Highly Responsive Mid-Infrared Metamaterial Enhanced Heterostructure Photodetector Formed out of Sintered PbSe/PbS Colloidal Quantum Dots

Efficient and simple-to-fabricate light detectors in the mid infrared (MIR) spectral range are of great importance for various applications in existing and emerging technologies. Here, we demonstrate compact and efficient photodetectors operating at room temperature in a wavelength range of 2710–4250 nm with responsivities as high as 375 and 4 A/W. Key to the high performance is the combination of a sintered colloidal quantum dot (CQD) lead selenide (PbSe) and lead sulfide (PbS) heterojunction photoconductor with a metallic metasurface perfect absorber. The combination of this photoconductor stack with the metallic metasurface perfect absorber provides an overall ∼20-fold increase of the responsivity compared against reference sintered PbSe photoconductors. More precisely, the introduction of a PbSe/PbS heterojunction increases the responsivity by a factor of ∼2 and the metallic metasurface enhances the responsivity by an order of magnitude. The metasurface not only enhances the light–matter interaction but also acts as an electrode to the detector. Furthermore, fabrication of our devices relies on simple and inexpensive methods. This is in contrast to most of the currently available (state-of-the-art) MIR photodetectors that rely on rather expensive as well as nontrivial fabrication technologies that often require cooling for efficient operation.


1) CQD Synthesis Synthesis of PbS Nanocrystals
In a typical synthesis of 8 nm PbS nanocrystals 1 , we mix 1.8 g of PbO with 60 ml of oleic acid and 20 ml of 1-octadecene in the three-neck flask. The reactor is then connected to the Schlenk line set-up and heated to 150°C for 1 hour in vacuum. During this time, the mixture is purified from oxygen and water residues, while simultaneously PbO and oleic acid react forming Lead (II) oleate, a precursor of Pb. Afterwards, the reactor is put under a nitrogen stream (1 bar), while the temperature is kept at 150°C. At these conditions, a sulfur precursor mixture, consisting of 0.84 ml of hexamethyldisilthiane and 40 ml of 1-octadecene, is introduced to the reaction via fast injection. The mixture turns brown within a few seconds, indicating a formation of PbS nanocrystals.
To achieve a good size distribution of PbS nanocrystals, the reaction is taken through a optimized temperature profile, namely (i) a natural cooling to 120°C during the first minute of reaction; (ii) slower cooling to 100°C within the next 3 minutes; (iii) annealing at 100°C for 5 minutes; (iv) fast cooling to room temperature to terminate the reaction and mass transfer processes.
As synthesized nanocrystals are then transferred air free to the glovebox and purified by the addition of hexane and ethanol solvents. This step is followed by centrifugation, the overall purification cycle is repeated 3 times.

Synthesis of PbSe Nanocrystals
Synthesis of PbSe nanocrystals is derived from the PbS synthesis above, replacing the sulfur precursor with trioctylphosphine selenide and adding diphenylphosphine as nucleation promoter 2 .
In a typical synthesis of 4 nm PbSe nanocrystals, we mix 2.2 g of PbO with 7.86 ml of oleic acid and 34.26 ml of 1-octadecene and heat this mixture at 150°C and vacuum to form Lead (II) oleate. Afterwards, the injection mixture of 30 ml of 1M trioctylphosphine selenide and 0.2 ml of diphenylphosphine is added, triggering the formation of PbSe nanocrystals. The reaction time is shortened to 30 seconds, after what the reaction is terminated by fast cooling to the room temperature. Post-synthetic purification of PbSe nanocrystals is carried out in complete analogy to the PbS synthesis above.  Figure S2. (a) PbSe XRD spectra for different annealing times. It can be seen that the peak height decreases with increasing annealing time, indicating a decrease of crystallinity for longer annealing times. (b) PbS XRD spectra for different annealing times. It can be seen that the peak height increases with increasing annealing time, indicating an increase in crystallinity for longer annealing times. (c) XRD pattern a single layer of PbS (black) and a single layer of PbSe (red) annealed for 1 min at 310 °C. The arithmetic sum of the annealed PbSe and PbS pattern (green) and the XRD pattern of the stacked PbSe/PbS layers. It can be seen that the shape of the single sum of the single layers matches very well the shape of the PbSe/PbS bilayer stack. The difference in peak intensity is attributed to a difference in the thickness of the measured layers.  Figure S4. Illustration of the optical setup used for the electro-optical characterization.

5) Discussion on the Origin of the Photoconductivity in the Annealed PbSe CQD Layers
The annealed PbSe shows a significant photoresponse in the MIR without any sensitization. Recent works suggest that the formation of trap states at grain boundaries, which are often accredited to oxides, play a significant role for the photoconductance of PbSe 3,4 . Although, no oxides are expected to form in the annealed PbSe CQD layer, we suspect that trap states exist at the grain boundaries.
More precisely, layers consisting of CQDs have a very large surface due to the large surface to volume ratio of the individual CQDs. The surfaces of CQDs usually have a large number of defects which is why a lot of effort has been put into different ligand exchange methods to passivate the surface and surface traps. We expect that during the annealing most of ligands evaporate and several CQDs fuse into larger grains. Despite the fusing we suspect that a large number of the original traps on surface of the individual CQDs are still present at the grain boundaries of the annealed layer. Since the CQDs crystallize into small grains many boundaries with the accompanying traps exist in theses layers, which lead to an increased photoresponse.

6) Discussion on the Presences of a Heterojunction
In the main text we have already argued that the heterojunction formation can be observed by the fact that the current decreases in the PbSe/PbS stack when applying a forward current due to the depletion of carriers at the interface. Here we give more arguments for the formation of a heterojunction. It can be expected that the two materials form a heterojunction since bulk PbSe has an electron affinity of and a bulk band gap of which is known to form a χ PbSe, bulk ≈ 4.7 eV E g, PbSe ≈ 0.27 eV, heterojunction with PbS ( [5][6][7][8] . Although these values can differ for χ PbS, bulk ≈ 4.55 eV, E g,PbS ≈ 0.4 eV) CQDs 7,9-11 and the exact values are unknown for the annealed CQD layers presented in this work, the formation of a heterojunction between individual PbS and PbSe crystal domains may be expected due to the differences in the bandgap.
A further indicator of the presence of a heterojunction can be seen by analyzing the normalized I-V curves shown in Figure S5. PbS forms an ohmic contact with Au which results in the nearly linear I-V, whereas PbSe forms a Schottky type contact. This difference in contact type indicates that the barrier height between Au and the pristine materials PbS and PbSe are likely to be different. In such a case it can be argued that energetic band position of PbS and PbSe differ as well, which also should lead to the formation of a heterojunction if the materials are brought into contact with each other.

7) Time Dependent Photoresponse
In Figure S6 (a) the photoresponse over a long period of time is shown and in Figure S6 (b) the photocurrent at a modulation frequency of 4 Hz, which was used to measure a responsivity of ~250 A/W. The ratio of the maximum current of these measurements was used to calculate the maximum responsivity of ~375 A/W at a wavelength of 2710nm. Both measurements were performed under identical illumination conditions.

10) Comparison of Photodetectors
The below presented table provides an overview of recent publications of photodetectors operating in the MIR. The responsivities and detectivities are used as a figure of merit. This comparison focuses on 2D and CQD materials and their combination and is not complete. A single representative responsivity and detectivity value was selected if several values were given in the referenced publications