Gas permeance properties of asymmetric carbon hollow fiber membranes at high feed pressures

Highlights

• High-pressure permeances of CH₄, N₂, CO₂ & C₂H₆ through CHFMs were measured.

• For CO₂ & C₂H₆ the increase of the feed pressure provides maxima in the permeance.

• Maxima can be explained by orientational correlations of confined molecules.

• Optimum operation condition: the pressure where a gas presents highest permeances.

Abstract

Asymmetric carbon hollow fiber membranes (CHFs) were prepared and their gas permeance properties were investigated at high pressures up to ∼60 bar. The main target of this work was to study the effect of the gas feed pressure on the permeance through carbon membranes. Specifically, nitrogen, carbon dioxide, methane and ethane were chosen for studying the permeance performance of the developed membranes at conditions of industrial interest. It has been observed that CO₂ obtains a maximum permeance of 412 GPU at 27.5 bar and 765 GPU at 22 bar for M1 and M3 membranes respectively while C₂H₆ presents maximum permeance values of 482 and 825 GPU at 22.45 and 16.13 bar for M1 and M3 membranes respectively. These maxima in permeances can be explained in terms of orientational correlations of the confined molecules, originating from inter–molecular quadrupolar interactions. Discussion on the diffusion mechanism of all the studied gases is also provided. Note that natural gas (NG) treatment, i.e. sweetening and purification processes, is energetically and economically convenient if it takes place under the conditions where the NG stream is extracted from the wells (i.e. pressure 30–60 bar). This practically means that if it is known the pressure where the used membranes provide the highest permeability factors, concerning the gas of interest, then the pressure can be fixed at this grade without pointless depressing actions. The presented testing methodology is very useful for identifying the optimum operation conditions towards an efficient gas separation process.

Introduction

The permeation of gases and liquids through microporous membranes has been studied for many decades. A landmark year was 1962 when Loeb and Sourirajan published a paper where high-permeable asymmetric cellulose acetate membranes were developed and studied for sea water demineralization based on the osmotic phenomenon (Loeb and Sourirajan, 1962). The first commercial scale separation process had been presented in the 1960s, achieving the desalination of brackish water by reverse osmosis via utilization of cellulose acetate membranes that selectively retained the salt and allowed pure water to permeate (Sourirajan, 1970). Since the late 60's many types of polymeric and inorganic membranes have been reported as candidate materials for liquid and gas separation applications (Baker, 2004). Gas separation technology is an “outstanding” worldwide challenge for both industry and academic society which can also offer a significant reduction of the environmental impact and the cost in several industrial processes. However, several difficulties are usually encountered, especially in cases where molecules with similar sizes (kinetic diameters) should be separated. The study of gas separation has a long history from the age of Thomas Graham who gave the first description of the solution-diffusion model and his work on porous membranes led to Graham's law of diffusion (Graham, 1866).

During the last few years many types of membranes such as metal (Deveau et al., 2013), polymeric (Lee et al., 2011), ceramic and zeolitic (Smart et al., 2010, Varoon et al., 2011) and mixed matrix membranes (Mahajan and Koros, 2000, Mahajan and Koros, 2002, Song et al., 2008, Favvas et al., 2014a, Favvas et al., 2014b) have been studied and used in gas separation applications. Polymeric membranes have been also studied for high-pressure permeability experiments, but with limited applicability (dependent on the membrane materials) due to swelling phenomena. As swelling phenomenon is described the process of dissolution of a polymer in a defined solvent, liquid or any fluid which provides good solvent properties. At first, the solvent molecules, for example the CO₂, slowly diffuse into the polymer to produce a swollen gel. If the polymer–polymer intermolecular forces are high, thanks to crystallinity, crosslinking, or strong hydrogen bonding, this is all what happens. But, if these forces are overcome by the introduction of strong polymer–solvent interactions, a second stage, the relaxation even the dissolution of the polymer, can take place (Wessling et al., 1991, Izák et al., 2007). This phenomenon is also reported as membrane plasticization and is well-known to natural gas industrial membrane systems (Wind et al., 2002, Visser et al., 2005). The gas is typically treated at relatively high pressures of 30–60 bar. Note that at these high pressures of natural gas operations, membrane materials absorb 30–50 cm³(STP) of CO₂/(cm³ polymer) (Baker and Lokhandwala, 2008).

The reported studied materials are many, including polyimide, polyamide–imide, polyethersulfone/polyimide, polysulfone etc in which swelling is usually observed in streams of high carbon dioxide content (Wind et al., 2003a, Wind et al., 2003b, Wind et al., 2004, Kapantaidakis et al., 2003, Kosuri and Koros, 2008, Scholes et al., 2010). On the other hand, inorganic membranes have become a promising material for gas separation technology with many potential applications such as the purification of hydrogen rich streams, the separation of olefin/paraffin mixtures, the recovery of CO₂ from natural gas, the enrichment of syngas as well as the water treatment and the production of oxygen enriched air (Saufi and Ismail, 2004, Ismail and David, 2001, Koros and Mahajan, 2000, Favvas et al., 2007, Favvas et al., 2011). Nanoporous carbon membranes constitute another promising type of inorganic membranes for gas separation applications thanks to their good separation properties (Favvas and Mitropoulos, 2008, Swaidan et al., 2013, Yoshimune and Haraya, 2013, Favvas et al., 2015). These membranes can be produced by carbonization, under inert atmosphere or vacuum. Lately, numerous synthetic precursors have been used to form carbon membranes, such as polyimide and its derivatives, polyacrylonitrile (PAN), phenolic resin, polyfurfuryl alcohol (PFA), polyvinylidene chloride–acrylate terpolymer (PVDC–AC), phenol formaldehyde, polyetherimide (PEI) and polyvinylpyrrolidone (PVP), cellulose, sulfonated poly(phenylene oxide) (SPPO) phenolic resols, numerous co-polymers and others (Song et al., 2008, Pandey and Chauhan, 2001, Salleh and Ismail, 2013, Yoshimune and Haraya, 2013, Li et al., 2014). To this point of view, Salleh at al., 2011, published a review paper entitled “Precursor Selection and Process Conditions in the Preparation of Carbon Membrane for Gas Separation: A Review” where they presented condensed information regarding the sufficient materials and processes for the preparation of carbon gas separation membranes (Salleh at al., 2011). Certainly, Karvan et al. reported an interesting work regarding the carbon molecular sieve hollow fiber membrane manufacturing in a pilot-scale system (Karvan et al., 2013). Note that the one of the major reasons that the carbon hollow fiber gas separation membranes are not commercially available yet is the difficulty to be produced in a pilot-scale dimension. At this direction, from materials to application, another recent work was focused where cellulose acetate (CA) was chosen as carbon gas separation hollow fiber membrane precursor, while HYSYS simulation was also conducted to evaluate the process feasibility of CO₂ capture by HFCMs in a post combustion process (He and Hägg, 2013).

In the present work, two types of CHFs were examined for their permeance properties at high pressures. This work is a continuation of our recent studies where the permeation properties of carbon dioxide at pressures up to 55 bar had been investigated for the same type of membranes (Favvas, 2014). Particularly gas permeances of high purity N₂, CO₂, CH₄ and C₂H₆ gases were measured at pressures up to ∼60 bar at 307 K. A maximum of the CO₂ and C₂H₆ permeance vs pressure was observed, especially in the case of CHFs with wide micropores. The described methodology in this work is proved to be a very valuable tool for elucidating the gas diffusion mechanism in asymmetric CHFs and for defining the optimum operating window for an efficient gas separation process.