当前位置:首页 >> 能源/化工 >>

Clarification of pomegranate juice by microfiltration with PVDF membranes


Desalination 264 (2010) 243–248

Contents lists available at ScienceDirect

Desalination
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e /

d e s a l

Clari?cation of pomegranate juice by micro?ltration with PVDF membranes
Hossein Mirsaeedghazi a,?, Zahra Emam-Djomeh a, Sayed Mohammad Mousavi a, Abdolreza Aroujalian b, Mahdi Navidbakhsh c
a b c

Transport Phenomena Laboratory, Department of Food Science and Engineering, Biosystems Faculty, Agriculture Engineering Campus, University of Tehran, Karaj, Iran Department of Chemical Engineering, Amir Kabir University of Technology, Tehran, Iran Department of Mechanical Engineering, Iran University of Science and Technology, Narmak, Tehran, Iran

a r t i c l e

i n f o

a b s t r a c t
Micro?ltration was used to clarify pomegranate juice using two polyvinylidene ?uoride membranes with pore sizes of 0.22 and 0.45 μm. Changes were studied in the chemical properties of the juice after passing through each membrane. Characteristics such as turbidity changed for both membranes (more than 95%). The permeate ?ux decreased over time as a result of membrane fouling. The degree of decline in the membrane with pore size of 0.22 μm was greater than another one. In both membranes, fouling resistance increased over time from 5 × 109 m2/kg to 4.43 × 1010 m2/kg for 0.22 μm and to 1.29 × 1010 m2/kg for 0.45 μm after 45 min. This increase had a sharp slope in the ?rst stages of the testing. Fouling index changes over time showed similar behavior. Scanning electron microscopy images showed that the cake layer had the greatest impact on membrane fouling after processing by preventing turbid components from entering pores. ? 2010 Elsevier B.V. All rights reserved.

Article history: Received 27 January 2010 Received in revised form 10 March 2010 Accepted 12 March 2010 Available online 10 April 2010 Keywords: Micro?ltration Pomegranate juice Clari?cation Polyvinylidene ?uoride Membrane Fouling

1. Introduction The pomegranate (Punica granatum, Punicaceae) is a native fruit of Iran. Pomegranate juice has potential health bene?ts such as antiatherogenic effects in healthy humans and atherosclerotic effects in mice that may be attributable to its anti-oxidative properties [1]. It has been reported that this fruit juice has other nutritional and health advantages [2–4]. Thus, this fruit juice is consumed in large quantities all over the world. One of the greatest hindrances to the marketability of this juice is its turbid appearance, which makes it undesirable to consumers. Polyphenols, which are responsible for haze formation and the undesirable color of pomegranate juice, are one of the most unfavorable components from a marketing point of view. These must be removed from the fresh juice to improve its appearance. Removal has been accomplished by adding enzymes that agglomerate these components and then removing the large particles by ?ltration. However, this process requires expensive enzymes and a signi?cant period of time. Micro?ltration is a pressure-driven membrane separation process that can help clarify fruit juice without enzymatic treatment. There are several studies on membrane processing of fruit juice for clari?cation. Borneman et al. [5] removed polyphenols in apple juice using PES/PVP membranes in a single ultra?ltration process. Vaillant

? Corresponding author. Tel.: +98 21 66404600; fax: +98 261 2248804. E-mail address: mirsaeed@ut.ac.ir (H. Mirsaeedghazi). 0011-9164/$ – see front matter ? 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.desal.2010.03.031

et al. [6] clari?ed melon juice by cross-?ow micro?ltration and the resulting juice was highly similar to the initial juice except for insoluble solids and carotenoids. When acerola juice was clari?ed using micro?ltration, 84% of consumers reported that they liked the clari?ed juice [7]. Vaillant et al. [8] studied micro?ltration of six tropical fruit juices (mango, pineapple, naranjilla, castillas blackberry, passion fruit, and tangerine). When estimating the total cost of producing clari?ed passion fruit juice, they found that a juice to volumetric reduction ratio (VRR) speci?cation exists at which costs are minimized. For juices with high pulp content, these optimal economic costs are reached at a relatively low VRR. They showed that, by controlling VRR, it was possible to increase the soluble solid (SS) content until it reached the same concentration level as found in the raw juice. Under these conditions, the retentate was found to be very similar to the initial juice and could be reintroduced into the pulpy juice processing line. In another study, de Barros et al. [9] studied fouling mechanisms in pineapple juice clari?cation using ultra?ltration and showed that a hollow ?ber membrane separation process is controlled by a caked ?ltration mechanism and that a pore blocking fouling mechanism controls ceramic tubular membranes. They also modeled fouling mechanisms in their study. The juice of the umbu (fruit native to Brazil) was clari?ed using a polypropylene membrane with a pore size of 0.22 μm. The authors evaluated different resistances and concluded that greater enzyme concentration and cross-?ow velocity cause less resistance as a result of the polarized layer. It was also shown that the interaction between

244

H. Mirsaeedghazi et al. / Desalination 264 (2010) 243–248

them affected polarized layer resistance; however, the major factor affected in fouling resistance was transmembrane pressure. An increase in the latter factor increases the fouling resistance [10]. These authors also tested tamarind juice and obtained similar results [11]. There has not been, to our knowledge, any study on the clari?cation of pomegranate juice by micro?ltration. In this work, the feasibility of pomegranate juice clari?cation using micro?ltration was studied. However, the main objective of this study was to evaluate the fouling mechanisms occurring during membrane processing. 2. Materials and methods 2.1. Juice extraction Pomegranate fruits, (Punica granatum L.) vt. malase saveh, were obtained from a local market in Saveh, Iran. They were washed and then manually cut up; the outer leathery skin that encloses hundreds of ?eshy sacs was removed. The juice was extracted by manual pressing of the sacs. Large particles in the juice were removed using a mesh ?lter (no. 9) and the juice obtained was poured into PET bottles and stored at ? 25 °C until processing. 2.2. Membranes Two sheet membranes with pore sizes of 0.22 (M1) and 0.45 μm (M2), made of polyvinylidene ?uoride (PVDF) (Milipore, USA), with a total effective ?ltration area of 137.5044 × 10? 4 m2 were used in this study. The experiments were performed in a batch mode using a laboratory-scale plant (Fig. 1). The feed temperature was adjusted by circulating water in a two-layered tank. A centrifugal pump was used to introduce the feed above the membrane surface in a cross?ow mode. A transmitter coupled with an inverter was used to maintain the feed pressure at the desired levels. Pressures were measured on the feed and retentate sides but not on the permeate side, which had very low pressure. After membrane processing, the permeate ?owed into a permeate vessel and its weight was measured as a function of time. The retentate was recycled to the feed tank. 2.3. Analytical methods The soluble solids content of the initial juice, permeate and retentate samples were measured with a refractometer (Atago, HSR500, Japan) at 25 °C and expressed in °Brix. To perform that, a drop of sample was placed on the surface of the refractometer and the TSS value was recorded. Color intensity was measured by UV–vis.
Fig. 2. Permeate ?ux over time in the membrane during the ?ltration of pomegranate juice.

adsorption at 510 nm (maximum λ of total anthocyanins) using a spectrometer (Cecil, CE 2502, England) after suitable dilution with distilled water. The instrument was zeroed before all experiments [12]. The turbidity of the samples was measured with a portable turbidimeter (WTW, 350 IR., USA) at 25 °C after distilling 2.5 × 10? 6 m3 (2.5 cm3) fruit juice with 5 × 10? 6 m3 (5cm3) water. Titratable acidity was calculated as the percentage of citric acid by titrating 2.5 × 10? 3 kg of pomegranate juice with a solution of NaOH (0.1 N) until it reached pH 8.1. The pH was measured with a digital pH meter (Metrohm, 1.744.0010, Switzerland) [13]. Particle size was measured using a particle size analyzer (Malvern, Hydro 2000 S, England) and analyzed using Mastersizer 2000 software. Viscosity was measured at 25 °C using a viscometer (Brook?eld, DV-II + pro, USA) with a ULA spindle at 38 rpm to 140 rpm and analyzed using Rheocalc V3.1-1 software. Scanning electron microscopy (SEM) (Philips, XL30, The Netherlands) was used for analyzing fouling layer morphology after coating samples with gold using a physical vapor deposition method with a sputter coater (BAL-TEC, SCDOOS, Switzerland). Surfaces and cross sections of the membranes were scanned before and after the clari?cation of pomegranate juice to evaluate the effect of juice treatment on membrane fouling. 2.4. Cross-?ow experiments The pressure on the feed side (pa) and retentate side (pb) were 5 × 104 Nm? 2 (0.5 bar); the pressure on the permeate side (pc) was approximately zero. This made the transmembrane pressure 5 × 104 Nm? 2 (0.5 bar), as measured by following relation: pb + pa ?pc ; 2

ΔP =

?1?

Permeate ?uxes (J) of treated solutions were calculated as:
mp ρ p



J=

A×t

= ?m=h?;

?2?

Fig. 1. The plate and frame micro?ltration that is used in this study.

where A is the membrane area [m2], mp is the permeate weight [kg], ρp is the permeate density [kg/m3] and t is time [h]. Before the testing, the unit was run with water at an inlet pressure of 5 × 104 Nm? 2 and the permeate ?ux was measured under these conditions.

H. Mirsaeedghazi et al. / Desalination 264 (2010) 243–248 Table 1 Changes in chemical properties of the pomegranate juice clari?ed by membrane processing. Membrane M1 Sample Initial juice Permeate Permeate Initialjuice Initial juice Permeate Permeate Initialjuice Total soluble Solids (°Brix) 16.27 15.13 0.93 18.1 17.5 0.967 Turbidity (NTU) 484 3.21 0.0066 497.6 14.78 0.0297 pH 3.38 3.44 1.018 3.27 3.33 1.018 Color intensity (Absorbance at 510 nm) 1.473 0.562 0.38 1.605 0.423 0.26 Acidity (g citric acid/100 g) 1.15 1.14 0.99 1.33 1.20 0.90

245

Viscosity (cp) 1.70 1.45 0.85 2.13 2.07 0.97

M2

2.5. Analysis of resistance Membranes have two types of resistance against permeate ?ux: membrane resistance and fouling resistance. Membrane resistance is resistance against permeate ?ux caused by the membrane itself and can be computed using the following equation [14,15]: Rm = ΔP ; μ w Jw ?3?

where Lp1 is the hydraulic permeability after washing with water, alkaline and acid detergents. Rfrev = 1 ?Rm ?Rfirr ; μw Lp2 ?7?

where Lp2 is the hydraulic permeability after washing with water. Caked resistance can be measured using following equation: Rc = Rt – Rm – Rfrev ? Rfirr ; ?8?

where Rm is membrane resistance, ΔP is transmembrane pressure (Nm? 2), μw is water viscosity (Nsm? 2) and Jw is the permeate ?ux of the membrane before testing (kg/m2s). After treating the juice by membrane processing, fouling resistance can be measured using the following equation: ΔP RF = ?Rm ; μp Jp ?4?

2.6. Measurement of fouling index The fouling index was calculated over time according to the following relation: J = J0 × t
?b

;

?9?

where μp is the viscosity of the permeate in the pomegranate juice membrane process (Nsm? 2) and Jp is permeate ?ux at any time (kg/m2s) [16]. Fouling resistance is included in caked (Rc), reversible (Rfrev) and irreversible (R?rr) resistances. To calculate each resistance, hydraulic permeability (Lp) was measured after washing membrane with water, alkaline and acid detergents as follows: ΔP ; Lp = Jw ?5?

where J0 is the water ?ux before beginning the main test (kg/m2s), J is the permeate ?ux in the membrane processing of pomegranate juice at time t (kg/m2s) and b is the fouling index [18]. 2.7. Statistical analysis All tests were performed in triplicate and the statistical analysis was performed using Minitab 15 software. 3. Results and discussion

where Jw is water ?ux after each washing. Each resistance was calculated according to Cassano et al. [17] as: Rfirr = 1 ?Rm ; μ w Lp1 ?6?

3.1. Flux behavior during testing Permeate ?ux decreased over time as shown in Fig. 2. The permeate ?ux at the beginning of the testing was greater than at other times because of fouling in both membranes. Initial permeate

Fig. 3. Particles size distribution in pomegranate juice before and after membrane process in a membrane with a pore size of 0.45 μm (M2).

Fig. 4. Changes in fouling resistance during the membrane processing with two membranes, M1 and M2.

246

H. Mirsaeedghazi et al. / Desalination 264 (2010) 243–248

These results show that membrane fouling produced in early stages of testing was greater than in the ?nal stages; thus the slope of the permeate ?ux decline during the primary stages was greater than at another times for both membranes. This phenomenon can be also observed in results obtained from membrane fouling and fouling index changes over time. The smaller pore size in M1 compared to M2 caused membrane fouling after the early stages of processing in M1 to be greater than for M2. This is because the ?ux difference between the two membranes in the primary stages of membrane processing was smaller than for the last stages. This agrees with the results of membrane fouling. 3.2. Changes in chemical properties of pomegranate juice after membrane processing As shown in Table 1, fruit juice turbidity decreased after passing through the membrane with the removal turbid components such as phenolic compounds, which is the aim of membrane processing. Statistical analysis of the samples showed that the permeate turbidity value was different for the initial juice for both membranes (p b 0.05). The results illustrated that the turbidity decline in M1 was greater than for M2 because the lower pore size caused greater fouling layer which act as a secondary membrane and this thicker fouling layer had more potential to remove turbid components such as phenolic components. The total soluble solids (TSS) of the fresh pomegranate juice and permeate are shown in Table 1. Statistical analysis showed that

Fig. 5. Changes in fouling index during the membrane processing with two membranes, M1 and M2.

?uxe was 0.017 m/h but dropped to 0.005 m/h and 0.003 m/h after 45 min in membranes with pore sizes of 0.45 μm (M2) and 0.22 μm (M1), respectively. As the experiment proceeded, the slope of the ?ux decreased in both membranes.

Fig. 6. SEM images of M2: (a) new membrane (× 1000); (b) processed membrane (×1000); (c) cross section of new membrane (×2000); (d) cross section of processed membrane (× 1000).

H. Mirsaeedghazi et al. / Desalination 264 (2010) 243–248

247

Fig. 7. SEM images of M1: (a) new membrane (× 2000); (b) processed membrane (×2000); (c) cross section of new membrane (×500); (d) cross section of processed membrane (× 1000).

permeate TSS was signi?cantly different from fresh juice TSS for both membranes (p b 0.05). The decline of fruit juice TSS after passing through the membrane was a result of the removal of components that cause turbidity in fresh juice. As already mentioned, the elimination of turbid components in M1 was greater than for M2 because the decrease in TSS in M1 was greater than for M2. Color intensity of the fresh juice before and after membrane processing is illustrated in Table 1. As seen, color intensity decreased after permeation by the membrane and the removal of large particles that may play a role in color intensity. The acidity and pH of the M1 samples in M1 and in both M1 and M2, respectively, (Table 1) were signi?cantly different (p b 0.01) because of the impact of the membrane on the elimination of components responsible for the acidity of the samples.
Table 2 Amount of hydraulic diffusivity at each stage in M1. Stage After pomegranate juice processing After water washing After alkaline and acidic solutions washing Hydraulic diffusivity (× 10? 7 m s? 1 kPa? 1) 0.27 0.69 1.67

Viscosity of the juice decreased after passing through the membrane and the removal of the large particles and increase in the dilution of the permeates in both membranes.

3.3. Role of membrane processing in change of juice particle sizes Particles size distributions change during membrane processing. Fig. 3 shows the fresh juice particle size distribution in M2. As seen in this ?gure, particle size predominantly ranged from 0.3 to 24.6 μm. Also, permeate particle size distribution was between 0.39 and 1.78 μm. Thus, it can be deduced that membrane processing plays a major role in reducing particle size in pomegranate juice. The particles removed were most likely those causing turbidity. There were no

Table 3 different resistance produced in clari?cation of pomegranate juice with M1 after 120 min process.

Rt (× 1013) (m? 1) 3.21

Rc (× 1013) (m? 1) 1.99

Rfrev (×1013) (m? 1) 0.74

R?rr(×1013) (m? 1) 0.47

Rc/Rt (%) 61.99

Rfrev/Rt (%) 23.05

R?rr/Rt (%) 14.64

248

H. Mirsaeedghazi et al. / Desalination 264 (2010) 243–248

detectable particles in permeate obtained with M1 because of its smaller pores compared to M2. 3.4. Membrane fouling measurements To measure membrane resistance, the membrane was run with water and its ?ux was found to be 0.7 kg/m2s in M1 and 10.6 kg/m2s in M2. Thus, according to Eq. (3), membrane resistance was 6.14 × 107 m2/kg for M1 and 4.06 × 106 m2/kg for M2. Water permeate ?ux decreases after running with pomegranate juice, as illustrated by membrane fouling. Permeate ?ux decreased over time and, thus, membrane fouling increased over time and became 4.43 × 1010 m2/kg for M1 and 1.29 × 1010 m2/kg for M2 after 45 min. Fig. 4 shows changes in fouling resistance during membrane processing. It can be deduced that the fouling produced in the ?rst stages of processing was greater than in the ?nal stages of membrane treatment and the difference in fouling resistance between M1 and M2 in the early stages of membrane processing was smaller than for the ?nal stages. This agrees with the permeate ?ux difference between the two membranes, as illustrated in Fig. 2. Measurement of the fouling index in both membranes showed that this index decreased over time; because, the cake formation as a major reason of ?ux decline was created in the ?rst stages of juice treatment. During the membrane clari?cation, this mechanism of fouling is replaced with other mechanisms which have fewer effects on membrane fouling [19]. Decreasing of the fouling index supported the fact that fouling intensity decreases over time. In the ?nal stages, fouling reaches a steady state, as shown in Fig. 5. 3.5. Analysis of fouling by SEM Analysis of the membranes by SEM revealed the membrane microstructure. Fig. 6(a) shows the new membrane microscopic structure in M2. As seen, the membrane pores are open and show no fouling. After processing with pomegranate juice, the membrane surface was again analyzed by SEM, as shown in Fig. 6(b). All pores were fouled by turbid components. To understand whether these components enter the pores, a cross section of the membrane was examined by SEM, as shown in Figs. 6(c) and (d). Here, the turbid components do not enter the pores and thus the fouling forms a caked layer. These results are similar to the SEM analysis of M1 (Fig. 7). As illustrated, the predominant fouling was caked fouling, which agrees with results obtained from experiments determining resistance in M1 after processing with pomegranate juice (120 min). There, caked resistance made the greatest contribution (62%) to total resistance (Tables 2 and 3). Caked and reversible fouling can be removed by washing with water, alkaline and acid detergents. Tables 2 and 3 show that this resistance is part of the 85% of total resistance. Only 15% of total resistance could not be removed by washing. 4. Conclusion The results of this study show that micro?ltration can clarify pomegranate juice without causing signi?cant changes in its chemical properties when compared to enzymatic methods. Clari?ed pome-

granate juice obtained from membrane processing has a more desirable color than fresh juice and can improve the marketability of the product. Fouling is the greatest hindrance to the industry's adoption of membrane use. As illustrated, a major component of fouling produced on the membranes was caused by formation of a caked layer. However, reversible fouling that can be easily removed by washing with water, acid and alkaline solutions. The recovered membrane can then be reused, since its ?ux was shown to be similar to that of the ?rst process. Nomenclature Symbol pb pressure before the membrane, Nm? 2 pa pressure after the membrane, Nm? 2 A membrane area, m? 2 mp permeate weight, kg T time, s Jp permeate ?ux, m(h? 1) ΔP transmembrane pressure, Nm? 2 J0 Initial permeate ?ux, m(h? 1) Rm membrane resistance, m? 1 μw viscosity of water, Nsm? 2 Jw permeate ?ux of water, m(h? 1) μp viscosity of permeate, Nsm? 2 Rf fouling resistance, m? 1 b fouling index Rc cake resistance Lp hydraulic permeability R?rr irreversible fouling resistance Rfrev reversible fouling resistance

References
[1] M. Maskan, J. Food Eng. 72 (2006) 218–224. [2] G. Türk, M. S?nmez, M. Aydin, A. Yüce, S. Gür, M. Yüksel, E.H. Aksu, H. Aksoy, Clin. Nutr. 27 (2008) 289–296. [3] M. Aviram, M. Rosenblat, D. Gaitini, S. Nitecki, A. Hoffman, L. Dornfeld, N. Volkova, D. Presser, J. Attias, H. Liker, T. Hayek, Clin. Nutr. 23 (2004) 423–433. [4] B. Fuhrman, N. Volkova, M. Aviram, J. Nutr. Biochem. 16 (2005) 570–576. [5] Z. Borneman, V. G?kmen, H.H. Nijhuis, Sep. Purif. Technol. 22–23 (2001) 53–61. [6] F. Vaillant, M. Cisse, M. Chaverri, A. Perez, M. Dornier, F. Viquez, C.D. Mayer, Innov. Food Sci. Eng. Technol. 6 (2005) 213–220. [7] M.V. Matta, R.H. Moretti, L.M.C. Cabral, J. Food Eng. 61 (2004) 477–482. [8] F. Vaillant, A. Millan, M. Dornier, M. Decloux, M. Reynes, J. Food Eng. 48 (2001) 83–90. [9] S.T.D. de Barros, C.M.G. Andrade, E.S. Mendes, L. Peres, J. Membr. Sci. 215 (2003) 213–224. [10] F.Y. Ushikubo, A.P. Watanabe, L.A. Viotto, Desalination 200 (2006) 546–548. [11] A.P. Watanabe, F.Y. Ushikubo, L.A. Viotto, Desalination 200 (2006) 339–340. [12] M.I. Gil, F.A. Tomas-Barberan, B. Hess-Pierce, D.M. Holcroft, A.A. Kader, J. Agric, Food Chem. 48 (2000) 4581–4589. [13] A. Fadavi, M. Barzegar, M.H. Azizi, M. Bayat, Food Sci. 11 (2005) 113–119. [14] M.C.V. Vela, S.?. Blanco, J.L. García, Desalination 184 (2005) 347–356. [15] M.C.V. Vela, S.?. Blanco, J.L. García, E.B. Rodríguez, Desalination 198 (2006) 303–309. [16] K. Riedl, B. Girard, R.W. Lencki, J. Membr. Sci. 139 (1998) 155–166. [17] A. Cassano, L. Donato, E. Drioli, J. Food Eng. 79 (2002) 613–621. [18] R. Ferrarini, A. Versari, S. Galassi, J. Food Eng. 50 (2001) 113–116. [19] H. Mirsaeedghazi, Z. Emam-Djomeh, S.M. Mousavi, A. Aroujalian, M. Navidbakhsh, Int. J. Food Sci. Technol. 44 (2009) 2135–2141.


相关文章:
更多相关标签: