**1. Introduction**

Carbon capture and storage (CCS) technology is essential for rapid CO2 mitigation. The geological storage of carbon dioxide (CO2) is a highly effective, long-term mitigation solution for the large quantities of CO2 emissions [1,2]. For these reasons, to date, CO2 geological sequestration (CGS) technology has been developed by several leading countries. However, most of existing CGS methods worldwide require particular geological structures to work, such as a highly pervious rock formation (e.g., sandstone layer) imbedded in impermeable layers (i.e., cap-rocks). This requirement leads to CGS application difficulties such as a shortage of proper sites, challenges in the long-range transport of CO2, deep drilling and injection, and restricted storage capacity, which substantially increases the cost of using CGS. To overcome these limitations, several CGS methods that do not need cap-rock, such as carbonated water injection (CWI), have been suggested [3–6].

CO2 can be stored in unconsolidated sediments under CO2 hydrate-bearing sediments. CO2 hydrates are ice-like solid lattice compounds composed of hydrogen-bonded water cages that encapsulate guest molecules of CO2. CO2 hydrates are formed in the seabed under low temperatures and high pressures [7,8]. Previous studies on natural gas hydrate-bearing sediments [9] and preliminary studies on CO2 hydrate-bearing sediments [10,11] have shown that CO2 migration is significantly hampered by the formation of gas hydrates, resulting in a self-trapping mechanism. Furthermore, the self-preservation response of CO2 hydrates slows the CO2 hydrate dissociation process [12], which serves to mend unintended fractures of CO2 hydrate-bearing sediments, thereby severely diminishing the transport of CO2 fluids [13,14]. Thus, it has been suggested that CO2 hydrates can be used as primary or secondary safety factors for CO2 geological storage in marine unconsolidated sediments [15–17]. Furthermore, unconsolidated sand sediments have advantages over consolidated rocks (e.g., sandstones) in that the CO2 storage capacity of the former is higher than that of the latter due to the high porosity of unconsolidated sandy sediments (40–60%). In addition, the CO2 injectability of unconsolidated sand sediments is superior because of their high permeability (0.1–10 darcys) resulting from wide and well-connected pore spaces.

Tohidi et al. (2010) performed experimental CO2 leakage simulations through each type of unconsolidated sediment (glass-bead, sand, and sand–clay mixture), and using electrical resistance measurements and a CO2 concentration analysis they confirmed the existence of the self-trapping effect of CO2 hydrates [11]. Massah et al. (2018) demonstrated the sequestration of CO2 through horizontal injection into a laboratory scale reservoir and revealed the large storage density of CO2 hydrate formations [18]. Gauteplass et al. (2018) described CO2 hydrate formation caused by liquid CO2 injection into cold, water-saturated sandstone and reported that hydrate formation in the pore space resulted in blockage of CO2 flow under most conditions [19]. However, more direct and comprehensive experimental data including temperature–pressure relations, elastic wave velocity, and dissociation tests are required for a better understanding of the behavior of CO2 and CO2 hydrate formation in unconsolidated sediments.

The objective of this study was to simulate CO2 geological storage into marine unconsolidated sediments using CO2 hydrates as a cap-rock. A large reaction cell was used to experimentally verify the CO2 self-trapping mechanism in marine sediments and to evaluate the behavior of CO2-stored unconsolidated sediment during CO2 hydrate formation and dissociation.

#### **2. Experimental Program**

#### *2.1. Soil Used*

The strata of unconsolidated marine sediments typically consist of multiple layers of a different sediment types such as sand-rich sediment layers and fine-grained sediment layers. The permeability of fine-grained sediments is very low (i.e., 10<sup>−</sup>3–10−<sup>7</sup> darcys; [20]), therefore, fine-grained sediments can be practically considered as impermeable layers, which obstruct the upward flow of CO2. Meanwhile, sand-rich sediments are suitable as CO2 storage host sediments because of their relatively high permeability (0.1–10 darcys), while they are ready for permeation of CO2. The effect of the self-trapping mechanism on a sand-rich layer is, therefore, important to CO2 geological storage into unconsolidated sediments. In this study, fine sand (Ottawa F110; mean particle size = 120 μm, specific gravity = 2.65, permeability = 5–6 darcys, quartz 99%) was used as the host sediment sample.

#### *2.2. Experiment Setup*

The experimental design simulated CO2 injection into a shallow marine sediment (i.e., high water pressure, low temperature) and CO2 hydrate formation. The experimental design used in this study is shown in Figure 1. A cylindrical and rigid-wall reaction cell was made of an aluminum alloy (duralumin, AA2024). The inner diameter of the cell was 20 cm, the height of the interior was 100 cm, and the internal volume was 31.4 L. The reaction cell was originally developed for an experimental simulation of thermal stimulation on gas hydrate-bearing sediments [21]. Water and liquid CO2 were injected from the bottom of the reaction cell using a water pump and gas booster. Pressure inside the reaction cell was controlled using a back-pressure regulator at the top of the cell. The quantities of CO2 gas and water that flowed out of the reaction cell were measured using a water substitution system.

Various types of sensors were installed at predetermined layers (every 10 cm) within the reaction cell as shown in Figure 2. The cell contained five T-type thermocouples for temperature measurements of the cell interior, five pressure transducers for fluid pressure measurements, five pairs of piezoelectric ceramic disks (diameter: 20 mm) for compressional wave (P-wave) measurements at layers A1–A5, and four pairs of electrodes for electrical resistance measurements at layers B1–B4. For P-wave velocity measurements, square-shaped pulses with amplitude of 10 V (peak-to-peak) were used for excitation, and the input frequency ranged from 1 to 10 kHz. The electrodes were connected to an LCR meter in order to measure the electrical resistance (frequency = 50 kHz).

Cool water was circulated through copper tubes that coiled around the reaction cell. The temperature inside the reaction cell was controlled by two water coolers, which had different temperatures (i.e., Water cooler 1 at 3 ◦C, and Water cooler 2 at 15 ◦C). Figure 3 shows the temperature gradient of the inside of the reaction cell, which was formed by two separated cooling systems. The CO2 hydrate stability zone was developed in the middle part of the reaction cell (i.e., height of 0.35–0.75 m in the reaction cell).

**Figure 2.** *Cont.*

**Figure 2.** Conceptual drawing of the high-pressure cell used in this study; PT: Pressure transducer, TC: Thermocouple, VP: Piezoelectric plates for P-wave velocity measurements, ER: Electrode for electrical resistance measurements. (**a**) Vertical cross-sectional drawing of the cell. (**b**) Horizontal cross-sectional drawing of the layers A1–A5 and B1–B4, respectively.

**Figure 3.** Temperature distribution of the high-pressure cell obtained from the preliminary test result with distilled water. The maximum CO2 hydrate equilibrium temperature is approximated using the second quadruple point of CO2–water mixture (the intersection of the water–CO2 vapor–CO2 liquid, and water–CO2 hydrate–CO2 vapor equilibrium line) because the water–CO2 hydrate–liquid CO2 equilibrium line is essentially vertical in the pressure–temperature diagram [22].

#### *2.3. Experimental Procedure*

The experiment involved three procedures: (1) The preparation of a water-saturated sample; (2) the injection of the CO2 liquid; and (3) the depressurization of the cell. When the CO2 liquid was injected from the bottom of the cell, the injected CO2 moved upward in the water-saturated sediment sample due to its buoyancy. Then, CO2 hydrates formed within the CO2 hydrates stable zone, which was located at the middle part of the cell (refer to Figure 3). The depressurization test was performed after CO2 hydrate formation to evaluate the behavior of CO2 during hydrate dissociation. These procedures are detailed within this subsection.
