WorksPhysics and Astronomy
Development of 15kA/cm Fabrication Process for Superconducting Integrated Digital Circuits
Abstract
A new fabrication process for superconducting integrated digital circuits is reported. We have developed the "SIMIT Nb04" fabrication technique for superconducting integrated circuits with Nb-based Josephson junctions based on the validated "SIMIT Nb03" process and Chemical Mechanical Planarization (CMP) technology. Seven Nb superconducting layers and one Mo resistor layer are included in the "SIMIT Nb04" process with 19 mask levels. The device structure is composed of active layers including junctions at the bottom, two passive transmission line (PTL) layers in the middle and a DC power layer at the top. The circuit fabrication started with the fabrication of Mo resistors with a target sheet resistance Rsh of 3 , followed by the deposition of Nb/Al-AlO/Nb trilayer Josephson-junction with a target critical current density Jc at 15 kA/cm. To increase the Al-AlO barrier layer etching's repeatability, an additional barrier protection layer was applied. To accomplish high-quality planarization, we created a planarization procedure coupled with dummy filling. To assess the process dependability and controllability, a set of process control monitors (PCMs) for monitoring fabrication and design parameters was designed and monitored. The successful manufacturing and testing of a few small-scale circuits, like our standard library cells, further attests to the viability of our fabrication process for superconducting integrated circuits.
INTRODUCTION
Rapid single flux quantum (RSFQ) circuits have the advantages of high operating frequencies and lower power consumptions. Numerous efforts have been made to develop RSFQ large-scale integrated (LSI) circuit fabrication technologies mainly based on Nb/Al-AlO/Nb Josephson junctions (JJs). For example, MIT Lincoln Laboratory [1]-[3] and HYPRES [4]-[6] in the USA, National Institute of Advanced Industrial Science and Technology in Japan [7]-[9], Leibniz Institute of Photonic Technology in Germany [10] have reported such circuits.
Using our usual fabrication method, we have been manufacturing several superconducting single flux quantum (SFQ) circuits (SIMIT Nb03 [11][12]). In this paper, we report on the development of our new fabrication process (SIMIT Nb04). Seven Nb superconducting layers and one Mo resistor layer are included. The process started with the fabrication of Mo resistors with a target sheet resistance of 3 , followed by the deposition of Nb/Al-AlO/Nb trilayer Josephson-junction with a target critical current density at 15 kA/cm. The wafer process flow is briefly discussed. Successful fabrication and testing of a few small-scale circuits are demonstrated.
FABRICATION PROCESS
Overview
The target parameters of the layers and the minimum feature sizes are summarized in Table 1. The process includes a resistor layer RN0, a trilayer MN0/AN0/JN0, a wiring MP1, two passive transmission line (PTL) layers in the middle, two ground plane MP2 and MP5, a DC power layer at the top, a contact pad metallization PP2. Nb/Al-AlO/Nb trilayers were fabricated in a multi-chamber dc magnetron sputtering system, in which Nb and Al films were deposited in two separated vacuum chambers, and the load-lock chamber was used for aluminum oxidation. The thickness of AlO was controlled by exposure time and the O partial pressure, which corresponds to the target critical current density for 15 kA/cm. Mo resistors with the target sheet resistance of 3 was deposited by DC magnetron sputtering. SiO interlayer insulators are deposited with a plasma-enhanced chemical vapor deposition (PECVD). The patterning is achieved by using a new i-line stepper Canon FPA-3000 EX4 with 248 nm exposure wavelength. Nb film is etched by Inductively Coupled Plasma (ICP) etching system with endpoint detector, while the SiO insulator is etched by Reactive Ion Etching system (RIE).
| Layer | Material | Description | Thickness (nm) | Min. size (μm) | Min. space (μm) |
| RN0 | Mo | Resistor | 30 ± 5 | 1.6 | 1.0 |
| CN0 | SiO | Insulation | 100 ± 10 | 1.0 | 1.0 |
| MN0 | Nb | Base electrode | 150 ± 5 | 1.0 | 1.0 |
| AN0 | Al-AlO | Junction barrier | 12 ± 0.5 | 1.5 | 2.0 |
| JN0 | Nb | Counter electrode | 150 ± 5 | 0.5 | 1.0 |
| IN0J | SiO | Insulation | 70 ± 10 | 0.7 | 1.0 |
| IN0M | SiO | Insulation | 220 ± 10 | 1.0 | 1.0 |
| MP1 | Nb | Superconductor | 300 ± 15 | 1.0 | 0.8 |
| IP1 | SiO | Insulation | 120 ± 10 | 1.0 | 1.0 |
| MP2 | Nb | Superconductor | 200 ± 10 | 1.0 | 0.8 |
| IP2 | SiO | Insulation | 120 ± 10 | 1.0 | 1.0 |
| MP3 | Nb | Superconductor | 200 ± 10 | 1.0 | 0.8 |
| IP3 | SiO | Insulation | 120 ± 10 | 1.0 | 1.0 |
| MP4 | Nb | Superconductor | 200 ± 10 | 1.0 | 0.8 |
| IP4 | SiO | Insulation | 120 ± 10 | 1.0 | 1.0 |
| MP5 | Nb | Superconductor | 200 ± 10 | 1.0 | 0.8 |
| IP5 | SiO | Insulation | 120 ± 10 | 1.0 | 1.0 |
| MP6 | Nb | Superconductor | 200 ± 10 | 1.0 | 0.8 |
| PP2 | Ti/Au | Contact pad | 108 ± 5 | 3.0 | 3.0 |
Table 1. LAYER PARAMETER AND MASK INFORMATION
For the junction barrier Al-AlO layer etching, most of the fabrication processes uses ion beam etching (IBE) [9] or wet etching [14][15]. However, sidewalls formed in the IBE process may create additional conductive paths in the insulation layer leading to a current leakage in the junctions. Moreover the automatic wet etching used in standard process (Nb03) can avoid this problem, but it may produce some black reaction products. To solve these problems, we used a very thin SiO film as a barrier protective layer and optimized IBE etching parameters to improve the reproducibility of Al-AlO barrier layer etching.
Barrier protective layer process
An additional barrier protective layer was carried out to improve the reproducibility of Al-AlO barrier layer etching and the reliability of the junctions, as shown in Figure 1. Following the inductively coupled plasma RIE (ICP-RIE) formation of the Nb/Al-AlO/Nb trilayer, a 10-nm SiO additional barrier protective layer (SiO) was applied using low-power PECVD. Following the optimized IBE, the AlO tunnel barrier (AN0) was patterned and then etched by RIE with CHF at 2 Pa. After the trilayer patterning, a 600-nm SiO was deposited using PECVD, then the wafer was planarized by CMP. Finally vias in IN0J and IN0M were separately made by RIE.

Figure 1. Schematic illustration of the process steps for a Josephson junction. (a) counter electrodes Nb etching (JN0); (b) a thin SiO film as an additional barrier protective layer was deposited using PECVD; (c) the AlO tunnel barrier (AN0) was patterned, than etched by RIE and IBE; (d) base electrode Nb etching (MN0); (e) SiO deposition using PECVD; (f) planarization by CMP; (e) SiO etching for vias in IN0J.
Circuit fabrication
After 30-nm Mo (RN0) was first fabricated onto 100mm diam Si wafer with 300nm-thick thermal oxidation SiO on the surface by dc magnetron sputtering and reactive ion etching (RIE), a 100-nm SiO film (CN0) was deposited by PECVD, and etched using RIE with CHF at 2 Pa. Then Nb/Al-AlO/Nb trilayers was deposited. First, an 150-nm-thick Nb film (MN0) was deposited in 0.7Pa Ar gas atmospheres and 2.0 A sputtering current with water cooling the substrate during sputtering. Then a 12-nm-thick Al layer was deposited at = 0.5 Pa and a sputtering current of 0.5 A. The Al film surface wad oxidized by Ar and 15%-O mixture gas in load-lock chamber at pressure of 4 Pa for 10 to 20 mins. The counter Nb layer with 150nm thickness was deposited on the AlO to finish junction. The counter Nb film was etched by inductively coupled plasma RIE (ICP-RIE). Then a 10-nm SiO additional barrier protective layer (SiO) was deposited using low-power PECVD and tched by RIE
with CHF at 2 Pa. The optimized IBE was used to pattern the AlO tunnel barrier (AN0).
After 600-nm SiO deposited and the planarized, vias in IN0J and IN0M layers were separately etched by RIE. The Nb wiring (MP1) with 300-nm thickness was deposited and patterned. Then another 600-nm SiO layer (IP1) was deposited, and vias in IP1 were fabricated with the same parameters after the planarization. A 200-nm Nb ground plane (MP2) was deposited and patterned. IP2-MP6 was formed using the same processing conditions as IP1 and MP2. Figure 2 shows a scanning electron microscope image of a cross-section of the fabricated circuit.

Figure 2. Cross-section of a device fabricated by SIMIT standard fabrication process.
During the wafer fabrication process, the process parameters such as lithography resolutions and alignment shifts, etched depths, and actual film thicknesses have been monitored using the on-wafer process control monitor (PCM) patterns. Details of the PCM fabrication technology evaluations are presented elsewhere [16].
TEST RESULTS
The fabrication process has been evaluated by electrical testing using PCM circuits [16]. The probability of the flaws as well as circuit parameters like and was selected as evaluation criteria. It has become a standard to provide feedback on the PCM results to advance the fabrication technology.
By submerging chips in liquid helium at 4.2 K, the current-voltage (I-V) of circular junctions of various diameters was measured using a four-wire measurement method. Figure 3 shows typical I-V characteristics for Nb/Al-AlO/Nb junction. Parameters related to the junction property were extracted from current-voltage curves (I-V curves) of un-shunted JJs.
Figure 3. Typical I-V characteristics for Nb/Al-AlO/Nb junctions fabricated by our standard process.
Josephson junctions’ critical current density has an exponential relationship with those conditions as shown in Fig. 4. can be empirically fitted to [17], where denotes oxidation conditions (product of oxidation pressure and oxidation time) of Al layer in Nb/Al-AlO/Nb. By fitting experimental data on empirical formula from the process line, an accurate result of designed critical current density can be easily gotten, which is a useful guide for following SFQ device development and fabrication process. It should be noticed that the oxidation curves of different oxidation layers are distinct and partial pressure of oxygen also have a different exponential relationship at high or low region. Utilizing pure O for the oxidation process, the resulting in the low- region changes to 1.5 in the high- zone using oxidations in 15%O-Ar mixtures. The value in the high- region is lower than our earlier reported data [11]. The reason for this is that to make the fitted curve more realistic, we have incorporated a lot more experimental data.
Small-scale circuits were designed and fabricated using our newly 15-kA/cm Nb junction technology SIMIT Nb04. OCTOPUX, an automated system specifically for evaluating superconducting circuits, was used to test the circuits at low frequencies. The margin comparison of the same circuit fabricated using various processes is shown in Figure 5. The typical experimental DC bias margins were in the range of -40% to +40% for most of the gates. The same cell using different processes shows similar bias margins. This clearly illustrates the suitability of the novel technique for superconducting integrated digital circuits; it may enhance circuit integration without lowering circuit quality.

Figure 5. Margin comparison of the same circuit fabricated by different process.
CONCLUSION
Based on the validated "SIMIT Nb03" method and CMP technology, we have developed a new 15 kA/cm fabrication technique "SIMIT Nb04" featuring Nb/Al-AlO/Nb Josephson junctions for superconducting integrated circuits. The critical current density and minimum diameter of the junction achieved are 15 kA/cm and 0.5 m, respectively. The fabrication parameters for the “SIMIT Nb04” process have been systematically optimized. To increase the consistency of Al-AlO barrier layer etching and the reliability of sub-micron connections, a second barrier protection layer was used. We investigated the dependence of on O exposure for high current density , whose values were determined with PCM junctions. Several small-scale circuits have been successfully constructed and tested at low frequencies, like RSFQ logic gates. The typical experimental operating margins for the bias current were found to be similar to those from “SIMIT Nb03” process. As a result, it is demonstrated that the superconducting integrated digital circuits can be produced using our innovative technique, "SIMIT Nb04."
ACKNOWLEDGMENT
The authors would like to thank Masaaki Maezawa for the valuable discussion. The fabrications were performed in the Superconducting Electronics Facility (SELF) of Shanghai Institute of Microsystem and Information Technology.
REFERENCE
References
- [3]S. K. Tolpygo, E. B. Golden, T. J. Weir, and V. Bolkhovsky, “Inductance of superconductor integrated circuit features with sizes down to 120 nm”, Supercond. Sci. Technol., vol. 34, 085005, June 2021;DOI
- [6]S. K. Tolpygo, D. Amparo, “Fabrication process development for superconducting VLSI circuits: minimizing plasma charging damage”, J. Phys.: Conf. Ser., vol. 97, 01227, 2008;DOI
- [9]M. Hidaka, S. Nagasawa, K. Hinode, and T. Satoh, “Device Yield in Nb-Nine-Layer Circuit Fabrication Process”, IEEE Trans. Appl. Supercond., vol. 23, No. 3, 1100906, June 2013;DOI
- [10]Rapid Single Flux Quantum (RSFQ) – Design Rules for Nb/Al₂O₃-Al/Nb-Process at Leibniz IPHT, Version 10.03.2017: RSFQ1H-1.6, available online: https://www.fluxonics.de/fluxonics-foundry/;
- [11]L. Ying, X. Zhang, M. Niu, J. Ren, W. Peng, M. Maezawa, and Z. Wang, “Development of Multi-Layer Fabrication Process for SFQ Large Scale Integrated Digital Circuits”, IEEE Trans. Appl. Supercond., vol. 31, No. 5, 1301504, August 2021;DOI
- [12]X. Gao, Q. Qiao, M. Wang, M. Niu, H. Liu, M. Maezawa, J. Ren, and Z. Wang, "Design and Verification of SFQ Cell Library for Superconducting LSI Digital Circuits", IEEE Trans. Appl. Supercond., vol. 31, No. 5, 1101105, August 2021;DOI
- [15]M. Maezawa, M. Chiaramonte, F. Hirayama, and M. Suzuki, “Design and Operation of RSFQ Cell Library Fabricated by Using 10-kA/cm² Nb Technology”, IEEE Trans. Appl. Supercond., vol. 17, No. 2, pp. 500 – 504, June 2007;DOI
- [16]X. Zhang, L. Ying, et. al., “Process Control Monitoring for Fabrication Technology of Superconducting Integrated Circuits”, IEEE Trans. Appl. Supercond., vol. 31, No. 5, 1101206, August 2021;DOI
- [17]A.W. Kleinassser, R.E. Miller, W.H. Mallison, “Dependence of Critical Current Density on Oxygen Exposure in Nb-Al-Nb Tunnel Junctions”, IEEE Trans. Appl. Supercond., 5, 26-30, 1995.
