Remus ALBU¹

Abstract. A cryogenic electronics lab is of strategic and economic importance for Romania. It is relatively simple (compared to CMOS, III-V, II-VI, carbon, diamond) technology, it uses simple, robust, verified models. More importantly, the processing know-how, required tools and related professionals are located in Romania. The local universities could easily extend their curriculum to include the Single Flux Quantum (SFQ) and Rapid Single Flux Quantum (RSFQ) devices, circuits and systems lectures, as clearly disseminated (public domain) by Stonybrook New York State University.

We glimpse over the technical advantages of superconductive electronics, applications, technology details and the estimated investment the implementation of a cryogenic electronics lab in Romania.

SOME EXISTING LABS

• Superconductive Electronics Research Lab – Ireland

• Hypres – NY, USA

• MIT Lincoln Lab – Superconductive Electronics – MA, USA

• RSFQ @ SUNY Stony Brook (sunysb.edu)

• Superconducting electronics | NIST

• Superconducting Materials and Devices (nasa.gov)

• CCAS – Coalition for the Commercial Application of Superconductors

• The European Society for Applied Superconductivity –

• SQEL | Superconducting Quantum Electronics Lab (cnr.it)

• TOBB ETÜ Superconductivity Electronics Laboratory (etu.edu.tr)

Yes, these ones are very expensive, with tens of years of accumulated expertise. Yet, as Dr. Sergey Tolpygo – the head of MIT LL lab – mentioned in his 2015 paper², “SFQ superconductor electronics (SCE) has well documented advantages over semiconductor electronics for energy efficiency, clock speed and potential for reversible computing. Over the years, however, these advantages have largely remained un-realized because of low scale of integration and limited sophistication in functionality of superconducting circuits. While the semiconductor industry maintained an exponential growth of electronic circuit density, the density of JJ circuits has remained nearly stagnant for the last 25 years. This stagnation has been attributed in part to the gross disparity in funding of SCE compared with CMOS and the resultant lack of access to modern fabrication tools and systematic process development”.

The same “SCE stagnation” message along with a positive investment outlook forecast is conveyed in the Outlook for Superconductor Computing³: “Superconducting processor research goes back to 1970s, but you don’t build something unless you have to, and CMOS operating at room temperature was much easier to develop first. However, now that Moore’s Law for CMOS is ending, the necessary cryogenics is becoming worth the billion-dollar investment needed to develop superconducting computers.”

ROMANIAN ACADEMY ON ADVANCED TECHNOLOGIES,
INCLUDING SUPERCONDUCTIVE ELECTRONICS

The European Union Council is allocating funds for Romania technological and economic development, including the superconductive electronics area.

The working group under the aegis of the Romanian Academy regarding the ”Important Project of Common European Interest” for MicroElectronics, “IPCEI-ME-2” mentions:

On December 7, 2020, the ministers of member states of the European Union Council (including Romania), signed a Joint Declaration by which “The signatory states have committed to collaborate in order to strengthen in Europe the chain of value creation in electronics and «embedded» systems … to confine in Europe the capabilities for leading the arena of chip design and manufacturing… Investments will be needed from the EU budget, from national budgets (including through the Plans for National Recovery and Resilience – PNRR) and from the private sector.

Microelectronics … is already identified for investments in the RRF. 20% of the NRPS must be allocated to the “digital transition”; the total can amount to €145 billion for the next 2-3 years.”

SUPERCONDUCTIVITY ELECTRONICS ADVANTAGES

At low temperatures – below the superconductivity critical temperature (e.g. 9.25K for Niobium) – the charge scattering (electron – phonon) within a conductive material are reduced, the carriers mobility increase, the electrical resistance decrease to 0. The maximum operation frequency for these superconductive devices and electronic systems is increasing up to hundreds of GHz, while the power dissipation decreases orders of magnitude compared to the room temperature semiconductor based (Ge, Si, SiGe, GaAs, GaN, C, InSb …) counterparts.

Additionally, a plethora of superconductivity effects and associated devices permit the handy implementation of artificial atoms – the background for superconductive quantum computing – and high resolution sensors.

A published assessment of cryogenic computing forecasts the energy con-sumption advantages: “The Superconducting computers could extend Moore’s Law beyond the limits of complementary metal oxide semiconductors (CMOS) by cutting power requirements 100-fold—down to kilowatts for exascale supercomputers, compared to the megawatts required today”.

APPLICATIONS

Cryogenic Computing – “Classic” digital systems

Any High Performance Computer (HPC) figure of merit combines the speed of operations (FLOPS), memory size, and consumed power. As mentioned in the 2015 IARPA C3 research program (https://www.iarpa.gov/index.php/newsroom/article/c3):

“Power and cooling for large-scale computing systems are rapidly becoming unmanageable problems for the enterprises which depend on them. The trend towards large, centralized computing facilities to house supercomputers, data centers, and special purpose computers continues to grow, driven by cloud computing, support of mobile devices, Internet traffic volume, and computation-intensive applications. Conventional computing systems, which are based on complementary metal-oxide-semiconductor (CMOS) switching devices and normal metal interconnects, appear to have no path to be able to increase energy efficiency fast enough to keep up with increasing demands for computation.

Superconducting computing could offer an attractive low-power alternative to CMOS with many potential advantages. Josephson junctions, the superconducting switching devices, switch quickly (~1 ps), dissipate little energy per switch (< 10^-19 J), and communicate information via small current pulses that propagate over super-conducting transmission lines nearly without loss. While, in the past, significant technical obstacles prevented serious exploration of superconducting computing, recent inno-vations have created foundations for a major breakthrough. Studies indicate that superconducting supercomputers may be capable of 1 PFLOP/s for about 25 kW and 100 PFLOP/s for about 200 kW, including the cryogenic cooler.

Superconducting computing research currently consists of a few, scattered efforts with no initiative focused on advancing the field overall. Major research challenges include insufficient memory, insufficient integration density, and no realization of complete computing systems. …

[Cryogenic Computing research] success will pave the way to a new generation of superconducting computers that are far more energy efficient than end-of-roadmap CMOS and scalable to practical application.”

Cryogenic Quantum Computing

Quantum computing emerges as a far better alternative to classic deterministic irreversible computing systems (von Neumann machines). Superconductive devices operated at about 0.1K (dilution ³He/⁴He refrigerators) are commonly used for imple-menting the required superconducting qubits, with relatively high coherence time.

Cryogenic Circuitry for Quantum System Interfaces

It seems practical to handle the flow of information surrounding the quantum engine with superconducting “classic” circuitry. Most of the present solutions (Intel, Google, IBM, D-Wave) are using cooled CMOS circuitry, with high dissipation and low operating frequencies (at this time qubits operate in GHz range, yet the natural trend would be to increase it, therefore Si CMOS will have problems dealing with it …).

Cryogenic Circuitry for mm waves

The frequency cutoff for a superconducting Josephson device (the basic switching non-linear device) is around 1 THz (Nb Al_xO_y Nb structure). That enables the imple-mentation of classic analog and digital circuitry operating up to THz frequencies, dissipating low power, including highly sensitive magnetic sensors, non-linear system, parametric amplifiers, circulators …

Quantum devices

performing as qubits, interferometers are implemented with superconductor Josephson devices. Moreover, most of the quantum qubits implementations – other than the superconducting qubits – do require near 0K operations, i.e. knowhow of cryogenics technology, interface with room temperature systems, measurement and characterization, packaging etc.

MANUFACTURING PROCESSES

Superconductive electronics technology and processing is much simpler compared to any semiconductor device / system manufacturing process. It requires:

• thin film deposition: metals (e.g. Nb, Au, Mo) and dielectrics (SiO₂, Si₃N₄, Al₂O₃, Al_xO_y),

• lithography (machines with 250nm wavelength sufficient … no nm lithography capabilities required),

• etching,

• planarization (with Chemical Mechanical Planarization – CMP) are required.

NOT USED:

• impurity doping (e.g. ion implantations),

• thermal diffusion and annealing (RTP, furnaces),

• oxidations,

• CVD, LPCVD.

One of the most advanced manufacturing processes to date developed at MIT Lincoln Lab⁴ detailed by Tolpygo in a EUCAS2015⁵ paper:

  Figure 1: MIT Lincoln Lab SFQ5ee process cross sections. 8 layers Nb

Classification

Total: 6,400 m² (Class-10: 740 m²; Class-100: 910 m²)

Production-class 90-nm CMOS, 200-mm tool set

Cluster metallization (sputter, MBE, CVD), etch, CMP, …

Advanced lithography: i-line, 248-nm, 193-nm, e-beam

Full SPC, electronic traveler, area engineers, technicians

Three-shift operation (24×5) from 01/2015

~10,000 wafer starts per year

Superconductor electronics (SFQ) fabrication

200-mm tool set

Deep submicron junctions and wiring features

Dedicated metal and dielectric deposition tools

Dedicated metal and dielectric etch tools

Dedicated CMP tool

248-nm and 193-nm photolithography tools (shared)

Metrology and defect inspection tools (shared)

Superconducting multichip module fabrication

Room-temperature testing

Three semi-automatic waferprobers

Switch matrix and test equipment

Automation software

3,000 test structures per PCM

9 PCM chips per wafer

Cryogenic testing

14-chip He immersion probes

5-mm die attached to PCB and wired with automatic wire bonder

Switch matrix and test equipment

Automation software

Dilution refrigerator with 40-cm samples space for mK testing

CryoxCMOS-enabled cryogenic switching

ESTIMATE FOR A SUPERCONDUCTIVITY ELECTRONICS LAB

Below, the estimates for the investment required for building a scaled down version of the SFQ electronics lab described in Tolpygo’s EUCAS2015⁶ paper

Building

TOTAL: $ 3.58 Mil

Processing

TOTAL: $ 7.33 Mil

Cryogenic testing

TOTAL: $ 3.2 Mil

Overall, the estimated investments for a Cryogenic Electronics Research and Design facility is $14.11 Mil.

CONCLUSIONS

The Superconductive Electronics technology is relatively simple (compared to CMOS and other semiconductor variants) and inexpensive.

The scary “it’s close to 0K” is to be changed into “at this time cryogenic tech is expensive, yet clear and established, and all quantum computing devices will need it. Therefore novel technologies are expected to emerge and produce cheaper, power efficient and smaller cryogenic refrigerators”.

A superconductive electronics R&D lab in Romania:

• Generates know-how and specialists to understand and contribute to the expected superconductive supercomputers wave – the “classic” ones take too much power, their future performances are limited by Moore’s law.

• R&D for qubits, quantum computing and support/interface circuits

• Allows the R&D for photonic detector arrays, circuits and systems – with space and military applications

• Enables research and development in the new “circuits quantum electrodynamics” (CQED), which allows to build quantum systems with superconductor devices, and replaces the Kirchhoff equations and SPICE like simulations with a quantum mechanics models. Qubits and quantum computing systems are modeled with CQED.

• Call for strong interactions with universities, where the novel “superconductivity and applications” curriculum is to be developed, for superconductivity theory, devices, circuits and systems and quantum computer software languages, compilers etc.

This endeavor would call for a relatively large investment and time and efforts for growing a mature team of experts. Nevertheless, there are so many foreseeable applications, opportunities and exciting challenges it would be rather unwise not trying to catch this train.