Interfacial Chemistry of Particulate Flotation
Interfacial Chemistry of Particulate Flotation
Interfacial Chemistry of Particulate Flotation
P. Somasundaran
Henry Krumb School of Mines
Columbia University
New YorkI New York 10027
}. large variety of chemical speciesranging from molecules again the basis for separation, but the adsorbed material is col.
and ions to micro-organisms and minerals can be separated lected selectively with a liquid (instead of air) that is immisci-
from one another or concentrated from solutions using flota. ble with the bulk aqueous solution. At present, froth flotation
tion processes. Even though the mechanismsinvolved in the is the only technique that has significant industrial application,
flotation or the nonflotation of each specieswith various reo and is therefore emphasized in this paper. Precipitate flota.
agents are not fuUy established, the flotation pro~sses can be tion, which appears to hold promise for future applications, is
considered to be primarily the result of the tendency of cer. also discussed.
tain surfa~ active speciesto concentrate at the liquid/gas
interfa~ and the tendency of some other speciesor particles METHODS
to associatewith or adsorb these surface active species. The
For froth flotation. a pulp of the particulates is first con.
large number of flotation processestested have been classified
ditioned with the appropriate reagents. It is then agitated us-
(Table I) according to the size (molecular, microscopic, or
ing impellers in a cell such as the one shown in Figure I. Air
macroscopic) of the flotated speciesand the mechanism (con-
is sucked in or sometimes fed into the cell near the impeDer
sistent with natural or acquired surface activity) of flotation.
lone. The air bubbles. now dispersed by the impeller. collide
Thus, there is foam fractionation for the separation of surface
with particles and are attached to those that are hydrophobic
active speciessuch as.detergentsin aqueous solutions, and
or have acquired hydrophobicity and rise to the cell top where
foam flotation for that of naturally surface active organisms
they are removed by skimming. Figure 1 is a schematic repre-
and proteins. Techniques based on the tendency of various
sentation of a typical industrial flotation cell.
speciesto associatewith surfactants are: ion flotation for the
As opposed to froth flotation which usesturbulent condi-
separation of ions, micro flotation and ultraflotation for very
fme particulates, and froth flotation for the separation of min- tions. foam separation techniques generally consist of aeration
erals that possessmostly polar surfaces. In addition to these at a low flow rate. and the separation of the foam containing
the collected material is followed by its breaking using various
techniques, there are also certain nonfoaming flotation
methods such as oil.flotation (2), bubble fractionation (3) and chemical, thermal or mechanical methods (5). A stripping
mode in which the descending feed is introduced into the
solvent sublation (4), for which adsorption at the interface is
foam and an enriching mode in which part of the foamate is
recycled to the top of the flotation column are recommended
Henry Krumb School of Mines, Columbia University, New York, for increasing the recovery and grade. respectively. of the
New York 10027. product (see Figure 2).
2 AIChE SYMPOSIUM SERIES
Size Range
Mechanism Molecular MiuOtCOt)ic Mecroscopic
Natllrat surfec8 activity Fo.m fractionation F08m flOClcion Froth flot.tion of nonoo~
IX: IX: miner."
dltlrgents from aqueous micrOOfpni."s. prOteins ex:
solutions sulfur
In essoc~tion with surf8C8 Ion flot8tiOn. molecul.r Microfloutlon. colloid IIo~tion. Froth flotation
Ktiw agentl floutlOn, Idsorbing ultra flotation ex:
colloid flotation ex: miner.'s such .s silic.
,.: P8rticuletes in _ste weter. clay. Precipitate flotation
5rl+, Pbl+, Hgl +, cyanide micrOOf't8nisms {1st.nd 2nd kind}
.x:
f.ric hydroxide
ReprInted from CtNn SurfKft 1970. Courtesy of Marcel ~ker Inc
BASIC PRINCIPLES
SELECTIVE HYDROPHOBICITY
As mentioned earlier. froth flotation of minerals or precipi-
tates is possible if they can be preferentially wetted by gas
rathe. than by water. Only a very smaUfraction of the min-
erals such as sulfur is naturaUy hydrophobic. Precipitates of
the second kind (6). formed by reaction between ions and cer-
tain organic reagents. are also hydrophobic. Hydrophobicity
has to be imparted to most other minerals and precipitates in
order to float them. Towards this purpose. a surfactant that
will selectively adsorb on the material to be floated is added to
the suspension in water and conditioned by agitation. This
surfactant. called a collector. possessesat least one nonpolar
and one polar portion. Owing to chemical. electrical. or phys-
ical attraction between the polar portions and the surface sites.
the collectors adsorb on the particles ~'ith their nonpolar ends
oriented towards the bulk solution. thereby imparting hydro-
phobicity to the particles. Collectors that are commonly used
include short chain alkylxanthates for base metal sulfides; long
chain fatly acids and their alkali soapsfor phosphates. berna-
No. ISO, Vol. 11 SOMASUNDARAN
this connection that due to the porosity of natural plena even under conditions of zero surface charge on the mineral. As-
freshly ground galena will possessoxidized layers, and hence, suming that potential differences due to dipoles remain con-
results reported for such systems do not necessarily determine stant, the total double layer potential is considered zero when
the effects of oxidation of the galenasurface on flotation. the surface charge is zero; such a condition is called the point
The reaction of xanthate with the oxidation products of of zero CM'Ke (pzc). For oxides, the activities a. and a~ will
galenaon the surface through an ion-exchange process is con- be that of hydrogen ions in the solution under consideration
sidered to be the major adsorption mechanism responsible for and that at the point of zero charge, respectively. Similarly,
the xanthate flotation of galena. Previous identification of Q- and a~ will be the activities of hydroxyl ions under corre-
dixanthogen on plena is attributed by GranviUe et al. to its sponding conditions. Thus the oxides will carry a positive
formation during the evaporation of lead xanthate in ether or charge in solutions that are more acidic than that a pzc and a
CSz onto a surface. Their tests, however, do not conclusively negative charge in solutions that are more alkaline. Sin~ the
identify the speciesresponsible for flotation, becausethe system IS a whole must be electric:a1lyneutral, the medium
amount of dixanthogen observed on the mineral could have surrounding the particles must contain an equivalent amount
been sufficient to causeflotation. Adsorption mechanisms for of ions, called counter ions, of charge opposite to that on the
sulfide flotation systems have been discussedelsewhere by surfa~ of the particle. Owing to the attraction by the charged
Finkelstein et al. (9) and hence are not covered in this paper. swface sites, these counter ions will not be uniformly distrib-
Unlike xanthates on galena,alkyl amines, sulfonates, and uted in the solution phase, but will be adsorbed at the oxide-
sulfates are considered to adsorb on minerals primarily due to solution interfa~. This gives rise to an electrical double layer
electrostatic attraction between the polar head of the coUector consisting of one layer of surface charge and another layer of
and the charged surface sites on the mineral. These electro- counter ions. However, becauseof thermal agiution, this sec-
static forces, which are not u strong as the chemical forces in- ond layer extends as a diffuse layer over a finite distan~ from
volved in xanthate adsorption, are assistedby the associative the particle surface. A schematic represenution of this elec-
interaction among the long alkyl chains that normally contain trical double layer is given in Figure 3.
8 to 20 carbon atoms (10-12). Becauseelectrostatic forces are Point of zero charge of a mineral is an important dwacter-
the basic cause for the selective adsorption in these cases,it is istic becausethe adsorption of various organic and inorganic
important to understand the electrical nature of the mineral/ ions will be governed by the location of the solution properties
solution interface and the mechanismsgoverning its origin. (as pH) with respect to the pzc. Typical pzc values are given
The electrical nature of the particle/solution interfal% is the in Table 2. It is important to note that these values are af-
result of either a preferential dissolution of lattice ions u in
the caseof silver iodide, or of the hydrolysis of the surface SURFACE
speciesfollowed by pH-dependent dissociation of the surface CH4Rj STERN
hydroxyls as in the caseof silica (I I) and alumina (13):
PLANE
- M(HzO)Surtace
--
--
- MOHSwf8c:8 ~OH- .-
H. ~:
- M°Sutt.~ + H2O 0
e 0 e
The lattice ions are considered as (surface) p>tential determin-
ing ions for AgI type solids, whereas H+ and OH- are the cor-
0 0 0
resp>ndinl ions for oxide minerals. For minerals sudt as ::; 0
0
calcite (J 4) and apatite (J 5, J 6), both of the above mecl1a. 0
nisms can be operative since the lattice ions can undergo pt'Cf.
erential dissolution as wen as reaction with H+ and OH-. In
the caseof these mmerab, the lattice ions H+ and OH- and cn~ SHEAR e POTENTIAL-
certain complexes of the lattice ions with H+ or OH- can be PLANE \!) DETERMINING IONS
.
p>tential determining. Silicate minerals with layered structure '110 HYDRATED
such as that of clays, on the other hand, are negatively charled ..J 0 COUNTER IONS
~
under most natural conditions due to the substituion, for ex.
t='
"'a 0 NEGATIVE CO-IONS
ample, of AlJ+ for Si4+ in the silica tetrahedra.
The surface p>tential '" a in these systems is given by
z
w ~
...
0
-RTF Inoa. or -RT
~o. Z z F ino
a- a..
. a. - a-
.Isoelectric point (iep) refers to die conditions under which the elec.
trokinetic potentia!, commonly termed zeta potential (the potentia!
that manifests in die region of shear between die liquid and the solid
when one is moving relative to die other), as determined by electro- such as leaching in acidic (35) and hot solutions (36) have also
kinetic experiments is zero. The surface potentia! tJI0 need not be zero been found to alter the interfacial properties drastically. Fig-
when the zeta potential is zero, particularly in the presenceof $pecif- ure 4, for example, gives values reported for the zeta potential
icaUy adsorbing ions. Therefore pzc and iep need not be the same. of quartz leached with different acids; the variations when us-
TABLE 3. POINT OF ZERO CHARGE OR ISOELECTRIC POINT VALUES REPORTED BY VARIOUS WORKERS FOR ALUMINA 1281
r, = 2rCexp.t~.!. trostatic (see Figure 5). When the concentration of the re-
RT
agent is increased, the adsorbed ions begin to associateto form
where' is the effective radius of the adsorbed ion, C is the two-dimensional aggregatescalled hem;m;ce/les. Unde{ this
bulk concentration in mole/crn3, ~ is the standard free condition, the adsorption is due, in addition to the electro-
energyof adsorpti~ is the gasconstant,and T the abso- static forces, to favorable energetics of removal of alkyl chains
lute temperature.~G~I is the driving force for adsorption from aqueous solutions (42). The above adsorption mecha-
and canbe consideredto be madeup of a numberof terths. nism is formulated on the basis of the experimental observa-
eachterm for a giventype of interaction that is responsiblefor tion that interfacial properties such as adsorption density, zeta
the adsorption(40,41). potential, flotation recovery, contact angle, sedimentation
rate, etc. undergo a marked change at a given surfactant con-
~ =dG:'ec + dG~ + dG:-c + dG:-s centration that is dependent upon pH and hence upon the sur-
face ootential of the oxide particles (12), solution temperature
+4Gft+4Gfl.o (43), chain length (10,11,44,45), and chemical composition
~ is the electrostaticinteraction term equal to zFIlI6 of the surfactant (45) and the solution (46 to 48). Adsorption
data for the alumina/dodecylsulfonate system is given in Fig-
wherez is the valencyof the adsorbate,F the Faradaycon-
stant and 1t16 the potentialat the S plane;~G~em is the chem- ure 6. Sharp increase in the slope of the adsorption isotherm
ical term due to any covalentbond formation betweenadsor- along with those of other parameters at an adsorption density
bateand adsorbent;~G~-c is due to the cohesivechain-chain of 10-11 mole/cm1 can be clearly seen L"l this figure.
The mode of adsorption of the previously-mentioned sur-
interactionthat could occurbetweenthe surfactantspecies
upon adsoprtion;~G~-I is similar to ~G:-c and is the vander factants on soluble salts such as sylvite is not established.
Waalsinteractionbetweenthe hydrophobicchain and the hy- Fuerstenauand Fuerstenau(50) haveproposedthat adsorption
in these casesis governed by matching the size of the polar
drophobicsiteson the mineralsurfa~; ~Gf. is due to hydro-
genbonding;~Gl\o is due to hydration or dehydrationof the head of the collector with the constituent ion of the solid
adsorbates or adsorbentsupon adsorption. Oneor more of the which carries the same charge as the polar head of the coUec-
abovetermscanpredominatefor eachsystem. l'hus, the tor. Thus amine adsorbs on sylvite (KCl) and not on halite
chemicalterm is consideredto be the predominantterm for (NaCl), due to comparable sizesof the aminium ion and K.
the adsorptionof xanthateson galena,and the electricaland ion. This theory, however, fails to explain. for example, why
the chain-chaininteractiontermsareimportant for the adsorp- the anionic aikylsulfate as a collector should distinguish be.
tion of alkylsulfonatesor aikylsulfateson oxidessuchas tween KCl and NaCl; it floats the former but not the latter.
alumina. According to an alternative theory proposed by Rogers and
The adsorptionof surfactantson aluminahasbeenrecently Schulman (51), adsorption is governed by the hydration prop-
studiedin detail (12,41). At low concentrationsof surfactant erties of the solid; the one with the larger negative heat of
and low surfa~ potentials,the surfactantions areindividually solution is the most amenable to flotation. This theory cannot
adsorbedand the for~ of attraction, responsiblefor adsorp- explain why a particular mineral such as KCl will be floated by
tion, betweentheseions and the mineralsurfa~ is mainly elec- certain collectorslike alkylsulfatebut not by carboxylatesor
6 ADVANCESIN INTERFACIAL PHENOMENA AIChE SYMPOSIUMSERIES
kinetia of adsorption at the liquid/air interface for the tion using short-chained alkylxanthates invariably requires the
hematite-oleate system is discussedelsewhere(54). The mi- addition of a frother. The desired concentration of the frother
gration of collector speciesfrom the bubble/liquid interface in the system is approximately that at which there is a signifi-
to the solid/gas interface caMot be expected to be nonselec- cant changein surface tension with concentration so that a
tive with respect to various minerals becausethe adsorption restoring force is available to prevent the rupture of bubbles
density at the interface of the bubble and the mineral particle subsequentto any local extension of the surface (see Figure 8).
(with possibly severallayers of water molecules strongly at- It is also necessarythat the diffusion of the surfactant species
tached) will be essentially detennined by the surface prop- from the subjacent surface to the locally extended surface
erties of the particle. region is not fast enough to reduce the difference between the
From Young's equation relating the three interfacial ten- surface pressurein the vicinity of the extended region and
sions '11,.,'1"., and '1" to the contact angle 9, one obtains the within it becausethis difference is mainly responsible for the
-
condition that '1", '1" must be smallerthan '11,.for a large restoring force.
contact angle and thereby good flotation (56). The larger the In addition to providing froth stability, the frother can play
value of '11,.,the lower is the froth stability the above condi- another role that must be taken into consideration in the
tion can be fulfilled only by keeping '1". small and '1" as large study of flotation mechanisms. Similarly to collector species
as possible. Becausethe adsorption of surfactant at any inter- adsorbed on the bubble, frother speciesalso can migrate from
face will only decreasethe interfacial tension there, it is clear the bubble surface to the particle/solution/bubble contact line
that the above condition can best be satisfied by allowing the and particle/bubble interfacial regions and co-adsorb on the
transfer of surfactant speciesfrom the bubble surface (or from particle along with the collector species(60 to 62). Such
the gaseousphase) to the solid/gas interface. Perhaps for this frother adsorption can be favorable for flotation, possibly be-
reason,Wada (58), on passingthe collector into the cell in the causethe neutral frother molecules adsorbing among the ionic
fonn of an aerosol with the gasstream, obtained optimum collector speciescan shield each adsorbed collector ion from
flotation at reagent concentrations as low as lis to I/aO of repulsion by others (63) and thereby enhance the overall ad-
those used in the conventional process. sorption of surfactant.
FROTHING ACTIVATION
Nonionic surfactants. generally slightly soluble, monohy- Activators enhance the flotation of a mineral by collectors
droxylated compounds such as cresol, are added to induce the that will not float it in their absence. Examples include cal-
desired froth stability during flotation. particularly when the cium activated flotation of quartz using oleate, and copper
chain length of the collector is relatively small. Sulfide flota- sulfate activated flotation of sphalerite with xanthate at rel-
atively high pH values. Activators normally act by adsorbing
at the mineral-solution interface, thereby providing sites for
z adsorption of the collector species. Adsorption of the acti-
Q vators is attributed to electrostatic attraction between them
(/)
Z and the mineral surface in such casesas calcium adsorption on
LiJ
..- negatively charged quartz. Bivalent ions such as calcium, upon
LiJ adsorption on minerals, are capable of reversing the sign of the
U
~ Stern potential and thereby cause the adsorption of collectors
~ that have a charge of the same sign as that of the mineral. In
(/) some cases,activators have been effective becauseof their re-
actions with collectors to form compounds of low solubility
product. This aspect is discussedin a paper by Fuerstenau
(64).
DEPRESStON
Reagentsthat prevent the adsorption of collectors on min-
erals and thereby retard their flotation are known as depres-
sants. Uke an activator, a depressant often reversesthe Stern
potential of the mineral, but the resulting potential in this case
is of same sign as that of the charge of the collector ion so that
collector adsorption is reduced.
CONCENTRATION OF SOLUTE Polymers such as starch can also depressflotation, but as
Fig. 8. Diagram illustratin& the correlation discussedin detail later, in this casecollector adsorption is en-
between froth stability and surface hanced by the depressant (65). The decreasein flotation ob.
tension lowerin& due to the addition tained in spite of the increased collector adsorption can be
of a surfacunt (after Cook (.59)).
attributed to the peculiar structure of the coUector-stardt
Reproduced by permission of John
WUey-lnterscience. clathrates formed with a hydrophilic exterior-
8 ADVANCES IN INTERfACIAL PHENOMENA AIChE SYMPOSIUM SERIES
DEACTIVATION
Deactivators prevent activation, generally, by interacting
with the activators to form an inert speciesof a stable ionic
complex. Addition of cyanides for the deactivation of copper
in the xanthate flotation of sulfides is a well-known example
of this type of deactivation reaction.
EFFECT OF VARIABLES
There is a large number of variables that affect flotation
processes. These variables originate from the physical and
chemical nature of the ore, its storage, its preparation includ-
ing grinding, the chemical constituents that may be present in
water, reagents added and the type of flotation cell used. It is
relevant to examine here only the physicochemical variables
that affect the interfacial processes. The nature of the min-
erals and other components in the ore wiU indeed govern the
interfacial processes,but there is very little control that one Fig. 10. Flotation of alumina as a function of pH with CF )(CF1)6
can exercise on this. However, with the composition of the COOK and CH)(CH1)6 COOK as collectors in 2 x 10-)
mole!1 KNO) solutions (after Somasundaranand Kulkalni
ore known, a knowledge of the effect of controUable variables
(45) J. Reproduced by permission of Institute of Mining and
can be helpful to obtain maximum processefficiency. Metallurgy (IMM) London.
CHAIN LENGTHOf THE COLLECTOR
Increase in length of the nonpolar part of a surfactant in- considerably (45). In the caseof perfluoro-carboxylates, an
creasesits adsorption at interfaces, thus generaUyenhancing additional advantage is their better pH tolerance, compared
the flotation of minerals. Results in Figure 9 for the flotation with any corresponding hydrocarbon homologue, attributed to
of quartz at natural pH using aikylammonium acetates of vary- the reduced tendency of the former to hydrolyze in acidic
ing chain length show an increase in flotation with an increase solutions (see Figure 10). However, their current cost is pro-
in chain length (10). The chain length effect on flotation was hibitive for any application in flotation processing. Uke per-
ascribed to the increased adsorption at the solid/liquid inter- fluoro compounds, silyl reagents are also reported to be highly
face owing to lateral interaction among the adsorbed species surface active (66). While geometrical modifications in the
to form two-dimensional aggregatesand at the liquid/gas inter- structure such as distribution of the methyl groups into var-
face. The length of the chain is often limited by solubility ious branches in general reduces the surface activity, addition
which decreaseswith increasing chain length. If very long of active groups at appropriate points in the surfactant mole-
chain surfactants are to be introduced into the flotation ceU, cule, so that the distance between the groups will correspond
it can be done through an oil emulsion. to the distance between active sites on the mineral surface, can
be beneficial. The role of Zwitter ions in flotation is discussed
CHEMICALSTRUCTUREOf THE SURfACTANT
by Gupta and Smith (67). A better understanding of the role
Substitution of hydrogen in CH1 or CH3 groups of the sur- of chemical constitution of the surfactant in flotation will in-
factant with fluorine increasessurface activity and flotation
deed be highly beneficial for developing tailor-made flotation
reagents with desired selective adsorption properties.
COLLECTORCONCENTRATION
eo! 18C
16C~
It can be seen from Figure 9 that flotation recovery is also
~ i
pH
Fig. 12. Flotation of calcite with dodecylammonium acetate (DDAA)
at lower coUector con~ntration. An ex~ss of coUector has and sodium dodecylsulfate (DDSO4) solutions (after
Somasunda.ranand Alar (14»). From Journal CoUoid and
been reported to reduce the flotation of minerals (45) and pre-
Interface Science 24, 439 (1967). Reproduced by permis$ion
cipitates (73). In particulate flotation this is sometimes due to of Academic PressInc.
a reduction in the size of the bubbles to such a level that they
are not capable of levitating the large number of particles that
caUect on them (45). Adsorption of a second layer of coUec-
tor (at high con~ntrations) with an orientation opposite to
that of the flfst layer, or adsorption as micelles can also de-
creaseflotation, but this is less likely since only a small frac.
tion of the surface need to be hydrophobic for flotation to
occur (see Figure 11).
SOLUTIONpH
pH is a major controlling factor in the separation of oxides,
silicates, etc. Proper choice of pH and the type of coUector is
in fact a requirement for selective flotation of one mineral
from another. The effect of pH iUustrated in Figure 12 is for
the flotation of calcite with an anionic and a cationic collector
as a function of pH at two concentrations (14). The isoelec-
tric point of calcite is in the pH range 8 to 9.5 (14). It is evi.
dent that significant flotation with an anionic collector is pos-
sible only below the isoelectric point when the particles are
positively charged, and with a cationic collector only above
the isoelectric point when the particles are negatively charged.
The pH will also influence flotation through its effect on
collector hydrolysis. A typical example of this is the amine
flotation of quartz in basic solutions. A critical pH-coUector
con~ntration curve for the incipient flotation of quartz with
dodecylammonium acetate is given in Figure 13. Quartz is
negatively charged above pH 2, and therefore it should be pos-
sible to float it with a cationic collector above this pH. It can
be seen from Figure 13 that flotation ceasesonce the pH ex-
ceeds 12. This observation was ascribed to the fact that at 12
2 4 6 8 ...0
1)0
102 I I I I I I I
-
~
Q) 80 f ~
J
:: -~
~
0
E
Z
-
10
-4
ae60
Q
IAJ
~
g40.
..J '
II
I
"
, 1I
1
Q 10 MAXIMUM :"
~
~ RNH2
20
~
cr- If . .-:-~~~--:~~-Jin(OH I' n(OH)z(s)'. (. ) -
I- _£, . ,"
..'
0.2
Z
W
lO-
/
~ .
0 2 4 pH 6 8 10 12
u I
Z I AI FiC.15, Efrector pH on zinc notation usinasodiumlaurylsulfatcat
I coUectorratiosor 0.2. I, Z.and ) from I x 10-4 molc/bter
I zinc (II) solution. In the halchedreSion,nolalion or
.j
I
lAM Zn(OH)2precipitateoccurs(ICter RubinInd upp (78)1.
... I Reprintedrrom Se'pGrgrion SC;,ltce'6 )57 (1971). ColUtesy
~. I or Marcd DekkerInc.
0" -
- (j.eo"0 8I J I
00"
I I I I I will be ion fl~tation or precipitate flotation; thus zinc is re-
10 10 12 14
moved from solution by ion flotation below pH 8 and by pre-
cipitate flotation above pH 8 (see Figure I S). It can also be
pH seen from Figure 1S that ion flotation is influenced by collec.
Fil. 14. Concenuation of neutral molecules (M) and ions (I) in a 2 tor concentration to a greater extent than precipitate flotation.
2 x 10-i mole!1 (A) and 2 x 10-4 mole!1 (8) dodecyiammo- Grieves found precipitate flotation to be most efficient when
mum chloride solution as a function or pH. Solid lines indi- the sign of the charge of the precipitate was opposite to that
cate the maximum solubility (after Gaudin l76»). From of the collector and when the amount of soluble specieswas at
110141;011,2nd Edit., by A. M. Gaudin (1957). Used with per-
a minimum (1). Microflotation of organisms is also reported
mission or McGraw-Hill Book Co.
to be highly sensitive to pH (79. 80); thus. flotation of E. coli
collectors when present along with ionic surfactant species. using lauric acid and alcohol is maximum in the pH range of 4
This is supported by the observation that maximum flotation to 8 (79).
of hematite with oleate was obtained in the PoHrange in which IONIC STRENGTH
half of the oleate was present in the neutral form (57). This If the adsorption of the collector on the particles is pri.
was attributed to the significantly higher surface activity of marily due to electrostatic attraction of the coUector species
the I : I acid-soap complex formed in this pH range and its ad. to the charged mineral surface. a significant increase in ionic
sorption at various interfaces. Fuerstenau and Yamada (63) strength will decreaseits adsorption on solids becausethe ad-
obtained an increase in flotation of alumina with dodecylsul- sorption of the collector ions must take place in competition
fate upon the addition of dodecylalcohol. They ascribed the with other ions that are similarly charged. This effect is
flotation increase to an increase in the total adsorption of the clearly shown by the results obtained recently for the cat-
surfactant speciesat the solid/liquid interface facilitated by ionic flotation ofquartl (46) (see Figure 16). In this case.
the screening of the repulsion among the ionic parts of the potassium nitrate, added to increase ionic strength. acu as a
adsorbed surfactant speciesby the neutral alcohol molecules depressant. An increase in ionic strength by the addition of
coadsorbed between. There is, however, no direct experi- uni-univalent electrolytes has in general been found to be det-
mental evidence available yet to support such a hypothesis. rimental to froth flotation. An exception has been noted reo
There has been observed an increase in flotation of alumina, cently for hematite flotation using oleate (81). In this case,
but not hematite, owing to the dissolution of methane and an increase in KNO) concentration from 8 X 10-5 to 2 X 10-1
butane in sodium dodecylsulfonate solutions (76, 77). The mole/liter was found to increase flotation from about 28 to
reason for this selective activation is not yet established. 9O'roat an oleate concentration of 1.5 X 10-5 mole/liter. Con-
The effect of pH on ion flotation and precipitate flotation ditioning was carried out at 2SoC and pH 8.0. When the con-
has been discussedin an earlier publication (1). The selected ditioning was carried out at 95°C. the effect of an increase in
pH of the solution may in fact determine whether the process ionic strength was. however. found to be detrimental. If the
~
No. 150.Vol. 71 SOMASUNDARAN ii
100
80
If
0601
w 1
too
940l
~
1
1.5 .
1O-4~.OOA
QUARTZ
20 pH- 5.8
0
K5 K5 K53 Kj"- K)-t
K~3 CONCENTRATION, mole/ liter
.eOr
6 NozSO.
80 > I
e +40 ~
- I
~ I
- 60 ~ :.
~
Q ~ ,
!AI Z '
~ IAI 0
«
... I -~~
0 40 ~ r
..J
I&. . I ALUa8IA
20 4
ALUMINA
-s
. 10 M RSO. ~
... -40
IAl
N
r-
.
.
0
0
NoC1.
NoZSO.
RSO.No
pH6 -80r pH 6.5
0
-7
I
-6
I
-S
I
-.
I
-3
I
-I -1
i
-1
'.
-&
.
-s
.
-.
."
-s
.1
-Z -I
10 10 10 10 10 10 10 10 K) K) K) K) K) K)
ELECTROLYTE CONCENTRATION. mole/ life'
CONCENTRATION OF ADDED SALT, mole/liter
Fic. 17. Thedepression of notation of aluminaby NaCIand NatSO. Fil. 19. Effect of addition of NaCl.Na1S0.. and Na dodecylsulfo-
with sodiumdodecybulfatcIS collectorIt pH 6 [after Modi nateon the ~etl.potentialof aluminaat pH 6.S (after Aplan
and Fuerstenau (82)). Reproducedby permissionor Amer- and Fuerstenau(81)J. Reproducedby permissionof Amer-
ican Inst. Min. Met. Engrl. (AIME). ican 1nst.Min. Met. En,q. (AlME).
ADVANCESIN INTERFACIAL PHENOMENA AICbE SYMPOSIUM SERIES
80l
~ ~
- t~
60
~
-C 40
r
,
0 r-
~ ~
20l
0
2 4 6 8 10
pH
Fig. 21. Effect of additionof alumon the flotation of illite UDnIso-
dium laurylsulfateascollector. . -no alum;. -6.25 lng/I
alum;e-50 mg/l alum (after Rubinand Erickson(92}J. Re-
printed from Wale'Res.5437 (1911)by Rubinand Erickson.
Reproducedby permissionof Pergamon Pressinc.
The effect of ionic strength on precipitate flotation can also meric type reagentssuchasstarch. Figure22 illustratesthe ef.
be expected to be of a similar nature. Sheiham and Pinfold re- fect of starchon the flotation of calcite usingoleate(65). It
port a decreasewith increasing ionic strength in the precipitate canbe seenthat the addition of evena smallamountof starch
flotation of strontium using dodecylpyridinium chloride. hexa- decreases the flotation of calcite drastically. As mentioned
decyltrimethylammonium chloride, and a dialkylammonium earlier,starchdoesnot reduceflotation by inhibiting the ad.
chloride (87). As expected, they did not observe such an ef- sorption of the surfactanton calcite particles. In fact, the ad.
fect in the precipitate flotation of "the second kind" of nickel sorption of oleate on calcite was found to be higher in the
and palladium with nioxime (88,89) becausethe collector at- presenceof starchthan otherwise. Similarly, the adsorption
tachment to the colligend is a part of the precipitation process of starchwasalsoenhancedby oleate(seeFigures23 and 24).
and independent of electrostatic attraction between the col-
Thus even though the particlesadsorbedmore surfactant in
lector speciesand the precipitate. Ion flotation also should be
retarded by an increasein ionic strength becausethe colligend
ions will now have to compete with other inorganic ions for at-
tachment to collector ions. Rubin et al. (78, 90) and Grieves
(9 J) have observed the interference of ionic strength to the
ion flotation of zinc, copper, and orthosphosphate, respec-
tively. Indeed, in ion flotation and other foam separation
techniques, the effect of ionic strength on foam stability and
drainage can be a governing factor. The foam separation of
~
0
surfactants is assistedby an increase in ionic strength because
of the enhancement of their adsorption on the bubbles, at 0
~
least below the critical micelle concentration. ..-
<{
0
FLOCCULANTSAND DISPERSANTS .oJ
~
Complete flotation of a material can be achieved in several
casesby adding auxiliary" reagents. For example, Rubin et at.
found that the flotation of illite, titania, and B. cerew with
sodium laurylsulfate was increased to almost 100% by adding
alum (92-94). Figure 21 shows the typical effect of alum on
microflotation. It is not clear whether the increased flotation
is due to charge reversal of the coUigend particles or simply 0
0 4.5 9 13.5 18
due to their flocculation. In the froth flotation of minerals,
CONCENTRATION OF STARCH, ppm
flocculation is valuable only if it is selective. Such selective
flocculation followed by flotation has been found to be useful Fig. 22. The depression of calcite notation using sodium oleate by
for the large-scaleseparation of fine hematite from quartz (95). starch (after Sornasundaran(6.5»). From Journal Conoid and
Interface ScienceJ 1 (4). 19. Reproduced by permission of
Separation by flotation is affected by the addition of poly- Academic PressInc.
No. ISO. Vol. 71 SOMASUNDARAN 13
~
14 ADVANCESIN INTERFACIAL PHENOMENA AIChE SYMroSIUM SERIES
techniques can very weU be the causefor frequent misinterpre- 26. Parks. G A.. Ch~1PI. R~p., 65.177 (1965).
tation, particularly when one is dealing with heterogeneous, 27. D~ju, R. A., and R. B. Bhappu. "A Chcmiallnt~rprctation of
Surface Phenomena in Sitiate Minerals:' New M~~iQ) State
natural minerals. Electrokinetic experiments provide only an Bureau of Mines and Mineral ResourcesCirc. No. 89 (1966).
averagemeasurement for the whole solid particle. Further- 28. Somasunduan, P., in Cle.n Swfaces Th~;, P,~panrion.nd
more, the measured property depends upon the type of pre- o..,.cterlurion 10' /nrul.c;.1 Srudin, Matcd Dekker,
treatment and storage that the mineral has received. Spectro- p. 285-306 (1970).
29. Smith. R. W..and N. Trivedi, Tnns. A/ME. 156.69 (1974).
scopic techniques can also causemisinterpretation if it is not
30. Puks,G. A.,A~. Min~ra/0Iist,51. 1163 (1972).
recognized that the surface speciescan undergo significant 31. Modi, H. J.. and 0 W. Fuerstenau.J. Phys. CIr~m.. 61, 640
changesduring the preparation of the sample for the analysis. (1951).
E~n today, flotation practice leans rather more on expe- 32. Dobias. B.. Spurny, and E. Freudlova. Coll~ction Cl~ch. Ch~m.
rience than on a meaningful scientific undentanding. It ap- COnllnU1l..24, 3668 (1951).
pean that the effects and interactions of a large number of 33. Johansen. P. G.. and A. S. Buchanen.Austr'8l. J. CIr~m.. 10.3980
(1957).
variables, including the chemical constituents present in a nat-
34. $<:huylenborgh, J., and A. M. H. Sanaa, R~c. Trap. Clrim.. 68,
ural flotation system, need to be recognized. studied. and un- 999 (1949).
derstood if the knowledge gained from researchon the physical 35. Ku1kuni. R. D., and P. Somasundaran,paper presented at 164th
chemistry of the mineral-solution-gas system is to be of any meetina of A.m. Chem Soc., New York (1972).
significant application in actual froth flotation practice. 36. Somasundaran.P., and R. D. Kulkarni, J. Colloid /nterf.~ Sri.,
45.591 (1973).
37. Li. H. C., Sc D. thesis. Mass. Inst. Tech..C.mbridp (1958); Li.
H. C., and P. L. d~Bruyn. Swfac~ Sci., 5. 203 (1966).
LITERATURE CITED
38. Zucker, G. L., D.E.$<:.thesis, Columbia Univ.. New York (1959).
1. Somasundaran,P., S~poralion ond Purifico/ion M~llrod.r. I, 117 39 Cited in Joy. A. S.. and R. M. ~nscr, Trans. /MM. 15, Cl 94
(1972). (1966).
2. ui. R~. w.. ~nd D. w. Fucr~tenau. Trans. AI."E. 241. S49 40. Haydon. D. A.. and F. H. Taylor. hoc. 1~ /nt~m. COlli". ofS".
(19681; See~I~o R~8h~van.S.. and D. W. Fuentenau,AIClrE fac~Activity,Vol. 1.p.157(19601.
Symp. ~,. No.. 71 (197S). 41. Fuerslcnau. D. W.. in Th~ CIr~mistry of Bio~f«es, M. Hail.
3. Sh.h, G. N.. and R. Lcmlich,lnd En,.. Ch~m. FuIIdG~nlols, 9. (ed.).Vol. l,p. 143(1971).
35011970). 42. Lin, I. J., and P. Somasundaran.J. Colloid /nlerf.~ Sri., 31.731
4. Sheibm. I.. and T. A. Pmfold. ~pt1rvlion Sa., 7.43 (1972). (19711.
S. Goldber8. ~.. and E. Rubin. Ind. En,.. "'~m. houu D~sip 43. Somasundaran.P., and D. W. Fuerstenau, Trans. A1ME.152. 275
D~II~/op..6,195(1967). (1972).
6. Pinfold. T. A.. ~pGlV/ion Sri.. 5. 379 (1970). 44. Wakamatsu. T.. and D. W. Fuerstenau. in AdsorpfiOfl from So1&l-
7. Gutierrez. C..."in~rvISa. En,., 5,108 (19731. lion. Adv. Chern. ser.. No 79.161 (1968).
S. GrAnville. A., N. P Finkelstein. ~nd S. A. Allison. r,ons. IMM. 81, 45 So~sundaran, P., and R. D Ku1kuni. Trans. /MM (London). 82.
Cl (1972). C163 (1973).
9. Finkelstein. N P..S A. Allison. and V. M. ~veU.AIClrESymp. 46. Somasundaran.P.. Trans. A/ME. 255. 64 (1974).
S~,. No. 150.71,165 (197S). 47. Onoda, G. Y.. and D. W. Fuerstenau, 711r/nttn!. Mineral PI'oceu.
10. Fuenten~u. D. W., T. W. Healey, and P Somasunduan, rrvns. inl Con"., Vol. /, p. 301, Gordon and Bteach. New York
AJ."£. 229. 321 (1964). (1964).
II. Somasundaran.P.. T. W. Healy. and D. W. Fuer~tenau.J. Pltys. 48. HoPSlock, D. M.. and G. E. Apr, Tram. A/ME, 241. 466 (1968).
Ch~m..68,3562(1964). 49. Somasundaran,P.. PhD. thesis. Univ. California. Berkeley (1964).
12. Somasundaran.P..~nd D. W Fuerstenau.ibid.. 70.90 (1966). 50. Fuerslenau, D. W.. and M. C. Fuerslenau. Trans.A/ME. 204. 302
13. Yopp~, J. A.. and D. W. Fucntenau,J. Colloid Sa.. 19.61 (1964). (1956).
14. Somasund~ran,P., ~nd G. E. Agar,J. Colloid In/~'foc~ Sci.. 24. 51. Rogers, J.. and J. H. Schulman, "Proc. Secood Intern. Conar. of
433 (1967). Surflce Acuvity:' Vol. 3. p. 243, Butterworth;, London
IS. Somasundaran.P..ibid., 27.659 (1968). (1957).
16. Somasund~ran.P.,and G. E. Alar, ibid.. 252. 348 (1972). 52. Peck. A.. S.. and M. E. Wadsworth, Proc. 711r/nl~'n. Mineral Pro-
17. KulkarnI, R. D..~nd P. Sornasundaran,in "Oxide- Electrolyte cessi", Con, p. 259. Gordon and Breach. New York (1965>-
Interfaces:' Am. Electrochem. Soc., 31-44 (1972). 53. Peck. A. S., L. H. Raby. and M. E. Wadsworth, T,.ns. A/ME. 238,
18. Gaudin. A. M.. D. W. Fuentenau. r,ons. AIM£. 202. 66 (19S5). 301 (1966).
19. Iwasaki. 1., R. R. B. Cooke,~nd H. D Choi.ibid..120. 394 54. Kulkarni, R. D., and P. Somasundaran.A/ChE Symp. ~r. No.
(1961). 150,71 (1975)
20. Berube. Y. G.. and P. L deBruyn.J. Colloid In/~'fo« Sri., 27. 55. Sonnsen. E..J Colloid /nt~'faa Sci.. 45. 601 (1973).
305 (1968). 56. Somasundaran, P.. Trans. A/ME. 241.105. (1968).
21. Robinson. M.. J. A. Pask..and D. W. Fuerstenau. 1. A~'. C~rvmic 57. Kulkarni. R. D.. and P. Somasundaran.to be published.
Soc., 47, 516 (1964). 58. Wada.M.. paper presentedat the IV Intern. Auft)ereil..KoUoquium,
22. Saleeb. F. Z. and P L deBruyn. El~ctrocrnol Ch~m. fnl"foc. Forsch. Aufbereit.. Freiberl (Sachs). (1966).
E/~ctroclr~m., 37.99 (19721. 59. Cooke. S. R. B.. in Advanc~s in Colloid Sci~nc~,Vol. III, H. Mark,
23. Buchanan. A. S..and E. Heym~nn.Proc. Ro}'oISoc.. A195.1S0 and E. J. W. Verwey, (ed.), p. 326, Interscience, New York
(1948). (1950).
24. Overb~ek. J. Th. G., in Colloid Sci~nc~. H. R. Kruyt. (ed.), Vol. 1. 60. Schulman, J. H.. and J. Leja, KolIoid.Z.. 136. 107 (1954).
ElseVIer.New York (19~2). 61. Leja, J.,Proc. 2nd/nt~'n Con".. Swfac~AclilliIY. London, 3,
25. Freyberger. W. L. aad P. L deBruyn.J. PItyr. Chtm.. 63,1475 27) (1957).
(19S71. 62. Schulman. J. H.. and J. Leja, inSwf.cePltenomeM i" Ch~mistry
No. ISO,Vol. 71 SOMASUNDARAN IS
.1Id BkJIoo, J. F. ~W, et &1.(ed.), p. 236, Petpmon Press, 98. Grie9Cs.R. B.. and R. K. Wood. AIChE J., 10.456 (1%4);
New York (19S8). Grieves,G. B., and D. BMttacharya. J. A"'. Oil Ch~",., 41.114
63. Fuerstenau.
(1962). D. W., and B.
. J. Yamlda. r,.lIs. A/ME, 213, SO (1965); Grieves. R. B., and D. BhattacMrya,J. WGt~rPollut.
Co/lrrol FN.. )7.980 (1965).
64. Fuersteuu. M. C.. "The Ro~ of Metal Ion Hydrolysis ia Oxide 99. Mune. E. J.. and T. A. Pinfold.J. Appi. Ch~",.. 11.52 (1968).
and S&licateFlotation Systems:' A/ChE Sy"'p. Ser. No. 150. 100. -,ibid.. 140.
71,00(1975).
6S. Somasundann, P..I. Colloid /1I11r{4a Xi., 31, 557 (1969).
66. Yen, W. T.. R. S. OIaJaal,aM T. Salman.C-.d. Met. Q t.,12. DISCUSSION
131 (1973).
Emesto Valdes-Krei&: How do collector ratios studied in a
67. Smith, R. W., and S. C. Gupta, A/OlE Sy1/lP.Ser. No. /50,71,94
(1975). batch ~ll reprodu~ in a continuous system? We have ob-
68. Rubin. A. J.. J. D. Johnson. and J. C. Lamb, III. /Ild. EIII. ae".. served that true concentration in a batch cell are often not re-
Pro«ll. DesiIII Denlop., 5,368 (1966). corded, but rather their initial values are given. The adsorp-
69. Schoen.H. M.. E. Rubin,andD. Ghosh,I. Wallr Pollulioll Coli' tion capacity of the mineral and the continuous aeration may
tIOl Fed., 34,1026 (1962).
modify the conclusion derived.
70. Grieves,R. B., ibid.. 41, R336 (1970).
71. Schncf', R. W., E. L Gaden, E. Mirocnik, and E. SchonfekS,Ore".. P. Somasundann: PhysiQI chemistry of froth flotation has
E/lf. Pro"., 55, 42 (19S9). been studied in the past using batch tests with no correspond-
72. Robertson, G. H., and T. Vermeulen. "Foam Fractionation of ing continuous tests for comparison. The best comparison
Ru.wth Elemeau:' Lawren~ Radiation Lab. Report, that is possible is between laboratory batch tesU and large-
UCRL.19525 (1969). scale plant continuous tests. The ratio of collector to col-
73. Grieves, R. B., W. T. Suanp, and D. Bhattacharya. A/OrE Sy",p.
Ser. No. 150,71,40 (197S).
ligend is in general found to be rugher under p~t conditions
74. Gaudin. A. M.. and F. W. Bioecher. Jr.. r,./IJ. AlME, 187,499 than under the laboratory test conditions. For example, while
(1950). an oleate con~ntration of only S X Iv-4 mole!1 was needed
7S. Somasundaran,P., and D. W. Fuentenau, 7)v/U. AIME,141,102 to obtain nearly complete flotation of hematite at pH 1 to 8 in
(1968). a laboratory batch test using Denver ~ll, a ooncentration of
76. Gaudin, A. M., "Flotation." 2nd Edit.. p. 262. Mc-Craw Hill. New
S X 10-3 mole!1 is reportedly used in the plant during oondi-
York (1957).
77. Sornasundaran.P., and B. M. Moudli1./. Colloid /lIlerfaa Sa.. tioning for the notation of the same ore. We found concentra-
47,290 (1974). tion for similar flotation. but of a different type of hematite,
78. Rubin. A. J..and W. L. Lapp, SeparaliO/lSa.. 6, 357 (1971). in a HaIlimond ~ll to be S X 10-5 mole!l. The need for in-
79. Rubin, A. J.. E. A. Cassel.O. Hendenon. J. D. Johnson. and creaseddosageunder a larger-scalecondition might be due to
J.C. Lamb,fll,Biotech. E/If., 8,f31 (1966).
80. Rubin, A. J.. ibid.. 10,89 (1968). the presenceof various additional colligends (that compete for
81. Sornasundaran.P., and R. O. Kulkarni. paper presented at the the co~ctor), particularly slimes that might often be present
103rd Ann. AWE Mcctinl, Dallas (1974). in higher amounts in a large-scaleoperation.
82. Modi, H. J., and D. W. Fuerstenau, TraIlS.A/ME. 217, 381 (1960). Flotation results are usually reported in tenns of the initial
83. Aplan,F. F.. and D. W. Fuerstenau,
Froth Flotation, 50th All- .
surfactant con~ntration rather than the equilibrium bulk
niversaryvolumeAfME, P. 182(1962).
84. Fuerstenau, M. C., D. A. Ri~, P. Sornasundaran.and D. W. concentration. In an ordinary laboratory flotation test the
Fuerstenau. r,./IJ. /MM. 73, 381(1965). equilibrium bulk concentration will be lower than the initiai
8S. Fuerstenau, M. C.. C. C. Martin, andR. B. Bhappu,TrailS.A/ME, concentration and such differences could indeed affect the in-
126,449 (1963). terpretation to some extent. Reported effects of such vari-
86. Fuerstenau,D. W..~ Appi. Ore",.. 24,135 (1970).
87. ShciIwII, f.. andT. A. Pinfold,I. Appi. Orem., 18,217 (1968). ables as solution pH and interpretations based on them, how-
88. Nahnc,E. J., andT. A. PinfokS.tbld.,19,57 (1969). ever, can be expected to be in seneral valid for a given flotation
89. -. Che",.llId., 1299(1966). system. Moreover, in the caseof HaUimond tube tests (that
90. Rubin.A. J., J. D. Johnson.and J. C. Lamb,ill, /Ild. E~. ae".. have been widely used to study the physical chemistry of flo-
ProceuDesip DeHlop.. 5,368 (1966). tation), surface area of the solid available for surfactant de-
91. Grieves,R. B., in "AdsorptiYeBubbleSeparationTechniques,-
R. ~mlich (ed.),p. 175,AcademicPress,New York (1972). pletion by adsorption is very small, and hence equilibriwn
92. Rubin. A. J.. and S. F. Erickson,WGIIrRes..5, 437 (1971). surfactant concentration will be very close to that of the in-
93. Rubin, A. J..and D. C. Haberkost,SepaIVtion Xi., 8, 363 (1973). itial concentration. It might be noted that results obtained for
94. Rubin, A. J.. andS. C. Lackey.I Am. WaterWorb Assoc.,60, adsorption tests during flotation research(which use rugh sur.
1156(1968). face area systems) are usually reported in tenns of equilibrium
95. E/lf. MllIiIII/.. 175(11),140 (1974).
bulk concentration. It will indeed be advisable to check the
96. Bjome.C.. and J. Keeley,Millill, EIII., 65 (1964).
97. DorenfekS. A. C.. in "Froth F1ot~tion.50th A.nni'YersMy
Vol.. equilibriwn concentration for all tests if available time per-
D. W. Fuentenau,(ed.) p. 373, AlME, NewYork (1962). mits the required chemical analysis.