Superconducting qubits are one of the intensive approaches to fabricate quantum computing (QC). Qubits can be manufactured by the current Si-based semiconductor manufacturing processes. However, circuits should be superconducting material such as Nb (Tc=9.29 K), Al (Tc=1.18 K), Pb (Tc=7.2 K), In (Tc=3.4 K) , etc. In current technology, qubits are arranged in planar geometry in a chip, and each qubit should wire to control and read signal lines in the same plane. When a large number of qubits are integrated, these wires would be too constrained, and several problems would be actualized, such as wiring for qubits located at the chip center, crosstalk among wiring, and so on. The signal integrity for the QC devices is much better to have qubits at short distances from the control circuitry. Therefore, 3D wiring is required for a large-scale qubit integration. 3D integration of qubits with interposer chips by flip-chip bonding is the proof-of-concept of the large-scale qubit integration. The 3D integration approaches for qubit- based devices are the same as conventional electronic devices, including the through silicon via (TSV) process for interposers and micro bump flip-chip bonding. However, they use superconducting metals instead of copper and other non-superconducting materials. Most flip-chip microbump bonding processes require thermal compression at over 170 °C. Superconducting elements are easily degraded by heat. Consequently, we apply surface activated bonding (SAB) to a superconducting metal as a low-temperature bonding process for superconducting circuits.The test element group (TEG) for the current-voltage (I-V) measurement of the bonded interface was bonded by SAB method, and at the Nb SAB interface layer is a 2- to 3-nm-thick amorphous layer and the layer consist of Si-rich interfacial layer compared to the Nb bulk area. The bonding specimens were activated on jigs made of Si during the bonding process; then the Si sputtered by FAB was redeposited on the bonding surfaces. Furthermore, the bonded specimen was wired using Al wire to a chip carrier with a stage installed into a 0.3-K cryostat. As a result, the Tc of this path is 9.3 K, which is very similar to that of Nb bulk (9.2 K). However, a small step of the Tc at 9.1-9.4 K was also observed. (Fig.1) [1]In this study, we discuss the superconducting interface of direct bonded Nb-Nb by Tc, and also the interfacial structure measured by polarized neutron reflectometry (PNR). The results show clear spin asymmetry in the superconducting state, as splitting is observed in the fringes of the parallel neutron (neutron spins parallel to the external magnetic field 0.1 T) reflection R+ and the antiparallel (neutron spins antiparallel to it) reflection R−, where the latter has become larger. This well-known effect is related to the flux penetration in the superconducting state. Therefore, unique information about the superconducting state of the buried thin film can be measured by using the polarized neutron beam. For this project the major advantage of PNR will be to characterize the superconducting state at the deeply buried Nb interfaces, sandwiched between the two substrate wafers, which would be infeasible using alternate techniques.[1] M. Fujino, et. al., J. Appl. Phys., 133, 015301, 2023 Figure 1
Read full abstract