Subscriber access provided by Gothenburg University Library
Article
Characterization and Modeling of Operating Curves for Membrane Microseparators Lu Yang, Nopphon Weeranoppanant, and Klavs F. Jensen Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.7b03207 • Publication Date (Web): 03 Oct 2017 Downloaded from http://pubs.acs.org on October 8, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Industrial & Engineering Chemistry Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 11
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Characterization and Modeling of the Operating Curves of Membrane Microseparators Lu Yang1, Nopphon Weeranoppanant1,2 and Klavs F. Jensen1,* 1. Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA 2. Department of Chemical Engineering, Faculty of Engineering, Burapha University, Chonburi, 20131, Thailand ABSTRACT: The membrane microseparator is a milliliter‐scale flow chemistry module that continuously separates a bi‐ phasic flow via a PTFE microporous membrane. It has found a wide range of application in the continuous manufacturing of active pharmaceutical ingredients and fine chemicals, especially those involving multiple synthetic steps. Yet, it is tech‐ nically challenging to predict and control accurately the pressure balance needed for successful phase separation. In this paper, we present systematic modeling of the operating ranges of the membrane microseparator. We characterize the re‐ tention and breakthrough phenomena of the device, and develop two new analytic models for retention and breakthrough by taking into consideration the tortuosity factor and pore size distribution. The new models are shown to be better pre‐ dictors of the experimental results compared to the original theoretical models based on simple Young‐Laplace equation and straight‐channel Hagen‐Poiseuille equation.
1. Introduction With the rapid development of flow chemistry technol‐ ogy in the recent two decades, a wide variety of microflu‐ idic modules have been developed that facilitate the exe‐ cution of chemical synthesis and workup on the mi‐ croscale.1 As a ubiquitous unit operation, continuous mul‐ tiphase separation has become an indispensable compo‐ nent downstream of multiphase reactors in microscale flow chemistry applications. Due to the decrease in length scale, phase separation by gravity is no longer feasible on the microscale. Instead, researchers have exploited the dominant capillary force to achieve successful continuous separation driven by the difference in wetting characteris‐ tic of the two fluids. Specifically, Gunther et al. fabricated a capillary microseparator that includes an array of mi‐ cron‐scale capillaries as side channels, which were able to absorb the wetting phase while leaving the non‐wetting phase in the main channel.2 Similar designs have since been adopted by researchers to perform a wide array of tasks, including reaction workup, extraction, and demulsi‐ fication. 3‐7 While the capillary microseparator enjoys many bene‐ fits, including well‐defined straight pores that prevent clogging and facilitate CFD modeling, it is rather difficult to scale up to accommodate higher throughput. The maxi‐ mum permeate removal flow rate through the capillary mi‐ croseparator is equal to the flow rate of the wetting fluid
through each capillary, , multiplied by the total num‐ ber of capillaries, n. is directly proportional to the pressure drop across the capillary, which is limited by the breakthrough pressure threshold.8 Therefore, without , the permeate removal flow much flexibility around rate scales linearly with the total number of capillaries, n. For higher flow rates, it becomes uneconomical or even im‐ practical to fabricate microseparator devices with hun‐ dreds or thousands of capillaries in parallel. Separators employing microporous membranes offer al‐ ternatives to microcapillary arrays for achieving separation on a larger scale. The application of microporous mem‐ branes as phase separators and contactors originated dec‐ ades ago in the field of analytical chemistry. Nord spear‐ headed the use of microporous membrane as a sample preparation method for an extraction manifold in the flow injection analysis (FIA).9 He also tested a wide array of membrane materials, and recommended the use of PTFE membrane with polyethylene support for future studies, due to its high durability and solvent compatibility. Valcar‐ cel and de Castro reviewed the application of microporous membranes as a tool for continuous separation for FIA,10 and Jonsson and Mathiasson summarized the use of mem‐ brane technology in sample enrichment as applied to ana‐ lytical methods, including FIA, gas chromatography (GC) and high‐performance liquid chromatography (HPLC).11 Following this line of work, Cai et al. fabricated a mem‐ brane‐based contactor in a microfluidic chip, in which the
ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
aqueous and organic phase each flow on one side of the microporous membrane, so that the analyte in the aqueous phase could be extracted into the organic phase without the need for further separation.12 Kralj et al. applied microporous PTFE membrane to re‐ place microcapillary arrays in achieving phase separation in flow chemistry applications (Figure 1a).13 Compared to microcapillary separators, the membrane separator enjoys several distinct advantages: (1). The membrane can be eas‐ ily scaled up using similar configurations, and can accom‐ modate a much wider range of flow rates. While the micro‐ separators are mostly designed for lab‐on‐chip demonstra‐ tions and microchemical assays, the membrane separator is compatible with larger‐scale flow chemistry applications and has promising applications in small scale production. (2). The membrane‐based separators are easier to assemble as it eliminates the microfabrication of each capillary. (3). It also provides better reusability, as a membrane can be easily be substituted if it becomes clogged, fouled or oth‐ erwise dysfunctional, while the entire capillary microsepa‐ rator chip would have to be replaced under the same cir‐ cumstance. (4). The membrane separator has much smaller pores (0.1 µm to 10 µm) compared to the capillary separator (20 µm), which becomes essential when separat‐ ing systems with low interfacial tensions.
Page 2 of 11
There are also three distinct operating regimes in which the membrane separator (Figure 1c) operates, the retention regime, the complete separation regime and the break‐ through regime.14 The three regimes are controlled by one variable, the pressure drop across the membrane, ∆ . If ∆ is too low, the pressure gradient would not be enough to push all the organic phase through the PTFE membrane, thus resulting in retention of the organic phase in the aqueous stream. This pressure threshold is deemed the retention pressure threshold, ∆ . On the other hand, if ∆ is too high, both the organic phase and the aqueous phase will permeate through the membrane, re‐ sulting in the contamination of the organic stream by the aqueous phase. This pressure limit is the breakthrough pressure threshold, ∆ . When ∆ ∆ ∆ , complete separation takes place, where the biphasic flow from the inlet is continuously sep‐ arated into an organic phase and an aqueous phase. Kralj et al. designed a polycarbonate module with a PTFE membrane inside to demonstrate the separation concept, then fabricated a silicon device that integrates multiphase mixer, contactor and separator into a single chip.13 The membrane microseparator has since then been applied by many researchers as an indispensable workup tool for
Figure 1. (a) A scheme showing the composition of a membrane separator, where there is one biphasic feed inlet and two outlets, one for the aqueous stream and one for the organic stream.13 (b) An SEM image of the active surface of the PTFE membrane15 (c) Schemes demonstrating the three operating modes of the membrane separator14 (d). Comparison of the theoretical limit (solid line) and experimental measurements (dots). The yellow region denotes the retention regime, the green region denotes the com‐ plete separation regime, and the red region denotes the breakthrough region. In the retention and complete separation regime, the permeate phase consists solely of the organic phase, whereas in the breakthrough regime, the permeate is a mixture of organic and aqueous phases. The normalized permeate flow rate is the ratio of the permeate flow rate to the incoming organic phase flow rate, and its value should be equal to 1 when complete separation occurs.14
ACS Paragon Plus Environment
Page 3 of 11
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
achieving multistep continuous synthesis.16‐26 For instance, the membrane separator was applied by Sahoo et al. to in‐ tegrate three chemical reactions continuously with separa‐ tion steps in between.27 Alimuddin et al. employed the membrane separator to design an automated, high‐ throughput method for measuring the oil/water distribu‐ tion coefficient of a molecule to replace the traditional shake‐flask method.28 Hartman et al. expanded the appli‐ cation of the membrane separator to the field of microscale distillation, adding yet another useful module to the toolbox of continuous flow chemistry synthesis.29,30 Heider et al. further demonstrated the scalability and practicality of the membrane separators by using scaled‐up membrane separators in the continuous synthesis of a pharmaceutical product, where each separator unit was equipped to han‐ dle flow rates between 10 – 1,000 mL/h, and the membrane ran successfully for 100 h before fouling occurred.31 Based on the multitude of applications, Cervera‐Padrell et al. evaluated different liquid‐liquid separation technologies in the context of large‐scale industrial applications for con‐ tinuous pharmaceutical production, and concluded that the microporous membrane separator is one of the most suitable, flexible and robust device for this purpose.32 Despite its versatility and adaptability, an inevitable dif‐ ficulty in the inline application of the membrane separator is the stringent requirement for pressure control. In order for the device to function properly, the pressure at both the wetting and non‐wetting outlets must be maintained me‐ ticulously, as deviations in each value would result in er‐ rors in ∆ , thus hindering the smooth operation of the device. This is especially difficult given that the separator, when used in the middle of a sequence of flow chemistry modules, is frequently subject to pressure fluctuations from upstream and downstream. In addition, two sets of pressure control valves take up extra space and make the system cumbersome, especially when multiple separation steps are involved. To address this problem, Adamo et al. designed a membrane separator device coupled with a de‐ formed diaphragm that provided a constant pressure dif‐ ference across the membrane, regardless of the outlet con‐ ditions.14 By fine‐tuning the configuration of the dia‐ phragm, consistent and complete separation performance could be guaranteed. The same module has been employed as an integral part of a refrigerator‐sized “flow chemistry factory” to synthesize and formulate pharmaceuticals on demand33 and has been used as the separator in the con‐ tinuous solvent‐free oxidation of alcohols.34 Moreover, the independent nature of the units enabled their integration into counter‐current extraction.35 Regardless of the pressure control mechanism, the key issue in operating the membrane separator is to under‐ stand the pressure ranges over which complete separation occurs. While theoretical expression exist for the break‐ through and retention threshold, based on the capillary flow model, they largely fail to predict the operating ranges of the membrane separator (Figure 1d)14 This behaviour re‐ sults from the polydispersity and tortuosity of the fluid
path in the membrane (Figure 1b) as compared to the well‐ defined micromachined straight channels in the capillary separator. The polydispersity of the membrane pore sizes reduces the breakthrough pressure threshold, as break‐ through would occur through the largest pore, and conse‐ quently narrows the operating window. In addition, the tortuous fluid path in the membrane increases flow re‐ sistance and raises the retention pressure limits compared to the straight and uniform channels in the microfabri‐ cated capillary separators. The discrepancy between the theoretical framework and experimental measurements was observed by Adamo et al (Figure 1d).14 Specifically, while the theoretical model was able to qualitatively cap‐ ture the three operating regimes of the device, it tends to under‐predict the onset of the retention and over‐predict the start of the breakthrough, thus overestimating the pressure range for complete separation. Recognizing the strong need for an increased under‐ standing of the operation the device, we seek to establish an analytic model for understanding and predicting the operating ranges of the membrane microseparator based on systematic experimental characterization. We will pre‐ sent the derivation of a new set of analytic models, com‐ pare the results of the new model with that of the original ones, and demonstrate the use of the new model in pre‐ dicting the performance of the membrane separators. 2. Methods 2.1 Experimental Measurement The membrane testing unit (Figure 2) was machined in perfluorinated polymers (HDPE), which was embedded in a metal shell for higher pressure rating. The membrane used was made of PTFE with a nominal pore diameter of 1.0 µm (Pall, USA). It had a supporting layer of PTFE. FEP coated o‐rings were used to provide a leak proof seal. The membrane area was 4.0 × 0.8 cm2. Subtracting from it the areas occupied by the supporting islands through which no permeation occurred, the active membrane area was 157 mm2. This unit was used for both breakthrough and reten‐ tion studies. To generate a breakthrough curve, the unit was installed with pressure gauges and back pressure reg‐ ulators (BPR) on the two outlets (Figure 2, highlighted with red). Two liquid‐liquid (immiscible) combinations were tested: (a) water‐toluene, (b) water‐hexane. The two BPRs were connected to adjustable compressed gas tanks so that the BPR set points could be changed dynamically during experiment. The two immiscible phases, delivered into the system by peristaltic pumps, were mixed in a T‐ junction and then introduced into the membrane testing unit. Due to high hydrophobicity of the membrane, the or‐ ganic phase tended to wet and permeate through the membrane while the aqueous phase would be retained. However, provided a sufficiently high pressure difference across the membrane, the aqueous phase could partially, or even fully, permeate through the membrane too. By ad‐ justing the regulators of the compressed gas tanks, we
ACS Paragon Plus Environment
Industrial & Engineering Chemistry Research
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
could set the pressures on the BPRs. At each adjustment, the setup was allowed to equilibrate as the reading on the gauges stopped changing. We then recorded the actual pressures of the two gauges, and used them to calculate the pressure difference between the two sides of the mem‐ brane (∆ ). The flow rates of both aqueous and organic phases in the two outlets were measured simultaneously using graduated cylinders and timers. Based on repeated runs under the same experimental condition, we were able to quantify the random errors in this experimental setup. For the breakthrough curve measurement, the relative standard error of ∆ is 2.5%, and that of the normalized permeate flow rate is 7.1%, which is likely due to the pulsa‐ tile nature of the pump.
2.2.1 Single phase flow through the membrane un‐ der retention limit The retention limit occurs when there is just enough pressure to propel the organic phase through the mem‐ brane, and could be described by the Hagen‐Poiseuille equation, if the pores were straight, uniform in size, and did not intersect: ∆
∆
,
(2)
Here L is the thickness of the membrane, is the flow rate of the organic phase through the membrane, is the total number of capillaries, and is the averaged radius of the membrane pores.
In this testing unit, the retention occurred at very low pressure differences (