Operation of a 25 KW_th Calcium Looping Pilot-plant with High Oxygen Concentrations in the Calciner

The experimental set-up is shown in Figure 3. The plant comprises two interconnected fluidized-beds. Namely, the carbonator is a CFB with 4.3 m height and 0.1 m internal diameter (ID); while the calciner is a BFB with 1.2 m height and 0.165 m ID. The solid transport from one reactor to the other is controlled by two loop-seals fluidized with nitrogen. Both reactors are fed a mixture of gas through a preheating line, and both are electrically heated; moreover, the calciner is fed with natural gas in order to produce by combustion the heat needed for the endothermic calcination and to heat the circulating sorbent. The carbonator distributor plate has 8 nozzles, each of them with twenty 2 mm holes, while the calciner has 20 nozzles with six 1 mm holes each.

The results from three different experiments are discussed in this section. These tests provide an overview with regards to running the pilot plant from air (~20% vol O₂) to 100% vol O₂ at the calciner inlet. This work also explores the results of using different particle size distributions in this operational mode to see if this parameter has an impact on the overall performance of the system. The limestone used in this study has a minimum content of 98.25% CaCO₃.

Experiment 1: Flue gas (15% vol CO₂) with limestone (200-300 µm) 30% vol O₂

This first test with the limestone fraction 200-300 µm was performed to test the rig with limestone circulating between the two reactors within the rig as a starting point for performance optimization. During this test, a capture efficiency of 45% was achieved (Figure 4). This capture efficiency, E_carb, was calculated using the following formula²³: where F_CO2 is the molar flow rate of CO₂ entering the carbonator and F_carb is the molar flow rate of CO₂ leaving the carbonator.

(2)

Figure 4: Concentration of CO₂ at the carbonator inlet and outlet and the capture efficiency (Ecarb) for 200-300 µm limestone with 30% O₂. Please click here to view a larger version of this figure.

The inferior capture efficiency during this experimental run was caused mainly by the insufficient supply of heat to calcine all the limestone residing in the BFB. This caused a decrease of the CaO/CaCO₃ ratio in the carbonator feed. Another probable reason was the deactivation and carry-over of the lime particles from the calciner, which reduced the total bed inventory and the amount of sorbent present in the system. After the experiment, a material balance of the inventory of the reactor was performed (Table 1). A shift towards smaller fractions can be observed, because of the attrition of the material in both fluidized beds.

Table 1: Weight balance of the material inputs and outputs for 200-300 µm limestone with 30% O₂.

Experiment 2: Flue gas (15% CO₂) with limestone (100-200 µm) 100% O₂

In this test, the main aim was to use of smaller limestone particles in order to investigate their possible beneficial effect on the system performance. The secondary aim was to provide more heat to the calcination process in the BFB calciner by combusting the natural gas in highly concentrated oxygen, ideally up to 100% at the inlet.

In this experiment, we successfully tested the possibility of using pure O₂ at the inlet of the calciner, which offers the possibility of completely eliminating the recycle of off-gas needed for a standard oxy-fuel process. This is made possible by the heat consumption in the form of a circulating fluidized bed material and the continuous calcination reaction.

The use of smaller particles did not have a beneficial effect on the carbonation process, most likely because of the higher contact area between particles and the gas. However, there is some controversy in this matter as smaller particles have shown decreased reactivity due to higher content of impurities²⁴. Almost all the added limestone that was < 150 µm was very quickly elutriated from the calciner to the downstream cyclone. Therefore, it was very difficult to maintain the necessary inventory of lime in the rig needed to achieve higher capture efficiency. The results of the capture efficiency are in Figure 5.

Figure 5: CO₂ concentration at the carbonator inlet and outlet and the corresponding capture efficiency (E_carb) for 100-200 µm limestone with 100% O₂. Please click here to view a larger version of this figure.

The balance of the inputs and outputs of the solids was performed after the test (Table 2), and it was discovered that most of the material introduced into the rig during the experiment ended up in the calciner cyclone. This was probably the main cause of the low measured capture efficiency during this test.

Table 2: Balance of recovered material and its sieve analysis for 100-200 µm limestone with 100% O₂.

To conclude, we successfully tested the use of pure O₂ at the inlet of the calciner in order to completely eliminate the recycle of the off-gas needed for a standard oxy-fuel process. This is possible due to the heat sink provided in the form of circulating fluidized bed material and the continuous calcination reaction. However, the small particle size distribution (100-200 µm) was not beneficial for the capture process, due to the elutriation of the particles. It was extremely difficult to maintain the bed inventory needed to obtain a higher capture efficiency. Therefore, we decided to investigate the use of bigger particles in the next experimental campaign.

Experiment 3: Flue gas (15% CO₂) with limestone (300-400 µm) 100% O₂

During this trial, the performance of the fraction 300-400 µm was tested so as to reduce the high material losses from the fluidized bed of the calciner as seen in previous runs. This was expected to enable retention of the necessary lime/limestone inventory needed for its efficient circulation and carbon capture. Due to use of an appropriate amount of sorbent circulating in the rig and sufficient heat provided by the combustion of natural gas in pure oxygen (releasing 18 kW), a stable capture efficiency of ~ 70% was achieved for more than 3 h; this is a very good result when considering the relatively short carbonator reactor tube and the consequent short contact time between the sorbent and CO₂. The concentration of CO₂ at the outlet of the carbonator was maintained below 5% vol, and fresh limestone (in 0.5 L batches) was added to the calciner when the concentration of CO₂ at the outlet of the carbonator exceeded this value. A stable experimental run was achieved with optimized conditions.

The process started with the standard procedure; i.e., first the reactor was heated up to 700 °C, then 2.9 L of limestone was added into the calciner and heated up. The temperatures and gas concentrations in the calciner are shown in Figure 6. Note that the numbers below correspond to those steps in Figure 6. 1) The air flow was replaced by a flow mixture of 40% O₂ and 60% N₂ and the combustion of natural gas in the fluidized bed was initiated (9.1 kW). The limestone in the fluidized bed was heated above 800 °C and 3 more batches of limestone (1 L) were added to the calciner. 2) While the limestone was calcining in the fluidized bed, the circulation of the lime/limestone was started by flowing preheated N₂ through the carbonator (at a velocity of 2.5 m/s at 650 °C). An additional 0.9 L limestone was added and 3) a fresh O₂ cylinder was connected to the inlet of the calciner. 4) After reconnecting the oxygen, the combustion was initiated again, this time in an inlet O₂ concentration of 70% (and 30% N₂), which led to a consumption of 14 kW of natural gas to reach an O₂ concentration at the outlet of ~ 5% (in wet gas). 5) Pure O₂ was introduced at the calciner inlet, which led to the heat release of 18 kW into the calciner, and 6) the carbonation was initiated in the carbonator by injecting 15% of CO₂. The efficiency of the carbonation (Figure 7) was the highest yet on this reactor design (~ 70%). 7) The gas velocity flowing through the BFB of the calciner had to be lowered to 0.30 m/s (required by the desired temperature) to maintain the temperature of about 930 °C generated by the combustion of natural gas in pure O₂ (while maintaining the O₂ concentration in the off-gas to an industrially acceptable level below 5% vol).

Figure 6: Temperature of the BFB of the calciner and the temperature and composition of the off-gas at its exit. Please click here to view a larger version of this figure.

The temperature in the fluidized bed was maintained constant during the test by regulating the flow rate of natural gas and consecutively adjusting the flow rate of O₂ to achieve a calciner velocity of ~0.30 m/s. During the constant operation, the following energetic balance was observed: in the calciner, the oxygen was preheated to only 300 °C for safety reasons; hence, providing 0.5 kW in latent heat. Around 15.5 kW was released by the combustion of natural gas in the fluidized bed, but only 5.3 kW was needed for a continuous calcination (when no fresh material was added), 3.6 kW to cover the latent heat of the off-gas and ~7 kW for heating the circulating adsorbent and to cover the heat losses. In the carbonator, 3.2 kW was provided by preheating the gas (to 350 °C), and 5.4 kW was released by the carbonation process. Almost 5 kW was carried out of the carbonator as latent heat of the off-gas and 3.6 kW needed to be dissipated by heat losses to cover the heat released by the exothermic reaction while maintaining the temperature at 650 °C. When the carbonation process starts (and is effective), the temperature in the carbonator approaches 700 °C, which shifts the attainable thermodynamic equilibrium concentration of the outlet gas to higher values. This theoretical minimum concentration, dependent on the carbonator temperature, is depicted alongside the concentration measured during testing in Figure 7.

The carbonation process was initiated with ~6.7 L of limestone present in the rig. An additional 0.54 L at the beginning of the carbonation (13:45 in Figure 7) resulted in the decrease of residual CO₂ in the gas below 5% vol, and this level was maintained throughout the test by additions of 0.5 L limestone batches (when the concentration of CO₂ at the exit of carbonator exceeded 5%). The time intervals between the additions of fresh limestone to the calciner were 15, 20, 50, 45 and 50 min. Therefore, we concluded that a stable operation requires the addition of fresh sorbent corresponding to 0.5 L (750 g) of limestone approximately every 50 min, which is equivalent to a make-up ratio (F₀/F_CO2) of 6%, as described elsewhere²⁵. The make-up ratio in these experiments was mainly influenced by the limestone (reactivity decay and elutriation). The value of this parameter was chosen based on the CO₂ concentration at the carbonator outlet, i.e., adding more limestone when it reached a 5% vol CO₂.

Based on the sieving analysis, we conclude that the original limestone fraction was retained mostly in the rig circulation, while particles mostly < 250 µm were carried over to the calciner cyclone. These particles resulted predominantly from the breakage/attrition of larger particles.

Table 3: Mass balance of solids and sieve analysis of the inputs (limestone) and outputs (other) for 300-400 µm limestone with 100% O₂.

Table 4: Molar balance estimate (10% humidity of the raw limestone, 75% wt of the output in calcined state) for 300-400 µm limestone with 100% O₂.

Figure 7: CO₂ concentration at the carbonator inlet and outlet and the corresponding capture efficiency (E_carb) for 300-400 µm limestone with 100% O₂.  Please click here to view a larger version of this figure.