Microsoft Word - 164.docx


CHEMICAL ENGINEERING TRANSACTIONS 

VOL. 56, 2017 

A publication of 

The Italian Association 
of Chemical Engineering 
Online at www.aidic.it/cet 

Guest Editors: Jiří Jaromír Klemeš, Peng Yen Liew, Wai Shin Ho, Jeng Shiun Lim 
Copyright © 2017, AIDIC Servizi S.r.l., 

ISBN 978-88-95608-47-1; ISSN 2283-9216 

The Effect of Air Gap on the Morphological Properties of 
PSf/PVP90 Membrane for Hemodialysis Application 

Noresah Saida, Hasrinah Hasbullaha, Ahmad Fauzi Ismaila, Muhammad Nidzhom 
Zainol Abidina, Pei Sean Goha, Mohd Hafiz Dzarfan Othmana, Siti Hamimah 
Sheikh Abdul Kadirb, Fatmawati Kamalb, Mohd Sohaimi Abdullaha, Be Cheer Nga
aAdvanced Membrane Technology Research Centre (AMTEC), Universiti Teknologi Malaysia, 81310 UTM Skudai, Malaysia 
bInstitute of Molecular Medicine and Biotechnology, Faculty of Medicine, Universiti Teknologi MARA Sungai Buloh Campus, 
 Jalan Hospital, 47000 Sungai Buloh, Selangor, Malaysia 
 hasrinah@petroleum.utm.my 

Membrane morphology plays an important role in achieving high flux and excellent uremic toxin removal for 
efficient hemodialysis therapy. Hemodialysis membrane morphology correlates to spinning parameters applied 
during fabrication of membrane. The effect of air gap on the morphology and liquid separation performance of 
the polysulfone (PSf) hemodialysis membrane is investigated. PSf hollow fibre membranes were prepared via 
dry-wet spinning process from dope solution comprises of 18 wt% PSf and 4.8 wt% polyvinylpyrrolidone in N-
methyl-2-pyrrolidone. The membrane morphology was characterised using a scanning electron microscope 
(SEM) before tested with the ultrafiltration system to measure the pure water flux (PWF) and protein rejection 
using bovine serum albumin (BSA). SEM analysis revealed that the air gap does change the structure of the 
membranes due to the elongation stress because of the gravitational pull on the PSf hollow fibres. At low air 
gap (3 cm), the lower average pore size on the outer surface reduced the PWF while at high air gap (50 cm), 
the larger average pore size of membranes permitted water molecules to pass through easier and faster. It 
was observed that the PWF of the membrane increased significantly with air gap due to the increasing pore 
size. Membrane fabricated at 50 cm air gap obtained better PWF (28.45 Lm-2h-1) and protein rejection (94.47
%) compared to the membranes fabricated at 3 and 30 cm air gap. The effect of air gap enhanced the 
morphology and the performance of PSf membrane for hemodialysis application.  

1. Introduction

In hemodialysis, the membranes used usually determine the success of the treatment. Hemodialysis removes 
excess moisture and metabolic waste (such as urea and creatinine) by diffusion and convection across the 
porous membrane and restores fluid and solute balance within the patient to some extent (Abidin et al., 2016). 
Polysulfone (PSf) is one of the promising polymeric materials that is widely used in separation field. PSf 
membranes show good heat resistance, physiochemical stability, resistance to chlorine, oxidation, and 
chemical compatibility over a wide range of pH (Yamashita and Sakurai, 2015). Although PSf membrane has 
been prevalently used, its hydrophobic character remains as the main disadvantage for liquid-liquid separation 
like hemodialysis. Many studies have concluded that the membrane fouling is related to hydrophobicity as 
reviewed by Khulbe et al. (2009). Modifying the pristine PSf has become the focus point before this to improve 
the hydrophilic character, which could enhance the biocompatibility of the PSf based-membrane. 
Polyvinylpyrrolidone (PVP), which is a hydrophilic polymer, is known to show good blood compatibility and has 
been commonly used in hemodialysis membrane formulation (Bowry et al., 2011). 
In hemodialysis, the use of hollow fibre membrane is more preferable since it has a larger contact area with 
blood and the dialysate in spite of its small volume as a whole to exhibit excellent mass transfer efficiency 
(Feng et al., 2013). However, the preparation of the hollow fibre membrane is more complex than that of flat 
sheet because it involves more controlling parameter during the membrane spinning. The preparation of the 
hollow fibres often requires internal and external coagulants for the polymer to solidify and involves more 

 

   

DOI: 10.3303/CET1756266

 
 
 
 
 
 

 
 
 
 
 
 
 
 
 
 

 

 
 
 
 
 
 
 
 
 
 
 
 

 

 

 

 

 
 

Please cite this article as: Said N., Hasbullah H., Ismail A.R., Abidin M.N.Z., Goh P.S., Othman M.H.D., Kadir S.H.S.A., Kamal F., Abdullah 
M.S., Ng B.C., 2017, The effect of air gap on the morphological properties of psf/pvp90 membrane for hemodialysis application, Chemical 
Engineering Transactions, 56, 1591-1596  DOI:10.3303/CET1756266   

1591



controlling parameters than those of flat sheet membranes such as the dimension of spinneret, dope viscosity, 
type of internal and external coagulants, flow rate of the bore fluid, type of bore fluid, dope extrusion rate, 
length and humidity of the air gap, wind-up speed, and fibre take-up speed (Korminouri et al., 2014). 
Air gap has been reported before to play an important role in the development of membrane morphology with 
improved separation performance (Khayet, 2003). The fabrication of hollow fibre with a desirable morphology 
for a specific application is a tedious process and the effect of air gap on the hollow fibre membrane 
morphology and performance reported in the literature often provides contradict results especially in terms of 
the changes in morphology and separation performance. In fact, no study has ever discussed specifically the 
impact of applying different air gaps towards the membrane in the context of hemodialysis. 
To resolve the confusion, this work attempts to do a quick study on the effects of air gap on the development 
of a desired PSf hemodialysis hollow fibre membrane. The main objective of this study is to understand the 
influence of air gap on the morphology of PSf/PVP90 hollow fibre membranes and figure out the membrane 
morphology that is needed to improve the membrane separation performance. 

2. Experimental 

2.1 Materials 

All chemicals used were reagent grade. PSf (Udel-P1700) was purchased from Solvay Advanced Polymers; 
PVP-K90 (molecular weight = 360,000 g/mol) and bovine serum albumin (BSA) of > 98 % purity were obtained 
from Sigma Aldrich; NMP (98 %; molecular weight = 99.1 g/mol) was purchased from Merck. 

2.2 Fabrication of PSf/PVP90 hollow fibre membranes 

PSf polymer was dried in oven at 50 °C for 1 d to remove moisture content prior to dope solution preparation. 
Dope solution of 18 wt% PSf and 4.8 wt% PVP in NMP was prepared. The mixture was intensively stirred for 
one day to obtain a homogeneous solution. Before spinning, the dope was degassed for 6 h. The hollow fibre 
membranes were spun at room temperature by dry-wet spinning technique under different air gaps. The dope 
solution passed through a spinneret with an orifice diameter/inner diameter (OD/ID) of 0.6 / 0.3 mm, pressured 
using nitrogen gas, N2 with the help of a gear pump to regulate dope extrusion rate (DER). The bore fluid, 
which is distilled water, was channelled into the spinneret using a constant-flow syringe pump. Coagulation 
bath used for phase inversion was tap water. The spinning parameters used were listed in Table 1. The 
fabricated membranes were treated in tap water for three days to remove the remaining solvent. The 
membranes were kept in 10 wt% glycerol to prevent the membrane structure from collapsing. After drying, the 
fibres were stored in sealed-plastic bags. 

Table 1: Spinning parameters for hollow fibre spinning process 

Air gap (cm) 3, 30, 50 
Dope extrusion rate (mL/min) 1 
Collection speed (m/min) 10 
Bore fluid flow rate (mL/min) 1 

2.3 Membrane characterisations 

2.3.1 Scanning electron microscopy (SEM) 

The structural morphology of PSf/PVP90 polymeric membrane was examined using SEM. The polymeric 
membrane was snapped in liquid nitrogen to obtain clear and smooth cross-section, before sticking onto metal 
plate using double-sided carbon tape at lateral sides. For surface studies, a small piece of each membrane 
was mounted on the metal plate horizontally. The membrane sample was sputter-coated with 
platinum/palladium mixture before being analysed. The images of outer surface and cross-section of the 
membrane were collected. 

2.3.2. Porosity and pore size measurement 

Dry-wet weight method was used to measure membrane porosity, ε. The membrane was immersed in pure 
water for 1 h. Next, the membrane weight was measured in both wet and dry condition. The porosity of the 
membrane was defined in the following Eq(1): 

ε =
w1 − w2

Vρw
 (1) 

where ε is the porosity of the membrane (%), w1 is the mass of the wet membrane, w2 is the mass of the dry 
membrane, V is the volume of the membrane and ρw is the density of water (1.0 g/cm3). The average pore 

1592



radius, r (m) was determined by the filtration velocity method (Abidin et al., 2016), using Guerout-Elford-Ferry 
Eq(2): 

r =  √
(2.9 −  1.75 ℇ)  ×  8ƞƖQ

ε ×  A ×  ΔP
 (2) 

where ƞ is the water viscosity at 25 °C, Ɩ is the membrane thickness (m), Q is the volume of the permeate 
water per unit time (m3/s), A is the effective area of membrane (m2) and ΔP is the operational pressure (Pa). 
Pore size (diameter) of membrane was determined by multiplying r by 2. 

2.4 Evaluation of membrane separation performance 

The PWF, J and BSA rejection, R of the membranes were measured to evaluate the membranes separation 
performance. Membrane modules consist of 10 fibres (10 cm in length) were potted using epoxy. The PWF 
was obtained at the trans-membrane pressure of 0.74 bar using the following Eq(3): 

J =  
V

A ×  ∆t
 (3) 

where J represents the PWF (L m-2 h-1), V is the volume of permeate (L), A is the effective surface area (m2) 
and Δt is the permeation time (h). For protein rejection test, 500 ppm BSA solution was used. The percent 
rejection was calculated by Eq(4). 

R (%)  =  (1 −  
Cp

Cf
)  ×  100 % (4) 

where Cp and Cf are the BSA concentrations of the permeate and feed. The concentration of BSA was 
measured using UV-Vis Spectrophotometer (DR 5,000). 

3. Results and discussion  

3.1 Morphologies of PSf/PVP90 membranes 

Based on Table 2, both inner and outer diameters of the membrane decreased with the increasing air gap. 
This was due to the elongation stress created on the fibre by gravitational pull. The stress may affect the 
molecular orientation during membrane formation which leads to significant changes in membrane 
characteristics. The subsequent effect of elongation stress on the membrane properties is further discussed in 
Section 3.2. 
Figure 1 represents the SEM micrographs taken from the cross section and the outer surface of the 
PSf/PVP90 hollow fibre membranes fabricated at the different air gaps. Membranes spun at 30 cm and 50 cm 
air gap have led to the formation of one long array of finger-like structure, which originated from the inner side 
of membrane. The finger-like void for membrane spun at 50 cm air gap was longer and bigger compared to 
the membrane fabricated at 30 cm air gap. This was due to the greater elongation force exerted at higher air 
gap which reduced the thickness of polymer solution, allowing faster solvent-non solvent exchange at the 
inner side of fibre (Khayet, 2003). The formation of porous sponge-like structure at the outermost layer of the 
membrane was due to the exchange of solvent and air moistures during residence time. The dense selective 
skin layer at the innermost region of the membrane was formed by instantaneous de-mixing between solvent 
and non-solvent (Abdelrasoul et al., 2015). The results are in accordance with the desired properties of 
hemodialysis membrane, in which the inner surface possessed the smallest pores to prevent clogging of 
solutes inside the pores (Yamashita and Sakurai, 2015). The nano-sized pores at the selective layer are able 
to minimise the transport of human albumin while permitting other smaller toxin molecules to pass through. 

Table 2: Dimensional changes of hollow fibre membranes as a function of air gap 

Air gap 
(cm) 

Residence 
time (s) 

OD (cm) ID (cm) OD/ID Ratio Cross-Section Area 
Change Ratioa 

3 0.38 0.045 0.036 1.25 3.70 
30 3.82 0.035 0.023 1.52 3.80 
50 6.36 0.031 0.020 1.55 4.81 
aRatio of the spinneret cross-section area to a fibre cross-section area. 
 
The membrane fabricated at 3 cm air gap produced a sandwich-like structure which consists of two selective 
skin layers and two arrays of finger-like voids. At very low air gap, the exchange of solvent and moistures at 

1593



the membrane outer surface cannot take place due to very short residence time. As the flow of bore fluid 
solidifies the membrane inner surface instantaneously, the same thing has happened at the outer surface of 
the nascent hollow fibre membrane as the membrane entered straight away into the coagulation bath. The 
identical dense skin layer to the inner surface was formed at the outer layer of the membrane. This kind of 
morphology is not recommended for hemodialysis application since it could promote the entrapment of protein 
molecules inside the membrane matrix. 
 

 

Figure 1: The cross-sectional SEM images of membrane fabricated at air gap of 3 cm, 30 cm, and 50 cm 

From Figure 2, the pore size at the membrane outer surface became larger as the air gap is increased. The 
increasing air gap enhanced the formation of larger pores at outer surface. Moisture induced phase inversion 
which occurred at high air gap created pores at the outer side of membrane before the membrane enters the 
coagulation bath. The longer the membrane takes to reach the coagulation bath, the larger the pore size. 
Polymer nucleation was allowed to happen longer at high air gap. In hemodialysis application, the larger 
membrane pore size at the outer surface promotes the removal of bigger uremic toxins through convection.  

 

Figure 2: The SEM images of PSf/PVP90 membrane outer surface fabricated at air gap of 3 cm, 30 cm, and 

50 cm 

3.2 Average pore size and porosity of PSf/PVP90 membranes 

Table 3 shows the results of average pore size and porosity of the membrane fabricated at different air gaps. 
The average pore size of PSf/PVP90 hollow fiber membrane was calculated based on water permeation data. 
The membrane experienced a substantial increment of average pore size with the increase of air gap. The 
porosity also increased with increasing air gap due to the changes in membrane morphology as shown in 
Figure 1 and 2. Higher air gap means that the membrane has been exposed longer to gravitational pull, 
causing the huge effect towards membrane structure and eventually produces a bigger pore size compared to 
membrane fabricated at lower air gap. Macromolecules inside the solution experienced swelling and relaxation 
once the solution is extruded from spinneret, changing the orientation of macromolecules inside the 

1594



membrane. This later produced a larger membrane average pore size which could permit water to pass 
through easier and faster.  
In terms of porosity, membrane spun at 50 cm and 30 cm air gap were by far more porous as shown in Table 
3. The sufficiently strong elongation stress during spinning has pulled the polymer molecular chains apart at 
the beginning of phase separation and created porous structure. This resulted in the increase of the 
membrane free volume (Cao et al., 2004). The formation of longer and bigger finger-like structures as the 
result of faster phase inversion also helped to accommodate pure water. On the other hand, the weak stress 
exerted on the membrane spun at 3 cm air gap had induced the unfavourable change in molecular orientation 
which reduced membrane porosity or free volume (Khayet, 2003). 

Table 3: Results of porosity and average pore size 

Air gap (cm) Porosity (%) Average pore size (µm) 
3 17.13 0.018 
30 44.13 0.039 
50 46.67 0.044 

3.3 The effect of air gap on membrane separation performance 

A more porous membrane structure is able to facilitate water movement across the membrane. As stated by 
previous study (Santoro and Guadagni, 2010), a high flux membrane can achieve a better removal of middle 
molecular uremic toxins by the mean of convection. 
The results of PWF and BSA rejection with respect to various air gaps are shown in Figure 3. Significant 
improvement in PWF was reported as the air gap is increased. The longer and larger finger-like voids helped 
in facilitating pure water more efficient.  The PWF of membrane spun at 3 cm air gap was very low due to its 
dual-skin layer property and small average pore size. 
As for the protein rejection, all the membranes obtained more than 90 % rejection of BSA which was quite 
compulsory for all hemodialysis membranes to prevent albumin loss. Too much albumin loss could render 
albumin loss syndrome to the dialysis patient (Irfan et al., 2014). The main reason would be the compact 
arrangement of nano-sized pores at the membrane skin layer, which restrained the movement of BSA 
molecules across the membrane. The results (Figure 3) proved that BSA rejection can be influenced by 
different membrane morphologies. The trend shows the enhanced BSA rejection as the PWF increases. This 
happened due to the improved route for water molecules to pass through, leaving the large and hydrophobic 
BSA molecules behind. As the pure water moving faster, the hydrophobic nature of BSA with the help of the 
membrane selective layer had weakened the interaction between the protein molecules and water molecules. 

 

Figure 3: PWF and BSA rejection of the membranes 

4. Conclusion 

In this work, air gap has been manipulated during the fabrication of PSf/PVP90 hollow fibre membranes. The 
membrane spun at low air gap (3 cm) showed a sandwich-like structure while membranes spun at higher air 

1595



gap (30 cm and 50 cm) possessed an asymmetrical structure with skin layer, compact spongy sublayer and 
finger-like structure with macro-voids. The porosity and pore size of the membrane increased with the 
increase in air gap. The same trend could be seen on the PWF of the membrane while maintaining excellent 
protein resistance. All the results indicated that air gap is one of the spinning parameters that should be 
focused on during hollow fibre membrane fabrication process to acquire the desired membrane morphology 
for hemodialysis application. 

Acknowledgements 

We gratefully acknowledge the financial support from the Universiti Teknologi Malaysia and Ministry of Higher 
Education Malaysia (Grant no: 01G46 and 4F652).  

References 

Abdelrasoul A., Doan H., Lohi A., Cheng C.H., 2015, Morphology control of polysulfone membranes in 
filtration processes: a critical review, Chemical and Biomolecular Engineering Review 2, 22–43. 

Abidin M.N.Z., Goh P.S., Ismail A.F., Othman M.H.D., Hasbullah H., Said N., Kadir S.H.S.A., Kamal F., 
Abdullah M.S., Ng B.C., 2016, Antifouling polyethersulfone haemodialysis membranes incorporated with 
poly (citric acid) polymerized multi-walled carbon nanotubes, Materials Science and Engineering: C 68, 
540–550. 

Bowry S.K., Gatti E., Vienken J., 2011, Contribution of polysulfone membranes to the success of convective 
dialysis therapies, Eds. Saito A., Kawanishi H., Yamashita A.C., Mineshima M., High performance 
Membrane Dialyzers, 1st Ed., Karger Publishers, Basel, Switzerland, 110–118. 

Cao C., Chung T.S., Chen S.B., Dong Z.J, 2004, The study of elongation and shear rates in spinning process 
and its effect on gas separation performance of Poly(ether sulfone) (PES) hollow fiber membranes, 
Chemical Engineering Science 59 (5), 1053-1062.  

Feng C.Y., Khulbe K.C., Matsuura T., Ismail A.F., 2013, Recent progresses in polymeric hollow fiber 
membrane preparation, characterization and applications, Separation and Purification Technology 111, 
43–71. 

Irfan M., Idris A., Yusof N.M., Khairuddin N.F.M., Akhmal H., 2014, Surface modification and performance 
enhancement of nano-hybrid f-MWCNT/PVP90/PES haemodialysis membranes, Journal of Membrane 
Science 467, 73–84. 

Khayet M., 2003, The effects of air gap length on the internal and external morphology of hollow fiber 
membranes, Chemical Engineering Science 58, 3091–3104. 

Khulbe K.C., Feng C., Matsuura T., 2009, The art of surface modification of synthetic polymeric membranes, 
Journal of Applied Polymer Science 115, 855–895. 

Korminouri F., Rahbari-Sisakht M., Rana D., Matsuura T., Ismail A.F., 2014, Study on the effect of air gap 
length on properties and performance of surface modified PVDF hollow fiber membrane contactor for 
carbon dioxide absorption, Separation and Purification Technology 132, 601–609. 

Santoro A., Guadagni G., 2010, Dialysis membrane: From convection to adsorption, NDT Plus 3, 36–39. 
Yamashita A.C., Sakurai K., 2015, Dialysis Membranes-physicochemical structures and features, Ed. Suzuki 

H., Updates in Hemodialysis, 1st Ed., Intech, Rijeka, Croatia, 159–187. 

1596