Minerals Engineering 45 (2013) 170–179

Contents lists available at SciVerse ScienceDirect

Minerals Engineering

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

From amine molecules adsorption to amine precipitate transport by bubbles:A potash ore flotation mechanism q

Janusz S. Laskowski ⇑Norman B. Keevil Institute of Mining Engineering, University of British Columbia, Vancouver, BC, Canada

a r t i c l e i n f o a b s t r a c t

Article history:Received 3 November 2012Accepted 11 February 2013

The paper is dedicated to the late ProfessorJan Leja

Keywords:SylvitePotash oreFlotationPotash ore flotationAminesPrecipitation

0892-6875/$ - see front matter � 2013 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.mineng.2013.02.010

q Based on 2010 Gaudin lecture which was presenteDenver, February 28, 2011.⇑ Fax: +1 604 822 5599.

E-mail address: jsl@apsc.ubc.ca

Recent investigations summarized in this review have been conveniently grouped into (i) those dealingwith the mechanism of action of the reagents applied in the flotation of potash ores, and (ii) those focusedon the flotation properties of salt-type minerals and explanation of the remarkable selectivity betweenfloatable sylvite, and non-floatable halite. This paper is confined to the first group. It is argued that in dis-cussing the mode of action of long-chain primary amines in the flotation of potash ores account must betaken of the way in which these amines are applied by industry. Because they are water insoluble theyare melted by heating up to 70–90 �C and then they are dispersed in acidified aqueous solution. Onceadded to the flotation pulp, the hot amine dispersion rapidly cools down to a temperature far belowthe Krafft point. The rapid conversion from a hot emulsion to a cold precipitate is a very severe transfor-mation. Since nothing is known about the kinetics of these changes and phase instability only the labtests in which the adopted reagent preparation procedures closely follow the industrial practice havebeen considered in this review.

� 2013 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1702. Early research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1713. Krafft point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1724. Electrical charge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1735. Use of amines in commercial potash ore flotation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1736. Amine precipitate in sylvite flotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1747. Molecular films. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1758. The mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1769. Frothers in potash flotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17610. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178

1. Introduction

Applying flotation to treat ores containing water-soluble salts –as pointed out by Gaudin in his monograph (Gaudin, 1957) – was

ll rights reserved.

d at the SME Annual Meeting,

not considered until after Jeanprost (1928) showed that the flota-tion must be conducted in a saturated solution of such minerals.The minerals sylvite (KCl) and halite (NaCl), two major compo-nents of the most important potash ore – the sylvinite ore – canbe separated by flotation that is carried out in a NaCl–KCl saturatedbrine. At 20 �C, 1.450 kg of the KCl–NaCl saturated solution con-tains about 0.300 kg of NaCl, 0.150 kg of KCl and 1 kg of water(Gaska et al., 1965). Thus the NaCl–KCl saturated brine is approxi-mately 6 mol/L solution of these two salts. The NaCl concentration

Fig. 1. Schematic representation of the mechanism of collection of sylvite and lackof collection of halite by dodecylammonium ions (D.W. Fuerstenau and M.C.Fuerstenau, 1956).

Table 1Some important contributions in the area of potash ore flotation.

J.S. Laskowski / Minerals Engineering 45 (2013) 170–179 171

in seawater is about 0.6 M, there is thus a huge difference in theelectrolyte concentrations between potash ore flotation pulpsand pulps in other flotation operations. Only now is it becomingapparent that dissimilarity between the potash ore flotation andother flotation systems results mostly from differences in the ionicstrength.

This paper reviews recent advances made in understanding thenature of phenomena taking place in the flotation carried out inNaCl + KCl saturated brine, that is in the potash ore flotation pro-cess. The purpose of this paper is to draw a common threadthrough seemingly disparate pieces of evidence, and to reconcilevarious experimental results and various theories put forward inthe area of potash ore flotation.

For the sake of discussion, the papers dealing with various as-pects of potash ore flotation have been grouped into:

(i) Those which discuss the mechanism of action of long-chainflotation collectors in saturated brine that leads to a goodflotation of KCl in such an environment.

(ii) Those studying the differences between flotation propertiesof various water soluble salts which make some of themfloatable while the others are not.

The latter group will not be considered in this review, the read-er is referred to several excellent publications on this topic. Rogersand Schulman (1957) and Rogers (1957) were the first to considerhydration as the phenomenon responsible for the surface proper-ties of alkali halides. They pointed out that ions like Na+, K+, andCl�, etc., are strongly hydrated and that the properties of the sur-faces of the minerals containing such ions in water are to a largeextent determined by these ions’ hydration. This model was fur-ther developed by Miller et al. (Hancer et al., 2001; Cao et al.,2010; Cao et al., 2011; Ozdemir et al., 2011), who were able toshow that hydration phenomena at salt crystal surfaces provide agood explanation for the flotation properties of floatable sylviteand non-floatable halite (advancing contact angle on KCl was mea-sured to be 7.9 ± 0.5� while such an angle measured on NaCl was0�). The differences in the hydration between KCl and NaCl surfacesis believed to affect the adsorption of flotation collectors on thesesalts.

Year Contributors

1937 J.E. Kirby1956 D.W. Fuerstenau, M.C. Fuerstenau1957 J. Rogers, J.H. Schulman1951–1988 R. Bachman, A. Singewald, H. Schubert1968 R.J. Roman, M.C. Fuerstenau, D.C. Seidel1977–1982 J. Leja, V.A. Arsentiev1985 D.A. Cormode1988 V.A. Arsentiev, T.V. Dendyuk, S.I. Gorlovski1982 – present S.N. Titkov1990 – present J.D. Miller and co-workers1986 – present J.S. Laskowski and co-workers

2. Early research

It was not until the 1950s that the first detailed papers on theadsorption of aliphatic amines on halides were published by Fuer-stenau and Fuerstenau (1956) – papers that represent the begin-ning of scientific research on the flotation fundamentals ofpotash ores. In this flotation process, two isomorphous minerals– sylvite and halite – are separated by flotation. The best separa-tion between these minerals, which differ only in the cation, areobtained with cationic collectors.

The only interpretation that makes sense (Gaudin, 1957) is thatthe ammonium ion fits in the place of potassium at the sylvite sur-face but does not fit in the place of sodium at the halite surface.Fig. 1 shows schematically the mechanism proposed by Fuerstenauand Fuerstenau (1956).

Some of the benchmarks in the development of the potash oreflotation process are listed in Table 1. While the author has triedto tabulate all important developments in this area, only those thatare consistent with the view presented in this paper will be dis-cussed further. The list begins with Kirby’s patent (US Patent2,088,325), which introduced straight-chain primary amines intothe technology of potash ore flotation as a universal collector.

Many ideas – especially in the early period of the process’ devel-opment were more or less directly transplanted from other flota-

tion processes. However, a comparison of some fundamentalcorrelations found in froth flotation, for example the relationshipbetween collector concentration and recovery, for the common flo-tation systems and the potash ore flotation system is stunning. It isknown that the recovery curve, plotted versus collector concentra-tion, drops to zero when collector concentration approaches thecritical micelle concentration. This is shown in Fig. 2.

As Fig. 2 demonstrates, whenever the collector concentrationapproaches the critical micelle concentration, micelles appear inthe solution and flotation drops to zero. This is not surprising sincemicelles are colloidal hydrophilic entities and their accumulationon the mineral surface must render such a surface hydrophilic.

Fig. 5. Effect of electrolyte concentration on the Krafft point of dodecylammoniumchloride (Laskowski et al., 2007).

-6 -5 -4 -3 -2 -1Concentration (log C)

0

50

100

Rec

over

y (%

)

C14

C12C10

CMC

Fig. 2. Effect of concentration of sodium alkyl-sulfonates on flotation of barite:empty circles, sodium tetradecyl sulfonate (C14); filled circles, sodium dodecylsulfonate (C12); empty squares, sodium decyl sulfonate (C10). Vertical arrowsindicate the critical micelle concentrations for the three studied alkyl-sulfonates at40 �C. (Dobias, 1986). Right-hand side insert shows a micelle on solid surface.

Fig. 3. Relationship between KCl recovery and amine addition (Roman et al., 1968).

Fig. 4. Schematic representation of the solubility of ionic surfactants versustemperature.

172 J.S. Laskowski / Minerals Engineering 45 (2013) 170–179

What is surprising is that this relationship – as shown by Romanet al. (1968) – is quite different in the potash flotation (Fig. 3).

Fig. 3 shows that the flotation of sylvite commences when thesolubility limit of amine1 is exceeded. Together, Figs. 2 and 3indicate that critical micelle concentration (c.m.c.) should not beconfused with the saturation concentration of the hydrated crystals(Cases and Villieras, 1992).

3. Krafft point

For ionic surfactants, the solubility curve plotted as a function oftemperature reveals two large domains (Fig. 4). At temperaturesbelow the Krafft point (TK), the solubility curve describes the satu-ration concentration of a hydrated crystal in equilibrium withmonomers (single surfactant molecules) in solution. At T < TK,when concentration of the surfactant increases (over solubilitylimit) the hydrated crystals start appearing in the solution; thisphase has a lamellar structure, with the polar groups lying alongthe interface with water. At T > TK, when the surfactant concentra-tion is increased, the monomers associate to form micellar aggre-gates; the concentration at which micelles first occur is referredto as a critical micelle concentration (cmc). In the three distinctzones in Fig. 4 distinct species appear: in Zone I only single surfac-tant molecules (monomers); in Zone II hydrated crystals in equilib-rium with monomers; and in Zone III micelles in equilibrium withmonomers. At temperatures lower than the Krafft point, the solu-bility is too low for micellization. The Krafft point is also definedas the temperature at which the solubility curve reaches the criti-cal micelle concentration; further increases in temperature sharplyenhance the solubility of the surfactant due to the formation ofmicelles.

As this discussion reveals, the hydrated crystals which appear attemperature T < TK should not be confused with micelles presentwhen T > TK. Fig. 2 demonstrates that flotation drops to zero whenthe collector concentration approaches cmc, while Fig. 3 showsthat sylvite flotation commences when the amine (collector) con-

1 The term ‘‘amine’’ stands in this publication for alkylamine with the number ofrbon atoms higher than 12.

ca

Fig. 6. NaCl (average size 280 lm) and KCl (average size 35 lm) particles insaturated NaCl–KCl brine (Roman et al., 1968).

J.S. Laskowski / Minerals Engineering 45 (2013) 170–179 173

centration exceeds the solubility limit.The difference between Figs. 2 and 3 suggest the obvious: the

colloidal species which appear in these two systems must havevery different properties, and a further discussion of these phe-nomena requires some knowledge of the Krafft point of long-chainamines. Additionally, the effect of electrolyte concentration on theamine Krafft point must also be characterized.

The available information on the Krafft point of amines is ratherlimited. The Krafft point values for dodecyl (C12) and octadecyl(C18) ammonium chlorides were reported to be 26 �C and 56 �C,respectively (Brandrup and Immergut, 1975). For dodecyl ammo-nium chloride Dai and Laskowski (1991) provided a value of17 �C. Theoretic estimations (Laskowski, 1994) indicate that theKrafft point for amines should dramatically increase with brineconcentration (the term brine used in this text stands for theNaCl–KCl saturated brine at room temperature). Fig. 5 shows theeffect of brine concentration on the Krafft point of dodecylamine.It is 80 �C at 16% brine concentration (approximately 1 M solutionof NaCl–KCl). If in concentrated electrolyte solutions C12 and C18amines observe the same relationship between Krafft point andchain length, then it can be expected that in a 1 M solution ofNaCl–KCl the Krafft point for C18 amine exceeds 100 �C. This leadsto the conclusion that all commercial potash ore flotation plants, aswell as all lab flotation experiments, are carried out at tempera-tures much lower than the Krafft temperature of long-chain pri-mary amines. Micelles do not exist in such systems. What willappear in the pulp when amines precipitate are solid particles, hy-drated crystals.

Fig. 7. Effect of pH and concentration on the electrophoretic mobility of thecolloidal precipitate in aqueous solutions of dodecylamine (Laskowski et al., 1988).

4. Electrical charge

Another important contribution brought about in Roman et al.’s(1968) paper was evidence that the particles of sylvite and halitecarry an electrical charge. Postulation of the electrical charge onthe particles suspended in brine was considered a blasphemy atthe time, but evidence for it was quite convincing. Roman et al.showed that while the suspensions of fine sylvite particles sus-pended in their own brine were very stable, and the suspensionsof fine halite in their own brine were stable too, they both imme-diately coagulated when mixed together (Fig. 6). The obviousexplanation was that these particles carried different electricalcharge. Somewhat later, Miller et al. (1992) used Doppler electro-phoresis and proved that while sylvite particles in brine carry neg-ative electrical charge, halite particles are charged positively.

At that time it was already known (Castro et al., 1986; Laskow-ski et al., 1988) that the particles of the precipitating amine are alsoelectrically charged. As Fig. 7 demonstrates, colloidal amine parti-cles are characterized by the clear iso-electric point which is situ-ated at a pH of approximately 11. Thus, the amine particles arepositively charged below pH 11 and negatively charged whenpH > 11.

These parallel but independent observations of electrical chargeof both sylvite and the precipitating amine, made it possible to ex-plain (Laskowski, 1994) the sylvite flotation curves published bySchubert (1967, 1988) (Fig. 8). The analysis reveals that whileKCl floats with amines at pH < 10.5, the flotation of NaCl takesplace only when this pH is exceeded. A comparison of Fig. 7 withFig. 8 leads to the conclusion that the negative electrical chargeof sylvite particles and the positive charge of the amine precipitateresults first in Coulombic attraction which then leads to potash oreflotation.

Hancer et al. (2001) questioned the surface charge-controlledcollector adsorption model and showed that the flotation of solublesalts is dictated by the ability of respected cations and anions toorganize the structure of interfacial water. Small cations such as

Na+ strongly interact with interfacial water molecules and stabilizethe interfacial water layer at the structure-maker NaCl surface.Consequently, octa-decyl amine (ODA) adsorption by the replace-ment of interfacial of water molecules cannot take place. In thecase of KCl with the larger cation, K+, it is found that ODA adsorp-tion is possible by attachment of the positively charged polar headgroup at the structure breaker KCl surface defects (Cao et al., 2010).However, while this explains the difference between good flotationof KCl and poor flotation of NaCl it does not explain the correlationbetween flotation response and surface charge of colloidal amineparticles shown in Fig. 8.

5. Use of amines in commercial potash ore flotation

Long-chain primary amines utilized as a collector are practicallyinsoluble in water. Leja (1983) compared various experimentaldata on the solubility of long-chain primary amines and concludedthat for all Cn > 16 amines, solubility in brine converges to levelsbelow 10�8 mole/L. The poor solubility of primary amines was also

Fig. 8. The effect of pH on flotation of sylvite and halite with C12 and C16 primaryamines (Schubert, 1967).

174 J.S. Laskowski / Minerals Engineering 45 (2013) 170–179

evidenced by Rogers (1957). He observed that the surface tensionmeasurements with saturated NaCl–KCl brine to which dodecy-lammonium chloride crystals were added (equivalent of2.26 � 10�4 mole/L) did not froth even after a week and surfacetension dropped only by a few mJ/m2.

The dosage of amines in industrial plants in Saskatchewan is inthe range from 60 to 95 g/t (Strathdee et al., 1982) which corre-sponds to approximately 10�4 mole/L, thus exceeding by far theamine solubility limit. Before use the amines are melted by heatingthem up to 70–90 �C, and are then neutralized with hydrochloric oracetic acids. The hot emulsion/dispersion is introduced into the flo-tation pulp, which is at room temperature. Since this temperatureis much below the Krafft point of the utilized amines, precipitationensues. According to all reported observations, a white precipitateappears immediately when the hot emulsion of amine is added tothe potash flotation pulp, and that it accumulates on the surface ofbubbles.

In commercial plants, after crushing and mechanical deslimingthe potash ore still requires the subsequent application of ‘‘blind-ers’’ to depress residual slimes and conserve valuable collector(Arsentiev et al., 1988). Depressants (blinders) include: carboxy-methyl cellulose, guar gum, and starch, etc. This is followed by

Fig. 9. (A) DDA colloidal particles on the surface of the KCl, and (B) DDA colloid

treatment with the cationic collector. In addition – in the flotationof coarse fractions – an extender oil is also utilized. A frother (e.g.MIBC) is added just prior to the flotation cells (Strathdee et al.,2007).

6. Amine precipitate in sylvite flotation

As demonstrated by Leja (1983), ‘‘in quiescent environment nocontact angle or pick-up of sylvite particles was observed evenafter deposition of amine-alcohol paste on the surface of the un-stirred brine. However, after thorough stirring for a few minutes,KCl particles were picked up and contact angle was developed onKCl discs.’’ The bubble/sylvite attachment was clearly possiblewhen the collector was manually placed directly on the surfaceof the bubble. When a wire coated with the collector-alcohol pastewas placed in brine on the other hand, there was no bubble/sylviteparticle pick-up, even after hours of immersion. But as soon as theprobe was touched first to the captive bubble, and the latter wascontacted with a KCl disc, contact angle immediately developed.These findings point toward adequate agitation as an importantstep, without which the surfactants utilized in potash ore flotationare not able to perform their function.

Since long-chain amines are insoluble in brine, this led to theconclusion that the mechanism responsible for flotation in thisparticular case must be different from conventional flotation inwhich collector adsorption renders the treated mineral hydropho-bic. It was postulated that the collector in the potash ore flotationpulp is transported by bubbles.

The effects described by Leja were quantified by Burdukova andLaskowski (2009). Their tests were designed to clarify how the pre-cipitating amine particles function in a potash ore flotation system.The amine dispersion was placed either onto the surface of a bub-ble, which was then contacted with a KCl plate and measurementof contact angle followed, or the amine dispersion was placed ontothe surface of a KCl plate, which was then contacted with a bubbleto measure the contact angle (Fig. 9). Special procedures had to bedeveloped for these tests since, in the former case, any previouscontact of the KCl plate with the solution/air interface was to betotally prevented. In the latter case the bubble was not to be con-taminated by possible surfactant adsorption (Schreithofer and Las-kowski, 2006; Burdukova and Laskowski, 2009; Laskowski, 2010).In 1995, Wang et al. reported (Wang et al., 1995) that frother,depending on the way it is introduced into the pulp, may have astrong effect on potash ore flotation (this effect was also confirmedby other researchers, see for example Monte and Oliveira, 2004).The effect of the presence of MIBC was also studied in the tests dis-cussed in this paper. Dodecyl amine was heated (70 �C) and thendispersed in a hot (70 �C) diluted HCl solution. In separate tests

al particles on the surface of the bubble (Burdukova and Laskowski, 2009).

Fig. 10. Advancing contact angle measurements on KCl plates in NaCl–KClsaturated brine. The 1% amine dispersion was placed directly onto the surface ofthe capillary used to generate the bubble, or onto the surface of KCl plate(Burdukova and Laskowski, 2009).

Fig. 11. Standard deviation of advancing contact angle as a function of the locationof amine and the presence/absence of MIBC (Burdukova and Laskowski, 2009).

J.S. Laskowski / Minerals Engineering 45 (2013) 170–179 175

MIBC was added to the hot HCl solution before the amine was dis-persed in it.

The results of these contact angle measurements are shown inFig. 10 in the form of Gaussian distributions. It is worth noting thatthe contact angle values exhibit a very high degree of scatter andthis necessitated up to 30 repeat measurements per condition toobtain a representative contact angle distribution. The results ob-tained with the use of amine fall more or less distinctly into twogroups. In one group, represented by filled markers, the KCl surfacewas not very hydrophobic when the amine dispersion was placedon the KCl surface (mean contact angle was approximately 40�)and the surface was very heterogeneous. The other group, depictedwith unfilled markers, showed the largest contact angle values(50–60�); in these tests the amine was placed onto the bubble be-fore contacting it with KCl plate. The presence of MIBC in the amineclearly increased the hydrophobicity of the KCl plate, especially inthose tests in which amine was deposited onto bubbles. As the left-most curve indicates, the KCl plate in saturated brine is to some ex-tent hydrophobic, a fact first reported by Hancer et al. (2001).

This was further confirmed by Ozdemir et al. (2011), who wereable to show that sylvite in its saturated brine is slightly hydropho-bic (advancing contact angle on KCl was measured to be 7.9 ± 0.5�).

The mean contact angle values are not the only information thatcan be derived from the measured contact angle distributions. AsFig. 10 reveals and Fig. 11 summarizes, there is a significant differ-ence in the degree of spread of the results. The lowest standarddeviations are obtained when the amine collector is present onthe surface of bubbles. In this case the resulting collector coverageis apparently more even and uniform than when the collector isdeposited directly on the KCl surface. In the absence of MIBC,amine colloidal particles provide patchy and uneven surface cover-age. As rheologic tests demonstrate (Laskowski et al., 2008), theprecipitating amine is better dispersed when MIBC is present inthe system. This was confirmed by direct measurements of the par-ticle size distribution of the amine precipitate in brine, and by tur-bidity measurements of such disperse systems (Burdukova et al.,2009). These later tests reveal that MIBC serves as a strong dispers-ing agent, affecting both the size and the morphology of the amineparticles, when it is added to the system at the same time as thehot amine is dispersed in hot water.

Fig. 12 demonstrates that solid particles are present in theprepared amine dispersion. Furthermore, it reveals that while the

precipitate prepared with MIBC resembles rather fine amorphousparticles, the precipitate without MIBC seems to be more crystal-line. A more detailed study on the crystalline nature of such parti-cles is, however, missing. Several experiments showed that whilesuch amine particles easily attach to KCl plates, they do not attachto NaCl plates.

7. Molecular films

Amphipathic compounds, insoluble in water, can form molecu-lar films at the water/gas interface (Gaines, 1966). The stability ofmonolayers depends on solubility; fatty acids form stable mono-layers on water in acidic solutions when they are not ionized, whilelong-chain amines form stable monolayers on water over alkalinepH range (Gaines, 1982). It was reported that with long-chainamines (e.g. C22 amine) condensed monolayers are formed undera wide range of conditions. Spreading on concentrated salt solutionis enhanced.

Arsentiev and Leja (1976) studied the monolayers of primaryamines (from C12 to C18) spread and compressed on saturated saltsolutions. The spreading of surfactants as films was carried outusing solutions of these surfactants in hexane. Several findings dis-cussed in their paper deserve to be highlighted. The adhesion be-tween planar discs of KCl brought up from underneath and theamine molecular film were very strong; the adhesion betweenthe NaCl discs and the film was about three times weaker(Fig. 13). The ability of a surfactant to adhere to a KCl disc de-pended to a great extent on the degree to which the surfactantmolecules form a condensed film. With progressive compressionof the film (achieved using Langmuir balance), the adhesion forcesincreased to a maximum. In flotation systems, the structure of thecollector film at the liquid/gas interface is determined by the struc-ture of the surfactant molecule and the length of the hydrocarbonchain. Arsentiev and Leja reported no adhesion between the filmspread using the amines containing branch chains and the solidKCl crystal. This provides strong corroborative evidence for Bach-mann’s postulate (Bachmann, 1951), that the flotation of KCl usingprimary amines is due to a favorable correlation in the dimensionsof the KCl crystal lattice and the quasicrystalline structure of thecondensed amine film.

In 1983, O’Brien et al. (1983) and Leja (1983) provided furtherexperimental results regarding amine molecular films. Since in flo-tation particle-bubble collisions take place within milliseconds, the

Fig. 12. Images of the DDA particles precipitating in a rapidly cooled dispersion. On the left: with MIBC; on the right: without MIBC (Burdukova et al., 2009).

Fig. 13. Schematically adhesion of the amine film to either NaCl crystal or KCl crystal (Arsentiev and Leja, 1976).

176 J.S. Laskowski / Minerals Engineering 45 (2013) 170–179

kinetics of the studied phenomena largely determine whetherthese sub-processes are or are not important in the flotation. Inspreading rate measurements, it was shown that while the spread-ing rates on saturated salt solutions were negligible for pure long-chain amines, these rates increase dramatically when the aminesare mixed with alcohols (e.g. hexanol). The tested flotation fro-thers, such as MIBC and DF-250, behave similarly to hexanol. Thespreading rates for C16–C18 amines mixed with hexanol were inthe range from 300 to 400 mm/s, which – when the size of the bub-bles in flotation systems is considered – should be quite sufficientfor the spreading to play an important role in the potash flotationsystem.

8. The mechanism

In this chapter the author has endeavored to reconcile variousgroups of facts, some established, accepted principles, others ten-tative. The basis for the proposed reconciliation is the way thatthe amines are utilized in commercial potash ore flotation pants.

In industrial practice, the long-chain amines, since they areinsoluble in water, are melted by heating up to 70–90 �C. Theyare then neutralized with hydrochloric (or acetic) acid and such ahot emulsion/dispersion is introduced into the flotation pulpwhich is at room temperature, the temperature much lower thanthe Krafft point of the utilized long-chain amine. As a result, awhite precipitate immediately appears and the fine particles ofthe precipitating amine coat the surface of bubbles. The amine par-ticles also attach to the KCl particles. The amine deposited at theliquid/gas interface starts spreading into a molecular film. Thismechanism is schematically shown in Fig. 14. While spreading is

fast and efficient in the presence of a co-surfactant (e.g. frother),it is not so in its absence (Arsentiev and Leja, 1976). The appear-ance of an amine molecular film on the bubbles converts thesebubbles into ‘‘active bubbles,’’ which easily pick up KCl particles.

In the proposed mode of the action of amines in potash ore flo-tation, the precipitating amine coats bubbles but also attaches toKCl particles (Burdukova and Laskowski, 2009). The outcome isvery different in these two cases. While the amine coating bubblesstarts spreading into molecular film that activates the bubbles withregard to their ability to attach to KCl surfaces, such spreading isimpossible in the case of amine particles attached to KCl surfaces.As Fig. 10 demonstrates, the KCl particles attaching to the ‘‘active’’bubbles (bubbles with amine molecular films) become very hydro-phobic, but the KCl particles coated by the precipitating amine arenot. The proposed mechanism strongly depends on the ability ofthe precipitating amine to quickly spread at the liquid/gas inter-face into a molecular film. This depends on the presence of a co-surfactant, such as a flotation frother. If the frother is mixed withthe amine at the stage when the amine is dispersed in a hot, acid-ified aqueous solution, the spreading of amine into a molecularfilm is dramatically enhanced. Therefore, incorporation of thefrother into the amine-frother combination increases the propor-tion of the amine which acts as an active component in the potashore flotation process.

9. Frothers in potash flotation

On a practical level, it is of interest to point out how the pro-posed mechanism might be used to further improve the potashore flotation process.

Fig. 14. Schematic dispersion of amine and its introduction to the flotation pulp, spreading of the amine into molecular film on the surface of bubbles followed by theattachment of KCl particles to such active bubbles.

Fig. 15. Effect of MIBC and electrolyte concentration on bubble size in an open-topLeeds flotation cell (Laskowski et al., 2003).

J.S. Laskowski / Minerals Engineering 45 (2013) 170–179 177

The flotation process commonly requires various flotation re-agents, not only collectors but also frothers and various modifiers.The frothers are used to make dispersion of air into fine bubblespossible, to stabilize the froth to some extent, and to facilitatethe particle-to-bubble attachment, as postulated by Leja andSchulman in their penetration theory (Leja and Schulman, 1954).Various reagents utilized in potash ore flotation are added to thepulp in the following order: depressants (blinders), amine collec-tor, and frother (commonly MIBC) just prior to the flotation cells.In spite of the very important differences between potash ore flo-tation and conventional flotation processes, flotation technologyas used in the potash ore flotation was derived from its ore flota-tion sibling. The use of the frother is a typical example.

But if it is used as in any other flotation operation, then thequestion arises whether the frother is needed in brine to enhanceproduction of fine bubbles.

As Fig. 15 demonstrates, the use of a frother (e.g. MIBC) is abso-lutely necessary in distilled water (or, in general, in aqueous solu-tions of a low ionic strength). The bubbles coalesce when theycollide in the cell and MIBC is needed to stabilize them against coa-lescence. At concentrations exceeding the critical coalescence con-centration (c.c.c.), which is approximately 10 ppm for MIBC, thebubbles generated in a flotation cell do not coalesce. Under suchconditions, it is possible to produce fine bubbles that are essentialin the flotation process. Fig. 15 also shows that in brine (and also in50% brine), the bubbles are quite stable and do not coalesce even inthe absence of frother. Evidently, a frother is not needed in potashore flotation to enhance production of fine bubbles. There is quite alarge number of papers in which prevention of bubble coalescencewas studied in electrolyte solutions; this topic is outside the scopeof the present paper and will not be discussed further.

Since coarse concentrate from potash ore flotation can be soldto the fertilizer industry without compaction, it has a special value.The ore is therefore not extensively comminuted, and the coarse

(+0.8 mm) fraction is separately conditioned not only with aminebut also with extender oil (ESSO 904, a distilled refinery bottom-distillate, is commonly used by the Potash Corporation of Saskatch-ewan). In the paper presented at the 24th International MineralProcessing Congress in Beijing (Laskowski et al., 2008), the PCSLanigan potash ore crushed below 1 mm was used. The results ofthe additional lab bench flotation tests, carried out to furtherexamine the effect of the way of the collector and frother prepara-tion and addition, are given in Tables 2 and 3. When the ArmeenHTD (from AKZO Nobel) aqueous dispersion is prepared in the

Table 2Effect of the way MIBC is utilized on the flotation of �1 mm potash ore fraction. Testconditions: Addition of 300 g/t CMC was followed by the introduction of 100 g/t ofArmeen HTD and 50 g/t of MIBC either separately or as a combination (WI stands forwater insoluble minerals).

Method Assay Concentrate Tailings

Grade (%) Recovery (%) Grade (%) Loss (%)

Separate KCl 96.9 86.3 8.9 13.7WI 0.4 16.2 1.3 83.8

Combined KCl 94.1 96.0 3.3 4.0WI 0.6 24.9 1.5 75.1

Table 3Effect of the way MIBC is utilized in rougher and scavenger flotation of potash ore.Test conditions: Addition of 300 g/t CMC was followed by the separate introduction of100 g/t of Armeen HTD and 50 g/t of MIBC in rougher flotation. In the scavengerflotation 50 g/t of Armeen HTD and 35 g/t of MIBC were either added as a combinationor separately (WI stands for water insoluble minerals).

Flows. Assay Rougher conc. Scavenger conc. Tailings

Grade(%)

Recovery(%)

Grade(%)

Recovery(%)

Grade(%)

Loss(%)

A Separate addition Combined additionKCl 98.1 80.0 96.7 14.3 7.9 5.6WI 0.5 24.0 0.5 3.9 1.7 72.1

BKCl 98.0 80.9 98.7 2.6 19.4 16.5WI 0.4 23.6 0.5 0.8 1.4 75.5

0 0.2 0.4 0.6 0.8 1 1.2Particle size, mm

0

20

40

60

80

100

Size

reco

very

, %

combine

separate

Fig. 16. Comparison of KCl size-by size recoveries in the flotation of the �1 mmpotash ore fraction (Laskowski et al., 2008).

178 J.S. Laskowski / Minerals Engineering 45 (2013) 170–179

standard way at 70 �C and used at room temperature, the KClrecoveries in batch flotation tests are very different depending onwhether the collector and frother are used separately or in combi-nation (86.3% and 96.0%, respectively). If the rougher flotation isfollowed by a scavenger flotation in which the collector and frotherare added separately, the overall recovery improves from 80.9% to83.6%, thus only by 2.6%. When the rougher flotation is followed bythe scavenger flotation with the use of the amine-MIBC combina-tion, the overall KCl recovery increases by 14.3 % (Table 2).

As Fig. 16 reveals, the difference in KCl recovery in the flotationexperiments with either ‘‘combined’’ or ‘‘separate’’ use of ArmeenHTD and MIBC results from the much better flotation of coarse par-ticles in the tests with the Armeen HTD–MIBC combination. Theflotation of coarse particles is much more sensitive to the hydro-phobicity of the particles than in the flotation of relatively finepar-ticles (Trahar, 1981). While conditions far from the optimum are

still satisfactory for the flotation of fine particles, these conditionsare obviously not good enough for the flotation of the coarse par-ticles. The flotation of the coarse fraction strongly depends onthe proportion of amine which appears in the flotation system inits active form (on the surface of bubbles). These results imply thatthe coarse fractions of potash ore, which today are floated with theaddition of not only amine but also of extender oil, could be floatedusing the same amine and frother but prepared as a singlecombination.

10. Summary

While advances in experimental techniques for studying sur-face chemistry effects and the use of these techniques in examiningthe phenomena accompanying flotation of soluble salts are welldocumented the topic of the preparation of reagents that are usedin these experiments is shamefully avoided. The silent assumptionseems to be that this topic is not different from similar studies inother areas and does not need any special attention. But thisassumption is totally baseless and it is important to realize thatthere are some fundamental differences between the conditionsof commercial potash flotation and the most lab bench flotationtests reported in literature.

Since long-chain amines used in potash ore flotation as collec-tors are water-insoluble, in commercial practice, to improve itshandling, the amine is melted at 70–90 �C before dispersion inhot acidified water. Once added to the flotation pulp, the hot aminedispersion rapidly cools down to a temperature far below the Kafftpoint. All reported observations reveal that a white precipitateimmediately appears wen the hot amine dispersion is added tothe potash ore flotation pulp. The rapid conversion from a hotemulsion to a cold precipitate is a very severe transformation butnothing is known about the kinetics of these changes. As weshowed it (Burdukova and Laskowski, 2009)the precipitatingamine coats bubbles, but can also attach to KCl particles. In thepresence of co-surfactants (e.g. frothers), the amine spreads intoa molecular film at the liquid/gas interface, activating bubbles.Such active bubbles easily pick up KCl particles.

The proportion between the active form of amine (the part thatspreads into a molecular film on bubbles) and the non-active form(the part that attaches directly to KCl particles) depends on theway the amine and frother (co-surfactant) are utilized. Evidenceis presented that this proportion is particularly important for flota-tion of coarse sylvite particles.

Acknowledgements

This work has been supported by grants from the Natural Sci-ences and Engineering Research Council of Canada, the Potash Cor-poration of Saskatchewan, AKZO Nobel, Agrium Potash Inc., andIMC Kalium.

The research reported emanates from the papers and theses ofmany graduate students and post-docs of the University of BritishColumbia: Ms. Elena A. Alonso, Dr. Elizaveta Burdukova-Forbes, Dr.Marek Pawlik, Mr. Carlos Perucca, Dr.X.M. Yuan and Dr. Qun Wang.Extensive discussions with Dr. Graeme Strathdee are gratefullyacknowledged. Special thanks go to Ms. Sally Finora for her friendlyhelp with figures, and to Dr. Kornel Laskowski for the final lan-guage adjustment. The paper is dedicated to the late ProfessorJan Leja.

References

Arsentiev, V.A., Leja, J., 1976. Interaction of alkali halides with insoluble films offatty amines and acids. In: Kerker, M. (Ed.). Colloid Interface Sci. 5, 251–270.

J.S. Laskowski / Minerals Engineering 45 (2013) 170–179 179

Arsentiev, V.A., Dendyuk, T.V., Gorlovsky, S.I., 1988. The effect of synergism toimprove the efficiency of non-sulfide ores flotation. In: Forssberg, F. (Ed.), Proc.16th Int. Mineral Processing Congress, Elsevier, Amsterdam, pp. 1439–1449.,pp. 1439-1449.

Bachmann, R., 1951. Beitrag zur Theorie der Schwimmaufbereitung. Erzmetall 4,316–320.

Brandrup, J., Immergut, E.H. (Eds.), 1975. Polymer Handbook, second ed. Wiley, p. II489.

O’Brien, R.N., Visaisouk, S., Leja, J., 1983. A multiple surface balance use and resultsin selecting suitable flotation reagents. In: McKercher, R.M. (Ed.), PotashTechnology. Pergamon Press, pp. 619–622.

Burdukova, E., Laskowski, J.S., 2009. Effect of insoluble amine on bubble surface onparticle-bubble attachment in potash flotation. Can. J. Chem. Eng. 87, 441–447.

Burdukova, E., Laskowski, J.S., Forbes, G.R., 2009. Precipitation of dodecylamine inKCl–NaCl saturated brine and attachment of amine particles to KCl and NaClsurfaces. Int. J. Min. Proc. 3, 34–40.

Cao, Q., Du, H., Miller, J.D., Wang, X., Cheng, F., 2010. Surface chemistry features inthe flotation of KCl. Miner. Eng. 23, 365–373.

Cao, Q., Wang, X., Miller, J.D., Cheng, F., Jiao, Y., 2011. Bubble attachment time andFITR analysis of water structure in the flotation of sylvite, bischofite andcarnalite. Miner. Eng. 24, 108–114.

Cases, J.M., Villieras, F., 1992. Therodynamic model of ionic and non-ionic surfactantadsorption–abstraction on heterogeneous surfaces. Langmuir 8, 1251–1264.

Castro, S.H., Vurdela, R.M., Laskowski, J.S., 1986. The surface association andprecipitation of surfactant species in alkaline dodecylammine hydrochloridesolutions, Colloids & Surfaces, vol. 21, Cases and Villieras, 1992. P. 87–100.

Cormode, D.A., 1985. Insoluble slimes removal from sylvinite ore by selectiveflocculation and flotation, In: 87th Annual Meeting of CIM, Vancouver, April 21–25.

Dai, Q., Laskowski, J.S., 1991. The Krafft point of dodecylammonium chloride: pHeffect. Langmuir 7, 1361–1364.

Dobias, B., 1986. Adsorption, electrokinetics, and flotation properties of mineralsabove the critical micelle concentration, Phenomena in Mixed SurfactantSystems, In: ACS Symposium Series, 311, 216.

Fuerstenau, D.W., Fuerstenau, M.C., 1956. Ionic size in flotation collection of alkalihalides. Min. Eng. 224, 302–307.

Gaska, R.A., Goodenough, R.D., Stuart, G.A., 1965. Ammonia as a solvent. Chem. Eng.Prog. 61, 139–144.

Gaudin, A.M., 1957. Flotation. McGraw-Hill, New York.Gaines, G.L., 1966. Insoluble Monolayers at Liquid–Gas Interfaces. Interscience Publ.Gaines, G.L., 1982. Langmuir–Blodgett films of long-chain amines. Nature 298, 544–

545.Hancer, M., Celik, M.S., Miller, J.D., 2001. The significance of interfacial water

structure in soluble salt flotation systems. J. Colloid Interf. Sci. 235, 150–161.Jeanprost, C., 1928. French patent 669,100.Kirby, J.E., 1937. US patent 2,088,325.Laskowski, J.S., 1994. Flotation of potash ores, Reagents for Better Metallurgy, In:

Mulukutla, P.S. (Ed.), SME, p. 225–243.Laskowski, J.S., 2010. A new approach to classification of flotation collectors. Can.

Metall. Quart. 49, 397–404.

Laskowski, J.S., Vurdela, R.M., Liu, Q., The colloid chemistry of weak-electrolytecollector flotation, In: Forssber g K.S.E., (Ed.), Proc. 16th Int. Mineral ProcesingCongress, Elsevier –Developments in Mineral Processing, vol. 10A, 1988, p. 703-715.

Laskowski, J.S., Cho, Y.S., Ding, K., 2003. Effect of frothers on bubble size and foamstability in potash ore flotation systems. Can. J. Chem. Eng. 81, 63–69.

Laskowski, J.S., Pawlik, M., Ansari, A., 2007. Effect of brine concentration on theKrafft point of long chain primary amines. Can. Metall. Quart. 46, 295–300.

Laskowski, J.S., Yuan, X.M., Alonso, E.A., Potash ore flotation – how does it work?, In:Proc. 24th Int. Mineral Processing Congress, vol. 1, Beijing, 2008, p. 1270-1276.

Leja, J., On the action of long chain amines in potash flotation, Potash Technology,In: McKercher, R.M. (Ed.), Pergamon Press, 1983, p. 623–629.

Leja, J., Schulman, J.H., 1954. Flotation theory: molecular interactions betweenfrothers and collectors at solid-liquid-air interfaces. Trans. AIME 199, 221–228.

Miller, J.D., Yalamanchili, M.R., Kellar, J.J., 1992. Surface charge of alkali halideparticles as determined by Laser-Doppler electrophoresis. Langmuir 8, 1464–1469.

Monte, M.B.M., Oliveira, J.F., 2004. Flotation of sylvite with dodecylamine and theeffect of added long chain alcohols. Miner. Eng. 17, 425–430.

Ozdemir, O., Du, H., Karakashev, S.I., Nguyen, A.V., Celik, M.S., Miller, J.D., 2011.Understanding the role of ion interactions in soluble salt flotation withalkylammonium and alkylsulfate collectors. Adv. Colloid Interface Sci. 163, 1–22.

Rogers, J., 1957. Flotation of soluble salts. Trans. IMM 66, 439–452.Rogers, J., Schulman, J.H., A mechanism of the selective flotation of soluble salts in

saturated solutions, In: Schulman, J.H. (ed.), Proc. 2nd Int. Congress SurfaceActivity, vol. 3 London, Butterworths, 1957, p. 243–251.

Roman, R.J., Fuerstenau, M.C., Seidel, D.C., 1968. Mechanism of soluble salt flotation.Trans. AIME 241, 56–64.

Schreithofer, N., Laskowski, J.S., Investigation of KCl crystal/NaCl–KCl saturatedbrine interface and octadecylamine deposition with the Atomic ForceMicroscope, Interfacial Phenomena in Fine Particle Technology, In: Xu Z., Liu,Q. (Eds.), Proc. 6th UBC-McGill-UA Int. Symposium, Metallurgical Society ofCIM, Montreal, 2006, p. 17–32.

Schubert, H., 1967. What goes on during potash flotation. Eng. Min. J., 94–97.Schubert, H., 1988. The mechanism of collector adsorption on salt-type minerals

from solutions containing high electrolyte concentrations. Aufbereit.-Technik29, 427–435.

Strathdee, G.G., Haryett, C.R., Douglas, C.A., Senior, M.V., Mitchell, J., 1982. Theprocessing of potash ores by PCS, In: 14th International Mineral ProcessingCongress, Toronto, Paper no. V-12.

Strathdee, G., Gotts, B., McEachen, R., Potassium flotation practice, Froth Flotation,In: Fuerstenau, M.C., Jameson, G., Yoon, R.H. (Eds.), A Century of InnovationSME, 2007, p. 533–543.

Trahar, W.J., 1981. A rational interpretation of the role of the particle size inflotation. Int. J. Min. Proc. 8, 289–327.

Wang, Q. Alonso, E.A., Laskowski, J.S., The effect of frothers on potash ore flotation,In: Proc. 19th Int. Mineral Processing Congress, SME, San Francisco, vol. 3, 1995,p. 49–53.