You are on page 1of 15

Minerals Engineering 171 (2021) 107067

Contents lists available at ScienceDirect

Minerals Engineering
journal homepage: www.elsevier.com/locate/mineng

The interaction of flotation reagents with metal ions in mineral surfaces: A


perspective from coordination chemistry
Jianhua Chen
School of Resources, Environment and Materials, Guangxi University, Nanning 530004, China

A R T I C L E I N F O A B S T R A C T

Keywords: The selective interaction between reagents and mineral surfaces is the core basis of mineral flotation separation.
Mineral surface This paper creatively proposes a perspective from coordination chemistry to clarify the interaction mechanism of
Reagent reagents with mineral surfaces systematically and profoundly. The metal ion in mineral surface is far different
Coordination chemistry
from the free ion. The former is in a semi-constrained state, causing the properties of surface metal ions to be
π-backbonding
greatly affected by surface structures and properties of surrounding atoms. Based on coordination chemistry, the
π-backbonding model is advanced for the interaction between sulfhydryl collectors and sulphide minerals. It is of
interest that with more π electron pairs, the surface metal ion is more likely to interact with sulfhydryl collectors
containing unoccupied π orbitals; with greater polarizability, the metal ion is more prone to covalent interactions
with sulfhydryl collectors. In addition, the unoccupied orbitals play a crucial role in selectivity of depressants.
For example, the depressants NaCN and Ca(OH)+ containing unoccupied π orbitals can strongly depress pyrite
holding π electron pairs, but can hardly depress galena possessing no π electron pair. Furthermore, the crystal
field stabilization energy resulting from the interaction between reagents and surface metal ions can influence
the stability of reagent adsorption, and can adequately explain the order of flotation critical pH for sulphide
minerals. The coordination theory sheds new light on the interaction mechanism between flotation reagents and
mineral surfaces.

1. Introduction (2007) said, “Without reagents, there would be no flotation. Without


flotation, the mining industry as we know it today would not exist.”. The
Flotation is a great invention in the utilization of mineral resources. interaction mechanism of flotation reagents with minerals has attracted
However, it was not until 1925 when xanthate could be used as a high- considerable attention of scholars worldwide.
efficiency collector for sulphide ore (Keller et al., 1925) that mankind Gaudin (1957) considered the mechanism of mineral collection as
began to utilize mineral resources on a large scale, especially complex the central problem of flotation theory. As early as the period of
ores and low-grade polymetallic mineral resources. At present, it is 1926–1929, Gaudin and his colleagues carried out systematic in­
estimated that 4 billion tons of copper ore and up to 10 billion tons of vestigations on how reagents worked in the flotation of pure minerals
materials are processed by flotation industry each year. Mineral flota­ (Gaudin and Martin, 1928). He proposed that the mechanism of xan­
tion has become one of the most significant technologies for the recovery thates floating sulphide minerals instead of galena might involve the
and utilization of mineral resources. Flotation is used to recover most of adsorption of xanthate without further reactions (Gaudin, 1928). Wark
the non-ferrous metals and associated rare and precious metals, such as (1994) suggested that xanthate was adsorbed on mineral surface by
gold, silver, rhenium and indium In the flotation process, the hydro­ exchanging attached ions (hydroxyl and oxidized sulphur ions). The
philic and hydrophobic differences of various minerals are enhanced by critical pH curves proposed by Wark and Cox (1934a) played an
the selective adsorption of flotation reagents according to the physical important role in the early flotation theory, which illustrated the rela­
and chemical properties of the mineral surfaces, so that the target tionship between collector concentration and pH for incipient flotation.
minerals are selectively separated from the gangue minerals. The se­ During the period of 1953–1956, Fuerstenau began to develop the
lective interaction of flotation reagents with mineral surfaces does result electrostatic model of flotation (Fuerstenau, 1957, 1962). The basic idea
in the high efficiency and selectivity of mineral flotation. As Bulatovic was that when the charge of mineral surfaces was negative (positive),

E-mail address: jhchen@gxu.edu.cn.

https://doi.org/10.1016/j.mineng.2021.107067
Received 22 February 2021; Received in revised form 2 July 2021; Accepted 4 July 2021
0892-6875/© 2021 Elsevier Ltd. All rights reserved.
J. Chen Minerals Engineering 171 (2021) 107067

the cationic collectors (anionic collectors) would be adsorbed. Never­ ions in the flotation system. The interaction between reagents and sur­
theless, the negatively charged short-chain collector ions can adsorb on faces is essentially hetero-coordination, that is, the reagent molecule
a negatively charged surface by chemical interaction. For example, coordinates with the metal ion that has already been bound to sur­
xanthate anion can still adsorb on sulphide mineral surface, charged rounding atoms. Therefore, the reactivity of metal ions is determined by
negatively, in strong alkaline pulp. the surface structure of minerals and the properties of coordinated
However, Cook and Nixon (1950) disagreed with the adsorption atoms. Pradip (2003) believed that there is “structural/stereochemical
mechanism, and ascribed the process of adsorption to the hydrolysis of compatibility” between molecular architectures of collector molecules
the collector to produce electrically neutral free-acid molecules, as and mineral surface structures.
shown in Eqn 1: Computational simulation based on density functional theory (DFT)
is a powerful tool to study microscopic surface structure and adsorption
MS| + HXaq ⇔ MS|HXaq (1)
mechanism (Chen et al., 2010, 2011, 2014; Gao et al., 2017; Li et al.,
Stelninger (1967) revealed the relationship between the pH limit of 2011; Liu et al., 2013; Reich and Becker, 2006; Sarvaramini et al., 2016).
sphalerite flotation and the pKa of various thiol collectors, suggesting Over the past few years, molecular modelling has gathered interest to
that the chemisorption of neutral molecules indeed played a role in the understand the key mechanisms that are the basis of flotation (Foucaud,
mineral-collector flotation system. However, the occurrence of adsorp­ et al., 2019). Atomistic simulations can be applied to describe the solid/
tion at high pH values indicated that the xanthate ion, instead of the liquid interface and the adsorption mechanisms (Foucaud et al., 2018,
xanthate acid, was the active substance in this system. 2021; Hounfodji et al., 2020; Souvi et al., 2017; Chen et al., 2020).
Taggart et al. (1934) did not accept the idea “adsorption” to describe However, DFT simulation for flotation system in which water, minerals
the formation of a film of collector on the mineral surface, and claimed and flotation reagents, namely, a huge number of atoms, are involved, is
that the action capacity of a reagent could be evaluated using the sol­ computation-intensive and time-consuming. Moreover, DFT simulation
ubility product constant of the compound formed by the reagent with cannot satisfactorily deal with the cases of ions and charged systems.
the metal ion contained in the mineral. Poling and Leja (1963) adopted The principle of coordination chemistry can very well describe both
infrared spectroscopy to demonstrate the nature of collector reaction on metal ions in complexes and in mineral crystals, so it is suitable for
the mineral surface. Mellgren (1966) used microthermometry to test the describing the interactions of reagents with metal ions in mineral sur­
adsorption heat of ethyl xanthate on galena surface, suggesting that the face. The paper employs ligand field theory (LFT), which mainly deals
uptake of xanthate by oxidized galena was energetically equivalent to with metal ions containing d orbitals, for discussion. Therefore, minerals
the chemical reaction of forming lead ethyl xanthate from lead sulphate. containing d orbitals are investigated. In addition, coordination com­
However, the solubility product constant obtained by adsorption was prises two major interactions: σ-bonding and π-backbonding. The latter
quite different from that obtained by solubility. Sutherland and Wark belongs to covalent interaction and plays a crucial role in mineral-
(1955) proved that the monolayer adsorption of collectors was not reagent interactions. Sulphide minerals with covalent properties are
directly related to the usual solubility product constant. thus studied. In order to explore the effect of spin state of d electrons on
Adsorption is merely a phenomenal description, and it cannot give an mineral-reagent interactions, hematite (high-spin iron) and pyrite (low-
explanation for the processes of attachment in which the polar groups of spin iron) are selected to investigate the interaction with xanthate. This
the collector bind with the mineral. Mechanisms of neither adsorption work is of great significance for further understanding the interaction
nor chemical reaction can explain why metal ions exhibit different mechanism of flotation reagents and sheds new light on the develop­
reactivity to collectors on various mineral surfaces. For example, thio- ment of novel reagents based on mineral structures.
containing collectors respond strongly to sulphide minerals, but
weakly to oxide minerals. The properties of minerals have attracted 2. Samples and methods
people’s attention since 1953. Some scholars found that the flotation of
sulphide mineral was related to the pulp potential (Ewans and Ewers, 2.1. Mineral samples and flotation tests
1953; Salamy and Nixon, 1954). After that, the electrochemical prop­
erties of sulphide mineral and its surfaces began to gain widespread In order to investigate the influence of spin states on the interaction
attention. Allison et al. (1972) proposed a theory that the electrostatic between minerals and reagents, two minerals with different spin states
potentials determined the oxidation product of the collector. Electro­ were selected for comparison, i.e., pyrite (low spin, non-magnetic) and
chemical investigations of thiol-containing collectors interacting with hematite (high spin, magnetic). Samples of pyrite were obtained from
sulphide minerals have demonstrated that the three anodic processes - the Dachang Mine in Guangxi Province, China, and hematite from
chemisorption, the reaction to form a metal collector compound, and the Anshan Mine in Liaoning Province, China. High purity samples were
formation of a dithiolate, play a significant role in creating hydrophobic hand-picked from the mine, and a chemical analysis revealed that the
surfaces (Woods, 1971). Although considering the effects of mineral purity of the hematite sample was 98.2% and pyrite 98.5%. Pyrite and
properties and surface oxidation, electrochemistry cannot explain the hematite samples were dry-ground in a porcelain ball mill and dry-
mechanism of non-ionic collectors such as Z-200 (O-isopropyl-N-ethyl screened to obtain – 150 + 74 µm particles.
thionocarbamate), nor can it give the microstructure of the interaction Impurity-doped ZnS samples were prepared by means of a copreci­
between reagent molecules and mineral surfaces. pitation method (Vacassy et al., 2010). Certain amounts of sulfocarba­
Ogwuegbu and Chileshe (2000) considered that the processes of mide and zinc chloride were mixed into an aqueous solution, and after 2
leaching, solvent extraction and flotation were major methods of min­ h of stirring, the dopant (copper chloride, cadmium chloride and ferrous
erals processing under aqueous conditions in which coordination chloride) was added to obtain the precursor. Next, the solution was
chemistry might play an important role; the success of flotation should heated in a boiling water bath at 95 ℃ for 2 h, and then NH3⋅H2O was
be ascribed to the ability of the chemical reagents to form stable com­ added to adjust the pH to 13–14. After pH adjustment, the solution was
plexes with metal ions in aqueous solution or on mineral surfaces. further heated to 100℃, stirred for 2 h, and then cooled to ambient
However, as we have known, the metal ions in mineral surfaces are temperature. The products were washed with distilled water and abso­
completely different from free ions in solution. The former is in a semi- lute ethyl alcohol several times and then heated in a thermostatic drying
constrained state, coordinated with adjacent atoms in mineral surfaces, oven at 100 ℃.
and their properties are thus dependent upon the properties and coor­ The micro-flotation tests were performed using an XFG-II flotation
dination numbers of surrounding bonding atoms (Chen and Zhu, 2021). machine (Wuhan Prospecting Machinery Factory) with a 40 mL cell
It is noteworthy that most of collector molecules and flotation medium volume under 1850 rpm impeller speed. In each micro-flotation test, 5.0
(water) are typical ligands that can also coordinate with surface metal g of mineral powder (–150 + 74 μm) and 40 mL deionized water were

2
J. Chen Minerals Engineering 171 (2021) 107067

added to the cell. The suspension pH is 7.0. Potassium butyl xanthate


(KBX, as a collector), purchased from Zhuzhou Flotation Reagents Fac­
tory, was added and conditioned for 3 min, then methyl isobutyl
carbinol (MIBC, as a frother) was added and conditioned for another 2
min. Thereafter, the flotation was carried out for 4 min and both the
float and sink fractions were collected, dried and weighed. Recovery can
be calculated by the weight of concentrates and tailings. And the
equation is shown in Eqn 2.
mC
R= (2)
mT + mC
Where R denotes the recovery (%); mC denotes the mass of concen­
Fig. 1. The hetero-coordination model of R and M.
trates; and mT denotes the mass of tailings. All the flotation tests were
performed three times, and the average values were taken as the final
data. The error of flotation recovery is less than 5%. eV/Å, the maximum energy change of 2.0 × 10− 5 eV/atom and the
maximum stress of 0.1 GP, and the SCF convergence tolerance was set to
2.0 × 10− 6 eV/atom. In addition, the spin-polarization was used for all
2.2. Time-of-flight secondary ion mass spectroscopy (TOF-SIMS) test
calculations.
The KBX adsorption on pyrite and hematite surfaces was tested by an
3. Coordination model of reagents with metal ions in mineral
IONTOF TOF SIMS IV™ secondary ion mass spectrometer. This tech­
surface
nique was capable of providing elemental and molecular information
from the surface of outmost layers. The static mode of TOF-SIMS was
3.1. Hetero-coordination model
adopted. An isotopically enriched 209Bi+ primary ion beam operating at
25 kV, 0.3 pA was rastered across an area of interest on the sample
The metal ions in mineral surface are coordinated by surrounding
surface. The positive and negative secondary ions of regions of interest
atoms, which are in a semi-constrained state (Chen and Zhu, 2021).
were acquired by the bombardment of the surface with 209Bi+ primary
Therefore, the coordination of flotation reagents (including water mol­
ion beam and raster size was 300 μm. For the region of interest, the data
ecules) with mineral surface ions is different from traditional coordi­
presented (counts) were normalized by the total ion intensity (counts of
nation. The coordination interaction could be called hetero-
the recorded total mass spectrum). At least nine or ten representative
coordination (Pilipenko and Tananayko, 1988). Fig. 1 shows the
regions on each sample were measured and the normalized data were
hetero-coordination model of the reagent molecule R interacting with
plotted in the vertical box plots.
the 5-coordinate metal ion M in mineral surface.
The freshly polished sample surface was conditioned in 10− 4 mol/L
According to the hetero-coordination model, the effective hetero-
KBX solution at pH 7.0 for 10 min. The samples were removed from the
coordination of reagent molecules with metal ions in mineral surface
solution, blew with nitrogen and immediately introduced to the fore-
needs to meet the following conditions (Sun, 2014, 2020; Pilipenko and
vacuum chamber of the TOF-SIMS instrument.
Tananayko, 1988):

2.3. DFT calculation


(1) Spatial structure matching: metal ions in mineral surface should
have enough space for reagent molecules to interact with.
The unit cell optimization of pyrite and hematite and their surface
(2) Orbital symmetry matching: the symmetry of the molecular
models were performed using DFT with CASTEP program (Payne et al.,
orbital of the reagent must match that of the metal ion orbital in
1992), while the frontier orbital calculation was carried out with DMol3
mineral surface to achieve effective bonding.
(Delley, 2000) under the same condition. The calculation details can be
found in our published paper (Cui et al., 2020).
These are the fundamental differences between reagents coordi­
To obtain molecular orbitals of collectors (KBX) and depressants (Ca
nating with free metal ions and reagents coordinating with surface metal
(OH)+, OH–, CN–, HS− , SO2− 2−
3 and S2O3 ), all the calculations were
3 ions. First, metal ions in mineral surfaces need to provide enough space
performed by DMol program (Delley, 1990, 2000) in Materials Studio.
for reagent molecules, such as the adsorption of HS− on smithsonite
The exchange–correlation functional applied was the generalized
surface. The results from DFT calculations show that HS− interacts quite
gradient approach (GGA) (Perdew et al., 1996) of Perdew and Wang
weakly with the 5-coordinate zinc atoms on the surface of smithsonite.
(PW91) (Perdew et al., 1992). All electrons were used in core treatment,
The S-Zn distance reaches 2.84 Å, as shown in Fig. 2(A), well beyond the
and the DND basis set was used within a medium quality whose self-
sum of the radii of the sulphur ion and zinc ion by 2.58 Å. In general, the
consistent field (SCF) tolerance was set to 1.0 × 10− 5 eV/atom. The
inert structure of zinc ion (3d10) can explain the weak binding of HS− to
convergence tolerance of (a) the maximum energy was 2.0 × 10− 5 Ha,
zinc ions. However, steric hindrance can more clearly account for the
(b) the maximum force was 0.004 Ha/Å, and (c) the maximum
long distance between HS− and the zinc ions in smithsonite surface. As
displacement was 0.005 Å. The solvation model and spin-polarized set
displayed in Fig. 2(B), zinc atoms in smithsonite surface are surrounded
were considered for all calculations.
by oxygen and carbon atoms, and only less than 1/6 of the zinc atom
The spin values of common reagents, including butyl xanthate (BX− ),
space is left for the coordination of HS− . Since the radius of HS− (1.80 Å)
dithiophosphate (DTP), dithiocarbonate (DTC), H2O, O2, OH–, CaO+ and
is much larger than that of oxygen ion (1.42 Å), the maximum coordi­
CN–, were calculated by CASTEP program (Payne et al., 1992). The
nation number of zinc ion with sulphide ion is four, that is, a sulphide
exchange–correlation functional applied was the generalized gradient
ion needs at least 1/4 of the zinc ion space for effective interactions.
approach (GGA) (Perdew et al., 1996) of Perdew and Wang (PW91)
Obviously, there is not enough space for zinc ions in smithsonite surface
(Perdew et al., 1992) with plane wave cut-off energy of 10 eV and
to achieve hetero-coordination with HS− . Effect of steric hindrance on
Monkhorst-Pack (Monkhorst and Pack, 1976; Pack and Monkhorst,
the adsorption of collectors on sphalerite, chalcopyrite and pyrite sur­
1977) k-point sampling density of 1 × 1 × 1. All electrons were
faces has been reported by Chena et al. (2021) and Chenb et al. (2021).
considered using ultrasoft pseudopotentials (Vanderbilt, 1990). The
Owing to the semi-constrained state of the metal ion in mineral
convergence tolerances for geometry optimization calculations were set
surfaces, the coordinated atoms alter the electronic structure of metal
to the maximum displacement of 0.002 Å, the maximum force of 0.05

3
J. Chen Minerals Engineering 171 (2021) 107067

Fig. 2. (A) The adsorption configuration of 5-coordinate Zn2+ on smithsonite surface and (B) Corey–Pauling–Koltun (CPK) molecular model of smithsonite surface.

Fig. 3. Coordination model of X− with Fe2+ in octahedral strong field.

ions, and thus affecting the ability of hetero-coordination of metal ions


Table 1
with reagent molecules. The electronic structure of the reagent molecule
σ and π orbitals in different coordination structures.
needs to match that of the metal ion in mineral surface in order to
produce effective interaction. For example, the metal cations of sulphide C.N. Geometry σ orbital π orbital
minerals have strong covalent properties because of the coordination 2 Linear s px py pz dxz dyz
with sulphur anions, and thus metal cations can easily interact with 3 Trigonal planar s px py dxy pz dxz dyz
4 Square planar dx2-y2 dxz dyz
reagents containing S atoms with smaller electronegativity to form co­
Tetrahedral dxy dzz dyz dxy dxz dyz dx2-y2 dz2
valent bonds; the metal cations of oxide minerals possess strong ionic 5 Trigonal bipyramidal dz2 dxz dyz dx2-y2 dxy
properties owing to the coordination with oxygen anions, and the metal Square pyramidal dx2-y2 dz2 dxy dxz dyz
cations can readily interact with reagents containing O or N atoms with 6 Octahedral dx2-y2 dz2 dxy dxz dyz
greater electronegativity to form ionic bonds. According to the frontier 8 Dodecahedral dxy dxz dyz dxy dxz dzy dx2-y2 dz2

orbital theory, only two orbitals with similar orbital symmetry can allow
effective bonding. The discussion on the orbital symmetry matching of simultaneously. Hence, Fig. 3 assumes the d electron arrangement of
reagent molecules with mineral surfaces will be introduced in another metal ions in a strong field. Meanwhile, the orbital of the collectors has
paper. unoccupied π orbitals and lone pairs (σ electron pairs) simultaneously.
The interaction of metal ion (Fe2+ in pyrite) in a strong octahedral field
3.2. Reagents coordinating with metal ions in mineral surface with reagents (xanthate, X− ), as shown in Fig. 3, will be discussed in the
following.
According to ligand field theory, when the metal ion in mineral In the octahedral field, the d orbitals are split into eg orbitals and t2g
surface is coordinated, the five d orbitals (dxy, dxz, dyz, dx2-y2, dz2) of the orbitals, which are σ orbitals and π orbitals, respectively. In the strong
metal ion will be split into two different sets. This will result in changes octahedral field, Fe2+ (d6) has three pairs of π electrons in t2g orbitals,
in orbital properties and d-electron distributions, further influencing the and its eg orbitals are unoccupied orbitals. The LUMO of xanthate is π
interaction of reagents with metal ions in mineral surface. In order to orbitals. According to the ligand field theory, the reagent molecule do­
discuss the role of σ bonding and π-backbonding, the d orbital of the nates lone pairs to the unoccupied eg orbitals of the metal ion to form σ
metal ion requires to have unoccupied orbitals and π electron pairs bonding, denoted by E1; the π electron pairs in the t2g orbitals of the

4
J. Chen Minerals Engineering 171 (2021) 107067

100

80
FeS2

60

40

20
Fe2O3
0
-5.0 -4.5 -4.0 -3.5 -3.0

Fig. 4. Recoveries of hematite and pyrite at different concentrations of KBX at


pH = 7.0.
Fig. 5. The energy level and orbital shape of the xanthate molecule calculated
by DFT.
metal ion interact with the unoccupied π orbitals of the reagent to form
π-backbonding, denoted by E2. In the mineral-reagent interaction, the hydrophobicity, and iron hydroxide was formed on the surface of he­
above two interactions may occur simultaneously, or either of them
matite in the aqueous solution, which hindered the adsorption of the
takes a dominant role. The type of bonding is determined by the ge­
xanthate. It was assumed that xanthate could interact with hematite. In
ometry of bonding.
order to clarify the difference between xanthate-hematite interaction
The type of one orbital is determined by the geometry. Table 1 lists σ
and xanthate-pyrite interaction, the properties of iron ions in different
orbitals and π orbitals in different coordination structures. As shown in
coordination fields and their effects on xanthate are discussed below.
Table 1, s orbitals fall into merely one category: σ orbital, since the s
First of all, the molecular orbitals of xanthate are analysed. As shown
orbital is spherically symmetric. The types of the p orbital and d orbital
in Fig. 5, the HOMO of the xanthate molecule is mainly distributed on
are determined by certain geometry. The bonding atoms of flotation
the double-bonded sulphur atom, which has two 3p lone pairs and can
reagents are mostly linear, in which px is a σ orbital, and pz and py are π
coordinate with the unoccupied orbital of the metal cation. The LUMO
orbitals. As for minerals, 3-, 4-, 5-, and 6-coordinate structures are more
of xanthate molecule is mainly distributed on both double-bonded and
common. The surface structure of spatially tetrahedral minerals can be
single-bonded sulphur atoms. From the perspective of orbital shape, the
regarded as a trigonal plane. For example, sphalerite and chalcopyrite
LUMO of xanthate has the feature of “side by side”, indicating that it is π
crystals are spatially tetrahedral, and their surfaces are trigonal planar,
orbitals. Hence, the xanthate molecule can provide unoccupied π or­
where dxy is a σ orbital, and dxz and dyz are π orbitals. The surface
bitals to form π-backbonding with the π electrons of the metal cation.
structure of spatially octahedral minerals can be considered as trigonal
The molecular formula of pyrite could be written as FeS2, its ligand is
bipyramidal or square pyramidal. For instance, pyrite, arsenopyrite, and
[S2]2− , and the cationic valence is + 2; the molecular formula of he­
galena among other mineral crystals are spatially octahedral, and their
matite is Fe2O3, its ligand is O2–, and the cationic valence is + 3. The iron
surfaces are square pyramidal, where dx2-y2 and dz2 are σ orbitals, and
in both pyrite and hematite crystals has a 6-coordinate structure, and yet
dxy, dxz and dyz are π orbitals. We can determine the orbital type of re­
the ligand properties of these two minerals are quite different. Our
agents and minerals according to Table 1, thereby discussing the inter­
calculation results suggest that the magnetic moment of Fe3+ in hema­
action between their electrons and orbitals.
tite is 4.1 μB, while Fe2+ in pyrite is zero. According to the equation of
In the case of oxide minerals, the ligand is a weak-field one; the
the magnetic moment (without considering the orbital magnetic
d electrons of the metal ion are mainly in a high-spin state; the t2g or­
moment), as shown in Eqn 3:
bitals lack π electrons, and cannot interact with the unoccupied π or­
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
bitals of the reagent molecule. Therefore, the interaction is primarily σ μs = n(n + 2) (3)
bonding with “hard to hard” characteristics. For sulphide minerals, the
Where μs denotes the magnetic moment, the unit is Bohr magneton; n
ligand is a strong-field one; the d electrons of the metal ion are arranged
denotes the number of unpaired electrons. So long as the magnetic
in a low-spin state; the t2g orbitals are filled with π electrons. The metal
moment is not zero, unpaired electrons exist. Electrons tend to be pairs
electron can interact with the unoccupied π orbitals in the reagent
in the low-spin state, and thus the existence of unpaired electrons means
molecule and form π-backbonding. Thus, the interaction is mainly a
the electrons are in a high-spin configuration. Hence, iron in hematite
covalent interaction with “soft to soft” characteristics. To simplify the
(Fe3+) is in a high-spin state, and the ligand O2– is a weak-field ligand.
discussion, we can assume that oxide mineral flotation is dominated by σ
And iron in pyrite (Fe2+) is in a low-spin state, and the ligand [S2]2− is a
bonding and sulphide mineral flotation by π-backbonding.
strong-field ligand.
Fig. 6 presents the d-electron configurations of Fe3+ in a high-spin
4. Results and discussion
state and Fe2+ in a low-spin state in the octahedral field. As shown in
Fig. 6, the five d electrons of Fe3+ in hematite are distributed individ­
4.1. Spin states of metal ions and π-backbonding
ually, and there are no unoccupied orbitals in the σ orbital eg, and thus
Fe3+ cannot accept electron pairs from xanthate. Meanwhile, there is no
4.1.1. Interactions of xanthate with pyrite and hematite
electron pairs in the π orbital t2g of hematite; as a result, it cannot form
Fig. 4 shows the relationship between the concentration of KBX and
π-backbonding with xanthate. In contrast, the eg orbitals of Fe2+ in py­
the flotation recoveries of hematite and pyrite. It is found that KBX can
rite are not occupied, enabling themselves to accept the electron pairs of
collect pyrite effectively, but not hematite. Traditional view suggested
xanthate and form inner-orbital coordination. Moreover, the t2g orbitals
that the hydrocarbon group was too short to provide sufficient

5
J. Chen Minerals Engineering 171 (2021) 107067

Fig. 6. The d-electron configurations of Fe3+ in hematite and Fe2+ in pyrite.

Fig. 7. Adsorption configurations of xanthate on hematite surface (a) and pyrite surface (b).

Fig. 8. TOF-SIMS results before and after the interactions of KBX with hematite and pyrite.

can donate three π electron pairs to interact with the unoccupied π or­ understanding. Density functional theory is used to simulate the
bitals of xanthate, leading to relatively strong π-backbonding. Obvi­ adsorption of xanthate on the surfaces of hematite and pyrite, of which
ously, the Fe3+ in hematite cannot coordinate with xanthate, whereas results are displayed in Fig. 7. Fig. 7(a) shows the distance between the
Fe2+ in pyrite can. bonded sulphur atom of xanthate and the iron ion in hematite surface
The adsorption of xanthate on hematite (0 0 1) and pyrite (1 0 0) reaches 3.192 Å and 3.447 Å, far exceeding the sum of the radii of S2−
surfaces has been well interpreted in our published paper (Cui et al., and Fe3+ (2.44 Å). This indicates xanthate can only weakly interact with
2020). Some key points are recalled in this paper for better the iron ion in hematite surface. In Fig. 7(b), the distance between the

6
J. Chen Minerals Engineering 171 (2021) 107067

Fig. 9. Crystal structures of pyrite (a), marcasite (b) and pyrrhotite(c).

sulphur atom of xanthate and the iron atom in pyrite surface is 2.225 Å It is found in practice that the floatability of marcasite is similar to and
and 2.220 Å, which is smaller than the sum of the radii of S2− and Fe2+ slightly better than that of pyrite. Pyrrhotite has the worst floatability.
(2.54 Å). Therefore, xanthate molecules can effectively bond with iron Generally, the floatability order of these minerals using xanthate as a
ions in pyrite surface. collector is as followed (Harada, 1964): marcasite > pyrite > pyrrhotite.
Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS) was Although these iron(II) sulphides have similar molecular formulas, they
used to examine the surfaces of hematite and pyrite after the adsorption have different crystal structures, as shown in Fig. 9.
of KBX, and the results are shown in Fig. 8. The finding shows that KBX is Pyrite and marcasite, of which formulas are both FeS2, are of iden­
hardly adsorbed on hematite surface, while it is strongly adsorbed on tical components with different structures. Pyrite is of an equiaxed
pyrite surface. The results of TOF-SIMS further confirmed the prediction crystal system with Fe-S bond length of 2.226 Å. Marcasite is of an
based on coordination theory: xanthate interacts weakly with hematite, orthorhombic crystal system, and two of the six Fe-S bonds are short and
whereas strongly with pyrite. four long, with lengths of 2.231 Å and 2.251 Å, respectively. The mo­
According to the d-electron distribution of Fe3+ in hematite crystal, lecular formula of pyrrhotite is Fe1-xS, which belongs to the monoclinic
the following deduction can be made on the basis of ligand field model: system. And two of the six Fe-S bonds are long and four short, with
strong-field collectors containing sulfhydryl (-SH and C = S) fail to lengths of 2.515 Å and 2.241 Å, respectively. Regarding the properties of
interact with hematite, that is, sulphide ore collectors cannot be used for these iron-sulphur minerals, pyrite and marcasite are similar. However,
hematite flotation. This conclusion is consistent with the research results our DFT calculation results show that the magnetic moment of Fe2+ in
of Rao and Finch (2003), which show that hematite cannot be floated pyrite and marcasite is zero, but in pyrrhotite is not zero, indicating that
even with octyl xanthate as a collector. Moreover, Wang (2008) also the Fe2+ in pyrite and marcasite is in a low-spin state. Therefore, both
proposed: however high the concentration of the collectors, such as pyrite and marcasite are strong-field ligands. The Fe2+ in pyrrhotite is in
xanthate, dithiophosphate, ethyl thiocarbamate, and Z-200, they cannot a high-spin state, and thus pyrrhotite belongs to a weak-field ligand. In
collect hematite at any pH; on the contrary, those collectors can be used addition, the strength of the ligand can also be judged from their iron-
for pyrite flotation. sulphur bond lengths. The smaller Fe-S bond length indicates that the
π-backbonding and the ligand field are stronger; conversely, the larger
4.1.2. Crystal structure affecting the coordination ability of iron(II) Fe-S bond length suggests that the π-backbonding and the ligand field
Pyrite, marcasite and pyrrhotite are common iron-sulphur minerals. are weaker. Accordingly, pyrite and marcasite are strong-field ligands,

Fig. 10. Arrangements of d electrons of Fe2+ in the ligand fields of pyrite, marcasite and pyrrhotite.

7
J. Chen Minerals Engineering 171 (2021) 107067

Fig. 11. Crystal structures of copper sulphide minerals.

and pyrrhotite is a weak-field ligand. The arrangements of d electrons of σ-bonding and π-backbonding of pyrrhotite with xanthate are weaker
Fe2+ in the ligand fields of pyrite, marcasite and pyrrhotite are shown in than those of pyrite and marcasite, and pyrrhotite has the worst
Fig. 10. floatability.
As shown in Fig. 10, three pairs of π electrons occupy the t2g orbitals In addition, from the perspective of coordination theory, Fe2+ in
of Fe2+ in pyrite crystal, and so in marcasite crystal. The difference pyrrhotite is in a high-spin state, lacks π electron pairs and forms weak
between pyrite and marcasite is the t2g orbitals in pyrite are of the same π-backbonding with collectors, Thus, for collector design, we should
energy, while not in marcasite. According to π-backbonding model, both consider the ability to donate electron pairs (i.e., σ bonding), as well as
pyrite and marcasite can provide the same amount of π electron pairs, its capability to provide unoccupied π orbitals (i.e., π-backbonding), in
meaning the π-backdonation of them with xanthate is comparable, order to enhance the interaction of collectors with the 4 s and 4p orbitals
hence similar floatability of the two minerals. When we consider the of Fe2+ in pyrrhotite.
transition probability, two pairs of π electrons in marcasite have higher
energy level than the three π electron pairs in pyrite, and thus marcasite
can more easily interact with the unoccupied π orbitals in xanthate. 4.2. The amount of π electron pairs in metal cations and the strength of
Moreover, the Fe-S bond lengths of marcasite (2.231 Å and 2.251 Å) are π-backbonding
shorter than that of pyrite (2.269 Å), suggesting the ability of marcasite
to donate π electron pairs is slightly stronger than that of pyrite. 4.2.1. Floatability of copper sulphide minerals
Considering the above three factors, we can conclude that marcasite Common copper sulphide minerals include chalcopyrite, chalcocite,
interacts more actively with xanthate than pyrite does. bornite and covellite. Among them, chalcopyrite is the most important.
Pyrrhotite has completely different d electron configuration from The floatability order of these minerals using xanthate as a collector is as
pyrite and marcasite. The d electrons of pyrrhotite are in a high-spin followed (Wark and Cox, 1934b):
state, and merely one π electron pair is distributed in the lowest Chalcocite > Covellite > Bornite > Chalcopyrite
orbital of t2g, i.e., dyz. As a result, pyrrhotite has the weakest ability to The chemical formulas of these copper minerals are: Cu2S (chalco­
provide π electron pairs and to form π-backbonding with xanthate. In cite), CuS (covellite), Cu5FeS4 (bornite), CuFeS2 (chalcopyrite). It is
addition, there are two unoccupied orbitals in the eg orbitals of pyrite generally believed that the more Cu in mineral surface, the stronger the
and marcasite, which can interact with the lone pair of xanthate to form interaction of xanthate with Cu2+. Actually, the component of copper
inner-orbital coordination. However, the eg orbitals of pyrrhotite has minerals is far more complicated than their chemical formulas. For
two single electrons and no unoccupied orbital, resulting in the inca­ example, bornite and covellite have both Cu2+ and Cu+; chalcopyrite
pability of pyrrhotite to form inner-orbital coordination with xanthate possesses two configurations(Buckley and Woods, 1984): Cu+Fe3+S2
but the ability to form outer-orbital coordination. Therefore, both the and Cu2+Fe2+S2; Bornite also presents two structures (Goh et al., 2006):
Cu+4Cu2+Fe2+S4 and Cu+5Fe3+S4. How do these possibly existing ions

8
J. Chen Minerals Engineering 171 (2021) 107067

Fig. 12. Cleavage planes of copper sulphide minerals.

Fig. 13. Distributions of d electrons in a high-spin state of Cu2+, Cu+, Fe3+ and Fe2+ in the tetrahedral field.

like Cu+, Cu2+, Fe2+, and Fe3+ interact with xanthate? And what are the
Table 2
differences between the interactions? The following will discuss the
The number of π electron pairs donated by the metal cations of copper sulphide
questions according to disparity in ligand fields of minerals.
minerals.
The crystal and coordination structures of different copper sulphide
Mineral Formula Amount of π electron pairs Average
minerals are presented in Fig. 11. It is found that the structures of
chalcopyrite and bornite crystals are relatively simple, in which Cu and Cu+ Cu2+ Fe2+ Fe3+
Fe are 4-coordinate atoms. The copper atoms in chalcocite are mostly 3- Chalcocite Cu2S 5 5.0
coordinate with few of a 2-coordinate structure, meaning Cu mainly has Covellite Cu2+S2⋅Cu+2S 5 4 4.67
one valence state in chalcocite crystal. Covellite has both 3-coordinate Bornite Cu+4Cu2+Fe2+S4 5 4 1 4.16
Cu+5Fe3+S4 5 0 4.16
and 4-coordinate Cu, suggesting two valence states of copper exist in
Chalcopyrite Cu2+Fe2+S2 4 1 2.5
covellite crystal. In general, Cu2+ is blue and Cu+ is greyish purple. It Cu+Fe3+S2 5 0 2.5
can be inferred from the colour that the content of Cu2+ in covellite is
greater than in chalcocite, while Cu+ is dominant in the latter. The
chemical formulas of the two also reflect the relationship between surfaces, and correspondingly the ligand fields become tetrahedral.
colour and valence: Cu+ 2S
2−
(chalcocite), Cu2+S2− (covellite). Fig. 13 illustrates the high-spin distribution of d electrons of Cu2+,
The stable cleavage surfaces of four copper sulphide minerals are Cu+, Fe3+ and Fe2+ in the tetrahedral field, where the e is pure π orbitals
shown in Fig. 12. It is observed that the stable cleavage surfaces of and the t2 is both σ and π orbitals. Cu2+ has four π electron pairs while
chalcocite, chalcopyrite, and bornite are all of a 3-coordinate structure. Cu+ possesses five in the tetrahedral field. Although Cu+ is of a d10
Although there are 3- and 4-coordinate structures in covellite crystal, the configuration, the d10 → d9s1 energy barrier is relatively low, with
cleavage surface of covellite (0001) presents solely 3-coordinate struc­ strong d orbital activity of Cu+. This can also be confirmed by the
ture. As a result, after collector molecules interact with these minerals, smaller solubility product constant of Cu+ ethyl xanthate (Ksp =
the originally 3-coordinate Cu atoms turn to 4-coordinate ones in the 10− 19.28) (Kakovsky, 1957). In the weak tetrahedral field, Fe3+ has no π

9
J. Chen Minerals Engineering 171 (2021) 107067

100 Table 3
Solubility product constants of compound formed by ethyl-xanthate with
different metal ions (Glembotskii, 1981).
Zn(Cu)S
80 Ion Mn2+ Fe2+ Zn2+ Cu+ Cu2+ Cd2+
Zn(Cd)S
− 9.9 − 7.1 − 8.2 − 19.28 − 24.1 13.59
ksp 10 10 10 10 10 10−

60
%

Fig. 14 indicates that the floatability of sphalerite without impurities


Recovery

is poor, with the recovery of only 14%; the doping of iron and manga­
40
nese ions has a relatively small effect on the recovery, while that of
Zn(Fe)S
copper and cadmium ions greatly increases the recovery. From another
20 perspective, the interaction of xanthate with sulphide minerals is an
Zn(Mn)S electrochemical process, and the electrochemical adsorption of xanthate
Zns and the reduction of oxygen are conjugated reactions. The sphalerite is
0 an insulator, of which width of forbidden band reaches 3.6 eV. Thus,
0 1 2 3 4 5 oxygen cannot be adsorbed on sphalerite surface, thereby depressing the
impurity content % electrochemical adsorption of xanthate. In fact, iron and manganese
impurities can greatly enhance the conductivity of sphalerite, and even
Fig. 14. Relationship between the recoveries and doping concentrations of make sphalerite transform from an insulator into a conductor. However,
impurity ions (KBX: 6.0 × 10− 4mol/L). iron and manganese do not improve the interaction of sphalerite with
xanthate, and the floatability of sphalerite with manganese and iron
electron pair, and Fe2+ has merely one. impurities is still poor. The electrochemical theory of flotation can
Table 2 lists the amount of π electron pairs that metal cations can hardly explain the influence of impurities on the floatability of
provide. Chalcocite mainly contains Cu+, with too small content of Cu2+ sphalerite.
to be noticed, thereby providing 5 pairs of π electrons. Covellite is blue Table 3 lists the solubility product constants of metal xanthates
due to the overwhelming presence of Cu2+. It is generally believed that (sodium butyl xanthate). The Ksp of copper xanthate and cadmium
one-third of copper in covellite has a + 2 valence state, and two-thirds of xanthate is the smallest, consistent with the fact that sphalerite con­
copper possess a + 1 valence state. The quantity ratio of 3-coordinate taining copper and cadmium impurities has better floatability. However,
copper to 4-coordinate copper in chalcocite crystal is 1:2. Moreover, the Ksp of manganese xanthate, iron xanthate, and zinc xanthate does not
the probability of 3- and 4-coordinate copper atoms appearing on the correspond to their floatability. The Ksp of iron xanthate is larger than
(0001) cleavage surface of chalcocite is also 1:2. Therefore, the ratio of that of manganese xanthate and zinc xanthate and Fe2+ interacts most
Cu2+ and Cu+ in the chemical formula can be used to calculate the π weakly with xanthate, but sphalerite containing iron impurity has
electron pairs chalcocite provides, and the calculation result is 4.67. The higher floatability than other impurity-bearing sphalerite. This indicates
valence states of copper and iron in bornite and chalcopyrite have been that the interaction of metal ions with xanthate alone cannot explain the
controversial. In practice, different valence states and structures may flotation behaviour; the influence of the coordination structure of min­
exist. As a result, when calculating the amount of π electron pairs, the erals on the properties of metal ions should also be considered.
contribution of different valence states should be considered. According Sphalerite is of a tetrahedral structure. The tetrahedral field is
to the chemical formulas, the calculated result of π electron pairs is 4.16 commonly a weak field due to its relatively small splitting energy.
pairs for bornite and 2.5 for chalcopyrite. Therefore, the d electrons of metal cations are distributed in a high-spin
According to previous discussion, xanthate possesses unoccupied π state in the tetrahedral field. Fig. 15 shows the d orbital splitting of
orbitals; with more π electron pairs, copper sulphide is more likely to Mn2+(3d5), Fe2+(3d6), Cu2+(3d9), and Zn2+(3d10) and their electron
form π-backbonding with xanthate, thereby improving floatability of arrangements.
minerals. As found in Table 2, the amount order of the π electron pairs of In the tetrahedral field, the e is π orbitals, and the t2 is π and also σ
copper sulphide minerals is as followed: orbitals. As seen in Fig. 15, the amount of π electron pairs of Mn2+ in the
Chalcocite > Covellite > Bornite > Chalcopyrite tetrahedral field is 0, resulting in its incapability to form π-backbonding
Accordingly, the π-backbonding of xanthate with chalcocite is the with xanthate, hence the worst floatability of Mn-doped sphalerite; Fe2+
strongest, followed by covellite and bornite, and chalcopyrite the has a pair of π electrons and can form weak π-backbonding with
weakest. This is consistent with the floatability order of copper sulphide xanthate, and thus the sphalerite containing iron impurity has better
using xanthate as a collector. floatability than that containing manganese; Cu2+ has four pairs of π
electrons and strong ability to form π-backbonding with xanthate, hence
4.2.2. Floatability of sphalerite containing impurities the best floatability. In addition, although Zn2+ has five pairs of π
It has been found in industry that sphalerite from different deposits electrons, the floatability of natural sphalerite is relatively poor. This is
or separate sections of the same deposit has diverse colours because of because Zn2+ is of a d10 configuration with fully occupied orbitals, it is
distinct impurities, ranging from light green, tan, and dark brown to inert, it is strongly localized, and it provides π electrons poorly.
steel grey. The floatability of sphalerite of different colours varies Cd2+ is also of a d10 configuration, and similarly its d orbitals are
greatly. Pure sphalerite is almost colourless since the 3d orbitals of the inactive. However, sphalerite containing cadmium impurity shows
zinc ion are fully occupied and no electronic transition occurs. When better floatability according to the results in Fig. 14. This is because the
sphalerite crystals contain impurity ions, especially transition metal ions polarizability of Cd2+ (1.09 Å3) is greater than that of Zn2+ (0.288 Å3),
whose 3d orbitals are not fully occupied, such as iron, manganese, which enhances the covalent interaction between the cations and li­
copper, and the like, sphalerite will show diverse colours. Impurity ions gands. According to the ligand field theory, xanthate with unoccupied π
greatly influence the flotation of sphalerite. Generally, impurities such orbitals belongs to soft base and interacts readily with soft acid of high
as manganese and iron will reduce the floatability of sphalerite, while polarizability. Therefore, the greater the polarizability of a metal cation,
those such as copper and cadmium will improve. We used the water bath the more easily it changes the electron cloud shape, and the stronger
method to synthesize sphalerite with different doping contents of covalent interaction between the metal ion and xanthate. It can be
manganese, iron, copper and cadmium ions, and the corresponding re­ predicted that the metal cations of large polarizability such as Hg2+
coveries are shown in Fig. 14. (1.24 Å3), Au+ (1.88 Å3), Pb2+ (2.00 Å3), Bi3+ (1.74 Å3), Tl+ (3.13 Å3),

10
J. Chen Minerals Engineering 171 (2021) 107067

Fig. 15. The d-electron configurations of Mn2+, Fe2+, Cu2+, Zn2+ in the tetrahedral field.

Fig. 16. Molecular orbitals of common depressants.

Ce3+ (1.03 Å3), Bi3+ (1.74 Å3), and Pt2+ (1.43 Å3) activates sphalerite in an octahedral field. It can provide three π electron pairs, thereby
effectively. Sutherland and Wark (1955) studied the activation effect of easily interacting with the unoccupied π orbitals of Ca(OH)+ to form
the metal ions in sphalerite, and the result is in line with the prediction π-backbonding, which enhances the interaction between Ca(OH)+ and
based on the polarizability. In particular, the activation of thallium ions pyrite. On the contrary, OH– cannot provide unoccupied π orbitals to
is in perfect agreement with the prediction. The solubility product form π-backbonding with Fe2+ on pyrite surface; the interaction be­
constant of thallium xanthate is 2.9 × 10− 8, close to that of zinc xanthate tween OH– and pyrite surface is merely σ-bonding, which is weak.
(4.9 × 10− 8) (Glembotskii, 1981). Polarizability can well explain the (2) Lime depresses pyrite, arsenopyrite, pentlandite, chalcopyrite
activation effect of thallium ions, whereas solubility product constant and sphalerite strongly whereas it depresses galena weakly. The valence
cannot. shell configuration of Pb2+ in galena is 6s26p0 without d orbitals, and the
s and p orbitals are σ orbitals in an octahedral field. Thus, galena cannot
provide π electron pairs or form π-backbonding with Ca(OH)+. In
4.3. Unoccupied π orbitals and selectivity of depressants
contrast, pyrite, arsenopyrite, pentlandite, chalcopyrite and sphalerite
have d orbitals and π electron pairs in the t2g (octahedral) or e (tetra­
The density functional theory was used to calculate the molecular
hedral) orbitals. The electron pairs can interact with the unoccupied
orbitals of common depressants, and the results are shown in Fig. 16.
orbitals of Ca(OH)+ to form π-backbonding and thus improve the
OH– is a hard base and has no unoccupied π orbitals; it can provide two
depressing ability of Ca(OH)+.
pairs of σ electrons, and only the σ-bonding occurs in adsorption. Ca
(3) NaOH is not as effective as lime in depressing pentlandite. The
(OH)+ has an unoccupied π orbital at the secondary level, making it a
d electron arrangement of Ni2+ in pentlandite is (t2g)6(eg)2. There are
softer base than OH–; it provides two lone pairs, and both σ-bonding and
two electrons in the eg orbital; without unoccupied orbitals, Ni2+ can
π-backbonding exist in adsorption. HS− has no unoccupied π orbitals for
merely form outer-orbital coordination with OH–, instead of inner-
LUMO; it provides two lone pairs; the σ-bonding is dominant in
orbital coordination. The two electrons in the eg orbital repel OH–
adsorption. The LUMO of CN– is π orbitals, suggesting it is a typical soft
strongly, weakening the adsorption of OH–. Three π electron pairs in the
base; there is one lone pair in the anti-bonding orbital, and the π-back­
t2g orbital of Ni2+ enable it to form π-backbonding with Ca(OH)+ easily,
bonding plays a dominant role in adsorption. The LUMO of SO2− 3 is π
and thus lime is more effective in depressing pentlandite than NaOH is.
orbitals, while the LUMO of S2O2− 2−
3 is σ orbitals, and thus SO3 can accept
π electron pairs from metal ions to form π-backbonding, but S2O2− 3
4.3.2. NaCN
cannot.
Galena crystal is of a 6-coordinate structure and its ligand field is
octahedral. The s and p orbitals are σ orbitals, and the t2g orbitals (of
4.3.1. Ca(OH)+ and OH–
d orbitals) are π orbitals. However, the outer-shell configuration of Pb2+
Due to the difference between the unoccupied π orbitals of Ca(OH)+
in galena is 6s26p0 and its 5d orbitals are completely blocked. As a result,
and OH–, the following conclusions can be obtained according to the
galena cannot donate π electrons to form π-backbonding with cyanide,
coordination theory:
and thus NaCN does not virtually depress galena. In addition, bismu­
(1) Lime depresses pyrite more efficiently than NaOH. The Fe2+ of
thinite (Bi3+: 5d106s2) and antimonite (Sb3+: 4d105s2) resemble galena.
pyrite is in a low-spin state, and its d electron arrangement is (t2g)6(eg)0

11
J. Chen Minerals Engineering 171 (2021) 107067

Table 4
The spin values before and after adsorption of reagents on pyrite surface.
BX− DTP DTC H2O O2 OH– CaOH+ CN–

Spin Value (μB) Before 0 0 0 0 0 0 0 0


After 0 0 0 0 1.22 1.0 0 0

Their d orbitals are blocked by valence electrons and cannot provide π


Table 5
electron pairs. Therefore, bismuthinite and antimonite are hardly
The critical pH of sulphide minerals in the presences of ethyl xanthate.
depressed by cyanide.
Transition metal sulphide minerals, such as pyrite, chalcopyrite, Minerals Critical pH Literature
values
sphalerite, pentlandite and skutterudite, are of 3d 6− 10 4 s0 structures.
Their 3d orbitals are active and can provide π electron pairs, which Pyrrhotite 6.0 (Wark and Cox, 1934a),
8.0 (Sutherland and Wark, 1955),
interact with the unoccupied π orbitals of CN– to form π-backbonding.
(Kakovsky and Grebnev, 1959)
Thus, those minerals are subject to the depression by cyanide. Pyrite Arsenopyrite 8.4 (Wark and Cox, 1934a)
among those minerals is depressed most severely as it provides not only Pyrite 10.5 (Wark and Cox, 1934a),
three π electron pairs (π-backbonding) but unoccupied eg orbitals to 10.35 (Sutherland and Wark, 1955),
form inner-orbital complexes with cyanide (σ bonding). (Kakovsky and Grebnev, 1959)
Galena 10.4 (Wark and Cox, 1934a),
In addition, minerals containing these precious metal cations, such as
11.0 (Sutherland and Wark, 1955),
Ag+(4d10), Pt+(5d9), Au+(5d10), Au3+(5d8), Pd2+(4d8) and so forth, are (Kakovsky and Grebnev, 1959)
prone to the depression by cyanide. Although some cations are of an Chalcopyrite 11.8 (Wark and Cox, 1934a),
inert configuration (d10), they can easily form π-backbonding with cy­ 12.4 (Sutherland and Wark, 1955), (Kakovsky and
Grebnev, 1959)
anide, because the splitting energy rises with the amount of valence
Activated 12.3 (Wark and Cox, 1934a),
shells. sphalerite (Sutherland and Wark, 1955)
Covellite 13.2
4.3.3. Na2S Chalcocite 13.8 (Wark and Cox, 1934a),
The LUMO of Na2S is σ orbitals, instead of π orbitals; theoretically, 13.8 (Sutherland and Wark, 1955),
(Kakovsky and Grebnev, 1959)
Na2S cannot accept π electron pairs from metal ions to form π-back­
Bornite 14 (Wark and Cox, 1934a),
bonding. Due to the large polarizability (10.2 Å3), S2− is inclined to form (Sutherland and Wark, 1955)
a covalent interaction with metal ions with high polarizability in min­
eral surfaces. Pb2+ (2.0 Å3) is of great polarizability, and thus S2− is
easily adsorbed on galena surface and depresses galena most strongly. As the larger the spin value, the more unpaired electrons in the d orbital.
for copper sulphide and iron sulphide, the interactions of Na2S with It is found that after the adsorption of CN–, Ca(OH)+, BX, DTP and
them are determined by the energy level of unoccupied orbitals, owing DTC, the spin values of the iron are all 0, and the d electrons are ar­
to the low polarizability of copper ions and iron ions. For example, the ranged in a low-spin state, indicating those reagents are strong-field li­
d orbitals of pyrite are in a low-spin state, among which two unoccupied gands. However, after OH– and O2 are adsorbed, the spin values of the
orbitals can accept the lone pairs of Na2S to form inner-orbital coordi­ iron are not 0, and the d electrons are arranged in a high-spin state,
nation; hence, pyrite is prone to depression by Na2S. meaning that OH– and O2 are weak-field ligands. Water molecules are
special. Although they are typical weak-field ligands, the spin value of
4.3.4. Na2SO3 and Na2S2O3 iron after the adsorption is still 0 and the iron is in a low-spin state, that
The LUMO of Na2SO3 is π orbitals and the LUMO of Na2S2O3 is σ is, the adsorption has little effect on the spin state of iron. This is possibly
orbitals. Accordingly, the former can accept π electron pairs from metal because the surface of sulphide ore is of strong hydrophobicity, and the
ions to form π-backbonding whereas the latter cannot. In the d orbitals of interaction between water molecules and pyrite surface is relatively
pyrite are three pairs of π electrons, able to interact with the π orbitals of weak.
Na2S2O3. Na2S2O3 can thus depress pyrite effectively while Na2S2O3 is From the perspective of flotation solution chemistry, the Ksp of iron
an inferior depressant for pyrite. In lead–zinc mineral flotation, Na2S2O3 (II) xanthate is 8.0 × 10− 8, and that of iron(II) hydroxide is 8.0 × 10− 16.
is generally used for depressing pyrite. Accordingly, it is hard for xanthate to be adsorbed on pyrite surface
under alkaline condition, but actually the reverse is true in flotation
experiments. Spin state can offer reasonable explanation. For pyrite-OH–
4.4. Crystal field conversion barrier and flotation critical pH
interaction, iron has to change from low spin to high spin and to over­
come high crystal field conversion barrier (-20Dq). For pyrite-xanthate
A sulphide ore collector needs to have the following two character­
interaction, the spin state of iron remains unchanged and the crystal
istics to be a strong-field ligand. First, it belongs to soft bases with
field conversion barrier is zero. Therefore, xanthate is more likely to be
phosphorus and sulphur atoms in the molecular structure. Second, it has
adsorbed on pyrite surface than OH– is.
unoccupied π orbitals. Strong-field ligands not only have a great impact
Flotation practice shows that every sulphide mineral has a critical pH
on the spin state of metal cations, but also affect the crystal field stabi­
for flotation. When the critical pH is exceeded, sulphide mineral flota­
lization energy. For example, for common sulphide ore collectors such as
tion will be depressed. Table 5 presents the flotation critical pH of
butyl xanthate (BX), dithiophosphate (DTP) and diethyl dithiocarba­
common sulphide minerals under the condition of ethyl xanthate as the
mate (DTC), DFT calculation results have found that when they are
collector. As seen in Table 5, the critical pH values of copper sulphide,
adsorbed on pyrite surface, the iron ion still remains in a low-spin state
such as chalcopyrite, chalcocite, covellite, bornite, and activated
virtually without any alteration; however, when OH– is adsorbed on
sphalerite (it is coated with a copper-bearing film: covellite) are higher,
pyrite surface, the iron turns into a high-spin state. Table 4 lists the spin
followed by galena and pyrite, and those of arsenopyrite and pyrrhotite
values before and after adsorption of reagents on pyrite surface by DFT
are the lowest.
calculation. The spin value is 0, suggesting that there is no unpaired
The stronger the interaction between the mineral surface and OH–,
electron in the d orbital and the electrons are all paired, and the iron is in
the more easily the mineral is depressed by OH–, and the lower the
a low-spin state; the spin value is not 0, indicating that there are un­
critical pH for flotation; on the contrary, the weaker the interaction, the
paired electrons in the d orbital, and the iron is in a high-spin state. And

12
J. Chen Minerals Engineering 171 (2021) 107067

Table 6 iron in pyrite surface is 5-coordinate and turns 6-coordinate after


The solubility product constant of metal hydroxides. adsorbing OH–; the ligand field changes from a tetrahedron to an octa­
Metal Pb Cu CuOH Fe Fe Zn hedron after the adsorption, as shown in Fig. 17.
hydroxide (OH)2 (OH)2 (OH)2 (OH)3 (OH)2 For convenience of discussion, we only consider the influence of
KSP 1.2 × 2.2 × 2.2 × 8.0 × 4.0 × 7.1 × ligand field changes on CFSE, rather than the influence of OH– adsorp­
10− 15 10− 20 10− 14 10− 16 10− 38 10− 18 tion on spin states of surface metal ions. In addition, the valence shell of
Pb2+ in galena has no d orbitals, thereby no CFSE for galena (PbS).
Data from J. A. Dean Ed. Lange’s Handbook of Chemistry, 13th. edition 1985.
Hence, the critical pH of galena is irrelevant to the CFSE, and the
interaction of galena with OH– will not be discussed here. Table 7 lists
higher the critical pH. The solubility product constants of metal hy­
the energy changes of chalcopyrite, pyrite, Cu-activated sphalerite,
droxides are listed in Table 6. The solubility product constants of metal
arsenopyrite and pyrrhotite after adsorption of OH–.
hydroxides are not significantly related to the flotation critical pH. For
The metal ions need to overcome the potential barrier ΔE for the
example, the Ksp of copper hydroxide is smaller than those of lead hy­
conversion in different fields. The smaller the barrier, the easier the
droxide and iron hydroxide, indicating that the interaction of hydroxide
conversion; in other words, a negative barrier means that the electron
with Cu2+ is stronger than those with Pb2+ and Fe2+; copper-bearing
rearrangement is conducive to the adsorption of OH– on mineral sur­
minerals should be more susceptible to depression by hydroxide than
faces, while a positive barrier suggests the electron reconfiguration has
pyrite and galena should, but in fact the critical pH of copper-bearing
an adverse impact on the adsorption of OH–. Table 7 shows that the ΔE
minerals is higher than that of pyrite and galena. Most importantly,
for chalcopyrite and copper-activated sphalerite are positive after their
the solubility product constant cannot explain the interactions of hy­
surfaces adsorb OH–. This indicates that the interactions of OH– with the
droxide with iron in different minerals, such as pyrite, arsenopyrite and
copper ions on the surfaces of chalcopyrite and copper-activated
pyrrhotite; although they all consist of Fe2+, the critical pH of them is far
sphalerite are weak, and thus chalcopyrite and copper-activated sphal­
varying.
erite can hardly depressed by OH– and their critical pH is relatively high.
The following will analyse the interaction of OH– with mineral sur­
The ΔE of pyrite, arsenopyrite and pyrrhotite are negative, suggesting
face ions from the perspective of coordination theory. For example, the
that OH– can easily interact with the surface of sulphide minerals, hence

Fig. 17. Changes of the coordination number and ligand field in pyrite surface before and after adsorption of OH–.

Table 7
Effect of adsorption of OH– on ligand field structures and CFSE.
Mineral Ligand field Ion CFSE ΔE
2+
Chalcopyrite Surface Trigonal planar Cu − 5.48 Dq 3.70Dq
Adsorption Tetrahedral Cu2+ − 1.78 Dq
Surface Trigonal planar Fe2+(HS) − 3.87 Dq 1.20Dq
Adsorption Tetrahedral Fe2+(HS) − 2.67 Dq
Cu-activated sphalerite Surface Trigonal planar Cu2+ − 5.48 Dq 3.70Dq
Adsorption Tetrahedral Cu2+ − 1.78 Dq
Pyrite Surface Tetragonal pyramidal Fe2+(LS) − 20 Dq − 4.0Dq
Adsorption Octahedral Fe2+(LS) − 24 Dq
Arsenopyrite Surface Tetragonal pyramidal Fe2+(LS) − 20 Dq − 4.0Dq
Adsorption Octahedral Fe2+(LS) − 24 Dq
Pyrrhotite Surface Tetragonal pyramidal Fe2+(HS) − 4.57 Dq -(η-0.57Dq)
Adsorption Monoclinic octahedral Fe2+(HS) − 4.0Dq-η

Note: HS denotes the high spin state and LS denotes the low spin state.

13
J. Chen Minerals Engineering 171 (2021) 107067

by the adjacent atoms, the interaction of reagent molecules with the


metal cations in mineral surfaces is hetero-coordination. The coordina­
tion properties of surface metal ions determine its flotation behaviours
and interactions with reagents.
In hematite, iron(III) is high-spin and has no π electron pair in d or­
bitals, unable to form π-backbonding with xanthate. Conversely, in py­
rite, iron(II) is low-spin and possesses three pairs of π electrons, prone to
π-backbonding with xanthate. In pyrrhotite, iron(II) is high-spin, with
one π electron pairs in d orbitals; thus, π-backbonding of xanthate with
pyrrhotite is milder than with pyrite and marcasite. From the perspec­
tive of ligand field strength, a monoclinic octahedral field (marcasite) is
stronger than a regular octahedral field (pyrite). Hence, π-backbonding
of xanthate with marcasite is somewhat more intense than with pyrite.
Similarly, the interaction mechanisms of xanthate with copper sulphide
minerals and with impurity-containing sphalerite can be well explained
by the π-backbonding model. Moreover, greater polarizability of metal
Fig. 18. The d electron configurations of Cu2+ in trigonal planar field. ions indicates better activating effect on sphalerite.
The unoccupied π orbitals of depressants determine the depressive
the easier depression and the lower critical pH. effect on sulphide minerals, since the orbitals can accept π electron pairs
For arsenopyrite and pyrite, although their ΔE is the same, their from surface metal ions to form π-backbonding. Ca(OH)+, CN–, HS− and
splitting energy (Dq) differs. The ligand of pyrite is [S2]2− , and that of SO2−
3 that all contain unoccupied π orbitals are effective depressants for
arsenopyrite is [AsS]2− . They both have unoccupied π orbitals and pyrite with three π electron pairs. Without unoccupied π orbitals, OH–
belong to strong-field ligands. However, since the electronegativity of and S2O2− 3 barely depress pyrite.
arsenic is smaller than that of sulphur, the splitting energy of arsenic Flotation critical pH is related to crystal field conversion barrier
ligand is greater than that of sulphur ligand, and the splitting energy of caused by adsorption of OH– on mineral surfaces. The greater the bar­
arsenopyrite is greater than that of pyrite. Therefore, the ΔE for arse­ rier, the higher the critical pH.
nopyrite is more negative than for pyrite, and the critical pH of arse­
nopyrite is lower than of pyrite. Declaration of Competing Interest
For pyrrhotite, the η is less than 4Dq, the iron in pyrrhotite is in a
high-spin state, the ligand field is a weak field ligand, and the splitting The authors declare that they have no known competing financial
energy is smaller. As a result, the potential barrier for pyrrhotite is interests or personal relationships that could have appeared to influence
smaller than for pyrite and arsenopyrite (-4Dq). In addition, the reason the work reported in this paper.
why pyrrhotite is easily depressed by OH– is relevant to the electron
arrangement in the weak octahedral field. The electron configuration of Acknowledgements
Fe2+ in the weak field is: (t2g)4(eg)2. There are two electrons in the eg
orbital, which repel OH–; meanwhile, there is only one pair of π electrons This research was funded by State Key program of the Joint Funds of
in the t2g orbital, and the π-backbonding with xanthate is very weak. the National Natural Science Foundation of China (Grant No.
Hence, pyrrhotite is most easily depressed by OH–, and its critical pH is U20A20269). The author is thankful for the support.
the lowest.
In addition, given the transition of Fe2+ from low-spin to high-spin References
on mineral surfaces after the action of OH–, we can calculate the po­
tential barriers of spin state transition for different iron-bearing min­ Allison, S.A., Goold, L.A., Nicol, M.J., Granville, A., 1972. A determination of the
products of reaction between various sulfide minerals and aqueous xanthate
erals, which are 20 Dq for pyrite, 20 Dq-η for arsenopyrite, and 0 for solution, and a correlation of the products with electrode rest potentials. Metall.
pyrrhotite. A smaller potential barrier for spin state transition indicates Mater. Trans. B. 3, 2613–2618. https://doi.org/10.1007/BF02644237.
easier adsorption of OH–, a result that is also consistent with the order of Buckley, A.N., Woods, R., 1984. An X-ray photoelectron spectroscopic study of the
oxidation of chalcopyrite. Aust. J. Chem. 37, 2403–2413. https://doi.org/10.1071/
flotation critical pH.
CH9842403.
Moreover, as shown in Table 5, the flotation critical pH values of Bulatovic, S.M., 2007. Handbook of Flotation Reagents: Chemistry. Theory and Practice,
copper sulphide minerals, including Cu-bearing sphalerite, are relatively Elsevier, Amsterdam.
Chen, J.H., Long, X.H., Ye, C., 2014. Comparison of multilayer water adsorption on the
high, and OH– seems to interact weakly with those surfaces. OH– can
hydrophobic galena (PbS) and hydrophilic pyrite (FeS2) surfaces: a DFT study.
barely coordinate with copper ions on the surfaces because of the higher J. Phys. Chem. C 118 (22), 11657–11665. https://doi.org/10.1021/jp5000478.
barrier of CFSE, as shown in Table 7. In addition, Jahn-Teller effect also Chena, J.H., Wang, J.M., Li, Y.Q., Liu, M., Liu, Y.C., Zhao, C.H., Cui, W.Y., 2021. Effects
affects the interaction of OH– with copper on the mineral surface. The of surface spatial structures and electronic properties of chalcopyrite and pyrite on Z-
200 selectivity. Miner. Eng. 163, 106803 https://doi.org/10.1016/j.
copper in copper sulphide minerals, for instance, chalcopyrite, covellite, mineng.2021.106803.
bornite and chalcocite, has a 4-coordinate structure and the surface is a Chen, J.H., Wang, L., Chen, Y., Guo, J., 2011. A DFT study of the effect of natural
3-coordinate trigonal plane. Fig. 18 presents the d electron arrange­ impurities on the electronic structure of galena. Int. J. Miner. Process. 98 (3–4),
132–136. https://doi.org/10.1016/S1003-6326(09)60169-2.
ments of copper ions in trigonal planar field. Chen, J.H., Xu, Z.H., Chen, Y., 2020. Electronic Structure and Surfaces of Sulfide
As found in Fig. 18, for trigonal planar field, there is one pair of Minerals: Density Functional Theory and Applications. Elsevier.
electrons in the dz2 orbitals. Therefore, it has a larger electron density in Chen, J.H., Zhu, Y.G., 2021. Study of semi-constrained properties of metal ions on
mineral surface of flotation system. Journal of China University of Mining &
the z-direction to repel OH–, which results in a weaker interaction be­ Technology, 50(6), 1-8. (In Chinese) https://doi.org/10.13247/j.cnki.jcumt.001271.
tween OH– and copper ions on the mineral surface. According to this Chen, Y., Chen, J.H., Guo, J., 2010. A DFT study on the effect of lattice impurities on the
result, it can be speculated that the more copper ions on the mineral electronic structures and floatability of sphalerite. Miner. Eng. 23 (14), 1120–1130.
https://doi.org/10.1016/j.mineng.2010.07.005.
surface, the higher the critical pH of mineral flotation.
Chenb, Y., Liu, X.M., Chen, J.H., 2021. Steric hindrance effect on adsorption of xanthate
on sphalerite surface: A DFT study. Miner. Eng. 165, 106834 https://doi.org/
5. Conclusion 10.1016/j.mineng.2021.106834.

As the metal ions in mineral surfaces have already been coordinated

14
J. Chen Minerals Engineering 171 (2021) 107067

Cook, M.A., Nixon, J.C., 1950. The theory of water-repellent films on solids formed by Monkhorst, H.J., Pack, J.D., 1976. Special points for Brillouin-zone integrations. Physical
adsorption from aqueous solutions of heteropolar compounds. J. Phys. Colloid Review B 13 (12), 5188–5192.
Chem. 54, 445–459. https://doi.org/10.1021/j150478a002. Ogwuegbu, M.O.C., Chileshe, F., 2000. Coordination chemistry in mineral processing.
Cui, W.Y., Chen, J.H., Li, Y.Q., Chen, Y., Zhao, C.H., 2020. Interactions of xanthate Miner. Process. Extr. Metall. Rev. 21, 497–525. https://doi.org/10.1080/
molecule with different mineral surfaces: a comparative study of Fe, Pb and Zn 08827500008914176.
sulfide and oxide minerals with coordination chemistry. Miner. Eng. 159, 106565 Pack, J.D., Monkhorst, H.J., 1977. Special points for Brillouin-zone integrations - A reply.
https://doi.org/10.1016/j.mineng.2020.106565. Physical Review B 16 (4), 1748–1749. https://doi.org/10.1103/PhysRevB.16.1748.
Delley, B., 1990. An all-electron numerical method for solving the local density Payne, M.C., Teter, M.P., Allan, D.C., Arias, T.A., Joannopoulos, J.D., 1992. Iterative
functional for polyatomic molecules. J. Chem. Phys. 92 (1), 508–517. https://doi. minimization techniques for ab initio total energy calculation: molecular dynamics
org/10.1063/1.458452. and conjugate gradients. Rev. Mod. Phys. 64 (4), 1045–1097. https://doi.org/
Delley, B., 2000. From molecules to solids with the DMol3 approach. J. Chem. Phys. 113 10.1103/RevModPhys.64.1045.
(18), 7756–7764. https://doi.org/10.1063/1.1316015. Perdew, J.P., Burke, K., Ernezerhof, M., 1996. Generalized gradient approximation made
Ewans, L.E., Ewers, W.E., 1953. Recent developments in mineral dressing. (London), Ind. simple. Phys. Rev. Lett. 77 (18), 3865–3868. https://doi.org/10.1103/
Eng. Chem. Inst. Min. Metall. 46, 2420. PhysRevLett.77.3865.
Foucaud, Y., Badawi, M., Filippov, L., Filippova, I., Lebègue, S., 2019. A review of Perdew, J.P., Chevary, J.A., Vosko, S.H., Jachson, K.A., Pederson, M.R., Singh, D.J.,
atomistic simulation methods for surface physical-chemistry phenomena applied to Fiolhais, C., 1992. Atoms, molecules, solids, and surfaces: applications of the
froth flotation. Miner. Eng. 143, 106020 https://doi.org/10.1016/j. generalized gradient approximation for exchange and correlation. Phys. Rev. B 46
mineng.2019.106020. (11), 6671–6687. https://doi.org/10.1103/PhysRevB.48.4978.2.
Foucaud, Y., Badawi, M., Filippov, L.O., Filippova, I.V., Lebègue, S., 2018. Surface Pilipenko, A.T., Tananayko, M.N., 1988. Hetero-ligand and hetero-metal complexes and
properties of fluorite in presence of water: an atomistic investigation. J. Phys. Chem. their applications in analytical chemistry. Fudan University Press, Shanghai (in
B 122 (26), 6829–6836. https://doi.org/10.1016/S1003-6326(09)60169-2. Russian).
Foucaud, Y., Canevesi, R.L.S., Celzard, A., Fierro, V., Badawi, M., 2021. Hydration Poling, G.W., Leja, J., 1963. Infrared study of xanthate adsorption on vacuum deposited
mechanisms of scheelite from adsorption isotherms and ab initio molecular films of lead sulfide and metallic copper under conditions of controlled oxidation.
dynamics simulations. Appl. Surf. Sci. 562, 150137 https://doi.org/10.1016/j. J. Phys. Chem. 67, 2121–2126. https://doi.org/10.1021/j100804a036.
apsusc.2021.150137. Pradip, B.R., 2003. Molecular modeling and rational design of flotation reagents. Int. J.
Fuerstenau, D.W., 1962. Froth flotation 50th anniversary volume. American Institute of Miner. Process. 72, 95–110. https://doi.org/10.1016/S0301-7516(03)00090-5.
Mining, Metallurgical. and Petroleum Engineers. Rao, S.R., Finch, J.A., 2003. Base metal oxide flotation using long chain xanthates. Int. J.
Fuerstenau, D.W., 1957. Correlation of contact angles, adsorption density, zeta Miner. Process. 69 (1), 251–258. https://doi.org/10.1016/S0301-7516(02)00130-8.
potentials, and flotation rate. Trans. AIME 208, 1365–1367. Reich, M., Becker, U., 2006. First-principles calculations of the thermodynamic mixing
Gao, Z., Li, C., Sun, W., Hu, Y., 2017. Anisotropic surface properties of calcite: a properties of arsenic incorporation into pyrite and marcasite. Chem. Geol. 225 (3–4),
consideration of surface broken bonds. Colloids Surf. A Physicochem. Eng. Asp. 520, 278–290. https://doi.org/10.1016/j.chemgeo.2005.08.021.
53–61. https://doi.org/10.1016/j.colsurfa.2017.01.061. Salamy, S.G., Nixon, J.C., 1954. Reaction between a mercury surface and some flotation
Gaudin, A.M., 1957. Flotation. McGraw-Hill. reagents: and electrochemical study. I. Polarization curves. Aust. J. Chem. 7,
Gaudin, A.M., 1928. Flotation mechanism, a discussion of the functions of floatation 146–156. https://doi.org/10.1071/CH9540146.
reagents. Trans. Am. Inst. Min. Metall. Eng. 79, 50–76. Sarvaramini, A., Larachi, F., Hart, B., 2016. Collector attachment to lead-activated
Gaudin, A.M., Martin, J.S., 1928. Flotation fundamentals. Part 3. The flotation of the sphalerite - experiments and DFT study on pH and solvent effects. Appl. Surf. Sci.
carbonates of copper: Malachite and Azurite. Univ. Utah US Bur. Mines, Tech. Pap. 367, 459–472. https://doi.org/10.1016/j.apsusc.2016.01.213.
no. 5. Souvi, S.M.O., Badawi, M., Virot, F., Cristol, S., Cantrel, L., Paul, J., 2017. Influence of
Glembotskii, B.A., 1981. Foundation of Physical Chemistry in the Process of Flotation. water, dihydrogen and dioxygen on the stability of the Cr2O3 surface: A first-
Metallurgical Industry Press, Beijing. principles investigation. Surf. Sci. 666, 44. https://doi.org/10.1016/j.
Goh, S.W., Buckley, A.N., Lamb, R.N., Rosenberg, R.A., Moran, D., 2006. The oxidation susc.2017.08.005.
states of copper and iron in mineral sulfides, and the oxides formed on initial Sun, C.Q., 2014. Compounds and nanocomposites: hetero-coordination. In: Relaxation of
exposure of chalcopyrite and bornite to air. Geochim. Cosmochim. Acta 70, the Chemical Bond (Springer Series in Chemical Physics, pp. 621-646). Singapore:
2210–2228. https://doi.org/10.1016/j.gca.2006.02.007. Springer Singapore.
Harada, T., 1964. Effects of oxidation of pyrrhotite, pyrite and marcasite on their Sun, C.Q., 2020. Electron and Phonon Spectrometrics. Springer, Singapore.
flotation properties: fundamental studies on flotation of iron sulphide minerals (2nd Sutherland, K.L., Wark, I.W., 1955. Principles of flotation. Australasian Institute of
Report). J. Min. Inst. Japan. 80, 669–674. https://doi.org/10.2473/ Mining and Metallurgy.
shigentosozai1953.80.914_669. Taggart, A.F., Del Guidice, G.R.M., Ziehl, O.A., 1934. The case for the chemical theory of
Hounfodji, J.W., Kanhounnon, W.G., Kpotin, G., Atohoun, G.S., Lainé, J., Foucaud, Y., flotation. Trans. AIME 112, 32.
Badawi, M., 2020. Molecular insights on the adsorption of some pharmaceutical Vacassy, R., Scholz-Odermatt, S., Dutta, J., Plummer, C., Houriet, R., Hofmann, H., 2010.
residues from wastewater on kaolinite surfaces. Chem. Eng. J. (Lausanne, Synthesis of controlled spherical zinc sulfide particles by precipitation from
Switzerland: 1996), 407. https://doi.org/10.1016/j.cej.2020.127176. homogeneous solutions. J. Am. Ceram. Soc. 81, 2699–2705. https://doi.org/
Kakovsky, I.A., 1957. Physicochemical properties of some flotation reagents and their 10.1111/j.1151-2916.1998.tb02679.x.
salts with ions of heavy non-ferrous metals, in: Proceedings of the Second Vanderbilt, D., 1990. Soft self-consistent pseudopotentials in generalized eigenvalue
International Congress of Surface Activity, London. IV: 225-237. formalism. Physical Review B 41 (11), 7892–7895. https://doi.org/10.1103/
Kakovsky, I.A., Grebnev, A.N., 1959. The concepts of flotation critical pH. Ore Dressing PhysRevB.41.7892.
5, 6–10 (in Russian). Wang, F., 2008. Fundamental research of flotation on typical sulphides/carbonates/
Keller, C.H., Lewis, C.P., U.S. Pat., 1554216 and 1554220, 1925. oxides of Cu, Pb, Zn and Fe, A Dissertation in Mineral Processing. North East
Li, Y.Q., Chen, J.H., Jin, G., 2011. DFT study of influences of As, Co and Ni impurities on University. (in Chinese). https://doi.org/10.7666/d.y1844586.
pyrite (1 0 0) surface oxidation by O2 molecule. Chem. Phys. Lett. 511 (4–6), Wark, I.W., Cox, A.B., 1934a. Principles of flotation. I. an experimental study of the effect
389–392. https://doi.org/10.1016/j.cplett.2011.06.078. of xanthates on contact angles at mineral surfaces. Trans. AIME 112, 189–232.
Liu, J., Wen, S., Chen, X., Bai, S., Liu, D., Cao, Q., 2013. DFT computation of Cu Wark, I.W., Cox, A.B., 1934b. Principles of flotation, II-an experimental study of the
adsorption on the S atoms of sphalerite (110) surface. Miner. Eng. 47, 1–5. https:// influence of cyanide, alkalis and copper sulphate on the effect of potassium ethyl
doi.org/10.1016/j.mineng.2013.03.026. xanthate at mineral surfaces. Trans. AIME 112, 245–266.
Mellgren, O., 1966. Heat of adsorption and surface reactions of potassium ethyl xanthate Wark, I.W., 1994. The chemical basis of flotation, in: Proceedings of the Australasian
on galena. Trans. AIME 235, 46–59. Institute of Mining and Metallurgy. Parkville, Vic.: The Institute, 90, 61-72.
Woods, R., 1971. Electrochemistry of sulphide flotation. Australian Mining 63, 68.

15

You might also like