2.4.1. Test A Method: One Cycle of Cryogenic Temperature Condition Up to −196 ◦C

The test specimens were subject to a single cycle at very low temperature and subsequently tested. The test results were compared with the concrete characteristic strength (40 MPa). The description and specimens of the test were specified as shown in Table 9. Test specimens were directly immersed into liquid nitrogen for 15, 30 and 60 min. at −196 ◦C and put out from the insulated storage. The cooled specimens were stored at the curing room (23 ◦C and RH 95%) where the temperature and moisture could be constantly controlled for 48 h. Reference specimens were kept in the curing room at the same time. After 2 days, the mechanical properties of cryogenic concrete were measured in terms of compressive strength, elastic modulus, poisson ratio, splitting tensile strength, length change for thermal expansion coefficient and thermal conductivity. For measuring the temperature automatically, two thermocouples were installed inside a spare specimen and one sensor was set up outside the spare specimen. The installed locations of thermocouples from the surface of a concrete sample were 25 mm and 75 mm, respectively. The temperature was recorded with a data logger and the temperatures of 15 min, 30 min and 60 min were corresponded to test specimens exposed to the cryogenic condition. The temperature recording was finished before immersing all test specimens into liquid nitrogen. As shown in Figure 3, the cryogenic tests were carried out as follows: compressive strength, tensile strength, elastic modulus, moisture content, length change and thermal conductivity.

(**a**) Preparation of liquid nitrogen (**b**) Preparation of specimen (**c**) Immersed into liquid nitrogen

(**d**) Contained for 15/30/60 min. (**e**) Pilot specimen (**f**) Curing in chamber

(**g**) Compressive strength test (**h**) Elastic modulus, Poisson ratio (**i**) Splitting tensile strength

(**j**) Thermal conductivity (GHP) (**k**) Specimens of thermal conductivity (**l**) Length change ratio

2.4.2. Test B Method: Cyclic Temperature Condition between 5–−20 ◦C

In Figure 4, test specimens were subject to 50 cycles between 5–−20 ◦C and subsequently tested and test results were compared with the concrete characteristic strength (40 MPa). After finishing freezing and thawing, all samples were maintained in the curing chamber and crushed at the same time as the cooled specimen. The temperature change rate might not be greater than 10 ◦C per hour. The freeze-thaw cycling test was carried out according to the ASTM C 666. The cooling and heating procedure of the cycle met conditions defined in ASTM C 666. The temperature of specimens was monitored throughout the test used by the thermocouples. The cycled specimens were taken out from the chamber and cycled and un-cycled specimens were crushed at the same time. After finishing the compressive strength tests, the moisture content was determined with the crushed debris dried in an oven for 24 h.

**Figure 4.** Test B method under cryogenic condition: (**a**) freezing-thawing equipment, (**b**) test set-up with specimens and (**c**) freeze-thaw cycles.

#### **3. Test Results and Discussions**

#### *3.1. Selection of Optimum Mix Design with Slump and Compressive Strength*

During 120 min after casting, slump tests with admixture Daracem 208 were performed every 30 min, and the testing value of cryogenic concrete was described in Figure 5. The initial slump of the C40-1 mixture exceeded the allowable tolerance, but the other slumps were in the range of it. On the other hand, the slump value of the C40-2 mixture was satisfied during the retention time of 120 min.

The chemical admixture of C40-3 and C40-4 was used with a polycarboxylate type (Baxel PC 650). The values of slump flow for 120 min were demonstrated in Figure 6. Based on the results, the initial slump flow of the C40-4 mixture with 65% GGBS was satisfied with the tolerance, but the C40-3 mixture with 60% GGBS was not. After 120 min, both of them were in the range of slump flow, 595 mm and 570 mm, respectively. These values were contained within the target range of 620 ± 75 mm. The tolerance of slump flow was within ±75 mm and the C40-4 mixture was more suitable rather than C40-3 one, in terms of retention time and workability.

All test samples were cast in accordance with ASTM C 172 and several times of compressive strength tests were conducted for 28 days [49]. The target compressive strength was determined to be more than 50 MPa, including a 10 MPa margin. As shown in Figure 7, the target strengths of all mix designs were sufficiently developed at the age of 7 days. All specimens at the age of 28 days already satisfied the compressive strength of more than 60 MPa. For describing in detail, the binder content of C40-1 and C40-2 was greater than that of C40-3 and C40-4 by 15 kg/m3, but the compressive strength of C40-1 and C40-2 was lower than that of C40-3 and C40-4. Therefore, with respect to the workability of fresh concrete, C40-2 and C40-4 were better than C40-1 and C40-3. In view of the development of compressive strength, C40-2 and C40-4 were superior to C40-1 and C40-3. In the next step for cryogenic tests, C40-2 and C40-4 mixture were chosen.

**Figure 5.** Slump test (**a**) C40-1 (GGBS 50%, Daracem 208) (**b**) C40-2 (GGBS 65%, Daracem 208).

**Figure 6.** Slump flow test (**a**) C40-3 (GGBS 60%, Baxel PC 650) (**b**) C40-4 (GGBS 65%, Baxel PC 650).

**Figure 7.** Compressive Strength of concrete mix design over time.

*3.2. Mechanical, Thermal and Durability Properties under Cryogenic Condition*

3.2.1. Mechanical and Thermal Properties after Exposed to a Cryogenic Condition

To figure out the characteristics of the concrete exposed to the cryogenic condition, firstly, two thermocouples in the specimen were installed as shown in Figure 8a. Then, the concrete specimen was put into liquid nitrogen to measure the temperature variation over

time. As the result, the initial temperature was started from 22 ± 2 ◦C and after 45 min, the temperature was dropped down up to −190 ± 2 ◦C as shown in Figure 8b. The measured temperature variation was regarded as the temperature of the other specimens without thermo-couples. That is, after 15 min, the temperature of the concrete surface went down up to −100 ◦C. If the specimen was immersed in the liquid nitrogen storage for 15 min, it would be equally considered to be exposed to −100 ◦C.

**Figure 8.** Thermocouple installation and temperature variation: (**a**) sample size and sensor location and (**b**) temperature variation over time.

In Table 11, the compressive strength and elastic modulus of concrete specimens under cryogenic temperature tended to be decreased over time. When the specimen was frozen under very low temperature, the expansion of ice crystal made some cracks in the capillary pores. Because the compression was loaded to concrete specimens after melting the ice crystal under ambient temperature, compressive strength of the specimens decreased. However, this trend was not the same as mentioned in ACI 376. That is, many studies had shown that compressive strength and elastic modulus increased as the temperature went down [28,50–53]. As the temperature decreased, the moisture contained in the capillary pores was changed into ice, and the internal structure of the concrete become tight and dense, resulting in an effect of increasing strength. The compressive strength of concrete exposure to cryogenic temperature rose up to about three times compared with ambient temperature. That is, the increment rate in compressive strength of concrete increased as the moisture content increased and the temperature decreased. In particular, below −120 ◦C, the deviation of compressive strength increased and the increment rate decreased. This was because the volume of ice crystals rapidly reduced around −120 ◦C or below [52].



In Table 11, the reason for the decline of mechanical properties was the expansion of volume. Freezing moisture expanded its volume by about 10%. The expanded volume of ice in the capillary pores caused pressure increase and the excess of the tensile strength of the pore walls impacted on the occurrence of cracks of concrete microstructure. Additionally, the degree of water saturation had a significant effect on the frost resistance of the concrete mix. As shown in Figure 9, the compressive strength and elastic modulus declined over time. That is, in the case of compressive strength, the strength reduction of the C40-2 mixture was about 15% in Figure 9a, but in the case of elastic modulus, 10% of reduction of the C40-4 mixture was observed in Figure 9b. This is because freezing moisture in concrete pores induced the cracks in the pore walls and this cracking resulted in the reduction of mechanical properties [50].

**Figure 9.** Normalized mechanical properties exposed to cryogenic temperature: (**a**) normalized compressive strength over temperature and (**b**) normalized elastic modulus over temperature.

For calculating the coefficient of temperature expansion (CTE), the length change of concrete specimens was measured as shown in Table 12. The CTE of cryogenic concrete was derived as shown in Equation (1) and the CTE of C40-2 and C40-4 was −1.503 and −1.605 × <sup>10</sup><sup>−</sup>6/◦C, respectively.

$$C = \frac{(R\_h - R\_I)}{G \cdot \Delta T} \tag{1}$$

where, *C* = coefficient of linear thermal expansion of the concrete (10−6/◦C), *Rh* = length reading at higher temperature (mm), *Rl* = length reading at lower temperature (mm), *G* = gage length between inserts (mm) and Δ*T* = difference in temperature of specimen between the two length readings (◦C).


**Table 12.** Test results of length change and splitting tensile strength.

For the measure of the thermal conductivity, GHP (Guarded Hot Plate) 456 Titan manufactured by NETZSCH in Selb, Germany was used in Figure 10 and was employed with standardized guarded hot plate technique according to ASTM C 177 [54]. The temperature range of the equipment was −160–(+250) ◦C and the range of thermal conductivity was 0.003 to 2 W/(m·K). The two samples of each mix design were prepared and the size of it was a square with 300 mm sides and with 90 mm thick due to the measurement

limit of equipment. GHP (Guarded Hot Plate) was based on the absolute measurement method without calibration and correction. The thermal conductivity value resulted in the stationary state and was derived from Equation (2) as follows:

$$
\lambda = \frac{\dot{Q} \times d}{2A \times \Delta T} \tag{2}
$$

**Figure 10.** GHP (Guarded Hot Plate) 456 Titan and set-up of the sample, thermal sensors and plate.

In Equation (1), . *Q* is precisely measured total power input into the hot plate, *d* is average sample thickness, *A* is measurement area and Δ*T* is mean temperature difference along the sample.

In ACI 376, the moisture content in concrete have an effect on the thermal conductivity. As temperature goes down, the thermal conductivity rises up linearly. In detail, the thermal conductivity of partially saturated normal-weight concrete increases from approximately 3.2 W/(m·K) at 25 ◦C to 4.71 W/(m·K) at −155 ◦C [28]. Table 13 demonstrated the thermal conductivity of concrete exposed to ambient and cryogenic temperature.


**Table 13.** Test results of thermal conductivity.

As a result, the thermal conductivity of C40-2 and C40-4 at ambient temperature (20 ◦C) was about 1.5 W/(m·K) and that of C40-2 and C40-4 at very low temperature (−160 ◦C) was 0.643 and 0.723, respectively. This result was opposite to what ACI 376 mentioned. The factors affecting the thermal conductivity were the ratio of aggregate volume, water-cement ratio, moisture content and curing period. That is, as the volume fraction of aggregate and moisture content increased as well as water-cement ratio and curing period deceased, the thermal conductivity tended to be increased [55]. On the contrary, the test error could be decreased with the thicker specimen. For verifying this tendency, additional thermal conductivity tests were carried out with thinner sample as shown in Figure 11. As the sample thickness was decreased up to 50 mm, the thermal conductivity was down up to 50% or more.

**Figure 11.** Effect of sample thickness on thermal conductivity.

3.2.2. Mechanical Properties after Exposed to Cyclic Low Temperature

The test B indicated that the cyclic temperature was repeated for 50 cycles in the range of 5 ◦C to −20 ◦C, in accordance with ASTM C 666. After exposed to the freeze-thaw conditions, the compressive strength and elastic modulus tests were carried out with two types of mix designs, C40-2 and C40-4. In Table 14, as the freeze-thaw cycles increased, the compressive strength and elastic modulus decreased because the volume expansion of ice crystal in the pores induced the microcracks under a low temperature [7,20]. Thus, the mechanical properties of the C40-4 mixture were better than those of the C40-2 mixture, in terms of the reduction rate of compressive strength and elastic modulus. The reduction rate of normalized mechanical properties was shown in Figure 12. In detail, the strength reduction was about 10%, and in the case of elastic modulus, the reduction was about 5% less.


**Table 14.** Test results of compressive strength, elastic modulus, poisson ratio and absorption.

**Figure 12.** Normalized mechanical properties exposed to cyclic low temperature: (**a**) normalized compressive strength over F-T cycles and (**b**) normalized elastic modulus over F-T cycles.
