UC Berkeley

Waste heat recovery from the human body provides opportunities to power electronics with a source that is cheap and readily available. Thermoelectrics utilize the Seebeck effect to recover useable electrical energy from this waste heat, but are limited due to material parameters being inversely coupled in the bulk. We investigate the role of novel physics at interfaces in order to develop new fundamental understanding of thermoelectrics. The goal is to discover systems where the Seebeck coefficient and the electrical conductivity are not inversely correlated. We investigate thermoelectric transport in organic-organic systems such as scanning tunneling microscope molecular break junctions on the nanoscale, gold nanocrystal arrays on the mesoscale and polymeric ion and mixed conductors at the macroscale.

The STM molecular junctions studied in this work can provide design rules to positively couple the Seebeck coefficient and the electrical conductance. Since STM molecular junctions are one-dimensional systems, by minimizing the gap between the molecular orbital energy level and the electrode Fermi energy, the power factor S2σ can be optimized. I built a toolbox of chemical structures by first understanding the role of the interface coupling to alkylthiol binding groups of thiophene-based molecules. With this understanding, I designed small molecules based on the monomer unit of donor-acceptor polymers and other conductive polymers. Molecules with very high HOMO levels or low LUMO levels were studied, and the corresponding energy levels were examined using spectroscopic techniques.

I then present our work on scaling these molecular junctions to the macroscale using ligand-exchanged gold nanocrystal arrays. Beginning with a model system of alkanethiols and alkanedithiols, I show that the electrical conductivity scales with ligand length exactly as observed in single molecule junctions, and the Seebeck coefficient follows a similar trend. By showing that gold nanocrystal arrays are a good model platform to scale molecular junctions to the macroscale, I aim to bring the same Landauer transport physics observed in single molecule junctions to macroscale devices. I further investigate the role of different binding groups, where I find that the character of the ligand binding group directly affects whether the sign of the Seebeck coefficient of the overall array.

Finally, I elaborate on our work with conductive polymers as thermoelectric materials, and specifically investigate the thermoelectric effect in polymers that are solely ion conductors or are capable of both ions and electrons transport. I demonstrate that thermoelectrics based ionic or mixed conduction can be well-described using the developed fundamental equations, and by using these rules we develop positive or negative Seebeck coefficient materials. I present our work on the newly developed thermoelectric AFM technique and how we use the thermoelectric response of different domains to understand the morphology of PEDOT:PSS. By fully understanding the thermoelectric properties of organic-inorganic hybrids from the monomer level, we can extend this same understanding to develop polymer thermoelectrics.