Print Go Back Next Page

MONITORING MOISTURE IN CONCRETE

WITH AN
EMBEDDED TRANSMISSION LINE
FS Malan & K Ahmet
University of Luton England

Summary: In monitoring concrete structures, long-term moisture monitoring is important as the

moisture plays a pivotal role in the chemical reactions of the corrosion process. The paper
describes the development of two versions of a low cost robust transmission line sensor operating at
1 GHz and capable of sensitive tracking. The sensor can be placed either directly into freshly cast
concrete or can be embedded within a porous medium for pre-calibration prior to installation
within existing structures. While the performance of the sensors is very encouraging, results show
that knowledge of the relationship in moisture transfer between the porous media is essential for
reliable moisture content measurement.

Keywords: moisture, concrete, transmission line

1 INTRODUCTION AND BACKGROUND
Moisture related problems are a leading cause of degradation in many building materials. Despite the importance of
measurement and monitoring, the accuracy, reproducibility and reliability of measurement is often limited especially where
hand-held moisture meters are employed. Problems in measurement using hand-held instrumentation were extensively
discussed previously, for example in Ahmet et al (1996). Remote moisture sensors are increasingly being used in conjunction
with associated electronics and circuitry to monitor moisture. Various types of wood-sensor are commonly employed where
the principle is based on the measurement of electrical resistance, which is a function of the moisture content. Tracking using
a datalogger provides useful information about the variation in moisture content and examples of such sensors were
extensively discussed in Dai and Ahmet (1999). Unfortunately, experience has shown that such sensors tend to be reliable for
short- and medium-term monitoring. For long-term monitoring (many years) necessary for concrete structures, the authors
vigorously sought and investigated alternative principles for measurement. A very promising solution is discussed in this
paper.
Long term moisture monitoring can provide early warning of changes inside concrete structures that may lead to possible
damage requiring remedial action. Extensive literature reviews concluded that there is an acute requirement in the construction
industry for a reliable, non-destructive and cost-effective method to track moisture variations in concrete and to monitor the
moisture content over many years or even decades (Parrot L. J., 1990) (Ahmet K., et al, 1999). It is recognised that while a
range of methods exists for the monitoring of moisture in construction materials, whether based on the determination of
moisture content or the relative humidity, these methods are predominantly aimed at either one-off measurements, or relatively
short term monitoring over a few weeks or months. Measurements should be made at various key points within the concrete
and not confined to surface or near-surface monitoring. Further, sensors should be capable of functioning reliably over the
long-term without the need for re-calibration and must be low-cost to be employed in large numbers (Ahmet K. and Malan F.
S., 2000).
The transmission line type sensor developed appears to fulfil the stringent requirements for long term monitoring. The
remainder of this paper justifies the various configurations of sensors investigated.
2 METHODOLOGY
The design was based on the investigations and results obtained by calibration with time domain reflectometry (TDR)1
measurements using a HP 8752C Network Analyzer. Permittivity measurements were done with a 5-conductor semi-coaxial
transmission line cast into the material under test (Fig. 1). There are two versions of the final probe transmission line, one with
and one without moisture absorbent material. Calibration was carried out using TDR and transmission velocity measurements
at discrete frequencies for the transmission line to determine the relative permittivity of the material plotted against moisture
content determined by oven drying and weighing.

1
TDR is the measurement of the time elapse between transmission of an electromagnetic pulse along a transmission line and
the reception of the reflection from the end of the line.

9DBMC-2002 Paper 141 Page 1

 Go Back Next Page

Volumetric moisture content measurements were used throughout and gravimetric results were converted to volumetric
moisture content for calibration purposes . This method is often used in microwave aquametry (Topp G.C. and Ferre P.A.,
2001).

16.0 43.4
Sensor line
Conductors BNC connector

Material under test

To TDR
Short circuit end plate

53.9

Figure 1Calibration line cast into material under test (dimensions in mm)

Cement
Conductor
PTFE

Screen

Insulation

Figure 2Line with cement paste filler around conductor

The final probe design consists of a radio frequency oscillator (operating at about 1GHz) that uses the transmission line as the
tuning device to determine the frequency. In this setup the frequency of oscillation is converted to an output voltage, which can
be monitored by means of a voltage datalogger. The output voltage is calibrated in terms of moisture content. Energy
consumption is minimal as it is supplied from an external source only when a reading is taken. Details were provided in Ahmet
and Malan (2000).
One major drawback of the short-circuited closed coaxial line used in the initial calibrations is that the material under test has
to be enclosed inside the “cage”. Three ways were investigated to overcome this problem.
a) The material inside the cage is allowed to take up the moisture from or to release it to the surrounding structure to measure
the moisture content (Fig. 1). This was the calibration setup.
b) The design is such that only part of the material in the cage is sensitive to moisture from the concrete (Fig. 2 shows that
the rest of the volume is filled with PTFE).
c) The transmission line is designed in such a way that it will measure the moisture in the wall of the drilled hole i.e. it is
sensitive to the material immediately surrounding the sensor (Fig. 3). In this case the electric field changes are affected by
the material surrounding the sensor. The dielectric material (PTFE) inside the sensor remains unchanged.

Line conductor
PTFE

MUT

Screen

Figure 3Off-set or “skew” line

2
The ratio of weight loss and therefore the volume of water lost to the volume of the sample.

9DBMC-2002 Paper 141 Page 2

 Go Back Next Page

3 MOISTURE IN CONCRETE
Although there are many types of concrete available, this discussion will be restricted to a common type, produced using
“Ordinary Portland Cement” (OPC) mixed with aggregates of varying particle sizes and reinforced with steel.
The hardened cement paste that is obtained by mixing cement powder with water is characterised by a porous microstructure
with high surface area (200 m2/g when fully hydrated, based on measurement by water adsorption) (Consolati G, et al, 2001).
Because mature hardened concrete is porous it absorbs water quite readily, but with such a mixture of materials the pores can
vary greatly in size from clearly visible air and/or water spheres, cracks and fissures to nanometre sized gel pores in the cement
paste. The types of pores can be divided into two main categories namely:
a) Spherical pores and other larger pores, cracks and voids in the concrete mix.
b) Capillary pores and gel pores.
Aggregates used in the concrete mix can vary in porosity from highly porous material to virtually non-porous sand grains and
crushed rock.

3.1 Spherical pores and cracks

Moisture can be trapped in residue material around the aggregate pieces and in cracks and fissures outside as well as inside the
cement paste. The authors used electron microscopy to qualitatively investigate these effects. An example is shown in Figure 4.
Trapped air in the mix forms spherical pores. The cement hydration process absorbs most of the water in the mix, but excess
water will diffuse through the matrix and evaporate, leaving capillary pores in the matrix. The capillary pressure (P, measured
in Pa) in a capillary tube (adapted from Lorenzo A.,1928) is given by:
P = 2γ/ r Equation 1
where: γ is the water/cement adhesion force (N m ) -1

r is the capillary radius or hydraulic radius (m)

Equation 1 shows that the capillary pressure increases as the inverse of the radius. Clearly, for nanometre scale pores this
represents a tremendous negative pressure (suction) that a capillary exerts on free water although the flow rate is extremely
slow. The implication is that with an increase in moisture the larger pores are the last to fill up with water, conversely they are
the first ones to become dry when there is a reduction in moisture content.

Figure 4. SEM photograph of a polished cement paste surface. Scale 8000:1.

Figure 5. Diagrammatic representation of the pore system in the cement paste

9DBMC-2002 Paper 141 Page 3

 Go Back Next Page

3.2 Gel and capillary pores in the cement paste.

Figure 4 gives an impression of the range of pore sizes found in cement paste (possibly near a crack in this case). Figure 5 is a
simplified diagrammatic representation of the pore system in the cement paste.
Gel pores are nanometre size pores (<10 nm), which form continuous pathways in the hydration product. The capillary pores
(10 nm < r < 1000 nm) are between the hydration product regions where excess water is trapped forming together with the gel
pores a continuous network across the material (Consolati G. et al, 2001)(Garboszi E. J., Bentz D. P., 1992). The implications
for the sensor design are discussed later.
4 MEASUREMENT OF MOISTURE IN CONCRETE
In the methodology used moisture content (MC) is calculated by determining how the relative permittivity εr of the material
under investigation varies with MC. The water has a much higher εr than the dry material. Extensive TDR measurements were
used for the calibration of the sensors to determine the reflection time from the end of the line in picoseconds and εr is
calculated from:
εr = (measured TDR time/calculated reflection time for the line in free space)2 Equation 2

4.1 Volumetric moisture content.

In this paper volumetric moisture content is used throughout, providing an ‘absolute moisture level’. Briefly the reasoning is
that as concrete consists of a mixture of different materials, any meaningful measure of moisture in a mix of materials should
be independent of the density of the material. That is why the measured moisture content (MC), defined in Eq. 3 below, is in
terms of a percentage of total volume (absolute moisture level).
MC = 100×(total volume of water/ total bulk volume)% Equation 3
3
In the rest of the paper absolute moisture content will be referred to as MC .

4.2 Relative humidity (HR) as a measure of moisture content

Under normal environmental conditions, concrete structures are frequently subjected to running or standing water. Except for a
small amount of trapped air even the mature concrete could be fully saturated with water. It is common practice to use
measurement of HR inside a sealed cavity as an indication of moisture content or dryness of concrete. It was shown in section
3.1 that as the concrete dries out, the moisture becomes ‘concentrated’ into the smaller and smaller capillaries. The relationship
between HR and pore size filled with water is given by the following formula of equation 4 (adapted from Bentz D.P., et al ,
1999)
ln( HR/100) = −2Vm γ/rRT Equation 4
where
HR = relative humidity (%)
Vm = molar volume of water (m3 mol-1)
γ = adhesion force between water and solid (N m-1)
r = capillary radius (m)
R = Universal gas constant (J K-1)
T = Absolute temperature (K)
At constant temperature in a homogenous porous solid, Equation 4 can be reduced to:
HR =100 e-C/r Equation 5
where C = 2Vm γ/RT is constant for given values of γ and T.
Using Equation 5 it can be shown that HR will reach close to 100% when the largest capillaries containing water are still in the
nanometre region. Equation 5 further shows that HR is a measure of the size of gel pores (r) still filled with water. HR
measurement can be translated to moisture content when the total volume of pores smaller than diameter r is known. Although
HR is therefore a good qualitative indication of the “dryness” of the concrete, accurate measurement of relative permittivity εr
is superior; it provides a more direct indication of moisture content and can measure effectively any level of moisture content.

3
Volumetric moisture level can be converted to gravimetric moisture level by dividing the volumetric moisture level by the
density of the dry material in g.cm-3.

9DBMC-2002 Paper 141 Page 4

 Go Back Next Page

4.3 Change in relative permittivity of concrete with moisture.

The dipolar character of water molecules results in a relative permittivity much higher than that of the dry material (about 76
compared to 4 for the dry material). As already mentioned moisture content can therefore be very effectively determined by
measuring the relative permittivity of the material under test. Figure 6 are typical results showing the variation of relative
permittivity as a function of the moisture content for 50% sand/cement mortar and cement paste. Measurements were carried
out with time domain reflectometry (TDR) using the fully embedded transmission line shown in Fig. 1. The w/c ratio was 0.5
for both the cement paste and the mortar. Although the best fit is a fourth order polynomial, a straight-line fit as shown is good
enough for practical purposes.

4.4 Advantages and disadvantages of the “skew line” sensor.

Figure 3 shows the position of the line conductor in this type of sensor, which determines how much of the electromagnetic
(EM) field lines penetrate the material under test (MUT). When the line conductor is placed nearer to the outside edge, the EM
field lines will penetrate deeper into the MUT. With the EM field penetrating more of the surrounding material the sensor will
be more sensitive to changes in moisture content. However, it would also be more sensitive to the presence of disturbances like
steel reinforcement and other discontinuities in the concrete. If the sensitivity is reduced (the conductor moved towards the
centre of the probe) the probe will not detect very low moisture levels. Optimum configurations are currently being
investigated.
For unambiguous calibration the distance between the line and the MUT must be constant. The probe is positioned with the
line conductor side of the sensor pushed firmly against the side of the hole to achieve this.

4.5 Advantages and disadvantages of transmission line with absorbent element.

The probe in Fig. 2 can be calibrated accurately for moisture content in the absorbent material of the probe. The probe is also
unaffected by nearby metallic objects (steel reinforcement) or proximity to a sudden break or change in the MUT. It can be tied
to the steel reinforcement if the probe is to be cast permanently into the concrete or placed close to the surface or edge of the
MUT.
The main disadvantage of using a porous material in the sensor that absorbs moisture from the concrete under test, is the long
time delay to reach equilibrium. This arrangement is therefore suitable for long term moisture monitoring and not for a quick
assessments of moisture levels.
Equation 1 shows that the moisture content in the MUT and the probe absorbent depend on the relative capillary sizes. The
moisture content is also be a function of the ratio of porous and non-porous materials in the MUT. The porosity of aggregates
varies and could be of the same order of magnitude as the porosity of the cement paste. At very high levels of moisture close to
saturation the probe material is likely to be saturated before the MUT if the latter contains cracks and voids.
Using the probe on a different material to that in the line requires knowledge of the relative pore sizes of the MUT. From
equation 1 it is evident that cement paste, which has pore sizes in the nanometre region, could absorb all the moisture in the
MUT with larger pores, indicating a much higher moisture content than is actually the case. This type of sensor is material
specific, which means it can only be used to measure moisture in concrete if the filler material is cement paste.
5 TEST RESULTS
TDR measurements were carried out on the calibration line (Fig. 1) using the HP 8752C Network Analyzer. Results showing
the correlation between the relative permittivity and moisture content are shown in Figures 6. The shorter dotted line is for
mortar with 1:1 ratio of sand and cement. In both materials the w/c ratio was 0.5. Notice the excellent correlation between the
relative permittivity and the moisture content. Further, the calibration line for the mortar is very close to that of the cement
paste.

Although the best fit trend line is a fourth order polynomial, the straight line approximation is good enough in both cases for
most practical purposes, where R2 > 0.995. The small pore sizes imply that the diffusion rate through the material under test
would be very slow. With a drying out cycle the moisture at the core could have been higher compared to a position near the
outer surface, in which case the relative permittivity would be measured higher than it should be. A wetting cycle should show
the opposite effect. The actual measurement of relative permittivity (εr) could be expected to be half way between the cycles.
Even if a relatively long time is allowed to reach a homogeneous moisture distribution, the fact that diffusion can only take
place when there is a moisture gradient in the material means that some hysteresis effect will always be present. Another
reason why the wetting and drying cycle may differ is the fact that there are clumps of hydration product (gel) in the material
surrounded by larger gaps (see Fig. 4). The result is that on a drying cycle some moisture may be trapped in the nanometre
pore system of the gel. The trapped moisture will eventually dry out as the humidity decreases, but the process is very slow.
The εr for the dry material in both cases is very close namely 4.96 and 4.98 for the cement paste and the mortar respectively.
Using the linear approximation for εr at 10% moisture level the values are 7.78 and 7.70 respectively, i.e. excellent agreement
for the two materials.

9DBMC-2002 Paper 141 Page 5

 Go Back Next Page

Results of cement paste and mortar test with TDR

12

11

R2 = 0.9971
10
Relative permittivity

9 Cement paste
Mortar
Linear (Mortar)
8 Linear (Cement paste)
R2 = 0.9952

5
0% 5% 10% 15% 20% 25%
Moisture content

Figure 6 Test results with cement paste and mortar dielectric with TDR

The results also reveal that the saturation level for the cement paste is about twice as high as for the mortar, showing that the
sand in the mortar with a 1:1 sand to cement mixture ratio took up a negligible amount of moisture (tested at normal
atmospheric pressure). This is in accordance with expectations.
Finally it is important to realise that TDR measurements are broad band frequency measurements. The relative permittivity
changes with frequency so that the value obtained in Figure 6 with TDR is an average of all the relative permittivity values in
the frequency band 300 kHz to 6 GHz employed on the instrument.
6 CONCLUSION
This paper has described a method to determine the moisture content of concrete as part of a continuous monitoring
programme. Used in conjunction with sensors to detect for example physical movements, the very useful information will give
early indication of deterioration of the structure.
There are a number of configurations in which the probe could be employed and two possibilities have been summarised in
this paper. Present work is being carried out to find the optimum solution for each application. Indications are that there is no
single optimum solution that will satisfy all applications or moisture conditions simultaneously. As an array of sensors would
normally be used in a large structure, these sensors could be a mixture of types with different dynamic ranges to cover MC
from dry to saturation levels, depending on the range of interest.
This transmission line MC probe appears to be a promising technique to determine the absolute moisture content at a specific
location in the concrete. The “skew line” design has the potential to measure MC up to virtually full saturation levels. In this
arrangement no time delay would have to be allowed for the probe to absorb moisture. With the relative low cost of the probe
and the fact that it can be left in situ for many years due to its very robust nature, it appears to overcome many of the problems
with present day techniques. Despite this, knowledge of the limitations and nature of moisture content equilibrium in various
concrete types or other porous media is required to determine the absolute moisture content. Although the intention was to
design sensors to be mounted in hardened concrete, both types of sensors described could be embedded in the concrete when it
is cast in order to monitor the setting and drying of the concrete and then left in situ for long term monitoring of the structure.
7 REFERENCES
1. Ahmet K. et al. 1999, The long term monitoring of moisture in concrete structures, Project report for the Concrete
Bridge Development Group.
2. Ahmet K and Malan F. S. 2000, Measurement of Moisture in Concrete at Microwave Frequencies using a transmission
line sensor, 20th Cement and Concrete Science Conf., Institute of Materials, Sheffield, 67 – 70.
3. Ahmet K. Hall G., Jazayeri S. 1996, Standardisation of conductance-type timber moisture meters, Proceedings of the
7th International Conference on the Durability of Building Materials and Components, E. & F. N. Spon, 673-681
4. Bentz D.P., Garboczia E.J., Haecker C.J., Jensen O.M., 1999, Effects of Cement Particle Size Distribution on
Performance Properties of Portland Cement-Based Materials, Cement and Concrete Research, 29 (10), 1663-1671.

9DBMC-2002 Paper 141 Page 6

 Go Back

5. Consolati G., Dotelli G., Quasso F. 2001, Drying and Rewetting of Mature Cement Pastes Studied through Positron
Annihilation Lifetime Spectroscopy, J. of the American Ceramic Soc., 84 (1)
6. Dai G. and Ahmet K. 1999, Mathematical Modeling of Sorption in Timber, Proceedings of the 8th International
Conference on the Durability of Building Materials, Vancouver, NRC Research Press, 704-713
Garboszi E.J. and Bentz D.P. 1992, Computer Simulation of the Diffusivity of Cement-Based Materials. J. of Material
Science, 27, 2083-2092
8. Lorenzo A. 1928, The Usefulness of Capillary Potential to Soil Moisture and Plant Investigators, J. of Agric. Research,
37, 719-742
9. Parrot L. J. 1990, A review of methods to determine the moisture conditions in concrete, British Cement Association,
BCA C/7, 28.
10. Topp G.C., Ferre P.A., 2001, Electromagnetic Wave Measurement of Soil Water Content: a State-of-the-Art, Proc. 4th
International Conference on Electromagnetic Wave Interaction with Water and Moist Substances, Weimar Germany,
327-335.

9DBMC-2002 Paper 141 Page 7