*2.3. CCC Sample Preparation*

The compacted clay samples were prepared according to the wopt and *ρ*dmax of the clay materials. The soil can reach the maximum degree of compaction only when it satisfies the wopt and *ρ*dmax. This is because Ralph R. Proctor proposed a compaction test, where a soil sample is compacted by means of a set of blows of a hammer per lift, which prove that the maximum dry density (*ρ*dmax) of soil is related to certain moisture, called the optimum moisture content (wopt) [45]. The soil samples with the size of 61.8 × 20 mm were prepared by the static compression method. The dry density and moisture content of all the compacted clay samples were 1.78 and 25.6%, respectively. The moisture content was in accordance with the optimum water content (wopt).

The degree of compaction was designed based on the authors' previous study [46], which proved that compacted clays with 90% degree of compaction had superior gas barrier performance than those with lower degrees of compaction. The specific ZP-modified compacted samples were prepared as follows: (1) thoroughly mix the air-dried in-situ soils with a certain amount of ZP powder. The ZP dosage is shown in Table 2 (2) distilled water was added until the moisture content reached 25.6%, and the soil–water mixture was thoroughly stirred using a glass rod until uniform; (3) the above soil–water mixture was added into a rigid mold with a size of 61.8 × 20 mm, and it was statically compacted using a hydraulic jack to the designed degree of compaction of 90%; (4) CCC samples were sealed with plastic bags and cured in the curing room for 14 days until the samples reached the moisture balance.

The CCC samples used in the gas permeability and gas diffusion tests were prepared with the same method with that used in the WRC tests samples. It is noted that the samples were subjected to drying at various times of 0 h, 3.5 h, 7 h, 10.5 h, 14 h, and 17.5 h, respectively, and the drying temperature was 100 ◦C (see Table 2).

### *2.4. Test Methods*

#### 2.4.1. WRC Tests

After soil sample preparation, the compacted clay samples were placed in aluminum boxes. The opening percentage of the aluminum box was greater than 85%. The opening percentage is the ratio of hole area to the total lid area. Studies [47] show that PAM solutions can maintain at least half their original viscosity for more than 8 years at 100 ◦C and for approximately 2 years at 120 ◦C. Its backbone can remain stable at high temperatures. So, the soil samples were dried in an oven with a temperature of 100 ◦C in order to save time. The quality changes in the samples were recorded every 3.5 h, and Equations (1) and (2) were used to calculate the moisture content and water loss rates after drying.

The water loss rates of the clay samples after drying were calculated as follows:

$$\text{W}\_{\text{s}} = \frac{\text{m}\_{0} - \text{m}}{\text{m}\_{\text{s}} \times 25.6\%} \times 100\% \tag{1}$$

where w is the moisture content; ms is the sum of the mass of the clay and polyacrylamide (g); m0 is the mass of a compacted clay sample before being baked (g); mt is the mass of a compacted clay sample after drying for time t (g), and WS is the water loss rate (%).

## 2.4.2. Clay Gas Permeability Tests

The CCC samples used in the gas permeability tests were prepared with the same method as those used in the WRC samples. It is noted that the samples were subjected to drying at various times of 0 h, 3.5 h, 7 h, 10.5 h, 14 h, and 17.5 h, respectively, and the drying temperature was 100 ◦C. Gas permeability is defined according to Darcy's equation as the factor of proportionality between the ratio of gas flow and the pressure gradient along the flow distance. The gas permeability of the CCC samples was measured immediately after drying for 10.5 h at 100 ◦C. The air permeability of soil was measured using the Eijkelkamp-type air permeability apparatus (model 08.07, Eijkelkamp Agrisearch Equipment, The Netherlands) with the following procedures: firstly, the instrument should be tested for tightness before the test. The preliminary test results revealed that it had good airtightness and no air leakage. Secondly, a thick rubber sealing ring (inside diameter: 50 mm, outside diameter: 70 mm, thickness: 10 mm), a perforated plate (diameter: 53 mm, thickness: 1 mm, the diameter of the hole: 1 mm), and a compacted soil sample were placed into a sample holder (inside diameter: 105 mm, outside diameter: 150 mm, height: 50 mm, material: stainless steel) in turn and fixed with a clamp for subsequent sealing. Finally, the flow meter (range: 0.1~10 L/min, accuracy: 1.25%) was switched on by turning the button counterclockwise to a vertical position to control the air pressure within an acceptable range for measurement. Each measurement was repeated for 5 replications within a max. of 10 min and enabled the quantification of pneumatic soil properties.

The testing time should be kept as short as possible during the experiments. This is because gas will dry out the soil sample. Three identical samples were prepared for testing. The air pressure was set as 10 kPa. It is noted that the gravimetric moisture content change in the samples before and after the tests was found to be insignificant, i.e., within 3%. This is attributed to the relatively low gas flow rate (0.5 L/min) and short testing time (10 min). Rouf et al. have proven that a maximum variation in gravimetric moisture content of ±5% was deemed acceptable during the gas permeability tests [48].

The gas permeability (Kp) of the compacted clay samples in the tests was derived in accordance with the published studies [49,50].

$$\mathbf{K}\_{\mathbf{p}} = \frac{\mathbf{k} \cdot \boldsymbol{\rho}\_1 \cdot \mathbf{g}}{\mu} \tag{2}$$

$$\mathbf{Q} = \frac{\mathbf{k} \cdot \mathbf{A} \cdot \mathbf{P}}{\mu \cdot \mathbf{L}} \tag{3}$$

$$\mathbf{Q} = \mathbf{v} \cdot \mathbf{A} = \frac{\mathbf{v}}{\mathbf{t} \cdot \mathbf{A}} \cdot \mathbf{A} = \frac{\mathbf{v}}{\mathbf{t}} \tag{4}$$

$$\mathbf{K}\_{\mathbf{P}} = \rho\_1 \cdot \mathbf{g} \cdot \frac{\mathbf{V} \cdot \mathbf{L}}{\mathbf{t} \cdot \mathbf{P} \cdot \mathbf{A}} \tag{5}$$

where Kp is the gas permeability of a clay sample (m2); k is the permeability coefficient (m/s); *ρ*<sup>l</sup> is the air density (kg/m3); t is the test time (s); V is the amount of air passing through the sample within time *t* (m3); L is the thickness of the compacted clay sample to be tested (m); p is the actual pressure value (MPa), and A is the bottom area of the sample (m2).

### 2.4.3. Gas Diffusion Tests

The CCC samples used in the gas permeability tests were prepared with the same method as that used in the WRC tests samples. It is noted that the samples were subjected to drying at various times of 0 h, 3.5 h, 7 h, 10.5 h, 14 h, and 17.5 h, respectively, and the drying temperature was 100 ◦C. The testing apparatus used for this study is schematically shown in Figure 2. The apparatus consisted of a 3D printing gas diffusion chamber, an oxygen sensor (KE-25, Figaro Inc., Tokyo, Japan) with the accuracy of ±1%, and a datalogger (CR1000, Campbell Scientific, Inc., Logan, UT, USA), and a computer. The oxygen sensor was used to measure the concentration of oxygen with a unit of % (the concentration defined based on the content of oxygen in the atmosphere, 21%). The location of this oxygen sensor is in the bottom of the diffusion chamber. The size of the diffusion chamber is shown in Figure 2. The soil sample is placed at the top of the diffusion chamber; thus, the oxygen gas can only migrate into the chamber through the soil sample. A soil sample can be rapidly inserted into the apparatus and sealed to prevent the loss of gases and water and can be removed without further soil disturbance. The steps of gas tightness self-tests were conducted as follows: (1) an air-tight PVC lib was covered on the top of the chamber, and (2) the nitrogen gaseous were released to fill the chamber. The oxygen concentration should be decreased to the threshold of 0.3~0.6%. (3) It can be considered that the chamber has good air tightness when the variation range of oxygen concentration is less than 0.3%. The specific practical steps were listed as follows: (1) the samples were placed at the top of the diffusion chamber; (2) the silicon grease was evenly applied to the contact part between the samples and the inner wall of the diffusion chamber. Applying the silicon grease was found to be effective in preventing gas migration through the gap in the contact part; (3) the air inlet and outlet valves were opened, and the nitrogen gas container was opened to introduce nitrogen into the diffusion chamber through the inlet. Furthermore, we adjusted the flow control valve to make nitrogen enter the diffusion chamber evenly and steadily through the inlet. The oxygen sensor was installed to monitor the oxygen concentration change in the diffusion chamber until all the oxygen was discharged, and (4) the oxygen was released when the oxygen concentration decreased to the threshold of 0.3–0.6%. The inlet was closed after injecting nitrogen for 5~15 s.

**Figure 2.** The 3D printing gas diffusion chamber.

In this research, a data acquisition instrument was applied to record the oxygen concentration in the diffusion chamber (Ct) every 5 min until the concentration was equal to the oxygen in the atmosphere (C0), i.e., a steady state was reached.

The gas diffusion coefficients Dp of the compacted clay samples were calculated based on the previous studies [51,52] and combined with Fick's first law.

$$\mathbf{D}'\_{\mathbf{P}} = \mathbf{h}\_{\mathbf{s}} \cdot \mathbf{h}\_{\mathbf{c}} \cdot \mathbf{k} \tag{6}$$

$$\ln\left(\frac{\Delta\mathbf{C}\_{\mathbf{t}}}{\Delta\mathbf{C}\_{0}}\right) = \mathbf{k} \cdot \mathbf{t} \tag{7}$$

$$\mathbf{K}\_{\mathbf{j}} = \frac{\mathbf{D}\_{\mathbf{P}}}{\mathbf{D}\_{\mathbf{P}}^{\prime}} = \frac{\varepsilon}{\alpha\_1^2} \cdot \frac{1}{\mathbf{h}\_{\mathbf{s}} \cdot \mathbf{h}\_{\mathbf{c}}} \tag{8}$$

$$\varepsilon = 1 - \frac{\rho\_{\rm b}}{\rho\_{\rm s}} - \Theta\_{\rm V} \tag{9}$$

where Dp is the correction gas diffusion coefficient (m2/s); Dp' is the gas diffusion coefficient before calibration (m2/s); hs is the height of the compacted clay sample (m); hC is the height of the diffusion chamber (m); k is the slope of the straight line in the scatter diagram of ln(ΔCt/ΔC0) (dimensionless); ΔCt is the difference between OC at both ends of the clay sample at time t (g/cm3); ΔC0 is the difference between OC at both ends of the clay sample at time t0 (t0 means initial stage); Kj is the corrected coefficient introduced with changes in the storage capacity of OC; α<sup>1</sup> is the first solution of the equation (α × hS)tan(α × hS) = (hS × ε)/hC greater than 0; ε is the gas-filled porosity [53]; *ρ*<sup>b</sup> is the bulk density of the soil sample (g/cm3); *ρ*<sup>s</sup> is the particle density of the clay (g/cm3); and θ<sup>v</sup> is the volumetric moisture content of the clay sample.

#### 2.4.4. Liquid Limit Test

The liquid limit test was conducted according to ASTM D4318 [39]. The liquid limit and plastic limit data were obtained from a liquid–plastic combined tester according to the standard test method SL237-007-1999 [54].

#### **3. Results and Discussion**

#### *3.1. Liquid Limit*

Figure 3 shows the relationship between the liquid limit of the modified compacted clay and the content of ZP. The liquid limit increases with the increasing ZP, which is higher than the unmodified clay (in Table 1). The liquid limits of the compacted clay were 45%, 49%, 52%, 54%, and 55%, when the corresponding ZP content ranged from 0.2 to 1%. Previous studies have demonstrated [55–57] that the WRC of clay increases with LL values. This is because clay with high LL can retain more content of water in the soil pore, which in turn can reduce gas permeability.

To compare the effects of different modification materials on LL, a dimensionless parameter, LLd, is proposed, which was defined as the ratio of LL to LLck, where LLck is the liquid limit of un-treated clay. The comparison results of LLd are shown in Figure 3b. The results show that the ZP-modified soil has higher LLd values when compared with other modifiers.
