Skip to main content

Optoelectronic detection of single spin in silicon

Periodic Reporting for period 1 - ODESI (Optoelectronic detection of single spin in silicon)

Reporting period: 2016-06-01 to 2018-05-31

The growing sophistication and shrinking size of silicon-based electronic have underpinned the revolutionary developments in information technology over the past decade. Nowadays, actual device performance is compromised by the emergence of quantum effects at the nanoscale. On the other hand, quantum information processing (QIP) provides a way to exploit these limitations and develop a new paradigm that will offer a new technological platform with greater computational power. The building block of QIP is the so-called quantum bit (qubit), which is generally made of a two level system such as an electron spin. A good qubit must be writeable and readable, interact with other qubits, and remain coherent long enough for error correcting protocols to be applied. Out of the large variety of approaches to building a quantum computer currently being pursued, those based in silicon are able to build upon the advanced methods used in silicon microelectronic industry, as well as the control over material purity. During the last decade architectures for spin-based QIP have been proposed such as the Kane’s one which consists in exploiting the nuclear spin of donor atoms implanted into silicon. Interest in donor-based spin-qubits in silicon has been motivated by their exceptionally long electron spin coherence times, exceeding one second in isotopically enriched 28Si. Additionally the donor electron spin can be a gateway to access the donor nuclear spin, which has longer coherence times, even at room temperature and the potential to serve as a quantum memory. Moreover, the single shot read-out of a single electron and nuclear spins, a milestone for donor-based quantum computing, has been recently demonstrated in nanoelectronic silicon devices.
We aim at studying spin-qubits in silicon that will be the basis of a building block for QIP with a great potential for scalability. For this reason, the devices will be designed to be compatible with CMOS industry process and will be fabricated in an industry-oriented cleanroom: the CEA-LETI. More specifically, the main goal is the study of donor spin Qbits in silicon. As a result, the main unit is composed of two elements: a donor atoms and a charge detector. The latter is used to measure the spin of the donor thanks to a conversion of quantum information from spin to charge degree of freedom.
We have built two electrometers able to sense single charge loading/unloading of a nearby capacitively coupled dot. The first spin read-out scheme that we have presented is based on transport measurements of a donor. This technique has allowed us to measure spin states of the sensed dot. In particular, we have studied a singlet/triplet system and have used energy selective read-out to show a T1 = 13.5 ms characteristic relaxation time, consistent with typical values reported for silicon electron pairs. Afterward, we have moved to another spin read-out scheme based on RF reectometry gate-sensing. This dispersive read-out has been used to demonstrate charge read-out of a sensed dot. For this technique, we need the connection to only one reservoir while the transport measurements need two.
Moreover, the presented RF gate-sensing is a promising step toward an easily scalable spin Qbit read-out, completely disconnected to reservoirs. Indeed, we are currently using this dispersive read-out to directly probe a charge exchange between two quantum dots. While here, we have built an electrometer with a dot that need to exchange particle with a reservoir. Thus, we will get rid of any reservoir for the sake of scalability. Once this spin read-out is demonstrated, the next goal is to implement spin-Qbit manipulation schemes. For this purpose, we have developed microwave cavities that will allow us to perform single spin manipulation on many Qbits in parallel.
We have demonstrated the performance an ultra-compact device fabricated in foundry-compatible Si MOS technology, with a built-in charge detector (SET) capacitively coupled to two Gate-defined QDs. Thanks to an energy-selective detection scheme, we have demonstrated single-shot readout of the spin-state in one of the QDs, which is an essential requirement to implement fault-tolerant quantum computing. Further optimization of the readout speed/fidelity trade-off are under investigation. On FDSOI, the possibility of using the back-gate as an additional handle is an asset. Another longer-term advantage is the perspective of reducing the parasitics by seamlessly co-integrating Si qubits with conventional control electronics circuitry.
Measure of spin lifetime in a CMOS device
single shot spin measurement
Artistic view of the donor-dot device