Use of ultrafine marble and brick particles as alternative raw materials for clinkerization

1 Introduction

Scientists have long been investigating alternative methods to control the cement quality for chemical reasons and to develop new methods to provide understanding of the chemical compositions of cement easily. The investigations focussed on such topics as hydraulic modulus (HM), silica modulus (SM), alumina-iron ratio (AR), lime saturation (LS or “free lime”) and lime saturation factor (LSF) as well as the chemical composition of cement.

Today, OPC clinker shows a range, respectively on the hydraulic modulus between 1.7 and 2.3, and of silica from 2.2 to 2.7, an alumina-iron ratio between 1.5 and 2.5, a free lime module of 0.90 to 0.95 and a variability of the lime saturation factor of 0.7 to 1, whilst higher LSF values indicate that free lime is to be present in the clinker.

Additionally, some standard limits have been noted by analysing the chemical composition of OPC showing MgO, which should not exceed 5 % and chlorine, which should be less than 0.20 %, while after preparing with water the pH should vary in a range of 12-13.

In order to monitor cement quality, in cement plants generally the most prevalent module features are LS and LSF, which display the optimum lime content in the cement. Moreover, the larger the LSF is, the better cement quality is with the view to the chemical composition. However, the results on silica module analysis show as larger the SM is the burnability deteriorates in the rotary kiln. In addition to these scientific realities, an LS more than 100 % will also exhibit a quick setting-time of the cement. Therefore, gypsum is added to cement clinker with an amount of 3-5 % to extend the setting-time according to state-of-the-art standards [1, 2, 3, 4, 5].

However, since more than 100 years of research have been conducted to foster the world’s natural mineral deposits rather than to investigate alternative materials in place of various cement raw materials.

These studies will show how such by-products like ore containing waste (OW), fly ash (FA), silica fume (SF), steel waste (SW), and ground granulated blast furnace slag (GGBFS) demonstrate its benefit on mineralisation and its pozzolanic effects. These present studies have shown that cement could be manufactured with by-products [6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18]. Although, these tested materials show a high content of main oxides required for clinker and cement manufacturing it is observed that they can increase some undesired chemical components in the cementitious system such as magnesium,  sulphates and alkali, which are inducing expansion, ettringite or alkali-silica reaction etc.

On the other hand, some by-products have chemical components very similar to cement raw materials. Marble powder (MP) and brick powder (BP) are well-known examples of by-products from a steady increase of an industrial output of approximately 2 592 000 t/year marble powder and 3 800 000 t/year brick powder all over the world [19]. And MP and BP have the same benefit as in powdered form. This implies that MP and BP can be easily ground to form ultrafine particles. Since re-use systems for MP and BP have not been established, they are stored on agricultural land. So, these priceless chemical compositions are lost without reuse. For this reason the study is designed to use the potential of ultrafine marble particles (UMP) and ultrafine brick particles (UBP) as raw materials in cement manufacturing. It does not address either the decrease of by-products volumes or the development of cement strength. Therefore, this paper presents new original experimental results including the chemical compositions, physical properties, pozzolanic activity and mineralogy by XRD, SEM and EDS, and new cements made by clinkering of UMP and UBP.

2 Experimental program

2.1 Materials

These two by-products used in the current experimental research are UMP and UBP sent for Tekmar marble factory in Bilecik and brick factory in Kaman as cement raw materials [19].

2.2 Cement raw materials preparation with UMP and UBP, burning process and cement production

Sample preparation consists of three stages. The first stage is the preparation of UMP and UBP in order to specify their characteristic properties. The mineral waste of marble and bricks are comminuted by ball mill at thirty minutes to form ultrafine particles in laboratory until Blaine is equal to a range between 5800 and 6300 cm²/g.

In the second stage the powders of UMP and of UBP respectively are formed to create wiry stick samples. A medium planetary mixer is used to blend every single meal using the following procedure:

1. add 79 % UMP and 21 % UBP for UMP-UBP-F1,

and 77 % UMP and 23 % UBP for UMP-UBP-F2,

and 81 % UMP and 19 % UBP for UMP-UBP-F3.

2. mix at a low speed to homogenize and

3. add 30 % water at low speed again until a plastic consistency is formed,

4. finally test specimens that are formed.

The last stage deals with the burning of cement clinker samples. The applied burning process is common for all meals. The meals are first formed into small spheres, with a diameter of 1 cm, and dried at 105 ± 5°C until constant mass. Subsequently they are placed in an electrical furnace at 1450°C for 30 min, using a temperature increase rate equal to 6°C/min, starting from 800°C. At the end of the burning processes, the clinker samples coded UMP-UBP-C1, UMP-UBP-C2, and UMP-UBP-C3 are rapidly cooled in air.

The work starts by manufacturing the UMP- and UBP-clinkers in a laboratory scale; determining the ratios and water in meal batches; generating fine wiry sticks of the meals and burning the fine wiry stick meals; creating UMP- and UBP-clinkers and grinding them down to cement.

2.3 Methods

This research is divided into three experimental methods to understand how the properties of UMP and UBP effect clinker raw materials (meals) and cements in various percentages but with the same fineness.

The three methods of experiments included:

1. Characteristics properties of UMP and UBP are determined through a set of relevant experiments such some properties as chemical, physical, pozzolanic and rheological to understand whether they could be used as raw materials for cement production or not.

2. Complementary chemical experiments are carried out on the meals, clinkers and cements prepared in the second and third stage to classify the meals, clinkers and cements more precisely and demonstrate the burnability of UMP and UBP.

3. With these chemical results the modules are calculated by means of the known equations in the literature such as Eqs.1-Eqs.5.

2.3.1 Chemical properties of UMP and UBP (wet analysis)

The TS EN 196-2 standard is known for wet chemical analysis and is followed to determine the chemical compositions of UMP and UBP. For each mineral by-product in each chemical experiment, three samples are analysed and the average value of six samples is taken to be the representative chemical properties [20].

2.3.2 Physical properties of UMP and UBP

The physical properties of UMP and UBP specimens such as fineness, density, specific surface and water absorption are carried out according to TS EN 196-6:2000, ASTM C188-09, TS EN 1097-6:2002. For each by-product in each experiment, three specimens are tested and the average value of twenty four specimens is taken to be representative for fineness, density, specific surface and water absorption [21, 22, 23].

2.3.3 Pozzolanic activity of UMP and UBP

The ASTM C593-95 standard is the specified pozzolanic properties of UMP and UBP. Mortars are prepared with tap water, the mix of lime and UMP (180 g and 360 g) or the mix of lime and UBP (180 g and 360 g), and CEN standard sand (1.480 g). For preparation of mortars a medium planetary mixer is used during 5 min as follows:

1. add the mix and CEN standard sand in a bowl,

2. mix them for 60 s,

3. add water until 65-75 % fluidity,

4. mix again for 180 s,

5. fill the specimens immediately into prism-moulds (40x40x160 mm) as three layers,

6. vibrate each layer 60 times,

7. cover the moulds air-tightly,

8. place them in damp controlled cure cabinet at 54 ± 2°C and 98 % relative humidity for a 7-day period,

9. at the end of the 7-day period, allow to acclimatise to room conditions,

10. submerge the specimens in water for 21 days,

11. remove them and allow draining and testing within 1 h after emerging.

For each by-product a flexural strength test is carried out by testing six prisms (40x40x160 mm) with an ELE hydraulic testing machine – the loading rate of the flexural strength test is 50 ± 10 (kPa/s) with one-point loading at 28 d. Compressive strengths are determined at cubic mortar samples which are pieces left from the prisms broken in the flexural test at 28 d. Twelve samples are tested out with an ELE hydraulic testing machine – the loading rate of the test for compressive strength is 50 ± 10 (kPa/s) and the average values of twelve samples are taken to be the representative compressive strength [24].

2.3.4 Mineralogy of UMP and UBP

XRD monitored minerals in the UMP and the UBP. Samples are comminuted to a particle size smaller than 45 µm. The 2 diffraction angle is scanning from 5° to 60° using a step size of 0.020°. SEM equipped with an energy dispersive spectroscope (EDS) is used to analyse the mineralogical habitus and its elementary compositions of UMP and UBP. All samples are spattered with gold. The accelerating voltage and current are 15-20 kV and 10-20 µA, respectively. SEM observations on UMP and UBP samples are performed on a LEO scanning electron microscope.

2.3.5 Chemical composition of meals and clinkers

The analyses are carried out regarding TS EN 196-2 standard to analyse the chemical compositions of each meal and clinker. Three specimens are analysed and the average value of eighteen samples are taken to be representative for the chemical composition [20].

2.3.6 Modulus calculations of meals and cements

The chemical compositions of each meal and of clinker are preconditions and put into the equations to calculate the module properties as follows [3].

HM —————————————————            (SiO₂ + Al₂O₃ + Fe₂O₃)       CaO  ⇥(1)

SM  ————————————        (Al₂O₃ + Fe₂O₃)     SiO₂⇥(2)

AR  ——————   Fe₂O₃   Al₂O₃ ⇥(3)

LS   ———————————————————————————————                       (2,8 · SiO₂) + (1,1 · Al₂O₃) + (0,7 · Fe₂O₃)                    100 · (CaO + 0,75 · MgO)  ⇥(4)

LSF ————————————————————————————————                          (2,8 · SiO₂) + (1,2 · Al₂O₃) + (0,65 · Fe₂O₃)                  (CaO – 0,7 · SO₃)⇥(5)

3 Results and discussion

3.1 Chemical composition of UMP and UBP

The chemical data of UMP and of UBP, which are used to prepare clinker, are given in Table 1. It is observed that UMP and UBP samples did not have the same oxide compositions as natural raw materials. The sum of CaO, SiO₂, Al₂O₃ and Fe₂O₃ of UMP is larger than 54 %, whilst UBP consisted of 82 % in the sum of CaO, SiO₂, Al₂O₃ and Fe₂O₃ (Table1) [19]. UMP shows the highest content of calcium oxide and the lowest of the oxides of aluminium, silica, ferrite, sulphur, magnesium and sodium. Contrary to the last results of UMP, the oxides of silica, aluminium and ferrite in UBP are the largest in these mineral by-products.

In UBP the calcium oxide content is the lowest and can be denominated unnatural pozzolana according to EN206-1 and ASTM C593-95. Regarding TS EN 197-1 standard minor oxides of MgO, SO₃ and others in UMP and UBP are not exceeding limited values for use as a cement clinker raw material [24, 25, 26].

However, the loss of ignition in UMP is based on several reasons. One of the reasons is the high CaCO₃ content and the other one, UMP is not exposed to any heat treatment. On the contrary UBP has no more confinement water from inside of the clay particles, which is now evaporated at 750°C in the brick furnace. These infer that no noxious impacts will be expected on cement manufacturing with UMP and UBP [3].

3.2 Physical properties of UMB and UBP

Table 2 shows the physical properties of UMP and UBP. To be able to evaluate the effect of fineness of UMP and UBP on cement raw material reactivity and burnability, grain size distributions are carried out with sieves for all raw materials. UMP is consistently smaller than 40 µm, while UBP is smaller than 90 µm. UMP shows a higher fineness than UBP after grinding for 30 min. The fineness results imply an easy burnability as well as pozzolanic reactivity. The bulk density of UMP and UBP are to be found in a relatively narrow range of 2.1-2.5 gr/cm³ because they are ground in the same ball mill [19].

UMP shows the highest bulk density of the by-products. Additionally, the specific surface of UMP is larger with respect to UBP. And each of them had larger specific surfaces than of OPC. The water absorption ratio of UMP and UBP are found in a relatively wide range of 16-41 %, whilst UMP with its higher fineness shows a lower water absorption ratio, the highest density and the largest specific surface than UBP [19].

3.3 Pozzolanic activity of UMP and UBP

The pozzolanic activity of UMP and UBP are achieved by determination of the compressive strength. In Table 3 the results are shown after 28 days of hydration the pozzolanic reaction of UMP and UBP is almost completed. The compressive strength of UMP and UBP at 28 d is in a relatively wide range of 1-18 MPa and of flexural strength between 1-4 MPa, respectively. The compressive strength of UMP is lower in comparison with UBP. With the view on flexural strengths UBP is about 3.5 times stronger than that UMP. A similar situation to flexural strength between UMP and UBP is to be found in compressive strength, too. The compressive strength of UBP is approximately 17 times higher than of UMP. From the point of compressive strength development UBP has the biggest pozzolanic effect on cement hydration [1, 17, 19].

The difference between the hydraulic activity of UMP and UBP can be found in the variation of their crystalline structures: the bonding between Ca-ions and CO₃-molecules on the UMP surface inhibits the pozzolanic reaction. However, a prior step of breaking the Si-O bonds on UBP surface, which is relatively quick, makes the reaction faster. When assuming UMP and UBP takes part in the pozzolanic reaction, a CaO-SiO₂ ratio between 2.29 and 3.5 can be calculated [19].

3.4 Mineralogy of UMP and UBP

The X-ray diffraction of UMP is presented in Figure 3. It is observed that UMP and UBP content completely different minerals and consequently different elements and oxides. UMP is containing mainly calcite (CaCO₃) and some other minor minerals are also found for instance quartz, dolomite, hematite and magnetite.

XRD-pattern in Figure 4 show UBP which consists mainly of alumosilicates, feldspars, kaolinite, magnesium orthosilicate, illite and magnesium silicate. There are also some minor minerals to be found in UBP which are montmorillonite, antigorite, calcite, olivine, hematite and ilmenite. The X-ray diffraction pattern of UBP is similar to the four main phases of cement clinker C₃S, C₂S, C₃A and C₄AF [19].

The SEM figures of the micro structure of UMP and UBP and their EDS element scan are shown in Figure 1a, Figure 1b, Figure 2a and Figure 2b [19].

Most of the UMP consists of nearly spherical particles with a quite smooth surface. Obviously, UMP is composed of various sized particles ranging up to several dozens of microns (Figure 1a). Additionally, UMP’s micro structure indicates the same crystallinity and is slightly less hollow. Due to its chemical composition Ca, C and O element peaks are detected by EDS (Figure 1b) [19].

With the view on XRD, SEM and EDS the UMP is composed of a medium pure chalk with calcium oxide, and contains low concentration of magnesium, sodium, potassium and sulphur oxide as defined as minor oxide for the clinker burning process and additionally, UMP may not cause harmful chemical damages such as ettringite or alkali-silica reaction [19].

SEM observations on UBP indicated a different habitus in a large variation of grain sizes of annular and flat particles, which might be amorphous as well (Figure 2a). Grains of UBP exhibit irregular shapes with flat surfaces covered by small debris which are similar to ground cement as in mortar. UBP samples are covered with immature needle-like products. Obviously, the amount of these pile products is significantly higher on UBP than on UMP and the products are more compact [19].

Noticeable differences in the UBP mineralogy are observed than in UMP. Contrary to UMP observations, UBP includes plenty of rich silicate, aluminate, calcite and hematite minerals which reduces detrimental effects on cement and provide for a much more stable chemical composition. There are mainly Ca, Si, Fe, S, Al element peaks in EDS to be observed on UBP which underlines the similarity to the chemical cement composition (Figure 2b) [19].

3.5 Chemical and module properties of meals

Table 4 presents the chemical compositions of UMP-UBP-F1, UMP-UBP-F2 and UMP-UBP-F3. Additionally, modules of the test meals are given in Table 5.

On the one hand, if UMP composed 77 mass-% of meal, the average of CaO in UMP-UBP-F1 is >2 % and larger of UMP-UBP-F2 and >2 % less with respect to UMP-UBP-F3.

However, on the other hand the SiO₂- and Al₂O₃-based recipes of UBP require the increase of UMP up to 81 % in the clinker raw meal to decrease SiO₂, Al₂O₃, Fe₂O₃, as well as MgO, SO₃, K₂O and Na₂O.

When the raw meal containing 81 mass-% UMP, the hydraulic module in UMP-UBP-F3 is more than 3 % and 13 % larger than in UMP-UBP-F1 respectively in UMP-UBP-F2 as well as the other modules are varying.

The increase of UMP up to 81 % in the raw meal caused the increase of the other modules in the meals recipe. The increasing of the module will cause to exhibit easily strength development of the cement produced with UMP and UBP. The more important inference than last is that the meals are being burned easier. So, they do not require more energy than making OPC clinker. The last inference is that lime content of cements could be optimized, and modified in order to enhance cement quality effectively (Table 5).

3.6 Chemical and module properties of cements

Table 6 presents chemical compositions of UMP-UBP-C1, UMP-UBP-C2, and UMP-UBP-C3 cements. As seen in Table 6, new cements were shown to be appropriate to the standard limits required for OPC.

In view of the chemical conversion from meal to clinker at 1.450°C, the chemical composition in UMP-UBP-C1 increased in relatively wide ranges of CaO from 41-51 %; 9-16 % of SiO₂; 1.9-8.75 % in Al₂O₃; 1.3-11.2 % Fe₂O₃; 3-3.85 % MgO; 1-4.9 % SO₃; 0.4-0.9 % K₂O and 0.9-1 % Na₂O.

Contrary to the increasing oxides, LOI decreased in UMP-UBP-C1 in a relatively wide range from 4 to 41 %. The oxidation of metallic iron to Fe₂O₃ in UMP-UBP-C1 originates the largest increase, which is more than 840 %, whilst the lowest increase is to be observed in K₂O in the chemical compositions (Table 4, Table 6). The increasing of UBP up to 23 % in UMP-UBP-C1 may expect the highest SiO₂ content in the cements. The results imply that the cement clinkers could be burned at 1.450°C at 30 min in laboratory scale, effectively and easily (Table 6).

Due to the chemical compounds the modules are calculated and show some variations according to BOGUE results, UMP-UBP-C1, UMP-UBP-C2, and UMP-UBP-C3 are in the celite cement class. Negative results on chemical compounds imply that further research is also required to calculate the chemical composition of new cement. However, the chemical composition results infer an increasing of UMP up to 81 % in cement caused the increasing of LoI, CaO and K₂O, and the decrease of SiO₂, Al₂O₃, Fe₂O₃, MgO, SO₃, and Na₂O (Table 6).

The cement clinker module of UMP-UBP-C1, UMP-UBP-C2, and UMP-UBP-C3 are presented in Table 7. Despite the oxides of the silica, aluminum and ferrite of UMP-UBP-C1, UMP-UBP-C2 and UMP-UBP-C3 show relatively wide ranges the resulting HM, SM, and AR module are in a narrow range of 1.4-1.7 and 0.7-0.8, respectively 0.7-0.9.

3.7 Burnability of cement clinker raw materials

The reactivity of the meals is evaluated on the basis of the LoI after sintering at 1450°C (Table 4, Table 6).

Burnability Equation of Clinker (BEC) is suggested to calculate burnability of the cement clinker raw materials in Eq. 6 following:

BEC_Lol = 1.2891x²_Lol - 102.95x_Lol + 2058.5

(r² = 1)⇥(6)

BEC_LoI and X_LoI are the arithmetic averages of clinker LoI and respectively of meal LoI.

The variables of the equation are UMP and UBP. Invariable includes the manufacturing data at 1450 °C and 30 min retention time. The relationship between cement clinker LoI and raw meal LoI are shown as burnability level and is r² in Eq. 6.

4 Conclusions

The research is carried out for understand the effects on manufacturing new cements with UMP and UBP. The chemical compositions, physical properties, pozzolanic activity, mineralogy and modules are determined by international standards and newest knowledge from the literature. The results of these experiments can be summarized as follows:

1. Chemical composition and physical properties:

The content of the main elements is in a range of natural raw materials. Minor oxides (MgO, SO₃ etc.) of UMP and UBP fall below the standard limited values for using them as cement raw materials. The mineral by-product UMP is consistently finer than 40 µm, while UBP is smaller than 90 µm. The fine grain size provides easy burnability. UMP has the highest density of the mineral by-products. And the specific surface of UMP is larger with respect to its of UBP, and both of them show larger specific surface than ground OPC. Additionally, the water absorption of UBP and UMP has been tested, which has the lowest ratio of both by-products. 

2. Compressive strength development:

UMP has no active latent hydraulic properties, while UBP shows hydraulic properties and provides the strength development in cement. UBP has approximately more than 17 times a higher pozzolanic effect than UMP. The reason is the chemical participation on the reaction. UMP only shows a physical filler effect by reducing the water demand and making a small contributing to increasing strength development, while UBP really took place at the pozzolanic reaction.

3. XRD, SEM and EDS analysis:

UMP – as ultrafine marble – mainly consists of calcite (CaCO₃) and some other inclusive minerals like quartz, dolomite, hematite and magnetite as found in the X-ray diffraction pattern. Additionally, UMP consists of medium pure calcium oxide.

UBP – as a former clay – mainly contains silicate, aluminate, feldspars, kaolinite, magnesium orthosilicate, illite and magnesium silicate, and it has some minor minerals such as montmorillonite, antigorite, calcite, olivine, gypsum, hematite and ilmenite.

UMP consists of near-spherical particles and the spheres are quite smooth, whilst UBP turned out big and little annular and flat particle according to SEM images. The element peaks of EDS confirm Ca, C and O in UMP, whilst UBP consists of Ca, Si, Fe, S, Al which was also analysed by XRD.

4. Chemical composition of raw meals

Increasing of UMP up to 81 % caused an increasing LoI and CaO, and a decreasing of SiO₂, Al₂O₃, Fe₂O₃, MgO, SO₃, K₂O and Na₂O. Module calculations of each artificial raw meal confirm its chemical compositions. This implies some important findings: an increase of modules causes an artificial cement clinker which can be produced out of UMP and UBP, and can be burned easily and provides strength development quickly. They do not need more energy than OPC and the lime content of cements could be optimized and modified for a good quality of cement very effectively.

5. Chemical composition of cement

An increase of UMP in the cement recipe up to 81 % caused an increase of LoI, CaO and K₂O, and a decrease of SiO₂, Al₂O₃, Fe₂O₃, MgO, SO₃, and Na₂O. This influences the burnability of the clinker as well. The results of the BOGUE calculation show some fluctuation because of the clinker burning process at 1450°C at 30 min (Editor’s note: In the case of using alternative materials, the Rietveld method turned out to be superior to the BOGUE calculation). Nevertheless, these results indicate that all cements produced with recipe with UMP and UBP are in the class of celite cement.

6. The study indicates a good quality of the new cements, which are close to OPC values known from literature. Although, UMP-UBP-C1 and UMP-UBP-C3 have a lower module than UMP-UBP-C2 regarding HM, SM, AR, LS and LSF when its calcium oxide is approximately 54 %. Therefore, the best quality of cement is UMP-UBP-C2 (Table 7). Burnability Equation of Clinker (BEC) is estimated on the basis of LoI of the recipe after sintering at 1450°C. The high level relationship between LoI of meal and LoI of clinker indicates the best burnability is achievable from the recipe made of UMP-UBP-C2, too.

7. The research study shows UMP and UBP have a potential of use for clinker and cement making technology. Additional investigations will be necessary to develop the best conditions for clinker burning with UMP and UBP at different temperature and time as done here. Further research is also required to monitor the strength development of the new cements, as well as to examine the resistance of these new cements against a number of detrimental factors.