Highly efficient technologies for the conversion of solid fuels, such as coal and biomass, to electricity are required for clean energy production and sustainable energy supply. Direct carbon fuel cells (DCFCs), which convert the chemical energy in solid carbon directly into electricity without gasification, have attracted much attention because of their theoretical ability to achieve 100 % efficiency. However, solid carbon particles have disadvantages in terms of fluidity when compared to gaseous fuels. Thus, carbon particles are dispersed into molten carbonate to form a carbon/carbonate slurry in which the carbon particles can move. Then, the anode is inserted into the carbon/carbonate slurry for discharge in the DCFC [1-5]. Carbon dispersed into the molten carbonate is easily transported to the anode, leading to an improvement of the formation of triple phase boundary (carbon, electrolyte, and anode). During discharge, carbon is supposed to be consumed through the complete electrochemical oxidation (C+2CO3 2-→3CO2+4e- (R1)) which releases four electrons per carbon. However, several studies have suggested a possible formation of CO from partial electrochemical oxidation (C+1/2CO3 2-→3/2CO+e- (R2)) which releases one electron per carbon, or even from the Boudouard reaction (C+CO2→2CO (R3)) where CO2 produced during discharge reacts with carbon particles [3,6]. Although the partial oxidation (R2) and the Boudouard reaction (R3) reduce the DCFC efficiency, little effort has been made to study the reaction mechanisms in the DCFC. The aim of this study is to investigate the anodic reaction mechanisms in the DCFC using carbon/carbonate slurry experimentally and theoretically. Figure shows setup of the DCFC using carbon/carbonate slurry stirred by gas bubbling. The working electrode (WE), counter electrode (CE), and reference electrode (RE) were made from gold sheets. Cell discharge was controlled with a potentio/galvano-stat. The reactor was heated using an electric furnace to 1073 K. Activated carbon particle with an average diameter of 44.1 μm was used as fuel. Ternary molten carbonate (Li2CO3/Na2CO3/K2CO3 (= 16.9/25/58.1 mol%)) was used as electrolyte. The carbon content in the carbonate was set to 1.0 wt% or 2.0 wt%. The carbon/carbonate slurry was contained in the porous alumina tube having an average pore diameter of 120 nm. This allowed the separation between the carbon/carbonate slurry and the molten carbonate electrolyte. In the setup, the carbon/carbonate slurry was able to be stirred by gas bubbling. CO and CO2 concentrations in off-gases were measured by FID with methanizer, and CO and CO2 production rates were determined at the open circuit mode or constant discharge mode at 20 mA/cm2. Moreover, to study the impact of Boudouard reaction (R3) on CO formation, CO2 bubbling with a flow rate of 1.0 ml/min, which was approximately 5 times of theoretical CO2 produced during discharge at 20 mA/cm2, was used in the carbon/carbonate slurry at the open circuit condition, and CO concentration in off gases was measured. As a result, an amount of CO was formed when the carbon/carbonate slurry was not stirred by Ar gas bubbling during the constant discharge at both the carbon contents. At the open circuit condition, CO was not formed by CO2 bubbling (1.0 ml/min) simulated CO2 produced during discharge. This indicated that the impact of Boudouard reaction (R3) was insignificant in the DCFC. Moreover, measured CO and CO2 production rates were well balanced on the calculation based on two electrochemical oxidations (R1 and R2). And, the partial electrochemical oxidation accounted for approximately 80 % in the carbon consumption at the carbon content of 2.0 wt%, whereas that accounted for approximately 40 % at the carbon content of 1.0 wt%. It was shown that the ratio of partial electrochemical oxidation depended on the carbon concentration near the anode. On the other hand, when the carbon/carbonate slurry was stirred by Ar bubbling with a flow rate of 50 ml/min, only CO2 was formed and the production rate of CO2 corresponded to the theoretical rate based on R1 at both the carbon contents. This indicated the carbon particles were consumed only by the complete electrochemical oxidation (R1). It was shown that stirring the carbon/carbonate slurry was effective to enhance the complete electrochemical oxidation. Reference [1] N. J. Cherepy et al., J. Electrochem. Soc., 152 (2005) A80-A87. [2] X. Li et al., Ind. Eng. Chem. Res., 47 (2008) 9670-9677. [3] H. Watanabe et al., J. Power Sources, 273 (2015) 340-350. [4] H. Watanabe et al., Energy and Fuels, 30 (2016) 1835-1840 [5] H. Watanabe et al., J. Chem. Eng. Jpn., 49 (2016) 237-242 [6] T.M. Gür, J. Electrochem. Soc., 157 (2010) B751-B759 Figure 1