Structures 61 (2024) 105973

Contents lists available at ScienceDirect

Structures
journal homepage: www.elsevier.com/locate/structures

Study on the sulfate freeze–thaw resistance of recycled coarse aggregate

self-compacting concrete
Chuanlei Zheng a, Luoyin Li b, Zhouhong Zong a, *
a
Engineering Research Center of Safety and Protection of Explosion & Impact of Ministry of Education, Southeast University, Nanjing 211189, China
b
School of Civil and Hydraulic Engineering, Ningxia University, Yinchuan 750021, China

A R T I C L E I N F O A B S T R A C T

Keywords: Concrete buildings in outdoor environments in northern China will suffer from sulfate freeze-thaw effects year-
Recycled coarse aggregate self-compacting round, which will seriously threaten the long-term safe use and service life of concrete. The construction industry
concrete is considered to be one of the key sectors for achieving sustainable development goals, and this view is widely
Sulfate freeze3thaw
recognized in industry and academia. In this paper, the sulfate freeze–thaw resistance values of recycled coarse
Mechanical properties
Stress3strain curve
aggregate self-compact concrete (RCASCC) under the influences of three recycling coarse aggregate (RCA)
Constitutive model replacement rates, four freeze–thaw environments, and six freeze–thaw cycles were studied. It was concluded
that the larger the RCA substitution rate was, the more severe the surface degradation and failure morphology of
RCASCC. Among the four freeze-thaw environments, the surface deterioration of RCASCC was the most serious in
the 5% MgSO4 solution freeze-thaw environment, and the surface deterioration and failure morphology of
RCASCC were the slightest in the clear water freeze-thaw environment. In addition, the compressive and splitting
tensile strength loss rate of RCASCC decreases the most in a 5% mass fraction MgSO4 solution freeze-thaw
environment. Among the six groups of RCASCC, RCASCC with RCA substitution rate of 0 had the largest rela­
tive peak stress and the smallest relative peak strain in the clear water freeze-thaw environment. RCASCC with
RCA substitution rate of 50% had the smallest relative peak stress and the largest relative peak strain in the 5%
mass fraction MgSO4 solution freeze-thaw environment. The uniaxial compressive stress–strain constitutive
model was used to fit the stress–strain curves of RCASCC before and after sulfate freeze–thaw cycling. The
elastic modulus and Poisson’s ratio values of RCASCC in the six groups decreased with increasing number of
freeze–thaw cycles. The larger the substitution rate of RCA was, the smaller the elastic modulus and Poisson’s
ratio of RCASCC. The elastic modulus and Poisson’s ratio of RCASCC in clear water freeze-thaw environment are
the largest, and the elastic modulus and Poisson’s ratio of RCASCC in 5% MgSO4 solution freeze-thaw envi­
ronment are the smallest.

1. Introduction concrete with a self-compacting concrete function, which is made by

using RCA obtained by crushing, screening, and grading waste concrete
To date, many scholars have performed much research on making to replace stone (abbreviation: NCA) mixing [9]. In recent years, many
recycled concrete from recycled aggregates [1–4]. Recycled concrete has scholars have studied RCASCC. Overall, replacing the NCA with the RCA
been used in engineering [5–8]. In previous studies, the research topic of reduces the working performance, mechanical properties, and durability
combining recycled concrete and self-compacting concrete has attracted of the RCASCC [15–17]. This reduction occurs because RCA has higher
the attention of many scholars [9–11]. Self-compacting concrete can fill water absorption, a larger fragmentation index, and a more irregular
all corners of the formwork without vibration in the process of pouring. grain shape than NCA [18,19]. The performance drawbacks of RCA limit
This material is a kind of high-performance concrete with the advan­ the engineering applications of RCASCC in load-bearing structures.
tages of compactness, stability, and homogeneity [12]. Self-compact Sulfate attack and freeze–thaw cycles are two adverse environ­
concrete has been widely used in structures with dense reinforcement mental problems that hydraulic concrete structures often encounter, and
and complex shapes [13,14]. RCASCC is a high-performance green they are important factors impacting concrete aging [20,21]. Compared

* Correspondence to: Zhouhong Zong,.

E-mail address: [email protected] (Z. Zong).

https://doi.org/10.1016/j.istruc.2024.105973
Received 12 November 2023; Received in revised form 18 January 2024; Accepted 25 January 2024
Available online 1 February 2024
2352-0124/© 2024 Institution of Structural Engineers. Published by Elsevier Ltd. All rights reserved.
C. Zheng et al. Structures 61 (2024) 105973

to durability failure under the influence of a single factor, the damage is urban tap water. Water reduction agent is a powder polycarboxylate
caused by the coupling effect of sulfate erosion and freeze–thaw cycles high performance water reducing agent. The appearance of the water
(abbreviation: sulfate freeze–thaw) on concrete is more severe, and the reduction is light gray, the water reduction rate is about 32%, the
failure mechanism is more complex. When sulfate freeze–thaw cycling density is 500–600 g/L, the solid content is 98 ± 1%, the pH is 6–8, the
is applied to concrete, the degree of destruction of concrete is not a chloride ion content is less than 0.01%, and the sulfate ion content is
simple superposition but a mutual promotion or inhibition of destructive about 2.6%. The NCA comes from the Yinchuan Lanshan Sandstone
effects. Therefore, the study of the impacts of sulfate freeze–thaw cycles Factory. There are two types of NCA, namely, large stones and small
on concrete cannot simply rely on experience to determine the damage stones. The particle sizes of large stones are 10–20 mm, and the particle
effect, and conclusions need to be drawn through experimental research. sizes of small stones are 5–10 mm. When mixing concrete, the large and
Due to the full recognition of the weakening effect of sulfate freeze small stones are mixed in a mass ratio of 7:3. The RCA needed for the test
–thaw on the durability of concrete, many scholars have conducted is made of scrap pavement concrete. The appearance of sodium sulfate
some research on the sulfate freeze–thaw resistance of concrete and (abbreviation: Na2SO4) powder is a white crystalline powder with
have achieved some research results [22–24]. Jia studied the effect of moisture absorption that is soluble in water and insoluble in ethanol.
sulfate freeze–thaw on the durability of high-performance recycled The magnesium sulfate (abbreviation: MgSO4) powder is a colourless
concrete [22]. The results show that the high porosity of RCA is the main crystal or white crystalline powder that is salty and bitter, soluble in
factor that makes recycled concrete increasingly susceptible to freeze water and glycerin, slightly soluble in ethanol, and insoluble in acetone.
–thaw damage and sulfate attack. Niu studied the erosion test of sulfate The performance parameters of NCA and RCA are shown in Table 1. The
freeze–thaw cycling on concrete [23]. The results show that the erosion performance parameters of the Na2SO4 and MgSO4 powders are given in
products of concrete under the action of sulfate freeze–thaw cycles are Table 2. The characteristics of NCA, RCA, Na2SO4 powder, and MgSO4
mainly ettringite and gypsum. The formation of substances such as powder are shown in Fig. 1. The particle size distribution of cement,
ettringite and gypsum will consume a large amount of calcium hy­ NCA, RCA, big stone, small stone and sand is shown in Fig. 2.
droxide in the hydration products of concrete; this process reduces the
pH value of concrete and causes the decomposition of the concrete hy­
dration products [24]. 2.2. Experimental design
RCASCC has limited structural applications worldwide. The poor
durability of RCASCC affects the long-term safe use and service life of Table 3 shows the RCASCC mix proportions. Four kinds of freeze
RCASCC in engineering applications of bearing structures. Many re­ –thaw environments are designed in the experiment: water, 5% Na2SO4
searchers have carried out research on the durability damage of RCASCC salt solution, 5% MgSO4 salt solution, 5% Na2SO4, and 5% MgSO4 mixed
under the influence of a single factor [21,25–26]. Tuyan’s results show salt solution. For the convenience of the subsequent writing of the paper,
that when the RCA substitution rate of the RCASCC is 100%, the the letters H, N, M, and NM are used to represent these four freeze–thaw
RCASCC suffers severe damage and cracks after 300 freeze–thaw cycles environments. Six freeze–thaw cycles (0, 25, 50, 75, 100, and 125 cy­
[25]. However, the results of Huda suggest that the use of RCA instead of cles) are designed. Three RCA substitution rates (0, 50%, and 100%) are
NCA has a certain enhancement effect on the pore properties of RCASCC, designed.
which may enhance its frost resistance [21]. Modani’s results show a
good linear correlation between carbonation depth and compressive 2.2.1. Freeze–thaw cycle test
strength [26]. After curing the specimen, the specimen is immersed in water at 20
To date, achievements have been made concerning the durability ± 2 ◦ C for 4 days. To ensure that the specimen is always in a state of
damage of RCASCC under the influence of a single factor, but there are water saturation during freeze–thaw, the water surface of the soaking
few studies on the durability of RCASCC under the coupling of multiple specimen should be higher than that of the specimen 20–30 mm [27].
factors. Therefore, we focused on the study of the RCASCC sulfate After immersion is completed, the specimen is placed in a rubber bucket.
freeze–thaw resistance, while the number of freeze–thaw cycles, RCA Clean water, a 5% mass fraction Na2SO4 solution, a 5% mass fraction
substitution rate, and freeze–thaw environment were considered MgSO4 solution, or a mixed salt solution of 5% mass fraction Na2SO4 and
together. A series of damage indicators were tested and calculated, such 5% mass fraction MgSO4 are added to the sleeve barrel. Finally, the
as appearance, failure morphology, mass loss rate, dynamic elastic sleeve barrel is placed into the freeze–thaw machine, and the machine
modulus damage value, and ultrasonic wave velocity damage charac­ is started on the freeze–thaw cycle test. The test method of the
teristics. RCASCC is a new type of green concrete with many advantages freeze–thaw cycle test is the fast freezing method, and the test instru­
of self-compacting concrete prepared by replacing NCA with RCA made ment is the HDK-3 concrete fast freeze–thaw machine. Fig. 3 is a
of waste concrete, which is in line with the national policy of sustainable freeze–thaw machine working diagram.
development. China has a vast territory, and concrete buildings in cold
regions often suffer from sulfate freeze-thaw erosion. Studying the 2.2.2. Quality, dynamic elastic modulus, and ultrasonic wave velocity test
performance changes of RCASCC after sulfate freeze-thaw has important The mass of the specimen is measured by an electronic scale with a
theoretical guidance and exploration significance for the engineering sensitivity of 1 g and a range of 30 kg. A DT-20 dynamic elastic modulus
practice of RCASCC in cold regions. tester is used to measure the dynamic elastic modulus of the specimen.
Fig. 4 shows a schematic diagram of dynamic elastic modulus detection.
2. Test plan There are a total of 4 groups of measuring points. The final test results
take the average of the four groups of measuring points. The ultrasonic
2.1. Test raw materials wave velocity of the specimen is measured by a DJUS-05 nonmetallic
ultrasonic instrument. Fig. 5 shows a schematic diagram of ultrasonic
The cement used is ordinary Portland cement (P⋅O 42.5) produced by wave velocity detection. In total, there are 5 sets of measurement points.
a local cement plant. The density of cement is 3.15 g/cm3, the initial The final test results take the average value of 5 groups of measuring
setting time and final setting time are 142 min and 198 min, respec­ points.
tively, and the compressive strength of 3d and 28d are 28.2 MPa and Formula (1) is used to calculate the mass damage rate of RCASCC [9].
48.7 MPa respectively. The sand used is the washed medium sand of the G0 − GN
Lanshan gravel field in Yinchuan city. The apparent density, bulk den­ ΔWN = × 100% (1)
G0
sity, porosity, mud content, and fineness modulus of sand are 2718
kg⋅m− 3, 1630 kg⋅m− 3, 40%, 1.2%, and 2.94, respectively. The test water where ΔWN is the mass loss rate, in %; G0 and GN are the masses of

2
C. Zheng et al. Structures 61 (2024) 105973

Table 1
Performance parameters for NCA and RCA.
No Particle size range Apparent density Bulk density Void ratio/% Water absorption/% Moisture content/% Crush indicator/%
/mm / (kg⋅m− 3) / (kg⋅m− 3)

NCA 5-20 2708 1496 44.76 0.74 0.37 8.8

RCA 5-20 2632 1322 49.77 3.85 2 10

Table 2
Performance parameters of Na2SO4 and MgSO4 powders.
No Relative molecular mass Content/% pH Insolubility/% Cl/% PO4/%

Na2SO4 powders 142.04 ≥99.0 5-8 ≤0.005 ≤0.001 ≤0.001

MgSO4 powders 120.37 ≥99.0 5-8 ≤0.005 ≤0.002 ≤0.005

Fig. 1. Material appearance: (a) NCA, (b) RCA, (c) Na2SO4 powders, and (d) MgSO4 powders.

value of RCASCC [28].

100
cement V 2dN
NCA/RCA VN = 1 − (3)
V 2d0
80 big stone
small stone where VN is the ultrasonic wave velocity damage value and Vd0 and VdN
sand are the ultrasonic wave velocities of RCASCC, in km⋅s− 1.
Passing rate /%

60
2.2.3. Test for compression and split tensile strength
40 The compressive strength of the specimen is tested using an
HTSHT4106 microcomputer-controlled electrohydraulic servo universal
test machine. The maximum load of the test machine is 1000 kN with
20 0.001 N accuracy. The loading method of the test machine is stress
loading, and the loading speed is 0.5 MPa⋅s− 1. Fig. 6(a) shows the actual
operation diagram of the compressive strength test. The splitting tensile
0 (Log)
strength of the test specimen is tested by the CMT5305 microcomputer-
1E-4 0.001 0.01 0.1 1 10 100
Sieve aperture /mm controlled electrohydraulic servo universal test machine. The maximum
load of the test machine is 300 kN, and the loading speed is
Fig. 2. The particle size distribution of cement, NCA, RCA, big stone, small 0.05 MPa⋅s− 1. The actual operation of the splitting tensile strength test is
stone and sand. shown in Fig. 6(b).

RCASCC, in kg; G0 and GN are the masses of RCASCC after 0 and N 2.2.4. Testing of the stress–strain curve, elastic modulus, and Poisson’s
freeze–thaw cycles, in kg; and 0 and N are the number of freeze–thaw ratio
cycles. The stress–strain curve, elastic modulus, and Poisson’s ratio of the
Formula (2) is used to calculate the dynamic elastic modulus damage specimens are measured by an SHT4106 microcomputer-controlled
value of RCASCC [27]. electrohydraulic servo universal testing machine. The loading pattern
EN of the test machine is stress loading. The loading rate of the testing
DN = 1 − (2) machine is set to 0.5 MPa⋅s− 1 when measuring the stress–strain data,
E0
and the loading rate of the testing machine is set to 0.05 MPa⋅s− 1 when
where DN is the dynamic elastic modulus damage value and E0 and EN measuring the elastic modulus and Poisson’s ratio. Strain data are
are the dynamic elastic moduli of RCASCC, in MPa. collected using the JM3813 multifunctional static strain gauge. The
Formula (3) is used to calculate the ultrasonic wave velocity damage resolution of the strain gauge is 0.5 με, the range is ± 15000 με, and the

Table 3
RCASCC mix ratio.
No W/C Design strength RCA replacement rate/% Amount of material/ (kg⋅m− 3)

Water Cement NCA RCA Sand Water reduction

NC 0.41 C40 0 218 531 812 0 818 0.372

RC50 0.41 C40 50 218 531 406 406 818 0.531
RC100 0.41 C40 100 218 531 0 812 818 0.637

3
C. Zheng et al. Structures 61 (2024) 105973

Fig. 3. Freeze–thaw machine working diagram. (a) Front view of the testing machine. (b) Temperature curves of the specimen and the thermally conduc­
tive medium.

Fig. 4. Dynamic elastic modulus test. (a) Test drawing. (b) Measuring point layout.

Fig. 5. Ultrasonic wave velocity test. (a) Test drawing. (b) Measuring point layout.

accuracy is ± 0.2% FS ± 1 με. A 5G101 displacement sensor is used for 3. Experimental phenomena
the displacement sensor. The sensitivity of the displacement sensor is 90
με/mm, the range is 60 mm, and the accuracy is 0.3% FS. The lengths of 3.1. Surface degradation
the strain gauges are 100 mm and 50 mm, which are placed in a +
shape. The resistance of the strain gauge is 120 Ω, and the sensitivity By observing the changes in the apparent characteristics, we can
coefficient is 2.08%. The stress–strain curve, elastic modulus, and intuitively understand the damage to RCASCC. Fig. 8 shows the surface
Poisson’s ratio test results are shown in Fig. 7. degradation of RCASCC before and after the freeze–thawing of sulfate.
The size of the RCASCC shown in Fig. 8 is
100 mm × 100 mm × 300 mm. Fig. 8 shows that the damage caused to
the RCASCC by the freeze–thaw cycling of sulfate is a gradual process.
With the increase in the number of freeze–thaw cycles, the surface
condition of the RCASCC is gradually deteriorated. At the 0th

4
 C. Zheng et al. Structures 61 (2024) 105973

gradually decreases, and the diagonal crack forms a continuous long

crack under the action of normal stress and shear stress. Final RCASCC
instability failure. With a cracking sound, RCASCC is split into several
irregularly shaped concrete blocks. The failure morphology of the NC-H-
0 group RCASCC exhibits oblique shear failure characteristics.
The damage caused by the freeze–thaw action of sulfate on concrete
is similar to that caused by the freeze–thaw action but has unique
features [30,31]. As the number of freeze–thaw cycles increases, sur­
face cracks gradually increase after RCASCC failure, plastic deformation
becomes increasingly severe, and the failure mode gradually changes
from shear failure to internal splitting and tensile failure in RCASCC. In
the H-type freeze–thaw environment with the same number of freeze
Fig. 6. Compressive and split tensile strength tests. (a) Schematic of the actual –thaw cycles, the larger the RCA replacement rate is, the larger the
operation of the compressive strength test. (b) Schematic of the actual opera­ number of surface cracks after RCASCC disruption and the more severe
tion of the split tensile strength test. the damage morphology. Compared with NCA, RCA has a lower
apparent density, higher packing density, higher porosity, higher water
freeze–thaw cycle, the surface of the RCASCC is smooth and flat content, and higher crushing index. The physical properties of RCA are
without damage. After 50 cycles, slight mortar shedding and pock­ weaker than those of NCA, and mixing concrete with RCA instead of
marking appears on the surface of RCASCC, and the distribution is NCA can weaken the durability of concrete to some extent. When the
sparse. After 125 cycles, the surface of the RCASCC becomes uneven, replacement rate of RCA is 50% and the number of freeze–thaw cycles
and some edges and corners are damaged. In the H-type freeze–thaw is the same, the damage forms of RCASCC are the lightest in the
environment and the same number of freeze–thaw cycles, the higher freeze–thaw environment of the H class, and the difference between the
the RCA substitution rate is, the more serious the surface deterioration of damage forms of RCASCC in the freeze–thaw environments of the N, M
RCASCC. When the RCA replacement rate is 50% and the number of and NM classes are small. Sulfate can cause concrete corrosion, which
freeze–thaw cycles is the same, the surface degradation degree of can aggravate damage to concrete. Second, the hygroscopicity and water
RCASCC is the most serious in the class M freeze–thaw environment, saturation of sulfate increase the saturation, frozen water volume, and
the damage of RCASCC is the least in the class H freeze–thaw envi­ frost-heaving force of concrete, thus accelerating the damage of concrete
ronment, and the apparent characteristics of RCASCC in class N and NM so that the damage degrees of RCASCC in N, M, and NM freeze–thaw
freeze–thaw environments are slightly different. environments are greater than that of H. In addition, sulfate can reduce
the freezing point of pore solutions, thereby reducing the damage caused
by sulfate freeze–thaw on RCASCC to a certain extent.
3.2. Failure morphology
4. Analysis of the experimental results
A stress–strain curve test on RCASCC is conducted. Fig. 9 shows the
failure morphologies of RCASCC before and after sulfate freeze–thaw 4.1. Mass loss rate
cycling. The size of the RCASCC shown in Fig. 9 is 100 mm × 100 mm
× 300 mm. As an example, we describe the loading failure process of The calculated mass loss rate is used to draw the relationship curve
RCASCC for the NC-H-0 group. Observations show that there is no between the mass loss rate and the number of freeze–thaw cycles, as
visible change on the surface of RCASCC during initial loading, and the shown in Fig. 10. Fig. 10 shows that as the number of freeze–thaw
load increases approximately proportionally over time. With increasing cycles increases, the RCASCC mass loss first decreases and then in­
load, small oblique cracks begin to appear in the middle and corner of creases. After 125 freeze–thaw cycles, the RCASCC has a mass loss rate
the RCASCC and gradually expand into cracks. At this time, the curve of of less than 5%. During 0–50 freeze–thaw cycles, the mass loss rates of
the relationship between load and time tends to be flat. As the load RC50 and R100 are negative. The surface mortar shedding and internal
further increases, the cracks begin to widen and increase in number, and pore erosion of the RCASCC are relatively minor from 0 to 50 freeze
the cracks develop along the longitudinal and oblique directions. When –thaw cycles, at which point the large amount of water absorption by
the load increases to 0.8–0.9 times the peak load, the volume defor­ the RCASCC makes the mass loss rate negative. With the increase in the
mation of RCASCC reaches the maximum value, and the cracks gradu­ number of freeze–thaw cycles, the rate of mass loss increases gradually,
ally increase. After the load exceeds the peak load, the volume and the rate of increase accelerates gradually. Fig. 10(a) shows that
deformation of RCASCC gradually recovers, the bearing capacity

(b)
Test machine
(c)
F /N

roof
Centering Preload
Upper pressure Destroy
plate
Steel pillar 90 s 60 s 60 s 90 s
Fixture
Displacement
Fa

meter
Strain gauges

Lower pressure
plate

Test machine 90 s 60 s 60 s
floor 90 s
F0

t /s

Fig. 7. Stress–strain curves, elastic modulus, and Poisson’s ratio detection. (a) Test drawing. (b) Test schematic. (c) Elastic modulus and Poisson’s ratio test load
change curve with time.

5
­
C. Zheng et al. Structures 61 (2024) 105973

Fig. 8. Apparent features of RCASCC before and after freeze–thawing of sulfate: (a) 0 times, (b) 50 times, and (c) 125 times.

when the RCA replacement rate is high, the quality loss rate of RCASCC indicating high fitting accuracy. In Fig. 11, all the data points are
is high, and the quality loss rate grows relatively fast. Fig. 10(b) shows distributed around the fitted curve, indicating that the fitted results are
that different freeze–thaw environments have relatively small impacts well correlated and have some reference value.
on the mass loss rate of RCASCC. After 125 freeze–thaw cycles, the mass ⎧ ( ) 2
loss rates of NC-H, RC50-H, RC100-H, RC50-N, RC50-M, and RC50-NM ⎪ NC − H : ln 0.9789
⎪
⎪ ( + 0.0029N , R ) = 0.9649
⎪
⎪ RC50 − H : ln 0.9906 + 0.0034N ,)R2 = 0.9698
are 0.34%, 0.36%, 0.38%, 0.41% and 0.39%, respectively. ⎪
⎨ ( 2
RC100 − H : ln ( 0.9675 + 0.0049N ) 2 , R = 0.9671
DN = (4)
⎪
⎪ RC50 − N : ln ( 0.98 + 0.0038N , ) 2 0.9104
R =
⎪
⎪
4.2. Dynamic elastic modulus damage ⎪
⎪
⎩
RC50 − M : ln ( 0.9833 + 0.0042N ) , R = 0.9487
RC50 − M : ln 0.9858 + 0.0038N , R2 = 0.9716
The calculated dynamic elastic modulus damage values are used to
plot the damage to the brake elastic modulus versus the number of 4.3. Ultrasonic wave velocity damage
freeze–thaw cycles, as shown in Fig. 11. Fig. 11 shows that the damage
of the dynamic elastic modulus of RCASCC gradually increases with Fig. 12 shows a plot of the damage value of the ultrasound wave
increasing number of freeze–thaw cycles. The curve of the dynamic velocity versus the number of freeze–thaw cycles. Fig. 12 shows that
elastic modulus damage of RCASCC with the number of freeze–thaw the ultrasound velocity damage of the RCASCC increases with increasing
cycles can be divided into three stages: a steady growth stage, a linear number of freeze–thaw cycles. The variation curve of the ultrasonic
growth stage, and a relatively stable stage. After 125 freeze–thaw cy­ wave velocity damage value of RCASCC with the number of freeze
cles, the dynamic elastic modulus damage values of NC-H, RC50-H, –thaw cycles can be divided into three stages: a relatively stable stage, a
RC100-H, RC50-N, RC50-M, and RC50-NM increase to 0.3025, 0.3207, linear growth stage, and a steady growth stage. After 125 freeze–thaw
0.4555, 0.3495, 0.3812 and 0.3606, respectively. After 125 freeze cycles, the ultrasonic wave velocity damage values of NC-H, RC50-H,
–thaw cycles, the damage value of the dynamic elastic modulus of RC100-H, RC50-N, RC50-M, and RC50-NM increase to 0.4587, 0.4649,
RC100-H is already greater than 0.4, indicating that RC100-H is 0.5061, 0.5241, 0.562, and 0.5321, respectively.
destroyed [29]. There is an exponential relationship between the damage value of
There is a logarithmic relationship between the damage value of the the ultrasonic wave velocity and the number of freeze–thaw cycles, as
dynamic elastic modulus and the number of freeze–thaw cycles, as shown in Formula (5). The fitting curve is shown in Fig. 12. Formula (5)
shown in Formula (4). The fitting curve is shown in Fig. 11. Formula (4) shows that the fitting accuracy of the obtained fitting curve is high, and
shows that the correlation coefficients R2 are all greater than 0.9104, the highest correlation coefficient is 0.9934. The data points are all

6
C. Zheng et al. Structures 61 (2024) 105973

Fig. 9. Failure morphologies of RCASCC before and after freeze–thawing of sulfate: (a) 0 times, (b) 50 times, and (c) 125 times.

0.40 0.40
NC-H (a) RC50-H (b)
0.35 RC50-H 0.35 RC50-N
RC100-H RC50-M
0.30 0.30
RC50-NM
Mass loss rate /%

Mass loss rate /%

0.25 0.25
0.20 0.20
0.15 0.15
0.10 0.10
0.05 0.05
0.00 0.00
-0.05 -0.05
0 25 50 75 100 125 0 25 50 75 100 125
The number of freeze-thaw cycles /times The number of freeze-thaw cycles /times

Fig. 10. Curves of the mass loss rate versus the number of freeze–thaw cycles. (a) Relationship curve between different RCA substitution rates and mass loss rates.
(b) Relationship curve between different freeze–thaw environments and mass loss rate.

7
C. Zheng et al. Structures 61 (2024) 105973

0.5 0.5
NC-H (a) RC50-H (b)

Dynamic elastic modulus damage value

Dynamic elastic modulus damage value

RC50-H RC50-N
0.4 RC100-H 0.4 RC50-M
NC-H fitting curve RC50-NM
RC50-H fitting curve RC50-H fitting curve
0.3 RC100-H fitting curve 0.3 RC50-N fitting curve
RC50-M fitting curve
RC50-NM fitting curve
0.2 0.2

0.1 0.1

0.0 0.0
0 25 50 75 100 125 0 25 50 75 100 125
The number of freeze-thaw cycles /times The number of freeze-thaw cycles /times

Fig. 11. Relationship curve between the damage value of dynamic elastic modulus and the number of freeze–thaw cycles. (a) Relationship curves of different RCA
substitution rates and dynamic elastic modulus damage values. (b) Relationship curves between different freeze–thaw environments and dynamic elastic modulus
damage values.

0.7 0.7
NC-H (a) RC50-H RC50-H fitting curve (b)
Ultrasonic wave velocity damage value

Ultrasonic wave velocity damage value

RC50-H RC50-N RC50-N fitting curve
0.6 0.6
RC100-H RC50-M RC50-M fitting curve
NC-H fitting curve RC50-NM RC50-NM fitting curve
0.5 0.5
RC50-H fitting curve
0.4 RC100-H fitting curve 0.4

0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
0 25 50 75 100 125 0 25 50 75 100 125
The number of freeze-thaw cycles /times The number of freeze-thaw cycles /times

Fig. 12. Relationship curve between the damage of ultrasonic wave velocity and the number of freeze–thaw cycles. (a) Relationship curves of damage values of
ultrasonic wave velocity and different RCA substitution rates. (b) Relationship curves between damage values of ultrasonic wave velocity and different freeze–thaw
environments.

54 48
(a) (b)
51 46
Compressive strength /MPa

Compressive strength /MPa

48 44

45 42

42 40 RC50-H
NC-H
RC50-H RC50-N
39 RC100-H 38 RC50-M
RC50-NM
36 36
0 25 50 75 100 125 0 25 50 75 100 125
The number of freeze-thaw cycles /times The number of freeze-thaw cycles /times

Fig. 13. Relationship curve between the compressive strength and the number of freeze–thaw cycles. (a) Relationship curves between different RCA substitution
rates and compressive strengths. (b) Relationship curves between different freeze–thaw environments and compressive strengths.

8
C. Zheng et al. Structures 61 (2024) 105973

distributed close to the fitted curve, indicating that the fit is good and During freeze–thaw, SO2-4 reacts with Ca(OH)2 in the hydration ma­
has some reference value. terial inside RCASCC and generates gypsum (CaSO4⋅2H2O), which re­
⎧ duces the content of hydration material and the bonding force between
⎪
⎪ NC − H : − 1.4487 × e− N/318.806 + 1.437, R2 = 0.9765 aggregate and cement slurry, resulting in RCASCC collapse and spalling,
⎪
⎪ RC50 − H : − 0.6509 × e− N/87.6941 + 0.628, R2 = 0.9307
affecting the overall structure. SO2-
⎪
4 can react with calcium sulfoalumi­
⎪
⎨
RC100 − H : − 0.8013 × e− N/121.9473 + 0.7885, R2 = 0.9613
VN = (5) nate (3CaO⋅Al2O3⋅12H2O) to produce ettringite
⎪
⎪ RC50 − N : − 0.7257 × e− N/94.2211 + 0.7102, R2 = 0.9498
⎪
⎪
⎪ RC50 − M : − 0.8337 × e − N/107.4736 2
+ 0.8255, R = 0.9934 (3CaO⋅Al2O3⋅3CaSO4⋅32H2O), Al(OH)3 and other water-absorbing and
⎪
⎩
RC50 − M : − 0.8469 × e− N/123.1474 + 0.836, R2 = 0.9689 expansive substances. Under the combined action of freeze–thaw and
sulfate attack, the expansion of expansive materials inside RCASCC, the
freezing of solution, the crystallization of salt and the decrease in
4.4. Compressive strength cohesive force between aggregate and cement slurry causes irreversible
residual expansion and deformation of cracks inside RCASCC, which
Fig. 13 shows the relationship curve between the compressive makes RCASCC increasingly fragile and shows a decrease in strength at
strength and the number of freeze–thaw cycles. As shown in Fig. 13, the the macroscopic level.
compressive strength of the RCASCC in each group decreases as the
number of freeze–thaw cycles increases. Fig. 13(a) shows the rela­ 4.5. Splitting tensile strength
tionship curve between different RCA substitution rates and compres­
sive strength. As shown in Fig. 13(a), at 0 freeze–thaw cycles, the The measured split tensile strength is used to plot the curve of the
difference in compressive strength between NC-H and RC50-H is rela­ split tensile strength versus the number of freeze–thaw cycles, as shown
tively large, and the difference in compressive strength between RC50-H in Fig. 14. As shown in Fig. 14, the splitting tensile strength of RCASCC
and RC100-H is relatively small. The strength and elastic modulus values in each group gradually decreases with increasing freeze–thaw cycle.
of RCA are weaker than those of NCA, making the compressive strength Fig. 14(a) shows that the curves of NC-H, RC50-H, and RC100-H exhibit
of RC50-H and RC100-H weaker than that of NC-H. When the ratio of a significant layering phenomenon. After 125 freeze–thaw cycles, the
RCA replacing NCA is high, the actual water–cement ratio of RCASCC splitting tensile strength loss rates of NC-H, RC50-H, and RC100-H are
increases due to the large amount of water absorption by RCA, resulting 15.27%, 19.05%, and 23.01%, respectively. Fig. 14(b) shows that the
in a small difference in compressive strength between RC50-H and splitting tensile strength of RC50-M in the four groups of RCASCC
RC100-H. Each group of curves exhibits a clear layering phenomenon. decrease the fastest, and the decreasing rates of the splitting tensile
After 125 freeze–thaw cycles, the compressive strength loss rates of NC- strengths of RC50-H, RC50-N, and RC50-NM are not much different.
H, RC50-H, and RC100-H are 10.63%, 15.26%, and 20.24%, respec­ After 125 freeze–thaw cycles, the splitting tensile strength loss rates of
tively. Fig. 13(b) shows the relationship curves between different RC50-H, RC50-N, RC50-M, and RC50-NM are 19.05%, 21.96%, 28.04%,
freeze–thaw environments and compressive strengths. Fig. 13(b) shows and 19.84%, respectively.
that the compressive strengths of RC50-H and RC50-N decrease rapidly In the process of long-term use and crushing of waste concrete, the
after 75 freeze–thaw cycles, while those of RC50-M and RC50-NM original cracks expand, and new cracks appear. The basic properties of
accelerate after 25 freeze–thaw cycles. The difference in the compres­ RCA made from waste concrete, such as strength, elastic modulus, and
sive strength of RC50 after undergoing 25 and 50 freeze–thaw cycles in water absorption, are weaker than those of NCA. The addition of RCA to
four different freeze–thaw environments is small, and the difference the concrete decreases the mechanical and durability properties. The
gradually increases after exceeding 50 cycles. After 125 freeze–thaw pore cracks in the old cement mortar on the RCA surface extend
cycles, the compressive strength loss rates of RC50-H, RC50-N, RC50-M, continuously under the action of sulfate freeze–thaw, thus adversely
and RC50-NM are 15.26%, 18.42%, 23.03%, and 19.29%, respectively. affecting the properties of RC50 and RC100 [32]. The higher the ratio of
During the freeze–thaw process, the free water volume in the pores RCA replacing NCA is, the greater the numbers of pores and cracks in the
of RCASCC expands and contracts, thereby reducing the structural sta­ old cement mortar inside RCASCC, the more serious the damage after
bility of RCASCC. As the number of freeze–thaw cycles increases, new sulfate freeze–thaw cycles, and the greater the loss rates of compressive
microcracks continue to emerge and develop, and damage continues to strength and splitting tensile strength.
accumulate. Compared to NC, the bonding performance levels between
the old mortar layers and the matrices inside RC50 and RC100 are poor.
Poor bonding properties make RC50 and RC100 less resistant to freezing
than NC in the face of freeze–thaw damage.

4.0 3.8
(a) (b)
3.8 3.6
Splitting tensile strength /MPa

Splitting tensile strength /MPa

3.6
3.4
3.4
3.2
3.2
3.0
3.0 RC50-H
NC-H 2.8 RC50-N
2.8 RC50-H RC50-M
RC100-H RC50-NM
2.6 2.6
0 25 50 75 100 125 0 25 50 75 100 125
The number of freeze-thaw cycles /times The number of freeze-thaw cycles /times

Fig. 14. Relationship curves between the splitting tensile strength and the number of freeze–thaw cycles. (a) Relationship between different RCA substitution rates
and splitting tensile strengths. (b) The relationship curve between different freeze–thaw environments and splitting tensile strength.

9
C. Zheng et al. Structures 61 (2024) 105973

Fig. 15. Stress–strain curves of RCASCC after freeze–thaw of sulfate. (a) NC-H. (b) RC50-H. (c) RC100-H. (d) RC50-N. (e) RC50-M. (f) RC50-NM.

45 45 45
(a) NC-H (b) NC-H (c) NC-H
40 RC50-H 40 RC50-H 40 RC50-H
35 /RC50-N 35 RC100-H 35 RC100-H
/RC50-M RC50-N RC50-N
30 /RC50-NM 30 RC50-M Stress /MPa 30 RC50-M
Stress /MPa

Stress /MPa

25 RC100-H 25 RC50-NM 25 RC50-NM

20 20 20
15 15 15
10 10 10
5 5 5
0 0 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
Strain /×10-3 Strain /×10-3 Strain /×10-3

Fig. 16. Stress–strain curves of RCASCC after different numbers of sulfate freeze–thaw cycles. (a) 0 times. (b) 50 times. (c) 125 times.

5. Testing of the stress–strain curve, elastic modulus, and

Poisson’s ratio 44 NC-H NC-H fitting curve
RC50-H RC50-H fitting curve
42 RC100-H RC100-H fitting curve
5.1. Stress–strain curve RC50-N RC50-N fitting curve
40 RC50-M RC50-M fitting curve
Peak stress /MPa

Using the measured stress and strain data, we plot the stress–strain RC50-NM RC50-NM fitting curve
curve of RCASCC after freeze–thawing of sulfate, as shown in Fig. 15. 38
As shown in Fig. 15, the shape of the upper half of the stress–strain
36
curve for RCASCC in each group is similar, while the lower half has some
differences. As the number of freeze–thaw cycles increases, the peak of
34
the curve moves to the lower right, and the shape of the curve gradually
changes from a high-narrow shape to a short-wide shape. This finding 32
indicates that RCASCC shows a gradual loosening trend.
The stress–strain curves of RCASCC after different numbers of sul­ 30
2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
fate freeze–thaw cycles are shown in Fig. 16. Fig. 16 shows that with Peak strain /×10-3
the same number of freeze–thaw cycles, as the RCA substitution rate
increases, the peak point of the curve gradually shifts to the right and Fig. 17. Variations in peak stress with peak strain.
downwards. The concavity levels of the lower halves of the stress–strain
curves in the N, M and NM freeze–thaw environments are more obvious The relationship between the peak stress and peak strain is a linear
than those in the H freeze–thaw environment. This phenomenon is function, and the fitting formula and fitting graph can be obtained by
mainly related to the relatively great damage of RCASCC in the sulfate linear fitting, as shown in Formula (6) and Fig. 17, respectively.
freeze–thaw environment.

10
 C. Zheng et al. Structures 61 (2024) 105973

⎧ ( )
⎪
⎪ NC − H : − 8.2827E
( + 4 N)+ 1.0048, R2 = 0.9403
⎪
⎪
⎪
⎪ RC50 − H : − 9.4333E + 4 N + 1.0048, R2 = 0.8761
σ 0N ⎨
RC100 − H : − 0.0014N + 1.0042, R2 = 0.9866
= (7)
σ 00 ⎪
⎪ RC50 − N : − 0.0013N + 0.9955, R2 = 0.89
⎪
⎪
⎪
⎪ RC50 − M : − 0.0017N + 0.9954, R2 = 0.9954
⎩
RC50 − NM : − 0.0012N + 0.9775, R2 = 0.9047

where σ0N
σ00 is the relative peak stress.

5.1.2. Relative peak strain

The relative peak strain is the ratio of the peak strain of RCASCC after
N freeze–thaw cycles to the peak strain at 0 freeze–thaw cycles. Fig. 19
shows the relationship between the relative peak strain and the number
of freeze–thaw cycles. Fig. 19 shows that the relative peak strain of each
group of RCASCC increases with the increase in the number of freeze
–thaw cycles. When the number of freeze–thaw cycles is the same, the
relative peak stress of RC50-M is the largest and that of NC-H is the
Fig. 18. Relationships between the relative peak stresses and freeze
–thaw cycles. smallest. Compared with the samples after 0 freeze–thaw cycles, the
relative peak strains of NC-H, RC50-H, RC100-H, RC50-N, RC50-M, and
⎧ ( ) RC50-NM increase by 45.77%, 44.71%, 47.2%, 47.12%, 54.33%, and
NC − H : − ( 4.7437ε0N + 50.0424 ×) 10− 3 , R2 = 0.926
⎪
⎪
⎪ 50.48%, respectively, after 125 freeze–thaw cycles. The dependence of
⎪ RC50 − H : − 5.5945ε0N + 51.0389 × 10− 3 , R2 = 0.9855
⎪
⎨ RC100 − H : ( − 6.9239ε + 52.6743) × 10− 3 , R2 = 0.9128
⎪ the relative peak strain on the number of freeze–thaw cycles is a linear
σ 0N = ( 0N
) (6) function. Formula (8) is obtained by fitting.
⎪
⎪ RC50 − N : ( − 6.7687ε0N + 52.9515) × 10− 3 , R2 = 0.8398 ⎧
⎪
⎪ − 3 2
⎪ NC − H : 0.0033N + 0.9597, R2 = 0.8155
⎩ RC50 − M : (− 8.025ε0N + 55.7896 ×) 10 ,−R3 =2 0.8971
⎪ ⎪
⎪
⎪
RC50 − NM : − 6.0658ε0N + 50.8768 × 10 , R = 0.8177 ⎪ RC50 − H : 0.0031N + 0.9757, R2 = 0.809
⎪
⎪
ε0N ⎨ RC100 − H : 0.0034N + 0.9709, R2 = 0.8481
= (8)
where σ0N and ε0N are the peak stress and peak strain of RCASCC after N ε00 ⎪ ⎪
⎪
RC50 − N : 0.0032N + 0.9828, R2 = 0.8294
⎪ 2
freeze–thaw cycles, respectively. ⎪
⎪
⎩
RC50 − M : 0.0038N + 1.003, R = 0.8945
RC50 − NM : 0.0035N + 0.9895, R2 = 0.8505
5.1.1. Relative peak stress
where ε0N
is the relative peak strain.
The relative peak stress is the ratio of the peak stress of RCASCC after ε00

N freeze–thaw cycles to the peak stress at 0 freeze–thaw cycles. Fig. 18

shows the relationship between the relative peak stress and freeze 5.1.3. Ultimate strain
–thaw cycles. From Fig. 18, with the increase in the number of The ultimate strain is taken as the longitudinal strain corresponding
freeze–thaw cycles, the relative peak stress of RCASCC in each group to 0.5σ 0N in the descending section of the stress–strain curve [33].
gradually decreases. When the number of freeze–thaw cycles is the Fig. 20 shows the relationship curve between the ultimate strain and the
same, the relative peak stress of NC-H is the highest, and the relative number of freeze–thaw cycles. From Fig. 20, with the increase in the
peak stress of RC50-M is the lowest. Compared with the samples after number of freeze–thaw cycles, the ultimate strain of RCASCC in each
0 freeze–thaw cycles, the relative peak stresses of NC-H, RC50-H, group generally shows a trend of first increasing, then decreasing and
RC100-H, RC50-N, RC50-M, and RC50-NM decrease by 10.67%, finally increasing. As shown in Fig. 21 (a), the curves of NC-H, RC50-H
13.14%, 18.11%, 16.32%, 21.85%, and 17.33%, respectively, after 125 and RC100-H show stratification. From Fig. 20(a), the ultimate strains of
freeze–thaw cycles. The relative peak stress shows a good linear NC-H, RC50-H, and RC100-H show stratification with the increase in the
dependence on the number of freeze–thaw cycles. The equation ob­ number of freeze–thaw cycles. Compared with the samples after
tained by fitting is shown in Formula (7). 0 freeze–thaw cycles, the ultimate strains of NC-H, RC50-H, and
RC100-H increase by 30.54%, 41.1%, and 38.13%, respectively, after
125 freeze–thaw cycles of RCASCC. Fig. 20(b) shows that the curves of
RC50-H, RC50-N, RC50-M, and RC50-NM are intertwined. Compared
with the samples after 0 freeze–thaw cycles, the ultimate strains of
1.6 RC50-H, RC50-N, RC50-M, and RC50-NM increase by 41.1%, 38.35%,
NC-H NC-H fitting curve
RC50-H RC50-H fitting curve 40.1%, and 32.83%, respectively, after 125 freeze–thaw cycles.
1.5 RC100-H RC100-H fitting curve
RC50-N RC50-N fitting curve 5.1.4. Uniaxial compressive stress–strain constitutive model
RC50-M RC50-M fitting curve To date, many researchers have proposed the uniaxial compressive
Relative peak strain

1.4 RC50-NM RC50-NM fitting curve stress–strain constitutive model [33–38]. In this chapter, the uniaxial
compressive stress–strain constitutive model proposed by Professor
1.3 Guo is used to fit the stress–strain curve of RCASCC [34]. Fit the
stress-strain curves of each group of RACSCC and obtain the fitting
1.2 graph and fitting parameters as shown in Fig. 21 and Table 4. a and b are
the fitting parameters of the ascending and descending segments of the
1.1 stress–strain curve, respectively. In Table 4, the minimum correlation
coefficient R2 is 0.9036, and the maximum value is 0.9995, indicating
that the fitting effect of the model is good. The larger the RCA substi­
1.0
0 25 50 75 100 125 tution rate is, the smaller a is, and the larger b is. The larger the number
The number of freeze-thaw cycles /times of freeze–thaw cycles is, the smaller a is, and the larger b is. This finding
shows that the slope of the ascending section of the curve decreases
Fig. 19. Relationships between relative peak strains and freeze–thaw cycles.

11
­
C. Zheng et al. Structures 61 (2024) 105973

6.0 6.0
(a) (b)

5.5
Limit strain /×10-3 5.5

Limit strain /×10-3

5.0 5.0

4.5 4.5

RC50-H
4.0 NC-H 4.0 RC50-N
RC50-H
RC50-M
RC100-H
RC50-NM
3.5 3.5
0 25 50 75 100 125 0 25 50 75 100 125
The number of freeze-thaw cycles /times The number of freeze-thaw cycles /times

Fig. 20. Relationship between the ultimate strain and the number of freeze–thaw cycles. (a) Relationship between different RCA substitution rates and ultimate
strain. (b) Relationship between different freeze–thaw environments and ultimate strain.

1.2 1.2 1.2

(a) NC-H (b) NC-H NC-H fitting curve (c) NC-H NC-H fitting curve
RC50-H/RC50-N/RC50-M RC-H RC-H fitting curve RC-H RC-H fitting curve
1.0 /RC50-NM 1.0 RC100-H RC100-H fitting curve 1.0 RC100-H RC100-H fitting curve
RC100-H RC50-N RC50-N fitting curve RC50-N RC50-N fitting curve
Relative peak stress
Relative peak stress

Relative peak stress

NC-H fitting curve RC50-M RC50-M fitting curve RC50-M RC50-M fitting curve
0.8 0.8 0.8
RC50-H/RC50-N/RC50-M RC50-NM RC50-NM fitting curve RC50-NM RC50-NM fitting curve
/RC50-NM fitting curve
0.6 RC100-H fitting curve 0.6 0.6

0.4 0.4 0.4

0.2 0.2 0.2

0.0 0.0 0.0

0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5
Relative peak strain Relative peak strain Relative peak strain

Fig. 21. Comparison of the fitting curve and test curve. (a) 0 times. (b) 50 times. (c) 125 times.

Table 4
Fitting parameters a and b.
No –Freezethaw cycles/times
0 25 50 75 100 125
Parameter a NC-H 1.99 1.72 1.52 1.28 0.97 0.86
RC50-H 1.95 1.68 1.44 1.22 0.9 0.71
RC100-H 1.88 1.6 1.37 1.02 0.77 0.62
RC50-N 1.95 1.67 1.42 1.2 0.93 0.65
RC50-M 1.95 1.65 1.37 1.17 0.84 0.63
RC50-NM 1.95 1.59 1.41 1.04 0.88 0.69
b NC-H 1.51 1.68 1.83 2.04 2.15 2.36
RC50-H 2.21 2.41 2.74 3.01 3.15 3.33
RC100-H 2.91 3.23 3.51 3.58 3.97 4
RC50-N 2.21 2.45 2.71 3.01 3.15 3.34
RC50-M 2.21 2.37 2.78 3.11 3.21 3.44
RC50-NM 2.21 2.38 2.72 3.06 3.17 3.33
Correlation coefficient R2 a NC-H 0.9949 0.992 0.989 0.9862 0.9921 0.9822
RC50-H 0.974 0.9961 0.9933 0.9846 0.9941 0.9993
RC100-H 0.9843 0.9874 0.9894 0.9871 0.9936 0.9888
RC50-N 0.974 0.9791 0.9928 0.9693 0.999 0.9875
RC50-M 0.974 0.9751 0.9612 0.9949 0.9995 0.999
RC50-NM 0.974 0.994 0.9824 0.9962 0.9979 0.9971
b NC-H 0.9971 0.9964 0.9758 0.9862 0.9886 0.9776
RC50-H 0.9968 0.9927 0.9626 0.9786 0.9036 0.9383
RC100-H 0.9954 0.9224 0.9631 0.966 0.9635 0.9927
RC50-N 0.9968 0.9852 0.971 0.9636 0.9491 0.9823
RC50-M 0.9968 0.9829 0.9803 0.9449 0.9755 0.9774
RC50-NM 0.9968 0.9835 0.9648 0.989 0.9895 0.9909

gradually, the shape of the descending section of the curve tends to be value and b as the maximum value.
gentle, and the plastic performance of RCASCC increases gradually. In All a and b values in Table 4 are analyzed by linear regression with
the four freeze–thaw environments, RC50-H has a as the maximum the peak stress. The fitting results are shown in Fig. 22, and the
value and b as the minimum value, and RC50-M has a as the minimum regression formula is shown in Formulas (9) and (10).

12
C. Zheng et al. Structures 61 (2024) 105973

2.8 NC-H NC-H fitting curve

4.4
(a) (b)
RC50-H RC50-H fitting curve 4.0
2.4 RC100-H RC100-H fitting curve
RC50-N RC50-N fitting curve 3.6
RC50-M RC50-M fitting curve
2.0 RC50-NM RC50-NM fitting curve 3.2
2.8
1.6
a

b
2.4
1.2 2.0 NC-H NC-H fitting curve
RC50-H RC50-H fitting curve
1.6 RC100-H RC100-H fitting curve
0.8 RC50-N RC50-N fitting curve
1.2 RC50-M RC50-M fitting curve
0.4 0.8 RC50-NM RC50-NM fitting curve
30 32 34 36 38 40 42 30 32 34 36 38 40 42
Peak stress /MPa Peak strain /×10-3

Fig. 22. Parameter changes with peak stress. (a) Variations in a with peak stress. (b) Variations in b with peak stress.

⎧
⎪ NC − H : 0.2617σ 0 − 8.8028, R2 = 0.9295 cycles. Fig. 23(a) shows that at the same number of freeze–thaw cycles,
⎪
⎪
⎪
⎪
⎪ RC50 − H : 0.242σ0 − 7.7416, R2 = 0.8768 the larger the RCA substitution rate is, the smaller the elastic modulus of
⎨
a=
RC100 − H : 0.187σ 0 − 5.3665, R2 = 0.961
(9) RCASCC. The main reason for the small elastic modulus of RCA-doped
⎪
⎪
⎪
RC50 − N : 0.1869σ 0 − 5.4748, R2 = 0.902 concrete is the low elastic modulus of RCA. Fig. 23(b) shows that the
⎪
⎪
⎪
⎩
RC50 − M : 0.1532σ0 − 4.1087, R2 = 0.9987 damage of RCASCC under a sulfate freeze–thaw environment is
RC50 − NM : 0.1946σ 0 − 5.6756, R2 = 0.9356 increasingly serious. When the number of freeze–thaw cycles is the
⎧ ( ) same, the elastic modulus of RCASCC in the freeze–thaw environment
⎪ NC − H : −( 0.1883ε0 + 9.2634 ×)10− 3 , R2 = 0.9305 of category H is the largest, and that of RCASCC in the freeze–thaw
⎪
⎪
⎪ RC50 − H : − 0.2144ε0 + 10.8347 × 10− 3 , R2 = 0.7728
⎪
⎨ RC100 − H : ( − 0.1591ε + 9.1278) × 10− 3 , R2 = 0.9336
⎪ environment of category M is the smallest. SO2- 4 and Mg
2+
react with the
b= ( 0
) (10) hydration products inside RCASCC in the freeze–thaw environment of
⎪
⎪ RC50 − N : ( − 0.1635ε0 + 8.7419 ) × 10− 3 , R2 = 0.8269 class M, resulting in a decline in the structural stability and elastic
⎪
⎪ RC50 − M : − 0.1481ε + 8.0504 × 10− 3 , R2 = 0.9471
⎪
⎪
⎩ ( 0
) modulus characteristics of RCASCC.
RC50 − NM : − 0.1774ε0 + 9.1329 × 10− 3 , R2 = 0.8412

5.3. Poisson’s ratio

5.2. Elastic modulus
Poisson’s ratio, like the elastic modulus, can evaluate the internal
The elastic modulus is an important evaluation index that reflects the force, deformation resistance, and crack resistance values of concrete
internal force, deformation resistance, and crack resistance of concrete structures. We calculate the Poisson’s ratio of RCASCC according to the
structures. The elastic modulus of concrete is calculated according to specification in Ref. [35]. Fig. 24 shows the Poisson’s ratio versus the
Formula 7.0.5 in ‘Standard methods for testing the physical and me­ number of freeze–thaw cycles. Fig. 24 shows that the Poisson’s ratio of
chanical properties of concrete’ [35]. The elastic modulus change curve each group of RCASCC decreases with the increase in the number of
of RCASCC after sulfate freeze–thaw is shown in Fig. 23. Fig. 23 shows freeze–thaw cycles. Compared with the samples after 0 freeze–thaw
that the larger the number of freeze–thaw cycles is, the smaller the cycles, the elastic odulus of NC-H, RC50-H, RC100-H, RC50-N, RC50-M,
elastic modulus of RCASCC. Compared with the samples after 0 freeze and RC50-NM decrease by 31.67%, 31.16%, 37.26%, 33.02%, 37.21%,
–thaw cycles, the elastic modulus values of NC-H, RC50-H, RC100-H, and 34.42%, respectively, after 125 freeze–thaw cycles. Fig. 24(a)
RC50-N, RC50-M, and RC50-NM decrease by 51.98%, 43.59%, shows that the Poisson’s ratio of RCASCC in each group presents a
58.99%, 59%, 63.95%, and 58.9%, respectively, after 125 freeze–thaw stratified phenomenon. When the number of freeze–thaw cycles is the

35 35
(a) (b)

30 30
Elastic modulus /GPa

Elastic modulus /GPa

25 25

20 20

RC50-H
15 NC-H 15 RC50-N
RC50-H RC50-M
RC100-H RC50-NM
10 10
0 25 50 75 100 125 0 25 50 75 100 125
The number of freeze-thaw cycles /times The number of freeze-thaw cycles /times

Fig. 23. Elastic modulus versus the number of freeze–thaw cycles. (a) Dependence of different RCA substitution rates on the elastic moduli. (b) Relationships
between different freeze–thaw environments and elastic moduli.

13
C. Zheng et al. Structures 61 (2024) 105973

0.23 0.23
(a) (b)
0.22 0.22
0.21 0.21
0.20 0.20
Poisson's ratio

Poisson's ratio
0.19 0.19
0.18 0.18
0.17 0.17
0.16 0.16 RC50-H
0.15 NC-H 0.15 RC50-N
RC50-H RC50-M
0.14 RC100-H 0.14
RC50-NM
0.13 0.13
0 25 50 75 100 125 0 25 50 75 100 125
The number of freeze-thaw cycles /times The number of freeze-thaw cycles /times

Fig. 24. Poisson’s ratio versus the number of freeze–thaw cycles. (a) Relationships between different RCA replacement rates and the Poisson’s ratios. (b) Re­
lationships between different freeze–thaw environments and the Poisson’s ratios.

same, the larger the RCA substitution rate is, the smaller the Poisson’s
ratio. Fig. 24(b) shows that when the number of freeze–thaw cycles is 42
the same, the Poisson’s ratio of the H-type freeze–thaw environment is
the largest, and the Poisson’s ratio of the M-type freeze–thaw envi­ 40
ronment is the smallest.
38
Peak stress /MPa
5.4. Relationship between the dynamic elastic modulus and ultrasonic 36
wave velocity and peak stress
34
5.4.1. Relationship between the dynamic elastic modulus and peak stress NC-H NC-H fitting curve
Fig. 25 shows the relationship between the peak stress and the dy­ 32 RC50-H RC50-H fitting curve
namic elastic modulus of each group of RCASCC after different numbers RC100-H RC100-H fitting curve
30 RC50-N RC50-N fitting curve
of sulfate freeze–thaw cycles. Fig. 25 shows that the peak stress increases
RC50-M RC50-M fitting curve
with increasing dynamic elastic modulus. The peak stress and dynamic RC50-NM RC50-NM fitting curve
elastic modulus values are fitted by linear regression. The regression 28
2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2
equation and fitting graph are obtained, as shown in Formula (11) and Ultrasonic wave velocity /km·s-1
Fig. 25, respectively.
⎧ Fig. 26. Relationship between the ultrasonic wave velocity and peak stress.
⎪
⎪ NC − H : 0.6978Ed + 27.8144, R2 = 0.9393
⎪
⎪
⎪ RC50 − H : 0.6668Ed + 27.3467, R2 = 0.7293
⎪
⎨
RC100 − H : 0.7498Ed + 24.6916, R2 = 0.9428 wave velocity of each group of RCASCC after different numbers of sul­
σ0 = (11) fate freeze–thaw cycles. Fig. 26 shows that there is a positive correlation
⎪
⎪ RC50 − N : 0.7693Ed + 24.7762, R2 = 0.7001
⎪
⎪
⎪
⎪
2
RC50 − M : 1.057Ed + 19.7229, R = 0.9283 between the peak stress and the ultrasonic wave velocity. There is a good
⎩
RC50 − NM : 0.8135Ed + 23.6075, R2 = 0.849 linear correlation between the peak stress and ultrasonic wave velocity.
The fitting figure and formula obtained by fitting are shown in Fig. 26
5.4.2. Relationship between the ultrasonic wave velocity and peak stress and Formula (12), respectively.
Fig. 26 shows the relationship between the peak stress and ultrasonic ⎧
⎪
⎪ NC − H : 3.6503Vd + 25.6812, R2 = 0.8855
⎪
⎪
⎪
⎪ RC50 − H : 3.5713Vd + 25.0961, R2 = 0.6725
⎨
42 σ0 =
RC100 − H : 5.5Vd + 16.68, R2 = 0.9211
(12)
⎪
⎪ RC50 − N : 4.5383Vd + 20.9045, R2 = 0.7539
40 ⎪
⎪
⎪ 2
RC50 − M : 6.0631Vd + 14.9891, R = 0.9773
⎪
⎩
38 RC50 − NM : 4.5313Vd + 20.2956, R2 = 0.8528
Peak stress /MPa

36 6. Conclusions
34
(1) The higher the RCA substitution rate is, the more severe the
32 degradation of the RCASCC surface and damage to the RCASCC
30 NC-H NC-H fitting curve morphology. In the four freeze–thaw environments, the surface
RC50-H RC50-H fitting curve degradation of RCASCC is the most severe in M and the least
28 RC100-H RC100-H fitting curve severe in H, and the difference is small between N and NM. In the
RC50-N RC50-N fitting curve
26 four freeze–thaw environments, the failure mode of RCASCC in
RC50-M RC50-M fitting curve
H is the slightest, and there are no significant differences in N, M,
24 RC50-NM RC50-NM fitting curve
and NM.
8 10 12 14 16 18 20
Dynamic elastic modulus /MPa (2) From 0 to 125 freeze–thaw cycles, the different freeze–thaw
environments have relatively small impacts on the quality loss of
Fig. 25. Dependence of peak stress on the dynamic elastic modulus. RCASCC. In the four freeze–thaw environments, the loss rates of

14
C. Zheng et al. Structures 61 (2024) 105973

the compressive and splitting tensile strengths of RCASCC in M [7] Wang C, Xiao J, Zhang C, et al. Structural health monitoring and performance
analysis of a 12-story recycled aggregate concrete structure. Eng Struct 2020;Vol.
decrease the most. The relationship between the dynamic elastic
205. id. 110102.
modulus loss rate of RCASCC and the number of freeze–thaw [8] Venkrbec V, Klanšek U. Suitability of recycled concrete aggregates from precast
cycles can be divided into three stages: steady growth stage, panel buildings deconstructed at expired lifespan for structural use. J Clean Prod
linear growth stage, and relatively stable stage. The change curve 2020;Vol. 247. id. 119593.
[9] Zheng CL, Li SX, Hou YF, et al. Frost resistance of fiber-reinforced self-compacting
of the ultrasonic wave velocity loss rate of RCASCC with the recycled concrete. Rev Adv Mater Sci 2022;Vol. 61(No. 1):711–25.
number of freeze–thaw cycles can be divided into three stages: [10] Aslani F, Ma G, Wan DLY, et al. Development of high-performance self-compacting
the relatively stable stage, linear growth stage, and steady growth concrete using waste recycled concrete aggregates and rubber granules. J Clean
Prod 2018;Vol. 182:553–66.
stage. In the four freeze–thaw environments, the loss rates of the [11] Zheng CL, Zhao YD, Jin BH. Mechanical properties of steel-polyvinyl alcohol fiber
compressive and splitting tensile strengths of RCASCC in M recycled coarse aggregate self-compacting concrete: orthogonal testing. Arab J Sci
decrease the most. Eng 2023;Vol. 2023:1–14.
[12] Khayat KH. Workability, testing, and performance of self-consolidating concrete.
(3) The depression of the upper part of the stress–strain curve of Mater J 1999;Vol. 96(No. 3):346–53.
RCASCC is increasingly obvious in a sulfate freeze–thaw envi­ [13] Muthu M, Kumar BG, Sadowski Ł. Experimental study on self-compacting concrete-
ronment. There is a linear functional relationship between the filled steel tube samples containing shrinkage-controlling admixtures. J Build Eng
2023;Vol. 70. id. 106465.
peak stress and peak strain. Among the 6 groups of RCASCC, the [14] Zhu W, Chen Y, Li FX, et al. Design and preparation of high elastic modulus self-
relative peak stress of NC-H is the largest and that of RC50-M is compacting concrete for pre-stressed mass concrete structures. J Wuhan Univ
the smallest. In 6 groups of RCASCC, RC50-M has the largest Technol (Mater Sci) 2016;Vol. 31(No. 3):563–73.
[15] Xiao J, Li J, Zhang C. Mechanical properties of recycled aggregate concrete under
relative peak strain, and NC-H has the smallest strain. The rela­
uniaxial loading. Cem Concr Res 2005;Vol. 35(No. 6):1187–94.
tionship between the relative peak stress and relative peak strain [16] Revilla-Cuesta V, Skaf M, Chica JA, et al. Quantification and characterization of the
and the number of freeze–thaw cycles is a linear function. As the microstructural damage of recycled aggregate self-compacting concrete under
number of freeze–thaw cycles increases, the limiting strain of cyclic temperature changes. Mater Lett 2023;Vol. 333. id. 133628.
[17] Neeraja VS, Sharma V. An overview of fresh and mechanical properties of recycled
RCASCC in each group shows a trend of first increasing, then coarse aggregate self-compacting concrete. Mater Today: Proc 2023;Vol. 264. id.
decreasing, and finally increasing. 118703.
(4) The uniaxial compressive stress–strain constitutive model is [18] Martínez-García R, Jagadesh P, Fraile-Fernández FJ, et al. Influence of design
parameters on fresh properties of self-compacting concrete with recycled
used to fit the stress–strain curves of RCASCC before and after aggregate—a review. Materials 2020;Vol. 13(No. 24). id. 5749.
sulfate freeze–thaw action. The results show that fitting [19] Manzi S, Mazzotti C, Bignozzi M. Short and long-term behavior of structural
parameter a gradually decreases and fitting parameter b gradu­ concrete with recycled concrete aggregate. Cem Concr Compos 2013;Vol. 37:
312–8.
ally increases with the increase in the number of freeze–thaw [20] Tang Z, Li W, Ke G, et al. Sulfate attack resistance of sustainable concrete
cycles. The larger the RCA substitution rate is, the smaller the a incorporating various industrial solid wastes. J Clean Prod 2019;Vol. 218:810–22.
value and the larger the b value. There is a linear relationship [21] Huda SB, Shahria Alam M. Mechanical and freeze-thaw durability properties of
recycled aggregate concrete made with recycled coarse aggregate. J Mater Civ Eng
between a and b and the peak stress. 2015;Vol. 27(No. 10):1–9.
(5) The elastic modulus and Poisson’s ratio values of RCASCC in the 6 [22] Jia P, Li L, Zhou J, et al. Performance evolution of recycled aggregate concrete
groups decrease with the increase in the number of freeze–thaw under the coupled effect of freeze-thaw cycles and sulfate attack. Appl Sci 2022;
Vol. 12(No. 14):1–14.
cycles. The larger the substitution rate of RCA is, the smaller the
[23] Niu D, Jiang L, Fei Q. Deterioration mechanism of sulfate attack on concrete under
elastic modulus and Poisson’s ratio of RCASCC. The elastic freeze-thaw cycles. J Wuhan Univ Technol-Mater Sci Ed 2013;Vol. 28(No. 6):
modulus and Poisson’s ratio values are the highest for RCASCC in 1172–6.
the Class H freeze–thaw environment, while those values are the [24] Tian B, Cohen MD. Does gypsum formation during sulfate attack on concrete lead
to expansion? Cem Concr Res 2000;Vol. 30(No. 1):117–23.
lowest for RCASCC in the Class M freeze–thaw environment. [25] Tuyan M, Mardani-Aghabaglou A, Ramyar K. Freeze-thaw resistance, mechanical
(6) The peak stress of the 6 groups of RCASCC increases with and transport properties of self-consolidating concrete incorporating coarse
increasing dynamic elastic modulus and ultrasonic wave velocity. recycled concrete aggregate. Mater Des 2014;Vol. 53:983–91.
[26] Modani PO, Mohitkar VM. Self-compacting concrete with recycled aggregate: a
solution for sustainable development. Int J Civ Struct Eng 2014;Vol. 4(No. 3):
Funding information 430–40.
[27] Yu TQ, Qian JC. Damage theory and its application. National Defense Industry
Publishing House,; 1998. p. 41–2.
Thanks for the help of project 51968060 and project CXXM202239. [28] Liu H, Ma Y, Ma J, et al. Frost resistance of desert sand concrete. Adv Civ Eng 2021;
Vol. 2021. id. 6620058.
[29] China Standards: GB/T 50082–2009. Standard for test methods of long-term
Conflict of interest performance and durability of ordinary concrete, 2009, p. 1.
[30] Tian W, Gao F. Damage and degradation of concrete under coupling action of
freeze-thaw cycle and sulfate attack. Adv Mater Sci Eng 2020;Vol. 2020. id.
There is no conflict of interest in this matter. 8032849.
[31] Kumar R, Verma M, Dev N. Investigation on the effect of seawater condition,
sulphate attack, acid attack, freeze–thaw condition, and wetting–drying on the
References
geopolymer concrete. Iran J Sci Technol, Trans Civ Eng 2022;Vol. 46:2823–53.
[32] Jia P, Li L, Zhou J, et al. Performance evolution of recycled aggregate concrete
[1] Yaowarat T, Horpibulsuk S, Arulrajah A, et al. Compressive and flexural strength of under the coupled effect of freeze-thaw cycles and sulfate attack. Appl Sci 2022;
polyvinyl alcohol–modified pavement concrete using recycled concrete aggregates. Vol. 12(No. 14):1–14.
J Mater Civ Eng 2018;Vol. 30(No. 4). id. 04018046. [33] China Standards: GB 50010–2010. Code for design of concrete structures, 2010, p.
[2] Rashid K, Rehman MU, de Brito J, et al. Multi-criteria optimization of recycled 1.
aggregate concrete mixes. J Clean Prod 2020;Vol. 276. id. 124316. [34] Guo Z.H. Strength and deformation of concrete-experimental basis and constitutive
[3] Wang B, Yan L, Fu Q, et al. A comprehensive review on recycled aggregate and relationship. 1997, pp. 35–38.
recycled aggregate concrete. Resour, Conserv Recycl 2021;Vol. 171. id. 105565. [35] China Standards: GB/T 50081–2019. Standard for test methods of concrete
[4] Gao D, Zhu W, Fang D, et al. Shear behavior analysis and capacity prediction for physical and mechanical properties, 2019, p. 1.
the steel fiber reinforced concrete beam with recycled fine aggregate and recycled [36] Bairagi N, Ravande K, Pareek V. Behavior of concrete with different proportions of
coarse aggregate. Elsevier 2022;Vol. 37:44–55. natural and recycled aggregates. Conserv Recycl 1993;Vol. 9(No. 1-2):109–26.
[5] Sadowska-Buraczewska B, Grzegorczyk-Frańczak M. Sustainable recycling of high- [37] Topcu I, Guncan N. Using waste concrete as aggregate. Cem Concr Res 1995;Vol.
strength concrete as an alternative to natural aggregates in building structures. 25(No. 7):1385–90.
Sustainability 2021;Vol. 13(No). id. 4286. [38] Mander J, Priestley M, Park R. Theoretical stress-strain model for confined
[6] Ibrahim Y. Durability and structural performance of recycled aggregate concrete: a concrete. J Struct Eng 1988;Vol. 114(No. 8):1804–26.
review. Int Rev Civ Eng 2019;Vol. 10(No. 3):135–41.

15