Conical nanopores fabricated via a pressure-biased chemical etch

Abstract

Controlling the size and shape of nanopores in polymer membranes can significantly impact transport of molecular or ionic species through these membranes. Here we describe a facile method to controllably form conical nanopores in ion-tracked polycarbonate membranes. Commercial polycarbonate ion-tracked membranes were placed between a concentrated alkaline solution and an acidic solution. By varying the height of the acidic solution, the hydrostatic pressure was controlled, regulating the acid flux through the nanopores. The resulting asymmetric etching of the membrane produced conical pores with controllable aspect ratios. Scanning electron microscopy of both the pores and nickel nanostructures electrolessly templated in the pores confirms their conical shape. This safe, straightforward approach obviates the need to use large voltages, currents, and/or plasma etching equipment traditionally employed to create conical nanopores.

---

1 Introduction

Nanoporous membranes continue to emerge as important tools for the filtration and separation of nanoscale materials, processes which are strongly influenced by the size, shape, and surface chemistry of the nanopores. For example, if the pores are sufficiently small, the electrochemical double layer formed on the interior surfaces of these pores may be manipulated so as to control the transport of charged species through the membrane, facilitating charge-mediated filtration, sensing, or even energy harvesting.¹ By altering the shape of the nanopore, the transport properties of the pore may be further tailored. Conically shaped nanopores, for instance, have been shown to rectify ionic currents through nanoporous polymer membranes.^2–6 Given a sufficiently small pore tip, these conical membranes could act as charge filters, allowing for increased discrimination when filtering charged species, be it ions, particulates or biological molecules. The method described herein provides a simple way to fabricate conically shaped nanopores at densities that allow macroscale volumes to rapidly be filtered.

Controlling the size and shape of these pores with nanoscale resolution, however, is technically challenging. Plasma etching has been shown to create conically shaped pores; however, this technique requires expensive vacuum equipment, and can be difficult to adapt for large-scale applications.⁷ A common alternative method used to shape the commercially available nanopores into cones involves placing the membrane between a concentrated alkaline solution and an acidic solution, while applying as much as 30 V across two electrodes, one on each side of the membrane.^8–11 The alkaline solution etches the membrane, while the acidic solution neutralizes any etchant that diffuses through the membrane. The electric field established across the membrane creates an energetic barrier, which alters the local hydroxide ion concentration. The resulting concentration gradient produces asymmetric chemical etching of the polymer membrane, ultimately creating conical pores.⁹ Typically, these materials have been used for fundamental research studies examining the electrical behavior of single conical nanopores and nanoporous membranes with low pore densities.^7,12–14 From an electroanalytical perspective, pore density is important, because when the nanopore density is very small, the large solution resistance through the membrane dominates the response of the electrochemical cell, and the charging of the electrochemical double layer at the electrodes is sufficient to provide the relatively small requisite current. If the nanopore density is high, however, the solution resistance through the membrane becomes quite small. To apply the same potential difference across the membrane, needed to induce the etching asymmetry, a larger current is required. To supply this current Faradaic processes occur at the electrodes, which at more than 1.5 V, include the production of potentially hazardous hydrogen and oxygen gases. At 30 V, the production of hydrogen and oxygen gases and the subsequent change in solution pH make this voltage-based process ill-suited for the conical shaping of high-density nanopores.

Here, we describe an effective, tunable, inexpensive, and safe alternative to this process, replacing the use of an electric field with a modest gravity-induced hydrostatic pressure to create the etching asymmetry needed for conical pore formation. A similar process was used by Dobrev et al. to fabricate templates for copper electron emitters.¹⁵ In this straightforward system, an ion-tracked polycarbonate membrane is placed between an alkaline solution and an acidic solution. The height of the acidic solution is varied, thus controlling the hydrostatic pressure applied across the membrane, as depicted in Fig. 1. This pressure regulates the acid flux through the nanopores and produces the asymmetric etching needed to create conical nanopores. By varying both the duration of etching and the hydrostatic pressure applied across the membrane, it is possible to tune the morphology of conical nanopores produced without the need for large voltages or plasma etching processing.

Fig. 1 Schematic of the cell used for etching the nanoporous membranes.

2 Experimental

Nanoporous “Nuclepore” polycarbonate membranes with a 0.015 μm pore size were obtained from Whatman, (GE Healthcare, Fairfield, CT). KCl, NaOH, HCl (37%), PdCl₂, CH₂Cl₂ were all of reagent grade and obtained from Fisher Scientific (Fairlawn, NJ). SnCl₂, trifluoroacetic acid (98%), and methanol were of reagent grade and purchased from Sigma-Aldrich (St. Louis, MO). Deionized water was purified to 18.2 MΩ cm. Membranes were sputter-coated with gold–palladium and examined in a Zeiss Supra 55VP scanning electron microscope (SEM) at 2 kV, 3–6 mm working distance. No measurable difference in pore size was noted for a membrane coated in gold–palladium versus one uncoated. Pore diameters were measured across several square millimeters of membrane; values reported are averages of all pores taken in the micrographs (minimum of 50 pores). Uncertainties and error bars throughout this work represent one standard deviation. Pore density, as measured in SEM, was 6 × 10¹² pores m⁻².

The as-received nanoporous membranes were cut into circles 19 mm in diameter, rinsed in deionized water, and soaked in fresh deionized water for at least 15 minutes. A membrane was loaded into a glass NW10 flange in a U-shaped cell (Adams and Chittenden, Berkeley, CA). 30 mL of 9 M NaOH was placed on one side of the membrane, and the pressure head was controlled by varying the volume of acidic solution added to the opposing side of the membrane, as depicted in Fig. 1. The acidic solution consisted of 1 M aqueous formic acid and 1 M KCl. After etching for 0.5, 1, or 2 hours, the membranes were removed, immersed in a fresh acid solution for 30 minutes, flushed with deionized water, and dried in air.

To reveal the shape of the nanopores throughout the thickness of the membrane, replicas of the pores were formed by filling the pores with electrolessly deposited nickel. Here a procedure similar to that developed by Tai, et al. was employed.¹⁶ The membranes were first sonicated in deionized water for 10 minutes, then preactivated for 3 minutes in a solution of 0.07 M trifluoroacetic acid and 0.026 M SnCl₂ in 1 : 1, by volume, methanol–water. After rinsing in deionized water, the membranes were activated for 3 minutes in 0.01 M HCl with 1.5 mM PdCl₂. The membranes were again rinsed in water before immersion in the electroless nickel plating bath preheated to 80 °C (16 : 1 : 3 by volume, H₂O:Part A:Part B. Nickel plating solution from Caswell, Lyons, NY). Deposition occurred over 5 minutes. One side of the membrane was then mounted in epoxy (Extra fast setting, Royal Adhesives & Sealants, Wilmington, CA). The nickel film on the opposite side of the membrane was peeled off with tape. The membrane was dissolved in methylene chloride, leaving nickel replicas of the nanopores extending from the epoxy surface.

3 Results and discussion

In the following discussion, “pore tip” is defined as the pore opening which faced the acidic solution during etch, and the “pore base” faced the alkaline solution. To calculate the pressure across the membrane, the static fluid pressure applied by the alkaline solution is subtracted from that applied by the acidic solution. Static fluid pressure may be written as p = hρg, where p is the static fluid pressure, h is the total fluid height, ρ is the fluid density, and g is the acceleration of gravity. With this convention, positive pressure indicates acid flow into the alkaline solution. As the alkaline solution is denser than the acidic solution, equal heights of acidic and alkaline solutions create a slight negative pressure (−20 Pa), increasing the flux of alkaline solution through the pores and accelerating the etching process, particularly at the pore base.

SEM micrographs in Fig. 2 illustrate the effectiveness of the asymmetric etching process. Unetched, the membranes contain pores with “tips” and “bases” with identical sizes, 23 ± 4 nm in diameter (Fig. 2a and b). The cylindrical shape of the nickel replicas electrolessly templated in these pores reveals uniform cross sections throughout the pores (Fig. 2c). Upon etching the pores for 2 hours using equal heights of acidic and alkaline solutions (−20 Pa), the uniform etching leads to widening of the pore tip (Fig. 2d) and the pore base (Fig. 2e) to 281 ± 25 nm and 331 ± 20 nm, respectively, as expected. Interestingly, these figures also reveal that etching of more than a few hundred nanometers results in overlapping pores. As with the unetched pores, nickel replicas of these pores appear essentially cylindrical (Fig. 2f). In contrast, however, membranes etched at 300 Pa for 2 hours display a large disparity between pore tip and pore base sizes. Fig. 2g shows pore tips of 24 ± 3 nm, statistically unchanged from the unetched pore diameter of 23 ± 4 nm. As seen in Fig. 2h, on the other hand, the pore bases were widened to 340 ± 27 nm, effectively the same diameter as the −20 Pa case. The nickel replicas of these pores (Fig. 2i) are no longer cylindrical, but are steeply tapered conical structures. Across all samples, the dimensions of the nickel replicas agreed well with the tip and base diameters of the polymer membranes. Together, the difference in the sizes of the pore tips and the pore bases, along with the conical morphology of the nickel pore replicas are strong evidence of the asymmetric etching produced by the pressure differential between the acidic and alkaline solutions.

Fig. 2 SEM micrographs of pore tips (A, D, and G), pore bases (B, E, and H), and nickel replicas (C, F, and I) of the pores unetched (A–C) and after etching for 2 hours at −20 (D–F) or 300 Pa (G–I) pressure differential across the membrane. Unetched and pores etched with minimal pressure bias produce symmetrical, cylindrical pores. With 300 Pa pressure, severely tapered, conical pores are produced.

Moreover, by varying the pressure across the membrane (as a function of acid solution height), the aspect ratio of the pores was tuned. To illustrate this effect, the lateral etch rate (how quickly the pore diameters expanded) was plotted against the calculated pressure across the membrane. Irrespective of applied pressure, pore diameters of etched pores were determined to increase linearly with respect to the etch times tested (0.5–2 hours); the lateral etch rate was determined by varying the etch times for each pressure tested and subsequently measuring pore sizes in electron micrographs, as tabulated in Table 1. The resulting data are plotted in Fig. 3. As the pressure is elevated, the increased flow of etchant-neutralizing acid through the pores decreases the etch rate at the pore tip. If sufficient pressure is applied (≈300 Pa), etching of the pore tip may be completely suppressed. The etch rate at the base of the nanopore, however, is not strongly affected by the pressure applied across the membrane. This behavior may be explained by the fact that the flux of 1 M acid through the nanopores has little effect on the 9 M hydroxide concentration at the interface between the pore base and the alkaline solution. (The slight increase in etch rate with respect to pressure may be attributed to the mechanical stress state of the membrane.) It is this discrepancy in the effective etch rates on the acidic and alkaline sides of these pores that is responsible for the formation of the asymmetric, conical pores.

Table 1 Summary of pore measurements in this work. “Ratio” indicates the quotient of the base diameter divided by the tip diameter

Pressure/Pa	Etch time/h	Base dia./nm	Tip dia./nm	Ratio
−20	0.5	86 ± 13	72 ± 17	1.2
−20	1.0	208 ± 14	169 ± 15	1.2
−20	1.5	331 ± 20.	281 ± 25	1.2
60	0.5	70 ± 15	14 ± 3	5.0
60	1.0	178 ± 12	49 ± 5	3.7
60	1.5	301 ± 23	103 ± 8	2.9
140	0.5	42 ± 13	13 ± 2	4.0
140	1.0	114 ± 12	15 ± 2	7.6
140	1.5	323 ± 17	30 ± 10	11
300	0.5	56 ± 10	15 ± 2	3.7
300	1.0	102 ± 14	13 ± 2	7.8
300	1.5	340 ± 27	21 ± 3	16
460	0.5	30 ± 5	14 ± 3	2.1
460	1.0	93 ± 27	19 ± 2	4.9
460	1.5	298 ± 20	24 ± 3	12

---

Fig. 3 Effect of etch pressure on the lateral etch rate of the pore tip and pore base.

These data indicate that not only can this pressure-mediated etch process be used to create conical pores, but that it is possible to tune the overall size and aspect ratio of these conical pores by varying the pressure applied across the membrane and the duration of the etch.

4 Conclusions

A pressure-biased chemical etch was developed to create conical nanopores in commercially-available polycarbonate ion-tracked membranes. Hydrostatic pressure applied across the nanoporous membranes was shown to affect the etch rate of the pore tips, independent of the etch rate at the pore base, creating conical pores. By varying both the etching time and the pressure applied across the membrane, the size and aspect ratio of the nanopores was tuned. This method provides a straightforward, low-cost approach to create high density conical nanopores while avoiding the risks and potential difficulties associated with more conventional processes such as voltage-based etching. Continued development of this tunable process and the conical nanopores created is expected to enable advances in a wide-range of technologies involving selective ion-transport, particle separations, or other molecular and ionic filtration processes.

Acknowledgements

The authors thank Bonnie B. Mckenzie for her efforts acquiring SEM images. This work was supported by the Laboratory Directed Research and Development (LDRD) program at Sandia National Laboratories. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-AC04-94AL85000.

References

1. J. Cervera, B. Schiedt and P. Ramirez, Europhys. Lett., 2005, 71, 35–41.
2. J. Cervera, P. Ramirez, S. Mafe and P. Stroeve, Electrochim. Acta, 2011, 56, 4504–4511.
3. C. Kubeil and A. Bund, J. Phys. Chem. C, 2011, 115, 7866–7873.
4. P. Apel, Y. Korchev, Z. Siwy, R. Spohr and M. Yosida, Nucl. Instrum. Methods Phys. Res., Sect. B, 2001, 184, 337–346.
5. Z. Siwy, I. Kosinska, A. Fulinski and C. Martin, Phys. Rev. Lett., 2005, 94, 048102.
6. Z. Siwy, E. Heins, C. C. Harrell, P. Kohli and C. R. Martin, J. Am. Chem. Soc., 2005, 126, 10850–10851.
7. N. Li, S. Yu, C. C. Harrell and C. R. Martin, Anal. Chem., 2004, 76, 2025–2030.
8. P. Apel, Radiat. Meas., 2001, 34, 559–566.
9. C. C. Harrell, Z. Siwy and C. R. Martin, Small, 2006, 2, 194–198.
10. P. Scopece, L. A. Baker, P. Ugo and C. R. Martin, Nanotechnology, 2006, 17, 3951–3956.
11. J. Wharton, P. Jin, L. Seton, L. Home, S. Sherill, W. Mino and C. Martin, Small, 2007, 3, 1424–1430.
12. J. Cervera, B. Schiedt, R. Neumann, S. Mafe and P. Ramirez, J. Chem. Phys., 2006, 124, 104706.
13. P. Ramirez, S. Mafe, V. Aguilella and A. Alacarez, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2003, 68, 011910.
14. M. R. Powell, L. Cleary, M. Davenport, K. J. Shea and Z. S. Siwy, Nat. Nanotechnol., 2011, 6, 798–802.
15. D. Dobrev, J. Vetter, R. Neumann and N. Angert, J. Vac. Sci. Technol. B, 2001, 19, 1385–1387.
16. Y.-L. Tai and H. Teng, Chem. Mater., 2004, 16, 338–342.

---