Charles M Earnest^{*}, Karla Gann and Britney Stong
Department of Chemistry, Berry College, Mount Berry, GA, USA
Received Date: 05 October, 2018
Accepted Date: 06 November, 2018
Published Date: 23 November, 2018
Earnest CM, Gann K, Stong B (2018) Improved Quantification of Gibbsite in Bauxite Ores by Thermogravimetric Methods (TGA and DTG). Adv Appl Chem Biochem 2018(1): 0917.
Copyright © 2018 Charles M Earnest et al. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and work is properly cited.
The major source of alumina is from bauxite ores. Bauxite ores are not pure clay mineral species but are composed of a mixture of mineral components. The components and amounts vary with location. The main aluminous components found in bauxite ores are gibbsite (Al(OH)_{3}), boehmite (γAlO(OH)) and diaspore (αAlO(OH)). These may occur individually or as mixtures of these components. This study offers an improved approach to the quantification of the gibbsite component in naturally occurring bauxite ores using the thermal method of analysis known as either Thermogravimetry (TG) or Thermogravimetric Analysis (TGA). The methodology used in this study involves the use of an experimentally determined empirical gravimetric factor, G, rather than the theoretical factor based on the stoichiometric equation for the complete dehydroxylation of gibbsite.
From the analytical results of this study, it was concluded that by developing and employing an empirical gravimetric factor “G,” it is possible to quantify the gibbsite content in bauxite ores to within an experimental error (±1%) of that given by the method of Xray Diffraction Spectroscopy (XRD).
2. Keywords: Bauxite Ores; Derivative Thermogravimetry (DTG); Gibbsite; Gravimetric Factor (G); Quantitative Analysis; Thermogravimetric Analysis (TGA)
Thermal analysis of bauxite ores is to great extent based on the dehydroxylation of gibbsite and boehmite (or diaspore). The gibbsite dehydroxylation occurs at a lower temperature (270^{o}C) than boehmite (500^{o}C). These are described by the chemical decompositions given below:
Gibbsite: Al(OH)_{3} (s)→ Al_{2}O_{3} (s) + AlO(OH) (s) + H_{2}O (g) (Decomposition One)
Boehmite: 2AlO(OH) (s) → Al_{2}O_{3} (s) + H_{2}O (g) (Decomposition Two)
In Differential Thermal Analysis (DTA) these decompositions are observed as endothermic peaks in the DTA thermal curves. Since these decompositions are associated with the evolution of water vapor, they are also observed as weight loss events in Thermogravimetry (TGA) and Derivative Thermogravimetry (DTG). It should be mentioned that the loss of a certain amount of adsorbed water is also observed by both TGA/DTG and DTA methods of analysis. This will be shown in the corresponding TGA/DTG thermal curves obtained for an “as received” bauxite specimenfrom room temperature to ca. 140 degrees Celsius when using a dynamic purge gas in the analyzer instrument. The amount of molecular water liberated in this desorption region can affect the decomposition temperatures, described above, for both gibbsite and boehmite components of the bauxite ores. This causes a shift to higher values with increasing amounts of water vapor.
The usual method for the determination of gibbsite in bauxite samples is performed with Xray diffraction spectroscopy, which is generally accurate to ±1% [1,2]. The method presented in this paper utilizes an experimentally determined gravimetric factor, “G,” to calculate the percent gibbsite in bauxite ores obtained from several continents when using the methods of Thermogravimetry (TGA) and/or Derivative Thermogravimetry (DTG). This method of quantitation is very similar to the “method of external standards” which is frequently employed in the analytical technique of gas chromatography and several spectroscopic techniques [3].
Bauxite Clays are composed of a mixture of minerals. The components of the mixture vary with the location of the bauxite deposit. The aluminum bearing components of bauxite ores are gibbsite, (Al(OH)_{3} (Exhibit 1.a)), boehmite (AlO(OH)) and diapore (AlO(OH)) (Exhibit 1.b ) as shown below. It is these components which determine the commercial value of these ores since it is either aluminum metal or aluminum oxide powder that is subsequently derived from the bauxite. These products are used in tile, coatings, aluminum cans, fillers, and ceramics. Other minerals that are sometimes found in bauxite ores include kaolinite clay mineral (Al_{2}Si_{2}O_{5}(OH)_{4}), goethite (FeO(OH)), hematite (Fe_{2}O_{3}), siderite (FeCO_{3}), anatase (TiO_{2}_{}), magnetite (Fe_{3}O_{4}) and silica (SiO_{2}).
Exhibit 1: a) is the chemical structure of gibbsite and b) is the structure of boehmite.
The crystalline gibbsite (monoclinic) and boehmite (orthorhombic) content of bauxite ores are usually determined by analyses performed using Xray Diffraction Spectroscopy (XRD), which has an accuracy of ±1 percent for this application. In this paper, an alternate method for determining one of these aluminous minerals (gibbsite) is presented using the techniques of Thermogravimetric Analysis (TGA) and Derivative Thermogravimetry (DTG [4]). This thermogravimetric method is based on the fact that both gibbsite and boehmite decompose upon heating, at characteristic temperatures, liberating water vapor. The mass of water lost in the 270^{o}C decomposition is due, although not stoichiometrically, to the gibbsite content of the bauxite ore. Also, it should be stated, that even though the expression given above for Decomposition 2 is completely stoichiometric, the mass loss of water vapor which is experimentally observed for this decomposition may not be directly related to the boehmite content of the bauxite clay specimen. This is due to the fact that not all of the gibbsite component is converted to aluminum oxide in the previous decomposition near 270^{o}C leaving the remainder as boehmite [58]. Thus, the use of the theoretical “G” factor [1] for the measured gibbsite mass loss would lead to a low value for the percent gibbsite in the Bauxite specimen under study. This problem is the basis of the study presented here.
Standard Reference Materials (SRM) of Bauxite were purchased from the National Institute of Standards and Technology (NIST). These were SRM 697 (Dominican), SRM 698 (Jamaican), SRM 69 b (Arkansas), and SRM 600 (Australian). The composition of these bauxite standards is given in percent by mass in table 1. A tank of high purity dry nitrogen gas was purchased locally and employed in all thermogravimetric analysis experiments.

SRM 69b 
SRM 697 
SRM 698 
Kaolinite 
30 
15 
– 
Gibbsite 
60 
50 
75 
Boehmite 
– 
10 
– 
Hematite 
– 
20 
20 
Siderite 
10 
– 
– 
Pyrite 
– 
– 
– 
Anatase 
Trace 
2.0 
0.50 
All values in the table are Percent by Mass.
These values are not certified by NIST. They were obtained by a single commercial laboratory with no connections to NIST.
All TGA analyses were performed with a PerkinElmer TGA7HT high temperature thermogravimetric analyzer. The analyzer was operated in conjunction with a PerkinElmer TAC 7 thermal analysis controller. The analyzer was equipped with an external furnace with vertical platinum hang down wire to platinum sample pan and supporting platinum stirrup. The TGA data and subsequent calculations were obtained using PerkinElmer Pyris 7 thermal analysis software. All TGA data were obtained using dynamic nitrogen atmosphere of 40 milliliters per minute and a heating rate of 20 degrees Celsius per minute. Temperature measurements were made using a Pt, Pt (10% Rh) Type Sthermocouple located near the TGA sample pan. All bauxite specimens were analyzed on an “as received” basis.
The Derivative Thermogravimetric (DTG) thermal curves were generated by the Pyris 7 software subsequent to the TGA thermal curve obtained by the TGA 7 instrument. The DTG thermal curves are displayed along with the TGA thermal curves for each of the bauxite specimens which was studied (Figures 15).
Figure 1: TG and DTG Thermal Curves for SRM 697 Dominican Bauxite.
Figure 2: TG and DTG Thermal Curves for SRM 69b Arkansas Bauxite.
Figure 3: TG and DTG Thermal Curves for SRM 698 Jamaican Bauxite.
Figure 4: TG and DTG Thermal Curves for SRM 600 Australian Bauxite.
Figure 5: TG and DTG Thermal Curves for “Bauxite Unknown” Analyzed Using a Single Reference Standard.
All of the TGA weight loss data were obtained by analyzing the sample on an “as received basis.” Thus, the measured percentage mass loss values had to be corrected to a dry weight (or moisture free) basis. This correction factor is described in table 2 below. The calculation of a correction factor in each case for the NIST bauxite standards (SRM 697, 69b, 698, and 600) is determined by subtracting the percentage water loss observed by the TGA for the heating range of room temperature and ca. 125^{o}C from 100%. That is, subtract the % adsorbed water (given in the first column in table 2) from 100% and divided this corrected value into 100%. This results in the factor given in the right hand column which has no units.
Specimen 
% Adsorbed Water (TG Data) 
Factor 
Dominican (SRM 697) 
1.44 
1.015 
Arkansas (SRM 69b) 
0.583 
1.006 
Jamaican (SRM 698) 
1.08 
1.011 
Australian (SRM 600) 
1.39 
1.014 
Corrected Value (Dry Weight Basis) = Observed Value x
Dry Weight Conversion Factors for Bauxite Specimens
As is shown in the equation, any measuredanalyte value obtained on an “as received” basis may be converted to a “dry weight basis” by simply multiplying by this factor given in table 2. One can observe that the moisture content of these bauxite specimens lead to very similar factors when compared to one another. This is usually related to the particle sizedistribution for most clay mineral specimen. However, due to the variation in composition of bauxite ores, this is not always the case for bauxites.
5.2. TGA analysis results and degree of dehydroxylation of the gibbsite component
All four of the NIST standard bauxite specimens were analyzed by TGA (and DTG) at 20^{o}C per minute in flowing nitrogen (40 cc/min). The resulting thermal curves for these heating runs are given in figures 14. The DTG thermal curves are displayed along with the TGA curves. The percentage mass losses were assigned for the gibbsite (Al(OH)_{3}) decomposition observed near 270^{o}C for all four NIST bauxite specimens. The XRD analyses of the gibbsite component was furnished for three of the four bauxite standards. These values are given in table 3 along with the TGA weight loss assignments calculated by the Pyris 7 software used with the TGA instrument. The observed mass loss values were converted to “dry weight” values, using the factors previously given in Section 5.1, prior to listing them in table 3.
Specimen 
% Gibbsite 
Observed WT Loss (Dry WT Basis) 
Theoretical % Weight Loss 
Degree of Dehydroxylation (%) 
Dominican 
50 
16.58 
17.325 
95.70 
Jamaican 
75 
24.06 
25.99 
92.61 
Arkansas 
60 
19.58 
20.79 
95.48 
Australian 
– 
18.36 
– 
– 
2Al(OH)_{3} (s) → Al_{2}O_{3} (s) + 3H_{2}O (g)
Theoretical Weight Loss for 100% Dehydroxylation
% H_{2}O =
= 34.56%
Degree of Dehydroxylation =
As will be seen in table 3, the theoretical mass loss for the complete dehydroxylation of pure Al(OH)_{3} is calculated based on the stoichiometric conversion of the gibbsite to alumina, Al_{2}O_{3} (s), and water vapor (H_{2}O (g)). The calculated percentage mass loss for 100% dehydroxylation is shown to be 34.56%. Thus, the degree of dehydroxylation is calculated, as shown in table 3, by expressing the observed experimental mass losses as a fraction of the theoretical value of 34.56%.
One will note that the degree of dehydroxylation tends to decrease somewhat as the percentage of gibbsite increases in the bauxite specimens. This is best explained by the increase in the partial pressure of water, P_{H2O}, in the TGA furnace tube with increasing amounts of water vapor arising from the decomposition of the gibbsite content. This effect will also be related to the volume of the TGA measuring chamber and the flow rate of purge gas employed in the experiment.
The important observation is that in none of these TGA analyses of the three bauxite standards did we obtain complete dehydroxylation of the gibbsite component. Furthermore, the degree of dehydroxylation would tend to vary somewhat with particle size [9], instrumental design for different instrument manufacturers and, to a lesser degree, on sample size and flow rate of the dynamic purge gas.
For these reasons, practicing thermal analysts have been advised to becautious in assigning the amount of the gibbsite component of bauxite ores, using either TGA mass loss values or DTG peak areas. As can be seen by the values given in table 3, the degree of dehydroxylation observed in our laboratory is approximately ninety five percent of the expected stoichiometric amount. Therefore, we cannot apply the theoretical gravimetric factor to quantitatively assign the gibbsite content of bauxite ores using thermogravimetric data. In the following section of this paper, an alternate approach is used to employ an experimentally determined “empirical” gravimetric factor (G) to overcome this error due to the nonstoichiometric behavior of the gibbsite dehydroxylation.
5.3. Determination of an “empirical” gravimetric factor for the dehydroxylation of gibbsite in bauxite ores
5.3.1. Analytical use of gravimetric factors: A “gravimetric factor,” G, is a multiplication factor which is used in many gravimetric methods for converting an experimental mass loss (or mass gain) to the amount of analyte present. Thus, the factor, G, has units of mass of analyte per mass of weighable form. Obviously, our weighable form is the water lost due to thermal dihydroxylation of the gibbsite. For completely stoichiometric processes, the theoretical G factor is simply calculated from the balance chemical equation. For a nonstoichiometric process, such as the thermal decomposition of Al(OH)_{3} (s) observed in the heating of bauxite ores by thermogravimetric methods, the use of such theoretical values for G will lead to incorrect values for the analyte content.
The first equation (Decomposition One) describes the incomplete decomposition of Al(OH)_{3} (s) to Al_{2}O_{3} (s) and water vapor, H_{2}O (g). As a consequence of the incomplete (or nonstoichiometric) dehydroxylation of the gibbsite, Al(OH)_{3}, a certain amount of boehmite, AlO(OH) (s)_{, }is formed. Thus, one would expect a low value for the calculation of gibbsite content if one were to apply the theoretical “G” value of this mass loss event.
5.3.2. “Empirical” gravimetric factor: To overcome this error due to nonstoichiometric, or incomplete, dehydroxylation of the gibbsite, an experimental “empirical” gravimetric factor was calculated using NIST standards of known gibbsite content. This calculation is shown in table 4 using the three NIST standards of known gibbsite content as previously assigned by XRD spectroscopy. The “empirical” G factor, having units of % gibbsite per % H_{2}O, was calculated to have a mean value of 3.06. When this small data set is treated statistically, an average deviation of ±0.043 and a standard deviation of ±0.056 for the mean is obtained.
Specimen 
% Gibbsite 
% Weight Loss (Dry WT Basis) 
G (% Gibbsite/% H_{2}O) 
Dominican 
50 
16.58 
3.02 
Jamaican 
75 
24.06 
3.12 
Arkansas 
60 
19.05 
3.03 


Mean Value of G 
3.06 
This “empirical” G factor may now be used to calculate the percentage gibbsite (by mass) in bauxite ores from the experimentally determined % H_{2}O obtained from the dehydroxylation of gibbsite when using a PerkinElmer TGA7 thermogravimetric analyzer. The instrument should be used with the high temperature (1500^{o}C) furnace and a flow rate of 40^{o}C/min of purge gas when employing the empirical G factor of 3.06 for the quantification of gibbsite content.
It should be noted, at this point, that the “theoretical” (stoichiometric) G value for the complete dehydroxylation of Al(OH)_{3} (s) is 2.89% gibbsite per % H_{2}O vapor evolved. This value is 5.56% (relative) lower than the “empirical” G value of 3.06. Since the G factor is used as a direct multiplier, a corresponding lower value would be obtained for the percentage gibbsite content if the theoretical G value were employed.
5.4. Assignment of gibbsite content of Australian Darling Range (SRM 600) bauxite specimen
Once the empirical gravimetric factor is established, for the TGA system and operating parameters (purge gas, heating rate, etc.), the analyst is then ready to make quantitative assignments of gibbsite content based on the amount (or percentage) water vapor liberated from the dehydroxylation of the analyte. The calculation is demonstrated in table 5 of this study. Since the units of the “G” factor were calculated to be in % gibbsite per % water, the calculation simply involves the multiplication of “G” and the % water vapor assigned from the TGA thermal curve. As can be seen in table 5, the 18.36% H_{2}O is converted to a gibbsite content of 56.18% when multiplied by the “G” factor of 3.06. Since the “G factor” was rounded to three significant figures, the correctly reported value to 56.2% gibbsite for this red colored Australian bauxite ore. As can be seen by visually observing the powdered samples, the Australian Darling Range Specimen (SRM 600) is not as deeply colored as the Dominican (SRM 697) specimen which contains twenty percent by mass of hematite (Fe_{2}O_{3}).
% Gibbsite = (% Liberated from Dehydroxylation) x G
For Australian SRM 600: Gibbsite = *18.36% H_{2}O x 3.06% = 56.18%
*From TG Experiment (corrected to Dry Weight Basis)
The accuracy of this assigned value of 56.2% gibbsite can be no better than the accuracy of the XRD values from which the empirical “G” value was determined.
5.5. Quantification using a single bauxite standard
The empirical G factor which was employed for the assignment of the gibbsite content of the Australian Darling Range (SRM 600) bauxite (Table 5) was obtained experimentally using three different bauxite standards. The next step in this study was to assign an empirical G factor from the use of multiple TGA analyses of a single bauxite standard. For this task, the Dominican (SRM 697) bauxite standard, which contains 50% gibbsite, was chosen.
The two student coauthors (Stong and Gann) were assigned to use the SRM 697 bauxite specimen as their reference standard. The student researchers then experimentally obtained an empirical G factor of 3.08% gibbsite/% H_{2}O for this instrument standard using the Dominican (SRM 698) bauxite standard. This calculated G value represents the mean of five TGA experimental assignments obtained from the SRM 697 standard. The percentage water of dehydroxylation, after correction to a dry weight basis, are given in table 6. The mean value of these five TGA runs was calculated to be 16.2 % H_{2}O. The average and standard deviations of this mean value were 0.146% and 0.184% respectively.
A. Determination of Experimental G Value Using STM 697 Bauxite Standard  
TGA RUN # 
% H_{2}0(Converted to Dry Weight Basis) 
1 
16.44 
2 
16.24 
3 
15.98 
4 
16.08 
5 
16.32 
Mean Value = 16.21 +/ 0.146 % H_{2}O
Standard Deviation, S = 0.184 % H_{2}O
G=50.0 % gibbsite/16.21 % H_{2}O G Value is calculated to be 3.08 % gibbsite per % H_{2}O
B. Calculation of % Gibbsite in Bauxite Unknown
% Gibbsite = G Value x Mean % H_{2}O (From TGA Runs)
The students were then given a “blind” unknown bauxite specimen to assign the gibbsite content thermogravimetrically using the G factor which they had obtained using the SRM 697 standard. The student researchers obtained a mean value from the TGA analysis of the bauxite unknown (after correcting to a dry weight basis) dehydroxylation of 19.62% H_{2}O. Figure 5 shows one of the original TGA and DTG thermal curves along with the graphical weight loss assignment. As is shown at the bottom of table 6, this corresponds to a percentage gibbsite component of 60.43%. The “blind” unknown which was issued had an XRD value for the gibbsite content of 60.0%. Thus, the gibbsite value obtained gives an absolute difference of 0.43, or a relative difference of 0.72%, from the known XRD value. Considering the fact that this thermogravimetric analysis was performed by student researchers with limited experience with the instrumental technique, it is the opinion of the senior researcher that this represents the expected accuracy for this particular application of the methodology.
The thermogravimetric techniques of TGA and DTG were employed to quantify the gibbsite content of several bauxite ore specimen. These thermogravimetric methods are not meant to replace the use of XRD spectroscopy for this purpose, but rather as a supplemental tool for the analysis of bauxites. The methodology used in this study, to quantify gibbsite using TGA/DTG, is dependent upon the use of at least one bauxite specimen of known gibbsite content as a reference standard. The value for the percentage gibbsite in the standard is usually assigned by XRD spectroscopy. This means that analytical values for the gibbsite content obtained using the thermogravimetric methodology described here cannot exceed the accuracy of the XRD value of the bauxite reference standard.
The thermogravimetric method used in this study offers an experimental means of overcoming the less than stoichiometric dehydroxylation of the gibbsite, as observed by the TGA and DTG thermal curves, when heated to temperatures greater than 300^{o}C. This quantitative assignment of the gibbsite content was accomplished through the development and subsequent use of an experimental (“empirical”) gravimetric factor rather than the usual theoretical gravimetric factor of 2.89% gibbsite/% water .
Such empirical gravimetric factors, G, must be determined prior to analysis of the bauxite ores, using the same TGA equipment and operating parameters as that to be employed in the analysis of the bauxite specimens. Due to reasons which were discussed earlier in this paper, there is no universal value of G for use in all thermogravimetric assignments of gibbsite in bauxite ores. There is, however, a minimum value of G (2.89% gibbsite per % water) which corresponds to complete, or 100%, dehydroxylation of the gibbsite contained in the bauxite ores.
It has been observed in this laboratory that the numerical value of the empirical gravimetric factor, G, will decrease with decreasing particle size, sample size, and heating rate employed in the experiment. As these three experimental factors continue to decrease, a lower limit of 2.89% gibbsite per % water evolved may eventually be reached. This corresponds to the theoretical G value and 100% dehydroxylation of the gibbsite component. In support of these observations, one should consult selected references [911] in the Bibliography.
This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright © 2019  All Rights Reserved  ocimumpublishers.com