QUT Digital Repository:

http://eprints.qut.edu.au/

Greenwood, Judy and Rainey, Thomas J. and Doherty, William O.S. (2007) Light
scattering study on the size and structure of calcium phosphate/hydroxyapatite
flocs formed in sugar solutions. Journal of Colloid and Interface Science
306(1):pp. 66-71.

Copyright 2007 Elsevier

Light scattering study on the size and structure of calcium

phosphate/hydroxyapatite flocs formed in sugar solution
by
Judy Greenwood
Formerly Department of Chemical Engineering, University of Sydney, Sydney, New
South Wales, Australia
Tom Rainey and William O. S. Doherty1
Sugar Research and Innovation, Queensland University of Technology, Brisbane,
Australia
ABSTRACT
The formation, flocculation and sedimentation of calcium phosphate particles are among
the main physico-chemical reactions that occur during sugar manufacture. The
mechanisms through which processes occur in sugarcane juice are still poorly
understood. This study (being part of a comprehensive investigation to unravel these
mechanisms) reports on the size and structure of calcium phosphate particles and
aggregates in water and sugar solutions at 20 oC using the small angle laser light
scattering technique. The average size of the primary calcium phosphate particles was in
the range 10.4 m to 17.5 m and the scattering exponents, which describe the structure
of the calcium phosphate flocs, varied from 1.97 to 2.76. The flocs formed without
flocculant are more compact in water than those formed in sugar solution. The
compactness of the flocs was also affected by pH of the solutions. This effect has been
explained by considering the electrical double layer phenomenon.
INTRODUCTION
There are numerous areas in the processing of cane sugar manufacture that are not clearly
understood. As outlined by Doherty and Edye1, one of those areas is the mechanism of
cane sugar clarification. Poor understanding of juice clarification is due to the complex
physico chemical reactions occurring during the process. However, the main chemical
reaction during the process is that between soluble phosphate in the juice and added free
calcium ions (in the form of lime/lime saccharate) to raise the juice pH and prevent
sucrose degradation. This reaction results in the formation of calcium phosphate flocs.
An anionic polymer, a co-polymer of acrylamide and acrylic acid, is added to bridge the
flocs and enhance their sedimentation.
Deben2 determined that the precipitation reaction is controlled by the initial reaction:

Ca 2+ + HPO42 CaHPO4

The calcium phosphate formed is insoluble in water and is unstable, undergoing a

secondary reaction to form a more stable compound (hydroxyapatite).

To whom correspondence should be addressed. Fax: 61 7 4952 7699

Email: w.doherty@qut.edu.au

Honig3, Deben2 and Jourani et al.4,5 found that the dominant calcium phosphate phases
formed were as follows:

Ca 2 + + 2 H 2 PO4 Ca (H 2 PO4 )2

Ca 2 + + HPO42 CaHPO4

2CaHPO4 + 2Ca3 (PO4 )2 Ca8 H 2 (PO4 )6

Ca3 (PO4 )2 + 2Ca 2 + + HPO42 + H 2O Ca5 (PO4 )3 OH + 2 H +

Tetra calcium phosphate (Ca4(OH)2(PO4) may also form.

There are a number of issues associated with calcium phosphate precipitation. Firstly, the
nature of the phase of the calcium phosphate precipitate is highly dependant on the
calcium and phosphate concentrations, pH, and the surface that precipitation is occurring
on6. Secondly, in the presence of organic acids (such as citric acid and isocitric acid), the
precipitation of calcium phosphates is inhibited6. Also, carbonates and magnesium ions,
even at low concentrations, will affect the rate of precipitation of phosphates6,7.
The structure of flocs is important in solid/liquid separation systems. A floc consisting of
a more open structure will firstly help to trap insoluble components in the fluid and
secondly it may allow for an increased settling rate as fluid can flow through the floc as
well as around it. Sutherland and Tan8 tested the hypothesis that an open floc structure
allowed for the flow of fluid through the pores. They found that under most conditions
the assumption that the floc contains immobilised fluid holds true, except in the case
where hindered settling occurs. As the solution becomes thicker and settling velocities
fall, the importance of permeation velocity increases. This is relevant in the case of
settling of calcium phosphate flocs in sugarcane juice, as settling occurs by the hindered
settling mechanism9. Alternatively, it may be that the higher the porosity the greater the
drag force. That is, the greater the surface area for interaction between the solution and
the floc, the higher the expected drag force. Experimentally, it has been shown that
porous spheres experience higher drag than impermeable spheres. Surface area effects
rather than internal fluid flow have explained this. This was detected at higher Reynolds
numbers where the surface roughness plays an increased role10.
There have been numerous studies that sought to characterise flocs as fractal objects
using small angle laser light scattering (SALLS). Examples include studies of the
structure of waste water flocs11, silica flocs12, polystyrene13, hematite14 and flocculated
polystyrene latex15 aggregates. Compact aggregates have been found to have fractal
dimensions of 2.3-2.5 (or higher if restructuring has occurred) and loose aggregates
typically have fractal dimensions of 1.7-1.814,16.

Determination of fractal dimensions using light scattering

There are various techniques that allow the measurement of the fractal dimension of
particles. They range from small angle X-ray scattering, small angle neutron scattering to
SALLS. All of the mentioned techniques, except SALLS, can be used to study colloids.
Due to limitations of the techniques, they can not be used successfully to measure the
aggregates formed due to flocculation. The reason for this is that the aggregates formed
during flocculation are larger than colloids and can range in size from a few microns to a
few millimetres.
The theory underlying the interpretation of the scattering data is the Rayleigh-GansDebeye (RGD) theory17. The theory involves the assumptions that the primary particle
size is comparable to the wavelength of the incident beam, that the particles have low
refractive index and that multiple scattering is negligible. The interpretation of SALLS
raw data involves finding the scattered light energy momentum transfer vector, Q, from
the scattering intensity of the flocs with respect to their scattering angle, as shown in
Equation 6.

I (Q ) M (Rg ) P(Q )S (Q )
2

Where I(Q) is the scattering intensity, P(Q) is the form factor of the primary particle and
S(Q) is the interparticle structural factor. Q is related to the refractive index of the
particles, the wavelength of the light and the angle of the detector to the incident beam, as
shown in Equation 7.

4n
Q=

Where

sin
2

- refractive index of the liquid

- wavelength of the laser
- angle between the laser beam and detector

RGD theory allows the fractal dimension of the floc to be determined from the following
simplified relation.18

I Q dF

where I is the scattering intensity, Q is the momentum and dF is the fractal dimension.
The power law scattering behaviour of a fractal object can be determined from a log-log
plot of the scattering intensity as a function of the momentum transfer, where the slope of
the line gives the fractal dimension.
Recent work by19 has developed a model to describe the settling behaviour of fractal
aggregates of calcium phosphate. Compact calcium aggregates, with fractal dimensions
varying from 2.3 to 2.8 were used in this study. Though the study reported producing
calcium phosphate flocs in sucrose systems, the effect the sucrose molecule on the size
and structure of calcium phosphate particles was not presented nor discussed. This
present study is an extension of the work of Tang and co-workers19 and examines the
effect of pH and sucrose on the size and structure of calcium phosphate.
EXPERIMENTAL

Preparation of calcium phosphate flocs

Calcium hydroxide suspension was prepared by mixing 5.6 g of analytical grade calcium
hydroxide with 25 mL of nano-pure water. The nano-pure water was prepared by using
deionised water that was filtered through activated carbon, ion exchange resin and a
membrane filter.
Calcium phosphate flocs in water were prepared by dropwise addition of freshly prepared
calcium hydroxide suspension to 500 mL of 5.74 gL-1 phosphoric acid to reach the
endpoint of either pH 7.6 or pH 7.8. The solution was stirred continuously during the
addition of the calcium hydroxide suspension. pH values of 7.6 and 7.8 were chosen
because in sugar milling, clarification is carried out around these pHs. The working
temperature selected for this investigation was ~20 oC.
Calcium phosphate flocs in sucrose solution were prepared by adding freshly prepared
calcium hydroxide suspension to 500 mL of 5.74 gL-1 phosphoric acid containing 70 g of
analytical grade sucrose.
A stock solution (5 gL-1) anionic flocculant was prepared with nano-pure water. This was
diluted to 1 gL-1 prior to use.
Analysis with small angle laser light scattering (SALLS)

The Malvern Mastersizer 2000 was used to obtain light scattering data from which size
and fractal dimension could be calculated.
A preliminary evaluation on the optimum particle concentration for accurate calculation
of the floc size and structure indicated that 3 mL of calcium phosphate solution (prepared
as described in the previous section) in 500 mL of nano-pure water adjusted to pH of 7.6
or 7.8 was required.
To ensure accurate measurements of the primary particle size of the unflocculated
calcium phosphate particles, a solution of the particles was sonicated for 7 minutes
immediately before measurement in the Malvern Mastersizer. For the measurement of
aggregate size and structure the technique above was modified because the aggregates
were delicate in nature and easily broken up. The sample cell contained nano-pure water
adjusted to the required pH. The aggregates were then extracted from the beaker with a
pipette with a broad spout to prevent damaging them. Caution was used also when
injecting them into the cell so that no great shear forces were exerted on them that could
potentially damage them.
The calcium phosphate flocs were investigated by SALLS both in the presence and
absence of a flocculant, and in water and sucrose solutions.

A Leica zoom 2000 dissecting microscope and an Olympus C800L digital camera was
used to capture black and white photographs of calcium phosphate flocs. Due to the
three-dimensional structure of the flocs not all of the flocs were able to be in focus
simultaneously. As a result it was only possible to focus on different levels of the floc.
The structures were investigated at three different magnifications (2.8, 6.3 and 13.2
times).
X-ray powder diffraction of calcium phosphate flocs

The crystallographic forms of the calcium phosphate flocs were determined by X-ray
powder diffraction (XRD) using a PANalytical XPert MPD with CuK radiation of
wavelength 1.5418 .
RESULTS
X-ray powder diffraction results

The XRD spectra of the samples are presented in figure 1. All the patterns show weakly
diffracting microcrystalline samples. Although the sample prepared in water at pH 7.8
show no identifiable compound, the other samples contained hydroxyapatite as the major
component. The results of figure 1 show that the presence of sucrose had no influence on
the crystallographic form of calcium phosphate formed.

Ca phosphate pH = 7.6 water

200

Intensity(Counts)

Ca-phosphate pH = 7.8 water

200

0
Ca-phosphate pH = 7.6 sucrose
200

0
Ca-phosphate pH = 7.8 sucrose
200

0
SQR(I)

09-0432> Hydroxylapatite, syn - Ca5(PO4)3(OH)

10

20

30

40

50

60

Two-Theta (deg)

Figure 1. X-ray powder diffraction pattern of calcium phosphate flocs

Calcium phosphate without flocculant addition
Particle size

70

To ensure measurement of primary particles only, and not aggregates, shear forces were
applied to the particles by stirring the solution before measurement in the Mastersizer.
The 50th percentile range diameter (D (v,0.5)) of the particles over 10 measurements was
recorded over a period of 5 minutes in water and sugar solutions. The 50th percentile
diameter allows for a comparison of average sizes under differing conditions. Table 1
gives the size of the particles in water and sugar solutions.
Table 1.

Average particle size and scattering exponent of calcium phosphate

formed in the absence of flocculant

System pH

D (v,0.5)

Scattering exponent

(m)
Water

Sugar solution

Water

Sugar solution

7.6

10.4

13.6

2.76

2.66

7.8

13.6

17.5

2.62

2.61

The results indicate that larger sized particles were formed in sugar solutions and at
higher solution pH. The differences observed between the particle sizes of calcium
phosphate in water and sugar solution is probably due to increased solubility of lime in
sucrose. This effect would increase calcium phosphate supersaturation, reduce the
number of nucleating sites and thereby increase the crystal growth of hydroxyapatite.
The increase in calcium phosphate particle size from a solution of pH 7.6 to a solution of
pH 7.8 is due to increase in hydroxyl ion concentrations. The rate of formation of
hydroxapatite from amorphous tricalcium phosphate is dependent on the hydroxyl ion
concentration20.
As the sizes of the primary particles are not comparable in size to the wavelength of the
incident light, this assumption in RGD theory is no longer valid. Consequently, it was
not possible to determine the fractal dimension of the particles. However, the scattered
light still contains information about the structure of the floc and can be used for
comparative purposes. To distinguish our measurement from a true fractal dimension,
the measured quantity will be called the scattering exponent.
Scattering exponent

The power law scattering behaviour of flocs can be used to investigate the structure of
flocs, by a log-log plot of the scattering intensity (I) as a function of the momentum
transfer (Q). Figures 1 and 2 illustrates the I vs Q trends. Figure 1 shows that for the
calcium phosphate flocs in water (at both pH 7.6 and 7.8), there is a kick in the I vs Q
plot. The kick observed here is not as large as that found by Wong et al.21 with
polyacrylamide flocs. Wong and co-workers found that this behaviour was unique to
flocs produced in neutral or basic pH with polymers of low charge per unit length and
with a high molecular weight. They postulated that the kick is caused by the

compression of the electrical double layer. This is therefore likely to have occurred with
the calcium phosphate flocs in this study.
The results of the scattering exponents for the calcium phosphate flocs are given in table
1. The scattering exponent decreased as the pH increased for both the water and sugar
systems. Jung and coworkers (1995) found that for similar sized primary particles, a
higher scattering exponent implies a higher packing density. In this work as the primary
particles do not vary significantly under the two different pH values, it can be assumed
that the increased scattering exponent directly implies a higher packing density.
There are two factors that influence the scattering exponent of aggregates formed in the
absence of a flocculant: speciation and compression of the zone of influence. Speciation
is not an influencing factor for the aggregates formed in this project because the main
component is hydroxyapatite. However, compression of the zone of influence could
potentially account for the discrepancy between the sucrose and water solutions at the
same pH. The reason is that there is the possibility of ionisation of the hydroxyl ions on
the sucrose molecule. Hence, the sucrose solution would have a relatively greater
number of hydroxyl ions that could lead to compression of the electrical double layer
allowing particles to approach each other more closely and be dominated by Van der
Waals forces. This will lead to an increase in the rate of flocculation, and hence impact
on the type of floc structure.

10 00 0 0

10000

1000

I
100

10

1
1 . 0 0 E -0 6

1 . 0 0 E -0 5

1 . 0 0 E -0 4

1 . 0 0 E -0 3

1 . 0 0 E -0 2

0.1

Q (/n m )

Figure 2.

Scattering intensity vs momentum of calcium phosphate flocs in a

water solution at pH 7.8

100 00

100 0

I
100

10

1
1 .0 0 E - 0 5

1.00E -04

1 .0 0 E - 0 3

1 .0 0 E - 0 2

0 .1

Q (/n m )

Figure 3.

Scattering intensity vs momentum of calcium phosphate aggregates in

a sugar solution at pH 7.8

Size of flocs

There was a great variety in the sizes of the aggregates formed without the use of a
flocculant. Microscopic analysis of the flocs confirmed that the aggregates formed were
loosely held. Figure 4 shows the structure of the aggregates formed in water at pH of 7.8.
The aggregates varied between 300 m to > 1000 m (past the detection limit of the
Mastersizer). With stirring it was noted that any large aggregates that had formed were
broken up by shear forces. For instance, aggregates formed in water (pH 7.8) had a mean
diameter of 390 m without shear forces. With shear forces the average size of the
aggregates was reduced to 54 m within 30 s of the stirrer in the Mastersizer being
switched on. Over longer time intervals the results indicate that aggregates increase in
size again until a balance between shear forces and aggregate size is achieved.
The radius of gyration was calculated for the aggregates formed in the water solution at
pH 7.6 and 7.8 and was 262.3 m and 293.6 m respectively. The radius of gyration of
the aggregate formed in a pH of 7.8 was approximately 12% higher than that at 7.6. The
larger size of the aggregates formed in a solution pH of 7.8 may be explained by double
layer compression due to availability of hydroxyl ions. As the particles can then
approach each other more closely, there is the enhanced probability of Van der Waals
forces playing a role in keeping the particles together.

Figure 4.

Calcium phosphate flocs formed in water in the absence of a

flocculant

Calcium phosphate flocs formed in the presence of a flocculant

Size and structure of flocs

Calcium phosphate formed on addition of calcium hydroxide slurry to a solution of

phosphoric acid was flocculated with a co-polymer causing aggregation of the primary
flocs into secondary flocs. The sizes of many of the flocs were ~1 mm in diameter, and
because of this, no accurate floc size could be obtained with the Mastersizer. The
Mastersizer only has the capability for measuring particles in the range 1 m to 900 m.
Figure 5 illustrates calcium phosphate flocs formed in sucrose solution in the presence of
a flocculant. These flocs are denser than those of figure 4 formed in the absence of
flocculants.

Figure 5

Calcium phosphate flocs formed in sucrose solution in the presence of

flocculant

As in the case with calcium phosphate flocs formed without flocculant, the fractal
dimension of the floc could not be measured because of the large primary particle sizes.
However, the scattering exponents could be calculated. Figure 6 is a typical plot of the
scattering intensity, I, vs Q for calcium phosphate particles with flocculant. The plot
illustrates that at low Q values, deviation from the power law behaviour occurred. Wong
and coworkers (1988) found that with flocs produced with polyacrylamide, interparticle
interference dominated at low Q values while intraparticle behaviour occurred at high Q
values. Hence the observed behaviour in this study could be accounted for by
interparticle interference.

-2 .2 8 8 8

y = 6 E -0 8 x
2

R = 0 .99 8 9
1000

100

10

1
1 .0 0 E -0 5

1 .0 0 E -0 4

1 .0 0 E -0 3

1 .0 0 E -0 2

0 .1

0 .0 1

Q (/n m )

Figure 6.

A typical plot of scattering intensity vs. momentum for calcium

phosphate flocs formed in the presence offlocculants

Effect of pH and flocculant dosage

Table 2 gives the results of the scattering exponents formed under different conditions.
The results show that, as expected, the use of a flocculant results in less compact flocs.
(cf. table 1). Also, there is an increase in the scattering exponent with increase in
flocculant concentration. Contrary to the results obtained in the absence of flocculant,
the calcium phosphate flocs formed in sugar solutions were in general less compact than
those formed in water.
The results of table 2 also show that, in general, the scattering exponent was lower at the
higher pH, similar to the studies were no flocculant was used.

Table 2

Scattering exponent of flocculated calcium phosphate aggregates

Flocculant concentration

System pH

3 mg.kg-1

8 mg.kg-1

Water

Sugar solution

Water

Sugar solution

7.6

2.21

1.98

2.44

2.12

7.8

1.97

2.02

2.34

2.00

Two of the major considerations in the flocculation of the particles are the amounts of
flocculant needed for effective flocculation and the effect of restructuring. It is expected
that the speed of formation of the floc would be inversely proportional to the density of
the floc and hence speed of formation will impact on floc structure. A floc that forms
rapidly would be expected to be less dense as the particles have not had time to arrange
themselves to form the best packed structure.
The amount of flocculant (based on surface area) would also be a significant factor to
consider because, as bridging flocculation occurs in these systems, the flocculant must
contact the particles and attach to them. Hence, to achieve a high proportion of removal
of the particles in solution it is necessary to ensure that there are enough sites for the
flocculant to attach to in order for well-structured flocs to form.
In a typical sugar milling factory the solid load to the clarifier varies depending on a
number of factors including the cane variety, the soil type and whether the cane was
harvested after a rainfall. Therefore, dosing flocculant based on solid size and loading
can probably achieve a more effective flocculation than dosing by solution volume as is
presently practised by sugar factories. It is therefore necessary to develop an on-line
device to measure the amount of solid loadings in secondary juice prior to clarification
and relate it to flocculant dose.
CONCLUSION

The diameter of the primary particles of calcium phosphate in water and sugar solutions
were in the range 10.4 and 17.5 m, and so could not be used to calculate the fractal
dimension of calcium phosphate aggregates. However, the SALLS technique was used to
determine the scattering exponent and hence describe the structure of the calcium
phosphate flocs.
Calcium phosphate aggregates formed without flocculant are more compact than those
formed with flocculant. However, the flocs formed with flocculant are stronger and
hence are able to resist shear forces. The results show that the compactness of the flocs is
affected pH and the precipitating medium and have been explained using the electrical
double layer phenomenon.

Further investigation is required on the effect of sugarcane juice impurities (soluble and
insoluble) on calcium phosphate floc size and structure.
ACKNOWLEDGMENTS

The authors thank Sugar Research Institute, Mackay, Australia and the Australian
Research Council for financial support of the project. The laboratory work was
conducted by Nicole Marchhardt of the University of Sydney.

REFERENCES
[1] Doherty W.O.S., Edye L. A., Proc. Aust. Soc. Sugar Cane Technol., 1999, 21: 381-388.
[2] Deben J. G., International Sugar Journal, 1976, 78: 3-7, 35-39, 73-75.
[3] Honig P., Principles of Sugar Technology, Elsevier Scientific Publishing, Amsterdam, 1953.
[4] Jourani A., Bunahmidi T., Laguerie C., International Sugar Journal, 1995, Vol 97, no 1162, pg. 586592.
[5] Jourani A., Bunahmidi T., International Sugar Journal, 1995, Vol 97, no 113, 633-639.
[6] Nancollas, G. H., Amjad, Z. and Koutsoukos, P., ACS symp. Ser., 1979, 93: 475-497.
[7] Feenstra K., Hop J., De-Bruyn .P L., J. Colloid Int. Science, 1981, 83, 583-588.
[8] Sutherland D. N and Tan C. T., Chemical Engineering Science, 1970, Vol 25 p1948-1950.
[9] Crees, O. L., PhD thesis, James Cook University, 1988.
[10] Masliyah J., Polikar M., Canadian Journal of Chemical Engineering, 1980, Vol 58, p299-302.
[11] Li D. and Ganczarczyk J., Environ. Sci. Technol., 1989, vol.23 no11, 1385-1389.
[12] Tang P., Colflesh D. E., Chu B., Journal Colloid Interface Science, 1998, 126, 304.
[13] Zhou Z., and Chu B/, Journal Colloid Interface Science, 1991, 143, 356.
[14] Amal R., Raper J. A., Waite T. D., Journal Colloid Interface Science, 1990, 140, 158.
[15] Burns J. L., PhD Thesis, University of Newcastle, 1999.
[16] Guan J., Water Sci. Tech., 1998, Vol 38 no 2 9-15.
[17] Jung S. J., PhD Thesis, University of NSW, 1997.
[18] Jung S., Amal R., Raper J., Part. Syst. Charact, 1995, 12, 274-278.
[19] Tang P., Greenwood J., Raper J.A., J. Colloid Interface Science, 2002, 247, 210-219.
[20] Doherty W.O.S., Crees, O. L., Senogles E., Cryst. Res. Technol, 1996, 31, 281-286.
[21] Wong K., Cabane B., Duplessix R., Journal Colloid and Interface Science, 1988, 123, 466.