betaFeedback

WorksPhysics and Astronomy

Quantum Hall effect in InAsSb quantum wells at elevated temperatures

Abstract

We have characterized the electronic properties of a high-mobility two-dimensional electron system in modulation doped InAsSb quantum wells and compare them to InSb quantum wells grown in a similar fashion. Using temperature-dependent Shubnikov-de Haas experiments as well as FIR transmission we find an effective mass of m∗≈m^{\ast} \approx 0.022mem_{e}, which is lower than in the investigated InSb quantum well, but due to a rather strong confinement still higher than in the corresponding bulk compound. The effective gg-factor was determined to be g∗≈g^{\ast} \approx 21.9. These results are also corroborated by k⋅pk \cdot p band structure calculations. When spin polarizing the electrons in a tilted magnetic field, the gg-factor is significantly enhanced by electron-electron interactions, reaching a value as large as g∗g^{\ast} = 60 at a spin polarization P = 0.75. Finally, we show that due to the low effective mass the quantum Hall effect in our particular sample can be observed up to a temperature of 60 K and we propose scenarios how to increase this temperature even further.

Introduction

Since its discovery in 1980 by von Klitzing et al. , the quantum Hall effect (QHE) has been observed in a variety of high-mobility 2D systems, ranging from MOS-FETs [1], quantum wells (QWs) [2, 3] as well as intrinsic 2D materials [4, 5, 6, 7]. This resulted in the definition of the resistance quantum, which enabled an incredibly precise determination of the fine-structure constant. Another breakthrough was achieved with the discovery of room temperature (RT) QHE in graphene, making it a convenient platform for performing metrological experiments [8, 9, 10, 11, 12, 13]. Since the robustness of the QHE is, amongst others, governed by the Landau level (LL) spacing (ΔLL∼2000\Delta_{LL} \sim2000 K at 30 T for graphene), InSb with its small effective mass (mbulk∗=0.014mem^*_{\mathrm{bulk}} = 0.014m_e), would conceptually also be a good material for the observation of a RT QHE. Optimally alloying InSb with InAs can reduce the effective mass even further, by utilizing the band bending commonly observed in ternary III-V compound semiconductors [14, 15]. The QHE in InSb QWs has been investigated from several perspectives, e.g. FQHE [16, 17], spin effects [16, 17], high-current breakdown [3] and Ising quantum Hall ferromagnetism [18], whereas the literature on InAsSb QWs is less abundant. However, this material has become the subject of renewed interest because of its strong spin-orbit interaction [19, 20] and small direct band gap [21].

In this paper, we present magneto-transport and far infrared (FIR) transmission data of InSb and InAsSb modulation doped QWs, enabling a direct comparison between the parent compound and the alloy. Firstly, the temperature dependence of Shubnikov-de Haas (SdH) oscillations enables the determination of the effective mass. Simultaneously, we acquire the temperature where the QHE disappears. The results for the effective mass are consistent with those obtained from cyclotron resonance (CR) measurements. Coincidence experiments in tilted magnetic fields are used to extract the gg-factors, which are consistent with band-structure calculations based on k⋅pk \cdot p theory. Additionally, an interaction-induced enhancement of the gg-factor for low filling factors, i.e. high spin polarization, is observed.

Samples and Methods

Both the InSb as well as the InAsSb QW discussed in this manuscript, were grown by molecular beam epitaxy (MBE). The sample structure is based on the work of Lehner et al. [22], who systematically investigated the effect of employing buffer layers to overcome the lattice mismatch between (100) GaAs substrates and the InSb QWs [23]. On the other hand, the InAsSb QW was grown on a (001) GaSb substrate, thereby removing the AlSb/GaSb intermediate transition. The underlying In1−y_{1-y}Aly_ySb/In1−x_{1-x}Alx_xSb (x=0.42x=0.42, y=0.54y=0.54) buffer was adjusted to the InAsSb QW lattice constant. The InSb (In0.38_{0.38}Al0.62_{0.62}) QW was surrounded by an Al0.1_{0.1}In0.9_{0.9}Sb (Al0.42_{0.42}In0.58_{0.58}Sb) confinement barrier, which contained an asymmetric doublesided Si δ\delta-doping layer, located 30 nm (44 nm) above and below the QW. The layer stack of both QW structures, as well as an investigation of the InAsSb QW and its As/Sb ratio via XRD, TEM, and EDX, are shown in Figure 1 and 2 of the supplemental information [24].

Magnetic field dependence of $\rho_{xx}$ and $\rho_{xy}$ in an InAsSb QW at different temperatures, and temperature dependence of $\sigma_{xx}$ at multiple filling factors

Figure 1. (a) Magnetic field dependence of ρxx\rho_{xx} and ρxy\rho_{xy} in an InAsSb QW at different temperatures. (b) Temperature dependence of σxx\sigma_{xx} at multiple filling factors. Using the relation σxx∝exp⁡(−Eact/2kBT)\sigma_{xx} \propto\exp(-E_{\mathrm{act}}/2k_{\mathrm{B}}T), we extracted the activation energies from the linear fits.

For the magneto-transport measurements both heterostructures were processed into a 880 ×\times 25 μm2\mu\mathrm{m}^{2} Hall bar using wet chemical etching. Ar-milling the top surface removes any contaminants and ensures that the successive layers of Ge/Ni/Au evaporated onto it, provide high quality Ohmic contacts. More information about the microfabrication can be found in Refs. 17 and 25. The FIR transmission measurements were performed with an unprocessed piece of the same wafers, warranting a better signal-to-noise compared to the Hall bar. In order to suppress Fabry-Pérot interference effects in the transmitted light, the back of the sample was wedged under an angle of ∼4∘\sim4^\circ.

The transport measurements were performed in a 4He{}^{4}\mathrm{He} bath cryostat, containing a variable temperature insert (VTI) with T=1.4T = 1.4 - 300 K, which was placed inside a 38 T Florida-Bitter magnet. By mounting the sample on a rotating platform, it was possible to change the angle of the magnetic field in-situ. Standard low-frequency lock-in techniques were used to acquire ρxx\rho_{xx} and ρxy\rho_{xy} of a 310 ×\times 25 (405 ×\times 25) μm2\mu\mathrm{m}^{2} section of the InAsSb (InSb) Hall bar.

A Fourier transform infrared (FT-IR) spectrometer (Bruker VERTEX 80v) with a Globar broadband source, was used for the FIR transmission experiment.

The IR radiation was coupled into a quasi-optical beamline, which guided the radiation towards a 33 T Florida-Bitter magnet. At the end of the beamline the radiation was focused into an oversized brass waveguide and passed through a polyethylene window, into a new-silver waveguide that was inside a 4He{}^{4}\mathrm{He} bath cryostat. Before reaching the sample, the radiation was focused using a brass cone. The transmitted IR radiation was collected by a second brass cone, passed through another polyethylene window onto a Si bolometer that was kept at superfluid helium temperatures. The FT-IR spectra were measured at different fixed magnetic fields and normalized with a background spectrum, which was constructed from all spectra by taking the 90th90^{\mathrm{th}} percentile of each frequency, to obtain the transmittance T(B)=Tm(B)/TrefT(B) = T_{\mathrm{m}}(B)/T_{\mathrm{ref}}.

Results

In the following section only the InAsSb QW will be discussed, though comparisons will be made between these results and those obtained with the InSb QW. The results from the InSb can be found in the supplemental information (SI).

SdH and quantum Hall measurements

As illustrated in Figure 1a, the magneto-transport measurements performed on a InAsSb QW Hall bar show clear SdH oscillations in ρxx\rho_{xx} and quantized plateaus in ρxy\rho_{xy} at h/νe2h/\nu e^{2}, where ν\nu is the filling factor. A carrier concentration of n=5.39×1011 cm−2n = 5.39 \times10^{11}\ \mathrm{cm}^{-2} and n=4.67×1011 cm−2n = 4.67 \times10^{11}\ \mathrm{cm}^{-2} was extracted from the classical Hall effect in small magnetic fields for the InAsSb and InSb QW, respectively. The corresponding mobility μ\mu for both the InAsSb and InSb QW were 4.24×104 cm2V−1s−14.24 \times10^{4}\ \mathrm{cm}^{2}\mathrm{V}^{-1}\mathrm{s}^{-1} and 3.37×105 cm2V−1s−13.37 \times10^{5}\ \mathrm{cm}^{2}\mathrm{V}^{-1}\mathrm{s}^{-1}. In an InAsSb QW at the lowest temperatures we can resolve ν=12\nu= 12 and ν=8\nu= 8 in ρxx\rho_{xx} and ρxy\rho_{xy}, respectively. This is considerably lower than for our high-mobility InSb QW where ν=42\nu= 42 was observed. As the condition μB≫1\mu B \gg1 must be met in order to observe quantum oscillations, one would expect the onset of SdH oscillations to be at B∼2.4 TB \sim2.4\ \mathrm{T}. This prediction is in good agreement with the experimental data presented in Figure 1a. On the other hand, spin-splitting occurs at a magnetic field of 3.2 T3.2\ \mathrm{T} corresponding to ν=7\nu= 7. This difference may be attributed to alloy disorder scattering, which is intrinsic for ternary compounds, or a possible alloy inhomogeneity in the QW [19]. In Figure 1b, we extract the quantum Hall activation energies of different filling factors, which will be discussed in greater detail in the upcoming section.

Effective mass, g-factor and lifetime analysis

We start our investigation of the effective mass by introducing the Ando’s expression for the quantum oscillation amplitude in 2D systems [27]

Δρxx(B,T)ρ0(B)∝AT(B,T)AD(B)As,(1)\frac{\Delta\rho_{xx}(B,T)}{\rho_{0}(B)} \propto A_T(B,T)A_D(B)A_s, \tag*{(1)}

where

AT(B,T)=2Cπ2kBTℏωcsinh⁡(2π2kBTℏωc),(2)A_T(B,T) = \frac{2C\pi^{2}k_B T}{\hbar\omega_c \sinh\left(\frac{2\pi^{2}k_B T}{\hbar\omega_c}\right)}, \tag*{(2)}
AD(B)=exp⁡(−πωcτq),(3)A_D(B) = \exp\left(\frac{-\pi}{\omega_c\tau_q}\right), \tag*{(3)}
As=cos⁡[2π(g∗m∗2me−12)],(4)A_s = \cos\left[2\pi\left(\frac{g^{*}m^{*}}{2m_e}-\frac{1}{2}\right)\right], \tag*{(4)}

ωc=eB/m∗\omega_c = eB/m^{*} is the cyclotron frequency and CC is a constant related to the oscillation amplitude at 0 K. In order to extract the effective mass from the temperature dependent suppression of the SdH oscillations, we first subtract a smooth magneto-resistance background ρ0\rho_0 from the ρxx\rho_{xx} data shown in Figure 1a, giving us the actual quantum oscillation amplitude Δρxx\Delta\rho_{xx} (see Figure 2a). Figure 2b shows the suppression of some local maxima and minima of Δρxx\Delta\rho_{xx} with increasing temperature, which was fitted with AT(B,T)A_T(B,T) to obtain an effective mass of m∗≈0.022mem^{*} \approx0.022m_e (see inset Figure 2b). In comparison, the effective mass of the InSb QW (m∗≈0.028mem^{*} \approx0.028m_e) was larger, confirming that alloying InSb with InAs indeed reduces the mass (see inset Figure 4b in the SI [24]).

Temperature dependence of the SdH oscillations in an InAsSb QW and temperature dependence of quantum oscillation amplitudes at different magnetic fields

Figure 2. (a) Temperature dependence of the SdH oscillations in an InAsSb QW after subtraction of a magnetoresistance background. Shown is only the field range were no spin splitting is observed. (b) Temperature dependence of quantum oscillation amplitudes at different magnetic fields. The solid lines represent fits to Eq. 1 and the effective mass deduced from them is depicted in the inset.

Quantum lifetime determined from the linear fit of the exponential Dingle term

Figure 4. Quantum lifetime determined from the linear fit of the exponential Dingle term in the Ando formula for the Shubnikov-de Haas effect in an InAsSb QW.

It should be noted that both masses exceed what is reported in literature for a similar InSb heterostructure, where m∗≈0.017mem^* \approx0.017m_e while the well width was comparable [17]. We attribute this to the fact that asymmetric doping can skew the potential well, effectively reducing its width and resulting in a higher mass due to stronger confinement. In the upcoming section we will give a more comprehensive analysis of the confinement enhancement based on band structure considerations

Alternatively, one can also extract the effective mass from FIR transmission experiments, by tracking the CR as function of magnetic field. As shown in Figure 3a we observe clear dips in the transmittance, the signature of a CR, between 1-4 T before the sample becomes opaque due to the InSb phonon absorption band. The position of the resonance was determined by fitting the spectra with a Lorentzian and was plotted against magnetic field in Figure 3b. By performing a linear fit of ECRE_{\mathrm{CR}} vs BB, we can determine the effective mass, which was m∗≈0.022mem^* \approx0.022m_e. This result is in perfect agreement with the mass obtained from the temperature dependent SdH measurements.

Transmittance as a function of energy and cyclotron resonance energy as a function of magnetic field

Figure 3. (a) Transmittance as a function of energy for different magnetic fields in an InAsSb QW. The traces have a constant offset of 0.2. In the region between 3–6 meV the signal is fairly noisy due to lower signal strength and the pick up of electronic/mechanical noise. Due to the Reststrahlen band, the sample becomes completely opaque between 22–24 meV [26]. The arrows indicate the position of the CR determined from the Lorentzian fits (gray dashed line). (b) Cyclotron resonance energy as function of magnetic field.

It should also be noted that the CRs are quite broad, considering that the width goes from 2.4 to 4.3 meV due to the interaction with the phonon band. These line widths correspond to a scattering time of the order of 0.1 ps, which is comparable to the Drude scattering time determined below. For comparison, the width of the resonances in the InSb QW only increases from 1 to 1.7 meV within this field range (see Figure 6a in the SI [24]), corroborating our assumption that alloying introduces

Schematic representation of a coincidence measurement

Figure 6. Schematic representation of the principle behind a coincidence measurement, where one can see the opening and closing of the energy gap between even and odd Landau states. By including the effect of gg-factor enhancement, one can see that coincidences (grey arrows) no longer occur at the same fields. Above the Landau fan diagram one can see cross-sections of the density-of-states at different angles, which correspond to the shaded areas between the Landau levels.

additional short range disorder as the anion atoms will be randomly distributed over the available sublattice sites, which is in line with the lower mobility.

In Figure 4 we extract the quantum lifetime τq\tau_q from the slope of the linear fit of 1/B1/B vs ln⁡(Δρxxρ0AT(B,T))\ln\left(\frac{\Delta\rho_{xx}}{\rho_0} A_T(B,T)\right). Using the effective mass calculated above, τq\tau_q was found to be 0.027 ps and was constant up to at least 20 K (see Figure 4). The Drude scattering time τD=0.53\tau_D = 0.53 ps, which is determined with the classical Drude model, has the same order of magnitude as the scattering time extracted from the line width of the CRs. Knowing τD\tau_D enables us to calculate the Dingle ratio as τD/τq≈20\tau_D/\tau_q \approx20. The fact that this value is significantly larger than 1, means that long-range potential fluctuations are the dominant scattering mechanism [28].

As the temperature dependence of the longitudinal conductivity σxx\sigma_{xx} at integer filling factors follows σxx∝exp⁡(−Eact/kBT)\sigma_{xx} \propto\exp(-E_{\mathrm{act}}/k_BT), we can determine the activation energy Eact,νE_{\mathrm{act},\nu} of the individual filling factors. From the linear fits presented in the inset of Figure 1a, we derive that Eact,(1,2,4,6)E_{\mathrm{act},(1,2,4,6)} are 28.2, 23.4, 7.4 and 3.2 meV, respectively. By realizing that for odd filling factors Eact,odd=g∗μBB−ΓE_{\mathrm{act,odd}} = g^*\mu_B B - \Gamma and neglecting the effect of the LL broadening Γ\Gamma, we can calculate the effective gg-factor. Therefore, at filling factor ν=1\nu= 1 we have g∗∼21.9g^* \sim21.9. Plotting Eact,evenE_{\mathrm{act,even}} as a function of BB and fitting this linear relation with the expression Eact,even=ℏωc−g∗μBB−ΓE_{\mathrm{act,even}} = \hbar\omega_c - g^*\mu_B B - \Gamma, gives us an alternative method to determine the effective mass as well as a rough estimate of the broadening (see Figure 5a in the SI [24]). Due to the fact that we underestimate g∗g^{*}, the effective mass m∗≈0.029mem^{*} \approx0.029m_{\mathrm{e}} is slightly larger than the one obtained from the temperature dependence of the SdH oscillations and FIR transmission. The LL broadening Γ\Gamma, given by the offset of Eact,evenE_{\mathrm{act,even}} vs BB, is found to be 7.4 meV. Here, we did not take into account any magnetic field dependence of Γ\Gamma.

Quantum well confinement as a function of Sb content

Figure 5. Assuming an effective mass of 0.022me0.022m_e, we calculated the quantum well confinement with the three and five band k⋅pk \cdot p theory as function of the Sb content in InAs1−x_{1-x}Sbx_x. A schematic depiction of the QW potential in the shaded area was included, showing the stronger confinement.

Next, we compare the activation energies derived above, with the temperature where the QHE disappears. In literature there are reports that the thermal energy at which the Hall resistivity is still quantized is roughly eight times larger than the energy gap between the levels [29]. Based on this phenomenological condition, we would expect the QHE to survive up to 41 or 34 K for ν=1\nu= 1 and 2, respectively. Our magnetotransport measurements show that quantized quantum Hall plateaus survive up to 40 K, whereas signatures of the effect itself are present till 60 K (see Figure 1a). Consequently, we can conclude that the robustness of the QHE is nicely in line with the activation energies of the lower filling factors.

Determination of confinement

Next, with calculations based on multiband k⋅pk \cdot p theory, we quantify confinement due to potential well skewing. First of all, we have to realize that the effective band gap EgE_{\mathrm{g}} will be increased due to the confinement of the QW, which, in the infinite potential approximation amounts to Econ=π2ℏ22L2mbulk∗E_{\mathrm{con}} = \frac{\pi^{2}\hbar^{2}}{2L^{2}m^{*}_{\mathrm{bulk}}}, where LL is the QW width and mbulk∗m^{*}_{\mathrm{bulk}} the effective mass of the bulk material. By only including the first order k⋅pk \cdot p terms, one uses a three-level model neglecting all other bands except the conduction and valence band, giving us an expression as in Ref. 28,

mem∗=1+2meP2ℏ2(Eg+Econ).(5)\frac{m_{\mathrm{e}}}{m^{*}} = 1 + \frac{2m_{\mathrm{e}}P^{2}}{\hbar^{2}(E_{\mathrm{g}} + E_{\mathrm{con}})} . \tag*{(5)}

If all k⋅pk \cdot p terms of the five closest bands at the Γ\Gamma point including the spin orbit interaction are taking into account, one can express the effective mass as follows,

mem∗=1+132meP2ℏ2(2Eg+Econ+1Eg+Δ0+Econ)\frac{m_{\mathrm{e}}}{m^{*}} = 1 + \frac{1}{3}\frac{2m_{\mathrm{e}}P^{2}}{\hbar^{2}}\left(\frac{2}{E_{\mathrm{g}} + E_{\mathrm{con}}} + \frac{1}{E_{\mathrm{g}} + \Delta_{0} + E_{\mathrm{con}}}\right)
−132meP′2ℏ2(2Eg′−Eg+Δ0′+1Eg′−Eg),(6)- \frac{1}{3}\frac{2m_{\mathrm{e}}P'^{2}}{\hbar^{2}}\left(\frac{2}{E'_{\mathrm{g}} - E_{\mathrm{g}} + \Delta'_{0}} + \frac{1}{E'_{\mathrm{g}} - E_{\mathrm{g}}}\right), \tag*{(6)}

where EgE_{\mathrm{g}}, Eg′E'_{\mathrm{g}}, Δ0\Delta_{0}, Δ0′\Delta'_{0}, PP and P′P' are band-edge parameters which can be found in literature [15, 28]. All band-edge parameters have a quadratic dependence on the alloy composition, which deviates from the linear interpolation between two binary compounds by defining a so-called Bowing parameter [15].

Figure 5 shows the dependence of the confinement on Sb content xx when m∗=0.022mem^{*}=0.022m_{\mathrm{e}} for both the 3- and 5-band model. Here we assumed that the Bowing parameter for EgE_{\mathrm{g}} is similar to that of Eg′E_{\mathrm{g}}', PP and P′P'. In both cases, we see an abrupt increase in confinement as xx is increased from 0 to 0.1, after which the dispersion starts to flatten off and eventually decreases again. EDX scans, shown in Figure 2 of the SI [24], confirm the presence of both As and Sb, which in conjunction with the fairly flat dispersion around the desired Sb content of 0.62, suggest that the confinement length will most likely lie between 8 and 10 nm. This rather strong confinement would also explain the relatively high effective masses we have reported earlier on.

Coincidence measurement

In this section we will discuss the interaction-induced enhancement of the gg-factor as function of spin polarization. Due to this enhancement, the gg-factor can become much larger than predicted by k⋅p\mathbf{k}\cdot\mathbf{p} theory in the single particle picture [16]. To investigate this effect, we have performed coincidence measurements. The principle of a coincidence measurement, is to change the ratio between the LL spacing and the Zeeman energy

by varying the angle θ\theta between the direction of sample normal and total applied magnetic field BtotB_{\mathrm{tot}} (see inset Figure 7a). Here we utilize the fact that the LL energy ℏωc=ℏeB⊥m∗\hbar\omega_{\mathrm{c}}=\frac{\hbar e B_{\perp}}{m^{*}} is proportional to the perpendicular field component B⊥B_{\perp}, whereas the Zeeman energy depends on BtotB_{\mathrm{tot}}. Let us define the parameter rr, which is the ratio between these two energies, i.e., r=g∗μBBtot/ℏωcr=g^{*}\mu_{\mathrm{B}}B_{\mathrm{tot}}/\hbar\omega_{\mathrm{c}}. Since B⊥=cos⁡(θ)BtotB_{\perp}=\cos(\theta)B_{\mathrm{tot}} and using the definition of the Bohr magneton (μB=eℏ/2me\mu_{\mathrm{B}}=e\hbar/2m_{\mathrm{e}}), we can rewrite this expression as rcos⁡(θ)=g∗m∗/2mer\cos(\theta)=g^{*}m^{*}/2m_{\mathrm{e}}. For the case r=1r=1, the Zeeman splitting g∗μBBtotg^{*}\mu_{\mathrm{B}}B_{\mathrm{tot}} equals the LL spacing ℏωc\hbar\omega_{\mathrm{c}}, meaning that for even filling factors the Landau states with opposite spins have the same energy, closing the energy gap between them. A similar reasoning holds for r=2r=2, but now the spin splitting is twice as large as the LL spacing, resulting in the suppression of the odd filling factors. Thus, the longitudinal resistivity minima appear only at odd or even filling factors for r=1r=1 and r=2r=2, respectively. The opening and closing of the energy gaps is schematically depicted in Figure 6. fig:7a shows the result of our coincidence measurement, where we measured ρxx\rho_{xx} as a function of B⊥B_{\perp} at angles ranging from θ=53∘\theta=53^\circ to 79∘79^\circ. At each filling factor, indicated by the dashed lines, we monitor the evolution of ρxx\rho_{xx} as the angle is increased. The cross-sections of these traces are plotted in Figure 7b and 7c for odd and even filling factors, respectively. It becomes apparent from the position of the maxima that the gg-factor is not universal for the individual filling factors, as the coincidences occur at different angles. With the expression for rr we can relate these coincidences to g∗m∗g^{*}m^{*}, which is proportional to the spin susceptibility χ=g∗m∗/2πℏ\chi=g^{*}m^{*}/2\pi\hbar in 2D systems [30], or equivalently to g∗g^{*} after normalizing it with m∗m^{*}. Figure 8 shows the dependence of g∗m∗g^{*}m^{*} on the spin polarization PP, defined as P=r/νP=r/\nu. The data displays a monotonic increase with spin polarization and g∗g^{*} can become as large as 70 when the system is fully spin polarized at P=1P=1. Such enhancement of the gg-factor was also reported in Ref. [16], where an undoped InSb QW was investigated. The linear increase of g∗g^{*} can be expressed as g∗=g0∗+Eex0P/μBBtotg^{*}=g_{0}^{*}+E_{\mathrm{ex}}^{0}P/\mu_{\mathrm{B}}B_{\mathrm{tot}}, where g0∗g_{0}^{*} is the bare gg-factor and Eex0E_{\mathrm{ex}}^{0} is the exchange parameter [31]. Eex0E_{\mathrm{ex}}^{0} initially increases as one increases PP, but starts to saturate around 70 meV at the highest spin polarization. For comparison, in graphene the exchange parameter amounts to 11.2 meV at 10 T, whereas in the InAsSb QW it is 28.3 meV at 10.5 T, which points towards stronger electron-electron interactions in this system [31]. At the same time we see that the value for the gg-factor obtained from the activation energies when the field was perpendicular to the sample, should be treated as the bare gg-factor considering that g∗=24.1g^{*}=24.1 at P=0P=0. Next, we will compare this result to k⋅pk\cdot p theory, where g0∗g_{0}^{*} is given by

Angle dependence of $\rho_{xx}$ and $\rho_{xx}$ as a function of $\theta$

Figure 7. (a) Angle dependence of ρxx\rho_{xx} in an InAsSb QW, where the angle was determined by aligning the ν=5\nu= 5 minima. The curves have a constant offset of 750 Ω\Omega. (b) and (c) Show ρxx\rho_{xx} as a function of θ\theta for odd and even filling factors, respectively.

Plots of $g^{*}m^{*}$ and exchange energy as functions of $P$

Figure 8. (a) g∗m∗g^{*}m^{*} as a function of PP. Using the effective mass obtained by from the mass analysis, m∗=0.022mem^{*}=0.022m_{\mathrm{e}}, one can determine the effective gg-factor. The corresponding g∗g^{*} values are shown on the right-hand side axis. (b) Eex0E_{\mathrm{ex}}^{0} as function of PP. The exchange parameter was extracted from the linear fit of g∗g^{*}.

g0∗=2−232meP2ℏ2(1Eg+Econ+1Eg+Δ0+Econ)+232meP′2ℏ2(1Eg′−Eg+1Eg′−Eg+Δ0′),(7)\begin{aligned} g_{0}^{*} ={}& 2-\frac{2}{3}\frac{2m_{\mathrm{e}}P^{2}}{\hbar^{2}}\left(\frac{1}{E_{\mathrm{g}}+E_{\mathrm{con}}}+\frac{1}{E_{\mathrm{g}}+\Delta_{0}+E_{\mathrm{con}}}\right) \\ &+\frac{2}{3}\frac{2m_{\mathrm{e} }P'^{2}}{\hbar^{2}}\left(\frac{1}{E'_{\mathrm{g}}-E_{\mathrm{g}}}+\frac{1}{E'_{\mathrm{g}}-E_{\mathrm{g}}+\Delta'_{0}}\right), \tag*{(7)} \end{aligned}

such as in Ref. 28. From this expression we get g∗=39.9g^{*}=39.9 if we take L=8L=8 nm and x=0.62x=0.62. Like in the previous section, the band-edge parameters were determined using a quadratic interpolation between the two binary compounds. The obtained gg-factor is larger than the experimentally determined value. This could be explained by the uncertainty in the two independent parameters, as g∗g^{*} strongly depends on the values used for the calculation.

Finally, we return to our discussion of the QHE observed at rather large temperatures and propose scenarios to increase this temperature even further. The large splitting between spin levels within one Landau level, will significantly reduce the activation energy of even filling factors and with it also the temperature where the QHE disappears. The robustness of the QHE can be increased by placing the sample under the coincidence condition r=1r=1 for filling factor ν=2\nu=2, where by definition the energy gap for ν=1\nu=1 is equal to the full LL spacing. In this particular case the energy gap is 58.4 meV, which would suggest that the QHE disappears at ∼85\sim85 K if we take into account the condition Eact≫kBTE_{\mathrm{act}}\gg k_{\mathrm{B}}T. In our case the carrier concentration is too high to reach this condition, but in theory this method can be used to increase the robustness. To illustrate this principle, we can again determine the activation energies of the QHE at r=1r=1 for filling factor ν=2\nu=2. But one can already see in Figure 9 that the energy gaps between the Landau states for even filling factors are significantly reduced, whereas the odd filling have become more robust. Eact,(2,3,4,5,6,7)E_{\mathrm{act},(2,3,4,5,6,7)} are found to be 0.54, 10.4, 0.42, 3.9, 0.082 and 1.1 meV, respectively. From the linear fit of Eact,oddE_{\mathrm{act},odd} vs BB in Figure 5b of the SI we get g∗≈37.8g^{*}\approx37.8 and Γ≈5.85\Gamma\approx5.85 meV [24]. The same can be done for the even filling factors, giving us m∗≈0.024mem^{*}\approx0.024m_{\mathrm{e}}, which is in good agreement with the value extracted from the temperature dependent SdH measurement (see Figure 5c in the SI [24]).

Magnetic field dependence of $\rho_{xx}$ and $\rho_{xy}$ at different temperatures

Figure 9. Magnetic field dependence of ρxx\rho_{xx} and ρxy\rho_{xy} in an InAsSb QW at different temperatures when θ=60.48∘\theta=60.48^\circ. This angle correspond to r=1r=1 for ν=2\nu=2.

Conclusions

In summary, we have shown that by alloying InSb with InAs we can reduce the effective mass. The masses extracted from the CRs are in good agreement with those obtained from the temperature dependent SdH measurements. The fact that the masses are larger than the bulk values can be ascribed to a strong 2D confinement. The QHE is observed up to 60 K, which is in line with the obtained activation energies. The robustness of the QHE could be drastically improved by increasing the confinement length and bringing the sample into coincidence at r=1r=1 for filling factor ν=2\nu=2. Though, for the latter we would need to decrease the carrier concentration to still be able to reach ν=1\nu=1 with our current setup. Also, an interaction-induced enhancement of the gg-factor was observed with the coincidence method. At the largest spin polarization, gg can be as large as 60. From the linear fit of the gg-factor as function of spin polarization we can extract values for the exchange parameter, which are significantly larger than those reported for graphene at comparable magnetic fields. To conclude, we have presented important proof-of-principle experiments on InAsSb QWs, a promising system with the potential of realising a low mass, high gg-factor system.

Acknowledgments

This work was supported by HFML-RU/NWO-I, member of the European Magnetic Field Laboratory (EMFL). It is part of the research programme “HFML-FELIX: a unique research infrastructure in the Netherlands. Matter under extreme conditions of intense infrared radiation and high magnetic fields” with project number 184.034.022 financed by the Dutch Research Council (NWO). We also acknowledge financial support from the Swiss National Science Foundation (SNSF) and the National Center of Competence in Science ”QSIT-Quantum Science and Technology”.

Supplemental information

Sample characterization

The layer stack of InSb, along with the InAsSb quantum wells (QWs), is illustrated in Figure 11. Further insights into the characteristics of the InSb QW can be found in the works of Christian Lehner [22, 23]. Utilizing these investigations, adjustments were made to the buffer structure of the InAsSb QW sample in order to achieve a nearly lattice-matched InAlSb barrier enveloping the QW. For the intended InAs0.38_{0.38}Sb0.62_{0.62} QW, this adaptation resulted in an In0.46_{0.46}Al0.54_{0.54}Sb/In0.58_{0.58}Al0.42_{0.42}Sb buffer structure.

Layer stack of the InSb QW structure and the InAsSb QW structure

Figure 11. Layer stack of the InSb QW Structure (a) (with x=0.1,y=0.3) and the InAsSb QW structure (b) (with x=0.42, y=0.54).

The fabrication of the samples took place using a molecular beam epitaxy (MBE) Veeco Gen II system, with As and Sb being supplied via valved cracker sources. The temperatures of the As and Sb crackers were set at 750 °C and 700 °C, respectively. A pyrometer was employed to measure the substrate temperature. The growth rates of group III elements were determined through the analysis of Reflection High-Energy Electron Diffraction (RHEED) oscillations. The V/III ratios were then deduced from these growth rates of group V and group III elements.

The growth rate of As (Sb) was derived by tracking As (or Sb) oscillations using the following method: Initially, the As (Sb) flux was terminated, and approximately 10 monolayers (MLs) of Ga were deposited onto the surface of GaAs (GaSb) at a rate of 0.5 Å/s and a temperature of 570 °C (480 °C). Immediately thereafter, As (Sb) was introduced at a selected valve setting, resulting in distinct and well-defined RHEED oscillations. For the InAsSb QW, an As (Sb) flux correspond-

ing to a GaAs (GaSb) growth rate of 0.7 Å/s (4.8 Å/s) was employed. The In0.46_{0.46}Al0.54_{0.54}Sb/In0.58_{0.58}Al0.42_{0.42}Sb buffer and the InAsSb QW were grown with a constant InAs (InSb) growth rate of 1.4 Å/s (1.7 Å/s) at a substrate temperature of 430 °C.

The composition of InAsx_xSb1−x_{1-x} layers is influenced not only by the appropriate As/Sb flux ratio but also by other growth conditions, including the growth temperature and the specific As and Sb species from the valved cracker sources [22]. To ensure accurate composition control, thicker InAsx_xSb1−x_{1-x} layers on an In1−y_{1-y}Aly_ySb/In1−x_{1-x}Alx_xSb buffer were utilized as a reference, and their composition was calibrated using (004) X-ray diffraction measurements.

Figure 10a displays X-ray rocking curves of the final InAsSb QW heterostructure in the symmetric (004) direction along with the complementary directions (115−) and (115+). The anticipated peak of the InAs0.38_{0.38}Sb0.62_{0.62} QW either coincides with the peak of the In0.58_{0.58}Al0.42_{0.42}Sb layer or exhibits a weak signal due to the relatively thin nature of the QW in comparison to the buffer structure. A simulation of the X-ray rocking curve for this heterostructure (Figure 10b) confirms the validity of this assumption. The faint satellite maxima emerge due to In0.46_{0.46}Al0.54_{0.54}Sb/In0.58_{0.58}Al0.42_{0.42}Sb superlattices in the buffer (SL2) and an AlSb/GaSb superlattice in the GaSb layer (SL1). The InAs0.38_{0.38}Sb0.62_{0.62} QW is examined via TEM and EDX. The EDX linescan qualitatively confirms the presence of As atoms in the QW layer.

X-ray diffraction measurements, simulation, and transmission electron microscopy image of the InAsSb quantum well

Figure 10. (a) The x-ray diffraction measurements depict symmetric (004), asymmetric (115-), and (115+) scans. These X-ray rocking curves reveal the diffraction peaks of the GaSb substrate, including the AlSb/GaSb superlattice (SL1), the In0.46_{0.46}Al0.54_{0.54}Sb/In0.58_{0.58}Al0.42_{0.42}Sb buffer, and satellite peaks of the In0.46_{0.46}Al0.54_{0.54}Sb/In0.58_{0.58}Al0.42_{0.42}Sb superlattice (SL2). The peak corresponding to the InAsSb quantum well (QW) either coincides as expected with the In0.58_{0.58}Al0.42_{0.42}Sb layer peak or is relatively weak. The asymmetric rocking curves reveal that the In0.58_{0.58}Al0.42_{0.42}Sb layer is slightly strained (with 95% relaxation). (b) A simulation of the symmetric (004) X-ray rocking curve of the InAs0.38_{0.38}Sb0.62_{0.62} QW heterostructure is presented. (c) A transmission electron microscopy (TEM) image demonstrates the InAsSb QW surrounded by InAlSb wells. Qualitative evidence of As atoms in the QW layer is substantiated by the energy-dispersive X-ray (EDX) linescan.

A qualitative assessment of the strain in the QW and the buffer can be inferred from measurements using the asymmetric diffraction geometry. The non-coincident maxima positions of In0.58_{0.58}Al0.42_{0.42}Sb indicate a slightly strained pseudomorphically grown QW (95% relaxation). Conversely, the maxima of In0.46_{0.46}Al0.54_{0.54}Sb overlap, indicating relaxation in these layers.

Effective mass analysis InSb

fig:12a shows the magnetic field dependence of a 405 × 25 μm2\mu\mathrm{m}^2 section of the InSb QW Hall bar at different temperatures. At the lowest temperatures we can resolve ν=42\nu= 42 and ν=22\nu= 22 in ρxx\rho_{xx} and ρxy\rho_{xy}, respectively. This is in good agreement with the condition μB≫1\mu B \gg1, which predicts that the onset of SdH oscillations will occur at 0.3 T, whereas in the ρxx\rho_{xx} trace at 1.4 K they start around 0.44 T.

After subtracting a magneto-resistance background ρ0\rho_0, we get the actual quantum oscillation amplitude Δρxx\Delta\rho_{xx} (see Figure 13a). At this point we again monitor the oscillation amplitude as function of temperature and fit the resulting points with AT(B,T)A_T(B,T) to determine the value of the effective mass, which was m∗≈0.028mem^* \approx0.028m_e (see inset Figure 13b).

Temperature dependence of the SdH oscillations in an InSb QW and temperature dependence of quantum oscillation amplitudes

Figure 13. (a) Temperature dependence of the SdH oscillations in an InSb QW after subtraction of a magnetoresistance background. Shown is only the field range were no spin splitting is observed. (b) Temperature dependence of quantum oscillation amplitudes at different magnetic fields. The data was fitted with AT(B,T)A_T(B,T), which describes the thermal damping of quantum oscillations, to obtain the effective mass (see inset).

The activation energy, which was determined from the longitudinal resistivity ρxx\rho_{xx} at integer filling factors, for ν=1\nu= 1 and ν=2\nu= 2 are 26.8 and 21.4 meV, respectively (see Figure 12b). Based on these activation energies we would expect the QHE to survive up to 39 or 31 K for ν=1\nu= 1 and ν=2\nu= 2, respectively. Our magnetotransport measurements show that quantized quantum Hall plateaus survive up to 40 K (see Figure 12a). Consequently, we can conclude that the robustness of the QHE is nicely in line with the activation energies of the lower filling factors. A direct comparison between the ρxy\rho_{xy} traces at 60 K for both types of QWs reveals that due to the smaller mass the QHE is more pronounced in InAsSb.

Magnetic field dependence of $\rho_{xx}$ and $\rho_{xy}$ in an InSb QW at different temperatures, and temperature dependence of $\sigma_{xx}$ at multiple filling factors

Figure 12. Magnetic field dependence of ρxx\rho_{xx} and ρxy\rho_{xy} in an InSb QW at different temperatures. (b) Temperature dependence of σxx\sigma_{xx} at multiple filling factors. Using the relation σxx∝exp⁡(−Eact/2kBT)\sigma_{xx} \propto\exp(-E_{\mathrm{act}}/2k_{\mathrm{B}}T), we extracted the activation energies from the linear fits.

Field-dependence of the activation energy in InAsSb

Figure 14. (a) Field-dependence of the activation energy in an InAsSb QW for even filling factors at θ=0∘\theta= 0^\circ. (b) and (c) Field-dependence of the activation energy in an InAsSb QW for odd and even filling factors at θ=60.5∘\theta= 60.5^\circ, respectively. The red lines represent the linear fits used to determine m∗m^* and g∗g^*.

Field-dependence activation energy InAsSb

As the activation energy of both even and odd filling factors have a linear dependence on field, if we neglect the field dependence of LL broadening, we can extract values for m∗m^* and g∗g^* from a linear fit of the data. fig:14a displays a fit of the even filling factors activation energies (ν=2\nu= 2, 4 and 6) when the sample is oriented perpendicular to the magnetic field. With the g∗g^* value calculated from Eact,1E_{\mathrm{act},1}, we can determine the effective mass from the slope and we find that m∗≈0.029mem^* \approx0.029m_e. The same can be done for θ=60.5∘\theta= 60.5^\circ, corresponding to the coincidence condition r=1r = 1 for ν=2\nu= 2. From the magnetic field dependence of the odd and even filling factor in fig:14b and c, respectively, we get g∗≈37.8g^* \approx37.8 and m∗≈0.024mem^* \approx0.024m_e.

FIR transmission InSb

The transmittance spectra contain CR features all the way up to 30 T (see Figure 15a). In the field range between 11-18 T there are two distinct CR dips present due to the non-equidistant LL spacing stemming from the non-parabolicity of the bands. This non-parabolicity can be clearly seen in the non-linear magnetic field dependence of the CR energy (see Figure 15b). As the transition probability for CRs depends on the occupancy of the Landau levels, which changes drastically with magnetic field as they are pushed through the Fermi level, certain transitions will only be observed within a specific field range. The effective mass can be determined from the low-field data in Figure 15b, where the dispersion is still linear, giving us m∗≈0.025mem^* \approx0.025m_e. This value is noticeably smaller than the value extracted from the transport data. Such a discrepancy could originate from the fact that the FIR transmission measurements were performed with a different piece of the wafer that distinctive values for the effective mass.

Transmittance and cyclotron resonance energy in InSb

Figure 15. (a) Transmittance as a function of wavenumber in an InSb QW for different magnetic fields. The traces have a constant offset of 0.2. The phonon absorption bands have been grayed out. In the field range between 11-18 T there are two CR features corresponding to different LL transitions. (b) Cyclotron resonance energy as function of magnetic field. Here, we used a multilayered Mylar beamsplitter for ν~=5\tilde{\nu} = 5 - 650 cm−1^{-1} and a KBr beamsplitter for ν~=400\tilde{\nu} = 400 - 1000 cm−1^{-1}.

References

References

  1. [1]K. v. Klitzing, G. Dorda, and M. Pepper, New method for high-accuracy determination of the fine-structure constant based on quantized hall resistance, Physical Review Letters 45, 494 (1980).
  2. [2]H. Störmer, R. Dingle, A. Gossard, W. Wiegmann, and R. Logan, Electronic properties of modulation-doped gaas-alxgal-xas superlattices, by BLH Wilson Inst. Phys., Bristol , 557 (1979).
  3. [3]J. A. A. Alexander-Webber, A. M. R. Baker, P. D. Buckle, T. Ashley, and R. J. Nicholas, High-current breakdown of the quantum hall effect and electron heating in insb/alinsb, Phys. Rev. B 86, 045404 (2012).
  4. [4]K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, M. I. Katsnelson, I. V. Grigorieva, S. Dubonos, and a. Firsov, Two-dimensional gas of massless dirac fermions in graphene, nature 438, 197 (2005).DOI
  5. [5]D. A. Bandurin, A. V. Tyurnina, G. L. Yu, A. A. Mishchenko, V. Zólyomi, S. V. Morozov, R. K. Kumar, R. V. Gorbachev, Z. R. Kudrynskiy, S. Pezzini, et al., High electron mobility, quantum hall effect and anomalous optical response in atomically thin insb, Nature nanotechnology 12, 223 (2017).DOI
  6. [6]L. Li, F. Yang, G. J. Ye, Z. Zhang, Z. Zhu, W. Lou, X. Zhou, L. Li, K. Watanabe, T. Taniguchi, et al., Quantum hall effect in black phosphorus two-dimensional electron system, Nature nanotechnology 11, 593 (2016).
  7. [7]F. Tang, P. Wang, M. He, M. Isobe, G. Gu, Q. Li, L. Zhang, and J. H. Smet, Two-dimensional quantum hall effect and zero energy state in few-layer zrt e5, Nano Letters 21, 5998 (2021).DOI
  8. [8]K. S. Novoselov, Z. Jiang, Y. Zhang, S. Morozov, H. L. Stormer, U. Zeitler, J. Maan, G. Boebinger, P. Kim, and A. K. Geim, Room-temperature quantum hall effect in graphene, science 315, 1379 (2007).
  9. [9]G. W. Small, B. W. Ricketts, P. C. Coogan, B. J. Pritchard, and M. M. R. Sovierszoski, A new determination of the quantized hall resistance in terms of the nml calculable cross capacitor, Metrologia 34, 241 (1997).DOI
  10. [10]A. Hartland, R. G. Jones, B. P. Kibble, and D. J. Legg, The relationship between the si ohm, the ohm at npl, and the quantized hall resistance, IEEE Transactions on Instrumentation and Measurement IM-36, 208 (1987).DOI
  11. [11]A.-M. Jeffery, R. Elmquist, L. H. Lee, J. Q. Shields, and R. F. Dziuba, Nist comparison of the quantized hall resistance and the realization of the si ohm through the calculable capacitor, IEEE transactions on instrumentation and measurement 46, 264 (1997).DOI
  12. [12]B. Jeckelmann and B. Jeanneret, The quantum hall effect as an electrical resistance standard, Reports on Progress in Physics 64, 1603 (2001).
  13. [13]T. Janssen, J. M. Williams, N. Fletcher, R. Goebel, A. Tzalenchuk, R. Yakimova, S. Lara-Avila, S. Kubatkin, and V. Fal'Ko, Precision comparison of the quantum hall effect in graphene and gallium arsenide, Metrologia 49, 294 (2012).DOI
  14. [14]R. Stradling, Magnetotransport and magnetooptics of inas _1-x sb _x and heterojunctions combinations formed between inas and gasb, alsb or inas _1-x sb _x, Acta Physica Polonica A 92, 851 (1997).
  15. [15]I. Vurgaftman, J. á. Meyer, and L. R. Ram-Mohan, Band parameters for iii–v compound semiconductors and their alloys, Journal of applied physics 89, 5815 (2001).
  16. [16]Z. Lei, E. Cheah, K. Rubi, M. E. Bal, C. Adam, R. Schott, U. Zeitler, W. Wegscheider, T. Ihn, and K. Ensslin, High-quality two-dimensional electron gas in undoped insb quantum wells, Phys. Rev. Res. 4, 013039 (2022).arxiv.org/abs/2110.11002
  17. [17]Z. Lei, C. A. Lehner, K. Rubi, E. Cheah, M. Karalic, C. Mittag, L. Alt, J. Scharnetzky, P. Märki, U. Zeitler, W. Wegscheider, T. Ihn, and K. Ensslin, Electronic q factor and magnetotransport in insb quantum wells, Phys. Rev. Res. 2, 033213 (2020).
  18. [18]J. C. Chokomakoua, N. Goel, S. J. Chung, M. B. Santos, J. L. Hicks, M. B. Johnson, and S. Q. Murphy, Ising quantum hall ferromagnetism in insb-based two-dimensional electronic systems, Phys. Rev. B 69, 235315 (2004).
  19. [19]S. Metti, C. Thomas, D. Xiao, and M. J. Manfra, Spin-orbit coupling and electron scattering in high-quality insb1−xasx quantum wells, Phys. Rev. B 106, 165304 (2022).DOI
  20. [20]Y. Jiang, M. Ermlaoev, S. Moon, G. Kipshidze, G. Belenky, S. Svensson, M. Ozerov, D. Smirnov, Z. Jiang, and S. Suchalkin, g-factor engineering with inassb alloys toward zero band gap limit, Phys. Rev. B 108, L121201 (2023).DOI
  21. [21]S. Svensson, W. Beck, W. Sarney, D. Donetsky, S. Suchalkin, and G. Belenky, Temperature dependent hall effect in inassb with a 0.11 ev 77 k-bandgap, Applied Physics Letters 114 (2019).DOI
  22. [22]C. A. Lehner, On the Limitations and Prospects of MBE grown high-mobility InSb quantum wells, Doctoral thesis, ETH Zurich, Zurich (2019).DOI
  23. [23]C. A. Lehner, T. Tschirky, T. Ihn, W. Dietsche, J. Keller, S. Fält, and W. Wegscheider, Limiting scattering processes in high-mobility insb quantum wells grown on gasb buffer systems, Phys. Rev. Mater. 2, 054601 (2018).DOI
  24. [24]"See Supplementary Information at URL_will_be_inserted_by_publisher for a description of the growth process and the analysis of the InSb QW."
  25. [26]Y. B. Li, S. S. Dosanjh, I. T. Ferguson, A. G. Norman, A. G. de Oliveira, R. A. Stradling, and R. Zallen, Raman scattering in inas1-xsbx alloys grown on gaas by molecular beam epitaxy, Semiconductor Science and Technology 7, 567 (1992).DOI
  26. [27]N. Miura, Physics of Semiconductors in High Magnetic Fields, Series on Semiconductor Science and Technology (OUP Oxford, 2007).
  27. [28]T. Ihn, Semiconductor Nanostructures: Quantum states and electronic transport (OUP Oxford, 2009).DOI
  28. [29]T. Khouri, M. Bendias, P. Leubner, C. Brüne, H. Buhmann, L. W. Molenkamp, U. Zeitler, N. E. Hussey, and S. Wiedmann, High-temperature quantum Hall effect in finite gapped hgte quantum wells, Phys. Rev. B 93, 125308 (2016).DOI
  29. [30]J. Zhu, H. L. Stormer, L. N. Pfeiffer, K. W. Baldwin, and K. W. West, Spin susceptibility of an ultra-low-density two-dimensional electron system, Phys. Rev. Lett. 90, 056805 (2003).DOI
  30. [31]E. V. Kurganova, H. J. van Elferen, A. McCollam, L. A. Ponomarenko, K. S. Novoselov, A. Veligura, B. J. van Wees, J. C. Maan, and U. Zeitler, Spin splitting in graphene studied by means of tilted magnetic-field experiments, Phys. Rev. B 84, 121407(R) (2011).DOI

Paper details

vlnc:W_sh57zjmy
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents