Preparation and characterization of inexpensive kaolin hollow fibre membrane (KHFM) prepared using phase inversion/sintering technique for the efficient separation of real oily wastewater

2.4.1 XRD analysis

X-ray diffraction (XRD, X’pert Pro α1, Philips, Amsterdam, The Netherlands) was conducted in ambient conditions with CuKα radiation (λ = 1.5406 Å) to measure the crystallinity of kaolin powder and KHFM. All samples diffraction patterns were recorded from 10° to 80° 2θ with a step size 0.026° and step time of 50 s, operated at 40 kV and 30 mA with a fixed 1/4° anti-scatter slit.

2.4.2 FTIR analysis

Fourier transform infrared spectroscopy (FTIR, Spectrum 100, Perkin Elmer, Waltham, MA) was used to measure the infrared spectrum of absorption for kaolin powder and KHFM. The samples were measured without the addition of KBr in the wavelength range of 4000–600 cm⁻¹ with a resolution of 2 cm⁻¹. The sample compartment was evacuated during acquisition and the contact between the sample and the attenuated total reflectance (ATR) diamond crystal was 2 mm diameter.

2.4.3 Thermogravimetric analysis

The thermal stability of the various phases in kaolin powder in the transformation into KHFM through sintering process was studied by thermogravimetric analysis (TGA, TG8120, Rigaku, Japan). The powder was thermally analyzed by placing approximately 25 mg of the powder in an alumina crucible, subjected to a linear heating ramp between 50 and 800 °C at a rate of 10 °C/min and a cooling rate of 50 °C/min.

2.4.4 BET analysis

The specific surface area of kaolin powder was determined by N₂ adsorption/desorption isotherms at − 196 °C with a Micromeritics instrument (BET, Belsorp Max, BEL Japan Inc., Osaka, Japan). The measurements were performed in the relative pressure range from 0 to 0.99. The classic relative pressure range (P/P₀) of 0.05–0.30 was chosen to determine the specific BET surface area. The adsorption/desorption isotherms were used to calculate the pore size distributions with Barrett–Joyner–Halenda (BJH) method.

2.4.5 Morphological studies

The structures of the cross-sectional areas and surfaces of KHFMs were observed using a scanning electron microscopy (SEM) and field emission scanning electron microscopy (FESEM), respectively, operating at 30 kV at various magnifications. Prior to the SEM and FESEM testings, all KHFMs were sputter-coated with a 0.5 nm gold layer.

2.4.6 Mechanical strength

The mechanical strength of KHFMs were determined by three-point bending strength method. The three-point bending tests were carried out with an Instron Micro Tester 5848 with a load cell of 2 2 kN (Instron Calibration Laboratory, United Kingdom). Each KHFM with an average length of ∼25 mm was placed on a span of 5 cm and was loaded at a crosshead speed of 0.25 mm/min until fracture occurred. The step was repeated for three times for each sample. The mechanical strength of each KHFM was calculated using the following equation:

2.4.7 Porosity and pore size distribution analysis

Mercury intrusion porosimetry (MIP, AutoPore IV, Micromeritics, USA) was used to measure the porosity and pore size distribution of the KHFMs prepared at different sintering temperatures and addition of pore agents. The KHFMs were broken into pieces and transferred to the 5 cc penetrometer sample holder (Micrometeritics, USA) which was then pressurized from 38.6 to 4.2 × 10⁶ mbar for mercury intrusion.

3.2.1 Effect of kaolin content

Fig. 3 shows the SEM images of KHFM prepared from ceramic suspensions with kaolin contents of 35, 37.5, and 40 wt%. The ceramic suspensions with kaolin contents of less than 35 wt% and more than 40 wt% were also investigated, but unfortunately, the reduction of the kaolin content in the ceramic suspension below 35 wt% resulted in an extremely low viscosity, inhibiting the ceramic suspension to be formed into hollow fiber configuration during phase inversion-based extrusion process. Meanwhile, the ceramic suspensions containing kaolin content of more than 40 wt% produced a very viscous suspension that clogged the spinneret. Therefore, the SEM images represent the successfully fabricated KHFMs through phase inversion/sintering technique, with the kaolin content of 35–40 wt%. A detailed discussion on the viscosity of the ceramic suspension at different kaolin contents was presented in our previous work (Hubadillah et al., 2017a). The KHFM with 35 wt% kaolin content consisted of an asymmetric structure of long finger-like voids originating from the inner KHFM surface and occupying up to 70% of the KHFM thickness, with the remaining 30% KHFM region was occupied by a sponge-like layer. The KHFM with 37.5 wt% kaolin content exhibited the same asymmetric structure as that of the KHFM with 35 wt% kaolin content. However, the finger-like voids decreased from 70% to about 40%, thus increasing the viscosity of ceramic suspension. Finally, the finger-like voids were almost diminished when 40 wt% kaolin content was added into the ceramic suspension.

Close

Fig. 3 SEM images of kaolin hollow fiber membrane (KHFM) prepared at different kaolin contents.

Export to PPT

In order to study the effect of kaolin content towards the membrane structure, the viscosity of the ceramic suspension containing kaolin content of 35–40 wt% were further measured using a viscometer test as shown in Fig. 4. As reported by Kingsbury and Li (2009), an increase in viscosity of ceramic suspension could inhibit the mechanism of viscous fingering, thus reducing the formation of finger-like voids in the hollow fiber structure. In fact, these finger-like voids could be diminished when the alumina suspension reached its viscosity threshold of 12,000 cP at the shear rate of 30 s⁻¹. Another similar structure was also obtained by Othman et al. (2010) in which its viscosity threshold was 39,100 cP for nickel oxide-based suspension. Consequently, the viscosity threshold obtained in this work for kaolin suspension was 16,111 cP at the shear rate of 30 s⁻¹ due to finger-like voids-free structure in KHFM prepared using 35 wt% ceramic loading in the dope suspension.

Close

Fig. 4 Viscosity of ceramic suspension at different kaolin content.

Export to PPT

The mechanical strength of the KHFM prepared at different kaolin contents with different structures is shown in Fig. 5. As the kaolin content in the suspension was increased where the finger-like voids size decreased, the mechanical strength increased. The KHFM prepared at 35 wt% showed the lowest mechanical strength compared to that of the KHFM prepared at 40 wt%, in which the finger-like voids were absent. This indicates that the symmetric structure with no finger-like voids, at the same time, improved the integrity of the sponge-like voids region, and more likely to enhance the mechanical strength of the membrane. Increasing the kaolin content from 35 to 37.5 wt% showed a small enhancement of mechanical strength with an increment of 1.95 MPa only. Thereby, the KHFM prepared at 35 wt% was chosen to minimize the production cost of the KHFM.

Close

Fig. 5 Mechanical strength of kaolin hollow fiber membrane (KHFM) prepared at different kaolin contents.

Export to PPT

3.2.2 Effect of bore fluid flow rate

Fig. 6 illustrates the SEM images of the KHFM prepared at different bore fluid flow rates which were 3, 5, 10, 15, and 20 mL/min at a fixed dope extrusion rate of 10 mL/min, air-gap distance of 50 mm, kaolin content of 35 wt%, and sintering temperature of 1200 °C. The results indicate that all the KHFMs possessed finger-like voids originating from the inner surface and a sponge-like layer. The KHFM prepared at the bore fluid flow rate of 3 mL/min exhibited the thickest cross section. Consistent with the findings by Bonyadi et al. (2007) and Hubadillah et al. (2016b), an irregular shape of hollow bore was obtained. An improvement of hollow bore shape with long finger-like voids was obtained when the bore fluid flow rate of 5 mL/min was applied. Therefore, a perfect circular hollow bore of KHFM was obtained at a flow rate of 10 mL/min. For the purpose of studying the effect of bore fluid flow rate on the morphology of KHFM, the bore fluid flow rate was increased to 15 and 20 mL/min. As a result, the thickness of the KHFMs and finger-like voids were both decreased. There are two factors attributed to this phenomenon (Li et al., 2016): (1) hydrodynamic force as described schematically in Fig. 7 and (2) solidification rate. With reference to hydrodynamic force, insufficient bore fluid flow rate cannot withstand a hydrodynamic force from inside the lumen (V_BF) against dope extrusion (V_E), inhibiting the perfect circular shape of hollow bore. This study found a ratio of V_BF:V_E of 1:1 as to prepare a successful ceramic membrane in perfect hollow fiber configuration through phase inversion technique. As for solidification rate, insufficient bore fluid flow rate disturbs the demixing process at the inner surface of nascent KHFM and allows only partial solidification to take place. In other words, the bore fluid flow rates of 3 and 5 mL/min were not high enough to withstand the inward radial forces, and vice versa for the bore fluid flow rate of 15 and 20 mL/min. Meanwhile, SEM images of KHFM’s inner surface showed decreasing pore size as function of increasing bore fluid flow rate. In line with the SEM result, Alobaidy et al. stated that lower BF lead to delayed liquid–liquid demixing process of phase inversion, thus induce larger pore size (Alobaidy et al., 2017) and vice versa for higher BF.

Close

Fig. 6 (A) SEM images and (B) mechanical strength of kaolin hollow fiber membrane (KHFM prepared at different bore fluid flow rates.

Export to PPT

Close

Fig. 7 Schematic picture of the exertion of outward force from the inner surface of nascent KHFM by the bore fluid flow rate.

Export to PPT

Fig. 8 shows the mechanical strength of the KHFM prepared at different bore fluid flow rates from 3 to 20 mL/min. Other than finger-like voids, the mechanical strength is also dependent on the membrane thickness as the mechanical strength decreases with the decrease of membrane thickness caused by the increment of bore fluid flow rate. The KHFM prepared at 3 mL/min bore fluid flow rate had the highest mechanical strength of 17.1 MPa as well as the thickest cross section. In terms of geometrical and mechanical strengths, the KHFM prepared at 10 mL/min was sufficient to be used in water application. The use of 10 mL/min of bore fluid flow rate is equal to the dope extrusion flow rate (10 mL/min) which gives a ratio of 1:1 as discussed previously. Our previous work also used the same ratio to produce the ceramic membrane from rice husk ash with perfect hollow fiber configuration (Hubadillah et al., 2017b).

Close

Fig. 8 Mechanical strength of kaolin hollow fiber membrane (KHFM) prepared at different bore fluid flow rates.

Export to PPT

3.2.3 Effect of sintering temperature

Fig. 9 illustrates the cross-sectional area and surface SEM images of the KHFM prepared at kaolin content of 37.5 wt% and sintered at different sintering temperatures of 1200–1500 °C. With increasing sintering temperature, the sponge-like voids layer became denser whereas no significant changes were observed for finger-like voids. The KHFM sintered at 1200 °C showed the most porous structure for both cross section and surface. Also, the surface was rough when the KHFM was sintered at 1200 °C. When the sintering temperature was increased to 1300 °C, the pore size of the sponge-like voids in the KHFM cross section tended to be reduced slightly and isolated pores started to grow on the surface of the KHFM. Increasing the sintering temperature up to 1400 and 1500 °C caused the sponge-like region to melt and became completely dense with isolated pores on the surface. This can be proved by the thermal analysis of kaolin powder (Section 3.1). A careful observation on the sponge-like region of the KHFM sintered at 1400 and 1500 °C reveals that many closed pores were obtained, which is undesired for water permeation.

Close

Fig. 9 SEM images of kaolin hollow fiber membrane (KHFM) prepared at different sintering temperatures.

Export to PPT

According to Gitis and Rothenberg (2016), ceramic particles induce three important stages during the sintering process which are initial neck formation, subsequent growth, and final form formation. Practically, based on the SEM images of the KHFM surface (Fig. 9C), sintering can be viewed as the rearrangement of particles on surface grains. Accordingly, diffusion between kaolin particles took part in the KHFM structure and dependable on the sintering temperature and diffusion path. As a result, the pore size of the KHFM decreased, in line with the results of most of previous studies (Tan et al., 2014; Norfazliana et al., 2016).

Sintering temperature is the crucial parameter in ceramic membrane production. In this study, the effect of sintering temperature of 1200–1500 °C on the mechanical property of the KHFM is shown in Fig. 10. The value of the mechanical strength increased from 5.64 to 14.26 MPa when the sintering temperature was increased from 1200 to 1300 °C, while the strength recorded a fourfold increase when the sintering temperature was further increased to 1400 °C. Only 12% increment was observed when the sintering temperature was increased from 1400 to 1500 °C. Two temperatures were involved in the sintering KHFM, which were 600 °C to burn off the binder and target temperature of 1200, 1300, 1400, and 1500 °C to control the pore structure and size. As stated by Silva et al. (2015), increasing the sintering temperature will lead to the pores densification, thus increasing the mechanical strength. At higher temperatures of 1400 and 1500 °C, the kaolin particles melted and fused to each other, resulting in a high mechanical strength of 55.31 and 61.86 MPa, respectively.

Close

Fig. 10 Mechanical strength of kaolin hollow fiber membrane (KHFM) prepared at different sintering temperatures.

Export to PPT

The investigation on the macrostructure of the KHFM prepared at different sintering temperatures was further demonstrated by a comparison of the mercury intrusion data as shown in Fig. 11. Fig. 11A shows the porosity of the KHFM where the porosity value decreased with the increasing sintering temperature. The highest porosity was 69.1% for the KHFM sintered at 1200 °C and this value decreased to 52.3%, 10.3%, and 7.8% for the KHFM sintered at 1300, 1400, and 1500 °C, respectively. This value can be related to the densification of pores as shown in the SEM images (Fig. 9) and the pore size distribution of the KHFM prepared at different sintering temperatures (Fig. 11B). It is important to note that the porosity is highly depending on the pore size of the KHFM. According to the pore size distribution data, two types of pores were recognized for the KHFM sintered at 1200 and 1300 °C, whereas only one peak was obtained for the KHFM sintered at 1300 and 1400 °C. For the KHFM sintered at 1200 °C, the data show a pore size distribution consisting of a peak at approximately 2.26 µm, which represents the finger-like voids and a smaller peak at 0.21 µm is believed to represent the sponge-like pores. As the sintering temperature was increased, both peaks of finger-like and sponge-like voids were still the same but with reduced intensity, indicating the densification of pores. At the sintering temperatures of 1300 and 1400 °C, the sponge-like voids peak at 0.21 µm was totally absent and the finger-like voids peak at 2.26 µm shifted to 1.51 and 1.21 µm, respectively, indicating a further pore densification.

Close

Fig. 11 (A) Porosity and (B) pore size distribution of kaolin hollow fiber membrane (KHFM) prepared at different sintering temperatures.

Export to PPT

3.4.1 Effect of calcium carbonate

To accomplish the aim of having high flux by enlarging the pore size of KHFM, the effect of CaCO₃ as pore agent added into the ceramic suspension was investigated by modifying the KHFM containing 35 wt% of kaolin: CaCO₃ at the ratios of 6:1, 5:2, 4:3, and 3:4. For example, the ratio of 6:1 represents 30 wt% of kaolin and 5 wt% of CaCO₃. Fig. 14 shows the SEM images of the KHFM prepared at different calcium carbonate contents of 5, 10, 15, and 20 wt% at the bore fluid flow rate and sintering temperature of 10 mL/min and 1300 °C, respectively. Based on the SEM image, no formation of finger-like voids can be observed in the KHFM cross section. Besides, only sponge-like voids with increasing in size were obtained, which produced KHFM with symmetric structure. The KHFM prepared with 5 wt% of CaCO₃depicted a similar structure of sponge-like voids compared to the structure of the KHFM prepared without CaCO₃ Fig. 3. The role of CaCO₃as pore agent in this work succeeded in enlarging the sponge-like voids, but not the finger-like voids. Consequently, the addition of 10 and 15 wt% of CaCO₃ increased the pose size of the KHFM. A similar trend was also found by Kaniganti et al. that prepared a flat sheet ceramic membrane from kaolin through pressing method (Kaniganti et al., 2015). This is also supported by Bouzerara et al. which revealed that higher calcium carbonate addition may be caused by the coalescence of formed smaller pores that were responsible for the formation of the larger pores (Bouzerara et al., 2006). However, further addition of 20 wt% of CaCO₃ showed a reverse trend. From the SEM image, the KHFM containing 20 wt% of CaCO₃ depicted a smaller pore size compared to that of the KHFM containing 20 wt% of CaCO₃. Interestingly, a formation of pin-hole voids on the particle neck growth of kaolin and calcium carbonate was observed.

Close

Fig. 14 SEM images of kaolin hollow fiber membrane (KHFM) prepared at different calcium carbonate addition as pore agent. Bore fluid flow rate of 10 mL/min and sintering temperature of 1300 °C.

Export to PPT

Fig. 15 displays the result of mechanical strength for the KHFM prepared with different calcium carbonate contents. The mechanical strength of the KHFM significantly decreased with the increasing calcium carbonate content, except for the KHFM prepared with 20 wt% of CaCO₃. As expected, the mechanical strength of the KHFM prepared with calcium carbonate was lower than that of the KHFM prepared without calcium carbonate. The amount of 5 wt% of CaCO₃decreased the mechanical strength by 39% from its initial value due to the pore size enlargement. A further decrease of 84% was obtained when 10 wt% of CaCO₃ was added. Increasing the calcium carbonate content to 15 wt% resulted in a nearly constant mechanical strength value, which might be due to the contradictory computational effects of grain growth and pore coalescence (Harabi et al., 2014). Thereby, the pore coalescence became a predominant factor when 20 wt% of CaCO₃ was added, giving the highest value of mechanical strength of 17.34 MPa. This value is comparable to the value obtained for the KHFM prepared without calcium carbonate (14.26 MPa). However, no rejection can be performed due to the larger pore size of the KHFM. Herein, it should be mentioned that mercury porosity analysis cannot be measured for KHFM prepared with CaCO₃ due to induce poor mechanical strength.

Close

Fig. 15 Mechanical strength of kaolin hollow fiber membrane (KHFM) prepared at different calcium carbonate addition as pore agent. Bore fluid flow rate of 10 mL/min and sintering temperature of 1300 °C.

Export to PPT

3.4.2 Effect of polyethylene glycol

Since calcium carbonate as an inorganic pore agent failed to control the desired structure of the KHFM, the addition of polyethylene glycolHe (PEG) as a polymeric pore agent at different molecular weights of 400, 1500, and 30,000 Da was investigated and the cross-sectional SEM images are shown in Fig. 16. In this study, the addition of PEG into the ceramic suspension was modified by adjusting the amount of kaolin from 35 to 33 wt% and a fixed amount of PEG at 2 wt%. The modification of the PESf content is not done because the PESf disturbance at 5 wt% will reduce the amount of binder in the ceramic suspension. Thus, kaolin powders are unable to hold each other during the phase inversion process and tend to disperse into the coagulant bath, hindering the fabrication process. Therefore, the PESf amount was remained at 5 wt%. From the SEM image, it is clearly observed that the finger-like voids increased with the increasing PEG molecular weight. The KHFM prepared wiht PEG 400 showed no formation of finger-like voids, but the formation of sponge-like voids seemed to be homogeneous. This might be because the low molecular weight PEG can easily leach out from the ceramic suspension during the extrusion process. A similar situation was faced by Wang et al. which studied the effect of PVP molecular weight (Wang et al., 1999). In addition, a high magnification on the sponge-like voids of the KHFM with PEG 400 revealed the occurrence of neck growth between the kaolin particles. At higher molecular weight, PEG behaved inversely and facilitated a rapid phase inversion process due to the enhancement of thermodynamic immiscibility (Hilal et al., 2015), enhancing the finger-like voids formation as exhibited by the KHFM added with PEG 1500 PEG 30000. These results indicate that the desired structure of KHFM can be controlled through the addition of different molecular weight of PEG.

Close

Fig. 16 (A) SEM images and (B) mechanical strength of kaolin hollow fiber membrane (KHFM) prepared at different PEG molecular weights as pore agent. Bore fluid flow rate of 10 mL/min and sintering temperature of 1300 °C.

Export to PPT

Fig. 17 shows the increasing trend of mechanical strength of the KHFM prepared with different PEG molecular weights. The additions of PEG 400 and PEG 30,000 gave the lowest and highest mechanical strengths of 42.31 and 105.34 MPa, respectively. Surprisingly, this study shows a reverse trend compared to the trend in most of the previous studies. Othman et al. (2010) indicated that the mechanical strength of the asymmetric structure of the ceramic hollow fiber membrane with finger-like voids is likely to diminish by reducing the integrity of the sponge-like structure region. This might be due to the presence of PEG in kaolin suspension that enhances the liquid–liquid demixing during the phase inversion of the KHFM, densifying the sponge-like voids. PEG is highly hydrophilic in nature and well miscible with water and the existence of PEG in dope suspension accelerates the phase inversion process (Hilal et al., 2015). An accelerated phase inversion process will create a larger finger-like voids but possess a denser surface with small pores which contributes to high mechanical strength. Thereby, the high mechanical strength of the KHFMs prepared with the addition of PEG as pore agent is well above the operating pressure used in microfiltration, ultrafiltration, nanofiltration, and reverse osmosis. It may be further increased by sintering at above 1400 °C. The mechanical strengths of the KHFM prepared with the addition of PEG as pore agent in this study are close to those reported for the expensive alumina membrane (72 MPa) (Norfazliana et al., 2016) and are significantly higher than those prepared from kaolin in flat sheet (15 MPa) through pressing method (Kaniganti et al., 2015; Bouzerara et al., 2006; Dong et al., 2009).

Close

Fig. 17 Mechanical strength of kaolin hollow fiber membrane (KHFM) prepared at different PEG molecular weights as pore agent. Bore fluid flow rate of 10 mL/min and sintering temperature of 1300 °C.

Export to PPT

The porosity and pore size distribution for the KHFM prepared with different molecular weights of PEG as pore agent is presented in Fig. 18. The increasing PEG molecular weight from 400 to 1500 and 30,000 Da reduced the porosity from 64.27% to 55.32% and 50.09%, respectively (Fig. 18A). This trend agrees well with the effect of PEG molecular weight on the flexural strength shown in Fig. 17. As stated by Fung and Wang (2014), the enhancement in mechanical strength of ceramic membrane came with a decrease in porosity. In addition, the porosity is not influenced by the formation of finger-like voids. This can be explained by the liquid–liquid demixing mechanism that enhanced the finger-like voids formation, densifying the sponge-like voids and resulting in a high mechanical strength. A similar result was also obtained by Wang et al. (1999) whom reported the effect of different PVP molecular weight. The pore size distribution curve obtained from the mercury intrusion porosimetry for the KHFM prepared with different molecular weights of PEG is shown in Fig. 18B. A significant peak at 0.063 µm that is believed to be sponge-like void, was observed in the KHFM prepared with PEG 400. A bimodal peak was obtained when high molecular weight of PEG was used, resulting from mercury intrusion of sponge- and finger-like voids. Therefore, it is concluded that the KHFM structure can be controlled by the addition of PEG as pore agent. Theoretically, the formation of finger-like voids will enhance the porosity. However, it should be mentioned that the theory is commonly referred to the membrane with dense structure (Othman et al., 2010). In our study, it was found that KHFM with no finger-like voids induced highest porosity compared to that with finger-like voids. From Fig. 19 below, it is clearly explained that KHFM prepared with PEG 400 have spongy structure for 100% its cross section. Whereas, at PEG 1500, the formation of finger-like voids were started to show but the thickness spongy structure reduce about half from its cross section. At PEG 30000, the thickness was further reduced to 20%. Another interesting observation is that the pore size of the spongy structure in KHFM prepared at PEG 1500 and PEG 30,000 are denser compared to KHFM prepared at PEG 400.

Close

Fig. 18 (A) Porosity and (B) pore size distribution of kaolin hollow fiber membrane (KHFM) prepared at different PEG molecular weights.

Export to PPT

Close

Fig. 19 Schematic diagram on the formation of KHFM’s voids as function of different PEG molecular weight.

Export to PPT