Journal of Building Engineering 82 (2024) 108317

Contents lists available at ScienceDirect

Journal of Building Engineering

journal homepage: www.elsevier.com/locate/jobe

Rapid chloride penetration test: An evaluation of corrosion

resistance in ultra-high performance concrete
Tawsif Mohammad Hasan a, Srinivas Allena a, *, Levi Gilbert b
a
Cleveland State University, Department of Civil and Environmental Engineering, United States
b
Washington State University, Department of Civil and Environmental Engineering, United States

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

Keywords: Six Ultra-High Performance Concrete (UHPC) mixtures were tested to determine their chloride
Ultra-high performance concrete ion ingress potential using the rapid chloride penetration test (RCPT). The UHPC mixtures
Corrosion showed low to negligible chloride ion penetrability, depending on their water to cementitious
Chloride diffusion materials (w/cm) ratio and curing regimen. Electrical resistivity and conductivity of the mixtures
Electrical resistivity were determined based on the recorded electric current during the RCPT test. The electrical re­
sistivity varied from 49 to 1419 Ω-meter, while conductivity ranged from 0.0007 to 0.0013 milli-
Siemens/meter, indicating a moderate to very low expected corrosion rate for the mixtures. The
chloride diffusion coefficient was calculated using the Nernst-Plank Equation, making corrections
for Joule effect and temperature adjustment due to the temperature increase of the electrode
solution during the RCPT test. The chloride ion diffusion coefficients of the UHPC mixtures
ranged from 0.53–3.01 × 10− 15 m2/s. These results suggest that UHPC mixtures with a w/cm
ratio as high as 0.30 can exhibit excellent corrosion resistance.

1. Introduction
Ultra-high performance concrete (UHPC) is a modern type of concrete that exhibits exceptional mechanical and durability prop­
erties, including compressive strength greater than 120 MPa, high ductility when fiber reinforced, and excellent resistance to abrasion,
alkali-silica reaction, frost damage, and corrosion of reinforcing. Numerous studies have confirmed the excellent durability charac­
teristics of UHPC which include, but not limited to, excellent resistance to frost damage, alkali-silica reaction very low permeability
and negligible chloride ion penetrability [1–4]. The dense microstructure of UHPC impedes the ingress and movement of deleterious
ions through the cementitious matrix, particularly the movement of chloride ions, which is a primary factor in evaluating the
serviceability of a structure due to its interaction with steel.
The presence of chloride ions in the pore solution of concrete poses a significant threat to steel rebars and fibers, and their
penetration is a major cause of corrosion initiation in reinforced concrete (RC) structures, leading to premature degradation [5,6]. As
chloride ions ingress into cement-based materials, a portion of the penetrated chloride can become bound by the hydration products,
resulting in the loss of its ability to migrate. The remaining chloride, often referred to as free chloride, continues to diffuse deeper into
the cement matrix through cracks and the interconnected pore network [7,8]. Chloride ions may enter concrete through the concrete
mixture constituents and external environmental sources, and they can interact with steel reinforcement through a range of trans­
portation mechanisms, such as diffusion, migration, capillary sorption (or convection), and permeation [9–11]. These complex

* Corresponding author.
E-mail addresses: t.hasan11@vikes.csuohio.edu (T.M. Hasan), s.allena@csuohio.edu (S. Allena), l.gilbert189@gmail.com (L. Gilbert).

https://doi.org/10.1016/j.jobe.2023.108317
Received 3 August 2023; Received in revised form 5 December 2023; Accepted 12 December 2023
Available online 16 December 2023
2352-7102/© 2023 Elsevier Ltd. All rights reserved.
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

physical-chemical reactions of chloride ions with cement hydration products result in the deterioration of the concrete microstructure
[12]. Such reactions have a direct impact on the compressive strength and damage characteristics of the concrete. Chloride ions
interact with cement hydration products both physically and chemically. Physical interaction involves their adsorption by calcium
silicate hydrate (C─S─H), leading to the development of a cellular and porous structure, which contributes to the fracture of concrete
structures. Additionally, chloride ions chemically react with cementitious materials, leading to the formation of compounds like
Friedel’s salt and Kuzel’s salt [11,13,14] which can increase the chloride binding capacity of portland cement [15].
Chloride ingress is affected by the pore structure, interactions with the pore walls and interactions with the other ions in the so­
lution [16,17]. The pore structure depends on factors such as the water-to-cement ratio, the inclusion of supplementary cementitious
materials (SCM), and the concrete’s age. The effect of SCMs on compacting the pore structure and influencing the movement of
chloride ions has been extensively studied [18–20]. The chloride ions may interact with the pore walls in two ways: chemically, where
chloride predominantly binds to the aluminate phases, or physically, where surface forces hold the chlorides to the pore walls [21–23].
The measurement of chloride transport parameters can be performed using variety of test methods such ASTM C1202, rapid chloride
permeability test (RCPT) [24] and ASTM C 1556, bulk diffusion test [25]. Nonetheless, these testing methods only offer an indication
of concrete’s resistance to chloride transport, without the ability to characterize the microstructure of concrete, which in turn affects
parameters like chloride diffusion, electrical resistivity, and conductivity. Electrical impedance spectroscopy, a potent technique for
analyzing the microstructure of cement-based materials [26,27], has been employed to assess the diffusion of chloride in concrete [16,
28,29]. Loche et al. [30] and Sánchez et al. [31] employed electrical impedance to observe the microstructural changes taking place
during and after chloride penetration. Loche et al. [30] demonstrated that the material’s electrical capacity is primarily influenced by
the exchange surface rate between the solid phase and the electrolyte within the pores, rather than the cross-sectional area of the tested
sample. Sánchez et al. [31] also used the technique of electrical impedance to study the microstructural changes during chloride
migration and reported an important finding regarding pore size. They observed a reduction in pore diameter attributed to the for­
mation of new solid phases at the pore walls due to chloride migration.
The bulk chloride diffusion coefficient of concrete can be determined using ASTM C1556 [25], which involves submerging a
concrete specimen in a chloride solution for an extended period. At the end of the test, the chloride content at different depths is
measured, and the diffusion coefficient is calculated using Fick’s law. Another method for assessing the potential for chloride ion
ingress is the rapid chloride ion penetration test. Andrade [32] demonstrated that the chloride ion diffusion coefficient can be pre­
dicted from RCPT data using the Nernst-Einstein and Nernst-Plank equations, which assume steady-state flow conditions, no Joule
effect, and no chemical reaction between chloride ions and hydration products. Zhang and Gjory [33] proposed another equation that
considers the known chloride migration rate. Berke and Hicks [34] also developed an empirical equation for predicting the chloride ion
diffusion coefficient using RCPT test data. While Andrade [32] conducted a comparison between the Nernst-Einstein and Nernst-Plank
equations, Zhang and Gjory [33] proposed an ionic interaction correction factor for Chloride ions at varying concentration levels in
sodium chloride solution. This correction factor improved the accuracy of predicting the chloride ion diffusion coefficient. Conversely,
Berke and Hicks [34] demonstrated the validity of Fick’s diffusion equation for predicting the chloride ion diffusion coefficient. They
developed an empirical equation by correlating these diffusion coefficients with corresponding RCPT values at ambient temperature.
Notably, they also identified that the equation becomes invalid for RCPT values exceeding 2000 Coulombs, as the temperature of the
solution deviates from ambient conditions, subsequently affecting the chloride diffusion coefficient.
A challenge with predicting the diffusion coefficient from RCPT test data is the Joule effect. As electricity passes through the test
specimen, the temperature of the electrode solution can increase, which may lead to higher current flow and larger charge passed
values. Julio-Betancourt and Hooton [35] proposed a method for adjusting the total charge passed value in a 6-h RCPT test to account
for the Joule effect. However, a limitation of the RCPT test is that the predicted diffusion coefficient is dependent on the test’s final
temperature. Diffusion coefficient is sensitive to temperature, and a 10 ◦ C change can result in a doubling of the coefficient [36]. For
comparison, the estimated diffusion coefficient value must be converted to a reference temperature of 21 ◦ C. An Arrhenius correction
factor can be applied to achieve this [37]. Ghosh et al. [37] emphasized the significance of these two correction factors to accurately
predict the chloride diffusion coefficient. Although numerous studies have been conducted on utilizing the RCPT test to predict
chloride diffusion properties in conventional concrete, the application of this method in the context of UHPC research is relatively
scarce. In the case of concrete reinforced with steel fibers, such as UHPC, the corrosion of steel fibers is a critical issue. Steel fibers are
added to enhance the concrete’s tensile performance and ductility, but if compromised, it can potentially shorten the structure’s
service life. To estimate the service life, it is necessary to determine resistivity, conductivity, and chloride diffusion coefficient.
Electrical resistance and conductivity represent inherent characteristics of materials, and they can be assessed to determine the
durability of concrete.
In this study, the objective was twofold: to evaluate the corrosion resistance of six different UHPC mixtures, and to determine the
parameters required for predicting their service life. These parameters include the chloride diffusion coefficient, porosity, permeable
porosity, electrical resistivity and conductivity, and cumulative chloride mass. Four different curing methods were used to determine
their impact on UHPC’s corrosion resistance. The fresh UHPC was subjected to a flow test, and the compressive strength and permeable
porosity were determined after 28 days of curing in specified curing regimens as explained in section 2.3. Additionally, RCPT testing
was conducted according to ASTM C1202 [24]. The resulting RCPT test data were analyzed to determine the electrical resistivity,
conductivity, cumulative chloride mass, and chloride ion diffusion coefficient of these UHPC mixtures.

2. Materials and experimental methods

This section presents the materials utilized in the UHPC mixtures, as well as the experimental methods employed to assess both the

2
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

fresh and hardened properties of the UHPC.

2.1. Materials
Type I - II portland cement and commercially available dry densified silica fume with a specific gravity of 3.15 and 2.24,
respectively, were used in this study. Table 1 presents the chemical compositions and physical properties of the cement and silica fume.
Natural sand with particle sizes ranging between 75 and 600 μm was used, and the specific gravity and absorption value of the sand
were 2.78 and 2.35%, respectively. To obtain this particle size range, the sand was first scalped through an ASTM No. 30 sieve and then
washed over an ASTM No. 200 sieve to remove deleterious material and dust. The washed sand was kept in an oven for 24 h at 105 ◦ C
to achieve a 0% moisture content. Steel fibers, 13 mm long with an aspect ratio of 65, were added to the mixture to improve tensile
ductility. A commercially available polycarboxylate-based high-range water-reducing admixture (HRWRA) was used to achieve the
desired workability.

2.2. Mixture proportions

Six UHPC mixtures were investigated in this study, and their mix proportions are presented in Table 2 which were proposed by
Hasan et al. [38]. The mixtures were divided into three categories: A, B, and C based on the silica fume content and
water-to-cementitious materials ratio (w/cm). Each category consisted of two mixtures: one plain mixture without steel fibers and one
fiber-reinforced mixture with 1.5% steel fibers by volume. The volume of steel fibers added was equivalent to the volume of sand in the
mixture. The mixtures were identified using two-digit symbols: the first digit represented the category, and the second digit indicated
the type (plain or fiber-reinforced). For example, mixture AP was a plain mixture from category A, while mixture AF was a
fiber-reinforced mixture from the same category.

2.3. Specimen preparation and curing

This section outlines the mixing method, specimen preparation, and curing process used for the UHPC samples.

2.3.1. Mixing and specimen preparation

A vertical shaft mixer with a paddle speed of 38 rpm was used, and it took 15–20 min to prepare each UHPC mixture. The mixing
procedure was outlined in Fig. 1. The workability of the fresh UHPC mixtures was evaluated using a flow table test according to ASTM
C1437 [39].
50 mm cubes and discs with a thickness of 50 mm and a diameter of 100 mm were prepared for each mixture. A high-frequency
vibrating table was used to consolidate the specimens for 30–60 s, depending on the consistency of the mixture. Molds containing the
fresh mixture were covered with plastic sheets to prevent moisture loss and were kept at room temperature for 24 h. After demolding,
specimens were cured under four different curing regimens for 27 days.

2.3.2. Curing
The first curing method was standard water immersion curing (SC), where specimens were immersed in water at room temperature
until the day of testing. The second method was moist curing (MC), where the specimens were kept in a sealed plastic container with
water at the bottom and a screen separating the specimens from the water to prevent direct contact. The container was sealed to
prevent moisture loss and kept at room temperature until the day of testing. The other two methods involved elevated temperature
curing at 90 ◦ C until the day of testing. The first method was warm bath curing (WB), where specimens were placed in a container full
of lime-saturated water and kept inside an oven at 90 ◦ C. Water was refilled every day to prevent exposure to dry heat. The second
method was heat curing (HC), which was a modification of WB. The specimens were initially placed in WB and then removed two days
before the testing date and placed in the oven at 90 ◦ C for dry curing until the day of testing.

2.4. Workability
The workability of freshly mixed UHPC was measured using a flow table test, in accordance with ASTM C1437 [39]. During the test,
the freshly mixed UHPC was poured into a mold and lifted vertically. Two perpendicular diameters were measured during the test, and
their average was determined and reported as the static flow of UHPC. Next, the flow table was dropped 25 times in 15 s, and the
average spread diameter was measured and reported as the dynamic flow of UHPC. The test arrangement and the flow of UHPC
resulting from the test are shown in Fig. 2.

2.5. Compressive strength testing

The compressive strength of each mixture under each curing regimen was determined using 50 mm cubes as per ASTM C109 [40]
on the 28th day. The average of strengths from three 50 mm cubes was reported as the compressive strength for each mixture.

Table 1
Chemical compositions and physical properties of cementitious materials.

Properties Chemical Properties (%) Physical Properties

SiO2 Al2O3 Fe2O3 CaO MgO SO3 Na2O K2O Relative density Moisture content (%)

Cement 22.3 3.40 3.10 64.30 2.30 2.50 0.22 0.47 3.15 –
Silica fume 96.90 0.20 0.20 0.30 0.20 0.10 0.20 0.30 2.24 0.04

3
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

Table 2
Mixture proportions of UHPC mixtures.

Mixture Category A B C

Mixture Designation AP AF BP BF CP CF
Cement (kg/m3) 890 890 890 890 890 890
Silica fume (kg/m3) 223 223 178 178 169 169
Fine sand (kg/m3) 592 549 982 939 1059 1016
Steel fiber (kg/m3) 0 119 0 119 0 119
HRWRA (L/m3) 50 50 50 50 59 59
Water (kg/m3) 334 334 214 214 180 180
w/cm 0.30 0.30 0.20 0.20 0.17 0.17
Flow diameter (mm) Static >250 230 >250 230 >250 230
Dynamic >250 >250 >250 >250 >250 >250

Fig. 1. Mixing procedure for UHPC.

Fig. 2. Flow table test for workability of UHPC mixtures; (a) test arrangement and (b) static flow.

2.6. Permeable porosity

To determine the permeable porosity of the UHPC mixtures under each curing regimen, two-disc specimens were used. The method
described in Ref. [41] was followed for this test. On the 28th day after mixing, the specimens were removed from the curing regimens
and submerged in water at a temperature of 23 ± 1.5 ◦ C for 24 h. The submerged water mass (mw ) was then determined. Next, the
specimens were air-dried and the mass of the specimen at saturated surface dry condition (ms ) was determined. The specimens were
then placed in an oven at a temperature of 90 ◦ C for 24 h to achieve an oven-dry condition. The mass of the specimen at the oven-dry
condition (md ) was determined, and the permeable porosity (φ) was calculated using Equation (1).
ms − md
φ= × 100 (1)
ms − mw

2.7. Rapid chloride penetration test (RCPT)

RCPT testing was conducted following the procedure outlined in ASTM C1202 [24]. Three disc-shaped specimens were
vacuum-saturated on the 28th day after mixing. The specimens were then mounted on test cells. One half of each cell was filled with a
3% sodium chloride solution, while the other half was filled with a 0.3 N sodium hydroxide solution. The negative terminal of an
electric circuit was connected to the sodium chloride solution, while the positive terminal was connected to the sodium hydroxide
solution.
A temperature sensor was placed inside the sodium chloride solution to monitor the temperature, and both the solution’s tem­
perature and the electric current flowing through the circuit were recorded every minute. A potential difference of 60 V was applied for
a duration of 6 h. The total electric charge passed through each specimen was determined based on the measurements, and the average
value of the three specimens was recorded as the total charge passed for that particular mixture.

3. Results and discussion

This section discusses the results obtained from the laboratory experiments and analytical calculations. It should be noted that the

4
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

primary objective of this work is to establish the applicability of the RCPT test data for predicting the electrical properties and chloride
diffusion coefficient of UHPC. The presentation of mechanical and physical properties in this paper serves to demonstrate that our
mixtures conform to the UHPC class standards outlined in ASTM C1856.

3.1. Workability
The flow diameters of the fresh UHPC mixtures obtained from the flow table test are presented in Table 2. The plain UHPC mixtures
resulted in an overflow of the standard flow table, while the fiber-reinforced mixtures exhibited a flow diameter larger than 225 mm.
As a result, all of the mixtures met the minimum workability requirement of 200 mm flow diameter, as defined in ASTM C1856 [42] for
UHPC-class materials.

3.2. Compressive strength

28-day compressive strengths of UHPC mixtures cured under different curing regimens are depicted in Fig. 3. It can be observed
from Fig. 3 that AP and AF mixtures achieved a compressive strength greater than the minimum required compressive strength for
UHPC, as per ASTM C 1856 [42], under HC curing regimen BP and CP mixtures showed similar strength across all curing regimens
except for specimens cured under the SC regimen. However, BF and CF mixtures achieved compressive strength greater than 120 MPa
in all curing regimens. Mixtures in category A had the highest w/cm ratio, whereas category C mixtures had the lowest. Therefore,
elevated temperature curing was necessary to achieve ultra-high strength for mixtures with a higher w/cm ratio. Nevertheless, the heat
treatment could cause thermal cracking upon removing the specimens from the curing container. This could be the reason for the
occasional increase and decrease in compressive strength when cured under WB and HC curing regimens. Additionally, the specimens
of the AF mixture, when cured under the WB regimen, showed a significant decrease in compressive strength. This could be due to the
overconsolidation of the material for which the fibers settled down at the bottom of the specimen. Thus, it created a weaker top zone,
which could be further deteriorated by the heat thermal cracking that could led to early failure.

3.3. Permeable porosity

As shown in Fig. 4, category A mixtures had the highest permeable porosity, which ranged from 7.4% to 10.5% and 8.7%–16.7% for
AP and AF mixtures, respectively. On average, BP and CP mixtures exhibited a 61% and 69% decrease in permeable porosity compared
to AP mixtures, respectively. For BF and CF mixtures, the reduction was 68% and 78%, respectively, compared to AF mixtures. This
reduction can be attributed to the very low w/cm ratio of category B and C mixtures.
In Fig. 4, it can be seen that specimens cured under the SC regimen had the highest permeable porosity for all mixtures. On average,
MC-cured specimens showed a 9% and 12% reduction of permeable porosity for plain and fiber-reinforced mixtures, respectively. The
entrapped heat inside the curing container of the MC regimen might have accelerated the hydration reaction, causing more silica fume
to react with primary hydration products. As a result, more secondary calcium-silicate-hydrate gel might have formed [43] and filled
the pores. However, when cured under the MC regimen, the mixtures CP and BF did not conform to this finding. They exhibited higher
permeable porosity than the corresponding SC specimens, which needs to be studied further to confirm the trend.
WB and HC regimens exhibited greater reduction in permeable porosity, although the performance varied for mixture type. On
average, compared to SC-cured specimens, plain mixtures achieved a 37% and 27% lower permeable porosity in WB and HC regimens,
respectively. For similar conditions in fiber-reinforced mixtures, the reduction was 18% and 48%, respectively. The dry curing
associated with the HC regimen might have caused thermal cracks in specimens of plain mixtures, reducing the effectiveness of the HC
regimen when compared to corresponding fiber-reinforced mixtures. The thermal crack could be the reason for the higher permeable
porosity of AP mixtures when cured under an HC regimen. However, the specimens of AF mixture cured under the WB regimen also
exhibited similar permeable porosity as their SC cured counterparts. This could be due to the imperfections mentioned in the previous
section 3.2.

Fig. 3. 28-day compressive strength of UHPC mixtures and.

5
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

Fig. 4. Permeable porosity of UHPC mixtures.

3.4. Rapid chloride permeability testing (RCPT)

Fig. 5 depicts the average charge passed values obtained from the RCPT test. The maximum values were observed for category A
mixtures, namely AP and AF. The range of charge passed values varied between 122 and 1055 Coulomb depending on the mixture type
and curing regimen. According to the guideline in ASTM C1202 [24], these values indicate low to very low chloride ion penetrability
for the aforementioned mixtures. Conversely, mixtures from category B and C exhibited charge passed values between 61 and 685
Coulomb. As per ASTM C1202, their chloride ion penetrability was classified as very low to negligible. On average, BP and CP mixtures
had 77% and 88% lower charge passed values than AP mixture, respectively. On the other hand, BF and CF mixtures had a reduction of
66% and 54%, respectively, compared to AF mixtures. Since the permeable porosity of category A mixtures was higher than that of
category B and C mixtures, higher chloride ion penetrability was observed for category A mixtures.

3.4.1. Effect of curing regimens

The recorded charge passed value is dependent on the permeable porosity of the mixture, as shown in Fig. 6. The permeable
porosity has a significant impact on the RCPT (Rapid Chloride Permeability Test) results. A higher permeable porosity is associated
with a greater charge passed value. For instance, in mixture category B, the BP mixture exhibited the highest permeable porosity
(5.9%) and also had the highest RCPT value (685 Coulombs) when cured under the SC regimen. When the permeable porosity was
reduced to 3.7% under the MC regimen, the RCPT value decreased to 117 Coulombs. Further reduction of permeable porosity was
observed for accelerated curing regimens WB and HC, where the permeable porosity obtained was 2.9% and 2.8%, respectively.
Consequently, the corresponding RCPT values were 36 Coulombs and 72 Coulombs, respectively. This indicates that the change in
RCPT value is proportional to the change in permeable porosity. Similar trends were observed for all other mixtures.
The choice of curing regimen exerts a substantial influence on both permeable porosity and RCPT values of the mixtures. Spe­
cifically, the SC regimen tends to yield higher permeable porosity and RCPT values, while the MC, WB, and HC regimens result in
progressively lower permeable porosity and corresponding RCPT values (as depicted in Fig. 6). This observation can be attributed to
the slope of the correlation lines drawn for the dataset. For instance, under the SC regimen (Fig. 6a), the AP mixture displays the
highest permeable porosity and charge passed value, while the CP mixture exhibits the lowest values for both parameters. The BP
mixture falls in between, resulting in a correlation line with a slope of 105.55. Under the MC regimen (Fig. 6b), the slope decreases to
36.352, which is 65.6% less steep than the SC regimen. Likewise, for the WB and HC regimens, the slopes of the correlation lines are
reduced by 72.5% (Figs. 6c) and 86.7% (Fig. 6d) compared to the corresponding SC regimen. The behavior of fiber-reinforced mixtures
generally follows this trend, except in the case of the HC regimen, where the slope of the correlation line is steeper than that of the SC

Fig. 5. Total charge passed values obtained from RCPT test.

6
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

Fig. 6. Relationship between charge passed value and permeable porosity for different curing regimens: (a) SC, (b) MC, (c) WB, and (d) HC.

regimen. This discrepancy could be attributed to an increased electric current passing through the mixtures, potentially due to the
fibers creating a continuous flow path. However, further testing is required to substantiate this hypothesis.
UHPC mixtures with greater permeable porosity may have a higher number of interconnected voids, which can lead to a higher
ingress of chloride ions into the material [44]. As a result, the material becomes more susceptible to chloride ion-induced corrosion,
which leads to a higher RCPT charge passed value. In other words, the higher the permeable porosity, the easier it is for chloride ions to
penetrate the material, leading to a higher charge passed value. As can be seen from Fig. 6, a direct correlation between permeable
porosity and RCPT charge passed value in UHPC mixtures.

3.4.2. Effect of steel fibers

Fig. 6 also shows that the fiber reinforced mixtures exhibited higher charge passed values compared to their plain counterparts. On
average, AF, BF, and CF mixtures showed 2.4, 2.1, and 2.8 times higher charge passed values compared to AP, BP, and CP mixtures,
respectively. The presence of steel fibers might have improved the electrical conductance of the fiber reinforced mixtures, leading to a
higher amount of electric charge conducted. As a result, higher maximum solution temperatures were observed for the fiber reinforced
mixtures (Fig. 7). On average, AF, BF, and CF mixtures showed a 20%, 22%, and 13% increase in maximum solution temperature
compared to AP, BP, and CP mixtures, respectively. It should be noted that ASTM C1202 [24] cautions that specimens containing steel
fibers or other embedded electrically conductive materials may yield inaccurate results in the RCPT test, showing artificially high total
charge passed and low surface resistivity values. Nevertheless, the relatively very low RCPT values (less than 800 coulombs) suggest
that the possibility of a conductive path forming between the two ends of the specimen is unlikely. Several studies by other researchers
have also confirmed that steel fibers in UHPC are typically dispersed and discontinuous, lacking a direct pathway to complete an
electric circuit [45,46].

Fig. 7. Maximum solution temperature recorded during RCPT test.

7
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

3.4.3. Electrical properties of UHPC mixtures

The electrical properties of UHPC mixtures were determined by calculating the electrical resistivity and conductivity using data
obtained from the rapid chloride penetration test (RCPT). The findings of this analysis are presented and discussed in this section.
3.4.3.1. Electrical resistivity. Electrical resistivity is a fundamental parameter of a material that indicates how easily electric current
can pass through it. The higher the resistivity, the lower the current that can pass through the material. During the RCPT test, the
electric current (I) passing through the UHPC specimens was recorded at 1-min intervals for 6 h. All specimens had the same cross-
sectional area (A) and thickness (t), while the potential difference (V) was kept constant at 60 V. The electrical resistivity (ρ) can
be calculated using Equation (2).
VA
ρ= (2)
It
Fig. 8 shows the average resistivity values for each mixture cured under all four curing regimens. The average resistivity is an
important parameter to evaluate the corrosion likelihood of embedded steel in concrete. As per Broomfield [47], when the resistivity is
within 50–100 Ohm-m, the corrosion likelihood is high. This was observed for AP, BP, and CF mixtures cured under the SC regimen,
and for AF mixture cured under all four curing regimens. On the other hand, MC cured specimens of AP and CF mixtures showed
average resistivity between 100 and 200 Ohm-m, indicating low corrosion likelihood. When the average resistivity is higher than 200
Ohm-m, the corrosion likelihood is negligible [47], which was observed in the remaining cases.
3.4.3.2. Electrical conductivity. Electrical conductivity (σ ) is a measure of a material’s ability to conduct electric current. It is the
inverse of electrical resistivity and can be calculated by equation (3).
1
σ= (3)
ρ
Average conductivity of the mixtures calculated from average resistivity was shown in Fig. 9.
As can be seen from Fig. 9, the conductivity of fiber-reinforced mixtures is higher than that of the corresponding plain mixtures. The
addition of steel fibers to UHPC can greatly enhance its electrical conductivity. The present finding aligns with earlier studies [48,49]
that suggest the electrical conductance of UHPC mixtures can be improved by incorporating steel fibers.

3.4.4. Chloride migration properties

In this section, the chloride migration properties of the UHPC mixtures, including the cumulative chloride mass and chloride ion
diffusion coefficient, are presented and discussed based on the calculated results.
3.4.4.1. Cumulative chloride mass. As the electric current was recorded at fixed intervals during the test, this information can be
utilized to determine the quantity of chloride mass that deposited or penetrated through the specimens. The cumulative chloride mass
associated with the test can be calculated using Equation (4), which considers the total charge passed during the test (Q), the molar
mass of chlorides (MCl), and the Faraday constant (F) of 96,485 Coulombs/mole.
Q MCL
M= (4)
F
Fig. 10 shows the cumulative chloride masses calculated for each mixture using Equation (4). The specimens cured under SC and
MC conditions exhibit greater cumulative chloride mass than those subjected to elevated temperature curing conditions. The increased
chloride mass in SC and MC-cured specimens could be attributed to greater hydration of the cementitious materials, which results in a
more porous and permeable matrix. This increased porosity can make the concrete more vulnerable to chloride ingress [50,51].
Furthermore, the presence of moisture on the surface of the specimens during the curing process could enhance the dissolution and
transport of chlorides into the concrete matrix, contributing to the higher cumulative chloride mass observed in SC and MC-cured
specimens [52].
The relationship between permeable cumulative chloride mass and permeable porosity has been investigated and is presented in

Fig. 8. Average resistivity values for plain and fiber reinforced UHPC mixtures cured under different curing regimens.

8
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

Fig. 9. Average conductivity values for plain and fiber reinforced UHPC mixtures cured under different curing regimens.

Fig. 10. Cumulative chloride mass calculated from RCPT test data.

Fig. 11. Relationship between cumulative chloride mass and permeable porosity.

Fig. 11. As shown in the figure, there is a linear correlation between cumulative chloride mass and permeable porosity for both plain
and fiber reinforced UHPC mixtures. There is a positive correlation can be seen between permeable porosity and cumulative chloride
mass in UHPC [9]. This means that as the permeable porosity of concrete increases, the amount of chloride ions that can penetrate into
the concrete also increase.
3.4.4.2. Chloride ion diffusion coefficient. The total charge passed values (Q6) obtained from 6 h RCPT test for different mixtures under
different curing regimens were adjusted for joule effect using Equation (5) [35]. In Equation (5), ∂T represents the temperature
increment in Kelvin, Qo denotes the charge passed value adjusted for joule effect, and β is an experimental constant equal to 1245 [35].
[ ( )]
ln(Q6 )+β 1 1
∂T −
(5)
273
Q0 = e

9
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

After adjusting the total charge passed values, the equivalent chloride ion diffusion coefficients (D) were calculated using the
Nernst-Plank equation (Equation (6)) [32].
RTKV
D= ( ) (6)
z C0 F EL A

where, R is the ideal gas constant (R = 8.314 J/mole/K), T is the final solution temperature in Kelvin, V is the volume of the chloride
ion solution (V = 0.25 L), z is the valency of the chloride ion (z = 1), C0 is the chloride ion concentration (C0 = 2.05 mol/L), E is the
electric potential (E = 60V), L is the thickness of the specimens, and A is the cross-sectional area of the specimens. The migration rate
(K) of chloride ion was calculated using Equation (7) [32], where J is the flux of chloride ion, and I is the electric current.
JA I
K= = (7)
V zFV
As Equation (6) considers the final temperature of the solution at the end of RCPT, and the final temperature was different for each
mixture (Fig. 12), an Arrhenius correction factor (Equation (8)) was used to convert the diffusion coefficient to room temperature
(21 ◦ C) [37].
( )
( ) U 1 1
294 R 273+T − 294
D2 = D1 e (8)
T

where, D2 is the chloride ion diffusion coefficient at 21 ◦ C, D1 is the chloride ion diffusion coefficient at temperature T◦ C, U is the
activation energy, and R is the gas constant.
The final chloride ion diffusion coefficients (D2) are presented in Fig. 13. As shown in the figure, the D2 value of the AP mixture was
the highest (3.01 × 10− 15 m2/s) when cured under the SC regimen, while the lowest value (0.53 × 10− 15 m2/s) was observed when the
same mixture was cured under the HC regimen. The other two curing regimens produced D2 values falling between these two extremes.
For the AF mixture, the D2 value was within the range of the aforementioned values, regardless of the curing regimen used to produce
the specimens. Consequently, the range of D2 values for the mixtures in category A was 0.53–3.01 × 10− 15 m2/s. On the other hand, the
ranges for category B and C mixtures were 0.20–2.19 × 10− 15 m2/s and 0.16–2.69 × 10− 15 m2/s, respectively. This indicates that the
D2 values of category B and C mixtures were reduced by 51% and 55%, respectively, compared to category A mixtures. The lower
permeable porosity of category B and C mixtures may be the reason for the reduced chloride diffusion in these mixtures.
The chloride ion diffusion coefficients for the UHPC mixtures were found to be in the range of 10− 15 to 10− 16 m2/s. In a similar
study, Provete Vincler et al. [53] reported diffusion coefficients in the range of 10− 14 to 10− 15 m2/s, while Wang and Kim [54] reported
a range of 10− 12 m2/s for UHPC mixtures. However, it should be noted that these studies did not apply corrections for joule effect and
temperature effect, even though these effects were reported. For plain UHPC mixtures, Thomas et al. [55] and Pierrard et al. [56]
reported diffusion coefficients in the range of 10− 13 m2/s. For UHPC mixtures containing 2% steel fibers, Tanaka et al. [57] showed
that the diffusion coefficient after a 5-year immersion test was in the range of 10− 15 to 10− 16 m2/s. Therefore, the values obtained from
the current study are consistent with the literature.
These studies serve to validate the diffusion coefficient obtained through the proposed method for UHPC mixtures. Notably, even in
the case of mixture category A with a w/cm ratio of 0.30, the diffusion coefficients were found to fall within the range of 10− 15 to 10− 16
m2/s. As a result, it can be concluded that achieving UHPC class durability is attainable at a w/cm ratio of 0.30.
The relationship between the measured charge passed values and adjusted chloride ion diffusion coefficients was derived using the
linear regression method. Two quadratic equations, Equations (9) and (10), were proposed for plain and fiber-reinforced UHPC,
respectively.

y = − 0.000001x2 + 0.0042x; R2 value is 0.99 (9)

y = 0.0000006x2 + 0.0026x; R2 value is 0.83 (10)

where, x represents the charge passed value in Coulombs obtained from a 6-h RCPT test, and y represents the calculated chloride ion
diffusion coefficient (in units of x 10− 15 m2/s). Fig. 14 shows the relationship between the coefficient of diffusion and charge passed
values. Equations (9) and (10) can be used to determine the chloride ion diffusion coefficients for UHPC materials when the charge
passed values are known from a 6-h RCPT test, thus avoiding the need for rigorous calculations.
The relationship between the Chloride diffusion coefficient and the permeable porosity was also analyzed for both plain and fiber
reinforced UHPC and is depicted in Fig. 15. As can be seen from Fig. 15, there is a linear relationship between diffusion coefficient and
permeable porosity. An increase in permeable porosity can lead to an increase in the coefficient of diffusion of concrete [58]. This is
because as the volume of interconnected voids in concrete increases, there is a greater potential for fluids to flow through the material.
This increased flow can result in a higher coefficient of diffusion. However, the relationship between permeable porosity and coef­
ficient of diffusion can be influenced by many factors, such as the composition of the concrete, the size and shape of the voids, and the
properties of the fluid in question [59]. Moreover, the method used to determine the permeable porosity of the mixtures could not
differentiate between the pore systems that had openings on faces exposed to the NaOH and NaCl solutions during the RCPT test and
the pores that had openings on faces that were not exposed to the chemical solutions. As a result, a few cases (AP-HC, CP-MC, and

10
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

Fig. 12. Solution temperature at the end of the RCPT test.

Fig. 13. Chloride diffusion coefficients of plain and fiber reinforced UHPC.

Fig. 14. Relationship between chloride diffusion coefficient and charge passed value.

BF-MC) were identified where higher permeable porosity values did not exhibit higher charge passed values. Therefore, it is not always
a straightforward relationship and may require further analysis or testing to fully understand.

4. Conclusions
In this study, the electrical and chloride ion migration properties of six UHPC mixtures were investigated. The RCPT test was
conducted, and equations available in the literature were used to calculate these properties from the obtained data.
1. Nernst -Plank equation can be used to predict the chloride ion diffusion coefficient when the correction for Joule effect and
Arrhenius correction for temperature effect were made.
2. Based on the RCPT test data, two empirical equations for predicting the chloride ion diffusion coefficients of both plain and fiber-
reinforced UHPC mixtures were proposed.

11
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

Fig. 15. Relationship between chloride diffusion coefficient and permeable porosity.

3. A positive correlation exists between the diffusion coefficient, cumulative chloride mass, and permeable porosity.
4. The chloride ion diffusion coefficients of the UHPC mixtures ranged from 0.53–3.01 × 10− 15 m2/s, indicating that UHPC mixtures
with a w/cm ratio as high as 0.30 can exhibit excellent corrosion resistance. This finding has the potential to stimulate new research
in UHPC class concrete with higher w/cm ratios that can reduce the use of expensive admixtures. Additionally, such mixtures may
be less susceptible to shrinkage issues which is a primary concern with UHPC in construction.

CRediT authorship contribution statement

Tawsif Mohammad Hasan: Data curation, Formal analysis, Investigation, Writing - original draft. Srinivas Allena: Conceptu­
alization, Data curation, Formal analysis, Methodology, Project administration, Resources, Supervision, Investigation, Writing - review
& editing. Levi Gilbert: Data curation, Investigation.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing in­
terests: Srinivas Allena reports financial support was provided by Washington River Protection Solutions LLC.

Data availability

Data will be made available on request.

Acknowledgements
We gratefully acknowledge the financial support provided by Washington River Protection Solutions for this research project.

References
[1] J. Muro-Villanueva, C.M. Newtson, B.D. Weldon, D.V. Jauregui, S. Allena, Freezing and thawing durability of ultra high strength concrete, J. Civil Eng. Archit. 7
(8) (2013) 907–915.
[2] C. Magureanu, I. Sosa, C. Negrutu, H. Bogdon, Mechanical properties and durability of ultra-high-performance concrete, ACI Mater. J. 109 (2) (2012) 177–183.
[3] J. Li, Z. Wu, C. Shi, Q. Yuan, Z. Zhang, Durability of ultra-high performance concrete – a review, Construct. Build. Mater. 255 (2020).
[4] E. Ghafari, M. Arezoumandi, H. Costa, E. Júlio, Influence of nano-silica addition on durability of UHPC, Construct. Build. Mater. 94 (2015) 181–188.
[5] P. Nguyen, E. Bastidas-Arteaga, O. Amiri, C. El Soueidy, An efficient chloride ingress model for long-term lifetime assessment of reinforced concrete structures
under realistic climate and exposure conditions, Int. J. Conc. Struct. Mater. 11 (2) (2017) 199–213.
[6] J. Upkata, O. Ogirigbo, L. Black, Resistance of concretes to external chlorides in the presence and absence of sulphates: a review, Appl. Sci. 13 (1) (2022) 182.
[7] C. Arya, N. Buenfeld, J. Newman, Factors influencing chloride-binding in concrete, Cement Concr. Res. 20 (2) (1990) 290–300.
[8] M. Florea, H. Brouwers, Chloride binding related to hydration products: Part I: ordinary Portland Cement, Cement Concr. Res. 42 (2) (2012) 282–290.
[9] C. Andrade, J. Alvarez, I. González, Relationship between resistivity and the chloride diffusion coefficient in concrete, Cement Concr. Res. 28 (12) (1998)
1825–1836.
[10] M.G. Richardson, Fundamental of Durable Concrete, Spon Press, Taylor & Francis Group, New York, 2002.
[11] T. Dyer, Concrete Durability, CRC Press: Taylor & Francis Group, Boca Raton, FL, 2014.
[12] F. Zhang, F. Wei, X. Wu, Z. Hu, X. Li, L. Gao, Study on concrete deterioration and chloride ion diffusion mechanism by different aqueous NaCl-MgSO4
concentration, Buildings 12 (11) (2022) 1843.
[13] F.P. Glasser, A. Kindness, S.A. Stronach, Stability and solubility relationships in AFm phases: Part I. Chloride, sulfate and hydroxide, Cement Concr. Res. 29 (6)
(1999) 861–866.
[14] H. Li, N. Farzadnia, C. Shi, The role of seawater in interactioon of slag and silica fume with cement in low water-to-binder ration pastes at the early age of
hydration, Constr. Build. Mater. 185 (2018) 508–518.
[15] W. Long, X. Zhang, G. Feng, J. Xie, F. Xing, B. Dong, K. Khayat, Investigation on chloride binding capacity and stability of Friedel’s salt in graphene oxide
reinforced cement paste, Cement Concr. Compos. 132 (2022), 104603.
[16] N. Neithalath, J. Jain, Relating rapid chloride transport parameters of concretes to microstructural features extracted from electrical impedance, Cement Concr.
Res. 40 (7) (2010) 1041–1051.
[17] G. Glass, N. Buenfeld, Chloride-induced corrosion of steel in concrete, Prog. Struct. Eng. Mater. 2 (4) (2000) 448–458.

12
T.M. Hasan et al. Journal of Building Engineering 82 (2024) 108317

[18] M. Hisada, S. Nagataki, N. Otsuki, Evaluation of mineral admixtures on the viewpoint of chloride ion migration through mortar, Cement Concr. Compos. 21
(5–6) (1999) 443–448.
[19] K. Gruber, T. Ramlochan, A. Boddy, R. Hooton, M. Thomas, Increasing concrete durability with high-reactivity metakaolin, Cement Concr. Compos. 23 (6)
(2001) 479–484.
[20] M. Ahmed, O. Kayali, W. Anderson, Chloride penetration in binary and ternary blended cement concretes as measured by two different rapid methods, Cement
Concr. Compos. 30 (7) (2008) 576–582.
[21] T. Luping, L. Nilsson, Chloride binding capacity and binding isotherms of OPC pastes and mortars, Cement Concr. Res. 23 (2) (1993) 247–253.
[22] H. Ziabara, R. Hooton, M. Thomas, K. Stanish, Influence of the C/S and C/A ratios of hydration products on the chloride ion binding capacity of lime-SF and
lime-MK mixtures, Cement Concr. Res. 38 (3) (2008) 422–426.
[23] Q. Yuan, C. Shi, G. De Schutter, K. Audenaert, D. Deng, Chloride binding of cement-based materials subjected to external chloride environment–a review,
Construct. Build. Mater. 23 (1) (2009) 1–13.
[24] ASTM C1202, Standard Test Method for Electrical Indication of Concrete’s Ability to Resist Chloride Ion Penetration, ASTM International, 2022.
[25] ASTM C1556, Standard Test Method for Determining the Apparent Chloride Diffusion Coefficient of Cementitious Mixtures by Bulk Diffusion, ASTM
International, 2022.
[26] B. Christensen, T. Coverdale, R. Olson, S. Ford, E. Garboczi, H. Jennings, T. Mason, Impedance spectroscopy of hydrating cement-based materials: measurement,
interpretation, and application, J. Am. Ceram. Soc. 77 (11) (1994) 2789–2804.
[27] W. McCarter, T. Chrisp, G. Starrs, J. Blewett, Characterization and monitoring of cement-based systems using intrinsic electrical property measurements,
Cement Concr. Res. 33 (2) (2003) 197–206.
[28] M. Shi, Z. Chen, J. Sun, Determination of chloride diffusivity in concrete by AC impedance spectroscopy, Cement Concr. Res. 29 (7) (1999) 1111–1115.
[29] B. Diaz, X. Nóvoa, M. Pérez, Study of the chloride diffusion in mortar: a new method of determining diffusion coefficients based on impedance measurements,
Cement Concr. Compos. 28 (3) (2006) 237–245.
[30] J. Loche, A. Ammar, P. Dumargue, Influence of the migration of chloride ions on the electrochemical impedance spectroscopy of mortar paste, Cement Concr.
Res. 35 (9) (2005) 1797–1803.
[31] I. Sánchez, X. Nóvoa, G. De Vera, M. Climent, Microstructural modifications in Portland cement concrete due to forced ionic migration tests. Study by
impedance spectroscopy, Cement Concr. Res. 38 (7) (2008) 1015–1025.
[32] C. Andrade, Calculation of chloride diffusion coefficients in concrete from ionic migration measurements, Cement Concr. Res. 23 (3) (1993) 724–742.
[33] T. Zhang, O.E. Gjørv, Effect of ionic interaction in migration testing of chloride diffusivity in concrete, Cement Concr. Res. 25 (7) (1995) 1535–1542.
[34] N.S. Berke, M.C. Hicks, Estimating the life cycle of reinforced concrete decks and marine piles using laboratory diffusion and corrosion data, Corr. Form. Control
Infrastruct. (1992) 207–231.
[35] G.A. Julio-Betancourt, R.D. Hooton, Study of the Joule effect on rapid chloride permeability values and evaluation of related electrical properties of concretes,
Cement Concr. Res. 34 (6) (2004) 1007–1015.
[36] N.S. Berke, M.C. Hicks, Predicting Chloride Profiles in, vol. 93, Corrosion, Houston, TX, 1993.
[37] P. Ghosh, A. Hammond, P.J. Tikalsky, Prediction of equivalent steady-state chloride diffusion coefficients, ACI Mater. J. 108 (1) (2011) 88–94.
[38] T.M. Hasan, L. Gilbert, S. Allena, J. Owusu-Danquah, A. Torres, Development of non-proprietary ultra-high performance concrete mixtures, Buildings 12 (11)
(2022) 1856.
[39] ASTM C1437, Standard Test Methd for Flow of Hydraulic Cement Mortar, ASTM International, 2020.
[40] ASTM C109, Standard Test Method for Compressive Strength of Hydraulic Cement Mortar (Using 50 Mmcube Specimens), ASTM International, 2020.
[41] R. Yu, P. Spiesz, H.J.H. Brouwers, Mix design and properties assessment of ultra-high performance fibre reinforced concrete (UHPFRC), Cement Concr. Res. 56
(2014) 29–39.
[42] ASTM C1856, Standard Practice for Fabricating and Testing Specimens of Ultra-high Performance Concrete, ASTM International, 2017.
[43] D.G. Jeong, H.S. Lee, The effect of pozzolanic reaction under different curing temperatures in strength development of RPC, in: Proceedings of the International
Conference on Sustainable Building Asia, 2010. Seoul, Korea.
[44] Q. Li, Y. Li, Y. Zhou, Y. Wang, Y. Lu, Effects of permeability and fiber content on the transport properties of UHPC, Construct. Build. Mater. 106 (2016) 293–301.
[45] S. Mousavinezhad, G. Gonzales, W. Toledo, J. garcia, C. Newtson, S. Allena, A comprehensive study on non-proprietary ultra-high-performance concrete
containing supplementary cementitious materials, Materials 16 (7) (2023) 2622.
[46] B. Graybeal, Ultra-high Performance Concrete; Publication No. FHWA-HRT-11-038, Federal Highway Administration, 2011.
[47] J. Broomfield, Corrosion of Steel in Concrete, 2 ed., Taylor and Francis, London, 2007, p. 270.
[48] J. Liu, D. Wang, C. Ma, J. Li, Electrical conductivity and rheological properties of steel fiber-reinforced ultra-high-performance concrete, Materials 11 (5) (2018)
812.
[49] Z. Li, J. Li, D. Wang, Y. Zhang, Electrical conductivity of steel fiber-reinforced concrete under compressive loading, Construct. Build. Mater. 111 (2016) 95–102.
[50] Y. Zhang, J. Yang, B. Huang, Effect of curing condition on chloride permeability of concrete: a review, Construct. Build. Mater. 174 (2018) 651–661.
[51] J. Xiao, J. Li, J. Li, Influence of curing condition on chloride diffusion in concrete: experimental investigation and modeling, Construct. Build. Mater. 208 (2019)
135–145.
[52] H. Zhang, Y. Ma, Y. Zhang, Effects of curing regimes on the chloride penetration resistance of concrete: a review, Construct. Build. Mater. 119 (2016) 157–167.
[53] J.P. Vincler, T. Sanchez, V. Turgeon, D. Conciatori, L. Sorelli, A modified accelerated chloride migration tests for UHPC and UHPFRC with PVA and steel fibers,
Cement Concr. Res. 117 (2019) 38–44.
[54] J. Wang, Y.J. Kim, Functional characteristics of ultra-high-performance concrete comprising various fibers, ACI Mater. J. 117 (5) (2020) 179–191.
[55] M. Thomas, B. Green, E. O’Neal, V. Perry, S. Hayman, A. Hossack, Marine performance of UHPC at treat island, in: Proceedings of the 3rd International
Symposium on UHPC and Nanotechnology for High Performance Construction Materials, Kassel, Germany, 2012.
[56] J. Piérard, B. Dooms, N. Cauberg, Evaluation of durability parameters of UHPC using accelerated lab tests, in: Proceedings of the 3rd International Symposium
on UHPC and Nanotechnology for High Performance Construction Materials, Kassel, Germany, 2012.
[57] Y. Tanaka, H. Musha, S. Tanaka, M. Ishida, Durability performance of UFC sakata-mira footbridge under sea environment, Proc. Framcos (2010) 1648–1654.
[58] J. Khatib, F. Bayomy, Permeability and diffucivity of concrete as related to the degree of hydration, Cement Concr. Res. 26 (2) (1996) 287–294.
[59] C. Shi GaoX, G. De Schutter, The relationship between poer structure and permeability and diffucivity of concrete: a Review, Cement Concr. Res. 76 (2015)
1–14.

13