1: \documentstyle[aps,epsfig]{revtex}
2:
3: \makeatletter
4:
5: \begin{document}
6:
7: {\par\centering \textbf{\Large Angular Correlation in Double Photoionization
8: of Atoms and the Role of the Detection Process }\Large \par}
9:
10: \vspace{0.15in}
11: {\par\centering \textbf{\large Dipankar Chattarji and Chiranjib Sur }\large \par}
12: \vspace{0.1in}
13:
14: {\par\centering Department of Physics, Visva-Bharati, Santiniketan 731 235, INDIA\emph{ }\par}
15: \vspace{0.2in}
16:
17: {\small The problem of angular correlation in the double photoionization (DPI)
18: of rare gas atoms is considered in some depth. We refer particularly to the
19: efficiency operator for the detection of an electron by a detector having cylindrical
20: symmetry. The different factors in the efficiency operator are discussed in
21: detail keeping in mind the fundamental epistemological question of the role
22: of the detection process in such experiments.}{\small \par}
23:
24: \textbf{\small PACS No} {\small : 32.80.H, 32.80.F, 03.65.T,79.20.F }{\small \par}
25: \vspace{0.2in}
26:
27: In this paper we wish to consider the problem of angular correlation between
28: the two electrons emitted by an atom when it is doubly ionized by a photon.
29:
30: Consider a randomly oriented rare gas atom in a \( ^{1}S^{e} \) state. The
31: atom absorbs an unpolarized photon and after a certain time interval emits a
32: photo-electron from one of the inner shells giving a singly ionized atomic state.
33: This intermediate ionic state now de-excites by emitting an Auger electron,
34: typically from an outer shell, giving rise to a two-vacancy final atomic state.
35: We can denote this sequence of events as follows.
36:
37: \begin{equation}
38: \label{one}
39: h\nu +\mathbf{A}\longrightarrow \mathbf{A}^{+}+e_{1}^{-}\longrightarrow \mathbf{A}^{++}+e_{1}^{-}+e_{2}^{-}.
40: \end{equation}
41: A polar plot of the observed distribution of coincidences between the two emitted
42: electrons as a function of the angle~ between their directions of emission shows
43: a clear periodic behaviour {[}\ref{schmidt}{]}. The question we ask is: what
44: is the origin of this angular correlation? Could it have anything to do at all
45: with the process of detecting the electrons? On the face of it, this last
46: question may not seem so obvious. It will, however, become clearer as we proceed
47: with our discussion.
48:
49: Double photoionization (DPI) occurs when an atomic target like the one described
50: above is irradiated with a monochromatic photon beam from an advanced light
51: source, e.g. a synchrotron. Along with single photoionization (PI), there may
52: be events in which~ two electrons are emitted by an atom in quick succession.
53: In case the time interval between the successive emission of the two electrons
54: is substantially longer than the time taken by the first electron to leave the
55: interaction zone, DPI may be regarded as a two-step process {[}\ref{we1}{]}.
56: In other words, the emission of the two electrons may be regarded as being clearly
57: separated in time. This in its turn will depend on the energy imparted to the
58: atom by the incident photon.
59:
60: We wish to obtain an angular correlation function for the two emitted electrons
61: in terms of the angle between their directions of emission. We shall do this
62: by considering an ensemble of such atomic systems belonging to all possible
63: quantum mechanical states \( Q \). Each state \( Q \) is labeled by the total
64: angular momentum \( J \), its projection \( M \), and the remaining set of
65: quantum numbers \( \alpha \).
66:
67: The angular correlation function \( W(\theta ) \) for the two emitted electrons
68: is the probability that the angle between their directions of emission is \( \theta \).
69: Evidently this is a statistical quantity, and \( W(\theta ) \) would have to
70: be the ensemble average of the above probability.
71:
72: Now, how do we determine this probability? Hopefully, we let the atomic system
73: attain the final state given in Eq.(\ref{one}), we set up two detectors at
74: a suitable distance from the reaction zone with their axes making an angle \( \theta ^{\prime } \)
75: with each other, and we try to detect coincidences between the photo- and Auger
76: electrons. The number of coincidences we can hope to detect will depend on two
77: distinct factors.
78:
79: (i) There is a certain probability for the atomic system to attain the final
80: state. This is described by the appropriate matrix element of the density or
81: statistical operator \( \rho \) {[}\ref{fano},\ref{blum}{]}.
82:
83: (ii) Even if the system goes over to the final state, because of the finite
84: size of the detectors and other important limiting factors, a coincidence event
85: may or may not be detected. There is thus a finite probability \( \epsilon \, (0\leq \epsilon \leq 1) \)
86: that the event will be detected by the system of detectors. This probability is represented by the efficiency
87: operator \( \varepsilon \). It will depend on the size, position and geometrical
88: configuration of the detectors, but not their internal physical or chemical
89: nature provided that there is full absorption of an electron within the material
90: of a detector {[}\ref{rose}{]}.
91:
92: Obviously, \( W(\theta ) \) will be given by the joint probability of the formation
93: of the final state and its detection by the system of detectors, i.e. by the
94: product of \( \rho \) and \( \varepsilon \). For a given state \( Q \)
95: this joint probability will be \( \rho \varepsilon \) . The average probability
96: \( \overline{\varepsilon } \) for the ensemble will be given by the trace of
97: the product matrix. We write
98:
99: \begin{equation}
100: \label{two}
101: \begin{array}{cc}
102: \overline{\varepsilon } & =\sum _{Q}\varepsilon _{Q}\left\langle Q\right| \rho \left| Q\right\rangle \\
103: & =\sum _{Q}\varepsilon _{Q}\rho _{QQ}\\
104: & =Tr(\varepsilon \rho )\\
105: & =Tr(\rho \varepsilon )\, .
106: \end{array}
107: \end{equation}
108: Here \( \varepsilon _{Q} \) is the efficiency or probability of detection
109: of the state described by the quantum numbers \( Q \), and \( \rho _{QQ} \)
110: the probability of the system being in the particular state \( Q \) {[}\ref{coester}{]}.
111: Both \( \varepsilon \) and \( \rho \) are tensor operators.
112:
113: Since the angular correlation function happens to be the trace of a matrix {[}\ref{coester},\ref{ferguson}{]},
114: it will be invariant under a unitary transformation in Hilbert space.
115:
116: Another property of the system arises from the random orientation of the rare
117: gas atoms. The electrons emitted by them are unpolarized. And we take the detectors
118: to be insensitive to electron polarization. Hence the angular correlation function
119: itself can depend only on scalar invariants formed of the unit momentum vectors
120: of the two emitted electrons \( \widehat{\mathbf{p}_{1}} \) and \( \widehat{\mathbf{p}_{2}} \).
121: These invariants are given by the scalar product of spherical tensors {[}\ref{satchler}{]},
122: as follows:
123:
124:
125: \begin{equation}
126: \label{three}
127: \begin{array}{cc}
128: \mathbf{C}_{k}(\widehat{\mathbf{p}_{1}})\cdot \mathbf{C}_{k}(\widehat{\mathbf{p}_{2}}) & =\sum _{m}C_{km}(\widehat{\mathbf{p}_{1}})C^{\star }_{km}(\widehat{\mathbf{p}_{2}})\\
129: & =P_{k}(\widehat{\mathbf{p}_{1}}\cdot \widehat{\mathbf{p}_{2}})=P_{k}(cos\theta )\, .
130: \end{array}
131: \end{equation}
132: In Eq.(\ref{three}) \( P_{k}(cos\theta ) \) is a Legendre polynomial. The
133: index \( k \) will be restricted to the allowed values of the resultant of
134: the angular momenta \( j_{1} \) and \( j_{2} \) of the two emitted electrons.
135: Out of these, odd values of \( k \) will drop out because they would give odd
136: parity.
137:
138: Now, from the elements of statistical mechanics, we know that \( \overline{\varepsilon } \)
139: is the expectation value (or average value) of the efficiency operator \( \varepsilon \)
140: {[}\ref{terharr}{]}. It needs to be pointed out that this expectation value
141: is a function of \( \theta \). Thus the angular correlation function \( W(\theta ) \)
142: is, to within a multiplying factor, just the expectation value of the efficiency
143: operator for a given value of \( \theta \). To be more precise, it represents
144: the angle-dependent factor in the expectation value~ of the efficiency operator
145: for the entire detecting system.
146:
147: Now, based on physical considerations, can we find an expression for the efficiency
148: operator?
149:
150: We begin by noting that the efficiency operator for a single electron represents
151: the attenuation of the probability of detecting the signal
152: caused by certain geometrical properties of the detector as well as certain
153: intrinsic limitations of the detection process to be discussed later. Let us
154: now try to write down an expression for the efficiency operator for a single
155: detector in detecting an electron in terms of such factors.
156:
157: Obviously, a co-ordinate representation would be the most appropriate for the
158: discussion of these factors. But what kind of co-ordinate system shall we use?
159:
160: Because of the spherical symmetry of the system in the interaction region, we
161: use spherical polar co-ordinates for our calculation {[}\ref{we2}{]}. However,
162: as soon as an emitted electron begins to interact with a detector, it will acquire
163: a symmetry appropriate to the detector. In the present paper our objective is
164: to examine the geometrical properties of the electron detector and to see how
165: they affect its efficiency. So let us see what kind of symmetry exists in the
166: detection region.
167:
168: Let us start by considering our options with regard to the shape and size of
169: the detector. Could we, for example, start with the limiting case of a point
170: detector {[}\ref{ferguson32}{]}? By considering the signal to noise ratio,
171: it may be shown that one must use a finite size detector. In order to find the
172: direction of emission of an electron it must have axial symmetry. Furthermore,
173: the photo-electron as well as the Auger electron is characterised by a well
174: defined energy and a well defined orbital angular momentum. The well defined
175: energy implies that we need an electron spectrometer which is a differential
176: energy analyser. And the well defined orbital angular momentum indicates that
177: the spectrometer should be an angle integrated device. The cylindrical mirror
178: analyser (CMA) meets all these requirements {[}\ref{URL},\ref{dc}{]}. Hence
179: it seems to be the obvious candidate. Recent angular correlation measurements
180: {[}\ref{schmidt2}{]} using the CMA clearly support this view.
181:
182: A CMA receives incident electrons through a circular aperture of finite radius
183: \( r \). Hence the angle of incidence \( \beta _{i} \) of an electron as measured
184: with respect to the cylinder axis varies from \( 0 \) to a small finite value.
185: Here the index \( i=1 \) for photo-electrons and \( 2 \) for Auger electrons.
186: But, whatever its
187: actual value, because of the angle integrated character of the CMA the observer
188: has no way of knowing \( \beta _{i} \). He identifies the direction of emission
189: of the electron with the axis of the CMA.
190:
191: As a result of the identification of the direction of emission of the electron
192: with the cylinder axis there is an effective rotation \( \Re _{i}(=0\beta _{i}0) \)
193: of its direction of emission. Thus the efficiency operator representing the detection process
194: will contain a rotation matrix \( D_{\kappa ^{\prime }_{i}\kappa _{i}}^{k_{i}}(\Re _{i}) \).
195: This rotation matrix will be a factor in the expression for the efficiency operator of a detector.
196:
197: In our paper under reference {[}\ref{we2}{]} we have shown that the expression
198: for the efficiency operator of a single detector contains a rotation matrix
199: element. It will be seen from the formal derivation given there that the angle-dependence
200: of the angular correlation function arises directly from this rotation
201: matrix.
202:
203: Note that the angle integrated character of the electron spectrometer is a requirement
204: imposed by the fact that it has to receive electrons with a well defined orbital
205: angular momentum. In other words, the complete indeterminacy in the direction
206: of emission of an electron is a dynamical requirement and not a matter of technical
207: deficiency of the detector. Since the rotation matrix in the expression for
208: the angular correlation function originates from this indeterminacy, evidently
209: the angular correlation arises directly from it.
210:
211: This takes care of the most important factor in the expression for the efficiency
212: operator, namely the rotation matrix element. We shall call this the attenuation
213: factor due to rotation. Before we go on to the other factors, we take a brief
214: look at the geometrical arrangement of the detectors.
215:
216: In Fig.1, \( A_{1} \) represents the axis of the detector set up to detect the photo-electron,
217: and \( A_{2} \) the axis of the detector receiving the Auger electron.
218: The angle between the two axes is \( \theta ^{\prime } \). The directions
219: \( D_{1} \) and \( D_{2} \) are the actual directions of emission of the photo-
220: and Auger electrons respectively. The angle \(\beta_{1} \) is the angle between \(A_{1}\) and \(D_{1}\), i.e. the angle of incidence of the photo-electron as defined above. Similarly, \(\beta_{2}\), the angle between \(A_{2}\) and \(D_{2}\), is the angle of incidence of the Auger electron.
221:
222: Each CMA has a circular aperture of radius \(r\) for receiving electrons. In a DPI experiment the base of each of the two CMAs used to detect the photo- and Auger electrons is placed at a distance \(h\) from the centre of the reaction zone.
223: The angular width of the aperture in each CMA as seen from the centre of
224: the target is \( 2\gamma \), where \( tan\, \gamma =\frac{r}{h}\, (\sim .01) \).
225:
226: We can now go back to the form of the efficiency operator corresponding to a
227: single detector detecting an electron. It is a tensor operator of rank \( k \)
228: with \( (2k+1) \) components. Following reference {[}\ref{we2}{]} we write
229: a reduced matrix element of the component labeled by \( \kappa \) of the tensor
230: as
231:
232: \begin{equation}
233: \label{four}
234: \varepsilon _{k\kappa }(jj^{\prime })=\sum _{\kappa ^{\prime }}z_{k}c_{k\kappa ^{\prime }}(jj^{\prime })D_{\kappa \kappa ^{\prime }}^{k}(\Re )\, .
235: \end{equation}
236: Here \( \kappa \) and \( \kappa ^{\prime } \) are projection quantum numbers
237: corresponding to the angular momentum \( k \). \( z_{k} \) is the attenuation
238: factor due to the finite size of a detector. It is different for different values
239: of \( k \). The factor \( c_{k\kappa ^{\prime }}(jj^{\prime }) \) arises from
240: the change of symmetry as the electron goes to the detection zone from the reaction
241: zone. Let us now look at these factors.
242:
243: (a) \emph{Attenuation due to absorption in a detector of finite size}. Let us
244: first consider the case of a single detector detecting, say, a photo-electron.
245: Here we can think in terms of an angular distribution measurement. The angular
246: distribution~ too can be written out as a Legendre polynomial expansion. The
247: attenuation factor multiplying \( P_{n}(cos\beta ) \) will be {[}\ref{rose}{]},
248:
249: \begin{equation}
250: \label{five}
251: z_{n}=\frac{J_{n}}{J_{0}},
252: \end{equation}
253: where
254:
255:
256: \begin{equation}
257: \label{six}
258: J_{n}=\int ^{\gamma }_{o}P_{n}(cos\beta )sin\beta d\beta \, .
259: \end{equation}
260:
261:
262: For an angular correlation experiment with two detectors having finite size
263: attenuation factors \( z_{n}(1) \) and \( z_{n}(2) \) the total attenuation
264: factor for the \( n \)th term will be
265:
266:
267: \begin{equation}
268: \label{seven}
269: Z_{n}=z_{n}(1)z_{n}(2)\, .
270: \end{equation}
271:
272:
273: (b) \emph{Attenuation factor corresponding to the state of polarization}. We
274: have already discussed the axial symmetry acquired by unpolarized electrons
275: as they enter the detection zone. What does this do to their quantum mechanical
276: state?
277:
278: Let us first consider the semi-classical vector model. Axial symmetry about
279: the detector axis implies that the angular momentum vector of an electron can
280: only lie in the plane perpendicular to that axis, i.e. the \( xy \) plane.
281: Obviously, the \( z \)-component of its angular momentum will be zero. Now
282: going over to quantum mechanics, only those states will survive for which the projection quantum number \( \nu =0 \).
283: This calls for a projection operator having the form
284:
285: \begin{equation}
286: \label{ten}
287: c_{n\nu }(jj^{\prime })=N_{jj^{\prime }n}C^{jj^{\prime }n}_{\frac{1}{2}-\frac{1}{2}\nu}\delta _{\nu 0}\, ,
288: \end{equation}
289: where \( N_{jj^{\prime }n} \) is a normalizing factor which turns out to be
290: \( \frac{\sqrt{2j+1}\sqrt{2j^{\prime }+1}}{4\pi }(-1)^{j-\frac{1}{2}+n} \)
291: {[}\ref{we2}{]}. In other words, only the factor \( c_{n0}(jj^{\prime }) \)
292: enters into our expression for the efficiency operator.
293:
294: A formal derivation of this result is given in our paper under reference {[}\ref{we2}{]}.
295: However, that derivation does not quite relate to the attenuation properties
296: of a detector. On the other hand, we feel that our present approach is physically
297: more transparent. It also throws some light on a couple of questions of fundamental
298: epistemological interest. Does the detection process have a role in this type
299: of experiment? If so, what is that role like? Obviously, such questions can
300: be important from the standpoint of the theory of measurement.
301:
302: Calculation of the angular correlation function can now go through as in reference
303: {[}\ref{we2}{]}. We finally get
304:
305:
306: \begin{equation}
307: \label{11}
308: \begin{array}{ccc}
309: W(\theta ) & = & \sum _{k}z_{k}(1)z_{k}(2)(-1)^{j_{1}+j_{2}}c_{k0}(j_{1}j^{\prime }_{1})c^{\star }_{k0}(j_{2}j^{\prime }_{2})\\
310: & & \times w(J_{b}J^{\prime }_{b}j_{1}j^{\prime }_{1},kJ_{a})w(J_{b}J^{\prime }_{b}j_{2}j^{\prime }_{2},kJ_{c})P_{k}(cos\theta )\, ,
311: \end{array}
312: \end{equation}
313: where the \( w \)'s are Racah coefficients, \( j_{1} \) and \( j_{2} \)
314: are the angular momenta of the photo- and Auger electrons, \( J_{a},J_{b},J_{c} \)
315: the angular momenta of the atom in its initial, intermediate and final states
316: respectively. Here \( k \) is an even integer ranging from \( 0 \) to \( k_{max} \),
317: \( k_{max} \) being defined as follows. Let \( \left\{ \left\{ j_{1}+j^{\prime }_{1}\right\} _{max},\left\{ j_{2}+j^{\prime }_{2}\right\} _{max}\right\} _{min}=p \).
318: Then
319:
320: \begin{equation}
321: \label{12}
322: \begin{array}{ccc}
323: k_{max} & = & p\, \, if\, p\, is\, even,\\
324: & = & (p-1)\, if\, p\, is\, odd.
325: \end{array}
326: \end{equation}
327: The set of primed angular momentum quantum numbers represent virtual states
328: which may arise from possible interaction with other atoms and electrons.
329:
330: To sum up, our main finding in this paper is that the angular correlation function
331: \( W(\theta ) \) arises directly from an unavoidable indeterminacy in the actual
332: directions of emission of the two electrons. From our analysis above, it should
333: be clear that this is not a matter of technical imperfection but a basic
334: restriction imposed on the detection process by the dynamical nature of the
335: problem.
336:
337: Our results for the double photoionization of xenon are discussed in reference
338: {[}\ref{we2}{]} in some detail.
339:
340: One of the authors (CS) is indebted to the University Grants Commission of India
341: for support in the form of a junior research fellowship.
342:
343: \begin{thebibliography}{}
344: \bibitem{1}\label{schmidt}B. K\( \ddot{a} \)mmerling and V. Schmidt, J.Phys.B \textbf{26},
345: 1141(1993).
346: \bibitem{2}\label{we1}D. Chattarji and C. Sur, J. of Electron Spect. and Rel. Phen. \textbf{114-116},
347: 153(2001).
348: \bibitem{3}\label{fano}U. Fano, Phys. Rev. \textbf{90}, 577(1953).
349: \bibitem{4}\label{blum}K. Blum, Density Matrix Theory and Applications (Plenum Press,
350: New York, 1981).
351: \bibitem{5}\label{rose}M. E. Rose, Phys. Rev. \textbf{91}, 610(1953).
352: \bibitem{6}\label{coester}F. Coester and J. M. Jauch, Helv. Phys. Acta \textbf{26}, 3(1953).
353: \bibitem{7}\label{ferguson}A. J. Ferguson, Angular Correlation Methods in Gamma-ray Spectroscopy
354: (North-Holland, Amsterdam, 1965).
355: \bibitem{8}\label{satchler}D. M. Brink and G. R. Satchler, Angular Momentum, 2nd ed.,
356: 55(Oxford University Press, Oxford, 1968).
357: \bibitem{9}\label{terharr}D. ter Haar, Elements of Statistical Mechanics, 150 (Holt, Reinhart
358: and Winston, New York, 1960).
359: \bibitem{10}\label{we2}D. Chattarji and C. Sur, \texttt{\textbf{physics/0106095}} (To be
360: published in Phys. Rev. A).
361: \bibitem{8}\label{ferguson32}A. J. Ferguson, Angular Correlation Methods in Gamma-ray
362: Spectroscopy, 32(North-Holland, Amsterdam, 1965).
363: \bibitem{12}\label{dc}D. Chattarji, The Theory of Auger Transitions, 234(Academic Press,
364: London, 1976).
365: \bibitem{9}\label{schmidt2}H. Darenbach et al, Nucl. Inst. Meth. \textbf{A 260}, 258 (1987).
366: \bibitem{10}\label{URL}http://srs.dl.ac.uk/SSL/equipment/CMA
367: \end{thebibliography}
368: \vspace{0.3cm}
369: {\par\centering \resizebox*{!}{0.5\textheight}{\rotatebox{-90}{\includegraphics{detector.ps}}} \par}
370: \vspace{0.3cm}
371:
372: {\par\centering Fig 1 : Geometrical arrangement of the detectors\par}
373:
374: \end{document}
375: