In front of a huge cream-colored machine are four researchers, covered from head to toe in white overalls. Only their eyes can be seen and no sounds are heard, stopped by the glass that protects the Clean Room. Technicians have to be careful when they talk or cough, but also when they breathe: the tiny particles of water released into the air, as well as the dust, are about the size of the objects produced in this laboratory which until a few years ago was a garage and today it is one of the few places in the world where sensors and microchips are built with components as large as one micron: one millionth of a meter, one hundredth of a hair.
Before arriving in front of the thick door of the clean room, walking along the gray corridors of the Bruno Kessler Foundation, it seems to be in a small branch of a company in Silicon Valley. In reality, the foundation, which is a research institution, is located in Povo, a hamlet of Trento three kilometers from the center. It is a quiet place and only students from the university or foundation researchers are seen on the streets. It is not easy to distinguish them, except for the badges hanging around the neck.
From the outside one gets the impression that the large structure with rounded corners has been enlarged piece by piece, like a Lego; from the inside, the sensation is confirmed in the network of paths organized in an indefinite number of floors, between suspended walkways that connect three different buildings simply called North, East and West.
It is a morning in mid-May and some developers work in silence, concentrated in front of their monitors, surrounded by the glass walls that divide the spacious offices. They deal with artificial intelligence, teaching machines to perform complex tasks such as analyzing the health data of thousands of people and predicting possible diseases. On the stairs of the West building, in front of the vending machines, there is less silence: in the chatter of a coffee break the comparison is stimulating, informal, and involves researchers from different departments.
The collaboration between over 400 researchers and researchers allows the Bruno Kessler foundation to develop cutting-edge projects in many fields, including machine learning and deep learning, cybersecurity, digital industry, materials science and energy sustainable. There is not much talk about it in the media: in the last year we have read about the foundation especially for the analysis of epidemiological data carried out for the Higher Institute of Health. But these laboratories are unique for other reasons: one of them is the Clean Room.
Pierluigi Bellutti has a pass that opens all doors, even those with a red stamp and the words “reserved entrance”. He walks fast and has so much to say that it's hard to keep up with him. He has the natural gift of telling extremely complex things with a few simple images, even if he often has to stop and think for a few seconds to find the right words.
Bellutti is the head of the MNF unit, “micro and nano facility”, of the device and sensor center. He is a chemist and as a chemist he knows the precious characteristics of silicon well, with which all the sensors produced inside the clean room are built. “If silicon disappeared from the periodic table of elements, we would be directly in the 1940s,” he says. “The electronic devices we use every day have silicon components, and at the moment no alternatives are available”.
Bellutti proudly displays a sensor inside a transparent box. It is one of the radiation detectors made for HERMES, an experiment of the National Institute of Astrophysics, which in 2022 plans to send into orbit a constellation of six nano satellites with sensors capable of detecting and localizing gamma and X rays emitted by known sources like Gamma Ray Bursts, the gamma-ray bursts. The goal is to observe more precisely what happens to cosmic sources in the periphery of the universe: it is a much anticipated experiment.
The growth of the projects developed in the clean room was only possible thanks to decisions that were not taken for granted, such as that of avoiding competition on the aggressive market of microprocessors and memories, and addressing a promising niche: the production of devices necessary for the execution of the experiments. scientific. In recent years, sensors have been built for large international experiments such as ATLAS and ALICE, at CERN in Geneva, and for the international space station. It means that the foundation is not involved in the serious shortage of microprocessors which in recent months has created many difficulties for the worldwide production of electronic devices and at the moment it wants to stay away from this problem.
– Read also: The microchip crisis is getting worse
If HERMES is the present, the past is called MAIA, the Advanced Model of Artificial Intelligence created over thirty years ago, in 1988.
To demonstrate that it was able to develop an integrated artificial intelligence system, the researchers designed a concierge robot: MAIA was able to understand, interact and accompany a person in all offices thanks to the ability to recognize the environment in which he was moving. One of the first sensors made was a camera, MAIA's eye. It was anything but a game: working on that robot we started talking about artificial intelligence and the need to create a laboratory to build microelectronic chips.
The clean room arrived a few years later. Finding a space large enough to accommodate the bulky machinery was not easy. In the end, the choice fell on the garage, as if to confirm one of the well-known clichés about the origins of many successful technology companies.
The clean room is called that because it is really clean. The ceiling, the floor and all horizontal surfaces such as the tables are perforated: the air falls down, is sucked in, recirculated and continuously filtered to be cleaned. This allows for a laminar flow that breaks down particles of sweat or saliva, rich in sodium and potassium and harmful to silicon. Throughout the clean room there are between 10 and 100 dust particles of less than one micron diameter every 0.02 cubic meters: in a city street there are usually 35 thousand and in an operating room about 10 thousand.
Furthermore, water called “ultrapure” is used, free of any foreign element, including bacteria eliminated by special filters. Many of the machines in the clean room are automatic, they allow to engage a limited number of technicians, about twenty, as well as as many researchers and researchers, and allow to limit human errors as much as possible.
The basis of all the processes is always the same, the silicon, which appears in the form of a disk, called a “slice”, slightly larger than an old CD. It has a diameter of fifteen centimeters and a thickness of half a millimeter. In the clean room you can see many cars: it is not easy to understand what they are for and even the captions hanging up for the students on a trip must be read carefully. The discs move from one side of the chamber to the other, moved by the technicians very carefully. They are very fragile, easy to break, and hundreds of them can be thrown away in the development of a new sensor. The technicians do not seem to be in a hurry even if in recent years production has increased and new people and new spaces would be needed.
The first steps are identical for all sensors produced inside the clean room, then the process varies according to the final result: there are sophisticated sensors, such as those produced for the HERMES project, and others simpler such as pressure sensors or optical sensors. “Our strength is the great quality of the work, whatever the product we produce”, says Bellutti. “They called us a boutique, the high craftsmanship of microelectronics”.
Before the production phase, work begins in the research labs where a chip is designed. Once made, it is apparently flat. It is actually made up of various overlapping layers connected to each other to form a precise architecture. This architecture is defined by several passages in the lithographic area, the heart of the clean room. Here, a photosensitive polymer is applied to the slices which uniformly covers the entire surface of the slice: it is photosensitive because it reacts to light. The silicon then passes into the printing machine where it is illuminated with an ultraviolet light that is passed through a transparent plate, on which there is the design of the project obtained with a metallic film: as in a lithographic print, the light passes where there is no chromium and modifies the structure of the photosensitive polymer.
The final result is a silicon wafer in which the design designed by the researchers has been faithfully reproduced. This process is repeated several times, layer by layer, depending on the complexities of the device. This is how the architecture is composed.
One of the most interesting phases is called doping, doping, and consists in altering the conductive properties of silicon by introducing different atoms that break its symmetry, transforming it from insulator to conductive. To do this, in the clean room of the Bruno Kessler foundation there is an ion implant, in which ions of boron and phosphorus are accelerated in an electric field and end up bombarding the surface of the silicon wafer.
This process breaks the ordered symmetry of the silicon which must be recovered in order to activate the inserted atoms. You can use different machines: there are ovens that bring the slices up to 1200 degrees in a process that lasts several hours, or you use a machine called Rapid Thermal Processing that reaches 1200 degrees in a few seconds and guarantees greater precision of the final result. .
The most complex processes also involve more than 200 steps and over three months of work. Among these, there are many control steps under the microscope, sometimes even the electronic one, to ensure that the details not visible to the naked eye have been made in the correct form.
One of the most interesting insights of recent years has been the collaboration between the foundation and the national institute of nuclear physics which has many offices and laboratories throughout Italy. Its researchers are constantly in need of new sensors for experiments and until a few years ago they were forced to turn to the foreign market, especially Switzerland, France, Germany and Japan. When the foundation realized that it would be more useful to concentrate efforts on the development of sensors for research, a relationship of mutual support was born.
The collaboration has lasted for almost twenty years and has made it possible to study and design new types of sensors, led to the purchase of some machinery with a co-investment, and made it possible to keep funds for research in Italy. “Before, Italian resources and researchers were forced to go to Switzerland, Germany and France,” explains Bellutti. “We have intercepted this brain and capital drain”.
On the wall hangs a billboard with the inscription “DRIE: deep reactive ion etching”: it explains the procedure designed to deeply engrave the already very thin slices of silicon. It is the technological development that allowed the foundation to win a tender for the supply of latest generation sensors for the ATLAS experiment to monitor collisions between particles.
The poster is there to remember the efforts made to create that particular sensor: the first attempts date back to 2003 and it took ten years of work to obtain a satisfactory result. But these ten years of experience also served to develop and adapt these technologies to other areas, such as the digital and automotive industries. This is the answer to the question that many are asking about whether to invest money in large physics projects. “By developing sensors for CERN or for the international space station or for scientific satellites, highly qualified skills are created that we also use in other sectors,” says Bellutti. “Que sto also allows us to find the resources necessary to continue to be competitive “.
There is an interesting aspect regarding space projects: everything that must be achieved in this field must have a very high reliability, given the current impossibility of being able to repair or replace parts of satellites.
This leads to an immediate availability of use on the “terrestrial” market of solutions based on these developments. This is the case of new sensors and chips for satellite telecommunications, microsensors for assisted driving of cars and for controlling braking systems. But it is also the case of the bio-medical sector for which, with silicon, miniaturized systems for monitoring physiological parameters are developed through “wearable” sensors, microsystems for carrying out clinical analyzes and also for use in the important agri-food sector with the possibility to carry out food quality analysis.
For those who produce microprocessors, the challenge is to reduce the size of the product more and more. Go from micro to nano. For the researchers of the foundation, however, it is not the priority, but the beginning of a new experimentation and the first attempts are already underway. In a new part of the clean room there is another large room where a new machine is at work. It manages to produce chips that are rather simple in the composition of the materials, but with dimensions of less than 50 nanometers. As already mentioned, a micron is one millionth of a meter, while a nanometer equals one billionth of a meter. It is a first work in the field of “quantum technology”, quantum technology.
The limit was touched by the IBM company, which in early May announced the production of a two-nanometer chip, more advanced than the “old” seven-nanometer chip and with a significant reduction in energy consumption. In Europe there are no companies that can afford the development of such small chips because considerable resources are needed, also to cover management costs, and a very advanced technological capacity. Investments are inversely proportional to size: the smaller the chips, the more money is needed to design and manufacture them.
In Europe, however, something has begun to move: with the financing of projects called IPCEI, Important Projects of Common European Interest, we will try to support and keep the development of micro and nano electronics on the continent.
