Terahertz photodetection in scalable single-layer-graphene and hexagonal boron nitride heterostructures

The unique optoelectronic properties of single layer graphene (SLG) are ideal for the development of photonic devices across a broad range of frequencies, from X-rays to microwaves. In the terahertz (THz) range (0.1-10 THz frequency) this has led to the development of optical modulators, non-linear sources, and photodetectors, with state-of-the-art performances. A key challenge is the integration of SLG-based active elements with pre-existing technological platforms in a scalable way, while maintaining performance level unperturbed. Here, we report on the development of room temperature THz detection in large-area SLG, grown by chemical vapor deposition (CVD), integrated in antenna-coupled field effect transistors. We selectively activate the photo-thermoelectric detection dynamics, and we employ different dielectric configurations on SLG on Al2O3 with and without large-area CVD hBN capping to investigate their effect on SLG thermoelectric properties underpinning photodetection. With these scalable architectures, response times ~5ns and noise equivalent powers ~1nWHz-1/2 are achieved under zero-bias operation. This shows the feasibility of scalable, large-area, layered materials heterostructures for THz detection.

The unique optoelectronic properties of single layer graphene (SLG) are ideal for the development of photonic devices across a broad range of frequencies, from X-rays to microwaves.In the terahertz (THz) range (0.1-10 THz frequency) this has led to the development of optical modulators, non-linear sources, and photodetectors, with state-of-theart performances.A key challenge is the integration of SLG-based active elements with preexisting technological platforms in a scalable way, while maintaining performance level unperturbed.Here, we report on the development of room temperature THz detection in largearea SLG, grown by chemical vapor deposition (CVD), integrated in antenna-coupled field effect transistors.We selectively activate the photo-thermoelectric detection dynamics, and we employ different dielectric configurations on SLG on Al2O3 with and without large-area CVD hBN capping to investigate their effect on SLG thermoelectric properties underpinning photodetection.With these scalable architectures, response times ~5ns and noise equivalent powers ~1nWHz -1/2 are achieved under zero-bias operation.This shows the feasibility of scalable, large-area, layered materials heterostructures for THz detection.Two-dimensional (2D) layered materials (LMs) and related heterostructures (LMHs) are a versatile platform for engineering optoelectronic and photonic devices 1,2,3 .They can be synthesised with wafer-scale methods 2,4,5 and stacked to form LMHs. Being compatible with Si and a wider range of III-V materials and substrates 2,[5][6][7] they open new prospects for emerging research domains such as future high-density optical communications 6 , high-speed datacom 8 , quantum nanophotonics 9 and optoelectronics 1,3 .In particular, LMs are a versatile platform for devising photodetectors (PDs) operating across a broad range of frequencies from microwaves 10 , to telecom 8 , visible [1][2][3] , and Xrays 11 .
A ~10 nm thick layer of Al2O3 is deposited on SiO2/Si substrate (resistivity~10kΩ•cm) by atomic layer deposition (ALD) (see Supplementary Information).SLG is grown in a hot wall CVD system using ~30 µm thick Cu foil as substrate.The foil, suspended on a quartz holder and loaded into the CVD system, is annealed at 1050 o C for 2h under H2 gas (100 sccm) at 760 Torr and cooled down to RT.For the growth, the foil is annealed at 1050 o C with 50 sccm hydrogen flow at 0.4 Torr for 2h. 5 sccm CH4 is introduced to start growth, which is completed in 30mins by stopping the CH4 flow.The system is then naturally cooled down to RT under 50 sccm H2 flow.As-grown SLG/Cu is spin-coated with poly (methyl methacrylate) (PMMA) (A4-950) at 1000 rpm for 1 min and baked at 80 o C for 10 mins.PMMA-coated SLG/Cu is kept in water overnight to oxidize Cu foil.PMMA/SLG is then electrochemically delaminated by applying 2 V between Pt anode and PMMA/SLG/Cu cathode in a NaOH aqueous electrolyte (~1M).The PMMA/SLG stack is cleaned in water and transferred on Al2O3/SiO2/Si substrates, which are then baked at 80 o C for 10mins after ~10 h natural drying.PMMA is removed by soaking in acetone and isopropyl alcohol (IPA).
hBN is grown on c-plane Al2O3 (0001) at 1400ºC, 500 mbar for 30 minutes in an AIXTRON CCS 2D reactor.10 sccm N2 is used to transport the single-source precursor, borazine, to the reactor.Before hBN growth, the sapphire substrates are annealed in H2 atmosphere for 5 mins, at 750 mbar and 1180ºC.As-grown hBN on c-plane sapphire is then spin-coated with PMMA (A4-950) at 1000 rpm for 1min and baked at 80 o C for 10 mins.PMMA-coated hBN on sapphire is kept in ~8% H3PO4 for ~10 h to delaminate PMMA/hBN.This is cleaned in water and transferred on SLG/Al2O3/SiO2/Si.
After natural drying, this is baked at 80 o C for 10 mins.PMMA is removed by soaking it in acetone and IPA.
The antenna arms are connected to the s and g electrodes.The second design features two top-gates, connected to the left and right arms of a 24 μm radius bow-tie antenna.The antenna dimensions are chosen to be resonant with a radiation frequency of 2.8 THz. 28These geometries are selected to activate the photo-thermoelectric (PTE) effect as dominant detection mechanism. 31,32,49This requires a spatial asymmetry along the SD channel. 3In single-gated systems, the asymmetry is established by the spatial gradient of the electronic temperature (Te) in the SLG channel, induced by the absorption of THz light in the antenna gap, which is located close to the s electrode. 31Instead, in p-n junctions, the electronic distribution is symmetrically heated at the centre of the s-d channel, where the antenna gap is located. 29Here, the asymmetry is determined by the longitudinal variation of the SLG Seebeck coefficient (Sb), whose profile along the s-d direction can be electrostatically defined by applying distinct gate voltages (Vg) on the left and right sides of the junction.
We activate a dominant PTE by design, since SLG displays unique thermoelectric properties, due to its low electronic specific heat (∼2000 kBμm −2 at RT, 50 where kB is the Boltzmann constant), ultrafast (~30 fs) carrier thermalization dynamics 51 and slower (~4 ps) electron-phonon cooling. 52,53us, it offers an outstanding route for performance optimization by means of e.g., carrier lifetime engineering 49 or coupling to plasmonic-polaritonic quasi-particles. 54PTE also allows for broadband, 49 room-temperature, zero-bias operation. 8e first characterize the devices electrically, by measuring the source-drain current (Isd) as a function of Vg.The resistance (R) curve for a single-gated device is in Fig. 2a.Instead, Fig. 2b shows the resistance map of a p-n junction device, measured as a function of the left-and right-gate voltages (VgL and VgR).We extract the field-effect mobility (μFE) for electrons and holes and the residual carrier density (n0), by using the fitting function 55 R= R0+(Lc/Wc)•(1/n2deμFE), where n2d=[n0 2 + (Cg/e•(Vg−VCNP)) 2 ] ½ is the carrier density, 55 Cg is the gate-capacitance per unit area and VCNP is the charge neutrality point.We get μFE=300-2000 cm 2 V -1 s -1 for Al2O3/SLG-based GFETs and μFE=3000-9000 cm 2 V -1 s -1 for hBN-capped ones.
Fig. 2c plots n0 as a function of μ for the complete batch of devices, where μ is the average field-effect mobility of electrons and holes: each GFET is represented by a coloured dot.From this comparison, hBN-capped samples show lower n0 and higher μ with respect to Al2O3/SLG/HfO2 devices.The field-effect measurements (Figs.2a,b) can be used to evaluate Sb, which determines the PTE response of SLG-based PDs.This can be done starting from the experimental conductivity (σ) and using the Mott equation, 30 which, however, is not accurate at low carrier densities, 56 i.e. close to VCNP.Thus, we theoretically calculate Sb using an effective medium theory (EMT) 56 in the framework of the linear Boltzmann equation 53,57 (see Supplementary Information).for the R curve in (a).
We then evaluate the PDs optical figures of merit: voltage responsivity (Rv), NEP and τ.The detectors are illuminated by a 2.8 THz quantum cascade laser (QCL), driven in pulsed mode (repetition rate 40 kHz, duty cycle 4%) delivering a peak power ~25 mW, corresponding to an average power ~1 mW.The antenna axis is oriented parallel to the linearly polarized electric field.The beam is focused by two TPX lenses onto a ~200 μm radius circular spot.We select an intermediate average power P0=0.4 mW to characterize the PDs, to avoid QCL overheating.The corresponding average intensity in the focal point is I0=0.32Wcm -2 .
We measure the photovoltage (Δu) at the d electrode, while keeping s grounded.Δu is amplified by a voltage preamplifier (DL Instruments, M1201, gain γ=1000) and sent to a lock-in (Stanford Research, 5210).We use a square-wave envelope with frequency fmod=1.333kHz as lockin reference and as triggering signal for the QCL pulse trains.Δu can be inferred from the demodulated lock-in signal (VLI) via the relation 30 Δu=(π√2/2)VLI/γ, where the pre-factor π√2/2 takes into account that the lock-in measures the root mean square of the fundamental Fourier component of the square wave produced by the QCL modulation.Rv is calculated from the ratio between Δu and the power Pa=I0Aeff impinging on the detector, with Aeff the detector effective area, assumed equal to the diffraction limited area 30 Aeff=λ 2 /4=2800 μm 2 , where λ is the free-space wavelength.We calculate the curve of Rv vs. Vg for each single-gated GFET and the map of Rv vs. VgL and VgR for each p-n junction.Typical examples of Rv vs. Vg plots are in Fig. 3a,b for a single-gated GFET and a p-n junction, respectively.The photoresponse in single-gated GFETs follows the profile of Sb, with an offset Sbu≅Sb (Vg=0 V), with PTE voltage 30 VPTE= ΔTe•(Sb -Sbu), where ΔTe is the Te gradient between s (hot) and d (cold) sides of the SLG channel, and Sbu the Seebeck coefficient of the ungated region between the s and g electrodes.Fig. 3a compares the measured Rv vs. Vg and the theoretical PTE responsivity RPTE vs. Vg, with RPTE inferred from VPTE by considering ΔTe/P0~1.5K/mW, showing good qualitative agreement between the two curves.The discrepancy between the theoretical and experimental responsivities at large positive Vg is ascribed to the fact that the adopted theoretical model (see Supporting Information) only includes electron scattering with charged Coulomb impurities as dominant effect limiting the conductivity, possibly neglecting additional contributions, e.g.phonon scattering or carrier inhomogeneities at the contacts.A dominant PTE detection mechanism is also observed in p-n junctions.EF in SLG can be tuned across the Dirac point by the electrostatic gating applied to the left and right sides of the junction.The non-monotonic dependence of Sb on EF leads to multiple sign changes in Rv, resulting in a six-fold pattern 29 in the Rv map (Fig. 3b), a distinctive feature of PTE. 26,29e sensitivity of THz detectors is evaluated through the NEP, 29 defined as the ratio between noise figure and responsivity.0][31] Thus, we measure the GFETs NSD with a lockin amplifier (Zurich Inst., UHFLI): the s electrode is grounded and the signal, demodulated by the lock-in, is collected at the d electrode, while a sweep of the modulation frequency is performed.The results are in Fig. 4c for a 50 Ω test resistor and for a prototypical SLG-based device.The white noise floor for the 50Ω resistor is dominated by the lock-in noise figure.The Johnson-Nyquist NSD formula gives 30 NJ=(4kBTR) ½ =0.91 nVHz -½ for a 50Ω resistor operated at RT, whereas our instrumental noise floor is ~8 nVHz -½ , as expected for the noise level of the employed lock-in. 58The NSD of one of the GFETs (R=9 kΩ in Fig. 4c) is dominated by the 1/f component 59 for modulation frequency <1kHz and flattens at NSD <14 nVHz -½ at higher frequencies, in agreement with the theoretically expected NJ=12.3nVHz -½ .The GFET NJ is thus the main contribution to the overall noise figure in our setup (with pre-amplifier NSD~7 nVHz -½ ).The measured NSD at 1.333kHz is then used to calculate NEP (Fig. 3d,e) as a function of the voltages applied to the gate electrodes.
We then characterize the detection speed by recording the time trace of Δu with an oscilloscope (Tektronix DPO520-4B, bandwidth 2 GHz).We use a THz pulse duration ~1.6 μs and we amplify the PD output with a high-bandwidth (1.1 GHz) voltage preamplifier (Femto, DUPVA-1-70) before the oscilloscope.We drive the QCL into the negative differential resistance regime, 30 which results in electronic instabilities that correspond to an intermittent output power: the QCL undergoes intensity fluctuations with characteristic time constants τqcl~0.9ns.This strategy allows us to test the bandwidth of our PDs up to a maximum (2πτqcl) -1 =180 MHz.Fig. 3f shows the waveform recorded by a single-gated GFET during an intensity fluctuation of the pulsed QCL.We evaluate  from exponential fits to the waveform (see Supplementary Information).We get =7-20 ns, with a mean value ~12 ns, corresponding to a bandwidth ~15 MHz.Statistical analysis is applied to 28 devices to evaluate performance variability and identify correlations between electrical and optical properties.We first consider NEP variability.For Al2O3/SLG/HfO2 devices, we get a mean value ~7.6 nWHz -½ and an interquartile range 28 (IQR) ~4.0 nWHz -½ .For hBN-capped PDs, we have mean NEP~3.0 nWHz -½ with IQR~1.4 nWHz -½ , which represents a variability improvement of a factor >2 with respect to SiO2/SLG 24 and Al2O3/SLG PDs.
We then evaluate correlations between NEP, Sb and n0 using the Pearson coefficient 60 (ρ) as a metric.ρ(v1,v2) represents the measure of linear correlation between two discrete variables v1 and v2: |ρ|=1 indicates an exact linear dependence and ρ=0 indicates no linear correlation.We get ρ(Smax,n0)= −0.95 for both hBN-capped and uncapped architectures, where Smax is the maximum |Sb| in the investigated Vg range, calculated with the EMT model.The scatter plot of Smax vs. n0 in Fig. 4a shows that hBN/SLG/Al2O3 LMHs have slightly larger Smax, even though they have significantly smaller n0.This is due to the different dielectric environment: the larger εr of HfO2 (with respect to hBN) on top of SLG is beneficial in terms of thermopower. 56This similarity in Smax is reflected in the detectors NEP, where the difference between the two material architectures is not as pronounced as the difference in n0 (Fig. 4b).However, in agreement with results obtained on SiO2/SLG/HfO2 heterostructures, 28 NEP increases for larger n0: ρ(log(NEP),log(n0))=0.4.These correlations confirm that the physical mechanism underpinning THz detection is, as expected, PTE.Thus, the Al2O3 termination alone does not show a significant performance improvement over SiO2/Si substrates, 28 whereas large-area HfO2/hBN/SLG/Al2O3 LMHs present advantages both in terms of absolute optical performance (average NEP~3.0 nWHz -½ ) and performance variability (IQR~1.40nWHz -½ ).It is worth mentioning that large area hBN-top-encapsulation significantly reduces the device performance variability by more than a factor 2 with respect to Ref. 28.
0][31] We demonstrate THz detection in a layered material heterostructure obtained by consecutive transfer of CVD graphene and CVD hexagonal boron nitride, a fabrication technique that is fully compatible with standard CMOS processing.This makes our PDs suited for real-time imaging and short-range (~10 m) THz communication applications, enabling

Figure 1 .
Figure 1.(a) Schematic cross-section of the GFETs.Top: double-gated Al2O3/SLG/HfO2.Bottom: singlegated Al2O3/SLG/hBN/HfO2.Dashed red circles indicate the position of the THz-induced field enhancement.(b) Raman spectra of as-grown SLG on Cu and SLG transferred on Al2O3 with and without hBN capping.(c,d) False colour scanning electron micrographs of a single-gated device.The inset shows the U-shaped channel.

Figure 2 .
Figure 2. (a) Channel resistance vs. Vg for an hBN-capped, single-gated GFET.(b) Map of R as a function of VgR and VgL for a hBN-capped p-n junction.(c) Chart of n0 vs. μFE for all the measured devices.Blue (lightblue) dots represent single-gated (double-gated) FETs without hBN-capping.Orange (red) dots represent hBNcapped single-gated (double-gated) FETs.(d) Sb calculated with the Boltzmann EMT51,56,57 for the R curve in (a).

Figure 3 .
Figure 3. (a) Rv as a function of Vg for an hBN-capped single-gated GFET.The experimental curve (black line) is compared with the theoretical PTE response (blue line), evaluated by EMT.(b) Rv map of an hBNcapped p-n junction, as a function of VgL and VgR.(c) NSD of a GFET and of a 50 Ω resistor, measured by sweeping the reference frequency of the lock-in from 100 Hz to 1 MHz.(d,e) NEP of a single-gated and a p-n junction GFETs as a function of gate voltage(s).(f) Time trace of an intensity fluctuation of the QCL.The rising and falling edges are fitted with exponential functions to retrieve .

TABLE I .
Performance of room-temperature THz detectors based on scalable, large-area graphene.