2016

Research Proposal

1. Summary
Aluminium production by the Bayer process is a vital part of the Australian mining industry.
However the high energy and environmental cost of the Bayer process necessitates the
development of an alternative. The most energy intensive step of the Bayer process, and the
subject of this research, is calcination. The proposed solution is a concentrated solar thermal
(CST) calciner. Due to fundamental differences between a conventional flash calciner and the
CST calciner, the proposed research aims to investigate the effect of reactor atmosphere on
the final calcined product. The researcher aims to perform a series of calcination experiments
in a tube furnace at 1000 oC. The effect of reactor atmosphere will be investigated by
performing calcination in different gaseous environments. Observation of the changes in
chemical reaction rates and analysis of the calcined material by x-ray diffractometry (XRD)
will illustrate the effect of gas composition. The results of this research will add to the
existing literature and will serve as a guide for future research that will eventually lead to a
solar calciner.
2. Background
The Bayer process is for the refinement of bauxite to smelter grade alumina that is a
precursor for aluminium metal. The first step is digestion of bauxite in caustic solution
followed by clarification to separate the insoluble red mud from the valuable liquid. The
liquid is cooled and gibbsite is precipitated. The gibbsite is washed before it enters the last
stage of the Bayer process; calcination. Calcination involves heating the aluminium hydrate
to approximately 1000 oC to and converts into aluminium oxide that is the feedstock for
aluminium smelters. The Bayer process accounts for nearly 40% of Australias mineral
processing CO2 emissions. In 2014 Alumina production consumed 220 PJ of energy in which
127 PJ was natural gas (International Aluminium Institute, 2016). The refining process alone
released an estimated 14 million tonnes of CO2-e in 2011 (Australian Aluminium Council,
2016). The majority of the natural gas is combusted in the energy intensive calcination step.
An alternative energy source would greatly reduce the environmental impact of aluminium
production. To replace the combustion of natural gas in calcination, the proposed solution is a
calciner powered by CST energy.
Concentrated solar thermal power systems generate electricity by concentrating solar
radiation into a small region, where a molten salt serves as a heat transfer medium. The heat
is used to generate steam which in turn drives an electric generator. The main barrier to
widespread adoption of these systems is the molten salt, which is required to efficiently store
heat for extended periods of time. The proposed solar calciner is not as complex as power
generation systems and is thus more feasible.The calciner would be a solar vortex reactor
mounted atop a tower surrounded by an array of heliostats that would concentrate sunlight
onto the calciner. CST calcination offers promise of being the solution because it can produce
the required energy, is suited to the Australian climate, emits no greenhouse gases and is
feasible with present technology.
Due to the commercial nature of the alumina calcination there are two problems associated
with research. Firstly, valuable data is proprietary information that is not made available to
researchers. Secondly, commercially funded research is oriented towards the narrow scope of
the sponsors interests. The aim of this research is to observe the effect of reactor gas
composition on the calcination product. A potential solar calciner will have a different
internal atmosphere than a traditional flash calciner, due to the absence of natural gas
combustion. This is important because the internal atmosphere affects the morphology of the
calcined product, consequently changing the reactivity which has ramifications downstream

when the alumina is converted into aluminium. Understanding the fundamental processes of
how the atmosphere in the reactor affects the calcined product will further understanding of
existing processes and will allow the creation of more accurate simulations eventually the
construction of a solar calciner.
3. Aims and Objectives
The CST calciner atmospheric conditions might be outside the parameters of previous
research. Therefore, the proposed research will be fundamental in nature and will use
parameters that are beyond the current scope of commercial Bayer process operations. The
objective of this research is to achieve the following:
1. Identify the presence and characteristics of different transition alumina after
calcination in different reactor atmospheres.
2. Generate data that will serve as a starting point for future work.

4. Literature Review
As gibbsite reacts it also undergoes a series of morphological changes before forming alumina. The intermediate oxide stages between gibbsite and -alumina are termed the
transition alumina. Brown et al. (1953) were the first to demonstrate that gibbsite decomposes
to either monohydrate boehmite or to -alumina, upon further heating the two products follow
independent routes to - Al2O3 as shown in Figure 1. The phases of ,,,,, alumina were
identified by using x-ray diffraction (XRD) techniques. Papee and Tertian (1955) identified
an amorphous - Al2O3, which forms under conditions of high vacuum rapid heating of
gibbsite. The formation of - Al2O3 is of limited relevance in the calcination process because
the resultant alumina rehydrates easily to form gibbsite or boehmite. The characteristics of
transition alumina are determined by reactor pressure, particle size, heating rate and reactor
atmosphere Alevra et al. (1972).

Figure 1: Common transformation sequences of calcined alumina (B Gan et al., 2009)

4.1 Formation of boehmite
Brown et al. (1953) and de Boer et al. (1954) identified after the calcination of large grained
gibbsite (Al(OH)3), one of the phases present, was boehmite (AlOOH). Further analysis
revealed that the boehmite was located in the interior of the calcine. Formation of boehmite
requires very high water vapor pressure, which was absent in the experiments. de Boer et al.
(1954) suggested that the interior of the gibbsite crystal was strong enough to withstand the
high pressures that promote the growth of the boehmite phase. In essence, each gibbsite
crystal acts as a high pressure reaction vessel for the boehmite. Tertian and Papee (1953,
1955) and Rouquerol et al. (1975) demonstrated the enhanced formation of boehmite with

increasing particle size of gibbsite (Al(OH)3), which was attributed to the greater structural
strength of the larger crystals.
Candela and Perlmutter (1992) examined the formation of boehmite under water vapor
pressures ranging from 100-3200 Pa over a range of 448 to 478K for particles between 40180 m. In that study, an internal nucleation model for the formation of boehmite was
proposed, where boehmite forms by random internal nucleation as shown in Figure 2. The
boehmite sites will keep forming as long as the high intracrystalline pressure is present. The
reaction rate starts to slow down when the internal conversion sites are connected to the
surface and the high pressure conditions are absent. The kinetics for the conversion were a
combination of the internal growth rate of boehmite sites and the rate at which water vapor
diffuses to the surface. Increasing water vapour partial pressure increased the final boehmite
yield.

Figure 2: Internal nucleation model for the formation of boehmite (Candela & Perlmutter,
1992)
Candela & Perlmutter (1992) is the only major study concerning the formation of boehmite
and is the generally accepted model. . Due to the nature of boehmites formation it is
considered a minor reaction pathway and is often neglected in large scale simulations. A
reaction order of approximately 1 based on experiments by Peri et al. (1996) and Zhu et al.
(2010) was determined by using the Kissinger equation. On the other hand, Candela and
Perlmutter (1992) determined a reaction order in the order of 2.5 indicating a combined
kinetic and diffusion controlled process. Further work is required to validate the approach and
hence the true role of reactor atmospheric conditions.

4.2 Decomposition of gibbsite and boehmite

The decomposition of boehmite is better understood than its formation because it is a surface
level reaction that can be monitored using the similar method to monitor the decomposition
of gibbsite. Callister et al. (1966) investigated the role of water vapor in the decomposition of
boehmite. The workers observed that increasing water vapor partial pressure led to a decrease
in the rate of decomposition. Callister et al. (1966) suggested two mechanisms to explain the
role of H2O but they were unable to distinguish between the mechanisms based on their

experimental methodology. The first mechanism was an adsorption-desorption process, which

is an irreversible decomposition reaction (1) occurs at active sites (S) on the crystal surface
and produces transition alumina. A competing reversible adsorption process also occurs on
the active sites (2). The increase in vapor pressure would increase the adsorption process and
slow decomposition of boehmite.
S

k1

products

(1)

S + H2O kb ka S.H2O

(2)

The other suggestion was the formation of an intermediate active oxide species that had to
diffuse to become part of the bulk crystal phase. The formation of this active oxide species is
an additional reversible reaction that occurs in conjunction with the normal decomposition.
The increase in water vapor pressure would favour the reverse reaction (3).
Al2O3H2O k2 k1 Al2O3x + H2O

(3)

Later work by Stacey (1987) also found that increasing vapor pressure decreased the
decomposition rate and identified a reaction order of n = -0.4 with respect to water vapor for
the decomposition of boehmite and n= -1.3 for the decomposition of gibbsite. An
investigation by Chen et al. (1989) on the decomposition of boehmite did not mention the
effect of water vapor but proposed a two-step sequential reaction with order n= 0.5.
An investigation by Hrabe et al. (1992) found that water vapor accelerated the phase
transformations of boehmite and catalyzed the nucleation of -Al2O3 at higher temperatures
(1020oC). Hrabe et al. (1992) postulated that water interacted on grain surfaces creating fast
diffusion paths resulting in nucleation and growth phases. Bagwell and Messing (1999)
conducted a similar experiment to study isothermal transformation kinetics and crystal
growth of boehmite to -Al2O3. The workers made the same observations about accelerated
kinetics due to increased vapor pressure. They also came to the same conclusion and
suggested increased diffusion due to water vapor as being responsible for the increased rate
of transformation. It is still unclear why this accelerating effect was not observed at the lower
temperature ranges (200-500 oC) of Callister et al. (1966) and Stacey (1987).
Attempts to model to the calcination process have applied models that were developed for
other gas-solid reactions like coal gasification Amiri et al. (2015). Most of the systems
described in literature are sharp interface models, which assume that the reaction takes place
on the surface of a spherical core which reacts and shrinks leaving behind an inert ash which
the product gas has to diffuse through. Models such as the unreacted shrinking core describe
systems with reactions of the form
aA(g)+bB(s)pP(g)+qQ(s)

(4)

where the diffusion rate of reactant gas, A, to the surface of reactant solid, B, is the rate
controlling step. In contrast the Bayer calcination reaction is of the form:
bB(s)pP(g)+qQ(s)

(5)

Diffusion mechanics in aluminium oxides are not well defined and the rate limiting species
has yet to be determined (Bagwell and Messing, 1999). Additionally, there is uncertainty
about the reaction order for each step of the calcination process. The derived order value
depends on the method chosen by the investigator, each of whom is able to provide
justification for their results. Therefore, the actual effect of gas composition on the

decomposition of gibbsite/boehmite is yet to be determined and most investigations are only

able to propose plausible mechanisms. To fully understand the effect of gas composition,
more research needs to be done to develop a complete model of alumina calcination.

5. Approach & Methodology

The experimental method in this research does not seek to directly replicate current Bayer
process conditions. At this preliminary stage, the research is exploratory in nature with
fundamental mechanisms and qualitative results being the end goal. Ultimately samples of
interest will serve as gateways for future research.
5.1 Required apparatus and materials
No.
1

Component
Gibbsite

2
3

X-ray diffractometer (XRD)

Thermogravimetric analyzer (TGA)

4
5

Tube furnace
Gas

6
7

Aluminium crucible
Thermocouple

Rotameter

Sieve shaker

10

Alcoa KA13

Description
50-200 m
>99.9% purity
For sample analysis
Measure dry weight of sample before
and after calcination
Primary calcination device
N2

CO2,
dry air
steam
Contains sample
Monitors temperature and heating rate
of the sample in tube furnace
For measuring gas flow rate into tube
furnace
For classification of raw gibbsite
according to size
70% alpha alumina
45-150 m

Table 1: Required materials

5.2 Gibbsite
The purpose of the gibbsite sample size (100-110 m) is to allow the formation of boehmite
to be observed.
5.3 Tube furnace
Past research by Johnson (1990) and Wang et al. (2006) on calcination has often used the
TGA to simulate the calciner However there are some disadvantages to this approach. Firstly
the TGA available at the University of Adelaide is incapable of varying the atmosphere as

required by the experiment. Secondly, the TGA process is very time consuming due to the
slow heating rates.
In comparison the tube furnace while lacking the real-time mass change data that a TGA can
provide is much suitable to this experiment. The tube furnace allows variation of the internal
atmosphere and induction of a wide range of heating regimes, up to 100 oC/min.

5.4 Variables
Particle size, calcination temperature, furnace pressure (atmospheric) and heating rate will be
kept constant since they would otherwise affect the calcination procedure.
5.5 Outline of experimental procedure

Obtain the dry weight of the raw gibbsite sample (100-110 m) by placing a small
amount of gibbsite into TGA and heating to approximately 100 oC for 5 minutes. This
will evaporate water adsorbed to the surface of the sample.
Tube furnace, at atmospheric pressure, is heated to 1000oC and gas is pumped into the
tube furnace to establish the required gas composition.
Sample is placed into crucible and placed into the furnace. Sample remains in furnace
for set period of time e.g. 5 minutes.
Sample is taken out and placed into TGA at 100 oC to obtain the dry weight after
calcination.
Above procedure is repeated for a range of calcination times between 5-30 minutes.
Gas atmospheres of pure N2, pure CO2 and different mixtures of steam and dry air will
be tested.

5.6 Results and analysis

The tube furnace experiments will produce mass change against time data. This will be used
to calculate the reaction rate. From the literature it is expected the different steam and air
mixtures will have different reaction rates. However the effect of pure N2 and CO2
atmospheres on the calcination process is still unknown.
When plotted, the mass change vs. time data will likely produce a decay curve with two
inflection points. The first inflection is due to the release of water when gibbsite reacts and
forms boehmite. The second inflection is due to the release of water when boehmite reacts
and forms - Al2O3. The different gas compositions will affect the position and size of the two
inflections, allowing comparisons to be made.
The calcined samples will then be tested using x-ray diffraction techniques. The resulting
diffraction patterns will be compared against reference samples. The different peak intensities
produced by each sample will allow identification of the different alumina phases present and
their relative quantities. The diffraction patterns will show the degree of crystallinity in the
form of well-defined peaks high intensity peaks. The diffraction patterns produced by the
samples calcined in steam-dry air mixtures are expected to have more distinct intensity peaks

as the amount of steam is increased. The diffraction patterns produced by the samples
calcined in N2 and are of particular interest because there is no precedent in the literature. The
Alcoa KA13, a commercially calcined product, will be used for comparison with the
experimental samples.

6. Planning
The expected timeframe for the project is 12 weeks divided into three phases; preliminary,
experimental and reporting. Weeks 1 to 2 are for preliminary arrangements. Weeks 3 to 8 are
for the furnace experiments and XRD analysis. The actual experimental period may be much
shorter depending on various factors, however eight weeks is the maximum allowable time
that can be spent on obtaining results. The final four weeks will be spent writing the final
report and preparing a poster presentation.
The researcher will be working in conjunction with another student who will be testing the
variables, heating rate and particle size. As part of the research process the two researchers
will be attending weekly meetings with their supervisor.
The main obstacle to timely completion of the project is access to the laboratory. Currently
the University of Adelaide only has one tube furnace available and it is required by other
researchers. The tube furnace experiments may be completed in a week but the four week
time frame in the timetable is meant to display the current uncertainty. At the present time it
is expected the researchers will have to make arrangements for laboratory access outside
normal hours. This is further complicated by the supervision requirements of the University
of Adelaide.
6.1 Safety
Due to the safety concerns associated with high temperature experiments, the researchers will
need to undergo a safety induction for use of the tube furnace and the XRD. Part of the
induction will include hazard management training and the completion of a risk assessment
document. Supervision by a PhD student is required for both the XRD and the furnace.
6.2 Materials
All the required materials except the raw gibbsite and the XRD standard reference material
(SRM) are available and ready to use. The raw gibbsite is easy to obtain and there are a range
of suppliers that will supply it. The XRD analysis requires (SRM) for calibration. The SRM is
currently unavailable at the University of Adelaide and arrangements to obtain it will be made
as soon as the second semester begins. Alternatively the samples can be sent to a third party
laboratory that specializes in XRD, this arrangement is more advantageous because it will
allow higher quality results to be obtained, however this will result in significant monetary
costs . A decision will be made early in the second semester on which approach will be
chosen.

6.3 Dissemination and Significance

The results of the research project will be summarized in a final report and research poster.
The poster will be presented to the public at Ingenuity 2016. Due to the nature of the
experiment and the limitations of the analysis techniques currently available to the researcher,
the final report will be fundamental in nature and will serve as guide for future research
directions. If this research results in novel observations, experimental alumina samples of
interest will be kept by the University of Adelaide for further in depth study. Previous
research by Johnson (1990) and Alevra et al. (1972) observed the effect of steam on the
surface area of the calcined material. The morphology of the calcined material might be a
topic for future research. Based on this research project, ideally in the future researcher will
engage in a more detailed study of the subject.

W1

W2

Lab Induction and

Equiptment
training
Acquire gibbsite &
KA13
Sieving sample
preparation
Tube furnace
experiments
Progress report
XRD analysis
Write report
Poster
Present report
Poster presentation

Table 2: Expected semester two timetable

W3

W4

W5

W6

W7

W8

W9

W 10

W 11

W 12

W 13

References

Alevra, V.; Ciomirtan, D.; Ionescu, M.(1972) Influence of the Preparation

Method on the Structure and Physical Properties of Alumina. 11, Aluminum
Oxides. Reo. Roum. Chim., 17 (8), 1379

Amiri, A., Ingram, G., Maynard, N., Livk, I., & Bekker, A. (2015). An Unreacted
Shrinking Core Model for Calcination and Similar Solid-to-Gas Reactions. Chemical
Engineering Communications, 202(9), 1161-1175.

Australian Aluminium Council. (2016). Alumina Refining Greenhouse

Performance. [ONLINE] Available at: http://aluminium.org.au/climatechange/refining-greenhouse-performance. [Accessed 02 May 2016].
Bagwell, R., & Messing, G. (1999). Effect of Seeding and Water Vapor on the
Nucleation and Growth of Al 2 O 3 from Al 2 O 3. Journal of the American
Ceramic Society, 82(4), 825-832.
J. F. Brown, D. Clark and W. W. Elliott, (1953) The thermal decomposition
of the alumina trihydrate, gibbsite, J. Chem. Soc., 84-88
Callister, W., Cutler, I., & Gordon, R. (1966). Thermal Decomposition Kinetics
of Boehmite. Journal of the American Ceramic Society, 49, 419-422
Candela, L., & Perlmutter, D. (1992). Kinetics of boehmite formation by
thermal decomposition of gibbsite. Industrial & Engineering Chemistry
Research, 31(3), 694-700
Gan, B K., Madsen, I C., & Hockridge, J G. (2009). In situ X-ray diffraction of
the transformation of gibbsite to [alpha]-alumina through calcination: Effect of
particle size and heating rate. Journal of Applied Crystallography, 42(4), 697705

Hrabe, Z., Komarneni, S., Pach, L., & Roy, R. (1992). The influence of water
vapor on thermal transformation of boehmite. Journal of Materials Research,
7(02), 444-449
International Aluminium Institute (2016) World Aluminium Metallurgical
Alumina Refining Fuel Consumption. [ONLINE] Available at: http://www.worldaluminium.org/statistics/metallurgical-alumina-refining-fuelconsumption/#data. [Accessed 02 May 2016]
Johnson, M. (1990). Surface area stability of aluminas. Journal of Catalysis,
123(1), 245-259
Molina, A., and Mondragon, F. (1998). Reactivity of coal gasification with
steam and CO2, Fuel, 77, 18311839.
D. Papee, R. Tertian, (1955). Bull. Soc. Chim. France. 983
Papee, D Tertian, R. (1953). Comp. Rend. Acad. Sci. 236, 1565
J. Peri, R. Krstulovi, and M. Vuak (1996) Investigation of dehydroxylation
of gibbsite into boehmite by DSC analysis, Journal of Thermal Analysis and
Calorimetry , 46(5), 1339-1347
Rouquerol, J., Rouquerol, F., & Ganteaume, M. (1975). Thermal
decomposition of gibbsite under low pressures: I. Formation of the boehmitic
phase. Journal of Catalysis, 36(1), 99-110

Stacey, M. (1987). Kinetics of decomposition of gibbsite and boehmite and the

characterization of the porous products. Langmuir, 3(5), 681-686.
Wang, H., Xu, B., Smith, P., Davies, M., DeSilva, L., and Wingate, C.(2006).
Kinetic modelling of gibbsite dehydration/amorphization in the temperature
range 823923 K, J. Phys. Chem. Solids, 67,25672582.
Zhu, Boquan, Fang, Binxiang, & Li, Xiangcheng. (2010). Dehydration
reactions and kinetic parameters of gibbsite. Ceramics International, 36(8),
2493-2498.