A review and future predictions

With the advent of quantum computing, the need for fault-tolerant peripheral logic control circuits has reached new heights. In classical calculation, the unit of information is a “1” or a “0”. In quantum computers, the unit of information is a qubit which can be characterized by a “0”, a “1” or a superposition of the two values ​​(called a “superimposed state”).

The control circuit of conventional computers is based on CMOS (semiconductor), due to its high performance and low power dissipation. The “1s” and “0s” of a typical computer can be manipulated, stored, and easily read using CMOS chips that operate at room temperature. Most quantum computers today operate at cryogenic temperatures, to ensure that the qubit remains coherent (in a superimposed state) for as long as possible. Coherence times are typically very short (nanoseconds to milliseconds) in a quantum computer, requiring control circuitry capable of high-speed, fault-tolerant operations. This requirement could be met by a conventional CMOS driver circuit if it could operate at cryogenic temperatures.

The first attempt to characterize semiconductor materials at cryogenic temperatures was made by AK Jonscher in his Proceedings of the 1964 IEEE publication, “Semiconductors at Cryogenic Temperatures”. [1]. His two basic conclusions were: 1) semiconductor devices have no major cryogenic application at this time due to “no real technological justification for going large scale to these extreme temperatures”, and 2) “The properties of semiconductor materials at cryogenic temperatures are so different from familiar properties at higher temperatures, that it is reasonable to expect that many more device applications will emerge as a result of continued efforts research and development in this direction”. A few years later, IBM became interested in the operation of semiconductor devices at low temperatures [2-3] and concluded that MOSFET semiconductor devices exhibit improved performance at cryogenic temperatures. With the advantages of low-temperature operation, downscaling of the cooling device is still a barrier to using semiconductor-based control circuits.

Enter quantum mechanics. In 1959, Richard Feynman challenged the scientific community to employ quantum mechanics in the design of information processing systems. He imagined new information systems and new functions that involved quantized energy levels and/or the interactions of quantized “spins” (angular momentum of quantum particles). His vision came to fruition in the 1980s, when it was shown that energy-based quantum mechanical equations could represent a universal Turing machine (computing). [4]. In 1994 it was shown that a quantum computer could factor integers much faster than a classical computer (“in polynomial time”) [5]. This discovery was the catalyst that sparked continued interest in building quantum computing systems. This interest continues today in many commercial, research and academic organizations.

Even with the strong interest in building quantum computers, the fact remains that proper functioning of this type of computer currently requires a cryogenic temperature environment. Quantum logic driver circuits will also need to operate at these cryogenic temperatures to operate effectively in this environment. Thus, we have seen a renewed interest in the cryogenic temperature performance of CMOS-based circuits.

Quantum computers do not require advanced CMOS circuitry, but CMOS devices operate differently at cryogenic and ambient temperatures. CMOS transistor performance (and associated IV performance) has recently been measured on 40 nm and 160 nm bulk CMOS devices, both at room temperature and at 4.2 degrees Kelvin (see Figure 1). The drive current increases at cryogenic temperatures due to an increase in silicon mobility at these temperatures. Unfortunately, other effects such as substrate freezing can limit the increase in drive current at these low temperatures.


Fig. 1. Measured IV characteristics of nMOS transistors fabricated in 160 nm (left) and 40 nm (right) CMOS. Room temperature operation is shown on the dashed curves, liquid helium operation is shown on the solid curves, and a Spice-compatible model fitted to the experimental data is shown on the dashed lines. (taken from [6])

The control circuits of quantum computers currently operate at room temperature. As mentioned earlier, this can be a problem due to the sensitivity of reading the “state” of qubits at higher temperatures. This challenge can be partially mitigated by operating the CMOS circuits at or near cryogenic temperatures, in the same cryogenic freezers as the quantum computer. This integration can be used to reduce latency and increase overall system scalability. Despite some second-order problems, low-temperature CMOS transistors can perform various functions needed to work with a quantum computer. These features include the ability to function as I/V converters, low pass filters, and A/D and D/A converters (see Figure 2).


Fig. 2. Silicon spin qubit centered in the dotted circle, control and readout signals (M, P, R, T and Q) are shown in the inset. Simplified diagrams of the quantum point contact and corollary circuits are presented. The voltage source is implemented as a room temperature digital-to-analog converter. (taken from [6])

To achieve the desired performance of a fault-tolerant quantum computing system, a new generation of deep submicron CMOS circuits will be needed, operating at deep cryogenic temperatures. [6]. Extrapolating this idea to its logical conclusion, we end up with a quantum integrated circuit (QIC) where the qubit array is integrated on the same chip as the CMOS electronics needed to read the state of the qubits. This integration would clearly be the ultimate goal in achieving scalable, reliable, and high-performance quantum computing.

In more futuristic applications, optical communications to and from the qubit may also be required. In this case, the CMOS integrated circuits will also have to include micro- and nano-optical structures, such as light guides and interferometers. These types of optical functions have been successfully demonstrated on CMOS devices at room temperature. Demonstration of this level of optical communications functionality at cryogenic temperatures may also be desirable in future quantum computing applications.

References
1.) AK Jonscher, “Semiconductors at Cryogenic Temperatures”, Proceedings of the IEEE, 1964.
2.) RW Keyes, et al., “The role of low temperatures in the operation of logic circuits”, Proc. IEEE, vol. 58, pages 1914-1932, 1970.
3.) FH Gaensslen, et al., “Very Small MOSFETs for Low Temperature Operation”, IEEE Trans. Electronic devices, vol. ED-24, p. 218-229, 1877.
4.) P. Benioff, “The Computer as a Physical System: A Microscopic Quantum Mechanical Hamiltonian Model of Computers Represented by Turing Machines,” J. Stat. Physics, vol. 22, no. 5, pages 563-591, 1980.
5.) P. Shor, “Algorithms for quantum computations: discrete log and factorization”, Proc. 35th Annual. Symp. Found. Calculation. Sci., Los Alamitos, CA, 1994, p. 124-134.
6.) E. Charbon, et al., “Cryo-CMOS for quantum computing”, 2016 IEDM, pp. 343-346.

Michael Hargrove

Michael Hargrove

(All posts)

Michael Hargrove is a semiconductor process and integration engineer at Coventor, a Lam research company. He has worked in the field of semiconductor technology development for over 30 years. He started his career at IBM, where he worked on the development of advanced CMOS technology. He then spent five years at Epson Research and Development, working on the design and characterization of high-speed/high-frequency devices. He then joined AMD, where he worked on high-k/metal gate technology. Hargrove then moved to GlobalFoundries Research and Development in Albany, NY. At Coventor, he focuses on 3D modeling of semiconductor processes. Hargrove earned his Ph.D. from the Thayer School of Engineering at Dartmouth College, Hanover, NH