Energy and Environment Research; Vol. 5, No.

1; 2015
ISSN 1927-0569 E-ISSN 1927-0577
Published by Canadian Center of Science and Education

Preparation and Characterization of Inorganic Membranes for

Hydrocarbon Separation from N2 for Environmental Applications
Mohammed N. Kajama1, Habiba Shehu1, Edidiong Okon1 & Edward Gobina1
1
Centre for Process Integration and Membrane Technology, (CPIMT), School of Engineering, The Robert
Gordon University, United Kingdom
Correspondence: Edward Gobina, Centre for Process Integration and Membrane Technology, (CPIMT), School
of Engineering, The Robert Gordon University, AB10 7GJ, Aberdeen, United Kingdom. Tel: 44-012-2426-2348.
E-mail: e.gobina@rgu.ac.uk

Received: March 25, 2015 Accepted: April 24, 2015 Online Published: May 20, 2015
doi:10.5539/eer.v5n1p110 URL: http://dx.doi.org/10.5539/eer.v5n1p110

Abstract
This paper reports the preparation and characterization of a tubular alumina and silica membrane for propylene
(C3H6) and nitrogen (N2) separation. Dip-coating technique was employed to prepare silica membrane. Nitrogen
adsorption measurements reveal that this membrane possesses mesostructure with pore diameter of 3.940 nm and
BET surface of 0.484 m2/g. Scanning electron microscopy (SEM) studies show that the layers are defect free and
energy diffraction X-ray analysis (EDXA) confirms the compositional element of silica. Propylene and nitrogen
permeation tests also confirmed that the membrane is defect free with C3H6/N2 selectivity of 1.79 higher than
ideal Knudsen selectivity.
Keywords: mesoporous membranes, silica membranes, nitrogen adsorption, pore size analysis, hydrocarbon
separation
1. Introduction
Hydrocarbons are among the most common air pollutants expelled from many industrial processes e.g. chemical
(Khan & Ghoshal, 2000), metal cleaning (Huang & Lin, 1997) among others. The venting of these hydrocarbons
to the atmosphere is an enormous contribution to poor air quality. From an environmental point of view, it is
imperative to separate these hydrocarbons from waste gas stream. Different countries in the world have
regulations in place that tightly regulate these emissions. The United States Environmental Protection Agency
proposed the air quality standards of a maximum 3 hour concentration of hydrocarbon content not to exceed 1.6
x 10-4 kgm-3 (0.24 ppm) (Ghoshal & Manjare, 2002). The European community has set a limit not to exceed 35 g
of total organic compounds (TOC) per cubic meter gasoline loaded (Ghoshal & Manjare, 2002). In addition, the
Gothenburg Protocol that was adopted by the EU in 1999 to abate acidification, eutrophication and ground level
ozone sets emission levels for sulphur, nitrous oxide, VOCs and ammonia hence, when fully implemented it is
expected that European VOC emission will be cut down by 40% when compared to 1990 levels (Derwent, et al.
2010). In many countries emissions of VOCs and non methane volatile organic compounds (NMVOCs) are not
subject to control, hence loss of hydrocarbons into the atmosphere can cause severe environmental pollution and
great monetary loss. A number of emerging techniques are being developed as a substitute to more established
processes such as absorption, adsorption and condensation and membrane processes (Khan & Ghoshal, 2000;
Huang & Lin, 1997).
Membrane processes has recently emerged as the state-of-the-art technology for such application. This technique
was earlier applied in the last five decades for desalination purposes (Khan & Ghoshal, 2000), volatile organic
compounds (VOC) removal (Khan & Ghoshal, 2000; Huang & Lin, 1997; Tahir & Koh, 1999; Saracco &
Specchia, 2000; Pina, et al., 1996; Pina, et al., 1997; Benard, et al., 2009; Benard, et al., 2010) and natural gas
processing (Kajama, et al., 2014). Polymeric membranes such as hollow fibre composite membrane of silicone
rubber have been used. Polymeric membranes have very low permeabilities for solvents but their selectivities are
quite high (Huang & Lin, 1997). Indeed, polymeric membranes cannot resist harsh chemical environments and
cannot withstand high temperatures. Therefore, ceramic membranes can substitute polymeric membranes since
these are chemically resistant and thermally stable. The driving force (partial or total pressure) is the pressure
difference across the membrane (Vercauteren et al., 1998). Gas transport mechanisms through porous membrane

110
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

include: surface diffusion, Knudsen diffusion, capillary condensation, Poiseuille or viscous flow, and molecular
sieving mechanisms (Coronas & Santamarıa, 1999; Lee & Oyama, 2002; Sidhu & Cussler, 2001). In Knudsen
diffusion mechanism, gas molecules diffuse through the pores of the membrane and then get transported by
colliding more frequently with the pore walls (Lee & Oyama, 2002; Phattaranawik, et al, 2003), Knudsen
mechanism also occurs if the mean free path (λ) of the permeating gas molecule is greater than pore diameter (d).
This indicates that the Knudsen number (K) is greater than 1, i.e. K = λ/d ˃ 1 (Sidhu & Cussler; 2001). Viscous
flow mechanism takes place if the pore radius of the membrane is greater than the mean free path of the
permeating gas molecule (Lee, et al. 2005; Benito, et al. 2005). Gas separation by molecular sieving mechanism
takes place when the pore dimensions of the inorganic ceramic membrane approach those of the permeating gas
molecules (Li, 2007; and Lee, et al. 2005). In capillary condensation mechanism, separation can takes place in
the pores of the membrane with mesoporous layer in the presence of condensable gas specie. Surface diffusion
mechanism occurs when the adsorption of the permeating gas molecule occurs on the pore surface of the
membrane material there by increasing the gas transport performance (Lee, et al. 2005; Javaid, et al. 2005).
Uhlhorn, et al. (1990; 1992) used gamma alumina ( γ -Al2O3) membranes with a pore diameter of 2.5 nm and
magnesium oxide (MgO) modified membranes with a lower pore size in order to separate propene from nitrogen
(C3H6/N2) mixtures. The membranes recorded both permeabilities and separation factors higher than those for
Knudsen diffusion.
The characterization of porous ceramic membranes is typically elaborated by the flux and the selectivity which
are influenced by the pore size distribution, intrinsic membrane surface properties and porosity (Smart, et al.
2013). According to the International Union of Pure and Applied chemistry (IUPAC) definition, membranes
pores structure are classified as mesoporous (2-50nm), microporous (< 2nm) and macroporous (> 50nm) layers.
Porous metallic membranes with the pore size greater than 0.3nm are normally used as sieves for larger
molecules and particles (Li, 2007; Guizard & Princivalle, 2009; Ismail & David, 2011).
Scanning electron microscopy (SEM) is a technique used to analyse the morphological surface of the membrane
and their cross-sections. It can also be employed for high magnification of the membrane separation layer.
Energy diffraction x-ray analysis (EDXA) is used to analyze the elemental compositions of the membrane. N2
adsorption-desorption isotherm (77 K) is a standard technique employed to characterize the pore structure and
network of a material. It is perhaps simple and the most widely used method to determine the specific surface
area, pore size and pore size distribution for meso and micro porous adsorbents (Smart, et al. 2013; Choma, et al.
2003). The gas adsorption isotherm is divided into six categories (Smart, et al. 2013; Choma, et al. 2003;
Weidenthaler, 2011). Five of the six categories are mentioned in the IUPAC recommendations (Choma, et al.
2003).
Type I isotherm is governed by adsorption in micropores sample (with < 2nm pore size) at low relative pressure,
while Type II isotherm is regarded as a non-porous or macroporous (with > 50nm pore size) adsorbents
formation of multilayers of adsobate on surfaces of adsorbent. Type III isotherm is also a non-porous or
macroporous adsorbents with weak adsorbent-adsorbate interactions. Type IV isotherm is a typical mesoporous
adsorbents with initial monolayer-multilayer coverage on external and mesopore surface and is followed by
capillary condensation in mesopores with different types of hysteresis loops which are observed depending on
the shape of pores. Type V isotherm is also a mesoporous adsorbents with weak adsorbent-adsorbate interactions.
An example of this uncommon isotherm is observed for water adsorption on activated carbon (Weidenthaler,
2011). Type VI isotherm is a highly uniform surface which shows a layer-by-layer adsorption (Smart et al. 2013;
Choma et al. 2003; Weidenthaler, 2011).
In this study, the permeation of propylene and nitrogen through silica membrane were carried out at various
pressures as the driving force and ambient temperature. N2 sorption at 77 K, SEM and EDXA were employed for
membrane characterization.
2. Experimental
2.1 Membrane Support
Commercially available porous alumina supports of tubular configuration supplied by CTI (France) consisted of
77% alumina + 23% TiO2 were employed for this experiment. The tubular alumina support used for this
experiment is depicted in Figure 1. The alumina support had an internal and outer diameter of 7 and 10 mm
respectively with a permeable length of 348 mm and a porosity of 45%.

111
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

Figure 1. Commercial tubular ceramic support

Pressures of 0.05 to 0.45 barg and temperatures of 22 °C (295 K) were applied. Crushed samples of the
unmodified and modified membranes (Figure 2 a and b) were used for characterization (nitrogen adsorption and
scanning electron microscopy respectively).

(a) (b)

Figure 2 (a). Crushed sample of the support (unmodified), (b). Crushed sample of silica membrane (modified)

2.2 Membrane Preparation

The solution use for membrane preparation was prepared by mixing 50 ml of silicone elastomer (Sylgard®) and
nine parts of isopentane contained in a glass tube to obtain a clear and colourless solution. A curing agent
(Sylgard®) equivalent to one-tenth of the elastomer was added and the resulting solution was mixed at room
temperature. The solution was then allowed to age for 30 minutes after which the ceramic support was immersed
for 30 minutes. The membrane was then oven dried at 65 0C for 24 hours (Gobina, 2006; Nwogu, et al. 2013) to
form an ultra-thin layer on the support. The same procedure was repeated for subsequent coatings.

112
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

2.3 Membrane Characterization

The Brunauer-Emmett-Teller (BET) method for surface area, Barrett-Joyner-Helanda (BJH) method for pore size
distribution, and the adsorption-desorption isotherm of unmodified support and silica modified membrane were
obtained from N2 sorption isotherm (Quantachrome instrument version 3.0). Prior to N2 sorption analysis,
weighted crushed samples were degassed for 180 min at 300 0C at a rate of 10 0C/min. The inner, outer and
cross-sectional surface morphologies of the support and membrane were analyzed by SEM (Zeiss EVO LS10).
The support was found to be defect free. EDXA was also employed to analyze the elemental compositions of the
material before and after modification.
2.4 Gas Permeation Measurements
Permeation behaviour of pure gases nitrogen (N2) and propylene (C3H6) were measured at different feed
pressures ranging from 0.05 to 0.45 bar. Permeate readings were recorded as l/min at ambient temperature.
3. Results and Discussion
3.1 Structure of Unmodified and Modified Membranes
Figure 3(a and b) depicts the SEM images of the inner and outer surface of the unmodified membrane. The
images exhibit a defect free support. Figure 4(a and b) depicts the inner and outer surface of the silica modified
membranes. It can be seen in Figure 4(b) that ultra-thin layer was formed on the outer surface of the membrane
and the silica layer is clearly visible from the micrograph. From this image, one can see a good bonding to the
support without any pinholes on the layer.

Figure 3(a). SEM image of the support (unmodified) inner surface

113
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

Figure 3(b). SEM image of the support (unmodified) outer surface

Figure 4(a). SEM image of silica membrane (modified) inner surface

Figure 4(b). SEM image of silica membrane (modified) outer surface

Figures 5(a and b) & Figure 6(a and b) depicts the EDX analysis for unmodified and modified membrane. These
Figures show the elemental compositions of the membrane before and after modification. It can be obviously
seen in Figure 5(a and b) that there was no sign of silica particle before modification. However, silica particles

114
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

were observed after modification as indicated on the EDXA result in Figure 6(b).

Figure 5(a). EDXA of the support (unmodified) inner surface

Figure 5(b). EDXA of the support (unmodified) outer surface

Figure 6(a). EDXA of silica membrane (modified) inner surface

115
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

Figure 6(b). EDXA of silica membrane (modified) outer surface

Figure 7(a and b) & Figure 8(a and b) depicts the N2 adsorption-desorption isotherms of unmodified and silica
modified membranes. The adsorption-desorption isotherms exhibits a characteristics of mesopores solids
(especially ceramics) resulting in type IV and V physisorption isotherm undergoing capillary condensation and
hysteresis during desorption (Smart, et al. 2013). Different parameters were obtained for the
Brunauer-Emmett-Teller (BET) method surface area and Barrett-Joyner-Helanda (BJH) method pore size
distribution. Table 1 quantitatively shows the physical properties obtained with nitrogen adsorption for the
unmodified alumina support and modified silica membrane. The BET surface area and BJH pore diameter were
0.364 m2/g and 4.171nm for the unmodified support and 0.439 m2/g and 4.179nm for the modified membrane
respectively. BET surface area increased to about 0.484 m2/g after subsequent dipping. BJH pore diameter also
decreased from 4.171 to 3.940nm which indicates that some amount of silica has penetrated into the support
pores which shows a typical mesopore diameter in the range 2 to 50 nm. This is consistent with the above SEM
and EDXA results.

Figure 7(a). N2 adsorption-desorption isotherm for the support (unmodified)

116
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

Figure 7(b). Pore size distribution for the support (unmodified)

Figure 8(a). N2 adsorption-desorption isotherm for the silica membrane (modified)

117
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

Figure 8(b). Pore size distribution for the silica membrane (modified)

Table 1. Physical properties for unmodified and modified membranes

Membrane BET surface area (m2/g) BJH pore diameter (nm)
Unmodified Ceramic Support 0.364 4.171
Silica Modified Membrane (5th Dip) 0.439 4.179
Silica Modified Membrane (9th Dip) 0.484 3.940

3.2 Nitrogen and Propylene Permeation

Figure 9 depicts gas flow rate through silica membrane as a function of the feed pressure at room temperature.
Nitrogen and propylene gas flow rates increases with the feed pressure. It can be seen that propylene has the
higher flow rate compared to nitrogen despite its higher molecular weight (N2: 28 g/mol; C3H6: 42 g/mol). This
result is consistent with the contribution of surface flow mechanism. Figure 10 shows the selectivity of C3H6
over N2. Highest selectivity of 1.79 at 0.05 barg was obtained. This selectivity increased by a factor of 2
compared to the ideal Knudsen selectivity (0.82) which corroborates with the literature Uhlhorn, et al. (1990;
1992). From the obtained gas permeation and selectivity, one can conclude that hydrocarbon could be separated
from nitrogen.

Figure 9. Gas flow rates of nitrogen and propylene as a function of feed pressure at room temperature

118
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

Figure 10. C3H6/N2 selectivity as a function of feed pressure at room temperature

4. Conclusion
Defect-free mesoporous silica membrane has been developed on the alumina substrate by dip-coating technique.
The BET surface area and BJH pore diameter of the membrane are 0.484 m2/g and 3.940 nm respectively. The
prepared membrane was defect free as confirmed by the propylene and nitrogen permeation data. The high
propylene selectivity obtained at low pressure drop is indicative of selective surface flow mechanism. Work is
currently being carried out to compare mixed gas selectivity to pure gas selectivity.
Acknowledgments
The authors gratefully acknowledge Petroleum Technology Development Fund (PTDF) Nigeria for funding this
research and School of Pharmacy & Life Sciences RGU Aberdeen for the SEM and EDXA results.
References
Benard, S., Giroir-Fendler, A., Vernoux, P., Guilhaume, N., & Fiaty, K. (2010). Comparing monolithic and
membrane reactors in catalytic oxidation of propene and toluene in excess of oxygen. Catalysis today, 156,
301-305. http://dx.doi.org/10.1016/j.cattod.2010.07.019
Benard, S., Ousmane, M., Retailleau, L., Boreave, A., Vernoux, P., & Giroir-Fendler, A. (2009). Catalytic
removal of propene and toluene in air over noble metal catalyst1. Can. J. Civ. Eng., 36, 1935-1945.
http://dx.doi.org/10.1139/L09-135
Choma, J., Kloske, M., & Jaroniec, M. (2003). An improved methodology for adsorption characterization of
unmodified and modified silica gels. Journal of Colloid and Interface Science, 266, 168-174.
http://dx.doi.org/10.1016/S0021-9797(03)00573-3
Gobina, E. Apparatus and Methods for Separating Gases. (2006). United States Granted Patent No. US 7048778,
May 23, 2006.
Huang, P., Xu, N., Shi, J., & Lin, Y. S. (1997). Recovery of volatile organic compounds from air by ceramic
membranes. Ind. Eng. Chem. Res., 36, 3815-3820. http://dx.doi.org/10.1021/ie960760b
Kajama, M. N., Nwogu, N. C., & Gobina, E. (2014). Hydrogen Separation Using Silica-Based Composite
Membranes. Advanced Materials Research, Trans Tech Publications, 1051, 107-111.
http://dx.doi.org/10.4028/www.scientific.net/AMR.
Khan, F. I., & Ghoshal, A. K. (2000). Removal of volatile organic compounds from polluted air. Journal of Loss
Prevention in the Process Industries, 13, 527-545. http://dx.doi.org/10.1016/S0950-4230(00)00007-3
Nwogu, N. C., Gobina, E., & Kajama, M. N. (2013). Improved carbon dioxide capture using nanostructured
ceramic membranes. Low Carbon Economy, 4, 125-128. http://dx.doi.org/10.4236/lce.2013.43013
Pina, M. P., Irusta, S., Menendez, M., & Santamaria, J. (1997). Combustion of Volatile Organic Compounds over
Platinum-Based Catalytic Membranes. Industrial Engineering Chemistry Resources, 36, 4557-4566.
http://dx.doi.org/10.1021/ie9700876
Pina, M. P., Menendez, M., & Santamaria, J. (1996). The Knudsen-diffusion catalytic membrane reactor: An

119
www.ccsenet.org/eer Energy and Environment Research Vol. 5, No. 1; 2015

efficient contactor for the combustion of volatile organic compounds. Applied Catalysis B: Environmental,
11, L19-L27. http://dx.doi.org/10.1016/S0926-3373(96)00081-1
Saracco, G., & Specchia, V. (2000). Catalytic Filters for the Abatement of Volatile Organic Compounds.
Chemical Engineering Science, 55(5), 897-908. http://dx.doi.org/10.1016/S0009-2509(99)00361-9
Smart, S., Liu, S., Serra, J. M., Diniz da Costa, J. C., Iulianelli, A., & Basile A. (2013). Porous ceramic
membranes for membrane reactors. http://dx.doi.org/10.1533/9780857097330.2.298
Tahir, S. F., & Koh, C. A. (1999). Catalytic destruction of volatile organic compound emissions by platinum
based catalyst. Chemosphere, 38(9), 2109-2116. http://dx.doi.org/10.1016/S0045-6535(98)00420-2
Uhlhorn, R. J. R., Keizer, K., & Burggraaf, A. J. (1990). Gas transport and separation properties of ceramic
membranes with pores of molecular dimensions. Paper presented at the International Congress on
membranes and membrane processes (ICOM 90), Chicago, IL, August 1990.
Uhlhorn, R. J. R., Keizer, K., & Burggraaf, A. J. (1992). Gas transport and separation with ceramic membranes
Part 1, Multilayer diffusion and capillary condensation. J. Membrane Science, 66, 259-269.
Weidenthaler, C. (2011). Pitfalls in the characterization of nanoporous and nanosized materials. Nanoscale, 3,
792-810. http://dx.doi.org/10.1039/C0NR00561D

Copyrights
Copyright for this article is retained by the author(s), with first publication rights granted to the journal.
This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution
license (http://creativecommons.org/licenses/by/3.0/).

120