source: ETALON/papers/2016_HDR_ND/Compton/mightylaser.tex @ 753

\input{./header}

\chapter{Compton scattering as a source of X-rays: MightyLaser}
\chaptermark{Compton scattering for X-rays: MightyLaser}

\section{The MightyLaser experiment}

The aim of the MightyLaser experiment at the KEK ATF is to demonstrate that it is possible to produce high flux of polarized $\gamma$ rays. Such flux could be used, for example, to produce polarized positrons at a high rate for the ILC. The experiment uses a Fabry-Perot cavity to enhance the laser power. It builds on previous efforts done at LAL~\cite{Baudrand:2010hp} to use Fabry-Perot cavities for Compton scattering. The MightyLaser cavity has the specificity of using a pulsed laser to increase the peak power that can be achieved. For better stability it uses 4 mirrors in a ``bow tie'' configuration (as shown on figure~\ref{fig:ml_cavity}). The cavity and the project have been described in details in~\cite{Bonis:2011hi}. I joined the project at the end of year~2010 when most of the hardware developments had already been done. They are documented in~\cite{Bonis:2011hi}. My main contribution to this project has been to lead the experimental campaign at the KEK ATF as described below.

\begin{figure}[htbp]
  \centering
  \includegraphics[width=0.6\linewidth]{Compton/kek_geo.jpg}
  \caption{Schematic drawing of the four-mirror cavity installed at ATF as published in~\cite{Bonis:2011hi}. The laser beam is represented by the red cones and the incident laser beam by a red arrow.}
  \label{fig:ml_cavity}
\end{figure}



\section{Experiment at the ATF}

The installation of an instrument in an accelerator bring a large number of additional constraints compared to a laboratory environment. The constraints include the inability to access the hardware (including oscilloscopes close from the experiment) during the operations, additional vibrations, thermal effects,...
On MightyLaser we had to adjust the experiment to these constraints by developing tools that allowed us to control and operate the cavity remotely. We also had to correlate our measurements with the accelerator status to show that what we measured was a real signal and not an unexpected background. Our first results (obtained before the earthquake that devastated Japan in 2011) are documented in~\cite{Delerue:2011nk,Akagi:2011hj}.

Together with a graduate student, Iryna Chaikovska, I developed most of these remote tools. The layout of data acquisition system that we developed over a few weeks to be ready for the data taking is shown on figure~\ref{fig:ml_daq}.

\begin{figure}[htbp]
  \centering
  \includegraphics[width=0.99\linewidth]{Compton/201012_cDAQ_architecture_v3.png}
  \caption{Layout of the data acquisition system developed for the MightyLaser experiment at KEK.}
  \label{fig:ml_daq}
\end{figure}

Examples of the data we obtained are shown in figure~\ref{fig:ml_data}. The full analysis is documented in~\cite{Akagi:2011hj} and in Iryna Chaivovska's thesis~\cite{Chaikovska:2012hja}.

\begin{figure}[htbp]
  \centering
  \includegraphics[width=0.45\linewidth]{Compton/14122010_file_data_locked2_3.pdf} 
  \includegraphics[width=0.45\linewidth]{Compton/TUPO002_peakheight_spectrum_20101214.pdf}
  \caption{Left: image from~\cite{Akagi:2011hj} showing the typical signal shape for the high energy $\gamma$ rays produced by Compton scattering observed in the MightyLaser experiment. Each spike on the picture corresponds to the $\gamma$ production after
successive bunch crossings over \SI{0.2}{ms}. Right: image from~\cite{Delerue:2011nk}, showing the $\gamma$ ray spectrum for different laser power (LP) stored in the Fabry-Perot cavity.}
  \label{fig:ml_data}
\end{figure}

 After the earthquake the lack of environmental control in the accelerator during several weeks required significant work to put the cavity back in working order. After attempts to do this work in Japan it was decided to do it in France before a reinstallation in Japan. It was only in December 2012 that we were ready to take data again with a significantly increased power enhancement in the cavity. This power enhancement was however limited by new problems such as the heat load on the mirrors which modified there curvature and hence the geometry of the cavity\footnote{These problems were later solved during the thesis of another graduate student (Pierre Favier\cite{ThesePierreFavier}) but I was not part of tis work.}. 
 Example of the results obtained at that time are shown on figure~\ref{fig:ml_DataScreenShot} and have been published in~\cite{Chaikovska:2016aa}.


%\ifnoemb
\begin{figure*}[htbp]
   \centering
\includegraphics*[width=70mm]{Compton/screenshot_data_taking_long473.png} 
\includegraphics*[width=70mm]{Compton/20131213_run2_long_data_datafile_473.eps} 
   \caption{ (left) Typical screen shot of a the data taken during one electron storage period with the laser power obtained in December 2012. The gamma ray signal from the calorimeter is shown in brown (this signal is negative because of the PMT bias), the photodiode signal giving the power in the cavity is shown in blue. The pink and green signals are used to monitor beam condition. (right) Example of data after processing: only the red area is considered as being signal for a given beam (the blue peak to the left corresponds to noise observed during the injection of the beam in the ring and the blue area to the right corresponds to signal from another beam injection).}
   \label{fig:ml_DataScreenShot}
\end{figure*}

%\else
%\embargo
%\fi

\section{Beam dynamics of Compton scattering at the ATF}

Another important point was to understand the effect of Compton scattering on the electron beam stored in the ring. Although it was not expected to have any significant effects on the beam with the laser power that we had in the cavity this was an important point to check. This work was done by Iryna Chaikovska under my guidance~\cite{Chaikovska:2012hja,Chaikovska:2011nj}. It showed that given the short storage duration no effect were to be expected and even the beam diagnostics in the extraction line would not be sensitive enough to see anything. After the 2012 upgrade when we took long data runs we noticed (see figure~\ref{fig:ml_LongStorage}) that we were clearly able to see in our data the effect of intrabeam scattering as predicted by~\cite{Kubo:2001ps} and this has to be taken into account when predicting the total flux that can be produced. These results have been published~\cite{Chaikovska:2016aa}.

This work also led me to similar work applied to the  ThomX  ring. It will be described in chapter~\ref{chap:thomx}.

%\ifnoemb

\begin{figure*}[htbp]
   \centering
\includegraphics*[width=70mm]{Compton/screenshot_long_data_taking8.png} 
\includegraphics*[width=70mm]{Compton/lifetime.png} 
   \caption{ (left) Data taken during one long electron storage period. The effect of the electron damping (from 50ms to 150ms) is clearly visible as well as the effect of intrabeam scattering after that. The gamma ray signal from the calorimeter is shown in brown (this signal is negative because of the PMT bias), the photodiode signal giving the power in the cavity is shown in blue. The pink and green signals are used to monitor beam condition. (Right)  Beam charge as a function of time with and without Compton. The effect of Compton scattering on particle losses is clearly visible.}
   \label{fig:ml_LongStorage}
\end{figure*}

%\else
%\embargo
%\fi


\section{Extension: ELI-NP-GS}

The success of the work on MightyLaser led to other applications. In particular the European Light Infrastructure project (ELI) for its Nuclear Pillar (NP) will use Compton scattering for its Gamma Source (GS).

At first we considered using a Fabry-Perot cavity for this application as had been done in MightyLaser however a quick calculation shows that given the high $\gamma$ flux required and the low repetition rate (128 pulses\footnote{At that time 32, 64 or even 128 crossing were considered; it was later decided to have only 32 crossings as it made the geometry easier.}  per macro pulse repeated at \SI{100}{Hz}) this solution is not favored.
Two reasons explain this: the average power stored in the cavity would induce significant thermal effects\footnote{When we did this work, in November 2011, we did not know that we would already see the thermal effects at the ATF experiment but we suspected that such effect would appear at an average power lower than what was required for ELI-NP-GS.} and loading a cavity for only 128 crossings would waste more power (due to the reflexion on the input mirror) than it would save thanks to the power enhancement inside the cavity, the laser system would therefore be more complicated. Instead of a cavity we therefore decided to use recirculator made of a large set of mirrors between which the pulses would bounce between each pass at the interaction point.

The beam crossing angle required significant attention: the Compton cross-section is much higher in the case of head-on collisions than when the beam has a higher crossing angle. However the mirrors need to be positioned around the interaction point and a small crossing angle prevent the positioning of a sufficient number of mirrors. This is illustrated on figure~\ref{fig:eli_cavity_nb_pass}. Although this points to a large number of pulses other parameters (and further optimisation) led to the choice of 32 pulses.

\begin{figure*}[htbp]
   \centering
\includegraphics*[width=0.45\linewidth]{Compton/nb_pass_vs_angle_500fs_step_2016.eps} 
\includegraphics*[width=0.45\linewidth]{Compton/lumi_vs_angle_500fs_step_2016.eps} \\
\includegraphics*[width=0.6\linewidth]{Compton/lumi_np_pass_vs_angle_500fs_step_2016.eps} 
   \caption{(top left) Number of passes possible as a function of the crossing angle. (top right) Luminosity as a function of the crossing angle (normalized to the maximum luminosity) for different laser pulse duration (FWHM laser pulse duration given in the legend in seconds). (bottom) Same figure but summed over the number of passes. }
   \label{fig:eli_cavity_nb_pass}
\end{figure*}

Further optimisation work was necessary to be able to adjust the timing between each pulse to the structure of the train. The resulting recirculator has been documented in~\cite{Dupraz:2014zha} (and is now being built but I no longer work on this project). The whole technical description of the project has been documented in~\cite{Adriani:2014cma}. This proposal was discussed and negotiated for more than a year with our Romanian counterparts and it is now under construction.

\input{./tail}