Shell Side Mass Transfer Research Articles

Recovery of ammonium nitrogen from human urine by an open-loop hollow fiber membrane contactor

Recovering nitrogen from separately collected urine can potentially reduce cost and energy consumption of wastewater nitrogen removal and fertilizer production. In this study, we developed an open-loop hollow fiber membrane contactor (HFMC) to recover ammonium nitrogen from hydrolyzed human urine. Ammonia capture performance at various feed velocities and pH values was investigated. Be different from closed-loop HFMCs, the feed flow velocity is a vital operating parameter. The ammonia capture efficiency decreased from 80.13% to 33.06% when the feed flow velocity increased from 2.61 × 10−5 to 46.83 × 10−5 m s−1. Increasing feed pH from 10.0 to 12.0 enhanced ammonia capture efficiency from 62.12% to 80.13% by promoting the feed solution free ammonia fraction to increase the driving force. Ammonia mass transfer was investigated based on resistance-in-series model. Shell side mass transfer dominated the ammonia capture in the open-loop HFMC. The empirical shell side Sherwood correlation, Sh=17.4Dh1-φLRe0.6Sc0.33, fitted the experimental data well. The used polypropylene membrane showed good ion rejection performance with more than 99.44% rejection efficiency of PO43− and K+ which can be further recovered. The membrane also showed good stability during the ammonia capture experiments. The hydrophobicity, mechanical and pore properties did not significantly change.

Hollow fiber gas membrane-based removal and recovery of ammonia from water in three different scales and types of modules

Ammonia present in many industrial process streams and effluent streams is beginning to be recovered by means of microporous hydrophobic hollow fiber-based membrane contactor devices with gas-filled pores; the process is often characterized as supported gas membrane (SGM) process. Ammonium sulfate is usually obtained in a sulfuric acid stream on the other side of the membrane. It is useful to develop a quantitative basis for the extent of ammonia removal in such devices. Unlike deoxygenation of aqueous streams in such devices, membrane resistance is quite important for ammonia transport. Ammonia transport modeling in such devices is hampered by the complexity of feed liquid flow in the shell side of commercially used devices and lack of information on membrane resistance where membrane tortuosity introduces considerable uncertainty. The approach adopted here involves studying ammonia transport with the feed solution flowing through the hollow fiber bore where the fluid mechanics is simpler than shell-side flows. Comparison of model-based predictions of overall mass transfer coefficient (ko) with experimentally observed values allows estimation of the membrane mass transfer coefficient (km). One can use such estimates of km to model the observed ammonia transport in small crossflow devices and develop an empirical guidance of the dependences of the shell side mass transfer correlations. Guided by such information and deoxygenation SGM literature, a model was developed for large modules used for ammonia recovery via SGM. Model predictions of performances of the large modules are likely to be useful for various process considerations including the effect of temperature and feed flow rate variations on ammonia removal.

Potentialities of a dense skin hollow fiber membrane contactor for biogas purification by pressurized water absorption

Gas purification technologies are a key step in the technological chain using energy sources such as natural gas, coal gas or biogas, and represent a significant part of production costs. Numerous separation technologies exist to achieve required purity specifications depending on the particular application (absorption, adsorption, cryogenic, membranes). The objective of this work is to present an evaluation of the potentialities of using a membrane contactor based on a dense skin Poly (Phenylene Oxide) (PPO) hollow fiber module for CO2 absorption from biogas by pressurized water through simulations and experiments. In this system, the liquid flowing on the shell-side is in direct contact with the dense skin, thus enabling pressurization and avoiding membrane wetting in order to maintain stable absorption performances.The overall mass transfer coefficient Kov, CO2 removal efficiency, and CH4 loss are evaluated for a range of liquid and gas flow rates. A one-dimensional (1D) model based on resistance in series provides very good predictions of the experimental results when the Graetz approach is used for shell-side mass transfer calculations.Membrane mass transfer resistance is shown to be negligible compared to the liquid resistance. Consequently, the overall mass transfer performances of the dense skin contactor are close to typical values of packed columns (10−5–10−4m/s). Thanks to the increased gas-liquid interfacial area offered by the membrane contactor, a significant volume reduction of the absorption unit (up to 15 process intensification factor) is potentially achievable. Moreover, a selective dense skin could minimize methane losses.

Modeling the Transverse Shell-side Mass Transfer in Hollow Fiber Membrane Contactors at Low Reynolds Numbers

A method for calculating the external mass transfer in a contactor with a transverse confined flow of a viscous incompressible liquid (gas) past hollow fibers at low Reynolds numbers is proposed. The method is based on the concept of regular arrays of parallel fibers with a well-defined flowfield. As a simplest model system, a row of parallel fibers is considered, for which dependences of a drag force and an efficiency of a solute retention on the inter-fiber distance, membrane mass transfer coefficient, Peclet and Reynolds numbers are computed. The influence of the fluid inertia on the mass transport is studied. It is shown that a linear Stokes equations can be used for as higher Re numbers, as denser is the fiber array. In this case the flow field is independent on the Re number, and analytical solutions for the flowfield and fiber sorption efficiency (fiber Sherwood number) can be used.

Supercritical CO2 Extraction of 1‐Butanol and Acetone from Aqueous Solutions Using a Hollow‐Fiber Membrane Contactor

AbstractPorocritical extraction was applied to the separation of 1‐butanol and acetone from aqueous solutions modeling a typical acetone‐butanol‐ethanol (ABE) fermentation broth. The efficiency of two lab‐scale hollow‐fiber membrane contactors of different was studied under various flow rates, solvent‐to‐feed ratios, and pressure conditions. Global and individual, i.e., tube side, membrane, and shell side mass transfer coefficients were determined. The main contribution to mass transfer resistance was found in the shell side in all cases, and experimentally determined shell side mass transfer coefficients were compared to values predicted by existing correlations.

Evaluation of mass transfer characteristics of non-porous and microporous membrane contactors for the removal of CO2

A comprehensive experimental investigation of membrane gas absorption for the removal of carbon dioxide (CO2) using non-porous silicone rubber and porous polypropylene hollow fibre membrane contactors was performed. Water, monoethanolamine (MEA) and diethanolamine (DEA) were used as absorbing solvents. Performances of membranes for gas absorption were evaluated as mass transfer through the membranes, selectivity and removal efficiency of CO2. Water flow rate affected the mass transfer through the porous membrane more than non-porous one. But, increase in flow rate of amine solutions has not affected the mass transfer for both membranes. Mass transfer, compared with water, increased four and eight times roughly for non-porous and microporous membranes, respectively, when amine solutions of 10wt.% were used. Amine solutions also increased the CO2 selectivity 2 and 4 folds roughly for non-porous and microporous membranes, respectively. Basic absorbents including MEA or DEA wetted the hydrophobic porous polypropylene membrane and increased the membrane resistance against the mass transfer. Increase in pressure difference between both sides of the membranes reduced the membrane CO2 selectivity. Whether water or amine solutions were used as absorbent, mass transfer through non-porous silicone rubber was mainly controlled by membrane itself. A mathematical model based on the effective permeability of the gaseous mixtures has been used to assess the performance of both membranes, and a correlation for the shell side mass transfer has been developed.

Experimental and theoretical mass transfer transient analysis of copper extraction using hollow fiber membrane contactors

Extraction of copper (II) from aqueous solutions using trifluoro-acetylacetone (TFA) diluted in 1-decanol, MIBK and tetradecane at 298 and 313 K was investigated experimentally and theoretically. Experimental part of the study concerns with the partitioning of copper (II) in both phases at equilibrium. Distribution of copper (II) in aqueous and organic phases was studied to investigate the hypothesis of constant partition coefficient for solvent extraction mass transfer analysis in hollow fiber membrane contactors (HFMCs). Partition coefficient of copper complexant was measured from extraction equilibrium analysis. Hatta number was calculated to define the kinetic regime. In the second part of experimentation, copper (II) extraction with TFA diluted in 1-decanol using hollow fiber membrane contactor closed system was examined based on aforementioned batch extraction kinetics. Previously developed theoretical model [M. Younas, S. Druon-Bocquet, J. Sanchez, Kinetic and dynamic study of liquid–liquid extraction of copper in a HFMC: experimentation, modeling and simulation, AIChE J. 56 (2010) 1469–1480] was modified to a more robust single-fiber flow model for mass transfer in non-dispersive solvent extraction in HFMC-based integrated plant. The model was, then, validated with the experimental results for dynamic responses of the solute concentration in reservoir. Experimental results were analysed with various shell side mass transfer correlations. It has been shown, by model simulations that the extent and speed of the extraction depend upon the copper (II) and TFA initial concentration, temperature and physical characteristics of solute and solvents. Likewise flow configuration affects the speed of extraction depending upon the nature of solute and solvent.

Kinetics of VOC absorption using capillary membrane contactor

The aim of this work was to experimentally study the acetone absorption in water by means of capillary membrane contactor. A “resistance in series” model was used to describe the rate of mass transfer. The experiments were designed in batch system for laminar and transient flow region. Results were approximated by mathematical model based on mass balance equations. The kinetic model enables to calculate two parameters: the equilibrium constant and mass transfer coefficient and takes into account the gas concentration profile along the equipment. The Wilson method was used to calculate membrane mass transfer resistance. Eight series of experiments for two types of modules and various flow arrangements were made for accurate approximation of membrane resistance and gas/liquid mass transfer coefficients. The results follow previous observations on advantages of the absorption process in the membrane contactors, additionally the process stability and a high performance compared to conventional absorption equipment were confirmed. However, there is still lack of appropriate description of shell side mass transfer in case of randomly packed polymeric fibers. A new correlation is proposed on the bases of conducted experiments.

Comparison of shell side mass transfer correlations in randomly packed hollow fiber membrane modules

In this work, the process of membrane-based solvent extraction was examined with 10% (v/v) tributyl phosphate in kerosene/phenol/water as experimental system, involving solute mass transfer from the aqueous phase (flowing in the shell or tube side) to the organic phase. Shell side mass transfer performance was determined in two randomly packed membrane modules with different packing densities. Experimental data of shell side mass transfer coefficient were compared with model predictions from nine existing empirical correlations. Results showed that the correlations of Gawronski et al. and Viegas et al. were found to give the good predictions in these modules.

Formaldehyde removal from air during membrane air humidification evaporative cooling: Effects of contactor design and operating conditions

Counter-current parallel flow membrane contactors were studied to estimate likely formaldehyde removal rates from air occurring during membrane air humidification and evaporative cooling applications. A mathematical model and experimental measurements of formaldehyde removal are presented to assess the impact of possible operating conditions and contactor configurations on expected formaldehyde removal rates. Removal efficiencies were predominately controlled by shell side mass transfer coefficients. However, removals decrease when water velocity within the fiber lumens is low and water supplied to the modules approach module evaporative losses. While removal rates were found to be adequate to control formaldehyde emissions in most space air-conditioning applications, large volumes of water usage appear to be necessary.

Mathematical modeling of gas–liquid membrane contactors using random distribution of fibers

This work presents a mathematical model for gas absorption in microporous hollow fiber membrane contactors by using a random distribution of fibers. The chemical absorption of carbon dioxide into aqueous amine solutions and sulfur dioxide into water were simulated by this model. The nonlinear mathematical expressions of the component material balance for the liquid, membrane, and gas were solved simultaneously by using a numerical method. The results from the model were compared with four sets of different experimental data in the literature. In addition, the contactors were modeled based on the assumption of regular arrangement of fibers in the shell side by using Happel's free surface as well as plug flow models. The plug flow model was employed to compare the various available equations in the literature for the shell side mass transfer coefficient. The results indicate that the channeling of gas in the shell side decreases the efficiency of contactor significantly. It was found that the random distribution of fibers is a suitable method to simulate the commercial modules. The results also indicate that, the regular Happel's free surface model and the plug flow model are more suitable for handmade modules. The influence of shell side channeling on the contactor performance were investigated in different fiber packing densities, and in various gas and liquid flow rates.

Separation of Zinc by a Non‐dispersion Solvent Extraction Process in a Hollow Fiber Contactor

A process for recovery of zinc from acid solution with di(2‐ethyl hexyl phosphoric acid) (D2EHPA) dissolved in iso‐dodecane was carried out at 20°C in a countercurrent tubular membrane extractor using a hollow fiber as solid support. Experiments were performed at different aqueous metal concentrations (0.1–1.0 g/L), pH 0.1–2.1, and D2EHPA concentrations (2–8 v%). It was found that both the flux of metal and the extraction extent was highly influenced by the extractant concentration and the pH of the feed solution. Overall mass transfer coefficients were determined and related to the tube side, the membrane, and the shell side mass transfer by varying the aqueous flow rate (0.38–0.80 L/min) and organic flow rate (0.22–0.57 L/min) in countercurrent flow. The overall mass transfer coefficient for zinc extraction ranged from 6.2×10−6 m/s to 25.3×10−6 m/s. It was concluded that extraction kinetics were a major contributor to the overall resistance to mass transfer.

Shell-side mass transfer of hollow fibre modules in osmotic distillation process

The effect of packing density of hollow fibre modules on mass transfer in the shell side of osmotic distillation process was studied. The osmotic distillation experiments were carried out with several modules of the packing densities ranging from 30.6 to 61.2%. It was found that the Reynolds number was a function of packing density and packing density affected mass transfer performance. The shell-side mass transfer coefficient increased with the brine velocity. The membrane permeability can be predicted from the experimental flux at the maximum brine velocity. The mass transfer correlation was proposed in order to determine the shell-side mass transfer coefficient in the randomly packed modules for osmotic distillation process. The empirical correlation proposed was fitted to the experimental results and it was found that the mass transfer coefficients calculated from the proposed correlation were in good agreement with those from the experimental data. Comparison of the results obtained from the proposed correlation with other correlations in the literature was discussed.

Theoretical estimation of shell-side mass transfer coefficient in randomly packed hollow fiber modules with polydisperse hollow fiber outer radii

The expression of average shell-side mass transfer coefficient (MTC) in hollow fiber modules is theoretically deduced in this paper. This expression considers the effects of both the polydispersity of hollow fiber outer radii and the randomicity of hollow fiber distribution on the shell-side mass transfer performance for the first time. In the expression, Gaussian function is used to model the polydisperse hollow fiber outer radii, Voronoi tessellation method is used to model the random hollow fiber distribution, and the two distributions are assumed to be independent. Then, the effect of polydisperse hollow fiber outer radii on average shell-side MTC is studied in detail.

A study of mass transfer in hollow-fiber membrane contactors—The effect of fiber packing fraction

Mass transfer in flow parallel to a bundle of fibers is considered. The study is focused on shell-side mass transfer and in particular on the effect of fiber packing fraction φ, a topic very inadequately treated in the literature despite its significance. The process of liquid–liquid membrane-based extraction is examined, involving solute transfer from an aqueous phase (flowing in the shell) to an organic solvent (hexane). Benzaldehyde and hexanal, at low concentrations in water, are the solutes employed, thus obtaining two systems characterized by high partition coefficient. The four modules used are made of hydrophobic polypropylene fibers covering the range of packing densities φ = 0.093–0.402. Quite complete sets of data are collected over sufficiently broad ranges of Reynolds numbers at shell-side ( Re w) and lumen ( Re o). The experimental study is backed by theoretical modelling of mass transfer, based on laminar boundary-layer type analysis at shell-side, presented in a companion paper. It is shown that the dependence of shell-side Sherwood number, Sh w, on Re w to 1/3 power satisfactorily describes the new data, throughout the packing fraction φ-range tested. The effect of φ on shell-side mass transfer (if any) appears to be very weak, probably buried within the narrow range of experimental error; thus, an appropriate correlation for Sh w is recommended, covering the φ-range of practical interest (0.05–0.45). The reliability of the new data (and of the proposed correlation) is supported by the theoretical model predictions, displaying fair agreement with no recourse to any adjustable parameters. Finally, the set of expressions developed and/or tested in the context of this study, for predicting stream exit concentrations in physical membrane-based extraction, are in very good agreement with all the new data.

A study of mass transfer resistance in membrane gas–liquid contacting processes

The change of mass transfer resistance with time has been examined for membrane-based carbon dioxide absorption in water. A commercial polypropylene hollow fiber membrane module was used with gas flow through the lumen side and liquid cross-flow on the shell side. Experiments were carried out for a prolonged period of time to evaluate absorption performance. Absorption flux decreased significantly with time due to a developing resistance to mass transfer. However, the initial flux value was restored after membrane drying, indicating that the additional resistance was reversible. A theoretical model was developed to analyze and predict flux deterioration with time in terms of partial resistances in series. The resistance against pure gas mass transfer was considered to be the sum of the liquid phase resistance and the resistance of the membrane; the latter arises from membrane pore filling by gas or liquid or both. A shell side mass transfer correlation was used for the prediction of the liquid phase resistance. The experimental absorption flux decline with time was attributed to gradual partial pore filling by liquid, and it was used to estimate the magnitude of the membrane resistance and its temporal variation. Model predictions, in good agreement with experimental results, permitted the estimation of liquid penetration into the membrane matrix. Although this was relatively low (∼13%), the resulting resistance of the liquid-filled pores accounted for over 98% of the membrane resistance and for 20–50% of the total resistance to absorption.

Corrections to Correlations for Shell-Side Mass-Transfer Coefficients in the Hollow-Fiber Membrane (HFM) Modules

Mass-transfer correlations for the shell-side mass-transfer coefficients in hollow-fiber membrane (HFM) modules have been published by several authors. The uneven distribution of shell-side flow is due to the random packing of fibers in the module. The results of different authors give apparently conflicting results. This work provides an approach to estimating empirical correlations for HFM shell-side mass transfer. This approach is then applied to explain several literature observations and to offer corrections to earlier published correlations. The dependence of the Sherwood number on both the Reynolds number and the packing fraction are considered. Further work on the packing fraction dependence is required.

Mass transfer in liquid–liquid membrane-based extraction at small fiber packing fractions

Results are presented of carefully controlled mass transfer experiments, complemented by theoretical analysis. A simple test section is employed, comprised of five microporous nearly parallel fibers (fixed within a tubular glass shell), to study mass transfer when the shell-side resistance is controlling. The feed phase, flowing in the shell side, is an aqueous solution of either benzaldehyde or 2-hexanal, whereas n-hexane is used as the organic solvent (in the lumen); the respective partition coefficients at 25 °C are 10 and 57. Accurate data are obtained by using ‘once-through’ flow of both fluids. For the class of problems studied here (i.e. small fiber packing fraction Φ and high Sc number at shell side), a boundary layer solution is obtained for the mass transfer rate to the outer fiber surface. This expression, model variations thereof, as well as similar theoretical results recently reported in the literature, are assessed by comparison with the new data. By combining the analytical solution for the shell with a known expression for mass transfer at the lumen side, a closed-form expression is derived for the solute concentration variation in the organic solvent along the fiber/module. Very good agreement is obtained between experimental data and theoretical predictions, with no recourse to adjustable parameters. An important conclusion of this work, concerning shell-side mass transfer, is the dependence of Sh· Sc −1/3 on Reynolds number to the 1/3 power. Furthermore, the flow rate in the lumen (though not controlling) is shown to be a significant parameter affecting process efficiency.

Shell side mass transfer in a transverse flow hollow fiber membrane contactor

The free surface model was introduced to describe the shell side fluid flow in a transverse flow hollow fiber membrane contactor, and a new method was developed to calculate the shell side hydraulic diameter, the effective average velocity, and the Reynolds number. An empirical shell side mass transfer correlation was presented for commercial Liqui-Cel ® Extra-Flow contactors on the basis of the experimental data reported by Sengupta et al. The data were correlated very well with maximum discrepancies of ±10% between the predicted and observed results.

Shell-Side Dispersion Coefficients in a Rectangular Cross-Flow Hollow Fibre Membrane Module

Membrane processes utilizing hollow fibre membrane modules are gaining increased interest in many industrial applications. However, these modules suffer from shell-side maldistribution and bypassing which results in a loss in efficiency. The shell-side mass transfer performance of these membrane modules strongly depends on the shell-side mixing and the shell geometry. In literature limited information is available on the shell-side mixing of hollow fibre membrane modules. In the present work, shell-side mixing of a rectangular cross-flow hollow fibre membrane contactor is investigated using gas-phase RTD measurements. A novel ultrasound based measurement technique was used to characterize the system. The shell-side mixing of the module is determined in terms of dispersion coefficients in three directions. The axial dispersion coefficient is found to have values between those applicable to pipe-flow and packed bed correlations. This can be attributed to the intermediate packing density of the membrane module. The dispersion in transversal directions, along and across the fibres, is significantly lower compared to that of the axial dispersion. The transversal dispersion coefficient across the fibre is higher and more sensitive to the shell-side velocity compared to the dispersion coefficient along the fibre due to the continuous splitting and remixing of shell-side flow across the fibres.