Journal of Food Engineering 158 (2015) 39–47

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier.com/locate/jfoodeng

Membrane fractionation of herring marinade for separation and recovery

of fats, proteins, amino acids, salt, acetic acid and water
Lene Fjerbæk Søtoft ⇑, Juncal Martin Lizarazu, Behnaz Razi Parjikolaei, Henrik Karring, Knud V. Christensen
University of Southern Denmark, Campusvej 55, 5230 Odense M, Denmark

a r t i c l e i n f o a b s t r a c t

Article history: In the production of marinated herring, nearly one ton of acidic saline marinade is produced per 1.5 tons
Received 14 November 2014 herring fillet. This spent marinade contains highly valuable compounds such as proteins and amino acids.
Received in revised form 11 February 2015 Membranes are suited to recover these substances. In this work, six membrane stages are employed:
Accepted 23 February 2015
microfiltration (MF) (0.2 lm), ultrafiltration (UF) (50, 20, 10 and 1 kDa) and nanofiltration (NF).
Available online 2 March 2015
The most promising stages are 50 kDa UF and NF based on SDS–PAGE analyses and total amino acid
concentration. The 50 kDa stage produces a protein concentrate (>17 kDa). NF produces a retentate con-
Keywords:
taining sugars, amino acids and smaller peptides and a NF permeate containing salt and acetic acid ready
Fish food waste
Herring brine
for reuse. 42% of the spent marinade is recovered to substitute fresh water and chemicals. The waste
Protein characterization water amount is reduced 62.5%. Proteins are concentrated 30 times, while amino acids and smaller pep-
Membrane filtration tides are concentrated 11 times.
Fractionation Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction separation, removal or recovery of organic material from fish

The seafood industry for human foods is a very water-intensive
industry (Afonso and Borquez, 2002; Almas, 1985; Matthiasson
and Sivik, 1978). The waste water is generally characterized by a
high organic load and a varying salt content (Vandanjon et al.,
2002). At the same time, there is a huge potential for recovery of
valuable compounds of marine origin and make-up water from
the waste fractions. This is important for the (1) better utilization
of valuable marine compounds and new value-added by-products,
(2) reduction in raw material consumption for improved process
economy and (3) reduction of the environmental impact of food
production.
A technology which can reduce water consumption in water-
intensive industries is membrane separation (Afonso and
Borquez, 2002; Almas, 1985; Matthiasson and Sivik, 1978).
Additionally, membrane technology can offer separation/recovery
of particles and molecules in very specific ranges making it very
interesting for byproduct separation.
Microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) and
electrodialysis are already seen as established technologies as dis-
cussed by Galanakis (2012), but are sensitive to fouling due to the
nature of the raw material (Galanakis, 2012). Several studies have
reported the use of MF, UF, NF and reverse osmosis (RO) for
⇑ Corresponding author. Tel.: +45 6550 7479.
E-mail address: lfj@kbm.sdu.dk (L. Fjerbæk Søtoft).
industry waste water streams (Dumay et al., 2008; Ferjani et al.,
2005; Matthiasson and Sivik, 1978; Li et al., 2006, 2008; Perez-
Galvez et al., 2011; Stine et al., 2012; Vandanjon et al., 2002, 2009).
Because membrane processes are carried out at a relative low
temperature, they offer an improved preservation of the concen-
trated compounds such as proteins compared to traditional ther-
mal or chemical processes (Dumay et al., 2008). Matthiasson and
Sivik (1978) were the first to demonstrate the usefulness of UF
and RO for processing various waste waters from herring process-
ing including spent herring marinade with 15–22 wt% salt. The aim
of that study was to recover a protein concentrate and reduce the
organic load in the waste water. The waste water COD (Chemical
Oxygen Demand) load was reduced by up to 97% and protein
was concentrated up to 10 wt% (Matthiasson and Sivik, 1978).
These authors suggested combining membrane filtration with
evaporation. Membrane filtration would then remove 65–90% of
the water and make a 15–20 wt% commercial protein concentrate,
which then by evaporation could be processed into 30–40 wt% pro-
tein (Matthiasson and Sivik, 1978).
Ceramic NF membranes (1 kDa) were used by Afonso and
Borquez (2003) to concentrate protein from fish meal waste water.
An economical assessment of a 10 m3/h fish meal waste water treat-
ment plant showed a rate of return of 17% and a feasible process
(Afonso et al., 2004). Since then, there has been a huge development
within membrane processes, an increased interest in value-added
marine products and a rise in environmental awareness.

http://dx.doi.org/10.1016/j.jfoodeng.2015.02.020
0260-8774/Ó 2015 Elsevier Ltd. All rights reserved.
40 L. Fjerbæk Søtoft et al. / Journal of Food Engineering 158 (2015) 39–47
It is expected that membrane fractionation of e.g. protein hydro-
lysates is not characterized by a sharp cut-off. Bourseau et al. (2009)
observed this in the fractions obtained after a UF (4000 Da cut-off)
and NF (300 Da cut-off) separation sequence of two fish protein
hydrolysates. Size-exclusion chromatography showed how larger
molecules were retained in the retentate and the smaller molecules
were present in the permeate, but with a mixed retention in the
intermediate molecular weight (MW) region (Bourseau et al.,
2009). Beaulieu et al. (2009) fractionated a herring hydrolysate with
similar results. A 50 kDa retentate had the highest total amino acid
content (74% dry matter), while RO retentate contained most
minerals (42% dry weight). The used stages were 0.3 lm, 50 kDa,
10 kDa, 1 kDa, 200 Da (NF) and <200 Da (RO).
The observed MW cut-off for a specific process is a combination
of the chemical and filtration properties of the membrane, sec-
ondary layer(s) and feed, which can cause differences between
nominal and observed cut-offs. Rejection of bovine serum albumin
by a 150 kDa ceramic membrane has been found to depend on pH
and whether or not 10 wt% NaCl is in the model solution. With
10 wt% NaCl and pH 9.0, the lowest rejection of 96.5% is obtained
while at lower pH (4.8 and 6.8) rejection is >99% also when NaCl
is present (Kuca and Szaniawska, 2009).
The presence of salts can reduce the value of protein or amino
acids concentrate, but diafiltration can be used to purify the con-
centrates even more. Diafiltration is a membrane process where
new solvent is added to an existing concentrate. The smaller
permeating molecules such as sugars or salts are then washed
away during a filtration as permeate, while the largest molecules
for instance proteins are retained and purified by the membrane.
This can be carried out as a continuous or batch process and can
be a way to remove smaller molecules from a retentate, when a
higher concentration of the largest retained molecules is of interest
as done by Taheri et al. (2014) on herring brine.
As organisms of marine origin are adapted to an environment
very different from the terrestrial, they produce numerous inter-
esting compounds such as pigments, proteins, polysaccharides
and lipids. Additionally, the market and interest in marine nutra-
ceuticals are growing significantly (Rasmussen and Morrissey,
2007). Examples could be marine proteases (Bougatef, 2014) and
other bioactive peptides (Picot et al., 2010).
The main aim of the present study is to recover valuable frac-
tions from a fish industry waste and characterize the properties
and potential use of each fraction. Additionally, the purpose is
to reduce the amounts of saline waste water with high organic
content discharged from the plant. This work is unique in the
number of product fractions (six consecutive membrane stages)
and the scale of the experiments with a starting volume of
120 L of spent herring marinade. The multiple objectives of this
work can be specified as (1) recovery of organic particles, (2)
recovery of proteins and amino acids, (3) recovery of fats, (4)
recovery of sugars, (5) recovery of water and chemicals of suffi-
cient quality for reuse and (6) volume reduction of waste for fur-
ther treatment.
2. Material and methods
2.1. Materials
A fresh sample of spent herring marinade (120 L) was pro-
vided by the Danish fish food producer LAUNIS Fiskekonserves
A/S. The sample was kept at 3–7 °C. Before addition of the herring,
the marinade is composed of water, acetic acid and salt. Thus, fish
meat residues, fat, proteins, peptides and free amino acids can be
expected to be present in the marinade in addition to acetic acid
and salt after marinating the herring under anaerobic conditions
and subsequent removal of the fillets.
An inspection of the marinade showed visible fish meat resi-
dues and a fat layer. This had to be removed prior to any filtration
to protect the membranes. Due to the low temperature, the fat
solidifies on top of the liquid and can be removed efficiently by
skimming (no visual oil droplets left).
Chemicals (acetonitrile: HiPerSolv Chromanorm from VWR,
PROLAB and water) for HPLC were of UV-grade, whereas cleaning
chemicals for the membrane setups (citric acid and NaOH, VWR)
were of food grade quality. Water used for membrane cleaning
was ion-exchange quality (conductivity below 10 lS cm1).
2.2. Methods
2.2.1. Sieving and membrane filtrations
A series of membrane filtrations with various pore sizes has
been carried out. Sieving is used as pretreatment and has been
done as manual batch sieving with dead-end sieves. Their charac-
teristics can be seen in Table 1.
The cross flow membrane unit used is a Labstack M20 for flat-
sheet membranes (Alfa Laval). The retentate is recycled to the feed
tank in order to concentrate the retentate. The setup is equipped
with a feed pump (Hydracell), an inline feed heat exchanger, two
feed side pressure gauges and a pressure control valve on the
retentate side. Feed flow rate was controlled by adjustment of
the pump speed. Permeate is collected separately during filtra-
tions. A weight (Dansk Vægt Industry A/S, 0–35 kg) was used to
measure the permeate flow rate. Temperature was controlled by
connecting a cooling unit (Heto HMT 200) to the heat exchanger.
Additionally, the feed tank was cooled by insertion into a cooling
tank during NF to reduce the temperature increase during opera-
tion (Fig. 1). During each run, only one type of membrane was used
at a time. The fractions and samples are cooled immediately at 3–
7 °C after a filtration. The membranes used in each run can be seen
in Table 2 specified with either pore size, molecular weight cut-off
(MWCO) or rejection.
The volume reduction (VR) for a specific fraction is calculated
as:
VR ¼ Concentrate volume=Initial volume ð1Þ
2.2.2. Dry matter and ash content
Dry matter (DM) and ash content measurements were per-
formed at 105 °C and 550 °C, respectively. The analyses were done
in triplicate according to DS 204:1980.
2.2.3. Protein quantification by absorbance at 280 nm
Total protein concentration in the different marinade fractions
was determined by measuring the absorbance at 280 nm (Beaven
and Holiday, 1952; Layne, 1957). For unknown complex protein
mixtures, it is commonly accepted that 1 absorbance unit at
280 nm equals 1 mg/mL protein (light path length of 1 cm). The
fractions were diluted in pure marinade solution (9.0 wt% NaCl,
2.0 wt% acidic acid, pH 4.15) to obtain absorbance values in the
range 0.1–0.7. Subsequently, absorbance at 280 nm (A280) was
measured for 1 mL samples using the pure marinade solution as
reference. Dilutions and A280 measurements were performed in
triplicates. Finally, the measurements were corrected for the dilu-
tion factor to obtain absolute A280 values for the fractions.
Table 1
Sieves used for pretreatment of the herring marinade.
Sieve area (mm) Material Supplier Mesh size (mm)
200  50 Stainless steel Retsch 0.5
200  25 Stainless steel Retsch 0.18
201  25 Stainless steel Retsch 0.145
202  25 Stainless steel Retsch 0.045
 L. Fjerbæk Søtoft et al. / Journal of Food Engineering 158 (2015) 39–47 41

Feed tank Valve

T
P2

Permeate tank
P1

Pump HX Membrane Weight

Fig. 1. Experimental setup (HX: heat exchanger, T: temperature, P1 and P2: retentate pressure gauges).

Table 2
Flat sheet membranes used for herring marinade fractionation. Table 3
Initial chemical composition and characteristics of herring marinade.
Process Membrane Material Pore size/MWCO/rejection
type Analysis Value

MF Alfa Laval- Fluoro polymer 0.2 lm Dry matter 10.6 wt%

MFP2 Proteina 3.9 wt%
UF50 DSS-GR51PP Polysulphone 50 kDa Fat 0.22 wt%
UF20 Alfa Laval- Fluoro polymer 20 kDa pH 4.36
FS61PP a
Measured with Dumas method for nitrogen determination.
UF10 DSS- Composite fluoro 10 kDa
ETNA10PP polymer
UF01 Alfa Laval- Fluoro polymer 1 kDa
ETNA01PP in order to plan the membrane fractionations accordingly. The ini-
NF Alfa Laval-NF Polyamide on Reject >98% MgSO4 tial chemical composition and characteristics of the marinade can
polyester (2000 ppm, 9 bar, 25 °C) be seen in Table 3 and Fig. 2.
An initial SDS–PAGE analysis of the untreated herring marinade
2.2.4. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis has been performed to characterize the sizes of the proteins pre-
(SDS–PAGE) sent. The results can be seen in Fig. 2. It shows a few predominant
The protein composition of the herring marinade fractions from bands in the range of 8–20 kDa and the presence of some proteins
the different sieving and filtration steps were analyzed by SDS– with a molecular weight of 25 kDa. Furthermore, several proteins
PAGE. Direct electrophoretic analyses of the raw fractions failed of various sizes between 30 kDa and 150 kDa are present in the
probably due to the high concentrations of salt and lipids in the marinade. Most of these bands are likely the result of proteolytic
samples. Therefore, the polypeptides were prepared for SDS–
PAGE by methanol–chloroform precipitation according to Wessel
and Flügge (1984). After precipitation of the proteins the air-dried
pellets were dissolved in SDS–PAGE sample buffer containing a final
concentration of 5 mM dithiothreitol. SDS–PAGE was performed
using hand-cast 4–16% gradient gels (7.3 cm  8.3 cm  1 mm)
and the SDS–PAGE Buffer System (Laemmli, 1970). The separated
proteins were visualized by Coomassie Brilliant Blue staining.
Relative quantification of three distinct and abundant protein
bands migrating at about 17 kDa (band a), 47 kDa (band b), and
64 kDa (band c) were achieved by densitometry using Adobe
Photoshop CS5.1 software (version 12.1  32) and according to Xu
et al. (2010). The staining intensities were normalized in relation
to the 28.5 kDa molecular marker band. Methanol–chloroform pre-
cipitation, separation by SDS–PAGE, and quantification by densito-
metry were performed in triplicates.

2.2.5. High performance liquid chromatography (HPLC)

Qualitative analyses for ethanol, glycerol, fructose, glucose, lac-
tose and sucrose were done by isocratic HPLC (Agilent series 1100)
using a Luna 5 lm NH2 100 Å 250  4.60 mm column with 65:35
(vol%) acetonitrile:H2O as eluent. The column temperature was
25 °C, injected sample size 5 lL, the eluent flow 0.5 mL/min and
the analysis time was 20 min. The analyses were done in duplicate.

3. Results and discussion

3.1. Initial characterization of herring marinade

The initial composition of the marinade prior to addition of fish Fig. 2. SDS–PAGE of the herring marinade. Left lane, the molecular weight marker
fillets is 9.0 ± 1.0 wt% NaCl, 2.0 ± 0.3 wt% acetic acid and a pH of which indicates masses in kDa. Right lane, herring marinade sample. Arrows
4.0–4.3. The herring marinade has additionally been characterized indicate the nominal MWCOs for the membranes used in the study.
42 L. Fjerbæk Søtoft et al. / Journal of Food Engineering 158 (2015) 39–47

Herring marinade Table 4

Process characteristics during filtrations of herring marinade fractions.
Fat fraction
Process Initial flux End flux Crossflow TMP (bar) Filtration
(kg/(m2h)) (kg/(m2h)) (kg/min) time (min)
MF 48 20 5 0.9 330
Sieving Fish meat UF50 95 30 5 7.8 215
UF20 250 150 5.5–6.6 5.9–6.0 45
UF10 120 80 6.5 5.4–6.3 70
UF01 35 5 6.5 4.2–6.7 410
NF 16 4 6.5 26.4–29.1 450

MF Major particulates
As pretreatment to reduce membrane fouling, solidified fat is
removed upon cooling and then the marinade is sieved to remove
visible fish meat particles. The sieving retained meat and proteins
Proteins on the sieves (visual inspection). Even though the marinade is still
UF50
turbid after the sieving, the marinade is processed by MF as the ini-
tial membrane filtration. Several UF MWCOs are chosen in order to
investigate the characteristics of the individual fractions. This is
Proteins specifically interesting for the protein/peptide content as it has a
UF20 very broad variation in molecular mass compared to e.g. free
amino acids. The permeate from the higher MWCO is then used
as feed for the following MWCO. Process characteristics can be
seen in Table 4.

UF10 Proteins
3.2.1. Microfiltration
MF has been carried out as a constant pressure filtration with a
transmembrane pressure (TMP) of 0.9 bar at room temperature.
Retentate is returned to the feed tank to further concentrate the
retained fraction. The final concentrate and permeate are then
UF01 Proteins
undergoing subsequent analysis. The permeate is feed for the fol-
lowing UF50 (Fig. 3). The results of the analyses are covered in
Section 3.3.
Sugars
amino acids An intermediate cleaning is performed with NaOH at pH 9,
NF Peptides before new feed was added to complete the treatment of all the
marinade. Clean water flux data before and after the cleaning show
that the membrane’s clean water flux of initial 190 kg/(m2h) at
0.9 bar decreases to 60 kg/(m2h) after MF and is restored to
Saline water 145 kg/(m2h) after the cleaning. This means that 75% of the initial
for reuse
flux is restored.

Fig. 3. Fractionation sequence. The number following UF indicates MWCO in kDa,

where i.e. UF50 is UF with 50 kDa MWCO.
degradation of larger protein fragments as shown by Andersen
et al. (2007). They observed significant proteolytic activity both
in the herring and brine during the ripening process leading to
fragmentation of major proteins such as actin (42 kDa) and myosin
(200 kDa), appearance of smaller proteins and peptides, and
smeared bands indicating heterogeneous proteolytic events.
Based on the initial characterization of the marinade and the
range of available commercial membranes a plan for the sequence
of membrane filtrations and subsequent characterization of the
fractions were developed (Fig. 3). These fractionations serve as a
basis to identify an optimized number of fractions for further eco-
nomic and qualitative analysis.
The initial aim is to separate valuable fractions of fats, fish meat
and organic particles, proteins, peptides and amino acids as well as
chemicals and water of sufficient quality for reuse combined with a
volume reduction of waste for further treatment.
3.2. Fractionation of herring marinade
A sequence of membranes with different MWCOs has been used
to get a broad array of fractions with different molecular size
ranges for later characterization.
3.2.2. Ultrafiltration
The permeate generated in the MF step is then fractionated
with UF. Four UF MWCOs have been used in series: 50 kDa,
20 kDa, 10 kDa, and 1 kDa. The overall observation for UF is a fast
initial flux decline followed by a more steady performance for the
later part of the filtration.
Due to the high number of stages, the feed for the UF processes
can be considered as efficiently pre-treated, but a high initial flux
decline is still observed for herring marinade filtrations. This is
especially severe for UF50 and UF01, while not as severe for
UF10 and UF20. This is caused by the higher amount of retained
materials during UF50 and UF01 (see Section 3.3) compared to
UF20 and UF10. A 0.2 lm MF pretreatment and subsequent UF10
filtration of herring marinade showed a large flux decline in the
UF10 stage and a 10 times lower flux (Lyder, 2014) than obtained
in this work where intermediate UF50 was done. An intermediate
stage of UF50 is therefore an option to increase the UF10 flux.
The fast initial decline can be due to accumulation of molecules
at the membrane surface, which increases local osmotic pressure
and/or builds a fouling layer and hence reduces the effective driv-
ing force. Clean water flux measurements after the filtrations show
83% recovery of the initial flux, so the effect is mainly reversible.
UF10 showed a higher initial clean water flux recovery compared
to the recovery of UF01, while the lowest clean water flux recov-
eries were seen for UF20 and UF50. This might be due to
 L. Fjerbæk Søtoft et al. / Journal of Food Engineering 158 (2015) 39–47
interaction with the membrane material as UF20 and UF01 are
hydrophobic fluoro polymer membranes, while the UF10
membrane is a composite hydrophilized fluoro polymer. It is
well-known that the membrane material can affect adsorption of
molecules on the membrane surface (Beier et al., 2007). Dumay
et al. (2008) filtered a surimi waste water to recover proteins
and lipids. Fouling was less severe for a hydrophilic regenerated
cellulose membrane (10 kDa MWCO) compared to some more
hydrophobic and open (40 kDa MWCO) membranes (polyethersulfone,
polyvinylidenefluoride and polyacrylonitrile). The flux was also
higher for the hydrophilic membrane during 5 h of waste water fil-
tration. Clean water rinsing was sufficient to recover the initial flux
for the regenerated cellulose, where the hydrophobic membrane’s
flux could not be restored neither by caustic nor acidic cleaning.
Based on the discussion above, a hydrophilic UF membrane
seems to be less prone to fouling in this case compared to more
hydrophobic membranes, but still appropriate pretreatment is
vital.
3.2.3. Nanofiltration
NF is performed without a clear initial flux decline. The overall
flux is low however and the process is terminated when the flux
decreases below 5 kg/(m2h). At this point, a severe amount of
white, floating precipitate indicating protein denaturation and
aggregation is also observed in the feed tank, so no further concen-
tration is carried out with NF.
As the NF permeate contains 8.47 wt% ash equivalent to an
osmotic pressure of 72.6 bar (Rao and Rizvi, 1995) RO is not a
viable process for further waste water concentration.
3.3. Characterization and evaluation of the membrane fractions
After removal of the fat, the fractions have been analyzed with
respect to DM and ash content (Fig. 4). The herring marinade con-
tains 13.8% DM and 10% ash after fat removal, which does not
change to a large degree until the NF stage. Here, the DM changes
from 13.8% in the NF feed to 9% in the NF permeate and to an end
retentate concentration of 20.8% DM in the batch filtration. The big
difference in DM from NF retentate to NF permeate can be due to
rejection of acetic acid.
Only minor differences in DM are seen for sieving and MF even
though visual inspection on sieve surface shows bigger fish meat
particles. This is however because of a high content of dissolved
dry matter and a relative smaller content of larger, visible particles.
As expected, only minor differences in the ash content are seen
prior to NF. The ash content is reduced to 8.37% in the NF
25.0
20.0
Content (wt%)
15.0
10.0
5.0
0.0
Fraction
Fig. 4. Dry matter and ash content in herring marinade and filtration fractions. The histogram shows the dry matter content (dark gray columns) and ash content (light gray
columns) of fresh herring marinade (after fat removal), marinade after sieving, and collected permeate (P) and retentate (R) samples during processing. The values are shown
in percent (%) of the total weight of the wet sample and with standard deviation based on triplicates.
43
permeate. This is to be expected since much of the ash is dissolved
NaCl which will pass nearly unhindered through sieves, MF, UF and
NF membranes.
3.4. Characterization of the protein content and composition
3.4.1. Quantification of total amino acid by absorbance at 280 nm
Aromatic amino acids exhibit absorbance at 280 nm (A280) and,
therefore, A280 measurements give a rough estimate of the total
amino acid content including proteins, peptides and free amino
acids. Except for a small increase for the UF50 retentate, the
A280 method revealed relatively constant total amino acid concen-
trations in the sieving, MF and UF retentates and permeates
(Fig. 5). Thus, the polypeptides present in the herring marinade
are basically not retained by membranes with MWCOs down to
1 kDa (UF01). However, according to the A280 measurements the
NF step exhibits high rejection of the total amino acids. While
the NF permeates at 200 min and 90 min have very small total
amino acid concentrations, the corresponding retentates have sig-
nificantly higher total amino acid concentrations compared to the
preceding filtration steps. Since the A280 spectrophotometric
analyses measure the total content of aromatic amino acids, the
results do not distinguish between free aromatic amino acids and
those in polypeptides. Therefore, SDS–PAGE is performed to inves-
tigate the protein composition of the different marinade fractions
in more detail.
3.4.2. Determination of the protein composition by SDS–PAGE
The protein composition of the different membrane filtration
fractions was studied by SDS–PAGE analyses under reducing condi-
tions (Fig. 6A). To evaluate the membranes’ ability to retain pro-
teins of different molecular weight, the relative intensity of three
distinct protein bands migrating at 17 kDa (band a, Fig. 6A and
B), 47 kDa (band b, Fig. 6A and C), and 64 kDa (band c, Fig. 6A
and D) were determined and compared between the fractions.
The 10–20 kDa area has the highest intensity as seen by SDS–
PAGE (Fig. 6A–D). For all three proteins, the concentration is sig-
nificantly reduced and increased in the UF50 permeate and reten-
tate, respectively. The results also indicate that the retention in the
UF50 step is more predominant for the larger proteins migrating at
47 kDa and 64 kDa than for the smaller 17 kDa protein (Fig. 6C and
D). Thus, the 47 kDa and 64 kDa proteins are almost absent in all
the fractions following the UF50 step, while some 17 kDa protein
is still present in the filtration fractions obtained with smaller
cut-off values as expected based on the MWCO. As an example,
the results indicate that the 17 kDa protein is to some degree
%DM %ASH
44 L. Fjerbæk Søtoft et al. / Journal of Food Engineering 158 (2015) 39–47
Fig. 5. Total amino acid concentration (proteins, peptides, and free amino acids) of the raw marinade and separated fractions determined by absorbance at 280 nm. The
amino acid concentration of the raw marinade and most of the fractions is close to 20 mg/mL. However, fractions UF50 R, NF R 200 min, and NF R 90 min have significant
higher protein concentrations while those of fractions NF P 200 min and NF P 90 min are very low.
present after UF01 and retained in the NF step (200 min and
90 min). Thus, our results show that the UF50 membrane to a high
degree rejects the proteins in the herring marinade, but that partly
depends on the molecular weight of the protein. This observation
differs from what is observed for the total amino acid concentra-
tion, where the retention of UF50 is not as high (Fig. 5). This is
explained by the fact that the majority of the aromatic amino acids
are present as polypeptides with a molecular weight below 20 kDa
as shown by SDS–PAGE (Fig. 6A) or as free amino acids and they
dominate the absorption signal from the total amino acid concen-
tration analysis.
The UF50 step is clearly the most efficient step in retaining the
largest proteins, also those of molecular weight lower than 50 kDa
though this is the MWCO of the membrane. This indicates that the
herring proteins either have lost their native and compact three-di-
mensional folds or somehow form non-covalent higher molecular
weight structures. Furthermore, it can be explained by a secondary
cake layer formation or partial pore blocking lowering the actual
MWCO of the membrane.
The A280 remains relatively constant throughout all the ultra-
filtration steps, but significantly increases in the NF step (Fig. 5).
Thus, some small molecules with absorbance at 280 nm are not
isolated in the UF steps, but are very efficiently retained by the
NF membrane which has a reported cut-off of 159 Da (Uzal et al.,
2010). Free amino acids have molecular weights from 75 to
204 Da. Most likely, the absorbance in the NF retentate is caused
by short peptides and free aromatic amino acids making this NF
fraction interesting as an amino acid source.
In conclusion, the UF50 membrane is sufficient to isolate and
concentrate the majority of proteins present in the herring mari-
nade while the NF membrane retains free amino acids/very small
peptides. Therefore, the UF50 retentate is a potential protein-rich
source content while the NF retentate is a promising fraction with
reference to high concentrations of free amino acids. Noteworthy,
the fractions have volumes 30 times (Retentate UF50) and 11 times
(Retentate NF) smaller than the original marinade waste water
(Table 5).
3.5. Further characterization
Another fraction of interest is the fat fraction. The herring fat is
a valuable source for x3 fatty acids and previous analyses from the
company have shown herring fat composition of 17–19 wt% x3
and 3 wt% x6 fatty acids giving a ratio between x3/x6 of 5.6–
6.3 (data not shown). The x3 fatty acids were mainly eicosapen-
taenoic acid (EPA) and docosahexaenoic acid (DHA) and to a minor
degree alpha-linolenic acid (ALA). The global market of x3 ingredi-
ents is expected to grow from around $1.56 billion in 2010 to $4
billion in 2018 (Transparency Market Research, 2012) as most
modern food products in e.i. Europe and US hold too much x6 over
x3 (Simopoulos, 2002). Dietary recommendations are ratios of 0.1
and 0.2 from WHO and the Federation of European Nutritional
Societies, respectively (DACH, 2000; Mahan and Scott-Stump,
2000).
Of the 120 L herring marinade, 3 L fat was collected. The size of
this fraction will though vary greatly with seasons (Henderson and
Almatar, 1989; Aro et al., 2000), so the actual available amount of
this fraction and hence the economy of recovery must be estimated
with care. However, this fraction can cause clogging and fouling at
low temperature in downstream equipment and as this fraction
will have high potential for sale, the removal must be considered
to improve economy, ease of operation and reduce waste volumes.
HPLC analyses (data not shown) revealed big differences
between NF retentate and permeate. The NF retentate showed
presence of sugars and glycerol, where the NF permeate only con-
tained minor amounts of glycerol, but no sugars of significance.
The glycerol is possibly formed by acid-catalyzed hydrolysis of
triglycerides forming free fatty acids and glycerol during the her-
ring curing process. This conclusion is supported by Andersen
et al. (2007) as oxidation of lipids into free fatty acids in herring
brine has been observed with time. Thus, Andersen and colleagues
reported an increase in the level of free fatty acids in herring brine
from an initial 20–30% to 57% during ripening (Andersen et al.,
2007).
3.6. Potential use of obtained fractions
A volume balance has been established based on the fractiona-
tion of the incoming 120 L herring marinade (see Table 5). Out of
the 120 L waste marinade, the amount of waste water which needs
further treatment is reduced to 44 L. The water for reuse (permeate
NF) is 50 L, which can substitute fresh water, NaCl and acetic acid
at the fish food producer. The NF permeate has high purity with
low presence of proteins, amino acids and sugars as well as no
smell from sulfur-containing compounds. Salt and acetic acid are
present in amounts close to the original content. It is therefore
 L. Fjerbæk Søtoft et al. / Journal of Food Engineering 158 (2015) 39–47 45

B 5.0
4.5
4.0
Relative intensity

3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0

Fraction

5.0
C
4.5
4.0
Relative intensity

3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0

Fraction

5.0
D
4.5
4.0
Relative intensity

3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0

Fraction

Fig. 6. SDS–PAGE analyses of herring marinade fractions. (A) The fractions obtained by sieving and different filtration steps including MF, UF, and NF were analyzed by SDS–
PAGE. Proteins migrating at 17 kDa (band a), 47 kDa (band b), and 64 kDa (band c) were quantified by densitometry. (B) Relative quantification of the 17 kDa protein (band a)
for the different fractions. (C) Relative quantification of the 47 kDa protein (band b) for the different fractions. (D) Relative quantification of the 64 kDa protein (band c) for the
different fractions. Error lines in the histograms indicate standard deviation of three replicates.
46 L. Fjerbæk Søtoft et al. / Journal of Food Engineering 158 (2015) 39–47
Table 5
Volumes and potential use of obtained membrane filtration fractions.
Incoming Process
120 L marinade Fat removal
Sieving and MF
Retentate UF50
Retentate NF
Permeate NF
Retentate UF20, UF10 and UF01
Saline waste watera
a
Here, saline waste water accounts for volume not collected during our procedure, but feed and dead volumes, etc. not processed completely and hence not included in the
fractions above.
believed that the quality of the NF permeate makes it highly suited
for reuse to replace fresh marinade and avoid discharge of 42% of
the original amount of waste alone.
Two fractions are of interest for protein (4 L UF50 retentate) and
smaller peptides/amino acids (11 L NF retentate). Daily intake of
proteins is vital for both humans and animals, where fish diets
typically consist of 20–55% crude protein (Ayadi et al., 2012).
Furthermore, amino acids are the largest feed additive group with
a global market expected to grow from $9.6 billion in 2010 to $18.8
billion in 2017 (BCC Research, 2012).
If required, the smaller molecules such as sugars and salts can
probably be removed from these fractions rich in proteins/peptides
and free amino acids using diafiltration. However, this has yet to be
investigated.
Furthermore, 3 L fat and 7 L fish meat can be raw material for
food, biogas or fodder.
In Denmark, the waste water treatment cost at municipal waste
water treatment plants is 4 €/m3. Our membrane fractionation will
have a profit of 39 €/m3 incoming waste water due to sales of pro-
teins and saving waste water treatment and fresh water intake.
This is based on prices of 1.5 €/kg 70 wt% protein, 0.07 €/kg NaCl,
0.45 €/kg acetic acid and 4 €/m3 fresh water with no operational
or capital costs included.
4. Conclusions
In the present work, membrane filtration has been applied to
separate and recover promising sources of value-added com-
pounds such as proteins, peptides and amino acids from spent her-
ring marinade. Among six consecutive membrane stages studied
(i.e. range from 0.2 lm MF, 50 kDa UF, 20 kDa, 10 kDa and 1 kDa
to NF), the stages of highest interest are 50 kDa UF and NF, but
the pretreatment by sieving and MF is vital for the performance
of the latter stages. The 50 kDa UF stage produces retentate con-
centrated in proteins (>17 kDa), while the NF stage produces a
retentate containing sugars, amino acids and smaller peptides. Of
the incoming marinade, 42% (the NF permeate) containing salt
and acetic acid and is ready for reuse to substitute fresh marinade.
The proteins in the spent marinade are concentrated 30 times
and the amino acids and smaller peptides are concentrated 11
times compared to the original amount of spent marinade. The
amount of waste water which needs further treatment is reduced
to 37.5%. The membrane treatment will earn 39 €/m3 incoming
waste water saving chemicals, fresh water, and selling protein
compared to a treatment cost of 4 €/m3 without membrane
fractionation.
Hence, the separation has been carried out successfully and the
important membrane stages have been identified (i.e. 50 kDa UF
and NF), but the processes need further investigations with respect
Product fractions Potential
3L Source of x3 and x6, biogas or fodder
7L Fish meat and major particulates for biogas or fodder
4L Proteins
11 L Sugars, amino acids and small peptides
50 L Water for reuse
16 L Further treatment
29 L Further treatment
to economy, price of the obtained fractions and long-term mem-
brane fouling.
Acknowledgements
This work is a part of the research within the Innovation
Consortium Natural Ingredients and Green Energy (NIGE)-with sus-
tainable purification technologies financially supported by Danish
Agency for Science Technology and Innovation to whom the
authors are indebted. The authors want to thank LAUNIS
Fiskekonserves A/S, Industrivej Nord 2 – 4, DK-9982 Aalbæk,
Denmark, for providing the herring marinade samples and insight
into the herring curing industry. In addition, Danish Technological
Institute, Kongsvang Allé 29, DK-8000 Aarhus C, Denmark, is
acknowledged for initial SDS–PAGE and fatty acid analysis.
References
Afonso, M.D., Borquez, R., 2002. Review of the treatment of seafood processing
wastewaters and recovery of proteins therein by membrane separation
processes – prospects of the ultrafiltration of wastewaters from the fish meal
industry. Desalination 142, 29–45.
Afonso, M.D., Borquez, R., 2003. Nanofiltration of wastewaters from the fish meal
industry. Desalination 151, 131–138.
Afonso, M.D., Ferrer, J., Borquez, R., 2004. An economic assessment of proteins
recovery from fish meal effluents by ultrafiltration. Trends Food Sci. Technol. 15,
506–512.
Almas, K.A., 1985. Applications of cross-flow membrane technology in the fishing
industry. Desalination 53, 167–180.
Andersen, E., Andersen, M.L., Baron, C.P., 2007. Characterization of oxidative
changes in salted herring (Clupea harengus) during ripening. J. Agric. Food
Chem. 55, 9545–9553.
Aro, T., Tahvonen, R., Mattila, T., Nurmi, J., Sivonen, T., Kallio, H., 2000. Effects of
season and processing on oil content and fatty acids of baltic herring (Clupea
harengus membras). J. Agric. Food Chem. 48 (12), 6085–6093.
Ayadi, F.Y., Rosentrater, K.A., Muthukumarappan, K., 2012. Alternative protein
sources for aquaculture feeds. J. Aquacult. Feed Sci. Nutr. 4 (1), 1–26.
BCC Research, 2012. Market Research Reports and Technical Publications. <http://
www.bccresearch.com> (retrieved 20.07.14).
Beaulieu, L., Thibodeau, J., Bryl, P., Carbonneau, M.E., 2009. Proteolytic processing of
herring (Clupea harengus): biochemical and nutritional characterisation of
hydrolysates. Int. J. Food Sci. Technol. 44, 2113–2119.
Beaven, G.H., Holiday, E.R., 1952. Ultraviolet absorption spectra of proteins and
amino acids. Adv. Protein Chem. 7, 319–386.
Beier, S., Enevoldsen, A.D., Kontogeorgis, G.M., Hansen, E.B., Jonsson, G., 2007.
Adsorption of amylase enzyme on ultrafiltration membranes. Langmuir 23,
9341–9351.
Bougatef, A., 2014. Trypsins from fish processing waste: characteristics and
biotechnological applications – comprehensive review. J. Cleaner Prod. 57,
257–265.
Bourseau, P., Vandanjon, L., Jaouen, P., Chaplain-Derouiniot, M., Massé, A., Guerard,
F., Chabeaud, A., Fouchereau-Peron, M., Le Gal, Y., Ravallec-Ple, R., Berge, J.P.,
Picot, L., Piot, J.M., Batista, I., Thorkelsson, G., Delannoy, C., Jakobsen, G.,
Johansson, I., 2009. Fractionation of fish protein hydrolysates by ultrafiltration
and nanofiltration: impact on peptidic populations. Desalination 244, 303–320.
DACH, 2000. Referenzwerte für die Nährstoffzufuhr. Frankfurt am Main, Umschau
Braus GmbH.
Dumay, J., Radier, S., Barnathan, G., Berge, J., Jaouen, P., 2008. Recovery of valuable
soluble compounds from washing waters generated during small fatty pelagic
surimi processing by membrane processes. Environ. Technol. 29, 451–461.
 L. Fjerbæk Søtoft et al. / Journal of Food Engineering 158 (2015) 39–47
Ferjani, E., Ellouze, E., Ben Amar, R., 2005. Treatment of seafood processing
wastewaters by ultrafiltration-nanofiltration cellulose acetate membranes.
Desalination 177, 43–49.
Galanakis, C.M., 2012. Recovery of high added-value components from food wastes:
conventional, emerging technologies and commercialized applications. Trends
Food Sci. Technol. 26 (2), 68–87.
Henderson, R.J., Almatar, S.M., 1989. Seasonal changes in the lipid composition of
herring (Clupea Harengus) in relation to gonad maturation. J. Mar. Biol. Assoc.
U.K. 69 (2), 323–334.
Kuca, M., Szaniawska, D., 2009. Application of microfiltration and ceramic
membranes for treatment of salted aqueous effluents from fish processing.
Desalination 241, 227–235.
Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the
head of bacteriophage T4. Nature 227, 680–685.
Layne, E., 1957. Spectrophotometric and turbidimetric methods for measuring
proteins. Methods Enzymol. 3, 447–454.
Li, Z., Kittikun, A., Youravong, W., 2008. Removal of suspended solids from tuna
spleen extract by microfiltration: a batch process design and improvement.
Biochem. Eng. J. 38, 226–233.
Li, Z., Youravong, W., Kittikun, A., 2006. Separation of proteases from yellowfin tuna
spleen by ultrafiltration. Bioresour. Technol. 97, 2364–2370.
Lyder, N., 2014. Fractionation of Value Added Components from Herring Pickle.
Master thesis, University of Southern Denmark.
Mahan, K.L., Scott-Stump, S., 2000. Nutricion y dietoterapia, de Krause. Mcgraw-Hill
Interamericana, Madrid, Spain.
Matthiasson, E., Sivik, B., 1978. New methods of recovering fish processing waters.
Nya möjligheter att tillvarata fiskindustrins vatten. Livsmedelsteknik 20 (5),
241–244.
Perez-Galvez, R., Guadix, E.M., Berge, J.P., Guadix, A., 2011. Operation and cleaning
of ceramic membranes for the filtration of fish press liquor. J. Membr. Sci. 384,
142–148.
Picot, L., Ravallec, R., Fouchereau-Peron, M., Vandanjon, L., Jaouen, P., Chaplain-
Derouiniot, M., Guerard, F., Chabeaud, A., Legal, Y., Alvarez, O.M., Berge, J.P., Piot,
47
J.M., Batista, I., Pires, C., Thorkelsson, G., Delannoy, C., Jakobsen, G., Johansson, I.,
Bourseau, P., 2010. Impact of ultrafiltration and nanofiltration of an industrial
fish protein hydrolysate on its bioactive properties. J. Sci. Food Agric. 90, 1819–
1826.
Rao, M.A., Rizvi, S.S.H., 1995. Engineering Properties of Food, second ed. Marcel
Dekker, New York, USA.
Rasmussen, R.S., Morrissey, M.T., 2007. Marine biotechnology for production of food
ingredients. Adv. Food Nutr. Res., 237–292
Simopoulos, A.P., 2002. The importance of the ratio of omega-6/omega-3 essential
fatty acids. Biomed. Pharmacother. 56 (8), 365–379.
Stine, J.J., Pedersen, L., Smiley, S., Bechtel, P.J., 2012. Recovery and utilization of
protein derived from surimi wash-water. J. Food Qual. 35, 43–50.
Taheri, A., Sabeena Farvin, K.H., Jacobsen, C., Baron, C.P., 2014. Antioxidant activities
and functional properties of protein and peptide fractions isolated from salted
herring brine. Food Chem. 142, 318–326.
Transparency Market Research, 2012. Global Omega 3 (EPA/DHA) Ingredients
Market – Industry Analysis, Market Size, Share, Growth and Forecast 2010–
2018. <http://www.transparencymarketresearch.com/global-omega-three-
market.html> (retrieved 20.07.14).
Uzal, N., Yilmaz, L., Yetis, U., 2010. Nanofiltration and reverse osmosis for reuse of
indigo dye rinsing waters. Sep. Sci. Technol. 45, 331–338.
Vandanjon, L., Cros, S., Jaouen, P., Quemeneur, F., Bourseau, P., 2002. Recovery by
nanofiltration and reverse osmosis of marine flavours from seafood cooking
waters. Desalination 144, 379–385.
Vandanjon, L., Grignon, M., Courois, E., Bourseau, P., Jaouen, P., 2009. Fractionating
white fish fillet hydrolysates by ultrafiltration and nanofiltration. J. Food Eng.
95, 36–44.
Wessel, D., Flügge, U.I., 1984. A method for the quantitative recovery of protein in
dilute solution in the presence of detergents and lipids. Anal. Biochem. 138 (1),
141–143.
Xu, C., Meng, S., Liu, X., Sun, L., Liu, W., 2010. Chicken cyclophilin A is an inhibitory
factor to influenza virus replication. Virol. J. 7 (372), 1–11.