1: %\documentclass[twocolumn,nofootinbib,showpacs,preprintnumbers,aps]{revtex4}

2: \documentclass [preprint,aps]{revtex4}

3: \usepackage{epsfig}

4: \usepackage{amssymb}

5: 

6: \begin{document}

7: \title{

8: EXPERIMENTS WITH PARAMETRIC X-RAY RADIATION (PXR) FROM

9: NON-RELATIVISTIC ELECTRONS}

10: 

11: \author{V.G.Baryshevsky, K.G.Batrakov, I.D.Feranchuk, A.A.Gurinovich,

12: A.O.Grubich, A.S.Lobko, A.A.Rouba, B.A.Tarnopolsky, P.F.Safronov,

13: V.I.Stolyarsky, A.P.Ulyanenkov}

14: 

15: \address{

16: Institute for Nuclear Problems, Belarus State University \\ 11

17: Bobrujskaya Str., Minsk 220050, Belarus }

18: 

19: \begin{abstract}

20: 

21: Interaction of non-relativistic electrons with single crystal target may

22: produce coherent x-rays. That is the result of interference between two

23: known x-ray generation mechanisms having orientational behavior, namely

24: parametric x-rays and coherent {\it Bremsstrahlung}. Experiments aimed to PXR research were

25: performed with 50-100 keV electrons and its distinctive features were

26: observed. Requirements to the experimental set-up, detector instrumental

27: response, and targets as well as experiment geometry are discussed in

28: detail. Series of PXR spectra in various conditions were recorded and their

29: distinctive features were observed. Tuning of radiation frequency with

30: crystal-target rotation was observed for the first time for low energy

31: electrons. Dependence of the x-ray frequency on the beam energy was

32: detected. Soft PXR peak with energy below 1~keV was observed for the first

33: time. Possible applications of PXR for structure analysis and

34: crystallography are discussed. These results are obtained in the framework

35: of ISTC project {\#}B626.

36: 

37: \end{abstract}

38: \maketitle

39: 

40: \section{INTRODUCTION}

41: 

42: Parametric x-rays produced by non-relativistic electrons passing

43: through a single crystal target were described in \cite{1} on the

44: basis of general theory of PXR \cite{2}. Due to the fact that the

45: angular distributions of x-rays originated from all radiation

46: mechanisms are almost isotropic in this case, peaks produced by

47: coherent effects, i.e. parametric x-rays and coherent {\it

48: Bremsstrahlung} (CBS), are the result of PXR and CBS interference

49: and may be observed on the intensive uniform background. Their

50: shape and intensity depend on the target thickness and the

51: detector spectral resolution. Some experiments with

52: non-relativistic electrons are currently known \cite{3}-\cite{5},

53: but their authors outlined the effect taking into consideration

54: only CBS mechanism. Model \cite{1} presumably describes x-rays

55: from non-relativistic electrons in crystal targets more

56: adequately.

57: 

58: To study x-ray properties followed from the theory \cite{1}, we

59: have performed experiments focused on the PXR research with

60: 50$\div$100~keV electrons of the electron microscope. Let us

61: briefly consider requirements of the PXR experimental observation.

62: In general case a charged particle passing though matter undergoes

63: elastic and inelastic collisions with the atoms of a medium. As a

64: result, the angular and energy distributions of electrons change

65: during beam passing through a medium. Influence of electron

66: scattering on the spectral-angular distribution becomes

67: significant if either the width of velocity distribution

68: $\frac{\Delta \nu }{\nu }$ or the width of angular distribution

69: $\Delta \theta _\ast $ becomes larger or equal to the width of the

70: spectral line: $\frac{\Delta \nu }{\nu }\ge \frac{1}{kL_\ast }$ or

71: $\Delta \theta _\ast \ge \frac{1}{kL_\ast }$, where

72: $k=\frac{\omega }{c}$ is the wave-vector of the radiated photon,

73: $L_\ast $ is the electron path in the crystal.

74: 

75: Angles of multiple scattering and influence of inelastic

76: scattering have been estimated for several crystals at low

77: electron energy. It was obtained, that to observe coherent

78: x-radiation in wide frequency range the rigid requirements for

79: target thickness and electron beam parameters should be fulfilled.

80: They are: the target thickness should not exceed $\sim $0.5~$\mu

81: $m and the energy of electron beam should be above 50-60~keV (in

82: this energy range angle of multiple scattering $<\theta_{s}^2>

83: \sim E^{-2})$. Initial energy and angular dispersion of the

84: electron beam should be less than 10$^{-1}$.

85: 

86: We propose to detect x-ray photons, which are radiated at small

87: sliding angle $\psi \ll 1$ with respect to the target surface in

88: the direction, which is opposite to electron velocity. In this

89: case the detector would register the photons from the electron

90: pass $\sim \psi/(k \chi^{\prime \prime})$ (if $\psi/(k

91: \chi^{\prime \prime}) < 1$), here $1/(k \chi^{\prime \prime})$ is

92: approximately equal to the absorption length of radiation in a

93: crystal, $\chi^{\prime \prime}$ is the imaginary part of crystal

94: susceptibility. Therefore, a part of electron beams with less

95: dispersion ($<\theta _{s}^2>\sim L_\ast$) radiates. This

96: experiment can be realized for such target and radiation

97: frequency, which correspond to small absorption length. For

98: example, absorption length for LiH crystal is $\sim $1~$\mu $m for

99: 10 $\AA$ radiation wavelength. Then, at the angle $\psi \sim $0.1

100: ($\sim 5.73^\circ$) one can observe radiation from the 0.1~$\mu $m

101: pass.

102: 

103: The PXR photons can be radiated at all directions. The frequency of PXR

104: photons depends not only on the angle between the electron velocity and

105: crystallographic plane {\it $\theta $}$_{H}$ but also on the angle $\theta $ between the

106: electron velocity and direction of registration

107: %

108: \begin{eqnarray}

109: \omega_{H}^{(n)}(\theta)=\frac{2 \pi v \cos{\theta_H}}{d (1-v/c

110: \cos{\theta})} n,~ n=1,2,... \label{1}

111: \end{eqnarray}

112: %

113: where $d$ is the interplanar interval, $v$ is the electron

114: velocity, $c$ is the speed of light. When the observation angle

115: $\theta $ is fixed, the PXR spectral width is defined by the

116: length of the coherent interaction of the electron in crystal,

117: that is before the multiple scattering becomes essential. So, the

118: effective observation of PXR with the non-relativistic electrons

119: is possible for the single crystal films with the thickness lower

120: than effective scattering length, $L<L_{sc}$. In this case the

121: relative spectral width of the PXR "harmonic" is defined by the

122: coherent interaction of the electrons with crystal and can be

123: calculated as follows:

124: 

125: \begin{eqnarray}

126: \frac{\Delta \omega_0}{\omega}=\frac{v}{L

127: \omega_{H}^{(n)}(\theta)}n

128:  \label{2}

129: \end{eqnarray}

130: 

131: \section{EXPERIMENT}

132: 

133: Taking into account conditions described above, we have performed

134: the experiment outlined in Fig. {\ref{Fig1}}. Narrow electron beam

135: falls on the surface of the single crystal target, $\vec{v}$ is

136: the velocity of electron beam ($\vec{v}\uparrow \downarrow OZ$).

137: Axis $OX$ is coplanar the target surface, $\vec{N}$ is orthogonal

138: to the crystal surface. Angle $\Theta_{0}$ is the angle between

139: $\vec{N}$ and $OZ$, which determines crystal rotation around the

140: axis $OX$. At $\Theta_{0}=0$ the target surface is orthogonal to

141: the velocity of electron beam. At crystal rotation around the axis

142: $OX$, $\vec{N}$ moves in the plane $ZY$. Detector window is placed

143: at the angle $\sim \pi /2$ to the velocity of electrons in the

144: direction $OY$.

145: 

146: \begin{figure}

147: \centering

148: \includegraphics [scale=1] {scheme1.eps}

149: \includegraphics [scale=0.6] {scheme2.eps}

150: \caption {Scheme and layout of the experiment} \label{Fig1}

151: \end{figure}

152: 

153: 

154: 

155: To provide requirements to the beam, we have applied electron beam of an

156: electron microscope. Electron beam injector was thoroughly tested

157: for stability and reliability of beam shaping systems and to optimize beam

158: parameters for CXR observation. Some measured beam parameters are as

159: follows:

160: 

161: \begin{itemize}

162: \item electron energy 50-100~keV

163: \item relative instability of accelerating voltage 2*10$^{-5}$

164: \item brightness 7*10$^{4}$ À/m$^{2}$ sr.

165: \end{itemize}

166: Careful selection of optimal currents of lenses and voltage between Venelt

167: cylinder and anode minimized parasitic scattering inside the microscope and

168: provide minimal x-ray background inside the experimental area. To reduce

169: background caused by electron scattered at the target, we have applied

170: electrostatic deflector made of two brass plates of the 50~mm length

171: positioned in the input window of the detector. The six kilovolts voltage

172: was applied between to them that decrease background rate significantly.

173: 

174: We have used Si(Li) detector with thermoelectric cooling and Si(Li) detector

175: with cryostat both with 20~mm$^{2}$ sensitive areas and thin polymer windows

176: for soft x-ray detection supplied by BSI (Riga, Latvia). Energy resolution

177: of detectors was evaluated as $\sim $170~eV. Detector angular aperture was

178: 0.2~mrad. ORTEC 2056-C 4096-channel analyzer collected detector output data.

179: Off-line spectra processing was performed after spectra transferring to the

180: computer. It involved calibration, smoothing, baseline subtraction, and

181: fitting by the set of Gauss peaks.

182: 

183: Analysis above shows that optimal target thickness should be below

184: half of micron. Applied target (Fig.{\ref{Fig2}}) is the silicon

185: crystal substrate of 2x2~mm dimensions and $\sim $200~$\mu $m

186: thickness with $\sim $0.5~$\mu $m thickness membrane of 1.0~mm

187: diameter. Basic plane has (100) or (111) orientation. Membrane

188: material is layer of pure epitaxial Si of $\sim $0.9--1.0~$\mu $m

189: thickness deposited on substrate of heavily doped p$^{+}$ Si of

190: KDB 0.01~$<$100$>$ grade. Choice of such structure was determined

191: by electrochemical etching technique, in which pure epitaxial Si

192: serves as termination layer. For membrane of other thickness one

193: should take structures with epitaxial layer thickness close to

194: desired one. Precise membrane thickness adjustment can be

195: performed by ion-beam etching with $\sim $10-15~nm/min rate.

196: 

197: \begin{figure}

198: \centering

199: \includegraphics [scale=0.6] {target.eps}

200: \caption {Sketch of Si (100) target of sub-micron thickness, E =

201: epitaxial layer} \label{Fig2}

202: \end{figure}

203: 

204: We have used relatively simple technique to measure thickness of such

205: ultra-thin Si targets. As membranes with thickness of about micron and below

206: are going to be semi-transparent in visible light range, one can record

207: their optical transmittance spectra. At this spectra interference fringes

208: can be observed and than thickness may be calculated. Dispersion of measured

209: thickness values normally is below 5{\%}.

210: 

211: \section{PRELIMINARY RESULTS AND DISCUSSION}

212: 

213: 

214: Experiments described below were performed with (111) Si target of $\sim

215: $410~nm thickness. Working current of $\sim $150~nA was chosen to provide

216: count rate at detector below 3~kHz and avoid peak distortion. Data

217: acquisition time was 5,000~seconds for the majority of spectra. Some spectra

218: were recorded at 10,000~seconds, but due to equipment instabilities obtained

219: spectra did not have considerable improvement.

220: 

221: Fragments of raw (just smoothed by Savitsky-Golay algorithm)

222: normalized PXR spectra are shown in Fig.{\ref{Fig3}}. One can see

223: dependence of PXR peaks frequency on beam energy. PXR frequency is

224: increasing with beam energy increase in conformity with formula

225: (\ref{1}). Group of characteristic x-ray peaks in right part of

226: spectra correspond to the microscope constructional materials

227: excited by electrons scattered by the target. Characteristic peak

228: of Si locates at $\sim $1.73~keV and it is not shown in picture

229: because it is very intensive in comparison with other peaks.

230: 

231: \begin{figure}

232: \centering

233: \includegraphics [scale=1] {energy1.eps}

234: \caption {Raw PXR spectra at 50, 75, and 100~keV electron energy}

235: \label{Fig3}

236: \end{figure}

237: 

238: \begin{figure}

239: \centering

240: \includegraphics [scale=1] {reflex1.eps}

241: \caption {Raw PXR spectra at 50, 75, and 100~keV electron energy}

242: \label{Fig4}

243: \end{figure}

244: 

245: 

246: After more detailed PXR spectra examinations we can attribute

247: peaks to corresponding crystallographic reflexes,

248: Fig.{\ref{Fig4}}. Most intensive peak in 2.5-2.9~keV region is

249: related to (220) reflexes with various indexes permutations.

250: Reflex (131) can correspond to peak at 3.5~keV. Region 4.9-5.2~keV

251: can be attributed to (331) and (511) including permutations.

252: Finally, region 5.4-6.0~keV is corresponding to (800) and (440)

253: reflexes.

254: 

255: \begin{figure}

256: \centering

257: \includegraphics [scale=1] {angle1.eps}

258: \caption {PXR frequency tuning with crystal target rotation}

259: \label{Fig5}

260: \end{figure}

261: 

262: 

263: 

264: As followed from Eq.(\ref{1}), PXR frequency must depend on beam

265: incidence angle. Experimentally observed frequency sifts are shown

266: in Fig.{\ref{Fig5}}. Basically, peaks should be splitted and

267: distance must increase depending on incidence angle. At 3~degrees

268: incidence angle split should be about 0.2~keV, at 10~degrees it

269: may be as high as 0.7~keV. One can see this split in some spectra.

270: Changes in spectra connected with incidence angle change are close

271: to calculated values.

272: 

273: 

274: \section{CONCLUSION}

275: 

276: Series of PXR spectra in various conditions were recorded and

277: their distinctive features were observed. Tuning of radiation

278: frequency with crystal-target rotation was observed for the first

279: time for low energy electrons. Dependence of the x-ray frequency

280: on the beam energy was detected. Despite of their relatively low

281: quantum yield, they can be considered as prospective source for

282: structure analysis and crystallography.

283: 

284: 

285: We have obtained and used big amount of targets made of various materials.

286: Spectra of acceptable quality were measured with only a few thin Si

287: membranes. Expecting application of PXR of non-relativistic electrons as

288: base for a source of tunable x-rays, problem of thin single crystal membrane

289: target must be a matter of high-tech challenge.

290: 

291: These results are obtained in the framework of ISTC project

292: {\#}B626.

293: 

294: \begin{thebibliography}{}

295: 

296: \bibitem{1} I.D. Feranchuk, A. Ulyanenkov, J. Harada, J.C.H. Spence //

297: Phys. Rev. E, v. 62, No3 (2000) 4225.

298: 

299: \bibitem{2} V.G.Baryshevsky, I.D. Feranchuk // J. de Phys. (Paris), v.

300: 44, (1983) 913.

301: 

302: \bibitem{3} Yu.S. Korobochko, V.F. Kosmach, V.I. Mineev

303: // Sov. JETP, v. 21 (1965) 834.

304: 

305: \bibitem{4} G.M. Reese, J.C.H. Spence, N. Yamamoto // Philos. Mag. A, v.

306: 49, No5 (1984) 697.

307: 

308: \bibitem{5} K.S. Vecchio, D.B. Williams // Journ. of Microscopy, v. 147,

309: pt. 1 (1987) 15.

310: \end{thebibliography}

311: 

312: \end{document}

313: