1: \subsection{First order rebinning}

2: \label{sec:rebin}

3: The problem of converting data obtained as pixels on a 2D detector to

4: pixels in reciprocal space, or equivalent to angular space defined by

5: pixels in a $\eta, 2\theta$ coordinate system, is general for many

6: applications of area detectors.  The major problem is that a square

7: pixel on the detector does not corresponds to a square pixel in

8: reciprocal space, and that the shape and size of the pixel in angular

9: space depends on the position of the pixel on the detector (figure

10: \ref{fig:PixelToResspaceEx} illustrates the differences in the case of

11: the present setup).

12: 

13: Multiple software packages (such as \citep{fit2d}) exist that are

14: able to perform the general transformation, known as rebinning. The

15: geometry of the present setup does, however, seems to give problems with

16: FIT2D as the beam center is located far outside the image obtained. A

17: \textit{first order rebinning} scheme has therefore been used.

18: 

19: \begin{figure}

20:   \centering

21:   \begin{tabular}{p{0.3\textwidth}p{0.3\textwidth}p{0.3\textwidth}}

22:  \includegraphics[width=0.3\textwidth]{AnalysisMethods/figs/pix_ang_space_1.eps} \vfill &

23:   \includegraphics[width=0.3\textwidth]{AnalysisMethods/figs/pix_ang_space_2.eps} \vfill &

24:   \includegraphics[width=0.3\textwidth]{AnalysisMethods/figs/pix_ang_space_3.eps}  \vfill     

25:   \end{tabular}

26:   \caption{The mapping of one pixel squared on the detector onto the

27:     $(\ve q_x,\ve q_y)$-plane for three horizontal pixel positions.

28:     Geometrical parameters are chosen within typical values of the

29:     actual setup. The mapped pixel are located at (left to right)

30:     pixel nr. $-1000$, $0$, $1000$ with respect to the horizontal

31:     beam center, and vertically close to the position of the 400

32:     reflection for Cu.}

33:   \label{fig:PixelToResspaceEx}

34: \end{figure}

35: 

36: The differences in area in reciprocal space between the different

37: pixels are minor, as the detector (and more important the interesting

38: reflection) only covers a small angular range in the present case.

39: This can also be seen on figure \ref{fig:PixelToResspaceEx}, where the

40: differences are illustrated in the extreme case.

41: 

42: For the typical case (as represented in the data presented in

43: \mycitet{acta}) the maximal difference in the size of the voxels that

44: corresponds to the pixels in the relevant region was calculated.  In

45: the azimuthal direction, $\ve q_x$, the change in length is at most

46: $0.33\%$ of the minimum length, and in the radial direction, $\ve

47: q_y$, the change is at most $0.05\%$ of the minimum length. In the

48: other azimuthal direction $\ve q_z$ the size is constant. This gives a

49: maximal difference in the volume of the voxels of $0.4\%$, and a

50: first order rebinning scheme, which does not take this volume

51: difference into account, is therefore appropriate.

52: 

53: The rebinning is not performed separately but is integrated into the

54: projections as described in the following. 

55: 

56: