Chemistry and applications of chelating agents in flotation and

flocculation

P. Somasundaran and D. R. Nagaraj

Synopsis recent approaches to understand chelation are discussed briefly.

Chelating agents, because of their metal specificity or Emphasis is given to the criteria for selecrioo of chelating agentS
selectivit)", can function as good flotaids and selective IS colleaors for various minerals and the prediCtabilit)" of the
flocculants. In addition to the metal selectivity, chelating behaviour of a chelating agent for a given mineral system.
agents offer certain advantages over conventional
mineral processing reagents even from the synthesis
point of,-je,,". With the major donor atoms-S, N, 0 and, Chelating agents
to some extent, P-numer<KJs possibilities exist to tailor-
make reagents for specific applications. For polymers the
Ch~latingagentsarecompoundsthat fonn metalcomplexes
characterizedb)' ring struaurcs ilIusuatedin Fig, I.I.~.j In type I
backbone offers funher possibilities for the
incorporation of the required propenies. th~ metalis coordinatedto th~ four nitrogcnsof two moleculesof
In general, the choice of a chelating agent or a group is ethylcn~diamine,giving rise to a d1argedcmubleringed compla
,,'ith cbIorid~nNtralizing th~ two charges,I Other typesof
made on the basis of its function in well-known analytical
metal separations. Such a choice is limited in that the dldating oomplcxesindudc t}"peII. "ith 81 inm-molccular
number orwell-known analytical separations is also hydrogenbridge.and type III. involving polynuclearhalogen
limited. In the present work, a new and effective do"or bridges.
atom approach is discussed for the design of mineral-
specific chelating groups. This generalized approach is
focused on the chemical properties of the imponant
donors both in the reagent and on the mineral. Data from .NH2-CHz
many chemical S<KJrceshave been summarized and .W
I
Cf2
applied to systems of relevance to mineral processing.
'NH2-C"z
Understanding of the chemical behaviour of donors is of
param<KJnt imponance in predicting the propenies of
chelating groups.
Numerous investigations have been made of the H3C\
application of chelating agents in mineral processing. I'c~o ..
The various fundamental aspects of the interactions HC H (D)
between a mineral and a chelating group are, however, " I
not clearly understood, although there is evidence to ~-O
suggest that some form of a metal complex is formed at H,C
the surface.

The separation of mineraJsfrom one another by flotation and

floccu1ation depends mainly on the sdeaive 8dsorption of .CI-
surfactants and polymers on them. Although many reagents have ...:,
~ng hem used in practi~ for the flotllion of mioera1s.most arc (ID)
not as selecti\'e E is required for the efficient separation, in
particular, of mineral fmes and ultrafmes. Reagents with ,ci.
chelating functional groups have received increased attention for
this purpose in recent years, being known to exhibit exceUent B B
metal selectirity in analytical separllions. MInY chelating agents FiB- I GeIxr8I sUUCtUresof clr1ating -cents I
have been tried E coUeCton for various mineral systems and
excellent seper8tions have been obtained in cenain czcs. The
mechanism~ that involve their colleCting action arc, however, not Chd~ng agentscm be cl~sirlCdon the b85isof donor atoms
adequate~' understood. In particular, the differences in their involvrd (0-0, N-O, N-N, 5-0, 5-5, 5-N)or ring size(4-,5-
mode of action at the mineral surface venus that in the bulk are or 6-memberrd)or chlrg~ on th~ complex(Inion, cation,
oot fully ~ and, as a resuh, the dcve~pment of neutral)or numberof bondsto the meuI for everydteJating
collecton based on their use in analytical separations is not molecule(bidentat~I: 1 /A;
ach~'ed easily. It is to be noted in this regard tha chelating M )
coUecton are seldom ~tal-specific or mineral-specific and the 'A
propenies of both the chelating agents and the mineral are
imponant in determining their colleCting action. The dIXlor
atoms on the chelating agents, as well as those associatedwith the
bidentatc1:2 ( A 'M/A) ; terdentBte A
mineral species, playa governing role in their intenctions on the
A/ 'A M~)
surface. In this paper the application of chelating agents in
mineral processing systems is re\'iewed on the blSis of the donor
'A)
properties of the c1'w:latinggroups and the metal species, and Specificityof chelatingagentsdependson the int~on

209
 Weal ckIDOratoUD Ind sunKe species on the panicle. A list of S.illa" UmuuDble
/lOr atoms Ind funaional groups a>ntlining the four major
lIDS(N, O. P and $) are given in Table 1. Halogen atoms, a. HO-CH-COOH CH2-COOH
I I
and I, panicipate in chelate ring formation in bridged HO-CH-COOH CH2-COOH

Table 1 Donor atomsand functional groups containing

major donor atoms)

H
OH OH
rI N
t:
---"'t

0 : I_!_-
I
, P s
~
r a: H~H rt:JfH
, I Br:
I
IAs ~
L__- Neutral complexes arr usuallyinsolublr in wltrr and.
Sb Tt Ii-.L_-J
I I I
I
therefore,promotr hydrophobicit},.For insolubilit}, in wltrr,
howr\"er,it is not sufrlcimt that thr chrlltr br a neutral complex.
A classicrxamplr is that of chrlatesof dimethylglyoximes
(DMG) (5« Fig. 2). The 1:2complrx of nickelwith DMG is
insolublr in wltrr. Thr coppercomplrx is, on the othrr hand,
rdai,'rly ~'.rr-soluble. This resultsfrom the subtlr differences
in thc shapesof the tWoIOOlecutcS and consequmtdiffrrmces in
-NH, the IOOdrof packingof thc tWomolecuJes in their resprctivr
0 -rooM =0 crystals(Fig. 2). The nickel compoundis planarand the
-OH (mol 01' phenol) -08 (ajQ)1XII5 molrculesarr stackrdin thr crystal,sothr Ni atomsarr co-linrar
-P(O:.(OHh md havr a wrak bond brtWrcn them. In the coppercompound,
s -5H howr\.er,thc chelaterings arenot co-planarbut makea small
-S-R
angleof 28° with onr anothrr andthere is no metal-metal
bonding.I The tetragonalpyramidalcopprr chrll(r molrculesarr
pairrd in thr crystalin sucha waythat ram copperatomhz an
polynuclearcomplexesthat an of minimal imponancein 0 atomirom itSIdjKent molrcuk a onr of its fivr immediate
mineralprocessing.As and Sefonn only a few usefulcomplexes neighboun, In thr Ni complexthr -OH md thr -0- are
and.tberef~, an seldomused. strOnlly internallyhydrogenbondrd Ind, bmce, lessreadily
sol"atrd. The coppercomplrx aCtUallyhasa highrr stability
Requirements for c:helatinl8JeDU constantthanthe Ni complexbut, sincr no intrrnal hYdrogrn
Thcr~ Ire. essentially, tWObasic:requir~mcnts that c:h~lating bondingis po5siblr,thr complexis easilysolvatrd.
agents must satisfy to form metal cbclates: the molecules should
hav~ suitabl~ functional groups and the functional groups must
be situated to permit the formation of a ring with a mctaI a the
closing m~mber.
These t"-o conditions are ncces~', but not sufficient for the
formation of a d1elat~ ring. For aample, under ccnain
conditions and under sufficiently to.. pH valum, a potential
chelating molecule may attach itself to a mcta1atom through only
on~ ligand atom:

H ooc. CH zNH z
"/ ,0

a
" Pt

NH~2COOH
or

HOOC.CH:S-M
I t hE been generaUyobserved that certain Iddioonal
requirements must be satisfied b}. chelating agents for mineral
processing applications: the chelates should, preferably. be
neutral, if the chelating agent is to function IS a collector, and the
dlelates should, preferably, be charged and very hydrophilic if
the chelating agent is to function as a depressant. On a pol}'Dleric
depressant or a flocculant this require~t is not stringent so
long IS the backbone hE other hydrophilic groups (seebelow).

Rcquirmtcuu for dr.pressanu:ionic chr.lMes,r..g.

lCo(DMG)~HzO~'. and bydrophilizinggroupson moJecuJr.

210
Specificity and selectivity
Most of the chelatinggroups form complexeswith almostall of
the transition and many non-transition metalsand, therefore,
specificity is not asabsoluteasis required for their selective
adsorptionon minerals.In practiceselectivity is achievedby
making useof differencesin stability constantsand chelate
formation under various solution conditions. As wasindicated
earlier, contribution from the donorsasthey arelocatedin the
minerallanice hasto be consideredin obtaining the selectivity.
Also, the solubility of the mineral, in addition to that of the
metal chelate,hasa pronouncedinfluenceon the selectivityand
the collectionpower of the chelatingagent.
Stability of complexes
_\1ajorfactorsthat determinethe stabilit~.constantare,in order
of decreasingimponance, pK. of the ligand molecule,
substin:ents,natureof donor atoms,central metalatom,and
ring size,number of rings.
The fIrStfour factorsarevery imponant in that they
determi~ the natureand strengthof bonds.pi(. of the che~
agenthasa direct influenceon the chelateformation sinceit
representsthe tendenC)'of the donor atomsto donateelectronsto
metalatoms(or acceptfrom ~a1s, in cenain cases)andthus
Ionn a chelate.Substitution on the chelatingmoleculehastWo
imponant effects:J it altersthep?<.and/orit introducessteric
Iactorsfor the formationof a c~late.
The influenceofthc natureof donor atom and the central
metal atomshouldbe consideredtogetherbecauseof the inter-
rclatio:nshipbetweentheir behaviour(seelater).
Table 2 Classificationof acceptorsbasedon unidentateligands
Mg
Sc Ti V ' G~
D CLASS
A Form stable
~ BORDER REGION
D Class B: Form stable
Properties of acceptors
Acceptorscanbe divided convenientlyinto two maingroups
basedon the unidentateligands:(a) thosewhidl form most
stablecomplexeswith N, 0 and F, Illd (b) thosewhidl form
moststabk complexeswith P, 5, and a (fable 2).
The stability constantsof metalcomplexesthat beJongto the
two groupsfollow the generalordersN»P>As>5b; 0>5;
F>CI>Br>I; Illd for similar liglllds N>O>F; differencesin the
seriesP>5>CI arefound to be lesspronounced.4It is seenthat a
vastmajority of elementsbelongto group (a), i.e. they
preferentiallycomplexwith 0- and N-containing liglllds.
A given metal, funhermore, may belongto group (a) or (b),
dependingon its oxidationstate:for example,Cu(I) belongsto
group (b), wherasCU(II)belongsto the overlap regionof (a)
and (b).
A major shoncomingof the abovegrouping is that it is based
on unidmtatt liglllds; the situation is complicatedwhenchelates
(or bidentates)areinvolved. Considerationmust be given in this
caseto chelatingagentsin which the donor atomsarethe same
(examplesareethylenediamine,oxalateor thiooxalate)Illd
chelatingagentsin which the donor atomsaredifferent
(examplesareo-aminothiophenoland o-arninophenol).For
nickel, tbiooxalateion is Believedto form a more stablenickel
chelatethlll the oxalateion. For zinc, stability constlllts of the 0-
aminophenoland o-aminothiophenolcomplexesshowthat the
substitutionof 5 for 0 doesha\'~a markedinfluenceon stability.
A moreusefulgroupingis that basedon the electronic
configurationof the metals.Thus metalscanbe classifiedinto
threegroups-those which contain (a) inen-gastype
configuration, (b) panially filled d or f orbitals and (c) ions of
metals~;th filled d orbitals (Ag., Zn2+,etc).
Metalsof group (a} reactprt:f~rentia1Jy ~;th O-containing
ligands,suchascarboxylateions and th~ anionsof quinalizarin,
Illd the p-diketones.
I ';:V/ ',"
8/1';( NO F
Ir,'
",
,
complexes with K,O,F
complexes with P.S.Cl
211
 0 a O' O' coUccton on their min~ral surfaces.7
Similarly, S donorsreact
pr~ferentiallywith sulphid~min~ra1s.It hasbeenpoint~dout
1f~
J.
~- ..,..C that a bond betweenm~tal ions and soft basesmakesth~ metal
ions soft acidsand vice v~na.8This is ~speciallyimportant to
bord~rlin~ cations.From this viewpoint sulphidizationof
oxidized min~ralsshould leadto an incr~asein the capacityof
th~ir reactiv~centresto reactwith sulphydryl coUccton.7Thus,
Transition-metalionsform morestablecomplexeswith th~ imponant contribution of donon on th~ min~ralsurfacesto
ligandsthat containpolarizableponions, suchasaminogroups int~ractionsbetweenmineralsand ch~latingag~ntsmust be
andheterocyclicnitrogenatoms. noted.Oxjd~ min~rals(which have0 donon) of soft-acid metal
Metals in group(c) alsoprefer highly polarizableligands, ions would preferentiallyadsorb0- and N-containing
especiallyif the latter havesuitablevacantorbitals into which collccton, wh~reasth~ir correspondingsulphideswould
someofthc d electronsfrom metalscanhc 'back-bonded'. pref~rentiallyadsorbS-containingcollectors.
Examplesof suchligandsare thosewhich contain S asthe donor
which is highly polarizableand basvacant3dorbitals.Thus,
Properties of donor atoms
thioanalideforms insoluble complexeswith Cu, Ag, Cd, Hg(ll), The donors o and N belong to the first row of elements in the
11, Sn(ll), Pb, As, Sb, Bi, Pt and Pd.
periodic table and follow the octet rule'-4,fo9"o,'1with valencies 2
and 3. The donors P and S belong to the second row, and these
0 SH
also follow the octet rule with valencies 3 and 2; however, P and
II I
~""I/::"T N H - C - CH 2
S can exhibit higher valence states also since they have easily
accessible vacant 3d orbitals. It is the electrons at or near the
~...J,,~ Thioanalide
surface of atoms and ions that are the most important in
determining their chemical andph)osical properties. This is
A reagent, such as dithizone, which oontains N and S dono~
especially true of donors P and S and transition metal ions. The
therefore reactSwith a large part of the transition metal and filled
laner have incompletdy filled d orbitals, which have quite large
d orbital series.To achieve the degree of selectivity that is
fractions of their tOtal volumes near the outsides of the ions, so
desirable in praCtical applications it is usually necesSar).to arrive
they are easily accessible for bond formation. Additionally, the
at the appropriate oombinations of dono~.
J orbitals are much more easily polarized, promoting a much
more favourable orbital overlapping.
Donor-acceptor relationships The important properties of the four donors 0, N, S and Pare
A useful classification of acceptors as well as donors is that based
summarized in Table 4.
on Pearson's concept of hard and soft acids and bas6 (rable
o and N have 2p electrons and no accessiblevacant d orbitals.
3).S.6.7
The acids (or electrophiles) are mostly (metal) cations.
S and P have 3p electrons in their outermost shells, but they
This group also includes such electrophiles as CO2' 0, a and N
have, in addition, easily acce:ssiblevacant 3d orbitals. The
and metals iI\ the zero oxidation state. The bases(or
electronegativities decreasein the order
nucleophiles) are non-metal anions, neutral atoms and
molecules. It should be noted that the donors 0 and N are hard
basesand donor S and P are soft bases.
0 N s P
,3.5) 3.07} :2.44: (2.06)
Table 3 Characteristicsof bard andsoft acidsand bases Consequently, 0 invariably forms ionic bonds with a majority of
Hard Soft elements in the periodic table. Thus, chelating Igents with 0-0
(acid or baK) (acid or base) donors often form chelates with a large number of metals and
Smallsize Large size are, therefore, lessselective. The selectivit). in general should
increase from 0 to P.
Orbitals involved far apan Orbitals closein ~rgy
The normal ,'alencies are 2, 3, 2 and 3 for 0, N, S and P.
in ener~'
respectively. 0 has tWo unpaired electrons or two elearon pairs
Ionic bonding coulombic Co,"alentbonding
to donate, but it seldom donates ooth pairs; only one pair is
attraction
active. It can form a maximum of four oonds. but seldom attains
High chargeions High chargeions all four (three oonds are common). 0 forms multiple oonds, N
Non-polarizabk Polarizable has five valence electrons. but only four orbitals and, therefore, a
Baseshave10"-proton maximum of four oonds can be formed, It has one lone pair of
affmity electrons to donate when three electron-pair bonds are formed
Hard likes hard Soft likes soft Like its neighoours C and 0, N also readily forms multiple
High elecrronegativity Lo'" elcctroncgativity oonds. In this respect N differs from P. S. As, Sb and Bi, S has
rn'o valence electrons and its normal valency is 2, But it has four
orbitals and easily accessibled orbitals. Consequently, S can
The bondingcharacteristicsof thesegroupsaregivenin Table torm two to six bonds. Similarly, P forms three to six oonds,
3..ln general,hard basespreferentiallyreactwith hard acids,and although its normal valency is 3 and it has five valence electrons.
similarly for soft bases.The former interactionis characterized .0\11
four donors have one active lone pair of electrons. Although
by ionic bondingsincethe orbitals in"olved arefar apartin 0 and N show strong prr-pir bonding (becauseof the pouter
energ}',therebypromoting coulombicattraction.On the other orbitals and no d orbitals). S and P show ver)'little or no
hand,the interactionbetweensoft basesand soft acidsis tendency for P70-p1T.On the other hand, only S and P show
characterizedby covalentbondingsincethe orbitalsinvolvedare strong d1T-d1Tbonding. In addition. S and P sho,,' d1T-p1T
closein energy. bonding (or back-oonding) since they can accommodate
It is questionablewhether this sort of classificationwill be ela.~rons from metals in their vacant d orbitals, This ability for
valid for mineralspecies,althoughthere is someevidencein the back-oonding for P and S, indeed, is an important distinction
literature that the metalcationsthat prefer to reactwith 0 that puts these donors in a special class. The polarizability of the
donorsalsotend to preferentiallyadsorbO-containing lone-pair electrons on these donors follo""5 an almost rC"'erse

212
 Table 4 Major propenicsof donor atoms0, N, S and P
0 tl s p

CONFIGURATION lS22S22pl& lS22S22P3 [Nij 3S23PQ3Do [HE]3SQp3300

ELECTRONEGATIVITY 3.5 3.07 2.44 2.11i
VALENCE ELECTRONS 2 5 2 5
NORMAL VALENCY 2 3 2 3
Ii OF ORBITALS II 4 4+D /I + 0

II OF BOUDS 3 II 2-6
(VALENCY EXPAaSION)

LONE PAIRS 2 1 1 1
pw-p. STRONG STRONG POOR NONE
Ow-p. NONE NONE STRONG $TRatG
{BACK-BONDING)
POLARIZABILITY NIL GOOO STRONG GOOD

H-BONDS STRONG STRONG VERY WEAK NONE

BONDS MORE IONIC LESS IONIC COVALENT COVALENT

STERIC LOW LOW HIGH HIGH

ACCESSIBILITY

ord~r ofth~ ~lectron~gativities.

Thus, for polarizability electron density and delocalization of electrons arise from the
resonance or mesomeric effects. With alkyl amines, for example,
~«.s-p the aJI,.-y1
group exertS an inductive effect on the nitrogen, making
its unshared electron pair more available for bonding and
0 and N have the distinction of having the ability to fonn very thereby increasing the basicity.
strong H bonds, whereas S and P show little or no such
tenden~'. Since P and S h8\'e d orbitals, and large size, these H H
donors Ire much more sterically accessiblefor bond formation
than90rN. R-N:+H'~ R- N H'
S is unique in its abilit). for catenation (forming bonds with
itself); its common OCCUiTence is in the fonn of an eight- H H
membered ring. S is also knO\\"Dto fonn higher polymers. This
unique featUre of S is Ver)' important in all its interactions. In carboxylicacidsthe inductive effectof analkyl group
0 and S can enter into chelation either through the ether or decreasesthe acidity; for formic acidtbe dissociationconstantK.
thiOetber fonn R-O-R, R-S-R, or the R-OH, R-SH. The -SH = 17.7x 10-5,whereasforacenc acidK. = 1.75x 10-5-almost
group is much more acidic than -OH and is highly polarizable, onc-tenthof the valuefor formic acid
but not an effective acceptor. The R-S-R group forms
~'ramidal bonds and is an effective drr acceptor. The R-SH
sho\\'Sa strong tendent1' for unidentate bonds, and the R-S-R #0
R -+- C' {-]
sho"'S a strong tendenC}' to form chelate rings. The C=S is more
ionizable and has a greater volume than C=O. S is a better donor
~, 0
than 0 with reglrd to sharing or donating the lone pair of
electrons. Similarly, N is also a better donor than O. [K. decreases)
It can be readily seen from the foregoing discussion that each
If, on the other hand. the R group exerts Ul oppositeinductive
donor has some unique properties. The acceptors also fall into
effect owing to certain substitUents on R such IS Cl or NOl' the
certain \\'ell-recognized patterns. A judicial combination of
acid strength of the carboxylic acid incrcascs.
donors for any given acceptor, taking into account the properties
of both acceptors and donors (on mineral as well as the chelating
agent), should provide the required selectivity in minerals 0
~'
processing applications, R -+ C?'
The "\\olecullr Orbital treatment is a powerful tool in ~,
understanding the donor-acceptor interactions. In this Q
approachI '.10.12.13 the individual atoms \\'ith their nuclei are all
(K. increzes]
placed in position and all the electrons concerned in bond
formation are allotted to the various molecular orbitals, This
The electroniceffectsaremuch strongerin aromatic
treatment, however, has scarcely been used in understanding
compounds.A benzenering oouldexertooth electron-donating
the mechanism of adsorption on minerals,
andelectron-withdrawingeffects.In benzoicacidthe aromatic
Substituent effects ring exertsan electron-withdra~ingeffectand,hence,it hasa
The electron-donating or \\'ithdrawing tendencies of organic much higherK. than that of aceticacid.An alkyl group in the
groups \\'ill influence the electron densities on donor atoms and para positionon the ring exertsan electron-releasing
effect.
the pKa of the molecule. ',1'.14-1"The inductive effect results from Groups suchas-OH or -OCH} areelectron-releasing o~.jng
electronegativity differences, Much stronger changes in to resonance,in spiteof the electronegative
oxygenand the
electron-withdrawinginductiveeffectsthat the groupscanbe Simultaneouslythe groupsRO and R INH could exert an
soownto ~sess. electron-donatingefTectsincethe C+-S- is more stablethan
The effectof electron-activesubstitUentson an aromaticring -C=S. This tautomerism,togetherwith the inductive effect,
is well exemplifiedby aromatichydroxy oximes.For anelectron- would delocalizcelectronsover the entire activegroup
releasingR group,suchz -CH) or -OCH)J the acidstrength
decreases, andtih. increaseS, the releze of electrons s
destabilizingthe phenolateion (the electrondensityon N may II
alsoincrease): -o-C-'-NH. Thus,

R R
! ! s
(;;"1-
-y ~/.--o."
RI
/ RI
RO' "C NHR

H~ The dectron-density order may follow S>O>N and,

OHNOH OH NOH considering the sizes and polarizability of the donors, the steric
accessibility would follow the order S>N,O. S may also show a
SubstitucntsR = 0, NO2ttc., ha\'eoppositeeffects.The tendenC)' to bond with S on the sulphide surface. Thus, S is a
incrc~e in electrondensityon donorscan(a) enhanceb~icity ve~' active donor in the thionocarbamate molecule. 7
and(b)thereforeenhancea bond stability, but (c)it may also Substituents introduced into the molecule influence the
decreaserr acceptorability of ligand and (a')thereforedecre~e electron distribution and, therefore, produce some interesting
donor rr-bond stability, effects. Russian investigators7.\) have done some extensive StUdy
Effects(c)and (a')could be pronouncedfor donorssuchasS on substitUted thionocarbamates-especially phenyl
sincethe donor-rr bond(or dativebond) is formedbetweenthe substitUents. The. V-benzoyl thiOnocarbamate has been stUdied
~tal andthe organicreagent(in additionto the coordinativeor the most:
I-bond) ~ a result of tranSferof electronsfrom d orbitalsof metal
into vacantd orbitalsof the donor.The formationof a dative S 0
bond,whidl reinforcesthe 0 bond, is moreprobablefor mttals Ii II
in their IIjweroxidationstatesincethesewould havemore CH, CH2O- C -NH- c-<Q)
electronsthan thosein the higheroxidationstate.Another
importantrequirementis that the donor atomshouldhaveeasily The aromaticring accentuates
the eleCtroniceffectscompared
accessible vacantd orbitals(S). 0 and N, whidl do not have with an alkyl analogue
thesed orbitals,do not form a dativebond.The nitrogenin
oximes,however,forms a donor-rr bond and,therefore,oximes s 0
arean exception.This explainsthe order of the ligand field II II
stabilizationeffectsobservedfor the chelatesbetweeen CH) CHzO- C NH c -CH;
substitutedsaliC)'ialdoximes (SALO) and Cu. Ni and Co (Table
5). In addition,the 0 of the C=O exertssomeelectron-withdrawing
effect,beingstronglyelectronegative.
The delocalizationof
Table S pK. of SALO II1d ligand field stabilizations 18 electronsis morepronounced.There is a decreasein electron
dcnsi~' on S andthe pi(..
pK. CH,.SALO SALO CLSALO NOrSALO
:to.OS 11.06 10.70 10.2S 8.72 s 0
,', 'I;~~~
CH} CH20dt'= NH =Jt.'V
7.S 8.4 10.1
0.6 1.5 2.6 3.8 ~ substitutionof a phenylgroup insteadof a benzOylgroup
0.8 1.1 2.4 3.5 mayleadto a different electroniceffect,sincethe phenylgroup
caneither beeJectron-donating or electron-withdrawing.
The ligand field stabilization(LFS) increasesfor eachmetal For thiol collectors(dithiophosphates, xanthates,and
asthe pK. decreases, i.e. asthe electrondensityon the donor dithiocarbamates'the electrondensitieson tbe activeS donors
atomsdecreases. This increasein LFS is attributed to the effect dependon the inductiveeffectsin the molecules.
of substituentson the 1T-acceptor capabilityof the ligands.18
The thionocaroonates(Z-200 t).pe)arenot only excellent s
sulphidecollecton but alsootTeran interestingcasefor study.
The moleculecontainsthe three iml'Ortant donors0, N and S. RO,"
P S- RO C;5
o andN arehard basesand S is a soft base. RO/ """'S- 1(/ 'So

CH, s DTP Xanthate DTC

,
c-o-c
II

NH .CH,CH) In DTP the {\\'o RO groupsexen an electron-withdrawing

inductive effectbecauseof the higherelectronegativityof the 0
/ atoms.This will delocalizethe electronandstabilizethe anion.
CH:
Funhermore,P is moreelectropositivethan C in xanthate.The
0 andN exen an electron-withdrawinginductive effect. electrondensityon S in DTP would be decreased and,
consequently,DTP is a strongeracid,weakercollectorand more
S selectivethan xanthate.Theseelectroniceffectsareless
II pronouncedin xanthatesincethereis one lessRO group andC is
R O-E-C~NHRJ lesseleCtropositivethan P. In DTC N is lesselectronegative than

214
0 and has. highertcndenq to donateelearons. As a resultthe
6 SALICYLALDOXIME
DTC would be . strongeroollector,weakeracid andless 0 O-HYOROXY ACETOPHENONE OXIME
selectivethan DTP (DTC<X <DTP). 0 O-HYDROXY BUTYROPHENONE OXIME
0 Z-HYDROXY-5-METHOXY-BENZALDOXIME
() Z-HYDROXY-5-METHYL-ACETOPHENONE OXIME
Table 6 StructUreand water solubility of various
~ 2-HYDROXY-)-NAPHTHAL:>OXIME
hydroxyoximesl9
.
.
O-HYDROXY CYCLOHEXANONE OXIME
a-HYDROXY BENZOPHENONE OXIME
pH- 4.8
1001

p~?
? p / /
at
80 f
>-
a: ~ p
f
f
!oj

~ r 1
u
60
!oj
I
a: ""
~
z
9;
~
~40
-
0
.J
"-

20

~
-.
1-11-
165 10-4 103
INITIAL COLLECTOR CCNC::NTR~7,ON.M
Fig. 3 Flotation of c~'socolla ~'ith sa1i~'laldoximcand various
subsutut~doxi~I'

benzene ring and the slower kinetics of adsorption of OHBePO,

A recentdetailedstudy \\ith severalwater-solublechelaIing
agents(fable 6) of the classof aromatichydroxy oxi~s showed
5e\'eralinterestingeffeCtSof substitution.Flotation results
obtainedwith thesechemicals,which canbe representedb). the
generalformula
RI
aregiven in Fig. 3. IfR~=H andR.=-H, -CH)or -CHzCHJ
CH}) the effea on the oolle-.""tor
efficient)' wasobservedto
increasein the order -H«-CH)<-CH~CH:CH,.
The introductionofth~ fl.'"St-CH2 in SALO resultsin a
much largerchangein polari~' of the moleculein comparison
with funher additionoft\\.o -CH~ groups.This is reflectedalso
in the water-solubilit).of the oximes-SALO, OHAPO and
OH-BuPO. Funhermore,the electron-donatingtendenC)'of
-CH: group mayalsofavour chelationr~ction.
Substitutionof a phenyl group on R. CRz=H)will makethe
tDOlecule much lesspolar than SALO, asreflectedin the ~'ater-
solubili~.. On this basisOHBePOcanbe predictedto havea
highercollectorefficien~.than SALO. The actualfmding ,,,'as,
ho~'ever,cont~. to the prediction,suggestedasbeingdue to
tbe possiblesterichindranceto chelationofferedby the second
Adsorption tests, ho\\'ever, snowed the adsorption of OHBePO
on dlrysocolla to be significantly higher than that of SALO at
the same oxime concentration, thereby suggesting that Steric
hindrance or slow chelation kinetics m~' not be responsible for
the lo~'er collector efficienC}' of OHBePO than that of SALO,
Bubble size reduCtion observed during flotation may be the
major faCtor in this case,
Substitution in R2 by -CHJ(R,=CH; or H, respeCtively) not
only decreasedthe \\'ater-solubility of the parem oximes but also
decreasedthe collector efficienC}', The decteasein "'ater-
solubili~' is to be expected from the decreasein ~larity of the
molecule as a resuh of substitution, The decreasein collector
efficien~' could be attributed to the decreasein acid strength of
the molecule o,,-ing to the increase in electron densi~' on the
phenolic oxygen caused by the nucleophilic substituents,20This
can, howe\'er, increase the metal-ligand stability, as is generally
observed, 3The effeCt of eleCtron-releasing groups on the ring is,
therefore, generally favourable for chelation, Again, the
reduCtion in bubble size appears to be the major faCtor in regard
to the lo"-er colleCtor efficienC}', This appears to hold for the low
collector efficiency obsen'ed in the caseof OHNAO, although
the solution concentrations stUdied \\'ere severely limited by its
low solubility,
Surface chelation
I t is now generallybeli~'ed that the adsorptionof xanthateon
galenaproceedsby the formation of a chelatecompoundPbX
on the surfaceof galena- asopposedto PbX2in the bulk.21
Thus, the fint layer on galenais belie,.edto be PbX (with a high

215
 tenacity)overwhich iayen of PbX1build up by 'sheerphysical
attachment'.n
Taggartand co-workers23 proposedthat surfacecompounds
differ greatlyin their propertiesfrom nortnally expectedbulk
compounds.This is becausethe lattice ionsor atomson the
sunlCehaveonly a part of their coordinationsphereto
contributeto the formation of a surfacecompound In addition,
stericeffectsassumea particularly significant role for reactions
on surfaceof mineral asopposedto thosein the bulk. It is
thereforeconceivablethat the first-layer compoundon mineral
maybe different from the compoundformed by a reactionin
the bulk. If it provesthat the bulk compoundis energetically
morefavourable(e\'enafter taking into accountthe fact that
adsorptionof collector is, in general,energericallyfavourable),
eitherthe collectormoleculesor the surfacecompound
comprisingthe collectormoleculesand latticeion or atom will be
scaled-offor detached.Of course,contribution to this
detachmentcould alsocomefrom physicalfactors.
It is not known whetherthere is a direct relationshipbetween
surfaceandbulk compounds.In manystudiesa I: I surface
chelatehasbeentacitly assumedwith little experimental
evidenceto suppon the assumption.Often elaboratestructures
for surfacechelateshavebeenproposed24.2S (tWoexamplesare
shownin Fig. 4).
The imponanceof distinguishingbetWeensurfaceand bulk
chelatesandquantitativelydeterminingthem in the samesystem
underflotation conditionswasclearlyshownby Nagarajand
So~undaran,19.2~.27 Surfacechelatewasfavouredunder certain
conditionsandtheseconditionscoincidedwith thosein which
flotationof tenoritewasobtainedwith SALO (Fig. 5). The bulk
chelateformedunder a ,,'ide rangeof conditionsand this chelate
wasineffectivein causingflotation whendispersedin the bulk
aqueousphase.This chelate,howe\'er,could aid flotation ,,'henit
is still attachedto the first layeron the mineral, asfor thiol
collectorson sulphides.Chanderand Fuerstenau28 discussed
the rolesof surfacereactionand bulk reactionin the system
chalcocite-DTP. Ananathapadmanabhan and Somasundaran,29
basedon detailedcalculationsof relevantmineral-collector
equilibria, haveclearlyshownthat much of the flotation results
in the literature canbe explainedon the basisof surface
reactionsor precipitation (asopposedto bulk reactionor
precipitation).
The influenceof bulk chelationasopposedto surfacechelation
hasnot beentakeninto considemionin the pastfor most
systems,evenwhenthe bulk chelationis Vcr)'pronounced.For
mineralsthat havefinite solubility (or speciation)in water, bulk
chelation(or precipitation)is inevitableif the kineticsof metal
chelation(or other complexformation)arereasonablyrapid Any
collectorassociated ,,'ith this chelatecanbe considered
essentiallywastedin flotation s)"Sterns.

Selective flocculation

The useof chelatingagentsfor selecti"eflocculation and

dispersionhasbeenclearly illustrated by several

:~"co.t4
HCi

~.ro,. !'
~ ".}=.I
,0"-/0\
~'C H + COZ + H2O
IC)
"C"-ro,-H
-1 .
~ l' ~
r-""\
J,o o
.~,
~ .;
--1
3~
~ Pd .tC'"
~
A
J ..-c
I H
'..j
-'..1
,~-Cu
~, 4
0 "'-to
.~~ +COz+HaO:?
- ~ 0 +COz+~O'
:j "
---

CuC03gq+~
(';::.yA ~ ~
('~OH fv.D~~ I
H I~~
I
+(01 + HaO

Fig. 4 Schematic illustrations «a)(l,ft}2' and (b) (abot.,f') of surfa~

chtlations

216

~
 invesngators.-- -- 1.nsome 01 these WOrkSCneJatlQggroups were
incorporated into a polymeric-type molecule) which acted IS
selective flocculants. For example, Sresty and
Somasundarann.)7 observed hydroxy propyl cellulose xantha1e
containing an active thiol group to produce good flocculation of
chalcopyrite with little effect on quartz (Fig. 6).

0
...

"
u
~
..

Reagcots
0-0 t~
Cupferron
Salicylaldehyde
Q-Nitroso~-naphthol
Acctylaccton~
.\1kylhydroxamicacid (IMSO)
Pbospbonic acids
.\'-<.JIy~
P-HydroX}'oximes
«-Hydroxyoximes
8-Hydroxyquinoline
Minerals
Cassiterite,uraninite, hematite
Cassiterite
Q)baltite
Malachite, chrysocolla
Chrysocolla,hematiteand minerals
aIrItainingTi, Y, La, Nb, Sn andW
Cassiterite
Fig. 6 Flocculation of chalcopyrite-andquartz fmcs in terms of per
cent solidsscttJedin 45 s asa function of hydrox}-propylcellulosc
Cu oxide minerals xanthate(re.ntizing time, 30sy'
Cu oxide minerals
Cerussite, pyrochiorc, chr}~ocoUa

.\' , Iypc TestS "lth synthetic mixtUres of these minerals showed both
Dipbenyiguanidine Cu minerals grade and recover}- of chalcopyrite in the sedimcnt portion to
Ni mi~ra1s improve with increase in the concentration of the above polymer,
Dimetbylglyoxime
Benzotriazo Ie Cu minerals but at Ver)- high dosagesentrapment of the quartz by the bulky
cha1cop~-riteflocs caused a decreasein the grade- This problem
S-S r_vpe can, however, be overcome easily by cleaning the product by
Xanthates redispersion follo,,-ed by settling-

Dithiophosphates All sulphid~ minerals

Dithiocarbamate5 Applications of chelating agents
.\"-S type Even though the relativeproportion of chelating-typereagents
.\ \crcaptobenzothiazole Sulphideand tarnished amongthe industrial non-thio collectorsis not significant, the
sulphides actualnumber of chelatingagentsthat havebeentested
Dithizone Sulphides successfullyascollectors,depressantsand flocculantsfor
variousmineral systems,at leaston a laboratoryscale,is large
S -0 ',vl'" (Tables 7,8 and 9). It is e,-identthat the actualuseof these
X-benzoyl a-alkyl Sulphides t).pcsof reagentsin the processingof mineralsis likely to
tbionoclrbamac undergoa significant rise in the coming decades.

217
 TableS Examplesof chelatingagentsasdepressants
Mineral
Reagent
Tanaric acid
Gallic acid
Alizarin red S
Starchxanthates
CtlIu~ xantbates
Reagents
Starchxanthates
Cellulosexanthates Selective flocculation and
depression of sulphides
Poly (4- and S-acr)'lamidosalyclic Cation-exchange polymer for
acids) Fe", Cu", Crt- and VOl"
Resincontaininghydroxy oximesgroup " ., " 20. Pltel R. P. andPatelR. D. Protonligand stability constantsof some
~l}:h~~xamic acids .. " ,.
Concluding remarks
I t is clearthat althoughchelatingagentscanbe usedeffectively
for the flotation of ores,they do not possessabsolutespecificity
towardsmineralspeciesand it is only the judiciouschoiceof the
chelatingagentsand conditionsfor selectiveseparationthat will
maketheir us~possiblefor the beneficiationof many
problematicores.In this regarda full understandingof the basic
mechanismsinvolved in their chemicalinteractionwith mineral
speciesin the bulk and on the surfacebecomesessential,even
though thosein the interfacial regionmight be understood
accuratelyonly by useof newexperimentalapproachesthat will
permit direct probing of this region.
The role of surfacechelationversusbulk chelationhasto be
takeninto accountin the developmentof any mechanismfor it
to be of significant usein flotation.
Also, more importantly, the role of surfacechemical
alterationsdueto either changein oxidation stateof surface
speciesor precipitation of variousdissolvedmineralspecieswill
haveto be consideredfor applicationof the aboveinformation
in aaual mineralprocessingsystems.
Acknowledgement
The authorswish to thank the mineralsand primar}' materials
programof the National ScienceFoundationfor the supportof
this work.
References
1. ~'Y~r F. P. and M~llor D. P.~ds.Chelaringagmua7ld_,alchela'es
(N~w York: Acadrmic Press,1964),530p.
2. CbaberekS. and Man~U A. E. °rtoni, sequeslfflng agellu(N~w
York: WiI~v, 1959).
3. Man~UA. E. and Calvin M. C},n,,;Sll:"of the_tal chelateCompoull4s
(EnglewoodCliffs, K.J.: PrenticeHall, 1952).
4. P~rrin D. D. Organi.-.;ompieringreagmls(New York: Wiley, 1964).
5. Pnr$oa R. A.J. Am. chnll. Soc.,85,1963, 3533;J. Alii. Chelll.Soc.,
89,1967, 103;J. chmI.Educ.,45, 1968,581;643.
6. K~nk S.F. A. CI'-OT.iinalion (N~. York:Appleton-
cOlllpou7lds
untUry-Crofts, 1~9),)20p.
7. Glembotskii A. V. Theoreticalprinciplesof foreclStingand
modifying collectorpro~nics. Tn.". Melalry, Mosk., 50, no. 91977,
61-4. (Russiant~Xt); TJ':~I.M~Iall.v,N. Y., 18,no. 91977,68-72.
8. Fkming 1.Fronlin or/Iilalsa7ldorganicchemicalreaclioru(London:
Wiley, 1976),258p.
9. Cotton F. A. and Wilkinson G. Adtoanced morganicchemistI:'.
(London: Int~rscience,1966).
10. Org~1L. E. All i",r£'duc,;on10transition_,al ChemiSll:"
(London:
Wiley, 1966).
11. Zolotov Y. A. Extract;"" ofchelatecompolnlds (Ann Arbor:
Humphrey Sciencc,1910).
12. WangT. T.J. Central-SouthInsc.Min. MecalJ.,4,no. 121980,
7-14. (Chineset~xt)
13. BogdanovO. S. et al. Reagentschemisorptionon mineralsasa
proccssof formation of surfacecompoundswith a coordinationbond. In
Prcxeedings of thel2ch incernationalmineralprocessing congrtss,1977
(SioPlu1o, Brazil: D.N.P.M., M.M.E., 1982),vol. 2,280-303.
14. Morrison R. T. and Boyd R. N. Organicchemistry(Boston:ADyn
and Bacon,1972).
15. March J. Adt-anc,dorganicchemiscry: rtactions,mechanisms and
st""cur,(N~. York: McGraw-HiD, 1968).
16. Millar I. T. and Springall H. D. Sidgwick'sorganicchtmistryof
mtrog,n,Jrdedll (Oxford: QarendenPress,1966).
17. Burger K. Organicreagmcs ill _ca! ana/_vIis(N~w York: P~rgamon,
1973).
18. Burger K. and EgyedI. J. Sometheoreticaland practicalproblems
in the useof organicreagentsin chemiClllna1~"Sis, V.J. inorg.IIucl.
Clatm.,27,1965,2361-70.
19. NagarajD. R. and Somasundaran P. Cbelatingagentsascollectors
in flotation: oximes-coppermin~ra1ss~..ems. "fin. Engng,N. Y., 33,
S~pt. 1981,1351-7.
ortbo-hydroX)'ph~nonesand their oximes.J. inorg.nucl.Chem.,32,
1910,2591-600.
21. Gran\.iUr A. Finkelst~in~. P. and Allison S. A. Rt~wof~oDS
in the flotation syst~ga1ena-xanthate-oxygen.TNIU. 1IUtnMin.
M,call. (Stct. C: Mintral Process. EXIT.M,call.), 81,1972,Cl-30.
22. RaoS. R. Xanthat,sand relatedcompounds (NtW York.:Marcel
D~kku, 1971).
23. Taggart A. F. TaylorT. C. and Ince C. R. Experimentswith
flotation r~agents.Trans.Am. 1nsc.Min. Engrs,87, 1930,285-368.
24. CecileJ. L. Utilisation d~ riactifs tn flotation. Doctoral thesis,
Univ~rsiteOrl~ans,France, 1978.
25. NagarajD. R. Cbelatingagentsasflotlids: hydroximts-copper
mineralsysten1s.Doctoraldissenation,Columbia lTniversity, 1979.
26. NagarajD. R. and SomasundlranP. Cbelatingagrotsasflotaids:
LIX8-coppu mineralsS}-St~ms. R,cmcDn.tlopmencs in Separation
Sc;".a,S, 1979,81-93.
27. NagarajD. R. and Somasundaran P. Commercialcbelating
~xtractantsascolltctors: flOtationof coppermin~ralsusing"LIX"
reagents.Trans.Alfl. Insc.Min. Engrs,266, 1979,1892-7.
28. Chandu S. and Fu~rstenauD. W. On th~ floatability of sulfid~
mineralswith thiol coUtctors:th~ chllcocitt/dittb~oJdithiophosphate
syst~m.In Proc,edings11thincernacional mintral processingcongreu,
Caglian, 19i5 (Cagliari: IstitUto di Ane Minuaria, 1975),583-604.
29. Ananthapadmanabhan K. P. and SomlsundaranP. Chemical
tquilibira in h~'droIYZ8ble surfactantsolutionsand their rolt in fk>tabon.
Paper~t~d to I 12thAI,\iE annul! meeting,Atlanta, March 1983.
30. Kitcb~ner J. A. Principlts of actionof pol}-m~ricflocculants.Br.
Po/.vmer J., 4, 1972,217-29.
31. Somasundaran P. StltctiVt flocculationoffints.ln Thephysical
chemistry' of ,nineral-reagmcinttraccionsin sulfid,/location: proceedings
of symporium,April 6-7 1978RichardsonP. E. Hy~ G. R. and Ojalvo
M. S. comps.lllform. Circ. [,'.S. Bur. Mines8818,1980,150-67.
32. Srcsi)' G. c. RajaA. and Somasundaran P. Selcctiveflocculationof
mintral slimesusingpolymers.R«mc Deo/."lopmmls in Separat;.,..
Sci,nu,4,1978,93-105.
33. Somasundaran P. Principlts of flocculation,dispersionand
selcctiv~flocculation. In Fin, parricltsproussingSomasundaran P. ed.
(Ne~' York: AIME, 1980),\'01.2,947-76.
34. Ania Y. A. and Fu~rstenauD. W. Principlesof separationof or~
minuals by selcctiveflocculation.In R«mc Dn',lopmmu in S'paratiorr
Scima, 4, 1978,51-69.
35. Ania Y. A.I. and Kitch~r J. A, Dtvelopment ofcompkxing
poly~rs for stkctive flocculationof coppermintrals. Ref~rence28,
1233-48.
36. BaudetG. tl Qt.Syntbes~~t caracterisationd~ flocculantssElectifsi
basedc derivesxanmes~ la ceUuloseet de ram~oJost. Induscnt
minir.-minhalurgie 1-78,1978,19-35.
37. Srtstv G. C. and Somasundaran P. S~ltcti\'e flocculationof
synthetic-mineralmixturesusing modifitd pol~.mers.lnl.J. Mintral
Proc,ss.,6, 1980,303-20.
38. ClaussC. R. A. Appltton E. A. and Vink J. J. Seltctiveflocculation
of cassittritt in mixtUreswith quartz usingI modified polyacrylamide
flocculant.lllc. J. Mintral Pr""ss., 3, 19;6,2;-34.
39. Ania Y. A. Synthtsisof PAMG cl1tlatingpol~.mtrsfor th~ selcctive
flocculationof copperminerals.Inc.J. Mineral PrOCtss., 4,1977,
191-208.
40. Rin~UiG. and Marabini A. M. A ne~.~~nt s~-stem for th~
seltctive flocculationof rutile. In Thirctmchinctmacionalminnal
proctssingcongress, Warsa11:, 1979LaskowskiJ. ~d. (Amsttrdam, ~tc.:
Else~r, 1981),vol. 1,316-45.

218

~
41. Rinelli G. and Marabini A. M. Dispersingpropeniesof tanning
agentsand ~sibilities of their usein flotation of fine minerals.
Reference33, 1012-33.
42 DrZ)omaIaJ. and LaskowskiJ. CJ1elating compoundsasflotation
reagents.Physico-ckmicalProbl~msof M;n"aIProc~ssing,13,1981,
39-64.
43. Barber)' G. and CecileJ. C. Complexingreag~ntsin flotation and
sclecti\.~flocculation: a ~\itW of possibilitiesand problems.Paper
presentedto CIM annualm~~ting,Winnipeg, 1983.

219