Graphene nanopores

DNA sequencing devices consisting of a nanopore within a graphene nanoribbon (GNR) will be discussed first. These devices function by detecting changes in the graphene current as DNA translocates through the pore in a direction perpendicular to the graphene plane [Fig. 1(b)]. These devices should not be confused with ionic current-based graphene nanopore sequencing devices that have been extensively described elsewhere.^[Reference Heerema and Dekker^19–Reference Garaj, Hubbard, Reina, Kong, Branton and Golovchenko^22] A schematic of the typical geometry for these devices is shown in Fig. 2(a). In particular, these devices typically consist of a SiN_x window that is suspended over a SiO₂ layer on a silicon substrate. Graphene and metal contact electrodes are patterned on the SiN_x with a passivation layer covering the metal electrodes. A nanopore in the graphene and underlying SiN_x window enables DNA translocation. These devices have been studied in many computational works and indeed nucleobase-specific changes in the graphene current have been predicted.^[Reference Nelson, Zhang and Prezhdo^24, Reference Saha, Drndić and Nikolić^33–Reference Girdhar, Sathe, Schulten and Leburton^36] Furthermore, several experimental studies have detected DNA translocation through a graphene nanopore via changes in the graphene current.^[Reference Traversi, Raillon, Benameur, Liu, Khlybov, Tosun, Krasnozhon, Kis and Radenovic^25, Reference Heerema, Vicarelli, Pud, Schouten, Zandbergen and Dekker^30, Reference Puster, Balan, Rodrguez-Manzo, Danda, Ahn, Parkin and Drndić^37] However, fabrication throughput and device yield, as well as experimental challenges such as graphene wetting and electrostatic discharge continue to hinder the development of these devices towards DNA sequencing.^[Reference Heerema, Vicarelli, Pud, Schouten, Zandbergen and Dekker^30]

Figure 2.

(a) Schematic representation of graphene nanopore electronic DNA sequencing devices. (b) Schematic of the steps involved in forming SiN_x windows. (c) Schematic of nanopore creation via (i) TEM drilling, (ii) electron beam lithography followed by reactive ion etching and (iii) delectric breakdown. (d) TEM images of a nanopore within a GNR created via TEM drilling.^[Reference Traversi, Raillon, Benameur, Liu, Khlybov, Tosun, Krasnozhon, Kis and Radenovic^25, Reference Puster, Rodrguez-Manzo, Balan and Drndić^38] Figures adapted with permissions from (d) Nature publishing group^[Reference Traversi, Raillon, Benameur, Liu, Khlybov, Tosun, Krasnozhon, Kis and Radenovic^25] and American Chemical Society.^[Reference Puster, Rodrguez-Manzo, Balan and Drndić^38]

This section will begin by discussing the fabrication of suspended SiN_x structures. These structures are necessary to enable nanopore formation and are also relevant for the graphene nanogap detectors discussed in the section “Graphene nanogaps.” The procedure for fabricating SiN_x windows has been described elsewhere^[Reference Grant, Hu and Kasemo^39] and typically consists of growing a SiO₂ layer on a high resistivity silicon <100> substrate then depositing low-pressure chemical vapor deposition (LPCVD) SiN_x films, followed by photolithography and subsequent etching [Fig. 2(b)]. Care should be taken to ensure etching along the correct crystal orientations. A protective layer made of an easily removed polymer may be used during etching to protect features on the front side of the wafer from leakages caused by defects when etching. The SiO₂ film is used to reduce capacitive coupling between the metal electrodes and silicon substrate and is generally between 0.1 and 5 µm thick. One result of this thick SiO₂ layer is that, due to its compressive stress, the resulting SiN_x windows can have an uneven surface topography.^[Reference Healy, Ray, Willis, Peterman, Bartel and Drndić^28] This can be partially reduced via thermal annealing, however, this may result in bulging of the window. Several studies have attempted to overcome these issues by growing thin SiN_x films on highly insulating substrates without a SiO₂ layer^[Reference Lee, Kumar, Park, Cho, Kim, Lim, Kim and Kim^40] or with a poly-Si^[Reference Yanagi, Ishida, Fujisaki and Takeda^41] layer, however, these devices may exhibit increased noise due to capacitive coupling to the substrate.

Metal electrodes are used to enable electrical contact to graphene nanodevices. Although not strictly necessary, patterning contact electrodes prior to graphene transfer is preferable as it exposes graphene to less processing. Metal electrodes are fabricated using photolithography coupled with electron beam lithography (for large and small feature sizes, respectively) followed by metal deposition and subsequent liftoff. Electron beam lithography is used to pattern electrodes over the SiN_x window with local alignment markers written during this step to allow for later alignment of graphene nanostructures. Note that patterning resist over the SiN_x window requires a higher dose than supported parts of the chip due to the decreased contribution of backscattered electrons. These metal electrodes must be passivated to (i) reduce faradaic currents resulting from electrochemical reactions between the electrodes and ionic solution^[Reference Healy, Ray, Willis, Peterman, Bartel and Drndić^28] and (ii) reduce noise resulting from capacitive coupling to the ionic solution. Several materials have been used for passivation such as SiO₂/SiN_x.^[Reference Ivanov, Instuli, McGilvery, Baldwin, McComb, Albrecht and Edel^42] Al₂O₃,^[Reference Traversi, Raillon, Benameur, Liu, Khlybov, Tosun, Krasnozhon, Kis and Radenovic^25] TiO₂,^[Reference Healy, Ray, Willis, Peterman, Bartel and Drndić^28] h-BN^[Reference Heerema, Vicarelli, Pud, Schouten, Zandbergen and Dekker^30] and various polymeric films. Care should be taken to ensure proper adhesion of the passivation to metal electrodes.^[Reference Temiz, Ferretti, Leblebici and Guiducci^43] In general, there are two considerations to be made regarding the thickness of the passivation layer. First, the film must be thick enough to avoid pinholes in the passivation. This thickness is largely dependent on the deposition process (e.g. atomic layer deposition, electron beam evaporation, CVD, photoresist spincoating) and the chosen dielectric. Second, the passivation thickness should be thick enough to reduce noise resulting from capacitive coupling to the ionic solution. Although dependent on the dielectric used and the device geometry, the authors have found that dielectric thicknesses greater than 500 nm are required to significantly reduce this noise. Electrode passivation is routinely performed after graphene transfer and patterning to ensure that the metal electrodes can be fully passivated as well as nonactive areas of the graphene.

Perhaps, the most challenging aspect of fabrication is forming the nanopore. These nanopores are typically formed using high-energy electron irradiation from a transmission electron microscope (TEM)^[Reference Traversi, Raillon, Benameur, Liu, Khlybov, Tosun, Krasnozhon, Kis and Radenovic^25, Reference Heerema, Vicarelli, Pud, Schouten, Zandbergen and Dekker^30, Reference Puster, Balan, Rodrguez-Manzo, Danda, Ahn, Parkin and Drndić^37, Reference Puster, Rodrguez-Manzo, Balan and Drndić^38] [Figs. 2(c)(i) and 2(d)]. However, this technique requires expensive equipment, expert training, and is time consuming. As such, techniques to reliably and rapidly form nanopores in graphene, SiN_x, and other two-dimensional materials continue to be researched. Other techniques to create nanopores in SiN_x membranes include electron beam lithography followed by reactive ion etching^[Reference Verschueren, Yang and Dekker^44] [Fig. 2(c)(ii)] and focused ion beam milling, including gas field ion sources such as helium ion microscopes and eutectic sources. While these techniques are only capable of reproducibly creating nanopores with diameters on the order of 15 nm and larger, nanopores shrinking methods can be used to reduce the diameter of these pores^[Reference Asghar, Ilyas, Billo and Iqbal^45] to sizes compatible with DNA sequencing applications. An alternate technique to scalably fabricate nanopores in SiN_x relies on electric field-induced breakdown. Namely, pores with subnanometer precision can be fabricated in SiN_x via feedback-controlled dielectric breakdown [Fig. 2(c)(iii)] whereby a large electric field is applied across the membrane while monitoring the leakage current.^[Reference Kwok, Briggs and Tabard-Cossa^46] Techniques to control the position and number of nanopores created via dielectric breakdown continue to be developed.^[Reference Pud, Verschueren, Vukovic, Plesa, Jonsson and Dekker^47, Reference Wang, Ying, Zhou, de Vreede, Liu and Tian^48] Such formation of SiN_x nanopores could then be followed by graphene transfer and nanopore formation by applying pulsed electric fields across the membrane^[Reference Kuan, Lu, Xie, Szalay and Golovchenko^49] to create self-aligned graphene nanopores. Alternately, this pulsed electric field breakdown technique could be used to create nanopores in freestanding graphene membranes, however, a coating layer would be required to avoid DNA sticking to the graphene prior to translocation.^[Reference Kuan, Lu, Xie, Szalay and Golovchenko^49]

Despite several promising studies demonstrating DNA detection via changes in the graphene current as DNA translates through a pore in the graphene,^[Reference Traversi, Raillon, Benameur, Liu, Khlybov, Tosun, Krasnozhon, Kis and Radenovic^25, Reference Heerema, Vicarelli, Pud, Schouten, Zandbergen and Dekker^30, Reference Puster, Balan, Rodrguez-Manzo, Danda, Ahn, Parkin and Drndić^37] recent works continue to cite low fabrication throughput and device yield as significant problems.^[Reference Heerema, Vicarelli, Pud, Schouten, Zandbergen and Dekker^30] This is no doubt in part due to the many fabrication steps required, each of which need to be optimized to return an acceptable yield. Additionally, nanopore formation using a TEM is a time-consuming process which significantly limits device throughput. Techniques to scalably create nanopores in graphene and SiN_x, some of which were highlighted above, will need to continue to be developed if this sequencing approach is to continue to advance over the coming years.

Graphene nanogaps

DNA sequencing techniques based on graphene nanogaps will now be discussed. These devices function by detecting changes in the tunnelling current across a graphene nanogap as DNA translocates through the gap in a direction perpendicular to the graphene plane^[Reference Puczkarski, Swett and Mol^50] [Fig. 1(c)]. Similar to the geometry discussed for graphene nanopores, these devices consist of a graphene nanogap aligned with a nanopore in the underlying SiN_x window. While it may be possible to use graphene as both the membrane and electrode material,^[Reference Postma^26] such a device would likely have limited control over the position of DNA during translocation.^[Reference Lagerqvist, Zwolak and Di Ventra^51] This is a particularly important consideration given the exponential dependence of tunnelling current on gap distance. Graphene nanogap sequencing devices have been studied in several computational works where nucelobase-specific changes in the tunnelling current were predicted due to different coupling strengths between the graphene and the molecular orbitals of each base.^[Reference Postma^26, Reference Prasongkit, Grigoriev, Pathak, Ahuja and Scheicher^52–Reference Amorim, Rocha and Scheicher^55] Additionally, experimental works have detected and discriminated nucleotides via changes in the tunnelling current in a scanning tunnelling microscope^[Reference Ohshiro and Umezawa^56–Reference Chang, Huang, He, Liang, Zhang, Li, Chen, Sankey and Lindsay^59, Reference Di Ventra and Taniguchi^60] and metal nanogaps.^[Reference Di Ventra and Taniguchi^60–Reference Pang, Ashcroft, Song, Zhang, Biswas, Qing, Yang, Nemanich, Bai, Smith, Reuter, Balagurusamy, Astier, Stolovitzky and Lindsay^65] No attempts to detect DNA using tunnelling currents across a graphene nanogap have been published to date.

Several techniques to fabricate graphene nanogaps have been demonstrated that could conceivably be used for nanogap-based DNA sequencing.^[Reference Prins, Barreiro, Ruitenberg, Seldenthuis, Aliaga-Alcalde, Vandersypen and van der Zant^66–Reference Bellunato, Vrbica, Sabater, de Vos, Fermin, Kanneworff, Galli, van Ruitenbeek and Schneider^69] One such technique is feedback-controlled electroburning.^[Reference Prins, Barreiro, Ruitenberg, Seldenthuis, Aliaga-Alcalde, Vandersypen and van der Zant^66, Reference Lau, Mol, Warner and Briggs^70, Reference El Abbassi, Pósa, Makk, Nef, Thodkar, Halbritter and Calame^71] Here, a voltage ramp is applied across a GNR while the current is simultaneously recorded. Large current densities in the graphene induce Joule heating which can lead to oxidation or sublimation of carbon atoms at the edges.^[Reference El Abbassi, Pósa, Makk, Nef, Thodkar, Halbritter and Calame^71] When this happens, the graphene resistance increases resulting in a drop in the current [asterisk in Fig. 3(a)] at which point the voltage is ramped back to zero. Repeating this process allows for the controlled formation of a nanogap across a GNR [Figs. 3(a) and 3(b)]. Recently, graphene nanogaps have also been fabricated via mechanically controlled break junctions.^[Reference Caneva, Gehring, Garca-Surez, Garca-Fuente, Stefani, Olavarria-Contreras, Ferrer, Dekker and van der Zant^67] Here, graphene was patterned on a flexible phosphor bronze substrate and loaded into a three-point bending configuration [Fig. 3(c)]. The graphene current is monitored during bending until a sharp drop in the conductance is observed signifying the formation of a nanogap. Another method that has been used to create graphene nanogaps relies on capillary force-induced breaking of suspended graphene.^[Reference Wang, Zheng, Wang, Qiu, Guo, Shen and Yu^68, Reference Patel, Carroll, Lopez, Sankararaman, Etienne, Kodigala, Paul and Postma^72] In this technique, graphene is suspended across a trench in the substrate, and broken due to the surface tension of the solvent during the drying process [Fig. 3(d)].

Figure 3.

(a) An example of the electroburning process. The first trace is shown in red. The asterisk shows the point where the graphene resistance increases indicating the onset of electroburning. The insets show schematics of the nanogap formation process during electroburning. (b) An atomic force microscope image of a graphene nanogap formed via electroburning on a SiO₂ substrate.^[Reference Lau, Mol, Warner and Briggs^70] The graphene has been patterned into a constriction to enable controlled positioning of the nanogap. (c) Schematic of the mechanical controlled break junction setup to form graphene nanogaps.^[Reference Sun, Diaz-Fernandex, Gschneidtner, Westerlund, Lara-Avila and Moth-Pouslen^98] (d) Schematic of an example fabrication process to produce a graphene nanogap via capillary force induce breaking.^[Reference Wang, Zheng, Wang, Qiu, Guo, Shen and Yu^68] Figures adapted with permissions from (b) PCCP Owner Societies,^[Reference Lau, Mol, Warner and Briggs^70] (c) Royal Society of Chemistry [relevant citation added above] (d) AIP Publishing.^[Reference Wang, Zheng, Wang, Qiu, Guo, Shen and Yu^68]

One of the major challenges in the fabrication of these devices is the alignment of graphene nanogaps with an underlying SiN_x nanopore. Graphene can be fabricated into a constriction to enable the controlled positioning of a graphene nanogap via the electroburning^[Reference Lau, Mol, Warner and Briggs^70] [Fig. 3(b)] and mechanically controlled break junction^[Reference Caneva, Gehring, Garca-Surez, Garca-Fuente, Stefani, Olavarria-Contreras, Ferrer, Dekker and van der Zant^67] techniques discussed above. Patterning of graphene into such geometries is generally done via electron beam lithography followed by etching with alignment of features down to 20 nm readily obtained using local alignment markers. Care should be taken with alignment marker position near the SiN_x window. However, alignment on the order of a few nanometers would likely be required to ensure that a single tunnel junction is reproducibly aligned with a nanopore. Moreover, formation of graphene nanogaps via feedback-controlled electroburning and mechanical-controlled break junctions have to date only been demonstrated on supported or semi-supported substrates and these techniques will need to be developed for suspended graphene membranes compatible with DNA sequencing. It is noted that DNA translocation through graphene nanogaps was recently demonstrated by aligning a “micropore” in a polymeric film with the nanogap.^[Reference Patel, Carroll, Lopez, Sankararaman, Etienne, Kodigala, Paul and Postma^72] By doing this, the authors were able to detect DNA passing though the nanogap via changes in the ionic current, however, simultaneous monitoring of the graphene tunnelling current was not reported. While such a nanogap/micropore assembly may be useful for preliminary studies into DNA detection using graphene tunnelling currents, it is likely that sequencing devices will require more stringent control over the position of DNA as it translates through the nanogap.

The integration of electric and fluidic components is another challenging aspect in the development of graphene-based electronic sequencing techniques. For the nanopore and nanogap devices discussed above, similar fluidic designs to that used in typical ionic current nanopore experiments can be envisioned. These fluidic designs can be broadly classified into three classes; (i) a fluidic reservoir approach, (ii) a microfluidic flow cell approach, and (iii) a nanocapillary approach. For the proposed sequencing devices, electrical connections will need to be incorporated to enable measurements of the graphene current. For the fluidic reservior and microfluidic approaches this could be done by wirebonding to a printed circuit board or the incorporation of pins within the fluidic cell to directly contact the chip. Microfluidic cells have the advantage over fluidic reserviors in that they enable liquids to be pumped through the nanopore enabling pore wetting and fluidic transfer. This is a particular problem for the devices considered here since graphene is not compatible with commonly used techniques that render a nanopore hydrophilic such as Piranha solution and oxygen plasma treatment.^[Reference Heerema, Vicarelli, Pud, Schouten, Zandbergen and Dekker^30] The use of nanocapillaries to fluidically contact devices^[Reference Arcadia, Reyes and Rosenstein^75] is an appealing approach as it may enable easy electrical connection. However, it is yet to be seen whether devices can be effectively contacted and the fluid localized at the nanopore without wetting of the electrical contacts. Additionally, a nanocapillary is not suited to multiplexed measurements.

While promising results have been reported on the detection and discrimination of nucleobases via tunnelling currents in a scanning tunnelling microscope and metal nanogaps, fabrication of graphene nanogaps for tunnelling current sequencing is yet to be reported. Indeed, several studies on metal nanogaps integrated within a nanopore have cited significant difficulties in the fabrication process,^[Reference Healy, Ray, Willis, Peterman, Bartel and Drndić^28, Reference Fanget, Traversi, Khlybov, Granjon, Magrez, Forró and Radenovic^29, Reference Ivanov, Instuli, McGilvery, Baldwin, McComb, Albrecht and Edel^42] which will likely remain for graphene-based devices. Promisingly, several scalable techniques to create graphene nanogaps have already been reported, however, these techniques will need to be developed for suspended graphene structures compatible with DNA seqeuncing. Moreover, the alignment of graphene nanogaps with a SiN_x nanopore to enable controlled DNA translocation poses a significant challenge in the future development of these devices.

Graphene nanoribbons

The last devices that will be discussed here consist of a GNR adjacent to a nanofluidic channel [Fig. 1(d)]. The proposed device works by detecting changes in the graphene current due to adsorption of DNA as it translates through the nanofluidic channel. A schematic of the proposed device geometry is shown in Fig. 4(a) with the envisioned devices consisting of a GNR bridging a nanofluidic channel on a SiN_x (or SiO₂) film. A sealing layer is placed over the nanochannel to enable fluid flow. These devices have been studied computationally where it was predicted that π − π stacking between nucleobases and a nanometer-wide graphene ribbon can lead to base-specific changes in the graphene current due to Fano resonances.^[Reference Min, Kim, Cho and Kim^27, Reference Cho, Min, Kim and Kim^76] Fabrication of devices capable of measuring the current in a nanometer-wide graphene ribbon while DNA translates through an adjacent nanochannel is yet to be demonstrated.

Figure 4.

(a) Schematic of the proposed device geometry. (b) Scanning tunnelling microscopy image of GNRs created via bottom-up synthesis. The inset shows the reaction scheme for the precursor monomers into GNRs.^[Reference Cai, Ruffieux, Jaafar, Bieri, Braun, Blankenburg, Muoth, Seitsonen, Saleh, Feng, Mllen and Fasel^73] (c) Transmission electron micrograph image of a GNR which has been sculpted using high-energy electrons produced in a TEM.^[Reference Xu, Wu, Schneider, Houben, Malladi, Dekker, Yucelen, Dunin-Borkowski and Zandbergen^74] (d) Schematic representation of the process to create nanofluidic channels in a substrate via electron beam lithography. Figures adapted with permissions from (b) Nature publishing group,^[Reference Cai, Ruffieux, Jaafar, Bieri, Braun, Blankenburg, Muoth, Seitsonen, Saleh, Feng, Mllen and Fasel^73] (c) American Chemical Society.^[Reference Xu, Wu, Schneider, Houben, Malladi, Dekker, Yucelen, Dunin-Borkowski and Zandbergen^74]

One of the major challenges in the fabrication of these devices is creating GNRs with widths on the order of 1 nm to ensure a limited number of nucleobases interact with the graphene at a given time. One technique to reliably create nanometer-wide graphene ribbons is via bottom-up synthesis. This method relies on the surface-assisted coupling of molecular precursors followed by their subsequent cyclodehydrogenation^[Reference Cai, Ruffieux, Jaafar, Bieri, Braun, Blankenburg, Muoth, Seitsonen, Saleh, Feng, Mllen and Fasel^73] [Fig. 4(b)]. The controlled alignment of these nanoribbons with electrodes and a nanofluidic channel is likely to be a challenge, however, this could potentially be overcome by developing self-alignment techniques such as dielectrophoretic deposition. Similar alignment issues may also be encountered with several other GNR fabrication techniques such as nanowire lithography^[Reference Bai, Duan and Huang^77] and unzipping of carbon nanotubes.^[Reference Jiao, Zhang, Wang, Diankov and Dai^78] GNRs with widths on the order of 20 nm have been fabricated via electron beam lithography,^[Reference Han, zyilmaz, Zhang and Kim^79] and their widths further reduced via gas-phase oxidation.^[Reference Wang and Dai^80] However, it is difficult to produce widths below 10 nm using this technique due to edge roughness introduced during electron beam lithography. Direct patterning of GNRs can also be achieved using energetic charged particles from a TEM^[Reference Xu, Wu, Schneider, Houben, Malladi, Dekker, Yucelen, Dunin-Borkowski and Zandbergen^74, Reference Fischbein and Drndić^81] [Fig. 4(c)] or helium ion microscope.^[Reference Bell, Lemme, Stern, Williams and Marcus^82]

The fabrication of nanofluidic channels as well as DNA translocation through these channels have been described elsewhere.^[Reference Xia, Yan and Hou^83, Reference Duan, Wang and Xie^84] Techniques such as electron beam lithography,^[Reference Hibara, Saito, Kim, Tokeshi, Ooi, Nakao and Kitamori^85–Reference Riehn, Austin and Sturm^87] interferometric lithography,^[Reference Xia, Ku, Lee and Brueck^88] and nanoimprint lithography are commonly used to create nanofluidic channels.^[Reference Guo, Cheng and Chou^89–Reference Liang and Chou^91] Briefly, these techniques generally involve patterning of a resist followed by etching of the substrate [shown in Fig. 4(d) for electron beam lithography]. These techniques create open nanofluidic channels that must be sealed to enable fluid flow. This can be done via, anodic bonding of a cover slide to the substrate. Metal electrodes may need to be patterned into the substrate to ensure a planar surface that allows for nanochannel sealing. This can be achieved by etching the substrate an appropriate thickness prior to metal deposition.^[Reference Maleki, Mohammadi and Ziaie^92] Note that further studies may need to be done to determine the channel dimensions required to ensure interaction between the DNA and graphene.

Finally, challenges in graphene transfer and contamination will be discussed. While this is clearly an important consideration for all the graphene-based sequencing techniques, it is discussed in this section as contamination may pose a significant issue for these devices as they rely on DNA adsorbing to graphene. Depending on the source of graphene, different graphene transfer methods may be used such as substrate etching, wedge transfer^[Reference Schneider, Calado, Zandbergen, Vandersypen and Dekker^93] and bubble transfer.^[Reference Wang, Zheng, Xu, Dubuisson, Bao, Lu and Loh^94] These techniques leave graphene suspended in an aqueous solution which can then be captured on the desired substrate. A polymeric film over the graphene is typically used during the transfer process to avoid the breaking of graphene. Graphene contamination may occur due to residues of this polymeric film, or other films introduced during lithographic steps. Such residues can be (partially) removed via current annealing^[Reference Moser, Barreiro and Bachtold^95] and thermal annealing.^[Reference Ishigami, Chen, Cullen, Fuhrer and Williams^96] Current and thermal annealing may also be used to remove atmospheric contamination on graphene, however, the surface will become quickly recontaminated when exposed to atmosphere. Techniques such as electrochemically cleaning graphene in solution have been studied and could potentially be used in situ to avoid graphene recontamination.^[Reference Fu, Feng, Panaitov, Kireev, Mayer, Offenhusser and Krause^97]

While the idea to sequence DNA using changes in the current through a GNR due to physisorption of DNA translating through an adjacent nanochannel was first suggested in 2011,^[Reference Min, Kim, Cho and Kim^27] there are yet to be any reports on the fabrication of these devices. This has likely been due to challenges in the fabrication of GNRs with widths on the order of 1 nm and their alignment with contact electrodes and a nanofluidic channel. Techniques such as dielectrophoretic deposition could potentially be used to align graphene and scalably produce devices, however, significant work will need to be done on the development of such techniques.

Conclusion and outlook

Graphene-based electronic DNA sequencing devices have received significant attention from the research community over the past decade. These devices are hoped to not only deliver on the advantages of ionic current nanopore sequencing, but also provide single base resolution, high bandwidth sampling, easy multiplexing and the potential for scalable manufacturing. In particular, three device designs have been suggested; (i) monitor the graphene current as DNA translates through a pore in the graphene, (ii) monitor the tunnelling current across a graphene nanogap as DNA translates through the gap and (iii) monitor the current through a GNR as DNA translates through an adjacent nanochannel. Despite the significant promise of these sequencing techniques, DNA sequencing using these devices is yet to be demonstrated. This has been partly due to the complex fabrication requirements of these devices resulting in low fabrication yields and limited throughput. Indeed, this will need to be addressed if DNA sequencing using these devices is to be realized. In particular, the development of robust, wafer-scale processes for substrate and electrode fabrication will be required. Additionally, techniques to scalably fabricate nanostructures in graphene, such as gaps and pores, as well as the alignment of these structures will need to be advanced. Given the relative infancy of graphene and the rapid development of nanofabrication techniques the authors are hopeful that such techniques will indeed by developed in the coming years. If this does indeed happen, it would represent a significant step toward graphene-based electronic DNA sequencing being realized.