Journal of Food Engineering 96 (2010) 427–432

Contents lists available at ScienceDirect

Journal of Food Engineering

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

Influence of gas sparging on clarification of pineapple wine by microfiltration

Wirote Youravong a,b,*, Zhenyu Li b, Aporn Laorko a,b
a
Department of Food Technology, Faculty of Agro-Industry, Prince of Songkla University, Hat Yai 90112, Thailand
b
Membrane Science and Technology Research Center, Prince of Songkla University, Hat Yai 90112, Thailand

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

Article history: A microfiltration process with a tubular ceramic membrane was applied for clarification of pineapple
Received 10 October 2008 wine. The process was operated with the membrane pore size of 0.2 lm at transmembrane pressure of
Received in revised form 17 August 2009 2 bar and crossflow velocity of 2.0 m/s. The effects of gas sparging on permeate flux, fouling and quality
Accepted 20 August 2009
of clarified wine were studied. It was found that a relatively low gas sparging rate could increase perme-
Available online 25 August 2009
ate flux up to 138%. Further increase of the gas sparging rate did not improve permeate flux compared
with that without gas sparging. Gas sparging affected the density of cake layer. Increasing gas sparging
Keywords:
rate led to an increase in specific cake resistance. It was observed that increasing gas sparging rate could
Microfiltration
Clarification
reduce reversible fouling rather than irreversible fouling. The turbidity of pineapple wine was reduced
Pineapple wine and a clear product with bright yellow color was obtained after microfiltration. The negative effect of
Gas sparging gas sparging which caused a loss of alcohol content in the wine was also observed.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction and Perez, 2002) and cajá (Spondias mombin L.) wine (Severo
et al., 2007).
Virtually all winemakers have a desire for their wines to be cos- It has been reported that microfiltration membranes with pore
metically appealing. The wine is needed to be clear of any haziness size of 0.1 and 0.2 lm are employed in most clarification processes
that may be caused by residual particulate matter. Some methods of wines by membrane filtration (Urkiaga et al., 2002). Similar to
have been conventionally applied for clarification of wine, such as other microfiltration processes, the major problem during microfil-
racking (removing the wine from sediments that settle at the bot- tration of wine is reduction of permeate flux due to concentration
tom of a fermentation or aging vessel), fining (introducing an fining polarization and membrane fouling. It is known that colloid parti-
agent such as egg white, gelatin) (Sims et al., 1995) and filtration cles, microorganism and macromolecules, especially polysaccha-
(running the wine through a pad or diatomaceous earth) (Ruediger ride and polyphenols in wine are important compounds which
et al., 2004). More recently, membrane filtration, especially micro- result in formation of fouling (Belleville et al., 1990; Goncalves
filtration is widely becoming employed in the winemaking process, et al., 2001; Vernhet and Moutounet, 2002). Many methods such
mainly as a clarification and microbiological stabilization (called as vortex promoter (Ding et al., 2002), backpulsing (Ma et al.,
cold sterilization) technique (Salazar et al., 2007). Advantages 2001) and critical flux operation (Li et al., 2008) have been con-
offered by membrane process clarification are as follows: a reduc- ducted to reduce fouling and to improve membrane filtration pro-
tion in clarification time; simplification of the clarification process; cesses. Compared with other methods, gas sparging has been
an increase in the amount of clarified juice; possibility of operation proved to be a simple and low-cost method to reduce membrane
at room temperature and preservation of juice freshness, aroma fouling by introducing gas into a module to promote local mixing
and nutritional value; an improvement in the quality of a final near membrane surface. In addition, gas sparging poses less risk
product through removal of extraneous substances and an to membrane and gas bubbles are easily separated from process
improvement in the production process (Cassano et al., 2003; stream (Cui and Fane, 2003).
Sa et al., 2003). Microfiltration has been employed as a clarification On the other hand, the pineapple industry in Thailand has
method during processing of ordinary red wine (Urkiaga et al., grown steadily since the first pineapple canning factory was estab-
2002), white wine (Goncalves et al., 2001), sherry wine (Palacios lished in 1967. Thailand is now one of the major producers and
exporters of pineapple. During the last three decades, the produc-
tion of pineapple in Thailand stood at about 2 million tons annually
(Anupunt et al., 2000). Except direct consumption of fresh pineap-
* Corresponding author. Address: Department of Food Technology, Faculty of
ple fruit, conventional products of pineapple include canned pine-
Agro-Industry, Prince of Songkla University, Hat Yai 90112, Thailand. Tel.: +66 7428
6321; fax: +66 7421 2889. apple, concentrated juice and dried pineapple chips. In recent
E-mail address: wirote.y@psu.ac.th (W. Youravong). years, pineapple wine as value-added product of pineapple is

0260-8774/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2009.08.021
428 W. Youravong et al. / Journal of Food Engineering 96 (2010) 427–432

becoming more popular because of its appealing flavor. Pineapple pump (Procon 814V230, Millipore, USA) combined with a retentate
wine is one of non-vintage wines made from pineapple juice, valve and measured by a digital flowmeter (Magflo5000, Danfoss,
which is produced and fermented in a manner similar to grape Denmark). According to the previous study using the same system
wines. Major steps during production of pineapple wine include in the laboratory, a fixed liquid crossflow velocity of 2.0 m/s and a
peeling and cutting of fruit, juice extraction, fermentation, clarifi- fixed transmembrane pressure of 2 bar were applied to minimize
cation, bottling and maturation. Membrane filtration could be a the effect of concentration polarization and to attain a relatively
potential method employed for clarification and cold sterilization high flux during microfiltration process. Microfiltration process
of pineapple wine. However, few results have been reported about was run under total recycle mode for 2 h. The feed was supplied
application of membranes in the processing of pineapple wine. The to the lumen side of the membrane. Both retentate and permeate
aim of this study is to examine the microfiltration process for clar- were recycled to a stainless steel feed tank (maximum capacity
ification of pineapple wine and the effect of gas sparging on pro- of 10 l) to maintain a constant feed property. The permeate from
cess performance. microfiltration process was collected and weighed continuously
using a digital balance (GF-3000, A&D, Japan). All filtration pro-
2. Materials and methods cesses were performed at ambient temperature (25 ± 1 °C).
A gas sparging technology was applied to enhance the permeate
2.1. Preparation of pineapple wine flux. The compressed nitrogen gas was injected into the inlet of
feeding pipe through a Y-tubular piece. The gas flow velocity was
Fresh pineapple wine (white wine) was prepared in our bio- controlled and measured by a gas flowmeter (RMB-53D-SSV,
process laboratory. It was made from fresh pineapple juice using Dwyer, USA) combined with a pressure gauge (2419-2C-P, CKD, Ja-
Saccharomyces cerevisiae as a starter. It was fermented and kept pan). The gas–liquid dual flow pattern depends on the gas injection
at room temperature (25–28 °C) for 6 months before filtration to factor (r) which equals to Ug/(Ug + Ul). Ug and Ul are the superficial
allow precipitation of large particles (>5.0 lm). Particle size distri- gas and liquid flow rate or flow velocity, respectively. The dual flow
bution in pineapple wine was detected by a Laser Particle Size Ana- pattern changes from bubble flow (0 < r < 0.2) over slug flow
lyzer (LS230, Beckman Coulter, USA). The minimum particle size (0.2 < r < 0.9) to annular flow (0.9 < r < 1.0) (Psoch and Schiewer,
was about 1.0 lm. The particle size ranged from 1.0 to about 2005). Gas flow velocity was varied from 0 to 1.1 m/s to achieve
70 lm (Fig. 1). r of 0–0.35 in this study.
Acidity (percentage of citric acid), color, alcohol content (%) and The experiments for determination of flux and fouling resis-
sugar (%) were measured by the methods of AOAC (1999). Viscosity tances were performed in triplicate with less than 5% of error.
was measured by a U-tube viscometer (Kapillarviskosimeter Three independent samples were collected for each analysis of
50904, Schott, Germany) at 20 °C. The pH was measured by a pH pineapple wine properties. The standard derivations of analysis
meter (PB-20, Sartorius, Germany). Turbidity was evaluated as were indicated in Table 1.
the absorbance at a wavelength of 340 nm by spectrophotometer
(Youravong et al., 2005). 2.3. Model of membrane fouling

2.2. Microfiltration and gas sparging setup Generally, a microfiltration process can be described by Darcy’s
Law (Eq. (1)):
The membrane used for microfiltration was a single-channeled DP
tubular ceramic membrane with pore size of 0.2 lm, length of J¼ ð1Þ
lRtol
40 cm, inner diameter of 6 mm, outer diameter of 10 mm and
effective area of 75 cm2 (Jiangsu Jiuwu Hi-Tech Co. Ltd., China). where: J is the volumetric flux of permeate across the membrane
Both membrane material and supporting material were a-Al2O3. (m/s), DP is the transmembrane pressure (Pa), l is the feeding fluid
Pressure transducers (MBS3000, Danfoss, Denmark) were used to viscosity (Pa s) and Rtol is the total hydraulic resistance (m1).
detect the pressure at the inlet and outlet of membrane, and that In practice, all fouling phenomena, e.g. adsorption, pore block-
of the permeate. The crossflow velocity was controlled by a feed ing and cake layer formation may occur simultaneously during
membrane filtration because of the complexity of the feed compo-
sition, operating conditions and membrane properties. However
5 one or more fouling mechanisms may be dominant in a certain
membrane filtration process or at different stages of a membrane
4.5
filtration. While filtering particulate matter, microfiltration mem-
4 branes undergo three different modes of flux loss: concentration
polarization, (followed by) pore blocking and aggregate cake for-
Differential volume (%)

3.5
mation (i.e. a cake of retained aggregates composed of many small
3 primary colloidal particles). The principal limitation of microfiltra-
tion lies in membrane fouling which is mainly associated with the
2.5
deposition of a cake layer onto the membrane surface. Cake filtra-
2 tion model has been considered for the study of membrane fouling.

1.5
Table 1
1 Effects gas sparging on the properties of cake layer during microfiltration of pineapple
wine at the crossflow velocity of 2.0 m/s and transmembrane pressure of 2 bar.
0.5
Gas injection factor, r t/V versus V kc
0
0 t/V = (1.0  103)V + 6.2 2.0  103
0.1 1 10 100 1000
0.15 t/V = (1.2  103)V + 3.6 2.4  103
Particle diameter (µm) 0.25 t/V = (2.7  103)V + 5.9 5.4  103
0.35 t/V = (4.0  103)V + 4.1 8.0  103
Fig. 1. Particle size distribution of original pineapple wine.
 W. Youravong et al. / Journal of Food Engineering 96 (2010) 427–432 429

Filtration time and permeate volume in this model can be related hance permeate flux. After 2 h of running time, it was observed
by Eq. (2) (Ye et al., 2005) that a steady flux was achieved in all runs with different r
(Fig. 2). The whole microfiltration process could be divided into
t lca V lRm two stages: declining flux stage and steady flux stage. The initial
¼  þ ð2Þ
V A2 DP 2 ADP declining flux stage could be mainly ascribed to the concentration
where V is the permeate volume (m3); t is the filtration time (s); c is polarization, adsorption of solutes and pore constriction, or combi-
the mass of the particles deposited in the filter cake per unit volume nation of these factors. The steady flux stage could be dominated
of the filtrate (kg/m3); a is the, specific cake resistance (m/kg); A is by the formation of cake layer. Varying r from 0 to 0.15 could in-
the effective membrane area (m2); Rm is the membrane resistance crease steady flux up to 138%. This could be due to the fact that
(m1). bubble-induced secondary flows promoted local mixing (or vor-
Based on Eq. (2), a cake filtration constant, kc is introduced into tex). An increase of wall shear stress due to introduction of bubbles
cake filtration model. It is a function of specific cake resistance gi- could minimize the accumulation of solutes and molecules on the
ven by Eq. (3) (Kim et al., 1993; Mccabe et al., 1993). membrane (e.g. concentration polarization and cake layer). Conse-
quently, the flux could be enhanced. In addition, mechanism of
lca enhancement in gas sparged membrane filtration relies heavily
kc ¼ ð3Þ
A 2 DP on the flow reversal effect caused by the presence of bubbles.
The fouling mechanism related to the cake layer can be ana- The flow reversal could be induced by the bubbles as they pass a
lyzed by fitting experimental data to Eq. (2). The linear relation be- point in the membrane. The flow reversal effect could enhance
tween t/V and V indicates the stage where cake layer is the the flux by reducing the thickness of the average mass transfer
dominant fouling during microfiltration process. The slope of the boundary layer (Smith and Cui, 2004). However, for controlled
linear line (i.e. kc/2) is linked to the cake property. gas sparged membrane filtration processes, it was found that gas
flow rate required for substantial improvement in permeate flux
2.4. Resistance analysis was very small (Taha et al., 2006). A relatively low gas injection
rate already successfully limited the accumulation of solutes onto
After microfiltration of pineapple wine for 2 h, the total resis- membrane, leaving less room for further enhancement by higher
tance (Rtol) to permeate flow during microfiltration was calculated gas injection rate. Therefore, further increase of r (i.e. r of 0.25
by Eq. (1). Rtol was divided into membrane resistance (Rm), resis- and 0.35) did not show any benefit to improve the flux in this
tance caused by reversible fouling (Rrf) and resistance caused by study. Actually the fluxes obtained from r of 0.25 and 0.35 were
irreversible fouling (Rif). Rtol was the sum of Rm, Rrf and Rif. Rm lower than that without gas sparging. The decay of steady flux with
was evaluated by measuring deionized water flux through clean increasing r could be explained by both solvent mass transporta-
membrane at transmembrane pressure of 2 bar for 10 min. After tion and solute mass transportation. Gas induced bubbles in the
each run of microfiltration, the fouled membrane was flushed by flow channel could increase the turbulence, leading to a flux
deionized water at the crossflow velocity of 2.0 m/s for 10 min to enhancement. However, the excessive injected gas could decrease
eliminate Rrf. Rif was calculated from the deionized water flux of the effective membrane area by replacing liquid mass by bubbles
flushed membrane at transmembrane pressure of 2 bar for 10 min. contacting to the membrane surface; therefore, less solvent passed
through the membrane (Mi-Jung et al., 2001). On the other hand,
the negative effect of gas sparging on the permeate flux could be
3. Results and discussion due to the change of fouling behavior by bubbles in the flow. It
was supposed that an increase in wall shear stress due to high
3.1. Effect of gas sparging on flux behavior gas injection rate could decrease cake thickness and porosity by
preferentially removing larger particles away from membrane
Gas sparging has been accepted as a simple and promising and allowing smaller particles to close to membrane, then make
method to reduce concentration polarization and fouling on a a formation of fouling layer thinner but more compactly packed
membrane surface. The results showed that gas sparging could en- (Mercier-Benin and Fonadeb, 2002).

3.2. Effect of gas sparging on cake layer

0.0001
Generally, the dominant fouling mechanism would be cake fil-
0.00009
tration at the end of filtration if the rate of flux loss is slow and
0.00008 steady. In this study, the slow and steady rates of flux loss were ob-
Permeate flux J (m/s)

0.00007 served at all operating conditions. The strong linear relation be-
0.00006
tween t/V and V could be seen (Fig. 3a–d). All data could be fitted
to a cake filtration model with defined coefficient of linear regres-
0.00005
sion (R2 P 0.99) at the end of microfiltration processes. Gas sparg-
0.00004 ing showed the influence on kc (Table 1). For example, kc was about
0.00003 2.0  103 for r of 0 and 8.0  103 for r of 0.35. The higher kc indi-
cated a more intensive cake layer, i.e. the specific cake resistance
0.00002
increased, or the effective membrane area available for filtration
0.00001 decreased. Gas injection may cause the formation of more com-
0 pactly packed solute layer on the membrane (Mercier-Benin and
0 1000 2000 3000 4000 5000 6000 7000 8000 Fonadeb, 2002). The more compact fouling layer led to higher kc.
Filtration time (s) In a crossflow microfiltration of submicro particles, the scanning
electron microscope analysis has illustrated that the existence of
Fig. 2. Effect of gas sparging on the permeate flux during the microfiltration of
pineapple wine at the crossflow velocity of 2.0 m/s and transmembrane pressure of
air bubbles caused the filter cake to be a more compact packing
2 bar (e, without gas; h, gas injection factor r of 0.15; D, gas injection factor r of structure and to a increase in the specific filtration resistance
0.25; , gas injection factor r of 0.35). (Hwang and Wu, 2008).
430 W. Youravong et al. / Journal of Food Engineering 96 (2010) 427–432

a 9 8000 b
8 7000

7
6000

Filtration time, t (s)

6
5000
t / V (s/ml)

5
4000
4
3000
3
2000
2

1 1000

0 0
0 200 400 600 800 1000 1200
Permeate volume, V (ml)

c 9 8000 d 9 8000

8 7000 8 7000
7 7
6000 6000
Filtration time, t (s)

6 6

Filtration time, t (s)

5000 5000
t / V (s/ml)

t / V (s/ml)

5 5
4000 4000
4 4
3000 3000
3 3
2000 2000
2 2
1 1000 1000
1
0 0
0 0
0 200 400 600 800 1000
0 200 400 600 800 1000
Permeate volume, V (ml)
Permeate volume, V (ml)

Fig. 3. Determination of cake filtration during the microfiltration of pineapple wine at the crossflow velocity of 2.0 m/s and transmembrane pressure of 2 bar ((a) without gas;
(b) gas injection factor r of 0.15; (c) gas injection factor r of 0.25; (d) gas injection factor r of 0.35).

that Rrf decreased with increasing r. The loss of reversible resis-

100
tance indicated that concentration polarization has been reduced.
90 Generally, the reversible fouling (e.g. concentration polarization or
loose layer that is created by accumulation of solutes on mem-
80
brane surface) is sensitive to hydrodynamic conditions and could
Normalized resistance ( R/Rm )

70 be reduced or eliminated by hydrodynamic techniques. Bubble-in-

duced secondary flows minimized concentration polarization and
60
accumulation of solutes on membrane by promoting local mixing
50 and turbulence in bubble wake (Cui and Wright, 1996), and in-
40
creased the mass transfer coefficient in the boundary layer. Thus,
reversible fouling could be reduced by gas sparging. On the other
30 hand, it was proved that gas sparging was more effective in a sys-
20 tem where flux decline was dominated by reversible fouling (such
as concentration polarization) than that coupled with irreversible
10 fouling (such as cake or gel layer, pore blocking) (Li et al., 1998).
0 Hence, r of 0.15 did not significantly reduce Rif compared with
0 0.1 0.2 0.3 0.4 that without gas sparging. However Rif was increased when r of
Gas injection factor, r 0.25 and 0.35 were applied. As mentioned above, the increase of
Rif could be due to the change of cake layer induced by higher r.
Fig. 4. Effect of gas injection factor on normalized resistances during microfiltration
of pineapple wine at the crossflow velocity of 2.0 m/s and transmembrane pressure The cake layer could become tighter and more compact with high-
of 2 bar (e, total resistance, Rtol; h, resistance caused by reversible fouling, Rrf; D, er r and it was difficult to be removed by water flushing. In this
resistance caused by irreversible fouling, Rif). study, the experiments were run over a relatively short time of
about 2 h. After the operation for longer time, the irreversible
3.3. Effect of gas sparging on filtration resistance fouling would have caught up or even overtaken all other forms
of fouling, negating effects of gas sparging. In further work, exper-
Fig. 4 shows the effect of gas sparging on normalized resis- iments should have been run over 24 h with one cleaning cycle to
tances during microfiltration of pineapple wine. It could be seen be of more value to the industry.
 W. Youravong et al. / Journal of Food Engineering 96 (2010) 427–432 431

Table 2
Effects of microfiltration and gas injection factor on the properties of pineapple wine.

Original wine before MF Permeate (r of 0) Permeate (r of 0.15) Permeate (r of 0.25) Permeate (r of 0.35)
a
Viscosity (MPa s) 2.03 ± 0.008 1.76 ± 0.005 1.78 ± 0.004 1.78 ± 0.005 1.77 ± 0.006
Turbidity (340 nm) 1.68 ± 0.053 1.05 ± 0.023 1.12 ± 0.019 1.15 ± 0.027 1.09 ± 0.020
Sugar (%) 6.94 ± 0.241 6.18 ± 0.197 6.35 ± 0.220 6.27 ± 0.187 6.44 ± 0.246
Acidity 0.26 ± 0.002 0.24 ± 0.002 0.23 ± 0.001 0.24 ± 0.002 0.24 ± 0.002
pH 3.14 ± 0.014 3.17 ± 0.013 3.15 ± 0.015 3.14 ± 0.013 3.18 ± 0.012
Alcohol (%) 10.81 ± 0.161 9.47 ± 0.103 8.83 ± 0.098 6.02 ± 0.091 5.85 ± 0.107
Color
L (lightness) 95.62 ± 0.175 97.31 ± 0.181 94.77 ± 0.163 97.18 ± 0.177 95.21 ± 0.161
A (redness) 1.86 ± 0.026 1.72 ± 0.021 1.83 ± 0.027 1.69 ± 0.016 1.78 ± 0.021
B (yellowness) 18.11 ± 0.043 19.48 ± 0.051 18.59 ± 0.046 19.31 ± 0.041 19.14 ± 0.044
a
Values represent standard deviations.

3.4. Effect of microfiltration on the properties of pineapple wine Subhadrabhandu, S., Chairidchia, P. (Eds.), Proceedings of The Third
International Pineapple Symposium. Wageningen, ISHS, pp. 99–107.
AOAC, Cunnif, P. (Eds.), 1999. Official Methods of Analysis of AOAC International,
After microfiltration, the pineapple wine recovered from per- 16th ed. AOAC International, Alrington, VA, USA.
meate was a clear liquor with bright yellow color. The turbidity Belleville, M.P., Brillouet, J.M., Fuente, T., Moutounet, M., 1990. Polysaccharide
and viscosity were reduced while other properties of wine were effects on cross-flow microfiltration of two red wines with microporous
aluminar membrane. Journal of Food Science 55, 1598–1602.
not changed significantly (Table 2). It should be mentioned that Cassano, A., Drioli, E., Galalaverna, G., Marchelli, R., Silvestro, R.D., Cagnasso, P.,
the gas sparging caused the loss of alcohol content of wine. The 2003. Clarification and concentration of citrus and carrot juices by integrated
loss of alcohol was more severe with higher r. Even though the membrane processes. Journal of Food Engineering 57, 153–163.
Cui, Z.F., Fane, A.G., 2003. The use of gas bubbling to enhance membrane processes.
gas sparging was not applied, a small loss of alcohol had still been Journal of Membrane Science 221, 1–35.
observed. This could be due to the operating mode used in this Cui, Z.F., Wright, K.I.T., 1996. Flux enhancement with gas sparging in downwards
study. The total recycle mode was employed, thus both retentate crossflow ultrafiltration: performance and mechanism. Journal of Membrane
Science 117, 109–116.
and permeate were repeatedly circulated in the system. The loss Ding, L.H., Al-Akoum, O., Abraham, A., Jaffrin, M.Y., 2002. Milk protein concentration
of alcohol should be due to the evaporation of alcohol during the by ultrafiltration with rotating disk modules. Desalination 144, 307–311.
process. The injection of gas promoted stir and vortex in the flow, Goncalves, F., Fernandes, C., Pinho, M.N., 2001. White wine clarification by
microultrafiltration: effect of removed colloids in tartaric stability. Separation
then enhanced the evaporation of alcohol. In an industrial opera- and Purification Technology 22–23, 423–429.
tion, the permeate is continuously removed as product from sys- Hwang, K.J., Wu, Y.J., 2008. Flux enhancement and cake formation in air-sparged
tem rather than repeatedly circulated in the system. In addition, cross-flow microfiltration. Chemical Engineering Journal 139, 296–303.
Kim, K.J., Chen, V., Fane, A.G., 1993. Ultrafiltration of colloidal silver particles: flux,
the operation can be controlled at lower temperature by cooling
rejection and fouling. Journal of Colloid and Interface Science 155, 347–359.
water system and lowering gas injection factor. It could be ex- Li, Q.Y., Ghosh, R., Bellara, S.R., Cui, Z.F., Pepper, D.S., 1998. Enhancement of
pected that the loss of alcohol would be reduced or avoided in ultrafiltration by gas sparging with flat sheet membrane modules. Separation
industrial operations. and Purification Technology 14, 79–83.
Li, Z.Y., Youravong, W., H-Kittikun, A., 2008. Removal of suspended particles from
extract of tuna spleen by microfiltration: experimental evaluation of a low
fouling condition. Journal of Food Engineering 85, 309–315.
4. Conclusion
Ma, H.M., Hakim, L.F., Bowman, C.N., Davis, R.H., 2001. Factors affecting membrane
reduction by surface modification and backpulsing. Journal of Membrane
Microfiltration can be employed for clarification of pineapple Science 189, 225–270.
wine with acceptable quality of product. Suspended particles in Mccabe, W.L., Smith, J.C., Harriott, P., 1993. Unit Operations of Chemical
Engineering. McGraw-Hill, New York. pp. 1019–1023.
pineapple wine were completely removed by microfiltration. Gas Mercier-Benin, M., Fonadeb, C., 2002. Air-sparged microfiltration of enzyme/yeast
sparging could be applied to enhance the permeate flux. A low mixtures: determination of optimal conditions for enzyme recovery.
gas injection factor could increase the permeate flux while higher Desalination 148, 171–176.
Mi-Jung, U., Seong-Hoon, Y., Chung-Hak, L., Kun-Yong, C., Jae-Jin, K., 2001. Flux
gas injection factor did not show any benefit on the permeate flux. enhancement with gas injection in crossflow ultrafiltration of oily wastewater.
Gas injection factor of 0.15 gave the best improvement of permeate Water Research 35, 4095–4101.
flux (up to 138%). Increasing gas injection factor tended to reduce Palacios, V.M., Perez, I.C., 2002. Comparative study of crossflow microfiltration with
conventional filtration of sherry wine. Journal of Food Engineering 54, 95–102.
reversible fouling but not irreversible fouling. In addition, gas Psoch, C., Schiewer, S., 2005. Long-term study of an intermittent air sparged MBR for
sparging also affected the fouling related to the formation of cake synthetic wastewater treatment. Journal of Membrane Science 260, 56–65.
layer onto membrane surface. The density of cake layer increased Ruediger, G.A., Pardon, K.H., Sas, A.N., Godden, P.W., Pollnitz, A.P., 2004. Removal of
pesticides from red and white wine by the use of fining and filter agents.
as the gas injection factor increased. The loss of alcohol in the pine-
Australian Journal of Grape and Wine Research 10, 8–16.
apple wine was observed during microfiltration under total recycle Sa, I.S., Cabral, L.M.C., Matta, V.M., 2003. Concentration of pineapple juice by
mode. However, it could be reduced or avoided in industrial pro- membrane separation processes. Brazilian Journal of Food Technology 6 (1), 53–
cess by changing mode of operation and lowering operating 62.
Salazar, F.N., Bruijn, J.P.F.D., Seminario, L., Guell, C., Lopez, F., 2007. Improvement of
temperature. wine crossflow microfiltration by a new hybrid process. Journal of Food
Engineering 79, 1329–1336.
Severo Jr., J.B., Almeida, S.S., Narain, N., Souza, R.R., Santana, J.C.C., Tambourgi, E.B.,
Acknowledgement
2007. Wine clarification from Spondias mombin L. pulp by hollow fiber
membrane system. Process Biochemistry 42, 1516–1520.
The authors gratefully acknowledge Bioprocess Laboratory, Fac- Sims, C.A., Eastridge, J.S., Bates, R.P., 1995. Changes in phenols, color, and sensory
characteristics of muscadine wines by pre- and postfermentation additions of
ulty of Agro-Industry, Prince of Songkla University for providing
PVPP, casein, and gelatin. American Journal of Enology and Viticulture 46, 155–
the pineapple wine. 158.
Smith, S.R., Cui, Z.F., 2004. Gas-slug enhancement hollow fibre ultrafiltration: an
experimental study. Journal of Membrane Science 242, 117–128.
References Taha, T., Cheong, W.L., Field, R.W., Cui, Z.F., 2006. Gas sparged ultrafiltration using
horizontal and inclined tubular membrane – a CFD study. Journal of Membrane
Anupunt, P., Chairidchia, P., Kongswat, A., Isawilanon, S., Subhadrabhandu, S., Science 279, 487–494.
Vasunun, S., Siripat, S., 2000. The pineapple industry in Thailand. In:
432 W. Youravong et al. / Journal of Food Engineering 96 (2010) 427–432

Urkiaga, A., Fuentes, L.D.L., Acilu, M., Uriarte, J., 2002. Membrane comparison for Ye, Y., Clech, L., Chen, V., Fane, A.G., Jefferson, B., 2005. Fouling mechanisms of
wine clarification by microfiltration. Desalination 148, 115–120. alginate solutions as model extracellular polymeric substances. Desalination
Vernhet, A., Moutounet, M., 2002. Fouling of organic microfiltration membranes by 175, 7–20.
wine constituents: importance, relative impact of wine polysaccharides and Youravong, W., Phukdeekhong, M., Taksinpatanapong, P., 2005. Influence of
polyphenols and incidence of membrane properties. Journal of Membrane membrane properties on pineapple wine clarification and fouling behavior.
Science 201, 103–122. Applied Membrane Science and Technology 1, 9–18.