The experimental device is composed of gas supply system, reactor for hydrate formation and decomposition, cooling water circulation system and detection system, as shown in Figure 1. The pure CH₄ gas and 40% CO₂ + 60% H₂, 40% CO₂ + 60% N₂, 20% CO₂ + 80% N₂ mixtures used in experiments are supplied by Foshan Huate Gas Co., Ltd. The deionized water used in experiments is supplied by Nanjing ultrapure water technology Co, Ltd. The Raman spectrometer (LabRam, Jobin Yvon) uses 50 times long focusing lens, a 600 grooves/mm monochromator and a multi-channel air-cooled electrically coupled device (CCD) detector. It can release 532 nm wavelength laser Ar ion laser source as the laser emission source. The single crystal silicon standard sample with Raman band at 520.7 cm⁻¹ is used to calibrate the Raman spectrometer.

The gas samples collected during the experiment were analyzed on Agilent 7890A, which is equipped with FID and TCD detector. The test method for gas samples is: the detector is heated from 298.15 to 523.15 K at a constant speed, the flow rate of combustion gas H₂ is 30 ml min⁻¹, the flow rate of combustion gas air is 400 ml min⁻¹, and the flow rate of carrier gas helium is 250 ml min⁻¹.

In order to compare the effects of different gas mixture on CH₄ recovery efficiency, all experiments were carried out at 275.15 K and 6.0 MPa. The volume of reactor is 100 ml, and the amount of water used to form hydrate is 60 ml. The CH₄ hydrate is generated by bubbling at the bottom of the reactor under magnetic stirring, and the gas mixture is injected through the bottom of the reactor. About 120 h later, the water in reactor has completely transformed to hydrate which is confirmed by Raman spectroscopy where there is no characteristic peak of water, as the same as the method used before (Ding et al., 2017; Ding et al., 2020). What should be noted is that the experimental data is the average value of two groups of experiments, because each experiment was carried out in two parallel reactors.

Experiment 1, 2 and 3 were conducted using 40% CO₂ + 60% H₂ to replace CH₄ hydrate. In Experiment 1, the mixture of 40% CO₂ + 60% H₂ was injected after the complete transformation of H₂O to hydrate which was confirmed by Raman spectroscopy, and the top vent valve of the reactor was opened at the same time to exhaust slowly (0.45 ml min⁻¹ of exhaust speed) to shift the CH₄ in gas phase to 40% CO₂ + 60% H₂. The top vent valve and the bottom inlet valve were kept open during the whole replacement process. What should be noted is that the pressure in reactor was remained at 6.0 MPa.

In Experiment 2, when the concentration of CH₄ in the gas phase was lower than 2% during gas exchange operation, the top vent valve of the reactor was shut off, and the time marked as the beginning of the replacement reaction. After the replacement reaction proceeded 12 h, one gas sample was collected, and then the gas exchange operation was carried out. Till that CH₄ concentration was lower than 2% again, another gas sample was collected as beginning of next 12 h of replacement reaction, and the top vent valve was shut off. Afterwards, the gas exchange operation was carried out every 12 h (the inlet and vent valves were opened simultaneously to inject 40% CO₂ + 60% H₂ gas mixture) to renew the gas in the gas phase till CH₄ concentration was lower than 2% again. The gas samples were collected at the beginning and end of the exchange process and detected by gas chromatography to determine the amount of CH₄ that were replaced out from hydrate phase within 12 h. Notedly, the inlet valve at the bottom of reactor was open during the whole replacement reaction. The only one difference between Experiment 3 and Experiment 2 is that the gas exchange operation was carried out every 36 h.

In order to study the reaction mechanism of CH4 hydrate replacement by CO₂ mixture containing small molecules, the replacement of CH₄ hydrate by CO₂/N₂ mixture with different concentrations was studied at different pressure. CH₄ hydrate was replaced by 40% CO₂ + 60% N₂ mixture at 275.15 K and 6.0 MPa in Experiment 4, 20% CO₂ + 80% N₂ mixture at 275.15 K and 6.0 MPa in Experiment 5, and 20% CO₂ + 80% N₂ mixture at 275.15 K and 8.0 MPa in Experiment 6. The comparison of experimental conditions of three experiments is also listed in Table 1. Gas samples were collected every 24 h during the replacement reaction, and the concentration changes of each gas component in the gas phase in the reactor were determined by gas chromatography. After the replacement reaction, the reactor was treated with liquid nitrogen, and then the hydrate was decomposed at room temperature. The decomposed gas was collected and each component concentration in the hydrate phase was determined by gas chromatography.

In Experiment 1, the gas was continuously discharged at the rate of 0.45 ml min⁻¹ during the whole replacement reaction (the inlet valve of 40% CO₂ + 60% H₂ was also open and the pressure in reactor was maintained at 6.0 MPa). After the replacement reaction, the gas phase was discharged quickly and the reactor was treated with liquid nitrogen. Immediately, the hydrate phase in reactor was decomposed at room temperature, and the content of each gas component originating from hydrate decomposition was determined by gas chromatography. In Experiment 1, the composition of the final hydrate decomposition gas is 89.59% CH₄ and 10.41% CO₂, that is, the recovery rate of CH₄ is 10.41%.

In Experiment 2, gas samples were collected at the beginning and end of the gas exchange operation, and the gas content in each gas sample were compared to have a deeper understanding of the replacement process. The increment of CH₄ and the decrement of CO₂ in the gas samples collected during the experiment are listed in Table 2, and is plotted in Figure 2.

It can be seen that the increment of CH₄ and the decrement of CO₂ in the gas samples collected every 12 h are gradually decreasing, and the decrement of CO₂ every 12 h is slightly higher than the increment of CH₄, indicating that more CO₂ is consumed due to the decomposition of hydrate in the replacement process. This result is consistent with the experimental results of CH₄ hydrate replacement with 40% CO₂ + 60% H₂ mixture, which are present in our previous work (Ding et al., 2017; Ding et al., 2020). The composition of final hydrate decomposition gas is 87.75 CH₄ and 12.25% CO₂, that is, the CH₄ recovery rate is 12.25%.

In Experiment 3, gas samples were collected at the beginning and end of the gas exchange operation every 36 h, the increment of CH₄ and the decrement of CO₂ in the gas samples collected during the experiment are listed in Table 3, and is plotted in Figure 3. It can be seen from the figure that the increment of CH₄ and the decrement of CO₂ decrease almost linearly, and the decrement of CO₂ is also slightly higher than the increment of CH₄ which is agreement with Experiment 2. The composition of the final hydrate decomposition gas is 72.64% CH₄ and 28.36% CO₂, that is, the CH₄ recovery rate is 28.36%.

The results of Experiment 1, 2 and 3 and experimental results in our previous work (Ding et al., 2020) signed as EP1, are compared in Table 4. It can be seen that different time intervals of gas exchange operation eventually led to different CH₄ recovery efficiency in replacement reaction. The lowest CH₄ recovery efficiency (10.41%) is obtained with continuous gas exchange meaning that the condition of continuous gas exchange, i.e., the mixture of 40% CO₂ + 60% H₂ passes through the hydrate area at a relatively faster speed, result in a shorter contact time between CO₂ or H₂ molecules and CH₄ hydrate. Thus, the replacement reaction was cannot effectively carried out, resulting in the final lower CH₄ recovery efficiency. The CH₄ recovery efficiency increased (12.25%) with the time interval changed to 12 h, and was significantly improved (32.24%) as the gas exchange interval increase to 24 h. However, when time interval increased to 36 h, the recovery efficiency decreased slightly (28.86%). The reason may be that as the time interval increases, CH4 gas that replaced from hydrate phase during the replacement reaction reformed CH₄ hydrate again or formed CH₄-CO₂ mixed hydrate together with CO₂.

With the above experimental results, it can be proposed that in the real process of using gas replacement method to exploit NGH, the replaced CH₄ needs to be discharged from the sediment in time to avoid the replaced CH₄ gas forming CH₄ hydrate again or forming CH₄-CO₂ mixed hydrate together with CO₂ and so affecting the exploitation efficiency. At the same time, it is necessary to control the frequency of gas extraction from hydrate sediment, so as to avoid the incomplete contact between CO₂ molecules or other small gas molecules and CH₄ hydrate, which makes the lower CH₄ recovery efficiency.

In Experiments 4, 5 and 6, CH₄ hydrate was replaced by 40% CO₂ + 60% N₂ at 6.0 MPa, and by 20% CO₂ + 80% N₂ at 6.0 and 8.0 MPa. Combined with the experimental results of replacing CH₄ hydrate with 40% CO₂ + 60% H₂, the replacement mechanism of CH₄ hydrate by CO₂ mixture containing H₂ or N₂ was discussed.

In Experiment 4, after the water in the reactor was completely converted into hydrate, the gas mixture of 40% CO₂ + 60% N₂ was injected into the reactor until the CH₄ concentration in the gas phase was less than 2%, the replacement reaction began. The gas exchange operation was carried out every 24 h. Table 5 shows the CH₄ and CO₂ content changes in the gas samples collected during the replacement process of Experiment 4, and these data are plotted in Figure 4. It can be seen that more CH₄ is replaced in the first 48 h of the replacement process (the increments of CH₄ concentration in the gas sample every 24 h were 8 and 4% respectively), while less CH₄ is replaced out in the subsequent replacement process. This result is quite different from that of CH₄ hydrate replacement with CO₂/H₂ mixture (more CH₄ is replaced in the first 5 days). Similar to that of CH₄ hydrate replacement with CO₂/H₂ mixture, the reduction of CO₂ in gas samples is slightly higher than the increment of CH₄ every 24 h. It is suggested that the decomposition of CH₄ hydrate may also occur in the process of CH₄ hydrate replacement with 40% CO₂ + 60% N₂, resulting in the decrement of CO₂ being higher than the increment of CH₄, as observed in CH₄ hydrate replacement with 40% CO₂ + 60% H₂.

Finally, the concentration of each gas component in the hydrate decomposition gas was detected by gas chromatography, and showed as 26.31% CO₂, 2.54% N₂ and 71.15% CH₄. In other words, at 275.15 K and 6.0 MPa, the CH₄ recovery efficiency is 28.85% by using 40% CO₂ + 60% N₂, which is lower than that using 40% CO₂ + 60% H₂ at the same temperature and pressure (32.24%). However, in the experiment of replacing CH₄ hydrate with 40% CO₂ + 60% H₂, there are only CO₂ and CH₄ in the final hydrate dissociation gas, and H₂ does not exist in the hydrate phase; In contrast, there is 2.5% N₂ in the final hydrate phase in the experiment of replacing CH₄ hydrate with 40% CO₂ + 60% N₂, indicating that N₂ molecules entered the hydrate lattice and occupied the hydrate cages.

The obtained low CH₄ recovery efficiency may be resulted from that the partial pressure of CO₂ reaches 2.4 MPa (above the pressure of CO₂ hydrate formation at 275.15 K) during the replacement process, which bring about CO₂ hydrate formed quickly on the surface of CH₄ hydrate and hindered the further contact between the injected gas mixture and CH₄ hydrate, resulting in the low final replacement efficiency. So, in the following Experiment 5, CH₄ hydrate was replaced by 20% CO₂ + 80% N₂ at 275.15 K and 6.0 MPa, and the partial pressure of CO₂ was only 1.2 MPa (below the pressure of CO₂ hydrate formation at 275.15 K).

After the water in the reactor is completely converted into hydrate, 20% CO₂ + 80% N₂ is injected and the CH₄ gas in the gas phase area of the reactor is discharged at the same time. The photos of the reactor taken during the experiment are shown in Figure 5. Figure a is the picture of the reactor before injecting 20% CO₂ + 80% N₂ mixture, and figure b, c, d, e and f are the picture of the reactor at 4, 6, 8, 10 and 12 h after injecting gas respectively. It can be seen that most of the hydrate decomposes within 12 h, and the hydrate completely decomposes within 24 h. It shows that the mixture of 20% CO₂ + 80% N₂ cannot react with the decomposed water to form stable hydrate at this temperature and pressure (275.15 K, 6.0 MPa).

On the basis of Experiment 5, the reaction of CH4 hydrate with 20% CO₂ + 80% N₂ mixture at 275.15 K and 8.0 MPa, where the partial pressure of CO₂ increased to 1.6 MPa, was carried out in Experiment 6. The experimental results are the same as those of Experiment 5. CH₄ hydrate is decomposed in 24 h and no new hydrate is formed in the reactor.

In addition to the above experiments, the interaction between CH₄ hydrate and pure N₂ was also carried out. The pictures taken during the experiment are shown in Figure 6, the pictures a, b, c, d and e in the figure are the pictures of the reactor taken before the start of reaction and after 4, 8, 12 and 20 h of reaction respectively. It can be seen that CH₄ hydrate gradually decomposes after the replacement process begins, and almost decomposes within 24 h. It shows that the contact of N₂ with CH₄ hydrate will lead to the destruction of hydrate lattice and release CH₄ gas.

Through the above experiments, it can be seen that CH₄ hydrate will be completely decomposed when the partial pressure of CO₂ in the mixture is too small to form hydrate; When the mixture can form hydrate stably, CH₄ hydrate will not decompose completely.