1: \documentstyle[psfig,aps,prl,amssymb]{revtex}
2:
3: \begin{document}
4: \title{Bures Geometry of the
5: Three-Level Quantum Systems. I}
6: \author{Paul B. Slater}
7: \address{ISBER, University of California, Santa Barbara, CA 93106-2150\\
8: e-mail: slater@itp.ucsb.edu, FAX: (805) 893-7995}
9:
10: \date{\today}
11:
12: \draft
13:
14: \maketitle
15:
16: \vskip -0.1cm
17:
18: \begin{abstract}
19: We compute --- using a formula of Dittmann --- the Bures metric tensor
20: ($g$) for the eight-dimensional state space of three-level quantum systems,
21: employing a newly-developed Euler angle-based
22: parameterization of the $3 \times 3$
23: density matrices. Most of the individual metric elements ($g_{ij}$) are
24: found to be expressible in relatively compact form, many of them in fact
25: being exactly {\it zero}.
26: \end{abstract}
27:
28: \vspace{.2cm}
29: \hspace{1.5cm} Keywords: Bures metric, three-level quantum systems, spin-1
30: systems, density matrices,
31:
32: \hspace{1.5cm} orthogonal parameters, Euler angles, unitary transformations
33:
34: \vspace{.15cm}
35:
36: \hspace{1.5cm} Mathematics Subject Classification (2000): 81Q70, 53Axx
37: \pacs{PACS Numbers 03.65.Bz, 02.40.Ky}
38:
39: \tableofcontents
40:
41: \vspace{.1cm}
42: \section{INTRODUCTION}
43:
44: The Bures metric is a distinguished member --- the {\it minimal}
45: one --- of the
46: (nondenumerable)
47: family of {\it monotone} metrics on the quantum systems \cite{petzsudar}.
48: Its contemporary
49: study was pioneered by Armin Uhlmann \cite{uhl1,uhl2}, along with several
50: of his associates at the University of Leipzig
51: \cite{hub1,ditt1,ditt2,ditt3,ditt4}.
52: In particular, Jochen Dittmann has derived several {\it explicit} formulas
53: (ones not requiring knowledge of the eigenvalues of density matrices) for the
54: Bures metric \cite{ditt3,ditt4}. Slater \cite{slat1} --- interpreting
55: the {\it volume element}
56: of the metric as a natural (unnormalized) measure on the
57: quantum systems --- applied this work to
58: certain low-dimensional subsets of the fifteen-dimensional set of
59: $4 \times 4$ density matrices to obtain ``exact Bures probabilities that
60: two quantum bits are classically correlated'' (cf. \cite{SLAT,ZHSL}).
61:
62: The Bures metric on the three-dimensional convex set of the $2 \times 2$
63: density matrices (making use of Cartesian coordinates ($x,y,z$)),
64: \begin{equation} \label{2by2}
65: \rho ={1 \over 2}
66: \pmatrix{1 + z & x + i y \cr x - i y & 1-z \cr}, \qquad (0 \leq x^2+y^2+z^2
67: \leq 1)
68: \end{equation}
69: has been intensively studied.
70: The corresponding metric tensor
71: \begin{equation} \label{hhe}
72: g = {1 \over 4 (1-x^2-y^2-z^2)} \pmatrix{1-y^2-z^2 & x y & x z \cr
73: x y & 1-x^2 -z^2 & yz \cr
74: x z & y z & 1-x^2-y^2 \cr}
75: \end{equation}
76: can be obtained by application of an (early) formula of Dittmann
77: \cite[eq. (3.7)]{ditt3},
78: \begin{equation} \label{fu}
79: d_{Bures}(\rho,\rho+ \mbox{d} \rho)^{2} = {1 \over 4} \mbox{Tr} \lbrace
80: \mbox{d} \rho \mbox{d} \rho + {1 \over |\rho|} (\mbox{d} \rho -\rho
81: \mbox{d} \rho) (\mbox{d} \rho - \rho \mbox{d} \rho) \rbrace.
82: \end{equation}
83: In spherical coordinates ($r,\theta,\phi$) the tensor (\ref{hhe})
84: takes a {\it diagonal} form
85: \begin{equation} \label{DIAG}
86: g = {1 \over 4} \pmatrix{{1 \over (1-r^2)} & 0 & 0 \cr 0 & r^2 & 0 \cr
87: 0 & 0 & r^2 \sin^{2}{\theta} \cr}.
88: \end{equation}
89:
90: The Bures metric can be viewed as the
91: standard metric on the surface of a three-sphere
92: \cite{hub1,bm}. As such, Hall \cite[p. 128]{hall} has written
93: that ``the Bures metric for
94: a two-dimensional system corresponds to the surface of a unit four-ball,
95: \linebreak
96: i. e., to the maximally symmetric three-dimensional space of positive
97: curvature (and may be recognized as the spatial part of the Robertson-Walker
98: metric in general relativity). This space is homogeneous and isotropic, and hence the Bures metric does not distinguish a preferred location or direction in
99: the space of density operators. Indeed, as well as rotational symmetry in
100: Bloch coordinates (corresponding to unitary invariance), the metric has a
101: further set of symmetries generated by the infinitesimal transformations
102: \begin{equation}
103: r \rightarrow r + \epsilon (1-r^2)^{1 \over 2} a,
104: \end{equation}
105: (where $a$ is an arbitrary three-vector [and $r$, radial distance in the
106: Bloch sphere of two-level quantum systems \cite{bm}]).''
107: Petz and Sud\'ar observed that in ``the case of the [Bures] metric
108: of the symmetric logarithmic derivative the tangential component is
109: independent of $r$'' \cite[p. 2667]{petzsudar}.
110:
111: A principal goal of the present study is to determine any such symmetries
112: possessed by the Bures metric when one proceeds from the study of the
113: two-level quantum systems to that of the three-level quantum systems.
114: One should be aware, though, that
115: Dittmann has noted that in this case, the space ``is not a space of constant
116: curvature and not even a locally symmetric space, in contrast to
117: what the case
118: of two-dimensional density matrices might suggest'' \cite{ditt3}.
119: (In a locally symmetric space, the sectional curvature is invariant under
120: parallel displacement, and the covariant derivative of the curvature
121: tensor field vanishes \cite{ditt1,helgason}. A formula for the scalar
122: curvature of the monotone metrics for general $n$-level quantum systems
123: is given in \cite{ditt11}, cf. \cite{ditt12}.) In other
124: work \cite{dittym}, Dittmann has shown that the gauge field
125: defining the Bures metric satisfies
126: the source-free Yang-Mills equation. Petz
127: \cite[Thm. 3.4]{petzmore} has established
128: that the Bures metric is the only monotone metric that is both
129: ``Fisher adjusted'' and ``Fubini-Study adjusted''.
130: \section{METHODOLOGY}
131:
132: Slater \cite{slat2} (cf. \cite[eqs. (6), (7)]{slat1}) applied a
133: formula (cf. (\ref{fu}))
134: of Dittmann \cite[eq. (3.8)]{ditt3} for the specific case
135: of the three-level quantum systems,
136: \begin{equation} \label{form1}
137: g^{B}_{\rho} = {1 \over 4} \mbox{Tr} \lbrace \mbox{d} \rho \mbox{d} \rho +
138: {3 \over 1 - \mbox{Tr} \rho^3}
139: (\mbox{d} \rho -\rho \mbox{d} \rho ) (\mbox{d} \rho -\rho
140: \mbox{d} \rho) +{3 |\rho| \over 1 - \mbox{Tr} \rho^{3} }
141: (\mbox{d} \rho - \rho^{-1} \mbox{d} \rho) (\mbox{d} \rho -\rho^{-1}
142: \mbox{d} \rho ) \rbrace,
143: \end{equation}
144: to the particular instance (a simple extension of the
145: two-level quantum systems (\ref{2by2})) of a {\it four}-dimensional subset,
146: \begin{equation}
147: \rho = {1 \over 2} \pmatrix{v+z & 0 & x- i y
148: \cr 0 & 2 -2 v & 0 \cr x + i y & 0 & v -z \cr},
149: \end{equation}
150: of the eight-dimensional convex set of $3 \times 3$ density matrices
151: \cite{bloore}. Now,
152: in the present study, we apply this same
153: formula (\ref{form1}) to the {\it full}
154: eight-dimensional convex set of the three-level quantum
155: systems itself. Of crucial
156: and central importance
157: here will be the use of a
158: recently-developed ``Euler angle'' parameterization of
159: these density matrices \cite{bb,us}. In this parameterization, one takes an
160: arbitrary density matrix ($\rho$) to be
161: expressed in the (``Schur/Schatten'') form \cite[sec. 3]{ditt4}
162: \cite[p. 3725]{twam}
163: \cite[p. 53]{hasegawa}
164: \begin{equation} \label{pte}
165: \rho = U \rho^{'} U^{\dagger}.
166: \end{equation}
167: Here
168: \begin{equation}
169: U= \mbox{e}^{i \lambda_{3} \alpha} \mbox{e}^{i \lambda_{2} \beta}
170: \mbox{e}^{i \lambda_{3} \gamma} \mbox{e}^{i \lambda_{5} \theta}
171: \mbox{e}^{i \lambda_{3} a} \mbox{e}^{i \lambda_{2} b},
172: \end{equation}
173: is a member of $SU(3)$, the three immediately relevant
174: (of the eight) Gell-Mann matrices \cite{po}
175: being
176: \begin{equation}
177: \lambda_{2} = \pmatrix{0 & - i & 0 \cr i & 0 & 0 \cr 0 & 0 & 0 \cr},\qquad
178: \lambda_{3} = \pmatrix{1 & 0 & 0 \cr
179: 0 & -1 & 0 \cr 0 & 0 & 0 \cr}, \qquad \lambda_{5} = \pmatrix{0 & 0 & -i \cr
180: 0 & 0 & 0 \cr i & 0 & 0 \cr}.
181: \end{equation}
182: Making use of spherical coordinates ($\theta_{1}, \theta_{2}$),
183: \begin{equation}
184: \rho^{'} = \pmatrix{\cos^2{\theta_{1}} & 0 & 0 \cr
185: 0 & \sin^{2}{\theta_{1}} \cos^{2}{\theta_{2}} & 0 \cr
186: 0 & 0 & \sin^{2}{\theta_{1}} \sin^{2}{\theta_{2}} \cr}.
187: \end{equation}
188: An appropriate set of ranges of the eight angles (by which all the
189: $3 \times 3$ density matrices can be reproduced without duplication) is
190: \cite[eqs. (11), (12)]{bb}
191: \begin{equation}
192: 0 \leq \alpha,
193: \gamma, a \leq \pi, \quad 0 \leq \beta,
194: \theta, b \leq {\pi \over 2}, \quad 0 \leq \theta_{1} \leq
195: \cos^{-1}{1 \over \sqrt{3}}, \quad 0 \leq \theta_{2} \leq {\pi \over 4}.
196: \end{equation}
197:
198: We have inserted the so-parameterized
199: $3 \times 3$ density matrix (\ref{pte}) into
200: formula (\ref{form1}) to obtain the $8 \times 8$ Bures metric tensor.
201: Since, by construction, we have explicit knowledge of the eigenvalues
202: ($\lambda$'s)
203: and eigenvectors of $\rho$, we could alternatively have directly
204: applied the {\it general} formula
205: for the Bures metric in the $n$-dimensional case \cite[eq. (10)]{hub1},
206: \begin{equation}
207: d_{Bures}(\rho,\rho + \mbox{d} \rho)^{2} =
208: {1 \over 2} \sum_{i,j =1}^{n} {|<i|\mbox{d} \rho|j>|^{2} \over \lambda_{i}
209: + \lambda_{j}},
210: \end{equation}
211: or that given by Proposition 4 in \cite{ditt4}.
212:
213: \section{ELEMENTS OF THE BURES METRIC TENSOR}
214:
215: Initially, all the entries of the tensor computed using
216: (\ref{form1}) --- implemented in MATHEMATICA --- were given
217: by extremely large
218: complicated expressions. However, in a number of cases, both through exact
219: computations and numerical experimentation, we were able to arrive
220: at certain relatively compact (if not simply strictly
221: {\it zero} themselves) expressions for the individual metric elements.
222:
223: The first remarkable item to note is that (as repeated numerical
224: experiments indicate) {\it all} the entries of the tensor
225: are {\it independent} of the Euler angle $\alpha$.
226: Further numerical investigations have convinced us that
227: many of the entries of the tensor are, in fact, zero (cf. \cite{tod,cox}).
228: (In the case of the two-level quantum systems, the off-diagonal
229: entries of the Bures metric tensor
230: (\ref{DIAG}) are zero, if spherical --- as
231: opposed to Cartesian --- coordinates
232: are employed.)
233: For example,
234: the
235: spherical coordinates $\theta_{1}$ and $\theta_{2}$ are both
236: orthogonal to the other seven coordinates. The diagonal entry
237: ($g_{\theta_{1} \theta_{1}}$) of the Bures
238: metric tensor ($g$)
239: corresponding to the pairing $(\theta_{1}, \theta_{1})$ is
240: simply 1, while the diagonal entry
241: ($g_{\theta_{2} \theta_{2}}$) corresponding to the pairing
242: $(\theta_{2},\theta_{2})$ is $\sin^{2}{\theta_{1}}$.
243:
244: Let us summarize our present state of
245: explicit knowledge regarding the
246: Bures metric elements ($g_{ij}$) for the three-level quantum systems.
247: We write the corresponding (symmetric) matrix, using the ordering
248: of coordinates (and hence rows and columns)
249: \begin{equation}
250: (\alpha, \gamma, a, \beta,b,\theta,\theta_{1},\theta_{2})
251: \end{equation}
252: as
253: \begin{equation} \label{gmatrix}
254: g =\pmatrix{? & ? & g_{13} & ? & g_{15} & g_{16} & 0 & 0 \cr
255: \cdot & g_{22} & g_{23} & g_{24} & 0 & 0 & 0 & 0 \cr
256: \cdot & \cdot & g_{33} & g_{34} & 0 & 0 & 0 & 0 \cr
257: \cdot & \cdot & \cdot & g_{44} & g_{45} & g_{46} & 0 & 0 \cr
258: \cdot & \cdot & \cdot & \cdot & g_{55} & 0 & 0 & 0 \cr
259: \cdot & \cdot & \cdot & \cdot & \cdot & g_{66} & 0 & 0 \cr
260: \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & 1 & 0 \cr
261: \cdot & \cdot & \cdot & \cdot & \cdot & \cdot & \cdot &
262: \sin^{2}{\theta_{1}} \cr}.
263: \end{equation}
264: Our specific element-by-element results are now presented.
265: \subsection{$g_{55} = g_{b b}$}
266: We have (Fig.~\ref{f1})
267: \begin{equation} \label{g55eq}
268: g_{55}=g_{b b} ={t^{2}
269: \over 16 u_{+}},
270: \end{equation}
271: where (cf. \cite[eq. (28)]{slat3})
272: \begin{equation}
273: t = 2 + 6 \cos{2 \theta_{1}} + \cos{2(\theta_{1}-\theta_{2})}
274: -2 \cos{2 \theta_{2}} +\cos{2 (\theta_{1}+\theta_{2})}
275: \end{equation}
276: and
277: \begin{equation}
278: u_{\pm}= 3 + \cos{2 \theta_{1}} \pm
279: 2 \cos{2 \theta_{2}} \sin^{2}{\theta_{1}}.
280: \end{equation}
281: \begin{figure}
282: \centerline{\psfig{figure=gbb.eps}}
283: \caption{Diagonal (5,5)-entry, corresponding to the Euler angle
284: $b$, of the Bures metric tensor (\ref{gmatrix}) for the three-level quantum
285: systems. This term --- which can be inverted (\ref{INVERT}) --- enters as
286: well
287: into many of the expressions for the other metric elements.}
288: \label{f1}
289: \end{figure}
290: \subsection{$g_{13}=g_{\alpha a}$}
291: \begin{equation}
292: g_{13} =g_{\alpha a} = {g_{55} \over 4} \lbrace (3 + \cos{2 \theta}) \cos{2 \beta}
293: \sin^{2}{2 b} + 2 \cos{2 (a + \gamma)} \cos{\theta} \sin{4 b} \sin{2 \beta}
294: \rbrace.
295: \end{equation}
296: \subsection{$g_{15} = g_{\alpha b}$}
297: \begin{equation}
298: g_{15} = g_{\alpha b} = g_{55} \cos{\theta} \sin{2 \beta}
299: \sin{2 (a + \gamma)}.
300: \end{equation}
301: \subsection{$g_{16}=g_{\alpha \theta}$}
302: \begin{equation}
303: g_{16} = g_{\alpha \theta} = {v \over 32 u_{-}} \sin{2 b} \sin{2 \beta}
304: \sin{2 (a + \gamma)} \sin{\theta},
305: \end{equation}
306: where
307: \begin{equation}
308: v= 15 + 28 \cos{2 \theta_{1}} + 21 \cos{4 \theta_{1}} +
309: 4 (7 + 9 \cos{2 \theta_{1}}) \cos{2 \theta_{2}} \sin^{2}{\theta_{1}} \quad -
310: \end{equation}
311: \begin{displaymath}
312: - \quad
313: 4 (5 + 3 \cos{2 \theta_{1}}) \cos{4 \theta_{2}} \sin^{2}{\theta_{1}} + 8
314: \cos{6 \theta_{2}} \sin^{4}{\theta_{1}} .
315: \end{displaymath}
316: In Fig.~\ref{kdy} we plot the Euler angle-independent part of $g_{16}$,
317: that is ${v \over 32 u_{-}}$.
318: \begin{figure}
319: \centerline{\psfig{figure=g16.eps}}
320: \caption{Euler angle-independent factor of metric element corresponding to
321: the (1,6)-entry of (\ref{gmatrix})}
322: \label{kdy}
323: \end{figure}
324: \subsection{$g_{22} = g_{\gamma \gamma}$}
325: \begin{equation}
326: g_{22} =g_{\gamma \gamma} = {1 \over 16 \kappa} \lbrace -g_{55} \kappa
327: \cos^{4}{b} (3 + \cos{2 \theta})^{2} + 4 \cos^{2}{b} (g_{55} \kappa +
328: \mu \cos^{2}{\theta} \quad +
329: \end{equation}
330: \begin{displaymath}
331: + \quad 4 (\kappa + \upsilon) \cos^{2}{2 \theta_{2}}
332: \cos^{4}{\theta} \sin^{2}{\theta_{1}}) + 16 \kappa \cos^{2}{2 \theta_{2}}
333: \cos^{2}{\theta} \sin^{2}{\theta_{1}} \sin^{2}{\theta} \rbrace,
334: \end{displaymath}
335: where
336: \begin{equation}
337: \kappa = 35 + 28 \cos{2 \theta_{1}} + \cos{4 \theta_{1}} -
338: 8 \cos{4 \theta_{2}} \sin^{4}{\theta_{1}}
339: \end{equation}
340: \begin{equation}
341: \upsilon = -4 (1 +3 \cos{2 \theta_{1}}) (7 +5 \cos{2 \theta_{1}})
342: \sec{2 \theta_{2}} -16 \cos{2 \theta_{2}} \sin^{4}{\theta_{1}},
343: \end{equation}
344: and
345: \begin{equation}
346: \mu= -\sin^{2}{\theta_{1}} \lbrace 1621 + 125 \cos{2 \theta_{2}} + 46 \cos{4 \theta_{2}} + 4 \cos{2 \theta_{1}} (261 + 49 \cos{2 \theta_{2}} + 10 \cos{4 \theta_{2}}) \quad +
347: \end{equation}
348: \begin{displaymath}
349: + \quad \cos{4 \theta_{1}} (151 + 63 \cos{2 \theta_{2}} + 42 \cos{4 \theta_{2}})
350: -768 \csc^{2}{\theta_{1}} + 8 (\cos{6 \theta_{2}} -\cos{8 \theta_{2}})
351: \sin^{4}{\theta_{1}} \rbrace.
352: \end{displaymath}
353: \subsection{$g_{23} = g_{\gamma a}$}
354: \begin{equation} \label{g23}
355: g_{23} =g_{\gamma a} = {g_{55} \over 4 } (3 +\cos{2 \theta}) \sin^{2}{2 b}.
356: \end{equation}
357: \subsection{$g_{24}= g_{\gamma \beta}$}
358: \begin{equation}
359: g_{24}= g_{\gamma \beta} = {3 t \cos{\theta} \sin{2 b} \sin{2 (a + \gamma)}
360: \sin^{2}{\theta_{1}} (\cos{2 \theta} - i \sin{2 \theta}) \over
361: 256 (-1 +\cos^{6}{\theta_{1}} +\sin^{6}{\theta_{1}} (\cos^{6}{\theta_{2}}
362: +\sin^{6}{\theta_{2}}))}
363: \end{equation}
364: \begin{displaymath}
365: \cos^{2}{\theta_{1}} (-p \cos{2 b} (-1 + 4 \cos{2 \theta_{2}} +
366: \cos{4 \theta_{2}}) +q (-1 + 7 \cos{2 \theta_{2}} - 3 \cos{4 \theta_{2}}
367: +\cos{6 \theta_{2}})) \qquad +
368: \end{displaymath}
369: \begin{displaymath}
370: +\quad 2 \cos^{4}{\theta_{1}} (p \cos{2 b} \cos^{2}{\theta_{2}}
371: (3 + \cos{2 \theta_{2}}) + q (4 -3 \cos{2 \theta_{2}} + \cos{4 \theta_{2}})
372: \sin^{2}{\theta_{2}}) + ( - p \cos{2 b} + q (-1 +2 \cos{2 \theta_{2}}))
373: \sin^{2}{2 \theta_{2}},
374: \end{displaymath}
375: where
376: \begin{displaymath}
377: p= 1 + 6 e^{2 i \theta} +e^{4 i \theta}, \quad q= (-1+e^{2 i \theta})^{2}.
378: \end{displaymath}
379: \subsection{$g_{33}=g_{a a}$}
380: \begin{equation}
381: g_{33} = g_{a a} = g_{55} \sin^{2} {2 b} .
382: \end{equation}
383: \subsection{$g_{34} = g_{a \beta}$}
384: \begin{equation}
385: g_{34} =g_{a \beta}= -{g_{55} \over 2 } \cos{\theta}
386: \sin{4 b} \sin{2 (a + \gamma)}.
387: \end{equation}
388: \subsection{$g_{44} = g_{\beta \beta}$}
389: \begin{equation}
390: g_{44} =g_{\beta \beta} = - g_{55} \cos^{2}{b} \cos^{2}{\theta} \sin^{2}{b}
391: \sin^{2}{2(a + \gamma)} + {\zeta \over 32 \kappa},
392: \end{equation}
393: where
394: \begin{equation}
395: \zeta = -\csc^{2}{\theta_{1}} \lbrace -101 +
396: 12 \cos{4 \theta_{1}} + 64 \cos{6 \theta_{1}}
397: +25 \cos{8 \theta_{1}} + 16 (61 + 100 \cos{2 \theta_{1}} + 31 \cos{4 \theta_{1}})
398: \end{equation}
399: \begin{displaymath}
400: \cos{2 \theta_{2}} \cos{2 \theta} \sin^{4}{\theta_{1}} - 64 (5 + 7 \cos{2 \theta_{1}}) \cos{4 \theta_{2}} \sin^{6}{\theta_{1}} \quad +
401: \end{displaymath}
402: \begin{displaymath}
403: + \quad 128 \cos{2 \theta_{2}} \cos{4 \theta_{2}} \cos{2 \theta}
404: \sin^{8}{\theta_{1}} + 2 \cos^{2}{b} \sin^{2}{\theta_{1}}
405: ( 242 + 445 \cos{2 \theta_{1}} + 286
406: \cos{4 \theta_{1}} \quad +
407: \end{displaymath}
408: \begin{displaymath}
409: \quad 51 \cos{6 \theta_{1}} + 4 ((125 + 196 \cos{2 \theta_{1}}
410: + 63 \cos{4 \theta_{1}}) \cos{2 \theta_{2}} -2 (29 + 28 \cos{2 \theta_{1}}
411: + 7 \cos{4 \theta_{1}})
412: \end{displaymath}
413: \begin{displaymath}
414: \cos{4 \theta_{2}}) \sin^{2}{\theta_{1}} +
415: 32 (\cos{6 \theta_{2}} + \cos{8 \theta_{2}}) \sin^{6}{\theta_{1}})
416: \sin^{2}{\theta} \rbrace.
417: \end{displaymath}
418: \subsection{$g_{45} = g_{\beta b}$}
419: \begin{equation} \label{g45}
420: g_{45} = g_{\beta b} = g_{55} \cos{2 (a + \gamma)} \cos{\theta}.
421: \end{equation}
422: \subsection{$g_{46} = g_{\beta \theta}$}
423: \begin{equation}
424: g_{46} =g_{\beta \theta} =t { \cos{2(a +\gamma)}
425: \sin{2 b} (2 \cos{2 \theta_{1}} - (\cos{4 \theta_{2}} -
426: 3 \cos{2 \theta_{2}})
427: \sin^{2}{\theta_{1}}) \sin{\theta} \over 8 u_{-}}.
428: \end{equation}
429: \subsection{$g_{66} = g_{\theta \theta}$}
430: \begin{equation}
431: g_{66} = g_{\theta \theta} = {32 \cos^{2}{b} \cos^{4}{\theta_{1}} \over
432: 6 + 2 \cos{2 \theta_{1}} + +\cos{2 (\theta_{1}-\theta_{2})}
433: -2 \cos{2 \theta_{2}} +\cos{2(\theta_{1}+\theta_{2})} } \quad +
434: \end{equation}
435: \begin{displaymath}
436: + \quad {1 \over 4} \lbrace -2 -
437: 4 \cos{2 \theta_{1}} + (- \cos{2 \theta_{2}} +
438: \cos{4 \theta_{2}}) \sin^{2}{\theta_{1}} -\cos{2 b} (6 \cos^{2}{\theta_{1}}
439: + (\cos{2 \theta_{2}} +\cos{4 \theta_{2}}) \sin^{2}{\theta_{1}}) \rbrace.
440: \end{displaymath}
441:
442:
443: Since the Euler angles $a$ and $\gamma$ seem only to appear in the
444: $g_{ij}$'s in the additive
445: combination $a+ \gamma$, we conducted a reparameterization of the form
446: $\gamma= \tau -a$. Then, we found that the entries of the
447: associated $8 \times 8$
448: Bures metric tensor
449: (again computed using (\ref{form1}))
450: were not only independent of $\alpha$, as before, but
451: also of the parameter $a$.
452: \section{CONCLUDING REMARKS}
453:
454:
455: We would like to express guarded optimism that, with sufficient expenditure
456: of computational resources
457: and/or added
458: ingenuity and insight, the question marks
459: in (\ref{gmatrix}) can be effectively
460: removed, and one proceed with
461: supplementary analyses, such as inversion of the Bures metric
462: tensor, for purposes of statistical estimation
463: \cite{gm} \cite[eq. (7)]{slat3}
464: and computation of the volume element of the metric, that is the
465: ``quantum Jeffreys' prior'' \cite{slat2,kwek}.
466: Let us note here that the inverse of the Bures metric tensor (\ref{hhe})
467: for the {\it two}-level quantum systems
468: takes the particularly simple form
469: \begin{equation}
470: g^{-1} = 4 \pmatrix{1-x^2 & - x y & - x z \cr
471: -x y & 1-y^2 & -y z \cr
472: -x z & - y z & 1-z^2 \cr}.
473: \end{equation}
474:
475: However, we have
476: confirmed that the remaining not
477: explicitly expressed $g_{ij}$'s in (\ref{gmatrix}) are not simply
478: products of two independent functions, one of the Euler angles
479: ($\alpha,\gamma,a,\beta,b,\theta$), and the other of the spherical angles
480: ($\theta_{1},\theta_{2}$). These three yet (relatively compactly)
481: unexpressed elements (that is, $g_{11} = g_{\alpha \alpha},
482: g_{12} = g_{\alpha \gamma}$ and $g_{14} = g_{\alpha \beta}$)
483: are independent only of
484: $\alpha$, and not of the other seven parameters.
485: If we set $\beta=b=0$, then $g_{14} =0$ and both
486: $g_{11}$ and $g_{12}$ reduce to (cf. \cite[eq. (28)]{slat3})
487: \begin{equation}
488: {(-2 -6 \cos{2 \theta_{1}} +\cos{2 (\theta_{1}-\theta_{2})} -2
489: \cos{2 \theta_{2}} +\cos{2 (\theta_{1}+\theta_{2})})^{2} \sin^{2}{2 \theta}
490: \over 64 (3 + \cos{2 \theta_{2}} -2 \cos{2 \theta_{2}}
491: \sin^{2}{\theta_{1}})}.
492: \end{equation}
493: If we set $\beta=\theta=0$, on the other
494: hand, then both $g_{11}$ and $g_{12}$ reduce to
495: $g_{33}$, that is $g_{55} \sin^{2}{2 b}$, while $g_{44}$ reduces to
496: $-g_{55} \sin{4 b} \sin{2 (a + \gamma)} /2$.
497:
498: We also can not rule out the
499: possibility that some of the
500: more complicated expressions we have presented here --- such as $g_{22}$
501: and $g_{44}$ --- have, in fact, considerably simpler forms than have
502: so far been uncovered.
503: In addition to the transformation $\gamma = \tau -a$, which as we have
504: already noted renders {\it all} the elements of the Bures metric tensor
505: independent of $a$, as well as $\alpha$, another quite interesting
506: reparameterization would be
507: based on the inversion of the relation (\ref{g55eq}), since the
508: element $g_{55}$ itself enters directly into the expressions
509: for many of the other elements. That
510: is, one has
511: \begin{equation} \label{INVERT}
512: \theta_{2} =\sec^{-1}{{2 \sqrt{2} \sin{\theta_{1}} \over
513: \sqrt{ 4 + g_{55} + 4 \cos{2 \theta_{2}} + \sqrt{g_{55}}
514: \sqrt{16 + g_{55} +16 \cos{2 \theta_{2}}}}}}.
515: \end{equation}
516: We have recomputed the Bures metric tensor, which we now denote
517: $\tilde{g}$, again
518: with (\ref{form1}), using $\tau$ and $g_{55}$ as
519: parameters, rather than $\gamma$ and $\theta_{2}$ as in our main
520: analysis and, indeed,
521: found that
522: $\tilde{g}_{b b}$ has the expected form, that is equalling
523: $g_{55}$, and, similar type results for $\tilde{g}_{\alpha a},
524: \tilde{g}_{\alpha b},
525: \tilde{g}_{a a},\tilde{g}_{a b}$ and $\tilde{g}_{\beta b}$. Also,
526: numerically
527: $\tilde{g}_{\tau a}=g_{\gamma a}$.
528:
529: Since M. Byrd has indicated that he will shortly present an Euler angle
530: parameterization of $SU(4)$, parallel to that of
531: $SU(3)$ \cite{us} used here, it will, at that point, be of interest to
532: similarly attempt to
533: recreate the
534: $15 \times 15$
535: Bures metric tensor for the {\it four}-level quantum systems --- which
536: are capable of describing the state of a pair of {\it qubits}
537: (cf. \cite{kus}). For this
538: task, rather than (\ref{form1}), it will be necessary to use one of the
539: other
540: ``explicit formulae for the Bures metric'' given by Dittmann in \cite{ditt4}.
541:
542: In part II of this paper, which is in preparation, we intend to report
543: further progress in the realization and simplification of
544: formulas for the entries of the
545: $8 \times 8$ Bures metric
546: tensor {\it and} of its inverse. These results will be
547: {\it applied} to the
548: study of the curvature properties of the metric
549: (cf. \cite{bilge}), following upon the
550: demonstration of Dittmann \cite{dittym} that the curvature tensor for the
551: Bures metric satisfies the Yang-Mills equation.
552: We will report additional highly interesting features of the curvature.
553:
554:
555: \acknowledgments
556:
557: I would like to express appreciation to the Institute for
558: Theoretical Physics for computational support in this research.
559:
560: \begin{references}
561: \bibitem{petzsudar} D. Petz and C. Sud\'ar, Geometries of quantum
562: states, J. Math. Phys. 37 (1996) 2662-2673 .
563: \bibitem{uhl1} A. Uhlmann, A gauge field governing parallel
564: transport along mixed states, Lett. Math. Phys. 21 (1991) 229-236.
565: \bibitem{uhl2} A. Uhlmann, Density operators as an arena for
566: differential geometry, Rep. Math. Phys. 33 (1993) 253-263.
567: \bibitem{hub1} M. H\"ubner, Explicit computation of the Bures
568: distance for density matrices, Phys. Lett. A 163 (1992) 239-242.
569: \bibitem{hub2} M. H\"ubner, Computation of Uhlmann's
570: parallel transport for density matrices and the Bures metric on
571: three-dimensional Hilbert space, Phys. Lett. A 179 (1993) 239-242
572: \bibitem{ditt1} J. Dittmann and G. Rudolph, On a connection
573: governing parallel transport along 2*2 density matrices, J. Geom. Phys.
574: 10 (1992) 93-106.
575: \bibitem{ditt2} J. Dittmann and G. Rudolph, A class of connections
576: governing parallel transport along density matrices,
577: J. Math. Phys. 33 (1992)
578: 4148-4154.
579: \bibitem{ditt3} J. Dittmann, On the Riemannian geometry of
580: finite dimensional mixed states, Sem. Sophus Lie 3 (1993) 73-87.
581: \bibitem{ditt4} J. Dittmann, Explicit formulae for the Bures
582: metric, J. Phys. A 32 (1999), 2663-2670.
583: \bibitem{slat1} P. B. Slater, Exact Bures Probabilities that Two Quantum Bits
584: are Classically Correlated,
585: Euro. Phys. Jour. B 17 (2000), 471-480.
586: \bibitem{SLAT} P. B. Slater, Hall normalization constants for the
587: Bures volume of the $n$-state quantum systems, J. Phys. A 32
588: (1999) 8231-8246.
589: \bibitem{ZHSL} K. \.Zyczkowski, P. Horodecki, A. Sanpera, and M. Lewenstein,
590: Volume of the set of separable states,
591: Phys. Rev. A 58 (1998) 883-892.
592: \bibitem{bm} S. L. Braunstein and G. J. Milburn, Dynamics of statistical
593: distance: quantum limits for two-level clocks, Phys. Rev. A 51 (1995)
594: 1820-1826.
595: \bibitem{hall} M. J. W. Hall, Random quantum correlations
596: and density operator distributions, Phys. Lett. A 242 (1998) 123-129.
597: \bibitem{helgason} S. Helgason, Differential Geometry and Symmetric
598: Spaces, Academic Press, New York, 1962.
599: \bibitem{ditt11} J. Dittmann, On the Curvature of Monotone Metrics and
600: a Conjecture Concerning the Kubo-Mori Metric, Lin. Alg. Applics. 315 (2000),
601: 83-112.
602: \bibitem{ditt12} J. Dittmann, The scalar curvature of the Bures metric
603: on the space of density matrices, J. Geom. Phys. 31 (1999) 16-24.
604: \bibitem{dittym} J. Dittmann, Yang-Mills equation and Bures metric,
605: Lett. Math. Phys. 46 (1998) 281-287.
606: \bibitem{petzmore} D. Petz, Information-geometry of quantum
607: states, Quantum Prob. Commun. 10 (1998) 135-157.
608: \bibitem{slat2} P. B. Slater, Quantum Fisher-Bures information of
609: two-level systems and a three-level extension, J. Phys. A 29
610: (1996) L271-L275.
611: \bibitem{bloore} F. J. Bloore, Geometrical description of the convex
612: sets of states for systems with spin-${1 \over 2}$ and
613: spin-1, J. Phys. A 9 (1976) 2059-2067.
614: \bibitem{bb} M. S. Byrd and P. B. Slater, Bures Measures over the
615: Spaces of Two and Three-Dimensional Density Matrices, quant-ph/0004055
616: (to appear in Phys. Lett. A).
617: \bibitem{us} L. J. Boya, M. Byrd, M. Mims, and E. C. G. Sudarshan,
618: Density Matrices and Geometric Phases for $n$-State Systems,
619: quant-ph/9810084.
620: \bibitem{twam} J. Twamley, Bures and statistical distance for
621: squeezed thermal states, J. Phys. A 29 (1996) 3723-3731.
622: \bibitem{hasegawa} H. Hasegawa, Exponential and mixture families in
623: quantum statistics, Rep. Math. Phys. 39 (1997) 49-68.
624: \bibitem{po} I. Lukach and Ya. A. Smorodinskii, On the algebra of
625: Gell-Mann's matrices for SU(3) group, Sov. J. Nuclear Phys.
626: 27 (1978) 1694-1702.
627: \bibitem{tod} K. P. Tod, On choosing coordinates to diagonalize the
628: metric, Class. Quant. Grav. 9 (1992) 1693-1705.
629: \bibitem{cox} D. R. Cox and N. Reid, Parameter orthogonality and
630: approximate conditional inference. With a discussion, J. Roy. Statist. Soc.
631: Ser. B 49 (1987) 1-39.
632: \bibitem{gm} R. D. Gill and S. Massar, State estimation for large
633: ensembles, Phys. Rev. A 61 (2000) 042312/1-16.
634: \bibitem{slat3} P. B. Slater, Quantum State Estimation Using
635: Non-Separable Measurements, quant-ph/0006009.
636: \bibitem{kwek} L. C. Kwek, C. H. Oh, and W. Xiang-Bin, Quantum Jeffreys
637: prior for displaced squeezed thermal states,
638: J. Phys. A 32 (1999) 6613-6618.
639: \bibitem{kus} M. Ku\'s and K. \.Zyczkowski, Geometry of Entangled
640: States, quant-ph/0006068.
641: \bibitem{bilge} A. H. Bilge, T. Dereli, and \c S. Ko\c
642: cak, Self-Dual Yang-Mills
643: Fields in Eight Dimensions, Lett. Math. Phys. 36 (1996), 301-309.
644:
645: \end{references}
646:
647: \end{document}