0710.2140/nc.tex
1: % 
2: % Proceedings for Leipzig Workshop July 2007
3: % Title: Noncommutative field theories from a deformation point of
4: % view 
5: %
6: 
7: \documentclass{birkmult}
8: 
9: %
10: % some macro packages
11: %
12: 
13: \usepackage[sort]{cite}       % nicer citations
14: \usepackage{amssymb}          % various ams stuff
15: \usepackage{latexsym}         % dito
16: \usepackage{amsmath}          % dito
17: \usepackage[latin1]{inputenc} % smart input of special characters
18: \usepackage{eucal}            % nicer caligraphic fonts
19: \usepackage{bbm}              % nicer bbm fonts
20: \usepackage{mathrsfs}         % nette kaligraphische Schrift
21: \usepackage{stmaryrd}         % noch ein paar mehr Symbole (Klammern)
22: 
23: 
24: 
25: \renewcommand{\mathbb}[1]{\mathbbm{#1}} % use nicer bbm fonts
26: 
27: \newcommand{\Lie}        {\operatorname{\mathscr{L}}}    
28: \newcommand{\cc}[1]      {\overline{{#1}}}              
29: \newcommand{\id}         {\operatorname{\mathsf{id}}}   
30: \newcommand{\Hom}        {\operatorname{\mathsf{Hom}}}   
31: \newcommand{\End}        {\operatorname{\mathsf{End}}}   
32: \newcommand{\SP}[1]      {\left\langle{#1}\right\rangle} 
33: \newcommand{\ver}        {\mathrm{ver}}                
34: \newcommand{\pr}         {\mathrm{pr}}                    
35: \newcommand{\I}          {\mathrm{i}}
36: \newcommand{\E}          {\mathrm{e}}
37: \newcommand{\D}          {\operatorname{\mathrm{d}}} 
38: \newcommand{\Anti}       {\Lambda}
39: \newcommand{\Schouten}[1]{\left\llbracket{#1}\right\rrbracket}
40: \newcommand{\Diffop}     {\operatorname{\mathrm{Diffop}}}
41: \newcommand{\ring}[1]    {\mathsf{#1}}
42: \newcommand{\HCdiff}     {\operatorname{\mathrm{HC}}_{\mathrm{diff}}}
43: \newcommand{\HHdiff}     {\operatorname{\mathrm{HH}}_{\mathrm{diff}}}
44: \newcommand{\Def}        {\operatorname{\mathrm{Def}}}
45: 
46: 
47: \newtheorem{theorem}{Theorem}[section]
48: \newtheorem{corollary}[theorem]{Corollary}
49: \newtheorem{lemma}[theorem]{Lemma}
50: \newtheorem{proposition}[theorem]{Proposition}
51: \theoremstyle{definition}
52: \newtheorem{definition}[theorem]{Definition}
53: \theoremstyle{remark}
54: \newtheorem{remark}[theorem]{Remark}
55: \newtheorem*{example}{Example}
56: 
57: 
58: \numberwithin{equation}{section}
59: 
60: \begin{document}
61: 
62: %-------------------------------------------------------------------------
63: % editorial commands: to be inserted by the editorial office
64: %
65: %\firstpage{1}
66: %\volume{228}
67: %\Copyrightyear{2004}
68: %\DOI{003-0001}
69: %
70: %
71: %\seriesextra{Just an add-on}
72: %\seriesextraline{This is the Concrete Title of this Book\br H.E. R and S.T.C. W, Eds.}
73: %
74: % for journals:
75: %
76: %\firstpage{1}
77: %\issuenumber{1}
78: %\Volumeandyear{1 (2004)}
79: %\Copyrightyear{2004}
80: %\DOI{003-xxxx-y}
81: %\Signet
82: %\commby{inhouse}
83: %\submitted{March 14, 2003}
84: %\received{March 16, 2000}
85: %\revised{June 1, 2000}
86: %\accepted{July 22, 2000}
87: %
88: %
89: %
90: %---------------------------------------------------------------------------
91: 
92: %
93: % Title author etc...
94: %
95: 
96: \title{Noncommutative field theories from a deformation point of view}
97: 
98: \author{Stefan Waldmann}
99: 
100: \address{%
101:   Fakult{\"a}t f{\"u}r Mathematik und Physik \\
102:   Albert-Ludwigs-Universit{\"a}t Freiburg \\
103:   Physikalisches Institut \\
104:   Hermann Herder Strasse 3 \\
105:   D 79104 Freiburg \\
106:   Germany
107: }
108: 
109: \email{Stefan.Waldmann@physik.uni-freiburg.de}
110: 
111: %
112: % classification, keywords, date
113: %
114: 
115: \subjclass{Primary 53D55; Secondary 58B34, 81T75}
116: 
117: \keywords{Noncommutative field theory, Deformation quantization,
118:   Principal Bundles}
119: 
120: \date{October 2007}
121: 
122: \begin{abstract}
123:     In this review we discuss the global geometry of noncommutative
124:     field theories from a deformation point of view: The space-times
125:     under consideration are deformations of classical space-time
126:     manifolds using star products. Then matter fields are encoded in
127:     deformation quantizations of vector bundles over the classical
128:     space-time. For gauge theories we establish a notion of
129:     deformation quantization of a principal fiber bundle and show how
130:     the deformation of associated vector bundles can be obtained.
131: \end{abstract}
132: 
133: %
134: % Title
135: %
136: \maketitle
137: 
138: %
139: % Introduction
140: %
141: 
142: \section{Introduction}
143: \label{sec:Introduction}
144: 
145: Noncommutative geometry is commonly believed to be a reasonable
146: candidate for the marriage of classical gravity theory in form of
147: Einstein's general relativity on one hand and quantum theory on the
148: other hand. Both theories are experimentally well-established within
149: large regimes of energy and distance scales. However, from a more
150: fundamental point of view, the coexistence of these two theories
151: becomes inevitably inconsistent when one approaches the Planck scale
152: where gravity itself gives significant quantum effects.
153: 
154: 
155: Since general relativity is ultimately the theory of the geometry of
156: space-time it seems reasonable to use notions of `quantum geometry'
157: known under the term \emph{noncommutative geometry} in the sense of
158: Connes \cite{connes:1994a} to achieve appropriate formulations of what
159: eventually should become quantum gravity. Of course, this ultimate goal
160: has not yet been reached but techniques of noncommutative geometry
161: have been used successfully to develop models of quantum field
162: theories on quantum space-times being of interest for their own.
163: Moreover, a deeper understanding of ordinary quantum field theories
164: can be obtained by studying their counterparts on `nearby'
165: noncommutative space-times. On the other hand, people started to
166: investigate experimental implications of a possible  noncommutativity
167: of space-time in future particle experiments.
168: 
169: 
170: Such a wide scale of applications and interests justifies a more
171: \emph{conceptual} discussion of noncommutative space-times and
172: (quantum) field theories on them in order to clarify fundamental
173: questions and generic features expected to be common to all examples.
174: 
175: 
176: In this review, we shall present such an approach from the point of
177: view of deformation theory: noncommutative space-times are not studied
178: by themselves but always with respect to a classical space-time, being
179: suitably deformed into the noncommutative one. Clearly, this point of
180: view can not cover all possible (and possibly interesting)
181: noncommutative geometries but only a particular class. Moreover, we
182: focus on \emph{formal} deformations for technical reasons. It is
183: simply the most easy approach where one can rely on the very powerful
184: machinery of algebraic deformation theory. But it also gives hints on
185: approaches beyond formal deformations: finding obstructions in the
186: formal framework will indicate even more severe obstructions in any
187: non-perturbative approach.
188: 
189: 
190: In the following, we discuss mainly two questions: first, what is the
191: appropriate description of matter fields on deformed space-times and,
192: second, what are the deformed analogues of principal bundles needed
193: for the formulation of gauge theories. The motivation for these two
194: questions should be clear.
195: 
196: 
197: The review is organized as follows: in Section~\ref{sec:NCSpacetimes},
198: we recall some basic definitions and properties concerning deformation
199: quantizations and star products needed for the set-up of
200: noncommutative space-times. We discuss some fundamental examples as
201: well as a new class of locally noncommutative space-times.
202: Section~\ref{sec:MatterFieldsVectorBundles} is devoted to the study of
203: matter fields: we use the Serre-Swan theorem to relate matter fields
204: to projective modules and discuss their deformation theory. Particular
205: interest is put on the mass terms and their positivity properties.  In
206: Section~\ref{sec:DeformedPrincipalBundles} we establish the notion of
207: deformation quantization of principal fiber bundles and discuss the
208: existence and uniqueness results. Finally, in
209: Section~\ref{sec:CommutantAssociated} we investigate the resulting
210: commutant and formulate an appropriate notion of associated (vector)
211: bundles. This way we make contact to the results of
212: Section~\ref{sec:MatterFieldsVectorBundles}. The review is based on
213: joint works with Henrique Bursztyn on one hand as well as with Martin
214: Bordemann, Nikolai Neumaier and Stefan Weiß on the other hand.
215: 
216: 
217: %
218: % Noncommutative space-times
219: %
220: 
221: \section{Noncommutative space-times}
222: \label{sec:NCSpacetimes}
223: 
224: In order to implement uncertainty relations for measuring coordinates
225: of events in space-time it has been proposed already very early to
226: replace the commutative algebra of (coordinate) functions by some
227: noncommutative algebra. In \cite{doplicher.fredenhagen.roberts:1995a}
228: a concrete model for a noncommutative Minkowski space-time was
229: introduced with commutation relations of the form
230: \begin{equation}
231:     \label{eq:xmuxnuthetamunu}
232:     [\hat{x}^\mu, \hat{x}^\nu] = \I \lambda \theta^{\mu\nu},
233: \end{equation}
234: where $\lambda$ plays the role of the deformation parameter and has
235: the physical dimension of an area. Usually, this area will be
236: interpreted as the Planck area. Moreover, $\theta$ is a real,
237: antisymmetric tensor which in
238: \cite{doplicher.fredenhagen.roberts:1995a} and many following papers
239: is assumed to be \emph{constant}: in
240: \cite{doplicher.fredenhagen.roberts:1995a} this amounts to require
241: that $\theta^{\mu\nu}$ belongs to the center of the new algebra of
242: noncommutative coordinates.
243: 
244: 
245: Instead of constructing an abstract algebra where commutation
246: relations like \eqref{eq:xmuxnuthetamunu} are fulfilled, it is
247: convenient to use a `symbol calculus' and encode
248: \eqref{eq:xmuxnuthetamunu} already for the classical coordinate
249: functions by changing the multiplication law instead. For functions
250: $f$ and $g$ on the classical Minkowski space-time one defines the
251: Weyl-Moyal star product by
252: \begin{equation}
253:     \label{eq:WeylMoyal}
254:     f \star g = \mu \circ \E^{\frac{\I\lambda}{2} \theta^{\mu\nu}
255:       \frac{\partial}{\partial x^\mu} \otimes
256:       \frac{\partial}{\partial x^\nu}}
257:     (f \otimes g),
258: \end{equation}
259: where $\mu (f \otimes g) = fg$ denotes the undeformed, pointwise
260: product. Then \eqref{eq:xmuxnuthetamunu} holds for the classical
261: coordinate functions with respect to the $\star$-commutator.
262: 
263: 
264: Clearly, one has to be slightly more careful with expressions like
265: \eqref{eq:WeylMoyal}: in order to make sense out of the infinite
266: differentiations the functions $f$ and $g$ first should be $C^\infty$.
267: But then the exponential series does not converge in general whence a
268: more sophisticated analysis is required. Though this can be done in a
269: completely satisfying way for this particular example, we shall not
270: enter this discussion here but consider \eqref{eq:WeylMoyal} as a
271: formal power series in the deformation parameter $\lambda$. Then
272: $\star$ becomes an associative $\mathbb{C}[[\lambda]]$-bilinear
273: product for $C^\infty(\mathbb{R}^4)[[\lambda]]$, i. e. a star product
274: in the sense of \cite{bayen.et.al:1978a}. It should be noted that the
275: interpretation of \eqref{eq:WeylMoyal} as formal series in $\lambda$
276: is physically problematic: $\lambda$ is the Planck area and hence a
277: physically measurable and non-zero quantity. Thus our point of view
278: only postpones the convergence problem and can be seen as a
279: perturbative approach.
280: 
281: 
282: With this example in mind, one arrives at several conceptual
283: questions: The first is that Minkowski space-time is clearly not a
284: very realistic background when one wants to consider quantum effects
285: of `hard' gravity. Here already classically nontrivial curvature and
286: even nontrivial topology may arise. Thus one is forced to consider
287: more general and probably even generic Lorentz manifolds instead.
288: Fortunately, deformation quantization provides a well-established and
289: successful mathematical framework for this geometric situation.
290: 
291: 
292: Recall that a star product on a manifold $M$ is an associative
293: $\mathbb{C}[[\lambda]]$-bilinear multiplication $\star$ for $f, g \in
294: C^\infty(M)[[\lambda]]$ of the form
295: \begin{equation}
296:     \label{eq:StarProduct}
297:     f \star g = \sum_{r=0}^\infty \lambda^r C_r(f, g),
298: \end{equation}
299: where $C_0(f, g) = fg$ is the undeformed, pointwise multiplication and
300: the $C_r$ are bidifferential operators. Usually, one requires $1 \star
301: f = f = f \star 1$ for all $f$. It is easy to see that $\{f, g\} =
302: \frac{1}{\I} (C_1(f, g) - C_1(g, f))$ defines a Poisson bracket on
303: $M$. Conversely, and this is the highly nontrivial part, any Poisson
304: bracket $\{f, g\} = \theta(\D f, \D g)$, where
305: \begin{equation}
306:     \label{eq:ThetaGlobal}
307:     \theta \in \Gamma^\infty(\Anti^2 TM),
308:     \quad
309:     \Schouten{\theta, \theta} = 0
310: \end{equation}
311: is the corresponding Poisson tensor, can be quantized into a star
312: product \cite{kontsevich:2003a, dewilde.lecomte:1983b}. Beside these
313: existence results one has a very good understanding of the
314: classification of such star products \cite{nest.tsygan:1995a,
315:   gutt.rawnsley:1999a, kontsevich:2003a}, see also
316: \cite{dito.sternheimer:2002a, gutt:2000a} for recent reviews and
317: \cite{waldmann:2007a} for an introduction.
318: 
319: 
320: With this geometric interpretation the Weyl-Moyal star product on
321: Minkowski space-time turns out to be a deformation quantization of the
322: \emph{constant} Poisson structure
323: \begin{equation}
324:     \label{eq:ThetaConstant}
325:     \theta = \frac{1}{2} \theta^{\mu\nu} 
326:     \frac{\partial}{\partial x^\mu}
327:     \wedge
328:     \frac{\partial}{\partial x^\nu}.
329: \end{equation}
330: On a generic space-time $M$ there is typically \emph{no} transitive
331: action of isometries which would justify the notion of a `constant'
332: bivector field. Thus a star product $\star$ on $M$ is much more
333: complicated than \eqref{eq:WeylMoyal} in general: already the first
334: order term is a (nontrivial) Poisson structure and for the higher
335: order terms one has to invoke the (unfortunately rather inexplicit)
336: existence theorems.
337: 
338: 
339: Thus answering the first question by using general star products
340: raises the second: what is the physical role of a Poisson structure on
341: space-time? While on Minkowski space-time with constant $\theta$ we
342: can view the finite number of coefficients $\theta^{\mu\nu} \in
343: \mathbb{R}$ as \emph{parameters} of the theory this is certainly no
344: longer reasonable in the more realistic geometric framework: there is
345: an infinity of Poisson structures on each manifold whence an
346: interpretation as `parameter' yields a meaningless theory. Instead,
347: $\theta$ has to be considered as a \emph{field} itself, obeying its
348: own dynamics compatible with the constraint of the Jacobi identity
349: $\Schouten{\theta, \theta} = 0$. Unfortunately, up to now a reasonable
350: `field equation' justified by first principles seems to be missing.
351: 
352: 
353: This raises a third conceptual question, namely why should there by
354: any Poisson structure on $M$ and what are possible experimental
355: implications? In particular, the original idea of introducing a
356: noncommutative structure was to implement uncertainty relations
357: forbidding the precise localization of events. The common believe is
358: that such quantum effects should only play a role when approaching the
359: Planck scale. Now it turns out that the quantum field theories put on
360: such a noncommutative Minkowski space-time (or their Euklidian
361: counterparts) suffer all from quite unphysical properties: Typically,
362: the noncommutativity enters in long-distance/low-energy features
363: contradicting our daily life experience. Certainly, a last word is not
364: said but there might be a simple explanation why such effects should
365: be expected: the \emph{global} $\theta$ (constant or not) yields
366: global effects on $M$. This was the starting point of a more refined
367: notion of noncommutative space-times advocated in
368: \cite{bahns.waldmann:2007a, heller.neumaier.waldmann:2007a} as
369: \emph{locally} noncommutative space-times. Roughly speaking, without
370: entering the technical details, it is not $M$ which should become
371: noncommutative but $TM$. Here the tangent bundle is interpreted as the
372: bundle of all normal charts on $M$ and for each normal chart with
373: origin $p \in M$ one constructs its own star product $\star_p$. The
374: crucial property is then that $\star_p$ is the pointwise, commutative
375: product outside a (small) compact subset around $p$. This way, the
376: long-distance behaviour (with respect to the reference point $p$) is
377: classical while close to $p$ there is a possibly even very strong
378: noncommutativity. In some sense, this is an implementation of an idea
379: of Julius Wess, proposing that the transition from classical geometry
380: to quantum geometry should be understood as a kind of phase transition
381: taking place at very small distances \cite{wess:misc}. Of course, the
382: conceptual question about the physical origin of the corresponding
383: Poisson structure on $TM$ as well as the convergence problem still
384: persists also in this approach.
385: 
386: 
387: Ignoring these questions about the nature of $\theta$, we shall assume
388: in the following that we are given a star product $\star$ on a
389: manifold $M$ which can be either space-time itself or its tangent
390: bundle in the locally noncommutative case. Then we address the
391: question how to formulate reasonable field theories on $(M, \star)$.
392: Here we shall focus on \emph{classical} field theories which still
393: need to be quantized later on. On the other hand, we seek for a
394: \emph{geometric} formulation not relying on particular assumptions
395: about the underlying classical space-time.
396: 
397: 
398: %
399: % Matter fields and deformed vector bundles
400: %
401: 
402: \section{Matter fields and deformed vector bundles}
403: \label{sec:MatterFieldsVectorBundles}
404: 
405: In this section we review some results from
406: \cite{bursztyn.waldmann:2005b, waldmann:2005a, waldmann:2001b,
407:   bursztyn.waldmann:2000b}.
408: 
409: 
410: In classical field theories both bosonic and fermionic matter fields
411: are given by sections of appropriate vector bundles. For convenience,
412: we choose the vector bundles to be complex as also the function
413: algebra $C^\infty(M)$ consists of complex-valued functions. However,
414: the real case can be treated completely analogously.  Thus let $E
415: \longrightarrow M$ be a complex vector bundle over $M$. Then the
416: $E$-valued fields are the (smooth) sections $\Gamma^\infty(E)$ which
417: form a module over $C^\infty(M)$ by pointwise multiplication. Thanks
418: to the commutativity of $C^\infty(M)$ we have the freedom to choose
419: this module structure to be a right module structure for later
420: convenience.
421: 
422: 
423: It is a crucial feature of vector bundles that $\Gamma^\infty(E)$ is
424: actually a finitely generated and projective module:
425: \begin{theorem}[Serre-Swan]
426:     \label{theorem:SerreSwan}
427:     The sections $\Gamma^\infty(E)$ of a vector bundle $E
428:     \longrightarrow M$ are a finitely generated and projective
429:     $C^\infty(M)$-module. Conversely, any such module arises this way
430:     up to isomorphism.
431: \end{theorem}
432: Recall that a right module $\mathcal{E}_{\mathcal{A}}$ over an algebra
433: $\mathcal{A}$ is called finitely generated and projective if there
434: exists an idempotent $e^2 = e \in M_n(\mathcal{A})$ such that
435: $\mathcal{E}_{\mathcal{A}} \cong e\mathcal{A}^n$ as right
436: $\mathcal{A}$-modules. More geometrically speaking, for any vector
437: bundle $E \longrightarrow M$ there is another vector bundle $F
438: \longrightarrow M$ such that their Whitney sum $E \oplus F$ is
439: isomorphic to a trivial vector bundle $M \times \mathbb{C}^n
440: \longrightarrow M$. Note that the Serre-Swan theorem has many
441: incarnations, e.g. the original version was formulated for compact
442: Hausdorff spaces and continuous sections/functions. Note also that for
443: our situation no compactness assumption is necessary (though it
444: drastically simplifies the proof) as manifolds are assumed to be
445: second countable.
446: \begin{remark}
447:     \label{remark:SerreSwan}
448:     The Serre-Swan theorem is the main motivation for noncommutative
449:     geometry to consider finitely generated and projective modules
450:     over a not necessarily commutative algebra $\mathcal{A}$ as
451:     `vector bundles' over the (noncommutative) space described by
452:     $\mathcal{A}$ in general.
453: \end{remark}
454: 
455: 
456: For physical applications in field theory one usually has more
457: structure on $E$ than just a bare vector bundle. In particular, for a
458: Lagrangean formulation a `mass term' in the Lagrangean is
459: needed. Geometrically such a mass term corresponds to a Hermitian
460: fiber metric $h$ on $E$. One can view a Hermitian fiber metric as a
461: map
462: \begin{equation}
463:     \label{eq:FiberMetric}
464:     h: \Gamma^\infty(E) \times \Gamma^\infty(E) 
465:     \longrightarrow C^\infty(M),
466: \end{equation}
467: which is $\mathbb{C}$-linear in the second argument and satisfies
468: $\cc{h(\phi, \psi)} = h(\psi, \phi)$, $h(\phi, \psi f) = h(\phi, \psi)
469: f$ as well as 
470: \begin{equation}
471:     \label{eq:Positive}
472:     h(\phi, \phi) \ge 0    
473: \end{equation}
474: for $\phi, \psi \in \Gamma^\infty(E)$ and $f \in C^\infty(M)$. The
475: pointwise non-degeneracy of $h$ is equivalent to the property that
476: \begin{equation}
477:     \label{eq:hNondegenerate}
478:     \Gamma^\infty(E) \ni \phi 
479:     \; \mapsto \; 
480:     h(\phi, \cdot) \in \Gamma^\infty(E^*)
481: \end{equation}
482: is an antilinear module isomorphism. Note that the sections of the
483: dual vector bundle $E^* \longrightarrow M$ coincide with the dual
484: module, i.e. we have $\Gamma^\infty(E^*) = \Hom_{C^\infty(M)}
485: (\Gamma^\infty(E), C^\infty(M))$.
486: 
487: 
488: In order to encode now the positivity \eqref{eq:Positive} in a more
489: algebraic way suitable for deformation theory, we have to consider the
490: following class of algebras: First, we use a ring of the form
491: $\ring{C} = \ring{R}(\I)$ with $\I^2 = -1$ for the scalars where
492: $\ring{R}$ is an \emph{ordered ring}. This includes both $\mathbb{R}$
493: and $\mathbb{R}[[\lambda]]$, where positive elements in
494: $\mathbb{R}[[\lambda]]$ are defined by
495: \begin{equation}
496:     \label{eq:PositivePowerseries}
497:     a = \sum_{r=r_0}^\infty \lambda^r a_r > 0
498:     \quad
499:     \textrm{if}
500:     \quad a_{r_0} > 0.
501: \end{equation}
502: In fact, this way $\ring{R}[[\lambda]]$ becomes an ordered ring
503: whenever $\ring{R}$ is ordered. More physically speaking, the ordering
504: of $\mathbb{R}[[\lambda]]$ refers to a kind of asymptotic
505: positivity. Then the algebras in question should be $^*$-algebras over
506: $\ring{C}$: Indeed, $C^\infty(M)$ is a $^*$-algebra over $\mathbb{C}$
507: where the $^*$-involution is the pointwise complex conjugation. For
508: the deformed algebras $(C^\infty(M)[[\lambda]], \star)$ we require
509: that the star product is Hermitian, i.e.
510: \begin{equation}
511:     \label{eq:HermitianStarProduct}
512:     \cc{f \star g} = \cc{g} \star \cc{f}
513: \end{equation}
514: for all $f, g \in C^\infty(M)[[\lambda]]$. For a real Poisson
515: structure $\theta$ this can be achieved by a suitable choice of
516: $\star$.
517: 
518: 
519: For such a $^*$-algebra we can now speak of positive functionals and
520: positive elements \cite{bursztyn.waldmann:2001a} by mimicking the
521: usual definitions from operator algebras, see e.g.
522: \cite{schmuedgen:1990a} for the case of (unbounded) operator algebras
523: and \cite{waldmann:2004a} for a detailed comparison.
524: \begin{definition}
525:     \label{definition:Positive}
526:     Let $\mathcal{A}$ be a $^*$-algebra over $\ring{C} =
527:     \ring{R}(\I)$. A $\ring{C}$-linear functional $\omega: \mathcal{A}
528:     \longrightarrow \ring{C}$ is called positive if $\omega(a^*a) \ge
529:     0$ for all $a \in \mathcal{A}$. An element $a \in \mathcal{A}$ is
530:     called positive if $\omega(a) \ge 0$ for all positive functionals
531:     $\omega$.
532: \end{definition}
533: We denote the convex cone of positive elements in $\mathcal{A}$ by
534: $\mathcal{A}^+$. It is an easy exercise to show that for $\mathcal{A}
535: = C^\infty(M)$ the positive functionals are the compactly supported
536: Borel measures and $\mathcal{A}^+$ consists of functions $f$ with
537: $f(x) \ge 0$ for all $x \in M$.
538: 
539: 
540: Using this notion of positive elements and motivated by
541: \cite{lance:1995a}, the algebraic formulation of a fiber metric is now
542: as follows \cite{bursztyn.waldmann:2005b, bursztyn.waldmann:2000b}:
543: \begin{definition}
544:     \label{definition:FiberMetric}
545:     Let $\mathcal{E}_{\mathcal{A}}$ be a right
546:     $\mathcal{A}$-module. Then an inner product $\SP{\cdot, \cdot}$ on
547:     $\mathcal{E}_{\mathcal{A}}$ is a map
548:     \begin{equation}
549:         \label{eq:InnerProduct}
550:         \SP{\cdot, \cdot}: 
551:         \mathcal{E}_{\mathcal{A}} \times \mathcal{E}_{\mathcal{A}}
552:         \longrightarrow \mathcal{A},
553:     \end{equation}
554:     which is $\ring{C}$-linear in the second argument and satisfies
555:     $\SP{x, y} = \SP{y, x}^*$, $\SP{x, y \cdot a} = \SP{x, y} a$, and
556:     $\SP{x, y} = 0$ for all $y$ implies $x = 0$. The inner product is
557:     called strongly non-degenerate if in addition
558:     \begin{equation}
559:         \label{eq:StronglyNondegenerate}
560:         \mathcal{E}_{\mathcal{A}} \ni x
561:         \; \mapsto \; \SP{x, \cdot} 
562:         \in \mathcal{E}^* =
563:         \Hom_{\mathcal{A}}(\mathcal{E}_{\mathcal{A}}, \mathcal{A})
564:     \end{equation}
565:     is bijective. It is called completely positive if for all $n \in
566:     \mathbb{N}$ and $x_1, \ldots, x_n \in \mathcal{E}_{\mathcal{A}}$
567:     one has $(\SP{x_i, x_j}) \in M_n(\mathcal{A})^+$.
568: \end{definition}
569: Clearly, a Hermitian fiber metric on a complex vector bundle endows
570: $\Gamma^\infty(E)$ with a completely positive, strongly non-degenerate
571: inner product in the sense of Definition~\ref{definition:FiberMetric}.
572: 
573: 
574: With the above definition in mind we can now formulate the following
575: deformation problem \cite{bursztyn.waldmann:2000b}:
576: \begin{definition}
577:     \label{definition:DeformationOfE}
578:     Let $\star$ be a Hermitian star product on $M$ and $E
579:     \longrightarrow M$ a complex vector bundle with fiber metric $h$.
580:     \begin{enumerate}
581:     \item A deformation quantization $\bullet$ of $E$ is a right
582:         module structure $\bullet$ for $\Gamma^\infty(E)[[\lambda]]$
583:         with respect to $\star$ of the form
584:         \begin{equation}
585:             \label{eq:RightModulephif}
586:             \phi \bullet f = \sum_{r=0}^\infty \lambda^r R_r(\phi, f)
587:         \end{equation}
588:         with bidifferential operators $R_r$ and $R_0(\phi, f) = \phi
589:         f$.
590:     \item For a given deformation quantization $\bullet$ of $E$ a
591:         deformation quantization of $h$ is a completely positive inner
592:         product $\boldsymbol{h}$ for $(\Gamma^\infty(E)[[\lambda]],
593:         \bullet)$ of the form
594:         \begin{equation}
595:             \label{eq:DeformedFiberMetric}
596:             \boldsymbol{h}(\phi, \psi) = \sum_{r=0}^\infty \lambda^r
597:             \boldsymbol{h}_r(\phi, \psi)
598:         \end{equation}
599:         with (sesquilinear) bidifferential operators
600:         $\boldsymbol{h}_r$ and $\boldsymbol{h}_0 = h$.
601:     \end{enumerate}
602: \end{definition}
603: 
604: 
605: In addition, we call two deformations $\bullet$ and $\tilde{\bullet}$
606: \emph{equivalent} if there exists a formal series of differential
607: operators
608: \begin{equation}
609:     \label{eq:EquivalenceTrafo}
610:     T = \id + \sum_{r=1}^\infty \lambda^r T_t:
611:     \Gamma^\infty(E)[[\lambda]] \longrightarrow
612:     \Gamma^\infty(E)[[\lambda]],
613: \end{equation}
614: such that
615: \begin{equation}
616:     \label{eq:Equivalencebullets}
617:     T(\phi \bullet f) = T(\phi) \tilde{\bullet} f.
618: \end{equation}
619: With other words, $T$ is a module isomorphism starting with the
620: identity in order $\lambda^0$ such that $T$ is not visible in the
621: classical/commutative limit. Conversely, starting with one deformation
622: $\bullet$ and a $T$ like in \eqref{eq:EquivalenceTrafo}, one obtains
623: another equivalent deformation $\tilde{\bullet}$ by defining
624: $\tilde{\bullet}$ via \eqref{eq:Equivalencebullets}. Similarly, we
625: define two deformations $\boldsymbol{h}$ and $\tilde{\boldsymbol{h}}$
626: to be \emph{isometric} if there exists a self-equivalence $U$ with
627: \begin{equation}
628:     \label{eq:Isometric}
629:     \boldsymbol{h}(\phi, \psi)
630:     = \tilde{\boldsymbol{h}}(U(\phi), U(\psi)).
631: \end{equation}
632: The relevance of the above notions for noncommutative field theories
633: should now be clear: for a classical matter field theory modelled on
634: $E \longrightarrow M$ we obtain the corresponding noncommutative field
635: theory by choosing a deformation $\bullet$ (if existing!) together
636: with a deformation $\boldsymbol{h}$ (if existing!) in order to write
637: down noncommutative Lagrangeans involving expressions like
638: $\mathcal{L}(\phi) = \boldsymbol{h}(\phi, \phi) + \cdots$.
639: 
640: 
641: Note that naive expressions like $\cc{\phi} \star \phi$ do not make
642: sense geometrically, even on the classical level: sections of a vector
643: bundle can not be `multiplied' without the extra structure of a fiber
644: metric $h$ unless the bundle is trivial \emph{and} trivialized. In
645: this particular case we can of course use the canonical fiber metric
646: coming from the canonical inner product on $\mathbb{C}^n$. We refer to
647: \cite{waldmann:2005a, waldmann:2001b} for a further discussion.
648: 
649: 
650: We can now state the main results of this section, see
651: \cite{bursztyn.waldmann:2005b, bursztyn.waldmann:2000b} for detailed
652: proofs:
653: \begin{theorem}
654:     \label{theorem:ExistenceDeformedE}
655:     For any star product $\star$ on $M$ and any vector bundle $E
656:     \longrightarrow M$ there exists a deformation quantization
657:     $\bullet$ with respect to $\star$ which is unique up to
658:     equivalence.
659: \end{theorem}
660: \begin{theorem}
661:     \label{theorem:ExistenceFiberMetric}
662:     For any Hermitian star product $\star$ on $M$ and any fiber metric
663:     $h$ on $E \longrightarrow M$ and any deformation quantization
664:     $\bullet$ of $E$ there exists a deformation quantization
665:     $\boldsymbol{h}$ of $h$ which is unique up to isometry.
666: \end{theorem}
667: The first theorem relies heavily on the Serre-Swan theorem and the
668: fact that algebraic $K_0$-theory is stable under formal deformations
669: \cite{rosenberg:1996a:pre}. In fact, projections and hence projective
670: modules can always be deformed in an essentially unique way. The
671: second statement follows for much more general deformed algebras than
672: only for star products, see \cite{bursztyn.waldmann:2005b}.
673: \begin{remark}
674:     \label{remark:Matter}
675:     \begin{enumerate}
676:     \item In case $M$ is symplectic, one has even a rather explicit
677:         Fedosov-like construction for $\bullet$ and $\boldsymbol{h}$
678:         in terms of connections, see \cite{waldmann:2002b}.
679:     \item\label{item:EndEDeformation} It turns out that also
680:         $\Gamma^\infty(\End(E))$ becomes deformed into an associative
681:         algebra $(\Gamma^\infty(\End(E))[[\lambda]], \star')$ such
682:         that $\Gamma^\infty(E)[[\lambda]]$ becomes a Morita
683:         equivalence bimodule between the two deformed algebras $\star$
684:         and $\star'$. Together with the deformation $\boldsymbol{h}$
685:         of $h$ one obtains even a strong Morita equivalence bimodule
686:         \cite{bursztyn.waldmann:2005b}.
687:     \item Note also that the results of the two theorems are more than
688:         just the `analogy' used in the more general framework of
689:         noncommutative geometry: we have here a precise link between
690:         the noncommutative geometries and their classical/commutative
691:         limits via deformation. For general noncommutative geometries
692:         it is not even clear what a classical/commutative limit is.
693:     \end{enumerate}
694: \end{remark}
695: 
696: 
697: %
698: % Deformed principal bundles
699: %
700: 
701: \section{Deformed principal bundles}
702: \label{sec:DeformedPrincipalBundles}
703: 
704: This section contains a review of results obtained in
705: \cite{bordemann.neumaier.waldmann.weiss:2007a:pre} as well as in
706: \cite{weiss:2006a}.
707: 
708: 
709: In all fundamental theories of particle physics the field theories
710: involve gauge fields. Geometrically, their formulation is based on the
711: use of a principal bundle $\pr: P \longrightarrow M$ with structure
712: group $G$, i.e. $P$ is endowed with a (right) action of $G$ which is
713: proper and free whence the quotient $P \big/ G = M$ is again a smooth
714: manifold. All the matter fields are then obtained as sections of
715: associated vector bundles by choosing an appropriate representation of
716: $G$.
717: 
718: 
719: In the noncommutative framework there are several approaches to gauge
720: theories: for particular structure groups and representations notions
721: of gauge theories have been developed by Jurco, Schupp, Wess and
722: coworkers \cite{jurco.et.al.:2001a, jurco.schupp.wess:2001a,
723:   jurco.schraml.schupp.wess:2000a, jurco.schupp.wess:2000a,
724:   jurco.schupp:2000a}. Here the focus was mainly on local
725: considerations and the associated bundles but not on the principal
726: bundle directly. Conversely, there is a purely algebraic and
727: intrinsically global formulation of Hopf-Galois extensions where not
728: only the base manifold $M$ is allowed to be noncommutative but even
729: the structure group is replaced by a general Hopf algebra, see e.g.
730: \cite{dabrowski.grosse.hajac:2001a} and references therein for the
731: relation of Hopf-Galois theory to noncommutative gauge field theories.
732: However, as we shall see below, in this framework which a priori does
733: not refer to any sort of deformation, in general only very particular
734: Poisson structures on $M$ can be used. Finally, in
735: \cite{waldmann:2002a} a local approach to principal $\mathrm{Gl}(n
736: \mathbb{C})$ or $\mathrm{U}(n)$ bundles was implicitly used via
737: deformed transition matrices.
738: 
739: 
740: We are now seeking for a definition of a deformation quantization of a
741: principal bundle $P$ for a generic structure Lie group $G$, arbitrary
742: $M$ and arbitrary star product $\star$ on $M$ without further
743: assumptions on $P$. In particular, the formulation should be
744: intrinsically global.
745: 
746: 
747: The idea is to consider the classical algebra homomorphism
748: \begin{equation}
749:     \label{eq:prpullback}
750:     \pr^*: C^\infty(M) \longrightarrow C^\infty(P)
751: \end{equation}
752: and try to find a reasonable deformation of $\pr^*$. The first idea
753: would be to find a star product $\star_P$ on $P$ with a deformation
754: $\boldsymbol{\pr^*} = \sum_{r=0}^\infty \lambda^r
755: \boldsymbol{\pr^*}_r$ of $\boldsymbol{\pr^*}_0 = \pr^*$ into an
756: algebra homomorphism 
757: \begin{equation}
758:     \label{eq:prAlgebraMorph}
759:     \boldsymbol{\pr^*}(f \star g) 
760:     = \boldsymbol{\pr^*}(f) \star_P \boldsymbol{\pr^*}(g)
761: \end{equation}
762: with respect to the two star products $\star$ and $\star_P$. In some
763: sense this would be the first (but not the only) requirement for a
764: Hopf-Galois extension. In fact, the first order of
765: \eqref{eq:prAlgebraMorph} implies that the \emph{classical} projection
766: map $\pr$ is a Poisson map with respect to the Poisson structures
767: induced by $\star$ on $M$ and $\star_P$ on $P$. The following example
768: shows that in general there are obstructions to achieve
769: \eqref{eq:prAlgebraMorph} already on the classical level:
770: \begin{example}
771:     \label{example:HopfFibration}
772:     Consider the Hopf fibration $\pr: S^3 \longrightarrow S^2$ (which
773:     is a nontrivial principal $S^1$-bundle over $S^2$) and equip $S^2$
774:     with the canonical symplectic Poisson structure. Then there exists
775:     \emph{no} Poisson structure on $S^3$ such that $\pr$ becomes a
776:     Poisson map. Indeed, if there would be such a Poisson structure
777:     then necessarily all symplectic leaves would be two-dimensional as
778:     symplectic leaves are mapped into symplectic leaves and $S^2$ is
779:     already symplectic. Fixing one symplectic leaf in $S^3$ one checks
780:     that $\pr$ restricted to this leaf is still surjective and thus
781:     provides a covering of $S^2$. But $S^2$ is simply connected whence
782:     the symplectic leaf is itself a $S^2$. This would yield a section
783:     of the nontrivial principal bundle $\pr: S^3 \longrightarrow S^2$,
784:     a contradiction.
785: \end{example}
786: \begin{remark}
787:     \label{remark:OtherHopfGalois}
788:     Note that there are prominent examples of Hopf-Galois extensions
789:     using quantum spheres, see
790:     e.g.~\cite{hajac.matthes.szymanski:2003a} and references therein.
791:     The above example shows that when taking the semi-classical limit
792:     of these $q$-deformations one obtains Poisson structures on $S^2$
793:     which are certainly not symplectic.  Note that this was a crucial
794:     feature in the above example. A further investigation of these
795:     examples is work in progress.
796: \end{remark}
797: 
798: 
799: The above example shows that the first idea of deforming the
800: projection map into an algebra homomorphism leads to hard obstructions
801: in general, even though there are interesting classes of examples
802: where the obstructions are absent. However, as we are interested in an
803: approach not making too much assumptions in the beginning, we abandon
804: this first idea. The next weaker requirement would be to deform
805: $\pr^*$ not into an algebra homomorphism but only turning
806: $C^\infty(P)$ into a \emph{bimodule}. This would have the advantage
807: that there is no Poisson structure on $P$ needed. However, a more
808: subtle analysis shows that again for the Hopf fibration such a
809: bimodule structure is impossible if one uses a star product on $S^2$
810: coming from the symplectic Poisson structure. Thus we are left with a
811: \emph{module structure}: for later convenience we choose a right
812: module structure and state the following definition
813: \cite{bordemann.neumaier.waldmann.weiss:2007a:pre}:
814: \begin{definition}
815:     \label{definition:DeformationPFB}
816:     Let $\pr: P \circlearrowleft G \longrightarrow M$ be a principal
817:     $G$-bundle over $M$ and $\star$ a star product on $M$. A
818:     deformation quantization of $P$ is a right $\star$-module
819:     structure $\bullet$ for $C^\infty(P)[[\lambda]]$ of the form
820:     \begin{equation}
821:         \label{eq:BulletPFB}
822:         F \bullet f = F \pr^*f + \sum_{r=1}^\infty \lambda^r
823:         \varrho_r(F, f),
824:     \end{equation}
825:     where $\varrho_r: C^\infty(P) \times C^\infty(M) \longrightarrow
826:     C^\infty(P)$ is a bidifferential operator (along $\pr$) for all $r
827:     \ge 1$, such that in addition one has the $G$-equivariance
828:     \begin{equation}
829:         \label{eq:Gequivariance}
830:         g^* (F \bullet f) = g^*F \bullet f
831:     \end{equation}
832:     for all $F \in C^\infty(P)[[\lambda]]$, $f \in
833:     C^\infty(M)[[\lambda]]$ and $g \in G$.
834: \end{definition}
835: Note that as $G$ acts on $P$ from the right, the pull-backs with the
836: actions of $g \in G$ provide a left action on $C^\infty(P)$ in
837: \eqref{eq:Gequivariance}. Then this condition means that the
838: $G$-action commutes with the module multiplications.
839: 
840: 
841: Note that the module property $F \bullet (f \star g) = (F \bullet f)
842: \bullet g$ implies that the constant function $1$ acts as identity.
843: Indeed, since $1 \star 1 = 1$ the action of $1$ via $\bullet$ is a
844: projection. However, in zeroth order the map $F \mapsto F \bullet 1$
845: is just the identity and hence invertible. But the only invertible
846: projection is the identity map itself. Thus
847: \begin{equation}
848:     \label{eq:FbulletEins}
849:     F \bullet 1 = F
850: \end{equation}
851: for all $F \in C^\infty(P)[[\lambda]]$, so the module structure
852: $\bullet$ is necessarily unital.
853: 
854: 
855: Finally, we call two deformation quantizations $\bullet$ and
856: $\tilde{\bullet}$ \emph{equivalent}, if there exists a $G$-equivariant
857: equivalence transformation between them, i.e. a formal series of
858: differential operators $T = \id + \sum_{r=1}^\infty \lambda^r T_r$ on
859: $C^\infty(P)[[\lambda]]$ such that
860: \begin{equation}
861:     \label{eq:EquivalenceG}
862:     T(F \bullet f) = T(F) \tilde{\bullet} f
863:     \quad
864:     \textrm{and}
865:     \quad
866:     g^* T = T g^*
867: \end{equation}
868: for all $F \in C^\infty(P)[[\lambda]]$, $f \in C^\infty(M)[[\lambda]]$
869: and $g \in G$.
870: 
871: 
872: We shall now discuss the existence and classification of such module
873: structures. For warming up we consider the situation of a
874: \emph{trivial} principal fiber bundle:
875: \begin{example}
876:     \label{example:TrivialPFB}
877:     Let $P = M \times G$ be the trivial (and trivialized) principal
878:     $G$-bundle over $M$ with the obvious projection. For any star
879:     product $\star$ on $M$ we can now extend $\star$ to $C^\infty(M
880:     \times G)[[\lambda]]$ by simply acting only on the $M$-coordinates
881:     in the Cartesian product. Here we use the fact that we can
882:     canonically extend multidifferential operators on $M$ to $M \times
883:     G$. Clearly, all algebraic properties are preserved whence in this
884:     case we even get a star product $\star_P = \star \otimes \mu$ with
885:     the undeformed multiplication $\mu$ for the $G$-coordinates. In
886:     particular, $C^\infty(M \times G)[[\lambda]]$ becomes a right
887:     module with respect to $\star$. So locally there are no
888:     obstructions even for the strongest requirement
889:     \eqref{eq:prAlgebraMorph} and hence also for \eqref{eq:BulletPFB}.
890: \end{example}
891: 
892: 
893: The problem of finding $\bullet$ is a global question whence we can
894: not rely on local considerations directly. The most naive way to
895: construct a $\bullet$ is an order-by-order construction: In general,
896: one has to expect obstructions in each order which we shall now
897: compute explicitly. This is a completely standard approach from the
898: very first days of algebraic deformation theory
899: \cite{gerstenhaber:1964a} and will in general only yield the result
900: that there are possible obstructions: in this case one needs more
901: refined arguments to ensure existence of deformations whence the
902: order-by-order argument in general is rather useless. In our situation,
903: however, it turns out that we are surprisingly lucky.
904: 
905: 
906: The following argument applies essentially to arbitrary algebras and
907: module deformations and should be considered to be folklore.  Suppose
908: we have already found $\varrho_0 = \pr^*$, $\varrho_1$, \ldots,
909: $\varrho_k$ such that 
910: \begin{equation}
911:     \label{eq:bulletk}
912:     F \bullet^{(k)} f 
913:     = F \pr^* f + \sum_{r=1}^k \lambda^r \varrho_r(F, f)
914: \end{equation}
915: is a module  structure up to order $\lambda^k$ and each $\varrho_r$
916: fulfills the $G$-equivariance condition. Then in order to find
917: $\varrho_{k+1}$ such that $\bullet^{(k+1)} = \bullet^{(k)} +
918: \lambda^{k+1} \varrho_{k+1}$ is a module structure up to order
919: $\lambda^{k+1}$ we have to satisfy
920: \begin{align}
921:     &\varrho_{k+1}(F, f) \pr^*g 
922:     - \varrho_{k+1}(F, fg) +
923:     \varrho_{k+1}(F \pr^*f, g)
924:     \nonumber \\
925:     &\qquad=
926:     \sum_{r=1}^k 
927:     \left(
928:         \varrho_r(F, C_{k+1-r}(f, g)) 
929:         - 
930:         \varrho_r(\varrho_{k+1-r}(F, f), g)
931:     \right)
932:     = R_k (F, f, g),
933:     \label{eq:varrhokpluseins}
934: \end{align}
935: for all $F \in C^\infty(P)[[\lambda]]$ and $f, g \in
936: C^\infty(M)[[\lambda]]$. Here $C_r$ denotes the $r$-th cochain of the
937: star product $\star$ as in \eqref{eq:StarProduct}. In order to
938: interpret this equation we consider the $\varrho_r$ as maps
939: \begin{equation}
940:     \label{eq:varrhorCochains}
941:     \varrho_r: C^\infty(M) \ni f
942:     \; \mapsto \;
943:     \varrho_r(\cdot, f) \in \Diffop(P)
944: \end{equation}
945: and similarly
946: \begin{equation}
947:     \label{eq:Rk}
948:     R_k: C^\infty(M) \times C^\infty(M) \ni (f, g)
949:     \; \mapsto \;
950:     R_k( \cdot, f, g) \in \Diffop(P).
951: \end{equation}
952: Viewing $\Diffop(P)$ as $C^\infty(M)$-bimodule via $\pr^*$ in the
953: usual way, we can now re-interpret \eqref{eq:varrhokpluseins} as
954: equation between a Hochschild one-cochain $\varrho_{k+1}$ and a
955: Hochschild two-cochain $R_k$
956: \begin{equation}
957:     \label{eq:deltavarrhokpluseins}
958:     \delta \varrho_{k+1} = R_k
959: \end{equation}
960: in the Hochschild (sub-)complex $\HCdiff^\bullet (C^\infty(M),
961: \Diffop(P))$ consisting of \emph{differential} cochains taking values
962: in the bimodule $\Diffop(P)$. Here $\delta$ is the usual Hochschild
963: differential. Using the assumption that the $\varrho_0, \ldots,
964: \varrho_k$ have been chosen such that $\bullet^{(k)}$ is a module
965: structure up to order $\lambda^k$ it is a standard argument to show
966: \begin{equation}
967:     \label{eq:deltaRkNull}
968:     \delta R_k = 0.
969: \end{equation}
970: Thus the necessary condition for \eqref{eq:deltavarrhokpluseins} is
971: always fulfilled by construction whence
972: \eqref{eq:deltavarrhokpluseins} is a cohomological condition: The
973: equation \eqref{eq:deltavarrhokpluseins} has solutions if and only if
974: the class of $R_k$ in the second Hochschild cohomology
975: $\HHdiff^2(C^\infty(M), \Diffop(P))$ is trivial.
976: 
977: 
978: In fact, we have also to take care of the $G$-equivariance of
979: $\varrho_{k+1}$. If all the $\varrho_0$, \ldots, $\varrho_k$ satisfy
980: the $G$-equivariance then it is easy to see that also $R_k$ has the
981: $G$-equivariance property. Thus we have to consider yet another
982: subcomplex of the differential Hochschild complex, namely
983: \begin{equation}
984:     \label{eq:HCdiffG}
985:     \HCdiff^\bullet (C^\infty(M), \Diffop(P)^G)
986:     \subseteq
987:     \HCdiff^\bullet (C^\infty(M), \Diffop(P)).
988: \end{equation}
989: Thus the obstruction for \eqref{eq:deltavarrhokpluseins} to have a
990: $G$-equivariant solution is the Hochschild cohomology class
991: \begin{equation}
992:     \label{eq:RkinHH}
993:     [R_k] \in \HHdiff^2 (C^\infty(M), \Diffop(P)^G).
994: \end{equation}
995: 
996: 
997: A completely analogous order-by-order construction shows that also the
998: obstructions for equivalence of two deformations $\bullet$ and
999: $\tilde{\bullet}$ can be formulated using the differential Hochschild
1000: complex of $C^\infty(M)$ with values in $\Diffop(P)^G$. Now the
1001: obstruction lies in the first cohomology $\HHdiff^1(C^\infty(M),
1002: \Diffop(P)^G)$.
1003: 
1004: 
1005: The following (nontrivial) theorem solves the problem of existence and
1006: uniqueness of deformation quantizations now in a trivial way
1007: \cite{bordemann.neumaier.waldmann.weiss:2007a:pre}:
1008: \begin{theorem}
1009:     \label{theorem:HHNull}
1010:     Let $\pr: P \longrightarrow M$ be a surjective submersion.
1011:     \begin{enumerate}
1012:     \item We have
1013:         \begin{equation}
1014:             \label{eq:HHdiffNull}
1015:             \HHdiff^k(C^\infty(M), \Diffop(P)) 
1016:             = 
1017:             \begin{cases}
1018:                 \Diffop_\ver(P) & \textrm{for} \; k=0 \\
1019:                 \{0\} & \textrm{for} \; k \ge 1.
1020:             \end{cases}
1021:         \end{equation}
1022:     \item If in addition $\pr: P \circlearrowleft G \longrightarrow M$
1023:         is a principal $G$-bundle then we have
1024:         \begin{equation}
1025:             \label{eq:HHdiffGNull}
1026:             \HHdiff^k(C^\infty(M), \Diffop(P)^G) 
1027:             = 
1028:             \begin{cases}
1029:                 \Diffop_\ver(P)^G & \textrm{for} \; k=0 \\
1030:                 \{0\} & \textrm{for} \; k \ge 1.
1031:             \end{cases}
1032:         \end{equation}
1033:     \end{enumerate}
1034: \end{theorem}
1035: The main idea is to proceed in three steps: first one shows that one
1036: can localize the problem to a bundle chart. For the local situation
1037: one can use the explicit homotopies from
1038: \cite{bordemann.et.al:2005a:pre} to show that the cohomology is
1039: acyclic. This is the most nontrivial part. By a suitable partition of
1040: unity one can glue things together to end up with the global
1041: statement. For a detailed proof we refer to
1042: \cite{bordemann.neumaier.waldmann.weiss:2007a:pre}.
1043: 
1044: 
1045: From this theorem and the previous considerations we obtain
1046: immediately the following result
1047: \cite{bordemann.neumaier.waldmann.weiss:2007a:pre}:
1048: \begin{corollary}
1049:     \label{corollary:DeformPFBOk}
1050:     For every principal $G$-bundle $\pr: P \circlearrowleft G
1051:     \longrightarrow M$ and any star product $\star$ on $M$ there
1052:     exists a deformation quantization $\bullet$ which is unique up to
1053:     equivalence.
1054: \end{corollary}
1055: In particular, the deformation for the trivial bundle as in
1056: Example~\ref{example:TrivialPFB} is the unique one up to equivalence.
1057: 
1058: 
1059: \begin{remark}
1060:     \label{remark:DeformationPFB}
1061:     \begin{enumerate}
1062:     \item It should be noted that the use of
1063:         Theorem~\ref{theorem:HHNull} gives existence and uniqueness
1064:         but no explicit construction of deformation quantizations of
1065:         principal bundles. Here the cohomological method is not
1066:         sufficient even though in
1067:         \cite{bordemann.neumaier.waldmann.weiss:2007a:pre} rather
1068:         explicit homotopies were constructed which allow to determine
1069:         further properties of $\bullet$.
1070:     \item In the more particular case of a symplectic Poisson
1071:         structure on $M$, Weiss used in his thesis \cite{weiss:2006a}
1072:         a variant of Fedosov's construction which gives a much more
1073:         geometric and explicit approach: there is a well-motivated
1074:         geometric input, namely a symplectic covariant derivative on
1075:         $M$ as usual for Fedosov's star products and a principal
1076:         connection on $P$. Out of this the module multiplication
1077:         $\bullet$ is constructed by a recursive procedure. The
1078:         dependence of $\bullet$ on the principal connection should be
1079:         interpreted as a global version of the Seiberg-Witten map
1080:         \cite{seiberg.witten:1999a}, now of course in a much more
1081:         general framework for arbitrary principal bundles, see also
1082:         \cite{barnich.brandt.grigoriev:2002a,
1083:           jurco.schraml.schupp.wess:2000a, jurco.et.al.:2001a}.
1084:     \item For the general Poisson case a more geometric construction
1085:         is still missing. However, it seems to be very promising to
1086:         combine global formality theorems like the one in
1087:         \cite{dolgushev:2005a} or the approach in
1088:         \cite{cattaneo.felder.tomassini:2002b} with the construction
1089:         \cite{weiss:2006a}. These possibilities will be investigated
1090:         in future works.
1091:     \end{enumerate}
1092: \end{remark}
1093: 
1094: 
1095: %
1096: % The commutant and associated bundles
1097: %
1098: 
1099: \section{The commutant and associated bundles}
1100: \label{sec:CommutantAssociated}
1101: 
1102: Theorem~\ref{theorem:HHNull} gives in addition to the existence and
1103: uniqueness of deformation quantizations of $P$ also a description of
1104: the \emph{differential commutant} of the right multiplications by
1105: functions on $M$ via $\bullet$: we are interested in those formal
1106: series $D = \sum_{r=0}^\infty \lambda^r D_r \in \Diffop(P)[[\lambda]]$
1107: of differential operators with the property
1108: \begin{equation}
1109:     \label{eq:Commutes}
1110:     D(F \bullet f) = D(F) \bullet f
1111: \end{equation}
1112: for all $F \in C^\infty(P)[[\lambda]]$ and $f \in
1113: C^\infty(M)[[\lambda]]$. In particular, if $D_0 = \id$ then
1114: \eqref{eq:Commutes} gives a \emph{self-equivalence}. Clearly, the
1115: differential commutant
1116: \begin{equation}
1117:     \label{eq:Kommutante}
1118:     \boldsymbol{\mathcal{K}} = 
1119:     \left\{
1120:         D \in \Diffop(P)[[\lambda]] 
1121:         \; \big| \; 
1122:         D \; \textrm{satisfies} \; \eqref{eq:Commutes}
1123:     \right\}
1124:     \subseteq \Diffop(P)[[\lambda]]
1125: \end{equation}
1126: is a subalgebra of $\Diffop(P)[[\lambda]]$ over
1127: $\mathbb{C}[[\lambda]]$.
1128: 
1129: 
1130: Note that there are other operators on $C^\infty(P)[[\lambda]]$ which
1131: commute with all right multiplications, namely the highly non-local
1132: pull-backs $g^*$ with $g \in G$. This was just part of the
1133: Definition~\ref{definition:DeformationPFB} of a deformation
1134: quantization of a principal bundle. However, in this section we shall
1135: concentrate on the differential operators with \eqref{eq:Commutes}
1136: only.
1137: 
1138: 
1139: Before describing the commutant it is illustrative to consider the
1140: classical situation. Here the commutant is simply given by the
1141: vertical differential operators
1142: \begin{equation}
1143:     \label{eq:ClassicalKommutant}
1144:     \Diffop_\ver(P) = 
1145:     \left\{
1146:         D \in \Diffop(P)
1147:         \; \big| \; 
1148:         D(F \pr^*f) = D(F) \pr^*f
1149:     \right\}
1150: \end{equation}
1151: by the very definition of vertical differential operators.
1152: Alternatively, the commutant is the zeroth Hochschild cohomology. More
1153: interesting is now the next statement which gives a quantization of
1154: the classical commutant, see
1155: \cite{bordemann.neumaier.waldmann.weiss:2007a:pre}.
1156: \begin{theorem}
1157:     \label{theorem:Commutant}
1158:     There exists a $\mathbb{C}[[\lambda]]$-linear bijection
1159:     \begin{equation}
1160:         \label{eq:varrhoprime}
1161:         \varrho': \Diffop_\ver(P)[[\lambda]] 
1162:         \longrightarrow
1163:         \boldsymbol{\mathcal{K}} 
1164:         \subseteq \Diffop(P)[[\lambda]]
1165:     \end{equation}
1166:     of the form
1167:     \begin{equation}
1168:         \label{eq:varrhoprimeExplicit}
1169:         \varrho' = \id + \sum_{r=1}^\infty \lambda^r \varrho'_r
1170:     \end{equation}
1171:     which is $G$-equivariant, i.e.
1172:     \begin{equation}
1173:         \label{eq:Gequivariantvarrhoprime}
1174:         g^* \varrho' = \varrho' g^*
1175:     \end{equation}
1176:     for all $g \in G$. The choice of such a $\varrho'$ induces an
1177:     associative deformation $\star'$ of $\Diffop_\ver(P)[[\lambda]]$
1178:     which is uniquely determined by $\star$ up to
1179:     equivalence. Finally, $\varrho'$ induces a left
1180:     $(\Diffop_\ver(P)[[\lambda]], \star')$-module structure $\bullet'$
1181:     on $C^\infty(P)[[\lambda]]$ via
1182:     \begin{equation}
1183:         \label{eq:bulletprimeDef}
1184:         D \bullet' F = \varrho'(D) F.
1185:     \end{equation}
1186: \end{theorem}
1187: The proof relies on an adapted symbol calculus for the differential
1188: operators $\Diffop(P)$: using an appropriate $G$-invariant covariant
1189: derivative $\nabla^P$ on $P$ which preserves the vertical distribution
1190: and a principal connection on $P$ one can induce a $G$-equivariant
1191: splitting of the differential operators $\Diffop(P)$ into the vertical
1192: differential operators and those differential operators which
1193: differentiate at least once in horizontal directions. Note that this
1194: complementary subspace has no intrinsic meaning but depends on the
1195: choice of $\nabla^P$ and the principal connection. A recursive
1196: construction gives the corrections terms $\varrho'_r(D)$ for a given
1197: $D \in \Diffop_\ver(P)$, heavily using the fact that the \emph{first}
1198: Hochschild cohomology $\HHdiff^1(C^\infty(M), \Diffop(P))$ vanishes.
1199: Since the commutant itself is an associative algebra the remaining
1200: statements follow.
1201: \begin{corollary}
1202:     \label{corollary:starprimeInvariant}
1203:     For the above choice of $\varrho'$ the resulting deformation
1204:     $\star'$ as well as the module structure are $G$-invariant, i.e. we
1205:     have
1206:     \begin{equation}
1207:         \label{eq:Ginvariance}
1208:         g^*(D \star' \tilde{D}) = g^*D \star' g^*\tilde{D}
1209:         \quad
1210:         \textrm{and}
1211:         \quad
1212:         g^*(D \bullet' F) = g^*D \bullet' g^*F
1213:     \end{equation}
1214:     for all $D, \tilde{D} \in \Diffop_\ver(P)[[\lambda]]$ and $F \in
1215:     C^\infty(P)[[\lambda]]$.
1216: \end{corollary}
1217: This follows immediately from the $G$-equivariance of $\bullet$ and
1218: the $G$-equivariance of $\varrho'$.
1219: \begin{remark}
1220:     \label{remark:Bicommutant}
1221:     A simple induction shows that the commutant of
1222:     $(\Diffop_\ver(P)[[\lambda]], \star')$ inside all differential
1223:     operators $\Diffop(P)[[\lambda]]$ is again
1224:     $(C^\infty(M)[[\lambda]], \star)$, where both algebras act by
1225:     $\bullet'$ and $\bullet$, respectively. This way
1226:     $C^\infty(P)[[\lambda]]$ becomes a $(\star', \star)$-bimodule such
1227:     that the two algebras acting from left and right are mutual
1228:     commutants inside all differential operators. Though this
1229:     resembles already much of a Morita context, it is easy to see that
1230:     $C^\infty(P)[[\lambda]]$ is \emph{not} a Morita equivalence
1231:     bimodule, e.g it is not finitely generated and projective.
1232:     However, as we shall see later, there is still a close relation to
1233:     Morita theory to be expected.
1234: \end{remark}
1235: \begin{remark}
1236:     \label{remark:BimoduleDeformation}
1237:     Note that classically $\pr^*: C^\infty(M) \longrightarrow
1238:     \Diffop(P)$ is an algebra homomorphism, too. Thus the questions
1239:     raised at the beginning of
1240:     Section~\ref{sec:DeformedPrincipalBundles} can now be rephrased as
1241:     follows: for a \emph{bimodule deformation} of $C^\infty(P)$ into a
1242:     bimodule over $C^\infty(M)[[\lambda]]$ equipped with possibly two
1243:     different star products for the left and right action, one has to
1244:     deform $\pr^*$ into a map
1245:     \begin{equation}
1246:         \label{eq:deformBimodule}
1247:         \boldsymbol{\pr^*}:
1248:         C^\infty(M)[[\lambda]] \longrightarrow
1249:         (\Diffop_\ver(P)[[\lambda]], \star')
1250:     \end{equation}
1251:     such that the image is a subalgebra. In this case, we can induce a
1252:     new product $\star'_M$ also for $C^\infty(M)[[\lambda]]$ making
1253:     $C^\infty(P)[[\lambda]]$ a bimodule for the two, possibly
1254:     different, star product algebras $(C^\infty(M)[[\lambda]],
1255:     \star_M')$ from the left and $(C^\infty(M)[[\lambda]], \star)$ from
1256:     the right. Note that this is the only way to achieve it since
1257:     $\star'$ is uniquely determined by $\star$. Thus it is clear that
1258:     we have to expect obstructions in the general case as there might
1259:     be no subalgebra of $(\Diffop_\ver(P)[[\lambda]], \star')$ which
1260:     is in bijection to $C^\infty(M)[[\lambda]]$. Even if this might be
1261:     the case, the resulting product $\star'_M$ might be inequivalent
1262:     to $\star$.  Note however, that we have now a very precise
1263:     framework for the question whether $\pr^*$ can be deformed into a
1264:     bimodule structure.
1265: \end{remark}
1266: \begin{remark}
1267:     \label{remark:DeftoDef}
1268:     As a last remark we note that changing $\star$ to an equivalent
1269:     $\tilde{\star}$ via an equivalence transformation $\Phi$ yields a
1270:     corresponding right module structure $\tilde{\bullet}$ by
1271:     \begin{equation}
1272:         \label{eq:equivalentstarbullet}
1273:         F \tilde{\bullet} f =  F \bullet \Phi(f),
1274:     \end{equation}
1275:     which is still unique up to equivalence by
1276:     Theorem~\ref{theorem:HHNull}. It follows that the commutants are
1277:     \emph{equal} (for this particular choice of $\tilde{\bullet}$)
1278:     whence the induced deformations $\star'$ and $\tilde{\star}'$
1279:     coincide. An equivalent choice of $\tilde{\bullet}$ would result
1280:     in an equivalent $\tilde{\star}'$. This shows that we obtain a
1281:     well-defined map
1282:     \begin{equation}
1283:         \label{eq:DeftoDef}
1284:         \Def(C^\infty(M)) 
1285:         \longrightarrow 
1286:         \Def(\Diffop_\ver(P))
1287:     \end{equation}
1288:     for the sets of equivalence classes of associative deformations.
1289:     In fact, the resulting deformations $\star'$ are even
1290:     $G$-invariant, whence the above map takes values in the smaller
1291:     class of $G$-invariant deformations $\Def_G(\Diffop_\ver(P))$.
1292: \end{remark}
1293: 
1294: 
1295: To make contact with the deformed vector bundles from
1296: Section~\ref{sec:MatterFieldsVectorBundles} we consider now the
1297: association process. Recall that on the classical level one starts
1298: with a (continuous) representation $\pi$ of $G$ on a
1299: finite-dimensional vector space $V$. Then the associated vector bundle
1300: is
1301: \begin{equation}
1302:     \label{eq:AssoVec}
1303:     E = P \times_G V \longrightarrow M,
1304: \end{equation}
1305: where the fibered product is defined via the equivalence relation $(p
1306: \cdot g, v) \sim (p, \pi(g) v)$ as usual. As the action of $G$ on $P$
1307: is proper and free, $E$ is a smooth manifold again and, in fact, a
1308: vector bundle over $M$ with typical fiber $V$. Rather tautologically,
1309: any vector bundle is obtained like this by association from its own
1310: frame bundle. For the sections of $E$ one has the canonical
1311: identifications
1312: \begin{equation}
1313:     \label{eq:SectionE}
1314:     \Gamma^\infty(E) \cong C^\infty(P, V)^G
1315: \end{equation}
1316: as right $C^\infty(M)$-modules, where the $G$-action of $C^\infty(P,
1317: V)$ is the obvious one.
1318: 
1319: 
1320: After this preparation it is clear how to proceed in the deformed
1321: case. From the $G$-equivariance of $\bullet$ we see that
1322: \begin{equation}
1323:     \label{eq:GammaE}
1324:     \Gamma^\infty(E)[[\lambda]] \cong C^\infty(P, V)^G[[\lambda]]
1325:     \subseteq C^\infty(P, V)[[\lambda]]
1326: \end{equation}
1327: is a $\star$-submodule with respect to the restricted module
1328: multiplication $\bullet$. It induces a right $\star$-module structure
1329: for $\Gamma^\infty(E)[[\lambda]]$ which we still denote by $\bullet$.
1330: This way we recover the deformed vector bundle as in
1331: Section~\ref{sec:MatterFieldsVectorBundles}.
1332: 
1333: 
1334: Moreover, we see that the $\End(V)$-valued differential operators
1335: $\Diffop(P) \otimes \End(V)$ canonically act on $C^\infty(P, V)$
1336: whence $\left((\Diffop_\ver(P) \otimes \End(V))[[\lambda]],
1337:     \star'\right)$ acts via $\bullet'$ on $C^\infty(P, V)[[\lambda]]$
1338: in such a way that the action commutes with the
1339: $\bullet$-multiplications from the right. By the $G$-invariance of
1340: $\star'$ we see that the invariant elements
1341: $\left(\Diffop_\ver(P)\otimes\End(V)\right)^G[[\lambda]]$ form a
1342: $\star'$-subalgebra which preserves (via $\bullet'$) the
1343: $\bullet$-submodule $C^\infty(P, V)^G[[\lambda]]$. Thus we obtain an
1344: algebra homomorphism
1345: \begin{equation}
1346:     \label{eq:MotherOfAll}
1347:     \left(
1348:         (\Diffop_\ver(P) \otimes \End(V))^G[[\lambda]], \star'
1349:     \right)
1350:     \longrightarrow
1351:     \left(
1352:         \Gamma^\infty(\End(E))[[\lambda]], \star'
1353:     \right)
1354: \end{equation}
1355: where $\star'$ on the left hand side is the deformation from
1356: Remark~\ref{remark:Matter}, part~\ref{item:EndEDeformation}.
1357: 
1358: 
1359: We conclude this section with some remarks and open questions:
1360: \begin{remark}
1361:     \label{remark:OpenQuestions}
1362:     \begin{enumerate}
1363:     \item The universal enveloping algebra valued gauge fields of
1364:         \cite{jurco.et.al.:2001a, jurco.schraml.schupp.wess:2000a} can
1365:         now easily be understood. For two vertical \emph{vector
1366:           fields} $\xi, \eta \in \Diffop_\ver(P)$ we have an action on
1367:         $C^\infty(P)[[\lambda]]$ via $\bullet'$-left multiplication.
1368:         In zeroth order this is just the usual Lie derivative
1369:         $\Lie_\xi$. Now the module structure says that
1370:         \begin{equation}
1371:             \label{eq:Ugvalued}
1372:             \xi \bullet' (\eta \bullet' F) 
1373:             - \eta \bullet' (\xi \bullet' F)
1374:             =
1375:             ([\xi, \eta]_{\star'}) \bullet' F
1376:         \end{equation}
1377:         for all $F \in C^\infty(P)[[\lambda]]$. Here $[\xi,
1378:         \eta]_{\star'} = \xi \star' \eta - \eta \star' \xi \in
1379:         \Diffop_\ver(P)[[\lambda]]$ is the $\star'$-commutator. In
1380:         general, this commutator is a formal series of vertical
1381:         differential operators but not necessarily a vector field any
1382:         more. Note that \eqref{eq:Ugvalued} holds already on the level
1383:         of the principal bundle.
1384:     \item For noncommutative gauge field theories we still need a good
1385:         notion of gauge fields, i.e. connection one-forms, and their
1386:         curvatures within our global approach. Though there are
1387:         several suggestions from e.g. \cite{jurco.schupp.wess:2000a} a
1388:         conceptually clear picture seems still to be missing.
1389:     \item In a future project we plan to investigate the precise
1390:         relationship between $(\Diffop_\ver(P)[[\lambda]], \star')$
1391:         and the Morita theory of star products
1392:         \cite{bursztyn.waldmann:2001a, bursztyn.waldmann:2002a,
1393:           bursztyn.waldmann:2000b}. Here \eqref{eq:MotherOfAll}
1394:         already suggests that one can re-construct all algebras Morita
1395:         equivalent to $(C^\infty(M)[[\lambda]], \star)$ out of
1396:         $\star'$.
1397:     \end{enumerate}
1398: \end{remark}
1399: 
1400: 
1401: %
1402: % Acknowledgment
1403: %
1404: 
1405: \subsection*{Acknowledgment}
1406: 
1407: It is a pleasure for me to thank the organizers Bertfried Fauser,
1408: Jürgen Tolksdorf and Eberhard Zeidler for their invitation to the very
1409: stimulating conference ``Recent Developments in Quantum Field
1410: Theory''. Moreover, I would like to thank Rainer Matthes for valuable
1411: discussions on Hopf-Galois extensions and Stefan Weiß for many
1412: comments on the first draft of the manuscript.
1413: 
1414: 
1415: %
1416: % references
1417: %
1418: 
1419: %\begin{footnotesize}
1420: %    \renewcommand{\arraystretch}{0.5} 
1421: %    \bibliographystyle{ewde}
1422: %    \bibliography{dqarticle,dqbook,dqprocentry,dqproceeding,preprints,misc,dqthesis,notes}
1423: %\end{footnotesize}
1424: 
1425: \begin{thebibliography}{10}
1426: 
1427: \bibitem {bahns.waldmann:2007a}
1428: {\sc Bahns, D., Waldmann, S.: }\newblock {\em Locally Noncommutative
1429:   Space-Times}.
1430: \newblock Rev. Math. Phys.  {\bf 19} (2007), 273--305.
1431: 
1432: \bibitem {barnich.brandt.grigoriev:2002a}
1433: {\sc Barnich, G., Brandt, F., Grigoriev, M.: }\newblock {\em Seiberg-{W}itten
1434:   maps and noncommutative {Y}ang-{M}ills theories for arbitrary gauge groups}.
1435: \newblock J. High Energy Phys.  .8 (2002), 23.
1436: 
1437: \bibitem {bayen.et.al:1978a}
1438: {\sc Bayen, F., Flato, M., Fr{{\o}}nsdal, C., Lichnerowicz, A., Sternheimer,
1439:   D.: }\newblock {\em Deformation Theory and Quantization}.
1440: \newblock Ann. Phys.  {\bf 111} (1978), 61--151.
1441: 
1442: \bibitem {bordemann.et.al:2005a:pre}
1443: {\sc Bordemann, M., Ginot, G., Halbout, G., Herbig, H.-C., Waldmann, S.:
1444:   }\newblock {\em Formalit{\'e} {$G_\infty$} adaptee et
1445:   star-repr{\'e}sentations sur des sous-vari{\'e}t{\'e}s co{\"{\i}}sotropes}.
1446: \newblock Preprint  {\bf math.QA/0504276} (2005), 56 pages.
1447: \newblock Extended version of the previous preprint math/0309321.
1448: 
1449: \bibitem {bordemann.neumaier.waldmann.weiss:2007a:pre}
1450: {\sc Bordemann, M., Neumaier, N., Waldmann, S., Weiss, S.: }\newblock {\em
1451:   Deformation quantization of surjective submersions and principal fibre
1452:   bundles}.
1453: \newblock Preprint   (2007).
1454: \newblock In preparation.
1455: 
1456: \bibitem {bursztyn.waldmann:2000b}
1457: {\sc Bursztyn, H., Waldmann, S.: }\newblock {\em Deformation Quantization of
1458:   Hermitian Vector Bundles}.
1459: \newblock Lett. Math. Phys.  {\bf 53} (2000), 349--365.
1460: 
1461: \bibitem {bursztyn.waldmann:2001a}
1462: {\sc Bursztyn, H., Waldmann, S.: }\newblock {\em Algebraic Rieffel Induction,
1463:   Formal Morita Equivalence and Applications to Deformation Quantization}.
1464: \newblock J. Geom. Phys.  {\bf 37} (2001), 307--364.
1465: 
1466: \bibitem {bursztyn.waldmann:2002a}
1467: {\sc Bursztyn, H., Waldmann, S.: }\newblock {\em The characteristic classes of
1468:   {M}orita equivalent star products on symplectic manifolds}.
1469: \newblock Commun. Math. Phys.  {\bf 228} (2002), 103--121.
1470: 
1471: \bibitem {bursztyn.waldmann:2005b}
1472: {\sc Bursztyn, H., Waldmann, S.: }\newblock {\em Completely positive inner
1473:   products and strong {M}orita equivalence}.
1474: \newblock Pacific J. Math.  {\bf 222} (2005), 201--236.
1475: 
1476: \bibitem {cattaneo.felder.tomassini:2002b}
1477: {\sc Cattaneo, A.~S., Felder, G., Tomassini, L.: }\newblock {\em From local to
1478:   global deformation quantization of {P}oisson manifolds}.
1479: \newblock Duke Math. J.  {\bf 115}.2 (2002), 329--352.
1480: 
1481: \bibitem {connes:1994a}
1482: {\sc Connes, A.: }\newblock {\em Noncommutative Geometry}.
1483: \newblock Academic Press, San Diego, New York, London, 1994.
1484: 
1485: \bibitem {dabrowski.grosse.hajac:2001a}
1486: {\sc Dabrowski, L., Grosse, H., Hajac, P.~M.: }\newblock {\em Strong
1487:   Connections and {C}hern-{C}onnes Pairing in the {H}opf-{G}alois Theory}.
1488: \newblock Commun. Math. Phys.  {\bf 220} (2001), 301--331.
1489: 
1490: \bibitem {dewilde.lecomte:1983b}
1491: {\sc DeWilde, M., Lecomte, P. B.~A.: }\newblock {\em Existence of Star-Products
1492:   and of Formal Deformations of the Poisson Lie Algebra of Arbitrary Symplectic
1493:   Manifolds}.
1494: \newblock Lett. Math. Phys.  {\bf 7} (1983), 487--496.
1495: 
1496: \bibitem {dito.sternheimer:2002a}
1497: {\sc Dito, G., Sternheimer, D.: }\newblock {\em Deformation quantization:
1498:   genesis, developments and metamorphoses}.
1499: \newblock In: {\sc Halbout, G. (eds.): }\newblock {\em Deformation
1500:   quantization}. \cite{halbout:2002a},   9--54.
1501: 
1502: \bibitem {dolgushev:2005a}
1503: {\sc Dolgushev, V.~A.: }\newblock {\em Covariant and equivariant formality
1504:   theorems}.
1505: \newblock Adv. Math.  {\bf 191} (2005), 147--177.
1506: 
1507: \bibitem {doplicher.fredenhagen.roberts:1995a}
1508: {\sc Doplicher, S., Fredenhagen, K., Roberts, J.~E.: }\newblock {\em The
1509:   Quantum Structure of Spacetime at the Planck Scale and Quantum Fields}.
1510: \newblock Commun. Math. Phys.  {\bf 172} (1995), 187--220.
1511: 
1512: \bibitem {gerstenhaber:1964a}
1513: {\sc Gerstenhaber, M.: }\newblock {\em On the Deformation of Rings and
1514:   Algebras}.
1515: \newblock Ann. Math.  {\bf 79} (1964), 59--103.
1516: 
1517: \bibitem {gutt:2000a}
1518: {\sc Gutt, S.: }\newblock {\em Variations on deformation quantization}.
1519: \newblock In: {\sc Dito, G., Sternheimer, D. (eds.): }\newblock {\em
1520:   Conf{\'e}rence Mosh{\'e} Flato 1999. Quantization, Deformations, and
1521:   Symmetries}, {\em Mathematical Physics Studies} no. {\bf 21},   217--254.
1522:   Kluwer Academic Publishers, Dordrecht, Boston, London, 2000.
1523: 
1524: \bibitem {gutt.rawnsley:1999a}
1525: {\sc Gutt, S., Rawnsley, J.: }\newblock {\em Equivalence of star products on a
1526:   symplectic manifold; an introduction to Deligne's {\v{C}}ech cohomology
1527:   classes}.
1528: \newblock J. Geom. Phys.  {\bf 29} (1999), 347--392.
1529: 
1530: \bibitem {hajac.matthes.szymanski:2003a}
1531: {\sc Hajac, P.~M., Matthes, R., Szyma{\'n}ski, W.: }\newblock {\em Chern
1532:   numbers for two families of noncommutative {H}opf fibrations}.
1533: \newblock C. R. Math. Acad. Sci. Paris  {\bf 336}.11 (2003), 925--930.
1534: 
1535: \bibitem {halbout:2002a}
1536: {\sc Halbout, G. (eds.): }\newblock {\em Deformation Quantization}, vol.~1 in
1537:   {\em IRMA Lectures in Mathematics and Theoretical Physics}.
1538: \newblock Walter de Gruyter, Berlin, New York, 2002.
1539: 
1540: \bibitem {heller.neumaier.waldmann:2007a}
1541: {\sc Heller, J.~G., Neumaier, N., Waldmann, S.: }\newblock {\em A
1542:   {$C^*$}-Algebraic Model for Locally Noncommutative Spacetimes}.
1543: \newblock Lett. Math. Phys.  {\bf 80} (2007), 257--272.
1544: 
1545: \bibitem {jurco.et.al.:2001a}
1546: {\sc Jur{\v{c}}o, B., M{\"o}ller, L., Schraml, S., Schupp, P., Wess, J.:
1547:   }\newblock {\em Construction of non-Abelian gauge theories on noncommutative
1548:   spaces}.
1549: \newblock Eur. Phys. J.  {\bf C21} (2001), 383--388.
1550: 
1551: \bibitem {jurco.schraml.schupp.wess:2000a}
1552: {\sc Jurc{\v{o}}, B., Schraml, S., Schupp, P., Wess, J.: }\newblock {\em
1553:   Enveloping algebra-valued gauge transformations for non-abelian gauge groups
1554:   on non-commutative spaces}.
1555: \newblock Eur. Phys. J. C Part. Fields  {\bf 17}.3 (2000), 521--526.
1556: 
1557: \bibitem {jurco.schupp:2000a}
1558: {\sc Jur{\v{c}}o, B., Schupp, P.: }\newblock {\em Noncommutative Yang-Mills
1559:   from equivalence of star products}.
1560: \newblock Eur. Phys. J.  {\bf C14} (2000), 367--370.
1561: 
1562: \bibitem {jurco.schupp.wess:2000a}
1563: {\sc Jur{\v{c}}o, B., Schupp, P., Wess, J.: }\newblock {\em Noncommutative
1564:   gauge theory for Poisson manifolds}.
1565: \newblock Nucl. Phys.  {\bf B584} (2000), 784--794.
1566: 
1567: \bibitem {jurco.schupp.wess:2001a}
1568: {\sc Jur{\v{c}}o, B., Schupp, P., Wess, J.: }\newblock {\em Nonabelian
1569:   noncommutative gauge theory via noncommutative extra dimensions}.
1570: \newblock Nucl.Phys.  {\bf B 604} (2001), 148--180.
1571: 
1572: \bibitem {kontsevich:2003a}
1573: {\sc Kontsevich, M.: }\newblock {\em Deformation Quantization of {P}oisson
1574:   manifolds}.
1575: \newblock Lett. Math. Phys.  {\bf 66} (2003), 157--216.
1576: 
1577: \bibitem {lance:1995a}
1578: {\sc Lance, E.~C.: }\newblock {\em {H}ilbert {$C^*$}-modules. A Toolkit for
1579:   Operator algebraists}, vol. 210 in {\em London Mathematical Society Lecture
1580:   Note Series}.
1581: \newblock Cambridge University Press, Cambridge, 1995.
1582: 
1583: \bibitem {nest.tsygan:1995a}
1584: {\sc Nest, R., Tsygan, B.: }\newblock {\em Algebraic Index Theorem}.
1585: \newblock Commun. Math. Phys.  {\bf 172} (1995), 223--262.
1586: 
1587: \bibitem {rosenberg:1996a:pre}
1588: {\sc Rosenberg, J.: }\newblock {\em Rigidity of K-theory under deformation
1589:   quantization}.
1590: \newblock Preprint  {\bf q-alg/9607021} (July 1996).
1591: 
1592: \bibitem {schmuedgen:1990a}
1593: {\sc Schm{\"{u}}dgen, K.: }\newblock {\em Unbounded Operator Algebras and
1594:   Representation Theory}, vol.~37 in {\em Operator Theory: Advances and
1595:   Applications}.
1596: \newblock Birkh{\"{a}}user Verlag, Basel, Boston, Berlin, 1990.
1597: 
1598: \bibitem {seiberg.witten:1999a}
1599: {\sc Seiberg, N., Witten, E.: }\newblock {\em String Theory and Noncommutative
1600:   Geometry}.
1601: \newblock J. High. Energy Phys.  {\bf 09} (1999), 032.
1602: 
1603: \bibitem {waldmann:2001b}
1604: {\sc Waldmann, S.: }\newblock {\em Deformation of Hermitian Vector Bundles and
1605:   Non-Commutative Field Theory}.
1606: \newblock In: {\sc Maeda, Y., Watamura, S. (eds.): }\newblock {\em
1607:   Noncommutative Geometry and String Theory}, vol. 144 in {\em Prog. Theo.
1608:   Phys. Suppl.},   167--175. Yukawa Institute for Theoretical Physics, 2001.
1609: \newblock Proceedings of the International Workshop on Noncommutative Geometry
1610:   and String Theory.
1611: 
1612: \bibitem {waldmann:2002b}
1613: {\sc Waldmann, S.: }\newblock {\em {M}orita equivalence of {F}edosov star
1614:   products and deformed {H}ermitian vector bundles}.
1615: \newblock Lett. Math. Phys.  {\bf 60} (2002), 157--170.
1616: 
1617: \bibitem {waldmann:2002a}
1618: {\sc Waldmann, S.: }\newblock {\em On the representation theory of deformation
1619:   quantization}.
1620: \newblock In: {\sc Halbout, G. (eds.): }\newblock {\em Deformation
1621:   quantization}. \cite{halbout:2002a},   107--133.
1622: 
1623: \bibitem {waldmann:2004a}
1624: {\sc Waldmann, S.: }\newblock {\em The Picard Groupoid in Deformation
1625:   Quantization}.
1626: \newblock Lett. Math. Phys.  {\bf 69} (2004), 223--235.
1627: 
1628: \bibitem {waldmann:2005a}
1629: {\sc Waldmann, S.: }\newblock {\em Morita Equivalence, {P}icard Groupoids and
1630:   Noncommutative Field Theories}.
1631: \newblock In: {\sc Carow-Watamura, U., Maeda, Y., Watamura, S. (eds.):
1632:   }\newblock {\em Quantum Field Theory and Noncommutative Geometry}, vol. 662
1633:   in {\em Lect. Notes Phys.},   143--155. Springer-Verlag, Berlin, Heidelberg,
1634:   2005.
1635: 
1636: \bibitem {waldmann:2007a}
1637: {\sc Waldmann, S.: }\newblock {\em Poisson-{G}eometrie und
1638:   {D}eformationsquantisierung. {E}ine {E}inf{\"u}hrung}.
1639: \newblock Springer-Verlag, Heidelberg, Berlin, New York, 2007.
1640: 
1641: \bibitem {weiss:2006a}
1642: {\sc Wei{\ss}, S.: }\newblock {\em Nichtkommutative {E}ichtheorien und
1643:   {D}eformationsquantisierung von {H}auptfaserb{\"u}ndeln}.
1644: \newblock master thesis, Fakult{\"{a}}t f{\"{u}}r Mathematik und Physik,
1645:   Physikalisches Institut, Albert-Ludwigs-Universit{\"{a}}t, Freiburg, 2006.
1646: \newblock Available at
1647:   {\texttt{http://idefix.physik.uni-freiburg.de/\~{}weiss/}}.
1648: 
1649: \bibitem {wess:misc}
1650: {\sc Wess, J.: }\newblock {\em Privat communication}.
1651: \newblock Discussion on noncommutativity as phase transition.
1652: 
1653: \end{thebibliography}
1654: 
1655: 
1656: \end{document}
1657: 
1658: %%% Local Variables: 
1659: %%% mode: latex
1660: %%% TeX-master: t
1661: %%% End: 
1662: