Regeneration of CO2 absorbent with membrane contactor via pressure swing

https://doi.org/10.1016/j.cherd.2021.01.007Get rights and content

Highlights

  • Membrane contactor was used as a gas-liquid separator in pressure swing process.

  • The two mass transfer mechanisms were analysed through calculation of viscous flux and mass transfer coefficients.

  • When integrated with membrane contactor for CO2 absorption, stable and consistent performance was observed.

Abstract

A new method to regenerate high pressure CO2 absorbent is presented in this paper, where PTFE hollow fiber membrane contactor was used as a gas–liquid separator in pressure swing process. CO2 desorption happened in the membrane contactor at the same time. The two mass transfer mechanisms were analysed through calculation of viscous flux and mass transfer coefficients. Experimental results showed that CO2 loading in regenerated absorbent increased with absorbent flow rate. This was due to decrease in membrane area available for CO2 desorption, as most of the membrane area was utilized for gas-liquid separation. Co-current configuration exhibited higher CO2 desorption driving force compared to counter-current configuration, hence its CO2 loading in regenerated absorbent was lower. When integrated with membrane contactor for CO2 absorption, stable and consistent performance was observed. Unlike conventional flash drum, membrane contactor did not have requirement on flashed gas velocity, because the membrane itself functioned as a screen to prevent fine liquid droplets from escaping the system along with flashed gas. Results from this study show that vessel size reduction up to 2.5 times could be achieved when PTFE membrane contactor was used, with absorbent regeneration energy as low as 0.94 GJ/ton CO2.

Introduction

CO2 removal from gas streams is a common operation in chemical industries such as natural gas processing (Marzouk et al., 2012) and post-combustion CO2 capture (Mulukutla et al., 2014). Many methods have been used in CO2 removal, including low temperature distillation (cryogenic separation), membrane separation, physical absorption, and chemical absorption. Among them, chemical absorption is a well understood and widely used technology, in which CO2 is removed by scrubbing gas stream in absorption column with aqueous absorbents such as potassium carbonate, monoethanolamine (MEA), 2-amino-2-methyl-1-propanol (AMP), diethanolamine (DEA), methyldiethanoamine (MDEA), and then the CO2-loaded aqueous absorbent is subsequently regenerated in regeneration column (Lu et al., 2011).

Chemical absorbents can be regenerated either via steam stripping or pressure swing (Gerhardt and Hefner, 1989). When the targeted CO2 content in treated gas is <50 ppm, which is normally the case in liquefied natural gas process, amines are used as absorbents and steam stripping regeneration is required in order to regenerate the amines into lean amines (Bahadori, 2014). Typical lean amine loading is 0.01 mol CO2/mol amine with 4.15 GJ/ton CO2 of regeneration energy (Li and Keener, 2016; Volkov et al., 2015). When the targeted CO2 content in treated gas is not that low, such as in post-combustion CO2 capture, extent of steam stripping in desorber can be reduced to produce semi-lean amine. Typical semi-lean amine loading is around 0.20 mol CO2/mol amine with 3.50 GJ/ton CO2 of regeneration energy (Jin et al., 2018). Pressure swing can be used to regenerate amines and other chemical absorbents into semi-lean condition, only in natural gas treating as post-combustion CO2 capture is not carried out at high pressure. For this pressure swing process, absorption has to be carried out at high temperature, so that the rich absorbent exiting absorption column only requires minimal heating before pressure swing desorption. This process typically consumes less than 1 GJ/ton CO2 when MDEA and PZ are used as absorbent.

Absorbent regeneration is carried out conventionally in packed column when steam stripping is required, or in flash drums when only pressure swing is required (Gerhardt and Hefner, 1989). Studies on absorbent regeneration using membrane contactor have been reported recently, and most of them adopted regeneration methods without steam stripping in order to reduce energy consumption (Fang et al., 2012; Khaisri et al., 2011; Listiyana et al., 2018; Simioni et al., 2011). For example, Khaisri et al. conducted CO2 desorption experiments using polytetrafluoroethylene (PTFE) hollow fiber membranes at 90–100 °C, with MEA as absorbent and N2 as stripping gas (Khaisri et al., 2011). Desorption temperature was lower compared to conventional steam stripping, but the resulting gas product was in N2-CO2 mixture which could not be utilized or stored directly. Besides that, water condensation in membrane pores might happen due to temperature difference between absorbent and N2 (Simioni et al., 2011). Listiyana et al. studied solvent regeneration using polypropylene (PP) hollow fiber membranes at 30–70 °C, with activated DEA (Listiyana et al., 2018). Vacuum was applied at gas side of the membrane contactor to reduce CO2 partial pressure. This technique was also employed by Fang et al. and they quantified the MEA and water that were lost during CO2 desorption (Fang et al., 2012). Additional energy and equipment might be required to recover the evaporated substances.

In order to avoid potential membrane wetting, some researchers used dense membrane instead of porous membrane. Shutova et al. carried out solvent regeneration with dense poly[bis(trimethylsilyl)tricyclononene] membrane at 100 °C, using a few amine solvents with 10 bar pressure difference between liquid phase and gas phase (Shutova et al., 2014). There was no membrane wetting, but the dense membrane exerted certain extent of mass transfer resistance. Kosaraju et al. carried out a study using porous poly(4-methyl-1-pentene) (PMP) membrane with an ultra-thin dense skin in order to prevent membrane wetting while minimizing mass transfer resistance. They found that the membrane was not able to eliminate permeation of MEA from liquid phase to gas phase completely (Kosaraju et al., 2005). Volkov et al. and Dibrov et al. conducted solvent regeneration using asymmetric poly(vinyltrimethylsilane) membrane (PVTMS) with a dense selective layer at 100 °C, and observed no liquid penetration through the membrane. They pointed out that water evaporation during the process could increase energy consumption, and their dense membrane was able to reduce water evaporation (Dibrov et al., 2014; Volkov et al., 2015). However, water flux through non-porous membranes including PVTMS were orders of magnitude greater than the corresponding CO2 flux according to study conducted by Scholes et al. (Scholes et al., 2015). Besides polymeric membrane, ceramic hollow fiber membranes which could better withstand high regeneration temperature without morphological change was also used to avoid membrane wetting (Koonaphapdeelert et al., 2009). However, its CO2 desorption efficiency was not encouraging.

Regardless of membrane types, most of the published works focused on using stripping gas or vacuum to replace steam stripping. To authors’ best knowledge, no publication has been sighted on using membrane contactor to carry out high temperature pressure swing so far. In fact, this is the most energy efficient conventional absorbent regeneration method in natural gas treating, consuming less than 1 GJ/ton CO2. Low pressure flash drum which functions as a gas–liquid separator in the conventional system can be replaced by membrane contactor as shown in Fig. 1. Fig. 1b shows that A and B are the two possible gas phase outlets (tube side of hollow fiber membranes). When A is open and B is close, flashed gas and amine flow in counter-current manner. When B is open and A is close, flashed gas and amine flow in co-current manner. Significant size reduction can be achieved as membrane contactor does not require large vessel diameter to avoid high absorbent loss in the form of fine liquid droplets. The micro-pores on hollow fiber membranes separate gas from liquid by only allowing gas phase to pass through. Detailed mass transfer mechanism in membrane contactor is discussed in Section 2.

The objectives of this study are to investigate feasibility and benefits of using membrane contactor to regenerate CO2 absorbent via pressure swing. Mass transfer mechanism was discussed and analyzed through calculation of viscous flux and mass transfer coefficients. Experiments were carried out using porous PTFE hollow fiber membranes and aqueous MDEA solution to investigate the effect of flow configuration under various flow rates. Continuous experiments combining CO2 absorption and absorbent regeneration were carried out to evaluate performance stability. Performance of membrane contactor was also compared against the conventional system. This study provides a new method of using membrane contactor for CO2 absorbent regeneration which has never been reported before, offering significant size reduction and low energy consumption benefits.

Section snippets

Theoretical background

Mass transfer in CO2 absorption had been discussed in previously published paper (Chan et al., 2020; Kang et al., 2017). Therefore, only absorbent regeneration was covered in this section.

Materials

Pre-mixed CO2-N2 gas cylinders containing 26 % CO2 was used in this study. The gas compositions of feed gas and treated gas were determined using gas chromatography (Agilent 490 Micro GC). Water solution of MDEA (44 wt.%) was used as absorbent in regeneration study. For integrated experiment of CO2 absorption and absorbent regeneration, water solution of MDEA and piperazine (28 wt.% and 5 wt.% respectively) was used in order to have similar solvent composition with conventional system.

Absorbent regeneration in membrane contactor

Table 2 shows the results of absorbent regeneration experiments carried out under different rich amine flow rates. All experiments were carried out under similar regeneration temperatures and initial rich amine CO2 loadings. Flashed CO2 was estimated using Aspen Hysys simulation software (Electrolyte Non-Random Two Liquids method provided in “Acid Gas Chemical Solvent” property package). It increased with rich amine flow rate due to increased amount of flashed gas generated during pressure

Conclusion

PTFE membrane contactor was used to regenerate CO2 absorbent successfully via a novel high temperature pressure swing process. It separated flashed gas from liquid absorbent and at the same time allowed CO2 to be desorbed from the absorbent. Viscous flux and mass transfer coefficients were used to describe the process. Theoretical overall mass transfer coefficients (Koverall) were in good agreement with experimental values (Kexp). Experiments with high amine flow rate had higher overall mass

Declaration of Competing Interest

The authors report no declarations of interest.

Acknowledgement

The authors would like to thank PETRONAS Research Sdn Bhd, Youth Innovation Promotion Association of the Chinese Academy of Sciences (2016171), Liao Ning Revitalization Talents Program (XLYC1807240), and The National Key R&D Program of China (2017YFB0603403) for the support.

References (36)

Cited by (6)

  • PTFE porous membrane technology: A comprehensive review

    2022, Journal of Membrane Science
    Citation Excerpt :

    Therefore, the membrane cleaning protocols and a large-scale implementation of desiccant regeneration under practical condition should be further studied in order to allow continuous liquid desiccant regeneration. In fact, PTFE membrane operated in MCs has been widely used in some practical production, also PTFE membranes also gradually appeared in bromine extraction [192], absorbent recovery [193,194], degassing and other fields. Compared with traditional distillation towers, evaporators, etc., it has the advantages of small size and weight, but due to the limitations of membrane fouling and limited operating pressure, further improvements are still needed in terms of membrane materials and membrane processes.

  • Process Intensification of CO<inf>2</inf>Desorption

    2022, Industrial and Engineering Chemistry Research
View full text