Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique

G Model

JTICE-849; No. of Pages 11 Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of the Taiwan Institute of Chemical Engineers journal homepage: www.elsevier.com/locate/jtice

Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique M. Madhumala a, D. Madhavi a, T. Sankarshana b, S. Sridhar a,* a b

Membrane Separations Group, Chemical Engineering Division, Indian Institute of Chemical Technology, Hyderabad 500 007, India University College of Technology, Osmania University, Hyderabad 500 007, India

A R T I C L E I N F O

A B S T R A C T

Article history: Received 19 July 2013 Received in revised form 29 January 2014 Accepted 15 February 2014 Available online xxx

Membrane distillation (MD) is a rapidly advancing process for separation of azeotropic and close boiling liquid mixtures besides dewatering of high boiling solvents. Recovery of hydrochloric acid (HCl) from a chloralkali industrial effluent and dehydration of glycerol/water mixture was performed using MD technique. HCl was recovered using chemically resistant polytetrafluoroethylene (PTFE) membrane of 0.22 mm pore size and 78% porosity. The effluent feed contained 32.8 wt.% of aqueous HCl with color forming Fe compounds and heavy hydrocarbon (C9–C14) impurities that gave an oily appearance. Permeate obtained was colorless aqueous HCl (33 wt.%) at a high flux with negligible impurity levels. An increase in permeate pressure from 5 to 15 mmHg resulted in a gradual reduction in flux from a high value of 8.57 kg/m2/h to a moderate 1.02 kg/m2/h at ambient temperature of 28 8C. Effect of feed composition in terms of acid and inorganic salt contents besides feed temperature (25–60 8C) on flux and separation efficiency was demonstrated at a constant downstream vacuum of 8–10 mmHg. Dehydration of glycerol was performed using novel indigenously synthesized ultraporous hydrophobic polystyrene (PS) membrane of 0.72 mm pore size. Permeate was found to contain pure water due to the low vapor pressure and larger molecular size of glycerol which cannot penetrate PS as it does not get wetted by water. This indicated a selectivity of infinity (1) which is associated with a reasonable water flux in the range 0.56–0.02 kg/m2/h at a vacuum of 5 mmHg for feed glycerol concentration varying from 10 to 90 wt.%. PS membrane was characterized by SEM, FTIR, XRD and TGA to assess surface and crosssectional morphologies, structural elucidation, crystallinity and thermal stability of the membrane, respectively. A detailed economic estimation of HCl recovery for a feed effluent capacity of 2 m3/h is presented. The study showed that commercial grade HCl and glycerol could be recovered from aqueous streams at a reasonable price by employing MD technique. ß 2014 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: Membrane distillation Hydrochloric acid Glycerol Hydrophobic polystyrene membrane Chloralkali industry Economic estimation

1. Introduction The separation of useful chemical entities such as solvents and acids from aqueous waste streams released from industries is important from both economic and environmental points of view [1–8]. Glycerol is a nontoxic, edible, bio-sustainable and biodegradable compound used as a versatile industrial organic solvent and more recently as a reactant for producing several important chemicals. It has a high boiling point of 290 8C and decomposition temperature. The hygroscopic nature of glycerol results in water absorption and formation of an aqueous solution of 80% glycerol

* Corresponding author. Tel.: +91 40 27193408; fax: +91 40 27193626. E-mail address: [email protected] (S. Sridhar).

which requires to be dehydrated before reuse [9]. With rapid growth in both population and industrialization, there is an urgent need to utilize biofuels such as biodiesel as alternative sources of energy since they are renewable, biodegradable and clean. Due to the large surplus of glycerol formed as a by-product during the production of biodiesel, many studies are focused on finding new applications of glycerol as a low cost feedstock for converting it into value-added chemicals. Glycerol is used in food and beverages, drugs, cosmetics, surface coating resins, textiles industries and in production of acrolein [10–14]. Several conventional methods, for instance, adsorption onto activated carbon [15,16], multiple-effect evaporation [17] and distillation [1,6,18–27] have been in use for recovery of solvents such as glycerol since past few decades. The technologies evaluated so far involve high capital and operating costs due to regeneration procedures and phase change.

http://dx.doi.org/10.1016/j.jtice.2014.02.010 1876-1070/ß 2014 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 2

M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

Inorganic acids such as HCl have manifold uses in a variety of industries including metal and woodwork, textile dyeing, petroleum, explosive and photography besides being employed as catalysts in several important chemical reactions. Several approaches have been proposed for the recovery of acids from effluents which includes ion retardation, diffusion dialysis, and electrodialysis [28–32]. Nevertheless, these regeneration processes produce lots of dilute acid solution. Thus, a novel efficient technology needs to be proposed to concentrate the acid solution. HCl is a strong inorganic acid having several applications in chemical and electroplating industries [5,33,34]. Aqueous HCl is used for industrial acidizing, refining ores of tin and tantalum, converting corn starch to syrup, etc. Since the past few decades, various techniques have been in use for separation of acids and solvents from aqueous solutions. In comparison to these processes, MD is an emerging technology due to inherent features of process safety and environmental friendliness for radioactive waste treatment, removal of volatile organic compounds, concentration of agro-based and organic solutions and other specific applications [34,35]. The key advantages of MD over conventional separation processes are relatively lower energy consumption compared to distillation, considerable separation of dissolved and non-volatile species, reduced vapor space as compared to conventional multistage flash distillation, lower operating pressure than pressuredriven membrane processes, lower operating temperature and corrosion related problems as compared to evaporation [36]. The large vapor space required by a conventional distillation column is replaced in MD by the pore volume of a microporous membrane, which is generally around 100 mm thick. While conventional distillation relies on extensive vapor–liquid contact, MD employs a hydrophobic microporous membrane to support a vapor–liquid interface. Membrane fouling is a smaller problem in MD than in other membrane processes because the pores are relatively large compared to the ‘pores’ or diffusional pathways in nanofiltration or ultrafiltration, and do not get easily clogged since vacuum is applied at downstream side in MD rather than pressure at the upstream surface [37]. Vacuum MD (VMD) uses porous hydrophobic membranes that act only as support for the vapor–liquid interface and do not contribute to the separation performance. On the contrary, pervaporation requires dense and selective membranes and the separation is based on the relative solubility and diffusivity of each component in the membrane material. Therefore, VMD typically achieves fluxes that are several orders of magnitude higher than pervaporation mass transfer rates even though selectivity in pervaporation is considerably higher than in MD for close boiling mixtures. This phenomenon is attributed to the fact that in VMD the selectivity is mainly determined by temperature gradient and vapor–liquid equilibrium (VLE) conditions at the membranesolution interface, although the diffusion step across the porous membrane may impart some assistance to the transport of lighter molecules [38]. MD is a thermally driven separation process, wherein a microporous hydrophobic membrane usually separates aqueous solutions maintained at different temperatures. The temperature difference across the membrane results in vapor pressure gradient, causing transfer of water vapor molecules though the pores of the membrane from the high to low vapor pressure side. Several authors have studied recovery of hydrochloric acid from aqueous streams using MD. Gryta et al. (2006) investigated the performance of MD process by treating effluents containing residual HCl along with soluble salts [39]. A significant enhancement in flux was observed by increasing acid concentration in the feed mixture. Another study was carried out by Tomaszewska et al. (2001) for concentration of acidic spent solutions containing HCl and salts using capillary PP membrane module [40]. The experiments were performed using model or real metal pickling

solutions of different compositions. The application of concentration or separation of HCl from spent acid solutions may have a limited practical significance therefore the studies on the influence of salts on HCl molar flux were demonstrated [41]. The presence of salts in feed solution improved the flux of HCl to a greater extent compared to the case wherein salt was absent. The same authors also studied mass transfer of HCl and water vapor through flat sheet membranes made of hydrophobic PTFE, PVDF and PP capillary membranes [42]. A transport model was developed by considering a set of possible equations. The study showed that experimental results obtained were in good agreement with theoretical predictions. Polyvinylidenefluoride (PVDF), polytetrafluoroethylene (PTFE) and polypropylene (PP) are the most common membrane materials for MD applications due to their hydrophobic nature and high thermal stability. PTFE membranes show a better performance compared to PVDF due to better chemical inertness but are far more expensive. Thus, there is enough scope to study polystyrene membrane for dehydration of glycerol since not much work pertaining to this particular membrane has been reported in literature so far. In the present study, ultraporous hydrophobic polystyrene was indigenously synthesized and used for dehydration of glycerol from aqueous solutions. Commercial PTFE membranes were preferred for recovery of HCl from a chloralkali industrial effluent, which contained concentrated and fuming HCl, water and traces of impurities in the form of higher hydrocarbons (C9-C14) and Fe compounds which gave an oily appearance to the aqueous HCl rendering the acid unfit for utilization. 2. Theory In MD a hot aqueous feed solution is brought into contact with one side (feed side) of a hydrophobic, microporous membrane. The hydrophobic nature of the membrane prevents penetration of the aqueous solution into the pores, resulting in a vapor–liquid interface at each pore entrance. Fig. 1 illustrates how the vapor– liquid interfaces are supported at the pore openings. The value of the contact angle u of a liquid droplet on an ideal smooth homogeneous surface is described by Young’s equation [43]:

g lv cosu ¼ g sv  g sl

(1)

where u contact angle between liquid and membrane, g lv liquid– vapor interfacial tension, g sv solid–vapor interfacial tension, g sl solid–liquid interfacial tension. A droplet of water on a hydrophobic surface of a polymer film made of polymer like PS, PP or PTFE will give a contact angle which is larger than 908. If the surface

Fig. 1. Transport mechanism in MD process.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

3

Table 1 Structure, porosity and pore size of the membranes used in the present study. Membrane type

Structure

PTFE

F

F

C

C

F

F

PS

Membrane pore size (mm)

Membrane porosity (%)

0.22

78

0.72

68

n

H C

C

H

H

n

Fig. 2. Schematic of liquid droplet on solid surface.

active agents or organic materials are dissolved in water, the surface tension of the liquid will decrease. As a result, the contact angle u will decrease, and if it becomes smaller than 908 the liquid will wet the solid surface. A schematic description of contact angle between air, liquid droplet and solid surface is shown in Fig. 2. 3. Experimental 3.1. Materials and methods PTFE membrane of 0.22 mm pore size and PS pellets were procured from Millipore and Sigma–Aldrich, USA, respectively. 2Methoxyethanol, N,N-dimethylacetamide (DMAc), NaOH, potassium hydrogen phthalate and Phenolphthalein were obtained from sd Fine Chemicals Ltd. Hyd., India. Glycerol was procured from Thermo Fisher Scientific India Pvt. Ltd., Mumbai, India. Properties like structure, porosity and pore size of PTFE and PS are illustrated in Table 1. 3.2. Synthesis of hydrophobic polystyrene (PS) membrane PS membrane was prepared by phase inversion technique. The polymer solution (23% (w/v)) was prepared by dissolving 23 g of PS pellets in mixture containing 73 ml of DMAc and 4 ml of 2methoxyethanol additive at an ambient temperature. The dope was de-aerated to obtain a bubble free solution which was cast to a

desired thickness using a doctor’s blade on a non-woven polyester fabric which was further affixed onto a glass plate. After casting, the glass plate was immediately immersed in an ice-cold nonsolvent water bath for 10 min to obtain a porous membrane. 3.3. Description of experimental set-up for membrane distillation A schematic diagram of the system used to carry out the MD experiments is shown in Fig. 3(a). The MD cell consists of two bellshaped B-24 size glass column reducers/couplers clamped together with external padded flanges by means of tie rods to give a vacuum tight arrangement. The top half was used as the feed chamber and the bottom one worked as the permeate chamber. The membrane was supported on a porous stainless steel plate embedded with a stainless steel 316 mesh/screen. Teflon gaskets were fixed by means of high-vacuum silicone grease on either side of the membrane and the sandwich was placed between the two glass column couplers and secured tightly to prevent any major vacuum leak. The effective area of the membrane used was 20 cm2. After fixing the membrane, the cell was installed in the manifold and connected to the permeate line by means of a B-24 glass cone fixed on the other end to a highvacuum glass valve followed by a glass condenser trap which consisted of a small detachable collector. The trap was placed in a Dewar flask containing liquid nitrogen or dry ice–acetone mixture for condensing the permeate vapors. A 0.75 HP rotary vacuum pump was used to maintain the low permeate side pressure, which was measured using an Edward’s Mcleod gauge (Scale: 10–0.01 mmHg).

Fig. 3. (a) Schematic of MD experimental set-up. (b) Photograph of laboratory MD cell.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

The upstream side of the membrane held the feed solution at atmospheric pressure while the downstream side was connected to a vacuum pump to enhance the rate of permeation. For heating the feed to higher temperature of 60 8C, a small heating coil comprising of Nichrome wire threaded through ceramic beads was wound around the feed chamber and connected to a Dimmerstat which controlled the temperature indicated by a thermometer inserted into a glass thermowell. Glass wool was used to cover the ceramic beads to prevent heat loss. The actual photograph of MD cell is shown in Fig. 3(b). Initially, a known volume of the feed mixture was introduced at the upper part of the cell, and vacuum was applied on the downstream side. Permeate was condensed and collected during a period of 3 h in the liquid nitrogen trap, and the permeation rate was determined from the weight of the collected sample. Concentration of individual components in the feed, permeate and retentate were determined by refractive index measurements using a calibration curve plotted from known compositions of the liquid mixture. The effect of operating parameters such as feed composition and temperature on selectivity and flux of the membrane was studied in detail. The concentration of glycerol in the feed mixture was varied from 10 to 90% (v/v) and that of HCl was altered from 19.8 to 34.2% (w/v). The collected permeate was weighed, after allowing it to attain room temperature in a Sartorius electronic balance (accuracy, 104 g) to determine the flux and then analyzed by refractive index and titration method to evaluate membrane selectivity. 3.4. Flux and selectivity calculations Membrane permeation flux, J (kg/m2/h), is determined as the amount of liquid transported though the membrane per unit time, per unit membrane area. J¼

W At

Yð1  XÞ Xð1  YÞ

3.6.3. Thermo gravimetric analysis (TGA) Thermal stability of the polymer films was examined using Seiko 220TG/DTA analyzer, Japan, in the temperature range 25 8C to 700 8C at a heating rate of 10 8C/min with continuous flushing using inert nitrogen gas flowing at the rate of 200 ml/min. The membrane was subjected to TGA to determine the thermal stability and polymer decomposition characteristics. 3.6.4. Scanning electron microscopy (SEM) SEM pictures of the surface and cross-section morphologies of the indigenously synthesized and commercially available membranes were scanned using a digital scanning electron microscope, JEOL JSM 5410, Tokyo, Japan. 3.7. Sorption experiments The active membrane layer was peeled off from the non-woven polyester fabric support before subjecting it to sorption experiments. A pre-weighed polymer sample was taken and soaked in pure water and organic solvent/acid solution as well as aqueous binary mixtures of different compositions and allowed to reach equilibrium over a period of 72 h at ambient temperature. The membranes were then removed from the solvent/solution mixture and quickly wiped to remove extra adhering liquid from the surface and immediately weighed to determine the amount of liquid absorbed by the film. The degree of swelling (DS) was calculated as follows: Degree of swelling ¼

(2)

where W represents the mass of permeate (kg), A is the membrane area (m2) and t represents the evaluation time (h). The selectivity (a) is defined as

a¼

is generated by a CuK-a source. The angle of diffraction is varied from 28 to 608 to identify the changes in the crystal structure and distance between the inter-segmental chains (d-spacing) in the polymer membranes.

Mass of swollen polymer ðM g Þ Mass of dry polymer ðM d Þ

(4)

The percentage sorption represents the amount of liquid absorbed by the membrane at equilibrium and is determined as: %Sorption ¼

Ms  Md  100 Md

(5)

(3)

where X and Y represent the mass fractions of the faster component which permeates preferentially though the membrane in the feed and permeate, respectively. 3.5. Analytical methods The feed and permeate samples for glycerol/water system were analyzed by a digital refractometer (Rudolph Research DSR-l) to take advantage of the large difference in the refractive indices of glycerol (1.471) and water (1.332). Concentration of acid in effluent was analyzed using standard acid–base titration method. 3.6. Membrane characterization 3.6.1. Fourier-transform infrared spectroscopy (FTIR) The FTIR spectra of membranes were studied for intermolecular interactions within the polymer matrix. The membranes were scanned in the range of 400–4000 cm1 wave numbers using a Nicolet-740, Perkin-Elmer-283B FTIR spectrophotometer (Boston, MA, USA) by KBr pellet method. 3.6.2. X-ray diffraction studies (XRD) In the present study, a Siemens D 5000 powder X-ray diffractometer (NJ, USA) is used to study the solid-state morphology of the membranes. X-ray of 1.5406 A˚ wavelengths

3.8. Calibration curve for glycerol/water system Different solutions of glycerol/water mixture were prepared by varying the concentration of glycerol from 10 to 90% by volume for analysis by digital refractometer. A standard graph is made by plotting refractive index values on Y-axis as against different known compositions of glycerol/water mixtures on X-axis. Fig. 4 represents standard RI curve for glycerol/water system. 1.48

Refractive Index

4

1.44

1.4

1.36

1.32

0

20

40

60

80

100

% Glycerol Concentration in Water (v/v) Fig. 4. Refractive index calibration curve for glycerol/water system.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

5

Fig. 5. FTIR spectra of polystyrene membrane.

out of plane CH bending vibration of hydrogen atoms in monosubstituted aromatic ring.

4. Results and discussion 4.1. Membrane characterization 4.1.1. FTIR FTIR spectrum of polystyrene is shown in Fig. 5. The most intense bands at 3058–3026 cm1 correspond to –CH stretching vibrations in the mono substituted aromatic cycle whereas bands at 2934–2850 cm1 represent CH2 asymmetric and symmetric stretching vibrations. There are aromatic C–C stretch bands representing carbon-carbon bonds in the aromatic ring at about 1600 cm1, which are strong and distinct. The other characteristic bands of polystyrene are seen at 1069 and 1027 cm1 which correspond to in-plane bending vibrations of hydrogen atoms of mono-substituted cycles whereas 755–698 cm1 are attributed to

4.1.2. XRD XRD spectrum of polystyrene is shown in Fig. 6. It can be noted that the XRD pattern of polystyrene membrane appears to be semicrystalline in nature with three distinct peaks at 148, 278 and 548 of 2u. A broad peak at 148 with the d-spacing value 7.32 A˚ represents the amorphous nature and corresponding narrow peaks at 278 and 548 with d-spacing values 3.29 A˚ and 1.70 A˚ represent crystalline regions of the polymer chains. 4.1.3. TGA studies TGA curve of polystyrene (PS) membrane is shown in Fig. 7. The thermal stability and degradation behavior of the sample was

Fig. 6. XRD spectra of polystyrene membrane.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 6

M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

Fig. 7. TGA curve of polystyrene membrane.

evaluated under an inert nitrogen atmosphere. TGA curve of the membrane shows its first weight loss stage occurring at 400 8C followed by final decomposition at 500 8C, which is due to the degradation of the main chain of PS. The study reveals that the membrane can be safely operated up to temperatures of 400 8C.

4.1.4. SEM Fig. 8(a) and (b) represents the surface and cross-sectional morphologies of PTFE membrane. The surface morphology of the PTFE (Fig. 8(a)) shows the presence of visible micropores which are distributed uniformly across the surface without any agglomerations.

Fig. 8. SEM pictures of MD membranes: (a) PTFE surface, (b) PTFE cross-section, (c) polystyrene surface and (d) polystyrene cross-section.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

12

water, % sorption value of the membrane was relatively low, due to the hydrophobic nature of the active polystyrene layer.

% Sorption

10

4.3. Effect of feed composition on glycerol dehydration

8 6 4 2 0

0

20

40 60 80 % Glycerol Concentration (v/v)

100

Fig. 9. Influence of glycerol/water feed composition on sorption behavior of polystyrene.

0.6 0.5 Flux (kgm-2h-1)

7

0.4 0.3 0.2

4.4. Recovery of hydrochloric acid from industrial effluent by MD through PTFE membrane

0.1 0

Dehydration study of glycerol/water mixtures by MD was performed using ultraporous polystyrene membrane. The effect of feed composition on membrane performance was studied by gradually increasing the glycerol concentration in the feed mixture at a constant downstream pressure of 5 mmHg and ambient feed temperature. The effect of feed composition on total flux is represented by Fig. 10. As the glycerol concentration increased from 10 to 90% in the feed, the flux decreased from 0.56 to 0.02 kg/ m2/h respectively, due to availability of fewer molecules of the more permeable water component. Glycerol is rendered more or less impermeable though the membrane due to its larger molecular size (MW = 92.09 g/mol), relatively nonvolatile nature (B.P. 290 8C) as well as hydrophobicity of the membrane which does not get wetted by water due to greater contact angle between air, membrane and water droplets. On the other hand, water is a smaller and more volatile molecule (MW = 18 g/mol, B.P. 100 8C) which can get vaporized though the membrane pores that possess larger diameter than the molecular size of water. Moreover, water also has much higher vapor pressure than glycerol at ambient and higher temperatures. Hence, as feed water concentration reduces, the overall flux falls correspondingly. A selectivity of infinity (1) was obtained for all the trials indicating that permeate consisted of only pure water. If the polymer gets wetted, water would also drag some glycerol molecules along with it to the permeate side due to sorption and momentum transfer. The hydrophobic nature of the membrane prevents wetting phenomenon.

0

20

40

60

80

100

% Glycerol Concentration (v/v) Fig. 10. Effect of glycerol/water feed composition on flux.

The approximate pore size determined from the SEM picture is around 0.25 mm. The cross-sectional view of membrane (Fig. 8b) shows the formation of a regular upper porous layer which is defectfree and supported on a porous mesh like structure of whitish shade. Fig. 8(c) and (d) reveals the surface and cross-sectional morphologies of ultraporous polystyrene membrane. The surface morphology of the membrane (Fig. 8c) signifies the visible pores distributed uniformly across the surface. The cross-sectional view of polystyrene (Fig. 8d) displays the formation of two different layers, in which the top layer indicates the presence of ultraporous polystyrene supported on a macroporous non-woven polyester fabric. 4.2. Equilibrium sorption of membrane in glycerol/water mixtures The polystyrene layer was peeled off from the non-woven polyester fabric support and subjected to sorption experiments. A known quantity of polymer sample was taken and immersed in pure water and glycerol as well as binary mixtures of different compositions and allowed to reach equilibrium at room temperature for duration of 72 h. The effect of varying composition of glycerol/water mixture on equilibrium sorption of the membrane is shown in Fig. 9. As the concentration of glycerol in the feed mixture increased, the percentage sorption also enhanced. In pure

The effluent sample sent by a chloralkali industry contained concentrated and fuming HCl, water and traces of impurities in the form of Fe salt and higher hydrocarbons (C9–C14) which gave an oily yellow appearance and rendered the HCl unfit for utilization. The presence of higher hydrocarbons gave an oily appearance to the HCl which needed to be separated in order to obtain a purified aqueous acid for recycle or sale. Fe salts gave a yellowish color to the effluent. Feed characteristics of the effluent are given in Table 2. The performance of commercial PTFE membrane was utilized for the MD process due to its high chemical resistance in acidic media. The membrane used is hydrophobic and microporous in nature with a pore size of 0.22 mm and active layer thickness of 0.02 mm (20 mm). 4.4.1. Effect of permeate pressure Fig. 11 shows the effect of permeate pressure on flux at room temperature, which illustrates that an increase in permeate pressure from 5 to 15 mmHg results in a reduction in flux from 8.57 to 1.02 kg/m2/h. With increase in permeate pressure or reduction in vacuum, the vapor pressure which is the driving force for permeation of the more volatile components (HCl + water)

Table 2 Characteristics of chloralkali industrial effluent. S. No

Name of the feed constituent

Concentration

1. 2. 3. 4.

HCl Water Hydrocarbons (C9-C14) Fe salt

32.8% (w/v) 67.18% 50 ppm 10–20 ppm

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

8

10

Flux (kgm-2h-1)

8 6 4 2 0

0

5 10 15 Permeate Pressure (mmHg)

20

Fig. 12. Actual photograph of (a) Feed HCl effluent and (b) MD permeate.

Fig. 11. Effect of permeate pressure on flux for HCl industrial effluent feed.

Table 3 Experimental data obtained for recovery of HCl from chloralkali industrial effluent. Vacuum (mmHg)

Flux (kg/m2h)

Time of operation (h)

Wt. of permeate obtained (g)

% wt. of HCl in permeate

5 6 8 10 15

8.57 6.03 3.46 2.23 1.02

0.5 0.5 0.5 0.5 1

8.57 6.03 3.50 2.23 2.04

33.60 31.90 35.30 35.04 34.65

Concentration of HCl in feed = 32.8 wt.%.

Flux

Flux (kgm-2h-1)

2.5

50

Permeate purity

2

40

1.5

30

1

20

0.5

10

0

10

15

20 25 30 HCl concentration (wt.%)

35

40

0

39

(b)1.2 1

38 0.8 37

0.6 0.4 Flux Permeate purity

0.2 0

HCl concentration in permeate (wt.%)

60

3

0

0.05

0.1 0.15 0.2 FeCl3 concentration (wt.%)

36

HCl concentration in permeate (wt.%)

4.4.2. Effect of feed composition The effect of feed HCl concentration on flux and permeate acid concentration was studied at a constant downstream pressure of 8 mmHg and ambient temperature of 28 8C (Fig. 13(a). Flux gradually increased from 0.63 to 2.52 kg/m2/h with corresponding concentration of pure HCl ranging from 7.46 to 37.9 wt.% when the feed acid concentration was increased from 19.8 to 34.2 wt.%. The concentration of HCl was initially much lower than that in the feed due to lower vapor pressure and concentration gradient of the acid. However, it eventually increased to a value greater than the feed acid content due to the rising driving force of concentration difference besides close-boiling compositions of HCl/water mix-

(a)

Flux (kgm-2h-1)

reduces across the barrier. As the mass flux is dependent on magnitude of the driving force, the overall flux reduces. Mass transfer resistance across the membrane increases because the transport mechanism is usually based on Knudsen’s diffusion. Selectivity depends on the relative volatility of the impurity and its molecular size and generally favors the more volatile component under conditions of decreasing vacuum. It was observed that throughout the range of permeate pressure studied (5–15 mmHg), complete separation of aqueous HCl from Fe salts and organic impurities was obtained as per the analysis. Table 3 illustrates the experimental observations made during selective permeation of aqueous HCl by MD though PTFE membrane. Permeate obtained was completely colorless in comparison to the oily yellow feed (Fig. 12). The concentration of HCl in the permeate was equal or marginally higher than that in the feed since the HCl– water bond is very strong and even forms a negative azeotrope at a particular composition (20.2% HCl). Water is the preferentially permeating component due to its smaller size and reasonable vapor pressure but they drag a large number of the highly volatile HCl gas molecules though the membrane pores due to transfer of momentum as well as strong HCl–water bond.

tures which meant increasingly stronger bonds between HCl and water were being formed, which led to transfer of momentum from the preferentially permeating H2O to acid molecules. The high purity of HCl in permeate which was devoid of any impurities can be attributed to the non-wetting phenomena of membrane pores by water molecules as well as negligible vapor pressure of Fe salt and heavy hydrocarbons. PTFE membrane exhibited high degree of inertness in the corrosive medium besides significant flux, which establishes commercial feasibility. The enrichment of HCl in permeate for original effluent composition meant that the entire

35 0.25

Fig. 13. Effect of feed composition on MD separation performance (a) Feed HCl concentration and (b) Feed Fe salt concentration.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

5

Flux (kgm-2h-1)

4 3 2 1 0

0

20

40 Temperature (oC)

60

80

Fig. 14. Effect of feed temperature on flux through PTFE membrane.

acid could be separated as permeate in a stage-wise manner with intermittent dilution of feed, if necessary.

9

32.8 wt.% HCl + 10  20 ppm Fe salt and 50 ppm hydrocarbon impurities. As the temperature was increased from 25 8C to 60 8C, the flux enhanced from 1.02 to 4.5 kg/m2/h. The driving force in MD is the temperature gradient which increases the flux due by enhancing vapor pressure and energy of the preferentially permeating mineral acid and water molecules. However, the separation efficiency still remains unaffected and nearly pure aqueous HCl is obtained in the permeate due to absence of vaporization of salt and heavy hydrocarbons. This study would be useful in cases where HCl effluent is available at higher temperatures. Fig. 15 describes the process flow diagram of pilot scale MD system using hollow fiber membrane module for recovery of HCl from chloralkali industrial effluent. In the proposed flow diagram, the effluent is initially passed though a prefilter to remove suspended particles and a preheater maintained at 50 8C to provide temperature gradient as well as enthalpy of vaporization followed by a hydrophobic hollow fiber PTFE membrane module of area 0.03 m2 for separating heavy hydrocarbons and color imparting Fe compounds. The other end of the membrane module is connected to a vacuum pump. The vapors are condensed and the permeate collected is further analyzed to determine HCl concentration. 5. Economic estimation of scaled-up MD system

4.4.3. Effect of salt concentration Fig. 13(b) shows the effect of Fe salt concentration on flux and pure permeate acid content at a constant downstream pressure of 10 mmHg and ambient temperature. Rising ferric chloride (FeCl3) concentration from 0.002 to 0.2 wt.% in the feed did not affect the separation performance of PTFE membrane due to the nonvolatile nature of the salt. FeCl3 salt could also bind with water molecules thus lowering their vapor pressure by Raoult’s Law that provides HCl gas molecules more freedom to detach from water and permeate across the barrier. Feed with higher concentrations of unknown heavy hydrocarbon impurities could not be prepared synthetically and is hence not reported. 4.4.4. Effect of feed temperature The influence of feed temperature on HCl + H2O flux through PTFE membrane is illustrated in Fig. 14 for feed containing

5.1. Capital investment and list of equipment The list of equipment and capital cost for MD system are given in Table 4, wherein unit price for all major accessories are included. The total capital investment for processing 2 m3/h of the chloralkali industrial effluent is approximately 6320.62 USD, a major chunk of which consists of membrane cost with housing amounting to approximately 2833.3 USD. The capital cost excludes storage tanks which are expected to be available with the industry. 5.2. Operation and maintenance cost Operation and maintenance costs of MD system are given in Table 5, which include membrane replacement, power and raw material costs. The feed capacity was assumed to be 2 m3/h with

Fig. 15. Schematic for scale-up of MD process to recover HCl from industrial effluent.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 10

M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

Table 4 Equipment list and capital cost of MD system for HCl recovery. Item

Capacity/Size

MOC

Quantity

Total Cost (USD)

Feed pump Vacuum pump with PTFE diaphragm PTFE membrane module Preheater (up to 50 8C) Condensor pump Condenser Vacuum gauge (mmHg) Prefilter Total cost

35 lpm 35 lpm 0.03 m2 49 kW/h 17 lpm – 0.01–20 –

SS – – – SS – – PP

1 1 1 1 1 1 1 1

120 3166 2833.3 60 41.66 50 41.66 8 6320.62

lpm: liter per minute; SS: Stainless steel; MOC: material of construction; PTFE: polytetrafluoroethylene.

Table 5 Operation and maintenance cost of 2 m3/h capacity MD system for HCl recovery. Operating and maintenance cost estimation for MD system Feed capacity (m3/h) Permeate capacity (m3/h) % Recovery

2 1 50

Power cost Feed pump (kW) Preheater (kW) Condenser pump (kW) Vacuum brandt (kW) Total power consumption (kW) Price per unit (USD) (6 Rs./unit) Total power cost (USD)

0.74 49 0.37 0.4 50.1 0.1 5.05

Cartridge replacement cost No. of cartridges Price per cartridge (USD) Total cartridge replacement (USD) Duration of replacement (days) No of working h per day Cost/h (USD)

1 8 8 150 22 0.0024

Membrane replacement cost PTFE membrane (USD) Duration of replacement (yrs) No. of working h per day Total membrane replacement cost per h (USD) Raw HCl cost per liter (USD) Total raw material cost per h (USD) Labor cost per h (USD) Total operating cost per h (USD) Total operating cost per year (USD) Depreciation cost (10% of capital cost)(USD) Interest on capital (5% of capital cost) (USD) Utility cost (5% of capital cost) (USD) Total cost per year (USD)

2833.3 2 22 0.176 0.016 33.2 0.74 39.05 313,547.41 632.06 316.03 316.03 314,811.53

Permeate Quantity (LPH) Operation time (h) Quantity of permeate generated in 1 year (L/year) Cost of permeate per liter (USD)

1000 22 8,030,000 0.038

If sold at 1 USD per liter Annual profit (USD) Payback period (years)

7,717,722.74 0.041

inversion technique and characterized by Scanning electron microscopy (SEM), Fourier transform infrared spectroscopy (FTIR), X-ray diffraction studies (XRD) and Thermogravimetric analysis (TGA) to study surface/cross-sectional morphologies, intermolecular interactions, nature of polymer and thermal stability of the membrane, respectively. Indigenously synthesized polystyrene membrane exhibited selectivity of infinity for dehydration of glycerol/water mixture, indicating that permeate obtained was pure water. The low vapor pressure and high boiling point of glycerol combined with the hydrophobic nature of polystyrene rendered the alcohol more or less impermeable. PTFE membrane exhibited commercial potential for the recovery of impurity-free aqueous HCl from chloralkali industrial effluent. The strong HCl–water bond, gaseous nature of HCl, lack of vapor pressure of dissolved solid Fe salts, poor volatility and large size of higher hydrocarbon impurities ensured permeation of only HCl and water though the hydrophobic PTFE pores. Pure colorless aqueous HCl was obtained as permeate. Lower feed concentration of HCl was not favorable due to less chemical potential whereas higher HCl concentrations (>30 wt.%) produced desired results in the form of enriched in acid concentration in permeate. By increasing permeate pressure from 5 to 15 mmHg a reduction of flux was expectedly observed due to lowering in rate of desorption and driving force, whereas permeate concentration was more or less constant. Enhancement in feed temperature produced higher flux without any compromise in separation performance due to the large differences in volatilities of desirable components as against properties of undesirable ones. The presence of FeCl3 in the feed favored the permeation of HCl through absorption of water resulting in higher acid flux compared to feed containing lower salt content. Both glycerol dehydration and HCL recovery by membrane distillation are feasible for scale-up since the polystyrene membrane can be prepared by phase inversion and PTFE by a combination of melt casting and phase inversion. Economic estimation revealed that 32–33% aqueous HCl could be recovered at a low operating cost of Rs. 2.28/- per L (0.038 USD) with a payback period of 0.041 years for effluent fed at the rate of 2 m3/h. Acknowledgment

50% (or 1 m3/h) recovery in permeate. Duration of operation was assumed to be 22 h per day and the duration of replacement is 2 years. Depreciation was taken as 10% of the total capital investment (TCI) whereas utility and interest costs was assumed to be 5% of TCI. The cost per L of purified aqueous HCl comes to 0.038 USD and the payback period was found to be 0.041 years.

The authors are thankful to Council of Scientific and Industrial Research (CSIR), New Delhi for granting funds to carry out this research work under the MATES Network Project pertaining to the XII Five Year Plan Program. Appendix A. Supplementary data

6. Conclusions Microporous poly(tetrafluoroethylene) (PTFE) and ultraporous polystyrene (PS) membranes were prepared by phase

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jtice.2014. 02.010.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010

G Model

JTICE-849; No. of Pages 11 M. Madhumala et al. / Journal of the Taiwan Institute of Chemical Engineers xxx (2014) xxx–xxx

References [1] Ban R, Liu M, Qin Y, Wang H, Cui D. Enrichment of semi-volatile organic acids from aqueous solutions by multiple-effect membrane distillation. Trans Tianjin Univ 2012;18:320–9. [2] Schlosser S, Kertesz R, Martak J. Recovery and separation of organic acids by membrane-based solvent extraction and per-traction: an overview with a case study on recovery of MPCA. Sep Purif Technol 2005;41:237–66. [3] Lewandowicz G, Bialas W, Marczewski B, Szymanowska D. Application of membrane distillation for ethanol recovery during fuel ethanol production. J Membr Sci 2011;375:212–9. [4] Mohammadi T, Akbarabadi M. Separation of ethylene glycol solution by vacuum membrane distillation (VMD). Desalination 2005;181:35–41. [5] Tomaszewska M. Membrane distillation - examples of applications in technology and environmental protection. Pol J Environ Stud 2000;9:27–36. [6] Banat FA, Al-Shannag M. Recovery of dilute acetone-butanol-ethanol (ABE) solvents from aqueous solutions via membrane distillation. Bioprocess Eng 2000;23:643–9. [7] Tomaszewska M, Gryta M, Morawski AW. Study on the concentration of acids by membrane distillation. J Membr Sci 1995;102:113–22. [8] Qureshi N, Meagher MM, Hutkins RW. Recovery of 2,3-butanediol by vacuum membrane distillation. Sep Sci Technol 1994;29:1733–48. [9] Yanlong G, Francois J. Glycerol as a sustainable solvent for green chemistry. Green Chem 2010;12:1127–38. [10] Sunitha K, Nikhitha P, Satyanarayana SV, Sridhar S. Recovery of hydrazine and glycerol from aqueous solutions by membrane separation techniques. Sep Sci Technol 2011;46:2418–26. [11] da Silva GP, Mack M, Contiero J. Glycerol: a promising and abundant carbon source for industrial microbiology. Biotechnol Adv 2009;27:30–9. [12] Ralf C, Bernd S, Udo S, Wolfgang D, Reetta K. Glycerol. Ullmann’s encyclopedia of industrial chemistry. 6th ed. Verlag: Wiley-VCH; 2006. [13] Olga Guerrero-Pe´rez M, Juana Rosas M, Jorge B, Jose´ R-M, Toma´s C. Recent inventions in glycerol transformations and processing. Recent Patent Chem Eng 2009;2:11–21. [14] Xiaohu F, Rachel B. Recent development of biodiesel feedstocks and the applications of glycerol: a review. Open Fuels Energy Sci J 2009;2. 109-100. [15] Ruhl MJ. Recover VOCs via adsorption on activated carbon. Chem Eng Prog 1993;89:37–43. [16] Brasquet C, Le Cloirec P. Adsorption onto activated carbon fibers: application to water and air treatments. Carbon 1997;35:1307–13. [17] Anirudhan TS, Sreekumari SS. Adsorptive removal of heavy metal ions from industrial effluents using activated carbon derived from waste coconut buttons. J Environ Sci 2011;23:1989–98. [18] Dinesh M, Kunwar PS. Single- and multi-component adsorption of cadmium and zinc using activated carbon derived from bagasse an agricultural waste. Water Res 2002;36:2304–18. [19] Ba¨rbel K, Sascha W. Novel distillation concepts using one-shell columns. Chem Eng Process: Process Intensif 2004;43:339–46. [20] Ligero EL, Ravagnani TMK. Dehydration of ethanol with salt extractive distillation – a comparative analysis between processes with salt recovery. Chem Eng Process: Process Intensif 2003;42:543–52. [21] Gopal PA. Glycerol Adv Biochem Eng Biotechnol 1990;41:95–128. [22] Cockrem Michael CM, Johnson PD, Recovery of lactate esters and lactic acid from fermentation broth. United States.5210296A;1993.

11

[23] Saha B, Chopade SP, Mahajani SM. Recovery of dilute acetic acid though esterification in a reactive distillation column. Catal Today 2000;60: 147–57. [24] Hua-Jiang H, Shi Ramaswamy Tschirner UW, Ramarao BV. A review of separation technologies in current and future biorefineries. Sep Purif Technol 2008;62:1–21. [25] Gil ID, Uyaza´n AM, Aguilar JL, Rodrı´guez GL, Caicedo A. Separation of ethanol and water by extractive distillation with salt and solvent as entrainer: process simulation. Braz J Chem Eng 2008;25:207–15. [26] Gurmukh MD. Comparison of membrane processes with distillation for alcohol/water separation. J Membr Sci 1982;12:1–26. [27] Seiler M, Ko¨hler D, Arlt W. Hyperbranched polymers: new selective solvents for extractive distillation and solvent extraction. Sep Purif Technol 2003;30:179–97. [28] Rongling L, Yingjie Q, Xiaojun L, Liqiang L. Concentrating aqueous hydrochloric acid by multiple-effect membrane distillation. Front Chem Sci Eng 2012;6:311–21. [29] Xu T. Electrodialysis processes with bipolar membranes (edbm) in environmental protection – a review. Resour Conserv Recycl 2002;37:1–22. [30] Sata T. Ion exchange membranes and separation processes with chemical reactions. J Appl Electrochem 1991;21:283–94. [31] Tina C, Michael Rajesh A, Amaranadh J, Amit KT, Ajay KS, Prakash S, et al. Stable ion-exchange membranes for water desalination by electrodialysis. Desalination 2011;282:2–8. [32] Archana A, Sahu KK. An overview of the recovery of acid from spent acidic solutions from steel and electroplating industries. J Hazard Mater 2009;171:61–75. [33] Regel Rosocka M. A review on methods of regeneration of spent pickling solutions from steel processing. J Hazard Mater 2010;177:57–69. [34] Bhausaheb LP, Sane MG, Parjane SB, Mahendra G. Vacuum membrane distillation for desalination of ground water by using flat sheet membrane. Int J Chem Bio Eng 2011;4:13–8. [35] Gryta M. Concentration of NaCl solution by membrane distillation integrated with crystallization. Sep Sci Technol 2002;37:3535–58. [36] Sharmiza A, Manh H, Huanting W, Zangli X. Commercial PTFE membranes for membrane distillation application: effect of microstructure and support material. Desalination 2012;284:297–308. [37] Lawson KW, Llyod DR, Membrane distillation. J Membr Sci 1997;124:1–25. [38] Khayet M, Matsuura T. Pervaporation and vacuum membrane distillation processes: Modelling and Experiments. AIChE J 2004;50:1697–712. [39] Gryta M, Karakulski K, Morawski A. Separation of effluents from regeneration of a cation exchanger by membrane distillation. Desalination 2006;197:50– 62. [40] Tomaszewska M, Gryta M, Morawski AW. Recovery of hydrochloric acid from metal pickling solutions by membrane distillation. Sep Purif Technol 2001;2223:591–600. [41] Tomaszewska M, Gryta M, Morawski AW. Mass transfer of HCl and H2O across the hydrophobic membrane during membrane distillation. J Membr Sci 2000;166:149–57. [42] Tomaszewska M, Gryta M, Morawski AW. The influence of salt in solutions on hydrochloric acid recovery by membrane distillation. Sep Purif Technol 1998;14:183–8. [43] Yuan Y, Lee TR. Contact angle and wetting properties. In: Bracco G, Holst B, editors. Surface science technologies. Verlag Berlin Heidelberg: Springer; 2013. p. 3–34.

Please cite this article in press as: Madhumala M, et al. Recovery of hydrochloric acid and glycerol from aqueous solutions in chloralkali and chemical process industries by membrane distillation technique. J Taiwan Inst Chem Eng (2014), http://dx.doi.org/10.1016/ j.jtice.2014.02.010