Access Official Technical PDF

Share Summary:LinkedInEmail

Silicon detector simulation can help engineers and research teams examine how detector geometry, bias conditions and particle trajectories affect electrical signals. A tool developed for silicon-pad detectors takes this approach by calculating electric potential, electric fields and transient induced currents in three dimensions. Its focus is a relatively simple detector structure, but its computational methods illustrate how numerical modeling can support detector design and analysis.

Fast Silicon Device Simulation is an open-source application built on CERN’s ROOT framework. It uses iterative and multigrid methods to calculate fields, then models charge-carrier movement and the resulting current. The reported current signals were compared with simulations from Silvaco TCAD and Garfield++, providing a basis for assessing the tool’s results under the conditions studied.

Why model a silicon-pad detector?

A silicon detector is based on a p-n junction operated with reverse bias. In the detector described here, a p-type region is formed on an n-type silicon wafer. The electric field across the junction creates a depletion region: a volume with few mobile charge carriers. Applying reverse bias widens this region, increasing the volume in which passing charged particles can generate a detectable signal.

When a charged particle travels through the silicon, it creates electron-hole pairs. The electric field drives electrons and holes toward their respective electrodes, inducing a current that can be measured. The signal therefore depends on the field within the device and on how the charge carriers move through it.

The modeled pad has a 2 cm by 2 cm area and a depth of 300 µm. The simulation accepts device dimensions, reverse-bias voltage, operating temperature, particle incidence angle and the specified generation rate of electron-hole pairs. This lets users examine how changes to geometry or particle path affect the calculated signal without treating the detector as a single, uniform electrical element.

How the silicon detector simulation calculates fields

Solving for electric potential

The starting point is Poisson’s equation, which relates electric potential to charge density and material permittivity. The detector volume is represented as a three-dimensional lattice. At each internal lattice point, the potential is updated from the values at neighboring points, subject to the boundary potentials set for the electrodes.

A straightforward iterative calculation can take many updates to propagate boundary conditions into the interior. The application reduces this workload with a multigrid approach. It first solves the potential on a coarser lattice, interpolates those values to create a finer lattice, and iterates again. The process continues until the lattice reaches a 1 µm spacing. In the reported calculation, the iteration at each level continued until potential values stabilized to the stated tolerance.

This sequence gives the fine-grid calculation a more informed starting point than initializing the entire detailed lattice from zero. The practical goal is faster convergence while still solving the same potential problem; the multigrid method does not change the underlying physical model.

Deriving the electric field

Once the potential is calculated, the electric field is obtained from its spatial gradient. The simulation evaluates the field components across the lattice, allowing the field’s direction and magnitude to be examined within the detector. For the studied reverse-biased device, the field along the depth direction varied as expected for a bulk region with constant doping density. The field profile was also compared with a Silvaco TCAD result.

From charge generation to induced current

The application models electron and hole transport using drift-diffusion equations. These account for carrier motion under the electric field as well as diffusion associated with carrier-density gradients. Carrier mobility varies with electric-field strength according to the Caughey-Thomas expression used in the study.

Charge continuity is considered alongside the field and transport calculations. In the reported implementation, recombination was ignored because its effect was considered very small for the modeled case, while electron-hole pair generation was specified at the start of the simulation. A minimum-ionizing particle was represented as generating charge carriers uniformly along its path.

The output includes transient electron, hole and total current. The two carrier types contribute differently over time because their mobilities differ. In the example, electrons were collected more quickly; after their collection, the hole current accounted for most of the remaining signal. That separation helps users understand the shape of a current waveform rather than only its peak magnitude.

Comparing results and exploring design variables

Current signals from Fast Silicon Device Simulation were compared with Silvaco TCAD and Garfield++. Under the study’s shared mobility model and simulation conditions, the total current was reported as similar across the three approaches. TCAD calculates device behavior using semiconductor models, while Garfield++ uses the Shockley-Ramo theorem to calculate induced signals; in this comparison, its electric and weighting fields were obtained from Silvaco TCAD.

The tool was also used to vary particle incidence angle and detector depth. As incidence angle increased, the particle traveled a longer distance through the silicon, so the generated charge and peak current followed the expected path-length dependence. The signal shape remained similar in the cases examined. Changing detector depth together with the reverse-bias voltage changed both signal magnitude and shape, with a more visible hole-current tail as depth increased.

These comparisons should be read in context: they support the implementation for the modeled detector and assumptions, not every silicon-device geometry or operating condition. Model inputs, boundary conditions and mobility assumptions remain important when interpreting results.

Practical implications

For teams evaluating silicon detector designs, the main value of this approach is the ability to connect geometry and operating conditions to fields and time-dependent signals in one workflow. A multigrid solver can make detailed potential calculations more efficient, while current comparisons provide a check against established simulation tools. The application was developed for research into silicon-detector properties, including work related to the LAMPS experiment; it is not presented as a general-purpose industrial device simulator.

In practice, simulation is most useful when its assumptions are explicit and its outputs are validated against suitable reference calculations. Engineers and technical decision-makers can use this kind of modeling to investigate how bias, detector depth and particle trajectory influence signal behavior, then use those findings to guide further design analysis.

Access Official Technical PDF

About The Author