Enzymatic fuel cells (EFCs) are bioelectronic devices based on redox enzymes, which convert chemical energy into electrical energy via biochemical reactions. A major difficulty to overcome is to successfully connect (using e.g., immobilization) the enzymes to the electrode surface. Since the immobilization process often stabilizes the enzyme, the electrode surface and the enzyme/electrode interface is of utmost importance for both the efficiency and stability of the EFC. In this work several different means of establishing the enzyme/electrode connection have been investigated.In order to construct a device that utilizes direct electron transfer the electrode surfaces were modified with nanostructures and, in some designs, self-assembled monolayers of thiols. The performance of the electrodes was evaluated by electrochemical methods, including potential sweeps and chronopotentiometry. Catalytic constants could be calculated mathematically by combining electrochemical methods with surface characterization methods, such as quartz crystal microbalance with dissipation and ellipsometry. All the fuel cells covered by this thesis are based on direct electron transfer processes. All designs also oxidize carbohydrates and reduce oxygen using cellobiose dehydrogenase and multi-copper oxidase, respectively.Our results revealed that the use of particular thiol had the capability to electrically connect cellobiose dehydrogenase to the electrode, equalling the commonly used two-thiol system. Both designs reached similar current densities, Le., about 20 jiA cm 2 with 5 mM lactose and the enzyme immobilized on thiolated gold nanoparticles (AuNPs). Both Bilirubin oxidase and Trichaptum abietinum Laccase could be directly immobilized on gold nanoparticles and current densities of up to 180 pA cm 2 were achieved. The 9- fold difference in currents with BOx and CDH reveals that the bioanode in this system requires more improvement to match the biocathode in performance. Upon doser inspection of the biointerface as regards the bioanode, it was concluded that a positive charge on the thiol was needed to create a direct (electric) contact between CDH and the electrode surface. Furthermore, the catalytic currents were nearly halved when the charged groups on the thiol were further modified with methyl groups.Biocompatibility of an implantable EFC design was evaluated using cell cultures of mammal cells, which was the first study of its kind. Toxicology tests revealed toxic by-products from the bioanode previously not reported in EFCs implanted in animals. The currents of the EFC was reduced by about half in cell culturing medium (10 1.1A cm') compared to PBS solutions, and was even more drastically reduced upon direct contact with fibroblast cells (1 jiA cm').