Abstract
A direct carbon fuel cell (DCFC) is a variation of the molten carbonate fuel cell (MCFC) which converts the chemical energy of carbon directly into electrical energy. Thus, the energy conversion efficiency is very high and correspondingly CO2 emission is very low for given power output. DCFC as a high temperature fuel cell performs better at elevated temperatures (>800 °C) but because of the corrosive nature of the molten carbonates at elevated temperatures the degradation of cell components becomes an issue when DCFC is operated for an extended period of time.
We explored the DCFC performance at lower temperatures (at 700 °C and less) using different sources of carbon, different compositions of electrolytes and some additives on the cathode surface to increase catalytic activity. Experiments showed that with petroleum coke as a fuel at low temperatures the ternary eutectic (43.4 mol % Li2CO3 - 31.2 mol% Na2CO3 - 25.4 mol % K2CO3) spiked by 20 wt % Cs2CO3 performed better than any binary or ternary eutectics described in the published work by other researchers. Maximum power output achieved at 700 °C was 49 mW/cm2 at a current density of 78 mA/cm2 when modified cathode was fed with O2/CO2 gases.
1. Introduction
World demand for electricity is ever increasing and the need for developing new energy generation technologies is increasing as well. One of these technologies is fuel cell (FC) technology. Several types of FCs are currently being developed and some of them are commercially available already. The most common fuel cells use hydrogen as fuel. Hydrogen can be produced from hydrocarbon resources as well as biomass via gasification. However, the present world production capacity for hydrogen can barely keep up with the demand by fertilizer production and petroleum processing. This is considered one of the disadvantages of the hydrogen fuel cell. Another type of fuel cell, which can use easily available and cheap fuel sources, should be considered. The direct carbon fuel cell (DCFC) provides a promising alternative. The DCFC uses solid carbon as fuel. Solid carbon fuel can be easily produced from many different resources including coal and petroleum coke e most abundant resources world-wide and particularly in North America.
The DCFC was first discovered by Sir William Grove in 1839. In 1896 the first DCFC was built by W.W. Jacques who used hydroxide melt as an electrolyte at 400 °C [1]. The major disadvantage of this type of cell was that the hydroxide melt turned to carbonate and the performance deteriorated with time, cutting short the life of the cell. This technology was investigated by other scientists in later years [2,3]. For example in the mid-1990s, researchers at Scientific Applica- tions & Research Associates, Inc. (SARA) have attempted to use “humidified water” or additives such as SiO2, As2O3 and MgO as well as oxy-anions such as pyrophosphate and per- sulfate that suppress the formation of carbonates [3,4].
Unfortunately neither the addition of humidity nor oxides achieved substantial reduction in carbonate formation at temperatures up to 650 °C. Much higher temperatures are needed for these effects to be significant. Consequently, this had a limited effect in lengthening the life of the cell and has led to a hybrid design using molten carbonates and the hydroxides in one cell compartment [3].
Later, hydroxide melt electrolytes were completely replaced by molten carbonate electrolytes by Broers and Ketelaar [5,6]. They adopted pure carbonate as a logical extension of the earlier explorations. R. Weaver et al. at Stanford Research Institute (SRI, CA) started developing a carbon fuel cell using the ternary eutectic of lithium, sodium and potassium carbonates. This eutectic was chosen because of its known scavenging capability for the pollutants produced by coal, namely ash and sulphur compounds [7]. Later, Weaver et al. found and confirmed that at temperatures near 700 °C in their ternary electrolyte system (32.1 wt % Li2CO3/33.4 wt % Na2CO3/34.5 wt % K2CO3) the electrochemical oxidation of carbon proceeded completely to carbon dioxide, thus providing full release of the energy of the carbon [8].
More recently Cooper et al. at Lawrence Livermore National Laboratory (LLNL) and other groups have rekindled interest in the DCFC [2,6e14,16,20e24]. Cooper et al. used a binary mixture of 38 mol % Li2CO3 - 62 mol % K2CO3.
They referred to this electrolyte composition as the “electrolyte, which is highly effective” [15] (we refer to it in this paper as “Eo”). In another publication by the same researchers 32 mol % Li2CO3 - 68 mol % K2CO3 eutectic composition was used [9]. They achieved 50-60 mW/cm2 power density when they used petroleum coke and 50-100 mW/cm2 with carbon black at 800 C. A high power density (187 mW/cm2) was reported by Chen M. et al. [24].They used the ternary eutectic mixture Li2CO3 - K2CO3 - Al2O3 (1.05:1.2:1 mass ratio). In their DCFC design they used a carbonaceous anode, which acted not only as an electrode, but also as a reactant which was consumed during the whole DCFC life.
Since the DCFC is a high temperature fuel cell and at high temperatures the electrolytes are corrosive, expensive mate- rials are required for the electrodes and the housing. The objective of this study is to lower the cell operating tempera- ture, while keeping its performance as high as possible when petroleum coke is used as a fuel.
Since the DCFC is a high temperature fuel cell and at high temperatures the electrolytes are corrosive, expensive mate- rials are required for the electrodes and the housing. The objective of this study is to lower the cell operating tempera- ture, while keeping its performance as high as possible when petroleum coke is used as a fuel.
2. Experimental
2.1. Design and principles of direct carbon fuel cell (DCFC)
The configuration and the operating mechanism of our DCFC are simple. It contains three compartments: anodeenickel foam (Marke Tech International Inc., US) and current collector; matrixezirconia felt (ZircarZirconia, Inc., US) filled with the molten electrolytes; and cathodeeoxidized nickel foam and current collector (Fig. 1). Current collectors, gold lead wires (from Technic Canada), are attached to the elec- trodes by spot welding.
At the cathode (NiO) which is wetted with the electrolyte, CO2 is catalytically converted to CO2— as the result of three phase interface reactions: solid phase (cathode), liquid phase (molten carbonates) and gas phase (CO2 and O2):
O2 + 2CO2 + 4e⁻ = ¼ 2CO₃²⁻
The resulting carbonate ions diffuse through the zirconia matrix saturated by the electrolytes and reach the anode (Ni) side. At the anode 2CO₃²⁻ ions react with carbon particles and release four electrons and CO2:
C + 2CO₃²⁻ = 3CO2 + 4e⁻
The net electrochemical reaction therefore is the oxidation of carbon:
C + O2 → CO2
with the flow of four electrons from the anode to the cathode.
It must be pointed out that the anode in our DCFC is simply a current collector and the real electrochemical reactions occur between carbon and carbonate ions in the electrolyte envi- ronment. As shown in Fig. 1, four electrons must be recycled through the external circuit. Likewise a fraction of product CO2 must be recycled from the gas exiting the anode chamber to the cathode side. For convenience, we mostly used the batch setup to study the cell performance. The carbon fuel was placed on the anode surface before the fuel cell was assembled. A typical DCFC operating temperature ranges from 650 °C to 900 °C and its performance is strongly dependent on temperatures.
To meet our objective and operate the high performance DCFC at lower (≤700 °C) temperatures we focused our study on two major issues: the lowering of the melting temperature of the electrolytes and the improving the catalytic properties of the cathode electrode. We also examined the usefulness of DCFC by testing the suitability of carbon particles from many sources.
2.2.Electrolytes
The key role of the electrolyte in the DCFC is to create a medium through which carbonate ions can travel from the cathode to the anode. Electrolytes can also act as a kind of filter, preventing undesirable ions or electrons from disrupt- ing the desired chemical reactions. Experiments have shown that, in order to achieve the desired ionic conductivity, the eutectic mixture must be heated approximately 150 °C above its melting temperature [4].
At these temperatures the carbonate mixtures are liquid and are good ionic conductors. The melting temperatures of electrolytes are highly dependent on their composition.
In general, binary carbonate (Li2CO3/K2CO3 or Li2CO3/ Na2CO3) eutectics have higher melting temperatures than ternary carbonate (Li2CO3 - Na2CO3 - K2CO3) eutectics [8], which can be decreased further by adding the right amount of other carbonates or some oxides. In order to find the right composition of the electrolytes which would work better at lower temperatures, we started our experiments using a ternary eutectic salt mixture consisting of 43.5 mol % Li2CO3 - 31.5 mol% Na2CO3 - 25 mol % K2CO3 which was identified as a potential gasification catalyst [17]. In this paper we refer to it as “E”. The carbonates Li2CO3 (99.9%), Na2CO3 (99%) and K2CO3 (99.7%) (all from Sigma Aldrich) were mixed together in a mass proportion of 32:33:35 respectively. This mixture was then placed in an oven and heated to 600 °C, higher than its melting temperature to ensure that the carbonates are completely melted and mixed well. The melt cooled inside the oven and the solid mass of “E” electrolyte was obtained.
2.3. Fuels
In order to ensure the usefulness of the DCFC we tested the suitability of carbon particles from many sources such as anthracite, charcoal, heat-treated lignite, activated carbon, coal-derived coke and petroleum coke. Our preliminary work showed that heat-treated Saskatchewan lignite resulted in the best performance at 700°C: a power density of 33 mW/cm2 at a current density of 72 mA/cm2 was achieved. But because of the objectives of our study, we conducted experiments using mostly petroleum coke, which was obtained from CanmetE- NERGY (Natural Resources Canada). Surface area of this coke was 1.6 m2/g (BET-micrometrics) and as its proximate analysis showed that it contained 11.12% volatiles. This coke was heat treated under N2 at 800 °C before using it in the cell. After the heat treatment the coke weight decreased by 12%. Semi-Quantitative analysis (WD-XRF) shows that after heat treat- ment there was little changes in the elemental composition. In each experiment 7 g of the treated coke was mixed with an electrolyte powder and loaded onto the anode electrode.
Fig. 2 - DSC graphs of the eutectic compositions: Eo, 38mol % Li2CO3-62 mol % K2CO3; E, 43.5 mol % Li2CO3-31.5 mol% Na2CO3-25 mol % K2CO3; E1, E + 20 wt% Cs2CO3; A purge gas is argon with a flow rate of 20 mL/min and a scanning rate of 10 C/min.
Fig. 3 - Effect of Cs2CO3 additive in electrolyte “E1” on DCFC performance; operating temperature is 700 °C; fuel is petroleum coke; “E” is a ternary electrolyte.
2.4. Electrodes
In our design a fuel is deposited on the surface of an anode where the actual oxidation takes place. It is important to have an anode with a good electrical conductivity since it also acts as a current collector with the attached gold wire. Unlike an anode, a cathode acts as a catalyst to produce CO₃²⁻ ions; for this reason, we examined the modification and the perfor- mance of the cathode compartment.
3. Results and discussion
3.1. Formulation of low temperature eutectic
The melting point of the eutectic mixture of “E” was identified as 397 °C which is about 100 °C less than melting point of the binary “Eo” (38 mol % Li2CO3 - 62 mol % K2CO3) eutectic (Refer to Fig. 2). These measurements were taken by Perkin Elmer Diamond DSC). To decrease the melting point even further, a series of different additives such as BaCO3, Cs2CO3, Rb2CO3, MgCO3, CaCO3, V2O5; CeO2; Ag2O, Fe2O3 were studied.
Fig. 4 e OCV of DCFC at 500 °C increases very slowly during first 24 h, then it reaches maximum of 0.74 V between 24 and 30 h
The lowest melting point —374 °C was obtained with the quaternary eutectic e “E1” (E รพ 20 wt% Cs2CO3). It was prepared as follows: 80 wt% of E powder was mixed with 20 wt% Cs2CO3 (99.9%) (from Sigma Aldrich), placed in the oven and heated to 600 °C. The melt was cooled inside the oven and ground up.
The influence of cesium carbonate on molten carbonate melting point is described in the work of Kojima et al. They found that the addition of rubidium and cesium carbonates to various binary carbonates decreased the surface tension of the molten carbonates which in turn increased gas solubility and decreased the melting point [18]. Decreased surface tension increases the distribution of an electrolyte between the porous cell components of the anode, cathode and electrolyte matrix [18], and consequently increases the DCFC performance.
Fig. 5 - DCFC runs at 700 °C with the physically modified cathodes; #1, reference - NiO (500 g/cm2); #2, NiO (800 g/cm2); #3, NiO (1450 g/cm2); #4, double #1; #5, Nickel powder and #1.
Fig. 6 - (a-c) SEM images of Ni foam in its starting state (a); NiO foam (b); Ni foam with NiO powder (c), and their close-up images.
The expected beneficial effect as discussed above was proved by our DCFC experiments as described below: two cell assemblies were prepared, one with the “E” electrolyte and another with the “E1” electrolyte. After 24 h, the first cell’s voltage was 0.674 V and the second 1.2 V. Consequently the power density from the first cell was 7.5 mW/cm2 at a cell current density of 20 mA/cm2 whereas the power density from the second cell was 25 mW/cm2 at a current density of 50 mA/ cm2. Results are shown in Fig. 3.
The DCFC with “E1” electrolyte performs 32.8% better than the DCFC with the “E” electrolyte. For the rest of our experi- ments we used only “E1” electrolyte.
3.2. Open circuit voltage (OCV) and operating temperature
Since our eutectic mixture E1 can melt at 374 °C, and since we proved it was performing better than DCFC with E electrolyte we attempted to operate the DCFC (with E1 electrolyte and petroleum coke) at the lowest possible temperature. But, as shown in Fig. 4, at 500 °C, OCV (the voltage measured with zero current) increases very slowly, and after about 30 h it reaches a plateau (0.74 V).
Therefore, we explored lowest temperature that we could use in our DCFC operation; the OCV at 700 °C reached 0.96 V in about 3 h. According to the OCV tests it was determined that our molten salts could function effectively at temperatures as low as 650 °C - 700 °C and achieve comparatively high DCFC performance.
3.3. Improving the cathode performance
Another way to increase DCFC performance is to improve catalytic properties of the cathode electrode. Fig. 5 summa- rizes the improvement of DCFC performance achieved by changing the cathode structure.
In this study the NiO foam cathode with a density of 500 g/ cm2 is as our reference cathode. The reference case resulted in a general power density of 22e25 mW/cm2 and a current density of 47 mA/cm2 - 50 mA/cm2 at 700 °C. Very similar results were obtained when we tested NiO foams with various densities. However, when two layers of the reference NiO foams were stacked, the cell performance increased due to the increased surface area of the cathode (by increasing the production of the carbonate ions), even though the distance that the oxidizing ions had to travel nearly doubled. This indicates that it is more critical to have increased surface area for CO2 conversion to create CO₃²⁻ ions than the distance these CO₃²⁻ ions must diffuse to react with carbon particles.
The cathode performance increased even more when the oxidized Ni powder was inserted into the Ni foam’s pores, which increased the effective catalytic surface area for the cathode reaction without increasing the diffusion path for the CO2— ions. Fig. 6aec shows the SEM images for the Ni foam, the NiO foam and the Ni foam with NiO powder. The surface of the oxidized Ni foam (NiO) appears rough (Fig. 6b) e therefore the catalytic surface area is more than the area of the refer- ence Ni foam by itself (Fig. 6a) which appears to be smooth. The third image (Fig. 6c) shows fine NiO powder that was created by immersing Ni foam in a mixture of fine NiO powder (from Aldrich Chemical Company Inc.) and polyvinyl alcohol. This Ni foam then was heat treated at 550 °C for 3 h. It is evident that the insertion of the NiO powder in the Ni pores was uniform and the catalytic surface area larger. This explains the higher performance of the nickel powder deposited cathode.
Fig. 7 - Comparison of the DCFC performance with: #1 reference cathode and Air/CO2 flow; #2 with CeO2 deposited cathode and Air/CO2 flow; #3 with Ce deposited cathode and O2/CO2 flow.
We also studied the catalytic activities of different addi- tives on the cathode with air/CO2 or O2/CO2 flow. Compara- tively high results (~30% higher) were obtained when the CeO2 was deposited on the cathode. This was achieved by immersing the already oxidized Ni foam into a 1% Ce (NO₃)₃ * 6H2O (from Sigma Aldrich) aqueous solution, holding it in about 10 s and drying at room temperature and then heating to 700 °C in the oven for 1 h. The cathode reaction was improved even more when the air/CO2 flow was replaced by the O2/CO2. This resulted in an increase of power output by more than 50% as shown in Fig. 7. It is evident that higher O2 concentration was beneficial in accelerating the formation of carbonate ions. However, considering the added cost of separating O2 from the air prior to feeding to the DCFC, careful economic comparisons of these two cases must be conducted.
4. Conclusion
Small batch direct carbon fuelcellswere constructed and tested. The new eutectic composition of the electrolyte “E1” with the low melting point (374 °C) was formulated and the performance of the DCFC was investigated. The DCFC performance with this electrolyte increased by almost 33% at 700 °C and by 24.5% at 800 °C when compared with the DCFC performance with the ternary eutectic “E” previously reported in our paper [19].
The modification of the cathode (with NiO powder or cerium additive), increased its catalytic activity and consequently the DCFC performance. The power density 33 mW/cm2 at a current density of 55 mA/cm2 was achieved when the air/CO2 gases were supplied to the cathode. When air was replaced by pure oxygen, the power density increased by nearly 50% and reached 49 mW/cm2 at a current density of 78 mA/cm2. More work has to be done in order to develop a more practical DCFC with continuous carbon feeding, which can operate at lower (≤700 °C) temperatures with high efficiencies.
Advertisement
The 10 largest coal producers and exporters in Indonesia:
- Bumi Resouces (BUMI)
- Adaro Energy (ADRO)
- Indo Tambangraya Megah (ITMG)
- Bukit Asam (PTBA)
- Baramulti Sukses Sarana (BSSR)
- Harum Energy (HRUM)
- Mitrabara Adiperdana (MBAP)
- Samindo Resources (MYOH)
- United Tractors (UNTR)
- Berau Coal






