Accepted Manuscript

Title: COMPARATIVE STUDY OF TANTALITE

DISSOLUTION USING DIFFERENT FLUORIDE SALTS
AS FLUXES

Author: M. Nete W. Purcell J.T. Nel

PII: S0022-1139(14)00146-8
DOI: http://dx.doi.org/doi:10.1016/j.jfluchem.2014.05.017
Reference: FLUOR 8329

To appear in: FLUOR

Received date: 26-3-2014

Revised date: 30-5-2014
Accepted date: 31-5-2014

Please cite this article as: M. Nete, W. Purcell, J.T. Nel, COMPARATIVE
STUDY OF TANTALITE DISSOLUTION USING DIFFERENT
FLUORIDE SALTS AS FLUXES, Journal of Fluorine Chemistry (2014),
http://dx.doi.org/10.1016/j.jfluchem.2014.05.017

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 COMPARATIVE STUDY OF TANTALITE DISSOLUTION USING DIFFERENT
FLUORIDE SALTS AS FLUXES

M. Nete1, W. Purcell1* and J.T. Nel2

t
1
Department of Chemistry, University of the Free State, Bloemfontein, 9300, South

ip
Africa

cr
2
The South African Nuclear Energy Corporation SOC Ltd. (Necsa), P.O. Box 582,
Pretoria 0001, South Africa

us
Abstract

an
Six different fluoride salts, NH4FHF, KF, KFHF, CaF2, NaF and HFNaF, were
evaluated as fluxes in the fusion dissolution of Ta/Nb containing minerals, and
compared with the commercially employed HF dissolution. The cold fusion products
M
were dissolved in dilute H2SO4 for both the dissolution and matrix matching purposes.
Analytical results indicated that the KF, KFHF and NH4FHF dissolution procedures
d

were the most effective for dissolving and recovering the main elements from the
te

tantalite mineral. CaF2, NaF and NaFHF yielded the lowest recoveries of Ta and Nb
from the tantalite mineral sample under study. This comparative study indicated that
p

the dissolution and separation of Ta and Nb from one another and from other elements
ce

in minerals depended heavily on fluoride concentration, acidity as well as temperature

for the formation of the metal-halogen complexes.
Ac

Dissolution of the minerals using KFHF and NH4FHF as fluxes are recommended as
alternatives to HF dissolution due lower energy demands, elemental recoveries, cost,
separation efficiency and ease of the dissolution step.

Keywords
Tantalite, fluoride salts, flux fusion, dissolution, analysis.

*To whom correspondence should be addressed: E-mail address: purcellw@ufs.ac.za

1
Page 1 of 24
1. Introduction
One of the most outstanding properties of Ta metal is its resistance to chemical attack.
For example, it used in the manufacturing of Cl2 containers used in large-scale CCl4
production [1]. The formation of a thin oxide layer (Ta2O5) on the surface of the Ta metal
makes it resistant to oxidation or acid attack [2]. Therefore, the chemical inertness of

t
Ta2O5 contributes to this highly valued property of Ta metal. The chemically inert metal

ip
oxides, Ta2O5 and Nb2O5, are predominantly present in many of the Ta/Nb containing

cr
mineral ores. Tantalum and niobium naturally occur together in more than 150 minerals.
The successful separation of Nb and Ta from each other and from other elements in a

us
complex mineral matrix depends primarily on the selection of appropriate sample
dissolution techniques. It is therefore not surprising that the choice or development of
effective dissolution processes for the chemical characterization and subsequent

an
separation of these chemically resistant minerals remains an interesting challenge.
M
Sample dissolution remains one of the most critical steps in the overall success of any
analytical process involving the chemical characterization of mineral samples such as
d

tantalite or its metals. The three main techniques available to dissolve these samples
are flux fusion (molten salts) [3,4,5], acid dissolution (wet ashing) [6,7] and microwave
te

digestion [5,8]. The ultimate goal using either of these methods is the conversion of the
highly inert mix of elements (always in their most stable oxidation states) in the mineral
p

to individual, water soluble chemical species (hydrated metal ions or metal complexes)
ce

[9]. The accurate quantification is the final process in the chemical characterization of
the mineral in question.
Ac

The solubilization of the elements in the minerals such as tantalite involve the chemical
attack on the strong metal-ligand bonds (normally oxides) present in the inert mineral
sample and the reaction can involve any one, or a combination of the three main types
of inorganic reactions, namely oxidation, neutralization and complexation. In the case of
wet ashing, simple protonation of the metal-oxide bonds can lead to dissolution while
acids with strong oxidizing properties (e.g. HNO3) can induce solubility by the oxidation
of the metal center to a new, more water soluble oxidation state. Flux fusion on the

2
Page 2 of 24
other hand, produces a new solvent (salt heated to temperatures above its melting
point) with either acidic, basic, oxidizing and/or complexation properties, which will
attack the metal-oxide bonds to render the elements water soluble. Microwave digestion
relies mainly on temperature and heat (changing kinetic properties) in the presence of
an acid to destroy the chemical structure of the mineral.

t
ip
Recent studies [5,10] have shown that lithium tetraborate, a combination of mono and

cr
disodium hydrogen phosphate as well as ammonium bifluoride as fluxing agents are
successful in dissolving Ta and Nb mineral. In the case of the phosphates, the formation

us
of linear polyphosphates is responsible for the successful dissolution of the mineral.
These poly phosphates are strong complexing agents and react with the metal oxides to
form soluble Nb and Ta complexes [10]. Lithium tetraborate (non-oxidizing acidic flux) is

an
successful in digesting different tantalite ore samples [5]. The dissolution, however,
requires that H2SO4 and methanol are added simultaneously after the fusion process to
M
prevent the formation of an insoluble boric acid in the reaction mixture. This step also
facilitates the removal of the excess boron as the volatile boron ester at ~ 40 oC prior to
d

sample analysis.
te

The most efficient and frequently applied dissolution method for Ta and Nb containing
mineral samples involves the use of HF as wet ashing reagent. HF is successfully used
p

on both naturally occurring (mineral ores) and artificial (metal alloys) sample dissolution,
ce

either alone or in combination with other acids at elevated temperatures. Experimental

results indicate the formation of different soluble metal fluoride complexes. These
Ac

include the possible formation of heptafluorides, hexafluorides or oxyfluorides, which is

extremely important in the subsequent separation and purification of the two metals.
Often the HF method does not completely dissolve the sample, but completely leaches
Nb and Ta from the mineral sample. Literature studies [8,11] have also shown that
tantalum and niobium dissolution can be improved by combining microwave digestion in
the presence of solvents such as H2SO4 and HF. Microwave digestion has the added
benefits of fast and uniform heating as a result of microwave energy absorption by the
sample. The improved recoveries using the combination of experimental conditions are

3
Page 3 of 24
attributed to the improved reaction rate between the reagents while the types of
complexes remain similar to those obtained from HF dissolution.

Despite its efficiency, the use of HF is discouraged in routine laboratories due to its high
corrosiveness and safety aspects [12,13]. Literature studies [13,14,15] have shown that

t
fluoride salts can be used for dissolution and fluorination of Ta and Nb in place of HF.

ip
Fluoride salts in general have an advantage of being less toxic, easy to handle and to

cr
store. However, it is important that such fluoride salts are evaluated for optimal
dissolution of a good representative of the Ta and Nb containing minerals.

us
The use of ammonium bifluoride both as a flux in a fusion technique [14] and as an
additive (fluorinating agent) in H2SO4/NH4FHF [13] tantalite dissolution were recently

an
reported. Very good Ta and Nb recoveries were reported, but some results pointed to
incomplete sample dissolution. A recent study by Nete et al. [15] demonstrated a
M
successful application of NH4FHF flux fusion procedure on various Ta/Nb containing
minerals originating from three different African countries. The possible use of other
d

fluoride containing salts such as KF, NaF and CaF2 was also suggested [13] for the
tantalite dissolution, but no experimental results have been reported. It is anticipated
te

that the high fluoride concentration depolymerizes the polymeric hydrolytic oxides in the
mineral to form monomeric fluoride complexes suitable for further analytical and
p

beneficiation procedures [16].

ce

The aim of this study was to evaluate the use of different fluoride salts, namely
Ac

NH4FHF, KF, KFHF, CaF2, NaF and NaFHF as fluxes for the dissolution of a Ta/Nb
mineral sample originating from Mozambique. Results obtained at optimum
experimental conditions for each fluoride flux were then compared with those obtained
using traditional dissolution techniques with special interest in energy consumption and
eco-friendliness.

4
Page 4 of 24
2. Results and discussion
2.1. Characteristic composition of the mineral samples
The ICP-OES results of the tantalite sample dissolved using the well-known HF/H2SO4
digestion method are given in Table 1. The results obtained in this part of the study
were used as the reference for the assessment of all the subsequent dissolution

t
procedures. These results were found to be in good agreement with those obtained by

ip
Nete et al. [5] on the same sample using the lithium tetraborate flux. However, SiO2

cr
could not be analyzed in this study due to the Si contamination caused by Si containing
components of the ICP-OES, such as the glass nebulizer.

us
Table 1: Chemical compositions of Ta and Nb mineral using HF/H2SO4 dissolution

an
Analyte Concentration (%)

Ta2O5 27.8(6)
M
Nb2O5 27.2(5)
TiO2 2.7(1)
Mn3O4 7.79(6)
d

Fe2O3 7.54(4)
Al2O3 1.9(2)
te

SnO2 1.4(3)
U3O8 2.92(2)
p

WO3 1.19(2)
ce

ThO2 0.48(6)
Ac

2.2. Comparison between different fluoride salts as fluxing agents

2.2.1. Influence of sample:flux ratio
Visual inspection indicated incomplete dissolution of the sample by all the flux fusions
at the initial 1:10 ratio. KFHF had the smallest amount of undissolved solids (~0.1%)
followed by NH4FHF (~1%). The analytical results for the digestion of the samples
using a sample:flux ratio of 1:10 are given in Table 2. Generally, lower concentrations
of metal oxides were obtained for most elements by the different fluxes at a sample:flux
ratio, 1:10 compared those obtained for the HF/H2SO4 dissolution. Good recoveries

5
Page 5 of 24
 were obtained for most of the elements using NH4FHF and KFHF compared to
HF/H2SO4, with exceptions for Ta2O5 [24.3(5)% and 26.3(6)% compared to 27.8(6)%,
respectively], ThO2 [0.3(1)% and 0.27(4)% compared to 0.48(6)%, respectively] and
U3O8 [0.9(2)% for NH4FHF compared to 2.92(2)%].

t
ip
Table 2: Quantification of elements in the dissolved portion at a sample:flux ratio of
1:10

cr
Fusion Concentration (%)
Flux temperature

us
and time Ta2O5 Nb2O5 TiO2 Mn3O4 Fe2O3 SnO2 WO3 ThO2 U3O8

250 oC, 0.5

NH4FHF 24.3(5) 28.1(6) 2.5(7) 7.7(6) 7.5(8) 1.5(3) 2.2(6) 0.3(1) 0.9(2)
h

KF 1000 oC, 1 h

1000 oC, 1 h
26.3(8)

6.7(2)
19(1)

18.0(6)
1.0(2)

3.3(8) an
2.88(9)

6.0(7)
5.2(2)

5.0(4)
0.6(1)

0.5(3)
2(1)

0.6(1)
0.7(1)

0.6(2)
3.1(2)

2.3(3)
M
NaF

NaFHF 1000oC, 1 h 1.6(4) 8.1(5) 1.2(1) 7.7(6) 6.8(4) 0.2(3) 1.3(6) 0.09(5) 2.3(3)
d

KFHF 600oC, 1 h 26.3(6) 28.7(6) 2.6(4) 8.1(5) 7.3(3) 1.2(4) 1.3(2) 0.27(1) 2.7(6)
te

HF/H2SO4 27.8(6) 27.2(5) 2.7(1) 7.79(6) 7.54(4) 1.4(3) 1.19(2) 0.48(6) 2.92(2)
p

The sample:flux ratio was changed to 1:20 and this alteration improved recoveries for
ce

most elements (Table 3). For example, Nb2O5 and Ta2O5 dissolutions by NaF increased
respectively from 18.02% to 23.68%, and 6.72% to 22.87%. Nearly complete recoveries
Ac

for most elements were obtained for KF, KFHF and NH4FHF as indicated by the
recoveries for Ta and Nb in Figure 1. Visual inspection of the samples indicated
complete sample dissolution using KFHF, but only partial to nearly complete dissolution
for NH4FHF and KF. The NaFHF fusion still produced poor recoveries with no obvious
improvement for most elements under these new set of experimental conditions.
However, there was a clear increase in Nb2O5 recovery from 8.08 to 21.41% (from 1:10
to 1:20 ratio) and Ta2O5 recovery from 1.63 to 9.60% under the same experimental
conditions for the NaFHF fusion.

6
Page 6 of 24
 Table 3: Quantification of elements in the dissolved portion at a sample:flux ratio of 1:20

Fusion Concentration (%)

Flux temperature
and time Ta2O5 Nb2O5 TiO2 Mn3O4 Fe2O3 SnO2 WO3 ThO2 U3O8

250 oC, 0.5

NH4FHF 27.3(2) 26.5(8) 2.5(3) 7.8(2) 7.4(1) 1.3(2) 1.8(2) 0.3(1) 0.9(1)

t
h

ip
KF 1000 oC, 1 h 28(1) 27.10(6) 3.6(8) 7.2(6) 7.5(3) 1.7(5) 2.1(5) 0.6(2) 1.9(1)

cr
NaF 1000 oC, 1 h 22.9(7) 23.7(9) 3.7(4) 5.9(6) 6.7(8) 0.5(1) 0.8(2) 0.8(3) 2.6(2)

1000oC, 1 h 9.6(3) 21.4(6) 2.1(6) 7.5(3) 6.6(2) 0.3(1) 1.3(5) 0.2(1) 2.5(4)

us
NaFHF

KFHF 600oC, 1 h 26.9(3) 28.76(6) 2.61(2) 7.9(3) 7.3(4) 1.3(2) 1.21(9) 0.29(3) 2.7(2)

an
HF/H2SO4 27.8(6) 27.2(5) 2.7(1) 7.79(6) 7.54(4) 1.4(3) 1.19(2) 0.48(6) 2.92(2)
M
Ta2O5
Nb2O5
d
p te
ce
Ac

NH
NH44FHF
F HF/H
HF/H2SO
2SO44

Figure 1: Quantitative analyses of Ta2O5 and Nb2O5 in tantalite sample dissolved by

HF/H2SO4 and flux fusion using five different fluoride fluxes at sample:flux ratio of 1:20.

The sample:flux ratio was increased again to 1:30 for those fluoride fusions (NaF and
HFNaF) that produced poor elemental recoveries for most of the elements in the
sample (Table 4). These results indicated no significant change in the dissolution of
many of the elements by these salts. The Al2O3 and SiO2 results for all the fluxes at all

7
Page 7 of 24
 three different flux concentrations were highly inaccurate and inconsistent and as such
were not reported. The inconsistency with regard to Al2O3 results is not clear, but the
literature [16] indicates that Al analysis is affected by the residual fluoride in solution.
The SiO2 results on the other hand could be affected by the fluoride reaction leaching
glass from the equipment components such as the ICP nebulizer (example of SiO2

t
reaction with NH4FHF is given in Equation 1) [17].

ip
cr
SiO2(s) + 4NH4FHF(aq) (NH4)2SiF6(aq) + 2NH4F(aq) + 2H2O(g) 1

us
Table 4: Quantification of elements in the dissolved portion at a sample:flux ratio of 1:30

Concentration (%)

an
Fusion
Flux temperature
and time Ta2O5 Nb2O5 TiO2 Mn3O4 Fe2O3 SnO2 WO3 ThO2 U3O8
M
NaF 1000 oC, 1 h 22.3(9) 22.7(5) 1.7(4) 7.0(5) 7.2(3) <0.01 2.5(8) 0.6(1) 2.5(6)

NaFHF 1000 oC, 1 h 20.8(3) 25.0(1) 2.4(6) 7.8(2) 7.0(9) 0.5(6) 1.4(5) 0.2(4) 2.6(1)
d

HF/H2SO4 27.8(6) 27.2(5) 2.7(1) 7.79(6) 7.54(4) 1.4(3) 1.19(2) 0.48(6) 2.92(2)
te

2.2.2. Efficiency of different fluoride fluxes on tantalite dissolution

p

The NaF and NaFHF fusions produced lower than expected elemental percentage
recoveries and generally showed similar behavior in dissolving the tantalite mineral.
ce

Elemental recoveries improved significantly with an increase in sample:flux ratio from

1:10 to 1:20. The elemental recoveries for NaFHF increased further with an increase in
Ac

sample to flux ratio (1:30 sample:flux), but the same tendency was not observed for
NaF. However, even this increase in sample to flux ratio for NaFHF recovered Ta2O5
and Nb2O5 well below their expected values (Table 1). This can be explained in terms of
the decomposition of NaFHF at elevated temperatures [18] as indicated in Equations 2
and 3. Reactions using KF, NaF and NaFHF fluxes were performed at 1000 oC, which
also makes these three salts as potential fluxes unpopular due to their high energy
requirements.

8
Page 8 of 24
 160 oC
NaFHF(s) NaF(s) + HF(s) 2

>160 oC
NaF(s) + HF(s) Na+(l) + HF2-(l) 3

The NH4FHF, KF and KFHF fusions produced good results for most elements at a

t
sample:flux ratio of 1:20. All the major elements values, particularly tantalum and

ip
niobium, were in good agreement with those obtained for the HF digestion. The U3O8

cr
values of 0.9(1)% for NH4FHF, 1.9(1)% for KF, and 2.7(2)% for KFHF were, however,
still lower than the expected value of 2.92(2)% (Table 1). The high fusion temperature

us
for KF makes this fluoride dissolution procedure a less desirable method. KFHF and
NH4FHF fusion methods on the other hand, were performed at relatively low fusion
temperatures of 600 and 250 oC, respectively. The KFHF has an advantage of

an
completely dissolving the tantalite sample even though the fusion temperature is higher
than that of NH4FHF. The NH4FHF fusion had the added advantages of shorter fusion
M
times (0.5 h) and the subsequent dissolution of the melt using only water. The benefit of
using water only is the simplified matrix matching step and acidity adjustment of the
d

solution. This is particularly important in further processing of the sample with well-
established separation techniques such as solvent extraction and anion exchange
te

[19,20,21,22].
p

One major disadvantage of the fluoride fusion methods is that fluoride resistant
ce

equipment is required for every step of the process. The high fluoride concentration
chemically attacks the surfaces of the different glass components (Equation 1), which
Ac

results in poor analysis (recoveries) of elements such as Si and B. For analytical

determinations using ICP-OES or ICP-MS, the fluoride solutions also have to be
extremely dilute to prevent any damage to the ICP glass components such as the
plasma torch. This requirement is anticipated to complicate the analyses of any trace
elements present in the mineral, which may result in some of the elements being
analyzed near or below the limit of detection (as a result of the large dilution factor) due
to a large concentration variation between the elements. CaF2 was also evaluated as

9
Page 9 of 24
potential fluxing agent. However its high melting point (1418 oC) resulted in poor mineral
dissolution and therefore also poor elemental recovery (results not shown).

The results obtained in this comparative study give insight into the dissolution
chemistry/mechanism for the two main elements present in the tantalite mineral. The

t
successful dissolution of Ta2O5 is definitely not a function of acid strength. The study

ip
performed by Theron et al. [23] indicated extremely poor Ta2O5 recoveries using

cr
concentrated strong acids such as H2SO4 (9.7%), H3PO4 (7%), HNO3 (4.2%) and aqua
regia (4%). Even the use of these acids in a microwave digester did not improve the

us
solubility of metal oxide, confirming its strong Lewis acidic properties. However, the
weaker HF acid succeeds, without the application of drastic experimental conditions, in
the complete dissolution of Ta2O5. The presence of the small sigma and pi donor ligand

an
F- seems to be the key to the complete dissolution using this weak acid. Complete
dissolution was also obtained for the bi-salts NH4FHF and KFHF, which also contain
M
an acid proton as well high [F-]. It seems that the key to complete Ta2O5 dissolution is
the presence of some acid protons to initialize metal-oxide bond cleavage (equilibrium
d

lies to left hand side of Reaction A in Scheme 1), which is then followed by metal-
fluoride complex formation (Reaction B) to form the very soluble tantalum fluoride
te

complexes. The depletion in the concentration of the intermediate obtained after the
protonation by its reaction with the fluoride ions ensure the continued formation of the
p

intermediate (re-establishment of the equilibrium) until all the metal oxide has been
ce

dissolved.
Ac

However, the results obtained in this study also indicate that complete dissolution can
be obtained in the absence of acid protons (Reaction C), but only at much higher
fluoride concentrations (increase in sample:flux ratio) and temperature (melting points of
salts such as NaF and HFNaF). It is anticipated that the decomposition of NaFHF at
high temperature to yield only NaF in the flux and no acid protons may be the reason for
its poorer dissolution ability of Ta2O5.

10
Page 10 of 24
 A
Ta2O5(s) + H+(aq) [Ta2O(OH)]*(aq)

C B 7F-(aq)
7F-(l)
TaF72-(aq)

t
ip
Scheme 1: Mechanistic presentation of the Ta2O5 dissolution using H+ and F- ions

cr
Different studies have suggested different dissolution properties for Nb2O5 [5,23,24].

us
Nete et al. [5] obtained excellent Nb2O5 recoveries in both pure inorganic salt as well
mineral using concentrated H2SO4 (87%) and H3PO4 (84%) in a microwave digester
confirming its superior Lewis basicity compared to Ta2O5. The complete dissolution of

an
Nb2O5 using the two mineral acids suggest that the protonation reaction forces the
equilibrium to right hand side (Reaction A, Scheme 2), producing substantial amounts
M
of the niobium hydrated species in solution, which then reacts with the fluoride anions
finally form the highly soluble fluoride complex (Reaction B). The same tendency of
d

product formation without protons was observed for Nb2O5. Interestingly, the Nb
recoveries were all much better compared to the Ta recovery under the same
te

experimental conditions, suggesting an equilibrium, which also lies to the right hand
side (Reaction C).
p
ce

A
Nb2O5(s) + H+(aq) [Nb2O(OH)]*(aq)

C B
Ac

5F-(aq)
5F-(l)
NbOF52-(aq)

Scheme 2: Mechanistic presentation of the Nb2O5 dissolution using H+ and F- ions

11
Page 11 of 24
2.2.3. Effect of temperature on element recovery using NH4FHF
The effect of lower temperatures with sample:flux ratio = 1:20 for 0.5 h on the digestion
and dissolution of the tantalite mineral ore using NH4FHF is given in Table 5. Results
indicated that fusion at 185 oC produced the lowest percentage metal oxides recoveries
for all the analytes suggesting that 185 oC and 0.5 h digestion time is insufficient for the

t
complete recovery of the elements in tantalite mineral. There was no marked

ip
improvement with an increase in temperature from 200 oC to 250 oC, but visual
inspection indicated that the amount of residue decreased significantly at the 250 oC

cr
fusion temperature suggesting a more complete reaction.

us
Table 5: Quantification of elements in the dissolved portion of as a function of

an
temperature with sample:flux ratio = 1:20 for 0.5 h

Concentration (%)
Temperature
(oC)
M
Ta2O5 Nb2O5 TiO2 Mn3O4 Fe2O3 SnO2 WO3 ThO2 U3O8

185 20.9(5) 24.5(6) 2.2(2) 6.5(1) 6.7(3) 1.0(1) 1.6(8) 0.3(1) 1.0(3)

200 28.1(2) 27.2(6) 2.4(3) 7.6(5) 7.4(1) 1.3(3) 2.0(3) 0.3(1) 0.7(1)
d

250 27.9(8) 26.6(9) 2.4(2) 7.8(5) 7.4(8) 1.2(1) 1.7(2) 0.3(1) 0.8(2)
te

HF/H2SO4 27.8(6) 27.2(5) 2.7(1) 7.79(6) 7.54(4) 1.4(3) 1.19(2) 0.48(6) 2.92(2)
p

2.2.4. Chemical characterization of the residue obtained after NH4FHF fusion and
ce

dissolution with water

The analysis of the residue obtained after the dissolution of the melt obtained after the
NH4FHF digestion (sample:flux ratio of 1:20) at 250 oC for 0.5 h was re-dissolved using
Ac

a different flux (1:1 mixture of Na2HPO4 and NaH2PO4H2O [10]) indicated the presence
of Sn, Ta, Mn, Fe and Ti. However, a quantitative analysis was done for all the elements
to confirm their presence or absence in the residue since the detection of some of the
elements such as U by ICP-OES might have been poor (due to high detection limits) at
the low concentrations in which U was present in the sample [25]. The results (Table 6)
indicated that the residue contained 1.25% and 1.86% of U3O8 and Mn3O4 respectively,
as the only significant amounts of the metal oxides in the residue. Most of the other

12
Page 12 of 24
metal oxides were successfully converted into their soluble fluoride salts as confirmed
by the concentrations in the residue (e.g. 0.06% Nb2O5 and 0.57% Ta2O5).

Table 6: Quantification of elements in the residual portion of a mineral sample after the
dissolution with NH4FHF

t
ip
(Concentration) %
Sample
TiO2 Mn3O4 Fe2O3 Nb2O5 SnO2 Ta2O5 WO3 ThO2 U3O8

cr
Residue 0.3(6) 1.2(9) 0.6(2) 0.06(5) 0.4(6) 0.6(5) 0.02(3) 0.4(7) 1.3(8)

us
2.3. Recommended fluoride flux fusion dissolution methods
KF, KFHF and NH4FHF procedures were the most effective as fluxes in the fusion

an
dissolution of tantalite. KFHF was the most efficient flux for U recovery and equal to
NH4FHF in dissolving the rest of the elements in this mineral. Both KFHF and
NH4FHF had an economic advantage due to their lower reaction temperatures and
M
hence the least energy demanding processes. However, the NH4FHF was among the
least efficient fluxes in recovering U, while all six fluoride fluxes investigated in this study
d

produced inaccurate Al analyses. In summary, the KFHF and NH4FHF fusion

procedures are recommended for digestion of Ta/Nb minerals and accurate
te

determination of the elements therein, particularly when there is a need for further
p

processing of the ore. However, the fluoride fusion procedure is not suitable for analysis
of Si especially when glass equipment is used during the processing and analysis of the
ce

samples.
Ac

2.3.1. Statistical comparison of the results

No significant differences (F Fcritical = 4.066 and p 0.05) existed between four
digestion methods in terms of the concentrations of Ta (F = 1.986, p = 0.19), Nb (F =
1.596, p = 0.27), Mn (F = 1.926, p = 0.20) and Fe (F = 2.476, p =0.14). A significant
difference was observed for the Ti (F= 33.268, p = 7.00x10-5) results. Even the larger
deviation obtained for Ti is not surprising based on the relative complexity of the mineral
sample and that small variations in homogeneity and particle size can influence the
recovery and therefor the outcome of the ANOVA result.

13
Page 13 of 24
3. Conclusion
Among the six fluxes evaluated in this study the KFHF flux fusion procedure was the
most effective in the total dissolution of the tantalite sample and recoveries of elements
therein. The KF flux fusion indicated partial dissolution of the mineral samples, but with
sufficiently comparable results for most of the elements including Ta and Nb, with

t
those obtained using the HF technique. NaF and NaFHF on the other hand, showed

ip
partial dissolution and poor results for most of the elements in the tantalite sample.

cr
Ammonium bifluoride also indicated incomplete sample dissolution, but good
agreement with results of HF as well as literature values. NH4FHF dissolution

us
procedure surpassed all the other digestion methods, which were investigated in this
study in terms of the cost effectiveness and energy demand. The KFHF and NH4FHF
dissolution procedures are therefore recommended for commercial processing of Ta
and Nb minerals and as alternative to HF.

an
M
4. Experimental
4.1. Equipment and reagents
d

A Shimadzu AW220 analytical balance was used for accurate weighing of samples
while the flux fusions were performed in a Thermo Scientific Thermolyne Compact
te

Benchtop Muffle Furnace. Grade B glass and polytetrafluoroethylene (PTFE) volumetric

flasks bought from Merck and Boeco SP series adjustable volume pipettes were used
p

for sample preparations. The PTFE flasks were used for all fluoride containing solutions.
ce

A Shimadzu ICPS-7510 ICP-OES sequential plasma spectrometer was used for the
elemental analysis of the sample solutions. KFHF, KF, as well as HF (48%) were
Ac

bought from Sigma Aldrich. NH4FHF, CaF2 and NaF were bought from Merck while
NaFHF was sourced Harshaw Chemical Co. ICP standard solutions containing 1000
mg/L Ta, Nb, Ti, Sn, W, Si as well as multi-element standard No. XXVI containing 1000
mg/L each of Mn, Al, Fe, and Ca was also bought from Merck. ICP standard solutions
containing 1000 mg/L Th and U were bought from De Bruyn Spectroscopic. Analytical
grade H2SO4 (97%) was bought from Associated Chemical Enterprises. Double distilled
water was used in all cases. The results in this study are reported as metal oxide
(content) for the most stable oxidation state of each element (for comparison purposes)

14
Page 14 of 24
and as an average of three replicate analyses. The reporting of the results was also
based on the standard deviation to indicate the uncertainty in the last digit of the
average value.

The tantalum/niobium-bearing mineral sample was supplied by the South African

t
Nuclear Energy Corporation SOC Ltd. (Necsa) as crushed rocks with particle sizes

ip
ranging from 0.1 to 5 mm and was subsequently ground to a fine powder using a

cr
vibratory disc mill in closed steel milling vessels. The mineral sample originated from the
Naquissupa mine in the north-west region of Alto Ligonha of the Zambesia province,

us
Mozambique. The tantalum and niobium bearing mineral deposits in this area are
deposited in the form of granitic rare-metal pegmatites and these are mined at Morrua,
Muiane and Marropino [26,27].

an
The mineral sample was characterized in previous studies [5,28] as predominantly
M
mangano tantalite with minor and accessory minerals such as quartz, and muscovite.
The elemental analyses of the sample indicated the presence of tantalum in
d

concentrations of 27.78(3)% Ta2O5 and niobium concentrations of 27.0(5)% Nb2O5 [5].

te

4.2. Mineral dissolution and ICP-OES analysis

In order to compare the efficiency of different fluoride salts to dissolve the tantalite, the
p

sample was initially dissolved using the well-known HF/H2SO4 digestion method and
ce

analyzed using ICP-OES. The next step was to evaluate the effectiveness of six
different fluoride salts (NH4FHF, KF, KFHF, CaF2, NaFHF and NaF) as fluxes for the
Ac

successful dissolution of the tantalite mineral. The optimum experimental conditions for
the dissolution of tantalite mineral were investigated and included fusion temperature,
ratio of solute to solvent as well as fusion time. The study was started with the
sample:flux ratio of 1:10 and at a fusion temperature above the known melting point of
the flux salt. The melting and boiling points [29] of the fluoride salts investigated in this
study are listed in Table 7. It is important to note that some salts such as NaFHF
decomposes before reaching its melting point (for example NaFHF decomposes to
form NaF and HF at 160 oC [30]. Therefore, the high temperature properties of NaFHF

15
Page 15 of 24
would be similar to those of the dissociated compounds namely NaF and HF and would
behave very similar to the single fluouride salt flux. Other experimental parameters such
as the fusion temperature were investigated where possible for some fluxes such as
NH4FHF.

t
Table 7: Melting and boiling properties of selected fluoride fluxes

ip
Salt Melting point (oC) Boiling point (oC)
NaF 996 1704

cr
KF 858 1502
NaFHF >160 decomposes
KFHF 239 decomposes

us
NH4FHF 125 240

4.2.1. Preparation of ICP-OES calibration solutions and measurements

an
Standard solutions for ICP-OES analysis were prepared by adding the appropriate
volumes of the 1000 ppm stock solutions to 10.0 mL 97% H2SO4 in 100.0 mL volumetric
flasks to prepare 1.0, 3.0, 5.0, 10.0, and 20.0 ppm concentrations. The blank solutions
M
were prepared by diluting 10.0 mL H2SO4 with water. The appropriate amount of flux
(fused) solution was added to the standard and the blank solutions to match the
d

matrices in the samples and the subsequent solutions were diluted to the 100.0 mL
te

mark of the volumetric flask with double distilled water. ICP-OES analyses were
performed at wavelengths of 309.418 nm for Nb and 240.068 nm for Ta. The detection
p

limits were determined as 0.0094 and 0.024 ppm for Nb and Ta respectively [11]. The
ce

ICP-OES conditions indicated in Table 8 were used for quantitative analyses.

Table 8: Experimental operating conditions for ICP-OES

Ac

Parameter Value
RF power 1.2 kW
Coolant gas flow 14.0 L/min
Plasma gas flow 1.2 L/min
Carrier gas flow 1.0 L/min
Sample uptake method Peristaltic pump
Glass cyclonic spray chamber with concentric
Spray chamber
nebuliser

16
Page 16 of 24
4.2.2. HF dissolution of tantalite samples
Approximately 0.1 g of the mineral sample was accurately weighed (to 0.1 mg) in a
platinum crucible. A 2 mL 48% HF + 8 mL 97% H2SO4 + 10 mL water mixture was
added and the crucible contents were subsequently heated with constant stirring for 1 hr
at 110 oC. Visual inspection indicated that this digestion procedure dissolved ~99% of

t
the ore sample. The solution was allowed to cool to the surrounding temperature. After

ip
cooling, the solids were separated from the solution by filtration and the filtrate was

cr
quantitatively transferred to a 100.0 mL PTFE volumetric flask and filled to the mark with
double distilled water. A 10.0 mL portion of the solution was pipetted into another

us
volumetric flask and diluted to 100.0 mL with double distilled water. The dilute solution
was analyzed using ICP-OES.

4.2.3. NaF fusion dissolution

an
Portions of 0.1 g (weighed to 0.1 mg accuracy) of the mineral sample were thoroughly
M
mixed with NaF (m.p. = 992 oC) in a specific sample:flux ratio in platinum crucibles.
Mixtures were fused at 1000 oC for 1 h in a high temperature oven. To obtain optimum
d

dissolutions using this flux, sample:flux ratios were investigated at a range of 1:10 to
1:30. A white turbid solid, which solidified into a white amorphous solid upon cooling,
te

was obtained. The cold melt was dissolved in 2.0 M H2SO4. Visual inspection indicated
that some sample remained undissolved. The residue was separated from the solution
p

by filtration. The filtrate was quantitatively transferred to 100.0 mL PTFE volumetric flask
ce

and then filled to the mark with distilled water. A 5.0 mL aliquot of the solution was then
transferred to another 100.0 mL volumetric flask. The acidity adjustment and dilutions of
Ac

the solutions were performed to match the blank and the standard solutions. The final
solutions were analyzed using ICP-OES.

4.2.4. NaFHF fusion dissolution

Portions of 0.1 g (weighed to 0.1 mg accuracy) of the mineral sample were
o
homogenously mixed with NaFHF (decomposition point = 160 C) in a specific
sample:flux ratio in platinum crucibles. Mixtures were fused at 1000 oC for 1 h in a high
temperature oven. To obtain optimum dissolutions using this flux, sample:flux ratios

17
Page 17 of 24
were investigated at a range of 1:10 to 1:30. The cold melt was cooled to the
surrounding temperature to form a glassy solid, which was subsequently dissolved in
2.0 M H2SO4. Visual inspection indicated that some sample remained undissolved. The
residue was separated from the solution by filtration. The filtrate was prepared for ICP-
OES analysis as described in Section 4.2.3.

t
ip
4.2.5. KF fusion dissolution

cr
Portions of 0.1 g (weighed to 0.1 mg accuracy) of the mineral sample were thoroughly
mixed with KF (m.p. = 857 oC) in a specific sample:flux ratio in platinum crucibles.

us
Mixtures were fused at 1000 oC for 1 h in a high temperature oven. A clear melt, which
solidified into a white amorphous solid upon cooling, was obtained. The cold melt was
dissolved in 2.0 M H2SO4. Visual inspection indicated that some sample had remained

an
undissolved. To obtain optimum dissolutions using KF flux, sample:flux ratios were
investigated for a range of 1:10 to 1:20. The residue was separated from the solution by
M
filtration. The filtrate was prepared for ICP-OES analysis as described in Section 4.2.3.
d

4.2.6. KFHF fusion dissolution

Portions of 0.1 g (weighed to 0.1 mg accuracy) of the mineral sample were thoroughly
te

mixed with KFHF (m.p. = 239 oC) in a specific sample:flux ratio in platinum crucibles.
Mixtures were fused at 600 oC for 1 h in a high temperature oven. A clear melt, which
p

solidified into a white amorphous solid upon cooling, was obtained. The cold melt was
ce

dissolved in 2.0 M H2SO4. Visual inspection indicated that almost the entire sample had
dissolved. Analytical results indicated incomplete recoveries for main elements a
Ac

sample:flux ratio of 1:10. To obtain optimum elemental recoveries using KFHF flux,
sample:flux ratio was changed to 1:20. The solutions were prepared for ICP-OES
analysis as described in Section 4.2.3.

4.2.7. NH4FHF fusion dissolution of tantalite

Portions of 0.1 g (weighed to 0.1 mg accuracy) of the mineral sample were thoroughly
mixed with 2 g NH4FHF flux (m.p. = 125 oC) in platinum crucibles. The mixture was
fused at 250 oC for 0.5 h in an oven. The melt was cooled to room temperature and

18
Page 18 of 24
dissolved in double distilled water. The undissolved solids were separated from the
solution by filtration. The filtrate was quantitatively transferred to a 100.0 mL PTFE
volumetric flask and diluted to the mark with double distilled water. The filtrated solution
was diluted 20 times and analyzed using ICP-OES. This procedure was repeated during
the investigation of lower fusion temperatures (185 and 200 oC).

t
ip
The undissolved portion of the sample (from NH4FHF flux fusion and dissolution) was

cr
quantitatively transferred to a pre-weighed platinum crucible by washing the sample
from the filter paper with some water. The water was evaporated and the samples were

us
dried at 100 oC for 5 h. The crucible was re-weighed to determine the sample mass.
The residual sample was dissolved using the phosphate flux [10].

4.2.8. Statistical comparison of the results

an
The level of differences between concentrations of the main elements (Ta, Nb, Mn, Fe
M
and Ti) obtained by HF/H2SO4 procedure and those obtained by KF, KFHF and
NH4FHF dissolution methods were analyzed by single factor analysis of variance
d

(ANOVA) [31]. Statistical significance was defined as F Fcritical = 4.066 and p 0.05.
te

Acknowledgements
p

The authors thank the Research Fund of the University of the Free State, the National
Research Foundation of South Africa, Necsa and the New Metals Development
ce

Network (NMDN) of the Advanced Metals Initiative (AMI) of the Department of Science
and Technology of South Africa (DST) for financial support.
Ac

19
Page 19 of 24
References
[1] D. Gussack, M. Gussack, Symposium 2012 on Tantalum and Niobium, Cape Town
(2012)
[2] K.M. Mackay, R.A. Mackay, W. Henderson, Introduction to Modern Inorganic

t
ip
Chemistry, fifth ed., Stanley Thornes (Publishers) Ltd., Cheltenham, 1996
[3] J.E.S. Uria, C.G. Ortiz, A.M. Garcia, A. Sanz-Medel, Mikrochim. Acta 92 (1987) 195

cr
202
[4] W. Kck, P. Paschen, JOM 41 (1989) 3339

us
[5] M. Nete, W. Purcell, E. Snyders, J.T. Nel, G. Beukes, J. S. Afr. Inst. Min. Metall. 112
(2012) 10791086

an
[6] H. Majima, Y. Awakura, M. Mashima, T. Hirato, Metall. Trans. B 19B (1988) 355363
[7] H.H. Htwe, K.T. Lwin, World Acad. Sci. Eng. Technol. 22 (2008) 125128
[8] R.A. Conte, J-m. Mermet, J.D.A. Rodrigues, J-L. Di Martino, J. Anal. At. Spectrom.
M
12 (1997) 12151220
[9] A. Agulyansky, The Chemistry of Tantalum and Niobium Fluoride Compounds,
d

Elsevier B.V., Amsterdam, 2004

te

[10] P.L. Mahanta, V.V. Hanuman, R. Radhamani, P.K. Srivastava, At. Spectrosc. 29
(2008) 172179
p

[11] M. Nete, W. Purcell, E. Snyders, J.T. Nel, S. Afr. J. Chem. 63 (2010) 130134
[12] Z-Y. Hseu, Bioresour. Technol. 95 (2004) 5359
ce

[13] R.P.S. Gaur, R.G. Mendenhall, United States Osram Sylvania Inc. (Danvers, MA)
Patent US 7182925 B2. 2007 Feb 27
Ac

[14] M.J. Kabangu, P.L. Crouse, Hydrometall. 129130 (2012) 151155

[15] M. Nete, W. Purcell, J.T. Nel, Proceedings of the 2013 Conference of Advanced
Metals Initiative. Precious Metals 2013, Cape Town (2013), pp 2126
[16] Multi-Agency Radiological Laboratory Analytical Protocols Manual, (2004),
http://www.epa.gov/safewater/radionuclides/training/resources/MARLAP_13.pdf
[17] A.R. Timokhin, L.A. Komarova, Glass Ceram. 42 (1985) 267269
[18] G.L. Flemmert, Inventor; Flemmert Goesta Lennart, assignee. Patent US 3551098
A. 1970 Dec 29

20
Page 20 of 24
[19] A.D. Damodaran, S.G. Deshpande, A.A. Majmudar, M.S. Sastri, P.K. Jena, D.K.
Bose, S. Chandra, C.K. Gupta, A.K. Taneja, Bhabha Atomic Research Center,
Trombay, Bombay, 36 (1969) 306318
[20] V.G. Maiorov, A.I. Nikolaev, V.K. Kopkov, Russ. J. Appl. Chem. 75 (2002) 1389
1393

t
[21] A. Agulyansky, L. Agulyansky, V.F. Travkin, Chem. Eng. Process. 43 (2004) 1231

ip
1237

cr
[22] O.M. El Hussaini, N.M. Rice, Hydrometall. 72 (2004) 259267
[23] T.A. Theron, M. Nete, J.A. Venter, W. Purcell, J.T. Nel, S. Afr. J. Chem. 64 (2011)

us
173178
[24] O.M. El-Hussaini, M.A. Mahdy, Hydrometall. 64 (2002) 219229
[25] J. Schibler, D. Moore, B. De Borba, PITTCON 2007 Presentation, Setting

an
Meaningful Detection and Quantitation Limits for Chromatography Methods (2007),
Dionex Corporation, Sunnyvale, CA, USA, http://www.dionex.com/en-
M
us/webdocs/52823-LPN-1926-Chromeleon.pdf
[26] R. Shaw, K. Goodenough, G. Gunn, T. Brown, D. Rayner, Niobium tantalum,
d

(2011), British Geologica Survey, http://tanb.org/tantalum

[27] Roskill Information, The Economics of Niobium, tenth ed., Roskill Information
te

Services Ltd., London, 2005

[28] M. Nete, Dissolution and analytical characterization of tantalite ore, niobium metal
p

and other niobium compounds, MSc thesis, University of the Free State, Bloemfontein,
ce

South Africa (2009)

[29] D.R Lide, Ed., CRC Handbook of Chemistry and Physics, eighty-ninth ed., CRC
Ac

Press, Boca Raton, FL, 20082009

[30] Sodium Hydrogendifluoride (Sodium Bifluoride), Technical Data Sheet, (2005),
http://www.solvaychemicals.us/SiteCollectionDocuments/tds/SBF.pdf
[31] D.A. Skoog, J.F. Holler, D.M. West, S.R. Crouch, Fundamentals of Analytical
Chemistry, eighth ed., Thomson Brooks/Cole, Australia, 2004

21
Page 21 of 24
 Ta2O5
Nb2O5

t
ip
cr
us
an
NH FHF
NH4 FHF HF/H
HF/H2SO
2SO44
4

Quantitative analyses of Ta2O5 and Nb2O5 in tantalite sample dissolved by HF/H2SO4

M
and flux fusion using five different fluoride fluxes at sample:flux ratio of 1:20.
d
p te
ce
Ac

22
Page 22 of 24
Graphical Abstract
Six different inorganic fluoride salts, NH4FHF, KF, KFHF, CaF2, NaF and NaFHF,
were evaluated as flux-fusion reagents for dissolution of tantalite. Results indicated that
quantitative recoveries were obtained for NH4FHF, KF, KFHF.

t
ip
cr
us
an
M
d
p te
ce
Ac

23
Page 23 of 24
Highlights

Fluoride salts were compared as fluxing reagents for tantalite dissolution

KFHF, KF and NH4FHF more efficient than NaF and HFNaF in tantalite dissolution

Nb2O5 dissolved more readily in the acidic fluoride matrix compared to Ta2O5

t
Both H+ and F- are important for complete dissolution of Nb2O5 and Ta2O5

ip
cr
us
an
M
d
p te
ce
Ac

24
Page 24 of 24