The coal samples were loaded into the sample cell, which was then carefully arranged in the experimental device per the schematic diagram (Figure 2). The sample cell was seamlessly positioned between a pressure gauge and a filter, and two valves efficiently regulated the gas flow in and out of the equipment. The sample cell was injected with CH₄ to initiate the adsorption process, and the pressure gauge accurately measured the resulting pressure. A filter was strategically

employed to prevent the coal powder from escaping with the gas from the equipment. Furthermore, the average weight of the equipment in a vacuum was meticulously recorded as 807.625 g.

Methane gas was injected into the equipment at room temperature while maintaining a pressure of 3 MPa. The pressure was kept constant for a period of 5 to 8 hours to ensure that an equilibrium state had been reached. Afterward, the weight of the equipment was measured using an electronic balance. The difference between the weight of the equipment in a vacuum state and the weight of the equipment when filled with methane was then recorded. Adsorption was taken into account during the process. The amount of CH₄ injected is calculated as [21] .

n = ( W − W 0 ) m (1)

where n is the amount of injected CH₄ (mol), W is the equipment weight after the gas injected and reached equilibrium (g), W 0 is the equipment weight in vacuum state (g), m is the gas molecular weight (g/mol).

Furthermore, it is possible to determine the quantity of gas that is not adsorbed to the coal by utilizing the following equation as [22] .

P V = Z n f R T (2)

where P is the gas pressure (MPa), V is the equipment internal volume (mL), Z is the compression coefficient, n f is the free gas amount (mol), R is the gas constant (J/K mol), T is the temperature (K).

The equation for determining the quantity of gas adsorbed on coal surface in the equipment has been formulated as [22] .

n ′ = n − n f (3)

where n ′ is the adsorbed gas amount (mol).

The equations can measure the quantity of gas adsorbed on coal but cannot determine the rate of adsorption. The only way to estimate the rate of adsorption is by making assumptions while considering other parameters such as pressure and temperature.

The research involved the adsorption of methane to coal, followed by the injection of carbon dioxide to extract the methane from the coal. The equipment used for both continuous and cyclic tests featured a microflow control valve and a water displacement device at the outflow side, as depicted in Figure 3. The microflow control valve was employed to adjust the outflow in small increments.

The experiment involved injecting CO₂ at a pressure of 5 MPa and room temperature. The process of displacement was being carried out while keeping the outflow constant. The experiment continued until steady state conditions were achieved, where the inflow and outflow gas composition and rate were equal. The duration required to attain a stable state was meticulously recorded during the experiment. The weight of the equipment was measured using an electronic balance to determine the amount of CO₂ injected.

As part of the continuous testing process, the CH₄ gas released from the equipment was efficiently collected in 10 mL sample bottles. To facilitate the gas collection process, a 100 mL cylinder was placed in a water displacement device, and the released gas was collected into the cylinder at a fixed outflow rate of 4.650 mL/min. Each collected gas sample was meticulously analyzed using a gas thermal conductivity detector (TCD) chromatograph. To accurately measure gas production, a needle was connected through the rubber stopper of the sample bottles, and the stopcock was opened to let the gas displace the liquid in the cylinder. Finally, gas samples were directly taken from the sample bottles using a specialized chromatography syringe. In the experiments assessing natural CH₄ desorption by CO₂ replacement, the desorption efficiency and desorption efficiency per unit time was defined as [8] .

D e = Q Q 0 (4)

D e p = D e T E (5)

where D e is the desorption efficiency (%), Q is the total desorption amount (mol), and Q 0 is initial adsorption amount (mol), D e p is desorption efficiency per unit time (%), and T E is desorption time (h).

In this experiment, the cyclic method was employed as a means of ensuring an adequate soaking time or cyclic phase between injection and adsorption. As such, the test involved the saturation of CO₂ for a duration of two weeks, during which the valves were kept firmly shut to create a sealed state. To obtain a comprehensive data set, samples of gas were collected every 12 hours, with a total of 28 samples being obtained over the two-week period.

Coal has the capability to adsorb methane easily, and it takes around 40 minutes to reach an equilibrium state. However, research into CH₄ adsorption simulation has shown that several micropores are closed and unable to adsorb CH₄ [23] . This condition leads to variations in the state of adsorbed and free CH₄ [8] . In this study, CH₄ adsorption was calculated by considering the excess CH₄ adsorption amount and free CH₄. As shown in Figure 4 higher injection pressure leads to a greater amount of CH₄ adsorption. A significant increase was observed between 0 and 1 MPa, with a slower increase between 1.5 and 3 MPa. The observed phenomenon results from increased pressure, which facilitates adsorption and enhances the degree of spontaneity of CH₄ adsorption [24] .

In this test, the CO₂ injection was continuous and recovery process was stopped after reaching steady state where the CO₂ inflow and outflow gas compositions

and rates are equal. To attain the steady-state condition, the collection of samples in bottles was carried out for a total of 8 hours. This process was repeated 15 times to ensure its accuracy and reliability. The effects of alterations in CH₄ concentration on the CO₂-ECBM behavior over time during gas injection. The results of the continuous testing, as shown in Figure 5, indicate that CO₂ initially caused a surge in methane production for 30 minutes, followed by a slight decrease for the subsequent 90 minutes, and then a gradual decline until methane production ceased at 300 minutes. In conclusion, the recovery rate of CH₄ is faster in the early stage and slower in the later stage. During the early stages, the high pressure of CO₂ facilitated a substantial amount of CH₄ to flow into the outflow system. As time passed and CO₂ injection rates remained constant, it became increasingly difficult to locate CH₄. By the total desorption calculation, the desorption efficiency from continuous test is 26% with desorption efficiency per unit time is 3.25%.

It was observed during cyclic testing that the rate of CH₄ adsorption was faster than that of CO₂. CO₂ injection under 5 MPa took about 5 hours to reach equilibrium. The results of the cyclic experiment were different from the continuous test. In the continuous test, the CH₄ concentration decreased over 150 hours and then slightly decreased until 350 hours (Figure 6). The CO₂ concentration was measured in this study to observe the sorption process on the coal matrix over the long term. The results showed that the CO₂ concentration slightly increased until 350 hours. The cyclic testing process showed different results from the continuous test, based on the duration of the test. In continuous testing, the peak of desorption occurs in a short time, while in cyclic testing, the peak of desorption remains stable over several hours. During continuous testing, CH₄ was only detected for 300 minutes, whereas cyclic testing continues for a longer period of time. According to calculations, the desorption efficiency from cyclic tests is 12%, with a desorption efficiency per unit time of 0.03%.

In continuous test, several stages occur during CH₄ production, as shown in Figure 7. During the initial phase, that methane CH₄ in its free state, adsorbed onto the coal surface and subsequently absorbed into the coal matrix. The second stage shows the significant impact of CO₂ pressure, which prompts the outflow of CH₄. Following the complete drying and vacuuming of the experimental sample, it was observed that the original coal sample lacked any water content except for adsorbed CH₄, thereby facilitating the flow of CH₄ to the outflow. Most of the remaining CH₄ was transported to the fracture space of the coal sample through large pores, where it transformed into a completely free state [8] . The process of CH₄ desorption from the coal matrix results in the contraction of the matrix, leading to a reduction in the horizontal stress and, subsequently, an increase in the permeability of cleats [25] . During the third stage, the phenomenon of Fickian diffusion takes place within the smaller pore sizes in the mesopores, leading to a sustained increase in mesopore diffusivity [26] . The current condition enables a greater accessibility of the coal matrix, thereby facilitating the adsorption of CO₂.

During the last stage of the process, the conditions stabilize, and no traces of methane are detected in the outflow. As the adsorption of CO₂ increases, there is also a corresponding increase in porosity [27] . As the permeability of coal decreases, the rate of CO₂ flow gradually slows down, leading to a breakthrough at the outlet, and results in a CO₂ concentration of 90% [28] . Based on the test results, it is evident that the continuous test successfully recovered CH₄. The continuous test successfully recovered CH₄, but recent research suggests that during the process of employing flooding CO₂ ECBM, a significant proportion of CH₄ is carried away by CO₂, instead of being removed primarily via the replacement method [28] . In addition to the adsorbed and absorbed methane, the swelling of the coal matrix can make it difficult to remove CH₄. Using continuous methods

can speed up the process, but it can also decrease the likelihood of CO₂ adsorbing and replacing CH₄ on the coal matrix.

During the observation of the CH₄ recovery process, it was found that the cyclic test demonstrated distinct stages that were not observed during the continuous test. Although the initial stage of both tests was the same, the subsequent stages of CH₄ recovery differed (Figure 8). During the second stage, CO₂ was injected and allowed to soak into coal samples, allowing CH₄ to desorb from coal into a free state. Coal samples can release CH₄ due to pressure differences and CH₄ could diffuses through the matrix pores faster than CO₂, indicating that CH₄ flows faster than CO₂ under the same pressure gradient [8] . In the case of CO₂, a stable condition was created because the CO₂ pressure in the coal near the injection well quickly reached the same level as the injection pressure and remained constant with the CO₂ concentration in the coal matrix gradually increased over time [16] .

On last stage where CH₄ decrease, and CO₂ slightly increase into the end of the process. This condition occurs when the gas concentration in coal increases gradually, causing a decrease in gradient between it and the boundary concentration, which in turn reduces the amount of CO₂ adsorbed by coal [16] . Additionally, once the process of adsorption is finished, the quantity of CO₂ that gets desorbed is measured [29] . On the other hand, the competition between different mechanisms of gaseous methane diffusion and the swelling of the coal matrix caused by gas adsorption was raised [26] . The desorption curve experiences a decline indicating that only a limited number of methane molecules can desorb, with the residual gas being retained [30] . This condition worsens when small pores reach maximum adsorption capacity, leading to

dominant swelling effects in the coal matrix [31] . Due to various factors, the CH₄ trapped in coal seams cannot be released smoothly, which may hinder CH₄ recovery improvement.

The results obtained from continuous and cyclic tests have revealed significant differences in the desorption efficiency and desorption per unit time, as shown in Figure 9. This calculation has been formulated considering the desorption efficiency of adsorbed methane on coal. The results show that the continuous method has better desorption efficiency and time than the cyclic method. However, the desorption efficiency in continuous test is not significant. This condition is caused by pressure, which pushes CO₂ to outflow faster, resulting in a shorter time for CO₂ to adsorb and absorb, thus leading to ineffective CH₄ desorption. During cyclic testing, the injection pressure no longer has an effect after a certain point due to the long time it takes for the process to complete. This results in CO₂ adsorption and swelling, which makes it difficult for CH₄ to desorb and flow out of the system. If this process were to be applied at an industrial scale, it is possible that the quantity of methane produced may fall below the threshold of economic viability, rendering the process unfeasible and requiring its discontinuation. Based on these results, it can be concluded that pressure and time are the primary factors impacting the efficiency of CH₄ recovery.

This study showcases the effectiveness of CH₄ desorption using an extreme example. Other publications have used cyclic testing, with a few hours of soaking time, repeated multiple times, to maximize CH₄ recovery [32] [33] . To optimize the research, a synergistic approach involving both continuous and cyclic methods can be employed, whereby CO₂ is subjected to a brief soaking period, followed by reinjection to attain maximum CH4 recovery. Furthermore, the research can be improved by considering coal seam permeability, porosity evolution, adsorption capacity, and geological conditions.