The fastest pure silicon bipolar transistor from FER's laboratory

A Croatian feat worthy of Silicon Valley can lower the cost of computer production!

Scientists from FER have developed the world's fastest HCBT transistor - the fastest pure silicon bipolar transistor. It is a bipolar transistor with horizontal current flow, as used today in more sophisticated analogue circuits, for optical and wireless communications, power supplies, operational amplifiers, hard disc and display drivers, explains the head of MiNEL, Prof. Tomislav Suligoj, PhD.

The fastest network, the fastest data transmission, the fastest chip, the fastest... News about breakthroughs that telecommunication companies and various IT companies are using to get customers to "give them money" as quickly as possible follow each other at lightning speed. The scientists and researchers at the Micro- & Nano-Electronics Laboratory (MiNEL) of the University of Zagreb Faculty of Electrical Engineering and Computing (FER), on the other hand, have only modestly celebrated their success, if at all: they have developed the fastest transistor in the world. Although they warn us to be precise to avoid even the slightest sensationalism, they say: we have the fastest pure silicon bipolar transistor. Globally. It is the Horizontal Current Bipolar Transistor (HCBT) as used today in more sophisticated analogue circuits, for optical and wireless communications, power supplies, operational amplifiers, hard disc and display drivers, explains the head of MiNEL, Prof. Tomislav Suligoj, PhD.

The team that worked on the project - Marko Koričić, Josip Žilak, Željko Osrečki and Filip Bogdanović - pointed to the possibility that the production costs for these assemblies, or chips, will be much cheaper. In the conditions of scarcity and rising prices of semiconductors and electronic components that form the basis of the modern world of technology, this sounds spectacular. Suligoj calmly explains to us what it is all about:

"HCBT is currently realised in pure silicon, which means a simpler production process, i.e., a cheaper technology. It has a cutoff frequency of 51 GHz (gigahertz), a maximum frequency of oscillations of 61 GHz and a breakdown voltage (collector-emitter) of 3.4 V (volts)," he says. The transistor itself, by the way, is quite tiny - the size of 1 micrometre (μm), with an active part of 50 nanometres (nm).

Such transistors consist of collectors, bases, and emitters - the collector and the base are in the silicon substrate, and in HCBT the base is on the sidewall of the silicon hill. The idea of the MiNEL experts was to arrange an optimal doping profile on this side wall in the intrinsic region of the transistor, which has not been seen before. As a result, the current does not flow vertically through the transistor, but horizontally.

"The concept of horizontal current flow means that individual transistor regions can be manipulated more easily. Circuits with different voltages on the same chip increase the flexibility of the technology and allow additional optimisation of circuits or the realisation of circuits with different functions in the same technology. This could mean that an electronic system that is realised today with different chips could be produced in one, thus reducing the cost of the entire system and improving its performance," the professor describes. He notes that all assemblies can be realised in HCBT as in other bipolar silicon technologies, but with lower production costs, greater possibilities of integration with other functions, potentially better efficiency, or lower consumption.

When we talk about transistors and semiconductors, their main area of application is the IT industry, i.e., the production of circuits for computers and all electronically operated devices.

At the moment, we are mainly focused on developing high-frequency circuits for use in communication devices such as mobile phones, base stations, remote controls and navigation devices that can use the speed of HCBT, but their market potential is very large," say MiNEL experts. However, they acknowledge that it is difficult to predict where it could best be used, given all the technical and economic aspects. It could also be added - and political, as some member states, including the EU itself, have recognised the need to develop semiconductor technologies on their soil, thus starting local production. MiNEL has protected innovation with more than 20 patents. Do they think they have the power to commercialise it in Croatia?

Crucial know-how

"For the production of HCBT, the developed technological process has to be implemented in a very complex and expensive production line for semiconductor components, mostly based on the CMOS process. Such a production line or factory does not exist in Croatia, the only possibility would be to build it. As tempting as this may sound, it is very difficult to realise this in the short term, not only because of the very high costs but also because of the technological knowledge, i.e., the existence of the crucial know-how and the entire environment for such a sophisticated production. However, in addition to technology, there are several accompanying activities in the chip development chain that are equally important for the final product (designing new elements and assemblies, developing their computer models, analysing the results, packing chips in housings, demanding measurements and tests...).

We hope that at this point Croatia will be involved in the improvement and development of European semiconductor technology in one of these activities, for which you do not have to have the technology and the processing itself can be done elsewhere," said Suligoj.

Therefore, he considers it more likely that foreign companies will "go crazy" for Zagreb's HCBT, and these heavyweights are worth tens of billions of dollars.
 

"It is more realistic to implement and possibly licence our technology to an existing semiconductor component factory. This procedure is very difficult and demanding because it relates to the most sensitive part of any manufacturer, namely the technology of the transistors themselves. The development of such processes is planned years in advance, and it is difficult for companies to choose a technology based on a significantly different concept that inevitably brings additional risks along with the benefits. There are intensive contact with various companies, but besides the presentation of innovative technology, many other factors are required, ranging from reliability and control of the technological process on a mass-production scale to intellectual property issues, market research and other activities that are not part of our primary competencies," concludes the head of FER's Micro- & Nano-Electronics Laboratory Tomislav Suligoj.

Cooperation with CERN

"In collaboration with CERN, we are working on very sensitive and precise particle detectors, optimising the structure of semiconductor detectors and designing circuits to read signals from these detectors, which need to be very sensitive, fast and precise, but also resistant to radiation from particles colliding in CERN detectors at speeds close to the speed of light. We are also working on silicon detectors with a layer of amorphous boron, which have shown record sensitivity to ultraviolet radiation and low-energy electrons and are attractive for use in various industrial and physical instruments," they say.

They also deal with the modelling and simulation of newer semiconductor materials and elements such as transistors with high electron mobility based on gallium nitride or MOS transistors based on gallium arsenide.

Part of the group also works on modelling the properties of two-dimensional materials - graphene, phosphorene and the like. In addition to the members mentioned in the text and picture, MiNEL also includes Dubravko Babić, Mirko Poljak, Tihomir Knežević, Borna Požar, Mislav Matić, Lovro Marković, Ivan Berdalović, Filip Bogdanović and Jyotiska Bharadwaj.