The past three years in Oxford have been very busy for me. I received an award from the Department of Chemistry for managing a research project independently when my superior was on leave. I won Oxford University’s poster prize and was selected to present my research in the British Parliament. I co-authored two publications and two more are on the horizon. I am building a huge apparatus that is 95% based on my ideas. So, I have completely forgotten about the skeleton in my cupboard: three unproductive years at the University of Freiburg.

The skeleton has resurfaced twice recently, both times with an annoying outcome, as if a reminder to let itself be known. My proposals for establishing an independent research group were rejected, and my publication gap was to be blamed every time. It is quite a paradox that those three years equipped me with the very skills that allowed me to get a job at a prestigious university. So should I regret those years or not?

On a personal level, I should not. My two children were born in the greenest part of Germany, close to Switzerland. Family trips to the Alps will be carved into my memory for ever. We lived in a subsidised flat, so – except for the period directly after moving to Germany – we were financially stable, even though I was employed on a part time contract.

However, from a professional point of view, gains from that period are very doubtful. As many postdocs know, having a part time contract does not mean that you work part time. Despite trying hard, I got no publishable experimental data but acquired a broad range of technical skills. Whenever the failures were too much to bear, I escaped to teaching duties that enhanced my knowledge of physics.

Nesting

The whole story begins with my graduation. In October 2013, I became jobless and was doomed to throw away all my technical skills because I could see no employer in the Czech Republic that would benefit from them. Eventually, I started my independent business as a Japanese/Czech interpreter and got a permanent customer soon after. However, translating corporate documents was not giving me the life satisfaction I desired, so I jumped on the first academic job that was offered to me from the University of Freiburg. That happened to be in a physical discipline I had nothing in common with, but it involved the usage of ion traps and plasma sources, which was exactly my specialisation.

When I arrived at the Freiburg lab in January 2014, I could see a jungle of cables and smelled burned dust from a huge diffusion pump. It was very silent – no experiments were running, as the only worker, a very demotivated postdoc, was going to leave the group soon. I got one month’s training on how to run the experiment and was left alone. I started doing experiments following the guidelines, but realised soon that I had no idea what was wrong every time the experimental apparatus, a 6 m long beast, did not want to listen to me.

I started making sense of the apparatus, tracking cable by cable, and eventually found out that it was indeed a complete mess. I made a huge decision to disconnect everything, mapping each connection, so that I could reconnect the devices in an optimal spatial configuration. Thanks to this, I learned quite a lot about the operation of double-reflectron time-of-flight mass spectrometers. After 6 months, everything was clean and tidy and I was much wiser.

Taking back control

People say that there are no negative results in science if all parameters are under control. Well, they were not in my case. I was hired to study phase transitions of the world’s smallest water droplets, so I needed to produce them reliably, with a reproducible size distribution. The production process had been completely incidental before I started digging into it. Based on an online literature survey, I constructed a bubbling water cluster source that would put a controlled amount of helium gas through a water vessel and generate a helium-water mixture. That served as a precursor for the production of droplets in a discharge. The flow was controlled by a mass flow controller I found on eBay, for which I had to have a driving box made. Unfortunately, this set-up was producing clusters too huge to be slowed down by existing ion optics. So I decided to break them down by heat, and winded a heating wire around the tubes. Later it turned out that I needed to set the pressure in the water vessel as well, and so installed a manual needle valve at the outlet from the vessel. This journey took me several months!

The principle of the phase transition measurements was one of the most elegant I had ever seen. I produced charged water droplets, cooled them down to a desired temperature in an ion trap, shot them into a flight tube, hit them with a strong laser light and observed a mass distribution of ice fragments. Recording a mass distribution for a range of initial ice temperatures, I was supposed to find out at which temperature a phase transition occurs.

The scientific charm of these experiments was hidden in the fact that such a sharp temperature value does not exist. It rather seemed that the phase transition occurs in a broad temperature region. Therefore, my boss came up with the idea that the ice particles are amorphous (have no periodic crystal structure). If it were so, it would mean that the structure the ice particles gain strongly depends on how quickly they freeze. Unfortunately, the freezing speed was totally out of control, so I designed a pre-cooling stage that would slow down the cooling process.

I managed to get only few sets of noisy data with my pre-cooler when an old computer operating an old-fashioned digitiser for a CAMAC crate got broken. Unfortunately, the software interface was written in a proprietary computer language, for which I had no interpreter compatible with new operating systems. I spent some time writing a new interface in LabVIEW. During the process, I even discovered a firmware bug in the digitizer!

Then a Deus ex machina came…

Really. I was told by my boss to abandon these fruitless experiments. Instead of that, I was now supposed to study how the negatively charged ice particles relax to their original structure once they are irradiated by a strong laser light. The final structure was to be judged from an energy spectrum of electrons detached from the ice by another (UV) laser. The problem here was that I had only a velocity-map imaging device available for the acquisition of such spectra, which has a much lower resolution than specialised time-of-flight spectrometers. Moreover, I was always getting a strange background noise in the images that reliably spoiled my data. I tried subtracting a periodically measured noise and recorded quite a lot of data. One data acquisition session even involved a cooperation with firefighters because one of the high voltage power sources started smoking so much that a smoke detector called two fire engines.

After several weeks, it was clear that the data were not good enough to be used for any conclusions. So my boss suggested tripling the frequency of our mighty 1J/pulse YAG laser to get a more intense UV light to increase the signal/noise ratio. However, that meant that I had to construct a frequency tripling stage by myself, learning how from a book. After doing so, I tried the experiment again and the background noise was still there! So I decided to clean the spectrometer. When I opened the velocity imaging spectrometer, I saw that one electrical connection was loose inside. That invalidated all my previous data, because I could not reliably state whether the loose connection affected them or not. Back to the beginning!

Battle with aliens

In winter 2016, the YAG laser got broken because algae growing in a cooling circuit over many years caused a short circuit in the laser that broke the electronics. This was clearly my fault because I trusted the former postdoc who told me that the red light on the laser alerting to that danger had been always on and it never affected the laser performance.

In order to identify a faulty electronic part, I had to use an oscilloscope probe as big as my forearm to check the generation of high voltage pulses. Oh gosh, I sweated a lot, for fear of getting an electric shock of several thousand volts!

Then I had to massacre the uninvited visitors. Standard procedure for doing so is to put a weak solution of peroxide into the cooling circuit where they live. After one attempt, a white powder of dead bodies got deposited on a mesh filter. But the red light indicating an over-threshold water conductivity did not turn off! So I repeated the procedure two, three, four, five times. I even put a UV light source into the water tank, but it started smelling funny so I stopped. The **stards survived! Peroxide again. After a month of committing a crime against living beings, the little creatures disappeared.

The end

In December 2016, I had my final interview in Oxford. I got a postdoc position and was saved. At least for three years. So what have I learnt from the three years in Freiburg? Maybe, that you need to be able to leave the job as soon as you realise it is too hard. In other words: know your limits. As someone said, it is extremely easy to do an extremely difficult research.