Proton exchange membrane (PEM) fuel cells offer an alternative as an efficient power source with low environmental impact. The heart of the fuel cell is the membrane, which conducts protons through an aqueous channel network. Proton transport is critically tied to the channel connectivity – disconnected channels do not participate in the overall electrochemical activity of the cell. Nafion, the current benchmark PEM, is a random statistical copolymer, characterized by a percolating network of cylindrical channels. In previous work, conductive-probe atomic force microscopy (cp-AFM) was used to image the conductance of Nafion. Although cp-AFM provides relevant information on which channels are connected, it provides no information on the disconnected “dead-end” channels at the surface. Electrostatic Force Microscopy (EFM) was used to analyze the structure and frequency of the “dead-end” channels. Applying a simple parallel plate model allowed us to assign differences in the EFM signal to particular channel shapes: connected cylindrical channels, “dead-end” cylinder channels, and bottle-neck channels.

Anion exchange membranes (AEMs), which conduct hydroxide, have attracted recent interest due to improved reaction kinetics in alkaline media, yet suffer from low conductivity and easily degrade. To this end, our AFM methodology was applied to analyze the channel connectivity of a commercial FumaTech AEM was investigated in its hydroxide form over a wide range of relative humidity (RH) by combining phase imaging and cp-AFM. At high RH, our AFM data indicates significant surface swelling. Lastly, we investigated a class of phosphonium-containing diblock copolymer AEMs that formed ordered morphologies. Although channels were observed to be well-connected in the bulk by TEM, channels aligned parallel at the surface leading to many “dead-end” channels shown by EFM. Correlating these findings with bulk measurements could offer insight toward AEMs with improved conductivity and chemical stability.

Protein misfolding has been implicated in a number of neurodegenerative diseases, including Alzheimer’s Disease (AD), Parkinson’s Disease, and amyotrophic lateral sclerosis (ALS). In these diseases, one or more proteins adopt a non-native conformation that is prone to aggregate into small soluble oligomers and larger fibrils, which typically present as abnormal protein deposits in the afflicted cells or tissues. Understanding the aggregation of these proteins is essential to the development of effective disease-modifying therapeutics. TAR DNA binding protein of 43 kDa (TDP-43) is one such protein that is believed to play a causative role in sporadic ALS and tau- and alpha synuclein-negative frontotemporal lobar degeneration (FTLD-TDP), and is also observable in a significant percentage of individuals with AD, where it is associated with increased cognitive decline. Atomic force microscopy (AFM) was used as a complementary technique to ion mobility-mass spectrometry (IM-MS) to elucidate the early-stage aggregation of the amyloidogenic core region of TDP-43, TDP-43(307-319). TDP-43(307-319) aggregation was shown to progress via a bifurcated pathway, with toxic oligomers off pathway from fibrillization. Two ALS-related mutations of TDP-43 increase aggregation while a synthetic non-toxic mutant prevents oligomerization but does not suppress fibrilization. Similar experiments with a fragment of the familial ALS-related protein superoxide dismutase-1, SOD1(28-38), show that the formation of ordered, non-fibrillar aggregates was tied to the presence of a potentially toxic corkscrew oligomer. TDP-43(307-319) aggregation was also modulated using two methods. First, computationally generated inhibitors molecules developed using the Join Pharmacophore Space (JPS) were shown to disrupt oligomers believed to play a toxic role in ALS but were not observed to impact off-pathway fibrillization. Further refinement of the JPS algorithm will be required if fibril disruption is shown to be necessary for effective treatment of ALS. Next, TDP-43(307-319) was co-aggregated with a fragment of the AD-implicated protein amyloid beta, Aβ(25-35). Coaggregation of TDP-43(307-319) and Aβ(25-35) results in increased formation of toxic oligomers for both peptides. Coaggregation was also conducted with a mutation of TDP-43(307-319) that suppresses toxic oligomers, and potentially toxic Aβ(25-35) oligomers were observed to increase fibrilization of TDP-43. These results indicate that cross talk between amyloidogenic proteins must be considered in the development of therapeutics for related diseases.

Vanadia catalysts supported on TiO2 belong to a class of mixed-metal oxides capable catalyzing the oxidative dehydrogenation (ODH) of alkanes into high commodity olefins. As olefins are central to many important industrial processes and their demand expected to increase, improving the reactivity and yield of vanadia catalysts has become an active area of research. Supported vanadia display unique catalytic properties not observed in their pure oxide counterparts; this unique reactivity and selectivity has been attributed to the complex local bond-coordination present at the interface. Thus, understanding the role of each metal oxide has on the reactive interface is critical to designing better catalysts.

Previous investigations into vanadia catalysts have found the activity for methanol ox- idation to formaldehyde increases with reducibility (sub-stoichiometric) of the support, however the nature in which support defects contribute to this synergy is unknown. This is due to the fact that the support is typically oxidized in the growth of vanadia catalysts. We circumvent this problem by soft-landing mass-selected cationic VO+ clusters onto single crystal rutile TiO2 110) - (1 × 1) surfaces to create well-defined model in-terfaces. By integrating temperature programmed desorption/reaction (TPD/R) studies with variable temperature scanning tunneling microscopy (VT-STM), we can correlate ensemble surface reactivity with sub-nanometer structures, reserved here for VO clusters and surface point-defects. First, we identify sub-surface Ti interstitial defects in bare reduced TiO2-x as the origin of a new methanol disproportionation reaction to methane and formaldehyde. These Ti interstitial defects are implicated as electron donors and thus contribute to the reduction of methanol. We find that the inclusion of mass-selected VO clusters results in a VO-mediated disproportionation reaction to formaldehyde and methanol at high temperature. Our STM results confirm the VO-mediated disproportionation reaction pathway by observation of a VO-(OCH3)n precursor intermediate. Furthermore, our methanol TPD/R and STM results indicate methanol adsorbs favor- ably over VO and five-fold coordinated Ti sites in-lieu of surface bridging oxygen vacancy defects and Ti interstitial defects. We propose VO clusters scavenge electrons from Ti interstitial and bridging oxygen vacancy defects, resulting in an effectively oxidized surface. The VO/TiO2 interface investigated here represents an interesting model towards understanding the synergistic effects of mixed-metal oxides catalysts.

The study of the oxidative dehydrogenation of methanol to formaldehyde by vanadia/TiO2 model catalysts has received a great deal of interest by the scientific community in recent years. However, due to the wide variety of catalyst preparation methods, there have been conflicting reports with formaldehyde production occurring anywhere from room temperature to 660 K. This is further complicated by the fact that the methods used often couple the oxidation state of the TiO2 support to that of the vanadia, creating many structures on the surface with uncertain stoichiometries.

In order to overcome these challenges, we have studied the oxidation chemistry of mass-selected Vx and VxOy clusters deposited on TiO2 (110) surfaces as a function of size and composition of the cluster as well as the oxidation state of the support. The cluster-decorated surface was characterized by scanning tunneling microscopy (STM) to determine the structure of the active catalyst and the chemistry was probed using temperature-programmed reaction (TPR). Density functional theory (DFT) was used to calculate the structure of the clusters on the surface and compared with experimental results.

We have shown that both V2O6 and V3O9 clusters are highly active in the oxidative dehydrogenation of methanol. V2O6 clusters catalyze the production of formaldehyde and methyl formate, the latter representing a previously undiscovered reaction path in the reaction of methanol and oxygen on supportedvanadia catalysts. Structural models from DFT and STM images have allowed us insight into the mechanism of this reaction. The models indicate that on a

stoichiometric surface, the V2O6 cluster contains a peroxyl group that is active for the production of methyl formate.

TPR has revealed that the production of methyl formate is highly dependent on the oxidation state of the surface. For the case of V2O6 clusters on the reduced surface, only formaldehyde production is observed. STM results also show a variation in the cluster structure as a function of the oxidation state of the TiO2, with the clusters on the reduced surface containing only vanadyl oxygen atoms due to the interaction with surface oxygen vacancies to dissociate the peroxyl group. The vanadyl oxygen has been previously shown to be active for production of formaldehyde by examining post-oxidized size selected VO clusters. In addition, we show that oxidative dehydrogenation of methanol on V3O9-decorated TiO2 surfaces results in methyl formate production at a lower temperature, and formaldehyde production with a significant increase in signal over that observed for V2O6 clusters.

Nafion represents the most commonly employed and well characterized proton exchange membrane (PEM) used in fuel cells, however structural models which explain its physical properties are oversimplified and incomplete. In this work we use a combination of tapping mode and conductive probe (cp) AFM to investigate the nanoscale morphology and proton conductivity of Nafion. By conducting imaging at a wide range of relative humidity, we were able to characterize Nafion from extensively dehydrated to liquid equilibrated states. Rather than an even swelling of hydrophilic clusters, we see an uneven dehydration of the surface at low hydration and a rearrangement to form a network of cylindrical inverted micelles at high water content. A statistical analysis of these features allowed us to match them with previous x-ray scattering data and develop a model of Nafion which is valid at large length scales and consistent over a wide range of water contents.

Comparison of this model system to new membrane materials has allowed us to better understand the structure-property relationship in these systems. We next investigated Hyflon, a short side chain analog of Nafion and found that in contrast to Nafion it is able to swell reversibly through a dilation and contraction of hydrophilic clusters with minimal rearrangement and retain more hydrophilic character at the surface at dehydrated conditions. Investigation of new multi-acid sidechain membranes was also conducted and showed that while these membranes have excellent water retention and proton conductivity, they swell and form a continuous hydrophilic phase at the surface which may be problematic during fuel cell operation. In addition to these characterization efforts, a number of advances to the cp-AFM technique were demonstrated which allowed for the investigation of new operating conditions and the extraction of new information on electrical and mechanical properties of the membrane.

In this research we present the novel technology that was developed using low-cost materials and 3D printing to allow for in operando nanoscale investigations of perfluorosulfonic acid membranes under vanadium oxygen fuel cell (VOFC) and direct methanol fuel cell (DMFC) conditions, using conductive atomic force microscopy. We detail the process of designing, iterating, fabricating, and validating our novel 3D printed imaging cell. The justifications for the design choices are explained in detail. Heating and relative humidity control are added to the cell as additional features for more accurate environmental simulation. Using the cell, we image Nafion membranes under VOFC conditions and compare our findings with previous research. We found that thinner membranes show higher current densities than thicker membranes. Furthermore, under VOFC conditions we observe the same shrinking behavior observed in previous research as the ionic strength of electrolytes that Nafion is exposed to increases. We imaged Nafion membranes under DMFC conditions and found that the operating temperature and anode catalysts play an important role in observing nanoscale proton conductivity. Finally, we detail the process of synthesizing low-valency vanadium/sulfuric acid electrolytes and end with an investigation into VO/TiO2 catalysts using density functional theory. This investigation helps to explain the atomic processes occurring on a model catalyst system that can catalyze the oxidative dehydrogenation of methanol to formaldehyde.