Remembering Stepan Mashnik

It is seven years, today, since Stepan Mashnik passed away. A nuclear physicist, Stepan was born in the Soviet Union during the last year of Stalin's regime. After completing his PhD, Stepan joined the Joint Institute for Nuclear Research (JINR) in Dubna and eventually started developing the Cascade-Exciton Model (CEM) of nuclear reactions.

The first original work was published in Russian, in 1980, on the JINR Communications journal. This work, coauthored with Konstantin Gudima and Viacheslav Toneev, was then published in English, on Nuclear Physics A, in 1983.

I will not go into the physics details, but let me just say that the predictive power of CEM was surprising. And so were the applications, from transmutation of nuclear waste to the design of spallation sources (CEM was in fact used in the design of the European Spallation Source ERIC), from radiation-therapy medical applications to the study of cosmic-radiation effects on astronauts and spacecrafts.

As Stepan loved to say, CEM has been his ticket to freedom. And that freedom passed via (my hometown) Trieste, where he had been invited to present his work, for the first time outside the Soviet Union, at the Abdus Salam International Centre for Theoretical Physics (ICTP).

Monte Carlo methods for radiation transport

Monte Carlo methods are computational algorithms allowing to obtain reliable numerical results from repeated random sampling. It is somehow the same principle of conducting a poll: you ask a question to a limited number of people, uniformly but randomly distributed in the general population, and you then assume that the answer offers a true representation of what the whole general population thinks.

Just that in the case of Monte Carlo, you don't ask people, but you simulate the behaviour of a limited number of particles (protons, neutrons, electrons, photons, etc...) at nuclear level. And then you assume that the way these particles behave is the actual behaviour of a much larger number of events "in real life".

Now, to make a very long story short enough to have you keep reading, to be successful in such simulation, you must know very well what a particle would do, how it would react. There are two ways to know that:

• make an experiment to learn how particles behave;
• have a model that describes how particles behave.

Experiments are challenging, expensive, often impossible. Also because of the very large number of variables that one would need to measure.

So we use Cross Sections

"Cross Sections" are the rules of the game. Cross Sections tell the computer code how each particle will behave. If you have good Cross Sections, then your calculated result will offer a solid representation of reality. Poor Cross Sections will lead to unreliable results.

What CEM does, in extreme synthesis, is to provide a way to calculate Cross Sections (and other parameters) for a set of reactions (particle behaviours) that are extremely important and at the same time extremely difficult to obtain experimentally.

Yes, there are many other models doing the same... but CEM had the potential to do better than any other model. The work of Stepan gained the attention of the scientific community, and finally in 1997, he moved with his family to the United States, to start working at the Los Alamos National Laboratory (yes, the same place of the Oppenheimer movie)

When I first met Stepan

I was up late at night writing my PhD thesis, sitting at my desk at Uppsala University, when I received an email from a person at the X-Division in Los Alamos. The sender was a name I read only in publications and conference proceedings. Yes, Stepan Mashnik. He read one of my papers (wow effect on my side!) and he was asking if I could share those data even before publication in my thesis. I replied immediately and, few moments later, Stepan called me. It was the beginning of our friendship.

In exchange for my data, that Stepan used as part of the benchmarking process of CEM in MCNP6, he agreed to provide me calculations to be included in my thesis. Now, maybe this sounds like nothing to you. But it was big for me.

MCNP6 was the most exciting -- NOT YET RELEASED -- version of the code. A code developed as part of the Manhattan Project (Oppenheimer again!), which allows to simulate almost everything: from radiaiton affecting the Mission to Mars, to the impact of a thermonuclear weapon against a large asteroid on its trajectory to destroy our planet (yes, I saw such simulations). It was like getting to play on a PlayStation 6 one year before it would be on the market (ok, it is still a nerd thing, but I am a nuclear physicist, what do you expect?)

Remembering a friend

We did not talk only about nuclear physics, but when we did, I felt the privilege of learning from him. As time passed, we were also sharing about our personal lives, the birth of my daughter, the wedding of his daugther, our experiences travelling in South America. Stepan was open about his long fight against cancer. His last email to me was just a month before passing away. He wrote "If my conditions improve, I will let you know." Unfortunately, I did not hear back from Stepan.

Today, CEM is the default model in the MCNP6 Monte Carlo code. Next time you will run a Monte Carlo calculation or will use some results from these calculations, I hope you will also remember the finest mind of a scientist who has been dedicated to his work until the very last days on this planet.