Research ArticleMATERIALS SCIENCE

Ultrahigh thermal isolation across heterogeneously layered two-dimensional materials

See allHide authors and affiliations

Science Advances  16 Aug 2019:
Vol. 5, no. 8, eaax1325
DOI: 10.1126/sciadv.aax1325
  • Fig. 1 Optical and STEM characterization of vdW heterostructures.

    (A) Cross-section schematic of Gr/MoSe2/MoS2/WSe2 sandwich on SiO2/Si substrate, with the incident Raman laser. (B) Raman spectrum of such a heterostructure at the spot indicated by the red dot in the inset optical image. Raman signatures of all materials in the stack are obtained simultaneously. The graphene Raman spectrum is flattened to exclude the MoS2 photoluminescence (PL) effect. arb.u., arbitrary units. (C to F) STEM cross-sectional images of four-layer (C) and three-layer (D to F) heterostructures on SiO2. In (D), MoSe2 and WSe2 are approximately aligned along the 1H [100] zone axis, and in (E and F), the layers are misaligned by ~21° with respect to the 1H [100] zone axis. The monolayer graphene on top of each heterostructure is hard to discern due to the much lower atomic number of the carbon atoms. (G) PL spectra of monolayer MoS2, monolayer WSe2, and a Gr/MoS2/WSe2 heterostructure after annealing. The PL is strongly quenched in the heterostructure due to intimate interlayer coupling.

  • Fig. 2 Electrical and scanning probe characterization.

    (A) Cross-sectional schematic of the test structure showing the four-probe configuration. Electrical current flows in the graphene top layer, and heat dissipates across layers, into the substrate. (B) Optical image of a four-probe test structure. Devices are back-gated by the Si substrate through 100-nm SiO2. (C) Measured transfer characteristics of three test structure stacks, Gr/MoS2/WSe2, Gr/WSe2, and Gr-only control devices in vacuum (~10−5 torr). All measurements display the ambipolar property of the top graphene channel. (D) KPM of an uncapped Gr/MoS2/WSe2 heterostructure device. The graph displays the surface potential along the channel (averaged across the channel width) at different bias conditions. The small potential jump near the Pd electrodes represents the relative work function difference (~120 mV). The KPM maps reveal no other heterogeneities in the surface potential, confirming the spatially uniform quality of these devices. The inset shows the zero-bias KPM map. (E) SThM thermal map of Gr/MoS2/WSe2 heterostructure, here capped with 15-nm Al2O3, revealing homogeneous heating across the channel. This confirms the uniformity of the thermal interlayer coupling in the stacks. The device dimensions are the same as in the (D) inset.

  • Fig. 3 Thermal resistance of the heterostructures.

    (A) Measured temperature rise ΔT versus electrical input power for each individual layer in a Gr/MoS2/WSe2 heterostructure, including the Si substrate, shown in the inset. Graphene (pink circles), MoS2 (blue diamonds), WSe2 (red triangles), and Si (black squares). All measurements are carried out at VG < 0 (see section S6). The slopes of the linear fits (dashed lines) represent the thermal resistance Rth between each layer and the heat sink. (B) Comparison of total thermal resistances (i.e., of the top graphene layer) measured by Raman thermometry and SThM for different vdW heterostructures. The Rth values obtained from these two techniques match within the uncertainty of the measurements. All devices have the same active area of ~40 μm2.

  • Fig. 4 Summary of TBC trends.

    (A) Schematic of all TBCs measured (in MW m−2 K−1) across heterostructures consisting of, clockwise from top left, graphene (Gr), Gr/MoS2, Gr/WSe2, and Gr/MoS2/WSe2, all on SiO2/Si substrates. (B) Measured TBC values of 2D/2D and 2D/3D (with SiO2) interfaces (red diamonds, left axis) and the calculated product of phonon density of states (PDOS), phonon transmission, and df/dT (blue circles, right axis). Calculated values are normalized to the minimum achieved for Gr/WSe2 (see table S2). The dashed line between the simulation symbols is a guide to the eye. Lower TBC is noted at interfaces between 2D/2D materials and those between materials with larger mismatch in mass density. Three devices were measured for each structure, at two or more distinct positions of the Raman laser. No significant TBC variation is seen between samples with different layer (mis)alignment, within the experimental uncertainty. All values are at room temperature.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/8/eaax1325/DC1

    Section S1. Multiple transfer technique

    Section S2. Atomic-resolution ADF-STEM

    Section S3. PL quenching in 2D heterostructures

    Section S4. SThM thermal maps

    Section S5. SThM temperature calibration

    Section S6. Raman temperature maps

    Section S7. Temperature-dependent Raman coefficients

    Section S8. Non–temperature-related Raman peak shifts

    Section S9. Thermal transport modeling and analysis

    Fig. S1. Transfer process and optical images.

    Fig. S2. Atomic-resolution ADF-STEM images.

    Fig. S3. PL quenching.

    Fig. S4. Uniform heating maps.

    Fig. S5. SThM calibration by metal line heaters.

    Fig. S6. Raman maps.

    Fig. S7. Temperature-dependent Raman coefficients.

    Fig. S8. Raman shift dependence on doping and electrical gating.

    Fig. S9. PDOS overlap.

    Table S1. Material parameters.

    Table S2. TBC comparison.

    References (3555)

  • Supplementary Materials

    This PDF file includes:

    • Section S1. Multiple transfer technique
    • Section S2. Atomic-resolution ADF-STEM
    • Section S3. PL quenching in 2D heterostructures
    • Section S4. SThM thermal maps
    • Section S5. SThM temperature calibration
    • Section S6. Raman temperature maps
    • Section S7. Temperature-dependent Raman coefficients
    • Section S8. Non–temperature-related Raman peak shifts
    • Section S9. Thermal transport modeling and analysis
    • Fig. S1. Transfer process and optical images.
    • Fig. S2. Atomic-resolution ADF-STEM images.
    • Fig. S3. PL quenching.
    • Fig. S4. Uniform heating maps.
    • Fig. S5. SThM calibration by metal line heaters.
    • Fig. S6. Raman maps.
    • Fig. S7. Temperature-dependent Raman coefficients.
    • Fig. S8. Raman shift dependence on doping and electrical gating.
    • Fig. S9. PDOS overlap.
    • Table S1. Material parameters.
    • Table S2. TBC comparison.
    • References (3555)

    Download PDF

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article