106 Proceedings of The South African Sugar Technologist!

;'Association - June 1978

SOME FACTORS AFFECTING THE PER'FORMANCE O F

POLYACRYLAMIDE FLOCCULANTS IN JUICE CLARIFliCATION
By G . S. SHEPHARD
Sugar Milling Research Institute, University of Natal, Durban.
Abstract mined by the extrapolation of a plot of mud volume against
The importance of various factors relating to the preparation reciprocal time8.
of polyacrylamide flocculant solutions was investigated. For op- The minimum time required for flocculant solution prepara-,
timum results, these solutions should be agitated for at least 2 tioti was determined on Superfloc A 130, dissolved in distilled
hours before use while water of high ionic content should be water at 0,05% and 0,5% concentration and agitated for periods
avoided. The quantity of calcium phosphate precipitated during between 15 min. and 6 hours before application. The flocculant
the clarification reaction and the calcium concentration in clear dosage was 0,8 pprn on juice. All flocculant solutions were
juice are shown to influence differently the behaviour of prepared by sprinkling the granular solid into the vortex of a
polyacrylamide flocculants of varying degrees of hydrolysis. stirrer. The effect of flocculant solution concentration on its ef-
ficiency in clarification was determined with Superfloc A130
Introduction (0,8 pprn on juice). Solutions were prepared in distilled water at
The lime-defecation process for the clarification of raw sugar various concentrations between 0,01% and 0,5%. The influence
cane juice relies on the in situ precipitation of calcium of dilution was investigated by diluting the higher concentrations
phosphate. This is achieved by the addition of lime (Ca (0H)z) immediately prior to application.
to raise the pH and thus precipitate the natural calcium and The influence of the nature of the water used as solvent was
phosphate content of the juice. This process flocculates the par- studied with regard to pH and ionic concentration. The effect of
ticles in the raw.juice by binding them into the precipitate via the the former was determined by adjusting the pH of flocculant
Ca2+ions adsorbed at the particle-liquid interface'. solutions with either 0,lM HC 1 or 0,lM KOH. The flocculants
The quality of the clear juice obtained by this method is thus investigated were Superfloc A1 10 (approx. 30% DH), Superfloc
dependent on the extent to which this flocculation process can A 150 (Approx. 60% DH) and Talosep A3 (approx. 40% DH).
scavenge the particles of the taw suspension. It has long been Flocculants were prepared in distilled water as 0,05% solutions
known 3q 51that the extent of defecation is dependent on the and applied at 3,3 pprn on juice. The effect of solvent ionic
amount of phosphate present in the original juice. Inorganic strength was studied by dissolving Talosep A3 (0,05% con-
phosphate is frequently added to juices deficient in this require- centration a plied at 1,7 p m on juice) in water whose ionic
ment in order to obtain clearer juice, faster mud settling and bet-
ter mud filtrationg. Since the adjustment of juice pH with lime
f
strength ha been adjustel with NaCI. Specific conductance
determinations were used as a measure of ionic concentration.
results in considerable addition of Ca *+ ions, the precipitation of Samples of distilled water, factory condensate and factory tap
the phosphate content of mixed juice is generally nearly com- water were also investigated. Juice for the above tests was ob-
plete. It has, however, been shown that addition of Ca *+ can tained from IL.
improve the clarity of clear juice ) * I 3 . The effect of the amount of phosphate precipitated was
In order to improve the clarification system, polyacrylamide studied at constant calcium concentration in clear juice (320
flocculants are added prior to the settling stage. These floc- pprn). The PzOscontent of a JR mixed juice sample was adjusted
culants are copolymers of acrylamide and sodium acrylate with H3P04 and then the juice was clarified with Ca (0H)z using
(generally known as partially hydrolysed polyacrylamides) and a liming pH of 7,5 (at 70C). The clarification behaviour of the
possess molecular weights in excess of one million. The percen- limed suspension was analysed without flocculant, with
tage of sodium acrylate groups is referred to as the degree of Superfloc A1 10, with Superfloc A130 (approx. 45O/0 DH) and
hydrolysis (DH). For their optimum use, care must be exercised with Superfloc A150. These flocculants !were prepared as 0,05%
in the preparation of the solution. The generally recognized prin- solutions in distilled water and dosed at 2 pprn on juice. The role
ciples of polymer solution preparation are a careful addition of played by the calcium concentration in clear juice was in-
the solid powder to the water so as to independently disperse the vestigated at a PzOs concentration of 320 pprn in mixed juice.
individual granules; gentle agitation of the solution for a suf- The calcium concentration in mixed juice was adjusted with
ficient period of time to ensure complete dissolution and calcium chloride and the pH raised to 7,5 with 2M KOH. Mixed
equilibration; temperatures below 50C; a two stage prepara- juice samples were obtained from IL and TS.
tion in which a stock solution is prepared and diluted prior to
use; and the avoidance of centrifugal pumps which can damage Results
flocculant molecules4. The relationship between initial settling rate and the time for
The purpose of this investigation was to determine what which the solution is agitated prior to use is shown in Figure 1
magnitude of influence polymer solution preparation parameters for Superfloc A130. It is clear that for this polymer an
have on clarification efficiency. The parameters considered were equilibrium in solution is attained after 2 hours and that further
time of agitation of the stock solution, solvent ionic strength, agitation does not bring any improvement in the settling rate.
flocculant solution concentration and solution pH. The second This equilibration time is unaffected by polymer concentration.
stage of this study concerned the influence of the amount of PzOs Although settling rates are affected by the period of agitation, a
precipitate and the concentration of calcium in clear juice on the steady turbidity value and final mud volume had been attained
behaviour of polyacrylamide flocculants. even after only 15 minutes. Due to their high viscosity, floc-
culant solutions are generally prepared at concentrations of
Experimental 0,5% or below. Low concentrations facilitate dosing and
The laboratory clarification unit and its operation have produce more efficient mixing with the limed juice. Figure 2 il-
already been describedlO. Turbidities of clear juice were deter- lustrates the effect that varying concentrations have on settling
mined as the absorbance at 800nm in a lcm cell using a Zeiss rates and shows that below a concentration of 0,05% a rapid
PM4 spectrophotometer. Initial settling rates of the limed decrease in flocculant efficiency occurs, whereas above this con-
suspension Were found graphically from the initial linear section centration a constant settling rate was observed. The variations
of the measured settling curves. Final mud volumes were deter- in flocculant solution concentration appear to have no effect on
 Proceedings of The South African Sugar Technologists' Association - June 1978 107

turbidity or final mud volume. below the optimum show very little change in activity with pH.
The conformation of the flocculant polymer in solution and Figure 4 shows the influence that dissolved ionic species in the
thus its efficiency in promoting clarificat~onis dependent on
the pH and ionic strength of the solution. Once again, these
parameters affect only the settling rates and do not produce a I C o n c e n t r a t i o n : 0.0s %
noticeable influence on turbidity or final mud volume. The D o s e : 3,s ppm
magnitude of the effect of pH of the flocculant solution on 181'
the settling rate depends on the DH of the flocculant. It can be .-c
seen from Figure 3 that the optimum flocculant (Talosep A3)
shows the greatest effect, increasing its activity as the pH of the
solution is raised. The other two flocculants with DH above and

Superfloc A110
-6
0

4 Superfloc Also
L I 1 I I I I I
5 7 9 11
pH
Figure 3: Effect of tlocculant solution pH on settling rate.

Flocculant : Superfloc A130 a Distilled water

D o s e : 0,8 ppm
b IL Tap water
1 2 3 4 5 6 c IL' Condensate
Time hrs.
d Laboratory - t a p water
Figure 1: Effect of solution agitation time on settling rate.

.- Flocculant : S u p e r f l o c A130
Dose : 0,8ppm

0-.
Q,

RI
16

ul
C
E~~-/
Q,
c,
0,2 % Diluted t o 0,025 % Flocculant : T a l o s e p A3
a D o s e : 1,7ppm
o 0,3 % Diluted to o,os %
- C o n c e n t r a t ion : 0 , 0 5 %
-
.m12-
.-uc
- I I I I I 1
0,l
I I 1

0-3
I I
0,5 -5 -4 -3 -2
Concentration /o
0 Log [ specific cond. mho/cm ]
Figure 2: Effect of tlocculant solution concentration on settling rate. Figure 4: Effect of dissolved salts in the tlocculant solution on settling rate.
 108 Proceedings of The South African Sugar Technologists' Association - June 1978

flocculant preparation water have on settling rates. Solutions at variation in clarification with flocculant dose. Settling rates are
relatively high pH (pH 8,5) are unaffected while those at lower very sensitive to the amount of flocculant used and increase to a
pH (pH 7) show a decline in settling rate at relatively high values maximum level. Turbidity decreases extremely rapidly and
of conductivity. The graph indicates that little is gained by using reaches an equilibrium value which ever increasing flocculant
distilled or deionized water in place of tap water as a solvent for amounts do not affect. Similarly, final mud volume shows an in-
these flocculants. It is clear that the amount of dissolved ionic itial rapid decrease followed by a slow decline. These results
species must be raised considerably before any significant show why the settling rate is more sensitive than turbidity or
decrease in eficiency is noted. mud volume to changes in the flocculant solution.
The above effects relating to flocculant efficiency should be The variation of final mud volume, turbidity and initial settl-
compared with the curves shown in Figure 5 which illustrate the ing rate for clarification withoiit flocculant aid at increasing
PzOs levels in mixed juice (100-400 ppm) and at a constant
calcium concentration in clear juice (320 ppm) are shown for a
typical juice in Figure 6. The increase in settling rate and clarity
obtained by raising the amount of P205 precipitated are the im-
provements expected of an increase in the inorganic precipitate 3'
and are achieved at the expense of an increase in final mud
volume.
Figure 7 illustrates the behaviour of three polyacrylamide floc-
culants of varying D H when used to flocculate the limed juice
samples shown in Figure 6. The flocculants of low (A1 10) and
intermediate (A130) D H show increases in settling rate with the
A 130 curve yielding a maximum near 300 pprn P2O5 and then
decreasing markedly. The curve for the flocculant with low DH
appears to rise to a maximum near 400 pprn P205. It is clear

L I I I I I I I Precipitated F$05 PPm

1 2 3 4 5
Dose pprn Figure 6: Effect of amount of PZOJprecipitated on clarification without
' Figure 5: Example of the effect of tlocculant dose on clarification. tlocculant. Amount Caf+ in clear juice was 320 ppm.
 Proceedings of The South African Sugar Technologists' Association - June

that whereas Superfloc A130 is the optimum flocculant at low The clarification properties of the limed, unflocculated
and intermediate phosphate levels, it is vastly inferior at high suspension at various calcium concentrations in clear juice are
P2Os levels. The high DH flocculant (A150) shows only a slight illustrated in Figure 8. Since the quantity of P205 in mixed juice
improvement in settling rate. All three flocculants show decreas- was held constant at 320 ppm, the initial changes are also af-
ing turbidity values up to 300 pprn Pz0s above which the clarity fected to some extent by the increasing P205 precipitation. At
remains nearly constant. It can be seen that even at high P205 100 pprn Ca 2+ in clear juice, 75% of the P205 had been
concentrations, at which stage it yields the best settling rate, the precipitated; at 200 ppm, 85% and at 450 ppm, 90%. The
low DH flocculant produces reasonable, but not optimum, suspension yields a decrease in settling rate as the calcium con-
clarity. The final mud volume appears to be less sensitive to in- centration increases. All the properties change gradually and
creases in P205 concentration in the flocculated case than it is reach a fixed value between 400 and 500 pprn Ca 2+ The
for the unflocculated suspension. experiments shown in Figure 8 , were repeated with
polyacrylamide flocculants of varying DH and the results with
respect to final mud volume and turbidity are shown in Figure 9
Superfloc A110 and initial settling rate in Figure 10. The former two parameters
behave in an analogous manner to the unflocculated suspension
0 Superfloc A130 although they achieve lower turbidities and lower mud volumes.
Figure 10 indicates that the flocculants of intermediate and high
A Superfloc A150 DH decrease in activity as the calcium concentration rises, the
effect for A130 being quiet substantial. On the other hand, low
18 DH A 1 10 increases the settling rate up to 1 000 pprn Ca2+.
aJ
C,

16
PC

0 l4
.E-
- 12
*.r
C,

$ lo-

26 -
"\ 24 -

16

100. 200 300 400 L L , 1 1 I I 1 I

200 2+ 400 600 800 1000 1200
Precipitated I?jO5 pprn Ca in Clear Juice PPm
Figure 7: Effect of amount PZOJprecipitated on clarification with floc-
culants of varying degrees of hydrolysis. Amount c a 2 + in clear juice was 320 Figure 8: Effect of the quantity of CaZ+ in clear juice on clarification
without flocculant. Amount P20, in mixed juice was 320 ppm.
PPm.
 1 10 Proceedings of The South African Sugar Technologists' Association - June 1978

These results indicate that the activity of polyacrylamide floc- in clear juice (380 ppm). The curves illustrated can thus be
culants of varying D H depends on the amount of inorganic divided into two sections. The first below pH 7 when phosphate
precipitate formed and on the quantity of calcium remaining in precipitation increases and the second above pH 7 when
solution. The relative shape and position of the curves shown in precipitation remains constant and all effects are due to pH
Figure 10 are expected to depend on the quantity of phosphate changes. As expected, mud volume rises to a maximum at pH 7
precipitated. Thus for each flocculant there is a family or set of and then remains unaltered. Similarly, clarity improves to a
curves, each curve characterizing a particular phosphate level. maximum but deteriorates at high pH values possibly due to the
Similarly the curves shown in Figure 7 will be sensitive to stabilization of some suspended matter. The behaviour of the
calcium concentration in clear juice. If the respective P2Os and flocculants with respect to settling rates is more complicated.
Ca2+ concentrations were to be plotted in the X- and Y- The suspension without flocculant appears to be pH insensitive
directions, then the settling rate obtained with a particular floc- and the flocculant of high DH (A150) shows little variation with
culant could be represented as a height above this plane. The pH. The flocculant of low D H (A1 10) shows a maximum rate at
result would be a surface in three dimensions indicating the floc- pH 7, 2, corresponding to a minimum produced by the floc-
culant activity under different conditions. culant of intermediate DH (A130). Thus the three dimensional
All the above tests were performed at a liming pH of 7,5 (at surface representing flocculant activity mentioned above varies
70C) which generally gave a clear juice of around pH 7,O (at according to the pH to which the suspension is raised.
25OC). At this level, the precipitation of P205 had reached its
limiting value, maximum precipitation generally being obtained Discussion
from a liming pH of 7 upwards. Figure 11 indicates the effect The polyacrylamide flocculants used in sugar manufacture
liming pH has on clarification at a constant P205 concentration are long chain copolymers of acrylamide and sodium acrylate.
in mixed juice (320 ppm) and a constant calcium concentration These long chain compounds are intertwined in the solid
phasell. 14. Thus, when a solution is prepared, time must be
allowed for the molecules to become hydrated and to disen-
3,4- tangle. Figure 1 shows the time required before an effective
Superfloc A l l 0 equilibrium is reached with respect to the clarification properties
of the flocculant. Even in this state, a network of entangled
0 Superfloc A130 polymer molecules exists in solution12 and has been associated
3,o
- with its flocculant activity4. The fall off in flocculant activity at
A Superfloc A150
very low concentration (Figure 2) may possibly be explained by
'.
a weakening of this network.
2,6 - The activity of a flocculant is dependent on its conformation
in solution. Since these polyacrylamides are required to act as
,.

.-
Y

I-

0.2 -
: . .- 0
-
O
A

A/A
/A
A-

0 -0

Y 7 - 9
I /
Is-8 1 I I I I 1
1 I 1 1 I I I 1
200 400 600 800 1'000 1200 200 400 600 800 1000 1200
c a 2 + in Clear Juice pprn c a 2 + in Clear Juice pprn
Figure 9: Effect of the quantity of Caa+ in clear juice on clarification with Figure 10: Effect of the quantity of Ca 1+ in clear juice on settling rates
flocculants of varying degrees of hydrolysis. Amount P205 in mixed juice was using flocculants of varying degrees of hydrolysis. Amount P 2 0 ~
in mixed juice
320 ppm. was 320 ppm.
 Rroceedings of The South African Sugar Technologists' Association - June

-
bridges between impurity particles during floc formation, they
Superfloc All0 have their best performance when the polymer chain is at its
maximum extension. This extended conformation is caused by
o S u p e r f loc / A130 the presence of intra-molecular repulsive forces generated by the
ionization of the acrylate groups along the chain. This dissocia-
tion is pH dependent, reaching its maximum at high pH and cor-
A ~ u ~ e r f l'* o eA150
i responds to the enhanced activity of the flocculants shown in
Figure 3. The extension of the polymer in solution is also af-
fected by the ionic content of the solvent. At high ionic con-
centrations, the anionic centres on the molecules gather around
themselves a sufficient ionic atmosphere of counter-charge to
become neutralized to a certain degree and the long polymer
does not achieve such a favourable extended conformation.
?'he results in Figure 5 indicate that the flocculant first
scavenges all the particles to which it can adsorb and rapidly
achieves an optimum clarity. With increasing amounts of floc-
culant larger flocs are formed but little of the residual turbidity is
removed.
The physico-chemical mechanism of juice clarification de-
pends on the precipitation of calcium phosphate in a primary
flocculation4 in which the impurities are bound into the floc. The
action of the polyacrylamide is to produce a secondary floccula-
tion, the extent of which depends on the D H of the polymer 6,'0.
The presence of an optimum DH has been explained in terms of
a balance between particle-flocculant bonding and the need for
the flocculant t o extend into the solution and act as a bridging
agent15. It is already known3 that an increase in the amount of
phosphate precipitated leads to the formation of larger flocs and
a more extensive primary flocculation. The overall result is an
increase in settling rate and clarity of clear juice (Figure 6).
Similar trends are observed with polyacrylamides (Figure 7),
although the presence of a maximum in settling rate suggests
that the interaction between particle and flocculant is dependent
on particle size. Thus as the primary floc size increases, the in-
dividual area available for polymer adsorption also enlarges with
the result that the flocculant of intermediate D H with its relative-
ly large proportion of carboxylate groups passes into the over-
bound condition. Similarly, the previously ineffective low D H
flocculant with its smaller proportion of carboxylate groups
becomes the optimum flocculant at high amounts of precipitated
phosphate.
Since both primary and secondary flocculation mechanisms
rely on the adsorption of calcium ionsZonto impurity particles,
an increase in calcium concentration aids flocculation and thus
leads to an improved clarity (Figures 8 and 9). Superimposed on
this is an increase in P20s precipitate of 40-50 ppm and a
change in Ca : PO4 ratio in the precipitate of 1,O to 1.6 which
occurs over the concentration range 100 to 400 ppm Ca2+in
clear juice. The increase in impurity removal and inorganic
precipitate explains the observed increases in mud volume. The
rapid improvement in primary flocculation indicated by the
rapid decrease in turbidity suggests that increasing amounts of
organic impurity are being flocculated with only this small in-
crease in the amount of inorganic precipitate. The primary flocs
will thus probably become smaller in size leading to the decrease
in settling rate. The variation of settling rate with CaZ'
concentration observed with the use of polyacrylamides appears
to be related to the specific interaction between polyacrylamide
and adsorbed Ca2+ions rather than to changes in primary floc
size. Whereas the high DH flocculant remains in the overbound
condition even at relatively low Ca2+ concentrations, the
intermediate D H flocculant tends to this state with increasing
Ca2+ concentration and the low DH flocculant achieves
6 7 8 9 optimum efficiency when the increase in adsorbed CaZ' ions
Clarification pH 70C produces stronger particle-flocculant binding.
Raising the pH of a mixed juice suspension at constant CaZ+
Figure 11: Effect of liming pH on clarification of a sample with fixed quan- concentration in clear juice produces changes in a number of
tity PZOJin mixed juice (320 ppm) and fixed quantity of Ca2+ in clear juice parameters and it is thus difficult to assign exact causes to the
(380 P P ~ ) . various effects noted (Figure 11). As the pH is raised increasing
 112 Proceedings of The South African Sugar Technologists' Association - June 1978

amounts of calcium phosphate are precipitated and the extent of 4. Bennett, M.C. (1975). Flocculation technology in sugar manufacture.
Ca2+adsorption also increases2. The similarity of the settling Sugar Ind. Tech. 34 : 22-32.
curves below pH 7 to those produced with increasing Ca2+ 5. Bond, J.D. (1925). Defecation in sugar cane manufacture Int. Sug.J., 27 :
concentration suggest that this is the primary cause behind the 311-317.
changes in flocculant behaviour. Above pH 7 the increasingly 6. Crees, O.L., ale; D.J., Whayman, E and Willersdorf. A.L. (1973). Floc-
culants in raw sugar production. Proc QSSCT 40 : 239-250.
alkaline medium might render some of the adsorbed calcium in-
active and a reverse in behaviour is noted. It is, however, clear 7. Farnell, R.G.W. (1926). J. Soc. Chem. Ind., 45 : 343T.
that the behaviour of the suspension in respect of the pH will be 8: Lionett, G.R.E. and Ravno, A.B. (1976). Flocculant assessment using a
very much dependent on the exact calcium and phosphate con- portable batch settling kit. SASTA Proc 50 : 176-178.
centrations. 9. Meade, G.P. (1963). Spencer-Meade cane sugar handbook. John Wiley
and Sons, Inc. New York and London. Ninth Edition : 95 and 99-100.
Acknowledgements 10. Murray, J.P. and Shephard, G.S. (1975) Some aspects of flocculant use in
The author wishes to extend thanks to Dr. M. Matic for help- clarification and filtration. SAS'I'A Proc 49 : 53-58.
ful discussion and to Mrs. A. Dunsmore and Mrs. J. Day-Lewis 11. Narkis, N..and Rebhun, M. (1966). Ageing effects in measurements of
polyacrylamide solution viscosities. Polymer. 7 : 507-5 12.
for analytical results.
12. Rochow, T. G. (1961). Resinography of high polymers. Analyt. Chem.
33 : 1810-1816.
REFERENCES 13. Sockhill, R.D. and James, G.P. (1962). The effects of added calcium in
1. Bennett, M.C. (1957). The physical nature and behaviour of cane sugar cane juice clarification. Proc. QSSCT 29 : 83-88.
juice. Int. Sug. J., 59 : 176-178 and 208-212.
14. Shyluk, W.P. and Stow, Jr., F.S. (1969). Ageing and loss of flocculation
2. Bennett, M.C. (1957). The binding of calcium ions by adsorbed proteins in activity of aqueous polyacrylamide solutions. J. Applied Polymer Sci., 13
cane sugar juice. Chem. & Ind. : 1098-1 100. : 1023-1036.
3. Bennett, M. C. and Ragnauth, J.M. (1960).The effects of calcium and 15. Whayman, E. and Crees, O.L. (1974). Mechanistic studies of cane mud
phosphate in cane juice clarification. Int. Sug. J., 62 : 13-16 and 41-44. flocculation. ISSCT Proc I5 : 1175-1182.