Relationship Among the Permeation Rate of Water into Concrete, the Mix Design, Curing, and the Degree of Drying

Abstract

Water permeation into concrete causes various types of deterioration in concrete structures such as rebar corrosion, frost damage, ASR, etc.; however, the evaluation of the water permeation rate is difficult. This paper proposes a simple equation to evaluate the water permeation rate into concrete, based on basic information such as the mix design, curing condition, and degree of drying. Concrete specimens are cast, cut into slices of different thicknesses, and the time taken for water penetration is measured. The permeation rate is calculated from the relationship between the obtained permeation time and the slice thickness. Additionally, a literature survey is conducted to collect the water permeation rate in various types of concrete. A prediction equation for the water permeation rate is proposed, based on the data. Although this data contains various types of concrete with/without the admixture material, different water-to-binder ratios, and different curing conditions, all the calculated values fall within a range of ±150% of the data from literature.

Full text

Correlation between Permeability of Concrete and Threshold
Pore Size obtained with Epoxy-Coated Sample
Yuya
Sakai
,
Choji
Nakamura
Kishi
Toshiharu
,
,
Journal of Advanced Concrete Technology,
volume ( ), pp.
11
8
189-195
Characterization and Modeling of Pores and Surfaces in Cement Paste: Correlations to Processing and
Properties
Hamlin Jennings,Jeffrey W.Bullard,
Jeffrey J.
Thomas,
Jose E.
Andrade
Journal of Advanced Concrete Technology,
volume ( ), pp.
6
2008
5-29
Effect of the key mixture parameters on tortuosity and permeability of concrete
Shamsad
Ahmed
,
Abul Kalam
Azad
Kevin F.
Loughlin
,
Journal of Advanced Concrete Technology,
volume ( ), pp.
10
2012
86-94
Development of High-Durability Concrete with a Smart Artificial Lightweight Aggregate
Jinhwan
Jeon
,
Haruki
Momose
Tetsushi
Kanda
,
,
Hirozo
Mihashi
Journal of Advanced Concrete Technology,
volume ( ), pp.
10
2012
231-239

Journal of Advanced Concrete Technolog y Vol. 11, 189-195, August 2013 / Copyright © 2013 Japan Concrete Institute189

Scientific paper
Correlation between Permeability of Concrete and Threshold Pore Size
obtained with Epoxy-Coated Sample
Yuya Sakai1, Choji Nakamura2 and Toshiharu Kishi1
Received 4 January 2013, accepted 6 August 2013 doi:10.3151/jact.11.189
Abstruct
The authors have proposed a new method, epoxy-coating on MIP (Mercury Intrusion Porosimetry) sample, to measure
threshold pore radius of concrete to obtain an indicator of pore structure which has correlation with air and water per-
meability. In this paper, first, the validity of the above method was studied through observation on splitting surface of
samples after MIP analysis and comparison with obtained threshold pore radius and permeability. Results showed that
the proposed method is suitable to extract threshold pore radius, and it showed good correlation with water permeability
and, if concrete is enough dried, air permeability. Good correlation was found even on samples prepared with overseas
concrete material and core samples taken from existing structures overseas. The above results indicate that pore struc-
ture governs both air and water permeability of concrete and that threshold pore radius can be an indicator of the per-
meability of concrete.
1. Introduction
For quality verification of newly constructed concrete
structures and efficient maintenance of existing concrete
structures, more attention is paid to evaluation of con-
crete durability. Considering the above purposes, the
evaluation method should be non-destructive. So far,
various methods have been proposed, however, all non-
destructive method to evaluate mass transfer resistance
of concrete gives us just qualitative information and it is
impossible or not reliable to estimate mass transfer or
deterioration rate of concrete members based on the
obtained results. In the first place, it is not understood
what information, pore size distribution, total pore vol-
ume, etc., is reflected to the results in those tests. The
authors (Nakamura et al, 2012) have developed a
method to extract threshold pore radius, one of the indi-
cators of pore structure. In this paper, first, it was con-
firmed that sudden intrusion of mercury occurs in MIP
(Mercury Intrusion Porosimetry) when the sample is
coated with epoxy resin. The corresponding pore radius
must be threshold pore radius according to the definition.
Next, pore structure and both air and water permeability
of concrete specimen was measured and correlation was
studied. Additionally, concrete specimen prepared with
Korean material and core sample taken from existing
structure in Korea was analysed and the applicability of
the proposed method was discussed.
2. New method to estimate threshold pore
size
So far, many researchers have pointed out good correla-
tion between threshold pore size and air and water per-
meability. Powers (1958) and Mehta (1980) studied re-
lationship between pore structures and water permeabil-
ity. Powers found correlation between volume of capil-
lary pore and water permeability, and Mehta reported a
good correlation between threshold pore size and water
permeability. Here, threshold pore size is the minimum
pore size which mass should pass to penetrate the objec-
tive, and pore size distribution is measured with MIP.
Halamickova and Detwiler (1995) reported that there
was a correlation between critical pore size, an indicator
of pore structure, and both water permeability and coef-
ficient of oxygen diffusion. Goto (1996) related thresh-
old pore size with hydration degree of cement. The indi-
cators of pore structures and permeability shown above,
however, doesn't show such high correlation in various
types of specimens since it is not easy to extract thresh-
old pore size correctly, particularly, in samples taken
from concrete specimens. One probable reason is that in
concrete, because of transition zone around gravel, pore
size distribution measured with MIP shows less sudden
intrusion as seen in Fig. 1, and as a result, it is difficult
to extract threshold pore size. Here, threshold pore size
in MIP means the largest pore size to push mercury into
the core of a sample, and when the pressure corresponds
to the threshold pore size is applied, much mercury is
intruded suddenly into the sample. In this paper, the
definition of threshold pore radius follows Winslow and
Diamond (1970), the corresponding pore radius where
cumulative pore volume curve shows the largest tangent.
At this time, in normal MIP method, samples are small
and mercury is intruded three-dimensionally. As a result,
the volume not intruded yet is small and the largest tan-
1Institute of Industrial Science, The University of Tokyo
Tokyo, Japan.
E
-mail: ysakai@iis.u-tokyo.ac.jp
2School of Engineering, The University of Tokyo,
Tokyo, Japan.

Y. Sakai, C. Nakamura and T. Kishi / Journal of Advanced Concrete Technology Vol. 11, 189-195, 2013 190

gent is not clear, as seen in another example, dotted line
in Fig. 2. Therefore, authors (2012) proposed a new
method in which special treatment was given to the cu-
bic samples, which usually used in MIP, 5mm in a side.
D-dried samples were covered with epoxy-resin leaving
small area, about 4mm2 as shown in Fig. 3, and MIP
analysis was conducted on them. In a pre-test with lump
of epoxy resin, it was confirmed that elastic deformation
of epoxy-resin affected the result when analysing pore
radius had been smaller than 10nm. As shown in Fig. 4,
this treatment was expected to keep larger not-intruded
area when the pressure reached the value which corre-
sponds to threshold pore radius. Figure 2 is a compari-
son between the measured results of a normal and an
epoxy-coated sample. In the figure, pore volume is
shown in per mass, not volume, since precise volume of
epoxy-coated sample was not easy to calculate. It is
obvious that epoxy-coated sample shows a clear jump
and the corresponding pore size indicates threshold pore
size. Smaller intrusion area in the epoxy-coated sample
causes smaller cumulative pore volume since some
pores are not easy to access because of epoxy-resin
coating.
3. Verification of MIP with epoxy-coated
samples to extract threshold pore size
3.1 Sudden intrusion of mercury at threshold
pore size
To confirm sudden intrusion in MIP with epoxy-coated
samples, some tests were done on mortar samples,
W/C=55%, C/S=30%. In the study, the splitting surfaces
of the samples were observed after MIP analysis. The
specimens were demolded 24 hours after casting, cured
under water until the age of 28 days, and dried in room
of 20 degrees, relative humidity was not controlled but
around 60% in average, for half year. 5mm cubic pieces
were taken from the specimen and immersed into ace-
tone for 24 hours, dried by D-dry method, and MIP
analysis was executed two times in each case. Figure 5
shows intruded mercury volume of samples with and
without coating. Since each two curves show almost
same behaviour, it can be said that this method has
enough reproducibility. Compared with normal samples,
epoxy-coated samples show sudden increase from
around 100nm. The tangent of the curves is shown in
Fig. 6. With and without epoxy-coating shows its peak
at around 45nm and 120nm, respectively, and 45nm is
Cumulative pore vol-
ume (ml/ml)
Pore radius (nm)
Fig. 1 Example of MIP results (concrete specimen,
W/C=40%).
Cumulative pore
volume (ml/g)

Normal
Epoxy
coated
Pore radius (nm)
Fig. 2 MIP results of normal and epoxy-coated samples.
Fig. 3 Epoxy-coated sample.
Intruded Mercury Intruded Mercury
Fig. 4 Conceptual diagram of mercury intrusion into normal and epoxy-coated sample.
(b) Epoxy-coated sample (a) Normal sample
Non-intruded area

Y. Sakai, C. Nakamura and T. Kishi / Journal of Advanced Concrete Technology Vol. 11, 189-195, 2013 191

the threshold pore radius measured with the proposed
method. To confirm the obtained threshold pore radius
is correct, i.e., sudden mercury intrusion occurred at
around 45nm, the intrusion was stopped at different
pressures, the samples were split, and the splitting sur-
faces were observed. The intrusion was stopped at the
pressures of 8.58MPa, 12.76MPa, and 43.45MPa which
correspond to the pore radius of 164nm, 100nm, and
30nm, respectively. The splitting surfaces are shown in
Fig. 7. In the case of normal samples, the colour of the
surface changes gradually as the pressure increases. On
the other hand, when the samples were coated, sudden
change occurred between 100nm and 30nm, and this
value corresponded to the obtained threshold pore ra-
dius, 45nm. The above results indicate that with epoxy-
resin-coated samples, sudden intrusion of mercury is
induced at threshold pore radius.
3.2 Correlation of threshold pore size with indi-
cators of pore structure
The validity of the obtained threshold pore radius as an
indicator of pore structure is discussed with concrete
specimen by comparing the correlation of water perme-
ability with other indicators of pore structure. The mix-
ing design and curing condition of the concrete is shown
in Ta bl e 1 . AE water reducing agent and AE agent is
used with the amount of 0.2% and 0.004% against ce-
ment weight, respectively. Specimens were demolded
24 hours after the casting and under-water, sealed, or in-
wind curing were given until the age of 28 days, and
after that, all specimens were cured in a room of 20 de-
gree Celsius. Here, in in-wind curing, specimens were
winded by a fan to accelerate drying. In water perme-
ability test, cylindrical specimen with 10cm in diameter
and 20cm in height were cut into 3.8cm thickness from
the middle of the specimens. Before the test, they were
Intruded mercury volume
(ml/g)
Normal
Coated
Fig. 5 Comparison between with and without coating.
Pore radius (nm)
(b) Epoxy-coated sample
(a) Normal sample
164nm 100nm 30nm
164nm 100nm 30nm
Fig. 7 Splitting surface after mercury intrusion (Numbers below figures indicate the pore size which corresponds to ap-
plied pressure in MIP before splitting).
Δ(intruded mercury)
Δ(log pore radius)
Normal
Coated
Fig. 6 Tangent of Fig. 5.
Pore radius (nm)

Y. Sakai, C. Nakamura and T. Kishi / Journal of Advanced Concrete Technology Vol. 11, 189-195, 2013 192

saturated with vacuum for 24 hours. 2.5 MPa pressure
was given to the specimens during the test. In surface
air permeability test, cylindrical specimen with 15cm in
diameter and 30cm in height were tested with Torrent
method. All tests were conducted when the age of the
specimens were 2.75 years. The relationship of water
permeability with total pore volume, threshold pore size
of normal sample and epoxy-coated sample is shown in
Fig. 8. Each measured value is shown in Tabl e 2 . Here,
total volume is the cumulative pore volume at the pore
radius of 1.5nm and was measured with normal samples.
Total pore volume and threshold pore size of normal
samples shows moderate correlation with water perme-
ability, on the other hand, threshold pore radius of ep-
oxy-coated specimen shows good correlation. It is the
same in surface air permeability, as shown in Fig. 9,
good correlation between surface air permeability and
threshold pore radius of epoxy-coated samples.
4. Applicability to existing structures over-
seas and specimen prepared with overseas
concrete material
Field study on existing structures, 11 years old, in Korea
was conducted. Design compressive strength and mix
proportion of the members are shown in Ta ble 3 . Sur-
face permeability test was executed on concrete mem-
bers at site and, after that, core samples were taken from
Table 1 Mix proportion of concrete specimens prepared in experimental room.
Unit : kg/m3
W/B (%) Curing condition W C FA or BFS S G
Threshold radius
(nm)
N40-1 40 Water 180 450 - 708 978 52.5
N55-1 55 Water 180 327 - 805 984 52.5
N70-1 70 Water 180 257 - 886 960 52.7
FB55-1 55 Water 172 251 62 791 1007 15.7
FC55-1 55 Water 169 216 92 783 1017 42.2
BA55-1 55 Water 179 260 65 787 1002 126.7
BB55-1 55 Water 174 159 159 792 1008 99.3
N70-3 70 Wind 180 257 - 886 960 301.2
FB40-1 40 Water 172 345 86 694 998 23.7
FB70-3 70 Wind 172 197 49 873 985 866.9
BB40-1 40 Water 174 218 218 695 1001 52.4
BB70-3 70 Wind 174 124 124 873 985 577.4
L55-2 55 Sealed 180 327 - 807 987 437.5
M55-2 55 Sealed 180 327 - 807 987 67.3
H55-2 55 Sealed 180 327 - 804 984 99.3

Table 2 Test and analysis results.
Total pore
volume (ml/ml)
Threshold pore
radius: normal (nm)
Threshold pore
radius: epoxy (nm)
Water permeability
(×10-9cm/s)
Surface air
permeability
(×10-16m2)
N40-1 0.169 99.4 52.5 2.50 0.83
N55-1 0.197 152.6 52.55 2.47 1.8
N70-1 0.214 51.9 52.7 10.33 1.65
FB55-1 0.181 41.6 15.75 2.77 0.665
FC55-1 0.218 99.2 42.25 6.50 1.4
BA55-1 0.182 488.8 126.75 18.14 3.8
BB55-1 0.185 871.6 99.3 24.96 2.4
N70-3 0.182 372.1 301.25 169.01 53
FB40-1 0.182 41.7 23.75 1.71 0.335
FB70-3 0.236 881.9 866.95 470.62 142
BB40-1 0.160 41.4 52.4 2.05 0.355
BB70-3 0.242 1108.9 577.45 356.18 493
L55-2 0.232 41.5 437.55 27.52 29
M55-2 0.200 41.5 67.35 18.62 11.5
H55-2 0.165 41.5 99.35 5.34 4.35


Y. Sakai, C. Nakamura and T. Kishi / Journal of Advanced Concrete Technology Vol. 11, 189-195, 2013 193

them to measure water permeability and pore structures.
Tab l e 4 shows the mixing design of concrete specimen
prepared in laboratory with Korean material. Specimens
were demolded 24 hours after the casting and under-
water, air curing were given until the age of 28 days,
and after that, all specimens were cured in a room of 20
degree Celsius. Surface permeability test, water perme-
ability test, and MIP were conducted on the specimen at
the age of 6 month. The results are shown in Ta b le 2
and Fig. 10 together with the results in Fig. 8. In the
figure, square plots are the results of specimen in Tab l e
1, triangle plots in Ta b l e 3, and asterisk in Ta ble 4 . It is
seen that in Fig. 10(a), the correlation between total
pore volume and water permeability is poor, the plots
distribute widely. On the other hand, in Fig. 9(b),
threshold pore radius and water permeability shows
good correlation, plotted in a line. In Fig. 11, the rela-
tionship between surface air permeability and total pore
volume and threshold pore radius is shown. Total pore
volume shows poor correlation. Correlation of threshold
pore radius is much better, however, the results meas-
ured at site on existing structures are deviated from the
tendency. The possible reasons will be moisture content
due to rain and the difference of test machine, Permea-
TORR in laboratory and TORRENT at site. Torrent
(2012) reported that the measured value in Permea-
TORR is smaller than TORRENT. In future, additional
data will be collected and the applicability of the pro-
Threshold pore radius (nm)
Water permeability (×10-9cm/s)
(b) Water permeability and threshold pore radius of normal
sample
Total pore volume (ml/ml)
Water permeability (×10-9cm/s)
(a) Water permeability and total pore volume
Threshold pore radius (nm)
Water permeability (×10-9cm/s)
(c) Water permeability and threshold pore radius of epoxy-
coated sample
Fig. 8 Correlation between water permeability and indi-
cators of pore structure.
Threshold pore radius (nm)
(b) Surface air permeability and threshold pore radius of nor-
mal sample
Ai
r
p
ermeabilit
y
(
×10-16
m
2
)
Total pore volume (ml/ml)
Air permeability (×10-16m2)
(a) Surface air permeability and total pore volume
Threshold pore radius (nm)
(c) Surface air permeability and threshold pore radius of ep-
oxy-coated sample
Air permeability (×10-16m2)
Fig. 9 Correlation between surface air permeability and
indicators of pore structure.

Y. Sakai, C. Nakamura and T. Kishi / Journal of Advanced Concrete Technology Vol. 11, 189-195, 2013 194

posed method will be discussed over specimens of vari-
ous conditions.
5. Conclusion
In this paper, first, it was confirmed that epoxy-resin
coating on MIP sample causes sudden intrusion at
threshold pore size. The obtained threshold pore size
with the proposed method showed good correlation with
water permeability and, when the concrete was enough
dry, air permeability. The above results indicate that
pore structure governs both air and water permeability
of concrete in the permeability test and that threshold
pore radius can be an indicator of the permeability.
Acknowledgement
The authors would like to thank for the advice provided
by Dr. Cheong haimoon at Expressway & Trans-
Table 3 Property and mixing proportion of measured existing structures.
Unit : kg/m3
Member Design compressive
strength (MPa) W/C (%) W C S G Water reducing
admixture (%)
Bridge railing 24 48 165 348 851 936 0.22
Bridge abutment 24 50 175 353 757 1025 0.3
Bridge pier 24 50 175 353 757 1025 0.3
Pavement 27-30 42.3 144 340 682 1192 0.15
Table 4 Property and Mixing proportion of concrete specimen.
Name W/B [%] Slump [cm] Air [%] Maximum aggregate size (mm) s/a [%]
W/C45% 45 15 5-7 25 45
W/C53% 53 15 5-7 25 43
FA 43 0-4 5-7 30 36
Unit [kg/m3]
Name W C FA S G AE (1) AE (2) SP
W/C45% 156 346 0 800 997 0 0 3.114
W/C53% 167 315 0 764 1032 2.205 0 0
FA 150 280 70 630 1150 0 0.07 1.05
Table 5 Test and analytical results of core samples or on members.
Total pore volume
(ml/ml)
Threshold pore radius:
epoxy (nm)
Water permeability
(×10-9cm/s)
Surface air permeability
(×10-16m2)
Bridge railing(1) 0.103 23.7 4.77 0.169
Bridge railing(2) 0.119 31.2 5.57 0.034
Bridge abutment 0.189 41.95 6.57 17.59
Bridge pier(1) 0.192 31.35 5.25 23.49
Bridge pier(2) 0.149 67.6 4.63 2.315
Bridge pier(3) 0.197 47.95 No data (cracked) 50.35
Pavement 0.146 10.4 0.49 0.241
Table 6 Test and analysis results.
Total pore volume
(ml/ml)
Threshold pore radius:
epoxy (nm)
Water permeability
(×10-9cm/s)
Surface air permeability
(×10-16m2)
W/C45% (Air) 0.167 10.8 No permeation 0.28
W/C53% (Air) 0.141 15.75 18.67 0.69
FA (Air) 0.221 52.6 7.90 1.4
W/C45% (Water) 0.153 2.45 No permeation 0.038
W/C53% (Water) 0.125 15.75 2.82 0.13
FA (Water) 0.242 126.65 7.32 0.25


Y. Sakai, C. Nakamura and T. Kishi / Journal of Advanced Concrete Technology Vol. 11, 189-195, 2013 195

portation Research Institute in Korea. The experimental
work in this paper was carried out with financial support
of international project (Establishment of durability as-
sessment system by inspecting the surface quality of
concrete) by Korea Expressway.
References
Goto, T., (1996). “Modelling of cement hydration
reaction and strength development.” Thesis (PhD).
The University of Tokyo.
Halamickova, P. and Detwiler, R., (1995). “Water
permeability and chloride ion diffusion in portland
cement mortars: Relationship to sand content and
critical pore diameter.” Cement and Concrete
Research, 25(4), 790-802.
Mehta, P. and Manmohan, D., (1980). “Pore size
distribution and permeability of hardened cement
paste.” 7th International Congress Chemistry of
Cement, Paris, 3, 71-75.
Nakamura, C., Sakai, Y. and Kishi, T., (2012). “Study on
the effects of chemical and physical properties of
concrete on the behavior of internal water.” 11t h
International Symposium on New Technology for
Urban Safety of Mega Cities in Asia [CD-ROM].
Powers, T. C., (1958). “The physical structure and
engineering properties of concrete.” Research and
Development, Bulletin of Portland Cement
Association, 90.
Torrent, R., (2012). “Non-destructive air-permeability
measurement: from gas-flow modelling to improved
testing.” The second international conference of
microstructural-related durability of cementitious
composites [CD-ROM].
Winslow, D. and Diamond, S., (1970). “A mercury
porosimetry study of the evolution of porosity in
Portland cement.” Journal of Materials, 5, 564-585.
Threshold pore radius (nm)
(b) Water permeability and threshold pore radius of epoxy-
coated sample
Total pore volume (ml/ml)
(a) Water permeability and total pore volume
Fig. 10 Correlation between water permeability and indicators of pore structure.
SpecimeninTable4
SpecimeninTable1SpecimeninTable3
Water permeability (×10-9cm/s)
Water permeability (×10-9cm/s)
*
Threshold pore radius (nm)
Specimen in Table 4
Specimen in Table 1 Specimen in Table 3 *
Air permeability (×10-16m2)
(b) Surface air permeability and threshold pore radius of ep-
oxy-coated sample
(a) Surface air permeability and total pore volume
Fig. 11 Correlation between surface air permeability and indicators of pore structure.
Air
p
ermeabilit
y
(
×10-16
m
2
)
Total pore volume (ml/ml)