Single-Molecule Sensing Using Nanopores in Two-Dimensional

5 days ago - We demonstrate a high-yield (60%), contamination-free, and alignment-free transfer method that involves their self-assembly at a liquidâ€...
0 downloads 0 Views 915KB Size
Subscriber access provided by ECU Libraries

Article

Single-Molecule Sensing Using Nanopores in TwoDimensional Transition Metal Carbide (MXene) Membranes Mehrnaz Mojtabavi, Armin VahidMohammadi, Wentao Liang, Majid Beidaghi, and Meni Wanunu ACS Nano, Just Accepted Manuscript • DOI: 10.1021/acsnano.8b08017 • Publication Date (Web): 07 Mar 2019 Downloaded from http://pubs.acs.org on March 8, 2019

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 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

Single-Molecule Sensing Using Nanopores in Two-Dimensional Transition Metal Carbide (MXene) Membranes Mehrnaz Mojtabavi Wanunu 4,5,*

1,#,

Armin VahidMohammadi

2,#,

Wentao Liang 3, Majid Beidaghi

2,*,

Meni

Department of Bioengineering, Northeastern University, Boston, MA 02115, USA. Department of Materials Engineering, Auburn University, Auburn, AL 36849, USA. 3 Kostas Advanced Nano-Characterization Facility, Northeastern University, Burlington Campus, 141 South Bedford Street, Burlington, MA 01803 4 Department of Physics, Northeastern University, Boston, MA 02115, USA. 5 Department of Chemistry and Chemical Biology, Northeastern University, Boston, MA 02115, USA # These authors contributed equally to the work. * Corresponding authors: [email protected]; [email protected] 1 2

ABSTRACT Label-free nanopore technology for sequencing biopolymers such as DNA and RNA could potentially replace existing methods if improvements in cost, speed, and accuracy are achieved. Solid-state nanopores have been developed over the past two decades as physically and chemically versatile sensors that mimic biological channels, through which transport and sequencing of biomolecules has already been demonstrated. Of particular interest is the use of two-dimensional (2D) materials as nanopore substrates, since these can in theory provide the highest resolution readout (60%) yield of transfer using this method, defined as the ratio of SiNx chips with holes completely covered with MXene flakes to the overall number of transfers attempted. We also found that increasing concentration of flakes in the initial dispersion increases the full-coverage yield; however, it also decreases the probability of obtaining single-layer membranes. In Figure 2b and Supplementary Movie 1, for illustration and technique demonstration, we used dispersion with relatively higher concentration (0.05 mg/ml). Figure S4 shows snapshots of 5-fold diluted MXene dispersion (~0.02 mg/ml) cast on chloroform during flakes self-assembly (top panel). To better represent the flakes, we have shown a higher contrast version of each image (bottom panel). In general, to obtain single layer MXene membranes, a 10-20 x dilution from the initial dispersion is needed. Figure S5 shows UV-VIS absorbance spectra of an as-synthesized Ti3C2Tx dispersion, a 3-fold diluted dispersion, and a 10-fold diluted dispersion. Inset shows photographs of all three dispersions. Apart from the high yield, this transfer method has several advantages over other fabrication methods used previously. First, yields obtained by directly drop-casting from the MXene dispersion, our initial method of deposition, were very poor and resulted in partial coverage, as well as the formation of multilayer films by piling up of sheets into the hole (see Figure S6), as also recently reported by Lipatov et al.82 In contrast to drop casting, we did not observe this effect for our transfer method, instead obtaining flat sheets (Figure S2m-o). Moreover, the lack of a need of a sacrificial polymer for transfer eliminates other materials from 6 ACS Paragon Plus Environment

Page 7 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

contaminating the resulting membrane, which is a plaguing problem in handling MXene82 and other 2D materials.42,44,48,51,87-89 As mentioned earlier, freestanding MXene membranes fabricated with this method could have different thicknesses depending on the initial flake concentration in the dispersion. Different techniques could be used to find the thickness of the membrane. Figure 2d shows an AFM image of Ti2CTx flakes transferred using the liquid-liquid interface method, and Figure 2e shows line height profiles from the image, where a thickness of about 2.4 nm for one layer of Ti2CTx was measured. To find the spacing between layers in the MXene sheets, we transferred flakes onto holey carbon filmed TEM grids using the same assembly method, and characterized the film using atomic-resolution TEM. Areas where folding was observed (Figure S7) were used to estimate the overall thickness. Figures 2g, 2h and Figure S8 are representative high-angle annular dark-field scanning transmission electron microscopy (HAADF STEM) images of Ti3C2Tx that show multilayer flake structure in a folded area. Based on the observed structure, we derive an empirical formula for the overall thickness as a function of the number of layers 𝑇(𝑁) = 1.0 + 1.8(𝑁 ― 1) nm, where N represents the number of layers and T the end-toend thickness of the layer. To confirm their composition, Figure S9 shows energy-dispersive Xray spectroscopy (EDS) elemental mapping of one of the folded regions. Unlike AFM imaging, which produces thickness artifacts from water presence, contrast measurements from TEM imaging is a robust method for thickness estimation, although a reduced stability of Ti2CTx flakes to electron beam exposure requires careful imaging to avoid hole formation. To investigate the performance of MXene freestanding membranes, we used a focused electron beam to fabricate nanopores with diameters ranging from 3-10 nm in both Ti2CTx and Ti3C2Tx membranes. Following nanopore fabrication, we assembled the devices in a fluidic cell that hosts two electrolyte chambers, such that the nanopore is the sole liquid junction between the chambers. Figure 3a shows schematic of the MXene nanopore setup, in which two Ag/AgCl electrodes are each in contact with an electrolyte chamber, and ion current through the pore that results from voltage application is measured using an electrometer. Figure 3b shows a HAADFSTEM image of a 4.2 nm diameter Ti3C2Tx nanopore that was fabricated using a STEM point probe. The image clearly shows the pore in black, a 1-layer region that defines the 4.2 nm pore, as well as a larger (6-7 nm) ring that outlines a region that was thinned during the drilling process which consists of a multilayer MXene film (3-4 layers). Figure 3c shows I-V curves for a Ti2CTx 7 ACS Paragon Plus Environment

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 26

nanopore with a measured cross-sectional area of 36.4 nm2 (red markers), as well as a Ti3C2Tx nanopore with a cross-sectional area of 19.2 nm2 (blue markers). Insets show bright-field TEM images of the pores, with the pores outlined in their respective colors. The triangular pore shape for the Ti3C2Tx pore is likely to be caused by the shape of the slightly out-of-focus electron beam. Other example TEM images and I-V curves are shown in Figure S10. Figure 3d shows currentvs.-time traces for both pores at 100 mV and 200 mV voltages, low-pass-filtered to 100 kHz. Corresponding power spectral density plots for the traces at 100 mV are shown in Figure 3e, and for 200 mV in Figure S11. For comparison, a typical power spectral density plot of a 5.7 nm diameter SiNx pore has been added to each figure (shown in grey). Noise characteristics for nanopores at low-frequency and high-frequency regimes has been studied before, shown to have 1 𝑓 noise in the low-frequency regime, whose amplitude is voltage dependent.90 In all transferred 2D nanopores studied so far, graphene,54 graphene-Al2O3,91 graphene-TiO2,41 MoS2,29 and BN,49 the dominant source of noise is in the low-frequency regime, known as 1 𝑓 noise, which owes to a number of factors that include mechanical fluctuation, surface charge variation, and pore contamination. Our results for MXene pores suggest a similarity in terms of noise characteristics to other nanopores in transferred 2D membranes, although the noise performance still lags behind transfer-free 2D membrane materials such as graphene55 and MoS2.45 The ion-conductance (𝐺 = 𝐼 𝑉) through a pore is dependent on several pore characteristics such as its diameter, thickness, and surface charge, as described by Equation 1.92 𝜋𝑑2𝑝𝑜𝑟𝑒

4

𝐺 = 4 𝐿𝑝𝑜𝑟𝑒 [(𝑖 + + 𝑖 ― )𝑛𝑒 + 𝑖 + 𝑑𝑝𝑜𝑟𝑒]

(1)

where, 𝑑𝑝𝑜𝑟𝑒 is the pore diameter, 𝐿𝑝𝑜𝑟𝑒 is the pore thickness, 𝑖 + and 𝑖 ― are electrophoretic mobilities of the cations and anions, respectively, n is the number density of anions or cations, e is the elementary charge, and  is the pore’s surface charge density. In the bracket terms of Equation 1, the left term is the conductance through the pore as a result of ion flow in the bulk, and is linearly dependent on the bulk salt concentration, whereas the right term shows the conductance through the pore as a result of electro-osmotic flow of counterions. Surface charge density is described by Grahame equation92 shown in Equation 2:

(𝜁) =

2𝜖𝜖𝑜𝐾𝐵𝑇𝜅 𝑒

𝑒𝜉

sinh (2𝐾𝐵𝑇)

8 ACS Paragon Plus Environment

(2)

Page 9 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

where 𝜁 is the zeta potential, 𝜀𝑜 is the absolute permittivity, 𝜀 is the relative solution permittivity, 𝐾𝐵 is Boltzman’s constant, T is the temperature, 𝜅 ―1 is the Debye length, and e is the elementary charge. We measured average zeta potentials of -40 mV for Ti3C2Tx at pH 5.5, -24 mV for Ti2CTx at pH 7.5, and -38 mV at pH 10 using Malvern Zetasizer NanoZSP (Malvern, UK). This corresponds to surface charge densities of -13.2 mCm-2 and -8.7 mCm-2 for Ti3C2Tx at pH 5.5 and Ti2CTx at pH 7.5 in 400 mM KCl, respectively. In order for electroosmotic flow to dominate the pore conductance, the condition n >> 2𝜎 𝑑𝑝𝑜𝑟𝑒𝑒 must be satisfied.92 For the smallest pore used here (dpore = 3 nm), the term 2𝜎 𝑑𝑝𝑜𝑟𝑒𝑒 is significantly smaller than n since our minimum experimental salt concentration was 400 mM. Therefore, we neglect the right term in Equation 1, and reason that conductance is governed by the bulk term of the equation. Aside from the dependence of conductance on pore and buffer characteristics, another factor that strongly influences conductance through the nanopore in the case of very thin membranes is access resistance,93 shown by Equation 3: 𝑅𝑎𝑐𝑐𝑒𝑠𝑠 = (𝑖 + +

1

(3)

𝑖 ― )𝑛𝑒𝑑𝑝𝑜𝑟𝑒

Taking this into account, the total ion-conductance through the pore is shown by Equation 4: 𝐺𝑡𝑜𝑡𝑎𝑙 =

{

1 2 𝜋𝑑𝑝𝑜𝑟𝑒 4 𝐿𝑝𝑜𝑟𝑒

(𝑖 + + 𝑖 ― )𝑛𝑒

+ ( 𝑖 + +

1

}

𝑖 ― )𝑛𝑒𝑑𝑝𝑜𝑟𝑒

―1

(4)

We have used Equation 4 to calculate effective pore thickness values based on measured conductance for 33 pores with TEM-measured effective diameters (approximated as discs with diameter d based on the TEM-measured pore areas), as shown in Figure 3f. Overlaid with the data points are contour lines that represent Equation 4 for different labeled pore thickness values in the range of 1-24 nm. Each dashed line shows access resistance contribution (Equation 3) to the total conductance through the pore for different pore dimeters at different buffer conditions (shown by different colors). In general, we find that most of the pores we tested fall within the range of 1-10 layers. For atomically thin pores, access resistance dominates total conductance through the pore, resulting in pore diameter being the only geometrical factor that affects the conductance through the pore. We predict that the data points in the vicinity of the access resistance lines belongs to the pores with monolayer-thick membranes. A few of the pores fell above the access 9 ACS Paragon Plus Environment

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

resistance lines, suggesting pore expansion between the fabrication and the measurement steps or formation of small holes during electron beam exposure to monolayer membrane as shown in Figure S12 (see experimental section for more information on pore drilling and imaging conditions). In general, by using this model, about 45% of MXene membranes shown in Figure 3f are single-layere. We note that in the absence of a fabricated pore, we obtained 10) comparable to the one reported for MoS2. Furthermore, we were able to decrease the translocation speed of dsDNA through pore considerably using 2 M LiCl. On the other hand, success rate of device fabrication with biomolecule sensing capability is still a challenge. This could be overcome by tailoring surface chemistry of MXenes to less negatively charged surface groups. Moreover, formation of several pores during electron beam exposure could be inhibited by using STEM probe which also yields a pore with more precise size and geometry.

11 ACS Paragon Plus Environment

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 26

Biomolecule Detection Type of 2D Material

Materials Synthesis Method

Nanopore Membrane Fabrication

Surface Properties

Graphene

CVD/ Exfoliation

Direct CVD/ Transfer

Hydrophobic

CVD/ Exfoliation

Direct CVD/ Transfer

SemiHydrophobic

MoS2

Applied Voltage Window

Success Rate of Device for Detection

Type

d/L† (nm)

Buffer

V (mV)

Translocation Speed (bp(nt)/µs)

Capture Rate

SNR

(s-1nM-1)

(@ 10 kHz)

[-500,

~ 10 %

dsDNA

8/2

33*

1

>2*

41

Not reported

dsDNA

22/0.3

1M KCl 1M KCl

100

500]93

200

18*

Not reported

>2*

42

>70 %

dsDNA

20/0.7

2M KCl

200

50

Not reported

>10

44

ssDNA

2.3/1.6

0.4 M KCl

200

10*

0.95

Not reported

45

dsDNA

4.4/0.7

3M KCl

400

25*

Not reported

>4*

51

dsDNA

~6/1.1

3M KCl

160

62*

Not reported

>2*

48

dsDNA

4.5/8

2M LiCl

500

0.5

0.042

>10

[-400, 400]44

Not reported

Not

Not

reported

reported

Hydrophobic

Not reported

reported

Hydrophilic

[-900, 900]

~ 10 %

WS2

CVD/ Exfoliation

Transfer

SemiHydrophobic

h-BN

CVD/ Exfoliation

Transfer

Ti2C

Exfoliation

Transfer

Not

Table 1. Comparison between 2D materials studied so far as nanopore membrane * values calculated using the data provided in the original manuscript. †

d=diameter, L= pore length

CONCLUSIONS We have carried out here a series of ion and DNA transport studies through MXene nanopores, produced by transferring sheets of TBAOH-exfoliated Ti2CTx MXenes, as well as Ti3C2Tx, onto hole-containing SiNx devices. To facilitate the MXene transfer process, we have developed a method for flake self-assembly at the liquid-liquid interface to mm-sized arrays of sheets, which results in high-yield (60%) and contamination-free transfer of flakes such that they completely seal the support hole. Key to this self-assembly process at the chloroform-water interface is the introduction of methanol to the aqueous phase, which rapidly drives the sheets to the inter-solution interface. Characterization of nanopores produced in these suspended MXene membranes show thickness values that range from 1-15 MXene layers, the thickness of which can be controlled by optimizing the assembly process. Furthermore, the MXene membranes exhibit

12 ACS Paragon Plus Environment

Ref

This study

Page 13 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

low ion current leakage, and nanopores in the MXene membranes show noise characteristics that are comparable to other two-dimensional membranes. Finally, we have demonstrated MXene nanopore-based sensing of DNA molecules in sub-10 nm pores, which show that DNA can transit through the Ti2CTx pores. Further refinement of this class of 2D materials, which are hydrophilic in nature and electronically conductive, can lead to distinct applications in biosensing and alternative sensing modalities such as electrically gating the pores and sensing DNA using transverse readout, the subject of future studies.

MATERIALS AND METHODS MXene Synthesis. Large flakes of Ti3C2Tx and Ti2CTx MXenes were synthesized according to minimally intensive layer delamination (MILD) method reported in literature.54 First the etching solution was prepared by adding 1 g of LiF powder (98.5%, Alfa Aesar) to 20 mL 6M HCl solution (ACS grade, BDH). The mixture was then stirred for 15 minutes to completely dissolve the LiF powder in the solution. Then, 1 g of MAX phase powder (Ti3AlC2 or Ti2AlC, synthesized according to previous work69,74) was slowly added to the etching solution. The addition of MAX phases should be very slow, and an ice bath should be used to avoid excessive heat generation due to exothermic nature of the reaction. The etching process was carried out for 24 h at 35 °C while stirring the solution at 550 rpm using a Teflon-coated magnetic bar. After the etching process was complete, the solution was divided into 4 different centrifuge vials and DI water was added (45 mL) to dilute them. The solutions were then centrifuged at 3500 rpm for 3 minutes and the supernatant was poured out. To obtain delaminated Ti3C2Tx dispersions, the washing process was continued by adding DI water, manual shaking of the solutions for 2 minutes and then centrifuging them at 3500 rpm for 3 minutes until a dark green supernatant was observed (pH >4.5). At this stage, the initial delaminated solution was poured out and DI water was added to the sediments. The solutions were shaken for another 2 minutes and this time centrifuged at 3500 rpm for 1h to collect the large flake size MXene solutions (pH ~ 5). As for Ti2CTx, the etching and washing process by manual shaking did not result in proper delamination. Therefore, to delaminate Ti2CTx, the synthesized multilayered powders after etching (pH of supernatant ~ 4.5) were collected and intercalated with tetra n-butyl ammonium hydroxide (TBAOH, 40 % w/w aqueous solution in water, Alfa Aesar) by adding 200 mg of Ti2CTx powder to 4 mL of TBAOH 13 ACS Paragon Plus Environment

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

solution and stirring for 2h at room temperature. Then, the obtained solution was diluted with DI water and hand shaken for 2 minutes, followed by centrifugation at 2000 rpm for 10 minutes. Again, the first supernatant (slightly dark red color) was poured out as it contained residual TBAOH. Then, DI water was added to the sediments, the solutions were shaken for 2 minutes followed by centrifugation at 2,000 rpm for 30 minutes. The obtained dark red supernatant was collected and used for device fabrication. MXene Transfer. To fabricate free-standing MXene membranes, MXenes flakes dispersed in water were mixed with equal volume of methanol (to obtain MXenes dispersion in 50:50 water: methanol). Then, small amount of dispersion was added to chloroform and flakes were self-assembled on the drop/chloroform interface. 5 × 5 mm2 chip with 50 nm thick SiNx membrane at the center, previously cleaned with hot piranha (H2SO4:H2O2 2:1) and deionized water) was dipped into the chloroform and lifted up through the drop. Afterward, chips were left on hotplate at 80 ºC for about five minutes to dry (the process is shown in supplementary video). For TEM imaging, the same transfer method was used but instead of using SiNx chips, a holey carbon film TEM grid was used. Device and Nanopore Fabrication. 5 × 5 mm2 chips with a 50 nm thick SiNx membrane at the center and a pre-fabricated 100 nm or 200 nm hole made with focused ion beam (FIB) or electron beam lithography (EBL) were used as substrates for device fabrication. The chips were cleaned prior to MXene transfer with hot piranha (H2SO4:H2O2 with a 2:1 ratio) and deionized water. After MXene transfer onto the chips, nanopores were fabricated on freestanding part of MXene flakes using JEOL 2010FEG at 1.5 Mx magnification and spot sizes in the range of 3-5. Pores were formed within 1-5 seconds (for