Latest revision as of 20:38, 21 August 2026

Abstract: To synergistically enhance the construction performance and durability of sulphoaluminate cement-based materials in cold regions, this study investigated the effects of cod antifreeze protein (AFP) on the workability and microstructural evolution of cement mortar. AFP was incorporated at dosages of 0%, 2%, and 4% by mass of cement, and the fluidity, consistency, setting time, and water absorption of the mortar were tested. Moreover, its action mechanism was analyzed combined with SEM-EDS, XRD, and amino acid composition analysis.

The results showed that AFP could effectively improve the workability of cement mortar. Compared with the control group, the incorporation of AFP increased the mortar fluidity by 5.9%–7.1% and consistency by 64.3%–79.5%, and extended the initial setting time from 16.3 minutes to 42.7 minutes. Meanwhile, AFP reduced the water absorption of mortar by 0.31%–0.43% and significantly slowed down the water absorption rate. The results of microcharacterization indicated that the addition of AFP altered the morphology of hydration products and affected the formation and distribution of ettringite (AFt), thereby promoting the densification of the internal structure of the paste. Combined with amino acid composition analysis, it is suggested that the hydrophilic amino acids in AFP may improve the particle dispersion state and retard early hydration through interactions with ions and particle surfaces, while the hydrophobic amino acids may participate in the regulation of the internal structure of the paste.

This study demonstrates that AFP has application potential in improving the workability and impermeability of sulphoaluminate cement mortar, and can provide a new insight for the performance optimization of cement-based materials in low-temperature environments.

Key words: Antifreeze protein; Cement mortar; Workability; Microstructure; Bio-inspired materials

1 Introduction

Cement-based materials are the key structural materials for construction engineering in cold regions, where their frost resistance and construction efficiency directly impact the durability of concrete structures. Traditional anti-freeze agents often come at the expense of compromising workability or causing environmental issues^[1-5], Antifreeze proteins (AFPs), as natural biomolecules, possess the characteristic of inhibiting ice crystal growth^[6,7], Widely applied in the food and pharmaceutical sectors^[8-12]. Research by the Leiter team^[13] on Type III antifreeze proteins in fish indicates, In low-concentration NaCl solutions (20–30 mmol/L), its inhibition of ice recrystallisation activity (IRI) is slightly enhanced compared to pure solutions. The research team further investigated the effect of 0.1 M NaOH (pH 11) on protein activity and found that the alkaline environment did not induce a significant change in ice recrystallization inhibition (IRI) activity. This study reveals the potential feasibility of antifreeze proteins to maintain their functional activity in non-native ionic environments, which is expected to facilitate the introduction of antifreeze proteins into cement-based materials and achieve the synergistic optimization of frost resistance and construction performance.

At present, sulfoaluminate cement has become a popular candidate for engineering in cold regions owing to its characteristics of rapid hardening, high strength and low alkalinity^[14,15]. Existing studies have optimized its performance by incorporating mineral admixtures (e.g., fly ash, nanomaterials) or chemical admixtures (e.g., water reducers), yet most of them focus on mechanical strength or frost resistance, with few systematic studies conducted on its construction performance^[16,17]. Meanwhile, the application of antifreeze proteins in cementitious materials remains in its exploratory phase. Existing research primarily focuses on verifying their efficacy in enhancing the freeze resistance of such materials. However, the influence mechanism of antifreeze proteins on the fluidity, setting time and microstructure of cement-based materials remains unclear.Special attention should also be paid to the changes in rheological properties in the field of cement concrete science induced by antifreeze proteins, which may directly determine the construction method, efficiency and quality of concrete。

Although the introduction of antifreeze proteins provides a new approach for optimizing the performance of cement-based materials, existing studies have the following shortcomings: First, the regulatory mechanism of antifreeze proteins on the workability (such as fluidity and consistency) of sulphoaluminate cement-based materials has not been systematically revealed; Second, the mechanism of antifreeze proteins retarding the hydration reaction and their impact on setting time lack quantitative analysis; Third, the research on the correlation between the amino acid composition of antifreeze proteins and the morphological evolution of hydration products is insufficient, making it difficult to clarify their micro-action mechanism. These shortcomings have restricted the engineering application of antifreeze proteins.

To address these issues, this study investigates the effects of AFP on the workability of low-alkalinity sulfoaluminate cement mortar. Particular attention is given to flowability, consistency, setting time, and water absorption. In addition, microstructural analyses (SEM-EDS and XRD) and amino acid composition analysis are conducted to clarify how AFP interacts with cement hydration and influences the morphology of hydration products. The objective is to establish a link between the macroscopic performance and the underlying microstructural evolution, thereby providing insight into the use of AFP as a bio-based modifier for cementitious materials in cold environments.

2 Test Protocol

2.1 Raw Materials

The cement used was low-alkalinity sulphoaluminate cement produced by Jiuqi Building Materials, whose basic technical properties and chemical composition are shown in Table 1 and Table 2, respectively. The antifreeze proteins adopted were cod antifreeze proteins manufactured by Shanghai Naijin Industrial Co., Ltd. The standard sand was supplied by Xiamen ISO Standard Sand Co., Ltd., and the water used was laboratory-made distilled.

2.2 Frost-resistant protein composition

The statistical classification results of the amino acid composition of the antifreeze protein are presented in Table 3. A total of 17 amino acids were identified. Based on side-chain polarity, hydrophilic and weakly polar neutral amino acids accounted for 79.21% of the total, which was markedly higher than the proportion of hydrophobic amino acids (20.79%), indicating that AFP is overall highly polar in nature. Among them, polar or potentially charged amino acids such as Asp, His, Lys, and Arg were present at appreciable levels, which may facilitate interactions with ions in the cement system and with particle surfaces. In contrast, hydrophobic amino acids may influence the interfacial behavior within the slurry, thereby contributing to the regulation of microbubble formation and microstructural development^[18-20].

2.3 Preparation of Frost-Resistant Protein-Modified Cement Mortar

Antifreeze protein-modified cement mortar was prepared in accordance with the mix proportions given in Table 4, following the Test Method for Strength of Cement Mortars (ISO Method) GB/T 17671-2021 ^[21]. The antifreeze protein was completely dissolved in distilled water before being added into the mortar mixture. Following vibration molding, the prepared specimens were covered with glass plates and placed in a curing chamber. After 24 hours, the specimens were removed from the molds and further cured in the chamber for 14 days.

2.4 Test Methods

2.4.1 Freeze-Thaw Resistance Protein Cement Mortar Flow Test

The flowability of modified cement mortar was tested in accordance with the method specified in the Test Procedure for Polymer-Modified Cement Mortar (DL/T 5126-2021)^[22].

2.4.2 Tests on Consistency, Setting Time and Water Absorption of Antifreeze Protein Modified Cement Mortar

The consistency, setting time, and water absorption rate of modified cement mortar were tested according to the methods specified in the Standard Test Methods for Basic Properties of Building Mortar (JGJ/T 70-2009)^[23]. The water absorption rate was calculated using Equation (1).

（1）

Where: WA denotes water absorption rate (%); G₀ denotes the dry mass of the specimen (g); Gn denotes the mass after n hours of water absorption (g).

2.4.3 Microscopic Morphology Test

The microscopic morphology of the specimens was characterized using a Thermo Fisher QuattroS scanning electron microscope (SEM) and an EDAX ELECTPIUS energy dispersive spectrometer (EDS). The middle part of each specimen was cut into blocks with smooth surfaces and dimensions of 8 mm × 8 mm × 4 mm using a cutting machine. These blocks were then immersed in anhydrous ethanol for 48 h to terminate hydration, followed by drying in an oven at 60 °C for another 48 h. Before the test, the samples were sputter-coated with gold and then placed on the test stage for characterization.

2.4.4 Phase Composition Test

The specimens were characterized using a Rigaku SmartLab SE X-ray diffractometer (XRD) manufactured in Japan. The specimens, whose hydration had been terminated with anhydrous ethanol, were dried in an oven at 60 °C for 48 h, then ground into powder using an agate mortar and sieved through a 0.075 mm square-hole sieve. A total of 30 mg of the uniformly mixed sample powder was subjected to X-ray diffraction scanning at a scanning rate of 5°/min and a scanning angle range of 10–80° (2θ). The phase composition analysis of the scanning results was performed using Jade 6.0 software.

A schematic overview of the experimental procedure is shown in Fig. 1.

3 Test Results

3.1 Analysis of Cement Mortar Fluidity

The relationship between freeze-resistant protein dosage and cement mortar flowability is illustrated in Figure 2. As shown, cement mortar flowability increases with rising freeze-resistant protein dosage. The flow value of cement mortar without freeze-resistant protein was 237 mm. At a 2% freeze-resistant protein dosage, the flow value reached 251 mm (a 5.9% increase). and at 4% freeze-resistant protein content, the flowability reached 254 mm (a 7.1% increase).

Figure 2 Relationship between Fluidity of Cement Mortar and Dosage of Antifreeze Protein

3.2 Analysis of Cement Mortar Consistency

Figure 3 presents the evolution of cement mortar consistency as a function of antifreeze protein content. It can be observed that the average consistency of mortar increases noticeably with rising antifreeze protein dosage.The average consistency of plain mortar without antifreeze protein is 66 mm. Upon adding 2% antifreeze protein, the average consistency rises markedly to 108.5 mm, representing an increase of 64.4%. When the dosage is further raised to 4%, the average consistency climbs to 118.5 mm, corresponding to a 79.5% improvement. In terms of developmental trend, the strengthening effect of antifreeze protein on mortar consistency is most pronounced within the low dosage range of 0%–2%. With a continued increase in dosage, although consistency still increases, its growth rate gradually slows down.This suggests that antifreeze protein exerts the most significant enhancement on the workability of cement mortar at low dosages, while its improving effect tends to level off at higher dosages. The inclusion of antifreeze protein can effectively raise the average consistency of cement mortar, especially under low dosage conditions.

Figure 3 Relationship between cement mortar consistency and frost-resistant protein dosage

3.3 Analysis of Water Absorption in Cement Mortar

Figure 4 depicts the time-dependent evolution of water absorption for specimens with various antifreeze protein dosages after 28 days of curing.As illustrated in Figure 7(a), the water absorption rate of specimen S0 (0% antifreeze protein) increased rapidly at the initial stage, reaching a peak of approximately 3.36% at 5 hours, and then remained constant thereafter.For specimens modified with antifreeze protein, the water absorption rate rose sharply within the first 5 hours, followed by a gradual slowdown in the growth rate.At a dosage of 2% antifreeze protein, the water absorption rate of specimen S2 stabilized after 16 hours, with a maximum value of roughly 3.05%.When the antifreeze protein dosage was increased to 4%, the water absorption rate of specimen S4 reached a steady state after 45 hours, with a peak value of about 2.93%.

Figure 4 Water absorption rate of cement mortar specimens (a) versus water absorption rate (b) versus frost-resistant protein dosage

As presented in Figure 4(b), the average water absorption of specimen S2 is reduced by 0.31% compared with specimen S0 after the addition of antifreeze protein, while that of specimen S4 is decreased by 0.43%.The introduction of antifreeze protein observably lowers the water absorption of aluminate cement specimens.This demonstrates that the protein can exert a pore-blocking effect during cement hydration, thus effectively reducing water absorption.Such characteristics are beneficial to improving the durability and frost resistance of cement-based materials.

3.4 Analysis of Cement Mortar Setting Time

Figure 5 Relationship between cement mortar setting time and frost-resistant protein dosage

Figure 5 illustrates the relationship between the setting time of cement mortar and the dosage of antifreeze protein.With an increase in antifreeze protein content, the initial setting time of cement mortar is obviously prolonged.As the dosage rises from 0% to 2%, the initial setting time increases from 16.27 minutes to 30.84 minutes.When the dosage is further raised to 4%, the setting time extends to 42.74 minutes, suggesting that antifreeze protein exerts an obvious retarding effect on cement setting.

Such retarding behavior is mainly attributed to the inhibition of antifreeze proteins on the early hydration of cement.At the initial stage of cement hydration, a large amount of Ca²⁺ is rapidly dissolved and released.The abundant carboxyl groups (-COOH) on the molecular chains of antifreeze proteins can coordinate with Ca²⁺ to form soluble complexes, which reduces the concentration of free calcium ions in the liquid phase and delays the nucleation and crystallization of hydration products such as C-S-H gel and ettringite.In addition, antifreeze proteins can adsorb onto the surface of cement particles to form an organic film, physically blocking the contact between water and cement particles and further slowing down the hydration reaction rate.

3.5 Microstructural Analysis

3.5.1 Phase Composition Analysis

The mineral composition of hydration products in cement mortar specimens was characterized by X-ray diffraction (XRD), and the corresponding results are displayed in Figure 6. It is revealed that the main hydration products of cement mortar include ettringite (AFt), calcite, and calcium silicate hydrate (C-S-H).The addition of antifreeze protein observably increases the diffraction peak intensity of AFt, suggesting a significant rise in its content, which means antifreeze protein contributes to the formation of ettringite.When the antifreeze protein dosage is 2%, the amount of AFt in specimen S2 reaches the maximum.The AFt peak intensity of specimen S4 is slightly lower than that of S2, but still higher than the blank control group, indicating a slight reduction in AFt content which remains superior to the reference group S0.In addition, the number of impurity peaks decreases and the characteristic peaks of ettringite become more distinct in S2 and S4 groups, demonstrating an improvement in the purity and crystallinity of hydration products.

3.5.2 Microstructural Analysis

Figure 7 and Figure 8 present the SEM and EDS results of the specimens after curing. SEM observations show that in specimen S0 without antifreeze protein, the hydration product ettringite (AFt) exists in the form of thick, long rod-shaped crystals. These crystals interlock with each other but form a loose structure containing a large number of internal pores. When the dosage of antifreeze protein is 2% (S2), the AFt crystals are significantly refined into needle-like and fine rod-like shapes with a denser distribution, and some voids are filled by gel products such as C-S-H. When the dosage increases to 4% (S4), the crystals are further refined into needle-like and flocculent aggregates, forming a tighter interlocking structure with obviously improved compactness and greatly reduced pores.

Combined with EDS analysis, the main elements in the red-marked regions of the three images are Ca, Al, S and O, corresponding to ettringite (AFt), a typical hydration product of sulphoaluminate cement.Trace amounts of Si and K are also detected, which is consistent with the morphological analysis.According to the mass fraction of the characteristic element S, the S contents of S0, S2 and S4 are 10.9%, 17.8% and 12.1% respectively, revealing that antifreeze protein at 2% dosage effectively promotes the formation of AFt.When the dosage rises to 4%, the AFt content decreases slightly but remains higher than that of the control group, which agrees with the SEM observations. The Ca/S ratio of S2 is closer to the theoretical value of ettringite, indicating higher product purity and more uniform element distribution, corresponding to its dense needle structure. The Ca/S ratio of S4 deviates slightly from the theoretical value. Combined with its flocculent morphology, it is inferred that excessive protein adsorbs on the surface of AFt and slightly affects the stoichiometric ratio of crystals.Nevertheless, AFt remains the dominant hydration product, and the structural compactness is still superior to that of the control group, consistent with the XRD findings.

Antifreeze protein can refine the morphology of AFt, promote its formation at an appropriate dosage and regulate the element proportion, thus optimizing the composition and distribution of hydration products.This serves as an important mechanism for its improvement on the microstructure of cement mortar.

4 Discussion

The improvement in the workability of cement mortar by antifreeze protein (AFP) can be interpreted via its synergistic effects in particle dispersion, hydration retardation, microbubble formation, and crystal morphology regulation^[24].

AFP enhances mortar fluidity mainly by improving particle dispersion. As presented in Figure 2, the fluidity rises from 237 mm in the control group to 251 mm and 254 mm with the addition of 2% and 4% AFP, respectively. This enhancement is mainly due to the adsorption of hydrophilic amino acids on cement particle surfaces, which generates electrostatic repulsion and restrains particle agglomeration^[25].

AFP also exerts an obvious retarding effect on cement hydration. The initial setting time is extended from 16.3 min for the control sample to 30.8 min and 42.7 min for mixtures containing 2% and 4% AFP (Figure 5). This effect is closely related to the reaction between carboxyl-rich amino acids and Ca²⁺ released in the early hydration stage, which lowers the free calcium ion concentration and slows the nucleation and growth of hydration products^[26]. Moreover, the adsorption of AFP on particle surfaces further prevents water from contacting cement grains^[27].

The notable increase in mortar consistency can be ascribed to the microbubble effect induced by AFP. As shown in Figure 3, the consistency increases from 66 mm to 108.5 mm and 118.5 mm when 2% and 4% AFP are added, respectively. Hydrophobic amino acids in AFP favor the formation of microbubbles during mixing, which expand the paste volume, reduce interparticle friction, and thus improve lubrication and consistency^[28].

At the microscale, AFP modifies the morphology of hydration products. SEM and EDS analyses (Figures 7 and 8) reveal that AFP promotes the transformation of AFt crystals from short columnar shapes into long rod-like or fibrous forms, constructing a denser three-dimensional network. This microstructural evolution corresponds to the reduced water absorption shown in Figure 4, where water absorption decreases by 0.31% and 0.43% in the 2% and 4% AFP groups, respectively. Although excessive AFP may slightly disturb the Ca/S ratio through adsorption on crystal surfaces, the overall microstructure remains denser than that of the control group^[29].

In summary, AFP improves the workability and compactness of cement mortar through a combined mechanism involving dispersion, retardation, lubrication, and microstructural optimization. These results provide a theoretical foundation for applying AFP as a bio-based admixture in cementitious materials for construction in cold regions.

5 Conclusion

Antifreeze protein (AFP) shows a significant influence on the workability and microstructure of low-alkalinity sulfoaluminate cement mortar. The addition of 2-4 wt.% AFP improves flowability (5.9-7.1%) and consistency (64.3-79.5%), while extending the initial setting time from 16.3 to 42.7 min. A reduction in water absorption (0.31-0.43%) indicates a more compact pore structure.

These changes are associated with the interaction between AFP and cement hydration. Hydrophilic components interact with Ca²⁺, reducing particle agglomeration and slowing hydration, whereas hydrophobic components promote microbubble formation, improving paste lubrication. In addition, AFP alters the morphology of AFt, leading to a denser microstructure.

An AFP dosage of around 2 wt.% provides a suitable balance between performance improvement and material efficiency. Overall, AFP can serve as an effective bio-based admixture for enhancing mortar performance, particularly under cold conditions.

Acknowledgements: This work was partially supported by the Basic Research Project of Higher Education Institutions in the Xinjiang Uygur Autonomous Region (Grant No. XJEDU2025J049) and the Talent Research and Development Project of Xinjiang Agricultural University (Grant No. 6660946/2522GCCRC).

.

References

[1] Sun Q. Study on Influence of Different Gas Content Performance of Hydraulic Concrete. Heilongjiang Hydraulic Science and Technology. 2024, 52 (8):1-4+42. https://doi.org/10.14122/j.cnki.hskj. 2024. 08. 001

[2] Ramachandran V S. Calcium chloride in concrete—Applications and ambiguities. Canadian Journal of Civil Engineering, 1978, 5(2): 213-221

[3] Yuan Q, Zhou D, Huang H, et al. (2020) Structural build-up, hydration and strength development of cement-based materials with accelerators. Construction and Building Materials 259: 119775. https://doi.org/10.1016/j.conbuildmat. 2020.119775

[4] Xunqin B, Linhua J, Ming-Zhi G, et al. (2022) Chloride erosion resistance of calcium formate incorporated cement mortar under chloride attack. Construction and Building Materials 314: 125611. https://doi.org/10.1016/j.conbuildmat. 2021.125611

[5] Yifei W, Lei L, Jianhui L, et al. (2022) Accelerators for normal concrete: A critical review on hydration, microstructure and properties of cement-based materials. Cement and Concrete Composites 134: 104762. [https: https:] //doi.org/10.1016/j.cemconcomp. 2022. 104762

[6] Griffith M, Ala P, Yang D S C, et al. Antifreeze protein produced endogenously in winter rye leaves. Plant Physiology, 1992, 100(2): 593-596.

[7] Duman J G, Olsen T M. Thermal hysteresis protein activity in bacteria, fungi, and phylogenetically diverse plants. Cryobiology, 1993, 30(3): 322-328.

[8] Xiangjian L, Yuxin P, Fenglin L, et al. (2021) A Review of the Material Characteristics, Antifreeze Mechanisms, and Applications of Cryoprotectants (CPAs). Journal of Nanomaterials 2021:9990709. https://doi.org/10.1155/2021/9990709

[9] Song H D, Kim M, Jin E, et al. (2019) Cryoprotective effect of an antifreeze protein purified from Tenebrio molitor larvae on vegetables. Food Hydrocolloids 94:585-591. https://doi.org/10.1016/j.foodhyd.2019.04.007

[10] Xiangli D, Tingting L, Hui Z, et al. (2020) Effect of Barley Antifreeze Protein on Dough and Bread during Freezing and Freeze-Thaw Cycles. Foods (Basel, Switzerland) 9(11):1698. https://doi.org/10.3390/FOODS9111698

[11] Xiang H, Yang X, Ke L, et al. (2020) The properties, biotechnologies, and applications of antifreeze proteins. International Journal of Biological Macromolecules 153(prepublish): 661-675. https://doi.org/10.1016/j.ijbiomac. 2020. 03. 040

[12] Tony C, Rosa M. (2019) Psychrophilic lifestyles: mechanisms of adaptation and biotechnological tools. Applied Microbiology and Biotechnology 103(7): 2857-2871. https://doi.org/10.1007/s00253-019-09659-5

[13] Leiter A, Rau S, Winger S, et al. (2016) Influence of heating temperature, pressure and pH on recrystallization inhibition activity of antifreeze protein type III. Journal of Food Engineering 187: 53-61. https://doi.org/10.1016/j.jfoodeng. 2016.04.019

[14] Shi YR.Nondestructive testing of strength and volume stability of sulphoaluminate cement mortar. Xining：Ningxia University, 2022.

[15] Gui Y. Study on Hydration Characteristics and Maturity of Calcium Sulfoaluminate Cement-based Materials. Wuhan: Wuhan University of Science and Technology, 2020.

[16] Qin Q, Miao FK, Zhao XZ, et al. Comparative Study on Influence of Different Replacement Methods Steel Slag Sand on Workability and Mechanical Properties of Dry-mixed Mortar. Journal of Xinjiang Agricultural University, 2022, 45(1): 79-86.

[17] Zhu S, Zhang X Y, Lee K C. (2025) Long-term mechanical properties of hybrid fiber-reinforced engineered cementitious composite. Structural Concrete 26(6):7076-7089. https://doi.org/10.1002/SUCO.202400902

[18] Jihui Z, Xuan G, Siying C, et al. (2022) Hydrophobic or superhydrophobic modification of cement-based materials: A systematic review. Composites Part B 243:110104. https://doi.org/10.1016/J.COMPOSITESB.2022.110104

[19] Friis D S, Ramløv H. Physicochemical Properties of Antifreeze Proteins[M]//Antifreeze Proteins Volume 2: Biochemistry, Molecular Biology and Applications. Cham: Springer International Publishing, 2020: 43-67.

[20] Hao Y, Zonglin X, Chuhan H, et al. (2021) Recent progress of hydrophobic cement-based materials: Preparation, characterization and properties. Construction and Building Materials 299: 124255. https://doi.org/10.1016/J. CONBUILDMAT. 2021. 124255

[21] GB/T 17671-2021. Test method of cement mortar strength (ISO method). Beijing, China: Standards Press of China; 2021.

[22] DL/T 5126-2021. Test Specification for Polymer-Modified Cement Mortar. Beijing, China: China Electric Power Press, 2021.

[23] JGJ/T 70-2009. Standard Test Methods for Basic Properties of Masonry Mortar. Beijing, China: China Architecture & Building Press, 2009.

[24] Gong J, Chen H, Xie G, et al. A Protein for Enhancing the Frost Resistance of Cement-Based Materials: Characterization, Performance, and Mechanism. Journal of Building Engineering, 2025: 114780.

[25] Xiang H, Yang X, Ke L, et al. (2020) The properties, biotechnologies, and applications of antifreeze proteins. International Journal of Biological Macromolecules 153(prepublish): 661-675. https://doi.org/10.1016/j.ijbiomac. 2020.03.040

[26] E D M, R N C, I M G. (2015) Rational, yet simple, design and synthesis of an antifreeze-protein inspired polymer for cellular cryopreservation. Chemical Communications (Cambridge, England) 51(65): 12977-12980. https://doi.org/10.1039/c5cc04647e

[27] Christof S, A. K E, Wanwipa S, et al. (2022) Recent progress in superabsorbent polymers for concrete. Cement and Concrete Research 151:106648. https://doi.org/10.1016/J.CEMCONRES.2021.106648

[28] Jihui Z, Xuan G, Siying C, et al. (2022) Hydrophobic or superhydrophobic modification of cement-based materials: A systematic review. Composites Part B 243:110104. https://doi.org/10.1016/J.COMPOSITESB.2022.110104

[29] Frazier D S, Matar G M, Osio-Norgaard J, et al. (2020) Inhibiting Freeze-Thaw Damage in Cement Paste and Concrete by Mimicking Nature’s Antifreeze. Cell Reports Physical Science 1(6): 100060. https://doi.org/10.1016/j.xcrp. 2020. 100060