Data Loading...
Electrostatic Conveyer for Excitons Flipbook PDF
Electrostatic Conveyer for Excitons A.G. Winbow, J.R. Leonard, M. Remeika, Y.Y. Kuznetsova, A.A. High, A.T. Hammack, and
113 Views
79 Downloads
FLIP PDF 366.32KB
PRL 106, 196806 (2011)
week ending 13 MAY 2011
PHYSICAL REVIEW LETTERS
Electrostatic Conveyer for Excitons A. G. Winbow, J. R. Leonard, M. Remeika, Y. Y. Kuznetsova, A. A. High, A. T. Hammack, and L. V. Butov Department of Physics, University of California at San Diego, La Jolla, California 92093-0319, USA
J. Wilkes, A. A. Guenther, and A. L. Ivanov Department of Physics and Astronomy, Cardiff University, Cardiff CF24 3AA, United Kingdom
M. Hanson and A. C. Gossard Materials Department, University of California at Santa Barbara, Santa Barbara, California 93106-5050, USA (Received 23 February 2011; published 12 May 2011) We report on the study of indirect excitons in moving lattices—conveyers created by a set of ac voltages applied to the electrodes on the sample surface. The wavelength of this moving lattice is set by the electrode periodicity, the amplitude is controlled by the applied voltage, and the velocity is controlled by the ac frequency. We found the dynamical localization-delocalization transition for excitons in the conveyers and determined its dependence on exciton density and conveyer amplitude and velocity. DOI: 10.1103/PhysRevLett.106.196806
PACS numbers: 73.63.Hs, 05.30.Jp, 78.67.De
An indirect exciton is a bound pair of an electron and a hole confined in spatially separated layers. Because of their long lifetimes, indirect excitons can travel over large distances before recombination [1–10]. Furthermore, indirect excitons have a built-in dipole moment ed, where d is the separation between the electron and hole layers, so their energy can be controlled by voltage: an electric field Fz normal to the layers results in the exciton energy shift U ¼ edFz [11]. This gives an opportunity to create inplane potential landscapes for excitons Uðx;yÞ ¼ edFz ðx;yÞ by laterally modulated voltage Vðx; yÞ. Excitons were studied in a variety of electrostatically formed potential landscapes: ramps [1,7], lattices [12–16], traps [15,17–20], and circuit devices [21–24]. In this Letter, we present an excitonic conveyer—a moving lattice created by a set of ac voltages. The excitonic conveyer realizes controlled transport of excitons as charged coupled devices (CCD) realize controlled transport of electrons [25]. Moving potential lattices can be created by surface acoustic waves (SAW). Transport of excitons, excitonpolaritons, and laterally separated electrons and holes via SAW is intensively studied [26–29]. The transport velocity in this case is defined by the sound velocity of SAW propagation 3 m=ns. In contrast, the velocity of the electrostatic excitonic conveyer can be controlled by the ac frequency and can be from well below to well above the sound velocity. The indirect excitons are created in a GaAs coupled quantum well structure (CQW) grown by molecular beam epitaxy [Fig. 1(a)]. An nþ -GaAs layer with nSi ¼ 1018 cm3 serves as a homogeneous bottom electrode. Two 8 nm GaAs QWs separated by a 4 nm Al0:33 Ga0:67 As barrier are positioned 0:1 m above the nþ -GaAs layer within an undoped 1 m thick Al0:33 Ga0:67 As layer. Positioning the CQW closer to the homogeneous electrode 0031-9007=11=106(19)=196806(4)
suppresses the in-plane electric field [15], which otherwise can lead to exciton dissociation [12]. The conveyer potential is created by a set of semitransparent 1 m wide 120 nm thick indium tin oxide (ITO) electrodes on the sample surface. The distance between the electrode centers is 2 m, the conveyer periodicity is 7 electrodes, and the wavelength of the conveyer potential is conv ¼ 14 m. The conveyer electrodes are covered by a
FIG. 1 (color online). (a) CQW band diagram. e, electron; h, hole. (b) Electrode pattern, contacts are shown by circles. (c)–(f) x-y and x-energy PL images for conveyer off and on. (g),(h) PL intensity profiles IðxÞ. Pex ¼ 20 W, Aconv ¼ 7:5 meV, vconv ¼ 0:7 m=ns.
196806-1
Ó 2011 American Physical Society
week ending 13 MAY 2011 4
b Exciton transport via conveyer, M1 (µm)
layer of transparent insulation (300 m thick SiO2 ). A set of connecting electrodes (10 m wide, 300 nm thick ITO) provides the contacts to the conveyer electrodes through 1 10 m etched openings in the insulating layer [Fig. 1(b)]. The conveyer length is 380 m, width is 80 m. The sample mounts in a He cryostat at 1.7 K. ac voltages to the conveyer electrodes are applied by coaxial cables with impedance-matching termination at the sample. The regime, where the indirect excitons have lower energy than spatially direct excitons in the CQW, is realized by dc bias Vbias ¼ 4 V supplied separately [30]. A set of differentially phase-delayed ac sine waves at frequency fconv creates a traveling potential lattice for indirect excitons—the excitonic conveyer. The amplitude of the conveyer potential for indirect excitons Aconv is controlled by the applied voltage [30]. The conveyer velocity is controlled by the ac frequency vconv ¼ conv fconv . The excitons are photoexcited by 788 nm Ti:sapphire laser focused to a spot 5 m in diameter. The exciton density is controlled by the laser excitation power Pex . Photoluminescence (PL) images of the exciton cloud are taken by a CCD with a bandpass filter 810 5 nm covering the spectral range of the indirect excitons. The diffractionlimited spatial resolution is 1:4 m. The spectra are measured using a spectrometer with resolution 0.18 meV. Figures 1(c)–1(f) show x-y and x-energy PL images for conveyer off and on. The PL intensity profiles IðxÞ, obtained by the integration of the x-energy images over the emission wavelength, are shown in Figs. 1(g) and 1(h). Exciton transport via conveyer is presented by the extension of the exciton cloud along the direction of the moving potential. We R quantify it R by the first moment of the PL intensity M1 ¼ xIðxÞdx= IðxÞdx, which characterizes the average transport distance of indirect excitons via conveyer. The spectrally broad emission at x ¼ 0 [Figs. 1(e) and 1(f)] and sharp peak in IðxÞ [Figs. 1(g) and 1(h)] originate from the bulk GaAs in the structure. To remove the contribution from the bulk, the shaded area is not included in the calculation of M1 in the analysis of exciton transport. Figure 2 presents exciton transport via conveyer as a function of the conveyer amplitude Aconv . For a shallow conveyer, the exciton cloud extension M1 is not affected by the conveyer motion indicating that the excitons do not follow the moving lattice, i.e., are dynamically delocalized in the lattice [Fig. 2(a)]. In contrast, at higher conveyer amplitudes, excitons are moved by the moving lattice, i.e., are dynamically localized in the lattice. At the dynamical localization-delocalization transition (DLDT), the exciton cloud starts to follow the conveyer and M1 changes from constant to increasing with Aconv . We define the conveyer amplitude at the DLDT, Aturn-on , as the point where the extrapolation of the growth of M1 to small Aconv becomes equal to the low-Aconv constant. The DLDT is a dynamical counterpart of the LDT for excitons in static lattices [16].
2 30 0 0
3
Αturn-on (meV)
PHYSICAL REVIEW LETTERS vconv=0.7 µm/ns
3.5 µm/ns
6
vconv (µm/ns) 15
6.3 µm/ns
0
a 0
2
4
6
8
Conveyer amplitude, Aconv (meV)
FIG. 2 (color online). (a) The average transport distance of indirect excitons via conveyer M1 as a function of the conveyer amplitude, Aconv . Lines are a guide to the eye, Aturn-on is Aconv at the line intersection. (b) The measured (points) and calculated (triangles) Aturn-on versus the conveyer velocity. Pex ¼ 20 W.
The control of fconv gives an opportunity to study exciton transport via conveyers and, in particular, the DLDT as a function of the conveyer velocity. Figure 2 shows that exciton transport via conveyer is less efficient for higher vconv . In particular, Aturn-on increases with vconv . The control of Pex gives an opportunity to study exciton transport via conveyers as a function of the exciton density. Figure 3(a) shows that excitons are hardly moved by the conveyer at low densities, efficient exciton transport via conveyer is achieved at intermediate densities, and exciton transport via conveyer becomes less efficient at high densities. The following nonlinear partial differential equation was used to model in-plane transport of indirect excitons subject to the applied conveyer potential Uconv ðxÞ: n @nx ¼ r½Dx rnx þ x nx rðu0 nx þ Uconv Þ þ x : opt @t (1) The first term in square brackets in Eq. (1) accounts for exciton diffusion; Dx is the diffusion coefficient. The second term accounts for exciton drift due to the dipole-dipole exciton interaction, which is approximated by u0 nx [31,32], and the conveyer potential Uconv ¼ edFz ðxÞ ¼ @ ed @z VðrÞ, where voltage VðrÞ originates from the voltage applied to the conveyer electrodes Vz¼0 ðxÞ. The mobility x is given by the generalized Einstein relationship Exciton transport via conveyer, M1 (µm)
PRL 106, 196806 (2011)
30
30
a
15
b
15
0 1
10
100
Excitation Density, Pex(µW)
0 0.1
1 9
10 -2
100 -1
Λ (10 cm ns )
FIG. 3. (a) The measured and (b) calculated average transport distance of indirect excitons via conveyer M1 as a function of density. Aconv ¼ 4:9 meV, vconv ¼ 0:7 m=ns.
196806-2
week ending 13 MAY 2011
PHYSICAL REVIEW LETTERS a
u0nx+Uconv
-30
-45
3
b
PL Intensity
x ¼ Dx ðeT0 =T 1Þ=ðkB T0 Þ, where T0 ¼ ð2@2 nx Þ= ðMx gkB Þ, Mx ’ 0:22m0 is the exciton mass, g ¼ 4 is the spin degeneracy [31]. Because of the geometry of the system, we use the approximation r ¼ @=@x and solve for the density of indirect excitons nx ðx; tÞ. The effect of disorder intrinsic to QWs is included using a thermionic model for the diffusion coefficient, Dx ¼ ð0Þ ð0Þ Dð0Þ x exp½U =ðkB T þ u0 nx Þ [31]. Dx is the diffusion coefficient in the absence of QW disorder and Uð0Þ =2 is the amplitude of the disorder potential. The temperature of indirect excitons T is approximated as T ¼ Tbath . The nonresonant photoexcitation causes heating of the exciton gas by a few kelvin. However, the hot excitons cool to the lattice temperature within a few microns of the excitation spot [10] justifying the approximation. The last two terms in Eq. (1) take account of the creation and decay of excitons. ðxÞ is the generation rate and opt is the optical lifetime. The increased exciton velocity due to transport via the conveyer can, in principle, shift the energy of excitons outside the photon cone and increase their optical lifetime. However, we evaluated opt using the expressions given by Eqs. (1)–(6) in [10], and found that for the studied range of parameters, the corrections to the lifetime are small and can be neglected. Therefore we used a constant opt independent of vconv . The results of the simulations are presented in Fig. 4. The gray (green) curve in Fig. 4(a) presents a snapshot of the conveyer potential for nx ¼ 0. The sinusoidal envelope of the conveyer potential with conv ¼ 14 m is modulated by 2 m-period ripples, which originate from the finite spacing between the conveyer electrodes. The amplitude of these ripples can be controlled by the spacing between the conveyer electrodes ds (the ripples essentially vanish for ds < 0:5 m for the structure). The repulsively interacting indirect excitons screen the external potential [31]. The snapshot of the exciton density distribution is shown by the gray (red) curve in Fig. 4(b) and the corresponding screened conveyer potential by the black (blue) curve in Fig. 4(a). The time-integrated exciton PL with the spatial resolution taken into account is presented by the black curve in Fig. 4(b). The average exciton transport distance via conveyer is evaluated by M1 , excluding the shaded area shown in Figs. 1(g) and 1(h) as for the experimental data. The obtained theoretical simulations of exciton transport via conveyer [Fig. 4(c)] are in qualitative agreement with the experimental data (Fig. 2) exhibiting the dynamical exciton delocalization in shallow conveyers and the DLDT with increasing conveyer amplitude. The simulated and measured conveyer amplitude at the DLDT, Aturn-on , are in qualitative agreement [Fig. 2(b)]. In order to simplify the calculations for the analysis of the exciton density dependence, we approximate the con ¼þ veyer potential by a cosine function Uconv Aconv cos½2ðx=conv fconv tÞ and treat the ripples in the same way as the disorder potential. The position of
Exciton transport via conveyer, M1 (µm) nx (1010 cm-2) Uconv (meV)
PRL 106, 196806 (2011)
2 1 0 0
15 x (µm)
15
c
30 vconv=0.7 µm/ns 3.5 µm/ns
10 6.3 µm/ns 5
0 0
2
4
6
8
Conveyer amplitude, Aconv (meV)
FIG. 4 (color online). Theoretical simulations. A snapshot of (a) the conveyer potential [gray (green) curve], (b) exciton density distribution [gray (red) curve], and (a) corresponding screened conveyer potential [black (blue) curve]. (b) PL intensity (black curve). Aconv ¼ 8 meV, vconv ¼ 0:7 m=ns. (c) M1 vs Aconv . Dxð0Þ ¼ 30 cm2 s1 , ¼ 109 cm2 ns1 , opt ¼ 50 ns.
the ripples is fixed, similar to the position of the CQW disorder potential. We approximate the effect of the CQW disorder and conveyer ripples on exciton transport within the thermionic model [31] via Dx ¼ ð0Þ þ Uð0Þ Þ=ðk T þ u n Þ. Here U ð0Þ Dð0Þ x exp½ðU B 0 x ripple ripple is
ð0Þ the ripple amplitude obtained by simulations. Uripple is nearly proportional to Uconv and, therefore, it is approxið0Þ ¼ CUconv (C is a fitting constant). The mated by Uripple simulated density dependence of exciton transport via conveyer is in qualitative agreement with the experimental data (Fig. 3). The results are discussed below. Conveyer amplitude dependence [Figs. 2(a) and 4(c)].— When the conveyer amplitude is smaller than the exciton interaction energy or disorder amplitude, excitons are not localized in the minima of the moving conveyer potential, and therefore are not moved by the conveyer. When the conveyer amplitude becomes larger than both the exciton interaction energy and disorder amplitude, excitons can localize in the minima of the moving conveyer potential. This results in efficient transport of excitons via conveyer. The saturation of M1 at large Aconv can be related to a device imperfection and can be studied in future works. Conveyer velocity dependence [Figs. 2 and 4(c)].— Excitons can efficiently follow the moving conveyer potential when the maximum exciton drift velocity in the conveyer is higher than the conveyer velocity, vdrift ¼ x ð@Uconv =@xÞmax * vconv . This leads to an estimate Aturn-on vconv conv =x , qualitatively showing that a higher conveyer amplitude is required for efficient exciton transport via conveyer at a higher vconv . The efficiency of exciton transport via conveyers can be improved by in-
196806-3
PRL 106, 196806 (2011)
PHYSICAL REVIEW LETTERS
creasing x . This can be achieved by reducing ds and, in ð0Þ turn, Uripple . A monotonic dependence of Aturn-on on vconv without abrupt changes at the sound velocity is consistent with the pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi thermal velocity of excitons, 2kB T=Mx 15 m=ns at T ¼ 1:7 K, being much higher than the sound velocity. Density dependence (Fig. 3).—At low densities, the excitons are localized in local minima of the disorder potential (given by the intrinsic disorder and ripples in the conveyer potential) and hardly follow the moving conveyer. At the intermediate densities, excitons effectively screen the disorder and can be efficiently moved by the conveyer. Exciton transport via conveyer becomes less efficient at the high densities when excitons screen the conveyer potential. The requirement for efficient exciton transport via conveyers x ð@Uconv =@xÞmax * vconv is relevant, where screening of disorder results in the enhancement of x while screening of the conveyer potential results in the reduction of the conveyer amplitude. In summary, we report on the realization of electrostatic conveyers for excitons and experimental and theoretical studies of exciton transport via conveyers. We thank Misha Fogler, Nikolai Gippius, and Egor Muljarov for discussions. This work was supported by the U.S. DOE Office of Basic Energy Sciences under Award No. DE-FG02-07ER46449. The development of the conveyer rf system and multilayer lithography was also supported by ARO under Grant No. W911NF-08-10341 and NSF under Grant No. 0907349. Cardiff group was supported by EPSRC, CUROP, and WIMCS. This work was performed using the computational facilities of the ARCCA Division, Cardiff University.
[1] M. Hagn, A. Zrenner, G. Bo¨hm, and G. Weimann, Appl. Phys. Lett. 67, 232 (1995). [2] L. V. Butov and A. I. Filin, Phys. Rev. B 58, 1980 (1998). [3] A. V. Larionov, V. B. Timofeev, J. Hvam, and K. Soerensen, Sov. Phys. JETP 90, 1093 (2000). [4] L. V. Butov, A. C. Gossard, and D. S. Chemla, Nature (London) 418, 751 (2002). [5] Z. Vo¨ro¨s, R. Balili, D. W. Snoke, L. Pfeiffer, and K. West, Phys. Rev. Lett. 94, 226401 (2005). [6] A. L. Ivanov, L. E. Smallwood, A. T. Hammack, Sen Yang, L. V. Butov, and A. C. Gossard, Europhys. Lett. 73, 920 (2006). [7] A. Ga¨rtner, A. W. Holleitner, J. P. Kotthaus, and D. Schuh, Appl. Phys. Lett. 89, 052108 (2006). [8] A. Ga¨rtner, L. Prechtel, D. Schuh, A. W. Holleitner, and J. P. Kotthaus, Phys. Rev. B 76, 085304 (2007). [9] X. P. Vo¨gele, D. Schuh, W. Wegscheider, J. P. Kotthaus, and A. W. Holleitner, Phys. Rev. Lett. 103, 126402 (2009). [10] A. T. Hammack, L. V. Butov, J. Wilkes, L. Mouchliadis, E. A. Muljarov, A. L. Ivanov, and A. C. Gossard, Phys. Rev. B 80, 155331 (2009).
week ending 13 MAY 2011
[11] D. A. B. Miller, D. S. Chemla, T. C. Damen, A. C. Gossard, W. Wiegmann, T. H. Wood, and C. A. Burrus, Phys. Rev. B 32, 1043 (1985). [12] S. Zimmermann, A. O. Govorov, W. Hansen, J. P. Kotthaus, M. Bichler, and W. Wegscheider, Phys. Rev. B 56, 13 414 (1997). [13] S. Zimmermann, G. Schedelbeck, A. O. Govorov, A. Wixforth, J. P. Kotthaus, M. Bichler, W. Wegscheider, and G. Abstreiter, Appl. Phys. Lett. 73, 154 (1998). [14] J. Krauß, J. P. Kotthaus, A. Wixforth, M. Hanson, D. C. Driscoll, A. C. Gossard, D. Schuh, and M. Bichler, Appl. Phys. Lett. 85, 5830 (2004). [15] A. T. Hammack, N. A. Gippius, Sen Yang, G. O. Andreev, L. V. Butov, M. Hanson, and A. C. Gossard, J. Appl. Phys. 99, 066104 (2006). [16] M. Remeika, J. C. Graves, A. T. Hammack, A. D. Meyertholen, M. M. Fogler, L. V. Butov, M. Hanson, and A. C. Gossard, Phys. Rev. Lett. 102, 186803 (2009). [17] T. Huber, A. Zrenner, W. Wegscheider, and M. Bichler, Phys. Status Solidi A 166, R5 (1998). [18] G. Chen, R. Rapaport, L. N. Pffeifer, K. West, P. M. Platzman, S. Simon, Z. Vo¨ro¨s, and D. Snoke, Phys. Rev. B 74, 045309 (2006). [19] A. A. High, A. T. Hammack, L. V. Butov, L. Mouchliadis, A. L. Ivanov, M. Hanson, and A. C. Gossard, Nano Lett. 9, 2094 (2009). [20] A. A. High, A. K. Thomas, G. Grosso, M. Remeika, A. T. Hammack, A. D. Meyertholen, M. M. Fogler, L. V. Butov, M. Hanson, and A. C. Gossard, Phys. Rev. Lett. 103, 087403 (2009). [21] A. A. High, A. T. Hammack, L. V. Butov, M. Hanson, and A. C. Gossard, Opt. Lett. 32, 2466 (2007). [22] A. A. High, E. E. Novitskaya, L. V. Butov, M. Hanson, and A. C. Gossard, Science 321, 229 (2008). [23] G. Grosso, J. Graves, A. T. Hammack, A. A. High, L. V. Butov, M. Hanson, and A. C. Gossard, Nat. Photon. 3, 577 (2009). [24] Y. Y. Kuznetsova, M. Remeika, A. A. High, A. T. Hammack, L. V. Butov, M. Hanson, and A. C. Gossard, Opt. Lett. 35, 1587 (2010). [25] G. E. Smith, Rev. Mod. Phys. 82, 2307 (2010). [26] C. Rocke, S. Zimmermann, A. Wixforth, J. P. Kotthaus, G. Bo¨hm, and G. Weimann, Phys. Rev. Lett. 78, 4099 (1997). [27] J. Rudolph, R. Hey, and P. V. Santos, Phys. Rev. Lett. 99, 047602 (2007). [28] S. Lazic´, P. V. Santos, and R. Hey, Physica (Amsterdam) 42E, 2640 (2010). [29] E. A. Cerda-Me´ndez, D. N. Krizhanovskii, M. Wouters, R. Bradley, K. Biermann, K. Guda, R. Hey, P. V. Santos, D. Sarkar, and M. S. Skolnick, Phys. Rev. Lett. 105, 116402 (2010). [30] See supplemental material at http://link.aps.org/ supplemental/10.1103/PhysRevLett.106.196806 for circuit schematic for the conveyer rf system, calibration of the conveyer amplitude, and simulation details. [31] A. L. Ivanov, Europhys. Lett. 59, 586 (2002). [32] A. L. Ivanov, E. A. Muljarov, L. Mouchliadis, and R. Zimmermann, Phys. Rev. Lett. 104, 179701 (2010).
196806-4