Enhanced Secondary Electron Detection of Single Ion Implants in Silicon Through Thin SiO2 Layers

TL;DR

This work tackles the challenge of deterministically placing single dopants in silicon by enabling real-time, non-destructive detection of single-ion implantation events via secondary-electron emission in a focused ion beam. The authors introduce a thin SiO2 capping layer that substantially boosts SE yield, achieving up to $98\%$ detection efficiency for Sb implants in Si, with ~30 nm spatial resolution and without electrical contacts. They couple calibrated ion-current measurements with a Poisson-based analysis and use TRIM simulations to optimize the oxide thickness, yielding a robust near-unity implant-success probability across practical conditions. The results establish SE-based detection as a scalable route for precise donor placement and extend deterministic implantation to a broad range of materials and quantum-device architectures, including potential removal of the oxide post-implantation.

Abstract

Deterministic placement of single dopants is essential for scalable quantum devices based on group-V donors in silicon. We demonstrate a non-destructive, high-efficiency method for detecting individual ion implantation events using secondary electrons (SEs) in a focused ion beam (FIB) system. Using low-energy Sb ions implanted into undoped silicon, we achieve up to 98% single-ion detection efficiency, verified by calibrated ion-current measurements before and after implantation. The technique attains ~30 nm spatial resolution without requiring electrical contacts or device fabrication, in contrast to ion-beam-induced-current (IBIC) methods. We find that introducing a controlled SiO2 capping layer significantly enhances SE yield, consistent with an increased electron mean free path in the oxide, while maintaining high probability of successful ion deposition in the underlying substrate. The yield appears to scale with ion velocity, so higher projectile mass (e.g. Yb, Bi etc) requires increased energy to maintain detection efficiency. Our approach provides a robust and scalable route to precise donor placement and extends deterministic implantation strategies to a broad range of material systems and quantum device architectures.

Figures (4)

• Figure 1: Secondary electron detection efficiency comparison for varying SiO${_2}$ capping-layer thicknesses (${\tau}$). (a) 25keV Sb$^+$. (b) 50keV Sb$^{2+}$. The plots show the relationship between the average ions per pulse (${\lambda}$) and the average number of detected events per pulse ($\nu={-\ln(p_0)}$, where $p_0$ is the fraction of empty pulses with no implant detections). Each dashed line represents a linear regression fit to the data points for a specific sample, found as described in the text. The detection efficiency (${\eta}$) is given by the gradient of the line, and indicated in the legend. For clarity of display, the intercept from each fit was subtracted off (both the fit and the data), and then a systematically increasing vertical offset was then added to each ${\tau}$ data set. The shaded region around each fit line depicts the standard deviation in the calculated gradient.
• Figure 2: Optimal $\text{SiO}_2$ capping-layer thickness ($\tau$) determination for maximal implant success probability. The secondary electron detection efficiency, $\eta(\tau)$ (blue points, right ordinate axis), data from Figure \ref{['fig:DE_comparison']}. The error bars on these points are too small to see. The blue dashed line is a guide to the eye of form $A-a\exp(-\tau/\tau_0)$. The probability of an ion arriving into the Si substrate, $P_I(\tau)$ (green points, left ordinate axis), was determined from SRIM simulations. The green line is a guide to the eye of form $1 /(1 + (\tau/\tau_0)^k)$. The total implant success probability, $P_S(\tau)$ (red line, left ordinate axis), from the produce of the green and blue lines (according to Eqn \ref{['eq:prob_success']}). This represents the probability of a single implant into the Si substrate and its simultaneous detection by the SE system. The orange shaded region illustrates the thickness range required for bulk SiO$_{2}$, for which all implanted atoms remain in the SiO$_2$, while the grey shaded region indicates the native oxide thickness range. The red shaded area defines the range of $\tau$ that yields a total implant success probability, $P_S(\tau)$, within 99.9% of its maximum value.
• Figure 3: TRIM simulations of Sb in the Si substrate after implantation at (a-b) 25keV and (c-d) 50keV through SiO$_2$. (a) and (c) show the implanted Sb depth profile as a function of SiO$_2$ thickness ($\tau$), with the colour intensity corresponding the normalised frequency density. (b) and (d) show 2D profiles of the normalised frequency density against lateral distance (x) and depth in Si for a 10nm SiO$_2$ thickness, which is at the optimum determined from Figure \ref{['fig:Optimal_tau_determination']}. The origin of the y axis corresponds where the Si interfaces with SiO$_2$.
• Figure 4: Detection efficiency measurements as a function of ion energy for a Si sample with (a) native oxide only and (b) an $\text{SiO}_2$ target.