A re-examination of creep mechanisms in hydrated cement systems

Publisher’s version / Version de l'éditeur:

Industria Italiana del Cemento, Sept. 757, pp. 688-697, 2000-10-01

READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE. https://nrc-publications.canada.ca/eng/copyright

Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la

première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.

Questions? Contact the NRC Publications Archive team at

PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.

NRC Publications Archive

Archives des publications du CNRC

This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.

Access and use of this website and the material on it are subject to the Terms and Conditions set forth at

A re-examination of creep mechanisms in hydrated cement systems Beaudoin, J. J.; Tamtsia, B. T.; Marchand, J.

https://publications-cnrc.canada.ca/fra/droits

L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.

NRC Publications Record / Notice d'Archives des publications de CNRC:

https://nrc-publications.canada.ca/eng/view/object/?id=1797c43d-2eca-4b3f-9bbd-4a2d360a9c29 https://publications-cnrc.canada.ca/fra/voir/objet/?id=1797c43d-2eca-4b3f-9bbd-4a2d360a9c29

A re-examination of creep mechanisms in

hydrated cement systems

Beaudoin, J.J.; Tamtsia, B.; Marchand, J.

A version of this paper is published in / Une version de ce document se trouve dans : L'Industria italiana del Cemento, no. 757, September 2000

www.nrc.ca/irc/ircpubs

A RE-EXAMINATION OF CREEP MECHANISMS IN HYDRATED CEMENT SYSTEMS

J.J. Beaudoina, B. Tamtsiab and J. Marchandc a

Institute for Research in Construction, National Research Council of Canada, Ottawa, Ontario, Canada K1A OR6

b

Department of Civil Engineering, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5

c

Department of Civil Engineering, Université Laval, Ste-Foy, Québec, Canada G1K 7P4

ABSTRACT

Mechanisms of creep of hydrated Portland cement paste are reviewed with reference to the role of water. The coupling of an a.c. impedance analyzer with a miniature loading system was used to follow real-time microstructural changes due to sustained load.

Various pre-drying treatments (including use of solvent exchange methods) were used to probe the sensitivity of the creep process to microstructural change and the presence or absence of moisture. Characterization of the water in the paste was carried out using a controlled environment differential thermogravimetric technique. Evidence suggests that the presence of water is not an 'à priori' condition for creep. A creep mechanism involving microsliding between C-S-H sheets appears to be compatible with the experimental evidence.

INTRODUCTION

Water is generally considered to have a seminal role in the majority of the hypotheses for the creep behavior of hardened cement paste [1,2]. For example it is integral to the seepage theory which describes time-dependent volume change due to applied load in terms of changes in the internal vapor pressure and hence in the so-called gel water content [3]. Variations of the seepage theory have been reported [4]. Other theories dependent on some form of water interaction include: the viscous shear theory (creep occurs through slip between C-S-H particles in a shear process in which water acts as a lubricant) [5]; the thermal activation theory (water plays an indirect role through its effect on disjoining pressure which, in turn, weakens interparticle bonds) [1]; microprestress-solidification theory (microprestress is generated by the disjoining pressure of the hindered adsorbed water in the micropores and by the very large and highly localized volume changes caused by hydration or drying) [6]. Physico-chemical processes during creep include: silica polymerization of C-S-H induced by loading or drying [7]; the interlayer consolidation (creep is a manifestation of the gradual aging of a poorly crystallized layered silicate material accelerated by drying or stress) [8]; microcracking near non-shrinking calcium hydroxide crystals [9].

A detailed description of the theories referred to above is provided in an excellent review by Neville [4]. There remains however, despite the various theories proposed, no universally accepted theory or mechanism for creep of hardened cement paste.

The experiments discussed in this paper were specifically designed to re-assess the importance of water and related volume change mechanisms in the creep process. Creep data was obtained on miniature specimens under environmentally controlled conditions. The monitoring of real-time changes in the microstructure of cement paste subjected to sustained load through a coupled a.c. impedance loading system facilitated the assessment.

EXPERIMENTAL PROGRAM

Specimen Preparation and Characteristics

The Portland cement paste used to fabricate the creep specimens was prepared at a water-cement ratio = 0.50 and hydrated for periods up to 30 years. The Portland water-cement had the following composition in percent: SiO2 (20.7); Al2O3 (5.9); Fe2O3 (3.1); CaO (62.7);

MgO (3.5); SO3 (2.2) and the lime (0.2). The Bogue composition in percent was as

follows: C3S (46.5); C2S (24.6); C3A (10.4) and C4AF (8.3). Mixing details are provided

elsewhere [10].

The C3S paste samples were prepared at a water-solid ratio of 0.4. The procedure was

similar to that for fabrication of the Portland cement paste. These samples were hydrated for 28 days. The C3S had the following composition in percent: CaO (73.9); SiO2 (26.2);

Al2O3 (0.08); free CaO (0.46).

"T" shaped specimens 25.4 mm long were cut from paste cylinders for the creep and shrinkage measurements. They had a cross-section 7.00 mm deep with a flange width of 12.70 mm and flange and web thicknesses of 1.27 mm. Details of the fabrication process are also given elsewhere [10].

Specimen Pre-treatment

Four series of test specimens were fabricated and conditioned to provide a wide variety of microstructural treatment and pore structures for the creep and shrinkage tests.

Series I (All tests were performed in a saturated condition)

The "T" shaped Portland cement paste specimens were conditioned (prior to testing) as follows:

(1) Untreated - samples were saturated surface dry and directly used for test without any further treatment.

(2) Drying at 37°C - samples were initially saturated surface dried, vacuum dried at 37°C for 24 hours and then re-saturated (with synthetic pore solution) under vacuum for 18 hours. Prior to the re-saturation process, the samples were vacuum dried for 3 hours in a dessicator at 20°C and 1x10-4 mm Hg.

(3) Methanol exchange - samples were soaked in methanol for 48hours, vacuum dried at 37°C for 24 hours, then re-saturated (with synthetic pore solution) under vacuum for 18 hours. Prior to the re-saturation process the samples were vacuum dried for 3 hours in a dessicator at 20°C and 1x10-4 mm Hg. (4) Isopropanol exchange - samples were conditioned in a manner similar to that

for the methanol exchange process.

Series II (All tests were performed in a dry condition)

The "T" shaped C3S paste specimens were subjected to the following drying treatments.

(1) Reference state - the specimens were all dried to 11% RH for 30 days in a dessicator and then re-saturated in lime-saturated water for 14 days. This was followed by D-drying.

(2) Methanol exchange - the specimens prepared as described for the reference specimens were exchanged with methanol after re-saturation with lime-water. This was followed by D-drying.

(3) Isopropanol exchange - the specimens prepared as described for the reference specimens were exchanged with isopropanol after re-saturation with lime-water. This was followed by D-drying.

Series III (All tests were performed in the saturated state)

The "T" shaped Portland cement paste specimens were conditioned (prior to testing) as follows:

(1) Reference state - specimens were tested in their original saturated state.

(2) Drying to 42% RH - reference specimens were dried to 42% RH and re-saturated with lime-water.

(3) D-drying - reference specimens were D-dried and then re-saturated with lime-water.

The series III enables the evaluation of intermediate drying and D-drying on the creep of saturated Portland cement paste.

Series IV (All tests were performed in a dry condition)

The "T" shaped Portland cement paste specimens were subjected to the following drying treatments.

(1) D-dry - specimens were oven-dried at 105°C for 3 hours.

(2) Drying at 37°C - specimens were vacuum dried at 37°C for 24 hours.

(3) Methanol exchange - specimens were subjected to a methanol exchange process and then vacuum dried at 37°C for 24 hours.

(4) Isopropanol exchange - specimens were subjected to a isopropanol exchange process and then vacuum dried at 37°C for 24 hours.

It is important to distinguish the conditions for solvent exchange in the test series II from those of the series I and IV. In series II all specimens are dried to 11% RH and then re-saturated prior to exchange. Pore coarsening effects are therefore similar for all preparations prior to exchange.

Thermal Analysis

Differential thermogravimetric Analysis (DTGA) was used to characterize the state of water in the samples. The method provided information on the effect of the various pre-treatments on the amount of water associated with the C-S-H phase. A Dupont 951 Thermal Analyzer placed in an environmentally controlled chamber was used for the tests.

Creep Measurement System

The a.c. impedance creep and shrinkage spectral responses were carried out by mounting the "T" shaped specimens (two per frame) in a miniature creep frame linking the specimens to a load cell through electrode interfaces which were connected to a Solatron 1260 frequency response analyzer. The creep frames were placed in environmentally controlled cells. Modified Tuckerman optical extensometers were used for length change measurements. They were mounted on the flanges of each of the two "T" shaped specimens. Creep strain was monitored with an accuracy of about 1 µstrain. Details of the measurement system are provided in reference 10.

RESULTS AND DISCUSSION

The total strain (creep + shrinkage) - time curves for the series I Portland cement paste specimens (w/c=0.50, stress-strength ratio=0.30) under sustained load are presented in figure 1. The total strain at 72 hours is 320, 450, 515 and 600 µ∈ for the control specimens and those dried at 37°C, solvent exchanged with isopropanol and solvent exchanged with methanol respectively prior to re-saturation with synthetic pore solution. The strain recovery for the control specimens and those dried at 37°C or exchanged with isopropanol prior to re-saturation with pore solution is about 100 µ∈. The value for the methanol exchanged specimens is about 200 µ∈. The increase in total deformation of the dried (37°C) or solvent exchanged specimens may be partly due to the pore coarsening effect. Nevertheless the collapse of the C-S-H structure on drying (a form of hindered aging) prior to re-saturation may occur to a lesser extent for pastes that have undergone solvent exchange. This may account for the higher strain values observed. The drying creep - time curves (not shown) display similar relative differences between the treated and untreated control specimens.

The a.c. impedance spectra for the total strain measurements of methanol exchanged specimens plotted in figure 2 are typical. The size of the high frequency arcs for the

unloaded specimens (not shown) is significantly greater than the corresponding arcs (at each specific time) for the loaded specimens.

The growth of the high frequency arc diameter (total strain conditions) for all the specimens subjected to the four pre-treatments is plotted in figure 3. The size of the arc at 72 hours is in the following order: untreated > isopropanol exchanged > methanol exchanged > drying at 37°C. The total strain (figure 1) was in the following order: methanol exchanged > drying at 37°C > isopropanol exchanged > untreated. These results are consistent with the relative pore coarsening effects due to the pre-drying treatments.

The smaller arc sizes for the total strain sequences (compared to shrinkage) suggests that processes involving displacement (slipping and sliding) of C-S-H sheets induced by the applied load are operative.

Previous work has shown that loading cement paste up to about 50 percent of the maximum has a little effect on the size of the high frequency arc [11]. This suggests that microcracking processes had little effect on the a.c. impedance response (stress/strength ratio = 0.30) obtained in these experiments.

The differences in high-frequency arc diameter between the shrinkage and the total strain impedance spectra (at corresponding times) generally increase with time. The differences at 24 hours are in the following order: methanol exchanged > isopropanol exchanged > untreated > pre-drying at 37°C. The large differences for the methanol and isopropanol exchanged samples are consistent with the large relative creep observed for these specimens.

The high-frequency arc is usually imperfect and depressed below the real axis by angle expressed as αd.π/2 where αd is an indicator of the extent of the arc depression. The

depression angle parameter n (1-αd) is plotted against time for the cement paste

specimens in figure 4. The range of values of n are 0 < n < 1 where n = 1 represents a perfect arc. The initial value of n is in the following order: untreated > methanol exchanged > isopropanol exchanged > pre-drying at 37°C. The order remains the same for the first few hours with the exception of n for the untreated specimen which decreases rapidly between 2 and 6 hours reaching a value significantly lower than for the treated

specimens. This suggests that initially the drying treatments perturb the surface. The dependence of the arc depression has been linked to fractal characteristics of surfaces [12]. The low values of drying creep for the untreated specimens may be related to the low values of the depression angle parameter. The fractal nature of the surface may affect the slipping-sliding behavior of the C-S-H sheets in cement paste under load.

Analysis of the specific total strain rate provides additional evidence for a slipping-sliding mechanism. The specific total strain rate versus time is plotted as a log-log relation in figure 5. Two straight lines describe the data. The line for the untreated control specimens lies below a line representative of the other specimens. The co-linearity of these type of curves (in other investigations) suggests that specific creep rate is independent of loading time and degree of hydration [13]. It has been shown that the long-term aging effect can be characterized by a decrease of the creep rate inversely proportional to the material age. This would suggest that the rate determining mechanism is associated with the behavior of the C-S-H sheets themselves, perhaps involving a slipping and sliding process as suggested previously.

The C3S paste "T" shaped specimens (series II) were dried at 11% RH then re-saturated,

immersed in an organic solvent and D-dried prior to loading. The results are presented in figure 6 in terms of specific basic creep i.e. micro-strain per unit stress. All specimens exhibit significant creep in the dry-state. The results indicate that solvent exchange with methanol before drying reduces creep. Methanol treated samples exhibit less specific basic creep than both the isopropanol exchanged and the untreated control specimens. It is especially important to note that the D-dried reference samples exhibit significant creep. This suggests that the presence of water is not an 'à priori' condition for creep to occur. Slipping and sliding of C-S-H layers may be an operative creep mechanism [10]. In these experiments all the specimens were initially dried at 11% RH (as indicated above) and therefore experienced a similar degree of pore 'coarsening' and C-S-H microstructural change prior to re-saturation and solvent exchange. Under these conditions the removal of water by a solvent prior to D-drying may alter the microstructural changes taking place in normal drying and therefore reduce the creep capacity of the hydrated C3S paste. This is significantly different to what occurs when

solvent exchange occurs prior to initial drying. In the latter case there appears to be a 'hindered' aging process that results in an increase of creep due to the exchange process. The DTGA results (figure 7 gives results for the methanol exchange only) indicate that immersion of C3S paste, treated as described above, may remove water from C-S-H and

therefore, bring the sheets into closer proximity once dried under vacuum. This may explain the smaller magnitude of creep in such C3S paste specimens (series II)

particularly those immersed in methanol.

Figure 1 Total strain (Creep+Shrinkage) and strain recovery of hardened cement paste (w/c=0.5) onditioned at about 96% relative humidity after resaturation from different drying pretreatment. 0 100 200 300 400 500 600 0 20 40 60 80 100 120 140 time, in hours m ic rostrai n, i n µµµµ m/ m Reference (untreated)

Dried at 37°C for 24 hours then resaturated with pore solution

Immersed in methanol, dried at 37°C then resaturated with pore solution Immersed in isopropanol, dried at 37°C then resaturated with pore solution

The magnitude of specific basic creep of saturated Portland cement paste specimens (series III) is dependent on the previous drying history (figure 8). Samples dried at 42% RH or D-dried have significantly greater creep at 25 days than the reference specimens. Creep of the saturated reference specimens is greater than or similar to that of series IV Portland cement paste specimens tested in the D-dry state (figure 9). The amount of creep of dry specimens can be very large if the specimens are immersed in organic solvents prior to the drying process. The removal of water by a solvent prior to vacuum drying may reduce the microstructural changes taking place in normal drying (a form of 'hindered' aging) and therefore increase the creep capacity of hardened cement paste as indicated by the results in figure 9 and the corresponding high surface area. The creep result (methanol exchange) can also be explained by the formation of a new complex from the reaction between the solvent and the cement paste.

Methanol has been shown to react with CH and C-S-H [14]. Such chemisorbed products and/or complexes that possibly form may weaken the C-S-H surface resulting in an increase in the sliding capacity between sheets.

Figure 2. AC impedance spectra: total strain of hcp methanol soaked vacuum dried at 37°C and resaturated with pore solution (w/c=0.5); specimens conditioned at ab out 96% relative humidity for 0, 1, 2, 3.66, 6.5, 9, 12.66, 24, 48, and 72 hours

-12000 -10000 -8000 -6000 -4000 -2000 0 0 2500 5000 7500 10000 12500 15000 17500 20000 22500 250 00 27500 30000 Re al, in ohm s Im a g ina ry , in oh m s

96% RH Arc diameter increasing

w ith time under load : 0 1h 2h 3.66h 6.5h 9h 12.66h 24h 48h 72h

Figure 3. High-frequency arc diameter R2 (following re-saturation of cement paste (w/c=0.50) after

several drying conditions) during a total strain (Creep + Shrinkage) test at 96% RH

0 5000 10000 15000 20000 25000 30000 35000 40000 45000 0 10 20 30 40 50 60 70 80 time, in hours s e m i-c ir c le di amet er R 2 , in o h m s untreated dried at 37°C methanol isopropanol

Figure 4. Depression angle parameter of high frequency arc following re-saturation of cement paste (w/c=0.50) after several drying conditions

0.6 0.7 0.8 0.9 1 0 10 20 30 40 50 60 70 80 time, in hours depr es si on angl e par a m e te r, n = (1-αd ) Reference (untreated)

Dried at 37°C for 24 hours then resaturated with pore solution Immersed in methanol, dried at 37°C then resaturated with pore solution Immersed in isopropanol, dried at 37°C then resaturated with pore solution

Figure 5. Specific total strain rate of hardened cement paste (w/c=0.50) conditioned at ~96% relative humidity after re-saturation from different drying pre-treatments

[ R2 = 0.8311 ] [ R2 = 0.8446 ] 0.1 1 10 0.1 1 10 100 time, in hours s p e c if ic t o ta l s tr a in ra te , in µ∈ µ∈ µ∈ µ∈ / M p a / h Power Power

Pre-drying (see text)

5512 . 0 ) ( 4199 . 2 ) , ( − − = t t_o dt o t t dJ 7322 . 0 ) ( 4145 . 5 ) , ( − − = t t_o dto t t dJ Control Control Pre-drying 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80

loading period, in days

Spec ific bas ic creep , µ∈µ∈µ∈µ∈ /Mp a Reference Methanol Isopropanol

Figure 6. Specific basic creep of D-dried C3S paste (74% hydrated)after alcohol

Plots of the compliance (total deformation per unit stress) versus time (log-log scale) for series III and IV were linear (as were the plots for the cement paste specimens series I that were monitored by a.c. impedance spectroscopy described earlier, figure 5). The curves are power functions that can be expressed as follows:

where J (t, to) is the compliance of hardened cement paste at age t loaded at to, α and β

are constants and (t - to) is the time elapsed under loading.

The range of variation for the β parameter in this study is small i.e. 0.71-0.94. The assumption that α is the only variable seems reasonable [13]. The values of α for the solvent exchanged specimens are more than 5 times those for the D-dried or vacuum dried specimens (37°C for 24 hours) and approximately 4 times the values for the other specimens tested in the saturated condition. This variation indicates that methanol and isopropanol exchanged specimens behave quite differently than the others. The

β

α

(

)

)

,

(

t

t

dt

t

t

dJ

−

=

comments referred to above concerning the independence of loading time and degree of hydration on specific basic creep rate also apply to these experiments. The results provide further evidence for the view that the rate determining mechanism is associated with the behavior of the C-S-H sheets themselves, perhaps involving a slipping and sliding process.

CONCLUSIONS

1. Creep of Portland cement paste subjected to drying is significantly affected by the solvent exchange pre-drying treatment. This also applies to creep recovery.

2. A.C. Impedance spectroscopy can detect real-time microstructural changes in Portland cement paste unloaded or subjected to a sustained load.

3. The observation of smaller high frequency arc sizes in the a.c. impedance spectra for total strain measurements relative to the arcs obtained for shrinkage measurements suggests that continuous processes involving slipping and sliding of the C-S-H sheets are operative when cement paste is under sustained load.

4. The value of the high-frequency arc depression angle in the a.c. impedance spectrum appears to reflect the relative creep potential of hardened cement paste.

5. The linear character of the log specific strain versus log time relation for Portland cement paste subjected to a sustained load suggests that the rate determining mechanism may be associated with slipping and sliding of the C-S-H sheets.

6. The D-dried tricalcium silicate reference paste exhibits a significant amount of creep confirming that creep mechanisms associated with water transport are not necessarily dominant.

7. Methanol and isopropanol exchanges of tricalcium silicate paste (after drying to 11% RH and re-saturation with water) reduce the magnitude of creep in the dry state. This may be due to irreversible interactions of the solvent with C-S-H.

8. Portland cement paste dried to an intermediate humidity and re-saturated creeps significantly more than saturated cement paste that has not been previously dried.

This is attributed to the pore coarsening effect due to drying and possible increase in creep sites due to increased layering of C-S-H.

9. Specific basic creep can be expressed as a power function. The parameters depend on the water content of the samples and/or on the drying condition and pre-treatment history. 0 25 50 75 100 125 0 5 10 15 20 25 30

Loading period, in days

spec ific bas ic cr ee p , in µ∈µ∈µ∈µ∈ /M Pa

Saturated (first state)

Dried at 42% r.h. then resaturated D-dried then resaturated (second state)

Figure 8. Specific basic creep of cement paste (w/c=0.5) after resaturation from different drying treatments

REFERENCES

[1] F. H. Wittmann, Influence of Moisture Content on the Creep of Hardened Cement, Reol. Acta. 9 (2) (1970) 282-87.

[2] V.S. Ramachandran, R.F. Feldman and J.J. Beaudoin, Concrete Science, Heyden & son, London, 1981, p427.

[3] T. C. Powers, Mechanisms of Shrinkage and Reversible Creep of Hardened Portland Cement Paste, International Conference On the Structure of Concrete, Cement and Concrete Association, London, England, 1968, pp. 319-44.

[4] A. M. Neville, Creep of Concrete: Plain, Reinforced and Prestressed, Chapters 10 and 11 (Mechanisms of Creep and Creep hypotheses), North Holland Publishing Co, Amsterdam, 1970, pp. 258-309. 0 40 80 120 160 200 240 280 320 360 400 0 10 20 30 40 50 60 70 80 90 100

Loading period, in days

s p ec ifi c creep , in µ∈µ∈µ∈µ∈ /M P a D-dried

Dried at 37°C for 24 hours

Meth. then dried at 37°C for 24 hours Isop. then dried at 37°C for 24 hours 0 45 90 135 180 0 1 2 3

Figure 9. Specific basic creep of cement paste (w/c=0.5) after different drying treatments

[5] W. Ruetz, A Hypothesis for the Creep of Hardened Cement Paste and the Influence of Simultaneous Shrinkage, International Conference On the Structure of Concrete, Cement and Concrete Association, London, England, 1968, pp. 365-87.

[6] Z. P. Bazant, A. B. Hauggard, S. Baweja, F. J. Ulm, Microprestress-Solidification Theory for Concrete Creep I: Aging and Drying Effects, Journal of Engineering Mechanics, 123 (11) (1997) 1188-94.

[7] A. Bentur, R. L. Berger, Jr F. V. Lawrence, N. B. Milestone, S. Mindess, J. F. Young, Creep and Drying Shrinkage of Calcium Silicates Pastes, III. A hypothesis of Irreversible Strains, Cem Concr Res 9 (1) (1979) 83-96.

[8] R. F. Feldman, Mechanism of Creep of Hydrated Portland Cement Paste, Cem Concr Res 2 (5) (1972) 509-20.

[9] O. Ishai, The Time-Dependent Deformational Behaviour of Cement Paste, Mortar and Concrete, International Conference On the Structure of Concrete, Cement and Concrete Association, London, England, 1968, pp. 345-64.

[10] B. Tamtsia and J.J. Beaudoin, Basic Creep of Hardened Portland Cement Paste: A Re-Examination of the role of water, Cement and Concrete Research, in press. [11] P. Gu, P. Xie and J.J. Beaudoin, Impedance Characterization of Microcracking

Behavior in Fiber-Reinforced Cement Composites, Cement and Concrete Composites, 15 (3), 1993, pp. 173-180.

[12] T. Pajkassy and L. Nyikos, Impedance of Fractal Blocking Electrodes, Journal of Electrochemical Society: Electrochemical Sci. and Tech., 133 (10), 1986, pp. 2061-2064.

[13] F. J. Ulm, F. Le Maou, C. Boulay, Creep and Shrinkage Coupling: New Review of Some Evidence, Revue Francaise de Génie Civil, 3 (3-4) (1999) 21-37.

[14] J. J. Beaudoin, Validity of using methanol for studying the microstructure of cement paste, Materials and structures, 20 (115) (1987) 27-31.