POSSIBILITIES OF USE OF ISOTHERMAL AND ISOPERIBOLIC CALORIMETRY TO STUDY THE EFFECT OF ZINC ON HYDRATION OF CEMENT BLENDED WITH FLY ASH

: Increasing utilization of secondary raw materials and alternative fuels results in increasing contents of metals in cements. One of elements, the content of which keeps rising in cement is zinc. It comes to cement with secondary raw materials such as slag or fly ash or by the utilization of used tires as an alternative fuel. Zinc ions significantly prolong the hydration process in cement. This work deals with the influence of zinc ions in the form of very poorly soluble ZnO salt and easily soluble ZnCl2 and Zn(NO)3 on the hydration of cement blended with fly ash. Zinc was dosed in the range of 0.05, 0.1, 0.5 a 1% of cement weight. Final products were next analyzed using X-Ray Diffraction.

also cement alternatives such as metalurgy slags or fly ash [6]. Zinc is known mainly for the retardation of cement hydration, for the change of structure and for strong decrease of initial strength [7].
The most common theory describing the mechanism of setting retardation by zinc ions suggests the formation of layer of poorly soluble hydroxides on the surface of hydrated grains, which forms a physical barrier for water transport towards clinker minerals. According to recent knowledge the main reason for the setting retardation is the conversion of hydroxide to zincate. This reaction leads to the consumtion of Ca 2+ and OHions from the pore solution which consequently inhibits the formation of C-S-H gel [2] and thus the hydration of silicate phases up to total hydrolysis and the conversion of clinker is inhibited. The reactions leading to the formation of zincate are given in Eqs.
1 -3 [5,8]. Another introduced theory is the poison effect. According to this theory, the retardation by zinc is caused by the poison effect, by which the nucleation of hydration products is delayed or blocked.
The poison improves the supersaturation of Ca(OH)2 and C-S-H in pore solution. After the poison effect is overcome, faster nucleation and growth occurs, compared to pure cement paste [9]. Odler and Schmidt [10] studied the influence of zinc on the formation of clinker minerals and the distrubution of zinc in between the portland cement phases. They discovered, that the free lime content decreases with increasing temperature and burn-out time as well as with increasing dosage of zinc oxide. Zinc cationts catalyze the reactions of clinker minerals formation. Increasing content of zinc changes also the phase distribution of clinker. The contents of alite (C3S) and interstitial phases (C3A and C4AF) increase at the expense of belite (C2S). The authors found two different reasons for this phenomenon. The first is the substitution of calcium ions by zinc ions and hence the release of more calcium ions for further reaction with silicates and the second one is increased content of C4AF at the expense of C3A.
Gineys et al. [5] in their work discovered, that zinc reacts preferentially with aluminates and affects the stability of C3A phase forming a new compound Ca6Zn3Al2O12. The limit of zinc content was described as the concentration at which the content of C3A phase is decreased and replaced by Ca6Zn3Al2O12 phase. The reactivity of C-Z-A phases was tested by Arceo H. a Glasser [11]. They prepared pure Ca6Zn3Al2O12 phase. Based on the results it can be assumed that despite the Ca6Zn3Al2O12 phase is reactive, zinc ions are blocked in insouble hydrates.
A similar experiment was performed by Gineys et al. [6]. They found out that insoluble hydrates are formed on the surface of zinc phase and hence there are only limited reactions of zinc with C-S-H gel. One of given explanations is non-optimized gypsum content, which was based on the content and reactivity of C3A phase, the amount of which, however, is not detectable in this sample.
Li et al. [12] studied the effect of zinc chloride and sulfate on the hydration and final compressive strengths of HPC. They found out that the setting retardation is stronger with chlorides than with sulfates.
Šiler et al. [13] dealt with the effect of zinc in the form of ZnO, Zn(NO3)2 and ZnCl2 on cement pastes and in [14] they doped cement with slag in the same experiments. They discovered that ZnO is the strongest retarder due to low solubility. In the cement paste the hydration was accelerated when 1 % ZnCl2 was added, but when combined with slag, the acceleration did not occur. Next, the synergic effect was observed, which retarded the hydration when zinc was combined with slag.
Fly ash (FA) is the product of coal combustion in thermoelectric plants [15]. Term fly ash was accepted as it is transported from combustion chamber by exhaust fuels. Fly ash can be marked as cementatious or pozzolanic [16,17]. The cementatious fly ash sets when mixed with water. The pozzolanic fly ash hardens when activated by alkaline substance prior to mixing it with water. The alkaline substance can be e.g. quick lime, commonly present in cementitious mixtures. These features predetermine fly ash to be used as a cement replacement in concrete and many other building applications. Fly ash contains many heavy metals, especially zinc, lead and copper [18]. Primary fly ash can contain 0.8 wt.% Zn [19]. Secondary fly ash which is formed by melting the primary fly ash can contain 2.2 % to 20.7 % Zn [18]. Fly ash is one of possible zinc donors for cementitious materials.

Materials and sample preparations
The mixtures were prepared from the cement CEM I 42.5 R Mokrá -Českomoravský cement, a.s., Heidelberg Cement, Czech Republic (x10 = 0,47 µm, x50 = 8,89 µm, x90 = 34,42 µm). Fly ash was from ČEZ Energetické produkty, s.r.o., Czech Republic (x10 = 8,36 µm, x50 = 112,57 µm, x90 = 298,37 µm). The particle size was determined by laser-diffraction method. The chemical and phase compositions of the components used are shown in Tables 1 and 2. Zinc was added in the form of poorly soluble ZnO and well soluble compounds Zn(NO3)2·6H2O and ZnCl2. The amount added was in the range from 0.05 to 1 wt. % of the cement substitution (the percentages of the replacement were always calculated to pure zinc amount in binder). The pastes were mixed with distilled water with water to binder ratio equal to 0.4.
A 15 wt % FA replacement of cement was chosen to study the effect of FA on the hydration of cement doped with zinc.
For the measurements in isothermal calorimeter the amount of 7g of mixture was dosed into glass ampoule and placed in the device, where it was immediately mixed. For the measurements in isoperibolic calorimeter the amount of 300 g of pre-mixed mixture was placed in styrofoam cup provided with a thermo-insulating jacket and a thermocouple and this mixture had been monitored until the main reaction of silicates was finished and the mixture temperature reached the ambient temperature.
Due to the sensitivity of measurement to X-ray diffraction analysis (XRD) the measurement was performed with samples containing 5 wt. % of zinc. Right after mixing the samples were placed in styrofoam cups and stored in a humid environment until the time of measurement. Next the samples were ground in a vibrating mill and the hydration was stopped by rinsing the sample with acetone. The samples were dried at the temperature of 50°C to remove residual acetone.
For the monitoring of zinc effect on the hydration the methods of isothermal and isoperibolic calorimetry were used. The main differences in results are caused mainly by the measurement conditions since in the case of isothermal calorimetry the measurement is performed under exactly defined conditions, while the isoperibolic measurements proceed under real conditions. The combination of those two methods brings an important information about the course of particular reactions (isothermal calorimetry) and the comparison with the process in real environment (isoperibolic calorimetry) [20].
The prediction and control of concrete temperature rise due to cement hydration is important for mass concrete structures since large temperature gradients between the surface and the core of the structure can lead to cracking, thus reducing durability of the structure [21].

Isothernal calorimetry
This method is often used for the monitoring of hydration of hydraulic binders, pozzolan materials and latent hydraulic properties [22,23]. The measurement proceeds at exactly set temperature inside the device together with reference sample (in this case the quartz sand).
Due to exactly defined conditions this method can be used for long lasting measurements. By the change of temperature the reaction rate can be changed and hence particular reactions can be studied in more detail, especialy during early stage of hydration where the heat rate is relatively high [13].
The measurement of hydration heat evolution was carried out using isothermal calorimeter TAM Air (TA Instruments). The measurement was based on ASTM C 1679. Quartz sand was used as a reference. The measurement was carried out at 25 °C. The introduction of the sample preparation is mentioned below.

Isoperibolic calorimetry
Isoperibolic calorimetry is a method which is often used for the monitoring of hydration of hydraulic binders, especialy cements [16,[24][25][26]. The principle is the measurement of temperature changes during hydration at constant ambient temperature. For the calculation of total released heat the subsequent numerical data integration is performed. Data required for this calculation was gained from the device calibration [20]. The hydration heat is evolved during the hydration and the sample heats up. During isoperibolic calorimetry the same heating occurs as during the hydration under real conditions. For the study of hydration reactions, it is therefore advantageous to use a combination of both these calorimetric methods [20].

X-ray diffraction analysis (XRD)
X-ray diffraction (XRD) is a non-destructive method used for testing the crystallic materials. X-ray diffractogram is a characteristics of investigated substance. It provides the information which can be used for the determination of quality and quantity of various crystalline phases [27,28]. The X-ray diffraction analysis was performed with X-Ray difractometer Empyrean (Panalytical) with Bragg-Brentano parafocusing geometry and using CuKa radiation. The measurements were done within the range from 5 to 120 °2Θ with angular step of 0.013 °2Θ and 25s duration using automatic divergence slits to maintain the constant irradiation of the sample area. The measurements were repeated four times and then summed.

Results and discussion
The addition of fly ash to cement brings about mainly further delay of hydration reactions due to the reaction of fly ash particles with Ca 2+ ions from cement. While the reactivity at lower temperature in isothermal calorimeter is lower also lower heat evolution occurs, which is due to the substitution of the amount of reactive amorphous phase [24]. Fly ash mostly contains rather non-reactive crystalline phase. The reaction of all these components is affected by the temperature of reaction environment.
Increasing concentration of zinc results in the increase of first peak of hydration curve [22]. This rise is caused mostly by the reactions of aluminate phases. This increase can be caused either by the effect of zinc, as smaller amount of ettringite is formed compared to the hydration of OPC (ordinary portland cement), or by high content of amorphous phase in fly ash (78 %), which can react e.g. by the pozzolanic reaction [16,[23][24][25].
When the concentration of zinc in the mixture is low the temperature is similar to that of reference  In samples with zinc concentration of 0.05 and 0.1 wt. % a small sulfate depletion peak was recorded on the hydration curve together with the main hydration peak [22] same as in reference samples in literature [13,14]. As the zinc content rises, the first peak rises too and the second peak decreases up it can be assumed that zinc does not significantly affect total evolved heat.

CEM I + ZnCl2
Isoperibolic calorimetry Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 October 2020 doi:10.20944/preprints202010.0233.v1 In samples containing 1 % of zinc in the form of ZnCl2 a noticeable effect on the acceleration of hydration caused by Clions was observed. The acceleration occurred under both, isothermal as well as real conditions, compared to pure cement [13], where this effect was observed only under real conditions during isoperibolic measurement -contrary to data introduced in literature [13], where no peak was found on the calorimetry curve for 1 wt % of zinc in the mixture, probably due to a low temperature in a calorimeter (25 °C). This effect was observed in pure cement, but not in samples with slag [14]. The reason is probably higher amount of alkali ions which make the course of hydration reactions in fly ash easier. Another reason is increased permeability and diffusivity of chloride anions which due to smaller ion radius than OHcan cause the increase of pressure gradient and further deterioration of layer of C-S-H gel [30,31]. In the chart presenting the measurement by isoperibolic calorimetry the rise of initial peak [22] with increasing zinc concentration can be  concentration. Due to this acceleration the sample with 1 % of ZnCl2 reaches the main hydration peak before the sample with 0.5 % of ZnCl2, contrary to Zn(NO3)2 in which, due to high inhibitory effect in the 1 wt% zinc sample, lower maximum temperature development and therefore slower hydration reactions occur (lower and wider peak). In samples with ZnCl2 lower values of heat flow maxima were observed, which is consistent with the results given in literature for pure cement [13] and for cement with slag [14]. The values of total released heat are similar for both soluble compounds; however for ZnCl2 the differences for various concentrations are very small.

CEM I + ZnO
Isoperibolic calorimetry The same as in pure cement [13] and in combination with slag [14] the strongest hydration retardation occurred in mixtures with ZnO, especially due to slow dissolution of this compound. This could be assigned to the longest induction periods in these samples.

The comparison of induction period length
The lengths of induction periods were determined from calorimetric measurements in the same way as in literature [13,14]. For the fitting of all curves the following equation was used: Since the t1 value was negative for all samples, we can rewrite this as equation 6. The results from isoperibolic measurement for the sample containing 1 % of Zn in the form of ZnCl2 was not included into the evaluation, since the acceleration of hydration, disalowing this way of evaluation, occured in large extent.
All curves created by fitting the recorded data with exponential function in figure 7 exhibit only minimal deviations from real data.
From these curves mainly similar behaviour of soluble compounds, positive effect of Clions on the hydration acceleration and unambiguously the strongest retardation ability of ZnO due to its low solubility can be read. Comparing these curves to the results from literature [14] for slag these curves are quite similar, for the combination with fly ash, however, the accelertion effect of Clions becomes more evident.

XRD results
The following chapter shows the results by XRD analysis. This analysis provides the information about the phase composition of crystalline materials. The evaluation was performed by Rietveld Analysis of XRD Patterns. The results are shown in Tables 3 -6.   When compared to literature [14] it was discovered that in the sample with only fly ash and no Zn content similar amounts of main cement phases were present, but also slightly higher amount of ettringite and higher amount of calcite after 28 days. In Zn(NO3)2 slightly lower amount of brownmillerite, but almost twice as high amount of ettringite were found. In the samples with ZnCl2 slightly lower amount of belite and higher amount of ettringite were found. For ZnO slightly higher amount of alite and brownmillerite, slightly higher amount of ettringite and lower content of portlandite detectable after 90 days just like for slag were found [14].

Conclusions
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 12 October 2020 doi:10.20944/preprints202010.0233.v1 The aim of this work was to study the effect of zinc in the form of soluble salts Zn(No3)2, ZnCl2 and very poorly soluble ZnO on the hydration of portland cement blended with 15% of fly ash.
15% addition of fly ash to cement causes mainly further retardation of hydration reactions due to the reactions of fly ash particles with Ca 2+ ions from cement. Due to lower reactivity at lower temperature in isothermal calorimeter lower heat evolution occurs which is caused by the substitution of very reactive cement grains by less reactive fly ash particles. Even the pozzolana reaction in later phase of hydration did not produce enough heat in any sample, to reach the same total released heat values as for the sample with pure cement. On the contrary, in real conditions of isoperibolic calorimeter, where the sample heats up, higher amounts of released heats were detected compared to pure cement, due to faster course of pozzolana reaction.
The strongest effect on the hydration retardation from all investigated compounds showed ZnO as it dissolves very slowly. On the contrary, for the dosage of 1% of zinc in the form of ZnCl2 significant acceleration of hydration occured.
The lengths of induction periods were assessed from detected calorimetric curves and from these lengths the curves were gained by fitting with the exponential function. These curves comply well with measured data. Again, significant prolongation when ZnO is added and the acceleration of hydration for the addition of ZnCl2 can be observed.
From the curves detected by isothermal calorimetry it can be seen, that for Zn(NO3)2·6H2O and ZnO also increased sulfate depletion peak and decreased main hydration peak occur. The addition of ZnCl2 does not increase the sulfate depletion peak above the main hydration peak.
Funding: This outcome has been achieved with the financial support by the project: GA19-16646S "The elimination of the negative impact of zinc in Portland cement by accelerating concrete admixtures", with financial support from the Czech science foundation.

Conflicts of Interest:
The authors declare no conflict of interest.