1: \subsection{Single peak analysis}
2: \label{sec:SinglePeakAna}
3: Individual peaks have been identified either from the raw 3D data
4: sets, or more frequently from azimuthal projections.
5:
6: The position of the maximum, $\ve q$, the full width at half maximum
7: (FWHM) in the three directions, $\Delta \ve q$, and the integrated
8: intensity, $I$, have been found for the individual peaks.
9:
10: \pagebreak
11:
12: The analysis is based on the full 3D intensity distribution in
13: reciprocal space, according to the following procedure:
14: \begin{itemize}
15: \item The voxel with maximum intensity was identified.
16: \item Three one-dimensional intensity profiles were gathered (along
17: approximately\footnote{The lines are taken along straight pixel lines on
18: the detector.} the $\ve q_x$, $\ve q_y$ and $\ve q_z$-directions)
19: going through the maximum voxel, and \emph{not} integrating in the
20: other directions.
21: \item A one-dimensional pseudo Voigt function plus a constant term
22: were fitted (using least square fitting) independently to each of
23: these three profiles (figure \ref{fig:fitex} show an example of such
24: a set of three fits).
25: \end{itemize}
26:
27: \begin{figure}[t]
28: \centering
29: \begin{minipage}{1.2\linewidth}
30: \begin{pspicture}(0,0)(15,4) %\psgrid[subgriddiv=1]
31: \rput[bl](0,0){
32: \begin{minipage}{1.0\linewidth}
33: \includegraphics[width=4.5cm]{AnalysisMethods/figs/fit_ex_qx.eps}
34: \includegraphics[width=4.5cm]{AnalysisMethods/figs/fit_ex_qy.eps}
35: \includegraphics[width=4.5cm]{AnalysisMethods/figs/fit_ex_qz.eps}\\
36: \end{minipage}}
37: \rput(1.5,3.5){$\ve{q_x}$}
38: \rput(6.2,3.5){$\ve{q_y}$}
39: \rput(10.8,3.5){$\ve{q_z}$}
40: \end{pspicture}
41: \end{minipage}
42: \caption{Example of fitting a single peak to three individual
43: pseudo-Voigt functions along the three reciprocal space
44: directions. Dots indicate measured intensities and lines the
45: fit obtained, the horizontal line indicates the full width at half
46: maximum. From \mycitet{acta}.}
47: \label{fig:fitex}
48: \end{figure}
49:
50: The fits directly give the position of the maximum in reciprocal
51: space and full width at half maximum of the peak in the three
52: directions.
53:
54: On figure \ref{fig:fitex} it is seen that the constant level in the
55: two azimuthal directions ($\ve q_x$ and $\ve q_z$) are approximately
56: the same, this indicates that the procedure in these directions
57: correctly separates the cloud and the peak. It is also seen that the
58: constant level in the radial direction ($\ve q_y$) is lower, in fact
59: it is almost at the detector background level. The reason for this is
60: that it is very hard to separate the cloud and the peak in this
61: direction as both tend toward $0$. This has the consequence that the
62: method has a tendency of overestimating the width in the $\ve
63: q_y$-directions. The uncertainty on the width is mainly attributed to
64: the uncertainty on the level of the cloud, and hence on the half
65: maximum level. It can be estimated individually from the fits, by
66: varying the constant level between the fitted values.
67:
68:
69: The integrated intensity in a peak is found from the fits to the
70: peaks. It is assumed that the average constant level and peak height
71: found in the azimuthal directions can be used in the radial direction.
72: The uncertainty of the total intensity was estimated by varying the
73: background level used and the peak height between the minimum and
74: maximum fitted values and constructing the corresponding upper and
75: lower limits of the intensity. The estimated uncertainty is $\approx
76: 15\%$.
77:
78: