF p. 181)). The higher value of JSCJ_{SC} in the CsFAMA based PSC is in good agreement with the external quantum efficiency (EQE) spectrum (Figure 7b). EQE spectrum shows a dip around 600 nm, similar to previous reports.[46, 47] In order to understand the origin of these PSCs EQE spectral profiles, reflection spectra were carried out using an integrating sphere in Bentham PVE 300. We observed the dip in the EQE spectrum correlates well with the hump in the reflection spectrum of the PSCs Figure 9(a) and (b). Hence, it is purely a micro-cavity effect for a multilayer thin-film device with back metal reflecting contact. The integrated JSCJ_{SC} over the EQE spectrum (Figure 7b) is 19.34 mA/cm2^2 and 21.88 mA/cm2^2 for MAPI and CsFAMA PSC respectively

Table 2.

Image not extracted

Figure 8 (PDF p. 148). (a) Absorption and PL spectra of MAPI and CsFAMA perovskite films. (b) Dark JJ-VV characteristics of MAPI and CsFAMA PSCs. (c) Built-in potential of MAPI and CsFAMA based PSCs. (d) Conductivity of MAPI and CsFAMA based PSCs.

The slightly lower value (∼5%\sim5\%) of integrated JSCJ_{SC} compared to the JSCJ_{SC} obtained from illuminated JJ-VV measurement can be attributed to either the edge effect from the active area (17.5 mm217.5\ \mathrm{mm}^{2}) of the devices or ion migration or pre-bias measurement condition.[48, 49] The lower value of JSCJ_{SC}, VOCV_{OC} and FFFF of MAPI based PSC compared to the CsFAMA based PSC is obtained, which will be discussed in the subsequent part. The hysteresis Index (HI)[50] of a solar cell indicates the ionic migration in the solar cell and can be calculated as

HI=PCE(reverse)−PCE(forward)PCE(reverse)(33)\mathrm{HI} = \frac{PCE(\mathrm{reverse})-PCE(\mathrm{forward})}{PCE(\mathrm{reverse})} \tag*{(33)}

The HI for MAPI based PSC is 4.32 % whereas for the CsFAMA based PSC it is 1.67 %. This can be correlated with the relatively higher degree of ion migration in MAPI based PSC due to the presence of defect states.[51, 52] Here, we draw attention to the fact that the devices are not treated with any kind of passivating agents purposefully in order to have an in-depth insight into the relaxation process in PSCs based on their crystal structure.

Image not extracted

Figure 9 (PDF p. 149). (a) EQE, IQE and reflection spectroscopy (R) of (a) MAPI and (b) CsFAMA based PSCs.

Image not extracted

Figure 10 (PDF p. 149). (a) Normalized in-phase part of the frequency dependent photo current of MAPI and CsFAMA PSCs. Normalized quadrature part of the frequency dependent photo current and phase delay of (b) MAPI and (c) CsFAMA PSC. Normalized quadrature part of the frequency dependent photo current on the active area and outside the active area of (d) MAPI and (e) CsFAMA PSC. (f) Normalized quadrature part of the frequency dependent photo current and phase of CsFAMAPbI3 PSC.

The IPh(f)I_{Ph}(f) measurement has been carried out to understand the dielectric relaxation processes in MAPI vs CsFAMA based PSCs. Figure 7c shows a typical frequency dependent normalized IPh,qdI_{Ph,qd} profile for MAPI and CsFAMA based PSCs. The normalized IPh,inI_{Ph,in} is shown in Figure 10a. The full scale (200 Hz to 102 kHz) of the IPh,qdI_{Ph,qd} along with phase for MAPI and CsFAMA based PSCs is shown in Figure 10b and 10c respectively. The resonance frequency for MAPI based PSC is at 13 kHz whereas, for CsFAMA based PSC, it is at 18.5 kHz.[35] We observe that the FWHM of the Iph,qd(f)I_{ph,qd}(f) profile is narrower for CsFAMA than MAPI. However, we are still looking for the specific reason behind the narrow FWHM of the Iph,qd(f)I_{ph,qd}(f) profile for CsFAMA.

Again, to validate the frequency depending on photocurrent results, we measured the frequency dependent capacitance of MAPI and CsFAMA based PSCs Figure 7d. We observed that the capacitance is lower for CsFAMA than MAPI based PSC which is in good agreement with the results obtained earlier for (FA1−xMAx)0.95Cs0.05PbI3(\mathrm{FA}_{1-x}\mathrm{MA}_{x})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} compositions Figure 1e.

We used Iph,qd(f)I_{ph,qd}(f) response to calculate the dielectric relaxation time constant (τd\tau_d) of the photogenerated charge carriers in PSCs. The τd\tau_d represents the response time in the dielectric material with respect to the externally applied sinusoidal electric field of 490 nm wavelength (electromagnetic) EM wave. Here, we have used the Lorentz peak fitting function to extract τd\tau_d from experimental results Figure 7c.

Iph,qd(f)=Iph,s−∞+2Ωδ4π(f−fmax)2+δ2(34)I_{ph,qd}(f) = I_{ph,s-\infty} + \frac{2\Omega\delta}{4\pi(f-f_{max})^2+\delta^2} \tag*{(34)}

Where Iph,s−∞I_{ph,s-\infty} is the photocurrent at very high or low (static) frequency, Ω=\Omega= area under the plot, δ=\delta= width of the plot, f=f = applied frequency, fmax=f_{max} = frequency at which the Iph,qd(f)I_{ph,qd}(f) is maximum termed as the resonance frequency.

The calculated τd\tau_d for MAPI and CsFAMA is (79.49 ± 0.32) μs and (55.14 ± 0.15) μs respectively. Hence, the relaxation process is faster in CsFAMA based PSC than that in MAPI.

It indicates that any optically perturbed charge carrier in the MAPI based PSC will take more time to come back to its equilibrium position than the CsFAMA based PSC, a representative of higher defect density in the former one. The Iph,qd(f)I_{ph,qd}(f) profile on the device and outside of the active area of the device for MAPI and CsFAMA based PSCs shows a similar resonance peak position (Figure 10d, e). Thus, it indicates that the photocurrent profile is a material property and does not depend on the spatial position of the device which is in good agreement with our earlier publication.[35] It is observed that the Iph,qd(f)I_{ph,qd}(f) profile outside the active area of the device is low compared to the Iph,qd(f)I_{ph,qd}(f) profile on the active area of the device. This can be understood by the lateral photovoltaic effect.[53] We do notice that CsFAMA based perovskite system contains Br⁻ and I⁻ both. In order to validate the effect of ‘A’ cation on Iph,qd(f)I_{ph,qd}(f) profile, we fabricated (FA0.83MA0.17)0.95Cs0.05PbI3(\mathrm{FA}_{0.83}\mathrm{MA}_{0.17})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} based PSC with only I⁻ as halide. The frequency dependent Iph,qd(f)I_{ph,qd}(f) measurement did not show any significant shift in the resonance frequency of (FA0.83MA0.17)0.95Cs0.05PbI3(\mathrm{FA}_{0.83}\mathrm{MA}_{0.17})_{0.95}\mathrm{Cs}_{0.05}\mathrm{PbI}_{3} based PSC in comparison to (FA0.83MA0.17)0.95Cs0.05Pb(I0.90Br0.10)3(\mathrm{FA}_{0.83}\mathrm{MA}_{0.17})_{0.95}\mathrm{Cs}_{0.05}\mathrm{Pb}(\mathrm{I}_{0.90}\mathrm{Br}_{0.10})_{3} based PSC (Figure 10f). However, it is acknowledged that we do not observe any significant shift in resonance frequency because the ratio of Br−\mathrm{Br}^{-} was quite less (10%) in CsFAMA based perovskite. The complete replacement of I−\mathrm{I}^{-} by Br−\mathrm{Br}^{-} may cause shift in the resonance frequency, and it can be a separate study.

Experimental validation of the frequency dependent photocurrent measurement

Resonance frequency determined by a direct frequency dependent photocurrent is correlated with an independent estimation of capacitance plus conductivity studies, which is found to be in good agreement.

The classical dielectric relaxation time can be expressed as[34]

τd=ε0εsσ(35)\tau_d = \frac{\varepsilon_0\varepsilon_s}{\sigma} \tag*{(35)}

The capacitance of a parallel plate capacitor can be expressed as

ε0εs=C⋅dA(36)\varepsilon_0\varepsilon_s = \frac{C \cdot d}{A} \tag*{(36)}

Thus, the dielectric relaxation time constant can be written as

τd=C⋅dσ⋅A(37)\tau_d = \frac{C \cdot d}{\sigma\cdot A} \tag*{(37)}

Where, CC = capacitance, dd = thickness, AA = device area, σ\sigma = conductivity, εs\varepsilon_s = dielectric constant, ε0\varepsilon_0 = dielectric permittivity in vacuum.

The thickness of MAPI is dMAPI=300d_{\mathrm{MAPI}} = 300 nm, and for CsFAMA is dCsFAMA=400d_{\mathrm{CsFAMA}} = 400 nm (Figure 5).

The conductivity of MAPI is σMAPI=1.6×10−5\sigma_{\mathrm{MAPI}} = 1.6 \times10^{-5} S. cm−1^{-1} and for CsFAMA is σCsFAMA=1.7×10−5\sigma_{\mathrm{CsFAMA}} = 1.7 \times10^{-5} S. cm−1^{-1} (Figure 8d (PDF p. 181)).

The capacitance of MAPI is CMAPI∼22C_{\mathrm{MAPI}} \sim22 nF/cm2^{2} and for CsFAMA is CCsFAMA∼12C_{\mathrm{CsFAMA}} \sim12 nF/cm2^{2} (Figure 7d).

The area of both the devices is the same A=17.5A = 17.5 mm2^{2}.

The ratio of dielectric relaxation time constants using equation (4.5) is,

τCsFAMAτMAPI=CCsFAMA⋅dCsFAMAσCsFAMA⋅σMAPICMAPI⋅dMAPIτCsFAMAτMAPI=0.68\begin{aligned} \frac{\tau_{\mathrm{CsFAMA}}}{\tau_{\mathrm{MAPI}}} &= \frac{C_{\mathrm{CsFAMA}} \cdot d_{\mathrm{CsFAMA}}}{\sigma_{\mathrm{CsFAMA}}} \cdot\frac{\sigma_{\mathrm{MAPI}}}{C_{\mathrm{MAPI}} \cdot d_{\mathrm{MAPI}}} \\ \frac{\tau_{\mathrm{CsFAMA}}}{\tau_{\mathrm{MAPI}}} &= 0.68 \end{aligned}

Further, the dielectric relaxation time constants from the frequency-dependent photocurrent measurement are τd,CsFAMA=(55.14±0.15)\tau_{d,\mathrm{CsFAMA}} = (55.14 \pm0.15) μ\mus and τd,MAPI=(79.49±0.32)\tau_{d,\mathrm{MAPI}} = (79.49 \pm0.32) μ\mus and the ratio of the time constant measured from the frequency-dependent photocurrent measurement is

τCsFAMAτMAPI=0.69\frac{\tau_{\mathrm{CsFAMA}}}{\tau_{\mathrm{MAPI}}}=0.69

Thus, the ratio of the time constant measured from two different experiments is almost similar, which validates our measurement that experimentally determined resonance frequency can provide a direct connection with semiconductor material properties.[34]

Measurement techniqueτd,CsFAMA/τd,MAPI\tau_{\mathrm{d,CsFAMA}}/\tau_{\mathrm{d,MAPI}}
Photocurrent Iph,qd(f)I_{\mathrm{ph,qd}}(f)0.69
CC-ff and dark JJ-VV measurement0.68

Table 3. The ratio of dielectric relaxation time measured from photocurrent response Iph,qd(f)I_{\mathrm{ph,qd}}(f) and CC-ff combined with JJ-VV measurement.

Defects in perovskite solar cells

Further, optical studies are carried out to gain deep insight into the defect states in MAPI versus CsFAMA based perovskite thin films. It is observed that the steady-state photoluminescence (PL) intensity of CsFAMA is ~2.38 times higher than that of MAPI based perovskite film (Figure 8a (PDF p. 181)) under identical conditions of excitation and measurement set-up. The higher PL intensity of the CsFAMA film could be due to either of two reasons: (a) higher thickness or (b) lower defect density than MAPI (Figure 5). To validate the higher PL intensity of CsFAMA, the time-resolved (TR) PL measurement is carried out for both films. Figure 11a represents the TRPL spectra for both films. The charge carrier lifetime for MAPI and CsFAMA based perovskite films is 20.36 ns and 49.06 ns respectively. Hence, the average charge carrier lifetime in CsFAMA film is in good agreement with steady-state PL intensity results (Figure 8a (PDF p. 181)). The reason for the higher TRPL lifetime for CsFAMA perovskite is due to higher photoluminescence quantum yield (PLQY). The PLQY of MAPI and CsFAMA perovskite films on the glass substrate is 4.06% and 9.78% respectively. Interestingly the PLQY ratio is 2.41 which correlates with the steady-state PL.[36, 54]

The VOCV_{\mathrm{OC}} of a solar cell can be expressed as [34]

VOC=nkTqln⁡(I)+constantV_{\mathrm{OC}}=\frac{nkT}{q}\ln(I)+\mathit{constant}
VOC=nkTq×2.303×log⁡(I)+constant(38)V_{\mathrm{OC}}=\frac{nkT}{q}\times2.303\times\log(I)+\mathit{constant} \tag*{(38)}

Where nn = ideality factor kk = Boltzmann constant, TT = temperature, qq = charge, II = intensity of the light.

Image not extracted

Figure 11 (PDF p. 153). (a) Time-resolved photoluminescence (TRPL) decay profile of MAPI and CsFAMA perovskite films, solid line is bi-exponential fit to the experimental data. (b) Intensity dependent VOCV_{\mathrm{OC}} of MAPI and CsFAMA PSCs, the solid line fits the Shockley diode equation.[55] (c) Transient photovoltage (TPV) decay profile of MAPI and CsFAMA PSCs under 1-Sun illumination condition. (d) Electroluminescence (EL) quantum efficiency as a function of injected current density for MAPI and CsFAMA PSCs, and the inset represents the normalized EL spectra of MAPI and CsFAMA PSCs at an injected current of 5 mA.

Figure 11b represents the intensity dependent VOCV_{\mathrm{OC}} of MAPI and CsFAMA PSCs. To calculate the ideality factor, we measured the slope of VOCV_{\mathrm{OC}} vs. log⁡(I)\log(I) in Figure 11(b) and used equation (38) to calculate the ideality factor. It is observed that the ideality factor (nn) for MAPI is n=1.61n=1.61 whereas for CsFAMA n=1.10n=1.10. The higher value of nn for MAPI indicates the presence of a higher number of trap-assisted recombination centers.[56, 57] Figure 11c represents the transient photovoltage (TPV) response for MAPI and CsFAMA based PSCs.[58] The calculated lifetime (τ\tau) of perturbed charge carriers for MAPI and CsFAMA are 0.33 μ\mus and 0.58 μ\mus respectively. The higher value of τ\tau for CsFAMA represents lower defect density and is responsible for higher VOCV_{\mathrm{OC}} despite of having the same bandgap of semiconductor (**Figure 6b).[59] Figure 11d represents the electroluminescence (EL) quantum efficiency (ELQE) of MAPI and CsFAMA PSCs. The inset of Figure 11d represents the EL spectra for MAPI and CsFAMA at 5 mA constant current injection. It is observed that EL intensity is ∼ 2.49 times higher in the case of CsFAMA than MAPI based PSCs (Figure 11d). Further, we found that the ELQE is dependent on injected current density up to Jinj=20J_{\mathrm{inj}} = 20 mA/cm2^{2} in the case of MAPI based PSCs whereas it is largely independent beyond Jinj∼5J_{\mathrm{inj}} \sim5 mA/cm2^{2} for CsFAMA PSCs.[60]

Thermal stability of the perovskite solar cells

Figure 12a represents the histogram of the MAPI and CsFAMA based PSCs over 24 devices. It is observed that the average PCE of MAPI based PSCs is 14.69%, whereas for CsFAMA based PSC the PCE is 18.83%. So, replacing the MAPI with CsFAMA perovskite, there is an increase in PCE of ∼ 28% even though the optical bandgap edge is the same. Further, the thermal stability of MAPI and CsFAMA based PSCs is investigated (Figure 12). The unencapsulated devices were kept under the continuous 1-Sun illumination for 14 hours, and the humidity was ∼ 40%. The illuminated JJ-VV characteristic is measured for both devices at regular time intervals. It is observed that MAPI based PSC retains 47.11% of its initial JSCJ_{\mathrm{SC}}, 105.56% of its initial VOCV_{\mathrm{OC}}, 97.34% of its initial FFFF, and 49.28% of its initial PCE after 14 hours of continuous 1-Sun illumination Figure 12b. However, the CsFAMA based PSC retains 85.89% of its initial JSCJ_{\mathrm{SC}}, 97.20% of its initial VOCV_{\mathrm{OC}}, 94.17% of its initial FFFF, and 78.61% of its initial PCE after 14 hours of continuous 1-Sun illumination Figure 12c. The retained PV parameters after 14 hours of continuous 1-Sun illumination are listed in Table 4. Thus, continuous illumination of the PSCs under the 1-Sun condition deteriorates the performance of the PSCs. The major change in the PCE of MAPI and CsFAMA based PSCs after 14 hours of continuous 1-Sun illumination is associated with the JSCJ_{\mathrm{SC}}. The JSCJ_{\mathrm{SC}} decreases for both MAPI and CsFAMA PSCs with time under 1-Sun continuous illumination, however, it is higher for MAPI based PSC. Reduction in JSCJ_{\mathrm{SC}} and a slight increase in VOCV_{\mathrm{OC}} under continuous illumination for MAPI based PSCs have been observed earlier too by Joshi et al.[61] MAPI based PSC showed the VOCV_{\mathrm{OC}} starts increasing after 6 hours of illumination and eventually, after 14 hours of illumination, it 105.56% of its initial VOCV_{\mathrm{OC}}. Whereas for CsFAMA, it is continuously decreasing, and eventually after 14 hours it remains 97.20 % of its initial VOCV_{\mathrm{OC}}. Even though the VOCV_{\mathrm{OC}} of MAPI based PSC is increasing with time during continuous illumination but the JSCJ_{\mathrm{SC}} is degrading very significantly. Hence, the overall PCE remains lower for MAPI based PSC after degradation.

Image not extracted

Figure 12 (PDF p. 155). (a) Histogram of the PCE of MAPI and CsFAMA based PSCs over 24 devices. (b) Thermal stability of PSCs under the continuous 1-Sun illumination for 14 hours of (a) MAPI and (c) CsFAMA based PSCs.

DeviceJSCJ_{\mathrm{SC}} (%)VOCV_{\mathrm{OC}} (%)FFFF (%)PCE (%)
MAPI47.11105.5697.3449.28
CsFAMA85.8997.2094.1778.61

Table 4. The kinetics of PV parameters in percentage (%) of its initial value (i.e. 100%) for MAPI and CsFAMA based unencapsulated PSCs after 14 hours of continuous 1-Sun illumination.

Discussion

Figure 1d suggests the role of FA vs. MA on dielectric resonance frequency while keeping absorber layer thickness fixed at ~ 500 nm. However, the device structure is not optimized for optimal PCE. We note that there is not a linear relationship or a correlation between the PCE vs. resonance frequency based on figure 1d and Table 1. This can be attributed to the multiple factors which govern the solar cell efficiency, like; compactness, i.e., the morphology of the thin-film, crystallinity and domain size, thickness optimization of the absorber, and electronic coupling with charge extracting layers.[7, 62] However, for an optimized PSC structure, dielectric relaxation studies consistently explain the influence on resonance frequency due to the presence of defects which are ions. Herein, we show a unique set of results which correlates the τd\tau_d with the defects present in MAPI and CsFAMA PSCs. Further, in semiconductor physics, it is widely known that the dielectric relaxation time constant (τd\tau_d) is proportional to the resistivity (ρ\rho) and dielectric permittivity (εs\varepsilon_s) i.e. τd=ρεs\tau_d = \rho\varepsilon_s.[34] Two independent methods are used to determine the ratio of τd,CsFAMA/τd,MAPI\tau_{d,\mathrm{CsFAMA}}/\tau_{d,\mathrm{MAPI}}, Iph,qd(f)I_{ph,qd}(f) and CC-ff combined with conductivity, which is found to be almost similar 0.69 and 0.68 respectively Table 3. This phenomenon can be understood by considering the dipoles oscillating in an external electric field.[40] When the 1/f≫τd1/f \gg\tau_d, the dipoles have enough time to respond to the electric field, and thus it shows some positive number of current. As we further increase the frequency and 1/f1/f becomes comparable to τd\tau_d, the resonance condition occurs and Iph,qd(f)I_{ph,qd}(f) attains a maximum value. Further, an increase in frequency results in a decrease in Iph,qd(f)I_{ph,qd}(f) (relaxation process) due to dipole lag behind the electric field. Perovskite materials are known to show ferroelectric behavior.[32] It is observed that ferroelectric dipoles of CsFAMA show a better response to an electric field than MAPI (Figure 7c). Hence, the relaxation process is also faster in the CsFAMA (τd=55.14±0.15 μs\tau_d = 55.14 \pm0.15\ \mu\mathrm{s}) than MAPI (τd=79.49±0.32 μs\tau_d = 79.49 \pm0.32\ \mu\mathrm{s}). This indicates that if an additional charge is perturbed in the system, it will come back to the ground state faster in CsFAMA than MAPI based PSCs. The faster relaxation process in CsFAMA is attributed to lower defect density, which is in good agreement with other optoelectronic measurements. The higher steady-state PL intensity (Figure 8a (PDF p. 181)), higher charge carrier lifetime (TRPL and TPV) (Figure 11a, c), and the injection current density independent higher ELQE_\mathrm{QE} behavior (Figure 11d) is evidence of lower defect density in CsFAMA based perovskite film.

When light interacts with absorbing semiconductor layers at different modulated frequencies, like 1/f≫τd1/f \gg\tau_d, then the dipoles have enough time to respond to the alternating electric field. We note that any perturbed charge carrier in the MAPI-based PSC takes a longer time to come back into its equilibrium position than the CsFAMA-based PSC. This suggests that at the microscopic level dipoles are interacting with defects, which are relatively hindering the dipole oscillations in MAPI based devices. Thermal stability studies suggested that there is a significant decrease in the JSCJ_{SC} which can be attributed to prominent ion migration in MAPI based PSCs than CsFAMA based PSCs as studied by Joshi et. al. using capacitance studies as a function of illumination dose and irradiation time duration.[61, 63] These two combined studies connect well macroscopic dielectric parameters to microscopic level properties of the semiconductor film in the form of the presence of relative ionic defects density. A significant decrease in JSCJ_{SC} and a slight increase in VOCV_{OC} is an established key feature of ion migration.[61, 64] Under constant illumination, a relative enhancement of VOCV_{OC} in MAPI based PSC can be understood by the presence of relatively more ionic defects. It has also been established previously by theoretical and experimental studies.[61, 64] Based on DFT calculation it is proven that due to presence of only 3 H-bonds from A-site cation with octahedral I−^{-} can cause the release of I−^{-} ions.[65] Whereas these H-bonding sites can be increased with the addition of FA+^{+} ion and this results in relatively better control over ion migration issues in pristine halide perovskites, i.e., CsFAMA studied here.[66] Hence, we believe that the low ionic/structural defect density in CsFAMA bulk due to the presence of relatively more number of H-bonding between cation to halide ions is key for stable absorber and hence PSCs than conventional MAPI based PSCs.

Conclusions

In conclusion, we have presented a comparative study on the dielectric relaxation process in MAPI vs CsFAMA in terms of Iph(f)I_{ph}(f) measurement. The Iph(f)I_{ph}(f) measurements allow us to determine τd\tau_d, and it is lower for CsFAMA than MAPI, which suggests the lower defect density in the former one. We also compared the defects in MAPI and CsFAMA by various optoelectronic measurements and observed that CsFAMA possesses lower defects than MAPI. We validated our measured τd\tau_d from Iph(f)I_{ph}(f) measurement and τd=ρε\tau_d = \rho\varepsilon and it is in good agreement. The longer frequency response of CsFAMA (lower τd\tau_d) dipoles to the electric field in comparison to the MAPI is attributed to lower defect density in the CsFAMA bulk. We also showed that low τd\tau_d in multi-cation CsFAMA is connected for stable absorber under illumination conditions and can be used as a tool to screen other ionic semiconductors for stable solar cell application.

Postscript

Overall, this chapter deals with the compositional engineering of hybrid organic-inorganic halide perovskite semiconductors and introduces the study of defects using various optoelectronic measurements. From the first principle calculation studies, it is shown that the intrinsic defects in the perovskite can be minimized by using multications in the perovskite crystal structure (discussed in Chapter 2). In the literature reports, the single monovalent cation perovskite MAPbI3_{3} has higher defects and lower stability compared to the FAPbI3_{3} or triple cation CsFAMA based perovskite, but the direct correlation was not established. In this chapter, we used frequency dependent photocurrent measurement technique and studied the dielectric relaxation process. The faster relaxation time for the CsFAMA based PSCs indicates lower defects, which are later verified using various optoelectronic measurements. Further, the CsFAMA-based perovskite solar cells show higher thermal stability, which is correlated with the halide ion migration and hydrogen bonding with the monovalent cation. More hydrogen bonding in the CsFAMA perovskite crystal structure is responsible for faster dielectric relaxation and higher thermal stability, hence the correlation established. However, we demonstrated the thermal stability studies under continuous 1-Sun illumination and relative humity of 40%, of our unencapsulated PSCs. The stability can be increased by encapsulating the device and carrying the stability measurement under N2N_{2} atmosphere. In addition, the reproducible efficiency and thermal stability of the PSCs can be further improved by choosing a self-assembled monolayer (SAM) based HTLs that facilitates good energy band alignment with the perovskite layer of bandgap 1.6 eV Figure 13.[67] We used a phosphonic acid anchored carbazol SAM based HTL and thermally stable triple cation CsFAMA perovskite (as discussed above) in the inverted architecture-based solar cells. The work function of the SAM is modified using a polyelectrolyte polymer and well matched with the perovskite work function, resulting in excellent interfacial charge transport quality. This work will be discussed in next Chapter 5.

Image not extracted

Figure 13 (PDF p. 159). Energy level band diagram of the pp-ii-nn architecture based perovskite solar cell where PTAA and Me-4PACz SAM are used as HTLs and CsFAMA perovskite (bandgap 1.6 eV) as an absorber. The energy level values are taken from the literature.[68, 69, 70]

References

Chapter 4: Role of Monovalent Cation …

Resolving the Hydrophobicity of Me-4PACz Hole Transport Layer for Highly Reproducible and Efficient (>20%) Inverted Perovskite Solar Cells

CHAPTER 5

Resolving the Hydrophobicity of Me-4PACz Hole Transport Layer for Highly Reproducible and Efficient (>20%) Inverted Perovskite Solar Cells

[4-(3,6-Dimethyl-9H-carbazol-9-yl)butyl]phosphonic acid (Me-4PACz) self-assembled monolayer (SAM) has been employed in perovskite single junction and tandem devices demonstrating high efficiencies. However, a uniform perovskite layer does not form due to the hydrophobicity of Me-4PACz. Here, we tackle this challenge by adding a conjugated polyelectrolyte poly(9,9-bis(3’-(N,N-dimethyl)-N-ethylammonium-propyl-2,7-fluorene)-alt-2,7-(9,9-dioctylfluorene))dibromide (PFN-Br) to the Me-4PACz in a specific ratio, defined as Pz:PFN. With this mixing engineering strategy of Pz:PFN, the PFN-Br interacts with the A-site cation is confirmed via solution-state nuclear magnetic resonance studies. The narrow full width at half maxima (FWHM) of diffraction peaks of perovskite film revealed improved crystallization on the optimal mixing ratio of Pz:PFN. Interestingly, the mixing of PFN-Br additionally tunes the work function of the Me-4PACz, as demonstrated by Kelvin probe force microscopy and built-in-voltage estimation in solar cells. Devices employing optimized Pz:PFN mixing ratio deliver open-circuit voltage (VOCV_{\mathrm{OC}}) of 1.16V and efficiency >20% for perovskites with a bandgap of 1.6 eV with high reproducibility and concomitant stability. Considering significant research on Me-4PACz SAM, our work highlights the importance of obtaining a uniform perovskite layer with improved yield and performance.

Introduction

In the previous chapter 4, we have demonstrated that the compositional engineering of the organic cations in the hybrid organic–inorganic metal halide perovskite is helpful in minimizing the defect states. The mixing the monovalent cations results in improved device performance and concomitant stability. However, in Chapter 4, we used PTAA as a hole transport layer (HTL). PTAA is hydrophobic in nature, and hence deposition of a uniform perovskite layer is challenging. Therefore, we used PFN-Br as an interlayer to deposit a uniform perovskite layer, which restrains direct contact of perovskite with the PTAA layer. In addition, the PTAA shows a band offset by 200–300 meV for ~1.6 eV perovskite, which further hampers charge collection efficiently.5 Hence, we used another HTL of self-assembled monolayer (SAM) in the p-i-n device architecture of perovskite solar cells (PSCs). So, in this chapter 5, we used the mixed cation perovskite CsFAMA as an absorber due to its enhanced performance and stability, as discussed in Chapter 4, and PCBM as an electron transport layer (ETL) in the p-i-n device architecture. The inverted or p-i-n device architecture is of significant interest using low-temperature processed charge transport layers (CTLs) such as poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT: PSS), poly (triaryl amine) (PTAA), fullerene (C60\mathrm{C}_{60}) are employed.[2, 3, 4, 5, 6] Recently, Al-Ashouri et al. used various phosphonic acid group anchored carbazole-based self-assembled monolayers (SAMs) such as 2PACz, MeO-2PACz, and Me-4PACz as hole transport layers (HTL), demonstrating superior device performance than the widely employed PTAA as HTL.[1, 7, 8] In particular, by employing Me-4PACz SAM as HTL, the same group, reported one of the highest efficiency in single junction devices and a certified PCE of ~29% in silicon-perovskite two-terminal tandem solar cells.[1] In comparison to other SAMs, the Me-4PACz-based device is reported to demonstrate the lowest density of interface traps, indicating suppressed non-radiative recombination at the perovskite/Me-4PACz interface.[9] These salient features make Me-4PACz SAM a promising HTL for PSCs. Motivated by this, we fabricate PSCs with Me-4PACz as HTL. Unfortunately, we (and others) observed poor perovskite layer formation on the Me-4PACz and is in line with recent reports.[10, 11, 12, 13] Strategies such as the incorporation of an Al2O3\mathrm{Al}_{2}\mathrm{O}_{3} insulating layer at the interface of Me-4PACz and perovskites and deposition of Me-4PACz by evaporation have been reported to improve the perovskite layer coverage on Me-4PACz.[10, 11, 14] These strategies are promising, but developing a strategy to deposit Me-4PACz by solution process and simultaneously obtain a uniform perovskite layer with improved device performance with a proper understanding remains a challenge.

Until recently, the interface engineering approach has been applied to address the issue of poor wetting of perovskite on underlying hydrophobic polymeric HTLs. For instance, PTAA is known to be a hydrophobic HTL and modification of its surface with poly(9,9-bis(3’-(N,N-dimethyl)-N-ethylammoinium-propyl-2,7-fluorene)-alt-2,7-(9,9dioctylfluorene))dibromide (PFN-Br), poly(methyl methacrylate) (PMMA):2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane (F4-TCNQ), polyvinyl oxide (PEO) and tetra-n-propylammonium bromide (TPAB), phenylethylammonium iodide (PEAI), etc. have been reported.[5, 15, 16, 17, 18, 19] Among others, modification with PFN-Br has been widely employed as it helps in obtaining a uniform perovskite layer with augmented device efficiency. In the previous chapter 4, we have also employed PFN-Br modified PTAA and reported PCE of >20%.[2] In addition to polymeric HTLs, modification of SAMs has also been reported.[20, 21] Recently, Li et al., employed a mixture of MeO-2PACz and 2PACz to improve the charge extraction and reported a PCE of 25% for all-perovskite tandem solar cells.[21] Deng et al., showed that co-assembled monolayers (a mixture of SAM and alkylammonium containing SAM) can help in simultaneously suppressing the non-radiative recombination and surface functionalization and the resultant device showed a PCE of 23.59%.[20] Very recently, Al-Ashouri added 1,6-hexylenediphosphonic acid with Me-4PACz to improve the perovskite layer deposition; however, a detailed investigation has not been performed.[13] Until present, the modification of carbazole-based SAMs has been limited to co-assemble with phosphonic or carboxylic acid anchoring group which might limit the exploration of SAM. Therefore, strategies need to be developed beyond co-assembly with SAM, to improve the perovskite layer coverage on Me-4PACz SAM.

In this chapter, we present a mixing engineering strategy by combining Me-4PACz SAM with conjugated polyelectrolyte PFN-Br to obtain a highly reproducible and improved performance of perovskite solar cells. With the aid of mixed Me-4PACz:PFN-Br, a uniform perovskite thin film is obtained and the device with optimized Me-4PACz:PFN-Br mixing ratio showed a PCE of >20% in 0.175 cm2^2 device active area and 19% over an active area of 0.805 cm2^2. To the best of our knowledge, the obtained device efficiency is one of the highest values reported for Me-4PACz with a triple cation perovskite composition having a bandgap (EgE_g) of 1.6 eV. More importantly, with the help of solution nuclear magnetic resonance (NMR), X-ray diffraction (XRD) and Kelvin probe force microscopy (KPFM) outcomes, we elucidate the triple role of PFN-Br. The PFN-Br interacts with the A-site cation and improves the crystallization. This also elevates the valence band position of Me-4PACz, leading to better interfacial energy level alignment with perovskite having a bandgap of 1.6 eV. We note that this improved performance could be achieved due to reduced interfacial localized states (low series resistance (RSR_{\mathrm{S}})) and bulk defects in the absorber layer (ideality factor n∼1n \sim1), which result not only in high fill factor (FFFF), short-circuit current density (JSCJ_{\mathrm{SC}}) and open-circuit voltage (VOCV_{\mathrm{OC}}) but also provide an excellent yield and stability for optimal PSCs. The unencapsulated device showed enhanced stability by retaining the initial device performance at T95_{95} even after 3000 hours when measured in ~ 40% humidity. Further, the operational stability at the 1-Sun condition and thermal stability at 85∘85^\circ C of the un-encapsulated device under N2_2 environment retains > 90% of the initial efficiency after 300 and 600 hours respectively.

Experimental Section

Materials: ITO-coated glass substrates (15 Ω\Omega/sq) were purchased from Lumtech. Lead iodide (PbI2\mathrm{PbI}_{2}), formamidinium iodide (FAI), cesium iodide (CsI), lead bromide (PbBr2\mathrm{PbBr}_{2}), and Me-4PACz, all were purchased from TCI chemical and used as received. Methyl ammonium Bromide (MABr) was purchased from Greatcell Solar. Phenyl-C61-butyric acid methyl ester (PC61_{61}BM) was purchased from Lumtech and was used as received. The poly(9,9-bis(3’-(N,N-dimethyl)-N-ethylammoinium-propyl-2,7-fluorene)-alt-2,7-(9,9-dioctylfluorene))dibromide (PFN-Br) ordered from Solarmer Material Inc. and used as received. Bathocuproine (BCP) was purchased from Sigma Aldrich and used as received. The list of all chemicals, companies, CAS, and product numbers is as follows.

ChemicalCompanyCAS numberProduct number
MABrGreat cell Solar6876-37-5MS301000
PbBr2_{2}TCI10031-22-8L0288
FAITCI879643-71-7F0974
PbI2_{2}TCI10101-63-0L0279
CsITCI7789-17-5C2205
Me-4PACzTCI2747959-96-0M3359

Table 7.

PFN-BrSolarmer Material Inc.889672-39-6ZJ710A
PCBMLumtec.160848-22-6LT-S905
BCPSigma4733-39-5699152
DMFSigma68-12-2227056
DMSOSigma67-68-5276855
MethanolSigma67-56-1322415
ChlorobenzeneSigma108-90-7284513
IPASigma67-63-019516
AgLocal company – Parekh Industry LimitedSilver wire – 1 mm diameterPurity 99.99%
ITO coated glassLumtec.ITO Non-Patterned 15Ω15\OmegaLT – G001
DMSO d6ACROS Organic2206-27-1166291000

Table 8.

Solution preparation: To make triple cation (FA0.83MA0.17)0.95Cs0.05Pb(I0.83Br0.17)3(\mathrm{FA}_{0.83}\mathrm{MA}_{0.17})_{0.95}\mathrm{Cs}_{0.05}\mathrm{Pb}(\mathrm{I}_{0.83}\mathrm{Br}_{0.17})_{3} perovskite solution, first we added 22.5 mg of MABr, 73.5 mg of PbBr2_2, 172 mg of FAI, and 507.5 mg of PbI2_2 in 1 ml of DMF: DMSO (4:1) and stirred at room temperature for 2 hours to make a premixed solution. Separately we made 1.5 (M) CsI solution in DMSO i.e. 100 mg of CsI in 257 μ\mul of DMSO and stirred for 2 hours. We filtered the premixed solution with a PTFE 45 mm filter in a separate vial. Finally, we added 950 μ\mul of premixed solution and 50 μ\mul of CsI solution to get the final triple cation(FA0.83MA0.17)0.95Cs0.05Pb(I0.83Br0.17)3(\mathrm{FA}_{0.83}\mathrm{MA}_{0.17})_{0.95}\mathrm{Cs}_{0.05}\mathrm{Pb}(\mathrm{I}_{0.83}\mathrm{Br}_{0.17})_{3} perovskite solution and stirred for 1 hour. To obtain mixed Me–4PACz: PFN-Br solution, first, we prepared 0.4 mg/ml Me–4PACz solution in anhydrous methanol and 0.4 mg /ml PFN-Br solution in anhydrous methanol and stirred overnight. Finally, one hour before spin-coating, we mix the Me–4PACz and PFN-Br solution in the desired volume ratio e.g. 6:4, 7:3, 8:2, 9:1, and 9.5:0.5.

20 mg PCBM is dissolved in 1 ml of chlorobenzene and stirred overnight. 0.5 mg BCP is dissolved in 1 ml of anhydrous isopropanol and stirred overnight at room temperature and 10 minutes at 70∘70^{\circ} C just before spin coating.

Perovskite solar cells fabrication: ITO-coated glass (10-15 Ω\Omega/square) substrate was patterned with Zn powder and HCl and then sequentially cleaned with soap solution, deionized (DI) water, acetone, and isopropanol for 10 minutes each. After drying the substrate with a nitrogen gun, we kept the ITO substrates at 80∘80^{\circ} C for 10 minutes and then took them inside the oxygen plasma ashing chamber for 20 minutes, and plasma ashing was done at an RF power of 18 watts. After plasma ashing, we immediately take the substrates inside the N2_2 environment glove box (O2_2<0.1 ppm, H2_2O<0.1 ppm) and spin-coat Me-4PACz: PFN-Br mixed solution at 4000 rpm for 30 seconds and then annealed at 100∘100^{\circ} C for 10 minutes. After that, we cool down the substrates for 5 minutes and then perovskite solution spin-coated on the ITO/ Me-4PACz: PFN-Br substrates. The CsFAMA perovskite was spin-coated at 5000 rpm for 30 seconds and at the last 5 seconds, we used 150 μ\mul of chlorobenzene as an anti-solvent treatment on a 1.5 cm by 1.5 cm substrate and then annealed at 100∘100^{\circ} C for 30 minutes. After that, PC61_{61}BM was spin-coated at 2000 rpm for 30 seconds and BCP was spin-coated at 5000 rpm for 20 seconds. Finally, 150 nm of Ag was deposited under a vacuum of 2×10−62 \times10^{-6} mbar using a metal shadow mask. Further, we fabricated PSC of the active area of 0.805 cm2^2 (i.e., 11.5 mm ×\times 7 mm) using ITO-coated glass substrates of sheet resistance 7 Ω\Omega/square.

Characterization:

All the photovoltaic measurements were carried out under ambient conditions. Photocurrent density versus applied voltage (J−VJ - V) measurement has been carried out using Keithley 4200 SCS and LED solar simulator (LSH-7320) after calibrating through standard Si solar cells provided by ABET, IIT Bombay. The J−VJ - V measurement was performed with a scan rate of 100 mV/s. EQE measurement has been carried out to measure the photo response as a function of wavelength using the Bentham quantum efficiency measurement system (Bentham PVE 300). The XRD measurements were carried out in Smartlab, Rigaku diffractometer with Cu Kα\alpha radiation (λ\lambda=1.54Å). Θ\Theta-2Θ\Theta scan has been carried out from 10∘10^{\circ}-45∘45^{\circ} with a step size of 0.001∘0.001^{\circ}. Morphological analysis was done using field emission scanning electron microscopy (FESEM). Optical absorption spectra were carried out using a spectrometer (PerkinElmer LAMBDA 950). Steady-state PL measurement was done on thin films in a vacuum at a pressure of 10−310^{-3} mbar in a custom-made chamber with a diode laser of wavelength 490 nm. Steady-state current-voltage–light characteristics were measured using a Keithley 2400 source meter, Keithley 2000 multimeter, and calibrated Si photodiode (RS components). TPV was measured by using a 490 nm TOPTICA diode laser, THORLABS white lamp S/N M00304198, ArbStudio 1104, and digital oscilloscope Tektronix DPO 4104B. The contact angle measurement is done in the GBX Digidrop instrument. The KPFM measurement was carried out in Asylum/Oxford instrument, MFP3D origin. KPFM images were also recorded in the dark using the same instrument in dual-pass mode under ambient conditions. Voltage was applied to the cantilever tip. The NMR spectra measurement was carried out in JEOL, ECZR series 600 MHz NMR spectrometer.

Results and Discussion

Hydrophobicity of the HTL and perovskite layer deposition

Me-4PACz has been used as HTL in a single junction and tandem (with silicon) devices, demonstrating one of the highest efficiencies.[1, 9, 22] In an attempt to fabricate the device using Me-4PACz SAM, we observed poor perovskite layer formation. The schematic illustration of perovskite deposition on the Me-4PACz coated ITO substrate and the photographic image of a perovskite thin film is shown in Figure 1a. The poor perovskite layer coverage can be attributed to the presence of non-polar groups such as methyl (−CH3\mathrm{-CH}_{3}) and long-alkyl chain (C4H8\mathrm{C}_{4}\mathrm{H}_{8}) in Me-4PACz SAM, which are in general responsible for hydrophobic nature.[23, 24, 25, 26, 27]

Image not extracted

Figure 1 (PDF p. 175). Schematic of the thin film deposition of the Me-4PACz, Me:4PACz:PFN-Br (abbreviated as Pz:PFN), and perovskite layer. (a) Me-4PACz and perovskite layer deposition. (b) Deposition of mixed Me-4PACz:PFN-Br and perovskite layers.

To understand this more clearly, we performed water contact angle measurements of Me-4PACz and compared it with widely used hydrophobic polymer and small organic molecules based HTLs such PTAA, poly[2,6-(4,4-bis-(2-ethylhexyl)-4H-cyclopenta [2 1-b, 3 4-b′]dithiophene)-alt-4,7(2,1,3-benzothiadiazole)] (PCPDTBT), poly(3-hexylthiophene-2,5-diyl) (P3HT) and 4,4′,4″-Tris[phenyl(m-tolyl)amino]triphenylamine (MTDATA). Figure 2 shows the molecular structure of the aforementioned HTLs and their respective water contact angle in comparison with Me-4PACz. As expected, the water contact angle for Me-4PACz is high and in the same range as compared to other HTLs, evidencing its hydrophobic nature.

Image not extracted

Figure 2 (PDF p. 176).

Image not extracted

Figure 2 (PDF p. 177). Molecular structure and corresponding contact angle of (a) Poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine (PTAA), (b) poly[2,6-(4,4-bis-(2-ethylhexyl)-4H-cyclopenta [2 1-b, 3 4-b′]dithiophene)-alt-4,7(2,1,3-benzothiadiazole)] (PCPDTBT), (c) poly(3-hexylthiophene-2,5-diyl) (P3HT) (d) 4,4′,4″-Tris[phenyl(m-tolyl)amino]triphenylamine (MTDATA), (e) [4-(3,6-Dimethyl-9H-carbazol-9-yl)butyl]phosphonic Acid (Me-4PACz), (f) [2-(9H-Carbazol-9-yl)ethyl]phosphonic Acid (2PACz) and (g) [2-(3,6-Dimethoxy-9H-carbazol-9-yl)ethyl]phosphonic Acid (MeO-2PACz)

We additionally washed the Me-4PACz coated substrate with methanol (solvent to dissolve SAM), in an attempt to reduce the hydrophobicity and obtain a uniform perovskite film. However, the perovskite layer showed non-uniform coverage, as shown in Figure 3. This indicates that the phosphonic acid group strongly binds with ITO, and a strong monolayer fingerprint is present even after the washing step, which is in line with previous reports.[7]

Image not extracted

Figure 3 (PDF p. 177). Photographic image showing non-uniform perovskite thin film being deposited on methanol washed ITO/Me-4PACz substrate.

To further confirm the role of non-polar groups that are present in Me-4PACz in preventing the perovskite layer formation, we fabricated device by employing 2PACz SAM as HTL. Because of the absence of the –CH3_3 group and long alkyl chain, not only a uniform perovskite layer was obtained but the resultant device showed a PCE of ~ 20%, this further confirms that other than the SAM all other layers of the device stack are working as expected Figure 4. The current density (JJ) – voltage (VV) curve of the best-performing device employing 2PACz SAM is shown in Figure 4. These results evidently suggest that Me-4PACz is sufficiently hydrophobic to prevent the formation of a uniform perovskite layer, and strategies need to be developed to improve the perovskite layer coverage on the Me-4PACz SAM.

Image not extracted

Figure 4 (PDF p. 178). (a) molecular structure of 2PACz hole transport material (HTL). (b) J−VJ - V characteristic curve of perovskite device employing 2PACz as HTL

Scan directionJSCJ_{\mathrm{SC}} (mA/cm2^2)VOCV_{\mathrm{OC}} (V)FFFF (%)PCE (%)
Forward21.75113780.3119.86
Reverse21.78113378.6219.40

Table 1 (PDF p. 178). Photovoltaic device parameters of perovskite device employing 2PACz as HTL.

Typically, researchers have incorporated PFN-Br as an interlayer to modify the interface of hydrophobic polymeric HTLs and obtain a uniform perovskite layer.[15, 28] We initially modified the surface of Me-4PACz with PFN-Br and obtained a uniform perovskite layer Figure 5.

However, the device performance was not significantly high and reproducible and showed a PCE of 15.54% and 16.52% in forward and reverse scans. The J−VJ - V curve and box plot of device parameters show the degree of non-reproducibility, as shown in Figure 5 and Figure 6 respectively. We initially assumed that the PFN-Br solution (in methanol) might corrode the Me-4PACz underlayer, leading to low device efficiency. However, the washing step of Me-4PACz with methanol solvent (as discussed earlier) suggests that Me-4PACz/PFN-Br (abbreviated as Pz/PFN) dual layer might not be suitable to achieve high device efficiency, unlike PTAA/PFN-Br as discussed in chapter 4.[29]

Image not extracted

Figure 5 (PDF p. 179). (a) PFN-Br interlayer deposition in between Me-4PACz and perovskite layers. (b) J−VJ - V characteristic curves of the device with Me-4PACz and PFN-Br dual layer.

Scan directionJSCJ_{SC} (mA/cm2^2)VOCV_{OC} (V)FFFF (%)PCE (%)
Forward21.84113062.9715.54
Reverse21.75112267.716.52

Table 2 (PDF p. 179). Photovoltaic device parameters of perovskite device employing Me-4PACz and PFN-Br dual layer.

Image not extracted

Figure 6 (PDF p. 180). Histogram of the PV parameters over 18 devices employing Me-4PACz and PFN-Br dual layer.

Previously, Levine et al. reported the lowest density of interface traps for Me-4PACz SAM.[9] This implies that high device performance can be obtained when perovskite is directly interfaced with Me-4PACz. However, in the Me-4PACz/PFN-Br interlayer modification, the perovskite is not directly interfaced with Me-4PACz SAM, which could be the reason for low device efficiency. Therefore, to make a direct interface between the Me-4PACz and perovskite, we performed a mixing engineering strategy.

Resolving the hydrophobicity and mixing engineering strategy

The mixing engineering strategy steps involving device stack layers deposition are schematically shown in Figure 1b. The Me-4PACz and PFN-Br were mixed in 6:4, 7:3, 8:2, 9:1, and 9.5:0.5 ratios followed by the perovskite layer deposition by one-step method (please see the experimental section above for more details). From now onwards, for our convenience, we term Me-4PACz:PFN-Br as Pz:PFN. Irrespective of all the mixing ratios, a uniform perovskite layer was obtained as shown in Figure 1b and Figure 7. Top surface scanning electron microscope (SEM) images, see Figure 8, showed no significant difference in the perovskite layer morphology with respect to the different Pz:PFN ratios.

Image not extracted

Figure 7 (PDF p. 181). Photographic image of perovskite films deposited on Pz:PFN mixed HTLs with different mixing ratios such as, Pz:PFN (6:4), Pz:PFN (7:3), Pz:PFN (8:2), Pz:PFN (9:1), and Pz:PFN (9.5:0.5).

Image not extracted

(a)

Image not extracted

(b)

Image not extracted

(c)

Image not extracted

(d)

Image not extracted

(e)

Image not extracted

(f)

Image not extracted

(g)

Figure 8 (PDF p. 181).

Image not extracted

Figure 8 (PDF p. 182). Top surface SEM images of perovskite thin film deposited on (a) Pz_z/PFN interlayer, (b) to (f) at different mixing ratios of Pz_z:PFN, (g) Me-4PACz.

Moreover, we observed indistinct perovskite layer morphology on both the Me-4PACz/PFN-Br (term as Pz/PFN) Figure 8a (PDF p. 181) dual layer and Me-4PACz (where perovskite layer is deposited on the substrate Figure 1a) (Figure 8g (PDF p. 181)) compared to the mixing ratio cases. This further indicates that the perovskite layer crystallization is not influenced by the PFN-Br mixing or interlayer modification. The X-ray diffraction (XRD) diffractograms, as shown in Figure 9, showed no traces of residual lead iodide and no significant difference in the perovskite crystalline layer. The combined outcomes of SEM and XRD indicate that perovskite crystallizes similarly in all cases without any significant changes.

Image not extracted

Figure 9 (PDF p. 182). XRD diffractograms of perovskite thin films deposited on Me-4PACz (without and with different ratios of PFN-Br) coated ITO substrates.. Me-4PACz and PFN-Br are referred to as Pz_z and PFN respectively. The XRD peaks due to the indium tin oxide layer of ITO-coated glass substrates are marked with a star (*).

After the perovskite layer deposition, phenyl-C61_{61}-butyric acid methyl ester (PCBM) and bathocuproine (BCP) were deposited via the solution process as an electron transport layer (ETL) and buffer layer respectively, and the silver electrode was thermally evaporated to complete the cell. Figure 10a depicts the best-performing device J−VJ - V curves in forward bias under 1 Sun illumination for all the mixed Pz:PFN based perovskite solar cells. The J−VJ - V curves under forward and reverse scan directions are shown in Figure 11b and the device parameters are tabulated in Table 3. With a change in mixing ratio, the device performance first increased and then decreased. We note that there is a change in JSCJ_{\mathrm{SC}} (also integrated JSCJ_{\mathrm{SC}} from IPCE) with

Image not extracted

Figure 10 (PDF p. 183). (Colour line figures – Pz:PFN (6:4) is red colour, Pz:PFN (7:3) is blue colour, Pz:PFN (8:2) is magenta colour, Pz:PFN (9:1) is olive colour, Pz:PFN (9.5:0.5) is wine colour). (a) Current density vs. Voltage (JJ – VV) characteristics of the photovoltaic devices under 1-Sun (100 mW/cm2^{2}) condition in the forward scan direction. The forward and reverse JJ – VV scans of the representative devices are shown in Figure 11b. The inset figure represents the power output of the Pz:PFN (9:1) PSC device taken under maximum power point tracking with a stabilized efficiency is 20.14% (at 970 mV). (b) IPCE spectrum of the best performed Pz:PFN (9:1) PSC including the reflection (RR), transmission (TT), and integrated current density (Int. JSCJ_{\mathrm{SC}}). (c) Boxplot of short circuit current density (JSCJ_{\mathrm{SC}}), open circuit voltage (VOCV_{\mathrm{OC}}), fill factor (FFFF), and power conversion efficiency (PCE%) over 30 devices.

Image not extracted

Figure 11 (PDF p. 184). (a) J−VJ - V characteristics of the photovoltaic devices under 1-Sun (100 mW/cm2^{2}) condition in the forward and reverse scan direction for 0.805 cm2^{2} active area (11.5 mm × 7 mm) PSC device. The inset figure represents the PSC of active area 0.805 cm2^{2}. (b) J−VJ - V characteristic curves of photovoltaic devices device under AM1.5G illumination conditions employing different mixing ratios of Pz:PFN HTL, under forward and reverse scan for 0.175 cm2^{2} active area (2.5 mm × 7 mm) PSC device. (c) The power output of the PSCs taken under maximum power point tracking for Pz:PFN (6:4) with a stabilized efficiency of 17.32% (at 920 mV). The inset figure represents the PSC of active area 0.175 cm2^{2}.

respect to the compositional ratio of HTL, which will be discussed later along with KPFM results. The device obtained from Pz:PFN with a 9:1 ratio showed the highest PCE of 20.67% with a JSCJ_{SC} of 22.54 mA/cm2^{2}, VOCV_{OC} of 1.16 V, FFFF of 79.12% with slight hysteresis (please see Table 3). Note that the obtained efficiency is one of the highest values reported with Me-4PACz based 1.6 eV bandgap PSCs device Figure 13, Table 4. The photovoltaic device performance over 0.805 cm2^{2} active area (11.5 mm x 7 mm) PSC device is 19.06% with JSCJ_{SC} = 22.25 mA/cm2^{2}, VOCV_{OC} = 1.145V, and FFFF = 74.81% Figure 11a. Figure 10b shows the incident photon to current efficiency (IPCE) spectrum of our best-performing device along with the reflection (R), and transmission (T) spectrum. The IPCE spectra of the devices based on different Pz:PFN ratios are shown in Figure 12. The dip in the EQE spectrum correlates well with the hump in the reflection spectra. The loss in the current density due to reflection can be calculated from the external quantum efficiency (EQE) and internal quantum efficiency (IQE) spectrum using equations (41) and (42). The integrated JSC,EQEJ_{SC,EQE} is calculated by integrating the EQE spectrum over the 1-Sun spectrum. There is a current density mismatch by less than 5% of the JSC,EQEJ_{SC,EQE} to JSCJ_{SC} measured from the J − V measurement, and this can be understood by edge effect from the active area (17.5 mm2^{2} inset figure 11c) of the device or pre-bias measurement condition.[30, 31]

Image not extracted

Figure 12 (PDF p. 185). (a) EQE/IQE & reflectance (R) spectra and (b) EQE & integrated JSCJ_{SC} of PSC devices employing different mixing ratios of Pz:PFN HTLs.

Following are the equations used to calculate the integrated current density (Int. JSCJ_{SC}) from the EQE and IQE spectra to calculate the loss in the current due to reflection.

IQE(%)=EQE(%)1−0.01∗R(%)−0.01∗T(%)(39)IQE(\%) = \frac{EQE(\%)}{1 - 0.01 * R(\%) - 0.01 * T(\%)} \tag*{(39)}
IQE(%)=EQE(%)1−0.01∗R(%)(40)IQE(\%) = \frac{EQE(\%)}{1 - 0.01 * R(\%)} \tag*{(40)}
JSC(EQE)=q∫300 nm850 nmϕ(λ)EQE(λ) dλ(41)J_{SC}(EQE) = q \int_{300\,\mathrm{nm}}^{850\,\mathrm{nm}} \phi(\lambda)EQE(\lambda)\,\mathrm{d}\lambda \tag*{(41)}
JSC(IQE)=q∫300 nm850 nmϕ(λ)IQE(λ) dλ(42)J_{SC}(IQE) = q \int_{300\,\mathrm{nm}}^{850\,\mathrm{nm}} \phi(\lambda)IQE(\lambda)\,\mathrm{d}\lambda \tag*{(42)}

where T (%) and R (%) are the transmission and reflection spectra in percentage. ϕ(λ)\phi(\lambda) represents the 1-Sun spectrum, and qq is the elementary charge.

T (%) is <1%, so it can be neglected in the equation (39) and represents equation (40).

The int. JSC,EQEJ_{SC,EQE} is 24.28 mA/cm2^2 and int. JSC,IQEJ_{SC,IQE} is 21.41 mA/cm2^2. The loss in the current density (JSC,EQEJ_{SC,EQE}) is due to reflection from the opaque device. This is a significant loss that we can overcome by the anti-reflection (ARC) coating.

HTLJSCJ_{\mathrm{SC}} (mA/cm2^2)VOCV_{\mathrm{OC}} (V)FFFF (%)PCE (%)Int. JSCJ_{\mathrm{SC}} (mA/cm2^2)Hysteresis index
Pz:PFN (6:4) –F22.12112373.9918.3821.116.92
Pz:PFN (6:4) –R22.01111170.3017.1921.116.92
Pz:PFN (7:3) –F22.33113473.3318.5721.143.05
Pz:PFN (7:3) –R22.06112372.7418.0221.143.05
Pz:PFN (8:2)–F22.37113776.2319.3921.130.87
Pz:PFN (8:2) –R22.38113976.7319.5621.130.87
Pz:PFN (9:1) –F22.54115979.1220.6721.412.22
Pz:PFN (9:1) –R22.65115177.5620.2221.412.22
Pz:PFN (9.5:0.5) –F22.64113474.6719.1721.432.89
Pz:PFN (9.5:0.5) –R22.67113376.8519.7421.432.89
Pz:PFN (9:1) –F (0.805 cm2^2)22.06113772.1218.0921.255.09
Pz:PFN (9:1) –R (0.805 cm2^2)22.25114574.8119.0621.255.09

Table 3 (PDF p. 187). Table 3: Photovoltaic device parameters, integrated JSCJ_{\mathrm{SC}} (measured from IPCE), and hysteresis index of Pz:PFN HTLs over the active area of 0.175 cm2^2 and 0.805 cm2^2 derived from Figure 11 and Figure 12. Where F and R represent forward and reverse scan directions respectively.

Image not extracted

Figure 13 (PDF p. 187). Figure 13: The performance of the solar cells corresponds to the active area of the device and has a bandgap of the perovskite layer of Eg∼(1.60±0.04)E_{\mathrm{g}} \sim(1.60 \pm0.04) eV.

SL. #.Bandgap (eV)HTLsJSCJ_{SC} (mA/cm2^{2})VOCV_{OC} (V)FFFF (%)PCE (%)Active area (cm2^{2})ref. in this chapterref. in the Figure 13
1.~ 1.60PEDOT:PSS20.111.1180.618.00.11321
2.~ 1.60P3CT-Rb21.671.14482.7820.52-332
3.~ 1.60PTAA20.721.238321.150.06343
4.~ 1.60PTAA22.611.0918220.200.08354
5.~ 1.59PTAA/PEAI24.511.1582.123.10.09365
6.~ 1.62PTAA/PEAI21.001.1678220.1-376
7.~ 1.60PTAA24.131.16284.623.720.04557
8~ 1.58PTAA22.891.16984.5522.620.073388
9~ 1.56PTAA23.861.1438222.340.07399
10~ 1.58Me-4PACz/Al2_2O3_323.01.0979.419.90.094010
11~ 1.58Me-4PACz/ PFN-Br20.501.0874.716.50.094010
12~ 1.60Me-4PACz:PFN-Br22.541.15979.1220.670.175This workThis work

Table 4 (PDF p. 188). List of the PV parameters of PSCs for bandgap EgE_g ~ (1.60 ±\pm 0.04) eV with corresponding active area of the device and the hole transport layer used.

Transition energy levels of intrinsic defects in FAPbI3 and in the mixed organic cation FA0.52MA0.48PbI3 perovskite

Figure 13. Transition energy levels of intrinsic defects (a) FAPbI3_3 and (b) in the mixed organic cation FA0.52_{0.52}MA0.48_{0.48}PbI3_3 perovskite. (adapted from ref.81).

To verify the reproducibility of the device performance with Pz:PFN mixing engineering strategy, 30 devices of each mixing ratio were fabricated using the device procedure outlined in the experimental section above. Figure 10(c) summarizes the distribution of photovoltaic parameters under forward scan. The average performance improved from ~17% for Pz:PFN (6:4) to >20% for Pz:PFN (9:1) ratio case with the most narrow distribution among 30 PSCs, i.e., highly reproducible efficient PSCs. The performance improvement is attributed to the increase in JSCJ_{SC}, VOCV_{OC}, and FFFF. The stabilized efficiency under maximum power point tracking for the Pz:PFN (9:1) PSC device is 20.14% (at 970 mV) inset Figure 10(a) ( and for Pz:PFN (6:4) PSC device is 17.32% (at 920 mV) Figure 11 (c) )

Image not extracted

Figure 14 (PDF p. 189). (a) Structure of PFN-Br showing non-polar groups responsible for hydrophobic nature and (b) water contact angle showing hydrophobicity of PFN-Br.

Image not extracted

Figure 15 (PDF p. 189). Solubility of PFN-Br in DMF and DMSO solvents.

Understanding of the buried interface of perovskite solar cells

Numerous research reports have employed PFN-Br as an interlayer to overcome the hydrophobicity of polymeric HTLs such as PTAA to obtain a uniform perovskite film and even as an additive in the perovskite solution to improve its performance.[15, 28, 41] On the other hand, by looking at the structure of PFN-Br (Figure 14a), one can see the presence of a non-polar long alkyl chain. This can cause severe wetting issues and, therefore, raises similar concerns about forming a non-uniform perovskite layer. Moreover, the high water contact angle of PFN-Br, as shown in Figure 14b, corroborates the hypothesis. However, a uniform perovskite layer was formed after the incorporation of PFN-Br. This raises the point of how the perovskite layer is formed when a highly hydrophobic PFN-Br is introduced. In literature, the role of PFN-Br in helping to form a uniform perovskite layer has not been focused on yet. For instance, Wang et al. incorporated PFN-Br interlayer in between the PTAA HTL and perovskite absorber layer and assumed the possibility of PFN-Br redissolution during the deposition of the perovskite. layer, thereby improvement in the device performance.[41] Therefore, it is of utmost importance to understand the underlying mechanism of a uniform perovskite layer formation on PFN-Br interlayer or mixed Pz:PFN HTL. We noticed that PFN-Br is readily soluble in DMF and DMSO solvents (Figure 15) and allows us to make good quality films as shown in Figure 8 and domain analysis in Figure 16, with almost similar average domain sizes. Further, we deposited MAPbI3\mathrm{MAPbI}_{3} (without Br−\mathrm{Br}^{-}) independently on MeO-2PACz and Pz:PFN-coated glass substrate. The MAPbI3\mathrm{MAPbI}_{3} perovskite film deposited on MeO-2PACz and Pz:PFN (9:1) is indistinctly uniform in Figure 17.

Image not extracted

Figure 16 (PDF p. 191). (a) XRD pattern of normalized (110) peak of the CsFAMA perovskite film deposited on bilayer Pz_z/PFN and Pz_z:PFN(9:1) coated substrates. The grain size was calculated from the FESEM images of the perovskite films (using Image-J software) deposited on bilayer (b) Pz_z/PFN and (c) Pz_z:PFN(9:1) coated substrates.

Image not extracted

Figure 17 (PDF p. 192). Photographic image of the MAPbI3MAPbI_3 perovskite film deposited on the (a) MeO-2PACz and (b) Pz:PFN (9:1) HTLs coated glass substrates. FESEM images of the MAPbI3MAPbI_3 perovskite film deposited on the (c) MeO-2PACz and (d) Pz:PFN (9:1) HTLs coated glass substrates.

The XRD diffractograms were carried out for both films as shown in Figure 18a. The XRD pattern of MAPbI3MAPbI_3 deposited on Pz:PFN HTL showed a peak with lower full width at half maximum (FWHM). A slight shift (~0.0024º) in XRD peak is observed between MAPbI3MAPbI_3 film deposited on MeO-2PACz coated glass substrate vs MAPbI3MAPbI_3 film on Pz:PFN (9:1) coated glass substrate. The reason for this small shift is investigated via spectroscopic technique as the spectroscopic probe is known to be more sensitive than the structural probe. As per previous reports by us and others, micro-structure synchrotron studies were found to be less conclusive than spectroscopic studies, where spectroscopy could provide essential insight to correlate with solar cells performance.[42, 43] These samples are tested for PL spectroscopy. The raw PL data showed small peak-shift and FWHM differences. FWHM of MAPI film being broad on MeO-2PACz substrates could be explained based on the FWHM of the XRD peak. However, peak-shift which is found to be red-shifted in raw data, interestingly overlapped once an O.D. (or self-absorption) correction is introduced Figure 18 (b) and Figure 19.[44]

Image not extracted

Figure 18 (PDF p. 193). (a) XRD pattern and (b) PL spectra (self-absorption corrected) of the MAPbI3_3 perovskite thin film ( thickness ~ 250 nm ) deposited on the MeO-2PACz and Pz:PFN(9:1) coated glass substrates.[44] (c) 1^1H NMR of the CsFAMA, CsFAMA +PFN-Br and PFN-Br solution prepared in DMSO d6 solvent. (d) 1^1H NMR of the MAI, MAI +PFN-Br and PFN-Br solution prepared in DMSO d6 solvent. (e) 1^1H NMR of the FAI, FAI +PFN-Br, and PFN-Br solution prepared in DMSO d6 solvent. (f) The work function of the Pz:PFN HTLs was measured using KPFM study.

Image not extracted

Figure 19 (PDF p. 194). (a) The photoluminescence (PL) spectra and the (b) absorption spectra of the MAPbI3\mathrm{MAPbI}_{3} perovskite film deposited on the MeO-2PACz and Pz:PFN (9:1) HTL. (c) The PL spectra after self-absorption correction of the MAPbI3\mathrm{MAPbI}_{3} perovskite films deposited on the MeO-2PACz and Pz:PFN (9:1) HTLs.

This suggests that bromine containing Pz:PFN is not incorporate the Br⁻ ion into the absorber layer within the sensitivity limits of used structural and spectroscopic probes. A small peak shift on the higher degree side in the XRD peak falls near the resolution limit of the instrument, which is 0.0012°.

To further understand the interaction between PFN-Br and perovskite, a series of liquid-state ¹H nuclear magnetic resonance (NMR) measurements in solution (deuterated DMSO) were performed, and the results are shown in Figure 18c – e. Initially, an ¹H NMR spectrum of triple-cation perovskite solution dissolved in deuterated DMSO was recorded (Figure 18c), showing a resonance peak at around 9 ppm, which corresponds to protons bound to the nitrogen atoms of formamidinium (HC(NH₂)₂⁺ or FA⁺) and methylammonium (CH₃NH₃⁺ or MA⁺). Upon addition of PFN-Br into the perovskite solution in deuterated DMSO, the peak at around 9 ppm splits into two new peaks due to interaction between cationic HC(NH₂)₂⁺ (FA⁺)/CH₃NH₃⁺ (MA⁺) and PFN-Br. To further examine the interaction between MA⁺ and PFN-Br, an additional ¹H NMR experiment of only MAI without and with PFN-Br was carried out. As shown in Figure 18d, no peaks are split and shifted, indicating that the peak split in the perovskite solution with PFN-Br addition (Figure 18c) is related to the change in the FA⁺ cation in the presence of PFN-Br. To further confirm this conclusion, we recorded a ¹H NMR spectrum of only FAI without and with PFN-Br. As shown in Figure 18e, this peak around 9 ppm of FAI solution with PFN-Br additive gets split into two peaks in a way similar to the splitting observed in the perovskite solution with PFN-Br additive. The splitting of FA⁺ protons into two new signals has also been observed by other research groups and is ascribed to the formation of hydrogen bond complexes of the amidinium moiety of the FA⁺ cation.[45, 46, 47, 48]

Considering all NMR results, it is thus proposed that a hydrogen bond between FA⁺ and PFN-Br is existed, which might help in perovskite layer formation during its spin coating step and crystallization and might also remain present in the final perovskite film and can be beneficial for prolong device stability. In addition to the perovskite layer formation with an improved interfacial property, we hypothesized that the addition of PFN-Br might influence the electronic properties of Me-4PACz. Therefore, we measured the work function of Me-4PACz (without and with PFN-Br different ratios), and perovskite using Kelvin Probe Force Microscopy (KPFM) technique. The work function of the films is determined using gold as a reference and the results are shown in Figure 18f, Figure 20, and Figure 21.

Image not extracted

Figure 21. (a) Work function of Me-4PACz without and with PFN-Br interlayer and with different Pz:PFN mixing ratios, perovskite, PFN-Br, ITO, and Au metal electrode measured using KPFM. (b) The valence band maximum and the Fermi-level of the Me-4PACz and PFN-Br mixed HTLs are measured with respect to the vacuum using ultraviolet photoelectron spectroscopy (UPS).

Image not extracted

Figure 20 (PDF p. 196). Figure 20: (a) The KPFM scanning image of a standard gold (Au) sample deposited on ITO substrate. (b) KPFM scanning image of perovskite film deposited on the ITO substrate. The

work function of the Pz:PFN HTL at two different locations and the average of them are listed in Table S5. All the Pz:PFN HTL samples were deposited on the ITO substrates. (c) and (d) are the KPFM scanning image at two different points for the Pz:PFN (6:4). (e) and (f) are the KPFM scanning image at two different points for the Pz:PFN (8:2). (g) and (h) are the KPFM scanning image at two different points for the Pz:PFN (9:1). (i) and (j) are the KPFM scanning images at two different points for the Pz:PFN (9.5:0.5).

The work function of the samples is calculated using the following formula.

VCPD=WFtip−WFsample−e(43)V_{\mathrm{CPD}} = \frac{WF_{\mathrm{tip}}-WF_{\mathrm{sample}}}{-e} \tag*{(43)}

Where WFtipWF_{\mathrm{tip}} is the work function of the tip and WFsampleWF_{\mathrm{sample}} is the work function of the sample. VCPDV_{\mathrm{CPD}} is the contact potential difference.[49, 50]

SubstrateAverage VCPDV_{\mathrm{CPD}} (mV)Work function (eV)
ITO/Au1405.1
ITO/Perovskite4965.46
ITO/Pz:PFN (6:4)-124.95±0.0184.95 \pm0.018
ITO/Pz:PFN(8:2)3475.307±0.0265.307 \pm0.026
ITO/Pz:PFN(9:1)3555.315±0.0345.315 \pm0.034
ITO/Pz:PFN(9.5:0.5)3135.273±0.0055.273 \pm0.005

Table 5. List of the average of the contact potential difference (VCPDV_{\mathrm{CPD}}) for the Au, perovskite, Pz:PFN HTLs, and corresponding work functions.

Though the pristine Me-4PACz shows a better energy level alignment with the perovskite layer Figure 21a. However, forming the perovskite layer on pristine Me-4PACs is a challenge. Because of the effect of interface dipoles caused by the electrolyte moiety present in PFN-Br, the work function of Me-4PACz showed dramatic changes with the mixing ratios. For the particular Pz:PFN (9:1) case, the energy level alignment matches closely with the Me-4PACz (without any PFN-Br) and also with the work function of the perovskite layer, implying better hole extraction efficacy. Moreover, we used ultraviolet photoelectron spectroscopy (UPS) on the above mentioned HTLs and found that the Fermi level values are well aligned with KPFM results, as shown in Figure 21b. Further, the built-in potential difference ((VbiV_{\mathrm{bi}}) calculated from the dark current characteristics of diodes show a clear trend in reduced barrier-voltage for the case of Pz:PFN (9:1) composition as compared to the other mixing ratio composition, as seen in Figure 22. We observed a slight relative change in the EQE and corresponding JSCJ_{\mathrm{SC}} (whether integrated or measured using an AM1.5G solar simulator light source), which can be seen modulating with respect to (VbiV_{\mathrm{bi}} values of diodes. (VbiV_{\mathrm{bi}} values modulate with respect to different Fermi-levels of ITO/(Pz:PFN::X:Y) layer, as ETL is common in all diodes (Figure 21).[51] This also explains the reason behind the high device efficiency Pz:PFN (9:1) case compared to the other cases despite similar perovskite crystal structure and top surface morphology.

Image not extracted

Figure 22. The built-in potential (VbiV_{\mathrm{bi}}) is estimated from the dark current-voltage characteristics for the diode of ITO/Pz:PFN(X:Y)/Perovskite/PCBM/BCP/Ag.

Device physics of perovskite solar cells

HTLDark current JdJ_{\mathrm{d}} @ -0.3 (mA/cm2^2)J0J_0 (mA/cm2^2)
Pz:PFN (6:4)3.72 X 10−410^{-4}1.07 X 10−810^{-8}
Pz:PFN (7:3)8.04 X 10−510^{-5}4.47 X 10−910^{-9}
Pz:PFN (8:2)5.10 X 10−510^{-5}1.43 X 10−910^{-9}
Pz:PFN (9:1)5.28 X 10−610^{-6}2.20 X 10−1010^{-10}
Pz:PFN (9.5:0.5)2.04 X 10−510^{-5}5.10 X 10−1010^{-10}

Table 6. The JdJ_{\mathrm{d}} and J0J_0 values of all the devices with different Pz:PFN mixing ratios derived from Figure 23a.

Image not extracted

Figure 23. (a) The dark current density plotted in semi-log scale as a function of applied bias. (b) The log (JSC/J0J_{\mathrm{SC}}/J_0) vs. mixing ratio Pz:PFN (X:Y), where X and Y represent the volume ratio for Me-4PACz and PFN-Br respectively.

The dark current measurement can also give insights into the recombination at the perovskite and Pz:PFN interface by an estimation of reverse saturation current density J0J_0.{sup}52 Therefore, we measured the dark current of the devices with different ratios of Pz:PFN and observed the lowest dark current density $J_{\mathrm{d}}$ at $-0.3$ V of $5.28 \times 10^{-6}$ mA/cm$^2$ and the reverse saturation current density $J_0$ of $2.20 \times 10^{-10}$ mA/cm$^2$ for the Pz:PFN (9:1) ({xref}Figure 23`) whereas a typical

References

  1. [1]1 A. Kojima, K. Teshima, Y. Shirai, and T. Miyasaka, “Organometal halide perovskites as visible-light sensitizers for photovoltaic cells,” J. Am. Chem. Soc. 131(17), 6050–6051 (2009).DOI
  2. [2]2 J. Jeong, M. Kim, J. Seo, H. Lu, P. Ahlawat, A. Mishra, Y. Yang, M.A. Hope, F.T. Eickemeyer, M. Kim, Y.J. Yoon, I.W. Choi, B.P. Darwich, S.J. Choi, Y. Jo, J.H. Lee, B. Walker, S.M. Zakeeruddin, L. Emsley, U. Rothlisberger, A. Hagfeldt, D.S. Kim, M. Grätzel, and J.Y. Kim, “Pseudo-halide anion engineering for α-FAPbI3 perovskite solar cells,” Nature 592(7854), 381–385 (2021).DOI
  3. [3]3 H. Min, D.Y. Lee, J. Kim, G. Kim, K.S. Lee, J. Kim, M.J. Paik, Y.K. Kim, K.S. Kim, M.G. Kim, T.J. Shin, and S. Il Seok, “Perovskite solar cells with atomically coherent interlayers on SnO2 electrodes,” Nature 598(7881), 444–450 (2021).DOI
  4. [4]4 X. Zhang, S. Yuan, H. Lu, H. Zhang, P. Wang, X. Cui, Y. Zhang, Q. Liu, J. Wang, Y. Zhan, Z. Sun, and W. Huang, “Hydrazinium Salt as Additive to Improve Film Morphology and Carrier Lifetime for High-Efficiency Planar-Heterojunction Perovskite Solar Cells via One-Step Method,” ACS Appl. Mater. Interfaces 9(42), 36810–36816 (2017).DOI
  5. [5]5 Z. Cheng, C. Gao, J. Song, D. Ding, Y. Chen, J. Wang, D. Zhang, L. Chen, X. Wang, Z. Yang, F. Liu, H. Liu, and W. Shen, “Interfacial and Permeating Modification Effect of n-type Non-fullerene Acceptors toward High-Performance Perovskite Solar Cells,” ACS Appl. Mater. Interfaces 13(34), 40778–40787 (2021).DOI
  6. [6]6 B. Sharma, S. Singh, K. Hossain, Laxmi, S. Mallick, P. Bhargava, and D. Kabra, “Additive engineering of 4, 4’-Bis (N-carbazolyl)-1, 1’-biphenyl (CBP) molecules for defects passivation and moisture stability of hybrid perovskite layer,” Sol. Energy 211, 1084–1091 (2020).DOI
  7. [7]7 M. Saliba, J.P. Correa-Baena, C.M. Wolff, M. Stolterfoht, N. Phung, S. Albrecht, D. Neher, and A. Abate, “Erratum: How to Make over 20% Efficient Perovskite Solar Cells in Regular (n-i-p) and Inverted (p-i-n) Architectures (Chem. Mater. (2018) 30:13 (4193-4201) DOI: 10.1021/acs.chemmater.8b00136),” Chem. Mater. 31(20), 8576 (2019).
  8. [8]8 C. Ma, and N.G. Park, “Paradoxical Approach with a Hydrophilic Passivation Layer for Moisture-Stable, 23% Efficient Perovskite Solar Cells,” ACS Energy Lett. 5(10), 3268–3275 (2020).DOI
  9. [9]9 F. Li, X. Deng, F. Qi, Z. Li, D. Liu, D. Shen, M. Qin, S. Wu, F. Lin, S.H. Jang, J. Zhang, X. Lu, D. Lei, C.S. Lee, Z. Zhu, and A.K.Y. Jen, “Regulating Surface Termination for Efficient Inverted Perovskite Solar Cells with Greater Than 23% Efficiency,” J. Am. Chem. Soc. 142(47), 20134–20142 (2020).
  10. [10]D. Corteccia, H.A. Dewi, J. Yin, A. Bruno, S. Chen, T. Baikie, P.P. Boix, and M. Gra, “Lead-Free Hybrid Perovskites,” Inorg. Chem. 55(3), 1044–1052 (2016).
  11. [11]L.T. Schellhas, Z. Li, J.A. Christians, A. Goyal, P. Kairys, S.P. Harvey, D.H. Kim, K.H. Stone, J.M. Luther, K. Zhu, V. Stevanovic, and J.J. Berry, “Insights into operational stability and processing of halide perovskite active layers,” Energy Environ. Sci. 12(4), 1341–1348 (2019).
  12. [12]W. Yang, D. Zhong, M. Shi, S. Qu, and H. Chen, “Toward highly thermal stable perovskite solar cells by rational design of interfacial layer,” Iscience 22, 534–543 (2019).DOI
  13. [13]T.A. Berhe, W.-N. Su, C.-H. Chen, C.-J. Pan, J.-H. Cheng, H.-M. Chen, M.-C. Tsai, L.-Y. Chen, A.A. Dubale, and B.-J. Hwang, “Organometal halide perovskite solar cells: degradation and stability,” Energy Environ. Sci. 9(2), 323–356 (2016).DOI
  14. [14]B. Li, Y. Li, C. Zheng, D. Gao, and W. Huang, “Advancements in the stability of perovskite solar cells: degradation mechanisms and improvement approaches,” RSC Adv. 6(44), 38079–38091 (2016).DOI
  15. [15]A.D. Sheikh, R. Munir, M.A. Haque, A. Bera, W. Hu, P. Shaikh, A. Amassian, and T. Wu, “Effects of High Temperature and Thermal Cycling on the Performance of Perovskite Solar Cells: Acceleration of Charge Recombination and Deterioration of Charge Extraction,” ACS Appl. Mater. Interfaces 9(40), 35018–35029 (2017).DOI
  16. [16]Y. Lv, H. Zhang, R. Liu, Y. Sun, and W. Huang, “Composite Encapsulation Enabled Superior Comprehensive Stability of Perovskite Solar Cells,” ACS Appl. Mater. Interfaces 12(24), 27277–27285 (2020).DOI
  17. [17]F. Corsini, and G. Griffini, “Recent progress in encapsulation strategies to enhance the stability of organometal halide perovskite solar cells,” JPhys Energy 2(3), 31002 (2020).DOI
  18. [18]C. Ran, J. Xu, W. Gao, C. Huang, and S. Dou, “Defects in metal triiodide perovskite materials towards high-performance solar cells: Origin, impact, characterization, and engineering,” Chem. Soc. Rev. 47(12), (2018).DOI
  19. [19]P. Fassl, Y. Zakharko, L.M. Falk, K.P. Goetz, F. Paulus, A.D. Taylor, J. Zaumseil, and Y. Vaynzof, “Effect of density of surface defects on photoluminescence properties in MAPbI 3 perovskite films,” J. Mater. Chem. C 7(18), 5285–5292 (2019).DOI
  20. [20]S. Zhang, P.E. Shaw, G. Zhang, H. Jin, M. Tai, H. Lin, P. Meredith, P.L. Burn, D. Neher, and M. Stolterfoht, “Defect/Interface Recombination Limited Quasi-Fermi Level Splitting and Open-Circuit Voltage in Mono-and Triple-Cation Perovskite Solar Cells,” ACS Appl. Mater. Interfaces 12(33), 37647–37656 (2020).
  21. [21]D. Meggiolaro, E. Mosconi, and F. De Angelis, “Formation of Surface Defects Dominates Ion Migration in Lead-Halide Perovskites,” ACS Energy Lett. 4(3), 779–785 (2019).DOI
  22. [22]G.E. Eperon, S.D. Stranks, C. Menelaou, M.B. Johnston, L.M. Herz, and H.J. Snaith, “Formamidinium lead trihalide: A broadly tunable perovskite for efficient planar heterojunction solar cells,” Energy Environ. Sci. 7(3), 982–988 (2014).DOI
  23. [23]Q. Han, S. Bae, P. Sun, Y. Hsieh, Y. Yang, Y.S. Rim, H. Zhao, Q. Chen, W. Shi, and G. Li, “Single crystal formamidinium lead iodide (FAPbI3): insight into the structural, optical, and electrical properties,” Adv. Mater. 28(11), 2253–2258 (2016).DOI
  24. [24]Z. Li, M. Yang, J.-S. Park, S.-H. Wei, J.J. Berry, and K. Zhu, “Stabilizing perovskite structures by tuning tolerance factor: formation of formamidinium and cesium lead iodide solid-state alloys,” Chem. Mater. 28(1), 284–292 (2016).DOI
  25. [25]P. Luo, S. Zhou, Y. Zhou, W. Xia, L. Sun, J. Cheng, C. Xu, and Y. Lu, “Fabrication of CsxFA1-xPbI3 Mixed-Cation Perovskites via Gas-Phase-Assisted Compositional Modulation for Efficient and Stable Photovoltaic Devices,” ACS Appl. Mater. Interfaces 9(49), 42708–42716 (2017).DOI
  26. [26]N. Liu, and C.Y. Yam, “First-principles study of intrinsic defects in formamidinium lead triiodide perovskite solar cell absorbers,” Phys. Chem. Chem. Phys. 20(10), 6800–6804 (2018).DOI
  27. [27]Z. Yang, C.-C. Chueh, P.-W. Liang, M. Crump, F. Lin, Z. Zhu, and A.K.-Y. Jen, “Effects of formamidinium and bromide ion substitution in methylammonium lead triiodide toward high-performance perovskite solar cells,” Nano Energy 22, 328–337 (2016).DOI
  28. [28]Z. Hu, Q. An, H. Xiang, L. Aigouy, B. Sun, Y. Vaynzof, and Z. Chen, “Enhancing the Efficiency and Stability of Triple-Cation Perovskite Solar Cells by Eliminating Excess PbI2from the Perovskite/Hole Transport Layer Interface,” ACS Appl. Mater. Interfaces 12(49), 54824–54832 (2020).DOI
  29. [29]M.I. Saidaminov, J. Kim, A. Jain, R. Quintero-Bermudez, H. Tan, G. Long, F. Tan, A. Johnston, Y. Zhao, and O. Voznyy, “Suppression of atomic vacancies via incorporation of isovalent small ions to increase the stability of halide perovskite solar cells in ambient air,” Nat. Energy 3(8), 648–654 (2018).DOI
  30. [30]C. Wu, K. Chen, D.Y. Guo, S.L. Wang, and P.G. Li, “Cations substitution tuning phase stability in hybrid perovskite single crystals by strain relaxation,” RSC Adv. 8(6), 2900–2905 (2018).DOI
  31. [31]G. Sharada, P. Mahale, B.P. Kore, S. Mukherjee, M.S. Pavan, C. De, S. Ghara, A. Sundaresan, A. Pandey, and T.N.G. Row, “Is CH3NH3PbI3 polar,” J. Phys. Chem. Lett 7, 2412–2419 (2016).
  32. [32]Y. Rakita, O. Bar-Elli, E. Meirzadeh, H. Kaslasi, Y. Peleg, G. Hodes, I. Lubomirsky, D. Oron, D. Ehre, and D. Cahen, “Tetragonal CH3NH3PbI3 is ferroelectric,” Proc. Natl. Acad. Sci. U. S. A. 114(28), E5504–E5512 (2017).DOI
  33. [33]R. Pandey, G. SB, S. Grover, S.K. Singh, A. Kadam, S. Ogale, U. V Waghmare, V.R. Rao, and D. Kabra, “Microscopic origin of piezoelectricity in lead-free halide perovskite: application in nanogenerator design,” ACS Energy Lett. 4(5), 1004–1011 (2019).DOI
  34. [34]S.M. Sze, and S.M. Sze, Selected Solutions for Semiconductor Devices: Physics and Technology (Wiley, 1985).
  35. [35]S. Singh, G. Banappanavar, and D. Kabra, “Correlation between Charge Transport Length Scales and Dielectric Relaxation Time Constant in Hybrid Halide Perovskite Semiconductors,” ACS Energy Lett. 5(3), 728–735 (2020).DOI
  36. [36]J.C. De Mello, H.F. Wittmann, and R.H. Friend, “An improved experimental determination of external photoluminescence quantum efficiency,” Adv. Mater. 9(3), 230–232 (1997).DOI
  37. [37]S. Singh, B. Sharma, G. Banappanavar, A. Dey, S. Chakraborty, K.L. Narasimhan, P. Bhargava, and D. Kabra, “Investigation on Organic Molecule Additive for Moisture Stability and Defect Passivation via Physisorption in CH3NH3PbI3 Based Perovskite,” ACS Appl. Energy Mater. 1(5), 1870–1877 (2018).DOI
  38. [38]J. Panidi, A.F. Paterson, D. Khim, Z. Fei, Y. Han, L. Tsetseris, G. Vourlias, P.A. Patsalas, M. Heeney, and T.D. Anthopoulos, “Remarkable Enhancement of the Hole Mobility in Several Organic Small-Molecules, Polymers, and Small-Molecule:Polymer Blend Transistors by Simple Admixing of the Lewis Acid p-Dopant B(C6F5)3,” Adv. Sci. 5(1), 1700290 (2018).DOI
  39. [39]M. Bag, and K.S. Narayan, “Universality in the intensity-modulated photocurrent in bulk-heterojunction polymer solar cells,” Phys. Rev. B 82(7), 75308 (2010).DOI
  40. [40]S. Khatavkar, K. Muniappan, C. V. Kannan, V. Kumar, K.L. Narsimhan, P.R. Nair, J.M. Vasi, M.A. Contreras, M.F.A.M. van Hest, and B.M. Arora, “Measurement of Relaxation Time of Excess Carriers in Si and CIGS Solar Cells by Modulated Electroluminescence Technique,” Phys. Status Solidi Appl. Mater. Sci. 215(2), 1700267 (2018).DOI
  41. [41]W. Ke, C.C. Stoumpos, and M.G. Kanatzidis, ““Unleaded” Perovskites: Status Quo and Future Prospects of Tin-Based Perovskite Solar Cells,” Adv. Mater. 31(47), 1803230 (2019).DOI
  42. [42]D. Ghosh, P. Walsh Atkins, M.S. Islam, A.B. Walker, and C. Eames, “Good vibrations: Locking of octahedral tilting in mixed-cation iodide perovskites for solar cells,” ACS Energy Lett. 2(10), 2424–2429 (2017).
  43. [43]C.H. Chiang, and C.G. Wu, “A Method for the Preparation of Highly Oriented MAPbI3 Crystallites for High-Efficiency Perovskite Solar Cells to Achieve an 86% Fill Factor,” ACS Nano 12(10), 10355–10364 (2018).DOI
  44. [44]M. Saliba, T. Matsui, J.Y. Seo, K. Domanski, J.P. Correa-Baena, M.K. Nazeeruddin, S.M. Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt, and M. Grätzel, “Cesium-containing triple cation perovskite solar cells: Improved stability, reproducibility and high efficiency,” Energy Environ. Sci. 9(6), 1989–1997 (2016).pubmed.ncbi.nlm.nih.gov/27478500
  45. [45]C.M. Proctor, and T.Q. Nguyen, “Effect of leakage current and shunt resistance on the light intensity dependence of organic solar cells,” Appl. Phys. Lett. 106(8), 23_1 (2015).DOI
  46. [46]J.A. Chang, S.H. Im, Y.H. Lee, H.J. Kim, C.S. Lim, J.H. Heo, and S. Il Seok, “Panchromatic photon-harvesting by hole-conducting materials in inorganic-organic heterojunction sensitized-solar cell through the formation of nanostructured electron channels,” Nano Lett. 12(4), 1863–1867 (2012).DOI
  47. [47]A.A. Syed, C.Y. Poon, H.W. Li, and F. Zhu, “A sodium citrate-modified-PEDOT:PSS hole transporting layer for performance enhancement in inverted planar perovskite solar cells,” J. Mater. Chem. C 7(18), 5260–5266 (2019).DOI
  48. [48]S. Singh, R.J. Shourie, and D. Kabra, “Efficient and thermally stable CH3NH3PbI3 based perovskite solar cells with double electron and hole extraction layers,” J. Phys. D. Appl. Phys. 52(25), 255106 (2019).DOI
  49. [49]M. Saliba, and L. Etgar, “Current density mismatch in Perovskite solar cells,” ACS Energy Lett. 5(9), 2886–2888 (2020).DOI
  50. [50]S.N. Habisreutinger, N.K. Noel, and H.J. Snaith, “Hysteresis Index: A Figure without Merit for Quantifying Hysteresis in Perovskite Solar Cells,” ACS Energy Lett. 3(10), 2472–2476 (2018).DOI
  51. [51]D. Seol, A. Jeong, M.H. Han, S. Seo, T.S. Yoo, W.S. Choi, H.S. Jung, H. Shin, and Y. Kim, “Origin of hysteresis in CH3NH3PbI3 perovskite thin films,” Adv. Funct. Mater. 27(37), 1701924 (2017).DOI
  52. [52]J. Chen, D. Lee, and N.G. Park, “Stabilizing the Ag Electrode and Reducing J-V Hysteresis through Suppression of Iodide Migration in Perovskite Solar Cells,” ACS Appl. Mater. Interfaces 9(41), 36338–36349 (2017).DOI
  53. [53]D. Kabra, S. Shriram, N.S. Vidhyadhiraja, and K.S. Narayan, “Charge carrier dynamics in organic semiconductors by position dependent optical probing,” J. Appl. Phys. 101(6), 64510 (2007).DOI
  54. [54]I. Poli, G.W. Kim, E.L. Wong, A. Treglia, G. Folpini, and A. Petroz​​za, “High External Photoluminescence Quantum Yield in Tin Halide Perovskite Thin Films,” ACS Energy Lett. 6(2), 609–611 (2021).DOI
  55. [55]S.S. Li, Semiconductor Physical Electronics (Springer Science & Business Media, 2012).
  56. [56]G.J.A.H. Wetzelaar, M. Kuik, and P.W.M. Blom, “IdenTIfying the nature of charge recombination in organic solar cells from charge-transfer state electroluminescence,” Adv. Energy Mater. 2(10), 1232–1237 (2012).
  57. [57]M.H. Futscher, J.M. Lee, L. McGovern, L.A. Muscarella, T. Wang, M.I. Haider, A. Fakharuddin, L. Schmidt-Mende, and B. Ehrler, “Quantification of ion migration in CH 3 NH 3 PbI 3 perovskite solar cells by transient capacitance measurements,” Mater. Horizons 6(7), 1497–1503 (2019).DOI
  58. [58]S. Singh, and D. Kabra, “Comparative Study of Recombination Dynamics in Optimized Composition of Sn- Versus Pb-Based Perovskite Solar Cells,” ACS Appl. Mater. Interfaces 13(35), 42297–42306 (2021).DOI
  59. [59]D.W. Miller, G.E. Eperon, E.T. Roe, C.W. Warren, H.J. Snaith, and M.C. Lonergan, “Defect states in perovskite solar cells associated with hysteresis and performance,” Appl. Phys. Lett. 109(15), 153902 (2016).DOI
  60. [60]W. Tress, N. Marinova, O. Inganäs, M.K. Nazeeruddin, S.M. Zakeeruddin, and M. Graetzel, “Predicting the open-circuit voltage of CH3NH3PbI3 perovskite solar cells using electroluminescence and photovoltaic quantum efficiency spectra: the role of radiative and non-radiative recombination,” Adv. Energy Mater. 5(3), 1400812 (2015).
  61. [61]P.H. Joshi, L. Zhang, I.M. Hossain, H.A. Abbas, R. Kottokkaran, S.P. Nehra, M. Dhaka, M. Noack, and V.L. Dalal, “The physics of photon induced degradation of perovskite solar cells,” Aip Adv. 6(11), 115114 (2016).DOI
  62. [62]Y. Liu, B. Jin, H. Zhang, Y. Zhang, Y. Kim, C. Wang, S. Wen, B. Xu, C. Im, and W. Tian, “Efficiency of MAPbI 3 -Based Planar Solar Cell Analyzed by Its Thickness-Dependent Exciton Formation, Morphology, and Crystallinity,” ACS Appl. Mater. Interfaces 11(16), 14810–14820 (2019).DOI
  63. [63]M. Samiee, S. Konduri, B. Ganapathy, R. Kottokkaran, H.A. Abbas, A. Kitahara, P. Joshi, L. Zhang, M. Noack, and V. Dalal, “Defect density and dielectric constant in perovskite solar cells,” Appl. Phys. Lett. 105(15), 153502 (2014).DOI
  64. [64]V. Nandal, and P.R. Nair, “Predictive Modeling of Ion Migration Induced Degradation in Perovskite Solar Cells,” ACS Nano 11(11), 11505–11512 (2017).arxiv.org/abs/1704.06603
  65. [65]I. Alkorta, and J. Elguero, “A theoretical study of perovskites related to CH 3 NH 3 PbX 3 (X= F, Cl, Br, I),” New J. Chem. 42(16), 13889–13898 (2018).
  66. [66]M.T. Weller, O.J. Weber, J.M. Frost, and A. Walsh, “Cubic perovskite structure of black formamidinium lead iodide, α-[HC (NH2) 2] PbI3, at 298 K,” J. Phys. Chem. Lett. 6(16), 3209–3212 (2015).
  67. [67]A. Al-Ashouri, E. Köhnen, B. Li, A. Magomedov, H. Hempel, P. Caprioglio, J.A. Márquez, A.B.M. Vilches, E. Kasparavicius, J.A. Smith, N. Phung, D. Menzel, M. Grischek, L. Kegelmann, D. Skroblin, C. Gollwitzer, T. Malinauskas, M. Jošt, G. Matič, B. Rech, R. Schlatmann, M. Topič, L. Korte, A. Abate, B. Stannowski, D. Neher, M. Stolterfoht, T. Unold, V. Getautis, and S. Albrecht, “Monolithic perovskite/silicon tandem solar cell with >29% efficiency by enhanced hole extraction,” Science (80-. ). 370(6522), 1300–1309 (2020).
  68. [68]A. Al-Ashouri, A. Magomedov, M. Roß, M. Jošt, M. Talaikis, G. Chistiakova, T. Bertram, J.A. Márquez, E. Köhnen, E. Kasparavičius, S. Levcenсo, L. Gil-Escrig, C.J. Hages, R. Schlatmann, B. Rech, T. Malinauskas, T. Unold, C.A. Kaufmann, L. Korte, G. Niaura, V. Getautis, and S. Albrecht, “Conformal monolayer contacts with lossless interfaces for perovskite single junction and monolithic tandem solar cells,” Energy Environ. Sci. 12(11), 3356–3369 (2019).
  69. [69]“https://www.lumtec.com.tw/products-view.php?ID=8990,” (n.d.).
  70. [70]O. Fernandez-Delgado, E. Castro, C.R. Ganivet, K. Fosnacht, F. Liu, T. Mates, Y. Liu, X. Wu, and L. Echegoyen, “Variation of Interfacial Interactions in PC61BM-like Electron-Transporting Compounds for Perovskite Solar Cells,” ACS Appl. Mater. Interfaces 11(37), 34408–34415 (2019).

Paper details

vlnc:W_s8cvgbc5
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 9

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents