Cross-flow filtration of fruit juice

9. Ultrasounds

9.1 Introduction
9.2 Experiments
9.3 The experimental setup
9.4 Results
9.5 Conclusions

9.1 Introduction

Separation processes are most important operations in a great deal of industries, especially food production. In fruit juice production separation operations are constituing the major part.

High cost, due to energy consuming techniqes, cleaning, use of chemicals for floculation and sedimentation, the later now considered a health-risk and undesired impact to natural environment, has encouraged development of new microfiltration technology based on new types of membranes for liquid filtration.

The most predominant problem in membrane filtering is fouling, which in filtration of cherry juice and similar fluids incorporates a chain of events, linked to a complex composition of ingredients as polysaccarides, cromo-phenols, fat, proteins, salt and small particles and clusters of aroma carrying molecules in water.

Earlier experimental work on micro-filtration of cherry juice using reverse-asymmetric hollow-fibre-based filters has shown that low temperatur filtering is essential to obtain an aquired quality due to the fact that part of the phenols dissolve easily at normal temperature, and therefor unwantedly precipitate after cooling, when not filteret at low temperature (below approx. 8 deg.C).

The chain of fouling is considerd to involve all of the well established processes;

concentration-polarisation layer formation, building of gel-layer, forming of a sediment layer.

Membrane resistance control in cross-flow microfiltration by the use of high-power ultrasound acting on the membrane, has been reported by E.S.Tarlton and R.J.Wakeman on china-clay and anatase and by Yutaka Matsumoto, on bakers yeast and Bovine Serum Albumin BSA.

Both groups are concluding that high-power ultrasonic irradiation is efficient in controlling fouling in some cases, and state that the effect is due to local cavitational events.E.S.Tarlton and R.J.Wakeman also reported very promissing results from using electrostatic fields combined with ultrasound.

Yutaca Matsumoto recorded the important observation that major improvement in performance of the filtration can be obtained when the trans-membrane pressure is shifted to 0 (Bar) during ultrasonic irradiation.

9.2 Experiments

The two types of filters used are shown in sketches in Fig. 9.1., together with the overview of the filters, membranes and variations in experimental parameters in tabel 9.1.

Figure 9.1. The two types of filters.

Tabel 9.1. OVERVIEW OF FILTERS MEMBRANES AND VARIATION IN EXPERIMENTAL PARAMETERS

All filters have approximately the same 100cm2 active filter area.

The cherry juice was standard food-grade, press-filtered and cleared to approximately 14.3 Brix (215 FNU) and kept at 4 deg.C.
The juice was mixed of one part ultrafiltrated (sterile) water to one part juice, before being used in the experiments.

9.3 The experimental setup

The instrumental setup of the experiments was very simple. Principles of the two experimental setups for filtering with reverse-asymmetric hollow-fiber membrane filter and for the yarn-wound filters respectively are shown below. The ultrasonic irradiation-cell was produced together with Reson A/S Denmark.

The ultrasonic transducers were powered in 5 steps (0,37,74,100 and 180Watts) from a generator with the power load controlled by a pulsgated cut-off on 23kHz with 100Hz amplitude modulation, feeding a resonant load of four

(60mmŲ) transducers in parallel, (corresponding to a theoretical field of approx. 0.3-1.5Watt/cm sq. Where 0.3W/cm sq. is near and below cavitation limit in the actual case.)

Coherent values of Trans-Membrane-Pressure difference (TMP), flow and temperature were recorded manually with regular intervals.

*text in brakets is added after conf.

Figure 9.2. Principle of experimental 1.

Figure 9.3. Principle of experimental 2.

Figure 9.4. Illustration of the irradiation.

9.4 Results

A plot of Flow versus TMP (Fig.9.5.) shows a pressure driven behaviour for both types of filters.

The three displayed time-series of Flow/TMP in Fig. 9.6. (a-c) present characteristic variations in Flow-rate correlated with the impact of ultrasound.

Data for the time-series of the filtering using the low-flow octalobal yarn-wound filter has been used to analyse the influence from the previous impact on this type of filter. The result is presented in the last plot (Fig.9.7.), indicating a trend, that the larger the power of the ultrasonic field is in the actual impact, compared to the power of the last contained, previous event, the stronger the influence of the ultrasound will be.
Figure 9.5. Flow versus TMP

Figure 9.6.a Characteristic responce of Flow/TMP to ultrasound

Figure 9.6.b Characteristic responce of Flow/TMP to ultrasound

Figure 9.6.c Characteristic responce of Flow/TMP to ultrasound

9.5 Conclusions

- In filtration of natural cherry juice, the filtration resistance is observed to decrease significantly for hollow-fibre membranes as well as for yarn-wound membranes when ultrasonic power is applied above cavitation level.

- The observed decrease in filtration resistance enhance with increasing ultrasonic power.

- The effect of the cavitation seems to depend on the level of the last irradiation.

- The filter membranes are expected to be influenced only when cavitation takes place in the vicinity of the membrane surface and the fouling material.

- There exists an upper limit of the ultrasound field having effect, expected to dependent on the fouling speed and filter design. Above and near this upper limit the filter membrane material is susceptible to fatigue, as no fouling protect the membrane surface

- No upper limit could be reached with the present equipmemt for the yarn-wound membranes. However a limit of 0.7W/cm2, was found for the hollow-fiber membrane (polysulphone)

- The standard quality parameter of clearity (Brix and Fnu-numbers) showed 3 x reduction using the hollow-fibre membranes. The high flux yarn-wound membrane resulted in 20% reduction, whereas the low-flux yarn-wound membrane resulted in 2.5 x reduction at low TMP (0.03Bar).

- Based on subjective evaluation, no significant change in quality of smell or taste was found. However, the low-flux yarn-wound membrane seems to result in a small fade in taste.

- High-power ultrasound, as a means to control fouling in microfiltration membranes seems be a theoretically feasible method. However, further investigation ( with special emphasis on materials in membrane and filter design, involving various types of membranes with different filter design techniques and applications), is needed to surmount the non-trivial technical difficulties in design of larger filters.