applied

sciences
Article
Internal Curing Effect of Waste Glass Beads on High-Strength
Cement Composites
Sujeong Pyeon 1 , Gyuyong Kim 1 , Sangsoo Lee 2 and Jeongsoo Nam 1, *

1 Department of Architectural Engineering, Chungnam National University, 99 Daehak-ro, Yuseong-gu,

Daejeon 34134, Korea
2 Department of Architectural Engineering, Hanbat National University, 125 Dongseo-daero, Yuseong-gu,
Daejeon 34158, Korea
* Correspondence: j.nam@cnu.ac.kr

Abstract: High-strength concrete (HSC) uses binders and microfillers with ultrafine particles, such
as silica fume. The resulting dense internal hydration structure rapidly decreases HSC humidity,
causing shrinkage cracks and affecting internal hydration. Herein, the hydration degree inside high-
strength cement composites (HSCCs) was examined using waste glass beads (WGBs) as lightweight
aggregates (LWAs). Moreover, unreacted hydrate reduction and hydrate formation tendencies were
investigated. WGBs with particle sizes within 2.00–6.00 mm were added at ratios of 5%, 10%, and
20% after pre-wetting. The increased number of hydrates inside the specimens were examined under
steam curing (80 ◦ C) and room temperature curing (25 ◦ C). The strength decreased as the WGB
content increased. Thermogravimetric, X-ray diffraction, and Si nuclear magnetic resonance analyses
revealed that the hydration degree of Si inside HSCCs changed when the content of pre-wetted LWAs
changed. A visual inspection of the specimen cross-section and scanning electron microscopy–energy-
dispersive X-ray spectrometry (SEM–EDS) analysis revealed the moisture trapped inside WGB pores
and the hydration tendency. Under steam curing and room temperature curing, the paste contained
different amounts of hydrates, depending on WGB content. Moreover, water-absorbed WGBs were
Citation: Pyeon, S.; Kim, G.; Lee, S.;
continuously desorbed through SEM–EDS, and hydrates were present in WGB pores.
Nam, J. Internal Curing Effect of
Waste Glass Beads on High-Strength
Keywords: high-strength cement composites; densification; waste glass bead; Si hydration; correlation
Cement Composites. Appl. Sci. 2022,
analysis
12, 8385. https://doi.org/10.3390/
app12168385

Academic Editors: José A.F.O.

Correia and Daniel Dias 1. Introduction
Received: 7 July 2022 High-strength concrete (HSC) has been widely used as a construction material owing
Accepted: 18 August 2022 to its high workability, strength, and durability [1]. These properties result from using
Published: 22 August 2022 pozzolanic materials, such as silica fume, as supplementary cement materials. Although
Publisher’s Note: MDPI stays neutral
silica fume is expensive, it can improve the mechanical performance and durability of
with regard to jurisdictional claims in
concrete [2]. Silica fume comprises spherical particles with an average diameter of 0.03
published maps and institutional affil- to 0.1 µm, which is approximately 100-fold smaller than cement particles. This makes the
iations. specific surface area approximately 20,000 m2 /kg, which is 10 to 20-fold larger than that
of other pozzolanic materials [3]. Based on these properties, silica fume with ultrafine
particles and low density can contribute to the densification and subsequent strengthening
of concrete. For HSC and ultrahigh performance concrete (UHPC) subjected to high-density
Copyright: © 2022 by the authors. packing, their dense internal structures cause a sharp reduction in internal humidity, which
Licensee MDPI, Basel, Switzerland. may cause autogenous shrinkage. Thus, drying must be reduced to control autogenous
This article is an open access article shrinkage and cracking at early ages. However, the method of spreading external moisture
distributed under the terms and to the internal structure through conventional water curing or water spray curing for
conditions of the Creative Commons concrete is ineffective for supplying moisture into HSC and UHPC [4].
Attribution (CC BY) license (https://
Various studies have focused on reducing the shrinkage of HSC and UHPC with
creativecommons.org/licenses/by/
dense internal structures by improving internal hydration. Bentz et al. reported that the
4.0/).

Appl. Sci. 2022, 12, 8385. https://doi.org/10.3390/app12168385 https://www.mdpi.com/journal/applsci

Appl. Sci. 2022, 12, 8385 2 of 17

internal curing technique for concrete using porous lightweight aggregates (LWAs) has
various benefits, such as reducing shrinkage and cost [5]. In a related study, Justs et al.
increased the degree of hydration using a method similar to increasing the water-to-
cement ratio inside the matrix by a hydration process using superabsorbent polymers
(SAPs) and investigated the resultant internal hardening [6]. Mechtcherine et al. and
Mönnig et al. verified the mechanism through which SAP absorbs and releases moisture
inside concrete [7–9]. Rößler et al. applied the moisture absorption property of rice husk
ash to UHPC and confirmed its effect on improving strength [10]. Liu et al. applied porous
aggregates comprising calcined bauxite and basalt aggregates to UHPC and proposed a
method to improve matrix strength, excluding fibers [11]. Sun et al. investigated the effect
of moisture transport of porous aggregates on the shrinkage reduction and microstructure
optimization of UHPC and reported a method for designing a UHPC matrix using a high-
density particle packing model (Andreasen and Andersen model) [12]. To study hydration
inside UHPC with LWA for strength improvement, the hydrates generated inside specimens
were detected using thermogravimetric analysis (TGA) and scanning electron microscopy–
energy-dispersive X-ray spectrometry (SEM–EDS). Based on the aforementioned results,
many studies have improved the hydration degree inside UHPC using porous materials.
Most of the internal curing materials currently used in UHPC are porous. The addition
of SAPs and LWAs is the most commonly used method because these materials can signif-
icantly inhibit the increase in the autogenous shrinkage of UHPC and fully compensate
for the initial shrinkage of concrete [13–15]. However, SAPs and LWAs contain numerous
pores, which may degrade the mechanical properties and durability of UHPC by increasing
the number of pores. Therefore, pre-wetted LWAs can effectively minimize the extent of
drying and autogenous shrinkage of UHPC [16]. Bentz et al. reported that the moisture
introduced into concrete by LWA can inhibit internal shrinkage and cracking as well as
improve the hydration degree of the surrounding mortar, thereby improving the strength
of concrete [17–19]. Akcay et al. and Henkensiefken et al. analyzed the influence of LWA
particle diameter on the internal curing efficiency and reported that hydration increases
as diameter decreases [20–22]. Van et al. proposed a mechanism to improve mechanical
properties via the shrinkage reduction in concrete using rice husk ash as pre-wetted LWA
because it can delay the relative humidity reduction inside UHPC and promote the contin-
uous hydration of cement [23]. Wang et al. examined an effective method of mitigating the
autogenous shrinkage of UHPC using coral aggregates [24]. Yang et al. conducted research
to increase the internal relative humidity and decrease autogenous shrinkage by adding
zeolite into cement paste with a low moisture content [25]. The degree of hydration of
UHPC through the internal curing effect of LWA has been primarily analyzed using TGA,
X-ray diffraction (XRD), and SEM–EDS; however, studies on the hydration degree of the
individual elements in paste, such as Si, in high-strength cement composites (HSCCs), such
as UHPC, remain limited.
In previous studies [6–25], the hydration effect of various porous materials was ex-
amined to improve the internal hydration of HSC and UHPC. However, as the studies on
hydration analysis using LWA have not performed thermodynamic modeling based on the
hydration degree of certain elements formed in the HSCC paste, they have failed to identify
the hydration tendency by age. Therefore, this study examined whether non-hydrated reac-
tants inside HSCCs can be reduced using waste glass beads (WGBs) as LWAs based on the
internal hydration mechanism of high-density concrete. In addition, WGB was subjected
to pre-wetting to examine the improvement in internal hydration by water-containing ag-
gregates, and the hydration tendency by age was identified via thermodynamic modeling.
The specimens prepared by adding WGBs at ratios of 5%, 10%, and 20% were subjected to
steam curing and room temperature curing. For each test group, the internal hydration
effect of HSCCs was examined via uniaxial compression testing, XRD/Rietveld-based
quantitative phase analysis, TGA, solid-state 29 Si magic-angle spinning nuclear magnetic
resonance (MAS NMR), and thermodynamic modeling. In addition to understanding the
hydration level of the paste, the internal state of WGB pores was confirmed through a
 Appl. Sci. 2022, 12, x FOR PEER REVIEW 3 of 18

Appl. Sci. 2022, 12, 8385 3 of 17

to understanding the hydration level of the paste, the internal state of WGB pores was
confirmed through a visual inspection of the cross-section of a test specimen mixed with
visual
WGBsinspection of the
and through cross-section
SEM–EDS of a test specimen mixed with WGBs and through
analysis.
SEM–EDS analysis.
2. Materials and Methods
2. Materials and Methods
2.1. Experimental Procedure
2.1. Experimental Procedure
CEMType
CEM Type1 (SsangYong
1 (SsangYong C&E,
C&E, Seoul,
Seoul, Korea)
Korea) of ASTM
of ASTM C 150-85 C 150-85
and silicaandfumesilica fume
(Elkem
(Elkem Silicones)
Silicones) were used wereas used
binders as binders for the materials
for the materials used to used
prepareto prepare the specimens.
the specimens. Silica
Silica powder (SAC Materials, Pure silica S-SIL 30) was used as
powder (SAC Materials, Pure silica S-SIL 30) was used as a filler, and a polycarboxylate a filler, and a polycarbox-
ylate superplasticizer
superplasticizer for UHPC for UHPC
(FLOWMIX (FLOWMIX3000S) 3000S)
was used wasfor
used for effective
effective mixing.mixing.
WGB, which WGB,
which was used as an internal moisture supply medium to
was used as an internal moisture supply medium to reduce unreacted hydrates inside the reduce unreacted hydrates
inside the
matrix, hadmatrix,
a particlehaddiameter
a particleofdiameter
2.00 to 6.00of 2.00
mmto 6.00
and mmdensity
a dry and a dry density
of 0.89 g/cm of3 ,0.89
as
g/cm 3 , as shown in Figure 1. WGB is fabricated through a special
shown in Figure 1. WGB is fabricated through a special heat treatment process at 900 C or heat treatment process
◦
at 900 after
higher °C orthe
higher
waste after
glass theiswaste glass
collected is collected
and crushed intoandacrushed into a finer
finer powder. In this powder.
instance, In
this instance, the Si component of waste glass causes internal
the Si component of waste glass causes internal expansion, resulting in numerous pores expansion, resulting in nu-
merous
inside. pores
This inside. This
fabrication processfabrication processresources;
aims to recycle aims to recycle
the poresresources;
generated theinpores gener-
the process
atedtoinathe
lead process
higher water lead to a higher
absorption water
rate absorption
of 13.04%. WGBratewasofsoaked
13.04%. in WGB
waterwasfor 48soaked
h, and in
water
the for 48 h,surface
saturated and thedry saturated surface
(SSD) state wasdryused.
(SSD)Table
state was used.
1 lists the Table
physical1 lists
andthechemical
physical
and chemical
properties properties
of WGB. In thisof study,
WGB. In this study,
reinforcing reinforcing
fibers and silicafibers
sand,andwhich
silica sand,
may act whichas
may act as
variables, variables,
were excluded were excluded toobserve
to effectively effectively observe the
the formation of formation
hydration of hydration
products in
products in specimens as a function of curing time. Table 2 lists
specimens as a function of curing time. Table 2 lists the chemical compositions of the the chemical compositions
of the binders
binders used to used
prepare to prepare
specimens, specimens,
and Figure and Figure the
2 shows 2 shows
particle thesizes
particle
of raw sizes of raw
materials.
materials.
The particleThe particle distribution
distribution of the materialof theused
material used for
for HSCCs is HSCCs
presented is presented
to understandto under-the
stand theconstituting
materials materials constituting
the dense the dense microstructure
microstructure in HSCCs,inwhichHSCCs, which
exhibit exhibitinternal
a dense a dense
internal with
structure structure with nanostructured
nanostructured material. Table material. Table
2 lists the 2 listsresults
analysis the analysis
performed results
with per-
the
Bruker
formed(D8 ADVANCE)
with the Brukerequipment
(D8 ADVANCE) of Chungnam
equipmentNational University.
of Chungnam This study
National aimed
University.
to examine
This studythe aimedimprovement
to examine in internal curing by identifying
the improvement the changes
in internal curing in the tightly
by identifying the
closed
changes internal structure
in the tightly as ainternal
closed result ofstructure
mixing WGB withof
as a result the internal
mixing WGB structure
with the ininternal
which
several
structure nanosized
in whichmaterials exist.
several nanosized materials exist.

Figure1.1.Shape
Figure Shapeof
ofwaste
wasteglass
glassbeads
beads(WGBs).
(WGBs).

Table1.1.Physical
Table Physicalproperties
propertiesofofWGBs.
WGBs.

Saturated Density
Saturated Densityinin
Absorption
Absorption Loss
Lossonon Ig-
Fineness Stability
Fineness Stability Surface
Surface
DryDry Absolute
Absolute ClayLump
Clay Lump Chloride
Chloride
Rate
Rate Ignition
nition
Modulus
Modulus (%) (%) Density
Density Dry State
Dry State (%)
(%) (%)
(%)
(%)
(%) (%)
(%)
(g/cm3 ) 3 (g/cm3 )3
(g/cm ) (g/cm )
WGB 5.82 4.50 1.00 0.89 13.04 0.13 0.00 0.10
WGB 5.82 4.50 1.00 0.89 13.04 0.13 0.00 0.10

Table 2. Chemical composition of binders.

(wt%) CaO SiO2 Al2 O3 MgO Na2 O K2 O SO3 Fe2 O3 TiO2

Silica fume 0.02 99.0 0.30 0.01 0.05 0.02 0.05 0.05 0.06
White cement 66.79 16.14 4.40 2.47 0.08 1.18 3.23 3.83 0.34
Silica powder 0.05 99.0 0.30 0.05 0.05 0.05 0.05 0.05 0.06
 Table 2. Chemical composition of binders.

(wt%) CaO SiO2 Al2O3 MgO Na2O K2O SO3 Fe2O3 TiO
Silica fume 0.02 99.0 0.30 0.01 0.05 0.02 0.05 0.05 0.0
Appl. Sci. 2022, 12, 8385 White cement 66.79 16.14 4.40 2.47 0.08 1.18 3.23 4 3.83
of 17 0.3
Silica powder 0.05 99.0 0.30 0.05 0.05 0.05 0.05 0.05 0.0

Figure 2. ParticleFigure
size distribution anddistribution
2. Particle size cumulative and
passing rate of binders
cumulative passingand
rateWGBs.
of binders and WGBs.

Table 3 lists theTable

mix proportions
3 lists the mix of proportions
specimens. of Notations
specimens.are Notations
listed to identify
are listed each
to identify
specimen according to theaccording
specimen curing condition and mixing
to the curing ratioand
condition of WGB.
mixingWGB ratiomixing
of WGB. ratios
WGB of mixing r
5%, 10%, and 20% were selected to activate and investigate internal curing while minimizing
of 5%, 10%, and 20% were selected to activate and investigate internal curing while m
strength degradation
mizingthat occurs
strength when a large
degradation thatamount of lightweight
occurs when aggregate
a large amount is mixed. aggreg
of lightweight
This was carriedmixed.
out byThis
referring to a previous study [26]. A study conducted
was carried out by referring to a previous study [26]. A study by Elango et al.conducte
produced self-compacting concrete using diatomite aggregate and observed
Elango et al. produced self-compacting concrete using diatomite aggregate and obse significant
deterioration insignificant
the mechanical properties
deterioration in and durability when
the mechanical the aggregate
properties was replaced
and durability when the aggre
at 5%, 10%, 15%, andreplaced
was 20% levels.
at 5%, The binders
10%, 15%, were
and 20%subjected
levels. to
Theanbinders
anhydride weremix for 5 min.
subjected to an anhyd
Next, water andmix admixtures
for 5 min.were
Next,added; moreover,
water and additional
admixtures mixingmoreover,
were added; was performed for mixing
additional
2 min to ensureperformed
complete mixing.
for 2 minWGB to was thencomplete
ensure added and mixedWGB
mixing. for another 2 min.
was then Upon
added and mixe
completion of theanother 2 min. Upon completion of the mixing process, the fresh mixturemm
mixing process, the fresh mixture was poured into a Ø100 × 200 was poured
mold. A 50 × 50a Ø100
× 50 mm× 200test
mm piece
mold.wasA additionally
50 × 50 × 50 mm prepared to confirm
test piece the cross-section
was additionally prepared to con
of the specimen. The mold was sealed with a film and cured under
the cross-section of the specimen. The mold was sealed with a film natural conditions
and cured under
for 24 h. SP, S5,ural
S10, and S20 specimens
conditions for 24 h. SP,(hereinafter
S5, S10, and called “S-group(hereinafter
S20 specimens specimens”) were
called “S-group
subjected to steam curing ◦
at 80subjected
C for 72toh [27] and then curing at room
imens”) were steam curing at 80 °C for 72temperature
h [27] and then(Temp. curing at r
20 ◦ C, Hum. 55%), whereas(Temp.
temperature RP, R5,20 R10,
°C,and
Hum. R2055%),
specimens
whereas (hereinafter
RP, R5, R10, called
and R20“R-group
specimens (he
specimens”) were subjected to room temperature curing from the beginning.
after called “R-group specimens”) were subjected to room temperature curing from In this study,
steam curing was performed
beginning. to study
In this study,the activation
steam curing of was internal curing
performed of UHPC,
to study a high- of int
the activation
strength and high-performance concrete, which was compared with room temperature
curing. Two types of curing environments were applied to examine the applicability of PC
products in the actual field later.
Appl. Sci. 2022, 12, 8385 5 of 17

Table 3. Mix proportions of specimens (per cubic meter of specimens, kg/m3 ).

Mixture ID. Cement Silica Fume Silica Powder Water Superplasticizer WGB
SP 0
S5 21
S10 42
S20 81
322 106 128 304 20
RP 0
R5 21
R10 42
R20 81

2.2. Testing Methodology

For each specimen, the compressive strength was measured at the ages of 1, 3, 28,
and 56 d; the density was measured at the age of 28 d; and XRD, TGA, and MAS NMR
spectroscopy were conducted at the ages of 3 and 56 d. HSCC fluidity was measured in
the unhardened state before a specimen was placed in the mold. The flow table method
(ASTM C 1437) was used in which the mixture in a fresh state was placed on a flow table
and measured after the flow stopped. The density of HSCC specimens was measured
according to ASTM D 792. For compressive strength measurements, the compressive
strengths of three cylindrical specimens were measured, and the average was selected as
the compressive strength in accordance with ASTM C109/C109a.
Powder samples that passed through a 75 µm sieve were collected from the specimens
used for compressive strength and hydrate composition measurements. When WGB was
added, a 50 × 100 mm area was selected at the center of the Ø100 × 200 mm specimen,
and then paste powder samples were collected. A flake-type sample was taken from
the specimen destroyed by a compressive strength test using stainless steel tweezers.
Subsequently, WGB was removed from the flake-type sample with stainless steel tweezers
and was powdered using a mortar. To stop further hydration of the collected samples
for Rietveld-refinement-based quantitative analysis, the solvent exchange method using
isopropanol and ethyl ether was used as the standard condition. The samples were kept
dry using silica gel until testing. The test conditions for hydration product composition
analysis are as follows.
XRD was performed using a PANalytical X’Pert PRO-MPD device (DA101 at the KBSI
Daegu Center), and the scan range was set to 0.02◦ . Measurements were obtained over
a range of 5◦ to 70◦ at intervals of 3.0◦ . Figure 3 shows the XRD patterns of the binders
used in this study. TGA was conducted at a heating rate of 10 K/min in a N2 environment
using a thermal analyzer (TA Instrument, Q600, PH407, KBSI Pusan Center). Solid-state
29 Si MAS NMR was analyzed using an INOVA 600 spectrometer (KBSI Western Seoul

Center). The 29 Si MAS NMR spectrum was obtained at a rotation speed of 10.0 kHz, pulse
width of 2.2 µs, and relaxation delay of 22 s using a 5 mm HX-CPMAS probe and 5 mm
zirconia rotor.
To estimate the extent of hydration, thermodynamic modeling was performed based
on the clinker reactivity obtained using the Gibbs energy minimization software GEM-
Selektor v3.3 and CEMDATA 18.01 database [28]. The reactivity model of UHPC mixed with
WGBs was designed using the Rietveld refinement method and Parrot Killoh’s hydration
model [29]. This method, which tracks hydration dynamics as a function of hydration
time, was used to model HSCC mixed with WGBs. The aforementioned database contains
thermodynamic data as well as data for cement minerals, such as ettringite (Aft), single-
phase aluminum-iron (Afm), hydrotalcite, hydrated garnet, and calcium silicate hydrate
(C–S–H). The chemical composition of cement and the hydration reaction conditions
defined and entered in GEMS were combined with the Parrot–Killoh model to obtain the
Appl. Sci. 2022, 12, 8385 6 of 17

relationship between cement hydration degree and time. The Parrot–Killoh empirical
model is expressed in Equations (1)–(3). Equation (1) concerns nucleation and growth.
Equation (2) concerns diffusion, and Equation (3) concerns the formation of hydration
phase. The degree of hydration α at time t (in days) is expressed as αt = αt − 1 + ∆t·Rt − 1.
The parameter values used to calculate the degree of hydration were obtained from relevant
literature [30–35]. Equations (1)–(3) are the same as the values of K1 , N1 , K2 , K3 and N3
arranged by Parrott [32].

K1
Rt = (1 − αt )(− ln(1 − αt ))(1− N1 ) (1)
N1
2
K2 × ( 1 − α t ) 3
Rt = 1
(2)
1 − (1 − α t ) 3

Rt = K3 × (1 − αt ) N3 (3)
The cross-section of the specimen was analyzed by cutting the produced specimen to
Appl. Sci. 2022, 12, x FOR PEER REVIEW
halfits size. SEM–EDS imaging was then performed by taking a 3 × 3 mm sample6 from
of 18

the specimen. For SEM–EDS analysis, a HITACHI S-4800 (KBSI, Smart openlab) was used.

Figure 3.
Figure X-raydiffraction
3. X-ray diffraction(XRD)
(XRD)patterns
patternsofofbinder
binder (A:
(A: alite,
alite, C:C: calcite,
calcite, D: D: dipotassium
dipotassium sulfate,
sulfate, E:
E: ettringite, G: gypsum, Q: quartz, and R: rutile).
ettringite, G: gypsum, Q: quartz, and R: rutile).

3. Results and Discussion

To estimate the extent of hydration, thermodynamic modeling was performed based
3.1.the
on Flow Table and
clinker Densityobtained using the Gibbs energy minimization software GEM-
reactivity
Table
Selektor v3.34 summarizes
and CEMDATA the constructability
18.01 database [28].and density of HSCCs.
The reactivity model Theof HSCCs
UHPC mixedin this
with WGBs was designed using the Rietveld refinement method and Parrot Killoh’stable
study were fabricated using a high-flow admixture (superplasticizer), and the flow hy-
methodmodel
dration was performed
[29]. This after
method,the flow
which stopped. The flow table
tracks hydration of theasfresh
dynamics mixture
a function ofstate
hy-
was measured
dration time, was as used
220 mm when HSCC
to model WGBs mixed
were not incorporated.
with As the incorporation
WGBs. The aforementioned rate
database
of WGB thermodynamic
contains increased, liquidity datatended
as welltoasdecrease. S20 andminerals,
data for cement R20, with a WGB
such incorporation
as ettringite (Aft),
rate of 20%, showed a decrease in the fluidity of approximately
single-phase aluminum-iron (Afm), hydrotalcite, hydrated garnet, and calcium 50 mm. WGB is pre-wetted
silicate
and has (C–S–H).
hydrate a very lowThe probability
chemicalof absorbing formulation
composition of cement and water.
the The frictionreaction
hydration with thecondi-
paste
due to the surface texture of WGB reduces the overall workability.
tions defined and entered in GEMS were combined with the Parrot–Killoh model to obtain Although a WGB is
spherical, it is produced by expanding waste glass powder and has
the relationship between cement hydration degree and time. The Parrot–Killoh empirical a rough surface with
many pores.
model According
is expressed to a previous
in Equations study
(1)–(3). [36], the
Equation (1)surface texture
concerns of aggregate
nucleation affects
and growth.
the surface friction properties of the mixture and consequently affects
Equation (2) concerns diffusion, and Equation (3) concerns the formation of hydration the fluidity of the
mixture in the fresh mixture state. Therefore, as the mixing rate of WGBs increases, the
phase. The degree of hydration α at time t (in days) is expressed as αt = αt − 1 + ∆t·Rt − 1.
fluidity decreases owing to the friction between paste and WGBs. The HSCC density tended
The parameter values used to calculate the degree of hydration were obtained from rele-
to decrease as the amount of WGBs mixed increased. Since WGBs have a low density,
vant literature [30–35]. Equations (1)–(3) are the same as the values of K1, N1, K2, K3 and N3
the HSCC density decreases with an increasing incorporation rate. The densities of the S-
arranged by Parrott [32].
and R-group specimens were different; generally, R-group specimens had a lower density
than S-group specimens. Therefore, 𝐾1 the density decreased as the age of the WGB-mixed
𝑅𝑡 = (1 − 𝛼𝑡 )(− ln(1 − 𝛼𝑡 ))(1−𝑁1) (1)
specimen increased and as the𝑁1room temperature curing proceeded.
2
𝐾2 × (1 − 𝛼𝑡 )3 (2)
𝑅𝑡 = 1
1 − (1 − 𝛼𝑡 )3
 tended to decrease as the amount of WGBs mixed increased. Since WGBs have a low den-
sity, the HSCC density decreases with an increasing incorporation rate. The densities of
the S- and R-group specimens were different; generally, R-group specimens had a lower
density than S-group specimens. Therefore, the density decreased as the age of the WGB-
Appl. Sci. 2022, 12, 8385 mixed specimen increased and as the room temperature curing proceeded. 7 of 17

Table 4. Flow table and density.

Table 4. Flow table and

Density density.
(g/cm 3) Density (g/cm3)
ID. Flow Table (mm) ID. Flow Table (mm)
Age of 3 d Age of 28 d Age of 3 d Age
3 of 28 d
Density (g/cm3 ) Density (g/cm )
ID.
SP Flow Table
220(mm) 2.62 2.51 ID.RP Flow Table220
(mm) 2.46 2.38
Age of 3 d Age of 28 d Age of 3 d Age of 28 d
S5
SP
205
220
2.44
2.62
2.16
2.51 RP
R5 220
205 2.46
2.20 2.08
2.38
S10
S5 180
205 2.42
2.44 2.02
2.16 R5R10 205 180 2.202.07 1.96
2.08
S10 180 2.42 2.02 R10 180 2.07 1.96
S20 175 2.26 1.88 R20 175 1.94 1.81
S20 175 2.26 1.88 R20 175 1.94 1.81

3.2. Compressive Strength

3.2. Compressive Strength
Figure 4 shows the compressive strength test results of the HSCC specimens mixed
Figure 4 shows the compressive strength test results of the HSCC specimens mixed
with WGBs. The average compressive strengths of SP, S5, S10, and S20 specimens were
with WGBs. The average compressive strengths of SP, S5, S10, and S20 specimens were
120.3, 85.6, 71.8, and 67.4 MPa, respectively. As shown in Figure 4, the average compres-
120.3, 85.6, 71.8, and 67.4 Mpa, respectively. As shown in Figure 4, the average compressive
sive strengths of RP, R5, R10, and R20 specimens were 62.1, 50.3, 39.6, and 35.5 MPa, re-
strengths of RP, R5, R10, and R20 specimens were 62.1, 50.3, 39.6, and 35.5 Mpa, respectively.
spectively. The compressive strengths of cast-off specimens at 1 d of age were similar for
The compressive strengths of cast-off specimens at 1 d of age were similar for the two test
the two test groups. The error bars in Figure 4 indicate the 95% confidence interval.
groups. The error bars in Figure 4 indicate the 95% confidence interval.

Figure 4.
Figure Compressivestrengths
4. Compressive strengthsofof specimens
specimens subjected
subjected to steam
to steam curing
curing and and
roomroom temperature
temperature cur-
curing as a function of curing time.
ing as a function of curing time.

The compressive strength of S-group specimens was approximately twice that of

R-group specimens. As shown in Figure 4, the strengths slightly increased with age
even after the completion of steam curing. Since typical HSCCs are subjected to steam
curing at early ages, the maximum hydration reaction occurs in the matrix during this
period [36,37]. Therefore, high-temperature and high-pressure curing can significantly
improve the mechanical properties of HSCCs [38].
Figure 4 shows the strengths of R-group specimens. The compressive strength of
RP specimen at the age of 56 d is comparable to that of S20 specimen at the age of 56 d
in Figure 4. HSCC specimens contain large amounts of silica fume and silica powder.
Moreover, the reaction between them is activated only at high temperatures; consequently,
steam curing at earlier stages leads to faster strength development [38,39]. Hence, the HSCC
specimens subjected to room temperature curing showed slower strength development
owing to the insufficient activation reaction of internal silica components.
WGBs reduced the compressive strengths of specimens to a certain level, and speci-
mens with a maximum content of 20% developed approximately half the strength of the
Appl. Sci. 2022, 12, 8385 8 of 17

specimens without WGBs. In general, concrete mixed with LWAs contributes to a decrease
in density, which tends to reduce mechanical strength [40,41]. The reduction in compressive
strength results from the low mechanical strength of WGBs as LWAs, and the destruction
of aggregates results from the increasing volume content of LWAs [38].
If WGBs are to be used in mixing high-strength concrete and cement composites, an
additional review on the mixing rate and curing method of WGBs is needed.

3.3. XRD Analysis

Figure 5 shows the representative XRD patterns of the HSCC specimens that used
WGBs at the ages of 3 and 56 d. Belite was the major hydration product of most samples.
In some peaks, calcite, which is part of the cement raw material, and portlandite, which
is generated by humidity during curing, were observed. Rutile was observed along with
quartz, and it was used as an internal standard sample to ensure the internal standard
mixing amount [42]. S-group specimens had similar hydration products regardless of
age. The peak increase or decrease was similar because the initial hydration reaction
was maximized during steam curing. For S-group specimens, the peak intensity of the
specimens with a WGB content of 10% or 20% was observed to increase with age. The
peak of portlandite at 18.0◦ changed with increasing WGB content and age. This appeared
conspicuously in R-group specimens, and the internal moisture is expected to remain
during curing. In R-group specimens, some belite was found regardless of age. Belite was
measured at 16.2◦ , and the peak occurred regardless of WGB content. The proportion of
belite generated in samples at the age of 3 d tended to increase with WGB content; this can
be observed at 32.5◦ , 51.6◦ , and 38.0◦ . The proportion of belite detected in samples at the
age of 56 d was the highest in the R10 sample; this can be noticed at 29.5◦ , 32.0◦ , 32.3◦ , 34.8◦ ,
and 52.0◦ . R10 specimen exhibited somewhat higher peak intensities than other samples at
Appl. Sci. 2022, 12, x FOR PEER REVIEW
the ages of 3 and 56 d, probably because the WGB content of 10% improved the hydrate 9 of 18
formation rate in the paste matrix.

Figure 5.5. XRD

Figure XRDpatterns
patternsofof
samples
samplesat the ages
at the of 3of
ages and 56 d (A:
3 and 56 dalite,
(A: C: calcite,
alite, D: dipotassium
C: calcite, sulfate,
D: dipotassium
sulfate,
E: E: ettringite,
ettringite, G: gypsum,
G: gypsum, P: portlandite,
P: portlandite, Q:and
Q: quartz, quartz, and R: rutile).
R: rutile).

For the samples analyzed using the internal standard method, the quantification re-
sults of calcium hydroxide obtained through the normalization of anhydride are as fol-
lows. For SP, S5, S10, and S20, the measurements at the age of 3 d were 0.2%, 0.2%, 0.3%,
and 1.5%, respectively; moreover, the measurements at the age of 56 d were 0.1%, 0.3%,
0.2%, and 0.2%, respectively. For RP, R5, R10, and R20, the measurements at the age of 3
Appl. Sci. 2022, 12, 8385 9 of 17
Figure 5. XRD patterns of samples at the ages of 3 and 56 d (A: alite, C: calcite, D: dipotassium
sulfate, E: ettringite, G: gypsum, P: portlandite, Q: quartz, and R: rutile).

For
Forthe samples
the samples analyzed
analyzedusing the the
using internal standard
internal method,
standard method,the quantification
the quantification re-
sults of calcium hydroxide obtained through the normalization of
results of calcium hydroxide obtained through the normalization of anhydride are asanhydride are as fol-
lows. For SP,
follows. For S5,
SP, S10, andand
S5, S10, S20,S20,
the the
measurements
measurements at the ageage
at the of 3ofd3were
d were 0.2%,
0.2%,0.2%,
0.2%,0.3%,
0.3%,
and 1.5%, respectively; moreover, the measurements at the age of
and 1.5%, respectively; moreover, the measurements at the age of 56 d were 0.1%, 56 d were 0.1%, 0.3%,
0.3%,
0.2%,
0.2%,andand0.2%,
0.2%,respectively.
respectively.For
ForRP,
RP,R5,
R5,R10,
R10, and
and R20,
R20, the
the measurements
measurements at the age
at the age of
of 33 d
d of
of age
agewere
were2.9%,
2.9%,1.9%,
1.9%,4.2%,
4.2%,and
and4.5%,
4.5%,respectively;
respectively;moreover,
moreover,the the measurements
measurements at at
the
the age of 56 d were 1.5%, 1.0%, 1.2%, and 1.2%, respectively. Figure 6
age of 56 d were 1.5%, 1.0%, 1.2%, and 1.2%, respectively. Figure 6 shows the results ofshows the results
ofquantitative
quantitativeanalysis
analysisaccording
accordingtotothetheRietveld
Rietveld results
results of of XRD.
XRD. TheThe fractions
fractions of clinker
of clinker and
and hydrate
hydrate production
production werewere obtained
obtained at at
thethe ages
ages of of 3 and
3 and 56 56
d; d; moreover,
moreover, SP,SP,
S20,S20,
RP,RP,
and
and
R20R20werewere analyzed.
analyzed. When
When thethe mixing
mixing amount
amount of WGBs
of WGBs andand
thethe amount
amount of water
of water cap-
captured
tured
werewere the same,
the same, the of
the level level of internal
internal constituent
constituent materials
materials verifiably
verifiably differeddiffered
based onbased
steam
oncuring
steamand curing
room and room temperature
temperature curing. curing.

Quantitative
Figure6.6.Quantitative
Figure calculation
calculation results
results forclinker
for clinkerconstituents
constituents andproduct
and producthydrates
hydratescalculated
calculated
from
from the
the resultsofofXRD
results XRDquantitative
quantitativeanalysis.
analysis.

3.4. TGA
Figure 6 shows the derivative TG/TGA (DTG/TGA) results of HSCC specimens at
the ages of 3 and 56 d, respectively. The weight loss at 100 ◦ C or less was caused by the
dehydration and decomposition of bound water and ettringite [43]. At the age of 3 d, SP and
R10 specimens showed the highest weight losses, which were 2.53% and 3.69%, respectively.
At the age of 56 d, S20 and R10 specimens exhibited the highest weight losses, which were
3.24% and 2.79%, respectively. The weight loss observed in the range of 250–280 ◦ C is due
to the decomposition of aluminum hydroxide [44]. Among S-group specimens at the age of
56 d, S10 and S20 specimens showed thermal decomposition and mass losses of 2.22% and
2.32%, respectively. As WGB content increased, the weight reduction in samples increased,
and a more dynamic difference was observed in room temperature curing than in steam
curing. According to Figure 7, at the age of 3 d, the SP sample had the highest weight loss.
However, according to Figure 8, at the age of 56 d, the S20 sample had the largest weight
reduction with a WGB content of 20%. Pre-wetted WGBs reduced the strength owing to
the nature of LWAs, but the water contained inside contributed to the hydration reaction in
the matrix, as confirmed by the difference in the TGA of the generated hydrates.
 in samples increased, and a more dynamic difference was observed in room temperature
curing than in steam curing. According to Figure 7, at the age of 3 d, the SP sample had
the highest weight loss. However, according to Figure 8, at the age of 56 d, the S20 sample
had the largest weight reduction with a WGB content of 20%. Pre-wetted WGBs reduced
the strength owing to the nature of LWAs, but the water contained inside contributed to
Appl. Sci. 2022, 12, 8385 10 of 17
the hydration reaction in the matrix, as confirmed by the difference in the TGA of the
generated hydrates.

Appl. Sci. 2022, 12, x FOR PEER REVIEW 11 of 18

Figure 7. Thermogravimetric analysis (TGA) of samples at the ages of 3 and 56 d (the derivative
TG Figure
(DTG) 7. Thermogravimetric
and weight reduction analysis
curves of(TGA) of samples
the steam curing at
andtheroom
agestemperature
of 3 and 56 dcuring
(the derivative
specimensTG
(DTG) and weight reduction curves of the steam curing and room temperature curing specimens
are classified according to age, and the combustion materials are selected according to the tempera-
are classified according to age, and the combustion materials are selected according to the temper-
ture range).
ature range.).
 Figure 7. Thermogravimetric analysis (TGA) of samples at the ages of 3 and 56 d (the derivative TG
(DTG) and weight reduction curves of the steam curing and room temperature curing specimens
Appl. Sci. 2022, 12, 8385 11 of 17
are classified according to age, and the combustion materials are selected according to the temper-
ature range.).

29
Figure8.8. 29Si magic-angle
Figure magic-angle spinning
spinning nuclear
nuclear magnetic
magneticresonance
resonancespectrum
spectrum(MAS
(MASNMR)
NMR)results
resultsbased
based
on
onage
agefor
forS20S20and
andR20
R20specimens
specimensafter
aftersteam
steamcuring.
curing.

The R10 sample exhibited a weight loss similar to that of R20 at an early age, but it had
the highest weight loss of all samples at the age of 56 d. This indicates that most hydrates
were formed in the R10 sample, which is obvious in its XRD peak intensity. Moreover, the
hydration degree of the specimen with a WGB content of 10% was higher in the case of
room temperature curing. In addition, in S-group specimens, those specimens with the
highest weight losses at the ages of 3 and 56 d were SP and S20, respectively. This indicates
that internal curing is more effective as age increases and is considered to influence the
improvement of the internal curing of WGB after steam curing. R-group specimens showed
proportional results with increasing weight loss and increasing WGB mixing rate as age
increased. Thus, internal curing by WGBs was performed in R20, in contrast to RP, which
was sealed inside as age increased.
Considering a major weight loss, the AFm phase-related dehydration of C–S–H and
ettringite occurs at 100 ◦ C or less; moreover, the evaporation of bound water in the phys-
ically bound matrix occurs at 130 ◦ C. Dehydroxylation of portlandite then occurs until
400–450 ◦ C, followed by the decarbonation of calcium carbonate and unreacted clinker
between 500 and 700 ◦ C [45–48]. At an early age, portlandite was apparently consumed for
some level of hydration as age increased, confirming the occurrence of pozzolanic reaction.
This led to a larger change in R-group specimens, and the consumption of portlandite
increased as WGB content increased. All samples showed similar tendencies, indicating
that significant initial hydration occurred because of the curing environment, such as steam
curing. Table 5 summarizes the relative weight loss due to the dehydration of C–S–H and
ettringite, dehydroxylation of portlandite, and decarbonation of unreacted clinker. The
weight loss trend below 200 ◦ C clearly decreased with long-term age, and the decrease
rate was larger in the case of steam curing specimens. The ratio of Ca(OH)2 produced was
similar for both steam curing and room temperature curing on the 56th day. The amount of
CaCO3 clearly decreased with time due to long-term aging, and the decrease was large in
the specimens subjected to steam curing.
Appl. Sci. 2022, 12, 8385 12 of 17

Table 5. Weight loss of specimens by temperature range (%).

Below 200 ◦ C 400–500 ◦ C 650–750 ◦ C

Evaporation of H2 O from C–S–H
ID. Dehydroxylation of Ca(OH)2 Decarbonation of CaCO3
Dehydration of AFm/AFt
Age of 3 d Age of 56 d Age of 3 d Age of 56 d Age of 3 d Age of 56 d
SP 7.93 6.93 2.59 2.24 2.73 2.06
S5 6.60 7.32 3.06 2.00 2.83 2.53
S10 5.01 9.39 2.68 1.84 2.76 2.36
S20 5.43 9.78 2.07 1.74 2.75 2.61
RP 4.37 6.97 1.56 2.38 2.46 1.54
R5 4.23 7.75 2.28 1.99 2.27 1.60
R10 5.47 9.14 2.02 1.60 2.59 1.68
R20 5.17 8.05 3.04 1.83 2.67 1.72

3.5. Solid-State 29 Si MAS NMR Spectroscopy

Figure 8 shows the 29 Si MAS NMR analysis spectrum of HSCCs mixed with WGBs.
Spectral deconvolution was performed using OriginPro software after selecting the number
of characteristic peaks by referring to the existing literature. The simulation was repeated
until the chi-square estimated by emulating each peak using Gaussian and Lorentzian
functions reached 10−6 [49].
Since typical HSCCs are subjected to steam curing, the NMR of S20 and R20 samples
was measured based on the sample with a WGB content of 20% that exhibited the highest
C–S–H hydrate formation rate among S-group specimens. This choice was based on the
dehydration of AFm and the mass reduction rate of the AFt phase that occurred with
increasing age, as listed in Table 4.
Among the raw materials used for producing HSCCs, the amorphous component of
cement showed a narrow resonance at −71.3 ppm related to Q0 , which is the peak related
to belite (C2 S) [49,50]. The relative Q0 site area of the S20 sample shown in Figure 8 (3 d)
significantly decreased after 56 d. Hydrates that occur at the beginning of curing show that
most of the hydration was completed during steam curing, and all remaining hydrates
that occur with increasing age indicate the completion of hydration. The R20 sample in
Figure 8 (56 d) showed a relatively larger area than that of the S20 sample in Figure 8
(3 d); moreover, initial hydrates were observed due to the hydration of cement during the
standard curing period.
The resonance of Q1 and Q2 occurred at −79 and −85 ppm, respectively. This section
confirmed the existence of Si chains in the C–S–H gel. For the S20 sample in Figure 8 (3 d)
and the R20 sample in Figure 8 (56 d), the relative area increased with age. For S20 specimen,
Q1 /Q2 were 2.78 and 5.10 at the ages of 3 and 56 d, respectively; for R20 specimen, Q1 /Q2
were 4.79 and 1.84 at the ages of 3 and 56 d, respectively. Generally, a more pozzolanic
reaction occurs as Q1 /Q2 decreases for HSCCs [50–54]. Q1 /Q2 values show that the initial
extent of pozzolanic reaction was higher for S20, and the extent of pozzolanic reaction
gradually increased with age for R20. Hence, the peak intensity near 420 ◦ C related to
CH decomposition can be observed by referring to the DTG analysis in Figure 7 (R-Deriv.
weight). C–S–H decomposition near 420 ◦ C in Figure 7 (R-Deriv. weight) decreased with
age, which portrays the decreasing extent of the internal pozzolanic reaction [50]. In the
29 Si MAS NMR analysis results for HSCCs, silica fume displayed a peak at −109.1 ppm

related to Q4 , and silica powder displayed a peak at −112.5 ppm, representing the presence
of micro-silica [54,55]. In Figure 8 (56 d), a significant amount of micro-silica remained
owing to the low pozzolanic reaction of the silica powder component in the R20 sample, but
more pozzolanic reaction occurred as age increased. A similar tendency can be observed
from the S20 sample in Figure 8 (3 d).
For the S20 sample, the hydration degree of Si at the age of 3 d tended to increase with
age; in particular, it was approximately 4-fold higher at the age of 56 d than that at the age
Appl. Sci. 2022, 12, 8385 13 of 17

of 3 d. For the R20 sample, the hydration degree of Si was approximately 5-fold higher at
the age of 56 d than that at the age of 3 d. In the NMR results for the same WGB content
and age, Q1 , Q2 , and Q3 had higher intensities in the steam curing environment than that in
the standard curing environment. This indicates that the moisture inside pre-wetted WGBs
can improve the hydration degree of Si in the specimen in a steam curing environment.

3.6. Thermodynamic Modeling

Figure 9 shows the thermodynamic modeling results of the S20 and R20 samples as
a function of age. Thermodynamic modeling excluded the influence of the temperature
parameter applied to specimens during the initial steam curing and simply calculated the
total amount of hydrates produced in specimens. The major hydration products of the two
samples were predicted to be ettringite, mono sulfate, C–S–H, and portlandite; moreover,
the analysis was conducted based on the Rietveld method. The results showed that the
volume of the solid in the gel phase increased as the hydration reaction progressed inside
the samples mixed with pre-wetted WGBs, which agreed with the XRD and 29 Si MAS
NMR results.
The total volume of the S20 sample was approximately 10% larger than that of the
R20 sample at the age of 56 d because the generated hydrates had different volumes. The
volume of C–S–H gel was larger in the S20 sample at all ages. The change in the14Ca/Si
Appl. Sci. 2022, 12, x FOR PEER REVIEW of 18
ratio of C–S–H is closely related to the presence and consumption of portlandite, which
shows that the ratio begins to decrease as portlandite is completely consumed [56,57]. This
change occurred earlier in the S20 sample. Some discrepancies between the experimental
results and thermodynamic calculations can be considered errors caused by the absence
results and thermodynamic calculations can be considered errors caused by the absence
of the AFm phase [20,48]. When the same level of WGB was added, the difference in the
of the AFm phase [20,48]. When the same level of WGB was added, the difference in
volume content of solid hydrates caused by the difference in the curing environment
the volume content of solid hydrates caused by the difference in the curing environment
could be observed, and the generation of C–S–H gel is expected to increase as the curing
could be observed, and the generation of C–S–H gel is expected to increase as the curing
period increases. In addition, the finding that a certain amount of internal moisture was
period increases. In addition, the finding that a certain amount of internal moisture was
maintained indicates that pre-wetted WGBs contributed to continuous internal hydration
maintained indicates that pre-wetted WGBs contributed to continuous internal hydration
during the curing period by releasing moisture.
during the curing period by releasing moisture.

(a) (b)
Figure
Figure 9.
9. Thermodynamic
Thermodynamic modeling
modeling of
of samples:
samples: (a)
(a) S20
S20 and
and (b)
(b) R20
R20 specimens.
specimens.

3.7. Cross-Sectional
3.7. Cross-Sectional Analysis
Analysis of
of Specimens
Specimens
Figure 10
Figure 10 shows
showsthe
thecross-section
cross-sectionofof a specimen
a specimen mixed
mixed withwith
WGBsWGBs
at the atage
theofage56 of
d,
56 d, confirming the distribution of WGBs inside the specimen and the moisture
confirming the distribution of WGBs inside the specimen and the moisture inside WGBs. inside
WGBs.
For S20,For S20, numerous
numerous white products
white products wereinside
were formed formed insideThese
WGBs. WGBs. These were
products products
het-
were heterogeneous substances with gel-like aggregates. Some WGBs still
erogeneous substances with gel-like aggregates. Some WGBs still had residual moisture,had residual
moisture,
which which
could could be confirmed
be confirmed with eye.
with the naked the naked eye. WGB
WGB pores poreswith
were filled werehydrates
filled withby
hydrates by the moisture captured in WGBs. SEM analysis confirmed internal
the moisture captured in WGBs. SEM analysis confirmed internal voids; a large amount voids; a large
amount
of C–S–Hofgel
C–S–H gel wason
was formed formed on surface.
the pore the poreFor
surface. For R20, was
R20, moisture moisture was confirmed
confirmed to remain
to remain inside WGBs when the cross-section of the specimen was
inside WGBs when the cross-section of the specimen was visually checked. SEM visually checked.
and
SEM–EDS analyses of WGBs confirmed that the intensity of the spectrum was lower than
that for S20. These results were similar to those of thermodynamic modeling. Thermody-
namic modeling confirmed the hydration tendency of paste except in the case of WGB;
however, even in the case of WGB, hydration tendency can be considered similar to that
Appl. Sci. 2022, 12, 8385 14 of 17

SEM and SEM–EDS analyses of WGBs confirmed that the intensity of the spectrum was
lower than that for S20. These results were similar to those of thermodynamic modeling.
Thermodynamic modeling confirmed the hydration tendency of paste except in the case of
WGB; however, even in the case of WGB, hydration tendency can be considered similar to
that inside the pores of paste and WGBs connected by micropores. In addition, when steam
curing was performed and when the absorbed water in WGB was mixed, the hydration
tendency inside WGB was higher. Therefore, microstructure hydration could be improved
by combining the mixing of captured WGBs and steam curing to effectively improve the
internal curing of HSCCs. In addition, this trend was confirmed by the tendency for
forming hydrates in pores similar to that of Beyene et al. and Zhang et al. [58,59]. Beyene
et al. identified the hydrate form and content level through SEM-EDS, which was similar to
that found around R20 WGB analyzed in this study. S20 has high silica (Si) and the calcium
(Ca), magnesia (Mg), and alumina (Al) contents are up to 3-fold higher than that in R20. It
was found that there was a difference in the amount of hydrate formed in the surface pores
of WGB as hydration was promoted by steam curing. A study by Zhang et al. confirmed
the formation of abundant hydration products in the vicinity of LWA due to water released
from LWA. This study found that the analysis results of the formation of hydrates generated
around WGB and surface voids were similar to the results of thermodynamic modeling
Appl. Sci. 2022, 12, x FOR PEER REVIEW
analyzed in Section 3.6. Accordingly, the curing water released by the WGB promotes 15 of 18

hydration of the cement particles around it and makes the particle spacing denser.

(a)

(b)
Figure
Figure 10.
10. Cross
Cross sections
sections of
of test
test specimens
specimens mixed
mixed with
with WGBs
WGBs that
that confirm
confirm the
the state
state of
of the
the moisture
moisture
trapped inside WGBs: (a) S20; and (b) R20 specimens.
trapped inside WGBs: (a) S20; and (b) R20 specimens.

The experimental
experimentalresults
resultsofofthis
thisstudy
studyrevealed
revealed that WGBs
that WGBs used in HSCCs
used in HSCCscould im-
could
prove
improveinternal hydration
internal hydrationreactions. TheThe
reactions. curing method
curing for HSCCs
method for HSCCsmixed withwith
mixed WGBs was
WGBs
demonstrated
was to affect
demonstrated the number
to affect the numberof hydration products
of hydration basedbased
products on age,
on which may may
age, which also
influence the extent
also influence of strength
the extent development.
of strength In the In
development. case
theofcase
steamof curing, strength
steam curing, devel-
strength
opment and internal
development hydrate
and internal formation
hydrate levels at
formation earlyat
levels ages were
early almost
ages weretwice
almostthat of room
twice that
temperature curing at curing
of room temperature the ageatofthe
56 d.
ageEven after
of 56 completing
d. Even steam curing,
after completing the proportion
steam curing, the
proportion
of unreactedofsilica
unreacted
inside silica inside was
was partially partially
reduced, andreduced, and moisture
the internal the internal moisture
content was
content wasatmaintained
maintained at a by
a certain level certain level byWGBs.
pre-wetted pre-wetted WGBs.
Therefore, Therefore,
WGBs WGBs could
could contribute to
contributeunreacted
reducing to reducing unreacted
hydrates hydrates
inside HSCCs. inside HSCCs.

4. Conclusions
This study used WGBs as LWAs to improve the internal hydration degree of HSCCs
that use ultrafine particles. Moreover, internal hydration characteristics were investigated
for different aging durations. The correlation between the presence of pre-wetted WGBs
Appl. Sci. 2022, 12, 8385 15 of 17

4. Conclusions
This study used WGBs as LWAs to improve the internal hydration degree of HSCCs
that use ultrafine particles. Moreover, internal hydration characteristics were investigated
for different aging durations. The correlation between the presence of pre-wetted WGBs
and internal hydration was established. The results of this study are summarized as follows.
1. The compressive strength of HSCC specimens decreased as WGB content increased,
regardless of the curing environment. This was due to the low mechanical strength of
the added WGBs; strength decreased as the volume content of WGBs increased.
2. From XRD pattern analysis results, S-group specimens showed similar reactant peak
tendencies regardless of age. In R-group specimens, as WGB content increased
according to age, the production of products, such as belite, also increased.
3. From TGA results, S20 specimen exhibited the highest weight loss at the age of 56 d
in the section below 200 ◦ C. Based on these results, NMR spectrum analysis was
conducted for S20 and R20 specimens; the extent of pozzolanic reaction increased as
the Q1/Q2 ratio decreased.
4. The interactions among the major hydration products inside the matrix were iden-
tified through thermodynamic modeling. The volume of the generated hydrates
differed depending on the curing environment, and a large difference in the amount
of C–S–H gel generated was confirmed at an early age. R20 specimen exhibited
hydrate proportions similar to those of S20 specimen with increasing age.
5. The residual moisture trapped inside WGB was confirmed through cross-sectional
imaging of the specimen and SEM–EDS analysis. The amount of hydrates produced
by S20 and R20 confirmed through EDS analysis was found to have a difference of up
to 3 fold in the Si, Ca, and Al content. Moreover, the hydration tendency inside WGB
was similar to that in thermodynamic modeling.
The XRD, TGA, 29 Si MAS NMR, and SEM–EDS results of the HSCC paste mixed
with WGBs subjected to steam curing and room temperature curing showcased varying
formation of hydrates, including C–S–H gel. These results agreed with the results of
thermodynamic modeling. This study also confirmed the effect of internal curing by WGBs.
However, the compressive strength tended to decrease with the addition of WGBs owing
to the prior damage and poor mechanical properties of WGBs. Therefore, an analysis of the
hydration products formed inside WGB micropores is required. Moreover, further research
is required in the future to decrease the influence of WGBs on strength by adjusting the
geometry and size of WGBs or by selecting a more suitable LWA than WGB. This will help
improve the hydration degree in a highly densified matrix, such as UHPC.

Author Contributions: Conceptualization, S.P. and J.N.; methodology, S.P.; investigation, S.P.; formal
analysis, G.K. and J.N.; data curation, S.P.; visualization, S.P.; writing—manuscript draft, S.P.; writing—
reviewing and editing, S.P., S.L. and J.N.; validation, S.P., G.K., S.L. and J.N.; supervision, G.K., S.L.
and J.N. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the National Research Foundation of the Ministry of Science
and ICT of Korea in 2020 (grant number 2020R1C1C101403812).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Pedro, D.; de Brito, J.; Evangelista, L. Evaluation of high-performance concrete with recycled aggregates: Use of densified silica
fume as cement replacement. Constr. Build. Mater. 2017, 147, 803–814. [CrossRef]
2. Neville, A.; Aïtcin, P.-C. High performance concrete—An overview. Mater. Struct. 1998, 31, 111–117. [CrossRef]
3. Neville, A.M. Properties of Concrete; Longman: London, UK, 1995.
Appl. Sci. 2022, 12, 8385 16 of 17

4. Ozawa, M.; Subedi; Parajuli, S.S.; Uchida, Y.; Zhou, B. Preventive effects of polypropylene and jute fibers on spalling of UHPC at
high temperatures in combination with waste porous ceramic fine aggregate as an internal curing material. Constr. Build. Mater.
2019, 206, 219–225. [CrossRef]
5. Bentz, D.P.; Lura, P.; Roberts, J.W. Mixture proportioning for internal curing. Concr. Int. 2005, 27, 35–40.
6. Justs, J.; Wyrzykowski, M.; Bajare, D.; Lura, P. Internal curing by superabsorbent polymers in ultra-high performance concrete.
Cem. Concr. Res. 2015, 76, 82–90. [CrossRef]
7. Mechtcherine, V.; Reinhardt, H.W. Application of Super Absorbent Polymers (SAP) in Concrete Construction: State-of-the-Art Report
Prepared by Technical Committee; Springer Science and Business Media: Berlin/Heidelberg, Germany, 2012; p. 225.
8. Mönnig, S.; Lura, P. Superabsorbent Polymers—An additive to Increase the Freeze-Thaw Resistance of High Strength Concrete.
In Advanced Construction Materials; Springer: Berlin/Heidelberg, Germany, 2007; pp. 351–358.
9. Paul, A.; Murgadas, S.; Delpiano, J.; Moreno-Casas, P.A.; Walczak, M.; Lopez, M. The role of moisture transport mechanisms on
the performance of lightweight aggregates in internal curing. Constr. Build. Mater. 2021, 268, 121191. [CrossRef]
10. Van, V.-T.A.; Rößler, C.; Bui, D.-D.; Ludwig, H.-M. Rice husk ash as both pozzolanic admixture and internal curing agent in
ultra-high performance concrete. Cem. Concr. Compos. 2014, 53, 270–278. [CrossRef]
11. Liu, Y.; Wei, Y. Effect of calcined bauxite powder or aggregate on the shrinkage properties of UHPC. Cem. Concr. Compos. 2021,
118, 103967. [CrossRef]
12. Sun, Y.; Yu, R.; Shui, Z.; Wang, X.; Qian, D.; Rao, B.; Huang, J.; He, Y. Understanding the porous aggregates carrier effect on
reducing autogenous shrinkage of Ultra-High Performance Concrete (UHPC) based on response surface method. Constr. Build.
Mater. 2019, 222, 130–141. [CrossRef]
13. Bentz, D. A review of early-age properties of cement-based materials. Cem. Concr. Res. 2008, 38, 196–204. [CrossRef]
14. Chidiac, S.E.; Mihaljevic, S.N.; Krachkovskiy, S.A.; Goward, G.R. Efficiency measure of SAP as internal curing for cement using
NMR & MRI. Constr. Build. Mater. 2021, 278, 122365. [CrossRef]
15. Liu, Y.; Wei, Y. Internal Curing by Porous Calcined Bauxite Aggregate in Ultrahigh-Performance Concrete. J. Mater. Civ. Eng.
2021, 33, 04020497. [CrossRef]
16. Liu, K.; Yu, R.; Shui, Z.; Li, X.; Ling, X.; He, W.; Yi, S.; Wu, S. Effects of Pumice-Based Porous Material on Hydration Characteristics
and Persistent Shrinkage of Ultra-High Performance Concrete (UHPC). Materials 2018, 12, 11. [CrossRef] [PubMed]
17. Bentz, D.; Snyder, K. Protected paste volume in concrete: Extension to internal curing using saturated lightweight fine aggregate.
Cem. Concr. Res. 1999, 29, 1863–1867. [CrossRef]
18. Bentz, D. Influence of internal curing using lightweight aggregates on interfacial transition zone percolation and chloride ingress
in mortars. Cem. Concr. Compos. 2009, 31, 285–289. [CrossRef]
19. Bentz, D.P. Influence of Shrinkage-Reducing Admixtures on Early-Age Properties of Cement Pastes. J. Adv. Concr. Technol. 2006,
4, 423–429. [CrossRef]
20. Akcay, B.; Tasdemir, M.A. Optimisation of using lightweight aggregates in mitigating autogenous deformation of concrete. Constr.
Build. Mater. 2009, 23, 353–363. [CrossRef]
21. Akcay, B.; Tasdemir, M.A. Effects of distribution of lightweight aggregates on internal curing of concrete. Cem. Concr. Compos.
2010, 32, 611–616. [CrossRef]
22. Henkensiefken, R.; Nantung, T.; Weiss, J. Internal Curing-from the Laboratory to Implementation. In Proceedings of the
International Bridge Conference, Pittsburgh, PA, USA, 14–17 June 2009.
23. Van Tuan, N.; Ye, G.; van Breugel, K.; Copuroglu, O. Hydration and microstructure of ultra high performance concrete
incorporating rice husk ash. Cem. Concr. Res. 2011, 41, 1104–1111. [CrossRef]
24. Wang, X.; Yu, R.; Shui, Z.; Song, Q.; Zhang, Z. Mix design and characteristics evaluation of an eco-friendly Ultra-High Performance
Concrete incorporating recycled coral based materials. J. Clean. Prod. 2017, 165, 70–80. [CrossRef]
25. Lyu, Y.; Huang, H.; Ye, G.; De Schutter, G. Autogenous Shrinkage of Zeolite Cement Pastes with Low Water-Binder Ratio. In
Proceedings of the 14th International Congress Chemia Cement (ICCC 2015), Beijing, China, 13–16 October 2015; pp. 1–8.
26. Elango, K.; Vivek, D.; Anandaraj, S.; Saravanakumar, R.; Sanfeer, J.; Saravanaganesh, S. Experimental study on self compacting
concrete using light weight aggregate. Mater. Today: Proc. 2022, 60, 1362–1366. [CrossRef]
27. Arel, H.Ş. The influence of steam curing on early-age compressive strength of pozzolanic mortars. Arab. J. Sci. Eng. 2016,
41, 1413–1420.
28. Lothenbach, B.; Kulik, D.A.; Matschei, T.; Balonis, M.; Baquerizo, L.; Dilnesa, B.; Miron, G.D.; Myers, R.J. Cemdata18: A chemical
thermodynamic database for hydrated Portland cements and alkali-activated materials. Cem. Concr. Res. 2019, 115, 472–506.
[CrossRef]
29. Swaddiwudhipong, S.; Chen, D.; Zhang, M.H. Simulation of the exothermic hydration process of Portland cement. Adv. Cem. Res.
2002, 14, 61–69. [CrossRef]
30. Wang, Y.; Chen, Y.; Guo, B.; Zhang, S.; Tong, Y.; Niu, D. Study on the Strength and Hydration Behavior of Sulfate-Resistant
Cement in High Geothermal Environment. Materials 2022, 15, 2790. [CrossRef]
31. Parrot, L.J.; Killoh, D.C. Prediction of cement hydration. Proc. Brit. Ceram. Soc. 1984, 35, 41–53.
32. Parrot, L.J. Modeling of hydration reactions and concrete properties. Mater. Sci. Concr. 1989, 1, 181–195.
33. Hou, J.J.; He, X.; Ni, X.; Ma, B. Optimization and Hydration Mechanism of Ecological Ternary Cements Containing Phos-phogypsum.
2022. Available online: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4123782 (accessed on 31 May 2022).
Appl. Sci. 2022, 12, 8385 17 of 17

34. Peng, Y.; Tang, S.; Huang, J.; Tang, C.; Wang, L.; Liu, Y. Fractal Analysis on Pore Structure and Modeling of Hydration of
Magnesium Phosphate Cement Paste. Fractal Fract. 2022, 6, 337. [CrossRef]
35. Zhang, Y.; Çopuroğlu, O. The role of hydrotalcite-like phase and monosulfate in slag cement paste during atmospheric and
accelerated carbonation. Cem. Concr. Compos. 2022, 132, 104642. [CrossRef]
36. Zhang, M.-H.; Gjørv, O.E. Microstructure of the interfacial zone between lightweight aggregate and cement paste. Cem. Concr.
Res. 1990, 20, 610–618. [CrossRef]
37. Park, J.-S.; Kim, Y.J.; Cho, J.-R.; Jeon, S.-J. Early-Age Strength of Ultra-High Performance Concrete in Various Curing Conditions.
Materials 2015, 8, 5537–5553. [CrossRef] [PubMed]
38. Wu, Z.; Shi, C.; He, W. Comparative study on flexural properties of ultra-high performance concrete with supplementary
cementitious materials under different curing regimes. Constr. Build. Mater. 2017, 136, 307–313. [CrossRef]
39. Aldred, J.M.; Holland, T.C.; Morgan, D.R.; Roy, D.M.; Bury, M.A.; Hooton, R.D.; Olek, J.; Scali, M.J.; Detwiler, R.J.; Jaber, T.M.
Guide for the Use of Silica Fume in Concrete; ACI—American Concrete Institute—Committee: Farmington Hills, MI, USA, 2006;
p. 234.
40. Richard, P.; Cheyrezy, M. Composition of reactive powder concretes. Cem. Concr. Res. 1995, 25, 1501–1511. [CrossRef]
41. Aslam, M.; Shafigh, P.; Nomeli, M.A.; Jumaat, M.Z. Manufacturing of high-strength lightweight aggregate concrete using blended
coarse lightweight aggregates. J. Build. Eng. 2017, 13, 53–62. [CrossRef]
42. Klug, H.P.; Alexander, L.E. X-ray Diffraction Procedures: For Polycrystalline and Amorphous Materials; Wiley & Sons: Hoboken, NJ,
USA, 1974.
43. Park, S.M.; Seo, J.H.; Lee, H.K. Thermal evolution of hydrates in carbonation-cured Portland cement. Mater. Struct. 2018, 51, 1–8.
[CrossRef]
44. Winnefeld, F.; Lothenbach, B. Hydration of calcium sulfoaluminate cements—Experimental findings and thermodynamic
modelling. Cem. Concr. Res. 2010, 40, 1239–1247. [CrossRef]
45. Lothenbach, B.; Durdzinski, P.; De Weerdt, K. Thermogravimetric analysis. Pract. Guide Microstruct. Anal. Cem. Mater. 2016,
1, 177–211.
46. Yoon, H.N.; Seo, J.; Kim, S.; Lee, H.K.; Park, S. Characterization of blast furnace slag-blended Portland cement for immobilization
of Co. Cem. Concr. Res. 2020, 134, 106089. [CrossRef]
47. Borges, P.H.R.; Costa, J.O.; Milestone, N.B.; Lynsdale, C.J.; Streatfield, R.E. Carbonation of CH and C–S–H in composite cement
pastes containing high amounts of BFS. Cem. Concr. Res. 2010, 40, 284–292. [CrossRef]
48. Jeong, Y.; Park, H.; Jun, Y.; Jeong, J.-H.; Oh, J.E. Microstructural verification of the strength performance of ternary blended
cement systems with high volumes of fly ash and GGBFS. Constr. Build. Mater. 2015, 95, 96–107. [CrossRef]
49. Seo, J.; Kim, S.; Jang, D.; Kim, H.; Lee, H. Internal carbonation of belite-rich Portland cement: An in-depth observation at the
interaction of the belite phase with sodium bicarbonate. J. Build. Eng. 2021, 44, 102907. [CrossRef]
50. Lee, N.K.; Koh, K.; Kim, M.O.; Ryu, G. Uncovering the role of micro silica in hydration of ultra-high performance concrete
(UHPC). Cem. Concr. Res. 2018, 104, 68–79. [CrossRef]
51. Walkley, B.; Provis, J. Solid-state nuclear magnetic resonance spectroscopy of cements. Mater. Today Adv. 2019, 1, 100007.
[CrossRef]
52. Lothenbach, B.; Scrivener, K.; Hooton, R. Supplementary cementitious materials. Cem. Concr. Res. 2011, 41, 1244–1256. [CrossRef]
53. Love, C.A.; Richardson, I.G.; Brough, A.R. Composition and structure of C–S–H in white Portland cement–20% metakaolin pastes
hydrated at 25 ◦ C. Cem. Concr. Res. 2007, 37, 109–117. [CrossRef]
54. Girão, A.V.; Richardson, I.G.; Taylor, R.; Drummond-Brydson, R. Composition, morphology and nanostructure of C–S–H in 70%
white Portland cement–30% fly ash blends hydrated at 55 ◦ C. Cem. Concr. Res. 2010, 40, 1350–1359. [CrossRef]
55. Seo, J.; Park, S.; Lee, H. Evolution of the binder gel in carbonation-cured Portland cement in an acidic medium. Cem. Concr. Res.
2018, 109, 81–89. [CrossRef]
56. Schachinger, I.; Hilbig, H.; Stengel, T.; Fehling, E. Effect of Curing Temperature at an Early Age on the Long-Term Strength
Development of UHPC. In Proceedings of the 2nd International Symposium Ultra High Performance Concrete, Kassel University,
Kassel, Gemany, 5–7 March 2008; pp. 205–213.
57. Kim, S.; Lee, N.; Lee, H.K.; Park, S. Experimental and theoretical studies of hydration of ultra-high performance concrete cured
under various curing conditions. Constr. Build. Mater. 2021, 278, 122352. [CrossRef]
58. Beyene, M.A.; Munoz, J.F.; Meininger, R.C.; Di Bella, C. Effect of Internal Curing as Mitigation to Minimize Alkali-Silica Re-action
Damage. ACI Mater. J. 2017, 114, 417–428.
59. Zhang, Y.; Sun, X. Influence of multi-walled carbon nanotubes on the multi-scale performance of internally cured concrete
containing pre-wetted lightweight aggregate. J. Build. Eng. 2022, 58, 104986. [CrossRef]