Can one graphene detector be sensitive, fast and broadband at once?
Narrow graphene stripes flanked by gold patches let one detector respond strongly from visible light to the far infrared while staying extremely fast.
Source
Gold-patched graphene nano-stripes for high-responsivity and ultrafast photodetection from the visible to infrared regime
Study at a glance
- Design
- Other — Single fabricated device characterised with lasers and a filtered infrared source across wavelength, gate voltage, bias and modulation frequency, plus FDTD simulations.
- N
- Device-physics experiment; no sample of units — results come from one fabricated photodetector design.
- Population
- Monolayer CVD graphene nano-stripes with gold patches on high-resistivity silicon/SiO2
- Outcome
- Photodetector responsivity across 0.8 to 20 micrometre wavelengths, gate dependence, noise-equivalent power and high-frequency response
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The authors patterned monolayer graphene into nano-stripes narrower than the metal-graphene junction regions and attached arrays of gold patches that funnel light onto the graphene. They simulated the light concentration, then measured responsivity with a supercontinuum laser (visible/near-infrared) and a Globar source with bandpass filters (mid/far infrared), varying gate voltage, bias and optical power. Speed was tested by beating two lasers at 783 and 785 nm to create modulation up to tens of gigahertz.
What they found
Responsivity ranged from 0.6 A/W at 0.8 micrometres up to 11.5 A/W at 20 micrometres with an optimised gate voltage, and the response showed no roll-off up to 50 GHz, the limit of the test equipment. Gate voltage changed responsivity in the way band-structure physics predicts: Pauli blocking reduced short-wavelength response at lower Fermi energies, while more available states raised infrared response. Photoconductive gain was higher at short wavelengths, consistent with carrier multiplication.
The limits
What it doesn't show
The results come from one device design, and the paper does not report how much performance varies across many fabricated devices. The detector has a relatively large dark current because it is photoconductive, and its low noise figure assumes chopped, lock-in detection above 1 kHz rather than measuring noise directly in all conditions. The upper speed and wavelength limits were set by the test equipment, so the true cutoff (predicted, not measured, near 425 GHz) remains unverified. Responsivity also falls at higher optical powers.
Key terms
- Responsivity
- Output photocurrent per watt of incident light (A/W); a key measure of detector sensitivity.
- Photoconductive gain
- The number of charge carriers collected per absorbed photon; above one when carriers are multiplied or recirculate before recombining.
- Pauli blocking
- Absorption is suppressed when the electron states a transition would fill are already occupied, so low-energy photons cannot be absorbed below twice the Fermi energy.
- Interband vs intraband absorption
- Interband absorption moves an electron between valence and conduction bands; intraband (free-carrier) absorption moves it within one band and dominates graphene's infrared response.
- Noise-equivalent power (NEP)
- The optical power that gives a signal equal to the noise in a 1 Hz bandwidth; lower is better.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
What mainly gives this detector its high photoconductive gain?
Common questions
Why doesn't graphene's short carrier lifetime ruin the sensitivity here?
The stripes are so narrow that carriers reach the gold patches faster than they recombine, so gain comes from a short transit time rather than a long lifetime.
Why does changing the gate voltage matter?
The gate shifts graphene's Fermi level, which changes which optical transitions are allowed (Pauli blocking) and how many states are available for infrared absorption.
Is 50 GHz the detector's actual speed limit?
No; it is the limit of the probes and spectrum analyser used. The electrical model predicts a higher cutoff, but that was not directly measured.
More on Two-dimensional materials