cond-mat0404371/arx.tex
1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
3: 
4: \documentclass{modified}
5: 
6: \begin{document}
7: 
8: 
9: \title{LINE SHAPE ANALYSIS OF LINEAR X RAY MAGNETIC\\
10: SCATTERING COPT THIN FILMS
11: }
12: 
13: \author{\footnotesize E.V.R.CHAN}
14:  
15: 
16: \address{ University of Washington,Box 351560\\
17: Seattle, Washington, 98195-2420, United States.\\
18: evr@u.washington.edu}
19: 
20: \maketitle
21: 
22: 
23: \begin{abstract}
24: Data analysis of the CCD files from x ray magnetic 
25: resonance scattering linearly polarized in transmission 
26: geometry produces information about the radial and
27: azimuthal intensities.  In a series of measurements 
28: of increasing photon energies trends in data are
29: analyzed. 
30: \end{abstract}
31: 
32: \section{Introduction}
33: 
34: Magnetic thin film systems and multilayer systems have
35: been studies very actively because of their magnetic
36: properties and possible application for practical
37: devices, such as magnetic recording media technologies.
38: 
39: \section{Experiment}
40: 
41: Samples were grown on smooth, low-stress, 160 nm. thick
42: SiNx membranes by magnetron sputtering; they all had
43: 20 nm. thick Pt buffer layers and 3 nm. thick Pt caps. 
44: Between the buffer layer and the cap, the samples had 
45: 50 repeating units of a 0.4 nm. thick Cobalt layer and
46: a 0.7 nm. thick Pt layer.  Experiments used linearly 
47: polarized x rays from the Advanced Light Source at 
48: Lawrence Berkeley Laboratory (supported by USDOE),
49: the thirteenth harmonic of the beamline 9, undulator  
50: gap of 54 mm. near the resonant Cobalt L edge.  
51: To achieve transverse coherence, the raw 
52: undulator beam was passed through a 35 micron diameter 
53: pinhole before being scattered in transmission by 
54: the sample.  The distance from the sample to the CCD 
55: is 118 cm.  The resonant magnetic scattering was
56: collected by the Princeton soft x-ray CCD camera 
57: 1024 X 1024 pixels in an area one inch by one 
58: inch.  The intensity of the raw undulator beam
59: was 2 X 10 \verb_**_ 14 photons/sec., the intensity of the 
60: coherent beam was 2 X 10 \verb_**_  12 photons/sec., and
61: the intensity of the scattered beam, was 2 X 10 \verb_**_  7 
62: photons/sec.  Each speckle pattern was measured 
63: for 30 to 100 seconds, so the total number of
64: photons in each CCD image 1024 X 1024 pixels 
65: is about 10 \verb_**_  9.  The speckle patterns may
66: be used to reconstruct the magnetic domain 
67: structure of the sampl$e^{1-3}$; 
68: this is but one of a general class of the old 
69: inverse or phase retrieval problem$s^4$. 
70: 
71: 
72: 
73: \begin{figure}[th]
74: \centerline{\psfig{file=evrFig1.eps,width=11cm}}
75: \vspace*{8pt}
76: \caption{(left) Average intensity versus filenumber 
77: for the series of images.  
78:  (right) Azimuthally averaged
79: radial intensity of image file 122.} 
80: \end{figure}
81: 
82: 
83: \section{Data Analysis}
84: 
85: The data appears as a series of CCD image files 
86: 726Axxx.SPE (of increasing photon energy) that
87: is read into the data processing computer program.
88: The file numbers can be converted to photon energy
89: since the difference between the two maxima files 122
90: (photon energy 779.2 eV.) and 145 is 15.8 eV.
91: Figure~1 (left) shows the magnetic resonance has
92: two peaks.  Each piece of data has an image file
93: associated and that file has a 1024 X 1024 matrix 
94: containing intensity values ( if plotted in 3D 
95: it has the shape of a centered crown ). 
96: The Princeton CCD camera 
97: file is read into the freely available Matlab 
98: data analysis program by the following code 
99: fragment:  
100: \begin{verbatim}
101: %auto-ignore 
102: fid=fopen('nameOfFile.SPE','r');
103: header=fread(fid,2050,'uinit16'); %half of 4100 bytes
104: ImMat=fread(fid,1024*1024,'uint16');
105: Z=reshape(ImMat,1024,1024);
106: fclose(fid);Z=double(Z);
107: [X,Y]=meshgrid(1:1024,1:1024);
108: mesh(X,Y,Z); %display 3D plot
109: axis square; axis tight; view(90,90);
110: print -djpeg99 nameOfFile.jpg %highest resolution saved
111: %send email  evr@u.washington.edu
112: \end{verbatim}  
113: Figure~1(right) shows the variation of the intensity in 
114: file 122 in the radial direction that has been 
115: azimuthally averaged (consider slicing through 
116: the center out past the edge of the crown).  The
117: data analysis calculates the radius of each pixel, 
118: places it into appropriate bins and finds the 
119: average intensity; since each bin is only one 
120: pixel wide the bin number is the radius rounded 
121: to integer.  The profile was fitted to a gaussian
122: using non-linear Levenberg-Marquardt least squares.  
123: The difference between the maximum and minimum 
124: calculated values is the height, the distance of 
125: the maximum from the origin is called the peak 
126: center, and the FWHM (full width at half maximum) 
127: is called the width.  More about the variation
128: of the height, width and center with filenumber
129: will appear at the end.  
130: 
131: 
132: \section{Results and Discussion}
133: 
134: The series of files 726Axxx.SPE are analysed for normalized
135: cross-correlatio$n^5$ using 2D matrices of the image 
136: files from which the average pixel values have been subtracted.  
137: The sum of the product of pixel values from two images divided
138: by the square root of the product of the sum of the pixel values 
139: squared of each image (makes it normalized) is gamma 
140: (normalized cross-correlation) of the two images.  If 
141: gamma is 1 they are correlated, if zero uncorrelated 
142: and if -1 anti-correlated.  The normalized 
143: cross-correlation function, gamma, is related to the 
144: coherence functio$n^6$. Table 1 shows the values of 
145: gamma for four of the files selected as file 
146: 117(middle of first peak), file 122 (top of first peak), 
147: file 133 (minimum between the two peaks) and file 145 
148: (top of second peak).
149: 
150: 
151: 
152: 
153: \begin{table}[pt]
154: \tbl{  Normalized      Cross-Correlation}
155: {\begin{tabular}{@{}ccccc@{}}\toprule
156:    &  117 & 122 & 133 & 145 \\
157: \colrule
158: 117 & 1.0 & 0.8 & 
159: -0.8 & -0.8 \\
160:  122 & 0.8 & 1.0 & 
161:   -0.9 & -0.9 \\
162: 133 & -0.8 & -0.9 & 1.0 & 
163: \hphantom{0}0.9 \\
164: 145 & -0.8 & -0.9 & 0.9 & \hphantom{0}1.0\\ \botrule
165: \end{tabular}}
166: \end{table}
167: 
168: 
169: 
170: 
171: The CCD images were processed so as to remove the burns,
172: remove anomalous charge scattering, remove the blocker arm, 
173: centered in the image and the central disk darkened to 
174: remove the burns in that area also.  The bright spots on 
175: the image are fixed burns in the CCD camera that cause a 
176: few pixels to be unusable; these are the ones that 
177: appear very bright. At the end of a series of SPE image
178: files a CCD burns only image was taken and a burns and 
179: anomalous charge scattering image was taken.  The difference 
180: of these two images was used to subract off the anomalous 
181: charge scattering pisels from each image file pixels 
182: (provided they are above the minimum background value 
183: of the image).  After the anomalous charge scattering is 
184: removed, the burns (or hot spots) are removed.  The hot 
185: spots are masked off, and after rotating around the 
186: middle of the 1024 X 1024 image, new background replaces 
187: the hot spots.  Next, the circle of maximum intensity is 
188: found by having an imaginary turtle going out from beyond 
189:  the edge of the blocker disc in rays every 1 degree 
190: finding maximum intensity. The intensities are sorted, 
191: the lowest 20 points out of 360 dropped and a least 
192: squares fit of the 340 points is done yielding the circle 
193: of maximum intensity.  The approximate edges of the 
194: blocker arm (about 7 degrees) are located by finding 
195: those points on the circle of maximum intensity where 
196: the second derivative changes sign.  A mask is created, 
197: a rotation picks up a patch, and the pixels of the 
198: blocker arm are replaced.  Missing background is fille 
199: in on the edges to make 1024 X 1024.  All pixels further 
200: out from the center, beyond the circle of maximum 
201: intensity were used to find the centroid and then the 
202: whole pattern was relocated to the center of the image 
203: matrix.  
204: 
205: 
206: 
207: 
208: 
209: \begin{figure}[th]
210: \centerline{\psfig{file=evrFig2.eps,width=11cm}}
211: \vspace*{8pt}
212: \caption{(upper left)Polar plot of average azimuthal angular
213: intensity looking down the z axis at file 122 (lower left)
214:  azimuthal angular intensity of file 122  (upper right )
215:  V1 versus filenumber where V1 is the difference between the 
216: first maximum azimuthal radially averaged intensity and the 
217: lowest minimum (lower right) V2 versus filenumber where V2
218: is the difference between the second maximum azimuthal
219: radially averaged intensity and the lowest minimum}
220: \end{figure}
221: 
222: 
223: 
224: 
225: Due to the use of coherent light from synchrotron radiation, 
226: the scattering patterns produced are highly speckled.  
227: Each speckle is the sum of light scattered from all 
228: the illuminated magnetic domains.  So small changes 
229: in the microscopic orientation of the magnetic domains 
230: can have a large effect on the speckle pattern.  The bulk 
231: magnetization of Cobalt usually is in the x-y plane 
232: but when there are only a few Cobalt atoms in the layer 
233: the magnetication becomes perpendicular to the plane of 
234: the layers.  The observed intensity is related to the 
235: charge density, polarization, photon energy, magnetization, 
236: atomic scattering factor and scattering geometry; there are 
237: terms due to magnetic resonance and electronic structur$e^7$.   
238: 
239: 
240: The azimuthal variation of intensity is calculated by 
241: looking at the angle from the center for each pixel, 
242: classifying it as belonging to bins (1 to 360) and 
243: finding the average of each bin; this is plotted 
244: versus the bin number with each bin being one degree 
245: wide.  In Fig.~2 the difference between the maxima and the 
246: lowest minimum is V1, V2.  The azimuthal variation of the 
247: intensity is described by an equation for magneto-crystalline 
248: anisotropy energy that includes a term proportional to 
249: the sine of twice the azimuthal angle square$d^8$.
250: The amplitude has real and imaginary charge and magnetic 
251: anomalous scattering factors which are tensors in  the 
252: general cas$e^9$.
253: 
254: The right side of Fig.~2 pretty much follows the shape of 
255: average intensity versus file number in Fig.~1(left) 
256: and illustrates the variation of the magnetic anisotropy 
257: difference (maximum minus lowest minimum) with photon 
258: energy(increases with file number).  The average of 
259: the value of first maximum and second maximum would be 
260: a reasonable estimate of the variable.  The reason the 
261: magnetic anistropy has these variations is because 
262: the atomic factor is a tenso$r^{10}$.
263: 
264: 
265: 
266: 
267: 
268: 
269: \begin{figure}[th]
270: \centerline{\psfig{file=LevrFig3.eps,width=11cm}}
271: \vspace*{8pt}
272: \caption{(upper left)contour plot of autocorrelation 
273: file 117 halfway up first peak, (lower left)
274: contour plot of autocorrelation file 133 minimum
275: between peaks,  (upper right )contour plot of autocorrelation
276: file 122 tope of first peak, (lower right) 
277: contour plot of autocorrelation file 145 top of second peak}
278: \end{figure}
279: 
280: 
281: 
282: 
283: 
284: 
285: 
286: The autocorrelation has a maximum at the origin.  You 
287: could think of it as the convolution of the complex 
288: conjugate of f(-x,-y) and f(x,y).  If f(x,y) has a 
289: Fourier transform F(s,r), then its autocorrelation 
290: function has the transform absolute value squared of 
291: F(s,r) and has no phase information(Wiener-Khinchin 
292: theorem).  In the data analysis, the autocorrelation 
293: is evaluated by taking the Fourier transform of the 
294: reverse complex conjugate of the image and the Fourier 
295: transform  of the image, then taking the real part 
296: of the inverse Fourier transform of the product of 
297: these two Fourier transforms.  The three dimensional 
298: plot of the autocorrelation looks like a mountain with 
299: a narrow spike in the middle.  The threshholded (25-30\%) 
300: contours(5) in the x-y plane of the different files 
301: show more variation and different symmetries.  
302: Figure~3(upper left) File 117, halfway up the first 
303: peak, looks like it has four 2-fold axes.  
304: Figure~3(upper right) File 122, at the top of the 
305: first peak, may have one 4-fold and two 2-fold axes.  
306: Figure~3(lower left) File 133, the minimum between 
307: the peaks, appears to have one 4-fold axis and two 
308: 2-fold axes.  Figure~3(lower right) File 145, top 
309: of the second peak, seems to have four 2-fold axes.  
310: 
311: 
312: 
313: \begin{figure}[th]
314: \centerline{\psfig{file=evrFig4.eps,width=11cm}}
315: \vspace*{8pt}
316: \caption{(top) Peak height from the fitted gaussian versus
317: filenumber (middle) Inverse of the full width at half 
318: maximum versus filenumber (bottom) Inverse of center of
319: peak versus filenumber}
320: \end{figure}
321: 
322: 
323: 
324: In Fig.~4(top), the height of the peak verus file number is
325: plotted and is in agreement with Fig.~1(left).  The correlation 
326: length is related to the inverse of the width.  In 
327: Fig.~4(bottom), the inverse of the peak center is plotted 
328: versus file number; the spacing of the magnetic domains is 
329: inverse to the peak center.  In all cases the curves 
330: follow Fig.~1(left) more or less; the last one doesn't
331: follow the second peak very well.
332: 
333: 
334: 
335: 
336: 
337: \section{Conclusions}
338: 
339: It is possible to see azimuthal intensity variation 
340: through data anaylsis of CCD images.  Because of the odd 
341: behavior of the second peak and the differences in the 
342: autocorrelation functions(reflecting differences in 
343: electronic configurations) of the first and second peaks, 
344: there appears to be a difference in the nature of the 
345: first and second peaks.
346: 
347: 
348: 
349: 
350: 
351: %\section{References}
352: 
353: 
354: \begin{thebibliography}{0}
355: \bibitem{1}
356: T. O. Mentes, C. Sanchez-Hanke and C. C. Kao,
357: {\it J. Sync. Rad.} {\bf 9}, 90 (2002).
358: %J. Callaway, {\it Phys. Rev.} {\bf B35}, 8723 (1987).
359: 
360: \bibitem{2}
361: %M. Tinkham, {\it Group Theory and Quantum Mechanics}
362: %(McGraw-Hill, New York, 1964).
363: F. Yakhou, A. Letoublon, F. Livet, M. de Boissieu and 
364: F. Bley, {\it J. Magn. Magn. Mater} {\bf 233}, 119 (2001).                          
365: \bibitem{3}
366: %T. Tel, in {\it Experimental Study and Characterization of
367: %Chaos}, ed. Hao Bailin (World Scientific, Singapore, 1990), p.
368: %149.
369: A. Rahmim, M.Sc. Thesis, Univ. of British Columbia, Vancouver,
370: 2001.
371: 
372: \bibitem{4}
373: %P. P. Edwards, in {\it Superconductivity and Applications
374: %--- Proc. Taiwan Int. Symp. on Superconductivity}, ed.  P. T. Wu
375: %{\it et al.} (World Scientific, Singapore, 1989), p. 29.
376: N. E. Hunt, {\it Phase Retrieval and Zero Crossings: 
377: Mathematical Methods in Image Reconstruction} (Kluwer,
378: Dordrecht, 1989).
379: 
380: \bibitem{5}
381: A. Rahmim, S. Tixier, T. Tiedje, S. Eisebitt, 
382: M. Lorgen, R. Scherer, W. Eberhardt, J. Luning 
383: and A. Scholl, {\it Phys. Rev. } {\bf B65}, 
384: 235421 (2002). 
385: 
386: \bibitem{6}
387: %W. J. Johnson, Ph.D. Thesis, Univ. of Wisconsin, Madison, 1968.
388: J. Benesty, D. R. Morgan and J. H. Cho, {\it IEEE Transactions
389: on Speech and Audio Processing} {\bf 8}, 168 (2000).
390: 
391: \bibitem{7}
392: %P. F. Marteau and H. D. I. Arbabanel, ``Noise reduction in
393: %chaotic time series using scaled probabilistic methods'',
394: %UCSD/INLS preprint, October 1990.
395: M. Blume, in {\it Resonant Anomalous X-ray Scattering}, ed. 
396: G. Materlik {\it et al.} (Elsevier Science, Amsterdam, 1994), 
397: p. 495. 
398: 
399: \bibitem{8}
400: R. C. O'Handley, {\it Modern Magnetic Materials} 
401: (Wiley, New York, 2000).
402: 
403: \bibitem{9}
404: D. H. Templeton, {\it Acta Cryst.} {\bf A54}, 158 (1998).
405: 
406: \bibitem{10}
407: S. Di Matteo, Y. Joly, A. Bombardi, L. Paolasini, F. de 
408: Bergevin and C. R. Natoli, {\it Phys. Rev. Lett.} 
409: {\bf 91}, 257402-1, (2003).
410: 
411: 
412: 
413: \end{thebibliography}
414: 
415: \end{document}
416: 
417: 
418: