Wat. Res. Vol. 35, No. 15, pp.

3665–3669, 2001
# 2001 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
PII: S0043-1354(01)00083-5 0043-1354/01/$ - see front matter

PURIFICATION OF OILY WASTEWATER BY

HYBRID UF/MD
M. GRYTA*, K. KARAKULSKI and A. W. MORAWSKI
Department of Water Technology and Environmental Engineering, Institute of Inorganic Chemical
Technology and Environmental Engineering, Technical University of Szczecin, ul. Puľaskiego 10,
70-322 Szczecin, Poland

(First received 3 August 2000; accepted in revised form 31 January 2001)

Abstract}Investigations on the treatment of oily wastewater by a combination of ultrafiltration (UF) and

membrane distillation (MD) as a final purification method have been performed. A tubular UF module
equipped with polyvinylidene fluoride (PVDF) membranes and a capillary MD module with
polypropylene membranes were tested using a typical bilge water collected from a harbour without
pretreatment. The permeate obtained from the UF process generally contains less than 5 ppm of oil. A
further purification of the UF permeate by membrane distillation results in a complete removal of oil from
wastewater and a very high reduction of the total organic carbon (99.5%) and total dissolved solids
(99.9%). # 2001 Elsevier Science Ltd. All rights reserved

Key words}ultrafiltration, tubular membrane, membrane distillation, oily wastewater, bilge water

NOMENCLATURE complex composition because it may contain oil

c concentration, wt% (mineral, vegetable or synthetic), fatty acids, emulsi-
D distillate fiers, corrosion inhibitors, bactericides and other
d day chemicals. The common industrial practice of col-
F feed lecting sewage from various sources in one storage
MD membrane distillation
m mass flux of liquid, kg s
1 tank usually multiplies these problems.
N permeate flux, kg m
2 s
1 The traditional methods for the separation of oil
in inlet value emulsion can be classified as chemical, mechanical
DP pressure difference, Pa and thermal (Hupka and Miller, 1999). The chemical
RO reverse osmosis
T temperature, K
methods are primarily based on the neutralisation of
TDS total dissolved solids, mg dm
3 detergents (emulsion’s stabilisers), and a change of
TOC total organic carbon, mg C dm
3 pH solution. This results in the acceleration of a
v flow velocity, m s
1 separation process due to the coalescence effect. The
UF ultrafiltration
water phase from chemical treatment contains
residual oil, therefore, it should be subjected to
further purification in order to meet today’s effluent
standard for the sewage systems. The mechanical
INTRODUCTION
methods of oily wastewater treatment are based on
Oily wastewater generated by various industries and the phenomenon of gravitational emulsion breaking.
subsequently discharged into the natural environ- The efficiency of this method is mainly dependent on
ment creates a major ecological problem throughout the dispersion degree of oil (size of oil droplets) as
the world. This wastewater occurs most frequently in well as the difference in density between water and
a form of oil-in-water emulsion (Hupka and Miller, oil. The emulsion breaking may be achieved by
1999; Karakulski et al., 1995). A particularly stable heating the wastewater, which results in a significant
emulsion is generated by the metal-finishing or metal- increase of the size of oil droplets. The thermal
working industry during mechanical operations such process of oily wastewater treatment requires a large
as grinding, rolling, alkaline degreasing and trans- amount of energy, therefore this process is not cost-
portation. Oily wastewater is characterised by a effective. One of the recently utilised solutions in the
treatment of oily wastewater is the biological
*Author to whom all correspondence should be addressed.
method. However, this method possesses several
Fax: +48-91-43-30-352; e-mail: margryta@mailbox. disadvantages such as a low efficiency, operational
tuniv.szczecin.pl difficulties, and high operation costs.

3665
3666 M. Gryta et al.
The traditional methods of oily wastewater treat-
ment described above are frequently not efficient
enough, especially when the oil droplets are finely
dispersed and their concentration is very low. The
treatment process is additionally complicated due to
different compositions of this wastewater. Therefore,
there is still lack of an universal methods for oil
removal from such wastewaters which will be efficient
in all these cases.
The membrane techniques such as ultrafiltration,
or a combined UF/RO system seem to be a
promising method for the separation of emulsion
(Jönsson and Trägârdh, 1990; Karakulski et al.,
1995, 1998; Ueyama et al., 1987). The advantage of
the membrane methods is a high efficiency of the
separation in conjunction with a high quality of the
permeate, not achievable by other methods, as well as
low operating costs.
Ultrafiltration is a pressure driven membrane
process which can concentrate and fractionate
macromolecular solutes and separate suspended
species from water. UF provides a non-destructive
separation which can be performed without a phase
transition. This process is operated at relatively low
pressure (0.7–7 bar) because the osmotic pressure
exerted by the high molecular weight solutes is
negligible, and the membranes are designed to
separate such solutes (Bodzek and Konieczny, 1992;
Karakulski et al., 1995). The concentrate may
contain up to 50% of oil. The oily concentrate
can be further separated by centrifugation. The
recovered oil is not recyclable, but it can be burned
to eliminate its pollution impact. Membranes made
from various polymers and inorganic materials are
most frequently used in the UF process. The
application of a given membrane type requires
detailed pilot plant studies in order to determine
the optimal process parameters.
A permeate with residual oil is produced during
ultrafiltration of oily wastewater. The combination of
UF with reverse osmosis permits to obtain a
permeate which does not contain petroleum deriva-
tives (Karakulski et al., 1995). In the work presented
here, the purification of UF permeate by the
membrane distillation has been elaborated.
Fig. 1. A schematic diagram of a hybrid UF/MD system for treatment of oily wastewater.
Membrane distillation is an evaporation process
through the pores of a hydrophobic membrane
(Tomaszewska et al., 1995). The driving force in the
MD process is generated by the pressure difference of
vapour due to a difference in temperature of the
solution on both sides of the hydrophobic mem-
brane. The presence of only a vapour phase in the
membrane pores is a necessary condition for
membrane distillation. The separation mechanism
of the MD process is based on the vapour/liquid
equilibrium of the liquid mixture. During MD of
solutions with non-volatile solutes, only water
vapour is transferred through the membrane, there-
fore, the obtained distillate comprises demineralised
water. The important applications of membrane
distillation can be seen in water treatment, sea water
desalination, the production of high purity water and
the concentration of aqueous solutions (Tomaszews-
ka et al., 1995, 1998). The application of MD for
purification of oily wastewater was also reported
(Gryta and Karakulski, 1999).
Hybrid ultrafiltration-membrane distillation system
The complete removal of oil from wastewater, e.g.
bilge water originating from an engine-room, may be
achieved by the application of ultrafiltration and
further purification of the UF permeate (with a trace
amount of oil) by membrane distillation. Figure 1
shows a schematic layout of an integrated UF/MD
membrane process.
In the process presented, the UF permeate is
heated in a heat exchanger prior to the MD
installation. The hot feed flows through the MD
modules connected in a parallel mode. The evapora-
tion of water through the membrane causes an
increase of the concentration of the oil phase in the
feed. The concentrated emulsion is returned to the
feed of the ultrafiltration plant. Pure water (MD
distillate) and an oil concentrate (UF retentate) are
produced in the hybrid system.
EXPERIMENTAL
Bilge water collected from the Szczecin harbour (Poland)
was used for the UF studies without pretreatment. The
 Oily wastewater
content of organic compounds (primarily oil) in the feed
and permeate was determined as total organic carbon
(TOC) using a TOC-Analyser multi N/C (Analytic Jena).
Additionally, the oil content was measured in the samples
using an OCMA 220 Horiba analyser. The oil content in the
bilge water was in the range of 124–360 ppm, as determined
by a Horiba analyser.
The electrical conductivity and the total dissolved solids
(TDS) of the examined solutions were measured by a 6P
Ultrameter (Myron L Company).
Ultrafiltration experiments were carried out with tubular
membranes made from PVDF (type FP 100, PCI Membrane
System). The membrane with an internal diameter of
12.5 mm and a length of 1.2 m were mounted in the B1
module (membrane area 0.9 m2). The retention character-
istic of the PVDF membrane was determined with dextran
solutions, and expressed as a molecular weight cut off
(MWCO), which amounted to 100 kDa. The detailed
procedure for determination of the separation and transport
properties of used membranes and the UF experiment
performed on a pilot plant are described elsewhere
(Karakulski et al., 1995, 1998).
The performance of the membrane distillation was
studied using a 1 wt% solution of NaCl or the UF permeate
with an oil content within 3–10 ppm as the feed.
The experimental set-up used in the MD investigations is
presented in Fig. 2. A capillary module made from
polypropylene membranes (dout =din ¼2.6/1.8 mm) was used
in this installation. These membranes had a pore size with a
nominal and maximum diameter of 0.2 and 0.6 mm,
respectively, and a porosity of about 73%. The membrane
working area amounted to 220 cm2. Within the closed
system, the feed and distillate stream was flowing through
the module in a counter-current mode. In the MD
experiments the feed passed inside the capillaries, whereas
the distillate through the intertubular space. The distillate
temperature at the module inlet amounted to 293 K, and
was kept constant during the MD process. The inlet
temperature of the feed was varied in the range of
313–353 K. The MD experiments with the effluent from
the UF pilot plant (permeate) containing 3.2 ppm of oil were
conducted at various inlet temperatures of the feed. During
the MD experiments, the total pressure on the distillate side
was 1000 Pa higher than that of the feed side.
RESULTS AND DISCUSSION
Treatment of bilge water was performed in two
stages using a hybrid UF/MD system. Wastewater
was treated by an ultrafiltration process and the
resulting UF permeate was further purified by means
of membrane distillation.
Fig. 2. The experimental set-up of the MD process utilised
for the purification of UF permeate: 1}MD module,
2}feed reservoir with stirrer, 3}distillate reservoir, 4 and
5}control system for the feed and distillate temperature,
6}pump, 7}manometer.
3667
Fig. 3. Flux during UF treatment of bilge water.
Table 1. Reduction of pollutant parameters obtained by UF
treatment of bilge water
Type of solution Oil content TOC TDS
(Horiba)
(mg dm
3) (mg C dm )
3
(mg dm
3)
Feed 1 124 138 3707
Permeate 1 2.4 6.7 3570
Feed 2 360 400 3790
Permeate 2 4.9 8.6 3700
The UF study was carried out at a constant feed
concentration, achieved by recycling of the permeate
to the feed tank. The changes of the permeate flux
during 50 h of running the ultrafiltration process with
bilge water containing 124 or 360 ppm of oil are
shown in Fig. 3. A decrease of the permeate flux was
observed during the operation of the UF plant. The
initial value of the flux for wastewater with an oil
content of 124 ppm was 1300 kg m
2 d
1, and a
decrease to 875 kg m
2 d
1 after 50 h was observed.
In the case of wastewater with a higher content of oil
a decrease of the flux from 768 to 528 kg m
2 d
1 was
found. This variation of permeate flux during the
treatment of oil wastewater is typical for an UF
process, and results from both fouling and concen-
tration polarisation (Jönsson and Trägârdh, 1990).
Each UF cycle (50 h) was completed by rinsing the
UF plant with permeate followed by cleaning with a
1 wt% solution of detergent (IMPUREX) and final
rinse of the installation with tap water. The UF
membrane cleaning was carried out for 1 h. The
effectiveness of this procedure was evaluated by the
measurement of the water flux under specified
conditions and comparison with the initial water flux
and amounted to 98%.
The quality of the UF permeate was practically
constant for the wastewater studied. The average
values of the permeate composition are presented in
Table 1. The amount of TOC determined in the
samples was higher than the oil content determined
with a Horiba OCMA 220 instrument. The experi-
mental results indicate that the wastewater subjected
3668 M. Gryta et al.

Table 2. A comparison of permeate quality after UF and UF/MD system utilised for purification of bilge water

Process/feed Oil content (Horiba) TOC TDS Reduction (%)

(mg dm
3) (mg C dm
3) (mg dm
3)

Oil TOC TDS

Bilge water 360 400 3790 } } }

UF 4.9 8.6 3700 98.64 97.85 2.4
UF/MD 0 1.8 1.4 100 99.55 99.96

Fig. 4. Variation of the flux and the quality of permeate as a

function of feed concentration. Fig. 5. The influence of the flow rate, the feed temperature
and the oil content on the MD permeate flux.

to the treatment contains several other organic

constituents besides oil. Hence, the application of
both measurement methods allows the accurate
determination of the content of organic compounds
in wastewater.
In the second part of studies, the bilge water was
continuously concentrated in the UF process from
200 to 1400 mg C dm
3 (TOC). The permeate flux as
a function of the feed concentration is shown in
Fig. 4. The initial value of the flux, equal to
1600 kg m
2 d
1, gradually decreases with the in-
crease of the feed concentration. The flux decreased
to 1150 kg m
2 d
1 after a seven-fold concentration
of the feed. The oil content in the permeate was lower Fig. 6. The influence of the process conditions of the MD
than 12 mg C dm
3 in spite of a significant increase of and the type of feed on the permeate flux.
the oil concentration in the feed. The retention
coefficient determined for the TOC was very high,
and varied within the range of 99.5–99.8%.
The effluent from the UF process was further influence of the temperature polarisation (Tomas-
purified by the membrane distillation. zewska et al., 1995). The maximum permeate flux of
The MD experiments were initially performed with 250 kg m
2 d
1 was obtained for a feed temperature
a reference solution of 1% NaCl in order to of 353 K and a flow rate of 0.0067 kg s
1.
determine the pure water flux prior to the experi- A slightly lower permeate flux was obtained
ments with UF permeate containing a trace amount (broken line in Fig. 5) for an UF permeate with
of oil. The experimental results presented in Fig. 5 3.2 ppm of oil. The oil rejected by the MD process is
show that the permeate flux increases along with an accumulated in the layer adjacent to the membrane,
increase of the feed temperature and the flow rate of therefore the water evaporation was impeded (Gryta
the solutions in the module. The observed variation and Karakulski, 1999). This phenomenon becomes
of the permeate flux is a consequence of an enhanced more severe with an increasing concentration of oil in
mass driving force as a result of the higher the feed (Fig. 6). Application of a relatively high feed
temperature and mass flow and a less significant temperature, e.g. 353 K, allows to limit (to a certain
 Oily wastewater
degree) a decline of the permeate flux. An increase of
the oil content in the feed has practically no influence
on the quality of the permeate. The distilled water
produced in the MD process exhibited an electrical
conductivity within the range of 1.5–2.5 mS cm
1, and
contained 1.5–1.8 mg C dm
3. The final quality of the
permeate obtained in the hybrid UF/MD system and
the reduction of the pollution parameters for both
processes are compiled in Table 2.
The results reported in Gryta and Karakulski
(1999) indicate that no water evaporation through
the membrane occurs for wastewater with an oil
content exceeding 2000 ppm. The permeate flux at a
level of 120 kg m
2 d
1 was obtained for a feed with
1000 ppm of oil at a feed temperature 353 K.
However, a drop of the permeate flux to
48 kg m
2 d
1 with an increase of the oil content to
2000 ppm was observed. The above problem may be
solved by a combination of an UF plant with an MD
system. Recycling of MD retentate to the UF plant
permits to maintain a low concentration of oil in the
MD feed, thereby a higher efficiency of the mem-
brane distillation process can be achieved.
CONCLUSIONS
The application of the UF process for the
treatment of bilge water allows to substantially
reduce the content of oil to less than 5 ppm. The
TOC measurements indicate that the level of organic
pollutants was slightly higher with maximum value of
8.6 mg C dm
3 was achieved. The utilisation of a
membrane distillation as a supplementary second
stage of treatment permitted the complete removal of
oil pollutants as well as the remaining soluble
substances in the feed. The produced MD permeate
was characterised by an electrical conductivity below
3669
3 mS cm
1 and a very low value of total organic
carbon}1.8 mg C dm
3. The experimental results
confirm the effectiveness of an UF/MD system for
the purification of bilge water.
Acknowledgements}This work was supported by the State
Committee for Scientific Research, Poland.
REFERENCES
Bodzek M. and Konieczny K. (1992) The use of ultrafiltra-
tion membranes made of various polymers in the
treatment of oil emulsion wastewaters. Waste Manage.
12, 75–84.
Gryta M. and Karakulski K. (1999) The application of
membrane distillation for the concentration of oil–water
emulsions. Desalination 121, 23–29.
Hupka J. and Miller J. D. (eds) (1999) Environmental
technology for oil pollution. In Proceedings of the Second
International Conference on Analysis and Utilization of
Oily Wastes AUZO’99, August 29–September 3, Gdańsk,
Poland.
Jönsson A. S. and Trägârdh G. (1990) Ultrafiltration
applications. Desalination 77, 135–179.
Karakulski K., Kozľowski A. and Morawski A. W. (1995)
Purification of oily wastewater by ultrafiltration.
Sep. Technol. 5, 197–205.
Karakulski K., Morawski A. W. and Grzechulska J. (1998)
Purification of bilge water by hybrid ultrafiltration and
photocatalytic processes. Sep. Purif. Technol. 14,
163–173.
Tomaszewska M., Gryta M. and Morawski A. W. (1995)
Study of the concentration of acids by membrane
distillation. J. Membr. Sci. 102, 113–122.
Tomaszewska M., Gryta M. and Morawski A. W. (1998)
The influence of salt in solutions on hydrochloric acid
recovery by membrane distillation. Sep. Purif. Technol.
14, 183–188.
Ueyama K., Fukuura K. and Furusaki S. (1987) Oil-phase
permeation behaviour of o/w emulsion through a porous
polytetrafluoroethylene membrane. J. Chem. Eng. Jpn.
20, 618–622.