henry krumb school of mines columbia university new york. ny …ps24/pdfs/principles of...

PRINCIPLES OF FLOCCULATION, DISPERSION, AND SELECTIVE FLOCCULATION

P. Somasundaran

Henry Krumb School of MinesColumbia UniversityNew York. NY 10027

ABSTRACT

Processing of fine particles in suspension is markedly affected bytheir state of dispersion/aggregation. In addition, tendency for agiven component of the system to undergo selective aggregation orselective dispersion is an important property which may be put to usefor processlng. In this paper, principles governing flocculation,dispersion, and selective flocculation of mineral fines are discussedand then the role of major solution properties such as pH, ionic

strength, etc. is examined.

INTRODUCTION

Fines and u1trafines are invariably produced during various indus-trial processes. In certain processes it is the aim to produce asfine a material as possib1~ and to have it in a totally dispersed statewhereas in some other processes too fine a material than desired isoften produced as a part of the overall process. In all cases the na-ture of dispersion or aggregation of fine particles affects s.1gnificant-1y the performance of processes involving them as well as quality ofproducts manufactured. Thus, in paints and varnishes it is necessaryto have the particles in a dispersed state for long periods, whereasin waste water treatment it is the objective to get the particles inan aggregated and sedimented form. Because of the highly complex man-ner in which fine particles interact while in liquids, in order to ach-ieve the desired dispersed or aggregated state it is necessary .to havean accurate understanding of the physico-chemical properties that playa governing role in determining such interactions and of means bywhich one can control them. In this paper, a discussion of basic prin-ciples governing dispersion and aggregation of fine particles will be

947

948 FINE PARTICLES PROCESSING

presented after defining terms that are commonly used.

TEBKINOLOGY

Following the IUPAC guidelines (1), aggregation refers to the pro-cess or the result of the formation of a group of particles held to-gether in any manner. Coagulation refers to the specific type of ag-gregation that leads to formation of aggregates that are compact where-as flocculation refers to that which leads to aggregates that areloose or open. Usually addition of inorganic salts or acid or alkalycauses a coagulum to be formed and addition of polymers gives rise tothe formation of flocs.

As far as the particles themselves are concerned, there are a vari-ety of terms loosely used to indicate their size leading to consider-able confusion between scientists and engineers in various disciplines.They include fines, ultrafines, colloids, slimes, sludges, sediment,suspension, oozes, slurries, mud, gel, suspensoid, and sub-sieve sizeand sub-micron size particles. The following definitions are offeredhere on the basis of their size. Kajor interactions such as van derWaals attraction, electrostatic attraction, and gravitational sedimen-tation are dependent on the size of the particles. The classificationsgiven below are based on the magnitude of these interactions.

Colloid: colloid is composed of particles that are less than 1Ddcron in size.

Ultrafines: these are particles that are not separated by conven-tional non-gravity processes such as flotation and are less than 10microns in size but greater than 1 micron in size.

It is recognized that for certain minerals such as coal, the mini-mum size for flotation might be much greater than that indicated above

Fines: particles that are not easily separated by gravity processesand are less than 100 microns in size but not less than 10 microns insize.

Coarse particles: these are defined as particles that are largerthan 1000 microns (16 mesh).

Slime: slime is defined as a mixture of colloidal and ultrafinesize waste mineral matter generated during milling operations and hasslow settling characteristics.

Sludge: this is colloidal mineral or non-~ineral matter which has atendency to form a highly porous three-dimensional netwQrk Btructureand has, as a result, very slow deli~uoring characteristics.

PRINCIPLES OF FLOCCULATION. DISPERSION. SELECTIVE FLOCCULATION

BASIC PRINCIPLES

When fines are produced during mining or milling operations theparticles gain energy in the form of surface energy and deformationenergy which gets stored in the sub-surface region. The system caneliminate this increase in energy partly by permiting the new solidsurface to undergo interactions with the solvent and solute molecules.In addition, most systems can also be expected to eliminate as much ofthe increase in energy by means of re-aggregation. The rate of thisre-aggregation can however be extremely slow depending upon the natureof interactions between the particles as well as energy changes in-volved in, for example, the deso1vation or desorption of any species inthe interfacial region upon re-aggregation. It is mostly this ratethat can be controlled in order to achieve the desired dispersion oraggregation of the system. In this paper therefore kinetic aspects ofaggregation will also receive emphasis in addition to the equilibriumprocesses.

The rate of aggregation will depend essentially on probability ofcollisions between particles, probability of attachment during suchcollisions and probability of detachment of particles from the aggre-gates subsequently. While the probability of collision will depend onthe Brownian motion determined essentially by the temperature of thesystem and on fluid flow motions determined by the viscosity of thepulp and external stirring, probabilities of adhesion and detachmentare dependent on the type of physico-chemical interactions between theparticles and, to some extent, on velocity gradients in the pulp.

Both Brownian and velocity gradient aggregations are described bySmoluchowski treatment that relates the rate of aggregation to theradius of the particle and the total number of particles in the systemin the following manner. Aggregation due to Brownian motion, known asperikinetic aggregation, leads to a decrease of particles governed bythe following second order equation if every collision is successfulin creating an adhesion (2).

dBt -s 2- - 'ffD r N

:dt lIt

where Nt is the total number of particles at time t. D is the diffu-sion coefficient for the particle, and r is the radius of the particle.Integrated form of equation (1) is

NONt

-1 + 8wDlrlRot

a

This derivation takes into account formation of only doublets fromoriginal particles. In reality primary particles can also aggregatewith multiplets. Solution for all particles aggregates containing iparticles is given as

FINE PARTICLES PROCESSING950

NOr Hi - 1 + 4'R'DlrlHoti-I

(3)

Slow stirring can indeed promote aggregation. Such aggregation underthe influence of velocity gradient is known as orthokinetic aggrega-tion. Rate of orthokinetic aggregation is given by

3.dN 16Gr 1 p 2k-!-- - 1: Nkdt 3 k=l

G is velocity gradient, Nk is number of k particles per unit volume,k is particles formed by aggregation of two particles, p is particleof limiting size in shear gradient G. As mentioned earlier, aboveformulations assume that every collision is fruitful. In reality,only a fraction of the collisions might be effective and as a resultthe aggregation might be slow. The magnitude of this fraction is de-pendent upon the nature of interactions between particles and, to someextent, viscous properties of the liquid film between the particles.If V is the total energy of interaction between particles of radii riand rj when they are separated by a distance d the fraction of effec-tive collisions, a, according to Fuchs (3) is given by the followingexpression

~ V dS! - 2 I' exp (iT) ~ 2a 2

As an example, for-5 -9,10 ,and 10 ,

2dk is Boltzmann constant and S is equal to ri +rj.

-2maximum V equal to 5, 15, and 25 kT, a is 2.5 x 10

respectively.

The interactions between particles that are important in determin-ing a are made ~p of London-van der Waals attractive forces (Va)' elec-trical double layer forces (Ve) that can be attractive or repulsive innature, bridging forces (Vb) that are always attractive and stericforces (Vs) that arise from the overlap of adsorbed layers. Stericforces can be repulsive or attractive depending on the solubility pro-perties of the adsorbed layer.

v-v +V +Vb +V a e s

~~tractive Fo~c~s

Va, the attractive interactions,arise from London dispersion forces,Keesom forces due to interactions between permanent dipoles and induceddipoles in the particles and Debye forces due to interactions amongst

PRINCIPLES OF FLOCCULATION, DISPERSION, SELECTIVE FLOCCULATION 951

1induced dipoles. These forces vary as ~ Where d is the intermolecularspacing. For macroscopic bodies, the forces When summed over all the

1molecules in the body vary as dIi. n.has a value less than 6, dependingupon the shape of the interacting bodies. For example, for two flatplates n has a value of 2 and for between two spheres it has a valueof 1. Thus

v between two plates = - (7)2aA

121fD

v between two unequal spheres = -a

AR6.n

between a plate and a sphere =-va

A is a constant for a given system of materials and is known as theHamaker constant. It is dependent essentially on the polarizing pro-perties of the molecules comprising various components of the system.Hamaker constants have been compiled for a number of systems by Healy(4) recently and are given below.

Table 1. - Approximate Hamaker constants (non-retarded cases) for

systems as shown (4)

All the above values refer to the case where the particles are fairlyclose to each other so that there is no real lag in the propagation ofelectro-magnetic waves between them. On the other hand, if they are


separated by a very large distance (5) the van der Waals attrac- 1tive energy of interaction is retarded approximately by a factor of D.Hamaker constant has been experimentally determined recently for themica/water/mica system and was found to be (2.2 z 0.3) x 10-13 erg forlOX < D < 65X in fair agreement with the value given above for quartz/water/quartz system (6). Interestingly, this value is found to beindependent of salt concentration below 1 M, whereas other forces suchthe electrical double layer force that will be considered subsequently,are very much dependent on ionic strength, particularly in the 10-5 to1 M range.

Elec~rical Double Lay~rForces

Most particles in liquids are electrically charged, the nature andmagnitude of which are dependent on the mechanisms responsible for thecharge genera'tion on their surfaces and the properties of solids andsolution comprising the system. Predominating mechanisms of chargegeneration on the surfaces are preferential dissolution, or hydrolysisof surface species followed by pH dependent dissociation of surfacehydroxyl groups (7 to 12). Alternate mechanisms that have been pro-posed include dissolution of lattice ions followed by hydrolysis in thebulk and subsequent adsorption of the resulting cJmplexes. Adsorptionof other charged species that are present in the system or impuritiescan also give rise to a charge on the surface. Thus, for simple saltssuch as silver ~odide, preferential dissolution and adsorption ofla~tice ion.s, ~+ and I-are considered to be the governing mechanism.Ag+ and 1- ions that determine the surface charge are called potentialdetermining ions. The surface will be positively charged under lowpAs conditions and negatively charged under high pAg conditions. ThepAs (or pI) at which the net charge of the surface is zero is calledthe point of zero charge (pzc). The point of zero Charge is an impor-tant experimentally accessible property of the system since its loca-tion with respect to solution conditions will essentially determinethe repulsion or lack of it between particles and consequently thestability of a suspension of the particles. For oxides, the hydroly-sis of the surface species followed by pH dependent dissociation isconsidered to be the major governing mechanism and hence H+ and OH-are potential determining ions. For salt-type minerals such as cal-cite and apatite, preferential hydrolysis' of surface species as wellas preferential dissolution of ions which is often accompanied by re-actions with the solution constituents and possible uptake by thesolid have been proposed to be important for determining the surfacecharge (10,12-16). Thus for calcite, both Ca++, C03--, ~, and OH-can be considered to be potential determining ions even though Ca ++ has

not been found to exhibit all the characteristics of such ions. Incontrast to the above minerals, clay minerals exhibit charge character-istics that are dependent upon the special structural characteristicsof these minerals. Most natural sediments have clays in them and itis often the properties of the clay components that determine the be-havior of these sediments. Also, clay is present in most ore bodiesand has a major influence on the processing of ores. Ciay minerals,because of their unique structure consisting of sheets of 5i04 tetra-


hedra and Al06 octahedra with possible substitutions, for example of8i4+ with Al 3+ , exhibits a negative charge on the face of the mineral

platelets that is pH independent, and charge on the edges that is pHdependent due to hydroxylation and ionization of the broken 8i-0 andAl-O bonds at the edges. The point of zero charge in such a case isdetermined approximately by the algebraic sum of face and edge charges.It is interesting to note that at the point of zero charge both thesides and faces of clays can be charged. Because of these interestingheterogeneous charge make-up of clays, they show peculiar flocculationproperties. For example, clay minerals under appropriate so1ution con-ditions can adhere to each other by means of face to edge attachmentand give rise to the so:called card-house structure.

The charge of the mineral/solution interface can indeed be modifiedfor controlling the flocculation by adsorption of ionized chemical spe-cies or polymer. This is possible since it is not the surface poten-tial that is directly controlling the interactions between particles,but an interfacial potential at some distance away from the surfacethat the particle will have to work against as they approach eachother. Control of this potential, which can be assumed for conveni-ence to be the potential at the shear plane, known as z;-potential, ispossible by adjusting the ionic composition and pH of the medium andthis will be considered further subsequently.

Steric Forces

Steric forces arise from the overlap of the adsorbed layers and canbe repulsive or attractive in nature depending on whether or not theoutermost layers on the particles prefer to be in contact with the sol-vent. If the solvent power of the medium for the exposed portions ofthe exterior layer, for example those of an adsorbed polymer layer, issatisfactory, they will be compatible with the medium and the suspen-sion will remain stable. On the other hand if the solvent power forthe adsorbed layer is minimal, there will be a tendency for the ex-posed portions of the adsorbed layer on one particle to interpenetrateinto those of a layer on another particle and thereby promote aggrega-tion. ~

Interpenetration and aggregation is possible only if the net changein Gibbs free energy due to interpenetration of the polymer chains isnegative (17). The Gibbs free energy change is considered in suchcases to be determined essentially by the change in entropy due to re-lease of solvent molecules and due to the decrease in randomness ofthe polymer chain and by the enthalpy of deso1vation of the polymerchains.' Briefly, for flocculation to occur according to this mechan-ism, the increase in entropy due to the release of solvent moleculesshould outweigh the loss of entropy due to interpenetration of polymerchains and increase in enthalpy of deso1vation (see Figure 1). It isclear that the deso1vation characteristics of the adsorbed polymer spe-cies and the dependence of it on solution properties such as tempera-ture and ionic strength will be important in determining aggregationbY this mechanism.


(..:

..<+

(0) COMPLETE COVERAGE

AGr. !J.Hr - T!J.Sr

t ~from polymer increase from releaseddesolvation solvent molecules;

decrease from interpene-trating polymers

Flocculation under complete coverage is possible only ifthe free energy of interpenetration of adsorbed layers,~GI' is negative.

t..v<.

\.,r-+-<-- (><.(.

PARTIAL COVERAGE(bl

Figure 1. Flocculation by polymers under complete coverage and partial

coverage conditions.

-",,<,

J ) (,. .c. ')

PRINCIPLES OF FLOCCULATION, DISPERSION, SELECTIVE FLOCCULATION

~ridging Forces

A long chain adsorbate with several active sites on it can induce ag-gregation by attaching itself to two or more particles. Polymers canprovide such bridging between particles particularly under conditionswhere particles are not totally coated by the polymeric species (seeFigure 1). If particles are already fully covered with polymers. bridg-ing can take place only if there is either detachment of some portion ofthe polymer already on a particle to provide sites for attachment of poly-mer fractions adsorbed on other particles or polymer-polymer bonding it-self. It has been suggested by various authors including Healy and LaMer that maximum flocculation and filtration rates occur when the frac-tion of particle surface covered by p~lymer molecules is close to 0.5(18-20). It is to be noted that. to this author's knowledge. this hypo-thesis has not. however. been experimentally proven. As indicated earli-er. bridging should be possible even when particles are fully covered ifsome detachment and reattachment of the adsorbed polymer is possible. Ad-sorption of polymers is in general found to be- irreversible (21.22). Butin the author's opinion. that is. except in the occasional case of poly-mers with functional groups that form covalent bonds with the surfacespecies. only due to the cumulative effect of adsorption of a singlepolymer molecule on several sites. Detachment and reattachment of ~-tions of such a polymer should, however, be possible and such a processshould facilitate bridging. Absence of bridging might then be inter-preted to suggest the presence of strong repulsive steric interactionsbetween dangling portions of the adsorbed polymers.

AGGREGATION/DISPERSION AND ROLE OF ADSORPTION IN IT

As mentioned earlier, aggregation between particles can take placeif the interparticle repulsive force does not exceed in magnitude thecorresponding attractive force. As it is the electrostatic force thatis most often responsible for the repulsion, one can achieve aggrega-tion if solution properties could be adjusted to minimize such forces.On the other hand, if dispersion is desired one would need to maximizethe electrostatic repulsive fo.rces between particles. On the basis ofthese considerations, aggregation can be expected to occur under theconditions discussed in the following sections and dispersion under alother conditions. :.

A. Homocoagulation

As indicated earlier, the total interactions between the particleswill be determined by the summation of forces that arise from the over-lap of similarly charged electrical interfaces and from that of. theadsorbed layers that prefer to be in contact with" the solvent.

V=V +V+V b +Va ~e s

Excluding Vb and V s' the following expression can be written for thetotal energy of interaction between two particles separated by a dis-tance HO in terms of the average Hamaker constant A and surface poten-tial ~l and ~2 of two particles of radii rl and r2 respectively (23).


v - v +va e-Arlr2(1- +

6(rl + r2)HO

1 +exp( -KUO) 2 2- 4(rl + r2} [21/1 11112J.n 1- exp(::KH~) +('1 + 1/12) 1n{1- exp( -2KHO) }

£ is the dielectric constant, l/K is the doUble layer thickness and (1is a factor that takes into account retardation of London forces underconditions of aggregation in secondary minimum indicated in Figure 2(24) . It can be seen from the above expression that V will be attrac-tive at large and small distances since exponential of -KHO will besmaller than the inverse power of H for large and small Ho. Typicalvariation of V as a function of dis~ance between particles is illus-trated in Figure 2. The bottom curve is for a case with negligible

rlr2£(10)

+REPULSION

ENERGY

ATTRACTIONENERGY

Figure 2. Diagram illustrating variation of V as a function of distancefrom the surface.

electrical repulsion between particles and the upper curves for condi-tions with increasing repulsion. On the basis of this, it is clearthat one can obtain aggregation by controlling the charge on the par-ticles so that repulsive energy between them will be less than that ofthe attractive energy. Healy has recently summarized zeta potentialconditions to meet the above re9uirements (4). It has been shown thataggregation can be obtained if Iclis less than about 14 mV. Zeta po-tential of a mineral can be adjusted in a number of ways:


1 Control of the Surface Char "e b the Addition of the Re uired Poten-tial Determining Ions. ex: HCl or NaOH for titania, quartz, aluminaand clays; AgNO3 or II for silver iodide.

This method is illustrated in Figure 3 where zeta potential andstability ratios of alumina are given as functions of pH (25). Stabi-lity is minimum near zero zeta potential conditions.

+6 (a)I

PIC0

. WATERAKCIO.oKCICKNO3

103M

>-

~ +4

8e2~u> +2

,~Q.~~ 0

f~0&~... -2t- .~ueUJ.JUJ -4

-6

w'-

~~Z

.J~~I&Jcn

I&J>.::c{.JI&J~

10

8

6

4

2

7 8 9 10pH

Figure 3. Comparison of stability of alumina suspensions with zetapotential of the particles at various pH values (8).

b~ ~ompres~ion of the Double Layer by the Addition of Required Amounts~(Inorganic Electrolytes. ex: KNO3' CaCl2' AINO3 for quartz, silveriodide, alumina, etc. The effect of the compression of the doUblelayer is illustrated in Figure 4 where results obtained by Wiese andHealy for the addition of KNO3 to titania suspensions are given asfunctions of pH (26). Zeta potential of titania is shown in this


40

20

0>E

-Jctt-Z\4Jt-oa..

~\4JN

-14mV. \'",\ °,

" "\ .

-20

-40 \ "\ .,,

10-2" " ', "

"--60 ,, ---10-3", "

10-4-80

I I I I I r I I I I I I . ,I

4 5 6 7 8 9 10 11

pH

-2 -3 -4Figure 4. Zeta potential-pH isotherms for Ti02 in 10 t 10 and 10KN03 solutions. pH values at which rapid coagulation wasobserved are also shown (26).

figure along with pH values at each ionic strength at which coagula-tion was observed. The region between two horizontal lines describesthe experimentally determined pH-KN03 conditions for coagu~ation. Itis interesting to note that zero zeta potential is not required for

obtaining coagulation. Even when the zeta potential is +10 mV, coagu-lation is obtained suggesting that the van der Waals attractive forcebetween particles is higher than the electrostatic repulsive force pro-duced under such potnetial conditions. Multivalent ions can also in-deed reduce interfacial potential and cause aggregation, but in thiscase it is necessary to exercise careful control of the addition sincethey ean speci.fically: adsorb to reverse the zeta potential ~nd then evenincrease it to very high values at larger additions. With certain

multivalent ions, the effect on coagulation has also been attributedto their increased adsorption when they are present in hydrolyzed form


chain surfactants has been attributed to lateral association betweenadjacent hydrocarbon chains of the adsorbed surfactant ions. Suchassociation to form two-dimensional aggregates is analogous to micelleformation and results from the favorable energetics of the removal ofthe alkyl chain partially from the aqueous environment. Mineral aggre-gation by surfactant is illustrated in Figure 6 with alumina/dodecyl-sulfonate system as the example. In this figure, stability ratio(which is inversely proportional to coagulation) for this system isgiven along with sulfonate adso~tion density and electrophoretic mo-bility. Similarity of the adsorption isotherm to stability isothermis clearly evident in this figure. It should be noted that the ionicstrength of the system is maintained at a constant level of 2 x 10-3 Munder all conditions and therefore the observed changes in electrokin-etic and stability isotherms is not due to the double layer compres-sion that one normally gets with the addition of a salt. The increasedadsorption of the anionic sulfonate due to the two-dimensional aggre-gation and consequent reduction in zeta-potential of the particles iswhat is considered to be mainly responsible for the observed change instability. It is however to be noted that alumina does not disperseeven after the surface charge is reduced to zero and then becomesnegative. In other words, maximum aggregation is not obtained at con-ditions corresponding to zero zeta potential. Continued increase inaggregation even after the system has undergone charge reversal wasattributed to the fact that it may be possible for particles with theirsurface partially covered to be bridged to each other with the help ofthe adsorbed surfactant species. Such an aggregation would help to re-move some of the hydrocarbon chains of adsorbed surfactant moleculesout of the aqueous e~vironment and thereby lower the free energy ofthe system further.

Inorganic electrolyte such as sodi\Dn silicate and sodilnll hexameta-phosphate are added often to increase the dispersion of a system. Theexact mechanism by which these reagents work is not known even thoughit is speculated that the charge of the particle does possibly increasewith the addition of these ions.

B. Heteroaggrega~ion

Heteroaggregation between two different types of particles willalso be governed by equation 10 which relates the interactiOn energybetween the two particles to their surface potential and the Hamakerconstant. Hamaker constant A in this case is the average for, say,particles 1 and 2 in medium 3 and is given by

(11)A =. A12 + A33 - A13 - A23

The easiest way to get aggregation in this system is to select condi-tions such that one type of particles will be charged oppositely tothe other type, i.e., 1111 will be opposite in sign to 1112. James et al

(31) have tested this possibility by mixing titania 'and ,A1Z03 in one


Eu-0>~

G)

Q..

U

G)

cn

cnC0~u.-E

):'~-J-a)0~

S::2I-W~0:J:a..0~I-ULLJ...JLLJ

6.-3

a-2

-1/20 CF A

- ~~N.:::!~::'.:.~~-

aI-I

A a

0

1+ IA

+20 ADSORPTIONDENSITY

A SETTLINGINDEX

c MOBILITY

~

E(,)

OJ

0

E 10"-alI.Jma:0(/)0<!

~ 1012::)0~<!

+3

>-..-(/)ZlLJI-Z-

100'5lLJ(,!)Z«--..A-

UI L1J

>10.0 ~ -

..Jl.&.Jcr

-XlLJ0Z-

1.0 ~..Jt- ~-I-4I-- .

L1J! U)

IONICSTRENGTH: 2 x 163N

pH 7.2--+5

-1310

Figure 6. Settling index, adsorption density and electrophoretic mobi-lity of alumina-sodium dodecylsulfonate system as a functionof the concentration of the sulfonate (28).

EQUILIBRIUM CONCENTRATION, molel liter

FINE PARTICLES PROCESSING962case and amphoteric latex colloids in another case at various pH val-ues; maximum coagulation was obtained when equal amounts of the twocolloids were mixed at a pH that is intermediate betWeen their isoelec-

tric points.

£:_polymer Flocculati~

Forces responsib1e for adsorption result mainly from three types ofbonding, i.e., electrostatic, hydrogen and covalent bonding. The pre-dominance of any of the above mechanism over the others depends on theparticular mineral/polymer system and the properties of the aqueousmedium. Under favorable conditions, more than one type of mechanism

can be operative.

~lectrostatic_Bondin~. In the case of a polymer with a large numberof charged units on it, electrostatic bonding is the predominant mech-anism by which the polymer adsorbs on particles. Such polymers canneutralize and even cause a reversal of surface charge at very lowdosages (33,34). This effect is illustrated in Figure 7. Polystyrenesulfonate adsorption on positively charged hematite particles is foundto reduce the charge on it to zero and then at larger additions tomake it negative. Ability of the anionic sulfonate to adsorb inamounts higher than those needed for complete neutralization and to ad-

PR

INC

IPLE

S O

F F

LOC

CU

LAT

ION

, D

ISP

ER

SIO

N, S

ELE

CT

IVE

FLO

CC

ULA

TIO

N

A ~

'1~

I.lN31O

d 'vr13Z

0It)- 00(\J0 0&n

00- -U)

-U)

<C

CD

U)

C-10U)

>-

a:a-~a..

Q..-

UZ0UlLJ

~Z0LL-J:::)(/)

lLJ

ZlLJ

a:>-

..-(/)~0a..

IQI

CQI

~>-

~VI

>-

r-40~s~~0VI

11-1

0C0~~Cd

~~CGJ

UC0u~

0c0~~u§11-1

Cd

VI

Cd

VI

Q)

C~~Q)

~~~aor:.

~0.

r-4-

Cd

0\

~~~-QQ

) Q

I

~~0

Cd

~~

0

Cd~

~~

N

VI

.r-..Q

)~=

'~orf

~

963


sorb on even negatively charged hematite particles results from thefact that there are other mechanisms in addition to electrostaticmechanism that are operative in this case.

Hydrogen Bonding. As mentioned above, anionic polystyrene sulfonateis able to adsorb even on negative hematite particles. Also it iswell known that even non-ionic polymers can adsorb in measurableamounts on mineral particles. This is attributed to the presence of-NH2 and -DB groups on the polymer that can hydrogen bond with thesurface oxygen species of the mineral~ Flocculation by interparticlebridging follows such adsorption if the electrical double layer repul-sive force between the particles is not very strong. It is to benoted htat adsorption by hydrogen bonding alone is relatively non-selective and polymers that adsorb due to hydrogen bonding are there-fore ideally suited for bulk flocculation rather than selective floc-culation.

Covalent Bonding. Selective adsorption of polymers can be achievedif adsorption is due to covalent bonding between the ac~ive groups ofthe polymer and the cations of the mineral surface. Such flocculationof kaolin by polyacrylamides owing to the formation of salt-type com-pounds by reaction between the polymer and the Ca++ ions present onthe kaolin has been repor~ed by Michaels and Morelos (40). A good ex-ample in this regard is the flocculation of chalcopyrite that is ob-tained by the addition of hydroxypropylcellulose xanthate (see Figure8) (33,41). Xanthates are known to Chemisorb on chalcopyrite and evi-dently the better performance of the clelulose xanthate on chalcopy-rite than on quartz is due to similar chemical adsorption of the xan-thated polymer on the chalcopyrite particles.

Kinetics of Adsorption. Rate of uptake of polymer by minersls is gov-erned mostly by the diffusion of the polymeric species to the surfaceof the particles and its pores, which in turn is influenced by themolecular weight, structure, and configuration of the polymer molecule,polymer concentration, agitation, temperature, ionic strength, and pH(22). As can be expected, the rate of adsorption increases with in-crease in agitation and polYmer concentration (42,43). The increasein molecular weight of the polymer, on the other hand, has been re-ported to produce an increase in some cases and a decrease in rate ofadsorption in others (44-46). In general, when the suspending mediumis a poor solvent for the polymer, the rate of adsorption of the poly-mer is greater. Both the kinetics and the extent of adsorption arevery much dependent on the porous nature of the particles. Equilibri-um is attained of course only slowly if the particles are porous (47,48). It is interesting to note that with porous particles, the ex-tent of adsorption can vary also due to any change in configuration(coiling) that the polymer might undergo with changes in solutionproperties such as pH and ionic strength.

PRINCIPLES OF FLOCCULATION. DISPERSION, SELECTIVE FLOCCULATION 965

~

75

0&&I..J..-

t,(/)

* CHALCOPYRITE pH = 3.6i

.. QUARTZ .pH = 7.0cn0.J0(/)

at

100 200

XANTHATE CONC., PPM (DRY SOLIDS BASIS)

3000

Figure 8. Percentage solids settled of chalcopyrite fines and quartzfines as a function of concentration of hydroxypropylcellu-lose xanthate (39).


HYDRODYNAMIC ASPECfS

As mentioned earlier, flocculation is dependent both on the prObs-bility of collision and the probability of adhesion. Some of the fac-tors that govern the probability of adhesion have been discussed ear-lier. Probability of adhesion during collision can be enhanced alsoby providing sufficient kinetic energy by means of stirring, this ef-fect being particularly dependent on the size of particles. For par-ticles that are larger than 1 micron, kinetic energy provided by stir-ring can be high enough to overcome the electrostatic repulsion energyThis becomes evident upon examining the results of Warren given inTable 2 (4).

Table 2. - Energies of particles in suspension due to various inter-actions (4).

Under turbulent conditions, collisions can occur either due to dif-fusion while entrained in eddies or due to inertia. Levidh's theorypredicts the collision rate ~nder turbulent conditions, Jturb' to be

11)Jturb - 12WBr3n12(~)~

where ~ is a constant, ni is the initial concentration of particles,EO is the energy loss per second per unit volume, and v is the kine-matic viscosity. Interestingly, under conditions selected by Warren(49) on the basis of the above formula for equal collision rates be-tween fines and fines and between fines and coarse, greater aggrega-tion was observed by him between coarse and fines than between finesand fines. It was suggested that the probability of adhesion is lar-ger for collision between a co~rse and a fine particle than betweentwo fine particles.

It is to be noted that the type of mixing and agitation can affectflocculation using polymers significantly also because of the effect of


agitation on the distribution of the polymer uniformly in the pulp. Inthis regard, the point of addition of the polymer, the concentration ofthe polymer solution and the manner of addition can be expected to havean influence on the polymer uptake and flocculation.

SELECTIVE AGGREGATION OR DISPERSION

The extent of selective aggregation or dispersion that can beachieved with minerals depends on the liberation of individual mineralparticles from each other and adsorption of flocculants on one or more,but not all, mineral types followed by their aggregation. Separationof the flocculated particles from the others can then be achieved byusing conventional techniques such as flotation, elutriation or sedi-mentation with care to produce minimum redispersion of flocs duringsuch processes. In the case of flotation, it is also necessary totake into account the interactions of flocculatns with flotationagents.

An examination of various particle-particle interactions discussedearlier suggests following criteria for selective interactions (aggre-gation or dispersion) betWeen, say, particles of component in a mix-ture of 1 with many other component minerals (2,3, etc.)

1. Particles of one should carry a charge similar to as many othermajor minerals as possible so that there will be no aggregation be-tWeen them. The magnitude of the charge should be such that therepulsive energy will be larger than the attractive energy. As men-tioned earlier, the zeta potential of the particles should be greaterthan 15-20 mV in magnitude. l~lland 1~21 > 15 to 20 mv; ~1/~2 = +.

2. Charge on the particles to be aggregated, say, of component 1,should be such that the repulsive energy between them will be lessthan the attractive energy. For this zeta potential of these par-ticles should be~ at elast after adsorption of the coagulants orflocculants, less than 1~'mV in magnitude. 1~11 < 15 mV. An ap-parent contradiction witp requirement 1 is evident now since it re-quires 1~11 to be greater than 20 aV. This requires, at least inprinciple, that It11 be, say, near 10 mV and It21, to counteract thereduction of potential on particle 1, be about 30 mV or so. It isinteresting to note that in practice, however, selective aggregationis rarely obtained even when solution conditions meet the above re-quirenents (35).

3. Adsorption of coagulants or flocculants should be selective. Inthis regard, it might even be sufficient to have differences in justthe rate of adsorption of reagents on various minerals components.In such a case, the differences in kinetics have to be utilized byopttaizing reagentizing ttaes for different polymer concentrations.

Reagents that are used for inducing selectivity include polymers,dispersants and activators. Adsorption of polymers is usually not as


selective as that of surface active or inorganic reagents. Partlythis is due to the hydrogen bonding nature of the polymer adsorption.Selective adsorption of polymer molecules can be achieved by adjustingthe chemical composition of the suspending medium and thereby the sur-face potential on the mineral or by introducing into the polymer activefunctional groups that will form complexes or salts with the metal spe-cies on the surface of the desired mineral. In the former case the de-pendence of adsorption of ionic polymers on surface charge is explo1ted.The dependence of polymer flocculation on electrostatic interaction isclearly illustrated in Figure 9 which gives data for settlinp of floc-

100

.. NALCOLYTE - 610

... SEPARAN - AP30CI&J..Jf-f-I&JC/) 7~

C/)c~0C/)

et

pH = 3.5 TO 3.6

~

500:::1-.~- - _J -

60 , 120 180POLYMER CONC, PPM (DRY 5a..IDS BASIS)

Figure 9. Diagram illustrating settling of synthetic silica in thepresence of cationic Nalcolyte-610 and anionic Separan AP-30

culated silica in the presence of an anionic and a cationic polymer.It can be seen that only the cationic Nalcolyte is able to produce anysignificant flocculation of the negatively charged quartz particles.Based on such a premise, selective separation of quartz from its mix-tures with hematite, calcite, etc. has been achieved by several in-vestigators (50-52) using, for example, anionic polyacrylamide.

As mentioned earlier, selective flocculation can also be obtainedby incorporating groups that can form covalent bonding with the miner-al surface species. It is well known that adsorption of collectors onminerals deoends on the oresence of functional groups such as carboxy-


late and mercaptan in them. It should be possible to obtain selectivi-ty during flocculation by incorporating such groups that are alreadyknown to adsorb selectively. Such an approach has been successfullytested in the past by a number of workers (33,38,40,53-55). For ex-ample, tests with hydroxypropylcellulose xanthate containing mercaptanas the active group has been observed to produce good flocculation ofchalcopyrite with practically no effect on quartz (see Figure 8). Re-sults of tests with synthetic mixtures of these minerals are shown inFigure 10 as a function of concentration of the xanthate (39). An ex-cellent separation index of 0.75 was obtained after a one-stage clean-ing operation in which quartz particles that have been trapped in thebulk chalcopyrite flocs, particularly at high polymer concentrations,were washed away.

Activators and Dispersa!!ts

Selective flocculation or dispersion of natural ores is often madedifficult owing to interference from the dissolved ions. Like in thecase of flotation, additives that can complex with such dissolved ionsor adsorb on mineral particles selectively to modify them can be usedto enhance selectivity in such cases. Thus, separation of hematite-quartz mixtures using anionic polyacrylamide has been reported to bepromoted by the addition of Calgou (principally sodium hexametaphos-phate) and sodium fluoride (56). Similarly, flocculation of heavyminerals has been depressed by the addition of sodium sulfide, po1y-phosphates and po1yacrylates (41). Po1yphosphates and polysi1icatesare widely used also as bulk dispersants. Such dispersants are par-ticularly effective when excessive slimes are present in the systemsince these colloidal particles do have a tendency to coat the mineralparticles and thereby reduce the selectivity of a flocculation opera-tion (57).

In this regard it is important to note that selection of any dis-persant has to be made so that there will be minimum interference bythese reagents on subsequent adsorption of flocculants. Indeed allthese reagents including th~ flocculant itslef should not interferewith any downstream process! such as flotation, filtration or pelletiz-ing. !

An interesting new approach of activation has been recently deve-loped by Rubio and Kitchener to induce selectivity in flocculation(58). Selective flocculation was obtained by these authors using'hydrophobic polymers' on minerals that are preconditioned with sur-factants to produce a hydrophobic surface.

Spherical Agglomerati~

Minerals reagentized with surfactants can also be aggregated usingoil. In this process, fines when tumbled in an aqueous solution con-taining oil aggregate due to the formation of capillary bridges be-tween surfactant coated particles. Such aggregation has been obtainedby Puddington It al (59,60) for graphite, chalk, zinc sulfide, coal,


x~0ZH

Z0

~~«Q..~cn

1000 15000 500

XANTHATE CONC., PPM (DRY SOLIDS BASIS)

Figure 10. Separation index achieved in selective flocculation ofchalcopyrite-quartz mixture as a function of concentrationof hydroxypropylcellulose xanthate (39). '


iron ore and tin ore in aqueous solutioos. Similarly good separationhas been reported by Read and Hollick (57) for a lead/zinc ore wherethe slime fraction of the ore was dispersed by a combination of sodiumsilicate, sodium carbonate, and a polyacrylate and subsequently condi-tioned with copper sulfate and then potassium amylxanthate.

Garrier Flot-.-ation

A selective aggregation process, known as carrier flotation as wellas ultraflotation and piggy-back flotation, has been successfully usedfor more than fifteen years for the removal of anatase impurity fromkaolin (61). In this method an auxiliary mineral is used as a carrierfor the fine particles to be separated. The carrier particles arecoarser than the material to be floated. It is speculated that thefines form a coating on the carrier mineral and are then floated alongwith the coarser carrier particles. Recently Wang and Somasundaran(62) have obtained evidence that anatase can be floated without thecarrier also even though to a lesser degree than in the presence ofthe calcite carrier. Carrier calcite particles activate the processapparently both due to the action of dissolved calcium and due to theenhanced size-dependent aggregation between the calcite particles andanatase particles (62).

CLOSING RmfARKS

In principle~ bulk as well as selective flocculation and dispersionof almost all minerals should be possible by the control of the ioniccomposition of the pulp, pre-treatment and pre-conditioning of themineral and by the use of properly modified polymers, activators anddispersants. While basic mechanisms governing such aggregation ordispersion in simple mineral systems are fairly well understood, verylittle is known on the behavior of complex natural ore systems. Majorproblems that exist in the area of aggregation, dispersion, and selec-tive aggregation have been discussed in detail by the author recently(63,64,34). It has become very clear that interactions that are ofserious consequences do occur in multi-mineral systems and that it isessential to fully develop our understanding of such interactions,particularly of minerals with dissolved mineral species as well asthose of minerals and reagents with dissolved ions that are naturallypresent in plant water.

ACKNOWLEDGMENT

Support of the Division of Applied Research of the National ScienceFoundation (DAR-79-09295) is gratefully acknowledged.

I/)

0>1/

)a

~P

oo

. n~

=.

~'i.

=

."'=

~

~P

o ~

~P

o.

o.'i

t"'~

'i.

18

.~

~

. ~

. ~

... ..'"

~..'

U.

.~.

. :1

8.

'U

I...

I/)

=~

.

0 It

~~

;t

~

~'.<

It .t"

'.

~...

.0

..

In

=

s:.~

~.

d I~

.~

~

. ~

~<

I..

Po

. 0\

18

~'i~

0 =

...

.. n'i

n -0

...

0\

. .

~

. ~

~...

II

.0

"""=

'i "'I

t.:s

.:s

...

o.~

. W

it..

~:s

s:O

\n.

~

Nit.

II I

'i W

O:

""

~...

I/)C

D

01

I:O

\X'i

II .

II ...

'<

n..

~n .

..~

.-0

...

~~

;..

'5\0

~>

cn'D

l1o

i0-

.0-

0.

fi. 'a

C'a

rtC

S.

...

~'D

O.

OC

Sfi

. .

.. =

'rt

.

rt'a

~. It

..>

='

~11

0

i 'o

j",

c=

' It

.. fi

~

~.

Itrt

~It

..fi

C .~

0

rt

..

Cfi

.

. ()

-ff

:() ~

ICo,

..

. =

' ...~

.~

S.

0.

...

In>

~

~

I~ . .

. C ~

I-.J

'"

:'8. rt .N t»

N ... f rt ... II ~ n . N . ,~ Po

. IS Po II fn n . I';. ... \D ... .., . '0 . lo

t.., ...

N

N

N

N0\

v.

,po

t.-.

.. .

~

>-<

~

0:

1:IG

'" 0

...

"'0'U

'U.

~v.

. :I"

'U

. n~

.. ..

.. O

..

,. .

'U

>

. n

c..

n O

. . c.

. ...

,po

8-

- Po:

l:N

_'"

. ...

1'.

.- .1'. -

Po

I' ~

'<

. =

P

o I'

'oj

.P

o Ic

.. ~

:I:

'U

o'

i...

ID

... :I:

O

i.Oi

.In. . ~

. 0"

'< . o

.'<

""1'

I'

. ..

o'i

..c.

-I

' o'i

1o-

~

=

Po

. n . ... ~

- lo'i 'oj ~ .. Oi

C

. - ~ lo'i ...

. . Oi

I'Oi

1""

-.

.. I~

;C

= ~

. ';

o'i .

~I'

n"":

I"

..10 IG

.. iOi c

0\

...,p

oN

Oi

Po

0 -.

. .

. ~ to.

Oi

Po .

Po

'UO

. ~

I~

~ 'g

1o

Ig

.:

Po

0 ;

. ~

n

. ~

".

CI'

-1

0\ "

" 0

IN

C

.. ...

0\

..

Oi

0\""

...

n -

. :I"

IG

n

n 0\

0.

11

10

01-

~

~1o

. 0

C n . 'U""

...

0 . . . ""U

" .Iv.

,po

0\

0 \0

t.-1'

. 0 ~ . .

...

......

0

\0.

. .

-O~

(')

0 0"

'"g.

)C~

~

O

)C:#

'~O

...ft

"'CD

"".

~~

"~

~

~

...

~~

.--

...

C

I"

" ft

.-

CD

~

~~

p...

Q

- ~

".

..p.

:~

C

D"'-

18>

ft

"ft

ft.

'.c-

...O

~--

.

~

ft~.

- 0'

".:

(')

Pt:l

- >

O

~

CD

':.0

~

CD

'">

CD

0-

"-

DoO

...

t' ~

P

<~

~

~

p~~

-.

...ft

t' ~

>

~00

"0

p~

ft"'O

~C

: ...

CD

88f

t C

Dft

_ft.

...

ft-

" ID"'~

~"

.Q

8CD

PID

".

-.

~(')

~

~

- .

:#'n

.

_Oft

1:#

' 0-

8."

~

t'...

: ".

0-8

.0-I

"'C:

:#'

ft.~

...IN

ID

0

ft .

ft .

'<

- N

10

0 It

n :

IDa

...ft

:#'

ft'"

.00

!P:#

'

...

g.

. II

.0

...~

.

>-

.M

-0

.o~

...

...

0~'"

Ig.

~

..,

pOll

...-I

D""

" "O

~..

. c

"-

0 -

.0.

~...

g. t'

W

Iw f

A 0

...~

w

n 0.

..-

I:#'

-a

~~

~

>

..,

.ft

..,

o~

.o...

~...

, .

II ...

OP

wP

~

...~

0'

.N

)C

...

- 0

...

:PI

N N

~f:;

: ~.

.-'r

t .$.

(n 00<

~

C.

..

(nZ

.

~.

.o<

~O

~t1 ~

(n.

ID t1 ~\O

lD

"'$ N<

. . . t'" :c ~ I?

.

~ '"' 0 ~ ~

n'C

ou=

...

:11=

"""0

0 .

~ID

OlD

..

1'."

'1'"

OU

f)'"

ID..

:s-n

"'l"'1'f)1' P

inP

iCl.

.0O

ID"'I

ID

. I'~

8Q

Pi

':".-

n

->

.8

:S-O

'.an

> c:

Q

.. 8

.

'.1'

0"".

-

Q.

C:.

()

~.

. P

i1

0\

0 t"

">

1""1

0\

"'.{n

SC

l.n

-~O

. II

S:C

...Pi{n- C:

C:

\OP

i"

",f)

...n

O\~

- ~ol

'on

<0

O.

n.

..g~

- ,?

:S-~

,..

.-~

""

1'\o

""~

Q.

\0:

~f)

0 w

>~

o.

...

{n

I.n .

a :s

-o.

.."'I

!0

C:.

. ~

{nf)

0>

0".

co.

f)"'I

C

..~

0

ID!Q

. nP

ing

. O

'C"'I

""P

i Q..

{nID

I' 00

~

- ""

1'S

O

Q.

'O~

~

..0

0 ...

~.,c

- Q

.~

SO

: ""

'0)

...>

Q

. 'C

O

01'.

1Df)

1'

0 P

if),,"

8C

>II'

C

0.-

no.

.:...

I'n0

0 P

in~

:0

'CO

I'

I go

~ID

...

0\

1.11

, ,

0-1"

" ~

(I)

01.

.~

.01

01=

'. .

.. ."

'too'

.

. (I

)~t)

~

..

a .

<

"=I~

~

".to

o'

.....

.

~-'.

D

o~

~

=(1

)~

' ."

'. Z

to

o'

01,

.1

~

"..

.~

...

\Q P

o P

o.\1

:>

.:,.

,Do"

" ; ...

'.

~.

. n 0

~ , ~ , . ~ 9 I (I) 0 t) . ~~

0P

o ...

. '=

'... . ,

N 0-

- -

-\0

~

Q

\.

.. .

~Ix

:~

ot'"

s~~

~tn

.It

It.

~

..P

o.

'tOO

\1

011-

'"'1

-' I-

':E~

.'0

B

rtlt

.'<

.. ...

~.

'0-

n 't

:'t

. 0'

. n'

t-

- rt

>-

't ~

~

OIt..

'tO

O

.. ItB

<

. ~

.~

...s:

B

O.

. -

rt.

o.

.1-'"

ni

x:

"'N0

- It.

O

.Iw

rt.

. n

1-'-

ItO

'~

-s:.

II-'

-..

'<D

o w

""~

0:

-

~.

It P

o ...

~

-:a.

~t'"

.a

. I-

' \0

O.

- Ix

: 0

""~

Po

1-'-

\O.1

t R

on

NI-

'It'<

- Q

\rt.

n -.

S.

w

I-""

'S:

n~It'

t .

rt'<

~

I-'

'Ort

O1

't-

~v

rt.

'O'ts

:.

~It

. rt

.0

'<-

<

. ~

j...

.~

:.

tn.

tnO

\O

~-

~

-O~

n~

\0

.-t..

. -I

-'.

...

II-'~

. -

N"'v

.

tn

- .

. a.

v

rt

-rtt'

"n

- ,:

c ~

0

..i

t'" >

Iw P

o . n

-I-

'=t

...a.

N

'" .

1-'.

.0.

-.

1-':

0 s:

0 .

tn...

~

...'t

- ns

:a.

~

Po'

O

\00

~'t

It rt

"""'"

",tn

't 0

..n~

~

O.tn

.n

rt~

rt

.

.rt

. '0

. .

't~.'t

~

~n

"'1-'

...

Po'

t\- .

. 0

- n

0\0

nn

o.

""

'd-

0 Itn

Q

\I-'

~.

...

-S:.I

-' rt

't-

I-'

'<

Itrt

\O~

~.

't...

Q\

Itrt

tn

~O

<

...

rt

..-

n .

0 .

nn

~

~

...0

'Os:

~

...

.~

C

" ~

~

. 'ti

t ...

>

Wrt

It.

"

~

N'"

n .

I-'~

W

0 IR

o rt

a.

.

..~

...

nl

-'-

0 ~

".

..~

~

0

't~i

. ::<

Y

'

,

... ""

c..n

lJl

. ;,0

..

n"

0.- r0

-O

0

. n~

.

=-

H'<

~~

~r't

"0..

0 .

~~

~

. ..

t1~

OO

-

.~>

0 nM'~

'.:

.- ~

.I~

~

=

'N

~

- ~

0...

CC

\00'

"..,

rit1

"'.'"

- 00

- .O

OM

'c..

00

. r't

~..

.N

-

""... o=

' :

~>

;IQ

\oO

"

:--i~

00 ;~r't

0 0 0

I-'H

\C r't

N8

a. -

s "" Z m "V >- ~ n r- m cn "V ~ 0 n m cn cn Z C

)

"V ~ z (') ::u r- m cn 0 "" "" r- 0 (') (') c E s z c u; "V m ~ cn S z cn m r- ~ ~ <:

m "" r- 0 (') (') c ~ ~ s Z G ... W

N \II N . .0 N . '0 '0 W .p-

m I W \II N

"" ." ... \0 0- ~ ~ ~ 0- '" \0 .!,.

ao ao

w

oo~

?;.~

.:1'

.""-

~I~

o.0

-- :Q

~w

'~

"-

~

~

...0

. 0

W"

0--

0 ~ ... ... " - " § P

o

~ ... ... Po

§ ~ Po

. " ~ " 0 . 0 Ii' - N

~

~

~

!~

~

~.o

~1D

01

~1D

0.

~

.O

\~D

o ""

Do

w

It w

..~

..

"oa

>~

>

0 (I

).

0 ~

ItN

O1~

~

~O

. It

.N

~

. n.

. ID

. ...

n:s

It-.

~

=0

Do

. <

0-

1

<

w.

~It

=

It01

0

I.-os

~~

w

~

C:

~

o.O

~

~nl

t...

n \O

n~

~

O-O

Sw

N

~'"

~

w.

~n

"=..

oa~

~.

0 ~

~0-

1 :s

~..

NO

1w

\llfA

~"

"'80

~=

.o

a~nf

A

.It~

It O

1nIt~

.~

.. "~

. n

~.

I;~

3;..I

t O

lt~

...

~. -

os

"~

.~

IfA

0

":lg

~

~. ~

:s

0I-

" ID

I-"

~.

01~

~ "

. .

to"

~n

; 3

~~

~~

1~0-

I M

lw

II

"n

~~

"D

oI

It It

. .

"" ...,

u. 0\V

i V

iV

i ~

- -

t"0~

...

:C~

(')

~...

~

000"

""

. ..,

~O

.H

rn.

Q\t"

n=~

n.

- ...

n.O

Dl~

t"

=-

~ ~

t" ~

n-

-~.

O~

n(

')'i.

H

...~

t"-

..-

..,00

- ~

~~

. 'o

C-

It~(')

.0

~~

. no

-.>

,. ~

'1

-H

o cn

...

'oC

(')-

t"n. n.o

!->

t"~

~

g-~

DIn

- t"

t"~

n~

~n'

"t"

~

..,

.o~

. '1

n~

. ~

t"0

- ~

c..

"'n~

:t"-

0"-

<-

n10-

1 .

n t"

tII

~nc

.. =

~-

.t"~

...

~<

- .:c

>...

~

t"-

-..

n"~

-

n.I

>.

~

: =

~

H-

~~

~

i :

Ii 3

~~

:2

- c~

...~

; t"

~~

.. '1

Ic

.. n

-~

-

~-

-.

c..

. ~

:(

.0-

...~

- ~

=O

On

""D

lc..

Q\.c

n -

'1-

00

- n-

-.

~N

~Z

'1

:~

c.-

Odc

n-

- In

-

.~

-

t"'"

~

: -

~It

~O

o-l

.n-

n ~

""

n

nit

- D

lt"=

<

"'d

...

-~

..

~

~

~

~

\.0\.0

N

I-

' 0

-0-

- -

- -

I-'

. ~

Iz:

~

1>

<.

>-

"~

oa

(I)

(I)

-or't

~

~>

-r't

r't

.""

C'<

O1

0 r't

DoO

1 r't

01r't

0"

O

"~tD

0-""

'. -.

0 ~

I-

'~

I-'r'

t..

a 08

a"

'. ...

~r't

~-

1~

01

tr1""

"'-

~.

".'<

oa...

co

atD

~

o "'1

-' .r

't.-"

n 01

r't01

r't

O1O

<

r'ta

r't'"

r't-

~

- -

.. ...

"C)

\.oc-

o.

r't

.

0-.."

01

:\

(1)

~I~

~

>

- :

>-

~

~~

. r't

~

~

..-

- ...

n\.0

.;t

. 0

- I-

'\

~

-

- .~

...

- (I

).

~(I

) ~

tD.

. ~

~

..~

~

- 01

r't>

- -

~-

I~

I-'~

-.

I D

o .

0-

01.~

1-'~

..I

I-'

0""

\i'.

r't_~

0 0-

'<,0

"r

't'"

~

--~

01

N.C

r't

~

",

(1)

ooa

...

<"

- -

Ic.."

I...

r't

r't.

~.~

-

- r't

0 tD

~

C

~

0 -0

01

C

~

'01 ~

19

01I~

01

.I~

"'.

. ...

"",

~..c

r-

-0.

0100

1 S

'" ~

.~...

0 .0

-"

. -

0 .

~

.0'"

~>

-."

~.

01C

::.C

oa

-

cc..

- .W

O1

...~ -

C) n

~ ~

o I-

'~..

..,,~

~

~...

.."

.-

~r't

>-

- -

~~

. '<

~

01

~a

-.

...0

...

o~~

rtr't

-

~

.~:

-OJ

"0-

-'<

0 01

tj

"'"V

..~

.01

01

~~

...

OtD

"'0

W

'"~

<

a.

..:.~

1-

'01

tD

'<

"",,(

1)"

. ...

n.

.. oa

tD

~

tD

.

III

"-00

0 ~

....

~

.tj

.."oa

~

. -

tD:0

-

""~

"

~"

...

~

~r't

or

'tIc

.. '<

~

(I

) 01

~

0

01

I-'

...-

~-

oa

r't"'0

'«

,..

-~

8"~

. >

-.

~O

'<

tD

~.

...~

...

~...

r't

01

'01

r't

~~

N

. ..

art!

8 \.o

O1n

.0

-'<

...~

\.0

.0

>-

:"...

~

. -

1-'.

Do

"I

n .0

1 .o

a m

.C

~

.pC

" .1

-' 0

"

! I

:..

:..

01

"t'

gr't

01

" oa

Itr

't-

~

Ic...

I~

...

r't

oa

...

.. -

01~

'" r't

OI

"I

~

o~

0 ,,~

..,,0

1 ~

- "

~

~

0'<

"

- ~

0

W(I

) 0

0"

I -

- I-

' '<

0

. '<

01

.

VI

VI

VI

'" N

I-

'.

. .

I>

0

n rt

~

:E

0<

0 ~

p.0

rt.

~'1

In~

..~

lip

...

'1'1

'1.

=

..I

I-'

rt

>

I-'

\0.

CI)

I~.

. ~

"' '1

..

.~

.0

=~

~

. .

I-' C

I)

. .

.I~

B

.

~~

~I. i

rl.

S"

. n.

D

' 1-

'. "

""'C

I).II

n

. . n

n

liD

' D

' ~

rt

tI

0-

0 c:

10

...

~

\O~

I-'

. <

..

'-I"

. '1

. =

. I-

'1'-1

.i

\O'O

S

\0~

'-I

I-'

. c..

. I-

' 0

."

. n

I-'

""O

S

n \0

>".

."

C:

'-I.

.. .

I-'

o.p.

..

.~

n"

=~

~

N'"

"'CI)

~

~o'

.t"

.

~

I-'

.0...

n.

.c:

~

rn

Nn

.~

. "'"

'1

'1

."

rtO

.

<

..-

'1

II.

.rn

. "

'OS

c:

...

I-'

.tIS

0

0."

. ~

n

II .

n~

S:

c:..

I-'

..,.

0 .

0 <

rt

~=

0

....~

I-

' 0

.D'

<

~=

tI ... ~

0

I~

tIS

~

...

'1'" ~...

=

5 .

..-

n g

...

D!

"'OV

(qfn

:lOtiO

>g.

..1~

.

53::5

~

OtP

o'U

fn

Ot:l

Ot

~...

~

-:I

:I()

'OS

.0.

..0.~

='U

...~

~.

CltP

o.B

.

- ...

:fn

"""'O

S..0

00O

tOt:l

-...

.

.ovn

-.~

O.

'< o

. ~ Po

()

Ot

.1t'O

S

~:I~

()

0fn

lt...

- .

~O

rt _:I'" ~ :

8:::

.;i

..

0->

~

0

. O

t n

:'<

Ot

. ...

~

rt\Q

OV

'"...

Oto

O=

.()I

t

rt'"

:0

()

o-It~

Itt9

~P

o(q

=

-0- ~ ... ~ =

s- It Po

t .

Vt

~

~

~

~0

\0

m

...

0-.

. .

. .

~

~

o.s d

cn

a:

... ~

cn

~,.

I~ cn

tt.

n.

N""

'" ...

tt V

tO.

~O

OO

O

Ott I

.».tt

..

- ~

r+~

,.,

'0

~

a...

-..

Q

n C

8

0~.a~

aa ,.8

,... "

,., tt

- ...

~II.»

0-

- -

\0

,.,...

~,.,

~

- m

_.

: 0-

000-

1.

0 ~

~

,.,

~...

I...

a ""

4.o.~...

.~ ,"",o.~

~. -

.-

I. \0

-

0.0

. "'0

..

'Orn

o-l

'O"'O

~

oc..

,.,~

~

.~.

. ...

~

0..

...00

.~P

o ...

~'O

~

~

~-

\0"

~

\00.

.~~

.. ~

o

~

0 ~

M~

Ot~

8:

...

.

. C

iDO

'" rn

-

cn

~.~

-0

n ~

0'

" C

.In

0

.l:c

n-

...~

'O

0-~

:1'~

"'0

-1...

~C

.8

..Po

".

.I.»

~"'~

.~

,.,

8-

.-

,.,

N."

. ,.,

.tt~

- ..n

~

~

- c.

. D

,.,

0 ...

n~

. .D

o\0

... .:c

cn

. ..

~I~

-0-

. .

'4'"

: ~

O<

m-

- O

t .

~

- -

>

~ 0

.. :~

.,.,D

:a

D

o ,0

.o.

o-.-

~

~:1

'I

8-I

\00

~~

~

o.s~

-

f:c

Vt..

. po

0.

...O

~:.

0..

Vt'O

I.

...,-

=

' ..-

-,

., cn

-,.,

-1

0-1

0.

Ot:

~

0

:1'..

.>O

t ~

>

'00

...0-

1

o-..c

n 0.

. ~

~

.=

':1'

o.s.

P

o-=

' .

0...0

I,.,

... ~

.

C

m...

...

:.,.

, O

t .'0

0

~0-

1,.

.. ,."

., \0

~

,.,

:1

'I~

..

='

:z ~

~

...

..

..=

,.o.~

...

f)0

. O

t to;

-0

.='

8 ...

. O

~

f)

- ,.,

....

:1'

mf)

0

if)

8 o.

s ~

1.

»'"

.0

lO

t 0-

1 ~

1m

-

Ot

0 :1

' o.

s O

t :'0

f).-

...

,.,

0 .

I,.,

. ""

IO""

~""

to

Ot"

'tt~

...O

"'

I0

.0~

. V

t='

- =

' =

'oc

=-

- ~

~.

~

00..:

'O

~

\0.

~~

...tt

rn.

0-

Ot

.~.

r+

...0

0-

o.s

-=""

n:

1'

-='

0 .0

. \0

" =

'I

~

f) \0

'0 rn~

,.,..

f)~

~

...

.

~

~yO

~

...O

~

(1)=

\D""

" 00

yO

~

...1:

\Dr-

'...

~.

.00

..01

~01

...

c.

~0I

I!'-3

.n

I C

)...

~t.,

(I

).w

.. 0

o\~

(I)

...~

. .

Doc

.~

.0

H:C

...t.,

~

n1

01

..n.

.. 0

'1

~'1

. .~ ~

n.

..

0 ..

~

C)

..

~n.

It.

,...

.,

'1'"

0-.

~=

0'1

~>

~.

c. . .~

>

. 0 '1

~:c

l&

l ~

,...

....

.. .

3

It.,

~

..

~

0...

,~

.. "'-

~...

.'w

. 0

.. .c;

\D~

10

II ~

I

~g

. '1

a"

'0.

. >

- ..

.c.

rt.

N1

0\

0 \D

rt

1.1'

1 ~

.'0

~..

rt

.... 0 ~

. ~ ." Z m 'V . ~ -4 n ... m cn 'V ~ 0 (') UI

cn z C) "V ~ z (') ::;

:r m 0 ." ." r 0 (') (') c S 0 ~ c ~ "V m ~ rn 0 z rn m r m (') -4 '< m ." r 0 (') (') C ~ -4 0 Z ~ U

I

I~~

~

~i~

I~: ::

I~;~

!!:t

.~:,

;!& oc-

,~

"':1 !~

~'"'

i!

~~

,? Ii' \~~ I"

~ . ~ W 0\

O",

oi""

""'.

.II~

,<O

"g..

"~""

..O>

~...

.0

. nN

Q,C

.

C..O

.O""

.. ~

.O~

.. "~

~.

~8

"18

Q,"

Q,..

C

.. "'N

.

"'~n.

.o-.

IIP.p

. ~

. ~

Q,

..~

~

~

PO

""P

"'",

."'II

~

"'11

<"

8"",

.

..0

.0

O"~

.~...

n ~

~

~.."

.."""

""~

no...

..<.."

~~

~

.. .Q

,""

"'Q,

0 ...

~...

0.

~~

...~

--

....

Q,"

n 0

~.."

~=

."

Z

~"'C

""...

n "t

o..

....

~O

"~..

n~~

.0

"".o

~...

.~

...~

...o

."

~

~... O

.O"~

~

..

..~

'U~

Q,.

"~II

" 'U

..=

"..~

.Z~

O.~

~.

.Q,"

8.

"n..~

~..<

S'<

n.

. 8

IIn

",<

.~

~

."O

~

~~

P

.. ~

O.O

"..~

.~~

~

~

.~

~

~

"~

-..

~C

~.

gNO< ..~

. ~

~.II

.0...

""

...Q

,~

~...

~...

"""~

~.n

"C

.

n .O

'U...

0 '<

C

. .

n0

I or

.0.n

"~~

~~

n.".

~"~

n'a'

a~0

.~~

.~

iP

N~

O"~

..~~

o~~

~~

~~

~P

. "

~~

..~.O

~C

~..

.."

~

c.~

~

~"P

C~

p~"~

an"

.~~

-..

~~

.~

""~

~.

0"

.P~

E."

cn~

w

""

..~~

op~

~~

" ~

...~

~O

.~P

''l

~~

"<

. ~

O

.P.

I.~

n~

o 'a

"0

._"c

.~~

~"~

"~

~r

" ~

" R

P

.. .~

~.~

".

~

~

O.O

~

"O~

..n~

..~"

,~

. ~

rE"~

~"

'~.~

~o

.~~

.P

f~

.~.

".p~

c'"'

.u

..0...

.~

"~C

~P

f"~

.. c.

"~

'a".

D~

~..~

~

.. n.

o~~

. D

~.~

.~.~

'U.

('a

O"P

f ~

'a~

'a~

. ~

...P

f..~

.",

.. P

f~

.I

~~

".

~~

c.!,

nPf"

t:'"g

.D

~

.~

~P

f ~

D.O

s.

..gig

o."0

.D.~

.~

'R.'

C"C

o ~

~."

",

.oiC

O

~

""~

.""

1 ~

~~

C"

.~

~~

~O

O~

~

."'O

~oi

.~"~

O~

C

o~

"C~

~

..D~

O~

.g

o"",

~"~

i~

."

" ...

"'..

.~.o

..

~.

~~

.

.Pf"

.~..~

~~

I'a.p

.~~

c~~

~.~

".

~~

'O".

Pf~

'O~

Q\.O

~

..0t

en

" C

P

f.'

'a"'~

.-

~

c~.

0oa

Pf~

"n~

Pfw

. go

e.

..~~

~

".

"'~

"'~"~

Ol

.~"~

".R

~

0...

~~

.C".

O'O

"'~P

f. ..'

a0

"P.

D~

D~

CP

f~~

.O~

I~

r~

~O

B."

.'"

'~"~

"C~

,",~

0

~~

.~.~

~.-

~

.~C

~

~

~.

"Cr

"Pf"

""

O"O

~

D~

.a~

~~

...O

. .

~~

.=c

~

" .C

~O

0

.~

...C

"-

~"

f.c

"<

.B~

~~

~

.. ..

"'P~

-

.O~

.~

.~"

Dn.

.O

~"

.~ID

~..n

,"

"~.O

~"O

nC

~~

O

"O~

".~

~.,.

. .."

00.C

~1

~~

0

..~

p"""

~"~

Pf.P

~

""a

."~

~

~

~...

D~

."

nDD

.a..

~'n

.op

~".

~~

DO

"~~

O~

O

"'~

Pf.

..0"'"

, ~

"~.~

D

D

.fn.~

. ..~

...".

.~D

.~.O

t<".

"'"

D

C~

C

~O

D~

C"~

"C

.~

.0.~

..nO

R.

,..~

".

~~

.<.

I.

..."~

~

n.~

.~go

..~

~~

"'~

ID

~

n.O

"."'~

Pf~

"

. ~

O"O

..

~

. .a

~.

~~

C...

~~

~.~

~

~

~~

.'aC

"C~

"~'a

".

.~01

00~

~

-~

C~

. ."

't

o .~

'a~

i""

""a.

~~

n..,

"~~

...

~

~O

O~

~"

o~

~~

"'C:

11

c:.

~'""

~n.

.goo

"a

..noo

i.'G

"

~ .I.

~i

~'a

.

~ "

.

0 ~

.

.. ~

go-

. C

~

~

~

.

~

. 0

~oi

. O

~Z

C~

~.~

~~

"D

C~

.rt"

_.

.~

."P

I~~

~",

..".

~C

O~

. .

>M

~

'a~

~

...'

--c

o "~

~"n

_"D

"..

~.o

- c

.. R

~..~

D

z~..l

i~"C

DO

n .'a

rt

~

I~~

~.o

. .~

'O"n

l'"

.. .

"-

.g~

..".<

. ..~

~

~

. ~

"~

a..

1 .1

.~

~

. ..-

-XC

I)~

"'i

. ~ . . - P

i. . c

~"'8

....

~.

Jg

S.

...,,- . . .~

d 0 ~

...

-

<-

ow!

. ...

- C

A~

~i:

O\c

Pt

- ~

ft

~...

ou.n

".

Pi..

.

. .

~.

N-

~

Oi

. ~

., . .

- .

~

i~

~ft

..,- C

:

CD

Pt~

. .

ft. . It

~

.

. P

i

fSo ?I[

..:i.

- IS

'0- ..

rn~

"'I

. n

. ...

. ..

I~~i il'~ .

10. M

ol'

...

.

oo:C ...~

~00

1'.

. ... '1 .

:~ .,

I~

~

~D

I'

n ..

...

.oa

. .

rn.

0 I~ . ~

rn

D.

~=

-

., n

. rt

:<

.

~ ~E

8... .

~.

a.~

:::-8

.. '1

:

Zl

~

. .. I' ,0- W

a .. '0 . ., ! ...a r

;'o Po.

..

...

fft9

!'-~

!M

g'.

io~

'<.

.. 0 m

c

N~

"0-

""'"

o~ 0... ..

f ... . c ., ~

:ot

.n i i ~ ~ ~ Po

0 -a . Po

~ '< . '<

0\

0\

00

UI

UI

N

0 \D

~

. .

. .

.

0 C

C

aN

>-.

.I

n N

~

.

, ~

_.

a'i

~

.c~

rt

. G

oGo

., rt

o-n~

.,.

~

"'l

. ~

.~

_

I..,

. 00

. ...

~~

.'<

0<i

.~

.. ~

.,.

~

~

oc..

: .N

IW

. n

=

. \D

rtc.

. .

rt.

,. >

-:c.

o.

.c

0 .

c.

...~

.,.

~.

~a.

0-.

rt-

\D.

rt"

II.

~...

0\

0 -

I.

a'

. ., "

c.

n .~

'0

~

-!

- rt

"'r

tC

0.a.

0

~n

n0-

0.

~O

0~

.tJ

\D

~

~

O

C

...~

00

0.

~i

.. .,~

\D

~-

\D~

. 8

~C

...

.

=--

..,

.~

0-

. =

It

.-a.

~

~.,n

.

..i

c...

~

:c...

.,.

Z

n .

. ro

. ""

.Nl

o-I

~.~

00

.,>

--

-.

.a

0-.

..:.

c..

>-

~

.,.

...~

-.

..

10-

1:

a.

a.

Go

~

'...i

. .,0

~

~

.. 0,

. H

~

~ =

- ;

g,

I '2

Ii

i.

n ~

a.

~~

~

C4 ..

.~,

rtIi

~<~

~

r-

.C

.G

o...~

0.

..~

.

~..

o~~

.'C

00

~

rt

c..,.

~

~

0 0.

..0

of

.. ,.~

~

O...

IN

0

: .,

- '"'

~

n ~

;.

In

~

8I-

~.

- .

..~.

., ~

~

N

. 0

., ..

.to!

-.

...

\DO

~

- ..

n.

00 .

.. :

rt.,

N.

: ~

...

rt

Ic...

C4

In

...i

...

<~

.. .~

.~

...

~

O

~

~.-

0 .

~In.

~

n

~..

~.,~

. ~

~

w

o.

.. .

~

.0rt

~

n rt

n

.. \D

..

Ic...

n

no.

\D

C~

""

.

~ B cn cn ~ ~ 0 'OS

~ ~ 8 ~ I 'OS

~ ~ cn

"" Z III 'V >- ~ (5 r- III (It 'V ~ 0 n III ~ Z Q!

henry krumb school of mines columbia university new york. ny …ps24/pdfs/principles of...

Documents