WATER POLLUTION CONTROL RESEARCH SERIES
14010 DRB 05/71
     Flocculation and Clarification
                   of
          Mineral Suspensions
ENVIRONMENTAL PROTECTION AGENCY • WATER QUALITY OFFICE

-------
       WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Reports describe
the results and progress in the control and abatement
of pollution in our Nation's waters.  They provide a
central source of information on the research, develop-
ment, and demonstration activities in the Environmental
Protection Agency, Water Quality Office, through inhouse
research and grants and contracts with Federal, State,
and local agencies, research institutions, and industrial
organizations.

A tripilicate abstract card sheet is included in the
report to facilitate information retrieval.  Space is
provided on the card for the user's accession number
and for additional uniterms.

Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Head, Project Reports
System, Office of Research and Development, Environmental
Protection Agency, Water Quality Office, Washington, D. C.
20242.

-------
Flocculation and  Clarification of Mineral Suspensions
                                by
                 Mineral  Resources Research  Center
                      University of Minnesota
                  Minneapolis, Minnesota   55455
                              for the

                   ENVIRONMENTAL PROTECTION AGENCY

                         WATER QUALITY  OFFICE
                        Grant  No.  14010 DRB
                             May  1971
  For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $1.25
                           Stock Number 6501-0055

-------
               EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents neces-
sarily reflect the views and policies of the Environ
mental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement
or recommendation for use.

-------
                        ABSTRACT
     A study of the flocculation and clarification of quartz
and goethite suspensions using starch and calcium chloride as
flocculants was made by determining the settling rates of the
suspensions, the amount of suspended solids in the supernatant
liquid, and the concentrations of the residual starch and cal-
cium ion.  These determinations, supplemented with measure-
ments of streaming potential, adsorption density, and visco-
sity, were used to formulate a comprehensive expression of the
mechanism involved in the use of starch flocculants.   No par-
ticular difficulties were experienced either in the floccula-
tion or in the clarification of goethite suspensions.  Quartz
suspensions, particularly when aged in alkaline solutions,
were extremely difficult to clarify.  A cationic starch of a
certain molecular size and degree of substitution or a com-
bination of a polymeric compound and a cation capable of re-
versing the charge of the quartz showed some promise in the
clarification of quartz suspensions.  The maximum settling
rates and the minimum turbidity of the supernatant liquid for
a given condition were observed at a point where an excess of
starch began to appear in the solution.  The most effective
flocculation and clarification condition may be observed when
a polymeric compound is adsorbed uniformly in sufficient quan-
tity and in a stretched-out conformation, and when the zeta
potential of the resulting suspension is made as close to zero
as possible.  This report was submitted in fulfillment of Grant
No. 14010 DRB from the Federal Water Quality Administration
to the University of Minnesota.

-------
                                CONTENTS


Section

  I          SUMMARY AND CONCLUSIONS

  II         RECOMMENDATION                                          xviii

  III        INTRODUCTION                                             1

  IV         EXPERIMENTAL MATERIALS                                   3

                  Quartz                                              3
                  Goethite                                            6
                  Starches                                            7
                  Synthetic Polymers                                  7
                  Dodecylammonium Chloride                            7
                  Chemical Reagents                                   8
                  Water                                               8

  V          EXPERIMENTAL PROCEDURE                                   9

                  Preparation of Starch Solution                      9
                  Preparation of Synthetic Polymer Solutions          9
                  Flocculation Test Procedure                         9
                  Filtration Test Procedure                          10
                  Analytical Methods                                 12
                     Determination of Starch                         12
                     Determination of Calcium Ion                    12
                     Determination of Dodecylammonium Chloride       12
                  Zeta-Potential Measurements                        15
                     Streaming-Potential Measurements                15
                     Electrophoretic Mobility Measurements           15
                  Viscosity Measurements                             19

  VI         EXPERIMENTAL RESULTS                                    20

               Flocculation Behavior with Starches                   20
                  Effect of Corn Starch, Calcium Chloride and pH     20
                     Flocculation Tests on Quartz                    20
                     Flocculation Tests on Goethite                  24
                     Filtration Tests                                31
                  Effect of Size Distribution and Pulp Density       37
                  Effect of Temperature                              43
                  Effect of Chemical Modification of Starch          43
                     Effect of Cationic Starch on Quartz             48
                     Effect of Cationic Starch on Goethite           53
                     Effect of Anionic Starch on Quartz              53
                     Effect of Anionic Starch on Goethite            53
                  Effect of Dodecylammonium or Aluminum Ion          53
               Flocculation Tests Using Synthetic Polymers           57
                                   VII

-------
Section

               Flocculation Tests on an Artificial Mixture of
                Quartz and Goethite                                 65
                  Total Flocculation - Clarification                65
                  Selective Flocculation - Desliming                65
               Zeta-Potential Measurements                          71
                  Zeta Potential of Goethite                        71
                  Zeta Potential of Quartz                          74
               Viscosity Measurements                               83

  VII        DISCUSSIONS                                            90

                  Properties of Starch Dispersions                  90
                  Surface Conditions of Quartz and Goethite
                   in Water                                         92
                  Adsorption Mechanism of Starches                  93
                  Flocculation Mechanism with Starches              98
                  Physical Factors Affecting Floe Formation        101
                  Flocculation Behavior of a Mixture of Minerals   102
  VIII       ACKNOWLEDGEMENTS

  IX         REFERENCES                                            105
                                   Vlll

-------
                                 FIGURES
No.                                                                 Page

 1           (a) Size Distribution of Quartz as a Function
             of Grinding Time,  (b) Size Distribution of
             Goethite as a Function of Grinding Time                  4

 2           Grinding Time - Size Modulus Relationships
             for Quartz and Goethite                                  5

 3           (a) Settling Curves of Quartz with Corn Starch 0.09
             Ib per ton, Calcium Chloride 3.65 Ib per ton, and
             pH 11.55,  (b) Settling Curve of Quartz Using a
             Photometer                                              11

 4           Filtration Test Data of Quartz and Goethite with
             Corn Starch 0.5 Ib per ton, Calcium Chloride 2.0
             Ib per ton, and pH 9.5                                  13

 5           Calibration Curve for the Colorimetric Analysis
             of Corn Starch                                          14

 6           Calibration Curve for the Colorimetric Analysis
             of Calcium Ion                                          14

 7           Calibration Curve for the Colorimetric Analysis
             of Dodecylammonium Chloride                             16

 8           (a) Schematic Diagram of Streaming Potential Cell,
             (b) Streaming Potential of Goethite as a Function
             of Pressure                                             17

 9           Apparatus for Electrophoresis Mobility Measurements     18

10           Response Contours of Quartz Flocculation Tests          23

11           Flocculation Test Results on Quartz as Functions
             of Starch Addition at pH 9.5 and Calcium
             Chloride Addition of 2 Ib per ton                       25

12           The Effect of pH on the Flocculation Test Results on
             Quartz as a Function of Corn Starch Addition; Calcium
             Chloride 2 Ib per ton                                   26

13           Response Contours of Quartz Flocculation Tests          28

14           Response Contours of Goethite Flocculation Tests        30

-------
No.

15           Flocculation Test Results on Goethite as
             Functions of Starch Addition at pH 9.5 and
             Calcium Chloride Addition of 2 Ib per ton               32

16           Response Contours of Quartz Filtration Tests            35

17           Response Contours of Goethite Filtration Tests          36

18           Adsorption of Corn Starch on Gbethite and
             Quartz of Different Size Distributions at
             pH 9.5 and with 2 Ib of Calcium Chloride per ton        38

19           Settling Rates of Goethite and Quartz of
             Different Size Distributions as a Function
             of Amount of Starch Addition                            39

20           Response Contours of Flocculation Test Results
             on Goethite as a Function of Size Modulus and
             Concentration of Solids                                 44

21           Response Contours of Flocculation Test Results
             on Quartz as a Function of Size Modulus and
             Concentration of Solids                                 45

22           Flocculation Test Results on Goethite as a
             Function of Pulp Temperature (pH 9.5, CaCl  2 Ib
             per ton, and Corn Starch 0.075 Ib pei ton)              46

23           Flocculation Test Results on Quartz as a Function
             of Pulp Temperature (pH 9.5, CaCl  2 Ib per ton,
             and Com Starch 0.075 Ib per ton)                       47

24           Response Contours of Flocculation Test Results
             on Quartz as a Function of Amount of Cationic
             Starch (0.067 D.S.) and of pH                           50

25           The Effect of pH on the Flocculation Test
             Results on Quartz as a Function of 0.067 D.S.
             Cationic Starch Addition                                51

26           Graphs Illustrating the Settling Rate and the
             Amount of Suspended Solids in Supernatant
             Water of Quartz as Functions of the Levels
             of Two Different Cationic Starches at pH 9.5            52

27           Flocculation Test Results on Goethite as a
             Function of 0.067 D.S. Cationic Starch
             Addition at pH 9.5                                      54

-------
No.

28           Flocculation Test Results on Quartz as
             Functions of 0.93 D.S. Carboxymethyl
             Starch Addition at pH 9.5; Calcium
             Chloride 2 Ib per ton                                   55

29           Flocculation Test Results on Quartz as a
             Function of Calcium Chloride Addition at
             pH 9.5; 0.93 D.S. Carboxymethyl Starch
             0.05 Ib per ton                                         56

30           Flocculation Test Results on Quartz as a
             Function of Dodecylammonium Chloride
             Addition at pH 9.5; Unmodified Corn Starch
             0.075  Ib per ton                                        58

31           Flocculation Test Results on Quartz as a
             Function of Synthetic Polymer Addition at
             pH 9.5                                                  59

32           Flocculation Test Results on Quartz as a
             Function of Synthetic Polymer Addition at
             pH 9.5                                                  61

33           Flocculation Test Results on Quartz as a
             Function of Synthetic Polymer Addition at
             pH 9.5; Calcium Chloride 2 Ib per ton                   62

34           The  Effect of pH on the Flocculation Test
             Results on Quartz as a Function of
             Separan NP-10 Addition                                  63

35           The  Effect of pH on the Flocculation Test
             Results on Quartz as a Function of
             Separan NP-10 Addition; Calcium Chloride
             2 Ib per ton                                            64

36           Flocculation Test Results on the Artificial
             Mixture of Goethite and Quartz in various
             Proportions                                             66

37           Selective Flocculation and Partial Upgrading
             Results on an Artificial Mixture of Goethite
             and  Quartz; (a) Effect of pH at a Constant
             Starch Level of 0.375 Ib per ton, (b) Effect
             of the Level of Starch Addition at pH 9.5               68

38           Zeta Potential of Goethite as a Function of
             Corn Starch Concentration                               72

-------
No.
39           Zeta Potential of Goethite as a Function of
             pH in the Absence and in the Presence of 50
             mg of Corn Starch per liter                             73

40           Zeta Potential of Goethite as a Function of
             Calcium Chloride Concentration in the Absence
             and in the Presence of 50 mg of Corn Starch
             per liter                                               75

41           Comparison of Zeta Potentials of Various
             Quartz Samples as a Function of pH in
             Distilled and Demineralized Water                       76

42           Electrophoretic Mobility and Zeta Potential
             of a Ground St. Peter Sand and a Brazilian
             Quartz as a Function of pH                              78

43           Zeta Potential of Brazilian Quartz as a
             Function of Corn Starch Concentration                   80

44           Zeta Potential of Quartz as a Function of
             pH in the Absence and in the Presence of
             50 mg of Corn Starch per liter                          81

45           Zeta Potential of Brazilian Quartz as a
             Function of Calcium Chloride Concentration              82

46           Zeta Potential of Brazilian Quartz as a
             Function of Calcium Chloride Concentration
             in the Presence of 50 mg of Corn Starch
             per liter                                               84

47           Zeta Potential of Brazilian Quartz as a
             Function of 0.067 D.S. Cationic Starch at
             pH 5.0, 9.5, and 11.2                                   85

48           (a) Relative Viscosity of 0.4-Percent Corn
             Starch Solution as a Function of pH,
             (b) Relative Viscosity of 0.4-Percent Corn
             Starch Solution as a Function of Calcium
             Chloride Concentration                                  86

49           (a) Flocculation Test Results on Quartz as a
             Function of Starch Addition at Different
             Concentrations of Calcium Chloride at pH 7
             and pH 11, (b) Maximum Settling Rate of Quartz
             as a Function of Calcium Chloride Concentration         88
                                    XI1

-------
No.                                                                 Page

50           (a) Flocculation Test Results on Goethite
             as a Function of Starch Addition at Different
             Concentrations of Calcium Chloride at pH 7
             and 11 (b) Maximum Settling Rate of Goethite
             as a Function of Calcium Chloride Concentration         89

51           (a) Adsorption of Corn Starch on Hematite and
             Quartz, (b) Adsorption of Cationic Starch on
             Hematite  and Quartz                                     94

52           A Model Showing Mode of Adsorption by Various
             Starches  on Oxide Mineral Surfaces                      96

53           Schematic Representation of the Three Limiting
             Cases of  Conformation of Adsorbed Starch
             Molecules                                              100
                                    Xlll

-------
                                 TABLES
No.                                                                 Page

 1           Composite Design and Product Data for
             Flocculation Tests on Quartz                            21

 2           Composite Design and Product Data for
             Flocculation Tests on Quartz                            27

 3           Composite Design and Product Data for
             Flocculation Tests on Goethite                          29

 4           Composite Design and Product Data for
             Filtration Tests on Quartz                              33

 5           Composite Design and Product Data for
             Filtration Tests on Goethite                            34

 6           Hexagonal Design and Product Data for
             Flocculation Tests on Goethite                          41

 7           Hexagonal Design and Product Data for
             Flocculation Tests on Quartz                            42

 8           Hexagonal Design and Product Data for
             Flocculation Tests on Quartz                            49

 9           Selective Flocculation and Partial Upgrading
             Results on an Artificial Mixture of Goethite
             and Quartz                                              69
                                   xav

-------
                             SECTION I
                      SUMMARY AND CONCLUSIONS
     Flocculation and clarification behaviors of quartz and goethite sus-
pensions having identical size distribution were studied by determining
the settling rates, the amounts of suspended solids in supernatant liquid,
and the residual calcium and starch concentrations using a combination of
calcium chloride and corn starch.  The effects of the size distribution
and the pulp dilution of the minerals, of the temperature, and of the chem-
ical modification of the starch structure were also investigated.   Stream-
ing-potential measurements, correlated with the adsorption data of starches
on minerals and with the viscosity measurements on starch solutions, clari-
fied the role each additive plays in flocculation.  The information thus
gathered was utilized also to the selective flocculation of a mixture of
quartz and goethite.

     The following mechanism of flocculation using starches has been de-
duced from the experimental results and from the previous knowledge ap-
pearing in literature and used in the discussion of the results.  Floc-
culation of a mineral suspension with a starch is effected by bridging
of particles by the adsorbed starch molecules.  The adsorption density
and the conformation of the adsorbed starch governs the flocculation be-
havior.  The adsorption of starch molecules on oxide minerals results
from hydrogen bonding, but is strongly influenced by the electrostatic
interaction between the charged functional groups in a starch molecule
and the charged mineral surface.  The electrical double-layer theory in
conjunction with the adsorption measurements and the streaming-potential
measurements provides a useful guide in interpreting the adsorption be-
havior.  The conformation of a starch molecule may be altered through an
introduction of charged functional groups in the starch structure, or
through an addition of an electrolyte to the solution.  A stretched-out
conformation facilitates the bridging, but an excessive introduction of
the charged functional groups interferes with bridging through intermo-
lecular repulsion.  A high concentration of an electrolyte makes the starch
molecules coil up and the bridging less effective.  Viscosity measurements
provide a convenient technique for investigating the conformation of starch
molecules.  The clarification of the fine turbidity by the suspended sol-
ids may be effected when the zeta potential is reduced to near zero.  A
polymeric flocculant of too high molecular weight often results in an ex-
cess local concentration which produces irreversible stabilization by
protective action.  In this respect corn starch appears to be more ideally
suited in studying the flocculation and clarification behaviors of mineral
suspensions than high molecular-weight synthetic polymers.

     The results of the present investigation lead to the following con-
clusions :

          1.  Goethite suspensions could be readily flocculated with
              clear supernatant liquid using starch over a wide range
              of pH.
                                JCV

-------
 2.   Quartz suspensions  could not  be  flocculated with  starch
     in the absence  of calcium ion.   The  supernatant liquid
     was extremely difficult  to clarify even when  an exces-
     sive amount  of  calcium ion was added along with a starch
     or a polyelectrolyte.

 3.   The condition for maximum settling rate and minimum  tur-
     bidity after settling  was observed at a starch  level where
     an excess  of starch begins to appear in the supernatant
     liquid.  The critical  amount  of  starch for this condition
     was dependent on the type and the mesh-of-grind of mineral,
     the type of  starch, the  pulp  pH  and  the calcium ion  con-
     centration.

 4.   Adsorption of starch molecules on an oxide mineral depends
     on the molecular size  and the substituted charged functional
     group in the starch structure, the pulp pH and  the calcium
     ion concentratration.  Goethite  adsorbed  appreciably more
     starch than  quartz for a given condition.

 5.   The isoelectric points of quartz and goethite were deter-
     mined by streaming-potential  measurements  to  be at pH near
     2 and 6.7  respectively.   Negatively-charged quartz and
     goethite surfaces may  be made positive in calcium chloride
     solutions  only  in highly alkaline solutions.  This obser-
     vation supports a view that a hydrolyzed  species  CaOH  is
     responsible  for the reversal  of  charge.

 6.   Streaming-potential measurements indicated that the  adsorp-
     tion of corn starch on goethite  was  irreversible, and the
     adsorption on quartz in  the acid pH  range appeared to be
     also irreversible.

 7.   Cationic starch was particularly effective in clarifying
     quartz suspensions. An  excessive substitution  of the
     starch structure with  cationic groups, however, appeared
     to interfere with flocculation due to interparticle  re-
     pulsion.

 8.   Anionic starch  (0.93 D.S.) was ineffective in clarifying
     quartz suspensions irrespective  of pH and of  calcium
     chloride concentration.   It was  also ineffective  in  floc-
     culating goethite suspensions of highly alkaline  pH.

 9.   The floe size appeared to be  determined by the  amount  and
     the conformation of starch adsorbed, thereby  controlling
     the settling as well as  the filtration rates.

10.   The clarities of supernatant  liquids were, in general,  not
     related to the  settling  rate, and appeared to be  governed
     by the zeta  potential.  Too strong an affinity  of poly-
                        xvi

-------
     meric flocculant due to high molecular weight or to a
     favorable chemical modification resulted in high tur-
     bidity in the supernatant liquids.   Locally excessive
     concentration of the polymeric flocculant caused ir-
     reversible stabilization, particularly of fine parti-
     cles, by protective action.

11.   Viscosity of a corn starch dispersion decreased as the
     concentration of calcium ion in solution increased.   This
     electroviscous effect was more pronounced at higher pH.

12.   The maximum settling rate of a goethite suspension floc-
     culated with corn starch at natural, near neutral pH and
     at pH 11 decreased with increasing concentration of cal-
     cium ion due to coiling of the adsorbed starch molecules.

13.   The maximum settling rate of a quartz suspension floc-
     culated with corn starch at natural, near neutral pH
     decreased, whereas the rate at pH 11 increased markedly
     with increasing concentration_of calcium chloride.  This
     increase in the rate up to 10 ^N calcium chloride was due
     to increased adsorption of the starch, but the rate even-
     tually decreased at a high concentration of calcium ion
     due to coiling of the adsorbed starch molecules.

14.   The maximum settling rate of a flocculated suspension for
     a given condition increased somewhat as the size distri-
     bution of that mineral became coarser.  Pulp dilution had
     a pronounced effect on the maximum settling rate; whether
     the maximum settling rate increased or decreased with the
     pulp dilution appeared to depend on the affinity of starch
     towards the mineral surface.  In either case, dilute sus-
     pensions were shown to be difficult to clarify.

15.   Both the settling rate and the supernatant liquid improved
     with increasing temperature.  The effect of temperature,
     however, was less than that expected from the decrease in
     viscosity alone.

16.   Selective flocculation of a mixture of goethite and quartz
     could be effected using corn starch as a flocculant.  The
     pulp pH and the level of starch addition were identified  to
     be some of the important variables.
                       xvi i

-------
                            SECTION II
                          RECOMMENDATIONS
     Of immediate interest in applying the findings of the present inves-
tigation is the clarification of the siliceous tailings from magnetite-
taconite plants and of the dispersed siliceous slimes as well as the water
reuse in the selective desliming-flotation process on nonmagnetic taco-
nites.  As reported in this article, quartz suspensions, particularly
those aged in alkaline solutions, are extremely difficult to clarify.

     It is recommended that test programs on the treatment of magnetite-
taconite tailings and of dispersed siliceous slimes from selective de-
sliming operations be considered in a small-scale pilot plant in order
to develop a sensing device for monitoring the clarification behavior in
a continuous stream.  Based on the results of the present investigation,
measurements of the residual concentrations of starch and calcium ions,
and of the pulp pH may be related to the clarity of the supernatant li
quid from a thickener, leading eventually to an automatic control system.
In this manner the clarification of the supernatant liquid will be maxi-
mized, yet the pollution due to excess starch and calcium ion in the ef-
fluent will be minimized.

     Though a certain degree of success in clarifying quartz suspensions
has been attained using cationic starch or a combined use of dodecylam-
monium ion and starch, a search for more effective means of clarifying
quartz suspensions should be considered.  Some of the flocculants sug-
gested in this report are cationic starches of different molecular weights
and degrees of substitution, and such modifying agents in conjunction with
starch as magnesium, aluminum, ferric,and dodecylammonium ions.

     In addition, an extension of the present approach to calcite, another
major constituent mineral in many iron ores and nonferrous metal ores,
becomes of both theoretical and practical interest.  Calcite is a salt-
type mineral and the characteristics of its adsorption of polymeric
flocculants differ greatly from those of oxides.  Furthermore,calcite
releases calcium ions into pulp liquors.  The presence of calcite in mag-
netite-taconite tailings may influence their flocculation behavior and
may also determine the effectiveness of separation in the selective de-
sliming of nonmagnetic taconites.
                                XVlll

-------
                            SECTION III
                           INTRODUCTION
     The clarification of plant effluents, particularly those from mineral
processing operations, is becoming of much concern for the prevention of
stream and lake pollution.  For example, magnetite-taconite concentrators,
base metal operations, such as copper; lead, and zinc, and uranium leach-
ing plants, produce siliceous tailings.   Coal washeries reject argilla-
ceous tailings, and oxidized iron ore plants discharge red iron ore slimes.
The effects of such flocculants as multivalent cations1 and polymeric com-
pounds2^3 have been the subject of much investigation.  The former is ex-
plained by the lowering of the electrokinetic potential and the latter
usually by polymer bridging.  The effectiveness of the combined use of a
multivalent cation and a polymeric compound has often been quoted in lit-
erature, 1+J5J6 ,»7 but little information is available on their individual
roles and their interaction with respect to sedimentation rate and clari-
fication of the supernatant water.  The beneficial effect of the presence
of coarse particles in suspension has been recognized,8 but the relation
between the size distribution characteristics and the flocculation be-
havior has not been well established.

     In a recent study at the Mineral Resources Research Center* on the
use of calcium chloride and starches in iron ore beneficiation,9 it was
noted that the sedimentation rates and the filtration rates of red iron
ore slimes and of magnetite-taconite tailings were strongly dependent on
the starch concentration, whereas the clarity of the supernatant water
depended on the calcium-ion concentration and the pH, and the extent of
their interaction was determined by the mineralogy of the suspension.  It
was speculated that the adsorbed starch through polymer bridging governed
the floe size and the calcium-ion concentration and the pH controlled
the zeta potential of the mineral, thereby controlling the clarity of the
supernatant solution.  Another notable observation was that an excess starch
tended to redisperse the particles due presumably to its protective ac-
tion.  Not only for efficient utilization of these flocculating agents on
the settling and clarification of mineral suspensions, but also for mini-
mizing the pollution of effluents by these reagents it is desirable to
formulate the nature and extent of their adsorption and to correlate with
their flocculation and clarification behavior using certain pure mineral
suspensions and their mixtures.

     The present study was undertaken in order to examine the mechanism
of a combined use of a multivalent cation and a polymeric compound, to
arrive at a most effective choice in their use as flocculants for the
control of water pollution problems, and to establish the effect of size
     *Formerly known as the Mines Experiment Station.

-------
distribution on floe formation and clarification.   For prototype solids,
quartz and goethite of several different size distributions were chosen,
since their grinding characteristics10 and their surface properties11,12.13
are fairly well  established.    At the same time they constitute the main
components of the ore minerals of immediate interest for the iron ore in-
dustry of Minnesota.  For flocculants, calcium ion and unmodified corn
starch were used mainly because their analytical procedures are well es-
tablished1^ and a certain amount of adsorption6*1^ and flocculation data,
particularly on ores and tailings,7 are available.  The investigation was
later extended to include chemically modified starches and polyelectro-
lytes.

-------
                            SECTION IV


                      EXPERIMENTAL. MATERIALS
     Quartz and goethite samples used in the tests were prepared in the
following manner.
Quartz

     For flocculation and filtration tests a quartz sample was prepared
from St. Peter sand by first agitating a 1000-gram sample in a Fagergren
laboratory flotation machine using distilled water for 10 mnutes,  stop-
ping the agitation to allow all plus 20-micron particles to settle, and
then removing the supernatant slimes by siphoning.  The cycle of brief
agitation, sedimentation, and siphoning was repeated six times.   The de-
slimed products thus prepared were heated overnight in a hot solution of
1 N hydrochloric acid and then repeatedly washed with distilled water un-
til there was no response to the silver nitrate test for chloride ion.
Approximately 40 pounds of the quartz sample was prepared in this manner
and dried at 110° C.

     The quartz samples for flocculation and filtration tests were  pre-
pared by dry-grinding 600-gram lots of the cleaned St. Peter sand in an
Abbe" porcelain mill for 30, 60, 120, and 240 minutes.  The size distribu-
tions of the ground products, as determined by the Andreas en pipette
method, are shown in Figure 1 (a).  As apparent in the figure, the  ex-
perimental data points are represented with straight lines having the same
slope, but are shifted laterally as the grinding time is increased.  Each
line could be represented by the  Schuhmann equation

                                      m
                           y = 100 @

where y is the cumulative percent weight finer than size x, and m and k
are respectively the distribution and size moduli.  The distribution
modulus, or the slope, is seen to remain constant at 0.79 and the size
moduli at 330, 138, 60, and 26 microns, respectively.  Figure 2, in which
the size moduli are plotted against the grinding time, shows another
straight line relationship.10  The Elaine surf ace areas of the last  three
samples were 1320, 2860, and 5110 cm2 per gram.

     For streaming-potential measurements three types of quartz samples,
viz., St. Peter sand, Montana pegmatite quartz,and Brazilian rock crystal,
were compared in a series of preliminary tests, and eventually Brazilian
quartz was selected for a detailed study of the effect of starch, cal-
cium chloride, and pH.  All the samples used in the streaming-potential
measurements were prepared by screening out a 48/65-mesh fraction,  treat-
ing it in hydrochloric acid solutions of different strengths for a speci-
fied length of time, and then washing it with distilled water until there

-------
o:
UJ
o

UJ
    100
u
  ^
h-

o
LU
                                 (a) QUARTZ , DRY GRINDING
                     PARTICLE  SIZE .MICRONS

       FIGURE l(a).  Size Distribution of Quartz as a Function of Grinding
                 Time
    100
LU
Z
I_L
 o
o

LU
>

§

=>


U

H
I


LU
      10
                                                 i  i  r i i i i
                              (b) GOETHITE.WET GRINDING  I
               i — i  i i  1 1 1 1
                              i _ i _ i i  i 1
                                                    i	'  i 111
      FIGURE l(b).
                        10              100

                   PARTICLE SIZE, MICRONS


                 Size Distribution of Goethite as a Function of
                 Grinding Time
                           4
                                                      1OOO

-------
Z)
Q
O

LU
N
to
        ;    \
00

g  100
U
to
Z)
O
       1

                    ^
        %
I   I  I  I I  INI

       10               100             1000
        GRINDING TIME,MINUTES
 FIGURE 2.  Grinding Time - Size Modulus Relationships for
          Quartz and Goethite

-------
was no response to the silver nitrate test for chloride ion.   The details
of the cleaning procedure for the three types of quartz samples and the
results of the preliminary tests are given in Experimental Results sec-
tion.  All the samples were stored either in distilled or in demineral-
ized water in a stoppered Pyrex glass jar.
Goethite

     For a goethite sample, relatively pure lumps, having a radiating fib-
rous structure, were obtained near the Maroco pit on the Cuyuna Range,
Minnesota.  The crude sample was crushed, and approximately 50 pounds of
high-purity goethite crystals were carefully handpicked.   It was first
crushed in a jaw crusher and then in a rolls crusher.   The product was
screened at 65 mesh, and the oversize was stage-ground dry in an Abbe" pfcr-
celain mill until all the material passed through this size.

     Samples of goethite were prepared for flocculation and filtration
tests by a dry-grinding procedure similar to the procedure used for the
quartz samples, but the preparation encountered some unexpected problems.
The ground samples packed tightly on the inside wall of the mill as well
as on the pebbles„ and the size distribution could not be characterized
as readily.  The use of a laboratory, steel, rod mill and the drying of
the sample in an oven prior to grinding, or frequent drying between suc-
cessive stages in 15-minute intervals, did not materially alleviate the
above difficulties.  Wet grinding in the steel rod mill at 50-percent
solids was found to work satisfactorily, and the size distribution of the
goethite samples ground wet for 15, 30, 60, and 120 minutes, determined
by the Andreasen pipette method, are shown in Figure 1 (b).  The size
distribution lines showed characteristics similar to those in quartz, and
could be represented by a constant distribution modulus of 0.80 and size
moduli of 60, 37, 22, and 12 microns, respectively.  The size moduli,
plotted against the grinding time, showed a straight line relationship,
as seen in Figure 2.

     For streaming-potential measurements, a goethite sample was prepared
by separating a 48/65-mesh fraction, treating it in a warm solution of
1 N hydrochloric acid for one-half hour, and then washing it with dis-
tilled water until there was no response to the silver nitrate test for
chloride ion.  The product was stored in distilled water in a stoppered
Pyrex glass jar.

     For most of the tests on flocculation and filtration, quartz and
goethite samples  that had approximately the same size modulus were chosen,
namely, the quartz sample dry-ground for 120 minutes Hnd the goethite sam-
ple wet-ground for 15 minutes.  Both samples had a size modulus of 60 mic-
rons and a distribution modulus, coincidentally, of approximately 0.8.
The ground sample of goethite was filtered and split into 50-gram  (dry
basis)  lots.   Each lot was kept moist in a sealed container until it was
used.   The chemical composition of the goethite sample thus prepared was
as follows:

-------
                        Fe            57.10%
                        Mn              1.39
                        SiO?            3.11
                        H 0 (400°C)     8.02
                            (800°C)   10.53
                        Al 0            0.75
Starches
      The following four starches received from Corn Products Company,
Argo, Illinois, were used for the present investigation.

                 Corn Starch (Globe Pearl No. 3001)
                 0.023 D.S. Cationic Starch
                 0.067 D.S. Cationic Starch
                 0.93  D.S. Carboxymethyl Corn Starch, Sodium Salt

Globe Pearl Starch 3001 is a corn starch which has been prepared with a
minimum alteration.  The cationic starches are laboratory samples prepared
by derivatizing unmodified, pregelatinized corn starches to introduce ami-
noethyl groups into the starch structures.  Anionically modified carboxy-
methyl starch is also a laboratory sample prepared in a similar manner as
above.16  The definition of the degree of substitution (D.S.) is given on
page  90.
Synthetic Polymers

      The four synthetic polymers used in the present investigation were
received from Dow Chemical Company, Midland, Michigan, and were reported
to have the following characteristics.17

                               Ionic           Nominal
                             Character    Molecular Weight

          Separan NP-10      Non-ionic         1 a 106

          Separan NP-20      Non-ionic         2 oc 106

          Separan AP-30      Anionic          >2 oc 106

          NC 1733            Cationic          1 -x 106
Dodecyl ammonium Chloride

      Dodecylammonium chloride was prepared-12 by bubbling dry hydrogen
chloride gas through a 20-percent solution of high-purity dodecylamine,
received from Armour and Company, in anhydrous benzene.  The crystals of
dodecylammonium chloride (DAC), precipitated on cooling to room tempera-
ture, were filtered, washed thoroughly with benzene, and transferred to a
flask with a reflux water condenser.  The material was heated over a water

-------
bath until all the solids had been dissolved.  It was cooled  filtered
and washed with benzene, and the process of recrystallization was  re-
peated twice.  Finally, a 5-percent solution of methanol in benzene was
used to produce large, well-defined crystals.  The amine salt was  fil-
tered, dried,and evacuated in a vacuum desiccator to remove the adhering
benzene and methanol.
Chemical Reagents

     The reagents used in regulating the pH, such as sodium hydroxide and
hydrochloric acid, anhydrous calcium chloride, potassium chloride, and
aluminum nitrate,were of analytical reagent grade.
Water

     Distilled water was used for flocculation and filtration tests.   De-
mineralized water^ containing less than 0.1 ppm of salts as NaCl, was
used for streaming-potential measurements and unless stated otherwise in
the preparation of all solutions and for all test work.

-------
                             SECTION V
                       EXPERIMENTAL PROCEDURE
     The procedures used for preparing starch and polyelectrolyte solu-
tions, for testing the flocculation and filtration behaviors, and for
measuring the zeta potentials and viscosities were as follows.
Preparation of Starch Solution

     Since the pulp* pH was to range between 3 and 12 in the test program,
and since the presence of excessive inorganic ions added for the adjust-
ment of the pH was expected to influence the flocculation behavior as well
as the zeta-potential measurements, it was thought necessary to avoid the
use of caustic soda for the solubilization of corn starch.  The starch
solution was, therefore, prepared by heating in an autoclave at 125° for
30 minutesJ followed by a rapid cooling in a water bath, and then homo-
genizing in a Waring blender for 30 seconds.

     The starch solution thus prepared was examined under a microscope to
confirm that all the starch granules had been ruptured to form a colloi-
dal dispersion.^  In addition, a few preliminary flocculation tests were
made with this solution and the results were compared with a causticized
starch solution.  Virtually identical results justified the present prep-
aration method for the following tests, unless stated otherwise.  To mini-
mize the effects of microbiological decomposition, fresh starch solutions
were prepared each day.  Chemically modified starches, both cationic and
anionic, were solubilized in the same manner.
Preparation of Synthetic Polymer Solutions

     The Separan samples were solubilized to 0.1-percent solutions by ag-
itating them in a beaker on a magnetic stirrer.  Two to four hours were
required for complete solubilization; the higher molecular weight Sepa-
rans took longer times.  The cationic polymer NC 1733 tended to clump in
water and, therefore, the concentration of the stock solution was reduced
to 0.01-percent.  Fresh synthetic polymer solutions were prepared each
week.
FlQ-cculation Test Procedure

     The flocculation test procedure was standardized as follows.   Fifty
grams (dry basis) of a sample was placed in a 1000-ml graduated cylinder
*The word 'pulp1 refers to a jnixture of a ground ore with water.

-------
and diluted with make-up water to near the 1000-ml mark so that  the  sub-
sequent addition of reagent would give a total pulp volume of 1000 ml.
At this point the pH of the pulp was adjusted to a desired value using
sodium hydroxide.  A predetermined amount of calcium chloride was then
added and mixed thoroughly by inverting the cylinder ten times.   This was
followed by the addition of starch solution in three equal portions, mix-
ing being performed between each addition by gently inverting the cylin-
der three times between each addition.  After the final addition of  starch,
the cylinder was inverted five times, set down, and the settling rate was
determined.

     To investigate the effect of agitation on settling rates, the pulp
from the initial settling test was re-agitated for a second and a third
settling test by inverting the cylinder five times between the successive
tests.  Typical experimental results thus obtained are shown in Figure  3(a),
The settling rates were obtained from the straight line portion.

     For quartz suspensions, particularly when the amount of calcium chlo-
ride was low, a well-defined pulp line was not observed and the above
procedure could not be applied.  A 1000-ml graduated cylinder was complete-
ly covered with a black tape except for an opening for the incident  light
from a 40-watt bulb and for the photocell of Photovolt No.  200 Universal
Photometer at 19.4 cm below the 1000-ml mark.   The pulp was agitated in
the identical manner as above and the photometer reading was recorded  as
a function of time.  A typical result is shown in Figure 3(b), from  which
the time required for the pulp line to fall the said distance is estimat-
ed to be 2.5 minutes, and the settling rate 3.06 inches per minute.  When
no calcium chloride was added the curve thus obtained showed no discon-
tinuity.  It was interpreted, therefore, that the pulp did not flocculate.
Under such a condition the settling of the particles was expected to be
extremely slow, and the settling rate of zero was assigned arbitrarily.

     After the settling test, approximately 300 ml of the supernatant
water was siphoned out and some of the solution samples were centrifuged
at a speed sufficient to remove suspended solids without appreciably sedi-
menting the starch.  The concentration of starch in the solutions was  de-
termined colorimetrically18 and the concentration of calcium ion also
colorimetrically using the purpurate method.19  The amounts of the sus-
pended solids were determined by pipetting 50 ml from the well-agitated
supernatant solution, evaporating to dryness, and weighing the residue.
The weights were corrected for soluble salts by comparing the dry weights
with equal volumes of the respective solutions centrifuged to remove the
suspended solids.
Filtration Test Procedure

     For the filtration tests a Baroid No.  300 filter press with a cross-
sectional area of 0.05 square foot was used.   Except that one hundred
grams (dry basis) of the sample was pulped to a total volume of 400 ml,
the conditioning procedure of adjusting the pH, of adding first the cal-
                                 10

-------
  1000i
  800-
to
   600-
LJ
   40O-
ID
a.
   200-
                        A First Settling Rate

                        O Second Settling Rate

                        D Third  Settling Rate
                           —AD-
                                          -A—
       0      30     60     90
                   TIME  (sec)
                      120
     FIGURE 3(a).
Settling Curves of Quartz with Corn
Starch 0.09 Ib/ton, Calcium Chloride
3.65  Ib/ton, and pH 11.55
                                                        1.4
                                     1.2
                                     1-0
                                                       0.8
                                  CD
                                  ^0.6
                                                    o
                                                    UJ
                                                    or

                                                    or
                                                    LU
                                     0.4
                                                       0.2
0       1



FIGURE 3(b).
     2345

 TIME  (min)


Settling Curve of Quartz Using  a Photometer.
Corn Starch 0.91 Ib/ton, Calcium Chloride
0.35 Ib/ton, and pH 7.45

-------
cium chloride and then the starch was  essentially identical to the proce-
dure used for the flocculation tests.   The final mixing of the pulp by
the inversion of the cylinder was repeated 15 times to simulate the three
successive settling tests.  The conditioned pulp was then poured into a
filter press, nitrogen gas at a pressure of 20 psig was applied, and the
volume of filtrate was recorded as a function of time.   The filtrate vol-
ume thus collected was almost invariably a linear function of time.   This
is shown in Figure 4, from which the filtration rates were determined.
The filtrate was analyzed for the residual concentrations of calcium ion
and starch colorimetrically using the  same analytical procedure.  The cake
thickness varied in the narrow range of 1.4 to 1.5 cm for goethite and
of 1.6 to 1.8 cm for quartz.
Analytical Methods

     Determination of Starch.   Phenol in the presence of sulfuric acid
can be used for the quantitative colorimetric microdetermination of star-
ches.18  The analytical procedure was standardized in the following way.
Two millileters of starch solution was pipetted into a 10-ml beaker and
1 ml of a 5-percent phenol solution was added to it.   Then 5 ml of con-
centrated sulphuric acid was added rapidly, the stream of acid being di-
rected against the liquid surface rather than against the side of the
beaker for good mixing.  The beakers were allowed to stand for 10 minutes
and then cooled in a water bath at 20° C to 30° C before readings were
taken in a Beckman DU spectrophotometer.  The absorbance of the charac-
teristic yellow orange color was measured against a blank at 490 my ab-
sorption maxima.  The calibration curve thus obtained is shown in Figure
5.  It is apparent that Beer's law is obeyed up to 100 yg of the starch
taken.

     Determination of Calcium Ion.  Purpurate ion combines with calcium
in basic solution to form complexes ranging in color from yellow-orange
to red.1^  For the calcium analysis the ammonium purpurate solution was
prepared by dissolving 40 mg of ammonium hydrogen purpurate (murexide)
in 75 ml of water and diluting with 175 ml of denatured alcohol.  Up to
50 ml of a solution was pipetted into a 100-ml volumetric flask, the vol-
ume adjusted to about 50 ml, 2 ml of 0.1 N sodium hydroxide added, fur-
ther diluted to 90 ml, and mixed.  To this solution 10 ml of the ammo-
nium purpurate solution was added and shaken for 5 minutes, and the ab-
sorbance was read against a blank immediately at 506 my.  The standard
curve thus prepared is shown in Figure 6, which is linear between 0 and
0.2 mg of calcium taken.

     Determination of Dodecylammonium Chloride.  Concentrations of dode-
cylammonium chloride in an aqueous solution may be determined colormetri-
cally by precipitating it as amine picrate and extracting the precipitate
with chloroform.20  The procedure was standardized as follows.  Up to 5
ml of aqueous dodecylammonium chloride solution (1-100 yg) was pipetted
out into a 10-ml graduated cylinder and the volume was adjusted to 5 ml.
To this sample 5 ml of a 0.2-percent solution of picric acid in chloro-
form was added and the cylinder was shaken vigorously for one minute.
                                12

-------
 E 300

Q
LU
H
U
LU

_l

8 200

LU
CC
LL  100
O

LU
0
            I     I    I     I     I     I    I     I     I
                                            /_'
                         —O— QUARTZ

                         —•— GOETHITE
            i     i     i    i     i     ii    i     i
                20       40      60       80     100

              FILTRATION TIME ,  SECONDS
     FIGURE 4.  Filtration Test Data of Quartz and Goethite with
              Corn Starch 0.5 Ib per ton, Calcium Chloride 2.0
              Ib per ton, and pH 9.5
                             13

-------
   1.:
   o.9
to
LU
Q 0.6
(J
h-
Q_
O
   0.3
                   i
               I
                    5W = 0.020 mm
                      1cm  COREX CELL
                                          I
     0
                 40         80        120
                       CORN STARCH,Jjg
                                     160
       FIGURE 5.  Calibration Curve for the Colorimetric Analysis of
                Corn Starch
u.o
> 0.6
I
r
co
-7
^0.4
1
§0.2
Q_
O
I I I I I
0^
/^ X= 506 mjj
y° SW = 0.035 mm
- / IcmCOREX CELL
/ I i i i i
     0
0.1        0.2      0.3       0.4

    CALCIUM   ION,   mg
                                                      0.5
       FIGURE 6.  Calibration Curve for the Colorimetric Analysis of
                Calcium Ion
                           14

-------
The two phases separated out on one to two hours standing.  The chloroform
layer was pipetted out and the absorbance was determined against a blank
at 410 my.  The standard curve, shown in Figure 7, follows Beer's law up
to 100 yg of the amine taken.
Zeta-Potential Measurements

     The electrokinetic behaviors of quartz and goethite in aqueous so-
lutions were investigated mainly by the streaming^potential measurement
technique, but electrophoretic mobility measurements were also used to
check the behavior of ground quartz samples.

     Streaming-Potential Measurements were made by forcing a liquid through
a porous plug of mineral particles under known applied pressure and measur-
ing the potential difference generated across the plug.  The zeta potential
was calculated according to the H«lmholtz-Smoluchowski equation

                           4irn   AE
                         ~
where  t, is the zeta potential, n is the viscosity, e is the dielectric
constant, A is the specific conductance of the solution, E is the poten-
tial difference, and P is the applied pressure.  The cell assembly was
similar to that described by Fuerstenau,21 and is schematically shown in
Figure 8(a).  The streaming potential was measured with a Hewlett-Packard
Model  412A DC vacuum tube voltmeter together with a Leeds and Northrup
Speedmax AZAR recorder, and the specific conductance was determined by
measuring the resistance of the plug in_ situ with an Electromeasurements
Model  250 Impedance Bridge.

     The effect of pressure on the streaming potential developed across
a goethite plug was investigated in a solution containing 50 mg of corn
starch per liter.  The results, plotted in Figure 8(b) agree with the
theory that the ratio E/P is a constant and is independent of the direc-
tion of liquid flow.

     Electrophoretic Mobility Measurements.  A flat type, vertical, micro
electrophoresis cell, having dimensions of 40 mm in length, 18 mm in
height, and 2.46 mm in width, was used with a petrographic microscope.
A rectangular cell containing a dilute copper sulfate solution was in-
serted between the light source and the microscope to minimize convec-
tion due to heating in the cell.  Two pieces of No. 22 gauge platinum
wire were used as electrodes.  The experimental set-up is shown schema-
tically in Figure 9.  The microscope was focused at the zero velocity
layer  of the liquid corresponding to 21.2 percent of the inner width of
the cell.1  The cellos filled with a conditioned suspension, and a po-
tential of 22.5 volts was applied, and the rate of motion of particles
was determined by .measuring the time in seconds required to travel a giv-
en distance.  The polarity of electrodes was then reversed, the rate of
particle motion was determined in the reverse direction, and the two val-

-------
0.8
                           SW= 0.040 mm
                            1cm COREX CELL
  0        20      40      60      80      100
        DODECYLAMMONIUM  CHLORIDE,jug
  FIGURE 7.  Calibration Curve for the Colorimetric Analysis of
          Dodecylammonium Chloride
                    16

-------
                                            (1)  THERMOMETER

                                            (2)  AIR INLET  AND OUTLET


                                            (3)  500-ml FLASK


                                            (4)  18/9 BALL AND SOCKET  JOINT

                                            (5)  STOPCOCK (2mm bore)
                                            (6)  PLATINUM  WIRE
                                            (7)  TUNGSTEN WIRE

                                            (8)  24/4O GROUND JOINT
                                            .(9)  PLATINUM  ELECTRODE
                                            (10)  PLUG OF MINERAL PARTICLES
                                                (12x60 mm)
                   FIGURE 8(a).  Schematic Diagram of
                                Streaming-Potential Cell
1.
; -120
<
I-
2
LU
&
(D
UJ
a:
i-
10
    -80
    -40
       0
                                        50 mg/l CORN STARCH
                                        pH 7.65 (26.7°C)
                                        R=727.5
                                         O LEFT  TO  RIGHT
                                         • RIGHT TO LEFT
                                                 I
                      5            10            15
                       PRESSURE,cm  Hg
               FIGURE 8(b).  Streaming Potential of Goethite as
                            a Function of Pressure
                                   17

-------
s
A
< *nj mm gj *
1

                                                     ELECTROPHORESIS  CELL
                                               (B)
00
                                                                                        (D)
                                              I      i
                               FIGURE 9.   Apparatus  for  Electrophoresis Mobility Measurements

                                          (A)  Fluorescent  Lamp,  (B) Cell Filled with Copper
                                          Sulfate  Solution,
                                          (D)  Microscope
(C)  Electrophoresis  Cell, and

-------
ues were averaged.  The observations were repeated until ten pairs of
the readings were obtained, and with the harmonic average and the val-
ues of applied potential the electrophoretic mobility was calculated
from the equation

       ,-   /    /,  , .  /  -,   Path length (cm) x Distance between electrodes (cm)
      u (cm/sec/volt/an) - Time ^ /potential difference in cell (volts)
     Though the value of zeta potentials evaluated from the electropho-
retic mobility data is mentioned to be open to serious doubt due to the
relaxation correction,22 the data are often converted to zeta potentials
using the von Smoluchowski equation
                                         u
Restrictions imposed on this equation are that the double layer must be
thin as compared to the particle size and that the surface conductance
is negligible.  Another expression for the electrophoretic mobility was
given by Htickel on an ion surrounded by an ionic atmosphere as follows.
Colloid particles  exhibit some intermediate values with a numerical fac-
tor dependent upon the particle size and the reciprocal thickness of the
double layer.

     Viscosity Measurements.  A Cannon-Fenske capillary viscometer No. 50
immersed in  a Sargent thermistor-controlled oil bath was used for the
measurements.  A self-levelling viscometer clamp was used to suspend the
viscometer in the  bath.  All viscosity measurements were made at 25° C,
the equipment controlling the temperature to within 0.1° C.  From a few
preliminary  tests  the starch concentration was fixed at 0.4 percent for
all the measurements.  The  efflux times for the sample solution  (t ) and
solvent  (t ) thus  determined were then converted to the relative visco-
sities according to23
                                 19

-------
                             SECTION VI
                         EXPERIMENTAL RESULTS
                 Flocculation  Behavior With  Starches
 Effect  of Corn  Starch,  Calcium  Chloride,and  pH

      Flocculation  Tests  on  Quartz.   To  investigate the effects  of the three
 major variables  on the  flocculation  behavioi  of a quartz suspension, a
 three-factor  orthogonal  composite design was  used.   From a few  prelimi-
 nary tests  the  ranges of the  three variables  were selected as shown in
 Table 1.   The quartz sample used in  this series  of tests was  that ground
 for 120 minutes, having  a size  modulus  of  60  microns.   The results of the
 settling  rates,  the suspended solids, the  residual  starch, and calcium ion
 concentrations  are given in Columns  Y    Y  ,  Y ,  and Y.,  respectively, in
 the same  table.                                       4

      The  experimental points  were fitted witn second-order equations by
 the method of least squares.  The regression  equations thus generated
 were as follows:

 (Settling Rate)  =  3.73  - 1.20X  + 1.99X9 + 2.14X  + 0.64X2
                          O         O
                   j. nO^Y  4- 9 PHY   j_n  D1VY     1  1 £V V
                   f U.UO^2  + £.OUA, +  U.OlA.A- -  1.1OA_A
                   + 0.82X2X3;   r = 0.93    (<£  = 10)                  [1]

 (Suspended Solids)  = 2559 + 1020X  - 1401X,,  + 223X7 -  421X 2
                            o   J-   rt     ^        O       1
                     +  825X2  - 123X3   -
                      - 409X2X3;  r = 0.93    ($ =  10)                {2]
 (Starch Concentration) = 8.35 + e.SlXj -  1.61X2 + 0.54X
                                   + 0.63X22 -
                         - 0.93X^2 + O.gSX^   0.69X2X3;
                         r = 1.0    (4. = 10)
 (Calcium Concentration) = 15.25 + 0.15X  + 14.76X9 - 2.26X
                            0.04X12 + 1.75X22 - 0.63X32
                                     - 0.21X^2 -  1.84X2X3;
                          r = L.O  (^ = 10)                         [4]

where X^ = starch addition (Ib per ton)
      X2 = calcium chloride addition (Ib per ton)
      X3 = pH
                                20

-------
           TABLE 1.   COMPOSITE DESIGN AND PRODUCT DATA FOR
                     FLOCCULATION TESTS ON QUARTZ
Factors*
Y-.
X2
X3
r\
= Corn Starch, pound per ton *
- Calcium Chloride,
= pH


puuau per uon ^





Responses :
"I
Y3
Y4
= Settling
= Suspended
= Residual
= Residual
Rate, inch per minute
Solids in supernate,
Starch Concentration,
Calcium
Concentration
Design Levels

-1
1
-1
1
-1
1
-1
1
-1
1
0
0
0
0
0
1
-1
-1
1
0
-1
Xl

-


-
_


.215
.215
-




_

_


.215
*2
1
1
1
1
1
1
1
1
0
0
1.215
1.215
0
0
0
1
1
1
1
0
0
X3
-1
-1
-1
-1
1
1
1
1
0
0
0
0
-1.215
1.215
0
-1
-1
1
1
0
0

0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
ppm
/
• i
ttlf
A
ppm w
*
Factor
Xl
.09
.91
.09
.91
.09
.91
.09
.91
.0
.0
.5
.5
.5
.5
.5
.91
.09
.0,9
.91
.5
.0

0
0
3
3
0
0
3
3
2
2
0
4
2
2
2
0
3
0
3
2
2
ppm

Levels
X2
.35
.35
.65
.65
.35
,35
.65
.65
.0
.0
.0
.0
.0
.0
.0
.35
.65
.35
.65
.0
.0
X
7
7
7
7
11
11
11
11
9
9
9
9
7
12
9
7
7
11
11
9
9
t
>/2
)
*/
f m

./*

i







Responses
3 Yl
.45 5.31
.45 3.06
.45 5.79
.45 5.54
.55 10.5
.55 2.31
.55 14.1
.55 11.0
.5 2.38
.5 5.19
.5 0*
.5 5 . 75
.0 4.60
.0 9.40
.5 5.37
.45 3.25
.45 6.84
.55 11.0
.55 10.3
.5 4.80
.5 2.34
Y2 Y3 Y
960 1 2
2570 13 3
570 1 38
1130 12 38
1230 1 1
6010 20 1
680 0 27
680 12 27
670 0 16
3510 15 13
7120 12 0
740 6- 34
970 7 14
800 6.5 13
2440 10 16
4230 16 2
560 1 32
1070 2 0
620 14 27
2200 7.5 17
1010 0 16
4
.5




.5









.5





*Flocculation not apparent and settling rate of zero assigned arMtrarily.
                                  21

-------
       r = multiple correlation coefficient

       4> = degrees of freedom

   Using  an F-test, all the equations were tested by comparing the lack
of fit variance with the experimental error variance, estimated for the
duplicate tests.  Equation 1 was found to be statistically significant,
but the multiple correlation coefficient was-high enough so that the e-
quation was considered to represent the experimental results satisfac-
torily.  The experimental errors in all the other equations were consid-
erably smaller than the variances due to the lack of fit of the equations

     Response contours were plotted by substituting values for X.^, X2,

and X_ in the above regression equations and solving for the responses
     O
Y , Y  , Y , and Y  with a digital computer.  The results are plotted in

Figure 10.  As seen in Figures 10(a) and 10(b)  the settling rate and the
amount of suspended solids are complexly dependent on all three of the
variables under investigation.  It is apparent  in Figure 10(c) that the
abstraction of starch by quartz is influenced by the calcium-ion concen-
tration, but is relatively independent of pH in the presence of calcium
ion in solution.  The abstraction of calcium ion, however, is affected
by the pH, whereas it is virtually independent  of the starch concentra-
tion,  as seen in Figure 10(d).  This observation is in good agreement
with the adsorption studies reported previously.6,15

     The flocculation and clarification behaviors of quartz suspensions
may be more closely related to the calcium-ion  concentration and the pH,
and the increase in these two parameters may be fulfilled most econom-
ically through the use of lime.  There appears  to be a tendency in Fig-
ures 10(a) and 10(b) for both the settling rate and the clarification of
the supernatant water to be impaired by the addition of starch.  In the
presence of a measurable excess concentration of starch a mineral sus-
pension may be redispersed due to its protective action..

     In order to substantiate the above hypothesis a number of tests were
performed with various amounts of starch, particularly in the low range,
but keeping all the other variables constant.   From the results in Fig-
ure 10 the pH was fixed at 9.5 and the level of calcium chloride addition
at 2 pounds per ton.  This level of addition resulted in a residual cal-
cium-ion concentration of 15 to 18 ppm, a typical value of Minneapolis
tap water.  The test results, which are shown graphically in Figure 11,
indicate that, as the amount of starch is increased, the settling rate
increases initially reaching a maximum at about 0.075 pound per ton, and
then decreases to a constant rate at about 0.2  pound per ton.  The amount
of the suspended solids appears to undergo a minimum in the neighborhood
of the starch level where the settling rate is  at a maximum.  The resi-
dual starch concentration is a trace as shown in the figure.  Beyond 0.1
pound of starch per ton a measurable quantity of starch, in excess of
several ppm, appears in the solution, the settling rate remains constant,
and the amount of suspended solids increases with the starch addition.
As expected, the residual calcium-ion concentration remains nearly con-
                                22

-------
(a) SETTLING RATE, inch/min.
    (c) RESIDUAL  STARCH CONC., ppm
 (b)  SUSPENDED SOLID, ppm
                                                   CO1
                                                    LGW
(d)  RESIDUAL CALCIUM ION CONC.,  ppm
                   FIGURE 10.  Response Contours  of Quartz Flocculation Tests
                   Settling  rate (a) .Suspended Solids Concentration in Supernate (b),
                   Residual  Starch Concentration  (c), and Residual Calcium-ion Concen-
                   tration (d), as Functions of pH, Calcium Chloride, and Starch

-------
UJ
Si
t- -C
I— 
-------
stant throughout the tests.  The response Contours shown in Figure 10 are
in good agreement with the test results, except that the contours fail to
indicate the presence of a maximum in the settling rate due presumably to
the lack of a sufficient number of experimental points within this range.
Figure 12 shows the effect of pH under otherwise identical test conditions.
The settling rates are seen to be strongly dependent on pH; the higher the
pH, the higher the settling rate  again in good agreement with the results
presented in Figure 10.  The starch level where the settling rate is at a
maximum at pH 11 is seen to be located near that at pH 9.5.  At pH 4 the
maximum that was observed at the other two pH is not observed.  It appears
that the decreased adsorption of calcium ion at low pH resulted in the de-
creased adsorption of starch and the condition corresponding to protective
action is not realized at this adsorption density.

     It then became of interest to perform a three-factor orthogonal com-
posite experiment in which the starch level ranged from 0 to 0.1 pound
per ton.  After the experiment had been carried out, the experimental points
were fitted with second-order equations by the method of least squares and
the equations were plotted graphically.

     The results of the above experiment are presented in Table 2 and the
graphs of the equations are shown as before in Figures 13(a), (b) ,  (c), and
(d).  The contours of the residual starch concentrations are not shown
since virutally all the starch added had been abstracted by the quartz sam-
ple.  The settling rates at the three levels of pH are seen to depend very
strongly on the level of starch addition, and, as expected, the settling
rates increase with an increase in the starch level.  It is also noted
that the settling rates increase as the pH is raised, but are virtually
independent of the level of calcium chloride addition.  The amounts of sus-
pended solids are seen to depend more strongly on the calcium chloride lev-
el than the other parameters.  By comparing the contour lines of Figures
13(b) and (_d~) it is apparent that a close correlation exists between.the
amount of suspended solids and the residual calcium-ion concentration.
This particular point will be discussed in great detail together with the
results of the streaming-potential measurements later in Section VII.

     It is noted in Figures 10 through 13 that the amount of suspended sol-
ids even at the lowest level was in excess of 500 ppm and that the super-
natant solution appeared turbid.  It is well known** that quartz suspen-
ded, particularly in an alkaline medium, is extremely difficult to floc-
culate with a synthetic polymer.  In fact, it has been mentioned that sil-
ica particles in water might act as a hydrophilic colloid.25

     Flocculation Tests on Goethite.  To investigate the effect of the same
three variables on the flocculation of goethite suspensions, an identical
procedure was used.  The design levels and the product data are given in
Table 3, and the regression equations are shown graphically in Figures
14(a) , 14 (b) , and 14(d) .  The contours of the residual starch concentra-
tion are not shown since virtually all the starch added has been abstracted
by the goethite sample.  The responses are seen to be quite different from
those of quartz.  The settling rates are strongly dependent on the level
                                  25

-------
c
E

_c
u
c
 ^
UJ
o
UJ
CO
      12
8
           (a)

             A"	"A>
   /o— o-
                                        	1	
                                         -o-pH 4.0

                                         -A- pH 9.5

                                         -n-pH 11.0
                                                     A
                                                    -A-


                                                    •o-
                 0.1       0.2        0.3       0.4
                      CORN STARCH, Ib/ton
                                                    0.5
FIGURE 12.
                 The Effect of pH on the Flocculation Test Results on
                 Quartz as a Function of Corn Starch Addition;  Calcium
                 Chloride 2 Ib per ton
                            26

-------
                TABLE 2.  COMPOSITE DESIGN AND PRODUCT DATA FOR
                          FLOCCULATION TESTS ON QUARTZ
Factors:

Xj = Corn Starch, pound per ton
X« = Calcium Chloride, pound per ton

X3 = PH

Responses:

Y.. = Settling Rate, inch per minute
Y_ = Suspended Solids in supernate, ppm
Y_ = Residual Starch Concentration, ppm
Y. = Residual Calcium Concentration, ppm
    Design Levels
Factor Levels
Responses
xl
-1
1
-1
1
-1
1
-1
1
-1.215
1.215
0
0
0
0
0
1
-1
-1
1
0
X2
-1
-1
1
1
-1
-1
1
1
0
0
-1.215
1.215
0
0
0
-1
1
-1
1
0
X3
-1
-1
-1
-1
1
1
1
1
0
0
0
0
-1.215
1.215
0
-1
-1
1
1
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
xl
.009
.091
.009
.091
.009
.091
.009
.091
.0
.1
.05
.05
.05
.05
.05
.091
.009
.009
.091
.05
X2
0.35
0.35
3.65
3.65
0.35
0.35
3.65
3.65
2.0
2.0
0.0
4.0
2.0
2.0
2.0
0.35
3.65
0.35
3.65
2.0
X3
7.45
7.45
7.45
7.45
11.55
11.55
11.55
11.55
9.5
9.5
9.5
9.5
7.0
12.0
9.5
7.45
7.45
11.55
11.55
9.5

3
4
2
5
6
10
4
14
2
8
0
7
3
10
7
5
2
5
11
7
Yl
.2
.55
.42
.79
.53
.5
.66
.1
.38
.4
.0*
.2
.6
.7
.4
.31
.52
.45
.0
.8
Y2
1620
2180
1360
570
4710
1230
760
680
670
1740
10290
1760
1800
1200
2150
960
1400
4290
700
2160
Y3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Y4
3
3.5
32
38
1.5
1
26
27
16
14
0
32
14
12
17
2.5
36
2
28
17.5
"Flocculation not apparent and settling rate of zero assigned arbitrarily.
                                      27

-------
                                                            ABSTRACTION  OF STARCH

                                                            NEARLY  COMPLETE IN

                                                            THE  REGION (c)
        (a) SETTLING RATE, inch/min.
N>
00
       (b)  SUSPENDED SOLID,  ppm
                                                                                               0.1
(d)  RESIDUAL CALCIUM ION CONG., ppm
                           FIGURE 13.  Response Contours of Quartz Flocculation Tests
                           Settling Rate (a), Suspended Solids Concentration In Supernate (b),
                           Residual Starch Concentration (c), and Residual  Calcium-ion Concen-
                           tration (d), as Functions of pH, Calcium Chloride, and Starch

-------
           TABLE  3.    COMPOSITE DESIGN AND PRODUCT DATA FOR
                       FLOCCULATION TESTS ON GOETHITE
Factors:
X^ = Corn Starch, pound per ton
X« = Calcium Chloride, pound per ton
X3 = pH

Responses:
YI = Settling Rate, inch per minute
Y2 = Suspended Solids in supernate, ppm
Y, = Residual Starch Concentration, ppm
Y. = Residual Calcium Concentration, ppm
    Design Levels
Factor Levels
Responses
xl
-1
1
-1
1
-1
1
-1
1
-1.215
1.215
0
0
0
0
0
1
-1
-1
1
0
X2
-1
-1
1
1
-1
-1
1
1
0
0
-1.215
1.215
0
0
0
-1
1
-1
1
0
X3
-1
-1
-1
-1
1
1
1
1
0
0
0
0
-1.215
1.215
0
_]_
-1
1
1
0
Xl
0.09
0.91
0.09
0.91
0.09
0.91
0.09
0.91
0.0
1.0
0.5
0.5
0.5
0.5
0.5
0.91
0.09
0.09
0.91
0.5
X2
0.35
0.35
3.65
3.65
0.35
0.35
3.65
3.65
2.0
2.0
0.0
4.0
2.0
2.0
2.0
0.35
3.65
0.35
3.65
2.0
X
7
7
7
7
11
11
11
11
9
9
9
9
7
12
9
7
7
11
11
9
3
.45
.45
.45
.45
.55
.55
.55
.55
.5
.5
.5
.5
.0
.0
.5
.45
.45
.55
.55
.5
Yl
3.
20.
5.
17.
15.
21.
5.
21.
2.
19.
15.
18.
16.
21.
18.
20.
5.
14.
18.
18.

70
1
78
4
0
0
64
5
36
6
0
8
7
5
8
1
71
6
8
8
Y2
220
40
200
80
5480
690
360
380
260
100
600
150
100
880
130
20
270
4280
380
130
Y3
0
0
0
0
>0
>0
0
0
0
0
0
0
0
1.5
0
0
>0
>0
>0
0
Y4
6
6
39
39
0
0
7
7
17
15
0
32
25
0
14
8
37
0
7
, 14
>0 indicates starch concentration somewhat higher than 0 ppm.
                                   29

-------
                                                  ABSTRACTION  OF STARCH

                                                  NEARLY COMPLETE  IN THE

                                                  REGION
(a) SETTLING RATE, inch/min.
(b) SUSPENDED SOLID, ppm
(d)  RESIDUAL CALCIUM ION CONC., ppm
             FIGURE 14.  Response Contours of  Goethite Flocculation Tests
             Settling Rate (a), Suspended Solids Concentration in Supernate (b),
             and Residual  Calcium-Ion Concentration  (d) as Functions of pH,
             Calcium Chloride and Starch

-------
of starch addition, and, to a less extent, on pH and the level of calcium
chloride addition.  The clarity of the supernatant water improves as the
levels of both calcium chloride and starch addition are increased, and the
amount of the suspended solids is the lowest at near neutral pH.   Since
the isoelectric point of goethite is known to be at pH 6.7 (see Figure 39),
the electrical charge on the solid particles is at minimum in the above pH
region.  The interaction of calcium ion and pH is apparent in Figure 14(d).
The flocculation behavior described abov§ appears to be readily interpreted
in the light of the adsorption behavior6'35 and the streaming-potential
data described in Section VII.

     To study the effect of an excessive addition of the starch flocculant,
a series of flocculation tests was made extending the level of starch ad-
dition beyond one pound per ton, but keeping the other variables  constant.
The pH was fixed at 9.5 and the level of calcium chloride addition at 2
pounds per ton, concordant with the test conditions used on quartz in pre-
paring Figure 11.  The results thus obtained are presented in Figure 15.
The settling rate increases initially and appears to reach a maximum at
approximately 0.75 pound per ton.  Beyond this point, a measurable quan-
tity of starch, in excess of several ppm, appears in the supernatant so-
lution, and the settling rate tends to decrease.  The amount of suspended
solids undergoes a minimum again at about a point where a measurable quan-
tity of the starch appears in solution.  The residual concentration of
calcium ion in solution remains  more or less constant throughout the
range, as expected.  The present observation is again in good accord with
a view that an excessive amount of starch present in solution tends to
hinder flocculation due to protective action although the level of the
starch addition is an order of magnitude greater than that for quartz.

     Filtration Tests.  To investigate the factors governing the filtra-
tion as well as the flocculation behavior of quartz and goethite suspen-
sions, a series of filtration tests was carried out on the same samples
used in the flocculation tests.  To facilitate the comparison, three-fac-
tor orthogonal composite experiments, identical to those used for the
flocculation tests were followed.  The design and the product data of the
experiments for the quartz and goethite samples are given in Tables 4 and
5, respectively, from which regression equations were generated and response
contours were calculated with a digital computer.  The results thus ob-
tained are given in Figures 16 and 17, respectively-

     It is interesting to note that the general trend of the filtration
rate contours of the quartz sample shown in Figure 16(a) are markedly
similar to the settling rate and the suspended solid contours of the floc-
culation tests shown in Figures 10(a) and (b).  The factors governing the
filtration and flocculation behaviors of quartz, therefore, may be the
same.  The most obvious factor is the floe size, which is confirmed qual-
itatively in the experimental observation.  Since the abstraction of both
calcium ions and starch by quartz is relatively small, the residual con-
centrations of these two reagents are seen to be considerably higher at
the corresponding levels of addition in the present case than in the floc-
culation tests due evidently to the higher percent solids used.
                                 31

-------
      40
LU

!<  0)

°=  D

CD  -g
      20
•-  r
uj  .£
co
CO
 I

 or
 <

 co
        0
3   1000 -
8
Q  E
LU  CL
Q  a
z
LU
Q_
CO

CO
   E
   Q.
   Q.
   <
   QC.
O ^
9 O
CO Z
LU O
or o
S&
o cc


CCL
     500-
      100
      50
          I
       20
        0
              1
1
I
          02468
              CORN STARCH  ADDITIONJb/ton
                                                         10
             FIGURE 15.  Flocculation Test Results on
                      Functions of Starch Addition
                      Calcium Chloride Addition of
                                              Goethite as
                                              at pH 9.5 and
                                              2 Ib per ton
                        32

-------
                TABLE 4.  COMPOSITE DESIGN AND PRODUCT DATA
                          FOR FILTRATION TESTS ON QUARTZ
Factors:
Xj = Corn Starch, pound per ton
X2 = Calcium Chloride, pound per ton
X3 = pH
Responses:
Y-^ = Filtration Rate, ml per minute
Y~ = Residual Starch Concentration, ppm
Y, = Residual Calcium Concentration, ppm
    Design Levels              Factor Levels
Responses
xl
-1
1
-1
1
-1
1
-1
1
-1.215
1.215
0
0
0
0
0
1
-1
-1
1
0
x2
-1
-1
1
1
-1
-1
1
1
0
0
-1.215
1.215
0
0
0
-1
1
-1
1
0
X3
-1
-1
-1
-1
1
1
1
1
0
0
0
0
-1.215
1.215
0
-1
-1
1
1
0
Xl
0.09
0.91
0.09
0.91
0.09
0.91
0.09
0.91
0.0
1.0
0.5
0.5
0.5
0.5
0.5
0.91
0.09
0.09
0.91
0.5
x2
0.35
0.35
3.65
3.65
0.35
0.35
3.65
3.65
2.0
2.0
0.0
4.0
2.0
2.0
2.0
0.35
3.65
0.35
3.65
2.0
X3
7.45
7.45
7.45
7.45
11.55
11.55
11.55
11.55
9.5
9.5
9.5
9.5
7.0
12.0
9.5
7.45
7.45
11.55
11.55
9.5
Yl
273
1.8
300
145
32
0.6
343
316
273
94
0.9
205
245
400
226
1.8
300
35
333
218
Y2
1.5
38.5
1.5
55
0.5
26.5
0.5
22
0
59
13
22
25
0.5
24
39
1.5
0.5
28
22
Y3
15
15
170
175
1
1
142
142
92
91
0
185
92
55
90
14
171
1
148
90
                                    33

-------
                TABLE 5.  COMPOSITE DESIGN AND PRODUCT DATA
                          FOR FILTRATION TESTS ON GOETHITE
Factors :
X  = Corn Starch, pound per ton
X~ = Calcium Chloride,  pound per ton
X3 = PH
Responses :
Y.. = Filtration Rate, ml per minute
Y_ = Residual Starch Concentration, ppm
¥„ = Residual Calcium Concentration, ppm
Design Levels
                               Factor Levels
Responses
xl
-1
1
-1
1
-1
1
-1
1
-1.215
1.215
0
0
0
0
0
1
-1
-1
1
0
X2
-1
-1
1
1
-1
-1
1
1
0
0
-1.215
1.215
0
0
0
-1
1
-1
1
0
X3
-1
-1
-1
-1
1
1
1
1
0
0
0
0
-1.215
1.215
0
-1
-1
1
1
0
Xl
0.09
0.91
0.09
0.91
0.09
0.91
0.09
0.91
0.0
1.0
0.5
0.5
0.5
0.5
0.5
0.91
0.09
0.09
0.91
0.5
X2
0.35
0.35
3.65
3.65
0.35
0.35
3.65
3.65
2.0
2.0
0.0
4.0
2.0
2.0
2.0
0.35
3.65
0.35
3.65
2.0
X3
7.45
7.45
7.45
7.45
11.55
11.55
11.55
11.55
9.5
9.5
9.5
9.5
7.0
12.0
9.5
7.45
7.45
11.55
11.55
9.5
Yl
106
214
112
197
1.3
273
128
214
102
209
10
182
177
177
188
226
119
1.7
214
185
Y2
1
1
1
1
1
1.5
0.5
0.5
0
1
1
1
1.5
1
0.5
1
0
1.5
0
1
Y3
38
38
201
195
1
1
43
47
87
87
0
185
129
O.E
80
37
117
1
47
80
                                 34

-------
tn
           (a)  FILTRATION  RATE,  ml/min.
                                                                (b)  RESIDUAL  STARCH CONC.,  ppm
                                                                                              •*** I
                                                                                                      ton
                                                              (c)  RESIDUAL  CALCIUM  ION  CONC.,  ppm
                              FIGURE 16.   Response Contours of Quartz Filtration Tests
                              Filtration  Rate (a), Residual Starch Concentration (b),
                              and Residual Calcium-ion Concentration (c) as Functions
                              of pH, Calcium Chloride and Starch

-------
ON
           (a)  FILTRATION RATE,  ml/min.
                                                             ABSTRACTION OF STARCH
                                                             NEARLY  COMPLETE IN
                                                             THE  REGION
                                                       pH
                                                           (c) RESIDUAL  CALCIUM  ION  CONC.. ppm
                           FIGURE 17.  Response Contours of Goethite Filtration Tests
                           Filtration Rate (a), and Residual Calcium-ion Concentration
                           (c) as Functions of pH, Calcium Chloride, and Starch

-------
     A similar observation may be made on the goethite sample when the
filtration behavior shown in Figure 17 and the flocculation behavior in
Figure 14 axe compared.  From the foregoing discussions it becomes ap-
parent that corn starch aids both the flocculation and the clarification
of mineral suspensions when it is used in optimum amounts, whereas an ex-
cessive addition has an adverse effect.  The optimum level depends on the
adsorption behavior of the mineral species and also on the total surface
area available in the suspension for the adsorption of the starch.  In
practice, the optimum level may be determined by analyzing the residual
concentration of the starch, say with an automatic analyzer.


Effect of Size Distribution and Pulp Density

     In a previous section it was reported that corn starch as a floccu-
lant for both goethite and quartz suspensions gave maximum settling rates
at the point where an excess of starch began to appeer in the supernatant
solution.  The amount of suspended solids after settling appeared to under-
go a mimimum in the neighborhood of the starch level where the settling
rate was at a maximum.  In studying the effects of size distribution and
pulp density on the flocculation behavior, a careful selection of the levels
of starch addition becomes important since the abstraction of starch would
depend not only on the amount but also on the size distribution of the
minerals in the pulp.  The levels of starch addition corresponding to the
maximum settling rates would probably be best suited for the comparison of
flocculation behavior.

     In order to establish the levels of starch addition for samples with
different size distributions, two series of preliminary tests were per-
formed:  Cl) the adsorption isotherms of starch and (2) the settling rates
as functions of the level of starch addition both on goethite and quartz
samples having size moduli of 60, 37, and 22 microns.  From the results
of the three-factor orthogonal composite experiment reported in Figures
10, 13, and 14, the pH was fixed at 9.5 and the level of calcium chloride
addition at 2 pounds per ton, which resulted in a residual calcium-ion
concentration of 15 to 18 ppm.

     The results of the adsorption isotherms are shown in Figure 18.  All
the isotherms showed, more or less, the same trend.  The abstraction of a
starch at low levels of addition was nearly complete, resulting in a steep
rise in the adsorption density, and then approached saturation coverage
corresponding to a particular size distribution.  Since the size modulus
may be shown to be inversely proportional to the specific surface area,
the increase in the amount of abstraction with smaller size moduli is to
be expected.

     Figure 19 shows the settling rates as functions of the level of
starch addition.  The maximum settling rates for the goethite samples of
the size moduli of 60 and 22 microns are located at 0.75 and 1.5 pounds
per ton, respectively.  From these figures and the amounts of starch ab-
stracted at saturation given in Figure 18(a), it is estimated that the
maximum settling rates are at a constant coverage of approximately 15
percent of saturation.  Similarly, the maximum settling rates for quartz
having the same moduli are located at 0.075 and 0.11 pounds per ton, re-


                                  37

-------
o>


a.

f
 •>
O
UJ
I-
o
<
I
o
            1 - 1 - 1 - 1
                             1 - 1 - 1
              1 - 1 - 1
                                     k =
            ..--""v"


                                             k = 60yU
                                 (a)  GOETHITE
en      0       20       40       60       80      100      120

            RESIDUAL STARCH CONCENTRATION,  mg per I
    0.20
o»

Q>
a.

o»
o
UJ
h-
o
CO
CD
O
a:
<
    0.15
                                      i     i    i     i    i
                                               k = 22/A   V
    V


	A-
                                                    k =
    O.IOX-
   0.05
     0
                                  (b)  QUARTZ

                                 j	i	i    i
                                                        i
       0        2        4        6        8        10

             RESIDUAL STARCH CONCENTRATION, mg per I
                                                           12
      FIGURE 18.  Adsorption of Corn Starch on Goethite (a) and Quartz (b)  of
      Different Size Distributions at pH 9.5 and with 2 1b of Calcium Chloride
      per ton
                               38

-------
                              (a)  GOETHITE
                246

              STARCH  ADDITION,  Ib per ton
               0.2           0.4            0.6

               STARCH  ADDITION,  Ib  per ton
 8
                              (b)  QUARTZ
0.8
FIGURE 19. Settling Rates of Goethite (a) and Quartz (b) of Different Size
         Distributions as a Function of Amount of Starch Addition
                           39

-------
spectively.  By combining the results of Figure 18(b)  these levels of
starch addition may be shown to occur at approximately 33 percent of
saturation.  The above figures of the surface coverage were used to es-
timate the amounts of starch addition in the hexagonal design experi-
ments described below.

     To investigate the effects of size distribution and pulp density
these two parameters were chosen to conform to an orthogonal design in
a hexagonal arrangement (see Tables 6 and 7), so that the experimental
points could subsequently be fitted with a second-order equation by the
method of least squares.  For the goethite sample the ranges of the above
two parameters, namely, the solid concentration in the pulp and the size
modulus, were limited to 10 to 170 grams per liter and 11 to 110 microns,
respectively.  It should be noted that the intervals of the size moduli
in this series of experiments were inadvertently chosen in logarithmic
scale.  The pH of the pulp was kept constant at 9.5, and the residual
concentration of calcium ion in solution was also kept constant near 16
ppm.

     The design levels and the product data are shown in Table 6, from
which the regression equations were generated.   (See the bottom of Table
6.)  Using an F-test, all the equations were tested by comparing the lack
of fit variance with the experimental error variance.   The experimental
error was estimated from the duplicate test results.  The experimental
errors of the settling rates were considerably smaller than the varian-
ces due to the lack of fit of the equation, but the equations were con-
sidered to represent the experimental results satisfactorily since the
multiple correlation coefficient based on four degrees of freedom was
quite high (0.99).  The multiple correlation coefficient of the equation
for the suspended solids was rather low (0.72), but through an F-test
the equation was found to be statistically significant, and was also
thought to represent the experimental results within tolerable limits.

     Response contours were then plotted by substituting values for X..
and X_ in the regression equation and solving for the responses, YI and
¥„, with a digital computer.  Only those curves in and around the experi-
mental domain are shown in Figure 20.  The settling rate contours given
in Figure 20(a) show a rapid decrease as the concentration of solids is
increased indicating that the crowding of the floes interfered with the
settling.  A slight tendency for the canonical axes to be tilted clock-
wise implies a beneficial effect of coarser particles presumably aiding
the settling rate by increasing the effective density of the floes.  The
amount of suspended solids in the supernatant water is seen in Figure
20(b) to decrease as the concentration of solids is increased.  This may
be interpreted to imply that more crowding by the floes and slower set-
tling rate must have filtered out the suspended fines more effectively.

     A similar approach was used to study the effect of size distribution
and pulp density on the flocculation behavior of quartz suspensions. In
this series of tests the ranges of the solid concentration and the size
modulus were chosen to cover 10 to 170 grams per liter and 13 to 69 mi-
crons, respectively.  The design levels and the product data are given
in Table 7.  The pH of the pulp was again kept constant at 9.5, and the
                                40

-------
         TABLE 6.  HEXAGONAL DESIGN AND PRODUCT DATA
                   FOR FLOCCULATION TESTS ON GOETHITE
Factors
       Solid Concentration,
       gram per liter

       Logarithm of Size
       Modulus in Microns
Responses:

  Y! = Settling Rate, inch
       per minute

  Y2 = Suspended Solids in
       Supernate, ppm
   Design Levels
                          Factor Levels
Responses
1.000
0.500
-0.500
-1.000
-0.500
0.500
0
0
1.0
-1.0
0.0
0.866
0.866
0.0
-0.866
-0.866
0
0
0
0
170
130
50
10
50
130
90
90
170
10
36
60
60
36
22
22
36
36
36
36
3.0
6.8
19.6
34.7
13.7
3.8
7.8
8.2
3.7
32.2
84
84
68
636
28
84
70
68
74
170
*Lists the size moduli of the samples used.
Regression Equations :

Yj = 28.66 - 0.34Xi - 6.31X2 + O-

           Y = 0.99     ( = 4)
Y2 = -2063 - 4.6X1 + 3051X2 + 0.027XJ

           Y = 0.72    ( = 4)
                                           7.61X22   0.082X1X2
                                     2 - 933X2 -
                             41

-------
         TABLE 7.   HEXAGONAL DESIGN AND PRODUCT DATA
                   FOR FLOCCULATION TESTS ON QUARTZ
Factors:
  Xj = Solid Concentration,
       gram per liter

  X2 - Size modulus, microns
Responses:

  Y! = Settling Rate, inch
       per minute

  Y2 = Suspended Solids in
       Supernate, ppm
   Design Levels
                          Factor Levels
Responses
xl
1.000
0.500
-0.500
-1.000
-0.500
0.500
0
0
1.0
-1.0
*2
0.0
0.866
0.866
0.0
-0.866
-0.866
0
0
0
0
Xl
170
130
50
10
50
130
90
90
170
10
X2
41
60
60
41
22
22
41
41
41
41
Yl
2.0
3.4
7.3
0.2
5.2
2.3
4.4
4.6
2.1
0.2
Y2
2840
1520
1740
350
5290
6170
2100
2300
2960
340
Regression Equations:

Y: = 1.383 + 0.109X!   0.131X2 -

           y = 0.96    ( = 4)

Y2 = 10898 + 45Xj- 444X2 - 0.090X!

           Y = 0.99    O = 4)
                                            + 0.0025X2  - 0.00033X^2
                                       4.5X2  - 0.36X^2
                             42

-------
-iC


O
Q
O
LU
N
co

o
Q
     1.8
     1.7
£   1.6
     1.5
1.4
     1.3
        - 32 28 24 20  16   12    86   4
                                                     I   I   T

                                                        a
                                                      \
                                                     3.4 (
           I	I
                                         3.8



                                        i  i   I   i   i
                                                         70


                                                         60


                                                         50



                                                         40




                                                         30
      to
      o
      cc
                                                                  (f)
      Q
      O
      UJ
      N
      (f)
                                                              20
        0            50            100            150

             CONCENTRATION  OF  SOLIDS,   GRAM / LITER
 en
 o
Q
O
UJ
CO

O
Q
      1.8
      1.7
co   1.6
     1.5
1.4
     1.3
            403
                                                     79  -
  70


  60


  50



  40




  30
                                                            co
                                                            -z
                                                            o
                                                            a:
                                                            o
                                                                  co
                                                                  o
                                                                  2

                                                                  UJ
                                                                  N
                                                                  CO
          J	\	L
                         J	I	\	I	I	I	L
-20

J
        0             50             100            150

             CONCENTRATION OF  SOLIDS,   GRAM / LITER

       FIGURE 20.   Response  Contours of  Flocculation Test
                   Results on  Goethite as  Functions of Size
                   Modulus and Concentration  of Solids
                   (a) Settling Rate, (b)  Suspended Solids
                   Concentration  in  Supernate
                            43

-------
residual concentration of calcium ion in solution near 16  ppm.   The  regres-
sion equations are also given in Table 7.   In both equations  the experimen-
tal errors were considerably smaller than the variances  due to  the  lack of
fit of the equations.   These equations, however,  were thought to represent
the experimental points satisfactorily in view of sufficiently  high  multi-
ple correlation coefficents.

     In Figure 21(a),  the response contours of the settling rate show an in-
crease with increasing solid concentration initially, reach a maximum and
then decrease.  Here again, the settling rate tends to be  somewhat  greater
for coarser size distribution.  In Figure 21(b)  it is seen that the  finer
the size distribution and the greater the solid concentration,  the higher is
the amount of suspended solids in the supernatant water.   It appears as if
the fines in the quartz suspensions remain dispersed though a mud line is
observed under the conditions of experimentation.  It appears,  therefore,
that the effect of the two parameters, namely, size distribution and pulp
density,, cannot be treated in the same manner for goethite and  quartz.

Effect  of Temperature

     To investigate the effect of the pulp temperature on the flocculation
behavior^ a number of tests were performed on goethite and quartz samples
with size moduli of 60 microns.  The concentration of solids was fixed at
50 grams per  liter, the pH of the solution at 9.5, and the level of calcium
chloride addition at 2 pounds per ton.  The level of the starch addition
for the goethite sample was 0.75 pound per ton and for the quartz sample,
0.075 pound per ton.

     The resulting settling rates and amounts of solids suspended in the
supernatant water for goethite are shown in Figure 22 and for quartz in
Figure  23.  In both cases the settling rates increased, whereas the amount
of suspended  solids decreased quite appreciably with the pulp temperature.
The increased settling rates,, however, cannot be explained by the correc-
tion for the viscosity change alone as indicated by the systematic decrease
of the  product of the settling rate and viscosity.  In view of the decreased
amount  of suspended solids, presumably the tendency of the fines to floc-
culate  was promoted through an increased probability of collision, and pos-
sibly the floe size was decreased due to a denser structure with closer con-
tact of particles.


Effect  of Chemical Modification of Starch

     Whereas goethite suspensions could be readily flocculated with corn
starch  alone, the flocculation of quartz suspensions required the presence
of calcium ion in solution.  The amount of suspended solids in the super-
natant  liquor was in the range of several hundred ppm even when a fairly
large quantity of calcium chloride was added.  The calcium ion adsorbed on
the quartz surface increased the adsorption of starch and hence the floc-
culation.  It became of interest, therefore, to  study the uses of cationic
starch, particularly in the flocculation of quartz.  The adsorption density
                                    44

-------
      0         50         100       150
         CONCENTRATION OF SOLIDS, GRAM/LITER
O
o:
(J
3
to
z>
_l
ID
Q
O
2
Lu
N
UO
70

60


50


40


30


20

  ol
           500
                          II
      0         50         100        150

          CONCENTRATION OF SOLIDS, GRAM/LITER

      FIGURE 21.  Response Contours of Flocculation Test Results on
      Quartz as Functions of Size Modulus and Concentration of Solids
      (a)  Settling Rate, (b) Suspended Solids Concentration in Super-
      nate
                     45

-------
UJ
H
<
o:

e>
I-
UJ
o
o
   20
I-
<
o:

o


I]
h-
h-
UJ
en
           10
                   (a)
   SETTLING RATE , inch /minute



— SETTLING RATExVISCOSITY,

    inch /min x cp
                       10
                          20
          30
          40
    E
    o.
    Q.
    CO
    o
    UJ
    Q
    z
    UJ
    Q.
    CO
    ^
    CO
         200
         100
                         PULP  TEMPERATURE,  °C
                10
20
30
40
                         PULP  TEMPERATURE,  °C
      FIGURE 22.  Flocculation Test Results on Goethite as a Function of Pulp

                Temperature  (pH 9.5, CaCl2  2 Ib per ton, and Corn Starch

                0.75 Ib per  ton)
                                 46

-------
LU
I-
o
I-
UJ
CO
£ 12
CO
o
o
CO

>

x  8

LU
        _J
        I-
        I-
        UJ
        CO
          (a)
        SETTLING RATE


 — •—  SETTLING RATE  x VISCOSITY


	I          I         I
                10        20       30        40


                  PULP  TEMPERATURE,  °C
        2000
     E
     o.
     OL
    CO
    O

    _l
    o
    CO


    3  1500
    UJ
    Q.
    CO


    CO
                                 T
              -   (b)
                                   T
                                    I
                             T
                             I
                       10        20       30       40


                          PULP  TEMPERATURE,   °C
       FIGURE 23.  Flocculation Test Results on Quartz as a Function of Pulp
                 Temperature (pH 9.5, CaCl2 2 Ib per ton, and Corn Starch
                 0.075 Ib per ton)
                                  47

-------
of a cationic starch on quartz is reported to be considerably higher than that
of the unmodified corn starch,15 due presumably to the electrical interaction.

     Effect of Cationic Starch on Quartz.  To investigate the effects of the
level of the cationic starch addition and of the pH on the settling rates and
the quality of the supernatant liquor, an experiment of hexagonal design was
performed on a quartz sample having a size modulus of 60 microns.  The stan-
dardized test procedure described previously was followed except that the
addition of calcium chloride was omitted.  The design levels and the product
data, and the resulting regression equations are given in Table 8.  The re-
sponse contours calculated with a digital computer are shown in Figure 24.

     The response contours show that the settling rate increases and the a-
mount of the suspended solids decreases with the level of the cationic starch
addition, but they are relatively independent of pH.  It is also noted that
the settling rates are appreciably higher and the amount of suspended solids
is lower as compared to the results observed with the unmodified corn starch.
Even at the most favorable conditions, however, the turbidity of the super-
natant water was high--in excess of 100 ppm in the region of experimentation.
These results are in accord with those of previous investigators who reported
that a quartz suspension becomes non-flocculable after aging in an alkaline
medium.

     The effect of pH on the flocculation characteristics of quartz when the
cationic starch was used is shown in Figure 25.  At pH 4 and 9.5 there appears
to be very little difference in the settling rate, whereas at pH 11.0 it is
appreciably higher.  It is interesting to note that the amount of suspended
solids may be lowered quite effectively with the starch at pH 9.5 and 11.0,
but not at pH 4.0.

     To test the effect of the degree of substitution on the flocculation be-
havior of the quartz suspension, a cationic starch of 0.023 D.S. was used.
The test results at pH 9.5 are compared in Figure 26.   The standardized test
procedure was followed as before, and again the addition of calcium chloride
was omitted.  It is interesting to note that the maximum settling rates were
observed at approximately 0.5 and 1.2 pounds per ton for 0.067 D.S. and
0.023 D.S. cationic starches, respectively, more or less as anticipated from
their adsorption behaviors.   The higher the degree of substitution, the stron-
ger was the electrical interaction with the negatively charged quartz surface,
and also the electrical repulsion among the adsorbed starch.  As a result, the
saturation coverage was reached at a lower level of the starch addition.   The
higher overall settling rates with the 0.023 D.S. cationic.-starch might have
resulted from a larger floe size due presumably to a more effective polymer
bridging attributable either to the optimum degree of substitution or to the
molecular size change during derivatization.  The 0.023 D.S. cationic starch
was also somewhat more effective in removing the suspended solids.  This  is
a point of considerable importance since the clarification of the quartz sus-
pension is known to be difficult.  Further investigation through adsorption
measurements and streaming potential measurements using cationic starches of
different degrees of substitution becomes of interest.
                                   48

-------
         TABLE 8.  HEXAGONAL DESIGN AND PRODUCT DATA
                   FOR FLOCCULATION TESTS ON QUARTZ
Factors:
      = 0.067 D.S. Cationic Starch,
        Pound per ton
   X  = pH
Responses:

   Y- = Settling Rate, inch per
        minute

   Y_ = Suspended Solids in
        Supernate, ppm
   Design Levels
Factor Levels
Responses
xl
1.000
0.500
-0.500
-1.000
-0.500
0.500
0
0
1.0
-1.0
Regression
YI = 39.23
X2
0.0
0.866
0.866
0.0
-0.866
-0.866
0
0
0
0
Equations :
+ 45.81X1 -
Xl
1.0
0.75
0.25
0.0
0.25
0.75
0.5
0.5
1.0
0.0

9.84X2 - 2:
X2
9.5
11.0
11.0
9.5
8.0
8.0
9.5
9.5
9.5
9.5

") Q f\V j_ f\
£ • OVJAi " \J
Yl
11.2
12.6
12.6
0
8.4
10.6
11.6
10.6
10.4
0

.61X22 -
2
230
450
640
3080
220
160
210
180
270
2940

1.47X][X2
       Y = 0.96    (4> = 4)

   = -6248 - 7175X  + 1734X
       Y = 0.95    ( = 4)
                             49

-------
      12
      10
  X
  0.
       8
                           10.8
          (a)
                 I
         I
        I
        I
                0.2      0.4      0.6      OS      1.0
                  CATIONIC  STARCH,  LB/TON
      12
      10
  X
  Q.
      8
                T
        T
        T
                          .250
          (b)
                 I
I
I
                                          160
I
I
        0       0.2      0.4      0.6       0.8      1.0
                   CATIONIC  STARCH,   LB/TON
                                 1.2
                                 1.2
FIGURE  24,  Response Contours of Flocculation Test Results  on Quartz
           as a Function of Amount of Cationic Starch (0.067 D.S.)
           and of pH
           (a) Settling Rate, (b) Suspended Solids Concentration in
           Supernate
                             50

-------
20
c
E
u
c
UJ
i 10
o
z
!^
l—
l—
LJ
CO
0
3000
E
Q.
Q.
§ 2000
0
CO
Q
UJ
Q
£ 1000
CO
H>
CO
I I I I
(a)
^^^^
X1^"
" /
/ AQ 	 p. p.
/ /^ % 	 °
~ AO A
/
A 1 1 1 1

f (b)
-o- pH 4.0
-A- pH 9.5
1 -D- pH 11.0
_\
\
\\
KL^__ 	 o. -
CM^-— 	 • — O
1 1 1 1
 0                    1.0                  2.0
       0.067 D.S. CATIONIC STARCH Ib/ton
FIGURE 25.   The Effect of pH on the Flocculation Test
           Results  on Quartz as a Function of 0.067
           D.S. Cationic Starch Addition
                51

-------
     20-     (a)
c
E

.c
u
c

LU

5
cr

O
LU
I/")
                            1.0                 2.0
                       CATION 1C  STARCHJb/ton
   4000
 E
 Q.
 Q.
(D
Q
Q
LU
O
Z
UJ
Q_

ID
(Jl
   2000-
0.023 D.S.
0.067 D.S.
                            1.0                  2.0

                      CATION 1C STARCH,  Ib/ton

         FIGURE  26.  Graphs Illustrating (a) the Settling  Rate and (b)  the
                   Amount of Suspended Solids in Supernatant Water of Quartz
                   as Functions of the Levels of Two Different Cationic
                   Starches at pH 9.5
                            52

-------
     Effect of Cationic Starch on Goethite.  The flocculation behavior of a
goethite suspension with the 0.067 D.S. cationic starch was investigated
briefly.  The experimental procedure was the same as that used previously:
50 grams of a goethite sample per liter at pH 9.5 and no addition of calcium
chloride.  The results are shown in Figure 27.  As expected, the cationic
starch was as effective as the unmodified corn starch in flocculating and
clarifying the goethite suspension.  No further attempt was made, therefore,
to elaborate the flocculation behavior of goethite suspensions with cationic
starches.

     Effect of Anionic Starch on Quartz.  Quartz suspensions cannot be floc-
culated with either an unmodified corn starch or an anionic starch alone.
Only when these reagents are used in combination with calcium ions can quartz
suspensions be flocculated.  To investigate the effect of the level of ad-
dition of 0.93 D.S. carboxymethyl starch, a number of tests were performed
on a quartz suspension of 50 grams per liter at pH 9.5.  The standardized
test procedure including the initial conditioning with 2 pounds of calcium
chloride per ton was followed.  The results are given in Figure 28.  It is
noted that the maximum settling rate was observed at 0.04 pound per ton,
which is approximately one-half of the level of unmodified corn starch ad-
dition of 0.075 pound per ton.  (See Figure 11.)  The amount of suspended
solids remained in excess of 1000 ppm in the presence of the starch.

     The effect of the level of calcium chloride addition was investigated
next by  fixing the starch level at 0.05 pound per ton.  The results thus
obtained are shown in Figure 29.  The maximum settling rate and the mini-
mum level of the suspended solids were observed at 3 pounds oj£ calcium
chloride per ton, beyond which both the settling rate and the amount of sus-
pended solids remained more or less constant.  It appears, therefore, that
the clarity of the supernatant water may not be improved beyond 1000 ppm
when an  anionic starch is used as a polymeric flocculant.

     Effect of Anionic Starch on Goethite.  Goethite suspensions at pH 9.5
could not be flocculated with the 0.93 D.S. carboxymethyl starch at 0.1 or
0.5 pound per ton in the absence of calcium chloride.  This was, more or
less, as expected since both the goethite surface and the starch are nega-
tively charged.


Effect of Dodecylammonium or Aluminum Ion

     In  a previous section it was mentioned that a quartz surface might be-
come hydrophilic in water, and that a quartz suspension might become non-
flocculable.  It is well known in ore flotation that a quartz surface may
be made hydrophobic with a long-chain aklylammonium salt, thereby rendering
the quartz particle floatable upon introduction of air bubbles.  At this
point the sign of the zeta potential on quartz is reversed.26  These two
points, namely: the hydrophobic surface and the reversal of the sign of the
zeta potential, become of particular interest in inducing the flocculation
condition for quartz suspensions.

     A few preliminary tests were performed by conditioning a^quartz suspension
                                    53

-------
      15
Ul
      10
UJ
CO
 E
 ex
 CL
    1500
CO
Q  1000
-l
O
CO
O
UJ
O
£L
CO
3
CO
500
                  I
                        I
                            1.0                  2.0

                  CATIONIC  STARCH  ADDITION, Ib/ton
           (b)
                            1.0                  2.0

                      CATIONIC  STARCH,   Ib/ton
        FIGURE 27.  Flocculation Test Results on Goethite as a Function of
                  0.067 D.S. Cationic Starch Addition at pH 9.5
                             54

-------
o
^c


uT
I-

cc
I-

I-
UJ
CO
 E
 o.
 Q.
   4000
   3000
O
eo


O
LJ
Q
   2000
CO
r>
en
   1000 -

     .
               (a)
                                              O
                          O.I                  0.2

                CARBOXYMETHYL   STARCH, Ib/ton
               (b)
       0                  O.I                  0.2

                CARBOXYMETHYL  STARCH, Ib/ton



       FIGURE 28.  Flocculation Test Results  on Quartz as a Function of
                 0.93 D.S. Carboxymethyl Starch Addition at pH 9.5;
                 Calcium Chloride 2 Ib per  ton
                            55

-------
u
UJ
LJ
CO
 a.
       10
       8
       2



       0
    5000
 .  4000
CO
o

O  3000
CO

o

o
   2000
UJ
0.
CO

CO
    1000
                (b)
                                                           o-
                            468

                        CALCIUM  CHLORIDE,  Ib/ton
                                                           10
                            468

                       CALCIUM  CHLORIDE, Ib/ton
                                                           10
        FIGURE 29.  Flocculation Test Results on Quartz as a Function of
                  Calcium Chloride Addition at pH 9.5; 0.93 D.S.
                  Carboxymethyl Starch 0.05 Ib per ton
                             56

-------
initially with different amounts of dodeeylammonium chloride instead of with
calcium chloride, and then adding 0.075 pound of unmodified corn starch per
ton.  The pH of the suspension was kept at pH 9.5.  The results of the floc-
culation tests are shown in Figure 30.  In the bottom of the figure the re-
sidual concentrations of dodecylamine in the supernatant solution, which were
determined colorimetrically, are included.

     If these results are compared with the results presented in Figure 11,
it is readily apparent that the amount of suspended solids is markedly less,
although further optimization studies are needed to ascertain the minimum
attainable.  An increase in the amount of corn starch to 0.15 pound per ton
resulted in increased turbidity of the supernatant water.

     As will be shown in the chapter covering the streaming-potential mea-
surements, calcium ion is not effective in reversing the sign of the zeta
potential of quartz at near neutral pH.  It is possible only when at rela-
tively high concentrations of calcium chloride and at high pH.  Since alu-
minum nitrate is known to reverse the sign of the zeta potential on quartz
at a relatively low concentration (2.5 x 10~6 M),11 a few cursory tests
were performed to investigate the effect of this electrolyte together with
corn starch, particularly for the clarification of the supernatant solution.
Due to the hydrolysis of aluminum nitrate the flocculation behavior of
quartz appeared to be complexly dependent not only on the aluminum ion and
starch concentrations, but also on the pH.  It becomes of much interest to
pursue this line of approach further for developing a method of clarifying
quartz suspensions, particularly after an alkaline treatment.


                Flocculation Tests Using Synthetic Polymers


     To compare the relative efficacies of synthetic polymers with starch
products for their flocculation and clarification behaviors on quartz the
following four polyacrylamides, viz., Separan NP-10, Separan NP-20, Separan
AP-30, and NC 1733 were selected on the basis of their ionic character and
their nominal molecular weight.17  The flocculation and clarification be-
haviors of a quartz sample with a size modulus of 60 microns were determined
in the same manner as for the starch flocculants, and the results in the
absence and in the presence of 2 pounds of calcium chloride per ton are
presented in Figure 31.  The pH of the suspensions was fixed at 9.5 for com-
parison with the previous data using starch flocculants.

     It is noted in the figure that non-ionic Separan NP-10 is the best floc-
culant for the quartz suspension although anionic AP-30 in conjunction with
calcium chloride approaches the results with Separan NP-10 at low levels of
their addition.  Though the addition of calcium chloride appeared to affect
the settling rate very little, the amount of suspended solids in superna-
tant water, particularly when anionic Separan AP-30 was used together with
calcium chloride, was lowered appreciably.  Cationic NC 1733 was found to be
the least effective of the three synthetic polymers tested.  The level of
the polymer addition at which the maximum of the settling rate is observed
                                    57

-------
.c
o
^c

 «
Ul
      15
      10
UJ
CO
                                         Corn  Starch


                                         0.075 Ib/ton


                                         0.15  Ib/ton
                     1.0          2.0          3.0

             DODECYLAMMONIUM  CHLORIDE,  Ib/ton
                    1.0           2.0          3.0

            DODECYLAMMONIUM  CHLORIDE,   Ib/ton
                                                           4.0
        FIGURE 30.  Flocculation Test Results  on Quartz as a Function
                  of Dodecylammonium Chloride Addition at pH 9.5;
                  Unmodified  Corn Starch 0.075 Ib per ton
                                 58

-------

•| 30
E
_c
u
uj 20
\-
ac.
(D
z 10
i—
i-
Lll
LO
o
XM /
1QOOO
E
CL
Q_
, 8000
LO
Q
Lo 6000

Q
LU
Q
z 4000
UJ
Q_
LO
M 2000

O
i i i i
_ ' . 	 o 	
/"^ * ^^^
Q ^""^ O
/J. 	 &r-.^ *^^^^
A "-A A ^
A fc* /\ ^^ — — , /\ (•«• -^» ••• •«>
— ^^_0-— "~ _
^
/°


/
/O
/ I I I I
(b) f NO 2lb/t
_ / CaCI2 CaCI2 _
A / NP-10 — 0— — •—
I /
i / /\p ~3Q 	 ^ 	 	 A_

- 1
1 / NC 1733 — 0 	
1
I
- I
\
\ /

A ^ ** I
— v —
I
^
^^ ^— — — — ^ *~
"^>«c^ 	 "3Z— — o—
Q^ ^f^--.* -^^flr 0
1 1 1 1
         0.1        0.2       0.3      0.4      0.5

  SYNTHETIC  POLYMER  ADDITION, Ib/ton
FIGURE 31.  Flocculation Test Results on Quartz as a Function
          of Synthetic Polymer Addition at pH 9.5
               59

-------
parallels the level where the minimum of the suspended solids is  recorded.
All these phenomena appear to be explained in terms of the adsorption den-
sity and the conformation of the particular polymer at the interface.

     Quartz surfaces being negatively charged would attract the cationic
functional groups, and, although the adsorption density might be  high, the
conformation of the polymer chain at the interface would not be stretched
out enough to promote effective polymer bridging.  This appears to be the
case particularly when the addition of NC 1773 is low.  With Separan AP-30
the adsorption density must be low for an equivalent addition, but the
chains  are stretched out fully and the polymer bridging results  at much
lower levels of addition.  The fact that the maximum in the settling rate
occurs at very low levels of its addition is in line with this interpreta-
tion.  The presence of calcium ion in solution presumably 'activates1 the
quartz surface for more effective adsorption of the anionic Separan AP-30.
It is readily understood then why the non-ionic Separan NP-10 has the best
overall quality as a flocculant for quartz suspensions.

     Referring back to the flocculation test results on the same  quartz sus-
pension using the unmodified corn starch (Figure 11), cationic starches of
two different D.S. (Figure 26), and anionic starch (Figure 28), it is noted
that the settling rates at maximum are seen to be in the range of 7 to 15
inches per minute which are about one half of those observed with the syn-
thetic polymers.  This difference must be attributable to the floe size,
governed by the molecular size of the synthetic polymers and of the starches.
It becomes of interest to test at least another synthetic polymer of dif-
ferent molecular weight, say, Separan NP-20 with the nominal molecular weight
of 2 x 106.

     For this reason two series of tests were carried out with Separan NP-20
at pH 9.5:  in one series in the absence of calcium chloride and  in another
in the presence of 2 pounds of calcium chloride per ton.  The results in the
absence of calcium chloride are plotted in Figure 32 together with those with
Separan NP-10 for comparison.  The settling rate curve of Separan NP-20 ap-
pears to be somewhat higher than that of Separan NP-10 over the concentration
range investigated, which is in line with the trends of the effect of mole-
cular weight mentioned previously.  The suspended solids in the supernatant
liquor, however, remained appreciably higher with Separan NP-20 than NP-10.
In the presence of 2 pounds of calcium chloride per ton (Figure 33) the two
polymers appeared to behave almost identically towards the quartz suspension.
In all cases, however, the amount of suspended solids was quite high and re-
mained in excess of 600 ppm throughout the range.

     The effect of pH on the flocculation of quartz with Separan NP-10 was
studied briefly by performing two additional series of tests at pH 4 and 11.
The results together with those at pH 9.5 in the absence and in the presence
of 2 pounds of calcium chloride per ton are presented in Figures  34 and 35,
respectively.  In all, the effect of pH is seen to be relatively unimportant.
except at pH 11 and in the absence of calcium chloride where the amount of
suspended solids is rather high.
                                   60

-------

c 30
£
"*H
u
c
, r 20
UJ
h-
cc
o
| 10
I—
\—
UJ
CO

0
O/^v/^v/^
oOOO

E
a.
DL
-4000
to
O

o
CO
S 2000
a
-X'
UJ
Q_
CO
Z)
CO
, . \ \ \ \
(a)
— A^""^— • — " — ° — "^""•^cr """•"- -^ ~
/Qx^^O ^ ^-^^^^
A/ °-
/a
/
— —



—o— Separan NP-10 -
—A— Separan NP-20
No CaCl2
\ \ \ \
0.1 0.2 0.3 0.4

\ \ \ \
(b)
,-~A~~
^ **
•• ••
— ~'~~ ~
/\«*^
x ^
i . "
A
i
- *» -A
\ V ' ^-^*1 	
O l\f ^^r*^*^"^
V -^ ^^"^'^
O ^-ii^1^^*
*X*-^-Q_ — 	

1 1 1 1
0         0.1       0.2       0.3      0.4
    SYNTHETIC  POLYMER ADDITION, Ib/ton
 FIGURE 32.  Flocculation Test Results on Quartz as a  Function
           of Synthetic Polymer Addition at pH 9.5
                   61

-------
c
E
2:
u
c
LU
I—
cr
o
LU
CO
30
        20
        10
                            —•— Separan NP-10
                            —A— Separan NP-20
                              2 Ib/ton CaCl2
         0
            0.1
0.2
0.3
0.4
          0         0.1        0.2       0.3      0.4
                SYNTHETIC POLYMER ADDITION. Ib/ton
ouuu
E
CL
Q.
to
9 4000
0
CO
Q
LU
§ 2000
LU
Q_
CO
=)
CO

1 1 1 1
(b)



— —

^
^ — """""T
^
- j • ^^^^
* ^ — •**
». A^^^
»_^, ^


i i i i
            FIGURE 33.  Flocculation Test Results on Quartz as  a Function
                     of Synthetic Polymer Addition at pH 9.5; Calcium
                     Chloride 2 Ib per ton
                          62

-------

| 3O
U
c
u 20
h-
o:
o
i io
UJ
CO
0
E6000
Q.
Q.
CO
§ 4000
CO
SUSPENDED
IV)
O
O O
(a) '
- D^A^""*30^— -^^ °
D/0 0 *D-
_/0

— o--pH4
-A- pH 9.5
NoCaCL
i i i L i
0.1 0.2 0.3 0.4
(b) '
,n"'"
/
D/
/ D
/
\ /
\ /
n /
— \x i 	 A—
1 1 1 1
0.1 0.2 0.3 0.4
SEPARAN NP-10 ADDITION Ib/ton
FIGURE 34.   The Effect of pH on the Flocculation  Test  Results
            on Quartz as a Function of Separan  NP-10 Addition
                     63

-------
30
c:
E
u
•s 20
UJ
t—
O
| 10
1—
Ld
(S)
0
8: 6000
co"
Q
_J
^ 4000
Q
LU
Q
&i 2000
LO
1 1 1 1
(a)
/ 	 A*^. 	 •
du : .• — «^-
^/r
-•^ PH 4
-A- pH 9.5
-•- pH 11
2 Ib/ton CaCt2
i i i i
0.1 0.2 0.3 0.4
i i i i
_(b)

— _
• 	
^ 	 *~ A
A^. 	
~\ ^-" • ~~
\ ^^"***^ A A
A x"**^ * *
^•^••A
I I I I
0         0.1        0.2      0.3       0.4
    SEPARAN   NP-10  ADDITION,  Ib/ton
 FIGURE 35.  The Effect of pH on the Flocculation Test Results
           on Quartz as a Function of Separan NP-10 Addition;
           Calcium Chloride 2 Ib per ton
                   64

-------
                Flocculation Tests on an Artificial Mixture
                	of Quartz and Goethite	


     Two series of preliminary flocculation tests were performed on an arti-
ficial mixture of quartz and goethite.  In one series an attempt was made to
flocculate both the quartz and the goethite.  The possibility of mutual floc-
culation of the two minerals and its effect on the settling rate and clarity
were also explored.  In the other series the selective flocculation of goethite
and partial upgrading by desliming were tested.


Total Flocculation - Clarification

     Destabilization of silica suspensions may be brought about by a proper
choice of the concentration of an iron salt and of pH.^7  in the alkaline
pH range a three-dimensional network formation due to the sorption of hydro-
lyzed iron (III) polymers was thought to be responsible for the particle ag-
gregation.  To ascertain if a mixture of quartz and goethite might result in
an improvement of the flocculation behavior, particularly of quartz, a series
of artificial mixtures of the two minerals with size moduli of 60 microns was
prepared.  The experimental conditionsselected were based on the information
reported previously, namely, 50 grams of the artificial mixture per liter at
pH 9.5, two pounds of calcium chloride per ton, and a level of corn starch
addition proportioned to the ratio of quartz and goethite at their maximum
settling rates.  The corn starch level was shown to be 0.075 pound per ton
for the quartz sample and 0.75 pound per ton for the goethite sample (Fig-
ure 19).  The results thus obtained are shown in Figure 36(a).  The experi-
mental points, both of the settling rate and suspended solids, appear to be
represented with lines convexly upwards, i.e., the settling rate of the mix-
ture appears to have been accelerated by the presence of goethite, but the
presence of goethite did not seem to affect the clarity of the supernatant
water.

     Previous tests have demonstrated that cationic starches are effective
flocculants for both quartz and goethite and that cationic starches with
high degrees of substitution are most efficacious (Figures 26 and 27).  Sim-
lar tests were performed on a series of quartz and goethite mixtures under
essentially identical conditions except for the omission of calcium chloride
addition.  For the flocculant,0.067 D.S. cationic starch was selected and
the level of its addition was proportioned according to the ratio of the two
minerals based on the condition observed at the maximum settling rates.  These
levels were 0.5 pound per ton for the quartz and 1.0 pound per ton for the
goethite (Figures 26 and 27).  The results are shown in Figure 36(b).  It is
apparent that the curve for the settling rate is convex upwards and that for
the suspended solids concave upwards, showing that there is a beneficial ef-
fect of goethite on the flocculation of the quartz suspension.


Selective Flocculation - Desliming

     In some of the flotation tests with the Mesabi iron ores the addition
                                   65

-------

LU
O
LU
LO
 o
 c
or

O
LU
CO
    0
         (a) CORN STARCH
         v '
\     —I
 \
   \
    \
                                              1OOOQ
                                                    Q
                                              500
                                                    LU
                                                    LO
   10
      8
         (b) 0.067 DS CATIONIC STARCH
    0         25        50      75     100
    PERCENT GOETHITE IN THE MIXTURE
                                              o
     FIGURE 36.  Flocculation Test Results on  the
     Artificial Mixture of Goethite and Quartz (k=60y)
     in various Proportions

     (a) pH 9.5, CaCl2  2 Ib per ton, and Corn Starch
         0.075 to 0.75 Ib per ton
     (b) pH 9.5, 0.067 DS Cationic Starch 0.5  to 1.0
         Ib per ton
                   66

-------
of causticized corn starch was observed to result in two relatively well-
defined layers of sediment upon standing.9  In the present investigation
it was shown that goethite can be flocculated with corn starch in the ab-
sence of calcium ion whereas quartz cannot.  It appears, therefore, that
goethite may be selectively flocculated with corn starch and that quartz
particles remaining in suspension may be removed by decantation.   Based
on the results presented in Figures 10, 13, and 14 the test conditions were
selected as follows.  Fifty grams of an artificial mixture of goethite and
quartz samples with a size modulus of 60 microns was placed in a 1000-ml
graduated cylinder and diluted with make-up water to near the 1000-ml mark
so that the subsequent addition of reagent would give a total pulp volume
of 1000 ml.  At this point the pH of the pulp was adjusted either with so-
dium hydroxide or hydrochloric acid, which was followed by the addition of
starch in three equal portions, mixing being performed between each addition
by gently inverting the cylinder three times between each addition.  After
the final addition of starch, the cylinder was inverted five times and set
down.  The settling of the goethite and quartz mixture began immediately.
After one minute the mud line had reached the  bottom. The tip of a siphon
was then lowered to the 900-ml mark and the supernatant water above the mark
was siphoned off.  Another 900 ml of water3 whose pH had been preadjusted
to the same value, was added to the remaining solids.  The cylinder" was in-
verted five times, set down, and the settling and siphoning procedure was
repeated.  The desliming process was performed six times.

     Initially, a series of three tests was carried out to explore the effect
of the pulp pH by keeping the level of starch addition constant at 0.375
pound per ton:  the first at natural near neutral pH, the second at pH 9.5,
and the third at pH 11.  The weights of each slime fraction and their ana-
lytical results are presented in Table 9, Sections A, B, and C, and are
depicted in Figure 37(a) , in which the percent iron recovery in the sand
fraction is plotted against the percent iron in the sand fraction.  The ver-
tical line at 28.3 percent iron shows the line of no selectivity.   The ex-
perimental curves extend markedly towards higher grades of the sand frac-
tion, indicating the selectivity of the desliming operation.  As seen in the
figure, the grade-recovery curves are displaced further to the right as the
pulp pH is raised indicating that the selective desliming operation is rea-
sonable since the quartz suspensions are known to become stable and difficult
to clarify after an exposure to the alkaline medium.  Though this condition
assists the selectivity of the desliming operation, a rather serious problem
may be anticipated when the water is to be recycled or to be released from
the plant.  It is also noted in the curves at pH 9.5 and 11 that after about
three decantations the line tends to become steeper indicating no effective
upgrading is obtained through further desliming.

     The effect of the level of starch addition was investigated next in a
series of tests at pH 9.5.  A control test was performed without the starch
addition.  The test results are given in Table 9, Sections D, E, and F, and
depicted in Figure 37(b).  It is readily apparent that in the absence of the
starch there is no selectivity in the desliming operation.  In fact, the
experimental curve slopes towards lower grades of the sand fraction somewhat,
indicating the reversal in the selectivity of the separation.  With the ad-
                                   67

-------
TABLE 9.  SELECTIVE FLOCCULATION AND PARTIAL UPGRADING RESULTS
          ON AN ARTIFICIAL MIXTURE OF GOETHITE AND QUARTZ
          (k = 60y)


1st Desliming
2nd Desliming
3rd Desliming
4th Desliming
5th Desliming
6th Desliming
Sand
Composite

1st Desliming
2nd Desliming
3rd Desliming
4th Desliming
Sand
Composite

1st Desliming
2nd Desliming
3rd Desliming
4th Desliming
5th Desliming
6th Desliming
Sand
Composite
% Wt
A,
3.06
2.30
2.01
1.64
1.44
1.45
88.10
100.00
B.
7.63
10.27
8.94
5.24
67.92
100.00
c.
21.11
10.94
5.09
3.92
3.80
3.36
51.78
100.00
% Fe
. pH ^
2.09
8.97
13.85
17.23
19.88
22.62
30.83
28.61
pH 9.5
4.97
4.11
9.74
24.95
37.03
28.13
pH 11;
4.59
6.76
16.54
32.20
40.89
45.41
41.57
28.42

Slime
Cum
% Wt
Fraction
Cum Cum
% Fe Fe Rec
Sand Fraction
Cum
% Wt
Cum
% Fe
Cum
Fe Rec
7; Corn Starch 0.375 Ib/ton
3
5
7
9
10
11


; Corn
7
17
26
32


Corn
21
32
37
41
44
48


.06
.36
.37
.01
.45
.90


2.09
5.04
7.45
9.23
10.69
12.15


Starch 0.375
.63
.90
.84
.08


Starch
.11
.05
.14
.06
.86
.22


4.97
4.48
6.23
9.29


0.375
4.59
5.33
6.87
9.29
11.96
14.29


0.22
0.94
1.92
2.91
3.91
5.05


Ib/ton
1.35
2.85
5.94
10.59


Ib/ton
3.41
6.01
8.97
13.42
18.88
24.25


96
94
92
90
89
88



92
82
73
67



78
67
62
58
55
51


.94
.64
.63
.99
.55
.10



.37
.10
.16
.92



.89
.95
.86
.94
.14
.78


29.44
29.94
30.29
30.53
30.70
30.83



30.04
33.29
36.16
37.03



34.79
39.31
41.15
41.74
41.80
41.57


99
99
98
97
96
94



98
97
94
89



96
93
91
86
81
75


.78
.06
.08
.09
.09
.95



.65
.15
.06
.41



.59
.99
.02
.58
.12
.75


                               68

-------
TABLE 9.  (CONTINUED)
Slime Fraction


1st Desliming
2nd Desliming
3rd Desliming
4th Desliming
5th Desliming
6th Desliming
Sand
Composite

1st Desliming
2nd Desliming
3rd Desliming
4th Desliming
5th Desliming
6th Desliming
Sand
Composite

1st Desliming
2nd Desliming
3rd Desliming
4th Desliming
5th Desliming
6th Desliming
Sand
Composite
% Wt

20.13
4.65
1.40
0.65
0.36
0.47
72.34
100.00
E
23.84
10.54
4.97
3.75
3.02
1.49
52.39
100.00
F.
14.95
9.11
5.89
3.61
2.98
2.51
60.95
100.00
% Fe
D. pH 9
32.48
32.16
32.64
32.06
31.41
28.85
27.01
28.49
. pH 9.5;
11.99
31.87
46.69
49.23
49.18
49.84
29.62
28.13
pH 9.5;
2.54
2.66
5.88
9.02
16.18
20.77
42.74
28.35
Cum
% Wt
. 5 ; Corn
20.13
24.78
26.18
26.83
27.19
27.66


Cum Cum
% Fe Fe Rec
Starch
32.48
32.42
32.43
32.42
32.41
32.35


Corn Starch 0.1
23.84
34.38
39.35
43.10
46.12
47.61


11.99
18.08
21.70
24.09
25.74
26.49


Corn Starch 0.65
14.95
24.06
29.95
33.56
36.54
39.05


2.54
2.59
3.23
3.86
4.86
5.88


None
22.95
28.20
29.81
30.54
30.93
31.41


Ib/ton
10.16
22.10
30.35
36.91
42.19
44.83


Ib/ton
1.34
2.19
3.42
4.56
6.27
8.10


Sand Fraction
Cum
% Wt

79.87
75.22
73.82
73.17
72.81
72.34



76.16
65.62
60.65
56.90
53.88
52.39



85.05
75.94
70.05
66.44
63.46
60.95


Cum
% Fe

27.48
27.19
27.09
27.04
27.02
27.01



33.18
33.39
32.30
31.19
30.18
29.62



32.88
36.51
39.09
40.72
41.87
42.74


Cum
Fe Rec

77.05
71.80
70.19
69.46
69.07
68.59



89.84
77.90
69.65
63.09
57.81
55.17



98.66
97.81
96.58
95.44
93.73
91.90


          69

-------
   100
    90
8   80
LU
OX

Z
o

s   70
\—
Z
LU
O
QL
LU   c: r\
0-   60
	 IU




:n I
V~^o-— -i_ i i i i i
% — °— A-___
°x ~~~"~~A--^
\ X
0 \
X A
>, \
> A
u '
Q; /
~3T /
1) '
1/1 / \ A
o (a)Effect of pH /
c (Corn Starch: 0.375 Ib/ton)
H— —
I -0— PH-7
c
-J — 0 — pH 9.5
— A — pH 11
I I I I I I I
IUU



90


80



70


60

»,
T—— ^Q-^^^-— -1 	 V 1 I' 1 1
^"^ 	 o 	 ^^^ ~~v 	 -v_
^ ^-^_ v "^-^JT
i \ ^^-. n AA. ^
t i \ U. ^V7
i
1
i
— '
i
i
1
i
1
I
i
i
i x X -v
\A \
1 A o
\
i
* i
- ! i / ^
f>
i
i
yx
~K
!

—
o A
-§ / (b) Effect of Starch Level
w / (pH 9.5)
° /
^ A7
"° / 	 O— - None
0) /
•_§ ^f —A—- 0.1 Ib/ton
/ — 0 — 0.375 Ib/ton
s
&' 	 V 	 0.65 Ib/ton
/
nl 5 , i i i i i i i i
0 28 30 32 34 36 38 40 42
PERCENT IRON IN SANDS
     FIGURE 37.
Selective Flocculation and Partial  Upgrading  Results on an Artificial
Mixture of Goethite and Quartz (k=60p);  (a) Effect of pH at a Constant
Starch Level  of  0.375 Ib per ton, (b)  Effect  of the Level of Starch
Addition at pH 9.5
                                        70

-------
dition of the starch, the curves extend towards higher grades of the sand
fraction.  The higher the level of the starch addition, the further the
shift of the grade-recovery curves to the right.  This again is in good
agreement with the test results on single mineral systems; a quartz sus-
pension cannot be flocculated with starch in the absence of calcium ion.
It becomes of interest to investigate the effect of calcium ion in solu-
tion, which is inevitably present in all the iron ores and causes quartz
particles to flocculate in the high pH region.
                       Zeta-Potential Measurements
     Zeta potentials of goethite and quartz in the absence and in the pre-
sence of corn starch and/or calcium chloride at different pH were determined
by the streaming-potential measurement technique.  To compare the electro-
kinetic behavior of the ground samples of St. Peter sand and Brazilian quartz,
which were too fine to apply the above method, the electrophoretic mobility
measurement technique was used.  The results are reported under the headings
of goethite and quartz.


Zeta Potential of Goethite

     Figure 38 shows the zeta potential of goethite as a function of corn
starch at neutral pH and at pH 10.7.  Both curves show the same trend—much
of the change in the zeta potential occurs when the starch concentration is
less than 10 mg per liter.  This is in good agreement with the adsorption
isotherms of the same starch on both quartz and hematite.15  The abstrac-
tion at low levels of addition is nearly complete, resulting in a steep rise
in the adsorption density, and then approaches saturation coverage corres-
ponding to a particular pH of the solution    (Figure 51), The zeta potentials
in the plateau region correspond to the adsorption densities at the satura-
tion coverages.

     Marked difference in the zeta potential in the plateau region at the
two pH values appears to provide an implication on the conformation of starch
molecules on the goethite surface.  When the goethite sample was equilibrated
with a solution  containing 50 mg of starch per liter at pH 10.7, the zeta
potential was highly negative and was about  -33 mV.  Lowering of the solu-
tion pH, but maintaining the constant starch  concentration of 50 mg per  liter,
however, resulted in the decrease in the zeta potential  approaching a plateau
of about -8 mV in the near neutral pH as shown in Figure 39.  This value of
zeta potential corresponds to that at pH 6.7 in Figure 38.  When the direc-
tion of the pH change is reversed, but at a  starch concentration of 50 mg
per liter, the zeta potential increases only to  a range  of -10 to -12 mV.
This hysteresis  effect may be readily understood if the  irreversibility  of
the starch  adsorption  is  assumed.  The  zeta  potential .values  of  th-e
"increasing pH"  cycle may represent  those of tLe adsorbed  starch itself.
The starch molecules adsorbed  from the high-pH  solution mist be  highly
stretched,  and the  zeta potential observed  at that pH  appears  to be  influ-
enced by the highly negative surface of goethite.  The adsorption  density
                                    71

-------
-60
-50
-40
> -30
_f
P
LU ~20
1 	
0
Q_
£ -10
LU
N
0
+10*
<
A
\\
A^^A pH 10.7


"

Q Q 	 O —
8'*"^ pH 6.6 (duplicate)
f
I
I I I I I
20      40       60
    CORN  STARCH, mg/
                                        80
100
         FIGURE 3&  Zeta Potential of Goethite as  a Function of
                  Corn Starch Concentration

-------
   -60
   -40
UJ
S
Q_
UJ
N
-20
      0
    20
                 NO
               STARCH
t
                           DECREASING pH   _
                     50 mg/l CORN STARCH
                                            T-V-
                                               ^H

                                               INCREASING
                     7
                            8
                            10
11
                                 PH
         FIGURE 39.  Zeta Potential  of Goethite as a Function of pH in the
                   Absence and in  the Presence of 50 mg of Corn Starch
                   per Liter
                                73

-------
increases and the conformation of the adsorbed starch becomes  less  stretched
as the pH of the solution is lowered  and the zeta potential  is  also lowered.

     Figure 40 shows the effect of calcium chloride concentration on the
zeta potentials at the two pH values  both in the absence and the  presence  of
starch.  At pH 6.6 in the absence of  starch the zeta potential  remained posi-
tive irrespective of the calcium chloride concentration.  At pH 10.7 the goe-
thite surface was negative, but in calcium chloride solution in excess  of
4 x 10"1* M, the sign of the zeta potential was reversed to become positive.
In the presence of starch, however, the zeta potential remained negative over
the calcium concentration and the pH  range investigated.  The  decrease  in  the
zeta potential above 10~5 M may be due partly to the effect  of calcium  ion on
the electrochemical property of starch and partly to the compression of the
electrical double layer.


Zeta Potential of Quartz

     Initially, the streaming-potential measurements were unexpectedly  ham-
pered by the choice of the quartz sample and by the use of an  ordinary  dis-
tilled water.  The cause of the anomalous behavior was explored by  changing
the surface cleaning procedure, by using a different quartz  sample, and fi-
nally by using demineralized water instead of distilled water.

     Three types of quartz, viz., St. Peter sand, Montana pegmatite quartz,
and Brazilian rock crystal, were used.  The St.  Peter sand was  cleaned  in
two ways.  One portion was treated in a warm solution of 1 N hydrochloric
acid (Sample A), and another was boiled in concentrated hydrochloric acid
(Sample B), each for one-half hour.  The Montana pegmatite quartz was treat-
ed in a hot concentrated hydrochloric acid for one-half hour^  and the Bra-
zilian quartz in the same manner for  one-quarter hour.

     Zeta potentials were determined  as a function of pH, adjusted with so-
dium hydroxide.  When distilled water was used to prepare the  solution, the
zeta potentials of St. Peter sand (Sample A)  never exceeded  -90 mV and  a
peculiar break in the curve occurred  near pH 10 (Figure 41).   The zeta  po-
tentials beyond pH 10 remained nearly constant at -60 mV. After  a severer
treatment of St.  Peter sand (Sample B), the zeta potentials  were  appreciably
higher and reached a maximum near pH  9, then decreased rapidly towards  the
steady value of -60 mV.  The Montana  quartz followed a similar trend, which
suggested that this anomalous behavior might originate from  certain impuri-
ties in the solution.  In fact, at the pH where the break in the  curves was
observed, a faint yellow coloration was seen on the quartz sample.   This
might be attributed to the presence of a minute quantity of  iron  in the dis-
tilled water.  It should be remembered, however, that the coloration was due
to a gradual accumulation after the multiple passage of a large volume  of
the solution during the streaming-potential measurements.

     In order to check this particular point the measurements  were carried
out on the Montana pegmatite quartz and the Brazilian quartz,  using demin-
eralized water.  For the Brazilian quartz the measurements were replicated
                                   74

-------
   -60'
>
    -40
           No Starch pH 10.7
         50 mg/l  Starch  pH 10.7
 LU
 H
 O
 Q_
 LU
 N
-20
         50mg/l  Starch pH 6.7
    o.	o   _
      0
         No  Starch  pH 6.6
     20
       10-7       ft)'6       10'5      10'4        10'3       10'2       10'1

                CALCIUM  CHLORIDE CONCENTRATION, M
           FIGURE 40.  Zeta Potential of Goethite as a Function of Calcium Chloride Concentration
                    in the Absence and in the Presence of 50 mg of Corn Starch per Liter

-------
-140
-120
-100
-80
1
_i
< _£}Q
I—
Z
LU
i 	
0
°- -40
LU
N
-20


0


i ill
o ^
^%r-k ° ^v\ V "
/ N\_ v
/A^-~~ r^^ \ ^
V'"""*~~"~\ "X \T
/O '* NN \ x
4 X A X

A A"A
— —
Distilled Dem in.
Water Water
St. Peter Sand (A) —A 	 A — -
St. Peter Sand (B) — v 	 v—
Montana Quartz --f — — 0 —
Brazil Quartz — O —
Increasing pH — € —
Decreasing pH — 3 —
ill i
6 7 8 9 10 11
                                  PH
FIGURE 41.   Comparison of Zeta Potentials  of  Various Quartz Samples as a
            Function of pH in Distilled and Demineralized Water
                               76

-------
a few times, once completing a cycle of increasing the pH and then decreasing
the pH on an identical sample for a hysteresis effect.  All the results are
plotted in Figure 41 together with those obtained in distilled water for com-
parison.  It is readily apparent that the general shapes of the zeta-potential
curves in demineralized water are quite different from those obtained in dis-
tilled water, and that the data of the Montana pegmatite quartz and those of
Brazilian quartz are in good agreement..  In fact, the present results with the
Brazilian quartz are virtually identical to those reported in literature on
the quartz sample from the same source and conductivity water.11  Furthermore,
the zeta potentials of the increasing pH cycle, indicating the absence of the
hysteresis effect.

     The difference between Samples A and B of St. Peter sand in demineralized
water might be attributed to the presence of a minor quantity of impurities
present on the original sand grains which was removed by boiling the sand in
concentrated hydrochloric acid.  It becomes of interest to compare the zeta
potentials of a finely ground St. Peter sand with that of Brazilian quartz
using an electrophoresis technique.  Here the freshly fractured surfaces
of different quartz samples may be compared.  Furthermore, when a quartz sam-
ple is ground fine and a limited amount of distilled water is used, the effect
of a minor quantity of impurities present in distilled water might become
negligible.

     In order to Ascertain this particular point, the electrophoretic mo-
bilities of ground samples of St. Peter sand dispersed both in distilled
and in demineralized water were determined as functions of pH.  The results
were then compared with the electrophoretic mobilities of ground Brazilian
quartz in demineralized water.  During the course of the experimental work
it was noted that the electrophoretic mobility increased gradually with time.
It appeared that the suspension in demineralized water required the longest
period, extending from a few days for the moderately alkaline solution to
well over a week for others.  Three series of determinations were made to
test the reproducibility of each condition.  The results are plotted in Fig-
ure 42(a).  Although the scatter of the experimental points is relatively
small within each series of determinations, the curves appear to be displaced
vertically depending on the concentration of the suspended solids; the thin-
ner the suspension, the higher the mobility.  Nevertheless, all the points
corresponding to the three sets of conditions are seen to scatter about a
line drawn through the experimental points, indicating that the electrophor-
etic mobility of quartz does not depend on the types of quartz or on the kind
of water used.  Furthermore, it is noted that, unlike the streaming potentials
shown in Figure 41,above a pH of 9 in distilled water the electrophoretic
mobility did not decrease.

     The mobility data were then converted to the zeta-potential values
using two limiting cases22 of the von Smoluchowski and Htlckel equations.
In Figure 42(b) the zeta-potential values calculated on the basis of Htickel's
treatment (with a factor of 6) are plotted.  A broken-line curve was drawn
through the plotted values.  The same curve, but corresponding to the other
limiting case of von Smoluchowski's treatment (with a factor of 4) was drawn
with a dotted line.  It is readily apparent that the former is in closer
agreement with the zeta-potential curves of the Brazilian quartz and of the
                                   77

-------
>
1
— 1
CD
O
^
U ,_
j= E
UJ^
o: —

^~§
Q_ CD
O^
H ^
O
1 1 1

UJ

-8


-6



-4



-2




O

i i i i i i
(a)

0°
O A 	 O A *•
A A ___ — '$£~~ ^ /^+
A AO^A-- "° o^ ° «^ \
— Q'^^'^ J • *\ ~
A ^ 	 0 • 0™
• ^ — O
-" ° Distilled Demin.
^ Water Water

^ 4- r^r~i •fr~ii-^"~imiH A _.^
oi. reier oana • o

Brazil Quartz — A —

i i i i i i
   -120




> -100



p  -80
~z.
UJ

§  -60
uj  -40
    -20
         (b)
                           St. Peter  Brazil
                             Sand   Quartz
 ._,  ,   ,      .   y STTTJ
 Electropheresis  £=   € 7u	o--	A--

       Streaming  Potential	
      Ol	L
           6
   7
8
9
10
11
                                pH
        FIGURE 42.
Electrophoretic Mobility and Zeta Potential of a Ground
St. Peter Sand and a Brazilian Quartz as a Function of pH
                              78

-------
St. Peter sand, Sample B, in demineralized water, determined by the stream-
ing-potential technique.  This was, more or less, to be expected since the
particles observed in the measurements were extremely small.

     To further confirm the effect of using distilled and demineralized water
on the flocculation behavior of quartz, a few tests were performed at pH 9.5
and 11.5 under corresponding conditions.  Both the settling rates and the a-
mount of suspended solids in supernatant liquor were compared.   Again, pre-
sumably to the large surface area of the quartz sample and the  relatively
small quantity of water used in the flocculation tests, the flocculation be-
havior was hardly affected by the type of water used.

     Having ascertained that Brazilian quartz in demineralized  water behaved
satisfactorily, the effects of corn starch, calcium chloride, and pH on the
zeta potentials were investigated.  Figure 43 shows the effect  of corn starch
on the zeta potentials at three pH values in the absence of calcium chloride.
The zeta potentials both at pH 9.5 and 11.1 are affected very little by the
presence of corn starch, which is in line with the adsorption data that the
adsorption density of corn starch on quartz is extremely small.   At pH 5,
however, the zeta potential decreases as the concentration of the starch is
increased, more or less, contrary to that at pH 9.5 or 11.1. This behavior
is indicative of some adsorption of the starch at this pH.  In  fact, the
adsorption of starch on quartz is known to increase as the pH is decreased.15

     The effect of pH on the zeta potentials of Brazilian quartz in the pre-
sence of 50 mg of corn starch per liter was determined, and the results are
presented in Figure 44 together with those in the absence of starch.  The pH
of the solution was adjusted using either hydrochloric acid or  sodium hydrox^-
ide.  In the absence of starch the zeta potential is seen to remain negative
in the pH range from 2 to 11, indicating that the point-of-zero-charge of
quartz is located at a pH below 2.  This observation is in agreement with the
results reported by Li and de Bruyn.12

     In the presence of starch the zeta potentials were lowered somewhat over
the pH range investigated.  In the alkaline pH range the experimental points
determined in the direction of increasing pH and of decreasing  pH are seen to
be nearly coincident, indicating that the hysteresis effect observed on goe-
thite is absent on quartz in this pH range due presumably to rather low ad-
sorption of the starch molecules,  In the acid range, however,  the zeta po-
tentials of increasing pH are seen to be appreciably lower than those of de-
creasing pH, showing, more or less, the same trend as observed  on goethite.

     Figure 45 shows the effect of calcium chloride concentration on the
zeta potentials at four pH values in the absence of corn starch.  It is in-
teresting to note that the zeta potentials remained negative at pH 4 and near
7, irrespective of calcium chloride concentration, but at pH 9.5 and in a
solution of calcium chloride exceeding 10 2 M the sign of the zeta potential
on the quartz was reversed to become positive.  At pH 11 the solution con-
centration of calcium chloride, at which the sign_of the zeta potential is
reversed is seen to be considerably lower than 10 2 M, and is estimated at
2.5 x 10"3 M.  The zeta potentials under identical conditions,  but in the
                                   79

-------
  -120
 - -80
H
Z
LU

o
Q_
       D
   -60
I-
LJ -40
N
   -20
     0
       0
         1
1
                                             9.5
                                             i
        20       40       60      80
             CORN STARCH, mg/l
                                                     100
       FIGURE 43.  Zeta Potential of Brazilian Quartz as a Function of
                Corn Starch Concentration
                            80

-------
-140
-120
-100
 -80
_J
<
I—
z
UJ


Q_

<
\-
LU
N
 -60
 -40
 -20
   0
                       DECREASING  pH
                         INCREASING pH
                                 	O— NO STARCH, INCREASING pH

                                 — O— NO STARCH, DECREASING  pH  _

                                 —A— 50mg/l STARCH, INCREASING pH

                                 — T— 50mg/l STARCH DECREASING pH
                                      I
                                      8
                                                   10
12
                                PH
     FIGURE 44.  Zeta Potential of Quartz as a  Function of pH in the
                Absence  and in the Presence of 50 mg of Corn Starch
                per Liter
                             81

-------
    -120
      i         r
—v—pH 4
—o— pH~7
          9.5
                                        —n— pH 11.1
LU
5
Q_
UJ
N
        10'7     10'6      1Cr5      10'4      10'3      10'2
         CALCIUM  CHLORIDE  CONCENTRATION, M
              FIGURE 45.  leta Potential of Brazilian Quartz as a
                       Function of Calcium Chloride Concentration
                           82

-------
presence of 50 mg of corn starch per liter, are presented in Figure 46.  As
evident, the zeta potential did not reverse its sign up to 10 2 M in all four
cases, but their magnitude is affected appreciably.  Due presumably to the
stronger interaction with calcium ion at high pH, the quartz surface adsorbed
a greater amount of corn starch, thereby approaching the zero zeta-potential
line at lower concentrations of calcium chloride.  Here again, strong inter-
action of the adsorbed calcium ion and starch adsorption is indicated.  These
observations are in line with the results of the flocculation tests given in
Table 1 and Figure 10:  that the most effective flocculation condition is ob-
tained in highly alkaline solution at high calcium chloride concentrations.

     Since the cationic starch was found to be quite effective in clarifying
the quartz suspensions, cursory streaming-potential measurements were made on
the Brazilian quartz using a 0.067 D.S. cationic starch.  Three series of ex-
periments were performed in which the solution pH was kept constant at 5.0,
9.5, and 11.2.  The results are shown in Figure 47.  It is readily apparent
that the zeta potential changes its sign at very low concentrations of the
cationic starch, and that the zeta-potential readings remain constant at
higher concentrations, although the constant values are dependent on the pH.
Since the saturation coverage by the cationic starch of quartz is known to
increase with the solution pH,^5 the difference in the zeta potentials at the
three pH values may be attributed to the difference in the adsorption den-
sities.  It becomes of much interest to test the effects of the de-
gree of substitution of the cationic groups and of the cationic polyelectro-
lyte.


                          Viscosity Measurements


     To investigate the effect of pH and calcium chloride concentration on
viscosity, a stock starch solution was diluted and its pH was adjusted to a
desired value so that a final concentration of 0.4 percent by weight was ob-
tained.  Hydrochloric acid and sodium hydroxide were used as pH regulators.
The results of the relative viscosities as functions of pH are given in Fig-
ure 48(a).  A sharp viscosity minimum is observed near pH 3.5, which may
correspond to the isoelectric point of the solubilized starch.28

     The effect of calcium chloride concentration on the relative viscosity
was determined at pH 7 and 11, and the results are given in Figure 48(bJ.
It is interesting to note that the relative viscosity at pH 7 remained  more
or less  constant, regardless of the calcium chloride concentration.  At pH
11, however, the relative viscosity decreased as the calcium chloride con-
centration was increased indicating that an interaction between the nega-
tively-charged starch and the positively-charged calcium ion is more pro-
nounced at high pH.  This behavior, which is known as the electroviscous ef-
fect,28 implies that the starch molecules become more tightly coiled as the
calcium-ion concentration is increased.

     The change in the conformation of starch molecules at a high concentra-
tion of calcium chloride may influence the flocculation behavior.   To check
                                   83

-------
                                             T
   -120-
  -100'
   -80
                  •v.
                    v\
                                           T-PH4

                                          -•— pH ~

                                           A   pH
E
_T
<
h-
LU

o
Q_
LU
Kl
                         \ \
   -60
    -40
   _20
      0
       10
         "7      10'6      10"5      10"4      10"3      10"2

             CALCIUM  CHLORIDE  CONCENTRATION, M

        FIGURE 46.  Zeta Potential of Brazilian Quartz as a Function of
                  Calcium Chloride Concentration in the Presence of
                  50 mg/1 Corn Starch
                          84

-------
   -12O
       8
   -1OO
LU
2
LU
rsi
    -80
   -60
       A
2  -40
   -20
      0
        i •
        0-0
    20
A
 D-
-A-
    40
        1
                         -A-pH 5.0
                         -D-pH9.5
                         -O-PH 11.2
                     •O-
-O-
                              -D-
                              A-
-D-
-A-
_J	
       0       20       40       60       80      100
             OO67  D.S.  CATIONIC  STARCH, mg/liter
         FIGURE 47. Zeta Potential of Brazilian Quartz as a Function
                  of 0.067 D.S. Cationic Starch at pH 5.0, 9.5 and
                  11.2.
                          85

-------
1.6
i—
10 1 A
o 1 .4
o

LU
1 12
LU
rr
1.0
I I I I I 1
c/
-o-o o^°
\ p— - — o o-^
\ / °
V


i i i i i i
                                6        8
                                    PH
                                                    10       12
      FIGURE 48(a).  Relative Viscosity of 0.4-percent Corn  Starch  Solution As
                   a Function  of pH
   1.6
o '
o  '7
LO
LU
>
LU
   1.2
   1.0
               pH 7
             1
            10~4       10~3         10~2
                CALCIUM  CHLORIDE  CONCENTRATION, N
      FIGURE 48(b).  Relative Viscosity of 0.4-percent Corn Starch Solution As
                   a Function of Calcium Chloride Concentration
                                  86

-------
this particular point several series of flocculation tests were performed
with quartz samples and with goethite samples.  In each series the cal-
cium chloride concentration was kept constant and the pulp pH was main-
tained either at natural  near neutral pH, or at 11.  The settling rates
of quartz and goethite thus obtained as functions of the level of starch
addition were plotted in Figures 49(a) and 50(a) , respectively.  As seen
in these figures, the levels of the starch addition where the maximum set-
tling rates were obtained varied widely with the calcium chloride concen-
tration.  The best comparison of the flocculation behavior then may be made
by the maximum settling rates.

     In Figure 49 (b) the maximum settling rates for the quartz sample were
plotted against the calcium chloride concentration.  As seen in the figure,
the settling rate at pH 7 decreased gradually with increasing calcium chlo-
ride concentration as expected.  At pH 11, however, the settling rate in-
creased markedly to 10 1 N calcium chloride.  The increase up to this con-
centration may be attributed to the increased adsorption of starch accom-
panying the increased adsorption of calcium ions on the quartz.  This point
is reflected in the increased level of starch addition where the maximum
settling rates are observed.  The decrease in the maximum settling rate a-
bove a calcium chloride concentration of 10 1N must then be caused by the
change in the conformation of the starch.  The somewhat lower level of starch
addition where the maximum is observed might also suggest some change in the
charge characteristics of the starch molecules themselves.  It certainly be-
comes of interest to study the zeta potential of starch under similar con-
ditions using the moving boundary method.

     For the goethite sample the maximum settling rates both at pH 7 and 11
are seen in Figure 50(b) to decrease gradually with increasing calcium chlo-
ride concentration.  Since the goethite surface is less negative than the
quartz surface, the effect of calcium ion on the adsorption of starch is ex-
pected to be less^ and the decrease as shown must then reflect the effect of
the conformation of starch more or less directly.  The maximum settling rate
at pH 11 in the presence of 10-1N calcium chloride, however, appears to be
somewhat higher than the curve drawn through the points.  This might indi-
cate a similar effect of calcium ion as observed for the quartz sample at pH
11, but to a much less extent.  The results of streaming-potential measure-
ments on the goethite sample show that the sign of the zeta potential may be
reversed at pH 11 when the calcium chloride concentration exceeds 4 x 10    M
(Figure 40).
                                   87

-------
   15
                                                    N CaCl
u
•E,  10
UJ

<
QL

CD
UJ
CO
         A- 10"N CaCl
                                                s-2,
•*•
 I
                                _A._10 N CaCl2
                                 • _10"1N CaCl2
                                •*-1N CaCl2
                                              i
                                         i
    0
0.2     0.4     0.6     0.8     1.0
    STARCH ADDITION,  Ib/ton
                                 1.2
     FIGURE 49(a).
      Flocculation Test Results on Quartz as a Function
      of Starch Addition at Different Concentrations of
      Calcium Chloride at pH 7 (Lower Curves) and pH 11
      (Upper Curves)
 c
 E
O
LLl
X
<
   10
   0
           o
               -3
                                     10"'         1
           CALCIUM CHLORIDE  CONCENTRATION,  N

     FIGURE 49(b).  Maximum Settling Rate of Quartz as a Function of
                  Calcium Chloride Concentration
                       88

-------
.E  25
 u
    20
LU
!<  15
o:
?  10
h-
LLI   R
00   °
     0
                                 ID'2 N	A	A—

                                 10-1  N —•	Q—
                    1            2            3
                   STARCH  ADDITION, Ib/ton
      FIGURE 50(a).   Flocculation Test Results on Goethite as a Function
                    of Starch Addition at Different Concentrations  of
                    Calcium Chloride at pH 7 and 11
_c
u 25
c
£20
O 15
I-
H 10
LU IW
00
< 5
0
i i i i i i i
n
\^ pH 11

• ^^ O 0
pH 7
— —
I i 1,1
10~3 10~2 10~1 1
CaC 2. N
      FIGURE 50(b).
                   Maximum Settling Rate of  Goethite as  a Function
                   of Calcium Chloride Concentration
                           89

-------
                                SECTION VII
                                DISCUSSIONS


     A theory by Derjaguin and Landau,,  and by Verwey and Overbeek5 on the
stability of lyophobic colloids based on an interplay of double layer re-
pulsion and van der Waals'attraction is well known.   Though an empirical
formulation based on the critical zeta potential and the Schulze-Hardy
rule may be explained by the above theoretical treatment, recent investi-
gations by Matijevic and his coworkers29 emphasize the effect of hydroly-
sis of the precipitating ion of an electrolyte.

     Flocculation by natural and synthetic polymers  is explained in terms
of polymer bridging,, and the effect of electrostatic interaction between
solid particles is considered to be of secondary importance.14'30'31  Their
adsorption, however, depends on the balance between  hydrogen bonding and
electrostatic interaction.  *15  Increased molecular  size of polymeric floc-
culants and a favorable condition for their electrostatic interaction with
mineral surface promote adsorption.  In some cases,  a chemical interaction
between, for example, the phosphate group of potato  starch and the calcium
or other cations of a phosphate slime forming insoluble phosphates may be-
come predominant.32  The flocculation condition may  not be directly related
to the adsorption densities since the bridging depends on the conformation
of the polymer chain at the interface.^'9  The role  the conformation of a
polymer plays at the interface is clearly indicated  by the detrimental ef-
fect of an excessive agitation on a suspension.   Here, the adsorbed poly-
mer chain would presumably be coiled around the particles rather than
stretched out into the aqueous phase for effective bridging.3>33  The in-
fluence of electrolytes on the viscosity of hydrophilic sols, such as agar
and starch, is commonly referred to as  the electroviscous effect2^ and re-
flects their influence on the conformation of these  sols.  The action of
a starch as a flocculant when used in conjunction with a multivalent electro-
lyte may be best understood, therefore, through an understanding of the
properties of starch dispersions in aqueous media, the surface conditions
of a pure mineral in water, the adsorption behavior  of a starch and its
derivatives at the mineral-solution interface in the absence and in the pre-
sence of multivalent cations, and the overall correlation of the above re-
sults with the flocculation and clarification behavior.
Properties of Starch Dispersions

     Starch is a highly polymeric carbohydrate built up of D-(+)- glucose
units
                                    90

-------
and its molecular weight ranges from 50,000 up to several millions.  Corn
starch consists of two types of polymers:  a linear amylose composed of 200
to IjOOO pyranose rings bonded together through oxygen atoms in positions
1 and 4, and a branched amylopectin with 1,500 or more pyranose rings con-
nected with periodic branching at positions 1 and 6.  The proportions of
each are reported to be 25 percent amylose and 75 percent amylopectin. 3Lf
In chemically modified starches the substituent groups, -0-CH COO~Na+ and
                      +  _                                   ^
-O-CH^-CHOH-CH N(CH )  Cl , have been introduced in varying amounts mainly

at positions 2 and 6 to the extent indicated by the degree of substitution.
The maximum degree of substitution attainable is 3.0 including the OH-  gxoup
at position 3.

     The structure of starch molecules in aqueous solutions is complicated
by their association through hydrogen bonding.  The infrared spectra of
starch indicate that the hydroxyl groups are strongly hydrogen bonded.35
The location of the hydrogen bonding in the starch molecules depends large-
ly on the spatial configuration involving the hydroxyl groups in the pyra-
nose ring.  The hydroxyl groups at carbon atom No. 6 form perhaps the stron-
gest 0-H .... 0 bonds _, whereas the other hydroxyl groups at carbon atoms No.
2 and 3 are generally more occluded than the protruding hydroxyls at carbon
atom No. 6, especially in the linear starch molecules.  The branched chain
aggregates of starch form relatively stable starch-water complexes due  to
the availability (for spatial reasons) of some more hydroxyl groups at  car-
bon atoms No. 2 and 3.3^

     An unmodified corn starch is negatively charged due presumably to  the
presence of a small amount of fatty acid, esterified with the carbohydrate.37
lonization of the hydroxyl groups, particularly at positions 2 and 6 of the
glucose unit, is also thought to  contribute towards the development of  a
charge on the starch.3®  In the electrophoretic mobility measurements on the
granules of unmodified corn starch15 the sign of the electrical charge  re-
mained negative in the pH range from 3 to 11 indicating that the isoelectric
point was not reached, although the electrophoretic mobility decreased  at high-
er   and lower pH values due presumably to the compression of the electrical
double layer around the starch granules.  The viscosity of a starch disper-
sion, as shown in Figure 48,undergoes a minimum point near pH 3.5, which may
correspond to the isoelectric point.28  However, a cursory measurement  indi
cated that in the presence of corn starch the streaming potential of quartz
remains negative in this pH region.  Since the adsorption density of starch
is not too high, the zeta potential may be more strongly influenced by  the
negatively charged quartz surface.  It becomes of interest to study the zeta-
potential of a solubilized starch in more detail using the moving-boundary
technique.  The isoelectric point of potato starch is reported to be at 4.6.39

     The electrophoretic mobility measurements also determined that the cat-
ionically and anionically modified starches used in the present investigation
were respectively positively and negatively charged, and that there were no
changes in the sign of electrical charge in the pH range from 3 to II.15
Evidently, these functional groups remain ionized over the stated pH range.

     Viscosity measurements provide a most convenient method of determining
                                     91

-------
relative molecular sizes of starch in solutions.23  Both mechanical shearing
and prolonged heating are reported to decrease viscosity markedly, and, there-
fore, the average molecular weight, and also a starch solution left standing
in a flask lowers its viscosity over a period of a few days probably through
biological degradation.6  The standardization of the preparation procedure
becomes imperative for the starch solution in order to have consistent mo-
lecular weight throughout the test program.   For a given starch solution the
change in viscosity reflects the change in the conformation of the starch mole-
cules.  An increase in viscosity with increasing pH (Figure 48) would indicate
the ionization of the hydroxyl groups, thereby the starch structure resulting
in more stretdied-out conformation.  The presence of an eacess concentration
of calcium ion neutralizes the negative charge on the starch structure and makes
the starch molecules to coil more tightly.  The effect of calcium ion on the
viscosity of corn starch is noted to be more pronounced at pH 11 than at pH
7,  a phenomenon parelleling the action of calcium ion on quartz.


Surface Conditions of Quartz and Goethite in Water

     In developing a theory of the stability of lyophobic colloids Verwey
and Overbeek22 differentiated between the adsorption behaviors of ions in
solution  on a given solid surface and classified them into three groups,
namely, potential-determining ions, indifferent ions, and specifically-ad-
sorbed ions.  Those ions which establish equilibrium at the interface and
which determine the potential drop between the solid and the liquid phases
are termed  potential-determining ions and are responsible for the electri-
cal charge observed on the solid surface.

     For oxide minerals, it is postulated1*0  that the oxygen ions in the sur-
face undergo the following reaction:

                    OT   r   •> + H00  	«•>  20H~,     .   ,
                     (surface)    2   •*	  ^IL (solution)

It follows that hydroxyl ions (and, accordingly, hydrogen ions) are involved
in the electrolytic reaction at the interface and are thought to be the po-
tential-determining ions.  Streaming-potential measurements on a number of
oxides as a function of pH and in the presence of different concentrations
of an indifferent electrolyte substantiated this.1*1;1*2  An increase in pH
drives the reaction to the left and increases the concentration of oxygen
ions at the surface, thereby making  the surface more negative.  Conversely,
a decrease in the pH should cause the reaction to go to the right and the
surface oxygen-ion concentration decreases.   At a certain pH, the surface
becomes electrically uncharged, and is referred to as the isoelectric point.
In a solution with a pH below this value the surface becomes positively charged.
The isoelectric points of the goethite and the Brazilian quartz samples are
seen in Figures 39 and 44 to be at pH 6.7 and near pH 2, respectively.  These
values are in good agreement with those reported previously.13,1*0  Schematic
illustrations of the origin of electrical charge on silica26 and hematite1*3
based on this concept, are given in the literature.

     The Schulze-Hardy rule relates the power of active ions coagulating sols
                                   92

-------
with the valence of the ions.  In its original form the valence of active
ions was taken directly.  Matijevic29 emphasized the importance of the hy-
drolyzed species of metal ions in effecting the reversal of charge on lyo-
phobic colloids.  In line with his view streaming-potential measurements on
quartz as a function of calcium chloride concentration (Figure 45) showed
that the zeta potentials remained negative irrespective of calcium chloride
concentration at pH 4 and near 7.  At pH 9.5 and 11, however, the sign of
the zeta potential on the quartz was reversed at calcium chloride concentra-
tions of 10~2 M and 2.5 x 10 3 M, respectively, suggesting that CaOH  might
be the effective form for adsorption.  In fact, Clark and Cooke49 arrived
at the same conclusion by comparing the equilibrium concentrations of the
different species of calcium ion with the adsorption data in connection
with the anionic flotation of this mineral.  The reversal of the sign of the
zeta potentials was also noted on goethite at pH 10.7 (Figure 40).  This ob-
servation is of particular interest in the anionic silica flotation of iron
ores, since the ability of calcium ion to reverse the zeta potential of goe-
thite in the highly alkaline pH range might activate this mineral along with
quartz for possible flotation with an anionic collector.50  This clearly in-
dicates the need of a starch depressant in such a system.

     In the discussions to follow it will be shown that the adsorption of a
starch flocculant is governed by the electrical interaction between the
starch and the charged mineral-solution interface.  For the negatively charged
corn starch to be effective the mineral surface must be positively charged.
It then becomes of interest  to investigate the effects of magnesium ion which
is known to form MgOH  at a much lower pH,49 of aluminum ion which forms poly-
meric hydrous oxide ion [Alg(OH)  ]  H at near neutral pH,51 and of ferric ion
which forms hydroxlyzed iron  (IIIJ polymers.27


Adsorption Mechanism of Starches

     In the flocculation of mineral suspensions by natural and synthetic poly-
mers, not only the adsorption density of the polymers, but also the manner by
which the polymers are adsorbed at the interface becomes important.   Several
different approaches have been proposed for a theoretical treatment of poly-
mer adsorption.44*45  These theories appear to differ primarily in their de-
pictions of the method by which the polymers are secured at the interface, but
'it is generally accepted that a polymer molecule is attached with short se-
quences of segments on the surface and that the unattached segments extend in-
to the solution as loops.  The relative number and arrangement of the attached
and unattached segments define the conformation of the adsorbed molecule.

     No single theoretical treatment has been shown to cover the whole pic-
ture even with polymers of well-defined composition and molecular weight.
The distribution of the molecular weight of starches prepared in the manner
described in the present investigation was necessarily undefined and, there-
fore, precluded the theoretical analysis based on these theoretical treat-
ments.  Nevertheless, the adsorption isotherms of corn starch and a cationic
starch on quartz and hematite,1^ shown in Figures 5l(a) and (b), are seen to
follow the general shape of the Simha-Frisch-Eirich plot.45
                                   93

-------
    3.0

    2.0

C\J
 E  1.0
a:
LU
o_
ch
E
o"
LU
CO
   0.1
g  .05
   .01
                                  H; pH 7
                                  H: pH 9

                                 -H: pH11
            20
                    O: PH7
                    40
60
80
100
120
                                                        E
                                                        o:
                                                         01


                                                        S 3
                                                        u
                                                        en
                                                        CO
                                                        m 2
                                                        0
                                                           If
                                                              x
20
                                                                                  H:  pH 11
                                                                                  H:  pH 9
                                                                                  H:  pH 7
40
60
80
                               RESIDUAL  CONCENTRATION.mg. PER  LITER
100
       FIGURE 51(a).  Adsorption of Corn Starch on
                    Hematite (H) and Quartz (Q)15
                                                             FIGURE 51(b).  Adsorption of Cationic Starch on
                                                                          Hematite (H) and Quartz (Q)15

-------
     In Figure 51 it is readily apparent that corn starch is more strongly
adsorbed on hematite than it is on quartz, but the adsorption densities of
the cationic starch are higher on quartz than on hematite.  If a closer, ex-
amination of the individual isotherms is made, it will be noted that the ad-
sorption of the corn starch decreases with increasing pH, whereas that of
the cationic starch increases with increasing pH both on quartz and on hema-
tite.   Quartz is more electronegative than hematite in aqueous suspensions
in the pH range 7 to 11 since the isoelectric points of quartz and hematite
suspensions exist at a pH near 2 and 6.7, respectively, and corn starch is
negatively charged in aqueous solutions in the same pH range.   The diametri-
cally-opposite adsorption behavior of the corn starch and the cationic starch
on these two oxide minerals may be understood if an electrostatic interaction
exists between the starch and the mineral surface.  The electrostatic repul-
sive forces acting between the oxide mineral surface and the corn starch mole-
cules  serve to hinder adsorption.  It is obvious then that the adsorption is
caused by some nonionic process.

     The hydrogen bonding between hydrogen in starches (COOH,  OH, NH, etc.)
and oxygen on the oxide mineral surface is generally accepted as being re-
sponsible for their adsorption, similar to the adsorption of polyelectro-
lytes  postulated by Michaels.^  Evidence of hydrogen bonding in the adsorp-
tion of polyacrylamide on silica was found by infrared spectres copy.47  Al-
though the hydrogen bond is relatively weak, since thousands of bonds are
formed per polymer molecule, the total bonding energy per polymer molecule
becomes very large.31  The adsorption of starches has been schematically
viewed, as shown in Figure 525by a model incorporating the electrostatic
and hydrogen-bonding phenomena. 5  The adsorption by hydrogen bonding is
considered to be the same for cationic, unmodified corn, and anionic star-
ches,  assuming that the molecular weights are the same for all three star-
ches.   Because of the combined effect of electrostatic attraction and the
hydrogen-bonding forces, the adsorption of the cationic starch by quartz
and hematite is quite high.  The adsorption of unmodified corn starch, be-
ing negatively charged, and in particular, of an anionic starch will be low-
ered because of electrostatic repulsion.  The so-called "squeezing-out" ef-
fect,  postulated for the adsorption of relatively large-sized anions, such
as NO ~ and C10.~, at a mercury-solution interface,1*8 may also have to be
taken into consideration for the adsorption of large starch molecules at a
solid-solution interface.

     A strong affinity of corn starch toward hematite is indicated in Fig-
ure 51 in the steep rise in the adsorption density at low concentration; the
isotherms then reach a saturation coverage corresponding to a particular pH.
The adsorption isotherm on hematite and the zeta-potential curve on goethite
in the presence of corn starch  (Figure 38) show strikingly similar trends.
The constant values of zeta potentials in the high concentration range of
starch may reflect the saturation coverage observed in the adsorption iso-
therms.  The irreversibility of the adsorption of corn starch, shown by the
hysteresis of zeta potentials as a function of pH on goethite (Figure 39)
and possiby on quartz at low pH (Figure 44), affords additional evidence that
corn starch is strongly adsorbed on uncharged goethite via hydrogen bonding.
                                   95

-------
        \///\ HYDROGEN  BONDING



        KV\1 ELECTROSTATIC  INTERACTION
      ANIONIC
 I
 a
   o
   Co
 (
l> >V|
IxV >i
I >v X '
           UNMODIFIED
CATIONIC
                                           I
                                           a ^


                                           O O
                                           ? H
                                           LO U

                                           tu tr
                                           cc i-
                                           O b
            OXIDE  MINERAL  SURFACE
FIGURE 52   A Model  Showing Mode of Adsorption
            by  Various  Starches on Oxide  Mineral
            Surfaces.i5
                     96

-------
     In the present discussion the adsorption behaviors of starches on hema-
tite and goethite are mentioned, more or less, interchangeably.   Since the
isoelectric points of both hematite and goethite are located at  the same pH
of 6.7, and since their flotation behaviors have been shown to be nearly
identical,52 the surface properties of these two minerals may be regarded as
behaving in a similar manner.  Goethite was chosen for the streaming-poten-
tial measurements because difficulty was encountered during the  experiment
with hematite in the determination of the zeta potential due presumably to
the surface conductance.53  Measurements of the surface area of  goethite by
the krypton adsorption method, however, became somewhat superfluous for ex-
pressing the adsorption density of giant polymer molecules, such as starch
and polyacrylamide, due to the extremely porous structure of goethite.  In
fact, it was reported that the surface area of goethite determined by the
nitrogen adsorption method was virtually independent of the fineness of the
grind.54

     The number of hydrogen-bonding groups in a starch molecule  and possibly
the squeezing-out effect of the starch molecule will be dependent on the
molecular size.  The adsorption density would then be expected to decrease
with the molecular size.  The homogenization of a starch dispersion and the
pyrolytic degradation of a corn starch into a British gum reduce the average
molecular size, which may be attested by the viscosity measurements.  As a
result, the abstraction of homogenized corn starch and of British gum by
quartz and hematite was reported to be appreciably lower.6,15

     The presence of calcium ions in solution enhances the adsorption of corn
starch both on quartz and on hematite appreciably.6  This phenomenon is shown
in the response contours of starch abstraction by quartz in Figure 10 (c).  Ac-
cording to the present interpretation based on the electrical interaction be-
tween the charged mineral surface and the starch, the effect of  calcium ions
is obviously to lower the zeta potential of negatively charged minerals, there-
by the electrostatic repulsion, as illustrated schematically in  Figure 52, is
reduced.  When the zeta potentials of quartz as functions of calcium chloride
concentration in the absence (Figure 45) and in the presence of  starch (Fig-
ure 46) are compared, the most pronounced effect of starch is seen where the
sign of the zeta potential of quartz reverses at high concentration of cal-
cium chloride and at high pH.  This was the condition where the  most effec-
tive flocculation of quartz was shown to occur in Figure 10.  Presumably the
reversal of the sign of the surface charge due to adsorbed calcium ion pro-
moted the adsorption of the starch for flocculation.

     In addition, the interaction of calcium ions with starch molecules in
solution should not be overlooked.  With an increasing calcium chloride con-
centration, the starch molecules become more tightly coiled, as  evidenced by
the decrease in viscosity (Figure 48).  It follows, therefore, that the pre-
sence of calcium ions in solution influences the adsorption behavior of
starch not only by increasing its adsorption density but also by altering the
conformation of the starch molecules at the interface.  The interaction of
starch molecules with calcium ions either in solution or at the  surface of a
mineral may also involve some chemical bond formation with a minor quantity
of fatty acid inherently present in the starch structure.38
                                   97

-------
Flocculation Mechanism with Starches

     In the flocculation of mineral suspensions with a polymeric reagent,
natural or synthetic, two separate processes may be identified, namely, the
adsorption of the polymeric flocculant and the formation of floes.   The ad-
sorption of a sufficient quantity of the flocculant appears to be a neces-
sary condition, but it does not always lead to an effective flocculation of
mineral suspensions.  A unique geometrical arrangement, or conformation, of
the polymeric molecules at the mineral surfaces is needed for successful
bridging in the formation of large and stable floes.  Almost invariably an
optimum dosage level exists for a given polymeric flocculant, and an excess
of the flocculant tends to redisperse the floes.  A number of examples may
be seen in the experimental results with both starches and polyacrylamides
in Figures 11, 12, 15, 19, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34,  and 35.
This is attributed to an increased hydration of the mineral surface due to
an excessive adsorption of a hydrophilic colloid.  The phenomenon is re-
ferred to as protection.28

     The optimum level and the effectiveness of floe formation depend on a
number of factors:  the type, the mesh-of-grind, and the concentration of
the minerals; the type , the functional group, and the molecular weight of
the polymeric flocculants; the pH of the suspension; the type of and con-
centration of the electrolytes in solution; and the temperature of the sus-
pension.  For a suspension of a mineral, in particular a mixture of minerals,
the proper  selection and use of a flocculant would, therefore, be an ex-
ceedingly complex subject.

     In Figures 11, 14, and 15, it is apparent that the maximum settling rate
and clarity of the supernatant solution are obtained both on quartz and goe-
thite suspensions at apoint where an excess of starch begins to appear in the
supernatant solution.  The adsorption isotherms given in Figure 18 indicate
that the maximum settling rate is observed when near saturation coverage for
this particular experimental condition is attained.  The critical levels of
starch for quartz and goethite, however^ differ nearly ten-fold.  The impor-
tance of the adsorption density in flocculation is indicated by an observa-
tion that the maximum settling rate for goethite is three times greater than
that for quartz.   In the absence of calcium ions corn starch does not floc-
culate quartz due evidently to an insufficient adsorption.  Here, both quartz
and corn starch are negatively charged.  In the presence of calcium ions the
settling rate of a quartz suspension improves as the pH is raised (Figure 12)
A strong interaction of calcium ions with quartz in highly alkaline solutions
was indicated in the adsorption studies49 and in the streaming potential mea-
surements (Figure 45).  Though the electrical charge on the mineral surfaces
may be made more favorable for the adsorption of the negatively charged
starch molecules, the effect of calcium ions on the conformation of the
starch molecules  (Figure 48) should not be overlooked.  In addition, the
electrical charge on the starch molecules may also be affected.  The decrease
in the maximum settling rate of quartz and of goethite with an increase in
the calcium chloride concentration (Figures 49 and 50) affords evidence that
the conformation at the interface is important.  The increase in the maximum
settling rate up to 10 *N at pH 11 must presumably be due to an increased
                                   98

-------
adsorption  of  starch resulting from the increased adsorption of calcium ion.
Obviously}  further experimental evidence is needed to confirm this particu-
lar point.

     The mechanism of flocculation is often studied by examining the set-
tling rate, the clarification, the sediment volume, the refiltration rate,
and the thickness of the compacted bed, but these parameters may not be in-
terrelated directly.  Polymers produce larger and tougher floes than do sim-
ple electrolytes; a close parallelism between the settling rates and the fil-
tration rates  (Figures 10 and 16 for quartz, and Figures 14 and 17 for goe-
thite) reflects such a flocculation behavior.  In fact, Smellie and LaMer55
attempted to describe the flocculation behavior quantitatively from the re-
filtration rates, although their approach has recently been discredited.56
Quite frequently, the settling rate and the clarity of supernatant solutions
appear to be governed by different parameters.9>24  Polymers do not replace
the metal salt necessary for the charge neutralization,57 and, therefore, are
not effective in removing the fine turbidity of suspended solids.   Linke and
Booth3 reported that the fine suspended solids in the supernatant solution
carried a disproportionately high dosage of flocculant, and that this result-
ed in the irreversible stabilization of these particles by protective action.24

     Since the electrostatic interaction between starch and the mineral sur-
face appears to govern the adsorption of the starch and, hence, the floccu-
lation, the cationic modification of the starch structure is an obvious al-
ternate on highly-negative quartz surfaces(Figure 51).  Anionic modification
of the starch structure was found to be totally ineffective, or, at best,
appreciably less effective for quartz, irrespective of the level of the starch
or of the calcium chloride addition (Figures 28 and 29).  In addition,  the
introduction of charged sites in a polymer structure uncoils the chain  (in-
tramolecular repulsion) and makes the molecule stretch out, thereby facili-
tating bridging.  The results of the flocculation tests on quartz (Figures
25 and 26) tend to substantiate this interpretation.  The maximum settling
rates observed on quartz suspensions are appreciably higher than those  with
corn starch due evidently to the increased adsorption.  As a result, the
starch levels, where the maximum settling rates are observed, are appreciably
higher also.

     A closer examination of the settling rate data shows that the cationic
starch with higher degree of substitution is less effective as a flocculant.
If it is assumed that the two cationic starches have approximately the  same
molecular weight, too high a degree of substitution with a cationic group
may tend to make the adsorbed starch lie down more closely to the negatively
charged quartz surface and the conformation of the adsorbed starch would then
be less favorable for bridging.  The fact that the level of addition of the
0.067 D.S. starch, where the settling rate is maximum, is notably lower than
the level of 0.023 D.S. starch supports this interpretation.  After a satura-
tion coverage has been reached, the chain interaction between the starch-
coated quartz particles might also be hindered due to the electrical repul-
sion of the cationic groups (intermolecular repulsion).   The fact that  the
maximum settling rate of a goethite suspension when treated with a cationic
starch  (Figure 27) is somewhat lower than when treated with corn starch (Fig-
ure 15) may possibly be accounted for by the unfavorable chain interaction in
bridging the particles.
                                   99

-------
      The three  limiting  cases  of  the  conformation  of  adsorbed starches  thus
 tar discussed may be  schematically represented  as  shown in Figure 53.
        coiled
     (less chance
     for bridging)
         Figure  53,
         stretched out
     (ideal  for bridging)
 stretched out, but
too strongly adsorbed
  (less chance for
      bridging)
Schematic Representation of the Three Limiting Cases
of Conformation of Adsorbed Starch  Molecules
 In summary,,  an  ideal  condition for an efficient flocculation of a quartz sus-
 pension  entails sufficient adsorption of a starch either through the combined
 use of such  modifying agents as calcium ion, or through a cationic modifica-
 tion of  the  starch structure, and ,a stretched-out conformation at the interface
 for effective bridging.  The zeta potential of the suspended matter in the
 supernatant  solution  must also be as close to zero as possible in order to
 effect the clarification of the supernatant solution.

      Thus, virtually  clear supernatant solutions were obtained when goethite
 suspensions were treated with an optimum dose of corn starch at near neutral
 pH (Figures  14 and 15).  The supernatant solutions of quartz suspension, how-
 ever, remained turbid, containing in excess of 500 ppm of solids in suspen-
 sion  irrespective of  the pH or the amounts of calcium chloride and corn starch
 added (Figures 10, 11, 12, and 13).  In addition to unfavorable zeta poten-
 tials, quartz suspensions aged in an alkaline solution are known to become
 unusually stable.   In fact, they are reported to become non-flocculable with
 nomonic polyacrylamide.24  According to Matijevic,25 colloidal silica may
 behave as a  lyophilic colloid and its destabilization by electrolyte proceeds
 by  a mechanism different from what would be expected for lyophobic colloids.
 One way to counter the lyophilic nature of the quartz surface may be to change
 it  into a hydrophobic surface with a surfactant.   It is well known in ore
 flotation that a quartz surface can be made hydrophobic with a long-chain
 alkylammonium salt, thereby rendering the quartz particle floatable upon in-
 troduction of air  bubbles.   At this point the sign of the zeta potential on
 quartz is reversed.12-26   These two points, namely, the hydrophobic surface
 and the reversal of the sign of the zeta potential, become of particular in-
 terest in inducing the flocculation condition for quartz suspensions, although
 it is also known   that dodecylammonium ion may interact with a starch mole-
 cule to modify its conformation at the quartz surface.   The results of a few
 cursory flocculation tests, presented in Figure 30,showed some promise in
 clarifying the supernatant  solution.   Another effective approach in clarify-
ing the supernatant solution of a quartz suspension is  through the use of a
cationic starch (Figure 26) .   The  role of the cationic  starch in reducing the
                                  100

-------
zeta potential of quartz has already been discussed.  It is rather puzzling,
however, that the cationic polyacrylamide NC 1733 was virtually ineffective
in this respect  (Figure 31).  This may presumably be attributed to a locally
excessive concentration of the polymer due to its high adsorb ability, partic-
ularly on small particles.  From a series of tests using a number of cationic
polyethyleneimine polymers with different molecular weights, Dixon et al59
also reached a conclusion that polymers of too high a molecular weight are
undesirable.
Physical Factors Affecting Floe Formation

     The surface conditions of solids including the adsorbed starch molecules
discussed above provide a necessary condition for flocculation; the thermal
and mechanical motions of these particles bring them together.   The presence
of coarse particles in a suspension increases the probability of collision on
their descent and also gives weight to each floe for faster settling.8  The
effect of size distribution on flocculation behavior has not been reported in
the literature, although occasional mention is made on the flocculation of
sized fractions.

     The response contours of settling rates for goethite suspensions (Figure
20) and for quartz suspensions (Figure 21) as functions of size moduli and
concentrations of solids indicate that, although the settling rates for these
two suspensions at given concentrations of solids are seen to increase with
increasing size moduli, as anticipated, the settling rates are more strongly
dependent on the concentration of solids.  No simple relationship may be
deduced between the settling rate and the concentration of solids since the
settling rate of the goethite suspension decreased monotonically with the con-
centration of solids whereas that of the quartz suspension increased initially
and then decreased.  Of course, these observations are at one specific condi-
tion only, namely, at pH 9.5 and in the presence of 2 Ib of calcium chloride
per ton, and the flocculation behavior may change if either the pH or the
level of calcium chloride are changed.  The difference between the goethite
and quartz suspensions appears to lie in their affinity towards a starch flo-
culant.  In a dilute goethite suspension, having high affinity towards starch,
only a portion of the suspended particles acquire the flocculant in a relatively
high dose, thereby resulting in a high settling rate.  The irreversibility of
the starch adsorption discussed in connection with the streaming-potential
measurements (Figure 39) and the large amount of suspended solids remaining in
the supernatant solution appear to support such an interpretation.  At higher
concentrations of solids the starch may tend to be distributed more evenly and
also the mutual interference of the resulting floes tend to retard the set-
tling rate.  The crowding floes would also have a scouring effect on the sus-
pended matter, particularly when the zeta potential is within that of the
critical value.

     In the case of quartz the affinity towards starch is not too high, and
the distribution of the flocculant throughout a quartz suspension may tend
to be more uniform than in a goethite suspension.  As a result, the settling
rate of a quartz suspension increases with increasing concentration of solids
                                   101

-------
due to increased rate of collision of the particles.   Eventually,  the mutual
interference of the floes discussed above becomes  more significant.   It is
readily apparent, therefore, that very dilute suspensions  of either  mineral
are very difficult to clarify.

     In the coagulation* of mineral suspensions with inorganic electrolytes
it is often noted that the clarity of supernatant  solutions may be improved
with increasing temperature.  Increased rate of collision  of the particles
and increased thermal energy acquired by each particle in  surmounting the
energy barrier of the interacting double layers1 might account for this.   In
the flocculation of quartz and goethite suspensions,  however, the  settling
rate corrected for the viscosity did not increase  with the temperature (Fig-
ures 22 and 23).  The size and the physical properties of  floes must have
been affected in a complex manner with the temperature.  With an increase in
temperature, it is expected that the adsorption, being exothermic, will be
adversely affected, and that the conformation of the  starch would  also be
altered due to the thermal motion of the water molecules.
Flocculation Behavior of a Mixture of Minerals
     The addition of corn starch to suspensions of certain iron ores resulted
in two relatively distinct layers of sediments when the suspensions  were per-
mitted to stand.  Presumably, the layers were due to the preferential adsorp-
tion of the starch on the iron oxide minerals causing them to flocculate and
settle more rapidly than the siliceous gangue. In fact, it has been  reported
in the literature that most starches flocculate aqueous suspensions  of hema-
tite, but do not flocculate similar suspensions of quartz.6°  This selective
flocculation has been applied in practice to the differential desliming of
pulps during the anionic silica flotation of iron ores, resulting in superior
metallurgy and lower reagent cost.61  Recent attempts to extend selective
flocculation to other systems, involving such minerals as quartz, calcite,
galena, pyrite,and sphalerite, have also been reported.62>63  This process is
considered to be one of the new and most promising approaches to the treat-
ment of minerals that are so finely divided that they are usually either
discarded or are tolerated in their interference with conventional processes.63

     For the successful removal of unwanted siliceous slimes from an oxidized
iron ore by selective flocculation, the optimum conditions are extremely sen-
sitive to the mineralogical characteristics of the ore and the mesh-of-grind.
It has also been noted that tapioca flour gave better results than corn starch.9*61
To apply this technique to a variety of refractory iron ores or to cope with
the fluctuations in the ore being fed to a plant, the factors governing the
selective flocculation must be formulated.  The flocculation behaviors of the
minerals in the ore, such as are described in the present report, provide the
*    In accordance with LaMer's usage, the term "coagulation" is used here
for destabilization due to increased ionic strength or to decreased surface
potential of particles with inorganic electrolytes, and "flocculation" for
a case in which a bridging agent is involved.
                                   102

-------
basic information as to the type and the amount of the flocculant, the pH,
and the possible effect of impurities either accidentally present or inten-
tionally added.  For example, polymers with too high a molecular weight are
inherently unsuitable for selective flocculation, and the presence of cal-
cium ions released by the ore may activiate quartz for flocculation with
starch.  (Although in the Bureau of Mines process calcium chloride is in-
tentionally added.61) Cationic starches and polyelectrolytes would be totally
ineffective for selective flocculation.

     The cursory test results on an artificial mixture of goethite and quartz
presented in Figure 37 identify a few important parameters, such as the pH and
the level of the starch addition, which may be directly correlated with the
individual flocculation behaviors and the surface properties given in this
report.  In a similar manner the most desirable type and molecular weight of
the polymer to be used, the effect of the presence of calcium and magnesium
ions released by the ore, and the mesh-of-grind may be formulated.  Such a
process, however, will certainly result in a well-dispersed siliceous slime
which must be disposed as tailing and the clarified water returned to the
circuit for reuse.6
                                   103

-------
                              SECTION VIII


                            ACKNOWLEDGEMENTS
     The support of this project by the Environmental  Protection Agency and
the assistance and suggestions provided by Dr.  S.  A.  Hannah, the Grant Project
Officer, are gratefully acknowledged.

     Thanks are also due to the Research Department of the Corn Products
Company, Argo, Illinois, particularly  to Drs.  H.  J. Roberts and S.  A.  Parmerter,
for their help in sponsoring the earlier investigations which laid  the founda-
tion for the present project and for providing  the supply of various starch
products.

     This work was performed by Iwao Iwasaki  and  Rodney J. Li pp.
                                   104

-------
                             SECTION IX


                             REFERENCES


 1.  E. J. W. Verwey and J. Th. G. Overbeek:  '"Theory of the Stability of
     Lyophobic Colloidsj" Elsevier, New York  (1948).

 2.  G. R. Gardner and K. B. Ray:  Trans AIME 134  146  (1939).

 3.  W. F. Linke and R. B. Booth:  Trans AIME 217  364  (1959).

 4.  S. K. Kuzkin and V. P. Nebera:  "Synthetic Flocculants in Dewatering
     Processes,," National Lending Library for Science and Technology,
     Boston Spa, England (1966).

 5.  F. Spaldon and K. Tkacova:  Mining and Metallurgy Quarterly, No. 4,
     p 45 (1966).

 6.  I. Iwasaki and R. W. Lai:  Trans AIME 232  364  (1965).

 7.  A. S. Teot and S. L. Daniels:  Env Sci $ Tech 3_  825   (1969).

 8.  A. M. Gaudin:  "Principles of Mineral Dressing^" McGraw-Hill, New
     York  p 326 (1939).

 9.  I. Iwasaki, W. J. Carlson, Jr., and S. M. Parmerter:   Trans AIME
     244  88  (1969).

10.  R. J. Charles:  Trans AIME 208_  80  (1957).

11.  A. M. Gaudin and D. W. Fuerstenau:  Trans AIME 202_  66  (1955).

12.  H. C. Li and P- L. de Bruyn:  Surface Science 5_  203    (1966).

13.  I. Iwasaki, S. R. B. Cooke and A.  F.  Colombo:  USBM RI 5593 (1960).

14.  I. Iwasaki:  Trans AIME 232_  383  (1965).

15.  S. R. Balajee and I. Iwasaki:  Trans  AIME 24£  401  (1969).

16.  S. M. Parmerter, Corn Products Co., Argo, Illinois, private communications

17.  Dow Chemical Company:  "Laboratory Evaluation of Dow  Flocculants."

18.  M. Dubois, K. A. Gilles, J. K. Hamilton, P. A.  Rebers and F. Smith:
     Anal Chem 28_  350  (1956).

19.  E. B. Sandell:  "Colorimetric Determination of Traces of Metals"
     Interscience, New York (1959).
                                 105

-------
20.   I.  Iwasaki and P.  L.  de Bruyn:  J Surface Science 3_  299 (1965).

21.   D.  W.  Fuerstenau:   Min Eng 8  834  (1956).

22.   J.  Th.  G.  Overbeek:  "Colloid Science"3 Vol. I (H. R. Kruyt, editor),
     Elsevier,  New York (1952)  p 207.

23.   R.  L.  Whistler (editor):  "Methods in Carbohydrate Chemistry", Vol. VI
     "Starch",  Academic Press,  New York (1964).

24.   J.  A.  Kitchener:   Filtration and Separation, Sept/Oct (1968) p 558.

25.   L.  H.  Allen and E. Matijevic:  J Coll and Interface Sci 5l_  287.
     (1969).

26.   A.  M.  Gaudin and D. W. Euerstenau:  Trans AIME 202  958 (1955).

27.   C.  R.  O'Melia and W.  Stumm:  J Coll and Interface Sci Z3  437   (1967).

28.   H.  B.  Weiser:  "Colloid Chemistry"., Wiley, New York  (1949)  p 298, 305.

29.   E.  Matijevic:  "Charge Reversal of Lyophobic Colloids", in "Principles
     and Applications  of Water Chemistry" (S. D. Faust and J. V.  Hunter,
     editors),  Wiley,  New York (1967) p 328.

30.   A.  P.  Black:  JAWWA 52_  492  (1960).

31.   Warren Spring Laboratory:   "Flocculation of Suspensions of Solids with
     Organic Polymers—A Literature Survey", Mineral Processing Information
     Note No. 5  (1965).

32.   V.  K.  LaMer:  "Coagulation Versus the Flocculation of Colloidal
     Dispersions by High Polymers", in "Principles and Applications of
     Water Chemistry"  (S.  D. Faust and J.  V. Hunter, editors), Wiley,
     New York (1967) p 246.

33.   M.  F.  McCarty and R.  S. Olson:  Trans AIME 214  61   (1959).

34.   H.  J.  Roberts:  Corn Products Co., Argo., Illinois, private communication

35.   M.  Samec:   J Poly Sci 23_  801   (1957).

36.   R.  W.  Kerr:  "Chemistry and Industry of Starch", Academic Press
     New York (1950).

37.   J.  A.  Radley:  "Starch and Its Derivatives", Chapman and Hall,  London
     (1968)  p 20.

38.   P.  Somasundaran:   J Coll and Interface Sci 3^  557   (1969).

39.   R.  D.  Snouffer:  Discussion in Trans AIME 134  166   (1939).
                                  106

-------
40.  W. D. Harkins:   "Physical  Chemistry  of Surface  Films",  Reinhold,
     New York  (1952).

41.  F. F. Apian and  D. W.  Fuerstenau:  "Principles  of Nonmetallic
     Mineral Flotation" in  "Froth Flotation",,  50th Anniversary Volume,
     AIME, New York  (1962).

42.  G. H. Parks:  Chem Rev 65_  177   (1965).

43.  G. A. Parks and  P. L.  de  Bruyn:  J Phys  Chem 66_  967   (1962).

44.  R. R. Stromberg:  "Adsorption of Polymers"  in "Treatise on Adhesion
     and Adhesives"  (R. L.  Patrick, editor),  Dekker, New York  (1967).

45.  R. Simha, H. L.  Frisch and F. R. Eirich:  J Phys Chem  57_   584   (19S3).

46.  A. S. Michaels and 0.  Morelos:   Ind  Eng  Chem 47_, No. 9, 1801 (1955).

47.  0. Griot and J.  A. Kitchener:  Trans  Far Soc 61_  1026   (1965).

48.  D. C. Grahame:   J Am Chem  Soc 79_ 3006  (1957).

49.  S. W. Clark and  S. R.  B.  Cooke:  Trans AIME 241  334   (1968).

50.  I. Iwasaki, S. R. B. Cook  and H. S.  Choi:   Trans AIME  217 237   (1960)

51.  E. Matijevic, K. G. Mathai, R. H. Ottewill  and  M.  Kerker:  J Phys
     Chem 65_  826  (1961) .

52.  I. Iwasaki, S. R. B. Cook  and Y. S.  Kim:  Trans  AIME 223   113   (1962).

53.  S. R. Balajee:   MS Thesis, University of Minnesota (1968).

54.  Y. Takahashi, I. Iwasaki  and H.  Kahata:  J  Iron and Steel Inst  of
     Japan 49_  1262   (1963) .

55.  R. H. Smellie, Jr., and V.  K. LaMer:   J  Coll Sci 13_ 589   (1958).

56.  R. W. Slater and J. A.  Kitchener:  Disc  Far Soc 42_ 267   (1966).

57.  C. L. Fitzgerald, M. M. Clemens  and  P- B. Reilly,  Jr.:  Chem Eng
     Prog 66_  36  (1970) .

58.  S. R. Balajee and I. Iwasaki:  Trans  AIME 244   407  (1969).

59.  J. K. Dixon, V.  K. LaMer,  C. Li, S.  Messinger and H. B. Linford:
     J Coll and Interface Sci  23_ 465   (1967) .

60.  S. R. B.  Cooke,  N. F.  Schulz and E.  W. Lindroos:  Trans AIME 193
     697  (1952).

61.  D. W. Frommer:   Mining Eng 16_  67   (1964).
                                  107

-------
62.   L.  Usoni, A.  M.  Marabini and G.  Ghigi:   Proceedings of VIII
     International Mineral Processing Congress, Leningrad (1968), Paper
     D-13.

63.   B.  Yarar and J.  A.  Kitchener:  Trans IMM  76_  C23  (1970).

64.   D.  W. Frommer:  "Aspects of Water Reuse in Experimental Flotation of
     Nonmagnetic Taconite:, AIME preprint 70-B-13 (1970).
                                 108

-------
1
Accession Number
w
5
2

Subject Field & Group
01B, 05D
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
       Minnesota Univ., Minneapolis, Mineral Resources Research Ctr
     Title
         FLOCCULATION AND CLARIFICATION OF MINERAL SUSPENSIONS,
10

Authors)
Iwasaki, Iwao
Lipp, Rodney J.
16

21
Project Designation
14010 DRB
Note
 22
 Citation
   Water  Pollution  Control Research Series
   53 fig.,  9  tab,  64 ref.
                                                 ORD-14010DRB10/70, 108 p., May 1971
 23
 Descriptors (Starred First)
 *Waste water  treatment,  *Mineral industry, *Flocculation, *Suspension, *Quartz,
 Settling velocity,  Filtration,  Particle size, Zeta-potential, Viscosity,
 Adsorption, Regression analysis, Colorimetry, Calcium chloride,
 25
 Identifiers (Starred First)
 *Goethite,  *Starch,  Polyacrylamide,  Streaming potential,  Selective flocculation
 27
Abstract
A study of the  flocculation and clarification of quartz and goethite suspensions using
starch and calcium  chloride as  flocculants was made by determining the settling rates
of the suspensions,  the  amount  of suspended solids in the supernatant liquid,  and
the concentrations  of the  residual starch and calcium ion.   These determinations,
supplemented with measurements  of streaming potential, adsorption density,  and visco-
sity, were used to  formulate a  comprehensive expression of the mechanism involved in
the use of starch flocculants.   No particular difficulties  were experienced either
in the flocculation or in  the clarification of goethite suspensions.  Quartz sus-
pensions were difficult  to clarify.   A cationic starch of a certain molecular size
and degree of substitution or a combination of a polymeric  compound and a cation
capable of reversing the charge of the quartz showed some promise in the clarifi-
cation of quartz suspensions.   The most effective flocculation and clarification
condition may be observed  when  a polymeric compound is adsorbed uniformly in
sufficient quantity and  in a stretched-out conformation, and when the zeta-potential
of the resulting suspension is  made close to zero.  This report was submitted in
fulfillment of  Grant No. 14010  DRB from the Federal Water Quality Administration
to the University of Minnesota.
Abstractor
                              Institution
                                       University of Minnesota
 WB.102 (REV. JULY 1969)
 WRSIC
                        SEND WITH COPY OF DOCUMENT, TO; WATER RESOURCES SCIENTIFIC INFORMATION CENTER
                           '                      U.S. DEPARTMENT OF THE INTERIOR
                                                  WASHINGTON, D. C. 20240
                                                                              * GPO: 1970-389-930

-------