Journal of Advanced Concrete Technology Vol.

20, 277-286, April 2022 / Copyright © 2022 Japan Concrete Institute 277

Technical report

Hydration Characteristics and Mechanic Properties of Basic Magnesium

Sulfate Cement Containing Steel Slag
Zhengyi Huang1, Lei Huang2* and Xuan He3

Received 6 October 2021, accepted 18 March 2022 doi:10.3151/jact.20.277

Abstract
The hydration characteristics and mechanical properties of basic magnesium sulfate (BMS) cement containing steel slag
(SS) are studied in the work. The heat release of the composite binders during hydration process was monitored, the
phase compositions of hydration products were analyzed, the microstructures of hardened pastes were investigated,
moreover, the compressive strength was tested. The results show that the hydration exothermic rate of BMS is slowed
and the cumulative heat of BMS is suppressed by SS addition. The incorporation of SS leads to the formation of C-S-H,
CaSO4·2H2O and Mg(OH)2, and suppresses the formation of 5-1-7 phase. Compared to plain BMS, the pore structure of
the composite paste with SS addition is much coarser, the crystallization and growth of Mg(OH)2 lead to the occurrence
of microcrack and large pores of the paste. The mechanical property of BMS is decreased by SS addition, especially
when the content of SS addition reaches 20%, the compressive strength is significantly weakened in all tested ages.

1. Introduction into the liquid solution, a hydration shell around MgO

grain with positive charge formed in early stage. In the
Basic magnesium sulfate (BMS) cement is a kind of absence of chemical additives, this hydration shell could
modified magnesium oxysulfate cement, which is made react with the dissolved OH- soon and therefore lead to
of active magnesium oxide, magnesium sulfate, water the formation of Mg(OH)2. In the presence of chemical
and admixture (Walling et al. 2016). It is concerned by additives, the negative charged ions dissolved from the
researchers and engineers for its excellent mechanical additives are absorbed on the shell, which can reduce
properties (Runčevski et al. 2013), such as high com- the interface energy of the shell and suppress the reac-
pression resistance, high tensile strength and high tion of the shell and the OH- in solution, the formation
toughness (Zeng et al. 2017; Zhang et al. 2019). In ad- of Mg(OH)2 is restrained. The ions dissolved into solu-
dition to mechanical properties, BMS is also a new con- tion is continuously accumulated until the solution is
struction material of potential use for its characters of over saturated with respect to 5MgSO4·Mg(OH)2·7H2O
light weight, low thermal expansion, high fire resistance (5-1-7 phase), the main hydration products phase of
etc. BMS.
However, there are also some unfavorable characters Many researchers focus on the modification of BMS
which limit the application of BMS, including the water properties and the reduction of its raw material cost
resistance and the high cost of the raw material. Chemi- through the incorporation of supplementary cementi-
cal additives are the first choice to modify the properties tious material (SCM). As a kind of excellent SCM used
of BMS. Up to now, kinds of chemical additives have in Portland cement, fly ash has been incorporated into
been proven to be effective for the modification of the BMS to replace part of the light-burned magnesia
properties of BMS, such as sodium citrate, citric acid, (LBM), the general source of active magnesium oxide.
sodium silicate, tartaric acid, phosphates, phosphoric It has been reported that a high dosage incorporation of
acid etc. (Wang et al. 2018; Wu et al. 2014, 2015). The fly ash could lead to a prominent decrease in compres-
mechanism of the chemical additives effect has been sive of BMS (Wu et al. 2014), some other researchers
thoroughly revealed (Wu et al. 2017), when the active found that the effect of fly ash on the properties of BMS
MgO is immersed into MgSO4 solution, OH- is released was determined by the calcium content of fly ash
(Zhang et al. 2019, 2020), the incorporation of
low-calcium fly ash can optimize the pore structure and
1 improve the water resistance of BMS, whereas, due to
Postgraduate student, School of Civil Engineering, the formation of CaSO4 2H2O, the addition of
Qinghai University, Xining 810016, China. high-calcium fly ash can lead to a compressive strength
2
Associate Professor, School of Civil Engineering, loss and a weaker water resistance.
Qinghai University and Qinghai Provincial Key The hydration process and properties of BMS with
Laboratory of Energy-saving Building Materials and some other SCM also have been studied. It has been
Engineering Safety, Xining 810016, China. found the incorporation of raw rice husk has little in-
*Corresponding author, E-mail: leihyx@foxmail.com fluence on the hydration products of BMS (Qin et al.
3
Postgraduate student, School of Civil Engineering, 2018), the interfacial transition zone between the rice
Qinghai University, Xining 810016, China.
Z. Huang, L. Huang and X. He / Journal of Advanced Concrete Technology Vol. 20, 277-286, 2022 278

Table 1 Chemical composition of LBM and SS (%).

Composition MgO SiO2 Fe2O3 CaO Al2O3 K2O Na2O SO3 MnO Total
LBM 83.10 6.12 3.01 5.93 0.79 0.07 0 0.23 0.21 99.46
SS 11.05 20.06 16.24 41.76 8.78 0.52 0.43 1.16 0 100

husk and matrix was improved. The incorporations of composite cementitious material, the mechanical prop-
pulverized fuel ash and incinerated were found to erties of the composite cementitious material were
change the needle-like morphology of 5-1-7 into a tested, the hydration process was monitored, the phase
lath-like structure (He et al. 2020), the change resulted compositions and the microstructures of the hydration
in the improvement of its water resistance. Circulating products were investigated to explore the mechanism
fluidized bed combustion ash addition was found to be behind.
beneficial to compressive strength of BMS at low dos-
age (Xu et al. 2020), the dosage of 20% was the optimal 2. Materials and methods
dosage for the mechanical property. The incorporation
of silicon fume can extend the induction and accelera- 2.1 Raw materials
tion period of BMS and therefore delay its setting (Tan The chemical composition of LBM and SS used in the
et al. 2021), due to the filling effect, the incorporation of study were analyzed by X-ray fluorescence and shown
silicon fume could also increase the compressive in Table 1, the active content of LBM is 66.24%, tested
strength of BMS, especially when the dosage was 10% according to the standard hydration method (Chau et al.
the compressive strength reached the highest level, 2008). Accurately weigh 2.0 g of LBM sample and
moreover, magnesium silicate hydrate (M-S-H) gel was place it into a 25 × 40 mm glass weighing bottle, the
found to form in the system. The formation of M-S-H sample was added with 20 ml distilled water and hy-
gel was also confirmed by other researcher (Li et al. drated for 24 hours in the bottle under the conditions of
2020), it was proposed that the pozzolanic reaction led 20 ± 2°C and 70 ± 5% relative humidity. Then the bottle
by SCM added to BMS resulted in the formation of was placed into an oven and the temperature was kept at
M-S-H, which could be further accelerated by CO2 cur- 100°C. After 24 h, the sample was dried to constant
ing, the coupling utilization of CO2 curing and poz- weight at 150°C. Afterwards, the sample was cooled to
zolanic materials could enhance the water resistance of room temperature in a dryer and weighed, the active
BMS. MgO content in the sample could be calculated accord-
Similar to SCM mentioned above, steel slag (SS) is a ing to Eq. (1).
by-product of steelmaking, statistics show that annual
discharge of steel slag in the world is more than 200 W2 − W1
W= × 100% (1)
million tons (Zhuang et al. 2021), the utilization rate of 0.45W1
SS is less than 20%. Many studies have focused on the
utilization for SS as SCM in Ordinary Portland Cement where W is the content of active MgO in LBM (%); W1
(OPC), its effects on hydration progress (Wang et al. is the mass of LBM sample (g); W2 is the mass of the
2011; Han et al. 2015), setting behavior (Zhuang et al. hydrated LBM sample (g); 0.45 is the molecular weight
2021), mechanical property (Sun et al. 2020) and the ratio of H2O to MgO.
mechanisms behind have been well revealed. Consider- The XRD patterns of LBM and SS are shown in Fig.
ing the mineral composition of SS (mainly calcium sili- 1. The particle size distribution of the LBM and the SS
cates), the incorporation of SS into BMS is desirable to were analyzed by a laser particle sizer and are shown in
low its cost and achieve energy savings. Fig. 2. Analytic grade MgSO4·7H2O and critic acid were
In this work, SS was incorporated into BMS to form a used in the study.

1 1 − Porlandite 2 − Quartz 3− C2S 4− C3S 5-Wustite

1−Periclase 2−Magnesite
6 −Calcite 7 −Periclase 8 −Feroxyhite 9 −C2F
3−Kieserite 10 −Todorokite 11 −Forsterite
1
12−Magmesium Iron Oxide

2
1 3+4
2 66 7
5 6
2 12 127
3 1 10 9 11 3 33 1 8
2

10 20 30 2θ / °40 50 60 10 20 30 40 50 60
2θ / °
(a) LBM (b) SS
Fig. 1 XRD patterns of LBM and SS.
Z. Huang, L. Huang and X. He / Journal of Advanced Concrete Technology Vol. 20, 277-286, 2022 279

Table 2 Mix proportions (g). recordation was conducted for 120 hours continuously.
Samples MgO MgSO4·7H2O H2O Steel slag Citric acid Before testing, the calorimeter was equilibrated at 20°C
BMS 200 227.1 0 0 1 for 6 hours.
SS5 190 215.7 5.54 15.8 0.95 When curing time came to the scheduled ages the
SS10 180 204.3 11.0 31.6 0.90 prepared pastes were crushed and dried by a vacuum
SS20 160 181.6 22.1 63.2 0.80 drying oven at 105°C for 6 h, then the dried pastes were
ground for the testing of XRD analysis and TG analysis.
2.2 Sample preparation XRD patterns of the samples were collected by an X-ray
The molar ratio of MgO, MgSO4 and H2O was fixed at diffractometer (D/max-2500PC, Rigaku Corporation,
7:1:20, the replacement levels of the LBM were 5%, Tokyo, Japan) with an acceleration voltage of 30 KV,
10% and 20% respectively. the dosage of citric acid the 2θ was in the range of 5 to 65° and the scanning
added was 0.5% by weight of the light-burned MgO. speed was kept as 6°/min. A thermal gravimetric ana-
The exact mix proportions are shown in Table 2. lyzer (HCT-1, HENVE, Beijing, China) was used to
MgSO4·7H2O was dissolved in deionized water firstly determine the thermal properties of the samples, the
to obtain MgSO4 solution, the molar ratio of heating rate was 5°C/min and the samples were heated
H2O/MgSO4 was 20:1, then citric acid was added into from 30°C to 1050°C. The pore distribution of the sam-
the solution, after that, pre-weighted powder was mixed ples was measured through MIP method with an auto-
together with the prepared solution to form BMS ce- matic mercury porosimeter (AutoPore IV 9500, Micro-
ment, the obtained paste was casted into steel mold with meritics, Norcross, Georgia, USA). The microstructures
the size of 20 × 20 × 20 mm3, cured at 20 ± 2°C and of the samples were observed and analyzed by
95% relative humidity for 1 day. After demolding, the SEM-EDS. (JSM-6610LV, JEOL Ltd., Tokyo, Japan).
specimens were cured in the same condition till the The compressive strength of the samples were tested at
scheduled ages (t = 1, 3, 7 and 28 days). 1, 3, 7 and 28 days, respectively.

2.3 Test methods 3. Results and discussion

An isothermal calorimeter (Calmetrix-4000HPC) was
used to monitor the hydration heat evolution rate and 3.1 Hydration process
the cumulative hydration of the composite pastes, the It has been proposed that the hydration of BMS could be
divided into five stages (Wu et al. 2017; Miao et al.
2020), including pre-induction, induction, acceleration,
SS deceleration and stable period, two heat evolution peaks
4
LBM present in exothermic rate curves. The first heat evolu-
tion peak results from the hydrolysis reaction of MgO,
3 the second one could be attributed to the reaction be-
Volume / %

tween the ions in solution, including Mg2+, SO42-, OH-

2 etc. to form 5-1-7 phase. Figure 3(a) shows the exo-
thermic rate of the composite binders, it can be seen that
1 there are five stages and two heat evolution peaks in the
curves, in line with the findings of other researchers.
0 The lengths of the induction periods and the heights of
the two heat evolution peaks are obviously altered by
0.1 1 10 100 SS addition. Compared to BMS, the heights of the first
Particle size / μm
Fig. 2 Particle size distribution of LBM and SS.
heat evolution peaks are increased 1.18 mw/g, 4.01

12 BMS
2.5 BMS 250
10 SS5 SS5
SS10
Exothermic rate / (mw/g)

SS10 200
Cumulative heat / (J/g)

2.0 8
SS20
SS20
6
1.5 150
4

1.0 2 100
0
0.5 0.00 0.25 0.50 50

0.0 0
0 24 48 72 96 120 0 24 48 72 96 120
Time / h Time / h
(a) (b)
Fig. 3 Exothermic rate and cumulative heat of BMS and the composite binders containing SS.
Z. Huang, L. Huang and X. He / Journal of Advanced Concrete Technology Vol. 20, 277-286, 2022 280

mw/g and 5.26 mw/g for the composite binders with 5%, binders with SS addition.
10% and 20% SS additions respectively, indicating the
hydrolysis reaction of MgO is promoted by SS addition. 3.2 Hydration products analysis
Moreover, the induction periods are prolonged by SS 5-1-7 phase in the hydration products of BMS is the
addition, the occurrences of the second heat evolution main carrier of its mechanical properties, while the for-
peaks of the composite binders with 5%, 10% and 20% mation of Mg(OH)2 is thought to be harmful to the me-
SS are delayed 1.31, 2.98 and 6.29 hours respectively, chanical properties of BMS, because its formation is
meanwhile, the heights of the second heat evolution magnesium-consumed and restrain the formation of
peaks decline with the increase in the amount of SS ad- 5-1-7 phase. As shown in Fig. 4, the diffraction peaks of
dition. 5-1-7 phase are present in the patterns of BMS cement
5-1-7 phase crystallizes and grows at the solid-liquid in 3 days, another prominent diffraction peaks in the
interface during induction period (Wu et al. 2017), be- patterns of BMS is MgO, even when the curing age
cause the nucleation of 5-1-7 phase at the begging of comes to 28 days, the diffraction peaks of MgO still
this stage is limited, which cannot provide enough space exist, meaning that part of MgO in BMS have not been
for the growth of 5-1-7 phase, the hydration rate of depleted. As introduced in Section 2.1, the active con-
BMS keeps at a low level in the induction period. When tent of LBM used in the study is 66.24%, the remainder
a sufficient number of 5-1-7 nucleus are accumulated, 33.76% MgO in LBM is inactive. Thus, the diffraction
which can support a rapid growth of the hydrates, the peaks of MgO in the patterns are mainly due to the dif-
induction period finishes and the hydration process goes fraction of the inactive MgO, besides, it could also be
into the acceleration period. The prolonged induction due to the diffraction of the residual active MgO in
periods of the composite binders led by SS addition, BMS. Moreover, the diffraction peaks of MgCO3 are
shown in Fig. 3(a), result from the suppression on the shown in the patterns.
crystallization of 5-1-7 phase. The content of active The diffraction peak intensity of 5-1-7 phase is
MgO in the samples with SS addition is lower than that weakened by SS addition in the patterns of the samples
of BMS, which could lead to a lower concentration of both in 3 and 28 days, resulting from the decrease in
ions (including Mg2+ and OH-) resulting in a lower su- active MgO content, whereas the diffraction peak inten-
persaturation with respect to 5-1-7 phase, and therefore sity of MgCO3 is enhanced by the SS addition both in 3
the formation of 5-1-7 phase is restrained. Besides, the and 28 days, which might result from the increase in
nucleation of 5-1-7 phase tends to occur on the carbonate led by SS addition. Moreover, the SS addition
solid-liquid surface and prior to forming on the surface leads to the formation of other hydrates, the diffraction
of MgO grain compared to that of SS, the SS addition peak of C-S-H is shown in the patterns, indicating that
leads to a decrease in the interface favoring the nuclea- the SS added into BMS is hydrated in 3 days. Another
tion of 5-1-7 phase, which also suppresses formation of hydration product should be noted is CaSO4·2H2O, it
5-1-7 crystal nucleus, thus, the induction period of the can be seen both in Figs. 4(a) and 4(b) that the diffrac-
composite binders is prolonged by SS addition. When tion peaks intensity of CaSO4·2H2O are enhanced by SS
the hydration process goes into acceleration period, the addition.
growth rate of the 5-1-7 phase is mainly determined by The mineral phases of SS include C3S, C2S, Portland-
the dissolution rate of MgO, the higher active MgO ite etc., and C3S is the most reactive phase among them
content could account for the higher second heat evolu- (Sun et al. 2020). With the progress of hydration, the
tion peaks. dissolution of C3S leads to the release of Ca2+ and SiO32-
Due to the decrease in heat release rate in early age, into pore solution. The solution of BMS is rich in SO42-
the cumulative hydration heat of the composite binders and Mg2+ (Wu et al. 2020), when Ca2+ concentration
is suppressed by SS addition. As shown in Fig. 3(b), the reaches to a certain point, Ca2+ and SO42- would be
cumulative hydration heat of the composite binders de- consumed to form CaSO4·2H2O. Thus, the molar ratio
creases from 196.11 to 135.85 J/g when the dosage of of magnesium to sulfur of the solution is increased,
SS increases from 0 to 20% at the age of 48 hours. With which is disadvantageous to the formation of 5-1-7
the prolong of curing age, the gap of the cumulative phase and results in the formation of Mg(OH)2 , as
hydration between the samples reduces, which de- shown in Figs. 4(a) and 4(b).
creases from 60.26 at 48 hours to 36.44 J/g at 120 hours, In addition to the formation of CaSO4·2H2O, Ca2+
meaning that the restrained effect led by SS addition on was also consumed by the formation of C-S-H when the
hydration is weakened. Compared to OPC, the reactivity content of Ca2+ and SiO32- dissolved in solution reach to
of SS is much slower, which often exerts an inhibition the certain supersaturation with respect to C-S-H. It has
effect on the early age hydration of cement (Zhuang et been found that C-S-H would be eroded by magnesium,
al. 2021). the decrease in cumulative hydration heat led when C-S-H exposes to a solution rich in Mg2+, Ca2+ in
by SS addition obviously due to the low reactivity of SS. C-S-H could be substituted by Mg2+ (Huang et al. 2021),
With the increase in curing age, the hydration of SS leading to the transformation of C- S-H to M-S-H, a
progress gradually, compensating the gap of the cumu- non-cohesive magnesium silicate hydrate. Thus, it can
lative hydration heat between BMS and the composite be speculated that the formation of C-S-H in the com-
Z. Huang, L. Huang and X. He / Journal of Advanced Concrete Technology Vol. 20, 277-286, 2022 281

posite system is temporality, with the progress of hydra- (lose 3H2O), the loss of free water and the dehydration
tion, the formed C-S-H would decalcify which might of CaSO4·2H2O (lose 1.5H2O) also contribute to the
cause the degradation of mechanical property for the mass loss in this temperature interval. The reactions
samples with SS addition in long ages. resulting the occurrences of P2, P3, P4, P5 and P6 are
The mass loss curves of BMS cement paste with dif- shown in Eqs. (2), (3), (4), (5) and (6), respectively.
ferent SS content with respect to the change of tem-
5Mg(OH) 2 ⋅ MgSO 4 ⋅ 4H 2 O(s)
perature are shown in Fig. 5. Five peaks are present in (2)
the DTG curves, P1 which occurs below 100°C has little ⎯⎯⎯
146 ℃
→ 5Mg(OH) 2 ⋅ MgSO 4 + 4H 2 O
indication on the composition of hydration products,
because in addition to the dehydration of 5-1-7 phase

∇ − 5· 1· 7 phase ⊗ − MgCO3 ⊕ − MgO a ♣− CaSO4•2H2O b ♥ − C-S-H c ◊− Mg(OH)2

♥ ◊
⊕ ♣
∇ ⊗
∇ a b c
∇ SS20
∇

SS10

SS5

BMS

10 20 30 40 50 60 70 11 12 13 28 29 30 38 39 40 41
2θ / °
(a) Aged 3 days
∇ − 5-1-7 phase ⊗ − MgCO3 ⊕ − MgO a ♣− CaSO4•2H2O b ♥ − C-S-H c◊ ◊− Mg(OH)2

⊗ ♣
♥
∇ ⊕
∇ a b c
∇ SS20
∇

SS10

SS5

BMS

10 20 30 40 50 60 70 11 12 13 28 29 30 38 39 40 41
2θ / °
(b) Aged 28 days
Fig. 4 XRD patterns of BMS and the composite pastes containing SS.

100 100 BMS SS5

BMS SS5
SS10 SS20
SS10 SS20
98 98
Weight / %
Weight / %

96 96

94 94

92 92
Deriv weight / (%/°C)
Deriv weight / (%/°C)

0.0 0.0
P3 P5 P5
-0.2 P4 -0.2 P3
P1 P4
P2 P6 P6
-0.4 -0.4 P1
P2
200 400 600 800 1000 200 400 600 800 1000
Temperature / °C Temperature / °C
(a) Aged 3 days (b) Aged 28 days
Fig. 5 TG and DTG curves of BMS and the composite pastes containing SS.
Z. Huang, L. Huang and X. He / Journal of Advanced Concrete Technology Vol. 20, 277-286, 2022 282

CaSO 4 ⋅ 0.5H 2 O(s) ⎯⎯⎯

360 ℃
→ CaSO 4 + 0.5H 2 O (3) It can be seen in Fig. 6(a) that the samples with differ-
ent SS content have little difference in the cumulative
5Mg(OH) 2 ⋅ MgSO 4 (s) pore volume in 3 days, which are between 0.31 and
(4)
⎯⎯ ⎯
420 ℃
→ 5MgO(s) + 5H 2 O + MgSO 4 (s) 0.325 ml/g. Whereas the characteristics of pore size
distribution are quite different, as shown in Fig. 6(b),
MgCO3 (s) ⎯⎯⎯
550°C
→ MgO(s) + CO2 (5) the most probable pore size of SS20 is the largest,
greater than 100 nm, the other three samples have the
MgSO 4 (s) ⎯⎯⎯
950 ℃
→ MgO(s) + SO3 (6) similar size of the most probable pore which are around
100 nm. Based on the size of pore diameter, the pores
As shown in Fig. 5, the intensity of P2 and P4 both in existing in cement paste could be divided into three
3 and 28 days decrease with the increase in SS content, types (Wu et al. 2021): gel pores (< 20 nm), small inter-
which means less bounder water loss for the samples granular pores (20 to 100 nm) and large intergranular
containing SS compared to BMS, confirming that the pores (> 100 nm). The majority of the pores formed in
incorporation of SS lead to a decrease in 5-1-7 forma- SS20 paste belong to large intergranular pores, com-
tion, in line with the results of XRD analysis. It should pared to SS20, the large intergranular pores formed in
be noted that P3 does not occur in DTG curves of BMS, SS10 paste are a small part, the pores formed in SS5 and
indicating there is no CaSO4·2H2O formed in the hydra- BMS pastes are mainly small intergranular pores. Then
tion products of BMS. Similar to the variation trend of it can be obtained that the SS addition promote the for-
P2 and P4, the intensity of P6 both in 3 and 28 days are mation of large intergranular pores. The intergranular
weakened by SS addition, because MgSO4, the desul- pores are formed by the interlock and crisscross of the
phurization of which accounts for P6, originates from reactants and hydration products (Wu et al. 2021), the
the thermal decomposition of 5-1-7 phase, the variation lager hydration products lead to the formation of lager
trend of P6 furtherly confirm that there is less 5-1-7 intergranular pores. It has been found in Fig. 3(a) that
phase formed in SS containing samples. the SS addition promote the formation of CaSO4·2H2O,
MgCO3 and Mg(OH)2, all of which have larger crystal
3.3 Pore structure size compared to 5-1-7 phase, so the pore structure of
The results of MIP analysis are shown in Figs. 6 and 7. the pastes with high SS addition is coarser than that of

0.35 BMS
Cumulative pore volume / (ml/g)

BMS
Incremental pore volume / (ml/g)

0.05 SS5
0.30 SS5
SS10 SS10
0.25 SS20 0.04 SS20

0.20 0.03
0.15
0.02
0.10
0.01
0.05
0.00 0.00
1 2 3 4 1 2 3 4
10 10 10 10 10 10 10 10
Pore diameter / nm Pore diameter / nm
(a) (b)
Fig. 6 Pore structure of BMS and the composite pastes containing SS, aged 3 days.

0.35 BMS BMS

Cumulative pore volume / (ml/g)

Incremental pore volume / (ml/g)

SS5 0.05 SS5

0.30
SS10 SS10
0.25 SS20 0.04 SS20

0.20 0.03
0.15
0.02
0.10
0.01
0.05
0.00 0.00
1 2 3 4 1 2 3 4
10 10 10 10 10 10 10 10
Pore diameter / nm Pore diameter / nm
(a) (b)
Fig. 7 Pore structure of BMS and the composite pastes containing SS, aged 28 days.
Z. Huang, L. Huang and X. He / Journal of Advanced Concrete Technology Vol. 20, 277-286, 2022 283

Table 3 Results of EDS analysis. formed pore structure, so the pores already formed in
Weight percentage Atomic percentage pastes with the high content of SS addition is getting
Region (%) (%) larger.
Mg O Mg O
1 39.7 60.4 30.2 69.8 3.4 Morphology of the hydration products
2 39.9 60.1 30.4 69.6 The morphology of the hydration products of BMS and
3 37.2 62.8 28.1 71.9 the sample with 20% SS additions in 3 days are shown
in Fig. 8. For BMS, the typical need-like 5-1-7 phase is
BMS pastes. formed, it can be seen that the 5-1-7 phase formed is
With the prolong of curing age, the gaps between the compact and homogeneous. Compared to BMS, the
cumulative pore volume of the samples are enlarged. As 5-1-7 phase formed in SS20 is sparse. Besides, some
shown in Fig. 7(a), the cumulative pore volume of BMS other flocculent aggregates and big crystal products are
is 0.26 ml/g in 28 days, whereas the cumulative pore formed [region 1 in Fig. 8(b)]. Taking into the mor-
volume of SS20 is higher than 0.3 ml/g, indicating the phology characters and the results of XRD analysis, the
refining of the pore structures of the BMS paste is sup- flocculent aggregates could be C-S-H. Based on the
pressed by SS addition. The most probable pore size of EDS analysis, shown in Table 3, the crystal products
BMS, SS5 and SS10 paste has no obvious change, could be identified as Mg(OH)2.
shown in Fig. 7(b), whereas the pores around 100 nm The morphology of BMS in 3 and 28 days are shown
increase. It might be due to the continuous consumption in Fig. 9, it can be seen that the matrix of BMS in 3
of MgO for hydrating, the spaces initially occupied by days is very compact, the amount of the large pores in
MgO cannot be fully fill by the new formed hydrates. the matrix in 28 days is more than that of the matrix in 3
The small pores, including some small intergranular days, consistent with the results of MIP analysis, the
pores and gel pores, formed in 3 days disappear when new formed 5-1-7 phase can’t fully fill the large pores
the age comes to 28 days. The needle-like 5-1-7 phase formed in early age [Fig. 9(c)]. The matrix of SS20 in
accounts for this filling effect. Besides, the C-S-H 28 days is porous compared to the matrix of SS20 in 3
formed in the pastes with high content of SS addition days. As shown in Fig. 10(c), tabular crystals (regions 1
might also contribute to this filling effect. and 2) are formed in the intervals between the hydration
The most probable pore size of SS20 is getting larger products of SS20, which has been verified as Mg(OH)2
when the age comes to 28 days. For OPC, the pore size based on EDS analysis shown in Table 3. Besides, large
is getting smaller and the pore structure tends to become cracks related to the formation of Mg(OH)2, are also
denser with the continuous hydration of cement (Han et observed in the hydration products of SS20.
al. 2018; Zhang et al. 2016), because the large pores
formed in early age are partly filled by the new hydra- 3.5 Mechanical properties
tion products, mainly C-S-H. The enlarged pores in The compressive strengths of BMS containing SS pastes
SS20 pastes may result from the formation of Mg(OH)2. in different ages are shown in Fig. 11. In 1 day, the
In OPC, MgO is regarded as a harmful component for compressive strength of BMS is higher than 16 MPa,
its expansion character and thus be constrained (Cao et 5% SS addition has not significantly changed the com-
al. 2018), the expansion character of MgO comes from pressive strength of the pastes, when 10% and 20% SS
the crystallization and growth pressure of Mg(OH)2, the are added into the pastes, compared to BMS, the com-
hydration product of MgO, some researchers use this pressive strengths decrease 28.9% and 63.1% respec-
expansion character of MgO to compensate the shrink tively. The addition of SS exerts the similar effect on the
of OPC (Cao et al. 2021). It can be inferred that the strength development in different ages. In 28 days, the
crystallizations and growths of Mg(OH)2 exert pressures compressive strength of BMS paste reaches to 75 MPa,
on the nearby pore walls, which could damage the 5%, 10% and 20% SS addition decrease the compres-

(a) (b)
Fig. 8 Morphology of the main hydration products for BMS and SS20, aged 3 days.
Z. Huang, L. Huang and X. He / Journal of Advanced Concrete Technology Vol. 20, 277-286, 2022 284

sive strength 11.8%, 19.6% and 44.2% respectively. cal properties of the BMS pastes is significantly de-
The MIP analysis has verified that SS addition leads creased by SS addition. In addition to the pore structure,
to the formation of a coarser pore structure which is the type, content and crystallinity the of hydration
more fragile and vulnerable to damage, so the mechani- products also affect the mechanical properties of BMS

(a) Aged 3 days (Magnification: 500)

(b) Aged 28 days (Magnification: 500) (c) Aged 28 days (Magnification: 2000)
Fig. 9 The matrix of BMS in different ages.

(a) Aged 3 days (Magnification: 500)

(b) Aged 28 days (Magnification: 500) (c) Aged 28 days (Magnification: 2000)
Fig. 10 The matrix of SS20 in different ages.
Z. Huang, L. Huang and X. He / Journal of Advanced Concrete Technology Vol. 20, 277-286, 2022 285

80 1d 3d property taking into account the low reactivity of SS,

which need be further studied. Meanwhile, more prop-
Compressive strength / MPa

70 7d 28d
erties should be concerned, including water resistance
60 and volume stability to achieve a rational utilization of
SS in BMS.
50
40 Acknowledgments
The authors would like to acknowledge the financial
30 support of the Natural Science Foundation of China
20 (Grant No.52002202) and the Youth Science Research
Fund Project of Qinghai University (Grant No.
10 2020-QGY-7).
0
BMS SS5 SS10 SS20 References
Fig. 11 Compressive strengths of BMS and composite Cao, F. Z., Miao, M. and Yan, P. Y., (2018). “Hydration
pastes containing SS. characteristics and expansive mechanism of MgO
expansive agents.” Construction and Building Materials,
(Wu et al. 2021), in other words, the content and crys- 183, 234-242.
tallinity of 5-1-7 phase have a significant effect on the Cao, F. Z., Liu, Y. and Yan, P. Y., (2021). “Properties
mechanical properties of BMS. As shown in Fig. 8, the and mechanism of the compound MgO expansive
crystallinity of 5-1-7 phase has not been modified by SS agent (CMEA) produced by calcining the mixture of
addition obviously, however, with the increase in SS dolomite and serpentine tailings.” Construction and
addition the content of 5-1-7 phase formed in pastes Building Materials, 277, 122331.
decrease, which could also lead to the decrease in com- Chau, C. K. and Li, Z. J., (2008). “Accelerated reactivity
pressive strength. Then it is understandable that though assessment of light burnt magnesium oxide.” Journal
the pore structure of SS20 is getting coarser in 28 days of the American Ceramic Society, 91(5), 1640-1645.
compared to that of in 3 days, the compressive strength Han, F. H. and Zhang, Z. Q., (2018). “Properties of
of SS20 still increases, it is attributed to the continuous 5-year-old concrete containing steel slag powder.”
formation of 5-1-7 phase. Powder Technology, 334, 27-35.
Han, F. H., Zhang, Z. Q., Wang, D. M. and Yan, P. Y.,
4. Conclusions (2015). “Hydration heat evolution and kinetics of
blended cement containing steel slag at different
Based on the results of isothermal calorimetry, XRD, temperatures.” Thermochimica Acta, 605,43-51.
TG, MIT, SEM-EDS analysis and the mechanical testing. He, P. P., Poon, C. S., Richardson, I. G. and Tsang, D. C.
The following conclusions can be drawn: W., (2020). “The mechanism of supplementary
(1) The hydration exothermic rate of BMS is slowed cementitious materials enhancing the water resistance
down and the cumulative heat of BMS is suppressed of magnesium oxychloride cement (MOC): A
by SS addition, 5-1-7 phase formed in paste is de- comparison between pulverized fuel ash and
creased by SS addition. incinerated sewage sludge ash.” Cement and
(2) The incorporation of SS which is mainly composed Concrete Composites, 109, 103562.
of portlandite and C2S introduces a generous amount Huang, Q., Zhu, X. H., Xiong, G. Q., Zhang, M. T.,
of calcium to BMS, and therefore change the com- Deng, J. X., Zhao, M. and Zhao, L., (2021). “Will the
position of hydration products significantly. The in- magnesium sulfate attack of cement mortars always
troduction of calcium leads to the formation of be inhibited by incorporating nanosilica?” Construction
C-S-H and CaSO4·2H2O directly and the formation and Building Materials, 305, 124695.
of Mg(OH)2 indirectly. Li, Q. Y., Zhang, L. C., Gao, X. J. and Zhang, J. Y.,
(3) Compared to plain BMS, the pore structure of the (2020). “Effect of pulverized fuel ash, ground granulated
paste with SS addition is much coarser, the crystal- blast-furnace slag and CO2 curing on performance of
lization and growth of Mg(OH)2 led to the occur- magnesium oxysulfate cement.” Construction and
rence of microcrack and large pores of the paste. Building Materials, 230, 116990.
(4) The mechanical property of BMS is decreased by Miao, M., Wu, C. Y., Tan, Y. S. and Yu, H. F., (2020).
the SS addition, especially when the content of SS “Mechanistic study of the effects of magnesia
addition reaches 20%, the compressive strength is reactivity on setting and hardening of basic
significantly weakened in all tested ages. magnesium sulfate cement.” Journal of Advanced
Based on the above work, the mechanical perform- Concrete Technology, 18(11), 678-688.
ance of BMS with a moderate dosage (5%, and 10% in Qin, L., Gao, X. J. and Chen, T. F., (2018). “Recycling
the study) of SS addition is acceptable, a high tempera- of raw rice husk to manufacture magnesium oxysulfate
ture curing might be helpful to improve its mechanical cement based lightweight building materials.”
Z. Huang, L. Huang and X. He / Journal of Advanced Concrete Technology Vol. 20, 277-286, 2022 286

Journal of Cleaner Production, 191, 220-232. Wu, C. Y., Yu, H. F., Dong, J. M. and Zheng, L. N.,
Runčevski, T., Wu, C. Y., Yu, H. F., Yang, B. and (2014). “Effects of material ratio, fly ash, and citric
Dinnebier, R. E., (2013). “Structural characterization acid on magnesium oxysulfate cement.” ACI Materials
of a new magnesium oxysulfate hydrate cement phase Journal, 111(3), 291-297.
and its surface reactions with atmospheric carbon Wu, C. Y., Yu, H. F., Zhang, H. F. Dong, J. M., Wen, J.
dioxide.” Journal of American Ceramic Society, 96, and Tan, Y. S., (2015). “Effects of phosphoric acid
3609-3616. and phosphates on magnesium oxysulfate cement.”
Sun, J. W., Zhang, Z. Q., Zhuang, S. Y. and He, W., Materials and Structures, 48(4) 907-917.
(2020). “Hydration properties and microstructure Xu, X., Hu, Z. and Duan L., (2020). “Investigation of
characteristics of alkali-activated steel slag.” high volume of CFBC ash on performance of basic
Construction and Building Materials, 241, 118141. magnesium sulfate cement.” Journal of Environmental
Tan, Y. S., Yu, H. F., Sun, S. K., Wu, C. Y. and Ding, H., Management, 256, 109878.
(2021). “Properties and microstructure of basic Zeng, X. C. and Yu, H. F., (2017). “Study on large
magnesium sulfate cement: Influence of silica fume.” eccentric compression column of basic magnesium
Construction and Building Materials, 266, 121076. sulfate cement concrete.” Journal of Harbin
Walling, S. A. and Provis, J. L., (2016). Engineering University, 38, 852-858.
“Magnesium-based cements: A journey of 150 years, Zhang, J. Y., Wen, Y. X. and Chen, L. D., (2019).
and cements for the future?” Chemical Reviews, “Fatigue properties of basic magnesium sulfate
116(7), 4170-4204. cement reinforced concrete beams: Based on response
Wang, N., Yu, H. F., Bi, W., Tan Y. S., Zhang, N., Wu, C. surface methodology.” Fresenius Environmental
Y., Ma, H. Y. and Hua, S., (2018). “Effects of sodium Bulletin, 28(10), 7655-7660.
citrate and citric acid on the properties of magnesium Zhang, N., Yu, H. F., Wang, N., Gong, W., Tan, Y. S. and
oxysulfate cement.” Construction and Building Wu, C. Y., (2019). “Effects of low- and high-calcium
Materials, 169, 697-704. fly ash on magnesium oxysulfate cement.” Construction
Wang, Q., Yan, P. Y. and Feng, J. W., (2011). “A and Building Materials, 215, 162-170.
discussion on improving hydration activity of steel Zhang, N., Yu, H. F., Gong, W., Liu, T., Wang, N., Tan,
slag by altering its mineral composition.” Journal of Y. S. and Wu, C. Y., (2020). “Effects of low- and
Hazardous Materials, 186, 1070-1075. high-calcium fly ash on the water resistance of
Wu, C. Y., Chen, W. H., Zhang, H. F., Yu, H. F., Zhang, magnesium oxysulfate cement.” Construction and
W. Y., Jiang, N. S. and Liu, L. X., (2017). “The Building Materials, 230, 116951.
hydration mechanism and performance of modified Zhang, Z. Q., Wang, Q. and Chen, H. H., (2016).
magnesium oxysulfate cement by tartaric acid.” “Properties of high-volume limestone powder concrete
Construction and Building Materials, 144, 516-524. under standard curing and steam-curing conditions.”
Wu, C. Y., Miao, M. and Yu, H. F., (2021). “Effect of Powder Technology, 301, 16-25.
MgO activity and ratio on the strength of basic Zhuang, S. Y. and Wang, Q., (2021). “Inhibition
magnesium sulfate cement and its mechanism.” mechanisms of steel slag on the early-age hydration
Journal of Building Materials [online]. Available of cement.” Cement and Concrete Research, 140,
from: <https://web02.cnki.net/kmobile/Journal/detail/ 106283.
DKCJ/JZCX20210305004>. (in Chinese)