ARTICLE

https://doi.org/10.1038/s41467-022-31607-7 OPEN

Waveguide-integrated mid-infrared photodetection

using graphene on a scalable chalcogenide glass
platform
Jordan Goldstein 1, Hongtao Lin1,3, Skylar Deckoff-Jones1, Marek Hempel1, Ang-Yu Lu 1,

Kathleen A. Richardson2, Tomás Palacios1, Jing Kong 1, Juejun Hu 1 & Dirk Englund 1✉
1234567890():,;

The development of compact and fieldable mid-infrared (mid-IR) spectroscopy devices

represents a critical challenge for distributed sensing with applications from gas leak
detection to environmental monitoring. Recent work has focused on mid-IR photonic inte-
grated circuit (PIC) sensing platforms and waveguide-integrated mid-IR light sources and
detectors based on semiconductors such as PbTe, black phosphorus and tellurene. However,
material bandgaps and reliance on SiO2 substrates limit operation to wavelengths λ ≲ 4 μm.
Here we overcome these challenges with a chalcogenide glass-on-CaF2 PIC architecture
incorporating split-gate photothermoelectric graphene photodetectors. Our design extends
operation to λ = 5.2 μm with a Johnson noise-limited noise-equivalent power of 1.1 nW/Hz1/2,
no fall-off in photoresponse up to f = 1 MHz, and a predicted 3-dB bandwidth of f3dB > 1 GHz.
This mid-IR PIC platform readily extends to longer wavelengths and opens the door to
applications from distributed gas sensing and portable dual comb spectroscopy to weather-
resilient free space optical communications.

1 Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, USA. 2 University of Central Florida, 4000 Central Florida

Boulevard, Orlando, FL 32816, USA. 3Present address: State Key Laboratory of Modern Optical Instrumentation, College of Information Science and Electronic
Engineering, Zhejiang University, 310027 Hangzhou, China. ✉email: [email protected]

NATURE COMMUNICATIONS | (2022)13:3915 | https://doi.org/10.1038/s41467-022-31607-7 | www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31607-7

M
id-IR absorption spectroscopy is a critical tool for Supplementary Fig. 1a depicts this beam-path in more detail.
chemical sensing and analysis, especially for inert gases We operate the device under zero-bias voltage to avoid
that evade detection by chemical reaction-based sen- introducing electronic shot noise and to prevent channel
sors. Many such gases derive their inertness from halogenated conductivity fluctuations from manifesting as 1/f noise15. For
chemistries and thus exhibit global warming potential due to the following low-frequency measurements we use a lock-in
carbon-halogen stretching modes resonant in the thermal IR1,2. amplifier to measure the photovoltage directly with no
To facilitate sensor deployment for greenhouse gas leak detection preamplification.
and other chemical sensor application areas, there exists a strong Figure 2a, b, and c plot the photovoltage, resistance, and
need to transition from co-packaged discrete components to transmission lock-in signals versus both gate voltages for one
compact and chip-integrated sensors. such photodetector (“Device A”). Here, we modulate the
To address this challenge, mid-IR photonic integrated cir- λ = 5.2 μm QCL source at 3.78 kHz with a guided “on” power
cuit (PIC) platforms have been investigated to reduce optical gas of 11 μW at the detector input. From our photovoltage and
sensors to the size of a chip. Recent work has demonstrated resistance maps, alongside the power and waveguide loss
integrated optical methane3 and volatile organic compound4 calibrations described in Supplementary Note 1, we infer the
sensing, but required coupling to off-chip sources and detectors. gate voltage pairs that optimize the voltage responsivity, current
However, integrating the detector on-chip is more compact and responsivity, and NEP with respect to Johnson noise, indicated
can improve sensitivity by reducing the volume of active material with green markers in Fig. 2. For these, we arrive at 1.5 V/W,
able to generate thermal noise. Su et al. achieved integration of a 10. mA/W, and 1.1 nW/Hz1/2, respectively. The observed photo-
PbTe photoconductor and demonstrated methane sensing at a voltage gate map indicates a PTE response mechanism, evidenced
wavelength of λ = 3.31 μm5, but their platform is limited to by the six-fold sign change pattern around the graphene channel’s
λ ≲ 4 μm due to absorption in the SiO2 substrate6 and by PbTe’s charge neutral point11. Figure 2d, e, and f show line slices of the
absorption cutoff7. Waveguide-integrated detectors based on voltage maps as indicated by the dashed lines in Fig. 2a, b, and c
narrow-gap 2D materials black phosphorus8 and tellurene9 have of the same color. Figure 2d, in particular, highlights the changes
also been demonstrated, but they too are bandgap-limited to in slope associated with PTE-based detectors11.
λ ≲ 4 μm.
Here we exceed the wavelength limit of previous demonstra- Photothermoelectric device model. To confirm our under-
tions using graphene-based detectors on an extended- standing of device operation and elucidate the prospects for
transparency waveguide platform. While graphene integrated performance improvement, we apply the formalism introduced in
detectors have shown promise at telecom wavelengths10, the Song et al.12 to calculate the electronic temperature distribution
material’s advantages are magnified further at longer wavelengths and Seebeck photovoltage in the graphene channel under illu-
due to the thermal nature of the photothermoelectric (PTE) mination. Figure 3a, b compare our measured and modeled
response mechanism11,12 and due to the impact of optical plas- voltage responsivities using calculations described in the Methods
mon scattering at short wavelengths13. Integrated photodetection section. The performance of our device depends on several fitting
with graphene has been demonstrated at wavelengths up to parameters, whose definitions and approximate values (derived
3.8 μm6 and with chalcogenide glass waveguides14, but on SiO2 from our measured data) we provide in Table 1. We describe our
platforms. To access longer wavelength operation and achieve fitting process in Supplementary Note 3. Critically, all features of
good sensitivity at zero-bias, we introduce a Ge28Sb12Se60 (GSSe)- the modeled responsivity map in Fig. 3b up to an overall scale
on-CaF2 waveguide platform supporting gated PTE-based gra- factor from τeph are established a priori from fitting parameters
phene photodetectors. These key changes allow us to extend extracted from the device transmittance and resistance maps,
operation to a wavelength of λ = 5.2 μm while achieving a with only τeph obtained by matching the scales of the measured
Johnson noise-limited noise-equivalent power (NEP) of 1.1 nW/ and modeled responsivities. The resemblance between Fig. 3a and
Hz1/2. By comparing the gate voltage maps of our device’s b thus reflects the validity of our PTE model and is not due to
resistance, transmittance, and responsivity with a photothermo- over-fitting. In Fig. 3c, we plot the solution to Eqn. (6), ΔTel(x), as
electric model, we extract material quality parameters of the _
well as the source term QðxÞ. The thermal transport model pre-
graphene channel, revealing a path to further reduce the device’s
dicts that 9 μW of guided power raises the temperature of the
NEP by shrinking the optical mode size in tandem with the
graphene channel’s electron gas by as much as 1 K along the
graphene channel.
center of the device.

Results Device bandwidth and noise performance. Current modulation

Device design and responsivity measurement. Figure 1a, b of our QCL source permits frequency response measurements up
illustrate the platform and photodetector design. The device to its modulation bandwidth of 1 MHz. To account for the
consists of a single-mode GSSe waveguide fabricated on top of a modulation response of our laser, we measure the photovoltage of
5.4 μm wide by 300 μm long, CVD-grown graphene channel, Device A alongside that of a fast InAsSb photodiode. The com-
flanked on either side by source and drain contacts placed far parison shown in Fig. 4a indicates that our device is faster than
enough away from the optical mode to avoid excess loss. Beneath our laser’s modulation bandwidth. We thus use a COMSOL
the graphene channel are pair of CVD graphene back-gates, model to find the actual RC contribution to our device’s fre-
separated by a 400 nm gap and used to electrostatically induce a quency response, plotted in the inset of Fig. 4a. We also plot the
pn-junction along the center of the channel. We use HfO2 as the product of the RC-limited frequency response and the τeph-lim-
0:5
gate dielectric and as an airtight capping layer. The device is ited frequency response with an assumed ð1 þ ð2πτ eph f Þ2 Þ
fabricated on a CaF2 substrate, transparent up to λ = 8 μm. Fig- dependence, which applies as the electron-phonon cooling length
ure 1c depicts the resulting waveguide mode at λ = 5.2 μm. qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
We use lock-in measurement to characterize our detectors, ‘ ¼ κτ eph =Cel  230 nm is narrower than our device
focusing light from a λ = 5.2 μm QCL source into our chip’s input channel12. We thus predict a 3-dB cutoff frequency of f−3dB ≈
facet. Light exiting the chip is focused onto an InAsSb 1.3 GHz, dominated by the capacitance between the graphene
photodetector and amplified for transmission measurement. back-gates.

2 NATURE COMMUNICATIONS | (2022)13:3915 | https://doi.org/10.1038/s41467-022-31607-7 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31607-7 ARTICLE

a Ge28Sb12Se60
Capping dielectric
waveguide
Drain contact
Graphene channel

Source contact Gate dielectric

Right gate
Left gate
contact
contact
y
Graphene
x
CaF2 substrate backgates
z

b Source Gates c
Waveguide Mode |E|2
1.00

1
0.75

y [μm]
0.50
0
0.25

100 μm
0.00
-2 0 2
x [μm]
Waveguide Drain

Fig. 1 Device geometry. a Illustration of the device cross-section perpendicular to the waveguide axis. The optical mode supported by the GSSe waveguide
evanscently couples to and is absorbed by the graphene channel, which is gated by two graphene back-gates to induce a pn-junction. b Optical image of the
device depicting source, drain and gate contact pads. c Depiction of the optical guided mode at λ = 5.2 μm.

Fig. 2 Gate voltage maps. a Measured zero-bias photovoltage produced by the device as a function of the two gate voltages. b Total device resistance as a
function of the two gate voltages. c Lock-in signal reflecting power measured by an InAsSb photodetector at the focal point of our output facet collection
lens, used to monitor transmission of the device as a function of the gate voltages. The star, triangle, and cross symbols on each gate voltage map
represent the optimum operating points for maximum voltage responsivity, maximum current responsivity, and minimum NEP, respectively. The power-
normalized transmittance is plotted in Supplementary Fig. 3b. d, e, f Plots of line sections indicated with dashed lines in panels a, b, and c, respectively.

NATURE COMMUNICATIONS | (2022)13:3915 | https://doi.org/10.1038/s41467-022-31607-7 | www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31607-7

a Measured Responsivity (V/W) Table 1 Device parameters and approximate values.

6

τDC Drude scattering time at DC ≈3.5 fs

4 τIR Drude scattering time at IR 30–50 fs
σn Standard deviation of native carrier 1.5–2.5 × 1012 cm−2
concentration due to spatial inhomogeneity
2 EFc Native Fermi level of graphene channel ≈ 0.17 eV
Gate Voltage 2 (V)

EFg Native Fermi level of graphene gates ≈ 0.48 eV

0
τeph Electron-phonon cooling time ≈ 50 ps
αe Excess light attenuation within device 2–3 mm−1

-2

a
-4
VG1 = -2V
VG2 = 3V

-6

-6 -4 -2 0 2 4 6
Gate Voltage 1 (V)
b
Modelled Responsivity (V/W)
6

4
Gate Voltage 2 (V)

b
0

-2

-4

-6

-6 -4 -2 0 2 4 6
Gate Voltage 1 (V)

c .
Modelled ΔTel and Q per guided power

100 2

75 1.5 Fig. 4 Bandwidth and noise properties. a Comparison of the frequency

ΔTel/P (K/mW)

Q/P (105 cm-2)

response of our photodetector with that of the laser current modulation

50 1
itself. The consistency between the two indicates that the photodetector
frequency response exceeds 1 MHz. Inset: Simulated GHz-range
photodetector frequency response, with and without considering the
.

25 0.5
impact of the electron-phonon cooling time τeph. b Measured noise spectral
density versus resistance and corresponding Johnson noise spectral density
0 0
of Device B, without illumination, for the 49 pairs of gate voltages {Vg1, Vg2}
-2 -1 0 1 2 where each Vgn is varied from −6 V to 6 V in steps of 2 V. Measurement
x (μm)
was performed at T = 293 K.
Fig. 3 Experiment/model comparison. a, b Contour plots of the a measured
and b modeled responsivity maps of our device, evaluated with τDC = 3.5 fs, To investigate our device’s noise performance, we modulate the
τIR = 40 fs, σn = 2 × 1012 cm−2, τeph = 50 ps, and αe = 2.5 mm−1. c Electron QCL current at 30 kHz, amplify the photovoltage with a low-
temperature increase ΔTel and absorbed optical power per area Q_ profiles in noise preamplifier and inspect using a signal analyzer. As shown
the graphene channel per guided optical power at gate voltages of in Supplementary Fig. 8, we observe in Device A no broadening
{−2.35 V, 0.35 V}, chosen to maximize the modeled photoresponse, and of the 30 kHz photoresponse peak at offset frequencies as low as
other parameters as above. 0.1 Hz, indicating long-term responsivity stability. We then
measure the un-illuminated noise spectral density and resistance
versus both gate voltages. Figure 4b shows the resulting data for a

4 NATURE COMMUNICATIONS | (2022)13:3915 | https://doi.org/10.1038/s41467-022-31607-7 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31607-7 ARTICLE

Table 2 Comparison of our detector with inferred room-temperature performance metrics for two HgCdTe photodiodes
optimized for two different wavelengths (from ref. 31) and a VOx bolometer (from ref. 32) available off the shelf. For the
photodiodes, the NEP is extrapolated from the specified detectivity for a detector scaled down to match the size of a diffraction-
limited spot with NA = 0.3, which is the acceptance NA of these detectors. For the bolometer, we give the NEP of a single
17 × 17 μm bolometer pixel as calculated from the specified noise-equivalent temperature difference as described in Rogalski7.

HgCdTe PD HgCdTe PD VOx bolometer This work

λopt = 5.0 μm λopt = 10.6 μm
pffiffiffiffiffiffi
NEP ½pW= Hz 1, λ = 5.2 μm 10, λ = 5.2 μm 0.9 1100
0.2, λ = 5.0 μm 40, λ = 10.6 μm
f−3dB [MHz] 1.3 106 10 Hz 1300 (pred.)
Vacuum required? No No Yes No
Waveguide-integrated? No No No Yes

Device B of identical design to Device A, organized by resistance waveguide platform with an NEP of 1.1 nW/Hz1/2 and a band-
and compared to the expected Johnson noise spectral density. We width exceeding f−3dB = 1 MHz. We have modeled the bandwidth
observe excellent consistency between the measured and to approach 1.3 GHz and we predict similar performance at
predicted noise, with a 2 − 4 dB discrepancy consistent with the longer wavelengths for scaled-up devices enabled by the trans-
specified noise figure of our preamplifier, corroborating our parency of GSSe beyond λ = 10 μm21. Finally, we have shown that
earlier claim of Johnson-noise-limited NEP. our device and waveguide platform would enable NO detection at
To demonstrate our device’s utility, we analyze its predicted concentrations comparable to its REL. Substantial improvements
gas-sensing performance, summarized from Supplementary are likely using metal-insulator-metal10 or dielectric slot wave-
Note 6. The minimum detectable gas concentration for a given guides to concentrate the optical mode to within a cooling length
waveguide platform and photodetector is given by16: of the pn-junction, which would also increase the attenuation of
α NEP the guided mode and thus decrease the device footprint needed to
pgas;min ¼ e base ; ð1Þ absorb an optical signal. Gapped bilayer graphene may also be
ang ΓE I 0
investigated as an alternative to monolayer graphene to reduce
where I0 is the source power, αbase is the waveguide attenuation thermal noise22. The PIC platform further promises to support a
coefficient in the absence of gas, a is the specific attenuation full toolkit of mid-IR active devices including on-chip quantum
coefficient of the gas, ng is the guided mode group index, ΓE is the cascade light sources23, and may even leverage the same graphene
confinement factor of electric field energy within the gaseous material platform for devices such as graphene modulators14 and
medium, and e ¼ expð1Þ. For detection of nitric oxide (NO), with hot-electron-based24 or gapped bilayer graphene light sources.
an absorption peak at λ = 5.24 μm and a specific attenuation of The platform could also be adapted to alternative mid-IR wave-
approximately a ≈ 70 m−1atmp −1 at low concentrations17, we
ffiffiffiffiffiffi
guide approaches, such as suspended Ge, as necessary to reach
arrive at pgas;min ¼ 74 μatm= Hz for a 1 mW illumination longer wavelength ranges25. Chalcogenide glass could then sup-
source. Assuming a measurement bandwidth of 0.1 Hz over plement such a platform by enabling designs where the graphene
which we have measured our photoresponse to be stable, we find channel is sandwiched between the Ge and high-index glass to
pgas;min ¼ 23 ppm, roughly equal to the National Institute of increase overlap with the optical mode. This research represents
the first foray into waveguide-integrated detectors operating
Occupational Safety and Health (NIOSH) recommended expo- beyond λ = 4 μm, paving the way towards 2D-material-enabled
sure limit (REL) of 25 ppm18. Removing the slightly lossy HfO2 integrated mid-IR microsystems for gas sensing, spectroscopy20
dielectric underneath the gas-light interaction waveguide could and free-space optical communications26.
decrease pgas;min considerably, as waveguide losses down to
0.7 dB/cm have been demonstrated at the same wavelength using
a similar chalcogenide glass and liftoff process19. Methods
Photodetector fabrication. A continuous monolayer graphene film was grown on
Cu foil (99.8%, Alfa Aesar, annealed, uncoated, item no. 46365) cut to a size of
Discussion 15 × 2 cm2 in a 1-inch-diameter quartz tube furnace under atmospheric pressure.
Although our demonstration is limited to λ = 5.2 μm by light The furnace was heated to 1060 ∘C over 30 min under 500 sccm of Ar flow;
source availability, the optical conductivity of our graphene afterwards, 15 sccm of H2 and 10 sccm of dilute CH4 (1% in Ar) were introduced as
reducing gas and carbon source, respectively, and flowed for 4 h to ensure the
inferred from the fitting parameters in Table 1 remains relatively continuity of the graphene film. Finally, the furnace was allowed to cool to 100 ∘C
constant and even increases at longer wavelengths due to intra- without modifying the gas flow before the CVD graphene was removed from the
band absorption as shown in Supplementary Fig. 7. We thus chamber. Our devices were fabricated on a 1" diameter by 1.0 mm thick (111)-cut
expect our platform to scale to λ = 10 μm and beyond, perhaps CaF2 substrate (MTI Corporation, item CFc25D10C2). We first coated our sub-
requiring a BaF2 substrate for extended transparency, with little strate with a PMMA bilayer for liftoff (495 PMMA A6 followed by 950 PMMA
A2), which features a slightly re-entrant sidewall profile after developing. We then
reduction in performance owing to the PTE effect’s thermal performed e-beam lithography using an Elionix FLS-125 125 keV electron beam
nature. In Table 2 we compare our device’s performance with lithography system to pattern alignment marks on our substrate, followed by
various off-the-shelf detectors. Although its NEP is not yet on par room-temperature development in 3:1 isopropanol:methyl isobutyl ketone for 90 s
with commercial options, its predicted bandwidth may be useful and isopropanol rinse for 120 s (“development process”), e-beam evaporation of
5 nm Ti/100 nm Au (Temescal VES2550) (“metal evaporation process”), and liftoff
for dual-comb spectroscopy-based integrated gas analyzers20. using a 4-h room-temperature acetone soak (“liftoff process”). To transfer the first
Additionally, the vacuum requirement of VOx bolometers may layer of graphene, we first coated one side of the CVD graphene-on-Cu sheet with
complicate co-packaging and introduce coupling losses, and the PMMA and removed the graphene from the other side using 90 s of oxygen RIE
high cost of HgCdTe may preclude use in broadly deployed (16 sccm He and 8 sccm O2 at a pressure of 10 mTorr and an RF power of 100W,
“oxygen RIE process”). We then etched away the Cu using a FeCl3-based etchant,
sensor networks. followed by 2 DI water rinses, a 30-min clean in 5:1 DI water:HCl 37% in water to
In conclusion, we have demonstrated a PTE-based graphene reduce metal ion contamination, and two more DI water rinses. After letting the
photodetector, integrated in a scalable chalcogenide glass graphene film sit overnight in the final evaporating dish of water, we scooped it out

NATURE COMMUNICATIONS | (2022)13:3915 | https://doi.org/10.1038/s41467-022-31607-7 | www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31607-7

with our CaF2 substrate, blew N2 on the film to eliminate most of the trapped where e is the elementary charge, f d ðεÞ ¼ ðexpððε  EF Þ=kB TÞ þ 1Þ1 is the Fermi-
water, and then baked the sample at 80∘ for 30 min followed by 160∘ for 2 h Dirac distribution and kB is Boltzmann’s constant. As I will show below, graphene’s
(“graphene transfer process”). We then removed the PMMA from the graphene low-frequency conductivity σDC and infrared conductivity σIR affect various
using acetone at room temperature, rinsed it in isopropanol and blew it dry intermediate model parameters; σDC and σIR themselves depend strongly on EF,
(“PMMA removal process”), and baked the sample at 200 ∘C in N2 for 1 h to which features spatial variation due to the back-gates. For the graphene channel, we
improve adhesion. To pattern the graphene back-gates, we spun on a layer of 950 assume a constant Nc = N0,c + e−1CgVg in the region above each gate, where Nc is
PMMA A6, exposed the gates in the Elionix and developed using “development the carrier concentration in the channel (positive for positive EF, negative for
process”, etched away the exposed graphene using “oxygen RIE process” for 45 s, negative EF), N0,c is the native carrier concentration at zero gate voltage, Cg is the
and removed the PMMA using “PMMA removal process”. We then spun on capacitance per area of the gate dielectric, and Vg is the voltage applied to the gate
another 495 PMMA A6/950 PMMA A2 bilayer, exposed the metal contacts to the in question (using a set of test devices, we measure Cg = 34. fF/μm2 on our chip,
graphene gates using the Elionix FLS-125, and repeated “development process”, corresponding to a back-gate dielectric constant of K ≈ 12; this is described in more
“metal evaporation process”, and “liftoff process”, but using a 2 nm Ti adhesion depth in Supplementary Note 4). In the part of the graphene channel above the gap
layer in the Ti/Au stack rather than 5 nm. After this, we evaporated 1.5 nm Al between the two gates, we assume a linear slope between Nc,1 and Nc,2. For the
(Temescal VES2550) as an ALD seed layer, allowed the thin Al layer to oxidize in gates, Ng = N0,g − e−1CgVg, with Ng and N0,g defined similarly to Nc and N0,c. In
ambient, and deposited 300 cycles ≈ 30 nm of HfO2 ALD at 200 ∘C (Cambridge general, the graphene’s Fermi level and carrier concentration are related by
Nanotech Savannah 200). To define the graphene channel, we performed another pffiffiffiffiffiffiffiffiffi
ffi
EF ¼ _vgr πjNj signðNÞ, where vgr is graphene’s Fermi velocity. To incorporate
“graphene transfer process”, “PMMA removal process”, 1 h N2 ambient 200 ∘C the blurring of the graphene’s Fermi level-dependent properties due to spatial
bake, 950 PMMA A6 spin-coating, Elionix FLS-125 exposure of graphene channel carrier concentration variations, we convolve the Kubo formula with a Gaussian as
pattern, “development process”, “oxygen RIE process” for 45 s, and “PMMA follows:
removal process”. To define the channel contacts, we spin-coated another 495
Z 1 1ðnNÞ2
PMMA A6/950 PMMA A2 bilayer, exposed the graphene channel contacts using 1 
the Elionix FLS-125, and performed another “development process”, “metal eva- σ DC ðNÞ ¼ pffiffiffiffiffi e 2 σ 2n σð0; EF ðNÞ; τ DC ; T 0 Þ dn ð3Þ
σ n 2π 1
poration process”, and “liftoff process”, but using a 2 nm Ti adhesion layer in the
Ti/Au stack rather than 5 nm. We then evaporated another 1.5 nm Al seed layer and similarly for σIR(N) using ω = 2πc/λ instead of 0 and τIR instead of τDC. Finally,
using the Temescal VES2550 and 150 cycles of HfO2 ALD at 200 ∘C using the we have R ¼ σ 1DC , κ ¼ π kB T 0 σ DC =3e via the Wiedemann-Franz law, and
2 2 2

Cambridge Nanotech Savannah 200 to protect the graphene channel. Finally, to S ¼ dðlog σ DC Þ=dEF 28. Cel is obtained by convolving the heat capacity of pristine
pattern the GSSe waveguides, we coated the chip with 495 PMMA A11, used the graphene with a Gaussian of standard deviation σN as in Eqn. (3), where the
Elionix FLS-125 to define the waveguides, and developed in room-temperature 3:1 pristine heat capacity is given by28,29:
isopropanol:methyl isobutyl ketone for 120 s followed by an isopropanol rinse for Z 1
2jεj ∂f ðε  EF ðNÞÞ
120 s. The longer development time is mandated by the thicker resist film. We then C el ðNÞjσ n ¼0 ¼ ε 2 2 d dε: ð4Þ
evaporated 750 nm of Ge28Sb12Se60 followed by a quick liftoff in boiling acetone 1 π_ vgr ∂T
(~20 min), IPA rinse and N2 blow-dry, and cleaving of the chip to expose wave- We use a waveguide eigenmode solver to find the mode profile of our
guide facets. waveguide at λ = 5.2 μm, using refractive indices of 1.4, 2.6, and 1.88 for the CaF2,
GSSe, and HfO2, respectively. The resulting mode profile enters into our expression
Measurement conditions. The maps in Fig. 2a, b, and c were measured by for Q_ el as follows30:
sequentially measuring each data point column by column, bottom to top from left  
to right. SR830 lock-in amplifiers were used for all measurements. Prior to each jEx ðx; yc Þj2 þ jEy ðx; yc Þj2 σ IR;c ðxÞ
_Qel ¼ P RR : ð5Þ
data point collection, both gate voltages were reset to −7 V for 80 ms to reset the  ^
gate dielectric hysteresis (see Supplementary Note 2), then set to the desired gate R2 ReðE ´ H Þ  z dx dy

voltages and allowed to dwell for 200 ms for the lock-in signal to stabilize. The Here, yc is the y-coordinate of the graphene channel, and yg would be the y-
lock-in filter was set to a 30 ms time constant with a 12 dB/octave falloff. The R W=2
coordinate of the graphene gates. We may then write αc ¼ P1 W=2 Q_ el ðxÞ dx.
detector photovoltage in Fig. 2a was measured directly by the lock-in amplifier with
Similar expressions hold for αg in terms of σIR,g(x), noting of course that
no additional amplification. For the resistance map in Fig. 2b, we used our lock-in
σIR,g(x) = 0 for x within the gap between the gates where there is no graphene.
amplifier to bias the device with a 1 VRMS sine wave at 3.78 kHz through a 100 kΩ R W=2
resistor to act as a current source and measured the voltage across the device with Finally, ρΩ ¼ W=2 RðxÞ dx.
the lock-in. To produce the frequency response plots in Fig. 4a, we apply a sinusoid Having thus obtained expressions for κ(x), Cel(x), Q_ el ðxÞ, S(x), Π(x), αc, αg and
of variable frequency to the current modulation input of our QCL and measured ρΩ as a function of the gate voltages, as well as τDC, τIR, σn, EFc, EFg, τeph, αe, and ρc,
the calibration and photoresponse signals with a SR844 RF lock-in amplifier. For we then solve for the increase in electronic temperature per guided power ΔTel(x)/
the laser modulation response (indicated in red in Fig. 4a), we couple the laser light P = (Tel(x) − T0)/P using the equation:
through a single-mode waveguide on our chip with no devices on it and directly  
d d ΔT el _ dΠ
measure the amplified transmission signal produced by the fast InAsSb detector on  κ þ τ 1
eph C el ΔT el ¼ η Qel  J x ; ð6Þ
the output side of our chip. For the photovoltage signal (blue curve in Fig. 4a), we dx dx dx
amplify the photovoltage produced by our detector by 40 dB using a preamplifier where κ is the 2D electronic thermal conductivity of the graphene, τeph is the
and measure this amplified signal with our lock-in. In all cases, we used a dwell
time of 1.5 s, and the filter of our lock-in was set to 100 ms with a 12 dB/octave electron-phonon cooling time, Q_ el is the absorbed optical power per area, η is the
falloff. To measure the un-illuminated noise spectral density in Fig. 4b, we amplify conversion efficiency of absorbed optical power to electronic heat after initial
the noise produced by the device using a 60 dB preamplifier and analyze the output electron-phonon scattering12, Jx is the line current density in the x-direction, and Π
on an FFT signal analyzer while controlling the gate voltages applied to the device. is the Peltier coefficient. We are approximating the electric field to run exclusively
We choose to measure the averaged noise spectral density between 22 and 32 kHz in the x-direction, valid for sufficiently gradual light absorption. We assume η = 1,
where we find no electromagnetic interference-related spectral peaks in our lab as has been previously reported in pump-probe experiments at this wavelength
range13. The thermal electromotive force (EMF) arising from the Seebeck effect is
environment. At the same time as the noise measurement, we also use a lock-in
amplifier to measure the device resistance by recording the voltage across the then given by:
device while biased with 1 VRMS through a 100 kΩ resistor, albeit at a higher Z W=2
d ΔT el
frequency so as to not produce a signal in the noise measurement range. We use Ex ¼  S dx; ð7Þ
W=2 dx
our signal analyzer’s band averaging feature to measure the noise spectral density
for each data point. To produce the final plot, we manually record the resistance where W = 5.4 μm is the channel width and S is the Seebeck coefficient. In Eqns.
and noise spectral density for all gate voltage pairs from −6 V to 6 V in steps of 2 V. (6) and (7), κ, Cel, S, and Π = STel ≈ ST0 (for small ΔTel) are all dependent on the
local Fermi level EF of the graphene, and thus have a gate-tunable x-dependence,
Device modeling. We use the Kubo formula reproduced here from Hanson27 to which we account for in our calculations. Combining the equations, the ηQ_ el source
model graphene’s conductivity at DC and infrared frequencies (albeit with different term in Eqn. (6) gives rise to a proportional photo-induced EMF, whereas the
values of the Drude scattering time τ for the different frequency ranges): Peltier term J x dΠ
dx gives rises to a current-dependent EMF, which appears as a
resistance in series with the Ohmic and contact resistances of the channel. We can
thus write:
je2 ðω  jτ 1 Þ 
σðω; EF ; τ; TÞ ¼
V ¼ Rv αc PðzÞ  ρΩ þ ρΠ þ ρc J x ðzÞ ð8Þ
 π_
2
Z 1  
1 ∂f d ðεÞ ∂f d ðεÞ
´ ε  dε where V is the voltage across the contacts, Rv is the photovoltage per absorbed
1
ðω  jτ Þ 0 2
∂ε ∂ε ð2Þ
# power per length of a cross-sectional slice of the device (i.e., dimensions of V/(W/
Z 1 m)), αc is the component of the waveguide power attenuation coefficient arising
f d ðεÞ  f d ðεÞ
 dε from absorption in the graphene channel, P(z) is the guided power at a position
0 ðω  jτ 1 Þ2  4ðε=_Þ2 along the waveguide, and ρΩ, ρΠ, ρc are the Ohmic, Peltier, and contact line

6 NATURE COMMUNICATIONS | (2022)13:3915 | https://doi.org/10.1038/s41467-022-31607-7 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31607-7 ARTICLE

resistivities (dimensions of Ω ⋅ m), respectively. Averaging over z along the length 17. Saier, E. L. & Pozefsky, A. Quantitative determination of nitric oxide and
of the waveguide we obtain: nitrous oxide by infrared absorption. Anal. Chem. 26, 1079–1080
(1954).
R v αc  
V¼ 1  eαtot L Pin  RΩ þ RΠ þ Rc I; ð9Þ 18. National Institute for Occupational Safety and Health. Nitric oxide. https://
L αtot www.cdc.gov/niosh/npg/npgd0448.html, accessed: 28 March 2021.
where I is the current produced by the photodetector, thus describing a Thévenin 19. Lin, H. et al. Demonstration of high-Q mid-infrared chalcogenide glass-on-
equivalent source. Here, αtot = αc + αg + αe is the total guided power attenuation silicon resonators. Opt. Lett. 38, 1470–1472 (2013).
coefficient within the detector, including contributions not only from the graphene 20. Coddington, I., Newbury, N. & Swann, W. Dual-comb spectroscopy. Optica 3,
channel but also from the graphene gates (αg), as well as a gate-independent excess 414–426 (2016).
loss αe associated with scattering and absorption from organic or metallic 21. Klocek, P. & Colombo, L. Index of refraction, dispersion, bandgap and light
impurities attached to or trapped underneath the graphene sheets. Thus, the total scattering in gese and gesbse glasses. J. Non-Crystalline Solids 93, 1–16 (1987).
device resistance is equal to R = RΩ + RΠ + Rc, and the voltage responsivity is given 22. Kim, M.-H. et al. Photothermal response in dual-gated bilayer graphene. Phys.
by: Rev. Lett. 110, 247402 (2013).
23. Tsay, C., Toor, F., Gmachl, C. F. & Arnold, C. B. Chalcogenide glass
R v αc 
Rv ¼ 1  eαtot L ; ð10Þ waveguides integrated with quantum cascade lasers for on-chip mid-IR
L αtot photonic circuits. Opt. Lett. 35, 3324–3326 (2010).
which we plot versus both gate voltages in Fig. 3b for the best-fit device parameters 24. Kim, L., Kim, S., Jha, P. K., Brar, V. W. & Atwater, H. A. Mid-infrared
given in Table 1 obtained as described in Supplementary Note 3. All calculations radiative emission from bright hot plasmons in graphene. Nat. Mater. 20,
are carried out in Mathematica. 805–811 (2021).
25. Osman, A. et al. Suspended low-loss germanium waveguides for the longwave
infrared. Opt. Lett. 43, 5997–6000 (2018).
26. Soibel, A. et al. Midinfrared interband cascade laser for free space optical
Data availability communication. IEEE Photon. Technol. Lett. 22, 121–123 (2010).
The datasets generated during and/or analyzed during the current study are available in 27. Hanson, G. W. Dyadic green’s functions and guided surface waves for a
the FigShare repository at https://doi.org/10.6084/m9.figshare.c.5514759.v1. surface conductivity model of graphene. J. Appl. Phys. 103, 064302
(2008).
28. Grosso, G. & Parravicini, G. P. Solid State Physics. 2nd edn (Academic Press,
2014).
Code availability 29. Castro Neto, A. H., Guinea, F., Peres, N. M. R., Novoselov, K. S. & Geim, A. K.
The Mathematica document used to simulate photodetector performance metrics is
The electronic properties of graphene. Rev. Mod. Phys. 81, 109–162
available in the FigShare repository at https://doi.org/10.6084/m9.figshare.c.5514759.v1.
(2009).
30. Snyder, A. W. & Love, J. D. Optical waveguide theory. 1st edn (Chapman and
Received: 27 March 2022; Accepted: 16 June 2022; Hall, 1983).
31. Thorlabs, Inc. Mid-IR Photovoltaic Detectors, HgCdTe (MCT). https://www.
thorlabs.com/newgrouppage9.cfm?objectgroup_id=11319, accessed: 29
March 2021.
32. Li, C. C., Han, C.-J. & Skidmore, G. D. Overview of DRS uncooled VOx
infrared detector development. Optical Eng. 50, 1–8 (2011).
References
1. Stuart, B. H. Infrared Spectroscopy: Fundamentals and Applications. Analytical Acknowledgements
Techniques in the Sciences. 1st edn (Wiley, 2004). We would like to thank the MIT.Nano and MIT Nanostructures Laboratory staff for
2. Ramanathan, V. Greenhouse effect due to chlorofluorocarbons: Climatic maintaining the cleanroom facilities used to fabricate these devices, in particular Mark
implications. Science 190, 50–52 (1975). Mondol, Jim Daley, and Dave Terry. We also would like to thank Sebastián Castilla of
3. Pi, M. et al. Design of a mid-infrared suspended chalcogenide/silica-on-silicon ICFO for helpful discussions. We acknowledge support from the Army Research Office
slot-waveguide spectroscopic gas sensor with enhanced light-gas interaction via the MIT Institute for Soldier Nanotechnologies University-Affiliated Research Center
effect. Sens. Actuators B: Chem. 297, 126732 (2019). (ISN UARC) under grant number W911NF-18-2-0048 (J.G., A.-Y.L., J.K., D.E.), the U.S.
4. Jin, T., Zhou, J., Lin, H.-Y. G. & Lin, P. T. Mid-infrared chalcogenide Army Research Office MURI program under grant number W911NF-18-1-0431 (J.K.),
waveguides for real-time and nondestructive volatile organic compound the Air Force Office of Scientific Research (AFOSR) MURI-FATE program under grant
detection. Anal. Chem. 91, 817–822 (2019). number FA9550-15-1-0514 (M.H., T.P., J.K.), and NSF Award numbers 1122374 (J.G.),
5. Su, P. et al. Monolithic on-chip mid-IR methane gas sensor with waveguide- 1453218 (H.L., J.H.) and 2023987 (S.D.-J., J.H.). Any opinions, findings, and conclusions
integrated detector. Appl. Phys. Lett. 114, 051103 (2019). or recommendations expressed in this material are those of the author(s) and do not
6. Yadav, A. & Agarwal, A. M. Integrated photonic materials for the mid- necessarily reflect the views of the National Science Foundation.
infrared. Int. J. Appl. Glass Sci. 11, 491–510 (2020).
7. Rogalski, A. Infrared Detectors. 3rd edn (CRC Press, 2019).
8. Huang, L. et al. Waveguide-integrated black phosphorus photodetector for Author contributions
mid-infrared applications. ACS Nano 13, 913–921 (2019). J.H., D.E., and J.G. conceived the experiments. J.G. designed, fabricated, and measured
9. Deckoff-Jones, S., Wang, Y., Lin, H., Wu, W. & Hu, J. Tellurene: A the devices, with the exception of chalcogenide glass deposition, performed by H.L. and
multifunctional material for midinfrared optoelectronics. ACS Photon. 6, S.D.-J. under the supervision of J.H., and graphene growth, performed by M.H. and A.-
1632–1638 (2019). Y.L. under the supervision of J.K. and T.P. K.R. provided the chalcogenide glass sources
10. Ma, P. et al. Plasmonically enhanced graphene photodetector featuring 100 for thermal deposition. J.G. and D.E. wrote the manuscript. All work was supervised
Gbit/s data reception, high responsivity, and compact size. ACS Photon. 6, by D.E.
154–161 (2019).
11. Gabor, N. M. et al. Hot carrier–assisted intrinsic photoresponse in graphene.
Science 334, 648–652 (2011).
Competing interests
The authors declare no competing interests.
12. Song, J. C. W., Rudner, M. S., Marcus, C. M. & Levitov, L. S. Hot carrier
transport and photocurrent response in graphene. Nano Lett. 11, 4688–4692
(2011). Additional information
13. Tielrooij, K. J. et al. Photoexcitation cascade and multiple hot-carrier Supplementary information The online version contains supplementary material
generation in graphene. Nat. Phys. 9, 248–252 (2013). available at https://doi.org/10.1038/s41467-022-31607-7.
14. Lin, H. et al. Chalcogenide glass-on-graphene photonics. Nat. Photon. 11,
798–805 (2017). Correspondence and requests for materials should be addressed to Dirk Englund.
15. Hooge, F. N. 1/f noise sources. IEEE Trans. Electron Device. 41, 1926–1935
(1994). Reprints and permission information is available at http://www.nature.com/reprints
16. Siebert, R. & Müller, J. Infrared integrated optical evanescent field sensor for
gas analysis: Part i: System design. Sens. Actuators A: Phys. 119, 138–149 Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
(2005). published maps and institutional affiliations.

NATURE COMMUNICATIONS | (2022)13:3915 | https://doi.org/10.1038/s41467-022-31607-7 | www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31607-7

Open Access This article is licensed under a Creative Commons

Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
article’s Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.

© The Author(s) 2022

8 NATURE COMMUNICATIONS | (2022)13:3915 | https://doi.org/10.1038/s41467-022-31607-7 | www.nature.com/naturecommunications