- Email: [email protected]
Surface modification at the molecular level in mineral beneficiation
Int. J. Miner. Process. 72 (2003) 129 – 140 www.elsevier.com/locate/ijminpro
Surface modification at the molecular level in mineral beneficiation Ela Mielczarski *, Jerzy A. Mielczarski ‘‘Laboratoire Environnement et Mine´ralurgie’’, UMR 7569 du CNRS, INPL/ENSG, 15 Avenue du Charmois, BP 40, Vandoeuvre-le`s-Nancy, Cedex 5450, France Received 20 September 2002; received in revised form 25 October 2002; accepted 30 June 2003
Abstract Direct information at a molecular level about surface products, their structure and surface distribution is vitally important to perform an efficient separation processes for mineral beneficiation. This information is equally important for the modification of surface properties of industrial minerals for various specific applications such as: fillers, composites, weathering resistance, etc. The understanding of the mechanisms and kinetics of interaction of the first adsorbed organic molecules with mineral surfaces is the fundamental requirement to make possible the prediction, control and modification of the macroscopic surface properties that govern the efficiency of separation technologies or new material formulations. The developed infrared external reflection technique has a very unique ability to study interface phenomena at a molecular level on heterogeneous substrates. The variety, precision and reliability of information about interface phenomena delivered by this technique are superior to other single techniques. The experiments are fast and nondestructive. High sensitivity (starting from 20% of monolayer), in situ collected information in a multiphase system even in the region of a strong absorption of substrate, makes this technique a very valuable experimental tool. The complexity of the recorded reflection spectra, their sensitivity to any variations of the optical properties of all bulk and surface components and their spatial distribution in the system under investigation are in fact the major strength of the technique. In this paper, a few examples of application of this multidiagnostic technique for monitoring surface modifications of sulfide minerals in aqueous solutions will be discussed in detail. D 2003 Elsevier B.V. All rights reserved. Keywords: surface modification; minerals; molecular orientation; infrared spectroscopy; adsorption
1. Introduction The flotation behavior of minerals depends on the nature, structure and surface distribution of their surface hydrophobic and hydrophilic species. The possibility of monitoring surface phenomena at mo-
* Corresponding author. E-mail address: [email protected] (E. Mielczarski). 0301-7516/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0301-7516(03)00093-0
lecular and atomic levels at the interfaces of natural minerals contacted with aqueous solution is vitally important for the understanding of the mechanisms and kinetics of surface phenomena and then to design efficient flotation technologies. Improvement of selective separation of ore components is achieved by manipulation of solution conditions such as pH, Eh (aeration, reducing or oxidizing reagents), addition of different types of collectors at different concentrations and modifiers such as activators or depressants. All these changes in solution compositions are performed
130
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
in one aim—to make the valuable mineral components very hydrophobic while keeping gangue components hydrophilic. This explains why the information collected directly from mineral surfaces is so important for understanding kinetic and thermodynamic relationships between the solution conditions and the nature and structure of the outermost layers that play a crucial role in the surface modification for separation or other processes. Ores are usually complex, multicomponent, and heterogeneous systems and cannot be described by a single value. They interact with aqueous solution through complex dissolution/precipitation, adsorption/desorption or oxidation/reduction processes. Monitoring the formation of surface species at submonolayer and monolayer levels, and their structures, requires the application of the appropriate experimental techniques. Infrared spectroscopy is particularly well suited to determine surface composition at a molecular level. Infrared is functional group selective, so it is particularly appropriate to detect small changes of the molecular microenvironment properties as those emerging at the interfaces. Experiments can be performed in situ at both gas – mineral and aqueous solution – mineral interfaces. There is only very gentle interaction of the infrared beam with the examined sample. The recent instrumental development of infrared spectroscopy is contributing significantly to the increased emphasis being placed on molecular level surface characterization. To perform the proper interpretation of reflection spectra for a more detailed picture of the interfacial structure, it is vitally important to combine the spectroscopic measurements with a spectral simulation. The importance of such combination is reinforced by the anticipated sensitivity of surface infrared absorbance not only to surface concentration but also to adsorbate structure, molecular orientation, surface distribution, lateral interaction, surface diffusion, molecular recognition and so-called optical effects. Several unique advantages of the developed infrared external reflection technique are overviewed in this work. The different examples of sulfide mineral – aqueous solution systems for which detailed qualitative and quantitative evaluations of the produced surface layers were performed are discussed in this paper.
2. Experimental 2.1. Developing of the infrared external reflection technique One of the major problems to solve was developing proper experimental procedures for collecting representative and appropriate information directly from mineral surfaces contacted with aqueous solution of organic and inorganic species. The complexity of the system under investigation, heterogeneity of mineral samples and the produced surface species and their stability require application of the appropriate experimental techniques. During these studies, we developed an infrared external reflection technique that let us overcome experimental problems and collect reliable data to monitor and understand surface phenomena at a molecular level (Mielczarski, 1993). A schematic diagram of the interaction of electromagnetic waves with a simple three-phase system is presented in Fig. 1. For polarization perpendicular to the plane of the incident beam (s-polarization), there is only one electric vector, E?Y, parallel to substrate plane. Hence, only molecular groups of the adsorbed species having a dipole transition moment parallel to the interface in y direction can interact with the incident radiation and produce an absorbance band A?Y. For example, in the case of the adsorption of a xanthate molecule involving both sulfur atoms with the same distance from interface (Fig. 1) and assuming a uniaxial system (no difference in orientation in x –y plane), it is expected that the absorbance assigned to the SCS group, A?Y, will show the highest value. This results from the most favorable configuration for the interaction of the asymmetric stretching vibration of the SCS group and electric field vector E?Y. For a molecule inclined to interface the interaction is lower and the recorded absorbance will decrease depending on the orientation of this molecular group versus interface. No interaction and obviously no absorbance band is observed when the SCS group is turned 90j from the position presented in Fig. 1 when the dipole transition moment of the asymmetric vibration of the group becomes vertical to interface. For parallel polarization (p-polarization), there are two electric field vector components at the interface, one parallel EOX and one perpendicular EOZ to the
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
131
CH3 CH2
EII
νas
O C S
z
νas
E ⊥Y S
θ
EII Z
n1 , k1
ambient
n2 , k 2
adsorption layer
n3 , k3
mineral
EII X d
x y
Fig. 1. Schematic diagram of the interaction of electric field vectors in three directions with a simple three-phase system. Ethyl xanthate molecule with marked dipole transition moments of the asymmetric stretching vibrations of the SCS group (parallel to interface) and of the asymmetric stretching vibrations of the COC group (almost vertical to interface) is also presented.
substrate plane (Fig. 1). Therefore, for p-polarization, the molecular groups showing dipole transition moments parallel and vertical to interface in x and z directions can produce absorbance bands. Using the example of xanthate molecules at the surface presented in Fig. 1, two absorbance bands, the AOX for the SCS group vibration and the AOZ related to the COC group vibration, will be presented in the recorded spectrum. As it is indicated in Fig. 1, the asymmetric stretching vibration of the COC group occurs almost vertically to the surface. It is also possible to distinguish these two components because they show reverse absorbance, whereas one produces positive absorbance, the second one is negative or vice versa, depending on the angle of incidence (see the discussion below). The intensities of the absorbance band recorded for the two polarizations A? and AO depend on the nature of the vibration of the particular molecular group, the amount of surface species, the angle of the incident beam and the orientation of molecules in the surface layer. The absorbance components in three directions for particular wavenumber can be calculated as follows (Mielczarski and Yoon, 1989): 16p cosh n2 k2 d2 A?Y ¼ ln10 n23 1 k 2 16p cosh n n2 k2 d2 ANX ¼ 34 ln10 n32 n43 cos2 h k n3 16p cosh sin2 h n2 k2 d2 ANZ ¼ 2 4 2 2 ln10 n3 n3 cos h ðn2 þ k 22 Þ2 k
where n3=(nˆ3 n1sin2h)1/2, n3 = nˆ3 + ik3, h is the incident angle and k is the wavelength of the incident beam. The presented approximate equations are used to calculate absorbance components in three directions and then to compare them with the A? and AO experimental values that are recorded for different angles. The results obtained for a 2-nm-thick isotropic layer of polyethylene aliphatic chain for the stretching vibration of the CH2 group at 2920 cm 1 versus an angle of incidence are presented in
Fig. 2. Calculated absorbance components in three directions A?Y, AOX, AOZ for a three-phase system: air, n1 = 1, k1 = 0; aliphatic chain, n2 = 1.576, k2 = 0.246; substrate, n3 = 1.657, k3 = 0.0002.
132
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
Fig. 2. There are tremendous changes in the absorbance components just with variation of the angle of the incident beam as is shown in Fig. 2. For spolarization, only the negative absorbance band A?Y is observed and its intensity is highest at the lowest angle of incident. Therefore, for s-polarization, the best experimental conditions to record reflection spectrum are with an angle of incidence close to normal to the interface. For p-polarization, there are two absorbance components AOZ and AOX. Below the Brewster angle, which is for this system 58.7j, the AOZ component is negligible for angles of incidence between 0j and 20j, then it shows a positive value with strong increases of intensity in the vicinity of the Brewster angle. Above this angle, the AOZ component shows a negative value, at first very high, then decreasing quickly to zero at 90j. The AOX component has reverse behavior; it is negative below the Brewster angle and positive above this angle. Moreover, the AOX component is negative with the same intensity as the A?Y component in the angle range from 0j to 20j. As a consequence, the best spectroscopic conditions and sensitivity can be obtained for an incident angle near 20j where pure A?Y and AOX components can be recorded for s- and p-polarization, respectively. For p-polarization, the component AOZ is always combined with the component AOX, but they show opposite signs, if the AOZ is positive, the AOX is negative and reverse. This allows distinguishing them from each other. The best experimental conditions to observe the AOZ component are for an angle of incidence 10 – 15j below and above the Brewster angle. This ensures strong sensitivity to study an adsorbed layer as thin as 0.3 nm with a maximum of confidence in the interpretation of experimental results (Mielczarski and Mielczarski, 1995). The calculated absorbances for various wavenumbers give the simulated spectrum that can be compared to experimental reflection spectrum. The incident infrared beam reflected from mineral surface carries all the information about surface composition and structure. The recorded reflection spectra of the investigated sample at different incident angles and polarizations of the incident beam are sufficient to describe qualitatively and quantitatively the surface phenomena taking place at mineral surfaces. Each spectrum recorded in different optical conditions carries specific spatial information about the composition
and structure of the surface layer. With proper manipulation of the experimental optical conditions (incident angle and polarization), it is sufficient to record three spectra, which, together, give a three-dimensional ‘‘picture’’ of the species present at the mineral surface. The technique is based on the comparison of the experimental spectra with the simulated spectra of the hypothetical surface layer with the assumed composition and structure as is shown below. The good agreement (fitting) between all recorded experimental spectra and the simulated ones allows us to evaluate qualitatively and quantitatively the investigated surface structure. Lack of good agreement requires continuation of spectral simulation with different assumed composition, structure or adsorbed quantity. The orientation of a molecular group (assuming a uniaxial orientation model of the transition dipole moment) can be described for a single wavenumber (see Fig. 3) by the following equation: AN;h1 AX ;h1 tan2 u þ 2AZ;h1 ¼ AN;h2 AX ;h2 tan2 u þ 2AZ;h2 where AX,h1, AZ,h1, AX,h2 and AZ,h2 are calculated absorbances components determined for 1 nm adsorbed layer, and h1 and h2 are the incident angles. AO,h1 and AO,u2 are absorbances determined experimentally. The thickness of the ultrathin surface layer can be determined from the same spectra using simple relations such as the one described below. AN;h1 ¼ 3=2nðAX ;h1 sin2 u þ 2AZ;h1 cos2 uÞ where n is the thickness of adsorbed layer in nanometers. It is also important to emphasize that the simulation of isotropic adsorbed layer is performed before any experiments are carried out, allowing to predict the best experimental conditions that ensure optimal spectral sensitivity and the maximum confidence in the interpretation of experimental results (see Fig. 2 and the discussion above). This also significantly speeds up the experimental procedure.
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
The developed technique has very unique properties compared with other known infrared techniques such as transmission, diffuse reflectance, attenuated total reflection or photoacoustic. This technique, supported by the spectral simulation of surface species, allows obtaining almost all the details about the mineral –aqueous solution interactions, including:
the nature of the adsorbed products (by which functional group adsorption takes place), the adsorbed quantities of different surface products (starting from 20% of monolayer), the surface distribution of the adsorbed species (uniform layer or patches with determined thickness), molecular orientation of the adsorbed species (orientation of particular functional groups), molecular recognition (selective adsorption on specific surface sites), and dynamic phenomena such as kinetics of adsorption/desorption, stability of surface products, surface mobility of the adsorbed species.
All of this information leads to the determination of the mechanisms and dynamics of surface phenomena and the proposition of surface modifications for numerous applications where mineral surface properties play important roles. This multidiagnostic technique shows important advantages in the study of the surface phenomena on solids:
the prediction of the most efficient experimental conditions, the investigation can be carried out in situ, which is particularly important when the produced surface species are not stable after removal of the sample from aqueous solution and drying, any type of mineral sample can be investigated; there is no limitation, from the transparent to nontransparent for infrared radiation, it is very easy to distinguish the absorbance bands of substrate from the bands due to the adsorption layer; they are, at certain optical conditions, positive and negative, respectively, the technique shows the ability to study mechanisms of the formation of monolayers on substrates
133
showing very strong absorbance in the region of the characteristic vibration of the surface species, for example, carboxyl group interactions with a calcite surface. This ability could be very helpful to study other very complex systems, for example, biological ones, and the surface distribution of the adsorbed species in multicomponent and heterogeneous mineral systems was successfully investigated (spatially resolved spectroscopic analysis). It was demonstrated that it is possible to determine exactly the type of mineral on which adsorption takes place even if the same product could be formed on different mineral components of ore sample. This unique advantage can find a wide application for studying the interfacial properties at a molecular level for multicomponent and heterogeneous mineral samples. All of the mentioned features were experimentally evaluated in our original studies. Examples of optical considerations and spectra simulations for different multilayer stratified systems on solid substrate can be found in several papers (Yen and Wong, 1988; Mielczarski and Yoon, 1989; Parikh and Allara, 1992; Mielczarski, 1993; Mielczarski and Mielczarski, 1995, 1999; Mielczarski et al., 1995, 1996; Hoffmann et al., 1995; Brunner et al., 1997). This multidiagnostic technique has been applied extensively to study the interaction of different minerals such as: semisoluble minerals (Mielczarski and Mielczarski, 1995, 1999; Mielczarski et al., 1998b, 1999, 2002) and sulfides (Mielczarski et al., 1995, 1996, 1997, 1998a) with various aqueous solutions. 2.2. Materials Natural mineral samples of chalcocite, pyrite and chalcopyrite (from Ward’s Natural Science) with dimensions of around 10 20 mm with one side polished were used in this study. The samples were high-purity with very small inclusions of other sulfide minerals. X-ray diffraction confirmed the polycrystalline structure of the mineral samples. The potassium ethyl and amyl xanthates used in these studies were synthesized from CS2, KOH and ethyl or amyl alcohol, and then purified by recrys-
134
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
tallization from acetone and ether. Other reagents were all of an analytical grade. Distilled water from the Millipore (Milli-Qplus, 18 MV cm) system was used throughout the experiments. 2.3. Adsorption studies The mineral samples were polished with emery paper and alumina powder. The final polishing was made with the use of 0.05 Am alumina and the polished samples were washed with water. The mineral sample was immersed in 200 ml of solution containing the desired amount of collector and/or modifier. The pH of all the used solutions, including washing water, was between 8.0 and 10.0 F 0.1. Immediately after adsorption, the sample was immersed in water with a pH exactly the same as during adsorption for about 1 s, in order to remove possible drops of collector or modifier solution, and then placed instantly in an FTIR spectrophotometer to record the reflection spectra. Electrochemical measurements were performed by the use of potentiostat/ galvanostat 263A from EG&G Instruments and reported potential values are versus saturated hydrogen electrode (SHE).
3. Results and discussion 3.1. Ethyl xanthate interaction with chalcocite Adsorption of ethyl xanthate on cuprous sulfide has been performed at open circuit potential (OCP) using solutions with different concentrations and pHs. The reflection spectra were recorded at angles of incidence 70j and 85j for p-polarization and 20j for s-polarization (Fig. 3). These are the optical conditions that ensure the determination of all the possible vibrations characteristic for the adsorbed molecules at appropriate spectral sensitivity. As expected, the spectra recorded from the same sample at various optical conditions are very different. There
2.4. Infrared analysis The infrared reflection spectra of slab samples were recorded on a Bruker IFS55 FTIR spectrometer equipped with an MCT or DTGS detector and a reflection attachment (Seagull). A wire-grid polarizer was placed before the sample and provided p- or spolarized light. These accessories were from Harrick Scientific. For each adsorption layer sample usually three reflection spectra were recorded by the use of sand p-polarized light and different angles of incidence. An optimized optical reflection system permits to detect adsorbed amounts as low as about 30% of a statistical monolayer of copper xanthate on a chalcocite surface, which is equivalent to a 0.2-nm-thick uniform layer of copper xanthate. The unit of intensity was defined as log(R/Ro), where Ro and R are the reflectivities of the systems without and with the investigated adsorbed layer, respectively. Both sample and reference spectra are averaged over the same number of scans, from 200 to 2000 scans, depending on energy throughput.
Fig. 3. Reflection spectra of ethyl xanthate adsorbed on chalcocite from solution of 5 10 5 M at pH 9.2. (a) Spectra recorded immediately after adsorption; (b) sample (a) after 30 min of immersion in water.
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
are strong changes in absorbance intensities as well as «negative» or «positive» peaks are observed. These experimental spectra were compared to the simulated spectra of the cuprous sulfide with assumed 1-nmthick adsorbed cuprous ethyl xanthate layer (Fig. 4). The comparison shows that the major features of the experimental spectra such as relative intensity, band shapes and positions could be predicted by theoretical calculation and explained by changes of three electric field components within the adsorbed nanolayer. This similarity indicates that the adsorbed product is cuprous ethyl xanthate and the adsorbed molecules are not well organized at chalcocite surface because the spectral calculation was performed with the assumption of an isotropic surface structure of cuprous xanthate. Nevertheless, important differences are also
Fig. 4. Simulated reflection spectra of an isotropic 1 nm cuprous ethyl xanthate on chalcocite for s-polarization and 20j and for p-polarization and incident angles of 70j and 85j.
135
observed that has to be explained by more than just optical effects. One of the major differences is the presence of the band at about 1225 cm 1 in the spectra of all samples after a short adsorption time, when a close to monolayer coverage was produced. The results obtained (Fig. 3a) clearly show that the band at about 1225 cm 1 is related to the formation of a surface product limited to submonolayer coverage. This band practically disappears from the reflection spectra after longer adsorption when multilayer coverage is produced (spectra not shown). Furthermore, because the band is observed at the incident angles of 70j and 85j at p-polarization and practically negligible at s-polarization (Fig. 3a), it is due to the molecular dipole moment oriented almost parallel to the surface normal. Very interesting changes in the structure and composition of the adsorbed layer were observed after holding the cuprous sulfide sample with the adsorbed xanthate layer in water over a period of about 30 min (Fig. 3b). It can be seen that the band at 1225 cm 1 disappeared. At the same time, the absorbance bands characteristic of the cuprous xanthate complex became sharp and more intensive, and a negative band at 1035 cm 1 in the spectrum recorded at 70j and ppolarization is found. These observations indicate that the already adsorbed xanthate molecules undergo surface diffusion and reorganization on the surface of cuprous sulfide forming a well-ordered structure. The performed quantitative evaluation enables the determination of the average orientation of the molecular groups of xanthate molecules, i.e., COC and SCS, in the adsorption layer. The results of the calculation are presented in Fig. 5. The calculated average thickness of this adsorption layer is 2.0 nm, which is about two oriented monolayers. A quantitative evaluation of the sample presented in Fig. 3a gives a value of about 1.8 nm. The good agreement between surface coverages found for the same sample, before and after immersing in water, supports the above conclusion about surface diffusion of adsorbed cuprous xanthate molecules and their reorganization in a well-organized structure, with simultaneous removal of the surface product with a band at 1225 cm 1. These results indicate that the first step of the formation of the self-assembled cuprous xanthate layer is governed by a very high affinity between
136
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
Fig. 5. Determined molecular arrangement of adsorbed ethyl xanthate molecules on chalcocite.
spectrum presented in Fig. 6a shows no adsorption of amyl xanthate at the potential of 90 mV, which is the open circuit potential of chalcopyrite in the system under investigation. The absorption bands at about 1200 and 1034 cm 1, are visible at a potential of 110 mV (Fig. 6b), indicating the formation of the cuprous amyl xanthate complex. At a potential of 10 mV higher (Fig. 6c), an additional band at 1269 cm 1 is observed, indicating the formation of a small amount of dixanthogen. A future small increase of potential (5 mV) to a value of 125 mV (Fig. 6d) results in a very strong increase of the amount of amyl dixanthogen produced at mineral surface while the amount of cuprous ethyl xanthate complex remains almost un-
xanthate ions and cuprous sulfide and the formation of the surface product with a band at 1225 cm 1, both of which hamper the formation of the ordered surface layer. The surface product with the band at 1225 cm 1 can act as an impurity, which disturbs the formation of the organized surface structure. Removal of this product and reorganization of the structure of the xanthate molecules at the cuprous sulfide surface required certain conditioning time to be successfully completed with creation of a well-ordered structure. This can be achieved by immersing the cuprous sulfide with adsorption layer in water or, even better, in a low-concentration xanthate solution. The band at 1225 cm 1 is assigned to surface decomposition product(s) that are produced at a very early stage of xanthate adsorption when the most energetic surface sites interact with xanthate molecules. Monothiocarbonate is a possible product at this step of surface reaction. 3.2. Amyl xanthate interaction with chalcopyrite Results of spectroelectrochemical studies of the surface composition of chalcopyrite contacted with xanthate solution for 10 min at various potentials are presented in Fig. 6. These spectra were recorded at the optical conditions that ensure the best sensitivity for the system under investigation, i.e., p-polarization and an incident angle of 70j. This is above the Brewster angle for the system under investigation (see explanation above). The highest potential applied, 300 mV, was in the region of oxidation of chalcopyrite. The
Fig. 6. Experimental reflection spectra of chalcopyrite contacted with amyl xanthate solution of 2 10 4 M for 10 min at various potentials. The spectra are recorded for p-polarization and incident angle of 70j.
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
changed. The adsorbed dixanthogen is in its liquid like form (Mielczarski et al., 1996) that excludes any rigid orientation at mineral surface, nevertheless, some differences between simulated (not shown) and experimental spectra could indicate some organization especially at lower coverage. It was interesting to find more details about the kinetic limitations of xanthate adsorption. Extension of contact time to 1 h significantly changes the surface composition (Fig. 7). Whereas at a potential of 80 mV after 10 min of adsorption, no xanthate surface product was detected (Fig. 7d), a prolongation of the adsorption to 1 h results in the formation of a significant amount (6-nm-thick layer) of amyl dixanthogen (bands at 1270, 1047 and 1031 cm 1) and a small amount (about 0.6 nm) of cuprous amyl xanthate (band at 1204 cm 1) (Fig. 7e) with isotropic structure. At a potential of 5 mV lower after 1 h of adsorption (Fig. 7b), the amount of dixanthogen is several times lower while the amount of cuprous xanthate shows only a small decrease. A prolongation
137
of the contact time to 2 h (Fig. 7c) does not change the surface composition noticeably, indicating an equilibrium or semi-equilibrium state. A rough estimation shows that the intensity of the band at 1204 cm 1 is very similar to that expected for a 40% of monolayer coverage (0.4 nm) of cuprous xanthate complex while dixanthogen forms a layer about 1.3 nm thick. The amount of these two products together suggests the formation of a nearly monomolecular two-component layer at the potential of 75 mV. At a potential of 70 mV, both adsorption products, amyl dixanthogen and cuprous amyl xanthate, are at the detection limit, i.e., below 0.2 nm for cuprous amyl xanthate and 0.4 nm for amyl dixanthogen estimated from spectra of isotropic structures. The above results clearly show that there are significant kinetic limitations of surface product formation with an inert time in the potential region close to the thermodynamic potential. Amyl xanthate produces two adsorption products simultaneously on the chalcopyrite surface; the adsorption begins at about 75 mV for an extended adsorption time of 1 h. Therefore, the potential of 75 mV could be assumed as the thermodynamic potential of the initial amyl xanthate surface layer formation on chalcopyrite.
4. Interaction of amyl xanthate with pyrite and chalcopyrite in the presence of metabisulfite (MBS) modifier
Fig. 7. Experimental reflection spectra of chalcopyrite contacted with amyl xanthate solution of 2 10 4 M for extended adsorption times at potentials close to thermodynamic potential of xanthate adsorption.
MBS or SO2 is used in flotation practice as a modifier, but the influence of the reagent on collector adsorption and surface hydrophobization is not well known. The modifier could interact with mineral surface in different ways: It could change potential of mineral in solution, be adsorbed on mineral surface, form complexes with collector in solution or on mineral surface, remove the adsorbed collector from mineral or modify surface charge, etc. Pyrite and chalcopyrite in MBS solution show lower potentials than those observed in electrolyte solution of 0.05 M sodium sulfate alone. The difference is around 140 mV for pyrite and 40 mV for chalcopyrite at 5 10 3 M concentration of MBS in solution. The spectroscopic results of xanthate adsorption on pyrite at different concentrations of MBS in xanthate solution are presented in Fig. 8. The influ-
138
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
Fig. 8. Reflection spectra of pyrite contacted 30 min with amyl xanthate solution of 5 10 4 M without and with MBS at various concentrations.
Fig. 9. Reflection spectra of chalcopyrite contacted 30 min with amyl xanthate solution of 5 10 4 M without and with MBS at various concentrations.
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
ence of the modifier on xanthate adsorption at a concentration of 10 3 M and lower is very small. The adsorption product formed in the presence of MBS is the same as without it. There is only dixanthogen. The amount of adsorbed dixanthogen is lower by about 20% compared to the result without the modifier. It is necessary to increase the concentration of MBS to 5 10 3 M to suppress dramatically xanthate adsorption (Fig. 8d). This high addition of modifier also causes lowering of pyrite potential in the solution by about 50 mV. At lower concentrations of MBS, 10 3 M (Fig. 8c), the potential changes by about 5 mV, which is a negligible difference. It should be noted that the lowest potential recorded at which xanthate adsorption is almost completely suppressed, i.e., 141 mV (Fig. 8d), is still much higher than 80 mV, which is the thermodynamic potential of dixanthogen formation at the xanthate concentration used in the studies. Hence, it is not only reduction of piryte potential that plays a decisive role in the suppression of xanthate adsorption. Similar experiments were performed with chalcopyrite and the results are presented in Fig. 9. In this case, OCP of chalcopyrite in the xanthate solution does not change significantly with MBS addition. It is around 60 mV even at the highest concentration of MBS. The observed adsorption product is dixanthogen, and its amount remains the same from lack of modifier in solution up to 10 4 M concentration. At a concentration of 1 10 3 M, a dramatic change is observed, the adsorption of xanthate is almost totally suppressed (Fig. 9d). Almost the same situation was found at a concentration of MBS that is five times higher (Fig. 9e). These results indicate that the use of MBS as a modifier to depress pyrite versus chalcopyrite cannot be effective, the suppression of the formation of hydrophobic product of dixanthogen is observed at a dosage of MBS a few times lower for chalcopyrite than for pyrite. In the real separation flotation process, pyrite and chalcopyrite are in the same solution together and they undergo galvanic interactions that significantly modify the thermodynamics and kinetics of xanthate adsorption as was already documented experimentally by numerous authors. The results of spectroscopic studies of multimineral systems with galvanic effect are subject of separate report.
139
5. Conclusions The developed infrared external reflection technique has a very unique ability to study interface phenomena at a molecular level for heterogeneous and multicomponent substrates such as natural minerals. The variety, precision and reliability of information about interface phenomena delivered by this technique are superior to other single techniques. The experiments are fast and nondestructive. High sensitivity (starting from 20% of monolayer) and in situ collected information in a multiphase system even in the region of strong absorption of substrate make this technique a very valuable experimental tool. As was demonstrated here, the complexity of recorded reflection spectra, their sensitivity to any variations of the optical properties of all investigated phases in the system and their spatial distribution, are in fact the major strength of the technique. The ability to perform good research in very complex systems is probably the most unique capability of the technique. It could be concluded that the direct monitoring of the surface of minerals with changing solution chemistry is extremely important to understand and determinate the critical parameters that control surface phenomena. These findings are important from the flotation point of view and found already industrial applications.
References Brunner, H., Mayer, U., Hoffmann, H., 1997. External reflection infrared spectroscopy of anisotropic adsorbate layers on dielectric substrates. Appl. Spectrosc. 51, 209 – 217. Hoffmann, H., Mayer, U., Krischanitz, A., 1995. Structure of alkylsiloxane monolayers on silicon surfaces investigated by external reflection infrared spectroscopy. Langmuir 11, 1304 – 1312. Mielczarski, J.A., 1993. External reflection infrared spectroscopy at metallic, semiconductor and nonmetallic substrates: I. Monolayers films. J. Phys. Chem. 97, 2649 – 2663. Mielczarski, J.A., Mielczarski, E., 1995. Determination of molecular orientation and thickness of self-assembled monolayers of oleate on apatite by FTIR reflection spectroscopy. J. Phys. Chem. 99, 3206 – 3217. Mielczarski, J.A., Mielczarski, E., 1999. Infrared external reflection spectroscopy of adsorbed monolayers in a region of strong absorption of substrate. J. Phys. Chem. 103, 5852 – 5859. Mielczarski, J.A., Yoon, R.H., 1989. FTIR external reflection study of molecular orientation in spontaneously adsorbed layers on low-absorption substrates. J. Phys. Chem. 93, 2034 – 2038.
140
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
Mielczarski, J.A., Mielczarski, E., Zachwieja, J., Cases, J.M., 1995. In situ and ex situ infrared studies of nature and structure of thiol monolayers adsorbed on cuprous sulfide at controlled potential. Langmuir 11, 2787 – 2799. Mielczarski, J.A., Mielczarski, E., Cases, J.M., 1996. Interaction of amyl xanthate with chalcopyrite, tetrahedrite and tennantite at controlled potentials. Simulation and spectroelectrochemical results for two component adsorption layers. Langmuir 12, 6521 – 6529. Mielczarski, J.A., Mielczarski, E., Cases, J.M., 1997. Infrared evaluation of composition and structure of ethyl xanthate monolayers produced on chalcopyrite, tetrahedrite and tennantite at controlled potentials. J. Colloid Interface Sci. 188, 150 – 161. Mielczarski, J.A., Mielczarski, E., Cases, J.M., 1998a. Influence of chain length on adsorption of xanthates on chalcopyrite. Int. J. Miner. Proc. 52, 215 – 231.
Mielczarski, E., Mielczarski, J.A., Cases, J.M., 1998b. Molecular recognition effect in monolayer formation of oleate on fluorite. Langmuir 14, 1739 – 1747. Mielczarski, J.A., Mielczarski, E., Cases, J.M., 1999. Dynamic of fluorite – oleate interactions. Langmuir 15, 500 – 508. Mielczarski, E., Mielczarski, J.A., Cases, J.M., Rai, B., Pradip, 2002. Influence of solution conditions and mineral surface structure on the formation of oleate adsorption layers on fluorite. Colloids Surf. 205, 73 – 84. Parikh, A.N., Allara, D.L., 1992. Quantitative determination of molecular structure in multilayered thin films of biaxial and lower symmetry from photon spectroscopies. J. Chem. Phys. 96, 927 – 945. Yen, Y.S., Wong, J.S., 1988. Intriguing absorption band behavior of IR reflectance spectra of silicon dioxide on silicon. Appl. Spectrosc. 42, 598 – 604.
Surface modification at the molecular level in mineral beneficiation Ela Mielczarski *, Jerzy A. Mielczarski ‘‘Laboratoire Environnement et Mine´ralurgie’’, UMR 7569 du CNRS, INPL/ENSG, 15 Avenue du Charmois, BP 40, Vandoeuvre-le`s-Nancy, Cedex 5450, France Received 20 September 2002; received in revised form 25 October 2002; accepted 30 June 2003
Abstract Direct information at a molecular level about surface products, their structure and surface distribution is vitally important to perform an efficient separation processes for mineral beneficiation. This information is equally important for the modification of surface properties of industrial minerals for various specific applications such as: fillers, composites, weathering resistance, etc. The understanding of the mechanisms and kinetics of interaction of the first adsorbed organic molecules with mineral surfaces is the fundamental requirement to make possible the prediction, control and modification of the macroscopic surface properties that govern the efficiency of separation technologies or new material formulations. The developed infrared external reflection technique has a very unique ability to study interface phenomena at a molecular level on heterogeneous substrates. The variety, precision and reliability of information about interface phenomena delivered by this technique are superior to other single techniques. The experiments are fast and nondestructive. High sensitivity (starting from 20% of monolayer), in situ collected information in a multiphase system even in the region of a strong absorption of substrate, makes this technique a very valuable experimental tool. The complexity of the recorded reflection spectra, their sensitivity to any variations of the optical properties of all bulk and surface components and their spatial distribution in the system under investigation are in fact the major strength of the technique. In this paper, a few examples of application of this multidiagnostic technique for monitoring surface modifications of sulfide minerals in aqueous solutions will be discussed in detail. D 2003 Elsevier B.V. All rights reserved. Keywords: surface modification; minerals; molecular orientation; infrared spectroscopy; adsorption
1. Introduction The flotation behavior of minerals depends on the nature, structure and surface distribution of their surface hydrophobic and hydrophilic species. The possibility of monitoring surface phenomena at mo-
* Corresponding author. E-mail address: [email protected] (E. Mielczarski). 0301-7516/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0301-7516(03)00093-0
lecular and atomic levels at the interfaces of natural minerals contacted with aqueous solution is vitally important for the understanding of the mechanisms and kinetics of surface phenomena and then to design efficient flotation technologies. Improvement of selective separation of ore components is achieved by manipulation of solution conditions such as pH, Eh (aeration, reducing or oxidizing reagents), addition of different types of collectors at different concentrations and modifiers such as activators or depressants. All these changes in solution compositions are performed
130
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
in one aim—to make the valuable mineral components very hydrophobic while keeping gangue components hydrophilic. This explains why the information collected directly from mineral surfaces is so important for understanding kinetic and thermodynamic relationships between the solution conditions and the nature and structure of the outermost layers that play a crucial role in the surface modification for separation or other processes. Ores are usually complex, multicomponent, and heterogeneous systems and cannot be described by a single value. They interact with aqueous solution through complex dissolution/precipitation, adsorption/desorption or oxidation/reduction processes. Monitoring the formation of surface species at submonolayer and monolayer levels, and their structures, requires the application of the appropriate experimental techniques. Infrared spectroscopy is particularly well suited to determine surface composition at a molecular level. Infrared is functional group selective, so it is particularly appropriate to detect small changes of the molecular microenvironment properties as those emerging at the interfaces. Experiments can be performed in situ at both gas – mineral and aqueous solution – mineral interfaces. There is only very gentle interaction of the infrared beam with the examined sample. The recent instrumental development of infrared spectroscopy is contributing significantly to the increased emphasis being placed on molecular level surface characterization. To perform the proper interpretation of reflection spectra for a more detailed picture of the interfacial structure, it is vitally important to combine the spectroscopic measurements with a spectral simulation. The importance of such combination is reinforced by the anticipated sensitivity of surface infrared absorbance not only to surface concentration but also to adsorbate structure, molecular orientation, surface distribution, lateral interaction, surface diffusion, molecular recognition and so-called optical effects. Several unique advantages of the developed infrared external reflection technique are overviewed in this work. The different examples of sulfide mineral – aqueous solution systems for which detailed qualitative and quantitative evaluations of the produced surface layers were performed are discussed in this paper.
2. Experimental 2.1. Developing of the infrared external reflection technique One of the major problems to solve was developing proper experimental procedures for collecting representative and appropriate information directly from mineral surfaces contacted with aqueous solution of organic and inorganic species. The complexity of the system under investigation, heterogeneity of mineral samples and the produced surface species and their stability require application of the appropriate experimental techniques. During these studies, we developed an infrared external reflection technique that let us overcome experimental problems and collect reliable data to monitor and understand surface phenomena at a molecular level (Mielczarski, 1993). A schematic diagram of the interaction of electromagnetic waves with a simple three-phase system is presented in Fig. 1. For polarization perpendicular to the plane of the incident beam (s-polarization), there is only one electric vector, E?Y, parallel to substrate plane. Hence, only molecular groups of the adsorbed species having a dipole transition moment parallel to the interface in y direction can interact with the incident radiation and produce an absorbance band A?Y. For example, in the case of the adsorption of a xanthate molecule involving both sulfur atoms with the same distance from interface (Fig. 1) and assuming a uniaxial system (no difference in orientation in x –y plane), it is expected that the absorbance assigned to the SCS group, A?Y, will show the highest value. This results from the most favorable configuration for the interaction of the asymmetric stretching vibration of the SCS group and electric field vector E?Y. For a molecule inclined to interface the interaction is lower and the recorded absorbance will decrease depending on the orientation of this molecular group versus interface. No interaction and obviously no absorbance band is observed when the SCS group is turned 90j from the position presented in Fig. 1 when the dipole transition moment of the asymmetric vibration of the group becomes vertical to interface. For parallel polarization (p-polarization), there are two electric field vector components at the interface, one parallel EOX and one perpendicular EOZ to the
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
131
CH3 CH2
EII
νas
O C S
z
νas
E ⊥Y S
θ
EII Z
n1 , k1
ambient
n2 , k 2
adsorption layer
n3 , k3
mineral
EII X d
x y
Fig. 1. Schematic diagram of the interaction of electric field vectors in three directions with a simple three-phase system. Ethyl xanthate molecule with marked dipole transition moments of the asymmetric stretching vibrations of the SCS group (parallel to interface) and of the asymmetric stretching vibrations of the COC group (almost vertical to interface) is also presented.
substrate plane (Fig. 1). Therefore, for p-polarization, the molecular groups showing dipole transition moments parallel and vertical to interface in x and z directions can produce absorbance bands. Using the example of xanthate molecules at the surface presented in Fig. 1, two absorbance bands, the AOX for the SCS group vibration and the AOZ related to the COC group vibration, will be presented in the recorded spectrum. As it is indicated in Fig. 1, the asymmetric stretching vibration of the COC group occurs almost vertically to the surface. It is also possible to distinguish these two components because they show reverse absorbance, whereas one produces positive absorbance, the second one is negative or vice versa, depending on the angle of incidence (see the discussion below). The intensities of the absorbance band recorded for the two polarizations A? and AO depend on the nature of the vibration of the particular molecular group, the amount of surface species, the angle of the incident beam and the orientation of molecules in the surface layer. The absorbance components in three directions for particular wavenumber can be calculated as follows (Mielczarski and Yoon, 1989): 16p cosh n2 k2 d2 A?Y ¼ ln10 n23 1 k 2 16p cosh n n2 k2 d2 ANX ¼ 34 ln10 n32 n43 cos2 h k n3 16p cosh sin2 h n2 k2 d2 ANZ ¼ 2 4 2 2 ln10 n3 n3 cos h ðn2 þ k 22 Þ2 k
where n3=(nˆ3 n1sin2h)1/2, n3 = nˆ3 + ik3, h is the incident angle and k is the wavelength of the incident beam. The presented approximate equations are used to calculate absorbance components in three directions and then to compare them with the A? and AO experimental values that are recorded for different angles. The results obtained for a 2-nm-thick isotropic layer of polyethylene aliphatic chain for the stretching vibration of the CH2 group at 2920 cm 1 versus an angle of incidence are presented in
Fig. 2. Calculated absorbance components in three directions A?Y, AOX, AOZ for a three-phase system: air, n1 = 1, k1 = 0; aliphatic chain, n2 = 1.576, k2 = 0.246; substrate, n3 = 1.657, k3 = 0.0002.
132
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
Fig. 2. There are tremendous changes in the absorbance components just with variation of the angle of the incident beam as is shown in Fig. 2. For spolarization, only the negative absorbance band A?Y is observed and its intensity is highest at the lowest angle of incident. Therefore, for s-polarization, the best experimental conditions to record reflection spectrum are with an angle of incidence close to normal to the interface. For p-polarization, there are two absorbance components AOZ and AOX. Below the Brewster angle, which is for this system 58.7j, the AOZ component is negligible for angles of incidence between 0j and 20j, then it shows a positive value with strong increases of intensity in the vicinity of the Brewster angle. Above this angle, the AOZ component shows a negative value, at first very high, then decreasing quickly to zero at 90j. The AOX component has reverse behavior; it is negative below the Brewster angle and positive above this angle. Moreover, the AOX component is negative with the same intensity as the A?Y component in the angle range from 0j to 20j. As a consequence, the best spectroscopic conditions and sensitivity can be obtained for an incident angle near 20j where pure A?Y and AOX components can be recorded for s- and p-polarization, respectively. For p-polarization, the component AOZ is always combined with the component AOX, but they show opposite signs, if the AOZ is positive, the AOX is negative and reverse. This allows distinguishing them from each other. The best experimental conditions to observe the AOZ component are for an angle of incidence 10 – 15j below and above the Brewster angle. This ensures strong sensitivity to study an adsorbed layer as thin as 0.3 nm with a maximum of confidence in the interpretation of experimental results (Mielczarski and Mielczarski, 1995). The calculated absorbances for various wavenumbers give the simulated spectrum that can be compared to experimental reflection spectrum. The incident infrared beam reflected from mineral surface carries all the information about surface composition and structure. The recorded reflection spectra of the investigated sample at different incident angles and polarizations of the incident beam are sufficient to describe qualitatively and quantitatively the surface phenomena taking place at mineral surfaces. Each spectrum recorded in different optical conditions carries specific spatial information about the composition
and structure of the surface layer. With proper manipulation of the experimental optical conditions (incident angle and polarization), it is sufficient to record three spectra, which, together, give a three-dimensional ‘‘picture’’ of the species present at the mineral surface. The technique is based on the comparison of the experimental spectra with the simulated spectra of the hypothetical surface layer with the assumed composition and structure as is shown below. The good agreement (fitting) between all recorded experimental spectra and the simulated ones allows us to evaluate qualitatively and quantitatively the investigated surface structure. Lack of good agreement requires continuation of spectral simulation with different assumed composition, structure or adsorbed quantity. The orientation of a molecular group (assuming a uniaxial orientation model of the transition dipole moment) can be described for a single wavenumber (see Fig. 3) by the following equation: AN;h1 AX ;h1 tan2 u þ 2AZ;h1 ¼ AN;h2 AX ;h2 tan2 u þ 2AZ;h2 where AX,h1, AZ,h1, AX,h2 and AZ,h2 are calculated absorbances components determined for 1 nm adsorbed layer, and h1 and h2 are the incident angles. AO,h1 and AO,u2 are absorbances determined experimentally. The thickness of the ultrathin surface layer can be determined from the same spectra using simple relations such as the one described below. AN;h1 ¼ 3=2nðAX ;h1 sin2 u þ 2AZ;h1 cos2 uÞ where n is the thickness of adsorbed layer in nanometers. It is also important to emphasize that the simulation of isotropic adsorbed layer is performed before any experiments are carried out, allowing to predict the best experimental conditions that ensure optimal spectral sensitivity and the maximum confidence in the interpretation of experimental results (see Fig. 2 and the discussion above). This also significantly speeds up the experimental procedure.
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
The developed technique has very unique properties compared with other known infrared techniques such as transmission, diffuse reflectance, attenuated total reflection or photoacoustic. This technique, supported by the spectral simulation of surface species, allows obtaining almost all the details about the mineral –aqueous solution interactions, including:
the nature of the adsorbed products (by which functional group adsorption takes place), the adsorbed quantities of different surface products (starting from 20% of monolayer), the surface distribution of the adsorbed species (uniform layer or patches with determined thickness), molecular orientation of the adsorbed species (orientation of particular functional groups), molecular recognition (selective adsorption on specific surface sites), and dynamic phenomena such as kinetics of adsorption/desorption, stability of surface products, surface mobility of the adsorbed species.
All of this information leads to the determination of the mechanisms and dynamics of surface phenomena and the proposition of surface modifications for numerous applications where mineral surface properties play important roles. This multidiagnostic technique shows important advantages in the study of the surface phenomena on solids:
the prediction of the most efficient experimental conditions, the investigation can be carried out in situ, which is particularly important when the produced surface species are not stable after removal of the sample from aqueous solution and drying, any type of mineral sample can be investigated; there is no limitation, from the transparent to nontransparent for infrared radiation, it is very easy to distinguish the absorbance bands of substrate from the bands due to the adsorption layer; they are, at certain optical conditions, positive and negative, respectively, the technique shows the ability to study mechanisms of the formation of monolayers on substrates
133
showing very strong absorbance in the region of the characteristic vibration of the surface species, for example, carboxyl group interactions with a calcite surface. This ability could be very helpful to study other very complex systems, for example, biological ones, and the surface distribution of the adsorbed species in multicomponent and heterogeneous mineral systems was successfully investigated (spatially resolved spectroscopic analysis). It was demonstrated that it is possible to determine exactly the type of mineral on which adsorption takes place even if the same product could be formed on different mineral components of ore sample. This unique advantage can find a wide application for studying the interfacial properties at a molecular level for multicomponent and heterogeneous mineral samples. All of the mentioned features were experimentally evaluated in our original studies. Examples of optical considerations and spectra simulations for different multilayer stratified systems on solid substrate can be found in several papers (Yen and Wong, 1988; Mielczarski and Yoon, 1989; Parikh and Allara, 1992; Mielczarski, 1993; Mielczarski and Mielczarski, 1995, 1999; Mielczarski et al., 1995, 1996; Hoffmann et al., 1995; Brunner et al., 1997). This multidiagnostic technique has been applied extensively to study the interaction of different minerals such as: semisoluble minerals (Mielczarski and Mielczarski, 1995, 1999; Mielczarski et al., 1998b, 1999, 2002) and sulfides (Mielczarski et al., 1995, 1996, 1997, 1998a) with various aqueous solutions. 2.2. Materials Natural mineral samples of chalcocite, pyrite and chalcopyrite (from Ward’s Natural Science) with dimensions of around 10 20 mm with one side polished were used in this study. The samples were high-purity with very small inclusions of other sulfide minerals. X-ray diffraction confirmed the polycrystalline structure of the mineral samples. The potassium ethyl and amyl xanthates used in these studies were synthesized from CS2, KOH and ethyl or amyl alcohol, and then purified by recrys-
134
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
tallization from acetone and ether. Other reagents were all of an analytical grade. Distilled water from the Millipore (Milli-Qplus, 18 MV cm) system was used throughout the experiments. 2.3. Adsorption studies The mineral samples were polished with emery paper and alumina powder. The final polishing was made with the use of 0.05 Am alumina and the polished samples were washed with water. The mineral sample was immersed in 200 ml of solution containing the desired amount of collector and/or modifier. The pH of all the used solutions, including washing water, was between 8.0 and 10.0 F 0.1. Immediately after adsorption, the sample was immersed in water with a pH exactly the same as during adsorption for about 1 s, in order to remove possible drops of collector or modifier solution, and then placed instantly in an FTIR spectrophotometer to record the reflection spectra. Electrochemical measurements were performed by the use of potentiostat/ galvanostat 263A from EG&G Instruments and reported potential values are versus saturated hydrogen electrode (SHE).
3. Results and discussion 3.1. Ethyl xanthate interaction with chalcocite Adsorption of ethyl xanthate on cuprous sulfide has been performed at open circuit potential (OCP) using solutions with different concentrations and pHs. The reflection spectra were recorded at angles of incidence 70j and 85j for p-polarization and 20j for s-polarization (Fig. 3). These are the optical conditions that ensure the determination of all the possible vibrations characteristic for the adsorbed molecules at appropriate spectral sensitivity. As expected, the spectra recorded from the same sample at various optical conditions are very different. There
2.4. Infrared analysis The infrared reflection spectra of slab samples were recorded on a Bruker IFS55 FTIR spectrometer equipped with an MCT or DTGS detector and a reflection attachment (Seagull). A wire-grid polarizer was placed before the sample and provided p- or spolarized light. These accessories were from Harrick Scientific. For each adsorption layer sample usually three reflection spectra were recorded by the use of sand p-polarized light and different angles of incidence. An optimized optical reflection system permits to detect adsorbed amounts as low as about 30% of a statistical monolayer of copper xanthate on a chalcocite surface, which is equivalent to a 0.2-nm-thick uniform layer of copper xanthate. The unit of intensity was defined as log(R/Ro), where Ro and R are the reflectivities of the systems without and with the investigated adsorbed layer, respectively. Both sample and reference spectra are averaged over the same number of scans, from 200 to 2000 scans, depending on energy throughput.
Fig. 3. Reflection spectra of ethyl xanthate adsorbed on chalcocite from solution of 5 10 5 M at pH 9.2. (a) Spectra recorded immediately after adsorption; (b) sample (a) after 30 min of immersion in water.
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
are strong changes in absorbance intensities as well as «negative» or «positive» peaks are observed. These experimental spectra were compared to the simulated spectra of the cuprous sulfide with assumed 1-nmthick adsorbed cuprous ethyl xanthate layer (Fig. 4). The comparison shows that the major features of the experimental spectra such as relative intensity, band shapes and positions could be predicted by theoretical calculation and explained by changes of three electric field components within the adsorbed nanolayer. This similarity indicates that the adsorbed product is cuprous ethyl xanthate and the adsorbed molecules are not well organized at chalcocite surface because the spectral calculation was performed with the assumption of an isotropic surface structure of cuprous xanthate. Nevertheless, important differences are also
Fig. 4. Simulated reflection spectra of an isotropic 1 nm cuprous ethyl xanthate on chalcocite for s-polarization and 20j and for p-polarization and incident angles of 70j and 85j.
135
observed that has to be explained by more than just optical effects. One of the major differences is the presence of the band at about 1225 cm 1 in the spectra of all samples after a short adsorption time, when a close to monolayer coverage was produced. The results obtained (Fig. 3a) clearly show that the band at about 1225 cm 1 is related to the formation of a surface product limited to submonolayer coverage. This band practically disappears from the reflection spectra after longer adsorption when multilayer coverage is produced (spectra not shown). Furthermore, because the band is observed at the incident angles of 70j and 85j at p-polarization and practically negligible at s-polarization (Fig. 3a), it is due to the molecular dipole moment oriented almost parallel to the surface normal. Very interesting changes in the structure and composition of the adsorbed layer were observed after holding the cuprous sulfide sample with the adsorbed xanthate layer in water over a period of about 30 min (Fig. 3b). It can be seen that the band at 1225 cm 1 disappeared. At the same time, the absorbance bands characteristic of the cuprous xanthate complex became sharp and more intensive, and a negative band at 1035 cm 1 in the spectrum recorded at 70j and ppolarization is found. These observations indicate that the already adsorbed xanthate molecules undergo surface diffusion and reorganization on the surface of cuprous sulfide forming a well-ordered structure. The performed quantitative evaluation enables the determination of the average orientation of the molecular groups of xanthate molecules, i.e., COC and SCS, in the adsorption layer. The results of the calculation are presented in Fig. 5. The calculated average thickness of this adsorption layer is 2.0 nm, which is about two oriented monolayers. A quantitative evaluation of the sample presented in Fig. 3a gives a value of about 1.8 nm. The good agreement between surface coverages found for the same sample, before and after immersing in water, supports the above conclusion about surface diffusion of adsorbed cuprous xanthate molecules and their reorganization in a well-organized structure, with simultaneous removal of the surface product with a band at 1225 cm 1. These results indicate that the first step of the formation of the self-assembled cuprous xanthate layer is governed by a very high affinity between
136
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
Fig. 5. Determined molecular arrangement of adsorbed ethyl xanthate molecules on chalcocite.
spectrum presented in Fig. 6a shows no adsorption of amyl xanthate at the potential of 90 mV, which is the open circuit potential of chalcopyrite in the system under investigation. The absorption bands at about 1200 and 1034 cm 1, are visible at a potential of 110 mV (Fig. 6b), indicating the formation of the cuprous amyl xanthate complex. At a potential of 10 mV higher (Fig. 6c), an additional band at 1269 cm 1 is observed, indicating the formation of a small amount of dixanthogen. A future small increase of potential (5 mV) to a value of 125 mV (Fig. 6d) results in a very strong increase of the amount of amyl dixanthogen produced at mineral surface while the amount of cuprous ethyl xanthate complex remains almost un-
xanthate ions and cuprous sulfide and the formation of the surface product with a band at 1225 cm 1, both of which hamper the formation of the ordered surface layer. The surface product with the band at 1225 cm 1 can act as an impurity, which disturbs the formation of the organized surface structure. Removal of this product and reorganization of the structure of the xanthate molecules at the cuprous sulfide surface required certain conditioning time to be successfully completed with creation of a well-ordered structure. This can be achieved by immersing the cuprous sulfide with adsorption layer in water or, even better, in a low-concentration xanthate solution. The band at 1225 cm 1 is assigned to surface decomposition product(s) that are produced at a very early stage of xanthate adsorption when the most energetic surface sites interact with xanthate molecules. Monothiocarbonate is a possible product at this step of surface reaction. 3.2. Amyl xanthate interaction with chalcopyrite Results of spectroelectrochemical studies of the surface composition of chalcopyrite contacted with xanthate solution for 10 min at various potentials are presented in Fig. 6. These spectra were recorded at the optical conditions that ensure the best sensitivity for the system under investigation, i.e., p-polarization and an incident angle of 70j. This is above the Brewster angle for the system under investigation (see explanation above). The highest potential applied, 300 mV, was in the region of oxidation of chalcopyrite. The
Fig. 6. Experimental reflection spectra of chalcopyrite contacted with amyl xanthate solution of 2 10 4 M for 10 min at various potentials. The spectra are recorded for p-polarization and incident angle of 70j.
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
changed. The adsorbed dixanthogen is in its liquid like form (Mielczarski et al., 1996) that excludes any rigid orientation at mineral surface, nevertheless, some differences between simulated (not shown) and experimental spectra could indicate some organization especially at lower coverage. It was interesting to find more details about the kinetic limitations of xanthate adsorption. Extension of contact time to 1 h significantly changes the surface composition (Fig. 7). Whereas at a potential of 80 mV after 10 min of adsorption, no xanthate surface product was detected (Fig. 7d), a prolongation of the adsorption to 1 h results in the formation of a significant amount (6-nm-thick layer) of amyl dixanthogen (bands at 1270, 1047 and 1031 cm 1) and a small amount (about 0.6 nm) of cuprous amyl xanthate (band at 1204 cm 1) (Fig. 7e) with isotropic structure. At a potential of 5 mV lower after 1 h of adsorption (Fig. 7b), the amount of dixanthogen is several times lower while the amount of cuprous xanthate shows only a small decrease. A prolongation
137
of the contact time to 2 h (Fig. 7c) does not change the surface composition noticeably, indicating an equilibrium or semi-equilibrium state. A rough estimation shows that the intensity of the band at 1204 cm 1 is very similar to that expected for a 40% of monolayer coverage (0.4 nm) of cuprous xanthate complex while dixanthogen forms a layer about 1.3 nm thick. The amount of these two products together suggests the formation of a nearly monomolecular two-component layer at the potential of 75 mV. At a potential of 70 mV, both adsorption products, amyl dixanthogen and cuprous amyl xanthate, are at the detection limit, i.e., below 0.2 nm for cuprous amyl xanthate and 0.4 nm for amyl dixanthogen estimated from spectra of isotropic structures. The above results clearly show that there are significant kinetic limitations of surface product formation with an inert time in the potential region close to the thermodynamic potential. Amyl xanthate produces two adsorption products simultaneously on the chalcopyrite surface; the adsorption begins at about 75 mV for an extended adsorption time of 1 h. Therefore, the potential of 75 mV could be assumed as the thermodynamic potential of the initial amyl xanthate surface layer formation on chalcopyrite.
4. Interaction of amyl xanthate with pyrite and chalcopyrite in the presence of metabisulfite (MBS) modifier
Fig. 7. Experimental reflection spectra of chalcopyrite contacted with amyl xanthate solution of 2 10 4 M for extended adsorption times at potentials close to thermodynamic potential of xanthate adsorption.
MBS or SO2 is used in flotation practice as a modifier, but the influence of the reagent on collector adsorption and surface hydrophobization is not well known. The modifier could interact with mineral surface in different ways: It could change potential of mineral in solution, be adsorbed on mineral surface, form complexes with collector in solution or on mineral surface, remove the adsorbed collector from mineral or modify surface charge, etc. Pyrite and chalcopyrite in MBS solution show lower potentials than those observed in electrolyte solution of 0.05 M sodium sulfate alone. The difference is around 140 mV for pyrite and 40 mV for chalcopyrite at 5 10 3 M concentration of MBS in solution. The spectroscopic results of xanthate adsorption on pyrite at different concentrations of MBS in xanthate solution are presented in Fig. 8. The influ-
138
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
Fig. 8. Reflection spectra of pyrite contacted 30 min with amyl xanthate solution of 5 10 4 M without and with MBS at various concentrations.
Fig. 9. Reflection spectra of chalcopyrite contacted 30 min with amyl xanthate solution of 5 10 4 M without and with MBS at various concentrations.
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
ence of the modifier on xanthate adsorption at a concentration of 10 3 M and lower is very small. The adsorption product formed in the presence of MBS is the same as without it. There is only dixanthogen. The amount of adsorbed dixanthogen is lower by about 20% compared to the result without the modifier. It is necessary to increase the concentration of MBS to 5 10 3 M to suppress dramatically xanthate adsorption (Fig. 8d). This high addition of modifier also causes lowering of pyrite potential in the solution by about 50 mV. At lower concentrations of MBS, 10 3 M (Fig. 8c), the potential changes by about 5 mV, which is a negligible difference. It should be noted that the lowest potential recorded at which xanthate adsorption is almost completely suppressed, i.e., 141 mV (Fig. 8d), is still much higher than 80 mV, which is the thermodynamic potential of dixanthogen formation at the xanthate concentration used in the studies. Hence, it is not only reduction of piryte potential that plays a decisive role in the suppression of xanthate adsorption. Similar experiments were performed with chalcopyrite and the results are presented in Fig. 9. In this case, OCP of chalcopyrite in the xanthate solution does not change significantly with MBS addition. It is around 60 mV even at the highest concentration of MBS. The observed adsorption product is dixanthogen, and its amount remains the same from lack of modifier in solution up to 10 4 M concentration. At a concentration of 1 10 3 M, a dramatic change is observed, the adsorption of xanthate is almost totally suppressed (Fig. 9d). Almost the same situation was found at a concentration of MBS that is five times higher (Fig. 9e). These results indicate that the use of MBS as a modifier to depress pyrite versus chalcopyrite cannot be effective, the suppression of the formation of hydrophobic product of dixanthogen is observed at a dosage of MBS a few times lower for chalcopyrite than for pyrite. In the real separation flotation process, pyrite and chalcopyrite are in the same solution together and they undergo galvanic interactions that significantly modify the thermodynamics and kinetics of xanthate adsorption as was already documented experimentally by numerous authors. The results of spectroscopic studies of multimineral systems with galvanic effect are subject of separate report.
139
5. Conclusions The developed infrared external reflection technique has a very unique ability to study interface phenomena at a molecular level for heterogeneous and multicomponent substrates such as natural minerals. The variety, precision and reliability of information about interface phenomena delivered by this technique are superior to other single techniques. The experiments are fast and nondestructive. High sensitivity (starting from 20% of monolayer) and in situ collected information in a multiphase system even in the region of strong absorption of substrate make this technique a very valuable experimental tool. As was demonstrated here, the complexity of recorded reflection spectra, their sensitivity to any variations of the optical properties of all investigated phases in the system and their spatial distribution, are in fact the major strength of the technique. The ability to perform good research in very complex systems is probably the most unique capability of the technique. It could be concluded that the direct monitoring of the surface of minerals with changing solution chemistry is extremely important to understand and determinate the critical parameters that control surface phenomena. These findings are important from the flotation point of view and found already industrial applications.
References Brunner, H., Mayer, U., Hoffmann, H., 1997. External reflection infrared spectroscopy of anisotropic adsorbate layers on dielectric substrates. Appl. Spectrosc. 51, 209 – 217. Hoffmann, H., Mayer, U., Krischanitz, A., 1995. Structure of alkylsiloxane monolayers on silicon surfaces investigated by external reflection infrared spectroscopy. Langmuir 11, 1304 – 1312. Mielczarski, J.A., 1993. External reflection infrared spectroscopy at metallic, semiconductor and nonmetallic substrates: I. Monolayers films. J. Phys. Chem. 97, 2649 – 2663. Mielczarski, J.A., Mielczarski, E., 1995. Determination of molecular orientation and thickness of self-assembled monolayers of oleate on apatite by FTIR reflection spectroscopy. J. Phys. Chem. 99, 3206 – 3217. Mielczarski, J.A., Mielczarski, E., 1999. Infrared external reflection spectroscopy of adsorbed monolayers in a region of strong absorption of substrate. J. Phys. Chem. 103, 5852 – 5859. Mielczarski, J.A., Yoon, R.H., 1989. FTIR external reflection study of molecular orientation in spontaneously adsorbed layers on low-absorption substrates. J. Phys. Chem. 93, 2034 – 2038.
140
E. Mielczarski, J.A. Mielczarski / Int. J. Miner. Process. 72 (2003) 129–140
Mielczarski, J.A., Mielczarski, E., Zachwieja, J., Cases, J.M., 1995. In situ and ex situ infrared studies of nature and structure of thiol monolayers adsorbed on cuprous sulfide at controlled potential. Langmuir 11, 2787 – 2799. Mielczarski, J.A., Mielczarski, E., Cases, J.M., 1996. Interaction of amyl xanthate with chalcopyrite, tetrahedrite and tennantite at controlled potentials. Simulation and spectroelectrochemical results for two component adsorption layers. Langmuir 12, 6521 – 6529. Mielczarski, J.A., Mielczarski, E., Cases, J.M., 1997. Infrared evaluation of composition and structure of ethyl xanthate monolayers produced on chalcopyrite, tetrahedrite and tennantite at controlled potentials. J. Colloid Interface Sci. 188, 150 – 161. Mielczarski, J.A., Mielczarski, E., Cases, J.M., 1998a. Influence of chain length on adsorption of xanthates on chalcopyrite. Int. J. Miner. Proc. 52, 215 – 231.
Mielczarski, E., Mielczarski, J.A., Cases, J.M., 1998b. Molecular recognition effect in monolayer formation of oleate on fluorite. Langmuir 14, 1739 – 1747. Mielczarski, J.A., Mielczarski, E., Cases, J.M., 1999. Dynamic of fluorite – oleate interactions. Langmuir 15, 500 – 508. Mielczarski, E., Mielczarski, J.A., Cases, J.M., Rai, B., Pradip, 2002. Influence of solution conditions and mineral surface structure on the formation of oleate adsorption layers on fluorite. Colloids Surf. 205, 73 – 84. Parikh, A.N., Allara, D.L., 1992. Quantitative determination of molecular structure in multilayered thin films of biaxial and lower symmetry from photon spectroscopies. J. Chem. Phys. 96, 927 – 945. Yen, Y.S., Wong, J.S., 1988. Intriguing absorption band behavior of IR reflectance spectra of silicon dioxide on silicon. Appl. Spectrosc. 42, 598 – 604.