Colloids and Surfaces A: Physicochemical and Engineering Aspects, 83 (1994) 129-141 09271157/94/$07OQ 0 1994 - Elsevier Science B.V. All rights reserved.

129

Complexing agents as modifiers in mineral flotation - mechanism studies

Erhard Orthgiess, Bohuslav DobiBS* Institut fiir Physikalische und Makromolekulare Chemie, Universitdt Regensburg, D-93040 Regensburg, Germany

(Received 7 May 1993; accepted 12 October 1993)

Abstract

Several complexing agents containing both oxygen and nitrogen as donor atoms were tested with respect to their influence on the floatability of the salt-type minerals calcite, barite and fluorite. Apart from solving practical flotation problems, emphasis was placed above all on mechanistic aspects. For this purpose, flotation tests with anionic and cationic surfactants as collectors were performed, extensive adsorption experiments were carried out and the dissolution and electrokinetic behaviour of the minerals were studied. It was found that some of the complexing agents tested act as selective depressants in the flotation with anionic collectors. During flotation with cationic collectors one can often observe activating effects. In spite of their normally strong influence on floatability, no significant adsorption of chelating agents could be demonstrated. A single mechanism which explains all effects of the complexing agents does not exist. It turned out that there are several modes of action depending on which components represent the flotation system.

Key words: Adsorption; Chelating agents; Collectors; Complexing agents; Dissolution; Flotation; Ionic surfactants; Modifiers; Salt-type minerals; Zeta potential

Introduction

Mineral flotation is a method for the selective separation of mineral components out of polymin- era1 dispersions in water. The selectivity of this process is based on the different adhesion of hydro- phobized and hydrophilic mineral particles to gas bubbles ascending to the top of the flotation cell. A mineral is hydrophobized by adsorption of a suitable surface-active compound (the collector) at the surface of the mineral component to be floated. All other non-hydrophobized particles remain dis- persed in the suspension.

Whereas sulfide mineral flotation is a well- established commercial process, and also in oxide and silicate mineral flotation most problems seem

*Corresponding author.

to be resolved, the separation of salt-type minerals is very complex and in several cases the required selectivities are not easily achievable.

A possible way of improving flotation selectivity is the addition of appropriate reagents (called modifiers) to the mixture. The modifiers which represent many different classes of substances, can either improve (“activators”) or inhibit (“depres- sants”) the floatability of a particular mineral component.

In this paper, complexing agents were tested with respect to their influence on floatability. The use of this type of reagent as collector has been discussed for several years [l-3]. In the case of complexing or chelating agents as modifiers, how- ever, only a few systematic investigations can be found in the literature [1,4,5], so the present investigation was undertaken in order to increase

SSDI 0921-1151(93)02612-2

130 E. Orthgiess and B. DobiriS/Colloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-141

our knowledge in this field. Emphasis was placed sodium oleate (NaOL), and the cationic surfactant above all on mechanistic aspects rather than on cetyltrimethylammonium bromide (CTAB). Their practical flotation problems. formulae are given in Fig. 1.

Experimental CTAB (analytical grade) was obtained from

Merck (Darmstadt, Germany) and was used with- out further purification.

Minerals

Our studies were performed with the salt-type minerals calcite, fluorite and barite. These minerals are slightly soluble in water, which often causes problems concerning the selectivity of the flotation process, because lattice ions in the pulp can interact with the added reagents (e.g. a precipitation of the collector can result) and a modification and a levelling of the mineral surfaces can be found also [6].

NaDBS was synthesized in our laboratory [8] because all purchasable products of NaDBS are mixtures of isomers and homologues, which are normally not well defined and are polluted by different substances. The isomer that we produced and characterized was the sodium salt of p-( 1-pentylheptyl)benzenesulfonate (see Fig. 1).

The mineral samples were crushed after washing with distilled water and then classified by means of sieves. Particles of the fraction 0.20-0.30 mm were used for the flotation tests, and particles smaller than 0.20 mm were dry-ground in an agate mortar for 3 h in order to obtain a suitable material for the adsorption experiments. The specific surface area of the ground minerals was determined by a modified BET method with use of the Micromeritics areameter model 210 apparatus.

NaOL was made by neutralization of oleic acid (purity greater than 99%; Roth, Karlsruhe, Germany) with NaOH. The NaOL solutions, which have a shelf life of some days only, were adjusted to pH 9 in order to clarify them.

Complexing agents

Compounds containing both oxygen and nitro- gen as donor atoms served as complexing agents for our studies. These substances were chosen on

W W -(W&s - 7” - (CY),- CH,

Some important characteristics of the minerals used are listed in Table 1.

I H,C - N*- CH, CTAB

I Br- CH,(CH,),,CH,

NaDBS

Collectors cl=;=0

I

H H NaOL 0-k+

As collectors we used the anionic surfactants sodium dodecylbenzenesulfonate (NaDBS) and

H,C-(G-!&-i=&(CH2),-COONa+

Fig. 1. Formulae of the surfactants used as collectors.

Table 1 Properties of the mineral samples used

Mineral Formula Origin Purity Specific surface area’ cm2 g-‘1

IEP

(PI-U

Solubility product [ 71

Calcite CaCO, Fluorite CaF, Barite BaSO,

‘Of the ground material.

Iceland VENTRON Pfalz, Germany

>99% approx. 100% approx. 99%

4.2 10.5 4.7 x 1O-9 M2 2.1 10.2 1.7 x lo-“’ M3 2.2 7.0 1.5 x 1O-9 M2

E. Orthgiess and B. DobiciSjColloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-141 131

the one hand because of their great binding vari- ability with metal ions both in solution and at the mineral/water interface, and on the other hand because there have not been many investigations up to now dealing with these reagents as modifiers during mineral flotation.

The formulae of the complexing agents used are shown in Fig. 2. NTA, EDTA and CDTA belong to the well-known group of complexones and were obtained, like SO (salicylaldoxime) and 8-HQ (8-hydroxyquinoline), from Merck in analytical grade quality. Acetohydroxamic acid (AHA) was obtained from Janssen (Geel, Belgium). p- Hydroxybenzaldehydoxime (HBO) was synthe- sized and characterized [8] in our laboratory.

With the exception of the latter substance, all complexing agents mentioned are capable of form-

0 N,g eo-;_CH,

f ,a+-C-oe +a

Na . co-$-w >N-W-W-N,

ciq-00 540

0 EDTA 0

i?

0

FI No. eo-c-c&\

NrP co-pi/

N N’

a+-c-cf %a

‘a+,-c-oe %a

0 CDTA Id

Ho HsC-C,~,H

AHA ‘OH P

:I -l N

OH 8-HO

q-W H&Y

HOsHO SO

Fig. 2. Formulae of the complexing agents tested as modifiers.

ing chelates and were used without further purification.

Electrophoretic measurements

To obtain the 5 potential of the ground mineral samples, 0.05% mineral dispersions were prepared and after pH adjustment the dispersions were conditioned for 60 min. The electrophoretic mobili- ties were then measured using the Mark II appara- tus (Rank Brothers, Bottisham, UK) at a constant ionic strength of 10e3 M NaCl in a flat quartz cell. The mobilities were converted into 5 potentials with use of the Smoluchowski equation [9].

Adsorption experiments

The amount of surfactant or complexing agent adsorption was determined from the difference between the initial and equilibrium concentration of these reagents in the solution. The experiments were performed by shaking the mineral dispersions in 100 ml Erlenmeyer flasks with a solid/liquid ratio of 1:25 (by mass). After equilibration, the suspensions were filtered (0.2 l.trn filter units) to separate the mineral. The clear filtrates were then analysed for surfactants, complexing agents and lattice ions.

In the presence of CTAB, however, the filtration process was substituted by centrifugation (10 min; 5000 rev min-‘) to avoid an uncontrollable adsorption of the cationic surfactant at the nega- tively charged surface of the filter material.

Preliminary studies of the adsorption kinetics showed that the adsorption of surfactants was a fast process, reaching a quasi-equilibrium within 5 min [8]. This finding allows the conditioning time to be restricted to 20 min and moreover ensures the correlation between the results of the adsorption experiments and the flotation tests (maximal time 12 min).

The equilibrium concentration of the surfactants was determined by different methods, which are compiled in Table 2.

132

Table 2

E. Orthgiess and B. DobiciSfColloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-141

Methods used for the determination of the surfactant concentration

Surfactant Method Range (mol I-‘)

Reference

NaDBS

NaOL

CTAB

Two-phase titration Rhodamine method Two-phase titration Gregory method Two-phase titration

(1 x lo-*) - (1 x 10-q 10 (1 x 10-S) - (1 x lo-‘) 11 (1 x 10-q - (5 x lo-s) 12 (5 x lo-s) - (1 x 10-e) 13, 14 (1 x lo-*) - (1 x 10-s) 10

The analytical procedures used for the complex- ing agents are listed in Table 3.

Dissolution measurements

The dissolution tests were carried out at the same time as the adsorption experiments by analys- ing the filtrates for lattice ions. In order to be able to distinguish between different species of metal ions (free cations, associated ions, complexed ions) several different analytical methods were used: the total metal concentration was measured by atomic absorption spectrometry; to determine all the cat- ionic species of a metal except the chelated ones compleximetric titration was used; the concen- tration of the free cations (actually their activity) was analysed using a metal-ion-selective electrode. In the study of anionic lattice ions, only the fluorite system was examined as an example. For this purpose an F-ion-selective electrode was used. Further details of the analytical methods men- tioned are given in Ref. 8.

Flotation tests

The flotation tests were performed in a Hallimond tube modified by Dobiig [ 19,201. A portion (0.5 g) of the mineral sample was intro- duced into the tube and was conditioned for 2 min in a solution containing all the desired reagents. Flotation was then started by supplying nitrogen and, unless stated elsewhere was halted after 10 min.

Generally in our investigations, distilled water was used for all experiments, the pH was adjusted with HCl or NaOH and the ionic strength was controlled by adding 10m3 mol 1-l NaCl.

Results and discussion

Flotation

First we performed extensive flotation tests to obtain an idea of the influence of the complexing agents selected on floatability. The results are

Table 3 Methods used for the determination of the complexing agent concentration

Complexing agent Method Range (mol 1-r)

Reference

NTA, EDTA, CDTA

HBO, SO AHA S-HQ

Reverse complexometric titration Cr complex method Spectrophotometry Colorimetry Colorimetry

(1 x 10-Z) - (1 x 10-q 8 (8 x 10-3) - (5 x 10-4) 8, 15, 16 (1 x 10-d) - (5 x 10-6) (2 x 10-4) - (1 x lo-s) 1; (3 x 10-d) - (5 x 10-6) 18

E. Orthgiess and B. DobiciS/Colloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-141 133

summarized in Table 4. For all these single-mineral experiments a collector concentration was chosen which would give a recovery of about 70% in the absence of modifiers. In this way, after the addition

Table 4 Results of the flotation tests (a) Effect of complexing agents on flotation with NaDBS

Complexing agent Calcite” Bariteb Fluorite”

NTA - _ --

EDTA - -- --

CDTA 0 _ ++ so * * 0 HBO 0 0 _

AHA - + -

8-HQ 0 + ++

(b) Effect of complexing agents on flotation with NaOL

Complexing agent Calcite* Barite’ Fluorite’

NTA 0 -- -

EDTA 0 -- -

CDTA 0 ++ SO + + 0 HBO + 0 0 AHA + 0 0 8-HQ + + ++

(c) Effect of complexing agents on flotation with CTAB

Complexing agent Calcites Fluoriteh

NTA + ++ EDTA + ++ CDTA + ++ SO + 0 HBO 0 -

AHA 0 0 8-HQ 0 ++

Key: 0, almost no effect; *, not determined; +, minor activation;

-, minor depression; + +, strong activation; - -, strong depression. “With 2 x lo-“ mol 1-l NaDBS; at pHz9. bWith 5 x 10e5 mol 1-l NaDBS; at pH ~8. “With 2 x 10m6 mol 1-l NaDBS; at pH z 8. *With 5 x 1O-4 mol 1-l NaOL; at pHz 10. e With 4 x 10m6 mol I-’ NaOL; at pH x 10. ‘With 8 x lo-’ mol 1-l NaOL; at pHz 10. BWith 1 x 10m4 mol 1-l CTAB; at pHz9. “With 1 x 10m4 mol 1-l CTAB; at pHx8.

of complexing agents, both activating and depres-

sant effects could be observed. When the flotation data are examined, the

following results are significant. (i) The effect of the complexing agents on float-

ability is not uniform: besides depressant effects, activations could also be observed.

(ii) The chelating agents do not affect calcite flotation on a larger scale.

(iii) Strong effects of the complexones are observed mainly on barite and on fluorite.

(iv) Some complexing agents show, in spite of a similar molecular structure, quite different effects (EDTA-CDTA on fluorite).

(v) Floatability is often affected by the modifiers in a similar way for both the anionic collectors used.

(vi) During floatation with cationic surfactants one can often observe activation (but never depression).

(vii) CDTA and 8-HQ are effective stimulants of the fluorite flotation with all types of collector.

Hydrophilization of the mineral surface by

adsorption of complexing agents

Experiments carried out on fluorite with the complexing agents HBO and SO in combination with the anionic collector NaDBS showed that the two molecules, which differ from each other only in the position of the hydroxy group on the benzene ring, have a significantly different effect on the floatability of the mineral. While SO is almost ineffective, HBO acts as a depressant in the concen- tration range 10-5-10-3 mol 1-l (cf. Table 4(a)).

In agreement with these observations it was found that HBO, but not SO, adsorbs on fluorite (Fig. 3). Owing to this adsorption the mineral surface becomes hydrophilized and this is the reason for the depressant effect observed.

One can assume that the HBO molecule interacts with the mineral surface through its oxime group, while the OH group, as a second hydrophi- lic group, is directed to the bulk phase. This can mean that in this case a mechanism operates which

134 E. Orthgiess and B. DobiciSJColloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-141

,o.s 2 3 ‘ 1o_* 2 3 4

= nao/so (mol,,l 10“ * 3 4 10-z

Fig. 3. Adsorption isotherms of HBO and SO on fluorite at 298 K (r is the adsorption density).

has been postulated as the standard mechanism for complexing agents acting as depressants [21] and which is schematically described in Fig. 4.

In contrast to HBO, SO is capable of chelating Ca2+ ions in the bulk phase, but not those of the mineral surface (probably because of steric hin- drance) and that is why adsorption does not take place. In this context it is interesting to note that in further experiments also, with other types of chelating agent, only a very small adsorption ten- dency was found. The weaker complexing agent HBO, however, forms spatially simpler complexes (no chelates) and so the complexation of cations in the mineral/water interface is possible. This

with mmplexing agent

,“zYv%:H mmpleting &lent

Fig. 4. The mechanism of the action of complexing agents as modifiers, shown schematically. Case 1: adsorption of complex- ing agents competing with the surfactant adsorption. (For a detailed explanation see text.)

leads to the adsorption (surface complexation) observed in our experiments with the fluorite-HBO system.

On fluorite flotation with NaOL, almost no depressant effect of HBO was evident (cf. Table 4(b)), in contrast to the experiments with NaDBS. This can be linked to the special adsorp- tion mechanism of oleate on salt-type minerals [22]. In contrast to NaDBS, here the adsorption takes place more in the way of a surface precipita- tion that can hinder the adsorption of HBO or at least make it ineffective as far as the floatability of the fluorite is concerned.

Using CTAB as a collector, a certain depressant effect of HBO on the fluorite recovery is found, but it should be mentioned that there is only a poor floatability of fluorite with CTAB in the absence of depressants.

Modijication of the collector adsorption by

complexing agents

Collector NaDBS In the flotation tests shown in Fig. 5, EDTA

behaves as an effective depressant in barite flota- tion, whereas AHA exhibits almost no effect. The results of the corresponding adsorption experiment are given in Fig. 6. The observation of no signifi- cant changes in NaDBS adsorption density in the

Fig. 5. Flotation recovery R of barite as a function of the concentration of EDTA or AHA (collector, NaDBS).

E. Orthgiess and B. DobiriS/Colloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-l 41 135

l.Or i

Fig. 6. Adsorption density r of NaDBS on barite as a function of the EDTA or AHA concentration.

presence of AHA, but a distinct decrease in the collector amount adsorbed with increasing EDTA concentration, correlates well with the flotation tests.

In further experiments we examined the adsorp- tion of the complexing agents themselves at the barite surface. It turned out that there is no signifi- cant adsorption either of AHA or of EDTA. To explain the phenomenon of the reduced adsorption density of NaDBS in the presence of EDTA (which is not adsorbed itself), electrophoretic measure- ments were carried out. They showed (see Fig. 7) that the c potential of barite becomes more nega- tive in the presence of EDTA, whereas the effect of AHA is small. This could be due to the removal

RHA

Fig. 7. The c potential of barite, fluorite and calcite as a Fig. 8. The total barium concentration in a barite suspension function of the EDTA or AHA concentration. as a function of the EDTA of the AHA concentration.

of Ba’+ ion from the mineral surface and from immediately adjacent layers by the strong chelating agent EDTA (AHA is not capable of this Ba2+ ion removal).

The described process should be accompanied by an increase in the total Ba2+ ion concentration in the bulk phase. This could be confirmed experi- mentally: Fig. 9 demonstrates distinctly different effects of EDTA and AHA on the composition of the solution, which is, according to further experi- ments, not greatly altered by an increasing NaDBS concentration.

Owing to the more negative c potential in the presence of EDTA the adsorption of the anionic surfactant is made more difficult (energetically less favourable) by electrostatic repulsion. Moreover, binding of the surfactant ions promoted by Ba2+ ions in the compact part of the electrical double layer cannot occur.

To confirm the above conclusions up to now, a “reverse” experiment was performed. Instead of the addition of a complexing agent (reduction of free cation concentration) the addition of Ba2+ ions was examined in terms of its effect on the [ potential, collector adsorption and floatability. As expected, it turned out that the barite surface was rendered more positive, the surfactant adsorption increased and the floatability improved.

The mechanism of the action of the chelating

Barite 1_1 EDTA

136 E. Orthgiess and B. DobiciS/Colloids Sufaces A: Physicochem. Eng. Aspects 83 (1994) 129-141

agents used here is presented schematically in barite-CDTA, fluorite-NTA and fluorite-EDTA, Fig. 9 and can be summarized as follows. The a reduced adsorption density of NaOL in the complexing agent removes cations from the presence of the chelating agents suggests a similar mineral/water interface. This renders the 5 poten- mechanism of the depressant action of these rea- tial more negative, the adsorption of the anionic gents as for the flotation with NaDBS. To explain collector is reduced, and a depressant effect on the the effects in the case of calcite, for which the floatability results. This mechanism occurs especi- NaOL adsorption is generally not influenced by ally with chelating agents that have a strong affinity the addition of complexing agents, it is necessary to free (i.e. hydrated) cations, but are not well to look again at the adsorption mechanism for fitted for complexation of cationic surface sites NaOL in systems containing sparingly soluble because of steric hindrance. minerals.

In the systems with NaDBS as a collector, the more or less depressant effect of chelating agents might be caused by the described process in the following cases (see Table 4(a)): barite-NTA, barite-EDTA, barite-CDTA, fluorite-NTA, fluorite-EDTA, calcite-NTA, calcite-EDTA.

Oleate adsorption in the presence of polyvalent lattice cations may be understood, at least at higher concentrations, as a precipitation of metal-oleate associates of low solubility to the mineral surface. This surface precipitation, which occurs in the compact part of the electrical double layer, starts at lower oleate concentrations than those for the precipitation of, for instance, CaOL, or BaOL, in the bulk phase [23].

Collector NaOL

One should expect that in mineral flotation with the collector NaOL, which is like NaDBS an anionic surfactant, similar effects of the complexing agents tested occur. Comparing Tables 4(a) and 4(b), it is evident that this is true for barite and fluorite. The floatability of calcite, however, is affected in a quite different manner. The com- plexones show no significant effect and the other modifiers cause a light activation; depressant effects are not observed at all.

In the systems barite-NTA, barite-EDTA,

without wmplexing agent with mmplexing agent

-+Wvv ;“m:, +vqH

dmk awiadant c=wm went 68 compb3x

Fig. 9. The mechanism of the action of complexing agents as modifiers, shown schematically. Case 2: depression of the NaDBS adsorption by chelating agents. (For a detailed explana- tion see text.)

If the composition of the solution of a system such as calcite-EDTA or barite-EDTA is consid- ered (Fig. lo), it is striking that free Ca’+ ions are available in the case of calcite also in the presence of the chelating agent. In this way the surface precipitation is not depressed by the addition of EDTA and the oleate adsorption is not affected.

= EOTR ~mol/ll

Fig. 10. The composition of the solution of the systems calcite-EDTA and barite-EDTA as a function of the EDTA concentration (T=298 K, pH 9.0 + 0.5). Note that in the pres- ence of calcite the concentration of free EDTA is always zero.

E. Orthgiess and B. DobiciSjColloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-141 137

In contrast to this finding, in barite (and fluorite) systems the concentration of free cations soon decreases to zero with increasing amount of chelat- ing agent. As a result, the formation of BaOL, (or CaOL,) is hindered and oleate adsorption is reduced.

The reason for this different availability of free cations could be the faster dissolution kinetics of calcite compared to those of barite and fluorite [24], so a rapid substitution of cations that have been consumed to form a complex is possible only in the case of calcite.

A summary for the mechanism for the effect chelating agents on flotation with oleate can be given as follows (Fig. 11): The complexing agent binds free cations in the mineral/solution interface, the formation of metal-oleate associates is inhib- ited if a fast substitution of the complexed cations from the mineral surface does not occur, the oleate adsorption due to surface precipitation is dimin- ished and so hydrophobicity and floatability are reduced.

Collector CTAB The more negative 5 potential in the presence

of the chelating agents NTA, EDTA and CDTA (see Fig. 7) causes (with the exception of the calcite-NaOL system) a reduction of the adsorp- tion density of the anionic collectors and a decreased floatability. Conversely, using cationic

wiihola amlpleldng apents I

with complaxing agents

Fig. 11. The mechanism of the action of complexing agents as modifiers, shown schematically. Case 3: depression of the NaOL adsorption by chelating agents. (For a detailed explanation see text.)

surfactants a promoted collector adsorption and

an activated flotation should be expected. This could be confirmed in the adsorption (Fig. 12) and flotation experiments (see Table 4(c)), and accord- ing to these findings one can explain the action of the modifiers tested in flotation with cationic sur- factants by a mechanism in reverse of that found for systems with NaDBS as a collector.

However, it should turn out that a second mech- anism exists for the improved floatability of calcite, and especially fluorite, with CTAB in the presence of chelating agents, as discussed in the next section.

Surfactant-complexing agent interaction

If cationic surfactants are present in the systems examined along with anionic complexing agents, a surfactant-complexing agent interaction is likely for electrostatic reasons. The first indication of such interactions was given by surface tension measurements (Fig. 13). The chelating agent (EDTA) is not surface active at pH 9, but in the presence of a constant amount of CTAB, a further reduction in the surface tension is found with increasing EDTA concentration. This can be explained by the formation of a possibly uncharged associate consisting of CTAB (cationic) and EDTA (anionic) which is more interface active than CTAB alone. A different explanation for this behaviour could be a simple counterion effect of EDTA with respect to CTAB molecules adsorbed at the surface,

’ EDTR l.Ol/ll

Fig. 12. Adsorption density of CTAB on calcite or fluorite as a function of the EDTA concentration.

138 E. Orthgiess and B. DobiciSIColloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-141

60

65 -

7 2 so- c ss- *

CTAB (0.1 mmol/l)

b

50 -

45 -

+ 40 -

EDTA

35 -

Fig. 13. Surface tension a as a function of the EDTA or CDTA concentration in the absence of surfactant (upper curves) and in the presence of 10e4 mol 1-l CTAB (lower curves).

but there are further findings indicating strong CTAB-EDTA interactions.

If relatively concentrated solutions of CTAB and EDTA (both 1 x lOA mol 1-r or more) are mixed together, first a turbidity occurs and then, after some minutes, a colourless precipitate forms. This precipitate was examined analytically after filtra- tion and drying. The IR spectrum shows only a superposition of the CTAB and EDTA spectra without any new bands. The elemental analysis (Table 5) correlates best with a mixture of EDTA-CDTA associates in ratios of 2:l and 3:l. A surplus of hydrogen can be explained by traces of water while the nitrogen loss is not immediately detectable.

Surfactant-chelating agent complexes, which could be observed with all complexones investi- gated and CTAB, should as strong interface-active compounds act as collectors, too. In this way a second explanation exists as to why chelating agents activate the mineral floatability using cat- ionic collectors. The additional collector adsorp-

Table 5 Elemental analysis (%) of the CTAB-EDTA precipitate

tion found in these cases could be due to the adsorption of the CTAB-complexing agent associ- ate at the mineral surface and not of CTAB alone.

This proposed mechanism is also compatible with the results of electrophoretic measurements which were performed with an initially negatively charged mineral in order to obtain clearer results concerning the effect of cationic surfactants like CTAB. According to Fig. 14 an increasing amount of CTAB causes, in the absence of EDTA, a more and more positive c potential, in accordance with an increasing surfactant adsorption. In the presence of EDTA no distinct effect of the chelating agent (rendering the potential more negative) occurs until the CTAB concentration exceeds about 2 x 10e5 mol 1-l and this happens simultaneously with a markedly increased CTAB adsorption. An explanation of these phenomena may be a preferen- tial adsorption of the uncharged (or even anionic) CTAB-EDTA complex over that of the cationic CTAB.

The second mechanism of the reaction of chelat-

50

Hornblende f

EDTA 0 40

S 30

E 20

EDTAlmmoUl

cCTIB lmolll)

Fig. 14. The c potential of hornblende as a function of the CTAB concentration.

Element Found Na,H(EDTA)(CTA) NaH(EDTA)(CTA), H(EDTA)(CTA),

C 61.5 + 0.2 56.20 65.42 10.42 H 12.5 f 0.2 8.95 11.20 12.26 N 5.6 k 0.1 6.78 6.35 6.13

E. Orthgiess and E. DobiciSjColloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-141 139

ing agents in flotation with cationic surfactants can be summarized as follows. The cationic surfac- tant and the anionic chelating agent form a com- plex, which is adsorbed as a strongly surface-active compound at the mineral surface thereby increas- ing hydrophobicity and activating floatability of the mineral component. The question as to which of the two mechanisms proposed for the action of modifiers during flotation with CTAB is predomi- nant in which system and in the end responsible for the effects observed cannot be answered on the basis of the investigations made hitherto. It is conceivable, too, that both mechanisms are effec- tive at the same time.

Interactions with complexing agents can occur also in the case of anionic collectors. This was demonstrated both with NaDBS and NaOL and the chelating agents NTA, EDTA and CDTA by surface tension measurements. The other complex- ing agents used (AHA, GHC, SO, HBO), however, do not interact with anionic surfactants according to our investigations.

Since it is not visible that these interactions play a role in mineral flotation with anionic collectors, further experiments were not performed in this direction. Incidentally, a turbidity increase or even a precipitation as observed with CTAB does not occur using anionic surfactants.

Complexing agents acting as collectors

In Fig. 15 the c potential of fluorite particles is plotted versus the EDTA and CDTA concen- tration. With increasing concentration of the che- lating agents the mineral surface becomes more and more negative. The effect of EDTA is some- what stronger than that of CDTA, but the same tendency can be seen in both cases.

It is to be expected that the adsorption density of anionic collectors decreases under these circum- stances. Figure 16 shows for NaDBS that this supposition is confirmed by experiment. The find- ings up to now make us presume that both EDTA and CDTA act as depressants in fluorite flotation, the effect of EDTA being perhaps a little stronger.

Fluorite

Fig. 15. The [ potential of fluorite as a function of the EDTA or CDTA concentration.

1.0

0.9 -

0.6 -

0.7 -

2 0.6-

: 05- . 2

c 0.4-

0.3 -

0.2 -

0.1 -

Fig. 16. Adsorption density of NaDBS on fluorite as a function of the EDTA or CDTA concentration.

Flotation tests revealed, however, a strong depres- sant activity only in the case of EDTA, whereas the floatability of fluorite was even activated by CDTA (Fig. 17).

Obviously there is no explanation for these results just on the basis of the collector adsorption density or of the surface charge of the mineral particles. Further investigations showed that nei- ther the composition of the flotation pulp nor the state of the liquid/gas interface could be responsible for these flotation results, because there are no significant differences to be seen in the behaviour of EDTA and CDTA. Also, kinetic studies concern- ing the surfactant adsorption and the mineral

140 E. Orthgiess and B. DobiriS/Colloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-141

100, I 90

80 t

70 -

60 -

s so- az

LO -

30 -

20 t

,---- /

---__a

------=--w-- f -----I’

A’ cDTA

\

i :

NaDBS 2pmoUI

: * 10mln + 2 min

Fig. 17. Flotation recovery of fluorite as a function of the EDTA or CDTA concentration (collector, NaDBS).

dissolution did not provide any indications as to the explanation of the phenomena observed.

However, when we tested the collector effects of the complexing agents themselves, it turned out that fluorite shows a good floatability with CDTA (even in the absence of any surfactant), where as EDTA does not operate as a collector at all. According to Fig. 18, it is evident that in a more alkaline medium with a higher degree of dissoci- ation, CDTA is more effective, but the undissoci- ated form (H,CDTA) also seems to work as a reasonable collector.

In further experiments we found that the collect-

Fluorite

/ I I I_ 0 12 3 4 5 6 7 8 3 10

time (mln)

Fig. 18. Floatability of fluorite with CDTA ( 10m4 mol 1-l) at Fig. 19. Adsorption tests on the system fluorite-CDTA (pH 9; different pH values. initial concentration of CDTA, 10m4 mol 1-l).

ing effect of CDTA is very selectively limited to fluorite. This selectivity remains preserved in polymineral flotation, too, although the collecting effect with respect to fluorite was distinctly reduced in the presence of minerals with fast dissolution kinetics (like calcite), which shows that only the free CDTA species are effective collectors.

A possible mechanism of the collector action of CDTA is certainly the selective adsorption of this chelating agent on the fluorite surface. The hydro- philic parts of the molecule interact with the min- eral surface, whereas the hydrophobic cyclohexane ring is directed towards the solution. In this way the already relatively hydrophobic fluorite surface may become sufficiently modified to ensure the floatability.

However, experimental verification of the pro- posed mechanism by various adsorption tests did not prove significant adsorption of CDTA on fluorite (see, for example, Fig. 19). According to these measurements an adsorption density in the range up to 0.5 umol m-’ is probable, however. This would correspond to a maximal degree of monolayer coverage of 30% (calculated with an assumed area of 1 nm2 for each CDTA molecule at the fluorite surface), which might be sufficient to provoke the phenomena observed.

Like CDTA, 8-HQ acts as an activator for fluorite flotation with all three surfactants used as collectors without affecting the adsorption density

E. Orthgiess and B. DobiciSIColloids Surfaces A: Physicochem. Eng. Aspects 83 (1994) 129-l 41 141

of the surfactants. Flotation tests in the absence of

the surfactants show that also in this case the observed activation is related to the selective collec- tor effect of 8-HQ for fluorite.

In this connection it should be stressed that the molecular structure of 8-HQ is similar to that of CDTA: both molecules contain hydrophobic ring structures as well as hydrophilic groups.

Conclusions

In this work, complexing agents containing both oxygen and nitrogen as donor atoms were tested for their influence on the floatability of salt-type minerals.

Some of the reagents used act as selective depres- sants in the flotation with anionic collectors. During flotation with cationic collectors one can often observe an activating effect. Furthermore, it turned out that structurally closely related com- plexing agents such as EDTA and CDTA can affect the floatability in a very different manner.

In order to explain these phenomena, extensive mechanistic studies were performed. We found that complexing agents in the flotation pulp generally affect (i) the amount of collector adsorbed at the mineral surface, (ii) the dissolution of the mineral components and therefore the composition of the solution and (iii) the structure of the mineral/water interface and the electrical double layer.

In spite of their normally strong influence on floatability, no significant adsorption of chelating agents at the mineral surface could be demon- strated during this work, probably due to steric hindrance. The conditions for chelate formation in the bulk phase and at the mineral/water interface seem to be very different and therefore it is very difficult to apply knowledge of complexing agents gained in analytical chemistry to flotation systems.

Finally it should be emphasized that a single mechanism for complexing agents in mineral flota- tion does not exist. Rather, we could propose several different modes of action during our investigations.

Acknowledgements

The authors wish to thank Claudia Mehler and Barbara Heller for experimental help. The financial support provided by the German Bundesministerium fiir Forschung und Technologie (BMFT) is gratefully acknowledged.

References

1

12 13 14

15 16

17

18 19 20

21

22

23

24

D.R. Nagaraj, in P. Somasundaran and B.M. Moudgil (Eds.), Reagents in Mineral Technology, Surfactant Sci. Ser. 27, Marcel Dekker, New York, 1987, p. 257. Pradip, Miner. Metall. Process., (1988) 80. X. Wang and E. Forssberg, J. Colloid Interface Sci., 140 (1990) 117. H. Baldauf, Freiberg. Forschungsh. A, 619 (1980) 5. T.L. Wei and R.W. Smith, Int. J. Miner. Process., 21 (1987) 93. H.S. Hanna, P. Somasundaran, in M.C. Fuerstenau (Ed.), Flotation - A.M. Gaudin Memorial Volume, American Institute of Mining Engineers, New York, 1976, p. 197. A.F. Hollemann, E. Wiberg, Lehrbuch der anorganischen Chemie, de Gruyter, Berlin, 1985, p. 213. E. Orthgiess, Thesis, Universitlt Regensburg, Germany, 1991. PC. Hiemenz, Principles of Colloid and Surface Chemistry, Marcel Dekker, New York, 1986, Chapter 13.4. V.W. Reid, G.F. Longman and E. Heinerth, Tenside Surf. Deterg., 4 (1967) 292; 5 (1968) 90. B. DobiBS, E. Orthgiess and D. Schieder, Repr. 8th Int. Symp. on Surfactants in Solution, Gainesville, FL, 1990, p. 191. R. Diederich, Seifen Ole Fette, Wachse, 97 (1971) 914. G.R.S.C. Gregory, Analyst, 91 (1966) 251. E.W. Fowler and T.W. Steele, Report of the National Institute for Metallurgy, Vol. 443, Johannesburg, 1968, p.3. P.J. Cherney, Anal. Chem., 26 (1954) 1806. J. Rubio and E. Matijevic, J. Colloid Interface Sci., 68 (1979) 408. B. Kakac and Z.J. Vejdelek, Handbuch der photomet- rischen Analyse organischer Verbindungen, Vol. 1, Verlag Chemie, Weinheim, 1974, Chapter 8.1.2. Ref. 17, Chapter 2.18.7. B. DobilS, Colloid Polym. Sci., 259 (1981) 775. B. Dobi&$ Trans. Inst. Min. Metall. (Sect. C: Miner. Process. Extr. Metal].), 92 (1983) 164. H. Schubert, Aufbereitung fester mineralischer Rohstoffe, Vol. 2, Deutscher Verlag fiir Grundstoffindustrie, Leipzig, 1978, Chapter 4. R. Sivamohan, P. de Donato and J.M. Cases, Langmuir, 6 (1990) 637. K.P. Ananthapadmanabhan and P. Somasundaran, Colloids Surfaces, 13 (1985) 151. C. Litsakes and P. Ney, Fortschr. Mineral, 63 (1985) 117.