Sei sulla pagina 1di 21

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover

RILEM TC 189-NEC: State-of-the-Art Report May 2007

CHAPTER 2. TRANSPORT MECHANISMS AND REFERENCE TESTS


J. Kropp (1) and M. Alexander (2)
(1) Institute for Building Materials, Hochschule Bremen, Neustadtswall 30, D-28199
Bremen, Germany
(2) Department of Civil Engineering, University of Cape Town, Private Bag Rondebosch,
7701 Cape Town, South Africa

During the past decades extensive research has been dedicated to degradation phenomena
of concrete and reinforced concrete structures in order to reveal the governing mechanisms,
and explain and quantify the deleterious effects of aggressive substances. For structures under
service conditions, the environment, i.e. the natural atmosphere, soil, water or user-made
conditions, imposes sustaining actions on the surfaces of structural members. The durability
of a concrete or reinforced concrete member is then controlled by the resistance of the
outermost layers or concrete skin of a cross section to the external attacks [2.1].
The external attacks may cause a deterioration of the skin of a member which is
immediately exposed to the actions, and with prolonged exposure the corrosion front
progresses deeper into the section. Assuming that the already corroded material in the surface
layer is instantaneously removed, the progress of the corrosion front is controlled by the
reaction rate of deleterious species with concrete compounds. If, however, the reaction
products remain in place, the attacking species have to penetrate the corroded concrete zone
before they can reach and react with non corroded material either concrete compounds or
steel. In this case the progress of the corrosion front is controlled by the rate of transfer of
aggressive species to the reaction front.
The second case includes the penetration of deleterious substances which do not cause a
degradation of the concrete surface zones directly, e.g. the ingress of chloride ions, or do not
immediately cause degradation of the concrete integrity, but they result in changes of the
physical-chemical properties, e.g. carbonation. These processes, however, may initiate further
corrosive actions in deeper sections of a structural member, such as the anodic dissolution of
embedded steel reinforcement. Also in these cases the external reactants associated with a
corrosion reaction have to penetrate a concrete section of a given thickness first. Thus, the
corrosive action is controlled by the availability of reactants, i.e. their rate of transfer.
The first case of a purely reaction-rate controlled degradation is not usually of importance
for concrete structures under typical service conditions, except in rare cases of acid attack in
flowing media, to soft water attack accompanied by erosion, or mechanical wear. Most
corrosive actions on concrete and reinforced concrete members are encountered when
aggressive species penetrate the concrete surface zone, i.e. the concrete cover of the
reinforcement. Examples of such degradation mechanisms are [2.2]:

13

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

the corrosion of the steel reinforcement by anodic dissolution of the iron, initiated
by the ingress of chloride ions which depassivate the steel surface after a critical
threshold concentration has been reached. The rate of transfer of chloride ions
controls the time to corrosion.
the corrosion of the steel reinforcement by anodic dissolution of the iron after
depassivation of the steel surface as a consequence of the loss of alkalinity due to
carbonation. The penetration of carbon dioxide determines the depth of
carbonation at a given time t.
the formation of secondary ettringite from calcium aluminate hydrates (hydration
products of the cement) with sulphate ions penetrating from external sources. The
availability of sulphate ions is rate-controlling, provided the internal reactants are
available.
expansive alkali-aggregate reactions supported by the ingress of external alkalies
and/or water. Although all necessary reactants for AAR may be introduced into the
concrete by the concrete making materials the penetration of alkali ions and water
into deeper concrete sections increases the severity of the attack.
frost action, eventually enhanced by de-icing agents, which cause disruptive forces
upon freezing of water in a water saturated pore system. Saturation of the pore
system is attained only by up-take of water from the exposed surfaces.

Based on these considerations the transport properties of concrete for water, gases and
dissolved species have attracted increasing interest during recent years in order to analytically
explain, quantify or predict a corrosive action on a concrete structure or describe the
perviousness of a concrete section to various media [2.3-2.5].
In addition to durability considerations, the imperviousness of concrete sections to liquids
and gases is often a service requirement for advanced structures in the nuclear energy sector,
for retaining or barrier structures, and for underground or underwater infrastructure [2.6-2.8].
Some related tasks associated with the transport properties of concrete are:
-

estimation of the degradation progress of a concrete section/member under given


exposure conditions
transfer of laboratory experiments on accelerated corrosion tests to the behaviour
of structures under field exposure
prediction or extrapolation of the development of degradation of a concrete or
reinforced concrete structure under given exposure to assess its remaining service
life (service life prediction)
assuring a planned service life of a structure by controlled materials behaviour and
structural detailing (service life design)
evaluation and design of barrier structures for environmentally hazardous
compounds such as Polycyclic Aromatic Hydrocarbons (PAHs) and containments
for dangerous wastes, e.g. radioactive materials and derived products
quality control testing of concrete production and execution of concrete structures
in a performance concept.

14

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

For an analytical treatment of these tasks the transfer of mass either in a gaseous state,
liquid state or dissolved state, must comply with the physical model describing a transport
process in which the mechanism of transfer is characterized by a driving force, by properties
of the moving species as well as by properties of the penetrated material. Under simplifying
assumptions the transport properties of the penetrated material are taken as a materials
characteristic or intrinsic value, which is often regarded as a constant but indeed may be a
function of several internal and/or external parameters. Consequently, apparent or effective
transport parameters are then often defined as a simplification of the real process.
As flow paths for transport through concrete, the capillary porosity of the matrix, the
paste/aggregate interface as well as microcracks are considered. Concrete is an aging material:
due to on-going hydration of the cement the porosity as well as the continuity of the capillary
porosity decreases with time, microcracking may increase due to drying shrinkage, and corrosive
interactions with the environment may increase or decrease the porosity of the matrix. Similar
considerations must be taken into account for the interaction of the penetrating species with the
porous material, i.e. the degradation process itself may change the porosity. Therefore, the
transport characteristics also change with time, i.e. aging is considered by time functions. This is
of considerable importance when transport properties are assessed in laboratory experiments on
young concrete specimens to provide input data for a long term assessment of structures, e.g. a
service life prediction. Sensitivity analyses on prediction models have demonstrated the
dominating effect of ageing parameters on the final result of the modelling [2.9].
Also the selection in isolation of a single transport mechanism for the ingress of a particular
substance may represent an over-simplification of the real transport process since more than one
transport mechanism may be active at a given time, either in parallel or in different sections
along the flow paths, so called mixed modes of transport. They play an important role in the
ingress of dissolved species, e.g. chloride ions, when capillary suction takes place in nonsaturated surface zones, whereas in deeper sections diffusion governs their transport.
It is common practice to distinguish those transport mechanisms which are important for
the above mentioned performance characteristics of concrete structures as diffusion,
permeation, capillary suction and migration. The distinction between these mechanisms
depends on the driving forces for the transport.
2.1

Diffusion

Brownian free motion of atoms, ions or molecules results in a net flow of mass from regions
of higher concentration to regions of lower concentration. The flux per unit area of a plane
perpendicular to the flux, q,

q=

dm 1
dt A

(2.1)

where q = mass flux (g/m2s)


m = mass of substance flowing (g)
t = time (s)
A = area (m2)

15

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

q is proportional to the concentration gradient dc/dx as the driving force for the transport, and
to the diffusion coefficient D. This relation is expressed in Ficks first law of diffusion [2.10]
q = D

dc
dx

(2.2)

where D = diffusion coefficient (m2/s)


c = concentration (g/m3)
x = distance (m)
The negative prefix considers that the flux occurs along a negative concentration gradient.
In these equations the concentration c refers to the phase in which the transport takes place,
e.g. the amount of gas molecules in the total volume of gas, the amount of dissolved ions in
the total volume of liquid etc. The diffusion coefficient D then describes the movement of the
species in the fluid phase. The same relations are applied also to porous media, where the
flow can only occur in the interconnected pore spaces, either filled with gas (air) or a liquid
(water). Then, the parameters D and c should refer only to the pore space and the fluid phase
contained there, considering an effective porosity of the solid material [2.11]. This effective
porosity is unknown in most cases and therefore, a homogenization of c and D is frequently
assumed, referring c to the gross volume and D to the transport characteristic of the porous
material as a simplification.
Ficks first law of diffusion in its general form is valid at any point in time and space for an
on-going diffusion process if local parameters are considered. For a steady state transport
problem with a constant diffusion coefficient D and constant boundary (surface) conditions c1
and c2, Ficks first law of diffusion can be re-written as
c c
q = D 2 1
(2.3)
x
where

c1, c2 = surface concentration at side 1 and side 2 of a separating material


x = distance of separation

The application of Ficks first law of diffusion to non-steady state transport processes leads
to the mass balance equation which describes the change of concentration in a unit volume
with time, i.e.
Change of concentration c with time = flux into element minus flux out of element (Fig. 2.1):

16

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

Fig. 2.1 - Mass balance for volume element in non steady state transport condition
flux into element:

c
q x = D
x x

(2.4a)

flux out of element:

c
q (x +x) = D
x x +x

(2.4b)

change of concentration:

c
c
c
x = D
+ D
t
x x +x
x x

(2.4c)

or

c
c
+ D
D
c
x x +x
x x
=
t
x

(2.4d)

and for x 0
2c
c
= D 2
t
x

(2.5a)

Equation (2.5a) is valid for D=const. If D is a function of the local concentration c or the
space variable x or time t then

17

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

c

c
= D

t
x
x

(2.5b)

The mass balance equation (2.5a) is usually referred to as Ficks second law of diffusion
[2.12]. Explicit solutions exist for this partial differential equation (PDE) for simple
geometries and constant boundary conditions as well as constant material properties, in
particular D = const. For instance

x
c(x, t) = c0 1 erf

2 Dt

(2.6)

where erf is the error function and the initial conditions (t = 0):
c = 0 for x > 0
c = c0 for x = 0
Equation (2.6) is often used to evaluate observed concentration profiles on laboratory test
specimens as well as on real structures after an exposure time t by means of a regression
analysis with respect to D, e.g. a least squares fit.
The ingress of ions into concrete is treated in general as a diffusion process. Nilsson [2.11]
points out that in this process, not only a transfer of mass occurs but also a transport of
electrical charge. In order to maintain electro-neutrality in the system the flow of anions must
be balanced by a corresponding flow of cations. If the transfer of charges cannot be balanced
a voltage difference will build up. In this electrical field one type of ion may be accelerated
while others can be slowed down in their movement. In this state the diffusion of ions will
overlap with migration as described in section 2.4 .
2.2

Permeation

Liquids and gases can percolate through interconnected pore spaces or crack networks of
cementitious materials under the driving force of an absolute pressure gradient [2.13, 2.14].
2.2.1 Gas permeability

Experiments on gas permeability are typically performed on specimens mounted in


pressure cells with a distinct difference in absolute gas pressure on the opposite ends of the
test specimen. Based on the assumption of a laminar flow in the capillaries of the test
specimens the permeation experiments are evaluated according to equation (2.7), which is
based on the Hagen-Poiseuille relation for a laminar flow of Newtonian fluids through
cylindrical tubes. Considering the mass conservation of the flowing gas it is essential to relate
the measured volume of flow to a corresponding pressure of the gas [2.14].
QL
2p
(2.7)
k=
t A ( p1 p 2 )( p1 + p 2 )
where k

= coefficient of gas permeability (m2)

18

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

= volume of gas flowing (m3)

= viscosity of gas (Ns/m2)

= thickness of permeated section (m)

= permeated area (m2)

= pressure at which volume Q is measured (N/m2)

p1

= pressure at entry of gas (N/m2)

p2

= pressure at exit of gas (N/m2)

= time (s)

There is experimental evidence, however, that eq. (2.7) does not account for all mechanisms
acting in a porous solid, since the observed coefficient of permeability kobs will depend on the
pressure level used in the experiments. Additional effects such as non-Newtonian fluid
characteristics, inertia effects causing turbulent flow patterns and Knudson or slip flow are
discussed in [2.15-2.17]. While non-Newtonian behaviour is not considered to have a major
effect on the flow volume [2.16], a turbulent flow pattern is likely to occur in the capillary pore
system due to the pore sizes, the tortuosity and the roughness of the pore walls. If the size of the
pores percolated is in the range or smaller than the mean free path length of the gas molecules
an additional flow component occurs, i.e. slip flow or Knudson flow. The mean free path length
of gas molecules at ambient temperature and atmospheric pressure is in the range of 10-7 m, and
it is inversely proportional to the gas pressure. In particular test methods applying pressures less
than 1 atmosphere should be affected more by slip flow than tests in the high pressure range,
since the free path length of the gas molecules then extends more into the range of capillary
pores contributing to the flow. In any case the observed coefficient of permeability kobs depends
on the applied pressure and it is no longer a material constant.
It is proposed to correct the observed coefficient of permeability kobs by the Klingenberg
term, to yield the intrinsic coefficient of permeability kint as follows [2.15, 2.17]:
kobs = kint (1+ /pm)
with

(2.8)

kobs

= observed coefficient of gas permeability (m2) at pm

pm

= mean absolute pressure (N/m2)

= Klingenberg number

The Klingenberg number can be derived from a series of permeability experiments at


different pressure levels pm. Plotting the observed coefficient of permeability kobs against
1/pm, a straight line should be observed with the intercept at kint for pm approaching infinity,
i.e. 1/pm approaching zero, see Figure 2. (N.B.: In [2.15, 2.17] the observed coefficient of
permeability kobs is referred to as apparent coefficient of permeability kapp).

19

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

kobs

non-viscous flow
. kint
kint

viscous flow

1/pm

Fig. 2.2 - Relationship between mean pressure pm and observed coefficient of permeability
kobs. [2.17]
In addition to the above mentioned influences from the pore structure and/or fluid
characteristics, the flow of gases through the pore spaces depends on the water content in the
pore system, since only the empty free pore space is available as flow paths. Therefore, the
apparent coefficient of permeability still depends on the degree of saturation of the pore
system. Since the slip flow is also affected by the water content, the Klingenberg number
depends on the degree of saturation as well. Both effects must be assessed experimentally.
The effect of the degree of saturation on the coefficient of gas permeability has been
investigated in numerous experiments: In the moisture range of oven dried specimens to
specimens in equilibrium with air of r.h.> 95 %, a variation of kobs and kapp has been observed
over three orders of magnitude [2.16, 2.18-2.20]. The effect of the water content on the
Klingenberg number is less pronounced: Carcasses et al. [2.20] reported a decreasing
contribution of the slip flow to overall flow with increasing water saturation.
The experimental assessment of the moisture effect on the gas permeability of concrete
relies on the assumption that the moisture is uniformly distributed in the volume of concrete
under test, i.e. no moisture gradients prevail along the flow paths. Therefore, in the past
different methods for preconditioning of test specimens were proposed aiming at a
homogenization of the water held in the specimens [2.16, 2.19, 2.21]. Concrete sections that
were subjected to drying will always exhibit moisture profiles, e.g. the near surface regions of
a concrete member, and the observed flow of gas by permeation will not be equal to the
volume flowing through a corresponding concrete section with the same mean moisture

20

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

concentration but uniformly distributed [2.16, 2.22]. Verdier and Carcasses [2.22] considered
the physical analogy of permeation with electrical conductance in order to analyze the effect
of known moisture profiles on the overall gas flow with the help of moisture dependent
coefficients of permeability. Tauscher [2.16] also found that the gas flow through sections of
variable moisture concentrations cannot be described by the flow parameter at a mean
moisture concentration but is strongly affected by the highest moisture content prevailing.
Drying may also induce microcracking: If cracking shows a preferential orientation, e.g.
perpendicular to the prevailing stresses, concrete sections will exhibit an anisotropic
behaviour with respect to gas permeability: Microcracks in parallel to the flow direction
drastically increase the volume flow while crack opening normal to the direction of gas flow
has a far lesser effect [2.23]. This effect must be addressed in gas permeability tests on
concrete surfaces exposed to drying conditions.
2.2.2 Water permeability

In contrast to gases, liquids may be considered as incompressible fluids. If the viscosity of


the liquid is taken into consideration and a laminar flow is assumed the coefficient of
permeability K is given by
K=

QL
t A p

(2.9)

where K = coefficient of permeability (m2)


Q = volume of liquid flowing (m3)
t = time (s)
L = thickness of penetrated section (m)
A = permeated area (m2)
= viscosity (Ns/m2)

p = pressure differential across specimen (N/m2)


K is a material parameter independent of the liquid considered. In view of concrete
durability as well as several serviceability aspects among those liquids penetrating into
concrete, water represents the most important fluid. Often, a coefficient of permeability is
then measured for water and the viscosity is neglected. The evaluation of corresponding
experiments follows an empirical approach developed by DArcy [2.24]:
Kw =

QL 1
t A h

(2.10)

with Kw = coefficient of water permeability (m/s)

h = pressure head applied (corresponds to the height of a water column) (m)

21

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

By equating (2.9) and (2.10) a relation can be found between the coefficient of permeability K
and DArcys coefficient as shown in equation (2.11):
Kw = K

(2.11)

with = density of water (1000 kg/m3)


g = gravity (9,81 m/s2)
= viscosity of water (0.001 Ns/m2 at 20oC)
The proportionality factor between Kw and K should be in the range of 107. Experiments
on the water permeability of cementitious materials do not confirm this proportionality, since
a continuous decrease of the percolating flow volume of water and thus of Kw is observed
even after extended test durations of several months [2.25, 2.26]. A similar phenomenon can
be observed in capillary suction experiments conducted with water. In section 2.3.2 likely
effects will be discussed.
Kw has the dimension of a velocity and it may be understood as an average velocity of
water flowing through the test specimen or concrete section with cross section A.
Considering that in porous media a flow can only occur in the interconnected pore spaces the
effective porosity can be taken into account. Thus,
K
K w eff = w
(2.12)
eff
may be understood as the average flow velocity of water in the pores.
In equation (2.12) eff represents the effective porosity.
2.3

Capillary Suction

2.3.1 General Mechanism and Parameters

Even in the absence of an external absolute pressure, porous media such as concrete can
take up liquids by the action of capillary forces. Surface forces of the liquids and solids are
responsible for this action which leads to a wetting of the internal solid surface in the capillary
pores (Fig. 2.3). The capillary pressure of the liquid (pore water pressure) is described by the
Laplace equation
2
(2.13)
pca p = pl pg =
r1
and r1 = r/ cos , then
pca p =

2 cos
r

(2.14)

22

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

where pcap = capillary pressure (N/m2)


p l = pressure in liquid (N/m2)
pg = pressure in gaseous phase
above liquid column (N/m2)

= surface tension (N/m)

= wetting angle

= radius of capillary (m)

r1 = radius of curvature of meniscus (m)


Fig. 2. 3 - Mechanism of capillary rise
With the cross sectional area of the capillary A= r2 the pore water pressure results in a
capillary force
f ca p =

2 cos r 2
= 2 r cos
r

(2.15)

For a wetting angle between 0 and 90 degrees fcap is a tension force.


Brauer [2.27] has discussed in depth further forces acting in the capillaries when a porous
section is in contact with a liquid, e.g. a horizontal surface covered with a liquid film.
Corresponding situations are given by spill of hazardous fluids or a de-icing salt solution on a
pavement. Aside from internal forces acting he also considered a low external pressure due to
a liquid level representing e.g. a spill of hazardous liquids in retaining structures such as a
concrete floor, c.f. Fig. 2.4.

23

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

fa = external force
fg = force induced by mass of liquid column
fr = force caused by friction at pore wall
fi = inertia force
fcap= capillary force
h = height of liquid level outside specimen
z = penetration depth
z = velocity in direction z
z = acceleration in direction z
r = capillary radius
= viscosity of liquid
liq= density of liquid
Fig. 2.4 - Forces acting in capillaries
The balance of forces in a capillary results in
(2.16)

fa + fg fr fi + fcap = 0

and
fa = liq a h r2

(2.16a)

fg = liq a z r2

(2.16b)

fr = 8/r2 z z r2

(2.16c)

fi = liq r2 z z

(2.16d)

In equations (2.16c) and (2.16d) z and z represent the first and second derivative of z
with time t.
The differential equation (2.16) can be solved for time t and penetration depth z and
subsequent integration yields
2
2 cos
r
+ pa
z=
t
r

(2.17)

24

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

In equation (2.17) an external pressure pa is considered for the liquid, e.g. given by the
density and the height of the liquid level.
Eq. (2.17) does not serve directly for the evaluation of a suction experiment although the
influencing parameters are demonstrated. However, the equation postulates a square root of
time relation for the development of penetration depth, or, assuming a homogeneously
distributed porosity, a square root of time relation for the absorbed volume of liquid. The
proportionality factors may represent
z(t) = B t
(2.18)

with

z (t) = penetration depth at time t (mm)


B

= penetration coefficient (mm/ t )

or
Ml(t) = C

(2.19)

Ml (t) = mass of liquid absorbed at time t per unit area of surface exposed to liquid
(g/m2)
C

= absorption rate (g/m2 t )

2.3.2 Capillary suction of water

In experimental studies with water as absorbed liquid it has often been observed that the
penetration depth or absorbed water Ml does not strictly follow a square root of time relation.
Empirical corrections are made then [2.19, 2.28] with
Ml(t) = D + C t

(2.20)

with D = initial absorption (g/m2)


or
Ml(t) = D + C tn ,

n < 0,5

(2.21)

Hall et al [2.29] have compared the capillary suction behaviour of cementitious materials
for water and a number of organic liquids that pose environmental risks and therefore are of
interest for retaining structures. Their results clearly indicate that the penetration of organic
liquids follows the expected square root of time relation with the surface tension and the
viscosity of the liquid as rate controling parameters, i.e (/ )1/2. However, water exhibited
the above mentioned anomalous penetration behaviour. They conclude that chemomechanical interactions of water with the cement paste occur, causing microstructural
changes in the paste: for pre-dried or re-saturated specimens on-going hydration of cement
particles may account for some changes in the pore system, as well as swelling of the paste.

25

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

Additional investigations showed that the decrease in penetration of water is accompanied by


external expansion. Since there is no noticeable effect on the total porosity it has been
concluded that the connectivity of the capillaries is affected by alterations in the paste.
It has been postulated that the same mechanisms act in permeability tests on concrete,
where a decay of the volume flow with time occurs for percolating water, whereas organic
liquids result in constant flow rates [2.29].
2.4

Migration

In section 2.1 it was pointed out that the diffusion of ions in a liquid provokes an electrical
field if the transport of cations is not balanced by a corresponding counter flow of anions.
Considering that in natural conditions as well as in most experimental set-ups not only one type
of ion is moving, and different ions have different mobilities, the build up of an voltage
difference is likely to occur. These phenomena are of particular interest in studying the ingress
of chloride ions into concrete, with a concrete pore solution containing various ions.
It has been claimed that in these cases simple diffusion tests evaluated on the basis of
Ficks diffusion laws do not reflect the transport mechanism prevailing in the actual process
and thus do not yield a valid material characteristic. The Nernst Planck equation (2.22)
describes the flow of mass due to the simultaneous action of a concentration gradient, an
electrical field and a flow of the solvent, i.e. convection [2.11, 2.30, 2.31]. The simultaneous
transport of multiple species and their charges in an arising electrical field has been studied
therefore, and a set of equations known as the Nernst-Planck equations must be solved for the
individual ions i participating in the transport process:
c z F
q i = D i i + i ci
+ cv
x RT x

(2.22)

with
q = mass flux (g/m2s)
D = diffusion coefficient (m2/s)
c = concentration (g/m3)
x = distance (m)
z = electrical charge
F = Faraday constant (J/V mol)
R = gas constant (J/mol K)
T = absolute temperature (K)
= electrical potential (V)
v = velocity of capillary flow (m/s)

26

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

Aside from the claimed shortcoming in the description of the real transport process, natural
diffusion tests in the diffusion cell require long test durations, especially when steady state
transport conditions are to be reached on concrete specimens with a relevant thickness.
Accelerated test methods impose an external electrical field as an additional driving force
for the transport of ions and the evaluation of the materials transport properties is performed
on the basis of the Nernst Planck equation. Water saturated specimens are used and the
convection term in equation (2.22) can then be omitted. This issue is further discussed in
Section 5.2: Ion Migration Methods.
2.5

Reference tests

Due to the increasing interest in transport phenomena in concrete, transport processes have
been extensively studied in recent decades. Following the distinction of processes according
to the driving forces the investigations were focused on diffusion processes, permeability
studies and capillary suction tests. The interest in migration processes in cement based
materials has arisen only recently.
Although the general principles of test methods were available from other disciplines, e.g.
ceramic industry, soil sciences, petrochemical industry, hydraulics etc. their application to
concrete involved a number of additional problems due to the specific properties of concrete, e.g.:
- the ageing of concrete due to on-going hydration
- reactivity of concrete with penetrating substances studied, for instance water,
carbon dioxide, chloride ions etc.
- variability of concrete properties with moisture content of concrete
- sensitivity of concrete pore structure to preconditioning, e.g. micro cracking upon
drying
- pore water composition, its effect on, and interaction with, transport processes.
It is evident that test methods must aim at controlling these effects, but also that
simplifying assumptions must be introduced and shortcomings will remain in the simulation
of the real process. The large number of influencing parameters may also be responsible for
the variety of test methods that have emerged each of them emphasizing some particular
aspects of transport by a special set up, test parameter or pre-treatment of the specimens,
while the general principles of the tests largely remain the same. In consequence, however, we
are far from having Reference Test Methods for the individual transport cases in a sense of
describing the true mechanism acting in reality, reflecting prevailing service conditions or
having a general acceptance in the research or engineering community. Despite these
limitations national and international standards and draft standards exist for some methods.
In RILEM Technical Committee 189-NEC, on-site test methods for transport mechanisms
were evaluated with regard to their suitability to indicate concrete durability characteristics,
taking into account their ease of handling, rapidity of performing a test as well as
uncertainties in the test conditions, in boundary conditions and in the state of the tested
concrete. In this context and for the purpose of comparison, reference is made to test methods
that draw on well defined test conditions and well known materials characteristics. However,
due to the additional efforts required, these are limited to laboratory application only. In [2.2]
a separate chapter presents well established laboratory test methods for transport parameters.

27

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

Furthermore, Alexander et al [2.32] have established a guide on the use and interpretation of
corresponding tests.
2.5.1 Diffusion tests

The general set-up of a diffusion test cell consists of a container which is separated by the
test specimen of given thickness into two chambers in such a way that the chambers contain
media of known concentrations c1 and c2 for the species under investigation, c.f. Fig. 2.5.

Fig.2.5 - Typical set-up of a diffusion cell


Due to the imposed concentration gradient, matter will flow from the chamber of higher
concentration (c2 in upstream chamber) through the test specimen into the chamber of lower
concentration (downstream chamber with c1). In the steady state of transport, i.e. dq/dt = 0, the
diffusion coefficient is calculated on the basis of Ficks first law of diffusion, c.f. section 2.1.
So-called ponding tests monitor the ingress of substances into a test specimen in non
steady state transport conditions: For instance, the test specimen is submerged into or ponded
with a test solution of concentration c for the substance under investigation, and after an
exposure period t the penetration of the substance is observed, either by determination of
penetration profiles or by a penetration front. The derivation of a transport coefficient can be
accomplished using Ficks second law of diffusion and regression of analytical solutions, e.g.
equation (2.6). Corresponding procedures for the actual tests are described in NT BUILD 443
[2.33] and prEN 13396 [2.34].
The cited standards deal with the ingress of chloride ions into cementitious materials. It
must be kept in mind that in non steady state conditions the effect of chloride binding in the
solid cannot be separated from the transfer of ions in the liquid phase but is implicitly
considered in the transport parameter. The calculated parameter is then defined as an apparent
diffusion coefficient.
2.5.2 Permeability

The test method developed by CEMBUREAU [2.35] is widely used for the measurement of the
gas permeability and procedures for the conduction of the test itself as well as for an appropriate
preconditioning of the test specimens have been developed, c.f. [2.16, 2.19, 2.21, 2.36].
Fig. 2. 6 gives an image of the set-up, with details of the pressure cell demonstrated.

28

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

2.5.3 Capillary suction

Capillary suction experiments are designed in such a way that the driving force of an
absolute external pressure is excluded or minimized. Therefore, the test specimen is exposed
with the bottom end in contact with the liquid under investigation; the immersion is limited to
a few millimetres. The lateral sides of the test specimens are sealed in order to prevent
disturbances and evaporation; also the top surface of the specimen should be sealed to prevent
evaporation from the interior or condensation from the ambient atmosphere. However, the
sealing must not be rigid, thus no pressure in the capillaries can arise in front of the
penetrating liquid front due to compressed air.
A typical set-up is given in Fig. 2.8, a description of the test and recommendations on the
specimen preparation are given in [2.32, 2.36-2.38].

Fig. 8 - Capillary suction test set-up


2.5.4 Migration

The layout of a migration test follows the same basic considerations as a diffusion test as
described in section 2.5.1, with the addition of an electrical field acting across the test
specimen. The electrical field is introduced into the test container by two electrodes which are
connected to a potentiostat. In chloride migration experiments the cathode is installed in the
up-stream cell and immersed in the catholyte, while the anode is immersed into the anolyte of
the down-stream cell. No general agreement has been reached on the voltage applied for the
electrical field: the desired accelerating effect of a high voltage is limited by possible heatbuild-up of the sample due to a high current. Voltages between 12 and 60 volts have been
recommended [2.32, 2.39-2.41].
The measurements during the operation of the test include monitoring of the flow of ions
by chemical analysis of the anolyte (and catholyte) or the conductivity of the anolyte as well
as the potential drop across the test specimen. In steady state conditions the diffusion
coefficient can be calculated according to the corresponding steady-state solutions of equation
(2.22). The ASTM C1202-97 method [2.42] measures the electrical charge passing the test
specimen.
With the same test set-up or at the termination of the migration experiment the resistivity
of the test specimen can be measured. The resistivity is related to the diffusion coefficient by
the Einstein relation

30

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

Deff =

k cl
= k cl
sa t

(2.23)

where Deff = effective diffusion coefficient for chloride ions in solution


kcl = test parameter depending on external ionic concentration
sat= resistivity of saturated concrete
= conductivity of saturated concrete
Instructions for the measurement of the resistivity are given for example in [2.43, 2.44].
Fig. 2.9 gives an overview for a typical set-up of a migration test. Fig. 2.10 presents the
details of the migration cell.

Fig. 2.9 - Experimental arrangement for migration tests

31

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

Fig. 2.10 - Migration cell with upstream and downstream compartment separated by the test
specimen.
Migration can also be studied in the non steady state condition, i.e. a penetration process is
monitored under the effect of an electrical field, and a penetration front is measured after a
test duration t. For chloride migration tests the penetration front of chlorides can be detected
by colorimetric tests after splitting the test specimens, e.g. spraying of silver nitrate. This
procedure is described in NT BUILD 492 [2.45].
REFERENCES

[2.1]
[2.2]
[2.3]
[2.4]

[2.5]

Basheer, L., Kropp J. and Cleland D. J., Assessment of the durability of concrete from its
permeation properties, Construction and Building Materials, vol. 15 (2001), pp. 93-103.
Kropp, J. and Hilsdorf, H. K. (Eds.), Performance Criteria for Concrete Durability,
RILEM Report No. 12, F&N Spon, London, 1995.
Helland, S., Performance of lightweight aggregate concrete in marine environment; 6th
CANMET/ACI Int. Conference on durability of concrete, Thessaloniki, June 2003.
Gehlen, Ch., Probabilistische Lebensdauerbemessung von Stahlbetonbauwerken
Zuverlssigkeitsbetrachtungen zur wirksamen Vermeidung von Bewehrungskorrosion,
Schriftenreihe des Deutschen Ausschusses fr Stahlbeton, Heft 510 (2000), Beuth Verlag
Berlin.
Andrade, C. and Kropp., J., (Eds), Proceedings 3rd Int. RILEM Workshop Chloride
ingress into Concrete, Madrid 2002, RILEM pro038, 462 p.

32

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

[2.6]
[2.7]
[2.8]
[2.9]
[2.10]
[2.11]
[2.12]
[2.13]
[2.14]
[2.15]
[2.16]
[2.17]
[2.18]
[2.19]
[2.20]
[2.21]
[2.22]
[2.23]
[2.24]
[2.25]
[2.26]
[2.27]

Verdier, J., Carcasses, M. and Ollivier, J. P., Modelling of a gas flow measurement
Application to nuclear containment vessels, Cement and Concrete Research, Vol. 32,
(2002), pp. 1331-1340.
Castellote, M., Andrade C., Alonso, C., Non destructive decontamination of mortar and
concrete by electro-kinetic methods: Application to the extraction of radio-active heavy
metals, Environment Science technology, Vol. 36, 2002, pp. 2256 2261.
Sosoro, M., Modell zur Vorhersage des Eindringverhaltens von organischen
Flssigkeiten in Beton, Schriftenreihe des Deutschen Ausschusses fr Stahlbeton, Heft
446, (1995), Beuth Verlag, Berlin.
Lazar, P.I.; Sensitivity analyses on prediction models, draft report on task 5 of EU
research project CHLORTEST, G6RD-CT-2002-00855.
Lykow, A. W., Transporterscheinungen in kapillarporsen Krpern, Akademieverlag,
Berlin, 1958.
Nilsson, L.-O. and Carcasss, M., Models for chloride ingress into concrete a critical
analysis, draft report on task 4.1 of EU project ChlorTest G6RD-CT-2002-00855.
Crank, J., Mathematics of diffusion, Oxford University Press, Oxford, 1970.
Carman, P. C., Flow of gases through porous media, Butterworth Scientific
Publications, London, 1956.
Zagar, L., Die Grundlagen zur Ermittlung der Gasdurchlssigkeit von feuerfesten
Baustoffen, Archiv fr das Eisenhttenwesen, Vol. 26, (1955), No. 12, pp. 777-782.
Abbas, A., Carcasses, M. and Ollivier, J.-P., The importance of gas permeability in
addition to the compressive strength of concrete, Magazine of Concrete Research, Vol 52,
(2000), pp. 1-6.
Tauscher, F., Einfluss des Wassergehaltes auf die Gaspermeabilitt von Mrtel und
Beton, Dissertation, Universitt Essen, 2005.
Abbas, A., Carcasses, M. and Ollivier, J.-P., Gas permeability of concrete in relation to
its degree of saturation, Materials and Structures, Vol. 32, Jan./Feb. 1999, pp. 3-8.
Jacobs, F., Permeability to gas of partially saturated concrete, Magazine of Concrete
Research, Vol. 50, (1998), pp. 115-121.
Kropp, J., Concrete durability an approach towards performance testing, Materials and
Structures, Vol 32, (1999), pp. 163-173.
Sugiyama, T., Bremner, T. W. and Tsuji, Y., Determination of chloride diffusion
coefficient and gas permeability of concrete and their relationship, Cement and Concrete
Research, vol. 26, (1996), pp. 781-70.
Carcasss, M., Abbas, A., Ollivier, J.P. and Verdier, J., An optimised preconditioning
procedure for gas permeability measurement, Materials and Structures, Vol. 35, (2002),
pp. 22-27.
Verdier, J. and Carcasses, M., Equivalent gas permeability of concrete samples subjected
to drying; Magazine of Concrete Research, 2004, Vol 56, No. 4, pp. 223 230.
Burlion, N., Skoczylas, F. and Dubois, Th.; Induced anisotropic permeability due to
drying of concrete, Cement and Concrete Research, Vol. 33, (2003), pp. 679-687.
DArcy, H. P. G.; Les fontaines publiques de la ville de Dijon; Dalmont, Paris, 1856.
Mills, R. H. and Kropp, J., Unpublished data on water permeability studies
Hearn, N., Saturated permeability of concrete as influenced by cracking and selfsealing, PhD Thesis, University of Cambridge, 1992.
Brauer, N.; Analyse der Transportmechanismen fr wassergefhrdende Flssigkeiten in
Beton zur Berechnung des Medientransports in ungerissene und gerissene Betondruckzonen;

33

Non-Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover


RILEM TC 189-NEC: State-of-the-Art Report May 2007

[2.28]
[2.29]

[2.30]
[2.31]
[2.32]
[2.33]
[2.34]
[2.35]
[2.36]
[2.37]
[2.38]
[2.39]

[2.40]
[2.41]
[2.42]
[2.43]
[2.44]
[2.45]

Schriftenreihe des Deutschen Ausschusses fr Stahlbeton, Heft 538 (2002), Beuth Verlag
Berlin.
Hall, C., Water sorptivity of mortars and concrete a review, Magazine of Concrete
Research, Vol 41, (1989), No. 147, pp. 51-61.
Hall, C., Hoff, W. D., Taylor, S. C., Wilson, M. A., Beom-Gi, Yoon, Reinhardt, H. W.,
Sosoro, M., Meredith, P. and Donald, A. M., Water anomaly in capillary liquid
absorption by cement- based materials, Journal of Material Science Letters 14 (1995) pp.
1178-1181.
Truc, O., Ollivier, J. P., A new way for determining the chloride diffusion coefficient in
concrete from steady state migration experiments; Cement and Concrete Research, Vol.
30, (2000), pp. 217-226.
Bockris, J. O. M. and Reddy, A. K. N., Modern Electrochemistry, Plenum Press, New
York, 1974.
Alexander, M. G., Mackechnie, J. R. and Ballim, Y., Guide to the use of durability
indexes for achieving durability in concrete structures, Research Monograph No. 2 ,
University of Cape Town and Witwatersrand, South Africa, 1999.
NT BUILD 443, Concrete, hardened: Accelerated Chloride Penetration. NorTest, VTT,
Finland 1995.
prEN 13396-1 (2002) Products and systems for the protection and repair of concrete
structures Test Methods, Part 1: Measurement of chloride ion ingress by diffusion of
repair mortars and concrete.
Kollek, J. J., The determination of the permeability of concrete to oxygen by the
CEMBUREAU Method a recommendation, Materials and Structures, Vol. 22, (1989).
Kropp, J., Preconditioning of concrete test specimens for the measurement of gas
permeability and capillary absorption of water, Materials and Structures, Vol. 32,
(1999), pp. 174-179.
Kropp, J., Determination of capillary absorption of water of hardened concrete, Materials
and Structures, Vol. 32, (1999), pp. 178-179.
Fagerlund, G., The critical degree of saturation method of assessing the freeze/thaw
resistance of concrete, Materials and Structures, Vol. 10, (1977).
Castellote, M., Andrade, C. and Alonso, C., Measurement of the steady state and non steady
state chloride diffusion coefficients in a migration test by means of monitoring the
conductivity in the anolyte chamber; Cement and Concrete Research, Vol. 31 (2001), pp.
1411-1420.
Luping, T., Chloride transport in concrete measurement and prediction, Chalmers
University of Technology, Publication P-96:6, 1996.
Andrade, C., Castellote, M. and Page, C.L., Are service life models appropriate for use in
design of reinforced concrete structures for exposure to chloride salts?, Proc. 3rd Int.
RILEM workshop on Chloride Ingress in Concrete, Madrid 2002, RILEM pro038, 462 p.
ASTM C1202-97 Standard method for electrical indication of concretes ability to resist
chloride ion penetration.
Monfore, G. E.; The electrical resistivity of concrete, Journal of the PCA Research and
Development Laboratories, May 1968, pp. 35-48.
RILEM Recommendation Test method for on-site measurement of resistivity of
concrete, Materials and Structures, vol. 33, Dec. 2000, pp. 603-611.
NT Build 492: Concrete, mortar and cement based repair materials: Chloride Migration
Coefficient from non steady state migration experiments, 1999.

34

Potrebbero piacerti anche