Abstract

Using the Green’s function method, we study the modulation of the conductance in zigzag graphene nanoribbon (ZGNR) junctions by the gate voltages. As long as the difference between the gate voltages applied on the left and right ZGNRs (ΔV) remains unchanged, the conductance profiles for different cases are exactly the same, except to a displacement along EF-axis. It is found that the transmission of electrons from the upper/lower edge state of the left ZGNR to the lower/upper edge state of the right ZGNR is forbidden, therefore, the width of the conductance gap increases first and then decreases as |ΔV| increases. The upper/lower edge states and conduction/valence subbands of ZGNR under higher/lower gate voltage (VH/VL) determine step positions of the conductance when EF >VH/EF < VL. But when VL ≤ EF ≤ VH, the conductance profile is mainly determined by the upper and lower edge states, a few lowest conduction subbands/topmost valence subbands of ZGNR under lower/higher gate voltage. These results are helpful to the exploration and application of a new kind of field effect transistor based on ZGNR junctions.

1. Introduction

Graphene has excellent electrical, optical, and thermal properties, and has been considered as a perspective base for the postsilicon electronics since the successful fabrication in experiment [14]. Around the Dirac points, the band structure of graphene presents a linear dispersion, and the quasiparticles obey the massless Dirac equation and relativistic-like behaviors appear [37]. If graphene is patterned into graphene nanoribbon (GNR), interesting properties emerge [8]. For example, the edge states are found in graphene nanostructures with zigzag edge [4, 915]. In fact, many interesting phenomena in graphene and other novel two dimensional honeycomb lattice materials are in nature associated with the edge states [4, 9, 1520], e.g., the valley-filtered transport [10, 2123] and magnetism [1113].

Furthermore, by interconnecting two semi-infinite GNRs with different widths, a GNR junction can be obtained [17, 23]. The conductance of metal–semiconductor GNR junctions, which is the key elements in all-graphene circuits, has attracted extensive research recently [2426]. Due to the mismatch between conducting channels in the left and right GNRs, a traveling carrier is strongly scattered at the junction interface and a finite junction conductance is induced [17, 23, 27, 28]. Therefore, the junction conductance strongly depends on the geometry of the junction interface [24].

Experimentally, the Fermi level (EF) in the GNR can be above or below the Dirac points by tuning the gate voltage, then the global or local charge carriers can be easily tuned from electron like to hole like and vice versa [1, 4, 5, 29, 30]. Moreover, the transport of topological edge states can be manipulated by adjusting the gate voltage embedding on the surface of two-dimensional topological insulator systems [31], and the spin polarization [32], spin inversion [33], and valley polarization [34] in silicene nanoribbons can also be manipulated by the gate voltage.

However, to the best of our knowledge, modulation of the electronic and transport properties of zigzag graphene nanoribbon (ZGNR) junctions by the gate voltages applied on the left and right ZGNRs (VgL and VgR, with ΔV = VgLVgR) has not been studied, which is crucial for designing all-graphene junctions and circuits [35]. Thus in this paper, to explore how the gate voltage affects the conductance channels (the edge states, the conduction, and valence subbands) and the mode matching between the left and right ZGNRs, we study how the conductance of the ZGNR junction depends on the gate voltage applied on the left and right ZGNRs. The GEF curves in all cases can be clarified from the energy band structures of the left and right ZGNRs. For VgL = 0 (VgR = 0), the G profiles for opposite VgR (VgL) are symmetric to each other with respect to EF = 0, and they move further away from each other with increasing |ΔV|. For a given ΔV, the G profiles for different VgL or VgR are identical, except to a displacement along EF-axis. The transmission probability of electrons from the upper/lower edge state of the left ZGNR to the lower/upper edge state of the right ZGNR is found to be 0. As a result, the width of the conductance gap increases first and then decreases with increasing |ΔV|.

The rest of the paper is organized as follows. In Section 2, the Green’s function method, the tight-binding model, and the geometrical structure of the ZGNR junction are introduced in brief. In Section 3, the dependance of the conductance of the ZGNR junction on the gate voltages applied on the left and right ZGNRs is studied and analyzed in detail. Finally, conclusions are given in Section 4.

2. Theory and Model

Figure 1 shows the geometry of the zigzag ZGNR junction connecting the left and right leads (semi-infinite ZGNR). From the top to down, atoms in a unit cell are labeled as 1, 2, …, NL/NR. Here NL/NR denotes the width of the left/right ZGNR.

The Hamiltonian for the ZGNR junction reads [5, 3639]

Here t = 2.75 eV is the transfer energy of the nearest neighbor hopping, <ij> represents the nearest neighbors, α denotes the spin index, and EF is the gate voltage applied on the ZGNR.

The conductance of the ZGNR junction can be calculated from the Landauer–Büttiker formula [40, 41],

Here is the unit quanta of conductance considering the spin degeneracy [26]. is the retarded Green function [42, 43]. is the self-energy, is the line width function, and is the surface Green function of the left/right semi-infinite ZGNR [4447].

3. Results and Discussion

In this section, assuming NL = 40 and NR = 20, the dependance of the conductance of the ZGNR junction on the gate voltages applied on the left and right ZGNRs is studied and analyzed in detail. The GEF curves are clearly clarified by analyzing the energy band structures of the left and right ZGNRs. First, the conductance of the ZGNR junction for VgR = 0 and different VgL are explored. Second, for VgL = 0, how G depends on VgR is analyzed. Next, assuming VgL = −VgR = Vg, the variation of G as a function of Vg is considered. Finally, for a given ΔV, the G profiles of the ZGNR junction under different VgL or VgR are compared.

3.1. Conductance of the ZGNR Junction for VgR = 0 and Different VgL

In Figures 2(a)2(d) and 3(a)3(d), we show G versus EF in the ZGNR junction for VgR = 0 and VgL = 0, 0.018, ±0.4, ±0.8, 0.8964, and 1.2 eV, and the corresponding energy band structures of ZGNRs.

When VgR = 0, the G profile for VgL = Vg is symmetric to that for VgL = −Vg with respect to EF = 0, i.e., , because the conduction and valecne subbands are just reversed when the sign of the gate voltage applied on the left ZGNR is reversed. With the increase of |VgL|, the profile of moves further away from that of , and G decreases as a whole, because the energy mismatch of the conducting channels increases.

For VgL > 0 (VgL < 0) and EF > VgL (EF < VgL), G decreases with the increase of |VgL| as a whole, and step positions of G are determined by the upper (lower) edge state and conduction (valence) subbands of ZGNR under higher (lower) gate voltage. For VgL > 0 (VgL < 0) and EF < 0 (EF > 0), G for different VgL are in proximity to that of VgL = 0, especially for a lower |EF|, since step positions of G are mainly determined by the lower (upper) edge states and the valence (conduction) subbands of ZGNR under lower (higher) gate voltage. When VgL = VgR, step positions of G are just decided by the subbands of the right narrower ZGNR.

When 0.018 ≤ |VgL| ≤0.8964 eV and VgL > 0 (VgL < 0), there is a conductance gap at the positive (negative) direction of EF-axis. By increasing |VgL|, the width of the conductance gap increases first and then decreases. These originate from that the transmission of electrons from the upper/lower edge state of the left ZGNR to the lower/upper edge state of the right ZGNR is forbidden. In fact, when the zigzag-chain number N is even (here NL = 40 and NR = 20), the electron transport in the ZGNR should satisfy the pseudoparity conservation and the valley valve effect appears [10, 46, 48], so electrons of the lower/upper edge state in the left ZGNR cannot transmit to the upper/lower edge sate in the right ZGNR.

When VgL = 0.018 eV, the profile of G is almost the same as that of VgL = 0, but G decreases abruptly to 0 at EF = 0.018 eV, then increases rapidly to a 1·G0 plateau at EF = 0 eV. With further increasing VgL, the conduction gap begins to form and the width of conduction gap increases. This is determined by the lower edge state of ZGNR under higher gate voltage and lower edge state of ZGNR under lower gate voltage.

When VgL = 0.4 eV, with the decrease of EF, G decreases to a 1·G0 plateau at EF = 0.7 eV, and decreases to a 0 plateau at EF = 0.4 eV, which is determined by the conduction subbands and the upper and lower edge states of ZGNR under higher gate voltage. Then G increases to a 0.5·G0 plateau at EF = 0.1 eV, which is determined by the topmost valence subband of ZGNR under higher gate voltage. So the conductance gap is in the EF interval [0.1, 0.4 eV], rather than [0, 0.4 eV], which is determined by the topmost valence subband and lower edge state of ZGNR under higher gate voltage. Finally, G shows a dip at EF = 0 and increases step by step as EF decreases, which is determined by the lower edge state and valence subbands of ZGNR under lower gate voltage.

When VgL = 0.8 eV, with the decrease of EF, G decreases to a 0.5·G0 plateau at EF = 0.8 eV. Then G decreases to a 0 plateau at EF = 0.59 eV, and increases to a 0.8·G0 plateau at EF = 0.5 eV, which is determined by the lowest conduction subband of ZGNR under lower gate voltage and topmost valence subband of ZGNR under higher gate voltage. So the conductance gap is in the EF interval [0.5, 0.59 eV], rather than [0, 0.8 eV].

When VgL = 0.8964 eV, the edge position of the lowest conduction subband of ZGNR under lower gate voltage just coincides with that of the topmost valence subband of ZGNR under higher gate voltage, so G decreases to 0 and increases rapidly to a 0.8·G0 plateau at EF = 0.59 eV with the decrease of EF. Therefore, the conduction gap begins to disappear, rather than lies in the EF interval [0, 0.8964 eV]. With further increasing VgL, there is no conduction gap.

When VgL = 1.2 eV, with the decrease of EF, G increases to a 1.4·G0 plateau at EF = 0.91 eV, and decreases to a 0.9·G0 plateau at EF = 0.59 eV. This is also determined by the topmost valence subband of ZGNR under higher gate voltage and the lowest conduction subband of ZGNR under lower gate voltage.

3.2. Conductance of the ZGNR Junction for VgL = 0 and Different VgR

In Figures 4(a)4(d) and 5(a)5(d), we show G versus EF in the ZGNR junction for VgL = 0 and VgR = 0, 0.018, ±0.4, ±0.8, 0.8964, and 1.2 eV, and the corresponding energy band structures of ZGNRs.

Here the G profiles for VgL = 0 and different VgR can be discussed similarly as that in Section 3.1 for VgR = 0 and different VgL, and the G profiles are found to be the same. For example, the G profiles for VgR = 0 and VgL = −0.8, −0.4, 0.4, and 0.8 shown in Figure 2(a) are the same to that for VgL = 0 and VgR = 0.8, 0.4, −0.4, and −0.8 shown in Figure 4(a), respectively. By comparing the G profiles for the above cases, it is found that as long as the difference between the gate voltages applied on the left and right ZGNRs (ΔV) remains unchanged, the conductance profiles for different cases are exactly the same, except to a displacement along EF-axis. This general rule and detailed reasons will be further discussed in Section 3.4.

3.3. Conductance of the ZGNR Junction for VgL = −VgR

Figure 6 shows G versus EF in the ZGNR junction for VgL = −VgR = Vg = 0.009, ±0.4, 0.4482, and 0.8 eV and Figures 7(a1–a4) and 7(b1–b4) show the corresponding energy band structures of ZGNRs.

The G profile for VgL = −VgR = Vg is symmetric to that for VgL = −VgR = −Vg with respect to EF = 0, i.e., , because the conduction and valence subbands are just reversed when the sign of the gate voltage applied on the left/right ZGNR is reversed. With the increase of |Vg|, the profile of moves further away from that of , and G decreases as a whole, because the energy mismatch of the conducting channels increases.

For Vg > 0 (Vg < 0) and EF > Vg (EF < Vg), step positions of G are determined by the upper (lower) edge states and conduction (valence) subbands of ZGNR under higher (lower) gate voltage. For Vg > 0 (Vg < 0) and EF < −Vg (EF > −Vg), step positions of G are mainly determined by the lower (upper) edge states and the valence (conduction) subbands of ZGNR under lower (higher) gate voltage.

When 0.009 ≤ |Vg| ≤ 0.4482 eV and Vg > 0 (Vg < 0), there is a conductance gap at the positive (negative) direction of EF-axis. With the increase of |Vg|, the width of the conductance gap first increases and then decreases. As discussed above, these originate from that the transmission of electrons from the upper/lower edge state of the left ZGNR to the lower/upper edge state of the right ZGNR is forbidden as a result of the pseudoparity conservation and the valley valve effect [10, 46, 48]. Here the G profiles are the same to that for 0.018 ≤ |VgL| ≤ 0.8964 eV and VgR = 0, because ΔV in the above two cases is the same.

When Vg = 0.009 eV, the profile of G is almost the same as that of Vg = 0, but G decreases abruptly to 0 around EF = 0.009 eV, then increases rapidly to a 1·G0 plateau at EF = −0.009 eV, namely, the 1·G0 plateau shows a conductance dip to 0 at EF = 0. With further increasing Vg, the conduction gap begins to form and the width of conduction gap increases. This is determined by the lower edge state of ZGNR under higher gate voltage and upper edge state of ZGNR under lower gate voltage. Here the G profiles are the same to that for VgL = 0.18 eV and VgR = 0, because ΔV in the above two cases is the same.

When Vg = 0.4 eV, with the decrease of EF, G decreases to a 1·G0 plateau at EF = 0.7 eV, and decreases to a 0.5·G0 plateau at EF = 0.4 eV, which is determined by the conduction subbands and the upper and lower edge states of ZGNR under higher gate voltage. Then G decreases slowly to a 0 plateau at EF = 0.2 eV, and increases to a 0.8·G0 plateau at EF = 0.1 eV, which is determined by the lowest conduction subband of ZGNR under lower gate voltage and the topmost valence subband of ZGNR under higher gate voltage. So the conductance gap is in the EF interval [0.1, 0.2 eV], rather than [0, 0.4 eV]. Finally, G shows a dip and increases to a 1·G0 plateau at EF = −0.4 eV, then increases step by step with the decrease of EF. This is determined by the upper and lower edge states and valence subbands of ZGNR under lower gate voltage. Here the G profiles are the same to that for VgL = 0.8 eV and VgR = 0, because ΔV in the above two cases is the same.

When Vg = 0.4482 eV, the edge position of the lowest conduction subband of ZGNR under lower gate voltage just coincides with that of the topmost valence subband of ZGNR under higher gate voltage, so G decreases to 0 and increases rapidly to a 0.9·G0 plateau at EF = 0.15 eV with the decrease of EF. Therefore, the conduction gap begins to disappear, rather than lies in the EF interval [0, 0.4482 eV]. As |Vg| increases, there is no conduction gap. Here the G profiles are the same to that for VgL = 0.8964 eV and VgR = 0, because ΔV in the above two cases is the same.

When Vg = 0.8 eV, with the decrease of EF, G increases to a 1.5·G0 plateau at EF = 0.48 eV, and increases to a 2.25·G0 plateau at EF = 0.27 eV. This is determined by the topmost two valence subbands of ZGNR under higher gate voltage. When −0.233≤EF ≤ 0.27 eV, G is determined by the upper edge state, a few lowest conduction subbands of ZGNR under lower gate voltage and a few topmost valence subbands of ZGNR under higher gate voltage. Finally, G decreases to a 0.9·G0 plateau at EF = −0.233 eV, which is determined by the lowest conduction subband of ZGNR under lower gate voltage.

3.4. Conductance of the ZGNR Junction for VgLVgR = ΔV ()

In this Section 3.4, the G profiles of the ZGNR junction for the same ΔV but different VgL or VgR are compared. Taking ΔV = 0.8 eV as an example, the G profiles of the ZGNR junction for different VgL or VgR are demonstrated in Figure 8. The G profiles for the same ΔV are exactly the same, i.e., for any Vg. If Vg > 0 (Vg < 0), the G profile for VgL can be obtained by translating that for VgLVg by |Vg| along the positive (negative) direction of EF-axis. In fact, the gate voltage applied on the ZGNR will shift the position of the energy subbands (conducting channels), the transmission coefficient of electrons or holes from one channel in the left ZGNR to that of the right one is just determined by the relative positions of the above conducting channels.

Therefore, as long as the difference between the gate voltages applied on the left and right ZGNRs (ΔV) remains unchanged, the conductance profiles for different cases are exactly the same, except to a displacement along EF-axis. Moreover, as discussed above, the width of the conductance gap increases first and then decreases as |ΔV| increases.

These results are helpful to the exploration and application of a new kind of field effect transistor (FET) based on ZGNR junctions, in which the conducting channels involved in the transmission of electrons or holes are controlled by the gate voltage, and their functions are similar to those of semiconductor FETs.

4. Conclusion

In summary, by adjusting the gate voltages applied on the left and right ZGNRs, the conductance of the ZGNR junction is studied. The GEF curves for all cases can be clarified from the energy band structures of the left and right ZGNRs. When VgL = 0 (VgR = 0), the G profiles for opposite VgR (VgL) are symmetric to each other with respect to EF = 0, and they move further away from each other with the increase of |ΔV|. For a given ΔV, the G profiles for different VgL or VgR are exactly the same, except to a displacement along the positive or negative direction of EF-axis. Because the transmission of electrons from the upper/lower edge state of the left ZGNR to the lower/upper edge state of the right ZGNR is not allowed, the width of the conductance gap increases first and then decreases as |ΔV| increases. The width of the conduction gap is mainly determined by the upper and lower edge states, a few lowest conduction subbands of ZGNR under lower gate voltage and a few topmost valence subbands of ZGNR under higher gate voltage. Step positions of G are determined by the upper edge state and conduction subbands of ZGNR under higher gate voltage (EF) when EF > VH, and by the lower edge state and the valence subbands of ZGNR under lower gate voltage (VL) when EF < VL. These results are helpful to the exploration and application of graphene-based FETs.

Data Availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This work was supported by the Scientific Research Innovation Team of the Xuchang University (grant no. 2022CXTD005) and the “316” Project Plan of the Xuchang University.