Augmented Microfluidic Dye Sensitized Solar Cells - Longevity and Performance Efficiency Investigation In Novel Architectures Through Synergizing Materials and Microfluidics

Abstract

Dye-sensitized solar cells (DSSCs) have garnered increasing attention from scientists and engineers due to the potential of low cost and flexibility in terms of choice of constituent materials, along with increasing emphasis on developing renewable energy technologies. Investigations into the individual components of the device reveal various limitations which effect the performance and longevity of the device thus limiting its capability to compete as a viable alternative to traditional photovoltaic systems. Of these limitations, general consensus shows that the electron collection mechanism between the the photosensitized metal-oxide and the electrolyte where electron diffusion is a relatively slow process, creates a bottleneck in the component interplay. In addition, the photoactive dye molecules adsorbed to the surface of the metal-oxide are prone to degradation whilst exposed to light sources for extended periods of time. The former has a direct impact on immediate performance whilst the latter limitation prevents the device from functioning at full capacity after various periods of time thus limiting longevity. Both of these limitations must be overcome in order for DSSCs to become competitive in a commercial market. Based on the nature of these limitations, the introduction of microfluidic capability into various DSSC architectures would directly address the slower transport and long-term stability issues. The microfluidic circulation of electrolyte introduces a complementary transport mechanism, convection, to help alleviate the bottlenecked diffusion. If a DSSC was capable of sustaining microfluidic electrolyte flow, it simultaneously also has the capability to receive payloads of fresh dye molecules to supplement or replace the older, degraded dye. In this study a fabrication workflow was developed that would allow microfluidic flow to be integrated into TiO2 and ZnO nanowire (NW) based DSSCs. The workflow included optimizing film-deposition techniques, NW synthesis, photolithographic microfluidic channel imprinting and implementation of a sealing mechanism. Of these, the synthesis of ZnO NWs was an important juncture in the study. ZnO NWs offer the potential to exponentially increase the amount of dye-loading sites based on the scale of their density and 3D geometry alongside favorable conductive properties when contacted with electrolyte. Therefore the study explored their optimization in terms of their growth with respect to control over precursor concentration, growth time and effect of synthesis additives. The most optimal concentration for the precursors used in NW synthesis was tuned over 2.5-100 mM for the zinc and hydroxide sources. At the lower end of the spectrum, the concentration was not high enough to create a large number of nucleation sites on a ZnO thin-film and thus resulted in the growth of single-crystal structures in a non-oriented manner. The hydrothermal synthesis performed at 25 mM resulted in a well-oriented, vertii cal, dense array of uniform ZnO NWs, characterized by scanning-electron microscopy (SEM). The density improved at this concentration with 66 million nanowires per square cm. When the concentration was increased further, the uniformity of the size distribution widened resulting in nanostructures with diameters ranging from 70-140 nm with heavy precipitate. The solution at 100 mM was highly supersaturated and resulted in a more common poly-crystal structure rather than an array of NWs. Additionally to concentration, longer growth times were found to be associated with longer NW arrays. An activity limit was found for the precursor after 2 hrs where growth would reach a threshold and thus replacing the growth solution with fresh precursors every 2 hrs was the most optimal strategy to maintain a linear growth rate. Consistent synthesis conditions with new precursor introduced every 2 hrs led to higher, average NW lengths at 1.69 µm with an average length increase of 0.42 µm every 2 hrs. The aspect ratio of the NWs increased from 14 to 28 between the 2 and 4 hr mark however at the final time of synthesis, decreased to 25 indicating that growth in the lateral direction was starting to become less inhibited. To inhibit lateral growth, polyethyleneimine (PEI) was used as a capping agent for the lateral NW faces. The use of PEI reduced the NW diameter by 30% and the aspect ratio increased 26.5%. Both TiO2 and ZnO NWs were tested in the same DSSC configurations where the effect of microfluidic electrolyte circulation was investigated. The effect of electrolyte flow rate on cell performance was first investigated via IV characterization of the DSSCs. TiO2 and ZnO NW DSSCs were systematically characterized with static electrolyte and then with increasing flow rates from 0.5-10mL/min. The fabrication of the devices and the clamping mechanism mitigated fluid leakage during circulation. With the addition of a convective transport mechanism via microfluidics, the short-circuit current (Jsc) increased as flow rate was elevated. Between static electrolyte and a max flow rate of 10 mL/min the difference in current was 38% for TiO2 and 13% for ZnO NW based DSSCs. Increased conductivity also appears simultaneously alongside more efficient transport where iodide diffusion from the bulk into the metal-oxide network and the triiodide production via oxidation at the counter-electrode is aided by rapid, convective transport. The consistent increase in Jsc correlated directly with higher efficiencies and max power of the devices. Photolithographic manipulation of the anodic surface by addition of microfluidic channels was also investigated. The addition of channel flow limits the available active area by addition of a photoresist or PDMS intermediary layer between anode and cathode, but concentrated delivery of the electrolyte aids in electron transport. Linear, series and parallel flow channels were implemented and showed that despite active area being limited, both TiO2 and ZnO NW cells were able to perform consistently alongside other open architecture constituents. By implementing microfluidic channels where the flow is controlled over a specified geometry, the efficiencies reached peaks of 6.09% for TiO2 and 1.04% for ZnO NW based DSSCs. The ZnO NW DSSCs have an advantage over the TiO2 based DSSCs because the length of the NWs can be controlled during synthesis. The mesoporous TiO2 film cannot be scaled ii vertically as the nanoparticle network will surpass the diffusion length of the electrons thus limiting the transport further. ZnO NWs are unique as their geometry provides direct electrical pathways for electrons which is further exploited by the addition of convection from microfluidic flow. Hence the last microfluidic electrolyte trial studied the effect of ZnO NW length on device performance. The Jsc of the longest NW array at 2.5 µm showed the highest photocurrent at 1.0 mA/cm2 , which doubles the Jsc measured for 1.5 µm NWs. The nanowire length directly corresponds to device performance thus creating the necessity for aspect ratios to be greater than 40 for viable cells. Aside from the need for higher aspect ratios, the ZnO NW DSSCs performed poorly compared to the TiO2 based devices due to higher shunt resistance in most of the cells. The shunt could potentially manifest from lower adsorption rates of dye onto the ZnO NW scaffolding. Despite the lower efficiency in the ZnO NW devices, both architectures showed a propensity to perform better when convection was introduced via microfluidic electrolyte. The last set of trials performed dealt with investigating how well the DSSCs would perform after intense ultra-violet degradation. The devices were degraded and then tested after a payload of fresh dye molecules was delivered to the metal-oxide layer. Current-voltage (IV) characterization showed that after degradation and regeneration, microfluidic electrolyte raised the efficiency of the DSSC of TiO2 cells by 2.2%. The study revealed that simple circulation of new dye was not sufficient as the degraded dye was not being desorbed. After stripping the degraded dye through a basic/acidic wash cycle a new payload of dye was introduced which resulted TiO2 cells being able to recover more than 100% of the original photocurrent. ZnO NW cells only partially recovered 83% of the pristine condition photocurrent. When the activation step was not performed the cells were not nearly efficient after 1 hr of degradation with only 24% (TiO2) and 35% (ZnO NWs) less than the original test. The addition of microfluidic capability to traditional and novel TiO2 and ZnO DSSC architectures shows that convective transport through microfluidic electrolyte consistently boosts performance when compared to static electrolyte. Microfluidic integration offers the added benefit of dye regeneration techniques which can be employed to mitigate degenerative limitations, previously unavailable for static DSSCs.