Feasibility of using clays from Southeast European deposits in limestone calcined clay cements

. Recently, the cement industry has faced new challenges in addition to the environmental constraints of the last decade. The decline in availability and current inconsistent prices of common supplementary cementitious materials (SCMs), such as by-products from the iron industry or coal-fired power plants, have opened the search for more reliable materials. Research on cements containing calcined clays now serve as a possible solution to the forementioned problems. Clays containing the mineral kaolinite in sufficient quantities, when calcined and mixed with limestone powder, produce hydration products that can improve the strength and durability of concrete. In addition, the production of limestone calcined clay is reported to be less CO2 intensive, but this eco-efficient solution is viable only if the materials are locally available. For this reason, this study investigates the possibility of using natural clays from the Southeast European region (SEE) as cement replacements. A systematic experimental study was conducted on 18 different clays from 13 different deposits to determine the physical and mineralogical composition of the raw clays, their reactivity and mortar strength. The results were then related to the environmental contributions they might have in comparison with ordinary Portland cement in concrete.


Introduction
Supplementary cementitious materials are siliceous and aluminous materials that serve as cement replacements in concrete.Most commonly used SCMs are by-products from industry such as steel slag or fly ash from burning coal.In western EU countries majority of those resources is being rapidly depleted as we move to more ecological energy resources and increase the recycling of steel [1].While majority of CO2 emissions come from the production of clinker, the use of alternative SCMs such as calcined clays serve as most favourable solutions in reaching carbon neutrality by 2050 [2].
The reactivity of clays in cements is led by their possibility of amorphization, i.e. the dehydroxylation of the clay minerals which makes them pozzolanic.Kaolinite is the most advantages clay mineral with a 1:1 silica and alumina layer formation.This structure allows a more easy decomposition of the hydroxyl groups than in the case of a 2:1 layer minerals such as illite or montmorillonite [3].In nature, clays are found in different forms and with different mineralogical compositions, mostly containing mineral illite, kaolinite and montmorillonite but also other impurities such as quartz, feldspars, mica, calcite and others [4].Even though kaolinite requires the lowest calcination temperature (500-800°C), other minerals can exhibit pozzolanic properties with a suitable temperature [3].Studies have shown that even a medium range kaolinite content of about 50% can obtain comparable compressive strength results to OPC mixes after only 7 days with a clinker substitution of 45% (30% calcined clay and 15% limestone) [5].It is outlined in the study that calcined clays can subitise Portland cement with a major cost and energy reduction during production due to the lower calcination temperature.
This study is a part of a project aimed to investigate suitable ecological and low-cost calcined clay cementitious materials, containing local low grade kaolinite clays for the development of alternative binders.By screening regionally available clay materials for the use as SCMs a more comprehensive conclusion can be drown out regarding their feasibility.It is also crucial to find a relationship between properties of different natural clays and its performance in concrete, not only for understanding the interaction with cement, but also to address the durability and strength development.

Materials and methodology
This study evaluates 18 different natural clay samples taken from 13 different deposits.All materials originate from the Southeast Europe region and were collected as a part of the "Advanced low CO2 cementitious materials, ACT" project.Before characterisation and testing all materials were dried in an oven for 24 ± 2 h at 60 ± 5°C and then grinded in a disc mill for 30 seconds.

Characterisation methods
The chemical composition of the selected materials was determined by X-ray fluorescence (XRF) (Fig. 3).The density of particles after drying was measured by Le Chatelier flask method according the ASTM C188-17 standard [6].The particle size distribution (PSD) was was determined using the Mastersizer 2000 instrument with a wet laser diffraction procedure by dispersing the particles in different solvents depending on the material type.Samples of 50 ± 5 mg of materials were heated from 30 °C to 1000 °C, with a constant heating rate of 10 °C and a nitrogen flow of 30 mL/min.The resulting DTG curves were used to calculate the kaolinite content as the primary indicator of pozzolanic properties of clays.
The amount of kaolinite was calculated according to the weight loss between approximately 350 and 650°C (WLdehydroxylation), using the molecular masses of kaolinite [7], according the "Eq (1)".The weight loss for this calculation was obtained with the tangent method.
Error! Reference source not found.
X-ray diffraction (XRD) was used for quantification of amorphous content using the Philips X'Pert Pro (Malvern Panalytical, Malvern, UK), θ-θ configuration, wavelength CuKα1,α2.Powder patterns were recorded between 5 and 70° 2θ range, with 0.002° 2θ step size and a sampling time per step of 30 s. X'Pert HighScore Plus software (v3.0e,Malvern Panalytical, Malvern, UK) was applied for mineral phase identification and Rietveld refinement.For amorphous quantification high crystalline rutile was used as external standard method.

Pozzolanic reactivity
The R 3 test was caried out to assess the possible pozzolanic properties of collected material.This test, developed for the RILEM TC-267 committee, was caried out using isothermal calorimetry, obtaining the total 7-day heat release at 40°C of pastes composed of the SCMs.Prior to the test, all the SCMs and reagents were weighted, mixed and held at 40 ± 2 °C for 24 h.The formulation of the solid blends relies on the ratio of Ca(OH)2/SCM and CaCO3/SCM of 3 and ½ respectively.In addition, an alkali solution 3M of K were prepared with KOH and K2SO4.

Fig. 5. Reactivity test (R3) results for clay samples
The pastes were prepared in a high shear mixer at 1600 ± 50 rpm for 2 min until get a homogenous paste.Immediately, they were cast in glass vial and placed into the isothermal calorimeter

Mortar strength
Prior mixing of mortars all clays have been calcinated in a furnace at 800°C for 1 h.In this study, 20 mortar mixtures have been considered, in which the water to binder ratio and the content of total binder was 0.5 and 450 kg/m 3 , respectively.Portland cement CEM 1 45.2 R was used for the reference sample (OPC) and the mixes with SCMs.Binary blended mortars consisted of 70% of OPC + 30% of calcined clays and one mixture with 30% of quartz powder.Standardized sand was used as aggregate and polycarboxylate (PCE) was added to obtain adequate workability (up to 1% on weight of binder).For each mixture, a volume of 1,5 litre was mixed following the norm EN 196-1 (2016) [8], after which samples of 40 x 40 x 160 mm were cast.The Samples were covered with a plastic wrap after casting and held 24 h in laboratory conditions.When demoulded they were cured in humidity chamber (RH 95%) until testing time.Compressive strength test was carried out on 2 prisms after 2, 7 and 28 days of curing age, according to the norm mentioned above.

Life cycle assessment
The LCA was performed according to the ISO 14044:2016 standard [9] to compare the environmental impact of two different cement productions.Label CEM I indicate a mix of cement clinker with 6% of gypsum while the LCC shows a mixture of limestone calcined clay cement contains 50% clinker, 30% calcined clay, 15% limestone and 5% gypsum.The binders were modelled with a cradle-to gate approach (Fig. 6), taking into account the material acquisition, preparation and transport.The inventory data for the cement productions were taken from the Ecoinvent database, obtaining the realistic distances from raw material acquisition to the cement plant ( Fig. 1).

Fig. 6. System boundary for CEM I and LCC production
The inventory input/output data is shown in the Table 1.The functional unit (FU) chosen for this study was one kg of cement binder that is assumed to be used for standard concrete strength production.For the Life Cycle Impact assessment characterisation factors were adopted from the CML database developed by the Centrum voor Milieuwetenschappen Leiden [10].(EP), global warming potential and photochemical oxidation potential (POP).choice impact categories is based on a study that indicates categories mostly affected by the production of blended cements [11].This includes raw material excavation, use of fossil fuels, transport, electricity, and the atmospheric emissions from the cement kiln.

Results and discussion
The chemical composition of clays differs from the OPC, having a much higher silica and alumina amount.The biggest difference between clays is the alumina/silica ratio and the amount of iron oxide.The particle size distribution of clays is somewhat similar to the Portland cement, having a higher amount of smaller particles in the range of 0,1 to 1.The outputs of the TG measurement indicate the kaolinite content of each sample (Table 2).The range goes from 9.7 to as high as 43.1 %, meaning the clays from this region can be categorised as low and medium range kaolinitic clays.The DTG curve also shows peaks around the temperature of 740°C which is a typical peak that describes the decomposition of limestone (CaCO3) [13].Clays NC1 and K3 have a distinct peak in that area.The XRD quantification showed the clays from this area consist mostly of quartz, muscovite and a combination of other clay minerals such as kaolinite, illite and in some cases zeolite.Other impurities can be found and the amorphous content of raw clays varies System boundary: Cradle -to -gate from 1 to even 47% (TOP_1).Interestingly, the samples containing higher amount of kaolinite have also a higher amount of these smaller particles.Density of all particles was in the range from 2,1 to 2,6 g/cm 3 .The pozzolanic reactivity test (R 3 ) corelates with the kaolinite content of the samples, pointing out the differences between low and medium kaolinite clays.Fig. 5 presents the results of this study, where samples K1 and ZOR clearly show the highest reactivity with a heat release of almost 400 J/g of SCM after 7 days which corresponds to the higher kaolinite content.Medium range clay of this study, D-5, shows medium reactivity while the rest of the clays result with a heat release between 150 and 250 J/g of SCM.
The relative compressive strength results are shown in Error!Reference source not found.showing 2day, 7-day and 28-day compressive strength compared to the OPC reference.All mixtures show pozzolanic properties when compared to the "inert" quartz mix.Mixes containing clays with a higher kaolinite content show a faster strength development with the highest relative strength after 7 days.Surprisingly, the highest relative strength after 28 days is obtained with the clay MAR, a middle range kaolinitic clay, followed by the two clays with the highest kaolin content -ZOR and K1.It is observed that the clays with the higher kaolinite content obtain a higher development of early strength.The life cycle assessment clearly shows a reduction of all impact categories when comparing CEM I and the LCC mix.

Conclusion
This study focuses on the characterisation of different regionally available clay for the use as suitable SCMs.The correlation between composition of the samples to the reactivity and compressive strength provides a better understanding of the material and the feasibility of their usage.The key findings of the study are:  There are significant differences in the composition of clays, mainly in the amounts of alumina, silica and iron oxides, and the kaolinite content. Clay samples with the highest amount of kaolinite (around 40%) show smaller amount of iron oxides and a higher amount of smaller particles (<1 µm) after the same treatment as others. All clays show pozzolanic reactivity when compared to the inert quartz powder mix, outperforming in compressive strength of mortars test. Clays with higher amount of kaolinite exhibit higher reactivity within the R3 test which correlates to the fastest contribution to early strength (2-and 7-day strength). One clay sample (MAR) reached the compressive strength of OPC after 28 days even though the kaolinite content of 17% and the R3 test do not support this result.It could be that the higher illite content of this clay also contributes to the pozzolanic reaction after calcination at 850°C [14]. The LCA study showed a significant reduction of environmental impact when implementing 50% of cement replacement with a combination of limestone calcined clay cement, lowering the GWP by 40% and all other indicators by more that 30%.
Exception is the abiotic depletion potential which has the slightest reduction due to clay excavation.
Further studies should be conducted to investigate which other mechanisms play a role in the compressive strength development of low-grade kaolinite clays.There is also a need to of the most viable calcined clays from this study to the formulation of hydration products.This should be expanded eventually to assess the durability benefits of such clay in concrete mixtures.

Fig. 1 .
Fig. 1.Map of the clay excavation points and distances from one cement plant Characterisation of raw materials was performed in the Laboratory of Construction Materials at EPFL while the mortar properties were tested on the Faculty of Civil Engineering, University of Zagreb.

Fig. 2 .
Fig. 2. Particle size distribution of dry and grinded clay samples, quartz and Portland cement determined by The solvent was chosen according to the recommendations from Scrivener et.al.(2016) [7], 0.01wt.%polyacrylic acid for clay testing, isopropanol for cement and water for quartz.Results of the PSD are shown in the Fig. 2.The mineralogical composition was studied using thermogravimetric analysis (TGA) and X-ray diffraction (XRD).TGA was performed with TGA/SDTA 851 apparatus (Mettler Toledo, Columbus, OH, USA).

Fig. 3 .
Fig. 3.Chemical composition of clay samples compared to the Portland cement

Fig. 4 .
Fig. 4. Mass loss and DTG curve for each clay sample

Fig. 7 .Fig. 8 .
Fig. 7. LCA of CEM I and LCC cement The highest reduction of environmental impact related to the Global warming potential (GWP) where calcined clay and limestone combination shows a decrease of almost 40%.The smallest reduction in abiotic category which mainly by the extraction natural resources such as

Table 1 .
Inventory data for the LCA of CEM I and LCC

Table 2 .
Characterisation results of density measurement, TGA and XRD