stripping of phenol

Chemical Engineering Journal 308 (2017) 727–737

Contents lists available at ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Importance of uniform distribution of impregnated trioctylphosphine oxide in hollow fiber membranes for simultaneous extraction/stripping of phenol Kreeti Das, Prashant Praveen, Kai-Chee Loh ⇑ Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117576, Singapore

h i g h l i g h t s  Trioctylphosphine oxide impregnation in polypropylene membranes was optimized.  Air drying conditions were found to be crucial to achieve uniform distribution.  Morphological analysis indicated uniform distribution under optimal conditions.  Impregnated membranes were used for simultaneous extraction and stripping.  High removal rates, efficiency and stability were achieved during phenol extraction.

a r t i c l e

i n f o

Article history: Received 19 July 2016 Received in revised form 9 September 2016 Accepted 22 September 2016 Available online 23 September 2016 Keywords: Hollow fiber membranes Phenol Solvent extraction Supported liquid membrane Trioctylphosphine oxide

a b s t r a c t Trioctylphosphine oxide (TOPO) was impregnated in polypropylene hollow fiber membranes. The impregnation process was optimized and a uniform distribution of TOPO was achieved within the membrane walls, when the membranes were treated with a carrier solution containing 600 mg/L TOPO, and air-dried for 30 min at a low air flow rate (Reair = 9.2). The resulting extractant impregnated hollow fiber membranes (EIHFM) were characterized by scanning electron microscopy, water entry pressure, gas permeability and mercury porosimetry analysis, all of which showed significant structural and morphological changes in the EIHFMs; pore size, porosity and tortuosity were estimated to be 0.5 lm, 0.09 and 33, respectively. The EIHFMs exhibited high mass transfer rates and removal efficiency during simultaneous extraction and stripping of phenol. At an initial phenol concentration of 200 mg/L, 99% phenol was extracted from the wastewater within 7 h, whereas more than 91% phenol was recovered in the stripping solution, yielding a concentration factor of 1.79. The performance of the EIHFMs did not change significantly during consequent operations under identical conditions, indicating the stability of impregnation. These results suggest that uniformly impregnated TOPO-based EIHFMs can be promising in the recovery of phenols from wastewater. Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction Phenol and its derivatives are aromatic compounds of high commercial interest as they are used as raw materials in manufacturing various chemicals, plastics, pulp/paper, dyes, pigments, adhesives and pharmaceutical products [1]. Unfortunately, they are also common pollutants in industrial wastewater. The presence of phenols in industrial effluents is a matter of concern as they exert adverse effects on aquatic ecosystem due to their high toxicity, molecular ⇑ Corresponding author at: Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117585, Singapore. E-mail address: [email protected] (K.-C. Loh). http://dx.doi.org/10.1016/j.cej.2016.09.105 1385-8947/Ó 2016 Elsevier B.V. All rights reserved.

recalcitrance and high reactivity, which may result in formation of even more hazardous compounds [2–4]. Hence, it is important that phenol and its derivatives be removed from industrial effluents before their discharge into natural water streams. Liquid-Liquid extraction using immiscible organic solvents is widely used in the removal of phenol from wastewater [5–8]. However, this technique exhibits certain limitations associated with phase dispersion and emulsion formation, which leads to difficulty in phase separation at later stages [9]. Membrane-based liquid/liquid extraction, especially the use of supported liquid membranes (SLM), can prevent phase dispersion and offers the advantages of compact design, low solvent and energy requirement, low operating cost and high mass transfer rates. However,

728

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737

SLMs are inherently unstable as the organic solvent is gradually lost from the membranes due to erosion [10,11]. Non-dispersive solvent extraction can also be performed using solvent encapsulated polymeric microcapsules [12,13]. However, these microcapsules cannot perform simultaneous extraction and stripping. Recently, Praveen and Loh [14–17] developed extractant impregnated hollow fiber membranes (EIHFMs) for the extraction of phenol from wastewater by impregnating polypropylene hollow fiber membranes with a solid organophosphorus extractant, trioctylphosphine oxide (TOPO) [18]. The EIHFMs exhibited high stability, high mass transfer rate and high partitioning capacity for phenol. Moreover, a very small amount of organic solvent was required in EIHFM preparation, whereas no organic solvents were needed during EIHFM-based extraction, which made this technique ‘solventless’. However, the EIHFMs failed to perform simultaneous extraction and stripping of phenol due to the nonuniform distribution of TOPO within the supporting membranes, which resulted in a high mass transfer resistance for phenol diffusion through the membranes. So far, the approach used in EIHFM preparation has been based on soaking hydrophobic membranes in dichloromethane (DCM) enriched with a high concentration of TOPO; DCM was subsequently evaporated from the membranes to facilitate TOPO impregnation in the membrane pores and surfaces [14–17]. The major drawback of this approach was that DCM evaporation was not performed under controlled conditions. The carrier solvent exited the membranes mainly through the outer surface resulting in a much higher deposit of TOPO on the outer membrane surfaces, as compared to that in the membrane walls or the inner membrane surfaces. We anticipated that this limitation in EIHFM preparation could be addressed by controlling the solvent evaporation from the membranes through careful adjustment of the drying time and the air flow rate during drying. DCM could also be forced to exit the membranes through the inner membrane surface by pumping air through the tube sides of the membranes, or by slowing DCM removal through the outer membrane surface by submerging the solvent soaked membranes in water. Using this approach, EIHFMs with uniform extractant distribution could be prepared which could be used for simultaneous extraction and stripping of phenol. The resulting EIHFMs would then be used as both, a membrane contactor, as well as a partitioning phase. Apart from the changes in the impregnation strategy, a systematic and detailed characterization of the EIHFMs is also needed for better understanding of the changes in the morphology and mass transfer properties of the membranes. The objective of this research was to design and optimize the membrane drying process in EIHFM preparation to achieve a uniform distribution of TOPO within the polypropylene membranes. A detailed characterization of the EIHFMs was undertaken to determine the optimal conditions which could maximize TOPO impregnation. The enhanced performance of the EIHFMs with uniformly distributed extractant was investigated through simultaneous extraction and striping of phenol from synthetic wastewater using a contactor module packed with the EIHFMs. The novelty of this research lies in developing a methodology for uniform TOPO impregnation within polymeric membranes that would result in better performance to simultaneously extract and strip phenol in the phenol recovery process. 2. Materials and methods 2.1. Reagents and analytical methods All the chemicals used in this research were of analytical grade and were purchased from Merck or Sigma-Aldrich. The chemicals were used as received from the suppliers.

Phenol concentration in the aqueous phase was determined by measuring absorbance at 270 nm using a UV spectrophotometer (UV-1240, Shimadzu, Japan). Phenol concentration in the solid phase was calculated using mass balance from the difference between initial phenol concentration in the wastewater and total phenol concentration in the two aqueous phases (wastewater and stripping solution) at any time. The concentration factor was calculated as the ratio of the concentration of phenol recovered in the stripping solution to the initial feed phenol concentration. 2.2. EIHFM preparation and characterization 2.2.1. Membrane module The membrane modules were prepared by potting Accurel PP 50/280 polypropylene membranes (Membrana GmbH, Germany) using epoxy resin into contactors that resembled shell-and-tube heat exchangers as described in Loh et al. [19]. The specifications of the modules are given in Table 1. 2.2.2. TOPO impregnation The carrier solvent (DCM) containing TOPO was circulated through the shell side at 5 ml/min for 2 h to impregnate TOPO into the membranes. This was followed by the drying process wherein water was circulated in the shell side to prevent DCM evaporation through the outer surface of the membranes, while air was pumped through the lumen side at a controlled flow rate. The EIHFM preparation process was varied by changing TOPO concentration in DCM (200, 400 and 600 g/L), air flow rate (Reair = 4.6, 9.2 and 18.4) and drying time (15, 30, 60 and 90 min) to determine the optimum impregnation conditions. Reair was calculated using air velocity in the lumen of the membrane module, the inner diameter of the membranes, and the density and viscosity of air at room temperature. After DCM had been evaporated, the lumen of the hollow fiber membranes was flushed with water to remove loosely attached TOPO from the membrane surface followed by a leakage test. All the experiments were conducted in triplets for reproducibility. 2.2.3. EIHFM characterization The EIHFMs were characterized by measuring the amount of TOPO impregnated, by visualizing the cross sections and the surfaces of the EIHFMs under SEM, and by performing liquid entry pressure of water (LEPw) test, gas permeability test and mercury porosimetry to observe any structural or morphological changes and to determine the pore size, porosity and tortuosity of the EIHFMs. The amount of TOPO impregnated into the fibers was determined by measuring the weight of the module before and after the impregnation of TOPO. The distribution of TOPO in the membrane cross-sections and on the surfaces was analyzed using Scanning Electron Microscopy (SEM) (JEOL JSM 5600-LV) after sputtering the membrane samples with platinum.

Table 1 Specifications of the membrane contactor. Characteristics

Values

Casing material Casing inner diameter Membrane inner diameter Membrane thickness Pore size Porosity Effective Fiber Length Number of Fibers Effective shell volume Effective lumen volume

Glass 0.7 cm 280 lm 50 lm 0.2 lm 0.5 20 cm 50 6.5 ml 0.62 ml

729

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737

The LEPw was estimated using an experimental setup as described by Lalia et al. [20]. Briefly, a salt solution (3.5% NaCl in water) was passed through the lumen side of the membrane module, keeping the shell side outlet dipped into a reservoir filled with distilled water. The pressure of the salt solution was increased stepwise by 0.5 bar using nitrogen and the pressure was maintained for 30 min at each condition. The conductivity of the water in the reservoir was measured after 30 min using a conductivity meter. An increase in the conductivity indicated that salt solution had permeated through the membranes and the corresponding pressure was considered the LEPw. The gas permeation test was conducted in an apparatus as described by Wang et al. [21]. In essence, 5 pieces of membranes of length 10 cm each were potted inside an aluminum holder at one end, while the other end of the fibers was sealed with epoxy resin. The membranes were then inserted into a steel casing acting as the shell which was sealed with aluminum holders using rubber O-rings. Nitrogen was pumped into the shell side from a gas cylinder at upstream pressure in the range of 14–30 psi. Nitrogen permeation rate through the membranes was measured at room temperature using a flow meter. The pore size and effective porosity were determined from the relationship between nitrogen permeation flux and pressure using equations described elsewhere [21–23]. Porosity and pore size analysis of the membranes were conducted using Micromeritics mercury porosimeter (AutoPore IV 9500 V1.09). The membranes (about 100 mg) were cut into small sections (4–5 mm) and each section was then cut longitudinally to expose the inner surface area of the membranes. The samples were dried, placed in a container and degassed. While the container was still evacuated, mercury was allowed to fill the container. The volume of mercury intruded into the membrane pores was measured as a function of increasing pressure, which was used to compute pore size using equations described elsewhere [24]. 2.3. Extractive recovery of phenol using EIHFMs In reference to Fig. 1, for the simultaneous extraction and stripping of phenol, a 250 mL Erlenmeyer flask containing 100 ml phenolic wastewater at pH 4–6 was used as the feed solution, whereas a 100 mL Erlenmeyer flask containing 50 ml of 0.2 M sodium hydroxide was used as the stripping solution. The feed and the stripping solutions were pumped on the opposite sides of the membranes in concurrent mode. The flow rates on the lumen and shell sides were maintained at 7.5 and 14 ml/min, respectively. Samples were periodically withdrawn to determine phenol concentrations. At the end of every experiment, both the shell and tube sides of the contactor were washed with 0.2 M sodium hydroxide to remove any remaining phenol in the EIHFMs. This was followed by washing with water. All experiments were carried out in triplets. 3. Results and discussion 3.1. EIHFM preparation and characterization Fig. 2 shows the SEM images of cross-sections of two EIHFM samples collected from two different fibers in a single membrane module prepared using 600 g/L TOPO solution at air flow rate of Reair = 4.6 and drying duration 30 min. TOPO in the EIHFMs appeared as white solid deposits on the membrane surfaces and cross sections and TOPO distribution also varied significantly from one membrane to another. It was seen that while one fiber had uniform distribution of TOPO (Fig. 2a) the other fiber had non-uniform distribution (Fig. 2b). These fiber-to-fiber variations were caused

EIHFM module

Phenol feed

Stripping solution

Magnetic stirrer Peristaltic pump

Peristaltic pump Fig. 1. Schematic diagram of the experimental setup for simultaneous extraction and stripping.

by the non-uniform exposure of the membranes to air during the drying process due to the large number of the membranes in the contactor module, as well as the differences in the roughness of the inner surfaces of the membranes. When air was pumped in the lumen side and DCM was evaporated from the membranes, thin layers of TOPO formed on the inner surfaces of the membranes. As a result, there was fiber-to-fiber variation with regards to inner surface roughness. It has been reported that in tubes of diameter less than 0.6 mm, pressure drop during flow is significantly affected by surface roughness [25]. The changes in pressure drop in different fibers led to a difference in the drying condition in different membranes, resulting in the non-uniform distribution of TOPO among them. The variation in TOPO distribution between the membranes within the same module was undesirable, and it could affect the consistency of extraction when using the EIHFMs. Although TOPO distribution in the membranes improved on changing the drying conditions, not all the cross-sections showed the same kind of distribution. Since there was no absolute uniformity and the EIHFMs were prepared under different conditions, a distribution consistency (DC) indicator was formulated to describe the nature and uniformity of TOPO distribution within the EIHFMs. DC was defined as the ratio of number of fibers showing uniform distribution to the total number of fiber samples observed under SEM. Optimization of the drying conditions was then performed to maximize DC, as well as the weight of the impregnated TOPO. Fig. 3 shows the effects of varying TOPO concentration in the carrier solvent on the impregnation of TOPO in the EIHFMs. Three concentrations of TOPO in DCM were investigated: 200 g/L, 400 g/L and 600 g/L, while maintaining a constant air flow rate (Reair = 4.6) and drying duration (30 min). The amount of TOPO impregnated in the EIHFMs and DC increased when the amount of TOPO in DCM was increased. The amount of TOPO impregnated using 200 g/L TOPO/DCM was 260 mg, which increased to 500 mg when 600 g/ L TOPO/DCM was used. On the other hand, TOPO distribution in the EIHFMs obtained using 200 g/L TOPO/DCM was poor and inconsistent, with a DC close to zero, but the DC improved linearly with rising TOPO concentration, and reached 22% and 45% at TOPO/ DCM concentrations of 400 and 600 g/L, respectively. The increase in DC with rising concentration of TOPO in DCM was a result of the changes in the density and viscosity of the organic solvent. The addition of TOPO increased the viscosity of the DCM solution and slowed the movement of DCM through the membrane pores

730

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737

(a)

Uniform impregnation of membrane surface and walls

(b)

Impregnation on membrane surfaces, not in the wall

Fig. 2. SEM images of the EIHFM showing TOPO distribution within the membranes: (a) uniform distribution as TOPO is spread evenly within the thickness of fibers throughout the cross-section, and; (b) non-uniform distribution as TOPO is evenly spread in some parts and non-evenly spread in some parts of the cross-section.

600

0.5

Weight Distribution

0.4 0.35

400

0.3 300

0.25 0.2

200

0.15 0.1

100

Distribution consistency

Weight of TOPO (mg)

500

0.45

0.05 0

0

200

400

600

TOPO conc. in DCM (g/L) Fig. 3. Effect of increasing TOPO concentration in DCM on the amount of TOPO impregnated, and DC of the EIHFMs.

towards the inner membrane surface during the drying process; a combined effect of higher TOPO content, density and viscosity therefore led to a greater accumulation of TOPO within the membrane walls, and an increase in the weight of TOPO impregnated in the EIHFMs. When a higher amount of TOPO was impregnated in the EIHFMs (hence a higher density of TOPO per unit volume of the membranes), higher DC of the EIHFMs was obtained. It was also observed that although the increase in the weight gain

was comparatively small when TOPO concentration was increased from 400 to 600 g/L, this increase in the amount of impregnated TOPO doubled the DC value. The next attempt was to obtain better and more uniform distribution of TOPO within the EIHFMs; experiments were performed to improve the evaporation of DCM from the membranes during air drying through optimizing the air flow rate and the drying time. In these experiments, only 600 g/L TOPO/DCM was used for EIHFM preparation. Fig. 4 shows the weight gain and the DC in the EIHFMs as a function of the air flow rate (Reair), and the drying time. At Reair = 4.6, the amount of TOPO impregnated in the EIHFMs increased with increase in the duration of air drying from 30 to 60 min. However, the amount of impregnated TOPO decreased when the drying time was increased further to 90 min. A similar trend was observed in the DC, which improved when the drying time was increased from 30 to 60 min but decreased drastically when the drying time was increased to 90 min. While the changes in DC was attributed to the changes in the TOPO content of the EIHFMs, the decrease in the amount of impregnated TOPO at higher drying time was a result of the loss of TOPO from the inner surfaces of the membranes; 90 min drying duration was quite high and TOPO was dragged out from the membranes when DCM was evaporated through the inner membrane surface. This was corroborated by the large deposits of TOPO in the lumen side as well as on the tubing used to pump air into the lumen side. As drying the membranes for 90 min at Reair = 4.6 resulted in the loss of TOPO from the EIHFMs, drying durations up to 60 min were considered in subsequent experiments conducted at higher air flow rates.

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737

600

0.7 0.6

500

0.5

400

0.4 300

0.3

200

0.2

100

0.1 0

0 4.6, 30

4.6, 60 Reair , drying time (min)

(b) 900 Weight of TOPO (mg)

Distribution Consistency

0.8

Weight Distribution

4.6, 90

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Weight

800

Distribution

700 600 500 400 300 200 100 0 9.2, 15

9.2, 30

Distribution consistency

Weight of TOPO (mg)

(a) 700

9.2, 60

Reair , drying time (min)

(c) 600

0.25 Distribution

500

0.2

400 0.15 300 0.1 200 0.05

100 0

Distribution Consistency

Weight of TOPO (mg)

Weight

0 18.4, 15

18.4, 30

18.4, 60

Reair , drying time (min) Fig. 4. Changes in the amount and DC of the impregnated TOPO in the EIHFMs withvariations in drying duration at air flow rates (Reair) of: (a) 4.6 (b)9.2, and; (c) 18.4.

A similar trend in the variation of TOPO weight and DC was observed in the EIHFMs when air was pumped into the lumen side of the EIHFM module at a higher Reair of 9.2. The maximum weight gain and DC of 730 mg and 92%, respectively, were observed when the drying time was 30 min. The DC values were greater than 80% for drying durations of 15 and 30 min but it decreased to 35% at higher drying time of 60 min. These changes in the TOPO weight and DC could be explained based on the removal of TOPO from the membranes when the drying time was higher. Since this loss in the TOPO weight at Reair = 9.2 was observed at a lower drying time as compared to that in case of Reair = 4.6, it could be inferred that the drying process was influenced by both the air flow rate, as well as the flow duration. Therefore, the drying duration required was lower when air was pumped into the module at higher flow rate. It was also observed that the amount of TOPO impregnated in the EIHFMs did not necessarily increase with increasing flow rate for a given drying duration. For example, the weight of TOPO impregnated increased from 500 to 730 mg when the flow rate was

731

increased from Reair = 4.6–9.2, at a drying time of 30 min, but the weight decreased from 610 to 580 mg when the flow rate was increased at the same drying time of 60 min. Contrary to the results and trend obtained at Reair of 4.6 and 9.2, poor impregnation of TOPO was obtained in the EIHFMs when Reair was increased further to 18.4. At this flow rate, the amount of TOPO impregnated in the EIHFMs was independent of the drying time and the DC values were very low. Large quantities of TOPO were observed in the lumen side and in the tubing used for gas supply, indicating that TOPO had been dragged out of the membranes under these conditions. Higher air flow rate above Reair = 9.2 was unsuitable for EIHFM preparation. Based on these results, it was concluded that air flow at Reair = 9.2 at drying durations of 15–30 min were the most suitable conditions to obtain optimum TOPO impregnation with consistently uniform distribution. SEM analysis of the cross-sections, internal and external surfaces of the EIHFMs prepared under these conditions are shown in Fig. 5. It can be seen that these EIHFMs exhibited uniform TOPO distribution, within the membrane walls as well as along the membrane length. Consequently, in all the subsequent experiments, EIHFMs were prepared under these conditions by drying at Reair = 9.2 for 30 min. To examine the structural integrity of the EIHFMs after TOPO impregnation, the LEPw was determined for the pristine and impregnated membranes. The LEPw for the pristine membranes was found to be 3.5 bar whereas that for the EIHFMs was 1.5 bar. The changes in LEPw are usually a result of changes in membrane hydrophobicity or in membrane pore size [20]. Since TOPO is hydrophobic, the presence of TOPO in the membrane walls was not likely to compromise the hydrophobicity of the polypropylene membranes. Therefore, the decrease in LEPw was a result of widening of the EIHFM pores when they were subjected to stretching during the drying process, when a large volume of solid TOPO was trapped within the membrane walls. This was evinced by surface analysis of the EIHFMs using SEM at high magnification (x1400), which showed the EIHFM pore size to be about 0.5 lm, while the pore size of the pristine fibers was in the range of 0.2– 0.3 lm. In order to confirm these results further, gas permeability and mercury porosimetry tests were performed. Table 2 lists the results of these experiments for both the pristine membranes and the EIHFMs. It can be seen that the results obtained through these techniques are consistent and they indicate significant changes in the membrane properties after TOPO impregnation. The pore size of the membrane increased from 0.2–0.3 lm (pristine) to 0.49– 0.51 lm, whereas membrane porosity decreased from 42 to 9%. These results indicate that TOPO impregnation resulted in the blockage of about 80% of the membranes pores, whereas the remaining pores experienced significant stretching and expanded. TOPO impregnation in the membrane also resulted in about 10fold increase in tortuosity from 3.5 to 33. It is important to note that the values of pore size, porosity and tortuosity of the untreated membranes, obtained through these tests, were comparable to those provided by the manufacturer, confirming the reliability of our results. The drastic decrease in the porosity of the EIHFMs as compared to the untreated membranes was expected. Since there were no appreciable changes in the dimensions of the membranes upon impregnation, the deposition of TOPO was mostly within the membrane pores. Given that TOPO is a low density solid and a large amount of TOPO was impregnated into the membranes under the optimized conditions, it was anticipated that significant blocking of the membrane pores by the solid TOPO would occur, leading to decreased porosity of the EIHFMs. This porosity decrease would mean that the path going from one side of the membranes to the other side would become more tortuous because of the limitation

732

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 5. SEM images of the outer surface, cross-section and inner surface of pristine polypropylene fibers (a, c and e); and the optimized EIHFMs (b, d and f). Upon impregnation, changes in membrane morphology and uniform white deposits of TOPO on membranes can be visualized.

Table 2 Gas permeation and mercury porosimeter tests. Membrane type Provided by manufacturer

Pristine membrane

EIHFMs

Lp

0.2 0.5 3 0.27 3330

NA NA NA 0.49 52

s

0.21 0.42 3.54

0.51 0.09 33

rp (lm) e

s

Gas Permeation Mercury Porosimeter

rp (lm) e (m 1)

rp (lm) e

of free space inside the EIHFMs. The reduction in porosity and the increase in tortuosity were also reflected in the effective porosity value obtained through gas permeability test, which decreased by 98%, from 3333 to 52 m 1, after impregnation. 3.2. Simultaneous extraction and stripping of phenol 3.2.1. Effects of flow orientation The EIHFMs were used to demonstrate the simultaneous extraction and stripping of phenol in a contactor module. The experimental setup was first operated with the feed solution circulating in the shell side, whereas the stripping solution was circulated in the

lumen side. Fig. 6a shows temporal phenol concentration profiles in the feed solution, stripping solution and the EIHFMs, during simultaneous extraction and stripping of 200 mg/L phenol in the EIHFM module. When the respective solutions were pumped into the contactor, phenol concentration in the feed solution decreased rapidly from the initial level of 215 mg/l, resulting in accumulation of phenol inside the EIHFMs. A gradual increase in phenol concentration was observed in the stripping solution. Phenol accumulation in the EIHFMs was highest after 3 h of equilibration when about 8.1 mg/g of phenol was sequestered within the solid phase, while the amount of phenol measured in the feed and stripping solution at this stage were 56 and 198 mg/l, respectively. Phenol concentration in the feed solution reached a stable value after about 10 h of operation when the amount of phenol present was 28 mg/l. At the end of operation, about 311 mg/l phenol had been recovered in the stripping solution and the remaining 4.2 mg/g of phenol was in the solid phase. Thus, about 87% of phenol was extracted from the feed solution, whereas 83% of the total phenol was recovered using sodium hydroxide with a concentration factor of 1.45. While the results in Fig. 6a manifest the potential application of the EIHFMs in simultaneous extraction and stripping of phenol, the mass transfer rates, as well as phenol recovery, were rather low. This was in contrast with the excellent performance of the EIHFMs

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737

9 300

8

250

7 6

200

5 150

4 Feed

100

3

Stripping

2

EIHFM

50

Phenol in EIHFM (mg/g)

Aq. phenol conc. (mg/l)

(a)

1 0

0 0

2

4

6

8

10

Aq. phenol conc. (mg/l)

(b)

400

16

350

14

300

12 Feed

250

10

Stripping 200

8

EIHFM

150

6

100

4

50

2

0

Phenol in EIHFM (mg/g)

Time (hr)

0 0

2

4

6

8

10

Time (hr) Fig. 6. Simultaneous extraction and stripping of 200 mg/L phenol with feed solution circulating in the: (a) shell side, and; (b) lumen side.

demonstrated when extraction and stripping were carried out in two separate steps [14]. This lower efficacy could be attributed to a number of reasons: (a) low packing density and low mass transfer area; (b) relatively poor performance of the EIHFM contactor due to channeling resulting in a poor contact between the feed solution and the EIHFMs, and/or; flow orientation of the solutions fed to the contactor module [10]. In order to further investigate the performance of the EIHFM module in extractive recovery, extraction/stripping of 200 mg/L phenol was again carried out in the module by changing the flow orientation so that the feed solution was pumped into the lumen side, while the stripping solution was circulated in the shell side of the module. Fig. 6b shows the changes in the amount of phenol in the three phases with time. In this orientation, phenol removal from the feed solution was much faster, and the amount of phenol in the feed solution decreased from 211 to 77 mg/l within 30 min of operation. At this stage, phenol concentration in the EIHFMs was highest at 13.4 mg/g, whereas 68 mg/l phenol had been recovered in the stripping solution. It was also observed that the extraction of phenol from the feed solution to the EIHFMs was faster than the stripping of phenol from the EIHFMs to the stripping solution. For example, more than 85% phenol had been extracted from the feed solution within 2 h of equilibration, but only 52% phenol had been recovered in the stripping solution at this stage. By continuing the separation further, about 99% phenol could be removed from the feed solution after 7 h of operation, while 91% phenol was recovered in the stripping solution. The operation was continued for 10 h and in the end, the concentration of phenol in the feed

733

solution and stripping solution were 2.3 and 378 mg/L, respectively, whereas that in the EIHFMs was 2.7 mg/g. Thus the phenol recovery was more than 91% with a separation factor of 1.79. The extraction/stripping performance of the EIHFM module was much better when the feed solution was passed through the lumen side than that when the feed solution was pumped through the shell side. This difference could be explained based on channeling when the wastewater was pumped into the shell side. Since there was poor contact between the wastewater and the EIHFMs, phenol absorption into the EIHFMs was slow. The slow mass transfer rate during extraction also resulted in a slow accumulation of phenol in the EIHFMs and low concentration gradient for mass transfer during stripping. On the contrary, good contact was achieved between the wastewater and the EIHFMs when the wastewater was pumped into the lumen side. Consequently, phenol was rapidly absorbed into the EIHFMs, generating a relatively large concentration gradient for the mass transfer of phenol from the EIHFMs to the stripping solution. Hence, the extraction/stripping performance of the EIHFM contactor depended on the distribution of the fibers on the shell side. Since there is no organic solvent to wet the fibers, it would be better if the fibers are spaced properly, or modules with low packing density are prepared. 3.2.2. Effects of phenol concentration Having successfully demonstrated the effects of flow orientation on EIHFM performance in simultaneous extraction and stripping, experiments were performed to demonstrate the performance of the EIHFM module at higher concentrations of phenol. Fig. 7a shows recovery of 400 mg/L phenol through the EIHFMs, with the wastewater circulating in the lumen side. The results showed trends similar to those observed in the treatment of 200 mg/L phenol. The amount of phenol in the feed solution dropped from 425 to 136 mg/l after 45 min of equilibration, whereas the corresponding increase in the phenol content of the EIHFMs and the stripping solution were 29.5 mg/g and 142 mg/l, respectively. Phenol concentration in the wastewater fell by about 98% to reach below 14 mg/L after 9 h of operation. On the other hand, phenol concentration in the stripping solution rose to 683 mg/L, which translated to about 80% recovery with a concentration factor of 1.61. About 9.5 mg/g phenol remained in the EIHFMs at the end of operation. Similar removal and recovery trends were observed when the feed phenol concentration was increased further to 600– 1000 mg/l, although changes in phenol concentrations affected EIHFM performance. Fig. 7b summarizes the effects of feed phenol concentration of the simultaneous extraction and stripping by the EIHFMs. Both phenol removal from feed solution to the EIHFMs, and phenol recovery from the EIHFMs to the stripping solution decreased with increasing feed phenol concentration. It was observed that the extraction of phenol into the EIHFMs was quite fast, and the rate of extraction was not affected significantly by the rise in phenol concentration. The extraction efficiency varied in the range of 94–99% when initial phenol concentration was varied between 200 and 1000 mg/L. However, the rate of stripping from the EIHFMs was relatively slow and it did not increase proportionately with increasing phenol concentration. In fact, phenol recovery decreased from 91 to 70%, when feed phenol concentration was increased from 200 to 1000 mg/l. The slow release of phenol from the EIHFMs was a consequence of the high mass transfer resistance of the EIHFMs. As shown in Table 2, impregnation of TOPO in the EIHFMs resulted in a dramatic change in membrane properties of porosity and tortuosity. The significant decrease in membrane porosity and a large increase in membrane tortuosity were highly unfavorable for phenol diffusion through the membrane and phenol diffusion through the EIHFMs was slow [26]. When phenol concentration in the feed solution was increased, a

734

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737

Feed

Stripping

35

EIHFM

700

30

600

25

500 20 400 15 300 10

200

Phenol in EIHFM (mg/g)

Phenol in aquesous phase (mg/l)

(a) 800

5

100 0

0 0

2

4

6

8

10

Time (hr)

(b) 120

2 Removal

Recovery

CF

1.8 1.6 1.4

80

1.2 60

1

CF

Removal/Recovery (%)

100

0.8 40

0.6 0.4

20

0.2 0

0 200

400

600

800

1000

Phenol Concentration (mg/l) Fig. 7. Effects of feed phenol concentration on EIHFM performance during simultaneous extraction and stripping: (a) performance with 400 mg/L phenol, and (b) performances with different concentrations.

larger amount of phenol had to diffuse through the same number of membranes at a fixed area; hence, a longer period was required for the diffusion which resulted in a comparatively slower recovery. It should also be noted that the experiments were carried out for a maximum of 10 h, which could have contributed to relatively low recovery of phenol at higher concentrations. During extractive recovery of 200–1000 mg/l phenol through the EIHFMs, the concentration factors decreased at increasing phenol concentration between 1.4 and 1.8. Since these concentration factors were inextricably linked with phenol recovery, this change in concentration factors was also a result of the slow diffusion of phenol through the EIHFMs, which was responsible for the low recovery of phenol at higher concentrations. The concentration dependent recovery of phenol could also be attributed to the distribution coefficient of phenol in TOPO-based two-phase processes,

wherein the distribution coefficient decreased with rising concentration of feed phenol [15,27]. 3.2.3. Effects of stripping solution concentration Since mass transfer through the EIHFMS was not only affected by feed phenol concentration, but also by concentration of the stripping agent, experiments were performed to assess the effects of 0.2–1 M NaOH(as stripping agent) on EIHFM performance at initial phenol concentrations of 800 and 1000 mg/L. Fig. 8 and Table 3 summarize the effects of NaOH concentration on phenol extraction/stripping by EIHFMs. It can be seen that there was no difference in phenol concentrations in the feed or stripping solutions under different NaOH concentration during the first 2 h of extraction, which was dominated mainly by the movement of phenol from the feed solution to the EIHFM. Changes in phenol

735

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737

70

1400

60

1200

50

1000

40

800 30

600

20

400

3.2.4. Stability EIHFM design was based on impregnation of solid TOPO in supporting membranes. In such extractant-impregnated systems, the stability of impregnation is crucial to achieve sustainability over long-term operation [14]. Moreover, there may be batch-to-batch variations in impregnation efficiency, which may affect the mass transfer performance. In order to ascertain that the methodology developed in this research for TOPO impregnation in polypropylene membrane was reliable and the results were reproducible, a new EIHFM module was fabricated under optimized conditions, and simultaneous extraction/stripping were carried out at feed phenol concentrations of 200 and 400 mg/l, with stripping solution comprising of 0.2 M NaOH. Fig. 9 shows the extraction efficiency of the EIHFMs during six consecutive runs, the first three were carried out at feed phenol concentration of 200 mg/l, whereas the next three were performed at a higher feed phenol concentration of 400 mg/l. It can be seen that the extraction performance was consistent and reproducible and the extraction efficiency at 200 and 400 mg/l phenol were obtained as 99.39 ± 0.27% and 94.61 ± 0.45%, respectively. These results were comparable to those achieved earlier under similar experimental conditions (Fig. 7). Thus, these results indicated that the EIHFMs-based extractive separation was reproducible, with little batch-to-batch variations, and no extractant loss over time. While the results obtained in this study are consistent with previous research on EIHFMs [14–17], this study is different from the previous studies in several ways: (1) the methodology used in TOPO impregnation has been improved to facilitate uniform TOPO distribution; (2) the uniformly impregnated EIHFMs allowed phenol diffusion from one side to another side of the membranes, and supported simultaneous extraction and stripping, and; (3) it

10

200

0.2 M

0.5 M

1M

0

0 0

2

4

6 Time (hr)

8

10

12

(b) 1800 Phenol in stripping phase (mg/l)

noted that phenol recovered through alkaline treatment is in the phenolate form; additional neutralization is needed to recover the phenol in this process.

Phenol in EIHFM (mg/g)

1600

80

1600

70

1400

60

1200

50

1000 40 800 30

600

20

400 200

0.2 M

0.5 M

Phenol in EIHFM (mg/g)

Phenol in stripping phase (mg/l)

(a)

10

1M

0

0 0

2

4

6 Time (hr)

8

10

12

Fig. 8. Phenol concentration profiles in membrane and stripping solution for 0.2, 0.5 and 1 M NaOH with initial concentration (a) 800 mg/l and (b) 1000 mg/l. Solid lines indicate concentration in stripping phase. Dotted lines indicate concentration in EIHFM.

concentration profiles in the feed and stripping solution at different NaOH concentration appeared after 2 h, and rate of phenol stripping from the EIHFMs improved with increasing NaOH concentration. At 800 mg/L feed phenol concentration, phenol accumulation in the EIHFMs dropped by 28% when NaOH concentration was increased from 0.2 to 0.5 M; the accumulation dropped further by 18%, when NaOH concentration was increased from 0.5 to 1 M NaOH. Therefore, increasing NaOH concentration from 0.2 M to 1 M resulted in a total drop of 42% in phenol accumulation in the EIHFMs. The concentration factor (CF) obtained using 1 M NaOH was 1.7 as compared to 1.45 using 0.2 M NaOH, which translated to an increase of 18%. Similar trends were observed for 1000 mg/l phenol. There was 22% increase in CF from 1.4 at 0.2 M NaOH to 1.69 at 1 M NaOH. On the other hand, phenol accumulation in the EIHFMs dropped by 46% when NaOH concentration was increased from 0.2 M to 1 M. these results are consistent with literature, where an NaOH to phenol molar ratio of 10 or above is reported for maximizing stripping efficiency [9], especially at high phenol concentrations above 800 mg/l. It should be

Fig. 9. EIHFM stability during consequent operation at phenol concentrations of 200 and 400 mg/L in a different membrane module.

Table 3 Effects of stripping solution strength on EIHFM performance. Feed phenol conc.

800 mg/l

1000 mg/l

Performance

Removal (%) Recovery (%) CF Removal (%) Recovery (%) CF

NaOH conc. in stripping solution 0.2 M

0.5 M

1M

96 ± 0.298 72 ± 0.510 1.45 ± 0.01 94 ± 0.386 69 ± 1.247 1.38 ± 0.025

97.4 ± 0.673 81.4 ± 1.23 1.63 ± 0.033 95.5 ± 1.01 80 ± 0.724 1.60 ± 0.015

98 ± 0.315 85 ± 1.468 1.7 ± 0.062 96 ± 0.443 83 ± 1.118 1.69 ± 0.054

736

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737

resulted in a configuration similar to that of SLM, which can be operated under continuous operation. These advantages compensate for the relatively low mass transfer rate of the uniformly impregnated EIHFMs (arising from low porosity and high tortuosity), as compared to previously used EIHFMs. Based on these results, the EIHFMs can be used as an efficient, stable and sustainable technology, with little solvent requirement, in the extraction of aromatic compounds from wastewater. Although the release of phenol from the EIHFMs during simultaneous extraction and stripping was slower as compared to that in separate extraction and stripping process [14], the results presented in Figs. 6 and 7 proved conclusively that the application of EIHFMs cannot be limited merely to sorption [15] and these membranes can be used to perform simultaneous extraction and stripping. It should also be noted that the membrane contactors used in this study had a relatively low mass transfer area of 0.012 m2, when compared against the membrane contactors typically used in membrane-based solvent extraction [9,28–30]. Despite that, the EIHFM-based solventless contact between the feed and stripping solution was able to extract about 73% phenol from 800 mg/l feed solution in 1 h, resulting in an average mass flux of 5.25 g/m2h. These results are a clear demonstration of the excellent mass transfer properties of the EIHFMs. It should also be noted that the performance of the EIHFM-based separation can be easily improved by increasing the number of EIHFMs in the membrane contactor. Since the EIHFMs provided both, the mass transfer area and the hydrophobic phase during extraction/ stripping, it is anticipated that increase in the membrane area would lead to a significant increase in the performance of the EIHFM contactor. Additionally, the modular configuration of EIHFM-based separation allows connecting multiples membrane contactors in series or parallel to improve performance or capacity of the process. In the past few decades, several configurations have been suggested to address the instability concerns of SLMs in membranebased solvent extraction, these included hollow fiber contained liquid membrane, hollow fiber renewal liquid membranes and pseudo-emulsion-based hollow fiber strip dispersion [7,10,31]. Although these new configurations were successful in improving SLM stability, some of the excellent advantages of the SLM such as the dispersion-free diffusion or the simplicity of design had to be compromised. On the contrary, EIHFM based extraction/stripping of phenol retained most of the advantages of SLM-based extraction, and the stability of TOPO impregnation in the EIHFMs have already been demonstrated under various operating conditions [14,15]. In this study, the experiments were carried out in triplets under identical conditions using the same EIHFM contactors. The results indicated that the batch-to-batch variations in the EIHFM contactor performance were negligible, as evident from the error bars in Figs. 6 and 7. Therefore, these results also signify the stability of the EIHFMs, and indicate that the EIHFMs can be an effective and stable alternative to the SLM-based separation processes. Apart from the excellent mass transfer performance and prolonged stability, EIHFM-based extractive separation also has the unique advantage of the process being ‘‘solventless”. While the amount of the carrier solvent and the extractant required during impregnation was quite low (5 ml), no solvent was required to sustain the extraction/stripping process. This ‘‘solventless” nature of EIHFM-based separation makes the process more economical by reducing solvent costs, whereas the absence of organic solvents also reduces the risks of any secondary waste generation and improve the sustainability of solvent extraction [16]. Since TOPO is a popular extractant in the separation of several organics and metals [32–34], the EIHFM contactor can be used in the separation of several other pollutants.

4. Conclusions EIHFMs with high TOPO density and uniform TOPO distribution have been prepared and the preparation process has been optimized. The EIHFMs were characterized; their pore size, porosity and tortuosity were compared with those of the pristine membranes. Simultaneous extraction and stripping of phenol has been performed using the EIHFMs prepared by the optimized technique. The EIHFMs showed excellent performance in phenol extraction and recovery, and offered the advantages of ‘solventless’ system. The results indicated that the EIHFMs can be an effective, stable and sustainable alternative to membrane-based non-dispersive solvent extraction.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cej.2016.09.105.

References [1] G. Collins, C. Foy, S. McHugh, T. Mahony, V. O’Flaherty, Anaerobic biological treatment of phenolic wastewater at 15–18°C, Water Res. 39 (2005) 1614– 1620. [2] C.D. Becker, D.W. Crass, Examination of procedures for acute toxicity tests with the fathead minnow and coal synfuel blends, Arch. Environ. Contam. Toxicol. 11 (1982) 33–40. [3] N.C. Saha, F. Bhunia, A. Kaviraj, Toxicity of phenol to fish and aquatic ecosystems, Bull. Environ. Contam. Toxicol. 63 (1999) 195–202. [4] D. Raghu, H. Hsieh, Considerations in disposal of phenolic waters, Int. J. Environ. Stud. 30 (1987) 277–285. [5] G. Busca, S. Berardinelli, C. Resini, L. Arrighi, Technologies for the removal of phenol from fluid streams: A short review of recent developments, J. Hazard. Mater. 160 (2008) 265–288. [6] M.C. Figueroa, M.E. Weber, Aqueous micellar solvent extraction of phenol from wastewater, Sep. Sci. Technol. 40 (2005) 1653–1672. [7] Z. Lazarova, S. Boyadzhieva, Treatment of phenol-containing aqueous solutions by membrane-based solvent extraction in coupled ultrafiltration modules, Chem. Eng. J. 100 (2004) 129–138. [8] C.M. Santana, Z.S. Ferrera, M.E.T. Padron, J.J.S. Rodriguez, Methodologies for the extraction of phenolic compunds from environmental samples: New approaches, Molecules 14 (2009) 298–320. [9] M.T.A. Reis, O.M.F. de Freitas, M.R.C. Ismael, J.M.R. Carvalho, Recovery of phenol from aqueous solutions using liquid membranes with Cyanex 923, J. Membr. Sci. 305 (2007) 313–324. [10] A. Hasanoglu, Removal of phenol from wastewaters using membrane contactors: Comparative experimental analysis of emulsion pertraction, Desalination 309 (2013) 171–180. [11] N.M. Kocherginsky, Y. Qian, L. Seelam, Recent advances in supported liquid membrane technology, Sep. Purif. Technol. 53 (2007) 171–177. [12] X.C. Gong, G.S. Luo, W.W. Yang, F.Y. Wu, Separation of organic acids by newly developed polysulfone microcapsules containing trioctylamine, Sep. Purif. Technol. 48 (2006) 235–243. [13] J. Serasols, J. Poch, I. Villaescusa, Expansion of adsorption isotherms into equilibrium surface: Case I: solvent impregnated resins (SIR), React. Funct. Polym. 48 (2001) 37–51. [14] P. Praveen, K.-C. Loh, Two-phase biodegradation of phenol in trioctylphosphine oxide impregnated hollow fiber membrane bioreactor, Biochem. Eng. J. 79 (2013) 274–282. [15] P. Praveen, K.-C. Loh, Trioctylphosphine oxide-impregnated hollow fiber membranes for removal of phenol from wastewater, J. Membr. Sci. 437 (2013) 1–6. [16] P. Praveen, K.-C. Loh, Solid/liquid extraction equilibria of phenolic compounds with trioctylphosphine oxide impregnated in polymeric membranes, Chemosphere 153 (2016) 405–413. [17] P. Praveen, K.-C. Loh, Solventless extraction/stripping of phenol using trioctylphosphine oxide impregnated hollow fiber membranes – Experimental and modeling analysis, Chem. Eng. J. 255 (2014) 641–649. [18] W. Cichy, J. Szymanowski, Recovery of phenol from aqueous streams in hollow fiber modules, Environ. Sci Technol. 36 (2002) 2088–2093. [19] K.C. Loh, T.S. Chung, W.F. Ang, Immobilized-cell membrane bioreactor for high-strength phenol wastewater, J. Environ. Eng. 126 (2000) 75–79. [20] B.S. Lalia, E. Guillen, H.A. Arafat, R. Hashaikeh, Nanocrystalline cellulose reinforced PVDF-HFP membranes for membrane distillation application, Desalination 332 (2014) 134–141. [21] D. Wang, K. Li, W.K. Teo, Effects of temperature and pressure on gas permselection properties in asymmetric membranes, J. Membr. Sci. 105 (1995) 89–101.

K. Das et al. / Chemical Engineering Journal 308 (2017) 727–737 [22] K. Li, J.F. Kong, D. Wang, W.K. Teo, Tailor-made asymmetric PVDF hollow fibers for soluble gas removal, AICHE J. 45 (1999) 1211–1219. [23] M. Khayet, C.Y. Feng, K.C. Khulbe, T. Matsuura, Preparation and characterization of polyvinylidene fluoride hollow fiber membranes for ultrafiltration, Polymer 43 (2002) 3879–3890. [24] L. Moscou, S. Lub, Practical use of mercury porosimetry in the study of porous solids, Powder Technol. 29 (1981) 45–52. [25] S.G. Kandlikar, S. Joshi, S. Tian, Effect of channel roughness on heat transfer and fluid flow charaacteristics at low reynolds number in small diameter tubes, in: 35th Nat. Heat Tr. Conf., Anaheim, California, 2001. [26] A. Gabelman, S.T. Hwang, Hollow fiber membrane contactors, J. Membr. Sci. 159 (1999) 61–106. [27] S. Lux, M. Siebenhofer, Investigation of liquid–liquid phase equilibria for reactive extraction of lactic acid with organophosphorus solvents, J. Chem. Technol. Biotechnol. 88 (2013) 462–467. [28] C.J. Tompkins, A.S. Michaels, S.W. Peretti, Removal of p-nitrophenol from aqueous solution by membrane-supported solvent extraction, J. Membr. Sci. 75 (1992) 277–292.

737

[29] M.J. Gonzalez-Munoz, S. Luque, J.R. Alvarez, J. Coca, Recovery of phenol from aqueous solutions using hollow fiber contactors, J. Membr. Sci. 213 (2003) 181–193. [30] S. Bocquet, A. Torres, J. Sanchez, G.M. Rios, J. Romero, Modeling the mass transfer in solvent-extraction processes with hollow-fiber membranes, Fluid Mech. Transp. Phenom. 51 (2005) 1067–1079. [31] P. Praveen, K.-C. Loh, Simultaneous extraction and biodegradation of phenol in a hollow fiber supported liquid membrane bioreactor, J. Membr. Sci. 430 (2013) 242–251. [32] M. Matsumoto, Extraction kinetics of organic acids with tri-n-octylphosphine oxide, J. Chem. Technol. Biotechnol. 67 (1996) 260–264. [33] T.-C. Huang, C.-C. Huang, D.-H. Chen, Transport of Chromium (VI) through a supported liquid membrane containg Tri-octylphosphine Oxide, Sep. Purif. Technol. 33 (1998) 1919–1935. [34] A.K. Pabby, R. Haddad, F.J. Alguacil, A.M. Sastre, Improved kinetics-based gold cyanide extraction with mixture of LIX79 + TOPO utilizing hollow fiber membrane contactors, Chem. Eng. J. 100 (2004) 11–22.