acceleration voltage

Acceleration voltage is the electrical potential used to give implanted dopant ions their kinetic energy before they strike the wafer, and it is the single parameter with the most direct control over how deep those ions come to rest in the silicon lattice. Doubling the acceleration voltage roughly doubles the ion's kinetic energy for a fixed charge state, and higher kinetic energy translates into deeper average penetration before the ion loses enough energy to nuclear and electronic collisions to stop, so acceleration voltage is the primary lever process engineers turn when a device structure calls for a shallower or deeper doped region. The relationship between voltage and depth is not perfectly linear, however, because the two stopping mechanisms that slow the ion down — nuclear stopping and electronic stopping — have different energy dependences, and getting the depth right at the increasingly shallow implants modern devices require means understanding which mechanism dominates at the energy in use. **The ion's kinetic energy is set directly by acceleration voltage and charge state, which is why voltage is the primary depth-control lever even before any discussion of stopping mechanisms.** For an ion of charge state $q$ accelerated through potential $V$, the kinetic energy delivered is $$ E = qV, $$ so a singly charged ion accelerated through a given voltage receives exactly that voltage's worth of energy in electron volts, while a doubly or triply charged ion receives a proportionally larger energy at the same accelerating voltage — a distinction implanters exploit deliberately, using higher charge states to reach higher effective energies without needing an correspondingly higher physical accelerating voltage. **Nuclear stopping arises from direct collisions between the implanted ion and the nuclei of target atoms, transferring energy in relatively large, discrete steps and dominating at low ion energies where the ion moves slowly enough for a close nuclear encounter to matter.** Electronic stopping, by contrast, arises from the ion's interaction with the target's electron cloud — a continuous drag-like energy loss rather than discrete collisions — and it generally dominates at higher ion energies where the ion moves fast enough that electron-cloud interactions accumulate more total energy loss than the comparatively rare nuclear collisions. Because these two mechanisms trade dominance as a function of energy, and because that crossover point itself depends on the ion species and target material, the projected range as a function of acceleration voltage is not a simple straight line: it has a shape set by which stopping mechanism controls energy loss across the swept voltage range, and Stopping and Range of Ions in Matter (SRIM) or TRIM Monte Carlo simulations are the standard tools for predicting range and straggling for a specific ion-target-energy combination rather than relying on a closed-form formula. **Projected range straggling — the statistical spread in final stopping depth even for ions implanted at identical energy — arises because each ion's individual collision history is random, and this straggling sets a practical floor on how sharp a doping profile edge can be, independent of any equipment precision limitation.** A narrower straggling distribution produces a steeper, more abrupt doping profile at the target depth, which matters directly for how sharply a source/drain or channel doping boundary can be defined; straggling generally scales with the projected range itself and depends on ion mass, target mass, and energy, so lighter ions such as boron typically show proportionally larger straggling relative to their range than heavier species such as arsenic at comparable implant conditions. Because straggling is a statistical property of the stopping process rather than a controllable equipment parameter, achieving an abrupt junction profile at ultra-shallow depths has pushed the industry toward lower acceleration voltages, different ion species (molecular or cluster ion implantation), and post-implant anneal strategies that minimize additional diffusion rather than relying on voltage control alone to sharpen the as-implanted profile. **Ultra-shallow junction formation for advanced logic devices has driven acceleration voltages down to a regime where beam-line implanters face genuine equipment challenges in delivering a well-controlled, high-current beam.** Very low energy ion beams are more susceptible to space-charge effects — mutual Coulomb repulsion between ions in a dense beam — which can defocus the beam and reduce achievable dose rate exactly in the energy range where shallow junctions demand it, creating a persistent tension between the low energy needed for shallow depth and the beam current needed for production throughput. Plasma-based and cluster-ion implant techniques, which deliver dopant atoms bound in a larger molecular or cluster ion that then fragments at the target surface, sidestep some of this tension because the effective per-atom energy is a fraction of the accelerating voltage applied to the whole cluster, allowing higher acceleration voltages (with their more tractable beam transport) to still deliver very shallow per-atom implant depth. | Acceleration voltage regime | Approximate typical depth (dopant-dependent) | Dominant stopping mechanism | Primary application | |---|---|---|---| | Sub-1 keV to few keV | A few nanometers | Nuclear stopping, strong straggling sensitivity | Ultra-shallow source/drain extensions, halo implants | | Few keV to tens of keV | Tens of nanometers | Mixed nuclear/electronic, species-dependent | Source/drain, well and channel implants | | Tens to hundreds of keV | Hundreds of nanometers | Electronic stopping increasingly dominant | Deep well formation, retrograde well profiles | | Hundreds of keV to MeV | Micrometers | Electronic stopping dominant | Deep isolation structures, some power device implants | **Tilt angle and acceleration voltage interact because the effective stopping distance an ion travels along the beam direction is not the same as the vertical depth beneath the wafer surface once the beam is deliberately tilted off the surface normal, which is a standard technique for controlling channeling and for reaching under gate or spacer structures.** At a tilt angle $\theta$ from vertical, the vertical depth for a given along-beam projected range $R_p$ is approximately $R_p \cos\theta$, so increasing tilt angle at fixed acceleration voltage reduces the effective vertical implant depth — a second, geometric lever on depth that process engineers combine with voltage selection, and tilt is also chosen specifically to avoid crystallographic channeling directions where ions can travel anomalously deep along open lattice channels rather than losing energy at the expected rate. ```flowchart Determine target junction depth and profile abruptness from the device design → Select ion species based on required dopant type and diffusion behavior → Choose acceleration voltage using SRIM/TRIM simulation or empirical calibration for the target projected range → Select tilt angle to manage channeling risk and reach the desired structure geometry → Set beam current and dose to hit the target areal dopant concentration → Implant and monitor beam current stability, especially at low-energy, space-charge-sensitive conditions → Measure as-implanted or post-anneal profile by SIMS, SRP, or calibrated electrical methods → Compare measured depth and profile shape against the target specification → Feed voltage, tilt, or species corrections back into the recipe if depth or abruptness is off target → Requalify whenever implanter beam-line configuration, species, or target structure changes materially ``` **Acceleration voltage does not act alone in determining final dopant depth, because subsequent thermal processing — spike, millisecond, or furnace anneal — redistributes the as-implanted profile through diffusion, so the voltage selected during implant and the anneal recipe applied afterward must be co-qualified rather than treated as independent, sequential steps.** A profile implanted at a given voltage to hit a specific as-implanted depth can end up measurably deeper after a diffusion-heavy anneal, which is why shallow-junction process integration increasingly pairs low acceleration voltage with minimal-diffusion anneal strategies (spike or millisecond annealing) rather than relying on voltage selection alone to hit the final target depth after all thermal budget has been spent. Read implant acceleration voltage through a stopping-mechanism lens: the same voltage change produces a different depth response depending on whether nuclear or electronic stopping dominates at that energy for that ion-target pair, so choosing acceleration voltage is inseparable from knowing which physical stopping regime the implant actually sits in, not just reading a depth number off a single calibration curve.