Mechanisms of Amine–Quartz Interaction in the Absence and Presence of Alcohols Studied by Spectroscopic Methods

Journal of Colloid and Interface Science 256, 59–72 (2002) doi:10.1006/jcis.2001.7895

Mechanisms of Amine–Quartz Interaction in the Absence and Presence of Alcohols Studied by Spectroscopic Methods A. Vidyadhar,∗ K. Hanumantha Rao,∗,1 I. V. Chernyshova,† Pradip,‡ and K. S. E. Forssberg∗ ∗ Division of Mineral Processing, Lule˚a University of Technology, SE-971 87 Lule˚a, Sweden; †St. Petersburg State Technical University, Polytechnicheskaya 29, 19521 St. Petersburg, Russia; and ‡Tata Research Development & Design Centre, 54 Hadapsar Industrial Estate, Pune 411 013, India Received May 13, 2001; accepted July 30, 2001; published online October 5, 2001

in the past, mostly by indirect methods, such as measurement of contact angle, zeta-potential, surface forces, and recovery response (1–5). There is a general consensus that the mechanism of amine adsorption on silicate depends on pH. Within the pH 2–7 region, where the surface potential of silicate is negative, the literature is consistent with the Gaudin–Fuerstenau– Somasundaran model (6, 7), where the ammonium ions undergo physisorption at the silicate–water interface much below the critical micelle concentration (CMC). They are electrostatically held in the stern layer as individual counterions behaving almost as an indifferent electrolyte. At a certain critical value of the total amine concentration in the double layer, referred as the critical hemimicelle concentration (CHC), the electrostatic repulsion between positively charged head groups becomes lesser than the force resulting from the removal of the hydrocarbon chains from water. This involves a lowering of free energy of the system by the interactions of the alkyl chains into two-dimensional aggregates called hemimicelles. At this point, a steep rise in the adsorption curves, the sign reversal of zeta-potential, and a sharp increase in floatability are observed simultaneously. It is implied that the head groups of the hemimicelles are directed to the surface and the hydrocarbon tails are oriented toward the solution. Because of the negative surface charge of silicates, the local concentration in the double layer is higher than that in the bulk concentration attaining the CMC at the surface region and the hemimicellization is regarded as a superficial analog of micellization. According to the model calculations, at the bulk concentration of amine equal to the CHC (4 × 10−4 M for dodecyl amine at pH 6), the concentration of amine near the quartz surface is 0.014 M, which is higher than the CMC (0.013 M). The CHC/CMC ratio is approximately 0.01–0.04. Other mechanisms were proposed for the adsorption of amine from alkaline solutions when some part of the surfactant is present in molecular form. According to Smith and Scott (2), the amine solubility limit (2 × 10−5 M) can be exceeded within the quartz-solution interface, provided the temperature is below the Krafft point (26◦ C for dodecylammonium chloride). As a result, surface precipitation takes place, rendering the surface hydrophobic. At the bulk concentration equal to CHC of dodecylamine (2 × 10−4 M), the calculations showed that the pH

The adsorption mechanism of long-chain alkylamines and their acetate salts on quartz was investigated using Hallimond flotation, zeta-potential, Fourier transform infrared, and X-ray photoelectron spectroscopy (XPS) studies at neutral pH 6–7. The influence of long-chain alcohols on the adsorption of amines in mixed amine– alcohol on quartz was also examined. It is shown by infrared spectroscopy that in differentiation to the electrostatic adsorption model of Gaudin–Fuerstenau–Somasundaran, amine cations form strong hydrogen bonds with the surface silanol groups. The XPS spectra revealed the presence of neutral amine molecules together with protonated ammonium ions at and above the critical hemimicelle concentration. The acetate counterions were found to influence the amine adsorption. Possible mechanisms of adsorption based on these observations were discussed. For the first time it was proved spectroscopically that coadsorption of long-chain alcohols along with amine cations leads to formation of a closely packed surface layer, as compared to the case of adsorption of pure amine alone at the same concentration. The highest order and packing at the surface are observed when the alkyl chain length of mixed amine and alcohol are the same. The condition of same chain length of amines and alcohols adsorbing at the surface corresponds to maximum flotation recovery. The results also confirmed the synergistic enhancement of amine adsorption in the presence of alcohols. The mechanism for mixed long-chain amine and alcohol adsorption onto quartz is consistent with the primary adsorption species of alkylammonium– water–alcohol complex, where deprotonation of ammonium groups in the adsorbed layer leading to two-dimensional precipitation of molecular amine was illustrated. C 2002 Elsevier Science (USA) Key Words: quartz; adsorption; flotation; zeta-potential; FTIR; XPS.

1. INTRODUCTION

The flotation of silicates is invariably carried out with unsubstituted long-chain alkyl amines. The primary alkylammonium salts are used as flotation collectors principally because of their amphilitic character and their relatively high solubility. The adsorption of amines on silicate minerals was studied extensively

1 To whom correspondence should be addressed. E-mail: Hanumantha. [email protected]. Fax: +46 920 97364.

59

0021-9797/02 $35.00

 C 2002 Elsevier Science (USA)

All rights reserved.

60

VIDYADHAR ET AL.

of surface solubility is 7.4. Laskowski (3) attributed the increased adsorption of amine ions at higher pH values to the coadsorption of the neutral amine molecules and ions, which allows one to understand why hemimicellization is enhanced in the pH range of 8–10 and a closely packed monolayer is formed. At higher concentrations and in neutral and slightly alkaline conditions, hydrophobicity ceases. The phase-separated amine precipitate, the second layer with the amine heads oriented toward solution, bilayer aggregates, and flip-flop orientation have been attributed as the possible structures on the surface layer that explain this effect. Cases and co-workers (8, 9) advocated the two-dimensional condensation theory to explain the stepwise increase in the shape of the adsorption isotherms of alkylammonium ions on biotite, which is the basis of the admicelle hypothesis by Scamehorn et al. (10). The basic proposition is that the break in the adsorption isotherm at CHC corresponds to two-dimensional condensation of the surfactant at the interface. The chemical identity of the two-dimensional phase is considered to be like the threedimensional hydrated crystal state or micelles of the surfactant at temperatures below and above the Krafft point, respectively. Since the real surface is heterogeneous, it is assumed that such a surface can be represented as a sum of several homogeneous domains with different surface energetic sites. Allowing different specific values for two-dimensional condensation on these different energetic homogeneous domains, the stepped isotherm is accounted for by the successive condensation, with the most energetic domains filled first with increasing surfactant concentration. The theory predicts that the formation of bilayer through tail–tail interactions is always characterized by the pronounced vertical step in the adsorption isotherm since the surface is rendered energetically homogeneous owing to the formation of the first layer. The two-dimensional condensation theory is very general in its approach and differentiates the situations below and above the Krafft point and takes into account the adsorbate– adsorbent and adsorbate–adsorbate interactions and the possibility of surface condensation. The macroscopic characteristics of the abundant surfactant adsorption data available in the literature can be interpreted within the frameworks of different models (11–14). Recent investigations showed that the origin of the counterion, more precisely its affinity to the amine head group and specific absorptivity to the mineral surface, affects the adsorption of cationic surfactants. In silicate flotation, the acetate ammonium salts are commonly used and the effect of an acetate counterion has not been distinguished even though previous studies have been performed for both the acetate and halogen salts (2, 4). At the same time it is well-known that an acetate ion can be hydrogen bonded (15) or chemisorbed (16) on a silicate surface. In the presence of nonionic surfactants (e.g., alcohols), the hemimicelle phenomenon takes place at a smaller concentration of amine. The intrusion of nonionics between the amine cations shields the adjacent amine head–head repulsion (17) and low-

ers the CMC of the amine hemimicelles. The coadsorption of neutral molecules results in a closer pack and, therefore, more hydrophobic coating of the surface, improving the flotation process further. The amines do not form a close-packed monolayer on silicate surface since the cross-sectional area of the aminecharged head is about 0.25 nm2 (for dodecylamine ion (18)) and the area occupied by each hydrocarbon chain in a closely packed monolayer is about 0.20 nm2 (19), while the area around the negative-charged oxygen hexagon on the silicate surface is about twice as large, 0.48 nm2 (20, 21). The increase in the contact angles of quartz (22), the increase in the hydrophobic force between mica surfaces (21), and the increase in the flotation response of corundum (23) with the addition of neutral alcohol molecules in dodecylamine revealed the coadsorption of alcohol where it is not adsorbed on the quartz surface by itself. The most hydrophobic effect was observed for coadsorption of the pair dodecyl amine–dodecyl alcohol from the surface force investigations (21). Although the general mechanism of the coadsorption of the long-chain amines and alcohols on silicates was studied, there was no direct spectroscopic observation of coadsorption of amines and alcohols on silicates or of dependence of the adsorbed layer structure on the relationship between chain lengths of both components. Moreover, the individual adsorption of the primary amine acetate and alcohol on quartz has yet to be studied spectroscopically. The aim of the present work is to study individual and coadsorption of the long-chain amines, their acetate salts, and alcohols on quartz by using direct Fourier transform infrared (FTIR) spectroscopic methods (diffuse reflectance Fourier transform (DRIFT) and infrared reflection–absorption spectroscopy (IRRAS)) and X-ray photoelectron spectroscopy (XPS). It is intended to compare the spectroscopic data to flotation and zetapotential results so as to distinguish the spectral characteristics for qualitative assessment of the adsorbed layer for surface hydrophobicity or hydrophilicity. The alkyl chain-length effect in the coadsorption of long-chain amines and alcohols onto quartz is also investigated. 2. EXPERIMENTAL

2.1. Materials The pure crystalline quartz sample provided by Mevior S. A., Greece, originating from their Thessaloniki deposit was used. The sample was crushed, ground, and wet sieved to obtain size fractions of −150 + 38 µm. The −38-µm-size particles were ground further in an agate mortar and microsieved in an ultrasonic bath to a −5-µm-size fraction. The size fractions of −150 + 38 µm and −5 µm having the BET specific surface areas of 0.09 and 1.30 m2 g−1 were used for flotation tests and zeta-potential measurements, respectively. The chemical analysis showed that the sample is quite pure, containing more than 99.0 wt % SiO2 and traces of aluminum (Al2 O3 , 0.1 wt %), calcium (CaO, 0.09 wt %), and iron impurities (Fe2 O3 , 0.046 wt %).

SPECTROSCOPIC STUDIES OF AMINE–QUARTZ INTERACTION

2.2. Reagents The C8 , C12 and C16 alkyl amines with 99% purity were from Akzo Nobel AB, Sweden. The acetate salts of C12 and C16 amines were prepared in benzene solvent by mixing equimolar amounts of the respective amine and acetic acid. The acetate salt crystallizes below the freezing temperature and it was purified thrice by recrystallization using fresh benzene each time. The pure amine, when dissolved in the presence of HCl, is referred to as amine solution in the text instead of amine hydrochloride salt solution. The SIGMA spectroscopic-grade alcohols of different alkyl chain lengths C8 , C10 , C12 , C14 and C16 were purchased from Kebo, Sweden. The pure alcohol solutions were prepared first by dissolving the alcohol in ethanol and the resultant aqueous solution of specified concentration contained 5% ethanol. Analar grade NaOH and HCl were used for pH adjustment and deionized water was used in all the experiments. 2.3. Flotation Tests The single mineral flotation tests were made using a Hallimond cell 100 ml in volume. Exactly 1.0 g of the mineral sample was conditioned first in a predetermined concentration of amine solution for 5 min and the suspension was transferred to the flotation cell. The flotation was conducted for 1 min at an air flow rate of 8 ml min−1 . All the tests were performed at a neutral pH region of 6–7. When the tests were performed in a mixed amine/alcohol system, the mineral was conditioned in a solution containing both the reagents at a specified concentration ratio. 2.4. Zeta-Potential Measurements Zeta-potentials were determined using a Laser Zee Meter (Pen Kem Inc., model 501) equipped with video system apparatus employing a flat cell. A 1.0 g L−1 quartz suspension was prepared in 10−3 KNO3 supporting electrolyte solutions, conditioned for 1 h at room temperature (22◦ C) in the presence of a predetermined concentration of reagents and pH. The pH of the suspension at the time of measurement was reported in the results. After the measurements, the suspension was filtered through Millipore filter paper (pore size 0.22 µm) and the solids were air-dried before recording the DRIFT infrared spectrum. 2.5. Diffuse Reflectance FTIR Measurements The infrared spectra were registered for all samples after zetapotential measurements on the air-dried −5-µm powder. The FTIR spectra were obtained using a Perkin–Elmer 2000 spectrometer with its own diffuse reflectance attachment. The typical spectrum was an average of 200 scans measured at 4 cm−1 resolution with a narrow-band liquid nitrogen cooled mercury cadmium telluride (MCT) detector. At the concentration of reagents used, the intensity of the bands with respect to adsorbed layers was low and the samples were not mixed with KBr since dissolution further lowers the sensitivity. Thus the untreated quartz powder was used as reference and the absorbance units were defined by the decimal logarithm of the ratio of initial quartz

61

reflectance to the sample one. The atmospheric water was always subtracted from the sample spectrum. The area under the alkyl chain bands was measured with the facility available within spectral manipulation. 2.6. Reflection Absorbance FTIR Measurements A plate with dimensions of about 20 × 20 mm2 was cut from a single quartz crystal and the working surface was prepared by polishing successively with SiC papers down to the 0.5-µm size and washing thoroughly with deionized water. The surface was conditioned with the required reagent solution for about 5 min and the excess solution was removed from the surface with a filter paper. The IRRAS spectra of the surfactant film at the quartz–air interface were obtained using a Harrick, Inc., IRRAS accessory immediately after contact with the solution. All the spectra were collected by coadding 600 scans at 4 cm−1 resolution with a narrow-band liquid nitrogen cooled MCT detector. The angle of incidence of unpolarized radiation was 10◦ . At this angle practically only the modes with the transition dipole moment (TDM) parallel to the surface contribute to the IRRAS spectrum, and according to the spectral simulations the absorption bands of adsorbates appear negative (24). 2.7. XPS Measurements The XPS spectra of quartz powder and fractured surfaces, and both the surfaces treated with primary alkyl amines and alcohols, were recorded with an AXIS Ultra (Kratos) electron spectrometer under Al monoirradiation and Mg irradiation with sample cooling. The vacuum in the sample analysis chamber during measurements was 10−8 Torr. A value of 285.0 eV was adopted as the standard C(1s) binding energy. 3. RESULTS AND DISCUSSION

3.1. Flotation Studies The flotation of quartz as a function of C8 , C12 , and C16 amine chlorides at pH 6–7 is shown in Fig. 1. The flotation response of quartz with C12 and C16 amine acetate salts is also presented in the same figure. Numerous microflotation studies of quartz have been made using various cationic and anionic collectors, and the present tests are intended to compare the response of quartz when floated with amine and its acetate salt of the same alkyl chain length to its response when floated in the presence of amine and alcohols of varying alkyl chain lengths. Furthermore, since the recent infrared studies of various carboxylic acids, including acetic acid, onto a set of common mineral substrates indicated the adsorption of an acetate ion through weak H-bonding on the surface of quartz (15), the effect of an acetate counterion on flotation was targeted for investigation. The effect of alkyl chain length on the flotation behavior of quartz shown in Fig. 1 is precisely the same as was presented by Fuerstenau et al. in 1964 (25). The onset of hemimicelle formation for C8 , C12 and C16 amines corresponds to about

62

VIDYADHAR ET AL.

FIG. 1. Flotation of quartz with C8 , C12 , and C16 amines as a function of concentration at pH 6–7.

1 × 10−4 , 1 × 10−5 , and 2 × 10−6 M, respectively, illustrating that the critical concentration decreases about one order magnitude with each four-carbon increase in alkyl chain length. The counterions of amine appear to have no influence since the flotation response is comparable to amine chloride and acetate solutions. By outlining the fact that the flotation recovery is not sensitive to surface coverage of collector and that one hydrophobic patch on the surface is enough for the particle to adhere to an air bubble and float, the effect of an acetate ion on amine adsorption cannot be ruled out. Figure 2 shows the effect of alcohol chain length on quartz flotation when mixed with C12 and C16 amine acetate solutions. The results in a mixed composition of a particular C12 amine and alcohol concentration as a function of alcohol chain length indicate that the quartz flotation reaches maximum at C12 alcohol, matching the amine chain length. At alcohol chain lengths greater than C12 , the quartz recovery decreased, although insignificantly. The results at the three concentrations of alcohols with varying chain length in the presence of 5 × 10−6 M C12

FIG. 2. Effect of alkyl chain length in mixed amine/alcohol on quartz flotation at pH 6–7.

amine demonstrate the same behavior. The same behavior of maximum recovery of quartz at equal chain lengths of amine and alcohol in mixed composition is also observed with C16 amine. However, in this case 80% recovery is almost reached at C12 alcohol and 90% is attained at C16 alcohol. Recognizing the fact that the flotation response is not directly related to the degree of collector coverage and thus hydrophobicity, the results signify the alkyl chain length compatibility in mixed surfactant systems at the quartz–water interface. The adsorption in a system containing both charged and neutral ions should be greater than one containing only surfaceactive ions, since the neutral molecule heads can actually screen the repulsion between the charged heads of the ions at the interface by coadsorption between the collector ions. Regarding this aspect, it is worthwhile to remember that the maximum flotation of minerals either with cationic amine or with anionic oleate collector occurs at a pH range where the surfactant exists as ions and neutral molecules (26, 27). This shows the greater surface activity of a neutral molecule in the presence of its ions. Being a single-surfactant system but existing in neutral and ionic form with the same alkyl chain length, the adsorbed layer could be very packed, thereby increasing the surface hydrophobic character and flotation. When the two components differ in chain length, the adsorbed layer expands and the spacing between adjacent molecules is found to increase at the liquid–air interface, presumably caused by the thermal motion of the unequal chain lengths in a mixed monolayer (28) (i.e., formation of gauche defects near the end of chains and longitudinal displacement of the chains). The same phenomena is expected to occur at the solid– liquid interface which explains the observed flotation response of quartz at equal alkyl chain lengths of amine and alcohol compared to unequal chain lengths in mixed composition. 3.2. Zeta-Potential Studies In Fig. 3, the zeta-potentials of quartz under conditions identical to that of flotation studies are presented. The hemimicelle concentration for C12 amine is reached at about 4 × 10−4 M,

FIG. 3. Zeta-potential of quartz and adsorption of amine as a function of C8 , C12 , and C16 amines concentration at pH 6–7.

SPECTROSCOPIC STUDIES OF AMINE–QUARTZ INTERACTION

and for acetate salt, it occurs at a slightly higher concentration. These concentrations are more than one order of magnitude higher when compared to flotation recovery curves. Similarly, the critical concentration for C16 amine where the zeta-potential raises toward positive is about 2 × 10−5 M in comparison to 1 × 10−6 M, where the onset of flotation is observed. For C8 amine, the zeta-potentials are either the same or increased in magnitude of negative charge until 2 × 10−3 M concentration, showing that the hemimicelle concentration is not attained in the concentration range studied. Neverthless, there is the onset in flotation with C8 amine at about 1 × 10−4 M. However, a good correlation between the adsorption of amine as evidenced from the area under the alkyl chain bands (3000–2800 cm−1 ) and zeta-potentials is noticed (Fig. 3). The results show that there is a difference in the critical hemimicelle concentration (CHC) in the zeta-potential and flotation curves. Considering the respective specific surface areas of the two size fractions (i.e., 1.30 and 0.09 m2 g−1 ), the total surface area involved in either zeta-potential or flotation studies is approximately the same. The good correlation between flotation response, zetapotential, and adsorption density in the C12 amine/quartz system reported by Fuerstenau (29), was from the zeta-potentials obtained by streaming potential measurements using coarser particles. The onset amine concentration (CHC) of the steep increase in the zeta-potential curves with amines and their salts is varied, with the acetate salt occurring at a slightly higher concentration. Coadsorption of an acetate ion would suppress the adsorption of amine so that the CHC for the ammonium acetate salt would be attained at a higher bulk amine concentration. The infrared studies presented later showed that the acetate ions are adsorbed on a quartz surface. The adsorption of acetate ions on a silicate surface through hydrogen and chemical bonding was reported (15). The XPS studies of the halogen salts of hexadecyltrimethylamine (CTA) showed that the coadsorption of counterions suppressing the amine adsorption onto mica is in the same order of I−  Br− > Cl− as that of the binding energies of these ions to CTA+ (30).

63

FIG. 4. DRIFT spectra of dodecyl alcohol, dodecyl amine acetate, and dodecyl amine.

and protonated forms can be identified with the appearance and disappearance of the band at 3333 cm−1 . The dodecyl alcohol spectra show a broad band centering around 3318 cm−1 , characteristic of the H-bonded OH . . . O of the ν(OH) group (31) in addition to the bands characterizing alkyl chains. The spectra also display several bands in the region 1750–1000 cm−1 but in the same region quartz exhibits strong absorption and thus it is difficult to identify any bands under mono- or submonolayer adsorption. Thus this region is not found in the present investigations because of the low concentrations employed. Figure 5 shows the spectra of quartz treated with 1 × 10−3 M C12 amine acetate at different pH values. Above pH 8.6, the spectra exhibit an intense absorption band corresponding to ν(NH2 ) and at lower pH values this band is totally absent. The ionization of dodecyl amine in aqueous media (pK 10.63) and the solution chemistry was studied thoroughly. In almost all cases the

3.3. Infrared Diffuse Reflectance Spectral Studies The reference DRIFT spectra of dodecyl amine, its acetate salt, and dodecyl alcohol are shown in Fig. 4. The bands characteristic of alkyl chains, νas (CH3 ), νas (CH2 ), and νs (CH2 ) groups, for dodecyl amine are identified at 2956, 2920, and 2852.5 cm−1 , respectively (31). The corresponding bands for its acetate salt are observed at 2950, 2921.6, and 2851.6 cm−1 . However, the spectrum of acetate salt is composed of several other bands (i.e., 2963, 2933, 2900, 2890, and 2870.7 cm−1 ) in the alkyl chain region owing to the acetate ion and, probably, to another packing of C–H chains in the amine acetate crystallites. A sharp intensity band at 3333 cm−1 in the amine spectrum is due to νas (NH2 ) and νs (NH2 ), while in the case of acetate salt two broad bands at 3000 and 2600 cm−1 are observed, which are assigned + to νas (NH+ 3 ) and νs (NH3 ), respectively (31–33). The reference spectra show that the amine molecule in the bulk phase in neutral

FIG. 5. DRIFT spectra of quartz treated with 1 mM C12 amine acetate at different pH values.

64

VIDYADHAR ET AL.

explanations for zeta-potential and flotation results were based mainly on the thermodynamic equilibrium solution chemistry model of the predominance of the ionized form, RNH+ 3 , in the acidic-to-neutral range and the molecular form, RNH2 , in very alkaline solutions without direct methods of investigations (26, 27). The present spectra assert that the molecular amine precipitate is on the surface of quartz in the basic pH region. The precipitated amine is comprehended in the severalfold increase in the intensity of alkyl chain bands at pH 9.5 and 10.6, as well as in the appearance of an intense molecular amine band. It is of importance that the analogous dependence of the spectra is observed with increasing amine concentration at pH 6–7. Figure 6 shows the selected spectra of quartz with an increasing initial concentration of C16 amine acetate in the region of 4000–2600 cm−1 . The intensity of alkyl chain bands increases with increasing amine concentration, corroborating increased adsorption of amine. It is interesting that at and above 5 × 10−5 M initial amine concentration, the spectra displayed a molecular amine band and the alkyl chain bands are intense. This is not the case with C12 amine acetate. The presence of molecular amine at higher concentrations of C16 amine acetate might be related to its low solubility in comparison to dodecyl amine acetate. The solubility of dodecyl amine was 2 × 10−5 M and the solubility decreases about one-third for each additional hydrocarbon group added to the alkyl chain (34). The spectra also showed a negative band at 3745 cm−1 , which is present in the initial quartz spectrum used as reference. This band is assigned to the stretching vibration of hydroxyl groups arising from the surface silanol groups (33, 35). The negative band indicates that the silanol groups are interacting with the adsorbed amine. The surface silanol groups are closely examined as a function of amine concentration. The normalized νas (CH2 ) peak area in between the wavenumbers 2945 and 2902 cm−1 , and the SiOH peak area in between 3770 and 3660 cm−1 wavenumbers as a function of C12 amine

FIG. 6. DRIFT spectra of quartz with increasing initial concentration of C16 amine acetate at pH 6–7.

FIG. 7.

Influence of C12 amine adsorption on quartz silanol groups.

and its acetate salt concentration, are shown in Fig. 7. Although the CH2 peak area curves characterizing the absorbed amine appear to be the same with increasing amine concentration until the break corresponding to the onset of CHC, the negative silanol groups area increases. This increase in negative area reached maximum at a certain bulk amine concentration. It signifies that the surface free silanol groups concentration is diminished because of the amine interaction with the silanol groups. The bulk amine concentration up to which the silanol groups are affected and the break in the adsorption curve (i.e., CHC) are found to be different in the case of amine and amine acetate. It suggests the influence of an acetate counterion on amine adsorption. The DRIFT spectra with increasing C12 amines concentration also revealed that at 2 × 10−3 M amine, the molecular amine band 3330 cm−1 appeared which is absent in the case of amine acetate in the entire concentration range studied. Since pH 6–7 is the same for the cases of both amine and amine acetate, the absence of a molecular amine identifying no surface precipitation in the case of amine acetate up to 1 × 10−2 M concentration studied seems to be irregular. This effect can be attributed to coadsorption of an acetate ion with a dodecyl amine cation. The DRIFT spectra adsorption results with C16 amines are presented in Fig. 8. The disappearence of surface silanol groups with increasing amine concentration is analogous to the case of C12 amines. The spectra showed additional bands at 1650, 1570, 1506, 1486 cm−1 , and so forth, above 2 × 10−5 M, and the 3330 cm−1 band at and above 5 × 10−5 M (Fig. 6). In the case of C12 amine, these spectral features are observed only at 2 × 10−3 M and were absent in the case of C12 amine acetate. The extra bands when compared to the DRIFT spectra of molecular and protonated forms of amines (Fig. 4) disclose that these are connected with the bulk phase of the molecular amine. The IR spectra displayed in Fig. 9 reveal that the probable interaction of amine with surface silanol groups is through H-bonding. A broad structural band centered at 3250–3000 cm−1 is distinctive of Hbonded ν(N+ -H), ν(O-H), and ν(N-H) stretching vibrations.

SPECTROSCOPIC STUDIES OF AMINE–QUARTZ INTERACTION

65

coadsorption and totally disappears for C12 amine–C16 alcohol and C8 amine-C16 alcohol pairs. Since the negative intensity of the silanol band arises from the interaction of silanols with the adsorbate, as explained earlier, the free silanols are less involved in the coadsorption as well in the adsorption of the alcohol, when compared to the adsorption of pure amines. 3.4. IRRAS Studies

FIG. 8.

Influence of C16 amine adsorption on quartz silanol groups.

The DRIFT spectra of quartz powder conditioned in 1 : 1 mixed solutions of alkyl ammonium acetate and alkyl alcohol of different chain lengths at a total concentration of 1 × 10−5 M are shown in Fig. 10. The reference quartz spectrum (Fig. 10a, curve 0) shows a narrow band at 3745 cm−1 due to OH stretching vibrations of surface-isolated silanol groups and the complex absorption band in the region 3000–3700 cm−1 due to adsorbed H-bonded hydroxyls and water (33). The spectra of quartz conditioned with mixed solutions of C16 amine acetate–C12 alcohol and with reverse chain lengths (Fig. 10a, curves 1 and 2) and binary solutions of C16 amine acetate–C8 alcohol and reverse chain lengths (Fig. 10b, curves 1 and 2) display characteristic alkyl chain bands due to adsorbed surfactants. The intensities of the CH stretching bands are nearly the same within the experimental error for the reverse chain lengths, illustrating that the quantity of adsorbed surfactants is the same. However, the adsorbed amount increased for the amine–alcohol pair when the sum chain length was more, which is in agreement with the general tendency of increasing adsorption with increasing alkyl chain length. When the CH stretching band intensities of the adsorbed pure amine and alcohol (Fig. 10c, curves 3 and 4) are compared, the adsorbed amine is approximately three times larger than that of alcohol at the same alkyl chain length. In the case of competitive adsorption of monomers from the binary solutions of amine and alcohols with reverse chain lengths, the surface coverage is expected to be larger for a mixture with longer amine homologues. However, if the adsorbed species are the 1 : 1 solvated associates, the surface coverage should be similar, which is observed. It could be that these solvated associates are the ion– molecule pairs where the alcohol hydroxyls act as base while the ammonium head group behaves as acid. Such dimers have been suggested to be formed between the protonated amine and alcohol (26). The intense negative band at 3745 cm−1 due to free silanols in the spectrum of the quartz conditioned with the solution of the pure amine (Fig. 10c, curve 3) diminishes in the case of

It is well established in numerous studies (36, 37) that the frequencies of the CH2 and CH3 stretching bands of hydrocarbon chains are extremely sensitive to the conformational ordering of the chains in a layer. When the chains are highly ordered (all-trans Zigzag conformation), the narrow absorption bands asym sym appear at 2918 (CH2 ) and 2848 cm−1 (CH2 ) in the infrared spectrum of the layer. On the other hand, if conformational disorder is included in the chains, they shift upward to 2926 and 2856 cm−1 , depending on the gauche conformers in the average orientation. Although these frequency shifts are small, modern

FIG. 9. DRIFT spectra of quartz conditioned with (A) C12 amine at 2 × 10−4 M (solid line, the spectrum multiplied by 2.5) and 5 × 10−4 M (dashed line), and (B) C16 amine acetate at 1 × 10−5 M (solid line, the spectrum multiplied by 2) and 2 × 10−5 M (dashed line).

66

VIDYADHAR ET AL.

FIG. 10. DRIFT spectra of (a, curve 0) initial quartz and the quartz conditioned in the 1 : 1 binary solutions (pH 6–7) of alkyl ammonium acetate and alkyl alcohol of different chain lengths at a total concentration of 1 × 10−5 M. The component chain lengths are as follows. (a) 1, C16 amine–C12 alcohol; 2, C12 amine–C16 alcohol. (b) 1, C16 amine–C8 alcohol; 2, C8 amine–C16 alcohol. (c) 1, C16 amine–C16 alcohol; 2, C12 amine–C12 alcohol. For comparison, the quartz conditioned in the solutions of the pure (3) HAAc and (4) C16 alcohol. The absorbance scale is the same for all the graphs.

FTIR spectrometers permit their routine determination with a precision of better than ±0.1 cm−1 . It was found that generally the hydrocarbon order decreases with decreasing chain length (38). Another important characteristic of the molecule organization in a layer is the absorption bandwidth, which is proportional to the degree of the molecule mobility within the layer. It was found that the bandwidth increases with temperature when the chain packing disorder increases (39). Figure 11 shows the IRRAS spectra in the stretching C–H mode region of dodecyl amine coadsorbed with long-chain alcohols of 8, 12, and 16 carbons on quartz from the 1 : 1 aqueous (natural pH) solutions of the total concentration 2 × 10−5 M. asym sym All the spectra display the distinct CH2 and CH2 bands. An

important point is that the lowest frequencies and the narrowest widths of these bands are exhibited by the spectrum of dodecyl amine coadsorbed with dodecyl alcohol. This observation testifies to the highest order and packing of the adsorbed species, when the chain lengths of the coadsorbents are equal and can be regarded as a spectroscopic confirmation of the results of the surface force measurements by Yoon and Ravishankar (20, 21). They found that the coadsorption of dodecyl amine with dodecyl alcohol on mica yields the most hydrophobic effect (the strongest hydrophobic force) as compared to the other cases studied. Using the molar absorptivity of the hydrocarbon chains, the coverage by the adsorbed molecules was roughly estimated to be equal to about 0.5 monolayer in all the cases considered (40, 41). 3.5. XPS Studies

FIG. 11. IRRAS spectra of quartz surface after 5-min conditioning in a 1 : 1 solution of dodecyl amine and (1) octyl alcohol, (2) dodecyl alcohol, and (3) hexadecyl alcohol with the total concentration 2 × 10−5 M at natural pH.

The XPS spectra of the reference compounds and the quartz powder and fractured surface treated with alkyl amines and binary solutions of alkyl amine and alcohol were recorded and the results are summarized in Table 1. It should be noted that the quartz sample is free from nitrogen impurity and is essential for the present investigations. The XPS spectral results demonstrate the appearance of the N(1s) signal of the amino groups and an increase in the intensity of the signal of C(1s) peak by the adsorption of amines on quartz with a simultaneous decrease in the peak intensities of quartz’s silicon and oxygen. The C(1s) spectra of the quartz powder conditioned with C12 amine and its acetate salt at 2 × 10−4 M concentration and with C16 amine at 2 × 10−5 M are presented in Fig. 12. The curve fitting unveiled that the spectra consist of two components, at 286.8 and 288.6 eV. The first component is assigned to carbon in the C–N, CH3 and C–O bonds and the second one to the carbon in the carboxylate group. The C12 amine adsorbed

67

SPECTROSCOPIC STUDIES OF AMINE–QUARTZ INTERACTION

TABLE 1 XPS Characterization of Reference Compounds (Solid DA, DAAc, and the Initial Quartz) and Quartz Treated with Pure Amines and Alcohols and in Mixed 1 : 1 Amine–Alcohol Solutions Element at % (BE, eV) Substrate

Powder

Sample

N

C

O

Si

Natural quartz Solid DA

0 4.65 (399.5)

57.29 (532.8) 2.58 (531.2)

29.94 (103.6) 0

Solid DAAc

1.46 (399.4) 3.78 (401.1)

2 × 10−4 M DACl

0.64 (399.5) 0.24 (401.6) 0.39 (400.1) 0.35 (399.8) 0.2 (402.0)

12.77 (285.0) 89.36 (285.0) 0.07 (286) 0.02 (287.7) 76.19 (285) 5.66 (286.9) 3.94 (288.4) 24.33 (285.0)

49.11

25.66

15.70 (285.0) 16.53 (285.0) 1.48 (286.8) 0.63 (288.6) 19.6 (285.0) 1.56 (286.9) 0.83 (288.6) 11.53 (285.0) 1.64 (286.6)

54.9 53.86

29.01 27.24

51.06

25.88

56.38 (532.8)

30.11 (103.5)

60.41 (532.7)

31.07 (103.5)

29.69 (532.8) 2.45 (531.6)

18.2 (103.5)

37.31 (532.4) 2.63 (530.7) 29.69 (532.8) 2.45 (531.6)

20.16 (103.2)

4 × 10−5 M DAAc 2 × 10−4 M DAAc

Fracture Surface

2 × 10−5 M HACl

0.38 (399.4) 0.24 (401.6)

DAAc/C16 -alc. (1 : 1) 1 × 10−5 M HAAc/C12 -alc. (1 : 1) 1 × 10−5 M 4 × 10−5 M DACl

0.24 (400.1) 0.10 (402.2)

2 × 10−4 M DACl 2 × 10−4 M DAAc 1 × 10−4 M HAAc 2 × 10−4 M HAAc 1 × 10−4 M HAAc HAAc/C16 -alc. (1 : 1) 5 × 10−5 M HAAc/C16 -alc. (1 : 1) 1 × 10−4 M

0.19 (400.1) 0.12 (402.0) 1.98 (400.1)

2.29 (400.1) 0.55 (401.9) 1.23 (400.1)

0.39 (399.3) 0.32 (401.5) 0.64 (399.5) 0.54 (401.4) 0.39 (399.3) 0.32 (401.5) 0.29 (399.7) 0.12 (402.2) 0.28 (399.5) 0.18 (401.6)

6.56 (285.0) 1.18 (286.6) 0.47 (288.3) 43.08 (285.0) 3.23 (286.0) 2.12 (288.6) 33.10 (285.0) 3.97 (286.4) 43.08 (285.0) 3.23 (286.0) 2.12 (288.6) 23.09 (285.0) 1.95 (286.0) 33.43 (285.0) 3.86 (286.0) 1.57 (288.1) 23.09 (285.0) 1.95 (286.0) 21.44 (285.0) 1.59 (286.4) 0.22 (286.6) 23.37 (285.0) 1.58 (286.4)

7.99 (531.2) 0.99 (533)

0

18.2 (103.5)

47.5 (532.5)

26.75 (103.3)

0.84 (530.8) 37.07 (532.5)

22.02 (103.2)

47.5 (532.5)

26.75 (103.3)

48.64 (532.8)

27.71 (103.6)

44.85 (532.5)

29.38 (103.5)

Note. DAAc, C12 amine acetate; DACl, C12 amine; HACl, C16 amine; HAAc, C16 amine acetate.

spectrum indicate a weak Cl(2p) peak. Thus the XPS data assert the coadsorption of amine counterions, with the acetate ions being more adsorbed compared to chloride ions. The acetate ions are found to be coadsorbed approximately with amines by 1 : 1 at 2 × 10−4 M from the peak intensities of the whole N(1s) peak and the 288.6 eV component of the C(1s) peak. The difference in the zeta-potential and adsorption curves of amine and its acetate salt can thus be substantiated by the coadsorption of acetate counter ions (Fig. 3).

Figure 13 shows the N(1s) spectra of the quartz powder treated with 4 × 10−5 and 2 × 10−4 M C12 amine acetate, and 2 × 10−4 M C12 amine. At amine concentration below CHC (curve 3), the spectrum is composed of one component at 400.1 eV. The spectra exhibit two components of N(1s) peak at CHC, with binding energies of 399.5 and 401.5 eV. These values, respectively, correspond to the binding energies of amino and ammonium groups when compared to the spectra of crystalline C12 amine and C12 amine acetate (Table 1). Thus, the

68

VIDYADHAR ET AL.

FIG. 12. XPS C(1s) spectra of quartz powder conditioned with (1) 2 × 10−4 M DACl; (2) 2 × 10−4 M DAAc; and (3) 2 × 10−5 M HACl.

mode points to the H-bond strengthening. Thus, the transition from Eq. [2] to Eq. [1], while strengthening H-bonds occurring at CHC, supports the above assignment of the N(1s) spectra. Thus, the extensively cited three regions of adsorption (10, 44) can further be exemplified as follows: (i) An amine cation is H-bonded to surface silanol groups and this bond becomes stronger after the break in the adsorption characteristics (isotherm, zeta-potential, flotation); (ii) at the break the origin of the adsorbed amine species changes qualitatively, and the molecular amine species appear together with an alkyl ammonium ion attached to the deprotonated silanol group and thereby forming monolayer-thick patches of well-oriented and densely packed adsorbed amine species, rendering the surface highly hydrophobic; and (iii) at higher amine concentration, bulk precipitation of amine takes place. The XPS spectra of the powdered quartz treated with mixed solutions with the reverse chain lengths of C12 amine acetate– C16 alcohol and C16 amine acetate–C12 alcohol at the same total concentration (1 × 10−5 M) with 1 : 1 ratio are given in Fig. 14. The N(1s) spectra show the same total atomic concentrations of nitrogen and the relative amounts of the ionic and molecular forms within the experimental errors. This observation is consistent if the adsorbing species are the 1 : 1 associates of the amine and alcohol formed in the solution. Similar to the spectra in Fig. 14, the spectra recorded with equal chain lengths

adsorbed layer is composed of molecular alkyl amine and alkyl ammonium ions. The adsorption of amine on silanol groups can be realized with the amino group H-bonded to surface silanol attributed to the first component while the second component is attributed to the ammonium group formed owing to a charge transfer in strong H-bond between nitrogen of the amino and silanol groups (40): ≡ SiOH . . . . H2 N-R ↔ ≡ SiO− . . . H+ 3 N-R.

[1]

The above equilibrium exists at the interface once the molecular amine appears at the surface corresponding to CHC. The single peak with binding energy at 400.1 eV (Fig. 13, curve 3) and at concentrations less than CHC is characterized by the ammonium group H-bonded to the silanols as H | ≡ Si − O . . . H − + NH2 − R .

[2]

The NH . . . O bonds are weaker than the N. . . HO bonds (42), and the proton-donating property of surface silanol groups is higher than its proton-accepting property in the H-bonding with a water molecule and acetic acid–acetate ion pair (43). The transition from Eq. [2] to Eq. [1] is observed from infrared spectra where the broad band is centered at 3250 cm−1 at CHC (Fig. 9). It is well-known that a redshift of any H-bonded stretching

FIG. 13. XPS N(1s) spectra of quartz powder conditioned with (1) 2 × 10−4 M DACl; (2) 2 × 10−4 M DAAc; and (3) 4 × 10−5 M DAAc.

SPECTROSCOPIC STUDIES OF AMINE–QUARTZ INTERACTION

69

FIG. 14. XPS N(1s) spectra of a fracture quartz surface conditioned in a 1 : 1 binary solution (pH 6.5) of C16 amine and C16 alcohol at total concentrations of (1) 5 × 10−5 M and (2) 1 × 10−4 M.

FIG. 15. XPS N(1s) spectra of a quartz powder conditioned 10 min in a 1 : 1 binary solution (pH 6.5) of (1) C16 alcohol–C12 amine and (2) C12 alcohol–C16 amine at a total concentration of 1 × 10−5 M.

of mixed surfactants of C16 amine acetate–C16 alcohol at total 5 × 10−5 and 1 × 10−4 M concentrations with 1 : 1 composition, the N(1s) component is split into two, with the peaks at the binding energies of 401.6 and 399.6 eV due to molecular amine and protonated amine, respectively, similar to the adsorption of pure amines (Fig. 15). The relative content of the neutral amine in the mixed film is significantly higher than in the pure amine film at the same total concentration. This implies that the long-chain alcohol promotes the precipitation of the amines. The most probable explanation of the appearance of the neutral amine in the monolayer adsorbed onto quartz either with pure amines or with the binary solutions of amine and alcohol is the precipitation of amine, but in two-dimensional space. The significantly higher amount of neutral amine in the presence of alcohol can be understood if one assumes that the precipitating species is a soluble associate of the protonated amine and alcohol. This species is expected to have lower solubility compared to the separated amine owing to a greater number of hydrocarbon groups at the same hydrophilicity of the head group, as the hydrophilicity of the attached alcohol hydroxyls is negligible when compared to that of the ammonium group. The hemimicelle model cannot explain the presence of neutral amine molecules in the adsorbed film below monolayer coverage and a strong H-bonding between amino head groups and the surface silanols. The present results can be interpreted within the framework of a two-dimensional condensation model if one substitutes the precipitation phenomena for the condensation phe-

nomena. However, before the transition to increased adsorption, the ammonium groups are hydrogen bonded to the negatively charged silanols (Eq. [2]), and when the local concentration at the interface approaches a critical value, the adsorbed layer transforms into a crystalline state owing to precipitation of neutral

FIG. 16. Adsorption according to the two-dimensional/three-dimensional precipitation mechanism.

70

VIDYADHAR ET AL.

amine. At the first step, the process is two-dimensional and the adsorbed neutral amine establishes the equilibrium shown in Eq. [1]. Screening the electrostatic repulsion between head groups, the neutral molecules change the structure of the adsorbed layer substantially, increasing the density of the monolayer. The second phase transition of three-dimensional precipitation occurs when the bulk solubility limit is reached at the surface. This adsorption model is illustrated in Fig. 16 (40). The coadsorption of amine and alcohol via alkylammonium– alcohol associates could explain (i) the intermixing of the surfactants in the adsorbed monolayer, which follows from the observed difference in the tilt angles of the hydrocarbon chains in the mixed and separated monolayers of the amine and alcohol (41), and (ii) the practical independence of the amount of the adsorbed hydrocarbon groups and the precipitated amine from the fact that the given two hydrocarbon chains are distributed between the amine and alcohol. The diameter of a well-packed hydrocarbon chain (4.3– ˚ (45) is larger than the diameter of the ammonium head 4.5 A)

FIG. 17.

˚ (46) and the diameter of the amine head group group (3.7 A) ˚ (3.38 A) (18) and is much larger than that of the alcohol hydroxyl ˚ (47). It follows that if the chains associate by hydropho(ca. 1 A) bic interaction in the soluble amine–alcohol complex, the head ˚ apart (Fig. 17a), which prevents their cogroups are 4.3–4.5 A ordination directly to each other (for comparison, the distance between the terminal atoms in the strong O . . . H–O H-bonds is ˚ (43). It follows that such a complex is rather weak. ca. 2.7 A) However, steric factors allow formation of a strong complex in which both the H-bonding and the hydrophobic interaction contribute if there is a water molecule in between the ammonium and alcohol head groups, as depicted in Fig. 17b. The assumption that the elementary adsorbing species has this or a similar structure allows attribution of the extra band at 3630–3698 cm−1 in the DRIFT spectra of the mixed film to the νOH mode of this incorporated water. The assumption that the elementary adsorbing species is “ammonium–water–alcohol” allows understanding of why the average tilt angle of the hydrocarbon chains in the mixed films

Schemes of the long-chain amine–alcohol complexes.

SPECTROSCOPIC STUDIES OF AMINE–QUARTZ INTERACTION

is larger than that in the pure films of the amine and alcohol (41). The tilt angle of 37◦ implies that the distance between the ˚ which hydroxyl oxygen and the amine nitrogen is 5.3–5.6 A, is realistic for such a complex (Fig. 17b). The following twodimensional precipitation consisting of the splitting off of a proton of the ammonium group can yield the “amine–water– alcohol” species depicted in Figs. 17c and 17d. However, in the species shown in Fig. 17c the lateral H-bonds are rather strong, so the distance between the terminal atoms in the N . . . H–O and ˚ (the oriO . . . H–O hydrogen bonds is on the order of 2.7–3.0 A gin of the H-bonded groups and cooperative enhancement of the H-bond strength due to the mediating water molecule justify this assumption) (43). As a result, the distance between the hydroxyl ˚ Moreoxygen and the amine nitrogen is on the order of 4.9 A. over, the functioning of the alcohol oxygen as a proton acceptor in the H-bond with the incorporated water prevents repeating this unit at the surface. At the same time, the species shown in Fig. 17d permits the head groups to accommodate each other ˚ and makes feasibile the at a distance on the order of 5.4–5.6 A repeating of this motif along the surface. Thus, we suggest that at the silicate surface the two-dimensional precipitation consists of transition from species b to species d shown in Fig. 17. The involvement of the surfactant head groups in the lateral Hbonds reduces interaction between them and the surface silanols, which explains the much smaller number of interacting free surface silanols observed in the DRIFTS spectra. This process explains the synergetic behavior of these surfactants in rendering the surface hydrophobic at much lower total concentrations compared to the concentration of the pure amine solution. From the above positions, the maximum in chain order observed when the chain lengths of the cosurfactants are the same can be explained in the following way. When the chains are of different lengths, the terminal part of the longer chain extends outward past the well-packed two-dimensional precipitate. This part can freely form the gauche-conformers. Since the maximum hydrophibicity is observed when CH3 groups rather than CH2 groups are directed toward the aqueous phase (48), one can expect that in the case of different chain lengths, the mixed monolayer will be less hydrophobic, which agrees with the surface force data (20, 21). 4. CONCLUSIONS

The FTIR spectra revealed that at low bulk amine concentrations, the surface silanol groups interact with ammonium groups through hydrogen bonds. After hemimicelle concentration, the XPS spectra showed neutral amine molecules along with the protonated ammonium ions coordinated to deprotonated silanol oxygen anions on the surface. Because of this, adsorption steeply increases. At higher concentrations, the molecular amine precipitates onto the surface, which is characterized by the typical absorption band at 3330 cm−1 . The later two regions of adsorption, where the formation of neutral amines takes place, is affected by the acetate counterions. However, the acetate counterions have

71

no influence on flotation response. The results are interpreted in terms of successive two-dimensional and three-dimensional precipitation of amine on a quartz surface after the initial phase of adsorption. The presence of alcohol enhanced the two-dimensional precipitation of amine at the same total concentration. The precipitating species is deduced to be the soluble associate of the protonated amine and alcohol. When the alkyl chain lengths are the same, the adsorbed layer is closely packed, leading to increasing hydrophobicity and thus maximum floatability. The total concentration in mixed composition is reduced to one order of magnitude when compared to the amine concentration alone for achieving the same flotation response. The neutral molecule coadsorption in between charged amine heads shielded its repulsion and thereby the adsorption is increased owing to lateral asym sym and CH2 bands in tail–tail hydrophobic bonds. The CH2 reflection–absorption spectroscopy spectra occurred at the lowest frequencies with the narrowest widths when the coadsorbents were dodecyl amine and dodecyl alcohol, validating the highest order and packing of the adsorbed species at equal alkyl chain lengths. The structure of a coadsorbed layer of long-chain amines and alcohols is elucidated in terms of deprotonation of an ammonium group in the adsorbed “alkyl ammonium–water–alcohol” complex leading to two-dimensional precipitation of molecular amine. The synergetic adsorption of these surfactants is thus explained. ACKNOWLEDGMENT The authors thank Dr. A. V. Shchukarev, Dept. of Inorganic Chemistry, Ume˚a University, Sweden, for the XPS measurements.

REFERENCES 1. Fuerstenau, D. W., and Raghavan, S., in “Proc. XII Int. Miner. Process. Congr,” Vol. II, p. 368. D.N.P.M., Sao Paulo, 1980. 2. Smith, R. W., and Scott, J. L., Miner. Process. Ext. Metall. Rev. 7, 81 (1990). 3. Laskowski, J. S., in “Challenges in Mineral Processing” (K. V. S. Sastry and M. C. Fuerstenau, Eds.), p. 15. SME, Littleton, CO, 1989. 4. Novich, B. E., and Ring, T. A., Langmuir 1, 701 (1985). 5. Cases, J. M., and Villieras, F., Langmuir 8, 1251 (1992). 6. Gaudin, A. M., and Fuerstenau, D. W., Trans. Soc. Min. Eng. AIME 202, 958 (1955). 7. Somasundaran, P., Healy, Th. W., and Fuerstenau, D. W., J. Phys. Chem. 68, 3562 (1964). 8. Predali, J. J., and Cases, J. M., in “Proc. X Int. Miner. Process. Congr,” (M. J. Jones, Ed.), p. 473. IMM, London, 1974. 9. Cases, J. M., Bull. Miner. 102, 684 (1979). 10. Scamehorn, J. F., Schechter, R. S., and Wade, W. H., J. Colloid Interface Sci. 85, 463 (1982). 11. Chen, Y. L., Chen, S., Frank, C., and Israelachvili, J., J. Colloid Interface Sci. 153, 244 (1992). 12. Scales, P. J., Grieser, F., and Healy, T. W., Langmuir 8, 227 (1992). 13. Koglin, E., Tarazona, A., Kreisig, S., and Schwuger, M. J., Colloids Surf., A 123–124, 523 (1997). 14. Bitting, D., and Harwell, J. H., Langmuir 3, 500 (1987). 15. Kubicki, J. D., Schroeter, L. M., Itih, M. J., Nguyen, B. N., and Aptiz, S. E., Geochim. Cosmochim. Acta 63, 2709 (1999).

72

VIDYADHAR ET AL.

16. Eischens, R. P., Science 146, 486 (1964). 17. Rosen, M. J., Ed., “Structure/Performance Relationships, in Surfactants.” ACS, Washington, DC, 1984. 18. Leja, J., “Surface Chemistry of Froth Flotation.” Plenum, New York, 1982. 19. Gericke, A., and Huhnerfuss, H., J. Chem. Phys. 97, 12899 (1993). 20. Yoon, R. H., and Ravishankar, S. A., J. Colloid Interface Sci. 166, 215 (1994). 21. Yoon, R. H., and Ravishankar, S. A., J. Colloid Interface Sci. 179, 391 (1996). 22. Smith, R. W., Trans. SME/AIME 226, 427 (1963). 23. Fuerstenau, D. W., and Yamada, B. J., Trans. AIME/SME 223, 50 (1962). 24. Chernyshova, I. V., and Hanumantha Rao, K., J. Phys. Chem. B 105, 810 (2001). 25. Fuerstenau, D. W., Healy, T. W., and Somasundaran, P., Trans. Soc. Min. Eng. AIME 229, 321 (1964). 26. Somasundaran, P., and Ananthapadmanabhan, K. P., in “Solution Chemistry of Surfactants” (K. L. Mittal, Ed.), Vol. 2, p. 777. Plenum, New York, 1979. 27. Laskowski, J. S., in “Advances in Flotation Technology” (B. K. Parekh and J. D. Miller, Eds.), p. 59. SME/AIME, Denver, CO, 1999. 28. Shah, D. O., and Shiao, S. Y., in “Monolayers” (E. D. Goddard, Ed.), p. 153. ACS, Washington, DC 1975. 29. Fuerstenau, D. W., Mining Eng. 208, 1365 (1957). 30. Lindman, B., and Wennerstrom, H., “Topics in Current Chemistry,” Vol. 87. Springer-Verlag, New York, 1980. 31. Bellamy, L. J., “The Infrared Spectra of Complex Molecules.” Wiley, New York, 1975. 32. Przeslawska, M., Melikowa, S. M., Lipkowski, P., and Koll, A., Vibrat. Spectrosc. 20, 69 (1999).

33. Koretsky, C. M., Sverjensky, D. A., Salisbury, J. W., and D’Aria, D. M., Geochim. Cosmochim. Acta 61, 2193 (1997). 34. Scott, J. L., and Smith, R. W., Miner. Eng. 4, 141 (1991). 35. Vansant, E. F., Van Der Voort, P., and Vranchen, K. C., “Characterisation and Chemical Modification of the Silica Surface.” Elsevier, Amsterdam, 1995. 36. Horn, A., in “Spectroscopy for Surface Science” (R. J. H. Clark and R. E. Hester, Eds.), Vol. 26, p. 273. Wiley, New York, 1998. 37. Mendelsohn, R., Brauner, A., and Gericke, A., Annu. Rev. Phys. Chem. 46, 305 (1995). 38. Park, S. Y., and Franses, E., Langmuir 11, 2187 (1995). 39. Katayama, N., Enomoto, S., Ozaki, T., and Kuramoto, N., J. Chem. Phys. 97, 6880 (1993). 40. Chernyshova, I. V., Hanumantha Rao, K., Vidyadhar, A., and Shchukarev, A., Langmuir 16, 8071 (2000). 41. Chernyshova, I. V., and Hanumantha Rao, K., Langmuir 17, 2711 (2001). 42. Kowalczyk, D., Slomkowski, S., Chehimi, M. M., and Delamar, M., Int. J. Adhesion Adhesives 16, 227 (1996). 43. Pimentel, G. C., and Mc Clellan, A. L., “The Hydrogen Bond.” Freeman, San Francisco, 1960. 44. Somasundaran, P., and Fuerstenau, D. W., J. Phys. Chem. 79, 90 (1966). 45. Schwartz, D. K., Surf. Sci. Rep. 27, 241 (1997). 46. Smith, R. W., in “Reagents in Mineral Technology” (P. Somasundaran and B. M. Moudgil, Eds.), p. 219. Dekker, New York, 1988. 47. Natal-Santiago, M. A., and Dumesic, J. A., J. Catal. 175, 252 (1998). 48. Zisman, W. A., Ind. Eng. Chem. 55, 19 (1963).