Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
171–181
DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH
LITHIUM SLAG UNDER FREEZE-THAW CYCLES
TRAJNOST RECIKLIRANEGA BETONA Z LITIJEVO @LINDRO
POD OHLAJEV ALNO-SEGREV ALNIMI CIKLI
Yongjun Qin
*
, Jiejing Chen, Ke Liu, Yi Lu
College of Architectural Engineering, Xinjiang University,1230 Yanan Road, Urumqi 830000, China
Prejem rokopisa – received: 2020-07-07; sprejem za objavo – accepted for publication: 2021-12-29
doi:10.17222/mit.2020.126
A water freeze-thaw cycle and sulfate freeze-thaw coupling cycle were explored experimentally to evaluate the durability of re-
cycled concrete with lithium slag (LS). The damage-deterioration law was studied from the aspects of mass-change rate, relative
dynamic modulus of elasticity, and cube’s compressive strength. Based on the relative dynamic modulus of elasticity, the dam-
age-degree equation of the concrete was fitted, and a mechanical-attenuation model related to this parameter and the cube’s
compressive strength was established and verified. The damage mechanism under the action of the sulfate freeze-thaw cycle was
revealed through scanning electron microscopy (SEM). The combination of recycled coarse aggregate (RCA) and LS was bene-
ficial to the anti-deterioration ability of the concrete. During the cycle experiments, the mass and relative dynamic modulus of
elasticity increased initially and then decreased, while the cube’s compressive strength declined continually. The concrete with a
30 % RCA substitution rate and 20 % LS exhibited the optimal comprehensive durability, and specimens with excessive LS
showed more susceptibility to sulfate erosion. The residual compressive strength of concrete structures can be evaluated by mea-
suring the relative dynamic modulus of elasticity as the two parameters are ideally correlated.
Keywords: recycled concrete, lithium slag, freeze-thaw, attenuation model, SEM
Avtorji so v preizkusu ugotavljali trajnost recikliranega betona z izbrano vsebnostjo litijeve `lindre med procesom zdru`enega
ohlajevalno-ogrevalnega (angl.: freeze-thaw cycle) in sulfatno ohlajevalno-ogrevalnega cikla. Zakone slabenja in po{kodbe
betona so {tudirali s stali{~a hitrosti spreminjanja mase, relativnega dinami~nega modula elasti~nosti in kubi~ne tla~ne trdnosti.
Na osnovi relativnega dinami~nega modula elasti~nosti so prilagodili ena~bo za posamezne stopnje po{kodb betona. Postavili in
verificirali so model mehanskega slabenja betona v povezavi s tem parametrom in kubi~no mehansko trdnostjo. Mehanske
po{kodbe pod vplivom sulfatnega ohlajevalno-ogrevalnega cikla so opazovali z vrsti~nim elektronskim mikroskopom (SEM).
Ugotovili so, da ima kombinacija grobega recikliranega agregata (RCA) in litijeve `lindre (LS), pozitivni u~inek na zmanj{anje
slabenja betona. Med eksperimentalnimi cikli sta masa in relativni dinami~ni modul elasti~nosti najprej nara{~ala in nato za~ela
padati, medtem ko je kubi~na tla~na trdnost zvezno padala. Beton s 30 % RCA in 20 % LS je imel optimalno trajnost.
Preizku{anci s prebitkom LS pa so bili bolj ob~utljivi na sulfatno erozijo. Avtorji ugotavljajo, da se preostala tla~na trdnost
betonskih struktur lahko oceni z merjenjem relativnega dinami~nega modula elasti~nosti in da se oba parametra idealno
medsebojno ujemata.
Klju~ne besede: reciklirani beton, litijeva `lindra, ohlajevalno-segrevalni cikel, vrsti~ni elektronski mikroskop, model slabenja
1 INTRODUCTION
Extensive use of concrete materials in the world of
architecture necessitates the evaluation of their durability
and service life. The damages inflicted to concrete struc-
tures as a result of environmental conditions are associ-
ated with a variety of complex physiochemical pro-
cesses. Engineering often encounters freeze-thaw cycles,
carbonation, chloride erosion, sulfate erosion or multiple
damages caused by two types of synergistic damage re-
actions, of which the mechanism of action is not a sim-
ple superposition.
1,2
Buildings in severely cold areas are
often damaged by the freeze-thaw cycle, which is gener-
ally considered to be caused by the pressure due to wa-
ter-freezing expansion of concrete pores as well as seep-
age pressure.
3
So far, scholars have conducted numerous
researches on the freeze-thaw resistance of various con-
cretes and established relevant freeze-thaw damage mod-
els.
4–7
In addition to the cold environment, local soils
contain abundant sulfate in many parts of the world (e.g.,
northwest China, the Alps) where concrete engineering
is vulnerable to the damage of the coupled freeze-thaw
and sulfate corrosion. The coupled degradation mecha-
nism of sulfate and freeze-thaw cycle is complex. On the
one hand, the freezing of free water during the freeze-
thaw cycle can prevent the entry and reaction of sulfate
ions with cement hydration products that form ettringite
and gypsum, leading to a disintegration of a gelatinized
material. On the other hand, the microcracks formed by
the freeze-thaw cycle provide channels for the salt-solu-
tion entry as well as further corrosion of the concrete in-
terior, aggravating the concrete deterioration with the
repetition effect.
8
Based on the concept of šdouble waste recycling’, re-
cycled concrete with lithium slag (LS) is a new type of
environmentally friendly building materials with broad
application prospects. Its macroscopic mechanical prop-
erties have been developed extensively over the years.
Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181 171
UDK 67.017:502.17:669.015.8:669.884:66.040 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 55(2)171(2021)
*Corresponding author's e-mail:
13144180978@163.com (Yongjun Qin)

LS, the waste from the large-scale production of lithium
salt, is a mineral admixture with a composition similar to
fly ash and silica fume. Studies confirmed the pozzolanic
activity of LS and its micro-aggregate effect after a treat-
ment under certain conditions.
9
With its large production
and low costs, LS is considered ideal for recycled con-
crete. Moreover, experiments proved that recycled con-
crete with an appropriate amount of LS exhibits better
physiomechanical properties than ordinary concrete.
10
Generally, the difference between recycled aggregate and
natural aggregate is a distinction in durability between
recycled concrete and ordinary concrete, and relevant re-
search has been carried out continuously in the field. V .
Bulatovi} et al.
11
found that the appropriate type of ce-
ment could help recycled concrete to exhibit the favor-
able resistance to sulfate attacks. B. Lei et al.
12
con-
ducted a coupling experiment on recycled concrete
during mechanical loading and a freeze-thaw cycle in a
salt solution to analyze the influence of a complex envi-
ronment on its durability. Mineral admixtures were sug-
gested to be used for the improvement of the durability
of concrete under certain conditions. Wang et al. experi-
mentally found that an addition of 25 % of fly ash and
5–8 % of silica fume can significantly improve the resis-
tance of recycled concrete to the freeze-thaw and sulfate
attack in a 5 % sodium sulfate solution.
13
W. Tian and F.
Gao
14
found that under the coupling action of the
freeze-thaw cycle and sodium sulfate solution, the con-
crete with 10 % of fly ash outperformed ordinary con-
crete in terms of the freeze-thaw resistance ina5%s o -
dium sulfate solution.
14
A thorough investigation of the durability of recycled
concrete with LS is of a practical significance, promoting
its engineering application. In this research, through an
indoor accelerated freeze-thaw method, the freeze-thaw
cycle and the coupling effect of freeze-thaw and sulfate
were tested on recycled concrete with LS. The effects of
the cycle time, RCA substitution rate and LS amount on
the mass-change rate, relative dynamic modulus of elas-
ticity (RDME) and cube’s compressive strength of the
concrete were studied, and a mechanical attenuation
model related to the RDME and cube’s compressive
strength was established. The damage mechanism was
further studied using scanning electron microscopy
(SEM). This research provides a design basis for the fu-
ture engineering construction and life prediction.
2 MATERIALS AND EXPERIMENTAL
PROCEDURES
2.1 Materials
Ordinary Portland cement of grade 42.5 was used in
this experiment. The LS from the Urumqi lithium-salt
plant was put into use after being dried and ground. The
density and specific surface area of LS were 2.48 g/cm
3
and 417 m
2
/kg, respectively. Table 1 shows the chemical
composition of cement and LS.
Fine aggregate is natural medium-coarse sand with a
fineness modulus of 3.5, and a polycarboxylic super-
plasticizer with a 25 % water-reducing rate was adopted
in the experiment. RCA came from the demolition site of
abandoned buildings in Urumqi. After having been sec-
ondarily crushed by a jaw crusher, the particle size was
screened and adjusted to a 5–20 mm continuous grada-
tion. The physical-performance indexes of natural coarse
aggregate (NCA) and RCA are illustrated in Table 2.
2.2 Specimen preparation
A total of 672 specimens were prepared for this ex-
periment. The water-cement ratio of the specimens was
0.45 and the design strength of C30 was selected. The
mix proportions of the specimens are listed in Table 3
with reference to the Chinese standard DG / TJ
08-2018-2007 and the researches by R. Liang and Li.
15–17
The calculation of additional water was based on the wa-
ter-absorption rate of RCA used in the experiment.
A total of 96 prismoid specimens with dimensions of
(100×100×400) mm were used to measure the mass and
transverse fundamental frequency alongside 576 cube
specimens with dimensions of 100×100×100 mm used to
measure the cube’s compressive strength. The number of
the specimens used for the experiment meets the require-
ments in accordance with the Chinese standard
GB/T50081-2002.
18
Cement, LS, natural sand and natu-
ral pebbles were poured sequentially into a HJW60 sin-
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
172 Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181
Table 1: The chemical composition of cement and LS /%
Materials CaO SiO 2 Al 2O 3 Fe 2O 3 SO 3 MgO Loss R 2OK
2ON a
2O
Cement 55.3 25.4 7.1 2.9 2.8 2.3 2.2 0.9 0.7 0.5
LS 22.0 41.7 18.1 1.2 15.1 0.5 0.4 - 0.3 0.1
Note: LS represents lithium slag
Table 2: Properties of coarse aggregate
Type
Bulk density
/kg/m³
Apparent density/
kg/m³
Water absorption
/%
Crush value
/%
Mud content
/%
Fine powder con-
tent /%
NCA 1536 2687 0.52 - 0.2 0.2
RCA 1472 2417 3.46 14 0.4 0.5
Note: NCA represents natural coarse aggregate; RCA represents recycled coarse aggregate

gle-shaft mixer to mix them evenly. Afterwards, the pre-
viously wetted RCA and water were added to the mixer,
and the materials were mixed for 3 min. The formwork
was removed after 24 h when the concrete specimens
were shaped, and the standard curing was carried out.
2.3 Experimental equipment
A concrete fast freeze-thaw experiment machine and
DT-W18 dynamic elastic modulus meter from Beijing
Digital zhiyilong Instrument Co., Ltd., were employed in
this experiment. The principle of the dynamic elastic
modulus meter is to induce a mechanical vibration of the
specimens. If the external-force frequency is equal to the
fundamental frequency of the specimen, resonance is
generated and the amplitude reaches its corresponding
maximum. In this way, the dynamic elastic modulus
could be calculated using the fundamental frequency.
The mass of a specimen was measured with a Hengxin
ACS series electronic weighing scale; the cube’s com-
pressive strength was gathered by a WAW-1000 elec-
tric-hydraulic servo universal test machine.
2.4 Experimental procedure
Half of the specimens from each mix were immersed
in a 5-% Na
2
SO
4
solution and the other half in water for
4 days after the 24-day standard curing. The liquids had
to be kept 20 mm above the surface of the specimens. A
complete freeze-thaw cycle lasted for about 4 h. Accord-
ing to the relevant provisions from the Chinese standard
GB/T 50082-2009, the average melting time in a
freeze-thaw cycle was about 1 hour and 20 minutes and
the central temperature in the conditioning chamber was
between 7 °C and 19 °C.
19
The specimens were taken out
every 25 cycles in order to dry their surfaces and collect
the data about the cube’s compressive strength, mass
change, RDME and obtain SEM pictures. The failure of
the concrete specimens was identified when the RDME
decreased to 60 % or the mass-change rate reached 5 %.
2.5 Data collection
In this research, three damage indexes were used: the
mass change rate, RDME and cube’s compressive
strength used to analyze the durability of recycled con-
crete with LS. These indexes can describe the process,
from its compactness to looseness, within the concrete
structure, representing the overall damage of the speci-
mens to some extent.
20
The mass-change rate is calcu-
lated as Equation (1):
ΔW
WW
W
Ni
iN i
i
=
−
0
0
·100 % (1)
Where W
Ni
is the mass-change rate (%) of a speci-
men after n cycles; W
0i
(g) is the initial mass; and W
Ni
(g)
is the mass of a specimen after N cycles.
The RDME is calculated as follows:
E
WL F
a
d
=××
−
13 244 10
4
32
4
, (2)
p
E
E
i
N
=
d
d0
·100 % (3)
where E
d
(MPa) is the RDME of a specimen; a(mm) is
the side length of a square section; L (mm) represents
the length; W (kg) is the mass of a specimen; f (Hz) is
the fundamental frequency of a specimen in transverse
vibration; p
i
represents the RDME reading; E
dN
(MPa)
represents the dynamic modulus of elasticity after N cy-
cles; and E
d0
(MPa) stands for the initial dynamic elastic
modulus of a specimen. The cube’s compressive
strength of specimens was examined under a loading
rate of 0.5 MPa/s, and the average value of each group
was taken as the final result.
3 RESULTS AND DISCUSSION
3.1 Mass-change rate
The mass-change rate of the specimens increased
gradually as the number of freeze-thaw cycles rose. The
constant tensile and compressive stress inside the con-
crete occurred due to the freezing and expansion of free
water that resulted in a fatigue failure and specimen mass
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181 173
Table 3: Design of mix proportions
Code
RCA
/%
LS
/%
Mix proportion /kg m
–3
Water
Additional
water
Cement LS NCA RCA Sand
R0L0 0 0 195.00 0.00 433.00 0.00 1221.40 0.00 523.50
R0L20 0 20 195.00 0.00 346.40 86.60 1221.40 0.00 523.50
R30L0 30 0 195.00 7.84 433.00 0 854.98 366.42 523.50
R30L15 30 15 195.00 7.84 368.05 64.95 854.98 366.42 523.50
R30L20 30 20 195.00 7.84 346.40 86.60 854.98 366.42 523.50
R30L25 30 25 195.00 7.84 324.75 108.25 854.98 366.42 523.50
R50L20 50 20 195.00 13.07 346.40 86.60 610.70 610.70 523.50
R70L20 70 20 195.00 18.30 346.40 86.60 366.42 854.98 523.50
Note: R0L0 represents natural concrete; R represents recycled coarse aggregate; L represents lithium slag

changes. Unlike the water freeze-thaw cycle (which is an
erosion-failure behavior), the curves for the sulfate
freeze-thaw cycle (a complex coupling of surface scaling
and internal damage) show an obvious mass increase be-
fore the 25
th
cycle, suggesting that these two cycles have
different mechanisms. In the sulfate freeze-thaw cycle,
Na
2
SO
4
reacts with cement hydration products to pro-
duce expansive substances, including gypsum (CaSO
4
·
2H
2
O) and ettringite (3CaO·Al
2
O
3
·CaSO
4
·32H
2
O).
21–23
In
the early stages of the sulfate freeze-thaw cycle, the
amount of erosion causes a mass decrease, thus account-
ing for a negative growth of the mass-change rate. The
chemical-reaction formula of the ettringite sulfate ero-
sion is as follows:
Na
2
SO
4
·10H
2
O + Ca (OH)
2
 CaSO
4
·2H
2
O + 2NaOH + 8H
2
O
4CaO·Al
2
O
3
·13H
2
O+3(CaSO
4
·2H
2
O) +14H
2
O 3CaO·Al
2
O
3
·3CaSO
4
·32H
2
O+ Ca (OH)
2
In the later period of the cycle experiment, the R0L20
group showed good resistance against the mass change,
which proved that the addition of LS moderately im-
proved the compactness of concrete, making the concrete
cementation system more stable and resistant against the
erosion of freeze-thaw. The change rate of mass is far
from a simple linear relationship as it includes a combi-
nation of relatively stable and rapid changes, as shown in
Figure 1. This curve trend is basically consistent with
the theory of water segregation and stratification pro-
posed by A. R. Collins who suggests that the frost ero-
sion of concrete is progressive among the surface lay -
ers.
24
Under a continuous experiment, an ice layer gradu-
ally forms on the surface of the concrete in contact with
water. When the frost-heaving force caused by the
freeze-thaw environment exceeds its ultimate tensile ca-
pacity, the ice layer falls off, thereby changing the con-
crete quality.
With the same LS amount, a higher substitution rate
for RCA leads to a greater mass change and a faster dete-
rioration speed. This is because the water saturation of
concrete is one of the most important influential factors
when it comes to frost resistance. On the one hand, the
initial defects of RCA reduce the compactness and
impermeability of concrete, causing more voids and pro-
viding the condition for the salt solution and water mole-
cules to enter and accumulate in the interior of the con-
crete. On the other hand, a large number of interface
transition areas of new and old sand slurry further weak-
ens the concrete cohesion and the overall expansion re-
sistance. The experimental group including RCA is less
resistant to the mass change than the R0L0 group, and
the R30L20 group is the closest to the mass change of
the R0L0 group, with a difference in the mass-change
rate of only 7 % and 9 %, respectively, after 100 water
freeze-thaw cycles and 100 sulfate freeze-thaw cycles.
From another perspective, the difference in the mass
change between the water freeze-thaw cycle and sulfate
freeze-thaw cycle may be due to the periodic change in
the temperature, which leads to a decrease in the solubil-
i t yo f5%N a
2
SO
4
, whose solubility in a saturated solu-
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
174 Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181
Figure 1: Mass-change rate and cycle experiment: a) R0L0 and specimens with 20 % L, b) R0L0 and specimens with 30 % RCA

tion (100 g water) at 20 °C is 20.5 g, while in a saturated
solution at 10 °C, it decreases to 9.5 g so that crystal pre-
cipitation occurs and fills in the pores and microcracks.
The specific reaction equation is as follows:
Ca (OH)
2
+N a
2
SO
4
 Ca
2+
+S O
4
2–
+N a
+
+O H
–
Ca (OH)
2
+N a
2
SO
4
+2 H
2
O CaSO
4
·2 H
2
O + 2NaOH
The mass-change rate of the specimens showed an
obvious difference between the sulfate freeze-thaw and
water freeze-thaw, especially in the later stage of the cy-
clic experiment. After 100 cycles, the mass-change rate
of almost all the experimental groups exposed to the sul-
fate freeze-thaw cycles was higher than for those ex-
posed to the water freeze-thaw cycles. Amongst them,
R50L20, R70L20 and R30L25 were intensely damaged
under the sulfate freeze-thaw experimental conditions so
that no effective data could be obtained, indicating a seri-
ous deterioration effect of the sulfate corrosion. It is also
worth mentioning that after 100 cycles, the mass-
change-rate difference for the R0L20 group under water
freeze-thaw cycles and sulfate freeze-thaw cycles could
be reduced to 6 %, which proves that 20 % of LS could
ensure that the specimens exhibited an adequate sulfate
resistance to a certain extent.
3.2 Relative dynamic modulus of elasticity (RDME)
3.2.1 Trend analysis
The R30L20 group with the slowest RDME decline
rate in the first 25 cycles under the two conditions still
maintained the maximum value after 125 cycles, which
was 11 % and 12 % higher than that of ordinary con-
crete, respectively, displaying better durability, as shown
in Figure 2. The RDME increases with the addition of
pozzolanic LS in an amount below 20 %. The secondary
hydration reaction between LS and Ca (OH)
2
, the prod-
uct of cement hydration, produces large amounts of
densely structured C-S-H and C-A-H cementitious mate-
rials, which can obviously change the transverse funda-
mental frequency. Simultaneously, the cementitious ma-
terials can effectively fill in the pores of concrete,
making up for the RCA-caused decline in the concrete
permeability and reducing the damage caused by the
early freeze-thaw cycle.
25,26
The most serious damage occurred in R30L25 with
an accompanied slow hydration rate. In the middle and
later periods of the sulfate freeze-thaw cycle, the RDME
of R50L20 and R70L20 decreases sharply, falling below
the levels of R0L0. The RDME degradation rate of the
70-% RCA is evidently faster than that of ordinary con-
crete. According to the classical critical water saturation
theory, concrete is easily damaged when the water satu-
ration of concrete mortar reaches a certain level.
27
RCA
with old mortar induces water absorption by micro-
cracks, reaching the critical saturation level, and the in-
terface transition zone between old and new mortars can
easily produce the osmotic pressure, aggravating the
damage to the concrete. However, after the water
freeze-thaw cycle, the RDME of R50L20 is higher than
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181 175
Figure 2: Relative dynamic modulus of elasticity and the cyclic experiment: a) R0L0 and specimens with 20 % LS, b) R0L0 and specimens with
30 % RCA

that of R0L0, which supports the theory proposed by
M. Mao et al.,
28
according to which the porous and loose
microstructure of RCA can play a role in dispersing the
condensation-water pressure to a certain extent. There-
fore, during a practical application of the recycled con-
crete mixed with LS, its proportion issue, in combination
with environmental factors, needs to be considered.
According to the basic principle of the damage me-
chanics of concrete, p
i
is the freeze-thaw damage vari-
able.
20
Damage factor D can be defined as:
DN
EE
E
N
() =
−
0
0
·100 % (4)
where E
N
is the dynamic elastic modulus after N cycles
of the sulfate freeze-thaw; and E
0
is the initial dynamic
elastic modulus. The relative damage degree of a con-
crete specimen after N freeze-thaw cycles can be de-
fined as G
i
:
G
DN
DN
i
i
=
()
()
0
(5)
where D
0
(N) and D
i
(n) are the damage factors of the
concrete specimen after N cycles of the water
freeze-thaw and Na
2
SO
4
freeze-thaw, respectively. G
i
re-
veals the difference between the mechanisms of the wa-
ter freeze-thaw cycle and sulfate freeze-thaw cycle.
Before the 25
th
cycle, the mechanism of water
freeze-thaw is almost opposite to that of sulfate freeze-
thaw, as shown in Figure 3. After the 50
th
cycle, as a re-
sult of the salt solution having a lower freezing point
than water, the actual freezing duration of the specimens
is far shorter than that of the liquid water in the same cy-
cle, hence limiting the early sulfuric acid ion erosion ef-
fect.
29
Damages made to most of the specimens are quite
similar, so the relative damage degrees of most groups
are close to 1. As the cycle experiment continues, the rel-
ative damage is gradually intensified, indicating that the
negative effect of the sulfate erosion is increasingly obvi-
ous.
25,30
In the experimental process, the
sulfate-enhanced erosion occurred the earliest in the
R50L20 group, and the R30L25 group failed the earliest.
After 100 cycles, the relative damage degrees of the
R30L20 group and R30L0 group were the largest, prov-
ing that the positive effect of 20 % of LS on the RDME
was gradually offset by the sulfate corrosion. Hence,
most of C-S-H and C-A-H were decalcified and decom-
posed at this time, and the negative effect of sulfate was
dominant. The increase in the relative damage degree in
the later stage shows that the damage caused by sulfate
freeze-thaw is much faster than that caused by water
freeze-thaw.
3.2.2 RDME damage model
In accordance with the changing law of the RDME of
the recycled concrete with LS under freeze-thaw cycle
and the classical freeze-thaw damage model, the follow-
ing exponential relationship can be established:
31
Pe
N
N
=  (6)
where p
i
is the RDME of the material (%); N is the num-
ber of the freeze-thaw cycles; is the material coeffi-
cient; and  is the attenuation coefficient. The fitted
curve and correlation fitting parameters are shown in
Table 4.
As far as the goodness-of-fit is concerned, the RDME
model based on the attenuation coefficient and material
coefficient is not as applicable to the sulfate freeze-thaw
cycle as to the water freeze-thaw cycle, especially for the
group with a high LS amount and high RCA substitution
rate. Specifically, the fitting coefficient for R30L25 is
only 0.731, and for R70L20, it is only 0.817. In this re-
gard, Zhao proposed a more universal quadratic polyno-
mial function:
32
PA xB xC
N
=⋅+⋅+
2
(7)
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
176 Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181
Figure 3: Relative-damage-degree trend: a) R0L0 and specimens with 20 % LS, b) R0L0 and specimens with 30 % RCA

where A, B are the coefficients; and C is the intercept.
The fitting effect is shown in Table 5.
The quadratic function outperforms the exponential
function in its description of the damage process for the
recycled-concrete mixed with LS under a sulfate attack.
The average optimization degree of R
2
is 12.32 % with a
maximization of 30.59 %, which is more consistent with
the reference requirements of practical engineering.
3.3 Cube’s compressive strength
The cube’s compressive strength reflects the overall
failure of specimens.
33
Unlike the ascending curves of
the mass-change rate and RDME emerging in a sulfate
freeze-thaw cycle, the cube’s compressive strength of the
specimens decreases throughout the cycle experiment,
with a greater decline range in the later cycles, as shown
in Figure 4. With the increase in the number of cycles,
the change of mortar and the intensification of internal
cracks led to a reduction in the concrete stress area and
horizontal binding force, causing the surface to peel off
with more ease from the whole specimen upon compres-
sion. Apparently, the specimens were severely denudated
after multiple cycles, resulting in the exposed surface to
be uneven. After repeated cycles, at the macro-strength
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181 177
Table 4: Correlation fitting parameters (1)
Group R0L0 R0L20 R30L0 R30L15 R30L20 R30L25 R50L20 R70L20
 100.74 101.66 100.70 101.46 101.47 99.40 100.57 99.69
 /10
–3
water –2.36 –2.31 –2.02 –1.93 –1.65 –2.53 –1.77 –2.37
sulfate –2.54 –2.49 –2.58 –2.24 –2.03 –2.07 –2.35 –2.46
R
2
water 0.991 0.978 0.988 0.977 0.978 0.988 0.986 0.996
sulfate 0.864 0.885 0.888 0.902 0.892 0.731 0.843 0.817
Note: a represents the material coefficient; represents the attenuation coefficient; R
2
represents the goodness of fit
Table 5: Correlation fitting parameters(2)
Group R0L0 R0L20 R30L0 R30L15 R30L20 R30L25 R50L20 R70L20
A/10
–3
–1.60 –1.72 –0.90 –1.05 –1.25 –4.40 –1.42 –1.18
B/10
–3
–76.09 –43.55 –156.43 –101.49 –60.70 –63.87 –118.19 –163.23
C 101.70 100.96 102.34 101.82 101.46 100.77 102.06 102.49
R
2
0.973 0.988 0.952 0.968 0.978 0.955 0.921 0.902
Figure 4: Cube’s compressive strength and cyclic experiment: a) R0L0 and specimens with 20 % LS, b) R0L0 and specimens with 30 % RCA

level, the cube’s compressive strength of most recycled
concrete with LS was evidently higher than that of ordi-
nary concrete.
Generally, the specimens with a larger relative
strength have a higher relative-strength surplus after the
cycle experiments. The anti-sulfate freeze-thaw strength
degradation is optimized with the addition of 30 % of
RCA, but excessive RCA will enlarge the spalling-transi-
tion area inside the specimens where cracks easily widen
and propagate. In terms of the failure form, however,
bone destruction rarely occurs in the specimens. The use
of LS contributes to the alleviation of the deterioration of
compressive strength. Secondary hydration reaction
plays a positive role in both water freeze-thaw and sul-
fate freeze-thaw.
34,35
However, an excessive use of LS, in-
stead of cement, is equivalent to an increase in the effec-
tive water-cement ratio, resulting in the alteration of the
concrete’s internal-water saturation, which is an impor-
tant influence factor for the freeze-thaw damage. In the
middle and later periods of a cycle experiment, the de-
velopment trend of the cube’s compressive strength in
each group is similar to that of the RDME.
3.4 Compressive-strength attenuation model
The freeze-thaw damage of concrete is evaluated pri-
marily with two indexes: the RDME and mass-change
rate, which mostly require non-destructive tests. Destruc-
tive tests are still needed to understand the whole me-
chanical-degradation law of concrete under the action of
freeze-thaw. Results show the RDME to be sensitive to
the number of freeze-thaw cycles, ultimately defining the
freeze-thaw damage of concrete. The deterioration of the
cube’s compressive strength with freeze-thaw cycles can
be expressed with a function including RDME.
The cube’s compressive strength is inversely propor-
tional to the number of freeze-thaw cycles, and the curve
can be regarded as a continuous differentiable function.
36
If the cube’s compressive strength of a specimen after 0
cycles is f
0
and it is f(N) after N cycles, then the decay
rate of the compressive strength afterN+ N cycles is
as follows:
fN N fN
fN
N
() ( )
()
+−
=
Δ
Δ  1
(8)
where 1
(< 0) is the constant representing the loss of
compressive strength per freeze-thaw cycle.
Equation (8) can be transposed as:
d
d
fN
N
fN
()
) = 1
( (9)
After integration:
fN
f
e
N
()
0
1
=
 (10)
Similarly, the RDME of concrete is inversely propor-
tional to the number of freeze-thaw cycles, hence the re-
lationship between the two functions can be regarded as
a differentiable function. By referring to the above
freeze-thaw damage-index model (6), the formula can be
transposed as:
N
P
a
N
=
1
 ln (12)
Integrating the relationship between the RDME and
cube’s compressive strength yields:
fN
fa
P
N
()
0
1
=
   (13)
Where 1/a = k
1
and 1
/ = b
i
. The relative residual
strength of a concrete specimen is:
fN
f
kP
i N
b
()
0
1
= (14)
Clearly, the relationship between the relative residual
compressive strength and RDME follows a power func-
tion. For a given RDME, the cube’s compressive strength
and the number of freeze-thaw cycles of the specimens
do not strictly obey the inverse proportional function;
therefore, k
i
and b
i
can be determined according to the
above model fitting correction based on the experiment
results. The specific values are shown in Table 6.
As seen from Table 6, the fitting coefficient R
2
is
more than 0.88 for all groups, exceeding 95 % in 62.5 %
of the groups under water freeze-thaw and 50 % of the
groups under sulfate freeze-thaw, indicating an ideal cor-
relation between the performance indexes and verifying
the rationality and reliability of the model. Therefore, the
residual strength of concrete structures in similar envi-
ronments can be estimated with the RDME.
4 MICROSCOPIC-MECHANISM ANALYSIS
Based on the above damage-index analysis, three
most representative groups, R30L25, R30L20 and
R70L20, were selected for a microstructure analysis to
further explore the damage mechanism after erosion due
to the sulfate freeze-thaw cycles. After 25 cycles, fewer
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
178 Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181
Table 6: Fitting-coefficient value
Group R0L0 R0L20 R30L0 R30L15 R30L20 R30L25 R50L20 R70L20
Medium
water
k
i
100.94 100.33 98.52 101.97 102.34 91.73 99.06 93.67
b i
0.52 0.50 0.40 0.54 0.53 0.33 0.40 0.28
NaSO
4
k i 118.64 110.06 110.59 110.91 109.61 105.16 106.41 114.58
b i
0.81 0.64 0.73 0.67 0.65 0.38 0.50 0.59

microcracks formed in each group of the concrete speci-
mens, and the interiors showed a uniform continuum, as
shown in Figures 5a to 7a. A good spatial network
structure was formed between hydration products and
aggregates with smooth and regular pore walls, suggest-
ing the necessity of the internal space within the speci-
mens for stress relief due to a volume expansion.
SEM images show the microstructure of R30L25
with incomplete hydration to be relatively loose and not
as smooth as the other two groups, facilitating the pene-
tration of sodium sulfate solution and providing an ex-
planation for the largest mass increase of R30L25.
37
Af-
ter 75 cycles, the microstructures of the three groups
were obviously damaged, especially in R30L25 where
abundant corrosion products appeared, as seen on Fig-
ures 5b to 7b. An energy spectrum analysis showed that
the main elements are Al, Si, Ca and O, confirming that
the needle-like corrosion crystals are mainly ettringite.
38
Both R30L20 and R70L20 have a loose and flocculent
colloidal structure, while microcracks are densely dis-
tributed and mutually crossing, indicating the macro-
scopic performance involving an increased mass-change
rate, and decreased RDME and cube’s compressive
strength.
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181 179
Figure 6: SEM pictures of R30L20: a) after 25 cycles, b) after 75 cy-
cles, c) after 100 cycles
Figure 5: SEM pictures of R30L25: a) after 25 cycles, b) after 75 cy-
cles, c) after 100 cycles

With further sulfate freeze-thaw corrosion, a mass of
corrosive products appeared in all three groups. Expan-
sive products (e.g., acicular ettringite and short-column
Ca
2
SO
4
·2H
2
O), gathering in large quantities and occupy-
ing the pore space were formed after cement hydration
(Figures 5c to 7c extending predominantly beyond the
range of the pore interior.
39
Among them, internal corro-
sive products evidently occupy a large part of the internal
structure of R30L25. Based on the three damage indexes,
it is easily proven that R30L25 failed first. Further devel-
opment of internal cracks in the other two groups led to a
reduction in the cohesion and compactness of the con-
crete, leading to a decreased macroscopic performance
of the concrete and accelerated inward entrance of the
solution. The increase in the SO
4
–
concentration indicates
augmented quantities of expansible products (e.g.,
ettringite and flaky gypsum), which accelerated the cor-
rosion and fundamentally destroyed the specimens. The
deterioration of the internal structure was more intense in
R70L20 than in R30L20, which had a lower macroscopic
damage index.
5 CONCLUSIONS
The survey of recycled concrete with LS investigates
different elements of durability under the conditions of
water freeze-thaw cycles and sulfate freeze-thaw cycles.
In the early stage of cycles, water freeze-thaw and sulfate
freeze-thaw have similar effects on the specimens; how-
ever, with the increase in the number of cycles, the dam-
age of sulfate freeze-thaw was accelerated, showing a
more serious deterioration.
In the sulfate freeze-thaw, the mass-change rate and
RDME increase in the early stage, while the cube’s com-
pressive strength is unaffected by certain fine cracks and
declines. R30L20 displays the optimal durability, whilst
R30L25 exhibits an inadequate antifreeze performance.
Additionally, the new mechanical attenuation model,
related to the RDME and cube’s compressive strength,
was verified with respect to its rationality and reliability.
This model can be used to evaluate the residual cube’s
compressive strength of the recycled concrete with LS in
cold regions.
The microstructure analysis illustrates the conditions
for an incomplete hydration reaction caused by excessive
LS and its impact on the acceleration of the inter-
nal-structure deterioration of concrete under external
conditions.
Acknowledgment
This work was fully supported by a grant from the
National Natural Science Foundation of China (Study on
the performance and constitutive relation of recycled
concrete with lithium slag as a mineral admixture),
China (Project No. 51668061).
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
180 Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181
Figure 8: Elemental analysis of R30L25
Figure 7: SEM pictures of R70L20: a) after 25 cycles, b) after 75 cy-
cles, c) after 100 cycles

6 REFERENCES
1
S. Zhang, B. Zhao, Research on the performance of concrete materi-
als under the condition of freeze-thaw cycles, Eur. J. Environ. Civ.
En., 17 (2013) 860–871, doi:10.1080/19648189.2013.826601
2
J. Gong, J. Cao, Y . Wang, Effects of sulfate attack and dry-wet circu-
lation on creep of fly-ash slag concrete, Constr. Build. Mater., 125
(2016), 12–20, doi:10.1016/j.conbuildmat.2016.08.023
3
R. Zhao, Y. Yuan, Z. Cheng, T. Wen, J. Li, F. Li, Z. J. Ma,
Freeze-thaw resistance of Class F fly ash-based geopolymer con-
crete, Constr. Build. Mater., 222 (2019) 474–483, doi:10.1016/
j.conbuildmat.2019.06.166
4
P. J. Tikalsky, J. Pospisil, W. MacDonald, A method for assessment
of the freeze–thaw resistance of preformed foam cellular concrete,
Cement Concrete Res., 34 (2004) 889–893, doi:10.1016/
j.cemconres.2003.11.005
5
A. E. Richardson, K. A. Coventry, S. Wilkinson, Freeze/thaw dura-
bility of concrete with synthetic fibre additions, Cold Reg. Sci.
Technol., 83–84 (2012) 49–56, doi:10.1016/j.coldregions.2012.
06.006
6
C. Karakurt, Y . Bayazýt, Freeze-thaw resistance of normal and high
strength concretes produced with fly ash and silica fume, Adv. Mater.
Sci., Eng., 2015 (2015) 1–8, doi:10.1155/2015/830984
7
Y . Fu, L. Cai, W. Yonggen, Freeze–thaw cycle test and damage me-
chanics models of alkali-activated slag concrete, Constr. Build. Ma-
ter., 25 (2011) 3144–3148, doi:10.1016/j.conbuildmat.2010.12.006
8
H. A. F. Dehwah, Effect of sulfate concentration and associated cat-
ion type on concrete deterioration and morphological changes in ce-
ment hydrates, Constr. Build. Mater., 21 (2007), 29–39,
doi:10.1016/j.conbuildmat.2005.07.010
9
W. Yiren, W. Dongmin, C. Yong, Z. Dapeng, L. Ze, Micro-morphol-
ogy and phase composition of lithium slag from lithium carbonate
production by sulphuric acid process, Constr. Build. Mater., 203
(2019) 304–313, doi:10.1016/j.conbuildmat.2019.01.099
10
Y . Qin, J. Chen, Z. Li, Y . Zhang, The mechanical properties of recy-
cled coarse aggregate concrete with lithium slag, Adv. Mater. Sci.
Eng., 2019 (2019), 1–12, doi:10.1155/2019/8974625
11
V . Bulatovi}, M. Mele{ev, M. Radeka, V . Radonjanin, I. Luki}, Eval-
uation of sulfate resistance of concrete with recycled and natural ag-
gregates, Constr. Build. Mater., 152 (2017) 614–631, doi:10.1016/
j.conbuildmat.2017.06.161
12
B. Lei, W. Li, Z. Tang, V . W. Y . Tam, Z. Sun, Durability of recycled
aggregate concrete under coupling mechanical loading and
freeze-thaw cycle in salt-solution, Constr. Build. Mater., 163 (2018)
840–849, doi:10.1016/j.conbuildmat.2017.12.194
13
D. Wang, X. Zhou, Y . Meng, Z. Chen, Durability of concrete con-
taining fly ash and silica fume against combined freezing-thawing
and sulfate attack, Constr. Build. Mater., 147 (2017),398–406,
doi:10.1016/j.conbuildmat.2017.04.172
14
W. Tian, F. Gao, Damage and degradation of concrete under coupling
action of freeze-thaw cycle and sulfate attack, Adv. Mater. Sci. Eng.,
2020 (2020) 1–17, doi:10.1155/2020/8032849
15
DG/TJ 08-2018-2007: The technical code on the application of recy-
cled concrete, China, 2007
16
R. Liang, Study on the performance and application of recycled
coarse aggregate of waste concrete in Urumqi, China, Xinjiang Uni-
versity, 2014
17
L. Li, Design and performance of recycled concrete mix of mineral
admixture in the cold region of Xinjiang, China, Xinjiang University,
2014
18
GB/T 50081-2002: Test method for mechanical properties of ordi-
nary concrete, China, 2003
19
GB/T 50082-2009: Standard test methods for long-term performance
and durability of ordinary concrete, China, 2009
20
X. Ren, X. M. Wang, T. J. Zhao, Research progress and model re-
view of freezing-thawing and salt-freezing degradation mechanism
of concrete, Concrete, 9 (2012) 15–18, doi:10.3969/j.issn.1002-
3550.2012.09.006
21
W. Qiu, F. Teng, S. Pan, Damage constitutive model of concrete un-
der repeated load after seawater freeze-thaw cycles, Constr. Build.
Mater., 236 (2020), 117560, doi:10.1016/j.conbuildmat.2019.117560
22
W. L. Jia, H. Li, Q. Zhang, Frost resistance of recycled aggregate
concrete under sulfate environment, Silicate Bulletin, 35 (2016),
1816–1820, doi:10.16552/j.cnki.issn1001-1625.2016.06.031
23
H. Siad, M. Lachemi, M. ªahmaran, K. M. A. Hossain, Precondi-
tioning method for accelerated testing of concrete under sulfate at-
tack, ACI Mater. J., 113 (2016), 222–225, doi:10.14359/51688705
24
A. R. Collin, The destruction of concrete by frost, Journal of Institu-
tion of Civil Engineers, 23 (1944), 29–41, doi:10.1680/ijoti.1944.
14086
25
Y . Li, R. Wang, S. Li, Y . Zhao, Y . Qin, Resistance of recycled aggre-
gate concrete containing low- and high-volume fly ash against the
combined action of freeze–thaw cycles and sulfate attack, Constr.
Build. Mater., 166 (2018), 23–34, doi:10.1016/j.conbuildmat.2018.
01.084
26
F. U. A. Shaikh, S. W. M. Supit, Compressive strength and durability
properties of high volume fly ash (HVFA) concretes containing
ultrafine fly ash (UFFA), Constr. Build. Mater., 82 (2015), 192–205,
doi:10.1016/j.conbuildmat.2015.02.068
27
G. Fagerlund, The significance of critical degrees of saturation at
freezing of porous and brittle materials, Durability of Concrete:
American Concrete Institute, Detroit 1975, 13–65
28
M. Mao, D. Zhang, Q. Yang, W. Zhang, Study of durability of con-
crete with fly ash as fine aggregate under alternative interactions of
freeze-thaw and carbonation, Adv. Civ. Eng., 2019 (2019) 1–15,
doi:10.1155/2019/4693893
29
Q. H. Xiao, Q. Li, Z. Y . Cao, W. Y . Tian, The deterioration law of re-
cycled concrete under the combined effects of freeze-thaw and sul-
fate attack, Constr. Build. Mater., 200 (2019) 344–355, doi:10.1016/
j.conbuildmat.2018.12.066
30
L. Jiang, D. Niu, L. Yuan, Q. Fei, Durability of concrete under sul-
fate attack exposed to freeze–thaw cycles, Cold Reg. Sci. Technol.,
112 (2015) 112–117, doi:10.1016/j.coldregions.2014.12.006
31
J. G. Niu, F. L. Zuo, J. L. Wang, C. B. Xie, Freeze-thaw damage
model of plastic-steel fiber reinforced lightweight aggregate con-
crete, Journal of Building Materials, 21 (2018), doi:10.3696/j.issn.
1007-9629.2018.02.010
32
Y . R. Zhao, J. Gao, L. Wang, Z. L. Guo, Study on frost resistance of
basalt fiber concrete under unilateral freezing and thawing, Concrete,
4( 2020), 70–73, doi:10.3969/j.issn.1002-3550.2020.04.017
33
M. Sun, D. Xin, C. Zou, Damage evolution and plasticity develop-
ment of concrete materials subjected to freeze-thaw during the load
process, Mech. Mater., 139 (2019), 103192, doi:10.1016/j.mechmat.
2019.103192
34
D. Zhang, M. Mao, Q. Yang, W. Zhang, P. Han, Experimental inves-
tigation of neutralisation of concrete with fly ash as fine aggregate in
freeze-thaw environment, Adv. Civ. Eng., 2019 (2019), 1–12,
doi:10.1155/2019/6860293
35
L. J. Malvar, L. R. Lenke, Efficiency of fly ash in mitigating al-
kali-silica reaction based on chemical composition, ACI Mater. J.,
103 (2016), 319–326, doi:10.3166/rmpd.7.543-554
36
C. D. Xu, Z. Li, X. L. Zhu, Z. P. Yao, M. Y . Wang, P. Zhang, Study
on mechanical properties of concrete under salt-frost erosion, Yellow
River, doi:10.3696/j.issn.1000-1379.2019.09.028
37
D. Zhang, M. Mao, S. Zhang, Q. Yang, Influence of stress damage
and high temperature on the freeze–thaw resistance of concrete with
fly ash as fine aggregate, Constr. Build. Mater., 229 (2019), 116845,
doi:10.1016/j.conbuildmat.2019.116845
38
H. Nguyen, T. Chang, J. Shih, C. Chen, Influence of low calcium fly
ash on compressive strength and hydration product of low energy
super sulfated cement paste, Cem. Concr. Comp., 99 (2019) 40–48,
doi:10.1016/j.cemconcomp.2019.02.019
39
M. F. Najjar, M. L. Nehdi, A. M. Soliman, T. M. Azabi, Damage
mechanisms of two-stage concrete exposed to chemical and physical
sulfate attack, Constr. Build. Mater., 137 (2017) 141–152,
doi:10.1016/j.conbuildmat.2017.01.112
Y. QIN et al.: DURABILITY PROPERTIES OF RECYCLED CONCRETE WITH LITHIUM SLAG ...
Materiali in tehnologije / Materials and technology 55 (2021) 2, 171–181 181