ArticlePDF Available

Measuring Freeze and Thaw Damage in Mortars Containing Deicing Salt Using a Low Temperature Longitudinal Guarded Comparative Calorimeter and Acoustic Emission (AE-LGCC)

Authors:

Abstract and Figures

Deicing salts are often applied to the surface of pavements and bridge decks in the winter to melt ice; thereby improving safety for the travelling public. In this paper, the influence of NaCl deicing salt on freezing and thawing temperatures of pore solution and corresponding damage of mortar specimens were investigated. A low-temperature longitudinal guarded comparative calorimeter (LGCC) was developed to cool down a mortar sample at a rate of 2 °C/h and to re-heat the mortar at a rate of 4 °C/h. Heat flux during freezing and thawing cycles was monitored, and the temperatures at which freezing and thawing events occurred were detected. During cooling and heating, acoustic emission activity was measured to quantify the damage (cracking) due to aggregate/paste thermal mismatch and/or phase changes. The results show that NaCl solution in a mortar sample freezes at a lower temperature than the value expected from its bulk phase diagram due to under-cooling. Conversely, the frozen solution in mortar melts at the same melting temperature as the bulk frozen NaCl solution. As the salt concentration increases, the freezing temperature is lowered. For samples containing more highly concentrated solutions, an additional exothermic event is observed whose corresponding temperature is greater than the aqueous NaCl liquidus line in the phase diagram. Damage also begins to occur at this temperature. For mortar samples saturated by solutions with 5 % and 15 % NaCl by mass, greater freeze/thaw damage is observed. The AE calorimeter developed herein is applicable for investigating damage behavior during freezing and thawing of different phases in pore solution (in mortars).
Content may be subject to copyright.
0
Advances in Civil Engineering Materials, 3 (1), 23 pp., 2014.
1
Measuring Freeze and Thaw Damage in Mortars
2
Containing Deicing Salt Using a Low Temperature
3
Longitudinal Guarded Comparative Calorimeter and
4
Acoustic Emission (AE-LGCC)
5
6
Yaghoob Farnam(1), Dale Bentz(2), Aaron Sakulich(3), Daniel Flynn(4), and Jason Weiss(5)
7
8
9
10
(1) Graduate Research Assistant, Ph.D. Student, School of Civil Engineering, Purdue University, 550
11
Stadium Mall Dr., West Lafayette, IN 47907, USA, yfarnam@purdue.edu
12
13
(2) Chemical Engineer, Materials and Structural Systems Division, National Institute of Standards and
14
Technology, 100 Bureau Drive, Stop 8615, Gaithersburg, MD 20899, USA, dale.bentz@nist.gov
15
16
(3) Assistant Professor, Department of Civil and Environmental Engineering, Worcester Polytechnic
17
Institute, 100 Institute Road, Worcester, MA 01609, USA, arsakulich@wpi.edu
18
19
(4) Research Associate, Energy and Environment Division, National Institute of Standards and
20
Technology, 100 Bureau Drive, Stop 8615, Gaithersburg, MD 20899, USA, daniel.flynn@nist.gov
21
22
(5) Professor, Director of Pankow Materials Laboratory, School of Civil Engineering, Purdue University,
23
550 Stadium Mall Dr., West Lafayette, IN 47907, wjweiss@purdue.edu
24
25
26
ABSTRACT
27
28
Deicing salts are often applied to the surface of pavements and bridge decks in the winter to melt ice;
29
thereby improving safety for the travelling public. In this paper, the influence of NaCl deicing salt on
30
freezing and thawing temperatures of pore solution and corresponding damage of mortar specimens were
31
investigated. A low-temperature longitudinal guarded comparative calorimeter (LGCC) was developed to
32
cool down a mortar sample at a rate of 2 °C/h and to re-heat the mortar at a rate of 4 °C/h. Heat flux
33
during freezing and thawing cycles was monitored, and the temperatures at which freezing and thawing
34
events occurred were detected. During cooling and heating, acoustic emission activity was measured to
35
quantify the damage (cracking) due to aggregate/paste thermal mismatch and/or phase changes. The
36
results show that NaCl solution in a mortar sample freezes at a lower temperature than the value
37
expected from its bulk phase diagram due to under-cooling. Conversely, the frozen solution in mortar
38
melts at the same melting temperature as the bulk frozen NaCl solution. As the salt concentration
39
increases, the freezing temperature is lowered. For samples containing more highly concentrated
40
solutions, an additional exothermic event is observed whose corresponding temperature is greater than
41
the aqueous NaCl liquidus line in the phase diagram. Damage also begins to occur at this temperature.
42
For mortar samples saturated by solutions with 5 % and 15 % NaCl by mass, greater freeze/thaw damage
43
is observed. The AE calorimeter developed herein is applicable for investigating damage behavior during
44
freezing and thawing of different phases in pore solution (in mortars).
45
46
47
Keywords: Acoustic Emission, Calorimeter, Concrete, Deicing Salts, Freeze and Thaw, Heat Flow,
48
Mortar, NaCl.
49
1
1.0 Introduction
1
Concrete can be damaged when first, it is sufficiently wet (has a high degree of saturation) and second, it
2
experiences temperature cycles that enable freezing and thawing. The damage that occurs due to
3
freezing and thawing can lead to premature deterioration, costly repairs, and early replacement of
4
concrete infrastructure elements. In cold climates, when the degree of saturation exceeds 86 % to
5
88 % [1], freeze-thaw damage is inevitable, even after only a few freeze/thaw cycles. To reduce this
6
freeze/thaw damage, engineered air voids (e.g., air entrainment) can be considered in mixture
7
proportioning [2].
8
Deicing salt is used to depress the freezing point of the water on the concrete surface. Although addition
9
of deicing salts on the surface of concrete elements can melt the ice and increase the safety of
10
infrastructure. The molten salty solution can be absorbed into concrete pores and can also cause severe
11
damage to concrete structures by inducing corrosion of reinforcing steel (as many deicing salts contain
12
chlorides), scaling of the concrete, or salt crystallization which causes damage to the concrete.
13
14
1.1 Mechanisms of freeze and thaw damage in Concrete
15
During freeze/thaw deterioration of concrete exposed to deicing salt, a variety of complex damage
16
mechanisms occur [29]. First, the formation of ice inside the concrete pore structure produces hydraulic
17
pressure [3]. Recognition of the role of hydraulic pressure in freeze/thaw damage led to the initial
18
introduction of air voids into concrete [2]. In addition to hydraulic pressure caused by 9 % volume
19
expansion of water freezing in large cavities, the osmotic pressure resulting from partial freezing of
20
solutions in capillaries can be another source of deterioration in concrete [4]. In fact, the difference in
21
solution concentration between partially frozen solution in larger pores and unfrozen solution in smaller
22
pores causes the transport of water from smaller pores to larger pores causing an additional pressure
23
(i.e., osmotic pressure). Another form of destructive expansion was observed by Beaudoin and McInnis
24
[5] where expansion of the cement paste occurred when the cement paste pore structure was saturated
25
by benzene instead of water, which contracts upon freezing. Another hypothesis in frost damage
26
described by Litvan [6] was large-scale migration of unfrozen water from small pores (high entropy region)
27
2
to large cavities filled by frozen water (low entropy region). In fact, the difference in energies in the two
1
regions causes large-scale migration which could produce further expansion and more damage.
2
In addition, when ice crystals form in pores, stresses are generated by their crystallization that depend on
3
pore size, the energy of the interface between the pore wall and the crystal, and the yield stress of the
4
crystal [7]. An increasing concentration of salt produces a greater depression of the freezing point;
5
however, crystallization of salt may occur inside the pores. The pressure caused by salt crystallization is
6
strong enough to cause severe damage in concrete [8,9].
7
Most of the hypotheses that have been introduced to explain freeze/thaw damage are only partially
8
validated; more research needs to be performed to understand the role of deicing salt in the damage
9
mechanisms of paste, mortar, and concrete. To do so, powerful experimental techniques are required to
10
detect and define the damage mechanisms caused by different components inside a porous cementitious
11
material exposed to deicing salt. This can be done by either heat flow analysis of phase changes in pore
12
solution or acoustic emission detection of microcrack and crack formation. Therefore, these two
13
techniques are combined into a single experimental setup in the present paper.
14
15
1.2 Background on Heat Transitions of Liquid/Solid Phase Changes in Concrete Pore Solution
16
Pure water freezes at 0 °C at atmospheric pressure. However, water inside concrete pores is not pure, as
17
it contains several soluble species such as various alkalis, chloride, and hydroxide. The presence of ions
18
in pore solution causes that pore solution to freeze at a temperature lower than what is expected of pure
19
water [9]. In addition, a further decrease of the freezing point of pore solution in porous materials occurs
20
due to surface interactions, confinement, and the presence of capillary pressures [7,10].
21
In cold climates, when deicing salts are added to the surface of pavements and bridge decks, the ions
22
concentration in the pore solution is increased. In general, using more salt produces a higher
23
concentration of ions in pores and leads to greater depression of the freezing point. Fig. 1 shows a phase
24
diagram of aqueous NaCl. In this phase diagram, four phases exist including aqueous NaCl solution, ice,
25
NaCl, and NaC2H2O (hydrohalite). When one phase transforms to another phase, latent heat must
26
either be released or absorbed. Therefore, energy is transferred, which can indicate that a freezing or
27
3
thawing phase change is occurring [6,11]. The amount of latent heat and its corresponding freezing or
1
thawing temperatures are specific material properties that can be used to differentiate different materials
2
and different phases in a composite system.
3
Low temperature heat flow studies of porous cementitious materials have conventionally been performed
4
on very small cement paste samples [6,10,12]. Low temperature differential scanning calorimeters have
5
been often used to investigate phase transformations of water and NaCl solution in hardened cement
6
paste. The low temperature calorimeter has been also used to define the freezing and thawing behavior
7
of small mortar samples [13,14]. Despite the ability of these techniques to accurately measure heat flows
8
of different phase transitions, the small size of samples prohibits the testing of composite systems
9
containing fine and coarse aggregates. A longitudinal guarded comparative heat flow technique [15] is
10
often used to quantify the heat flow and thermal properties of composites, and could be applied for mortar
11
and concrete samples. In this research, this heat flow technique is used to quantify the heat flow in mortar
12
samples. It is anticipated that this can be extended to concrete samples.
13
14
15
Fig. 1 Phase diagram of aqueous NaCl.
16
17
4
1.3 Background on Use of Acoustic Emission (AE) to Quantify Freeze-Thaw Damage
1
Different acoustic emission (AE) techniques are conventionally employed to quantify the damage in
2
concrete caused by destructive phenomena [1619]. They can also be used to detect damage caused by
3
freeze/thaw cycles in concrete [1,2026]. A resonant frequency technique or a pulse/wave velocity
4
technique can be used to quantify the extent of damage in concrete samples exposed to freezing and
5
thawing [2023]. The main drawback of the resonant frequency and the pulse/wave velocity techniques is
6
that they are conventionally limited to discrete measurements of bulk sample deterioration. In other
7
words, continuous measurement is almost impossible using these techniques and they give almost no
8
information about the moment at which micro/macro cracks are generated, and whether they are
9
generated within paste, aggregate, or interfacial transition zones. In an advanced method, a passive
10
acoustic emission technique is used alongside an active or pulse velocity technique to understand the
11
behavior of concrete cracking during freezing and thawing [1,8,2426].
12
13
1.4 Research Significance
14
Most of the hypotheses described for freeze/thaw damage are only partially validated and more
15
investigations are required. In addition, a testing technique which can detect both damage and heat flow
16
during liquid/solid phase changes inside a composite pore structure is not readily available. When deicers
17
(such as NaCl) are used to depress freezing points, complex damage phenomena may occur that cannot
18
be described by existing experimental devices.
19
The main objectives of this study are: first, to develop an experimental technique that can characterize
20
freeze/thaw damage action; second, to determine and correlate damage and phase change events by
21
monitoring the acoustic and thermal behavior of mortar samples containing NaCl solution; and third, to
22
detect how NaCl influences the damage mechanisms, freezing temperature, and melting temperature of
23
different phases in saturated mortar samples.
24
25
5
2.0 Experimental Program
1
2.1 Materials and mixture proportioning
2
Ordinary Type I portland cement (OPC) with a fineness of 375 m2/kg was used in this study. The chemical
3
composition of this cement is indicated in Table 1. Aggregates used in this study consisted of natural
4
sand with a maximum size of 4.75 mm, specific gravity of 2.61, and an absorption value of 2.2 % by
5
mass. A single mortar mixture was used with a sand volume fraction of 55 % and a water-to-cement ratio
6
(w/c) of 0.42 by mass. No chemical admixtures or supplementary cementitious materials were used.
7
Mixture proportioning for making the mortar samples is shown in Table 2.
8
9
Table 1Properties of Type I ordinary portland cement (OPC).
10
Item
Percent by mass (%)
Silicon Dioxide (SiO2)
19.43
Aluminum Oxide (Al2O3)
5.39
Ferric Oxide (Fe2O3)
3.18
Calcium Oxide (CaO)
63.45
Magnesium Oxide (MgO)
2.97
Sulfur Trioxide (SO3)
3.38
Loss on Ignition
0.88
Sodium Oxide
0.35
Potassium Oxide
0.77
Insoluble Residue
0.25
Total Equivalent Alkali as Na2O
0.86
Tricalcium Silicate (C3S)
60
Dicalcium Silicate (C2S)
10
Tricalcium Aluminate (C3A)
9
Tetracalcium Aluminoferrite (C4AF)
10
11
Table 2Mixture proportions of mortar (Saturated-Surface-Dry)
12
Item
Type
Unit
Cement
Type I
kg/m3
Aggregate
Sand (0-4.75mm)
kg/m3
Water
-
kg/m3
w/c
-
Ratio by mass
6
2.2 Sample Preparation and Conditioning
1
The mortar was prepared in a standard mortar mixer in accordance with ASTM C305-12 [27]. The cement
2
was mixed with water for 30 s at the low speed. Sand was then slowly added to the mixer within 30 s,
3
while the mixture was being mixed at the low speed. The entire mixture was mixed for another 30 s at
4
medium speed. The mixing was stopped for 90 s as the sides and bottom of the mixing bowl were
5
scraped. A final 90 s of mixing was at medium speed. The mortar was cast in 25.4 mm × 25.4 mm × 300
6
mm (1 in × 1 in × 11.81 in) molds that were demolded after 24 h. All mortar bars were then sealed in
7
double plastic bags and cured for 28 d in these sealed conditions at 23 °C ± 0.5 °C. After 28 d of curing,
8
the mortar bars were cut using a wet saw to 25.4 mm × 25.4 mm × 50.8 mm (1 in × 1 in × 2 in) samples.
9
Samples were then placed in a vacuum oven at 65 °C ± 1 °C and a pressure of 20 mm Hg ± 5 mm Hg for
10
7 d. The ± 1 °C and ± 5 mm Hg are indicative of the nominal operating range encountered when running
11
the experiment.
12
The samples were then placed in a desiccator using two small spacers underneath each sample to
13
provide a small gap between the bottom of the container and the lower surface of sample. The samples
14
were evacuated to a pressure of 10 mm Hg ± 5 mm Hg for 3.5 h. After evacuation and while still under
15
vacuum, de-aerated NaCl solution (de-aerated by vacuuming the solution for 15 min) was introduced into
16
the desiccator to cover the samples for 1 h. Soaked samples inside NaCl solution were transferred to a
17
23 °C ± 0.5 °C chamber before testing, where they were kept in a beaker for 3 d. This condition was
18
considered as fully saturated (i.e., assuming 100 % degree of saturation).
19
After conditioning, samples were wrapped with a thin plastic sheet to protect the samples against
20
subsequent moisture exchange with their surrounding environment (preventing them from absorbing or
21
releasing water during the freeze-thaw process). The top and bottom cross section plastic covers were
22
removed to ensure better connection with thermal pads during the freeze/ thaw tests. In addition, a small
23
circular hole was made in the side plastic of each sample to attach the AE sensor.
24
25
7
3.0 Testing Design and Procedure
1
Different samples were prepared for freeze/thaw testing. Samples were saturated by different NaCl
2
solutions. Pure salt and de-ionized water were used to prepare NaCl solutions with (0, 0.7, 3, 5, 6, 8, 10,
3
13, 15, 18, 23.3, and 26) % NaCl by mass. Dry control samples were also prepared for comparative
4
testing.
5
6
3.1 Low Temperature Longitudinal Guarded Comparative Calorimeter (LGCC)
7
During solid/liquid phase changes, a large amount of heat is released or absorbed. The amount and rate
8
of this heat exchange is an indication of different components and their volume fractions within a system.
9
Under steady state thermal conditions, the heat flow of a material can be measured by means of the
10
longitudinal guarded comparative heat flow technique [15,28]. In this technique, one-dimensional heat
11
flow is produced by using a heat sink and a longitudinal guarded (insulated) shell. Two samples of known
12
thermal properties are used as meter bars to comparatively measure the heat flow through a sample with
13
unknown thermal properties.
14
In this study, a LGCC was constructed according to ASTM E1225-09 [15] and ASTM D5470-12 [28] to
15
produce and to quantify heat flow. A test sample was inserted between two 25.4 mm × 25.4 mm ×
16
25.4 mm (1 in × 1 in × 1 in) meter bars with known thermal properties. In this study, Pyroceram Code
17
9606 material
1
was used as the meter bar material. The thermal conductivity of this material as a function
18
of temperature can be calculated by Eq. 1 [29]:
19
 
(1)
where PC is the thermal conductivity of the Pyroceram (W/(m·K)) and T is its temperature (°C).
20
A temperature gradient was established in the test sample by using a two stage cold plate (Cascade
21
CCP-221 [30]) as shown in Figs. 2a and 2b. The two stage cold plate is capable of decreasing the
22
temperature to as low as -58 °C in a 25 °C ambient temperature environment. To establish one-
23
1
Certain commercial products are identified in this paper to specify the materials used and procedures employed. In
no case does such identification imply endorsement or recommendation by the National Institute of Standards and
Technology, or Purdue University, nor does it indicate that the products are necessarily the best available for the
purpose.
8
dimensional heat flow, heat losses in lateral directions were minimized by using a longitudinal guard
1
having approximately the same temperature gradient. In addition, thermal insulation was placed between
2
the longitudinal guard and the sample. Insulation was also placed in the environment surrounding the
3
longitudinal guard to further prevent heat losses (Fig. 2).
4
Thermally conductive pads (ThermaCool TC30081) with a thickness of 3 mm, thermal conductivity of
5
3.0 W/(m·K), and an operating temperature range of -55 °C to 200 °C were used at the Pyroceram/mortar
6
sample and Pyroceram/cold plate interfaces. Seven thermocouples with an accuracy of ± 0.1 °C were
7
positioned at different heights (Fig. 2c) to measure corresponding temperatures and heat flow through the
8
LGCC. It should be noted that the heat flow measured by this technique is an approximate measurement,
9
since the designed LGCC is operating under quasi-steady state conditions.
10
11
12
Fig. 2Experimental setup: (a) mortar sample and Pyroceram specimens on the two stage cold plate; (b)
13
entire experimental setup with insulation; (c) schematic of mortar sample inserted between Pyroceram
14
material inside LGCC with AE sensor.
15
(a)
(b)
(c)
9
3.2 Acoustic Emission Measurements
1
Acoustic emission (AE) activity was measured during freezing and thawing. The acoustic events refer to
2
sound waves that are produced when a material undergoes cracking, resulting in stress waves due to the
3
energy release in a material. Acoustic waves and corresponding energy could be released by generation
4
of cracks or microcracks under several types of internal or external loading. The detected AE can cover a
5
wide range of inaudible and audible frequencies [31]. Therefore, a threshold value is often assigned for
6
frequencies or amplitudes to filter environmental noise. Piezoelectric sensors are often used to convert
7
the captured acoustic waves into electrical signals. The strength of the signals (the amplitude and
8
duration of the waves) generally depends on the amount of released energy, distance, and orientation of
9
the source with respect to the sensor, and the nature of the transferring medium [31,32]. The signals are
10
then amplified and recorded in a data acquisition system. The area under the absolute value of the
11
surface displacement can be calculated and considered as the AE signal energy’ that is released. This
12
AE signal energy is proportional to the fracture energy [26,33,34].
13
AE can be performed in either a passive or active mode. In the passive mode, AE transducers are
14
attached to samples to capture acoustic waves generated by the formation of cracks or microcracks
15
during the test. These acoustic waves are then analyzed to determine the quantity and behavior of the
16
resulting damage. In the active mode, coupled transducers are attached at a known distance from one
17
another on the surface of the samples. One of the coupled transducers generates known pulses and the
18
other transducers record the sent pulses. The changes in pulse properties and its speed through the
19
sample during the experiment indicate damage occurring in the sample.
20
In this study, both passive and active techniques were used to measure freeze/thaw damage in mortar
21
samples containing NaCl solution. A Vallen AMSY51 acoustic emission system with the capability of wave
22
transient recording (TR) was used. Therefore, a complete waveform diagram of any captured wave could
23
be recorded and then analyzed. VS375-M cylindrical broadband sensors with a diameter of 20.3 mm and
24
a height of 14.3 mm were used in this study. This type of transducer is a high frequency transducer with
25
its peak sensitivity at 375 kHz. It is ideally suited for detecting waves due to crack growth in noisy
26
environments. During testing, the acoustic waves generated due to the crack/microcrack formation were
27
10
captured by the AE sensor and then converted to electrical signals. The electrical signals were processed
1
and magnified by preamplifiers. AE pre-amplifiers are used to transform the high-impedance signal of a
2
sensor to a low-impedance signal suitable for transmission through long cables. A data acquisition
3
system with the capability of streaming TR data up to 10 MHz was used to record the results. A noise
4
threshold of 40 dB was considered for all AE sensors to exclude surrounding environmental noise.
5
Before and after the freeze/thaw test, two coupled AE sensors were used to perform active AE
6
measurement through the length of the sample. During the freeze and thaw test, one AE sensor was
7
used to record a continuous passive AE measurement. All sensors were installed using high vacuum
8
silicone grease (Dow Corning1) containing polydimethylsiloxane, amorphous silica, dimethyl siloxane, and
9
hydroxyl-terminated. This grease has good resistance to water, chemicals, and high and low
10
temperatures; and it was found to be stable over the temperature range of the test. A slight force was
11
applied to the sensor to have a better contact at the interface between the sensors and the sample.
12
13
3.3 Temperature Used for Freeze-Thaw Testing
14
In this study, each mortar sample was exposed to one freeze/thaw cycle in order to determine freezing
15
and thawing temperatures of solution in mortar samples by using the two stage cold plate. Samples
16
saturated with different concentrations of NaCl solution (up to 26 % by mass) were tested. Temperature
17
was controlled to vary from 24 °C to -40 °C. Fig. 3 shows the temperature of the cold plate for this test.
18
The initial temperature of the test was set to remain at 24 °C for one hour to allow the sample to
19
equilibrate. After the initial temperature became stable, the bottom surface was cooled by controlling the
20
temperature of the cold plate. The cold plate was cooled at a rate of 2 °C/h within 32 h. At -40 °C, the
21
temperature was kept constant for 4 h to allow the sample to again reach equilibrium. Then, the
22
temperature was increased to 24 °C at a rate of 4 °C/h within 16 h. It should be noted that other
23
temperatures and conditions can be applied for freezing and thawing tests.
24
25
11
1
Fig. 3Temperature of cold plate used for freeze/thaw testing.
2
3
4.0 Results and Discussion
4
4.1 Temperature and Heat Flow during Cooling and Heating
5
Heat flow is the rate of energy (i.e., heat) passing through a sample due to a temperature gradient. At
6
steady state, the temperature gradients along the sections can be calculated from temperatures
7
measured at the top and bottom of the two Pyroceram meter bars and the sample. Heat flow can also be
8
calculated by knowing the thermal conductivity of Pyroceram, Eq. 1. All calculations are done in
9
accordance with ASTM E1225-09 [15], D5470-12 [28], and C1045-07 [35]. It should be noted that these
10
calculations are for steady state conditions, while the current experimental program was performed in a
11
quasi-steady state condition, since the temperature of the mortar sample was decreased or increased
12
gradually. To reach almost steady state conditions, the heating and cooling rates were assigned to be
13
relatively low. Heat flow per unit area for the top and bottom Pyroceram meter bars can be calculated by
14
Eq. 2 and Eq. 3 as follows:
15
16


(2)


(3)
17
12
where:
1
Heat flow per unit area (W/m2) through top Pyroceram meter bar at (T6+T7)/2 when the sample
2
temperature is at Tave,
3
Heat flow per unit area (W/m2) through bottom Pyroceram meter bar at (T2+T3)/2 when the sample
4
temperature is at Tave,
5
 Thermal Conductivity (W/(m·K)) of Pyroceram at its average temperature calculated by Eq. 1 [29],
6
 Thickness of Pyroceram meter bar (m),
7
Temperature (°C) measured by thermocouple at location i (see Fig. 2), and
8
 .
9
10
After calculating the heat flow per unit area through the top and bottom Pyroceram meter bars, the heat
11
flow through the mortar sample can be calculated by Eq. 4 and Eq. 5:
12
13

(4)
 
(5)
14
where:
15
 Average heat flow per unit area (W/m2) through mortar sample at ,
16
 Average heat flow (W) through mortar sample at Tave, and
17
A = Cross-sectional area of the sample (m2).
18
19
In addition, ΔQsample, the heat flow that is consumed or released by the mortar sample at Tave (i.e., heat
20
flow inward or outward sample) can be estimated by Eq. 6:
21
22
 
(6)
23
24
13
Fig. 4 shows the temperatures of the sample and Pyroceram meter bars at seven different locations (as
1
indicated in Fig. 2c). Heat flows (i.e., QSample and QSample) were also calculated and shown in Fig. 4 and
2
Fig. 5 for one dry sample and for wet samples containing 0 %, 5 %, or 23.3 % by mass NaCl solution. For
3
samples containing solutions with NaCl concentrations less than or equal to 10 %, a single exothermic
4
peak was observed during cooling that corresponds to the freezing point of pore solution (containing
5
NaCl). However, for samples containing solutions with NaCl concentration greater than 10 %, two
6
exothermic peaks were observed during cooling. The first exothermic peak became larger with increasing
7
salt concentration. The second exothermic peak occurred gradually and it was therefore difficult to
8
determine its exact onset time.
9
During the heating process, one endothermic peak was observed for all concentrations of NaCl solution.
10
This peak corresponds to the thawing of ice in contact with the pore solution. No peaks were detected for
11
the dry sample during cooling or heating, indicating that no heat was released or absorbed due to
12
solid/liquid phase changes, as would be expected. The difference between cooling and heating processes
13
for heat flow versus temperature curves (Fig. 5) can be attributed to the difference in cooling and heating
14
rates, and the quasi-steady-state nature of heat flow in the experimental setup.
15
It should be noted that a temperature rise was observed at exothermic peaks. The temperature increase
16
was created by the large transfer of latent heat to the surrounding environment (meter bars and
17
thermocouples). Such a temperature rise at the freezing point is often seen when larger samples are
18
frozen in low temperature calorimetry experiments [8,36]. A similar phenomenon was detected during
19
thawing at which temperature remained constant for a short period of time due to the melting phase
20
change absorbing energy.
21
22
14
1
Fig. 4 Temperatures of sample and Pyroceram meter bars at different locations together with
2
corresponding consumed/released heat flow (QSample) versus time for (a) dry sample; (b) sample
3
saturated with deionized water; (c) sample saturated with 5 % NaCl solution; and, (d) sample saturated
4
with 23.3 % NaCl solution.
5
6
The slope of heat flow versus temperature increases with an increase of the salt concentration (Fig. 5).
7
This means that the heat flow is more sensitive to the temperature change for samples saturated with a
8
greater concentrated NaCl solution.
9
10
15
1
Fig. 5Heat flow through sample (QSample) versus average temperature for (a) dry sample; (b) sample
2
saturated with deionized water; (c) sample saturated with 5 % NaCl solution; and, (d) sample saturated
3
with 23.3 % NaCl solution.
4
5
4.2 Freeze and Thaw Temperatures
6
Water and salt solutions may freeze at a temperature below their freezing point. This phenomenon is
7
known as super-cooling or under-cooling. In general, a liquid below its freezing point crystallizes only in
8
the presence of a seed crystal or nucleus. However, lacking any such seeds or nuclei or presence of any
9
inhibitor forces, the liquid phase can be maintained down to the temperature at which homogeneous
10
crystal nucleation occurs [37]. In the opposite case, a solid will almost always melt at its melting
11
16
temperature (i.e., it cannot be “superheated”). For this reason, it is the melting point which is usually
1
utilized to identify the liquidus line in a phase diagram [38].
2
NaCl-rich solution in the pores of the mortar sample could freeze at a lower temperature than expected
3
from the NaCl solution phase diagram, but should thaw at the same liquidus temperature as indicated on
4
the phase diagram. Indeed, under-cooling action was observed during the cooling process for mortar
5
samples saturated with NaCl solutions. Freezing and thawing points for samples saturated with solution
6
of different dosages of NaCl salt are indicated in Fig. 6; they are also compared to freezing and thawing
7
points for a bulk NaCl solution derived from the NaCl-H2O phase diagram. Since the heat flowed
8
downwards during cooling, the pore solution is likely to have frozen first at the bottom, propagating
9
upward. Therefore, the freezing point at the bottom of the sample may be more representative of the
10
actual freezing point of the sample.
11
While the freezing points at the bottom and the top of the mortar samples were observed to be different,
12
their thawing points were quite similar. For samples containing solutions with a NaCl concentration of
13
more than 13 %, two freezing points were reported in Fig. 6 corresponding to the two exothermic events
14
observed during cooling. For samples containing solution with NaCl concentration less than or equal to
15
8 %, the point of the single observed exothermic event was considered as the first freezing point, while for
16
the sample with a solution of 10 % NaCl concentration, the single observed exothermic event was
17
reported as the second freezing point. This was done for consistency with the general trend for freezing
18
points versus NaCl concentration. For 10 % NaCl concentration, the first exothermic point might not have
19
been detected due to a very small heat release (small freezing event).
20
The second freezing point for high concentrations could be attributed to formation of ice in the pore
21
structure, while the origin of the first freezing point is unknown, as its temperature is above the NaCl
22
solution liquidus line. The possible explanations for this exothermic behavior could be 1) the formation of
23
NaCl·2H2O at concentrations less than the expected eutectic concentration due to confinement, capillary
24
pressure, and surface interaction within the pore structure; 2) the generation of an alternative phase
25
change due to the composition of pore solution and chloride binding; 3) the inhomogeneous distribution of
26
17
NaCl within the specimen microstructure during the saturation process or, 4) the formation of NaCl salt on
1
the pore structure surface.
2
Table 3 shows the freezing and thawing temperatures of the top and bottom surfaces of the samples as
3
well as the amount of under-cooling with respect to the liquidus line. For high concentrations, the second
4
freezing point was used to calculate the amount of under-cooling, as it may be a better representative of
5
ice formation. For samples with solution concentrations between 5 % and 8 % NaCl, an increase in
6
freezing temperature was observed which can be followed by the first freezing points observed for
7
concentrations greater than 10 % NaCl. In addition, for samples saturated with highly concentrated NaCl
8
solutions, less depression (under-cooling) was observed compared to samples saturated with low
9
concentration solutions. This can be attributed to the reduction of pore solution concentration due to the
10
early phase change at the first freezing point. Further research is needed to understand the observed
11
phenomenon at the first freezing point.
12
13
14
Fig. 6 Freezing and thawing points of samples saturated with different NaCl solutions compared to the
15
freezing points of bulk NaCl solution.
16
18
Table 3 Freezing and thawing points at the bottom and top surfaces of samples (°C).
1
NaCl solution
concentration
(%)
Freezing
point of
bulk NaCl
solution
Freezing point
Thawing point
Amount of
under-cooling
w.r.t. liquidus
line+
First peak
Second peak
Bottom
Top
Bottom
Top
Bottom
Top
0 (Dry)
-
-
-
-
-
-
-
-
0 (DI Water)
0.0
-6.1
-3.8
-
-
0.3
0.0
-6.1
0.7
-0.4
-7.0
-5.2
-
-
-0.1
-0.6
-6.7
3
-1.6
-8.3
-6.5
-
-
-1.1
-1.6
-6.7
5
-3.0
-10.8
-7.9
-
-
-2.3
-2.9
-7.8
6
-3.7
-10.2
-6.9
-
-
-3.9
-4.1
-6.6
8
-5.1
-9.3
-6.7
-
-
-4.8
-5.0
-4.2
10
-6.5
-
-
-12.0*
-9.3*
-5.3
-5.8
-5.5
13
-9.1
-6.4
-4.2
-12.9**
-10.6**
-8.1
-8.7
-3.8
15
-10.9
-6.1
-5.3
-15.5**
-13.0**
-10.1
-10.9
-4.5
18
-14.3
-3.0
-1.3
-17.0**
-15.0**
-15.0#
-15.5#
-2.7
23.3
-21.1
0.1
2.1
-26.0**
-23.5**
-20.7
-21.0
-4.9
26
-21.1++
-2.4
-0.8
-26.0**
-23.5**
-20.6
-20.8
-4.9
* For samples containing 10 % NaCl solution, one peak was observed and considered as the second freezing
2
point. It was assumed that the first freezing point was not large enough to detect.
3
** For these concentrations, approximate values are reported since it was difficult to find the exact onset value
4
for heat transfer.
5
+ To calculate the amount of under-cooling, the lowest freezing point is subtracted from the freezing point of the
6
corresponding bulk NaCl solution.
7
++ For this dosage, there are two freezing points according to the aqueous NaCl phase diagram (Fig. 1). Here,
8
the second freezing point is reported as it corresponds to the ice formation.
9
# For this concentration, two thawing points were observed (Fig. 6). The one which is reported in this table
10
corresponds to the liquidus line. The other melting point was observed around -20.6 °C which corresponds to
11
the eutectic temperature.
12
13
4.3 Passive Acoustic Emission
14
Plots of AE event amplitude versus temperature for the dry sample and samples saturated with DI water,
15
5 % NaCl solution, and 15 % NaCl solution are shown in Fig. 7 during cooling and heating. For the dry
16
sample, relatively very few AE events were recorded during cooling and heating (Fig. 7a). These acoustic
17
events are most likely attributed to the coefficient of thermal expansion (CTE) mismatch between
18
aggregate and paste, and they disappear in subsequent freeze and thaw cycles [1]. A greater number of
19
AE events were observed for the solution-saturated samples during freezing. While the saturated
20
samples had AE activities associated with the CTE mismatch, greater levels of AE activity were detected
21
19
due to the damage and cracking caused by freezing and thawing. It should be noted that the amount of
1
damage varies with the concentration of NaCl solution.
2
For samples saturated with NaCl concentrations lower than 3 %, a cluster of AE activity was recorded
3
during freezing due to crack formation generated by ice formation and osmotic pressure. Another cluster
4
was observed during thawing (Fig. 7b) which may be attributed to additional crack formation due to
5
osmotic pressure or crack closing. When the NaCl concentrations increased from 3 % to 8 %, the AE
6
events during thawing diminished (Fig. 7c). For solutions with higher concentrations of NaCl, in addition
7
to AE events due to freezing damage, the AE activity was also observed between the first and the second
8
exothermic freezing temperatures. Moreover, a cluster of AE events was detected around -21.1 °C (the
9
eutectic point) during heating, which corresponds to the phase transformation (melting) of NaCl·2H2O
10
(Fig. 7d).
11
For samples saturated with solutions of (0, 3, 5, 6, 10, 15, and 23.3) % NaCl, cumulative AE energy
12
versus time is indicated in Fig. 8. While a dramatic increase in AE activity was observed for mortar
13
samples containing lower concentrations of NaCl solution, the increase of AE activity for higher
14
concentrations occurred gradually. This can also be explained by the slope of the AE cumulative energy
15
curve versus time at the moments of ice formation. In other words, damage propagation and the
16
generation of cracks or microcracks during cooling were more distributed for higher concentrations, while
17
damage and crack development happened more suddenly for lower concentrations.
18
Fig. 9 shows the total cumulative acoustic energy recorded during one freeze/thaw cycle for samples
19
saturated with different concentrations of NaCl solution. Two levels of high damage or AE activity (i.e.,
20
humps) can be seen. The first hump can be attributed to ice formation and osmotic pressure damage.
21
The decrease of cumulative AE energy with the increase in NaCl solution concentration may be due to a
22
reduction in the volume of ice formation, since the amount of a solvent (water) decreases in a solution
23
(aqueous NaCl) with an increase in dosage of a solute (NaCl). The second hump can be attributed to
24
damage due to the expansion of NaCl·2H2O during melting at -21.1 °C. Further research is needed to
25
investigate the effect of different phase changes on damage mechanisms in mortars during freezing and
26
thawing.
27
20
1
Fig. 7 Passive AE events versus temperature during cooling and heating
2
3
4
Fig. 8 Cumulative acoustic energy versus time for samples saturated with NaCl solution.
5
21
1
Fig. 9 Total cumulative acoustic energy due to freeze/thaw damage for samples saturated with different
2
NaCl solutions (the error bars indicate ± one standard deviation for two replicates).
3
4
4.4 Active Acoustic Emission
5
The velocity of a longitudinal stress wave (V) can be calculated using the measured length of the sample
6
(L) and the transit time (T) that the wave takes to pass through the length of sample (Eq. 7).
7
8
(7)
The velocity can be obtained for the sample before and after a freeze/thaw cycle as the sample
9
undergoes damage. The average transmission times of four single pulses was measured along the
10
sample in each direction. Knowing the distance between the two sensors, an average pulse velocity was
11
calculated by the acoustic emission system in each direction. Pulse velocities were then averaged and
12
used to calculate the relative dynamic modulus.
13
ASTM C597-09 [39] describes how the velocity of the longitudinal stress wave (V) measured in concrete
14
can be related to the elastic properties and density of the concrete using the following relationship:
15
16
22


(8)
1
where E, , and are the dynamic modulus of elasticity, the dynamic Poisson’s ratio, and the density,
2
respectively.
3
The relative dynamic modulus Et/Eo (the ratio of the dynamic modulus after freeze/thaw testing to the
4
dynamic modulus before the freeze/thaw test) was used to assess the damage level of the samples as a
5
result of freezing and thawing. It is assumed that the density and Poisson’s ratio remain constant during
6
testing. It has been shown that assuming a constant Poisson’s ratio during freezing does not significantly
7
influence the value of the dynamic modulus [20]. The damage parameter (D) is estimated using Eq. 9.
8
9
(9)
10
where Eo and Vo are dynamic elastic modulus and average pulse velocity, respectively, before a freeze/
11
thaw test; and Et and Vt are the dynamic elastic modulus and average pulse velocity, respectively, after
12
the freeze/thaw test.
13
The damage index (D) for mortar samples is shown in Fig. 10. Similar to the behavior observed for the
14
cumulative acoustic energy, two levels of high damage (i.e., humps) can be seen. The damage that
15
occurs at the lower concentrations of NaCl solution may be similar to what was observed during scaling
16
tests, where more damage occurred due to a pessimism salt concentration between 3 % and 5 % NaCl
17
[40]. Greater damage also occurred with a higher NaCl concentration solution, which appears to be
18
caused by the combination of ice formation during freezing, the osmotic pressure, the expansion of
19
NaCl·2H2O during melting, and the damage caused between the first and second exothermic freezing
20
points. The temperature at which NaCl·2H2O melts is the eutectic temperature (-21.1 °C). At this
21
temperature, all of the ice that has formed during freezing remains in the pores and cracks in a frozen
22
form. As NaCl·2H2O changes from a solid to a liquid, it expands, causing more damage in the samples.
23
23
This can be also observed in Fig. 7d as AE events were observed for melting of NaCl·2H2O. In addition,
1
another AE cluster observed for high concentration between the first and second freezing exothermic
2
behavior (Fig. 7d) implies an additional damage mechanism happening for high concentrations. Above a
3
concentration of 15 %, the damage parameter decreases, showing that the damage due to a combination
4
of ice formation, osmotic pressure, NaCl·2H2O melting, and the damage caused between the first and
5
second exothermic freezing points passes a critical combination. It should be noted that relatively lesser
6
AE energy was measured for the concentration of 15 % (Fig. 8 and Fig. 9).
7
8
9
Fig. 10Decrease of the relative dynamic elastic modulus (damage parameter) due to freeze and thaw
10
damage for samples saturated with different dosages of NaCl solution (the error bars indicate ± one
11
standard deviation for four replicates of pulse velocity testing).
12
13
5.0 Conclusions
14
In this paper, the freeze/thaw behavior of mortar samples that were saturated by solutions with various
15
concentrations of NaCl salt was investigated. A low-temperature longitudinal guarded comparative
16
calorimeter (LGCC) was developed and equipped with acoustic emission sensors to measure the heat
17
flux and acoustic emission activities (i.e., damage or cracking behavior) of the mortars during freezing
18
and thawing.
19
24
Mortar samples exhibited different exothermic behaviors during cooling (freezing) while showing only one
1
endothermic activity during heating (melting). For lower concentrations of NaCl, one exothermic peak was
2
observed in heat flux due to ice formation. For higher concentrations, however, two exothermic events
3
were seen. The second exothermic peak was due to ice formation, while the cause of the first exothermic
4
peak remains unknown and more investigation is needed to understand and validate possible causes.
5
Heat release due to ice formation occurred at a lower temperature than what would be expected from a
6
NaCl aqueous solution. The frozen solution which was created during cooling in mortar was observed to
7
melt at the same temperature as would be expected for the bulk NaCl solution. A greater freezing point
8
depression was observed for samples saturated with a lower NaCl salt concentration in comparison to
9
samples saturated with high concentration.
10
Two distinctive features (i.e., humps) were observed in the cumulative energy and damage index
11
obtained from acoustic activities. The first hump may be attributed to ice formation and the resulting
12
osmotic pressure that cause cracking and damage. The second distinct feature (second hump) may be
13
attributed to expansion of NaCl·2H2O during melting at its eutectic temperature (-21.1 °C), which can
14
result in further stress and cracking. Additional damage may occur due to destructive phenomenon
15
observed for high concentration between the first and second freezing exothermic behavior. For mortar
16
samples saturated with 5 % and 15 % of NaCl solution, higher freeze and thaw damage was observed.
17
18
6.0 Acknowledgements
19
The experiments reported in this paper were conducted in the Pankow Materials Laboratory at Purdue
20
University. The authors would like to acknowledge the support from the Joint Transportation Research
21
Program administered by the Indiana Department of Transportation and Purdue University. The contents
22
of this paper reflect the views of the authors, who are responsible for the facts and the accuracy of the
23
data presented herein, and do not necessarily reflect the official views or policies of the Indiana
24
Department of Transportation, nor do the contents constitute a standard, specification, or regulation.
25
26
25
7.0 References
1
[1] W. Li, M. Pour-Ghaz, J. Castro, J. Weiss, Water absorption and critical degree of saturation
2
relating to freeze-thaw damage in concrete pavement joints, Journal of Materials in Civil
3
Engineering. 24 (2012) 299307.
4
[2] T.C. Powers, The air requirement of frost-resistant concrete, Highway Research Board. 29 (1949)
5
184211.
6
[3] T. Powers, A working hypothesis for further studies of frost resistance of concrete, in: Journal of
7
the American Concrete Institute, Portland Cement Association, Detroit, Michigan, 1945: pp. 245
8
272.
9
[4] T.C. Powers, The physical structure and engineering properties of concrete, Research Department
10
Bulletin, Portland Cement Association. 90 (1958) 27 pages.
11
[5] J. Beaudoin, C. MacInnis, The mechanism of frost damage in hardened cement paste, Cement
12
and Concrete Research. 4 (1974) 139147.
13
[6] G. Litvan, Frost action in cement in the presence of de-icers, Cement and Concrete Research. 6
14
(1976) 351356.
15
[7] G. Scherer, Crystallization in pores, Cement and Concrete Research. 29 (1999) 13471358.
16
[8] J. Kaufmann, Experimental identification of damage mechanisms in cementitious porous materials
17
on phase transition of pore solution under deicing salt attack, École polytechnique fédérale de
18
Lausanne (EPFL), 2000.
19
[9] P.K. Mehta, P.J. Monteiro, Concrete: microstructure, properties, and materials, Third, McGraw-Hill,
20
New York, 2005.
21
[10] R. Beddoe, M. Setzer, A low-temperature DSC investigation of hardened cement paste subjected
22
to chloride action, Cement and Concrete Research. 18 (1988) 249256.
23
[11] B. Han, J.H. Choi, J. Dantzig, J.C. Bischof, A quantitative analysis on latent heat of an aqueous
24
binary mixture, Cryobiology. 52 (2006) 14651.
25
[12] R. Beddoe, M. Setzer, Phase transformations of water in hardened cement paste, a low-
26
temperature DSC investigation, Cement and Concrete Research. 20 (1990) 236242.
27
[13] B. Johannesson, Dimensional and ice content changes of hardened concrete at different freezing
28
and thawing temperatures, Cement and Concrete Composites. 32 (2010) 7383.
29
[14] S. Jacobsen, E. Sellevold, S. Matala, Frost durability of high strength concrete: Effect of internal
30
cracking on ice formation, Cement and Concrete Research. 26 (1996) 919931.
31
[15] ASTM E1225- Standard test method for thermal conductivity of solids by means of the guarded-
32
comparative-longitudinal heat flow technique, ASTM International, West Conshohocken, PA, 2009.
33
[16] D.-J. Yoon, W.J. Weiss, S.P. Shah, Detecting the extent of corrosion with acoustic emission,
34
Transportation Research Record. (2000) 5460.
35
26
[17] D.-J. Yoon, W.J. Weiss, S.P. Shah, Assessing damage in corroded reinforced concrete using
1
acoustic emission, Journal of Engineering Mechanics. 126 (2000) 273283.
2
[18] M. Pour-Ghaz, J. Weiss, Quantifying damage due to aggregate expansion in cement matrix,
3
Advances in the Material Science of Concrete. 270 (2010) 101114.
4
[19] S. Puri, J. Weiss, Assessment of localized damage in concrete under compression using acoustic
5
emission, Journal of Materials in Civil Engineering. 18 (2006) 325333.
6
[20] Z. Yang, W.J. Weiss, J. Olek, Water transport in concrete damaged by tensile loading and freeze
7
thaw cycling, Journal of Materials in Civil Engineering. 18 (2006) 424434.
8
[21] W.S. Park, J.H. Lee, Evaluation of freeze-thaw damage in concrete by one-sided stress wave
9
velocity measurement technique, Key Engineering Materials. 270-273 (2004) 16041609.
10
[22] M. Ohtsu, Nondestructive evaluation of damaged concrete due to freezing and thawing by elastic-
11
wave method, Journal of Advanced Concrete Technology. 3 (2005) 333341.
12
[23] W. Li, W. Sun, J. Jiang, Damage of concrete experiencing flexural fatigue load and closed
13
freeze/thaw cycles simultaneously, Construction and Building Materials. 25 (2011) 26042610.
14
[24] M.A. Taylor, Passive acoustic emission for quantitative evaluation of freeze thaw and alkali
15
aggregate reaction in concretes, in: F. Ansari, S. Sture (Eds.), Nondestructive Testing of Concrete
16
Elements and Structures, New York, NY, 1992: pp. 112.
17
[25] Y. Qian, Y. Farnam, J. Weiss, Using acoustic emission to quantify freeze-thaw damage of mortar
18
saturated with NaCl solutions, in: T. Miyagawa (Ed.), International Conference on Sustainable
19
Construction Materials & Technologies, Kyoto, Japan, 2013: pp. 110.
20
[26] H. Shimada, K. Sakai, G.G. Litvan, Acoustic emissions of mortar subjected to freezing and
21
thawing, in: Second International Conference on Durability of Concrete, Montreal, Canada, 1991:
22
pp. 263278.
23
[27] ASTM C305 - Standard practice for mechanical mixing of hydraulic cement pastes and mortars of
24
plastic consistency, ASTM International, West Conshohocken, PA, 2012.
25
[28] ASTM D5470- Standard test method for thermal transmission properties of thermally conductive
26
electrical insulation materials, ASTM International, West Conshohocken, PA, 2012.
27
[29] Thermal properties of pyroceram code 9606, Private Communication from the National Institute of
28
Standards and Technology (NIST), Gaithersburg, MD.
29
[30] www.teca-eu.com, Low temperature cold plate - CPV Cascade CCP-22, (2012) 12.
30
[31] M. Pour-ghaz, R. Spragg, J. Castro, J. Weiss, Can acoustic emission be used to detect alkali silica
31
reaction earlier than length change?, in: 14th International Conference on Alkali-Aggregate
32
Reaction, University of Texas; Austin, Texas, 2012: pp. 110.
33
[32] H. Vallen, AE testing fundamentals, equipment, applications, Nondestructive Testing. 7 (2002).
34
[33] E.N. Landis, L. Baillon, Experiments to relate acoustic emission energy to fracture energy of
35
concrete, Journal of Engineering Mechanics. 128 (2002) 698702.
36
27
[34] M. Pour-Ghaz, Detecting damage in concrete using electrical methods and assessing moisture
1
movement in cracked concrete, Purdue University, 2011.
2
[35] ASTM C1045- Standard practice for calculating thermal transmission properties under steady-
3
state conditions, ASTM International, West Conshohocken, PA, 2007.
4
[36] S. Jacobsen, D. Sœther, E. Sellevold, Frost testing of high strength concrete: Frost/salt scaling at
5
different cooling rates, Materials and Structures. 30 (1997) 3342.
6
[37] P.G. Debenedetti, H.E. Stanley, Supercooled and glassy water, Physics Today. 56 (2003) 4046.
7
[38] J.A. Ramsay, A new method of freezing-point determination for small quantities, The Journal of
8
Experimental Biology. 26 (1949) 5764.
9
[39] ASTM C597- Standard test method for pulse velocity through concrete, ASTM International, West
10
Conshohocken, PA, 2009.
11
[40] John Valenza, Mechanism for salt scaling, Princeton University, 2005.
12
13
... In ceramic materials, elastic waves are generated mainly during the formation of microcracks [13,14]. AE was previously employed to study the freeze-thaw behavior of concrete and mortars [15][16][17][18][19]. In [15], it was found that in mortar saturated with NaCl solution, most of the AE events occurred slightly after the temperature of the sample reached the liquidus line on cooling. ...
... In the same study, the impact echo method revealed a shift in the natural frequency of vibration after six freeze-thaw cycles as a consequence of microstructural changes. In [18,19], AE was used to study the microcracking in concrete saturated with NaCl solution in various concentrations. Cyclic tests (up to 3 cycles) revealed a decreasing amplitude of AE events with an increasing number of cycles. ...
... The observation of the AE activity during heating/ thawing of water in the pores is rather counter-intuitive. Such an effect was, however, observed also elsewhere [17][18][19]. The models which are standardly used to explain frost damage (crystallization pressure, hydraulic pressure, osmotic pressure) cannot explain microcracking on heating/thawing. ...
Article
Full-text available
The acoustic emission (AE) technique was employed to monitor crack formation in the water-saturated ceramic material subjected to freeze–thaw cycles. The samples with water accessible porosity ranging from 6 to 50 vol% were prepared from the illite-rich clay by the sacrificing template method. The experimental samples were fired to 1100 °C in the static air atmosphere. Afterward, they were saturated with distilled water and subjected to the freeze–thaw cycles in the temperature range from − 22 to 20 °C. The AE signals occurred repeatedly during freeze–thaw cycling and they were attributed to the formation of microcracks in the ceramic body. The intensity of microcracking increased with increasing porosity. The microcracks were created primarily during the freezing stages of the experiment. Nevertheless, a considerable number of microcracks also appeared during thawing, especially in the ceramic body with pores below 1 μm. The microcracking during thawing was attributed to the reabsorption of water by tiny pores.
... Chloride-rich environments such as de-icing chemicals and seawater may cause severe physical damage to concrete structures due to interactions between alkaline cement matrix and related ions and resultant expansive forces attending the formation of various products [1][2][3][4][5][6]. Farnam et al. [3] investigated the phase changes and damage development of the concrete exposed to CaCl 2 and confirmed the formation of calcium oxychloride at temperatures above freezing. ...
Article
Full-text available
Chloride-rich environments including seawater and de-icing salt solutions may lead to severe physical damage to cementitious materials, especially considering the special service environment such as semi-immersion exposure for some concrete structures. The physical degradation behaviour of cement mortars at three different relative humidity (RH) levels was investigated based on changes in physical appearances, dynamic elastic modulus and microstructural analysis. Experimental results showed that RH at 65 ± 5% resulted in obvious salt precipitations on the top region of the tested cement mortar specimens when semi-immersed in NaCl and seawater solutions. Besides, this RH level led to cracks and fractures of specimens with exposure to MgCl2 and CaCl2 solutions, which are closely related to the formation of expansive products, expressed as 0.4MgCO3·5.4 Mg(OH)2·MgCl2·6H2O and CaCl2·CaCO3·nH2O respectively, as confirmed by the microstructural analysis. Interfacial transition zones within binding matrices and sharp edges of specimens are more likely to suffer from salt precipitations. Corresponding different chloride-induced degradation processes owing to various cations or co-existence of various ions in seawater were also proposed and discussed.
... This phenomenon cannot be revealed by the theories mentioned above. Salt-freeze damage can be attributed to the phase change, including both ice formation [146] and salt crystallization [147]. The freeze-thaw damage of concretes may be related to the crystallization of Friedel's salt. ...
Article
In cold regions, concrete structures such as pavements, bridges, and tunnels can undergo freeze-thaw conditions, which may significantly deteriorate the performance of concrete and further pose a threat to the structures’ safety and shorten their service life. This paper displays a comprehensive review of the effects of freeze-thaw cycles (FTCs) on the thermophysical and mechanical properties of concrete, which include the mass loss, specific heat capacity, thermal conductivity, coefficient of thermal expansion, compressive strength, splitting tensile strength, flexural strength, elastic modulus, and stress-strain relationship. Meanwhile, the influences of water/binder ratio, air content, number of FTCs, saturation degree, and multiple factors are analyzed and discussed in detail. In addition, changes in the microstructure and constituents, available theories, test methods, and damage identification/evaluation methods to characterize the freeze-thaw damage are summarized. Accordingly, recommendations are proposed for future investigations to be carried out. The discussion indicates that multiple factors and multiple fields should be comprehensively considered to reveal the micromechanism of the freezing and thawing damage on concrete in real complex environments. Also, existing freeze-thaw damage theories and test methods are supposed to be considered to develop the multiscale models and corresponding techniques.
... e damage mechanism of saltfreezing is different from freezing and thawing damage. Deicing salt, employed to reduce the freezing point of water, alters the degree of saturation [4,5] and reacts with the hydrated ordinary Portland cement (OPC), creating expansive reaction products that lead to cracking and distress [6][7][8][9][10][11]. e deterioration due to the erosive effect of deicing salt can develop rather rapidly than the destructive effect of normal freezing and thawing cycle on concrete. ...
Article
Full-text available
The damage development trend of concrete with cracks in salt-freezing environment is systematically studied. The cracks are also tested in intact concrete for comparison, and crack characterization is introduced. The mass loss, the relative dynamic elastic modulus, and the change of crack width are analyzed. Results show that the crack width increases as the salt-freezing cycle progresses. Following the development trend of the cracks, concrete cracks can be divided into three categories: 0–40, 40–100, and 100–150 μm. The mass loss increases significantly, and the change of relative dynamic elastic modulus decreases in concrete with an initial crack compared with the intact concrete. When the crack width is 80 μm, a maximum mass loss rate of 0.19% and a minimum relative dynamic elastic modulus of 75.81% can be obtained. These test results prove that crack and freeze-thaw coupling can influence each other and accelerate the failure of concrete. Overall, this study can serve as a basis for the durability design and life improvement of concrete structures.
... A one-dimensional and well-distributed heat flow through the specimen was maintained by using aluminum plate (longitudinal guarded) and meter bars (Pyoceram ceramic), respectively. Temperatures at different locations were monitored using thermocouples for the heat flow calculations [48]. The applied thermal cycle was varied from 25 • C to − 10 • C within 24 h as follows: (1) temperature equilibrium at 25 • C for 1 h, (2) cooling to − 10 • C at a rate of 3 • C/h, (3) isotherm at − 10 • C for 3 h, and (4) heating to 25 • C at a rate of 3 • C/h. ...
Article
Interactions between sodium chloride (NaCl) and cementitious materials have shown to facilitate premature deteriorations of concrete in cold environments. NaCl chemically interacts with tri‑calcium aluminate (C3A) and monosulfate (AFm) resulting in a chemical phase change during thermal cycling. The chemical phase change can create damage in porous cementitious materials. Experimental and thermodynamic modeling were conducted in this study to understand the thermo-chemo-physical interactions between NaCl and cementitious materials and identify the source of the chemical phase change. Results indicates that the destructive chemical phase change is the formation of mirabilite (Na2SO4.10H2O). It was found that the phase change temperature of mirabilite is strongly influenced by the concentration of NaCl in the solution. At high concentration NaCl solutions (>10% by mass), mirabilite is formed due to the release of a small amount of sulfate ions from the concrete matrix.
... Freeze-thaw (F-T) cycles may adversely, as well as very quickly affect the durability of concrete structures. From a research point of view, it is appropriate to monitor the behaviour of concrete already during the actual exposure to freeze-thaw cycles [3][4][5]. ...
Article
Full-text available
This manuscript deals with a complex analysis of acoustic emission signals that were recorded during freeze-thaw cycles in test specimens produced from air-entrained concrete. An assessment of the resistance of concrete to the effects of freezing and thawing was conducted on the basis of a signal analysis. Since the experiment simulated testing of concrete in a structure, a concrete block with the height of 2.4 m and width of 1.8 m was produced to represent a real structure. When the age of the concrete was two months, samples were obtained from the block by core drilling and were subsequently used to produce test specimens. Testing of freeze-thaw resistance of concrete employed both destructive and non-destructive methods including the measurement of acoustic emission, which took place directly during the freeze-thaw cycles. The recorded acoustic emission signals were then meticulously analysed. The aim of the conducted experiments was to verify whether measurement using the acoustic emission method during Freeze-thaw (F-T) cycles are more sensitive to the degree of damage of concrete than the more commonly employed construction testing methods. The results clearly demonstrate that the acoustic emission method can reveal changes (e.g., minor cracks) in the internal structure of concrete, unlike other commonly used methods. The analysis of the acoustic emission signals using a fast Fourier transform revealed a significant shift of the dominant frequency towards lower values when the concrete was subjected to freeze-thaw cycling.
Article
The sodium chloride (NaCl) can chemically interact with the monosulfate (AFm) within the concrete matrix, leading to a detrimental sulfate-based phase formation during thermal cycling. This paper studies the potential of reducing the sulfate-based phase in the cementitious pastes made using different types of supplementary cementitious materials (SCMs) used as a partial replacement of ordinary portland cement (OPC). A low-temperature differential scanning calorimetry was used to quantify the amount of the sulfate-based phase. The results indicated that the addition of SCMs as a partial replacement of OPC revealed different effects on forming the sulfate-based phase through pozzolanic activities and/or dilution effects. The addition of slag or class F fly ash can lower the formation of the sulfate-based phase through the dilution effect, while the addition of silica fume can substantially lower the formation of the sulfate-based phase by both dilution and pozzolanic activities. The use of class C fly ash, however, showed negative effects through increasing the formation of the sulfate-based phase. Thermodynamic modeling was used to predict the phase equilibrium of hydrated cementitious systems. Results from thermodynamic predictions were in agreement with the experimental results, indicating that the formation of the sulfate-based phase was associated with the amount of AFm present in the cementitious systems.
Article
Interactions of sodium chloride (NaCl) with cement hydrates can cause chemical and physical changes in cement pastes, leading to premature deterioration, particularly in cold climates. The present study explores the conditions responsible for the damage and investigates potential chemical changes in cement paste exposed to different concentrations of NaCl solutions (0 to 4 M) at temperatures of 25 °C and 5 °C. Thermogravimetric analysis (TGA) was used to investigate the effect of salt concentrations on the leaching of calcium hydroxide (CH) and Friedel’s salts (FS) formation. X-ray diffraction (XRD) was used to study phase transformations and final phase products associated with the chemical changes. Thermodynamic modeling was performed to predict changes in the hydrated phase assemblage and corresponding volumetric changes due to NaCl ingress. A noticeable reduction in the compressive strength was observed with the increase in NaCl concentrations at 5 °C compared to 25 °C. It was found that temperature plays a vital role in reducing the compressive strength in the presence of NaCl. Leaching of CH was not observed, indicating the stability of the CH phase in the presence of NaCl solution. The FS formation was found to increase as salt concentrations increase up to a point (2 M NaCl) where it started to plateau at higher salt concentrations. Observations from XRD revealed that the secondary ettringite formation is the final phase product of chemical changes that occurred in cement paste exposed to high NaCl concentration and lower temperature. Since the TGA confirmed the stability of FS and CH phases at high NaCl concentrations, the reduction in the compressive strength appeared to be mainly due to secondary ettringite formation. Results from thermodynamic calculations were in excellent agreement with the experimental results, indicating that the damage may be attributed to expansions associated with the secondary ettringite formation.
Article
Full-text available
Calcium oxychloride (CAOXY) formation is a deterioration mechanism known to cause joint damage in concrete pavements. The accepted CAOXY mitigation threshold in paste is 15 g/100 g paste; however, this limit was developed using flexural strength testing of pastes. This investigation seeks to verify this threshold using compressive strength and mass change testing. Cementitious pastes were cast with portland cement and fly ash (up to 50% mass). Specimens were stored in a 30% mass CaCl2 at 5 °C to accelerate deterioration. Visual observations, thermogravimetric analysis, low-temperature differential scanning calorimetry, mass change, and compressive strength results were recorded. Partial replacement of cement with fly ash reduces the Ca(OH)2, CAOXY, and damage levels. Using mass change, a threshold at which no damage occurs is established at 20 g/100 g paste, indicating that the current threshold is conservative. For compressive strength, while increasing the amount of fly ash reduces damage, damage is noted in all specimens tested; however, strength loss noted in this work is similar to the research which established the current threshold. Therefore, the compressive strength results generally validate the CAOXY threshold level of 15 g/100 g paste. Careful correlations between paste and concrete damage are needed in order to further verify the threshold value.
Article
Externally bonded fiber-reinforced polymer (FRP) is a promising tool to use for either preserving the integrity of new concrete infrastructure or mitigating the deteriorations of aged concrete infrastructure. This laboratory study evaluates the durability performance of carbon FRP-wrapped concrete with a 1.5 wt% montmorillonite nanoclay (NC) modified siloxane epoxy adhesive, under a simulated cold and chloride-laden environment. Five types of concrete samples were fabricated and tested, i.e., reference concrete, FRP-wrapped concrete, NC-modified FRP-wrapped concrete, FRP/pre-aged concrete, and NC-modified FRP/pre-aged concrete, respectively. Both before and after the salt scaling test in 3 wt% NaCl, the NC-modified FRP-wrapped concrete samples exhibited notably better mechanical properties (higher compression strength and elastic modulus) and significantly better transport properties at the interface (reduced water absorption and gas permeability), related to their unmodified FRP-wrapped counterparts. These benefits of nano-modification, along with better resistance to salt scaling, are attributable to the improved microstructure and reduced hydrophilicity of the adhesive, as revealed by microscopic investigations using a high-resolution optical microscope and a water contact angle tester. The Fourier transform infrared spectroscopy, differential scanning calorimetry, and thermogravimetric analysis revealed that the admixed NC chemically reacted with the adhesive and increased its crosslinking density and thermal stability.
Conference Paper
Full-text available
Mortar samples were saturated with NaCl solutions of various concentrations and subjected to freeze-thaw cycles. Passive and active acoustic emission (AE) testing was conducted. The freezing temperature of the NaCl solutions in mortar corresponded with the sudden observation of passive AE events. The acoustic energy and damage parameter were calculated to evaluate the extent of freeze-thaw damage. The influence of the NaCl solution concentration and whether the solution freezes on freeze-thaw damage are discussed. © 2014 4th International Conference on the Durability of Concrete Structures.
Article
Concrete is a composite of aggregates in a cement paste matrix. Dissimilar volume changes in these constituent materials may result in localized stress development. This is particularly problematic when the aggregate expands more than the surrounding paste. This expansion results in tensile stress development in the cement paste matrix which can lead to micro-cracking in the cement paste matrix. These micro-cracks can eventually coalesce and localize in visible cracking. Quantifying this type of damage can be difficult. This paper describes a conceptual model and physical simulation of this damage considering the expansion of polymeric inclusions (i.e., aggregates) in cement paste matrix subjected to temperature changes. Thermal loading (i.e., temperature change) was selected since it provides a method to control the expansion. Physical experiments were performed where continuous length change measurement and acoustic emission measurements were carried out. These experimental methods are used to better understand the mechanics of the damage. The experimental results indicate that a deviation from classical composite behavior occurs when damage develops which can be seen in the length change measurements. This deviation can be used to quantify the extent of damage. A numerical model is used to interpret the experimental results. An Eshelby misfit approach was used to determine the pressure created by the expanding aggregate. This enables the stresses that develop in a composite material to be determined. A linear fracture mechanics failure criterion is used to calculate the onset of damage formation. Results are in agreement with length change measurements and acoustic emission measurements. A composite damage model for direct calculation of the extent of damage from length change measurements is proposed.
Article
A study of supercooled and glassy water was carried out. Cold, noncrystalline states play an important role in understanding the physics of liquid water. Supercooled water is also important for life at subfreezing conditions, for the commercial preservation of proteins and cells, and for the prevention of hydrate formation in natural gas pipelines.
Article
Fluid ingress is a primary factor that influences freeze-thaw damage in concrete. This paper discusses the influence of fluid ingress on freeze-thaw damage development. Specifically, this paper examines the influence of entrained air content on the rate of water absorption, the degree of saturation, and the relationship between the saturation level and freeze-thaw damage. The results indicate that whereas air content delays the time it takes for concrete to reach a critical degree of saturation it will not prevent the freeze-thaw damage from occurring. The results of the experiments show that when the degree of saturation exceeds 86-88%, freeze-thaw damage is inevitable with or without entrained air even with very few freeze-thaw cycles. DOI: 10.1061/(ASCE)MT.1943-5533.0000383. (C) 2012 American Society of Civil Engineers.
Article
Over the past 60 years, concrete infrastructure in cold climates has deteriorated by “salt scaling,” which is superficial damage that occurs during freezing in the presence of saline water. It reduces mechanical integrity and necessitates expensive repair or replacement. The phenomenon can be demonstrated by pooling a solution on a block of concrete and subjecting it to freeze/thaw cycles. The most remarkable feature of salt scaling is that the damage is absent if the pool contains pure water, it becomes serious at concentrations of a few weight percent, and then stops at concentrations above about 6 wt%. In spite of a wealth of research, the mechanism responsible for this damage has only recently been identified. In this article, we show that salt scaling is a consequence of the fracture behavior of ice. The stress arises from thermal expansion mismatch between ice and concrete, which puts the ice in tension as the temperature drops. Considering the mechanical and viscoelastic properties of ice, it is shown that this mismatch will not cause pure ice to crack, but moderately concentrated solutions are expected to crack. Cracks in the brine ice penetrate into the substrate, resulting in superficial damage. At high concentrations, the ice does not form a rigid enough structure to result in significant stress, so no damage occurs. The morphology of cracking is predicted by fracture mechanics.
Book
Fully revised to address the latest advances in concrete technology, the Fourth Edition of Concrete: Microstructure, Properties, and Materials provides in-depth, scientific details on concrete -- the most widely used structural material. This authoritative resource discusses the microstructure and properties of hardened concrete; concrete-making materials; concrete processing; and current developments in concrete technology, mechanics, and non-destructive methods.