Surface Optimization of Aluminum Resonators for Robust Quantum Device Fabrication

TL;DR

This work addresses dielectric losses in aluminum-based superconducting resonators under industry-relevant 24 h cooldown delays by evaluating surface treatments that mitigate TLS-related losses. It combines XPS surface characterization with cryogenic microwave benchmarking to compare oxygen- and fluorine-based passivation and selective chemical etching. Remote oxygen plasma can modestly reduce dielectric loss, while fluorination tends to increase losses at longer Ar milling times; sequential HF vapor and phosphoric acid etching achieves the most substantial reduction, reaching $\delta_{{\mathrm{LP}}} \approx 5.2\times10^{-7}$ ($Q_{{\mathrm{LP}}} \approx 1.9\times10^{6}$) for the lower quartile after a 24 h delay. The results offer a practical pathway for scalable, high-Q aluminum-based quantum devices compatible with industrial fabrication workflows.

Abstract

Aluminum remains the central material for superconducting qubits, and considerable effort has been devoted to optimizing its deposition and patterning for quantum devices. However, while post-processing of Nb- and Ta-based resonators has been widely explored, primarily focusing on oxide removal using buffered oxide etch (BOE), post-treatment strategies for Al resonators remain underdeveloped. This challenge becomes particularly relevant for industry-scale fabrication with multichip bonding, where delays between sample preparation and cooldown require surface treatments that preserve low dielectric loss during extended exposure to ambient conditions. In this work, we investigate surface modification approaches for Al resonators subjected to a 24-hour delay prior to cryogenic measurement. Passivation using self-limiting oxygen and fluorine chemistries was evaluated utilizing different plasma processes. Remote oxygen plasma treatment reduced dielectric losses, in contrast to direct plasma, likely due to additional ashing of residual resist despite the formation of a thicker oxide layer on both Si and Al surfaces. A fluorine-based plasma process was developed that passivated the Al surface with fluorine for subsequent BOE treatment. However, increasing fluorine incorporation in the aluminum oxide correlated with higher loss, identifying fluorine as an unsuitable passivation material for Al resonators. Finally, selective oxide removal using HF vapor and phosphoric acid was assessed for surface preparation. HF vapor selectively etched SiO2 while preserving Al2O3, whereas phosphoric acid exhibited the opposite selectivity. Sequential application of both etches yielded dielectric losses as low as $δ_\mathrm{LP} = 5.2 \times 10^{-7}$ ($Q\mathrm{i} \approx 1.9\,\mathrm{M}$) in the single photon regime, demonstrating a promising pathway for robust Al-based resonator fabrication.

Figures (8)

• Figure 1: Design of the resonator chips, containing nine CPW meanders with frequencies between 4 GHz and 5.6 GHz.
• Figure 2: Exemplary plot of $\delta$ vs. photon number with measured data (blue) from a 4.36 GHz resonator and a fit using Eq. (1) (red).
• Figure 3: Schematic representation of the different post-treatments applied to the resonator chips. Approach A utilized oxygen plasma under different conditions. In approach B, the Al surface was cleaned in-situ using Ar ion milling, followed by plasma processing with CF$_4$. In approach C, HF vapor and H$_3$PO$_4$ acid were applied both individually and in combination.
• Figure 4: Low-power loss $\delta_\mathrm{LP}$ of Al resonators after different oxygen plasma treatments.
• Figure 5: Atomic composition versus Ar ion milling duration prior to CF$_4$ plasma treatment. The surface was cleaned for 1 min using Ar ion milling before the XPS measurements were carried out.