Skip to content
Escape
Meridian

The future is still open.

Li, Xiangying · Nov 9, 2026

arXiv eess.SY: Li proposes near-junction cooling for power transistors

Markdown · JSON · Original source

Xiangying Li proposes a horizontal near-junction active cooling topology for RF power GaN HEMTs, whose hotspots the abstract says exceed 5 kW/cm^2. The design places a thermoelectric thin film beside the device drain on a diamond substrate. Li’s proposed arrangement absorbs heat at an interface while bulk Joule heat leaves vertically. In Li’s lateral model, the horizontal film has a recirculation factor gamma of about 0.03, compared with 1 for a vertical pillar. At equal resistance, the model raises maximum cooling from 13 mW to 500 mW, and the critical current from 0.65 A to 24 A. The abstract also describes -constant and linear stability analysis to define a working window and safety margins for microsecond pulse cooling. These are model results and a proposed design path; the supplied abstract does not describe an experimental demonstration.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

arXiv eess.SYArticle · arxiv.orgOpen original source

Original source

eess.SY · Article

Source details

Original source date: Nov 9, 2026 (2026-11-09).

Date basis: Date basis not provided.

Added to Escape Meridian: 2026-11-10T21:18:44.237Z. First observed: 2026-11-10T08:34:31.351Z. Retrieved: 2026-11-10T21:17:41.114Z.

Referenced by

Recorded citations and public file attachments. A reference alone does not establish agreement or influence.

No public references recorded here yet.

Reference data (JSON)

Limits

The supplied evidence is the abstract; it reports a model and analysis but does not describe an experimental demonstration. The abstract gives proposed performance figures and safety-margin analysis; it does not report measured results from an operating device.

Open question

How would measurements in an operating GaN HEMT compare with the model’s predicted cooling and critical current?

Source-stated

Li proposes horizontal near-junction active cooling for RF power GaN HEMTs, whose hotspots the abstract says exceed 5 kW/cm^2.

We study the hardest case, RF power GaN HEMTs, whose hotspots exceed 5 kW/cm^2, and propose a horizontal near-junction active cooling topology

Source-stated

The topology places a thermoelectric thin film beside the device drain on a diamond substrate, allowing bulk Joule heat to leave vertically.

a thermoelectric thin film lies flat on a diamond substrate beside the device drain (the electron exit), so bulk Joule heat leaves vertically into the substrate.

Source-stated

The lateral fin model gives the horizontal film a recirculation factor gamma of about 0.03, compared with 1 for the vertical pillar.

the vertical pillar yields gamma = 1, while the horizontal film yields gamma_max = sqrt(2) L_z/L_x, about 0.03 for L_z = 10 um, L_x = 500 um.

Source-stated

At equal resistance, the abstract reports model results of 500 mW maximum cooling and a 24 A critical current for the horizontal film, compared with 13 mW and 0.65 A.

At equal resistance, the same film lifts maximum cooling from 13 mW to 500 mW, and the critical current from 0.65 A to 24 A.

Source and contributor context

Li, Xiangying · Article author

These are credits stated by the source; they do not establish a profile match or authorship of every passage.

Source media

No suitable media was found by the supported page extraction.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

arXiv

arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.

Fin

Fin is Intercom's AI customer service product, whose site publishes guidance on support teams, organizational structure and careers.

TIME

TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.

Editorial provenance

Version 1. Publication ID: 9c288a32-a9c6-4b1f-844f-f58f25bf4854.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T21:18:44.237Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: material_development.

Observation: 486ceb3dbf8206f1e102e55914e8557ed3436de0db6c44c52aad6d30854cf9cf. Source text hash: 2998002e7dcbdc4febe860dd4102f8d8a78fcca0f678c4536f9da95121b0de52.

Revision and source-check history

Version 1 · 2026-11-10T21:18:44.237Z

arXiv eess.SY: Li proposes near-junction cooling for power transistors

Xiangying Li proposes a horizontal near-junction active cooling topology for RF power GaN HEMTs, whose hotspots the abstract says exceed 5 kW/cm^2. The design places a thermoelectric thin film beside the device drain on a diamond substrate. Li’s proposed arrangement absorbs heat at an interface while bulk Joule heat leaves vertically. In Li’s lateral model, the horizontal film has a recirculation factor gamma of about 0.03, compared with 1 for a vertical pillar. At equal resistance, the model raises maximum cooling from 13 mW to 500 mW, and the critical current from 0.65 A to 24 A. The abstract also describes -constant and linear stability analysis to define a working window and safety margins for microsecond pulse cooling. These are model results and a proposed design path; the supplied abstract does not describe an experimental demonstration.

Original source

eess.SY · Article

Source details

Original source date: Nov 9, 2026 (2026-11-09).

Date basis: Date basis not provided.

Added to Escape Meridian: 2026-11-10T21:18:44.237Z. First observed: 2026-11-10T08:34:31.351Z. Retrieved: 2026-11-10T21:17:41.114Z.

Limits

The supplied evidence is the abstract; it reports a model and analysis but does not describe an experimental demonstration. The abstract gives proposed performance figures and safety-margin analysis; it does not report measured results from an operating device.

Open question

How would measurements in an operating GaN HEMT compare with the model’s predicted cooling and critical current?

Source-stated

Li proposes horizontal near-junction active cooling for RF power GaN HEMTs, whose hotspots the abstract says exceed 5 kW/cm^2.

We study the hardest case, RF power GaN HEMTs, whose hotspots exceed 5 kW/cm^2, and propose a horizontal near-junction active cooling topology

Source-stated

The topology places a thermoelectric thin film beside the device drain on a diamond substrate, allowing bulk Joule heat to leave vertically.

a thermoelectric thin film lies flat on a diamond substrate beside the device drain (the electron exit), so bulk Joule heat leaves vertically into the substrate.

Source-stated

The lateral fin model gives the horizontal film a recirculation factor gamma of about 0.03, compared with 1 for the vertical pillar.

the vertical pillar yields gamma = 1, while the horizontal film yields gamma_max = sqrt(2) L_z/L_x, about 0.03 for L_z = 10 um, L_x = 500 um.

Source-stated

At equal resistance, the abstract reports model results of 500 mW maximum cooling and a 24 A critical current for the horizontal film, compared with 13 mW and 0.65 A.

At equal resistance, the same film lifts maximum cooling from 13 mW to 500 mW, and the critical current from 0.65 A to 24 A.

Source and contributor context

Li, Xiangying · Article author

These are credits stated by the source; they do not establish a profile match or authorship of every passage.

Source media

No suitable media was found by the supported page extraction.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

arXiv

arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.

Fin

Fin is Intercom's AI customer service product, whose site publishes guidance on support teams, organizational structure and careers.

TIME

TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.

Editorial provenance

Version 1. Publication ID: 9c288a32-a9c6-4b1f-844f-f58f25bf4854.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T21:18:44.237Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: material_development.

Observation: 486ceb3dbf8206f1e102e55914e8557ed3436de0db6c44c52aad6d30854cf9cf. Source text hash: 2998002e7dcbdc4febe860dd4102f8d8a78fcca0f678c4536f9da95121b0de52.

uncertain · 2026-11-10T21:18:44.237Z · anthropic/claude-opus-5.5

Publication: 9c288a32-a9c6-4b1f-844f-f58f25bf4854

attribution — pass: The draft correctly attributes the paper to Xiangying Li on arXiv eess.SY and keeps the figures framed as Li's proposal and model results rather than established facts.

Authors:Xiangying Li

summary — uncertain: Central jargon is unexplained for a general reader: RF power GaN HEMTs, near-junction, Joule heat, recirculation factor gamma, and critical current (the source calls it critical runaway current). The draft also omits the abstract's framing of why this matters, that chips are limited because cores make heat faster than it can be removed. It drops the core argument too: Peltier cooling grows with current but Joule heating grows with its square, and a vertical pillar sends about half the Joule heat back. Smallest repair: add one plain-language sentence each on the bottleneck, the Peltier-versus-Joule contest, and what gamma and runaway current mean.

Chip performance is limited by a shared bottleneck: the core produces heat faster than it can be moved out.

date — pass: originalPublished is 2026-11-09. This matches the arXiv v1 submission date and is not the later feed date.

[Submitted on 9 Nov 2026]

support — uncertain: The claims are source-stated, but the gamma figure loses its conditions. The source gives gamma_max, an upper bound, of about 0.03 only for L_z = 10 um and L_x = 500 um. The draft presents it as the film's gamma without those dimensions. Repair: say the maximum gamma is about 0.03 for the stated film dimensions. The limitations, which note no experimental demonstration, and the question are appropriately supported.

the horizontal film yields gamma_max = sqrt(2) L_z/L_x, about 0.03 for L_z = 10 um, L_x = 500 um.

media — pass: The selected media is the verified direct arXiv abstract page for this paper, which is the most direct access to the source.

Challenge or correct this reading

Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.