Electroluminescent Cooling in III-V Intracavity Diodes: Efficiency Bottlenecks

Toufik Sadi*, Ivan Radevici, Pyry Kivisaari, Jani Oksanen

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

13 Citations (Scopus)


Recent advances in the photoluminescent cooling of doped glasses provoke the question of whether similar progress is possible in electroluminescent cooling (ELC), and if so, what are the conditions for observing it at high powers. Here, we establish a simulation framework for III-V intracavity double-diode structures (DDSs) intended for studying ELC and introduce and analyze the most relevant figures of merit for the recently measured devices exhibiting the highest reported quantum efficiency of 70%. In essence, the DDSs optically couple a GaInP/GaAs double heterojunction light-emitting diode (LED) and a GaAs p-n homojunction photodetector (PD), integrated as a single device. The modeling framework couples the drift-diffusion charge transport model with a photon transport model and uses our recent experimental measurements for validation and the extraction of important material parameters. Results show that the model can accurately describe the experimental behavior over many orders of magnitude and suggest that the internal efficiency of the LED already exceeds the cooling threshold. Directly observing cooling in the presently studied devices, however, is still hindered by bottlenecks arising from the surface recombination at the LED walls and recombination losses in the PD.

Original languageEnglish
Article number8693895
Pages (from-to)2651-2656
Number of pages6
JournalIEEE Transactions on Electron Devices
Issue number6
Publication statusPublished - 1 Jun 2019
MoE publication typeA1 Journal article-refereed


  • double-diode structures
  • electroluminescent cooling
  • III-As
  • light-emitting diodes (LEDs)


Dive into the research topics of 'Electroluminescent Cooling in III-V Intracavity Diodes: Efficiency Bottlenecks'. Together they form a unique fingerprint.

Cite this