A Path to Ultra-Selectivity: Support Layer Properties to Maximize


A Path to Ultra-Selectivity: Support Layer Properties to Maximize...

0 downloads 55 Views 1MB Size

Subscriber access provided by Kaohsiung Medical University

Remediation and Control Technologies

A Path to Ultra-Selectivity: Support Layer Properties to Maximize Performance of Biomimetic Desalination Membranes Jay Ryan Werber, Cassandra J Porter, and Menachem Elimelech Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b03426 • Publication Date (Web): 14 Aug 2018 Downloaded from http://pubs.acs.org on August 17, 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 31

Environmental Science & Technology

1 2 3 4 5 6 7 8 9

A Path to Ultra-Selectivity: Support Layer Properties to Maximize Performance of Biomimetic Desalination Membranes

10 11 12 13 14 15 16 17 18 19 20 21 22

Environmental Science & Technology Revised: August 9, 2018

Jay R. Werber, Cassandra J. Porter, and Menachem Elimelech*

Department of Chemical and Environmental Engineering, Yale University, New Haven, Connecticut 06520-8286

23 24 25 26 27 28 29

* Corresponding author: Menachem Elimelech, Email: [email protected], Phone: (203) 432-2789

1 ACS Paragon Plus Environment

Environmental Science & Technology

30

ABSTRACT

31

Reverse osmosis (RO) has become a premier technology for desalination and water purification.

32

The need for increased selectivity has incentivized research into novel membranes, such as

33

biomimetic membranes that incorporate the perfectly-selective biological water channel

34

aquaporin or synthetic water channels like carbon nanotubes. In this study, we consider the

35

performance of composite biomimetic membranes by projecting water permeability, salt

36

rejection, and neutral-solute retention based on the permeabilities of the individual components,

37

particularly the water channel, the amphiphilic bilayer matrix, and potential support layers that

38

include polymeric RO, nanofiltration (NF), and porous ultrafiltration membranes. We find that

39

the support layer will be crucial in the overall performance. Selective, relatively low-

40

permeability supports minimize the negative impact of defects in the biomimetic layer, which are

41

currently the main performance-limiting factor for biomimetic membranes. In particular, RO

42

membranes as support layers would enable >99.85% salt rejection at ~10,000-fold greater

43

biomimetic-layer defect area than for porous supports. By fundamentally characterizing neutral-

44

solute permeation through RO and NF membranes, we show that RO membranes as support

45

layers would enable high rejection of organic pollutants based on molecular size, overcoming the

46

rapid permeation of hydrophobic solutes through the biomimetic layer. A biomimetic membrane

47

could also achieve exceptionally high boron rejections of ~99.7%, even with 1% defect area in

48

the biomimetic layer. We conclude by discussing the implications of our findings for biomimetic

49

membrane design.

50 51

INTRODUCTION

52

Water scarcity is one of the foremost challenges of the 21st century, with severe water scarcity

53

affecting 4 billion people for at least one month a year.1 While responsible water management is

54

essential, water-treatment technologies that increase potable water supplies are similarly crucial.

55

In particular, membrane-based processes such as reverse osmosis (RO) are finding extensive

56

application.2 RO has emerged as a premier technology for not just the desalination of seawater

57

and brackish water, but also the treatment and reuse of municipal and industrial wastewaters due

58

to the ability of RO to remove a wide variety of inorganic and organic contaminants.3, 4 The

59

central component in RO is the membrane itself. Because of the importance of RO in addressing

2 ACS Paragon Plus Environment

Page 2 of 31

Page 3 of 31

Environmental Science & Technology

60

global water challenges, extensive research efforts are on-going to develop novel membrane

61

materials for RO.2

62

The current state of the art is the fully aromatic polyamide thin-film composite (TFC)

63

membrane. Current membranes can achieve up to 99.85% salt (NaCl) rejection5, 6 in standard

64

seawater RO (SWRO) test conditions (55.1 bar, 32000 ppm NaCl) with water permeabilities of

65

1−3 L m-2 h-1 bar-1.2,

66

performance from an energy- and cost-efficiency standpoint, with increased water permeability

67

projected to yield minimal practical impact.7,

68

attainable by TFC membranes would be highly beneficial in desalination and water treatment.7

69

For example, increased rejection of boron and chloride would eliminate the need for additional

70

RO passes during SWRO when the product water is eventually used for irrigation.7 Similarly,

71

increased rejection of organic micropollutants in advanced wastewater treatment would allow for

72

effective potable reuse without the advanced oxidation step, which is currently used to degrade

73

recalcitrant species in the RO permeate but has the potential to itself produce toxic by-products.9

7

These water permeabilities are sufficient for near-optimal SWRO 8

In contrast, increased selectivity above levels

74

A promising class of “next-generation” membrane materials that have the potential to

75

achieve dramatically improved selectivity are so-called biomimetic desalination membranes.2, 10

76

Design of these membranes seeks to mimic the structure and separation performance of

77

biological cell membranes by incorporating biological water channels (i.e., the membrane protein

78

aquaporin11) or chemically-designed synthetic water channels2, 10 within a bilayer of lipids or

79

amphiphilic block copolymers. Amphiphilic block copolymers (macromolecules with discrete

80

hydrophilic and hydrophobic polymer segments) are strongly preferred due to their chemical and

81

mechanical stability.12,

82

commercial interest due to its relative stability14 and, more importantly, its high water

83

permeability and essentially perfect selectivity. AqpZ and certain other aquaporin types are near-

84

perfect barriers to all solutes, even small neutral solutes like ammonia and urea.15, 16

85 86

13

Bacterial aquaporin-Z (AqpZ) has received the most research and

Synthetic water channels would ideally yield similar transport performance as aquaporin, but with improved stability and processability.17, 19, 20

18

The most prominent example is the carbon

87

nanotube porin (CNTP),

88

selectivity for water over salt if the CNTP diameter is well-controlled.20, 21 Research on synthetic

89

water channels is still mostly in the early stage, with channel permeabilities and/or selectivities

which is extremely permeable to water and could yield very high

3 ACS Paragon Plus Environment

Environmental Science & Technology

90

still substantially inferior to that of aquaporin.2 For synthetically-designed channels, current

91

research largely focuses on the rational design of improved channel structures.22

92

In terms of membrane fabrication, there have been two approaches used thus far, mostly with

93

AqpZ as the water channel. The first method incorporates AqpZ-containing vesicles within the

94

polyamide layer of TFC membranes, with the vesicles theoretically improving the water

95

permeability.23 While this method has recently been commercialized,24 performance is still

96

largely dependent on the polyamide film, and for this reason, only minor performance gains may

97

be possible. The second, more radical method is to form a solid-supported planar bilayer to serve

98

as an integral selective layer. Following the example of TFC membranes, for which the

99

polyamide selective layer is formed on ultrafiltration-type porous supports, most efforts to form

100

planar biomimetic membranes have used porous support layers.10 In contrast to the vesicle

101

approach, the planar-bilayer approach would fully utilize the transport properties of AqpZ (or

102

other selective water channels).

103

We recently reported that a defect-free biomimetic selective layer with ~3% AqpZ (by area)

104

would have revolutionary performance, with similar water permeability as TFC membranes,

105

near-perfect salt rejection (salt passage of 1 since ABM is dependent on ASL from eq 6. The normalized

268

pure water flux through defects is the ratio of the water flux through defects, Jw,def, and the total

269

water flux, Jw,total, and is given by

270

J w,def

J w,total

=

A 'θ 1− θ + A 'θ

[8]

271

As shown in Figure 2B, the normalized pure water flux through defects is much greater for

272

support layers with higher permeability. Decreasing A’—by maximizing the permeability of the

273

biomimetic layer or by using a relatively low-permeability support layer—is clearly important

274

for minimizing the deleterious impact of defects. If flow through defects is magnified by a high

275

value of A’, then even small defect areas will be catastrophic for performance.

276

Potential for Unprecedented Salt Rejection. Biomimetic membranes arguably have

277

the greatest potential among next-generation RO membrane materials for dramatic improvements

278

in salt rejection.25 If a channel is used that perfectly rejects ions (such as aquaporin), then ion

279

flux through a defect-free active layer can only occur through the amphiphilic matrix. We

280

recently measured NaCl permeability through PB-PEO and MDM bilayers to be ~4 × 10-9 L m-2

281

h-1, as compared with 0.03–0.06 L m-2 h-1 for the most-selective TFC RO membranes.7

282

Correspondingly, NaCl passage of 99.9999999%) is expected for

283

a defect-free biomimetic membrane. The discrepancy between this ultra-low value and the

284

relatively poor rejections observed experimentally10 must be due to the presence of defects.

10 ACS Paragon Plus Environment

Page 11 of 31

Environmental Science & Technology

285

In order to quantitatively estimate the impact of defects, we consider the local water and salt

286

fluxes that would occur in defect-free regions and through defects. We use the water and salt

287

performance equations from the solution-diffusion model, which models molecular transport

288

through the membrane as a combination of partitioning from the feed into the membrane,

289

diffusion down its chemical potential gradient across the membrane thickness, and departitioning

290

into the permeate.39 The performance equations incorporate boundary-layer effects using film

291

theory, which can then be used to solve for the local water flux and the local salt flux, Js:

292

  J  J w,x = Ax ∆p − ∆π exp  w,x    kNaCl 

[9]

293

 J  J s,x = Bx ∆C exp  w,x   k NaCl 

[10]

294

The subscript x refers to the local membrane condition: a defect-free region of the biomimetic

295

membrane (subscript “BM”) or a defect (subscript “def”). The osmotic pressure difference, ∆π,

296

and the salt concentration difference, ∆C, refer to the bulk feed and permeate solutions (e.g., ∆C

297

= Cb − Cp). The total flux, J, for water or salt is then calculated as the area-weighted average of

298

the local fluxes:

299

J = (1 − θ ) J BM + θ J def

[11]

300

The observed rejection, Ro, can then be calculated from these total water and salt fluxes, as the

301

permeate concentration, Cp, is the ratio of the fluxes (i.e., Cp = J s J w ), and Ro = 1 − Cp Cb .

302

Figure 3 shows the projected NaCl rejections during SWRO for defect-containing

303

biomimetic membranes (with 3% AqpZ) formed on various support layers. The dashed line

304

shows the highest NaCl rejection attainable using current TFC membranes (99.85%).5, 6 Because

305

the biomimetic active layer can theoretically attain near-perfect rejection, biomimetic

306

membranes formed on each of the considered supports can theoretically outperform this rejection

307

level. The key difference is the acceptable defect area. For example, according to our model, a

308

biomimetic membrane formed on the most selective, lowest permeability support—RO-HR,

309

which itself is already a SWRO membrane—could reach 99.85% rejection with a defect density

310

of 16%, whereas a biomimetic membrane formed on the porous UF-3 support would need a 11 ACS Paragon Plus Environment

Environmental Science & Technology

311

~10,000-fold lower defect density (0.002%). The intermediate permeability support layers, RO-

312

HP and NF, would require defect densities to be less than 4% and 0.06%, respectively. Similarly,

313

for a given defect density, the corresponding projected rejections vary dramatically. For example,

314

if a defect density of 0.06% can be achieved to enable 99.85% rejection using NF as a support,

315

the corresponding salt rejection when using RO-HR as a support would be 99.9994%, a more

316

than 200-fold increase in separation efficiency. While increasing defect area increases water

317

permeability (Figure 2B), in the performance range of interest (Ro ≥ 99.85%) the effect on

318

selectivity is much greater (Figure S1).

319

The difference in projected performance stems from two factors: the salt-rejecting

320

performance of the support layer and the ratio of water permeabilities, A’, that was discussed in

321

the previous section. For example, if we arbitrarily increase the salt permeability of RO-HR to

322

1000 L m-2 h-1 but keep the water permeability at 2.20 L m-2 h-1 bar-1, then the defect density

323

would need to be ~0.05% to achieve 99.85% salt rejection, roughly the same density as projected

324

when using the partially salt-rejecting NF and 25-fold greater than the density projected for UF-3

325

(0.002%). Conclusions are very similar for a biomimetic active layer containing 10% AqpZ

326

(Figure S2).

327

The dramatic defect-mitigating effect of moderately-permeable, salt-selective support layers

328

in biomimetic membranes is strikingly similar to the effect of “caulking” that first enabled the

329

commercial application of membranes for gas separations.34, 35 In gas separation membranes, the

330

inevitable presence of defects drastically reduced achievable separation factors due to the sharp

331

permeability difference between flow in the pores (defects) and solution–diffusion-driven flow

332

through the polymeric selective layer. The caulking approach solved this problem by casting a

333

relatively thick (>1 µm) silicone rubber layer on top of the defect-containing membrane to block

334

the defects. This silicone layer, which was slightly more permeable than the base membrane and

335

not selective between the relevant gases, sharply decreased the permeability of the defected area.

336

The reduced defect permeability from this composite membrane structure allowed for ~105-fold

337

higher defect tolerance (by area), with only minor reductions in overall permeability and

338

selectivity compared to a defect-free base membrane.34

339

FIGURE 3

12 ACS Paragon Plus Environment

Page 12 of 31

Page 13 of 31

Environmental Science & Technology

340

Neutral Solute Retention. While salt removal is typically the most prominent task in RO,

341

especially in seawater desalination, neutral solute removal is often important. For example, in

342

SWRO, removal of the neutral boric acid (boron) to very low levels (less than ~0.5 ppm) is often

343

necessary when the product water is used downstream for irrigation purposes.7 In the large-

344

volume production of ultrapure water for the semiconductor and power industries, RO is used as

345

the central step to remove the bulk of the dissolved ion content.7 Extremely high salt rejections

346

like those projected in Figure 3 would be very useful in this application, but the need for low

347

organic content also necessitates neutral solute removal. Undoubtedly though, neutral solute

348

removal is most important in the potable reuse of municipal wastewater, which will play a major

349

role in the coming years in addressing water scarcity issues.3 The most critical requirement of

350

RO in wastewater reuse is the removal of small micropollutants that are difficult to remove using

351

other technologies. Micropollutants are species that are harmful to human or ecological health at

352

very low concentrations, and typically are relatively hydrophobic.40

353

In our recent study, we found that the permeability of neutral solutes through lipid and block

354

copolymer bilayers was solubility-dependent over 10 orders of magnitude.25 In other words, the

355

more soluble a species is in the nonpolar core of the bilayer, the more permeable that species is.

356

Solute size played a relatively minor role. Permeability of solutes through the membrane matrix,

357

PM, was related to the octanol-water partition coefficient, Kow, a commonly used metric of solute

358

hydrophobicity:25

359

PBL ≈ PM = a ( K ow )

b

[12]

360

PBL ≈ PM for all solutes because of the low areal density of water channels25 and the perfect

361

selectivity of aquaporin (i.e., Pc = 0 for all species except for water). For P in units of m/s, the

362

coefficients a and b in eq 12 were 1.78 × 10-6 and 2.05 for PB-PEO and 1.33 × 10-4 and 2.19 for

363

MDM. While this relation does not give quantitative predictions, it allows for the permeability of

364

a given solute to be estimated. In particular, a hydrophobicity cut-off of log Kow ~ −0.2 was

365

determined for biomimetic active layers using PB-PEO, the less permeable (and therefore more

366

selective) amphiphilic matrix.25

367

All of the support layers considered (RO, NF, and UF) predominantly retain neutral solutes

368

based on size. Therefore, the permeation behavior of neutral solutes through a composite 13 ACS Paragon Plus Environment

Environmental Science & Technology

Page 14 of 31

369

biomimetic membrane will involve both the hydrophobicity-based selectivity of the biomimetic

370

layer and the size-based selectivity of the support layer. To obtain size-based relationships for

371

solute retention, the permeation of small neutral molecules was assessed for RO-HR, RO-HP,

372

and NF.

373

For RO, transport was assumed to follow the solution-diffusion model.39,

41

Using real

374

rejections measured for model neutral solutes (Figure 4A, Table S1), solute permeabilities were

375

determined for RO-HR and RO-HP (Figure 4B, Table S1). Model solutes were chosen to give a

376

range of size as well as hydrophobicity, including hydrophilic sugars and hydrophobic solutes

377

such as the aromatic thymol (log Kow of 3.3). Rejection behavior was relatively similar for RO-

378

HR and RO-HP, both of which had molecular-weight cut-offs of ~80 g/mol. By free-volume

379

theory, diffusive permeability, Pi, for species i through a polymer film is given by30

380

 −b  Pi = Ki ai exp  i   〈vf 〉 

[13]

381

where Ki is the partition coefficient of species i into the membrane, 〈vf〉 is the average free

382

volume in the polymer, and ai and bi are adjustable parameters. The parameter bi is proportional

383

to the solute size. In the original formulation by Cohen and Turnbull,42 bi = γ v * , where γ is a

384

fitting parameter and v* is the minimum void volume that allows entrance or exit of the penetrant.

385

Based on this volume-based condition, a comparison of log P with the van der Waals volume of

386

the species of interest is perhaps most appropriate. The van der Waals volume is the minimum

387

volume that a species occupies,43, 44 which is relevant considering that the solution-diffusion

388

model often assumes diffusion of individual molecules through the film.39 While the measured

389

permeabilities for RO-HR and RO-HP fit fairly well with the van der Waals volume (Figure S3),

390

the correlation is stronger with the van der Waals radius, rvdw (Figure 4B), which is the radius of

391

a sphere with the equivalent van der Waals volume. Because of the relatively strong correlations

392

(r2 of 0.88 for RO-HR and 0.76 for RO-HP) and the more intuitive nature of the solute radius, we

393

relate the permeability of species i to its van der Waals radius, rvdw,i, using the following

394

adaptation of eq 13:

395

Pi = a exp ( −brvdw,i )

[14]

14 ACS Paragon Plus Environment

Page 15 of 31

Environmental Science & Technology

396

Here, a and b are constants that are characteristic of a given membrane. For permeability in m/s

397

and rvdw in nm, a and b were 0.319 m/s and 49.8 nm-1 for RO-HR, and 0.0558 m/s and 40.0 nm-1

398

for RO-HP.

399

FIGURE 4

400

Loose NF membranes like NF-270 are generally considered to have discrete pores.29 For this

401

reason, instead of calculating permeabilities using the solution-diffusion model, the NF

402

membrane was assessed using a pore flow model (hindered transport theory), which is fully

403

described elsewhere.29, 45 The key outcome of this model is an equation for real rejection, Rr, that

404

largely relies on λ, which is the ratio between the solute radius, rvdw, and the pore radius, rp (i.e.,

405

λ = rvdw rp ):

(1 − λ ) K c Rr = 1 − 2 1 − exp ( − Pe ) 1 − (1 − λ ) K c    2

406

[15]

407

Here, Kc is a hydrodynamic coefficient and the membrane Peclet number for species i, Pei, is

408

defined as

409

Pei =

K c J wδ m K dε Di

[16]

410

where Kd is also a hydrodynamic coefficient. Kc and Kd are functions of λ and can be determined

411

analytically as described elsewhere.29, 46 The membrane thickness, δm, of NF was taken to be 30

412

nm based on published thickness measurements of the isolated active layer.47 The porosity, ε,

413

was taken to be 8% to yield reasonable water permeability values by eq 4.

414

Using eqs 15 and 16, rejection data for select neutral solutes was fitted to determine the

415

membrane pore size (Figure S4, Table S2). The hydrophilic, non-aromatic solutes glucose,

416

erythritol, and dioxane yielded a consistent pore radius of 0.40−0.41 nm, matching published

417

results for this membrane determined using similar solutes.29, 47 However, when a hydrophobic

418

aromatic solute (thymol) was used as the organic tracer, rejection was anomalously low,

419

resulting in a substantially larger fitted pore radius of 0.51 nm. For aromatic acids assessed at pH

420

3 (Figure S5), fitted pore sizes were similarly greater than the 0.40-nm pore radius obtained

421

using the hydrophilic solutes. Anomalously low rejection of hydrophobic contaminants by NF

15 ACS Paragon Plus Environment

Environmental Science & Technology

Page 16 of 31

422

membranes (NF-270) has similarly been observed for hydrophobic endocrine disrupting

423

compounds, with the increased permeation attributed to enhanced solute partitioning within the

424

semi-aromatic active layer.29, 48

425

From the relationships above, the permeation of neutral solutes through composite

426

biomimetic membranes can be estimated based on the solute size (rvdw) and hydrophobicity (Kow).

427

For RO membranes as support layers, the composite permeability can be determined with eq 6

428

and used to calculate rejection using the solution-diffusion model (analogs to eqs 1, 9, and 10).

429

For NF and UF membranes as support layers, the real rejection by the support layer, Rr,SL, can

430

first be determined based on Jw and λ with eqs 15 and 16. For NF, the larger fitted pore size of

431

0.51 nm based on thymol was used, as the majority of the solutes of interest are aromatic and/or

432

hydrophobic. The calculated real rejection by the support layer can then be combined with the

433

biomimetic layer permeability, PBL, to calculate the overall observed rejection, Ro:

434

435

J  R  −J  Ro =  w + r,SL  exp  w  1 − Ro  PBL 1 − Rr,SL   ki 

[17]

A derivation of eq 17 is provided in the Supporting Information.

436

Projected rejections of neutral solutes, based on rvdw and Kow, are shown in Figures 5 and S6

437

for biomimetic membranes using PB-PEO and MDM, respectively. Fluxes were calculated for

438

applied pressures of 15.5 bar and 1000 ppm NaCl. Overlaid on the rejection contours are the size

439

and hydrophobicity of neutrally-charged micropollutants that are relevant for wastewater reuse.25

440

Almost all of the micropollutants would be poorly rejected when UF-3 is used as a support, since

441

hydrophobic species will rapidly permeate the biomimetic active layer and 3-nm pores are too

442

large to meaningfully retain solutes of interest. The main exception is the hydrophilic

443

pharmaceutical iopromide (log Kow of −2.05), which should be highly rejected by the biomimetic

444

active layer. Two important partial exceptions are N-nitrosodimethylamine (NDMA; log Kow of

445

−0.57) and 1,4-dioxane (log Kow of −0.27), both of which have caused shutdowns of advanced

446

wastewater treatment plants due to detection in the product water.7,

447

octanol/water partition coefficients, NDMA and 1,4-dioxane would be rejected at ~98% and

448

~94%, respectively, by a defect-free PB-PEO-based biomimetic active layer. Clearly, however,

449

the largest gains can be made by using a highly selective support layer. If RO-HR is used as a

16 ACS Paragon Plus Environment

49

Based on their

Page 17 of 31

Environmental Science & Technology

450

support, all of the micropollutants considered are projected to be at least 90% retained, with

451

many retained at >99.9%.

452

FIGURE 5

453

The above discussion was for a defect-free biomimetic layer. Using the same methodology as

454

for salt, we now consider rejection with biomimetic layers with 1% defect area for the important

455

neutral solutes boron and 1,4-dioxane. Explicitly-measured boron permeabilities from our

456

previous study25 were used for PB-PEO (0.018 ± 0.004 L m-2 h-1) and MDM (0.86 ± 0.07 L m-2 h-

457

1

458

m-2 h-1 for MDM, as permeation was too rapid for measurement (Figure S7). Boron and dioxane

459

permeabilities were also explicitly measured for the model support layers (Figure 6A). Figure 6B

460

shows the projected rejections using biomimetic membranes in SWRO for boron and in

461

wastewater RO for 1,4-dioxane. For boron, the biomimetic membrane offers tremendous

462

advantage, even with 1% defect area. With RO-HR as a support, projected boron rejections are

463

99.7% (PB-PEO) and 97.5% (MDM), compared with 87.1% for the support alone. Even with NF

464

as a support, which only has 5.4% boron rejection by itself, rejections of 94.6−96.5% are

465

projected, which exceed rejections (93% at pH 8) of the most-selective SWRO membranes

466

For dioxane, the biomimetic layer only plays a significant role when the support is NF and the

467

selective layer uses PB-PEO, with a projected rejection of 89% compared to 37% for the support

468

alone. For RO-HR and RO-HP, the support plays the dominant role in dioxane retention.

469

). 1,4-dioxane permeability was estimated using eq 12 to be 1.8 L m-2 h-1 for PB-PEO and 123 L

5, 6

.

FIGURE 6

470

Implications for Membrane Design. Our analysis projects that composite biomimetic

471

desalination membranes could achieve ultra-selectivity for desalination (removal of salt), even

472

with meaningful amounts of defects in the biomimetic layer. Similar, if not better, performance

473

would be expected for inorganic contaminants such as heavy metals or arsenic. Our models also

474

predict dramatic differences in performance based on the support layer. In particular, design

475

strategies using RO or NF membranes as support layers are highly advantageous. Higher

476

permeability UF membranes offer marginal benefits for water permeability, sharply increase the

477

penalty of defects on salt rejection, and do not aid with neutral solute retention. For increased

478

removal of salt and contaminants, using the most selective support layer (i.e., RO-HR) is always

479

preferred. The main drawback is the moderate decrease in water permeability, although the 1.18 17 ACS Paragon Plus Environment

Environmental Science & Technology

Page 18 of 31

480

L m-2 h-1 bar-1 permeability projected for RO-HR with 3% AqpZ would be adequate for SWRO.

481

Increased water permeability may incentivize usage of RO-HP (or even NF), depending on the

482

application and the achievable channel density in the biomimetic layer.

483

Size-selective TFC membranes may also be advantageous for fabricating biomimetic layers.

484

The chemistry can be tailored to provide desired functional groups for anchoring or attaching of

485

the biomimetic layer. Additionally, the dense surface should provide enhanced mechanical

486

integrity, compared to porous surfaces, for the soft materials (proteins and block copolymers)

487

that comprise the biomimetic layer. Another interesting consideration from our results deals with

488

the channel itself. Since the support layer will be essential for neutral solute removal, channels

489

that perfectly reject salt but allow passage of small neutral solutes, such as potentially carbon

490

nanotubes20, would be nearly as effective in biomimetic membranes as the perfectly-selective

491

aquaporin. One important exception would be the increased boron rejection projected when

492

using aquaporin.

493

It is also important to consider the assumptions inherent in our modeling, which not only

494

affect interpretation of our results but also how these membranes should be designed. First, we

495

assumed that there is minimal lateral transport between the biomimetic layer and the underlying

496

layer. As a result, any defects that are present only affect the relevant defect area. If a liquid-

497

filled gap or a gutter layer with extremely high permeability is used between a low-permeability

498

support (e.g., RO-HR) and a biomimetic layer, this assumption would not apply. Second, we

499

assumed that defect edges have negligible influence on transport, which is likely an accurate

500

assumption considering that the vesicle-rupture technique that has been most commonly used can

501

often produce defects >10 nm in diameter.10, 50 Lastly, we neglected roughness in the support

502

layer, which is most relevant for fully-aromatic polyamide RO membranes that typically have a

503

rough, crumpled-nanofilm morphology.51, 52 Our models considered the projected area, thus not

504

taking into account roughness and the correspondingly increased actual surface area.

505

Additionally, the crumpled-nanofilm morphology would likely complicate the fabrication of a

506

biomimetic membrane. Recently developed techniques to make smooth polyamide films53,

507

would likely be essential.

508

18 ACS Paragon Plus Environment

54

Page 19 of 31

Environmental Science & Technology

509

ASSOCIATED CONTENT

510

Supporting Information

511

Derivation of eq 17; trade-off between water permeability and salt rejection (Figure S1);

512

projected salt passage for membranes with 10% AqpZ (Figure S2); solute permeability

513

relationships with van der Waals molar volume for TFC RO membranes (Figure S3); pore size

514

estimation for NF-270 at pH 7 (Figure S4); pore size estimation for NF-270 at pH 3 (Figure S5);

515

projected solute retention for MDM-based biomimetic membrane (Figure S6); dioxane

516

permeability measurements for PB-PEO bilayers (Figure S7); solutes assessed and measured real

517

rejections and permeabilities for RO membranes (Table S1); estimated pore sizes for different

518

solutes for NF-270 (Table S2); physicochemical characteristics of neutrally-charged

519

micropollutants (Table S3). This information is available free of charge via the Internet at

520

http://pubs.acs.org/.

521 522

AUTHOR INFORMATION

523

Corresponding Author

524

*Phone: (203) 432-2789; e-mail: [email protected].

525

Notes

526

The authors declare no competing financial interest.

527 528

ACKNOWLEDGMENTS

529

We acknowledge financial support received from the National Science Foundation (NSF)

530

through the Engineering Research Center for Nanotechnology-Enabled Water Treatment (EEC-

531

1449500) and via Grant CBET 1437630. We also acknowledge the NSF Graduate Research

532

Fellowships awarded to J.R.W. (DGE-1122492) and C.J.P. (DGE-1752134), and the Abel

533

Wolman Fellowship from the American Water Works Association awarded to J.R.W.

19 ACS Paragon Plus Environment

Environmental Science & Technology

534

REFERENCES

535 536

(1)

Mekonnen, M. M.; Hoekstra, A. Y., Four billion people facing severe water scarcity. Sci. Adv. 2016, 2, (2), e1500323.

537 538

(2)

Werber, J. R.; Osuji, C. O.; Elimelech, M., Materials for next-generation desalination and water purification membranes. Nat. Rev. Mater. 2016, 1, (5), 16018.

539 540 541

(3)

National Research Council. Water Reuse: Potential for Expanding the Nation's Water Supply Through Reuse of Municipal Wastewater. National Academies Press: Washington, DC, 2012.

542 543

(4)

Elimelech, M.; Phillip, W. A., The future of seawater desalination: energy, technology, and the environment. Science 2011, 333, (6043), 712-7.

544

(5)

Technical Data Sheet for DOW FILMTEC™ SW30XHR. www.dow.com

545

(6)

Technical Data Sheet for LG SW SR. www.lgwatersolutions.com

546 547 548

(7)

Werber, J. R.; Deshmukh, A.; Elimelech, M., The critical need for increased selectivity, not increased water permeability, for desalination membranes. Environ. Sci. Technol. Lett. 2016, 3, (4), 112-120.

549 550 551

(8)

Cohen-Tanugi, D.; McGovern, R. K.; Dave, S. H.; Lienhard, J. H.; Grossman, J. C., Quantifying the potential of ultra-permeable membranes for water desalination. Energy Environ. Sci. 2014, 7, (3), 1134-1141.

552 553 554

(9)

Karci, A., Degradation of chlorophenols and alkylphenol ethoxylates, two representative textile chemicals, in water by advanced oxidation processes: the state of the art on transformation products and toxicity. Chemosphere 2014, 99, 1-18.

555 556

(10) Shen, Y. X.; Saboe, P. O.; Sines, I. T.; Erbakan, M.; Kumar, M., Biomimetic membranes: A review. J. Membr. Sci. 2014, 454, 359-381.

557 558 559

(11) Kumar, M.; Grzelakowski, M.; Zilles, J.; Clark, M.; Meier, W., Highly permeable polymeric membranes based on the incorporation of the functional water channel protein Aquaporin Z. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, (52), 20719-24.

560 561

(12) Mecke, A.; Dittrich, C.; Meier, W., Biomimetic membranes designed from amphiphilic block copolymers. Soft Matter 2006, 2, (9), 751.

562 563 564 565

(13) Ren, T.; Erbakan, M.; Shen, Y.; Barbieri, E.; Saboe, P.; Feroz, H.; Yan, H.; McCuskey, S.; Hall, J. F.; Schantz, A. B.; Bazan, G. C.; Butler, P. J.; Grzelakowski, M.; Kumar, M., Membrane protein insertion into and compatibility with biomimetic membranes. Adv. Biosyst. 2017, 1, (7), 1700053.

566 567 568

(14) Borgnia, M. J.; Kozono, D.; Calamita, G.; Maloney, P. C.; Agre, P., Functional reconstitution and characterization of AqpZ, the E. coli water channel protein. J. Mol. Biol. 1999, 291, (5), 1169-79.

569 570 571

(15) Borgnia, M. J.; Agre, P., Reconstitution and functional comparison of purified GlpF and AqpZ, the glycerol and water channels from Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, (5), 2888-93. 20 ACS Paragon Plus Environment

Page 20 of 31

Page 21 of 31

Environmental Science & Technology

572 573

(16) Hub, J. S.; de Groot, B. L., Mechanism of selectivity in aquaporins and aquaglyceroporins. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 1198-1203.

574 575 576 577

(17) Licsandru, E.; Kocsis, I.; Shen, Y. X.; Murail, S.; Legrand, Y. M.; van der Lee, A.; Tsai, D.; Baaden, M.; Kumar, M.; Barboiu, M., Salt-excluding artificial water channels exhibiting enhanced dipolar water and proton translocation. J. Am. Chem. Soc. 2016, 138, (16), 54039.

578 579 580 581

(18) Shen, Y. X.; Si, W.; Erbakan, M.; Decker, K.; De Zorzi, R.; Saboe, P. O.; Kang, Y. J.; Majd, S.; Butler, P. J.; Walz, T.; Aksimentiev, A.; Hou, J. L.; Kumar, M., Highly permeable artificial water channels that can self-assemble into two-dimensional arrays. Proc. Natl. Acad. Sci. U.S.A. 2015, 112, (32), 9810-5.

582 583 584 585

(19) Geng, J.; Kim, K.; Zhang, J.; Escalada, A.; Tunuguntla, R.; Comolli, L. R.; Allen, F. I.; Shnyrova, A. V.; Cho, K. R.; Munoz, D.; Wang, Y. M.; Grigoropoulos, C. P.; Ajo-Franklin, C. M.; Frolov, V. A.; Noy, A., Stochastic transport through carbon nanotubes in lipid bilayers and live cell membranes. Nature 2014, 514, (7524), 612-5.

586 587 588

(20) Tunuguntla, R. H.; Henley, R. Y.; Yao, Y. C.; Pham, T. A.; Wanunu, M.; Noy, A., Enhanced water permeability and tunable ion selectivity in subnanometer carbon nanotube porins. Science 2017, 357, (6353), 792-796.

589 590

(21) Corry, B., Designing carbon nanotube membranes for efficient water desalination. J. Phys. Chem. B 2008, 112, 1427-34.

591 592 593

(22) Song, W.; Lang, C.; Shen, Y.-x.; Kumar, M., Design Considerations for Artificial Water Channel–Based Membranes. Annu. Rev. Mater. Sci. 2018, 48, (1), annurev-matsci-070317124544.

594 595 596 597

(23) Zhao, Y.; Qiu, C. Q.; Li, X. S.; Vararattanavech, A.; Shen, W. M.; Torres, J.; Helix-Nielsen, C.; Wang, R.; Hu, X.; Fane, A. G.; Tang, C. Y., Synthesis of robust and high-performance aquaporin-based biomimetic membranes by interfacial polymerization-membrane preparation and RO performance characterization. J. Membr. Sci. 2012, 423, 422-428.

598 599

(24) Perry, M.; Madsen, S.; Jørgensen, T.; Braekevelt, S.; Lauritzen, K.; Hélix-Nielsen, C., Challenges in Commercializing Biomimetic Membranes. Membranes 2015, 5, 685-701.

600 601

(25) Werber, J. R.; Elimelech, M., Permselectivity limits of biomimetic desalination membranes. Sci. Adv. 2018, 4, (6), eaar8266.

602 603

(26) Saeki, D.; Yamashita, T.; Fujii, A.; Matsuyama, H., Reverse osmosis membranes based on a supported lipid bilayer with gramicidin A water channels. Desalination 2015, 375, 48-53.

604 605 606

(27) Wang, M.; Wang, Z.; Wang, X.; Wang, S.; Ding, W.; Gao, C., Layer-by-layer assembly of aquaporin Z-incorporated biomimetic membranes for water purification. Environ. Sci. Technol. 2015, 49, (6), 3761-8.

607 608 609

(28) Werber, J. R.; Bull, S. K.; Elimelech, M., Acyl-chloride quenching following interfacial polymerization to modulate the water permeability, selectivity, and surface charge of desalination membranes. J. Membr. Sci. 2017, 535, 357-364.

21 ACS Paragon Plus Environment

Environmental Science & Technology

610 611 612

(29) Nghiem, L. D.; Schafer, A. I.; Elimelech, M., Removal of natural hormones by nanofiltration membranes: measurement, modeling, and mechanisms. Environ. Sci. Technol. 2004, 38, (6), 1888-96.

613 614

(30) Geise, G. M.; Paul, D. R.; Freeman, B. D., Fundamental water and salt transport properties of polymeric materials. Prog. Polym. Sci. 2014, 39, (1), 1-42.

615 616 617

(31) Hoek, E. M. V.; Kim, A. S.; Elimelech, M., Influence of crossflow membrane filter geometry and shear rate on colloidal fouling in reverse osmosis and nanofiltration separations. Environ. Eng. Sci. 2002, 19, 357-372.

618 619

(32) Sutzkover, I.; Hasson, D.; Semiat, R., Simple technique for measuring the concentration polarization level in a reverse osmosis system. Desalination 2000, 131, (1-3), 117-127.

620 621 622

(33) Tang, C. Y.; Kwon, Y.-N.; Leckie, J. O., Effect of membrane chemistry and coating layer on physiochemical properties of thin film composite polyamide RO and NF membranes. Desalination 2009, 242, (1-3), 149-167.

623 624

(34) Henis, J. M. S.; Tripodi, M. K., Composite hollow fiber membranes for gas separation: the resistance model approach. J. Membr. Sci. 1981, 8, 233-246.

625 626

(35) Koros, W. J.; Fleming, G. K., Membrane-based gas separation. J. Membr. Sci. 1993, 83, (1), 1-80.

627 628 629

(36) Sagle, A. C.; Van Wagner, E. M.; Ju, H.; McCloskey, B. D.; Freeman, B. D.; Sharma, M. M., PEG-coated reverse osmosis membranes: Desalination properties and fouling resistance. J. Membr. Sci. 2009, 340, 92-108.

630 631 632 633

(37) Van Wagner, E. M.; Sagle, A. C.; Sharma, M. M.; La, Y.-H.; Freeman, B. D., Surface modification of commercial polyamide desalination membranes using poly(ethylene glycol) diglycidyl ether to enhance membrane fouling resistance. J. Membr. Sci. 2011, 367, 273287.

634 635 636

(38) Horner, A.; Zocher, F.; Preiner, J.; Ollinger, N.; Siligan, C.; Akimov, S. A.; Pohl, P., The mobility of single-file water molecules is governed by the number of H-bonds they may form with channel-lining residues. Sci. Adv. 2015, 1, e1400083.

637 638

(39) Wijmans, J. G.; Baker, R. W., The solution-diffusion model: a review. J. Membr. Sci. 1995, 107, (1-2), 1-21.

639 640 641

(40) Schwarzenbach, R. P.; Escher, B. I.; Fenner, K.; Hofstetter, T. B.; Johnson, C. A.; von Gunten, U.; Wehrli, B., The challenge of micropollutants in aquatic systems. Science 2006, 313, (5790), 1072-7.

642 643 644

(41) Shen, M.; Keten, S.; Lueptow, R. M., Dynamics of water and solute transport in polymeric reverse osmosis membranes via molecular dynamics simulations. J. Membr. Sci. 2016, 506, 95-108.

645 646

(42) Cohen, M. H.; Turnbull, D., Molecular Transport in Liquids and Glasses. J. Chem. Phys. 1959, 31, 1164-1169.

647

(43) Bondi, A., van der Waals volumes and radii. J. Phys. Chem. 1964, 68, (3), 441-451. 22 ACS Paragon Plus Environment

Page 22 of 31

Page 23 of 31

Environmental Science & Technology

648 649 650

(44) Zhao, Y. H.; Abraham, M. H.; Zissimos, A. M., Fast calculation of van der Waals volume as a sum of atomic and bond contributions and its application to drug compounds. J. Org. Chem. 2003, 68, (19), 7368-73.

651 652

(45) Deen, W. M., Hindered transport of large molecules in liquid filled pores. AIChE J. 1987, 33, 1409-1425.

653 654

(46) Bungay, P. M.; Brenner, H., The motion of a closely-fitting sphere in a fluid-filled tube. Int. J. Multiphase Flow 1973, 1, 25-56.

655 656 657

(47) Boo, C.; Wang, Y.; Zucker, I.; Choo, Y.; Osuji, C. O.; Elimelech, M., High Performance Nanofiltration Membrane for Effective Removal of Perfluoroalkyl Substances at High Water Recovery. Environ. Sci. Technol. 2018, 52, (13), 7279-7288.

658 659

(48) Nghiem, L. D.; Schäfer, A. I.; Elimelech, M., Nanofiltration of Hormone Mimicking Trace Organic Contaminants. Sep. Sci. Technol. 2005, 40, (13), 2633-2649.

660 661 662

(49) Mitch, W. A.; Sharp, J. O.; Trussell, R. R.; Valentine, R. L.; Alvarez-Cohen, L.; Sedlak, D. L., N-nitrosodimethylamine (NDMA) as a drinking water contaminant: A review. Environ. Eng. Sci. 2003, 20, (5), 389-404.

663 664

(50) Richter, R.; Bérat, R.; Brisson, A., Formation of solid-supported lipid bilayers: an integrated view. Langmuir 2006, 3497-3505.

665 666 667 668

(51) Kłosowski, M. M.; McGilvery, C. M.; Li, Y.; Abellan, P.; Ramasse, Q.; Cabral, J. T.; Livingston, A. G.; Porter, A. E., Micro-to nano-scale characterisation of polyamide structures of the SW30HR RO membrane using advanced electron microscopy and stain tracers. J. Membr. Sci. 2016, 520, 465-476.

669 670 671

(52) Pacheco, F.; Sougrat, R.; Reinhard, M.; Leckie, J. O.; Pinnau, I., 3D visualization of the internal nanostructure of polyamide thin films in RO membranes. J. Membr. Sci. 2016, 501, 33-44.

672 673

(53) Karan, S.; Jiang, Z.; Livingston, A. G., Sub-10 nm polyamide nanofilms with ultrafast solvent transport for molecular separation. Science 2015, 348, 1347-1351.

674 675 676

(54) Park, S.-J.; Ahn, W.-G.; Choi, W.; Park, S.-H.; Lee, J. S.; Jung, H. W.; Lee, J.-H., A facile and scalable fabrication method for thin film composite reverse osmosis membranes: duallayer slot coating. J. Mater. Chem. A 2017, 5, 6648-6655.

677 678 679

(55) Epsztein, R.; Cheng, W.; Shaulsky, E.; Dizge, N.; Elimelech, M., Elucidating the mechanisms underlying the difference between chloride and nitrate rejection in nanofiltration. J. Membr. Sci. 2017, 548, 694-701.

680

23 ACS Paragon Plus Environment

Environmental Science & Technology

681 682

Table of Contents Graphic

24 ACS Paragon Plus Environment

Page 24 of 31

Page 25 of 31

Environmental Science & Technology

683 684

Figure 1. The two main designs of composite biomimetic desalination membranes considered in

685

this study. (A) The biomimetic active layer, comprising amphiphilic block copolymers and

686

biological or synthetic water channels, is formed directly on an ultrafiltration (UF) membrane.

687

Defects in the biomimetic layer expose the underlying pores. UF membrane pore diameters of 3,

688

5, and 8 nm are considered. (B) The biomimetic active layer is formed on a “dense” size-

689

selective membrane, such as fully-aromatic polyamide reverse-osmosis (RO) membranes and

690

semi-aromatic polyamide nanofiltration (NF) membranes. Defects in the biomimetic layer

691

expose the polyamide layer.

692

25 ACS Paragon Plus Environment

Environmental Science & Technology

693 694

Figure 2. Projected water permeation through biomimetic membranes with various support

695

layers. (A) Water permeability in defect-free regions of a composite biomimetic membrane, ABM,

696

as a function of the support-layer permeability, ASL. Measured support-layer permeabilities

697

(Table 1) and a biomimetic layer permeability, ABL, corresponding to 3% aquaporin-Z (AqpZ) by

698

area were used in eq 6 for calculations. A water permeability of 2.54 L m-2 h-1 bar-1 was used for

699

the biomimetic active layer, which was calculated from the single-channel permeability of

700

AqpZ.25, 38 (B) Defect-driven flow of pure water as a function of support layer type and the total

701

defect area in the biomimetic active layer. Normalized flow through defects was determined

702

using eqs 7 and 8 for the labeled support layers and an active layer comprising 3% AqpZ.

26 ACS Paragon Plus Environment

Page 26 of 31

Page 27 of 31

Environmental Science & Technology

703 704

Figure 3. Projected salt rejection, Ro, for composite biomimetic desalination membranes

705

comprising a defect-containing biomimetic active layer with 3% aquaporin-Z formed on

706

different support layers. Transport was modeled at standard seawater RO (SWRO) test

707

conditions (55.1 bar applied pressure, 32000 ppm NaCl) with a mass transfer coefficient of 140

708

L m-2 h-1. The green shaded area corresponds to rejections greater than 99.85%, the highest

709

rejection currently achieved by polyamide thin-film composite membranes.5, 6 Dash-dotted drop

710

lines correspond to the maximum defect area, θ, able to achieve this rejection level when using a

711

porous support layer (blue, UF-3; 0.002%) and a highly-rejecting dense support layer (black,

712

RO-HR, 16%).

713 714

27 ACS Paragon Plus Environment

Environmental Science & Technology

715 716

Figure 4. Neutral solute permeation through aromatic polyamide reverse-osmosis (RO)

717

membranes considered as model support layers for biomimetic desalination membranes. (A)

718

Real solute rejection, Rr, as a function of molecular weight through a “high-rejection” RO

719

membrane (RO-HR, SW30XLE, Dow) and a “high-permeability” RO membrane (RO-HP, XLE,

720

Dow). (B) Solute permeability (P in m/s) as a function of the van der Waals radius, rvdw.

721

Regression lines were used to determine parameters for eq 14. Solute characteristics and

722

measurements are listed in Table S1. Some data for RO-HR are from ref 25 (Table S1).

723

Measurements were conducted at 25.0 ± 0.5 ºC, pH 7.0 ± 0.5, and 21 cm/s crossflow velocity,

724

with applied pressures of 15.5 bar for RO-HR and 8.6 bar for RO-HP.

725 726 28 ACS Paragon Plus Environment

Page 28 of 31

Page 29 of 31

Environmental Science & Technology

727 728

Figure 5. Projected observed solute rejections, Ro, for neutral solutes and micropollutants.

729

Rejection was calculated from hydrophobicity-based permeability relationships for the

730

biomimetic active layer and size-based permeability relationships for the support layer. For the

731

biomimetic active layer, eq 12 was used to estimate permeability from the solute octanol-water

732

partition coefficient (Kow), considering a defect-free layer with 3% aquaporin-Z and a polymer

733

matrix based on poly(1,2-butadiene)-poly(ethylene oxide) (PB-PEO). PB-PEO has a relatively

734

low permeability for bilayer structures.25 For reverse-osmosis membranes (RO-HP and RO-HR)

735

as model support layers, composite permeabilities were calculated using eq 14 with parameters

736

determined from Figure 4. For nanofiltration (NF) and ultrafiltration (UF-3) membranes as

737

model support layers, rejections were calculated using eqs 15−17 and pore radii of 0.51 nm for

738

NF (from thymol, Table S2) and 1.5 nm for UF-3. Overlaid are the properties of micropollutants

739

of interest during reuse of municipal wastewater, which are listed in Table S3. Rejection was

740

calculated for 15.5 bar applied pressure, 1000 ppm NaCl feed, and a mass transfer coefficient of

741

140 L m-2 h-1.

742

29 ACS Paragon Plus Environment

Environmental Science & Technology

743 744

Figure 6. Projected permeation of important neutral solutes. (A) Measured permeabilities of

745

boron (as boric acid) and 1,4-dioxane for thin-film composite RO and NF membranes as model

746

support layers for a biomimetic desalination membrane. Observed boron rejection by NF (NF-

747

270) was 5.3 ± 1.5% at 5.0 bar, corresponding to a permeability of 660 ± 180 L m-2 h-1 if the

748

solution–diffusion model is used. (B) Projected observed rejection of boron and 1,4-dioxane by a

749

composite biomimetic desalination membrane comprising various support layers and a selective

750

layer with 3% aquaporin-Z, an amphiphilic matrix of either poly(1,2-butadiene)-poly(ethylene

751

oxide) (PB-PEO) or poly(methoxazoline)-poly(dimethylsiloxane)-poly(methoxazoline) (MDM),

752

and 1% defect area. Boron rejections were calculated for standard seawater RO test conditions

753

(55.1 bar, 32000 ppm NaCl) using permeabilities explicitly measured for each material at pH 7.

754

Dioxane permeability was too rapid to be measured for the polymeric matrices using solution-

755

based techniques (Figure S7). Dioxane permeability was therefore estimated from its log Kow

756

(−0.27) using eq 12, and rejection was calculated for 15.5 bar applied pressure and a feed of

757

1000 ppm NaCl. Mass transfer coefficients of 140 L m-2 h-1 were used in all cases.

758

30 ACS Paragon Plus Environment

Page 30 of 31

Page 31 of 31

Environmental Science & Technology

759

Table 1. Measured water and salt permeabilities for membranes considered as model support

760

layers for a composite biomimetic desalination membrane. Permeabilities were measured for RO

761

and NF membranes and calculated for UF membranes.

support layer type

reverse osmosis (RO)

nanofiltration (NF) ultrafiltration (UF)

no support

considered example

SW30XLE, Dow (RO-HR) XLE, Dow (RO-HP) NF270, Dow (NF) 3 nm pores (UF-3)a 5 nm pores (UF-5)a 8 nm pores (UF-8)a --

water NaCl permeability, permeability, A (L m−2 h−1 B (L m−2 −1 bar ) h−1)

measured NaCl rejection, Ro (%)c

projected water permeability, A, with biomimetic layer (L m−2 h−1 bar−1)d 3% 10% AqpZ AqpZ

2.20 ± 0.02

0.26 ± 0.06

98.6 ± 0.3

1.18

1.74

5.57 ± 0.38

0.93 ± 0.08

95.8 ± 0.5

1.74

3.35

20.5 ± 0.5

52.7 ± 3.8b

38.7 ± 2.0

2.26

5.95

114

--

--

2.48

7.81

316

--

--

2.52

8.17

809

--

--

2.53

8.30

--

--

--

2.54

8.39

762 2 b

763

a

764

ion rejection varies with ionic strength and pH.55 The tabulated value was determined for 2000 ppm NaCl,

765

pH 7, and 25 ºC, measured at 3.5, 5.0, and 6.5 bar. c Measured using 2000 ppm NaCl at pH 7, 25 ºC and

766

21 cm/s crossflow at 15.5, 8.6, and 5.0 bar for SW30XLE, XLE, and NF270, respectively. d Calculated

767

using eq 6.

Assumed monodisperse pores with 10% surface porosity and 100-nm active layer thickness.

768 769 770 771 772

31 ACS Paragon Plus Environment

NF-270