Herein, we present the fabrication of multiplexed single-walled carbon nanotube (SWCNT) devices, where selected chiralities were separately immobilized on one chip with single-tube precision. Each chirality was subsequently electrically measured individually. Specifically, (6,5) and (7,5) SWCNT species were isolated via aqueous two-phase polymer systems, after which dielectrophoresis was used to precisely control the placement of each chirality, along with a metallic species, separately on prepatterned electrodes on a single chip.
I. INTRODUCTION
The use of single-walled carbon nanotubes (SWCNTs) as components in next generation nanoelectronic devices has been a major focus in the nanomaterials field for the past few decades, owing to their outstanding electrical properties.1–6 In particular, SWCNT-based field effect transistors (FETs) can be used for many applications, and a variety of FETs, ranging from dense arrays to individual SWCNT devices, have been fabricated to investigate the electrical properties of SWCNTs.7–15
SWCNTs can be metallic or semiconducting based on their structure. These discrete structures can be defined by the chirality index (n,m) that corresponds to the direction and magnitude of the rolling vector along the hexagonal carbon lattice.16 The chirality of a SWCNT has direct implications on its electrical and optical properties; moreover, the bandgap of semiconducting SWCNTs is inversely proportional to the tube diameter.17–20 Consequently, it is desirable to control the chirality and hence electronic properties of SWCNTs to tailor their use for specific applications.21–23
Various techniques have been proposed to produce SWCNTs with highly enriched single-chirality species, through either direct controlled growth24–32 or by separation post-synthesis.33–38 Although SWCNTs with defined chiralities can be achieved via chemical vapor deposition,18,25,27,39 different in-solution strategies have been developed for chirality enrichment such as chromotography,40–42 density differentiation,43,44 DNA recognition,45,46 and polymer systems.47–51
In particular, approaches based on polymer aqueous two-phase (ATP) systems have been demonstrated to be efficient for SWCNT separation.48,52–54 ATP systems work on the principle that each chirality of SWCNT will vary in solvation energy due to differences in surfactant or DNA interactions with the sidewall. Consequently, SWCNTs are unevenly distributed in the two phases, resulting in spontaneous separation of SWCNTs.48,53 By tuning the surfactants or DNA, and using modulating agents to push SWCNTs between phases, ATP systems allow the simultaneous separation of many individual species in a single experiment.55 This strategy is advantageous as it shows potential for large scale separation of SWCNTs due to its ease of processing.48,53
In this regard, while early electronic devices based on SWCNTs suffered from poor efficiency due to metallic SWCNT impurities,56 thin-film electronic devices made from SWCNTs of defined chirality are desirable for uniformity and reproducibility.23 Additionally, the optical properties of semiconducting SWCNTs are of great interest for optoelectronic devices with defined absorbances based on their chirality.
Different strategies have been pursued toward achieving single-chirality SWCNT field effect transistors (FETs),57,58 optical emitters,59 and optical detectors.60 However, a general strategy for the solution-processable fabrication of single-chirality multiplexed CNT devices has yet to be demonstrated, and few studies report controlling the placement of individual SWCNTs of chosen chirality.61–64
In this work, we present a solution-processable strategy for the fabrication of multiplexed single-chirality CNT-FETs. This was achieved by immobilizing individual SWCNTs of selected chirality on prepatterned electrodes via dielectrophoresis (DEP), namely (6,5), (7,5), and metallic SWCNTs. These were organized in device configuration on the same chip, where each distinct chirality was electrically addressed in isolation.
II. MATERIALS AND METHODS
A. Materials
Oligonucleotides (ssDNA) were purchased from Integrated DNA Technologies. Metallic SWCNTs were purchased from NanoIntegris. Dextran (DX), polyethylene glycol (PEG), and poly-vinylpyrrolidone (PVP) were purchased from Alfa Aesar. Au/Cr nanoelectrodes on silicon wafers, patterned with via electron beam lithography, were purchased from Con-Science. HiPco SWCNTs, CoMoCAT SWCNT, (6,5) enriched SWCNTs, and other chemicals used in this paper were purchased from Sigma-Aldrich.
B. SWCNT sorting
Highly enriched (6,5) SWCNTs were obtained by adapting the procedure of Lyu et al.53 Briefly, CoMoCAT SWCNTs (1 mg) were dispersed in MilliQ water (1 ml) by sonicating in the presence of ssDNA (2.5 mg/ml; TCTCCCTCTCCCTCT) with NaCl (30 mM) for 90 min. Following sonication, the SWCNT solutions were centrifuged at 13 000 rpm for 90 min (Heraeus Biofuge Pico), and the supernatant was extracted to isolate dispersed SWCNTs.
Aqueous solutions of DX (250 kDa; 20% w/w) and PEG (1.5 kDa; 60% w/w) were made. These form the basis of the aqueous two-phase (ATP) system.
In a typical purification, the 7:3 solution (containing 12.2% PEG and 14.4% DX) was vortexed and 140 μl was extracted immediately and added to a microcentrifuge tube. 45 μl of SWCNT solution was added to this, after which the mixture was vortexed for 20 s and centrifuged at 13 000 rpm for 2 min. The top phase was carefully removed by a pipet and stored in a clean microcentrifuge tube, and the same volume of top phase from the 10:0 solution (containing 9.23% PEG and 10.9% DX) was added. 0.3 μl of an aqueous PVP (10 kDa; 10% w/v) solution was added to the replenished SWCNT mixture and vortexed and centrifuged again. This was repeated for ∼15 separations until almost all the SWCNTs had moved into the top phase. Fractions were analyzed by UV-vis (300 nm–1400 nm; Shimadzu UV3600 Plus) and could be further purified using sodium thiocyanate (NaSCN) with blank bottom phase from the 10:0 stock.
Highly enriched (7,5) tubes were obtained via a similar method. Briefly, the SWCNT mixture was initially wrapped with DNA sequence (ATT)4. Subsequently, the DNA-wrapped SWCNT mixture was subsequently added to a [PEG + poly(ethylene glycol) diamine (PEG-DA)]/DX system. After vortex and centrifuge, a DX-rich bottom phase, in which (7,5) species was enriched, was collected as the starting materials for further purification. By repeating the purification process, a solution of highly enriched (7,5) species was obtained.
To remove DX and PEG from the purified SWCNTs, concentrated NaSCN was added to the fractions and the tubes were centrifuged for 10 min at 13000 rpm. The SWCNTs formed a pellet while the supernatant was removed and discarded. The pellet was washed very gently with water, and the pellet was redispersed by adding water. If the pellet could not be dispersed in water, it was sonicated again in the original DNA solution. Finally, SWCNT solutions were desalted using dialysis cartridges with a 20 kDa MWCO membrane (Millipore).
C. Fabrication of SWCNT devices
To immobilize an individual (6,5) species, the AC voltage of the generator was switched onto typically Vp-p = 2 V at f = 400 KHz. The as-prepared solution was diluted (ca. 100 ng/ml) and drop-cast onto the prepatterned electrode pairs with a pipet. After a duration of 15 s, the samples were rinsed with water and blow-dried gently with nitrogen gas.
Similarly, to immobilize an individual (7,5) species, the AC voltage of the generator was switched onto typically Vp-p = 2 V at f = 400 KHz. The as-prepared solution was diluted (ca. 100 ng/ml) and drop-cast onto the prepatterned electrode pairs with a pipet. After a duration of 10 s, the samples were rinsed with water and blow-dried gently with nitrogen gas.
To immobilize a metallic SWCNT, the AC voltage of the generator was switched onto typically Vp-p = 2.5 V at f = 5 MHz. The as-prepared solution was diluted (ca. 100 ng/ml) and drop-cast onto the prepatterned electrode pairs with a pipet. After a duration of 10 s, the samples were rinsed with water and blow-dried gently with nitrogen gas.
D. Characterization
1. AFM imaging
The SWCNT solutions were deposited onto freshly cleaved mica, which was pre-treated with 1M MgSO4 solution to enhance DNA adsorption, rinsed with water, and dried with nitrogen gas. Topographic analysis of the electrodes and the SWCNT samples was performed with a Bruker Dimension Icon atomic force microscope (AFM) with ScanAsyst Air tips in Peak Force Quantitative Nanomechanical Mapping (QNM) mode.
2. Electrical measurements
Before electrical measurements, annealing treatment was performed in a tube furnace under nitrogen gas to minimize the influence of DNA. The temperature was set to be 200 °C with an increase rate of 10 °C/min. After a duration of 2 h, the samples were cooled down to room temperature for further investigation.
Electrical measurements were performed using a probe station (PS-100, Lakeshore) equipped with a semiconducting parameter analyzer (Keithley, 4200SCS) at room temperature. To measure the current through CNT devices as a function of voltage (I–V curves), sweeping bias (−1 V to 1 V) was applied between the source and drain electrodes. To measure the transfer characteristic of CNT devices (Isd vs Vg), a constant voltage was applied between source and drain electrodes (1 V) while a sweeping bias was applied to the backgate electrode (−15 V to 15 V).
III. RESULTS AND DISCUSSION
The separation of DNA-wrapped SWCNTs by chirality was achieved using aqueous two-phase (ATP) systems, following previously published strategies.53,65 Figures 1(a) and 1(b) show the absorption spectra of the separated (6,5) and (7,5) species, respectively (also see Fig. S1 in the supplementary material). The absorption spectrum of the separated (6,5) species shows the E11 transition at 990 nm and the E22 transition at 570 nm, demonstrating the high enrichment of (6,5) tubes. The (7,5) species was separated and purified via a similar process. In the absorption spectrum of the separated (7,5) species the E11 transition peaks at 1050 nm and the E22 transition at 650 nm demonstrate the high purity of the (7,5) species.48 Atomic force microscopy (AFM) was used to characterize the morphology of the separated (6,5) tubes [Fig. 1(c)]. The average diameter of these SWCNTs was found to be 2.1 ± 0.3 nm, larger than the expected 0.76 nm diameter due to the presence of DNA66 (see Fig. S2 in the supplementary material).
Subsequently, DEP was performed to immobilize individual SWCNTs between electrodes fabricated via electron beam lithography. Individual SWCNTs were immobilized between electrodes pairs with DEP via careful optimization of voltage, time, and concentration during DEP;67 we achieved a yield of immobilization of individual SWCNTs of ca. 40%. Additionally, to ensure that SWCNTs can only be immobilized between selected electrodes pairs, AC bias was applied to a selected source electrode while the common (drain) electrode was grounded. Hence, different species with a defined chirality could be immobilized between separate electrode pairs on a single chip:68 in this case, individual (6,5), (7,5), and metallic SWCNTs were immobilized on a single chip, forming multiplexed single-chirality devices [Fig. 2(a)]. The overall yield of multiplexed single-chirality devices was found to be 30%.
Figure 2(b) shows representative AFM images of individual SWCNTs of different chiralities immobilized between electrodes on a single chip, where the channel length of the devices is 300 nm. The diameter of the SWCNTs immobilized between electrodes is measured to be 2.5 ± 0.7 nm, in good agreement with the measured diameter of the separated SWCNTs (Fig. S3 in the supplementary material).
The conductance of these SWCNT devices could not be measured directly after immobilization, primarily due to the influence of DNA. This was addressed by annealing the devices at 200 °C for 2 h under nitrogen gas, minimizing the presence of DNA between the SWCNTs and the metallic electrodes and improving electrical contact between the SWCNTs and gold electrodes.8,62
Figure 3 shows I–V curves through these individual SWCNT devices in which various gate voltages were applied. The devices containing individual (6,5) and (7,5) show the typical gate voltage dependence of semiconducting devices [Figs. 3(a) and 3(b)]. The current–voltage characteristics of metallic SWCNT devices are shown in Fig. 3(c); as expected, changing the gate voltage cannot induce changes in the current in these devices made from metallic CNTs.
To further confirm the electrical properties of these devices, we measured their transfer characteristics, shown in Figs. 4(a) and 4(b) for (6,5) and (7,5) tubes, respectively. These devices show p-type transport characteristics; the Ion/Ioff ratio of was found to be more than 103, indicating good FET behavior.7,61
Figure 4(c) shows the current through the devices made with metallic SWCNTs as a function of gate voltage. As expected, the current is independent on the gate voltage applied, demonstrating the metallic nature of these devices. However, the average contact resistance is in the order of magnitude of 10 MΩ, which is higher than previously reported results.7,8 This can be attributed to the remaining residues from the degraded DNA.
IV. CONCLUSIONS
In conclusion, we demonstrated the immobilization of individual (6,5), (7,5), and metallic SWCNTs on a same chip, forming multiplexed single-chirality SWCNT devices. Electrical measurements show how the (6,5) and (7,5) devices exhibit the expected FET behavior, while metallic tube devices exhibit metallic behavior, i.e., no gate dependence.
The ability to control the assembly from solution of single-chirality SWCNTs in device configuration is of general applicability for optoelectronic applications, from FETs to integrated circuits, as well as optical emitters and detectors. Moreover, the multiplexed devices described here may be beneficial in the development of multiplexing sensing platforms, where simultaneous measurements can be performed to select optimal chirality for tailored applications.
SUPPLEMENTARY MATERIAL
See the supplementary material for the histogram of diameter of the separated (6,5) SWCNTs and the histogram of diameter of SWCNT devices.
ACKNOWLEDGMENTS
We gratefully acknowledge financial support from the Air Force Office of Scientific Research under Award No. FA9550-16-1-0345. M.F. would like to thank Queen Mary University of London for the Impact Fund, and X.X. also gratefully thanks the China Scholarship Council for financial support.
DATA AVAILABILITY
The data that support the findings of this study are available within the article and its supplementary material.