Abstract
Donor spins in silicon provide a promising material platform for large scale quantum computing. Excellent electron spin coherence times of \({T}_{2}^{* }=268\) μs with fidelities of 99.9% have been demonstrated for isolated phosphorus donors in isotopically pure ^{28}Si, where donors are localareaimplanted in a nanoscale MOS device. Despite robust single qubit gates, realising twoqubit exchange gates using this technique is challenging due to the statistical nature of the dopant implant and placement process. In parallel a precision scanning probe lithography route has been developed to place single donors and donor molecules on one atomic plane of silicon with high accuracy aligned to heavily phosphorus doped silicon inplane gates. Recent results using this technique have demonstrated a fast (0.8 ns) twoqubit gate with two P donor molecules placed 13 nm apart in ^{nat}Si. In this paper we demonstrate a single qubit gate with coherent oscillations of the electron spin on a P donor molecule in ^{nat}Si patterned by scanning tunneling microscope (STM) lithography. The electron spin exhibits excellent coherence properties, with a \({T}_{2}\) decoherence time of 298 ± 30 μs, and \({T}_{2}^{* }\) dephasing time of 295 ± 23 ns.
Similar content being viewed by others
Introduction
Atom qubits in silicon^{1} rely on using the potential well naturally formed by the donor atom nucleus to bind the electron spin at cryogenic temperatures allowing for reproducible qubits with a low gate density^{2}. To date, electron spin qubits in ^{nat}Si have demonstrated \({T}_{2}^{* }=55\) ns^{3} increasing to \({T}_{2}^{* }=268\) μs in isotopically pure ^{28}Si^{4} by addressing one of the phosphorus atoms statistically present in an ion implant region^{5,6}. Atomic precision STM lithography allows us to pattern both individual single donor atoms, but also multidonor molecules in silicon devices. Whilst single donors can trap up to two electrons, donor molecules can host multiple electrons, depending on number of donors present^{7}. This ability to change the number of donors creates an enhanced tunability of the exchange interaction required for twoqubit gates^{8} and is known to reduce the impact of randomly placed charge impurities on both single and exchange coupled qubits^{9}. Donor molecules have also been shown to produce longer spin relaxation times^{10} with an inbuilt addressability of spin resonance transitions allowing for lower error rates in multiqubit control^{11}.
In this paper we present coherent control of an electron spin on a multielectron 1P–2P phosphorus donor molecule patterned by STM lithography in ^{nat}Si with a Rabi frequency of 1.2 MHz and a \({T}_{2}^{* }\) dephasing time of 295 ns. Utilising the fifth electron spin on this molecule we observe a single narrow (7 MHz full width half maximum (FWHM)) electron spin resonance (ESR) transition, which we attribute to the reduced hyperfine interaction in our multielectron system and potential nuclear spin pumping effects. Our results confirm that electron spin coherence times in ^{nat}Si can be maintained in atomically engineered devices in silicon^{12}, patterned close to heavily doped phosphorus leads and with multiple phosphorus nuclear spins in the qubit. These results are promising for precision engineered phosphorus in silicon architectures, where the transition to isotopically pure ^{28}Si^{13} promises significant increases in coherence times^{4}.
Results and Discussion
The donor qubit is fabricated using STM hydrogen lithography, where we employ an STM tip to selectively desorb hydrogen atoms from the silicon surface with subsequent dosing with phosphine gas and a high temperature incorporation anneal^{14}. The central device structure after lithography is shown in Fig. 1a. Control structures with a metallic doping density are shown in the bright regions. Bias voltages V_{L}, V_{C}, V_{R}, V_{B} applied to inplane gates provide electrostatic control over the charge state, and a singleelectron transistor (SET), which is biased at V_{SET} ≈ 300 μV, permits singleshot charge detection^{15} by monitoring the drain current I_{SET}. The SET also acts as an electron reservoir for the donor sites, patterned at a distance of 18.5 nm from the SET.
The dotted rectangular central device region is enlarged in Fig. 1b, c, which show the donor incorporation sites before and after phosphine dosing, respectively. By examining the height profiles we can identify all features in the STM micrographs (see Supplementary Note 1 for a full analysis). To the left of the SET, we observe one PH fragment in site L after dosing. To the right of the SET we observe two incorporation sites. The first site, R_{α} incorporated a single P atom, indicated by the P–Si heterodimer labelled surrounded by seven dangling bond sites. From previous experiments we know that a single dangling bond cannot absorb and incorporate P into the surface^{16}. The second site, R_{β} absorbed one PH_{2} and two PH species, giving rise to two incorporated P atoms. Collectively, these two sites R_{α} and R_{β}, separated by 8 nm, form a 1P–2P molecular state which can bind up to five electrons. Extrapolating from a previously obtained tunnel coupling of 4.3 GHz for a 2P–3P(3e) system, we expect the tunnel coupling for the fifth electron case in our molecular system to be in the hundreds of GHz taking into account the reduced separation between the quantum dots (16 nm → 8 nm), the additional electrons loaded onto the system (3e → 5e), and the lower number of P atoms comprising the quantum dots (2P–3P → 1P–2P)^{7}.
Following lithography, dosing and annealing, 50 ± 5 nm of epitaxial silicon was grown at 250 ^{∘}C and electron beam lithography used to pattern surface ohmic contacts and a microwave antenna for ESR control^{11}. Using etched registration markers, we align the antenna at a lateral distance of 300 ± 50 nm from the buried donors to produce an oscillating magnetic field B_{1} perpendicular to the substrate at the donors’ position. Measurements were performed in a dilution refrigerator at an electron temperature of ~250 mK and under the application of a static magnetic field of B_{0} = 1.45 T along the [110] crystallographic axis, within the plane as indicated in Fig. 1a.
With a positive potential on the two centre gates (V_{C} = 750 mV, V_{B} = 300 mV), we vary gate voltages V_{L} and V_{R} to obtain the charge stability diagram of the SET and donor system, shown in Fig. 1d. From the charge stability diagram we observe breaks in the SET current due to charge transitions from the three donor sites (L, R_{α} and R_{β}) to the SET with three distinct slopes (blue, green and pink). The difference in slope of these charge transition lines in the stability diagram equates to a unique relative lever arm, α of the left and right gate to each dot, confirming the presence of three different charge sites where phosphorus has been incorporated. We identify several accessible charge transitions within the achievable gate range, giving up to a total of six electrons in the system in the upper right corner. At more negative gate voltages (V_{L} < 0 and V_{R} < 0) the system is fully ionised, see Supplementary Note 2. Increasing the gate voltages lowers the electrochemical potential of the bound states, progressively adding electrons in the order R_{β}, R_{β}, R_{α}, R_{α}, R_{β}, L (along the V_{R} = V_{L} line).
From charge stability diagrams using different combinations of gates we extract the ratio of the lever arms of each gate set to the donor dots. We then use these ratios, in combination with electrostatic finite element modelling, to triangulate^{2} and thus confirm the positions of the three donor dot sites L, R_{α} and R_{β} (see Supplementary Note 3). We also calculate the addition energies from the stability diagram, and confirm that the incorporation sites R_{α}–R_{β} indeed consist of a 1P–2P molecular configuration, see Supplementary Note 4. We find that whilst each site has welldefined charge states there is strong capacitive coupling between them with a mutual charging energy \({E}_{{R}_{\alpha }{R}_{\beta }} \sim 40\) meV. Moreover, we find that the spin parity is determined by the total number of electrons across both R_{α}–R_{β} and not their individual electron number. For example, if R_{β} were an isolated quantum dot, spin readout^{17,18} should be possible when R_{β} is hosting 1 (odd) electron, e.g. in the [0,1,1] charge region. However, we observe that spin readout is only possible in the purple shaded regions in Fig. 1d where the total spin count of both R_{α} and R_{β} is odd. This result strongly suggests that the unpaired spin states on R_{α} and R_{β} form molecular singlet states, arising from the strong exchange coupling J between them. For spin readout we used a magnetic field of B_{0} = 2.5 T, allowing us to estimate that J > gμB_{0} ~ 70 GHz.
The ESR results we present next are obtained for the [L; R_{α}; R_{β}] = [0; 2; 3] fiveelectron charge state, at the position of the red triangle marker in Fig. 1d, where the first four electrons in the molecule occupy magnetically inactive singlet states. We note that we obtained ESR spectra for different electron occupations of the donor molecule with hyperfine values that are in agreement with tight binding calculations of the 1P–2P molecular state and this work will be published elsewhere.
Spinlattice relaxation in donor based qubits is known to be limited by acousticphonon mediated valley repopulation^{19}, with a dependence of the relaxation rate T_{1} on the magnetic field B_{0} of the form \({T}_{1}^{1}={K}_{5}{B}_{0}^{5}\). As shown in Fig. 2a, we find a spinlattice relaxation time of T_{1} ≈ 16 s (at B_{0} = 1.45 T), and a proportional constant K_{5} = 0.0072 T^{−5} Hz for the 1P–2P molecule. We can compare these results to earlier measurements for a single P atom where T_{1} ≈ 6 s^{17,18} and T_{1} ≈ 15 s for a 2P quantum dot at B_{0} = 1.5 T^{10}. Whilst the values are very similar we note that these earlier T_{1} measurements were recorded for the first electron on either the single donor or the 2P quantum dot. This is in contrast to the fifth electron on our molecular 1P–2P system. Since the fifth electron is most likely to be delocalised across the molecule, our results represent the first spin relaxation measurements of a molecular 1P–2P system and as such are distinct from previous results.
Since the long spin lifetime does not allow for fast initialisation by relaxation, we employ a fourlevel pulse sequence based on controlled tunnelling to and from the SET, which we show in Fig. 2b. The spin state is first emptied, leaving the molecule in a spin 0, even parity charge state (four electrons total). To initialise the spin, an odd parity electron is loaded deterministically spindown by aligning the SET Fermi energy between the two Zeeman split spin \(\frac{1}{2}\) states. This state is then isolated from the SET by electrostatically plunging far below the Fermi energy during the application of an ESR control pulse via the microwave antenna. Finally, the Zeeman split state is brought near the system’s Fermi energy, producing a detectable change in I_{SET} depending on the spin state. Each full sequence takes 24 ms, limited by the electron tunnel rate (~3 kHz) to and from the SET, with a single example time trace shown in Fig. 2c. We obtain statistics of the fraction of spinup outcomes p_{↑} by repeating this pulse sequence.
The molecule’s ESR spectrum for the fiveelectron state is shown in Fig. 2d. The figure plots p_{↑} as a function of microwave frequency f_{mw} for an external field of B_{0} = 1.45 T, microwave power of P_{MW} ~18 dBm (before ~70 dB attenuation in the transmission line) and microwave pulse duration of 330 ns using a single tone at the given frequency. As shown in more detail in the inset which focuses on the centre region (the shaded area marks the frequency range of the inset), we only observe one resonance at a frequency of f_{mw} = 40.394 GHz. Superimposed is a fit to a single gaussian peak (red line) with a FWHM of δf_{FWHM} = 7 ± 1 MHz. The linewidth can be attributed to Overhauser fluctuations of the background spin bath of ^{29}Si atoms in natural silicon^{20}.
The existence of a single peak in the ESR spectrum of Fig. 2d has two potential explanations. Firstly, that the hyperfine coupling between the P donor nuclear spins and the bound electron spin is small. It has been shown that the hyperfine coupling of the third electron on a 2P quantum dot is reduced by a factor of ≈10 compared to a single electron on a single P atom^{21}. Here the outer electron spin is screened from the positively charged donor cores by the spinpaired electrons of the inner orbitals and results in a lateral spread of the outer electron wavefunction and a reduced charge density and hyperfine interaction at the central nuclear sites^{11}. For the fifth electron on our 1P–2P molecular state an even larger reduction in hyperfine coupling might be expected due to the presence of four electrons shielding the inner core. Depending on the exact geometrical donor configuration^{21}, the hyperfine coupling in donor molecules with one electron typically range between tens to hundreds of MHz. With five electrons on our 1P–2P molecule the hyperfine coupling is likely to be no more than a few MHz. This value is comparable to the Overhauser field broadened ESR linewidth in ^{nat}Si and prevents us from resolving any splitting of the peak experimentally.
Whilst the observation of a single ESR line can be explained by the reduced hyperfine splitting from the increased lateral extent of the wavefunction of the fifth electron, we cannot rule out the alternate explanation that only one nuclear spin configuration is observed in our measurement. This could arise from a nuclear spin pumping effect driven by the repeated loading and excitation sequence^{22}, effectively forcing the phosphorus nuclei into a fixed spin configuration during the measurement.
We measure the electron gfactor by varying the external magnetic field B_{0} and observing the resonance condition hf_{mw} = gμ_{B}B_{0} (with h Planck’s constant and μ_{B} = 57.9 μeV/T the Bohr Magneton). We find g = 1.99 ± 0.02 in agreement with the accepted value for phosphorus donors in silicon^{23}, with an uncertainty limited by the calibration of our superconducting magnet, see Supplementary Note 5.
In the coherent driving experiments, we employ an interleaved measurement method where we constantly toggle the driving frequency between the two frequencies marked in Fig. 2d, f_{mw,on} = 40.394 GHz (on resonance) and f_{mw,off} = f_{mw,on} − 20 MHz (off resonance). In this way we constantly measure a background signal when exciting offresonance, which allows us to display the results as Δp_{↑} = p_{↑}(f_{mw,on}) − p_{↑}(f_{mw,off}), independent of classical effects such as heating, charge noise or bias voltage drift, which can alter the background spinup fraction over time. Both p_{↑}(f_{mw,on}) and p_{↑}(f_{mw,off}) are obtained using a singleshot readout protocol shown in Fig. 2c.
Coherent control is achieved by varying the pulse duration for fixed microwave power and frequency and extracting the spinup fraction using the singleshot spin readout method outlined above. The observed Rabi oscillations in Fig. 3a–d for different settings of microwave power, exhibit nonexponential decay envelopes due to the fluctuating Overhauser field of ^{29}Si nuclear spins. The instantaneous Overhauser field component parallel to B_{0} contributes an unknown detuning relative to the expected electron resonance condition, causing the spin to accumulate phase at an unknown rate. This detuning also modifies the rate of driven spin rotation during an applied ESR pulse, as greater detuning tilts the effective rotation axis out of the plane perpendicular to B_{0}.
For this experiment we are in the socalled weakdriving regime, where the Rabi frequency (≤1.2 MHz) is less than the width of the Overhauser field distribution (7 MHz)^{24,25}. In this regime, the decay profile for times beyond the first oscillation period is well described by a power law envelope together with a phase shift of \(\frac{\pi}{4}\). From fits to the formula:
with a, c, f_{R} as free parameters, indicated by dashed black lines in Fig. 3a–d, we derive the ensemble average Rabi frequency f_{R} for each power setting. Parameters a and c describe the visibility of the oscillations, and reflect the variation in Overhauser field values across the time ensemble of singleshot measurements together with state preparation and readout errors.
In total ≈11,400,000 singleshot traces are represented in these graphs, from which we extract the expected linear dependence between the amplitude of the driving field (\(\sqrt{{P}_{{\mathrm{MW}}}}\)) and the Rabi frequency as shown in Fig. 3e. At the highest power setting we see a Rabi frequency of f_{R} = 1.18 ± 0.02 MHz, and on average an initial π rotation within about 376 ns.
The Overhauser field variations can be partially overcome using spin echo techniques. In Fig. 3f we demonstrate a spin echo sequence (as indicated in the inset) consisting of a \(\frac{\pi }{2}\) rotation from the initial spindown state, a short delay of time τ = 50 ns, a refocusing π pulse which serves to correct for any unknown phase accumulated throughout the sequence, and following a second delay of τ, a final \(\frac{\pi }{2}\) rotation of varying phase ϕ relative to the first two pulses. For ϕ = 0, the three pulses together result in a total rotation of angle \(\frac{\pi }{2}+\pi +\frac{\pi }{2}=2\pi\) about the Xaxis and the final spin outcome is spindown. For ϕ = π, the final pulse rotates the state in the opposite direction, leaving the final state as spinup. For intermediate values of the phase angle the final rotation leaves the spin state as a superposition. The data is fit to a sinusoid:
superimposed as a black dashed line. The result in Fig. 3f demonstrates two axis control over the state of the electron spin, enabling any single qubit unitary transformation by a combination of X (ϕ = 0) and Y (\(\phi =\frac{\pi }{2}\)) axis rotations.
We also study the coherence properties of our spin system by applying compound pulse sequences. We perform a Hahn echo experiment, in which we initialise the electron spindown, then apply \(\frac{\pi}{2}, {\pi}\) and \(\frac{\pi}{2}\) pulses about the Xaxis, separated by time τ as shown in the inset to Fig. 4a before spin readout. Operating at P_{MW} = 18 dBm, we set the π time in the Hahn echo experiment to t_{π} = 376 ns and vary the free precession time 2τ in Fig. 4a. For short times, we recover the spindown state, and for very long evolution times, the spin state becomes completely randomised due to complete dephasing caused by uncontrolled fluctuations in the spin bath. The transition between the two regimes is described by a compressed exponential^{26} of the form:
shown as a black dashed line. While fit parameters c and a define the visibility, the coherence properties are captured by an exponent n = 2.6 ± 1 and the Hahn echo decoherence time T_{2} = 298 ± 30 μs which are both remarkably close to values measured for phosphorus in bulk ^{nat}Si^{26}. This result indicates that decoherence is mainly due to a slowly fluctuating bath of ^{29}Si nuclear spins^{27}.
Changing the timing of the last projection pulse by introducing a deviation of δτ from a fixed free evolution time τ = 3 μs (shown in the inset to Fig. 4b results in a Ramseytype spin echo envelope experiment. Phase coherence with the Larmor precession due to B_{0} is lost if the instantaneous Overhauser detuning is comparable to \(\frac{1}{\delta \tau }\). We sample this spin dephasing across the time ensemble of single shot measurements to determine the pure dephasing time \({T}_{2}^{* }\). Fitting the data in Fig. 4b to a gaussian distribution with standard deviation σ, we obtain \({T}_{2}^{* }=\sqrt{2}\sigma =295\pm 23\) ns. This value is about five times longer than observed in an ion implemented device measuring an individual P phosphorus atom^{20} in natural silicon. We note however that electron spin dephasing can be influenced by interaction with the nuclear spin bath, which depends on the time dynamics of the hyperfine interaction under a specific experimental pulse sequence^{28}. Accurately modelling the decoherence time would require a full simulation of the multielectron fewdonor wavefunction^{29}, along with a method such as correlated cluster expansion^{30} to account for nonclassical memory effects in the interacting ^{29}Si atoms. However, we can qualitatively attribute the extended dephasing time \({T}_{2}^{* }\) to the reduced electrostatic confinement and thereby more extended wave function of our fiveelectron molecular state. This results in a weaker coupling to a larger ensemble of nuclear spins, producing a narrower distribution of Overhauser field values. At the same time, our molecular confining potential does not significantly impact the concentration of ^{29}Si nuclear spins or the timescale of their flipflop dynamics, leaving the T_{2} decoherence time observed via Hahn echo comparable to that seen for a single phosphorus in bulk ^{nat}Si^{26}.
In conclusion, we have demonstrated coherent control of an electron spin bound to a 1P–2P donor molecule in isotopically natural silicon. We found that the inhomogeneous dephasing time \({T}_{2}^{* }\) is five times longer than that observed for a similar experiment performed on a single P donor. Along with the intrinsic addressability of donor molecules, and their advantages for achieving controllable interqubit exchange coupling, we confirm that multielectron states bound to donor molecules provide a longlived coherent spin resource for quantum computing. This work confirms that the presence of numerous dopant atoms within the planar conductive leads of a donordefined readout SET, along with multiple nuclei within the donor qubit, does not limit its coherence.
Methods
Device fabrication
The STM hydrogen lithography was done at a pressure below 1e^{−11} mbar with an Omicron Variable Temperature instrument. A chemically cleaned Si(001) wafer was passivated in a beam of atomic hydrogen after a 1100 ^{∘}C reconstruction anneal. The hydrogen mask was selectively removed by scanning with a tip voltage of 3–6 V and current setpoint of 1–10 nA. After lithography, phosphine dosing and imaging, the wafer was heated to 350 ^{∘}C before a 55 nm layer of epitaxial silicon was grown at a rate of 0.15 nm/min to encapsulate the incorporated donors. The donor layer was electrically contacted by depositing aluminium over contact vias produced by reactive ion etching. The contact structures, as well as the microwave antenna were all defined by electron beam lithography using a PMMA mask.
Cryogenic measurements
Spin resonance measurements were performed at 50 mK in a dilution refrigerator. A superconducting solenoid magnet provided the external magnetic field. DC voltage offsets were generated by Yokogawa 7651 and Stanford Research Systems SIM928 voltage sources. Timevarying voltage pulses were generated by a National Instruments USB6363 DAC/ADC device and added to the static offsets via resistive voltage dividers. The combined gate control signals were filtered by twostage lumped element RC filters inside the dilution fridge with a lowpass cutoff of 150 kHz. The microwave signals were supplied to the onchip antenna from a Keysight E8267D vector signal generator (with phase and pulse modulation signals supplied by a Tektronix 5014C arbitrary waveform generator) via a lossy stainless steel coaxial cable running from room temperature to 50 mK. The SET current readout signal was amplified by a Femto DLPCA200 transimpedance amplifier and then electrically decoupled and filtered by a Stanford Research Systems SIM910 JFET isolation amplifier and SIM965 Bessel filter before being digitised by the National Instruments USB6363 DAC/ADC.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
Kane, B. E. A siliconbased nuclear spin quantum computer. Nature 393, 133–137 (1998).
Pakkiam, P., House, M. G., Koch, D. M. & Simmons, M. Y. Characterisation of a scalable donor based singlettriplet qubit architecture in silicon. Nano Lett. 18, 4081–4085 (2018).
Pla, J. J. et al. A singleatom electron spin qubit in silicon. Nature 489, 541–545 (2012).
Muhonen, J. T. et al. Storing quantum information for 30 seconds in a nanoelectronic device. Nat. Nanotechnol. 9, 986–991 (2014).
Gamble, J. K. et al. Multivalley effective mass theory simulation of donors in silicon. Phys. Rev. B 91, 235318 (2015).
Song, Y. & Das Sarma, S. Statistical exchangecoupling errors and the practicality of scalable silicon donor qubits. Appl. Phys. Lett. 109, 253113 (2016).
He, Y. et al. A twoqubit gate between phosphorus donor electrons in silicon. Nature 571, 371–375 (2019).
Wang, Y. et al. Highly tunable exchange in donor qubits in silicon. npj Quant. Inf. 2, 16008 (2016).
Nielsen, E., Barnes, E., Kestner, J. & Sarma, S. D. Sixelectron semiconductor double quantum dot qubits. Phys. Rev. B 88, 195131 (2013).
Watson, T. F. et al. Atomically engineered electron spin lifetimes of 30 s in silicon. Sci. Adv. 3, e1602811 (2017).
Hile, S. J. et al. Addressable electron spin resonance using donors and donor molecules in silicon. Sci. Adv. 4, eaaq1459 (2018).
Fuechsle, M. et al. A singleatom transistor. Nat. Nanotechnol. 7, 242–246 (2012).
Abrosimov, N. et al. A new generation of 99.999% enriched 28si single crystals for the determination of avogadro’s constant. Metrologia 54, 599 (2017).
Rueß, F. J. et al. Narrow, highly pdoped, planar wires in silicon created by scanning probe microscopy. Nanotechnology 18, 044023 (2006).
Elzerman, J. M. et al. Singleshot readout of an individual electron spin in a quantum dot. Nature 430, 431–435 (2004).
Wilson, H. F. et al. Thermal dissociation and desorption of PH_{3} on Si (001): a reinterpretation of spectroscopic data. Phys. Rev. B 74, 195310 (2006).
Morello, A. et al. Singleshot readout of an electron spin in silicon. Nature 467, 687–691 (2010).
Watson, T. F., Weber, B., House, M. G., Büch, H. & Simmons, M. Y. Highfidelity rapid initialization and readout of an electron spin via the single donor D^{−} charge state. Phys. Rev. Lett. 115, 166806 (2015).
Hasegawa, H. Spinlattice relaxation of shallow donor states in ge and si through a direct phonon process. Phys. Rev. 118, 1523–1534 (1960).
Pla, J. J. et al. A singleatom electron spin qubit in silicon. Nature 489, 541–545 (2012).
Wang, Y., Chen, C.Y., Klimeck, G., Simmons, M. Y. & Rahman, R. Characterizing Si:P quantum dot qubits with spin resonance techniques. Sci. Rep. 6, 31830 (2016).
Pla, J. J. et al. Highfidelity readout and control of a nuclear spin qubit in silicon. Nature 496, 334–338 (2013).
Feher, G. Electron spin resonance experiments on donors in silicon. i. electronic structure of donors by the electron nuclear double resonance technique. Phys. Rev. 114, 1219 (1959).
Koppens, F. H. L. et al. Universal phase shift and nonexponential decay of driven singlespin oscillations. Phys. Rev. Lett. 99, 106803 (2007).
Yoneda, J. et al. Fast electrical control of single electron spins in quantum dots with vanishing influence from nuclear spins. Phys. Rev. Lett. 113, 267601 (2014).
Tyryshkin, A. M. et al. Coherence of spin qubits in silicon. J. Phys. Condens. Matter 18, S783–S794 (2006).
Witzel, W. M. & Sarma, S. D. Quantum theory for electron spin decoherence induced by nuclear spin dynamics in semiconductor quantum computer architectures: spectral diffusion of localized electron spins in the nuclear solidstate environment. Phys. Rev. B 74, 035322 (2006).
Madzik, M. T. et al. Controllable freezing of the nuclear spin bath in a singleatom spin qubit. Sci. Adv. 6, eaba3442 (2020).
Tankasala, A. et al. Twoelectron states of a groupv donor in silicon from atomistic full configuration interactions. Phys. Rev. B 97, 195301 (2018).
Yang, W., Ma, W.L. & Liu, R.B. Quantum manybody theory for electron spin decoherence in nanoscale nuclear spin baths. Rep. Prog. Phys. 80, 016001 (2016).
Acknowledgements
We thank Bill Coish, Sue Coppersmith, Samuel Gorman and Rajib Rahman for useful discussions. The research was supported by the Australian Research Council Centre of Excellence for Quantum Computation and Communication Technology (project number CE170100012), the US Army Research Office under contract number W911NF1710202 and Silicon Quantum Computing Pty Ltd. M.Y.S. acknowledges an Australian Research Council Laureate Fellowship. This work was performed in part at the NSW node of the Australian National Fabrication Facility.
Author information
Authors and Affiliations
Contributions
S.J.H. and P.P. fabricated the device. L.F., S.J.H., L.K. and Y.C. performed the measurements. L.F., S.J.H., L.K., Y.C., Y.H., M.G.H. and J.G.K. analysed the data. The manuscript was written by S.J.H., L.F., J.G.K. and M.Y.S. with input from all other authors. M.Y.S. planned and supervised the project.
Corresponding author
Ethics declarations
Competing interests
M.Y.S. is a director of the company Silicon Quantum Computing Pty Ltd. The remaining authors declare no competing interests.
Additional information
Peer review information Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Fricke, L., Hile, S.J., Kranz, L. et al. Coherent control of a donormolecule electron spin qubit in silicon. Nat Commun 12, 3323 (2021). https://doi.org/10.1038/s41467021236623
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467021236623
This article is cited by

Highfidelity initialization and control of electron and nuclear spins in a fourqubit register
Nature Nanotechnology (2024)

Spin Splitting and Band Gap Structure in Si[110] Nanowires Doped with Impurities
Journal of Electronic Materials (2024)

Uncovering anisotropic effects of electric highmoment dipoles on the tunneling current in \(\delta\)layer tunnel junctions
Scientific Reports (2023)

Noisy intermediatescale quantum computers
Frontiers of Physics (2023)

Engineering topological states in atombased semiconductor quantum dots
Nature (2022)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.