Users Guide to the 10 Basic Facts on Clinker

Raw mix rejects ................................................................................................................................ 1 1.1 ................................................................................................... 1 1.2 .......................................................................................................... 3 1.3 ................................................................................................... 4 Thermal profile ................................................................................................................................. 5 2.1 Port la Nouvelle ......................................................................................................................... 5 2.2 Le Teil ........................................................................................................................................ 6 Burning atmosphere and volatilization .............................................................................................. 7 3.1 Saint Constant ........................................................................................................................... 7 3.2 Brookfield .................................................................................................................................. 7 3.3 Le Teil ........................................................................................................................................ 8 Free lime content and setting time ................................................................................................... 9 4.1 Saint Constant ........................................................................................................................... 9 4.2 Bath ........................................................................................................................................... 9 4.3 Val dAzergues ........................................................................................................................ 10 4.4 Woodstock .............................................................................................................................. 10 Clinker C3S content ........................................................................................................................ 11 5.1 Villaluenga kiln 2302................................................................................................................ 11 5.2 Villaluenga kiln 1501................................................................................................................ 13 5.3 Whitehall.................................................................................................................................. 13 Clinker C2S content ........................................................................................................................ 14 6.1 Ocumare.................................................................................................................................. 14 6.2 Karsdorf ................................................................................................................................... 14 Alkalies and 28-day strength .......................................................................................................... 16 7.1 Cements from ................................................................ 16 7.2 ....................................................................................................... 17 Alkalies and short term strengths ................................................................................................... 18 8.1 .................................................................................................................. 18 8.1.1 ......................................................................................... 18 8.1.2 .................................................................................. 18 8.2 ................................................................................................................... 19 Alkali saturation .............................................................................................................................. 20 9.1 ................................................................................................................................. 20 9.1.1 . .................................................................................................................. 20 9.1.2 Alkalies in solid solution. .................................................................................................. 20 9.2 Ranteil ..................................................................................................................................... 21 9.3 Ste ......................................................................................................................................... 21

10 Excess of sulfate with respect to alkalies ....................................................................................... 22 10.1 ..................................................................................................... 22 10.1.1 Ste 1971 ......................................................................................................................... 23 10.2 ..................................................................................................... 23 10.2.1 Mekns............................................................................................................................. 23 10.2.2 Ste .................................................................................................................................. 24 10.2.3 La Couronne .................................................................................................................... 24 10.2.4 ......................................................................................................................... 25

1-Raw mix rejects

1
Reducing raw mix rejects lowers burning temperature and grinding energy. This is particularly the case with siliceous rejects. This action is also beneficial to strength properties.

Example:

When the amount of 100m rejects is reduced from 20 to 10%, the global raw mix + cement energy consumption is lowered by about 4 kWh per tonne of cement at a fineness 2 of 350 m /kg.

Raw mix fineness is generally characterized by the weight of rejects in one or several sieves (a 100m sieve is often used). The following examples illustrate these three points:

1.1

Effect on burning temperature

Observed: in 1991, the purchase of a precrusher allowed the Contes plant to improve raw mix fineness. The burning temperature (pyrometer reading or kiln outlet temp.), whose measurement integrates all changes occurring in the burning zone, shows a poor relationship with rejects, whereas fuel consumption shows a good relationship, as demonstrated in Figure 1, taken from four typical periods.

58.5

Oil (l/t Raw mix) 58

57.5 57 56.5 56

>100m residue (%)

55.5 8 12
Figure 1: Contes plant 1991 Several industrial raw mixes were characterized in lab burnability tests at temperatures between 1400 and 1550C (1000C/h heating rate and a 30 minutes hold point at final temperature), both as is and with regrinding of the rejects. The two graphs that follow (Figures 2 and 3), which give the observed free lime as a function of temperature, show the effect of rejects regrinding. We can see that: : the Martres raw mix, which is rich in quartz (> 10%), is harder to burn and more susceptible to fineness than the Karsdorf raw mix.

16

20

______________________________________

________________________________

1-Raw mix rejects

8.00 % free CaO

6.00

4.00

2.00

temperature C 0.00 1400 1450 1500 1550 1600

Figure 2: Free lime after burning for different raw mix finenesses (Martres lab test)

8.00 % free CaO

6.00
Raw mix

4.00

16% > 100m industrial as is 8% > 100m lab re grind

2.00

temperature C 0.00 1350 1400 1450 1500 1550 1600

Figure 3: Karsdorf lab test the % rejects criterion is not sufficient, of itself, to determine a given raw mixs burnability (different results for Martres according to whether the entire sample or only the rejects are reground).

______________________________________

________________________________

1-Raw mix rejects

1.2

Effect on grinding energy

In 1990, the Lexos plant had the chance to grind its raw mix to different finenesses (13% to 21% rejects at 100m) over a long enough period of time so that the plants monthly averages could be considered as meaningful. Figure 4 shows the increase in the raw m ills power consumption with the increased fineness. 25

kWh/t
24

y = -0.4267x + 29.759
r2 = 0.91

y = - 0.42 x + 29.7

23

22

21

20 12 14 16 18

% > 100 m residue

20 22

Figure 4: Raw mix grinding energy vs. fineness A study1 was done to determine the cement grinding power consumption for clinkers that correspond to different raw mix finenesses and for various manufactured products. Figure 5 shows the rise in power consumption for cement grinding corresponding to increasing raw mix rejects for a CEM I 52.5 R cement.
75

kWh/t

y = 1.38 x + 37.8
70

r2 = 0,58

65

60

% > 100 m residue

55 12 14 16 18 20 22

Figure 5: Clinker grinding energy vs. raw mix fineness If we look at the end result, for the production of a CEM I 52.5 R, an additional 1% of 100m rejects in the raw mix causes an increase in grinding energy of more than 0.5 kWh/t.

1 H. Geesen : Influence of raw meal fineness on cement grinding energy (Lexos plant)

______________________________________

________________________________

1-Raw mix rejects

1.3

Effect on mechanical strength

When raw meal fineness was changed at Contes2 (Figure 6), a small increase in strength at constant Blaine fineness was observed. This allowed the plant to slightly lower the Blaine (and increase mill production) while maintaining strength.

Raw
Month >100m >200m Blaine cm2/g 20.0 2.0 3850 3850 February March

CPA HPR
t/h mill 49.0 52.5 55.5 1d
MPa

2d
MPa

28d
MPa

22.5 24.0 22.6

35.7 37.0 37.0

70.0 71.0 70.0

13.0 April - May 13.0

0.8 0.8

3680

Figure 6: Contes 1991

2 R. Dupont : Contes plant 1992

______________________________________

________________________________

2-Thermal profile

2 Thermal profile
A short profile promotes grindability and strength development.

Note:

The optimum is achieved when the kiln torque is at the minimum value compatible with stable kiln operation.

By thermal profile, we mean the rate of heating and cooling of the product in the kiln and cooler. The burning zone length can also be assimilated to this concept. The thermal profile is affected by a number of factors: the raw mix burnability and the kilns heat consumption the type of fuel and its preparation the burner and its settings cooler operation (via secondary air temperature) kiln operation, especially the draught and fuel settings, but also rotational speed Generally speaking, it is somewhat difficult to compare a thermal profile from one kiln to the next. On the other hand, for a given system, several sensors provide readings as to the burning zone length: amps or torque of the drive motors, clinker temperature measured at the kiln outlet (or the NO in the kiln exit gases), temperatures in the preheater cyclones, pressure drop through the Lepol grate, shell scanners, etc. For each kiln, the most representative indicator should to be determined and analyzed. Two recent examples can be given :

2.1

Port la Nouvelle

kWh / t
75 70

BB 10

65 60 55 50 45 1050 1100 1150 1200 1250 1300 1350 1400

C
1450

Cooling zone temperature

Figure 7: Grindability vs ; kiln cooling zone temperature (Port-la-Nouvelle)

______________________________________

________________________________

2-Thermal profile

75 70 65 60 55 50 45 600 700 800 900 1000 1100 1200 1300 1400

BB10 (kWh/t)

Nm
1500

Kiln drive torque

Figure 8: Grindability vs. kiln drive torque (Port-la-Nouvelle) Weekly spot clinker sampling over a period of approximately one year. Simultaneous recording of kiln operating parameters. Clinker grindability measured by BB10. A short burning zone, characterized by a low kiln torque value and high burning zone temperature, affords the best grindability.

2.2

Le Teil
Kiln amps
500 450 400 350 300 250 200
30/05/1995 04:00 30/05/1995 12:00 30/05/1995 20:00 31/05/1995 04:00 31/05/1995 12:00

kWh / t
80

Grindability improvement

75 70 65

Kiln amps decrease

60 55 50

Figure 9: Grindability vs. kiln amps (Le Teil) Spot clinker sampling during a SHTS (~CEM I 52.5 R) production test over a two-day period. Simultaneous recording of kiln operating parameters. Clinker grindability measured by BB10. A lengthening of the burning zone early in test leads to a decrease in clinker grindability. A shorter burning zone (low kiln amp values) affords the best grindability. The technical literature and lab studies point to the favorable impact of a short thermal profile on strength, as well as, the beneficial effect of rapid quenching for strength and workability. This information has not been verified industrially.

______________________________________

________________________________

3-Burning atmosphere

3 Burning atmosphere and volatilization

Steady production requires an oxidizing atmosphere because a reducing atmosphere increases volatilization, causing both cyclical operations and sulfate and alkali fluctuations, hence producing a non-uniform clinker. Numerous industrial trials have demonstrated the effect of the burning atmosphere on volatilization, particularly for sulfur. We can mention the correlation seen in the Sulfur Synthesis3 which was derived using the industrial data from 73 volatile balances: vSO3 = K(K2O, % liquid phase) - 9.3 * O2% Therefore, one %of oxygen equals roughly ten points of sulfur volatilization coefficient. The standardization of high impulse burners and the overall increase in clinker sulfur levels has changed this relationship.

3.1

Saint Constant

The following results have been observed at the Saint Constant plant, where unground fluid coke with up to 2% rejects at 5mm is burned:
Fuel Fuel oil 100% Coke 23.8% Coke 47.7% Coke 47.7% Kiln exit oxygen (%) 1.8 2.4 1.1 2.1 v SO3 (%) 47.1 59.6 83 67 Clinker SO3 (%) 1.39 1.6 1.2 1.96 Clinker K2O (%) 0.94 1.04 0.98 1.15 Emissions SO2 ppmV 2 7 1278 200 Emissions SO3 ppmV 25 19 43 58

Figure 10: Sulfur emissions results at Saint Constant It was observed that the kiln remained stable at all coke percentages. It is evident that the coarse particles in the coke burn on the clinker load: the reducing atmosphere that results shows no effect on kiln stability (because, in the case of a long kiln, the volatilized sulfur has an exit via the stack), but has an important effect on sulfur volatilization and increased SO2 emissions at the stack. In this particular case, a 1% increase in the oxygen exiting the kiln, at constant coke input, translates into a 16 point decrease in the volatilization coefficient. The level of sulfates in the clinker is increased by 60%.

3.2

Brookfield

At Brookfield (long dry kiln), where raw mix SO3 is 1.5%, the kiln went through many cycles that were recognizable by: the cyclical variation of kiln torque the cyclical variation of clinker production, with surges every eight hours variation in the clinker SO3 and K2O content : SO3 between 1% and 4% Cyclical operations, with a period of about two hours, were observed on other long kilns such as Bath and Exshaw (kiln 4), where the raw mixes are rich in volatile elements. One solution used to reduce the sulfur cycle is to divert part of the electrostatic precipitator dust (the finest and highest in sulfur content) away from the kiln circuit. At Brookfield, some tests were carried out on the kiln exit oxygen values. An increase in the draught augmented the dust pick-up, which lead to a redesign of the chain section. A setting of 3.2% oxygen was decided upon instead of 1.9%. Of course, this had a negative impact on heat consumption, but a positive effect on the quantity of by-passed dust.
Kiln exit oxygen (%) volatilization SO3 (%) Dust eliminated (t/d) 1.9 78 90 3.2 40 4 32

Figure 11: Brookfield kiln 1

3 J.C. Guerche, Viviers, 1985.

______________________________________

________________________________

3-Burning atmosphere

3.3

Le Teil

The figure below shows the impact of oxygen levels on the volatilization of sulfates (clinker SO3 in the C4 material).

% SO3 C4 % exit 2 Preheater %O O2 sortie tour

% SO3 C4 4,5 4

%% SO SO3 clinker 3 Clinker

5,5

3,5 3

4,5 2,5 2 4 1,5 3,5 1 0,5 3

17/sep 09:00 17/sep 13:00 17/sep 17:00 17/sep 21:00 18/sep 01:00 18/sep 05:00 18/sep 09:00 18/sep 13:00 18/sep 17:00 18/sep 21:00

% OO2 exit sortie tour 2 preheater % SO 3 C4 SO3 C4 % SO 3 Clinker SO3 KK

Figure 12: Influence of the oxygen level at preheater exit on sulfate volatilization.

Figure 12 shows the switch from 60% to 100% coke on September 17 at 9 a.m. (beginning of chart). In cyclone 4, one can see the increase in SO3 attaining 4% at 1400h. The clinker SO3 remains low. The sulfur introduced as a result of the additional coke does not leave the system: there is a risk of plugging. At 1400h the preheater exit O2 was raised from 4.4% to 4.9%; the C4 material SO3 content goes down from 4% to 2.5%, the clinker SO3 content increases from 0.7% to 1.3%. The oxygen increase allows an acceptable sulfur balance to be reached with 100% coke thereby avoiding plugging in the cyclones, ring formation, etc.

______________________________________

________________________________

4-Free lime

4 Free lime content and setting time

Increasing the clinker free lime content reduces both initial and final setting times in the same proportions. Adding lime also accelerates both initial and final setting times. Order of magnitude: When free CaO increases from 0.5 to 1.5%, initial set decreases by about 40 to 50 minutes. This impact may vary greatly from clinker to clinker4 (-10 to 100 minutes).

4.1

Saint-Constant

A burning intensity study undertaken with St. Constants long dry k iln, allowed the plant to compare the relationship between clinker free lime and initial setting time, as shown in the table below:

Initial setting time (mortar)

( min. )

180 160 140 120 100 0.2 0.4 0.6 0.8 1 1.2 1.4 Clinker free lime
Figure 13: Setting time as a function of free lime The correlation is non-linear: the effect setting time is a lot stronger between 0.4 and 0.8% free lime than beyond 0.8%.

1.6

4.2

Bath

At Bath, another long dry process, numerous tests were carried out to reduce initial and final setting times in response to a customer request. These tests were all aimed at increasing the free lime content to 0.9%: fineness, mix composition, kiln speed/feed ratio, etc. None of these tests gave the desired results. The solution applied today consists of adding limestone at the kiln outlet, in the nose ring area. This addition could technically be done at the cooler inlet, however this would not be in accordance with ASTM standard which apply to part of the plants sales. The results are as follows:

4 Lime quality (specifically its burning temperature, its hydration level, etc.) and clinker quality have an influence on the results obtained.

______________________________________

________________________________

Sample

SSB ( / kg)

free lime (%)

Setting time 23C (plant)

165 125 120

Setting time 23C (CTS)

180 110 115

Concrete setting time (h:min)

5:45 4:55 4:40

Concrete setting time 24C / 2%CaCl2

3:30 3:10

Concrete setting time 10C / 2%CaCl2 4:45 4:25

Concrete set. time 24C 5:10 4:30

4-Free lime Concrete set. time 10C 8:30 7:15

Reference Test 1 Test 2

360 357 371

0.35 0.7 0.9

Figure 14: Injection of limestone at kiln outlet (Bath) Tests 1 and 2 were done with an injection of different addition rates of limestone at the kiln outlet, as reflected in the different levels of free lime. The final setting times were not recorded (they were estimated based on initial setting times). The gain in concrete setting time at 10C is especially noteworthy. .

4.3

Val dAzergues

Following a clients request (BDI), in order to improve the reactivity of Val dAzergues CEM I 52.5 R, particularly in heat-cured (HC) concrete, the level of free lime was increased by adding lime from different sources. The increase in short-term HC concrete strength is partially related to the initial setting time. The gain in short-term HC concrete strength (2h30) slightly penalizes the long-term (5h30). Two types of lime sources were tested industrially: The lime contained in the ash from Gardanne

The co-grinding of 5% ash containing 30% free lime and clinker results in a better performing cement with a setting time that decreases by 60 minutes. Cooler limestone

Some trials with limestone injection (6/10 mm) were done in the cooler throat (at a temperature of 1250 to 1350 C). By introducing 2.5% limestone, 0.95% added free lime was obtained and, for the most part, incorporated in the clinker. The gain in initial setting time was 40 to 60 minutes.

Normal % cm2 /g initial final 1d 28d

+5% Ash

Limestone 1 2,90 4600 110 170 24,5 66,9 6,0 25,3 2 2,50 4420 135 255 24,0 67,2 3,8 26,7

Free lime Blaine Setting time (min) Strength EN (MPa)

1,10 4320 175 245 24,7 69,1 0,5 24,1

2,10 4480 110 180 25,3 67,2 3,8 29,4

Heated concrete 2h30 80C (MPa) 5h30

4.4

Woodstock

At Woodstock, a comparable test was carried out at the kiln outlet, where the effect was minimal. This could be related to the level of lime saturation, which is lower in this plant. This conclusion was also reached in the Delta-free lime study carried out by LCR: the decrease in setting time through an increase in free lime is most effective when the degree of lime saturation in the raw mix is high.

______________________________________

10

_______________________________

5-Clinker C3S

5 Clinker C3S content

Increasing clinker C3S (to the detriment of C2S) improves strength at 1, 2, 3 and 7 days. After 28 days, the gain may be less because of the C2S contribution. +10% C3S +2 to +5 MPa

Order of magnitude:

in the short and medium terms

5.1

Villaluenga kiln 2302

Villaluenga plant. Production period from February to June 1992. Start up of kiln 2302 (kiln w/ AS precalciner): during this stage, there was a progressive increase in C3S (see Figure 15) which generated enough data to allow us to study its influence on strength.

65.00 60.00
% C3S

55.00 50.00 45.00 40.00

12/02/92

10/03/92

12/03/92

17/03/92

24/03/92

31/03/92

07/04/92

21/04/92

29/04/92

05/05/92

13/05/92

19/05/92

03/06/92

Figure 15: Evolution of C3S (Villaluenga plant)

Characteristics of clinker produced: The average clinker values are: C3S = 53.2, SR = 2.8, A/F = 1.75, SO3/alkalies molar ratio = 1.06. In addition to the variation of C3S, an increasing consumption of coke during the start up caused clinker SO3 to rise. This increase in SO3 resulted in an increase of the SO3/alkalies molar ratio from 0.8 to 1.24, and at the same time, a higher soluble SO3 content.

______________________________________

11

_______________________________

26/05/92

10/06/92

9/02/92

7/02/92

8/02/92

3/03/92

5-Clinker C3S

1.80 1.60 1.40 1.20

% SO3

1.00 0.80 0.60 0.40 0.20

12/02/92 10/03/92 12/03/92 17/03/92 24/03/92 31/03/92 07/04/92 21/04/92 29/04/92 05/05/92 13/05/92 19/05/92 03/06/92 26/05/92 10/06/92 9/02/92 7/02/92 8/02/92 3/03/92

0.00

Figure 16: Evolution of clinker SO3 (Villaluenga plant) The strength increase with increasing C3S, at all ages is clearly shown in Figure 17 (clinker sample lab ground at 3600 SSB with total sulfates held constant).

60.0 55.0 50.0 45.0

28d 7d

y = 0.42 x + 32.24 r2 = 0.31 y = 0.55 + 15.51 r2 = 0.58 y = 0.58x + 1.42 r2 = 0.67

N/mm 2

40.0 35.0 30.0 25.0 20.0 15.0 45.0 50.0 55.0

2d 1d

y = 0.45x - 2.59 r2 = 0.43

60.0

C3 S
Figure 17: Evolution of strengths as a function of C3S (Villaluenga plant) 1 day: gives a moderate correlation (r =0.43) which improves in the case of multiple regression using the proportion of sulfate in the clinker. Result: 4.5 MPa for 10% C3S 2 2 days: good correlation (r =0.67). Result: 5.8 MPa for 10% C3S 2 7 days: good correlation (r =0.58). Result: 5.5 MPa for 10% C3S 2 28 days: poor correlation (r =0.31). Result: 4.1 MPa for 10% C3S
2

It must be noted, however, that the SO3 and C3S evolved in the same manner during those four months and it is therefore difficult to distinguish their individual effects and the impact of optimum sulfate addition. This may explain why the slopes of the linear regressions (1 day and 28 days in Figure 17) are very similar.

______________________________________

12

_______________________________

5-Clinker C3S

5.2

Villaluenga 1501

Production in normal conditions for kiln 1501 (kiln w/ AS precalciner). This corresponds to the same period as the start up of kiln 2302 but the C3S is less variable. Clinker characteristics: the average values are C3S =57.9, SR=2.7, A/F=1.73, SO3/alkalies molar ratio = 0.92. The compressive strengths vs. C3S content for different ages are show in Figure 18:

60.00 55.00 50.00 45.00

y = 0.59x + 20.84 r 2 = 0.59

28d 7d 2d 1d
55.0 60.0
y = 0.68x + 5.61 r 2 = 0.66

N/mm 2

40.00 35.00 30.00 25.00 20.00 15.00 50.0

y = 0.54x + 0.771 r 2 = 0.59 y = 0.28x + 4.88 r 2 = 0.22

65.0

C3 S
Figure 18: Evolution of strengths as a function of C3S (Villaluenga plant, kiln 1501) 1 day: very weak correlation (r =0.22). Result: 2.8 MPa for 10% C3S 2 2 days: good correlation (r =0.59). Result: 5.4 MPa for 10% C3S 2 7 days: good correlation (r =0.66). Result: 6.8 MPa for 10% C3S 2 28 days: poor correlation (r =0.59). Result: 5.9 MPa for 10% C3S
2

Note the weak evolution of 1 day strength.

5.3

Whitehall

For reasons that have to do with the quarry, the Whitehall plant had to perform some clinker production tests with low C3S levels. The industrially produced clinkers were ground in the lab at constant sulfate addition. The table on the following page shows the test results.
40 35 30 25 1 d 3 d 7 d 28 d

MPa

20 15 10 5 0 10 20 30 40 50 60

C3 S
Figure 19: Evolution of strengths as a function of C3S

______________________________________

13

_______________________________

6-Clinker C2S

6 Clinker C2S content

For a given Blaine specific surface (SSB), grinding energy increases with C2S content. Conversely, it decreases with increasing C3S. +10% C2S, (or -10% C3S) +5 kWh/t (for 3500 c

Order of magnitude:

/g).

6.1

Ocumare

The Ocumare plant5 (FNC - Venezuela) recently modified its raw mix composition (February 1997) by increasing the lime saturation in order to improve clinker reactivity. All other parameters including the clinker free lime remained constant. The table below summarizes the situation over a consecutive twomonth period, before and after changing the mix.
% LSF Clinker C3 S C2 S C3 A C4 A F Free CaO kWh/t BB10 SSB BB10 ( cm2/g ) January 1997 94,50 56 22 10,20 8,20 1,27 46,70 3680 53,10 32,0 February 1997 98,60 64 15 9,60 8,00 1,30 44,40 3710 51,60 35,5

Feb. - Jan.
+ 4,10 +8 -7 = = = - 2,30 = - 1,50 + 3,5

kWh/t Industrial 28 d. Industrial (ASTM standards)

Figure 20: Evolution of grindability as a function of chemistry at Ocumare The effect on grinding energy is lesser in the plant than in the lab, although it has the same tendency to decrease. If we use the figures obtained in the lab (calculated for 350 /kg), we can evaluate the drop in 1.5 1.5 power consumption at 46.7 x (350/368) - 44.4 x (350/371) = 2.6 kWh/t for 7% less C2S, or:

- 3.8 kWh/t for - 10% C2S Note that during the same period, the increase in C3S content resulted in a gain of 3.5 MPa at 28 days on industrial cement.

6.2

Karsdorf

Until 1994, only one clinker with high lime saturation (KST = 98) was used for the entire cement product range. At that time, the new European standards lowered the upper limit of 28-day strength of CEM I 32.5 to 52.5 MPa from its previous 55 MPa. Meeting this demand was not easy. The solution, which consisted of reducing cement fineness, caused problems with bleeding, which was unacceptable for the users.

5 Presentation given by L. Corda, J.A. Sbardella from Gerencia Desarollo y Procesos, Cementos La Vega, TYTP Combustion Meeting on April 15-16, 1997 at Yozgat.

______________________________________

14

_______________________________

6-Clinker C2S

It was therefore decided to produce a second clinker less saturated in CaO (KST = 90), which gives less 28 day strength because of lower C3S content. The results obtained6 are as follows:
KST = 90/92 45 25 45 KST = 96/98 55 15 41 Difference -10 +10 +4

% C3 S % C2 S KWh/t BB10 @ 3500 SSB

The increase in the percentage of Belite causes the mill power consumption to increase by:

+ 4 kWh/t for + 10% C2S It must be stated that this solution is not satisfactory firstly because of increased energy cost and secondly because the strong reactivity of the Belite at Karsdorf doesnt allow for a significant reduction of 28 -day strength in the cement considering the clinkers high lime saturation. Figure 21 below shows the results of laboratory-ground cement made from two industrial clinkers with two different sulfate addition rates (2/3 gypsum 1/3 SH). Another method is being studied, which consists of modifying the sulfate addition of KST 98 clinker. The trouble here lies with the risk of rheological disturbances.

Figure 21: Laboratory cements made from Karsdorf clinkers

6 Presentation given by G. Cochet and G. Staupendahl at the CTI/CTEC Technical Days in Madrid, October 1996. Study of clinker grindability at Karsdorf by G. Cochet, CTI, September 1995.

______________________________________

15

_______________________________

7-Alkalies and 28-day strength

7 Alkalies and 28-day strength

Alkalies7, whatever their form, are never favorable to 28-day compressive strength.

Order of magnitude:

+ 0.1 % Eq Na2O total -1 N/m

at 28 days

It is usually very difficult to change the alkali content in a given plant without greatly altering other parameters, because the content in the individual raw materials tend to be relatively constant.

7.1

Cements from Lafarge Ciments and Lafarge Corp.

The measurements performed on 29 industrial cements 8 (from Lafarge Ciments and Lafarge Corp. as well as French competitors) confirm the above relationship. When the TOTAL alkalies increase from 0.2% to 1.8%, one notices that the mechanical strength at 28 days9 decreases from 66 MPa to approximately 45 MPa. The regression equation indicates a loss of 1.3 MPa for an increase of 0.1% in total alkalies, with a correlation coefficient of 0.88. The results are shown in the diagram below:
70 65
28 -day Strength (MPa)

60 55 50 45 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 % Total Alkalies

Fig ure 22: Influence of total alkalies on 28-day strength

7 The clinker alkalies may occur in different forms depending upon the degree of alkali saturation by sulfates. The total alkalies Na O+K 0 are present 2 2 in either or both of the following forms: Soluble alkalies The alkalies combined with sulfate are Na2SO4, K2SO4, K2SO4(CaSO4)2. These compounds can make their appearance during industrial production when the fuel is changed from a natural gas to a sulfur-bearing fuel like bunker oil or petroleum coke or when gypsum is added to the raw mix. In this case, there will be more soluble alkalies at the expense of alkalies in solid solution. Alkalies in solid solution The alkalies that are not combined with sulfates will enter the aluminate and silicate crystal lattices, modifying their reactivity. This can be a problem for the C3A (loss of workability) as mentioned in the ninth basic fact. 8 R. Guyot, R. Ranc, B. Cariou: Sulfates Synthesis Report , June 1983. 9 at a constant Blaine and with sulfate addition optimized for 28-day strength.

______________________________________

16

_______________________________

7-Alkalies and 28-day strength

7.2

Effects on a single clinker

Laboratory testing10 was carried out on a single clinker to measure the effects of alkaline sulfate addition in the water:
80 80

Rc M P a Rc M P a

28 d R 28j R 28j

70 70

28 d + K2O R 28j + K2O R 28j + K2O

60 60

7d ++ K2O R 7j K2O R 7j + K2O

50 50

d R77j R 7j
40 40

2R d+ 2jK2O + K2O R 2j + K2O

30 30

20 20

d R22j R 2j

10 10

1d+K2O R 1j + K2O R 1j + K2O

0 25 25

30 30

35 35

40 40

45 45

50 50

55 55

60 60

% C3S

1R d 1j R 1j
65 65

Figure 23: Influence of alkalies on strengths

10 M. Debos, G. Chaudouard

______________________________________

17

_______________________________

8-Alkalies and short-term strengths

8 Alkalies and short-term strengths

At optimum sulfate addition for early ages, soluble alkalies in the form of alkali sulfates improve early strength.

Order of magnitude:

+ 0.1 % Eq. Na2O soluble + 0.5 to 1.5 N/m

at 1 day

The soluble alkalies [Na2SO4, K2SO4, K2SO4(CaSO4)2] in the clinker are, as mentioned above, mainly in the form of K2SO4 although they are calculated on the basis of Na2SO4 equivalent. There are two ways to increase soluble alkalies: Increase the sulfates insofar as the alkalies are not yet saturated. In this case, the soluble alkalies will increase. This can be done either via the fuel (ex. : gas bunker oil or low sulfur content oil high sulfur content oil) or via the raw mix (adding sulfates to the mix). Increase the alkalies in the raw mix insofar as there are available sulfates. This can be done by using a siliceous sand that is rich in alkalies, for example (sea sand). These two examples will be further investigated.

8.1

Increasing sulfates

8.1.1 Adding gypsum to the raw mix

Two industrial cases from the Ranteil and Ste plants can be cited as examples.
Plant Ranteil Ste As is As is % SO3 Clinker 0.2 1.1 0.1 1.3 % K2O Total 1.15 1.15 0.6 0.6 %Na2O Total 0.1 0.1 0.2 0.2 SSB (m2/kg) 400 375 402 413 1d (MPa) 11.5 18.0 11.5 17.0

Gypsum addition Gypsum addition

Figure 24: Sulfate addition to raw mixes at Ranteil and Ste In both cases, it is very clear that an increase in the saturation of alkalies with sulfates (increase in soluble alkalies) leads to a very slight increase in 1-day strength. In both cases with the gypsum addition to the raw mix, the C3A takes on a cubic form, and no longer the orthorhombic form which is the case in the presence of alkalies in the C3A crystal structure.

8.1.2 Increasing the fuels sulfur content

Shown in the table below are results from the Martres and La Malle plants during a switch from a fuel with a low sulfur content to a sulfur-rich fuel.
Plant Martres La Malle Gaz Fuel oil low-s oil Hi-s oil % SO3 Clinker 0.20 1.00 0.60 1.00 0.95 0.10 % K2O Total 0.40 % Na2O Total 0.06 SSB (m2/kg) 370 350 360 340 1d (MPa) 12.5 15.5 18.0 21.0

Figure 25: Increase in fuel sulfur at Martres and La Malle

______________________________________

18

_______________________________

8-Alkalies and short-term strengths

For both plants, when one operated with gas and the other with low-sulfur oil, the alkalies were not totally saturated. There would have to have been 0.40% and 0.93% of SO3 in the clinker to saturate the alkalies at the Martres and La Malle plants, respectively. The switch to higher sulfur content oils in both cases produced the saturation of the alkalies and, therefore, an increase in the amount of soluble alkalies, with a positive impact on 1-day compressive strengths. In the lab, the increase in the portion of soluble alkalies was also tested, confirming industrial results. In these laboratory tests, 1.85% of K2SO4 (=1% K20) was added to clinker, equivalent to a 1.85% addition of K2SO4 to the cement during the mixing with water. The table below confirms once more the positive effect on early strength and the negative effect on longterm strength.

Term 1 day 7 days 28 days

As is 20.3 54.3

1.85 % K2SO4 clinker 28.1 53.0

1.85 % K2SO4 cement 28.2 53.8 66.3

74.3 66.2 Figure 26: Strength of cements spiked with K2SO4 in the lab

8.2

Increasing alkalies

At the Retznei plant, in the past, we used an additional source of SiO 2 containing alkalies with high amounts of Na2O between 2.2 and 2.5%, and K2O between 1.2 and 1.4% Analysis of the raw mix at that time showed that the alkali content was the highest of all cement plants in Austria, with 0.45% of Na2O eq. (in the raw mix). Today, with a source of silica that is poorer in alkalies (Na2O = O and K2O = 0.5%), we have brought the Na2O eq. values down to around 0.3%. The consequences of this drop in raw mix alkali content can be seen in terms of both short (lesson 8) and long-term (lesson 7) mechanical performances in the table below.

Rmeq Na2O=0.45% Measured (addition 16%) 1 day 28 days Number of tests 17.1 0.1 46.8 0.19 Caldulated at O% addition 21.1 0.1 57.8 0.19
1

Rmeq Na2O=0.3% Measured (addition 20%) 13.9 0.12 47 0.23 Calculated at 0% addition 18.5 0.12 62.7 0.23

143 100 Figure 27: Cement strength with varying levels of alkalies (Retznei)

Despite slight variability in the alkalies and the difference in cement composition (additive ratios of 16 and 20%), it remains possible to compare the results.

______________________________________

19

_______________________________

9-Alkali saturation

9 Alkali saturation
The molar saturation of alkalies by SO3 in the clinker facilitates workability control.

9.1

Reminder

SO3 molar saturation: % SO3 = 1.29 (% total Eq. Na2O). A low SO3/alkali ratio results in a small percentage of soluble alkalies and the presence of orthorhombic C3A. All Portland clinkers11 contain alkalies in greater or lesser quantities. But depending on the nature of the fuel used (oil or petcoke rich in sulfur or low-sulfur coal), the alkalies can be found in two different forms. When the clinker contains sulfates that come from the raw mix and/or fuel, a significant part of the alkalies are in the form of alkali sulfates. These alkalies are referred to as soluble . Non-soluble alkalies are incorporated in the silicate or aluminate crystals structure The sum of soluble alkalies and alkalies in the crystal strucutres is called total alkalies .

9.1.1 Alkali sulfates

Alkali sulfates improve initial strengths (1 and 2 days), but reduce long-term strength (28 days or more). There is no negative effect on the main cement properties.

9.1.2 Alkalies in solid solution

If the clinker does not contain enough sulfates to combine all the alkalies, then the alkalies enter the aluminate and silicate crystal structures. The defects created in the C3A crystals modify the lattice and its morphology assumes an orthorhombic rather than cubic form. This transformation is accompanied by an increase in its reactivity with water. The presence of alkalies in the crystal structure has numerous unfavorable effects : change in burnability rheological disturbances due to the slow formation of ettringites brought about by the hydration of very reactive orthorhombic C3A strong sensitivity to weathering effects, which accentuates the rheological defects expansion increased shrinkage during the plastic state (24 hours) increased shrinkage during drying at 28 days inferior 28-day strength without any improvement in initial strengths Given the fact that, in general, we are not able to control clinker alkali content, and given that alkali sulfates present many advantages, it is imperative12 that the SO3/alkali molar ratio be >1.0. There are very few recent examples because of: the nearly-generalized use of sulfur-rich fuels for many years now (thus limiting the cases of orthorhombic C3A). a lack of information on complaints or disputes from the plants that have SO3-poor clinkers (gas burning for example: Venezuela) for which the rheological control of cement and concrete is done either sporadically or not at all. some tests for raw mix sulfate addition were not carried out under well controlled conditions (Gabon).

11 M. Debos, G; CHAUDOUARD. Portland Cement: chemistry - mineralogy - properties of phases - reactivity (June 1991). 12 R. Ranc. Influence of alkalies on the physical and mechanical properties of Portland cements . Sept. 93.

______________________________________

20

_______________________________

9-Alkali saturation

The above observations are taken from examples of industrial and lab testing which compare the properties of normal clinkers without sulfates and clinkers made from a sulfate-rich raw mix as the result of the addition of gypsum. These tests were all carried out for the purpose of improving the rheological properties of cement (by eliminating the causes of false set).

9.2

Ranteil

The table below presents the laboratory results13 (burning and grinding) due to the addition of gypsum to the raw mix at Ranteil (non-saturated alkalies).
% raw mix % gypsum SO3 kk 100 0,35 0 97,95 1,40 2,05 96,0 2,40 3,95 SSB 2 ( cm / g ) 3830 3770 3920 Fluidity (%) 88 112 125 1d MPa 6,5 10,5 12,5

Clinker as is Sulf. raw mix Sulf. raw mix

SO3 cem. 2,0 2,50 2,90

W/C 28,6 26,0 26,2

Figure 28: Lab testing (burning and grinding) The cements prepared from the clinker made from a raw mix with gypsum addition are more fluid than the control clinkers. The higher the SO3 clinker content, the better the fluidity appears to be. .
Crystal form As is Sulf. raw mix C3A

Orthorhombic ( alkalies in crystal ) Cubic ( pure )

9.3

Ste

Industrial tests14 were carried out in 1970:

SO3kk K20 tot. Na2O tot. SO3 SSB
(cm/g)

Rsidue17(mm)
after 2 min after 30 min

1d
MPa

As is Cement (sulfated kk)

0,1 2,9

0,37 0,56

0,10 0,09

2,80 2,90

3510 3500

14 10

29 19

11,5 16,0

Figure 29: Results of raw mixes with gypsum addition at Ste Probe penetration tests show an improvement in the rheological characteristics of pure pastes and mortars that come from clinker where gypsum has been added to the raw mix.

13 Ray. Allgre. Study on the influence of the addition of gypsum to the raw mix at Ranteil (1972) 14 Ray. Allgre. CB N.20 IDSG industrial trial at the Ste plant - OS 11445 March 1971.

______________________________________

21

_______________________________

10-Excess of sulfates

10 Excess of sulfates with respect to alkalies

If clinker SO3 is increased beyond the molar saturation of alkalies, an increase in both clinker fineness and grinding energy is noted.

Order of magnitude:

+1% Excess SO315 + 5 kWh/t at 350

/kg

If the clinker excess SO3 is increased beyond the molar saturation of alkalies, the following are observed: an increase in clinker fineness an increase in grinding energy Each of these points will be dealt with separately.

10.1 Increase in clinker fineness

This observation has been reported time and time again, but has never been really quantified through granulometric analysis (see 1995 raw mix sulfate addition tests for Gabon). Nevertheless, the industrial testing done at Ciments Lafarge in the 1970s showed that the presence of sulfates in the clinker beyond the saturation of alkalies leads to a dustier clinker. Below are shown three examples from industrial testing done with raw mix sulfate addition in the Ste and Ranteil plants.

100 80

% passing

60 40 20 0 0 5 10 15 20 25

0,1 % S03 1,3 % SO3

mm

Figure 30: Comparison of clinker granulometries (Ranteil plant 72-73)

15 excess SO = SO clinker - 1.29 (% total Eq. Na O) 3 3 2

______________________________________

22

_______________________________

10-Excess of sulfates

10.1.1 Ste 1971

100 80

% passing

60 40 20 0 0 5 10 15 20 25 30 35

0,1 % SO3 2,9 % SO3

mm
100 80

% passing

60 40 20 0 0 10 20 30 40 50 60

0,1 % SO3 1,2 % SO3

mm

Figure 31: Comparison of clinker granulometries (Ste plant 1971) Figure 32: Comparison of clinker granulometries (Ste plant 197116) It is likely17 that in these tests, there was a significant volatilization factor due to a poorly controlled burning atmosphere and that the increase in fines could (partially) be the result of alkali sulfate volatilization. Recent tests have shown that the burning zone length has a significant effect on clinker particle size and that the level of SO3 in the kiln load influences burnability.

10.2 Increase in grinding energy

An excess of SO3 in the clinker beyond the saturation of alkalies worsens grindability. This is a fact that has been observed and reported time and again in numerous documents. The attempts to quantify the impact of clinker SO3 (total SO3, excess SO3, etc.) on grindability have been numerous, both industrially and through laboratory and statistical studies. For the statistical studies, the results are influenced by the parameters taken into consideration. Some results are reported below. For the studies carried out in the plants, the recent increase in high sulfur fuel usage should give us more data. These data are difficult to exploit, however, insofar as the clinker SO3 parameter was not the only one to fluctuate (burner, combustion, optimized mill operations, fineness changes, changes in cement additive ratios, etc.).

10.2.1 Mekns
At the Meknes18 plant, a gradual switch to coke was made without any major changes to equipment, raw mix or products.

16 Ray. Allgre / CB 20 OS n 11445 Industrial test JDSG at the Ste plant March 1971 17 M. Debos Sulfates conference (L'Isle d'Abeau) : Influence of sulfates on kiln operations Nov. 1993 18 Meknes plant. Market reports 1994/95

______________________________________

23

_______________________________

10-Excess of sulfates
coke coal fuel SO3 F2 0,45 0,95
kk

K2OT F3 0,75 1,18 ~ 0,35 ~ 0,35

Na2OT

C/K

Mill kW CPJ 45 38,52 40,29

1994 1995

47,37 74,14

51,01 21,30

4,62 3,56

#0 #0

1,222 1,237

Figure 33: Results of an increase in sulfur (Mekns plant) * CPJ 45 = CEM II 32.5 These results show that despite a slight increase in the additive ratio (limestone, easier to grind than clinker), cement grindability (CPJ 45) decreases with an increase in the percentage of coke.

10.2.2 Ste
In the industrial study16 on sulfate addition to the raw mix at the Ste plant (1971), cement grinding trials were carried out both in the lab and in the plant. The results are expressed in terms of: output for the industrial trials mill rotations and SSB (Blaine specific surface) for the lab trials.
SO3 kk
Industrial tests kk normal kk sulfated Lab.tests kk normal kk sulfated

K2OT

Na2OT

SO3 cem

Prod.
t/h

SSB # of mill rev

cm/g

0,1 2,9 0,1 2,9

0,37 0,56 0,37 0,56

0,10 0,09 0,10 0,09

2,8 2,9 -

21 17 -

3150 2700 2880 2580

2000 2000

Figure 34: Cement grinding results for raw mixes with gypsum addition (Ste plant)

10.2.3 La Couronne
The use of petcoke at the La Couronne19 plant produced an increase in clinker SO3 content and a reduction in terms of pure cement production rate from the various mills.
1991 SO3 kk 0,80 1992 1,45

K2 O

0,89

0,92

Production t / h (B0) (B3) (B3) CPA 55 CPA 55 CPA HPR 21,6 19,3 14,2 20,3 18,2 13,2

Figure 35: Results of sulfur increase at La Couronne The study done on 12 clinkers at La Couronne shows that the SO3/alkalies ratio correlates with clinker grinding energy. W 4000 SSB = 5.44 SO3 /alkal. + 55.7 with: SO3 /alkalies between 1 and 4 r =0.67
2

19 M. Debos Sulfates conference Nov. 18-19, 1993. : Grindability Nov. 1993

______________________________________

24

_______________________________

10-Excess of sulfates

10.2.4 Cantagalo Clinker

On a series of 17 industrial clinkers20 taken from the Cantagalo plant, it was determined that lab grinding energy is correlated with clinker SO3 (or with the excess of SO3), according to the equation: W (# of mill rotations BB 10) = 1008 SO3 kk + 3250 r =0.66
2

Figure 36

4800 4600 4400

rotations BB10

4200 4000 3800 3600 3400 0.2 0.4 0.6 0.8 SO3 kk 1 1.2 1.4 1.6

y = 1008 x + 3255
r 2 = 0.66

Figure 36: Grindability as a function of SO3 (Cantagalo plant)

20 P. Barriac : Cantagalo clinker , May 1995

______________________________________

25

_______________________________