Harnessing Clean Water from Power Plant Emissions Using

Subscriber access provided by Kaohsiung Medical University

Harnessing clean water from power plant emissions using membrane condenser technology Jeong F. Kim, Ahrumi Park, Seong-Joong Kim, PyungSoo Lee, Young Hoon Cho, Hosik Park, S. E. Nam, and You In Park ACS Sustainable Chem. Eng., Just Accepted Manuscript • DOI: 10.1021/ acssuschemeng.8b00204 • Publication Date (Web): 28 Mar 2018 Downloaded from http://pubs.acs.org on March 30, 2018

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 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 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.

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 22 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 Sustainable Chemistry & Engineering

1

Harnessing clean water from power plant emissions using

2

membrane condenser technology

3

Jeong F. Kim1, Ahrumi Park1, Seong-Joong Kim1,2, PyungSoo Lee1, YoungHoon Cho1, HoSik

4

Park1, SeungEun Nam1, YouIn Park1*

5 6 7 8

1

Research Center for Membranes, Advanced Materials Division, Korea Research Institute of Chemical Technology, 141 Gajeongro, Yuseong, Daejeon 34114, Republic of Korea 2

University of Science & Technology (UST), 176 Gajeongro, Yuseong, Daejeon 34114, Republic of Korea

9 10

*Corresponding Author Email Address: [email protected]

11

Keywords: membrane condenser, flue gas dehydration, capillary condensation, white plume,

12

ceramic membrane

13

Abstract

14

Power plants consume a major fraction of water to generate electricity, typically in the

15

range between 30 – 50% of all fresh water sources. Most of the water from plants are lost

16

with heat through stack and cooling towers. It has been reported that if 20% of these water

17

can be recycled, power plants can be self-sustainable, allowing them to be located with higher

18

flexibility. Membrane contactor can be an effective solution to harness this source of water,

19

but most of the work have been focused on dense vapor separation membranes with limited

20

success. In this work, we investigated potential application of membrane condenser

21

technology to harness fresh water from power plants. It has been shown that membrane

22

condenser configuration can be three orders of magnitude more effective in recovering water

23

compared to dense vapor separation membranes, with a reasonable water/SOx selectivity of

24

100. We have prepared suitable ceramic membranes as a proof-of-concept and achieved up to

25

85% dehumidification efficiency in a single-pass flow. A thorough energy balance indicates

26

that both heat and water flux must be carefully balanced to maximize the membrane

27

condenser performance, and an effective module design must be developed.

28

ACS Paragon Plus Environment

ACS Sustainable Chemistry & Engineering 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

29

Introduction

30

In the United States alone, power plants consume 40% of all available water sources (45%

31

in EU) 1. It has been calculated that if 20% of the evaporated water can be recovered from a

32

power plant, it can be self-sufficient from the process water point of view 2. The current

33

power plants on average consume approximately 1.6 L of water to generate 1 kWh of

34

electricity, which converts to 45,000 m3.hr-1 of water for a regular-sized 500 MW plant 3. As

35

illustrated in Figure 1, two main sources of emissions in power plants are from the stack and

36

the cooling towers. Streams emitting from a stack become saturated in the desulfurization

37

step (FGD), and the streams from cooling towers are typically river or seawater evaporated to

38

cool the steam cycle stream.

39 40 41

Figure 1. Power Plant Operation Schematics. Significant amount of water and energy are lost through stack and cooling towers.

42

The evaporated water (i.e. white plumes) also poses several downsides such as visual

43

pollution, frost damage, and corrosion of chimneys and stacks. The current practice now is to

44

intentionally heat up the emission stack to avoid corrosion 4, which consume additional

45

energy. If the evaporated water can be effectively recovered, it can be a fruitful source of

46

distilled water and latent energy, and it can relieve the exacerbating energy-water collisions,

47

particularly during drought or hot weather. In addition, the technology can be valuable to

48

other industries that employ water-cooling systems such as steel, semiconductors, and pulp

49

industry.


Page 2 of 22

Page 3 of 22 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


50

Several technologies to recover evaporated water exists such as heat exchangers,

51

absorption, adsorption, and membranes. Implementation of heat exchangers to condense out

52

the water vapor is an obvious option that has been considered for a long time 5, however, the

53

recovered water is typically of low quality that requires further treatment, and the

54

infrastructure is inherently prone to corrosion over time, appending additional maintenance

55

cost. The use of liquid absorbent (glycol-based) or solid adsorbent has been partly

56

implemented

57

required to regenerate the absorbent makes the overall process uneconomical at the current

58

level.

6-7

. The process yields high quality water that can be recycled, but the energy

59

Membrane technology has been researched as a promising option to recover evaporated

60

water. To reduce the water demand from the energy sector, US Department of Energy and

61

Gas Technology Institute (GTI) have developed membrane-based system to recover

62

evaporated water8. In Europe, CapWa2 and MATCHING grants aim to capture evaporated

63

water from power plants are going through pilot trials.

64

Many works exists that employ dense membranes that selectively permeate water vapor

65

over other gases with excellent selectivity and permeability, reaching above 107 (H2O/N2) and

66

105 barrer 9-10. However, the evaporated water through cooling tower and stack leaves at near

67

atmospheric pressure, requiring additional compression or vacuum to apply the necessary

68

driving force for separation. Hence, current vapor separation membranes are completely

69

limited by pressure ratio11 and mass transfer resistance9, not by the membranes. There has

70

been works that employ supported liquid membranes that combine advantages of absorbent

71

and membranes

72

analysis.

12

but the work is largely in the research stage without firm economic

73

Recent advances in membrane fabrication now allows more unique membrane

74

configurations to be developed such as membrane distillation 13-14, membrane crystallization

75

15

76

actively participate in separation but provides an interface to facilitate chemical potential

77

minimization (Figure 2). In the field of MC, a pioneering work was reported by GTI

78

company

79

Instead of permeating water vapor through a dense membrane, TMC exploits the capillary

80

condensation phenomenon to selectively condense out water vapor within small membrane

81

pores (vapor condenses at lower saturation pressures). With this innovative unit operation,

, membrane emulsifiers

8

16

, and membrane condensers (MC)17. These membranes do not

by introducing an innovative concept of transport membrane condenser (TMC).



82

Wang et al. 8 reported 19% reduction in greenhouse emission, 20% reduction in boiler water

83

consumption, and 40% enhancement of process efficiency. More recent works by Zhao et

84

al.18-19 and Chen et al.20-21 further investigated the efficacy of TMC configurations.

85 86 87 88 89 90

Figure 2. Illustration of membrane-based dehydration configurations: (a) vapor permeation using a dense membrane, (b) transport membrane condenser using a microporous membrane to selectively condense water vapor within capillary pores, (c) conventional membrane condenser configuration using a hydrophobic microporous membrane to pass gas while condensing water vapor on the surface.

91

Taking a slightly different approach, Drioli et al., 3, 17, 22-23 have reported conventional MC

92

configuration to recover water vapor using polymeric membranes. However, as the water

93

condenses outside of the pores and washes down, the quality of water was low. In addition,

94

this configuration requires a new membrane module design to drain the condensed water. For

95

capturing evaporated water from flue gas, many possible MC configurations exists22 but no

96

clear optimizations in terms of membranes and process configurations have yet been carried

97

out.

98

In this work, we investigated key parameters to maximize the TMC configuration

99

performance for capturing the evaporated water. We fabricated ceramic membranes and tested

100

the effect of independent parameters on recovered water quality, as well as process conditions

101

such as humidity, flowrates, and thermal gradients. Moreover, a full energy balance was

102

carried out to reveal that TMC performance is highly dependent on the temperature gradient

103

across the membrane, which can be tailored during membrane fabrication step. The obtained

104

TMC performance was compared against the dense vapor separation membranes to show that


Page 4 of 22

Page 5 of 22 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


105

TMC has a high potential to recover both water and energy from power plant emissions.

106



Page 6 of 22

107

Experimental

108

Materials

109

Ceramic alumina particle (α-Al2O3, d50=1.1 µm, 99.9%) for membrane fabrication was

110

purchased from Sumimoto (Japan). Polysulfone (Ultrason S6010, BASF, Germany) was

111

employed as polymer binder, DISPERBYK-190 (BYK, Altana, Korea) as dispersant,

112

PEG200 (Sigma-Aldrich, Korea) as pore former, Mg(OH)2 (Sigma-Aldrich, Korea) was used

113

as sintering aid. N-methlypyrrolidone(NMP, 99.9%) and ethanol (EtOH, 99.9%) were

114

purchased from Samchun Chemicals (Korea). All the reagents were used without further

115

purification. Hyflux20 ceramic membranes (mean pore size of 20 nm) were purchased from

116

Hyflux Ltd (Singapore).

117

Membrane Preparation

118

Ceramic alumina membranes was prepared by extruding a ceramic dope solution (70%

119

alumina particle, 6.5% polysulfone, 2% PEG200, 0.5% BYK-190, 0.3% Mg(OH)2, 20.7%

120

NMP) using an extruder (FM-P20, Miyazaki Iron Works Co., Ltd., Japan) through a nozzle

121

(OD: 3.2 mm, ID: 2.2 mm). Deionized water was employed as the bore solution. The

122

extruded dope solution passed through an air gap of 10 cm then immersed in waterbath for

123

phase inversion. The prepared pristine membrane was washed thoroughly in 20%

124

EtOH/water for 24 hrs then sintered at 1450 oC for 2 hrs.

125

Membrane Characterization

126

Membrane water permeability was measured using a crossflow system at the transmembrane

127

pressure of 1 bar. The membrane separation performance was tested against 1000 ppm

128

polystyrene (PS) particle (30 nm and 100 nm particle size) colloid solution at 1 bar. The

129

permeate and retentate stream was collected separately and their PS concentrations were

130

measured using a UV spectrometer (UV-2401PC, Shimadzu, Japan) at 200 nm. The rejection

131

of PS particle was calculated using Equation 1.

R (%) = 1 −

132

(Eq. 1)

133

where R stands for membrane rejection, and Cp and CR represent permeate and retentate

134

concentrations, respectively.

135

Pore size distribution was characterized using a capillary porometer (CFP-1200, PMI, USA).


Page 7 of 22 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


136

Each sample was first wetted with Galwick solution then tested up to 15 bar. Membrane

137

morphology was observed using scanning electron microscopy (TM-3000, Hitachi, Japan).

138

Flue gas dehydration apparatus & membrane modules

139

Air dehydration experiments were carried out using the apparatus illustrated in Figure 3.

140

Ceramic membrane (KRICT100 and Hyflux20) was first potted into a hollow fiber module

141

and loaded onto the apparatus. The effective surface area of KRICT100 and Hyflux20

142

module was 53.2 cm2 and 71.6 cm2, respectively.

143

A humidified air stream was prepared by bubbling dry air through a heated waterbath. The

144

relative humidity was controlled by mixing with a separate dry stream. The flowrate of each

145

stream was controlled using a mass flow controller. The humidified stream was fed into the

146

shell side of the membrane module. The water vapor condenses on the outer surface of the

147

membrane and subsequently permeates through the membrane. The temperature and relative

148

humidity of the retentate stream was then measured using a hygrometer. For the experiments

149

to assess SO2 permselectivity, a separate SO2 stream (1000 µmol/mol N2) was mixed with the

150

feed stream to give 100 ppm SO2 concentration. The feed stream gauge pressure was

151

approximately 0.03 bar.

152

A cooling water stream was circulated through the bore side of the membrane module. A

153

slight negative pressure was applied (-0.3 bar) using a booster pump (LFP1150S, PS holding

154

Ltd., South Korea) by placing the pump downstream of the module. The flowrate of the

155

circulating water was controlled using a bypass line. The temperature of the water was

156

maintained at 20 oC using a chiller. A reservoir tank was placed on a balance that is connected

157

to a computer, and the increase in mass reading was continuously measured to calculate the

158

water flux. A calibrated pH probe was immersed in the reservoir tank to measure the change

159

of pH over time which was used to calculate the SO2 flux. A summary of experimental

160

parameters are shown in Table I

161

Table I. Summary of Dehydration Experimental Parameters Parameter

Set Point

Unit

Air Flowrate (Fair)

1000, 2000, 3000, 4000, 5000, 6000

sccm (STP)

Air Temperature (Tair)

60, 70, 80

Air Relative Humidity (Hair)

50, 80

%

Cooling Water Flowrate (Fwater)

0.9

L.min-1


o

C


Cooling Water Temperature (Twater)

20.0 ± 0.2

Differential Pressure (∆P)

-0.3

Page 8 of 22

o

C

bar

162 163

164 165 166 167

Figure 3. Dehydration Experiment Test Apparatus. Black lines indicate hot gas stream flows, blue lines indicate cold liquid stream flows. MFC – Mass Flow Controller, F – flowmeter, T – thermometer, H – hygrometer, P – pressure gauge.

168


Page 9 of 22 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


169

Results and Discussion

170

Transport Mechanisms

171

In order to optimize the membrane condenser (MC) performance, it is important to

172

clearly distinguish the difference between condensable vapors and non-condensable gases

173

when discussing membrane transport mechanisms. Contrary to non-condensable gases (e.g.

174

O2, N2), condensable vapors (e.g., H2O) can condense within small pores and block the

175

transport of other gases. According to the Kelvin’s Equation (Eq. 2), the saturation pressure

176

of condensable vapors decreases with decreasing pore size. Hence, the vapors condense more

177

readily within pores. This phenomenon is commonly referred to as capillary condensation

178

and many interesting works have applied this phenomenon to separate alcohol/water 25-26

24

,

27-28

179

olefin/paraffin

180

phenomenon, it is possible to selectively condense vapors within the pores, and subsequently

181

permeate the condensed liquid.

182

, and water vapors

. Exploiting this capillary condensation

ln = −2

(Eq. 2)

183

where ρ is the condensate density, R is the gas constant, T is the temperature, M is the

184

molecular weight of the condensable compound, Pc is the capillary condensation pressure, Po

185

is the vapor saturation pressure at a planar interface, σ is the interfacial tension, θ is the

186

contact angle, and rp is the membrane pore radius.

187

In addition to the transport mechanism, researchers have tried different MC

188

configurations, as summarized in Figure 2(b), Wang et al.8 and Zhao et al.18 have exploited

189

the condensation phenomenon to extract water using the transport membrane condenser

190

(TMC) configuration. On the other hand, Drioli et al.3, 17, 23 have employed hydrophobic

191

membrane to condense out water on the membrane surface while allowing dehydrated gas to

192

permeate (Figure 2(c)). Compared to the TMC configuration, employing a hydrophobic

193

membrane to induce surface condensation requires non-conventional module design to collect

194

the liquid retentate, and typically results in low quality water, similar to using heat

195

exchangers. Therefore, TMC configuration holds more potential to recover water in higher

196

purity with a simpler module design.

197

Before investigating the performance efficiency of membrane condensers, one must

198

carefully consider the thermodynamic aspects of the overall process. As illustrated in Figure 1,



199

the water vapor emitting from the cooling towers were intentionally evaporated to utilize the

200

latent heat to cool the exothermic stream. Therefore, one must ask whether it is

201

thermodynamically logical to re-condense the evaporated stream, which also requires

202

considerable amount of cooling energy. One plausible explanation is that since the evaporated

203

water has been distilled to some degree, the energy input can be justified if high quality water

204

can be harnessed. In addition, as proposed by Wang et al.,8 the heat of condensation of

205

evaporated water can be re-utilized to heat the boiler feed stream.

206

It should be emphasized that capturing the evaporated water must be approached from

207

the environmental perspectives, as minimizing water consumption is one of the top priorities

208

for power plants. Therefore, it is crucial to develop an energy efficient process for capturing

209

evaporated water to relieve the energy-water collisions.

210

Membrane Characterization

211

To avoid condensate-induced corrosion, the flue gas from stacks typically exit at an

212

elevated temperature (70 ~ 90 oC) in a saturated state. Although polymeric membranes

213

currently dominate the membrane market, the long-term thermal stability of polymer

214

membranes in humid conditions is not guaranteed29. On the other hand, ceramic membranes

215

hold distinct advantages such as excellent thermal and chemical stability over long term.

216

Therefore, ceramic membranes have been tested in this work as a proof-of-concept.

217

The in-house fabricated ceramic membrane (KRICT100) and commercial ceramic

218

membrane (Hyflux20) was characterized using SEM, porosimetry, and rejection tests,

219

summarized in Figure 4 and 5.


Page 10 of 22

Page 11 of 22 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


220 221

Figure 4. SEM images of KRICT100 and Hyflux20 membranes. Hyflux20 membrane has a

222

γ-alumina coating layer in the inner side.

223 224

Figure 5. Pore size distribution data of KRICT100 and Hyflux20 membranes.

225 226

Table 2. Characterization of Ceramic Hollow Fiber Membranes

Memb.

PWP (Lm-2h-1bar-1)

Max pore size (nm)

Mean pore size (nm)

OD (mm)

ID (mm)

KRICT100

395

208

90

2.78

2.12

0.33

-

83.1%

Hyflux20

329

44

39

3.87

2.91

0.48

90.1%

-

227


Thickness Rejection Rejection (mm) (PS 30 nm) (PS 100 nm)


228

It can be seen that the ceramic hollow fiber membranes show isotropic morphology

229

throughout the membrane thickness. In comparison to the KRICT100 membranes, Hyflux20

230

membrane is coated with a γ-alumina layer on the inner surface of the fiber. The mean and

231

maximum pore size of Hyflux20 was measured to be 40 nm, and 44 nm, respectively. And 30

232

nm polystyrene particle rejection test gave 90.1% rejection.

233

On the other hand, KRICT100 membrane exhibited the mean and maximum pore size of

234

90 nm and 209 nm, respectively. And 100 nm polystyrene particle rejection test gave 83.1%

235

rejection. Hence, the overall pore size of KRICT100 membranes is higher than that of

236

Hyflux20 membranes. These two membranes have been tested for dehydration experiments.

237

Flue gas dehydration

238 239

Figure 6. Transport membrane condenser (TMC) water flux as a function of dry air flowrate

240

for (a) KRICT100 and (b) Hyflux20 membranes.

241 242

Figure 7. Transport membrane condenser (TMC) performance and dehumidification


Page 12 of 22

Page 13 of 22 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


243

efficiency as a function of feed water vapor flowrate.

244

The dehydration data are summarized in Figures 6 and 7. Several clear trends are

245

visible from the experiments. First, the water flux increases proportionally with respect to the

246

increasing feed flowrate, temperature, and humidity. Such trend is expected as increase in

247

temperature and humidity increases vapor pressure gradient at the membrane outer surface.

248

A better correlation can be obtained by plotting the membrane flux against the feed

249

vapor flowrate. As shown in Figure 7, there is a strong linear relationship between the feed

250

vapor flowrate and water flux through the membrane. On the other hand, there is no clear

251

difference in performance between KRICT100 and Hyflux20 membranes, despite the

252

difference in the pore size. Such trend clearly implies that the performance is limited by the

253

rate of condensation, not by the rate of permeation (membrane performance). More

254

specifically, the basic mechanism of TMC can be seen as condensation followed by

255

permeation, which can be semi-qualitatively described using the resistance-in-series model

256

(Eq. 3).

257

J =

! "#$ ' ()*+ ∆&#$

∆& )*+

=

,

-#∆.

!

'

, - )*+ ∆(

(Eq. 3)

258

where Rcond and Rperm are condensation and permeation resistances, respectively, kcon and

259

kperm are rate constants for condensation and permeation, and ∆µ, ∆T, ∆P represent chemical

260

potential, temperature, and pressure gradients, respectively.

261

The main driving forces for water condensation are the temperature and vapor

262

pressure gradients between the membrane surface and the feed vapor stream, and the driving

263

force for liquid water permeation through the membrane is the transmembrane pressure. As

264

shown in Table 2, the permeability (kperm) of KRICT100 and Hyflux20 membranes are 329,

265

395 L.m-2.hr-1.bar-1, respectively, indicating that these membranes are not permeation-limited.

266

Hence, it can be deduced that the rate constant of condensation (kcon) is significantly lower

267

than the permeability (kperm), making the overall process condensation limited, as observed in

268

Figure 7. Therefore, controlling the temperature and vapor pressure gradients at the

269

membrane surface is crucial to maximize the TMC performance.

270

Thirdly, the dehumidification efficiency plot exhibits a saturation trend, as shown in

271

Figure 7(b). It can be seen that using the current membrane module, the maximum

272

dehumidification efficiency approaches 85%. As the stream gets dehumidified within the



Page 14 of 22

273

membrane module, the driving force (vapor pressure gradient) for further dehumidification

274

proportionally decreases, plateauing the dehumidification efficiency at 85%. This trend

275

suggests the importance of module design and a delicate control of driving force (vapor

276

pressure gradient)

277

The performance obtained in this work is compared against other membrane-based

278

dehydration processes in Table 3. It should be noted that performance comparison between

279

different processes can be subjective; however, the overall process productivity can be

280

qualitatively compared in terms of dehumidification efficiency.

281

Table 3. Literature Data for Gas Stream Dehydration Year

Material

Pore Size

Mode

2008

P(AA-AMPS)-PVA

Dense

Vapor diffusion, pressure

2008

sPEEK

Dense

Vapor diffusion, vacuum

2008

PVA

Dense

Vapor diffusion, sweep

2012

PEBAX

Dense

2005

Ethyl Cellulose

2005

Application Propylene dehyd. Flue gas dehyd.

Highest Flux (kg/m2/hr)

Selectivity

Ref 30

0.03 1

H2O/N2 > 107

4

Air dehyd.

0.2

N/A

31


CH4 dehyd.

0.03

H2O/CH4 > 1500

Dense


N2 dehyd.

0.25

9

Polysulfone

Dense


N2 dehyd.

0.02

9

2005

PEO-PBT

Dense


N2 dehyd.

0.17

9

2012

Ceramic membrane

6 - 8 nm

TMC, water cooling

Flue gas dehyd.

4.5

N/A

8

2015

Ceramic membrane

20 nm

TMC, water cooling

Flue gas dehyd

22

N/A

18

2017

Ceramic membrane

20 nm

TMC, water cooling

Flue gas dehyd

7.5

N/A

20

2017

α-Al2O3 ceramic

20 - 100 nm

TMC, water cooling

Flue gas dehyd.

12

H2O/SO2 > 100

This work

32

282 283

As summarized in Table 3, most of the dehydration work have been performed using

284

dense vapor separation membranes with outstanding permeability. However, as mentioned

285

previously, emissions from power plants are high temperature streams (70 ~ 90 oC) near

286

atmospheric pressure, requiring either compression or vacuum to provide the necessary

287

driving force for vapor permeation. As expected, these high performance membranes with

288

outstanding permeability are completely limited by pressure ratio, resulting in poor

289

performance in real scales 4. On the other hand, TMC configuration can exploit the

290

temperature gradient to selectively condense out water vapors in high efficiency (85% in this

291

work). In addition, it can be seen in Table 3 that TMC membrane performance outperforms

292

the dense vapor separation membranes, from one order of magnitude to as high as three

293

orders of magnitude.


Page 15 of 22 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


294

Apart from the dehydration performances, emissions from power plant stacks contain

295

small percentage of SOx and NOx compounds, even after the flue gas desulfurization (FGD)

296

unit. According to Wang et al. 8 the condensate stream from TMC process is pure enough to

297

be fed directly into the boiler feed water. It was claimed that capillary condensation prevents

298

permeation of SOX species; however, it should be noted that SOX species immediately get

299

oxidized under the presence of O2 and exists as a sulfurous acid (and sulfuric acid) by

300

accepting a proton from water vapor. Hence, it is yet questionable whether the mode of

301

separation is by capillary condensation, or simply a competition between condensation (water)

302

and absorption (SOx).

303 304

Figure 8 summarizes the water and SOx flux through Hyflux20 membranes in TMC configuration. The selectivity (α) is calculated using Eq. 4. 2345)* 76 α01 = 289 345)* :; 689:

305

(Eq. 4)

306

It can be seen that water to SOx selectivity is approximately 100 independent of the vapor

307

flowrate, which is, interestingly, much higher than the simple Knudsen selectivity of 1.8

308

(square root of molecular weight ratio). With the measured water/SO2 selectivity of 100, the

309

overall process must be carefully designed to match the product stream quality. For instance,

310

the residence time for the cooling water stream must not exceed the point at which the

311

product quality is irreversibly compromised. It has been hypothesized that the membrane

312

pore size can affect the membrane SO2 selectivity 8; however, it is likely that SO2 flux is a

313

strong function of the liquid-gas interface area, hence it is necessary to induce vapor

314

condensation within the pores to form a meniscus. Nevertheless, the TMC selectivity against

315

SO2 still remains to be validated and the key parameters needs to be found to maximize the

316

selectivity.



317 318

Figure 8. Water and SO2 flux through Hyflux20 membrane as a function of water vapor

319

flowrate, selectivity (αTMC) value is written above each flux data.

320 321

Transport Membrane Condenser Heat Balance Profile

322

As shown in Section 3.3, the current membrane condenser technology is almost entirely

323

limited by the rate of condensation, which is a strong function of temperature gradient

324

between the membrane surface and the vapor feed stream. In order to design an effective

325

membrane and module, it is first necessary to understand the key parameters that determine

326

the overall heat transfer coefficient.

327

To model the temperature profile of the membrane and the system, a steady state energy

328

balance can be made around each fiber (See Supporting Information for detailed modeling

329

calculations). For analysis, the heat transfer coefficient at the air-membrane interface (outer

330

fiber surface) was estimated using the Colburn correlation, and the heat transfer coefficient at

331

the membrane-water interface (inner fiber surface) was estimated using the Sieder-Tate

332

correlation.


Page 16 of 22

Page 17 of 22 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


333 334 335 336

Figure 9. (a) Calculated membrane outer surface temperature as a function of feed air temperature for polymeric and ceramic membranes; (b) membrane temperature profile along the fiber thickness.

337

In order to maximize the driving force (temperature and vapor pressure gradient), it is

338

necessary to maintain a wide temperature gap between the feed stream and the membrane

339

outer surface temperature. Therefore, it is desired to keep the membrane outer surface

340

temperature as low as possible. Figure 9(a) clearly illustrates the effect of material thermal

341

conductivity on the membrane outer surface temperature. Assuming a membrane porosity of

342

70%, ceramic membranes (alumina) with high thermal conductivity (kalumina = 35 W.m-1.K-1)

343

can effectively maintain low surface temperature compared to typical polymeric membranes

344

(kPVDF = 0.19 W.m-1.K-1). Figure 9(b) illustrates the effect of feed temperature on the

345

temperature profile across the membrane cross-section. It can be seen that polymeric

346

membranes exhibit steeper temperature gradient along the thickness compared to ceramic

347

membranes, primarily due to the low thermal conductivity of the material itself.

348

Therefore, from the performance perspective, it certainly is more effective to utilize

349

ceramic membranes for membrane condenser applications. However, ceramic membranes are

350

brittle, rendering them difficult to handle in large scale. On the other hand, polymeric

351

membranes exhibit relatively low thermal stability but can be more cost effective. Since

352

membrane condenser process is a relatively new technology, it is yet unclear which

353

membrane material can give economic and process advantage in large scale, and both options

354

are being explored. Nevertheless, the main goal is to maintain low surface temperature to

355

maximize the driving force.



356 357

Figure 10. Calculated membrane surface temperature as a function of thickness and porosity,

358

for (a) polymeric membrane; and (b) ceramic membrane.

359

When fabricating membranes, apart from tailoring the pore size, independent parameters

360

under control include membrane dimensions (diameter and thickness) and porosity. Figure 10

361

illustrates the effect of these parameters on the membrane condenser performance (outer

362

surface temperature) for polymeric and ceramic membranes. As expected, it can be seen that

363

the membrane temperature increases with increasing thickness for both materials.

364

Interestingly, it was found that the two studied materials give opposite trends as a

365

function of porosity. For polymeric materials, membranes with higher porosity exhibit lower

366

membrane temperature. On the other hand, for ceramic membranes, low porosity display

367

lower membrane temperature. Such opposite trends results from the assumption that the open

368

pores are filled with water during membrane condenser operation, and water has a thermal

369

conductivity (k = 0.67 W.m-1.K-1) between that of ceramic and polymeric material.

370

The trends observed in Figures 9 and 10 can give an important direction to tailor the

371

membrane characteristics to improve the membrane condenser productivity. For polymeric

372

membranes, it is desirable to maximize the membrane porosity while reducing the thickness.

373

It should be noted that as membrane condenser operates at relatively low transmembrane

374

pressures, the mechanical integrity is not expected to be a significant issue. For ceramic

375

membranes, lower porosity is preferred yet has negligible effect on the membrane

376

temperature due to its high thermal conductivity. Instead, more focus can be placed on

377

controlling the membrane pore size to improve the condensed water quality.


Page 18 of 22

Page 19 of 22 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


378

With the escalating water shortage and exacerbating energy-water collisions, reducing the

379

water consumption from power plants is becoming an important topic from both

380

environmental and economic perspectives. Based on this work, it is envisaged that membrane

381

condenser technology can play a unique role to harness clean water from flue gas emissions,

382

and further research on large scale process feasibility will be the topic of our subsequent work.

383 384

Conclusions

385

In this work, an effective method to harness clean water from power plant emissions using

386

membrane condenser technology is proposed. Compared to dense vapor separation

387

membranes that suffer from low driving force, the proposed transport membrane condenser

388

(TMC) configuration exhibited water flux up to 12 kg.m-2.hr-1, as high as three orders of

389

magnitude higher compared to the vapor separation membranes. In addition, the TMC

390

process gave a reasonable water/SOX selectivity of 100, which is much higher than the

391

Knudsen selectivity of 1.8. It was determined that the current TMC process is completely

392

limited by the rate of condensation, and a better membrane and more effective module design

393

must be developed that enhances the vapor pressure gradient. The current limit of

394

dehumidification efficiency was determined to be approximately 85%, after which the driving

395

force cannot be maintained to induce water condensation. It is envisaged that further

396

dehydration can be achieved by controlling the membrane pore size to induce capillary

397

condensation. The exact nature of water/SOx selectivity has not been fully elucidated, but it is

398

speculated the SOX permeation is a strong function of contact surface area at which SOX can

399

absorb. In order to utilize the condensate product for boiler feed water, the process design

400

must minimize the cooling water residence time while maximizing the condensation rate.

401

This work showed a potential to apply TMC process to harness clean water form power plant

402

emissions, but further membrane development, preferably with polymeric materials, is

403

necessary to suit the TMC requirements.

404

Supporting Information

405

Supporting Information includes 7 equations with detailed heat balance simulations.

406

Acknowledgement

407

This project was supported by the R&D program of the institutional research program of



408

Korea Research Institute of Chemical Technology (KK1802-B00 and BS.K18M503)

409

References

410

1.

411

Macknick, J.; Madden, N.; Meldrum, J. Water-smart power: Strengthening the US electricity system

412

in a warming world; 2013.

413

2.

414

Nederland BV: 2013.

415

3.

416

streams: from environmental issue to source of water by using membrane condensers. CLEAN

417

2014, 42 (8), 1145-1153. D.O.I. 10.1002/clen.201300104

418

4.

419

dehydration using polymer membranes. J. Membr. Sci. 2008, 313 (1), 263-276. D.O.I.

420

10.1016/j.memsci.2008.01.024

421

5.

422

water from boiler flue gas. US Department of Energy 2008. D.O.I. 10.2172/952467

423

6.

424

flue gas; University Of North Dakota: 2006.

425

7.

426

membrane for harvesting water from the atmosphere. Environmental Science: Water Research &

427

Technology 2015, 1 (6), 753-760. D.O.I. 10.1039/C5EW00098J

428

8.

429

recovery. Applied Energy 2012, 91 (1), 341-348. D.O.I. 10.1016/j.apenergy.2011.10.003

430

9.

431

and inert gas mixtures through highly selective and highly permeable polymer membranes. J.

432

Membr. Sci. 2005, 251 (1), 29-41. D.O.I. 10.1016/j.memsci.2004.08.036

433

10.

434

polybutadiene membranes and their use in drying alcohols by pervaporation and vapor

435

permeation technologies. ACS Sustainable Chemistry & Engineering 2016, 4 (8), 4442-4450. D.O.I.

436

10.1021/acssuschemeng.6b01072

437

11.

438

separation processes. J. Membr. Sci. 2014, 463, 33-40. D.O.I. 10.1016/j.memsci.2014.03.016

439

12.

440

of a hydrophobic microporous membrane. J. Membr. Sci. 2000, 175 (1), 35-42. D.O.I.

441

10.1016/S0376-7388(00)00404-X

442

13.

443

scale-up of direct contact membrane distillation modules for seawater desalination. Desalination

444

2017, 402, 72-87. D.O.I. 10.1016/j.desal.2016.08.025

445

14.

Rogers, J.; Averyt, K.; Clemmer, S.; Davis, M.; Flores-Lopez, F.; Frumhoff, P.; Kenney, D.;

Daal, L. Capture of Evaporated Water with Novel Membranes (CAPWA FP7-NMP); KEMA Brunetti, A.; Santoro, S.; Macedonio, F.; Figoli, A.; Drioli, E.; Barbieri, G., Waste gaseous

Sijbesma, H.; Nymeijer, K.; van Marwijk, R.; Heijboer, R.; Potreck, J.; Wessling, M., Flue gas

Levy, E.; Bilirgen, H.; Jeong, K.; Kessen, M.; Samuelson, C.; Whitcombe, C., Recovery of Folkedahl, B. C.; Weber, G. F.; Collings, M. E. Water extraction from coal-fired power plant Roy, S.; Hussain, C. M.; Mitra, S., Carbon nanotube-immobilized super-absorbent

Wang, D.; Bao, A.; Kunc, W.; Liss, W., Coal power plant flue gas waste heat and water Metz, S.; Van de Ven, W.; Potreck, J.; Mulder, M.; Wessling, M., Transport of water vapor

Vane, L.; Namboodiri, V.; Lin, G.; Abar, M.; Alvarez, F., Preparation of water-selective

Huang, Y.; Merkel, T. C.; Baker, R. W., Pressure ratio and its impact on membrane gas Ito, A., Dehumidification of air by a hygroscopic liquid membrane supported on surface

Dong, G.; Kim, J. F.; Kim, J. H.; Drioli, E.; Lee, Y. M., Open-source predictive simulators for

Roy, S.; Humoud, M. S.; Intrchom, W.; Mitra, S., Microwave-Induced Desalination via Direct


Page 20 of 22

Page 21 of 22 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


446

Contact Membrane Distillation. ACS Sustainable Chemistry & Engineering 2017, 6 (1), 626-632.

447

D.O.I. 10.1021/acssuschemeng.7b02950

448

15.

449

J. F.; Drioli, E., Thermally rearranged polymer membranes for desalination. Energy Environ. Sci.

450

2016, 9 (3), 878-884. D.O.I. 10.1039/C5EE03768A

451

16.

452

Cellulose Microbeads via Membrane Emulsification. ACS Sustainable Chemistry & Engineering

453

2017, 5 (7), 5931-5939. D.O.I. 10.1021/acssuschemeng.7b00662

454

17.

455

technology for water recovery from humidified “waste” gaseous streams. Ind. Eng. Chem. Res.

456

2012, 52 (3), 1160-1167. D.O.I. 10.1021/ie203031b

457

18.

458

and heat recovery from gaseous streams: Performance evaluation. J. Membr. Sci. 2015, 484, 10-17.

459

D.O.I. 10.1016/j.memsci.2015.03.007

460

19.

461

water recovery from flue gas by membrane condensation: Experimental investigation. Appl. Therm.

462

Eng. 2017, 113, 843-850. D.O.I. 10.1016/j.applthermaleng.2016.11.101

463

20.

464

experiments of flue gas with using nanoporous ceramic membranes. Appl. Therm. Eng. 2017, 110,

465

686-694. D.O.I. 10.1016/j.applthermaleng.2016.08.191

466

21.

467

recovery from flue gas using hollow micro–nano porous ceramic composite membranes. Journal

468

of Industrial and Engineering Chemistry 2018, 57, 349-355. D.O.I. 10.1016/j.jiec.2017.08.042

469

22.

Macedonio, F.; Brunetti, A.; Barbieri, G.; Drioli, E., Membrane condenser configurations for

470

water

recovery

471

10.1016/j.seppur.2017.03.009

472

23.

473

membrane preparation for recovery of humidified gas streams using membrane condenser. React.

474

Funct. Polym. 2014, 79, 1-7. D.O.I. 10.1016/j.reactfunctpolym.2014.03.003

475

24.

476

membrane. J. Chem. Eng. Jpn. 1986, 19 (1), 72-77. D.O.I. 10.1252/jcej.19.72

477

25.

478

membranes. Part I. Multilayer diffusion and capillary condensation. J. Membr. Sci. 1992, 66 (2-3),

479

259-269. D.O.I. 10.1016/0376-7388(92)87016-Q

480

26.

481

membranes. Part II. Synthesis and separation properties of microporous membranes. J. Membr. Sci.

482

1992, 66 (2-3), 271-287. D.O.I. 10.1016/0376-7388(92)87017-R

483

27.

484

Vycor glass membrane. J. Colloid Interface Sci. 1986, 110 (2), 544-555. D.O.I. 10.1016/0021-

Kim, J. H.; Park, S. H.; Lee, M. J.; Lee, S. M.; Lee, W. H.; Lee, K. H.; Kang, N. R.; Jo, H. J.; Kim,

Coombs OBrien, J.; Torrente-Murciano, L.; Mattia, D.; Scott, J. L., Continuous Production of

Macedonio, F.; Brunetti, A.; Barbieri, G.; Drioli, E., Membrane condenser as a new

Wang, T.; Yue, M.; Qi, H.; Feron, P. H.; Zhao, S., Transport membrane condenser for water

Zhao, S.; Yan, S.; Wang, D. K.; Wei, Y.; Qi, H.; Wu, T.; Feron, P. H., Simultaneous heat and

Chen, H.; Zhou, Y.; Cao, S.; Li, X.; Su, X.; An, L.; Gao, D., Heat exchange and water recovery

Chen, H.; Zhou, Y.; Su, X.; Cao, S.; Liu, Y.; Gao, D.; An, L., Experimental study of water

from

waste

gases.

Sep.

Purif.

Technol.

2017,

181,

60-68.

D.O.I.

Drioli, E.; Santoro, S.; Simone, S.; Barbieri, G.; Brunetti, A.; Macedonio, F.; Figoli, A., ECTFE

ASAEDA, M.; DU, L. D., Separation of alcohol/water gaseous mixtures by thin ceramic Uhlhorn, R.; Keizer, K.; Burggraaf, A., Gas transport and separation with ceramic

Uhlhorn, R.; Keizer, K.; Burggraaf, A., Gas transport and separation with ceramic

Lee, K.-H.; Hwang, S.-T., The transport of condensible vapors through a microporous



485

9797(86)90407-8

486

28.

487

condensation effects on mass transport through porous membranes. Sep. Purif. Technol. 2003, 33

488

(3), 273-281. D.O.I. 10.1016/S1383-5866(03)00087-X

489

29.

490

on the gas separation performance of polymeric membranes for carbon dioxide capture. Energy

491

Procedia 2009, 1 (1), 311-317. D.O.I. 10.1016/j.egypro.2009.01.043

492

30.

493

composite hollow fiber membranes for propylene dehumidification. J. Membr. Sci. 2008, 323 (2),

494

395-403. D.O.I. 10.1016/j.memsci.2008.06.046

495

31.

496

PVA/LiCl blend membrane for air dehumidification. J. Membr. Sci. 2008, 308 (1), 198-206. D.O.I.

497

10.1016/j.memsci.2007.09.056

498

32.

499

K. A.; Merkel, T. C., Dehydration of natural gas using membranes. Part I: Composite membranes. J.

500

Membr. Sci. 2012, 413, 70-81. D.O.I. 10.1016/j.memsci.2012.04.009

Uchytil, P.; Petrickovic, R.; Thomas, S.; Seidel-Morgenstern, A., Influence of capillary

Scholes, C. A.; Kentish, S. E.; Stevens, G. W., The effect of condensable minor components

Pan, F.; Jia, H.; Jiang, Z.; Zheng, X.; Wang, J.; Cui, L., P (AA-AMPS)–PVA/polysulfone

Zhang, L.-Z.; Wang, Y.-Y.; Wang, C.-L.; Xiang, H., Synthesis and characterization of a

Lin, H.; Thompson, S. M.; Serbanescu-Martin, A.; Wijmans, J. G.; Amo, K. D.; Lokhandwala,

501 502

For Table of Contents Use Only

503 504

The currently unsustainable water and energy loss from power plants can be recovered using

505

membrane condenser technology


Page 22 of 22

Harnessing Clean Water from Power Plant Emissions Using

Recommend Documents