Journal of Advanced Concrete Technology Vol. 3, No.

3, 343-353, October 2005 / Copyright © 2005 Japan Concrete Institute 343

Scientific paper

Nondestructive Evaluation of Setting and Hardening of Cement Paste

Based on Ultrasonic Propagation Characteristics
Toshiro Kamada1, Shinya Uchida2 and Keitetsu Rokugo3

Received 30 May 2005, accepted 19 September 2005

Abstract
This paper describes the relationship between the setting/hardening properties of cement paste and ultrasonic propagation
characteristics from both a macroscopic and a microscopic point of view. The first experimental series was aimed at
evaluating changes in the physical properties of high-early-strength cement paste. In the experiment, a single cylinder
rotational viscometer was used for viscosity measurement. The second series of experiments was aimed at evaluating
changes in the chemical properties of ultra-rapid-hardening cement paste. In the experiment, scanning electron micros-
copy (SEM) and powder X-ray diffraction analysis were done to investigate the generation of hydration products.
As a result, it was confirmed that the maximum amplitude of the obtained waveform adequately reflects the changes in
shear resistance of cement paste. On the other hand, change in ultrasonic wave velocity is well correlated with the for-
mation of ettringite crystals as observed by SEM and powder X-ray diffraction analysis.

1. Introduction basis of elastic wave propagation behavior in suspen-

Test methods to evaluate the change of the properties of
cement-based materials with time after cement-water
mixing include the slump test, flow test, funnel
flow-through test, Vebe test, rotational viscometer test,
Vicat needle test, Proctor penetration resistance test, and
strength test. These methods have different ranges of
applicable material properties and are thus applicable to
different time ranges in the setting and hardening proc-
esses of cement, as shown in Fig. 1. Because of this
situation, different test methods are being used according
to the changes in material properties with time. In the
Vicat needle and Proctor penetration resistance tests, the
setting and hardening states of the materials in early ages
are investigated by means of the change in the material's
penetration resistance. The procedures of these conven-
tional methods are simple and easy to perform. However,
it is difficult to evaluate changes in the physical/chemical
properties of the materials based only on penetration
resistances. Therefore, the development of methods to
measure changes in the physical/chemical properties of
cement-based materials in early ages has been required.
Against this background, Whitehurst (1951) attempted
nondestructive evaluation using an ultrasonic technique
to investigate the changes in material properties with
time. Later, some studies evaluated the setting and
hardening properties of cement-based material on the
1
Associate Professor, Department of Civil Engineering,
Gifu University, Japan
E-mail: kamada@gifu-u.ac.jp
2
Doctoral Student, Department of Civil Engineering,
Gifu University, Japan.
3
Professor, Department of Civil Engineering, Gifu
University, Japan.
sions found by Biot (1956). Akashi and Yamaji (1958),
for example, found an increase in wave velocity with
time but no prominent inflection point in velocity gra-
dient. Kakuta and Akashi (1983) measured longitudinal
and transverse ultrasonic waves separately and calcu-
lated volume modulus and shearing modulus from the
velocities of these waves to investigate changes in ma-
terial properties associated with setting and hardening.
However, the instruments used in the above studies made
it difficult to record and analyze the obtained waveform
as digital data. Thus, ultrasonic measurement could not
be monitored at that time.
Advances in measurement technologies in later years
have allowed the obtained waveform to be recorded and
analyzed continuously, thus making it reasonably possi-
ble to monitor setting and hardening processes continu-
ously. Reinhardt et al. (1996a, 1996b, 2000, 2004)
evaluated changes in material properties associated with
concrete aging by using the velocity of elastic waves and
the energy and frequency distribution of obtained waves.
However, these studies did not address how these values
were related to changes in physical properties in the
setting and hardening processes. Grosse et al. (2000)
compared ultrasonic velocities with scanning electron
microscopy and nuclear magnetic resonance spectros-
copy as methods to evaluate changes in chemical prop-
erties caused by cement hydration. Their experiment
showed a good coincidence between an inflection point
in the time-dependence curve of relaxation time obtained
by nuclear magnetic resonance spectroscopy and an
inflection point in ultrasonic wave velocity gradient.
However, the relationship between generation of hydra-
tion products and ultrasonic wave velocity was not dis-
cussed in detail.
In addition to the above-mentioned studies, there are
344 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005
0
Suspension Semisolid
Semisolid
Slump test Vicat needle test
Flow test Proctor penetration resistance test
Vebe test
Rotational viscometer test
Fig. 1 Test methods to evaluate the change of the properties of cement-based materials.
also several studies that measured and utilized the re-
flection coefficient of shear waves for the evaluation of
cementitious materials in early ages (Valic 2000, Morin
2001, Smith 2002, Voigt 2004). Voigt et al. (2005)
compared longitudinal ultrasonic wave velocities with
in-situ temperature rises or adiabatic heat releases. As a
result, it was reported that temperature measurements
were useful to evaluate the setting and hardening prop-
erties along with ultrasonic measurements. The authors
(2003) had also pointed similar findings in previous
studies.
On the basis of currently available research results, in
order to investigate the relation between the setting and
hardening properties of cement paste and ultrasonic
propagation characteristics, two experimental series
were performed in this study. Experimental series I was
aimed at evaluating changes in the physical properties of
cement paste during its setting and hardening from a
macroscopic point of view. In the experiment, a single
cylinder rotational viscometer was used for viscosity
measurement, and the relationship between changes in
the physical properties of the cement paste and its ul-
trasonic propagation characteristics were discussed. The
experimental results were compared with the results of a
Proctor penetration resistance test. Experimental series II,
on the other hand, was aimed at evaluating changes in the
chemical properties of cement paste during its setting and
hardening from a microscopic point of view. The rela-
tionship between changes in chemical properties, in
particular generation of hydration products, and ultra-
sonic wave velocity was discussed. In the experiment,
scanning electron microscopy and powder X-ray dif-
fraction analysis were conducted in parallel with ultra-
sonic measurement to investigate the generation of hy-
dration products.
2. Experimental Program
2.1 Experimental series I
(1) Materials and mixture proportions
High-early-strength cement was used in experimental
series I. An oxycarboxylate-based liquid retarder was
used to confirm the effect of the retarder on ultrasonic
Solid
Solid
Elapsed time
Strength test
propagation. The percentages of the retarder of cement
weight were 0%, 0.2%, and 0.4%. The water-cement
ratio of the cement paste was 0.45 regardless of the dos-
age of the retarder.
Immediately following the addition of water to the
cement, the cement paste was mixed in a mixer for 3
minutes. The retarder was added to the water before the
mixing procedure.
The mixed cement paste was poured into forms for the
ultrasonic measurements, the Proctor penetration resis-
tance test and the beaker for the viscosity measurements.
These three tests were carried out always simultaneously.
(2) Ultrasonic measurement
The ultrasonic measurement setup is outlined in Fig. 2. A
steel form was used, and a hole was made at the center of
both sides of the form to allow propagation of the ultra-
sonic waves. Ultrasonic sensors (US sensors) were se-
cured on brass plates (50-mm long, 50-mm high, and
0.3-mm thick) with sensor holders to maintain contact
pressure with the specimen. The distance between the
sensors was 35 mm. The sensor distance was chosen so
that it would be larger than the wavelength of the ultra-
sonic waves.
The ultrasonic pulse output from a multifunction
synthesizer was amplified by using a bipolar power
supply and then emitted with a voltage of 240 V and a
pulse interval of 50 seconds. The ultrasonic wave re-
ceived by each US sensor after propagation through the
specimen was amplified by a preamplifier and then a
main amplifier by 40 dB each. Signals above a threshold
were converted to digital data at a sampling frequency of
2 MHz, and the obtained waveform was recorded by an
Ultrasonic measurement system (US measurement sys-
tem). The US sensors used for pulse transmission and
reception have a wide sensitive range of 0 to approxi-
mately 500 kHz and a resonant frequency of 140 kHz.
The ultrasonic measurement was carried out continu-
ously over a duration of 72 hours after mixing and at a
room temperature of 20qC.
The ultrasonic propagation characteristics were
evaluated in terms of wave velocity, maximum amplitude,
and frequency distribution. Wave velocity was calculated
 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005 345

Multifunction synthe-
sizer

US measurement sys- Bipolar amplifier

tem

Preamplifier
d: 35mm (Experimental series I)
d : 27mm (Experimental series II)

Plate
50
100
50
Sensor holder

US sensor Test material Form

(Side view) (Unit: mm)

Fig. 2 Test setup.

by dividing the sensor distance by the propagation time, 400 Pa˜s for the rotational viscometer, the spindle, and
which was given by the difference between the time of the rotational speed used in the test.
pulse input and the time when the signal detected by the
reception-side US sensor exceeded the threshold. The (4) Proctor penetration resistance test
maximum amplitude was obtained as the largest voltage A specimen for the Proctor penetration resistance test
amplitude in the obtained waveform, and the frequency was poured into a steel form with a length of 400 mm, a
distribution was calculated by fast Fourier transform height of 100 mm, and a width of 100 mm. Four types of
(FFT) of the obtained waveform. penetration needles having sectional areas of 100, 50, 25,
and 12.5 mm2 were used. The test started by using a
(3) Viscosity test
The single cylinder rotational viscometer shown in Fig. 3
was used to measure the change in specimen viscosity
with time. For the viscosity measurements, the fresh Single cylinder rotational viscometer
cement paste was placed into a 500-ml beaker having a
diameter of 90 mm and a height of 120 mm. The viscos-
ity measurements were started one hour after mixing of
the cement paste and then repeated at intervals of 10 to
30 min according to the state of the paste. Because ce-
ment paste is a non-Newtonian fluid, the results obtained
by the rotational viscometer indicate apparent viscosity
(ISO 2555).
Using ISO 2555 (1989) as a reference, apparent vis-
cosity at each time was measured with the following Beaker
procedure. A spindle with a diameter of 3.18 mm was
vertically immersed in the fresh cement paste and was
then rotated at 10 rpm for 1 min. Then, apparent viscosity Spindle Test material
was measured two times consecutively to confirm that
the difference between the two values was within 3%.
The average of the two measurements was adopted as a
measured value of apparent viscosity.
The upper limit of measurable apparent viscosity is Fig. 3 Determination of viscosity.
346 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005

100-mm2 needle, and the needle was sequentially re- Stage1

Stage Stage
Stage2 Stage
Stage3
placed with one having a smaller sectional area as the
cement paste hardened. In each test run, a needle appro-

(m/s) (m/s)
priate for the specimen's hardened state was first 3000
mounted on the penetration resistance test equipment.

Velocityvelocity
(m/s)
Then, the needle was inserted into the specimen as much
2000
as 25 mm, and the elapsed time (in hours) and the re-

Velocity
quired force (in N) were recorded. Penetration resistance

Propagation
(in N/mm2) was derived by dividing the force (in N) by 1000
the cross sectional area of the needle (mm2). According Velocity
to the ASTM C403-85 (1985), the penetration resistances
corresponding to the values of 500 psi (3.5 MPa) and 0 10 20 30 40 50 60 70
4000 psi (27.6 MPa) are defined as initial and final set- Elapsed time (hr)
tings, respectively. (a) Dosage
(a) Dosageof of
retarder 0%
retarder

2.2 Experimental series II Stage

Stage1 Stage
Stage2 Stage
Stage3
To confirm the applicability of the present method to
cement that shows rapid hardening, ultra-rapid-hardening

(m/s) (m/s)
cement was used in experimental series II. The mix 3000
proportions of the cement paste consisted in a wa-
ter-cement ratio of 0.30 and amounts of retarder and Velocityvelocity
(m/s)
2000
superplasticizer of 1.0% and 1.5% of the cement weight,
Velocity
Propagation

respectively. An US measurement system was used for

ultrasonic measurement, similarly to experimental series 1000
Velocity
I. The ultrasonic measurements were started when water
was added to the cement, and automatic measurement
0 10 20 30 40 50 60 70
was continued until the cement paste was hardened and
Elapsed time (hr)
the change in wave velocity became very gradual. In the
(b)
(b) Dosage
Dosageof retarder 0.2 %
of retarder
case of automatic measurements, the pulse interval was
0.8 seconds. Then, manual measurements were per-
formed every few hours until the lapse of 12 hours. The Stage
Stage1 Stage 2
Stage Stage
Stage3
sensor distance was set to 27 mm in experimental series
II. The other methods and conditions of the ultrasonic
(m/s) (m/s)

measurements were the same as those in experimental 3000

series I.
Velocityvelocity
(m/s)

To investigate the hydration products generated in the 2000

specimen, scanning electron microscopy (SEM) was
Velocity
Propagation

conducted intermittently in parallel with the ultrasonic

measurements. SEM samples were prepared at certain 1000
Velocity
time intervals by replacing water in the cement paste
with acetone to stop the hydration reaction. Powder
0 10 20 30 40 50 60 70
X-ray diffraction analysis was performed simultaneously
Elapsed time (hr)
with SEM by using samples in which hydration was (c) Dosage of retarder 0.4 %
stopped by acetone. CuK-alpha radiation was used for (c) Dosage of retarder
the analysis, with a scanning rate of 2 degrees/min and Fig. 4 Wave velocity and elapsed time.
diffraction angles ranging from 2 to 70 degrees.

3. Results and discussion (Experimental

series I)
3.1 Ultrasonic wave velocity
The relationship between ultrasonic wave velocity and
elapsed time is shown in Fig. 4. Time zero in the figure
means the time when cement and water were mixed; only
air existed in the form at that time. The measured veloc-
ity of approximately 340 m/s at time zero in the figure
indicates the wave velocity in the air. After a certain time,
ultrasonic waves could be obtained continuously. The
wave velocity then increased gradually up to levels close
to those found in water (approximately 1500 m/s) and an
inflection point appeared. The period between time zero
and an inflection point is defined as stage 1 in this study.
Subsequently, the wave velocity increased gradually
from approximately 1500 to 3000 m/s and then plateaued
at another inflection point. This behavior is explained by
a decrease in the volume of water and an increase in the
amount of hydration products with progress of cement
hydration in the cement paste. The time period between
 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005 347

the first inflection point and the second inflection point is
called stage 2, and the time period after stage 2 until the
end of measurement is called stage 3. When different
retarder dosages are compared ((a), (b), and (c) in Fig. 4),
the stage boundaries are delayed as retarder dosage in-
creases. This indicates that different setting states in
cement paste can be recognized by looking at ultrasonic
wave velocity and the time of each inflection point in
velocity gradient. Similar findings were obtained in a
previous study using retarder-added concrete (Reinhardt
and Grosse 1996a).
The relationship between ultrasonic wave velocity,
apparent viscosity and elapsed time is shown in Fig. 5,
which also shows the stage boundary defined by ultra-
sonic wave velocity (Fig. 4). The time axis in Fig. 5
amplitude ratio and apparent viscosity are delayed as the
retarder dosage increases, and the maximum amplitude
ratio and apparent viscosity exhibit similar gradient pat-
terns. It is inferred that the maximum amplitude ratio
increased suddenly as the shear stress resistance of the
cement paste increased and the energy of propagated
ultrasonic waves increased abruptly in stage 1.
3.3 Frequency distribution
The changes in frequency distribution with time when no
Stage
Stage 1 Stage2
Stage

Apparent viscosity (Pa࡮s)

covers the range between zero and 15 hours, during 3000 400
which the apparent viscosity was measured. At any re-

Velocity (m/s)
tarder dosage, the apparent viscosity increased abruptly 300
2000
and reached a maximum in stage 1 ((a), (b), and (c) in Fig.
200
5). This phenomenon implies that, as hydration pro-
gressed, aggregation and cohesion of cement particles 1000 Velocity 100
accelerated (Nawa et al. 1986) and shear stress resistance Viscosity
of the cement paste increased. The ultrasonic wave ve- 0
0 5 10 15
locity also increased in stage 1 but more gently than
Elapsed time (hr)
apparent viscosity. In stage 1, the time differential of
apparent viscosity increased with time whereas that of (a) Dosage
(a) Dosage of retarder
of retarder 0%
wave velocity decreased with time.
Changes in ultrasonic wave velocity and Proctor Stage1
Stage Stage
Stage 2
penetration resistance with time are comparatively
shown in Fig. 6. In any case, penetration resistance

Apparent viscosity (Pa࡮s)

showed the timing of initial and final settings (3.5 and 3000 400
27.6 MPa, respectively) in stage 2, during which wave
Velocity (m/s)

velocity increased monotonously. This phenomenon has 300

2000
already been pointed out by the authors (2002) and Voigt 200
et al. (2005). A comparison between the change of ap-
parent viscosity with time in Fig. 5 and the change of 1000 Velocity 100
Proctor penetration resistance with time in Fig. 6 shows Viscosity
that apparent viscosity increased abruptly before the 0
0 5 10 15
initial setting in all cases. The apparent viscosity meas- Elapsed time (hr)
ured with the rotational viscometer in this study is thus
considered to be a suitable index for evaluating the shear (b) Dosage
(b) Dosage of retarder
of retarder 0.2 %
stress resistance of the cement paste prior to its initial
setting determined by the penetration resistance test. Stage1
Stage Stage
Stage 2

3.2 Maximum amplitude

Apparent viscosity (Pa࡮s)

The relationship between the maximum amplitude ratio, 3000 Velocity 400
Viscosity
apparent viscosity and elapsed time is shown in Fig. 7.
Velocity (m/s)

300
The maximum amplitude ratio is defined as the ratio of 2000
the wave amplitude to the maximum amplitude observed 200
within the measurement period in each test case.
1000
As shown in Fig. 7, the maximum amplitude ratio was 100
small at the start of ultrasonic wave measurement, in-
creased abruptly and marked a clear inflection point in 0
0 5 10 15
stage 1, and plateaued in stage 2. This trend is in agree- Elapsed time (hr)
ment with the results obtained by Akashi and Yamaji
Dosage
(c) Dosage of retarder
of retarder 0.4 %
(1958) for the maximum amplitude of longitudinal
waves. Fig. 5 Wave velocity and apparent viscosity.
In Fig. 7, the inflection points of both the maximum
348 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005

Stage
Stage 1 Stage
Stage 2 Stage
Stage 1 Stage
Stage 2

Apparent viscosity (Pa࡮s)

Penetration resistance (MPa)
1.0

Maximum amplitude ratio

400
3000 30
0.8
Velocity (m/s)

300
2000 20 0.6
200
0.4
1000 Velocity 10 Amplitude 100
Penetration 0.2 Viscosity
resistance
0 0
0 10 20 30 0 5 10 15
Elapsed time (hr) Elapsed time (hr)
(a) Dosage
(a) Dosageof retarder 0%
of retarder (a) Dosage
(a) Dosageof retarder 0%
of retarder

Stage
Stage 1 Stage
Stage 2 Stage
Stage 1 Stage
Stage 2
Penetration resistance (MPa)

Apparent viscosity (Pa࡮s)

1.0
Maximum amplitude ratio
400
3000 30
0.8
Velocity (m/s)

300
2000 20 0.6
200
0.4
1000 Velocity 10 Amplitude 100
Penetration 0.2 Viscosity
resistance
0 0
0 10 20 30 0 5 10 15
Elapsed time (hr) Elapsed time (hr)
(b) Dosage
(b) Dosageof retarder 0.2 %
of retarder (b) Dosage
(b) Dosage of retarder
of retarder 0.2 %

Stage
Stage 1 Stage
Stage 2 Stage
Stage 1 Stage
Stage 2
Penetration resistance (MPa)

Apparent viscosity (Pa࡮s)

1.0
Maximum amplitude ratio

3000 30 Amplitude 400

Viscosity
0.8
Velocity (m/s)

300
2000 20 0.6
200
0.4
1000 Velocity 10
0.2 100
Penetration
resistance
0 0
0 10 20 30 0 5 10 15
Elapsed time (hr) Elapsed time (hr)

(c) Dosage
Dosage of retarder
of retarder 0.4 % (c) Dosage
(c) Dosage of retarder
of retarder 0.4 %

Fig. 6 Wave velocity and penetration resistance. Fig. 7 Maximum amplitude ratio and apparent viscosity.

retarder was added are shown in Fig. 8. A frequency
component very close to the resonant frequency of the
US sensor (140 kHz; B in Fig. 8) was dominant in stage 1.
An increase in the shear stress resistance of the cement
paste in this stage is also affirmed by the abrupt increase
in maximum amplitude ratio.
A frequency component around 29 kHz (A in Fig. 8)
became dominant at the lapse of 7 hours and 47 min in
stage 2. After a further lapse of time, the peak frequency
shifted to around 35 kHz and the spectral intensity in-
creased.
In stage 3, the intensity of the spectral peak around 35
kHz (A in Fig. 8) and the resonant frequency of the US
sensor (B in the figure) remained stable.
 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005 349

-4 -3
[u10 ] [u10 ]
2 2

Intensity

Intensity
1 1 BB

0 50 100 150 200 0 50 100 150 200

Frequency (kHz) Frequency (kHz)
Elapsed time:time:
Elapsed 2 hours 00 minutes
2 hours 00 Elapsed time:
Elapsed time: 5 hours
5 hours 00
00 minutes
(a) Stage
Stage11
-3 -3
[u10 ] [u10 ]
AA
2 2
Intensity

BB Intensity BB
1 1
AA

0 50 100 150 200 0 50 100 150 200

Frequency (kHz) Frequency (kHz)
Elapsed time: 7 hours 47 minutes Elapsed time:
Elapsed time: 2929 hours
hours 47
47 minutes
(b) Stage
Stage22
AA
-3 -3
[u10 ] [u10 ]
AA
2 2
Intensity

Intensity

1 BB 1 BB

0 50 100 150 200 0 50 100 150 200

Frequency (kHz) Frequency (kHz)
Elapsed time: 35 hours 47 minutes Elapsedtime:
Elapsed time:
7272 hours
hours 47
47 minutes
(c) Stage
Stage33
Fig. 8 Frequency distribution.
Stage
Stage1 Stage
Stage2 Stage
Stage3

4. Results and discussion (Experimental 5000

(m/s)(m/s)

series II)
4000
velocity
(m/s)

The change in ultrasonic wave velocity with time is

3000
shown in Fig. 9. The time axis in Fig. 9 covers the range
Velocity
Velocity

between zero and 4 hours, during which the ultrasonic 2000

Propagation

wave velocity leveled off. The wave velocity was in the

range of 1100 to 1600 m/s after the start of measurement 1000 Velocity
and then gradually increased to approximately 1800 m/s.
The wave velocity measured in this time region was 0 1 2 3 4
larger in experimental series II than in experimental Elapsed time (hr)
Fig. 9 Wave velocity and elapsed time.
350 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005

series I. This is due to the difference of the water-cement
ratio of the cement paste. This time period (elapsed
times: 0 min to 1 hour 40 min) is defined as stage 1.
The ultrasonic wave velocity increased abruptly after
stage 1 until the lapse of approximately 2 hours and then
plateaued with a clear inflection point; this time period is
defined as stage 2. As is the case with experimental series
I, the increase in wave velocity is explained by an in-
crease in the amount of hydration products in the cement
paste. The wave velocity was nearly constant at 4000 m/s
after the lapse of 2 hours until the end of the measure-
ment; this time period is defined as stage 3. The pla-
teaued level of velocity was larger in experimental series
II than in experimental series I because the water-cement

1.43Ǵm
1.43Ǵm 1.43Ǵm
1.43Ǵm

Elapsed time: 0 hour 04 minutes Elapsedtime:

Elapsed time: 1 00
1 hours hour 00
minutes
(a) Stage
Stage1 1

1.43Ǵm
1.43Ǵm 1.43Ǵm
1.43Ǵm

Elapsed time: 1 hour 40 minutes Elapsedtime:

Elapsed time: 2 00
2 hours hours 00
minutes
(b) Stage
Stage22

1.43Ǵm
1.43Ǵm 1.43Ǵm
1.43Ǵm

Elapsed time: 4 hours 00 minutes Elapsedtime:

Elapsed time: 12 hours
12 hours 00
00 minutes
(c) Stage
Stage33

Fig. 10 Samples of SEM image.

 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005 351

ratio of cement paste was larger in experimental series I
than in experimental series II.
Figure 10 shows SEM images obtained after different
elapsed times. In stage 1 (elapsed times of 4 min to 1
hour), hydration products were hardly found and only
cement particles were observed. After the lapse of 1 hour
and 40 min in stage 2, acicular crystals were observed.
The number of acicular crystals then increased abruptly
over the next 20 min until the lapse of 2 hours in stage 2.
Intensity/a.u.
However, the acicular crystals showed little change
during stage 3 (elapsed times of 4 to 12 hours). Hydration
products other than acicular crystals were rarely found in
the SEM.
To identify the hydration products observed as acicular
crystals in the SEM, X-ray diffraction analysis was car-
ried out; the results are shown in Fig. 11. There was little
difference in the diffraction pattern between the elapsed
times of 4 min and 1 hour in stage 1. During stage 2,

Intensity/a.u.
0 5 10 15 20 0 5 10 15 20
2ǰ/deg. 2ǰ/deg.
Elapsed time:
Elapsed time:00hours
hour 04 minutes
minutes Elapsed time:
Elapsed time: 1 hours
1 hours 00
00 minutes
Stage1 1
(a) Stage
Intensity/a.u.

Intensity/a.u.

0 5 10 15 20 0 5 10 15 20
2ǰ/deg. 2ǰ/deg.
Elapsed
Elapsed time:
time:11hours
hour 40 minutes
minutes Elapsed time:
Elapsed time: 2 hours
2 hours 00
00 minutes
Stage2 2
(b) Stage
Intensity/a.u.

Intensity/a.u.

0 5 10 15 20 0 5 10 15 20
2ǰ/deg. 2ǰ/deg.
Elapsed
Elapsed time:
time:44hours
hours00
00minutes
minutes Elapsed time:
Elapsed time: 12 12 hours
hours 00
00 minutes
Stage3 3
(c) Stage

Fig. 11 Samples of X-ray diffraction.

352 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005
diffraction peaks at approximately 9 and 15 degrees grew
markedly with time. However, there was no change in the
diffraction pattern during stage 3. The X-ray diffraction
patterns obtained for each stage are concordant with the
SEM results. By comparing the spectral peaks (ap-
proximately 9 and 15 degrees) obtained in this study and
the JCPDS (Joint Committee on Powder Diffraction
Standards) card published by the International Centre for
Diffraction Data, the hydration products found by the
SEM and the X-ray diffractometry were identified as
ettringite crystals (Uchida et al. 2003).
Comparison of Fig. 9 with Fig. 10 and 11 indicates
that the abrupt formation of ettringite crystals and the
sudden increase in the ultrasonic wave velocity both
occurred during the time period from 1 hour and 40 min
to 2 hours in stage 2. It was considered that the cement
paste texture became denser owing to the formation of
acicular ettringite crystals, so that initial increase in
strength was achieved (Nagataki and Yamamoto 2000)
and the ultrasonic wave velocity increased.
The cause of the wave velocity increase is thus clearly
associated with the formation of ettringite crystals in the
cement paste.
5. Conclusions
Changes in the physical and chemical properties of ce-
ment paste during its setting and hardening were inves-
tigated in association with its ultrasonic propagation
characteristics from macroscopic and microscopic points
of view. The following conclusions were obtained in this
study.
(1) In experimental series I, apparent viscosity meas-
ured with a rotational viscometer increases sharply
when the ultrasonic wave velocity remains stable at
approximately 1500 m/s.
Therefore, within the limits of this study, evaluation
of changes in apparent viscosity of cement paste
using wave velocity is thought to be difficult.
(2) The maximum amplitude ratio is considered to be
an ultrasonic wave propagation index that reflects
the increase in apparent viscosity associated with
changes in physical properties during the setting and
hardening of cement paste. The apparent viscosity is
an index for evaluating the shear stress resistance of
cement paste. Maximum amplitude is therefore a
characteristic value of propagation that adequately
reflects changes in shear stress resistance.
(3) Frequency distribution, like maximum amplitude
ratio, is a physical property index for evaluating the
shear stress resistance of cement paste. Unlike
maximum amplitude, frequency distribution does
not depend on elastic wave energy or the states of
setting and hardening.
(4) In experimental series II, ettringite crystals were
formed abruptly in a stage where ultrasonic wave
velocity increased suddenly. It is considered that the
chemical properties of the cement paste changed
abruptly and the cement paste texture became
denser owing to the formation of acicular ettringite
crystals, so that ultrasonic wave velocity increased.
The authors intend to extend their study to mortar
and concrete and pursue a method to estimate set-
ting time on the basis of existing physical and
chemical knowledge. With the aim of developing a
nondestructive method to evaluate the setting and
hardening behavior of concrete quantitatively, the
authors also intend to clarify the meanings of the
start of ultrasonic wave propagation, the gradients
of wave velocity and maximum amplitude when
they increase abruptly, and the points at which they
begin to plateau.
Acknowledgements
The authors gratefully acknowledge the financial support
of the Japan Society for the Promotion of Science (Grant
No. 15656105).
References
Akashi, T. and Yamaji, F. (1958). “On measurement of
setting time of concrete by ultrasonic method.” Trans.
of Cement Association of Japan, XII, 323-329. (in
Japanese)
ASTM C403-85 (1985). “Standard Test Method for Time
of Setting of Concrete Mixture by Penetration
Resistance.” Annual Book of ASTM Standards,
Philadelphia, USA.
Biot, M. A. (1956). “Theory of propagation of elastic
waves in a fluid saturated porous solid.” J. Acoust. Soc.
Am. 28, 168-191.
Grosse, C. U., Koble, S. U. and Reinhardt, H. W. (2000).
“Ultrasound, scanning electron microscopy and
nuclear magnetic resonance spectroscopy-comparison
of results investigating the hydration process in
cementitious materials.” Otto-Graf-Journal, 11,
141-156.
ISO 2555 (1989). “Plastics - Resins in the liquid state or
as emulsions or dispersions - Determination of
apparent viscosity by the Brookfield Test method.”
Geneva: International Organization for
Standardization.
Kakuta, S. and Akashi, T. (1983). “Study on physical
properties of cement paste and mortar at early age by
ultrasonic pulse method.” Journal of the Society of
Materials Science, Japan, 32 (353), 175-181. (in
Japanese)
Morin, V., Tenoudji, F. C., Feylessoufi, A. and Richard, P.
(2001). “Superplasticizer effects on setting and
structuration mechanisms of ultrahigh-performance
concrete.” Cement and Concrete Research, 31 (1),
63-71.
Nagataki, S. and Yamamoto, Y. (2000). “Illustrated
Cyclopedia for Concrete Terminology.” Tokyo:
Sankaido Publishing. (in Japanese)
Nawa, T., Eguchi, H. and Ohshima, H. (1986). “Effect of
early cement hydration on rheological behavior of
 T. Kamada, S. Uchida and K. Rokugo / Journal of Advanced Concrete Technology Vol. 3, No. 3, 343-353, 2005
cement paste.” Trans. of Cement Association of Japan,
40, 55-58. (in Japanese)
Reinhardt, H. W. and Grosse, C. U. (2004). “Continuous
monitoring of setting and hardening of mortar and
concrete.” Construction and Building Materials, 18
(3), 145-154.
Reinhardt, H. W. and Grosse, C. U. (1996a). “Setting and
hardening of concrete continuously monitored by
elastic waves.” In: J. P. M. Bartos, D. L. Marrs and D.
J. Cleland, Eds. Production methods and workability
of concrete Proceedings of the International RILEM
Conference, Paisley 3-5 June. Oxon: Spon Press,
415-425.
Reinhardt, H. W., Grosse, C. U. and Herb, T. (2000).
“Ultrasonic monitoring of setting and hardening of
cement mortar – A new device.” Materials and
Structures, 33 (233), 581-583.
Reinhardt, H. W., Grosse, C. U., Weiler, B., Bohnert, J.
and Windisch, N. (1996b). “P-wave propagation in
setting and hardening concrete.” Otto-Graf-Journal, 7,
181-189.
Smith, A., Chotard, T., Gimet-Breart, N. and Fargeot, D.
(2002). “Correlation between hydration mechanism
and ultrasonic measurements in an aluminous cement:
effect of setting time and temperature on the early
hydration.” Journal of the European Ceramic Society,
22 (12), 1947-1958.
353
Uchida, S., Kawamura, A., Kamada, T. and Hisada, M.
(2002). “Basic study on evaluation methods of
hardening behavior of concrete based on ultrasonic
measurement.” Proceedings of the Japan Concrete
Institute, 24 (1), 1569-1574. (in Japanese)
Uchida, S., Kamada, T., Kunieda, M. and Rokugo, K.
(2003). “Evaluation methods for hardening process of
cement paste based on ultrasonic propagation
behavior.” Proceedings of the Japan Concrete
Institute, 25 (1), 1631-1636. (in Japanese)
Valic, M. I. (2000). “Hydration of cementitious materials
by pulse echo USWR method, apparatus and
application examples.” Cement and Concrete
Research, 30 (10), 1633-1640.
Voigt, T., Grosse, C. U., Sun, Z., Shah, S. P. and
Reinhardt, H. W. (2005). “Comparison of ultrasonic
wave transmission and reflection measurements with
P-and S-waves on early age mortar and concrete.”
Materials and Structures, 38 (282), 729-738.
Voigt, T. and Shah, S. P. (2004). “Properties of early-age
Portland cement mortar monitored with shear wave
reflection method.” ACI Materials Journal, 101 (6),
473-482.
Whitehurst, E. A. (1951). “Use of the soniscope for
measuring setting time of concrete.” Proc. ASTM, 51,
1166-1183.