Frequently Asked Questions

What is in-Silicon simulation ?

In-Silicon technology deals with the design and fabrication of devices and structures that are found in the Silicon part of the full architecture. This part of the integrated circuit is also known as 'Front-End technology' and the design is traditionally supported by TCAD (Technology computer Aided Design). TCAD naturally falls in two segments: The first part deals with simulation of all fabrication steps such as implanting, diffusion, deposition and etching. The second part deals with the simulation of the devices. In the second part it is assumed that devices are connected to the outside world through the contacts or device ports that provide bias voltages or currents.

What is on-Silicon simulation ?

On-Silicon Technology deals with the design and fabrication of the metallic wires that connect the sea of transistors in a single integrated circuit. Its design is supported by a hierarchy of tools. At chip level, the focus is on routing and placing functional blocks. At device level one deals with the physical simulation of the interconnect schemes. Physical simulation can still be fine-grained into several levels of sophistication. At low frequencies, it often suffices to extract lumped-element characteristics, such as the resistance and capacitance the building blocks of the wiring. At higher frequencies, more refined simulations are needed to extract the effects of rapidly changing electro-magnetic fields. In-depth understanding of the high-frequency effects can be achieved by using EM (electromagnetic) field solvers.

Why integrate the in-Silicon and on-Silicon simulation ?

High-frequency effects show up in various ways. Besides the well-known skin effect that forces currents to be located only at the surface regions of the interconnect wires, there are effects such as proximity effect that redistributes the currents in neighboring wires due to exchange of electromagnetic energy. One form of substrate coupling deals with the electromagnetic interaction of different parts of the integrated circuit due to electromagnetic energy that is injected into the Silicon part of the chip. Whereas at low and moderately-low frequencies, the devices in the Silicon and the wires on the Silicon could be characterized separately, we have now entered the era in which an integrated approach is needed. At higher frequencies, the electric fields are for a large part inductive, i.e. a major contribution originates from the rapidly changing magnetic fields. The details of the substrate will then determine the interconnects characteristics, and vice versa the in-Silicon device characteristics will be strongly influenced by the on-Silicon structures (interconnects) that are around. The border that separates the Silicon from the layers above the Silicon dissolves when a characterization must be achieved at high-frequencies: An integrated characterization by simulation of the front-end and back-end structures is needed at high frequencies.

How does MAGWEL address the integrated in-Silicon and on-Silicon simulation ?

The standard decoupled TCAD approach is indicated in the following schema.

Currents and charges are solved independently, and then used as source for the electromagnetic fields. That these electromagnetic fields on their turn alter the sources is not taken into account, and therefore an inconsistency is incorporated in the solution. MAGWEL offers an approach to accurately simulate in-Silicon devices and on-Silicon structures by simultaneously solving the full set of Maxwell equations that describe the electromagnetic fields the drift-diffusion equations that describe the semiconductor physics Ohm's law that describes the currents in metallic domains.
The electromagnetic fields are determined by the current and charge distributions and vice versa, the currents and charge distributions are determined by the electromagnetic field. A reliable solution is obtained if all equations are satisfied with a single set of variable values. This is known as the self-consistent solution. It is common practice in TCAD that only the self-consistent solution is compliant with experiment.

Which physical effects does the MAGWEL software consider ?

The Magwel software adequately predicts quantitatively:

Skin effect is the tendency of a high-frequency electric current to distribute itself so that the current density near the surface of a (semi)-conductor is greater than that at its core. This effect in conductors and semi-conductors is predicted in a fully consistent way. This means that the skin effect is not added in an ad-hoc fashion, but comes out of the self-consistent simulation.

Current crowding is the tendency for high-frequency electric current in adjacent conductors to repel each other if the are running in parallel and to attract each other if they are running anti-parallel. This effect in conductors and semi-conductors is predicted in a fully consistent way.

Eddy Currents (also called Substrate currents or induced currents) are the currents that run in a conductive area (e.g. the substrate) due to the inductive (magnetic) effects of currents nearby. This effect in conductors and semi-conductors is predicted in a fully consistent way.

Semi-conductor junctions are due to different doping profiles in a semiconductor. These junctions make it possible to build active devices like transistors, and have to be described nonlinearly.

Which frequency range can be addressed by the MAGWEL tools ?

The MAGWEL tools are able to carry out full self-consistent simulation ranging from DC up to 100 GHz.

Can I use my standard lay-out files as input files for the MAGWEL tools?

MAGWEL tools are able to read your GDSII files. Of course also the technology information needs to be provided in order to reconstruct the detailed 3D geometry of the input structure. The MAGWEL tools do not deal with process information. They start with a design geometry to calculate the detailed electromagnetic behavior of the problem. Internally the input file of the MAGWEL tool is the open-source CODESTAR xml format. This means that the whole solving procedure is scriptable and standard optimizer tools can be used to enhance and optimize your results.

Can the MAGWEL tools produce for you an optimized design ?

Yes and no. There is no optimizer module available yet. However, internally the input file of the MAGWEL tool is the CODESTAR xml format. This means that the whole solving procedure is script driven and standard optimizer tools can be used to enhance and optimize your results.

Also, Magwel has recently acquired Kimotion Technologies which brings an opportunity to integrate an optimizer into the Magwel simulator.

Does the MAGWEL tool provide S-parameters ?

Yes indeed. Standard Touchstone files are the output of the extractor ExtractEM, so you can plug the result directly into your favorite SPICE solver.

Can I get a test version of the software ?

Contact us with your request and we will contact you.