From waste treatment to resource efciency in the chemical industry: recovery

of organic solvents from waters containing electrolytes by pervaporation

Vernica Garca
a,1
, Eva Pongrcz
a, b,
*
, Paul S. Phillips
c
, Riitta L. Keiski
a
a
University of Oulu, Department of Process and Environmental Engineering, Mass and Heat Transfer Process Laboratory, P.O. Box 4300, FI-90014, Finland
b
University of Oulu, Thule Institute, NorTech Oulu, P.O. Box 7300, FI-90014, Finland
c
University of Northampton, School of Science and Technology, Northampton NN2 6JD, UK
a r t i c l e i n f o
Article history:
Received 30 November 2011
Received in revised form
25 July 2012
Accepted 18 August 2012
Available online 31 August 2012
Keywords:
Chemical industry
Pervaporation
Resource efciency
Recovery
n-Butanol
Dichloromethane
a b s t r a c t
Organic solvents in chemical industry wastewaters are not only a pollution and health problem but also
a sign of inefcient use of resources. This paper reports a methodology based on pervaporation, using
rst hydrophobic then hydrophilic membranes, for the separation of n-butanol and dichloromethane
from sodium chloride containing aqueous mixtures into water, brine and a concentrate of organic
solvents. The aim of this research was to study the efciency of two hydrophobic and one hydrophilic
membranes for the recovery of the organic solvents.
The separation of organic solvents from aqueous mixtures was conducted using CMX-GF-010-D (Celfa
AG, Switzerland) and PERTHESE

500-1 (Perouse Plastic, France) membranes, and was promoted by high

temperatures. The addition of sodium chloride did not affect the permeation of the organic solvents
through the CMX-GF-010-D membrane while the salting out phenomena was observed when the
PERTHESE

500-1 membrane was used.

The dehydration was conducted using a polyvinyl alcohol and titanium dioxide membrane on
a support of polyacrylonitrile and polyphenylene sulde (supplied by Helmholtz-Zentrum Geesthacht
Zentrum fr Material und Kstenforschung GmbH, Germany). The membrane showed impermeable
features towards dichloromethane in the studied conditions. At the optimal dehydration conditions the
concentration of water was 97.6 0.1 %wt and the n-butanol content was 2.42 0.13 %wt.
Finally the article calls for a paradigm change, to view organic solvents in wastewaters as resources to
be recovered, rather than pollutants to be treated as such.
2012 Elsevier Ltd. All rights reserved.
1. Introduction
Chemical industry uses a remarkable quantity and variety of
organic solvents or Volatile Organic Compounds (VOCs) in the
manufacture of a wide range of products e.g. adhesives, fuels,
pharmaceuticals and cosmetics. The release of organic solvents to
aquatic media is of particular concern as they may be found in
drinking water supplies and affect the human health (Zogorski
et al., 2006). Further, the presence of organic solvents in chemical
efuents is a sign of inefcient use of resources, implying long-
lasting environmental problems, depreciation of natural capital
and valuable materials leaking out of economic systems.
The prime objective of European environmental policy is
pollution prevention. The general principle of Directive 2008/1/EC
on Integrated Pollution Prevention and Control (IPPC Directive) is
taking preventative measures through the application of best
available technologies (BAT), which includes furthering the
recovery and recycling of wastes. The IPPC Directives also supports
the waste management hierarchy (WMH) set out by the Waste
Framework Directive (WFD) (2008/98/EC). In its Article 4, the WFD
denes the waste management hierarchy as waste prevention,
preparing for re-use, recycling, other recovery (such as energy
recovery) and disposal. Waste prevention indicates measures to be
taken before the substance becomes waste and preparing for re-use
indicates operations by which components that have become waste
are prepared to be used again for the same purpose, without
reprocessing. Recycling means reprocessing the waste material or
component to be used for the same or other purpose. Recovery is
Abbreviations: PAN, Polyacrylonitrile; PVA, Polyvinyl alcohol; PPS, Poly-
phenylene sulde.
* Corresponding author. University of Oulu, Thule Institute, NorTech Oulu, P.O.
Box 7300, FI-90014, Finland. Tel.: 358 40 506 1623; fax: 358 8 553 3564.
E-mail address: eva.pongracz@oulu. (E. Pongrcz).
1
Present address: Department of Chemical Engineering and Inorganic Chemistry,
ETSII y T, University of Cantabria, Avda. de los Castros s/n, 39005 Santander, Can-
tabria, Spain.
Contents lists available at SciVerse ScienceDirect
Journal of Cleaner Production
j ournal homepage: www. el sevi er. com/ l ocat e/ j cl epro
0959-6526/$ e see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jclepro.2012.08.020
Journal of Cleaner Production 39 (2013) 146e153
a collective term meaning any operation the principal result of
which is waste serving a useful purpose. (2008/98/EC)
Inthecaseof wastewaters containingorganic solvents, thesuccess
of waste avoidance measures such as product re-design, substitution
of input materials and process changes may not be viable in some
chemical industry branches. For example, in the ne chemicals
industry, the waste-to product ratio can be up to 40; therefore, even
with drastic waste prevention measures, a signicant amount of
waste will still be produced (Garca et al., 2008). In many cases, VOCs
in wastewaters are disposed of by thermal destruction, after being
extracted fromwaste (Lilja and Liukkonen, 2008). When considering
the most sustainable waste management option, disposal should be
the least promoted, especially when the organic solvents have an
addedvalue. Energy recovery is more preferable tothis; however, on-
site recovery of solvents is the most sustainable strategy. In this way,
not only the organic solvents but also the water can be recovered and
thus contributing to overall resource efciency. Resource efciency is
a priority area of the United Nations Environment Programme and
a agshipinitiative of the EU, and refers to a sustainable use of Earths
resources. In our perspective, resource efciency represents taking
aviewbeyondwastes that needtobedealt with, towards resources to
be used sustainably.
The recovery of organic solvents from chemical efuents is
complicated and technically challenging. Most of the chemical
efuents are not highly concentrated in organic solvents. This
implies an extensive use of resources in order to obtain a stream of
a re-usable quality. Added difculties are the volatile nature of
VOCs and the characteristic of some organic solvents of forming
azeotropes. Further, the wastewaters are of a complex nature. It is
often the case where electrolytes are also part of those efuents and
this fact affects the recovery strategy. Membrane technology is
a clean alternative for the recovery of organic solvents from
aqueous systems containing electrolytes (Urtiaga et al., 2006;
Koltuniewicz and Drioli, 2008). Among the different membrane
processes, pervaporation has excellent potential because of the
advantages of energy efciency, selectivity, environmental benign,
easy operation and effective recovery of volatile components even
from diluted systems.
This article argues that there is a need for a paradigm change in
viewing wastes, even those characterized as hazardous, as
resources to be utilized rather than a problem that needs to be
amended. In this article, we suggest pervaporation for the recovery
of two particular organic solvents, n-butanol and dichloromethane,
from aqueous systems containing sodium chloride. The process is
based on previous results achieved by the authors. The main goal is
to present recovery as a sustainable management option for VOC
containing wastewaters and pervaporation as a clean technology
that promotes resource efciency.
2. Reclamation of organic solvents from efuents containing
electrolytes by pervaporation
The use of pervaporation for the recovery of organic solvents
from chemical efuents may be affected by the presence of elec-
trolytes. The electrolytes may have a positive inuence on the
pervaporation performance through the salting out effect. This
effect entails the rise of the activity coefcient due to the reduction
of the solubility of the component (Zemaitis et al., 1986) and causes
an increase of the partial pressure of the compound on the feed
side. The difference in partial pressure between both sides of the
membrane is hence enhanced and so is the pervaporation driving
force and the separation performance.
The presence of electrolytes may have also a negative effect on
the pervaporation process. The electrolytes may cause an increase
in the density and viscosity of the feed solution decreasing the mass
transfer rate. Further, the existing salts may penetrate the
membrane causing membrane fouling and may enhance the solu-
bility of the organic compound in the solvent, i.e., salting-in effect.
The salting-in effect would cause the decrease of the activity
coefcient, which reduces the pervaporation driving force.
Several authors have studied the separation of VOCs from
aqueous mixtures containing electrolytes using hydrophobic per-
vaporation membranes (Table 1). Generally, it was observed that
the presence of electrolytes had a positive effect on the pervapo-
ration performance of the membranes. Kujawski and Krajewski
(2007) observed an increase of 59% in the enrichment value of
PERVAP 1060 PDMS membrane towards methyl acetate at 25

C
when the feed mixture contained 0.5% of the organic solvent and
4 M of sodium chloride. Lipnizki et al. (2004) found that the
addition of salts and sugars increased the separation factor of
PERVAP 1060 towards 1-propanol in the following order: magne-
sium chloride > sodium chloride > glucose. Bddeker and
Bengtson (1990) also observed an enhancement of the enrich-
ment factor of a PEBA membrane towards phenol from approxi-
mately 68e105 when the sodium chloride was present up to 10 %
wt. In all the mentioned research, the improvement was caused
Table 1
Studies on the separation of VOCs from mixtures containing VOCs and electrolytes using hydrophic pervaporation membranes: literature and own results.
VOC Electrolyte Membrane T (

C) J
VOCs
(10
5
kg m
2
s
1
) b, a
a
Reference
n-Butanol (0.2 %wt) NaCl (<1 %wt) CMX-GF-010-D 40 0.47 0.03 39 1 Garca et al., 2009a
PERTHESE

500-1 40 0.06 0.09 42 7

DCM (0.5 %wt) NaCl (<1 %wt) CMX-GF-010-D 40 0.51 0.02 16 1 Garca et al., 2009b
n-Butanol (0.2 %wt), NaCl (<1 %wt) CMX-GF-010-D 40 But: 0.40 0.05 But: 37 3 Garca et al., 2009b
DCM (0.5 %wt) DCM: 0.42 0.03 DCM: 15 1
Methyl acetate (<3 %wt) NaCl (<4 M) Pervap 1060 25 5.56 < J < 44.4 26 < b < 105 Kujawski and Krajewski, 2007
1-propanol (5 %wt) NaCl, (0.1 M-1.5 M) Pervap 1060 80 22.2 < J < 34.7 8 < a < 16 Lipnizki et al., 2004
MgCl
2
(0.1 M-1.5 M) 80 19.4 < J < 45.8 7 < a < 17
Glucose (0.1 M-1.5 M) 80 19.4 < J < 34.7 7 < a < 16
NaCl, (0.1 M-1.5 M) Pervap 1070 80 10.6 < J < 19.2 12 < a < 26
MgCl
2
(0.1 M-1.5 M) 80 11.1 < J < 21.9 14 < a < 43
Glucose (0.1 M-1.5 M) 80 11.4 < J < 13.9 14 < a < 20
n-Butanol (1 %wt) (NH
4
)
2
SO
4
(<10 g L
1
) PDMS 40 13.9 < J < 15.0 Favre and Nguyen, 1996
Trichloroethylene (200,
300 and 400 ppm)
NaCl, KCl, Na
3
PO
4
FeCl
3,
CaCl
2
(<2.6 M)
Zeolite lled PDMS 50 C 200 ppm: 0.78
< J < 0.42; C 300 ppm: 0.61
< J < 1.44; C 400 ppm: 0.5
< J < 0.86
Dotrmont et al., 1994
Phenol (400 mg kg
1
) NaCl (<10 g L
1
) PEBA 40 50 0.13 < J < 0.19 65 < b < 105 Bddeker and
Bengtson, 1990
a
Unless specied data refers to b.
V. Garca et al. / Journal of Cleaner Production 39 (2013) 146e153 147
by the simultaneous increase of the permeate uxes of the organic
compounds and the decrease of the water ux. On the contrary,
Dotrmont et al. (1994) showed a decrease of 65% of the trichlo-
roethylene ux and no effect to the water ux through a zeolite-
lled PDMS membrane when potassium chloride, sodium chlo-
ride, sodium phosphate, ferric chloride or calcium chloride were
added to the organic solvent/water system. They attributed the
decline ux to an increase in the resistance to mass transfer and to
the formation of a salt-rich layer at the feed side of the membrane
that repelled the trichloroethylene (Dotrmont et al., 1994). In the
literature reviewed in Table 1, the membranes were found to be
impermeable to the electrolytes. Of these, Lipnizki et al. (2004)
showed evidences of the penetration of sodium chloride into the
structure of the membrane and its presence on the membrane
surface and Dotrmont et al. (1994) found traces of potassium and
chloride ions in the support layer.
Generally, the permeates obtained in a pervaporation process
are not enriched enough to be directly re-used. Consequently, the
dehydration of the organic solvents is needed. Organic dewatering
is an important and complicated step because many common
solvents such as n-butanol form azeotropes with water. The
dehydration of organic solvents by pervaporation is conducted
using hydrophilic membranes and overcomes the azeotropic
difculty.
3. Separation of n-butanol and dichloromethane from
sodium chloride containing mixtures
Among the variety of organic solvents that are present in
chemical efuents, dichloromethane and n-butanol are of interest
for being recovered on-site and further utilized. Dichloromethane
is classied as a carcinogen (category 3) by the CLP regulation (EC)
1272/2008. Some observed effects of dichloromethane after acute
and/or chronic exposures are reversible central nervous system
depression, narcosis, irritation of eyes, skin and respiratory tract,
lung oedema as well as production of elevated carboxyhemoglobin
levels in the blood (Dhillon and Von Burg, 1995). Further,
dichloromethane is included in the List of priority substances in the
eld of European water policy in the Water Framework Directive
2008/105/EC. Due to its toxicity, the conventional management for
wastewater containing dichloromethane is to reduce the hazard,
which would result in the destruction of the toxic compound.
Considering that DCM is a valuable solvent, our view was to treat it
as a displaced resource. The recovery of n-butanol is to be
considered as well, due to its potential use as a liquid transportation
fuel (Garca et al., 2011).
Fig. 1 illustrates a proposed methodology for the reclamation of
n-butanol and dichloromethane from the waters containing
sodium chloride using pervaporation. The process consists of
several pervaporation stages using hydrophobic and hydrophilic
membranes. After permeating the feed sample through the
hydrophobic membrane (Stage I), the amount of n-butanol and
dichloromethane in the initial sample is reduced. The permeate
does not contain the electrolyte, as sodium chloride is not a volatile
compound. Since only water permeates through the hydrophilic
membrane (Stage II), the result of this stage is a mixture of
dichloromethane and n-butanol as retentate, and water as
permeate.
The experimental conditions of Stages I and II are summarized
in Table 2. The pervaporation experiments were conducted using
a cross-ow laboratory scale membrane unit with an effective
membrane area of 2.8 10
3
m
2
(model P28, Celfa AG,
Switzerland).
The study of the separation of n-butanol and dichloromethane
from the quaternary system n-butanol/dichloromethane/sodium
chloride/water by pervaporation (Stage 1) was conducted using
two hydrophobic membranes: CMX-GF-010-D (CELFA AG,
Switzerland) and PERTHESE

500-1 (Perouse Plastie, France). The

CMX-GF-010-D membrane is made of polysiloxane polymer with
an active layer thickness of 0.01 mm and PERTHESE

500-1 is
a dense membrane of 0.125 mmmade of dimethyl and methyl vinyl
siloxane copolymers. In the work related to Stage I the effect of
several experimental variables (feed temperature and sodium
chloride feed concentration) on the pervaporation performance of
the two membranes was studied. The feed temperature range was
selected according to the membrane tolerance and the feed
concentration was chosen according to the concentration of n-
butanol and dichloromethane typically found in wastewaters.
In Stage II the hydrophilic membrane consisted of an active layer
of 1 mm of polyvinyl alcohol and titanium dioxide (PVA-TiO
2
) on
a support of polyacrylonitrile (PAN) and polyphenylene sulde
(PPS) (supplied by Helmholtz-Zentrum Geesthacht Zentrum fr
Material und Kstenforschung GmbH, Germany).
The effects of the following four factors on the efciency of the
pervaporation process were studied: feed temperature, initial feed
water concentration, initial feed dichloromethane concentration
and initial feed n-butanol concentration. The range of the factors
was based on preliminary trials, the tolerance of the membrane and
Fig. 1. Schematic diagram of the proposed approach for the reclamation of n-butanol and dichloromethane from aqueous mixtures containing sodium chloride (Garca, 2009).
V. Garca et al. / Journal of Cleaner Production 39 (2013) 146e153 148
the homogeneity of the feed solution. The conducted experiments
were designed according to the Response Surface Methodology
(RSM) using the MODDE 8.0 software (Umetrics) and are listed in
Garca et al. (2009c).
Feed and permeate samples were taken regularly in Stage 1 and
Stage 2. The quantication of n-butanol and dichloromethane in the
samples collected in Stage 1 was conducted using Gas Chromatog-
raphy with Flame Ionization Detector (Agilent, 6890N) while the
sodium chloride content was determined as sodium using Atomic
Absorption Spectroscopy (Perkin Elmer, Aanalyst 100). The concen-
tration of the VOCs in the feed samples gathered in Stage 2 was
quantied using Gas Chromatography with a Thermal Conductivity
Detector (Agilent, 6890N) whereas Gas Chromatography with Flame
Ionization Detector (Agilent, 6890N) was used in the determination
of the organic solvents in the permeate samples. Further, the salt
content in the hydrophobic membranes was studied by X-ray
excited photoelectron spectroscopy using Scienta SES-200 electron
energy analyser with commercial Specs X-ray source.
The performance of the pervaporation process was described by
the concentration of the components in the permeate, the uxes
through the membrane and the enrichment factor. Flux (J) repre-
sented the rate of permeation of feed components per unit of area
and time. Flux was dened as J m/Dt A, where m was the
amount of compounds collected in the permeate during the
experimental time interval Dt and A was the effective membrane
area. The enrichment factor (b) was determined by the expression
b y
1
/x
1
, where x
1
was the weight fraction of the minor compound
in the feed and y
1
in the permeate.
4. Results
4.1. Stage I: the separation of n-butanol and dichloromethane
rom waters containing sodium chloride using hydrophobic
pervaporation membranes
4.1.1. Effect of temperature on the pervaporation performance of
CMX-GF-010-D and PERTHESE

500-1 membranes
The effect of temperature on the composition of the permeate
obtained in the pervaporation of the n-butanol/dichloromethane/
sodium chloride/water system through the membranes CMX-GF-
010-D and PERTHESE

500-1 is illustrated in Fig. 2.

As observed, an increase in temperature when the pervapora-
tion of the system was conducted using the CMX-GF-010-D
membrane implied a rise in the concentration of the organic
solvents in the permeate (Fig. 2a). The temperature inuence was
moderately stronger for n-butanol than for dichloromethane. The
concentration of n-butanol in the permeate augmented 43%, from
4.42 to 6.32 %wt., whereas the dichloromethane content increased
30%, from 4.86 to 6.30 %wt. The enrichment of the permeate in
VOCs was due to the fact that temperature had a more intense
impact on the uxes of the organic solvents through the hydro-
phobic membrane than on the partial ux of water. The uxes of n-
butanol, dichloromethane and water through the CMX-GF-010-D
membrane were promoted 3.9, 3.5 and 2.6 times, respectively
when raising the temperature from 22 to 40

C (Garca et al.,
2009b). The promotion of the uxes of the organic solvents
through the CMX-GF-010-D membrane compared to water uxes
caused the increase of the enrichment factor towards n-butanol up
to 32% while it did not affect signicantly the enrichment factor
towards dichloromethane (Garca et al., 2009b).
Fig. 2b also indicates that the concentration of n-butanol and
dichloromethane in the permeate obtained in the pervaporation of
the quaternary system using the PERTHESE

500-1 membrane was

enhanced when the temperature increased from 22 to 40

C.
Further, the permeation of dichloromethane through the
membrane was slightly more affected by temperature than the
permeation of n-butanol through the same hydrophobic
membrane. The concentration of n-butanol in the permeate
increased from 6.63 to 13.20 %wt. and the content of dichloro-
methane was promoted from 4.43 to 9.52 %wt. The increase in
temperature also lead to an enhancement of the uxes through the
membrane, of n-butanol 3.2 times, of dichloromethane 3.6 times
and of water 1.4 times (Fig. 3a).
The enrichment factors of the PERTHESE

500-1 membrane
towards n-butanol and dichloromethane became also greater with
rising temperature (Fig. 3b). The enrichment factor towards n-
butanol increased 88% while the enrichment factor towards
dichloromethane increased 122% when rising the temperature
from 22 to 40

C.
The results shown in Fig. 3 and Table 1 indicate that the CMX-
GF-010-D membrane exhibited higher uxes than the PERTHESE

500-1 membrane. On the contrary, the pervaporation of the

mixtures using the PERTHESE

500-1 membrane resulted in

permeates more enriched in VOCs. These ndings were due to the
higher thickness of the PERTHESE

500-1 membrane.
When studying the effect of temperature on the pervaporation
performance of the membranes, the quantication of the perme-
ates informed that the electrolyte present in the feed solution did
not permeate through the hydrophobic membranes. XPS analysis
indicated no sodium above the estimated detection limit of 0.2% of
Table 2
Stages, aims and experimental conditions of the process for the reclamation of n-butanol and dichloromethane from waters containing sodium chloride by pervaporation.
Stage Aim Experimental conditions
I Separation of the initial efuent into brine solution
and an enriched mixture of organic solvents
Membrane CMX-GF-010-D (Celfa, Switzerland)
PERTHESE

500-1 (Perouse Plastie, France)

Temperature 22, 30 and 40

C
Feed concentration 0.2 %wt of n-butanol
0.5 %wt of dichloromethane
<1 %wt of sodium chloride
Vacuum pressure < 3 mbars
Feed ow velocity 3 l min
1
Operating time 4 h
II Dehydrating the organic mixture Membrane PVA-TiO
2
/PAN/PPS (GKSS, Germany)
Temperature 30, 40 and 50

C
Feed concentration n-butanol: 50e90 %wt
Dichloromethane: 5e45 %wt
Water: 5e20 %wt
Vacuum pressure <3 mbars
Feed ow velocity 3 l min
1
Operating time 4 h
V. Garca et al. / Journal of Cleaner Production 39 (2013) 146e153 149
the amount of silicon neither in CMX-GF-010-D (Garca et al.,
2009b) nor in PERTHESE

500-1. As the high temperatures

beneted the pervaporation of the n-butanol/dichloromethane/
sodium chloride/water mixture the effect of salt concentration was
studied at 40

C.
4.1.2. Effect of the presence electrolyte on the pervaporation
performance of CMX-GF-010-D and PERTHESE

500-1 membranes
Fig. 4a illustrates the effect of the addition of sodiumchloride on
the composition of the permeates obtained in the pervaporation
experiments using CMX-GF-010-D membrane. When sodium
chloride was incorporated to the n-butanol/dichloromethane/
water mixture the concentration of the VOCs in the permeate
decreased insignicantly. This trend was also observed with
increasing electrolyte concentration. The concentration of n-
butanol in the permeate when altering the sodium chloride feed
concentration from0 %wt. to 0.9 %wt. at 40

C was 6.03 0.55 %wt.
whereas the dichloromethane content was 6.38 0.32 %wt. The
fact that the concentration of the organic solvents remained about
constant was due to the negligible effect of the presence of sodium
chloride on the partial uxes of the solutes and water. Hence, no
salting-in or out phenomena was observed. The uxes of n-butanol,
and dichloromethane through the CMX-GF-010-D membrane and
enrichment factor towards the VOCs are indicated in Table 1. It is
important to note that no fouling was observed during the exper-
iments. Similar observations were done by Nguyen and Noble
(1987) when conducting the pervaporation of dichloromethane,
bromoethane, acetone and ethanol through silicone tubular
membranes in the presence of sodium chloride and potassium
iodide. Favre and Nguyen (1996) observed that adding up to
10 gdm
3
of ammonium sulphate to a 1-butanol/water mixtures
induced negligible uxes modications through a GFT homogenous
PDMS membrane. Kujawski and Krajecski did not found either any
salting in or out effect on the pervaporation of methyl acetate/water
solution when the concentration of the organic solvent was 3 %wt
and the concentration of the electrolyte was in the range of 0e4 M.
Figs. 4b and 5 show the effect of the addition of sodium chloride
on the pervaporation performance of the PERTHESE

500-1
membrane. Contrary to the experience with CMX-GF-010-D
membrane, the presence of the electrolyte in the feed solution
did affect remarkably the pervaporation of the quaternary mixture.
The addition of sodium chloride to the feed solution caused the
enrichment of the permeate in organic solvents. The concentration
of n-butanol was intensied 90.2% while the concentration of
dichloromethane increased 83.6% (Fig. 4b). The concentration of
water decreased up to 15.3%, from 85.1 to 72.1 %wt. The rise in the
concentration of the organic solvents in the permeate resulted in an
increase in the enrichment factor of the PERTHESE

500-1
membrane towards n-butanol and dichloromethane (Fig. 5a). The
enhancement of the enrichment factor towards dichloromethane
was a result of the salting out effect, i.e., promotion of the ux of
dichloromethane through the membrane and a decrease of the
water ux (Fig. 5b). The ux of n-butanol through the membrane
was, however, not altered by the presence of the electrolyte
(Fig. 5b). Consequently, the increase of the enrichment factor
towards n-butanol was due to the combination of intensication of
the dichloromethane ux and the diminution of the water ux
through the membrane.
The results achieved in Stage I indicated that the VOCs
concentration in the permeates obtained when conducting
the pervaporation of the n-butanol/dichloromethane/sodium
chloride/water system using CMX-GF-010-D or PERTHESE

500-
Fig. 2. Effect of temperature on the concentration of n-butanol and dichloromethane in the permeate obtained in the pervaporation of n-butanol/dichloromethane/sodium
chloride/water mixture using a) CMX-GF-010-D and b) PERTHESE

500-1 membranes. Feed composition: C

n-Butanol
0.2 %wt.; C
Dichloromethane
0.5 %wt.; C
Sodium chloride
0.5 %wt.
Fig. 3. Effect of temperature on a) n-butanol and dichloromethane uxes through the PERTHESE

500-1 membrane and b) enrichment factor towards n-butanol and dichloro-

methane for the n-butanol/dichloromethane/sodium chloride/water mixture. Feed composition: C
n-Butanol
0.2 %wt.; C
Dichloromethane
0.5 %wt.; C
Sodium chloride
0.5 %wt.
V. Garca et al. / Journal of Cleaner Production 39 (2013) 146e153 150
1membranes was not enough for the direct re-utililization of the
VOCs at the studied conditions.
4.2. Stage II: dehydration of mixtures containing n-butanol and
dichloromethane using hydrophilic pervaporation membranes
The dehydration of the water/n-butanol/dichloromethane
system was conducted at different feed compositions and
temperatures (Table 2). The dewatering effectiveness of the
membrane is inversely related to the amount of organic solvents
found in the permeate. The results indicated the impermeable
features of the PVA-TiO
2
membrane towards dichloromethane in
the studied conditions (Garca et al., 2009c). The membrane,
however, allowed the permeation of some quantities of n-butanol
due to the moderate hydrophilic nature of the compound given by
the alcohol group. The presence of n-butanol in the permeate
depended on the feed working conditions. The 3D contour plot
illustrated in Fig. 6 describes the effect of feed concentration on the
concentration of n-butanol in the permeate at the studied
temperatures 30, 40 and 50

C. As observed, at higher temperatures

the quantity of n-butanol in the permeate is lower. When perva-
poration was conducted at 50

C, the maximum observed
concentration was 2.94 wt.%, while it reached up to 10 wt.% when
the selected temperature was 30

C. This was due to the fact that
temperature enhanced the pervaporation driving force, and the
effect of temperature on uxes in this study was higher for water
than for n-butanol (Garca et al., 2009c). Similar observations were
done by Sommer and Melin (2005) when conducting the perva-
poration of water/n-butanol mixtures through tubular inorganic
membranes.
Additionally, according to the contour plots, the n-butanol
concentration in the permeate augmented with initial n-butanol
feed concentration. This was the outcome of the pervaporation
driving force being promoted with feed concentration and that
increasing n-butanol resulted in a decrease in the feed water
concentration, diminishing the driving force for water permeation.
Accordingly, increase in feed water concentration decreased the
n-butanol permeate concentration. From the subplots of Fig. 6
it may be observed that the initial concentration of dichloro-
methane needed to be altered considerably to cause a decrease in
the n-butanol permeate concentration.
The optimal dehydration conditions of Stage II, achieved in
a previous work, were a feed temperature of 49

C and initial feed
concentrations of water, dichloromethane and n-butanol of 6.64 wt
%, 41 wt% and 52.4 wt%, respectively (Garca et al., 2009c). At these
conditions, the resulting permeate did not contain dichloro-
methane and the concentration of water and n-butanol are
97.6 0.1 wt% and 2.42 0.13 wt%, respectively. The uxes of water
and n-butanol through the membrane were 4.39 10
5
0.20 10
5
and 0.11 10
5
0.08 10
5
kg m
2
s
1
.
5. Discussion
The preferred options of Cleaner Production are prevention or
minimization by better product design, process optimization,
monitoring, training and management, combined with improved
governmental policies (Klemes et al., 2012). Much progress has
been made towards this in the past decades, yet much more needs
to be done (Ibid.). Most products in the ne chemicals industry are
manufactured discontinuously by multiproduct plants. Batch
Fig. 5. Effect of the addition of sodium chloride on a) enrichment factor of PERTHESE

500-1 membrane towards n-butanol and dichloromethane and b) n-butanol and

dichloromethane uxes through the membrane for the n-butanol/dichloromethane/sodium chloride/water mixture. VOC composition in the feed: C
n-Butanol
0.2 %wt.;
C
Dichloromethane
0.5 %wt.
Fig. 4. Effect of the addition of sodium chloride on the concentration of n-butanol and dichloromethane in the permeate for the quaternary system n-butanol/dichloromethane/
sodium chloride water at 40

C using a) CMX-GF-010-D and b) PERTHESE

500-1 membranes. VOC composition in the feed: C

n-Butanol
0.2 %wt.; C
Dichloromethane
0.5 %wt.
V. Garca et al. / Journal of Cleaner Production 39 (2013) 146e153 151
processes are traditionally preferred due to their handling different
product varieties and reactant properties. On the other hand, their
disadvantage is large water consumption and wastewater produc-
tion (Grundemann et al., 2012). If it is not possible to transfer batch
processes to continuous production schemes, as suggested by
Grundemann et al. (2012), attention needs to be oriented to the
treatment of the wastewater. Negny et al. (2012) pointed out that,
for chemical industries, waste management needs to transition
from an environmental approach to an economical one: the waste
is not only a constraint to minimize but also a resource to optimize.
On the other hand, Giurco et al. (2011) argued that, as opposed to
the traditional waste management hierarchy, in case of industrial
water management, there is a greater tension between reducing
on-site use though water use efciency, if there is a water recycling
or reuse synergy. In some cases, as highlighted in our earlier work
(Garca et al., 2008), prevention is not viable and thus recycling
opportunities must be found. Negny et al. (2012) also suggest that
chemical industry must innovate to create processes to deliver
value fromwaste. Pajunen et al. (2012) asserted that a way to drive
process industry towards treating residue streams as material ows
appears to be missing also in other sectors.
In the case of hazardous efuents the main routes are special-
ized treatment to reduce the hazard and high temperature
destruction (Cristbal Andrade et al., 2012; Lilja and Liukkonen,
2008), although Andrade et al. have also considered (off-site)
recovery strategies for wastes considered hazardous. A more ef-
cient wastewater management would be to recover hazardous
components and transform them into resources and/or water
purication and re-use. Membrane technologies are mentioned as
one of the important problem solving technologies (Weinberger
et al., 2012).
Membrane technologies are very often used for the purication
of water from residues considered pollutants (Tahri et al., 2012).
Membrane ltering of process water was demonstrated to be an
economic and technological feasible technology in the textile
industry (Sndergrd et al., 2004 and references therein). The
economic benets of ltration technologies for wastewater treat-
ment in the polymer industry have also been demonstrated (Mutlu
et al., 2012). Some other works focused on the concentration of
hazardous components from wastewaters in order to be utilized.
Galiana-Aleixandre et al. (2011) proposed nanoltration for the
separation of chromium sulphate from the tanning washing
wastewater and re-use in the tanning drums.
Application of membrane technologies to VOC-containing
wastewaters could serve several purposes; reduction of the
amount and hazard of wastewater, recirculation of re-use quality
recovered water, as well as recovery of valuable (albeit hazardous)
components. The strategy should not be solely the avoidance of
direct environmental damages, but also a fundamental change in
perspective is needed, as suggested by Weinberger et al. (2012).
Pervaporation is mentioned as a CLEANER (Combining Lower
Emissions And Networked Energy Recovery) methodology to
remove pollutants fromend-of-pipe streams (Dunn and Bush, 2001).
Some of the challenges faced in this study were the low
concentration of recoverable organics (n-butanol and dichloro-
methane) in the efuent and the presence of the electrolyte. Here
lies the potential of pervaporation, for being efcient in diluted
streams as well as its ability to retain salt. Pervaporation has gained
renewed interest for its potential application for the dehydration of
biofuels (Garca et al., 2011). Our suggestion is to consider perva-
poration a clean technology, capable of delivering value also from
wastewater streams.
6. Conclusions
This paper reported on an extensive study of pervaporation,
with the goal of establishing this membrane technology as a clean
process for the recovery of volatile organic solvents from waste-
water. A synthetic mixture containing n-butanol, dichloromethane
and sodium chloride was studied, representing a discharge from
the ne chemical sector. This investigation suggests that, by
applying subsequent pervaporation stages of hydrophobic followed
by hydrophilic membranes, this hazardous and complex waste-
water can be transformed into a resource. With the two-stage
process the quarternary mixture was separated into a concentrate
of organic solvents, brine and an aqueous system.
Fig. 6. Contour plot for the effects of the initial feed concentration of water (C
w
), dichloromethane (C
DCM
) and n-butanol (C
b
) on the concentration of n-butanol in the permeate at
30, 40 and 50

C.
V. Garca et al. / Journal of Cleaner Production 39 (2013) 146e153 152
The reported work indicated that the removal of n-butanol and
dichloromethane from mixtures containing sodium chloride was
better conducted at high temperatures. The pervaporation of the
mixtures using PERTHESE

500-1 membrane resulted in permeates

more enriched in the organic solvents compared to the pervapo-
ration using CMX-GF-010-D. However, the latter exhibited higher
uxes. This work also concluded the existence of the salting out
phenomena when the PERTHESE

500-1 membrane was used.

The dehydration of the organic solvents using the PVA-TiO
2
membrane was best conducted at a feed temperature of 49

C and
at the following initial feed concentrations: water 6.64 wt%,
dichloromethane 41 wt% and n-butanol 52.4 wt% in order to obtain
a permeate with the minimum amount of VOCs.
The stream containing the organic solvents may be further
separated by distillation, as n-butanol and dichloromethane have
different boiling points, to be re-used internally or by another
factory. Additional processes could be considered for the treatment
of the resulting brine and aqueous system, so the recovered water
may be further utilized for industrial processes or environmental
applications.
While further investigation would be needed to establish the
applicability of this proposed process in industrial scale applica-
tions, the results achieved in laboratory scale experiments
demonstrated the potential of pervaporation as a viable wastewater
management option for the recovery of VOCs. The amount of
efuent to be treated and the high volatility of dichloromethane
maylimit the large-scale use of this technology. Yet we would like to
reiterate the key notion that there is a place and need for a paradigm
change in viewing even hazardous wastes as potential resources.
Acknowledgements
The authors would like to thank the nancial support of the
Academy of Finland (project no. 111416), the Maj and Tor Nessling
foundation (projects no. 2006262, 2007196 and 2008394), the
Oskar lund Foundation, the Finnish Foundation for Technology
Promotion, the Oulu University Foundation, the Tauno Tnning
Foundation, the Environet Graduate School, the Finnish Cultural
Fund and the Thule Institute. Further, the authors would like to
thank to the Helmholtz-Zentrum Geesthacht Zentrum fr Material
und Kstenforschung GmbH (Centre for Materials and Coastal
Research) for generously supplying the membranes used in the
experiments. The Electron Spectroscopy Group at the Department
of Physical Sciences, University of Oulu is gratefully acknowledged
for the X-ray excited photoelectron spectroscopy analysis of
membranes.
References
Bddeker, K., Bengtson, G., 1990. Pervaporation of low volatility aromatics from
water. J. Membr. Sci. 53, 143e158.
Cristbal Andrade, L., Gmez Mguez, C., Taboada Gmez, M.C., Bello Bugallo, P.M.,
2012. Management strategy for hazardous waste from atomised SME: appli-
cation to the printing industry. J. Clean. Prod. 35, 214e229.
Dhillon, S., Von Burg, R., 1995. Toxicology update. J. Appl. Toxicol. 15, 329e335.
Dotrmont, C., Van den Ende, S., Vandommele, H., Vandecasteele, C., 1994.
Concentration polarization and other boundary layer effects in the pervapora-
tion of chlorinated hydrocarbons. Desalination 95, 91e113.
Dunn, R.F., Bush, G.E., 2001. Using process integration technology for cleaner
production. J. Clean. Prod. 9, 1e23.
Favre, E., Nguyen, Q.T., 1996. Extraction of 1-butanol from aqueous solutions by
pervaporation. J. Chem. Tech. Biotechnol. 65, 221e228.
Galiana-Aleixandre, M.V., Mendoza-Roca, J.A., Bes-Pi, A., 2011. Reducing sulfates
concentration in the tannery efuent by applying pollution prevention tech-
niques and nanoltration. J. Cleaner Prod. 19, 91e98.
Garca, V., 2009, Reclamation of VOCs, n-butanol and dichloromethane, from sodium
chloride containing mixtures by pervaporation. Towards efcient use of resources
in the chemical industry. Doctoral dissertation, University of Orciaulu, Faculty of
Technology, Department of Process and Environmental Engineering. Acta Univ.
Oul. C 341, 2009, Oulu, Finland. http://urn./urn, ISBN: 9789514292644.
Garca, V., Pongracz, E., Phillips, P., Keiski, R., 2008. Factors affecting resource use
optimisation of the chemical industry in the Northern Ostrobothnia region of
Finland. J. Clean. Prod. 16, 1987e1994.
Garca, V., Pongracz, E., Muurinen, E., Keiski, R., 2009a. Recovery of n-butanol
from salt containing chemical wastewaters by pervaporation. Desalination 241,
201e211.
Garca, V., Pongracz, E., Muurinen, E., Keiski, R., 2009b. Pervaporation of dichloro-
methane from multicomponent aqueous systems containing n-butanol and
sodium chloride. J. Membr. Sci. 326, 92e102.
Garca, V., Landaburu-Aguirre, J., Pongrcz, E., Permki, P., Keiski, R., 2009c.
Dehydration of water/dichloromethane/n-butanol mixtures by pervaporation;
optimisation and modelling by response surface methodology. J. Membr. Sci.
338, 111e118.
Garca, V., Pkkil, J., Ojamo, H., Muurinen, E., Keiski, R., 2011. Challenges in bio-
butanol production: how to improve the efciency? Renew. Sustain. Energy
Rev. 15, 964e980.
Giurco, D., Bossilkov, A., Patterson, J., Kazaglis, A., 2011. Developing industrial water
reuse synergies in Port Melbourne: cost effectiveness, barriers and opportuni-
ties. J. Clean. Prod. 19, 867e876.
Grundemann, L., Gonschorowski, V., Fischer, N., Sholl, S., 2012. Cleaning waste
minimization for multiproduct plants: transferring macro batch to micro conti
manufacturing. J. Clean. Prod. 24, 92e101.
Klemes, J.J., Varbanov, P.S., Huisingh, D., 2012. Recent cleaner production advances
in process monitoring and optimization. J. Clean. Prod. 34, 1e8.
Koltuniewicz, A., Drioli, E., 2008. Membranes in Clean Technologies. Wiley-Verlag
GmbH & Co. KGaA, Weinheim, Germany.
Kujawski, W., Krajewski, S., 2007. Inuence of inorganic salt on the effectiveness of
liquid mixtures separation by pervaporation. Sep. Purif. Technol. 57, 495e501.
Lilja, R., Liukkonen, S., 2008. Industrial hazardous wastes in Finland e trends related
to the waste prevention goal. J. Clean. Prod. 16, 343e349.
Lipnizki, F., Hausmanns, S., Field, R., 2004. Inuence of impermeable components
on the permeation of aqueous 1-propanol mixtures in hydrophobic pervapo-
ration. J. Membr. Sci. 228, 129e138.
Mutlu, A., Lee, B.K., Lee, C.H., Kim, J.Y., 2012. Sludge minimization and production
improvement by changing ltration and reaction techniques in vinyl acetate-
ethylene copolymerization. J. Clean. Prod. 24, 109e116.
Negny, S., Belaud, J.P., Cortes Robles, G., Roldan Reyes, E., Barragan Ferrer, J., 2012.
Toward an eco-innovative method based on a better use of resources: appli-
cation to chemical process preliminary design. J. Clean. Prod. 32, 101e113.
Nguyen, Q.T., Noble, K., 1987. Extraction of organic contaminants in aqueous solu-
tions by pervaporation. J. Membr. Sci. 30, 11e22.
Pajunen, N., Watkins, G., Wierink, M., Heiskanen, K., 2012. Drivers and barriers of
effective industrial material use. Miner. Eng. 29, 39e46.
Sommer, S., Melin, T., 2005. Inuence of operation parameters on the separation of
mixtures by pervaporation and vapor permeation with inorganic membranes.
Part 1: dehydration of solvents. Chem. Eng. Sci. 60, 4509e4523.
Sndergrd, B., Hansen, O.E., Holm, J., 2004. Ecological modernisation and institu-
tional transformations in the Danish textile industry. J. Clean. Prod. 12, 337e
352.
Tahri, N., Masmoudi, G., Ellouze, E., Jrad, A., Drogui, P., Amar, R.B., 2012. Coupling
microltration and nanoltration processes for the treatment at source of
dyeing-containing efuent. J. Clean. Prod. 33, 226e235.
Urtiaga, A.M., Gorri, D., Ortiz, I., 2006. Pervaporative recovery of isopropanol from
industrial efuents. Sep. Purif. Technol. 49, 245e252.
Weinberger, N., Jrissen, J., Schippl, J., 2012. Foresight on environmental technol-
ogies: options for the prioritization of future research funding e lessons learned
from the project Roadmap Environmental Technologies 2010. J. Clean. Prod.
27, 32e41.
Zemaitis, J., Clark, D., Rafal, M., Scruvner, N., 1986. Handbook of Aqueous Electrolyte
Thermodynamics; Theory and Applications. American institute of Chemical
Engineering Inc., New York, USA.
Zogorski, J., Carter, J., Ivahnenko, T., Lapham, W., Moran, M., Rowe, B., Squillace, P.,
Toccalino, P., 2006. The Quality of Our Nations Waters e Volatile Organic
Compounds in the Nations Ground Water and Drinking-Water Supply Wells.
US Geological Survey, Virginia. URL: http://pubs.usgs.gov/circ/circ1292/pdf/
circular1292.pdf (accessed 24.07.12.).
V. Garca et al. / Journal of Cleaner Production 39 (2013) 146e153 153