betaFeedback

WorksPhysics and Astronomy

Reply to: Mobility overestimation in MoS2_2 transistors due to invasive voltage probes

Abstract

In this reply, we include new experimental results and verify that the observed non-linearity in rippled-MoS2_2 (leading to mobility kink) is an intrinsic property of a disordered system, rather than contact effects (invasive probes) or other device issues. Noting that Peng Wu's hypothesis is based on a highly ordered ideal system, transfer curves are expected to be linear, and the carrier density is assumed be constant. Wu's model is therefore oversimplified for disordered systems and neglects carrier-density dependent scattering physics. Thus, it is fundamentally incompatible with our rippled-MoS2_2, and leads to the wrong conclusion.

These authors contributed equally to this work.

  • Correspondence and requests for materials should be addressed to M. Y. (email: link text) or J. W. (email: link text).

In response to our recent paper[1] on the improvement of carrier mobilities in two-dimensional (2D) semiconductor transistors using rippled molybdenum disulfide (MoS2\mathrm{MoS}_{2}), we received comments from Peng Wu (hereafter referred as PW)[2] regarding a potential “overestimation” of mobility values in our experimental method due to contact issues from partially-invasive device configuration employed in our four-probe measurement setup. The main argument revolves around whether the mobility “kink” observed in our rippled devices arises from device configuration or an intrinsic property.

In our reply, we include new experimental results and verify that the observed non-linearity in rippled-MoS2\mathrm{MoS}_{2} (leading to mobility kink) is an intrinsic property of a disordered system, rather than contact effects (invasive probes) or other device issues. Noting that PW’s hypothesis is based on a highly ordered ideal system, transfer curves are expected to be linear, and the carrier mobility is assumed to be constant. PW’s hypothesis is therefore oversimplified for disordered systems and neglects carrier-density dependent scattering physics.[3, 4] Thus, it is fundamentally incompatible with our rippled-MoS2\mathrm{MoS}_{2}, and leads to the wrong conclusion.

First, experimental results for rippled- and flat-MoS2\mathrm{MoS}_{2} presented in our article were obtained using the same measurement method. If the observed mobility kink was caused by contacts or device configuration issues, we would expect to observe the same phenomenon in both rippled- and flat-MoS2\mathrm{MoS}_{2}. However, this is not the case. Additionally, we addressed the issue of contacts and demonstrated that our devices have good ohmic contact (Fig. S15 and S16 in the paper) that are consistent with literature,[5, 6] thus indicating that contacts minimally affects the transistor performance.

Second, according to PW’s hypothesis, if the shift in threshold voltage (VthV_{th}) were solely due to contacts, the mobility kink would be accompanied by a poorer performance and smaller source-drain current (IsdI_{sd}).[7, 8] However, in all rippled-MoS2\mathrm{MoS}_{2} devices, the IsdI_{sd} is at least an order of magnitude larger than in flat-MoS2\mathrm{MoS}_{2}, despite similar device configuration. For instance, at room-temperature, a rippled-MoS2\mathrm{MoS}_{2} device (channel width ∼\sim1 μ\mum, Vsd=0.5V_{sd}=0.5 V) is capable of producing IsdI_{sd} reaching hundreds of microamperes, which is difficult to achieve in flat-MoS2_2. Additionally, our results confirm no notable shift in VthV_{th} when comparing invasive and non-invasive voltage probes, as will be discussed later. These findings proves that the observed mobility kink is indeed attributed to changes in the intrinsic property of rippled-MoS2_2.

Third, it is evident that PW’s hypothesis loses its validity at high gate voltages, indicating an oversimplification. In materials with electronic polarization and high dielectric constants, an increase in carrier density causes a softening of phonon modes, leading to enhanced electron-phonon interactions.[9] This phenomenon is primarily responsible for the observed non-linearity in rippled-MoS2_2, and it explains the rapid decrease in mobility with increasing gate voltage. Similar non-linearities are commonly reported in 2D systems with high dielectric constant or high disorder, as extensively discussed in previous studies.[4, 10, 11, 12, 13, 14, 15] Take Bi2_2O2_2Se for example, its mobility initially peaks at ∼\sim2000 cm2^2 V−1^{-1}s−1^{-1} before decreasing to ∼\sim200 cm2^2 V−1^{-1} s−1^{-1} with increasing carrier density at room-temperature,[11] exhibiting analogous behavior to rippled-MoS2_2. This behavior arises because mobility is carrier-density dependent. Neglecting these carrier-density dependent scattering physics in theoretical models would inevitably result in the incorrect extraction of mobility values.

The influence of contacts on carrier mobility in transition metal dichalcogenide (TMD) transistors is widely recognized and extensively investigated by the 2D community.[16, 17] This is why studies on 2D transistors consistently assess and compare contact resistances with existing literature. Typically, linear I-V curves are presented to demonstrate ohmic contacts at the metal-semiconductor interface to ensure minimal impact on transistor performance. Depending on the intended application, the 2D community employs various device configurations (invasive, partially-invasive and non-invasive) for transistor measurements, each offering its own advantages and disadvantages.[11, 18, 19]

To provide further evidence, we followed PW's suggestion and performed additional experiments. Here, we present a direct comparison between two devices: Device 1 with a non-invasive configuration (as proposed by PW, Fig. 1a) and Device 2 with a partially-invasive configuration (as in our paper[1], Fig. 1d) made using the same MoS2_2 flake. We first measured the 2P mobility of both devices which represent fully-invasive configuration that induces the largest disturbance to electric potential distribution within the channel. The 2P mobility values are 536 and 618 cm2 V−1 s−1\mathrm{cm^2\,V^{-1}\,s^{-1}} respectively (Fig. 1c,f), consistent with our reported device variation. Next, we measured the 4P mobility for Device 1 (non-invasive) where disturbance to electric potential distribution within the channel is minimal, and for Device 2 (partially-invasive) where the disturbance to electric potential distribution is intermediate.[20] If contact effects are dominant as suggested by PW, the 4P mobility values in Device 1 (non-invasive) and Device 2 (partially-invasive) are expected to be substantially smaller than their 2P mobility values. In addition, a more significant drop in 4P mobility from its own 2P mobility is expected in Device 1 compared to Device 2 since it contains the largest disturbance to electric potential distribution within the channel. However, no notable differences between 4P and 2P mobility values in both devices are observed (Table 1). In Device 1 (non-invasive), the difference between 4P mobility (541 cm2 V−1 s−1\mathrm{cm^2\,V^{-1}\,s^{-1}}) and 2P mobility (536 cm2 V−1 s−1\mathrm{cm^2\,V^{-1}\,s^{-1}}) is only ∼1%\sim1\% (Fig. 1c). Similarly, in Device 2 (partially-invasive), the difference between 4P mobility (596 cm2 V−1 s−1\mathrm{cm^2\,V^{-1}\,s^{-1}}) and 2P mobility (618 cm2 V−1 s−1\mathrm{cm^2\,V^{-1}\,s^{-1}}) is only ∼4%\sim4\% (Fig. 1f). All within the reasonable variation due to the channel geometry measurement. This indicates that invasive voltage probe does not play a crucial role in determining the measured carrier mobility, as opposed to PW’s suggestion. Moreover, no significant shift in VthV_{\mathrm{th}} between transfer curves of invasive and non-invasive configurations is observed. This is supported by the consistent mobility peak shift of 2V2\mathrm{V} between 2P and 4P measurements in Device 1 (non-invasive) and Device 2 (partially-invasive) (Fig. 1b,e). It is also worth noting that the variation in mobility between the presented two devices is attributed to varying lattice distortion arising from the distribution of ripples, rather than an overestimation due to contact geometry. These findings reveal that PW's hypothesis is invalid in our rippled-MoS2_2.

2P mobility (cm2 V−1 s−1\mathrm{cm}^{2}\ \mathrm{V}^{-1}\ \mathrm{s}^{-1})4P mobility (cm2 V−1 s−1\mathrm{cm}^{2}\ \mathrm{V}^{-1}\ \mathrm{s}^{-1})Difference between 2P and 4P mobility
Device 1536.33 (invasive)541.13 (non-invasive)∼\sim1%
Device 2618.59 (invasive)595.69 (partially-invasive)∼\sim4%

Table 1. Comparison of mobility

Furthermore, PW misunderstood the definition of VthV_{\mathrm{th}} and its distinction from the zero carrier density position in Hall measurements. Due to residual carrier density, this position might not even exist in FETs with a small bandgap material at room temperature. As such, the extracted VthV_{\mathrm{th}} at the zero carrier density position from Hall measurements can be easily shifted, for example, to -200V for a MoS2_2 device that is already OFF at -10V gate voltage.[18] Hence, the VthV_{\mathrm{th}} estimated from Hall experiments by PW is invalid. This is corroborated by our results below where no notable shift in VthV_{\mathrm{th}} is observed in different device configurations, as well as the divergence of the calculated electrical conductivity at high gate voltages (>15>15 V) in PW’s hypothesis.

In summary, the enhanced carrier mobilities showcased in our paper[1] are supported by the comparison between non-invasive and invasive device configurations, demonstrating that invasive voltage probes do not play a significant role in charge transport within our rippled devices. Thus, we stand by our interpretation that the enhanced carrier mobilities are due to suppressed phonon scattering and enhanced intrinsic dielectric constant originating from the rippled-structure-induced electronic polarization.

Data availability

All data generated or analyzed during this study are included in the published article. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Six-panel comparison of Devices 1 and 2, showing non-invasive and partially-invasive device images, conductance curves, and mobility curves

Figure 1.

References

  1. [1]Ng, H. K. et al. Improving carrier mobility in two-dimensional semiconductors with rippled materials. Nat. Electron. 5, 489–496 (2022).DOI
  2. [2]Wu, P. Mobility overestimation in MoS₂ transistors due to invasive voltage probes. arXiv:2206.12787 (2022).arxiv.org/abs/2206.12787
  3. [3]Choi, H. H., Cho, K., Frisbie, C. D., Sirringhaus, H. & Podzorov, V. Critical assessment of charge mobility extraction in FETs. Nat. Mater. 17, 2–7 (2018).DOI
  4. [4]Adam, S., Hwang, E. H. & Das Sarma, S. Scattering mechanisms and Boltzmann transport in graphene. Phys. E Low-Dimensional Syst. Nanostructures 40, 1022–1025 (2008).DOI
  5. [5]Shen, P. C. et al. Ultralow contact resistance between semimetal and monolayer semiconductors. Nature 593, 211–217 (2021).DOI
  6. [6]Cheng, Z. et al. How to Report and Benchmark Emerging Field-Effect Transistors. Nat. Electron. 5, (2022).DOI
  7. [7]Liu, C. et al. Device Physics of Contact Issues for the Overestimation and Underestimation of Carrier Mobility in Field-Effect Transistors. Phys. Rev. Appl. 8, (2017).DOI
  8. [8]Liu, Y. et al. Promises and prospects of two-dimensional transistors. Nature 591, 43–53 (2021).DOI
  9. [9]Ma, J., Yang, R. & Chen, H. A large modulation of electron-phonon coupling and an emergent superconducting dome in doped strong ferroelectrics. Nat. Commun. 12, 1–9 (2021).DOI
  10. [10]Jang, C. et al. Tuning the effective fine structure constant in graphene: Opposing effects of dielectric screening on short- and long-range potential scattering. Phys. Rev. Lett. 101, 1–4 (2008).DOI
  11. [11]Wu, J. et al. High electron mobility and quantum oscillations in non-encapsulated ultrathin semiconducting Bi₂O₂Se. Nat. Nanotechnol. 12, (2017).DOI
  12. [12]Yang, F. et al. Gate-Tunable Polar Optical Phonon to Piezoelectric Scattering in Few-Layer Bi₂O₂Se for High-Performance Thermoelectrics. Adv. Mater. 33, 2004786 (2021).DOI
  13. [13]Wu, J. et al. Low Residual Carrier Concentration and High Mobility in 2D Semiconducting Bi₂O₂Se. Nano Lett. 19, 197–202 (2019).
  14. [14]Quhe, R. et al. High-performance sub-10 nm monolayer Bi₂O₂Se transistors. Nanoscale 11, 532–540 (2019).DOI
  15. [15]Li, T. et al. A native oxide high-κ gate dielectric for two-dimensional electronics. Nat. Electron. 3, 473–478 (2020).DOI
  16. [16]Chang, H. Y., Zhu, W. & Akinwande, D. On the mobility and contact resistance evaluation for transistors based on MoS₂ or two-dimensional semiconducting atomic crystals. Appl. Phys. Lett. 104, (2014).DOI
  17. [17]Nasr, J. R., Schulman, D. S., Sebastian, A., Horn, M. W. & Das, S. Mobility Deception in Nanoscale Transistors: An Untold Contact Story. Adv. Mater. 31, 1–9 (2019).DOI
  18. [18]Baugher, B. W. H., Churchill, H. O. H., Yang, Y. & Jarillo-herrero, P. Intrinsic Electronic Transport Properties of High Quality Monolayer and Bilayer MoS₂. Nano Lett. 13, 4212–4216 (2013).DOI
  19. [19]Radisavljevic, B. & Kis, A. Mobility Engineering and a Metal-insulator Transition in Monolayer MoS₂. Nat. Mater. 12, 815–820 (2013).DOI
  20. [20]Choi, H. H. et al. Accurate Extraction of Charge Carrier Mobility in 4-Probe Field-Effect Transistors. Adv. Funct. Mater. 28, (2018).DOI

Paper details

vlnc:W_489atgtk
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents