Docstoc

PASTEURIZATION OF ORANGE JUICE BY MEMBRANE FILTRATION

Document Sample
PASTEURIZATION OF ORANGE JUICE BY MEMBRANE FILTRATION Powered By Docstoc
					                                                             2nd Mercosur Congress on Chemical Engineering
                                                       4th Mercosur Congress on Process Systems Engineering


           PASTEURIZATION OF ORANGE JUICE BY MEMBRANE
                                                   FILTRATION
                                            Beatriz Castro* and Patricia Gerla
                         Departamento de Operaciones Unitarias en Ingeniería Química e Ingeniería
                         de Alimentos, Instituto de Ingeniería Química, Universidad de la República,
                                                     Montevideo-Uruguay




        Abstract. Fresh orange juices are usually pasteurized by heating, with the undesirable thermal effects that
        generate loss of vitamins and flavors, color degradation and a taste characterized as “cooked”. If the fresh
        juice is submitted to clarification by micro or ultrafiltration, the microorganisms and part of the orange pulp
        are retained by the membrane and the filtrate, containing the small molecules, vitamins, minerals, sugars and
        flavors, resulted pasteurized and subsequently any thermal treatment can be avoid. At this work, orange juice
        was clarified alternatively by microfiltration or ultrafiltration. A part of the concentrated was pasteurized by
        heat and after it mixed with the filtrate in order to reconstitute the natural appearance of the product. The
        resistances present during the microfiltration and ultrafiltration operations were studied, the suitable
        operation parameters that minimize the resistances were defined and a model for prediction of flux of
        filtrated was obtained. Instrumental and sensorial analyses were performed in the juices, in order to compare
        the stability and the nutritional and organoleptic characteristics of the products obtained with the different
        membranes. The products also were compared with fresh squeeze and thermal pasteurized orange juice. The
        suitable operation conditions, determined at this work, to minimize the resistances present under the
        ultrafiltration and microfiltration operations, the filtrate flux obtained with the selected conditions and the
        results of the comparison of stability and nutritional and organoleptic properties of the different products are
        presented.


   Keywords: Juice Pasteurization, Membrane Clarification, Juice Ultrafiltration


1. Introduction

   Fresh fruit juices are beverages of high nutritional value because they contain vitamins, minerals and
antioxidants. When the juice is pasteurized by heat application the thermal damage and the chemical oxidation
of the more sensitive components reduce the final quality of the product. The thermal pasteurization also
produces modifications in the juice that derive in taste and flavors deterioration. Microfiltration (MF) and
ultrafiltration (UF) membranes retain microorganisms while the smaller molecules like vitamins, sugar and
antioxidants pass trough the membranes with the water. Therefore, with a correct design of the membrane
operation, the thermal pasteurization of the permeate flow, the juice, can be avoided minimizing loss of
termolabile compounds, including aromatic substances.
   Clarified juices resulting of microfiltration or ultrafiltration processes can be the final product or can be
submitted to a concentration process in order to obtain a product suitable as ingredients in juices and beverages
as multi fruits cocktails. Today the research of integrated membrane processes in the fruit juice industry is in
constant expansion. News membrane technologies continuously arise and always in the direction of accomplish


    *
     Beatriz Castro.
    Address: Depto de Operaciones Unitarias en Ing. Química e Ing. De Alimentos, Instituto de Ingeniería Química,
    UDELAR, J. Herrera y Reissg 565, CP: 11300, Montevideo-Uruguay. E-mail: beatrizc@fing.edu.uy
                                                                                                                           1
                                                         2nd Mercosur Congress on Chemical Engineering
                                                   4th Mercosur Congress on Process Systems Engineering


clarification, pasteurization and concentration of the juices improving product quality and reducing energy
requirements (Cassano et al., 2003).
   There are products, like apple juices, that after clarification by membranes present a crystal clear aspect that
is suitable for the final commercialization. This is not the case of the orange juice that the final consumer
prefers to drink cloudy and with the inclusion of the natural orange pulp. When this juice is pasteurized by
ultrafiltration or microfiltration the membrane not only retains the microbial components but also the orange
pulp and consequently a part of the color. The solution to this problem, in order to preserve the pulp and the
aspect of the fresh juice, is the reincorporation in the filtrate of at least a fraction of the retentate. The microbial
charge has been concentrated in the retentate after the membrane filtration and it is necessary the thermal
pasteurization of the retentate before mixing it with the filtrate. The thermal degradation that can derive of this
operation will only affect a small fraction of the components of the juice because the bulk of the product, the
filtrate, is not submitted to the heat action.
   At this work, orange juice was clarified alternatively by microfiltration or ultrafiltration with membranes of
two different cut-offs. A part of the concentrated was thermal pasteurized and after it mixed with the filtrate in
order to reconstitute the natural appearance of the product. Instrumental and sensorial analyses were performed
in the juices, in order to compare the stability and the nutritional and organoleptic characteristics of the products
obtained with the different membranes. The products also were compared with fresh squeeze and thermal
pasteurized orange juice.
   The design of a membrane process able to substitute in the fruit juice industry the traditionally filtration and
pasteurization methods implies the optimization of filtrate flux that decides the size of the equipment or the time
of the process. The four principal parameters affecting filtrate flux in any membrane process are: applied
transmembrane pressure (PT), turbulence in the feed channel and feed concentration and temperature. In an ideal
membrane, with uniformly distributed and evenly sized pores, with no fouling and negligible concentration
polarization, it is generally believed (Cheryan, 1998) that the best description of fluid flow through porous
membranes is given by the Hagen-Poiseuille law as presented in Eq. (1).


        K
J=        PT                                                                                                        (1)
        µ


   where K is a constant value, dp2ε/32λ, for a given membrane, J denotes the flux rate through the membrane
(l/hm2), PT the applied transmembrane pressure (kPa), µ the filtrate viscosity (kg/hm), λ membrane thickness
(m), ε membrane porosity and dp membrane pore diameter (m). According to this model, filtrate flux is directly
proportional to the applied pressure and inversely proportional to the fluid viscosity. However, this is true only
under certain conditions such as low pressure, low feed concentration without foulants and high turbulence in
the feed channel. In the pressure controlled region the Resistance Model or “resistances in series” model that can
be expressed mathematically as shown in Eq. (2), seems to give a better description of the behavior of filtrate
flux.


                                                                                                                      2
                                                        2nd Mercosur Congress on Chemical Engineering
                                                  4th Mercosur Congress on Process Systems Engineering


            PT
J=                                                                                                               (2)
     R MS + R F + R G


   where RMS is the intrinsic membrane resistance to the feed solution, RF is the resistance due to the fouling,
and RG is the resistance due to concentration polarization.
   The concentration polarization resistance arises because species rejected by the membrane are accumulated
in a region close to the membrane surface. The feed flow acts over this region, sweeping away these molecules
from the membrane surface and therefore this resistance is depending on feed velocity.
   When fouling is the controlling resistance, the filtration velocity is not proportional to the applied pressure
and the behavior of filtrate flux is only predictable by fouling mathematical models.
     Two mathematical models represented by Eq. (3) and Eq. (3), are presented in the literature as successfully
to describe the decrease of flux in presence of fouling. Eq. (3), the Cake Filtration Model, assumes that a layer
of foulants covers the entire surface. Eq. (4), the Internal Pore-Plugging Model, assumes adsorption or
deposition of microsolutes inside the membranes pores (Cheryan, 1998).


                               1
                           −
         ⎛ 2AJ 0 bθ ⎞          2
Jθ = J 0 ⎜ 1 +
         ⎜          ⎟                                                                                            (3)
         ⎝     RM ⎟ ⎠
                      −2
         ⎛ J b´θ ⎞
Jθ = J 0 ⎜ 1 + 0 ⎟                                                                                               (4)
         ⎝     ελ ⎠


   where Jo denotes the initial flux (l/hm2), RM the intrinsic membrane resistance to pure water defined as: RM =
µwater / K (kPa/( l/hm2), A the membrane area (m2), Jθ the flux at any time θ (h), and b and b´ are constants that
characterize fouling process.
   At this work, the mass transfer resistances present during the microfiltration and ultrafiltration of orange
juice were identified, the suitable operation parameters that minimize the resistances defined and a model for
prediction of flux of filtrated was obtained.


2. Materials and Methods

   Raw orange juice coming from several industries of the country was frozen for conservation, and thawed at
room temperature at the moment of the clarification that was performed using the equipment schematized in Fig.
(1). Three different filtration cartridges, two provided with spiral membranes (ultrafiltration, cut-off 30 and 300
kDa, polysulfone, Millipore), and the third one supplied with an asymmetric plane membrane (microfiltration,
0.2 µm, polysulfone, Filterite) were utilized. The process was performed at bench scale (orange juice 20-50
liters, membrane area: 0.23 and 0.71 m2, spiral and asymmetric plane respectively). Applied transmembrane



                                                                                                                   3
                                                          2nd Mercosur Congress on Chemical Engineering
                                                    4th Mercosur Congress on Process Systems Engineering


pressure and recirculation rate were regulated by the valve indicate in the schema in Fig. (1), and by the feed
velocity, supplied for a variable speed pump from a feed tank provided with a thermostat.
   The pressure applied through the membrane was measured with two manometers and the flow of filtrate and
retentate was determined by measuring the filtrate and retentate volume collected during certain time. After each
filtration the membranes were cleaned in line according to the instructions provided by the producer and until
the original resistance of the membranes to the filtration of pure water was reached.
   The resistance to filtration was calculated as PT divided by the filtrate flux. The concentration factor (FC) was
evaluated as the gradient between the initial volume of orange juice and the volume of retentate.
   The experimental data of potentials and flux obtained during the different runs were used to determine the
nature of the principal resistances present during the filtration. The parameters of the theoretical models were
calculated with a regression program that optimized the fitting of the experimental data to the values predicted
by the model.




            Figure (1): Schema of the equipment used at ultra and microfiltration, T: thermometer, P: peristaltic
                                          pump, V: valve and M: manometer


   Four runs were performed: The raw orange juice was filtrated in experiment 1, through the 0.2 µm
membrane, in experiment 2 through the 300 kDa membrane and in experiments 3 and 4 the filtrate resulting of
runs 1 y 2 was filtrated through the 30 kDa membrane. The temperature of the process was kept between 20 and
22 ºC.
   All juices were stored at 8ºC in transparent PVC bottles, except by the commercial product that was stored in
its own Tetra pack carton, and before the orange juice clarified with membranes was submitted to the sensorial
panel, it was supplemented with 15% of the pulp obtained in the retentate and previously thermal pasteurized
(100º C, 10 minutes).




                                                                                                                    4
                                                                              2nd Mercosur Congress on Chemical Engineering
                                                                        4th Mercosur Congress on Process Systems Engineering


   The fresh, raw, commercial and filtrated juices were submitted to chemical analyses for determining Total
Soluble Solids, Ascorbic acid and pH and to sensorial analyses to compare the organoleptic properties. The
concentration of total solids (ST) was determined by duplicate using a Microwave Moisture/Solids Analyzer,
LabWave9000 (CEM Corporation) following the instructions provided by the fabricant and with 0.1 mg
accuracy. Vitamin C (ascorbic acid) was determined by HPLC using a SHIMADZU equipment with a
SPHEREX 10 NH2 column (Mobil phase: acetonitrile - NH4H2PO4 0.01 M (70:30), temperature: 25ºC, flow:
1ml/min, λ: 254 nm). The pH was measured with a Cole-Parmer 8350-series pH Data logger. A panel with 10
integrants that evaluated for each juice the following attributes: flavor, taste, palatability, color, acidity and
sweetness, as very acceptable, acceptable, fairly acceptable or unacceptable, performed the sensorial analyses
immediately after the filtration and seven days later. The results of the sensorial analyses indicate the average
percentage approbation for each evaluated attribute, considering the judgment very acceptable as 100 % of
approbation and the judgment unacceptable as 0% approbation. The overall quality of the product was
evaluated, at this work, by out average of the percentage approbation achieved for each attribute.


3. Results and Discussion

3.1 Resistances to filtration

   The filtration of raw orange juice, when performed with 0.2 µm plane asymmetric or 300 kDa spiral
membranes, originated a decrease of the initial filtration flux, Fig. (1), that could no be controlled by changing
pressure. These results pointed that flux decrease under operation was principally due to fouling and could be
described by models considering these phenomena, similar to the models presented at Eq. (3) and Eq. (4)



                                                                                                 experimental 0.2 micras
                                  80
                                                                                                 model prediction 0.2 micras
          Filtrate Flux (lh m )
         -2




                                                                                                 experimental 300 kDa
                                  60
         -1




                                                                                                 model predition 300 kDa


                                  40


                                  20


                                   0
                                       0.0        0.1         0.2         0.3        0.4         0.5         0.6        0.7    0.8

                                                                                 Time (h)


                                             Fig. 1. Experimental and predicted flux during filtration of raw orange juice




                                                                                                                                     5
                                                            2nd Mercosur Congress on Chemical Engineering
                                                      4th Mercosur Congress on Process Systems Engineering


   Figure 1 shows that in this case, the Eq. (4) predicts values that fit with the experimental data with errors
smaller than 5%. Eq. (4) becomes:


                                −2
Jθ = 300(1 + 21θ )                                                                                                  (5)
                                −2
Jθ = 150(1 + 24θ )                                                                                                  (6)


for the filtration with the 0.2 µm and the 300 kDa membranes respectively.
   The fit of the experimental data to the model prediction allowed the conclusion that under the concentration
of raw orange juice with the described membranes, decreasing in filtrate flux was principally due to fouling and
that the major fouling occurred inside the pores of the membrane.
   When the raw orange juice filtrated with the 0.2 µm or the 300 kDa membranes was submitted to ultra
filtration with the 30 kDa spiral membrane, the filtrate flux showed a performance characteristic of the presence
of a resistance originated by polarization of the concentration. The figure 2 shows how the resistance depended
under the ultrafiltration of the tangential feed velocity (vt) over the membrane.

                                4


                                3
               R (kPa l h m )
               2
               -1




                                2


                                1
                                                                                               160 l/h
                                                                                               100 l/h
                                0
                                     1   3                5                7               9               11

                                                                 FC

                  Fig. 2. Resistances during filtration with 30 kDa spiral membrane of prefiltrated orange juice.


   Figure 2 shows that the value of the resistance to the filtration, for a value of the concentration factor (FC)
approximately with 8, diminished almost 60%, when the tangential feed velocity was increased from 100 to 160
l/h. The Fig. 2 also shows how the value of the resistance to the filtration of the prefiltrated orange juice grows
under all the concentration slower at the higher tangential feed velocity than at the lower one.
   When the ultrafiltration of orange juice was performed at the highest tangential feed velocity (vt) provided
by the used equipment, 160 l/h, the model presented at Eq. 4 takes the form of Eq. (7) and Eq. (8) and described
again with an error lesser than 5%, the flux under the ultrafiltration with the 30 kDa membrane,



                                                                                                                     6
                                                                      2nd Mercosur Congress on Chemical Engineering
                                                                4th Mercosur Congress on Process Systems Engineering


                                   −2
Jθ = 40(1 + 1.6θ )                                                                                                          (7)
                                         −2
Jθ = 45(1 + 0.09θ )                                                                                                         (8)


of the orange juice prefiltrated with the 0.2 µm and the 300 kDa membranes respectively. The experimental
values and the values predicted by the models are presented in Fig.(3).




                                                                            model prediction 0.2 micras+30 kDa
                                         40
                                                                            experimental 0.2 micras+30kDa
                 Filtrate Flux (lh m )
                -2




                                                                            model prediction 300 kDa+30 kDa
                                         30
                -1




                                                                            experimental 300 kDa+30 kDa

                                         20


                                         10


                                              0
                                                  0.0   0.2   0.4     0.6         0.8         1.0         1.2

                                                                      Time (h)

         Fig. 3. Experimental and predicted flux during ultrafiltration of prefiltrated orange juice with a 30 kDa spiral



3. 2 Characteristics of the orange juices.
   Total soluble solids, (ST), pH and Ascorbic acid content were determined in the fresh, raw, commercial and
filtrated orange juices. Also sensorial analyzers were performed in order to compare the organoleptic quality of
the different processed products.
   The content in total soluble solids in the juices filtrated with membranes was only slightly lower that the
corresponding content in the raw juice and similar to he content on the fresh squeezed juice (Table 1). The
commercial orange juice was prepared, since declaration in the Tetra Pack carton, with concentrated orange
juice, water and sugar. Consequently it can be followed that the solid content coming from the fresh orange fruit
is in the commercial considerably lesser than in the raw and membrane filtrated juices.
   No difference between the pH values of the different juices were noted neither immediately after the
filtration nor after the storing at 8ªC. The pH values stay in all the cases between 3.4-3.5.
   The table 2 also shows the velocity of destruction of the acid ascorbic content in the different orange juices
stored at 8ºC during seven days. While the acid ascorbic content diminished only around 4% in the commercial
product, the lost of vitamin C reached values higher than 90% in the fresh and raw juices and values around
60% in the membrane filtrated juices

                                                                                                                              7
                                                              2nd Mercosur Congress on Chemical Engineering
                                                        4th Mercosur Congress on Process Systems Engineering



                             Table 1. Total soluble solids content in the different orange juices

                                                                                      Total Solids (%)
                  Fresh juice                                                              12.83
                  Raw juice                                                                13.17
                  Commercial juice                                                         12.52
                  Raw juice filtrated through 0.2 µm                                       12.80
                  Raw juice filtrated through 300 kDa                                      12.82
                  Raw juice filtrated through 0.2 µm +30 kDa                               12.76
                  Raw juice filtrated through 300 kDa +30 kDa                              12.77
   .


                                Table 2. Acid Ascorbic content in the different orange juices

                                                                                    Ascorbic acid (ppm)
                                                                                     Day 0 Day 7/8ºC
                      Fresh juice                                                       510         31
                      Raw juice                                                         288         17
                      Commercial juice                                                  474        450
                      Raw juice filtrated through 0.2 µm                               247         101
                      Raw juice filtrated through 300 kDa                               235         94
                      Raw juice filtrated through 0.2 µm +30 kDa                       213          89
                      Raw juice filtrated through 300 kDa +30 kDa                       212         87


   The sensorial panel preferred clearly at day 0 the flavor and the color of the raw juice before the
corresponding characteristics of the other juices. No significant differences were evidenced between the flavors
of the commercial juice and of the juices processed using membranes, but the sensorial panel found that the
color of the commercial product was better than the color of the juices filtrated with membranes. The
commercial juice had a higher approbation than the other juices both in taste and sweetness and no differences
were noted between the acidity content, except in the juice that was only clarified through the microfiltration
membrane.
   No one of the processed juices reached a total approbation, in the sense explained in section 2, by part of the
sensorial panel, the higher approbation was to the commercial product, 62 %, followed by 60 % for the raw juice
that was judged by the panel like having bad palatability (Table 3).


                     Table 3. Results of the sensorial evaluation for the different orange juices at day 0

                                                   R           C           0.2         300        0.2+30     300+30
          % Flavor approbation                     40         22            0           20          18         20
          % Taste approbation                      50         56            6           30          30         40
          % Palatability approbation              10          50           31           52          60          62
          % Color approbation                     100         44           17           20          22         20
          % Acidity approbation                   100         100          78          100          100        100
          % Sweetness approbation                  60         100          63           75          70         65
          % Average approbation                    60         62           33           50          50         51
       Note: Raw juice (R), Commercial juice (C), Raw juice filtrated through 0.2 µm, 300 kDa, 0.2 µm+30 kDa and 300 kDa+30
                                     kDa denoted as 0.2, 300, 0.2+30 and 300+30 respectively




                                                                                                                              8
                                                               2nd Mercosur Congress on Chemical Engineering
                                                         4th Mercosur Congress on Process Systems Engineering


   After the juices were stored seven days at 8ºC, the sensorial panel noted a significant decrease in the general
quality of the raw and of the commercial product as in the quality of the microfiltrated juice. The juices clarified
through ultrafiltration membranes maintained or improved slightly the quality. Almost all the characteristics of
the ultrafiltrated juices deserved a better qualification in the opinion of the panel at day seven when compared
with the day cero. The taste, the flavor and the color of the juice filtrated by the combination 300-30 kDa
ultrafiltration membranes were improved, after a week at 8ºC, in the opinion of the sensorial panel that at the
same time considered that these organoleptic characteristics had becoming worse for the raw juice.



                       Table 4. Results of the sensorial evaluation for the different orange juices at day 7

                                                      R            C          0.2         300       0.2+30      300+30
            % Flavor approbation                      25           25          6           20           0          40
            % Taste approbation                       10           10          0           30          44          52
            % Palatability approbation                0            0          25           52          56          62
            % Color approbation                       89           89          0           20          38          33
            % Acidity approbation                      0            0         25          100          50          100
            % Sweetness approbation                   20           20         25           75          81          65
            % Average approbation                     24           24         14           50          45          59
         Note: Raw juice (R), Commercial juice (C), Raw juice filtrated through 0.2 µm, 300 kDa, 0.2 µm+30 kDa and 300 kDa+30
                                       kDa denoted as 0.2, 300, 0.2+30 and 300+30 respectively




4. Conclusions

   The controlling resistance to filtrate flux under raw orange juice clarification both with the 0.2 µm
asymmetric microfiltration and with the 300 kDa spiral membrane was fouling. The filtrate flux under both
operations is well described by a model that assumes adsorption or deposition of microsolutes inside the
membranes pores. To optimize filtration velocity, high transmembrane pressures that promote fouling have to be
avoided and an enzymatic treatment to diminish pectin content can be of use.
   During the ultrafiltration with the 30 kDa spiral membrane, of the juices previously clarified through 0.2 µm
or 300 kDa, a resistance originated by polarization of the concentration was evidenced. This resistance could be
minimized by adequately selection of feed tangential velocity and applied transmembrane pressure. The applied
pressure cannot be so high that prevent the tangential flow to clean the zone, close to the filtration surface, of the
species rejected by the membrane and than otherwise accumulated over it. Inadequate combinations of those two
parameters will originate an important increase of the total membrane resistance to filtration. A growing
concentration polarization contributes to membrane fouling. After the polarization concentration resistance was
minimized the filtrate flux could be again well described by the model that assumes fouling inside the membrane
pores.
   All used membranes, 0.2 µm asymmetric plane and 30 or 300 kDa spiral, showed equal performance in the
recovery of the soluble solids of the orange juice and a similar and good performance in the recovery of ascorbic
acid. On the other hand the product obtained using only the 0.2 µm microfiltration membrane had, according to
the opinion of the sensorial panel, a poorer organoleptic quality than the product obtained with the ultrafiltration
membranes.
                                                                                                                                9
                                                          2nd Mercosur Congress on Chemical Engineering
                                                    4th Mercosur Congress on Process Systems Engineering


   The panel judged the color of the membrane filtrated juices no accord to the expectations for a commercial
product and more studies has to be done to optimize the amount of orange pulp to be included in the membranes
filtrated juice and to optimize the pasteurization process of this pulp.
   Regarding the organoleptic characteristics, ultrafiltrated juices showed a better livestock during storing at 8
ºC that raw, commercial, and microfiltrated juices and after it the panel judged that some of the properties of the
ultrafiltrated products such flavor, taste and color had been improved. Nevertheless the ascorbic acid content in
all the juice, except for the commercial one, descended hasty during storing at 8ºC. It can be concluded that the
ultrafiltration membranes, especially those with small cut-off, avoid some species associated to deterioration of
orange juices stability, to permeate with the clarified product. Regarding vitamin C stability more studies
concerning both the store of the ultrafiltrated products in packages similar to the Tetra pack containers and the
fortification of the product in vitamin C have to be done. Nonetheless the ultrafiltration of the raw juice provided
a certain protection of vitamin C content that had a less important decrease in the ultrafiltrated products than in
the raw juice.



References

Cassano, A., Drioli, E., Galaverna, G., Marchelli, R., Di Silvestro G., Cagnasso P. (2003). Clarification and Concentration
   of Citrus and Carrot Juices by Integrated Membrane Processes. Journal of Food Engineering 57, 153.
Cheryan, M. (1998). Ultrafiltration and Microfiltration Handbook. Technomic Publication, USA.




Acknowledgments

   The authors acknowledge the partial support of this work from the Comisión Sectorial de Investigación
Científica de la Universidad de la República (CSIC).
   The authors thank to Mrs. Sandra Hermida and to the Laboratorio de Análisis, IIQ, Facultad de Ingeniería,
for performing vitamin C determination.




                                                                                                                        10

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:111
posted:8/25/2011
language:English
pages:10