Electron transport in a 1.6~nm-thick double-gated (100) silicon nanosheet: A theoretical study accounting for phonon confinement and remote-phonon scattering

Bimin Cai; Dallin O. Nielsen; Edward Chen; Massimo V. Fischetti; Shoaib Mansoori

Electron transport in a 1.6~nm-thick double-gated (100) silicon nanosheet: A theoretical study accounting for phonon confinement and remote-phonon scattering

T0 review · 3 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read For a 1.6 nm double-gated (100) silicon nanosheet, phonon confinement and interface plasmon-phonon scattering set the room-temperature mobility near 200–360 cm²/Vs and can lower the saturation velocity below bulk silicon.

desk verdict A careful full-band Monte Carlo study whose main qualitative result—low saturation velocity from transfer into flat primed subbands—is likely right, while the quantitative mobility numbers depend on an approximate phonon boundary-condition model that needs independent validation. read the letter →

arxiv 2608.12214 v2 pith:SEKOT54E submitted 2026-08-12 cond-mat.mes-hall

Shoaib Mansoori , Bimin Cai , Edward Chen , Dallin O. Nielsen , Massimo V. Fischetti This is my paper · ORCID

classification cond-mat.mes-hall

keywords phononconfinementelectronmobilitysiliconnanosheetremotescatteringinterfaceplasmon-phononmodessaturationvelocityfull-bandMonteCarlohigh-kgatedielectrics

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Electrons in today's nanoscale transistors move through silicon sheets only a few atoms thick, and this work asks which scattering processes actually limit their speed. For a 1.6 nm double-gated (100) silicon nanosheet with SiO₂/HfO₂ gate stacks, the authors argue that phonons themselves are confined by the surrounding insulators, and that the way they are pinned at material interfaces matters more than any other modeling choice. With acoustic phonons clamped at the SiO₂/HfO₂ interfaces and optical phonons at the Si/SiO₂ interfaces, the computed room-temperature phonon-limited mobility is about $260\ \mathrm{cm^2/Vs}$ ($360\ \mathrm{cm^2/Vs}$ with dielectric screening), well below the $700\ \mathrm{cm^2/Vs}$ obtained from the usual bulk-phonon, elastic high-temperature approximation. Interface plasmon-phonon scattering lowers the low-field mobility further, to about $200\ \mathrm{cm^2/Vs}$, but raises the saturated drift velocity above $2\times10^7\ \mathrm{cm/s}$. The paper also predicts a saturation velocity below bulk silicon in the absence of such scattering, and explains it by high-field transfer into flat primed subbands; if correct, these results mean high-κ gate stacks need not degrade nanosheet transistors, but phonon confinement must be included in transport models.

What carries the argument

The load-bearing object is the elastic-continuum model of confined acoustic phonons, with branch-specific quantized transverse wave vectors $q_{z,m}=m\pi/(t_s+2z_0)$, where $t_s$ is the sheet thickness and $z_0$ is an effective penetration depth of the vibration into the SiO₂ layers, fixed by stress continuity at the interfaces. The electron-phonon coupling is carried by shape functions $G_m(z)$ derived from the divergence of the ionic displacement field; different boundary conditions — free-standing, clamped at Si/SiO₂, clamped at SiO₂/HfO₂ — change these shape functions and the number and zero-point energy of the acoustic branches. The second mechanism is the dielectric-continuum treatment of interface plasmon-phonon modes, where a secular equation yields symmetric and antisymmetric hybrid-mode dispersions and the scattering potential is quantized by equating the time-averaged field energy to the oscillator ground-state energy. A full-band empirical-pseudopotential band structure and a Monte Carlo transport solver complete the machinery, so the flat primed-subband dispersion that controls the saturation velocity is resolved directly rather than through an effective-mass approximation.

What would settle it

Measure the low-field electron mobility of a 1.6 nm double-gated (100) Si nanosheet with the exact 0.6 nm SiO₂/1.2 nm HfO₂ stack; a value well above $360\ \mathrm{cm^2/Vs}$ would show that the clamped outer-interface model over-suppresses phonon scattering. A complementary atomistic calculation of the phonon dispersions and electron-phonon couplings for the same stack would settle whether the elastic-continuum clamping picture is quantitatively correct.

Watch

Extended reading notes

Core claim

The central claim is that phonon confinement, not just electron confinement, controls transport in a 1.6 nm silicon nanosheet. Depending on how the confined acoustic phonons are pinned, the room-temperature mobility ranges from about $160\ \mathrm{cm^2/Vs}$ (free-standing boundary conditions at the Si/SiO₂ interfaces) to about $1060\ \mathrm{cm^2/Vs}$ (clamped at the Si/SiO₂ interfaces). The physically motivated choice — acoustic phonons clamped at the stiff outer SiO₂/HfO₂ interfaces, optical phonons clamped at the Si/SiO₂ interfaces — gives $260\ \mathrm{cm^2/Vs}$ unscreened and $360\ \mathrm{cm^2/Vs}$ with dielectric screening, instead of the $700\ \mathrm{cm^2/Vs}$ of the bulk-phonon elastic approximation. Scattering with interface hybrid plasmon-phonon modes reduces the low-field mobility to about $200\ \mathrm{cm^2/Vs}$ at $1.15\times10^{13}\ \mathrm{cm^{-2}}$ sheet density, yet because these inelastic small-angle collisions keep electrons cool, the saturated drift velocity reaches above $2\times10^7\ \mathrm{cm/s}$. Without those modes, the saturation velocity is $5.5\times10^6\ \mathrm{cm/s}$ ($8.9\times10^6\ \mathrm{cm/s}$ screened), below the bulk value, because electrons transfer at high field into the flat fourfold-degenerate primed subband ladder — an explanation the authors present for previously measured but unexplained low inversion-layer saturation velocities.

Load-bearing premise

The quantitative headline rests on the assumption that acoustic vibrations are pinned at the outer SiO₂/HfO₂ interfaces and can be described by a single approximate penetration depth for every branch; if the true interface condition is different, the computed mobility shifts between about $160\ \mathrm{cm^2/Vs}$ and $1060\ \mathrm{cm^2/Vs}$.

Editorial extensions

If this is right

For the 1.6 nm double-gated stack, the phonon-limited room-temperature mobility is predicted to be about $260\ \mathrm{cm^2/Vs}$ unscreened and $360\ \mathrm{cm^2/Vs}$ with dielectric screening, not the $\sim 700\ \mathrm{cm^2/Vs}$ bulk-phonon value.
The high-field saturation velocity is below the bulk value ($\sim 5.5\times10^6\ \mathrm{cm/s}$ unscreened, $\sim 8.9\times10^6\ \mathrm{cm/s}$ screened) because electrons transfer into the flat primed subbands; this supplies a theoretical explanation for old inversion-layer measurements.
Interface plasmon-phonon scattering lowers the low-field mobility to about $200\ \mathrm{cm^2/Vs}$ but raises the saturated velocity above $2\times10^7\ \mathrm{cm/s}$, so high-κ gate stacks need not degrade nanosheet transistors.
The mobility depends only weakly on electron density and gate bias in these ultra-thin sheets, because the external potential barely perturbs the band structure even at $10^{13}\ \mathrm{cm^{-2}}$ sheet density.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Relaxing the mode-independent penetration-depth approximation — letting $z_0$ vary per branch or per polarization — could shift the relative weights of dilatational versus flexural scattering and is a testable refinement of the model.
An atomistic phonon calculation for the same SiO₂/HfO₂ stack would test the clamping assumption; if interfacial stiffness is softer than assumed, the mobility could move toward the upper part of the 160–1060 cm²/Vs bracket.
The predicted high-field benefit of IPP scattering suggests short-channel transistors, where transport is more streaming-like, could show a larger relative velocity boost than the long-channel mobility alone indicates.
Since the model predicts weak mobility dependence on electron density, a density sweep in a 1.6 nm double-gate device could separate intrinsic phonon-limited behavior from extrinsic interface-roughness and charge-defect scattering.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The thing worth knowing about this paper is that it gives a physically plausible mechanism for the long-standing discrepancy between measured saturation velocities (5–6.5e6 cm/s) and earlier theoretical predictions near 1e7 cm/s in Si inversion layers and nanosheets. That mechanism—transfer into flat, fourfold-degenerate primed subbands at high field—comes out of a full-band Monte Carlo calculation with empirical pseudopotentials, and it is not present in the cited theoretical literature. The paper also treats remote-phonon (IPP) scattering in the full SiO2/HfO2 double-gate stack, including dynamic screening and Landau damping, and finds the effect on mobility is modest because the SiO2 interlayer and metal gates screen the high-k contribution. That is a useful, credible result for device people.

Credit where due: the calculation is internally consistent, the equations in Sections III–V support the reported trends, and the authors are refreshingly candid about the sensitivity of their mobility to boundary conditions. They do not fit any free parameter to the predicted mobility or velocity; material parameters are imported from the bulk literature. The comparison with experimental mobility values is honest—they note they cannot draw exact quantitative conclusions because of extrinsic scattering.

The soft spot is real and the stress-test note gets it right: the headline mobility numbers (260–360 cm2/Vs) are set not by the electron-phonon formalism but by the elastic-continuum phonon confinement model, specifically the boundary conditions and the m-independent penetration depth z0 replacing the actual roots of Eq. (20). The paper states those roots vary between 49% and 91% of the adopted value, and Table III shows a factor of ~6.6 spread across boundary-condition choices. That means the quantitative values should be treated with caution. The qualitative ordering—bulk > clamped-at-outer-interface > free-standing—is probably right, and the saturation-velocity explanation does not depend strongly on z0, since it follows from subband structure and occupation. But the mobility numbers could shift by tens of percent under a better-justified choice of z0.

The lack of code or data is a minor annoyance, not a fatal flaw. The citation pattern is fine: the authors build on their own earlier work, but the central claim is not circular.

This paper deserves a serious referee. I would send it to review, and the main request to the authors should be validation or at least a sensitivity study of the phonon boundary-condition model—ideally against atomistic simulations—and, if possible, release of the simulation code. The saturation-velocity result is important enough that it should not be buried by uncertainty in the mobility numbers.

Referee Report

3 major / 5 minor

Summary. The paper reports a full-band Monte Carlo study of low-field mobility and high-field drift velocity in a 1.6 nm-thick double-gated (100) Si nanosheet with SiO2/HfO2 gate stacks. The electron band structure is obtained from local empirical pseudopotentials, confined acoustic phonons are described by an elastic continuum model under three boundary-condition choices, and interface plasmon-phonon (IPP) scattering is treated in the dielectric continuum limit. The central quantitative claims are that the realistic assumption of acoustic phonons clamped at the SiO2/HfO2 interfaces yields a room-temperature phonon-limited mobility of about 260 cm2/Vs (360 cm2/Vs with static dielectric screening and about 200 cm2/Vs with IPP scattering), much lower than the 700 cm2/Vs obtained with bulk phonons in the elastic high-temperature approximation; that the saturation velocity is reduced below the bulk value because electrons transfer into the flat primed subbands; and that IPP scattering, while lowering the low-field mobility, increases the saturation velocity.

Significance. If the quantitative results are robust, the paper provides a plausible explanation for the long-standing observation of reduced saturation velocity in Si inversion layers and a careful account of remote-phonon effects in extremely scaled nanosheet transistors. The strengths include the use of full-band pseudo-potential electronic structure, explicit treatment of the hybridization of 2DEG plasmons with insulator optical phonons, selection rules for IPP scattering, and a comparison with available experimental data for ultra-thin Si films. However, the headline mobility and velocity numbers rest on an approximate elastic-continuum phonon-confinement model whose key parameter, the penetration depth z0, is replaced by a single m-independent value while the manuscript itself states that the exact roots of Eq. (20) vary between 49% and 91% of that value. Since Table III shows that the mobility changes by a factor of about 6.6 across boundary-condition choices, the central numerical claims are not yet anchored without a sensitivity analysis. No code or tabulated data are provided to make the numerical results independently checkable.

major comments (3)

[Sec. III.B, Eq. (20)] The m-independent replacement z0 ≈ t1(c_Si/c_SiO2)^2 is load-bearing for the headline mobility values. The text states that the smallest roots of Eq. (20) for m = 1..16 lie between 49% and 91% of this adopted value, and Eqs. (21)-(23) as well as the number of acoustic branches (~16) depend directly on z0. Because Table III shows that the mobility varies from 160 to 1060 cm2/Vs depending on the boundary-condition choice, a factor-of-two uncertainty in z0 can plausibly shift the central 260/360/200 cm2/Vs numbers by tens of percent. Please recompute the low-field mobility and saturation velocity using the exact branch-dependent roots of Eq. (20), or at least provide a sensitivity study over the stated 0.49-0.91 range, and report how the conclusions change.
[Sec. VI, Table III] The comparison between the 'realistic' CBC@SiO2/HfO2 results and the bulk-phonon value of 700 cm2/Vs conflates two separate approximations: the choice of phonon boundary conditions and the static-screening approximation. The manuscript acknowledges in Sec. IV that static screening is expected to give an upper bound for the mobility, but no uncertainty estimate is attached to the quoted 260/360/200 cm2/Vs values. Because the physical choice of clamped boundary conditions at SiO2/HfO2 is argued from elasticity tables rather than validated against experiment or atomistic simulation for this specific 1.6 nm geometry, the quantitative strength of the claim that the realistic mobility is 'much smaller' than the bulk value needs to be demonstrated by showing the range of mobility obtained when the elastic parameters in Table I and the screening model are varied within their accepted uncertainties.
[Sec. III.B, Eqs. (21)-(23)] The normalization constant Nm in Eqs. (21) and (22) is approximated by the m-independent form in Eq. (23) in the same spirit as the z0 approximation. This approximation enters directly into the deformation-potential matrix element of Eq. (25), and its accuracy is not separately checked. Since the claimed reduction of the electron-phonon coupling relies on the increased effective mass density of the Si/SiO2 sandwich, a quantitative check of Eq. (23) against the full expression in Eq. (22) for the relevant m values should be reported.

minor comments (5)

[Throughout] There are numerous typographical errors that should be corrected, e.g., 'fogr' after Eq. (23), 'SO2' instead of 'SiO2' in Sec. III.B and in the Fig. 10 caption, 'cmn/s' in Sec. VII, and 'A pgpl' in Ref. [63].
[Sec. VI] The phrase 'a high-kappainsulator' in the discussion of Table III is missing a space and should read 'a high-kappa insulator'.
[Sec. VI, Fig. 11 and Table III] The saturated-velocity values are given as '≳5.6' and '≳8.9' in Table III, but the text quotes specific values (5.5×10^6 and 8.9×10^6 cm/s). Please make the reporting consistent and clarify whether the tabulated entries are lower bounds or converged Monte Carlo estimates.
[Sec. IV, Eq. (32)] The static-screening correction in Eq. (33) is acknowledged to be an approximation, but the direction of the error is stated only qualitatively as 'presumably obtaining a lower bound for the scattering rates.' It would be helpful to state explicitly whether the screened form factor is always ≤1 for the matrix elements used, because the Monte Carlo implementation relies on this inequality.
[Conclusions] The statement that the results for the saturated velocity are 'the first to predict' the lower experimental values would be easier to evaluate if the paper briefly explained why the earlier effective-mass-based Monte Carlo studies cited in Refs. [84]-[88] failed to capture the primed-subband effect, given that the primed-subband dispersion was already known in that literature.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the phonon-confinement and IPP models are independently parameterized inputs; the mobility and velocity results are computed outputs.

full rationale

The paper's central numerical claims are not equivalent to any fitted input. The electron dispersion is computed from empirical pseudopotentials, the acoustic-phonon confinement parameters come from an elastic-continuum model whose material constants are taken from the bulk literature (Table I and Table II), and the IPP scattering rates are computed from a standard dielectric-continuum quantization with independently specified dielectric functions. No free parameter is fitted to reproduce the nanosheet mobility or saturated velocity. The z0 approximation in Sec. III.B is an acknowledged modeling approximation: the paper explicitly states that the exact roots of Eq. (20) vary between 49% and 91% of the adopted asymptotic value, and Table III exposes a roughly 6.6x sensitivity to boundary-condition choice. This is an uncertainty and sensitivity statement, not a disguised fit: the headline 260/360/200 cm2/Vs values are outputs of the model, not targets used to select z0. Self-citations to Refs. [30], [31], and [88] provide framework, dimensionality arguments, and historical context, but the load-bearing deformation potentials, phonon energies, sound velocities, and experimental comparisons are drawn from external sources or are clearly stated as model assumptions. No uniqueness theorem is imported to forbid alternatives, and the paper candidly notes that experimental data are scant and that 'we cannot draw any exact quantitative conclusion.' Thus the derivation chain is self-contained relative to its inputs; the main risks are modeling accuracy of the confinement boundary conditions, which is a scientific uncertainty rather than circularity.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central claim rests on imported empirical parameters and modeling approximations, not on new postulates. The most sensitive ingredient is the phonon boundary-condition choice, which changes the mobility by a factor of 6.6. No invented entities are introduced.

free parameters (6)

Acoustic deformation potentials (Delta_dil=10 eV, Delta_flex=2 eV) = 10 eV and 2 eV
Adopted from Refs. [55,56] and 'corrected' in Table II for spin degeneracy and sound velocity. These set the acoustic-phonon scattering strength and hence the mobility. Not fitted in this paper, but historically tuned in the bulk literature.
Optical deformation potential (DK)_op = 1.75e8 eV/cm
Taken from bulk Si transport literature (Table II). It controls the optical-phonon scattering rate and contributes to the central mobility values.
Phonon zone-edge energies = Dilatational 49.5 meV, flexural 33 meV, LO 66/49.5 meV, TO 66/64 meV
Imported from Ref. [43] (Table II) and used to construct the folded confined-phonon dispersion in Eqs. (7)-(8).
Confined-phonon penetration depth z0 = Roughly 2.5t1 (dilatational) and 3.1t1 (flexural), with t1=0.6 nm
In Section III.B the transcendental equation (20) is replaced by an m-independent approximation for all branches. This sets the effective confinement thickness, the shape functions, and the number of acoustic modes, directly controlling the mobility.
Number of confined acoustic modes = 16 dilatational, 16 flexural, 48 total acoustic
Derived from an atom count and the approximated z0 in Section III.B. The number of modes entering the scattering sum changes the total scattering rate.
Electron sheet density for the IPP case = 1.15e13 cm-2
Chosen to represent the on-state density of a transistor. It sets the 2DEG plasma response, the IPP dispersion, and the IPP scattering rates.

assumptions (6)

domain assumption Elastic continuum approximation for confined acoustic phonons
Section III replaces atomistic phonons with a long-wavelength elastic continuum model and a folded bulk dispersion. The approximation is stated as strictly valid only at long wavelength.
domain assumption Clamped boundary conditions at SiO2/HfO2 interfaces are physically realistic
Section III.B argues HfO2 is stiff enough to clamp acoustic vibrations. This choice is central to the headline 260/360 cm2/Vs mobility, but it is not directly validated against experiments or atomistic simulations.
domain assumption Local empirical pseudopotentials give accurate band structure for H-terminated Si nanosheets
Section II chooses local empirical pseudopotentials to match bulk gap and masses. The external gate potential is neglected after a smallness estimate, which the paper justifies with a parabolic-potential example.
domain assumption Scalar 2DEG dielectric response and approximate Landau damping
Section V uses a scalar dielectric function with form factors ignored and an approximate Q_LD expression in Eq. (48). This affects IPP dispersion and scattering rates.
ad hoc to paper Static screening of electron-phonon interactions is a valid lower-bound approximation
Section IV states that dynamic screening is replaced by static screening, 'presumably obtaining a lower bound for the scattering rates and an upper bound for the mobility'. This is an acknowledged ad hoc approximation for numerical cost.
domain assumption Ideal metal gates and symmetric gate stacks
Throughout the paper, gates are treated as ideal equipotential metals and the structure is assumed symmetric. This simplification is used in the IPP secular equation and the Poisson Green's function of Appendix A.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Electron transport in a 1.6~nm-thick double-gated (100) silicon nanosheet: A theoretical study accounting for phonon confinement and remote-phonon scattering." pith.science (2026). https://pith.science/paper/SEKOT54E

@misc{pith2026260812214,
  author       = {Pith},
  title        = {Pith review of: Electron transport in a 1.6~nm-thick double-gated (100) silicon nanosheet: A theoretical study accounting for phonon confinement and remote-phonon scattering},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SEKOT54E}},
  note         = {Machine review of arXiv:2608.12214}
}

read the original abstract

We study theoretically electron transport in an top-and bottom-gated (100) 1.6 nm-thin silicon nanosheet with SiO2/HfO2 gate stacks, focusing on the intrinsic physical processes that affect transport: the confinement of phonons and the presence of interface hybrid plasmon-phonon excitations (IPPs or `remote phonons'). The band structure is calculated using local empirical pseudopotentials; an approximated elastic continuum model is used to consider the confinement of acoustic phonons; the dielectric continuum limit is used to deal with the IPPs. We find that the electron mobility is affected significantly by the boundary conditions chosen to deal with phonon confinement. The more realistic assumption of phonons clamped at the SiO2/HfO2 interfaces and optical phonons at the Si/SiO2 interfaces results in a room temperature mobility much smaller than what is obtained using the common assumption of bulk phonons in the elastic, high-temperature approximation. We also find that, as a result of the complicated structure of the primed subbands, the high-field saturated velocity is significantly lower than its bulk value, as it had been measured in the past in the case of Si inversion layers but never explained theoretically. Finally, we find that IPP scattering does depress the low-field mobility but to a small extent, thanks to the presence of the interfacial SiO2 layers and to the proximity of the metal gates. Moreover, by keeping electrons `cooler', IPP scattering results in a higher saturated velocity. Therefore, the presence of high-kappa materials in the gate-insulator stacks should not affect negatively the performance of field effect transistors based on Si nanosheets.