Can the maturity concept be used to separate the autogenous shrinkage and
thermal deformation of a cement paste at early age?
Philippe Turcry a ,Ahmed Loukili a,*,Laurent Barcelo b ,Jean Michel Casabonne b
a
Laboratoire Ge
´nie Civil de Nantes St-Nazaire,Ecole Centrale de Nantes,BP 92101,44321Nantes Cedex,France b
LAF ARGE,Laboratoire Central de Recherche,95rue du Montmurier,38290St-Quentin Fallavier,France
Received 3January 2001;accepted 18March 2002
Abstract
The influence of temperature on the autogenous shrinkage of cement paste has been studied using the maturity approach based on Arrhenius’law.Application of this law requires knowledge of the apparent activation energy,E a ,of cement.In this work,E a has been determined by the ‘‘setting time method.’’The
external volume change of cement paste was measured by hydrostatic weighing.In order to separate the thermal and autogenous deformations,the thermal dilation coefficient (TDC)was determined at both 20and 30°C.Investigations have shown that maturity can be used to predict autogenous shrinkage under isothermal and realistic conditions as long as temperatures remain between 10and 40°C.Outside of this temperature range,the calculated autogenous deformation and measured isothermal shrinkage are quite different and,as a result,autogenous shrinkage appears to be dependent on more than hydration advancement alone.D 2002Elsevier Science Ltd.All rights reserved.
Keywords:Autogenous shrinkage;Temperature;Cement paste;Early age;Maturity concept
1.Introduction
At early age,thermal deformations and autogenous shrinkage occur simultaneously and may,if restrained,lead to the cracking of cementitious materials.These thermal deformations,e th ,result from the temperature rise caused by hydration reactions and are proportional to the thermal dilation coefficient (TDC)of the cement paste:e TH =TDC ÂD T .Autogenous shrinkage,on the other hand,is a consequence of chemical shrinkage:hydration products are denser than the unhydrated compounds.In the plastic state,this contraction is converted into settlement.Once the solid skeleton has been formed
and in the absence of an external source of water,hydration reactions continue through the consumption of capillary water.This phenomenon is called ‘‘self-desiccation.’’According to the Kelvin equation,this self-desiccation gradually increases the tensile stress in pore water,which then leads to a global compressive stress on the
solid skeleton.A deformation,called autogenous shrinkage,e as ,thereby occurs.
For the calculation of stresses at early age,autogenous shrinkage is assumed to be dependent solely on the degree of hydration,a ,with e as =e as (a ),and total deformation is taken as equal to the sum of thermal and autogenous ,e total =e as +e th [1].However,recent studies [2,3]have shown that the autogenous shrinkage amplitude of concrete is influenced by the temperature history.It seems that for a given degree of hydration,concretes with different temperature histories do not develop the same autogenous shrinkage.These results question the applicabil-ity of the maturity concept.
This paper focuses on the influence of temperature on the autogenous shrinkage of cement paste by use of the maturity concept [4].The rate of chemical reactions is affected by temperature,since cement hydration is thermally activated.Moreover,the microstructure itself will undergo change when formed at
different temperatures.The maturity con-cept serves to predict the effect of temperature on certain material properties with respect to the degree of hydration (e.g.,compressive strength).This concept is based on Arrhenius’law,which requires determining the thermal
0008-8846/02/$–see front matter D 2002Elsevier Science Ltd.All rights reserved.PII:S 0008-8846(02)00800-1
*Corresponding author.Tel.:+33-2-40-37-16-67;fax:+33-2-40-37-25-35.
E-mail address :ahmed.loukili@ec-nantes.fr (A.Loukili).
Cement and Concrete Research 32(2002)
1443–1450
sensitivity of hydration reactions,called the apparent activa-tion energy,E a.
This study deals first with the determination of apparent activation energy by using the setting time method[5,6]. Deformations of the cement paste at early age were meas-ured for different thermal conditions,both isothermal and ‘‘realistic.’’The method chosen to assess volume change was the measurement of buoyancy variations for a sample in a flexible latex mold through hydrostatic weighing,as presented in detail elsewhere in the literature[3].With this method,we also measured the TDC,which has allowed calculating thermal deformations in the realistic cases. Lastly,using the maturity approach,we have attempted to separate thermal deformation from autogenous shrinkage.
2.Materials and experimental methods
2.1.Materials
An ordinary type I Portland cement(CPA CEM I52.5 CP2),containing60%C3S,12%C2S,9%C3A,and8% C4AF,was used.The Blaine specific surface was332m2/kg. The water-to-cement(W/C)ratio was0.25.Cement and water were mixed for3min in order to ensure a homogen-eous mix.
2.2.Experimental methods
2.2.1.Determination of E a by means of the setting times method
The maturity concept determines the time required at a reference temperature for the cement paste to achieve the same level of development as that under the influence of the actual time–temperature history.This time is called the ‘‘equivalent age’’and can be calculated by a function based on Arrhenius’law(Eq.(1)):
t equ¼
Z t
0exp
E a
R
1
273þq ref
À
1
273þqðtÞ
d tð1Þ
where t equ is the equivalent age at the reference temper-ature,q ref(h),t is the paste age(h),q(t)is the paste temperature at time t(°C),q ref is the reference temperature (°C),E a is the apparent activation energy(J/mol),and R is the universal gas constant(J/mol/K).
The following proposed relation has been demonstrated theoretically[5]:
ln
D a
D t
¼À
E a
R
1
T0
þconstantð2Þ
where T0is the temperature of the paste(K)under isothermal conditions and D t is the time required to increase the degree of hydration by an increment D a at T0.By plotting the time D t for different temperatures T0,we can deduce E a.Nevertheless,a way must be found to character-ize the degree of hydration a.The setting time method presumes that both the initial set time,t is,and final set time, t fs,are reached for definite levels of hydration(at a given W/ C ratio).This method therefore consists of measuring these specific times(t is and t fs)for several isothermal conditions and,in the subsequent step,calculating an activation energy for each period,initial setting and setting(t fsÀt is).
Initial set time and setting time of cement pastes were measured using the Vicat needle apparatus,which operates in accordance with European Standard EN196-3.The samples were placed in a water bath at various temperatures (10,20,30,and40°C).The temperature inside the samples was measured through embedded thermocouples.
2.2.2.External volume change
Measurements of the external volume change of cement paste were conducted through hydrostatic weighing.Imme-diately after casting,a prophylactic without lubricant was filled with cement paste in order to form a spherical sample. The latex mold was closed using a thin stainless steel wire. The excess latex was then cut and the sample container was cleaned and weighed.The sample was hung under a balance and immediately immersed in the water bath at the chosen temperature.The sample mass was approximately90g.The bath temperature was controlled by a‘‘cryo-thermostat’’(Fig.1)with an accuracy of0.1°C.The external volume change results from a change in the buoyancy force and serves to alter the weight reading on the balance.Both the bath temperature and measured mass were continuously logged on the computer at5-min intervals for a period exceeding24h.
In realistic tests,the bath temperature is imposed by a cement paste sample placed in a quasi-adiabati
c enclosure after casting.The temperature is measured by a ther-mocouple introduced into the cement paste.The weighed sample therefore undergoes the same temperature history.A detailed description of this system is provided in Ref.[3].In order to obtain different temperature histories(Fig.2),the mass of the sample placed in the enclosure is
changed.
Fig.1.Experimental system.
P.Turcry et al./Cement and Concrete Research32(2002)1443–1450 1444
2.3.Calculation of the deformations
The external shrinkage,expressed in cubic millimeters per gram cement,was obtained from Eq.(3)as a result of the calculation already discussed in detail [3]:D V
M c
¼r iw r w ðT ÞÀ1
ÀM r V i r w ðT Þ
V i
M c
ð3Þ
where T is the bath temperature (°C),r iw is the initial density of water (g/cm 3),r w (T )is the density of water as a function of temperature (g/cm 3),V i is the initial sample volume (cm 3),D V is the volume change (cm 3),M r is the mass reading on the balance (g),and M c the cement mass in the sample (g).
3.Results and discussion
3.1.Determination of E a by the setting times method The results related to the initial set and setting times determined by the Vicat needle device have been listed in Table 1.Both the initial set and setting time increase with decreasing temperature.This finding has been previ-ously reported.
Figs.3and 4were drawn by assuming isothermal con-ditions until the initial set;a linear curve was fitted to the data points.According to Eq.(2),the computed apparent activa-
tion energies were 29kJ/mol for the time periods until the
initial set and 39kJ/mol between the initial and final sets.This procedure yields two activation energies,which is not unrealistic since E a reflects more of a mathematical than a chemical tool,thereby making it possible to estimate the global thermal sensitivity of all reactions taking place in the hydrating cement paste.
3.2.Isothermal tests at 10,20,30,and 40°C
Exposed to each temperature,the external shrinkage was determined as the mean value of three measurements
and
Fig.2.Different realistic temperature histories.
Table 1
Initial set and setting time at different temperatures Average
temperature (°C)Initial set time,t is (min)Setting time,t fs Àt is (min)12150138201109034705037
60
30
Fig.4.Arrhenius plot for setting
time.
Fig.3.Arrhenius plot for initial set times.
P .Turcry et al./Cement and Concrete Research 32(2002)1443–14501445
expressed in cubic millimeters per gram of cement,with a variation of less than 8%.
The external shrinkage at different temperatures versus real time has been plotted in Fig.5.It can be observed that temperature has an unsystematic effect on the autogenous deformation of cement paste.This result,which has also been reported by Bjøntegaard [2],is not surprising since the comparison was conducted in an inconsistent manner by virtue of not incorporating the temperature effect on the rate of cement hydration,which is responsible for auto-genous deformations.Comparing autogenous shrinkage curves versus real age is similar to comparing different materials;in this case,we need to apply the maturity concept.
Fig.6shows deformation curves versus equivalent age,which has been calculated for each temperature using Eq.(1).E a is taken as equal to 29kJ/mol before the initial set
and 39kJ/mol thereafter.The reference temperature is 20°C (as results obtained at 20°C do not change with time transformation).It is important to note that deformations must be initialized at the same equivalent ,for the same theoretical degree of hydration.The selected equival-ent time origin was 3.25h,in correspondence with the final set time (Fig.6).It can be seen that the four curves follow si
milar trends with just little scatter,which is caused by measurement accuracy levels.The maturity concept there-fore leads to predicting with reasonable accuracy the auto-genous shrinkage of cement paste under isothermal conditions over the 10–40°C range.3.3.External volume change in realistic cases
By considering the behavior of material to be isotropic,linear deformation represents one-third of the
volumetric
Fig.5.Comparison of deformation for different temperatures versus paste
age.
Fig.6.Autogenous deformation for different temperatures versus maturity,starting after 3.25h.
P .Turcry et al./Cement and Concrete Research 32(2002)1443–1450
1446
deformation calculated in Eq.(3).In this part,deformations are expressed in micrometers per meter.
3.3.1.Deformation kinetics
Three realistic temperature tests with initial temperatures of 10and 20°C were performed.A typical curve of a realistic temperature history (20–60°C)is illustrated in Fig.7.
The realistic deformation curve can be divided into four phases:AB,BC,CD,and DE.In the first phase,it is observed that the isothermal and realistic deformations display the same shape since the temperature is constant in both tests.When temperature increases in the realistic test (B),the curves diverge;during BC,no ‘‘realistic’’deforma-tion is recorded.It seems that thermal deformations com-pensate for autogenous shrinkage.In the case of the ‘‘realistic’’test,thermal expansion occurs until the t
truncated怎么解决emper-ature peak (CD),while the rate of autogenous shrinkage at 20°C is decreasing.During the cooling period (DE),a large deformation caused by both thermal contraction and auto-genous shrinkage can be observed.Moreover,when the temperature returned to 20°C,the magnitude of the ‘‘real-istic’’deformation corresponds to the autogenous shrinkage recorded under isothermal conditions.It can thereby be
stated that the total measured deformation under realistic conditions is equal to the sum of the measured deformation in the isothermal test and a thermal component related to temperature history.In order to separate thermal and auto-genous deformations,we have determined the evolution of the TDC with respect to time.3.4.TDC measurements
The TDC was measured with the method of spontaneous heating by using hydrostatic weighing [3].Tests were carried out under two isothermal conditions:20and 30°C.Every hour,the bath temperature was increased by 4°C within a period of about 7min and then immediately decreased to the isothermal temperature.Fig.8presents the evolution in weight given by the balance without any water density correction.Note that the 4°C peak effect is different at very early age and after 2h,which reveals an evolution in the TDC.
By using this method for the TDC calculation,it was assumed that the temperature rise was sufficiently fast for the deformation measured during heating to be
consid-
Fig.7.Comparison of the isothermal test at 20°C and the 20–60°C realistic
test.
Fig.8.Evolution in mass readings during the TDC
measurement.Fig.9.TDC evolution.
P .Turcry et al./Cement and Concrete Research 32(2002)1443–14501447

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。