Embed Size (px)
Citation preview
Univers
ity of
Cap
e Tow
nTHE SYNTHESIS OF ALUMINIUM HYDROXIDE AND OXYHYDROXIDE
by
Linda-Ann Louw
A thesis submitted for the degree of Master of Science in Applied Science to the Faculty of Engineering at the University of Cape Town.
Department of Civil Engineering University of Cape Town
March 1993
THE SYNTHESIS OF ALUMINIUM HYDROXIDE AND OXYHYDROXIDE
by
Linda-Ann Louw
A thesis submitted. for the degree of Master of Science in Applied Science to the Faculty of Engineering at the University of Cape Town.
Department of Civil Engineering University of Cape Town
March 1993
Univers
ity of
Cap
e Tow
n
The copyright of this thesis vests in the author. No quotation from it or information derived from it is to be published without full acknowledgement of the source. The thesis is to be used for private study or non-commercial research purposes only.
Published by the University of Cape Town (UCT) in terms of the non-exclusive license granted to UCT by the author.
Univers
ity of
Cap
e Tow
n
J
THE SYNTHESIS OF ALUMINIUM HYDROXIDE AND OXYHYDROXIDE POLYMORPHS
by
Linda-Ann Louw Department of Civil Engineering
University of Cape Town
January 1993 I.
Beta alumina is a ceramic with the ability to conduct certain ions. It is conventionally formed through the high temperature solid state reaction of a-alumina with soda and lithia, which results in a mixture of 6- and 6"-alumina phases. To maximise the amount of 6 "-alumina formed, a second heat treatment step is required. Beta alumina may also be synthesised directly from the various aluminium hydroxide ( Al(OH)J ) and aluminium oxyhydroxide (AIOOH) polymorphs. The particular oxygen sublattice of the starting material is critical in determining which phases of beta alumina are formed. Boehmite and bayerite are known to yield pure B" -alumina.
In this investigation, the differences in structure and the identification of the polymorphs is discussed. A review of the methods of synthesis and concomitant mechanisms of reaction proposed in the literature are presented. Of these, the synthesis of boehmite and bayerite are investigated experimentally.
Attempts to synthesise boehmite at ambient conditions resulted in the formation of pseudoboehmite, and confirmed that boehmite can only be formed by the introduction . of a hydrothermal step. However, it was possible to form bayerite at ambient conditions by sµnple chemical reaction. Control of the reaction pH and the removal of inhibiting ions were found to facilitate the formation of well crystallised bayerite. Although the factors governing the successful synthesis of bayerite need further investigation, the possibility of a direct synthesis route for fi "-alumina has been demonstrated.
)
Finally, the synthesis of 6 "-alumina by the solid state reaction of the hydroxide and oxyhydroxide polymorphs was successfully completed.
J
THE SYNTHESIS OF ALUMINIUM HYDROXIDE AND OXYHYDROXIDE POLYMORPHS
by
Linda-ADD Department of Civil Engineering
University of Cape Town
January 1993 I'
alumina is a the ability to certain ions. It is conventionally t-........ n&>L'll through the solid state of a-alumina with soda and
which results in a mixture of S- and S"-alumina phases. To the amount of SrI-alumina a second heat treatment step is required. alumina
also be synthesised directly from the various aluminium hydroxide ( AI(OH)] ) and aluminium oxyhydroxide (AlOOR) polymorphs. The particular sub lattice of starting material is critical in determining which phases of beta alumina are
Boehmite and bayerite are known to yield pure 8" -alumina.
investigation, in identification polymorphs is discussed. review of the methods of synthesis and COlllcollmtant mechanisms of reaction proposed in the literature are presented. Of these, the synthesis of boehmite and bayerite are investigated experimentally.
Attempts to synthesise boehmite at ambient conditions resulted in the formation of pseudoboehmite, and that boehmite can only be formed by the introduction . of a hydrothermal ,However, it was to form bayerite at amibleJnt
conditions hy ~le reaction. the reaction pH and removal of inhibiting ions were found to facilitate the of well crystallised baverite. Although the factors governing the successful synthesis of bayerite investigation, the possibility of a direct synthesis route for 6" -alumina demonstrated.
J
the synthesis of S .. -alumina by the solid reaction of the hydroxide and OX'1fDyarCJIXI<lle ..... "'h,'m""1''nh~ was successfully completed.
Univers
ity of
Cap
e Tow
n
TABLE OF CONTENTS I .INTRODUCTION 1
2. THE ALUMINIUM HYDROXIDE AND OXYHYDROXIDE PRECURSORS 3 2.1 NOMENCLATURE 3 2.2 STRUCTURE OF THE ALUMINIUM HYDROXIDES 4
2.2.1 Gibbsite 4 2.2.2 Bayerite 6 2.2.3 Nordstrandite 7
2.3 STRUCTURE OF THE OXIDE HYDROXIDES 7 2.3.1 Diaspore •7 2.3.2 Boehmite 8 2.3 .3 Pseudoboehmite 9
2.4 METHODS OF IDENTIFICATION 10 2.4.1 X-ray Diffraction 10 2.4.2 Thermal Analysis 12 2.4.3 Other Methods 16
3.MECHANISMS OF FORMATION OF ALUMINIUM HYDROXIDES AND OXYHYDROXIDES 17
3 .1 Review of synthesis procedures 17 3 .1.1 Aluminium hydroxides 17 3.1.2 Aluminium oxide hydroxides 19
3.2 Mechanisms of formation 21 3.2.1 Formation of the structural units 21 3.2.2 Formation of the Hydroxides 23 3.2.3 Formation of the Oxyhydroxides 32 3.2.4 Discussion of the proposed mechanisms 35 3.2.5 Summary 37
4.SYNTHESIS OF BOEHMITE AND BA YERITE 39 4.1 THE SYNTHESIS OF BOEHMITE 39
4.1.1 Effect of reaction temperature 39 4.1.2 Effect of ageing 41 4.1.3 Discussion 42
4.2. THE SYNTHESIS OF BA YERITE 43 4.2.1 Effect of final pH 43 4.2.2 Effect of washing 43 4.2.3 Effect of ageing 44
4.3 RESULTS 44 4.3.1 Changes in pH 44 4.3.2 Effect of final pH 44 4.3.3 Effect of washing 46 4.3.4 Effect of ageing 46
4.4 DISCUSSION AND CONCLUSIONS 48
TABLE OF CONTENTS LINTRODUCTION 1
2.THE ALUMINIUM HYDROXIDE AND OXYHYDROXIDE PRECURSORS 3 1 NOMENCLATURE 3
2.2 STRUCTURE OF tHE ALUMINIUM HYDROXIDES 4 2.2.1 Gibbsite 4
Bayerite 6 Nordstrandite 7
2.3 STRUCTURE OF OXIDE 7 1 Diaspore # 7
2.3.2 Boehmite 8 2.3.3 Pseudoboehmite 9
METHODS OF IDENTIFICATION 10 1 X -ray Diffraction 10
Thermal Analysis 12 Other Methods 16
3.MECHANISMS OF FORMATION OF ALUMINIUM HYDROXIDES AND OXYHYDROXIDES
1 Review of synthesis procedures 3.1.1 Aluminium hydroxides
1.2 Aluminium oxide hydroxides 19 3.2 Mechanisms of formation 21
3.2.1 Formation the structural units 21 3.2.2 Formation the Hydroxides 23
Formation the Oxyhydroxides Discussion of the proposed me::l1aIrnSIlrlS
3.2.5 Summary 37
4. SYNTHESIS BOEHMITE AND 39 4.1 THE SYNTHESIS OF BOEHMITE 39
4.1.1 Effect of reaction temperature 39 4. Effect ageing 4.1 Discussion 42
4.2. THE SYNTHESIS OF BAYERITE 43 4.2.1 of final pH 43 4.2.2 of washing 43
Effect of 44 4.3 RESULTS 44
4.3.1 Changes pH 4.3.2 Effect final pH 44 4.3.3 of washing 46 4.3.4 of' 46
4.4 DISCUSSION AND CON SIONS 48
Univers
ity of
Cap
e Tow
n
5.SYNTHESIS OF 6"-ALUMINA 5 .1 Synthesis
5 .1.1 Composition 5 .1.2 Powder preparation 5.1.3 Solid state reaction 5 .1. 4 Density determination 5.1.5 X-ray determination
5~2 RESULTS : 5 .2 .1 Observations
5.2.2 Density 5.2.3 X-ray Diffraction Analysis
5 .3 DISCUSSION
6.CONCLUSIONS
REFERENCES
i) Boehmite precursor ii) Bayerite precursor
50 50 50 50 51 51 51 56 56 56 59 59 59 64
65
67
5.SYNTHESIS 8"-ALUMINA 5.1 Synthesis
5 .1.1 Composition 5.1 Powder preparation
1.3 Solid state reaction 1 Density determination
5 .1.5 X-ray determination
1 Observations 5.2.2 Density
X-ray Diffraction Analysis
5.3 DISCUSSION
6. CONCLUSIONS
REFERENCES
i) Boehmite precursor ii) precursor
50 50
50 51 51 51 56 56 56
59 64
67
Univers
ity of
Cap
e Tow
n
TABLE OF FIGURES
CHAPTERl Fig.l Section through the idealised structure of 2
.6-A120 3 and .6"-A120 3 showing the stacking sequence along the c-axis.
CHAPTER2: Fig. 2.1 Hydroxyl bonding betwee~ sp3 hybridised 5
orbitals #
Fig. 2.2 ABBA-ABBA stacking in gibbsite 5
Fig. 2.3 Crystal Morphology 6
Fig. 2.4 ABAB-ABAB stacking in Bayerite 7
Fig. 2.5 ABBA-BABA stacking in Nordstrandite 7
Fig. 2.6 Schematic representation of OH-Al-0 chains 8
Fig2.7 Schematic representation of the crystal 9 structure (a) Diaspore (b)boehmite
Fig. 2.8 Comparative XRD patterns for aluminium 11 hydroxide and oxyhydroxide polymorphs
Fig. 2.9 DTA curves of aluminium hydroxides and 13 oxyhydroxides.
Fig. 2.10 DT A curves of NIMR standard samples 14
CHAPTER3: Fig. 3.1 Proposed structure of an hydroxy-aluminium 22
.. polymer having an average positive charge of 1 per Al-atom
Fig. 3.2 Variation in pH with time. 25
Fig. 3.3 Hydroxyl bridging mechanism of edge-Al is 26 hampered by the citrate.
Fig. 3.4 Effect of final pH and salt concentration on 27 polymorph formation.
TABLE OF FIGURES
CHAPTER 1 1
1
Fig. 2.2
Fig. 2.4
2.5
Fig.
Section through the idealised structure of S-AI20 3 and S"-AI20 3 showing the stacking sequence along the c-axis.
Hydroxyl bonding betwee~ sp3 hybridised orbitals
ABBA-ABBA stacking in gibbsite
Crystal Morphology
ABAB-ABAB ",,,,,,,.All,,#; in Bayerite
ABBA-BABA stacking in Nordstrandite
Schematic representation of ........... 4 .... chains
Schematic representation of the crystal structure (a) Diaspore (b)boehmite
Fig. 2.8 Comparative XRD patterns for aluminium hydroxide and oxyhydroxide polymorphs
Fig. 2.9 DTA curves of aluminium hydroxides and oxyhydroxides.
2.10 DTA curves of NIMR standard samPles
Fig. 3.1
Fig. 3.2
Fig. 3.3
Fig.
3: Proposed structure
.. polymer having an
of 1 per AI-atom
an hydroxy-aluminium positive charge
in pH with time.
Hydroxyl bridging mechanism of edge-AI is hampered by the citrate.
of final and salt concentration on polymorph formation.
2
5
5
6
7
7
8
9
11
13
14
25
26
27
Univers
ity of
Cap
e Tow
n
Fig.3.5 Effect of OH/ Al ratio and salt concentration 28 on polymorph formation.
Fig.3.6 Effect of temperature and ageing time on the 28 concentration of solid material, mononuclear ions and polynuclear ions.
Fig.3.7 Average degree of polymerization versus 30 OH/ Al ratio.
#
Fig.3.8 Mechanism of breakdown of polymers by acid. 31
Fig.3.9 Equilibrium diagram for the system AI20 3-H20 33
Fig.3.10 Equilibrium diagram for the system AI20 3-H20 33
Fig.3.11 Composition seque~ce of aluminium hydroxide 34 and oxyhydroxide polymorphs
CHAPTER4:
Fig.4.1 XRD Spectrum of pseudoboehmite obtained by reaction at 25°C and 1 oo<>c. 40
Fig.4.2 XRD spectra of batches aged at 200C and 700C. 41
Fig.4 3 XRD spectra Sample A - Reaction precipitate. Sample B - Reaction precipitate, pH adjusted. 45
Fig.4 4 XRD spectrum of Sample C Reaction precipitate, pH adjusted 45
Fig.4.5 XRD spectrum of Sample D Reaction precipitate, pH adjusted, washed 46
Fig.4 6 XRD spectrum of Sample E Reaction precipitate, pH adjusted, washed 47
Fig.4.7 XRD spectrum of Sample F Aged 3 days, then as for Sample E. 47
Fig.3.5 Effect of OH/ AI ratio and salt concentration 28 on polymorph formation.
Fig.3.6 Effect of temperature and ageing time on the 28 concentration of solid material, mononuclear ions and polynuclear ions.
Fig.3.7 Average degree of polymerization versus 30 OH/AI ratio.
Fig.3.8 Mechanism of breakdown of polymers by acid. 31
Fig.3.9 Equilibrium diagram for the system AI2OrH2O 33
Fig.3.10 Equilibrium diagram for the system AI20 3-H2O 33
Fig.3.11 Composition seque~ce of aluminium hydroxide 34 and oxyhydroxide polymorphs
CHAPfER4:
Fig.4.1 XRD Spectrum of pseudoboehmite obtained by reaction at 25°C and 1000C. 40
Fig.4.2 XRD spectra of batches aged at 200C and 700C. 41
Fig.4 3 XRD spectra Sample A - Reaction precipitate. Sample B - Reaction precipitate, pH adjusted. 45
Fig.4 4 XRD spectrum of Sample C Reaction precipitate, pH adjusted 45
Fig.4.5 XRD spectrum of Sample D Reaction precipitate, pH adjusted, washed 46
Fig.4 6 XRD spectrum of Sample E Reaction precipitate, pH adjusted, washed 47
Fig.4.7 XRD spectrum of Sample F Aged 3 days, then as for Sample E. 47
Univers
ity of
Cap
e Tow
n
I
CHAPTERS:
Fig 5.1 XRD pattern of (a) 6"-alumina and (b) 13-alumina 52
Fig 5.2(a) XRD pattern of standard .6-alumina. 55
Fig 5.2(b) XRD pattern of standard 13 "-alumina. 55
Fig. 5.3 Density vs Time - Timcerm. 57 #
,' Fig. 5.4 Density vs Time - Cera Hydrate. 57
Fig. 5.5 Density vs Time - Bayerite. 58
Fig. 5.6 6"-alumina from boehmite - Cera Hydrate. 60
Fig. 5.7 B "-alumina from boehmite - Timcerm. 61
Fig. 5.8 .B "-alumina from Bayerite - 4 hour soak. 62
Fig. 5.9 6 "-alumina from Bayerite - 1 hour soak.' 63
Fig. 5.10 Mixed 6 " - and .6-alumina from chemically precipitated bayerite 63
CHAPTER 5:
Fig 5.1 XRD pattern of (a) fi"-alumina and (b) ./3-alumina 52
Fig 5.2(a) XRD pattern of standard fi-alumina. 55
Fig 5.2(b) XRD pattern of standard ./3'" -alumina. 55
Fig. 5.3 Density vs Time - Timcenn. 57
, Fig. 5.4 Density vs Time - Cera Hydrate. 57
Fig. 5.5 Density vs Time - Bayerite. 58
Fig. 5.6 ./3"-alumina from boehmite - Cera Hydrate. 60
Fig. 5.7 6" -alumina from boehmite - Timcerm. 61
Fig. 5.8 6" -alumina from Bayerite - 4 hour soak. 62
Fig. 5.9 6" -alumina from Bayerite - 1 hour soak. 63
Fig. 5.10 Mixed ./3"- and ./3-alumina from chemically precipitated bayerite 63
Univers
ity of
Cap
e Tow
n
• Aluminium
f} Sodium
Q Oxygen
t3·Alumina 13"· Alumina
Fig.1: Section through the idealised structure of 6-Al20 3 and 6 n -Al203 showing the stacking sequence along the c-axis. (after Mosely 1985)
2
C
A
A'
C
e
A
c'
A
B
.. Alumini",m
o Sodl",m
0 0 ')'\111'1
l3'Alumino
C
A
B
C
,:
B
C
A
B
c'
A
B
13"· Alumina
1: Section through the idealised structure of S-A120 3 and S"-A120 3 showing the stacking sequence along the (after Mosely 1985)
2
Univers
ity of
Cap
e Tow
n
3
. 2.THE ALUMINIUM HYDROXIDE AND OXYHYDROXIDE PRECURSORS.
2.1 NOMENCLATURE.
Many crystalline compounds of aluminium and oxygen occur in nature. Apart from aluminium oxide (A120 3), known as corundum, there exists a range of aluminium hydroxides and oxide-hydroxides. All occur naturally in soils or in bauxitic deposits, though not to the same extent.
The primary motivation for study of these compounds has come from geochemists and soil scientists attempting to formulate a theory of the formation of bauxite deposits and the action of such minerals in soil environments, though more recently the commercial potential of alumina catalysts has provided fresh impetus to research. This chapter will describe the range of polymorphs and the means by which they may be identified.
Two main groups of compounds exist: aluminium hydroxides (Al(OH)J) and aluminium oxide hydroxides (AIOOH). These have in the past been referred to as hydrous oxides or hydrates, implying that the structure contains adsorbed water molecules. In fact, the water is bound as Oir ions; consequently the International Committee on Aluminium Hydroxides Nomenclature (ICHAN) (Newsome et al, 1960) recommended that Al(OH)J be referred to as aluminium trihydroxide and AIOOH as aluminium oxide-hydroxide. The latter are also referred to as oxyhydroxides.
The various systems nomenclatures are summarised and compared in Table 2.1 below. Researchers familiar with crystallographic conventions have also used alpha (a)- and gamma (r)- to refer to those hydroxides whose structures are hexagonal close packed or cubic close packed related respectively. Krischner (1971), finding this convention inadequate for the number of transition aluminas identified, grouped the various hydroxides, oxide-hydroxides and their degradation products according to their oxygen stacking sequence: a for ABAB-ABAB; 6 for ABAC-ABAC; r for ABC-ABC. This system has proved thorough in description and useful in predicting properties and behaviour.
3
. 2.THE ALUMINIUM HYDROXIDE AND OXYHYDROXIDE PRECURSORS.
2.1 NOMENCLATURE.
Many crystalline compounds of aluminium and oxygen occur in nature. Apart from aluminium oxide (AI20 3), known as corundum, there exists a range of aluminium hydroxides and oxide-hydroxides. All occur naturally in soils or in bauxitic deposits, though not to the same extent.
The primary motivation for study of these compounds has come from geochemists and soil scientists attempting to formulate a theory of the formation of bauxite deposits and the action of such minerals in soil environments, though more recently the commercial potential of alumina catalysts has provided fresh impetus to research. This chapter will describe the range of polymorphs and the means by which they may be identified.
Two main groups of compounds exist: aluminium hydroxides (AI(OH)3) and aluminium oxide hydroxides (AIOOR). These have in the past been referred to as hydrous oxides or hydrates, implying that the structure contains adsorbed water molecules. In fact, the water is bound as Off' ions; consequently the International Committee on Aluminium Hydroxides Nomenclature (ICHAN) (Newsome et al, 1960) recommended that AI(OHh be referred to as aluminium tribydroxide and AIOOH as aluminium oxide-hydroxide. The latter are also referred to as oxyhydroxides.
The various systems nomenclatures are summarised and compared in Table 2.1 below. Researchers familiar with crystallographic conventions have also used alpha (a)- and gamma (r)- to refer to those hydroxides whose structures are hexagonal close packed or cubic close packed related respectively. Krischner (1971), finding this convention inadequate for the number of transition aluminas identified, grouped the various hydroxides, oxide-hydroxides and their degradation products according to their oxygen stacking sequence: a for ABAB-ABAB; .B for ABAC-ABAC; r for ABC-ABC. This system has proved thorough in description and useful in predicting properties and behaviour.
Univers
ity of
Cap
e Tow
n
Table 2.1: Comparison of Nomenclature Systems (adapted from Van Zyl, 1987).
IC;HAN '
Aluminium trihydroxide Al(OH)3
gibbsite I hydragillite bayerite nordstrandite
Aluminium oxide-hydroxide AlOOH
boehmite diaspore
Previously called
Aluminium trihydrate
Alz03.3HzO
r-aluminium trihydrate a-aluminium trihydrate
----
Aluminium monohydrate AI2o3.H2o.
r-aluminium monohydrate a-aluminium monohydrate
2.2 STRUCTURE OF THE ALUMINIUM HYDROXIDES.
2.2.1 GIBBSITE.
4
~
The structure of Gibbsite was determined by Megaw in 1934. It is monoclinic, having a=8.659, b=5.077, c=9.703 A and 6=94°12' (ASTM 7-324).
The basic structural unit of gibbsite is simply two sheets of hydroxyl ions close-packed and bound together by the aluminium ions which occupy two thirds of the interstices, ie. a dioctahedral arrangement. As the double layers stack, they assume an openpacked structure, each hydroxyl layer lying directly above the one below. Schoen and Roberson (1970) report the following explanation for this:
Firstly the protons of the hydroxyl ions are displaced toward the neutral "holes" left by the dioctahedral filling of interstices. Secondly, the highly polarising effect of Al3+ leads to a tetrahedral hybridisation of the hydroxyl ion. Two aluminium ions bond with two negatively charged lobes, leaving one negative and one positive (with proton) lobe free. These can then bond with another hybridized hydroxyl group from another layer (Figure 2.1). This is called an hydroxyl bond and it leads to each superposed layer of gibbsite lying directly above the one below in an open-packed arrangement.(Figure 2.2). The hydroxyl ions are thus stacked according to the scheme
Table 2.1: Comparison of Nomenclature Systems (adapted from Van Zyl, 1987).
IC~AN ,
Aluminium trihydroxide
AI(OH)3
gibbsite I hydragillite bayerite nordstrandite
Aluminium oxide-hydroxide AlOOH
boehmite diaspore
Previously called
Aluminium trihydrate
A120 3·3H2O
r -aluminium trihydrate a-aluminium trihydrate
----
Aluminium monohydrate
Al20 3·H2O
r-aluminium monohydrate a-aluminium monohydrate
2.2 STRUCTURE OF THE ALUMINIUM HYDROXIDES.
2.2.1 GffiBSlTE.
4
•
The structure of Gibbsite was determined by Megaw in 1934. It is monoclinic, having a=8.659, b=5.077, c=9.703 A and 6=94° 12' (ASTM 7-324).
The basic structural unit of gibbsite is simply two sheets of hydroxyl ions close-packed and bound together by the aluminium ions which occupy two thirds of the interstices, ie. a dioctahedral arrangement. As the double layers stack, they assume an openpacked structure, each hydroxyl layer lying directly above the one below. Schoen and Roberson (1970) report the following explanation for this:
Firstly the protons of the hydroxyl ions are displaced toward the neutral "holes" left by the dioctahedral filling of interstices. Secondly, the highly polarising effect of A13+ leads to a tetrahedral hybridisation of the hydroxyl ion. Two aluminium ions bond with two negatively charged lobes, leaving one negative and one positive (with proton) lobe free. These can then bond with another hybridized hydroxyl group from another layer (Figure 2.1). This is called an hydroxyl bond and it leads to each superposed layer of gibbsite lying directly above the one below in an open-packed arrangement. (Figure 2.2). The hydroxyl ions are thus stacked according to the scheme
Univers
ity of
Cap
e Tow
n
5
ABBA-ABBA. On heating, gibbsite dehydroxylates to yield an oxygen sublattice stacking ABAC-ABAC, classified as a 13-series material by Krishner (1971).
H
H
Fig. 2.1: Hydroxyl bonding between sp3 hybridised orbitals
T <I(
~ oi II
ID. c ... I u
l ~Al at 113; 213
~ Al at V6; 616
Fig. 2.2: ABBA-ABBA stacking in Gibbsite
A
A
8
e
A
(from Schoen and Roberson, 1970)
These layers extend laterally by adding on more units, but because each unit is charge satisfied, no ionic forces exist to attach layers above and below. These add more
. slowly, giving rise to the characteristic plate-like shape of gibbsite crystals, described by Schoen and Roberson ( 1970) following their investigation by electron microscopy. Gibbsite forms as thin platelets whose surfaces are parallel to the basal planes {001} of its lattice, along which there is perfect cleavage (Figure 2.3.). Gibbsite is usually found as white or grey foliated masses.
5
ABBA-ABBA. On heating, gibbsite dehydroxylates to yield an oxygen SUbJlattic:::e stacking ABAC-ABAC, as a .8-series by Krishner (1971).
H
~
1: Hydroxyl oonlilmg between sp3 nybndl.sed
orbitals
Fig.
T A
" R A CJi II
CD.
IE B 00 I IJ
B
A
II
~ AI at 113; 2/3
Q) AI at V6; 5/6
stacking Gibbsite (from Schoen and Roberson, 1970)
layers extend laterally by adding on more units, but because each unit is charge sansnea, no ionic forces attach layers above and These add more
. slowly, giving rise to the plate-like shape of crystals, described by Schoen and Roberson (1970) following their by electron microscopy. iibllislte forms as thin surfaces are basal planes {OOI}
.......... """'. along which there cleavage (Figure Gibbsite is usually as white or grey foliated masses.
~
Univers
ity of
Cap
e Tow
n
6
.C• ••ii
_. b-•···
Fig. 2.3: Crystal Morphology (from Schoen and Roberson (1970) (a) Gibbsite (b) Bayerite (c) Nordstrandite
Ginsberg et al (1962) and Saalfeld et al (1968) regarded the sodium they consistently found in gibbsite as being necessary to stabilise the structure and consequently regard gibbsite not as a pure aluminium hydroxide, but as a ternary compound.
2.2.2 BA YERITE.
Bayerite is monoclinic, having a=5.062, b=8.671, c=4.713 A and 6=90.27° (ASTM 20-11).
The basic structural unit of bayerite is exactly the same as that of gibbsite, the only difference being in the manner in which the units are stacked. In bayerite, hydroxyl bonds do not form. There is still distortion owing to the neutral "holes", but overall the proton-oxygen bonds point upward from the sheet of ions. These protons repel each other, so that successive layers mesh together in a manner similar to closepacking. This gives the protons maximum separation. Although this structure resembles close-packing, the effective hydroxyl ion separation from the hydroxyl ions in the next layer is 1.57 A, whereas that of gibbsite was 1.39 A, showing that the structure is really much more open than the so-called "open-packed" one. This "open" packing leads to the hydroxyl ions being stacked in the pattern ABAB-ABAB (Figure 2.4). When heated, bayerite dehydroxylates to yield an oxygen sublattice stacking of ABC-ABC, therefore belonging to the r-series according to Krishner (1971).
-
Fig. 2.3: Crystal Morphology (from Schoen and Roberson (1970) (a) Gibbsite (b) Bayerite (c) Nordstrandite
6
I
Ginsberg et al (1962) and Saalfeld et al (1968) regarded the sodium they consistently found in gibbsite as being necessary to stabilise the structure and consequently regard gibbsite not as a pure aluminium hydroxide, but as a ternary compound.
2.2.2 BAYERITE.
Bayerite is monoclinic, having a=5.062, b=8.671, c=4.713 A and 6=90.27° (ASTM 20-11).
The basic structural unit of bayerite is exactly the same as that of gibbsite, the only difference being in the manner in which the units are stacked. In bayerite, hydroxyl
bonds do not fonn. There is still distortion owing to the neutral "holes", but overall the proton-oxygen bonds point upward from the sheet of ions. These protons repel each other, so that successive layers mesh together in a manner similar to closepacking. This gives the protons maximum separation. Although this structure resembles close-packing, the effective hydroxyl ion separation from the hydroxyl ions in the next layer is 1.57 A, whereas that of gibbsite was 1.39 A, showing that the structure is really much more open than the so-called "open-packed" one. This "open" packing leads to the hydroxyl ions being stacked in the pattern ABAB-ABAB (Figure 2.4). When heated, bayerite dehydroxylates to yield an oxygen sublattice stacking of ABC-ABC, therefore belonging to the r-series according to Krishner (1971).
Univers
ity of
Cap
e Tow
n
7
In bayerite, addition of new unitS is preferentially to the basal planes, with slower growing edges. This gives the characteristic shape of long, tapered rods, the long ( caxis) direction being perpendicular to the basal planes of the lattice. (Figure 2.3)
B
T <( A .-,..... v II B u
__t_
A
~aluminum
Fig. 2.4: ABAB-ABAB stacking in Bayerite
Fig. 2.5: ABBA-BABA stacking in Nordstrandite
(from Schoen and Roberson 1970)
2.2.3 NORDSTRANDITE.
Nordstrandite is triclinic, having a=8.893, b=5.004, c= 10.237 A and a=92°56' , .6= 110°23' , r=90°32' (ASTM 18-31).
Nordstrandite can be regarded as having a structure midway between that of gibbsite and bayerite. It is proposed that the structure is made up of alternating gibbsite and bayerite layers, as depicted in Figure 2.5. This means the hydroxyl ions are packed according to the pattern ABAB-BABA. According to Schoen and Roberson (1970), this type of mirrored layer structure is not unusual, particularly among clay minerals. When heated, nordstrandite dehydroxylates to the oxygen sublattice stacking ABCABC, therefore belonging to the r-series according to Krishner (1971).
Nordstrandite t.akes the form of long rods with hexagonal cross-section (Figure 2.3). Crystal growth is slowest in the a-axis direction.
2.3 STRUCTURE OF THE OXIDE HYDROXIDES.
2.3.1 Diaspore.
Diaspore is orthorhombic, having a=4.396, b=9.426, c=2.844 A (ASTM 5-355). It is isomorphous with Goethite (r-FeOOH). When heated, diaspore dehydroxylates to
7
In bayerite, addition new is preferentially to basal planes, with slower growing This gives characteristic shape of long, tapered rods, the long (c-axis) direction being perpendicular to the basal planes of the lattice. (Figure 2.3)
Fig.
B
A
B
A
I'
~aluminum
ABAB-ABAB stacking in Bayerite
Fig. 2.5: stacking in Nordstrandite
(from Schoen and Roberson 1970)
NORDSTRANDITE.
Nordsttandite is triclinic, having a=8.893, b=5.004, c= 10.237 A and a=92056' , 110~3' , r=90032' (ASTM 18-31).
Nordstrandite can be regarded as having a structure midway between that of gibbsite and bayerite. It is proposed that the structure is made up of alternating gibbsite and bayerite layers, as depicted in Figure This means the hydroxyl ions are packed according the pattern ABAB-BABA. According to Schoen and Roberson (1970), this type of mirrored layer is not unusual, particularly among minerals. When heated, nordstrandite dehydroxylates to the oxygen sublattice stacking ABC-
therefore belonging to the according Krishner (1971).
Nordstrandite takes form of long rods hexagonal cross-section 2.3). Crystal growth is slowest in the direction.
STRUCTURE OF OXIDE HYDROXIDES.
2.3.1 Diaspore.
Diaspore orthorhombic, having a=4.396, b=9.426, A (ASTM It is isomorphous with Goethite (r-FeOOH). When heated, diaspore dehydroxylates to
Univers
ity of
Cap
e Tow
n
the oxygen sublattice stacking ABAB-ABAB, therefore it belongs to the a-series, according to Krischner ( 1971).
Diaspore crystals are thin, brittle {010} plates and the powder is white or grey.
2.3.2 Boehmite.
8
Boehmite has an orthorhombic structure. Milligan and McAtee (1965) determined the unit cell parameters to be a= 3.690 A, b=12.227 A and c=2.868 A (ASTM 21-1307). Wilson ( 1979} and Lippens and Steggerda ( 1970) confirmed the structure of boehmite to be isomorphous with Lepidocrocite (FeOOH), in that each aluminium atom is surrounded by a distorted octahedron of oxygen atoms. These [ Al-06] octahedra polymerise to form an interlocking double molecular layer, which the above researchers preferred to model as zigzag HO-Al-0 chains. The successive double layers are held together by hydrogen bonds between the OH- groups on the outside surfaces of the layers. This is evidenced by the O-H-0 bond length of 2, 70 A, which is shorter than the length normal repulsion between oxygens would achieve. Extra interlamellar water disrupts the lattice and depending on the amount present, this may result in the formation of pseudoboehmite. ·
Diaspore and boehmite differ in their arrangement of the double chains, as is shown schematically in Figure 2.6.
··, . - --.. ---.. . . ........... ,
' 1' I
I
I
·-········ ··--····-······
I
I . b c );, d •( ,,
'!:.~
Fig. 2.6: Schematic representation of OH-Al-0 chains. (a) OH-Al-0 chain parallel to plane of drawing
0 OH
0 0
• Al
(b) 2 anti-parallel chains perpendicular to plane of drawing (after Lippens et al, 1970)
the oxygen sublattice stacking ABAB-ABAB, therefore it belongs to the a-series, according to Krischner (1971).
Diaspore crystals are thin, brittle {010} plates and the powder is white or grey.
2.3.2 Boehmite.
8
Boehmite has an orthorhombic structure. Milligan and McAtee (1965) determined the unit cell parameters to be a= 3.690 A, b= 12.227 A and c=2.868 A (ASTM 21-1307). Wilson (1979) and Lippens and Steggerda (1970) confirmed the structure of boehmite to be isomorphous with Lepidocrocite (FeOOH), in that each aluminium atom is surrounded by a distorted octahedron of oxygen atoms. These [Al-06] octahedra polymerise to form an interlocking double molecular layer, which the above researchers preferred to model as zigzag HO-AI-O chains. The successive double layers are held together by hydrogen bonds between the OH- groups on the outside surfaces of the layers. This is evidenced by the O-H-O bond length of 2,70 A, which is shorter than the length normal repulsion between oxygens would achieve. Extra interlamellar water disrupts the lattice and depending on the amount present, this may result in the formation of pseudoboehmite. '
Diaspore and boehmite differ in their arrangement of the double chains, as is shown schematically in Figure 2.6.
", • _ ~ _ .. ....... f . ············1 ,
1 ' 1
I
1 .......... .~ ~ ~ ..... ~ ..... ... 1 , , .
b C A;, d If ': '!:.!oo
Fig. 2.6: Schematic representation of OH-AI-O chains. (a) OH-Al-O chain parallel to plane of drawing
Q) OH
0 0
• AI
(b) 2 anti-parallel chains perpendicular to plane of drawing (after Lippens et al, 1970)
Univers
ity of
Cap
e Tow
n
' ' ' ' ' ' ~ ... ~
.. ··· '' ..... '
~~-~,~¥-S I / I I / , . I ,/ ·· .. :y\ ... : I . I I I ,- +- - - ' -1- - - s
0
' I ' I ,, ' '" "l
~----~ a I< b .. ,
I I ...... , I
f
..... -·~ ...... ---
,- - - - - -. - -, - - - - .. - - - ,. - -s
I
l .... ··: ••••• J I I I I
------'
-- . - .. , ' I
- - - - - ..J I I
I ' I
-, - -s
I
f
....... J J" -· ... , ... -.. --~ . . ..... ~
L ______ J
b b •I
9
0
Fig. 2. 7: Schematic representation of the crystal structure: (a) Diaspore ; (b) Boehmite
S represents cleavage planes (after Lippens et al, 1970)
The structure explains the crystal shape of orthorhombic plates, with easy cleavage on (010) corresponding to breaking of the weak hydrogen bonds. Boehmite forms a fine,
white powder. When heated, boehmite dehydroxylates to the oxygen sub lattice stacking ABC-ABC, therefore belonging to the r-series, according to Krishner (1971).
2.3.3 Pseudoboehmite.
There exists a poorly crystalline compound which has been called pseudoboehmite or gelatinous boehmite. It has basically the same structure as boehmite, but contains more water and has the overall appearance and behaviour of a gel or paste.
Often encountered by early researchers where a crystalline precipitate was expected, it was assumed to be an incompletely dehydrated form of boehmite. For example, Moeller (1952) reports that boehmite may simply be precipitated from boiling aluminium salt solutions by ammonia, or by ageing the amorphous gel precipitated by
the same method at room temperature.
MacKenzie (1957), from his differential thermal investigation into synthesised alumina gels, reports that all fresh precipitates of aluminium chloride solutions consist of
poorly-crystallised boehmite, which can be transformed to a crystalline form by ageing in the mother liquor at elevated temperatures.
Papee et al (1958) examined both fully crystalline and poorly crystalline boehmite by X-ray diffraction. They noted a considerable variation in the first reflection, from
"
o
a
" "
1<
"
b
" "
,.1
, I ------, ,
r
o - _o ___ ~
........ ~-
,- - - - - -. - -I
- - - - .,.. - - - T --S ,
l------l _____ J
1 I I ,
------,
- - ... - .. ~I . I
____ .. ,J
I I
I
. , -, --s
1
r
-----· OJ JO -- ---,
-.------~ . --,---~ l ______ J
b b -I
9
o
Fig. 2.7: Schematic representation of the crystal structure:(a) Diaspore; (b) Boehmite S represents cleavage planes (after Lippens et al, 1970)
The structure explains the crystal shape of orthorhombic plates, with easy cleavage on (010) corresponding to breaking of the weak hydrogen bonds. Boehmite forms a fine, white powder. When heated, boehmite dehydroxylates to the oxygen sublattice stacking ABC-ABC, therefore belonging to the r-series, according to Krishner (1971).
2.3.3 Pseudoboehmite.
There exists a poorly crystalline compound which has been called pseudoboehmite or gelatinous boehmite. It has basically the same structure as boehmite, but contains more water and has the overall appearance and behaviour of a gel or paste.
Often encountered by early researchers where a crystalline precipitate was expected, it was assumed to be an incompletely dehydrated form of boehmite. For example, Moeller (1952) reports that boehmite may simply be precipitated from boiling aluminium salt solutions by ammonia, or by ageing the amorphous gel precipitated by the same method at room temperature.
MacKenzie (1957), from his differential thermal investigation into synthesised alumina gels, reports that all fresh precipitates of aluminium chloride solutions consist of poorly-crystallised boehmite, which can be transformed to a crystalline form by ageing in the mother liquor at elevated temperatures.
Papee et al (1958) examined both fully crystalline and poorly crystalline boehmite by X-ray diffraction. They noted a considerable variation in the first reflection, from
Univers
ity of
Cap
e Tow
n
10
6 .11 A for the former to 6. 6-6. 7 A for the latter. The poorly crystalline material was then treated hydrothermally for up to 48 hours at 300°c, during which process the first reflection shifted from 6.6-6. 7 to 6.19 A, accompanied by a drop in the percentage
water content from 60% to 1 % . From this they concluded that pseudoboehmite is not simply a poorly crystalline form of AlOOH, with the excess water adsorbed on the surface, but rather a different structure, with the water situated between the structural layers.
2.4 METHODS OF IDENTIFICATION.
Two main methods of identification are used: X-ray Diffraction and Differential Thermal Analysis:
2.4.1 X-ray Diffraction.
The method of X-ray diffraction analysis is discussed more fully in chapter 5 .1.5.
Standard samples were supplied by the Nationat Institute for Materials Research for comparative purposes. All samples were examined using Cu-Ka radiation. Random
powder samples were used. All specimens were examined under the standard conditions:
Power: 45 kV ; 40 mA Slits: 1h : 1h : 1 Scan Rate: 2° 2e/ min Time Constant: 1 Range: 1*103
The results obtained are shown in figure 2.8 below.
X-ray powder diffraction was found to be able to differentiate between the polymorphs
easily and quickly, as well as supplying quantitative data. Furthermore, it gives information on the relative crystallinity of the sample. The main limitation of the method is that it.will not detect phases present in very low percentages.
10
6.11 for the former to A for the poorly crystalline 1'n!.ltpr1!.l was treated hydrothermally for up to 48 hours at 3000C, during which fIrSt
ren4eco()n shifted to 6.19 A, by a drop in the peI1CenltaJ!;e
content from 60% 1 %. From this they concluded that pseudoboehmite is not ......... , ... "iI a poorly crystalline form of AIOOH, with excess water adsorbed on the
but rather a different structure, with the situated between the structural layers.
METHODS IDENTIFICATION.
methods of lQenbtlcalt10n are used: ... "' .. , ........ Analysis:
1 X-ray Diffraction.
method of X-ray diffraction analysis is discussed more fully in chapter 1
Standard samples were SUllipJ1€!<1 by the NationallnSlIltu.lte comparative purposes. All were examined radiation. KallOOim
no~wC1f~r samples were All specimens were exalDlillled under the standard conditions:
Power: 45kV 40mA Slits: lh : : 1
Rate: 2° 2el Constant: 1
1*103
2.8 below.
powder diffraction was found to be able to differentiate between the polymorphs and quickly, as well as supplying quantitative data. Furthermore, it gives
information on the relative of the sample. limitation of the method that it will not detect present in very
Univers
ity of
Cap
e Tow
n[; (/) z w 1-z
GIBBSITE
BAYERi
BOEH MITE
PSEUDOBOEHMITE
10 15 20 25 30 35 40 45 50 55 60
DEGREES 2 THETA
Fig. 2.8: Comparative XRD patterns for aluminium hydroxide and oxyhydroxide polymorphs
11 I
BAYERI
BOEHMITE
PSEUDOBOEHMITE
30 35
DEGREES
2.8: Comparative XRD patterns for aluminium hydroxide and oxyhydroxide polymorphs
11\
Univers
ity of
Cap
e Tow
n
12
2.4.2 Thermal Analysis.
This technique is very sensitive, being able to detect very small percentages of phases present. In addition it can provide quantitative data, but it is usually used qualitatively to confirm or detect which polymorph is present. The fundamental investigation of this technique for clay minerals was done by MacKenzie (1957), who performed syntheses of the aluminium hydroxides and oxyhydroxides and examined the resulting powders by differential thermal analysis. He obtained the following results:
Mackensie found that the differential thermal curves for gibbsite all showed a strong #
endothermal peak between 3200C and 3300C, usually associated with two small endothermic peaks at 250-3000C and 525°C. The bayerite curves varied considerably, the main characterising feature being a strong endothermic peak at 310-315°c. He observed a single, strong endothermic peak in the region 450-580°C for boehmite, the position of which appeared to vary according to sample particle size. A smaller exothermic hump was sometimes observed between 850-950°C. Finally, Mackenzie prepared precipitates from AlC13 solution which he called poorly crystalline boehmite. These gave differential thermal curves characterised by broad endothermic peaks of intermediate strength, usually having maxima in the regions 100-150 °c and 400-5000C.
Van Zyl (1987) performed thermal analyses on his synthesised materials to characterise them. His results, reproduced in Figure 2.9 below, agree with those of Mackenzie, though the peak positions are consistently lower. This was probably as a result of the method of synthesis used by Van Zyl, which gives a product of smaller particle size. Mackenzie points out that smaller particles would tend to lower the peak temperature.
Differential thermal and thermogravimetric analyses were performed on the standard samples supplied by the National Institute for Materials Research. This was done in order to obtain comparative curves for work done at UCT and to gain experience with the technique, its limitations and potential.
All determinations were done under the same conditions: Approximately 20g of sample was weighed out accurately. The temperature was raised from zero to 700°C at l<>°C per minute in air flowing at 30 ml/min. Temperature differences and gravimetric cha~ges were plotted against temperature. The results obtained are shown in Figure 2.10.
12
2.4.2 Thermal Analysis.
This technique is very sensitive, being able to detect very small percentages of phases present. In addition it can provide quantitative data, but it is usually used qualitatively to confmn or detect which polymorph is present. The fundamental investigation of this technique for clay minerals was done by MacKenzie (1957), who performed syntheses of the aluminium hydroxides and oxyhydroxides and examined the resulting powders by differential thermal analysis. He obtained the following results:
Mackensie found that the differential thermal curves for gibbsite all showed a strong • endothermal peak between 3200C and 3300C, usually associated with two small endothermic peaks at 250-3000C and 525°C. The bayerite curves varied considerably, the main characterising feature being a strong endothermic peak at 31O-3150C. He observed a single, strong endothermic peak in the region 450-580oC for boehmite, the position of which appeared to vary according to sample particle size. A smaller exothermic hump was sometimes observed between 850-950oC. Finally, Mackenzie prepared precipitates from AICl3 solution which he called poorly crystalline boehmite. These gave differential thermal curves characterised by broad endothermic peaks of intermediate strength, usually having maxima in the regions 100-150 °c and 400-500°C.
Van Zyl (1987) performed thermal analyses on his synthesised materials to characterise them. His results, reproduced in Figure 2.9 below, agree with those of Mackenzie, though the peak positions are consistently lower. This was probably as a result of the method of synthesis used by Van Zyl, which gives a product of smaller particle size. Mackenzie points out that smaller particles would tend to lower the peak temperature.
Differential thermal and thermogravimetric analyses were performed on the standard samples supplied by the National Institute for Materials Research. This was done in order to obtain comparative curves for work done at UCT and to gain experience with the technique, its limitations and potential.
All determinations were done under the same conditions: Approximately 20g of sample was weighed out accurately. The temperature was raised from zero to 700°C at 100C per minute in air flowing at 30 ml/min. Temperature differences and gravimetric chap\ges were plotted against temperature. The results obtained are shown in Figure 2.10.
Univers
ity of
Cap
e Tow
n
13
0 0 0
= 0 ... -
0 0 0
~ ~ 0
~ -i-l ......
0 s 0 0
~ ~ 0
-i-l ,..c:: ~
...... ~ ~ 0 0 0
~ 0 .. 0
~ ,.Cl ..
ro 0
o::i 0 "O 0 0 .. ;:; 0 ..
~ 0 1fJ. 0 0-t
0 ... 0 ...
0 0 0 N
0 N
0 0 ... N 0 N ... 'f ... N 0 N I I
... .. I I I
0 0 0 0
= = 0 0 0 0 N ~ ...
0 0 0 0
~ 0 ~ ~ ...
-i-l -i-l ...... ...... 1fJ. 0 s 0
,.Cl 0 0 .. ..
,.Cl ,..c:: ...... ~ Q() 0 0
0 0 0 0 0
o::i
g 0 0 .. ...
0 0 0 0 N N
0 0 ... N 0 N 1 'f ... N 0 N ... 'I' I I I
Fig. 2.9: DTA curves of aluminium hydroxides and oxyhydroxides. (Van Zyl, 1987)
13
"" .. .. ... "" .... ... -1< 0 <:> <:>
~ ~ ...
8 "" CIJ :: .,. :: ...,
'l'""4 ~ <:> 0 <:>
CIJ .. '"
.. .. ~ ro 0
0:1 <> '"d <:> <:> .. ~
<:> ...
J CIJ .. r:n .. p... <:> ... <:> .... ::;::::,
IS 0 .. <> (II
<:> ...
"" .. .... . 1'1 "" Of ... ,. I
.. (II "" ~ T .. I
.. <:> .. <:> :! :!
"" <:>
"" <:> !l !l
"" 8 CIJ ! CIJ :: ..., ...,
'l'""4 'l'""4
r:n <:> S "" ,.Cl
<:> <:>
'" '" ,.Cl ~ 'l'""4 CIJ bJJ "" 0
<:> <:> <> ... ..
0:1
8 8 .... ...
8 "" (II 2
<:> <:> .. N .. ~ 1 , ... ... .. N .. ,. I I
Fig. 2.9: DTA curves of aluminium hydroxides and oxyhydroxides. (Van Zy1, 1987)
Univers
ity of
Cap
e Tow
n
c we +J ".::::-·- E Lo Cl> ~ >. 0
0 .E CD.=
E ti-0 ... ii E ~ QI ...
0 N ..,. co Ill I I I I
c
Cl> e +J ".::::-·- E (J)O
..Q ~
..Q 0 ·- c 0-c e
d-0 ...
0 N ..,. IO 111 I I I I
0 N ... co Ill 0 N ... ... "j "j ... N N I I I I I
(~)VJ.a
(:>6eP)V.UJ
Cl> 0 .µ 0 Ill
E c .c e
Cl>".::::-0 0 E
..Q g "' 0
I .. Q 0 0 .E 0 .E 0 :! ""O c ...
::i ::J e 1j .. Cl> ti-0 Q. (/) 0 E 0 0... ': .. "' " E Q. QI E qj
0 0 ... 0 UI ... N .&:
"' "i J
0 ..,. IO Ill IQ "j N N N ci I I I I
0 0 411
" 0 Q.) .!: 0 +J E IO ·-".::::-
0 EE "' .c ~ 0 "ti
.E Cl>< 0 .5 e CD .s .a 0 0 E
0 .. 8' ... " Q.
E u .. ii 0 E Q. 0 E ·ci 0 N
UI ... 0 .&: 0 "' N "i
~
Fig. 2.10: DTA curves of NIMR standard samples: (a) gibbsite (b) bayerite ( c) boehmite ( d) pseudoboehmite
~ N
"j I
14
IQ I') IQ ..,. iq Ml
l'i I ..; I ... I I I I
Co~ep)V.UJ
N-o-c."1.-IO I I I I I
(~)V.UI
0 0 co
0 0 ...
0 0 N
0
0 0 IO
~
0 0 N
0
"' 0 "ti
.E e ::i
1! 0 Q.
E 0 ~ 0 Q. E 0
UI
0
"' 0 "ti
.E e ::i
1! x. E 0 ... .. Q. E 0
UI
14
0 0
0 0 Ill!
'" E c: .e-(1):;;
0 o E 0
"" .og "" " " I Q ~
"(I
0 .E o c: 0 .E
0 CI -0';; 0 I!
~ .. .., ::>
:I .:lE e E (1)'-co " Q" eng Q.
E 0..": E
~ " " .... DO "'t
E " E <I
I'; Ci. "!
Ci.
(I E Ill! E
... 0 D ... 0 D
0 III 0 III
N ~ N
"" 'i J
0 0
0 N ~ Ill! III 0 N ~ Ill! III 0 N ~ 10 10 II) j' III N II) I') II) ... 111 II)
I I I I j' j' ... ... ... ~ N N N 'i' d j' I « I I"i I ... I
I I I I I I I I I I
{OhP)v.I.CI {OaGp)v.I.CI
0 0 III
0 0 III
" <:
(1)~ 0 0)5 0 +IE
+I -fD .-'-
.- E 0 EE 0
roo .o~
ClIO .c:~ "" " " .ou
"(I 'tl
.E 0)< 5 .E I! 05 ~ I! <: LIl.s ~
::I :I
0 ... E e D 0 .. g 0 ~ .. 1
0 Q. ... E .. f ... ~ .... ~ co E
<I
Ci. ~ Ci.
E '0 E
D '" 0
., III 0 III
0 ~ N 0 ClIO
'" i
Fig. 10: DTA curves of NIMR standard samples: (a) gibbsite (b) bayerite (c) boehmite (d) pseudoboebmite
Univers
ity of
Cap
e Tow
n
15
(i) Gibbsite:
There was no strong dehydration peak. Thermal and gravimetric changes
began at 230°C. A small thermal event was recorded at 240°C followed by a larger peak at 3I0°c. The peaks were smooth, sharp and well defined. The thermal event was completed at 360°C. There was a ~harp reduction in the rate of weight loss at 320°C. No subsidiary peaks were observed. This result agrees well with the results of both Mackenzie and Van Zyl.
(ii) Bayerite:
No initial dehydration took place. Thermal changes and weight loss began at approximately 240°C. One large endothermic peak was observed culminating just above 300°C. There was a slight change in slope of the curve at approximately 400°C, which may be evidence of a further thermal event. The peak was smooth, sharp and well defined. The main endothermic event was completed at approximately 340°C, and this corresponded to a sharp reduction in the rate of weight loss.
(c) Boehmite:
The boehmite differential thermal curve was more complex. Heat was immediately absorbed as shown by a broad, jagged peak reaching a maximum at 280°c. This was followed by a smooth, sharp peak at 560°C. This event was accompanied by a rapid weight loss up to 580°C whereafter the rate
decreased. The peak maximum is higher than that found by Van Zyl, but within the range given by Mackenzie.
(d) Pseudoboehmite:
It can immediately be seen that this is an amorphous form from the broad and jagged shape of the peaks. There is an immediate absorption of energy reaching a maximum at 80°-too0c. This is accompanied by a rapid gravimetric loss. The first event can be ascribed to H20 held in the gel being driven off. Van Zyl also found this peak to be strongest in pseudoboehmite. It is followed by a broad and very jagged peak, again with a maximum at about 270°-290°C, accompanied by a slow gravimetric loss. This may be ascribed to water held within the crystal structure being driven off. Finally there is a third peak which is less broad and less jagged. It begins at 330°C and reaches a maximu~ at 390°-400°C; there is a corresponding rapid gravimetric loss in this temperature range. This thermal event is completed at about 500°C. These results are consistent with those of Mackenzie and Van Zyl.
15
(i) Gibbsite:
was no dehydration peak. Thermal and gravimetric changes began at A small thermal was recorded at 240°C followed by a larger at 3100C. peaks were smooth, sharp and well defined. thermal event was completed at 360°C. was a &harp reduction the rate weight at 3200C. No subsidiary peaks observed. This result
well with the results both Mackenzie and Van Zyl.
(ii) Bayerite:
No initial dehydration took place. Thermal changes weight began at approximately 24QOC. One large endothermic peak was observed culminating just above 3()()OC. was a slight change in slope of the curve at approximately 400°C, which may evidence of a further thermal The peak was smooth, sharp and well defined. The main endothermic was completed at approximately 3400C, and corresponded to a reduction
the rate of weight loss.
(c) Boehmite:
boehmite differential thermal curve was more complex. was immediately absorbed as shown by a broad, jagged peak reaching a maximum at 2800C. was followed by a smooth, sharp peak at This was accompanied by a rapid loss up to 580°C the rate decreased. The maximum is higher than that found by Van Zyl, but within the given by Mackenzie.
(d) Pseudoboehmite:
It can immediately be seen that this is an amorphous from the broad and jagged shape of the peaks. There is an immediate absorption of energy reaching a maximum at 800-100oC. This is accompanied by a rapid gravimetric loss. The event can be ascribed to H20 held the being driven Van Zyl found peak to strongest in pseudoboehmite. It is followed by a broad and jagged again a maximum at about 270o-2900C, accompanied by a slow loss. may be ascribed to water held within the structure being driven Finally there is a peak is less broad and less jagged. It begins at 330°C and reaches a maximu~ at 390°-400°C; there is a corresponding rapid gravimetric loss this temperature range. This thermal event completed at about 5()()OC. These results are consistent with Mackenzie and Van Zyl.
Univers
ity of
Cap
e Tow
n
16
2.4.3 Other Methods of Identification.
All the compounds under discussion contain OH groups; consequently they may be identified by infrared absorption analysis as each shows characteristic OH bending and stretching bands. These characteristics are discussed by Ryskin (1974). This method was not used in this· study, as the previously outlined methods proved sufficient for all purposes. Furthermore, infrared data must be interpreted with caution, requiring the prior acquisition of such skills.
Identification may also be made based on the different rates of dissolution of the polymorphs. This is an extremely difficult technique, as solubility can be affected by, among others, particle size, crystallinity and impurities. Hsu (1977) found this method particularly useful in analysing aluminium hydroxide components in soils.
16
2.4.3 Other Methods of Identification.
All the compounds under discussion contain OH groups; consequently may be identified by infrared absorption analysis as shows characteristic bending and stretching bands. characteristics are discussed by Ryskin (1974). This method was not used in study, as the previously outlined methods proved sufficient for all purposes. Furthermore, infrared data must be interpreted with caution, requiring the
acquisition of such skills.
Identification may also be made based on the different of dissolution of the #
polymorphs. is an extremely difficult technique, as solubility can be affected by, among others, particle crystallinity and impurities. Hsu (1977) found this method particularly useful in analysing aluminium hydroxide components in soils.
Univers
ity of
Cap
e Tow
n
17
3. MECHANISMS OF FORMATION OF ALUMINIUM HYDROXIDES AND OXYHYDROXIDES.
3.1 REVIEW OF SYNTHESIS PROCEDURES.
As researchers have attempted to understand the differences between the various polymorphs and the mechanism by which they form, many different methods of synthesis have been attempted. More recently, the commercial importance of these materials has spurred efforts to find methods of synthesis which are relatively simple and give good results. It is not possible to be comprehensive in this review of the methods of synthesis, but the examples selected should indicate the range and types of methods available.
3.1. l Aluminium hydroxides.
Moeller (1952), in his "Textbook of Inorganic Chemistry", lists three methods by which aluminium hydroxides may be formed:
a) hydration of alumina, b) hydrolysis of aluminate solution, and c) reaction of aluminium salt solution with alkali.
Variations in the reaction conditions are said to yield the various crystal forms. For example, gibbsite may be formed by rapid hydrolysis of aluminate solutions, or by reaction with C02 at lOO°C, or by ageing bayerite in dilute alkali at 600C. Bayerite may be formed by rapid hydrolysis of aluminate solution.
Brauer (1963) describes two methods by which gibbsite may be obtained by dissolving existing aluminium hydroxide material and reprecipitating. Bayerite is precipitated by passing C02 gas through the solution.
The method of rapidly neutralizing sodium aluminate solution with Na2co3 at room temperature was successfully used by Petz (1968) in preparing samples of amorphous alumina gel. Wade and Bannister (1973) in "Comprehensive Inorganic Chemistry" say that gibbsite can be precipitated from sodium aluminate by passing through C02 and recrystallising at 800C, while bayerite results if the solution is held at 200C.
More recently Sato ( 1981) precipitated hydroxides from aluminate solution with varying concentrations of HCI and HN03, so obtaining pseudoboehmite or bayerite, with the formation of bayerite enhanced by slow addition of the acid, higher temperatures and higher concentrations. When the pseudoboehmite was aged in the mother liquor, it transformed to- gibbsite or bayerite, depending on the alkali concentration.
Many authors, in contributing to a definition of the mechanism of formation of these polymorphs, have adopted the method of synthesis involving the precipitation of the product from AlC13 solution by the addition of NaOH or other hydroxyl containing
17
MECHANISMS FORMATION OF ALUMINIUM HYDROXIDES AND OXYHYDROXIDES.
1 REVIEWOF JIlAJL,# .... u .... PROCEDURES.
researchers have attempted to understand the differences between the various polymorphs and the mechanism by which they many different methods of synthesis have been attempted. More recently, commercial importance of these materials has spurred to find methods of synthesis which are relatively SImple
good It is not possible to in this rpv',pu/
melIDOC1S of but examples selected should indicate the methods available.
1.1 Aluminium hydroxides.
the types of
Moeller (1952), in his of Inorganic Chemistry", lists three methods by which aluminium may be formed:
a) hydration of alumina, b) aluminate solution, and c) reaction aluminium salt solution with alkali.
Variations in the conditions are said to yield the various crystal For example, gibbsite may formed by rapid hydrolysis aluminate solutions, or by reaction with CO2 at l000C,. or by ageing in dilute alkali at 600C. may be formed by rapid hydrolysis of aluminate solution.
UTa.uer (1963) two methods by which may be obtained dissolving "' ...... "'lI.LAj.c aluminium hydroxide material and reprecipitating. Bayerite is precipitated by ..,a.>';'UAJ;; CO2 gas through solution.
The method of rapidly neutralizing sodium aluminate solution with Na2C03 at room temperature was used by Petz (1968) preparing samples of amorphous alumina gel. Wade and (1913) in "Comprehensive Inorganic '"'ll ........ "' ..... 7
that gibbsite can be from sodium by passing CO2 and recrystallising at bayerite solution is held at
More recently Sato (1981) precipitated hydroxides from aluminate solution with varying concentrations and HN03, so obtaining pseudoboehmite or bayerite, with the formation of bayerite enhanced by slow addition of the acid, ....... p, ..........
temperatures and higher concentrations. When pseudoboebmite was in the m01:ber liquor, it to gibbsite or depending on concentration.
Many authors, in contributing to a definition of mechanism of of these polymorphs, have adopted the method of synthesis involving the precipitation of the product from AICl3 solution by the addition of or other hydroxyl containing
Univers
ity of
Cap
e Tow
n
18
solutions. Among these researchers are Hsu and Bates (1964), Barnhisel and Rich
(1965), Turner and Ross (1970), Stol et al (1976), Ng Kee Kwong and Huang (1979), and Violante and Violante (1980).
Hsu and Bates (1964) refer to the fact that, although earlier researchers formed the various polymorphs by controlling the pH of tbe reaction solution, a variety of products can be obtained at the same pH by varying other reaction conditions more
widely, possibly as the result of changes in the reaction kinetics or the presence of extraneous anions. To study the mechanism of aluminium hydroxide formation, they
prepared gibbsite, bayerite and nordstrandite, or mixtures of these, by reacting NaOH' with AIC13 or Al2(S04)J over a range of ratios.
Schoen and Roberson (1970) prepared the hydroxides by mixing three aqueous solutions in various ratios. The three solutions contained:
a) H+ ClO = Na+ Af3+ ' 4 ' '
b) Na+ ClO = oH= ' 4 '
c) Na+, ClOt, Af3+.
They found that gibbsite precipitated slowly from solutions with pH < 5, 8 and bayerite precipitated rapidly from those with pH > 5,8. Nordstrandite formed from bayerite when aged at high pH values, while at lower values both gibbsite and bayerite formed. On aging, the gibbsite converted to bayerite.
In 1971 Chesworth proposed a method for rapidly synthesizing Al(OH)J under conditions of low temperature and pressure and pointed out that the method had been available in the literature since 1870. Aluminium dissolved in mercury (ie. amalgamated) is highly reactive and displaces hydrogen from water to form gibbsite. Other polymorphs are easily formed by varying the reacting medium, thus 0,5M NaCl, 2M NaOH, 0, IM Na2C03 and NH40Hconc. yield bayerite, pseudoboehmite, nordstrandite and gibbsite respectively at different temperatures. In the same year McHardy and Thomson prepared Al(OH)J gels by the hydrolysis of amalgamated aluminium and by precipitation from aluminium salt solutions with an anion exchange resin. These gels then crystallized to give the polymorphs.
In order to examine the rate of crystallisation, Aldcroft et al (1968,1969) successfully produced hydroxide powders by hydrolysis of aluminium s-butoxide solutions in benzene or ethylbenzene, with equal volumes of water or aqueous organic solvent. Yoldas (1973) described a method of synthesizing bayerite and boehmite by hydrolysing aluminium alkoxides of formula Al(OR)3 (R =CHMe2 or CHMeEt) at 200C and 75°C respectively. Bayerite is obtained by ageing the solution. Van Zyl (1987) extended this method successfully to produce a range of polymorphs from a stock alkoxide solution. He determined that this method produces pseudoboehmite, not boehmite as reported, with further processing necessary to obtain crystalline boehmite.
18
solutions. Among these researchers are Hsu and Bates (1964), Bamhisel and Rich (1965), Turner and Ross (1970), Stol et al (1976), Ng Kee Kwong and Huang (1979), and Violante and Violante (1980).
Hsu and Bates (1964) refer to the fact that, although earlier researchers formed the various polymorphs by controlling the pH of tl)e reaction solution, a variety of products can be obtained at the same pH by varying other reaction conditions more widely, possibly as the result of changes in the reaction kinetics or the presence of extraneous anions. To study the mechanism of aluminium hydroxide formation, they prepared gibbsite, bayerite and nordstrandite, or mixtures of these, by reacting NaOH' with AlCl3 or AI2(S04h over a range of ratios.
Schoen and Roberson (1970) prepared the hydroxides by mixing three aqueous solutions in various ratios. The three solutions contained:
a) H+ CIO::: Na+ Al3+ , 4' , b) Na+ CIO::: OW , 4, c) Na+, ClOt, Al3+.
They found that gibbsite precipitated slowly from solutions with pH < 5,8 and bayerite precipitated rapidly from those with pH > 5,8. Nordstrandite formed from bayerite when aged at high pH values, while at lower values both gibbsite and bayerite formed. On aging, the gibbsite converted to bayerite.
In 1971 Chesworth proposed a method for rapidly synthesizing AI(OHh under conditions of low temperature and pressure and pointed out that the method had been available in the literature since 1870. Aluminium dissolved in mercury (ie. amalgamated) is highly reactive and displaces hydrogen from water to form gibbsite. Other polymorphs are easily formed by varying the reacting medium, thus 0,5M NaCI, 2M NaOH, O,IM Na2C03 and NH40Hconc. yield bayerite, pseudoboehmite, nordstrandite and gibbsite respectively at different temperatures. In the same year McHardy and Thomson prepared AI(OHh gels by the hydrolysis of amalgamated aluminium and by precipitation from aluminium salt solutions with an anion exchange resin. These gels then crystallized to give the polymorphs.
In order to examine the rate of crystallisation, Aldcroft et al (1968, 1969) successfully produced hydroxide powders by hydrolysis of aluminium s-butoxide solutions in benzene or ethylbenzene, with equal volumes of water or aqueous organic solvent. Y oldas (1973) described a method of synthesizing bayerite and boehmite by hydrolysing aluminium alkoxides of formula AI(ORh (R =CHMe2 or CHMeEt ) at 200C and 75°C respectively. Bayerite is obtained by ageing the solution. Van Zyl (1987) extended this method successfully to produce a range of polymorphs from a stock alkoxide solution. He determined that this method produces pseudoboehmite, not boehmite as reported, with further processing necessary to obtain crystalline boehmite.
Univers
ity of
Cap
e Tow
n
19
3.1.2 Aluminium oxide hydroxides.
Confusion exists and persists between the crystalline boehmite and the gel-like pseudoboehmite. Researchers routinely refer to "boehmite" when in fact the synthesis has produced pseudoboehmite. This makes outlining of methods of synthesis very complex. In producing bo~hmite, it is usually necessary first to produce a gel or another polymorph; in addition, most syntheses of Al(OH)J polymorphs can yield pseudoboehmite too, as will be discussed more fully below. In compiling this summary of methods of synthesis, the relevant data was examined wherever possible to determine which form was actually being produced. The earlier papers are more likely to be problematic, but have been included for completeness.
Moeller (1952) writes that boehmite may be precipitated from boiling solutions of aluminium salts by ammonia, or by ageing the gels formed when the reaction takes place at room temperature, or else by the reaction of amalgamated aluminium with water at 6<>°C and by ageing the gibbsite in an autoclave at 350°C.
Brauer (1963) also describes a method using amalgamated aluminium metal. He reports that boehmite is also formed when Al(OH)J, which was precipitated with ammonia solution, is heated at 200°c for two hours in an autoclave.
In 1956 Milligan and McAtee determined the unit structure of boehmite using samples prepared firstly by precipitation from aluminium chloride or sulphate with dilute NH40H. Samples were then treated in an autoclave for seven days at 350°C and 340 atm. and for five days 400°C and 408 atm., after which a crystalline powder was recovered. Mukaibo et al (1969) report forming well crystallized boehmite by treating gibbsite in an autoclave at 300°C for 2 hours, while Juan and Lo (1975) found that boehmite formed easily between 302° and 374° Cat pressures between 680 and 1700 atm. from gels which they made by mixing Cao, AI20 3, and silicic acid in suspension. They felt these values to be inflated owing to the retarding influence of the other components on the rate of nucleation and growth of the boehmite phase.
Cylindrical boehmite fibres 0,06 to 0,1 µm long were prepared by Packter (1976) by the hydrolysis of aluminium isopropoxide. This gave an amorphous gel which was then aged in water at 80°C for 30 minutes to form the crystals. In 1982 Krivoruchko et al studied the mechanism of boehmite formation in aged Al(OH)J precipitates. They found that when the hydroxide is aged at 200°C, boehmite is always formed, irrespective of the pH of solution, but that its morphology is determined by both the pH and the phase composition of the hydroxide, sometimes leading to the formation of needle-shaped crystals.
Fanelli and Burlew (1986) synthesised boehmite by the hydrolysis of sec-butoxide; however, water was not added to the system, but formed through the H2S04 catalysed dehydration of the alcohol at the reaction temperature of 2500C. Basic aluminium sulphate was also formed.
19
1.2 Aluminium hydroxides.
Confusion exists and between the crystalline boehmite and the pseudoboehmite. routinely refer to when in fact the synthesis
produced pseudoboehmite. This makes outlining of methods of synthesis complex. In producing boehmite, it is usually first to produce a or , another polymorph; in addition. most syntheses Al(OH)3 polymorphs can yield pseudoboehmite too. as will be discussed more fully below. In compiling summary
methods of synthesis, relevant data was examined wherever possible aetemune which form was actually being earlier papers are more likely
problematic, included for '"'V.JllVU"' ......... JI"'.,.,
Moeller (1952) writes that boehmite may be precipitated from boiling solutions aluminium salts by ammonia, or by ageing the gels formed when the reaction place at room temperature, or else by the reaction amalgamated aluminium with
at (iOOC and by gibbsite in an autoclave at 350°C.
(1963) also .......... 'w ... ..., ...... a method using aluminium metal. that boehmite formed when Al(OH)3' was precipitated with
ammonia solution, is at 2000C for two hours in an autoclave.
In 1956 Milligan and McAtee determined the unit of boehmite using samples prepared firstly by precipitation from aluminium chloride or sulphate with dilute
Samples were treated in an autoclave seven days at 350°C 340 and for five days 4O()OC and 408 atm., after a crystalline was
Mukaibo et al (1969) report forming crystallized boehmite treating an autoclave at 2 hours, while Juan (1975) found that
boehmite formed easily 3020 and 3740 C at between 680 and 1700 gels which they made mixing Cao, Al20 3, silicic acid in suspension.
felt these values to be inflated owing to the retarding influence of the other components on the rate of nucleation and growth of the phase.
Cylindrical boehmite 0,06 0,1 lAm long were (1976) by of aluminium This gave an gel which was
water at 800C minutes to form crystals. In 1982 Krivoruchko studied the mechanism boehmite formation in aged Al(OH>J precipitates.
found that when the hydroxide aged at 200oC, boehmite is always formed, irrespective of the pH of solution, but that its morphology is determined by both the
and the phase composition of the hydroxide. sometimes leading to the formation needle-shaped crystals.
and Burlew (1986) synltbes:i.sea boehmite by nO\lveVj~ water was not to system, but formed dehydration of the alcohol the reaction temperature of sulphate was also formed.
of sec-butoxide; the H2S04 catalysed Basic aluminium
Univers
ity of
Cap
e Tow
n
20
In investigating the preparation of thin aluminium oxide films for use as catalyst supports by evaporation of aluminium in an oxidising atmosphere, Lamber ( 1986) observed that such films are easily hydrated by distilled water. Films of aluminium oxide were immersed in water at between 77°C and 87°C and held there for 1 to 2 minutes, after which electron micrographs revealed needle-shaped crystallites in an amorphous matrix.
The interest in boehmite as a material for use in catalysts has resulted in the filing of numerous patents covering various methods of synthesis. For example, in 1986 Misra and Sivakumar, of the Aluminium Company of America, published a patent for the #
preparation of boehmite by heating a solution containing caustic soda and AI20 3 to between 115°C and 145°C, then seeding the solution with boehmite or pseudoboehmite and ageing the seeded solution for six hours. There are references in the literature to many other patents of novel syntheses or manufacturing methods, from European countries, America, Canada, USSR, Japan and other wide-ranging sources. It was not possible to examine many of these, but they are plentiful enough to warrant care if patent infringements are to be avoided.
Apparently there has not been the same degree of interest in diaspore. Mukaibo et al (1969) report that diaspore can be prepared by treating boehmite in a NaOH solution, seeded with naturally occurring diaspore, in an autoclave at 380°C and 500 atm. for one week. Wade and Bannister, in the text "Comprehensive Inorganic Chemistry", quote Linsen ed.(1970) as saying that all aluminium hydroxides and oxides, when treated hydrothermally at pressures over 140 atm. and temperatures between 275° and 425°C, will be converted to diaspore if seeded with naturally occurring diaspore in this way.
Krause et al (1970) synthesized diaspore in a three stage process: firstly amorphous Al(OH)J was precipitated from Al2(S04)J with excess ammonia; secondly, liquid air was poured over the gel, which transformed it to gibbsite; finally, the gibbsite was refluxed for 450 hours in O,lM NaOH to form diaspore. They were also able to form boehmite by the same process, if the initial precipitation was performed with stoichiometric amounts of ammonia.
It can be seen that there are many similarities between all these procedures.
20
In investigating the preparation of thin aluminium films for use as catalyst supports by evaporation of aluminium an oxidising atmosphere, Lamber (1986) observed that such films are easily hydrated by distilled water. Films of aluminium oxide were immersed in water at between and 87°C and held there for 1 to 2 minutes, after which electron micrographs revealed needle-shaped crystallites in an amorphous matrix.
interest in boehmite as a material for use in catalysts has resulted in filing of numerous patents covering various methods of synthesis. example, in 1986 Misra and Sivakumar, of the Aluminium Company of America, published a patent for preparation of boehmite by heating a solution containing caustic soda and Al20) to between 115°C and 145°C, then the solution with boehmite or pseudoboehmite and ageing the seeded solution for hours. are in the literature to many other patents of novel syntheses or manufacturing methods, European countries, Canada, USSR, Japan and other wide-ranging sources. It was not possible to examine many of these, but they are plentiful enough warrant care if patent infringements are be avoided.
Apparently there not been same degree of interest diaspore. Mukaibo al (1969) report that diaspore can be prepared by treating boehmite in a NaOH solution, seeded with naturally occurring diaspore, an autoclave at 380°C and 500 for one week. Wade and Bannister, in the text "Comprehensive Inorganic Chemistry", quote ed.(1970) as saying all aluminium hydroxides and oxides, when treated hydrothermally at pressures over 140 atm. and temperatures between 2750 and
will be converted to diaspore if seeded with naturally occurring diaspore in this way.
Krause al (1970) synthesized diaspore in a three stage firstly amorphous Al(OHh was precipitated from A12(S04)3 with excess ammonia; secondly, liquid was poured over the gel, transformed it to gibbsite; finally, the gibbsite was ret]luxc:~d for hours in O,IM NaOH to form diaspore. They were also to form boehmite by the same process, if the initial precipitation was performed with stoichiometric amounts ammonia.
It can be seen there are similarities between all these procedures.
Univers
ity of
Cap
e Tow
n
3.2 MECHANISMS OF FORMATION.
In this section an overview of the attempts by many researchers to formulate a consistent and useful mechanism of formation of aluminium hydroxide and oxyhydroxide polymorphs is presented.
3.2.1 Formation of the structUral units.
Hsu and Bates (1964) proposed the following mechanism for the formation of aluminium hydroxides:
21
An Al3+ ion in solution is co-ordinated to six H20 molecules, contributing half a positive charge to each water molecule. When NaOH is added to such a solution, one of these H20 molecules is replaced by an Oir ion. This ion, with its charge being only half satisfied, is able to co-ordinate to a second Al3+ ion, so acting as a bridge. In this way the hydroxy aluminium ions polymerise, until they form a six-member ring unit, A16(0H)i2
6+, which is the basic structural unit of the various crystalline and . amorphous forms.
Hsu and Bates suggest that on adding NaOH to a solution containing aluminium ions,
the reaction
AI3+ + OH- = Al(OH)2+
immediately occurs and, being unstable with respect to charge, polymerization would immediately follow, the complete scheme being
[ Al(H20)5(0H)]2+ + H20 [ Al(H20)4(0H)2]1+ + H10 Al6(0H)i2
6+ .12 H20
Although the principle of polymerization was accepted widely, the compounds suggested as products varied a great deal. For example, Matijevic and Tezak (1953) suggested that the monomers [Al(H20)50H]2+ join to form
Brosset et al (1954) first proposed a ring-like structure AI6(H20)153+ as the main
product. Hsu and Rich ( 1960), from studies of aluminium fixation in cation exchange resins, proposed that various polymers form, but that the most stable and consequently predominant structure would consist of a single six aluminium ion ring surrounded by twelve water molecules (Figure 3 .1). They added that the single ring structure may be the result of the limited space available in the resin pores.
3.2 MECHANISMS OF FORMATION.
this section an overview of the attempts by many researchers to formulate a consistent and useful mechanism of formation aluminium hydroxide oxyhydroxide polymorphs is presented.
1 Formation of the structUral units.
Hsu and Bates (1964) proposed the following mechanism for the formation of aluminium hydroxides:
21
An Al3t ion in solution co-ordinated to six H20 molecules, contributing half a positive charge to each water molecule. When NaOH added to such a solution, one of H20 molecules is by an ion. This with its being only half satisfied, is able to co-ordinate to a second Al3t ion, so acting as a In this way the hydroxy aluminium ions polymerise, until they form a six-member ring unit, Al6(OH)126t• which is the basic structu:ra1 unit of the various crystalline and . amorphous forms.
Hsu and Bates suggest that on adding reaction
to a solution containing aluminium ions.
= Al(OH)2t
immediately occurs and, unstable with respect to charge, polymerization would immediately follow. the complete scheme being
[Al(H20)S(OH)]2t + H20 [Al(H20)4(OH)2]lt + A16(OHh2 6+ .12
Although the principle of polymerization was accepted widely, the compounds suggested as products varied a deal. example, Matijevic and (1953) su~:~ested that the monomers [Al(H20)50H]2t join to form
Brasset et al (1954) first proposed a structure Al6(H20)153t as the main product. Hsu and Rich (1960), from studies of aluminium in cation exchange resins, proposed that various polymers form, but that the stable and consequently predominant structure would consist of a single aluminium ion surrounded by twelve water molecules 3.1). They added that single structure be the result of limited available resin pores.
Univers
ity of
Cap
e Tow
n
22
Fig. 3.1: Proposed structure of an hydroxy-aluminium polymer having an average positive charge of 1 per Al-atom (after Hsu and Rich, 1960)
Hsu and Bates proposed that the OH- on the ring units would continue attaching other Al3+, forming larger and increasingly basic double, triple and higher order rings in preference to neutral Al(OH)3 molecules. However, because of the difference in charge.density between these rings and Al3+, very large polymers would not form until all the aluminium ions were formed into ring units. The charge per aluminium ion decreases from 3+ to 1 + as the OH/ Al ratio increases, but the overall charge per unit increases, causing the units to repel, so slowing the rate of polymerization. Only when the ratio reaches 3 and sufficient hydroxyl ions are available to cancel the net positive charge, are they able to cluster together, forming crystalline Al(OH)J.
·Hem and Roberson ( 1967) proposed that the strong positive charge of the aluminium ion repels the two protons of co:..ordinated water molecules. As the pH of the solution increases, a proton is eventually driven off, forming the complex ion [Al(OH)(OH2)5] 2+. Two of these may lose two water molecules and form a dimer:
Further dehydration and polymerization gives a ring structure of six octahedrally coordinated Al ions: [ Al6(0H)i2(H20)i2J6+. The further growth and stacking of these units gives rise to the various polymorphs.
In more alkaline solutions, the dominant species is thought to be [Al(OH)4]1-.
Precipitation occurs according to
accompanied by an increase in pH of the solution.
22
Fig. 3.1: Proposed of an hydroxy-aluminium polymer having an average positive charge of 1 AI-atom (after and Rich, 1960)
Hsu and Bates proposed the OH- on the ring units would continue attaching other AI3+, forming larger and increasingly basic double, triple and higher order rings preference to neutral Al(OH)3 molecules. However, because the difference in charge. density between these rings and AI3+, very large polymers would not until all the aluminium were formed into ring units. The charge per aluminium ion decreases 3+ to 1+ the OH/AI ratio increases, but overall charge per unit J.ll"'JL'-«,',.,..,. causing the units to so slowing rate of polymerization. Only when the ratio reaches 3 sufficient hydroxyl ions are available to cancel net positive charge, are they able to together, crystalline Al(OHh .
. Hem and Roberson (1967) proposed that strong positive charge of the aluminium ion repels the two protons of co ... ordinated water molecules. As pH of the solution increases, a proton is eventually driven off, forming complex ion [Al(OH)(OH2)sl2+ . Two of these may lose two water molecules and a dimer:
l-in'f'1r ... ",,?' dehydration and polymerization gives a ring structure of six octahedrally co-ordinated ions: [A16(OH)12(H20h2]6+. The further growth and stacking of these
gives rise to the various polymorphs.
more alkaline solutions, the dominant soeC:les is thought to be [Al(OH)4]1-. Precipitation occurs according to
accompanied by an increase in of the solution.
Univers
ity of
Cap
e Tow
n
23
It is known from synthesis experiments that acid environments favour the precipitation of gibbsite, while alkaline environments favour bayerite. Also, it has been stated above that the only difference between the various crystalline forms is the stacking sequence of the basic structural units. The question is: how may these factors be linked?
I.
Aluminium is amphoteric and can exist either as an anion or as a cation in solution, depending on pH. In acid solutions (pH 4 to 6) the cationic aluminium complex [Al(OH)(H20)5] 2+ predominates. In this complex the full polarising power of the AI3+ is felt by one hydroxyl ion, making it energetically favourable for it to undergo tetrahedral hybridisation. As this ion polymerises, the polarising effect is reinforced by the influence of additional aluminium ions. Consequently, the compound formed from an acid solution should contain highly polarised hydroxyl ions, leading to the formation of hydrogen bonds and to the ABBA-ABBA stacking, which is characteristic of gibbsite.
In alkaline solutions the anionic complex [Al(OH)4]1- predominates. Here the excess
negative charge reduces the polarising effect of the Al3+ considerably. Hybridisation is energetically unfavourable; consequently the precipitate from an alkaline environment should contain weakly polarised hydroxyl ions leading to an open structure.
3.2.2 Formation of the hydroxides.
Bye and Robinson (1964), studying the crystallisation processes of hydroxide gels from systems as free as possible from foreign ions, summarized the relationships as:
freshly precipitated --- > amorphous gel
pseudoboehmite-- > bayerite* I (high pH)
~ gibbsite*
* = nordstrandite may form
Each step is accelerated by increased pH and retarded by impurity ions. Progress of crystallisation was followed by determining changes of the specific surface area, which is affected by the following three factors:
a) Recrystallisation either without change in crystal form (Ostwald ripening), or with change to a more stable crystal form.
b) Aggregation of small particles. c) Fracture of particles from strain arising from internal reordering.
Processes. (a) and (b) decrease , but process (c) increases the surface area.
23
It known from synll:beS.ls experiments that favour gibbsite, while environments favour Also, it has
above that the only between the various crystalline forms is the Sta<:la1lg sequence of the units. The question how may these factors be linked?
i"
Aluminium is amphoteric can exist either as an anion or as a cation solution, depending on pH. In solutions (PH 4 to 6) the aluminium COlltlplt::X
[Al(OH)(H20)s]2t In this full polarising Al3+ by one hydroxyl for it to undergo
tetrahedral hybrldisation. As this ion polymerises, the polarising effect is reinforced by the influence of additional aluminium ions. Consequently, the compound formed from an acid solution should contain highly polarised hydroxyl ions, leading to formation of hydrogen bonds and to the ABBA-ABBA stacking, which is I"h~"~I"'t.,.... ... ttl"
gibbsite.
aJ..k;aJ1IIle solutions complex [Al(OH)4]1- predominates. Here excess npoc'!lItn:lpo charge reduces polarising effect of the A13t considerably. Hybridisation is energetically unfavourable; consequently the precipitate from an alkaline environment should contain weakly polarised hydroxyl ions leading to an open structure.
Formation of the hydroxides.
Robinson (1964), "' ... .,"'""'.n cry:Stallllsal[10n J)fOlces:ses of hydroxide ~~tpotTl~ as free as possible
freshly precipitated --- > amorphous gel
* = nordstrandite may
pseudoboehmite-- >
relationships as:
bayerite* I (high pH)
~
step is accelerated by pH and retarded by impurity ions. Progress of crystallisation was followed by determining changes of the specific surface area, which is by the following factors:
a) without change change to a more
particles.
RecrystaUisation ripening), or Aggregation Fracture n!lirhrll"'iI;1 from strain aU.~'LUE>
crystal form (Ostwald form.
.......... u ... ,'" reordering.
t'rCtCe~.ses (a) and (b) decrease, but process (c) increases the surface area.
Univers
ity of
Cap
e Tow
n
24
Two stages were noted: firstly, the surface area increased as pseudoboehmite appeared; secondly, it decreased as bayerite appeared. Increased surface area during initial ageing, ascribed to process (c), is said to confirm that pseudoboehmite is formed by condensation polymerization.
Aldcroft and Bye ( 1969) proposed that the condensation polymerization in the amorphous hydroxide suspension occurs by the formation of a boehmitic group, with net volume increase, as first suggested by Fricke and Meyring (1933) and Feitknecht (1954), for the formation of boehmite:
---->
furthermore, ligands may rearrange to form an hydroxide link
---->
The rate of this process was determined by Aldcroft and Bye and found to fit the Avrami-Erofe'ev equation: -log(l- a)= (kt)n where a = fractional volume change and t =time.
Aldcroft et al (1969) continued the investigation into the second stage, namely the conversion of pseudoboehmite to trihydroxide by dissolution and recrystallisation. They proposed that the rate of formation of a trihydroxide would initially be determined by the rate of random nucleation and growth, which in turn are governed by the rate of dissolution of the pseudoboehmite. They found the kinetics of the reactions to be first order, while at a later stage the rate of the mass transfer processes becomes the governing factor.
The effect of the ageing medium, in particular organic solvents, on the process was examined in all three studies, and it was found that ethanol promoted aggregation relative to water, while glycerol inhibited it. This variation was ascribed to their differing ability to form a salvation sheath on the aggregate surfaces. In addition, ethanol reduces the dielectric constant at the particle surfaces, promoting aggregation, which in turn appeared to promote the condensation process by which pseudoboehmite is formed. It should, at the same time, inhibit the dissolution process by reducing the exposed surface area. This wmlld have the net effect of stabilising pseudoboehmite relative to bayerite, which was in fact observed by Aldcroft et al (1969). However, a second type of gel is mentioned, pseudoboehmite-c, which has a more strained lattice. This strain appears to increase solubility to such an extent that crystallisation of bayerite is rapid and linear with time.
24
Two stages were noted: firstly, the surface area increased as pseudoboehmite appeared; secondly, it decreased as bayerite appeared. Increased surface area during initial ageing, ascribed to process (c), is said to confirm that pseudoboehmite is formed by condensation polymerization.
Aldcroft and Bye (1969) proposed that the condensation polymerization in the amorphous hydroxide suspension occurs by the formation of a boehmitic group, with net volume increase, as first suggested by Fricke and Meyring (1933) and Feitknecht (1954), for the formation of boehmite:
---->
furthermore, ligands may rearrange to form an hydroxide link
---->
The rate of this process was determined by Aldcroft and Bye and found to fit the Avrami-Erofe'evequation: -log(1- a)= (kt)n where a = fractional volume change and t = time.
Aldcroft et al (1969) continued the investigation into the second stage, namely the conversion of pseudoboehmite to tribydroxide by dissolution and recrystailisation. They proposed that the rate of formation of a tribydroxide would initially be determined by the rate of random nucleation and growth, which in turn are governed by the rate of dissolution of the pseudoboehmite. They found the kinetics of the reactions to be first order, while at a later stage the rate of the mass transfer processes becomes the governing factor.
The effect of the ageing medium, in particular organic solvents, on the process was examined in all three studies, and it was found that ethanol promoted aggregation relative to water, while glycerol inhibited it. This variation was ascribed to their differing ability to form a solvation sheath on the aggregate surfaces. In addition, ethanol reduces the dielectric constant at the particle surfaces, promoting aggregation, which in turn appeared to promote the condensation process by which pseudoboehmite is formed. It should, at the same time, inhibit the dissolution process by reducing the exposed surface area. This would have the net effect of stabilising pseudoboehmite relative to bayerite, which was in fact observed by Aldcroft et al (1969). However, a second type of gel is mentioned, pseudoboehmite-c, which has a more strained lattice. This strain appears to increase solubility to such an extent that crystallisation of bayerite is rapid and linear with time.
Univers
ity of
Cap
e Tow
n
_,
-~
25
McHardy and Thomson ( 1971) found that when aluminium hydroxide is precipitated in aqueous suspension, two different types of gel may be formed: In the absence of acids, an uncharged gel, which they called pseudoboehmite, forms and rapidly crystallises to bayerite. However, in acid solutions, a positively charged gel, which they called pre-gibbsite gel, may crystallise slowly to gibbsite if extraneous anions are removed. The pseudoboehmite was characterised by reflections at 6,5; 3,2; 2,35; 1,85; and 1,42 A, while the pre-gibbsite gel gave reflections at 1,85,and 1,439 A only. The variation of pH with time is given in Figure 3.2. In stage A to Ba positively charged colloid with basic counterion forms; the gel acts as a base and a reversible neutralizing reaction occurs. During stage B to C pre-gibbsite gel slowly • transforms to gibbsite, the gel acting as an acid toward the solution.
pH
8
3.8 ._____,.___.. _ __._ _ _._ _ _,__ _ _.___..__-..J._____._-.1 0 1 a a 4 e e T e 9 10
Time of ageing (days)
Fig. 3.2: Variation in pH with time.
Ng Kee Kwong and Huang (1977, 1979) and Violante and Violante (1980) investigated the influence of organic acids, such as may be found in natural systems, on the crystallisation of aluminium hydroxides. They found that the acids hindered the crystallisation of the hydroxides, because of their tendency to block the co-ordination sites on the charged edges of the hydroxyaluminium polymers through forming complexes, thus obstructing the hydrolysis mechanism by which the polymerization proceeds (see Figure 3.3).
25
McHardy and Thomson (1971) found that when aluminium hydroxide is in aqueous suspension, two different types of gel be formed: In
an uncharged they called pseudoboehmite, forms ......... 'u ........ to bayerite. in acid solutions, a positively charged which
they called pre-gibbsite may crystallise slowly to gibbsite if extraneous anions are removed. The pseudoboehmite was characterised by reflections at 6,5; 3,2; 1,85; and 1,42 A, while the pre-gibbsite gel gave reflections at 1,85 iand 1,439 only. The variation with time is given in In stage A to B a positively charged colloid with basic counterion the gel acts as a a reversible neutralizing occurs. During B to pre-gibbsite ttalllstc)rnlS to gibbsite, as an acid solution.
pH
B
3.8 '----'_--'-_--'-_-I......_.L.---I_--L._-I...._...J..----J
o I 3 458 1 I I ~
Time of ageing (days)
Variation in pH with time.
Kwong and Huang 1979) and Violante (1980) mVlesti:gat~~ influence of organic as may be found systems, on the
i)!.AIIUi)'tI,U'-'U of aluminium They found acids hindered the crystallisation of the hydroxides, because of their tendency to block the co-ordination
on the charged edges the hydroxyaluminium polymers through forming complexes, thus obstructing hydrolysis mechanism which the polymerization proceeds ( see Figure 3.3).
Univers
ity of
Cap
e Tow
n
26
0
~ C--CH1
II 0
Fig. 3.3: Hydroxyl bridging mechanism of edge-Al is hampered by the citrate.
This mechanism, together with pH, and the steric hindrance of the often large organic groups, promoted the formation of amorphous structures. The promotion of amorphous forms increases with increasing organic acid concentration, but also with increasing affinity of the chelating group for aluminium. The order of effectiveness was found to be: tartaric > citric > malic ~ salicylic > oxalic > aspartic > p-hydroxybenzoic acid.
Barnhisel and Rich ( 1965), working from the theories derived from the study of pure systems, investigated the effect of other factors encountered in natural systems, such as the presence of salts. They produced the phase diagram reproduced in Figure 3.4, showing the influence of salts on the phases which may form. The broad lines indicate a large degree of overlap, and the authors note that experimental conditions may change the shape of the diagram considerably.
26
"'", I H
AI
/ H
C-CH,COO o ~
C-CH!
II o
3.3: Hydroxyl ............ I'> .... ,LI'> mechanism of ........ "'''_ iL:I. .. is hampered by the cltt'8te
mechanism, together with pH, and the sterlc the often 2fCIUDS. promoted the formation of amorphous The promotion ....... ,n ....... , ......... '" forms increases increasing organic concentration, but also with mClrea~img affinity of the chelating group for aluminium. The order of ett(~ctt'ven,ess was found to be: tartaric > citric > malic ~ salicylic > oxalic > ~~i~rtll" > p-hydroxybenzoic
Barnhisel and Rich (1965), working from the from the study of investigated other factors in natural as
presem;e of the phase diagram reproduced in Figure the influence phases which may broad lines indicate
a degree of overlap, and authors note that conditions may cnanee the shape of the diagram considerably.
Univers
ity of
Cap
e Tow
nFig. 3.4:
NaCl (me./ g.) 1s.--~~~.;;;._~~~---,....--~~~~~~~~-.
N Nordatrandlte B Bayerite
12 A
8
8 7
pH
G Glbbaite
Na. B
8
,, ,,.,, ,,, ;,
/I , , 1 1 B
9 10
Effect of final pH and salt concentration on polymorph formation.
27
Hsu ( 1966) examined the effect of salts on the preferential formation of bayerite or pseudoboehmite, constructing the phase diagram reproduced in Figure 3.5. At low NaCl concentrations stable, well-crystallized bayerite formed quickly, while at high concentrations stable pseudoboehmite was formed. In the intermediate range both were formed, with the pseudoboehmite gradually transforming to bayerite with ageing. From this Hsu concluded there was an irreversible or slowly reversible step in the process of formation. He proposed that, if the salt concentration is very high in the early stages of formation, the salt, in forming its own hydration sheath, may dehydrate the
/OH°"' Al""OH/"'1\1 links in the gel to Al--0--Al;
the latter would rehydrate only very slowly. He reported the relative effectiveness of anions in stabilising pseudoboehmite as: er> c104- > No3- >Br-> r.
27
NaCI (me.! gJ 18
N Nordatrandlte /'-11 // I B Bayerite
// II G Gibbaite
12 A // II // II
// II // II NIB
/'/ II a ,./ ) I ,. ""'-::;. NIG / ",' 1/ ,'\ , "'" 1/ \ \ //
1/ \ /,
" .... /, ....
'/ N,G I /1
G :1 " /' B 0 I
3 " 15 8 1 a 9 10
pH
Effect of final and salt concentration on polymorph formation.
(1966) examined salts on the formation of or pseudoboebmite, phase diagram reproduced in Figure At low
concentrations stable, well-crystallized bayerite formed quickly, while at concentrations stable pseudoboehmite was formed. the intermediate range both formed, with the pseudoboebmite gradually transforming to bayerite with ageing.
this Hsu concluded was an irreversible or slowly reversible step in process of formation. He proposed that, if the salt concentration is very high
of formation, forming its own sheath,
the gel to AI--O--Al;
would rehydrate only very slowly. He .. p.nnri~>rl effectiveness of stabilising as:
> ClOt> N03- >
Univers
ity of
Cap
e Tow
n
28
NaCl cone., M
0.08 0.24 0.80 1.2 2.0 s.o 4.0 5.0 • • I
Pure . I
s.oo Paeudoboehmlte - I
s.oe Mixture NaOH/AI .;
~Fl. S.15
- Pure Bayerlte
s.so
at•r Hau, 1818
Fig.3.5: Effect of OH/ Al ratio and salt concentration on polymorph formation.
Interesting experimental evidence is presented by Turner and Ross (1970) to support their somewhat contradictory proposed mechanism. Analysing the amount of polynuclear hydroxyaluminium ions present in solution by the 8-quinolinolate extraction method, they found that the concentration of such ions was initially relatively low, increasing rapidly to a maximum, and then decreasing, while for the solids the inverse occurred, as shown by Figure 3.6.
Fig.3.6:
26 deg.C 70
60
50
40
30
20 Mononuclear Iona
10
O'--~-'-~~'--~-'-~~..._~_._~~'--~-'-~--'
0 10 20 30 40 50 60 70 80
Ageing time (days)
Effect of temperature and ageing time on the concentration of solid material, mononuclear ions and polynuclear ions.
The length of time to reach these maxima and minima, as well as for the polynuclear ion concentration to reach zero, increased with decreasing temperature and with increasing AICI3 concentration. Above 0,6 M AIC13 aluminium hydroxychloride and
28
NaCI conc., M
O.OB 0.24 O.BO 1.2 2.0 8.0 4.0 1.0 • • • Pure •
8.00 P.eudoboehmlte - J
Mixture .. 8.0e
NaOH/AI M.R. 8.11
~ Pure Bayerite
8.80
Effect of OHI Al ratio and salt concentration on polymorph formation.
Interesting experimental evidence is presented by Turner Ross (1970) to support their somewhat contradictory proposed mechanism. Analysing the amount of polynuclear hydroxyaluminium ions in solution by the 8-quinolinolate extraction method, they found that the concentration of such ions was initially relatively low, rapidly to a maximum, and then decreasing, while for the solids inverse occurred, as shown by Figure 3.6.
% AI in system .nll!' "IIImIII' I RON, 111811 80.-----~------------------~~~~==~
26 de;.C
10
O~--~--~----~--~----~--~--~--~ o 1D 20 aD 40 eo eo 10 80
Ageing time (days)
.................. of temperature and ageing time on the concentration of solid material, mononuclear and polynuclear ions.
The length of time to reach these maxima and minima, as well as for the polynuclear ion concentration to reach zero, increased with decreasing temperature and with increasing AlCI) concentration. Above M AlCI) aluminium hydroxychloride and
Univers
ity of
Cap
e Tow
n
29
not gibbsite crystallized. These findings agree with the proposal of Fripiat and Pennequin (1965) that the first solid to form on addition of a base to Alel3 solution contains er ions substituting for OH- ions, which prevent trihydroxide forming. If these er ions are washed out soon after addition, gibbsite crystallises rapidly.
Turner and Ross ( 1970) found that the first order rate constant for the reaction between the polynuclear ions and the 8-quinolinol mixture was constant at constant temperature and independent of ageing time or extent of neutralisation. This evidence is said to contradict the mechanism proposed by Hsu and Bates (1964), whereby the aluminium hydroxide is formed by polymerization, the polynuclear ions increasing in size and changing the OH/ Al ratio, which would cause the rate constant to vary with the ageing time or neutralisation. They suggest that only one polynuclear cationic species may exist, though it may be any of those proposed by other researchers.
Turner and Ross (1970) propose the following mechanism:
When base is added to an Alel3 solution, the most rapid reaction is:
An initial solid high in er and H20 is formed, which is soluble. Simultaneously, but more slowly, the polynuclear cations form:
The initial solid product is metastable with respect to the cations; consequently equilibrium shifts to the left in reaction I and the solid dissolves. In addition, Turner and Ross propose a third reaction consecutive to reaction I, which is a gradual solid phase reaction involving the elimination of er and H20 and rearrangement to gibbsite, which is insoluble and stable with respect to the cations:
The equilibrium of reaction II then goes to the left and solid trihydroxide particles begin to grow as the cation concentration decreases.
There is considerable experimental evidence presented to support this mechanism, and the authors are also able to interpret the findings of other researchers in the light of their mechanism.
In 1976 Stol, Van Heiden and De Bruyn published an investigation of the kinetics of the hydrolysis process. Improved techniques facilitated the investigation of the individual stages leading to precipitation. Continuous titrations, relaxation and dilution tests were used to investigate the effect of aluminium concentration, ionic strength and
temperature.
gibbsite crystallized. These findings agree with the proposal of Fripiat and pellllle<lIUm (1965) to form on of a base to solution
ions OH- ions, tribydroxide If cr ions are washed out soon after addition, gibbsite crystallises
and Ross (1970) found that the first order constant for the between the polynuclear ions and the 8-quinolinol mixture was constant at constant temperature and independent of or extent of neutralisation. This evidence is said to contradict the mechanism proposed by Hsu and (1964), whereby the aluminium
is formed ions increasing and changing the OHI AI which would cause to vary the ageing
or neutralisation. suggest that only one polynuclear cationic speC:les though it may be of those proposed by other researchers.
and Ross (1970) propose the following mechanism:
When base is to an AICl3 solution, the most rapid reaction
An initial solid CI- and H20 is which is soluble. Simultaneously, but more slowly, the polynuclear cations form:
The initial solid equilibrium in reaction I and the solid 111""<'1">1",."'.,
Turner and Ross a third reaction consecutive to reaction I, which is a gradual solid reaction involving elimination of cr and H20 and rearrangement to gibbsite, which is insoluble and stable with the cations:
The equilibrium particles begin to
AI(OHh + (3-n)HCl +
reaction IT then to as the cation concentration decreases.
There is considerable experimental presented to support mechanism, and authors are also able to interpret the findings of other researchers the light of their mechanism.
1976 Stol, Van and De Bruyn an investigation of kinetics of hydrolysis process. Improved techniques the mVestllgal10n
individual stages leading to precipitation. Continuous titrations, relaxation dilution were used to the effect of aluminium concentration, ionic strength and
temperature.
Univers
ity of
Cap
e Tow
n
30
In conjunction with these titrations, light scattering experiments were conducted to ascertain the size of particles involved in the various stages of reaction. The turbidities determined where OH/ Al < 2,5, were low and stable over an eight month period. In samples where OH/Al > 2,5, a sharp increase in turbidity was noted. Measured turbidity was translated into degree of polymerization by the method outlined by Tobias and Tyree (1959). The resulting curve is reproduced in Figure 3.7.
Average degree of polymerlsatlon 20.--~-"'"-~_.;;;..~~....;_~~~~~~-
16
10
6
o.._~_._~_...~~~~_._~~..__~~
0 0.6 1 1.6 2 2.6 3
OH/Al
Fig.3.7: Average degree of polymerization versus OH/Al ratio.
Interpreting the titration results, Stol et al propose that at low Af3+ concentrations (5x10-5 M), only monomeric species exist [Al(OH)2+; Al(OH)2+; Al(OH)3]. At higher concentrations, but within the range 0 > OH/Al < 0.5, only monomeric and dimeric species exist [Al2(0Hhl4+].
Increasing the ionic strength will promote the dimerisation reaction:
2Al3+ + 4H 0 - Al (OH) 4+ + 2H o+ 2 - 2 2 3
For higher concentrations and 0,5 < OH/ Al < 2,2, the composition of the solutions cannot be explained without introducing the concept of polymerization. Relaxation experiments and light scattering measurements both showed the polymerization which occurs on this first plateau to bC rapid and to increase in degree. The further increase in polymerization on the addition of salts is ascribed to the screening of the high positive charge on the polynuclear compounds by the added negative ions.
30
In conjunction with these titrations, light scattering experiments were conducted to ascertain the size of particles involved in the various stages of reaction. The turbidities determined where OH/AI < 2,5, were low and stable over an eight month period. In samples where OH/AI > 2,5, a sharp increase in turbidity was noted. Measured turbidity was translated into degree of polymerization by the method outlined by Tobias and Tyree (1959). The resulting curve is reproduced in Figure 3.7.
Average degree of polymerlsatlon 20r---~--~----~~------~--~
16
10
o~--~--~----~--~----~--~ o 0.6 1 Ui 2 2.6 3
OH/AI
Fig.3.7: Average degree of polymerization versus OH/AI ratio.
Interpreting the titration results, Stot et al propose that at low AI3+ concentrations (5xlO-5 M), only monomeric species exist [AI(OH)2+; AI(OH)2 +; AI(OHh]. At higher concentrations, but within the range 0 > OHI AI < 0.5, only monomeric and dimeric species exist [AI2(OHh]4+].
Increasing the ionic strength will promote the dimerisation reaction:
2AI3+ + 4H 0 - Al (OH) 4+ + 2H 0+ 2 - 2 2 3
For higher concentrations and 0,5 < OHI AI < 2,2, the composition of the solutions cannot be explained without introducing the concept of polymerization. Relaxation experiments and light scattering measurements both showed the polymerization which occurs on this first plateau to be rapid and to increase in degree. The further increase in polymerization on the addition of salts is ascribed to the screening of the high positive charge on the polynuclear compounds by the added negative ions.
Univers
ity of
Cap
e Tow
n
31
Further kinetic studies using the "go-stop-reverse-go" relaxation method showed that while the formation of polymers was fast relative to the rate of addition, their breakdown was slow, indicating that the two processes are not simply the inverse of each other. The breakdown of the polymers by acid was found to be a second order reaction and the proposed mechanism is shown in Figure 3.8.
n+ n•1 n+1
"v/ "v/ "V/ Al Al Al
/!'\,~ /l" alow /l"· faat 1 OH • H•-
"V/o~ -"v/ - H2o "V/o~ Al Al
/l" /:" Al
/!" I
----·-·-· ----
Fig.3.8: Mechanism of breakdown of polymers by acid.
The slow loosening of the Al-OH-Al bridge is followed by the fast breaking of the bond and addition of H20.
A structural model was proposed consistent with the experimental finding that within the range 0,5 < OH/ Al < 2,5 the majority of OH- ions are bridging. Possible compounds not resembling the crystalline structure were eliminated as options. The resulting structures are consistent with the proposals of Hsu (1964) and with the light scattering data.
In summary, Stol et al (1976) found that in the range 0 < OH/Al < 0,5 or for very low Al3+ concentrations, monomeric species dominate. Thereafter relatively small polymeric compounds are formed, with a rapid increase in the degree of polymerization. At OH/ Al = 2,5, hexameric polymers dominate, and although these may grow, a different process begins by which non-bridging OH- ions are attached to the edges of the polynuclear complex, reducing the high charge at the edges and thus permitting aggregation and the formation of the solid precipitate. This process consumes the excess OH- ions, causing the appearance of the second plateau, as seen in Figure 3.7. This may be regarded as the nucleation stage preceding precipitation. The prevalence of amorphous forms is understood in the light of the steric constraints on crystallisation imposed by the orientation of the hexameric plates relative to one another.
31
Further kinetic studies using the "go-stop-reverse-go" relaxation method showed that while the formation of polymers was fast relative to the rate of addition, their breakdown was slow, indicating that the two processes are not simply the inverse of each other. The breakdown of the polymers by acid was found to be a second order reaction and the proposed mechanism is shown in Figure 3.8.
11+ n+1 11+1
"v/ "V/ "V/ AJ AJ AJ
/!'\,~ /1" slow /1". faat
"H" -- --I OH - "V/O~ H2O "V/ "V/O~ AJ AJ
/1" /:" AJ
/1" I
----.~.-. ----
Fig.3.8: Mechanism of breakdown of polymers by acid.
The slow loosening of the AI-OH-AI bridge is followed by the fast breaking of the bond and addition of H20.
A structural model was proposed consistent with the experimental finding that within the range 0,5 < OHt AI < 2,5 the majority of OH- ions are bridging. Possible compounds not resembling the crystalline structure were eliminated as options. The resulting structures are consistent with the proposals of Hsu (1964) and with the light scattering data.
In summary, Stol et al (l976) found that in the range 0 < OHtAI < 0,5 or for very low A13+ concentrations, monomeric species dominate. Thereafter relatively small polymeric compounds are formed, with a rapid increase in the degree of polymerization. At OHtAI = 2,5, hexameric polymers dominate, and although these may grow, a different process begins by which non-bridging OH- ions are attached to the edges of the polynuclear complex, reducing the high charge at the edges and thus permitting aggregation and the formation of the solid precipitate. This process consumes the excess OH- ions, causing the appearance of the second plateau, as seen in Figure 3.7. This may be regarded as the nucleation stage preceding precipitation. The prevalence of amorphous forms is understood in the light of the sterlc constraints on crystallisation imposed by the orientation of the hexameric plates relative to one another.
Univers
ity of
Cap
e Tow
n
32
3.2.3 Formation of the Oxyhydroxides.
The oxyhydroxides form by solid state reaction when hydroxides are subjected to conditions of high temperature and pressure. The formation of the structural units is the same for oxyhydroxides as for hydroxides, the differences again arising in the relative stacking of these cell units as discussed fully in section 3 .2 .. 1.
I
In all the investigations where the presence of true boehrnite and diaspore have been confirmed, a hydrothermal treatment step has been reported. However, there is some controversy in the literature as to whether these forms are stable polymorphs or ~
metastable states induced by the hydrothermal conditioning. As with the amorphous gels, there is ambiguity about the phase boundaries defining the polymorphs.
Juan and Lo (1975) investigated the formation of boehmite and other natural deposits of aluminous material. They proposed that a bauxite deposit consisting primarily of boehmite would form by burial metamorphism of weathered argillaceous limestone. To simulate this, they prepared mixtures of Cao. AI20 3 and silicic acid and subjected these to a range of temperatures and pressures in a hydrothermal reactor for 120 to 168 hours. Temperatures ranged from 302°C to 374°C and the corresponding pressures from 68910 kPa (10000 psi) to 172 275 kPa (25000 psi), under which conditions boehmite was easily formed.
Laubengayer and Weisz (1943) reported that gibbsite converted to boehmite at 150°C, which in turn converted to diaspore at 275°C, and finally to corundum at temperatures above 400°C and pressure of 40 000 kPa (400 bars). They suggested that boehmite was not a stable phase, but that the rate of change to diaspore was very slow. On the other hand, Ervin and Osborn (1951) considered boehmite to be a stable phase which could be converted to corundum at temperatures above 300oC and pressures below 13782 kPa (2000 psi), but which would convert to diaspore at higher pressures (see figure 3. 9).
Kennedy (1959) found that gibbsite converted to boehmite below 260°C, while gibbsite converted to a mixture of diaspore and boehmite at temperatures above 260°C and pressures of 2,6 MPa (26 kbar) (see figure 3.10). He concluded that while boehmite was metastable, diaspore was the stable phase.
While Juan and Lo's work could not resolve this disagreement, they did find that boehmite was never converted directly to corundum, implying that a stable phase having a slow nucleation rate could exist between the phase fields of the two forms, and that this phase could be diaspore.
32
3.2.3 Formation of the Oxyhydroxides.
The oxyhydroxides form by solid state reaction when hydroxides are subjected to conditions of high temperature and pressure. The formation of the structural units is the same for oxyhydroxides as for hydroxides, the differences again arising in the relative stacking of these cell units as discussed fully in section 3.2.1.
I
In aU the investigations where the presence of true boehmite and diaspore have been confirmed, a hydrothermal treatment step has been reported. However, there is some controversy in the literature as to whether these forms are stable polymorphs or metastable states induced by the hydrothermal conditioning. As with the amorphous gels, there is ambiguity about the phase boundaries defining the polymorphs.
Juan and 1.0 (1975) investigated the formation of boehmite and other natural deposits of aluminous material. They proposed that a bauxite deposit consisting primarily of boehmite would form by burial metamorphism of weathered argillaceous limestone. To simulate this, they prepared mixtures of CaO. Al20 3 and silicic acid and subjected these to a range of temperatures and pressures in a hydrothermal reactor for 120 to 168 hours. Temperatures ranged from 302°C to 374°C and the corresponding pressures from 68910 kPa (10000 psi) to 172275 kPa (25000 psi), under which conditions boehmite was easily formed.
Laubengayer and Weisz (1943) reported that gibbsite converted to boehmite at 150°C, which in tum converted to diaspore at 275°C, and finally to corundum at temperatures above 400°C and pressure of 40000 kPa (400 bars). They suggested that boehmite was not a stable phase, but that the rate of change to diaspore was very slow. On the other hand, Ervin and Osborn (1951) considered boehmite to be a stable phase which could be converted to corundum at temperatures above 3000C and pressures below 13782 kPa (2000 psi), but which would convert to diaspore at higher pressures (see figure 3.9).
Kennedy (1959) found that gibbsite converted to boehmite below 260°C, while gibbsite converted to a mixture of diaspore and boehmite at temperatures above 260°C and pressures of 2,6 MPa (26 kbar) (see figure 3.10). He concluded that while boehmite was metastable, diaspore was the stable phase.
While Juan and 1.o's work could not resolve this disagreement, they did find that boehmite was never converted directly to corundum, implying that a stable phase having a slow nucleation rate could exist between the phase fields of the two forms, and that this phase could be diaspore.
Univers
ity of
Cap
e Tow
n
P(H20}MPa P(H20}bar
101.3 1000
diaspora
10.13
glbbslte boehmlte corundum
1.013
100 200 300 400 500
T (deg.C}
Fig.3.9: Equilibrium diagram for the system Al20 3-H20 (after Ervin and Osborn, 1951)
P(H20}MPa P(H20}kbar
100
10
40,52..-----------..--------. 40
3039 30
glbbslte
2026 20.
1(.13 10
100 200 300 400 500
T (deg.C)
Fig.3.10: Equilibrium diagram for the system AI20 3-H20 · (after Kennedy, 1959) .
33
Van Zyl (1987) reported quantitative conversion of gibbsite to boehmite at 100°c under the pressure of saturated water vapour. This shows the importance of water
P(H20)MPa P(H20)bar
I I
101.3 I r 1000
I diaspore
I I
10.13 I 100
I gibbsite I boel'1mlte corundum
1.013 I 10
I I j
100 200 300 400 600
T (deg.C)
Equilibrium diagram for the system Al20 3-H20 (after and Osborn, 1951)
P(H20)MPa P(H20)kbar 40,52.----------.--------.40
3039
gibbsite
2016
100 200 300
T (deg.C)
10:: Equilibrium diagram for (after Kennedy, 1959)
20
10
400 600
Van (1987) quantitative conversion of gibbsite to boehmite at l000e
33
under the pressure of saturated water vapour. This shows importance of water
Univers
ity of
Cap
e Tow
n
vapour pressure as a condition of formation along with temperature and pressure, although all these are linked, as highlighted by Voigt et al (1977).
34
The variations in the condition of formation of the oxyhydroxides, as reported by various researchers, could also be linked to grain size. Within larger grains of hydroxide material; high water vapour pressures may be generated at relatively low temperatures. Thi~ may create the correct conditions for the solid state reaction to oxyhydroxide on a micro-scale well before the bulk conditions are appropriate.
Less work has been published on the precise mechanisms of transformation of aluminium oxyhydroxides than on the mechanisms of formation of hydroxides. Van Zyl (1987) outlined a general mechanism for the hydrothermal decomposition of aluminium hydroxide, developed from the work of Feitknecht et al (1961) and Freund (1967). Initially on heating, protons diffuse between the layers and react with the bridging OH groups to form water. This removes the binding forces between the layers, facilitating structural and chemical changes within the layers. The inherent lattice structure, the interlayer spacing, and the presence of impurities all affect the mobility of the protons. The solid state reaction is further influenced by the pressure, atmosphere, heating rate and particle size of the hydroxide. The dehydroxylation and structural transformations of the aluminium hydroxides and oxyhydroxides have been summarised by Gitzen ( 1970) in Figure _3 .11.
Conditions favouring transformat~ons Conditions Path a Path b
Pressure Atmosphere I Heating rate
I
'"''''' :i" >l atm. Mai.st air >l 0 c/min >100 microns
~IBBSITE H CH.1
I ~--,. a I
i I 1 .... l _so_E_HM_IT_E ___ ___.H GAMMA
! a I 1----....__,b I BAYE:Rl TE ETA
!-------------. P SEUDOBOEHMITE
K;lPP.~
DELi~
THET.\
l atm. dry air ·
<l °C/min <10 microns
I ALPHA I
THETA I ALPHA I
I ALPHA I
DIASPORE ; ._I _____ A_L_PH_A ____ ___.
. . . ~ . 100 200 300 400 500 6~0 700 800 900 1000 1100 1200 °C
Note: Enclosed area indicates range of occurrence. Open area
indicates range of transi~ion.
Fig.3 .11: Composition sequence of aluminium hydroxide and oxyhydroxide polymorphs (after Gitzen, 1987)
vapour pressure as a condition of formation along with temperature and pressure, although all these are linked, as highlighted by Voigt et al (1977).
34
The variations in the condition of formation of the oxyhydroxides, as reported by various researchers, could also be linked to grain size. Within larger grains of hydroxide material; high water vapour pressures may be generated at relatively low temperatures. Thi~ may create the correct conditions for the solid state reaction to oxyhydroxide on a micro-scale wen before the bulk conditions are appropriate.
Less work has been published on the precise mechanisms of transformation of aluminium oxyhydroxides than on the mechanisms of formation of hydroxides. Van Zyl (1987) outlined a general mechanism for the hydrothermal decomposition of aluminium hydroxide, developed from the work of Feitknecht et al (1961) and Freund (1967). Initially on heating, protons diffuse between the layers and react with the bridging OH groups to form water. This removes the binding forces between the layers, facilitating structural and chemical changes within the layers. The inherent lattice structure, the interlayer spacing, and the presence of impurities all affect the mobility of the protons. The solid state reaction is further influenced by the pressure, atmosphere, heating rate and particle size of the hydroxide. The dehydroxylation and structural transformations of the aluminium hydroxides and oxyhydroxides have been summarised by Gitzen (1970) in Figure .3.11.
Conditions favouring transformat~ons Conditions Path a Path b
Pressure Atmosphere
I Heating rate
II BOEHMITE
I a I I I b
CHl
I
)1 atm. Hoist air >1 0 C/min
>100 mic4'ons
I
H GAMMA I
I K;lPP.~ I
I
DELi-l i
i THO-i
1 atm. dry air
<l°C/min
<10 microns
I ALPHA I
THETA I ALPHA I
IISAY(RI TE f ET A
i I
i I I ALPHA I . I
IlpSEUDOBOEHMITE I
DIASPORE I . I ALPHA
j •• ~ •
100 200 300 400 sao 6~0 700 800 900 10CG 1100 120G °C Nate: Enclosed area indicates ::ange of ocr.urrence. Open a;:-ea
indicates range of transi:ion.
Fig.3.11: Composition sequence of aluminium hydroxide and oxyhydroxide polymorphs (after Gitzen, 1987)
I I
Univers
ity of
Cap
e Tow
n
35
3.2.4. Discussion of the proposed mechanisms.
What emerges from this review of mechanisms proposed for the formation of the hydroxides of aluminium is that one is dealing with a complex system which is und~stood only in broad outline, and about which there is still much disagreement amottg researchers. This complexity arises from the occurrence of amorphous forms and gels as much as from the occurrence of polymorphs. While it is beyond the scope of this project to propose a mechanism of formation, certain observations can be made.
The basic mechanisms proposed by early researchers appear to be widely supported in the literature. They provide a satisfactory explanation for the origin of the polymorphs, although the exact chemical mechanisms by which they form are not elucidated, particularly the polymerization of the [Al(OH)4r1 anion. The formation of gels is not dealt with, but this is understandable, remembering that these were still regarded as immature forms of the crystalline precipitate.
The proposal that the polymorph type is determined at the polymerization stage by the chemical environment appears well founded and explains the observed reactions better than the consecutive mechanism proposed by Bye and Robinson (1964). However, the experimental work of Bye and Robinson contributes other useful information, particularly on the role of pseudoboehmite and the assertion that it does not rearrange to a crystalline product, but dissolves and recrystallises. The influence of solvents, salts and impurities can be understood in the light of this. Whether or not the salts dehydrate the linkages, block polymerization, promote aggregation, or prevent dissolution of pseudoboehmite, the net effect is to stabilise amorphous or gel forms relative to crystalline trihydroxide. To promote the latter, it is obviously necessary to remove all excess salts and impurities from the system as soon as possible.
The assertion of Turner and Ross ( 1970) that the first product contains Cr ions substituted for OH- ions appears well founded and explains why the precipitates are occasionally observed to dissolve in the mother liquor. However, the mechanism they propose is somewhat problematic for a number of reasons:
Firstly, the formation of the polynuclear ions is seen as merely an energetically favourable blind alley along which the reaction proceeds temporarily. This appears inconsistent with the degree of order achieved.
Secondly, the final trihydroxide structure is said to be formed by solid phase reaction and rearrangement. The energy required for rearrangement of this structure would seem intuitively to be more than that available under the given reaction conditions.
Thirdly, the mechanism does not acknowledge the occurrence of polymorphs or consider factors which may determine preferential formation of one above
35
3.2.4. Discussion of the proposed mechanisms.
What emerges from this review of mechanisms proposed for the formation of the hydroxides of aluminium is that one is dealing with a complex system which is und~rstood only in broad outline, and about which there is still much disagreement among researchers. This complexity arises from the occurrence of amorphous forms and gels as much as from the occurrence of polymorphs. While it is beyond the scope of this project to propose a mechanism of formation, certain observations can be made.
The basic mechanisms proposed by early researchers appear to be widely supported in the literature. They provide a satisfactory explanation for the origin of the polymorphs, although the exact chemical mechanisms by which they form are not elucidated, particularly the polymerization of the [Al(OH)4r1 anion. The formation of gels is not dealt with, but this is understandable, remembering that these were still regarded as immature forms of the crystalline precipitate.
The proposal that the polymorph type is determined at the polymerization stage by the chemical environment appears wen founded and explains the observed reactions better than the consecutive mechanism proposed by Bye and Robinson (1964). However. the experimental work of Bye and Robinson contributes other useful information, particularly on the role of pseudoboebmite and the assertion that it does not rearrange to a crystalline product, but dissolves and recrystallises. The influence of solvents, salts and impurities can be understood in the light of this. Whether or not the salts dehydrate the linkages, block polymerization, promote aggregation, or prevent dissolution of pseudoboebmite, the net effect is to stabilise amorphous or gel forms relative to crystalline tribydroxide. To promote the latter, it is obviously necessary to remove all excess salts and impurities from the system as soon as possible.
The assertion of Turner and Ross (1970) that the first product contains cr ions substituted for OH- ions appears well founded and explains why the precipitates are occasionally observed to dissolve in the mother liquor. However, the mechanism they propose is somewhat problematic for a number of reasons:
Firstly, the formation of the polynuclear ions is seen as merely an energetically favourable blind alley along which the reaction proceeds temporarily. This appears inconsistent with the degree of order achieved.
Secondly, the final tribydroxide structure is said to be formed by solid phase reaction and rearrangement. The energy required for rearrangement of this structure would seem intuitively to be more than that available under the given reaction conditions.
Thirdly, the mechanism does not acknowledge the occurrence of polymorphs or consider factors which may determine preferential formation of one above
Univers
ity of
Cap
e Tow
n
36
the other. More evidence than that provided by the authors is required to support this mechanism.
The finding of Turner and Ross that only one polynuclear cation exists and that the trihydroxide is not formed by polymerization does not find support in the literature and contradicts the findings of Stol et al ( 1967), who measured the degree of polymerisation directly. The concentrations and OH/ Al ratios used oy Turner and Ross are outside the ranges given by Stol et al wherein only monomers and small polymers exist. However, Stol et al do suggest that above OH/Al = 2,5 polymerization is not the dominant mechanism, but rather the attachment of nonbridging OH-, leading to aggregation.
Although the mechanism Stol et al propose does not discuss the formation of polymorphs, it does allow to a large extent for the conditions and mechanisms put forward in Hsu's various publications. Its principal weakness is that the formation and prevalence of gels are not discussed, though this is perhaps outside the scope of Stol et al's investigations.
The major problem in understanding this system remains the occurrence of gels and how they fit into the overall mechanism. Many researchers have attempted to elucidate these issues, yet many grey areas remain. McHardy and Thomson (1971) and Aldcroft et al (1969) each identified two different gel forms, one which crystallized rapidly to bayerite, and one which crystallized slowly to gibbsite. In the former paper, an uncharged gel is reported to crystallise quickly to bayerite, while in the latter a more strained lattice is given as reason for the same behaviour. However, it is not known whether these two sets are in fact the same phenomenon with differing interpretation, or merely the edge of a minefield of varying gel types. No systematic and comprehensive study of the aluminium hydroxide gels appears to be available, although attempts have been made to identify the phase limits.
An additional complication is that the term n gel" can be applied to anything from a disrupted water matrix, through to the well developed structure of pseudoboehmite. Furthermore, although gels are referred to as being amorphous, it would be incorrect to assume that they are uniformly amorphous throughout; within each there exist areas of greater and lesser long range order.
Another unresolved issue is whether the gel forms and redissolves before crystallising, or whether nucleation takes place within the areas of greater local order, or whether polymers aggregate directly into a crystalline structure, with gels being simply a frequent aberration owing to unfavourable conditions (or all of the above). The experimental evidence presented by all the various authors indicates a two stage process.
36
the other. More evidence than that provided by the authors is required to support this mechanism.
The finding of Turner and Ross that only one polynuclear cation exists and that the trihydroxide is not formed by polymerization does not find support in the literature and contradicts the findings of Stol et al (1967), who measured the degree of polymerisation directly. The concentrations and OHI Al ratios used by Turner and Ross are outside the ranges given by Stol et al wherein only monomers and small polymers exist. However, Stol et al do suggest that above OH/AI = 2,5 polymerization is not the dominant mechanism, but rather the attachment of nonbridging OH-, leading to aggregation.
Although the mechanism Stol et al propose does not discuss the formation of polymorphs, it does allow to a large extent for the conditions and mechanisms put forward in Hsu's various publications. Its principal weakness is that the formation and prevalence of gels are not discussed, though this is perhaps outside the scope of Stol et al's investigations.
The major problem in understanding this system remains the occurrence of gels and how they fit into the overall mechanism. Many researchers have attempted to elucidate these issues, yet many grey areas remain. McHardy and Thomson (1971) and Aldcroft et al (1969) each identified two different gel forms, one which crystallized rapidly to bayerite, and one which crystallized slowly to gibbsite. In the former paper, an uncharged gel is reported to crystallise quickly to bayerite, while in the latter a more strained lattice is given as reason for the same behaviour. However, it is not known whether these two sets are in fact the same phenomenon with differing interpretation, or merely the edge of a minefield of varying gel types. No systematic and comprehensive study of the aluminium hydroxide gels appears to be available, although attempts have been made to identify the phase limits.
An additional complication is that the term It gel" can be applied to anything from a disrupted water matrix, through to the well developed structure of pseudoboehmite. Furthermore, although gels are referred to as being amorphous, it would be incorrect to assume that they are uniformly amorphous throughout; within each there exist areas of greater and lesser long range order.
Another unresolved issue is whether the gel forms and redissolves before crystallising, or whether nucleation takes place within the areas of greater local order, or whether polymers aggregate directly into a crystalline structure, with gels being simply a frequent aberration owing to unfavourable conditions (or all of the above). The experimental evidence presented by all the various authors indicates a two stage process.
Univers
ity of
Cap
e Tow
n
37
3.2.5 Summary.
If all the proposed mechanisms of the discussed authors are integrated, the following sequence emerges:
On addition of base to A1Cl3 the most rapid reaction i~
Simultaneously, the first bridging reactions occur, giving rise to a variety of • monomers, dimers and polymers. As the first product is metastable with
respect to the polynuclear ions, equilibrium in the above reaction shifts to the left and the initial solid dissolves:
(AI(H20)6]3+ + OH- <-> [Al(OH)(H20)5]2+ + H 20
2 [Al(OH)(H20)5]2+ <- > [Al2(0H)2(H20)s]4+ + H20
or:
The environment determines which ions are stable and the relative equilibria.
Under certain conditions of ion concentration, impurities, ligand or solvent interference, gelation may possibly occur at this stage. This gel would be positively charged. If, on the other hand, conditions are favourable, nonbridging OH- can attach to the edges of the polymers, reducing the charge and allowing aggregation and rapid crystallisation to bayerite:
------------- > aggregation ------------- > Polymers Crystals
-- > gel -- > dissolution -- > nucleation -- >
It may be necessary for the fixed gel to redissolve and nucleate. If in this process the hydrated structure [Al(OH)(H20)s]2+ is maintained, the recrystallising solid may be gibbsite. If excess salt has stabilised the gel, removing it by washing may enable the trihydroxide to crystallise out. Another possibility is that the init:ial solid may form a gel. This would then need to redissolve before polymerising and aggregating to crystalline trihydroxide, in
which case washing would also be beneficial. Many gaps remain in this scheme, but it seems consistent with the experimental evidence. Obviously, specific mechanistic investigations would be necessary to confirm the sequence of formation, as proposed above.
37
3.2.5 Summary.
If aU the proposed mechanisms of the discussed authors are integrated, the foHowing
sequence emerges:
On addition of base to AICl3 the most rapid reaction i~
Simultaneously, the first bridging reactions occur, giving rise to a variety of • monomers, dimers and polymers. As the first product is metastable with
respect to the polynuclear ions, equilibrium in the above reaction shifts to the left and the initial solid dissolves:
[Al(H20)6]3t + OH- <-> [AI(OH)(H20)S]2t + H 20
2 [AI(OH)(H20)5]2t < -> [A12(OH)2(H20)s]4+ + H20
or:
The environment determines which ions are stable and the relative equilibria.
Under certain conditions of ion concentration, impurities, ligand or solvent interference, gelation may possibly occur at this stage. This gel would be positively charged. If, on the other hand, conditions are favourable, nonbridging OH- can attach to the edges of the polymers, reducing the charge and allowing aggregation and rapid crystallisation to bayerite:
------------- > aggregation ------------- > Polymers Crystals
-- > gel -- > dissolution -- > nucleation -- >
It may be necessary for the fixed gel to redissolve and nucleate. If in this process the hydrated structure [Al(OH)(H20)sl2t is maintained, the recrystallising solid may be gibbsite. If excess salt has stabilised the gel, removing it by washing may enable the trihydroxide to crystallise out. Another possibility is that the initial solid may form a gel. This would then need to redissolve before polymerising and aggregating to crystalline trihydroxide, in
which case washing would also be beneficial. Many gaps remain in this scheme, but it seems consistent with the experimental evidence. Obviously, specific mechanistic investigations would be necessary to confirm the sequence of formation, as proposed above.
Univers
ity of
Cap
e Tow
n
38
There appears to be consensus among researchers that true boehmite and diaspore are formed by solid state reaction under conditions of elevated temperature and pressure. This conclusion has been substantiated by attempts to synthesise these polymorphs.
Differences of opinion persist as to whether the two oxyhydroxide polymorphs are completely stable phases. However those authors who argue that boehmite is metastable acknowledge that its rate of transition is extremely slow, So that for practical purposes it may be regarded as stable. The variations in the optimal conditions specified by the various authors may be attributed to the inherent differences in lattice structure and grain size, and to differences in reaction parameters which #
always exist. The control of water vapour pressure as a critical reaction parameter has been highlighted more recently, and lack of consideration of this parameter may explain some of the differences in findings.
The general mechanism outlined for the formation of oxyhydroxide polymorphs, based of the diffusion of protons through the structural layers, appears straightforward and consistant with the experimental findings of most researchers. The structural transformations of the hydroxide and oxyhydroxide polymorphs have been well documented by Krischner (1971) and Van Zyl (1987). The interest in these transition aluminas, particularly for their catalytic properties, should generate increasingly detailed research and discussion into the conditions and mechanisms of their formation. This will greatly benefit the understanding of the general mechanisms of formation of these polymorphs.
38
There appears to be consensus among researchers that true boehmite and diaspore are formed by solid state reaction under conditions of elevated temperature and pressure. This conclusion has been substantiated by attempts to synthesise these polymorphs.
Differences of opinion persist as to whether the two oxyhydroxide polymorphs are completely stable phases. However those authors who argue that boehmite is metastable acknowledge that its rate of transition is extremely slow, so that for practical purposes it may be regarded as stable. The variations in the optimal conditions specified by the various authors may be attributed to the inherent differences in lattice structure and grain size, and to differences in reaction parameters which always exist. The control of water vapour pressure as a critical reaction parameter has been highlighted more recently, and lack of consideration of this parameter may explain some of the differences in findings.
The general mechanism outlined for the formation of oxyhydroxide polymorphs, based of the diffusion of protons through the structural layers, appears straightforward and consistant with the experimental findings of most researchers. The structural transformations of the hydroxide and oxyhydroxide polymorphs have been wen documented by Krischner (1971) and Van Zyl (1987). The interest in these transition aluminas, particularly for their catalytic properties, should generate increasingly detailed research and discussion into the conditions and mechanisms of their formation. This will greatly benefit the understanding of the general mechanisms of formation of these polymorphs.
Univers
ity of
Cap
e Tow
n
39
4. SYNTHESIS OF BOEHMITE AND BAYERITE.
The initial brief for this study was to synthesise pure 6 "-alumina from the appropriate polymorphs, and furthermore to synthesise those polymorphs from readily available starting reactants. Previous investigators, most notably Van Zyl (1987), had established that the oxygen sublattice of the hydroxide or oxyhydroxide starting material was critical for the successful and optimal synthesis of pure 1 B 11 -alumina. Based on the analysis of the polymorph structures presented in Chapter 2, the two polymorphs whose oxygen sublattice render them most likely to yield pure B 11 -alumina on reaction are boehmite and bayerite, both of whom have the required ABC-ABC stacking sequence in the oxygen sub-lattice.
4.1 THE SYNTHESIS OF BOEHMITE.
In accordance with the brief, synthesis methods reported in the literature which claimed to form boehmite by simple reactions were examined and tested. In particular, the influence of the reaction parameters on the efficiency of the process was investigated.
As a starting point, the precipitation of boehmite from an AlC13.6H20 solution with NH40H was selected. The equation for this reaction is:
An attractive feature of this method is that while the insoluble Al(OH)3 precipitates, the soluble NH4Cl can be separated by washing, or under certain conditions of temperature and pressure may dissociate to NH3 (gas) and Cl2 (gas) and escape thus.
4.1.1 Effect of reaction temperature.
Moeller ( 1952) reported that the reaction temperature determined which hydroxide polymorph was precipitated. Accordingly, known amounts of AIC13.6H20 were dissolved in known volumes of distilled water. The solutions were stirred vigorously and continuously with magnetic stirrers, while the reaction temperature of one solution was controlled at boiling point by means of a hot plate. To these vigorously stirred solutions stoichiometric quantities of NH40H were added dropwise from a burette.
In both cases, a thick, white.precipitate was immediately formed, which initially took the form of discrete globules in the solution. As more NH40H was added, the mixture became thick and pasty and the stirring sluggish. Stirring was continued for ten minutes after all the alkali had been added. The precipitate was very fine and did not settle to the bottom completely, but rather formed a diffuse, almost colloidal, cloud with ill-defined boundaries. This made separation very difficult. The final pH of the · solutions was tested by indicator paper and found to be in the range pH 9 to 10, as reported in the literature.
Once precipitation was complete, a further volume of distilled water was added to both the reaction vessels and the solutions vigorously stirred. The precipitates were allowed
39
4. SYNTHESIS OF BOEHMITE AND BAYERITE.
The initial brief for this study was to synthesise pure .B" -alumina from the appropriate polymorphs, and furthermore to synthesise those polymorphs from readily available starting reactants. Previous investigators, most notably Van Zyl (1987), had established that the oxygen sub lattice of the hydroxide or oxyhydroxide starting material was critical for the successful and optimal synthesis of pure i B" -alumina. Based on the analysis of the polymorph structures presented in Chapter 2, the two polymorphs whose oxygen sublattice render them most likely to yield pure B" -alumina on reaction are boehmite and bayerite, both of whom have the required ABC-ABC stacking sequence in the oxygen sub-lattice.
4.1 THE SYNTHESIS OF BOEHMITE.
In accordance with the brief, synthesis methods reported in the literature which claimed to form boehmite by simple reactions were examined and tested. In particular, the influence of the reaction parameters on the efficiency of the process was investigated.
As a starting point, the precipitation of boehmite from an AIC13.6H20 solution with NH40H was selected. The equation for this reaction is:
An attractive feature of this method is that while the insoluble Al(OH)3 precipitates, the soluble NH4CI can be separated by washing, or under certain conditions of temperature and pressure may dissociate to NH3 (gas) and Cl2 (gas) and escape thus.
4.1.1 Effect of reaction temperature.
Moeller (1952) reported that the reaction temperature determined which hydroxide polymorph was precipitated. Accordingly, known amounts of AlCI3.6H20 were dissolved in known volumes of distilled water. The solutions were stirred vigorously and continuously with magnetic stirrers, while the reaction temperature of one solution was controlled at boiling point by means of a hot plate. To these vigorously stirred solutions stoichiometric quantities of NH40H were added dropwise from a burette.
In both cases, a thick, white precipitate was immediately formed, which initially took the form of discrete globules in the solution. As more NH40H was added, the mixture became thick and pasty and the stirring sluggish. Stirring was continued for ten minutes after all the alkali had been added. The precipitate was very fme and did not settle to the bottom completely, but rather formed a diffuse, almost conoidal, cloud with ill-defined boundaries. This made separation very difficult. The final pH of the . solutions was tested by indicator paper and found to be in the range pH 9 to 10, as reported in the literature.
Once precipitation was complete, a further volume of distilled water was added to both the reaction vessels and the solutions vigorously stirred. The precipitates were allowed
Univers
ity of
Cap
e Tow
n
40
to settle and the clear liquid decanted. Owing to the diffuse boundaries, this proved very difficult. A second volume of distilled water was added, the solutions stirred, and the precipitates separated by filtration on a Buchner funnel. Because the precipitate was so fine and had so much water associated with it, it formed a translucent white gel at the point of drainage, which then considerably slowed the filtration process. The filtrates were then washed three times on the filter, after which the odour of ammonia could no longer be detected in the precipitate.
The filter cakes were dried overnight in a laboratory drying oven at 9o0c. The appearance of the filtrates changed, from that of a translucent gel to large, solid white' chips. The change is associated with a great reduction in volume, indicating the loss of associated water held within the precipitate. These large pieces were then ground down to a powder by hand in~ pestle and mortar for X-ray diffraction examination.
The X-ray diffraction spectra were compared with those obtained for standard materials. Both samples showed the flat, broad peaks typical of an amorphous material, as shown in Figure 4.1. The positions of the peaks corresponded to the known reflections of pseudoboehmite. These results indicated that, contrary to Moeller, reaction temperature did not appear to influence the form of aluminium hydroxide which would precipitate. Furthermore, it suggests that where early researchers refer to "boehmite", they are actually referring to pseudoboehmite.
IO!l.
100 deg C
25deg c
O+-~-.-~--.-~---.-~--.~~.--~...-~-.-~--.-~-.-~--t 10 15
Fig.4.1:
20 25 30 35 40 45 DEGREES 2 THETA
50 55
XRD spectrum of pseudoboehmite obtained by reaction at 2s0c and 100°c
60
40
to settle and the clear liquid decanted. Owing to the diffuse boundaries, this proved very difficult. A second volume of distilled water was added, the solutions stirred, and the precipitates separated by filtration on a Buchner funnel. Because the precipitate was so fme and had so much water associated with it, it formed a translucent white gel at the point of drainage, which then considerably slowed the filtration process. The filtrates were then washed three times on the filter, after- which the odour of ammonia could no longer be detected in the precipitate.
The filter cakes were dried overnight in a laboratory drying oven at 90°C. The appearance of the filtrates changed, from that of a translucent gel to large, solid whitechips. The change is associated with a great reduction in volume, indicating the loss of associated water held within the precipitate. These large pieces were then ground down to a powder by hand in ~ pestle and mortar for X-ray diffraction examination.
The X-ray diffraction spectra were compared with those obtained for standard materials. Both samples showed the flat, broad peaks typical of an amorphous material,as shown in Figure 4.1. The positions of the peaks corresponded to the known reflections of pseudoboehmite. These results indicated that, contrary to Moeller, reaction temperature did not appear to influence the form of aluminium hydroxide which would precipitate. Furthermore, it suggests that where early researchers refer to "boehmite", they are actually referring to pseudoooehmite.
II!:!. ..,..-------:------------------...,
100 deg C
25deg C
O+---~-~--~---r_--,_--_r--~----r_--,_--_i 10 15
Fig.4.1:
20 25 30 35 40 45 DEGREES 2 THETA
50 55
XRD spectrum of pseudoboehmite obtained by reaction at 25°C and lOOOC
60
Univers
ity of
Cap
e Tow
n
41
4.1.2 Effect of ageing.
Following the teported findings of Mackenzie (1957), it was decided to investigate the effect of ageing the precipitate in the mother liquor. It was also felt that an aged precipitate may separate more easily, the average particJe size being greater. Consequently, NH40H was added to boiling solutions of A1Cl3.6H20 to pH 10, as described above. The resultant precipitates were aged in the mother ljquor under reflux, under the conditions listed in Table 4.1, whereafter the precipitates were washed, filtered, dried and prepared for X-ray diffraction analysis as described above.
Table 4.1: Ageing conditions of precipitates
TEMP (°C) TIME (hr)
20 2 24 50 70 2 24 50
The resulting X-ray diffraction spectra for the 20°c and 70°C batches are shown in Figure 4.2. All the samples aged at 20°C showed the flat, broad peaks in the regions 10-18° 20, 27°-29° 20 and 37°-400 20 which are associated with pseudoboehmite.
Fig.4.2:
IOa -.....---------------....,
~ (/) z UJ
~
0
24 hr@ 700C
---------------~--~~ 2hr@£0CC
liO 55 50 ~5 <() 35 30 25 20 15 10 5 0 DEGREES 2THETA
XRD spectra of batches aged at 200C and 70°C
41
1 Effect of age:tng.
n .. ".ul .... ." the teported of it was decided to the of ageing the in the mother liquor. It was also felt that an
precipitate may separate more easily, the average size being O'f'S'''!:It", ..
Consequently, NH40H was added to boiling solutions of AICl).6H20 to 10, as des.cribed above. The resultant precipitates were the mother ljquor
under the conditions listed in Table 4.1, the precipitates were washed, filtered, dried and prepared for X-ray diffraction analysis as above.
1: Ageing CODidlbons of prec;ipit:ates
TEMP (OC) TIME (hr)
20 2 24 50 70 2 24 50
resulting X-ray amraCll10n spectra for the 20°C 70°C batches are "''''1'1 ..........
4.2. AU the at 20°C showed broad peaks in the 8° 29, 270-2f)O and 37°-4()O 28 which are Q3.,'''\"UILI.VU with pseudoboehmite.
lOa .,....---------------....,
o
21'1r.g> 7Cee
flO 55 so ~5' 4() 35 30 25 20 15 10 5 0 DEGREES 2 THETA
XRD spectra batches aged at 200C and
Univers
ity of
Cap
e Tow
n
42
There was no discemable increase in crystallinity with increased ageing time. The samples aged at 70°C showed slightly stronger reflections, however the peaks were still broad and flat, once again indicating a pseudoboehmite structure. There was a slight, but insignificant, increase in crystallinity between the unaged sample and those aged for long periods, that is 24 - 50 hours.
Ageing under these conditions does not appear to enhance the transfdrmation of amorphous pseudoboehmite to either bayerite or boehmite, as was reported by Mackenzie. It is possible that some small differences between the conditions or geometry of these investigations and those of Mackenzie are the source of the discrepancy, however, every effort was made to follow the same procedure as that given in the literature.
4 .1. 3 Discussion.
It would appear that those researchers who reported the formation of boehmite by the simple reaction given above were, in fact, forming pseudoboehmite. Neither reaction temperatures in the range suggested by Moeller, nor solution ageing as reported by Mackenzie were found to effect the synthesis of true boehmite. The failure to reproduce the findings of these researchers might be owing to the difficulty of repeating the exact conditions employed by them. However, the overwhelming evidence presented by later researchers, as discussed in Chapter 3, is that boehmite is a hydrothermally formed phase which can only be synthesised under conditions of simultaneous high temperature and high pressure. The above results support this conclusion.
It is more likely that these authors, writing before Papee et al (1958) established that pseudoboehmite was not simply an immature form of boehmite but a different structure, were using the two terms interchangeably. This is one example of how confusion over nomenclature has complicated the study of these compounds. To produce true boehmite following this method of synthesis, it would be necessary to subject the pseudoboehmite described above to the required conditions of temperature, pressure and water-vapour pressure as discussed in Chapter 3.
The motivation for investigating methods of direct synthesis of the appropriate precursors was to avoid the double-peak firing schedule employed in the synthesis of B "-alumina in order to avoid the costs incurred by the extra energy input. If boehmite is to be the chosen starting material, the anticipated cost savings will not be realised owing to the reintroduction of a costly hydrothermal treatment step during the synthesis of the precursor.
42
was no discemable in crystallinity with increased ageing samples aged at 70°C showed slightly stronger however the were still broad and flat, once indicating a pseudoboehmite structure. was a slight,
insignificant, crystallinity between unaged sample and those long periods, that is 50 hours.
Ageing under these conditions does not appear to enhance the transformation of amorphous pseudoboehmite to either bayerite or as was reported by Mackenzie. It is possible that some small differences between the conditions or 2ec.me:trv of these and those are the source
"'''''''.'''''' ...... ''''' however, same procedure as in the literature.
1 Discussion.
It would appear that those who reported the formation of boehmite by the simple reaction given in fact, forming pseudoboehmite. Neither temperatures in the range by Moeller, nor ageing as reported MalCkt~nz:le were found to the synthesis of boehmite. The failure
the findings of researchers might be to the difficulty repeating the exact conditions employed by them. However, the overwhelming evidence presented by researchers, as discussed in Chapter 3, is that boehmite a hydrothermally formed which can only be synthesised under conditions of simultaneous high temperature and high pressure. above results support conclusion.
It is more likely that these writing before Papee al (1958) established that pseudoboehmite was not an immature form of but a different
were using the two interchangeably. This is one example of how confusion over nomenclature complicated the study of compounds. To ftrl'1.r1U/'P. true boehmite following this method of synthesis, it would be necessary to
the pseudoboehmite above to the conditions of temperature, nreSSWl'e and water-vapour as discussed
motivation for of direct synthesis the appropriate nrp..1"11"1.n1"1:1 was to avoid double-peak firing schedule employed in the synthesis of B It -alumina in order to avoid incurred by the extra input. If boehmite
to be the chosen starting material, the anticipated cost win not be realised I'\u' .... n to the reintroduction of a costly hydrothermal step during the syl1lUlf:81S of the precursor.
Univers
ity of
Cap
e Tow
n
43
4.2. THE SYNTHESIS OF BA YERITE.
As bayerite has the same favorable oxygen sub-lattice for optimal synthesis of fi " -alumina, that is: ABC-ABC stacking, emphasis shifted to the investigation of the synthesis of this hydroxide polymorph. A study of the literature revealed two main guidelines for the formation of bayerite:
a) the system should be as free as possible from interfering anions, ligands or solvents which might promote the amorphous state, and
b) the formation of bayerite is favoured if the reaction takes place in an • alkaline environment.
The same basic experimental method as that used above was employed. Known quantities of AIC13.6H20 were dissolved in known volumes of distilled water as before. This solution was transferred to a burette. Stoichiometric quantities of NH40H were placed in a glass beaker on a magnetic stirrer. Thus the aluminium chloride solution was added dropwise to the stirred alkali, that.is, the geometry of the reaction was the inverse of that used before. This was done to satisfy condition (b) above. The rate of addition was controlled at approximately 5ml per minute. The pH of the solution was monitored throughout, with the final pH being four. Immediately on contact of the two solutions, a thick, white precipitate was formed, initially taking the form of discrete globules. As more aluminium chloride was added, the mixture became progressively thicker and the stirring sluggish. Stirring was continued for ten minutes after all the aluminium chloride solution had been added. Once stirring was stopped, no settling of the precipitate was observed.
4.2.1 Effect of final pH.
The precipitated mixture was split into two equal portions. One portion was filtered immediately on a Buchner filter . A very small amount of precipitate was recovered (Sample A). The liquid which passed straight through the filter, being cloudy, was allowed to stand overnight in the hope that more solid might be recovered. After twelve hours the liquid was completely clear, the unrecovered precipitate apparently having dissolved in the highly acidic medium. It was then decided to adjust the pH of this solution to the range previously successful (pH 8 to 10) with excess NH40H, which resulted in the formation of a more manageable paste, which was white and not translucent. This paste was washed on a Buchner filter, oven dried at 100°c, crushed and prepared for XRD analysis (Sample B).
4.2.2 Effect of washing.
The other portion of the original reaction mixture was also corrected to pH 8 with excess NH40H. This was again split into two equal parts to examine the effect of washing, that is the removal of NH4 + and er, on crystallinity. The one part (Sample C) was filtered unwashed, oven dried at 100°c, crushed and prepared for XRD analysis. The other part (Sample D) was treated as follows:
43
4.2. SYNTHESIS OF
As has the same favorable oxygen sub-lattice for optimal synthesis of 6 If_
alumina, that ABC-ABC stacking, emphasis shifted to the investigation of the synthesis of this hydroxide polymorph. A study of the literature two main guidelines the formation of h~'l1"""t.,,,,·
a) .,,,,,,rpm should as as possible from anions, ligands
or solvents which promote the amorphous and b) formation of bayerite is favoured if the rea.cU()D takes place in an •
alkaline environment.
The same basic experimental method as that above was employed. Known quantities AlC13.6H20 were dissolved volumes water as before. solution was transferred a Stoichiometric quantities of NH40H were in a glass on a magnetic stirrer. Thus aluminium chloride was added dropwise to stirred alkali, that' geometry of the reaction was inverse of that used This was done to satisfy condition (b) above. The rate of addition was controlled at approximately 5ml minute. The pH of the solution was monitored throughout, with the final pH Immediately on contact of two solutions, a thick, precipitate was initially taking the form of globules. As more aluminium chloride was mixture became thicker and sluggish. Stirring was continued for ten minutes aluminium chloride solution had been added. Once stirring was stopped, no of the precipitate was observed.
4.2.1 pH.
The precipitated was split into two equal portions. One portion was immediately on a Buchner filter . A very small of precipitate was 'f"s:>r",""'''','f''''''''' (Sample A). liquid which passed the filter, being cloudy, was
twelve having dissolved this solution to which resulted in translucent. and prepared for
4.2.2 Effect
JVCIW;gIU in the hope more solid might be recovered. was completely clear, unrecovered precipitate apparently highly acidic medium. It was then decided to adjust of
range previously successful (PH 8 to 10) with excess formation of a more paste, which was
was washed on a Buchner oven dried at analysis (Sample B).
The other portion of original reaction mixture was also corrected to pH 8 with excess NH40H. This was again split into two parts to examine the of washing, that is the removal of NH4 + and cr, on crystallinity. The one C) was filtered unwashed, oven dried at 1000C, crushed and prepared analysis. The other (Sample D) was treated as follows:
Univers
ity of
Cap
e Tow
n
44
In order to wash as much of the NH4Cl as possible from the paste, it was necessary to use a centrifuge to separate the liquid from the gel-like precipitate. Each centrifuge tube was half filled with gel, topped up with distilled water, and then stirred until the contents were well mixed. The homogenised mixture was then spun for thirty minutes at 3000 rpm, after which the gel and washing liquid were well separated. This process was repeated five times, after which the odour of ammonia could no longer be detected. The washed gel was dried first on a Buchner funnel and finally in an oven at 1 lOOC, whereafter the hard, white filtercake, much decreased in volume, was crushed
for XRD analysis.
4.2.3 Effect of ageing.
The entire synthesis procedure was repeated in order to vet?fy the previous findings and examine the effect of ageing on crystallinity. The above synthesis method was repeated exactly, and the final pH adjusted to pH 8 as before. The reaction mixture was again devided into two parts. The one part was washed as described for sample D above, filtered, oven dried at 100°c, and prepared for XRD analysis (Sample E). The second part was aged for three days in the pH adjusted mother liquor, thereafter it was washed, filtered and dried (Sample F).
4.3 RESULTS.
4.3.1 Changes in pH.
The pH of the reaction mixtures were monitored throughout and were found to undergo a sudden transition in the region pH 9 to 4. This transition corresponded to a marked thickening of the mixture, whereafter the mixture returned to its previous viscosity. The final pH before adjustment was in the region pH 3 to 5. Where excess NH40H was added the final pH was in the region of pH 8 to 10.
I
4.3.2 Effect of final pH.
All samples were examined by XRD as described fully in 5.1.5. The resulting spectra are presented in Figure 4.3 below.
Sample A was found to consist entirely of pseudoboehmite, showing the broad reflections typical of this poorly crystalline form. Sample B, for which the fmal pH of the reaction mixture had been adjusted back into the alkaline range, also showed the broad reflections in its XRD spectrum, but in addition revealed the existence of sharp, well-defmed peaks, which confirmed that crystalline bayerite had also been formed . .
44
In order to wash as much of the NH4CI as possible from the paste, it was necessary to use a centrifuge to separate the liquid from the gel-like precipitate. Each centrifuge tube was half filled with gel, topped up with distilled water, and then stirred until the contents were well mixed. The homogenised mixture was then spun for thirty minutes at 3000 rpm, after which the gel and washing liquid were well separated. This process was repeated five times, after which the odour of ammonia could no longer be detected. The washed gel was dried first on a Buchner funnel and finally in an oven at ll00C, whereafter the hard, white filtercake, much decreased in volume, was crushed
for XRD analysis.
4.2.3 Effect of ageing.
The entire synthesis procedure was repeated in order to ve~fy the previous fmdings and examine the effect of ageing on crystallinity. The above synthesis method was repeated exactly, and the final pH adjusted to pH 8 as before. The reaction mixture was again devided into two parts. The one part was washed as described for sample D above, filtered, oven dried at lOOOC, and prepared for XRD analysis (Sample E). The second part was aged for three days in the pH adjusted mother liquor, thereafter it was washed, filtered and dried (Sample F).
4.3 RESULTS.
4.3.1 Changes in pH.
The pH of the reaction mixtures were monitored throughout and were found to undergo a sudden transition in the region pH 9 to 4. This transition corresponded to a marked thickening of the mixture, whereafter the mixture returned to its previous viscosity. The fmal pH before adjustment was in the region pH 3 to 5. Where excess NH40H was added the fmal pH was in the region of pH 8 to 10.
I
4.3.2 Effect of final pH.
An samples were examined by XRD as described funy in 5.1.5. The resulting spectra are presented in Figure 4.3 below.
Sample A was found to consist entirely of pseudoboehmite, showing the broad reflections typical of this poorly crystalline form. Sample B, for which the fmal pH of the reaction mixture had been adjusted back into the alkaline range, also showed the broad reflections in its XRD spectrum, but in addition revealed the existence of sharp, wen-defmed peaks, which confmned that crystalline bayerite had also been fon:ned.
Univers
ity of
Cap
e Tow
n
Fig.4 3:
Fig.44:
~ en z w ~ 0
.:iso
900
~ Ci5 z w 1-z
60
60
55 50
XRD spectra:
45 35 30 25 DEGREES 2 THETA
Sample A - Reaction precipitate.
20
Sample B - Reaction precipitate, pH adjusted
55 50 45 35 30 25 DEGREES 2 THETA
XRD spectrum of Sample C: Reaction precipitate; pH adjusted.
20
45.
A
15 10
15 10
Fig.4 3:
Fig.4 4:
~ en z w !z 0
.. SO
55
XRD spectra:
45 35 30 25 DEGREES 2 THETA
Sample A - Reaction precipitate.
20
Sample B - Reaction precipitate, pH adjusted
A
15 10
900.-------------------------------------------.
~ en z w IZ
55 45 35 30 25 DEGREES 2 THETA
XRD spectrum of Sample C: Reaction precipitate; pH adjusted.
20 15 10
45
Univers
ity of
Cap
e Tow
n
46
4.3.3 Effect of washing.
Analysis of Sample C by XRD once again revealed the broad reflections typical of pseudoboehmite as shown in Figure 4.4. In contrast to Sample B, no sharp bayerite peaks were observed, although the sharper peaks indicate that the removal of salts by washing has assisted the formation of more ordered pseudoboehmite.,
Sample D, however, showed mainly the strong, sharp peaks typical of well-crystallised bayerite, but a slightly raised baseline in places indicates the possible presence of a very small amount of pseudoboehmite (Figure 4.5).
4.3.4 Effect of ageing.
XRD analysis of both Sample E and Sample F revealed the well-defined spectrum of bayerite, with very little difference between the two materials, as shown in figures 4.6 and 4. 7 respectively. However, in both spectra evidence can be seen of the superposition of the typical pseudoboehmite spectrum shown in figure 4.5. This is particularly noticeable in the raised backgrounds in the regions 46°-50° 20, 35°-42° 20 and 26°-32° 29. This indicates the formation or retention of more pseudoboehmite than was observed in Sample D.
~ Ci5 z UJ 1-z
Fig.4.5:
60 55 50 45 40 35 JO 25 DEGREES 2 THETA
XRD spectrum of Sample D: Reaction precipitate; pH adjusted; washed.
D
20 15 10
46
4.3.3 Effect of washing.
Analysis of Sample C by XRD once again revealed the broad reflections typical of pseudoboehmite as shown in Figure 4.4. In contrast to Sample B, no sharp bayerite
peaks were observed, although the sharper peaks indicate that the removal of salts by washing has assisted the formation of more ordered pseudoboehmite.,.
Sample D, however, showed mainly the strong, sharp peaks typical of well-crystallised
bayerite, but a slightly raised baseline in places indicates the possible presence of a very small amount of pseudoboehmite (Figure 4.5).
4.3.4 Effect of ageing.
XRD analysis of both Sample E and Sample F revealed the well-defined spectrum of bayerite, with very little difference between the two materials, as shown in figures 4.6 and 4.7 respectively. However, in both spectra evidence can be seen of the superposition of the typical pseudoboehmite spectrum shown in figure 4.5. This is particularly noticeable in the raised backgrounds in the regions 46°-50° 28, 35°-42° 28 and 26°-32° 29. This indicates the formation or retention of more pseudoboehmite than was observed in Sample D.
900
~ en z LU IZ
Fig.4.5:
60 50 45 35 30 25 DEGREES 2 THETA
XRD spectrum of Sample D: Reaction precipitate; pH adjusted; washed.
D
20 15 10
Univers
ity of
Cap
e Tow
n
900
~ (j) z w 1-z
Fig.4 6:
900
~ Ci5 z w 1-z
60
60
Fig.4.7:
55 50 45 40 35 JO 25 DEGREES 2 THETA
XRD spectrum of Sample E:
20
Reaction precipitate, pH adjusted, washed.
55 50 45 40 35 JO 25 DEGREES 2 THETA
XRD spectrum of Sample F: Aged 3 days, then as for Sample E.
20
47
15 10
F
15 10
900
~ (i5 z W IZ
Fig.4 6:
900
~ Ci5 z W IZ
00
00
Fig.4.7:
55 50 45 40 35 JO 25 DEGREES 2 THETA
XRD spectrum of Sample E:
20
Reaction precipitate, pH adjusted, washed.
55 50 45 40 35 JO 25 DEGREES 2 THETA
XRD spectrum of Sample F: Aged 3 days, then as for Sample E.
20
47
E
15 10
F
15 10
Univers
ity of
Cap
e Tow
n
48
4.4 DISCUSSION AND CONCLUSIONS.
While neither the pH of the reaction mixture, the specific surface area of the precipitates, nor the turbidity of the solutions were studied in detail as by previously discussed researchers, the course and outcome of the reaction discussed were observed to be consistent in all respects with their observations.
I
The inversion of the reaction, that is the addition of the A1Cl3 to the NaOH solution, such that the reaction takes place in an alkaline environment, is successful in promoting the formation of bayerite. This is consistant with the literature reviewed. • The stoichiometrically required volumes of NH40H do not appear to be adequate to maintain the alkaline environment required for the formation of bayerite under the conditions of batch processing, although it is possible that the correct ratio might be achieved more easily in a continuous processing configuration. Adjusting the final pH to pH 8 to 10 is effective in promoting the formation of bayerite over pseudoboehmite. Samples B and C, having been formed under similar conditions, were quite different. This can be ascribed to small differences in the reaction environment, in particular the concentration of er ions.
Washing to remove NH4 + and er significantly promotes the formation of, or transformation to, bayerite from pseudoboehmite. This is consistant with the findings of Barnhisel and Rich (1965) and Hsu (1967) that the presence of er ions stabilises pseudoboehmite relative to bayerite.
Ageing of the mixture in the pH adjusted mother liquor before washing did not appear to enhance the formation of bayerite, nor significantly affect the amount of pseudoboehmite retained. This seems to confirm that pseudoboehmite will not transform to bayerite under these conditions of ageing. However, it may do so after the conditions have been altered by washing as discussed by Barnhisel and Rich and Hsu. The regions of stability of the various phases as defined by these researchers were not explored in this exercise.
The two main guidelines derived from the literature study for the formation of bayerite appear to have been substantiated by this experimental investigation. It is possible to form fully-crystalline bayerite by the simple reaction of Alel3.6H20 with NH40H at room temperature. The reaction must take place in an alkaline environment, which must be maintained, in a batch processing configuration, by the addition of excess NH40H to bring the final pH to pH 8 to 10. Ageing at this stage produces no significant advantage. Washing to remove as much as possible of the er ions greatly aids and/or facilitates the formation of well-crystallised bayerite.
o the system is complex and the governing factors interact in such a way so as to produce varym under seemingly identical conditions. As a consequence, results are not always perfectly r ible. This is problematic if one is attempting to generate large amounts of the same materi . rmore, until the governing factors are better understood, and a consistant and credible me of reaction is derived, attempts to scale up the synthesis to any significant extent are like y
48
4.4 DISCUSSION AND CONCLUSIONS.
While neither the pH the reaction mixture, the specific area the precipitates, nor the turbidity of the solutions were studied in detail as by previously
discussed researchers, the course and outcome of the reaction discussed were observed to be consistent in all respects with observations. ,
The inversion of the reaction, that is addition of AICl3 to the NaOH solution, such that the reaction place in an alkaline environment, is successful in promoting the formation of bayerite. This is consistant with the literature reviewed. • The stoichiometrically required volumes of NH40H do not appear to be adequate to maintain the environment required the formation bayerite under the conditions of batch processing, although it is possible that the correct ratio might be achieved more a continuous processing configuration. Adjusting the fmal pH to pH 8 to is effective in promoting the formation of bayerite over pseudoboehmite. Samples B and C, having formed similar conditions, were quite different. This can be ascribed to small differences in reaction environment, particular the concentration of cr ions.
Washing to remove NH4+ and cr significantly promotes formation of, or transformation to, bayerite from pseudoboehmite. is consistant with fmdings of Bamhisel and Rich (1965) and (1967) that presence of cr ions stabilises pseudoboehmite relative to bayerite.
Ageing of the in the pH adjusted mother liquor before washing did not appear to enhance the formation bayerite, nor significantly the amount of pseudoboehmite retained. This seems to confirm that pseudoboehmite will not transform to bayerite under these conditions of ageing. However, it do so after the conditions have been altered by washing as discussed by Barnhisel and Rich and
The regions of stability of the various phases as defined by these researchers were not explored this exercise.
The two main derived from the literature study for the formation of bayerite appear to have substantiated by this experimental investigation. It is possible to form fully-crystalline bayerite by the simple reaction of with NH40H at room temperature. The reaction must take place in an alkaline environment, which must be maintained, a batch processing configuration, by the addition of excess NH40H to bring the fmal pH to pH 8 to 10. Ageing this produces no significant advantage. Washing to remove as much as possible the cr ions ",..."" .... 1 ..
aids and/or facilitates formation of well-crystallised bayerite.
and the governing factors in such a way so as to produce seemingly identical conditions. As a consequence, results are not al i'epftK~ble. This is problematic if one is attempting to generate large amounts of the same until the governing factors are better understood, and a consistant . of reaction is derived, attempts to scale up the synthesis to significant eXllent are
Univers
ity of
Cap
e Tow
n
49
However, the system is complex and the governing factors interact in such a way so as
to produce varying results under seemingly identical conditions. As a consequence, results are not always perfectly reproducible. This is problematic if one is attempting to generate large amounts of the same material. Furthermore, until the governing factors are better understood, and a consistant and credible mechanism of reaction is
derived, attempts to scale up the synthesis to any significant extent are likely to be fraught with difficulties. Nevertheless, if reproducibility is achieved;' significant cost savings over a hydrothermal synthesis path may be possible.
49
However, the system is complex and the governing factors interact in such a way so as
to produce varying results under seemingly identical conditions. As a consequence, results are not always perfectly reproducible. This is problematic if one is attempting to generate large amounts of the same material. Furthermore, until the governing factors are better understood, and a consistant and credible mechanism of reaction is
derived, attempts to scale up the synthesis to any significant extent are likely to be fraught with difficulties. Nevertheless, if reproducibility is achieved,! significant cost savings over a hydrothermal synthesis path may be possible.
Univers
ity of
Cap
e Tow
n
50
5. SYNTHESIS OF 6"-ALUMINA.
The ultimate aim of these investigations remained the synthesis of B " - Alumina; therefore it was important to subject the precursors examined in chapters 4 and 5 to the solid state reaction forming beta alumina and to examine the results. The procedures used were in accordance with the standard method described by Van Zyl (1987).
5 .1 SYNTHESIS.
Two samples of commercially available, hydrothermally formed boehmite were obtained for examination and comparison: Cera Hydrate from America, and Timcerm from Holland. The average particle size of the two powders were 28,9 microns and 2,6 microns respectively. Owing to the small vo~ume of bayerite produced experimentally in this study, a sample of bayerite was obtained from Germany for comparison with boehmite as a precursor for beta alumina. The average particle size of this powder was 27 ,4 microns. XRD analysis confirmed the composition of these powder samples.
5 .1.1 Composition.
Sodium and lithium were added to the aluminium hydroxide so as to obtain a composition of 84,0 mole percent AI20 3, 13,5 mole percent Na20 and 2,5 mole percent Li20. The sodium and lithium were added in the form of Na2C03 and Li2C03 respectively. Because the latter is hydroscopic, it was dried at 320° C for 5 hours to ensure consistent composition of the batches.
5 .1.2 Powder preparation.
The powder batches were prepared as follows: The boehmite or bayerite powder was weighed into a teflon beaker, ·and the sodium and lithium carbonates added in the correct stoichiometric ratio. To this 100 ml of distilled water and 300 g of 2 mm Zr02 grinding beads were added. The beaker was then attached to an attritor mill and the slurry milled at 850 rpm for two hours. After milling the slurry was separated from the Zr02 beads by seiving. The resultant slurry was well dispersed· and no settling was observed after five hours.
Because the slurry contained a soluble component (Na2C03), a partly soluble component (Li2C03), and an insoluble component (aluminium hydroxide), it was necessary to freeze-dry the mixture in order to preserve it's homogeneity. The slurry was transferred to round-bottomed freeze drying flasks which were immersed in a refrigerant solution and rotated until a uniform, solid shell of frozen slurry coated the vessel.
The flasks were then transferred to the freeze dryer where they were evacuated until all the water had sublimated. This method produced homogeneous, free-flowing powders
50
5. SYNTHESIS I"-ALUMINA.
The ultimate these investigations the synthesis of B" - Alumina; therefore it was important to subject the precursors examined in 4 and 5 to the
state beta alumina to examine the procedures used were with the standard described by (1987).
1
Two samples of commercially available, hydrothermally formed boehmite were obtained and comparison: Hydrate from and Timcerm
average particle of the two powders were microns and 2,6 microns Owing to the volume of produced
this study, a bayerite was Germany for comparison boehmite as a precursor for beta alumina. particle of this powder was 27,4 microns. analysis confirmed the composition of these powder samples.
5.1.1 Composition.
Sodium and ..... ", ....... were added to aluminium hydroxide so as to obtain a composition mole percent mole percent and mole percent The sodium and lithium were added in the of and L4"~,,,,, .. ".l respectively. Because the latter is bydroscopic, it was dried at 3200 C for 5 bours to ensure composition of the batches.
5.1.2 Powder preparation.
nr"uru • .,. batches were prepared as follows: The or bayerite powder was weighed a teflon beaker, . and and added in the correct stoichiometric ratio. To 100 ml of distilled 300 g of 2 mm grinding beads were added. was then attached to an attritor mill and slurry milled 850 rpm for two hours. After milling the slurry was separated from the Zr02 beads by seiving. The resultant slurry was well dispersed' and no settling was observed five hours.
contained a ""'v ...... .., .... component (Na2C03)' a partly soluble component (Li2C03), and an component hydroxide), it was nec:ess:ary to freeze-dry the in order to preserve homogeneity. The was to round-bottomed drying flasks were immersed in a refrigerant solution and rotated a uniform, solid of frozen slurry coated the vesseL
were then transferred the freeze dryer were evacuated until all sublimated. .......... ~ .. vu produced bOlnOl~enleou,s, free-flowing powders
Univers
ity of
Cap
e Tow
n
51
which were then stored in sealed containers in an oven at 110° C to prevent the readsorption of water.
The yield for this entire procedure always exceeded 95 percent.
5.1.3 Solid state reaction.
The prepared powders were formed into standard discs 1 Omm in diameter and 3 to 5 mm thick in a uniaxial press at 300 kPa. All test discs was accompanied by two discs above and two below to limit sodium ion loss during firing. Each disc was numbered• so that any differences which might emerge could be correlated with its position. The five discs were stacked neatly inside a magnesia crucible which was sealed with a lid.
The crucible was then placed inside a furnace and heated at 20° C per minute to the required reaction temperatures ranging from 1100° C to 1600°C, where it was held for periods ranging from 1 to 5. hours and thereafter cooled at 20° C per minu,te to room temperature. After reaction, the cooled discs were removed from the crucible and stored in sealed containers in a desiccator.
5.1.4 Density determination.
The density of each of the fired discs was determined according to the Archimedes principle by immersion in xylene.
Having accllrately obtained the weight in air of the discs, they were immersed in boiling xylene for 10 minutes to eliminate air from surface pores. The discs were then weighed while suspended in xylene, and the bulk density calculated from
(Massair x Xylene density) Bulk Density =
(Massair - Mas8xylene>
5.1.5 X-ray determination.
Many researchers have characterised the beta-aluminas by X-ray diffraction and there is good agreement between their results. For example, Schmid (1986) reported a full and detailed characterisation of the diffraction patterns of .B " - and .B-alumina. These are illustrated in Figure 5.1 (a) and (b) respectively:
51
which were then stored in sealed containers in an oven at 110° C to prevent the readsorption of water.
The yield for this entire procedure always exceeded 95 percent.
5.1.3 Solid state reaction.
The prepared powders were formed into standard discs 10mm in diameter and 3 to 5 mm thick in a uniaxial press at 300 kPa. All test discs was accompanied by two discs above and two below to limit sodium ion loss during firing. Each disc was numberedso that any differences which might emerge could be correlated with its position. The five discs were stacked neatly inside a magnesia crucible which was sealed with a lid.
The crucible was then placed inside a furnace and heated at 20° C per minute to the required reaction temperatures ranging from 1100° C to 1600oC, where it was held for periods ranging from 1 to 5 hours and thereafter cooled at 20° C per minute to room temperature. After reaction, the cooled discs were removed from the crucible and stored in sealed containers in a desiccator.
5.1.4 Density determination.
The density of each of the fired discs was determined according to the Archimedes principle by immersion in xylene.
Having accurately obtained the weight in air of the discs, they were immersed in boiling xylene for 10 minutes to eliminate air from surface pores. The discs were then weighed while suspended in xylene, and the bulk density calculated from
(Massair x Xylene density) Bulk Density =
(Massair - MasSxy lene)
5.1.5 X-ray determination.
Many researchers have characterised the beta-aluminas by X-ray diffraction and there is good agreement between their results. For example, Schmid (1986) reported a full and detailed characterisation of the diffraction patterns of .B" - and .B-alumina. These are illustrated in Figure 5.1 (a) and (b) respectively:
Univers
ity of
Cap
e Tow
n
52
r I lo) (
b5 bO 55 50 45 40 35 30 25 20 15 10
I 14 ~ 19 r 31 26 I 1l
, I J ,. 1 j I
(b)
· . ~· 3' j/ I ' JS6 J zu ·1· ~ ! J'V~ j ~ .. ,.:'" ~AJ, LJ "\JV~~ULL __ JJ~
70 bS bO 55 50 45 4o 35 30 25
Fig 5.1 XRD pattern of (a) 6"-Alumina and (b) 6-Alumina (Schmid, 1986)
20 15 10 70
(b)
b5 0'0 s's 5'0 ,
45 40 , I ,
35 30 25
Fig 5.1 XRD pattern of (a) 6 ft-Alumina and (b) 6-Alumina (Schmid, 1986)
, , 20 15
I 10
52
Univers
ity of
Cap
e Tow
n
53
Various researchers report using different peaks to characterise the beta phases. Some of these are presented in Table 5 .1.
Table 5.1: Characteristic peaks identified by researchers. (notation as quoted)
RESEARCHER 6 6" i
hid d (A) hid d(A)
Schmid ( 1986) 026 2.03 0111 2.60 027 1.94 0210 1.97 •
017 2.68 014 4.20 012 4.46
Hodge (1983) 017 2.68 0111 2.60 026 2.03 0210 1.97
K vachkov et al 2026 2.03 022.10 1.97 (1982) 2027 1.94
Pekarsky and 017 2.69 0111 2.60 Nicholson (1980) 026 2.04 0210 1.97
Johnson et al 012 4.46 104 4.20 (1979)
May (1978) 013 4.08 1014 4.20 0226 2.03 02210 1.97 0227 1.94
The peaks in these ranges are well defined and provide a convenient contrast between the patterns of the two phases, which is easily identifiable. Furthermore there is good separation of the peaks, reducing the need for complex mathematical manipulation of results. Another important factor is the clear background region on either side of both sets of peaks, which would facilitate the determination of absolute intensities.
Johnson et al (1979) determined, by examining the relevant JCPDS powder diffraction file cards for 6- and 6"-alumina (19-1173 and 25-775 respectively), that the fraction of 6 "-Alumina present in a mixture of the two phases may be calculated from the expression
53
Various researchers report using different peaks to characterise the beta phases. Some of these are presented in Table 5.1.
Table 5.1: Characteristic peaks identified by researchers. (notation as quoted)
RESEARCHER 13 13" i
hid d (A.) hid d(A.)
Schmid (1986) 026 2.03 0111 2.60 027 1.94 0210 1.97 •
017 2.68 014 4.20 012 4.46
Hodge (1983) 017 2.68 0111 2.60 026 2.03 0210 1.97
K vachkov et al 2026 2.03 022.10 1.97 (1982) 2027 1.94
Pekarsky and 017 2.69 0111 2.60 Nicholson (1980) 026 2.04 0210 1.97
Johnson et al 012 4.46 104 4.20 (1979)
May (1978) 013 4.08 1014 4.20 0226 2.03 02210 1.97 0227 1.94
The peaks in these ranges are wen defined and provide a convenient contrast between the patterns of the two phases, which is easily identifiable. Furthermore there is good separation of the peaks, reducing the need for complex mathematical manipulation of results. Another important factor is the clear background region on either side of both sets of peaks, which would facilitate the determination of absolute intensities.
Johnson et al (1979) determined, by examining the relevant JCPDS powder diffraction file cards for 13- and 13 ft-alumina (19-1173 and 25-775 respectively), that the fraction of 13" -Alumina present in a mixture of the two phases may be calculated from the expression
Univers
ity of
Cap
e Tow
n
54
f(.6") = 1.5 .6"1104
1.56 "1104 + 61012
Pekarsky and Nicholson (1980) modified this relationship to work with the normalised intensity values for the characteristic peaks of their choice; Morgan (1982), in discussing a paper by K vachkov et al ( 1981), disputed this: method b~cause the intensity of X-ray lines can be greatly affected by syntactic intergrowth effects. The latter refuted this, saying that the peaks selected were well separated and could be accurately measured. They also pointed out that the powder diffraction method ensured the averaging of intensities so that it was unlikely that syntactic intergrowths would affect the intensity ratios.
Hodge (1983) observed that diffraction lines corresponding to planes which are similar in the 6- and 6"-alum1na appeared to be sharp, whereas those unique to one phase were more diffuse. This type of X-ray pattern was earlier described by Morgan (1982), who attributed the pattern to intermixed layers of the two phases along the c-axis, and also by Poulieff et al (1978), who attributed it to faulty stacking inherited from the starting structure.
X-ray diffraction investigations were performed on a Philips diffractometer using Cu Ka radiation and the following standard conditions:
45 kV; 40 mA; Range = lxHP; Time constant = 1 Scan rate = 20 20/min
All samples were milled in a mechanical ball mill for a standard period of ten minutes and were mounted in standard aluminium side-loading sample holders. Preliminary trials of different milling times and mounting techniques showed no preferred orientation effects or other significant variations. Milled samples were stored in an oven at 11 OOC to prevent the adsorption of water.
After exploratory scans to determine the best experimental conditions, diffraction patterns were obtained on standard .6 " - and .6-aluminas obtained from the CSIR. The patterns corresponded well with those reported in the literature. They are presented in Figure 5.2 (a) and (b) respectively.
f(8") =
Pekarsky and Nicholson (1980) modified this relationship to work with the normalised intensity for the characteristic peaks their choice; Morgan (1982), discussing a paper by K vachkov et al (1981), disputed this method the intensity of lines can be greatly affected syntactic intergrowth effects. The latter refuted this, saying that the peaks selected were well separated and could be accurately measured. also pointed out that powder diffraction method ensured the averaging of so that it was unlikely syntactic intergrowths would affect the intensity
Hodge (1983) observed that diffraction corresponding to planes which are similar in the 6- and 6"-alumlna appeared to be sharp, those unique to one phase were more diffuse. This type of X-ray pattern was earlier described by Morgan (1982), who attributed the pattern to intermixed layers of the two phases along the and also by Poulieff et al (1978), who attributed it to faulty stacking inherited from starting SU1lClUI'e.
diffraction investigations were performed on a Philips diffractometer using Cu Ka and the following standard conditions:
kV; 40 rnA; Range = lxUY; Time constant = 1 Scan rate = 2°
All samples were milled in a mechanical ball mill for a standard period of ten minutes and were mounted in standard aluminium side-loading sample holders. Preliminary
of different milling and mounting techniques showed no preferred orientation effects or other significant variations. Milled samples were stored in an oven at 11 ooc to prevent the adsorption of water.
exploratory scans to determine the best experimental conditions, diffraction patterns were obtained on standard 6"- and .6-aluminas obtained from the CSIR. The patterns corresponded wen with those reported in the literature. They are presented in Figure 5.2 (a) and (b) respectively.
Univers
ity of
Cap
e Tow
n
~ en z w 1-z
~ en z w 1-z
0-+-~~-.--~~--.-~~--.-~~--..~~--.,.--~~r-~~-.---~~-;-~~--r-~~---i
60 55 50 45 40 35 . 30 25 20 15 0
60 55
DEGREES 2 THETA
< >
50 45 40 35 30
DEGREES 2 THETA
Fig 5.2 (a) XRD pattern of standard 6-alumina Fig 5 .2 (b) XRD pattern of standard 6 "-alumina
25 20 15 10
55
~ en z w ..... z
~ en z w ..... z
IOJ r------------------------------------------------------------A~
(
O+-----~----~----_.----~----~------r_----~----+_----._----~ 60 55 50 45 40 35 . 30 25 20 15 0
DEGREES 2 THETA
I05r-------------------------------~--------------------------__4 B
< >
O~----~----_r----~----~----~~----~----~----~----~----~ 55 45 40 35 30
DEGREES 2 THETA
Fig 5.2 (a) XRD pattern of standard fi-alumina Fig 5.2 (b) XRD pattern of standard fi"-alumina
25 20 15 10
55
Univers
ity of
Cap
e Tow
n
' 56
From examination of the patterns and of the literature the following identification peaks were selected for the reasons given previously:
Phase
.6-alumina
.6 "-alumina
Position
44,36° 29 46,03° 28
Identification
026 0210
Initially all patterns were recorded graphically; later the patterns were collected digitally, which allowed the data to be manipulated using standard software packages.
The .6-alumina fraction was calculated from the normalised intensity values for the characteristic peaks according to the method of Pekarsky and Nicholson, that is
f(.6) = 2,29 IB(2,04 A)
2,29 ls(2,04 A) + Is"(l,97 A)
Using the standard conditions specified, the f(.6) results calculated after repeated scans of the same material agreed within 0,5% of the mean f(.6).
The same standard conditions were used in identifying and characterising the various aluminium hydroxide precursor materials.
5.2 RESULTS.
5 .2.1 Observations.
As might be expected, the discs shrank during firing, the amount of shrinkage being proportional to the firing temperature. The maximum decrease in diameter was approximately tWenty percent. The crucible in which the discs were fired developed a light pink discoloration on the inside, indicating that some sodium had escaped from the discs during firing.
5.2.2 Density.
As shown by Figures 5.3 and 5.4 both Timcerm and Cera Hydrate achieved a maximum density of approximately 3,15 g/ml at a firing temperature 13500C and 5 hours. Lowering the soak timeto 2,5 hours decreased the density achieved at 1200°c, but did not have a significant influence at 1600°C.
, 56
From examination of the patterns and of the literature the following identification peaks were selected for the reasons given previously:
Phase
.6-alumina
.6 "-alumina
Position
44,36° 29 46,03° 29
Identification
026 0210
Initially all patterns were recorded graphically; later the patterns were collected digitally, which allowed the data to be manipulated using standard software packages.
The .6-alumina fraction was calculated from the normalised intensity values for the characteristic peaks according to the method of Pekarsky and Nicholson, that is
f(.6) = 2,29 18(2,04 A)
2,2918(2,04 A) + 18"(1,97 A)
Using the standard conditions specified, the f(.6) results calculated after repeated scans of the same material agreed within 0,5% of the mean f(.6).
The same standard conditions were used in identifying and characterising the various aluminium hydroxide precursor materials.
5.2 RESULTS.
5.2.1 Observations.
As might be expected, the discs shrank during firing, the amount of shrinkage being proportional to the firing temperature. The maximum decrease in diameter was approximately twenty percent. The crucible in which the discs were fired developed a light pink discoloration on the inside, indicating that some sodium had escaped from the discs during firing.
5.2.2 Density.
As shown by Figures 5.3 and 5.4 both Timcerm and Cera Hydrate achieved a maximum density of approximately 3,15 g/ml at a firing temperature 13500C and 5 hours. Lowering the soak timeto 2,5 hours decreased the density achieved at 1200oC, but did not have a significant influence at 1600oC.
Univers
ity of
Cap
e Tow
n
57
5 HR SOAK
DENSITY (g/ml) 3.s....-~~~~~~~~~~~~~~~~~~~~~~----.
3.3 I-
3.1 ..._
2.9 '-
2.7 ..._
2.SL----~~-'--'~~-1'--~~~1'--~~~1~~~~'~~~~·~~~
1000 1100 1200 1300 1400 1500 1600 1700
FIRING TEMPERATURE (deg C)
Fig. 5.3: Density vs Firing Temperature - Timcerm
3.3 ....
3.1 ...
2.9 ....
2.7 -+ 2.5 HOURS • 5.0 HOURS
2.5'---~~~·'--~--+-~·~~~~·~~~~·~~~~'~~~~'~~_._.j I
1000 1100 ;1200 1300 1400 1500 1600 1700
FIRING TEMPERATURE (deg C)
Fig. 5.4: Density vs Firing Temperature - Cera Hydrate
57
5 HR SOAK
DENSITY (gjml) 3.5.-----------------------------------------------,
3.3 I-
3.1 r-
2.9 -
2.7 r-
2.5~----~------~-----1~----~----~------~1----~
1000 1100 1200 1300 1400 1500 1600 1700
FIRING TEMPERATURE (deg C)
Fig. 5.3: Density vs Firing Temperature - Timcerm
DENSITY (g/ml) 3.5r-----------------------------------------------,
3.3
.3.1
2.9
2.7 + 2.5 HOURS • 5.0 HOURS
2.5~ ____ ~ ____ ~I------~J.------~i------~----~i----~
1000 1100 ;1200 1300 1400 1500 1600 1700
FIRING TEMPERATURE (deg C)
Fig. 5.4: Density vs Firing Temperature - Cera Hydrate
Univers
ity of
Cap
e Tow
n
58
The trends for the bayerite material were not as well defined as may be seen from Figure 5.5. The overall densities achieved were also lower, the average density at 1500°C and 4 hours being 2,9 g/ml.
3.1 D _E_N_S~l_T_Y~(~g/_m_I):__ ______________________________ ~-,
2.Q ... •
2.7 ~
2.3~
+
+
• •
• 1 HOUR + 6 HOUR
* I
•
2.1'--~~-.L.-·~~--'-'~~~i·~~~~·~~~~·~~~~·..__~~-
1000 1100 1200 1300 1400 1600 1600 1700 FIRING TEMPERATURE (deg 0)
Fig. 5.5: Density vs Firing Temperature - Bayerite
The trends for the bayerite material were not as wen defined as may be seen from Figure 5.5. The overall densities achieved were also lower, the average density at 15000C and 4 hours being 2,9 g/ml.
3.1 rDE_N_S~I_Ty~(g~/_m~I) ____________________________ ~~-.
2.9 •
2.1
2.6
2.3
+
+
• •
• 1 HOUR + I) HOUR
* •
2.1L------L-1-----i------~.------~----~.------~1~----~
1000 1100 1200 1300 1400 1600 1600 1700 FIRING TEMPERATURE (deg 0)
Fig. 5.5: Density vs Firing Temperature - Bayerite
Univers
ity of
Cap
e Tow
n
59
5.2.3 X-ray Diffraction Analysis.
The results of the X-ray diffraction analysis of the material obtained after firing were as follows:
i) Boehmite precursor: 1
An examination of Figures 5. 6 and 5. 7 showed that in the beta alumina synthesised from both Timcerm and Cera Hydrate by the method described above, only B 11
- alumina was present. There was no evidence at all of a peak a! 47° 28 which would have indicated the presence of 6-alumina.
At a firing temperature of 1100°C, the 13 11-alumina structure had not yet fully developed, as evidenced by the lower peak heights and increased peak width. However, at 1200°c the B "-alumina was fully crystalline, showing no further sharpening or increase in peak height at increasingly higher firing temperatures.
ii) Bayerite precursor: An examination of Figures 5. 8 and 5. 9 showed that the beta alumina synthesised from commercial bayerite by the method described above consisted entirely of 6 "-alumina. There was no evidence of a peak at 47 °28 which would have indicated the presence of .6-alumina. At a firing temperature of 1100 0c, the B 11 -alumina structure had not yet fully developed as evidenced by a slightly broadened peak at 46,0 °28. However, above 1200°C the .6"-alumina was fully crystalline, and showed little further sharpening or increase in peak height as the soak time and temperature increased. Some of the minor peaks, such as that at 41,65°20 showed an increase in sharpness with increasing firing temperature, while others such as at 44,5°20 remained unchanged throughout.
An examination of the XRD pattern obtained for the beta alumina synthesised from a sample of experimentally preeipitated bayerite type material showed that it contained both 6- and B "-alumina (Figure 5 .10).
59
5.2.3 X-ray Diffraction Analysis.
The results of the X -ray diffraction analysis of the material obtained after firing were as follows:
i) Boehmite precursor: i
An examination of Figures 5.6 and 5.7 showed that in the beta alumina synthesised from both Timcerm and Cera Hydrate by the method described
above, only fi" - alumina was present. There was no evidence at all of a peak at 47° 29 which would have indicated the presence of fi-alumina.
At a fIring temperature of 11 OOoC, the fi" -alumina structure had not yet fully
developed, as evidenced by the lower peak heights and increased peak width. However, at 12000C the filt-alumina was fully crystalline, showing no further sharpening or increase in peak height at increasingly higher fIring temperatures.
ii) Bayerite precursor: An examination of Figures 5.8 and 5.9 showed that the beta alumina synthesised from commercial bayerite by the method described above consisted entirely of fi" -alumina. There was no evidence of a peak at 47 OZ9 which would have indicated the presence of fi-alumina. At a firing temperature of 1100 °C, the fi"-alumina structure had not yet fully developed as evidenced by a slightly broadened peak at 46,0 OZ9. However, above 12000C the filt-alumina
was fully crystalline, and showed little further sharpening or increase in peak height as the soak time and temperature increased. Some of the minor peaks, such as that at 41,65°29 showed an increase in sharpness with increasing fning temperature, while others such as at 44,5°29 remained unchanged throughout.
An examination of the XRD pattern obtained for the beta alumina synthesised from a sample of experimentally preCipitated bayerite type material showed that it contained both fi- and fi" -alumina (Figure 5 .10).
Univers
ity of
Cap
e Tow
n
60
BETA ALUMINA FROM CERA HYDRATE 3•ttf'-.----------------------------,
-en a. 0
~G ~ 0 L-------w 1-z 0 1---------
44.5 45 45.5 46 46.5
DEGREES 2 THETA
Fig. 5.6: .6"-alumina from boehmite - Cera Hydrate
1600deg C
1520deg C
1400 deg c
1300deg C
1200 deg C
1100degC
47 47.5
BETA ALUMINA FROM CERA HYDRATE 31&10+.,..---------------------------,
-UJ Co ~o (f) z 0 W IZ 01--------
44.5 45 45.5 46 46.5
DEGREES 2 THETA
Fig. 5.6: filt-alumina from boehmite - Cera Hydrate
1600 deg C
1520 deg C
1400 deg C
1300deg C
1200 deg C
1100 deg C
47 47.5
Univers
ity of
Cap
e Tow
n
-"' 0. 2. ~o en z w 1-
BETA ALUMINA FROM TIMCERM
(\
z 0 1--------
16 O deg c
1400 deg c
HOO deg C
al=====:::::;==-~~-,--~~~,--~~_;.======~====~ 44.5 45 45.5 46 46.5 47 47.5
DEGREES 2 THETA
Fig. 5.7: .6"-Alumina from boehmite - Timcerm
61
-en 0-,g. ~o en z w I
BETA ALUMINA FROM TIMCERM
(\
Z 0 1---------
16 0 deg C
1400 deg C
1100 deg C
o!======;~----~----~----~~====T=====~ 44.5 45 45.5 46 46.5 47 47.5
DEGREES 2 THETA
Fig. 5.7: aft-Alumina from boehmite - Timcerm
61
Univers
ity of
Cap
e Tow
n
62
BETA ALUMINA FROM BAYERITE 4 HOUR SOAK
-o en 0. (.) -~o ~ en c Q) ...... c 0
0
i.
1600 deg C
40.0 41.0 42.0 43.0 44.0 45.0 46.0·47.0 48.0 49.0 50.0
Degrees 2 Theta
Fig. 5.8: 6"-Alumina from Bayerite - 4 hour soak
62
BETA ALUMINA FROM BAYERITE 4 HOUR SOAK
i-
:l.5JlIJ..,----------------------,
_ 0
C/)
0.. () -~o
.:t::! C/) C CD ........ C 0
o
o
1600 deg C
40.0 41.0 42.0 43.0 44.0 45.0 46.0-47.0 48.0 49.0 50.0
Degrees 2 Theta
Fig. 5.8: S"-Alumina from Bayerite - 4 hour soak
Univers
ity of
Cap
e Tow
n
BETA ALUMINA FROM BAYERITE 1 HOUR SOAK
2·')rJc3i-r---------------------------.
0+-~---,~~-.-~~..,..-~---..,.--~-.-~~-.--~---.~~-.-~~-.--~--1 40.0 41.0 42.0 43.0 44.0 45.0 46.0 47.0
Degrees 2 Theta
Fig. 5.9: .6"-alumina from Bayerite - 1 hour soak
BETA ALUMINA FROM CHEMICALLY PRECIPITATED BAYERITE
48.0 49.0
0+-~~~~~~~~~~~~~~~~~~~~~--"-..~---l 40.0 41.0 42.0 43.0 44.0 45.0 46.0 47.0 48.0 49.0 50.0
Degrees 2 Theta
Fig. 5.10: Mixed .6"- and .6-alumina from chemically precipitated bayerite
50.0
63
BETA ALUMINA FROM BAYERITE 1 HOUR SOAK
.1.')rJ~,-r-------------------------~
O+-----r---~----_r----~----~--~----~----~----~--~ 40.0 41.0 42.0 43.0 44.0 45.0 46.0 47.0
Degrees 2 Theta
Fig. 5.9: fin-alumina from Bayerite - 1 hour soak
BETA ALUMINA FROM CHEMICALLY PRECIPITATED BAYERITE
48.0 49.0
O+---~-----.----r---~----'----.----~----r---~--~ 40.0 41.0 42.0 43.0 44.0 45.0 4B.0 47.0 4B.0 49.0 50.0
Degrees 2 Theta
Fig. 5.10: Mixed .6"- and fi-aIumina from chemicaIIy precipitated bayerite
50.0
63
Univers
ity of
Cap
e Tow
n
Using the equation
%.6 = 2,29 1(026 ) I (2,29 1(026) + 1(0210))
derived from the work of Pekarsky and Nicholson (1980) as discussed above, the amount of .6-alumina present was estimated;' to be 68.5 percent.
5.3 DISCUSSION.
64
The amount of sodium lost during firing appears to have been limited sufficiently by • the method of stacking five discs together such that the stoichiometric composition of the central disc was maintained.
The densities achieved by the boehmite derived .6 11 -alumina are approximately the theoretical density of the material. There appears to be no need to employ a firing temperature higher than 1250 to 1300°C in order to ensure full densification of the powder compact.
The lower densities achieved by the bayerite derived .6 "-alumina may be attributed to the slightly coarser powder used as starting material. Kalsi et al (1984) found that the sintered density of .6 11 -alumina was independent of uniaxial compacting pressure, but decreased with an increase in particle size of the powder.
X-ray diffraction was found to be a simple yet precise way of determining the .616 11 -
alumina ratio. The beta alumina derived from both pure boehmite and pure bayerite was found to consist entirely of .6 11-alumina, as predicted by Van Zyl(1987). A fully crystalline .6 "-alumina structure could be achieved by firing at temperatures as low as 1200°C in both cases. Where a mixture of aluminium hydroxide precursors was reacted to form beta alumina, a mixture of .6- and .6 "-alumina was obtained, as expected from the investigations of Van Zyl (1987).
Using the equation
%fi = 2,29 1(026) I (2,29 1(026) + 1(0210))
derived from the work of Pekarsky and Nicholson (1980) as discussed above, the amount of fi-alumina present was estimated;' to be 68.5 percent.
5.3 DISCUSSION.
64
The amount of sodium lost during frring appears to have been limited sufficiently by • the method of stacking five discs together such that the stoichiometric composition of the central disc was maintained.
The densities achieved by the boehmite derived fi" -alumina are approximately the theoretical density of the material. There appears to be no need to employ a firing temperature higher than 1250 to 13000C in order to ensure full densification of the powder compact.
The lower densities achieved by the bayerite derived fi" -alumina may be attributed to the slightly coarser powder used as starting material. Kalsi et al (1984) found that the sintered density of fi" -alumina was independent of uniaxial compacting pressure, but decreased with an increase in particle size of the powder.
X -ray diffraction was found to be a simple yet precise way of determining the filfi"alumina ratio. The beta alumina derived from both pure boehmite and pure bayerite was found to consist entirely of fi"-alumina, as predicted by Van Zyl(1987). A fully crystalline fi n -alumina structure could be achieved by firing at temperatures as low as 12000C in both cases. Where a mixture of aluminium hydroxide precursors was reacted to form beta alumina, a mixture of fi- and fi"-alumina was obtained, as expected from the investigations of Van Zyl (1987).
Univers
ity of
Cap
e Tow
n
65
6. CONCLUSION.a.
The conventional route for the formation of beta alumina has involved the high temperature solid state reaction of a-alumina with soda and lithia or magnesia, and resulted in a mixture of B- and fi"-alumina. A second heat-treatment was then necessary to maximise the amount of 6 "-alumina in the mixture in order to ensure the optimal conductivity of the solid electrolyte material. Van Zyl (1987) demonstrated that by replacing the a-alumina component with an aluminium hydroxide precursor that has a compatible oxygen sub-lattice, it was possible to form virtually single phase B "-alumina directly. The two precursors which best fulfilled this condition were found to be boehmite and bayerite.
This study has investigated the possibility of synthesising the necessary aluminium hydroxide precursors directly from readily available chemicals in order to further simplify the production process of beta alumina. The differences in structure of the aluminium trihydroxides and oxyhydroxides have been highlighted and effective methods of identification discussed. In particular, historical problems of the incorrect identification of the polymorphs have been addressed.
A study of the methods of synthesis reported in the literature for these trihydroxide and oxyhydroxide polymorphs was conducted and several methods were examined experimentally. Efforts to synthesise boehmite by simple reaction at room temperature were unsuccessful, consistently yielding pseudoboehmite, and confirming that boehmite is a hydrothermally formed phase which can only be synthesised under conditions of simultaneous high temperature and pressure. The introduction of such a reaction step would re-incur the energy cost which this investigation sought to avoid, although this may still be lower than the cost of the a-alumina route depending on the economies of scale.
The synthesis of bayerite proved more successful, although not without difficulty. By manipulating the reaction conditions, particularly pH and washing to remove extraneous ions, well-crystallised bayerite was formed by simple reaction at ambient conditions. This offers a route for the direct synthesis of bayerite from readily available chemicals with no cost penalties from extra heat-treatment steps. It may also promise a route for the direct synthesis of 6 n -alumina from such starting materials without the intermediate precursor drying stage, if the efficient removal of undesirable ions can be achieved. This may be an area for future investigation.
Unfortunately, a mixture of bayerite and pseudoboehmite was frequently formed under seemingly identical conditions. This is problematic as it will lead to the formation of a mixture of .8- and .6 "-alumina. ·A. greater understanding of factors controlling the reaction is necessary in order to overcome this obstacle.
A study of the mechanisms of formation presented in the literature was conducted, and certain common factors identified. There appeared to be agreement that the polymorph type is determined at the polymerisation stage by the general chemical
6. CONCLUSION ....
The conventional route for the formation of beta alumina has involved the high temperature solid state reaction of a-alumina with soda and lithia or magnesia, and resulted in a mixture of B- and B"-alumina. A second heat-treatment was then necessary to maximise the amount of .8 n -alumina in the mixture in order to ensure the optimal conductivity of the solid electrolyte material. Van Zyl (1981) demonstrated that by replacing the a-alumina component with an aluminium hydroxide precursor that has a compatible oxygen sub-lattice, it was possible to form virtually single phase B "-alumina directly. The two precursors which best fulfilled this condition were found to be boehmite and bayerite.
This study has investigated the possibility of synthesising the necessary aluminium hydroxide precursors directly from readily available chemicals in order to further simplify the production process of beta alumina. The differences in structure of the aluminium trihydroxides and oxyhydroxides have been highlighted and effective methods of identification discussed. In particular, historical problems of the incorrect identification of the polymorphs have been addressed.
A study of the methods of synthesis reported in the literature for these trihydroxide and oxyhydroxide polymorphs was conducted and several methods were examined experimentally. Efforts to synthesise boehmite by simple reaction at room temperature were unsuccessful, consistently yielding pseudoboehmite, and confirming that boehmite is a hydrothermally formed phase which can only be synthesised under conditions of simultaneous high temperature and pressure. The introduction of such a reaction step would re-incur the energy cost which this investigation sought to avoid, although this may still be lower than the cost of the a-alumina route depending on the economies of scale.
The synthesis of bayerite proved more successful, although not without difficulty. By manipulating the reaction conditions, particularly pH and washing to remove extraneous ions, wen-crystallised bayerite was formed by simple reaction at ambient conditions. This offers a route for the direct synthesis of bayerite from readily available chemicals with no cost penalties from extra heat-treatment steps. It may also promise a route for the direct synthesis of .8" -alumina from such starting materials without the intermediate precursor drying stage, if the efficient removal of undesirable ions can be achieved. This may be an area for future investigation.
Unfortunately, a mixture of bayerite and pseudoboehmite was frequently formed under seemingly identical conditions. This is problematic as it will lead to the formation of a mixture of .8- and .8" -alumina. 'A greater understanding of factors controlling the reaction is necessary in order to overcome this obstacle.
A study of the mechanisms of formation presented in the literature was conducted, and certain common factors identified. There appeared to be agreement that the polymorph type is determined at the polymerisation stage by the general chemical
Univers
ity of
Cap
e Tow
n
66
environment. The assertion that an alkaline environment favours the formation of bayerite was supported by experimental evidence. Furthermore, the fact that any ions, solvent or impurities act to stabilise amorphous structures relative to crystalline forms was widely identified and confirmed in experiment.
However, the formation of gels and the role they play in the reaction mechanism remains poorly defined. This is an obstacle to effective control of the reaction to give the reproducible results necessary for larger scale production, and is an area in which more detailed study is required. A comprehensive and coherent mechanism of formation needs to be derived as the basis for further experimental investigation and • optimisation of the process.
Finally, the synthesis of 6 "-alumina by the solid state reaction of the hydroxide precursors with sodium- and lithium carbonate, was successfully completed. The beta alumina derived from both pure boehmite and pure bayerite was found to consist entirely of 6" -alumina.
Where a mixture of mixture of crystalline bayerite and amorphous pseudoboehmite was reacted to form beta alumina, a mixture of 6- and 6"-alumina was formed. No investigation of varying ratios of intimately mixed polymorphs on the ratio of 6- to 6 " -alumina was conducted. Future examination of this relationship may yield information concerning the effects of seeding on the solid state reaction and on the kinetics of the reaction. An understanding of this will be necessary if it proves impractical to obtain consistently pure bayerite by the method described.
In conclusion, it is possible by reaction of AIC13.6H20 with Na40H at ambient temperature to form bayerite, an aluminium trihydroxide polymorph having the ideal oxygen sub lattice structure for the direct formation of pure 6 "-alumina by solid state reaction. This offers the opportunity for significant cost savings in the production of beta alumina products by eliminating the need for a second heat-treatment stage to optimise the ratio of 6- to 6 "-alumina. The possibility of further simplifying the synthesis route by eliminating the drying stage, that is adding the ingredients for the solid state reaction directly to the polymorph slurry, may exist. However, further work on understanding and controlling the reaction conditions will be necessary before this could be achieved.
environment. The assertion that an alkaline environment favours the formation of bayerite was supported by experimental evidence. Furthermore, the fact that any ions, solvent or impurities act to stabilise amorphous structures relative to crystalline forms was widely identified and confmned in experiment.
However, the formation of gels and the role they play in the reaction mechanism remains poorly defined. This is an obstacle to effective control of the reaction to give the reproducible results necessary for larger scale production, and is an area in which more detailed study is required. A comprehensive and coherent mechanism of formation needs to be derived as the basis for further experimental investigation and • optimisation of the process.
Finally, the synthesis of fi II -alumina by the solid state reaction of the hydroxide precursors with sodium- and lithium carbonate, was successfully completed. The beta alumina derived from both pure boehmite and pure bayerite was found to consist entirely of fi" -alumina.
Where a mixture of mixture of crystalline bayerite and amorphous pseudoboehmite was reacted to form beta alumina, a mixture of 8- and 8"-alumina was formed. No investigation of varying ratios of intimately mixed polymorphs on the ratio of fi- to fi "_ alumina was conducted. Future examination of this relationship may yield information concerning the effects of seeding on the solid state reaction and on the kinetics of the reaction. An understanding of this will be necessary if it proves impractical to obtain consistently pure bayerite by the method described.
In conclusion, it is possible by reaction of AlCI3.6H20 with Na40H at ambient temperature to form bayerite, an aluminium tribydroxide polymorph having the ideal oxygen sublattice structure for the direct formation of pure fi" -alumina by solid state reaction. This offers the opportunity for significant cost savings in the production of beta alumina products by eliminating the need for a second heat-treatment stage to optimise the ratio of fi- to fi" -alumina. The possibility of further simplifying the synthesis route by eliminating the drying stage, that is adding the ingredients for the solid state reaction directly to the polymorph slurry, may exist. However, further work on understanding and controlling the reaction conditions will be necessary before this could be achieved.
Univers
ity of
Cap
e Tow
n
REFERENCES
ALDCROFT, D. and BYE, G.C. (1969): "Crystallisation processes in aluminium hydroxide gels-III". Proc. British Ceramics Society, 125-141.
ALDCROFT, D., BYE, G.C. ~d HUGHES, C. (1969): "Crystallisation Processes in aluminium hydroxide; gels-IV". J. Applied Chemistry, 19, 167-172.
: /
BARNHISEL, R.I. and RICH, C.I. (1965): "Gibbsite, Bayerite, and Nordstrandite formation as affected by anions, pH and mineral surfaces". Soil Science Proceedings, 531-534.
BRAUER, G. (1963): Handbook of Preparative Inorganic Chemistry, 2nd ed., 815-819.
BROSSET, C., BIEDERMAN, G. and SILLEN, L. (1954): Acta Chem. Scand., 8, 1917.
BYE, G.C. and ROBINSON, J.G. (1964): "Crystallisation processes in aluminium hydroxide gels". Kolloid 'Zeitschrift, 198, 53-60.
CHESWORTH, W. (1971): "Mineral synthesis in the system :an amalgamation technique". Nature Physical Science, 230, 65.
ERVIN, G. and OSBORN, E. (1951): "The system Al203-H20". lour. Geog., 59, 381-394.
FANELLI, A. and BURLEW, J. (1986): "Preparation of fine alumina powder in alcohol". J. Ceram. Soc. , 69(8], cl 7 4-175.
FEITKNECHT, W. (1954): Kolloid 'Zeitschrift, 136, 52-56.
FEITKNECHT, W., WITTENBACH, A. and BUSER, W. (1961): 4th Symp. Reactiv. Solids, Elsevier, 234-9.
FREUND, F. (1967): Ber. Deut. Keram. Ges., 44, 141-7.
FRICKE, V. and MEYRING, K. (1933): Z. Anorganische und Allgemeine Chemie, 214, 269-274.
FRIPIAT, J. and PENNEQUIN, M. (1965): Bull. Soc. Chim. Fr., 1655.
GINSBERG, H., HUTTIG, W. AND STIEHL, H. (1962): Z. Anorg. Alig. Chem., 318, 238-56.
GITZEN, W. (1970): "Alumina as a ceramic material". Amer. Ceram. Soc. , p253.
GLINKA, A., PACEWSKA,B., MICHALOWSKI,S. (1964): J. Thermal Analysis, 29, 953.
HEM, J.D. and ROBERSON, C.E. (1967): "Form and stability of aluminium hydroxide complexes in dilute solutions". U.S. Geological Survey Water Supply Papers, 1827-a, 55.
REFERENCES
ALDCROFT, D. and BYE, G.C. (1969): "Crystallisation processes in aluminium hydroxide gels-III", Proc. British Ceramics Society, 125-141.
ALDCROFT, D., BYE, G.C. ~d HUGHES, C. (1969): .. Crystallisation Processes in aluminium hydroxide;gels-IV". J. Applied Chemistry, 19, 167-172.
: /
BARNHISEL, R.I. and RICH, C.I. (1965): " Gibbsite, Bayerite, and Nordstrandite formation as affected by anions, pH and mineral surfaces". Soil Science Proceedings, 531-534.
BRAUER, G. (1963): Handbook of Preparative Inorganic Chemistry. 2nd ed., 815-819.
BROSSET, C., BIEDERMAN, G. and SILLEN, L. (1954): Acta Chem. Scand., 8, 1917.
BYE, G.c. and ROBINSON, J.G. (1964): "Crystallisation processes in aluminium hydroxide gels". Kolloid Zeitschrift, 198, 53-60.
CHESWORTH, W. (1971): "Mineral synthesis in the system :an amalgamation technique". Nature Physical Science, 230, 65.
ERVIN, G. and OSBORN, E. (1951): "The system Al203-H20". Jour. Geog., 59, 381-394.
FANELLI, A. and BURLEW, J. (1986): "Preparation of fine alumina powder in alcohol" . J. Ceram. Soc., 69[8], c17 4-175.
FEITKNECHT, W. (1954): Kolloid Zeitschrift, 136, 52-56.
FEITKNECHT, W., WITIENBACH, A. and BUSER, W. (1961): 4th Symp. Reactiv. Solids, Elsevier, 234-9.
FREUND, F. (1967): Ber. Dew. Keram. Ges., 44, 141-7.
FRICKE, V. and MEYRING, K. (1933): Z. Anorganische und Allgemeine Chemie, 214, 269-274.
FRIPIAT, J. and PENNEQUIN, M. (1965): Bull. Soc. Chim. Fr., 1655.
GINSBERG, H., HUTTIG, W. AND STIEHL, H. (1962): Z. Anorg. Allg. Chem., 318. 238-56.
GITZEN, W. (1970): "Alumina as a ceramic material". Amer. Ceram. Soc., p253.
GLINKA, A., PACEWSKA,B., MICHALOWSKI,S. (1964): J. Thermal Analysis, 29, 953.
HEM, J.D. and ROBERSON, C.E. (1967): "Form and stability of aluminium hydroxide complexes in dilute solutions". U. S. Geological Survey Water Supply Papers, 1827-a, 55.
Univers
ity of
Cap
e Tow
n
HODGE, J.D. (1983): Ceramic Bulletin, 62,2.
HSU, P.H. (1967): Soil Science, 103, 2.
HSU, P.H. ( 1977): Minerals in Soil Environments, Soil Science Society of America, , Wisconsin, USA, ch.4.
' HSU, P.H. and BATES, T.F. (1964): Mineral Magazine, 33, 749-768.
HSU, P.H. and RICH, C.I. (1960): Soil Science Proceedings, p21.
JUAN,V.C. amd LO, H.J. (1975): Acta Geologica Taiwanica, 13, 1-8.
JOHNSON, D., GRANSTAFF, S. and RHODES, W. (1979): "Preparation of 6-A1203 pressing powders by spray drying". Ceramic Bulletin, 58, 9.
KALSI, H. TANDON, R., SINGH, B., GOEL, R.and DAS, B. (1984): "Preparation and characterisation of 6"-A1203 11 • Bulletin of Materials Science, 6, 6.
KENNEDY, G. (1959): "Phase relatio!1ships in the sy~tem A1203-H10 at high temperatures and pressures . Am. Jour. Scz., 257, 563-573.
KRAUSE, A., SKUPIN, W. and LOMOZIK, L. (1970): Naturwissenschaften, 51, 2, 88-89.
KRISCHNER, H. (1971): "Structure and surface formation of aluminium oxides". Trav. Com. Int. Etude Bauxites, Oxydes, Hydroxydes Alum, 1, 107-116.
KRIVORUCHKO, 0., ZOLOTOVSKII, B. PLYASOVA, L., BUYANOV, R. and ZAIKOVSKII, V. (1982): React. Kinet. Catal. Lett., 21, 1-2, 103-8.
KVACHKOV, R., YANAKIEV, A., POULIEFF, C., BALKANOV, I., YANKULOV, P. and BUDEVSKI, E. (1982): J. Matera! Science, 17, 3672-3.
LAMBER, R. (1986): "Thin boehmite films: preparetion and structures". J. Materials Science Letters, 5, 177-8.
LAUBENGAYER, A and WEISZ, R. (1943): J. Amer. Chem. Soc., 65, 247.
LINSEN, B. (ed) (1970): Physical and Chemical Aspects of Adsorbents and Catalysts, Academic Press, London.
LIPPENS, B. and STEGGERDA, J. (1970): Active Alumina, p171-211. in Physical and Chemical Aspects of Adsorbents and Catalysts, Academic Press, London.
MACKENZIE, R.C. (1957): The Differential Thermal Investigation of Clays, p311-320.
MATIJEVIC, E.and TEZAK, B. (1953): J. Physical Chemistry, 51, 951-4.
MAY, G. and HOOPER, A. (1978): J. Material Science, 13, 1480-6.
HODGE, J.D. (1983): Ceramic Bulletin, 62,2.
HSU, P.H. (1967): Soil Science, 103, 2.
HSU, P.H. (1977): Minerals in Soil Environments, Soil Science Society of America, ; Wisconsin, USA, chAo
HSU, P.B:. and BATES, T.F. (1964): Mineral Magazine, 33, 749-768.
HSU, P.H. and RICH, C.I. (1960): Soil Science Proceedings, p21.
WAN,V.C. amd LO, H.J. (1975): Acta Geologica Taiwanica, 13, 1-8.
JOHNSON, D., GRANSTAFF, S. and RHODES, W. (1979): "Preparation of.6-Al203 pressing powders by spray drying". Ceramic Bulletin, 58, 9.
KALSI, H. TANDON, R., SINGH, B., GOEL, R.and DAS, B. (1984): "Preparation and characterisation of S"-Al203'" Bulletin o/Materials Science, 6, 6.
KENNEDY, G. (1959): "Phase relatio,?-ships in the sy~tem Al203-H20 at high temperatures and pressures. Am. Jour. SCI., 257, 563-573.
KRAUSE, A., SKUPIN, W. and LOMOZIK, L. (1970): Naturwissenschaften, 57, 2, 88-89.
KRISCHNER, H. (1971): "Structure and surface formation of aluminium oxides". Trav. Com. Int. Etude Bauxites, Oxydes, Hydroxydes Alum, 7, 107-116.
KRIVORUCHKO, 0., ZOLOTOVSKII, B. PLYASOVA, L., BUYANOV, R. and ZAIKOVSKll, V. (1982): React. Kinet. Catal. Lett., 21, 1-2, 103-ft
KVACHKOV, R., YANAKIEV, A., POULIEFF, C., BALKANOV, I., YANKULOV, P. and BUDEVSKI, E. (1982): J. Materal Science, 17, 3672-3.
LAMBER, R. (1986): "Thin boehmite films: preparetion and structures". J. Materials Science Letters, 5, 177-8.
LAUBENGAYER, A and WEISZ, R. (1943): J. Amer. Chem. Soc., 65, 247.
LINSEN, B. (00) (1970): Physical and Chemical Aspects 0/ Adsorbents and Catalysts, Academic Press, London.
LIPPENS, B. and STEGGERDA, J. (1970): Active Alumina, p171-211. in Physical and Chemical Aspects 0/ Adsorbents and Catalysts, Academic Press, London.
MACKENZIE, R.C. (1957): The Differential Thennal Investigation o/Clays, p311-320.
MATIJEVIC, E.and TEZAK, B. (1953): J. Physical Chemistry, 57, 951-4.
MAY, G. and HOOPER, A. (1978): J. Material Science, 13, 1480-6.
Univers
ity of
Cap
e Tow
n
McHARDY, W. and THOMSON, A. (1971): Mineralogical Magatine, 38, 358.
MEGAW, H. (1934): Z. Krist.,87,185-204.
MILLIGAN, W. and McAtee, J. (1956): J. Phys. Chem., 60, 273.
MISRA, C. and SNAKUMAR, T. (1986): Ger. Offen., DE 3528534 al, 19 June 1986, 18pp. I . :
MOELLER, T. (1952): Inorganic Chemistry: An Advanced Text, J.Wiley and Sons, 754-57.
MOISSEY, K., CZANDERNA,K. AND CARTER, C. (1984): Comm. American Ceram. Soc., c-88.
MORGAN, P. (1982): J. Materials Science, 17.
MOSELY, P. (1985): The solid electrolyte properties and characteristics, Ch. 2, The Sodium Sulphur Battery, (eds. J.L. Sudworth & A.R.Tilley), Chapman and Hall, London.
MUKAIBO, T., TAKAHASHI, Y. and YAMADA, K. (1969): 1st Int. Corif. Calorimetry and Thermodynamics, Warsaw, N4.
NEWSOME, J., HEISER, H., RUSSEL, A. and STUMPF, H. (1960): Aluminium Co. of America Tech. Rep. JO, 2nd Ed ..
NG KEE KWONG, K. HUANG, P. (1977): "Influence of Citric acid on the Hydrolytic reactions of aluminium". J. Soil Sd. Soc. Amer., 41.
NG KEE KWONG, K. HUANG, P. (1979): Soil Science, 128, 6
PACKTER, A. (1976): Colloid and Polymer Science, 254,1024-9.
PACKTER A. and DHILLON, H. (1975): Colloid and Polymer Science, 252, 249-56.
PAPEE, D., TERTAIN,R. and BIAIS, R. (1958): Bull. Soc. Chim. Fr., 1301-10.
PETZ, J. (1968): "Structure of Aluminium Hydroxide Gel". J. Chemical Physics, 48, 2.
PEKARSKY, A. and NICHOLSON, P. (1980): Materials Research Bulletin, 15, 1517.
POULIEFF, C., KAVCHKOV, Rand BALKANOV,I. (1978): Materials Research Bulletin, 13, 323.
RYSKIN, Y. (1974): "The vibration of protons in mineral" in ¥.Farmer (ed) Infrared Spectra of Minerals, Mineralogical Society, London.
SAALFELD, H. and JARCHOW,O. (1968): Neues Jahrb. Mineral. Abh., 109, 185-91.
SATO, T. (1981): J. Chem. Technology and Biotechnology, 31(11), 670-5.
W. and THOMSON, A. (1971): Mineralogical Magazine, 38, 358.
H. (1934): ,87,185-204.
MILLIGAN, W. and McAtee, J. (1956): 1. Phys.
MISRA, C. and SIVAKUMAR, T. (1986): Ger. 1986, 18pp.
,60,273.
• DE 3528534 a1, 19 I :
MOELLER, T. (1952): I1UU'onHU' Chemistry: An Advanced Text, J.Wiley and 754-57.
K., CZANDERNA,K. AND CARTER, (1984): Comm. American Ceram. Soc., c-88.
MORGAN, P. (1982): 1. Materials Science, 17.
(1985): The solid electrolyte properoes characteristics. Ch. The Sodium Sulphur (eds. J.L. Sudworth A.R.Tilley), Chapman and Hall, London.
MUKAIBO, ,TAKAHASm, and YAMADA, K. (1969): 1st Int. Corif. Calorimetry and Thermodynamics. Warsaw, N4.
NEWSOME, J., HEISER, , ... " .... n ................. . (1960): Aluminium of America Tech. Rep.
NG KWONG, K. HUANG, (1977): "Influence acid on the Hydrolytic reactions of aluminium". 1. Soil Sci. Soc. Amer., 41.
K. HUANG, (1979): Soil Science, 128, 6
A. (1976): Colloid and Polymer Science, 254,
PACKTER A. and DHILLON, Colloid and Polymer 3CU!nce. 249-56.
, 1301-10. and BlAIS, (1958): Bull. Soc.
PETZ, (1968): "Structure of Aluminium Hydroxide Gel". 1. Chemical Physics, 48,
and NICHOLSON, (1980): Materials Research Bulletin, 15, 1517.
POULIEFF, , KA VCHKOV, Rand BALKANOV,I. (1978): Materials Kesearir:h Bulletin, 13, 323.
RYSKIN, (1974): "The vibration of mineral" in IInr;mpI" (ed) Irifrared Spectra of Minerals, MUleraLlogu:ru Society, London.
SAALFELD, 91.
SATO,
JARCHOW,O. (1968): Neues lahrb. Mineral. Abh., 109, 185-
Chem. Technology and Biotechnology, 31(11), 670-5.
Univers
ity of
Cap
e Tow
n
SCHOEN, R. and ROBERSON,, C. (1970): "Structures of Aluminium hydrooxide and Geochemical Implications". The American Mineralogist, 55, 43-77.
SCHMID, C. (1986): "X-ray characterisation of 6- and 6"-alumina". J. Materials Science Letters, 5,263-6.
STOL, R., VAN BELDEN, A. and DE BRUYN, P. (1976): "Hydrolysis precipitation studies of Aluminium (III) solution". J. Colloid <ind Interface S<;ience, 57, 1. .
TOBIAS, R. and TYREE,S (1959): J. Amer. Chem Soc., 82, 6385.
TURNER, R. and ROSS, G. (1970): Canadian Journal a/Chemistry, 48, 723-9.
VANSTRATEN, H., HOLTZKAMP, B. and DE BRUYN, P. (1984): "Precipitation from supersaturated solutions -I". J. Colloid and Interface Science, 98, 2.
VAN ZYL, A. (1987): PHD Thesis, University of Cape Town.
VIOLANTE, A. and VIOLANTE, P. (1980): Clays and Clay Minerals, 28(6), 425.
VOIGT, vR., SEIFERT, K. and NEUHAUS, A. (1977): N. Jb.Miner.Mh., H5, 238-40.
WADE, K. and BANNISTER, A. (1973): Comprehensive Inorganic Chemistry, 1, 1032-6.
WILSON, S. (1979): J. Solid State Chemistry, 30, 257.
YOLDAS, B. (1973): J. Applied Chemistry and Biochemistry, 23(11), 803-9.
• •
·~
SCHOEN, R. and ROBERSON" C. (1970): "Structures of Aluminium hydrooxide and Geochemical Implications". The American Mineralogist, 55, 43-77.
SCHMID, C. (1986): "X-ray characterisation of.6- and fi"-alumina". J. Materials Science Letters, 5,263-6.
STOL, R., VAN HELDEN, A. and DE BRUYN, P. (1976): "Hydrolysis precipitation studies of Aluminium (Ill) solution". J. Colloid dnd Interface S~ience, 57, 1. .
TOBIAS, R. and TYREE,S (1959): J. Amer. Chem Soc., 82, 6385.
TURNER, R. and ROSS, G. (1970): Canadian Joumal o/Chemistry, 48, 723-9.
VAN STRATEN, H., HOLTZKAMP, B. and DE BRUYN, P. (1984): "Precipitation from supersaturated solutions -I". J. Colloid and Interface Science, 98, 2.
VAN ZYL, A. (1987): PHD Thesis, University o/Cape Town.
VIOLANTE, A. and VIOLANTE, P. (1980): Clays and Clay Minerals, 28(6), 425.
VOIGT, vR., SEIFERT, K. and NEUHAUS, A. (1977): N. Jb.Miner.Mh. , H5, 238-40.
WADE, K. and BANNISTER, A. (1973): Comprehensive Inorganic Chemistry, 1, 1032-6.
WILSON, S. (1979): J. Solid State Chemistry, 30, 257.
YOLDAS, B. (1973): J. Applied Chemistry and Biochemistry, 23(11), 803-9.
Univers
ity of
Cap
e Tow
n
SCHOEN, R. and ROBERSON,, C. (1970): "Structures of Aluminium hydrooxide and Geochemical Implications". The American Mineralogist, 55, 43-77.
SCHMID, C. (1986): "X-ray characterisation of 6- and -6"-alumina". J. Materials Science Letters, 5,263-6.
STOL, R., VAN HELDEN, A. and DE BRUYN, P. (1976): "Hydrolysis precipitation studies of Aluminium (III) solution". J. Colloid and Interface Sl:ience, 57, 1.
TOBIAS, R. and TYREE,S (1959): J. Amer. Chem Soc., 82, 6385.
TURNER, R. and ROSS, G. (1970): Canadian Journal of Chemistry, 48, 723-9.
VANSTRATEN, H., HOLTZKAMP, B. and DE BRUYN, P. (1984): "Precipitation from supersaturated solutions -I". J. Colloid and Interface Science, 98, 2.
VAN ZYL, A. (1987): PHD Thesis, University of Cape Town.
VIOLANTE, A. and VIOLANTE, P. (1980): Clays and Clay Minerals, 28(6), 425.
VOIGT, vR., SEIFERT, K. and NEUHAUS, A. (1977): N. Jb.Miner.Mh., H5, 238-40.
WADE, K. and BANNISTER, A. (1973): Comprehensive Inorganic Chemistry, 1, 1032-6.
WILSON, S. (1979): J. Solid State Chemistry, 30, 257 ..
YOLDAS, B. (1973): J. Applied Chemistry and Biochemistry, 23(11), 803-9.
R. and , C. (1970): "Structures of Aluminium hydrooxide and Geochemical Implications". The American Mineralogist, 55, 43-77.
SCHMID, C. (1986): "X-ray characterisation of Science Letters, 5,263-6.
.8" -alumina" .
R., VAN HELDEN, and DE BRUYN, (1976): "Hydrolysis precipitation studies of Aluminium (III) solution" . J. Colloid and Interface Stience, 57. L
and (1959): J. Amer. Chem Soc., 82, 6385.
TURNER, R. and ROSS, G. (1970): Canadian Journal of Chemistry, 48, 723-9.
VAN H., and (1984): from supersaturated solutions -I". J. Colloid Interface Science,
VAN (1987): University of Cape Town.
A. and (1980): Clays and Clay Minerals, 28(6), 425.
, SEIFERT, K. NEUHAUS, A. (1977): N. Jb.Miner.Mh. , H5. 40.
WADE, K. and BANNISTER, (1973): Comprehensive Inorganic Chemistry, 1, 1032-6.
WILSON, S. (1979): J. Solid Chemistry, 30, 257.
YOLDAS, (1973): J. Applied Chemistry and Biochemistry, 23(11), 803-9.
![Contents · PDF filecomplexing agents, like inorganic anions (fluoride and phosphates) or ... and solubility product of amorphous and crystalline aluminium hydroxide [Ohman 1988]](https://img.pdfslide.us/doc/110x75/5a9995067f8b9a30358d9c00/contents-agents-like-inorganic-anions-fluoride-and-phosphates-or-and-solubility.jpg)
