A novel fabrication technique for producing hybrid organic–inorganic perovskite impregnated poly(styrene) microfibers in situ in a melt electrospinner is presented. The relationship between the hybrid perovskite precursors, electrospun fiber morphology, and chemical phase domains was investigated using light microscopy, scanning electron microscopy, x-ray diffraction spectroscopy, and energy dispersive spectroscopy. The method was successful in producing composite microfibers and revealed information regarding the nature of encapsulated hybrid perovskites under higher-than-normal temperature conditions during the synthesis and subsequent annealing process.
I. INTRODUCTION
The existence of hybrid organic–inorganic perovskites (HOIPs) has been known since 1978.1 However, it was not until nearly 20 years later when the semiconducting nature of the HOIP class of materials was investigated.2 In 2006, HOIPs were first investigated as a light sensitizer in dye-sensitized solar cells (DSSC). At the time, HOIPs were synthesized as nanoparticles onto the mesoscopic TiO2 photoanode layer via spin coating. The HOIPs of interest at that time were CH3NH3PbBr3 and CH3NH3PbI3 with lead iodide providing a more favorable optical band gap. DSSCs produced using HOIPs were efficient and capable of high photovoltages; however, it was observed that the photocurrent decayed quickly in open air.3 Later, the DSSC model was abandoned in favor of a solid-state device in 2008 that utilized a polypyrrole/carbon black composite as the charge transport layer in the solar cell. From this point on, research into HOIP solid state solar cells started to increase rapidly. The power conversion efficiency (PCE) of HOIPs has risen rapidly from 3.81% in 2009 (Ref. 3) to 22.1% at the time of this writing.4 Despite the nearly 2% per year growth in the PCE, low cost of materials, and ease of processing, the perovskite solar cell is not widely implemented due to the short lifetime of the cells caused by chemical instability in the HOIP material.
Chemical instability drives the degradation of the HOIP and the functionality of the solar cell concomitantly. The degradation reaction that is generally accepted is given in reaction (R1). Methylammonium lead iodide perovskite (CH3NH3PbI3) degrades to leave solid lead iodide (PbI2) and highly volatile methyl amine (CH3NH2) and hydroiodic acid (HI).
The presence of water facilitates the above reaction, and the volatile nature of the degradation products further drives the degradation reaction to the right. The progress of the degradation can be directly observed as a color change from the black/brown CH3NH3PbI3 to the yellow PbI2. Degradation has been shown to be a surface phenomenon with H2O first forming monohydrates and dihydrates with CH3NH3PbI3. Complexation with H2O is reversible up to this point, and irreversible degradation does not occur until condensed water forms on the surface of HOIP crystals.5 There is a strong relationship between the surface area of HOIP and the rate of degradation when exposed to moisture. The efforts to mitigate the degrading effect of atmospheric moisture on HOIP films in solid state devices have been developed and most often involve the modification of the hole transport material (HTM).6 Modifications to the HTM are done to slow egress of water through the HTM to the active HOIP layer and include increasing the molecular weight of the hole transport polymer and increasing the hydrophilicity of layers above the HOIP layer to serve as moisture trapping layers.7 Modifications to the HTM are well suited to the solution processing techniques often used in the manufacturing of planar sequentially layered, spin-coated, and vapor deposited solid-state solar cells. In this study, a novel solutionless synthesis method8 developed previously by the authors is utilized in conjunction with melt electrospinning (M-ES) to fabricate microfibers of HOIP/poly(styrene) (PS) composite material that prevents moisture driven degradation by encapsulation of the perovskite material in a polymer matrix. The resultant structures were characterized using scanning electron microscopy (SEM), optical microscopy, powder x-ray diffraction, and energy dispersive spectroscopy (EDS).
II. EXPERIMENT
A. Materials
PbI2 (99%) was used as received from Sigma Aldrich. The CH3NH3I was prepared as described by Lee et al.9 in a 500 ml Erlenmeyer flask by adding 10 ml 57 wt. % HI in water dropwise to 24 ml 33 wt. % CH3NH2 in ethanol diluted with 100 ml absolute ethanol at RT (∼21 °C) while stirring for 2 h. The CH3NH3I was then crystallized using a rotary evaporator at 100 °C and dried overnight in a vacuum oven at 60 °C to produce a dry, white crystalline powder. Poly(styrene) was sourced from the U.S. Army Research Labs (Aberdeen, MD) as recycled petri-dishes, densified via mechanical milling. The molecular weight (229, 700 g/mol) was measured via gel permeation chromatography.
B. Methods
Microfibers were produced using a custom built M-ES apparatus seen in Fig. 1. The melt chamber is heated with electrically isolated heat tape, and the chamber itself is sealed by a pneumatically actuated piston. When the chamber is sufficiently heated to produce a polymer melt, the piston drives the melt through a grounded spinneret. A high strength electric field (∼5 kV/cm) is then applied which pulls the polymer melt from the spinneret tip via electrostatic elongation to the collection surface on a counterelectrode.
(Color online) Melt electrospinner apparatus. The counterelectrode was held at a high voltage (14 kV), and the spinneret was grounded. HOIP precursors and PS were introduced into the melt chamber and sealed in using the piston. Heat provided from the melt electrospinning process thermodynamically drove the reaction that formed HOIP crystallites.
(Color online) Melt electrospinner apparatus. The counterelectrode was held at a high voltage (14 kV), and the spinneret was grounded. HOIP precursors and PS were introduced into the melt chamber and sealed in using the piston. Heat provided from the melt electrospinning process thermodynamically drove the reaction that formed HOIP crystallites.
Mixtures of 40 wt. % HOIP precursors in PS in a molar ratio of 2:1 (CH3NH3I:PbI2) were first mixed thoroughly using sonication to break-up agglomerations of PbI2 in the mixture. A melt chamber was loaded with the mixture and then sealed with a pneumatically actuated piston and heated to approximately 200 °C; temperature was measured from the outside of the melt chamber as the sealed nature of the chamber does not permit internal measurement in situ. The spinneret was 3 cm from the counterelectrode, and a voltage of 14 kV was applied. The emitted composite microfibers were collected on glass slides. Additionally, fibers with only precursor materials doped into the PS melt were electrospun to study their influence and behavior in the microfibers and melt electrospinner.
A Mitutoyo FS60 light microscope was used to visually inspect the crystallites embedded throughout the composite fibers. Light microscopy allows a study of crystallite size and resultant effects on fiber morphology.
Electron micrographs were generated using a LEO SEM 1430 VP. Z contrast was added to images using Apollo 40 EDS detectors. Energy dispersive spectroscopy was achieved through the use of an EDAX system and was used to assess the stoichiometry of crystallites of interest. All samples were sputter coated with Au using a Denton Desk-1 at 25 mA and 100mTorr using atmospheric gas inlet for approximately 45s. Samples were fixed to SEM stubs using a conductive carbon tape.
Diffractograms were collected using a Rigaku Ultima IV x-ray diffractometer using a CuKα source and a Ni filter operating at 40 kV and 40 mA. Scans were carried out at a scan rate of 2°/min from 5° to 90° on a zero background quartz plate.
III. RESULTS AND DISCUSSION
A. Verification of embedded HOIP crystallites
1. Energy dispersive spectroscopy
Chemical composition of HOIP crystallites embedded in the electrospun composite fibers were investigated via SEM with EDS capabilities. SEM-EDS allows interrogation without damaging or altering the internal structure of the fiber. Understanding the placement of HOIP crystallites throughout the fiber and differentiating from unreacted precursors is crucial for future implementation of HOIP/PS microfibers, specifically in charge extraction. An embedded HOIP crystallite with measured relative atomic percentages can be seen in Fig. 2(c).
(Color online) Electron micrographs of melt electrospun poly(styrene) microfibers with (a)PbI2, (b) CH3NH3I, and (c) CH3NH3PbI3. An EDS spectrum was generated for each region of interest, denoted by a red cross in each micrograph. The relative atomic percentages of I and Pb in the perovskite microcrystallite (c) were measured as 75.82% and 24.18%, respectively. An I:Pb ratio of 3.14 corresponds to a slightly iodine-rich HOIP crystallite.
(Color online) Electron micrographs of melt electrospun poly(styrene) microfibers with (a)PbI2, (b) CH3NH3I, and (c) CH3NH3PbI3. An EDS spectrum was generated for each region of interest, denoted by a red cross in each micrograph. The relative atomic percentages of I and Pb in the perovskite microcrystallite (c) were measured as 75.82% and 24.18%, respectively. An I:Pb ratio of 3.14 corresponds to a slightly iodine-rich HOIP crystallite.
The crystallite seen in Fig. 2(c) appears to be spherical, although slightly out of focus due to imaging through the PS fiber. The shapes of the HOIP crystallites reveal how the precursors come together in the polymer melt. The temperature of the polymer melt is sufficiently high to sublimate the CH3NH3I precursor which has been reported to occur at 247 °C ± 26 °C.10 Therefore, it is possible that the CH3NH3I exists in the gaseous and solid states in the melt chamber. However, temperatures are not so high as to cause a change in state or phase in the PbI2 which has been recorded to melt at 402 °C (Ref. 11) and thermally decompose at 646 °C.10 The shape of the crystallite in the micrograph in Fig. 2(c) appears similar in size to the PbI2 crystallite shown in Fig. 2(a) which suggests HOIP formation about the PbI2 crystallites. The formation about the PbI2 crystallites is further supported by the mixed state of the CH3NH3I in the melt chamber. The gaseous CH3NH3I would be the fastest diffusing species in the polymer melt; a fact supported by the large amounts of CH3NH3I that appear to have phase separated from the polystyrene upon rapid solidification through cooling seen in Fig. 2(b).
2. Powder x-ray diffraction
The successful formation of HOIP crystallites was further validated with XRD analysis. Diffractograms seen in Fig. 3 reveal the presence of precursors in the composite material and a broadening caused by the PS. In addition to the precursor and polymer peaks, the composite diffractograms also demonstrate the characteristic peaks of tetragonal CH3NH3PbI3.12 It was desirable to further verify the presence of CH3NH3PbI3 due to the unprecedented high temperatures experienced by both the precursors and HOIP during synthesis in the polymer melt.
(Color online) Powder x-ray diffractograms of CH3NH3I (MAI), PbI2, poly(styrene), and the electrospun composite. All materials were powdered in a mortar and pestle before measuring. The composite diffractogram exhibits some of the precursor peaks indicating the synthesis did not progress to completion. However, the composite diffractogram also demonstrates peaks at 14.2°, 28.4°, and 31.9° corresponding well with the 110, 220, and 114 peaks of tetragonal CH3NH3PbI3 (Ref. 12).
(Color online) Powder x-ray diffractograms of CH3NH3I (MAI), PbI2, poly(styrene), and the electrospun composite. All materials were powdered in a mortar and pestle before measuring. The composite diffractogram exhibits some of the precursor peaks indicating the synthesis did not progress to completion. However, the composite diffractogram also demonstrates peaks at 14.2°, 28.4°, and 31.9° corresponding well with the 110, 220, and 114 peaks of tetragonal CH3NH3PbI3 (Ref. 12).
Temperatures experienced in the melt chamber exceed the temperatures established by Philippem et al.13 and Dualeh et al.10 as the onset temperature of thermal degradation of HOIPs, which are reported to be 100 °C in ultrahigh vacuum, and 234 °C in a N2 environment, respectively. The successful formation of HOIP crystallites in the melt chamber suggests that the rate at which the organic precursor and degradation products can volatilize is fundamental to driving thermal degradation of both the precursors and the HOIP. Surrounding the PbI2 crystallites with the polymer melt as they react to form HOIP effectively limits the volatilization of the degradation products allowing synthesis at a much higher temperature. The successful limitation of the sublimation of the organic precursor and the volatile degradation products can be attributed to high vapor pressure in the melt chamber described by the Clausius-Clapeyron relationship in the following equation:
where p is the vapor pressure of the system, ΔHsub is the energy required to sublimate, R is the gas constant, T is the temperature of the system, and Tsub is the onset temperature of sublimation.
The molar ratio of 2:1 (CH3NH3I: PbI2) allows for the sublimation of CH3NH3I while maintaining the necessary concentration to react with the PbI2 in the melt. As the CH3NH3I sublimates and the piston in the melt chamber is driven downward, the headspace is quickly saturated with gaseous CH3NH3I limiting further sublimation by increasing the energy requirement for sublimation, assuming the temperature of sublimation is a constant. As the gaseous CH3NH3I is drawn into the HOIP reaction, it is quickly replaced by the solid excess CH3NH3I in the melt chamber. It is proposed that the gaseous CH3NH3I limits the volatilization of the both the precursor CH3NH3I as well as the organic component of the HOIP through increasing the enthalpy of sublimation of both species. The gaseous CH3NH3I has also demonstrated a fast rate of diffusion in the polystyrene melt which facilitated the formation of the HOIP about the PbI2 crystallites. The movement of gaseous CH3NH3I through the melt electrospinner is further supported by the observation of a white smoke emitted from the spinneret that was not observed in the course of melt electrospinning PS only, or PS doped with PbI2 microfibers. Diffusion of the gaseous CH3NH3I through the polymer matrix facilitated by the pressure gradient14 from the head space to the spinneret of the electrospinner may be applicable in other systems with similar polymeric overlayers. Utilization of over-layers to slow the volatilization of degradation products would also allow for higher temperature annealing of semiconducting thin perovskite films, thereby facilitating growth of larger grains in HOIP films. Larger grain sizes would further mitigate degradation through minimization of grain boundaries at which degradation occurs. Furthermore, larger grains would lead to better optical performance of the HOIP material as well as improving the performance of HOIP solar cells through improvement of the charge transport properties15 and allow for a greater electrical continuity in the active layer improving the efficiency of devices.
B. Relationship between crystallite size and fiber morphology
1. Light microscope images
To elucidate the effects that each precursor would have on fiber morphology, initial M-ES attempts included precursors individually. The morphological properties of electrospun fibers are critical to the design of efficient interfaces for charge extraction, in addition to consistent optical properties in solar cells fabricated using the composite. A photomicrograph of an early attempt at melt electrospinning can be seen in Fig. 4.
(Color online) Photomicrographs of (a) CH3NH3I and (b) PbI2 crystallite in melt electrospun poly(styrene) fibers. Initial attempts produced large fibers but demonstrated how the CH3NH3I phase separates and the PbI2 precursor affected the diameter and shape of the electrospun fiber in image (b).
(Color online) Photomicrographs of (a) CH3NH3I and (b) PbI2 crystallite in melt electrospun poly(styrene) fibers. Initial attempts produced large fibers but demonstrated how the CH3NH3I phase separates and the PbI2 precursor affected the diameter and shape of the electrospun fiber in image (b).
The photomicrograph in Fig. 4 demonstrates that PbI2 crystallites result in bulging of fiber diameter approximately equal to the size of the crystallite itself. In the cases when the crystallite possessed at least one dimension larger than the diameter electrospun fiber, a pooling of PS occurred at the spinneret resulting in an “electrodripped” section of substantially larger diameter. The effect was exclusive to the PbI2 precursor and was not observed with the CH3NH3I, most likely because CH3NH3I existed in the liquid/gaseous phase at the spinneret.
2. Scanning electron micrographs showing precursors fibers
The effect each precursor had on the morphology of electrospun fibers was further studied using SEM-EDS. The elemental contrast overlaid on the electron micrograph allowed for the simultaneous analysis of both the morphology of electrospun fibers as well as the distribution of precursors in fibers. The morphology and chemical phase domains of the electrospun fibers can be seen in Fig. 5.
Electron micrographs of melt electrospun polymer fibers the in (a) PbI2 only, (b) CH3NH3I only, and (c) synthesized HOIP electrospun microfiber experiments. HOIP and PbI2 fibers demonstrate similar morphologies with crystallites embedded in the fiber. Whereas the CH3NH3I appears to have phase separated from the poly(styrene) upon cooling evidenced by the white powder external to the fiber in (b).
Electron micrographs of melt electrospun polymer fibers the in (a) PbI2 only, (b) CH3NH3I only, and (c) synthesized HOIP electrospun microfiber experiments. HOIP and PbI2 fibers demonstrate similar morphologies with crystallites embedded in the fiber. Whereas the CH3NH3I appears to have phase separated from the poly(styrene) upon cooling evidenced by the white powder external to the fiber in (b).
The morphology of the fiber as well as the size of HOIP crystallites produced are determined by the size of PbI2 crystallites initially added to the melt chamber, evidenced by the micrographs in Fig. 5. The phase separation of the CH3NH3I seen in Fig. 5(b) is ideal for fabrication purposes. The phase separation of the CH3NH3I from the PS will allow it to be used in excess in the polymer melt feed stock, as it will solidify outside the PS microfiber and can be removed with a simple solvent rinsing step after the electrospinning process. Excess CH3NH3I in the feedstock will allow the PbI2 to be the limiting reagent in the perovskite reaction minimizing the PbI2 remaining in the resultant fibers. Furthermore, the location of the HOIP crystallites in the fibers is conducive to mitigation of moisture driven degradation. Electrospun PS fiber mats have demonstrated superior hydrophobicity when compared to PS thin films.16 The increased hydrophobicity of the electrospun fiber mats should minimize the interfacial area between condensed water and the electrospun fiber mat limiting permeation of the water into the fiber superstructure.
IV. CONCLUSIONS
Alternative synthesis methods, modification of pre-existing material properties, and implementation of novel materials in perovskite solar cells are vital in furthering the development and eventual mid-to-large-scale deployment of this class of photovoltaic technology. Current research is directed at the development of synthesis techniques that produce photoactive HOIP materials to be incorporated into the pre-existing solar cell device architecture with cell lifetime sufficiently long enough to see a return on manufacturing costs. The work presented here represents a novel synthesis method paired with a scalable manufacturing technique demonstrated for the first time. Results demonstrate a successful incorporation of HOIPs into polymer fibers through utilization of melt electrospinning. Future work will focus on modifying electrospun fibers for efficient charge extraction and incorporation of HOIP/PS composite materials into functional solar cells, as well as accelerated ageing studies to quantify the improved stability, if any, in the composite HOIP/PS microfibers.
ACKNOWLEDGMENTS
The authors would like to thank Nicole E. Zander and the members of the U.S. Army Laboratory Strategic Material Command for providing and quantifying molecular weight of poly(styrene) for melt electrospinning. Research was sponsored by the Army Research Laboratory and was accomplished under Cooperative Agreement No. W911NF-15-2-0020. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Laboratory or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein.