This work experimentally demonstrates mid-infrared emittance spectra of dielectric and semi-conductor substrates with and without a germanium–antimony–tellurium (GST) film coating. The film experiences non-volatile phase changes at 140°C and 300°C. Impacts from amorphous, face-centered cubic, and hexagonal close packed phases on spectral emittance are demonstrated within the spectral range from 4 μm to 18 μm. The spectra are measured at 100°C, 200°C, 300°C, and 400°C to show temperature dependence. Close-to-total emittance is calculated for comparison. The GST film can reduce emittance from a SiO2 substrate, but it raises close-to-normal emittance as well as the spectral emittance at wavelengths 5 μm ≤ λ ≤ 18 μm for the doped Si substrate.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Phase change materials, vanadium dioxide (VO2)  and germanium–antimony–tellurium (GST) , can tailor emission because their refractive index (n) and extinction coefficient (k) vary with the phase. The VO2 is promising for applications at relatively low temperature (∼ 67°C) [3,4]. On the other hand, the GST is attractive to devices working at or experiencing higher temperature than the phase transition temperature. For example, Qu et al. were able to control thermal emission from a metallic substrate via a GST film and its transition between crystalline and amorphous phases . Thermal camouflage using coverage of a GST film was also realized .
Spectral emittance ε from a GST film is a fundamental optical response relying on not only phase but temperature and wavelength. However, quantitatively specifying the dependences of ε on temperature and three phases (amorphous, face-centered cubic (FCC), and hexagonal close packed (HCP)) are still unavailable. Moreover, little was known about the capability of a semi-transparent GST film tailoring IR emittance from commonly-seen dielectrics. The missing of total emittance hinders development of applications at extreme temperature.
This work is thus going to experimentally investigate ε spectra from a GST film on dielectric and semi-conductor substrates. All three phases will be looked into, and the emission wavelength λ is from 4 μm to 18 μm. The spectral range covers more than 75% of thermal emission from a blackbody at preset temperature (100°C, 200°C, 300°C, and 400°C) based on the Planck’s distribution . Four types of samples are carefully prepared. Two are plain dielectric and semi-conductor substrates, while the other two are the dielectric and semi-conductor substrates coated with a GST film. Oven annealing is employed to change its phase from amorphous to FCC or HCP . The glass (SiO2) and doped silicon (Si) are selected for substrates because of their popularity in semiconductor and display industry, respectively.
2. Sample preparation and characterization
Figure 1(a) shows the refractive index (n) and extinction coefficient (k) of SiO2 at 4 μm ≤ λ ≤ 18 μm . The n > k within most of the spectral range assures dielectric characteristics of employed substrate. The penetration depth (δ) is calculated via δ = λ/4πk  and plotted. Since the thickness of SiO2 substrate is 700 μm and much larger than δ at λ ≥ 5 μm, the substrate is opaque. The opaqueness is critical for ε measurements because it assures emission contribution from the substrate itself. Figure 1(b) shows n and k of highly-doped Si , the other employed dielectric. The dopant is Boron, and its doping concentration is very high (at the order of 1019 cm-3 based on the fitted optical constants of the doped Si wafer) such that the sheet resistance is smaller than 0.0015 Ω·cm. Thickness of the doped Si substrate is 350 μm, much larger than δ at 4 μm ≤ λ ≤ 18 μm. The substrate is thus opaque. Plain SiO2 and doped Si substrates are named as sample I and sample II, respectively. Their ε will serve as benchmarks. Sample III is the SiO2 substrate coated with a 122-nm-thick GST film. Sample IV is the doped Si substrate with a 80-nm-thick GST film. Insets of Fig. 1 exhibit sketches of the four samples and their dimensions. Figure 1(c) shows n, k, and δ spectra of GST at amorphous and FCC phases . Because δ is larger than the film thickness of samples, the GST film is semi-transparent. The ε from samples III and IV are contributed jointly from both the substrate and film emission.
Figures 2(a) and 2(b) show the measured diffraction patterns from our fabricated GST film supported by the SiO2 and doped Si substrates, respectively. The diffraction measurement is conducted using an x-ray diffractometer (Bruker, D8). Peaks of substrate crystalline structures are removed to clearly reveal the crystalline structures of GST films. Each diffraction angle peak at 2θ is associated with unique crystallographic plane . These peaks are exhibited at the FCC and HCP phases, but not amorphous phase. Planes of the FCC phase are (200), (220), and (222). In contrast, planes of the HCP phase are (103), (106), (110), and (203). Clearly, the GST film at amorphous, FCC, or HCP phase is successfully fabricated on the two substrates.
Figure 3 shows hemispherical absorptance A and normal transmittance T spectra of samples I and II to verify n and k of substrates. A numerical model is constructed using transformation matrix method  along with dimensions from SEM images. Numerically obtaining A and T from the model uses homemade codes . On the other hand, A and T are measured with a custom-designed optical set-up  and commercially-available FT-IR spectroscopy (Nicolet, iS50), respectively. When A is measured, the sample is mounted at center of an integrating sphere. The polar angle of incidence θi is set to 8°, suppressing leakage of specularly reflected light. The set-up has a long-wavelength limit at λ = 12 μm associated with its germanium detector. Conversely, the FT-IR spectroscopy is able to measure transmittance T through each sample at θi = 0°.
Figure 3(a) shows A and T spectra of sample I. Modeling and experimental results agree well with each other, which indicated the n and k of SiO2 and GST used for modeling are reliable. T is zero at λ ≥ 7.5 μm, assuring opaqueness of the sample. Even at 4 μm ≤ λ ≤ 7.5 μm, T is also trivial (< 0.2). The sample is close to opaque. For this sample, A is larger than 0.9, except at λ ≈ 10 μm. Figure 3(b) shows A and T spectra of sample II. The sample is opaque within the spectral range. A is larger than 0.8 at 4 μm ≤ λ ≤ 6 μm. It then monotonically decays to A = 0.25 at λ = 18 μm. Modeling and experimental results are again consistent, which verifies the n and k of doped Si and GST.
3. Emittance spectra
Spectral emittance ε from samples are measured using a custom-designed emissometer, which can directly measure ε at high temperature . The measurement uncertainty is around 0.05 in the infrared wavelength range. The ε of samples I and II serve as benchmarks to demonstrate impacts from the GST film coating on emission. Measurements are conducted at 100°C, 200°C, 300°C, and 400°C. At 100°C, all three phases of GST can exist. At 200°C, the amorphous phase cannot exist because it changes to FCC during measurements. Only FCC and HCP phases can sustain. Similarly, neither amorphous nor FCC can exist at 300°C and 400°C. The film can only be at HCP phase.
Figures 4(a) and 4(b) show ε spectra of samples with the SiO2 and doped Si substrate, respectively. Insets are the SEM images of samples III and IV. In Fig. 4(a), the ε spectrum of SiO2 substrate without GST coverage is close to previously shown A of the SiO2 substrate. The consistency comes from trivial optical constant dependence of SiO2 on temperature. The Kirchhoff’s law  assures the equality of ε and A. The four spectra in Fig. 4(a) show that the coated GST film can significantly modify ε from that of the SiO2 substrate. When the film is at amorphous phase, ε is lower than that of plain substrate at 4 μm ≤ λ ≤ 8 μm. But ε is enlarged to 0.9 at 9 μm ≤ λ ≤ 18 μm, As the phase becomes FCC, the ε spectrum is even lower than the above-mentioned two within most of the spectral range. ε is lower than 0.6 at 4 μm ≤ λ ≤ 8 μm, and it is about 0.8 at 9 μm ≤ λ ≤ 18 μm. The ε spectra for GST at amorphous and FCC phase exhibit similar trends. But the ε spectrum associated with the GST at HCP phase becomes unique. ε is about 0.6 at 4 μm ≤ λ ≤ 18 μm and then monotonically reduces to ε = 0.4 at λ = 18 μm. The GSP film at HCP phase actually leads to the largest reduction in ε from that of plain SiO2 substrate among three phases. The maximum reduction is Δε = 0.55 at λ = 18 μm.
In Fig. 4(b), ε spectrum of doped Si substrate shows a peak at λ ≈ 5 μm and then decays monotonically till ε = 0.2 at λ = 18 μm. The ε at 100°C largely differs from A at room temperature. One reason is the rough surface on the back side. It magnifies A considering the scattering effect inside an integrating sphere , but the roughness has trivial influences on emissometry measurements. The other reason for discrepancy is the variation of λ and k. As the temperature raises, the carrier (hole) concentration and scattering rate both increase . The doped Si substrate behaves more metallic during measurements for ε than those for A. The ε at 100°C is therefore lower than A at room temperature within most of the spectral region.
Four spectra in Fig. 4(b) depict that the GST film largely changes the ε spectrum of doped Si substrate. The phase variation further shifts the λ and expands the bandwidth of the emittance peak in the spectrum. Once the substrate is covered with an amorphous GST film, ε is enlarged within the whole spectral range. The peak bandwidth also expands. The ε is about 0.8 at 4 μm ≤ λ ≤ 8 μm, and ε monotonically decays with λ. As the film phase changes to FCC and HCP, the peak of ε red-shifts to λ = 7.5 μm. ε is further enlarged at 7 μm ≤ λ ≤ 18 μm. Two ε spectra are quite similar, while the FCC phase leads to a little larger ε at 4 μm ≤ λ ≤ 10 μm than the HCP phase. But the FCC phase results in lower ε at 10 μm ≤ λ ≤ 18 μm than the HCP phase. Large emittance peak and dips occur slightly above 4 μm in spectra of Fig. 4 are not due to any absorption/reflection/transmission enhancement mechanism. They are mainly attributed by the CO2 absorption in the atmosphere. The absorption significantly reduced the signal and thus caused notable uncertainties.
Unique emission spectra from samples at the same temperature are dominated by local structures of GST and their vibration modes. The amorphous phase of GST withholds defective octahedra, GeTe4-pGep (p = 0, 1, 2) edge and/or corner-sharing tetrahedral, and SbTe3 pyramid units. Because the phase is constructed from predominantly covalent bonds, the refractive index is relatively small . Conversely, the crystalline phase of GST withholds defective octahedra, corner-sharing GeTe4-pGep (p = 0, 1, 2), and hexagonal Sb2Te3 . The highly polarizable delocalized p-orbital resonant bonding results in a large refractive index . That is, the number of bonds and local orders of atoms magnified the optical constants as well as weakened emission spectrum with the transition from amorphous phase to crystalline phases. The difference in emission spectra can also be explained via larger amount of defects in amorphous GST. These defects result in higher resistivity than that of the crystalline GST as they will trap charge carriers . Therefore, the defects of GST can be considered as insulators at amorphous state, while crystalline grains act as metals relatively. The emittance difference of SiO2 substrates with and without amorphous GST film also has some interesting phenomena. At λ < 8 μm, the emittance is higher without GST film. Since the k of GST is very small within the investigated spectrum, the emission of the GST film is negligible. However, as GST has relatively low δ and it is also optically denser than SiO2, the GST film actually blocks the thermal emission from SiO2 and thus leads to the reduction of emission spectrum. At 8 μm < λ < 9.5 μm where the phonon band of SiO2 is located, the metallic behavior of SiO2 and the total internal reflection at the GST-air interface will form a Fabry-Perot cavity which ends up with a distinct absorption (emission) peak. At λ > 8 μm, the δ of GST is now large enough to contribute on the emission spectrum (more than twice of that at λ = 4 μm) and thereby the emittance is further enhanced at λ > 8 μm.
Figures 5(a) and (b) show the ε spectra of samples at 200°C. In Fig. 5(a), both FCC and HCP phases of the GST film can largely reduce ε at 4 μm ≤ λ ≤ 18 μm. Moreover, the HCP phase film can diminish ε at 8 μm ≤ λ ≤ 18 μm. But ε within this spectral range is a little higher than the counterpart at 100°C. In Fig. 5(b), the GST at both FCC and HCP phases enlarges ε from that of doped Si substrate at 6 μm ≤ λ ≤ 18 μm. Both ε spectra show a peak at λ ≈ 7.5 μm like those at 100°C. The maximum ε associated with the FCC phase is close to unity. But the maximum ε associated with the HCP phase is close to 0.8, almost the same as at 100°C.
Figures 5(c) and 5(d) exhibit ε spectra at 300°C and 400°C. The GST phase is HCP only. In Fig. 5(c), the GST film can suppress ε within the whole spectral range. As temperature increases from 300°C to 400°C, ε from the SiO2 substrate increases about 0.02. But ε from the substrate with a GST film decreases 0.05 at λ ≈ 8 μm. In Fig. 5(d), the GST film enlarges ε from that of doped Si substrate at 6 μm ≤ λ ≤ 18 μm. The difference in ε from two samples diminishes as temperature increases from 300°C to 400°C. The main reason is that ε from the doped Si substrate increases with temperature. But the high ε from the sample decreases with temperature.
4. Total emittance
Table 1 list εt from all samples. εt of the SiO2 substrate is high (εt ≥ 0.80) and trivially temperature dependent. Adding the GST film at amorphous phase dose not influence radiative heat transfer at 100°C because εt remains 0.88. But the film at both FCC and HCP phases can reduce dissipation/absorption of radiative power. The HCP phase even outperforms the FCC phase due to the low εt (0.58 < 0.68). At 200°C, the GST film at two phases still has lower εt (0.62 and 0.63) than that of SiO2 (0.79). But the difference between εt at FCC and HCP phases becomes trivial. At 300°C and 400°C, the GST film at HCP phase still diminishes heat dissipation/absorption thanks to the reduced εt, demonstrating potentials as a thermal fuse for SiO2. In contrast to SiO2, εt of doped Si substrate is low (εt = 0.37) at 100°C. It monotonically increases with temperature to εt = 0.55 at 400°C. Adding the GST film is able to facilitate radiative heat transfer at 100°C regardless of the film phase. The film at both FCC and HCP phases can better help heat dissipation/absorption of radiative power than the amorphous phase. At temperature higher than 100°C, the sample with a GST film still enhance radiative heat dissipation/absorption. But the relative enhancement gets trivial because εt of doped Si enlarges with temperature.
To sum up, this work demonstrates capabilities of a GST film and its amorphous, FCC, and HCP phases in modifying ε spectra from two dielectric substrates. The GST film at all three phases is able to largely reduce ε from a SiO2 substrate at 4 μm ≤ λ ≤ 8 μm. The HCP phase can even reduce ε within the whole spectral range. Close-to-total emittance εt is calculated to quantify impacts of GST film on emission. The GST film can benefit thermal camouflage or serve as a heat fuse for objects made of glass. Conversely, the GST film significantly enlarges ε from the doped Si substrate at 6 μm ≤ λ ≤ 18 μm. The ε peak red-shifts to λ = 7.5 μm, and its bandwidth expands. The film can facilitate radiative heat dissipation from semi-conductor devices composed of doped silicon. A GST film and its phase transition clearly show promising potentials in thermal management and spectral property manipulation.
Ministry of Science and Technology, Taiwan (109-2636-E-009-014, 110-2221-E-007-044-MY2, 110-2636-E-009-019, 110-3116-F-027-001).
We much appreciate help from Prof. Tsung-Eong Hsien and his team with NYCU for fabricating samples.
The authors declare no conflicts of interest.
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
1. Z. L. Gong, F. Y. Yang, L. T. Wang, R. Chen, J. Q. Wu, C. P. Grigoropoulos, and J. Yao, “Phase change materials in photonic devices,” J. Appl. Phys. 129(3), 030902 (2021). [CrossRef]
2. P. G. Guo, A. M. Sarangan, and I. Agha, “A review of germanium-antimony-telluride phase change materials for non-volatile memories and optical modulators,” Appl. Sci. 9(3), 530 (2019). [CrossRef]
3. M. C. Larciprete, M. Centini, S. Paoloni, I. Fratoddi, S. A. Dereshgi, K. Tang, J. Wu, and K. Aydin, “Adaptive tuning of infrared emission using VO2 thin films,” Sci. Rep. 10(1), 11544 (2020). [CrossRef]
4. L. S. Long, S. Taylor, and L. P. Wang, “Enhanced infrared emission by thermally switching the excitation of magnetic polariton with scalable microstructured VO2 metasurfaces,” ACS Photonics 7(8), 2219–2227 (2020). [CrossRef]
5. K. K. Du, Q. Li, Y. B. Lyu, J. C. Ding, Y. Lu, Z. Y. Cheng, and M. Qiu, “Control over emissivity of zero-static-power thermal emitters based on phase changing material GST,” Light: Sci. Appl. 6, e16194 (2017).
6. Y. R. Qu, Q. Li, L. Cai, M. Y. Pan, P. Ghosh, K. K. Du, and M. Qiu, “Thermal camouflage based on the phase-changing material GST,” Light: Sci. Appl. 7(1), 26 (2018). [CrossRef]
7. M. F. Modest, Radiative Heat Transfer (Elsevier Science, 2013).
8. M. Wuttig and N. Yamada, “Phase-change materials for rewriteable data storage,” Nat. Mater. 6(11), 824–832 (2007). [CrossRef]
9. E. D. Palik and G. Ghosh, Handbook of Optical Constants of Solids (Academic, 1998).
10. S. Basu, B. J. Lee, and Z. M. Zhang, “Infrared radiative properties of heavily doped silicon at room temperature,” J. Heat Transfer 132(2), 023301 (2010). [CrossRef]
11. I. Friedrich, V. Weidenhof, W. Njoroge, P. Franz, and M. Wuttig, “Structural transformations of Ge2Sb2Te5 films studied by electrical resistance measurements,” J. Appl. Phys. 87(9), 4130–4134 (2000). [CrossRef]
12. Z. M. Zhang, Nano/microscale Heat Transfer (McGraw-Hill, 2007).
13. Y. H. Chen, F. Y. Shih, M. T. Lee, Y. C. Lee, and Y. B. Chen, “Development of lightweight energy-saving glass and its near-field electromagnetic analysis,” Energy 193, 998–1006 (2020). [CrossRef]
14. Y. B. Chen, Y. C. Lee, Y. F. Chang, Y. H. Lin, and P. H. Chen, “Tempering hemispherical radiative properties with a resonance compilation,” Plasmonics 10(3), 595–603 (2015). [CrossRef]
15. M. Born, E. Wolf, and A. B. Bhatia, Principles of Optics: Electromagnetic Theory of Propagation, Interference, and Diffraction of Light (Cambridge University, 2019).
16. C. J. Fu and Z. M. Zhang, “Nanoscale radiation heat transfer for silicon at different doping levels,” Int. J. Heat Mass Transfer 49(9-10), 1703–1718 (2006). [CrossRef]
17. T. Cao, L. Zhang, R. E. Simpson, and M. J. Cryan, “Mid-infrared tunable polarization-independent perfect absorber using a phase-change metamaterial,” J. Opt. Soc. Am. B 30(6), 1580–1585 (2013). [CrossRef]
18. S. Guo, Z. G. Hu, X. L. Ji, T. Huang, X. L. Zhang, L. C. Wu, Z. T. Song, and J. H. Chu, “Temperature and concentration dependent crystallization behavior of Ge2Sb2Te5 phase change films: tungsten doping effects,” RSC Adv. 4(100), 57218–57222 (2014). [CrossRef]
19. K. Shportko, S. Kremers, M. Woda, D. Lencer, J. Robertson, and M. Wuttig, “Resonant bonding in crystalline phase-change materials,” Nat. Mater. 7(8), 653–658 (2008). [CrossRef]
20. M. Kaes, M. Le Gallo, A. Sebastian, M. Salinga, and D. Krebs, “High-field electrical transport in amorphous phase-change materials,” J. Appl. Phys. 118(13), 135707 (2015). [CrossRef]