Optimizing the Composition Design of Cement-Based Expanded-Polystyrene (EPS) Exterior Wall Based on Thermal Insulation and Flame Retardance

The use of thermal insulated decorative panel materials with low thermal conductivity and high flame retardance is a key step toward energy-saving buildings. However, traditional thermal insulation materials are always highly conductive and inflammable, which restricts their application for new buildings. This study aims to prepare the non-combustible, cement-based EPS mixtures with thermal conductivity lower than 0.045 and density less than 140 kg/m3 and characterize it with mechanical, thermal, and flame retardant properties. The effect of particle size, Silica coated and content of EPS on the physical, mechanical, thermal, and combustion performance are conducted in this paper. The comprehensive indoor tests including density, water absorbing, softening coefficient, compressive strength, tensile strength, moisture susceptibility, thermal conductivity, and scanning electron microscopy (SEM) along with combustion performance are reported to evaluate the effects of several variables on the investigated cement-based nonflammable EPS (CEPS)mixtures. The results show that small and gradation EPS particles significantly improve the comprehensive performance of mixtures. In addition, Silica coated ESP significantly improve the flame retardance of mixtures while reduce the mechanical characteristics slightly. These results contribute to the selection of appropriate materials to enhance the thermal insulation, flame retardance and mechanical properties of CEPS.


Introduction
Exterior wall insulation materials should be built with sufficient strength, thermal insulation, and heat insulation performance, to achieve building energy conservation and socially sustainable development [1]. Conventionally, two types of materials have been commonly used, those being organic and inorganic [2]. The organic materials are represented by EPS and extruded polystyrene plate (XPS), while typical representatives of inorganic materials are rock wool board and perlite [3]. However, both materials have some limitation. Lightweight EPS and XPS provide excellent thermal insulation performance, but it is flammable [4]. The EPS and XPS have low strength and are prone to aging, resulting in poor stability and durability [5,6]. In addition, the synthesizing organic materials is high-costed and difficult to recycle [7]. The inorganic materials with stable performance are fire-proof and flame retardant, but with poor thermal insulation performance, and its thermal conductivity is difficult to lower than 0.06. In the past 30 years, rock wool had been used in large quantities [8]. After absorbing water, its bulk density increases, and its thermal insulation characteristic deteriorates rapidly [9]. Worse more, it is unfriendly to construction personnel, so it has been listed as a prohibited material by the management department.
Recently, inflammable insulation board made of inorganic materials composite polystyrene, known as cement-based EPS exterior materials has gained increasing interest as an alternative to the conventional materials [10,11]. Since it is compressed by the mixture of EPS and cement-based materials, it has the advantages of completely bypassing the flammable and thermal insulation concerns [12]. In order to investigate the characteristic of the composite thermal insulation materials, various studies have been conducted [13]. Many studies on the relationship among the thermal conductivity the density, and the EPS content of EPS concrete showed that the thermal conductivity was in line with content [14][15][16]. Wang used cone calibration and butane torch flame burning tests to study the flame retardancy of EPS concrete coated with nanomaterials, the peak heat release rate (HRR), the total heat release (THR), and the total smoke release (TMR) of the composites [17]. In addition, some other studies raised the concerns on the potential effects of various factors, such as material composition, mixture ratio, and impact of additives on the comprehensive performance the mixture [18,19].
The previous studies on the EPS mixture with the density of more than 300 kg/m 3 focused on the fire resistance and its thermal conductivity which is generally greater than 0.06; meanwhile, the studies on the density less than 200 kg/m 3 focus on the thermal insulation performance [20,21]. To this end, this study aims to prepare the non-combustible cement-based EPS mixtures with thermal conductivity lower than 0.045 and density lower than 140 kg/m 3 . In order to achieve this objective, comprehensive laboratory tests, thermal conductivity tests, and combustion performance tests were conducted to evaluate the effects of several variables on the CEPS samples. In addition, the non-combustibility test was performed to investigate the fracture surface characteristics after flame. The microscopic features of the CEPS interface bonding characteristics were detected by scanning electron microscopy (SEM). This paper is expected to reveal the characteristic mechanism of CEPS exterior wall insulation material.

Materials and Sample Preparation
In this study, the EPS with granules of 2-3 mm and 3-4 mm in size (mean diameter) and a density of 5 kg/m 3 was provided by Hunan Yue yang Baling Petrochemical Co., Ltd. (Yueyang, Hunan province, China). The cementing materials with Portland cement P.O. of 42.5-80% by weight, was provided by Henan Xin sheng building energy saving decoration Co., Ltd. (Xinxiang, Henan province, China). The ratio of water to cementing materials (W/C) was kept 0.45 consistently for all of the CEPS mixtures. Tap water was used to mix the mixture. Seven CEPS mixtures were investigated in this paper. The control CEPS was made by EPS with particle size 3 mm-4 mm, cementing materials 70 kg/m 3 , and water to cementing materials (W/C) ratio 0.45. The other six EPS with different particle sizes, different contents of cementing materials, and mixing coated EPS with cementitious materials were designed to study the comprehensive properties. Table 1 provides the detailed description of the prepared CEPS. The material parameters in Table 1 are from manufacturer's recommendations and preliminary test results. The Coated-EPS-CEPS in Table 1 was first adhered to the surface of the EPS using epoxy resin and then mixed with the slurry.  Figure 1 shows the preparation process and test items of the CEPS. First, the Portland cement, fly ash, nitride, and aluminum oxide were used as the main raw materials with a composite binder to produce a cement-based slurry. Second, a special equipment was used to mix the EPS and slurry evenly, and then the specimens were made at the fixed compression ratio (generally 40%), and maintained at ambient temperature for 3 days to 7 days. Finally, the specimens were sawn to external wall insulation panels at a given size. The thickening agent in Figure 1 was mainly composed of sodium ethoxylated alkyl sulfide and sodium dodecyl sulfonate. mers 2022, 14, x FOR PEER REVIEW 3 of The thickening agent in Figure 1 was mainly composed of sodium ethoxylated alkyl s fide and sodium dodecyl sulfonate.   [5], as shown in Figure 2.
The compressive strength test was conducted in accordance with GB/T 5486. The compression specimen with size of 100 mm × 100 mm × 50 mm, were compressed at the deformation rate of 10%/min of specimen thickness. The peak compressive stress was taken as the compressive strength value, which shall not be less than 0.15 MPa. The mechanical properties of CEPS were evaluated by three tests, including compressive strength test, immersion tensile strength and freeze-thaw strength test, and flexural strength test [5], as shown in Figure 2. The compressive strength test was conducted in accordance with GB/T 5486. The compression specimen with size of 100 mm × 100 mm × 50 mm, were compressed at the deformation rate of 10%/min of specimen thickness. The peak compressive stress was taken as the compressive strength value, which shall not be less than 0.15 MPa.
The standard specification JG/T 287 was referred to for testing the tensile strength and moisture susceptibility of CEPS. The specimen with a side length of 50 mm was pasted onto the metal plate with high-strength resin adhesive, and the specimen was stretched to the strength at the time of destruction at the speed of 5 ± 1 mm/min [18]. A dry-wet tensile strength test, and freeze-thaw tensile strength test were conducted to comprehensively evaluate the moisture characteristics. After immersing the sample in water at 25 °C for 2 days, the surface was wiped and tested using the dry tensile strength test method. Similarly, after immersing the sample in water for 3 days, the sample was frozen at −20 ± 2 °C for 3 h, and then the surface of the sample was wiped for tensile strength test. Four replicates were prepared and tested.

Thermal Conductivity
The thermal conductivity of CEPS was measured using steady-state method in accordance with ISO 8302. The instrument is shown in Figure 3. The steady-state method is based on the principle of the heat transfer balance, that is, the heat transfer rate is consistent with the heat dissipation rate [22]. If the temperature gradient at both ends of the material is measured and the heat flow through its unit area is known, the value of the The standard specification JG/T 287 was referred to for testing the tensile strength and moisture susceptibility of CEPS. The specimen with a side length of 50 mm was pasted onto the metal plate with high-strength resin adhesive, and the specimen was stretched to the strength at the time of destruction at the speed of 5 ± 1 mm/min [18]. A dry-wet tensile strength test, and freeze-thaw tensile strength test were conducted to comprehensively evaluate the moisture characteristics. After immersing the sample in water at 25 • C for 2 days, the surface was wiped and tested using the dry tensile strength test method. Similarly, after immersing the sample in water for 3 days, the sample was frozen at −20 ± 2 • C for 3 h, and then the surface of the sample was wiped for tensile strength test. Four replicates were prepared and tested.

Thermal Conductivity
The thermal conductivity of CEPS was measured using steady-state method in accordance with ISO 8302. The instrument is shown in Figure 3. The steady-state method is based on the principle of the heat transfer balance, that is, the heat transfer rate is consistent with the heat dissipation rate [22]. If the temperature gradient at both ends of the material is measured and the heat flow through its unit area is known, the value of the thermal conductivity can be calculated through the Fourier formula shown in Equation (1).
where, λ m -thermal conductivity coefficient, W/(m. K), q-constant heat flux on one side of the heating plate, W/m 2 , ∆T-temperature difference between cold plate and hot plate, K.
where, -thermal conductivity coefficient, W/(m. K), q-constant heat flux on one side of the heating plate, W/m 2 , ΔT-temperature difference between cold plate and hot plate, K.

Combustion Tests
The combustion performance of CEPS was measured by two parameters according to the standard GB 8624, including heat of combustion and fire growth rate index (FIGRA).
Gross heat of combustion (PCS) was calculated by Equation (2) according to the standard GB/T 14402.
where PCS is gross heat of combustion (MJ/kg), E is the water equivalent other than water in calorimetric system (MJ/K), Ti and Tm is the starting and maximum temperature (K), respectively, b is the correction value of combustion calorific value of combustion supporting materials used in the test (MJ), c is temperature correction value of external heat exchange (K), and m is the weight of specimens (kg). The fire growth rate index (FIGRA) is calculated according to the GB/T 20284. FIGRA is the maximum value of the heat release rate (HRR) of the sample combustion to its corresponding time, which is used to classify combustion performance. The larger the FIGRA, the easier the material is to burn; the faster the fire grows, the higher the fire risk coefficient. The FIGRA0.2MJ is the combustion growth rate index when the heat released from the combustion of the sample reaches 0.2 MJ. THR600s is the total heat release of the specimen in the first 600 s (300 s ≤ t ≤ 900 s) after fired by the main burner.
The butane combustion test is an intuitive qualitative index to evaluate the flame retardance of exterior wall insulation materials [4]. In this paper, CEPS specimens were directly exposed to butane flame with temperature up to 1500 °C for 1 h, the combustion process was recorded and the difference of residues after combustion was compared.

Combustion Tests
The combustion performance of CEPS was measured by two parameters according to the standard GB 8624, including heat of combustion and fire growth rate index (FIGRA).
Gross heat of combustion (PCS) was calculated by Equation (2) according to the standard GB/T 14402.
where PCS is gross heat of combustion (MJ/kg), E is the water equivalent other than water in calorimetric system (MJ/K), T i and T m is the starting and maximum temperature (K), respectively, b is the correction value of combustion calorific value of combustion supporting materials used in the test (MJ), c is temperature correction value of external heat exchange (K), and m is the weight of specimens (kg). The fire growth rate index (FIGRA) is calculated according to the GB/T 20284. FIGRA is the maximum value of the heat release rate (HRR) of the sample combustion to its corresponding time, which is used to classify combustion performance. The larger the FIGRA, the easier the material is to burn; the faster the fire grows, the higher the fire risk coefficient. The FIGRA 0.2MJ is the combustion growth rate index when the heat released from the combustion of the sample reaches 0.2 MJ. THR 600s is the total heat release of the specimen in the first 600 s (300 s ≤ t ≤ 900 s) after fired by the main burner.
The butane combustion test is an intuitive qualitative index to evaluate the flame retardance of exterior wall insulation materials [4]. In this paper, CEPS specimens were directly exposed to butane flame with temperature up to 1500 • C for 1 h, the combustion process was recorded and the difference of residues after combustion was compared.

Scanning Electron Microscopy (SEM) Tests
In order to further investigate the interfacial transition zone (ITZ) of CEPS, SEM was employed. FEI Quanta 250 FEG (FEI, Hillsboro, OR, USA) apparatus was used for capturing the SEM images. The microstructures of the CEPS specimens were analyzed by SEM images.

Physical and Mechanical Properties
In this study, the main purpose is to meet the requirements of mixture density between 120 and 140 kg/m 3 ; the compressive strength, tensile strength, and flexural strength greater than 0.15 MPa, 0.1 MPa, and 0.2 MPa, respectively. The tensile strength ratio and softening coefficient equal or greater than 75% and 0.7, respectively; water absorption equal or less than 10%. Figure 4a presents the density results of different CEPS. As it can be seen, cement-based binder affects the density of mixtures the most. The density of single particle size-EPS (SS) is small. At the same compression ratio, the particle size of EPS has no effect on the density. The results have also been acknowledged by other researchers [11,23]. The water absorption test data of the seven CEPS groups are plotted in Figure 4b. It can be seen that the water absorption of SS, LC and LP are relatively large. In terms of the single particle size and high content of EPS particles, large voids in the structure mean high water absorption [24]. The water absorption of EPS with a large particle size is larger than that of EPS with a small particle size, and the particle size of larger diameter EPS cause larger voids. The softening coefficient test results are shown in Figure 4c. The greater the softening coefficient, the less the material is affected by the external environment, and the better the freeze-thaw resistance and aging resistance. The softening coefficient is most affected by cement-based binder. Sufficient content and reasonable distribution of cement-based binder are the key to improving the softening coefficient of CEPS. When EPS particle size is small and cement-based binder is evenly distributed, the softening coefficient will be higher after EPS is coated. The relationship between softening coefficient and water absorption is analyzed, and the results are shown in Figure 4d. The data show that the correlation between the two is not strong. High water absorption of the material does not mean low softening coefficient, and vice versa. That is because the content of cement-based binder and its distribution in CEPS are the key factors affecting the softening coefficient.
The mechanical test results of CEPS are shown in Figure 5. As illustrated in Figure 5a, the compressive strength and flexural strength of low EPS content are the highest, followed by the small particle size of high EPS content. With the increase of particle size, the compressive and tensile strength decrease. The compressive strength and tensile strength of the larger EPS particle size are lower than that of the higher EPS content [25]. There is a similar trend in the pull strength test results. It shows that the performance of cementbased binder is the primary factor affecting the strength.

Scanning Electron Microscopy (SEM) Tests
In order to further investigate the interfacial transition zone (ITZ) of CEPS, SEM was employed. FEI Quanta 250 FEG (FEI, Hillsboro, OR, USA) apparatus was used for capturing the SEM images. The microstructures of the CEPS specimens were analyzed by SEM images.

Physical and Mechanical Properties
In this study, the main purpose is to meet the requirements of mixture density between 120 and 140 kg/m 3 ; the compressive strength, tensile strength, and flexural strength greater than 0.15 MPa, 0.1 MPa, and 0.2 MPa, respectively. The tensile strength ratio and softening coefficient equal or greater than 75% and 0.7, respectively; water absorption equal or less than 10%. Figure 4a presents the density results of different CEPS. As it can be seen, cement-based binder affects the density of mixtures the most. The density of single particle size-EPS (SS) is small. At the same compression ratio, the particle size of EPS has no effect on the density. The results have also been acknowledged by other researchers [11,23]. The water absorption test data of the seven CEPS groups are plotted in Figure 4b. It can be seen that the water absorption of SS, LC and LP are relatively large. In terms of the single particle size and high content of EPS particles, large voids in the structure mean high water absorption [24]. The water absorption of EPS with a large particle size is larger than that of EPS with a small particle size, and the particle size of larger diameter EPS cause larger voids. The softening coefficient test results are shown in Figure 4c. The greater the softening coefficient, the less the material is affected by the external environment, and the better the freeze-thaw resistance and aging resistance. The softening coefficient is most affected by cement-based binder. Sufficient content and reasonable distribution of cement-based binder are the key to improving the softening coefficient of CEPS. When EPS particle size is small and cement-based binder is evenly distributed, the softening coefficient will be higher after EPS is coated. The relationship between softening coefficient and water absorption is analyzed, and the results are shown in Figure 4d. The data show that the correlation between the two is not strong. High water absorption of the material does not mean low softening coefficient, and vice versa. That is because the content of cement-based binder and its distribution in CEPS are the key factors affecting the softening coefficient.   The mechanical test results of CEPS are shown in Figure 5. As illustrated in Figure  5a, the compressive strength and flexural strength of low EPS content are the highest, followed by the small particle size of high EPS content. With the increase of particle size, the compressive and tensile strength decrease. The compressive strength and tensile strength of the larger EPS particle size are lower than that of the higher EPS content [25]. There is a similar trend in the pull strength test results. It shows that the performance of cementbased binder is the primary factor affecting the strength. As shown in Figure 5 b, the influence of the freeze-thaw cycle on the mixture: tensile strength ratio (TSR) with lower EPS content is the highest, followed by small particle size with higher EPS content. In terms of the high elasticity of EPS, the volume required for ice expansion after water enters the gap of CEPS is achieved by compressing EPS, so the freeze-thaw strength is significantly higher than that of ordinary concrete.   The mechanical test results of CEPS are shown in Figure 5. As illustrated in Figure  5a, the compressive strength and flexural strength of low EPS content are the highest, followed by the small particle size of high EPS content. With the increase of particle size, the compressive and tensile strength decrease. The compressive strength and tensile strength of the larger EPS particle size are lower than that of the higher EPS content [25]. There is a similar trend in the pull strength test results. It shows that the performance of cementbased binder is the primary factor affecting the strength. As shown in Figure 5 b, the influence of the freeze-thaw cycle on the mixture: tensile strength ratio (TSR) with lower EPS content is the highest, followed by small particle size with higher EPS content. In terms of the high elasticity of EPS, the volume required for ice expansion after water enters the gap of CEPS is achieved by compressing EPS, so the freeze-thaw strength is significantly higher than that of ordinary concrete.  As shown in Figure 5b, the influence of the freeze-thaw cycle on the mixture: tensile strength ratio (TSR) with lower EPS content is the highest, followed by small particle size with higher EPS content. In terms of the high elasticity of EPS, the volume required for ice expansion after water enters the gap of CEPS is achieved by compressing EPS, so the freeze-thaw strength is significantly higher than that of ordinary concrete.

Thermal Conductivity
The lower the thermal conductivity of the external wall insulation material, the less likely the heat is to penetrate the wall, and the better the energy-saving effect of the building. In general, the thermal conductivity of EPS particles is 0.039 W/(m. K) [26]. To make the thermal conductivity of the EPS mixture as close as possible to 0.039 W/(m. K), the cement-based binder should be added as less as possible. On the contrary, with less cement-based binder, the flame retardant effect will become worse [27]. In previous studies, the thermal conductivity of organic materials is generally about 0.039, even lower than 0.039 W/(m. K) after adding aerogel gel materials (high costed but low strengthened material), while the thermal conductivity of inorganic insulation materials is generally greater than 0.07 W/(m. K), as shown in the literature [1,24,28].
The target thermal conductivity of the mixture designed in this paper is less than or equal to 0.045 W/(m. K). The test results of thermal conductivity are presented in Figure 6. The thermal conductivity with higher EPS content is the lowest, followed by that with higher EPS content but smaller particle. With the decrease of particle size, the thermal conductivity decreases, the thermal conductivity with a certain gradation is better. In the unit volume of CEPS, the number of small particles is larger than large particles, and the compact packing density is larger. That is, during the heat transfer process, the pore interface between the pores of CEPS cement-based binder and EPS particles increases, so the heat conduction path is extended, which slows down the heat conduction rate of CEPS, and finally reduces the thermal conductivity. ment-based binder, the flame retardant effect will become worse [27]. the thermal conductivity of organic materials is generally about 0.0 0.039 W/(m. K) after adding aerogel gel materials (high costed but lo terial), while the thermal conductivity of inorganic insulation ma greater than 0.07 W/(m. K), as shown in the literature [1,24,28].
The target thermal conductivity of the mixture designed in this equal to 0.045 W/(m. K). The test results of thermal conductivity are 6. The thermal conductivity with higher EPS content is the lowest, fo higher EPS content but smaller particle. With the decrease of partic conductivity decreases, the thermal conductivity with a certain grada unit volume of CEPS, the number of small particles is larger than lar compact packing density is larger. That is, during the heat transfer p terface between the pores of CEPS cement-based binder and EPS pa the heat conduction path is extended, which slows down the heat cond and finally reduces the thermal conductivity.

Combustion Behavior
The combustion characteristics of the samples include the hea combustion value, total heat release within 600 s-THR600s, and the com index FIGRA0.2MJ. HRR represents the speed and size of heat release and it also reflects the ability of the fire source to release heat. The g more heat the combustion feedback gives to the surface of the mate accelerated pyrolysis speed of the material and the increase in the pr combustibles, thus accelerating the spread of the fire. Generally, the H into three stages: the initial growth stage, full development stage an [29]. The combustion value of building materials is an important param the potential fire risk of building materials, and it is the essential ba the heat released by the combustion of building materials and the fire

Combustion Behavior
The combustion characteristics of the samples include the heat release rate HRR, combustion value, total heat release within 600 s-THR 600s , and the combustion growth rate index FIGRA 0.2MJ . HRR represents the speed and size of heat released by the fire source, and it also reflects the ability of the fire source to release heat. The greater the HRR, the more heat the combustion feedback gives to the surface of the material, resulting in the accelerated pyrolysis speed of the material and the increase in the production of volatile combustibles, thus accelerating the spread of the fire. Generally, the HRR curve is divided into three stages: the initial growth stage, full development stage and weakening stage [29]. The combustion value of building materials is an important parameter to characterize the potential fire risk of building materials, and it is the essential basic data to calculate the heat released by the combustion of building materials and the fire load. It can be used to evaluate the potential fire risk of building materials products. The other two characteristics have been shown in Section 2.3. The HRR of the seven CEPS are shown in Figure 7. Similar to other exterior wall insulation materials, the CEPS, in the initial growth stage, will release a large amount of heat rapidly; but in the second stage, the full development stage the release will not maintain for a long time but directly weaken. The reasons for the performance are that EPS particles are wrapped by cement-based binder, and that the volume content is lower than that of other materials. The HRR of the sample with higher EPS content and larger particles is the highest, followed by the one with higher EPS content but smaller particles. With the increase of particle size, the combustion performance becomes worse, and the HRR of the sample with larger particles is lower than that of higher EPS content. The HRR of the CE is lower than that of any other materials because the coating material prevents the combustion, so the combustion performance is significantly improved. The finding is consistent with previous studies in which silica-coated EPS improved flame retardance, smoke suppression and mechanical strength [17].
to evaluate the potential fire risk of building materials products. The o istics have been shown in Section 2.3.
The HRR of the seven CEPS are shown in Figure 7. Similar to insulation materials, the CEPS, in the initial growth stage, will releas heat rapidly; but in the second stage, the full development stage the re tain for a long time but directly weaken. The reasons for the perfor particles are wrapped by cement-based binder, and that the volume c that of other materials. The HRR of the sample with higher EPS conte cles is the highest, followed by the one with higher EPS content but sm the increase of particle size, the combustion performance becomes wo the sample with larger particles is lower than that of higher EPS cont CE is lower than that of any other materials because the coating m combustion, so the combustion performance is significantly improved sistent with previous studies in which silica-coated EPS improved smoke suppression and mechanical strength [17]. As described in Section 2.3, the combustion value is calculated b target PCS designed in this paper is less than or equal to 3.0 MJ/kg. Th and THR600 test results are shown in Figure 8. The content of EPS is th ing PCs and THR600s. If the content is larger, both are larger. The seco EPS particles. When the particle size is larger, CEPS with the same vo face and is easier to burn. As described in Section 2.3, the combustion value is calculated by PCS. The mixture target PCS designed in this paper is less than or equal to 3.0 MJ/kg. The combustion value and THR 600 test results are shown in Figure 8. The content of EPS is the main factor affecting PCs and THR 600s . If the content is larger, both are larger. The second factor is the size EPS particles. When the particle size is larger, CEPS with the same volume has less interface and is easier to burn.
The target FIGRA 0.2MJ of the mixture designed in this paper is less than or equal to 120 W/s. The result of FIGRA 0.2MJ is shown in Figure 9. As is shown, FIGRA 0.2MJ of the sample with larger particles needs to be strictly controlled. In addition, the nano coating can be used to reduce the value of FIGRA 0.2MJ to a certain extent. The target FIGRA0.2MJ of the mixture designed in this paper is l 120 W/s. The result of FIGRA0.2MJ is shown in Figure 9. As is show sample with larger particles needs to be strictly controlled. In additio can be used to reduce the value of FIGRA0.2MJ to a certain extent. After being exposed in high temperature, EPS is melted, and th are shown in Figure 10. It can be seen that the residues of all test p honeycomb. As EPS is dispersed in the cement-based binder, no drop ing the combustion process, and the shape is not collapsed and defo the requirements of grade A2 flame retardance. The size and content different, showing different structural states. Figure 10d, g show that composed of inorganic substances present dense and uniform mor voids, which can effectively prevent heat and mass transfer between the condensing phase. It results in the reduction of HRR, PCS, THR600s better flame retardance. Figure 10c After being exposed in high temperature, EPS is melted, and the results of residue are shown in Figure 10. It can be seen that the residues of all test pieces are loose and honeycomb. As EPS is dispersed in the cement-based binder, no drops are produced during the combustion process, and the shape is not collapsed and deformed, which meets the requirements of grade A2 flame retardance. The size and content of EPS particles are different, showing different structural states. Figure 10d, g show that the residues mainly composed of inorganic substances present dense and uniform morphology and fewer voids, which can effectively prevent heat and mass transfer between the flame zone and the condensing phase. It results in the reduction of HRR, PCS, THR 600s and FIGRA 0.2MJ , and better flame retardance. Figure 10c,f,h show similar contour characteristics, except that EPS particle sizes and compositions are different; the flame retardance mechanism is similar, so there is no difference in its flame retardance performance. Figure 10e shows that the combustion residue of the sample with EPS coated is not significantly different from that of other samples, but the combustion HRR, THR 600s and FIGRA 0.2MJ are lower, which can provide a precious opportunity for personnel to escape and materials to be rescued in case of fire. In addition, there is no shape collapse after combustion, which is closely related to the addition of nitride, aluminum oxide, etc. in the cement-based binder. The addition of these flame-retardant materials prevents the open fire from further hidden combustion.

Microstructure Analysis
The microstructures of the CEPS are analyzed by SEM imaging to reveal the m nisms of modification. The cement-based binder plays an important role in the bon of EPS. If it is not sufficiently adhesive, it will lead to strength damage, freeze-thaw age, etc., especially in the interface area between EPS and cement-based binder. Thu study focuses on the interface region of binary materials and the internal structure ment-based binder. The microstructure images are shown in Figure 11a-d. Figur shows the microstructure of the cement-based binder after hydration. It can be seen the image that the cement-based binder is partially hydrated, which is composed o drated calcium silicate (C-S-H) gel, calcium hydroxide (C-H) and Ettringite (AFT) cry as well as some voids and cracks. The cement-based binder first forms the crystal nu of the hydration product in the hydration process [30]. The length of the crystal nu means that the hydration product grows up and adheres to each other in a staggered

Microstructure Analysis
The microstructures of the CEPS are analyzed by SEM imaging to reveal the mechanisms of modification. The cement-based binder plays an important role in the bonding of EPS. If it is not sufficiently adhesive, it will lead to strength damage, freeze-thaw damage, etc., especially in the interface area between EPS and cement-based binder. Thus, this study focuses on the interface region of binary materials and the internal structure of cementbased binder. The microstructure images are shown in Figure 11a-d. Figure 11a shows the microstructure of the cement-based binder after hydration. It can be seen from the image that the cement-based binder is partially hydrated, which is composed of hydrated calcium silicate (C-S-H) gel, calcium hydroxide (C-H) and Ettringite (AFT) crystals, as well as some voids and cracks. The cement-based binder first forms the crystal nucleus of the hydration product in the hydration process [30]. The length of the crystal nucleus means that the hydration product grows up and adheres to each other in a staggered way. After curing, the hydration product develops, grows, condenses, hardens, and further solidifies to attain the required strength. In a word, cement-based binder is the skeleton to obtain mechanical strength. As illustrated in Figure 11b, the EPS content of the sample is high, while the content of cement-based binder is low. When it is cut into the specimen, bond failure occurs, and the failure interface is located between EPS particles and cement-based binder. As shown in Figure 11c, the cement-based binder of the sample is relatively sufficient, the interface zones significantly improved. The cement-based binder is evenly distributed on the surface of EPS particles, making the interface adhesion tighter, so the composite strength of CEPS is higher. It can be observed that many EPS particles break at the damaged interface rather than at the interface zones. Figure 11d reveals the microstructure of the CEPS with nano materials coated on the surface of EPS. Compared with other groups of CEPS, CEPS coated with nano materials on the surface of EPS has better interface adhesion, more voids after hydration, more organized speed of heat transmission, and lower thermal conductivity. Therefore, this porous structure has great advantages for improving flame retardance.

Statistical Analysis
In order to statistically study the contribution of EPS particle size and content to the strength, thermal conductivity and flame retardance of CEPS, the influence of significant difference analysis is used. Table 2 summarizes the statistical results of strength, thermal conductivity, and flame retardance of different EPS compositions. The three letters H, M and N are used to represent high significant difference (p < 0.01), medium significant dif- As illustrated in Figure 11b, the EPS content of the sample is high, while the content of cement-based binder is low. When it is cut into the specimen, bond failure occurs, and the failure interface is located between EPS particles and cement-based binder. As shown in Figure 11c, the cement-based binder of the sample is relatively sufficient, the interface zones significantly improved. The cement-based binder is evenly distributed on the surface of EPS particles, making the interface adhesion tighter, so the composite strength of CEPS is higher. It can be observed that many EPS particles break at the damaged interface rather than at the interface zones. Figure 11d reveals the microstructure of the CEPS with nano materials coated on the surface of EPS. Compared with other groups of CEPS, CEPS coated with nano materials on the surface of EPS has better interface adhesion, more voids after hydration, more organized speed of heat transmission, and lower thermal conductivity. Therefore, this porous structure has great advantages for improving flame retardance.

Statistical Analysis
In order to statistically study the contribution of EPS particle size and content to the strength, thermal conductivity and flame retardance of CEPS, the influence of significant difference analysis is used. Table 2 summarizes the statistical results of strength, thermal conductivity, and flame retardance of different EPS compositions. The three letters H, M and N are used to represent high significant difference (p < 0.01), medium significant difference (0.01 < p < 0.05) and no significant difference (p > 0.05).