Letter pubs.acs.org/NanoLett
Ultrathin Free-Standing Bombyx mori Silk Nanofibril Membranes Shengjie Ling,†,‡ Kai Jin,† David L. Kaplan,*,‡ and Markus J. Buehler*,†,§,∥ †
Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States ‡ Department of Biomedical Engineering, Tufts University, Medford, Massachusetts 02155, United States § Center for Materials Science and Engineering and ∥Center for Computational Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, United States S Supporting Information *
ABSTRACT: We report a new ultrathin filtration membrane prepared from silk nanofibrils (SNFs), directly exfoliated from natural Bombyx mori silk fibers to retain structure and physical properties. These membranes can be prepared with a thickness down to 40 nm with a narrow distribution of pore sizes ranging from 8 to 12 nm. Typically, 40 nm thick membranes prepared from SNFs have pure water fluxes of 13 000 L h−1 m−2 bar−1, more than 1000 times higher than most commercial ultrathin filtration membranes and comparable with the highest water flux reported previously. The commercial membranes are commonly prepared from polysulfone, poly(ether sulfone), and polyamide. The SNF-based ultrathin membranes exhibit efficient separation for dyes, proteins, and colloids of nanoparticles with at least a 64% rejection of Rhodamine B. This broad-spectrum filtration membrane would have potential utility in applications such as wastewater treatment, nanotechnology, food industry, and life sciences in part due to the protein-based membrane polymer (silk), combined with the robust mechanical and separation performance features. KEYWORDS: Silk nanofibrils, ultrathin membrane, filtration, separation
N
nanoporous membranes with a thickness of 23−47 nm, but the separation performance left room for improvement; the rejection (calculated by dividing the amount of the molecules in the retentate or permeate relative to the original amount in the feed solution) of cytochrome c was 49.0±0.3% for 47 nm thick membranes.9 Silk nanofibrils (SNFs), as the fundmental nanoconstruction building blocks of silk fibers,13−15 have not been used in the preparation of filtration membranes because the methods to extract or prepare well-dispersed SNFs remain challenging. Here, we report a new free-standing ultrathin filtration membrane consisting of pure protein in the form of SNFs, exfoliated from native Bombyx mori (B. mori) silk fibers to retain structure and physical properties. The four-step route used to fabricate the SNF filtration membranes is summarized in Figure 1. The first three steps were utilized to exfoliate the SNFs from silk fibers, as found in our recent report.16 Briefly, The degummed B. mori silk fibers were immersed in hexafluoroisopropanol (HFIP) solution with weight ratio of 1:30 and were mixed by vortexing. Then the sealed silk fiber/HFIP mixture was incubated at 60 °C to partially dissolve the silk fibers to silk microfibrils (SMFs) (Supporting Information, Figure S1a). After 24 h, the resultant SMF pulp blend (Supporting Information, Figure S1b) was dried to evaporate the HFIP. After totally drying (about 4 h), the SMFs were placed in water with a weight ratio of
anoporous membranes are important for separation processes and in wastewater treatment, nanotechnology, food industry, and life sciences.1−3 However, available commercial membranes usually have a poor solvent resistance and a thick skin layer with a broad pore size distribution, leading to low permeation flux and poor cutoff properties.1,4 Recently, there has been a renewed focus on developing ultrathin filtration membranes that can provide maximal permeability while retaining highly separation efficiency. Zerodimensional nanoparticles (e.g., ferritin,4 gold,5 and polystyrene6), one-dimensional nanofibers (e.g., inorganic nanofibers or nanowires,3,7 carbon nanotubes,8 and cellulose nanofibers9) and two-dimensional laminar materials (e.g., graphene oxide,10 molybdenum disulfide,11 and tungsten disulfide12) have been used to fabricate ultrathin nanoporous membranes. However, the challenge remains to prepare ultrathin and low-cost filtration membranes while retaining mechanical strength and good separation performance. Ultrathin membranes made of one-dimensional nanofibers have several unique advantages to address this challenge, that is, high porosity, good flexibility, large surface area per unit volume, and interconnected open pore structures.7 However, most nanofiber porous membranes are prepared from inorganic and polymer nanofibers or nanowires. In contrast, biopolymer-based nanofiber ultrathin membranes of the same genre have not been widely explored due to the lack of approaches for the preparation of biopolymer nanofiber dispersions in the range of a few to tens on the nanometer scale. Recently, sub-10 nm wide cellulose nanofibers have been used to prepare ultrathin © XXXX American Chemical Society
Received: March 21, 2016 Revised: April 11, 2016
A
DOI: 10.1021/acs.nanolett.6b01195 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
1:200 with continuous stirring or shaking, followed by removal of undissolved silk. Finally, the SMFs/water mixture (Supporting Information, Figure S1c) was sonicated at 120 μm amplitude and 20 kHz frequency at intervals of 10 s. After 1h, the exfoliated SNF dispersion was harvested by centrifugation at 10 000 rpm for 20 min. This dispersion was transparent (insert in Figure 2a) and was stable over several months. Scanning electron microscopy (SEM) images revealed that the extracted SNFs had a diameter of 20 ± 5 nm and a contour length in the range of 300−500 nm (Figure 2a), similar to the diameter of single SNFs found in silk fibers.17 The fourth and the last step in the process was to assemble the SNFs to ultrathin membranes, using a vacuum filtration process that has been previously reported.18,19 The resultant membranes with thickness of ∼520 nm could be removed from the supporting substrate and appeared homogeneous and transparent with structural color on the surface (Figure 2b). These free-standing membranes were also robust and flexible and could be cut and bent without damage (Supporting Information, Figure S2). Compared with other ultrathin membranes that lack flexibility, such as inorganic nanofibers, nanowires, and nanosheet based membranes,3 the mechanical superiority of these SNF membranes permit use in pressure-driven filtration operations, even at high applied pressures. Different from cast silk fibroin membranes which will dissolve in water if not treated with alcohol or by water annealing to generate β-sheet secondary structures,20 these SNF membranes were stable in water without dissolution (Supporting Information, Figure S3). This property, as a critical role in filtration,
Figure 1. Schematic of the process used to design ultrathin SNF membranes. Step 1, silk fiber immersed in HFIP with a weight ratio of 1:30 and incubated at 60 °C for 24 h to obtain silk fiber/SMF slurries. Step 2, the dried silk fiber slurries transferred to H2O solution and precipitates removed. Step 3, SMF dispersion treated by ultrasound to extract SNFs. Step 4, SNF dispersion assembled to ultrathin SNF membranes via vacuum filtration.
Figure 2. Visual appearance and structural characterization of the SNF dispersion and membranes. (a) SEM image of exfoliated SNFs. The insert is a photograph of SNF dispersion at room temperature for 1 month. (b) Picture of a free-standing SNF membrane with a thickness about 520 nm under visual light with structural color. (c) FTIR spectra of SNF membrane and degummed silk fibers. (d−f) SEM images of SNF membranes with a thickness of 520 nm. The images (e) and (f) are top view and cross-sectional SEM images of membranes, respectively. (g) Pore size distribution of the SNF membranes with thicknesses of 40, 60, and 120 nm. The relveant SEM images can be found in Supporting Information. B
DOI: 10.1021/acs.nanolett.6b01195 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters ensures the stability of membrane without collapse in the filtration process. Fourier transform infrared spectroscopy (FTIR) was utilized to assess structural details of the SNF membrane (Figure 2c). The amide I band showed a similar shape to degummed silk fibers: a sharp peak at 1620 m−1 and a shoulder at 1695 cm−1, which are assigned to β-sheets and β-turns of the hairpin-folded antiparallel β-sheet structure,21 respectively, indicating that the SNFs were mainly composed of β-sheet. Such structure not only gives the membranes mechanical robustness but also contributes the excellent stability in water, as with native silk fibers.22−24 Structural insight into the mesoscopic structure of the membranes was obtained via SEM. A free-standing membrane with a thickness of 520 nm revealed a uniform fibrous structure (Figure 2d). The related surface (Figure 2e) and cross-section SEM images (Figure 2f) showed a uniform pore size distribution with diameters of 6 ± 2 nm with interconnected pores that were uniform without cracks or pinholes. The thickness of membrane was controllable by adjusting the concentration and volume of the SNFs dispersed during vacuum filtration. In a typical procedure, 5 mL of dispersion with a concentration of 0.1 wt % generated a membrane of about 1000 nm thick using a casting mold of 3.5 cm in diameter, and the volume of the SNF dispersion was linearly correlated to the membrane thickness (Supporting Information, Figure S4). Additionally, pore size could be tuned to some extent by choosing the appropriate filtration volume and concentration of SNF dispersion. For instance, the average pore size varied from 12 to 8 nm (Figure 2g) with the SNF membrane thickness increased from 40 nm (0.25 mL 0.1 wt % SNF solution, Supporting Information, Figure S5a,b) to 60 nm (0.5 mL 0.1 wt % SNF solution, Supporting Information, Figure S5c,d). The pore size distribution could also be narrowed through control of membrane thickness. For example, the 120 nm thick membrane had the narrowest pore size distribution when compared to the 40 and 60 nm thick membranes (Figure 2g and Supporting Information, Figure S5). These results suggest utility for these types of membranes related to the separation of colloidal particles and molecules of varying sizes (Figure 3a). To evaluate the permeation performance of these SNF membranes, pure water fluxes were assessed through membranes with different thicknesses from 40 to 1500 nm. Surprisingly, the flux of 40 nm thick SNF membrane was up to 13 000 L h−1 m−2 bar−1, more than 1000 times higher than fluxes of commercial filtration membranes25−27 and better than fluxes of most advanced recently reported ultrathin membranes (Figure 3b).4,6,9−12,28,29 By increasing the thickness of the membrane, the water flux sharply declined, consistent with the Hagen−Poiseuille theoretical models30 (red plot in Figure 3b) and other experimental reports (Figure 3b).4,6,9−12,28,29 However, the water flux was larger than 600 L h−1 m−2 bar−1, even with the thickness increased to 1500 nm, this flux was also faster than that of most commercial materials.25−27 The separation performance of the SNF membranes was measured through pressure-driven filtration. First, Rhodamine B was used to study the influence of membrane thickness on separation performance (Figure 4a). By increasing the membrane thickness, both visual and fluorescence color of the permeated solution became lighter and transparent, and no fluorescence was observed when the thickness reached 295 nm. The UV−vis spectra (Figure 4b) confirmed the visual results with an absorption at 554 nm that dropped gradually with no peak appearing for the 295 nm thick membrane. The
Figure 3. Filtration through SNF membranes. (a) Schematic showing how SNF membranes reject large molecules while allowing small molecules to pass through. (b) Thickness-dependent changes in permability to pure water. The red circles are fluxes determined by using polycarbonate membranes with an effective surface area of 0.962 cm2 (porosity, 10%). The red solid line is a fitted curve using the Hagen−Poiseuille equation.30 The comparison of pure water flux of SNF membranes with other materials used inultrathin filtration membranes is presented. Abbreviations: Cellulose NFs, cellulose nanofibers; SPEK-C, sulfonated polyetherketone with cardo groups; WS2, chemically exfoliated tungsten disulfide nanosheets; MoS2, atomthick molybdenum disulfide sheets.
calculation demonstrated that rejection reached equilibrium with a value of 96% with a thickness of 295 nm (insert plot in Figure 4b). Next, protein molecules (cytochrome c, Cyt. c; bull serum albumin, BSA), colloids (gold nanoparticles, 5 nm diameter; CdSeS/ZnS quantum dots, 6 nm diam.), small molecules (L-tryptophan, 1.1 × 0.5 nm) and ions (Cu2+),3,4,6,8−12,28,31−33 were selected to assess size-selectivity. Small Cu2+ ions (rejection: 3 ± 3%) and L-tryptophan (rejection: 10 ± 2%) freely passed through the channels (Table 1). Cyt. c, BSA, gold nanoparticles and CdSeS/ZnS quantum dots, as larger size objects, had a rejection of 99 ± 1, 100 ± 0, 99 ± 2, and 100 ± 1%, respectively. These results validated that pore size was a crucial factor for separations with the SNF membranes. We further monitored the separation with different dyes, which are usually unsatisfactory in terms of separation with most ultrathin filtration membranes. Their separation performance is summarized in Table 1. Size, shape, and charge of the molecules were key factors for rejection. For example, large size molecules, such as Alcian Blue 8GX (2.5 × 2.3 nm) and Brilliant Blue G (2.3 × 1.8 nm) had a 100 ± 0% rejection, versus small molecules, such as 77 ± 2% rejection for sulfonated naphthalene (ANTS) (1.1 × 1.0 nm), 82 ± 2% for Orange G (1.3 × 0.8 nm), and 84 ± 1% for Eosin B (1.1 × 1.0 nm). Linear molecules even with large molecule length (e.g., Brilliant Yellow with size of 2.4 × 0.8 nm; Direct Red 81 with size of 2.4 × 0.8 nm; Fluorescent Brightener 28 with size of 3.0 × 0.8) could partially permeate C
DOI: 10.1021/acs.nanolett.6b01195 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
Figure 4. Separation performance of SNF membranes. (a) The pictures before and after filtering Rhodamine B aqueous solutions using SNF membranes with different thicknesses. The top and bottom images are under visual and UV light, respectively. (b) UV−vis absorption changes of an aqueous solution of Rhodamine B after filtration with SNF membranes with different thicknesses. The insert plot shows the rejection of Rhodamine B aqueous solution with different thicknesses of the SNF membranes. (c,d) Comparison of separation performance of SNF membranes for cytochrome c (c) and 5 nm gold nanoparticle dispersion (d) with other ultrathin filtration membrane materials. The rejection is represented by the color of the pattern. The blue and red are 0% and 100% rejection, respectively. The hollow five-pointed star pattern is the separation performance of the SNF membranes. In these two figures, the separation performance of the SNF membranes with 40 and 60 nm thicknesses is listed for comparison. Abbreviations: Cd(OH)2, Cu(OH)2, and Zn(OH)2 are Cd(OH)2, Cu(OH)2, and Zn(OH)2 nanofibers, respectively; cellulose, cellulose nanofibers; PPy, polypyrrole-coated cooper hydroxide nanostrands; SWNCT-GO, single wall carbon nanotubes and graphene oxide sheets composites; GO, graphene oxide sheets; PS NP, polystyrene nanoparticles; P4VP, cross-linked poly(4-vinylpyridine); PP2b, (5,5′-bis(1-ethynyl-7 polyethylene glycol-N,N′-bis(ethylpropyl) perylene3,4,9,10-tetracarboxylic diimide)-2,2′-bipyridine; CNF, carbonaceous nanofiber; SPEK-C, sulfonated polyetherketone with cardo groups; WS2, chemically exfoliated tungsten disulfide nanosheets; MoS2, atom thick molybdenum disulfide sheets.
Table 1. Separation Performance of 120 nm Thick SNF Membranes for Dyes, Protein, and Nanoparticlesa Mw (g mol −1) Cu2+ L-tryptophan cytochrome c bull serum albumin gold nanoparticles CdSeS/ZnS quantum dots Alcian Blue 8GX Brilliant Blue G ANTS Orange G Eosin B Brilliant Yellow Direct Red 81 Fluorescent Brightener 28 Rhodamine B Congo Red
134.45 204.23 12400 ∼66 kDa
1298.88 854.02 427.33 452.4 624.06 624.5 675.60 916.98 479.01 696.66
size (nm)
concentration
analyte charge
rejection (%)
1.1 × 0.5 2.5 × 2.5 × 3.7 14 × 4 × 4 5 nm 6 nm 2.5 × 2.3 2.3 × 1.8 1.1 × 1.0 1.3 × 0.8 1.1 × 1.0 2.4 × 0.8 2.4 × 0.8 3.0 × 0.8 1.6 × 1.3 1.9 × 1.3
1 mM 979 μM 2.0 mg mL−1 2.0 mg mL−1 ∼5.5 × 1013 unit per mL 1 mg mL−1 185 μM 398 μM 548 μM 601 μM 320 μM 76 μM 414 μM 124 μM 5 μM 17 μM
+ − − − − − + − − − − − − − + −
3±3 10 ± 2 99 ± 1 100 ± 0 99 ± 2 100 ± 1 100 ± 0 100 ± 0 77 ± 2 82 ± 2 84 ± 1 64 ± 3 80 ± 1 85 ± 2 91 ± 1 86 ± 2
a
Abbreviations: ANTS, 8-aminonaphthalene-1,3,6-trisulfonate. Filtration experiments were conducted on polycarbonate membrane supports (porosity, 10%). Molecular sizes of dyes were calculated using Materials Studio 7.0, the sizes of proteins were calculated from their PDB files. It should be noted that the real sizes of the molecules would be a little bit larger than the measurements made with Materials Studio, because van der Waals radii of atoms were not taken into account in Materials Studio calculation. More details can be found in Supporting Information, Figure S6.
the membrane with rejection rates of 64 ± 3, 80 ± 1, and 85 ± 2%, respectively. In addition, because SNFs are negatively charged at neutral pH (the isoelectric point of silk fibroin is
4.5334), more positively charged molecules can be taken up by the membranes via electrostatic interactions. The positively charged molecules (e.g., Rhodamine B and Alcian Blue 8GX) D
DOI: 10.1021/acs.nanolett.6b01195 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters
and polycarbonate filtration membranes (pore size, 200 nm; diameter 47 mm; Sigma-Aldrich). The Separation Performance Measurements. The separation performances were performed on a vacuum filtration device (Sigma-Aldrich glass vacuum filtration assembly device, membrane diameter of 47 mm, inner diameter of funnel top 35 mm). Water (100 mL) was filtered across the membrane to measure the pure water flux (J, L m−2 h−1 bar−1) that is calculated by J = V/(Atp), where V is the volume of the water filtered (L), A is the effective membrane filtration area (m2), t is the filtration time (h), and p is the suction pressure across the membrane (bar). The filtration area of our filter holder is 9.62 cm2 and the porosity of the PC membrane is 10%. Then, the effective surface area in our case is 0.962 cm2. Dyes, proteins, and gold nanoparticles were used to evaluate the membrane rejection with the feed (20 mL) filtered across the membrane under 1 bar of applied pressure. Permeation was characterized by UV−vis spectrophotometer (SpectraMax M2, Molecular Devices, CA). The rejection (R, %) is calculated by
showed higher rejection than negatively charged molecules with similar sizes (e.g., Congo Red with rejection of 86 ± 2%). Besides the factors discussed above, hydrophobic interactions of SNFs and dye molecules also contributed to the high separation performance of the SNF membranes because most of the dye molecules have benzene rings (Supporting Information, Figure S6), which act with the hydrophobic domain to interact with the silk fibroin chains,35,36 (i.e., GAGAGS peptides that form β-sheets37). Hydrophobic interactions also relate to the mechanism by which carbon nanotubes8 and graphene oxide10 based filtration membranes can separate Rhodamine B from solution, whereas polymer-based filtration membranes38 with similar pore sizes cannot reject the dyes. Figure 4c,d summarizes the separation performance of our SNF membranes for Cyt. c (Figure 4c) and 5 nm gold nanoparticles (Figure 4d) compared with other ultrathin filtration materials from the literatures.3,4,6,8−12,28,31−33 The SNF membranes exhibit a better balance of thickness, flux, and separation performance. The thicknesses of membranes were in the range of 40−60 nm, comparable with most ultrathin membranes. The flux was 1.2−6.2 times higher than that of most inorganic and polymer membranes and comparable with ferritin4 and inorganic nanowires.3 In terms of separation performance, the rejection of protein and gold nanoparticles was higher than that of membranes with similar thickness. In summary, we report a facile strategy to prepare novel ultrathin nanoporous protein membranes consisting of natural SNFs, which retain the exquisite structure and superior physical properties of the natural silk fibers. The resultant free-standing membranes were mechanically robust and flexible with a thickness that is tunable from 40−1500 nm, pore sizes of 8−12 nm, and superior water permeation flux as high as 13 000 L h−1 m−2 bar−1 (40 nm thick membrane). Moreover, the SNF membranes exhibited excellent broad-spectrum separation performance for most dyes, proteins, and nanoparticles. Thus, we anticipate that these new ultrathin SNF membranes have utility in a wide range of applications in the wastewater treatment, nanotechnology, food industries, and life sciences. Methods. Preparation of Degummed Silk Fibers. Bombyx mori (B. mori) silkworm cocoon silk fibers were degummed by boiling in two 30 min changes of 0.5% (w/w) NaHCO3 solution. Then the degummed silk fibers were washed with distilled water and allowed to air-dry at room temperature.20 More details about materials information can be found in the Supporting Information. Liquid Exfoliation of SNFs. The degummed B. mori silk fibers were immersed in HFIP with a weight ratio of 1:30, and sufficiently agitated to make sure that all fibers were immersed. Then the airtight silk fiber/HFIP mixture was incubated at 60 °C and after 24 h the SMF pulp was dried in a fume hood to evaporate the HFIP. After about 4 h, the dry SMFs were added to water with weight ratio of 1:200 under continuous stirring or agitation, followed by removal of any undissolved material. Finally, the silk/water mixture was sonicated at 120 μm amplitude and 20 kHz frequency with interval of 10 s. After 1 h, the exfoliated SNF dispersion was harvested by centrifugation at 10 000 rpm for 20 min. All of the steps were operated in a chemical hood with the necessary precautions due to the use of the HFIP. The Preparation of SNF Membranes. All of the SNF membranes were fabricated by vacuum-filtrating the SNF dispersions through a Sigma-Aldrich vacuum filtration assembly
⎛ Cp ⎞ ⎟ × 100% R = ⎜⎜1 − Cf ⎟⎠ ⎝
(1)
where Cf and Cp are the concentrations of compound in the feed and permeate, respectively. Characterization. The morphology of SNFs and SNF membranes were characterized by SEM (Ultra 55 filed emission scanning electron microscope, Harvard University Center for Nanoscale Systems). The structures of the membranes were characterized by FTIR (Jasco FTIR-6200, Jasco Instruments, Easton, MD). More details can be found in the Supporting Information.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.6b01195. Detailed description of materials and characterization methods and additional figures. (PDF)
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected]. *E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS S.L., D.L.K., and M.J.B. acknowledge support from NIH (U01EB014976) and the AFOSR (FA9550-11-1-0199). REFERENCES
(1) Vandezande, P.; Gevers, L. E. M.; Vankelecom, I. F. J. Chem. Soc. Rev. 2008, 37, 365−405. (2) Shannon, M. A.; Bohn, P. W.; Elimelech, M.; Georgiadis, J. G.; Marinas, B. J.; Mayes, A. M. Nature 2008, 452, 301−310. (3) Karan, S.; Wang, Q. F.; Samitsu, S.; Fujii, Y.; Ichinose, I. J. Membr. Sci. 2013, 448, 270−291. (4) Peng, X. S.; Jin, J.; Nakamura, Y.; Ohno, T.; Ichinose, I. Nat. Nanotechnol. 2009, 4, 353−357. (5) He, J. B.; Lin, X. M.; Chan, H.; Vukovic, L.; Kral, P.; Jaeger, H. M. Nano Lett. 2011, 11, 2430−2435. E
DOI: 10.1021/acs.nanolett.6b01195 Nano Lett. XXXX, XXX, XXX−XXX
Letter
Nano Letters (6) Zhang, Q. G.; Ghosh, S.; Samitsu, S.; Peng, X. S.; Ichinose, I. J. Mater. Chem. 2011, 21, 1684−1688. (7) Liang, H. W.; Wang, L.; Chen, P. Y.; Lin, H. T.; Chen, L. F.; He, D. A.; Yu, S. H. Adv. Mater. 2010, 22, 4691−4695. (8) Gao, S. J.; Qin, H. L.; Liu, P. P.; Jin, J. J. Mater. Chem. A 2015, 3, 6649−6654. (9) Zhang, Q. G.; Deng, C.; Soyekwo, F.; Liu, Q. L.; Zhu, A. M. Adv. Funct. Mater. 2016, 26, 792−800. (10) Huang, H. B.; Song, Z. G.; Wei, N.; Shi, L.; Mao, Y. Y.; Ying, Y. L.; Sun, L. W.; Xu, Z. P.; Peng, X. S. Nat. Commun. 2013, 4, 2979. (11) Sun, L. W.; Huang, H. B.; Peng, X. S. Chem. Commun. 2013, 49, 10718. (12) Sun, L. W.; Ying, Y. L.; Huang, H. B.; Song, Z. G.; Mao, Y. Y.; Xu, Z. P.; Peng, X. S. ACS Nano 2014, 8, 6304−6311. (13) Giesa, T.; Buehler, M. J. Annu. Rev. Biophys. 2013, 42, 651−673. (14) Nova, A.; Keten, S.; Pugno, N. M.; Redaelli, A.; Buehler, M. J. Nano Lett. 2010, 10, 2626−2634. (15) Koh, L.-D.; Cheng, Y.; Teng, C.-P.; Khin, Y.-W.; Loh, X.-J.; Tee, S.-Y.; Low, M.; Ye, E.; Yu, H.-D.; Zhang, Y.-W.; Han, M.-Y. Prog. Polym. Sci. 2015, 46, 86−110. (16) Ling, S. J.; Li, C. M.; Jin, K.; Kaplan, D. L.; Buehler, M. J. unpublished data. (17) Lin, N.; Liu, X. Y. Chem. Soc. Rev. 2015, 44, 7881−7915. (18) Ling, S. J.; Li, C. X.; Adamcik, J.; Wang, S. H.; Shao, Z. Z.; Chen, X.; Mezzenga, R. ACS Macro Lett. 2014, 3, 146−152. (19) Ling, S. J.; Li, C. X.; Adamcik, J.; Shao, Z. Z.; Chen, X.; Mezzenga, R. Adv. Mater. 2014, 26, 4569−4574. (20) Rockwood, D. N.; Preda, R. C.; Yucel, T.; Wang, X. Q.; Lovett, M. L.; Kaplan, D. L. Nat. Protoc. 2011, 6, 1612−1631. (21) Ling, S. J.; Qi, Z. M.; Knight, D. P.; Shao, Z. Z.; Chen, X. Biomacromolecules 2011, 12, 3344−3349. (22) Termonia, Y. Macromolecules 1994, 27, 7378−7381. (23) van Beek, J. D.; Beaulieu, L.; Schafer, H.; Demura, M.; Asakura, T.; Meier, B. H. Nature 2000, 405, 1077−1079. (24) Keten, S.; Xu, Z. P.; Ihle, B.; Buehler, M. J. Nat. Mater. 2010, 9, 359−367. (25) Vankelecom, I. F. J.; De Smet, K.; Gevers, L. E. M.; Jacobs, P. A. Nanofiltration: Principles and Applications, Elsevier: New York, 2005. (26) Lu, Y. Y.; Suzuki, T.; Zhang, W.; Moore, J. S.; Marinas, B. J. Chem. Mater. 2007, 19, 3194−3204. (27) Braeken, L.; Van der Bruggen, B.; Vandecasteele, C. J. Phys. Chem. B 2006, 110, 2957−2962. (28) Deng, C.; Zhang, Q. G.; Han, G. L.; Gong, Y.; Zhu, A. M.; Liu, Q. L. Nanoscale 2013, 5, 11028−11034. (29) Lee, Y. M.; Jung, B.; Kim, Y. H.; Park, A. R.; Han, S.; Choe, W.-S.; Yoo, P. J. Adv. Mater. 2014, 26, 3899−3904. (30) Baker, R. W. Membrane Technology and Applications, 2nd ed.; Wiley: New York, 2004. (31) Wang, Q.; Samitsu, S.; Ichinose, I. Adv. Mater. 2011, 23, 2004− 2008. (32) Krieg, E.; Weissman, H.; Shirman, E.; Shimoni, E.; Rybtchinski, B. Nat. Nanotechnol. 2011, 6, 141−146. (33) Peng, X. S.; Jin, J.; Ichinose, I. Adv. Funct. Mater. 2007, 17, 1849− 1855. (34) Foo, C. W. P.; Bini, E.; Hensman, J.; Knight, D. P.; Lewis, R. V.; Kaplan, D. L. Appl. Phys. A: Mater. Sci. Process. 2006, 82, 223−233. (35) Tansil, N. C.; Li, Y.; Teng, C. P.; Zhang, S.; Win, K. Y.; Chen, X.; Liu, X. Y.; Han, M.-Y. Adv. Mater. 2011, 23, 1463−1466. (36) Mondia, J. P.; Amsden, J. J.; Lin, D. M.; Dal Negro, L.; Kaplan, D. L.; Omenetto, F. G. Adv. Mater. 2010, 22, 4596−4599. (37) Omenetto, F. G.; Kaplan, D. L. Science 2010, 329, 528−531. (38) Akthakul, A.; Salinaro, R. F.; Mayes, A. M. Macromolecules 2004, 37, 7663−7668.
F
DOI: 10.1021/acs.nanolett.6b01195 Nano Lett. XXXX, XXX, XXX−XXX
