Ultrasonic monitoring of setting of green concrete containing high cement substitution by mineral additions

NDTCE’09, Non-Destructive Testing in Civil Engineering Nantes, France, June 30th – July 3rd, 2009

Ultrasonic monitoring of setting of green concrete containing high

cement substitution by mineral additions

Muhammad Irfan Ahmad KHOKHAR1,2, Stéphanie STAQUET1, Emmanuel ROZIÈRE2, Ahmed LOUKILI2

1 Université Libre de Bruxelles, ULB, BATir Department, CP194/4, A. Buyl Avenue 87, 1050 Brussels, Belgium, [email protected] 2 Institut de Recherche en Génie Civil et Mécanique (GeM), UMR-CNRS 6183, Centrale Nantes, Université de Nantes, France

Abstract Setting and hardening are high importance properties for understanding the green

concretes behaviour at early age. The setting and hardening behaviour since casting time of

six green concrete mixtures containing high percentage of mineral additions were monitored

by applying non-destructive ultrasonic waves. During the test, the ultrasonic velocity, the

energy and the frequency spectrum (FFT algorithm) evolution as function of concrete age are

computed. The point corresponding to the first inflexion point on the velocity vs. age plot is

related to the initial setting time. Tests were carried out at two temperatures (20°C and 10°C)

for six mixtures proportions : a reference concrete with Portland cement and the others

containing various proportions of blast furnace slag (30%, 50% and 75% of the binder mass

content) and fly ash (30% and 50%). In order to check the results obtained with the ultrasonic

method, the initial setting time was compared with the Kelly Bryant method.

Résumé La prise et le durcissement sont des propriétés très importantes pour la compréhension du

comportement au jeune âge des éco-bétons. La prise et le durcissement ont été suivis par une

méthode de mesure d’ondes ultrasoniques utilisée depuis le coulage sur six compositions de

béton dont le liant contient des pourcentages élevés en additions minérales. Au cours de

l'essai, la vitesse, l'énergie et le spectre de fréquences (algorithme FFT) des ondes

ultrasonores sont calculés en fonction de l'âge du béton. Le point correspondant au premier

point d’inflexion sur la courbe de vitesse correspond au début de la prise. Les essais ont été

effectués à deux températures, 10°C et 20°C pour six compositions: un béton de référence

formulé avec du ciment Portland, trois bétons formulés avec du laitier de haut-fourneau (30,

50 et 75%) et deux bétons formulés avec des cendres volantes (30, 50%). Afin de vérifier les

résultats obtenus avec les mesures ultrasoniques, le début de prise a été comparé avec la

méthode Kelly Bryant.

Keywords Green concrete, ultrasonic, setting, fly ash, slag, early age

1 Introduction

One of the main goals of the Kyoto protocol is the reduction of greenhouse gases. The

production of one ton of Portland cement generates about one ton of CO2 due to the

combustion process and the de-carbonation of CaCO3. The use of supplementary

cementitious materials (mineral additions) as a partial replacement of Portland cement is one

way to decrease the environmental impact of the cement industry. Various types of by-


products and waste materials, such as coal fly ash and blast furnace slag, are already used as

mineral additions in concrete and provide competitive properties in terms of strength and

durability. Use of very limited percentage of mineral additions in concrete is already in

practice, but use of mineral additions in large proportions in concrete needs detailed

investigation of concrete behavior, specially, at early age which includes setting

phenomenon. An in-depth understanding of the setting process is essential for the study of the

effect of mineral additions on the concrete hardening.

To study the setting of cementitious materials, the methods practiced for years and

considered well established are [1]:

- determining the setting of cement paste by the needle Vicat;

- determining the time of setting of concrete by penetration resistance (penetrometer).

The main drawback of these methods is that they provide only instantaneous data on

hydration and setting properties of the mortar/concrete and not the hydration continuous

process.

For the ongoing study, the FreshCon method of transmitting ultrasound waves through

concrete and the Kelly Bryant mechanical method were used to investigate the setting

phenomenon of fresh concrete using different proportions of slag and fly ash.

2 Concrete mixtures proportions

Besides the mineral additions like slag and coal fly ashes, the concrete mixtures

proportions are composed by a Portland cement (CEM I 52.5 N CE CP2 NF), a

polycarboxylate type superplasticizer, crushed coarse aggregate (10/14 and 6/10 mm) and sea

sand (0/4). Wet sand was used so that corrections were made systematically to adjust the net

water quantity. Six different concrete mixtures proportions were tested at two temperatures

(10°C and 20°C), including a reference concrete with ordinary Portland cement (OPC), three

mixtures using slag with 30, 50 and 75% replacement ratio and two mixtures using fly ash

with 30 and 50% replacement ratio. Kelly Bryant test was carried out only at 20°C due to

problems of portability of testing apparatus. The concrete mixtures proportions tested are

given below in Table 1.

Table 1. Concrete mixtures proportions

Reference Slag Fly ash

(kg/m3) OPC 30% 50% 75% 30% 50%

Gravel 10/14 875 868 875 836 841 842

Gravel 6/10 211 209 211 202 203 203

Sand 0/4 855 848 855 817 822 823

Portland cement 303 219 163 103 241 174

Slag - 94 162 309 - -

Fly ash - - - - 103 175

Net Water 182 182 171 170 182 170

Vpaste (l/m3) 279 285 278 309 305 304

w/c 0.6 0.83 1.05 1.65 0.75 0.98

w/b 0.6 0.58 0.53 0.41 0.53 0.49

Superplasticizer (Sp) - 0.65 1.544 2.26 0.72 1.74

Sp (% of Binder) - 0.21 0.47 0.55 0.21 0.5


3 Mixing Technique

In the reference mix with OPC, the quantity of cement and net water was fixed to 303

kg/m3 and 182 kg/m

3 respectively. The green concretes had to satisfy the following

requirements: same rheological properties as the reference one and a minimal compressive

strength of 10 MPa at 2 days. To achieve these criteria, all the constituents of the different

concrete mixtures proportions had to be varied. Materials were added in the order: cement,

slag or fly ash, then sand and gravel. After mixing dry for 30s, water was added for next 30s,

then after mixing 1 min, superplasticizer was added and left to be mixed for next 90s.

4 Experimental Methods

4.1 Ultrasound method (FreshCon)

The FreshCon device, designed at the University of Stuttgart a few years ago [2,3],

enables the continuous monitoring of the early age concrete hardening behaviour and the

determination of the initial and final setting by ultrasonic measurements (Figure 1). This

device is constituted of two polymethacrylate walls (PMMA) 5.9 cm apart and held by four

screws. A foam material is used to make a U-shaped mould. Its high damping properties are

able to suppress waves passing trough the mould and moving around the concrete sample

(concrete volume equal to 450 cm³). A pulse signal with a width of 2.5µs of which the

amplitude was enhanced by an amplifier was generated at selected regular intervals during

this test. The ultrasonic longitudinal compression wave was then transmitted through the

concrete sample by means of a piezoelectric broadband transmitter. The signal was received

by an ultrasonic receiver after travelling through the sample and sent back to the data

acquisition card (DAQ card). During the test, ultrasonic velocity, energy and frequency

spectrum evolution are computed by the FreshCon software.

The tests were carried out in an air-conditioned room at 10°C and 20°C and the sample

was protected by a plastic foil against desiccation phenomenon and risk of uncoupling

between concrete and the mould walls due to the shrinkage development. The settings of the

different mixtures proportions are detailed in Table 2.

4.2 Kelly Bryant

The method of Kelly-Bryant described in [4] is a mechanical method and is used

exclusively in Belgium. The principle of the determination of the initial setting time of

concrete using this method is based on the measurement of the force required to wrench

stainless bars embedded in concrete over time. Concrete was placed in a rectangular prismatic

mould 10x10x60 cm placed horizontally. Stainless steel bars maintained vertically by guides

were disposed in the mould (Figure 2).

During the test, the sample was protected against evaporation using plastic plates. These

guides were removed after the concrete casting and the vibration phase. The bars were pulled

out one by one from concrete at regular intervals with a mechanical system. The tensile force

needed to pull out the bars from concrete is measured by a dynamometer.


Figure 1. FreshCon device: a

computer, a signal generator and a

concrete mould

Figure 2. Steel bar wrenching device for concrete

(Kelly Bryant Method)

5 Results and Analysis

The change of ultrasonic velocities as a function of time for concrete mixtures with

additions of slag and fly ash tested at 20°C and 10°C are presented in Figure 3 & Figure 4

respectively. The ultrasonic velocity evolutions result in S-shape curves. The first part is the

dormant period, characterised by a quasi constant low velocity value which can be attributed

to the presence of air entrained during the mixing procedure. Fresh concrete can be

considered as water-saturated media rather than suspension and already have an elastic

granular frame before the cement grains start to hydrate [5]. The determination of the signal

onset time and the wave velocity is less accurate during this early period (repeatability error

of 10%) due to larger signal attenuation compared to a more hardened material.

After this period, the velocity increases rapidly at first (second stage) and then gradually

(third stage) to finally reach an asymptotic value. This first stage can not be linked to the

initial setting. According to Scrivener [6], during the first three hours of cement hydration

mainly ettringite is formed outside of the primarily still unhydrated cement grains in the

shape of small needle. Although these needles do not create bonds between the cement

particles, but they do fill pore space that was previously occupied by water with solid

products. In contrast to the stiffening behaviour, the ultrasonic velocity is strongly affected by

the formation of the ettringite. Voigt et al. [3] investigated that, the very early increase in

ultrasonic velocity is not attributed to setting, but to the formation of ettringite and to internal

settling. Hydration products such as ettringite do not create connected particles and have no

or little influence on the stiffening process. However, they fill pore space as a result of which

the porosity and the air content decrease and the velocity increases. At the same time internal

settling due to gravity causes a better mechanical coupling of the particles without a real

bond. The major increase in velocity takes place due to start of percolation of cement

hydrates which form complete pathways of connected particles for the ultrasonic pulse wave

[7]. According to Robeyst et al. 2007 [8], the initial setting corresponds to a moment equal to

the inflexion point of the velocity curve, while we have determined this inflection point

mathematically by calculating the derivative of the velocity curve. In the last stage, the

velocity increases at a very low rate and finally reaches at an asymptotic value. This slow

increase in velocity (indicated by a constant value of the derivative of the velocity curve close

to zero) corresponds quite well with the final setting.


0

1000

2000

3000

4000

5000

0 6 12 18 24Age (h)

Vel

oci

ty (

m/s

ec)

OPC

Slag 30%

Slag 50%

Slag 75%

Fly ash 30%

Fly ash 50%

Figure 3. Evolution of ultrasonic velocity vs. concrete age at 20°C

0

1000

2000

3000

4000

5000

0 6 12 18 24 30 36Age (h)

Vel

oci

ty (

m/s

ec)

OPC

Slag 30%

Slag 50%

Slag 75%

Fly ash 30%

Fly ash 50%

Figure 4. Evolution of ultrasonic velocity vs. concrete age at 10°C

Frequency change calculated with a FFT-algorithm during ultrasound signal transmission

with time (Figure 5) provides information about the setting evolution. Very low frequencies

are observed at very early age when the cementitious particles are said to be in suspension. At

the percolation threshold, when the suspension state changes to solid state, frequency start to

appear very clear in the spectrum [9]. The values obtained are close to the initial setting

determined by the ultrasound velocity curves (Figure 3,Figure 4 & Table 2). With the

increase in mineral addition content, very low frequencies are observed for longer duration

before the frequency start to appear clearly in the spectrum. The same behaviour of the

frequency was observed at the low initial temperature of the test.


(a) (b)

Figure 5. Evolution of frequency spectrum for concrete mixture with only OPC at 20°C (a)

and 10°C (b)

Using Kelly Bryant method, the initial setting corresponds to the first significant slope

change on the force evolution curve as function of time (Figure 6). Actually the bond strength

occurs in two phases. Constant low values of pulling force can be observed in the first phase.

In the second phase the strength increases at a rapid rate due to stiffness. The stiffening

process of the cement paste as a constituent of the tested concrete is due to the development

of rigid connections between the cement grains caused by the hydration products [6]. It is this

gradual development of solid microstructure that causes the wrenching force to gain higher

values with increasing hydration time. This transition from low constant to the linear rapid

increase in bond strength is linked to the initial setting of concrete sample. The results

obtained from this method for initial setting are given in Table 2 and are close to the initial

setting determined by the ultrasound velocity curves (Figure 3).

0

50

100

150

200

250

300

350

400

450

2 3 4 5 6 7 8 9

Age (h)

Wre

nch

ing

Forc

e (N

)

10

OPC

Slag 30%

Slag 50%

Slag 75%

Fly ash 30%

Fly ash 50%

Initial setting

Figure 6. Strength evolution in the concrete mixtures (Kelly Bryant method)


Table 2. Initial and final setting times measured by different techniques Reference Slag Fly ash

Temperature (°C) 20 10 20 10 20 10 20 10 20 10 20 10

Time (h)/Mixture OPC 30% 50% 75% 30% 50%

Initial Setting

(Velocity analysis) 4.2 6.0 5.0 6.98 6.3 7.35 5.78 7.6 6.93 8.45 8.58 10.46

Initial Setting

(Kelly Bryant) 4.84 - 5.47 - 5.69 - 5.4 - 6.18 - 7.4 -

Initial Setting

(Frequency) 4.14 6.63 5.75 7.74 6.01 8.1 6.2 8.2 6.59 10 8.15 11.6

Final Setting

(Velocity) 7.45 10.08 8.66 11.56 10.05 12.43 8.86 12.35 12.6 12.86 15.16 16.55

5.1 Effect of replacement rate over setting time

The change in ultrasonic velocities as a function of concrete age is given in Figure 3 &

Figure 4 for concrete mixtures with different replacement percentages of slag and fly ash.

Setting time has been delayed with the increasing percentage of slag comparing to OPC

concrete. Since the ultrasonic velocity is a measure for the stiffness of the hardening concrete

sample, the stiffness seems to develop slower with increasing slag content. This delay is

attributed to the fact that the hydration of the slag is not initiated until the lime liberated

during the hydration of OPC provides the correct alkalinity [10,11]. The addition of fly ash

resulted in a longer dormant period, lower values of the pulse velocity at early ages and a

longer time necessary to asymptotically reach the maximum velocity. Fly ashes exhibit

pozzolanic activity: in finely divided form and in the presence of moisture, they chemically

react with calcium hydroxide to form compounds possessing cementitious properties. In

concrete, the Ca(OH)2 essential for the activation of fly ash mostly comes from the hydration

of Portland cement so that concrete with fly ashes shows a delay in microstructure

development and thus in setting and hardening.

5.2 Effect of temperature over setting time

The setting time is a function of the curing temperature and increases with lower initial

temperature, for all type of concretes. At 10°C, first the dormant period is longer and then

time to reach asymptotic value is delayed too due to low rate of hydration at low temperature.

Table 2 shows that for the reference mix, the initial setting time has been retarded from 4.2h

at 20°C to 6h at 10°C while the final setting time from 7.45h to 10.08h. For concretes with

slag content 30, 50 and 75%, the initial setting time has been retarded from 5, 6.3 and 5.78h

at 20°C to 6.98, 7.35 and 7.6h at 10°C and the final setting time from 8.66, 10.05 and 8.86h to

11.56, 12.43 and 12.35h. Likewise behaviour was observed for concretes with fly ash content

30 and 50%: 6.93 and 8.58h as initial setting time at 20°C to 8.45 and 10.46h at 10°C.

6 Conclusions

Concrete mixtures with mineral additions were studied for the evolution of setting

phenomenon at temperatures of 10 and 20°C. The non-destructive FreshCon method allows

monitoring of microstructural changes occurring in a concrete in a continuous way. It turned

out that the initial setting measured by the ultrasonic velocity coincides rather well with the

time of increase of the pulling force by Kelly Bryant method. Increase in mineral addition

content delays the setting phenomenon, in case of slag due to its latent hydraulic property and

due to slow pozzolanic reaction in fly ash. In comparison to mixtures with slag, where the

hydraulic properties of slag can manifest same as properties of Portland cement, the mixtures

with fly ash show slow setting due to their latent pozzolanic action (which can be seen after


some weeks). Initial temperature has an inverse effect on the setting of concrete, lower

temperature delays the setting notably.

Acknowledgements The authors are pleased to thank Nicolas Poncelet, a student from ULB for his collaboration. The support of

Higher Education Commission of Pakistan and Belgian National Foundation for Scientific Research is also

acknowledged.

References

1. Rilem 185-ATC (2005) "Advanced testing of cement based materials during setting and

hardening", Report 31, Ed by H.W. Reinhardt and C.U. Grosse, 341p.

2. Reinhardt, H. W., and Grosse, C. U. (2004) “Continuous Monitoring of Setting and

Hardening of Mortar and Concrete,” Construction & Building Materials, 18, pp. 145-154.

3. Voigt, T., Grosse, C. U., Sun, Z., Shah, S. P., Reinhardt, H. W. (2005) “Comparison of

ultrasonic wave transmission and reflection measurements with P- and S-waves on early

age mortar and concrete”, Materials & Structures 38 (282), pp.729–738.

4. Kelly, T.M., Bryant, D.E. (1957) “Measuring the rate of hardening of concrete by bond

pullout pins”, Proc. of the American Society for Testing and Materials, 57, pp.12.

5. Robeyst, N., Gruyaert, E., Grosse, C. U., De Belie, N. (2008) “Monitoring the setting of

concrete containing blast-furnace slag by measuring the ultrasonic p-wave velocity”,

Cement & Concrete Research 38, pp.1169–1176

6. Scrivener, K.L. (1989) “The microstructure of concrete' in Skalny, J. (ed.), Materials

Science of Concrete”, 1, American Ceramic Society, Inc., Westerville, Ohio, pp.127-161.

7. Ye, G., Lura, P., Breugel, K. V., Fraaij, A. L. A. (2004) “Study on the development of the

microstructure in cement-based materials by means of numerical simulation and ultrasonic

pulse velocity measurement”, Cement & Concrete Composites 26 (5), pp.491–497.

8. Robeyst, N., Gruyaert, E., De Belie, N. (2007) “Ultrasonic monitoring of setting and

hardening behaviour of concrete and mortar with blast-furnace slag cement”, 12ème

Congrès Int. de la chimie du ciment, 8-13 Juillet 2007, Montréal, Canada.

9. De Belie, N., Grosse, Ch. U., Robeyst, N., Gruyaert, E. (2006), "Evolution of different

ultrasound transmission parameters to monitor microstructure development in fresh

concrete or mortar”, 2nd

Int. Symp. on advances in concrete through science &

engineering, 11-13 September 2006, Quebec city, Canada, p.313.

10. Neville, A. M. (1995) “Properties of Concrete”, Longmen, Essex.

11. Taylor, H. (1990) “Cement Chemistry”, Thomas Telford Publishing, London.

Ultrasonic monitoring of setting of green concrete containing high cement substitution by mineral additions

Documents