Berlinite substitution in the cement clinker Cement

Cement and Concrete Research 92 (2017) 21–28

Contents lists available at ScienceDirect

Cement and Concrete Research journal homepage: www.elsevier.com/locate/cemconres

Berlinite substitution in the cement clinker Theodor Staněk a,⁎, Petr Sulovský b, Martin Boháč a a b

Research Institute for Building Materials, Hněvkovského 65, 617 00 Brno, Czech Republic Department of Geology, Faculty of Science, Palacky University, tř. 17. listopadu 12, 771 46 Olomouc, Czech Republic

a r t i c l e

i n f o

Article history: Received 16 December 2015 Received in revised form 6 September 2016 Accepted 10 November 2016 Available online xxxx Keywords: Cement (D) Ca3SiO5 (D) P2O5 Microstructure (B) EDX (B)

a b s t r a c t Incorporation of P2O5 in clinker minerals in laboratory samples of clinker with graded content of P2O5 were studied by means of electron microscopy. The most important outcome is that P2O5 enters partially the structure of clinker silicates together with Al2O3 via berlinite substitution (Al3+ + P5+ ↔ 2 Si4+). Therefore, experiments were also carried out with the pure components and the influence of P2O5 on clinker phase composition in the absence of one of the main clinker oxide (Fe2O3 and Al2O3) was examined. It was found out that raw mix without Fe2O3 eliminates to a certain level negative influence of P2O5 on alite formation. The work was completed with preparation of white cement with high content of P2O5 in clinker, which exhibited good technological properties. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction Character of cement manufacturing process allows using various raw and waste materials from other technologies. Among the waste materials utilizable in this way are meat and bone meal (MBM), waste oils and other processed animal wastes. Particularly MBM from BSE positive animals must be disposed at high temperature. It is possible in incinerators and cement rotary kilns. Burning MBM as alternative fuel in cement kiln [1,2] ensures effective degradation and disposal of this hazardous waste, as the temperature of the flame can reach up to 2000 °C. MBM ash formed in the kiln during the burning process is incorporated in the clinker minerals. The main problem of burning of MBM is its relatively high content of P2O5, which is present in the bone material in the form of hydroxyapatite Ca5(PO4)3(OH). The released P2O5 influences the properties of the clinker melt [3], especially its viscosity and polymerization, and subsequently influences the phase composition of clinker [4–7] and the final quality of cement [6,8]. In cement manufacture practice the MBM is used only in quantity which allows to safely avoid its adverse effect on clinker properties, which means to keep the P2O5 content in the clinker below 0.5% (rarely more [9]). It was experimentally proved that P2O5 has negative effect on the nucleation of alite crystals, stabilizing C2S. α-C2S forms a continuous solid solution with α (‘super α’) C3P above 1450 °C [10] with a miscibility gap at 1500 °C [11]. Negative effect of increased belite and free lime ⁎ Corresponding author. E-mail addresses: [email protected] (T. Staněk), [email protected] (P. Sulovský), [email protected] (M. Boháč).

http://dx.doi.org/10.1016/j.cemconres.2016.11.007 0008-8846/© 2016 Elsevier Ltd. All rights reserved.

content at the expense of alite was observed from 0.7 wt.% of P2O5. With the increasing content of P2O5 CaO does not completely react with belite to form alite even after 4 h of burning at 1450 °C, which can be considered as equilibrium burning. Alite formation is completely blocked at 4.5 wt.% of P2O5 in the clinker [12]. As observed in this and other studies ([13,6]) phosphorus is preferentially incorporated into belite, stabilizing it by lowering its free energy. As stated by Herfort [14], any minor component that is preferentially incorporated in one phase over another will effectively stabilize that phase by lowering its free energy due to the increased entropy which dominates over the enthalpy of mixing at low concentrations. In case of phosphorus, its lower incorporation in alite than in belite [12,15] destabilizes this phase. At the same time, P-doped belite becomes thermodynamically in equilibrium with lime, what leads to undesirably high contents of this phase in resulting clinker. Several substitution mechanisms were suggested for the incorporation of phosphorus into calcium silicate phases, e.g. - Ca2SiO4 ↔ Ca3(PO4)2 (e.g., [6,15,16]) - a structure with the exchange of Si for P and Ca2+ vacancies (Va) is suggested for the belite solid solution [17,18] via the chemical reaction 24Ca3SiO5 + Ca3(PO4)2 ↔ {4[(Ca0.75Va(0.25)(Si0.5P0.5)O4] + 22Ca2 SiO4}(ss) + 24CaO In a CaO–SiO2–P2O5 system, Hökfors in [18] has constructed a pseudo-ternary sub-solidus phase diagram for C2S-C3P-CaO from a series of clinker burnings. In the diagram, only a'C2S(P) and f-C coexist (see field IV in Fig. 1) at P2O5 concentrations above 3%; it does not take into account other possible substitutions, like:

22

T. Staněk et al. / Cement and Concrete Research 92 (2017) 21–28 Table 2 Chemical composition of raw materials used for raw mix.

SiO2 TiO2 Al2O3 Fe2O3 FeO P2O5 MnO MgO CaO Na2O K2O LOI SO3 total SO3 sulfate Cl− Humidity

Fig. 1. Pseudo-ternary sub-solidus phase diagram for C2S-C3P-CaO. The solid solutions are represented by bold blue lines. Area I – solid solutions of C3S and βC2S(P); area III – three phase area, where C3S, α'C2S(P) and CaO coexist; the clinkers with the highest P2O5 input reside in two-phase area IV, where α'C2S(P) coexists with CaO. After Hökfors [18]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

- in clinker containing increased amounts of Al and Fe, AlPO4 [12]/ Fe3+ PO4 [19] substitute 2SiO4. The effect of P2O5 is similar to that of SO3 [20], but the influence mechanism is different. Study of the kinetics of reactions and phase changes in clinker due to P2O5 allows to reach considerably higher contents of this oxide in the clinker without adverse effect on its properties and also to reach new findings on its effect on clinker minerals.

Limestone [wt.%]

Kaolinitic sand

Al2O3

MBM ash

0.48 0.13 0.13 0.15 – 0.12 0.012 0.48 55.14 0.010 0.026 43.20 0.07 0.05 0.007 0.11

92.84 0.26 3.45 0.06 1.00 0.20 0.004 0.07 0.80 0.07 0.29 0.92 0.01 – – 0.21

0.01 – 99.30 0.01 – – – – 0.01 – – 0.21 – – – –

1.10 b0.01 0.35 0.32 – 38.05 0.014 1.37 44.84 3.97 2.01 7.48 0.47 0.43 2.75 0.22

15 °C/min, end temperature 1450 °C and isothermal holding time 4 h, slow cooling in the furnace to 1200 °C. This procedure ensures the formation of homogenous clinker structure and slow cooling causes perfect and separated crystallization of C3A and C4AF from the clinker melt. In case of raw meals with MBM ash the non-equilibrium burning as the second type of burning was performed as well. Pellets 40 mm in diameter, pressed from 50 g of individual raw meals were burned in one stage process. The burning regime was as follows: the rate of temperature increase 15 °C/min, isothermal holding at 1050 °C for 1 h, the rate of temperature increase 15 °C/min to end temperature 1450 °C, isothermal holding for 2 h, fast cooling in air. The third type of experiment was focused on the kinetics of formation of clinker with MBM. In these non-equilibrium burning experiments, pre-formed sandwiches of pressed pellets made of ground

Table 3 Composition of raw meals with graded content of P2O5.

2. Material and methods

Ca3(PO4)2

2.1. Clinker preparation

RM-F-0P

RM-A-0P

Theoretical content of P2O5 in clinker

100 98.57 95.64 92.62 – – – –

– – – – 100 98.57 95.64 92.62

0 1 3 5 0 1 3 5

[wt.%]

Raw mix for laboratory tests was prepared from common raw materials (pure limestone, limestone with increased content of SiO2, argillaceous schist and iron correction). Granulometric analysis and chemical parameters are given in Table 1. In the first stage of research, finely ground Ca3(PO4)2 as the source of P2O5 was added in amounts graded to achieve calculated final content of 0.25–5 wt.% P2O5 in clinker. In the second stage MBM ash as the source of P2O5 was used. Individual raw meals with graded amounts of P2O5 were burned in electric super-kanthal furnace to equilibrium state. Pellets 40 mm in diameter, pressed from 50 g of individual raw meals were burned in two stages. The regime of the first burning was: the rate of temperature increase 15 °C/min, end temperature 1050 °C, isothermal holding time 2 h and fast cooling in air. Calcined pellets were ground to fineness below 0.09 mm. New pellets were prepared and burned in the second stage. The second burning regime was: the rate of temperature increase

Table 1 Sieve analysis and chemical parameters of raw meal. Sieve opening [mm] 0.045 0.063 0.090 Residue on sieve [wt.%] 34.8 25.8 17.3

Chemical parameters 0.125

0.200

SLP

Ms

Ma

7.8

2.1

98

2.3

1.7

RM-F-0P RM-F-1P RM-F-3P RM-F-5P RM-A-0P RM-A-1P RM-A-3P RM-A-5P

0 1.43 4.36 7.38 0 1.43 4.36 7.38

Table 4 Phase composition of white clinkers (WCl) and their main chemical parameters.

P2O5 content C3S C2S C3A C4AF Cvol C3Seq. C2Seq. SLP Ms Ma

WCl-0P [wt.%]

WCl-3P

0.05 62.6 13.5 22.9 0.1 1.0 66.8 10.2 96.2 2.82 11.00

3.07 45.0 33.3 18.7 0.3 2.7 56.4 24.6 101.8 2.38 13.20

where:C3Seq = 4.219 ∗ f-C + C3S calculated equilibrium content of alite. C2Seq = 100 – C3Seq – C3A – C4AF calculated equilibrium content of belite.

T. Staněk et al. / Cement and Concrete Research 92 (2017) 21–28 Table 5 Composition of raw mix for white cement.

WRM-0P WRM-3P

23

Table 7 Specific gravity and specific surface of prepared white cements (WC).

Limestone [wt.%]

Kaolinitic sand

80.65 76.40

14.27 13.52

Al2O3

MBM ash Specific gravity [kg·m−3] Specific surface [m2·kg−1]

– 5.27

5.08 4.81

clinker made of the same raw meal as in the above experiments; the adjacent pre-formed pellet was pressed from MBM ashed at 750 °C. The sandwiches were burned at 1450 °C for times ranging between 1 and 20 min, after which they have been removed from the oven and quenched in air. The interface region between clinker and MBM ash pellet was studied in detail by element mapping and linear microprobe spot analyses. As the results of the microprobe analysis of clinker minerals indicated important role of alumina in minimizing the adverse effect of phosphorus on clinker phase composition (see below), white cement clinker with high content of P2O5 was prepared. Analogously, for the burning of clinker without Al2O3, a raw meal (RM-F-0P) was prepared from pure CaCO3, pure quartz (SiO2) and chemically pure Fe2O3. For burning of clinker without Fe2O3 the raw meal (RM-A-0P) was prepared from chemically pure CaCO3, pure quartz (SiO2) and chemically pure Al2O3. Composition and main chemical parameters of raw meals are given in Table 2. From these raw meals new raw meals were prepared with graded P2O5 content. Pure Ca3(PO4)2 was used as the source of P2O5. Composition of raw meals with various content of P2O5 is given in Table 3. Pellets with diameter of 40 mm and weight of 50 g were pressed from these raw meals. The pellets were burned in two stages in a super-kanthal furnace to equilibrium. 2.2. White cement clinker preparation Chemical composition and composition of raw meals used for preparation of white clinkers without P2O5 and with 3 wt.% of P2O5 is given in Tables 4 and 5. Raw meals were prepared in amount of 6 kg and ground in a ball mill to the same residue on sieve with 0.090 mm openings (ca. 12 wt.%). Pellets with diameter of 40 mm were prepared from both raw meals. Pellets were burned in super-kanthal furnace at 1450 °C for 180 min. Quantitative phase composition of clinkers along with main chemical parameters of clinkers and total content of P2O5 are summarized in Table 6. White cements were prepared from burned clinkers with the addition of 6 wt.% of FGD-gypsum as setting retarder.

WC-0P

WC-3P

3102 319

3131 321

The lower limits of detection varied between ~ 0.01 wt.% (e.g. for P, S, Si, Al, Ca) and ~ 0.02 wt.% (e.g. Ca, K, Mn, Fe). The acquired counts, corrected for background, were recalculated to oxides using the PAP correction procedure included in the CAMECA PeakSight automation program. The EMPA results in wt.% oxides were recast into composition formulae in atoms per formula unit (apfu). Equilibrium burned clinkers were crushed below 1 mm. Polished sections were prepared from the 0.045–1 mm size fraction. Quantitative phase composition of prepared clinkers was determined by point– counting in polished sections [21]. For perfect differentiation of individual phases the surface of polished section was etched in fumes of glacial acetic acid [22]. Quantitative phase composition is expressed in vol.% and for calculation of wt.% the following densities of clinker minerals were used: C3S – 3.15; C2S – 3.28; C3A – 3.03; C2F – 4.04; and free CaO – 3.35 g·cm−3. 2.4. White cement testing Density of white cements was determined by the pycnometer method according to ČSN 72 2113 and specific surface by the Lee-Nurse permeability method according to British standard B. S. 12:1958. Clinkers with the addition of FGD-gypsum were ground to approximately same surface area in a laboratory cylindrical porcelain mill with porcelain balls as a grinding charge (Table 7). According to EN 196-1, mortar bars were prepared from the cements and the development of compressive and flexural strength was tested after 2, 7, 28 and 90 days of hydration. Normal consistency, start and time of the setting were determined according to EN 196-3. The temperature development during hydration of cement pastes from both white cements was monitored by semiadiabatic calorimeter [23] according to EN 196-9. Calorimetric curve gives information about intensities of exothermal processes and kinetics of reactions during hydration of various cementitious materials. Total heat after 24 h of hydration was calculated according to Eq. (1): Q 24 ¼

C total ðT−T 0 Þ m

ð1Þ

2.3. Clinker testing Electron microprobe CAMECA SX100 was used for the analysis of clinker phases. The analytical conditions were: accelerating voltage 15 kV, beam current 20 nA and spot size 2 μm. 10 to 20 crystals of alite as well as belite and interstitial phases were analyzed from each clinker sample. The standards used for internal calibration were natural standards - wollastonite (Ca, Si), barite (S), orthoclase (K, Al), andradite (Fe), jadeite (Na), rhodonite (Mn), pyrope (Mg), hydroxylapatite (P), and synthetic TiO (Ti); all elements were analyzed on their Kα lines. Table 6 Composition and main chemical parameters of raw meal RM-F-0P and RM-A-0P. CaCO3

SiO2

Fe2O3

Al2O3

14.83 14.06

5.92 –

– 5.63

[wt.%] RM-F-0P RM-A-0P

79.25 80.31

Main chemical parameters SLP

Ms

Ma

98 98

2.5 2.5

0.0 37.0

Fig. 2. Relation of P2O5 contents in belite and alite to the bulk contents of P2O5 in clinker from the series given in Table 8 (raw meals with P added as Ca3(PO4)2).

24

T. Staněk et al. / Cement and Concrete Research 92 (2017) 21–28

Fig. 3. Relationship between Al2O3 and P2O5 contents in belite from clinkers with different burning regime ad for of P2O5 addition.

where:Ctotal – total heat capacity of the calorimeter in J·K−1 (cement, water, empty calorimeter)m – weight of cement in gT, T0 – ending and initial temperature of mixture

Fig. 5. Relationship of Al and P in belite from clinkers burned without Fe2O3.

the so-called berlinite substitution, when berlinite AlPO4 (isostructural with quartz) substitutes SiO4 tetrahedra (Eq. (2)) – a phenomenon known from nature in feldspars [24], quartz [25], garnets [26] and some other silicates.

3. Results and discussion 3þ

Al 3.1. Electron microanalysis of clinker minerals Chemical composition of clinker minerals in samples of prepared clinkers was studied by electron microprobe spot analyses; the distribution of the main elements was revealed by X-ray maps. The elemental maps have shown that P does not form individual phases – it only occurs as a substituent in both Ca silicates and to a lesser extent in C3A and C4AF. P concentration in belite was in all cases 2.4 to 2.8 times higher than in alite. In belite as well as in alite the average P2O5 content grows with increasing P2O5 addition (Fig. 2), no matter in which form it was added, or which regime of burning (equilibrium/ non-equilibrium) was used (Fig. 3). The relationship between P2O5 and Al2O3 contents in belite as well as in alite appears to be linear over the whole studied range between 0 and 4.5% P2O5 (Fig. 2). This suggests that at least a part of P enters the structure of both Ca silicates via

Fig. 4. Relationship of average apfu values of Al and P in belite of clinkers from the series given in Table 9.

þ P 5þ ↔2Si4þ

ð2Þ

The berlinite substitution is the more important, the higher is the P content in the raw meal. When the raw meal contains besides Al also Fe, the regression equation calculated from the results of spot analyses of belite (Fig. 4) in atoms per formula unit (apfu) is P ¼ 4:13Al−0:21

ð3Þ

In this case, the part of phosphorus incorporated by substitution C2S – C3P appears to be larger than that incorporated via the berlinite substitution. When the raw meal does not contain iron, then much more P is incorporated as berlinite (Fig. 5), as evidenced by the regression equation P ¼ 1:78Al−0:11

Fig. 6. Relationship of Fe and P in belite from clinkers burned without Al2O3.

ð4Þ


25

Fig. 9. P2O5 contents in C4AF phase in relation to bulk P2O5 in clinker. Fig. 7. Relationship of P and Al in C2S – C3P solid solution obtained in kinetic experiments.

On the other hand, phosphorus in combination with trivalent iron is also known to form solid solutions with belite [19]. The corresponding phase, rodolicoite Fe3+ PO4, is isostructural with berlinite. In experimental clinkers containing Fe and no Al, the correlation between Fe and P is also high (Fig. 6) and the regression equation shows that the degree of C2S – C3P substitution is lower than in Fe-free clinker, yet higher than in the case when both Fe and Al are contained in the raw meal (Eq. (5)): P ¼ 3:25Fe−0:24

ð5Þ

Kinetic experiments indicated that the substitution between C2S and C3P is much broader, because they substituted each other practically over the whole range between pure dicalciumsilicate and pure calcium phosphate. The regression relationship here appears to be linear only in the range 0–0.3 a.p.f.u. P; above this value it is less steep. The whole relationship can be expressed by a third order equation (Fig. 7). This suggests the existence of a limit of solubility of Al in Ca silicophosphates with N 0.35 a.p.f.u. of phosphorus. The ratio of P2O5 concentration in belite to that in alite increases up to 0.75% P2O5 in clinker, then it moderately decreases (non-equilibrium

burning) or stays more or less unchanged; the ratio of concentrations is approximately 2.5:1 (2.4–2.8: 1) for contents higher than 1% P2O5 (Fig. 2).

Table 8 Chemical composition of alite in series with addition of Ca3(PO4)2; electron microprobe point analyses.

Bulk P2O5 in clinker SiO2 TiO2 Al2O3 Fe2O3 P2O5 MnO MgO CaO Na2O K2O SO3 Total

0P 0.25P [wt.%]

0.5P

0.75P

1P

2P

3P

4P

0.07 24.49 0.19 0.80 0.74 0.11 0.04 0.60 72.51 0.05 0.09 0.00 99.62

0.51 24.12 0.16 0.80 0.63 0.47 0.02 0.59 72.67 0.09 0.12 0.00 99.67

0.73 24.46 0.15 0.82 0.68 0.62 0.01 0.59 72.24 0.12 0.18 0.00 99.87

0.96 23.24 0.22 1.26 0.90 0.72 0.07 0.66 72.27 0.06 0.12 0.00 99.52

1.81 23.29 0.19 1.30 0.85 0.99 0.04 0.72 71.97 0.08 0.16 0.00 99.59

2.72 22.64 0.18 1.45 0.76 1.31 0.04 0.87 72.03 0.08 0.14 0.00 99.50

3.60 21.77 0.14 1.63 0.61 1.91 0.03 0.86 72.32 0.10 0.13 0.00 99.50

0.28 24.43 0.17 0.72 0.66 0.33 0.05 0.58 72.56 0.08 0.14 0.00 99.72

Table 9 Chemical composition of belite in series with addition of Ca3(PO4)2; electron microprobe point analyses.

Fig. 8. Relationship of average apfu values of Al and P in alite of clinkers from the series given in Table 8.

bulk P2O5 in clinker SiO2 TiO2 Al2O3 Fe2O3 P2O5 MnO MgO CaO Na2O K2O SO3 Total

0P 0.25P 0.5P [wt.%]

0.75P 1P

2P

3P

4P

5P

0.07

0.28

0.51

0.73

0.96

1.81

2.72

3.60

4.50

30.40 0.36 1.79 1.14 0.14 0.08 0.25 63.29 0.22 0.80 0.10 98.57

30.10 0.32 1.63 1.44 0.58 0.08 0.25 62.87 0.28 1.16 0.09 98.8

29.97 0.32 1.69 1.44 0.96 0.07 0.25 62.99 0.30 1.22 0.07 99.28

29.80 0.32 1.66 1.34 1.65 0.08 0.24 62.78 0.34 1.47 0.08 99.76

28.50 0.34 1.95 1.39 1.89 0.06 0.22 64.86 0.23 1.21 0.10 100.75

28.03 0.31 1.90 1.27 3.14 0.06 0.26 62.32 0.34 1.52 0.10 99.25

26.86 0.27 2.01 1.26 3.99 0.05 0.30 62.57 0.37 1.45 0.11 99.24

25.82 0.28 2.59 1.44 4.77 0.04 0.37 63.41 0.34 1.39 0.10 100.55

24.69 0.26 2.55 1.40 5.84 0.06 0.39 62.57 0.39 1.35 0.09 99.59

26


Table 10 Phase composition of equilibrium burned clinkers without Al2O3 (wt.%).

C3 S C2 S C2 F f-C C3Seq. C2Seq.

C-F-0P [wt.%]

C-F-1P

C-F-3P

C-F-5P

65.2 16.5 18.3 0.0 65.2 16.5

63.3 19.6 17.1 0.0 63.3 19.6

32.3 46.5 16.0 5.2 54.2 29.8

0.0 77.0 11.3 11.7 49.4 39.3

Table 11 Phase composition of equilibrium burned clinkers without Fe2O3 (wt.%).

C3 S C2 S C3 A f-C C3Seq. C2Seq.

C-A-0P

C-A-1P

C-A-3P

C-A-5P

57.5 22.0 20.5 0.0 57.5 22.0

56.7 23.3 20.0 0.0 56.7 23.3

51.3 29.1 19.1 0.5 53.4 27.5

2.5 74.4 13.1 10.0 44.7 42.2

Fig. 11. Microstructure of equilibrium burned clinker C-F-3P, blue crystals – alite, brown grains – belite, orange grains – free CaO, white laths among crystals and grains – C2F, dark areas – pores), reflected light, etched by fumes of acetic acid, magnification 145×. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

where: C3Seq. = 4.219 ∗ f-C + C3S calculated equilibrium content of alite. C2Seq = 100 – C3Seq - C2F calculated equilibrium content of belite or. C2Seq = 100 – C3Seq – C3A calculated equilibrium content of belite.

In alite, the trend of P increase with increase of Al is also linear (Fig. 8) and their correlation is very close (R2 = 0.983). The regression equation P ¼ 1:44Al−0:04

ð6Þ

shows that the role of the berlinite substitution is the highest of the discussed cases. If the berlinite substitution were the only mechanism of phosphorus incorporation into calcium silicates, the slope of the regression line would be 1. The C4AF phase contains up to 0.4% P2O5; the incorporation of P into it increases with increasing bulk P2O5 in clinker (Fig. 9). In Fe-free (white) clinkers the P2O5 content increases in C3A (up to 1 wt.%) with increasing bulk P2O5. Our findings do not correspond to findings of other authors [15], who did not find increased P concentrations in C4AF, which may be attributed to less precise method of analysis (EDS, analysis spots in lines without precise localization in analyzed grains).

Fig. 10. Microstructure of equilibrium burned clinker C-A-3P, blue crystals – alite, brown grains – belite, gray laths among crystals and grains – C3A, orange and red inclusions in alite – free CaO, dark areas – pores), reflected light, etched by fumes of acetic acid, magnification 145×. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The decrease of SiO2 accompanied with the increase of P2O5, Al2O3 and MgO contents in clinker silicates was observed in all series of burned clinkers (Tables 8 and 9). The increase of P2O5 is accompanied by the decrease of SiO2, TiO2 and MnO and an increase of Al2O3, MgO, Na2O and K2O in belite. All minor oxides except of MgO are bonded in the belite structure in larger amounts than in the structure of alite.

3.2. The influence of P2O5 on phase composition of clinker without Al2O3 or Fe2O3 Based on the above results further experiments were designed to reveal mechanisms related to P2O5 in the presence of Al2O3. Of the major oxides present in ordinary Portland raw meals, Al has greater mass balance effect on alite than iron. In the series of experiments with systems SiO2-CaO-P2O5-Al2O3 and SiO2-CaO-P2O5-Fe2O3, alite was present at bulk P2O5 concentration of 3 wt% in amount exceeding 50 wt.% (51.3%) only in the first system, while in the second (with Fe and no Al) at the same P2O5 content alite made only 32%; at the same conditions, free lime was only 0.5% in the system with Al and no Fe, while in the other its concentration was ten times higher (see Tables 10 and 11).

Fig. 12. Phase composition of the series of clinkers 0P, 0.25P, 0.5P, 0.75P, 1P, 2P, 3P and 4P from Tables 8 and 9 in dependence bulk P2O5 content in clinker in wt.%.


27

Table 12 Development of compressive (CS) and flexural strength (FS) of samples made from white cements with various content of P2O5 in clinker.

WC-0P WC-3P

Fig. 13. Phase composition of the series of clinkers without Fe in dependence on theoretical bulk content of P2O5 in wt.%.

Examples of the microstructure of equilibrium burned clinkers are given in Figs. 10 and 11. The results show that clinkers without Al2O3 have similar phase composition in dependence on P2O5 content to common clinker with all four main oxides [12]. The degree of conversion of free lime in reaction with belite to alite is higher at higher contents of P2O5. Since the Al2O3 is missing the interstitial matter is formed by C2F. Its P2O5 content increases with increasing bulk P content in the clinker – up to 0.95 wt.% at bulk 5 wt.% P2O5 in clinker. The development of phase composition is considerably different for clinkers without Fe2O3.Total content of alite in these clinkers is lower at the same values of lime saturation, but there is no adverse effect of P2O5 on formation of alite up to 3% of P2O5. Little amount of alite is formed even at 5% of P2O5 in clinker. Comparison of phase composition of three series (of clinkers is shown in Figs. 12–14.

2 days

7 days

28 days

56 days

90 days

FS CS [MPa]

FS

CS

FS

CS

FS

CS

FS

CS

3.0 3.1

5.7 6.4

36.4 33.4

6.7 7.4

52.0 51.6

8.3 7.7

59.2 59.2

7.4 7.6

60.3 58.7

19.5 13.4

3.3.2. Calorimetry Two peaks are present on calorimetric curves of white cements. Analogy of hydration of white cement can be found in hydration of ordinary Portland cement. The initial peak is attributable to a combination of exothermic wetting and the early stage reactions. Since initial reactions take place within first seconds and minutes of hydration it cannot be monitored by semiadiabatic method due to preparation of the sample which takes 5 min after adding of the water. A part of aluminate phase is consumed during initial period giving gelatinous coating and rods of AFt. Then the reactions of aluminate phase slow down almost to zero until the point of exhaustion of solid gypsum. Renewed formation of AFt as the result of hydration of aluminate phase is then present after depletion of sulfates in pore solution [27]. The second aluminate peak can be often seen as a shoulder after main silicate peak. The positions of these peaks were discussed in Lerch as early as 1946 [28]. Tested white cements have high content of aluminate phase so it expectable that the position of the second aluminate peak is moving to earlier times. In our case it is prior to main silicate peak. The relation of the position of the peaks and strength of resulting material from white cement is one of the topics for further research. Temperature maximum (T-max) a time of T-max (t-max) on calorimetric curve from semiadiabatic calorimeter a total heat Q total (24 h) for each cement are given in Table 14. The course of temperature during hydration of samples is depicted in Fig. 15.

3.3. Properties of white cements

4. Conclusions

3.3.1. Setting and strength development Results of strength development are given in Table 12. The results of testing of normal consistency, start and time of the setting are showed in Table 13. Results of technological parameters testing show that both cements have similar properties and the white cement with increased content of P2O5 has no anomalies in terms of strength and setting compared to white cement without P2O5.

P2O5 has significantly adverse effect on formation of alite crystals by entering the structure of C2S with which C3P forms solid solutions and thus stabilizing it. Significant increase of P2O5 content in structure of silicates was revealed by electron microanalysis, particularly in the belite structure, related to increase of total P2O5 in clinker. It was found that P enters the structure of both silicate clinker minerals partially via the so-called berlinite substitution: Al3+ + P5+ ↔ 2Si4+. Therefore new experiments were started in which the role of P2O5 on phase composition of clinker was studied. The elevated amounts of Al supported the formation of alite even at higher levels of P. Burning of raw meals with P and Fe and no Al also resulted in belite showing substitution analogous to the berlinite one (Fe3+ + P5+ ↔ 2Si4+) , but with not so beneficial effect on the phase composition. It was revealed that composition without Fe2O3, which is characteristic for white clinker, eliminates the adverse effect of P2O5 on formation of alite. Finally the white cement was prepared with 3 wt.% of P2O5 which exhibited good technological properties that were close to those of standard white cement without P2O5.

Table 13 Setting times and normal consistency of cement according to EN 196–3.

Fig. 14. Phase composition of the series of clinkers without Al in dependence on theoretical bulk content of P2O5 in wt.%.

WC-0P WC-3P

Normal consistency [%]

Start of setting [hours:min]

End of setting [hours:min]

25.3 25.7

4:00 3:40

5:10 5:00

28


Table 14 Hydration heat of cement pastes from semiadiabatic calorimetry.

WC-0P WC-3P

Tmax [°C]

tmax [hours]

Total heat, 24 h [J·g−1]

70.26 72.82

11.38 10.61

368 391

Fig. 15. Calorimetric curves of cements monitored by semiadiabatic calorimetry.

Acknowledgement This research was done within the project No. 16-08959J financed by the Czech Science Foundation. References [1] U. Kääntee, R. Zevenhoven, R. Backman, M. Hupa, Cement manufacturing using alternative fuels and the advantages of process modeling, R'2002 Recovery, Recycling, Re-integration, 2002 (Feb. 12–15, Geneva) http://www.wbcsdcement.org/pdf/tf2/ R2002_UK.pdf. [2] A. Scheuer, Verwertung von Sekundärstoffen in der Zementindustrie, Cement International 1 (2003) 48–66. [3] V.V. Timašev, A.P. Osokin, Physico-chemical bases of the structure formation and properties of clinker(in Russian) Cement 9 (1982) 4–6. [4] F. Goetz-Neunhoeffer, J. Neubauer, Effects of raw meal substitution by sewage sludge on OPC clinker studies by Rietveld analysis, Proc. 20th International Conference on Cement Microscopy, Guadalajara, Mexico 1998, pp. 130–138. [5] M. Pech, G. Douillet, Operational experience in co-incineration of animal meal in cement kilns, Proc. 5th International VDZ-Congress, Düsseldorf 2002, pp. 352–356. [6] S. Puntke, Auswirkungen des Phosphateintrages in Drehofenanlagen der Zementindustrie auf Klinkermineralogie und ZementeigenschaftenThesis (PhD) Clausthal University of Technology, 2004.

[7] T. Staněk, P. Sulovský, L. Tomancová, M. Hroza, The influence of P2O5 on the phase composition of Portland clinker(in Czech) Proc. Cement 2006, High Tatras, Slovakia 2006, pp. D1–D10. [8] W. Lieber, The influence of phosphates on the hydration of Portland cement, Proc. 6th Intern. Congr. Chem. Cem., Moscow, 110, 1974, pp. 3–13 Sept. 1974, Supplementary Paper, Section II-6. [9] Patent CZ 291729 B6, The Method of Disposal of Organic Waste, in Particular Meat and Bone Meal, Waste Fats and Other Processed Veterinary Waste of Livestock Origin, by Combustion in a Cement Rotary Kiln, patent owner: Považská cementáreň, a. s., Ladce, Slovakia, 2003. [10] R.W. Nurse, J.H. Welch, W. Gutt, High-temperature phase equilibria in the system dicalcium silicate–tricalcium phosphate, J. Chem. Soc. 1077-1083 (1959). [11] W. Gutt, High-temperature phase equilibria in the system 2CaO.SiO2 — 3CaO.P2O5 — CaO, Nature 197 (1963) 142–143. [12] T. Staněk, P. Sulovský, The influence of phosphorous pentoxide on the phase composition and formation of Portland clinker, Mater. Charact. 60 (2009) 749–755. [13] K. Kolovos, P. Loutsi, S. Tsivilis, G. Kakali, The effect of foreign ions on the reactivity of the CaO ± SiO2 ± Al2O3 ± Fe2O3 system. Part I. Anions, Cem. Concr. Res. 31 (2001) 425–429. [14] D. Herfort, Chemistry of the clinker production process, J. Chin. Ceram. Soc. 43 (2015) 1314–1323. [15] T. Ifka, M. Palou, J. Baraček, F. Šoukal, M. Boháč, Evaluation of P2O5 distribution inside the main clinker minerals by the application of EPMA method, Cem. Concr. Res. 59 (2015) 147–154. [16] H. Salge, P. Thormann, The Effect of P2O5 on the Constitution of Portland Cement Clinker, Building Research Establishment, Garston, 1974 (10 pp). [17] S.L. Poulsen, H.J. Jakobsen, J. Skibsted, Incorporation of phosphorus guest ions in the calcium silicate phases of Portland cement from 31P MAS NMR spectroscopy, Inorg. Chem. 49 (2010) 5522–5529. [18] B. Hökfors, Phase Chemistry in Process Models for Cement Clinker and Lime ProductionDoctoral Thesis Umeå University - Thermal Energy Conversion Laboratory, Department of Applied Physics and Electronics, Umeå, Sweden, 2014 (76 pp). [19] M.Y. Benarchid, A. Diouri, A. Boukhari, J. Aride, J. Rogez, R. Castanet, Elaboration and thermal study of iron–phosphorus-substituted dicalcium silicate phase, Cem. Concr. Res. 34 (2004) 1873–1879. [20] T. Staněk, P. Sulovský, The influence of the alite polymorphism on the strength of the Portland cement, Cem. Concr. Res. 32 (7) (2002) 1169–1175. [21] S. Chromý, Accuracy and precision of microscopic qualitative phase analysis of Portland clinker(in Czech) Silikáty 22 (1978) 215–226. [22] S. Chromý, Anfärben des freiem CaO und Silikate in anschliffen von Portlandklinker, Z-K-G 27 (1974) 79–84. [23] J. Brandštetr, J. Polcer, J. Krátký, R. Holešinský, J. Havlica, Possibilities of the use of isoperibolic calorimetry for assessing the hydration behavior of cementitious systems, Cem. Concr. Res. 31 (2001) 941–947. [24] D. London, The berlinite substitution, AlP = 2 Si, in alkali feldspars from differentiated peraluminous igneous rocks (granites, pegmatites and rhyolites), Abstracts with Programs. – Geol. Soc. Am, 22, Geol. Soc. of America, Annual meeting, Dallas, TX, United States 1990, p. 346 (Oct. 29–Nov. 1, 1990). [25] I.V. Veksler, R. Thomas, R. Wirth, Crystallization of AlPO4-SiO2 solid solutions from granitic melt and implications for P-rich melt inclusions in pegmatitic quartz, Am. Mineral. 88 (2003) 1724–1730. [26] K. Breiter, M. Novák, F. Koller, J. Cempírek, Phosphorous – an omnipresent minor element in garnet of diverse textural types of leucocratic granitic rocks, Mineral. Petrol. 85 (2005) 205–221. [27] P.L. Pratt, A. Ghose, Electron microscope studies of Portland cement microstructures during setting and hardening, Phil. Trans. R. Soc. London A310 (1983) 93. [28] W. Lerch, The influence of gypsum on the hydration reactions of Portland cement, Proc. ASTM 46 (1946) 1252–1292.