See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/372665915

A cost-effective approach to manufacturing ultra-high-performance

lightweight concrete via air-entraining

Article in Archives of Civil and Mechanical Engineering · July 2023

DOI: 10.1007/s43452-023-00751-2

CITATIONS READS

0 30

6 authors, including:

Hongyan Chu Fengjuan Wang

Nanjing Forestry University Southeast University (China)
32 PUBLICATIONS 557 CITATIONS 46 PUBLICATIONS 365 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Sacrificial concrete in nuclear power plants View project

All content following this page was uploaded by Hongyan Chu on 31 July 2023.

The user has requested enhancement of the downloaded file.

Archives of Civil and Mechanical Engineering (2023) 23:210
https://doi.org/10.1007/s43452-023-00751-2

ORIGINAL ARTICLE

A cost‑effective approach to manufacturing ultra‑high‑performance

lightweight concrete via air‑entraining
Hongyan Chu1 · Jianjian Qin1 · Li Gao1 · Jinyang Jiang2 · Fengjuan Wang2 · Liguo Wang2

Received: 23 March 2023 / Revised: 29 June 2023 / Accepted: 10 July 2023

© Wroclaw University of Science and Technology 2023

Abstract
Ultra-high-performance concrete (UHPC) has attracted wide interests in civil engineering but it has large deadweight,
which limits its application in lightweight structures and rehabilitation. Compromising the merits of UHPC and lightweight
concrete, ultra-high-performance lightweight concrete (UHPLC) is a potential solution to solve the drawback of high self-
weight. Based on particle densely packing theory, this study proposed a UHPLC manufacturing approach using shale ceramic
sand and air-entraining agent (AEA). An experimental study was conducted to systematically examine the effects of the
types and contents of AEA on various performances of UHPLC, including its workability, mechanical properties, durability
and microstructure. The results show the following: (1) The addition of AEA resulted in the decrease of UHPLC in elastic
modulus, compressive strength and flexural strength but bring UHPLC higher specific strength; (2) A void system that is
featured with homogeneous distribution of pores and an appropriate size and could be formed by adding a suitable amount
of AEA; (3) The optimum proportion of AEA in UHPLC was 0.01% in consideration of its influence on UHPLC from the
perspectives of microstructure, durability and mechanical properties.

Keywords Lightweight ultra-high-performance concrete · Mechanical property · Microstructure · Porosity · Air-entraining

agent

1 Introduction overall structure in practical engineering is an important fac-

tor that leads to high engineering costs, so using lightweight
As a novel type of cement-based material, UHPC has concrete can reduce costs. On the other hand, prefabricated
been found to have outstanding performance in durability, components made of lightweight building materials have a
workability and mechanical properties. Generally, UHPC smaller weight, and thus it is convenient for transportation
has 40–60 GPa elastic modulus and more than 100 MPa and installation [4]. Therefore, to promote the usage and
compressive strength [1]. Due to its excellence in durabil- promotion, it is pressing to figure out how to reduce the
ity and mechanical properties, UHPC is perceived to have deadweight of UHPC. Ultra-high-performance lightweight
great application value in the fields of bridge engineering, concrete (UHPLC), the currently researching hotspot, has a
underground engineering, military engineering and nuclear slightly lower mechanical properties than those of UHPC,
power engineering [2, 3]. However, the density of UHPC but its deadweight is much lower than that of UHPC, which
is usually more than 2500 kg/m3, which limits its applica- is a potential solution to improve the performance of UHPC
tion in engineering. On the one hand, the deadweight of the in deadweight [5].
Replacing traditional river sand and quartz sand with
* Hongyan Chu lightweight aggregates is one of the most common meth-
[email protected] ods for preparing UHPLC. Lightweight aggregates that can
* Jinyang Jiang be used in ordinary concrete include polyurethane foam
[email protected] [6], diatomaceous earth [7], etc., and be used to produce
lightweight concrete that had compressive strength and
1
College of Civil Engineering, Nanjing Forestry University, density in the range of 30–60 MPa and 1200–1600 kg/m3,
Nanjing 210037, China
respectively. Functionally graded construction materials
2
School of Materials Science and Engineering, Southeast [8], perlite [9] and fine fly ash [10] are useful in preparing
University, Nanjing 211189, China

13
Vol.:(0123456789)
210 Page 2 of 19 Archives of Civil and Mechanical Engineering (2023) 23:210

lightweight high-strength concrete, and the compressive of UHPC. This MAA model has been widely utilized to
strength and apparent density of high-strength concrete design UHPC in some related research [21–25]. In our previ-
prepared by these lightweight aggregates are usually in ous research, we explored the feasibility of using the MAA
50–90 MPa and 1500–1900 kg/m3, respectively. In recent model to design UHPLC, and we found that the MAA model
years, some researchers have shifted their attention from could be used for the mix design of UHPLC [26]. Therefore,
river sand or quartz sand and are looking for some other it is feasible to optimize the mix proportion of UHPLC by
lightweight aggregates for the preparation of UHPLC. the MAA model, because the raw materials of UHPLC are
Zhang et al. [11] prepared UHPLC by using ceramic sand also different.
as lightweight aggregate, to get more than 40 GPa elastic Air-entraining agent (AEA) is a common surfactant, and
modulus, 16.5 MPa flexural strength and 110 MPa compres- it can reduce the surface tension of water by preferentially
sive strength. Wang et al. [12] used expanded perlite sand to attaching at the air–water interface, to entrain tiny air bub-
replace partial cement and quartz sand, and the compressive bles in the concrete [27]. The stable and sufficient pore sys-
strength of UHPLC is 117.2 MPa. After preparing UHPLC tem in fresh or hardened concrete has great influence on
with ceramic sand, Xie et al. [13] found that it had appar- the properties of concrete, such as resistance to freeze–thaw
ent density of about 1950 kg/m3 and compressive strength cycles, workability, rheology, pumpability, mechanical
reaching 110 MPa. In addition, some researchers have pre- strength and fireproof performance of concrete [28, 29].
pared UHPLC-based materials. Lu et al. [14] designed a Zhang et al. [30] explored how the amount of AEA affects
kind of high-performance lightweight concrete by expanded hardened concrete with regard to its mechanical strength.
clay and expanded shale lightweight aggregates, whose com- The study showed that the quantitative increase of AEA in
pressive strength and density are more than 80 MPa and concrete was linearly accompanied by the decline of com-
about 2200 kg/m3, respectively. Using glass microspheres pressive strength. Zhang et al. [31] pointed out that appro-
and glass powder, Lu et al. [15] made an UHPLC-based priately stable, closed, well-distributed air bubbles in con-
material that had 120 MPa compressive strength and about crete can disrupt the connectivity of the pore system, thereby
1800 kg/m3 density. However, there are few research reports minimizing chloride ion migration. Research by Nili et al.
on UHPLC at present, and the methods for the light qualita- [32] indicated that concrete with recycled aggregates, when
tive of UHPLC should be further developed. added with AEA, had lower elastic modulus and compres-
Mixture design, which has direct effects on the durabil- sive strength. Zheng et al. [33] stated that the addition of
ity, mechanical properties and workability of the material, AEA resulted in the decline in the compressive strength of
is a typical way to expand and promote the usage of UHPC concrete, but also led to the significant improvement in its
[16]. The mixture design method for traditional UHPC is frost resistance. Schackow et al. [34] found in an experimen-
highly empirical and lacks scientific theoretical guidance, tal study that the density of concrete was decreased, when
which restricts the engineering application and development the content of AEA was increased. Li et al. [35] confirmed
of UHPC [17, 18]. Fuller et al. [19] proposed the concept that the frost resistance of lightweight concrete was signifi-
of Fuller gradation curve in order to optimize the design of cantly improved, but its compressive strength was consid-
concrete mixture and improve the compactness of concrete erably reduced, due to the addition of AEA. It can be seen
particle accumulation. However, Fuller grading curve is a that the use of AEA in common concrete is a hot spot in
semi-empirical and semi-theoretical model. For mixture the research of cement-based materials. However, how AEA
with different particle sizes, the designer needs to select the affects the performance of UHPLC should be systematically
corresponding blending ratio on the Fuller grading curve examined in view of the lack of studies concerning the uti-
to obtain the maximum packing compactness. Larrard and lization of AEA in UHPLC.
Sedran [20] introduced the compression packing model, and The lightweight aggregate used herein was shale ceramic
this model can consider the difference between the actual sand, while the cementitious materials included silica fume,
packing density and the virtual packing density. However, fine fly ash, and cement. The MAA model was followed
the compression packing model is only suitable for mixture to design the reference mixture of UHPLC. On this basis,
filling with a single particle size. Particle densely packing three different types of AEA with different dosages (0%,
theory is one of the useful methods for the mixture design of 0.01%, 0.02%, 0.03%, 0.04% of cement quality) were added
UHPC. On this basis, Andreasen and Andersen developed a to UHPLC, to explore the impacts of AEA on UHPLC
modified Andreasen and Andersen (MAA) model, a model in four dimensions, including microstructure, durability,
more practical in mixture design of UHPC. Compared to mechanical properties and workability. In addition, and the
Fuller model [19] and compression packing model [20], the optimum dosage of AEA in UHPLC was also determined via
MAA model can be utilized to mix design of cement-based experimental study. The findings of this study have guiding
materials composed of different particle size raw materials. significance for the preparation of UHPLC by AEA.
Therefore, the MAA model has advantage in the mix design

13
Archives of Civil and Mechanical Engineering (2023) 23:210 Page 3 of 19 210

Fig. 1  Process of obtaining the shale ceramic sand: a Mining raw material, b Preliminary screening, c High-temperature roasting, d Fine screen-
ing

2 Experimental section Table 1  Chemical composition of raw materials (wt%)

Materials Cement Fine fly ash Silica fume Ceramic sand
2.1 Experimental materials
SiO2 20.99 55.39 97.52 58.6
Coming from Yangchun Shanshui Cement Co., Ltd, Grade Al2O3 9.00 19.05 0.66 23.1
P·O52.5 Cement was utilized as the cement herein, which Fe2O3 3.60 6.72 0.19 11.4
had 28 days strength of 60 MPa, density of 3150 kg/m3, CaO 60.21 13.45 1.18 0.6
specific surface area of 434.2 m ­ 2/kg, and loss on ignition of MgO 4.25 0.10 0.02 3.9
1.4%. Coming from Henan Zhengzhou Haoda Environmen- SO3 3.32 0.22 0.06 0.2
tal Protection Co., Ltd., the fine fly ash utilized in this work K 2O 0.45 2.75 0.21 0.6
had specific surface area of 2120 m ­ 2/kg and packing density Na2O 0.18 2.32 0.16 1.6
3
of 450 kg/m . Coming from Shanghai Elken International
Co., Ltd., the silica fume utilized herein had S­ iO2 mass frac-
tion of roughly 97.52% and specific surface area of 1324
­m2/kg. Coming from Hubei Yicheng Huihong Energy Sav- 100
Cement
Fine fly ash
ing Technology Co., Ltd., shale ceramic sand was utilized Silica fume
Cumulative finer/(%)

Ceramic sand
herein as the fine aggregate, which had 1 h water absorption 80 Theoretical designed curve
Experiment curve
of 9.9%, cylinder compressive strength of 1.7 MPa, appar-
60
ent density of 1730 kg/m3, and particle size of 0–2.36 mm.
The process of obtaining the shale ceramic sand is shown in
40
Fig. 1. Table 1 gives the details of their chemical composi-
tion, while Fig. 2 demonstrates their particle size distribution 20
curve. The highly efficient polycarboxylic acid water-reduc-
ing agent from Jiangsu Sobute New Materials Co., Ltd. was 0
utilized as the water-reducing agent here and had higher-
than-33% water-reducing rate and 40% solid content. Com- 10-3 10-2 10-1 100 101 102 103
ing from Jiangsu Subote New Materials Co., Ltd., the AEAs Particle size/(μm)
utilized herein included GYQ-I concrete high-efficiency
AEA, GYQ-III concrete high-efficiency AEA, and GYQ-IV Fig. 2  Particle size distribution of the raw materials

13
210 Page 4 of 19 Archives of Civil and Mechanical Engineering (2023) 23:210

concrete high-efficiency AEA, respectively, and these AEAs recorded as ULC0, and the mixtures of UHPLC with
were recorded as AEA-A, AEA-B, and AEA-C, respectively. 0.01%, 0.02%, 0.03% and 0.04% of AEA-A were recorded
AEA-A was a modified rosin compound with solid content as ULCA1, ULCA2, ULCA3 and ULCA4, respectively, as
of 50%; AEA-B was a composite AEA with solid content showed in Table 2. Likewise, the mixtures of UHPLC with
of 35%; AEA-C was a strong air-entraining AEA with solid the same contents of AEA-B and AEA-C were labelled
content of 29%. All three AEAs could entrain a large num- as ULCB1, ULCB2, ULCB3, ULCB4, ULCC1, ULCC2,
ber of closed, uniform, and fine bubbles into the cement- ULCC3 and ULCC4, respectively.
base materials at a low dosage. Tap water available to the
population of people living in Xuanwu District, Nanjing was 2.3 Preparation and curing of UHPLC specimens
utilized as the water in this work.
The weighed silica, fine fly ash, and fume cement were thor-
2.2 Mix design oughly mixed for 4 min in the mixer. The mixture was then
added with shale ceramic sand in the weighed amount and
The optimal dosage of each raw material was calculated, continuously mixed for 3 min. The water-reducing agent was
according to the MAA model [36], which was shown as poured into roughly 3/4 of water. A glass rod was utilized
follows: to completely mix the water-reducing agent with the water.
q q The mixture in the mixer was added with the mixture of
Di − Dmin
P(Di ) = q q (1) water-reducing agent and water and stirred for 3 min. All
Dmax − Dmin the residues in the containers were rinsed with the rest 1/4
water and added into the mixer for 5 min-stirring. At last,
n
∑ the mixer was evenly added with the steel fiber weighed and
(2)
[ ( ) ( )]
RSS = Pmix Di − Ptar Di → min, stirred for 5 min. The molds that were made into different
i=1
sized were applied with mold lubricant. The mixture in the
where D stands for the particle size (μm); P(Di) refers to the mixer was poured into each mold and fully vibrated placed
ratio of all smaller-than-D particles; q represents the dis- on the platform vibrator for the purpose of removing bubbles
tribution modulus; Dmax stands for the largest particle size inside. Then, plastic film was used to cover the molds for the
(μm). In general, there should be a coarse mix full of coarse purpose of reducing water loss. The specimens were taken
particles if q > 0.5; there should be a fine mix full of fine from the molds 24 h later, and cured for 3, 7, 14, and 28 days
particles when q < 0.25 [37]. In this study, q was set to 0.23 as actually required in a standardized environment featured
in consideration of the preparation of UHPLC using more with higher-than-90% relative humidity and 20 ± 2 °C tem-
fine particles and also in consistence with the mix design of perature. After the completion of these steps, an experimen-
UHPC put forward in the existing research [38]. tal study was done to test their mechanical properties.
In accordance with the corresponding formula, Ptar refers
to the calculated target gradation, while Pmix is the synthetic 2.4 Experiment methods
mixture. Using a least-squares-based optimization algorithm,
the proportions of each filler and aggregate in the design Chinese National Standard GB/T 2419-2005 [39] was fol-
mixture were adjusted until the deviation between Pmix(D) lowed in the testing of fluidity to explore the link between
and Ptar(D) was minimized, and the optimal mix design of the workability of UHPLC and the addition of AEA. The
UHPLC could be obtained. The theoretical designed curve truncated cone circular mold was poured with the fresh
and the experiment curve of the UHPLC based on the MAA UHPLC mixture twice and made to jump for 25 times
model are also shown in Fig. 2. on the jumping table after the mixture in the mold was
According to the MAA model, the mixture of UHPLC tamped using a tamping rod. The diameters on the bottom
could be obtained. The initial mixture of UHPLC was surface in the two mutually perpendicular directions were

Table 2  Mix proportions of Mixture Cement Fine fly ash Silica fume Ceramic sand Steel fiber Water Water reducer AEA-A
UHPLC (kg/m3)
ULC0 671 206 159 752 157 197 20.7 0
ULCA1 671 206 159 752 157 197 20.7 0.671
ULCA2 671 206 159 752 157 197 20.7 1.342
ULCA3 671 206 159 752 157 197 20.7 2.013
ULCA4 671 206 159 752 157 197 20.7 2.684

13
Archives of Civil and Mechanical Engineering (2023) 23:210 Page 5 of 19 210

measured using a ruler. The result of fluidity was repre- 3 Results and discussion
sented by the mean value.
The apparent density test method was implemented 3.1 Fluidity
in line with what is described in Chinese national
standard GB/T 50081-2019 [40]. There were three Figure 3 is the slump flow of the UHPLC added with AEA
specimens in each of the groups, of which the size was at varying ratios. According to what can be observed,
100 × 100 × 100 mm. Meanwhile, the calculated result was the slump flow of ULC0 was 230 mm, the slump flow of
represented by the mean value. ULCA1–ULCA4 was 235, 240, 255, and 265 mm, respec-
Chinese National standard GB/T 17671-1999 [41] was tively; The slump flow of ULCB1–ULCB4 was 240,
followed in this work to measure the mechanical proper- 245, 255, and 260 mm, respectively; The slump flow of
ties of UHPLC, including its flexural strength and com- ULCC1–ULCC4 are 240, 255, 250, and 270 mm, respec-
pressive strength. There were three specimens in each tively. These results indicate that UHPLC, from the per-
of the groups, whose size was 40 × 40 × 160 mm. These spective of workability, is able to meet the fluidity-related
specimens underwent 3-day, 7-day, 14-day and 28-day cur- criterion of self-compacting mortar (240 ± 20 mm) [45]. For
ing under a standardized environment. The mean value UHPLC with AEA-A, its fluidity was observed to follow an
obtained in this measurement was taken to represent their upward trend with the rising content of AEA, with the high-
flexural strength and compressive strength. est slump flow found from ULCA4. AEA-A is a resin air-
Chinese national standard GB/T 50081-2019 [40] was entraining agent, and calcium ions in the slurry of UHPLC
followed to measure the elastic modulus of UHPLC. There immediately react with alkali salts in AEA-A to generate
were three specimens in each of the groups, of which the insoluble calcium salts or rosin salts that can entrain air
size was 100 × 100 × 300 mm. After 28-day curing in into the slurry. The gaseous emulsion could be stabilized
standardized conditions, their stress and strain were meas- under the action of such insoluble sols or solid particles.
ured to test their relationship. The result of elastic modu- As the temperature of the mixture increases, the tendency
lus measured was represented by the mean of the secant and speed of insoluble salt sol precipitation increase, and
modulus on the stress–strain curve that corresponded to thus the foaming ability of AEA-A weakens [46]. There-
the loading of 30–60% [42]. fore, the slump flow of the UHPLC with AEA-A was lower
CECS 02:2005 [43] was followed to measure ultrasonic than that of the UHPLC with AEA-B or AEA-C. The effects
pulse velocity (UPV). There were three specimens in each of AEA-B and AEA-C on the slump flow of UHPLC were
of the groups, of which the size was 100 × 100 × 300 mm. similar, that is, the slump flow of UHPLC with AEA-B or
These specimens underwent 28-day curing under a stand- AEA-C was also observed to follow an upward tendency
ardized environment. The value of UPV was read using the with the increasing content of AEA, suggesting that the
intelligent non-metallic ultrasonic detector BJNM-1 and fluidity of the UHPLC could be improved, because of the
was finally represented by the mean. addition of AEA. The reasons why the AEA could improve
Chinese standard GB/T 50082-2009 [44] was followed
in the use of the method of rapid chloride migration to
measure rapid chloride diffusion resistance. A cylindrical 300
specimen, with 50 mm in thickness and 100 mm in diam- ULCA ULCB ULCC
eter, was utilized in the measurement of rapid chloride
diffusion resistance. 280
Slump flow/(mm)

The environmental scanning electronic FEI Quanta 200

was employed to measure the microstructure of UHPLC. 260
The samples utilized in the measurement of microstructure
were taken from the UHPLC specimens with the curing
age of 28 days. 240
The mercury intrusion porosimetry Micrometrics Auto-
poreIV-9500 was employed to measure pore structure and 220
porosity. The samples utilized in the measurement of this
test were taken from the UHPLC specimens with the cur-
ing age of 28 days and in the size of 40 × 40 × 160 mm. 200
0 0.01 0.02 0.03 0.04
AEA content/(% of cement)

Fig. 3  Slump flow of UHPLC with various contents of AEA

13
 210 Page 6 of 19 Archives of Civil and Mechanical Engineering (2023) 23:210

the fluidity of UHPLC might as follows: (1) The air bubbles unit weight of UHPLC [49]; On the other hand, the distance
entrained by the AEA could play an air ball-bearing effect between the ceramic sand particles was increased, due to
[47]; (2) The AEA could entrain a large number of micro- the incorporation of AEA, which led to the reduction of
bubbles into the slurry of UHPLC, which could reduce the the content of aggregate and cementitious materials in the
area of the contact of the mortar matrix touching steel fibers, UHPLC mixture [50].
thereby reducing the adhesion force between them [48]; (3)
Due to the addition of AEA, a large number of air bubbles 3.3 Mechanical properties
with uniform size were entrained into the slurry, which could
increase the distance between the ceramic sand particles. 3.3.1 Compressive strength
These tiny bubbles could reduce the frictional resistance
between the slurry particles, and thus the fluidity of the Figure 5 is the compressive strength of UHPLC. The details
UHPLC could be improved [49]. of Fig. 5a demonstrate the upward trend of the compres-
sive strength of UHPLC added with AEA-A with the ris-
3.2 Apparent density ing age of curing; when the ratio of AEA-A was increased,
the compressive strength of UHPLC displayed a decreas-
Figure 4 shows the apparent density of UHPLC. Accord- ing trend. The 28 days compressive strength of ULC0
ing to the details given in the figure, the apparent density and ULCA1–ULCA4 was 130.4, 124.9, 117.8, 106.0, and
of ULCA1–ULCA4 was, respectively, 1968.0, 1926.4, 97.7 MPa, respectively, which suggests that UHPLC had a
1850.5 and 1720.0 kg/m 3, respectively. Compared to decrease of 4.2–25.1% in compressive strength compared to
ULC0, UHPLC had a decrease of 8.5–20% in appar- ULC0, due to the addition of AEA-A. The details of Fig. 5b
ent density, due to the addition of AEA-A. The apparent demonstrate that when the ratio of AEA-B was increased,
density of ULCB1–ULCB4 was 1986.6, 1855.5, 1744.8, the compressive strength of UHPLC at each curing age
and 1651.6 kg/m3, respectively. Compared to ULC0, the also showed a decreasing trend. The 28 days compressive
apparent density of UHPLC was decreased by 7.6–23.2%, strength of ULCB1–ULCB4 was 124.2, 114.3, 104.5, and
due to the addition of AEA-B. The apparent density 94.6 MPa, respectively. In comparison to ULC0, the com-
of ULCC1–ULCC4 was 1900.4, 1826.0, 1501.5, and pressive strength of UHPLC was decreased by 4.8–27.5%,
1419.4 kg/m3, respectively, which suggested the apparent due to the incorporation of AEA-B. The details of Fig. 5c
density of UHPLC was decreased by 11.6–34.0%, due to the demonstrate that with the rising content of AEA-C, UHPLC
addition of AEA-C. On the whole, the apparent density of at each curing age underwent continuous decline in com-
UHPLC decreased in the order of ULCA > ULCB > ULCC. pressive strength. The 28 days’ compressive strength of
The main reasons for the decrease in the apparent density ULCC1–ULCC4 was 119.6, 94.5, 68.2, and 66.3 MPa,
of UHPLC might be as follows: On the one hand, many respectively, which indicates that UHPLC had a decrease
air bubbles were entrained into the matrix of UHPLC, due of 8.3–49.1% in compressive strength compared to ULC0,
to the addition of AEA, which resulted in the decrease of due to the addition of AEA-C. The compressive strength of
UHPLC added with AEA at 0.01% was decreased slightly
(< 10%), and that of UHPLC added with AEA at 0.04% was
2200 ULCA decreased sharply (> 25%). The porosity of UHPLC could
ULCB be increased, due to the addition of AEA, which resulted in
Apparent density/(kg/m3)

ULCC the reduction of the effective stress area of UHPLC, eventu-

2000 ally leading UHPLC to have a decline in its compressive
strength [51].
Specific strength is a main indicator to measure the
1800 light weight and high strength of the material, measured
as the ratio of strength to apparent density. Studying the
specific strength of UHPLC could better characterize its
1600 light weight and high strength. Figure 6 shows the spe-
cific strength of UHPLC. The details of Fig. 6 demonstrate
that with the increasing content of AEA content, there is a
1400 change trend of rising followed by declining with regard
0.00 0.01 0.02 0.03 0.04
to the specific strength of UHPLC. When the content of
AEA content/(% of cement) AEA was 0.01%, UHPLC was observed to have the maxi-
mum specific strength. Meanwhile, the specific strength of
Fig. 4  Apparent density of UHPLC with various contents of AEA ULC0 was 0.0619; the specific strength of ULCA1–ULCA4

13
Archives of Civil and Mechanical Engineering (2023) 23:210 Page 7 of 19 210

140
ULC0 ULCA
Compressive strengths/(MPa) ULCA1 0.08

Specific strength/(N∙m/kg)
ULCA2
ULCB
120
ULCA3 ULCC
ULCA4
100 0.06

80
0.04
60

0.02
40
3d 7d 14d 28d
Age/(day)
0.00
(a) Compressive strength of UHPLC with various contents of AEA-A. 0.00 0.01 0.02 0.03 0.04
AEA content/(% of cement)
140
ULC0
Compressive strengths/(MPa)

ULCB1
120 ULCB2 Fig. 6  Specific strength of different kinds of UHPLC
ULCB3
ULCB4
100
UHPLC with the same AEA. Furthermore, it could be also
concluded that the optimal amount of AEA in UHPLC was
80
0.01%, from the perspective of specific strength.
The specific strength of UHPLC and UHPC in varying
60
types from the published papers is summarized in Fig. 7.
It should be highlighted that all the water/binder ratio of
40
3d 7d 14d 28d UHPLC is within 0.2 ± 0.05. As shown in Fig. 7, the specific
Age/(day) strength of ULCA1, ULCB1, and ULCC1 was higher than
that of UHPLC prepared in these studies [4, 5, 13], and was
(b) Compressive strength of UHPLC with various contents of AEA-B.
also greater than that of UHPC made in these papers [11, 13,
140 14], which suggested that the specific strength of UHPLC
ULC0
Compressive strengths/(MPa)

ULCC1 designed herein was improved by the utilization of AEA.

120 ULCC2
ULCC3
ULCC4
100

80
0.08
Specific strength/(N∙m/kg)

60
0.06
40

3d 7d 14d 28d 0.04

Age/(day)
(c) Compressive strength of UHPLC with various contents of AEA-C.
0.02

Fig. 5  Compressive strength of different kinds of UHPLC

0.00
4]
]

3]

U 1 [11]

U 2 [13]
[4

[5
1

3 [1
e [1
A

ch

Lu

was 0.0635, 0.0617, 0.0577, and 0.0568, respectively; the

LC

LC

LC

PC
PC

PC
Xi
ba
U

specific strength of ULCB1–ULCB4 was 0.0642, 0.0630,

H

H
m

U
U

0.0600, and 0.0573, respectively; the specific strength of Different kinds of UHPLC and UHPC
ULCC1–ULCC4 was 0.0629, 0.0579, 0.0477, and 0.0452,
respectively. Therefore, the specific strength of ULCA1, Fig. 7  Specific strength of different kinds of UHPLC and UHPC with
ULCB1 and ULCC1 was the maximum in each kind of w/b at 0.2 ± 0.05

13
210 Page 8 of 19 Archives of Civil and Mechanical Engineering (2023) 23:210

3.3.2 Flexural strength 20
ULC0
ULCA1
Figure 8 gives the flexural strength of varying types of

Flexural strength/(MPa)
ULCA2
16
UHPLC designed herein. With rising content of AEA, ULCA3
ULCA4
the flexural strength of UHPLC at each curing age dis-
12
played a decreasing trend; the flexural strength of UHPLC
with different contents of AEA was increased continually
when the curing age was increased. The 28 days flexural 8
strengths of ULC0 was 17.40 MPa. The 28 days flexural
strengths of ULCA1–ULCA4 was 15.33, 14.07, 12.11 and 4
11.46 MPa, respectively. This implies that as a result of the
addition of AEA-A, the flexural strength of UHPLC was
0
decreased by 11.7–34.1%; the 28 days flexural strength of 3d 7d 14d 28d
ULCB1–ULCB4 was 16.25, 14.53, 12.11, and 11.48 MPa,
Age/(day)
respectively, and UHPLC had a decrease of 6.6–34.0% in
its flexural strength, due to the addition of AEA-B; The (a) Flexural strength of UHPLC with
28 days flexural strengths of ULCC1–ULCC4 was 14.21, various contents of AEA-A.
14.06, 7.96, and 7.10 MPa, respectively. This indicates that 20
as a result of the addition of AEA-C, the flexural strength of ULC0
ULCB1
UHPLC was decreased by 18.3–52.3%. The reasons why the

Flexural strength/(MPa)
16 ULCB2
flexural strength of UHPLC was reduced due to the addition ULCB3
of AEA were similar with those of the compressive strength ULCB4
of UHPLC mentioned above. 12
The flexural strength of UHPLC and UHPC in varying
types from the published papers is summarized in Fig. 9. It 8
should be highlighted that the water/binder ratio of these
UHPLC and UHPC is within 0.2 ± 0.05. As shown in Fig. 9,
4
the flexural strength of ULCA1, ULCB1and ULCC1 was
higher than that of UHPLC prepared in these studies [11,
13, 52–54]. Particularly, the flexural strength of ULCA1 and 0
3d 7d 14d 28d
ULCB1 was even higher than that of UHPC presented in
reference [55], which suggested that the flexural strength Age/(day)
of UHPLC designed herein was satisfactory because of the (b) Flexural strength of UHPLC with various
utilization of AEA. contents of AEA-B.

20
3.3.3 Elastic modulus ULC0
ULCC1
Flexural strength/(MPa)

ULCC2
Figure 10 is the stress–strain curves of UHPLC. Accord- 16
ULCC3
ingly, the strain of UHPLC was basically linear with stress. ULCC4
The increase of AEA led to the decreasing slope of the 12
stress–strain curve. It can be seen that the incorporation
of AEA resulted in lower elastic modulus of UHPLC. The 8
elastic modulus of UHPLC could be obtained from the data
shown in Fig. 10. The elastic modulus of UHPLC added
with AEA at varying ratios is shown in Fig. 11. Obvi- 4
ously, UHPLC added with an increasing content of AEA-A
was found to have a decline in elastic modulus. The elas- 0
tic modulus of ULC0 was 41.1 GPa. The elastic modulus 3d 7d 14d 28d
of ULCA1–ULCA4 was 39.0, 36.9, 35.7, and 34.6 GPa, Age/(day)
respectively. ULCA1, ULCA2, ULCA3, and ULCA4 had (c) Flexural strength of UHPLC with various
a decrease of 5.1%, 10.3%, 13.2%, and 16.1%, respectively contents of AEA-C.
in elastic modulus when compared with ULC0; UHPLC
added with an increasing content of AEA-B was also found Fig. 8  Flexural strength of different kinds of UHPLC

13
Archives of Civil and Mechanical Engineering (2023) 23:210 Page 9 of 19 210

20 that AEA-C performed better than AEA-A and AEA-B in

terms of the air-entraining effect. After AEA was incorpo-
rated in UHPLC, there were many fine air bubbles in the
Flexural strength/(MPa)

15 slurry of UHPLC. And these fine air bubbles could connect

to form large bubbles, and then could escape from the slurry
of UHPLC, when the contents of AEA were too high [56].
10 Therefore, AEA did not have a strong air-entraining effect
at a lower content of AEA, and thus resulted in slight drops
in the mechanical properties of UHPLC, and vice versa.
5 In comprehensive consideration of the impacts of AEA
on UHPLC with respect to its mechanical properties and
specific strength, the optimal amount of AEA in UHPLC
0
was 0.01%. When UHPLC was added with AEA at 0.01%,
3]
1]

2]

3]

4]

]
the decrease in the apparent density of UHPLC was lower

5
1

1
g [1

e [1

[5

[5

g [5

[5
A

uo

uo

Su
LC

LC

LC

Xi

than that of compressive strength, and thus a higher specific

an

en
G

G
U

U
Zh

M
strength of UHPLC could be obtained, which suggested that
Different kinds of UHPLC and UHPC UHPLC produced herein fully conformed to the demand for
light weight and high strength.
Fig. 9  Flexural strength of different kinds of UHPLC and UHPC with
w/b at 0.2 ± 0.05 3.4 UPV

The UPV of UHPLC is presented in Fig. 12. As shown in

to have a decline in elastic modulus. The elastic modulus Fig. 12, the UPV of UHPLC was decreased, due to the addi-
of ULCB1–ULCB4 was 38.9, 38.0, 32.7, and 31.9 GPa, tion of AEA. The UPV of ULC0 was 4.55 km/s. The UPV of
respectively. ULCB1, ULCB2, ULCB3, and ULCB4 had ULCA1, ULCA2, ULCA3, and ULCA4 was, respectively,
a decrease of 5.5%, 7.0%, 20.6%, and 22.5%, respectively, 4.25, 4.23, 4.18 and 4.16 km/s. When the content of AEA-A
in elastic modulus when compared with ULC0; UHPLC was increasing, the UPV of UHPLC showed a decreasing
added with an increasing content of AEA-C was also found trend. The UPV of ULCB1, ULCB2, ULCB3, and ULCB4
to have a decline in elastic modulus. The elastic modulus was 4.40, 4.24, 3.84 and 3.82 km/s, respectively, which sug-
of ULCC1–ULCC4 was 38.7, 34.8, 32.1, and 31.0 GPa, gested that the UPV of UHPLC was decreased continually
respectively. ULCB1, ULCB2, ULCB3 and ULCB4 had a with the increase of AEA-B. The UPV of ULCC1, ULCC2,
decrease of 6.1%, 15.5%, 22.2%, and 24.8%, respectively, in ULCC3 and ULCC4 was 4.23, 3.93, 3.81 and 3.65 km/s,
elastic modulus when compared with ULC0. It could be seen respectively. The UPV of UHPLC added with an increas-
that there was a critical value for the content of the three ing amount of AEA-C was also found to decrease gradu-
kinds of AEAs. When the addition of AEA exceeded this ally. Analyzing the above experimental results, it could
critical value, the elastic modulus of UHPLC was decreased be seen that the influence of the three kinds of AEAs on
sharply. When the content of AEA was 0.01%, the reduction UHPLC was similar, but the influence of AEA-C on the
for the elastic modulus of UHPLC was rather low, which UPV was the most obvious, followed by AEA-B, and the
was consistent with the mechanical properties of UHPLC influence of AEA-A on the UPV was the smallest, which
(see Figs. 5 and 8). indicated that the air-entraining effect of the three AEAs was
It could be seen from the experimental results of mechan- as follows:AEA-C > AEA-B > AEA-A. This was basically
ical properties of UHPLC that the addition of AEA resulted consistent with the change of UHPLC in mechanical proper-
in the decreases of UHPLC in elastic modulus, compressive ties. As a result of the addition of AEA, substantial minor air
strength and flexural strength. The mechanical properties bubbles got into the matrix of UHPLC, which increased the
UHPLC with AEA-A or AEA-B were decreased gradually, porosity of the UHPLC and reduced the density of UHPLC,
when the content of AEA was between 0.01% and 0.03%. so that the UPV propagated in the UHPLC was decreased.
However, when added with AEA at a ratio 0.04%, UHPLC In this work, the changing trend of UPV was similar to
had sharp decreases in mechanical properties. The mechani- the changing trend of mechanical strengths of UHPLC.
cal properties UHPLC with AEA-C were reduced sharply, When it came to UHPLC, UPV and mechanical properties
when the content of AEA reached 0.03%, which indicated were found to have a certain association. In this case, UPV

13
 210 Page 10 of 19 Archives of Civil and Mechanical Engineering (2023) 23:210

Fig. 10  Stress–strain curves of ULC0

different kinds of UHPLC 100 ULCA1
ULCA2
80 ULCA3
ULCA4

Stress/(MPa)
60

40

20

0 500 1000 1500 2000 2500 3000

Strain/(με)

(a) Stress-strain curves of ULCA with various contents of AEA.

ULC0
100
ULCB1
ULCB2
80 ULCB3
ULCB4
Stress/(MPa)

60

40

20

0 500 1000 1500 2000 2500 3000

Strain/(με)

(b) Stress-strain curves of ULCB with various contents of AEA.

ULC0
100 ULCC1
ULCC2
80 ULCC3
ULCC4
Stress/(MPa)

60

40

20

0 500 1000 1500 2000 2500 3000

Strain/(με)
(c) Stress-strain curves of ULCC with various contents of AEA.

13
Archives of Civil and Mechanical Engineering (2023) 23:210 Page 11 of 19 210

was obtained through the use of Origin software to fit the

40 experimental results as shown as follows:
Elastic modulus/(GPa)

Fc = 26.22199V, (3)
30 where Fc is used to represent compressive strength, while V
is a symbol used to represent UPV.
The fitting curve was observed to have a correlation
20 coefficient (R2) of 0.98458, which indicated the good fitting
between compressive strength and UPV. Meanwhile, it is
observable that their relationship could be characterized by
10
Formula (3).
According to what is depicted in Fig. 13b, the flexural
0 strength of UHPLC climbed with the rise of UPV. It could
be inferred that the flexural strength and UPV of UHPLC
L 0
L 1
L 2
LC 3
U A4
L 0
L 1
L 2
LC 3
U B4
LC 0

4
has a certain connection. Their specific relationship was
L 1
LC 2
U C3
U LC
U CA
U CA
U CA

U LC
U CB
U CB
U CB

U LC

C
U C
U CC

LC
U

obtained through the use of Origin software to fit the exper-

Different kinds of UHPLC imental data, and the fitting result is shown as follows in
Formula (4:).
Fig. 11  Elastic modulus of UHPLC with various contents of AEA
Ff = 2.87268V (4)
In the above formula, Ff is used to represent flexural
strength, while V is a symbol used to represent UPV.
4.8
The fitting curve was observed to have a correlation coef-
ficient (R2) of 0.96710, which suggested that the relation
4.4 between flexural strength and UPV could be evaluated by
UPV/(km/s)

the fitting curve.

According to what is depicted in Fig. 13c, UHPLC
4.0
had increasing gradually elastic modulus when its UPV
increased. The association between elastic modulus and
3.6 UPV could be acquired through the use of Origin software
to fit the experimental data, and the fitting result is displayed
in Formula (5).
3.2
Ec = 8.79411V (5)
In the above formula, Ec is used to represent elastic modu-
L 0

LC 3
L 1
L 2

U A4
L 0

L 2

4
LC 1

LC 3
U B4
L 0
LC 1
L 2
LC 3
U LC
U CA
U CA
U CA

U LC

U LC
U CB
U B
U CB

U CC
U C
U CC
C

lus, while V is a symbol used to represent UPV.

U

Different kinds of UHPLC The fitting curve was observed to have a correlation
coefficient (R2) of 0.99901. It could be concluded that the
Formula (5) could be used to characterize the association
Fig. 12  The UPV of UHPLC with various contents of AEA
between elastic modulus and UPV.
Furthermore, concrete is regarded to have a series of
was used to predict mechanical properties by applying ultra- mechanical properties, typically represented by elastic
sonic testing technology. The relationships between UPV modulus and compressive strength. Previously, a large
and strengths of UHPLC are displayed in Fig. 13. number of researchers have focused on the identification of
According to what is depicted in Fig. 13a, the compres- how elastic modulus of concrete is correlated with its com-
sive strength of UHPLC climbed with the rise of UPV. It pressive strength, and the use of the compressive strength
could be inferred that the compressive strength and UPV of of concrete offers a convenient way to calculate its elastic
UHPLC has a certain connection. Their specific relationship modulus. For UHPC, there have been many studies propos-
ing empirical formulas to demonstrate how its compressive
strength is correlated with its elastic modulus [57–59]. In

13
210 Page 12 of 19 Archives of Civil and Mechanical Engineering (2023) 23:210

Fig. 13  The relationships between UPV and mechanical strengths: ▸

a the relationship between UPV and compressive strength; b the
relationship between UPV and flexural strength; c the relationship
between UPV and elastic modulus

terms of UHPLC, such a relationship was also built in this

study, and the result is displayed in Fig. 14, from which it
is easy to observe that the elastic modulus of UHPLC fol-
lowed an upward trend when its compressive strength was
increasing. The experimental data were fitted by the Origin
software, and the relationship between these two variables
was obtained, as shown in Formula (6).

Ec = 5.80028 × Fc 0.39650 , (6)

where Ec represents elastic modulus, and Fc represents com-

pressive strength. (a)
The fitting curve was found to have a correlation coeffi-
cient (R2) of 0.92308, revealing that Formula (6) performed
well in accurately assessing the relationship between the two Measured value
18 Fitting curve
properties examined. In addition, their relationship was evi-
dently a power function in this work, exactly in consistence Flexural strength/(MPa)
with the fitting results acquired by Nguyen et al. [60], Wang 15
et al. [61], and Jurowski et al. [62].

3.5 Chloride diffusion resistance 12

Equation y = a + b*x
Plot B

The details about the chloride ion permeation test and the 9
Weight
Intercept
No Weighting
0 ± --

chloride migration coefficients of UHPLC measured by Slope

Residual Sum of Squares
2.87268 ± 0.1416
73.12374

the RCM method are summarized in Table 3. The chlo- Pearson's r

R-Square (COD)
0.98341
0.9671

ride migration coefficients of UHPLC are also presented 6 Adj. R-Square 0.96475

in Fig. 15, from which it is quite easy to observe that the 3.6 3.8 4.0 4.2 4.4 4.6 4.8
chloride migration coefficients of UHPLC basically dis-
played a gradually increasing trend with the increase of UPV/(km/s)
AEA. This probably because the addition of AEA caused (b)
the expansion of the internal porosity of UHPLC. When 42
the content of AEA in UHPLC was same, the chloride Measured value
migration coefficients were expressed in the order of Fitting curve
39
Elastic modulus/(GPa)

ULCC > ULCB > ULCA, which indicated that AEA-C has
the best air-entraining effect in UHPLC. A study shows that
the air content and the water-binder ratio of concrete are 36
important factors that affect the chloride migration coeffi- Equation y = a + b*x
cients of concrete, and that there is an optimal amount for 33 Plot B
Weight No Weighting
air content of concrete [31]. In the meantime, when UHPLC 0 ± --
Intercept
was added with AEA at 0.01%, its chloride migration coef- 30 Slope 8.79411 ± 0.074
ficient climbed slightly; while if UHPLC was added with Residual Sum of Squa 19.97768
Pearson's r 0.9995
AEA at more than 0.01%, there was a sharp rise in its chlo- 27 R-Square (COD) 0.99901
ride migration coefficient. To maintain the better chloride Adj. R-Square 0.99894
diffusion resistance of UHPLC, the content of AEA in
3.6 3.8 4.0 4.2 4.4 4.6 4.8
UHPLC was 0.01%.
UPV/(km/s)
(c)

13
Archives of Civil and Mechanical Engineering (2023) 23:210 Page 13 of 19 210

42 that the utilization of AEA brought significant increase

Mersured value
in the porosity of UHPLC. The porosity of UHPLC with
Fitting curve
different kinds of AEA was followed in the sequence of
39
ULCC1 > ULCB1 > ULCA1, which indicated that the
Elastic modulus/(GPa)

air-entraining effect of the three AEAs was as follows:

36 AEA-C > AEA-B > AEA-A. Figure 16b shows the pore
size distribution of UHPLC, from which it is easy to
Model Allometric1 observe that UHPLC had a distinct peak in the curve of
33 Equation y = a*x^b
its pore size distribution, exactly corresponding to the
Plot B
a 5.80028 ± 0.991 threshold pore diameter of UHPLC. ULC0, ULCA1,
30 b 0.3965 ± 0.0362 ULCB1, and ULCC1 were found to have a threshold pore
Reduced Chi-S 0.93661
R-Square (COD 0.92308 diameter of 16.39, 17.94, 18.39 and 20.41 nm, respec-
Adj. R-Square 0.91608 tively. Accordingly, there was an increasing tendency in
27
60 80 100 120 140
the threshold pore diameter of UHPLC that was added
with more AEA. The threshold pore diameter of ULCC1
Compressive strength/(MPa) was the highest, which suggested that the AEA-C had the
best air-entraining effect. UHPC is generally believed to
Fig. 14  Relationship between elastic modulus and compressive have a threshold pore diameter around 10 nm [63]. Obvi-
strength of UHPLC ously, the threshold pore diameter of UHPLC produced
in this work was higher than that of UHPC, which might
3.6 Porosity and pore size distribution by one of the reasons for the lower mechanical properties
of UHPLC. Figure 16c is the cumulative pore volume of
For the purpose of exploring how the porosity and pore the UHPLC. As shown in Fig. 16c, the cumulative pore
structure of UHPLC are affected by different types of volume of UHPLC shared a similar changing trend with
AEAs, the MIP experiment was carried out on the speci- AEA. The order of ULC0 < ULCA1 < ULCB1 < ULCC1
mens of ULC0, ULCA1, ULCB1, and ULCC1. The MIP from the perspective of cumulative pore volume was con-
results of UHPLC are shown in Fig. 16. Figure 16a is sistent with the porosity of UHPLC.
the porosity of UHPLC, and the porosity of ULC0, Concrete is assumed to have the following three types
ULCA1, ULCB1, and ULCC1 was 7.9%, 9.3%, 10.5%, of pores: gel pores (less than 10 nm), capillary pores
and 12.6%, respectively. Compared with the ULC0, the (ranging from 10 to 5000 nm), and macro-pores (greater
porosity ULCA1, ULCB1, and ULCC1 was increased by than 5000 nm) [64]. Their contents in UHPLC are shown
17.7%, 32.9%, and 59.5%, respectively, which suggested in Fig. 16d. The proportion of capillary pores in ULC0,
ULCA1, ULCB1, and ULCC1 was 52.9%, 44.2%, 43.6%,

Table 3  Details and results of No Specimen Anodic solution Initial Impose voltage, Voltage DRCM/10−12
chloride diffusion resistance test thickness L/ temperature T/°C current/ new initial current, absolute value ­m2 s­ −1
(RCM) mm mA time U /V
Start End Avg /V I/mA t/h

ULC0 50.3 19.8 23.7 21.8 33.58 60.01 31.34 24 60.01 2.76
ULCA1 50.1 19.0 20.4 19.7 32.46 49.99 32.53 24 49.98 3.33
ULCA2 50.1 19.5 20.7 20.1 47.63 29.98 41.37 24 29.97 4.84
ULCA3 50.0 19.5 20.8 20.2 39.00 39.98 38.44 24 39.98 7.37
ULCA4 49.8 20.3 20.9 20.6 30.40 50.00 50.14 24 49.99 8.31
ULCB1 50.2 19.5 21.3 20.4 58.88 24.99 57.93 24 24.99 5.24
ULCB2 50.0 20.1 22.4 21.3 31.31 39.99 28.00 24 39.98 5.26
ULCB3 50.1 20.4 22.8 21.6 18.98 19.98 17.96 48 19.98 8.50
ULCB4 49.9 19.6 22.2 20.9 43.60 34.99 39.51 24 34.99 10.79
ULCC1 50.2 19.5 21.6 20.6 45.29 34.98 49.28 24 34.97 3.73
ULCC2 50.0 20.5 21.0 20.8 39.77 35.02 39.44 24 35.01 5.81
ULCC3 50.1 20.5 21.5 21.0 30.22 50.00 29.91 24 49.99 7.16
ULCC4 50.1 19.5 21.2 20.4 39.94 29.99 30.36 24 29.99 11.16

13
 210 Page 14 of 19 Archives of Civil and Mechanical Engineering (2023) 23:210

Chloride migration coefficient/ steadily increased, while the average size of pores was also
12 increased, which indicated that the degradation of matrix
was more and more serious, and followed the sequence of
10
ULC0 < ULCA1 < ULCB1 < ULCC1 (see Fig. 17). This
8
result was consistent with the mechanical properties, UPV,
(×10-12 m2/s)

porosity, threshold pore diameter, and cumulative pore

6 volume of the UHPLC. When the magnification is 2000,
the C–S–H in ULC0 were bonded to each other to form a
4 dense microstructure; The slurry near the interface transi-
tion zone (ITZ) was well bonded to the aggregate; no obvi-
2 ous pores and cracks were seen on the matrix of ULC0,
and thus the microstructure of ULC0 was compact. Some
0 C–S–H were loosely distributed on the surface of ULCA1
and ULCB1; The slurry and aggregate near the ITZ were
LC 0
L 1
L 2
LC 3
U A4
L 0
L 1
L 2
LC 3
U B4
L 0
L 1
L 2
LC 3
4
U LC
U A
U CA
U CA

U LC
U CB
U CB
U CB

U LC
U CC
U CC
U CC
C
not tightly packed, and some obvious pores and cracks could
U

Different kinds of UHPLC be seen on the matrix of ULCA1 and ULCB1. There were
more pores and cracks on the surface of ULCC1, and thus
the microstructure of ULCC1 was less dense. According to
Fig. 15  Chloride migration coefficient of UHPLC with various con-
tents of AEA what is depicted in Fig. 17, as a consequence of the incor-
poration of AEA, the compactness of the UHPLC matrix
was deteriorated and the porosity of UHPLC was increased,
and 42.7%, respectively, which suggested that UHPLC due to the addition of AEA. In addition, pores with different
added with more AEA had a drop in the content of capil- sizes could be seen on the surface of the UHPLC matrix,
lary pores. The decreasing trend of the ratio of capillary and these pores were evenly distributed on the matrix of
pores in UHPLC displayed that it was negatively corre- UHPLC. When the content of AEA was 0.01%, there were
lated with porosity. The ratio of macro-pores to total pores no too much macro-pores that have a critical impact on the
was 31.8%, 49.5%, 50.0%, and 51.0%, respectively, which matrix of UHPLC with regard to its mechanical properties,
indicated that UHPLC added with more AEA had a rise in as shown in Fig. 18.
the content of macro-pores. This might be the reason why As suggested by Powers [65], cement-based composites
the increase in porosity and chloride migration coefficients have entrained air from two primary sources. The first is the
of UHPLC, and the reduction in its mechanical proper- vortex action that in-folds the air when stirring the mixture
ties and UPV. In addition, the content of AEA should not into the matrix. The second source is known as three-dimen-
exceed 0.01% in UHPLC to guarantee its fine pore struc- sional screen, in which the cascading of aggregates upon
ture and low porosity. themselves in the process of being mixed leads to the for-
mation of air bubbles. The fresh concrete, air voids are not
3.7 Microstructure stable as a result of the free surface energy [44]. AEA plays
a role in preventing voids or air bubbles from coalescing,
Figure 17 shows the SEM results of UHPLC. When the collapsing or emerging to the surface, and also hindering the
magnification is 500, the matrix of ULC0 was compact, generation of larger surrounding bubbles from the diffusion
and the size of pores in ULC0 was small (see Fig. 17a). of the air [65]. Therefore, the hardened concrete is made to
While the matrix of ULCA1, ULCB1, and ULCC1 was have homogeneously distributed pores or voids in suitable
less compact than that of ULC0, and the size of pores in sizes through being added with an air entrainment agent.
them was larger than that of ULC0, which suggested that Figure 19 illustrates the action mechanism of AEA,
the microstructure of UHPLC was deteriorated, due to the according to Metha and Monteiro [66]. Two processes are
addition of UHPLC. In general, when UHPLC was added mainly contained in the action mechanism of AEAas fol-
with an increasing amount of AEA, the number of pores was lows: the first process is dominated by the air–water phase,

13
Archives of Civil and Mechanical Engineering (2023) 23:210 Page 15 of 19 210

Fig. 16  MIP results of different kinds of UHPLC: a Porosity; b pore ▸

size distribution; c cumulative pore volume; d pore volume fraction

in which the aqueous phase-oriented polar groups foster the

generation of bubbles, cut down the surface tension and pre-
vent the already formed bubbles from dissolving [67]. At the
same time, these polar groups are adsorbed to the cement in
the solid–water interface, rendering the hydrophobic cement
surface and keeping the aqueous phase orientation of the
nonpolar groups. As suggested by Fig. 19, the water could be
displaced with the air, which is constantly connected in the
form of bubbles to the solid particles. Therefore, a structure
of homogeneously distributed pores and voids in suitable (a)
sizes could be formed by using AEA.

Log differential Intrusion/(mg-1/mL)

0.20
ULC0
When it comes to the current study, if the content of AEA ULCA1
was 0.01%, an extremely stable structure of homogeneously 0.15
ULCB1
ULCC1
distributed pores and voids in suitable sizes formed in the
matrix of UHPLC, and eventually a high porous and stable 0.10
matrix was formed in UHPLC [65], which in turn improved
the workability (see Fig. 3) and the specific strength (see 0.05
Fig. 6) of UHPLC.
0.00

4 Conclusions 100 101 102 103 104 105 106

Pore diameter/(nm)
As a novel cement-based material, UHPLC is more and more (b)
widely utilized in super high-rise, long-span structures. This 0.15
ULC0
work presented a cost-effective approach of producing UHPLC
Cumulative pore volume(mL/g)

ULCA1
via air-entraining. In this study, the mixture of UHPLC was ULCB1
ULCC1
designed according to MAA model to explore the microstruc- 0.10

ture, durability, mechanical properties and workability of

UHPLC added with three different types and four different 0.05
contents of AEA.
Compared to the UHPLC without AEA, the UHPLC con-
taining AEA was proved to have lower apparent density, lower 0.00
mechanical property (compressive strength, flexural strength),
lower stiffness (elastic modulus), and worse chloride diffu- 100 101 102 103 104 105 106
sion resistance. However, the specific strength of UHPLC was Pore diameter/(nm)
improved when the content of AEA was 0.01%. A structure of (c)
homogeneously distributed pores and voids in suitable sizes 0.15
<10 nm
was formed by adding an appropriate amount of AEA, which 10-5000 nm
0.12 >5000 nm
might be the main reason why the specific strength of UHPLC
Pore volume/(mL∙g-1)

was improved by using AEA. In addition, the optimal dosage

0.09
of AEA in UHPLC was 0.01% in comprehensive consideration
of the impacts of AEA on UHPLC in various aspects including 0.06
microstructure, durability, mechanical properties, and work-
ability. Based on the findings of this paper, some suggestions 0.03
on how to improve the mechanical property and durability of
UHPLC were put forwarded as follows: (1) using the light- 0.00
ULC0 ULCA1 ULCB1 ULCC1
weight aggregate with excellent properties, (2) utilizing the
Different kinds of UHPLC
AEA that can produce fine and well-distributed pores and (d)
voids, and (3) incorporating some nanomaterials to UHPLC.

13
210 Page 16 of 19 Archives of Civil and Mechanical Engineering (2023) 23:210

Fig. 17  SEM results of UHPLC

with various contents of AEA

13
Archives of Civil and Mechanical Engineering (2023) 23:210 Page 17 of 19 210

Author contributions HC: conceptualization, funding acquisition,

data curation, formal analysis, investigation, methodology, resources,
validation, visualization, writing—original draft, writing—review and
editing. JQ: data curation, formal analysis, investigation, methodol-
ogy, resources, supervision, validation, visualization, writing—original
draft. LG: methodology, investigation, validation, resources, super-
vision. JJ: conceptualization, investigation, methodology, validation,
funding acquisition, visualization. FW: data curation, investigation,
methodology, validation, visualization, project administration. LW:
conceptualization, methodology, validation, visualization, writing—
review and editing.

Funding The research is finically supported by the National Natu-

ral Science Foundation of China, China (No. 52278262), and the
National Science Fund for Distinguished Young Scholars, China (No.
51925903), which are gratefully acknowledged.

Data availability The data that support the findings of this study are
available from the corresponding author upon reasonable request.

Declarations
Fig. 18  SEM results of ULCC1 Conflict of interest The authors declare that they have no known com-
peting financial interests or personal relationships that could have ap-
peared to influence the work reported in this paper.

Ethical approval This article does not contain any studies with human
participants or animals performed by any of the authors.

Fig. 19  Action mechanisms of air-entrainment from anionic surfactants

13
210 Page 18 of 19 Archives of Civil and Mechanical Engineering (2023) 23:210

References 19. Xiang J, Liu L, Cui X, He Y, Zheng G, Shi C. Effect of Fuller-fine

sand on rheological, drying shrinkage, and microstructural proper-
ties of metakaolin-based geopolymer grouting materials. Cement
1. Chang W, Zheng W, Hao M. Compression behavior of ultra-high-
Concr Compos. 2019;104: 103381.
performance concrete (UHPC) confined with high-strength recti-
20. Larrard F, Sedran T. Mixture-proportioning of high-performance
linear ties. Arch Civ Mech Eng. 2022;22:27.
concrete. Cem Concr Res. 2022;32:1699–704.
2. Yu Z, Wu L, Yuan Z, Zhang C, Bangi T. Mechanical properties,
21. Lim JLG, Raman SN, Safiuddin M, Zain MFM, Hamid R. Autog-
durability and application of ultra-high-performance concrete
enous shrinkage, microstructure, and strength of ultra-high per-
containing coarse aggregate (UHPC-CA): a review. Constr Build
formance concrete incorporating carbon nanofibers. Materials.
Mater. 2022;334: 127360.
2019;12:320.
3. Fan D, Yu R, Shui Z, Liu K, Feng Y, Wang S. A new development
22. Wang XP, Yu R, Shui ZH, Zhao ZM, Song QL, Yang B, Fan DQ.
of eco-friendly ultra-high-performance concrete (UHPC): towards
Development of a novel cleaner construction product: ultra-high
efficient steel slag application and multi-objective optimization.
performance concrete incorporating lead–zinc tailings. J Clean
Constr Build Mater. 2021;306: 124913.
Prod. 2018;196:172–82.
4. Umbach C, Wetzel A, Middendorf B. Durability properties of
23. Vatannia S, Kearsleya E, Mosterta D. Development of economic,
ultra-high -performance lightweight concrete (UHPLC) with
practical and green ultra-high performance fiber reinforced con-
expanded glass. Materials. 2021;14:5817.
crete verified by particle packing model. Case Stud Constr Mater.
5. Lu J, Shen P, Ali HA, Poon CS. Mix design and performance
2020;13: e00415.
of lightweight ultra-high-performance concrete. Mater Des.
24. Li PP, Yu QL, Brouwers HJH. Effect of coarse basalt aggregates
2022;216: 110553.
on the properties of ultra-high-performance concrete (UHPC).
6. Pi Y, Zhang W, Zou W, Zhang Y. Study on the mechanical behav-
Constr Build Mater. 2018;170:649–59.
iors and failure mechanism of polyurethane cement composites
25. Sun Y, Yu R, Shui ZH, Wang XP, Qian D, Rao BY, Huang J, He
under uniaxial compression and tension. Arch Civ Mech Eng.
YJ. Understanding the porous aggregates carrier effect on reduc-
2022;22:18.
ing autogenous shrinkage of ultra-high-performance concrete
7. Hasan M, Saidi T, Afifuddin M. Mechanical properties and
(UHPC) based on response surface method. Constr Build Mater.
absorption of light weight concrete using lightweight aggregate
2019;222:130–41.
from diatomaceous earth. Constr Build Mater. 2021;277: 122324.
26. Chu H, Qin J, Gao L, Jiang J, Wang F, Wang D. Effects of gra-
8. Geng Z, Pan H, Zuo W, She W. Functionally graded lightweight
phene oxide on mechanical properties and microstructure of ultra-
cement-based composites with outstanding mechanical per-
high-performance lightweight concrete. J Sustain Cem Based
formances via additive manufacturing. Addit Manuf. 2022;56:
Mater. 2022;11(4):2104757.
102911.
27. Liu Q, Chen Z, Yang Y. Effect of fly ash on the air void size
9. Behzad N, Ravi R, Jay S, Sayanthan R. Thermal and mechanical
distribution entrained by selected anionic, cationic and nonionic
properties of sustainable lightweight strain hardening geopolymer
surfactants in hardened cement mortars. Cem Concr Compos.
composites. Arch Civ Mech Eng. 2017;17:55–64.
2021;124: 104253.
10. Huang XY, Ravi R, Zhang Q, Ni W, Li VC. Mechanical and
28. Dabrowski M, Glinicki MA, Dziedzic K, Antolik A. Valida-
thermal properties of green lightweight engineered cementitious
tion of sequential pressure method for evaluation of the content
composites. Constr Build Mater. 2013;48:954–60.
of microvoids in air entrained concrete. Constr Build Mater.
11. Zhang GZ, Ge JC, Ding QJ, Yang J, Xiang WH, Hu J. Preparation
2019;227:116633.
and mechanism of lightweight ultra-high-performance concrete. J
29. Shah HA, Yuan Q, Zuo S. Air entrainment in fresh concrete and
Chin Ceram Soc. 2021;49(2):381–90 (in Chinese).
its effects on hardened concrete—a review. Constr Build Mater.
12. Wang XP, Wu D, Gen QH, Hou DS, Wang MH, Li LW, Wang
2021;274: 121835.
P, Chen DD, Sun ZP. Characterization of sustainable ultra-high-
30. Zhang H, Gao P, Zhang Z, Pan Y, Zhang W. Effects of parameters
performance concrete (UHPC) including expanded perlite. Constr
of air-void structure on the salt-frost durability of hardened con-
Build Mater. 2021;303: 124245.
crete. Appl Sci. 2020;10:632.
13. Xie YJ, Zhou QQ, Long GC, Chaktrimongkol P, Shi Y, Umar
31. Zhang P, Li D, Qiao Y, Zhang SL, Sun CT, Zhao TJ. Effect of air
HA. Experimental investigation on mechanical property and
entrainment on the mechanical properties, chloride migration, and
microstructure of ultra-high-performance concrete with ceram-
microstructure of ordinary concrete and fly ash concrete. J Mater
site sand. Struct Concr. 2021;2021:1–14.
Civ Eng. 2018;30(10):04018265.
14. Lu JX, Shen PL, Zheng HB, Ali HA, Poon CS. Development
32. Nili M, Sasanipour H, Aslani F. The effect of fine and coarse
of high-performance lightweight concrete using ultra-high-
recycled aggregates on fresh and mechanical properties of self-
performance cementitious composite and different lightweight
compacting concrete. Materials. 2019;12:1120.
aggregates. Cem Concr Compos. 2021;124: 104277.
33. Zheng X, Ge Y, Yuan J. Influence of air content and vibration
15. Lu JX, Shen PL, Zheng HB, Ali HA, Poon CS. Development and
time on frost resistance of air entrained concrete. Adv Mater Res.
characteristics of ultra-high-performance lightweight cemen-
2014;857:110–5.
titious composites (UHP-LCCs). Cem Concr Res. 2021;145:
34. Schackow A, Effting C, Folgueras MV, Güths S, Mendes GA.
106462.
Mechanical and thermal properties of lightweight concretes with
16. Shi CJ, Wu ZM, Xiao JF, Wang DH, Huang ZY, Fang Z. A review
vermiculite and EPS using air-entraining agent. Constr Build
on ultra-high-performance concrete: part I. Raw materials and
Mater. 2014;57:190–7.
mixture design. Constr Build Mater. 2015;101:741–51.
35. Li HY, Shen XD, Zou CX. Properties of air-entrained pumice
17. Fan DQ, Yu R, Shui ZH, Wu C, Wang J, Su Q. A novel approach
lightweight aggregate concrete and a freezing-resistance forecast-
for developing a green ultra-high-performance concrete (UHPC)
ing model. J Mater Civ Eng. 2016;28(3):04015144.
with advanced particles packing meso-structure. Constr Build
36. Ragalwar K, Heard WF, Williams BA, Ranade R. Significance of
Mater. 2020;265: 120339.
the particle size distribution modulus for strain-hardening-ultra-
18. Yu R, Spiesz P, Brouwers HJH. Mix design and properties assess-
high-performance concrete (SH-UHPC) matrix design. Constr
ment of ultra-high-performance fiber reinforced concrete (UHP-
Build Mater. 2020;234: 117423.
FRC). Cem Concr Res. 2014;56:29–39.

13
Archives of Civil and Mechanical Engineering (2023) 23:210 Page 19 of 19 210

37. Yang R, Yu R, Shui ZH, Gao X, Xiao XG, Fan DQ, et al. Feasibil- 55. Su Y, Wu C, Li J, Li Z, Li W. Development of novel ultra-high
ity analysis of treating recycled rock dust as an environmentally performance concrete: from material to structure. Constr Build
friendly alternative material in ultra-high-performance concrete Mater. 2017;135:517–28.
(UHPC). J Clean Prod. 2020;258: 120673. 56. Ma QM, Guo RX, Zhao ZM, Lin ZW, He KC. Mechanical prop-
38. Chu H, Gao L, Qin J, Jiang J, Wang D. Mechanical properties erties of concrete at high temperature—a review. Constr Build
and microstructure of ultra-high-performance concrete with high Mater. 2015;93:371–83.
elastic modulus. Constr Build Mater. 2022;335: 127385. 57. Hoang AL, Fehling E. Influence of steel fiber content and
39. GB/T2419-2005, Standard for test method for fluidity of cement aspect ratio on the uniaxial tensile and compressive behav-
mortar, general administration of quality supervision. Inspection ior of ultra-high-performance concrete. Constr Build Mater.
and Quarantine, P.R. China, 2005 2017;153:790–806.
40. GB/T50081-2019, Standard for test methods of concrete physi- 58. Wang Z, Wang J, Liu T, Zhang F. Modeling seismic performance
cal and mechanical properties, general administration of quality of high-strength steel-ultra-high-performance concrete piers with
supervision. Inspection and Quarantine, P.R. China, 2019. modified Kent–Park model using fiber elements. Adv Mech Eng.
41. GB/T17671–1999. Standard for test method of testing cements 2016;8(2):1–13.
determination of strength. State Bureau of Quality Technical 59. Alsalman A, Dang CN, Prinz GS. Evaluation of modulus of elas-
Supervision, P. R. China, 1999. ticity of ultra-high-performance concrete. Constr Build Mater.
42. Zhang WH, Zhang ZX, Liu PY, Zhang YS, Zhang CX, She W. 2017;153:918–28.
Uniaxial tensile and compressive stress–strain behavior of multi 60. Nguyen T, Thai H, Ngo T. Optimised mix design and elastic
scale fiber reinforced ultra-high-performance concrete. J Chin modulus prediction of ultra-high strength concrete. Constr Build
Ceram Soc. 2020;08:1155–67 (in Chinese). Mater. 2021;302: 124150.
43. CECS 02:2005, Technical specification for detecting strength of 61. Wang Y, Zhang H, Gen Y, Wang Q, Zhang S. Prediction of the
concrete by ultrasonic rebound combined method. China Associa- elastic modulus and the splitting tensile strength of concrete incor-
tion for Engineering Construction Standardization. Standard of porating both fine and coarse recycled aggregate. Constr Build
China Association for Engineering Construction Standardization, Mater. 2019;215:332–46.
P.R. China, 2005. 62. Jurowski K, Grzeszczyk S. Influence of selected factors on the
44. GB/T 50082-2009, Standard for test methods of long-term perfor- relationship between the dynamic elastic modulus and compres-
mance and durability of ordinary concrete, Beijing, China, 2009. sive strength of concrete. Materials. 2018;11:477.
45. European Project Group. Specification and guidelines for self- 63. Chen Y, Yu R, Wang X, Chen J, Shui Z. Evaluation and opti-
compacting concrete. UK: EFNARC; 2002. mization of ultra-high-performance concrete (UHPC) subjected
46. Du LX, Folliard KJ. Mechanisms of air entrainment in concrete. to harsh ocean environment: towards an application of layered
Cem Concr Res. 2005;35:1463–71. double hydroxides (LDHs). Constr Build Mater. 2018;177:51–62.
47. Kim HK, Jeon JH, Lee HK. Workability, and mechanical, 64. He ZH, Du SG, Chen D. Microstructure of ultra-high-per-
acoustic and thermal properties of lightweight aggregate con- formance concrete containing lithium slag. J Hazard Mater.
crete with a high volume of entrained air. Constr Build Mater. 2018;353:35–43.
2012;29:193–200. 65. Mendes JC, Moro TK, Figueiredo AS, Silva KDC, Silva GC.
48. Zhu JC. Effect of air entraining agent on uniaxial tensile properties Mechanical, rheological and morphological analysis of cement-
of PVA-ECC. Mater Sci Eng. 2018;409: 012029. based composites with a new LAS-based air entraining agent.
49. Chia KS, Zhang MH. Workability of air-entrained light- Constr Build Mater. 2017;145:648–61.
weight concrete from rheology perspective. Mag Concr Res. 66. Mehta PK, Monteiro PJM. Concrete: microstructure, properties
2007;59(5):367–75. and materials. 4th ed. New York: McGraw-Hill; 2014.
50. Salem MAA, Pandey RK. Effect of air entrainment on compres- 67. Lea FM. The chemistry of cement and concrete. 3rd ed. New
sive strength, density and ingredients of concrete. Int J Adv Mech York: Chemical Publishing Company; 1971.
Civ Eng. 2017;4(6):2394–827.
51. Bai R, Cai J, Wu J, Wei G. Influence of air content on the compres- Publisher's Note Springer Nature remains neutral with regard to
sive strength of concrete. Adv Mater Res. 2012;535–537:1790–3. jurisdictional claims in published maps and institutional affiliations.
52. Guo P, Meng W, Du J, Stevenson L, Han B, Bao Y. Lightweight
ultra-high-performance concrete (UHPC) with expanded glass Springer Nature or its licensor (e.g. a society or other partner) holds
aggregate: development, characterization, and life-cycle assess- exclusive rights to this article under a publishing agreement with the
ment. Constr Build Mater. 2023;371: 130441. author(s) or other rightsholder(s); author self-archiving of the accepted
53. Guo K, Ding Q. Effect of shale powder on the performance manuscript version of this article is solely governed by the terms of
of lightweight ultra-high-performance concrete. Materials. such publishing agreement and applicable law.
2022;15:7225.
54. Meng W, Khayat K. Effects of saturated lightweight sand content
on key characteristics of ultra-high-performance concrete. Cem
Concr Res. 2017;101:46–54.

13

View publication stats