Since the antifouling action of hollow fiber membranes mainly rely on the random fiber movement in two phase flow, fiber flexibility plays an important role. There are many factors affecting fiber flexibility: length, size (diameter), structure of fiber, intrinsic material flexibility, looseness, etc.
If fiber looseness are identical, fiber length determines the fiber amplitude. The longer the fiber is the wider the range of the movement is. When identical fibers were cut to three different lengths and their movements in air-water two phase flow were observed through a video camera, the longest fiberÔÇÖs amplitude was the biggest at the same fiber looseness of 1% (Fig.1). When fiber length increased from 50 cm to 90 cm, both theoretical maximum amplitude (shown as a solid line above the bars) and actual amplitude increased. It is noticeable that the actual amplitude increased unproportionally by more than 200% while the theoretical maximum amplitude increases by only 80%. As a result, the gap between theoretical and the actual amplitudes shrunk as fiber length increased. The increased fiber movement with longer fibers would be expected to improve antifouling action (Wicaksana, 2006).
The gains from the fiber movement of long fibers are at least partially offset by the larger internal pressure losses and greater axial variation in local flux as discussed here. When 50 cm and 70 cm fibers with 0.65mm fiber OD and 1% looseness were used to filter 5 g/L yeast at 30 LMH, 70 cm fiber was fouled significantly faster at 1.5 L/min air flow rate. This means the effect of increased internal pressure loss was more significant than the gains from the increased fiber movement under the condition employed in the experiment. However, this observation cannot be generalized for other conditions with different feed water since the result can turn out differently due to the interactions among fiber length, diameter, flexibility, aeration condition, etc.
Fig. 1. The variation of fiber amplitude with fiber length: 0.65 mm fiber OD, 1% looseness, 2L/min air flow rate and 1 mm diffuser (Wicaksana, 2006).
Fiber diameter is also one of the major factors affecting fiber flexibility. It is intuitive that smaller fibers are more flexible and have larger amplitudes than larger fibers. When two fibers with different OD were compared as shown in Fig. 6, the smaller fiber with 0.65 mm OD fouled slower than the larger fiber with 2.7 mm OD. This means, the fiber movement effect was more dominant than the internal pressure drop effect under the experimental condition. In fact the internal pressure drop in this experiment is supposedly not significant since fiber lengths used in the experiment were relatively short at 510 mm and 475 mm, respectively.
Table 1. Specifications of the hollow fiber membranes used for the experiment shown in Fig. 2
|0.65 mm||0.39 mm||0.2 ╬╝m||Polypropylene||510 mm||USFilter|
|2.7 mm||1.8 mm||0.2 ╬╝m||Polypropylene||475 mm||AKZO|
┬áFig. 2. Effect of fiber diameter on membrane fouling. Scouring air flow = 7.3 L/min, medium = 5 g yeast/L (Chang, 2002a).
Fig. 3 shows an apparent link among membrane fouling rate, fiber amplitude and fiber size (Fane, 2005a). As the fiber diameter increases, the fiber amplitude reduces from 30 mm for the 0.65mm fibre OD to around 10 mm for the 2.7mm fiber OD. The membrane fouling rate mirrors this observation with the bigger fibers foul faster than smaller fibers as indicated by TMP rising rates. However, the benefit of using small fibers is limited by the greater lumen-side pressure losses as mentioned earlier. Perhaps the relatively short fiber length (500 mm) used in the experiment is responsible for the less membrane fouling with smaller fibers since the fiber amplitude effect can dominate the lumen-pressure drop effect under the condition.
Fig. 3. Fiber displacement and TMP rise (dTMP/dt) versus fiber OD. Fiber length = 50 cm, tightness = 99%, MLSS = 5g/L yeast, flux = 30 LMH (Fane, 2005a).
In conclusion, rhe random fiber movement is largely affected by the flexibility of the fiber in a given two phase flow, which is determined by membrane chemistry, diameter, wall thickness, etc.,┬ábut it can be also affected by fiber length and the┬álooseness in a module (Henshaw, 1998).
┬ę Seong Hoon Yoon