1 Introduction

With continuous social development, annual consumption of non-renewable resources is increasing, and the economic cost of construction materials is rising. In China, huge amounts of agricultural and forestry solid waste are produced annually, but their utilization rate is low, and most are landfilled or incinerated, such as corn cobs, rice husks, wheat straw, fallen leaves, and waste wood. Such disposal methods cause considerable CO2 emissions, occupy large amounts of land, reduce land utilization and agricultural development, and are inconsistent with human sustainable development [1,2,3,4,5]. Therefore, utilizing agricultural and forestry solid waste to prepare biomass-recycled concrete has become a widespread concern. Waste wood and expanded perlite have lightweight, porous, and strong heat flow resistance characteristics, providing potential advantages for recycled aggregate selection.

Utilizing biomass aggregates for concrete preparation can effectively improve building energy consumption. Researchers have found that concretes prepared with biomass aggregates such as pine needles, canola straw, coconut husk, corn cob, and corn stover exhibit reduced density and thermal conductivity with increasing biomass content, along with improved durability and plasticity [6,7,8,9,10,11,12,13]. This indicates potential advantages of biomass aggregates in construction materials. For instance, using 40% coconut husk in concrete mixes can result in 22% lower strength and 7.5% lower density compared to conventional mixes [6]. Zhang et al. [14] investigated use of corn stover as aggregate for evaluating hygrothermal properties of concrete, finding addition of corn stover improved these properties. Moreover, shape and volume fraction of corn stover affected hygrothermal properties of concrete. Additionally, use of biomass ash to replace cement was investigated by Festus et al. [15,16,17], who found rice husk ash could partially or completely replace cement, reducing cement usage in concrete. Due to prismatic siliceous structure of rice husk ash under microstructure, its use reduced porosity, improved the fracture characteristics and brittleness of concrete, and enhanced the durability properties such as resistance to sulfate attack, acid attack, and chloride ion penetration. Moreover, compressive strength of 5% rice husk ash concrete reached over 70% of conventional concrete under different curing days. Vinícius et al. [18, 19] utilized sugar cane bagasse ash (SCBA) instead of blast furnace slag and cement for concrete preparation. SCBA improved workability and durability but adversely affected compressive and tensile strengths to some extent. Among them, the compressive strength of the SCBA system reached about 60 MPa after 270 days curing at 20 °C. Second, waste wood as a biomass aggregate for lightweight building material preparation has garnered extensive researcher attention due to advantages including good toughness, impact resistance, and heat flow resistance. Waste wood chip replacement of fine aggregates in mortar and concrete shows increasing wood chip content caused a gradual decrease in concrete density and compressive strength, a decrease in drying shrinkage, and an increase in crack resistance. Compressive strength decreased approximately 10% at 10% chip substitution and 40% at 60% substitution [20,21,22,23,24]. Whereas the preparation of wood fiber-based concrete using waste wood found that wood fiber addition can significantly increase the mechanical strength of concrete, especially the flexural strength. Secondly, the high friction between the surface of wood fibers and the liquid components had a significant effect on the fluidity of the paste [25]. On the effect of maintenance methods on wood fiber-based mortar, Kittipong et al. [26] found that the ratio of flexural-compressive strength of the mortar was relatively low under high-pressure curing conditions and high under conventional and low-temperature thermal curing conditions. Mortar with 5% wood fiber waste added under conventional curing conditions had the highest flexural-compressive strength ratio at 28 days.

Perlite is a type of volcanic rock, composed of silicate minerals. During heating, the water in perlite evaporates, causing the rock to expand into a porous granular material. Expanded perlite is thus a highly porous, low-density material with low thermal conductivity and good thermal insulation properties. It is widely used in construction, agriculture, industry, and horticulture. When expanded perlite is used as an aggregate in construction materials (e.g., concrete and mortar) it can reduce the thermal conductivity, increase the durability of concrete, improve its resistance to chemical attack, and the strength of concrete can be effectively adjusted according to the amount of perlite used [27,28,29,30].

The studies above have shown using wood chips, fibers, and expanded plastic (EP) as fine aggregates can effectively reduce weight, conserve construction energy, and benefit the environment. However, research on producing lightweight concrete with multi-source aggregates from agroforestry waste and other minerals remains relatively scarce. This study prepared lightweight concrete using waste wood as coarse aggregate and EP instead of natural fine aggregate. Its dry density, mechanical properties, thermal properties, chloride penetration resistance, and frost resistance were evaluated. The aim is to explore making multi–source aggregate–based lightweight concrete.

2 Experimental program

2.1 Materials

Sulphoaluminate cement (SAC) with a density of 2960 kg/m3 was selected as the cementitious material. Expanded Perlite (EP) is a kind of acidic volcanic lava with a glassy structure. It has a bulk density of 125 kg/m3, a thermal conductivity of 0.05 W/(mK), and a water absorption rate of 287%, as shown in Fig. 1c. The XRD test was used to determine the chemical composition of the various materials, as shown in Table 1. The fine aggregate was natural fine aggregate (NFA) with a bulk density of 2646 kg/m3, as shown in Fig. 1a.

Fig. 1
figure 1

a Natural fine aggregate, b Wood aggregate and c Expanded perlite

Full size image
Table 1 Chemical composition of EP and SAC (%)
Full size table

In the test, the wood aggregate (WA) was made of broadleaf wood, selected and collected from the furniture industry in Qingdao. Its specific chemical composition is shown in Table 2. This type of wood has low thermal conductivity, high toughness, and porous characteristics. The water absorption rate of WA was 80%. The WA was selected according to the GB/T 14685-2011 specification. To ensure the volume stability of lightweight thermal insulation recycled concrete (LTIRC), non-needle aggregates with a thickness greater than 0.4 times the average particle size were used. Therefore, two consecutive crushing processes were carried out, and then WA with a particle size of 4.75–9.5 mm was selected, as shown in Fig. 1b.

Table 2 The chemical composition content of wood agregate
Full size table

2.2 Mix design and preparation

To improve the thermal insulation performance and reduce the bulk weight, EP partially replaced NFA (20%, 30%, 50%). Additionally, WA doping ranged from 20 to 40%, with a 5% gradient increase. The water-reducing agent effectively improved LTIRC fluidity, so its dosage was 1.2% of cementitious material. Table 3 shows specific compositions.

Table 3 Mix design optimization (kg/m3)
Full size table

In preparing LTIRC, due to the high-water absorption of both WA and EP, SAC and NFA were first dry–mixed for 30 s to ensure thorough mixing. Then, 70% of the water was added slowly. Once sufficiently mixed, WA, EP, and water-reducing agents were added. The remaining 30% of water was added after 60 s of mixing to ensure homogeneity. The mixture was then poured into the test mold and vibrated for 8 s to avoid aggregate floating due to the low density of EP and WA. Figure 2 shows the preparation process.

Fig. 2
figure 2

Preparation process of LTIRC

Full size image

2.3 Testing method

2.3.1 Dry density measurement

A specimen with a size of 100 mm × 100 mm × 100 mm was selected for dry density determination. Firstly, the specimens were maintained for 28 d. In this case, the conservation method was standard conservation (temperature 20 ± 2 °C, relative humidity greater than 95%). In the second step, the specimens were put into the drying box to set the temperature at 105 ± 5 ℃ and then dried for 48 h. During this period, the quality was measured every 12 h. When the difference in quality between the front and back was not more than 0.5%, it was judged that the constant weight had been reached, as shown in Fig. 3b.

Fig. 3
figure 3

a Thermal conductivity instrument, b Measurement of dry density

Full size image

2.3.2 Mechanical property test

The test used 100 mm × 100 mm × 100 mm specimens to determine the mechanical properties of 3 d, 7 d, 14 d and 28 d age specimens. According to GB/T 50081–2019, the compressive resistance of three specimens in each group was determined and averaged (the deviation of three specimens was greater than 15%, which required retesting). In addition, the correction factor was 0.95.

2.3.3 Measurement of thermal conductivity

According to the relevant literature and specifications [31], the resistance of LTIRC to heat flow through can be expressed by Eq. (1), in which the average thermal conductivity was determined by a thermal conductivity tester, and the 300 mm × 300 mm × 35 mm specimen with standard curing for 28 d was placed into the instrument for the determination. The hot plate temperature was set to 35 °C and the cold plate temperature was set to 15 °C, while the testing time was set to 1.5 h, as shown in Fig. 3a. In addition, the thermal conductivity of gas is the smallest among the three phases: solid, liquid, and gas. Therefore, the thermal conductivity in the saturated state was found to be about 16.3–27.2% higher than that in the dry state in the pre-test. In this study, to accurately test the relationship between pores and thermal conductivity in porous materials (WA and EP), drying treatment is required before thermal conductivity testing. The treatment is the same as in the dry density test.

$$\lambda = \frac{\phi \times d}{{\Delta T \times S}}$$
(1)

λ is the thermal conductivity (W/m·k), φ is the average thermal conductivity (J/s), d is the thickness of the heat conduction plate (mm), here take 35 mm, △T is the temperature difference between the two surfaces of the specimen (°C), and s is the specimen area (m2).

2.3.4 Test of resistance to chloride ion penetration

The tests were conducted using the RCM method for chloride ion penetration test. According to the relevant specifications and literature [32, 33], the diameter of the test block was 100 mm ± 1 mm and the height was 50 mm ± 2 mm of the test block. Determination was carried out using the RCM test device after 28 d of standard curing, as shown in Eq. (2).

$$D_{RCM} = \frac{1}{3}\sum\limits_{{{\text{i}} = 3}}^{3} {\frac{0.0239 \times (273\; + \;T) \cdot L}{{(U - 2) \cdot t}}} \left( {E_{d} - 0.0238\sqrt {\frac{{(273\; + \;T) \cdot L \cdot E_{d} }}{U - 2}} } \right)$$
(2)

DRCM is the non–steady state chloride migration coefficient of LTIRC; U is the voltage used (V); T is the average temperature of the anode solution test (℃); L is the thickness of LTIRC (mm); Ed is the average value of chloride penetration depth (mm); t is the duration of the test (h).

2.3.5 Frost resistance test

Tests were conducted on specimens with dimensions of 100 mm × 100 mm × 400 mm to evaluate the mass loss and relative dynamic elastic modulus of LTIRC after multiple freeze–thaw cycles. It was noted that the specimens should be saturated with water absorption before the freeze–thaw cycles, and the initial mass and the initial value of transverse fundamental frequency should be determined, as shown in Eqs. (3) and (4).

$$\Delta m_{n} = \frac{1}{3}\sum\limits_{{{\text{i}} = 3}}^{3} {\frac{{m_{0} - m_{{\text{n}}} }}{{m_{0} }} \times 100}$$
(3)

∆mn is the mass loss rate (%) after N cycles; m0 is the initial mass of LTIRC (g); mn is the mass of LTIRC after N cycles (g)

$${\text{F}}_{n} = \frac{1}{3}\sum\limits_{{{\text{i}} = 3}}^{3} {\frac{{f_{{\text{n}}}^{2} }}{{f_{{0}}^{2} }}} \times 100$$
(4)

Fn is the relative dynamic elastic modulus (%) of LTIRC after N freeze–thaw cycles. f0 is the initial value of the transverse fundamental frequency (HZ) of LTIRC; fn is the transverse fundamental frequency (HZ) of LTIRC after N cycles.

3 Results and discussion

3.1 Dry density

Because of the low bulk density of EP and WA, EP is a glassy structure and WA is a porous structure. Therefore, with the increase of EP and WA dosage under the same water–to–cement ratio, the dry density decreases linearly, of which the dry density decreased by 11.49% on average when the EP dosage increased from 20 to 50%, and the dry density decreased by 17.38% on average when the WA dosage increases from 20 to 40%, as shown in Fig. 4a, b.

Fig. 4
figure 4

a Dry density and b Standard curing of LTIRC

Full size image

3.2 Mechanical properties

With the increase of WA, the compressive strength decreases. However, the rate of decrease gradually slows down with the increase of WA. The compressive strength of WA decreased by 15.3% from 20 to 30%, and by 8.77% from 30 to 40%, as shown in Fig. 5. It was due to the high-water absorption rate of both WA and EP, leading to partial absorption of water involved in hydration. Meanwhile, higher WA dosage increased internal soluble sugar content, causing some retardation of SAC and affecting cement hydration. Moreover, WA’s high toughness reduced its compression damage. Secondly, the high-water absorption and retention of WA and EP promoted slurry hydration around them, reducing shrinkage. This created tighter bonding between WA/EP and slurry, positively impacting compressive performance, as shown Fig. 6. However, the bond between WA and slurry interface wasn’t strong. The AFt and Ca(OH)2 crystals from hydration didn’t fully fill the analysis area, which was the focus of compressive damage, as shown Fig. 6.

Fig. 5
figure 5

a EP content of 20% compressive strength, b EP content of 30% compressive strength, and c EP content of 50% compressive strength

Full size image
Fig. 6
figure 6

a SEM images of the interface between EP and paste, b SEM at the EP–paste interface, c SEM at the EP–paste interface, d Overall microscopic morphology of EP, e SEM at the WA–paste interface, and f SEM at the WA–paste interface

Full size image

As LTIRC was plastic concrete (slump 50–60 mm), it led to crushing after large EP extrusion, causing the slurry to further bond with EP and reduce porosity, positively affecting compressive properties, as shown Fig. 6c, d. Moreover, EP active oxide content was about 40%, indicating some volcanic ash activity. Volcanic ash reactions produce C–S–H phases, improving EP-slurry interfaces and densifying bonding. C–S–H phases also partially filled crushed EP cavities, optimizing mechanical properties.

3.3 Thermal conductivity

According to the relevant literature, buildings account for 40% of global energy consumption [34]. Whereas the incorporation of WA and EP reduced the thermal conductivity by a minimum of 62.1% and a maximum of 76.5% as compared to conventional concrete, as shown Fig. 7b. It was due to the porous structure of WA and the glassy structure of EP, which increased the internal porosity with the increase of WA and EP admixture, while the gas was the least thermally conductive among the three phases of solids, liquids, and gases, which decreased the thermal conductivity of LTIRC. WA has a low conductor property, so it can be seen from Fig. 7 that WA doping and thermal conductivity were negatively correlated. However, it can be seen from Fig. 6c that EP was crushed by cement paste, changing its original morphology and decreasing LTIRC’s heat flow resistance. Therefore, using LTIRC in residential and commercial building envelopes can effectively reduce heat transfer, improve energy efficiency, and lessen HVAC (Heating, Ventilation, and Air Conditioning) system burden [35].

Fig. 7
figure 7

a Chloride diffusion coefficient of LTIRC, b Thermal conductivity of LTIRC, and c Thermal conductivity of different building materials

Full size image

Figure 7c shows that the thermal conductivity of lightweight concrete generally exceeded 0.6 W/(m·K) [35, 36], while biomass aggregates mostly ranged from 0.4 to 0.75 W/(m·K) [33, 37]. In contrast, foam concrete was significantly lower, usually 0.15 to 0.2 [38, 39]. Combined use of WF and EP effectively improved thermal insulation of LTIRC. This composite application strategy provides an effective technical way to improve building energy efficiency and indoor thermal environment.

3.4 Resistance to chloride ion permeability

Figure 7a shows that the non-steady state chloride migration of LTIRC increased with increasing EP. However, as WA and EP increased, the water used for hydration inside LTIRC decreased, reducing capillary pores. This reduced the rate of increase of the non-steady state chloride migration coefficient. Secondly, the non-steady state chloride migration coefficient of LTIRC decreased with increasing WA dosage. Specifically, it decreased by 6.26% on average when WA increased from 20 to 30%, and 1.79% on average when WA increased from 30 to 40%. The details are shown in Fig. 5f. This was due to WA inhibiting thermal crack generation from temperature differences between the slurry’s inside and outside, and reducing internal moisture, thereby decreasing pore structure within LTIRC. However, high WA dosing made WA more sensitive to chloride ion adsorption and penetration, exacerbating chloride erosion [40]. Therefore, WA doping made LTIRC more permeation-resistant, but this enhancement weakened at over 30% dosing. While the chloride ion permeability of LTIRC is inferior to conventional concrete [41], it still meets construction standards requirements. Improvements in other properties represent the main advantage of LTIRC.

3.5 Frost resistance

In the process of freezing and thawing, some of the free water in the bubble pores inside the LTIRC diffused to the capillary pores and then froze and generated hydrostatic pressure, causing the slurry to expand and crack. After the introduction of EP, the thin wall of EP and the semi-open pores generated by extrusion damage increased the bubble pores inside LTIRC, resulting in more significant expansion and cracking of LTIRC. In addition, due to the increasing number of freeze–thaw cycles and expansion within the EP, the thin-walled structure of the EP was further damaged internally; LTIRC internal damage intensified; and the relative dynamic modulus decreased significantly. However, the exterior was more intact and the mass loss rate was small, as shown in Figs. 8 and 9. Moreover, LTIRC exhibits a lower mass loss rate compared to previous studies [42, 43]. However, after freeze–thaw cycles, EP and WA can result in a more significant loss in relative dynamic elastic modulus. This suggests that special attention should be paid to long-term environmental effects on material properties during LTIRC design and application. After the introduction of WA, the internal moisture of WA promoted the hydration of the cement around WA, which made the surrounding slurry combine more closely with WA, and reduced the expansion cracking of the pores in the cement due to freeze–thaw cycles under the high toughness characteristics of WA. However, WA and EP have potential application in concrete as recycled materials and substitutes for natural minerals, but they may contain some hazardous substances. These substances can cause corrosion and damage to concrete, reduce its ability to resist expansion stresses, increase expansion cracking, and ultimately decrease its relative dynamic modulus of elasticity after freeze–thaw cycles. The effect of biomass aggregate and solid waste aggregate on frost resistance is a complex issue, and the practical application process affects concrete in many ways. As a result, it is necessary to consider the type of biomass aggregate, the handling method, and the concrete mix ratio [42, 44,45,46,47].

Fig. 8
figure 8

a The mass loss of WA content of 20%, b The mass loss of WA content of 25%, c The mass loss of WA content of 30%, d The mass loss of WA content of 35%, and e The mass loss of WA content of 40%

Full size image
Fig. 9
figure 9

a The relative dynamic elastic modulus after freeze–thaw cycles with WA content of 20%, b The relative dynamic elastic modulus after freeze–thaw cycles with WA content of 25%, c The relative dynamic elastic modulus after freeze–thaw cycles with WA content of 30%, d The relative dynamic elastic modulus after freeze–thaw cycles with WA content of 35%, and e The relative dynamic elastic modulus after freeze–thaw cycles with WA content of 40%

Full size image

4 Conclusions

In this study, lightweight aggregate insulated recycled concrete (LTIRC) was prepared using waste wood and expanded perlite as aggregates, reducing natural aggregate usage and concrete weight and costs. The paper investigates LTIRC’s dry density, thermal properties, mechanical properties, chloride ion permeability resistance, and frost resistance.

  1. (1)

    The unique characteristics of WA and EP, including their low density and porous nature, significantly impact the dry density of LTIRC. Consequently, with the incremental use of WA and EP, the dry density of LTIRC exhibits a gradual decrease. Specifically, the dry density was discerned to be 1402.6 kg/m3 when the dosage of WA was 40% and EP was 5%; however, the dry density increased to 1891.3 kg/m3 when the dosage of WA was lowered to 20% and EP was reduced to 2%. This observation suggests that the dry density is inversely proportional to the amount of WA and EP used. Therefore, LTIRC has met the dry density criteria for lightweight concrete.

  2. (2)

    Both WA and EP dosing increases caused the mechanical properties of LTIRC to deteriorate. However, as the addition of these two lightweight aggregates reduced the water within the LTIRC, the flow decreased, and the rate of mechanical property deterioration decreased.

  3. (3)

    The introduction of WA and EP can significantly reduce the thermal conductivity of LTIRC. The thermal conductivity of LTIRC was reduced by a maximum of 76.5% compared to that of concrete prepared with natural aggregates, which may be a positive contribution to the thermal insulation performance of future buildings.

  4. (4)

    Compared to natural aggregates, waste aggregates’ toughness and high-water absorption positively inhibits thermal and expansion cracking in concrete. Thus, increasing WA dosage retarded chloride ion diffusion and expansion cracking from freeze–thaw cycles. However, waste and natural aggregates’ high chloride or harmful substance content can corrode and damage concrete, and EP’s thin walls poorly resist expansion stresses in lightweight thermal insulation concrete.