Implant Depth

Implant depth, formally called projected range (Rp), and junction depth (Xj) are two related but genuinely distinct measurements that together define how deep an implanted dopant actually sits in silicon — Rp marking the average depth of the implanted ion distribution itself, and Xj marking the specific depth where that distribution's concentration falls to match the background doping level, which is the depth that actually defines where the electrical p-n junction forms. Confusing the two is a common mistake with real consequences, because Xj is always deeper than Rp, and it's Xj — not Rp — that directly determines transistor channel length, junction capacitance, and leakage behavior.

Projected range depends jointly on implant energy, ion mass, and target material, with higher energy producing deeper Rp and heavier ions at the same energy producing shallower Rp — boron at 10 keV reaches roughly 35nm while arsenic at the identical 10 keV reaches only about 7nm, directly illustrating how dramatically ion mass alone can shift this depth parameter even when energy is held fixed. This mass sensitivity is exactly why Rp alone can't be specified without also specifying which ion species is involved — quoting a target depth without the corresponding ion species leaves the actual implant energy needed to achieve that depth genuinely undetermined.

The implanted dopant profile approximates a Gaussian distribution centered at Rp with a standard deviation called straggle, and because a Gaussian distribution has tails extending in both directions from its peak, junction depth — defined by where the tail's concentration finally drops to match background doping — is necessarily always deeper than the peak concentration depth Rp itself. This Rp-versus-Xj gap isn't a measurement artifact or an approximation error; it's a direct, unavoidable mathematical consequence of the implanted profile's genuinely Gaussian shape, which is exactly why any specification or discussion of implant depth needs to be explicit about whether it's referring to Rp or Xj — treating the two as interchangeable produces a real, consequential depth error.

Junction depth increases significantly during post-implant annealing beyond whatever Xj the as-implanted profile initially showed, because thermal diffusion during anneal spreads dopants further from their implanted position — meaning the final, device-relevant junction depth is a combination of both the implant's initial profile and however much subsequent thermal diffusion the anneal step introduces. This thermal-diffusion contribution to final junction depth is exactly why implant energy and dose alone don't fully determine a device's final junction depth — the anneal recipe's time-temperature profile matters just as much, and advanced-node processes specifically favor millisecond annealing precisely to activate dopants while minimizing how much additional diffusion that thermal step adds on top of the as-implanted profile.

Advanced CMOS nodes at sub-7nm dimensions target source/drain extension junction depths of just 5 to 15 nanometers, which is only achievable through the combination of ultra-low implant energy in the 0.2 to 2 keV range, heavy ion species like BF2+ or As+ that resist channeling and diffuse slowly, and millisecond annealing that activates dopants with the minimum possible additional diffusion. This specific combination of requirements — extremely low energy, heavy species, and extremely short thermal budget — illustrates exactly how tightly every implant parameter has to be co-optimized at advanced nodes: no single parameter choice alone could achieve a junction depth this shallow without the other two parameters also being pushed to their respective extremes.

SIMS, spreading resistance profiling, and four-point-probe sheet-resistance measurement each characterize a genuinely different aspect of the implanted and annealed profile — SIMS measures total dopant chemical concentration regardless of electrical activation state, SRP measures actual carrier concentration which reflects only electrically active dopants, and sheet resistance provides an integrated electrical measurement related to junction depth through the relationship Rs equals the inverse of charge times mobility times carrier concentration times Xj for a uniform profile. Having all three measurement techniques available matters because they answer genuinely different questions: SIMS can show dopant atoms are physically present in the right place even if some fraction never activated electrically, while SRP and sheet resistance specifically capture whether those atoms are actually contributing to the device's real electrical behavior — a distinction that matters directly for diagnosing whether a disappointing device result traces back to implant or to incomplete activation during anneal.