Literature review TERI University Ph.D. Thesis, 2007 9 Literature review 3.1 Microfiltration (MF) and ultrafiltration (UF) applications in juice processing MF and UF are pressure driven membrane processes for the selective separation of two or more components from a fluid stream (Cheryan, 1998). The separation is primarily based on size differences with the retention of macromolecules and particles larger than 0.001-0.02μm for UF and 0.1-5μm for MF. The process involves continuous molecular separation without phase transfer; also under ideal conditions, it does not involve the addition of heat or chemicals. Clarification of fruit juice by UF was commercialized in the late 1970‘s (Jönsson and Trägårdh, 1990). Since then, maximum number of UF plants have been established for apple juice clarification; however others like grape, pear, pineapple, cranberry and citrus juices have also been processed. Freshly squeezed apple juice is cloudy due to the presence of protein and tannins, which remain in suspension because of the polysaccharide pectin (Mondor and Brodeur, 2002). Thus large quantities of pectinase and gelatin have to be added to induce clarification. UF not only improves the quality of juice but also increases juice yield by up to 8%. Juice flux can be further enhanced 2-3 fold by appropriate pretreatment with enzymes (Alvarez et al., 1998) and filter aids like bentonite and gelatin (Gökmen and Çetinkaya, 2007). In another recent work with a plate and frame UF system using 50 kD PES membranes, pasteurization alone was reported to result in higher and more stable juice flux over a 20h full scale operation when compared to a full scale tubular UF system with 200 kD membrane (He et al., 2007). The opposite effect was observed with juice heating prior to UF of orange juice using inorganic membranes (Merin and Shomer, 1999). The flux was reduced after heat treatment, possibly due to the interaction of the coagulated pectins and proteins with the membrane-filtering layer. In apple juice clarified by UF, haze formation is a problem because of the polymerization of phenols and its interaction with other components, e.g. proteins (Siebert et al., 1996). In particular, UF through higher molecular weight cut off (MWCO) membranes displayed increased turbidity with time as well as at higher storage temperature (Girard and Fukumoto, 1999). Also, UF with PES/PVP (polyvinylpyrrolidone) membranes exhibited better removal of 3
23
Embed
3.1 Microfiltration (MF) and ultrafiltration (UF ...shodhganga.inflibnet.ac.in/bitstream/10603/6274/9/09_chapter 3.pdf3.2 MF/UF of sugarcane juice 3.2.1 History and current status
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Literature review
TERI University Ph.D. Thesis, 2007 9
Literature review
3.1 Microfiltration (MF) and ultrafiltration (UF) applications in juice processing
MF and UF are pressure driven membrane processes for the selective separation
of two or more components from a fluid stream (Cheryan, 1998). The separation
is primarily based on size differences with the retention of macromolecules and
particles larger than 0.001-0.02μm for UF and 0.1-5µm for MF. The process
involves continuous molecular separation without phase transfer; also under
ideal conditions, it does not involve the addition of heat or chemicals.
Clarification of fruit juice by UF was commercialized in the late 1970‘s (Jönsson
and Trägårdh, 1990). Since then, maximum number of UF plants have been
established for apple juice clarification; however others like grape, pear,
pineapple, cranberry and citrus juices have also been processed.
Freshly squeezed apple juice is cloudy due to the presence of protein and
tannins, which remain in suspension because of the polysaccharide pectin
(Mondor and Brodeur, 2002). Thus large quantities of pectinase and gelatin
have to be added to induce clarification. UF not only improves the quality of
juice but also increases juice yield by up to 8%. Juice flux can be further
enhanced 2-3 fold by appropriate pretreatment with enzymes (Alvarez et al.,
1998) and filter aids like bentonite and gelatin (Gökmen and Çetinkaya, 2007).
In another recent work with a plate and frame UF system using 50 kD PES
membranes, pasteurization alone was reported to result in higher and more
stable juice flux over a 20h full scale operation when compared to a full scale
tubular UF system with 200 kD membrane (He et al., 2007). The opposite effect
was observed with juice heating prior to UF of orange juice using inorganic
membranes (Merin and Shomer, 1999). The flux was reduced after heat
treatment, possibly due to the interaction of the coagulated pectins and proteins
with the membrane-filtering layer.
In apple juice clarified by UF, haze formation is a problem because of the
polymerization of phenols and its interaction with other components, e.g.
proteins (Siebert et al., 1996). In particular, UF through higher molecular weight
cut off (MWCO) membranes displayed increased turbidity with time as well as at
higher storage temperature (Girard and Fukumoto, 1999). Also, UF with
PES/PVP (polyvinylpyrrolidone) membranes exhibited better removal of
3
Literature review
TERI University Ph.D. Thesis, 2007 10
yellowish brown color and polyphenols compared to commercial cellulose
membrane (Borneman et al., 1997).
To obtain high permeate flux at low membrane fouling, pretreatment prior to
UF is usually followed with most juice streams. For instance, a combination of
enzymatic treatment followed by adsorption using bentonite resulted in the
maximum permeate flux with mosambi (Citrus sinensis (L.) Osbeck) juice
ultrafiltered in a continuous stirred filtration cell using 50 kD thin film
composite polyamide membrane (Rai et al., 2007). Enzymatic pretreatment
alone has been employed in several instances. Flux increase was reported with
pectinase pretreatment prior to cross flow UF of West Indian cherry (Malpighia
glabra L.) and pineapple (Ananas comosus (L.) Meer) juice clarification using
100 kD PS hollow fiber and 0.01μm ceramic tubular membranes (De Barros et
al., 2004). In the UF of depectinized kiwi juice, flux was higher with a modified
polyetherether ketone (PEEK) hollow fiber membrane (MWCO>69 kD) when
compared to that obtained with commercial 10-50 kD tubular (PVDF) and
hollow fiber membranes (Tasselli et al., 2007). Enzymatic pretreatment has also
been successful in improving flux during the filtration of pulpy feeds such as in
the MF of pasteurized, diluted umbu (Spondias tuberosa Arr. Cam.) pulp in a
pilot unit equipped with 0.2 m polypropylene tubular membrane (Ushikobo et
al., 2007).
Apart from apple juice, UF has been applied effectively for pear juice
clarification (Kirk et al., 1983) and the process is employed commercially
(www.unipektin.com). These successes have encouraged trials on dark colored
fruit juices such as blackberries, redcurrants, raspberries, sour cherries,
strawberries and elderberries. Here, the permeate must retain its color while
being free from turbidity. In a pilot study on dark colored juices (cherry,
raspberry and redcurrant) with polymeric (PS, PVDF) and ceramic tubular
membranes, ceramic membranes reportedly reduced turbidity without affecting
the color (Bolduan and Lartz, 2002). Further, the permeate from UF ceramic
membranes showed good stability even after hot/cold tests. In another
investigation, blood orange juice was clarified by UF using tubular PVDF
membranes (Cassano et al., 2007a). West Indian cherry juice clarified using a
0.14 m tubular ceramic MF membrane retained a high concentration of the
ascorbic acid but the color became lighter (Wang et al. 2005). Unlike apple juice,
enzyme pretreatment for removal of phenolic compounds resulting in haze and
sediment formation is not necessarily recommended with dark juices. In trials
on pomegranate juice with laccasse treatment prior to UF, there was a loss of
In addition to the juice organic components, the lime defecation method
universally applied in sugarcane juice clarification enhances the calcium
concentration. This can be an additional source of membrane fouling (Cheryan,
1998). Calcium ions can precipitate on the membrane surface as phosphate salt.
The presence of calcium ions can lead to increased adsorption by electrostatic
charge shielding, complexation and/or bridging effects. It has been observed to
increase interactions between polysaccharide (alginate) molecules (Jermann et
al., 2007)
The conventional sugarcane juice clarification process results in almost
complete removal of irreversible colloids and waxy materials but only partial
removal of soluble gums and reversible colloids (Chen, 1993b). These
components, in turn, are carried through to the raw sugar. The clarification step
removes around 50% of the polysaccharides but the color components, which
are mainly plant pigments associated with polysaccharides, remain unaffected
(Godshall et al., 2002). Furthermore, a very high molecular weight
polysaccharide component of ~1000 kD was isolated during cane sugar refining
(Godshall, 1999). This polysaccharide, which was pale yellow with a turbid
appearance and had a tendency to be occluded in the sugar crystal, was believed
to be responsible for most of the color in white sugar.
Jacob and Jaffrin (2000) investigated membrane fouling in UF of brown cane
sugar solutions with 15 kD ceramic membranes. Using different models for
fouling at constant transmembrane pressure (TMP) (Hermia, 1982), they
observed that a single model was unable to explain the fouling for the entire
duration of filtration. Pore narrowing was dominant in the initial hour whereas
cake filtration model dominated subsequently. De Barros et al. (2003) adopted a
similar approach while studying UF fouling mechanism with depectinized
pineapple juice. Susanto and Ulbricht (2005) studied dextran-membrane
interaction in an attempt to better understand the nature of polysaccharide
fouling. The analysis indicated formation of a monolayer of dextran on the
membrane surface and there was very little dextran-dextran interaction.
3.2.4 Fouling mitigation strategies
In general, fouling control involves one or a combination of the following
strategies viz. adjusting the feed properties, optimizing the operation conditions
and modification of the membrane surface properties.
Literature review
TERI University-Ph.D. Thesis, 2007 24
3.2.5 Feed properties
UF of cane juice at high temperature (70-90°C) and neutral pH (around 7.5)
resulted in better clarification and relatively high permeate flux (Kishihara et al.,
1981; Verma et al., 1996; Saska et al., 1999; Balakrishnan et al., 2000, Ghosh et
al., 2000). The juice type (mixed or clarified) has a significant influence on
membrane flux and fouling. In experiments with 20 kD PES and modified PS
membranes, fouling was considerably higher with mixed juice (Ghosh et al.,
2000) in comparison to clarified juice (Balakrishnan et al., 2001). This was
attributed to higher content of non-sugar impurities in the mixed juice fraction,
which was precipitated out upon treatment with lime in the clarification step.
The effect of suspended solids on juice flux is unclear. Bagacillo present in the
feed stream after initial screening reduced the permeate flux in a cross-flow
system (Ghosh et al., 2000, Balakrishnan, 2000). This was contradictory to an
earlier observation wherein the flux reportedly improved in a stirred cell,
possibly due to scouring effect of the suspended bagasse particles (Kishihara et
al., 1983). Of the different feed pretreatment strategies viz. juice liming, liming
combined with boiling, α-amylase treatment, flocculation of limed and
untreated juice and centrifugation that have been investigated to improve the
juice permeability, liming alone is established to be an effective method of flux
enhancement (Kishihara et al., 1981; Balakrishnan, 2001).
3.2.5.1 Operation parameters
Operation parameters like temperature, trans-membrane pressure (TMP) and
cross-flow velocity (CFV) have a significant impact on fouling.
An increase in temperature from 30 to 60° C almost doubled the sugarcane juice
flux probably due to the decrease in viscosity (Kishihara et al., 1983) At higher
temperature the microbial activity is also reduced Therefore sugarcane juice
filtration was suggested to carry out at maximum possible temperature within
the tolerable range of Browning effect.
In feeds containing suspended particles, the effect of cross-flow velocity on
permeate flux depends upon the particle size distribution (Wakeman and
Tarleton, 1991). With fine suspension, the shear force generated at higher CFV
causes fewer particles to accumulate near the membrane surface and thus
results in lesser fouling. In contrast, larger particles are carried away at higher
CFV and the finer particles settle down onto the membrane surface to cause
Literature review
TERI University-Ph.D. Thesis, 2007 25
fouling resulting in reduced filtration rate. This behavior was observed in the
UF of mixed sugarcane juice using 20 kD PES and 50 kD PS membranes in a
cross-flow unit (Balakrishnan et al., 2000). As the CFV was increased, the
permeate flux decreased with increasing TMP. The membrane surface was
visibly covered with a brownish-green layer, attributed to the suspended solids
(bagacillo), which formed a secondary layer on the membrane surface. A
decrease in juice permeate flux upon an increase in TMP was attributed to the
increasing compactness of the fouling layer on the membrane surface. Kishihara
et al. (1981) reported a similar behavior in the UF of limed sugarcane juice in a
stirred cell with 5-200 kD PES and 300 kD cellulose ester membranes. Similarly
in raw sugar melt filtration, membrane fouling was prominent under conditions
of low CFV and high TMP (Dornier et al., 1994). In an investigation on the start-
up procedure of cross-flow MF, a progressive increase in TMP and cross-flow
velocity was observed to result in higher and consistent permeate flux (Dornier
et al., 1995).
3.2.5.2 Membrane surface characteristics
Most of the investigations on sugarcane juice MF/UF have been conducted on
commercial membranes, without any modification in the surface properties. The
effect of surface hydrophilization on sugarcane juice UF was studied by the
adsorption of 0.1% polyvinyl alcohol (PVA) on 20 kD PES membranes
(Balakrishnan et al., 2001). There was a perceptible reduction in fouling with the
PVA adsorbed membrane and the flux decline with time was less than with the
corresponding unmodified membrane. Moreover, the increase in permeate
purity (expressed as a percentage of polarizing substances in the total dissolved
solids) was significantly higher with the modified membrane, indicating a higher
rejection of non-sugar components. More recently, PEGMA (poly (ethylene
glycol) methacrylate) photo grafting on PES membrane surface was reported to
reduce fouling resistance in UF of sugarcane juice polysaccharide fraction
(Susanto et al., 2007); furthermore, the modified membrane also displayed
higher retention of high molecular weight components.
3.3 Membrane cleaning
Membrane fouling, due to deposition of the rejected material on the surface and
within the pores, results in both flux decline and change in membrane
Literature review
TERI University-Ph.D. Thesis, 2007 26
selectivity. Strategies like feed-pretreatment, adjusting the operation parameters
and membrane surface modification can minimize fouling but cannot completely
eliminate it. Thus, cleaning is an integral part of membrane applications.
Membrane cleaning can involve one or a combination of the following methods.
physical e.g. ultrasound (Muthukumaron et al., 2004), sponge balls
(Maartens et al., 2002 ), back pulsing (Mores and Davis, 2002)
biological e.g. enzymatic treatment (te Poele and Graaf, 2005) and
chemical using acids, alkalis, disinfectants, detergents (Lee et al., 2001;
Liikanen et al., 2002; Maartens et al., 2002; Kuzmenko et al., 2005;
Strugholtz et al., 2005).
For food processing applications like UF of milk, whey and juices, chemical cleaning is most common (Sayed Razavi et al., 1996; Gan et al., 1999; Rabiller-Baudry et al., 2006a; Kazemimoghadam and Mohammadi, 2007; Cassano et al., 2007c).
The efficiency of chemical cleaning is governed by the choice of the cleaning
agent(s) as well as the cleaning conditions. These include pH, ionic strength,
duration and temperature as well as the cross-flow velocity. Further, if multiple
cleaners are being employed, the sequence is also important. Appropriate
chemicals usage causes less damage to membrane surface thereby extending its
lifetime and reducing the frequency of membrane replacement. Thus,
developing an optimal membrane-cleaning strategy for a given application is
essential since it has a direct impact on the process economics.
For most chemical cleaners, 30-60 minutes is generally required for complete
action (Cheryan, 1998); in fact, prolonged chemical cleaning beyond optimal
time may actually refoul the membrane. Thus, there have been several reports
on short chemical cleaning cycles (up to 30 minutes) for membranes fouled by
various multi-component feed streams viz. food streams like passion fruit juice
(Chiang and Yu, 1987), aqueous extract of soy flour (Sayed Razavi et al., 1996),
apple juice (Borneman et al., 1997), milk (Kazemimoghadam and Mohammadi,
2007), effluents like oily wastewater (Lindau and Jonsson, 1994), spent sulfite
liquor (Weis et al., 2003; 2005), palm oil mill effluent (Ahmed et al., 2005) etc.
Literature review
TERI University-Ph.D. Thesis, 2007 27
Various chemicals have been employed for membrane cleaning in food
processing applications. These include sodium hydroxide (NaOH), acids like
hydrochloric acid (HCl), nitric acid (HNO3) and citric acid (C6H8O7.H2O),
oxidizing agents like hydrogen peroxide (H2O2), sodium hypochlorite (NaOCl).
Commercial membrane cleaning detergents and surfactants like Terg-a-zyme