Record breaking photo detector captures light in just 125 picoseconds

Mar 07, 2026 Last reply: 4 months ago 4 Replies

From:

formatting link
photodetector captures light in just 125 picoseconds Record-breaking ultrathin photodetector captures light across the entire spectrum in just 125 picoseconds, opening the door to powerful new imaging technologies. Date: March 4, 2026 Source: Duke University Summary: A new ultrathin photodetector from Duke University can sense light across the entire electromagnetic spectrum and generate a signal in just 125 picoseconds, making it the fastest pyroelectric detector ever built. The breakthrough could power next-generation multispectral cameras used in medicine, agriculture, and space-based sensing.



Is this real?


Slow as molasses. I had <20 ps edges from my antenna-coupled tunnel junctions 20 years ago.

Plus it seems to be a narrowband pyroelectric, which has usability problems due to being intrinsically AC coupled.

Fun project though.

Cheers

Phil Hobbs

Phil Hobbs snipped-for-privacy@electrooptical.net wrote: |-------------------------------------------------------------------| |"Jan Panteltje snipped-for-privacy@comet.invalid wrote: | |> From: | |>

formatting link
| |[. . .] | |> [. . .] space-based sensing. | |> | |[. . .] | | | |Slow as molasses. I had <20 ps edges from my antenna-coupled tunnel| |junctions 20 years ago. " | |-------------------------------------------------------------------|

Dear Doctor Hobbs, Bah! I work on an area of gamma-ray astronomy with a proposed standard measurment taking some 10**6 seconds (i.e. nearly

12 days)!

Cheers :) (S.

formatting link
fuer Kontaktdaten!)

That’s even longer than it took me to document the sub-20 ps response, which was limited by my Tek 11801C sampling scope. I was using a lamp-pumped pulsed Nd:YAG, 20 ps @20 Hz, driving a seeded optical parametric generator and difference frequency generator, which was continuously tunable from 400 nm to 10 microns with a couple of narrow gaps. A beautiful thing, but all that stuff going on meant that it had a pulse to pulse variation of about plus or -10%.

Sampling scopes give you one point per trigger, so one trace took 25.6 seconds, and averaging 100 traces to control the noise took that up to over

40 minutes per trace.

So I feel your pain. ;)

Cheers

Phil Hobbs

I was just going to say, talk to Phil, he can do that a lot faster :-)

[...]

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required