Computational simulation of adsorption and thermodynamic study of xanthate, dithiophosphate and dithiocarbamate on galena and pyrite surfaces

Minerals Engineering 46–47 (2013) 136–143

Contents lists available at SciVerse ScienceDirect

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

Computational simulation of adsorption and thermodynamic study of xanthate, dithiophosphate and dithiocarbamate on galena and pyrite surfaces Jianhua Chen a, Lihong Lan b,c, Ye Chen a,⇑ a

College of Resources and Metallurgy, Guangxi University, Nanning 530004, PR China College of Chemistry and Chemical Engineering, Guangxi University, Nanning 530004, PR China c School of Chemistry and Chemical Engineering, Key Laboratory of Chemical and Biological Transformation Process of Guangxi Higher Education Institutes, Guangxi University for Nationalities, Nanning 530006, PR China b

a r t i c l e

i n f o

Article history: Received 1 November 2012 Accepted 18 March 2013 Available online 6 May 2013 Keywords: Collector adsorption Computational simulation DFT Microcalorimetry

a b s t r a c t Computer modeling and the microcalorimetry method were employed to investigate the adsorption of xanthate, dithiophosphate (DTP) and dithiocarbamate (DTC) on the galena and pyrite surfaces. The calculated results show that the pyrite surface undergoes greater relaxation than galena, while galena has a more electronegative surface than pyrite. The pyrite Fe atom is more active than the galena Pb atom. The simulations of adsorption show that the adsorbates coordinate mainly to the surfaces through interaction between their S atoms with the surface Pb or Fe atoms. The analysis of the density of states (DOSs) suggests that the adsorption of xanthate on the pyrite surface is stronger than that on the galena surface, while that of DTP and DTC are stronger on the galena surface than on the pyrite surface. The heat of adsorption and kinetics parameters of DTC and DTP at the galena and pyrite surfaces differ greatly, suggesting that DTC and DTP exhibit good selectivity in the separation of pyrite and galena, while these two parameters for xanthate at the two minerals differ little, indicating the poor selectivity of xanthate. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction The selectivity of a collector is an essential prerequisite to achieve high quality concentrate and great recovery in the separation of complex sulfide ores (Taggart, 1945). Thiol collectors such as xanthate, dithiophosphate (DTP) and dithiocarbamate (DTC) are widely employed in sulfide minerals flotation. Xanthate is generally used in bulk flotation due to its powerful collecting property, but would increase the subsequent separation difficulty due to a lack of selectivity. In contrast, dithiophosphate and dithiocarbamate are usually used in selective flotation due to their good selectivity, especially in lead and copper sulfide mineral flotation. Although several mechanisms for the adsorption of these three collectors on sulfide minerals have been reported (Chander and Fuerstenau, 1974; Bradshaw et al., 1995; Crozier, 1991; Fuerstenau, 1990; Raju and Forsling, 1991; Guler et al., 2006). However, the micro-mechanisms of the selectivity of these three thiol collectors on different sulfide mineral surfaces are not fully understood. In the adsorption process, the heat of adsorption could characterize the intensity of adsorption between the reagent and mineral (Glembotskii, 1981). The kinetics and thermochemistry of the xanthate adsorption reaction on pyrite and marcasite were evaluated ⇑ Corresponding author. Tel.: +86 771 3232200; fax: +86 771 3233566. E-mail address: [email protected] (Y. Chen). 0892-6875/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.mineng.2013.03.015

by Haung and Miller (1978), and the rate of the adsorption reaction was found to be approximately one-half order with respect to the xanthate concentration. Mellgren (1966) investigated the heat of adsorption of ethyl xanthate on galena samples ‘‘as ground’’ or previously treated with potassium carbonate, sulfate or thiosulfate solutions and concluded that an ion exchange mechanism might be involved in the adsorption process. Sutherland and Wark (1955) reported that the adsorption of diethyl dithiocarbamate onto sulfide minerals is faster than that of ethyl xanthate and they attributed this phenomenon to the lower solubility of the dithiocarbamates salt. A computational method such as the density functional theory (DFT) method is an effective tool to acquire the microscopic details of configuration and electronic structure of the reagent adsorption on mineral surface, and to illuminate the fundamental aspects of adsorption at the atomic level. Hung et al. (2003) employed the DFT method to study the xanthate adsorption on the pyrite (1 1 0) and (1 1 1) surfaces. Yekeler and Yekeler (2004) presented the DFT results from the investigation of the structural properties and the frontier orbital energies of some thiol collectors and their interactions with the Ag (+1) ion. Yekeler and Yekeler (2006) calculated the relevant molecular properties of 2-mercaptobenzothiazole and its 6-methyl and 6-methoxy derivatives as collector using the DFT. Moreover, our previous works on the application of the DFT approach to the study of pyrite, galena and sphalerite

J. Chen et al. / Minerals Engineering 46–47 (2013) 136–143

surfaces (Chen and Chen, 2010a,b; Li et al., 2011a) have provided a foundation for the current investigation of collector adsorption. In this paper, simulations of three thiol collectors including xanthate, dithiophosphate (DTP) and dithiocarbamate (DTC) adsorptions on PbS (1 0 0) and FeS2 (1 0 0) surfaces were modeled by DFT. Meanwhile, the kinetics and thermochemistry of these three thiol collectors on galena and pyrite were measured by the microcalorimetry method. The computational simulation and thermodynamic results can provide the useful information for a better understanding of the interactions between different collectors and different minerals, and give insights into the essential distinction of the selectivity of different collectors. 2. Experimental and computational methods 2.1. Materials Single crystal galena and pyrite samples obtained from Fankou Mine, Guangdong province, China, were used in this experiment. Multi-element analysis shows that the galena and pyrite samples were of high purity only with a trace of the element Sb and Co. The XRD data (Fig. 1) also confirmed the results. The diethyl dithiocarbamate (DDTC) in analytical grade was purchased from Sinopharm Chemical Reagent Co., Ltd., and ammonium dibutyl dithiophosphate (ADDTP) in industrial grade was purchased from Zhuzhou Flotation Reagents Factory. Sodium butyl xanthate (SBX) was synthesized by reacting butyl alcohol with sodium hydroxide and carbon disulfide. 2.2. Microcalorimetry The microcalorimetric measurements were accomplished with an RD496-III type microcalorimeter. The detailed description about the structure and the technical parameters of this calorimeter has been given elsewhere (Ji et al., 2001). The operation temperature was kept constant at 298.15 K. Heat of the adsorption reaction was measured in an isothermal reactor. Firstly, 1.0 g mineral samples and 20 mL of distilled water were put into the erlenmeyer flask (50 mL). The mixture was vibrated ultrasonically for 5 min, then allowed to stand for a few minutes until the liquid was divided into two layers. The supernatant was replaced with 20 mL of distilled water to prepare the mineral solution sample. The concentrations of collector solution (SBX, ADDTP and DDTC) were 2  104 mol/L. Secondly, 1 mL of collector solution and 1 mL of

(b)

(a)

FeS2

PbS

137

mineral solution samples were injected respectively in a small cell (3 mL) and a big cell (6 mL) by a micro-injector. Then the small cell was put into the big cell, which were all put into a 15 mL stainless steel sleeve together. The sealed stainless steel sleeve was put into the calorimeter. Then the reaction parameters were set. When the baseline was stable, the small cell was pierced and distilled water flowed into the big cell. So the distilled water was mixed with mineral sample in the big cell. The thermal effect was then recorded automatically by a computer. 2.3. Computational method The simulations of collector adsorption have been done using the Cambridge Serial Total Energy Package (CASTEP) developed by Payne et al. (1992), which is a first-principle pseudopotential method based on density functional theory (DFT). The DFT calculation employed plane wave (PW) basis sets and ultrasoft pseudopotentials. The exchange–correlation functional applied was the generalized gradient approach (GGA) of Perdew and Wang (PW91) (Perdew and Wang, 1992; Vanderbilt, 1990). The interactions between valence electrons and ionic core were represented by ultrasoft pseudopotentials. Valence electrons configurations considered in the study included Pb 5d10 6s2 6p2, S 3s2 3p4, Fe 3d6 4s2, P 3s2 3p3, N 2s2 2p3, C 2s2 2p2, O 2s2 2p4, and H 1s1 states. The plane wave cutoff energies of 280 eV and 270 eV were used for galena and pyrite calculations, respectively. The surface Brillouin zone was sampled with a 1  2  1 k-point grid for PbS surface calculation and a 2  2  1 grid for FeS2 surface calculation (Monkhorst and Pack, 1976), which shows that the cutoff energy and the k-point meshes are sufficient for the system. For self-consistent electronic minimization, the Pulay Density Mixing method was employed with the convergence tolerance of 2.0  106 eV/atom. The energy tolerance was 2.0  105 eV/atom, the force tolerance was 0.08 eV/Å, and the displacement tolerance was 0.002 Å. The models xanthate (HOCS2), dithiophosphate (H2O2PS2) and dithiocarbamate (H2NCS2) were used in place of butyl xanthate, ammonium dibutyl dithiophosphate and diethyldithiocarbamate for efficient computation as the pre-tested results showed that the head groups gave little effect to property. The optimizations of xanthate (HOCS2), DTP (H2O2PS2) and DTC (H2NCS2) were calculated in a 15  15  15 Å cubic cell and the optimizations were performed at the gamma point in the Brillouin zone. After testing the slab thickness and vacuum slab thickness, we constructed a (4  2) PbS (1 0 0) surface with 8 atomic layers and 10 Å vacuum slab and a (2  2) FeS2 (1 0 0) surface with 15 atomic layers and 10 Å vacuum slab (Fig. 2) For the FeS2 surface, the six outermost atomic layers of the substrate were allowed to relax while the nine bottom-most atomic layers of the substrate were fixed to the bulk coordinates, and for the PbS surface the three outermost atomic layers of the substrate were allowed to relax while the five bottom-most atomic layers of the substrate were fixed to the bulk coordinates in the adsorption calculations. 3. Results and discussion 3.1. The electronic structure and properties of galena (1 0 0) and pyrite (1 0 0) surfaces

10

20

30

40

2 θ (°)

50

60

70

10

20

30

40

50

2 θ (°)

Fig. 1. XRD diffraction patterns of galena and pyrite.

60

70

In froth flotation, the mineral surface plays an important role in the process of reagent adsorption, which determines the geometry and strength of adsorption and the difference in properties of mineral surfaces is the premise of mineral separation by flotation. We employed the quantum method to model the galena (1 0 0) and the pyrite (1 0 0) surfaces and investigate the electronic structure and property of the surfaces. The results show that the dissociation of

138

J. Chen et al. / Minerals Engineering 46–47 (2013) 136–143

the deep layers (Pb5 and Pb7), the DOS of the surface Pb3 6p state is slightly stronger. It is clearly noted that the DOS of S1 and S3 atoms are quite different from that of S5 and S7 atoms, indicating that electronic properties of surface atoms are different from the deep layer atoms. Moreover, the PbS surface S1 and S3 states dominate around the Fermi level, while the S5 and S7 states have few contributions to the Fermi energy, suggesting that surface S atoms show greater reactivity than the bulk S atoms. For the pyrite surface, by comparison with the bulk Fe12 layer, the DOS peak of the outermost Fe3 states is obviously enhanced around 2.0 eV and Fe 3d state in the conduction band increases from one peak to two peaks. Compared to the deep layer S atoms (S5, S7 and S9), the surface S 3p states (S1 and S4) around 2.5 to 0 eV increase obviously, especially at 2.0 eV. By comparison to PbS, the Fe atom shows a greater sharp DOS peak at the Fermi level, indicating that the FeS2 Fe atom is more reactive than the PbS Pb atom.

Fig. 2. The slab models of a (4  2) galena (1 0 0) surface (a) and a (2  2) pyrite (1 0 0) surface.

a galena surface results in the breakage of Pb–S bonds, and the Pb and S atoms are changed from six-coordinated as in the bulk to five-coordinated in the surface. The pyrite Fe and S atoms are five-coordinated and three-coordinated respectively at the surface instead of six in the bulk. The decrease in the coordination number of surface atoms may lead to the variation of their reactivity. Surface relaxation of galena (1 0 0) and pyrite (1 0 0) surfaces had been discussed in our previous study (Li et al., 2011b) and the results showed that the top three layers of both galena and pyrite surfaces underwent the surface relaxation, however, pyrite underwent greater relaxation than galena. The Mulliken electron of each layer of PbS (1 0 0) and FeS2 (1 0 0) surfaces are shown in Fig. 3. The electron distributions for galena and pyrite are quite different. The outermost layer of both PbS and FeS2 surfaces is electronegative, but PbS carries more negative charge than FeS2 surface. The density of states (DOSs) of the galena Pb and S atoms and the pyrite Fe and S atoms are shown in Figs. 4 and 5, where the numeral beside the symbol of element represents the layer number. For example, S1 represents S atom in the 1st layer. Compared with

3.2. The geometry and electron density of collector adsorption To further investigate the differences of xanthate, DTP and DTC reacting with galena or pyrite, the adsorption of three thiol collectors on galena or pyrite surface has been simulated by the DFT method. Fig. 6 demonstrates the electron density of xanthate adsorption on the galena or pyrite surface, in which, the S1 and S2 refer to the sulfur atom with single bond and double bond, respectively. On the galena surface, the distances between the xanthate S1 and the surface Pb1 is 2.863 Å, which is close to the radius between Pb–S of 2.840 Å, while that of the xanthate S2 and the surface Pb2 is 2.935 Å, which is larger than 2.840 Å. It is indicating that the adsorption of xanthate on the PbS surface is mainly via bond formation between the xanthate S1 and the surface Pb1 atom. In addition, a strong electron overlap across S1–Pb1 and a weak electron overlap of S2–Pb2 are clearly visible in Fig. 6a. On the pyrite surface, distances between the S1–Fe1 and S2–Fe2 are 2.284 Å and 2.281 Å, respectively, which are smaller than the atomic radius of S–Fe (2.31 Å), indicating that the interaction between xanthate and pyrite is via the bonding of the two xanthate S atoms with the surface Fe atoms. Furthermore, charge overlaps between S1– Fe1 and S2–Fe2 are observed in Fig. 6b. It may explain the reason for the greater affinity of xanthate for the pyrite.

Fig. 3. Mulliken electron of each layer of galena (1 0 0) and pyrite (1 0 0) surfaces.

139

J. Chen et al. / Minerals Engineering 46–47 (2013) 136–143 1.0

Pb7

Pb 6s

Density of states (electrons/eV)

Density of states (electrons/eV)

Pb5

Pb 6s

Pb 6p

0.5

0.0

Pb3

0.0

Pb 6p

Pb 6s

0.5

Pb1 Pb 6p

Pb 6s

EF

S7

1 0.0

3p

3s

2

Pb 6p 0.5

0 2

3s

3p

S5

1 0 2

3p 3s

S3 1 0 2

3p

3s

S1

0.5 1 0.0 -10

-8

-6

-4

-2

0

0 -14

2

-12

-10

-8

Energy/eV

-6

-4

-2

0

Energy/eV

Fig. 4. DOS of galena Pb and S atoms in different layers.

s

1

Ef Fe 3d

Bulk Fe

S9

E

F

p

0

Density of states (electrons/eV)

Density of states (electrons/eV)

2 7 6 5 4 3 2 1 0 6 5 4 3 2 1 0 6 5 4 3 2 1 0

Fe7

1

S7

0 1

S5

0 1

S4

0

Fe3

1

S1

0 -4

-2

0

2

-15

-10

-5

0

Energy/eV

Energy/eV Fig. 5. DOS of pyrite Fe and S atoms in different layers.

Fig. 6. Electron density maps of xanthate adsorption on (a) PbS surface and (b) FeS2 surface.

Electron density maps of DTP adsorption on the PbS and FeS2 surface are shown in Fig. 7. On the PbS surface, the lengths of

Fig. 7. Electron density maps of DTP adsorption on (a) PbS surface and (b) FeS2 surface.

S1–Pb1 and S2–Pb2 are 2.860 Å and 2.881 Å, suggesting that the interaction between DTP and PbS is via the two S atoms in the

140

J. Chen et al. / Minerals Engineering 46–47 (2013) 136–143

DTP group bonding with the surface Pb atoms. The electron density suggests that the bonding of S1–Pb1 is little stronger than S2–Pb2 as shown in Fig. 7a. On the FeS2 surface, the lengths of S1–Fe1 and S2–Fe2 are 2.324 Å and 2.376 Å, which are a little longer than the radius of S–Fe (2.310 Å), indicating that the reactivity of DTP on the FeS2 surface is relatively weaker than xanthate. Electron density maps of DTC adsorption on the PbS and the FeS2 surfaces are given in Fig. 8. On the PbS surface, the lengths of S1–Pb1 and S2–Pb2 are 2.863 Å and 2.862 Å, suggesting that the interaction between DTC and PbS is via the two S atoms in the DTC group bonding with the surface Pb atoms. Compared with electron density maps of xanthate (Fig. 6a) and DTP (Fig. 7a), it is noted that the adsorption of DTC on galena is the strongest, which may lead to the greatest value of heat of adsorption. On the FeS2 surface, the lengths of S1–Fe1 and S2–Fe2 are 2.271 Å and 2.285 Å, which are smaller than the radius of S–Fe (2.310 Å), indicating strong bond formation between S1–Fe1 and S2–Fe2, which is also observed in the electron density map (Fig. 8b).

3.3. The analysis of density of states The analysis from density of states (DOSs) could give an insight into the interaction between different collectors and different minerals at the electronic level. The DOSs of the three reagents on the PbS and FeS2 surfaces before and after adsorption have been calculated by the CASTEP module and the Fermi energy is set to 0 eV (EF = 0) in the DOS curve. In this section, we will discuss the DOS from two aspects: one is the collector and the other is the mineral surface. The DOS of xanthate, DTP and DTC on the PbS and FeS2 surfaces before and after adsorption are shown in Fig. 9a–c, respectively. Before adsorption, it is shown that the partial DOSs of xanthate, DTP and DTC functional groups are similar near the Fermi level, which are composed of the S 3p orbital, indicating that the S 3p orbital is very active. In addition, the DOS of a S atom with a single bond is the same as the S atom with a double bond, indicating that the two S atoms in the thiol group have similar chemical reactivity, which may be ascribed to the conjugation effect of a pi bond. The main differences of DOS for the three reagents are caused by the O, C, P and N atoms in the functional group. As the P 6p and P 6s orbital are located in the deep site of the conduction and valence band, the influence of the P atom on the adsorption is small. After adsorption on mineral surfaces, DOSs of the three collectors change obviously. For xanthate adsorption, the DOS of xanthate changes more notably on the FeS2 surface than on the PbS surface, and especially the S 3p state depletes largely near the

Fermi level. It is indicated that the reaction of xanthate on the FeS2 surface is stronger than that on the PbS surface. For DTP adsorbed on the PbS surface, the two peaks of DTP S 3p state combine into one peak and a dramatic change near the Fermi level (EF = 0) occurs where the DOS peak of S 3p reduces to almost zero. On the FeS2 surface, although the S 3p state at the Fermi level decreases largely, the change of S 3p state is not as dramatic as that on the PbS surface. It is suggested that the adsorption of DTP on the PbS surface is stronger than that on the FeS2 surface. After adsorption on the FeS2 surface, the DTC S 3p state shifted to a lower energy level and the DOS peak reduced largely at the Fermi level, meanwhile, the three DOS peaks of C 2p and N 2p located at 3 to 6 eV merge into one DOS peak. On the PbS surface, the DTC S 3p state moves to the valence band and the two DOS peaks of S 3p merge into one peak, meanwhile the two DOS peaks of C 2p and N 2p located at 3 to 5 eV merge into one peak. The DOS of DTC on the PbS surface changes more obviously than that on the FeS2 surface, indicating that the adsorption of DTC on galena is stronger. DOSs of the PbS surface Pb atom and the FeS2 surface Fe atom before and after adsorption are demonstrated in Fig. 10. For the PbS surface, the Pb 6p and 6s states have changed after interaction with the collector, suggesting that the Pb atom participated in the bonding interaction. After interacting with DTC, the Pb 6s state has the highest DOS peak value at the Fermi level and shows the largest movement to the conduction band. In addition, a new DOS peak of the Pb 6p state appears at 4.2 eV. It is suggested that the interaction between DTC and PbS is the strongest. Due to the similar structure for xanthate (OCSS) and DTP functional groups (O2PSS), the DOSs of the Pb atom interacting with xanthate and DTP are similar, and there are only little differences in the Pb 6s state. After interacting with xanthate, the peak value of Pb 6s at the Fermi level is higher than that of Pb 6s interacting with DTP, indicating that the interaction of xanthate with the PbS is stronger than DTP. For the pyrite surface, the DOS of the Fe 3d orbital changes after interaction with the reagent, indicating that the Fe atom participates in the bonding interaction. Before adsorption, Fe 3d orbital splits into two peaks (named as p1 and p2, respectively) at the top of valence band. After adsorption, a significant change of p1 and p2 occurs. After interaction with the DTC, the change of Fe DOS is the largest, then xanthate and DTP. It can be found that the p1 and p2 of Fe DOS almost completely disappeared after the DTC interaction, while the change in p1 and p2 is relatively small for xanthate and DTP. 3.4. The heat of adsorption of collectors on galena and pyrite surfaces For an adsorption process, the Gibbs free energy can be defined (Eq. (1)):

DG ¼ DH  T DS

Fig. 8. Electron density maps of DTC adsorption on (a) PbS surface and (b) FeS2 surface.

ð1Þ

in which DH is the enthalpy, T is the temperature, DS is the entropy, a measure of disorder. For an adsorption process, since it proceeds from disorder into an orderly process, DS is less than zero and consequently TDS, the second item in Eq. (1), is greater than zero. Therefore, DS makes a negative contribution to DG, the numerical the value of the Gibbs free energy depends mainly on enthalpy DH, and the greater the value of the DH term, the more negative the value of DG becomes, and the more easily the adsorption reaction occurs. The heats of butyl xanthate (BX), dithiophosphate (DTP) and dithiocarbamate (DTC) adsorption on galena and pyrite surfaces are listed in Table 1. For the pyrite, BX shows the strongest collecting capacity, and followed by DTC, while that for DTP is the weakest. For the galena, DTC exhibits the strongest collecting capacity, then the BX and DTP. The DOS results show that the interaction be-

141

J. Chen et al. / Minerals Engineering 46–47 (2013) 136–143

(a)

6

EF

S 3p

-2

S 3p

O 2p S 2s

3

0

2

Free xanthate

S 3p

O 2p S 3p C 2p

C 2p

4

C 2p

0 -8

-6

-4

-2

0

2

6

4

Xanthate on FeS2

Density of states(eletrons/eV)

Density of states (eletrons/eV)

6

-4

S 3p

3

DTP on PbS

3

0 -6

EF

S 3p O 2p O 2p

C 2p

3

-8

(b)

6

Xanthate on PbS

0

-8

-7

-6

-5

-4

-3

O 2p S 3p

6

-2

-1

O 2p

0

1

2

3

4

P 6s

3

5

6

7

Free DTP

S 3p

P 6p P 6s

S 3p

O 2s 0

-8

-7

6

-6

-5

O 2p

-4

-3

O 2p

-2

-1

0

1

2

3

4

5

6

7

4

5

6

7

DTP on FeS2

S 3p

3

0

0

-8

-6

-4

-2

0

2

4

-8

Energy/eV

6

(c)

-7

-6

-5

-4

-3

-2

-1

0

1

2

3

Energy/eV

EF

S 3p

DTC on PbS

Density of states (eletrons/eV)

3

0 -6

-4

-2

0

N 2p

6

2

4

Free DTC

S 3p

C 2p S 3p S 3s C 2s

C 2p

S 3p

C 2p S 3p

C 2p

3

0 -6

-4

-2

6

0

2

4

DTC on FeS2

S 3p

3

0 -6

-4

-2

0

2

4

Energy/eV Fig. 9. DOS of xanthate (a), DTP (b) and DTC (c) on PbS and FeS2 surface before and after adsorption.

tween xanthate S 3p and pyrite Fe 3d is stronger than that with galena Pb 6sp, while for DTP and DTC, the interaction between their S 3p and the galena Pb 6sp is stronger than that with the pyrite Fe 3d. The activity differences in the S 3p in the three function groups may be caused by the atoms bonding to the S atom as analyzed before. The difference of DH between PbS and FeS2 could represent the selectivity of the collector, and the greater the absolute value of the difference of DH, the selectivity of the collector is stronger. It is found that the absolute value of the difference of DH for DTC is the greatest (1.538 J/m2), then the DTP (1.033 J/m2) and xanthate (0.991 J/m2), indicating that DTC and DTP will show a better selectivity than xanthate in the separation of galena and pyrite. 3.5. Kinetics of collectors adsorption on galena and pyrite surface The reaction rate constant and the reaction order of the adsorption reaction could be calculated by the following equation, according to the reference (Gao et al., 2002).

 ln

1 dHi  H0 dt



  Hi ¼ ln k þ n ln 1  H0

ð2Þ

    i was plotted as a function of ln 1  HH0i , and the intercept ln H10  dH dt of the curve is ln k, and then the values of reaction rate constant k and reaction order n are obtained and are listed in Table 2. As we all know, the reaction rate constant quantifies the speed of a chemical reaction and the reaction order can tell the relationship between the concentrations of species and the rate of a reaction. It is noted that the reaction order for butyl xanthate on the pyrite surface is 0.58, which is very close to that of 0.60 measured by Haung and Miller (1978), indicating that the results are reliable. As shown in Table 2, the value of k for BX on the PbS is larger than that on the FeS2, indicating that the adsorption rate for BX on the PbS is faster. In addition, the order of both reactions is approximately one-half, suggesting that the influences of xanthate concentration on the two adsorption processes are similar. The reaction order of DTP on PbS is unity (1.03) while that on FeS2 is one-half order (0.42), and the reaction rate constant of DTP on PbS is also greater than that on FeS2, suggesting that chemical kinetics of DTP favors PbS surface adsorption. For DTC, the reaction rate constant on PbS is smaller than that on FeS2, indicating that the adsorption reaction of DTC on the FeS2 surface is faster than that on the PbS surface. In addition, the reaction order of DTC on PbS is unity (1.34) while that on

142

J. Chen et al. / Minerals Engineering 46–47 (2013) 136–143

(a)

EF

Pb adsorbed DTC 2

Pb 6s

Density of states (eletrons/eV)

Pb 6p

0.4 0.2 0.0 0.0

0.1

0.2

0.3

0.4

0.3

0.4

0 -4 2

-2

0

Pb adsorbed DTP

2

0.4

4. Conclusions

0.2

Pb 6s

Pb 6p

0.0 0.0

0.1

0.2

0 -4

-2

0

2 Pb adsorbed xanthate

0.4

2

0.2

Pb 6s

Pb 6p

0.0 0.0

0.1

0.2

0.3

0.4

0.3

0.4

0 -4 2

-2

0

Pb before adsorption

2

0.4 0.2

Pb 6p

Pb 6s

0.0 0.0

0.1

0.2

0 -4

-2

0

2

Energy/eV

(b) 15

EF

Fe adsorbed DTC

p2 p1

10

Fe 3d

5

Density of states (eletrons/eV)

FeS2 is one-half (0.53). It is suggested that the DTC adsorption on galena is more likely be affected by the concentration than that on pyrite. The flotation practice also shows that galena could be collected by adding a small amount of DTC and the selectivity of DTC will drop in the presence of a large concentration.

15 0 -5 10

-4

-3

-2

Fe adsorbed DTP

-1

p2 p10

1

2

p2 p10

1

2

Fe 3d

5 0 15 -5

-4

-3

-2

-1

Fe adsorbed xanthate

10

In this work, the differences of xanthate, DTP and DTC adsorption on galena and pyrite have been discussed from four aspects: the surface electronic structure and properties, adsorption geometry, heat of adsorption and the kinetics of adsorption, and the conclusions can be drawn as following. The electronic structure and surface properties of PbS (1 0 0) and FeS2 (1 0 0) vary in the surface coordination, Mulliken charge and the density of states (DOSs). Moreover, the pyrite Fe atom is more active than the galena Pb atom. The adsorption results suggest that the interactions of three thiol collectors are via their S atoms bonding with the surface Pb or Fe atoms. The DOS of xanthate changes more apparently on a FeS2 surface than on a PbS surface, indicating that the adsorption of xanthate on FeS2 surface is stronger than that on PbS surface. However, the DOSs of DTP and DTC changes more obviously on a PbS surface than on a FeS2 surface, indicating that the interactions of DTP and DTC with PbS are stronger than that with FeS2. The results from thermodynamics and kinetics show that the difference in adsorption behaviors of xanthate on the galena and pyrite surfaces is relatively small while that for DTC and DTP differ greatly at the two mineral surfaces, which could explain the difference of these three collectors in selectivity.

Fe 3d 5

Acknowledgements

0 15 -5 10

-4

-3

-2

Fe before adsorption

-1 p2 p1 0

1

2

-1

1

2

Fe 3d

5 0 -5

-4

-3

-2

0

Energy/eV

This research was funded by National Natural Science Foundation of the People’s Republic of China (No. 51164001), New Century Excellent Talents in University (No. NCET-11-0925), Guangxi Natural Science Foundation (2011GXNSFB018010) and Scientific Research Foundation of Guangxi University, China (XBZ100459). The authors are thankful for these supports.

Fig. 10. DOS of PbS and FeS2 surface before and after adsorption.

References Table 1 The heat of adsorption values of butyl xanthate (BX), DTP and DTC adsorption on galena and pyrite surfaces.

DH (J/m2)

PbS FeS2 Difference of DH between PbS and FeS2

BX

DTP

DTC

2.976 1.985 0.991

2.122 1.089 1.033

3.219 1.681 1.538

Table 2 Kinetic parameters of BX, DTP and DTC adsorption on PbS and FeS2 surface. Collector

Mineral

k  103/s

n

R2

Butyl xanthate (BX)

Galena Pyrite

2.27 1.30

0.26 0.58

0.9938 0.9792

Dithiophosphate (DTP)

Galena Pyrite

3.17 2.64

1.03 0.42

0.9901 0.9858

Dithiocarbamate (DTC)

Galena Pyrite

0.55 2.56

1.34 0.53

0.9911 0.9390

Bradshaw, D.J., Cruywagen, J.J., O’Connor, C.T., 1995. Thermochemical measurements of the surface reactions of sodium cyclohexyl-dithiocarbamate, potassium n-butyl xanthate, and a thiol mixture with pyrite. Minerals Engineering 8 (10), 1175–1184. Chander, S., Fuerstenau, D.W., 1974. The effect of potassium diethyl dithiophosphate on the electrochemical properties of platinum, copper and copper sulphide in aqueous solution. Journal of Electroanalytical Chemistry and Interfacial Chemistry 56, 217–247. Chen, J.H., Chen, Y., 2010a. A first-principle study of the effect of vacancy defects and impurities on the adsorption of O2 on sphalerite surfaces. Colloids and Surfaces A: Physicochemical and Engineering Aspects 363, 56–63. Chen, Y., Chen, J.H., 2010b. The first-principle study of the effect of lattice impurity on adsorption of CN on sphalerite surface. Minerals Engineering 23, 676–684. Crozier, R.D., 1991. Sulphide collector mineral bonding and the mechanism of flotation. Minerals Engineering 4 (7–11), 839–858. Fuerstenau, M.C., 1990. Thiol collector adsorption processes. International Journal of Mineral Processing 29, 89–98. Guler, T., Hicyilmaz, C., Gokagac, G., Emekci, Z., 2006. Adsorption of dithiophosphate and dithiophosphinate on chalcopyrite. Minerals Engineering 19, 62–71. Glembotskii, B.A., 1981. Foundation of Physical Chemistry in the Process of Flotation. Metallurgical Industry Press, Beijing, pp. 191–198. Gao, S.L., Chen, S.P., Hu, R.Z., Li, H.Y., Shi, Q.Z., 2002. Derivation and application of thermodynamic equations. Chinese Journal of Inorganic Chemistry 18 (4), 362– 366. Haung, H.H., Miller, J.D., 1978. Kinetics and thermochemistry of amyl xanthate adsorption by pyrite and marcasite. International Journal of Mineral Processing 5, 241–266. Hung, A., Yarovsky, I., Russo, S.P., 2003. Density-functional theory studies of xanthate adsorption on the pyrite FeS2 (1 1 0) and (1 1 1) surfaces. The Journal of Chemical Physics 118 (13), 6022–6029.

J. Chen et al. / Minerals Engineering 46–47 (2013) 136–143 Ji, M., Liu, M.Y., Gao, S.L., Shi, Q.Z., 2001. Instrumentation Science and Technology 29 (1), 53. Li, Y.Q., Chen, J.H., Guo, J., 2011a. DFT study of influences of As, Co and Ni impurities on pyrite (1 0 0) surface oxidation by O2 molecule. Chemical Physics Letters 511, 389–392. Li, Y.Q., Chen, J.H., Chen, Y., Guo, J., 2011b. Density functional theory calculation of surface properties of pyrite (1 0 0) with implications for flotation. The Chinese Journal of Nonferrous Metals 21 (4), 919–926. Monkhorst, H.J., Pack, J.D., 1976. Special points for Brillouin-zone integrations. Physical Review B 13, 5188–5192. Mellgren, O., 1966. Heat of adsorption and surface reactions of potassium ethyl xanthate on galena. Transactions Society of Mining Engineers 235, 46–59. Payne, M.C., Teter, M.P., Allan, D.C., Arias, T.A., Joannopoulos, J.D., 1992. Iterative minimization techniques for ab initio total energy calculation: molecular dynamics and conjugate gradients. Review of Modern, Physics 64, 1045–1097. Perdew, J.P., Wang, Y., 1992. Accurate and simple analytic representation of the electron-gas correlation energy. Physical Review B 45, 13244–13249.

143

Raju, B.G., Forsling, W., 1991. Adsorption mechanism of diethyl dithiocarbamate on covellite, cuprite and tenorite. Colloids and Surfaces 60, 53–69. Sutherland, K.L., Wark, I.W., 1955. Principles of Flotation. Australasian Institute of Mining and Metallurgy, Melbourne. pp. 95–101. Taggart, A.F., 1945. Handbook of Mineral Dressing, Ores and Industrial Minerals, 2nd ed. Wiley, New York. Vanderbilt, D., 1990. Soft self-consistent pseudopotentials in a generalized eigenvalue formalism. Physical Review B 41, 7892–7895. Yekeler, H., Yekeler, M., 2004. Reactivities of some thiol collectors and their interactions with Ag (+1) ion by molecular modeling. Applied Surface Science 236, 435–443. Yekeler, H., Yekeler, M., 2006. Predicting the efficiencies of 2mercaptobenzothiazole collectors used as chelating agents in flotation processes a density-functional study. Journal of Molecular Modeling 12, 763– 768.