Letter

pubs.acs.org/NanoLett

Electrical Control of g‑Factor in a Few-Hole Silicon Nanowire MOSFET

B. Voisin,†,‡ R. Maurand,*,†,‡ S. Barraud,§ M. Vinet,§ X. Jehl,†,‡ M. Sanquer,†,‡ J. Renard,†,‡
and S. De Franceschi†,‡
†
Univ. Grenoble Alpes, INAC-SPSMS, F-38000 Grenoble, France
‡
CEA, INAC-SPSMS, F-38000 Grenoble, France
§
CEA, LETI, MINATEC Campus, 17 rue des Martyrs, 38054 Grenoble, France
*
S Supporting Information

ABSTRACT: Hole spins in silicon represent a promising yet

barely explored direction for solid-state quantum computation,
possibly combining long spin coherence, resulting from a
reduced hyperfine interaction, and fast electrically driven qubit
manipulation. Here we show that a silicon-nanowire field-effect
transistor based on state-of-the-art silicon-on-insulator technol-
ogy can be operated as a few-hole quantum dot. A detailed
magnetotransport study of the first accessible hole reveals a g-
factor with unexpectedly strong anisotropy and gate depend-
ence. We infer that these two characteristics could enable an
electrically driven g-tensor-modulation spin resonance with Rabi frequencies exceeding several hundred mega-Hertz.
KEYWORDS: Quantum dot, spin qubits, hole transport, Landé g-factor, silicon, MOSFET

L ocalized spins in semiconductors provide a natural way to

encode quantum information in a solid-state environ-
ment.1,2 This opportunity has been extensively investigated
spin coherence time an order of magnitude longer than in the
electron case.12 The second advantage has to do with the
existence in the valence band of a sizable spin−orbit (SO)
using the spin of electrons confined to quantum dots in GaAs/ interaction. By coupling spin and orbital degrees of freedom,
AlGaAs high-mobility heterostructures.3−5 In this system, and the SO interaction offers the possibility of fast all-electrical spin
in III−V materials in general, the electron spin coherence is manipulation as opposed to magnetic-field driven electron spin
rather short though. The presence of a large hyperfine coupling resonance. The SO mediated approach only requires a
with the nuclear spins of the host crystal elements leads to time microwave modulation on a gate electrode, and it can in
averaged coherence times, T2*, of only a few tens of principle yield very large Rabi frequencies. On the contrary, the
nanoseconds. This limitation can be overcome by considering magnetic-field driven ESR performed using oscillatory currents
spin-qubit implementations in group-IV semiconductors (Si, requires an additional manipulation circuit. This approach is
Ge, C), where the concentration of isotopes carrying nonzero also limited by the need to minimize local Joule heating, which
nuclear spins is naturally much smaller. As a matter of fact, the translates into rather limited Rabi frequencies, typically below
concentration of group-IV elements with nonzero nuclear spin 10 MHz. A third approach enabling all-electrical yet magneti-
can be even lowered by means of isotopic purification cally driven ESR was successfully demonstrated.13,14 It uses a
techniques.6 An outstanding T*2 of 120 μs, i.e., 4 orders of micromagnet to locally create an inhomogeneous magnetic field
magnitude longer than in III−V materials, was obtained in a in which the qubit electron is “shaken” by oscillating a gate
spin-qubit device based on isotopically purified silicon, with voltage. This approach has so far enabled Rabi frequencies as
over 99.9% of nuclear-spin-free 28Si, the most abundant silicon high as 100 MHz, but a recent theoretical work showed that the
isotope.7 magnetic field gradient induces extra dephasing.15 Finally a last
Most of the experimental work on silicon-based spin qubits electrical manipulation has been recently demonstrated on a P
has so far focused on electron spins, either bound to a donor donor in silicon using Stark shift for coherent rotation.16
atom8 or confined to a gate-defined quantum dot.7,9,10 The precise SO-related mechanism underlying spin manip-
However, hole spin qubits have been lagging behind, although ulation depends on the nature of the SO interaction. In the case
they may potentially present some advantages. The first of Rashba and Dresselhaus SO interactions, spin rotation is
advantage comes from the p-wave symmetry of the hole achieved via the so-called mechanism of electric dipole spin
Bloch functions. Having a node at the atomic site suppresses resonance (EDSR),17,18 experimentally demonstrated for
the direct-contact coupling to the nuclear magnetic moment,
which results in a weaker hyperfine interaction.11 Optical Received: July 24, 2015
experiments performed on hole spins in self-assembled InAs Revised: November 2, 2015
quantum dots have confirmed this expectation revealing a hole

© XXXX American Chemical Society A DOI: 10.1021/acs.nanolett.5b02920

Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters Letter

Figure 1. (a) Simplified schematics of a CMOS nanowire transistor. A quantum dot is formed below the gate inside a thin silicon on insolator (SOI)
channel. The dot is isolated from the silicided source and drain by the presence of spacers. (b) False colored transmission electron microscope
(TEM) micrograph of a nanowire transistor section, the silicon channel (in red) at the top of a SOI wafer with a 145 nm buried oxide (BOX) layer
(blue) is surrounded by a gate stack of SiO2 and high-k dielectric (in yellow). The gate (in gray) is realized in poly-silicon. (c) Measured differential
conductance taken at zero bias as a function of gate voltage (Vg) at a temperature of T = 4 K showing Coulomb oscillations (×5 and ×10 indicate
that the corresponding current peaks have been multiplied by, respectively, 5 and 10 times for visibility). Inset shows the transistor characteristic,
current versus Vg at room temperature at Vb = 10 mV. The subthreshold slope close to 60 mV per decade indicates an excellent electrostatic control.
(d) Stability diagram dI/dV vs (Vb, Vg) performed at 260 mK showing the addition of the first holes in the quantum dot. The first resonance appears
at the threshold voltage around Vg = −580 mV, and no feature is observed below.

electron qubits in GaAs,17 InAs,19 InSb,20 and carbon The measured transistors were fabricated on a comple-
nanotube21 quantum dots. The SO interaction can as well mentary metal-oxide semiconductor (CMOS) platform starting
result in gate-dependent and anisotropic g-factors. In this case, from 300 mm-diameter undoped silicon-on-insulator (SOI)
another possibility arises: spin manipulation via electrically wafers ((100) surface). Fabrication involves several steps of
driven g-tensor modulation resonance (g-TMR), as demon- optical lithography, metal/dielectric deposition, and etching.
strated for electrons spins in InGaAs-based heterostructures.22 The use of an industrial-level CMOS platform presents crucial
This latter approach could as well be applied to hole-type advantages: an excellent electrostatic control of the undoped
quantum dots. A few examples of gate defined hole quantum channel is achieved despite ultimate characteristic dimensions,
dots were reported in the recent years,23−28 and to the best of with a channel length of about 25 nm oriented along the [110]
our knowledge, the only experimental demonstration of a SO- direction, width, and thickness of about 10 nm (see Figure
mediated hole-spin resonance was reported for InSb nanowire 1a,b); the devices also benefit from low-resistance source and
quantum dots.28 The implementation of hole spin qubits in drain leads fabricated through a boron implantation process
silicon-based quantum dots would be particularly interesting followed by silicidation. These two features enable the
owing to the lower hyperfine coupling and the compatibility achievement of the few-hole regime reported here. The high
control and reproducibility of CMOS technology is finally
with mainstream silicon technology.
essential for qubit integration toward large-scale quantum
In this letter we show that state-of-the-art metal-oxide field-
computing architectures.
effect transistors (MOSFETs) can be operated as p-type
Devices were studied in a 3He measurement setup with a
quantum dots in the few-hole regime. So far, the few-hole base temperature of 260 mK. A 9 T superconducting magnet
regime in quantum dots had only been reached in bottom-up and a mechanical sample rotator were used to investigate the
nanowires, based on either InSb28 or silicon,29,30 while the hole g-factors and their angular dependences with an estimated
operation of top-down nanostructures was limited to large error of ±5°. Electrical measurements were performed using
numbers of holes.23,24,31,32 Using magneto-transport measure- both direct current (DC) and low-frequency lock-in techniques.
ments at low temperature, we study the Zeeman splitting of the In our setup, noise attenuation along the DC lines is carried out
first hole level accessible, evidencing a spin doublet with at room temperature, using pi-filters, and at low temperature,
predominant heavy-hole character. The anisotropy and gate- using low-pass RC filters (∼10 kHz cutoff frequency) thermally
dependence of the corresponding hole g-factor reveal that the anchored to the 3He pot. This enables electronic temperatures
studied system meets all the requirements to be used as a spin as low as the cryostat base temperature. Current is amplified at
qubit electrically manipulated by g-TMR. Simulations predict room temperature by means of a battery powered current−
Rabi frequencies exceeding several hundred mega-Hertz, more voltage converter before DC or lock-in detection.
than 2 orders of magnitude higher than those of recent works At 4 K the linear conductance of the device presents peaks
on electron spin qubits.7,9,33 above the threshold voltage as shown in Figure 1c. This is the
B DOI: 10.1021/acs.nanolett.5b02920
Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters Letter

signature that charge carriers are accumulated in a single island under a 6 T in-plane magnetic field, B∥, parallel to the channel
formed under the gate. The localized holes are weakly coupled (see Figure 1a). The corresponding g-factor is labeled as g0h/1h
∥ .
to the source and drain leads. Confinement along the wire The lever-arm parameter α, relating gate-voltage and electro-
direction owes to the presence of spacers separating the gated chemical-potential (μ) variations (i.e., α = dμ/dVg), can be
central part of the undoped channel from the heavily doped obtained from the diamond slopes. We estimate α = 0.85 eV/V:
source and drain leads (see Figure 1a,b). The stability diagram such a large value, close to unity, demonstrates the excellent
of the device, i.e., a two-dimensional map of the differential gate control over the channel electrostatic potential. Among
conductance dI/dV as a function of the gate voltage Vg and of some ridges invariant under magnetic field, which we ascribe to
the source-drain bias voltage Vb, is shown in Figure 1d. It shows resonances in the density of states of the leads,35 a strong ridge
clear Coulomb-blockade regions, signature of single-hole can be clearly observed in Figure 2a, highlighted by a black
transport through the quantum dot. Note here that two corner arrow. This ridge denotes the onset of tunneling via the one-
dots34 are not expected due to the narrow width (W = 10 nm) hole, excited spin state whose separation from the ground state
of the device. The first diamond is positioned just above the increases linearly with magnetic field as shown in Figure 2b.
room temperature threshold voltage (see inset of Figure 1c), at This data indicates that the first energy level is 2-fold
which charge carriers start to populate the valence band in the degenerate constituting a spin doublet. The degeneracy lifting
channel. No extra feature appears below this first resonance up between heavy and light holes observed here is a consequence
to 50 mV in bias (see Supporting Information S1), which of the strong confinement at the silicon interface.
means that the last hole in the channel has been most probably From the data of Figure 2b we extract the in-plane g-factor
observed.30 However, an irrefutable proof would be given by a ΔE
g 0h/1h = μ B for magnetic fields parallel to the channel (here
charge detection method. Details about electronic transport in B

the regime of a higher number of confined holes are presented ΔE is the Zeeman energy splitting between the two spin states
in the Supporting Information S2. and μB the Bohr magneton). We find g0h/1h ∥ ≈ 2.3. The in-plane
Figure 2a shows a high-resolution tunnel spectroscopy at the azimuthal angle, ϕ, between the magnetic field vector and the
charge transition between the zero- and one-hole state (0h and channel axis is then varied, and the corresponding change in g∥
1h, respectively). The measurement was performed at 260 mK recorded. The full angle dependence, shown in Figure 2c,
reveals g0h/1h
∥ values varying between 2.3 and 2.6. The g-factor
exhibits a strong variation also as a function of the polar angle θ
(see Figure 1a). The θ dependence measured in the plane
orthogonal to the channel axis (i.e., for ϕ = 90°) is illustrated in
Figure 2d. When the magnetic field is aligned along the vertical
axis (θ = 0°), perpendicular to the substrate plane, the g-factor
reaches its smallest value g0h/1h
⊥ = 1.5. Moreover, Figure 2c,d
seem to suggest a nonsinusoidal anisotropy. Due to an
experimental uncertainty of (±5°) on the angle between the
magnetic field and the channel axis, however, we cannot
conclude on the actual functional form of the angle
dependence.
We note that our results are compatible with earlier
measurements of the hole g-factor in a silicon−nanowire
quantum dot.30 In that work a g-factor of 2.3 was reported for
the first hole and a magnetic field perpendicular to the substrate
(and hence to the nanowire), but angular dependences were
not addressed. The g-factors obtained here are strongly
anisotropic and mostly below the expectation for heavy holes
in a two-dimensional silicon layer, namely, |g| = 6|κ| where the
Luttinger parameter |κ| ≈ 0.42 in silicon.36 This result is likely
due to the strong confinement and the associated mixing
Figure 2. (a) Two-dimensional color plot of the differential between heavy and light hole states. Since confinement is gate-
conductance dI/dV as a function of bias and gate voltages at a
voltage dependent we may expect g-factors to vary with Vg,
magnetic field of B = 6 T at the zero- to one-hole transition. The black
arrow indicates the conductance resonance through the Zeeman split which may open the possibility for spin manipulation using g-
excited state. The inset sketches a stability diagram around the one- TMR. This possibility, originally demonstrated for electrons in
hole charge state at finite magnetic field. Excited states resonances are InGaAs heterostructures,37 was recently investigated for the
plotted in red, while the dashed lines indicate schematically where the case of holes in SiGe self-assembled quantum dots.38 It was
magnetospectroscopies have been recorded. (b) Two-dimensional shown that for a fixed, odd number of holes in the quantum
color plot of the differential conductance dI/dV as a function of gate dot, a relatively small gate-voltage modulation could induce
voltage and magnetic field recorded at a fixed bias voltage of Vb = 3 Rabi oscillations at a frequency of 100 MHz.
mV. The gate voltage scale is Vg−Voffset with Voffset = −581 mV. The Following a similar approach, we focus on the first-hole
linear-in-field splitting of the conductance peaks is a consequence of
regime and investigate the gate-voltage dependence of the
the Zeeman effect. (c) Magnetic field orientation dependence of the g
factor for the zero- to one-hole (in red indicated as g0h/1h and one-hole parallel and perpendicular components of the hole g-factor. In
∥
to two-hole (in blue indicated as g1h/2h) transitions. An angle of ϕ = 0° order to evaluate the maximal variation of the g-factor
∥
corresponds to a magnetic field applied along the channel of the components within the one-hole regime we compare the
transistor. (d) Same as that in panel c but for the out-of-plane already discussed tunnel spectroscopy measurement of the
magnetic field measured at the 0h/1h transition for ϕ = 90°. Zeeman splitting at the 0h/1h transition with analogous
C DOI: 10.1021/acs.nanolett.5b02920
Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters Letter

Figure 3. Illustration of the g-tensor modulation technique. A different gate dependence for g⊥ and g∥ combined with an anisotropy (a) leads to an
equivalent oscillatory magnetic field (b) when a time-dependent gate voltage is applied together with a static magnetic field B0. This can drive
transition between spin states at a speed given by the Rabi frequency. This frequency depends on the system parameters and can exceed 600 MHz in
∂g
our system. The calculation (c) is performed using ∂Vg
≈ 0.025/mV , g∥ = 2.6, g⊥ = 1.5, B0 = 1 T, and Vac = 10 mV.

measurements at the 1h/2h transition, i.e., around Vg = −612 The optimal magnetic field angle θmax with respect to the
mV. To study the Zeeman splitting at the transition 0h/1h we ⎛ g ⎞
have recorded a conductance trace at finite positive Vb device plane is then given as θmax = arctan⎜ || ⎟ ≈ 50°. The
⎝ g⊥ ⎠
depending on Vg marked by a black dashed line in the inset
corresponding Rabi frequency exceeds 600 MHz, i.e.,
of Figure 2a. The same way we have recorded a conductance
significantly larger than Rabi frequencies reported for electric-
trace at negative Vb (black dashed line in the inset of Figure 2a)
field driven spin resonance in III−V semiconductor nanostruc-
to study the 1h/2h transition. The sign of Vb has been chosen
tures (up to 100 MHz in InSb nanowires39). In addition to a
to have the best contrast between the Zeeman excited state
large Rabi frequency, a conveniently long spin coherence time
conductance peak and the conductance peaks associated with
may be expected in the hole spin qubits studied here. Assuming
local density of states in the leads. We emphasize that both sets spin coherence to be limited by hyperfine coupling, we expect
of measurements, while corresponding to different charge T2* ≈ 1.5 μs for a hole spin in a [13 nm]3 silicon quantum dot,
transitions (0h/1h and 1h/2h), provide information on the which is 2 orders of magnitude larger than in III−V
Zeeman splitting of the one-hole ground state. The only semiconductors (see Supporting Information S4).
difference lies in the electrostatic confining potential, due to In conclusion, the few-hole regime has been achieved in a p-
different gate voltages. The in-plane angular dependence of type silicon quantum dot confined within the channel of a field-
g1h/2h
∥ is shown in Figure 2c (blue data points). As compared to effect transistor based on industrial level SOI technology. A
g0h/1h
∥ (red data points), g1h/2h
∥ is generally smaller and presents a detailed study of the first accessible hole reveals an anisotropic
roughly opposite anisotropy. While the precise origin of this and gate-dependent hole g-factor. The observed gate depend-
peculiarity remains unclear, this comparison reveals a strong ence is maximal for the in-plane g-factor component
gate dependence of the g-factor in-plane anisotropy. Going perpendicular to the channel of the transistor. Moreover we
from one side to the other of the one-hole Coulomb-blockade have discussed the revealed anisotropy and gate-dependence in
regime, the in-plane g-factor component along the channel is terms of possible g-TMR. We estimate that Rabi frequencies as
basically constant (around 2.3), while the in-plane component high as several hundred MHz could be reached with realistically
perpendicular to the channel varies from about 1.85 to about strong gate modulation. These frequencies are at least two
2.6. This strong gate voltage dependence of the g-factor could orders larger than in the case of electron spin qubits using
be due to the large variation of the confinement potential magnetically driven spin resonance.7,8,40 In view of the
enabled by the thin gate oxide of ∼1 nm. For magnetic fields demonstrated potential for fast electric-field driven spin
applied perpendicular to the plane we find an approximately manipulation, the reduced hyperfine interaction expected
constant g-factor, i.e., g0h/1h
⊥ ≈ g1h/2h
⊥ ≈ 1.5 (shown in the from the p-symmetry valence band states, and the full
Supporting Information S3). compatibility with mainstream microelectronics technology,
The anisotropy and gate dependence of the in-plane g-factors the silicon-based p-type quantum dots studied here appear to
alone are sufficient to allow for g-TMR. Following ref 38, for be an attractive option for spin qubit applications.

■
reasonably strong gate voltage modulation Vac = 10 mV and a
static magnetic-field amplitude of 1 T, we calculate the ASSOCIATED CONTENT
corresponding Rabi frequency as a function of magnetic field
orientation. The results are presented in Figure 3. First we *
S Supporting Information

select ϕ ≈ 90° in order to maximize the gate dependence of g∥ The Supporting Information is available free of charge on the
(from Figure 2c, g∥ goes from 2.6 at the 0h/1h transition to ACS Publications website at DOI: 10.1021/acs.nano-
1.85 at the 1h/2h transition). Assuming a linear evolution of g∥ lett.5b02920.
∂g∥
with Vg, we deduce ≈ 0.025/mV while g⊥ is constant at a Additional experimental data about the quantum dot
∂Vg
hole shell-filling and an estimation of the hyperfine
value of 1.5 (see Figure 3a). Finally we compute the Rabi coupling for holes in silicon are provided (PDF)
frequency at ϕ ≈ 90° with the following expression:

fRabi =
μB Vac ⎡ 1 ∂g ⎤
⎢ ⎥×
2h ⎢⎣ g ∂Vg ⎥⎦
g g⊥B B⊥ ■ AUTHOR INFORMATION
Corresponding Author
(g B )2 + (g⊥B⊥)2 (1) *E-mail: [email protected].
D DOI: 10.1021/acs.nanolett.5b02920
Nano Lett. XXXX, XXX, XXX−XXX
Nano Letters Letter

Notes (23) Dö t sch, U.; Gennser, U.; David, C.; Dehlinger, G.;
The authors declare no competing financial interest. Grützmacher, D.; Heinzel, T.; Lüscher, S.; Ensslin, K. Appl. Phys.

■
Lett. 2001, 78, 341.
(24) Grbić, B.; Leturcq, R.; Ensslin, K.; Reuter, D.; Wieck, A. D. Appl.
ACKNOWLEDGMENTS Phys. Lett. 2005, 87, 232108.
The authors thank Y.-M. Niquet and J. Li for fruitful (25) Hu, Y.; Churchill, H. O.; Reilly, D. J.; Xiang, J.; Lieber, C. M.;
Marcus, C. M. Nat. Nanotechnol. 2007, 2, 622−625.
discussions. The research leading to these results has been
(26) Klochan, O.; Micolich, a. P.; Hamilton, a. R.; Trunov, K.;
supported by the European Communitys Seventh Framework Reuter, D.; Wieck, a. D. Phys. Rev. Lett. 2011, 107, 076805.
under the Grants Agreement No. 323841 (http://www.sispin. (27) Hu, Y.; Kuemmeth, F.; Lieber, C. M.; Marcus, C. M. Nat.
eu/) and No. 610637 (http://www.the-siam-project.eu/) and Nanotechnol. 2012, 7, 47−50.
by the European Research Council under the ERC Grant (28) Pribiag, V.; Nadj-Perge, S.; Frolov, S.; van den Berg, J.; van
agreement No. 280043 (HybridNano). Weperen, I.; Plissard, S.; Bakkers, E.; Kouwenhoven, L. Nat.

■
Nanotechnol. 2013, 8, 170−174.
(29) Zhong, Z.; Fang, Y.; Lu, W.; Lieber, C. M. Nano Lett. 2005, 5,
REFERENCES 1143−1146.
(1) Loss, D.; Divincenzo, D. P. Phys. Rev. A: At., Mol., Opt. Phys. (30) Zwanenburg, F. A.; van Rijmenam, C. E. W. M.; Fang, Y.;
1998, 57, 120−126. Lieber, C. M.; Kouwenhoven, L. P. Nano Lett. 2009, 9, 1071−1079.
(2) Kane, B. E. Nature 1998, 393, 133−137. PMID: 19226128.
(3) Koppens, F. H. L.; Folk, J. A.; Elzerman, J. M.; Hanson, R. Science (31) Spruijtenburg, P. C.; Ridderbos, J.; Mueller, F.; Leenstra, A. W.;
2005, 309, 1346−1350. Brauns, M.; Aarnink, A. A. I.; van der Wiel, W. G.; Zwanenburg, F. A.
(4) Johnson, a. C.; Petta, J. R.; Taylor, J. M.; Yacoby, a.; Lukin, M. D.; Appl. Phys. Lett. 2013, 102, 192105.
Marcus, C. M.; Hanson, M. P.; Gossard, a. C. Nature 2005, 435, 925− (32) Li, R.; Hudson, F. E.; Dzurak, A. S.; Hamilton, A. R. Appl. Phys.
8. Lett. 2013, 103, 163508.
(5) Petta, J. R.; Johnson, A. C.; Taylor, J. M.; Laird, E. A.; Yacoby, A.; (33) Muhonen, J. T.; Dehollain, J. P.; Laucht, A.; Hudson, F. E.;
Lukin, M. D.; Marcus, C. M.; Hanson, M. P.; Gossard, A. C. Science Kalra, R.; Sekiguchi, T.; Itoh, K. M.; Jamieson, D. N.; McCallum, J. C.;
2005, 309, 2180. Dzurak, A. S.; Morello, A. Nat. Nanotechnol. 2014, 9, 986.
(6) Itoh, K. M.; Kato, J.; Uemura, M.; Kaliteevskii, A. K.; Godisov, O. (34) Voisin, B.; Nguyen, V.-H.; Renard, J.; Jehl, X.; Barraud, S.;
N.; Devyatych, G. G.; Bulanov, A. D.; Gusev, A. V.; Kovalev, I. D.; Triozon, F.; Vinet, M.; Duchemin, I.; Niquet, Y.-m.; Franceschi, S. D.;
Sennikov, P. G.; Pohl, H.-J.; Abrosimov, N. V.; Riemann, H. Jpn. J. Sanquer, M. Nano Lett. 2014, 14, 2094−2098.
Appl. Phys. 2003, 42, 6248−6251. (35) Pierre, M.; Wacquez, R.; Jehl, X.; Sanquer, M.; Vinet, M.; Cueto,
(7) Veldhorst, M.; Hwang, J.; Yang, C.; Leenstra, A.; de Ronde, B.; O. Nat. Nanotechnol. 2010, 5, 133−137.
Dehollain, J.; Muhonen, J.; Hudson, F.; Itoh, K.; Morello, A. Nat. (36) Winkler, R. Spin−Orbit Coupling Effects in Two-Dimensional
Nanotechnol. 2014, 9, 981−985. Electron and Hole Systems; Springer Science & Business Media: New
(8) Pla, J. J.; Tan, K. Y.; Dehollain, J. P.; Lim, W. H.; Morton, J. J. L.; York, 2003.
Jamieson, D. N.; Dzurak, A. S.; Morello, A. Nature 2012, 489, 541− (37) Kato, Y.; Myers, R. C.; Driscoll, D. C.; Gossard, A. C.; Levy, J.;
545. Awschalom, D. D. Science 2003, 299, 1201−1204.
(9) Kawakami, E.; Scarlino, P.; Ward, D. R.; Braakman, F. R.; Savage, (38) Ares, N.; Katsaros, G.; Golovach, V. N.; Zhang, J. J.; Prager, A.;
D. E.; Lagally, M. G.; Friesen, M.; Coppersmith, S. N.; Eriksson, M. A.; Glazman, L. I.; Schmidt, O. G.; De Franceschi, S. Appl. Phys. Lett.
Vandersypen, L. M. K. Nat. Nanotechnol. 2014, 9, 666−670. 2013, 103, 263113.
(10) Hao, X.; Ruskov, R.; Xiao, M.; Tahan, C.; Jiang, H. Nat. (39) van den Berg, J. W. G.; Nadj-Perge, S.; Pribiag, V. S.; Plissard, S.
Commun. 2014, 5, 3860. R.; Bakkers, E. P. A. M.; Frolov, S. M.; Kouwenhoven, L. P. Phys. Rev.
(11) Fischer, J.; Coish, W.; Bulaev, D.; Loss, D. Phys. Rev. B: Condens. Lett. 2013, 110, 066806.
Matter Mater. Phys. 2008, 78, 155329. (40) Koppens, F.; Buizert, C.; Tielrooij, K.-J.; Vink, I.; Nowack, K.;
(12) Brunner, D.; Gerardot, B. D.; Dalgarno, P. a.; Wüst, G.; Karrai, Meunier, T.; Kouwenhoven, L.; Vandersypen, L. Nature 2006, 442,
K.; Stoltz, N. G.; Petroff, P. M.; Warburton, R. J. Science 2009, 325, 766−771.
70−2.
(13) Tokura, Y.; van der Wiel, W. G.; Obata, T.; Tarucha, S. Phys.
Rev. Lett. 2006, 96, 047202.
(14) Pioro-Ladriére, M.; Obata, T.; Tokura, Y.; Shin, Y.-S.; Kubo, T.;
Yoshida, K.; Taniyama, T.; Tarucha, S. Nat. Phys. 2008, 4, 776−779.
(15) Beaudoin, F.; Coish, W. a. Phys. Rev. B: Condens. Matter Mater.
Phys. 2013, 88, 085320.
(16) Laucht, A.; Muhonen, J. T.; Mohiyaddin, F. A.; Kalra, R.;
Dehollain, J. P.; Freer, S.; Hudson, F. E.; Veldhorst, M.; Rahman, R.;
Klimeck, G.; Itoh, K. M.; Jamieson, D. N.; Mccallum, J. C.; Dzurak, A.
S.; Morello, A. Science Advances 2015, 1, 1−6.
(17) Nowack, K. C.; Koppens, F. H. L.; Nazarov, Y. V.; Vandersypen,
L. M. K. Science 2007, 318, 1430−1433.
(18) Bulaev, D.; Loss, D. Phys. Rev. Lett. 2007, 98, 097202.
(19) Nadj-Perge, S.; Frolov, S. M.; Bakkers, E. P. a. M.;
Kouwenhoven, L. P. Nature 2010, 468, 1084−7.
(20) van den Berg, J. W. G.; Nadj-Perge, S.; Pribiag, V. S.; Plissard, S.
R.; Bakkers, E. P. a. M.; Frolov, S. M.; Kouwenhoven, L. P. Phys. Rev.
Lett. 2013, 110, 066806.
(21) Laird, E. A.; Pei, F.; Kouwenhoven, L. P. Nat. Nanotechnol.
2013, 8, 565−568.
(22) Kato, Y.; Myers, R. C.; Driscoll, D. C.; Gossard, A. C.; Levy, J.;
Awschalom, D. D. Science 2003, 299, 1201−1204.

E DOI: 10.1021/acs.nanolett.5b02920
Nano Lett. XXXX, XXX, XXX−XXX