Pure electron-electron dephasing in percolative aluminum ultrathin film grown by molecular beam epitaxy
© Lin et al.; licensee Springer. 2015
Received: 12 December 2014
Accepted: 24 January 2015
Published: 18 February 2015
We have successfully grown ultrathin continuous aluminum film by molecular beam epitaxy. This percolative aluminum film is single crystalline and strain free as characterized by transmission electron microscopy and atomic force microscopy. The weak anti-localization effect is observed in the temperature range of 1.4 to 10 K with this sample, and it reveals that, for the first time, the dephasing is purely caused by electron-electron inelastic scattering in aluminum.
KeywordsWeak anti-localization Ultrathin metal films Pure electron-electron dephasing
Weak localization (WL) is the quantum correction to the conductance which occurs in weakly disordered systems due to coherent backscattering of electrons (or holes). As a results of spin-orbit coupling, weak anti-localization (WAL) may be observed in a weakly disordered electron (or hole) system . The theoretic derivation and experimental proof of WAL were extensively developed since 1980s, and the investigation on various materials in all dimensions has been a central topic in condensed matter physics for decades. In particular, a wide variety of experimental results of WAL were obtained in two-dimensional (2D) systems. 2D system is suitable for experimental study of WAL because of its stronger WAL contribution than three-dimensional (3D) ones and its easier sample fabrication than one-dimensional (1D) ones. Recently, due to its sensitivity to the electron dephasing and spin dephasing, WAL has been widely applied to studying the spin-orbit interaction in new materials, such as graphene, topological insulator, magnetic-doped semiconductor, and narrow-gap semiconductor, to evaluate the potential for spintronics devices [2-7].
In addition, the interplay between superconducting effect and WAL also attracts much attention and has been investigated extensively. Ebisawa et al. derived the relationship between the superconducting pair-breaking parameter δ and the inelastic scattering rate τ i −1 by δ = (πℏ/8k B T)τ i −1 . Their results have been used to study experiments of WAL in aluminum thin films [9-11]. However, it is a chanting challenge to grow an ideal 2D superconducting metallic sample. As the thickness of evaporated metal goes thinner, the discontinuity of the metal film un-avoidably appears due to large lattice mismatch between the template and metallic material as well as the surface non-uniformity of the bottom template. Previously, the reported metallic films were in the thickness of ten to a few tens of nanometers [9-14]. Although plenty of the theoretic works of WAL in two-dimensional systems have been published in the past few decades, the experimental proof toward WAL in an ideal two-dimensional metallic system is still lacking. In this work, we have used molecular beam epitaxy (MBE) system as the deposition technique to prepare ultrathin Al films. By using gallium-rich GaAs as the epi-template, we are able to successfully deposit ultrathin percolated Al film for studying WAL toward the 2D limit. Interestingly, we have observed a pure electron-electron dephasing in this sample over the whole temperature range that WAL effect exists. Note that all of our characterizations including structural and electrical assessments were fulfilled ex situ, and the continuity of the Al film remains even after the post-processing for the Hall device fabrication. Even though our sample is not thin enough to reach the ultimate two-dimensional limit such as monolayer graphene, our results still provide an experimental proof that Nyquist scattering becomes the dominant inelastic scattering mechanism at all temperatures when the system approaches an ideal two-dimensional one.
where β represents the interaction strength between the electron pair . Accordingly, β ~ 1/[ln(T/T c) − δ] at temperature close to T c . B T = 2k B Tln(T/T c)/πDe and D is the diffusion coefficient, D = 1/3v F l 0. Here, v F is the Fermi velocity and l 0 is the mean free path.
Bergmann experimentally studied WAL in various metallic thin films  and confirmed that the Maki-Thompson correction gives the most contribution of superconducting fluctuations and cannot be neglected even for temperatures far above T c. Therefore, in our following analysis, both WAL effect and Maki-Thompson term will be taken into account.
The epitaxial aluminum thin film was fabricated on a semi-insulating gallium arsenide (SI-GaAs) substrate in our Varian Gen II MBE system equipped with an arsenic cracking cell. First, a GaAs wafer was heated to 620°C for 20 min for de-oxidation under As flux, and then a 300-nm-thick undoped GaAs buffer layer was deposited at 590°C as the epitaxial template. The sample surface was turned into Ga-rich at 620°C in the absence of As flux. We kept the sample in high vacuum (approximately 3 × 10−10 Torr) until the background As vapor was pumped out. The Al film with an intentional thickness of 3 nm was then deposited at room temperature with a growth rate of 0.366 μm/h.
The 50-μm-wide Hall devices were processed using conventional lithography technique. Al was etched by 2% tetramethylammonium hydroxide (TMAH) for 20 s to transfer the Hall bar pattern from photoresist to our Al film. Ti/Au (30/300 nm) was deposited using e-gun evaporation as the contact electrode. The finished devices were wire-bonded on a ceramic carrier and loaded into 4He cryogenic system equipped with a superconducting magnet. A DC four-terminal electrical measurement has been performed in this work for studying WAL. We have used a Keithley 2602 multi-meter (Keithley Instruments Inc., Cleveland, OH, USA) as a constant current source as well as a voltage meter. Electrical current was set at 3 μA for clear signals and was low enough in order to avoid possible current heating effect. Voltage noise level was at about 5 μV.
Results and discussion
In our fitting procedure, B i and β were temperature-dependent fitting parameters. B so was chosen as a temperature-invariant constant because it is commonly accepted that the electron configuration in the half-filled conduction bands is insensitive to the temperature, so the angular momentum and the spin-orbit interaction remains unchanged at all temperature in metals [10-13]. In Figure 3, the solid lines represent theoretic fits of our experimental data at three temperatures as an example. It is clear that the theory can only be applied to small magnetic field. At higher temperatures, the fits can be extended to higher magnetic fields. This is because the Maki-Thompson term is only valid at B ≪ k B(T − T c)/4eD [15-17].
In our 3-nm-thick aluminum sample, weak anti-localization changes into weak localization when the temperature goes above 10 K. In the temperature range for existing WAL (T < 10 K), we found that 1/τ i is proportional to T, which agrees with the theoretic prediction of Nyquist scattering (dashed line) in the same order.
We would like to emphasize that our ultrathin Al film has a very high carrier concentration (n ~ 1029 m−3), so it is actually in mixed dimensions. That is, considering the density of states and diffusive motion, it is treated as a 3D system because the film thickness d is much larger than the Fermi wavelength λ F (d ≫ λ F ~ 10−1 nm) and its mean free path of electrons l 0 (d ≫ l 0 ~ 10−1 nm in our sample). However, in terms of electron dephasing, the film thickness is much less than the dephasing length l i (d ≪ l i ~ few tens of nanometers) so the system is two-dimensional. Previous experimental studies on clean and dirty Al thin films demonstrated that, only for the temperature near T c, Nyquist scattering was the major dephasing mechanism. When the temperatures were higher than about 4 K, electron-phonon scattering dominated the dephasing process [9-14]. In contrast, our results show a pure electron-electron dephasing all the way up 10 K. For a type I superconducting material, there is only one experimental result on clean titanium film exhibiting a full T −1 dependence of dephasing rate . However, our sample is in the dirty limit as the sheet resistivity of our sample is over 1.6 kΩ, at least one order higher than those all of their samples. On the other hand, for type II superconductors such as ZrRh and TaN, most experimental works did not observe the pure T −1 dependence of dephasing rate [26,27]. It is worth mentioning that the observation of Giannouri et al.  with their NbTa film is very similar to ours except that the sudden drop of dephasing rate at temperature approaching T c has not been seen even at 1.4 K in our experiment.
Dephasing process which purely comes from electron-electron scattering observed here is unusual but is relatively easy to be obtained in semiconductor-based 2D systems, such as 2D electron or hole gases (2DEG and 2DHG), formed by modulation-doped heterostructures. The carriers (electrons or holes) are confined by the triangular potential well caused by remote ionized dopants and limited in the few monolayers next to the heterostructure interface. The carrier concentration of 2DEG or 2DHG is controlled in the range of 1010 ~ 1012 cm−2, so the Fermi wavelength λ F easily exceeds the size of the confinement potential well and the system is quantized in this dimension to become two-dimensional. However, for metallic materials even down to a few nanometers, the carrier concentration is still as high as 1015 cm−2, which makes the λ F is much smaller than its thickness. Recently, the experimental results of WAL in various semiconductor heterostructures were published, including AlGaN/GaN, GaAs/InGaAs, InP/InGaAs, and AlGaAs/GaAs [29-35]. Most of these works demonstrated a full T −1 dependence of the dephasing rate up to 10 K and considered as consistent with the theoretic calculation of Nyquist scattering because of their 2D nature. In our Al film, the observed pure 2D-like dephasing indicates that the superconducting metallic film can be an ideal 2D system in inelastic process.
We wish to address the issue of zero-temperature saturation rate (1/τ i 0) that is the dephasing rate in the zero temperature limit. The origin of the saturation rate is still under debate for the time being . Some of the reported experimental works on superconducting metallic materials revealed a divergence dephasing rate when the temperature approached T c [9,10,13,26,27]. In contrast, such situations have not been observed in our sample that exhibits a full Nyquist scattering rate from liquid helium temperatures to 10 K, thus could be helpful for developing the related theory.
In Figure 4b, the electron-electron interaction strength parameter β in the Maki-Thompson correction term is plotted as a function of temperature. It is clear that β has a trend to diverge at temperature approaching T c, as expected with Larkin’s theory . The value of our sample decreases dramatically with the increasing temperature because the superconducting effect becomes less significant. It is noted that β does not vanish but converge to about 0.9 instead. A similar result was observed in ZrRh films . Further investigations on this non-zero β are certainly needed.
We note that the sheet resistivity of our sample is about 1.6 kΩ. The mean free path l 0 actually is 0.082 nm estimated by the Drude model, which is slight shorter than λ F. This appears to contradict to the basic assumption of the WAL theory that WAL occurs in a weakly disordered system (l 0 ≫ λ F). In a strongly disordered system, WAL should not be observed. We believe that, due to the percolative morphology of our sample, the actual electron path is much longer than the Hall bar size; therefore, the mean free path l 0 has been underestimated.
Similar to the WAL effect, the universal conductance fluctuations (UCFs) due to the electron (or hole) interference between two classical paths is closely related to the phase coherent length l i [37-40]. We have also examined the conductance fluctuations in our measured data, but no clear UCFs are observed, which could be due to the large size of our device and the limit of the measurement system noise (approximately 5 μV).
We have presented the structural and electrical characterization of the ultrathin percolating aluminum film grown by MBE. The TEM results indicate a superior crystal quality of the epitaxial aluminum film. WAL revealed by low-temperature magnetoresistance measurement showed its unusual dephasing mechanism. At all temperatures that WAL exists, a pure electron-electron scattering was observed, so the aluminum film behaves as an ideal two-dimensional system in this aspect. Based on this, we conclude that the MBE-grown aluminum films could achieve the two-dimensional limit of a superconducting metallic material.
We thank the financial support from the NSC and ATU program of MOE in Taiwan. The equipment help from the Center for Nano Science and Technology at the National Chiao Tung University and National Device Laboratory is highly appreciated. C-TL acknowledges the support from the MOST, Taiwan (grant numbers: MOST 103-2918-I-002-028 and MOST 102-2119-M-002-016-MY3).
- Hikami S, Larkin AI, Nagaoka Y. Spin-orbit interaction and magnetoresistance in two-dimensional random system. Prog Theor Phys. 1980;63:707.View ArticleGoogle Scholar
- Baker AMR, Alexander-Webber JA, Altebaeumer T, Janssen TJBM, Tzalenchuk A, Lara-Avila S, et al. Weak localization scattering lengths in epitaxial, and CVD graphene. Phys Rev B. 2012;86:235441.View ArticleGoogle Scholar
- Steiberg H, Laloë J-B, Fatemi V, Moodera JS, Jarillo-Herrero P. Electrically tunable surface-to-bulk coherent coupling in topological insulator thin films. Phys Rev B. 2011;84:233101.View ArticleGoogle Scholar
- Lang M, He L, Kou X, Upadhyaya P, Fan Y, Chu H, et al. Competing weak localization and weak antilocalization in ultrathin topological insulators. Nano Lett. 2013;13:48.View ArticleGoogle Scholar
- Neumaier D, Wagner K, Geißler S, Wurstbauer U, Sadowski J, Wegscheider W, et al. Weak localization in ferromagnetic (Ga, Mn)As nanostructures. Phys Rev Lett. 2007;99:116803.View ArticleGoogle Scholar
- Kallaher RL, Heremans JJ. Spin and phase coherence measured by antilocalization in n-InSb thin films. Phys Rev B. 2009;79:075322.View ArticleGoogle Scholar
- Kallaher RL, Heremans JJ, Goel N, Chung SJ, Santos MB. Spin-orbit interaction determined by antilocalization in an InSb quantum well. Phys Rev B. 2010;81:075303.View ArticleGoogle Scholar
- Ebisawa H, Maekawa S, Fukuwama H. Pair breaking parameter of two-dimensional dirty superconductors. Solid State Commun. 1983;45:75.View ArticleGoogle Scholar
- Gordon JM, Lobb CJ, Tinkham M. Divergent phase-breaking rate in aluminum films from magnetoresistance measurements. Phys Rev B. 1984;29:5232.View ArticleGoogle Scholar
- Gordon JM, Goldman AM. Electron inelastic scattering in aluminum films and wires at temperatures near the superconducting transition. Phys Rev B. 1986;34:1500.View ArticleGoogle Scholar
- Shinozaki B, Kawaguti T, Fujimori Y. Magnetoresistance near superconducting transition temperature of aluminum films. J Phys Soc Jpn. 1986;55:2364.View ArticleGoogle Scholar
- Santhanam P, Prober DE. Inelastic electron scattering mechanisms in clean aluminum films. Phys Rev B. 1984;29:3733.View ArticleGoogle Scholar
- Shinozaki B, Kawaguti T, Fujimori Y. Localization and interaction effects in weakly localized region in aluminum films. J Phys Soc Jpn. 1984;53:3303.View ArticleGoogle Scholar
- Santhanam P, Wind S, Prober DE. Localization, superconducting fluctuations, and superconductivity in thin films and narrow wires of aluminum. Phys Rev B. 1987;35:3188.View ArticleGoogle Scholar
- Maki K. Critical fluctuation of the order parameter in a superconductor. I Prog Theor Phys. 1968;40:193.View ArticleGoogle Scholar
- Thompson RS. Microwave, flux flow, and fluctuation resistance of dirty type-II superconductors. Phys Rev B. 1970;1:327.View ArticleGoogle Scholar
- Santos JMBLd, Abrahams E. Superconducting fluctuation conductivity in a magnetic field in two dimensions. Phys Rev B. 1985;31:172.View ArticleGoogle Scholar
- Larkin AI. Reluctance of two-dimensional systems. JETP Lett. 1980;31:219.Google Scholar
- Bergmann G. Quantum corrections to the resistance in two-dimensional disordered superconductors above T c : Al, Sn, and amorphous Bi0.9Tl0.1 films. Phys Rev B. 1984;29:6114.View ArticleGoogle Scholar
- Cho AY, Dernier PD. Single-crystal-aluminum Schottky-barrier diodes prepared by molecular-beam epitaxy (MBE) on GaAs. J Appl Phys. 1978;49:3328.View ArticleGoogle Scholar
- Ludeke R, Chang LL, Esaki L. Molecular beam epitaxy of alternating metal-semiconductor films. Appl Phys Lett. 1973;23:201.View ArticleGoogle Scholar
- Lin S-W, Wu J-Y, Lin S-D, Lo M-C, Lin M-H, Liang C-T. Characterization of single-crystalline aluminum thin film on (100) GaAs substrate. Jpn J Appl Phys. 2013;52:045801.View ArticleGoogle Scholar
- Ruffino F, Grimaldi MG. Island-to-percolation transition during the room-temperature growth of sputtered nanoscale Pd films on hexagonal SiC. J Appl Phys. 2010;107:074301.View ArticleGoogle Scholar
- Altshuler BL, Aronov AG, Khmelnitsky DE. Effects of electron-electron collisions with small energy transfers on quantum localization. J Phys C: Solid State Phys. 1982;15:7367.View ArticleGoogle Scholar
- Vangrunderbeek J, Van Haesendonck C, Bruynseraede Y. Electron localization and superconducting fluctuations in quasi-two-dimensional Ti films. Phys Rev B. 1989;40:7594.View ArticleGoogle Scholar
- Giannouri M, Rocofyllou E, Papastaikoudis C, Schilling W. Weak-localization, Aslamazov-Larkin, and Maki-Thompson superconducting fluctuation effects in disordered Zr1-xRhx films above T c . Phys Rev B. 1997;56:6148.View ArticleGoogle Scholar
- Breznay NP, Kapitulnik A. Observation of the ghost critical field for superconducting fluctuations in a disordered TaN thin film. Phys Rev B. 2013;88:104510.View ArticleGoogle Scholar
- Giannouri M, Papastaikoudis C, Rosenbaum R. Low-temperature transport properties of Nb1-xTax thin films. Phys Rev B. 1999;59:4463.View ArticleGoogle Scholar
- Lu J, Shen B, Tang N, Chen DJ, Zhao H, Liu DW, et al. Weak anti-localization of the two-dimensional electron gas in modulation-doped AlxGa1-xN/GaN heterostructures with two subbands occupation. Appl Phys Lett. 2004;85:3125.View ArticleGoogle Scholar
- Thillosen N, Schäpers T, Kaluza N, Hardtdegen H, Guzenko VA. Weak antilocalization in a polarization-doped AlxGa1-xN/GaN heterostructure with single subband occupation. Appl Phys Lett. 2006;88:022111.View ArticleGoogle Scholar
- Thillosen N, Cabañas S, Kaluza N, Guzenko VA, Hardtdegen H, Schäpers T. Weak antilocalization in gate-controlled Al1-xGaxN/GaN two-dimensional electron gases. Phys Rev B. 2006;73:241311(R).View ArticleGoogle Scholar
- Minkov GM, Sherstobitov AA, Germanenko AV, Rut OE, Larionova VA, Zvonkov BN. Antilocalization and spin-orbit coupling in the hole gas in strained GaAs/InxGa1-xAs/GaAs quantum well heterostructures. Phys Rev B. 2005;71:165312.View ArticleGoogle Scholar
- Minkov GM, Germanenko AV, Rut OE, Sherstobitov AA, Golub LE, Zvonkov BN, et al. Weak antilocalization in quantum wells in tilted magnetic fields. Phys Rev B. 2004;70:155323.View ArticleGoogle Scholar
- Studenikin SA, Coleridge PT, Ahmed N, Poole PJ, Sachrajda A. Experimental study of weak antilocalization effects in a high-mobility InxGa1-xAs/InP quantum well. Phys Rev B. 2003;68:035317.View ArticleGoogle Scholar
- Millo O, Klepper SJ, Keller MW, Prober DE, Xiong S, Stone AD, et al. Reduction of the mesoscopic conductance-fluctuation amplitude in GaAs/AlGaAs heterojunctions due to spin-orbit scattering. Phys Rev Lett. 1990;65:1494.View ArticleGoogle Scholar
- Lin JJ, Bird JP. Recent experimental studies of electron dephasing in metal and semiconductor mesoscopic structures. J Phys: Condens Matter. 2002;14:R501.Google Scholar
- Lee PA, Stone AD, Fukuyama H. Universal conductance fluctuations in metals: effects of finite temperature, interactions, and magnetic field. Phys Rev B. 1987;35:1039.View ArticleGoogle Scholar
- Li Z, Chen T, Pan H, Song F, Wang B, Han J, et al. Two-dimensional universal conductance fluctuations and the electron-phonon interaction of surface states in Bi2Te2Se microflakes. Sci Rep. 2012;2:595.Google Scholar
- Kahnoj SS, Touski SB, Pourfath M. The effect of electron-electron interaction induced dephasing on electronic transport in graphene nanoribbons. Appl Phys Lett. 2014;105:103502.View ArticleGoogle Scholar
- Weber B, Ryu H, Matthias Tan Y-H, Kilmeck G, Simmons MY. Limits to metallic conduction in atomic-scale quasi-one-dimensional silicon wires. Phys Rev Lett. 2014;113:246802.View ArticleGoogle Scholar
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.