 Open Access
 Total Downloads : 425
 Authors : Olubajo O. O., S. M. Waziri, B. O. Aderemi
 Paper ID : IJERTV3IS20732
 Volume & Issue : Volume 03, Issue 02 (February 2014)
 Published (First Online): 01032014
 ISSN (Online) : 22780181
 Publisher Name : IJERT
 License: This work is licensed under a Creative Commons Attribution 4.0 International License
Kinetics of the Thermal Decomposition of Alum Sourced from Kankara Kaolin
1Olubajo O. O., 2S. M. Waziri, 2B. O. Aderemi
1Department of Chemical Engineering, Abubakar Tafawa Balewa University Bauchi, Nigeria
2Department of Chemical Engineering, Ahmadu Bello University, Zaria, Nigeria
ABSTRACT This study investigated the kinetics of the thermal decomposition of alum produced from Kankara kaolin. The work involved preparation of alum as well as calcination of the prepared alum. The CoatsRedfern kinetic model was adopted in estimating the kinetic parameters for the decomposition reactions. The experiment was conducted using a muffle furnace and the rate data were obtained using thermogravimetric (TG) and Xray fluorescence (XRF) analyses. From the TGA experimental data, an average order of reaction, n of 1.3 with corresponding activation energy value of 321.03 kJ/molK and frequency factor 2.57E+14min1 gave the smallest residual deviation values and best fitting curves compared to other reaction orders tested within the range of 0 to 2, while equivalent values from XRF residual sulphate data are n = 1, activation energy = 337.22 kJ/molK and frequency factor of 5.27E+14 min1. The discripances in the corresponding values from the two modes of data generation is attributed to the differences in the reaction mechanisms captured by the different methods of analysis. While, the XRF solely followed the sulfate decomposition phenomenon, TGA was more encompassing including dehydroxylation and impurity decomposition reactions.
(KEYWORDS: Kankara Kaolin, decomposition, CoatsRedfern model, TGA, XRF, kinetics)

INTRODUCTION
The great interest in alumina is mostly due to its ever increasing applications as adsorbents, catalysts support, refractory, filler and as constituents of various household products. Invariably, bauxite has been the dominant source alumina, through the popular Bayers process [1]. However, notwithstanding the acclaimed bauxites relative economic advantages over all other alternative sources, its rapid depletion in quantity and quality globally is an impetus promoting intensive research activities on alumina extraction from clay materials worldwide [2]. In fact, this is a welcome development to many countries such as Nigeria that are not naturaly endowed with bauxite but having kaolin in abundance.
It must be admitted that in the last decade, a lot of research efforts have been expended in producing alums, alumina and silica for various purposes from Kankara kaolin clay [38]. However, it is of note that while much of the efforts were geared towards establishing favourable process conditions,
apparently, no such attention has been paid to the prevailing kinetics.
The work of Moselhy et al. (1994) [9] , on thermal treatment of aluminium sulfate hydrate made a fundamental contribution in addressing the peculiarity of alum decomposition including the testfitting of the CoatRedfern kinetic model. However, they limited their consideration to only the ideal solidstate reaction cases of order n= Â½, 2/3 and 1, thus ignoring the obvious peculiarity of alum. At the onset, it is obvious that a well crystalline alum behaves as a non porous material, in which the shrinking core reaction model ought to describe, howbeit, as reaction progresses and it is evacuated of the occluded free water and sulfate ions, it becomes a porous material to mimic a diffusionprogressive phenomenon. Even at that, decomposition of alum to release the bonded waters (water of crystallization) requires different energy level (activation energy) to that of sulfate decomposition, with no clear cut demarcation in time or space of the occurence of these duo phenomena.
The strenght of the CoatRedfern model resides in its development, which is free from the majority of the idealized models assumptions. It equally merged the kinetic constants with the thermodynamic parameters in a single equation, thus affording the evaluation of reaction order, activation energy and reaction frequency factor from same set of data. Its major shortcoming being that the order of reaction determination, inherently depends on the trial and error principle, whether by numerical or graphical approach.
The use of thermogravimetric data to evaluate kinetic parameters for solid state reactions involving weight loss has been investigated by a number of workers as noted by Coats and Redfern (1964) [10] , but sad enough, the single sample, rapid, and continous kinetics calculation over the entire temperature range offered by dynamic thermogravimetry is still not executable in Nigeria, five decades after, due to lack of requisite thermobalance facility in the neighborhood.
Hence, this paper attempts to estimate the kinetic parameters such as activation energy, preexponential factor and the order of reaction using CoatsRedfern kinetic model. Therefore, the present work attempts to explore the traditional cumbersome isothermal gravimetry under varying holding time and
temperature. The study also involves the preparation of single alum from local clay and the calcination of the produced alum at various temperatures and holding times. Gravimetric determination was employed in monitoring the overall rate data, while Xray Fluorescence analyses of the residual sulfate
dX = dX . dT = dX . 9
dt dT dt dT
Substituting Equation 9 into 1 and /, the subject
of the formula
concentration in the alumina were obtained to model sulphate decomposition specifically.
dX = dT
dX . 1
=
X 10
dt
Derivation of CoatRedfern Model
In general, for the reaction
dX = X dT 11
0
+ the disappearance of component A can be described by the formal kinetic expression:
g X = () = X dT X
12
dX g X = =
13
dt = kf X 1a
dX = f X 1b
dt
From Arrhenius,
From calculus:
Therefore, if
dX = 1 a
=
2
=
= 1
Where X is the fractional conversion; t is the time; A is the preexponential factor; E is the activation energy; R is the gas constant; T is the temperature in Kelvin and f (X) is the kinetic function which takes different forms depending on the particular reaction rate equation. In isothermal kinetic studies, the rate equation used to calculate the rate constant has the
form:
Differentiating the above with respect to temperature T gives
dx dt = T 2 dT = T2dx
Substituting dT = T 2dx into Equation13:
= 2 =
14
g X = t 3
0
0 2
Differentiating with respect to time t
d g X = 4
dt
Substituting the LHS term of Equation 4 into Equation 1a to
Since dx = 1 a
Then using integration by part
= 15
give
d g X f X =
5
Let = 1 2
= , then from Equation 14
dt
d g X X = 6
Making dg( X ) subject of the formula
0
2
= 1 2
0
+ 2 1 3 16a
g X =
7
0 0
= + 2 16b
Integrating with respect to X
0 2
2
0
3
g X = ()
0
8
Taken the integration by part a step furtherand rewrite Equation 16
However, nonisothermal methods are becoming more widely
used because they are more realistic than the classical isothermal methods. In nonisothermal kinetics the time
2
+ 2 .
2 3
2 +
3 =
3
17
0
dependence on the left hand side of Equation 1 is eliminated
dt
using constant heating rate = dT so that T = To + , where To
0 +
4
is the starting temperature and t is the time of heating. Using integral methods of analysis, from Equation 1:
Its obvious that the integration continues endlessly, however, from the generic definition of x = 1/T, x4 and above goes to
zero, rendering the equation undefined. Therefore, it becomes expedient to stop at the second term on the right hand bracket.
The equation has been written in the form:
Hence,
=
2
1 1 X 1n
T2 1 n
2
2
= In R 1 2RT
E RT (27)
2
+
a 0
3
+
2
3 .
18
E E +
Substituting a = and x = 1 T
gives Equation 19
into Equation 18
Equation 27 satisfies for n< 1or n > 1 but for n=1, Equation 27 becomes
= 1 X
E
RT2 T
= e RT
2R2T3
+
T
eE RT
19
2
T2
= In R 1 2RT E
(28)
E
E2
E E +
RT
And then factorize Equation 19 to give Equation 20
E
= RT2 eE RT 1 2RT 20
For Equations 27 and 28 to be amenable to graphical solution, the quantity 1 2RT is assumed to be close to 1, hence the need to verify the reasonability of this assumption [10].
E E
E E
= RT2eE RT 1 2RT 21
Incorporating the power law Equation 21 as follows:
X
n dX

MATERIALS
Raw kaolin used in this investigation was obtained from Kankara village, Katsina State, Nigeria, while beneficiated and metakaolin were obtained from processing of the raw clay
X =
1 X , then g X =
0
X
and calcination of the beneficiated clay respectively. Fresh alum produced in this work as described in the following
0
dX 1 X n
X
= 1
n
0
X
dX (22)
subsection meets general requirement of 69% wt of Al2O3/ wt of fresh alum. All other chemicals used were laboratory grade.
If Equation 22 was integrated to give:
EXPERIMENTAL PROCEDURE
1 1 X 1n
Preparation of Metakaolin from raw Kankara kaolin clay
=
0
1 n (23)
Substituting Equation 23 into Equation 21
1 1 X 1n
The raw clay was crushed using a mortar and pestle. The resultant product was then beneficiated by soaking in water and intermitent vigorious stirring for 3 days, after each day the spent water was replaced while sand particles were removed
=
1 n
RT2 1 2RT E E
= E e
RT
(24)
and discarded. The significance of removal of spent water is to faciliate the removal of soluble impurties. The kaolin suspension was then centrifuged and dried overnight at 120oC to remove free water [11]. The dried lump was crushed and screened with 315 microns sieve. The sieved clay powder was
Dividing Equation 24 through by T2
then weighed into crucibles and calcined in a muffle furnace at a temperature of 750oC for 2 hrs [2].
=
2
1 1 X 1n
T2 1 n
= R E 1 2RT E eE RT
(25)
Dealumination of metakaolin using sulphuric acid
The dealumination of metakaolin (Al2Si2O7) was performed by reacting 50 grammes of metakaolin with 168.03cm3 of
Taking the negative natural logarithm of both sides
sulphuric acid of 96wt% (H2SO4) to give a 60 wt% acid solution [12]. A simplified chemical reaction for dealumination process is presented as Equation 29:
=
2
1 1 X 1n
T2 1 n
+ 3 + 24
R
2RT
2 2 7 ()
2 4 ( ) 2
= In E 1
E + E RT (26)
2 4 3 .272() + 22 (29)
Distilled water (184.27cm3) was added to quench the reaction and to enhance the seperability of the alum ladden aliquot
2
1 1 X 1n
T2 1 n
1 T
(31)
from the residual silica. The crystal yield obtained from 50ml of filtrate was observed to increase with decrease in temperature. This is in agreement with the established fact that the solubility of alum decreases at reduced temperatures [13,14]. The crystallization stage was achieved by cooling at – 10oC for 3 hours to initate the growth of alum crystals and resultant content is then filtered [2]. The hydrated alum crystals (Al2(SO4)3.27H2O) were gradually dried in an oven between 40oC and 160oC for 12 hrs. to enhance size reduction and seperation, after which the resultant sample is heated at 350oC for 5 hours to remove the occluded acid and the chemically bonded water from the alum [9]. The dried alum was then ground and sieved with 315 microns [2]. Table 2 presents the compositional analysis for fresh alum and pretreated alums at 160oC and 350oC.
Calcination of the Pretreated Alum
40 grammes of ground alum in a crucible was calcined in a muffle furnace at 750oC, 800oC, 810o C, 850o C and 900oC and for holding time intervals of 20,60, 120,150 and 180 minutes respectively. The residues were weighed and also subjected to Xray flurorescence analysis after cooling to room temperature in a desiccator. A simplified decomposition reaction for aluminum sulphate is as shown in Equation 30.
2 4 3 () 23() + 3() (30)
KINETIC PARAMETER ESTIMATION USING COATSREDFERN MODEL
Kinetic parameter estimation using thermogravimetric data
The weight losses obtained were converted to conversion at different temperatures and time interval as shown in Table 5. The calculated conversion and temperature values were inputted into the CoatsRedfern model. Plot of
T2 1n T
1 1X 1n vs 1 were made. Regressing the left side of the
above equation using the least square criteria for different values of n, the slope is E/R and the intercept is equal to
R E; from these various plots, activation energy, E and the preexponential term, A were evaluated [9].
Discriminatory test for X ray Fluorescence results
The samples were then characterize using a Energy Dispersive
Xray Fluorescence machine to obtain the elemental composition by monitoring the extent of conversion of sulphate decomposition.
RESULTS AND DISCUSSION
The raw clay, beneficiated clay and metakaolin
component analysis
From the results shown in Table 1, it was observed that the alumina content of the beneficiated clay increased significantly over the raw clay from 41.74 wt % to 43.01 wt.
% while the silica content slightly decreased from 54.56 wt % to 53.95 wt %.
TABLE 1: CHEMICAL COMPOSITION OF RAW CLAY, BENEFICIATED CLAY AND METAKAOLIN
Component
Raw Clay
%
Beneficiated
clay %
Metakaolin
%
Al2O3
41.74
43.01
44.87
SiO2
54.56
53.95
52.41
CaO
0.44
0.34
0.32
Fe2O3
0.53
0.58
0.53
MgO
1.74
1.31
1.22
K2O
0.71
0.56
0.47
Na2O
0.04
0.05
0.04
TiO2
0.10
0.07
0.07
Cr2O3
0.01
0.01
0.01
Mn2O3
0.02
0.02
0.01
NiO
0.01
0.03
0.02
CuO
0.08
0.06
0.03
ZnO
0.00
0.01
0.00
This could be attributed to the loss of organic material and free silicia respectively during the process of beneficiation. The color of the calcined kaolin changed from white to brick red indicating the presence of Fe3+ [15].
1 1 X 1n
R
2RT E
2 =
T2 1 n = In E 1
E +
RT (26)
CoatsRedfern model was adopted to estimate the kinetic parameters such as activation energy, preexponential factor and the order of the reaction were obtained by plotting a graph [9]
Effect of drying on alum
Drying at 160oC evidently was able to drive off the adsorbed free water on the alum and possibly part of the occluded water, while heating to 350oC was believed to be potent enough not only to eliminate the remaining occluded water but also the occluded excess acid as well. Table 2 showed these effects resulting in increase in concentration of alumina in the two heating steps; while SO2 increased slightly in the first step, evidence of loss of free water and dropped partially in the second step, indicating loss due to occluded acid evacuation.
TABLE 2: ELEMENTAL COMPOSITION OF FRESH ALUM, THERMAL PRETREATED ALUMS AT 160OC AND 350OC
Components
Alum
Fresh Alum
Dried at 160oC
Dried at 350oC
Al2O3
8.41
8.7
16.02
SO3
60.4
61.4
60.81
CaO
0.46
0.38
0.2
Fe2O3
0.18
0.24
0.14
MgO
0.17
0.02
0.31
K2O
–
0.15
0.01
Na2O
0.03
0.09
0.11
TiO2
0.08
–
0.06
NiO
–
–
0.01
This observation is consistent with Moselhy et al. (1994) [9] differential thermal analysis (DTA) on a hydrated aluminium sulfate. The pretreated alum at 160oC comprised of about 8.41 wt/wt % of alumina which agrees with Alan et al, (2000) [16] that liquid alum contains about 8 wt/wt%.
Calcination of the partially dried alum (at 350oC)
The residual weight of the alum samples (in percentage of the initial quantity) obtained after calcination at different
temperatures and time intervals are shown in Figure 1. Figure 2 shows the percentage of sulphate decomposed at various temperatures from thermogravimetric analysis. From Figures
1 and 2, it could be observed that the minimum residual weight of alum samples stood at 12.67g, while the maximum sulfate decomposition stood at 97.37%. The two figures clearly illustrated that as the temperature increased from 750 900oC, the sensitivity of the decomposition reaction increased, this is evident by steeper initial rates which flattens out as the reaction progressed from 20 mins. to 180 mins.
KINETIC PARAMETER ESTIMATION USING COATSREDFERN MODEL
Kinetic parameter estimation from thermogravimetric data
The TGA data obtained from the decomposition of the pretreated alum were used to generate values to plot In g(X) against 1/T, by varying the order n, between 0 and 2. Figure 3 is representative of such CoatsRedfern plots for estimating the kinetic and thermodynamic parameters.
The summary of the regression analysis and the obtained kinetic parameters are shown in Table 3, indicating that reaction order of 1.3 for TGA gave the best average regression value of 0.9446 as compared with other orders within the range of 0 to 2 tested. The estimated average activation energy and preexponential factor corresponding to the reaction order
1.3 were 321.03 kJ/mol K and 2.57E+14 min1 respectively.
TABLE 3: AVERAGE VALUES COMPARISION OF REGRESSION ANALYSIS AS OBTAINED FROM COATSREDFERN METHOD OF ESTIMATING KINETIC PARAMETERS FROM TGA DATA
Order n
Equation y
Kinetic parameters
Regression
value R2
Activation
energy kJ/molK
Pre
exponential factor / mins
0.0
y = 2.386x 6.241
0.884
198.33
1.20E+08
0.5
y = 2.811x 10.412
0.9178
233.69
8.28E+09
0.67
y = 2.992x 12.168
0.9274
248.77
5.00E+10
1.0
y = 3.408x 16.163
0.9407
283.13
3.01E+12
1.2
y = 3.702x 18.983
0.9442
307.81
5.41E+13
1.3
y = 3.861x 20.504
0.9446
321.03
2.57E+14
1.4
y = 3.863x 22.097
0.94416
334.90
1.31E+15
1.5
y= 4.202x23.7587
0.9272
349.37
7.18E+15
1.6
y= 4.383x24.321
0.9409
364.43
4.2E+16
1.7
y=4.5709x27.274
0.9384
380.03
2.61E+17
1.8
y=4.765x29.118
0.9353
396.13
1.71E+18
1.9
y= 4.964x31.02
0.9317
412.71
1.19E+19
2.0
y= 5.169x32.97
0.9278
429.71
8.68E+19
Effect of reaction time on thermodynamic parameters values
Table 4 revealed that as decomposition reaction progresses as a function of time while the order of the reaction was held constant at n= 1.3, the thermodynamic parameters: activation energy and preexponential factor experiences decrease in their values. This may be attributed to decrease in the barrier to heat and mass transfer due to increase in voidage in the sample matrix as the reaction progresses. The higher activation energy at the early stage of calcinations could also be attributed to the energy required for hydroxylation process coupled with sulfate decomposition, while at later times, sulfate decomposition being the sole reaction, which requires lesser energy for its transformation.
TABLE 4: TYPICAL VARIATION OF THE THERMODYNAMIC PARAMETERS ESTIMATED WITH REACTION HOLDING TIME FROM TGA DATA AT CONSTANT REACTION ORDER OF 1.3
Time/ mins
Equation y
Regressio n value R2
Activation
energy kJ/molK
Pre
exponential factor / mins
20
0.9410
333.50
5.79E+14
60
y = 3.7744x 19.406
0.9432
313.80
8.09E+13
120
y = 3.6326x 18.535
0.9461
302.01
3.26E+13
150
y = 3.5632x 18.047
0.9513
296.24
1.96E+13
180
y = 3.5296x 17.817
0.9415
293.45
1.54E+13
Kinetic parameter estimation from Xray Fluorescence data Figure 4 is a representative of the CoatsRedfern plots for estimating kinetic parameters for reaction orders of 0, 0.5, 1.0,
1.3 and 1.5 respectively from XRF conversion data on sulfate decomposition.
The summary of the regression analysis and the obtained reaction parameters are shown in Table 5, where order of 1.0 for XRF gave the best average regression value of 0.9818.
TABLE 5: AVERAGE VALUES COMPARISION OF REGRESSION ANALYSIS AS OBTAINED FROM COATSREDFERN METHOD OF ESTIMATING KINETIC PARAMETERS FROM XRF DATA
Order N
Kinetic parameters
Equation y
Regression value R2
Activation
energy kJ/molK
Pre
exponential factor / mins
0.0
y = 2.431x 6.400
0.8619
202.14
1.38E+08
0.5
y = 3.246x 13.468
0.9698
269.85
2.48E+11
0.67
y = 3.494x 15.744
0.9758
290.50
2.61E+12
1.0
y = 4.056x 20.834
0.9818
337.22
5.27E+14
1.3
y = 4.509x 25.266
0.9238
374.86
5.47E+16
1.4
y = 4.739x 27.322
0.9244
394.01
4.79E+17
1.5
y = 5.114x 30.281
0.9780
425.19
1.08E+19
1.6
y= 5.229x 31.680
0.9241
434.70
4.78E+19
1.7
y=5.487x33.976
0.9235
456.17
5.41E+20
1.8
y=5.753x36.342
0.9225
478.34
6.59E+21
1.9
y=6.028x38.776
0.9213
501.16
8.64E+22
2.0
y=6.592x43.492
0.9731
548.04
1.28E+25
The estimated activation energy and preexponential factor for the best average regression analysis of order 1.0 were 337.22 kJ/mol K and 5.27E+14 min1 respectively. Table 6 clearly shows that as the calcinations progressed, the activation energy and preexponential factor decreased as the reaction order was held at n=1. This reduction in the activation energy and frequency factor could be attributed to decrease in the diffusion barrier to heat and mass transfer resulting from increase in voidage as the reaction progresses.
TABLE 6: TYPICAL KINETIC PARAMETERS VARIATION WITH REACTION TIME FROM XRF DATA
Time
/ mins
Equation y
Kinetic parameters
Regression value R2
Activation energy kJ/molK
Pre exponential
factor / mins
60
y = 4.270x 16.42
0.9726
355.01
3.76E+15
120
y = 4.144x 22.344
0.9977
344.52
1.68E+15
180
y = 3.754x 19.192
0.9752
312.12
6.50E+13
From Table 7, the kinetic and thermodynamic parameters obtained from XRF data were slightly higher than those from TGA data. This can be attributed to the fact that X ray fluorescence analysis monitors the sulfate decomposition process only, while that of TGA is more or less average of several reactions including the effect of other metallic sulfates acting as impurities (Table 2) present in the prepared alum which could decrease the overall observed activation energy. It worth mentioning that the kinetic parameters obtained from the XRF data were in good agreement with that of Moselhy et al. from derivatograph data using the standard equipment, dynamic thermogravimetric method for data generation (n=1, E
= 348 kJmol1 and A = 1.3E+16 s1)
Validation of CoatsRedfern model
To validate the CoatsRedfern model, the assumption of the term 1 2 1. It was observed that after substituing
typical activation energy , gas constant R, and temperature T.
The fraction 2RT/E 0.05, was slightly insignificant resulting in a value of the term in the bracket approximately equal to 1. Thus validating the assumption.

CONCLUSION
From this study, it can be concluded that reaction orders of 1.0 and 1.3 were most acceptable for XRF data and TGA data respectively; having activation energy and preexponential factor of 337.22 kJ/molK (5.27E+14 min1) and 321.03 kJ/molK (2.57E+14 min1) respectively. The large activation energy indicates that the reaction is temperature sensitive. It was observed that as the calcination progressed while the reaction order was held constant there was a reduction in the activation energy and preexponential factor for both TGA and XRF results. This can be attributed to decrease in the diffusion barrier to heat and mass transfer most likely due to increase invoidage as the reaction progresses. Based on the assumption considered for CoatsRedfern model, the model was found to fit the rate data obtained satisfactorily.
REFERENCES

Raw Materials Research and Development Council (2000), Raw Materials and Consumer industries in Nigeria, www.rmrdc.org.

Abdul, B. Aderemi B.O. and Ahmed T.O. (2009): Production of alumina from Kankara kaolinite clay for electrical insulation applications. Int. J. Sci. & Techn. Research 6(1&2) pp107114
EdomwonyiOtu, L.C; Aderemi, B.O. and Ahmed A.S. (2010): Beneficiation of Kankara kaolin for alum production. Nig. J. of Eng. 16(2) 2735

EdomwonyiOtu, L.C; Bawa, S.G. and Aderemi, B.O. (2010): Effect of Beneficiation of kaolin raw material on alumina yield and quality. Nig.
J. of Eng. 16(2) 3643

EdomwonyiOtu L.; Aderemi B.O and Ofoku, A.G. (2010): Studying the effect of calcination temperature and dealumination time on alum yield from Kankara kaolin. Afri. J. Nat. Sci. 13 pp 69 74

EdomwonyiOtu L.C.; Aderemi B.O.; Simo A. and Maaza M. (2012): Alum production from Nigerian kaolinite deposits. Int. J. Eng Res. in Africa 7 pp 13 19 Switzerland.

EdomwonyiOtu L.C., Aderemi B.O., Ahmed A.S., Coville N.J. and Maaza M. (2013): Application of alum from Kankara kaolinite in catalysis: A preliminary report. Ceramic Transactions 240 pp 167 174

EdomwonyiOtu L.C., Aderemi B.O., Ahmed A.S., Coville N.J. and Maaza M. Influence of thermal treatment on Kankara kaolinite Opticon1826 15(5) 15 (2013)

Moselhy H.J.; Madarasz, G; Pokol, S; Pungor, E.(1994). Aluminum Sulphate hydrates: Kinetics of the Thermal Decomposition of Aluminum Sulphate using different calculation methods Journal of Thermal Analysis 41:1,25.
 <>Coats, A. W., Redfern, J. P. (1964): Kinetic parameters from Thermogravimetric data. Nature Vol. 201. pp. 6869.

Aderemi, B.O, Oloro, E.F, Joseph, D, Oludipe, J. (2001) Kinetics of the Dealumination of Kankara Kaolin Clay, NJE vol.9 No1, pp 4044.

Aderemi, B.O. and Oludipe, J.O. (2000). Dealumination of Kankara Kaolin ClayDevelopment of Governing Rate Equation. Nig. J. of Eng. 8(2) 2230

MacZura G, Goodboy, K.P.; Koenig, J.J. (1978) Aluminium Sulfate and Alums in KirkOthmer (Ed), Encyclopedia of Chemical Technology, Wiley Interscience New York, Vol 2, pp 245250.

Perry R. H., Green, D. W. (1997) Perrys Chemical Engineering Handbook. McGrawHill Book Co. Pp 28 36, 2843.

Vogel A, I. (1999) Quantitative Inorganic Chemical Analysis. 4th edition pp.354.

Alan C., Ratnayaka D., Malcolm J. B., (2000), Water supply Technology & Engineering pp. 676.