Water Research 36 (2002) 1503–1512

Destabilisation of oil–water emulsions and separation by

dissolved air flotation
A.A. Al-Shamrania, A. Jamesa,*, H. Xiaob
a
The University of Manchester Institute of Science and Technology UMIST, Department of Chemical Engineering,
P.O. Box 88, Manchester M60 1QD, UK
b
The University of Manchester Institute of Science and Technology UMIST, Department of Paper Science, P.O. Box 88,
Manchester M60 1QD, UK
Received 6 July 2000; received in revised form 4 May 2001; accepted 4 July 2001

Abstract

The roles of aluminium and ferric sulphates as destabilising agents for oil–water emulsions that have been stabilised
by a non-ionic surfactant (Span 20) are investigated in terms of oil removal. The effects of coagulant dose, pH, and the
duration and intensity of both slow and fast mixing are considered. Electrokinetic measurements indicate that oil
droplets have a negative zeta potential that is weakly dependent on pH. The chosen coagulants are shown to be effective
in reducing the zeta potential of the oil droplets, and charge reversal was observed for aluminium sulphate. Oil removals
up to 99.3% at pH 8 and 99.94% at pH 7 are seen for aluminium sulphate and ferric sulphate respectively. Rapid
mixing times of around 120 s and flocculation times ranging from 15 to 20 min appear to be optimal for the DAF
separation. It is concluded that relatively low average mixing speeds for coagulation and flocculation are essential for
efficient operation. r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Emulsion; Flocculation; Flotation; Industrial wastewater; Zeta potential

1. Introduction alum and ferric chloride were used. In addition,

Dissolved air flotation (DAF) is an effective method
for removing low density particles from suspension and
for clarifying low turbidity, highly coloured waters
where light flocs are produced [1]. In addition, DAF
has been used in wastewater treatment processes for
thickening activated sludge [2]. In one of the earliest
reports of a DAF separation of oil from aqueous
dispersion, Boyd and co-workers [3] reported the results
of pilot plant studies into the separation of oil from
chemically treated wastewater, showing that effluent
with 10–15 mg l 1 oil can be produced following DAF.
Abo-Elela and Nawar [4] showed that DAF could be
effective for BOD, COD and turbidity reduction when
*Corresponding author. Tel.: +44+161-200-4368; fax: +44-
161-200-4399.
E-mail address: alec.james@umist.ac.uk (A. James).
Miskovic et al. [5] used DAF to treat oil refinery
wastewater although the effluent was still loaded with
dissolved oils. In other work, Moursy and Abo-Elela [6]
have concluded that DAF units are capable of removing
most of the oil in emulsions from refinery wastewater
that was chemically pre-treated.
Malley [7] found that the removal of dissolved organic
carbon (DOC), was independent of the separation
process but depended on the pre-treatment process.
The electrical charges on particles and bubbles are also
important in DAF [8]. It is clear that chemical pre-
treatment is an essential requirement for high efficiency
in DAF. Generally, high shear rates are necessary for
chemical dosing and lower shearing rates are needed to
promote the aggregation of fine particles through
flocculation. Takahashi et al. [9] noted that large-
diameter oil droplets in the presence of inorganic salts
are essential for the rapid separation of emulsified oils.

0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 3 - 1 3 5 4 ( 0 1 ) 0 0 3 4 7 - 5
1504 A.A. Al-Shamrani et al. / Water Research 36 (2002) 1503–1512
However, it is still not clear what the conditions
producing optimal flocculation for DAF might be.
Some workers indicated that prolonged flocculation
times are not needed in DAF as good separations can be
achieved with short residence times, typically 5–15 min
for 10–30 mm flocs [10]. In contrast, Bunker et al. [11]
suggested that DOC removal is increased with the time
allowed for flocculation, while Fukushi et al. [12]
remarked that large flocs are more desirable since the
greater number of bubbles attached to them will make
the separation more efficient. Other workers [13]
indicated that relatively high mixing intensities in the
flocculation stage produce larger flocs corresponding to
increased removal of suspended solids. However, it was
suggested that small velocity gradients in the floccula-
tion stage produce good floc size distributions for
efficient removal by DAF [14].
Although electrical charges on particles and bubbles
are important in flocculation, Ho and Ahmad [15] noted
that there are very few reports in the literature
concerning the effect of ionic additives on the stability
or electrokinetic properties of vegetable oil emulsions
stabilised by non-ionic emulsifiers. It appears that no
work has been reported on the DAF separation of
mineral oil–water emulsions stabilised by non-ionic
surfactants and flocculated by inorganic salts. To this
end the present work investigates the role of aluminium
and ferric sulphates as destabilising agents for oil
droplets stabilised by a non-ionic surfactant, and uses
the electrokinetic properties of the emulsion to interpret
the experimental results. The work also focuses on the
effect of duration and intensity of both rapid and slow
mixing prior to the DAF on separation efficiency.
2. Materials and methods
A bench scale DAF unit (Capital Controls, Oxford-
shire, UK) consisting of an oil-free air compressor; a 9 l
stainless steel unpacked saturator vessel and a 1.5 l
flotation cell was used in this study. The DAF unit has a
variable speed-controlled impeller providing rotational
speeds between 100–400 and 20–80 rpm for rapid mixing
and flocculation respectively. Chemicals used for coa-
gulation and pH control were aluminium sulphate
(Analar, BDH Chemicals Ltd, Poole, UK), ferric
sulphate, sodium hydroxide (SLR, Fisher Scientific,
UK) and hydrochloric acid (Analar, Fisher Scientific,
UK). Catenex oil 11 (Shell, UK) is a solvent refined
petroleum distillate of aliphatic hydrocarbons having a
density 815 kg m 3. Non-ionic surfactant, sorbitan
monolaurate, Span 20 having an 8.6 HLB (hydrophile-
lipophile balance) (Sigma-Aldrich Ltd, Dorset, UK),
was used as the stabilising agent at concentration of
0.3% (wt. on oil). Wastewater containing 1630 ppm oil
was prepared by mixing surfactant treated oil and
deionised water at 18000 rpm in a blender (Fisher
Scientific, UK) for 4 min. After pH adjustment, the
synthetic wastewater was introduced to the flotation cell
and mixed with the coagulant for a short time at high
shear rate, followed by slow mixing for a relatively long
time in flocculation. After flocculation, appropriate
quantities of saturated water were introduced and
flotation was allowed to proceed and samples were
collected for analysis. Oil concentrations were deter-
mined using a spectrophotometer (UV/vis/VIR, Perkin
Elmer, UK) to measure absorbency at 682 nm wave-
length. This procedure is similar to that described
elsewhere [16,17]. The zeta potential and diameter of
oil droplets were measured by a Zeta-Sizer 3000 HS
(Malvern Instruments, UK) Average values were ob-
tained from five and ten replications for zeta potentials
and droplet sizes, respectively.
3. Results and discussion
3.1. Effects of coagulant dose
The influence of aluminium sulphate and ferric
sulphate doses in the form of Al2(SO4)316H2O and
Fe2(SO4)39H2O respectively. The influence of coagulant
dosage at different pH was studied for fixed coagulation,
flocculation and flotation times. This involved flash
mixing at 200 rpm for 1 min, flocculation at 20 rpm for
15 min and flotation for 7 min using a saturation
pressure of 80 psi (551.2 kPa) and the addition of 10%
recycled water.
Fig. 1 shows the residual oil concentration at different
pH as functions of the amount of aluminium sulphate
added. The initial oil concentration was 1630 ppm. At
pH 5, 10 mg l 1 of aluminium sulphate reduces the oil
concentration to 1094 ppm. As the coagulant dose is
increased to 60 mg l 1 further reductions to 571 ppm oil
are seen. Higher doses of aluminium sulphate result in
slightly increased residual oil concentrations. At pH 6,
approximately the same trend is observed, and the
residual oil concentration decreases to the minimum of
240 ppm, when 50 mg l 1 aluminium sulphate is used. As
the dosage is increased to 60 mg l 1, a value of 400 ppm
is observed. At pH 7 a more dramatic reduction in
residual oil concentration to 82 ppm is observed when
20 mg l 1 aluminium sulphate is used. The concentration
increases to 240 ppm at 30 mg l 1 coagulant and then
falls again to 150 ppm at 50 mg l 1 coagulant. The
greatest removal of oil (99.3%) is reached at pH 8 when
100 mg l 1 aluminium sulphate was used. However,
ineffective oil removal (only 31%) is found at pH 9.
The results obtained using ferric sulphate are sum-
marised in Fig. 2. At pH 5 there is a minimum of
210 ppm in the residual oil concentration at 40 mg l 1.
After this, residual concentration increases slightly as
 A.A. Al-Shamrani et al. / Water Research 36 (2002) 1503–1512 1505

Fig. 1. The effect of aluminium sulphate on the residual oil concentration over a range of pH.

Fig. 2. The effect of ferric sulphate on the residual oil concentration over a range of pH.

the coagulant dose is increased to 80 mg l 1. At pH 6 reduction in residual oil concentration is observed at

there is a significant decrease in the residual oil pH 7 where almost 99.94% of oil is removed at a dosage
concentration, the minimum value of 107 ppm corre- of 120 mg l 1. At pH 8 the oil removal is also good and
sponds to a 20 mg l 1 dose of ferric sulphate. The worst residual oil concentration is reduced to 143 ppm using
removal is seen at pH 9, whereas the maximum 20 mg l 1 ferric sulphate.
1506 A.A. Al-Shamrani et al. / Water Research 36 (2002) 1503–1512
These results indicate the importance of pH and
coagulant concentration on the removal of oil. Measure-
ments indicate that the average oil droplet diameter in
the original emulsions is approximately 407 nm, which is
clearly in the colloidal range and suggests that for the
oil–water mixtures in this study, the performance of
DAF is related to the physico–chemical properties of the
system. The reductions in oil content of clarified water
may be sensitive to the pH and electrolyte concentration
since the colloidal stability of the emulsion is a function
of these quantities. In the presence of inorganic
electrolytes such as those used in the present study, the
stability of colloidal dispersions depends on the interac-
tion energy between particles or droplets [18]. Particles
are generally destabilised if they reach a separation
corresponding to either the primary minimum, where
the aggregation is assumed to be irreversible, or the
secondary minimum, where aggregation is assumed to
be rather weak and reversible. The latter state, reversible
flocculation in the secondary minimum, may be ob-
served in oil–water emulsions stabilised by a non-ionic
surfactant [19]. Some of the results shown in Figs. 1 and
2 suggest two destabilisation regions depending on the
concentration of the coagulating agent and the pH,
which may correspond to coagulation and flocculation.
The aggregates of oil droplets resulted from flocculation
might be separated by a thin film of the continuous
phase, and the shear forces resulting from injection of
the recycle water to the flotation cell might break down
these aggregates leading to poor oil separations. On the
other hand, it is possible that coalescence of oil droplets
Fig. 3. Residual oil concentration and oil droplet zeta potential as function of aluminium sulphate dosage at pH 8.
might occur when the oil droplets are brought to the
primary minimum where strong aggregates are supposed
to be formed and the performance of the DAF
separation is improved.
Both aluminium and ferric sulphates proved to be
effective destabilising agents in the present study,
facilitating the removal of most of the oil. As observed,
ferric is slightly more efficient than aluminium sulphate
in terms of oil removal. This finding is in agreement with
that reported by Vilage et al. [20]. However, the
comparative effectiveness cannot be explained in terms
of highly charged Al3+ and Fe3+ ions and the Schulze–
Hardy rule, since this is now known to be greatly
oversimplified. Because of hydrolysis, simple Al3+ and
Fe3+ ions do not exist in aqueous solution around
neutral pH. Instead, a range of hydrolysis products
exists and these are responsible for the destabilising
effects on colloidal impurities [21]. It was reported that
alum floc particles are more hydrophilic than other
coagulant floc particles [11]. This could make the
attachment between air bubbles, which are strongly
hydrophobic, and droplets flocculated by aluminium
sulphate less effective, resulting in slightly poorer
separations in comparison to those flocculated by ferric
sulphate.
The effect of aluminium sulphate on zeta potential of
oil droplet was investigated at pH 8 where the best oil
removal was seen. Results, presented in Fig. 3, indicate
that the zeta potential of oil droplets increases as the
concentration of aluminium sulphate increases. Charge
reversal is observed at a dose of about 50–100 mg l 1.
 A.A. Al-Shamrani et al. / Water Research 36 (2002) 1503–1512 1507

This corresponds to the optimal dosage for aluminium
sulphate and reflects the role of charge neutralisation in
the destabilisation of oil droplets. The results shown in
Fig. 3 suggest that further increase in coagulant dosage
appears to be unnecessary. Similar measurements were
conducted for ferric sulphate at pH 7 where the best oil
removal was obtained for this coagulant (see Fig. 4). A
similar trend to the aluminium sulphate results was
observed the main difference being that the zeta
potential increased as the dose of ferric sulphate rose
to 160 mg l 1. Although charge reversal was not seen,
ferric sulphate was still more effective in oil removal
than aluminium sulphate. This could imply that simple
DLVO theory does not fully explain the mechanism
behind the separation of destabilised oil droplets by
DAF. It is possible that hydrophobic interactions and
properties of the flocs might also play an important role.
3.2. The effect of pH
The zeta potential and average size of the dispersed oil
droplets at initial pH 3.6 are 6074.0 mV and 407 nm
respectively. If the oil droplets are assumed to be
spherical then the high zeta potential can be expected
to produce sufficient electrostatic repulsion between oil
droplets to maintain the stability of the system. Fig. 5

Fig. 4. Residual oil concentration and oil droplet zeta potential as function of ferric sulphate dosage at pH 7.

Fig. 5. Zeta potential of stabilised oil droplets before flocculation as a function of pH.
1508 A.A. Al-Shamrani et al. / Water Research 36 (2002) 1503–1512
shows the zeta potential of droplets in the oil–water
emulsion as a function of pH in the presence of 1 mM
NaCl. These results show that the Span 20-stabilised oil
droplets are negatively charged over the pH range of
interest to the present study and are consistent with
earlier findings [22,23]. The magnitude of the negative
zeta potential increases gradually with pH to reach a
maximum (z= 106 mV) at pH 9. The negative charges
are thought to result from adsorption of hydroxyl ions
at the oil–water interface [22]. These results support the
assertion that the oil droplets are stabilised by the non-
ionic surfactant over the pH range under investigation.
Before considering the effects of pH on DAF
performance it is worth remembering that the aqueous
chemistry of aluminium and iron in water can be
explained by considering numbers of monomers, several
polymeric species and solid precipitates. In particular,
the nature of coagulant species formed is influenced by
the amount of coagulant added, the solution pH and
ionic strength together with the concentration of organic
compounds in water to be treated [24]. The inability to
control the polymerisation of such coagulant species is
considered a major drawback when using aluminium
and ferric sulphates [21]. It is suggested that the
coagulants species can be identified from the various
pC–pH diagrams presented elsewhere [25,26]. However,
further work, which is beyond the scope of the present
investigation, is required to determine the exact char-
acteristics of the various hydrolysis species. This might
involve procedures for identifying the coagulant species
described elsewhere [27,28].
Figs. 6 and 7 show the effects of solution pH on the oil
removal for aluminium and ferric sulphates respectively.
Fig. 6. The effect of solution pH on the residual oil concentration for various doses of aluminium sulphate.
Similar removal patterns are seen for each coagulant.
The maximum reductions in the oil content occur at
almost the same pH for both aluminium and ferric
sulphate (pH 8 and 7 respectively). This is expected
because the maximum neutralisation of the oil droplets’
surface charge by hydrolysed aluminium and ferric
cations occurs in this range of pH. Consequently a
significant amount of de-emulsification is achieved and it
can be concluded that, for the present system, the best
destabilisation conditions are found in the neutral pH
range.
The present results suggest that, since changes in pH
affect the residual oil concentration more than changes
in the coagulant dose, the optimum conditions for the
best clarification are more sensitive to pH than
coagulant dose. These findings differ from those of
Kamel and co-workers, [19] but are in agreement with
those of Reay and Ratclif [29].
3.3. Influence of the duration and intensity of mixing
While there have been many investigations into the
effect of duration and intensity of slow mixing to
promote flocculation, few studies have been made into
effects arising from the duration and intensity of rapid
mixing [30]. Furthermore it appears that no studies have
been made on the influence of mixing intensity on the
coagulation and flocculation of oil droplets. This is
perhaps because most investigations have been carried
out on municipal wastewater and surface water. Some of
the effects of duration and intensity of both rapid and
slow mixing are now investigated under optimal removal
 A.A. Al-Shamrani et al. / Water Research 36 (2002) 1503–1512
Fig. 7. The effect of solution pH on the residual oil concentration for various doses of ferric sulphate.
conditions, namely a coagulant dose of 100 mg l 1
aluminium sulphate at pH 8.
The residual oil concentration was investigated as a
function of the average rapid mixing velocity gradient
over the range 60–250 s 1. Values of the average velocity
gradient as a function of stirring speed are available
from the manufacture of the flotation cell and stirrer.
The best oil removal was found using a velocity gradient
of 110 s 1. This is equivalent to a stirrer rotation rate of
200 rpm. For this amount of rapid mixing the concen-
tration of the residual oil is reduced from 1630 to 18 ppm
using a flocculation time of 15 min. The results are worse
when the velocity gradient is increased. This suggests
that in oil separation the average velocity gradients in
rapid stirring should be fairly small to ensure maximum
efficiency. The effect of varying the rapid mixing time is
shown in Fig. 8. The mixing time was varied from 30 to
150 s for each shearing rate. It is seen that increasing
rapid mixing times of up to 120 s improves the DAF
performance at almost every mixing speed.
On the other hand, regarding the slow mixing velocity
gradient, it has been reported that a relatively high
velocity gradient is required for dissolved air flotation
during the flocculation stage [13]. The results found in
this study are shown in Fig. 9. It is clear that increasing
the slow mixing velocity gradient above 12 s 1 reduces
DAF performance. This suggests that low average shear
rates are essential for efficient flotation as it is possible
that higher velocity gradients may re-disperse the
flocculated droplets. This observation has also been
made recently in surface water clarification studies by
other workers [14].
1509
Finally, the effects of changing the duration of the
flocculation stage were investigated and average velocity
gradients were set to 60 and 12 s 1 for the coagulation
and flocculation stages respectively. It was found that
the flocculation times ranging from 15 to 20 min are
optimal for good DAF performance where high separa-
tion is required as shown in Fig. 10.
4. Conclusions
The conclusions from this work are:
* Catenex oil in water emulsions stabilised by Span 20
provided a suitable model system for the current
investigation. The zeta potential measurements
clearly indicated that the oil droplets are negatively
charged, and the magnitude of zeta potential is
sufficiently large to stabilise the oil–water emulsions
electrostatically. The zeta potential also showed some
pH dependence, implying the adsorption of hydroxyl
ion on the surface of the oil droplets.
* The addition of highly charged cations in the form of
aluminium or ferric salts has effectively induced the
destabilisation of the emulsions, leading to the
significant oil separations (removal efficiency
>99%). Both coagulants have also effectively
reduced the zeta potentials of oil droplets. Charge
reversal was observed in aluminium sulphate. How-
ever, ferric sulphate was slightly more effective but
did not reverse the surface charge, indicating some
1510 A.A. Al-Shamrani et al. / Water Research 36 (2002) 1503–1512

Fig. 8. The effect of rapid mixing time on the residual oil concentration for a range of average velocity gradient in the DAF cell.

Fig. 9. Residual oil concentration and turbidity as a function of the average slow mixing velocity gradient in DAF cell.

other factors apart from charge neutralisation
influence the separation process.
* The stability of the emulsions is more sensitive to pH
than the coagulant concentration. This may be linked
to the optimum conditions for hydrolysis of the
aluminium and ferric ions. The best pH conditions
for destabilisation of the oil–water emulsion were
found in the neutral pH range.
* The best oil removal was found using a velocity
gradient of 110 s 1 for rapid mixing. Prolonging the
 A.A. Al-Shamrani et al. / Water Research 36 (2002) 1503–1512
Fig. 10. Residual oil concentration and turbidity as function of the time taken for flocculation before DAF.
rapid mixing time (up to 120 s) also improved the
separation performance of the DAF process when
coagulants were used. However, low velocity gradi-
ents are required for slow mixing as this improves
flocculation. Flocculation times ranging from 15 to
20 min are optimal for good DAF performance.
References
[1] Gregory R, Zabel T. In: Pontius FW, Water quality and
treatment, 4th ed. New York: McGraw-Hill, p. 367–453,
1990.
[2] Offringa G. Dissolved air flotation in Southern Africa.
Water Sci Technol 1995;31(3–4):159–72.
[3] Boyd JL, Shell GL, Dahlstrom DA. Treatment of oily
wastewater to meet regulatory standards. AIChE Con-
ference, Toledo, OH, 1971.
[4] Abo-Elela SI, Nawar SS. Treatment of wastewater from an
oil and soap factory via dissolved air flotation. Environ Int
1980;4:47–52.
[5] Miskovic D, Dalmacija B, Zivanov Z, Karlovic E, Hain Z,
Maric S. An Investigation of the Treatment and Recycling
of Oil Refinery Wastewater. Water Sci Technol
1986;18:105–14.
[6] Moursy AS, Abo-Elela SI. Treatment of oily refinery
wastes using a dissolved air flotation. Environ Int
1982;7:267–70.
[7] Malley Jr JP. Removal of organic halide precursors by
dissolved air flotation vs. conventional water treatment.
Environ Technol 1991;11:1161–8.
[8] Leppinen DM. A kinetic model of dissolved air flotation
including the effects of interparticle forces. J Water Suppl:
Res Technol-Aqua 2000;49(5):258–9.
1511
[9] Takahashi T, Miyahara T, Nishizaki Y. Separation of oily
water by bubble column. J Chem Eng Jpn 1979;12(5):394–9.
[10] Edzwald JK, Walsh JP, Kaminski GS, Dunn HJ.
Flocculation and air requirements for dissolved air
flotation. J Am Water Works Assoc 1992;84(3):92–100.
[11] Bunker Jr DQ, Edzwald JK, Dahlquist J, Gillberg L.
Pretreatment consideration for dissolved air flotation:
water type, coagulants and flocculation. Water Sci Technol
1995;31(3–4):63–71.
[12] Fukushi K, Matsui Y, Tambo N. Dissolved air flotation:
experiments and kinetic analysis. J Water Suppl: Res
Technol-Aqua 1998;47(2):76–86.
[13] Klute R, Langer S, Pfeifer R. Optimisation of coagulation
processes prior to DAF. Water Sci Technol 1995;31:59–62.
[14] Vlaski A, Van Breemen AN, Alaerts GJ. Optimisation of
coagulation conditions for the removal of cyanobacteria
by dissolved air flotation or sedimentation. J Water Suppl:
Res Technol-Aqua 1996;45(5):53–261.
[15] Ho CC, Ahmad K. Electrokinetic behaviour of palm oil
emulsion in dilute electrolyte solutions. J Colloid Interface
Sci 1999;216:25–33.
[16] Gaysoss RE. Experiments on oil–water separation by
induced air flotation. M.Sc. dissertation, Imperial College,
London, 1981.
[17] Nikolov AD, Randie M, Shetty CS, Wasan DT. Chemical
demulsification of oil in water emulsion using air flotation:
the importance of film thickness stability. Chem Eng
Comm 1996;152–153:337–50.
[18] Hunter RJ. Zeta potential in colloid science principles and
applications, 2nd ed. London: Academic Press, 1988.
[19] Kamel A, Sabet V, Sadek H, Srivasava SN. The role of
non-ionic surfactant in emulsion stability. Prog Colloid
Polym Sci 1978;63:33–40.
[20] Vilage A, Rose J, Masion A, Laine F, Bottero F.
Advanced coagulation of natural organic matter mechan-
1512 A.A. Al-Shamrani et al. / Water Research 36 (2002) 1503–1512
ism of interaction between NOM and Al and Fe.
Proceedings of the Annual Conference of the American
Water Works Association, Atlanta, 1997. p. 403–13.
[21] Jiang J, Graham NJD. Observations of the comparative
hydrolysis/precipitation behaviour of polyferric sulphate
(PFS) and ferric sulphate (FS). Water Res 1997;32(3):
930–5.
[22] Marinova KG, Alargova RG, Denkov ND, Velev ND,
Petsev DN, Ivanov IB, Borwankar RP. Charging of oil
water interface due to spontaneous adsorption of hydroxyl
ions. Langmuir 1996;12:2045–51.
[23] Sato Y, Murakami Y, Hirose T, Yamamoto H, Uryu Y.
Removal of emulsified oil particles by dissolved air
flotation. J Chem Eng Jpn 1979;12(6):454–9.
[24] Stephenson RJ, Duff SJB. Coagulation and precipitation
of a mechanical pulping effluent-I. Removal of carbon,
colour and turbidity. Water Res 1996;30(4):781–92.
[25] Dentel SK, Gossett JM. Mechanisms of coagulation with
aluminium salts. J Am Water Works Assoc 1988;80:187–98.
[26] Johnson PN, Amirtharajah A. Ferric chloride and alum as
single and dual coagulants. J Am Water Works Assoc
1983;75:232–9.
[27] Murphy PJ, Ponser AM, Quirk JP. Chemistry of iron in
soil. Aust J Soil Res 1975;13:189–201.
[28] Smith R. Relations among equilibrium and nonequili-
brium aqueous species of aluminium hydroxide complexes.
Adv Chem Ser 1971;106:250–79.
[29] Reay D, Ratclif GA. Removal of fine particles from water
by dispersed air flotation: effects of bubble size and particle
size on collection efficiency. Can J Chem Eng 1973;51:
178–85.
[30] Rossini M, Garcia J, Galluzzo M. Optimisation of the
coagulation–flocculation treatment: influence of rapid mix
parameters. Water Res 1999;33(8):1817–26.