Dd

DO

a

ARRAA

KDHSPP

1

aorscwtoTttsta

se

0d

Journal of Membrane Science 389 (2012) 380– 388

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science

j ourna l ho me pag e: www.elsev ier .com/ locate /memsci

esalination of brine and produced water by direct contact membraneistillation at high temperatures and pressures

hananjay Singh, Kamalesh K. Sirkar ∗

tto H. York Department of Chemical, Biological and Pharmaceutical Engineering, New Jersey Institute of Technology, Newark, NJ 07102, United States

r t i c l e i n f o

rticle history:eceived 6 August 2011eceived in revised form 25 October 2011ccepted 1 November 2011vailable online 7 November 2011

eywords:irect contact membrane distillationigh temperature and pressureteam-assisted gravity drainage processroduced water

a b s t r a c t

Direct contact membrane distillation (DCMD)-based desalination process is a thermally driven separationprocess where a hydrophobic microporous membrane separates a hot brine feed and a cold distillatewhich condenses the water vapor from the hot brine passing through the membrane pores. So far, DCMDhas been explored for hot brines and other aqueous solutions below 100 ◦C. For feed solutions above100 ◦C, DCMD has an extra advantage over other conventional separation processes like reverse osmosis(RO) which requires cooling of the feed solution requiring additional energy; further RO cannot utilize theheat available in the feed solution. Produced water obtained from steam assisted gravity drainage (SAGD)process is one such example where DCMD can potentially be a very useful process. In this paper, theDCMD technique is explored in the range of 80–130 ◦C for brines containing 10,000 ppm sodium chloridewith porous flat sheet polytetrafluoroethylene (PTFE) membranes. The pressure of the solution went up

olytetrafluoroethylene membrane to 2–3 atm. The performance of this membrane has also been explored with a saline feed containing3000 ppm NaCl, 45 ppm phenol, 45 ppm cresol and 10 ppm naphthenic acid, simulating the compositionof hot produced water obtained from the SAGD process. There was no leakage of salt under any conditions.The highest water vapor flux achieved was 195 kg/m2 h which is a few times larger than that for seawaterRO process. The water generated by the DCMD process may be used for steam generation in the SAGDprocess.

. Introduction

Membrane distillation (MD) process for removing water from anqueous solution of nonvolatile solutes involves passing hot aque-us solution on one side of a porous hydrophobic membrane andemoving water vapor passing through the pores from the otheride. In direct contact membrane distillation (DCMD) (Fig. 1), aold distilled water stream passing on the other side collects theater vapor by condensing it [1]. In vacuum membrane distilla-

ion, a vacuum existing on the other side of the membrane allowsne to withdraw the water vapor and condense it elsewhere [2].here are additional MD techniques such as air gap membrane dis-illation [3] where a cold condensing surface near the other side ofhe membrane condenses the water vapor directly. One can have aweep air stream on the other side of the membrane as well prioro condensation of the vapor in a separate condenser [4]. There are

number of advantages and disadvantages of each process.

Of these processes, the DCMD has been investigated exten-

ively by a number of groups [5–11]. Most groups havemployed commercially available hydrophobic flat films of

∗ Corresponding author. Tel.: +1 973 596 8447; fax: +1 973 642 4854.E-mail address: Sirkar@adm.njit.edu (K.K. Sirkar).

376-7388/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.memsci.2011.11.003

© 2011 Elsevier B.V. All rights reserved.

polytetrafluoroethylene, polyvinylidene fluoride and polypropy-lene as well as hollow fibers of polypropylene. We have extensivelystudied a new membrane structure consisting of a porous fluorosil-icone coating on a porous polypropylene hollow fiber substratehaving a high wall thickness of 150 �m [2,8–10,12–14]. We havedemonstrated that such a hollow fiber structure when employedwith cross flow of the hot brine over the fiber outside surface andcold distillate flow through the fiber bore yields extraordinary per-formance in terms of high flux, no pore wetting and reproducibleand scalable performance including pilot plant scale concentrationof sea water [10] to the level of ∼20% salt. A countercurrent cascadeof such hollow fiber membrane devices has demonstrated encour-aging energy efficiency in terms of GOR (gained output ratio) values[15,16].

However, DCMD investigators have generally limited their stud-ies to hot saline and other solutions at temperatures less than100 ◦C. Water vapor pressure however increases almost exponen-tially as temperature rises. Should the aqueous solution be availableat temperatures greater than 100 ◦C, we can achieve considerablyhigher water vapor fluxes. Since water vapor flux is driven by water

vapor pressure difference on two sides of the membrane, this couldbe an added advantage for DCMD vis-à-vis reverse osmosis fordesalting aqueous streams available at >100 ◦C since RO has notbeen implemented at such high temperatures; cooling down the

D. Singh, K.K. Sirkar / Journal of Membr

fopee

mas2tsipsfb

2

2

Paa6b(Nsoav

Fig. 1. Direct contact membrane distillation (DCMD) process.

eed solution is a prerequisite to RO. The inherent thermal energyf the hot feed stream could be a boon to the energy required for therocess. The above 100 ◦C membrane operation is demanding andliminates porous substrates like polypropylene which has beenxtensively used for lower temperature DCMD operation.

We have therefore studied here the DCMD performance oficroporous films of polytetrafluoroethylene (PTFE) subjected to

queous feeds at temperature in the range of 80–130 ◦C; corre-pondingly the pressures of the feed solution rises to the range of–3 atm. The pore sizes of the PTFE films are considerably smallerhan those used in conventional DCMD processes. The feed aqueousolution contained primarily salt at the level of 10,000 ppm. Exper-ments have also been done with saline feeds containing 45 ppmhenol, 45 ppm cresol and 10 ppm naphthenic acid. Such a compo-ition simulates the higher temperature produced water obtainedrom steam assisted gravity drainage (SAGD) process in oil–shaleased energy explorations [17].

. Experimental

.1. Approach

To study the DCMD behavior, experiments were performed withTFE flat sheet membranes in lower temperature range (<100 ◦C)s well as higher temperature range (>100 ◦C). All higher temper-ture experiments were performed in a large cell (Fig. 2) having

mm thick porous stainless steel disc to support the PTFE mem-rane. This large cell was originally used as a pervaporation cellmodel PTC-6, Carbone Lorraine, Salem, VA) and obtained from GFT,eunkirchen-Heinitz, Germany. On one side of the membrane, hot

olution of 1% (by wt) NaCl was introduced as the brine feed; thether side of the membrane was exposed to deionized water at

low temperature as the distillate stream condensing the waterapor. The flow rate used on either side of the membrane was

ane Science 389 (2012) 380– 388 381

generally high to reduce temperature polarization if any. With aflow rate of around 500 cc/min on the distillate side, the water vaporproduction rate was determined from a change in this flow rate viatwo flow meters (Fig. 3). Our earlier studies involved cells usingmuch smaller surface area; we had problem obtaining reproducibledata. Therefore we needed a large enough membrane surface areato generate enough vapor flux and employed the large cell. The gas-ket diameter of the cell was 12.8 cm and the effective membranearea was 128.6 cm2.

2.2. Materials and chemicals

We used the following materials and chemicals: PTFE flat sheetmembrane of 0.03 �m pore size (W. L. Gore and Associates), sodiumchloride (Sigma Aldrich), phenol (Sigma Aldrich), cresol (SigmaAldrich), o-phosphoric acid (HPLC grade from Fischer Scientific),methanol (HPLC grade from Fischer Scientific), naphthenic acid(Sigma Aldrich) and N-methyl-N-(tert.-butyldimethylsilyl) trifluo-roacetamide which contains 1% t-BDMS-chloride (MTBSTFA, Regis,Morton Grove, IL), chloroform (Sigma Aldrich), dichloromethane(Fischer Scientific) and hydrochloric acid (Fischer Scientific).

2.3. Experimental procedure

2.3.1. Gas permeation testGas permeation test was performed to measure the mean pore

size and effective surface porosity over the effective pore lengthof the porous membrane. Yasuda et al. [18] suggested a methodto determine the volumetric porosity of porous membranes. Later,Teo’s group [19,20] introduced a modified gas permeation methodto measure the mean pore size and the effective surface porosityover the effective pore length of an asymmetric membrane. Thetotal molar gas permeation flux per unit transmembrane pressuredifference across the porous membrane, (Ji/�p) is described by Eq.(1) where the first term represents the contribution from Knudsenflow and the second term is due to Poiseuille flow:

Ji�p

= 23

(8RT

�M

)0.5 1RT

rε

Lp+ p̄

8�iRT

r2ε

Lp(1)

Here ε is surface porosity, r is mean pore radius of the mem-brane, �i is gas viscosity, R is gas constant, p̄ is the mean pressure(average of feed and permeate side pressure), M is molecular weightof gas, Lp is effective pore length and T is temperature (K).

The gas permeation flux per unit driving force (Ji/�p) is calcu-lated as follows:

Ji�p

= Nt,i

At(2)

Here Nt,i is total molar gas permeation rate (mol s−1), �p is thetransmembrane pressure difference across the membrane area At.The total gas permeation rate through the membrane at differentpressures was measured using a soap bubble flow meter for lowflow rates of nitrogen and by another flow meter (Hewlett Packard)for high flow rates. From a plot of the nitrogen flux as (Ji/�p) againstthe mean gas pressure p̄, the mean pore size (r) and the effectivesurface porosity over pore length, ε/Lp, can be obtained from theslope So and the intercept Io as follows:

r = 163

(IoSo

) (8RT

�M

)0.5�i (3)

ε

Lp= 8�iRTSo

r2(4)

2.3.2. DCMD experimentThe experimental apparatus employed to study the DCMD

behavior for PTFE flat sheet membrane in the lower temperature

382 D. Singh, K.K. Sirkar / Journal of Membrane Science 389 (2012) 380– 388

l, (c) b

r[

sueCsdGtoo(ccFwc

t(tdt

Fig. 2. Photographs of (a) large cell used for experiments, (b) top of the cel

ange has been shown in earlier publications from our group12–14].

The schematic of the higher temperature/higher pressure DCMDetup is shown in Fig. 3. Brine from the brine tank (Alloys Prod-ct Corp, Waukesha WI) was pumped to the shell side of the heatxchanger (Titanium Fabrication Corp., Fairfield NJ) by a pump (‘QD’erampump, Fluid Metering, Syosset, NY) where it was heated byteam circulated on the tube side of the heat exchanger. Steam waselivered from a steam generator (MBA3A Sussman Electric Steamenerator, Long Island, NY) at different pressures corresponding to

he temperature of the feed solution. The heated brine was fed tone side of the PTFE membrane in the DCMD cell; on the other sidef the membrane, cold distilled water was circulated by a pump‘QD’ Cerampump, Fluid Metering, Syosset, NY). After the DCMDell, hot brine was recycled to the brine tank; distilled water wasirculated to the shell side of another heat exchanger (Titaniumabrication Corp., Fairfield NJ), where it was cooled down by coldater from a chiller (CH3000 Series Chiller, REMCOR, Anoka, MN)

irculated on the tube side of the heat exchanger.The cooled distilled water from the heat exchanger was recycled

o the distillate tank which was fitted with a liquid level controller

Low-Amp Liquid-Level Switch, McMaster-Carr, Robbinsville, NJ)o avoid overflow of water from the tank due to continuous con-ensation of water vapor in the distilled water stream acrosshe membrane from the hot brine feed solution. This liquid level

ottom of cell and (d) 6 mm thick stainless steel support for the membrane.

controller activated a pump (‘QD’ Cerampump, Fluid Metering Inc.,Syosset, NY) to take out extra water from the distillate tank into themakeup water storage tank which was also fitted with another liq-uid level controller activating the water refill pump to feed waterinto the brine tank for maintaining constant salt concentration inthe brine tank. Inlet and outlet temperatures of brine and distilledwater streams through the test cell were monitored by thermocou-ples (EW-08516-74, Cole-Parmer, Vernon Hills, IL) connected to adata acquisition logger (SPER SCIENTIFIC, Thermometer 4 channel,Cole-Parmer Instrument Co.). The flow rates of distillate in and dis-tillate out streams were monitored by two different flow meters(JV-KG series positive displacement flow meter, A W Lake Com-pany, Franksville, WI). PFA tubing and different types of PFA fittings(Cole-Parmer) were used to make connections in the setup.

Any experiment under given conditions was run for around 3 h.Each experiment was repeated three times to check the repro-ducibility.

2.3.3. Permeate analysis2.3.3.1. HPLC analysis for measurement of phenol and cresol on per-

meate side. Concentrations of phenol and cresol in the distillatetank were determined by HPLC (Water Alliance HPLC, Waters 2690Separation Module) having a C185U Econosphere column of length15 mm and a UV detector at �max = 218 nm. The mobile phase used

embrane Science 389 (2012) 380– 388 383

wr

2cAtiaM

2(DfiHawfosbwapwnsdeNd

Table 1Details of the used PTFE flat sheet membrane.

Material PTFE (W. L. Gore)Type Flat sheetNominal pore size (�m) 0.03Thickness (�m) 24Porosity (%) 65Effective porosity over pore length (ε/Lp) calculated from

N2 gas permeation (m−1)1.7 × 104

Tortuosity factor (calculated from N2 gas permeation) 1.58Tortuosity factor (calculated from theoretical model) 1.54

D. Singh, K.K. Sirkar / Journal of M

as 90/10 (v/v) methanol/water and 0.1% H3PO4 (v/v/) at a flowate of 1 ml/min.

.3.3.2. Atomic absorption spectrometer for measurement of sodiumhloride on permeate side. Permeate side samples were analyzed bytomic Absorption (AA) (AAnalyst 400, Atomic Absorption Spec-

rometer, Perkin Elmer, Norwalk, CT-06859) for sodium chloride;n addition the conductivity on distillate side was measured by

conductivity meter (Orion 115A+, Thermo Electron Corporation,A).

.3.3.3. Gas chromatography–electron impact-mass spectrometryGC–EI-MS) for measurement of naphthenic acid on the permeate side.ue to limitations of GC–EI-MS, naphthenic acids were extracted

rom water samples obtained from the distillate tank. Accord-ng to the method suggested by Merlin et al. [21], concentratedCl was first added to the water sample to maintain the pH at 2nd later 150 g NaCl was dissolved into the sample. This sampleas extracted three times with 60 ml of chloroform. The chloro-

orm phase was subjected to extraction three times with 10 mlf an alkaline aqueous solution of 4% (w/v) Na2CO3. The alkalineolution was acidified with concentrated HCl to pH 2 and free car-oxylic acids from alkaline solution were extracted three timesith 10 ml of dichloromethane (DCM). The extracted naphthenic

cid with DCM was dried under nitrogen and the remaining sam-le was transferred to fresh DCM. The extracted naphthenic acidsere derivatized by adding 100 �l of the MTBSTFA in 100 �l of aaphthenic acid standard (5 mg/ml) in dichloromethane [22]. Theamples were next heated at 60 ◦C for 20 min to get the t-BDMS

erivative. Gas chromatography–electron impact-mass spectrom-try (GC–EI-MS) (HP 6890 Series GC system with Agilent 5973etwork Mass Selective Detector) was used next for analysis of theerivatized samples prepared by the above method.

Fig. 3. Schematic diagram for higher tempera

Evaporation efficiency at 120 ◦C based on distillate streamin ideal conditions

∼80%

3. Results and discussion

3.1. Characterization of PTFE flat sheet membrane

PTFE flat sheet membranes used for experiments were charac-terized by scanning electron microscopy. It is clear from Fig. 4(a)and (b) that both sides of the membrane have a similar morphologyillustrating the symmetric nature of the membrane. Nitrogen gaspermeation flux through the flat sheet membrane was measured atdifferent pressures and plotted as (Ji/�p) against the mean pressurep̄. From the intercept and slope of the line, the effective porosity(ε/Lp) for the flat sheet membrane was found to be 1.77 × 104 m−1.The value of tortuosity factor was calculated to be 1.58 taking intoaccount the value for porosity of PTFE membrane (ε = 0.65) from themanufacturer (Table 1); it is also close to the value calculated fromfractal theories of random walks [23] which predicts a tortuosityfactor (�) as follows:

1

� =

ε(5)

Here ε is the porosity of the membrane; from this method thetortuosity factor comes out to be 1.54.

ture and higher pressure DCMD setup.

384 D. Singh, K.K. Sirkar / Journal of Membrane Science 389 (2012) 380– 388

sheet

3

3

raN5wwwsfla

Fig. 4. Scanning electron micrographs for PTFE flat

.2. DCMD performance for 1% NaCl solution

.2.1. Lower temperature rangeDCMD performances were explored first in lower temperature

anges (80–95 ◦C) in the setup designed for lower temperaturend lower pressure DCMD experiments. Feed brine containing 1%aCl was pumped over one side of the membrane at a flow rate of00 ml/min; on the other side of the membrane, cold distilled wateras passed at a rate of 500 ml/min. The temperature of the brineas varied between 80 ◦C and 95 ◦C; the temperature of the distilled

ater feed was maintained around 20 ◦C for all experiments. Steady

tate was assumed when the brine flow rate, the cold distilled waterow rate and the temperatures of the brine in and brine out as wells the temperatures of the distillate in and distillate out attained

membrane: (a) top surface and (b) bottom surface.

constant values. After steady state was reached, the increase in thewater level in the distillate tank over a certain time was used forcalculation of water vapor flux (Nv) from the following relation:

Nv (kg/m2h)

= volume of water transferred (l) × density of water (kg/l)membrane area (m2) × time (h)

(6)

Here, membrane area is calculated on the basis of inside diam-eter of the gasket used over the membrane surface.

Pore wetting is a common problem in DCMD experiments.Therefore, the conductivity of the distillate tank was monitoredcontinuously during the experiments. It was constant throughoutthe experiments, which indicated that the membrane pores were

D. Singh, K.K. Sirkar / Journal of Membrane Science 389 (2012) 380– 388 385

PTFE Membrane with 1% NaCl Solution

10

30

50

70

75 80 85 90 95 100

Temperature (deg C)

Wat

er V

apor

Flu

x (k

g m

-2 h

r-1)

Fl

npvtfl

33miDsmacttbqrvcFi1ibtosius

bd

N

Fh

2.00E-07

2.50E-07

3.00E-07

110 120 130Temperature (deg C)

Mas

s Tr

ansf

er

Coe

ffici

ents

(kg

m-2

s-1

Pa-

1)

it was observed that as the flow rate of brine was increased from50 ml/min to 850 ml/min; there was no significant change in thevalue of water vapor flux. This behavior is potentially the result ofthe special design of the cell on the brine side. From Fig. 2c, it is

90

100

, kPa

ig. 5. Variation of water vapor flux with temperature for 1% NaCl feed solution inower temperature DCMD setup for PTFE flat sheet membrane.

ot wetted by the 1% NaCl solution. Fig. 5 illustrates that as tem-erature of the feed brine was increased from 80 ◦C to 95 ◦C, waterapor flux increased from 14.5 kg/m2 h to 66 kg/m2 h. In these loweremperature experiments, there was ±1% variation in the value ofux; the variation has been shown in Fig. 5.

.2.2. Higher temperature range

.2.2.1. Membrane performances with 1% NaCl solution. The PTFEembrane was used above 100 ◦C and upto 3 atm gauge pressure

n the setup designed for the higher temperature/higher pressureCMD experiments. Hot 1% NaCl solution was pumped over one

ide of the membrane at 500 ml/min; the permeate side of theembrane was exposed to cold distilled water at a flow rate of

round 500 ml/min at temperatures of 25 ◦C to 30 ◦C. The method ofalculation of water vapor flux was based on the difference betweenhe flow rates of the distillate out and the distillate in. Sincehe membrane area was 128.67 cm2, the change in the flow rateetween the distillate out stream and the distillate in stream wasuite large. The change in the distillate in and the distillate out flowate was more than 25 ml/min at the higher temperatures; this pro-ided consistent and reliable data. Earlier experiments employingells having smaller membrane area did not provide reliable data.ig. 6 shows that the water vapor flux increased drastically withncreasing brine temperature; the flux went up from 105 kg/m2 h to95 kg/m2 h (a very high value compared to the flux levels achieved

n reverse osmosis desalination of brackish or sea water) as therine temperature was increased from 110 ◦C to 128 ◦C. Even athe high temperature of 128 ◦C, no pore wetting phenomenon wasbserved; there was no increase in the conductivity of the distillateide water. In these higher temperature experiments, the deviationn the data was ±3% as reflected in Fig. 6. A given membrane wassed successfully for many consecutive experiments without anyalt leakage.

Water vapor flux in the membrane distillation process is driven

y water vapor pressure difference across the membrane; it can beescribed as follows:

V = C�P = C(Pf − Pp) (7)

PTFE Membrane with 1% NaCl Solution

75

125

175

225

105 110 115 120 125 130

Temperature (deg C)

Wat

er V

apor

Flu

x (k

g m

- 2 h

r-1)

ig. 6. Variation of water vapor flux with temperature for 1% NaCl feed solution inigher temperature and higher pressure DCMD setup for PTFE flat sheet membrane.

Fig. 7. Variation of water vapor mass transfer coefficient with temperature of feedbrine in DCMD experiment for PTFE flat sheet membrane.

The proportionality constant C, a mass transfer coefficient,is determined by membrane properties like porosity, tortuosity,pore size, material and morphology of surface, etc. Although C isdependent on temperature and pressure, in many cases it is approx-imately constant, which was also observed for our system (Fig. 7).The transmembrane water vapor pressure difference, �P, equal tothe difference of feed side water vapor pressure Pf and the perme-ate side water vapor pressure Pp, is the driving force for water vaportransfer. The value of Pp was estimated at the average of inlet andoutlet distilled water temperatures; the maximum temperaturedifference between the inlet and the outlet was 30 ◦C. The enhance-ment in water vapor flux with temperature is a consequence of theexponential vapor pressure rise according to Antoine equation

P = exp(

23.238 − 3841Tm − 45

)(8)

Here P is the vapor pressure of water vapor in Pa and Tm is thetemperature in K. Using values of water vapor pressure, theoreti-cal values of water vapor pressure difference across the membranewere plotted against different values of brine temperature for fixedtemperature difference across the membrane in Fig. 8. Experimen-tally also the exponential growth in transmembrane vapor pressuredifference and correspondingly exponential growth in water vaporflux with temperature rise on feed side were observed; see Fig. 9.Such trends have been reported in other desalination studies forlower temperature ranges [24,25].

The effect of brine flow rate on water vapor flux for 1% NaClsolution was studied at different temperatures of brine (Fig. 10);

0

10

20

30

40

50

60

70

80

200 40 60 80 100 120 140 160

Tempeature, °C

Wat

er v

apor

pre

ssur

e di

ffere

nce

(p) T4= 10

T3= 8 °C

T2= 5 °C

T1= 2 °C

Fig. 8. Water vapor pressure difference at different values of feed water temperaturefor fixed temperature difference across the membrane.

386 D. Singh, K.K. Sirkar / Journal of Membrane Science 389 (2012) 380– 388

0

50

100

150

200

250

80 90 110 119 124Temperature (deg C)

Waa

ter V

apor

Flu

x (k

g m

- 2 h

r-1)

0

50

100

150

200

250

Vapo

r Pre

ssur

e D

iffer

ence

(kPa

)Water Vapor Flux (kg m-2 hr-1)Transmembrane PressureDifference (kPa)

Fig. 9. Water vapor flux and water vapor pressure difference across two sides ofporous PTFE flat sheet membrane at different hot brine temperatures.

75

125

175

105 110 115 120 125Temperature (deg C)

Wat

er V

apor

Flu

x(k

g m

- 2 h

r-1)

Flow Rate = 850 ml/minFlow Rate = 500 ml/minFlow Rate = 300 ml/minFlow Rate = 100 ml/minFlow Rate = 50 ml/min

Ftm

crgfla

lhdflmTerc

Hm

Table 2Simulated Composition of produced water.

Phenol (ppm) 45Cresol (ppm) 45

Fd

ig. 10. Variation of water vapor flux with different flow rates of 1% NaCl feed solu-ion in higher temperature and higher pressure DCMD setup for PTFE flat sheet

embrane.

lear that the cell on the brine side is not completely flat in shapeather it has a depth of 1.6 cm in the center of the cell which isradually reduced towards the periphery of the cell; this stagnationow creates enough shear on the brine side reducing the effect ifny of temperature and concentration polarization.

It is known that DCMD is a temperature driven process; heatoss either through the equipment surfaces to the environment oreat conducted through the membrane from the feed to the coldistillate results in a lower driving force and lower water vaporux through the membrane. Well insulated and properly designedembrane module can minimize the heat loss from the system.

hermal efficiency of the system is determined by evaporationfficiency () which is the ratio of the heat needed for water evapo-ation and the total heat exchanged between the hot brine and theold distillate. Evaporation efficiency () can be expressed as

NV �HV At NV �HV At

= .

Vdistillatecp�Tp

= .V feedcp�Tf

(9)

ere �HV is the heat of vaporization of water, At is the total effectiveembrane area,

.Vdistillate and

.V feed are the volumetric flow rates of

ig. 11. Chromatogram of phenol and cresol obtained from HPLC with mobile phase 90/etector at �max = 218 nm.

Naphthenic acid (ppm) 10Sodium chloride (ppm) 3000

the distillate and feed respectively, �Tf is the temperature drop onfeed brine side and �Tp is the temperature rise on the distillateside. Assuming negligible heat loss to environment, which is notthe case, evaporation efficiency at the higher temperatures for thissystem was found to be approximately 80% based on the distillatestream.

3.2.2.2. Membrane performances with simulated produced water. Inthe SAGD process [26], high pressure steam is injected into groundwhich requires high quality water input for steam generation. Cor-respondingly, huge amounts of produced water are obtained alongwith crude oil at temperatures around 150 ◦C. The DCMD perfor-mance of this PTFE membrane was also explored with a simulatedproduced water feed at high temperatures. The produced watercomposition employed is characteristic of the SAGD process [27]and is provided in Table 2. Water vapor fluxes measured were sim-ilar to those from a 1% NaCl feed solution for flat sheet membranes;it was unaffected by the presence of organic compounds in thefeed brine. The boiling points of phenol and cresol are 182 ◦C and191 ◦C respectively; it is expected that due to volatilization tracesof phenol and cresol will be present on the distillate side (Fig. 11).Water vapor was collected in the distillate tank and concentrationsof phenol and cresol would continuously increase with time. After30 min, samples were taken from the distillate tank for analysis byHPLC to measure the concentrations of phenol and cresol. Two dis-tinct peaks (Fig. 11) were obtained at retention times of 2.8 min and4.1 min for phenol and cresol respectively. The water vapor fluxesand concentration of phenol and cresol at high temperatures withproduced water are reported in Figs. 12 and 13; it is clear that con-centrations of phenol and cresol are below 5 ppm even when thebrine temperature is 125 ◦C. Although both phenol and cresol areorganic compounds with some hydrophobicity, they are quite polarand do not possess low enough surface tension to wet the PTFEmembrane pores. Therefore, their presence in the permeate with-out any salt indicates their transport via vaporization. It is generallysafe to use this water for steam generation for such low concentra-tions of phenol and cresol. No trace of sodium chloride was found

in the distillate stream during the experiments with the simulatedproduced water; this was verified by conductivity measurement aswell as by AA analysis.

10 (v/v) methanol/water and 0.1% H3PO4 (v/v) at the flow rate of 1 ml/min by UV

D. Singh, K.K. Sirkar / Journal of Membrane Science 389 (2012) 380– 388 387

0

2

4

6

8

10

95 105 110 115 125Temperature (deg C)

Phen

ol C

onc.

(ppm

)

50

100

150

200

Wat

er V

apor

Flu

x(k

g m

- 2 h

r-1)

PhenolFlux

Fig. 12. Concentration of phenol on permeate side along with water vapor flux at dif-ferent temperatures of simulated produced water in higher temperature and higherpressure DCMD setup for PTFE flat sheet membrane.

0

2

4

6

8

10

95 105 110 115 125Temperature (deg C)

Cre

sol C

onc.

(ppm

)

50

100

150

200

Wat

er V

apor

Flu

x (k

g m

- 2 h

r-1)

CresolFlux

Fig. 13. Concentration of cresol on permeate side along with water vapor flux at dif-fp

cirrbodbT

4

osppwcpwbtpNiIaberog

Nomenclature

At total permeation area (m2)cp specific heat (kJ/kg K)C mass transfer coefficient (kg m−2 s−1 Pa−1)�HV heat of evaporation (kJ/kg)Io interceptJi gas permeation flux (mol m−2 S−1)Lp effective pore length (m)M molecular weight of gas/vaporNt,i gas permeation rate (mol s−1)Nv water vapor flux (kg m−2 h−1)Pf feed side vapor pressure (Pa)Pp permeate side vapor pressure (Pa)P vapor pressure (Pa)p̄ mean pressure (Pa)�p transmembrane pressure difference (Pa)R gas constant (8.3144 m3 Pa mol−1 K−1)r mean pore radius of membrane (m)So slopeT absolute temperature (K)Tm temperature at membrane surface�Tp temperature rise on permeate side�Tf temperature drop on feed side.

Vdistillate volumetric flow rate of distillate (ml/min).

V feed volumetric flow rate of feed (ml/min)

Greek lettersε surface porosity

Separation and Purification Technology 43 (2005) 233–244.

erent temperatures of simulated produced water in higher temperature and higherressure DCMD setup for PTFE flat sheet membrane.

The analysis of naphthenic acid is extremely difficult due to itsompositional heterogeneity; as a result chromatogram of under-vatized naphthenic acid resulted in a large hump with very fewesolvable peaks [22]. But derivatized naphthenic acid componentseduce extent of molecular fragmentation and resulted in strongase peaks in spectra of derivatized naphthenic acids. The spectraf derivatized naphthenic acid extracted from the distillate tank atifferent temperatures have shown no traces of naphthenic acid;ut at 125 ◦C few ppm (<2 ppm) of naphthenic acids were detected.his is due to naphthenic acid having a boiling point of 212 ◦C.

. Concluding remarks

A microporous PTFE flat sheet membrane having a pore sizef 0.03 �m was employed in DCMD experiments with 1% NaClolution at lower brine temperatures as well as higher brine tem-eratures (>100 ◦C). A simulated feed representing SAGD processroduced water containing phenol, cresol, naphthenic acid alongith sodium chloride was also investigated. The membrane was

haracterized by scanning electron microscopy and nitrogen gasermeation test to study surface morphology and its properties. Aater vapor flux of up to 195 kg/m2 h was achieved at 128 ◦C for 1%

rine solution, which is around an order of magnitude higher thanhat for seawater reverse osmosis (RO) processes. Experiments withroduced water yielded water vapor fluxes similar to those for 1%aCl solution. Concentrations of phenol, cresol and naphthenic acid

n the permeate were analyzed successfully by HPLC and GC–EI-MS.t was found that only a few ppm (∼5 ppm) of phenol and cresol andround 2 ppm of naphthenic acid passed through the PTFE mem-rane at very high temperature with no leakage of NaCl. This isxpected since the boiling points of these compounds are in the

ange of 165–212 ◦C. Further such low permeate concentrationsf these compounds are not expected to be a problem for steameneration.

�i viscosity of gas (Pa s)� evaporation efficiency

In general, DCMD has been proved to be a very useful process,especially at high temperatures and correspondingly above-atmospheric pressures where it can utilize the considerable heatavailable in the feed. Compared to other conventional processeslike reverse osmosis, it may be economically viable in such cases,because DCMD does not require cooling which reduces additionalenergy requirements.

Acknowledgements

We acknowledge funding for this research from MAST (Mem-brane Science, Engineering & Technology) Center, ConocoPhillips,USA and National Science Foundation, USA. Kay Bjornen and XiaoyiGong of ConocoPhillips provided substantial help. We acknowledgeadditionally the following contributions: Uwe Beuscher from W. L.Gore & Associates especially for providing us the PTFE flat sheetmembrane; Irv Joffee from Pall Corporation for providing stainlesssteel mesh supports; Fei He designed the higher temperature setup.Liming Song developed Fig. 8 in 2006.

References

[1] D.W. Gore, Gore-Tex membrane distillation, in: Proceedings of the 10th AnnualConvention of Water Supply Improvement Association, Honolulu, Hawaii, July25–29, 1982.

[2] Baoan Li, K. Sirkar Kamalesh, Novel membrane and device for vacuum mem-brane distillation-based desalination process, Journal of Membrane Science 257(1–2) (2005) 60–75.

[3] C.M. Guijt, G.W. Meindersma, T. Reith, A.B. de Haan, Air gap membrane distil-lation. 1. Modeling and mass transport properties for hollow fiber membranes,

[4] C. Feng, K.C. Khulbe, T. Matsuura, R. Gopal, S. Kaur, S. Ramakrishna, M. Khayet,Production of drinking water from saline water by air-gap membrane distilla-tion using polyvinylidene fluoride nanofiber membrane, Journal of MembraneScience 311 (1–2) (2008) 1–6.

3 embr

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

88 D. Singh, K.K. Sirkar / Journal of M

[5] R.W. Schofield, A.G. Fane, C.J.D. Fell, Heat and mass transfer in membrane dis-tillation, Journal of Membrane Science 33 (3) (1987) 299–313.

[6] J. Phattaranawik, R. Jiraratananon, A.G. Fane, Heat transport and membranedistillation coefficients in direct contact membrane distillation, Journal ofMembrane Science 212 (1–2) (2003) 177–193.

[7] W. Lawson Kevin, R. Lloyd Douglas, Membrane distillation – review, Journal ofMembrane Science 124 (1–5) (1997) 1–25.

[8] B. Li, K.K. Sirkar, Novel membrane and device for direct contact membranedistillation-based desalination process, Industrial and Engineering ChemistryResearch 43 (2004) 5300–5309.

[9] L. Song, B. Li, K.K. Sirkar, J. Gilron, Direct contact membrane distillation-baseddesalination: Novel membranes, devices, larger-scale studies and a model,Industrial and Engineering Chemistry Research 46 (2007) 5300–5309.

10] L. Song, Z. Ma, X. Liao, P.B. Kosaraju, J.R. Irish, K.K. Sirkar, Pilot plant studies ofnovel membranes and devices for direct contact membrane distillation-baseddistillation, Journal of Membrane Science 323 (2008) 257–270.

11] E. Drioli, E. Curcio, G. di Profio, State of the Art and recent progresses in mem-brane contactors, Chemical Engineering Research and Design 83 (3) (2005)223–233.

12] F. He, J. Gilron, H. Lee, L. Song, K.K. Sirkar, Potential for scaling by sparinglysoluble salts in cross-flow DCMD, Journal of Membrane Science 311 (2008)68–80.

13] F. He, K.K. Sirkar, J. Gilron, Studies on scaling of membranes in desalination bydirect contact membrane distillation: CaCO3 and mixed CaCO3/CaSO4 systems,Chemical Engineering Science 64 (2009) 1844–1859.

14] F. He, K.K. Sirkar, J. Gilron, Effects of antiscalants to mitigate membrane scal-

ing by direct contact membrane distillation, Journal of Membrane Science 345(2009) 53–58.

15] J. Gilron, L. Song, K.K. Sirkar, Design for cascade of cross-flow direct con-tact membrane distillation, Industrial and Engineering Chemistry Research 46(2007) 2324–2334.

[

[

ane Science 389 (2012) 380– 388

16] H. Lee, F. He, L. Song, J. Gilron, K.K. Sirkar, Desalination with a cascade of cross-flow hollow fiber membrane distillation devices integrated with a hollow fiberheat exchanger, AIChE Journal 57 (7) (2011) 1780–1795.

17] C.V. Deutsh, J.A. McLennan, Centre for Computational Geostatistics (CCG) GuideBook Series, vol. 3, 2005.

18] H. Yasuda, J.T. Tsai, Pore size of microporous polymer membranes, Journal ofApplied Polymer Science 18 (1974) 805–819.

19] K. Li, J.F. Kong, W.K. Teo, Tailor made asymmetric hollow fibers for soluble gasremoval, AIChE Journal 45 (6) (1999) 1211–1219.

20] D. Wang, W.K. Teo, Preparation and characterization of polyvinylidene fluo-ride (PVDF) hollow fiber membranes, Journal of Membrane Science 163 (1999)211–220.

21] M. Merlin, S.E. Guigard, P.M. Fedorak, Detecting naphthenic acids in waters bygas chromatography–mass spectrometry, Journal of Chromatography A 1140(2007) 225–229.

22] P. Wayne, St. John, Jagdish Rughani, A. Sarah, Green, D. Gary, McGinnis, Analy-sis and characterization of naphthenic acids by gas chromatography–electronimpact mass spectrometry of tert.-butyldimethylsilyl derivatives, Journal ofChromatography A 807 (1998) 241–251.

23] T.E. Kohav, M. Sheintuch, D. Anvir, Steady state diffusion and reactionsin catalytic fractal porous media, Chemical Engineering Science 46 (1991)2787–2798.

24] A. Criscuoli, M.C. Carnevale, E. Drioli, Evaluation of energy requirementsin membrane distillation, Chemical Engineering and Processing 47 (2007)1043–1050.

25] L. Martinez, F.G. Florido-Diaz, Theoretical and experimental studies

on desalination using membrane distillation, Desalination 139 (2001)373–379.

26] C.V. Deutsch, J.A. McLennan, Centre for Computational Geostatistics (CCG)Guidebook Series, vol. 3, 2005.

27] Xi. Gong, ConocoPhillips, August 04, 2010, Personal communication.