Current-Voltage Characteristics of Long-Channel Nanobundle Thin-Film Transistors: A Bottom-up Perspective

J. Y. Murthy; M. A. Alam; N. V. Pimparkar; S. Kumar

REVIEW

Current-Voltage Characteristics of Long-Channel Nanobundle Thin-Film Transistors: A Bottom-up Perspective

N. V. Pimparkar, S. Kumar, J. Y. Murthy, M. A. Alam

Mesoscale and Nanoscale Physics · Submitted 1 Mar 2006 · 9 references

Pith's verdict: Plausible numerical scaling law for nanobundle transistors, but the universal claim rests on a deferred proof and an absent out-of-sample test.

T0 · AI review

How this review was made: An AI referee (deepseek-v4-flash) read the paper on August 28, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

T0 AI review An AI read the paper and answered a fixed list of questions.
T1 Audit Automatic checks that the references and files the paper relies on exist and match.
T2 Compute The paper's code or calculations re-run with the same results.
T3 Formal statement The claims written out in a formal math language a computer can read.
T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 4 major objections · 6 minor · Ask about this paper · The compiled record · Read the full review →

Challenge this review Comment (0) Read on arXiv ↗

Cite ▾

arXiv IDcond-mat/0603033 (version 1)

Pith IDRZJFIOUT, Pith's permanent record of this paper and its review

CitationJ. Y. Murthy, M. A. Alam, N. V. Pimparkar, S. Kumar. "Current-Voltage Characteristics of Long-Channel Nanobundle Thin-Film Transistors: A Bottom-up Perspective." arXiv:cond-mat/0603033 (2006). https://pith.science/paper/RZJFIOUT

Cite this reviewBibTeX for Pith's review of this paper

Claim paper

Why Pith says this

Highlight any text to comment on it.

Editor's note

read the letter

Colleague,

Quick take: this paper shows something genuinely interesting and narrow — that the exponent m in the drain-current vs channel-length scaling is the same at low and high bias for simulated long-channel nanobundle TFTs. That is a real numerical observation, and if it holds up it is useful for compact modeling. But the paper goes further and claims that measuring one device predicts the I-V of any other device with arbitrary geometry and bias. That claim is not supported by the evidence presented.

What's new: the short-channel version of the factorization was in their own IEDM paper [12]. Here they test it for LC > LS, where sticks don't bridge the channel and junction interactions matter. Their self-consistent drift-diffusion/Poisson simulations show that log(ID·LS) vs log(LC/LS) curves at two different bias points are parallel for several densities, and that the extracted m matches the low-bias percolation exponent (m≈1.8 near threshold, dropping to 1 at high density). They also get a square-law-like f(VG,VD) with β≈0.5. If that is robust, it's a nice compact model.

The soft spots are load-bearing. The justification for the factorization — that under the 1D Poisson approximation any random stick network reduces to a homogeneous thin-film strip of effective width W, with ξ measuring W independent of bias — is deferred to "details to be published elsewhere." That is exactly the step that makes the universal claim work. If W is bias-dependent near percolation threshold, the factorization collapses. The stress-test note is right that the parallel curves are necessary but not sufficient. You need to show that a device calibrated at one geometry actually predicts a different geometry's full I-V, not just that two ratios match. That direct test is absent. Also, m and β are fitted from the same simulations, so the "universal" scaling is not parameter-free. No error bars are given, though the data appear to be averaged over ~200 samples; the parameter range is limited (LC/LS up to 20, densities up to 20).

There is no experimental validation for long-channel devices. For a short conference-style paper this might be acceptable if the theory were established, but for the strength of the claim, it's thin. The authors are candid about the missing derivation, which is a point in their favor. This is not a deceptive paper; it's an overextended one.

Verdict: the numerical finding deserves a serious referee. I would send it to peer review, but I'd expect major revision: they need to provide the effective-width derivation (or at least a much more convincing argument), add an out-of-sample prediction test, and dial back the "single device predicts all" claim until it is demonstrated. I wouldn't cite it yet as a universal law, but I'd keep it in mind as evidence for bias-independent exponent.

Best.

Referee report

Major objections (4)

[Sec. III, Eq. (2)] The factorization ID = (A/LS) xi(LS/LC, rho_S L_S^2) f(VG,VD) is the central claim, but its justification for long channels is the statement, in the paragraph following Eq. (4), that under the 1D Poisson approximation any random stick network can be reduced to a single homogeneous thin-film strip of effective width W, with details deferred to 'to be published elsewhere.' This reduction is load-bearing: if W, and hence xi, depends on bias, the ratio rule IA/IA' = IB/IB' fails and Eq. (2) collapses. The manuscript provides no derivation and no direct numerical or experimental test of this reduction, so the central claim is currently unsupported.
[Sec. III, Eq. (3) and Fig. 2] The exponent m in Eq. (3) is not predicted or derived; the values m = 1.96, 1.7, 1.4, 1.2, 1.1 for rho_S L_S^2 = 5.4, 6, 8, 12, 20 are extracted from the same simulated I-V curves used to verify the scaling law. At rho_S L_S^2 = 5.4 the low-bias and high-bias panels report m = 1.96 and m = 1.9, respectively, which conflicts with the text's claim that the high-bias exponents are 'exactly the same' as the low-bias ones and underlines the absence of error bars. A scaling law whose key exponent is fitted from the validating data is not the parameter-free universal prediction advertised in the abstract.
[Sec. III, final paragraph; Abstract] The claim that measuring the full I-V characteristics of a single NB-TFT is sufficient to predict any other device is not supported by the stated procedure. The recipe requires m as a function of rho_S L_S^2 from Fig. 2(b), which is a family of simulations at many channel lengths and densities, not information available from a single device. The single-device rule would require an independently established universal curve m(rho_S L_S^2), which the paper does not provide.
[Fig. 2 and Sec. III] The numerical evidence for Eq. (2) consists of parallel log-log curves at two bias voltages for five densities near or above the percolation threshold. This is necessary but not sufficient for factorization, and no direct cross-prediction test is shown, such as using the I-V of one simulated network to predict the I-V of another network with different LC or rho_S. The paper also gives no error bars for the roughly 200 sample averages, no long-channel experimental data, and no deeply sub-threshold networks, despite the claim that Eq. (2) holds for arbitrary stick density above and below the percolation threshold.

Minor objections (6)

[Abstract and Introduction] The text contains a typo, 'ver last several years', and Eq. (1) is garbled in the typeset version; the governing equations should be rewritten with clear notation.
[Sec. III] The percolation threshold expression '4.236 2/pi' should read 4.236^2/pi (or an explicit numerical value) and should be checked against Ref. [18].
[Sec. III and Fig. 2] The text says the exponent m near threshold is about 1.8, while Fig. 2 lists 1.96 and 1.9 for rho_S L_S^2 = 5.4; these values should be reconciled and reported with precision or error estimates.
[Sec. III] The phrase 'the conductance exponent m (ID ~ 1/LC^m) ... is exactly the same' is misleading because m is not a single constant but varies with density; it should be written as m(rho_S L_S^2).
[References] Reference [12] is cited as a conference paper without page numbers, and Ref. [11] is given as 'In Press'; these should be updated to final publication data.
[Fig. 1] Fig. 1 would be more informative if it showed simulated conductance maps, since the contrast between Figs. 1(b) and 1(c) is the physical motivation for the paper.

How Pith checks papers → · Full audit of this paper →

The paper

The paper claims

This paper claims that a universal factorized current formula holds for long-channel nanobundle thin-film transistors, so a single calibrated device predicts the I-V behavior of any other device made from the same random stick network.

The reading

The paper claims that the drain current of a long-channel nanobundle thin-film transistor (NB-TFT), a transistor whose channel is filled with a random network of carbon nanotubes or nanowires, obeys a universal factorized formula: the current separates into a technology constant, a purely geometrical factor, and a purely bias-dependent factor. The bias factor has the familiar power-law form $(V_G-V_T)^\beta - (V_G-V_T-V_D)^\beta$, while the geometry factor depends only on stick length, channel length, and stick density. Because the two factors are independent, the ratio of currents at two different biases is the same for every geometry, so that measuring the full current-voltage curve of one transistor predicts the behavior of any other transistor made from the same material. The authors verify this factorization by self-consistently solving drift-diffusion and Poisson equations for simulated random networks, covering low and high bias and densities above and below the percolation threshold.

What carries the argument

The key object is the factorization identity $I_D = (A/L_S)\, \xi(L_S/L_C,\rho_S L_S^2)\, f(V_G,V_D)$, with $\xi = (\rho_S L_S^2)(L_S/L_C)^m$ and $f = (V_G-V_T)^\beta - (V_G-V_T-V_D)^\beta$. It is produced numerically by solving the generalized drift-diffusion and Poisson equations (Eq. 1) for random stick networks, using the parabolic back-gate approximation and stick-stick interaction terms; the sticky point is the claim, stated without proof, that the network's geometric effect on current is captured entirely by $\xi$, the effective width of an equivalent homogeneous thin-film strip, independent of bias.

Extended reading notes

Core claim

In the long-channel limit $L_C > L_S$, where individual sticks cannot bridge the channel and stick-stick interactions matter, the drain current still factorizes as $I_D = (A/L_S)\; \xi(L_S/L_C, \rho_S L_S^2)\; f(V_G,V_D)$. The geometry factor takes the explicit form $\xi = (\rho_S L_S^2)\,(L_S/L_C)^m$, with the same percolation exponent $m$ as in the low-bias homogeneous regime: $m \sim 1$ for densities far above the percolation threshold and $m \sim 1.8$ near it. The bias factor is geometry-independent with $\beta \sim 0.5$, matching the classical square law at extreme densities. The authors argue that this robustness follows because any random stick network can be reduced to a single homogeneous strip of effective width $W$, with $\xi$ measuring that width; the factorization then converts a single calibrated device into a complete description of all devices with arbitrary geometry and operating conditions.

Load-bearing premise

The entire scaling rule rests on the unproved claim that any random stick network can be mapped to a single homogeneous thin-film strip of fixed effective width, so that the geometric factor separates cleanly from the bias factor even near the percolation threshold.

Editorial extensions

If this is right

A single set of I-V measurements on one NB-TFT gives the current of any other NB-TFT made from the same material, for any channel length, stick length, stick density, and bias.
The formula provides a bottom-up effective mobility $\mu_{\rm eff} \sim (dI_D/dV_G/V_D)(L_S/\xi)/(L W C_{OX})$ that is independent of channel length, letting different laboratories compare data on a common footing.
The model allows computing the highest operating frequency $f_{\max} = I/(C V)$ for an ideal device, giving about 1 GHz for a $1\,\mu$m channel, free from hysteresis and interface-trap artifacts.
The scaling law can be used to organize experimental data and shorten technology development and characterization time for network transistors.

Reading between the lines

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

If the factorization is as robust as claimed, the same one-device calibration should apply to other percolating conductors, such as metal nanowire or graphene networks used as sensors, so a single strain or bio response curve could predict the response of any network geometry.
The effective-width reduction suggests a renormalization-group relation between a random stick network and a homogeneous film; a testable consequence is that the noise spectrum or subthreshold swing of a network device would match a homogeneous device with the same effective width, which the paper does not analyze.
Near the percolation threshold, where $m \sim 1.8$, the network's backbone is fractal; the factorization's validity there may fail for finite networks whose lateral size is below the percolation correlation length, a regime not covered by the simulations.

Original abstract

By generalizing the classical linear response theory of stick percolation to nonlinear regime, we find that the drain current of a Nanobundle Thin Film Transistor (NB-TFT) is described under a rather general set of conditions by a universal scaling formula ID = A/LS g(LS/LC, rho_S * LS * LS) f(VG, VD), where A is a technology-specific constant, g is function of geometrical factors like stick length (LS), channel length (LC), and stick density (rho_S) and f is a function of drain (VD) and gate (VG) biasing conditions. This scaling formula implies that the measurement of full I-V characteristics of a single NB-TFT is sufficient to predict the performance characteristics of any other transistor with arbitrary geometrical parameters and biasing conditions.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13-20260917 via OpenRouter, question bank abstract-p1-20260923, 2026-09-23: 1 request, 2,500 input tokens, $0.0001 reported.

What the abstract itself states, as probabilities: claims first 0.06 · states limitation 0.11 · human participants 0.01 · controlled comparison 0.02 · randomized assignment 0.01 · states quantitative result 0.13 · compares with prior results 0.10 · mentions code or data availability 0.02.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; show every role.

By generalizing the classical linear response theory of stick percolation to nonlinear regime, we find that the drain current of a Nanobundle Thin Film Transistor (NB-TFT) is described under a rather general set of conditions by a universal scaling formula ID = A/LS g(LS/LC, rho_S * LS * LS) f(VG, VD), where A is a technology-specific constant, g is function of geometrical factors like stick length (LS), channel length (LC), and stick density (rho_S) and f is a function of drain (VD) and gate (VG) biasing conditions. This scaling formula implies that the measurement of full I-V characteristics of a single NB-TFT is sufficient to predict the performance characteristics of any other transistor with arbitrary geometrical parameters and biasing conditions.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	By generalizing the classical linear response theory of stick percolation to nonlinear regime, we find that the drain current of a Nanobundle Thin Film Transistor (NB-TFT) is described under a rather general set of conditions by a universal scaling formula ID = A/LS g(LS/LC, rho_S * LS * LS) f(VG, VD), where A is a technology-specific constant, g is function of geometrical factors like stick length (LS), channel length (LC), and stick density (rho_S) and f is a function of drain (VD) and gate (VG) biasing conditions.	result 0.95	0.93	0.56	asserted 1.00
s1	This scaling formula implies that the measurement of full I-V characteristics of a single NB-TFT is sufficient to predict the performance characteristics of any other transistor with arbitrary geometrical parameters and biasing conditions.	interpretation 1.00	0.64	0.75	asserted 0.97

What kind of paper the classifier reports this is: theoretical result 1.00 · evidence from proof or derivation 0.97.

What was examined, and what was not

sources read

arXiv listing abstract v1

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

Figures

**Figure 1.** (a) A schematic of NBT: Geometry of nanobundle network transistor. Distribution of normalized conductance along the sticks in the random network in (b) low and (c) high bias, respectively. The conductance is uniform along the channel in low bias while it varies nonlinearly in the high bias condition resulting in an inhomogeneous percolating network and the contours of conductance are not straight lines parallel to S/D as in conventional MOSFET

**Figure 2.** (a) IDS-VDS for 2 different random networks A and B, Drain current vs. channel length for network transistors with channel length, LC > stick length, LS for (b) low and (c) high bias, respectively, in strong coupling limits. The current exponents, m, for high bias regimes are exactly same as low bias regime, and Eq. (2) requires that the ratio of currents at two different biasing conditions is independent of geometrical parameters for any particular percolating random network, i.e. IA/IA’=IB/IB’

Figures reproduced from arXiv:cond-mat/0603033 by the authors.

References

Works cited (9)

[1]

Y. X. Zhou, A. Gaur, S. H. Hur, C. Kocabas, M. A. Meitl, M. Shim, and J. A. Rogers, "p-channel, n-channel thin film transistors and p-n diodes based on single wall carbon nanotube networks," Nano Lett. , vol. 4, pp. 2031- 2035, 2004

View on Pith 2004
[2]

W. J. Liang, M. Bockrath, D. Bozovic, J. H. Hafner, M. Tinkham, and H. Park, "Fabry-Perot interference in a nanotube electron waveguide," Nature, vol. 411, pp. 665-669, 2001

View on Pith 2001
[3]

E. S. Snow, J. P. Novak, P. M. Ca mpbell, and D. Park, "Random networks of carbon nanotubes as an electronic material," Appl. Phys. Lett. , vol. 82, pp. 2145-2147, 2003

View on Pith 2003
[4]

A. Javey, J. Guo, M. Paulsson, Q. Wang, D. Mann, M. Lundstrom, and H. J. Dai, "High-field quasiballistic transport in short carbon nanotubes," Phys. Rev. Lett., vol. 92, 2004

View on Pith 2004
[5]

X. F. Duan, C. M. Niu, V. Sahi, J. Chen, J. W. Parce, S. Empedocles, and J. L. Goldman, "High-performance thin-film transistors using semiconductor nanowires and nanoribbons," Nature, vol. 425, pp. 274-278, 2003

View on Pith 2003
[6]

E. S. Snow, J. P. Novak, M. D. Lay, E. H. Houser, F. K. Perkins, and P. M. Campbell, "Carbon nanotube networks: Nanomaterial for macroelectronic applications," J. of Vac. Sci. and Tech. B, vol. 22, pp. 1990-1994, 2004

View on Pith 1990
[7]

L. Hu, D. S. Hecht, and G. Gr uner, "Percolation in transparent and conducting carbon nanotube networks," Nano Lett., vol. 4, pp. 2513-2517, 2004

View on Pith 2004
[8]

E. Menard, K. J. Lee, D. Y. Kha ng, R. G. Nuzzo, and J. A. Rogers, "A printable form of silicon for high pe rformance thin film transistors on plastic substrates," Appl. Phys. Lett., vol. 84, pp. 5398-5400, 2004

View on Pith 2004
[9]

S. Kumar, J. Y. Murthy, and M. A. Alam, "Percolating conduction in finite nanotube networks," Phys. Rev. Lett., vol. 95, 2005. [10]O. Marinov, M. J. Deen, and B. Iniguez, "Charge transport in organic and polymer thin-film transistors: recent issues," Iee Proceedings-Circuits Devices and Systems, vol. 152, pp. 189-209, 2005. [11]S. Kumar, N. Pimparkar, J. ...

View on Pith Pith review arXiv 2005

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.

[1] [1]

Y. X. Zhou, A. Gaur, S. H. Hur, C. Kocabas, M. A. Meitl, M. Shim, and J. A. Rogers, "p-channel, n-channel thin film transistors and p-n diodes based on single wall carbon nanotube networks," Nano Lett. , vol. 4, pp. 2031- 2035, 2004

View on Pith 2004

[2] [2]

W. J. Liang, M. Bockrath, D. Bozovic, J. H. Hafner, M. Tinkham, and H. Park, "Fabry-Perot interference in a nanotube electron waveguide," Nature, vol. 411, pp. 665-669, 2001

View on Pith 2001

[3] [3]

E. S. Snow, J. P. Novak, P. M. Ca mpbell, and D. Park, "Random networks of carbon nanotubes as an electronic material," Appl. Phys. Lett. , vol. 82, pp. 2145-2147, 2003

View on Pith 2003

[4] [4]

A. Javey, J. Guo, M. Paulsson, Q. Wang, D. Mann, M. Lundstrom, and H. J. Dai, "High-field quasiballistic transport in short carbon nanotubes," Phys. Rev. Lett., vol. 92, 2004

View on Pith 2004

[5] [5]

X. F. Duan, C. M. Niu, V. Sahi, J. Chen, J. W. Parce, S. Empedocles, and J. L. Goldman, "High-performance thin-film transistors using semiconductor nanowires and nanoribbons," Nature, vol. 425, pp. 274-278, 2003

View on Pith 2003

[6] [6]

E. S. Snow, J. P. Novak, M. D. Lay, E. H. Houser, F. K. Perkins, and P. M. Campbell, "Carbon nanotube networks: Nanomaterial for macroelectronic applications," J. of Vac. Sci. and Tech. B, vol. 22, pp. 1990-1994, 2004

View on Pith 1990

[7] [7]

L. Hu, D. S. Hecht, and G. Gr uner, "Percolation in transparent and conducting carbon nanotube networks," Nano Lett., vol. 4, pp. 2513-2517, 2004

View on Pith 2004

[8] [8]

E. Menard, K. J. Lee, D. Y. Kha ng, R. G. Nuzzo, and J. A. Rogers, "A printable form of silicon for high pe rformance thin film transistors on plastic substrates," Appl. Phys. Lett., vol. 84, pp. 5398-5400, 2004

View on Pith 2004

[9] [9]

S. Kumar, J. Y. Murthy, and M. A. Alam, "Percolating conduction in finite nanotube networks," Phys. Rev. Lett., vol. 95, 2005. [10]O. Marinov, M. J. Deen, and B. Iniguez, "Charge transport in organic and polymer thin-film transistors: recent issues," Iee Proceedings-Circuits Devices and Systems, vol. 152, pp. 189-209, 2005. [11]S. Kumar, N. Pimparkar, J. ...

View on Pith Pith review arXiv 2005