BAUXITE EXTRACTABLE PHASES IN THE BAYER HIGH

TEMPERATURE PROCESS:

RE-ASSESSMENT OF BOEHMITE CONTENT AND

ALUMINIUM SUBSTITUTION IN ALUMINO-GOETHITE

M. Authier-Martin and G.D. Fulford

Alcan International Limited

Kingston Research and Development Centre

P.O. Box 8400, Kingston, Ontario, Canada K7L 5L9

F. Feret

Alcan International Limited

Arvida Research and Development Centre

P.O. Box 1250, Jonquière, Québec, Canada G7S 4K8

ABSTRACT

The exact quality of the bauxite that will feed Alcan alumina refineries beyond the year 2000 is a subject of significant interest to the process plants. Concern arises from the mining of lower grade bauxite and the potential of high variability within a single deposit. Alcan's traditional approach in bauxite characterization, using sound wet chemistry methods and semi-quantitative X-ray diffraction, has been successful for many years. Nevertheless, a number of alternative future bauxites have presented new challenges that have stretched Alcan's characterization methodology to its limits. More specifically, boehmite levels in low-boehmitic bauxites (< 4 % b.Al2O3) and aluminium substitution in alumino-goethite have received careful re-assessment. For example, accurate determination of boehmite levels in bauxites is critical to the appropriate choice of processing strategy; with lower and lower grade bauxite available, alumino-goethite from Jamaica and Boke bauxites presents an appealing potential source of additional alumina while it also raises a number of issues and possible problems.

Various analytical techniques, among them XDBTM, BQUANTTM, deuterated FT-IR, DSC and dithionite leach, were used to characterize boehmite and alumino-goethite in a series of Jamaica, Weipa, Ely and Boke bauxites. Typical results are presented in this paper and the mining and process implications are highlighted. Both the consistencies and the disparities of the data recorded from various analytical techniques are used to shed new light on the future bauxite quality and its processability.

KEY WORDS:

analysis, phase quantification, bauxite, boehmite, alumino goethite.

BAUXITE EXTRACTABLE PHASES IN THE BAYER HIGH TEMPERATURE PROCESS: RE-ASSESSMENT OF BOEHMITE CONTENT AND ALUMINIUM SUBSTITUTION IN ALUMINO-GOETHITE

M. Authier-Martin, G.D. Fulford and F. Feret

1.0 INTRODUCTION

1.1 Bauxite Grade

The Bayer process can readily extract alumina through a caustic digestion at elevated temperatures from gibbsitic and mixed bauxites. To do so more economically, the Bayer plants process low-mono bauxites (boehmitic alumina, b.Al2O3 < 4% ) at low temperatures (> 102 to 150° C) and higher boehmite bauxites at high temperatures (> 200 to 265° C). This calls certainly for an accurate determination of the grade of the bauxite.

Traditionally, the Bayer operations were more interested in the low- and high-temperature extractable alumina phases, assuming that the difference could be assigned to boehmite. Declining reserves in low-mono bauxites and mining of poorer grades of bauxites at present bring forth more strongly than before two ever present concerns: the need to know accurately the levels of boehmite to minimize boehmite reversion in low-temperature plants and the impact of alumina recovery from potentially appealing secondary phases in high-temperature processes (e.g., from alumino-goethite). Therefore, the shift in bauxite analysis is to move from the determination of caustic extractable phases to the true quantification of mineralogical phases.

The next sections will describe the traditional wet chemical laboratory digestion methods, the current instrumental methods and the analytical tools under development along with their impact on the characterization of past, present and future bauxite feeds.

2.0 DESCRIPTION OF ANALYTICAL METHODOLOGY

2.1 Wet Chemical Methods

Traditionally, the Alcan standard methods of analysis for bauxite grade certification or assessment (shipments or process control) rely on Alcan solubility data for synthetic gibbsite and boehmite in pure sodium hydroxide. The analytical methods are hence carried out in pure caustic soda (2.5 N or 133 g/L as Na2CO3), and stipulate that the alumina -to-caustic ratio (A/C) at the end of the low-temperature digestion be maintained in a controlled window of 0.41 to 0.45 to ensure complete gibbsite dissolution while preventing any boehmite extraction or reversion.

Alcan analytical methodology calls for the simultaneous determination of alumina and silica extractable phases in a single digestion using a small rotating block heater and 45 mL pressure vessels (Authier-Martin, 1995). Digesters such as Alcan 16-place block heaters ensure uniform temperature control, good solid-liquid mixing and good analytical throughput. It is equipped with several safety features and the Parr pressure vessels are made of Nickel 200 wetted parts and have stainless steel closures, for additional safety (e.g., vessel integrity).

The need to quantify alumina and silica, under low and high temperature caustic digestions, has led to the use of more drastic conditions than called for if only the alumina phases were to be determined.

At low temperature, while maintaining the f40inal A/C ratio in the critical 0.41-0.45 window, the laboratory digestion temperature and time have been set at 150° C and 25 minutes to ensure complete gibbsite dissolution and kaolinite attack and re-precipitation of silica as Bayer sodalite. Alumina in solution (i.e., gibbsite) is measured by complexometry (CDTA/ZnSO4). By acidification, Bayer sodalite is then released into solution and the silica level in the acidified solution (i.e., kaolinite) is measured by colorimetry (silico-molybdate). For low-gibbsitic bauxites (gibbsitic alumina, g.Al2O< 41%), the digestion is carried out in a caustic sodium aluminate solution chosen to achieve a final 0.41-0.45 A/C ratio for a fixed bauxite weight.

At high temperature, the analytical conditions may be selected to only allow complete dissolution of gibbsite and boehmite; this is possible if the minimum following conditions are applied; temperature = 200° C, A/C ~ 0.26, time = 30 minutes. Also to ensure complete kaolinite and quartz attack and re-precipitation as Bayer sodalite, the conditions have to be set at A/C ~ 0.09, t = 60 minutes, T =  225° C. Alumina in solution is again measured by complexometry (CDTA/ZnSO4) and silica in solution, before and after acidification, by colorimetry (silico-molybdate).

For high grade tri-hydrate bauxites, the extractable species seem up to now to match, within the precision limits of the methods, the mineralogical phases as measured by other means. For lower grade bauxites, the match may become less and less perfect as other minerals than gibbsite and boehmite (or kaolinite and quartz) could be extracted. The introduction of secondary impurities into the caustic digestion solutions may also shift the solubility data gathered in pure caustic for gibbsite and boehmite, as shown in other related work carried out in our laboratories. Moreover, Jamaica fine bauxites have already been shown to behave differently from other tropical bauxites, if only to mention their high phosphorus content linked to crandallite, a mineral behaving similarly to gibbsite as regards to alumina extractability. So the drive to characterize bauxite by other means than wet chemical laboratory digestion methods is not new and will be addressed in the next sections.

2.2 Alternative Methods

2.2.1 X-Ray Diffraction

In the early 80’s, X-ray diffraction appeared as a very promising tool for phase quantification (Strahl, 1982), the capital costs and high expertise needed being somewhat outweighed by automation and access to information on additional phases (e.g., titania and iron minerals). XRD still remains a very powerful semi-quantitative tool, but over the years it was shown to be plagued by several inherent features making XRD less appealing for true quantitative applications, to name only a few: preferred orientation effects, crystallinity effects, calibration being deposit specific and relative technique based on wet chemical data. More recently, it was also found that XRD calibration may not be truly linear due to matrix composition effect while bauxite may exhibit heterogeneity in phase amorphicity. All these aspects may certainly be addressed to increase the reliability in terms of accuracy and precision of the XRD analysis, but this makes it a more complex tool, not easy to standardize and to implement.

2.2.2 Linear Regression

The task of processing thousands of exploration samples at the lowest possible costs presented a big challenge in the early 90’s, more so since the declining grade and increasing complexity of the future bauxites required more data related to phase composition than the previous core composite approach could allow. Feret and Giasson (1991) developed a mathematical approach i.e. a polynomial regression relating wet chemistry data with loss of mass (LOM, 105-1000oC) and elemental composition (Al, Si, Fe, Ti by XRF). This approach assumed that foot-by-foot samples were analyzed with the highest accuracy for LOM and major elemental oxides, the non-bauxitic samples be discarded and that a small set of representative samples be accurately analyzed by wet chemical laboratory digestions for extractable phases. A correlation was drawn for this set and then applied to the remaining composite bauxite samples. The linear regression method allowed high throughput at lower costs with higher reliability, but it was shown to be deposit specific and still relied on wet chemical digestion data.

2.2.3 XDBTM

During the same period, Sajó (1994) developed a method combining XRD and a mathematical approach. Sajó’s XDB software attempts to match simultaneously an XRD profile fit and a mineralogical mass balance, the former based on the XRD pattern and the latter on full elemental composition. This method was successfully applied to a series of Boke bauxites (Feret, Authier-Martin, Sajó, 1997)(6). XDB provides a complete quantitative mineralogical analysis for bauxites of different origins. It has proven to be a powerful R&D analytical tool, but the throughput is low (5-10 samples per day), and the expertise needed high, while it leaves room for a high degree of subjectivity to achieve a compromise match between the XRD fit and the mass balance. The know-how gained at Alcan through the linear regression approach and the XDB software led our scientists to develop a proprietary software for bauxite quantification.

2.2.4 BQUANTTM

BQUANTTM is an Alcan proprietary expert system, combining the expertise of a mathematician and several crystallographers, geologists and Bayer chemists. It is a mathematical approach that calculates, through Windows based software (Feret and Kimmerle, 1997), the mineralogical composition of bauxites from LOM, elemental analysis and organic carbon by attempting a full mass balance between the 15 elemental oxides measured by XRF and the assumed stoichiometric composition of several mineral species known to be present in bauxites. Unlike XRD and similarly to the linear regression, BQUANTTM has a very high throughput, involves very low analytical costs and shows better reliability than wet chemical digestion methods (less operator involvement). It does not have to rely on any calibration or wet chemical digestion data. Because BQUANTTM makes some basic assumptions on the "ideal" composition of the geological phases, this causes some limitations (e.g., level of Al2O3 in alumino-goethite fixed by the degree of aluminium substitution selected). As for any other analytical tool that reports mineralogical phases, some data interpretation needs to be made between BQUANTTM results (minerals) and the traditional wet chemical values (caustic extractable species). Overall though, BQUANTTM has so far led us to a better understanding of our Bayer chemistry and has re-opened some relevant R&D issues.

2.3 Instrumental Methods

Two areas used traditionally for qualitative or semi-quantitative analysis and phase information have been re-investigated more recently: thermal analysis (DSC) and infrared spectroscopy (FT-IR). Both techniques appeared initially very promising for direct determination of gibbsite and boehmite, hence possibly increasing the precisions encountered with wet chemical methods to calculate - usually by difference - alumina recovery from bauxite and red mud data.

2.3.1 Thermal Analysis

In thermal analysis, alumino-goethite interference with gibbsite analysis has been traditionally dealt with using an acid-reducing leach to remove goethite (Mg/SO2 leach). More recently, a less drastic bicarbonate-citrate-dithionite leach (Jackson, 1969) has been tested successfully to remove goethite and may well be replacing the more hazardous and difficult Mg/SO2 leach.

It was only recently that we extended the use of thermal analysis for full quantification of boehmite in bauxite. The sample is first pre-digested at a controlled A/C ratio (> 0.41) to selectively remove gibbsite while transforming the interfering kaolinite to Bayer sodalite. The same approach can be used for process red muds if they contain significant amounts of gibbsite. Thermal analysis gives a direct boehmite measurement in both matrices, enhancing precision and accuracy. It can also be used to detect and measure the amounts of native (> 500° C) and recrystallized (shoulder < 500° C) boehmite in red muds. It is still necessary to pre-remove significant amounts of gibbsite since part of the gibbsite may form additional boehmite on heating (~  300° C).

The application of thermal analysis has yet to be optimized for boehmite determination in red muds but it is a powerful technique to characterize several parameters of the same sample, as will be shown below.

2.3.2 Infrared Spectroscopy

FT-IR has been used for qualitative analysis of bauxites, but it is only with the discovery of selective deuteration (Farmer, 1974) that its full potential for quantitative analysis has been revealed. We have indeed shown that, under specific conditions, preferential deuteration of boehmite can be achieved. This allows the boehmite OH stretching peaks to be moved from the crowded area ~ 3000-3600 cm-1 to a cleaner area ~ 2300-2500 cm-1 (boehmite OD stretching peaks), making quantification much easier. Work carried out in our laboratories has shown that the calibration is linear in the 0-6% boehmite range and at least the method has been used successfully for quantitative boehmite analysis in red muds. This direct method applied to red muds shows an improvement in terms of precision and accuracy compared to the traditional wet chemical methods for boehmite working by difference.

Attempts to apply the deuteration FT-IR technique to boehmite analysis in bauxites has not led to a standardized method for three reasons:

    1. Different deuteration times (3 to 25 hours) are needed for bauxites of different origins.
    2. Possible hydrothermal transformation of gibbsite to boehmite may occur which then deuterates during boehmite deuteration in very fine bauxites (e.g., Jamaica).
    3. Possible deuteration of alumino-goethite in other bauxites (e.g., Boke bauxites).

The gibbsite transformation and Al-goethite deuteration tend to interfere with the OD peaks due to boehmite, while the variation in deuteration time is troublesome when processing "unknown" bauxites. The two interferences could be dealt with by sample pre-treatment (2 steps needed though, digestion to remove gibbsite and leach to remove Al-goethite), leaving an easy to analyze residue, but the extra steps make the technique certainly less attractive for bauxites than anticipated. On the other hand, the need to vary the deuteration time or the presence of gibbsite or Al-goethite interference could give an insight in the comparative reactivity (amorphicity ?) of different bauxites. At this stage, the FT-IR technique seems less attractive than the thermal analysis approach for boehmite quantification purposes, but it remains a promising R&D analytical tool.

3.0 RESULTS AND DISCUSSIONS

3.1 Boehmite Analysis

Traditionally, boehmite analysis was carried out within Alcan by a caustic digestion. For high mono-hydrate bauxites, boehmitic alumina was calculated by difference: high temperature - low temperature extractable alumina. For low-boehmite bauxites containing < 3% b.Al2O3, the boehmitic alumina determination was done through a sequential double digestion extraction (i.e., the direct method), first a low temperature digestion to remove gibbsite and kaolinite, followed by a mild high temperature digestion where the soluble alumina was simply associated to boehmite. As reported elsewhere, the boehmite levels obtained by wet chemistry, XDBTM and BQUANTTM agreed very well for a series of tri-hydrate bauxites, considering that the reproducibilities of the wet chemical methods, at the 95% confidence level, are 0.4% (direct method) and 1% (by difference):

Key parameters for these bauxites are presented in Table 1. The level of agreement for boehmitic alumina was on the average better that 0.5% while reaching a maximum difference of 1.5% for a few high boehmitic samples.

Table 1

Typical Compositions of Several Bauxites from Different Origins with Accurate

Boehmite Determination by Wet Chemical Laboratory Digestions

 

Samples

t.Al2O3

t.SiO2

t.Fe2O3

t.TiO2

t.P2O5

L.O.M.

g.Al2O3

b.Al2O3

(% w/w)

(% w/w)

(% w/w)

(% w/w)

(% w/w)

(% w/w)

(% w/w)

Approx.

200 samples from

13 locations

 

< 30-63

 

0.4-25

 

2-35

 

0.5-9.5

 

0-0.95

 

7.5-31

 

15-56

 

< 0.5-32

(*) ZnO: 0.002-0.025%, MnO: 0.01-0.32%; Cr2O3: 0.02-0.013%; V2O5: 0.04-0.19%.

When Caribbean bauxites are analyzed (Table 2), the level of agreement between the same techniques varies highly, but there is a definitive trend for wet chemistry to report higher values than either BQUANTTM or XDBTM. This trend is confirmed by instrumental techniques, as shown in Table 3.

Table 2

Re-Assessment of Boehmitic Alumina Levels in Caribbean Bauxites

Samples

b.Al2O3 (%w/w)

Wet Chem

BQUANT

XDB

Deposit #1

(low Al-goethite)

Average

Range

1.7

[1.4-2.2]

1.1

[0.3-1.6]

0.6

[0.4-0.8]

Deposit #2

(high Al-goethite)

Average

Range

4.2

[2.1-5.8]

2.9

[0.7-3.9]

3.0

[0.6-3.9]

 

 

Table 3

Boehmitic Alumina Levels in Predominantly Caribbean Bauxites:

Comparison between Several Techniques

Samples

b.Al2O3 (% w/w)

Wet Chem

BQUANT

XDB

FT-IR*

DSC*

CRM BXT-03

0.6**

0.2

0.0

0.0

0.0

CRM BXT-06

1.8

1.5

1.7

1.7

1.0

Yellow bauxite

3.7

3.3

2.6

--

--

Feed #1

1.7

0.3

0.9

0.5

0.2

Feed #2

2.1

1.0

0.5

0.6

0.3

Feed #3

2.7

0.6

1.9

--

--

Feed #4

3.4

2.1

2.1

4.6

2.7

CRM BXT-01

2.6

2.2

1.5

***

1.4

Feed #5

8.8

7.5

8.0

--

--

CRM BXT-09

20.0

19.7

18.9

***

18.7

Feed #6

11.0

11.0

11.2

--

--

Test pit

11.0

11.3

9.9

--

--

(*) After pre-digestion at low-temperature at a A/C ratio > 0.41.

(**) Close to limit of quantification of wet chemical method.

(***) Outside calibration range

Jamaica bauxites are low boehmitic bauxites and as such are usually analyzed within Alcan by the so-called direct method. This method was originally developed for low-mono Boke bauxites (0-3% b.Al2O3) and its application was later extended to bauxites from other origins up to a boehmitic alumina level of 20%. In retrospect, the digestion parameters used in this laboratory digestion may well be harsh enough to extract at high temperature other alumina bearing minerals than gibbsite and boehmite, e.g., part of Al2O3 from alumino-goethite. The bauxites analyzed in the past, especially the Jamaica bauxites, were low in alumino-goethite, making the possible discrepancies barely significant. Moreover, Jamaica bauxites are known to be fine and usually more highly substituted than other bauxites, both characteristics making the extraction of Al-goethite more likely. In fact, indirect evidence exists that, under the harsher laboratory conditions, such a side reaction may occur. The extent of goethite extraction and its relation to the degree of Al-substitution has not been clearly demonstrated during our laboratory work. However, it may be appropriate to infer from the work carried out by Golden (Golden, 1978) that finer highly-substituted goethites - with an increased surface area - may well extract more readily under caustic digestion than coarser low-substituted goethites. Moreover, as will be shown in the next section of this paper, Boke bauxites can be subjected to Al-goethite extraction under (milder) high temperature process conditions.

3.2 Aluminium Substitution in Alumino-Goethite

The discrepancies found for a large majority of Jamaica bauxites in what was assumed to be a boehmite determination opened the need to check the currently available analytical methodology to measure the degree of aluminium substitution in alumino-goethite.

The obsolete wet chemical method measuring that value by subtracting four different parameters had long been discarded because of its large imprecision. Two alternative methods have been since considered, i.e., X-ray diffraction (Schulze, 1984)(12) and thermal analysis (Ni and Khalyapina, 1978).

Goethite's main X-ray diffraction peak may shift by as much as 0.04 Å as the aluminium substitution increases to its usual maximum (~ 30% mol). Even though the shift is clearly significant, the peak due to goethite is small and not well defined and may even appear split. There is even question that water and aluminium substitution may shift the peak in opposite directions (Stanjek and Schwertmann, 1992). Quartz also interferes while gibbsite and kaolinite raise the background, hence limiting the limit of detection. All these aspects certainly raise serious difficulties in defining the peak maximum and displacement.

By comparison the endotherm shift in thermal analysis for alumino-goethite dehydration is even more significant (D ~ 90° C for maximum substitution, Table 4), assuming that one got rid of any excessive amounts of interfering gibbsite present in bauxites by a caustic pre-digestion.

Table 4

Shift in Dehydration Temperature for Goethite (Ni and Khalyapina, 1978)

Temperature of

Endotherm *

(oC)

Al2O3 in

Goethite

(mol. %)

Al2O3 in

Goethite

(wt. %)

310

0 0

325

5 ~3

340

10 ~6

340-355

15 ~9

370

20 ~12

400

30 ~18.5

(*) At a heating rate of 20-25° C.

Another means to look at the degree of aluminium substitution is to carry out a quantitative bicarbonate-citrate-dithionite (BCD) leach, as done in soil chemistry (Jackson, 1969). The BCD leach will selectively dissolves all iron (hydr)oxide minerals, including alumino-goethite, but excluding massive magnetite, which is normally not present in bauxites and red muds; the determination of the resulting dissolved alumina in solution, by a technique such as ICP-AES, may give an indication of the degree of aluminium substitution in goethite - assuming a priori no Al-hematite substitution, a valid assumption for normal tropical bauxites. By then comparing the thermal shift and the BCD leach result, one may draw some conclusions on the aluminium substitution in goethite, as shown in Table 5.

 

Table 5

Degree of Aluminium Substitution in Various Bauxite/Red Mud Samples:

Comparison Between Shift in Endotherms and BCD Leachates

 

Sample

No(*).

Endotherm

Temperature

(oC)

Al2O3

Content in

Fe Minerals

(wt. %)

Dissolved Al2O3

Dissolved Fe2O3

Al2O3

Content in

Fe Minerals

(wt. %)

Go/He

(***)

in BCD Leachates

(wt. %)

(wt. %)

1

334

~5

0.7

11.1

~6

17/83

2

342

~6

0.7

8.7

~7

24/76

3

350-355

~9

1.3

-

-

-

4

n.d.

0

-

-

-

-

5

370-380

~12

1.3

7.7-7.9

~14-15

78/22

330-340

~5

6(**)

370-380

~12

330-340

~5

1.9-2.5

22-25

~8-9

7

325

~3

-

-

-

8

342

~6

-

-

-

9

--

--

4.7

17.7

~21

97/3

10

350

~9

2.2

16.0

~12

70/30

11

350

~9

2.3

16.0

~13

76/24

12

335

~5

1.4

17.7

~7

65/35

n.d.: Not detected.

(*) Unless stated otherwise, muds generated in laboratory digestions using plant liquors and plant conditions; results for mud samples expressed on bauxite basis.

(**) Bayer process red muds.

(***) Goethite/Hematite ratio measured by XDB™.

Three special experiments were first carried out to check the validity of any forthcoming conclusion. Sample No. 4 was run as a quality control check to eliminate any possibility of "artifact". Samples No. 2 and 3 and samples No. 7 and 8 were processed under caustic digestions susceptible to alterate the aluminium substitution in alumino-goethite. As can be seen, under some laboratory specific conditions, Australian and Guinean bauxites can be processed to increase the aluminium substitution, hence supporting the earlier claim made by Pechiney that alumina losses to iron minerals may be encountered (Lamerant et al, 1998). However, when actual process red muds are analyzed (Sample No. 6, representing 5 muds from two plants), we may observe that the aluminium substitution in fact decreases from the bauxite feeds (Sample No. 5) to the process red muds (Sample No. 6), as clearly shown by the BCD leach results and the thermograms presented in Figure 2. Even if the same doublet is detected in the bauxite and the red mud samples, the intensities of the 370-380 peak (high substitution) and the 320-350 peaks (low substitution) are both drastically reduced from the bauxites to the red muds, especially when taking into account the concentration factor. These results clearly show that the aluminium substitution in process muds decreases (due to extraction of Al2O3 from alumino-goethite) rather than increases (i.e., alumina lost to iron minerals).

Moreover, Table 5 shows in general the same trends whether we consider endotherm shifts or BCD leachates results. The case of the Jamaica bauxites is also interesting when referring back to Table 3. Jamaica bauxites generally show both a higher level of dissolved iron minerals and a higher degree of aluminium substitution, hence hinting that Al-goethite may well be a more troublesome interference in wet chemical determinations of boehmite for these specific bauxites. We still need to consolidate these data for the two groups of bauxites presented in Table 2, where clear differences in wet chemical and BQUANTTM boehmite determinations were seen. We also want to look more closely at the BCD leachates to follow the possible release of undesirable impurities (e.g., Zn and Cr) when extracting alumina from alumino-goethite, following the dissolution of goethite, through a goethite-to-hematite transformation step.

4.0 CONCLUSIONS

There is a definite trend in analysis to move from caustic extractable species to more absolute values i.e., mineralogical phases, partly because traditional wet chemical laboratory digestion methods have been shown to analyze boehmite levels inaccurately. To achieve best practices and full business potential, it is critical that the boehmite levels be exactly known and be measured with more precision in order to better assess boehmite reversion and to more accurately assign the sources of alumina losses in Bayer operations.

A mathematical approach, through an expert system such as BQUANTTM, or instrumental methods such as thermal analysis - following some sample pre-treatment - are rapidly becoming valuable tools for mining, process control and R&D purposes, in terms of accuracy and precision as well as for throughput/cost reasons. However, modern mineralogical data from future bauxites must be ultimately related to caustic extractable species, through known correlations or through carrying out some kind of caustic digestions, well before these new in-coming bauxites are used as feed for any plant in order to avoid any nasty surprises.

Thermal analysis appears to be a very promising tool for bauxite and red mud analysis when used to determine in a single analysis boehmite and aluminium substitution in alumino-goethite, two key parameters for process optimization with current and future bauxite feeds. Coupled with microwave digestion for sample pre-treatment, modern thermal analyzers may become competitive techniques in terms of high sample throughput and low analytical costs. The BCD leach method may also bring more understanding of the nature and behavior of undesirable impurities thought to be substituted for iron in the non-stoichiometric alumino-goethite mineral.

Using thermal analysis and the BCD leach approach, we have so far found no evidence of alumina losses to iron minerals in Boke bauxites processed in high temperature Bayer plants. Moreover, it was shown that alumino-goethite extraction does occur to some limited extent, thus releasing valuable alumina into process liquors. Though the data were limited to 10 samples, they represent monthly composites for a 2-year period from two different plants.

Other interesting samples have already been identified to help consolidate the scattered data already collected. These will be characterized using BQUANTTM, XDBTM, thermal analysis and BCD leach. Along with previous samples, they will be further scrutinized with tools such as high-temperature X-ray diffraction, deuterated FT-IR and TEM.

The current drive in bauxite characterization is two-fold: (1) to minimize boehmite reversion and reprecipitation, hence improving alumina recovery, (2) to better our knowledge on iron minerals transformation in an attempt to optimize alumina recovery and mud handling while ultimately minimizing undesirable impurities in alumina.

 

 

ACKNOWLEDGMENT

We would like to thank Dr. F.M. Kimmerle, Dr. J. Doucet and Mrs. A. Taylor for their valuable comments.

NOTATION

For sake of simplicity throughout this paper, low-temperature extractable species bearing alumina are expressed as gibbsitic alumina, g.Al2O3; similarly all the additional high-temperature caustic extractable phases bearing alumina are referred to as boehmitic alumina, b.Al2O3. Boehmite and gibbsite refer specifically to the minerals.

A/C is the Alumina in g/L as Al2O3 over Caustic soda in g/L as Na2CO3.

REFERENCES

Authier-Martin, M, Bernier, J.-L., Kimmerle, F.M., ALExx, The Laboratory Tumbling Digester, 4-page Brochure, Alcan International Limited.

Strahl, EO (1982), Modern Analytical Methods in Bauxite Survey Programs, Proc. Bauxite Symposium, 5th, A. Lyew-Ayee Editor, pp. 188-203.

Feret, F., Giasson., G. (1991). Quantitative Phase Analysis of Sangaredi Bauxites (Guinea) Based on Their Chemical Composition, Light Metals, 187-191.

Sajó, I. (1994), Powder Diffraction Phase Analytical System, Version 1.7, Users Guide, Aluterv-FKI, Budapest, Hungary.

Feret, F., Authier-Martin, M., Sajó, I. (1997). Quantitative Phase Analysis of Bidi-Koum Bauxites (Guinea), Clays and Clay Minerals, 45 (3), 418-427.

Kimmerle, F.M. and Feret, F. (1997), BQUANT: Cost Effective Calculations of Bauxite Mineralogy, Light Metals, pp.1-8.

Jackson, M.L. (1969). Soil Chemical Analysis, Advanced Course (2nd ed.), Madison, Wisconsin, pp. 49-51.

Farmer, V.C. (Ed) (1974), The Infrared Spectra of Minerals, Mineralogical Society Monograph 4, Chapter 9.

Kimmerle, F.M., Feret, F., Paré, S., and Feret, B. (1998), BQUANT 98: An Efficient Software Tool for Bauxite Mine Explorations, Proceedings of 5th Alumina Quality Workshop.

Golden, D.C. (1978). Physical and Chemical Properties of Aluminum-Substituted Goethite, Ph.D. Thesis, North Carolina State University at Raleigh, pp.66-69.

Schulze, D.G. (1984). The Influence of Aluminum on Iron Oxides. VIII. Unit Cell Dimensions of Al-substituted Goethites and Estimation of Al from Them, Clays and Clay Minerals, 32, pp. 36-44.

Ni, L.P. and Khalyapina, O.B., (1978). Physical-Chemical Properties of the Raw Materials and Products of Alumina Production, p. 33, Alma-Ata, Izd-vo Nauka Kazakh. SSR.

Stanjek, H., and Schwertmann, U., (1992), The Influence of Aluminum on Iron Oxides, Part XVI: Hydroxyl and Aluminum Substitution in Synthetic Hematites, Clays & Clay Minerals, 40, pp. 347-354.

Lamerant, J.M., Le Gouic, F. and Pannetier, J. (1998). Alumina and Soda Losses During Bayer Digestion, Light Metals, pp.147-152.