Implant Dose
Implant dose is the total number of dopant ions delivered per unit area of wafer surface, expressed in ions per square centimeter, and it's the parameter that most directly controls the resulting dopant concentration — ranging across roughly five orders of magnitude from around 10^11 for the lightest threshold-voltage-adjustment implants up to 10^16 for the heaviest source/drain and contact implants. Where implant energy controls how deep dopants land, dose controls how many of them land there, and the two parameters together — not either alone — are what fully specify a given implant step's intended electrical result.
Dose measurement relies on a Faraday cup capturing the beam current during implant and integrating that current over time, then dividing by both the wafer area and the charge carried per ion — a direct, physical measurement rather than an indirect inference, which is exactly why dose accuracy can be held to roughly 1 to 2 percent across production implants. This tight accuracy matters because device electrical parameters — threshold voltage, sheet resistance, junction characteristics — depend sensitively on precise dopant concentration, and a dose error that seems small in percentage terms can still shift a device's electrical behavior enough to matter across a wafer or between lots.
Low-dose implants, typically in the 10^11 to 10^12 range, deliver extremely light doping specifically intended to fine-tune device characteristics like threshold voltage rather than to form a functional junction outright — a deliberately subtle intervention where getting dose accuracy right matters disproportionately, since the entire implant is calibrated to shift an electrical parameter by a precise, intentionally small amount. This is exactly the dose regime where even a modest percentage dose error translates into a meaningfully larger relative shift in the intended electrical effect, since the baseline doping level being adjusted is itself so light to begin with.
High-dose implants, reaching 10^15 to 10^16 for source/drain and contact regions, exist to create heavily doped regions with low resistance and strong ohmic contact behavior, and at this dose regime beam current and throughput become the dominant practical concern rather than the subtle precision concerns that matter more at low dose. Delivering this much dopant per unit area within a reasonable process time requires high beam current, and that higher beam current introduces its own tradeoffs — beam quality degradation and increased wafer heating — that process engineers have to manage specifically because they only become significant at the high-current levels these heavy-dose implants demand.
Dose uniformity across the wafer, typically targeted below 1 percent non-uniformity, depends on how the implant beam is scanned relative to wafer motion, because any systematic variation in how much time the beam effectively spends over different wafer regions directly translates into dose variation across that same wafer. This uniformity requirement exists because device parameters that depend on dose — threshold voltage, sheet resistance, junction depth after thermal diffusion — need to be consistent across every die on a wafer for a design's performance specification to actually hold uniformly across the entire chip population that wafer produces.
Post-anneal sheet resistance, measured by four-point probe, functions as the primary real-world electrical verification that both dose and subsequent dopant activation actually achieved their intended target, converting the Faraday-cup dose measurement taken during implant into a confirmed electrical outcome measured after the complete implant-and-anneal sequence. This two-stage verification — direct physical dose measurement during implant, followed by electrical sheet-resistance confirmation after annealing — exists because dose alone doesn't guarantee a correct electrical result; activation efficiency during anneal also matters, and sheet resistance is exactly the measurement that captures the combined, real-world effect of both steps together.