Enhanced Flux and Electrochemical Cleaning of ... - ACS Publications

Subscriber access provided by Eastern Michigan University | Bruce T. Halle Library

Article

Enhanced Flux and Electrochemical Cleaning of Silicate Scaling on Carbon Nanotube-Coated Membrane Distillation Membranes Treating Geothermal Brines Li Tang, Arpita Iddya, Xiaobo Zhu, Alexander V Dudchenko, Wenyan Duan, Craig Turchi, Johan Vanneste, Tzahi Y. Cath, and David Jassby ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b12615 • Publication Date (Web): 13 Oct 2017 Downloaded from http://pubs.acs.org on October 17, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Enhanced Flux and Electrochemical Cleaning of Silicate Scaling on Carbon Nanotube-Coated Membrane Distillation Membranes Treating Geothermal Brines

Li Tang a,⊥*, Arpita Iddyaa, ⊥, Xiaobo Zhu a, Alexander V. Dudchenko a, Wenyan Duan a, Craig Turchi b, Johann Vanneste c, Tzahi Y. Cath c, and David Jassby a,* a

b

c

Department of Chemical and Environmental Engineering, University of California, Riverside, California 92521, United States

National Renewable Energy Lab, Department of Energy, Golden, Colorado 80401, United States

Department of Civil and Environmental Engineering, Colorado School of Mines, Golden, Colorado 80401, United States

* Corresponding author: David Jassby, E-mail: [email protected], Phone: 1-951-827-6475 * Corresponding author: Li Tang, E-mail: [email protected] ⊥

These authors contributed to the work equally

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

Abstract

2

The desalination of inland brackish groundwater offers the opportunity to provide potable

3

drinking water to residents and industrial cooling water to industries located in arid regions.

4

Geothermal brines are used to generate electricity, but often contain high concentrations of

5

dissolved salt. Here, we demonstrate how the residual heat left in spent geothermal brines can

6

be used to drive a membrane distillation (MD) process and recover desalinated water. Porous

7

polypropylene membranes were coated with a carbon nanotube (CNT)/poly (vinyl alcohol)

8

layer, resulting in composite membranes having a binary structure that combines the

9

hydrophobic properties critical for MD with the hydrophilic and conductive properties of the

10

CNTs. We demonstrate that the addition of the CNT layer increases membrane flux due to

11

enhanced heat transport from the bulk feed to the membrane surface – a result of CNT’s high

12

thermal transport properties. Furthermore, we show how hydroxide ion generation, driven by

13

water electrolysis on the electrically conducting membrane surface, can be used to efficiently

14

dissolve silicate scaling that developed during the process of desalinating the geothermal brine,

15

negating the need for chemical cleaning.

16 17

Keywords:

18

cleaning.

Membrane distillation; silicate scaling; carbon nanotubes; electrochemical

19 20 21 22 23 24 25 26 27 28 29

2


Page 2 of 44

Page 3 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1 2

Graphical Abstract

3

4 5 6 7 8 9 10 11 12 13 14

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 2 3

1. Introduction

4

The desalination of inland brackish groundwater has the potential of addressing water

5

shortages across many regions. Over the last several decades, membrane-based desalination

6

technologies, such as reverse osmosis (RO), and membrane distillation (MD) have attracted

7

significant research attention. MD, which relies on thermal energy to drive the desalination

8

process, is an attractive and cost-effective technology when an abundant low-grade heat source

9

is available, or when high salinity brines require treatment 1-3. Such a heat source can be found

10

in geothermal brines, which are often exploited for power generation in geothermal power

11

plants

12

concentrations as high as 254,000 mg/l measured in deep brines near the Salton Sea, CA 5. In

13

the power generation process, heat from the brines is extracted and used to power a turbine and

14

generate electricity, with the spent brines pumped back into the ground 6. However, not all of

15

the heat is extracted in the process, and the spent brine is returned to the environment with

16

temperatures as high as 70 °C, which are sufficient to drive brine desalination 7. Thus, capturing

17

the residual heat in the spent geothermal brines and using it to drive a desalination process

18

offers an attractive option to arid inland regions that often struggle to secure potable water

19

resources.

4

; the salinity of these brines depends on the underlying geology, with TDS

20

In the direct-contact MD (DCMD) process, a hot brine flows over a hydrophobic

21

microporous membrane, with cold water flowing on the opposite (distillate) side of the

22

membrane; due to the temperature difference across the membrane, a vapor pressure difference

23

emerges, which drives water vapor from the hot feed to the cold distillate, generating a

24

desalinated stream. Because the membrane is hydrophobic, liquid water does not transport

25

across the membrane, and dissolved species, such as ions, remain in the hot feed and do not

26

cross the hydrophobic barrier. Because MD does not rely on pressure as a driving force, and

27

because water vapor pressure is only moderately impacted by salinity, the technology can be

28

used to desalinate high salinity brines that have an osmotic pressure that is too high for

29

desalination by RO. In addition, MD requires less specialized mechanical equipment compared

30

to high-pressure desalination systems 8-11.

4


Page 4 of 44

Page 5 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

Even though MD technology was first proposed in the late 1960s 12, it has yet to be applied

2

in large-scale industrial applications. One of the key technical challenges that hinder the wide-

3

spread use of MD is membrane fouling, and in particular, membrane mineral scaling

4

mineral scaling, supersaturated conditions along the membrane surface, which develop due to

5

water leaving the brine during the desalination process, lead to the nucleation and growth of

6

mineral deposits on the membrane surface (heterogeneous nucleation) as well as nucleation and

7

growth in the feed stream itself (homogeneous nucleation), which results in the deposition of

8

solids on the membrane surface, although there have been several demonstrations of

9

fluorosilicone surface coatings that have reduced (although not completely prevented) scaling, 13-16

13

. In

10

including silica scaling

11

of pore clogging or surface blockage. In addition, scaling exacerbates temperature and

12

concentration polarization, and can lead to a drop in salt rejection due to pore wetting and

13

membrane material damage

14

desalinating brackish groundwater

15

water 27-28. During the MD process, silicate monomers accumulate along the membrane surface,

16

and under certain concentration and pH conditions begin to polymerize

17

multivalent cations in natural waters (e.g., iron, aluminum, calcium, magnesium, and barium

18

ions) favors and promotes the polymerization process 29-30. This polymerization process leads to

19

the formation of a dense and sticky “gel” layer, which is a complex and amorphous product that

20

is difficult to remove. Membrane cleaning using a high pH solution provides an effective

21

method to recover membrane performance because of the dissolution of silicates at high pH

22

19, 26

23

interruptions, which leads to an increase in O&M costs 13, 17, 26, 31.

. Once scaling occurs, MD performance rapidly deteriorates because

15, 17-21

. Silicate scaling, in particular, is a challenge when

22-26

. Silicates of various forms are ubiquitous in natural 19

. The presence of

13,

. However, this cleaning process requires the use of chemicals and results in process

24

Recently, thanks to the discovery of the unique electrical, thermal, and mechanical

25

properties of carbon nanotubes (CNTs), interest in the utilization of these materials in

26

membrane separation processes has been increasing. In particular, the ability to form porous,

27

electrically conducting composite polymeric materials has attracted significant attention due to

28

the exceptional anti-fouling and self-cleaning properties enabled through the application of an

29

electrical potential to the membrane surface

30

cleaning mechanisms include simple electrostatic repulsion, electrochemical redox reactions,

31

micro-environmental pH changes, and micro-bubble generation 37-38. Among these mechanisms,

2, 32-37

5

. The electrochemical anti-fouling and self-



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 44

1

manipulating pH in the microenvironment of the membrane surface serves as a very promising

2

method to remove silicate scale in the MD process because silicate scales can be readily

3

dissolved under high pH conditions. This approach can be realized by forming a CNT/polymer

4

composite material that consists of a layer of electrically conducting CNTs deposited on a

5

microporous hydrophobic support. By applying an electrical potential to the membrane surface,

6

with the membrane as a cathode, water splitting can be facilitated, which generates hydroxide

7

ions and increases the pH on the membrane surface.

8

In addition to their electrical properties, CNTs also possess remarkable heat transfer

9

properties. The thermal conductivity of CNTs ranges between 2000-6000 W m–1 K–1 at room 39

10

temperature

11

feed and distillate streams; however, as water vapor passes through the hydrophobic membrane,

12

the feed solution cools down, resulting in the formation of thermal polarization

13

thickness of the temperature boundary layer is dependent on several factors, including

14

hydrodynamic (mixing) conditions along the membrane surface, the temperature difference

15

between the feed and distillate streams, and membrane morphological architecture

16

Ultimately, for the distillation process to proceed, heat from the bulk stream must transfer

17

across the thermal boundary layer to the membrane/feed interface to drive vapor formation 41, 43,

18

48

19

mixing, the rate of this conductive heat transfer is fundamentally determined by the thermal

20

conductivity of water, which is 0.6 W m–1 K–1 at room temperature

21

higher thermal conductivity of CNTs, a layer of CNTs deposited on a microporous hydrophobic

22

support and extending into the bulk stream could increase conductive heat transfer from the

23

bulk to the membrane surface, which would reduce thermal polarization and increase membrane

24

performance 45.

. The MD process relies on the maintenance of a thermal gradient between the 40-43

. The

44-47

.

. Ignoring advective transport of water from the bulk feed to the membrane surface through 49

. Considering the far

25

The main objective of this research was to examine the effectiveness of using a

26

CNT/polymer composite MD membrane to decrease thermal polarization and for in-situ

27

electrochemical cleaning to remove silicate scales from the membrane surface. In this study,

28

microporous polypropylene (PP) base membranes were spray-coated with a thin film of CNTs

29

by a custom-made spray coater, which are then crossed-linked with a hydrophilic polymer

30

(poly(vinyl alcohol) (PVA)). Because the CNT films are spray coated on the membrane surface,

31

the underlying hydrophobic properties of the PP support are maintained, and the composite

6


Page 7 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

material functions as an MD membrane, albeit having a top hydrophilic layer. A laboratory-

2

scale MD system was used to treat a synthetic geothermal brine that had high silicate

3

concentrations with and without electrochemical cleaning. Membrane flux and electrical

4

conductivity were continuously monitored and recorded in real time throughout the silicate

5

scaling and electrochemical cleaning processes. In addition, flux differences between the coated

6

and uncoated membranes were evaluated as a function of the temperature difference and

7

hydrodynamic conditions in the feed channel. Our results show that the membrane modified

8

with CNT thin films can facilitate rapid in-situ cleaning of silica scaling. In addition, we

9

demonstrate that the addition of the CNT film increases membrane flux by enhancing heat

10

transfer from the bulk to the membrane surface, compressing the thermal polarization layer.

11 12

2. Materials and methods

13

2.1

Materials

14

PP membranes having a nominal pore size of 0.45 µm were provided by 3M (Charlotte, NC)

15

and were used directly without any treatment as the base membranes on which CNT thin films

16

were spray coated. 50 wt% glutaraldehyde (GA) solution was purchased from Fisher Scientific

17

and used as received. 146,000–186,000 MW PVA, sodium dodecylbenzenesulfonate (DDBS),

18

Na2SiO3·5H2O, AlCl3·6H2O, BaCl2·2H2O, FeCl6· H2O, KCl, CaCl2·2H2O, MgCl2·6H2O, and

19

Na2SO4 were purchased from Sigma Aldrich and used as received. Multi-Walled CNTs

20

functionalized with carboxylic groups (-COOH) via plasma treatment were purchased from

21

CheapTubes (Cheaptubes Inc., Brattleboro, VT) and used without any further purification.

22

Based on the specification provided by the manufacture, the CNTs are reported to have an outer

23

diameter of 13-18 nm, a length of 1-12 µm and purity >99%, with a functional group content of

24

7.0 ± 1.5%.

25 26 27

2.2

Membrane surface modification CNT suspensions were prepared using previously reported methods

33-34

. Briefly, CNT

28

powder was suspended in deionized water at 1.0 wt% concentration with 1:10 ratio of CNT:

29

DDBS. The suspension was sonicated with a horn sonicator (Branson, Danbury, CT) for 30 min

30

to fully disperse the CNT powder. A 0.1 wt% PVA solution was prepared by dissolving PVA in

31

deionized water at 95-100 °C with stirring. To modify the base PP membranes with CNT thin

7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

films, CNT suspension were sprayed evenly on top of membrane surfaces with a surface mass

2

of either 0.09 mg cm–2 (CNT_009), 0.19 mg cm–2 (CNT_019), or 0.37 mg cm–2 (CNT_037). A

3

final thin layer of PVA was sprayed on top of the CNTs-modified membrane surface, with a

4

surface density of 0.004 mg cm–2. The prepared membranes were washed using deionized water

5

for at least 2 hrs to remove DDBS and then the membranes were dried in oven at 90 °C for 10

6

minutes. Following that, the dried membranes were submerged into a crosslinking solution

7

consisting of 11.3 g L–1 GA as the cross linker and 4.4 g L–1 hydrochloric acid as the catalyst

8

and heated at 70 °C for 1 hr. The membranes were then removed from the crosslinking solution,

9

dried at 90 °C for 10 min, and used without any further treatment.

10 11

2.3

Membrane characterization

12

Membrane surfaces and cross sections were imaged using scanning electron microscopy

13

(SEM, FEI NNS540). All samples were affixed onto SEM stubs with carbon tape and sputter

14

coated with Pd/Pt before imaging (Sputter coater 108 Auto, Cressington, UK). Quantitative

15

chemical analysis of the membrane surface was performed by energy dispersive X-ray (EDX)

16

spectroscopy. Water flux was measured and compared between the base membrane (bare PP)

17

and the CNT_037 membrane, as a representative of the CNTs-modified membranes, at three

18

different feed temperatures (50, 60, and 70 °C). Membrane surface roughness was measured

19

using atomic force microscopy (AFM, Bruker, Billerica, MA) in tapping mode. The membrane

20

sheet resistance was measured using a 4-point conductivity probe (Mitsubishi, MCP-T610).

21

Contact angle measurements were performed using a contact angle goniometer (Rame-Hart,

22

Model 250, Netcong, NJ), using the captive bubble method. Attenuated total reflectance Fourier

23

transform infrared spectroscopy (ATR-FTIR, Thermo Scientific, Waltham, MA) was performed

24

on all samples to analyze and compare the functional groups before and after membrane

25

modification.

26

Measurements of the relative potential of the membrane vs. a Ag/AgCl reference electrode

27

at an applied cell potential of 15 V were performed by connecting a membrane sample (working

28

electrode), a titanium (Ti) sheet (counter electrode), and a reference electrode (Ag/AgCl

29

solusion) to a potentiostat (CH Instruments, Austin, TX) following a similar procedure as

30

previously described

31

reference electrode were placed in an electrochemical cell, with the open-circuit potential on the

50

. Briefly, the membrane sample, the Ti counter electrode, and the

8


Page 8 of 44

Page 9 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

membrane surface measured once a 15 V cell potential was applied to the membrane and Ti

2

counter electrode using a DC power source (KORAD KA3005P).

3 4

2.4

Design and operation of the MD system

5

An automated laboratory-scale cross flow MD system (with a polycarbonate membrane flow

6

cell) was used to perform all of the experiments (Figure 1). Each portion of the flow cell has an

7

open channel with the dimensions of 100 mm in length, 40 mm in width, and 3.80 mm in

8

height. A Ti sheet, located 3 mm above the membrane surface, was used as the counter

9

electrode. A peristaltic pump (Cole Parmer, Pump Drive Model 7553-70, Pump Head Model

10

77200-50, Vernon Hills, IL) with temperature resistant tubing was used to circulate the hot feed

11

solution, while a gear pump (Greylor, PQ-12/24, Cape Coral, FL) was used to circulate the cold

12

distillate solution on the opposite side of the membrane. Feed and permeate spacers were placed

13

in the flow channels. The circulation flow rates were monitored using two rotameters. The

14

conductivity of the circulating distillate was measured continuously with a conductivity meter

15

(Thermo Scientific, Orion Star A322) to monitor salt rejection and membrane wetting.

16

The feed solution was heated using an immersion heater (Process Technology, VX1111,

17

Mentor, OH) immersed in the feed tank, with a temperature sensor (Vktech, DS18b20)

18

suspended in the hot water. Another temperature sensor was placed immediately prior to the

19

inlet of the feed channel. The two temperature sensors were connected to a temperature control

20

unit and the inlet temperature of the hot feed solution could be maintained at a constant value

21

with a PID cascade loop 51. A vertically mounted level float switch (Madison, M8000, Branford,

22

CT) was suspended in the hot water tank to maintain a constant water level throughout the

23

experiments. A temperature sensor was placed immediately before the inlet of the distillate

24

channel, and the inlet temperature of the cold distillate solution was maintained constantly at

25

20 °C using a chiller (PolyScience, 6500 Series, 1/2 HP) through a stainless-steel heat-exchange

26

coil. A digital balance (Fisher Scientific Education Precision Balance) was used to weigh the

27

excess produced distillate for the determination of water flux. When comparing the flux of the

28

CNT-coated membranes and the bare PP membranes, deionized water was used as the feed at

29

different temperatures (50, 60, and 70 °C), with the distillate maintained at 20 °C. In these

30

experiments, the membrane was tested in the presence and absence of a feed-side spacer, at

9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

three different flow rates (0.75, 1.5, and 2 L min–1); these flow rates correspond to a cross-flow

2

velocity of 10.4, 20.8, and 27.7 cm s–1, respectively.

Page 10 of 44

Control Program on PC Microcontroller

Chilling Unit Tcold1

Tcold2 Distillate Solution Tank

Membrane Liquid Level Sensor

Balance Thot1

Thot2

Feed Solution Tank

3 4

Buffer Tank

Figure 1. Schematic of the MD testing system.

5 6

2.5

Preparation of silicate scaling solution

7

A simulated brine solution, with a composition similar to that found in geothermal brines

8

used for power generation in Nevada, was prepared by dissolving representative salts in

9

deionized water; salts (type and concentrations) used to prepare the solution are summarized in

10

Table 1. In an effort to limit the time required for fouling experiments, the concentration of salts

11

was increased by a factor of six over what is naturally found in Nevada geothermal fluids. Thus,

12

the concentrations in Table 1 are 6X what is naturally found in the environment.

13

Table 1. Solution chemistry of the feed solution Salts NaSiO3·5H2O AlCl3·6H2O BaCl2·2H2O FeCl3·6H2O CaCl2·2H2O KCl MgCl2·6H2O Na2SO4

Concentration (mg L–1) 1906 10 2 6 176 85 102 92

Concentration (mM) 8.985 0.051 0.010 0.026 1.439 1.372 0.605 0.774

14 15

2.6

Silicate scaling and electrochemical cleaning procedure 10


Page 11 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

MD scaling experiments on the CNT-modified membranes were started with 10 L of the 6X

2

brine solution in the feed tank and 1 L of deionized water as the initial fluid in the distillate

3

tank. The temperatures of the feed and distillate were maintained at 60 °C and 20 °C,

4

respectively. Both the feed and distillate were circulated at a flowrate that resulted in a flow

5

velocity of 14 cm s–1 in the flow channels. The flux and the conductivity of the permeate was

6

recorded in real time.

7

The first electrochemical cleaning cycle was initiated when the flux declined by 40%

8

compared to its original value. During the cleaning, deionized water and the distillate streams

9

were circulated with the temperature difference maintained at 40 °C (60 °C feed and 20 °C

10

distillate). The cleaning was first conducted for 20 min without applying any electrical potential

11

until water flux was stable. Then, a constant DC current (20 mA) was applied across the

12

membrane (cathode) and the Ti counter electrode (anode) using a DC power source for 50 min

13

until the water flux was recovered to the original level. A freshly prepared feed solution, with

14

6X higher feed concentration than that found naturally in geothermal brines from Nevada, was

15

used to re-foul the membrane after it was cleaned electrochemically. Since the initial feed

16

concentrations in the feed were already near saturation with regards to silica, the system was

17

always operated in constant concentration mode, as increasing the concentration (by not

18

returning the permeate to the feed tank) would lead to silicate formation in the feed tank.

19

Following the cleaning, a new round of scaling was initiated on the same membrane with a

20

freshly prepared feed solution. The scaling experiment continued until the flux declined by

21

40%, and then the second cleaning was initiated. A round of scaling and cleaning followed. In

22

addition, a continuous silicate scaling experiment without any electrochemical cleaning was

23

performed as a control. All scaling experiments were conducted in duplicate using new

24

membranes in each round.

25 26

2.7

27

The driving force for mass transfer in DCMD is the partial vapor pressure difference of water

28

across the membrane. This difference arises from the difference in the feed and distillate

29

temperatures on the opposite sides of the membrane. When modeling the flux through MD

30

membranes, it is important to account for changes in the vapor pressure within the concentration

31

polarization layer, which can be accomplished using different modeling methods

Modeling vapor flux in DCMD

11


52

. In our


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 44

1

experiments, the feed stream was DI water, and as a result, we neglected changes in the partial

2

vapor pressure along the membrane surface. Mass transport occurs as water from the bulk of the

3

feed stream is brought to the membrane surface where it evaporates to form water vapor. The

4

vapor is then transported across the membrane to the distillate side, where it condenses to form

5

water.

6

The relationship of vapor pressure (P in Pa units) as a function of temperature (T in K units)

7

is given by the Antoine equation:

8

= exp ( −

9

)

(1)

where A, B, and C are Antoine constants 23.238, 3841.273 and 45, respectively, for water 53.

10

The temperature difference also leads to heat transport from the hot feed to the cold distillate.

11

This transport occurs over three regions: (1) bulk of feed to membrane surface, primarily

12

through convection, (2) through the membrane, primarily through conduction, and (3) from the

13

surface of the membrane to the bulk of distillate, primarily through convection.

The heat flux in each domain is given by Equations 2-4, where represents the heat

14 15 16 17 18

flux over the respective domain: the subscripts represent the feed (f), membrane (m), and

permeate (p). In addition, ℎ is the convective heat transfer coefficient, is the conductive heat transfer coefficients of the membrane, and is the thickness of the membrane. The bulk

temperatures of the feed and permeate are denoted by / , while & , denote the

19

temperatures at the membrane-feed and membrane-permeate interface, respectively (Figure 2a).

20

In Equation 3, J is the vapor flux, and H the enthalpy of vaporization.

21

= ℎ −

22 23 24

=

(2)

− + (! ∗ #)

(3)

= ℎ ( − )

(4)

25

At steady state, the heat flux across the membrane is constant for flat sheet DCMD membranes

26

45

27

:

= =

(5)

28

Substituting equations 2-4 into equation 5, yields the temperatures at the feed-membrane and

29

membrane-permeate interfaces:

12


Page 13 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


%

%

,∗-∗

%

,∗-∗

1

= − $ ( ∗ ) + ∗ ) − +

2

= + $ ( ∗ ) + ∗ ) − +

3

where U is the overall heat transfer coefficient, given by:

4

/ = $

5

The convective heat transfer coefficient for the feed and permeate is calculated using:

6

ℎ/ =

7

&'

*

&.

*

%

%

&'

+

+

%

&.

+

+

(

(6) (7)

(8)

01'/. ∗ '/.

(9)

2-

where, / is the conductivity of heat in water in the feed/permeate and dH is the hydraulic

8

diameter of the flow system. The Nusselt number (34/ ) is dependent on the flow conditions;

9

in our model we use one of the two following correlations to calculate the Nusselt number for

10

laminar and turbulent flow conditions 45, 54:

11

a. For laminar flow

12

34 = 4.36 +

13

b. For turbulent flow

9.9:;∗?∗(2-/@)

%A9.9%%[?∗)

(10)

CD G.H +] E

2-

34 = 0.036 ∗ JK 9.L ∗ M %/: ∗ ( @ )9.9NN

14

(11)

15

where Re and Pr are the Reynolds and Prandtl numbers, respectively. The vapor flux through

16

the membrane is obtained by modelling mass transfer across the membrane using the dusty gas

17

model for the DCMD 55. The criteria for the mode of vapor transport through the membrane is

18

given by the Knudsen number, which is the ratio of the mean free path of molecules to the

19

radius of the membrane pore. If the mean free path of molecules is of the order of the pore size,

20 21 22 23 24 25 26

i.e., OP >1, the molecule-pore wall collisions are dominant and Knudsen diffusion is responsible

for mass transfer. For the case where the mean free path is much smaller than the pore size, OP

Enhanced Flux and Electrochemical Cleaning of ... - ACS Publications

Recommend Documents