International RILEM Conference on Material Science MATSCI, Aachen 2010 Vol.

II, HetMat 43
WATER CONTENT OF CONCRETE IN NATURAL ATMOSPHERES
AND ITS IMPACT IN THE CORROSION PARAMETERS
C. Andrade, A. Castillo. Institute of Construction Science Eduardo Torroja, CSIC; Serrano
Galvache N4; 28033-Madrid-Spain
ABSTRACT: Concrete moisture is the most important parameter controlling the risk of
degradation in spite of which, there are very scarce works in which the water content of
concrete has been monitored. There have been very numerous studies on concrete moisture
and its evolution when changing the external conditions in controlled chambers in isothermal
conditions. However, the exposure to natural atmospheres subjected to climate changes
induce permanent non steady-state situations due to the evolution of temperature and events
such as snow and rain which make the water content very difficult to be predicted as it is
necessary first to have the prediction of the climate itself. In present paper the monitoring of
external and internal relative humidity, RH, temperature, T, and rain amount have been
measured together with the weight of several concretes submitted to different climates, in
order to model the concrete water content and the internal RH which is very much dependent
on the temperature regime. A set of equations are proposed as an algorithm for predicting the
water content in any climate. In parallel to the water content, the corrosion parameters were
monitored in embedded bars. The corrosion potential, resistivity and corrosion rate show their
strong dependence on the temperature and water content. Expressions were fitted to relate
degree of saturation and the corrosion parameters. The energy of activation of them results
not constant as was previously detected.
1 INTRODUCTION
Water is a key component of the cement paste and the study of his state has been an essential
part of hydration research [Pow48, Fel69, Wit82] and its related paste properties: shrinkage,
creep, etc. The presence of evaporable water is also the key aspect in concrete durability as
most of the degradation phenomena, including reinforcement corrosion, do not progress
without a certain degree of saturation of the concrete [Fag75]. Water state in cement paste has
been the subject of very numerous studies mainly in chambers with controlled humidity and
temperatures [Fag94, Hed93, Nil80, Bar98]. However, there are very few the studies
following its evolution in natural exposure conditions [Sel97, Par88, Ven97, And99, And02,
Cas05]. These studies have aimed at the advanced modelling of vapour diffusivity or water
absorption [Fag94, Hed93, Nil80, Bar98, Sel97], but under fixed hydrothermal conditions and
steady state. Nevertheless, natural weather is characterized in general by a constant evolution
during the day-night and seasonal cycles, and therefore the boundary conditions are in a
permanent non-steady state regime far from laboratory conditions.
In the present paper, certain aspects observed when measuring the evaporable water content
and the relative humidity inside the concrete in natural exposition are commented. Some of
them have been previously published [And99, And02, Cas05].The results of moisture are
related to the corrosion parameters simultaneously measured.
44 ANDRADE, CASTILLO: Water Content of Concrete in Natural Atmospheres and its Impact in the Corrosion
2 EXPERIMENTAL
Several types of structures are studied for following their evolution of moisture content and
corrosion rate under natural atmospheric exposition to rain or other climatic events. In present
paper there will be explained those specimens that have been longer monitored and in which
simultaneous moisture and corrosion measurements are made.
The specimens fabricated were of cylindrical shape of, 7.5cm in diameter x 15cm in height.
They had a centred bar as shown figure 3.1. They were cast in cylindrical moulds and the
moulds placed in a chamber of RH>95% and 20 2C during 24 h. Then, they were cured
under water for other 2 days. Then, they were exposed to air in laboratory conditions
(RH50% and 20 2C) up to 28 days. After that curing series P-2, P-4 and P-13 were
carbonated in a chamber at 100 % of CO
2
and 60 % RH until increase of weight was less than
0.1%. Full carbonation was required to allow for the active corrosion processes to occur.
Series P-23 having 3% CaCl
2
added in the mix was not carbonated, as active corrosion was
developed from the beginning due the presence of chloride ions.
After carbonation, in all series holes were drilled in order to obtain a cavity in which the
internal Relative Humidity (RH-IN) could be measured. Cylindrical cavity is about 20 mm in
diameter and around 3.5 cm in depth. A special metallic tube was introduced in the cavity to
seal its walls, allowing evaporation of water only from the bottom of the cavity. The cavity
borders at the surface of the specimens were sealed with silicone. The cavity remained sealed
by a plug (figure 3.1) between successive measurements. A suitable device (Vaisala humidity
probe) having almost the same diameter than the cavity is introduced periodically in the
cavity to reach its bottom, thus allowing for the humidity exchange of the concrete close to
the steel bar.
The concretes used were prepared with an OPC (CEM I-42,5R), river siliceous sand (0-6 mm)
and crushed granite 6-12 mm as coarse aggregate. Table 2.1 shows the compositions of four
types that will be used in present description.
Table 2.1. Designation of specimen type and composition of concretes used in them.
Concrete designation Cement content
(kg/m
3
)
w/c CaCl
2
Carbonated
P-13 250 0.6 NO YES
P-2-nos sheltered 350 0.50 NO YES
P-4-sheltered 350 0.50 NO YES
P-23 300 0.50 3 % NO
A 380 0.40 NO NO
B 280 0.59 NO NO
All the specimens were introduced during their first year to different constant Temperature, T,
and Relative Humidity, RH. chambers. After this year some of them were submitted to natural
weathering in the Madrid climate. During this time, the internal (IN) and external (EXT) RH,
T and weight, W, were monitored over time, together with the simultaneous measurement of
corrosion potential, electrical resistivity and corrosion rate. For measuring the RH and T, at
the environment of the particular exposition ambient, a small hygrometric datalogger, was
used to record the atmospheric evolution of RH and T (HOBO H08-003-02).
International RILEM Conference on Material Science MATSCI, Aachen 2010 Vol. II, HetMat 45
The reference moisture conditions selected were: dry (40 C until constant weight) and water
saturated (full saturation under vacuum). For dry condition it has been chosen 40C of
temperature in order to guarantee just the evaporation of the free water, avoiding the
evaporation of interstitial water and, then, damage in the microstructure. Thus, the evaporable
water in the sample related to its total absorption was named as S
w
(saturation degree) while
the ratio between the evaporable water content over the exposed surface of the sample was
named W
s
(Evaporable water content/by exposed surface).
The corrosion parameters were measured by means of a laboratory made potentiostat SVC-2.
It uses a potentiostatic step with a waiting period of 30 seconds for making the Polarization
Resistance measurements and applies the current interruption method for measuring the
ohmic drop. The reference electrode is a saturated calomel electrode.
In addition to these specimens other were fabricated to be placed in different climates ranging
from very dry and hot conditions, to very cold conditions, exposed to snow. Twelve site
stations in Spain were identified. In each station three climate exposures where selected:
North-sheltered from rain, North-exposed to rain and South-West exposed to sun and rain.
Cylinders of two sizes were exposed in the stations: larger ones of 15 x 30 cm in size and
smaller of 5 x 8 cm. Concrete mixes designed as A and B of Table 2.1 were used.
The specimens where weighed at least 4 times per year (summer, winter, spring and autumn)
in order to monitor the evaporable water content of the concrete. Monitoring of the
atmospheric climatic parameters was made by using the data of meteorological stations
located near the placement of the specimen stations or by placing the dataloggers previously
mentioned (HOBOS) very near the specimens.
3 RESULTS
3.1 Evolution with time
Figure 3.1 depicts the RH/T-EXT values recorded during the five years period of exposure to
Madrid climate. This town has a continental weather reaching temperatures around 0-5C in
winter and 35-40C in summer. The RH evolves from around 10% in summer to 60-70% in
winter as average values. Raining may appear around 75-100 days per year [And02] with
around 3155 hours/year of Time of Wetness, TOW [Cas05] and around 600-700 ml/m
2

collected annually. Snow may occur one to two times per year.
46 ANDRADE, CASTILLO: Water Content of Concrete in Natural Atmospheres and its Impact in the Corrosion

Exterior Temperature (T-C)
-20
-10
0
10
20
30
40
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Exterior Relative Humidity (RH-%)
0
10
20
30
40
50
60
70
80
90
100
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Precipitation (mm)
0
10
20
30
40
50
60
70
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02

Fig. 3.1. Evolution of RH, T (RH-EXT, T-EXT) and rain events. Climate of Madrid.
Although several specimens have been monitored, for the sake of simplicity, results of only
two of them, in particular, the fully carbonated samples will be shown. The results of T-INT,
RH-INT (in the cavity), and their corresponding corrosion potential, resistivity and corrosion
rate for the non-sheltered condition are shown in figure 3.2 while figure 3.3 shows the data
obtained for the sheltered condition. Several trends can be identified:
Internal Temperature (T-C)
-20
-10
0
10
20
30
40
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Internal Relative Humidity (RH-%)
30
50
70
90
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Resistivity (k.cm)
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
1.E+05
1-2-97 9-9-97 17-4-98 23-11-98 1-7-99 6-2-00 13-9-00 21-4-01 27-11-01
Corrosion Intensity (Icorr-A/cm)
0.001
0.01
0.1
1
10
100
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Saturation Degree (Sw-%)
0
20
40
60
80
100
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Corrosion Potential (Ecorr-mV)
-600
-400
-200
0
200
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02

Fig. 3.2. Left: Evolution of T-IN, RH-IN, Saturation degree, Corrosion Potential. Corrosion Rate
and resistivity in the specimen P-2 exposed to rain. Right: Evolution of T-IN, RH-IN and
Saturation degree, Corrosion Potential. Corrosion Rate and Resistivity in the specimen
P-4 sheltered from rain.
Internal Temperature (T-C)
-20
-10
0
10
20
30
40
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Internal Relative Humidity (RH-%)
30
50
70
90
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Corrosion Potential (Ecorr-mV)
-600
-400
-200
0
200
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Resistivity (k.cm)
1.0E+01
1.0E+02
1.0E+03
1.0E+04
1.0E+05
1.0E+06
1-2-97 9-9-97 17-4-98 23-11-98 1-7-99 6-2-00 13-9-00 21-4-01 27-11-01
Corrosion Intensity (Icorr-A/cm)
0.001
0.01
0.1
1
10
100
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
Saturation Degree (Sw-%)
0
20
40
60
80
100
1-2-97 20-8-97 8-3-98 24-9-98 12-4-99 29-10-99 16-5-00 2-12-00 20-6-01 6-1-02
International RILEM Conference on Material Science MATSCI, Aachen 2010 Vol. II, HetMat 47
Hydrothermal behaviour: The RH evolves following the opposite direction than the T
changes. The RH is much affected by raining events. The S
W
depends mainly on raining
amount in the non-sheltered specimen and on RH-EXT in the sheltered ones. S
w
values
are different whether the specimens are sheltered from rain or not.
Corrosion behaviour: The specimen sheltered remains with negligible I
corr
values due its
low S
W
. The specimens non sheltered exhibits the higher I
corr
values in winter and spring
presenting very low I
corr
values in summer due the drying induced by the higher
temperatures. Resistivity changes in a similar manner although in opposite direction than
I
corr
. The E
corr
evolves as well following the hydrothermal changes, reaching very different
values when wet or dry.
With respect to the specimens placed in different exposure stations, figure 4.1 in the left
shows the evolution during the testing time of the weight of the different specimens types
exposed, and in the right it depicts the maximum water content by square meter of exposed
surface. It can be deduced that the exposure to rain gives higher water contents by square
meter of exposed surface than the sheltered condition.
Ws (Kg/m2)
0
1
2
3
4
01/10/2000 19/04/2001 05/11/2001 24/05/2002
Cil 15x30
Cubic
Yogur
Cil 7.5x15

0
0.5
1
1.5
2
2.5
3
3.5
4
B
I
I

S
o
l
B
I
I

S
o
m
B
I
I

P
r
o
t
B
I

S
o
l
B
I

S
o
m
B
I

P
r
o
t
B
I
V

S
o
l
B
I
V

S
o
m
B
I
V

P
r
o
t
A
I

S
o
l
A
I

S
o
m
A
I

P
r
o
t
A
I
I

S
o
l
A
I
I

S
o
m
A
I
I

P
r
o
t
A
I
V

S
o
l
A
I
V

S
o
m
A
I
V

P
r
o
t

Fig. 3.3. Up: the evolution with time of the water content (W
s
) in K/m
2
of exposed surface of the
different specimens tested and Down: the maximum W
s
for the different exposures.
Those non sheltered from rain are showing higher water contents than the sheltered ones
(sol =exposed to sun, Som =sheltered from sun but not from rain, Prot =sheltered from
rain and sun).
4 DISCUSSION
In order to analyze the results it is convenient to visualize the concrete cover as having two
zones: the external skin and that close to the reinforcement. Also it is important to identify
that the degree of saturation due to drying-wetting cycles will differ following pore sizes. The
capillary pores larger than few nanometers will remain empty until RH exceeds 70-80% RH
or when rain water is falling onto the concrete.
48 ANDRADE, CASTILLO: Water Content of Concrete in Natural Atmospheres and its Impact in the Corrosion
Assuming an initial even distribution of the moisture in the concrete cover, in the sheltered
from rain specimens the processes related to water content can be summarized as follows: a)
when the EXT-RH >INT-RH a vapour transport is produced from the atmosphere to the
interior followed by a condensation in the pores of smaller diameters following Kelvins law
and b) EXT-RH<INT-RH evaporation in the pores larger than the Kelvins law equilibrium
and transport from the interior of concrete to the outside atmosphere. If the condensation due
to the penetration of water vapour induces higher evaporable contents in the concrete skin
zone, an additional capillary movement is expected towards the interior zone not reached by
the penetration of the external vapour. When the concrete is exposed to rain, the processes
present an additional capillary absorption from the skin zone towards the interior. That is,
water entries in liquid state and the transport are produced as two fluids: liquid water and
vapour.
Another observation is that an increase in temperature produces instantaneously an increase in
the RH-INT due to the evaporation. The result is represented in figure 4.1-left and the
schematic trend in figure 4.1-right. In the right part is also shown that the corrosion rate also
increases due to temperature rise and the lowering of the resistivity, but it the temperature is
maintained, the corrosion rate further decreases due to the evaporation.

Fig. 4.1. Effect of temperature in the RH-INT and in the corrosion rate.
For the study of the total amount of evaporable water annually stored in each concrete, in
figure 4.2 is depicted the degree of saturation in function of the RH-INT of the four
specimens P. The zero for determining S
w
corresponds to the weight at 40C and 30% RH-
IN which is typical of long periods of summer in Madrid. It is important to emphasize that
due to present results are not isothermal, the obtained result do not present the same shape of
a typical water isotherm curve reached in experiments under laboratory conditions.
The different slopes of the S
W
-RH-IN plots are due to the different porosity of the concretes
and to the fact that P4 specimen is sheltered from rain. Thus, while the concrete non-sheltered
from rain (P-2) presents S
w
values in winter near 50%, the concrete sheltered (P-4) never
shows values higher than around S
w
=20%, however the RH-IN presents very similar values in
many occasions.
International RILEM Conference on Material Science MATSCI, Aachen 2010 Vol. II, HetMat 49
It can be summarized that the main features of weather influence are: a) the rain (and snow) as
water suppliers and b) the effect of temperature: daily cycles and year cycle (seasonal
evolution). They have to be studied and related to the S
W
and in turn to the corrosion rate, I
corr
.
Additionally, regarding aggressivity classes in standards (Eurocode 2, EN 206, EHE-08), it
seems better to consider the conditions as sheltered and not-sheltered from rain than the
present classes of wetting and dry periods.
P-2 Jan 1999 - Dec 2002.
0
10
20
30
40
50
60
70
20 30 40 50 60 70 80 90
RH (%)
S
w

(
%
)
P-4 Jan 1999 - Dec 2002.
0
10
20
30
40
50
60
70
20 30 40 50 60 70 80 90
RH (%)
S
w

(
%
)
P-23 Jan 1999 - Dec 2002.
0
10
20
30
40
50
60
70
20 30 40 50 60 70 80 90
RH (%)
S
w

(
%
)
P-13 Jan 1999 - Dec 2002.
0
10
20
30
40
50
60
70
20 30 40 50 60 70 80 90
RH (%)
S
w

(
%
)

Fig. 4.2. Degree of saturation of the specimens P-2 (carbonated) and P-23 (3%CaCl
2
), P-4
carbonated specimen sheltered from rain and P-13 (carbonated more porous) exposed to
rain versus the RH-IN.
In order to represent more comprehensively the behaviour it has been developed a
modification of the psychometric graph that is represented in figure 4.3. The right part of the
graph contains the possibility to calculate the evaporable water content (as W
s
in Kg/m
2
of
exposed concrete surface).
0
5
10
15
20
25
0 10 20 30 40 50
External Temperature (C)
A
b
s
o
l
u
t
e

I
n
t
e
r
n
a
l

H
u
m
i
d
i
t
y

(
g
/
k
g
)
350 kg Carbonated
250 Kg Carbonated
3% CaCl
1% CaCl
0 0.2 0.4 0.6 0.8 1
Ws (kg/m2)
350 kg Carbonated-5 mm
350 kg Carbonated-20 mm
250 kg Carbonated-5 mm
250 kg Carbonated-20 mm
3% CaCl-5 mm
3% CaC-20 mm
1% CaCl-5 mm
1% CaCl-20 mm
Ti
Habsi
Ws i
Precipitation

Fig. 4.3. Modified psychometric graphs enabling the calculation of the evaporable water content
from the external RH and T.
50 ANDRADE, CASTILLO: Water Content of Concrete in Natural Atmospheres and its Impact in the Corrosion
4.1 Corrosion behaviour
The saturation degree, S
w
, (or its equivalent W
s
) and the temperature, T, are the main
influencing parameters in the corrosion rate I
corr
. The corrosion rate increases with both T and
S
w
; nevertheless, if the increase in T leads to water loss by evaporation, the resistivity can also
increase, lowering the corrosion rate. This behaviour explains why in Madrid, in spite of
having high T, as the RH-EXT is low in summer, the yearly average I
corr
is small. That is,
countries with lower temperatures induce longer periods of water condensation within
concrete pores and, in consequence, of higher annual averaged I
corr
values. In these cold
climates, in addition, the melted snow may become an additional source of liquid water.
4.2 Prediction of water content from the climatic parameters
These values together with those of the specimens exposed to several exposure stations at
different climates in Spain were used to develop the relation between Corrosion rate,
evaporable water content and the climatic parameters (RH, T and amount of rain) by means of
an algorithm, because a direct relation was not found. The algorithm was developed by fitting
the experimental data to estimate the evaporable water content from the three parameters: T,
RH and amount of rain (snow is assumed to be rain). From the evaporable water content the
Resistivity and the corrosion rate are estimated. In the algorithm I
corr
is a function of:
I
corr
=f (Resistivty) =f (Ws, Rain amount, T and absolute humidity)
5 CONCLUSIONS
The study presented has been performed in relatively small specimens. Although, they may
present a different behaviour than real-size structures, the trends enable to draw up the
following conclusions:
The main climatic parameters influencing the hydrothermal behaviour of concrete are rain
events and temperature. The most illustrative graphs to study this behaviour are the
psychometric chart and the S
W
-RH-IN plot.
Sheltered and not sheltered from rain conditions should be taken into account in exposure
classifications in codes or standards. They reflect better a difference in real moisture
conditions than the definition of wet dry-cycles.
An algorithm has been developed that enables the calculation of the evaporable water
content and the degree of saturation in any climate, providing the values of T, RH and rain
regime are known. If the validity of this algorithm is confirmed with a larger set of real
concretes, it will be possible to predict, in any climate, the water content and, in
consequence, the reinforcement corrosion.
The yearly averaged I
corr
values seem to be higher in cold than in hot weathers, due to the
lack of driving force for water evaporation in the cold ones and in consequence the longer
times of wetness periods in concrete, although rain events in hot climates periods can be
also providers of water to the concrete.
International RILEM Conference on Material Science MATSCI, Aachen 2010 Vol. II, HetMat 51
ACKNOWLEDGEMENTS
The authors thank to those collaborating in all the weather stations facilities for the support
provided to the work. They are also grateful to the funds provided by the Ministry of Science
and Innovation of Spain though the project INGENIO 2010-CONSOLIDER Project on
Safety and Durability of Structures: SEDUREC.
REFERENCES
[And02] Andrade, C., Sarra, J ., Alonso, C. Corrosion rate evolution in concrete structures
exposed to the atmosphere Cement and Concrete Composites 24 (2002) 55-64.
[And99] Andrade, C., Sarra, J ., Alonso, C. Relative humidity in the interior of concrete exposed
to natural and artificial weathering Cement and Concrete Research 29 (1999) 1249-
1259.
[Bar98] Baroghel-Bouny, V., Proceedings of the International RILEM Conference Concrete:
from material to structure, 11-12 sept., 1996, Arles, France (Edited by J .P. Bournazel &
Y. Malier, RILEM, (1998) 144-165.
[Cas05] Castillo, A. Modelado del efecto de las acciones climticas en la cantidad de agua en
estado lquido del hormign para la estimacin de la corrosin de la armadura. Tesis
doctoral. UPM-Escuela Tcnica Superior de Ingenieros de Caminos. Madrid, 2005.
[Fag94] Fagerlund, G., Div. of Building Materials, Lund Institute of Technology Report
TVBM-3059 (1994).
[Fag75] Fagerlund, G., Significance of critical degrees of saturation at freezing of porous and
brittle materials. American Concrete Institute -Pub SP-47, Detroit (1975).
[Fel69] Feldman, R.F. y Sereda, P.J . A model for hydrated portland cement pastes deduced
from sorption length change and mechanical properties- Materials and
StructuresV1(1969)509.
[Hed93] Hedenblad, G., Division of Building Materials Lund Institute of Technology, Doct.
Dissertation, March (1993).
[Nil80] Nilsson, L.O., TVBM-1003, Division of Building Material, Lund Institute of Technology,
Lund (1980).
[Par88] Parrott, L.J ., Advances in Cement Research, vol.1, no.3, J uly (1988).
[Pow48] Powers, T. C., Brownyard, T.L. - Studies of the physical properties of hardened cement
paste. - Portland Cement Ass. Res. Bull. 22, Chicago (1948).
[Sel97] Sellevold, J .E. Resistivity and humidity measured of repaired and non repaired areas in
Gimsostrauman bridge International symposium on Theory to Practice in Marine
Environment, Norway-Swevaer-May 1997 A.Blacknold Editor Norland: Norwegian
Public Road Administration, P. 283-95.
[Ven97] Vennesland, O. Electrochemical parameters of repaired and non repaired concrete at
Gimsoystrauman Bridge. Ibidem ref (5) Solvaer-Norway, May 1997, p.253-262.
[Wit82] Wittmann, F.H. Properties of hardened cement paste Proc. 7
th
Intern Congress
Chemistry of Cement, Paris 1982 vol 1, pg VI-2/10.