Loss and Coupling Tuning via Heterogeneous Integration of MoS2 Layers in Silicon Photonics

Layered two-dimensional (2D) materials provide a wide range of unique properties as compared to their bulk counterpart, making them ideal for heterogeneous integration for on-chip interconnects. Hence, a detailed understanding of the loss and index change on Si integrated platform is a prerequisite for advances in opto-electronic devices impacting optical communication technology, signal processing, and possibly photonic-based computing. Here, we present an experimental guide to characterize transition metal dichalcogenides (TMDs), once monolithically integrated into the Silicon photonic platform at 1.55 um wavelength. We describe the passive tunable coupling effect of the resonator in terms of loss induced as a function of 2D material layer coverage length and thickness. Further, we demonstrate a TMD-ring based hybrid platform as a refractive index sensor where resonance shift has been mapped out as a function of flakes thickness which correlates well with our simulated data. These experimental findings on passive TMD-Si hybrid platform open up a new dimension by controlling the effective change in loss and index, which may lead to the potential application of 2D material based active on chip photonics.

Transition metal dichalcogenides (TMDs), a family of materials with a general formula of MX2, where M is a transition metal (Mo, W, Re) and X is a chalcogen (S, Se or Te), provide a promising alternative to integrate them into atomically precise heterostructures combining conducting (graphene) and insulating (hBN) 2D materials. A stable member of the TMDs family, molybdenum disulfide (MoS2), has attracted widespread attention for a variety of next-generation electrical and opto-electronic properties such as high room temperature mobility, high switching characteristics, bandgap tunability, and high exciton binding energies [11][12][13]. Bulk MoS2 has an indirect bandgap of ~1.2 eV which, due to quantum confinement, crosses over to a direct bandgap of ∼1.8 eV, when the material is a monolayer [12]. Due to this tunable electrical and optical properties, MoS2 could be a prospect for future advances in the field of nano-optics and photonics [13,14]. This material is studied here as monolithically integrated with Silicon photonics as just one example, where we are interested in the respective impact of optical absorption and index-shift whence heterogeneously added to a Silicon photonic waveguide and microring resonator (MRR). Beyond MoS2, other TMDs could be studied as well in follow-up work on the same or similar integrated photonics platform, which is interesting, since the spectral distance of each TMD's exciton relative to the waveguide probing wavelength (here, l = 1550 nm) is different; thus, the real vs. imaginary part index decay away from the exciton resonance has a different impact on the telecom-operating photonic structures.
Silicon photonics is becoming an integration platform of large interest for optical datacom and telecom applications [15]. However, Silicon's weak electro-optic properties and indirect bandgap severely limit opto-electronic device functionality. In contrast, hybrid or heterogeneous photonic integration solutions offer an appealing approach, when combined with an optical low-loss, yet commercially accessible large volume and low-cost CMOS fabrication technology such as Si/SiN photonics [16][17][18][19]. Other active opto-electronic materials such as transparent conductive oxides, while showing high switching performance, usually introduce relatively high optical losses [20].
Whereas, because of the advent of sufficiently strong van der Waals (vdW) force, 2D materials can (in principle) be easily integrated with photonic chip, offering a rich variety of electronic and optical properties that enable light generation, modulation, and detection could be a promising platform for next-generation PIC [21][22][23][24]. In reality, the state-of-the-art of TMDs transfer techniques is not benign with taped-out chip technology due to the inability to place a single 2D material flake on the pre-fabricated photonics chip without incurring significant crosscontamination (e.g. transfer of undesired flakes). We recently provided a solution for this challenge developing a 2D material printer enabling cross-contamination-free transfers without impacting the underlying photonic waveguide structures reported in ref [25].
Here, we demonstrate a novel heterogeneous platform to study the physical properties of TMDs by newly developed 2D printer transfer technique, enabling rapid and precise transfer of 2D atomic layers on the integrated photonic chip without any cross contamination. Using the TMD-Silicon heterogeneous integrated platform, we perform a comparative study to determine the optical loss and refractive index change as a function of 2D material Silicon waveguide and microring resonator (MRR) coverage length and TMD (MoS2) thickness at a telecom wavelength. The effect of MRR-to-waveguide coupling has been mapped out in terms of resonance shift as a function of monolayer coverage analyzing the loss induced by monolayer MoS2 which is found to be about 0.005 dB/µm. We obtain a resonance shift per unit waveguide coverage length of 0.064 nm/µm as a function of thickness which matches our numerical results well. Together these experimental studies of integrating MoS2 with Silicon photonics shows an induced negligible loss, but relatively strong index-tuning potential thus paving the way for future studies of active opto-electronic device technology.

METHODS:
Here, we demonstrate a heterogeneous platform to study the effect of ultrathin TMDs towards building on-chip active device component. The study is performed on taped-out Si photonic chips (APPLIED NANOTOOLS INC.). It is important to keep all the physical parameters unchanged before and after the transfer of the 2D materials to single out the influence solely from atomic layered materials, thus; first, we coat a uniform layer of polymethyl methacrylate (PMMA) (~300 nm) as a cladding for the improvement of coupling efficiencies of the grating couplers (GC). Then, in order to keep a similar coupling efficiency and to eliminate the variation of light coupling, it is important to remove the PMMA layer and transfer the 2D material over the targeted opening areas of the devices. A box shape opening has been made by electron beam lithography for transferring 2D materials on top of the photonic devices (Fig. 2a).
In order to understand the potential of 2D TMDs at Telecom wavelength range, here we study MoS2 integrated Si photonic platform as a function of layer thickness. Few-layer MoS2 flakes are obtained using scotch tape exfoliation technique, whereas the monolayers are triangular flakes grown on Si/SiO2 substrate by CVD process [26,27].

RESULTS & DISCUSSION:
In order to realize MoS2 as an active material at telecom wavelength (here 1.55 µm), it is imperative to understand the interaction of monolayer and few layers system on integrated Si photonic devices (Figure 1). We study the loss and coupling in detail as a function of monolayer coverage and thickness of the MoS2. It is well-known that the optical properties of TMDs materials are dominated by excitons: bound electron-hole pairs with strong binding energy due to quantum confinement and weak screening of the Coulomb interaction [12]. For any active integrated device structures, one needs to characterize the individual material system systematically such as loss or effective index change at the targeted device operation wavelength range. Here, we investigate the loss impact and impact on the effective mode index of the Si waveguide hundreds of nanometers away from the excitonic transition of MoS2 (A exciton ~1.88 eV & B exciton ~2.06 eV) [13]. We anticipate a minimal loss impact but meaningful index change upon heterogeneous integration. While we are not modulating the TMD here electrically, the passive impact of the modes index and impact on an MRR provide fundamental insides in the potential of TMD-Si hybrid devices such as phase modulators [28][29][30]. We, therefore, study the waveguide bus-to-ring coupling change by shifting the MRR's phase upon adding MoS2. We measure the corresponding transmission spectra which show a coupling change as compared to bare ring indicating MRR mode index tuning (Figures 1c&f). The improvement of coupling can be attributed as shifting the MRR from the over coupled regime towards the critically coupled regime by inducing loss which are evident since the quality factor decreases from ~1500 to ~1100 as a function of the flake thickness, suggesting gradual increase of loss mainly due to scattering effect, caused by small impedance mismatch between bare and TMD covered sections of the ring.
Quantitative modeling and analyzing the MRR's resonance change, the fringe-visibility can be optimized (shifting towards critical coupling) in two ways: either by increasing coverage length or by increasing the thickness of flakes. So, to understand the coupling effect, it is important to extract coupling coefficients, especially round-trip transmission coefficients (a) as a function of coverage. The transmission, T, from an all-pass MRR (Figure 3a) is given by, where is the round-trip phase shift, r is the self-coupling coefficient and a is round-trip transmission coefficient related to the power attenuation coefficients by, ) = exp5− 89 (2 − )@ * exp (− BCD(89 * ) dB/µm and 0.04 dB/µm, respectively for Si, monolayer and multi-layer flakes, respectively (Table   1). Inserting these values into (2), we find the round-trip transmission coefficients (a) to be tuned as a function of TMD coverage (Figure 2d). The variation of a from 0.77 to 0.60 as a function of coverage for multilayer flakes and from 0.75 to 0.70 for monolayer flakes, respectively. The result explains the improvement of visibility inducing the transition from over-coupled to towards critically-coupled regime since a = 1 is the zero-loss condition of the ring. Hence, the loss of tunability in MRRs can be manipulated accordingly by controlling the coverage length and thickness. We conclude that for a given device, one can determine the coverage length and thickness by configuring the device at a critically coupled condition for optimized light-matterinteraction [31].
Si-based MRRs provide a compact and ultra-sensitive platform to find sensitive detection of an unknown analyte for various applications [32,33]. Here, the detection mechanism is mostly based  Since, here the MRR is partially covered by MoS2 flakes the effective refractive index of the ring can be formulated as an effective length-fraction index via [35], GII,'9"P = ()QR(S) * " LTT,UVKL &S * " LTT )QR (3) where R is the radius of the ring and l is the MoS2 coverage length. The effective index for monolayer covered waveguide is found to be 1.723 for TM mode. The effective index (neff) for flakes with different thickness can be found using FEM Eigenmode analysis. The refractive index of the MoS2 layer was taken from [34] and Si refractive index from [35]. Figures 3a &b represent the normalized in-plane (|Ex|) electric field distribution for TM mode along the device which integrates a layer of MoS2 of 20 and 50 nm thickness, respectively. It is possible to observe that the higher intensity of the electric field in correspondence of the MoS2 layer and the consequently decreased leakages in air and substrate suggest a higher confinement for multi-layer MoS2 (50 nm). At this stage, using eqn (3), we can obtain ∆ GII and hence the resonance shift (∆ ), which is showing an unequivocal correlation with our experimental data (Figure 3d). We map out the resonance shift (∆ ) as a function of MoS2 flake thickness (Figure 3d, (i)) upto 50 nm and observe a resonance shift per coverage length of 0.064 nm/µm, which is beyond the resolution limit of our spectrometer (~0.03 nm). The results indicate mono-multilayer MoS2 integrated on Si photonic platform could be an interesting choice for active modulator devices at telecom wavelength.

CONCLUSION:
We have demonstrated the interaction between mono to multi-layers of MoS2 heterogeneously integrated onto a Silicon photonic waveguides and microring cavities. The coupling regime of the ring can be tunable from over coupled regime to under coupled regime. The underlying physical mechanism of tunable coupling can be explained by extracting different coupling and loss coefficients as a function of coverage length and thickness. This study demonstrates a method to determine critical coverage for a given ring resonator which is an important parameter for obtaining maximum extinction ratio for the active modulator. We have mapped out resonance shift as a function of monolayer coverage and thickness of MoS2 flakes which shows a resonance shift of 0.64 nm/µm correlates well with our simulated result. These findings along with the developed methodology for placing MRRs into critical coupling for active device functionality and determining the refractive index of 2D materials could be useful tools in future heterogeneous integrated photonic and opto-electronic devices.