| Author: | Lo, Kin Shing Kenneth |
| Title: | Incorporation of shear-exfoliated graphene for dye-sensitized solar cell |
| Advisors: | Leung, Woon Fong Wallace (ME) |
| Degree: | Ph.D. |
| Year: | 2020 |
| Subject: | Dye-sensitized solar cells Solar cells Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Mechanical Engineering |
| Pages: | xxxiv, 140 pages : color illustrations |
| Language: | English |
| Abstract: | Dye-sensitized solar cells (DSSCs) are an emerging class of solar cells that has great promise. The advantages of DSSC are low-cost, non-toxic, flexible, semi-transparent, and could come in any shape, sizes, or colors. The typical DSSC uses a photoanode comprised of a 5 μm to 15 μm thick layer of mesoporous titanium dioxide (TiO₂) nanoparticles of approximately 20 nm in size. However, the current efficiency of just greater than 14 % for small devices is still too low to be viable for commercial use. The low efficiency arises from various bottlenecks that occurs in conventional DSSC. To lift the restrictions on DSSC, three areas are specifically in this study: the transport of electrons in photoanode, the surface traps of TiO₂, and the inefficient transport of triiodide ions in replenishing the oxidized dye. In the first part of our research, to address the inefficient electron transport, the conventional TiO₂ nanoparticles are replaced by the TiO₂ electrospun nanofibers to confine diffusion to one-dimensional (1D) transport. The restricted freedom of movement prevents electrons from straying as in transporting between TiO₂ nanoparticles. However, the transport mechanism of electron hopping and diffusion in TiO₂ nanoparticles is rather ineffective. Due to the slow transport, electrons frequently get absorbed by triiodide (I3 -) ions in the electrolyte, or recombine with the oxidized dye, which results in a loss of current and overall power conversion efficiency (PCE) in DSSC. The approach to solve the transport problem is to introduce highly conductive material inside the nanofibers, allowing the photogenerated electrons to be transported efficiently through the nanofibers to the fluorine-doped tin oxide (FTO) anode, and extracted to power an external load. Pristine, defect-free graphene are ideal carrier forcharges but are not commonly used in composite as it tends to aggregate in solution. Commonly used reduced graphene oxide (rGO) circumvents the problem by starting with a suspendable precursor. However, the defect-filled rGOs are not as conductive and the defect sites promote recombination. Carbon nanotube (CNT) has been combined with TiO₂ to form photoanode but the combination leads to discontinuous nanorods. Ideally, the nanofibers should be continuous for better conductive pathways. Recent progress in graphene production has found shear-exfoliation of graphene with surface-binder can produce defect-free graphene suspensions in large quantity. For the first time self-made, shear-exfoliated graphene are incorporated into electrospinning of TiO₂ nanofibers for photoanodes used in DSSC. Through the use of pristine graphene instead of rGO the photoanode introducing defects that serves as recombination sites. In addition, electrospinning caused the graphene to be rolled-up within a layer of TiO₂ further isolate graphene edges from the electrolyte. Polyvinylpyrrolidone (PVP) is chosen as the sacrificial polymer for its ability to bind with graphene surfaces and its common usage in electrospinning of TiO₂ nanofibers. Shear-exfoliated graphene are produced as a suspension, which is centrifuged with high centrifugal force at 'optimal centrifugation time' to control the size and concentration of graphene sheet to be incorporated in the nanofibers during electrospinning. At low centrifuge time, graphene flakes tend to unravel during calcination, destroying the nearby fibers. On the other hand, long centrifugation time leads to small graphene flakes that requires more electron hopping before extraction, potentially making transport less efficient. Through experimentation the optimal graphene processing condition for combining with TiO₂ nanofibers has been determined. The amount of graphene in the optimal TiO₂/graphene (TG) nanofibers is too small to be quantified with thermal gravimetric analysis (TGA), but the presence of graphene are confirmed through Raman spectroscopy. With the optimal graphene processing, photoanodes produced with TG nanofibers can reach 23 μm in thickness while still retaining good extraction, much thicker than the optimal thickness of 13 μm with TiO₂ nanofibers lacking graphene. The new thicker photoanode has the advantage of capturing a greater portion of incident light on DSSC even without the assistance of a scattering layer, which produces little current on its own. The higher charge generation increases PCE to 8.9 %, which is 22 % higher than that with TiO2 nanofibers. The embedded graphene, completely covered by TiO2 anatase crystals, shows no ill influence on dye-loading of N719 dye, and only enhances electron transport. Despite the increase in conductivity, the electrons still need to travel through trap-filled TiO₂. The surface defects of TiO₂ acts as electron traps that slows electron extraction, leading to higher recombination. Surface traps of TiO₂ are addressed through the incorporation of zinc oxide (ZnO) along with graphene in TiO₂ nanofibers. TiO₂ and ZnO are combined first to test the effects of ZnO additives on TiO₂ nanofibers. As the precursor of TiO₂ and ZnO are mixed in the electrospinning solution, co-doping or segregation may occur while processing the nanofibers. Normally, ZnO is coated on the surface of TiO₂ for passivation but ZnO is discovered to be capable of passivating traps of even without surface coverage. Despite mixing two different semiconductor-precursors, no co-doping occurs and the rendered TiO₂/ZnO (TZ) nanofibers have both crystals uniformly distributed. The passivation did not hinder dye-loading despite surface-conditioning. As no physical barrier exist, ZnO also did not interfere with electron injection. By passivating traps, diffusion in TZ nanofibers have less chance of getting capture and electron diffusion is greatly enhanced relative to just TiO₂ nanofibers. In addition, ZnO also does not introduce any significant recombination pathways for electrons in the TiO₂. The optimal amount of ZnO is first optimized with TiO₂ nanofibers without graphene. The PCE of TZ nanofiber photoanode improves to 7.33 % from a baseline of TiO₂ nanofibers at 6.05 %, a 21.2 % increase. However, once graphene is added at the second stage into TZ nanofibers, a coupling effect between graphene and ZnO decreases the fill factor (FF), which lowers the PCE despite an increase in short-circuit voltage (jSC). While the iodine-based electrolyte has been the standard for DSSC, the changing landscape of electrolyte has shifted away to better support their respective dyes. The various redox charge carriers generally are larger and therefore lower mobility than the gold standard (iodide/triiodide) and the previous enhancing chemical additives may no longer apply to the new dye-electrolyte combination. Even iodide/triiodide ionic transport is not high, resulting in inefficiency with regenerating the oxidized dye. The inefficient transport is addressed by self-made, electrospun graphene nanotubes (GNTs), which are open-ended nanotubes made of rolled-up graphene sheets. In nanofluidics, when ions passed through a nanochannel made of sp2 carbon material, e.g. CNT, the particles will speed through the tunnel due to electrokinetic phenomena. However, common CNT are small diameter, short length, and may not be open-ended to form a tunnel. With GNT, there is now a simple, inexpensive, long, and adjustable diameter sp2 carbon channel that can be used as a non-species dependent additive to enhance ionic transport of any electrolyte. GNTs are produced by electrospinning, followed by calcination to remove the PVP surface binder. A short centrifugation time under high speed is chosen to removed not-exfoliated graphites and ensure nanofiber connectivity. Nanotubes may fail to form after calcination if the centrifugation time is too long. None of the observed GNT has been found to be closed at either ends in field emission scanning electron microscopy images, and all suggest the graphene flakes are rolled up into a tube-like structure. As open-ended nanotubes, GNT utilizes electrokinetic effect and enhances diffusion to greatly increase the carrier mobility across the electrolyte. Separate tests done under alternating current (AC) and direct current (DC) shows a transport rate increase by 3-5 times. Relaxation test from a steady-state current reveals diffusion rate to increase by 2-3 times. While GNTs are dispersed within the electrolyte, it may influence the bordering photoanode, such as enhancing recombination. Luckily, test reveals GNT are practically inert to the photoanode. Short-circuiting occurs when GNT concentration is above the percolation threshold, e.g. 10 mg mL−1 for 97-nm diameter GNT. Under the percolation threshold, GNT facilitates the ionic transports of both iodide and triiodide through electrokinetic phenomena to move towards the counter-electrode, where the triiodide ions are reduced to iodide ions. The increase iodide (I-) concentration fuels a strong diffusion of iodide ions from the counter-electrode toward the photoanode, regenerating the oxidized dye. The strong diffusion is also further enhanced by the GNT. With faster transport of both I- /I3- redox pair in the electrolyte, the GNT results in higher jSC. As a result, the performance of DSSC rises from 6.96 % to 8.66 %, a 24.4 % increase for samples using 97-nm diameter GNT at 10 mg mL−1 nanotube concentration in electrolyte. Through the incorporation of self-made shear-exfoliated graphene with electrospinning, pristine defect-free graphene could be embedded within a nanofiber or used for GNT production. The electron diffusion length is enhanced by the inserted graphene with little to no influence on recombination. TiO2/ZnO/graphene (TZG) nanofibers is demonstrated to work in tandem to increase electron extraction, with ZnO passivating the trap states and graphene increasing photoanode conductivity. However, a negative coupling between ZnO and graphene lead to a reduction in performance with TZG nanofiber photoanode even though ZnO or graphene enhancement alone does not influence electron lifetime. With GNT suspended in electrolyte, the ionic conductivity is greatly enhanced through electrokinetic effect and diffusion without affecting photoanode performance. The various bottlenecks of DSSC are addressed and applicable to conventional DSSC. The above results are actually more general than it appears. Graphene can be incorporated into the nanofibers to improve electron transport of any material. Even if TiO2 gets replaced, or p-type DSSC is being explored, pristine graphene embedded nanofibers will still have an enhanced conductivity relative to its pure counterpart. As a physical enhancement, GNT is indifferent to the choice of redox mediator. GNT can be incorporated in any electrolyte to improve the ionic transport, accelerating any redox reactions. The non-species-specific improvement allows separate optimization between dye-electrolyte compatibility and ionic conductivity. Together, DSSC containing electrospun shear-exfoliated graphene can support a much higher current than conventional DSSC |
| Rights: | All rights reserved |
| Access: | open access |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/11043