1: \documentstyle[12pt,epsf,epsfig,amsmath]{article}
2: \hoffset -0.2in
3: \textwidth 6in
4: \textheight 8.5in
5: %\setcounter{page}{1} `
6: \parskip 7pt \openup1\jot \parindent=0.5in
7: \topmargin -0.5in
8:
9: % the stuff below defines \eqalign and \eqalignno in such a
10: % way that they will run on Latex
11: \newfont{\thiplo}{msbm10 scaled\magstep 2}
12: \newfont{\gothic}{eufb10 scaled\magstep 2}
13: \newfont{\unc}{eurb10}
14: \newskip\humongous \humongous=0pt plus 1000pt minus 1000pt
15: \def\caja{\mathsurround=0pt}
16: \def\eqalign#1{\,\vcenter{\openup1\jot \caja
17: \ialign{\strut \hfil$\displaystyle{##}$&$
18: \displaystyle{{}##}$\hfil\crcr#1\crcr}}\,}
19: \newif\ifdtup
20: \def\panorama{\global\dtuptrue \openup1\jot \caja
21: \everycr{\noalign{\ifdtup \global\dtupfalse
22: \vskip-\lineskiplimit \vskip\normallineskiplimit
23: \else \penalty\interdisplaylinepenalty \fi}}}
24: \def\eqalignno#1{\panorama \tabskip=\humongous
25: \halign to\displaywidth{\hfil$\displaystyle{##}$
26: \tabskip=0pt&$\displaystyle{{}##}$\hfil
27:
28: \tabskip=\humongous&\llap{$##$}\tabskip=0pt
29: \crcr#1\crcr}}
30: % eqalignnoleft is eqalignno positioned flush left on the page
31: \def\eqalignnoleft#1{\panorama \tabskip=0pt
32: \halign to\displaywidth{\hfil$\displaystyle{##}$
33: \tabskip=0pt&$\displaystyle{{}##}$\hfil
34: \tabskip=\humongous&\llap{$##$}\tabskip=0pt
35: \crcr#1\crcr}}
36: % \eqright causes display equation material between \eqright and \cr
37: % to be positioned flush right on the page. This is useful
38: % in breaking long lines in a display equation.
39: % \eqright is usually used in conjunction with \eqalignnoleft.
40: \def\eqright #1\cr{\noalign{\hfill$\displaystyle{{}#1}$}}
41: % \eqleft causes display equation material between \eqleft and \cr
42: % to be positioned flush left on the page.
43: \def\eqleft #1\cr{\noalign{\noindent$\displaystyle{{}#1}$\hfill}}
44: % The oldref and fig macros are for formatting
45: % references and figure lists at the end of the paper.
46: % If you type \oldref{1}Dirac, P.A.M. you will get
47: % [1] Dirac, P.A.M.
48: % Same goes for \fig except you get Figure 2.1
49: \def\oldrefledge{\hangindent3\parindent}
50: \def\oldreffmt#1{\rlap{[#1]} \hbox to 2\parindent{}}
51: \def\oldref#1{\par\noindent\oldrefledge \oldreffmt{#1}
52: \ignorespaces}
53: \def\figledge{\hangindent=1.25in}
54: \def\figfmt#1{\rlap{Figure {#1}} \hbox to 1in{}}
55: \def\fig#1{\par\noindent\figledge \figfmt{#1}
56: \ignorespaces}
57: %
58: % This defines et al., i.e., e.g., cf., etc.
59: \def\ie{\hbox{\it i.e.}{}} \def\etc{\hbox{\it etc.}{}}
60: \def\eg{\hbox{\it e.g.}{}} \def\cf{\hbox{\it cf.}{}}
61: \def\etal{\hbox{\it et al.}}
62: \def\dash{\hbox{---}}
63: % common physics symbols
64: \def\tr{\mathop{\rm tr}}
65: \def\Tr{\mathop{\rm Tr}}
66: \def\partder#1#2{{\partial #1\over\partial #2}}
67: \def\secder#1#2#3{{\partial~2 #1\over\partial #2 \partial #3}}
68: \def\bra#1{\left\langle #1\right|}
69: \def\ket#1{\left| #1\right\rangle}
70: \def\VEV#1{\left\langle #1\right\rangle}
71: \def\ME#1#2{\left\langle #1\right|\left. #2 \right\rangle}
72: \def\gdot#1{\rlap{$#1$}/}
73: \def\abs#1{\left| #1\right|}
74: \def\pr#1{#1~\prime}
75: \def\ltap{\raisebox{-.4ex}{\rlap{$\sim$}} \raisebox{.4ex}{$<$}}
76: \def\gtap{\raisebox{-.4ex}{\rlap{$\sim$}} \raisebox{.4ex}{$>$}}
77: % \contract is a differential geometry contraction sign _|
78: \def\contract{\makebox[1.2em][c]{
79: \mbox{\rule{.6em}{.01truein}\rule{.01truein}{.6em}}}}
80: % The command \sectioneq produces numbering of equations by section
81: \def\holdtheequation{\arabic}
82: \def\sectioneq{\def\theequation{\thesection.\arabic{equation}}{\let
83: \holdsection=\section\def\section{\setcounter{equation}{0}\holdsection}}}%
84:
85: % The commands \beginletts and \endletts delimit sections of
86: % text in which successive equation numbers are distinguished by
87: % sequentially appending lower case letters
88: \newcounter{holdequation}\def
89: \beginletts{\begingroup\setcounter
90: {holdequation}{\value{equation}}\addtocounter
91: {equation}{1}\edef
92: \holdtheequation{\theequation}\setcounter
93: {equation}{0}\def
94: \theequation{\holdtheequation\alph{equation}}}
95: \def\endletts{\endgroup\setcounter
96: {equation}{\value{holdequation}}\refstepcounter{equation}}
97: % The command \num provides automatic numbering in LaTex when used in
98: % place of (equation number) in PlainTeX-style equations
99: \def\num{(\refstepcounter{equation}\theequation)}
100: % \auto is shorthand for \eqno\num
101: \def\auto{\eqno(\refstepcounter{equation}\theequation)}
102: % The commands \begineq and \endeq provide for one vertically
103: % centered automatic number for multiline equations
104: \def\begineq #1\endeq{$$ \refstepcounter{equation}\eqalign{#1}\eqno
105: (\theequation) $$}
106: % The command \contlimit puts (a\rightarrow0)
107: % under \longrightarrow
108: \def\contlimit{\,{\hbox{$\longrightarrow$}\kern-1.8em\lower1ex
109: \hbox{${\scriptstyle (a\rightarrow0)}$}}\,}
110: % The command \centeron#1#2 backs up #2 so that it is centered
111: % over #1. \centerover and \centerunder work like \centeron,
112: % except that they raise or lower #2 to place it over or under
113: % #1.
114: \def\centeron#1#2{{\setbox0=\hbox{#1}\setbox1=\hbox{#2}\ifdim
115: \wd1>\wd0\kern.5\wd1\kern-.5\wd0\fi
116: \copy0\kern-.5\wd0\kern-.5\wd1\copy1\ifdim\wd0>\wd1
117: \kern.5\wd0\kern-.5\wd1\fi}}
118: %
119: \def\centerover#1#2{\centeron{#1}{\setbox0=\hbox{#1}\setbox
120: 1=\hbox{#2}\raise\ht0\hbox{\raise\dp1\hbox{\copy1}}}}
121: %
122: \def\centerunder#1#2{\centeron{#1}{\setbox0=\hbox{#1}\setbox
123: 1=\hbox{#2}\lower\dp0\hbox{\lower\ht1\hbox{\copy1}}}}
124: % The commands \lsim and \gsim provide symbols for
125: % `less than of order' and `greater than of order'
126: \def\lsim{\;\centeron{\raise.35ex\hbox{$<$}}{\lower.65ex\hbox
127: {$\sim$}}\;}
128: \def\gsim{\;\centeron{\raise.35ex\hbox{$>$}}{\lower.65ex\hbox
129: {$\sim$}}\;}
130: % The command \st (for stroke) puts a slash through the succeeding
131: % character in math mode
132: \def\st#1{\centeron{$#1$}{$/$}}
133: % The command \newcases works like \cases except that
134: % the baselines and type size are the same as for
135: % display equations
136: \def\newcases#1{\left\{\,\vcenter{\normalbaselines\openup1\jot \caja
137: \ialign{\strut$\displaystyle{##}$\hfil
138: &\quad##\hfil\crcr#1\crcr}}\right.}
139: % The command \super inserts the characters in its argument as a
140: % superscript with the correct spacefactor.
141: \def\super#1{\ifmmode \hbox{\textsuper{#1}}\else\textsuper{#1}\fi}
142: \def\textsuper#1{\newcount\holdspacefactor\holdspacefactor=\spacefactor
143: $^{#1}$\spacefactor=\holdspacefactor}
144: % The command \supercite redefines \cite so that it makes superscripted
145: % citation numbers. It is to be used in conjunction with the
146: % \label command (for example, with one of the list-making
147: % environments). The command \oldcite restores the original LaTeX
148: % \cite command.
149: \let\holdcite=\cite
150: \def\supercite{\def\cite{\newcite}}
151: \def\oldcite{\def\cite{\holdcite}}
152: \def\newcite#1{\super{\newcount\citenumber\citenumber=0\getcite#1,@, }}
153: \def\getcite#1,{\advance\citenumber by1
154: \def\getcitearg{#1}\def\lastarg{@}
155: \ifnum\citenumber=1
156: \ref{#1}\let\next=\getcite\else\ifx\getcitearg\lastarg\let\next=\relax
157: \else ,\ref{#1}\let\next=\getcite\fi\fi\next}
158: % The command \nskip gives a vertical skip of the specified
159: % dimension (in braces) without including any extra \baselineskip
160: % or \parskip.
161: \def\nskip#1{\vbox{}\vskip-\baselineskip\vskip#1\vskip-\parskip\noindent}
162: % The command \lskip skips vertically by one line, i.e.,
163: % the current \baselineskip. There is no indentation unless
164: % \indent is specified.
165: \def\lskip{\vskip\baselineskip\vskip-\parskip\noindent}
166: \def\np{Nucl.\ Phys.\ }
167: \def\pr{Phys.\ Rev.\ }
168: \def\prl{Phys.\ Rev.\ Lett.\ }
169: \def\pl{Phys.\ Lett.\ }
170: \def\arnps{Ann.\ Rev.\ Nucl.\ Part.\ Sci.\ }
171: \def\mn{\mu\nu}
172: \def\epm{e^+e^-}
173: \def\pom{{\rm P\kern -0.53em\llap I\,}}
174: \def\spom{{\rm P\kern -0.36em\llap \small I\,}}
175: \def\sspom{{\rm P\kern -0.33em\llap \footnotesize I\,}}
176: \def\gev{{\rm GeV}}
177: \def\mev{{\rm MeV}}
178: \def\parens#1{\left(#1\right)}
179: \relax
180: \def\contlimit{\,{\hbox{$\longrightarrow$}\kern-1.8em\lower1ex
181: \hbox{${\scriptstyle (a\rightarrow0)}$}}\,}
182: \def\upon #1/#2 {{\textstyle{#1\over #2}}}
183: \relax
184: \renewcommand{\thefootnote}{\fnsymbol{footnote}}
185: \def\srf#1{$^{#1}$\ }
186: \def\mainhead#1{\setcounter{equation}{0}\addtocounter{section}{1}
187: \vbox{\begin{center}\large\bf #1\end{center}}\nobreak\par}
188: \sectioneq
189: \def\subhead#1{\bigskip\vbox{\noindent\bf #1}\nobreak\par}
190: \def\rf#1#2#3{{\bf #1}, #2 (19#3)}
191: \def\autolabel#1{\auto\label{#1}}
192: \def\til#1{\centeron{\hbox{$#1$}}{\lower 2ex\hbox{$\char'176$}}}
193: \def\tild#1{\centeron{\hbox{$\,#1$}}{\lower 2.5ex\hbox{$\char'176$}}}
194: \def\sumtil{\centeron{\hbox{$\displaystyle\sum$}}{\lower
195: -1.5ex\hbox{$\widetilde{\phantom{xx}}$}}}
196: \def\sumtilt{\sum^{\raisebox{-.15mm}{\hspace{-1.75mm}$\widetilde{}$}}\ }
197: \def\gltext{$\raisebox{1mm}{\centerunder{$\scriptscriptstyle
198: >$}{$\scriptscriptstyle <$}}$}
199: \def\intcent#1{\centerunder{$\displaystyle\int$}{\raisebox{-2.2mm}{$ #1 $}}}
200: \def\kbar{\underline{k}}
201: \def\qbar{\underline{q}}
202: \def\kbarsl{\underline{\st k}}
203: \def\qbarsl{\underline{\st q}}
204: \def\parens#1{\left(#1\right)}
205: \renewcommand{\arraystretch}{1.5}
206: \def\pbar{\underline{p}}
207: \def\pbarsl{\underline{\st p}}
208: \def\q{\unc q}
209: \def\p{\unc p}
210: \def\gam{\hat{\gamma}}
211: \def\kt{\hat{k}}
212: \def\pt{\hat{p}}
213: \def\qt{\hat{q}}
214: \def\Pit{\hat{\Pi}}
215: %-------------------------------------------------------------------
216:
217: \newcommand{\bit}{\begin{itemize}}
218: \newcommand{\eit}{\end{itemize}}
219: \newcommand{\cl}{\centerline}
220: \newcommand{\beq}{\begin{equation}}
221: \newcommand{\eeq}{\end{equation}}
222: \newcommand{\beqa}{\begin{eqnarray}}
223: \newcommand{\eeqa}{\end{eqnarray}}
224:
225:
226: %\pagestyle{empty}
227:
228: \begin{document}
229:
230: \begin{titlepage}
231:
232: \rightline{\vbox{\halign{&#\hfil\cr
233: &ANL-HEP-PR-03-95\cr
234: &\today\cr}}}
235: \vspace{0.25in}
236:
237: \begin{center}
238:
239: {\large\bf Electroweak High-Energy Scattering}
240:
241: {\large\bf and the Chiral Anomaly }\footnote{Work
242: supported by the U.S.
243: Department of Energy under Contract
244: W-31-109-ENG-38}
245:
246: \medskip
247:
248:
249: Alan. R. White\footnote{arw@hep.anl.gov }
250:
251: \vskip 0.6cm
252:
253: \centerline{Argonne National Laboratory}
254: \centerline{9700 South Cass, Il 60439, USA.}
255: \vspace{0.5cm}
256:
257: \end{center}
258:
259: \begin{abstract}
260:
261: The effect of perturbative QCD interactions
262: on the high energy scattering of electroweak vector bosons,
263: when the exchanged channel has pion quantum numbers, is considered.
264: The chiral anomaly is shown to appear in the couplings of particular
265: transverse momentum diagrams,
266: producing an enhancement of the scattering amplitude
267: by a power of the energy. At $O(\alpha_s)$
268: a single large transverse momentum gluon is involved and, within the transverse
269: momentum diagram framework, there is no cancelation.
270: In higher orders, soft gluons, carrying both normal and
271: anomalous color parity, are also present.
272:
273: The manipulation of a transverse momentum cut-off
274: to replace the ultra-violet anomaly divergence by
275: infra-red divergences that can lead to confinement and chiral symmetry breaking
276: is briefly discussed. Possible implications for electroweak symmetry breaking
277: are noted.
278:
279:
280: \end{abstract}
281:
282:
283: \renewcommand{\thefootnote}{\arabic{footnote}} \end{titlepage}
284:
285: %\setcounter{page}{2}
286:
287: \mainhead{1. INTRODUCTION}
288:
289: The cancelation of the chiral anomaly in the electroweak
290: sector of the Standard Model is crucial for the existence of the model as a
291: well-defined short-distance field theory. In perturbation theory, the anomaly
292: is a large momentum contribution in axial vector triangle diagrams
293: that, if uncanceled, destroys the renormalizability of a left-handed gauge
294: theory such as the electroweak sector of the Standard Model.
295: In the regge limit the feynman diagrams of standard perturbation theory
296: contract to form transverse momentum diagrams and produce
297: a new perturbation expansion that can be organized
298: into reggeon diagrams\cite{fkl}-\cite{arw93}. Beyond leading-log order,
299: the external couplings of the
300: transverse momentum diagrams (which are also the couplings of the reggeon
301: diagrams) contain contracted loop diagrams involving
302: ``effective vertices'' that result from the contraction. The
303: effective vertices can then produce new ``anomalies'', not present in the normal
304: perturbation expansion. For QCD, this is demonstrated in the next-to-leading log
305: calculation\cite{fl1} of gluon scattering, in which
306: an infra-red (gluon) triangle anomaly is responsible for the helicity
307: non-conservation that occurs, and in
308: our work on the contribution of infra-red quark loop anomalies to
309: pion/pomeron\cite{arw02} and triple pomeron\cite{arw021} vertices.
310:
311: In the electroweak scattering problem considered in this paper,
312: the underlying left-handed theory contains
313: elementary axial vector vertices. It is natural, therefore, that (components of)
314: these elementary vertices will appear also in the
315: regge limit effective vertices. A priori, therefore, large momentum contributions,
316: directly analagous to the familiar triangle anomaly, can be expected within the
317: loop diagrams that contribute to (beyond-leading-order) transverse momentum diagram
318: couplings. Indeed, as we shall see, internal effective vertices,
319: resulting from longitudinal vector meson exchange, also
320: appear which are quark current components that have point
321: interactions only at infinite momentum. Such current components
322: do not appear in the original lagrangian and do not couple to
323: leptons. Consequently, in the electroweak regge limit,
324: we can anticipate a significantly expanded ``anomaly problem'' which the well-known
325: short distance cancelations, between quarks and leptons,
326: will not be sufficient to remove.
327:
328: In this paper we will show that the triangle anomaly does indeed
329: appear in the couplings of
330: transverse momentum diagrams that describe the high-energy
331: scattering of $W^{\pm,0}$ vector mesons. All the diagrams we consider
332: describe the exchange of a quark-antiquark pair, with the flavor quantum numbers
333: of the pion, together with some number of gluons.
334: We choose pion quantum numbers because our ultimate goal
335: is to understand the relationship between the anomaly and chiral symmetry breaking
336: in the context of high-energy scattering.
337: Since two fermion exchange is involved, we would expect
338: the energy dependence to be, at most, logarithmic. (Pion exchange
339: would have no energy dependence.) The signal of the anomaly
340: will be a power divergence of transverse momentum integrals that
341: produces an additional power of the energy in the full amplitude.
342:
343: Since the anomaly phenomenon we discuss
344: involves longitudinal vector meson states it is natural
345: to expect that
346: the underlying gauge invariance of the electroweak theory will
347: be responsible for some form of cancelation. We discuss this possibility
348: at some length in an Appendix
349: at the end of the paper. While it appears that
350: the anomaly is not completely eliminated,
351: these identities do produce cancelations that
352: can not be straightforwardly expressed in terms of
353: transverse momentum diagram divergences. Moreover, it is clear that
354: the large transverse momentum region producing the anomaly
355: could also contribute in an important way within superficially non-leading
356: feynman diagrams. If there is finally a cancelation,
357: then it most likely means
358: the failure of the transverse momentum diagram formalism for
359: the electroweak theory unless (as, in any case, we strongly advocate) a
360: transverse momentum cut-off is imposed from the outset.
361: In the main body of the paper our purpose will be
362: to study contributions to transverse momentum diagrams and, apart from
363: the discussion in Appendix D, and the related discussion in sub-section
364: {\bf 5.3}, we
365: will make only brief references to the possibility
366: that there could be important contributions outside of the transverse
367: momentum diagram formalism.
368:
369: The anomaly occurs only in the even signature amplitude, which is a sum
370: of scattering amplitudes for vector mesons with opposite and same sign
371: helicities, i.e.
372: $$
373: \eqalign{A^+(S)~&=~A^+(P_+P_-)~=~A_{-+}(P_+,P_-)~+~ A_{++}(-P_+,P_-)\cr
374: &=~A_{-+}(S)~+~ A_{++}(-S)}
375: \auto\label{esa0}
376: $$
377: with
378: $$
379: ~A_{-+}(S)~\centerunder{$\to$}{\raisebox{-4mm}{${\scriptstyle S\to \infty}$}}
380: ~~c~ S~,~~~~~~
381: A_{++}(S)~\centerunder{$\to$}{\raisebox{-4mm}{${\scriptstyle S\to \infty}$}}~~-c~S
382: \auto\label{esa01}
383: $$
384: In a vector theory the amplitudes $A_{-+}(S)$ and $A_{++}(S)$ would simply add
385: in a single helicity amplitude and the anomaly would cancel.
386:
387: Our calculations are carried out in a theory that is very close to, but is
388: not quite, the Standard Model. For simplicity,
389: as we discuss further in Section 2, we ignore both
390: leptons and the photon and consider only one doublet of quarks. We
391: discuss the general framework for our analysis in Section 3
392: and isolate the simplest diagram,
393: which is $O(\alpha_s)$,
394: that potentially gives an enhancement. Section 4 is devoted to a detailed
395: demonstration that the enhancement does indeed occur in this diagram.
396: It is generated
397: by the combination of an effective vertex due to the left-handed coupling of
398: a scattering vector meson, a quark-antiquark effective vertex
399: due to a longitudinal
400: massive vector intermediate state, and a single (large transverse momentum) gluon
401: vertex.
402: At $O(\alpha_s)$, there is only a small
403: number of possibilities for the anomaly to occur within the
404: transverse momentum diagram formalism and, as we discuss in Section 5,
405: it is clear that it does not cancel.
406:
407: As we show in Section 6, at $O(\alpha_s^2)$ there are contributions
408: in which an additional soft gluon
409: plays no kinematical role in the occurrence of
410: the anomaly and simply accompanies the $O(\alpha_s)$ process.
411: Not surprisingly, the soft gluon
412: produces a transverse momentum infra-red
413: divergence in individual diagram contributions. In tracing the corresponding
414: cancelation, we find new (but closely related)
415: processes which occur first only at $O(\alpha_s)^2$ and for which
416: infra-red properties of the anomaly are needed to fully determine their
417: contribution.
418: Also at $O(\alpha_s^2)$, ``anomalous'' (odd) color charge
419: parity two gluon exchange appears, involving one finite and one large
420: transverse momentum gluon.
421:
422: In this paper we will frequently refer to a multigluon transverse
423: momentum state which carries color zero and anomalous color charge
424: parity (not equal to the gluon number) as ``anomalous gluons''.
425: In the present context, such a state first appears at $O(\alpha_s^3)$. Three
426: gluons with even color parity and (separately) large, finite, and soft
427: transverse momentum are involved. In higher orders
428: various combinations of soft and finite transverse momentum gluons can
429: accompany the large transverse momentum gluon. Additional gluons could also share
430: the large transverse momentum, but we do not discuss this possibility.
431:
432: Reggeized gluon exchanges that are the outcome of perturbative
433: calculations\cite{fkl}-\cite{arw93} in a vector theory
434: carry normal color parity (even/odd for
435: an even/odd number of gluons).
436: However, we have argued (for a very long time\cite{arw84}) that
437: anomalous gluons play a crucial role\cite{arw02}
438: in the emergence of QCD confinement and chiral symmetry breaking,
439: in the context of high-energy scattering and reggeon diagrams.
440: In particular, we have argued that the
441: pomeron is formed from anomalous gluons and that
442: configurations of this kind
443: are an essential component of regge limit pions and nucleons. However,
444: while we have been able to show how\cite{arw021}
445: anomalies provide triple pomeron and pion/pomeron interactions
446: involving the anomalous gluons,
447: it has proven very difficult to find a simple
448: starting point in which the anomalous gluons couple directly and
449: from which a detailed description of hadron
450: amplitudes can be developed. It is very encouraging,
451: therefore, to see that anomalous gluons appear straightforwardly in
452: the anomaly contributions that dominate the electroweak scattering
453: amplitudes we consider.
454:
455: We have not explored the full consequences of the
456: power enhancement (\ref{esa01}) nor, as we discussed above, is it clear that
457: the possibilities for cancelation have been exhausted.
458: While no unitarity bound is violated,
459: we, nevertheless, believe that the enhancement severely threatens the
460: unitarity of the theory (at least in the $t$ - channel)
461: and should not be present
462: in physical amplitudes. Rather than looking for further cancelations,
463: we will argue (only very briefly in this paper) that
464: although the enhancement is actually unphysical, it selects the
465: physically relevant diagrams and, in doing so, anticipates chiral symmetry
466: breaking and confinement. In fact there is no sign, in the anomaly
467: amplitudes that give (\ref{esa01}), of either the
468: $s$ - channel or the $t$ - channel intermediate states
469: that are present in the diagrams from which they are calculated.
470: In Section 7, we suggest that the enhancement is obtained by using
471: a ``wrong procedure'' to evaluate the regge limit
472: contribution of diagrams. If a transverse momentum cut-off is initially imposed,
473: the energy enhancement will be eliminated. Instead, because the cut-off
474: produces a Ward identity violation, the anomaly diagrams
475: dominate because of infra-red transverse momentum divergences that appear
476: and infra-red properties of the anomaly come into play. We will argue that
477: these divergences should be analysed, and ``physical amplitudes'' extracted,
478: before the cut-off is removed. This is emphasized as a major conclusion
479: of the paper in Section 8, which also contains other conclusions.
480:
481: Initially,
482: a study of the infra-red anomaly contributions of diagrams, that
483: matches the present study of ultra-violet contributions, will be required.
484: After this, we anticipate, the analysis of infra-red divergences will
485: parallel our discussion\cite{arw02} of hadron scattering.
486: All-orders properties of the
487: divergences have to be combined with Reggeon Field Theory,
488: to obtain the ``physical amplitudes'' in which the cut-off
489: can be removed. In this paper
490: we will describe only the general arguments
491: that we believe should be employed. We expect that
492: the resulting amplitudes will have both confinement
493: and chiral symmetry breaking, in the sense that the scattering
494: will be describable as the exchange of a color zero, Goldstone boson, pion.
495: Although our hadron work provides the framework for our
496: general understanding, a major part of
497: the logic and justification for the procedure we outline can be appreciated
498: directly within the present context, without reference to the pomeron problem.
499: That the starting point is much more straightforward
500: than in the hadron case holds out the promise that
501: it will be correspondingly easier to carry the procedure through in detail.
502:
503: \newpage
504:
505: \mainhead{2. THE ALMOST STANDARD MODEL}
506:
507: For simplicity, we will consider a theory which, for our purposes,
508: is sufficiently close to the Standard Model, but which is actually
509: less complex. We will consider
510: a ``flavor SU(2)'' triplet of vector mesons $\{W^+,W^-,W^0\}$ with mass $M$
511: and left-handed couplings to a flavor doublet of quarks $\{u,d\}$
512: with the usual QCD interaction. We will effectively assume that the vector
513: mesons originate from a spontaneously-broken gauge theory, as in the Standard
514: Model, but apart from the
515: discussion of reggeization in this Section, their self-interaction will play
516: almost no role in our analysis. (There will be no gauge dependence in our
517: discussion because the vector mesons will always be on-shell and gluons
518: will only contribute in gauge-independent transverse momentum diagrams - although,
519: in effect, we evaluate gluon contributions in feynman gauge.)
520:
521: We ignore the extra complications of the photon and all mixing angles,
522: which could only lessen the possibilities for cancelation of the anomaly
523: phenomenon that we find. Since there is no photon, the usual electroweak
524: ultraviolet anomaly is absent and so we do not need to include leptons.
525: In fact, as we noted in the Introduction, the anomaly we discuss involves
526: (components of) QCD currents to which leptons do not couple and, therefore, could
527: not provide any possibility for cancelation.
528: If we give the quarks a small mass
529: $m~(<< ~M)$ any potentially singular infra-red contributions will be eliminated.
530: However, for much of our discussion we will be interested only in large internal
531: (transverse) momenta, where ``large'' is defined relative to $M$, and so $m$ will
532: be omitted. The absence of a quark mass has the technical
533: advantage that we will be able to exploit the considerable, regge limit,
534: simplifications of the feynman diagrams that describe a massless, chiral, theory.
535:
536: We will study the high-energy
537: scattering of the massive vector mesons via a quark-antiquark
538: exchange channel in which, potentially, a ``pion'' could appear as a bound state.
539: Perturbatively, the leading behavior of the amplitudes we study would be given
540: (if there were no anomaly enhancement)
541: by the exchange of vector mesons. However,
542: since the flavor symmetry is
543: non-abelian, the vector mesons will be reggeized by self-interactions.
544: Since $CP$ is conserved, signature is well-defined and reggeized
545: vector meson exchange will give high-energy behavior in the odd-signature
546: channel of the form
547: $$
548: A(S,0)~~\centerunder{$\sim$}{\raisebox{-4mm}{${\scriptstyle S\to \infty}$}}
549: ~~S^{\alpha(0)}
550: \auto\label{rbe1}
551: $$
552: where
553: $$
554: \alpha(0) ~=~1 - \frac{g^2}{16\pi^2} ~+~O(g^4) ~~<~1
555: \auto\label{rbe11}
556: $$
557: The even signature channel will be dominated by the exchange of two
558: reggeized vector mesons for which (apart from a logarithmic factor)
559: $$
560: A(S,0)~~\centerunder{$\sim$}{\raisebox{-4mm}{${\scriptstyle S\to \infty}$}}
561: ~~S^{2\alpha(0)-1}~~<~~S^{\alpha(0)}
562: \auto\label{rbe2}
563: $$
564: Therefore, if we sum
565: (in principle at least) all diagrams producing all self-interaction
566: reggeization effects then
567: the contribution of (any number of) exchanged vector mesons to flavor
568: exchange amplitudes will be smaller than the anomaly enhanced
569: quark-antiquark exchange amplitudes we discuss which give
570: $$
571: A(S,0)~~\
572: \centerunder{$\sim$}{\raisebox{-4mm}{${\scriptstyle S\to \infty}$}}~~
573: \frac{S}{M^2}
574: \auto\label{fpb}
575: $$
576:
577: Since the running of $\alpha_s$ does not enter our calculations, the relative value
578: of $M$, compared to the QCD scale, does not appear in our discussion. Therefore,
579: the instability of the vector mesons is not an issue. More generally
580: it also should not be a very significant issue. We can, of course, define
581: vector meson scattering amplitudes by going to complex poles and any
582: undesirable features of these amplitudes will feed back into the scattering
583: amplitudes of physical particles. Alternatively, we could exploit the
584: reggeization property of vector mesons and, although
585: it would be very obscure
586: for most readers, we could carry out our discussion in terms of reggeon amplitudes.
587: In this case, it would be rather straightforward to argue that the non-regge
588: nature of the anomaly enhancement energy behavior that we find will violate
589: $t$ - channel unitarity. However, since the simplest
590: diagrams we consider are already very high-order
591: in the electroweak coupling (~$O(\alpha_w^4)$~) their contribution is very small
592: at current energies and so any
593: unitarity problems could only be of physical relevance
594: at extremely high energies.
595:
596: We could also regard our calculations as academic and say that we are simply using
597: left-handed vector mesons to uncover properties of QCD. From this point of
598: view we could make the mass $M$ as small as we like.
599: Also, throughout the main body of the paper the origin of $M$
600: will be irrelevant and we will implicitly assume that it
601: originates from some mechanism which
602: is unrelated to the quarks we consider. It is important that the high-energy
603: behavior (\ref{fpb}) is obtained with internal vector mesons on mass-shell.
604: Because longitudinal states are involved this behavior
605: is scaled by $M^2$ and so can not be canceled by a physical
606: Higgs contribution of any kind.
607:
608: If we ignore reggeization, or if the vector mesons
609: are massless, then the anomaly enhanced amplitudes
610: will give high-energy behavior
611: comparable with that of (multiple) vector meson exchange.
612: This could be an additional reason,
613: to be added to those briefly mentioned in Section 5,
614: why bound-states of (higher-colored) quarks and antiquarks
615: should actually be responsible for the vector meson mass generation. If the
616: electroweak scale $M$ is actually a second QCD scale, as would then be the case,
617: we would surely expect unitarity to be just as important for the higher scale
618: as for the lower scale.
619:
620: \newpage
621:
622: \mainhead{3. $O(\alpha_s)$ - ONE GLUON DIAGRAMS}
623:
624: In this Section we describe the general framework within which we discuss
625: $0(\alpha_s)$ diagrams and focus on the simplest feynman diagram that, potentially,
626: produces an anomaly enhancement.
627:
628: \subhead{3.1 Transverse Momentum Diagrams}
629:
630: As is very well known, the leading (regge limit)
631: high-energy behavior of a feynman diagram is typically obtained
632: by routing the large light-cone momenta through the diagram in such
633: a way that the number of particles that are close to mass-shell and have
634: large, relative, longitudinal momentum separations
635: (i.e. large rapidity differences) is maximal. After
636: longitudinal integrations are carried out, close to the on-shell
637: configuration, the result is a transverse momentum
638: integral multiplied by logarithms of the energy.
639: The transverse momentum integral corresponds to a ``transverse
640: momentum diagram'' obtained by contracting all of the (close to) on-shell lines.
641: In general, there is one logarithm and one transverse
642: momentum loop for each large rapidity difference. Consequently,
643: the leading-log amplitude contains a transverse momentum
644: diagram with the maximal number of loops.
645: In Appendix B we provide a brief, non-technical,
646: review\cite{rk} of known results that apply to
647: the fermion exchange scattering amplitudes we will discuss.
648:
649: The relationship between transverse momentum diagrams
650: and the process of putting lines on-shell
651: in full feynman diagrams will dominate our discussion.
652: When two, or more, particles have finite relative rapidity,
653: fewer lines are placed on-shell in the reduction to a transverse momentum diagram,
654: and a non-leading log amplitude, with a smaller number of
655: transverse momentum loops, is obtained. In this case,
656: the couplings and interactions in the transverse momentum diagrams have
657: more structure. It is in (superficially)
658: non-leading amplitudes of this kind that the high-energy
659: behavior can be enhanced by the occurence of the triangle anomaly within
660: the couplings of the transverse momentum diagram.
661:
662: \subhead{3.2 Double Logs}
663:
664: A well-known
665: extra complication, in the application of the transverse momentum diagram
666: formalism to fermion exchange amplitudes, is
667: that transverse loops involving only fermion
668: propagators are generally logarithmically divergent\cite{rk}
669: at large transverse momentum.
670: These divergences, effectively, produce additional logarithms of the energy
671: and give rise to ``double logs'' that are
672: associated with single rapidity differences.
673: In the diagrams we discuss there is, potentially, a logarithmic divergence
674: of this kind but it is overwhelmed by the anomaly power
675: divergence that we find. Therefore, we will not be directly interested in
676: logarithmic transverse momentum
677: divergences and, in the main body of the paper, will refer to them only
678: for reasons of completeness. In Appendix B
679: we briefly discuss the possible physical
680: relevance of the anomaly with respect to the double logs.
681:
682: From the general viewpoint of this paper, however,
683: it is important that, because we can regard
684: the double logs as described by transverse momentum diagrams,
685: they do not represent high-energy behavior that is not anticipated
686: by this formalism. (Even though it might not be the most efficient method
687: for studying properties of the double logs.)
688: It is, perhaps, worth noting that,
689: since the divergences do not occur in reggeization
690: diagrams, they do not affect the reorganization of transverse momentum diagrams
691: into reggeon diagrams. In fact, this reorganization reduces the degree of
692: divergence. The divergences occur only in reggeon diagram loops containing
693: just reggeized quarks and antiquarks and, if the leading log form of the
694: trajectory
695: function is used, the presence of the reggeon propagator reduces the divergence
696: from log to log[log] form.
697:
698: \subhead{3.3 The Enhanced Transverse Momentum Diagram }
699:
700: As we will elaborate below, the lowest-order appearance of the
701: anomaly enhancement is associated with the
702: transverse momentum diagram shown in Fig.~1.
703: {\begin{center}
704: $~$\hspace{1in}
705: \epsfxsize=2.2in
706: \epsffile{ehes01.ps}
707: \hspace{0.6in}
708: \parbox{1.4in}{\epsfxsize=1.3in
709: \epsffile{ehes00.ps}
710: }
711:
712: Fig.~1 A Transverse Momentum Diagram
713: \end{center}
714: The remainder of this Section, and the following two Sections,
715: will be devoted to the
716: study of $O(\alpha_s)$ feynman diagrams which give a contribution
717: to the high-energy scattering of vector mesons
718: that contains this transverse momentum integral.
719:
720: We will use the diagrammatic notation of Fig.~1 -
721: for quarks, gluons, and vector mesons -
722: throughout the paper, in both feynman diagrams and transverse momentum diagrams.
723: (Almost all of our discussion will be concerned with feynman diagrams
724: and so there should be no confusion as to which kind of diagram is under
725: consideration.) For simplicity, in this and the following two Sections,
726: we will omit flavor and color quantum numbers and consider just the momentum
727: and spin structure of diagrams. In this case, a ``gluon''is effectively
728: a ``photon'', i.e. a massless vector particle with a vector coupling to a massless
729: (for most of the discussion) ``quark-antiquark'' pair. A
730: ``vector meson'' is a massive vector particle with a left-handed (right-handed)
731: coupling to the quark (antiquark). Because of the left-handed coupling,
732: the high-energy scattering of vector bosons with definite helicity
733: has a particularly simple diagrammatic structure.
734:
735: To avoid the introduction of an extra momentum scale,
736: we will consider forward scattering, i.e. zero momentum transfer. We should
737: emphasize, however, that this does not imply that our calculations are
738: invalidated for the simple reason that we consider an infra-red region in
739: which perturbation theory does not apply. It should become clear that, since
740: the phenomenon we discuss involves large internal transverse momenta, a momentum
741: transfer $t$ with $M^2 <<t<<S$ would not significantly affect our analysis.
742: In the forward direction, the integrand of Fig.~1 is a product of the
743: couplings $G_L(k_{\perp},k_{\perp}')$ and $G_R(k_{\perp},k_{\perp}')$
744: and transverse momentum propagators for the gluon, quark, and antiquark and
745: the integral has the simple form
746: $$
747: \int ~d^2k_{\perp}'~\int d^2k_{\perp} ~\frac{
748: Tr\bigl\{~\st{k}_{\perp}~G_L(k_{\perp},k_{\perp}')~\st{k}_{\perp}~
749: G_R(k_{\perp},k_{\perp}')~\bigr\}}
750: {{k_{\perp}'}^2~(k_{\perp}^2)^2}
751: \auto\label{tqqg}
752: $$
753:
754: The $G_L$ and $G_R$
755: should satisfy (reggeon) Ward identities so that, as discussed further in Section
756: 5, there is no infra-red divergence at ${k_{\perp}'}^2=0$ and a divergence
757: at ${k_{\perp}}^2=0$ would be eliminated by
758: adding either a momentum transfer or a quark mass, as discussed in the previous
759: Section. Conventionally, since two fermion exchange is involved, we would expect
760: the accompanying energy dependence to be only logarithmic. We would
761: also expect the large momentum behavior of $G_L$ and $G_R$ to be such that
762: the full integral is, at worst, logarithmically divergent (producing an additional
763: energy logarithm, as discussed above).
764: The signal of the anomaly will be that $G_L$ and $G_R$
765: actually grow at large transverse momentum,
766: in a manner that produces an additional power of the energy.
767:
768: \subhead{3.4 A Feynman Graph Producing the Anomaly Enhancement}
769:
770: In the next Section we will see
771: that the feynman diagram shown in Fig.~2 has a regge limit
772: contribution, involving the transverse momentum diagram of Fig.~1,
773: in which each of the hatched lines is placed on-shell.
774: {\begin{center}
775: \epsfxsize=2.2in
776: \epsffile{ehes0.ps}
777:
778: Fig.~2 A Feynman Diagram - the Hatched Lines are On-Shell
779: \end{center}
780: In effect,
781: the box graphs at either end of the full graph contract to give
782: triangle diagram contributions to the couplings, $G_L$ and $G_R$,
783: as shown in Fig.~3.
784: {\begin{center}
785: \epsfxsize=3.8in
786: \epsffile{ehes2.ps}
787:
788: $~$ \newline
789: Fig.~3 A Triangle Diagram Coupling with ``Effective Vertices''
790: \end{center}
791: If the left-handed nature of the interactions of the
792: scattering vector particles leads to a left-handed (vector) ``effective
793: vertex'' for the triangle diagram then, naively, it would
794: appear that the triangle anomaly is obviously present.
795: If this has the standard form of the
796: ultra-violet triangle anomaly, we would expect
797: a linear growth with $k'_{\perp}$
798: that would then produce a divergence of the
799: $k'_{\perp}$ integration.
800: In fact, since
801: effective vertices are not necessarily
802: simple local vertices or, if they are, the propagators
803: may no longer be elementary, much more discussion
804: is required to show that there is a contribution that is closely related to
805: the triangle anomaly.
806:
807: The diagram of
808: Fig.~2 has the minimal complexity needed to generate triangle couplings,
809: as in Fig.~3, for both $G_L$ and $G_R$. We will find that this is
810: necessary to obtain a non-zero contribution in the full amplitude.
811:
812: \subhead{3.5 Leading Logs, Non-Leading Logs, and the Anomaly }
813:
814: A priori, as described in Appendix B, we expect
815: the leading high-energy behavior of Fig.~2 to be $[ln~s]^4$, multiplied by the
816: transverse momentum diagram obtained, as illustrated in Fig.~4, by placing
817: all vertical lines on-shell.
818: \begin{center}
819: \parbox{4in}{\epsfxsize=1.7in
820: \epsffile{ehes241.ps}
821: \epsfxsize=2in
822: \epsffile{ehes242.ps}
823: }
824: \parbox{1in}{
825: $$ \times~ [lns]^4~$$
826: }
827:
828: Fig.~4 The Leading-Log Amplitude
829: \end{center}
830: (A hatch on a line will always imply that it is on-shell.)
831: In fact, this transverse momentum diagram
832: contains quark loops that are logarithmically divergent and generate
833: additional powers of $ln s$, as discussed above.
834:
835: The transverse momentum diagram of Fig.~1 should appear
836: at the next-to-next-to-leading log level
837: (formally with a factor of $[ln S]^2$).
838: In this contribution, it would be anticipated that
839: the dominant internal momenta,
840: within the $G_L$ and $G_R$ couplings, will be ``close to''
841: that of the corresponding fast external particle (in particular,
842: there should be no large internal rapidity difference.)
843:
844: The expectation would be that large (relative) internal momenta
845: within the $G_L$ and $G_R$ couplings are
846: suppressed because of Ward identity cancelations that are
847: a consequence of gauge invariance for the gluon
848: (giving either a finite or, at worst, logarithmically divergent integral).
849: As we will see, the anomaly
850: contradicts this expectation in that it is
851: a contribution to the $G_L$ and $G_R$ couplings
852: from, relatively, large internal momentum
853: in which the unhatched vertical quark lines of Fig.~2 (and, correspondingly, the
854: unhatched vertical line of Fig.~3) are far off-shell.
855: This does not, however, imply the failure of a Ward identity.
856: Rather, as we enlarge on further in Section 5,
857: the presence of the anomaly means that large internal momenta
858: play an important role in the Ward identity. (In Appendix C
859: we review the corresponding situation for the vector Ward
860: identities in the familiar axial-vector/vector/vector triangle diagram in which
861: the anomaly occurs.)
862:
863: In the course of our analysis we will
864: find that, in the low-order diagrams where the anomaly first occurs,
865: there are no additional logarithms (multiplying the power enhancement)
866: associated with lines that are only close to on-shell,
867: rather than actually on-shell, in the large transverse momentum divergence.
868:
869: \newpage
870:
871: \mainhead{4. THE ANOMALY ENHANCEMENT}
872:
873: Is this Section we study, in detail, the occurrence of the anomaly enhancement
874: in the high-energy behavior of the feynman diagram of Fig.~2. This diagram
875: is shown again in Fig.~5, together with the momentum notation that we will use.
876: \begin{center}
877: \epsfxsize=3in
878: \epsffile{ehes1.ps}
879:
880: Fig.~5 Momentum Notation for the Diagram of Fig.~2
881: \end{center}
882: We will consider the limit
883: $$
884: P_+ ~ \to ~\bigl( \frac{\sqrt{S}}{2},\frac{\sqrt{S}}{2},0,0\bigr)
885: ~,~~~~~~~
886: P_- ~ \to ~\bigl(\frac{\sqrt{S}}{2},\frac{-\sqrt{S}}{2},0,0\bigr)
887: \auto\label{hel}
888: $$
889: and will find that the anomaly
890: is a simple pole, of the feynman integral, at $S=\infty$, which results
891: from the combination of the asymptotic pinching
892: of mass-shell propagator poles (those hatched in Fig.~2)
893: with the large momentum behavior of off-shell propagators. The on-shell
894: propagators will be used to carry out longitudinal momentum
895: integrations and produce a reduction to the transverse momentum integral
896: of Fig.~1. The large momentum behavior will be a combination of the
897: transverse momentum dependence of the exchanged propagators and the internal
898: loop momentum dependence of the propagators in the left and right side
899: triangle diagrams corresponding to Fig.~3.
900:
901:
902: \subhead{4.1 Internal Momenta and the Quark Mass-Shell Conditions }
903:
904: We consider, first, the left-hand box sub-graph that appears in Fig.~5.
905: As shown, we direct $P_+$ along the left-most quark line and
906: use the $k_-'$ integration to put this line on-shell, i.e.
907: $$
908: \int d k_-' ~~ \frac{i~ \gamma \cdot (P_+~ +~\cdots)}{
909: (k_-' + p_-')P_+~ + i\epsilon +~\cdots }
910: ~~~~\centerunder{$\to $}{\raisebox{-4mm}{${\scriptstyle P_+ \to \infty }$}}
911: ~~
912: \pi~\gamma_- ~+ ~\cdots
913: \auto\label{k'os}
914: $$
915: By using $k_-'$ for this purpose, we keep the $p'$ integration as a
916: four-dimensional integral
917: that we can anticipate will have an anomaly contribution from the large
918: momentum region, as
919: $P_+ \to \infty$.
920:
921: Using the $k_-'$ integration as in (\ref{k'os}) has, however, the disadvantage
922: that it introduces additional $p'$ dependence into two of the propagators
923: forming the triangle diagram. Consequently the triangle diagram no longer
924: has the elementary structure known to generate the anomaly. We will avoid this
925: problem by considering only a limited part of the $p'$ integration, i.e.
926: we consider the region where
927: the components of $p'$ have the order of magnitude
928: $$
929: |p'_+| ~\sim ~\epsilon~
930: S^{\frac{1}{2}}~<<~ P_+ ~,~~~~
931: {p'_{\perp}}^2~\sim ~\epsilon~M~ S^{\frac{1}{2}}~
932: <<~ MP_+ ~,~~~
933: | p'_- |~\sim ~\epsilon~ M
934: \auto\label{intr1}
935: $$
936: where $\epsilon$ is small, but finite. As we will see, this will allow us to
937: ignore the $k_-'$ dependence of triangle diagram propagators while simultaneously
938: keeping only the transverse momentum dependence of the exchanged propagators
939: and also allowing the anomaly spin structure to emerge as a large
940: $k_{\perp}$ approximation.
941:
942: We will use powers of $\epsilon$ as a
943: simple way to impose inequalities amongst momenta
944: that we could equally well impose more abstractly.
945: We will integrate over a range of
946: momenta having the given order of magnitude. Since we are only interested in showing
947: that an anomaly power enhancement occurs, and will make no attempt to determine the
948: coefficient multiplying it, the use of
949: powers of $\epsilon$ will be sufficient to carry
950: through our arguments. Note that
951: as we explicitly discuss later, if we
952: allowed $p'_-$ to be slightly larger, i.e. $|p'_-| \sim M$,
953: a Lorentz transformation on (\ref{intr1}) would give all components the same
954: order of magnitude, i.e. $p_i' \sim
955: (\epsilon M)^{\frac{1}{2}}S^{\frac{1}{4}}$. In this case, however,
956: the approximations we
957: make in the following would be more marginal, and more discussion of their
958: justification would be required. For simplicity, therefore, we keep
959: $|p'_-| \sim \epsilon M$, although we believe the full anomaly generating region
960: includes $|p'_-| \sim M$.
961:
962: If $k'_{\perp}$ is also large, but small
963: compared to $p'_{\perp}$, say
964: $$
965: {k'_{\perp}}^2 ~~ \sim~
966: ~ \epsilon^{\frac{3}{2}} ~M ~S^{\frac{1}{2}} ~~<<~ {p'_{\perp}}^2
967: \auto\label{intr2}
968: $$
969: then the mass-shell condition (\ref{k'os}) becomes
970: $$
971: |k_-' + p_-'|~ \sim ~ ~\frac{ {p'_{\perp}}^2}{P_+}
972: ~~\sim~~\frac{\epsilon~M~S^{\frac{1}{2}}}{S^{\frac{1}{2}}}~~
973: \sim ~~ \epsilon ~M
974: \auto\label{msc1}
975: $$
976: which, together with (\ref{intr1}), implies that
977: $$
978: |k'_-| \sim ~\epsilon~M
979: \auto\label{msc11}
980: $$
981:
982: If we similarly direct $P_-$ along the right-most quark line in Fig.~5 and
983: consider the analagous region of the $p''$ integration, then using $k'_+$
984: to put this line on-shell we will obtain a similar constraint on $k'_+$.
985: Together, these constraints ensure that, in the momentum region we are
986: considering, transverse momenta dominate the gluon propagator.
987: (\ref{intr1}) and (\ref{intr2}), together with the corresponding range for
988: $p''$, define a range of internal transverse momenta
989: that is growing with the external energy but which is, nevertheless, close
990: to mass-shell for the hatched quark lines.
991:
992: \subhead{4.2 Adjacent Quark Numerators and the External Effective Vertices }
993:
994: Consider, next, the contribution of the quark numerators
995: that are adjacent to the fast quark line. When combined with (\ref{k'os}),
996: the $\gamma_-$ components give zero. Therefore, the leading
997: contribution, in the momentum region we are considering,
998: is given by the transverse numerator components. To discuss this contribution
999: we use the complex $\gamma$ - matrix formalism\cite{kms} described in Appendix A.
1000: We can then write both numerators in the form
1001: $$
1002: \st{p}_{\perp}~=~\frac{1}{2}[~( \pt'+\kt')\gam^* ~+~(\pt'+\kt')^*\gam~]
1003: \auto\label{qn1}
1004: $$
1005: We take the vector meson states to be transversely polarized. The
1006: helicity of each vector meson will be conserved, but the two helicities
1007: can be equal or opposite.
1008: Using (\ref{prj3}) and (\ref{prj4}), the $(1-\gamma_5)$ vector meson coupling
1009: implies that, as illustrated in Fig.~A1, there is just one combination of
1010: $\gam$ and $\gam^*$ numerator matrices that can contribute for each helicity.
1011: For helicity $\lambda=-1$, the resulting coupling is that
1012: illustrated in Fig.~6(a), while $\lambda=+1$ gives that of Fig.~6(b).
1013: \begin{center}
1014: \epsfxsize=4.8in
1015: \epsffile{ehes30.ps}
1016:
1017: Fig.~6 Effective Vertices (a) for Helicity $\Lambda=-1$
1018: (b) for Helicity $\Lambda=+1$
1019: \end{center}
1020: (In all the figures of this kind we follow the
1021: normal convention and multiply $\gamma$ - matrices
1022: in the direction of the quark arrow.)
1023:
1024: \subhead{4.3 The Internal Vector Mesons}
1025:
1026: The on-shell contribution of the internal vector meson propagator is, of course,
1027: gauge independent and can be obtained from the unitary gauge propagator
1028: $$
1029: \Gamma_{\mu\nu}(p'+k)~=~ \frac{\bigl(~g_{\mu\nu}~-~
1030: (p'+k)_{\mu}(p'+k)_{\nu}/{M^2}~\bigr)}{(p'+k)^2 ~-~M^2}
1031: \auto\label{vpr}
1032: $$
1033: It is the second part of the on-shell numerator,
1034: corresponding to longitudinal polarization of the intermediate state, that
1035: produces\cite{arw02} the vector-like (cross-channel)
1036: coupling needed to obtain the anomaly. Since we are looking for a large
1037: momentum contribution of the $p'$ integration, we note that
1038: we obtain a factor of $p'_+$ in this part of the numerator
1039: if there is a $\gamma_-$ at one of the vertices. This is possible
1040: at the upper vertex in Fig.~7 (see below), but is not possible
1041: at the lower vertex, since a $\gamma_-$ at the bottom vertex
1042: would (anti-)commute through the adjacent transverse numerator and give zero.
1043:
1044: As we will see, we obtain the factor of $p'_+$, in combination with the spin
1045: structure needed to obtain the anomaly in the resulting triangle diagram,
1046: if $k_{\perp}^2~>> ~ {p'_{\perp}}^2~$. Since we have already
1047: imposed (\ref{intr1}), we can achieve this by taking
1048: $$
1049: k_{\perp}^2 ~~\sim ~~\epsilon^{\frac{1}{2}}~M~S^{\frac{1}{2}}~~
1050: \sim ~\epsilon^{-1/2}~ {p'_{\perp}}^2
1051: \auto\label{intr3}
1052: $$
1053: The dominant contribution of the vector meson numerator is then
1054: as shown in Fig.~7.
1055: (We remove the $(1-\gamma_5)$ factors in the quark couplings of the internal
1056: vector meson by (anti-)commuting them around the quark loop.)
1057: \begin{center}
1058: \epsfxsize=5.5in
1059: \epsffile{ehes5.ps}
1060:
1061: Fig.~7 The Internal Vector Meson Numerator
1062: \end{center}
1063: The $\otimes$ notation indicates that the two $\gamma$-matrices are
1064: not multiplied.
1065:
1066: With $k_{\perp}^2~>> ~ {p'_{\perp}}^2~$,
1067: putting the vector meson propagator on-shell via the $k_-$ integration gives
1068: $$
1069: \int d k_- ~~ \frac{(\kt \gam^*~\otimes ~
1070: i~p'_+ \gamma_- ~ +~\cdots)/M^2}{ [(k_- + p_-')p'_+ ~\cdots] }
1071: ~~~~\centerunder{$\sim $}{\raisebox{-4mm}{$p'_+ \to \infty $}}
1072: ~~\pi ~\kt \gam^* ~\otimes
1073: ~\gamma_- /M^2
1074: \auto\label{kos}
1075: $$
1076: and, if (\ref{intr3}) is satisfied, the mass-shell constraint gives
1077: $$
1078: |k_-| ~~ \sim ~~ \frac{k_{\perp}^2}{|p_+'|} ~~
1079: \sim ~~\frac{\epsilon^{\frac{1}{2}}~M~S^{\frac{1}{2}}}
1080: {\epsilon~S^{\frac{1}{2}}}~~\sim~~
1081: \epsilon^{-1/2}~M
1082: \auto\label{msc2}
1083: $$
1084: The parallel discussion of the on-shell contribution of the right-side internal
1085: vector meson will give a corresponding constraint on $|k_+|$.
1086: The two constraints,
1087: taken together with (\ref{intr3}), imply that (as for the gluon)
1088: transverse momenta dominate
1089: the central quark and antiquark propagators. As we noted,
1090: ensuring that $k_-$ remains
1091: finite, with (\ref{intr3}) satisfied, provides part of the motivation
1092: for the the initial choice of (\ref{intr1}) as the $p'$ integration region.
1093:
1094: Note that,
1095: although we consider very large transverse momenta, because the vector mesons
1096: remain on-shell the high-energy behavior we will find will be scaled by $M^2$.
1097: Consequently, there is no possibility that it could be canceled by the
1098: contribution of a Higgs particle. (That is, if the Higgs mechanism were
1099: used to generate the vector meson mass.)
1100:
1101: \subhead{4.4 The Internal Quark Numerator and the Triangle Amplitude}
1102:
1103: For the remaining components of Fig.~3 that we have not yet discussed,
1104: the largest contribution (that also gives the $\gamma_-$ vertex as in Fig.~7)
1105: is obtained by taking the gluon coupling to be $\gamma_+$
1106: and taking the remaining quark numerator to also be transverse.
1107: The chirality then feeds through the propagator as illustrated in Fig.~8.
1108: \begin{center}
1109: \epsfxsize=2.8in
1110: \epsffile{ehes6.ps}
1111:
1112: Fig.~8 The Internal Quark Numerator
1113: \end{center}
1114:
1115: Combining Figs.~6, 7, and 8, and using (\ref{kos})
1116: we obtain an effective triangle diagram with the numerators
1117: and vertices shown in Fig.~9.
1118: In this figure we have also included the transverse quark and antiquark
1119: propagators ($\gam/\kt$ and $\gam^*/\kt^*$) that are external
1120: to the triangle diagram.
1121: \begin{center}
1122: \epsfxsize=1.9in
1123: \epsffile{ehes8.ps}
1124:
1125: Fig.~9 The Effective Triangle Diagram
1126: \end{center}
1127: The amplitude obtained from Fig.~9
1128: is (apart from an overall constant that we neglect - involving a numerical factor,
1129: factors of $\pi$, and powers of the coupling constants)
1130: $$
1131: \frac{\gam}{\kt}~G_L(k_{\perp},k_{\perp}')~\frac{\gam^*}{\kt^*}~~\sim~~
1132: \frac{\gam}{\kt}~\biggl(~\int d^4p'~
1133: \frac{\kt\gam^* ~({\pt}'+{\kt}')^*\gam
1134: ~\gamma_-~ ({\pt}'+ {\kt}')
1135: ~\gam^* ~ \gamma_+[\pt']^*\gam}{
1136: (p'+k')^4~(p')^2}~\frac{ \gamma_- }{M^2}\biggr)~\frac{\gam^*}{\kt^*}
1137: \auto\label{bxi}
1138: $$
1139: with the integration region specified by (\ref{intr1}).
1140:
1141: With (\ref{intr1})-(\ref{msc1}) satisfied, we can make the
1142: approximations
1143: $$
1144: \eqalign{|k'_+(p'_- +k'_-)| ~~&\sim ~~\epsilon^2~M^2~, \cr
1145: |p'_+(p'_- +k'_-)| ~~ &\sim
1146: ~~{\epsilon}^2~ M~S^{\frac{1}{2}}~,\cr
1147: ~~(p'+k')_{\perp}^2~~
1148: &\sim~~\bigl(\epsilon ~+~O(\epsilon^{\frac{5}{4}})\bigr)~M~S^{\frac{1}{2}}
1149: }
1150: \auto\label{asb1}
1151: $$
1152: and so
1153: $$
1154: (p'+k')^2 ~~\sim ~~(p'+k')_{\perp}^2
1155: \auto\label{asb2}
1156: $$
1157: Consequently, we can approximate (\ref{bxi}) as (again ignoring an overall constant)
1158: $$
1159: \frac{\gam}{\kt}~G_L(k_{\perp},k_{\perp}')~\frac{\gam^*}{\kt^*}~~\sim~
1160: ~\frac{\gam}{\kt}~~\biggl(~\int~ d^4p' ~\frac{\kt[\pt']^* }{
1161: [p'_{\perp}+k'_{\perp}]^2 ~[2p'_+p'_- - (p'_{\perp})^2 ]}~
1162: \frac{\gamma_-}{M^2} ~\biggr)
1163: ~~\frac{\gam^*}{\kt^*}
1164: \auto\label{bxi1}
1165: $$
1166: with the integration region still specified by (\ref{intr1}).
1167:
1168: Because we have eliminated $k_-'$ and $k_+'$ (which are functions
1169: of $p'$ and $p''$), the amplitude (\ref{bxi1}) is (almost) that of
1170: a conventional triangle feynman diagram - with local vertices. Achieving the
1171: elimination of $k_-'$ and $k_+'$ from (\ref{bxi1}) was, as we remarked,
1172: part of the motivation
1173: for the initial restriction to the momentum region (\ref{intr1})-(\ref{msc1}).
1174:
1175: \subhead{4.5 The Anomaly Contribution}
1176:
1177: Using (\ref{cta1}) we can write the numerator momentum factor of (\ref{bxi1}) as
1178: $$
1179: \kt [\pt']^*~=~k_{\perp}\cdot p_{\perp}'~+~i~k_{\perp}\times p_{\perp}'
1180: \auto\label{cta11}
1181: $$
1182: The first term is not special to a vector vertex
1183: fermion triangle diagram and is not related to the anomaly.
1184: We expect it to arise from (and to eventually
1185: be canceled by) a variety of contributions to the complete transverse
1186: momentum couplings of Fig.~1.
1187: It is the second term in (\ref{cta11}) that we expect to give an anomaly
1188: contribution. It's parity properties result directly
1189: from the product of an odd number of quark numerators and so we anticipate
1190: that it can only be canceled by effective triangle diagrams that contain
1191: three quark propagators.
1192:
1193: Keeping just the second term in (\ref{bxi1}) gives,
1194: for the integral within the brackets (apart from the factor of $\gamma_-/M^2$)
1195: $$
1196: \int~dp_+' dp_-' \int d^2p_{\perp}'
1197: ~\frac{i~k_{\perp}\times p_{\perp}' }{
1198: [p_{\perp}'+k'_{\perp}]^2 ~[2p_+'p_-' - (p_{\perp}')^2 ]}
1199: \auto\label{bxi3}
1200: $$
1201: To carry out the angular integration for $p_{\perp}'$
1202: we choose co-ordinates $(p_2',p_3')$
1203: such that $k_{\perp}'$ lies along the 2-axis. In this case,
1204: $$
1205: p_{\perp}'\cdot k_{\perp}' ~=~ |p_{\perp}'|~|k_{\perp}'|~ cos\phi~,
1206: ~~~~~p_{\perp}'\times k_{\perp}~=~|p_{\perp}'|~(k_3 cos\phi ~-~ k_2 sin \phi)
1207: \auto\label{bxi4}
1208: $$
1209: where $k_2$ and $k_3$ are projections of $k_{\perp}$ along and perpendicular
1210: to $k'_{\perp}$. We can, therefore, write
1211: $$
1212: \eqalign{\int~&dp_+' dp_-' \int \frac{d(p_{\perp}')^2}
1213: {[2p_+'p_-' - (p_{\perp}')^2 ]}
1214: ~\frac{k_{\perp}\times p_{\perp}' }{
1215: [p_{\perp}'+k'_{\perp}]^2 } \cr
1216: &= ~ \int~dp_+' dp_-' \int \frac{d(p_{\perp}')^2}
1217: {[2p_+'p_-' - (p_{\perp}')^2 ]}
1218: \int_0^{2\pi}d\phi~ \frac{|p_{\perp}'|(k_3cos\phi -k_2sin\phi) }
1219: {[(p_{\perp}')^2 + (k_{\perp}')^2+2 |p_{\perp}'||k'_{\perp}|~cos\phi]}}
1220: \auto\label{bxi30}
1221: $$
1222:
1223: Using
1224: $$
1225: \int_0^{2\pi}d\phi ~\frac{sin\phi}{a+b~cos\phi} ~
1226: =~-~ \frac{1}{b}~ \bigl[~a+b~cos \phi ~\bigr]_0^{2\pi}
1227: ~=~0
1228: \auto\label{bxi5}
1229: $$
1230: and
1231: $$
1232: \eqalign{\int_0^{2\pi}d\phi ~\frac{cos\phi}{a+b~cos\phi} ~~
1233: &\centerunder{$\sim$}{\raisebox{-4mm}{${\scriptstyle b~ <<~ a}$}}
1234: ~~\frac{1}{a}~\int_0^{2\pi}d\phi ~cos\phi~-~
1235: \frac{b}{a^2}\int_0^{2\pi}d\phi ~cos^2\phi ~+
1236: ~O\bigl(~\bigl[\frac{b^2}{a^3}\bigr]~\bigr)\cr
1237: &\sim 0~-~\frac{\pi b}{a^2}~+ ~O\bigl(~\bigl[\frac{b^2}{a^3}\bigr]~\bigr) }
1238: \auto\label{bxi6}
1239: $$
1240: we obtain
1241: $$
1242: \int d\phi
1243: ~\frac{k_{\perp}\times p_{\perp}' }{
1244: [p_{\perp}'+k'_{\perp}]^2 }
1245: ~~~\centerunder{$\sim$}{\raisebox{-4mm}{${\scriptstyle {k_{\perp}'}^2 <<
1246: ~ {p_{\perp}'}^2}$}}~~
1247: \frac{k_3~|k'_{\perp}|}{{p_{\perp}'}^2}
1248: ~~\sim~~\frac {k_{\perp}\times k_{\perp}'}{{p_{\perp}'}^2 }
1249: \auto\label{bxi7}
1250: $$
1251: (\ref{bxi30}) gives, therefore,
1252: $$
1253: k_{\perp}\times k_{\perp}' ~~
1254: \centerunder{\hbox{\huge $\int$}}{\raisebox{-4mm}{${\scriptstyle
1255: {p_{\perp}'}^2 \sim~ \epsilon M \sqrt{S}}$}}
1256: ~~\frac {d({p_{\perp}'}^2)} {({p_{\perp}'}^2)} ~~
1257: \centerunder{\hbox{\huge $\int$}}{\raisebox{-4mm}{${\scriptstyle
1258: |p_+'| \sim \epsilon \sqrt{S}~,~|p_-'| \sim ~\epsilon~ M}$}}
1259: ~~\frac{ dp_+'
1260: dp_-'}{[2p_+'p_-' - (p_{\perp}')^2 ]}
1261: \auto\label{bxi8}
1262: $$
1263: If we change variables to
1264: $$
1265: \sqrt{S}~x~=~{p_{\perp}'}^2~, ~~~~~~~ \sqrt{S}~y~=~p_+'
1266: \auto\label{scS}
1267: $$
1268: we obtain
1269: $$
1270: \eqalign{ \centerunder{\hbox{\huge $\int$}}{\raisebox{-4mm}{${\scriptstyle
1271: {p_{\perp}'}^2 \sim~ \epsilon M \sqrt{S}}$}}
1272: ~~&\frac {d({p_{\perp}'}^2)} {({p_{\perp}'}^2)} ~~
1273: \centerunder{\hbox{\huge $\int$}}{\raisebox{-4mm}{${\scriptstyle
1274: |p_+'| \sim~ \epsilon \sqrt{S},~|p_-'| \sim~ \epsilon M
1275: }$}}
1276: ~~\frac{ dp_+'
1277: dp_-'}{[2p_+'p_-' - (p_{\perp}')^2 ]} \cr
1278: &=~\centerunder{\hbox{\huge $\int$}}{\raisebox{-4mm}{${\scriptstyle
1279: x \sim~ \epsilon M }$}}
1280: ~~\frac {dx} {x} ~~
1281: \centerunder{\hbox{\huge $\int$}}{\raisebox{-4mm}{${\scriptstyle
1282: |y| \sim~ \epsilon ,~|p_-'| \sim~\epsilon M }$}}
1283: ~~\frac{ dy
1284: dp_-'}{[2yp_-' - x ]} }
1285: \auto\label{sc1}
1286: $$
1287: which is clearly a constant -
1288: that we do not need to evaluate.
1289:
1290: (\ref{bxi8}) is sufficient
1291: to conclude that the integration region on which we have focussed gives,
1292: for (\ref{bxi1}), the behavior
1293: $$
1294: \sim ~~\frac{\gam^*}{\kt^*}~~\biggl(~\frac {(k_{\perp}\times k_{\perp}')~
1295: \gamma_-}{M^2} ~\biggr)
1296: ~~\frac{\gam}{\kt}
1297: \auto\label{bxi11}
1298: $$
1299:
1300: \subhead{4.6 Behavior of the Full Amplitude}
1301:
1302: It is straightforward to obtain the behavior of
1303: the full amplitude that results from combining (\ref{bxi11})
1304: with the corresponding contribution from the right-side box graph in Fig.~5.
1305: As we have discussed above, the internal
1306: mass-shell conditions determine that
1307: the longitudinal momenta in the central propagators of Fig.~5 can be neglected.
1308: As a result the $k'_{\perp}$ and $k_{\perp}$ loop
1309: integrations produce, as anticipated, a
1310: transverse momentum integral of the form of (\ref{tqqg}) which we write,
1311: in complex $\gamma$ - matrix notation, as
1312: $$
1313: \eqalign{\int& ~\frac{d^2k_{\perp}'}{{k_{\perp}'}^2}~\int d^2k_{\perp} ~
1314: Tr\bigl\{~\frac{\gam}{\kt}~G_L(k_{\perp},k_{\perp}')~\frac{\gam^*}{\kt^*}~
1315: G_R(k_{\perp},k_{\perp}')~\bigr\}\cr
1316: &\sim ~\int ~\frac{d^2k_{\perp}'}{{k_{\perp}'}^2}~\int d^2k_{\perp} ~
1317: \biggl(~\frac{k_{\perp}\times k'_{\perp}}{M^2}
1318: ~ \biggr)^2~
1319: \frac{ Tr\{ \gamma_- \gam^* \gamma_+ \gam\} }
1320: {\kt \kt^*} \cr
1321: &\sim ~
1322: \int ~\frac{d^2k_{\perp}' d^2k_{\perp}}{{k_{\perp}'}^2 k_{\perp}^2}
1323: ~\biggl(~\frac{k_2k'_3 -k_3k'_2}{M^2}~\biggr)^2\cr
1324: }
1325: \auto\label{fam0}
1326: $$
1327: Since the foregoing analysis assumes that both
1328: (\ref{intr2}) and (\ref{intr3}) hold, it follows that both ${k'_{\perp}}^2$
1329: and $k_{\perp}^2$ can be integrated over a range of values,
1330: that are $O(MS^{\frac{1}{2}})$ without either the
1331: approximations that we have made breaking down or the transverse momentum
1332: approximation to the gluon and
1333: quark propagators being invalidated. Therefore, we obtain
1334: a contribution from (\ref{fam0}) of the form
1335: $$
1336: \eqalign{\int& ~\frac{d^2k_{\perp}'}{{k_{\perp}'}^2}~\int d^2k_{\perp} ~
1337: Tr\bigl\{~\frac{\gam}{\kt}~G_L(k_{\perp},k_{\perp}')~\frac{\gam^*}{\kt^*}~
1338: G_R(k_{\perp},k_{\perp}')~\bigr\}\cr
1339: &\sim ~\frac{1}{M^4}~
1340: \int_{O(M S^{\frac{1}{2}})}d({k_{\perp}'}^2)~\int_{O(M S^{\frac{1}{2}})}
1341: d(k_{\perp}^2)~~~\sim ~\frac{S}{M^2}
1342: }
1343: \auto\label{fam}
1344: $$
1345: (Note that, because the anomaly contribution to $G_L(k_{\perp}, k_{\perp}')$
1346: is linear in $k_{\perp}'$, if it is combined with a $G_R(k_{\perp}, k_{\perp}')$
1347: that does not have this contribution then integration over $k_{\perp}'$
1348: will give a cancelation of the enhancement effect. This is why we have considered
1349: a diagram which gives anomaly contributions to both $G_L$ and $G_R$.)
1350:
1351: As we noted above, because two fermion exchange is involved, we
1352: would have expected the amplitude to increase only as some power of $ln S$.
1353: However, we have now shown
1354: that the kinematic region of Fig.~5 that we have isolated actually
1355: produces a power enhancement of the expected high energy behavior. As we will
1356: see in the next Section, there are no accompanying logarithms in this lowest-order
1357: appearance of the anomaly.
1358:
1359: Clearly, if a fixed transverse momentum cut-off is imposed, i.e.
1360: $$
1361: {k'_{\perp}}^2,~k_{\perp}^2 < \lambda_{\perp}
1362: \auto\label{fco}
1363: $$
1364: then there will be no
1365: contribution of the form (\ref{fam}) when $S$ is sufficiently large. Therefore,
1366: a transverse cut-off eliminates the enhancement effect and restores
1367: the normal behavior expected for two fermion exchange. However, as we discuss
1368: at greater length in Section 7, a transverse cut-off violates gauge invariance
1369: Ward identities in a way that replaces the anomaly enhancement by transverse
1370: momentum infra-red divergences.
1371:
1372: \subhead{4.7 Comparison With the Axial Vector Vertex Anomaly}
1373:
1374: To see the relationship between the anomaly amplitude (\ref{bxi11}) that we
1375: have found and the familiar axial vector anomaly we proceed as follows.
1376: First we change variables from $p'$ to $q$ where
1377: $$
1378: q_+ ~=~\frac{p'_+}{\Lambda}~, ~~
1379: q_- ~=~ \Lambda~p'_-~,~~
1380: q_{\perp}~=~p'_{\perp}~, ~~~~~~~\Lambda~=~
1381: \bigl(\frac{\epsilon}{M}\bigr)^{\frac{1}{2}}~S^{\frac{1}{4}}
1382: \auto\label{cv1}
1383: $$
1384: If we also extend the integration region for $p_-'$
1385: to $p_-' \sim M$ (which, as we have already noted, would not
1386: significantly alter the above analysis) then
1387: (\ref{bxi1}) becomes (moving the $\gamma_-/M^2 $ outside of the brackets)
1388: $$
1389: ~\frac{\gam}{\kt}~~
1390: \biggl(~
1391: \centerunder{\hbox{\huge $\int$}}{\raisebox{-4mm}{${\scriptstyle
1392: |q_j|~ \sim ~(\epsilon M)^{\frac{1}{2}}S^{\frac{1}{4}} }$}}
1393: ~ dq_j ~\frac{\kt\qt^* }{
1394: [q_{\perp}+k'_{\perp}]^2 ~[2q_+q_- - q_{\perp}^2 ]}~ \biggr)
1395: ~ \frac{\gamma_-}{M^2}~\frac{\gam^*}{\kt^*}
1396: \auto\label{bxi2}
1397: $$
1398: where, as indicated the
1399: range of integration is now the same for all components of $q$.
1400: In the limit $S \to \infty$, the integration is over a
1401: four-dimensional large momentum region and, formally, the integral
1402: is linearly divergent. Also, a product
1403: of three othogonal $\gamma$ matrices is present
1404: - although there is no trace involved. Consequently, it is natural to expect
1405: a large momentum contribution of the form
1406: associated with the triangle anomaly.
1407:
1408: The strictly infinite momentum region contribution to (\ref{bxi2}) is
1409: $$
1410: \sim ~~\frac{\gam}{\kt}~\biggl(~ \int~dq_+ dq_- \int d^2q_{\perp} ~
1411: \biggl[~\frac{k_{\perp}\times q_{\perp}}{q_{\perp}^2}
1412: \biggr]~\frac{1}{[2q_+q_- - q_{\perp}^2]}~ \biggr)~\frac{ \gamma_-}{M^2}
1413: ~\frac{\gam^*}{\kt^*}
1414: \auto\label{lmr}
1415: $$
1416: which, integrating by parts with respect to $q_{\perp}^2$, we can rewrite
1417: as
1418: $$
1419: \sim~~ \frac{\gam}{\kt}~\biggl(~
1420: \int~dq_+ dq_- \int d^2q_{\perp} ~\frac{k_{\perp}\times q_{\perp}}
1421: {\bigl[2q_+q_- - q_{\perp}^2\bigr]^2} ~ \biggr)~ \frac{\gamma_-}{M^2}
1422: ~\frac{\gam^*}{\kt^*}
1423: \auto\label{lmr1}
1424: $$
1425: and, with a further integration by parts, as
1426: $$
1427: \sim ~~ \frac{\gam}{\kt}~\biggl(~
1428: \int~dq_+ dq_- \int d^2q_{\perp} ~\frac{q_{\perp}^2
1429: [~k_{\perp}\times q_{\perp}~]}
1430: {\bigl[2q_+q_- - q_{\perp}^2\bigr]^3} ~ \biggr)~ \frac{\gamma_-}{M^2}
1431: ~\frac{\gam^*}{\kt^*}
1432: \auto\label{lmr2}
1433: $$
1434: Undoing the $\gamma$ - matrix removal involved
1435: in going from (\ref{bxi}) to (\ref{bxi1})
1436: (or, equivalently, inserting $\gamma$ - matrices using
1437: $2 = \gamma_+\gamma_- + \gamma_-\gamma_+ = \gam\gam^* + \gam^*\gam$~) we
1438: can rewrite (\ref{lmr2}) as
1439: $$
1440: \eqalign{&\sim ~~ \gam ~\biggl(~\int_{|q_i|~> >~ O(M) }~ d^4q ~\frac{ ~
1441: \gam^* ~[\qt^*\gam] ~ \gamma_-
1442: ~ [\qt \gam^*]~ \gamma_+
1443: ~[\qt^* \gam]}{[~q^2~]^3 } \biggr)
1444: ~~\frac{\gamma_-}{M^2}~ \frac{\gam^*}{\kt^*}\cr
1445: & \sim ~~\gam \biggl(~\int_{|q_i|~> >~ O(M) }~ d^4q ~\frac{ ~
1446: \gam^*~ \st{q}_{\perp} ~(1- \gamma_5) \gamma_-
1447: ~ \st{q}_{\perp}~ \gamma_+
1448: ~\st{q}_{\perp} }{[~q^2~]^3 } \biggr)
1449: ~~\frac{\gamma_-}{M^2} ~\frac{\gam^*}{\kt^*} }
1450: \auto\label{bxi40}
1451: $$
1452:
1453: We recognize the integral, within the brackets, of (\ref{bxi40})
1454: as a left-handed transverse propagator contribution to a tensor component of
1455: the standard large momentum anomaly integral (apart from the feature that
1456: there is no trace of the $\gamma$ -matrices involved). Therefore, we could
1457: anticipate (\ref{bxi11}) directly from the familiar anomaly contribution to
1458: a three current vertex $T_{\mu \alpha \beta}(k_1,k_2)$ - the notation is that
1459: of Fig.~C1. In our case,
1460: $$
1461: k_1~=~ -k_2~= ~k'_{\perp}
1462: \auto\label{k1k2}
1463: $$
1464: and if we consider the decomposition into invariant
1465: amplitudes (\ref{inde}), (\ref{bxi11}) corresponds to the contribution of the first
1466: two terms, which are linear in $k_1$ and $k_2$.
1467:
1468: It is very well known that in the three-current vertex the ambiguity
1469: of the ultra-violet anomaly contribution
1470: is determined by vector current Ward identities that relate
1471: the anomaly contribution to infra-red triangle diagram contributions.
1472: In our case, we anticipate that
1473: there will be a (reggeized) gluon
1474: Ward identity which similarly determines the coefficient of the anomaly
1475: contribution we have found. We discuss this point further in Section 5.
1476: Note, however, that, in the special
1477: momentum configuration (\ref{k1k2}), all the other terms in (\ref{inde}) vanish
1478: - if there are no infra-red divergences to consider. Therefore, in the lowest-order
1479: graphs we are discussing, the ultra-violet anomaly contributions we are discussing
1480: can not be canceled by the contribution of infra-red transverse momentum regions.
1481:
1482: \newpage
1483:
1484: \mainhead{5. NON-CANCELATION OF THE ANOMALY}
1485:
1486: In this Section we consider other diagrams that are also $O(\alpha_s)$
1487: and similarly have anomaly enhancements that might
1488: produce an overall cancelation.
1489:
1490: \subhead{5.1 Reality of the Anomaly Amplitude}
1491:
1492: It is significant that the anomaly amplitude we have found, although
1493: calculated with internal lines on-shell, is real. Indeed there is no evidence, in
1494: the amplitude, of either the
1495: $s$ - channel or the $t$ - channel intermediate states
1496: that are present in the diagram from which it was calculated.
1497: At first sight this seems paradoxical since it
1498: would appear that the analysis of Fig.~2, in the previous Section,
1499: can be viewed as the
1500: calculation of an $s$ - channel discontinuity - via the unitarity cut corresponding
1501: to the dashed line in Fig.~10(a).
1502: \begin{center}
1503: $~~~~$\epsfxsize=4.5in \epsffile{ehes12.ps}
1504: \newline \centerline{(a)\hspace{2.4in}(b)}
1505:
1506: Fig.~10 Unitarity Cuts of (a) the Diagram of Fig.~2 (b) a Related Diagram
1507: \end{center}
1508: The tree amplitude that appears below the cut
1509: is integrated with the one loop amplitude that appears above the cut.
1510: That the calculation can be related to
1511: the evaluation of a discontinuity immediately justifies, in fact,
1512: our choice of lines to place on-shell.
1513:
1514: That there is ultimately no discontinuity
1515: associated with the anomaly is due to a second discontinuity
1516: contribution from the unitarity cut of a closely related
1517: graph shown in Fig.~10(b).
1518: Clearly Fig.~10(b) is so
1519: similar to Fig.~10(a) that our analysis carries over directly.
1520: In both cases, the intermediate
1521: state integration over $k_{\perp}$ produces an imaginary
1522: contribution of $i|S|^{\frac{1}{2}}$. However,
1523: the $k'_{\perp}$ integration, that also gives a factor
1524: of $i|S|^{\frac{1}{2}}$, is part of the integration
1525: within the loop amplitude.
1526:
1527: If we formulate the above analysis as a
1528: unitarity calculation
1529: then the amplitude on one side of the cut must be complex conjugated.
1530: As a result, the loop amplitude will have the opposite sign in the
1531: contributions
1532: from Figs.~10(a) and 10(b) and since all other parts of the diagrams
1533: contribute identically, adding the two will give a factor of
1534: $$
1535: (2\pi )^4~\bigl(~[i~|S|^{\frac{1}{2}}~]~[~i~|S|^{\frac{1}{2}}~]
1536: ~+~[~i~|S|^{\frac{1}{2}}~]~[- i~|S|^{\frac{1}{2}}~]~\bigr)
1537: ~~=~0
1538: \auto\label{disc}
1539: $$
1540: Alternatively, if we calculate the contribution of the two diagrams as that
1541: of amplitudes then the loop amplitude will have the same sign in both cases
1542: and (\ref{disc}) will be replaced by
1543: $$
1544: (\pi )^4~\bigl(~[~i~|S|^{\frac{1}{2}}~]~[~i~|S|^{\frac{1}{2}}~]
1545: ~+~[~i~|S|^{\frac{1}{2}}~]~[~i~|S|^{\frac{1}{2}}~]~\bigr)
1546: ~~=~2~\pi^4~S
1547: \auto\label{ampl}
1548: $$
1549: which will, indeed, give a real amplitude. The contribution of Fig.~10(b) simply
1550: doubles that of Fig.~10(a). The absence of a discontinuity implies that, as we
1551: anticipated earlier, there are no additional logarithms accompanying the power
1552: enhancement due to the anomaly. As we stated in the last Section, the anomaly
1553: is a simple pole at $S=\infty$ which results
1554: from the combination of the asymptotic pinching
1555: of the mass-shell poles of the hatched propagators
1556: with the large momentum behavior of the unhatched propagators.
1557:
1558: \subhead{5.2 Another Anomaly Generating Diagram}
1559:
1560: The diagram of Fig.~10(b) is obtained from that of Fig.~10(a) by simultaneously
1561: ``twisting'' both the left and right-side box diagrams.
1562: For much of our discussion (including the addition of extra gluons in the
1563: next Section) we will keep the right-side of the diagrams we
1564: consider, and therefore the corresponding $G_R$, fixed
1565: and discuss anomaly amplitudes entirely in terms
1566: of possible left-side contributions to $G_L$.
1567: In the simple case of the one gluon diagrams that we are presently discussing,
1568: the right-side
1569: coupling will be that of Fig.~2 (or Fig.~10(a) ) and it will be clear that, as in
1570: the above discussion, diagrams with the right-side coupling of
1571: Fig.~10(b) simply give parallel contributions. However, when we consider infra-red
1572: cancelations in the next Section,
1573: it will be essential to also consider all contributions to $G_R$.
1574:
1575: Consider, next, the diagram shown in Fig.~11 that is
1576: obtained from that of Fig.~2 by twisting the left half of the diagram
1577: relative to the right half.
1578: \begin{center}
1579: \epsfxsize=2.5in
1580: \epsffile{ehes68.ps}
1581:
1582: Fig.~11 A Twisted Diagram
1583: \end{center}
1584: By a similar application of the above analysis, which puts on-shell
1585: the hatched lines shown in Fig.~11,
1586: the transverse momentum diagram of Fig.~1 will again be
1587: generated. The $G_L$ shown in Fig.~12(a),
1588: obtained from the left part of Fig.~11,
1589: contains the effective triangle diagram shown in Fig.~12(b).
1590: \begin{center}
1591: \epsfxsize=5in
1592: \epsffile{ehes701.ps}
1593:
1594: (a)\hspace{3in}(b)
1595:
1596: Fig.~12 (a) The $G_L$ Generated by Fig.~11 (b) The
1597: Effective Triangle Diagram
1598: \end{center}
1599: A very similar expression to (\ref{bxi11}) will clearly be obtained.
1600: The differences in the analysis can be summarized as follows.
1601:
1602: \noindent (i) ~~~ The analogue of (\ref{bxi}) gives (\ref{bxi1})
1603: but with $\kt [\pt']^* \to
1604: \kt^*\pt'$ which leads to $k_{\perp}\times p'_{\perp} \to
1605: -~ k_{\perp}\times p'_{\perp}$ in (\ref{bxi3}) and the following.
1606:
1607: \noindent (ii)~ ~ A second change of sign arises from $k'_{\perp}
1608: \to - k'_{\perp}$ in (\ref{bxi1}).
1609:
1610: \noindent The net result is that an identical anomaly contribution,
1611: to that obtained from Fig.~2, is obtained from the diagram of Fig.~11.
1612:
1613: The lines placed on-shell, asymptotically,
1614: in Fig.~11 do not correspond to a simple cut of the
1615: diagram, as was the case for Fig.~2. However, Fig.~11 can also
1616: be represented as in Fig.~13(b), i.e. as an exchanged gluon attached to the
1617: off-shell lines of the cut amplitude of Fig.~13(a).
1618: The exchanged gluon has transverse momentum much less than
1619: the off-shell quark or antiquark
1620: to which it couples in the large momentum $p'$ and $p''$
1621: regions which generate the anomaly. Consequently, it does
1622: not interfere (kinematically) with either the quark/antiquark scattering process,
1623: or the asymptotic placing on-shell of the left side fast quark and the right side
1624: antiquark. Therefore, the justification for the choice of lines placed on-shell
1625: is closely related to
1626: the existence of the asymptotic physical region discontinuity of Fig.~13(a).
1627: \newline
1628: \parbox{2.6in}{
1629: \begin{center}
1630: \epsfxsize=1.8in
1631: \epsffile{ehes0671.ps}
1632: \end{center}}
1633: \parbox{0.6in}{$\to$}
1634: \parbox{2.6in}{
1635: \begin{center}
1636: \epsfxsize=1.8in
1637: \epsffile{ehes067.ps}
1638: \end{center}
1639: }
1640: \newline \centerline{(a)\hspace{3in}(b)}
1641: \begin{center}
1642: Fig.~13 Addition of an Exchanged Gluon to a Cut Amplitude
1643: \end{center}
1644: The asymptotic pinching of the particle poles that gives the discontinuity in
1645: Fig.~13(a), together with the large momentum behavior of the uncut propagators,
1646: is responsible for the pole at infinite momentum in Fig.~13(b)
1647: that corresponds to the anomaly.
1648:
1649: A second, essential, point related to the choice of on-shell lines
1650: is the following. According to multi-regge theory,
1651: the coupling $G_L$ can be evaluated by a double dispersion relation, represented
1652: schematically in Fig.~14 - where the cuts represent the discontinuities involved.
1653: As a consequence, $G_L$ can be expressed as a
1654: sum over dispersion integrals which give
1655: amplitudes corresponding to all possible double discontinuities plus,
1656: possibly but not necessarily, (generalized)
1657: subtraction terms containing just single discontinuities. The anomaly
1658: contributions we have found are, in fact, generalized subtraction terms and
1659: the contributions of Fig.~3 and Fig.~12, respectively, correspond to the
1660: two single discontinuity terms shown explicitly in Fig.~14.
1661: \begin{center}
1662: \epsfxsize=5in
1663: \epsffile{ehes69.ps}
1664:
1665: Fig.~14 Representation of the Double Dispersion Relation for $G_L$
1666: \end{center}
1667: However, since we are
1668: evaluating an amplitude, and not a discontinuity, to have a contribution
1669: with on-shell lines corresponding to particular single discontinuities
1670: of $G_L$ and $G_R$,
1671: these discontinuities must be present in the asymptotic
1672: kinematic region we are considering.
1673: In fact, the discontinuity line in Fig.~13(b) can be regarded as
1674: representing the combination of the relevant discontinuities of $G_L$ and $G_R$.
1675:
1676: \subhead{5.3 Possible Cancelation Mechanisms}
1677:
1678: If we consider just contributions to
1679: the transverse momentum diagram of Fig.~1, then
1680: Fig.~11 is the only diagram which contributes (via the coupling of Fig.~6(a))
1681: to the same helicity amplitude as Fig.~2 and which
1682: generates an appropriate effective triangle diagram,
1683: apart from the diagram obtained by similarly twisting Fig.~10(b).
1684: We can not twist just the quark-antiquark state since this would reverse the
1685: direction of the quark arrow along the fast quark line, requiring a change of the
1686: external helicity to obtain a coupling.
1687: We conclude, therefore, that the full anomaly contribution to $G_L$ is
1688: obtained by adding the two effective triangle diagrams of Fig.~15.
1689:
1690: \begin{center}
1691: \parbox{3.2in}{\epsfxsize=3in
1692: \epsffile{ehes23.ps}
1693: }
1694:
1695: Fig.~15 The Two Effective Triangle Diagrams Generating the Anomaly
1696: \end{center}
1697: Therefore, within the transverse momentum diagram of Fig.~1, the anomaly
1698: enhancement does not cancel.
1699:
1700: It is natural, however, to expect that there will be further cancelations.
1701: As we emphasized in the Introduction,
1702: because longitudinal vector meson contributions are
1703: involved, it is important to look for all possible
1704: cancelation mechanisms that could be associated with an underlying
1705: gauge invariance. In particular,
1706: because the left-side quark and right-side antiquark are asymptotically
1707: on-shell, we must consider whether asymptotic electroweak Ward identities
1708: could lead to the cancelation of the vector meson
1709: numerator contributions
1710: that are producing the anomaly enhancement.
1711:
1712: There are two obvious Ward identity related cancelations that we should consider.
1713: First, we consider the tree diagram that appears in the lower half of
1714: Fig.~10(a). At finite momentum, if the intermediate state
1715: quark and antiquark are strictly on-shell,
1716: there will be Ward identities involving this
1717: diagram and all other diagrams obtained by attaching the internal vector meson
1718: lines at all possible points. Examples of such diagrams, together with the initial
1719: diagram, are shown in Fig.~16.
1720: \begin{center}
1721: \epsfxsize=4.5in
1722: \epsffile{ehes79.ps}
1723:
1724: Fig.~16 Tree Diagrams Contributing to an Electroweak Ward Identities
1725: \end{center}
1726: In fact, because the intermediate state quark and antiquark are
1727: only asymptotically on-shell, we might expect that only
1728: the vector meson numerator components that are parallel to the
1729: asymptotic light-cone quark and antiquark momenta must decouple.
1730: This decoupling has
1731: already appeared in the analysis of sub-section {\bf 4.3}.
1732:
1733: In Appendix D we study in detail the Ward identity cancelations associated with
1734: the tree diagrams of Fig.~16. The essential
1735: part of the first diagram in Fig.~16 is, indeed,
1736: directly canceled by the contribution
1737: of the second diagram, which corresponds to the
1738: feynman diagram shown in Fig.~17.
1739: \begin{center}
1740: \epsfxsize=1.8in
1741: \epsffile{ehes82.ps}
1742:
1743: Fig.~17 A Feynman Diagram With a Canceling Anomaly Contribution.
1744: \end{center}
1745: However, the anomaly enhanced amplitude produced
1746: by Fig.~17 appears to not be representable as a transverse momentum
1747: diagram divergence.
1748:
1749: More surprisingly, perhaps, essentially the same anomaly enhanced
1750: amplitude then reappears via
1751: the contribution of the third diagram in Fig.~16, which corresponds to the
1752: feynman diagram shown in Fig.~18.
1753: This is a diagram that would normally be neglected because off-shell
1754: propagators are carrying large light-cone momenta.
1755: \begin{center}
1756: \epsfxsize=1.8in
1757: \epsffile{ehes99.ps}
1758:
1759: Fig.~18 The Feynman Diagram Corresponding to the Third Tree Diagram in Fig.~16
1760: \end{center}
1761:
1762: In a sense, therefore,
1763: nothing is gained by implementing the Ward identity cancelations.
1764: However, after this implementation it is apparent that
1765: the lack of anomaly cancelation is
1766: entirely due to the asymptotic nature of the placing on-shell
1767: of the quark and antiquark lines.
1768: Also, when the Ward identity cancelations are carried out several diagrams are
1769: included, in addition to Fig.~18, that would normally be considered non-leading.
1770: This makes it clear that there is a general phenomenon of superficially
1771: non-leading high-energy behavior contributing to the leading behavior
1772: because of large transverse momentum divergences.
1773:
1774: There is also
1775: a second Ward identity, involving the top part of Fig.~10(a) and other
1776: loop diagrams, some of which are shown in Fig.~19, that might be expected to
1777: lead to the decoupling of the top $\gamma_- p_+'$ vertex in Fig.~7, together with
1778: the corresponding $\gamma_+p_-''$ right-side vertex. These vertices
1779: are crucially important for our analysis.
1780: \begin{center}
1781: \epsfxsize=4.5in
1782: \epsffile{ehes78.ps}
1783:
1784: Fig.~19 Loop Diagrams Contributing to Electroweak Ward Identities
1785: \end{center}
1786: However, in this case it is
1787: the intervention of the asymptotic anomaly that invalidates the potential
1788: asymptotic Ward identities and, self-consistently, prevents
1789: the decoupling of the vector meson vertices that are involved.
1790:
1791: In conclusion, we can say that there is no cancelation
1792: of the transverse momentum coupling
1793: effective triangle diagram anomaly by another diagram with a similar
1794: anomaly. There may very well be a cancelation
1795: outside of the transverse momentum diagram formalism. However,
1796: in this paper at least, we will not pursue this possibility any further.
1797:
1798: \subhead{5.4 The Same Helicity Scattering Amplitude }
1799:
1800: To obtain an anomaly amplitude for the scattering of
1801: vector mesons which both have helicity $\lambda = +1$
1802: we include the left side coupling of Fig.~6(b) within a diagram that
1803: otherwise is the same as Fig.~2 or Fig.~11.
1804: The result is the two diagrams shown in Fig.~20.
1805: When displayed in the first form, it is clear that
1806: the only difference between these diagrams and, respectively, Figs.~2 and 11
1807: is that along the left-most vertical line
1808: $P_+ \to - P_+$. Therefore, if we evaluate the diagrams with the sign of $P_+$
1809: reversed, corresponding to a cross-channel physical region, the appropriate
1810: on-shell configurations will be present. The diagrams
1811: will be kinematically identical, respectively, to Figs.~2 and 11
1812: and will give identical anomaly contributions, but with $S \to -S$.
1813:
1814: The second form for the diagrams displayed in Fig.~20 is
1815: more transparent for discussing symmetry properties of the intermediate state. In
1816: particular, in this form, it is
1817: clear that Fig.~20(b) can be obtained from Fig.~2 by twisting the quark-antiquark
1818: intermediate state (together with the necessary redirection of the quark arrow
1819: in the left part of the diagram). Fig.~20(a) can similarly be obtained from
1820: Fig.~11.
1821: \begin{center}
1822: \epsfxsize=4.5in
1823: \epsffile{ehes14.ps}
1824:
1825: (a)\hspace{2.5in}(b)
1826:
1827: Fig.~20 The Scattering of Same Helicity Vector Mesons
1828: \end{center}
1829:
1830: \subhead{5.5 Cancelation in a Vector Theory}
1831:
1832: If the vector mesons we are considering had a vector coupling, rather than a
1833: left-handed coupling, to quarks, then the diagrams of Fig.~16 would appear also
1834: in the opposite sign helicity amplitude but with $(1+\gamma_5)$ couplings
1835: replacing the $(1-\gamma_5)$ couplings in Fig.~6(b). In this case, after the use of
1836: (\ref{prj5}), the relative
1837: minus sign discussed in the previous sub-section
1838: (resulting from $S\to -S$) produces a cancelation between the
1839: anomaly contributions from Fig.~2 and Fig.~20(a). Similarly, the contributions from
1840: Fig.~11 and Fig.~20(b) cancel. In a vector theory,
1841: this cancelation of right and left-handed
1842: coupling contributions would persist, even as we add more gluons as discussed in
1843: the next Section.
1844:
1845: \subhead{5.6 The Even Signature Amplitude}
1846:
1847: To form signatured scattering amplitudes we should add to, or subtract from,
1848: a particular helicity amplitude the amplitude
1849: obtained by a CPT transformation of one scattering state relative to the
1850: other. Therefore, if $A_{-+}$ and $A_{++}$ are the opposite sign and
1851: same sign amplitudes we have discussed,
1852: $$
1853: A^{\pm}(P_+,P_-)~=~A_{-+}(P_+,P_-)~\pm~ A_{++}(-P_+,P_-)
1854: \auto\label{esa}
1855: $$
1856: is an even/odd signature amplitude. This implies that the anomaly amplitudes
1857: arising from Fig.~2 and Fig.~20(a) are added in the even signature amplitude
1858: and subtracted in the odd signature amplitude (as are, also,
1859: the anomaly amplitudes arising from Fig.~11 and Fig.~20(b). Therefore,
1860: the anomaly cancels in the odd signature amplitude and
1861: is present only in the even signature amplitude. This will continue to be the
1862: case as we add more gluons in the next Section. It is directly related, via a
1863: generalization of the discussion of the previous two sub-sections, to the
1864: cancelation in a vector theory.
1865:
1866: \subhead{5.7 $C$ and $P$ Properties of the Transverse Momentum State}
1867:
1868: Since the intermediate state in Fig.~1 is completely transverse (or,
1869: equivalently, is a $t$ - channel intermediate state) the $T$ part of the
1870: $CPT$ transformation, defining the signature of an amplitude, has no effect on it.
1871: Therefore, we should be able to relate signature directly to the $CP$ properties
1872: of the transverse momentum state.
1873:
1874: The parity transformation reverses the transverse momentum of the gluon
1875: and so, because of the coupling (\ref{bxi11}),
1876: simply gives a minus sign. Without a color factor, the charge conjugation
1877: transformation also gives just a minus sign. Therefore,
1878: the gluon component of the intermediate state is even under $CP$.
1879: For quarks the left-handed coupling violates both $P$ and $C$.
1880: As a result, the quark-antiquark
1881: intermediate state only has simple
1882: transformation properties under the combined $CP$ transformation.
1883: Charge conjugation transforms
1884: a quark (antiquark) to the
1885: corresponding antiquark (quark), with the same helicity (opposite chirality).
1886: The parity part of the $CP$ transformation then
1887: reverses the helicities. In our case, the quark and antiquark have opposite
1888: helicities and so they will be simply
1889: interchanged by the $CP$ transformation. Individually, the diagrams we
1890: are discussing do not have simple symmetry properties with respect to
1891: quark/antiquark interchange. Not surprisingly, however, the full set of anomaly
1892: contributions in the even signature amplitude does have such a property.
1893:
1894: With the four diagrams, Figs.~2, 11, 20(a) and (b),
1895: added in the even signature amplitude, it is clear
1896: (using the second display form in Fig.~20) that the left-side
1897: coupling is symmetric, diagrammatically,
1898: with respect to the interchange of the quark and antiquark.
1899: The interchange relates Fig.~2 to Fig.~20(b) and Fig.~11 to Fig.~20(a).
1900: In addition to the
1901: reversal of the quark line, the contribution of
1902: Fig.~20(b) to the even signature amplitude differs kinematically from that
1903: of Fig.~2 in two ways that produce canceling sign changes. Firstly,
1904: $k_{\perp} \to -k_{\perp}$, and, secondly, the effect in the
1905: $k_-'$ integration of $P_+ \to -P_+$ resulting from the definition of
1906: the even signature amplitude. Therefore, in this
1907: amplitude, the quark/antiquark intermediate state is even under $CP$.
1908: Since the gluon state is also even under $CP$,
1909: the full transverse momentum state is indeed even under $CP$, as
1910: it should be.
1911:
1912: Note that the full even signature
1913: amplitude will contain, in addition to the four diagrams
1914: of Figs.~2, 11, 20(a) and 20(b), the four related diagrams obtained by
1915: substituting the right side of Fig.~10(b) for that of Fig.~10(a). In effect,
1916: in this second set of diagrams the twists are made on $G_R$ that are made on
1917: $G_L$ in the first set of diagrams. Each set of twists is sufficient to give
1918: an intermediate state with the appropriate $CP$ property. As a result, the
1919: discussion of each set of four diagrams can be made separately and
1920: is directly parallel. In higher orders it will sometimes
1921: be necessary to consider both sets of twists
1922: together to obtain an intermediate state with the right $CP$ property.
1923:
1924: \newpage
1925:
1926: \mainhead{6. COLOR FACTORS AND MORE GLUONS}
1927:
1928: We begin with a discussion of SU(2) flavor that will, essentially, allow us to
1929: ignore it in the following.
1930:
1931: \subhead{6.1 SU(2) Flavor}
1932:
1933: The SU(2) flavor symmetry will play only a minimal role in our discussion
1934: and we will introduce it in a very elementary manner.
1935: We consider the exchange of a quark-antiquark
1936: $\{I=1, I_z=0\}$ state that, in the standard model, would
1937: carry the quantum numbers of the $\pi^0$. Identifying $W^{\pm},W^0$ with the
1938: $\{I=1, I_z = \pm,0\}$ vector mesons and identifying
1939: $u,d$ with the $\{I= \frac{1}{2}, I_z= \frac{\pm1}{2}\}$
1940: quarks, we can add flavor quantum numbers to the discussion
1941: of the previous Section by using the vector meson / quark vertices of Fig.~21.
1942: \begin{center}
1943: \epsfxsize=3.5in
1944: \epsffile{ehes64.ps}
1945:
1946: Fig.~21 Vector Meson / Quark Vertices
1947: \end{center}
1948: The flavorless couplings of Fig.~6 are then replaced by
1949: the sums of couplings shown in Figs.~22(a) and (b) and the internal vector meson
1950: on-shell contributions are replaced by a similar sum.
1951: \begin{center}
1952: \epsfxsize=2.5in
1953: \epsffile{ehes180.ps}
1954: \hspace{0.5in}
1955: \epsfxsize=2.5in
1956: \epsffile{ehes181.ps}
1957:
1958: (a)\hspace{2.8in}(b)
1959:
1960: Fig.~22 Couplings with SU(2) Flavor Quantum Numbers
1961: \end{center}
1962: We then add all the diagrams obtained with this set of couplings.
1963: The most important feature of these couplings is that they are
1964: symmetric with respect to $u \leftrightarrow d$ and that this symmetry is preserved
1965: by the internal vector meson exchange interactions.
1966: (It is also important for the $CP$ properties of the diagrams we discuss
1967: that, in going from Fig.~22(a) to Fig.~22(b), left-handed quarks and right-handed
1968: antiquarks are interchanged, as was the case for Figs.~6(a) and (b).)
1969: Consequently, the addition
1970: of SU(2) flavor factors will not produce any diagram cancelations and we can
1971: leave, as implicit, the replacement of the couplings of Fig.~6 by those of Fig.~22.
1972:
1973: \subhead{6.2 Color Factor Diagrams and the One Gluon Color Factor}
1974:
1975: SU(3) color factors will also be relatively
1976: simple. In all the diagrams we discuss,
1977: there will be only one quark loop. There is
1978: no external color and so color is introduced into the quark loop only by the
1979: couplings to the internal gluons. Also, for the diagrams we consider, gluons
1980: will appear only as part of the exchanged transverse momentum state and will
1981: be attached within the corresponding $G_L$ and $G_R$ transverse momentum couplings.
1982: As a result, we can use a simple notation
1983: to describe color factors. We represent the quark loop as a rectangle,
1984: and attach gluons only to the vertical lines.
1985: The attachment of gluons to the left-side vertical line represents the order
1986: of attachment to the quark loop within the left-side transverse momentum
1987: coupling $G_L$, while the right-side vertical line similarly represents the
1988: order of attachment within $G_R$. For each gluon
1989: there is a color matrix $\lambda_i$ at each attachment point.
1990: The full color factor is the trace of
1991: the product of the $\lambda$ - matrices taken around the loop,
1992: and then summed over $i=1,..,8$ for each gluon.
1993: The notation is illustrated for various numbers of exchanged gluons in Fig.~23.
1994: \begin{center}
1995: \epsfxsize=3.5in
1996: \epsffile{ehes25.ps}
1997:
1998: Fig.~23 Color Factor Diagrams
1999: \end{center}
2000:
2001: All of the diagrams discussed in the previous Section contain just one
2002: gluon and have the same color factor. This is represented by Fig.~23(a) and is
2003: simply
2004: $$
2005: \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2006: {\scriptstyle i}$}}
2007: ~~Tr\{\lambda_i^2\}
2008: \auto\label{cf1}
2009: $$
2010: Since all diagrams have the same color factor (and flavor factor)
2011: all of the discussion in the previous Section is essentially unchanged, apart from
2012: the discussion of charge conjugation, which now has to include color charge
2013: conjugation.
2014:
2015: For a general gluon field with color matrix $M_{ab}$, color charge conjugation
2016: is defined as
2017: $$
2018: M_{ab}~~\to~~ \bigl[-M_{ab}~\bigr]^T ~~= ~~-M_{ba}
2019: \auto\label{gcc}
2020: $$
2021: For the hermitian color matrix vertices $\lambda_i$
2022: $$
2023: \bigl[-\lambda_i~\bigr]^T ~~= ~-~\bigl[\lambda_i ~\bigr]^*
2024: \auto\label{gcc1}
2025: $$
2026: where $[~]^*$ denotes complex conjugation. Therefore, in addition to the charge
2027: conjugation minus sign discussed in the last section, the coupling of the gluon
2028: to the quark line (within $G_L$, say) is complex conjugated. Correspondingly,
2029: for the quark-antiquark pair, in addition to the
2030: charge conjugation discussed in the previous Section,
2031: quark/antiquark interchange gives
2032: $$
2033: \bigl[\lambda_i ~\bigr]^* ~\to ~ \bigl[\lambda_i ~\bigr]
2034: \auto\label{qcc}
2035: $$
2036: in (\ref{gcc1}). Since the parity transformation is unchanged,
2037: the full gluon plus quark/antiquark transverse
2038: momentum state remains even under $CP$ when color charge
2039: conjugation is included.
2040:
2041: \subhead{6.3 The Addition of a Soft Gluon }
2042:
2043: Next, we look for feynman diagrams that contain two gluons and that also,
2044: potentially, contain the anomaly enhancement. We will
2045: assume that only one gluon is involved in the transverse momentum divergence
2046: and will consider two
2047: possibilities for the scale of the transverse momentum carried by the second
2048: gluon. It can either be ``soft'',
2049: i.e. it carries a very small momentum $k''$, with
2050: $$
2051: |k_{\perp}''|~<< ~M~<<~|k_{\perp}|, |k'_{\perp}| ~~~~
2052: \leftrightarrow ~~~~``soft~''
2053: \auto\label{soft}
2054: $$
2055: or ``finite'', i.e.
2056: $$
2057: |k_{\perp}''|~\sim ~M~<<~|k_{\perp}|, |k'_{\perp}| ~~~~
2058: \leftrightarrow~~~~``finite~''
2059: \auto\label{fnte}
2060: $$
2061: As we will see,
2062: in some diagrams soft gluon exchange is possible, in addition to the anomaly
2063: generation, while in others only finite gluon exchange is possible. In both
2064: cases, the second gluon will provide an important color factor.
2065: A soft gluon, however, will also produce an infra-red divergence.
2066: Since the full transverse momentum
2067: state carries zero color, such divergences must cancel. This will
2068: help us to locate other diagrams generating the anomaly.
2069:
2070: We consider the ``soft'' gluon case first and
2071: look for diagrams that contribute to the
2072: transverse momentum diagram of Fig.~24 which, as discussed further in Appendix B,
2073: would again be expected to
2074: contribute (formally) only at next-to-next-to-leading log.
2075: (Note that, when the anomaly is not present, this diagram again,
2076: potentially, includes a logarithmic transverse momentum divergence generating
2077: an additional energy logarithm.)
2078: \begin{center}
2079: \epsfxsize=2in
2080: \epsffile{ehes27.ps}
2081:
2082: Fig.~24 The Two Gluon Transverse Momentum Diagram - the Broken Line Denotes
2083: a Soft Gluon
2084: \end{center}
2085:
2086: We begin with the addition of a soft gluon to the diagram of Fig.~2.
2087: The anomaly will appear in the same manner as before if, in the high-energy limit,
2088: an effective triangle diagram is generated as in Fig.~3, but with the
2089: additional gluon attached, via a point coupling,
2090: to one of the three vertices of the diagram.
2091: The required local coupling could appear,
2092: in principle, if
2093: $k''$ can be directed through an adjacent quark line which can
2094: be put on-shell by the $k_-''$ integration. If this line carries
2095: (predominantly) a large light-cone momentum then, in analogy with
2096: (\ref{k'os}), the integration will produce
2097: couplings that are independent of $k_{\perp}''$.
2098: In Figs.~25(a), (b), and (c) we show how the extra gluon could be added
2099: to Figs.~6,7, and 8, respectively, with
2100: the final $\gamma$ - matrices remaining the same as in Fig.~9.
2101: \begin{center}
2102: \epsfxsize=4.5in
2103: \epsffile{ehes28.ps}
2104:
2105: Fig.~25 Adding a Gluon to the Vertices of (a) Fig.~6~, (b) Fig.~7~, and
2106: (c) Fig.~8
2107: \end{center}
2108: In Fig.~25(c) the soft and hard gluon can be interchanged, whereas in Figs.~25(a)
2109: and (b) there is no ambiguity as to where the soft gluon has to be attached,
2110: if the $\gamma$ - matrix structure is to remain the same.
2111:
2112: For Fig.~25(a) we can, essentially, apply (\ref{k'os}) directly.
2113: For the couplings of Figs.~25(b) and (c) there is, however, a problem
2114: if the extra gluon is soft and so
2115: carries only very small transverse momentum. In these cases,
2116: the propagator that has to be placed on-shell by the $k_-''$ - integration
2117: is adjacent to an off-shell propagator
2118: that, in the anomaly configuration,
2119: is carrying very large transverse momentum ($ p_{\perp}'$).
2120: In this case the mass-shell condition is
2121: $$
2122: k_-''~\sim ~ \frac{(p_{\perp}'~+~ k_{\perp}'')^2}{p_+'}~\sim~
2123: \frac{\epsilon M \sqrt{S}}{\epsilon \sqrt{S}} ~\sim~M
2124: \auto\label{k''os}
2125: $$
2126: which can not be satisfied with $k_-''~<<~ |k_{\perp}''|
2127: ~<<~M$. Therefore, if the vertex for the extra gluon is of the form of
2128: Fig.~25(b) or (c), in both the $G_L$ and $G_R$ couplings, then it can not carry
2129: $|k_{\perp}''|^2 ~<<~M^2$.
2130:
2131: Later, we will discuss potential contributions from
2132: vertices of the form of Fig.~25(b) and (c) when $k''$ is ``finite'', i.e.
2133: $|k_{\perp}''|^2 ~\sim~M^2$. For the moment,
2134: we consider only the vertex of Fig.~25(a).
2135: Generation of the corresponding triangle diagram
2136: is shown in Fig.~26
2137: \begin{center}
2138: \epsfxsize=4.5in
2139: \epsffile{ehes221.ps}
2140:
2141: Fig.~26 A Two Gluon Effective Triangle Diagram
2142: \end{center}
2143: and the full feynman diagram, with the extra gluon attached in the same
2144: manner to both sides of Fig.~5, is shown in Fig.~27.
2145: \begin{center}
2146: \parbox{2.5in}{\epsfxsize=2.1in
2147: \epsffile{ehes160.ps}}
2148: \parbox{0.5in}{$\leftrightarrow$}
2149: \parbox{2.5in}{\epsfxsize=2.1in
2150: \epsffile{ehes161.ps}}
2151:
2152: Fig.~27 A Two Gluon Feynman Diagram With Two Cuts
2153: \end{center}
2154: As illustrated, the lines put on-shell correspond to making a double cut
2155: of the diagram. This corresponds to double discontinuity contributions to
2156: the $G_L$ and $G_R$ couplings. (These contributions are again generalized
2157: subtractions in that the full dispersion relations for $G_L$ and $G_R$
2158: contain triple discontinuities.)
2159:
2160: \subhead{6.4 The Two Gluon Anomaly Amplitude}
2161:
2162: If $k_{\perp}''$ is much smaller than any other transverse momentum in the
2163: diagram of Fig.~23, the only significant
2164: $k_{\perp}''$ dependence will be in the $k''$ propagator.
2165: Hence, the $k_{\perp}''$ integration can be factored out from the
2166: remaining integrations and,
2167: before the inclusion of any color factor, the diagram of Fig.~23 gives,
2168: via the reduction of Fig.~25(a) and Fig.~26,
2169: a high-energy anomaly enhanced amplitude of the form
2170: $$
2171: \eqalign{ \int^{|k_{\perp}''|<< M }&
2172: ~\frac{d^2k_{\perp}''}{{k_{\perp}''}^2}~
2173: \int ~\frac{d^2k_{\perp}'}{{k_{\perp}'}^2}~\int d^2k_{\perp} ~
2174: ~~\biggl(~\frac{k_{\perp}\times k'_{\perp}}{M^2}
2175: ~ \biggr)^2~
2176: \frac{ Tr\{ \gamma_- \gam^* \gamma_+ \gam\} }
2177: {\kt \kt^*} \cr
2178: &\sim ~\biggl(~\int^{|k_{\perp}''|<< M } ~\frac{d^2k_{\perp}''}{{k_{\perp}''}^2}~
2179: \biggr)~\int ~\frac{d^2k_{\perp}' d^2k_{\perp}}{{k_{\perp}'}^2 k_{\perp}^2}
2180: ~\biggl(~\frac{k_2k'_3 -k_3k'_2}{M^2}\biggr)^2~\cr
2181: &\sim ~\biggl(~\int^{|k_{\perp}''|<< M } ~\frac{d^2k_{\perp}''}{{k_{\perp}''}^2}~
2182: \biggr)~\int_{O(MS^{\frac{1}{2}})}~\frac{d({k_{\perp}'}^2)}{M^2}~
2183: \int_{O(MS^{\frac{1}{2}})}~ \frac{d(k_{\perp}^2)}{M^2}~\cr
2184: &\sim ~S~/~M^2
2185: }
2186: \auto\label{amp2}
2187: $$
2188: We can similarly add a soft gluon to each of the one gluon diagrams
2189: discussed in the last Section and generate a high energy amplitude
2190: of the same form. There is, however, clearly a divergence at $|k_{\perp}''|^2
2191: =0$ that we must discuss. First, however, we discuss the relevant
2192: color factors.
2193:
2194: \subhead{6.5 Two Gluon Color Factors}
2195:
2196: There are two possible color factors for the transverse momentum diagram
2197: of Fig.~24. They are shown in Figs.~23(b) and (c) and have the form
2198: $$
2199: (b)~~ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2200: {\scriptstyle i,j}$}}
2201: ~~Tr\{\lambda_j\lambda_i\lambda_i\lambda_j\}~~~~~~~~
2202: (c)~~\centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{${
2203: \scriptstyle i,j}$}}
2204: ~~Tr\{\lambda_i\lambda_j\lambda_i\lambda_j\}
2205: \auto\label{cf2}
2206: $$
2207: We discuss these two factors in a manner that will generalize when
2208: further gluons are added. The essential formula we need is
2209: $$
2210: \lambda_i\lambda_j ~=~ \frac{2}{3}~\delta_{ij}~
2211: +~\centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2212: {\scriptstyle k}$}}~
2213: (if_{ijk} +
2214: d_{ijk})\lambda_k
2215: \auto\label{lij}
2216: $$
2217: which, since $f_{ijk}$ is antisymetric and $d_{ijk}$ is symmetric,
2218: implies that
2219: $$
2220: \lambda_j\lambda_i ~+~ \lambda_i\lambda_j ~=~ \frac{4}{3}~\delta_{ij}~
2221: ~+~2~ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2222: {\scriptstyle k}$}}~d_{ijk}\lambda_k
2223: \auto\label{lji+}
2224: $$
2225: and
2226: $$
2227: \lambda_j\lambda_i ~-~ \lambda_i\lambda_j ~=~
2228: -~2i ~\centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2229: {\scriptstyle k}$}}~f_{ijk}\lambda_k ~~~~~~~
2230: \auto\label{lji-}
2231: $$
2232: Therefore, the sum of the two color factors (\ref{cf2}) is given by
2233: $$
2234: \eqalign{ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2235: {\scriptstyle i,j}$}}
2236: ~~Tr\{\lambda_j\lambda_i\lambda_i\lambda_j ~+~
2237: \lambda_i\lambda_j\lambda_i\lambda_j\}
2238: &=~ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2239: {\scriptstyle i,j}$}}
2240: ~~Tr\{[~\frac{4}{3}~\delta_{ij} ~+ ~2~
2241: \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2242: {\scriptstyle k}$}}~
2243: d_{ijk}\lambda_k]~\lambda_i\lambda_j~\}\cr
2244: &= ~ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2245: {\scriptstyle i,j}$}}
2246: ~~Tr\{[~\frac{4}{3}~\delta_{ij}~
2247: +~2~
2248: \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2249: {\scriptstyle k}$}}
2250: d_{ijk}\lambda_k]~[~\frac{2}{3}~\delta_{ij} ~+~
2251: \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2252: {\scriptstyle l}$}}~d_{ijl}\lambda_l]~
2253: \}\cr
2254: &= ~ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2255: {\scriptstyle i,j}$}} ~[~\frac{8}{9}~\delta_{ij}~+\frac{4}{3}
2256: \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2257: {\scriptstyle k}$}}~ d_{ijk}^2]~=~20\frac{4}{9}
2258: }
2259: \auto\label{cf3}
2260: $$
2261: and the difference of the two gives
2262: $$
2263: \eqalign{ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2264: {\scriptstyle i,j}$}}
2265: ~~Tr\{\lambda_j\lambda_i\lambda_i\lambda_j ~-~
2266: \lambda_i\lambda_j\lambda_i\lambda_j\}
2267: &= \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2268: {\scriptstyle i,j,k,l}$}}
2269: ~~Tr\{[-2i~f_{ijk}\lambda_k]~\lambda_i\lambda_j\}\cr
2270: &= \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2271: {\scriptstyle i,j,k}$}}
2272: ~~Tr\{[2i~f_{jik}\lambda_k]~[i~f_{ijl}\lambda_l]\}\cr
2273: &= \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2274: {\scriptstyle i,j,k}$}} ~\frac{4}{3} f_{ijk}^2~=~12
2275: }
2276: \auto\label{cf4}
2277: $$
2278:
2279: Both (\ref{cf3}) and (\ref{cf4}) are expressed as a sum of squares of color
2280: factors where each individual term corresponds to a particular color for
2281: the gluon intermediate state. The states that contribute can be found by writing
2282: the left-side color factor as a sum over distinct intermediate states, as
2283: is effectively done in
2284: (\ref{lji+}) and (\ref{lji-}). The full color factor can then be written,
2285: relying on the orthogonality of the intermediate states, as a sum of squares of the
2286: left-side factors.
2287: The quark-antiquark intermediate state can
2288: only carry zero or octet color. Correspondingly, since
2289: the total color of the intermediate state is zero, the
2290: gluon sums must also contribute either zero or octet color. This will continue to
2291: be the case when more gluons are present in the
2292: transverse momentum state.
2293: For the two gluon case, as illustrated in Fig.~28,
2294: \begin{center}
2295: \epsfxsize=4in
2296: \epsffile{ehes2601.ps}
2297:
2298: \epsfxsize=3.5in
2299: \epsffile{ehes2602.ps}
2300:
2301: Fig.~28 Color Breakdown of the Intermediate State
2302: \end{center}
2303: in the symmetric color
2304: factor (\ref{cf3}) the gluon sums give zero and
2305: octet color contributions, which both have even color parity. In the antisymmetric
2306: factor (\ref{cf4}) only the color octet with odd color parity contributes.
2307:
2308: \subhead{6.6 The Color Factor for the Anomaly}
2309:
2310: Consider again the effective triangle diagram of Fig.~22.
2311: (As in the last Section, we will initially discuss only the various
2312: anomaly contributions to $G_L$, with $G_R$ kept fixed.)
2313: Comparing with
2314: the diagrams of Fig.~3 and Fig.~12(a), the twisted triangle
2315: diagram that should give an anomaly contribution to
2316: add to that of Fig.~26
2317: is that shown in Fig.~29(a) and the corresponding full diagram is that of
2318: Fig.~29(b).
2319: \begin{center}
2320: \epsfxsize=3in
2321: \epsffile{ehes7060.ps}
2322: \hspace{0.6in}
2323: \epsfxsize=2in
2324: \epsffile{ehes320.ps}
2325:
2326: (a)\hspace{3in}(b)
2327:
2328: Fig.~29 (a) A Twisted Effective Diagram (b) The Full Twisted Diagram
2329: \end{center}
2330: As illustrated in Fig.~30
2331: and in analogy with Fig.~13, Fig.~29(b) can be obtained
2332: by adding an exchanged gluon to the two cut amplitude of Fig.~30(a).
2333: \begin{center}
2334: \parbox{2.6in}{
2335: \begin{center}
2336: \epsfxsize=2in
2337: \epsffile{ehes06711.ps}
2338: \end{center}}
2339: \parbox{0.6in}{$\to$}
2340: \parbox{2.6in}{
2341: \begin{center}
2342: \epsfxsize=2in
2343: \epsffile{ehes06712.ps}
2344: \end{center}
2345: }
2346:
2347: (a)\hspace{3in}(b)
2348:
2349: Fig.~30 Addition of an Exchanged Gluon to a Two Cut Amplitude
2350: \end{center}
2351: Again, in the anomaly region, with $|k'_{\perp}|^2 ~<<~
2352: |p'_{\perp}|^2,~ |p''_{\perp}|^2$, the exchanged gluon does not interfere,
2353: significantly, with the kinematics of taking the double discontinuity of
2354: Fig.~30(a). It is also clear from
2355: Fig.~30, that the on-shell lines of Fig.~29(b)
2356: correspond to physical double discontinuities of the $G_L$ and $G_R$.
2357:
2358: If we consider the order of $\lambda$ matrix
2359: multiplication (following the quark arrow) we see that
2360: it is reversed for the two
2361: gluons in Fig.~29(a) compared to those in Fig.~26. As a result,
2362: if we add the two diagrams
2363: the anomaly is multiplied by the sum of the two color factors (\ref{cf3}).
2364: As we have discussed, in this case
2365: there are separate contributions corresponding to whether
2366: the quark-antiquark pair is
2367: in a color octet or a color zero state.
2368:
2369: \subhead{6.7 Signature and Color Charge Parity}
2370:
2371: To form signatured amplitudes we consider, with the diagrams of Fig.~27
2372: and Fig.~29(b), the corresponding diagrams for same helicity vector scattering.
2373: These are shown in Figs.~31(a) and (b), respectively. With flavor included,
2374: the left-hand couplings in these diagrams should be replaced
2375: by the full, flavor symmetric, couplings of Fig.~22(b), while the right-hand
2376: couplings should continue to be the analogue of Fig.~22(a).
2377: As in Fig.~18, we display the diagrams in two different ways, each of which will
2378: be simpler for particular arguments.
2379: \begin{center}
2380: \epsfxsize=2.5in
2381: \epsffile{ehes57.ps}
2382: \hspace{0.6in}
2383: \epsfxsize=2.5in
2384: \epsffile{ehes58.ps}
2385:
2386: (a)\hspace{3in}(b)
2387:
2388: Fig.~31 Same Helicity Scattering Diagrams With One Soft Gluon
2389: \end{center}
2390:
2391: To discuss signature
2392: we focus on the first forms displayed.
2393: Comparing with our discussion of the diagrams of
2394: Fig.~20, we note that Figs.~31(a)and(b) again
2395: differ, kinematically, from Fig.~27 and Fig.~29(b) (respectively)
2396: only in that $P_+ \to -P_+$
2397: along the left-most vertical line. Therefore, if we
2398: evaluate the diagrams of Fig.~31 in the cross-channel physical
2399: region - with the sign of $P_+$
2400: reversed, they will be kinematically identical, respectively, to Figs.~27 and 29(b)
2401: and will give identical anomaly contributions, but with $S \to -S$.
2402: Therefore, these diagrams have anomaly
2403: contributions with the opposite sign to those of Figs.~26 and 29(b)
2404: and in a vector theory would provide a cancelation.
2405: In the present case, since the color factors are the same, the anomaly
2406: contributions of the two diagrams of Fig.~31 add
2407: to those of Figs.~27 and 29(b) in the even signature amplitude and produce a
2408: cancelation in the odd signature amplitude.
2409:
2410: As in our discussion of the one gluon diagrams,
2411: we can also obtain the signature from
2412: the $CP$ symmetry properties of the intermediate transverse momentum state.
2413: We consider, first, properties of the gluon component.
2414: Since the gluons have only QCD vertices, they can have simple
2415: transformation properties under $C$ and $P$ separately.
2416: Applying color charge conjugation to (\ref{lji+}) gives
2417: $$
2418: \eqalign{ [\lambda_j\lambda_i ~+~ \lambda_i\lambda_j]
2419: ~&\to~ (-1)^2 [\lambda_j^*\lambda_i^* ~+~ \lambda_i^*\lambda_j^*] \cr
2420: &=~ [\lambda_j\lambda_i ~+~ \lambda_i\lambda_j]^*
2421: ~=~\frac{2}{3} \delta_{ij}~
2422: +~ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2423: {\scriptstyle k}$}}~d_{ijk}[\lambda_k]^*}
2424: \auto\label{lji-c}
2425: $$
2426: The parity transformation reverses the transverse momentum of the gluons
2427: and so,
2428: because of the coupling (\ref{bxi11}) for the large transverse momentum gluon,
2429: again gives a minus sign. Therefore, the full effect of the $CP$
2430: transformation of the gluon component of the intermediate state is given by
2431: $$
2432: ~\frac{2}{3} \delta_{ij}~
2433: +~ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2434: {\scriptstyle k}$}}~d_{ijk}[\lambda_k]~~\to ~~
2435: -~\frac{2}{3} \delta_{ij}~
2436: -~ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2437: {\scriptstyle k}$}}~d_{ijk}[\lambda_k]^*
2438: \auto\label{cpg}
2439: $$
2440: Comparing (\ref{lji-c}) and (\ref{cpg}) with (\ref{lji+}) we observe that
2441: the color zero two gluon state carries negative $CP$ and is
2442: separately even under $C$ and odd under $P$.
2443: It therefore has ``normal'' color charge parity (equal to the number of gluons
2444: in the state) but has ``anomalous'' negative parity, producing an ``anomalous''
2445: negative signature. The color parity of the octet gluon state is also
2446: well-defined if we ignore $[\lambda_k] \to [\lambda_k]^*$ (which is, of course,
2447: compensated for by the quark/antiquark interchange discussed below)
2448: and is similarly even.
2449:
2450: To discuss the $CP$ transformation of the quark/antiquark pair
2451: we compare with our discussion in sub-section {\bf 3.15} and note, first, that
2452: the left-side coupling of Fig.~31(b) can be obtained, diagrammatically,
2453: from that of Fig.~27 by quark/antiquark interchange.
2454: The left-side coupling of Fig.~31(a) can similarly be obtained
2455: from that of Fig.~29(b). In the even signature
2456: amplitude we must evaluate the diagrams of
2457: Fig.~31 with $P_+ \to -P_+$ compared to the other diagrams.
2458: Therefore, quark/antiquark interchange now gives three kinematic changes of sign
2459: $$
2460: \eqalign{& (i) ~~~~~~~~ k_{\perp} \to -k_{\perp} \cr
2461: &(ii)~~~~~ ~~ k_-' \to - k_-' ~~~\hbox{from}~~~ P_+ \to -P_+ \cr
2462: &(iii) ~ ~~~~~k_-'' \to - k_-''~~~ \hbox{from}~~~ P_+ \to -P_+}
2463: \auto\label{sc2}
2464: $$
2465: When all four diagrams are
2466: added the amplitude is, kinematically,
2467: antisymmetric under quark/antiquark interchange. As a result,
2468: when the quark/antiquark pair has color zero it is, straightforwardly,
2469: negative under $CP$.
2470: Combined with the negative $CP$ of the color zero two gluon state this gives
2471: no change under the full
2472: $CP$ transformation, as is necessary to obtain an even signature amplitude.
2473:
2474: When octet color is involved, the color
2475: effect of interchanging the quark-antiquark pair will, as we already noted
2476: above, again be
2477: $[\lambda_h]~ \to~ [\lambda_h]^*$. Therefore,
2478: with the negative sign coming from the kinematic interchange, the
2479: complete $CP$ transformation on the quark-antiquark pair again
2480: combines with the octet part of (\ref{cpg}) to produce an overall positive
2481: $CP$ result for the full two gluon quark-antiquark state.
2482:
2483: \subhead{6.8 Infra-Red Cancelations}
2484:
2485: Since (\ref{amp2}) contains an infra-red divergence (at
2486: ${k_{\perp}''}^2 \sim 0$), and there is no external color,
2487: there must be other anomaly contributions that
2488: cancel this divergence. Before discussing the possible diagrams
2489: that could be involved, it will be useful to first discuss the lower limit on the
2490: ${k''_{\perp}}^2$ integration in Fig.~27.
2491:
2492: The momentum flow through the two lines that are put on-shell by the
2493: $k_-'$ and $k_-''$ integrations is shown in Fig.~32.
2494: \begin{center}
2495: \epsfxsize=2.5in
2496: \epsffile{ehes31.ps}
2497:
2498: Fig.~32 Momentum Flow for the $k_-'$ and $k_-''$ Integrations
2499: \end{center}
2500: Since $k'_{\perp}$ is large and $k''$ is small,
2501: all the large momenta flow through both lines. Therefore, the large momenta are
2502: significantly constrained by the $k'$ mass-shell condition before the
2503: $k''$ mass-shell condition is imposed. If we, temporarily, introduce a quark
2504: mass $m$ then
2505: the $k_-'$ mass-shell condition is $(P_+ + p'+k')^2 =m^2$ and, with this
2506: constraint, the mass-shell condition for $k_-''$ becomes
2507: $$
2508: \eqalign{ k_-''~&\sim~ \frac{- 2k_{\perp}''(p'+k')_{\perp} + m^2}{P_+} \cr
2509: &\centerunder{$\sim$}{\raisebox{-4mm}{${\scriptstyle k_{\perp}'' \to 0}$}}
2510: ~~~~\frac{m^2}{S^{\frac{1}{2}}}}
2511: \auto\label{llmt}
2512: $$
2513: If $k_+''$ is similarly constrained, and (to justify
2514: the reduction to a transverse momentum diagram) we require
2515: ${k_{\perp}''}^2
2516: ~\centerunder{$>$}{\raisebox{-2mm}{$\sim $}}~|k_-''k_+''|$,
2517: the lower limit for the $k_{\perp}''$ - integration is
2518: $$
2519: |k_{\perp}''|^2 ~~\sim ~~\frac{m^4}{ S}
2520: \auto\label{llmt1}
2521: $$
2522: and so an infra-red divergence appears at $|k_{\perp}''|^2=0$,
2523: as $S\to \infty$, of the form
2524: $$
2525: ~\int_{|k_{\perp}''|^2\sim \frac{m^4}{S}} ~\frac{d^2k_{\perp}''}{
2526: {k_{\perp}''}^2}~~\sim ~~ln~S ~-~ln~ m^4
2527: \auto\label{ird1}
2528: $$
2529:
2530: As we discuss briefly in the next Section, we expect
2531: the cancelation of the infra-red divergence (\ref{ird1}) to be a consequence
2532: of a Ward identity that results from attaching the soft gluon at every possible
2533: point around the effective triangle diagram.
2534: Consider, therefore, the attachment of the soft gluon, in $G_L$, according to the
2535: other possibilities illustrated in Fig.~25. With the attachment
2536: shown in Fig.~25(b) we obtain the full diagrams shown in Fig.~33.
2537: \begin{center}
2538: \epsfxsize=2in
2539: \epsffile{ehes63.ps}
2540: \hspace{0.8in}
2541: \epsfxsize=2in
2542: \epsffile{ehes631.ps}
2543:
2544: (a)\hspace{2.8in}(b)
2545:
2546: Fig.~33 Diagrams Obtained with the Soft Gluon Attachment of Fig.~25(b)
2547: \end{center}
2548: The mass-shell condition now has the form (\ref{k''os}) and so,
2549: if the lower limit for the $k_+'$ integration still has
2550: the form (\ref{llmt}), requiring
2551: $|k_{\perp}'|^2 > k_+'k_-'$ gives the lower limit
2552: $$
2553: |k_{\perp}'|^2 ~~\sim ~~\frac{m^2~ M}{S^{\frac{1}{2}}}
2554: \auto\label{llmt3}
2555: $$
2556: Therefore, if the diagrams of Fig.~33 give anomaly contributions they will have,
2557: as a factor, an infra-red divergence of the form
2558: $$
2559: ~ \int_{|k_{\perp}''|^2\sim m^2M /S^{\frac{1}{2}} }
2560: ~\frac{d^2k_{\perp}''}{{k_{\perp}''}^2}~~
2561: \sim ~~~\frac{1}{2}~ln~S ~-~ln~ m^2~-~ln~M
2562: \auto\label{ird2}
2563: $$
2564: If, instead, one attachment of the soft gluon is that
2565: of Fig.~25(c) rather than that of Fig.~25(b), as we have just discussed,
2566: there will clearly be a similar infra-red divergence.
2567:
2568: There is, however, a major reason why the diagrams of Fig.~33, ultimately,
2569: do not give an anomaly contribution. Because of the location
2570: of the soft gluon attachment, the sign change $(iii)$ in
2571: (\ref{sc2}) will not be present when $P_+ \to -P_+$ in the diagrams appropriately
2572: related to those of Fig.~33 in the same sign helicity amplitude. As a result,
2573: in the even signature anomaly amplitude,
2574: the quark-antiquark intermediate state must be positive under $CP$.
2575: Alternatively, if we add all the diagrams related to
2576: those of Fig.~33 that give all possible contributions
2577: to each $G_R$, we will obtain $G_R$ couplings to the
2578: quark-antiquark intermediate state that have negative $CP$. Therefore,
2579: when all diagrams related to those of Fig.~33 are added in the even signature
2580: amplitude, the resulting $G_L$ requires positive $CP$ for the quark-antiquark
2581: state, while the $G_R$ requires negative $CP$
2582: Consequently, if there is an anomaly
2583: contribution from any of the diagrams, it must cancel in the sum.
2584:
2585: This last problem similarly applies
2586: to all diagrams in which one soft gluon attachment is as
2587: in Fig.~25(c) while the other attachment is that of Fig.~25(a).
2588: As we will shortly discuss there will, nevertheless,
2589: be important anomaly contributions from
2590: diagrams in which the attachments of the second gluon, in both $G_L$ and $G_R$,
2591: are either of the form of Fig.~25(b) or Fig.~25(c). In this case, however,
2592: the second gluon necessarily has ``finite'' transverse momentum and so
2593: can not produce an infra-red divergence.
2594:
2595: To look further for a divergence that could cancel that due to Fig.~27,
2596: we must consider
2597: whether there are any new kinematic configurations, generating the anomaly
2598: and involving a soft gluon,
2599: that can not be viewed as a soft gluon accompanying the one gluon enhancement
2600: diagrams. In fact, if we are considering the
2601: attachment of the soft gluon at every possible
2602: point around the effective triangle diagram, there is
2603: one possibility that we have not
2604: yet included or discussed. This is to
2605: interchange the momenta of the two gluons involved in Fig.~26.
2606: In Fig.~34 we show the full diagram obtained from Fig.~27 by
2607: interchanging $k'$ and $k''$ in one side of the diagram relative to the other.
2608: \begin{center}
2609: \epsfxsize=2.5in
2610: \epsffile{ehes190.ps}
2611:
2612: Fig.~34 The Diagrams Obtained by Interchanging $k'$ and $k''$ in Fig.~27
2613: \end{center}
2614: The soft gluon is again indicated by a broken line. (Note that Fig.~34 is
2615: symmetric with respect to $k'$ and $k''$, if we interchange the roles
2616: of $G_L$ and $G_R$.)
2617:
2618: If we interchange $k'$ and $k''$ in Fig.~26, then
2619: the large transverse momentum flows into the triangle diagram at the left-most
2620: vertex while the single gluon vertex carries only small transverse momentum.
2621: This does indeed give an anomaly contribution. This contribution has, however,
2622: some different properties compared to those we have so far discussed.
2623: In Fig.~35 we compare the
2624: large momentum route for $k'$ with the possible routes for $k''$,
2625: around the same triangle.
2626: \begin{center}
2627: \epsfxsize=4in
2628: \epsffile{ehes29.ps}
2629:
2630: Fig.~35 (a) Momentum Route for $k_{\perp}'$ (b)
2631: Possible Routes for $k_{\perp}''$
2632: \end{center}
2633: We observe, first, that while the route for $k'$ large was determined by
2634: the particles we wished to put on shell, there are two possible
2635: routes if $k''_{\perp}$ is large. As illustrated, it could flow through either
2636: one or two quark propagators. From (\ref{bxi6}) and (\ref{bxi7}), we see that
2637: the anomaly contribution is obtained from an expansion in powers of
2638: $|k''_{\perp}/p'_{\perp}|$ (in which the first term does not contribute).
2639: Consequently,
2640: if $k''_{\perp}$ appears in only one propagator, rather than two,
2641: the anomaly contribution
2642: will be reduced by a factor of $1/2$ (in the even signature amplitude where
2643: both chiralities are added for each propagator). The sign will also be opposite.
2644: This is the normal ambiguity
2645: of the ultra-violet anomaly, that occurs because of the choice of momentum
2646: routing, which we expect to be determined by a Ward identity.
2647:
2648: If $k'$ is small, the $k''$ mass-shell condition
2649: does not constrain the large transverse momentum ($p_{\perp}'$)
2650: involved in the $k_-'$ mass-shell condition. As a result, the $k_-'$ mass-shell
2651: condition gives a constraint similar to (\ref{k''os}), i.e.
2652: $$
2653: k_-'~\sim ~ \frac{{p'}^2}{P_+} ~~\sim~~\frac{\epsilon~ M ~S^{\frac{1}{2}}}
2654: {S^{\frac{1}{2}}}~~\sim ~~\epsilon~M
2655: \auto\label{llmt2}
2656: $$
2657: Since the $k_+'$ integration
2658: in $G_R$ has the lower limit (\ref{llmt}),
2659: we will again obtain an infra-red divergence of the form of
2660: (\ref{ird2}). In this case, however, since there is a gluon attached
2661: to both the left and right side quark lines, there is no
2662: $CP$ conflict. Also, from Fig.~36, it is clear that
2663: neither gluon interferes with the kinematics
2664: of the quark sub-amplitude, which remains such that the vector meson lines can
2665: consistently be placed on-shell.
2666: \begin{center}
2667: \epsfxsize=2.2in
2668: \epsffile{ehes672.ps}
2669:
2670: Fig.~36 Two Gluons Accompanying the Quark Sub-Amplitude
2671: \end{center}
2672: In fact, if the anomaly has the sign and magnitude
2673: obtained from the second routing of Fig.~35(b) (which is the ``normal''
2674: routing), there will be a cancelation
2675: with the divergence obtained from Fig.~27 when the diagram of Fig.~34 is
2676: added and the two contributions, from $k''$ large and $k'$ small, and
2677: $k'$ large and $k''$ small, are combined.
2678:
2679: At this point we note that
2680: $k'$ is the total momentum flowing in at a ``vector'' vertex of
2681: an effective triangle diagram with an anomaly. In this case, as we discuss in
2682: the next Section (and illustrate for the usual triangle anomaly in Appendix C),
2683: we expect that the appropriate Ward identity, which would determine which routing
2684: in Fig.~35(b) is correct, will involve both the large internal
2685: momentum region of the triangle diagram that generates the anomaly and a
2686: small internal momentum region that produces a very different kinematic
2687: form, containing the ``anomaly pole''.
2688: The discussion of the anomaly pole and pion wee gluon couplings
2689: that we have given in \cite{arw02} should, essentially,
2690: carry over to an infra-red analysis of effective triangle vertices
2691: that would be the analogue of the ultra-violet analysis presented in this
2692: paper. This analysis must be carried out before we can
2693: establish, in detail, how the (reggeon) Ward identities are satisfied,
2694: and that infra-red divergences are indeed eliminated when the diagram
2695: of Fig.~34(a) is added to that of Fig.~27.
2696:
2697: \subhead{6.9 Anomalous Color Parity Gluons}
2698:
2699: In the final form of two gluon anomaly contributions that we consider
2700: the second gluon transverse momentum $k''_{\perp}$ is
2701: neither very small, nor grows with $S$. It is ``finite'', i.e. $O(M^2)$.
2702: In the diagrams we consider, the above discussion implies that
2703: the kinematics of the anomaly prevent the
2704: second gluon from carrying very small transverse momentum.
2705: In particular, we consider diagrams of the form
2706: shown in Fig.~37, in which the second gluon attachment is the same
2707: in both $G_L$ and $G_R$, and has the form of either Fig.~25(b) or Fig.~25(c).
2708: \begin{center}
2709: \epsfxsize=1.5in
2710: \epsffile{ehes70.ps}
2711: \hspace{0.3in}
2712: \epsfxsize=1.5in
2713: \epsffile{ehes654.ps}
2714: \hspace{0.3in}
2715: \epsfxsize=1.5in
2716: \epsffile{ehes65.ps}
2717:
2718: (a)\hspace{1.8in}(b)\hspace{1.8in}(c)
2719:
2720: Fig.~37 Gluon Attachments (a) - as in Fig.~25(b), or (b) and (c) - as in Fig.~25(c)
2721:
2722: \end{center}
2723: The broken gluon
2724: line in these diagrams now indicates finite transverse momentum
2725: and the lines put on-shell correspond, as illustrated, to two cuts through
2726: each diagram. The combination of particle poles giving these two discontinuities,
2727: together with the off-shell loop, that occurs either at the top, in the middle,
2728: or at the bottom, in the three diagrams, is now responsible for the anomaly.
2729: (As usual, a closely related
2730: set of diagrams is obtained by twisting simultaneously both the $G_L$ and
2731: the $G_R$ in each of these diagrams. For Fig.~37(a) the same diagram is
2732: actually obtained, but the kinematic regions for the two gluons are
2733: interchanged.)
2734:
2735: To form signatured amplitudes, as before, we consider the same helicity
2736: scattering amplitudes shown in Fig.~38.
2737: \begin{center}
2738: \epsfxsize=1.5in
2739: \epsffile{ehes73.ps}
2740: \hspace{0.3in}
2741: \epsfxsize=1.5in
2742: \epsffile{ehes751.ps}
2743: \hspace{0.3in}
2744: \epsfxsize=1.5in
2745: \epsffile{ehes75.ps}
2746:
2747: (a)\hspace{1.8in}(b)\hspace{1.8in}(c)
2748:
2749: Fig.~38 Same Helicity Diagrams Related to the Diagrams of Fig.~37
2750: \end{center}
2751: Again we can argue that these diagrams are related to those of Fig.~37 by
2752: $P_+ \to - P_+$ and so the same anomaly amplitude is obtained, but with
2753: $S \to -S$. Therefore, these diagrams give anomaly contributions
2754: that add to those of Fig.~37 in the even signature amplitude.
2755: However, as in our
2756: discussion of the diagrams of Fig.~23, because the finite gluon
2757: is not attached to the fast left side quark, the sign change $(iii)$ in
2758: (\ref{sc2}) will not be present when $P_+ \to -P_+$ in the diagrams
2759: of Fig.~38. Consequently, the $G_L$ coupling obtained by adding the diagrams
2760: of Fig.~38 to those of Fig.~37, requires even $CP$ for the quark-antiquark
2761: intermediate state. When the corresponding diagrams are added,
2762: this argument will similarly apply to $G_R$. As a result, the quark-antiquark
2763: state is necessarily positive under $CP$. This can be
2764: consistent with even signature for the complete intermediate state
2765: only if the two gluon state is also
2766: even under $CP$. This can, in turn, only be the case if the two gluon state
2767: carries antisymmetric octet color and so has,
2768: ``anomalous'', odd color charge parity (not equal to the number of gluons).
2769:
2770: In the two gluon anomaly contributions that we have discussed in previous
2771: sub-sections, the two gluon state has carried normal color charge parity
2772: because the addition of twisted diagrams gave the symmetric color factor.
2773: We must consider, therefore, whether there are also ``twisted diagrams''
2774: related to those of Fig.~37 which cancel the antisymmetric part of the color
2775: factor.
2776:
2777: Twisted diagrams related to the diagrams of Fig.~37 are shown in Fig.~39.
2778: For Figs.~39(a) and (b), the color factor is indeed reversed compared,
2779: relatively, to Figs.~37(a) and (b). Also, it is clear that,
2780: in Figs.~39(a) and (b), the appropriate (hatched) lines can be consistently
2781: placed on shell. (This is not true for other diagrams that could, potentially,
2782: be related, by twisting, to either of Fig.37(a) or (b).)
2783: Therefore, as before, adding the diagrams of Figs.~39(a) and (b) to
2784: those of Figs.~37(a) and (b) gives the
2785: symmetric color factor and so, because the quark-antiquark state has even $CP$,
2786: the anomaly contribution must cancel in the sums of these diagrams.
2787: \begin{center}
2788: \epsfxsize=1.5in
2789: \epsffile{ehes72.ps}
2790: \hspace{0.5in}
2791: \epsfxsize=1.5in
2792: \epsffile{ehes721.ps}
2793:
2794: (a)\hspace{2in}(b)
2795:
2796: \epsfxsize=1.4in
2797: \epsffile{ehes655.ps}
2798: \hspace{0.5in}
2799: \epsfxsize=1.4in
2800: \epsffile{ehes652.ps}
2801:
2802: (c1)\hspace{2in}(c2)
2803:
2804: Fig.~39 Diagrams Related to the Diagrams of Fig.~33 by Twisting
2805: \end{center}
2806:
2807: Both Fig.~39(c1) and Fig.~39(c2) can be regarded as twisted relative to
2808: Fig.~37(c). For Fig.~39(c1) the color factor is reversed
2809: compared to Fig.~37(c), but the hatched lines clearly can
2810: not be consistently placed on-shell in the physical region.
2811: For Fig.~39(c2) it seems probable
2812: that the hatched lines can be placed on-shell consistently, even though
2813: the necessary cuts would cross. The issue is irrelevant, however, since
2814: Fig.~39(c2) has the same color factor as Fig.~37(c).
2815: In fact, there is no twisted diagram corresponding to Fig.~37(c) which has
2816: a reversed color factor and in which all the
2817: necessary hatched lines can be consistently placed on-shell.
2818:
2819: Since we can regard the color factor for Fig.~37(c) as
2820: the sum of the symmetric and antisymmetric factors,
2821: the antisymmetric, odd color parity,
2822: component will be selected for the
2823: anomaly contribution from this diagram (and it's same helicity counterpart).
2824: Similarly for Fig.~39(c2), if it contributes.
2825: Together with the contribution of corresponding diagrams obtained by
2826: twisting both $G_L$ and $G_R$, these will be the only
2827: anomaly contribution from diagrams in which
2828: the second gluon carries ``finite''transverse momentum. In all cases, the
2829: second gluon is
2830: attached as in Fig.~25(c), and not as in Fig.~25(b). The finite gluon
2831: contributions have the important property that the two gluon state carries
2832: ``anomalous color parity''. This is significant because,
2833: as we emphasized in the Introduction, reggeized gluon exchanges that appear in
2834: vector theory perturbative calculations\cite{fkl}-\cite{arw93}
2835: carry normal color parity. The appearance of anomalous color parity gluon
2836: states is, therefore, a direct consequence of the presence of the anomaly.
2837:
2838: \subhead{6.10 General Multigluon Color Factors}
2839:
2840: Before considering more complicated multigluon configurations
2841: we give a general discussion of multigluon
2842: color factors which generalizes the previously discussion
2843: of two gluon color factors.
2844:
2845: We note first that we can obtain (\ref{lji+}) and (\ref{lji-}) from
2846: (\ref{lij}) by a more general argument than just the symmetry
2847: and antisymmetry of the $d$ and $f$ tensors. Consider
2848: a product of $\lambda$ - matrices
2849: $$
2850: P_{1n}~=~\prod_{j=1}^n ~\lambda_{i_j}
2851: \auto\label{p1n}
2852: $$
2853: together with the product taken in the reverse order
2854: $$
2855: P_{n1}~=~\prod_{j=n}^1 ~\lambda_{i_j}
2856: \auto\label{pn1}
2857: $$
2858: Using (\ref{lij}) extensively we can write
2859: $$
2860: P_{1n}~=~A_{1n}~+~
2861: ~\centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2862: {\scriptstyle k}$}} ~B_{1nk}~\lambda_k
2863: \auto\label{p1n1}
2864: $$
2865: where $A_{1n}$ multiplies the unit matrix and
2866: both $A_{1n}= A_{i_1,i_2,..,i_n}$ and $B_{1n}= B_{i_1,i_2,..,i_n,k}$
2867: contain combinations of $f$ and $d$ tensors. Similarly, we can write
2868: $$
2869: P_{n1}~=~A_{n1}~+~
2870: ~\centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
2871: {\scriptstyle k}$}}~B_{n1k}~\lambda_k
2872: \auto\label{pn11}
2873: $$
2874: (\ref{p1n1}) and (\ref{pn11}) decompose $P_{1n}$ and $P_{n1}$, respectively,
2875: into a sum of color singlet and color octet contributions.
2876:
2877: It follows from the hermiticity of the $\lambda_i$
2878: that
2879: $$
2880: (P_{n1})^T~=~ \prod_{j=n}^1~(\lambda_{i_j})^T~=~ \prod_{j=n}^1~(\lambda_{i_j})^*
2881: ~=~(P_{1n})^*
2882: \auto\label{tcc}
2883: $$
2884: where $(~)^T$ denotes transposition and $(~)^*$ denotes complex conjugation.
2885: Equivalently,
2886: $$
2887: P_{n1}~=~\bigl(~(P_{1n})^T~\bigr)^*
2888: \auto\label{tcc1}
2889: $$
2890: As a result
2891: $$
2892: A_{n1}~=~A_{1n}^*~,~~~~~B_{n1k}~=~B_{1nk}^*
2893: \auto\label{tcc2}
2894: $$
2895: or, equivalently,
2896: $$
2897: \eqalign{P_{1n}~+~P_{n1}~&= ~2~Re(A_{1n}) ~+~2~Re(B_{1nk})~\lambda_k \cr
2898: P_{1n}~-~P_{n1}~&=~ 2i~Im(A_{1n})~+~2i~Im(B_{1nk})
2899: ~\lambda_k}
2900: \auto\label{tcc3}
2901: $$
2902: which gives ((\ref{lji+}) and (\ref{lji-}), as a very simple case.
2903:
2904: Since the $f$ and $d$ tensors are both real, it
2905: follows from (\ref{lij}) that a factor of $i$ is always
2906: accompanied by an $f$ tensor. Therefore, the real and imaginary parts
2907: of both $A_{1n}$ and $B_{1n}$ contain, repectively,
2908: even and odd numbers of $f$ tensors. If we then consider
2909: $$
2910: Tr\{\Sigma_{i_1,i_2,..,i_n}\bigl(P_{1n}~\pm~P_{n1}\bigr)\bigl(P_{1n}\bigr)\}
2911: \auto\label{tra}
2912: $$
2913: the distinct color of $A_{1n}$ and $B_{1n}$ and the distinct
2914: symmetry properties of the real and imaginary coefficients
2915: implies that
2916: $$
2917: Tr\{\Sigma_{i_1,i_2,..,i_n}\bigl(P_{1n}~+~P_{n1}\bigr)\bigl(P_{1n}\bigr)\}
2918: ~=~2~\Sigma_{i_1,i_2,..,i_n}\bigl([Re(A_{1n})]^2 ~+~[(Re(B_{1n})]^2~\bigr)
2919: \auto\label{tra2}
2920: $$
2921: which is a sum of the squares of color factors for color zero and color octet
2922: states which contain an even number of $f$ - tensors and so
2923: describe normal color parity gluon states.
2924: Similarly,
2925: $$
2926: Tr\{\Sigma_{i_1,i_2,..,i_n}\bigl(P_{1n}~-~P_{n1}\bigr)\bigl(P_{1n}\bigr)\}
2927: ~=~-2~\Sigma_{i_1,i_2,..,i_n}\bigl(~[Im(A_{1n})]^2~+~[Im(B_{1n})]^2~\bigr)
2928: \auto\label{tra3}
2929: $$
2930: which is a similar sum of squares of color factors which, because they
2931: contain an odd number of $f$ tensors,
2932: describe gluon states with anomalous color charge parity.
2933:
2934: In the two gluon states that we have so far considered, all color factors
2935: apart from the first term
2936: in (\ref{tra3}) have appeared. This term corresponds to a color zero
2937: anomalous color parity multigluon configuration. It will appear in the
2938: three gluon diagrams that we discuss next.
2939: We continue to
2940: confine our discussion to a single large transverse momentum gluon and consider
2941: only multiple soft or finite gluon contributions that do not
2942: involve any factors of $ln S$. (In general,
2943: we anticipate that higher order logarithms
2944: lead to the separate reggeization of each of
2945: the transverse momentum gluons in the diagrams we study.)
2946:
2947: \subhead{6.11 Two Soft Gluons}
2948:
2949: We begin with two soft gluons and consider diagrams
2950: that contribute to the transverse momentum diagram of Fig.~40.
2951: \begin{center}
2952: \epsfxsize=1.8in
2953: \epsffile{ehes40.ps}
2954:
2955: Fig.~40 A Transverse Momentum Diagram With Two Soft Gluons
2956: \end{center}
2957: We will straightforwardly
2958: obtain an anomaly enhanced amplitude, as before, from diagrams in which
2959: two soft gluons are attached to
2960: the external effective point vertices, as illustrated in Fig.~41.
2961: \begin{center}
2962: \epsfxsize=2in
2963: \epsffile{ehes43.ps}
2964: \hspace{0.7in}
2965: \epsfxsize=2in
2966: \epsffile{ehes74.ps}
2967:
2968: (a)\hspace{2.8in}(b)
2969:
2970: Fig.~41 Diagrams With Two Soft Gluons
2971: \end{center}
2972: Because all the large momenta pass through both soft gluon attachments and
2973: are constrained by the $k_-'$ integration, the infra-red scale is the same for
2974: both and is $m^2/S^{\frac{1}{2}}$, as in (\ref{llmt}) and (\ref{llmt1}).
2975: Infra-red cancelations can, presumably, be discussed in the same manner as
2976: in our discussion of one soft gluon diagrams.
2977: The twisted diagram of Fig.~41(b) again reverses the color matrix multiplication
2978: of Fig.~41(b) and so
2979: the sum of the two diagrams gives a color factor of the form of
2980: (\ref{tra2}), corresponding to normal color parity for the complete
2981: three gluon transverse state
2982:
2983: To check that the anomaly amplitude obtained is even signature we note that
2984: the anomaly
2985: diagrams in the same sign helicity amplitude
2986: will again be related to the opposite sign diagrams by $P_+ \to -P_+$.
2987: For example, the diagram of Fig.~37(a) will be related to
2988: a same sign helicity amplitude diagram as illustrated in Fig.~42
2989: \begin{center}
2990: \parbox{2in}{\epsfxsize=2in
2991: \epsffile{ehes43.ps}}\parbox{1in}{
2992: \centerline{$\longleftrightarrow$}}
2993: \parbox{2.2in}{\epsfxsize=2.2in
2994: \epsffile{ehes60.ps}}
2995:
2996: Fig.~42 Related Diagrams in Different Helicity Amplitudes
2997: \end{center}
2998: Since the color factors are the same, these amplitudes are again related
2999: by $S\to -S$ and will add in the even signature amplitude (and would
3000: cancel in a vector theory).
3001:
3002: As before, the full
3003: gluon state will have negative parity because of the single large
3004: transverse momentum gluon. Combined with the negative (normal) color parity
3005: this implies that the three gluon state is even under $CP$.
3006: Repeating the discussion of the quark/antiquark state that we gave for
3007: the one soft gluon amplitudes we find that there is an additional change of
3008: sign from $P_+ \to -P_+$ that results from the additional soft gluon. As a result,
3009: quark/antiquark interchange gives no kinematic
3010: change of sign and the color charge parity
3011: transformation simply gives $\lambda_h \to \lambda_h^*$. Consequently,
3012: the full quark/antiquark/multigluon intermediate
3013: state is even under $CP$, as required for even signature.
3014: (We should note, however, that although the color zero three gluon state
3015: has normal color charge parity, it is ``anomalous'' in that it has
3016: negative parity, giving an ``anomalous'' positive signature.)
3017:
3018: \subhead{6.12 Color Zero Anomalous Color Parity }
3019:
3020: Consider, next, adding a soft gluon
3021: to the diagrams of Figs.~37(c) and 38(c), in which there
3022: is already one finite transverse momentum gluon present.
3023: If the soft gluon is attached to the left and right side
3024: quark lines, the resulting
3025: opposite sign and same sign helicity
3026: diagrams are shown, respectively, in Figs.~43(a) and (b).
3027: \begin{center}
3028: \epsfxsize=1.8in
3029: \epsffile{ehes761.ps}
3030: \hspace{0.7in}
3031: \epsfxsize=2.1in
3032: \epsffile{ehes7610.ps}
3033:
3034: (a)\hspace{2.8in}(b)
3035:
3036: Fig.~43 Anomalous Color Parity Diagrams With One Soft and One Finite Gluon
3037: \end{center}
3038: The color factor for both diagrams is the same, i.e.
3039: $$
3040: \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
3041: {\scriptstyle i,j,k}$}}
3042: ~~Tr\{\lambda_k\lambda_j\lambda_i\lambda_i\lambda_j\lambda_k\}
3043: \auto\label{3gcf1}
3044: $$
3045:
3046: If we pick out the color zero intermediate state then the color factors on the
3047: left and right side of the diagrams must separately factor into traces.
3048: Therefore, we can write
3049: $$
3050: \eqalign{ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
3051: {\scriptstyle i,j,k}$}}
3052: ~~Tr\{\lambda_k\lambda_j\lambda_i\lambda_i\lambda_j\lambda_k\}
3053: &\sim~
3054: \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
3055: {\scriptstyle i,j,k}$}}
3056: Tr\{\lambda_j\lambda_i\lambda_k\}~
3057: Tr\{\lambda_i\lambda_j\lambda_k\} ~~+~\cdots\cr
3058: &\sim ~ \centerunder{$\hbox{\Large $\Sigma$}$}{\raisebox{-3mm}{$
3059: {\scriptstyle i,j,k}$}}~f_{ijk}^2~+~\cdots }
3060: \auto\label{3gcf3}
3061: $$
3062: where the term we have shown explicitly is the color zero,
3063: anomalous (even) color parity, term.
3064:
3065: Because, the diagrams of Fig.~43 contain one less gluon attached to the
3066: fast quark lines than the diagrams of Fig.~42, the quark-antiquark
3067: component of the transverse state, in the even signature amplitude,
3068: will be odd under $CP$. Therefore, to obtain even signature overall
3069: the gluon component must also be odd under $CP$, implying that is
3070: even under color charge conjugation. Consequently, the anomalous color parity term,
3071: shown explicitly in (\ref{3gcf3}),
3072: is selected for the color zero component of the
3073: combination of the diagrams of Fig.~43 in the even signature amplitude.
3074:
3075: \subhead{6.13 Multigluons}
3076:
3077: Clearly we could generalise the foregoing
3078: discussion to a variety of multigluon configurations involving
3079: combinations of soft and finite gluons, with effective vertices
3080: of the form illustrated in Fig.~44.
3081: \begin{center}
3082: \epsfxsize=3.5in
3083: \epsffile{ehes281.ps}
3084:
3085: (a)\hspace{1.2in}(b)\hspace{1.2in}(c)
3086:
3087: Fig.~44 Multi-gluon Effective Vertices
3088: \end{center}
3089: However, having established the coupling of color zero anomalous color parity
3090: gluons, we have all the general properties that we require for the discussion
3091: of the next Section.
3092:
3093: Since a large transverse momentum gluon can give a scaling contribution of
3094: the form
3095: $$
3096: \sim ~~~\int_{ (k_{\perp}'')^2\sim\sqrt{s}}~
3097: \frac{d^2 k_{\perp}''}{(k_{\perp}'')^2}
3098: \auto\label{scc}
3099: $$
3100: it is also possible for additional large transverse momentum gluons
3101: to participate in the enhancement effect. Potentially, this could be an
3102: elaborate phenomenon involving, presumably,
3103: the reggeization of both quarks and gluons and, in higher orders, scaling
3104: properties of reggeized gluon interactions, as well as
3105: the evolution of $\alpha_s$. However, since we will argue, in the next Section,
3106: that large transverse momentum anomaly contributions are unphysical, there
3107: seems little point in exploring the issue any further.
3108: In part, we discuss the analagous infra-red phenomenon in the next Section.
3109:
3110: \newpage
3111:
3112: \mainhead{7. REGGEON WARD IDENTITIES, CUT-OFFS, AND INFRA-RED DIVERGENCES}
3113:
3114: Our calculations in the previous Sections have demonstrated that the
3115: anomaly enhanced diagrams, some of which contain
3116: anomalous gluons, provide
3117: the dominant contribution in the exchange channel we have considered.
3118: However, as we remarked in the Introduction, we believe that the power enhancement
3119: involved should not be present in physical amplitudes. Assuming that there is no
3120: perturbative cancelation, via some mechanism
3121: that has yet to be elucidated, then obtaining ``physical'' amplitudes without the
3122: enhancement is, a priori, a challenging problem.
3123:
3124: In this Section we will briefly outline how we anticipate the desired physical
3125: amplitudes can be obtained. The essential point will be
3126: that the contribution of the anomaly
3127: diagrams is very different if we take
3128: the regge limit before or after the removal of an
3129: ultra-violet transverse momentum cut-off. This cut-off introduces
3130: infra-red divergences and if it is removed only at a
3131: very late stage, as we will propose, then
3132: the result obtained will also depend
3133: on whether all orders sums are performed before or after it is removed
3134: and on how, and at what stage,
3135: infra-red cut-offs (in the form of gluon and quark masses) are removed.
3136: This ambiguity
3137: is the essence of the anomaly and it would not be surprizing
3138: if there is a unique procedure that is necessary to obtain the right
3139: ``physical'' answer.
3140:
3141: As we have already noted, both in the Introduction and in the previous
3142: Section, a study of the
3143: infra-red anomaly contributions of the diagrams we have considered, that
3144: matches the present study of ultra-violet contributions, will be necessary
3145: before any detailed arguments can be carried through. To fully elucidate
3146: infra-red anomaly contributions it will surely be necessary to abandon
3147: the restriction to forward kinematics and transverse polarizations that has
3148: greatly simplified the foregoing discussion. Nevertheless, based on our
3149: experience with hadron scattering amplitudes\cite{arw02},
3150: we believe that a complete procedure for obtaining physical
3151: amplitudes can be developed utilizing the following, briefly summarized,
3152: properties. Our hope is that since the present starting point is
3153: much simpler than the hadronic problem, the analysis will be correspondingly
3154: more straightforward.
3155:
3156:
3157: \subhead{7.1 Ward Identity Consequences and a Transverse Momentum Cut-Off}
3158:
3159: Gauge invariance implies that a general amplitude $\VEV{A_{\mu}(q)~...~}$
3160: with all (external) lines
3161: on shell except for one gluon that carries four momentum $q$
3162: satisfies the simple Ward identity
3163: $$
3164: q_{\mu}~\VEV{A_{\mu}(q)~...~}~=~0
3165: \auto\label{wd2}
3166: $$
3167: This identity usually follows\cite{gth}, at a given order in perturbation theory,
3168: only after the zero momentum gluon has been attached to the remainder
3169: of the diagram at all possible points. It is well-known that this
3170: identity, in turn,
3171: implies that the gluon amplitude vanishes at zero four momentum.
3172: Also, from analagous Ward identities\cite{gth}, a
3173: similar result holds when more than one gluon carries vanishing four momentum.
3174: (If the gauge symmetry is spontaneously-broken then, of course, the Ward identities
3175: corresponding to (\ref{wd2}), such as the ``electroweak Ward identities''
3176: referred to in sub-section {\bf 5.3}, have additional mass terms which prevent
3177: the infra-red vanishing of amplitudes.)
3178:
3179: The vanishing of a loop amplitude when external momenta are
3180: small compared to (``finite'') internal
3181: momenta also implies, generally, a suppression of
3182: internal momenta that are large compared to, finite, external momenta.
3183: If, however, there is an external axial current
3184: producing an anomaly contribution in a loop, then the situation is different.
3185: In this case, as we briefly review in Appendix C, in addition to the
3186: well-known anomalous Ward identity \cite{awb} for the axial current, vector
3187: Ward identities require a cancelation between separate contributions,
3188: with different kinematic structure, from large and
3189: small internal momentum regions. In particular,
3190: the large momentum anomaly contribution (\ref{uvco})
3191: cancels with an infra-red term that, in special momentum
3192: configurations (and only when the quarks are massless),
3193: reduces to a simple pole in the axial vector channel.
3194: This is the ``anomaly pole'' that can, in the right circumstances, be
3195: interpreted as a Goldstone boson pole, signaling chiral symmetry breaking.
3196: In addition to the discussion in Appendix C,
3197: a detailed analysis of the anomaly pole, and the
3198: internal momentum region generating it, can be found in\cite{arw02}.
3199:
3200: In general, the above properties of gluon amplitudes, as functions
3201: of four-dimensional momenta, transfer directly\cite{arw98} to corresponding
3202: properties for the multi-gluon (multi-reggeon)
3203: transverse momentum couplings that we discuss in this paper, as functions of
3204: transverse momenta. The linear vanishing
3205: when transverse momenta are scaled to
3206: zero is sufficient to eliminate infra-red divergences in the
3207: transverse momentum diagrams that we consider. If there is an anomaly in a
3208: transverse momentum coupling then, as we already noted in our discussion
3209: of infra-red cancelations in {\bf 6.8}, there will be
3210: large/small internal momentum cancelations in the
3211: associated transverse momentum (reggeon) Ward identities that parallel the
3212: cancelations that take
3213: place in the four-dimensional Ward identities. We expect that
3214: infra-red divergences will be avoided, in part, by cancelation of the ultraviolet
3215: anomaly contributions we have found with infra-red ``anomaly pole'' contributions.
3216: In fact, the coefficient of the
3217: anomaly in a transverse momentum coupling (which we did not determine)
3218: should be fixed by this cancelation.
3219:
3220: If we impose a transverse momentum cut-off in all internal loop
3221: integrals of the diagrams we consider, this cut-off will be present in
3222: all transverse momentum diagram integrals and also within the loop integrals giving
3223: the external couplings. A cut-off in the transverse momentum diagram
3224: integrals is gauge invariant. A priori, however, in the external couplings
3225: such a cut-off is not gauge invariant. Therefore, if we take the regge limit with
3226: a transverse momentum cut-off imposed, it will be a serious question whether
3227: gauge invariance is restored by removing the cut-off after the limit. For
3228: the present we note only that in \cite{arw02}
3229: we argued that anomaly pole contributions to infra-red divergent amplitudes
3230: are gauge invariant.
3231:
3232: In the infra-red region, we anticipate that there will be
3233: effective triangle diagram contributions to $G_L$ and $G_R$ couplings
3234: in which small transverse momentum gluons
3235: couple at all three vertices, as illustrated in Fig.~45.
3236: \begin{center}
3237: \leavevmode
3238: \epsfxsize=2.5in
3239: \epsffile{ehes53.ps}
3240:
3241: Fig.~45 An Infra-Red Effective Triangle Diagram
3242: \end{center}
3243: Based on the analysis of the previous Section, both
3244: normal and anomalous color parity multigluon
3245: states should couple. With a transverse momentum cut-off imposed
3246: we expect that, when the total gluon transverse momentum vanishes
3247: the corresponding Ward identity will fail
3248: and there will be a non-zero coupling involving (when the quarks are
3249: massless) the anomaly pole.
3250: As a consequence, in transverse momentum diagrams of the form of Fig.~40 (with
3251: all gluons soft) there will be a logarithmic infra-red divergence of the form
3252: $$
3253: \int^{\lambda_{\perp}}~\frac{dQ^2}{Q^2}
3254: \auto\label{ind}
3255: $$
3256: where $Q$ is the sum of all gluon transverse momenta.
3257:
3258: A priori, the anomaly pole can appear in both $G_L$ and $G_R$. However, at $t=0$,
3259: where the pole should appear, a finite light-like momentum can be exchanged
3260: which can be parallel to either $P_+$ or $P_-$. We suspect that this light-cone
3261: momentum determines whether the pole appears via $G_L$ or $G_R$. Clearly,
3262: a detailed study, of the kinematics/polarizations
3263: and kinematic forms associated with the appearance of the anomaly pole will be
3264: needed to be sure that, in the full amplitude, there
3265: is a simple pole with the appropriate residue to be associated with a pion.
3266:
3267: \subhead{7.2 Transverse Momentum Infra-red Divergences}
3268:
3269: Since the divergence (\ref{ind}) is not removed by external couplings
3270: (with a transverse cut-off), we must consider the effect of
3271: (all orders) interactions
3272: amongst the gluons. In the lowest-order diagrams we expect the divergence to be
3273: present for both normal and anomalous color parity gluon states.
3274: There may also be additional divergent transverse momentum configurations.
3275: However, as we now describe,
3276: when we sum all infra-red divergences to all orders we expect
3277: that (\ref{ind}) is the only divergence that survives,
3278: and then only for anomalous gluons.
3279:
3280: We can summarize the general nature of gluon infra-red
3281: transverse momentum divergences and the role of a transverse
3282: momentum cut-off, very briefly, as follows. An expanded version of
3283: this summary can be found in \cite{arw02}. For reasons that will become
3284: apparent in the
3285: next sub-section we specifically discuss the case of SU(2) color, although
3286: all the properties we describe remain the same for higher gauge groups.
3287:
3288: The self-interactions of an $N$ gluon transverse momentum state
3289: $T_N$ are described by dimensionless ``kernels'' $
3290: K^I_N(\hdots,k_i,\hdots,{k_j}', \hdots)$, where $I$ denotes
3291: SU(2) color. (Each iteration of a kernel produces an additional factor of
3292: $ln~ S$, or $(J-1)^{-1}$ in the $J$ - plane, which we will not show explicitly.)
3293: When the $t$-channel
3294: color is non-zero the infra-red divergences related
3295: to reggeization do not cancel and
3296: $$
3297: \int~ \prod_{i=1}^N~ \frac{d^2k_i}{k_i^2 }
3298: ~K^I_N(\hdots,k_i,\hdots,{k_j}', \hdots)~\to ~\infty~,~~~Q^2, I \neq ~0
3299: ~~~~~\bigl(~Q=\Sigma_i k_i~\bigr)
3300: \auto\label{KIN}
3301: $$
3302: As a result, the sum of all diagrams in any colored
3303: channel exponentiates to zero as illustrated in Fig.~46.
3304: \begin{center}
3305: \leavevmode
3306: \epsfxsize=4.5in
3307: \epsffile{ehes50.ps}
3308:
3309: Fig.~46 Iteration of a Gluon Kernel $K_N$.
3310: \end{center}
3311: $G^I_N$ is an external coupling
3312: analagous to the $G_L$ and $G_R$ appearing in the previous Sections.
3313:
3314: When $I=0$ and $Q^2 ~\neq 0$, there is a cancelation of
3315: divergences in the $K^0_N$. (This is the infra-red
3316: finiteness property which is extensively exploited in BFKL applications.)
3317: At the leading-log level, infra-red finiteness leads directly to
3318: conformal scale invariance. When renormalization effects are introduced,
3319: scale invariance is lost in the ultra-violet region.
3320: Scale invariance properties may still be present in the infra-red region
3321: (in particular, they will be present
3322: if there is an infra-red fixed point for the gauge coupling).
3323: In this case, the kernels $K^0_N$ will scale canonically
3324: as $Q^2 \to 0$ so that, with a transverse momentum cut-off $\lambda_{\perp}$,
3325: $$
3326: \int_{|k_i|^2, |k^\prime_j|^2 ~<~\lambda_{\perp}} ~\prod_i \frac{d^2k_i}{ k^2_i}
3327: ~\prod_j \frac{d^2k^{\prime}_j}{ {k'}^2_j}~
3328: K^0_N ( k_1, \cdots k_N, k^{\prime}_1, \cdots k^{\prime}_N)
3329: ~ \sim ~ ~\int^{\lambda_{\perp}} \frac{dQ^2}{ Q^2}
3330: \auto\label{irscl}
3331: $$
3332: The presence of the cut-off ensures that this divergence is unambiguously isolated
3333: from ultra-violet divergences with which it might mix.
3334:
3335: This is the same divergence as (\ref{ind}), which appears in the lowest-order
3336: diagrams. The kernels $K^0_N$ have Ward
3337: identity zeroes which result in the special property that iteration of
3338: any $K^0_N$ does not increase the degree of divergence. Instead, there is
3339: a distinct contribution from each $T_N$
3340: and the residue of the divergence
3341: can be written in a factorized form, as illustrated in Fig.~47.
3342: \begin{center}
3343: \leavevmode
3344: \epsfxsize=5.9in
3345: \epsffile{ehes51.ps}
3346:
3347: Fig.~47 Isolation of the Divergence Associated with $T_N$.
3348: \end{center}
3349: If there is no
3350: Ward identity zero in the external couplings $G^0_N$, (\ref{irscl}) is
3351: a potential source of a simple infra-red divergence at $Q^2=0$.
3352:
3353: Similar properties to the above hold for
3354: the interactions of gluons with quarks. Crucially, however, there
3355: is no kernel describing a
3356: transverse momentum interaction between a quark/antiquark pair and an
3357: anomalous gluon state. This is because anomalous gluons couple
3358: only through an anomaly and anomalies can not occur
3359: within the two-dimensional kinematics that the kernels describe.
3360:
3361: \subhead{7.3 SU(2) Color and Reggeon Field Theory}
3362:
3363: If we consider all the diagrams discussed in previous Sections, generalized to
3364: include arbitrary numbers of gluons, and add both
3365: interactions amongst the gluons and between the quark-antiquark pair,
3366: we arrive at the set of transverse momentum
3367: diagrams shown in Fig.~48.
3368: \begin{center}
3369: \epsfxsize=4.5in
3370: \epsffile{ehes52.ps}
3371:
3372: Fig.~48 A Class Of Transverse Momentum Diagrams
3373: \end{center}
3374:
3375: If the gluons are anomalous and carry zero color, they will have no interaction
3376: with the quark-antiquark pair and the
3377: divergence (\ref{irscl}) will occur when the transverse momentum
3378: of all gluons is scaled to zero. As we discussed above, the anomaly pole
3379: should appear in the coefficient of this divergence, presumably, with
3380: the right kiematic structure to be interpreted as a pion pole.
3381: All other similar diagrams, in which either the
3382: color is non-zero or the gluons are not anomalous will be exponentiated to
3383: zero by interactions that iterate
3384: the divergence. However, since the cut-off has still to be removed
3385: and it is unclear how to handle the infra-red divergence, the result is still far
3386: from a sensible amplitude.
3387:
3388: To obtain a more sensible result, we have to use a more sophisticated
3389: treatment of the infra-red divergences. In particular, we
3390: initially take the SU(3) gauge symmetry of QCD to be partially broken to
3391: SU(2). We could motivate this by noting, first, that the structure of
3392: the anomaly diagrams is much simpler. Only odd numbers of anomalous gluons
3393: can carry color zero (because of the absence of
3394: the $d$ - tensor). An overall logarithmic infra-red
3395: divergence will still occur, as we have discussed in the previous sub-section,
3396: because of the unbroken SU(2) gauge symmetry. However, some gluons (that are
3397: massive and outside the SU(2) subgroup) will interact with the quark-antiquark
3398: pair. Also, we might hope to eliminate the divergence by
3399: averaging over the direction of the SU(2) subgroup within SU(3),
3400: as the transverse momentum cut-off is removed.
3401:
3402: With these last observations in mind, it is easy to appreciate why
3403: Reggeon Field Theory (RFT) should be applied
3404: to the problem. In formulating the study of the QCD pomeron using
3405: RFT we have argued\cite{arw93,arw84,arw98},
3406: that we should start from the reggeon diagrams (or, equivalently,
3407: transverse momentum diagrams) in which
3408: the gauge symmetry is completely broken. With a transverse momentum cut-off,
3409: the gauge symmetry can first be restored to SU(2) and
3410: the resulting reggeon diagrams can be described by Super-Critical Pomeron
3411: RFT - provided all infra-red divergences
3412: can be absorbed into a ``pomeron condensate''.
3413:
3414: For our present problem we anticipate applying RFT as follows. With
3415: SU(3) color broken to SU(2), we consider
3416: all diagrams of the form illustrated in Fig.~49. In these diagrams,
3417: anomalous gluons (within an SU(2) subgroup) accompany
3418: a quark-antiquark pair that is interacting with massive, reggeized,
3419: gluons. The massive gluons are outside the SU(2) subgroup and
3420: carry non-zero transverse momentum.
3421: \begin{center}
3422: $~~~~$\epsfxsize=3in
3423: \epsffile{ehes55.ps}
3424: \hspace{0.3in}
3425: \epsfxsize=1.8in
3426: \epsffile{ehes59.ps}
3427:
3428: Fig.~49 General Diagrams With a Divergence
3429: \end{center}
3430: This set should map completely on to Super-Critical RFT diagrams
3431: containing both pomerons (with the pomeron being
3432: a massive reggeized gluon plus
3433: anomalous SU(2) gluons) and a reggeized Goldstone boson pion. In this mapping,
3434: the physical significance of the logarithmic infra-red
3435: divergence would be clear. It would be identified as responsible for
3436: the appearance of a pomeron condensate.
3437:
3438: The restoration of SU(3) gauge symmetry would be described
3439: by the Critical Pomeron\cite{cri} interacting with a regge pole pion and (with
3440: the appropriate quark sector present\cite{arw93}-\cite{arw02})
3441: the transverse momentum cut-off can be
3442: removed as part of the critical phenomenon.
3443: Also, as part of the critical phenomenon,
3444: the SU(2) direction of the pomeron condensate would be randomized within SU(3)
3445: and disappear. In effect, the infra-red divergence, producing the condensate,
3446: would be eliminated by averaging over the
3447: direction of the SU(2) subgroup within SU(3).
3448:
3449: For hadron scattering it was important\cite{arw02} that the pomeron
3450: condensate could be related to an anomalous
3451: gluon component of the scattering states. This was necessary, firstly,
3452: because the $\gamma_5$ anomaly coupling of the pion to the pomeron
3453: is produced by a product of three orthogonal $\gamma$ matices. To obtain
3454: this product, it was essential to have anomalous gluon components
3455: in both the scattering pion and the exhanged pomeron. As we have seen,
3456: in electroweak
3457: scattering this requirement is absent because the vector
3458: mesons have elementary $\gamma_5$ couplings, which allow the anomalous gluons
3459: to appear in just the exchanged channel. However, the gluon components
3460: of the scattering states also seemed to be important for
3461: the higher-order pomeron interactions needed to obtain the Critical Pomeron.
3462: It may be, therefore, that
3463: RFT can only be consistently applied to the analysis
3464: of infra-red divergences if the scattering vector mesons are also ``hadronic''.
3465: That is, if they also have an anomalous gluon component,
3466: as they would have if they aquire their mass by absorbing Goldstone bosons
3467: resulting from QCD chiral symmetry breaking - with the quarks being color sextet
3468: quarks. The presence of the sextet quarks would produce\cite{arw93}-\cite{arw02}
3469: an infra-red fixed-point (in the massless quark theory) that would guarantee
3470: the infra-red scaling of gluon kernels producing (\ref{irscl}) and would also
3471: produce the ``quark saturation'' of QCD
3472: that we have argued is needed to obtain the Critical Pomeron with no transverse
3473: momentum cut-off. Perhaps,
3474: all these features are needed to obtain a self-consistent description
3475: of the regge limit for left-handed, massive, electroweak vector bosons.
3476:
3477: \newpage
3478:
3479: \mainhead{8. CONCLUSIONS}
3480:
3481: We have demonstrated that the triangle anomaly
3482: appears in the couplings of
3483: transverse momentum diagrams that describe the high-energy
3484: scattering of $W^{\pm,0}$ vector mesons. When the full amplitudes
3485: are directly evaluated, without any special cut-off procedure, the
3486: anomaly produces an enhancement, by a power of the energy, that threatens
3487: the unitarity of the theory.
3488:
3489: The most well-known consequence of a large momentum triangle anomaly is
3490: the famous anomalous Ward identity for axial currents\cite{awb}.
3491: Less emphasized is the feature that, in the presence of the anomaly, vector
3492: Ward identities are satisfied by a subtle cancelation between the contributions
3493: of large and small internal momentum regions. In the vector meson scattering
3494: we have discussed, an effective current component
3495: with an anomaly appears and
3496: it is the less emphasized feature that plays a crucial role. Even though
3497: there are no anomaly-related cancelations between large and small internal
3498: momenta in the finite momentum Ward identities, in a left-handed
3499: gauge theory, it appears that the regge limit enhances large transverse
3500: momentum regions such that there are cancelations of this kind in the
3501: transverse momentum (reggeon) Ward identities. There is
3502: then an ``anomaly problem'' in the sense that the regge limit result is
3503: very sensitive to the manipulation of ultra-violet and infra-red cut-offs,
3504: as we have described.
3505:
3506: In Appendix D, we raise the possibility that
3507: the occurrence of the anomaly
3508: enhancement phenomenon in the diagrams that we have discussed is related, via
3509: Ward identity contributions, to a more widespread phenomenon of large transverse
3510: momentum enhancement.
3511: If this is the case, then it is likely that the general transverse momentum
3512: diagram formalism will fail. Since there would then be no reggeon diagram
3513: formalism, $t$-channel unitarity
3514: is also likely to fail.
3515: The conclusion, which is really the main conclusion of this paper,
3516: is that to use the transverse momentum diagram formalism (and therefore
3517: to ensure $t$-channel unitarity) it is essential to initially
3518: employ a transverse momentum cut-off.
3519:
3520: In previous papers we have
3521: found that, for bound-state amplitudes in QCD, the occurrence
3522: of anomalies in multi-reggeon vertices (involving anomalous gluons)
3523: leads to an analagous sensitivity to
3524: infra-red and ultra-violet transverse momentum cut-offs. We have argued
3525: that an ultra-violet cut-off should be kept
3526: until after physical scattering amplitudes
3527: have been derived via an analysis of infra-red divergences. We anticipated
3528: that, without an initial
3529: cut-off, the ultra-violet anomaly effects would produce non-unitary power
3530: enhancement of the energy behavior of bound-state amplitudes.
3531: However, as we noted in the Introduction,
3532: accessing the anomalies in
3533: hadron amplitudes is very complicated and, therefore, it is much more difficult
3534: to appreciate their significance. In the electroweak amplitudes we
3535: have studied in this paper the anomaly appears immediately, because
3536: of the presence of elementary left-handed couplings. As a result the choice
3537: between bad, large transverse momentum based, high-energy behavior and
3538: infra-red anomaly domination producing ``non-perturbative'' dynamics, is also
3539: immediately clear.
3540:
3541: Potential non-unitary properties
3542: of electroweak high-energy scattering amplitudes may not be of great concern if,
3543: as is currently believed by many physicists, the gravitational interaction
3544: intervenes long before the relevant energies are reached. From this perspective,
3545: our study of electroweak amplitudes can be viewed as simply a technical exercise
3546: in which left-handed vector mesons are used to study how, with the cut-off
3547: manipulation we have described, the formation
3548: of QCD bound states, including confinement and chiral symmetry breaking,
3549: can take place via regge limit infra-red anomaly effects.
3550: Nevertheless, it seems hard to avoid the conclusion that
3551: if confinement and chiral symmetry breaking do
3552: not take place in this manner, then (assuming that it does not cancel)
3553: the power enhancement
3554: of quark-antiquark exchange by the ultra-violet anomaly
3555: will dominate any electroweak symmetry
3556: breaking mechanism that is perturbatively based.
3557:
3558: Our point of view is that the unitarity of the electroweak
3559: part of the Standard Model is a deep constraint. Indeed,
3560: it could be that obtaining
3561: consistent high-energy scattering amplitudes for massive vector mesons, with
3562: left-handed couplings to quarks, may actually
3563: require QCD confinement and chiral symmetry breaking to take place via the anomaly,
3564: and may even, perhaps, require
3565: that the chiral symmetry breaking (of higher color quarks) is responsible for
3566: electroweak symmetry breaking.
3567:
3568: We were led to the present investigation as an outcome
3569: of our study of the QCD pomeron. For a long period of time
3570: we understood the crucial
3571: role of the anomaly in producing unitary high-energy amplitudes within QCD,
3572: but were unable to find a simple starting point from which to begin construction
3573: of such amplitudes. Then, in our most recent paper\cite{arw02}, we showed that
3574: wee gluon properties of the pion, obtained from the anomaly, provide
3575: such a starting-point, at least in part. At the same time we realized that
3576: such properties should appear if the pion is extracted from
3577: the wee parton structure of an electroweak vector meson. This led us to
3578: the, a priori much simpler, problem of how an exchanged
3579: pion appears within
3580: the scattering of vector mesons. We now believe that the results of this paper
3581: will lead to an understanding of the pion which will,
3582: eventually, provide a simple starting point
3583: for the construction of QCD high-energy amplitudes.
3584:
3585: \subhead{Acknowledgement}
3586:
3587: I am grateful to Jochen Bartels for useful criticism of a preliminary version
3588: of this paper. I am also grateful to Bill Bardeen for valuable
3589: discussion of electroweak Ward identity cancelations.
3590:
3591:
3592: \newpage
3593:
3594: \renewcommand{\theequation}{A.\arabic{equation}}
3595: \setcounter{equation}{0}
3596: \vskip 1cm \noindent
3597: \noindent {\large\bf Appendix A. ~ Complex Notation for
3598: Transverse $\gamma$ - Matrices}
3599: \vskip 3mm \noindent
3600:
3601: To discuss high-energy vector meson
3602: scattering amplitudes involving (massless) fermion
3603: exchange, it is convenient to use a complex number notation\cite{kms}
3604: for both transverse momenta and $\gamma$ matrices. In this formalism,
3605: the consequences of
3606: chirality conservation and a left-handed gauge interaction are particularly
3607: apparent.
3608:
3609: In addition to using conventional
3610: light-cone momenta $k_{\pm} = (k_0 \pm k_1)/\sqrt{2}$ ,
3611: we write
3612: $$
3613: \kt~ =~ k_2~ +~i~k_3~, ~~~~~ \kt^*~ =~ k_2~ -~ik_3
3614: \auto\label{cta}
3615: $$
3616: to describe
3617: transverse momenta. We then have
3618: $$
3619: k_{\perp}^2 ~=~\kt\kt^*
3620: \auto\label{cta0}
3621: $$
3622: and
3623: $$
3624: 2 k_{\perp}\cdot q_{\perp}~=~\kt\qt^* ~+ ~\kt^*\qt
3625: \auto\label{cta01}
3626: $$
3627: We can also write
3628: $$
3629: \eqalign{\kt\qt^*~=~ (\kt^*\qt)^*~&= ~k_2q_2 + k_2q_3 ~+~i( k_2q_3 -k_3q_2)\cr
3630: &=~k_{\perp}\cdot q_{\perp}~+~i~k_{\perp}\times q_{\perp}}
3631: \auto\label{cta1}
3632: $$
3633: where
3634: $$
3635: k_{\perp}\times q_{\perp}~=~|k_{\perp}|~| q_{\perp}|~\sin \theta
3636: \auto\label{theta}
3637: $$
3638: with $\theta$ the angle between the two vectors.
3639:
3640: To describe transverse $\gamma$ - matrices, we similarly write
3641: $$
3642: \gam~=~(~\gamma_2~+~i~\gamma_3~)/\sqrt{2},
3643: ~~~~\gam^*~=~(~\gamma_2~-~i~\gamma_3~)/\sqrt{2}
3644: \auto\label{cga}
3645: $$
3646: We then have
3647: $$
3648: (\gam)^2~=~ (\gam^*)^2 ~=~0~, ~~~~
3649: \gam~\gam^*~+~ \gam^*~\gam ~= ~2
3650: \auto\label{cga1}
3651: $$
3652: and we can write
3653: $$
3654: 2 \st{k}_{\perp}~=~\kt\gam^* ~+~ \kt^*\gam
3655: \auto\label{cga2}
3656: $$
3657:
3658: In the regge limit the transverse part of
3659: an exchanged fermion propagator dominates, i.e. for a massless fermion
3660: $$
3661: \frac{\st{k}}{ k^2} ~~\to ~~ \frac{1}{ 2} \biggl(\frac{\gam^*}{ \kt^*}
3662: + \frac{\gam} {\kt} \biggr)
3663: \auto\label{frp}
3664: $$
3665: where the two terms represent the two different chiralities.
3666: For example, the transverse momentum integration of
3667: a quark-antiquark state with transverse momentum $q_{\perp}$ and
3668: equal chiralities (opposite sign helicities) takes the form
3669: $$
3670: \int~ d^2\kt_1 d^2\kt_2 ~\delta^2(\qt-\kt_1-\kt_2)
3671: \biggl( \frac{\gam}{ \kt_1 }\otimes \frac{\gam^* }{ \kt_2^* }
3672: ~~+~~ \frac{\gam^*}{ \kt_1^* }\otimes \frac {\gam}{ \kt_2 }\biggr)
3673: \auto\label{f2tr}
3674: $$
3675: where the $\otimes$ sign indicates that the two $\gamma$ - matrices are
3676: separately associated with the two fermion lines.
3677: The contribution of a two fermion state with opposite chiralities is clearly
3678: analagous. However, the distinct
3679: combinations of same sign and opposite sign chiralities
3680: are exchanged and interact separately\cite{kms,rk}.
3681: As we elaborate on briefly in Appendix B,
3682: the very different properties of the interaction of
3683: same sign and opposite sign chirality
3684: exchanges is of fundamental importance.
3685:
3686: If we also define
3687: $$
3688: {\Pit}_+ ~=~ -\frac{1}{2}\gam\gam^*~,~
3689: ~{\Pit}_-~ =~ -\frac{1}{2}\gam^*\gam~,~
3690: ~\Pi_+ ~= ~\frac{1}{2}\gamma_-\gamma_+~,~
3691: ~\Pi_- ~=~ \frac{1}{2}\gamma_+\gamma_-
3692: \auto\label{prj1}
3693: $$
3694: then we can write
3695: $$
3696: \gamma_5 = (\Pi_+ - \Pi_-) ({\Pit}_+ - {\Pit}_-)
3697: \auto\label{prj2}
3698: $$
3699: Spinors in the subspaces $\Pi_-{\Pit}_+$ and
3700: $\Pi_- {\Pit}_-$ (or $\Pi_+{\Pit}_+$ and
3701: $\Pi_+ {\Pit}_-$) carry opposite chirality, as
3702: is evident from the following relations
3703: $$
3704: \eqalign{\gamma_-\gam[1-\gamma_5]\gam^* ~&=~\gamma_-[1+\gamma_5]\gam \gam^*
3705: ~=~ \gamma_-[1-
3706: \Pi_- {\Pit}_+] \gam \gam^* \cr
3707: &=~ \gamma_- (1 + \frac{1}{2}\gam \gam^* )\gam \gam^*
3708: ~=~ 2 \gamma_- \gam \gam^* \cr
3709: ~& \cr
3710: \gamma_-\gam^*[1-\gamma_5]\gam ~&=~ \gamma_-[1+\gamma_5]\gam^*\gam
3711: ~=~ \gamma_-[1+
3712: \Pi_- {\Pit}_-]\gam^* \gam \cr
3713: &=~ \gamma_- (1 - \frac{1}{2}\gam^* \gam )\gam^* \gam
3714: ~=~ 0
3715: }
3716: \auto\label{prj3}
3717: $$
3718: Similarly, we can show that
3719: $$
3720: \gamma_+\gam^*(1-\gamma_5)\gam ~= ~0 ~,~~~~~
3721: \gamma_+\gam(1- \gamma_5)\gam^* ~=
3722: ~2 \gamma_+ \gam\gam^*
3723: \auto\label{prj4}
3724: $$
3725: $$
3726: \gamma_-\gam(1+\gamma_5)\gam^* ~= ~0 ~,~~~~~
3727: \gamma_-\gam^*(1+\gamma_5)\gam ~= ~2 \gamma_-\gam^* \gam
3728: \auto\label{prj5}
3729: $$
3730: $$
3731: \gamma_+\gam^*(1+\gamma_5)\gam ~= ~0 ~,~~~~~
3732: \gamma_+\gam(1+\gamma_5)\gam^* ~= ~2 \gamma_+ \gam\gam^*
3733: \auto\label{prj6}
3734: $$
3735:
3736: For a vector particle, with momentum along the 1-axis,
3737: the polarization vectors for states with helicity $\lambda = \pm 1$ are
3738: $$
3739: \eqalign{\epsilon^{\mu}(\lambda = +1)~&=~ - \frac{1}{\sqrt{2}}(0,0,1,i) \cr
3740: \epsilon^{\mu}(\lambda = -1)~&=~ \frac{1}{\sqrt{2}}(0,0,1,-i) }
3741: \auto\label{pov}
3742: $$
3743: A vector boson with helicity $\lambda=-1$
3744: can make a transition to a left-handed intermediate state quark via the
3745: emission of an antiquark. To calculate the scattering of a vector boson
3746: with helicity $\lambda=-1$ we introduce an initial coupling of
3747: $\bar{\psi}\gam^*(1-\gamma_5)\psi$ and a final state coupling of
3748: $\bar{\psi}\gam^(1-\gamma_5) \psi$. Utilising the above relations
3749: we find that, as illustrated in Fig.~A1,
3750: there is only one non-zero coupling to potential quark-antiquark
3751: transverse momentum states that could be exchanged.
3752: \begin{center}
3753: \epsfxsize=4.7in
3754: \epsffile{ehes3020.ps}
3755:
3756: Fig.~A1 Couplings to Quark-Antiquark Transverse Momentum States
3757: \end{center}
3758: As a consequence,
3759: if we consider the scattering of opposite helicity states there is only one
3760: possible lowest-order diagram, which is that shown in Fig.~A2(a).
3761: \begin{center}
3762: \epsfxsize=5.5in
3763: \epsffile{ehes040.ps}
3764:
3765: Fig.~A2 Lowest-Order Diagrams
3766: \end{center}
3767: The initial $\gam^*(1-\gamma_5)$ vertex on the right-hand side of Fig.~A1(a)
3768: represents the coupling of a vector boson with the same polarization, but
3769: opposite helicity (since it has opposite momentum along the $1$-axis) to that
3770: of the left-hand side vector boson.
3771:
3772: A simple way to see that the diagram of Fig.~1(a) contributes to
3773: opposite helicity scattering is to note that, because of the direction of the
3774: quark arrow, the intermediate state consists of a left-handed quark, which must
3775: be produced by a negative helicity vector boson, and a right-handed antiquark,
3776: which must be produced by a positive helicity vector boson. The direction of the
3777: arrow is fixed by choosing the left-hand
3778: vector meson to be the one with negative helicity.
3779: The diagram of Fig.~1(a) contributes to the $A_{-+}$ helicity amplitude while the
3780: diagram with the arrow reversed contributes to the $A_{+-}$ helicity amplitude.
3781: By similar reasoning, the diagram of Fig.~2 contributes to the
3782: $A_{-+}$ helicity amplitude.
3783:
3784: The diagram shown in Fig.~A2(b) is the only possibility for
3785: the scattering of states with equal, negative, helicities. In the cross-channel,
3786: in which the incoming and outgoing, right hand, vector mesons are interchanged,
3787: the diagram untwists to become the diagram of Fig.~A1(a).
3788: Fig.~A2(b) contributes to the $A_{--}$ helicity
3789: amplitude, while the corresponding diagram with the quark arrow reversed
3790: contributes to the $A_{++}$ helicity amplitude. For the amplitudes
3791: with $\pi^0$ quantum numbers in the $t$ - channel, that we discuss
3792: in this paper, it follows from $CPT$ invariance that
3793: $$
3794: A_{++}~=~A_{--}~~~~~~\hbox{and} ~~~~~~A_{-+}~=~A_{+-}
3795: \auto\label{eha}
3796: $$
3797: Note that, in all diagrams, only same sign chirality states are exchanged.
3798:
3799: \newpage
3800:
3801: \renewcommand{\theequation}{B.\arabic{equation}}
3802: \setcounter{equation}{0}
3803: \vskip 1cm \noindent
3804: \noindent {\large\bf Appendix B. ~ Review of Leading and Non-Leading Logs}
3805: \vskip 3mm \noindent
3806:
3807: As far as we know, the diagrams we discuss in this paper
3808: have not been discussed in
3809: detail in the literature. However, if we were to make the (wrong)
3810: assumption that the
3811: left-handed coupling does not affect the extraction of high-energy
3812: logarithms, or (more simply) if we impose a transverse momentum cut-off,
3813: there are a number of well-known results that would carry over,
3814: almost directly, into our problem. Just to put the discussion of this paper in
3815: context, we give here a very brief, non-technical, overview\cite{rk}
3816: of these results.
3817:
3818: All the results concern the extraction of leading and non-leading logarithms.
3819: If we organize the
3820: quark-antiquark exchange diagrams into distinct series, as
3821: illustrated in Fig.~B1, depending on the power of $\alpha_s$ (the QCD coupling)
3822: involved, then typical diagrams giving such logarithms are
3823: \newline \parbox{0.5in}{$~$}
3824: \parbox{4.6in}{\begin{center}
3825: \epsfxsize=4.5in
3826: \epsffile{ehes020.ps}
3827: \end{center}}
3828: \parbox{0.6in}{\begin{center}
3829: $$
3830: O(\alpha_s^0)
3831: $$
3832:
3833: $$
3834: O(\alpha_s)
3835: $$
3836:
3837: $$
3838: O(\alpha_s^2)
3839: $$
3840:
3841: \end{center}}
3842: \begin{center}
3843: Fig.~B1 Quark-Antiquark Exchange Diagrams Organized According to Powers of
3844: $\alpha_s$.
3845:
3846: \end{center}
3847: The first series contains purely electroweak diagrams that have
3848: a logarithmic expansion in $\alpha_w$ (the ``electroweak'' coupling).
3849: The second series contains $O(\alpha_s)$
3850: corrections to the first series, the third series contains $O(\alpha_s^2)$
3851: corrections to the first series, etc.
3852:
3853: All diagrams of the form shown in
3854: Fig.~B1 would be expected to give high-energy amplitudes of the form
3855: $$
3856: A(S,0)~\centerunder{$\sim$}{\raisebox{-5mm}{$S\to \infty $}}
3857: ~ \centerunder{\hbox{\large $\Sigma$}}{\raisebox{-3mm}
3858: {${\scriptstyle n,m,r} $}}
3859: ~~ a_{nmr}~\alpha_{w}^n~\alpha_s^m~ [lnS]^r
3860: \auto\label{ser}
3861: $$
3862: To make our discussion straightforward we can suppose that
3863: we, initially, introduce a transverse momentum cut-off so that we can ignore
3864: ultra-violet transverse momentum divergences - including
3865: both the anomaly power divergences that we discuss in this paper,
3866: and the logarithmic divergences that we discuss below. As a result,
3867: all the coefficients $a_{nmr}$ can be represented as (sums of) transverse momentum
3868: diagrams of the form illustrated in Fig.~B2.
3869:
3870: \noindent \parbox{0.6in}{$~$}\parbox{4.4in}{
3871: \epsfxsize=4.3in
3872: \epsffile{ehes02201.ps}
3873: }
3874: \parbox{0.9in}{\begin{center}
3875: $$
3876: O(\alpha_s^0)
3877: $$
3878: \end{center}}
3879: \newline \parbox{0.6in}{$~$}\parbox{4.4in}{
3880: \epsfxsize=4.3in
3881: \epsffile{ehes02202.ps}}
3882: \parbox{0.9in}{\begin{center}
3883: $$
3884: O(\alpha_s)
3885: $$
3886: \end{center}
3887: }
3888: \newline \parbox{0.6in}{$~$}\parbox{4.4in}{
3889: \epsfxsize=4.3in
3890: \epsffile{ehes02203.ps}}
3891: \parbox{0.9in}{\begin{center}
3892: $$
3893: O(\alpha_s^2)
3894: $$
3895: \end{center}}
3896: \begin{center}
3897: Fig.~B2 Transverse Momentum Diagrams Originating from the Diagrams of Fig.~B2
3898: \end{center}
3899:
3900: With the transverse momentum cut-off in place, the
3901: first two diagrams in Fig.~B1 give a leading log amplitude which contains
3902: the first diagram in Fig.~B2 multiplied by $ln S$ and a
3903: next-to-leading log amplitude which contains
3904: the same transverse momentum diagram but with no factor of $ln S$.
3905: The third diagram in Fig.~B1 gives a leading log amplitude which contains
3906: the second diagram in Fig.~B2 multiplied by $ln^2 S$
3907: and a next-to-leading log amplitude which contains
3908: the first diagram in Fig.~B2 multiplied by $ln S$ and so on. In general,
3909: the external couplings and the internal vertices in the transverse
3910: momentum diagrams aquire more and more structure (involving loop
3911: integrals) as first leading logs,
3912: then next-to-leading logs, next-to-next-to-leading logs, etc.,
3913: are included in the sum (\ref{ser}).
3914:
3915: The diagram of Fig.~2, appearing in Section 3,
3916: is the last diagram shown explicitly in the second row of Fig.~B1.
3917: It is first-order in $\alpha_s$ and, conventionally, as noted in Section 3,
3918: we would expect that it's
3919: leading-log contribution would contain the last $O(\alpha_s)$
3920: transverse momentum diagram
3921: shown explicitly in Fig.~B2 - with simple vertices. This diagram
3922: being obtained by placing all vertical lines on-shell, as in Fig.~5 using
3923: longitudinal momentum integrations. At the next-to-leading log
3924: level the second $O(\alpha_s)$
3925: transverse momentum diagram should be generated and the first $O(\alpha_s)$
3926: transverse momentum diagram, which is the
3927: diagram that appears in Fig.~1, should be generated by Fig.~2
3928: at the next-to-next-to-leading log level.
3929:
3930: The transverse momentum diagram of Fig.~17 is the first $O(\alpha_s^2)$
3931: diagram appearing in Fig.~B2 and would be generated, at leading log, by
3932: the first $O(\alpha_s^2)$ diagram in Fig.~B2. The diagram of Fig.~20 is the
3933: second $O(\alpha_s^2)$ diagram appearing explicitly in Fig.~B1. The
3934: anticipated leading log result for this diagram would be $ln^5 S$
3935: multiplied by the second $O(\alpha_s^2)$ transverse momentum
3936: diagram appearing explicitly in Fig.~B2 and
3937: the first $O(\alpha_s^2)$ diagram in Fig.~B2 would be generated at
3938: the next-to-next-to-leading log level, i.e. multiplied by a factor of $ln^3 S$.
3939:
3940: In the leading and non-leading-log studies of
3941: pure vector gauge theories\cite{fl} there is no problem with ultra-violet
3942: divergences, either in transverse momentum, or more generally.
3943: Only the normal
3944: (ultra-violet) divergences associated with renormalization have appeared
3945: in the non-leading-log vertices. The Ward
3946: identities of the gauge theory produce cancelations that lead always to
3947: convergent transverse momentum integrals, with the accompanying
3948: logarithms just those predicted by regge theory. (Even though, as is very
3949: well known by now, individual feynman
3950: diagrams produce transverse momentum divergences that, at first sight,
3951: produce additional logarithms beyond those anticipated by regge theory.)
3952: Equivalently, the complete sum of logarithms and transverse momentum
3953: diagrams can be rearranged\cite{bs} into sub-series represented by
3954: reggeon diagrams.
3955:
3956: When fermions are involved there is, as we already noted in Section 3,
3957: the extra subtlety of the logarithmic divergence of fermion transverse momentum
3958: integrals\cite{rk,kms}. Therefore, if the transverse momentum cut-off is removed,
3959: extra powers of $ln S$ will be generated and
3960: the series (\ref{ser}) must be rearranged appropriately.
3961: However, fermion reggeization is not affected
3962: by the divergences (since the relevant transverse momentum integrals involve
3963: combinations of fermions and gluons). Consequently, as we noted in Section 3,
3964: when the transverse momentum diagrams are organized into reggeon diagrams,
3965: the presence of the reggeon propagator reduces the divergence
3966: from log to log[log] form.
3967: Also, the reggeon kernel for opposite sign helicities gives convergent integrals.
3968: (As we noted in the previous Appendix, the distinct
3969: combinations of opposite and same sign helicities
3970: are exchanged and interact separately.) Only
3971: the kernel for opposite sign helicities (same sign chiralities)
3972: produces logarithmic divergences at large transverse momentum.
3973:
3974: In the diagrams discussed in this paper, the quark-antiquark states we consider
3975: are same sign chirality states. However, the anomaly enhancement overwhelms
3976: the logarithmic divergence that would otherwise result.
3977: We believe this is important, physically. If we start with a transverse momentum
3978: cut-off both the anomaly power divergence and the logarithmic divergence
3979: will be absent.
3980: When the confinement and chiral symmetry breaking described in Section 5 is
3981: implemented via the extraction of infra-red divergences,
3982: it may be (and the results of \cite{arw02} directly suggest this) that
3983: only (transverse momentum) convergent same sign helicity exchanges are involved
3984: in forming bound states. Since ``double logs''
3985: are, a priori, in conflict with regge theory,
3986: this is probably necessary for the bound states to be described by regge theory.
3987:
3988: \newpage
3989:
3990:
3991: \renewcommand{\theequation}{C.\arabic{equation}}
3992: \setcounter{equation}{0}
3993: \vskip 1cm \noindent
3994: \noindent {\large\bf Appendix C. ~ The Anomaly and Vector Ward Identities}
3995: \vskip 3mm \noindent
3996:
3997: To understand the special nature of Ward identities in the presence of
3998: the anomaly, it is helpful to recall some well-known
3999: properties of the one loop contribution, shown in Fig.~C1, of massless fermions to
4000: an axial-vector/two-vector three current vertex $T_{\mu \alpha \beta}(k_1,k_2)$.
4001: \begin{center}
4002: \leavevmode
4003: \epsfxsize=2in
4004: \epsffile{ehes54.ps}
4005:
4006: Fig.~C1 The Fermion Loop Contribution to $T_{\mu \alpha \beta}(k_1,k_2)$
4007:
4008: \end{center}
4009: After decomposition into invariant amplitudes
4010: $$
4011: \eqalign{T_{\mu \alpha \beta}(k_1,k_2) ~&= ~ A_1~
4012: {\hbox{\large $\epsilon$}}_{\sigma\alpha\beta\mu}~ k_1^{\sigma} ~+~ A_2~
4013: {\hbox{\large $\epsilon$}}_{\sigma\alpha\beta\mu} ~k_2^{\sigma}
4014: ~+~A_3~
4015: {\hbox{\large $\epsilon$}}_{\delta \sigma\alpha\mu}~
4016: k_{1\beta}k_1^{\delta} k_2^{\sigma} \cr
4017: ~~~& +~A_4~ {\hbox{\large $\epsilon$}}_{\delta \sigma\alpha\mu}
4018: ~ k_{2\beta}k_1^{\delta}
4019: k_2^{\sigma}~+~A_5~ {\hbox{\large $\epsilon$}}_{\delta \sigma\beta\mu}
4020: ~k_{1\alpha}k_1^{\delta}
4021: k_2^{\sigma}~+~A_6~ {\hbox{\large $\epsilon$}}_{\delta \sigma\beta\mu}
4022: ~ k_{2\alpha}k_1^{\delta}
4023: k_2^{\sigma} }
4024: \auto\label{inde}
4025: $$
4026: the vector Ward identities
4027: $$
4028: k_1^{\alpha}~\Gamma_{\mu \alpha \beta}~=0 ~,~~~
4029: k_2^{\beta}~\Gamma_{\mu \alpha \beta}~=0
4030: \auto \label{vwi}
4031: $$
4032: require
4033: $$
4034: A_2~=~k_1^2~A_5 ~+~k_1\cdot k_2 ~A_6
4035: \auto\label{vwi1}
4036: $$
4037: $$
4038: A_1~=~k_2^2~A_4 ~+~k_1\cdot k_2 ~A_3
4039: \auto\label{vwi2}
4040: $$
4041:
4042: The large momentum region (with appropriate regularization)
4043: gives an ``anomaly'' contribution to $A_1$ and $A_2$ of the form
4044: $$
4045: T_{\mu \alpha \beta}(k_1,k_2) ~= ~ \frac{1}{ 4 \pi^2}~
4046: {\hbox{\large $\epsilon$}}_{\sigma\alpha\beta\mu}~ k_1^{\sigma} ~+~
4047: \frac{1}{ 4 \pi^2}~
4048: {\hbox{\large $\epsilon$}}_{\sigma\alpha\beta\mu} ~k_2^{\sigma} ~~+~~\cdots
4049: \auto\label{uvco}
4050: $$
4051: leading to the well-known ``anomalous'' divergence equation
4052: $$
4053: (k_1 + k_2)^{\mu}~T_{\mu \alpha \beta}~=~
4054: \frac{1}{ 2 {\pi}^2 }~{\hbox{\Large $\epsilon$}}_{\delta\sigma\alpha\beta}
4055: ~k_1^{\delta} k_2^{\sigma}
4056: \auto\label{awi}
4057: $$
4058: For the vector Ward
4059: identities to hold in the presence of (\ref{uvco}), there must be
4060: related, infra-red singular, contributions to the other $A_i$.
4061: For example, when $k_1^2=0$, (\ref{vwi1}) becomes
4062: $$
4063: A_2~=~k_1 \cdot k_2~A_6 ~=~\frac{q^2 - k_2^2}{2} ~A_6
4064: \auto\label{vwi11}
4065: $$
4066: implying that there must be a pole in $A_6$, arising from the region
4067: of small internal momentum. In appropriate circumstances, this pole can be
4068: interpreted as a Goldstone boson pole, signaling chiral symmetry breaking.
4069:
4070: If we consider $k_1 \to 0$, and assume that all the
4071: $A_i$ are sufficiently non-singular, then (\ref{inde}) gives
4072: $$
4073: T_{\mu \alpha \beta}(k_1,k_2) ~\centerunder{$\to$}{\raisebox{-5mm}{$k_1 \to 0$}}
4074: ~~ ~ A_2~
4075: {\hbox{\large $\epsilon$}}_{\sigma\alpha\beta\mu} ~k_2^{\sigma}
4076: \auto\label{k10}
4077: $$
4078: which, if we keep only the ultra-violet anomaly term (\ref{uvco}), gives
4079: $$
4080: T_{\mu \alpha \beta}(k_1,k_2) ~
4081: ~\centerunder{$\to$}{\raisebox{-5mm}{$k_1 \to 0$}}
4082: ~ ~~\frac{1}{4\pi^2}~
4083: {\hbox{\large $\epsilon$}}_{\sigma\alpha\beta\mu} ~k_2^{\sigma}~~\neq~0
4084: \auto\label{k101}
4085: $$
4086: Alternatively, if we use (\ref{vwi1}), together with (\ref{k10}), we obtain
4087: $$
4088: T_{\mu \alpha \beta}(k_1,k_2) ~
4089: ~\centerunder{$\to$}{\raisebox{-5mm}{$k_1 \to 0$}}
4090: ~~~ (k_1^2A_5 + k_1\cdot k_2 A_6)~
4091: {\hbox{\large $\epsilon$}}_{\sigma\alpha\beta\mu} ~k_2^{\sigma} ~~\to 0
4092: \auto\label{k102}
4093: $$
4094: For consistency, again, there must be infra-red singular contributions
4095: to $T_{\mu \alpha \beta}(k_1,k_2)$ that cancel the ultra-violet
4096: anomaly contribution (\ref{uvco}) and produce the ``Ward identity zero''
4097: (\ref{k102}).
4098:
4099: From our point of view, therefore, the
4100: presence of the ultra-violet anomaly (\ref{uvco})
4101: has two consequences. The first is the
4102: anomalous Ward identity (\ref{awi}). The second is that the vector
4103: Ward identities require a cancelation between separate contributions
4104: (with different kinematic structure) from large and
4105: small internal momentum regions. As a consequence,
4106: if an explicit ultra-violet cut-off is introduced, (\ref{uvco}) will be
4107: modified and the vector Ward identities will no longer hold. The contribution,
4108: to the vector current divergence, of the pole term in
4109: $A_6$ will survive, however, since it is generated in the infra-red
4110: region\cite{arw02}.
4111:
4112: \newpage
4113:
4114:
4115: \renewcommand{\theequation}{D.\arabic{equation}}
4116: \setcounter{equation}{0}
4117: \vskip 1cm \noindent
4118: \noindent {\large\bf Appendix D. ~ Electroweak Ward Identity Cancelations}
4119: \vskip 3mm \noindent
4120:
4121: In this Appendix we consider whether Ward identity cancelations can
4122: remove the longitudinal polarization contributions of vector mesons
4123: that produce the anomaly enhanced high-energy behavior in the diagrams we
4124: have discussed.
4125: As described in sub-section {\bf 3.13},
4126: we are interested in Ward identity implications when we add all the diagrams
4127: that effectively replace
4128: the tree diagram that forms the lower part of Fig.~1 by another tree diagram.
4129: As in Section 3, it will be sufficient for our purposes to consider only
4130: the diagrams of an abelian theory.
4131:
4132: We focus on the same region of phase space as in Section 3, which will be
4133: the basis for all approximations we make. If we ignore (1-$~\gamma_5~$) factors
4134: (which are irrelevant for the present discussion), the lower part of Fig.~1
4135: gives the amplitude shown in Fig.~D1
4136: \begin{center}
4137: \epsfxsize=5in
4138: \epsffile{ehes90.ps}
4139:
4140: Fig.~D1 The Tree Diagram Obtained From Fig.~1.
4141: \end{center}
4142: We consider, first, the addition of tree
4143: diagrams in which the internal left-side vector boson line
4144: is attached at all possible points.
4145: We begin with the diagrams obtained by moving this line to the right.
4146:
4147: The subdiagram forming the left part of Fig.~D1
4148: can be split into two pieces as illustrated in Fig.~D2.
4149: \begin{center}
4150: \epsfxsize=4.5in
4151: \epsffile{ehes91.ps}
4152:
4153: Fig.~D2 Splitting Fig.~D1
4154: \end{center}
4155: where the dashed line indicates that additional momentum flows in to a vertex
4156: without changing the algebraic structure.
4157: It is the Ward identity cancelation for the second piece of Fig.~D2
4158: that involves moving the vector boson line to the right.
4159: Note, first, that if we combine this term
4160: with the right-side of the full diagram we obtain the amplitude shown in Fig.~D3
4161: \begin{center}
4162: \epsfxsize=4in
4163: \epsffile{ehes93.ps}
4164:
4165: Fig.~D3 The Amplitude Obtained From the Second Term of Fig.~D2
4166: \end{center}
4167: and so the relevant piece of Fig.~D1 is retained.
4168:
4169: The first tree
4170: diagram obtained by moving the left side internal vector meson to the
4171: right is the second diagram appearing in Fig.~16,
4172: which is the lower part of the feynman diagram appearing in Fig.~17.
4173: We consider the right part of this tree diagram and divide it into
4174: two pieces as in Fig.~D4.
4175: \begin{center}
4176: \epsfxsize=5in
4177: \epsffile{ehes94.ps}
4178:
4179: Fig.~D4 Splitting of Another Tree Diagram
4180: \end{center}
4181: The first piece gives an amplitude that directly
4182: cancels the amplitude of Fig.~D3. Therefore, it would appear that
4183: the amplitude involved in the anomaly enhancement is immediately
4184: eliminated. However, there are further cancelations that remain to be
4185: discussed.
4186:
4187: The second piece of Fig.~D4 has to be combined with the
4188: contribution of the tree diagram shown in Fig.~D5.
4189: \begin{center}
4190: \epsfxsize=2.4in
4191: \epsffile{ehes950.ps}
4192:
4193: Fig.~D5 Another Tree Diagram
4194: \end{center}
4195: Making the usual separation (into two pieces) of the right side of the tree diagram
4196: appearing in Fig.~D5 and removing the piece that
4197: cancels with the second piece of Fig.~D4 leaves the piece shown in Fig.~D6.
4198: \begin{center}
4199: \epsfxsize=5.8in
4200: \epsffile{ehes97.ps}
4201:
4202: Fig.~D6 Part of the Tree Diagram Appearing in Fig.~D5
4203: \end{center}
4204: This piece would be zero if the
4205: vertical antiquark were exactly, and not just approximately, on-shell.
4206:
4207: Since the amplitude of Fig.~D6 goes to zero as $P_- \to \infty$,
4208: with all internal momenta fixed, it is, superficially, a non-leading
4209: asymptotic contribution. However, it has worse large transverse momentum
4210: behavior than the original amplitude of Fig.~D2.
4211: In effect, we have replaced a leading asymptotic
4212: contribution that has manageable internal momentum behavior with a
4213: superficially non-leading contribution with bad internal momentum behavior.
4214: At this point, this substitution does not actually lead to any important
4215: effects, although this will not be the case for an analagous substitution
4216: that we make later. We obtain the maximal
4217: contribution from the amplitude of Fig.~D6 if we
4218: use the mass-shell condition $P_-k_-' \sim (p_{\perp}'' -k_{\perp}')^2$
4219: and combine the resulting amplitude with the second term in
4220: Fig.~D2. This gives
4221: $$
4222: \bigl(~\frac{{k'_-}^2}{k_-}~\bigr)~~\frac{1}{(\pt'+\kt')
4223: (p_{\perp}'' - k'_{\perp})}
4224: \auto\label{d1}
4225: $$
4226: This amplitude
4227: does not have the growth at large $k_{\perp}$ that the amplitude of
4228: Fig.~D1 has and so can be neglected.
4229:
4230: We now consider the contribution of the first term in Fig.~D2. This has
4231: to be combined with tree diagrams obtained by moving the left side internal vector
4232: meson line to the left.
4233: There is
4234: only one such diagram, which is shown in Fig.~D7.
4235: \begin{center}
4236: \epsfxsize=1.8in
4237: \epsffile{ehes100.ps}
4238:
4239: Fig.~D7 Another Tree Diagram
4240: \end{center}
4241: Normally this contribution to the high-energy
4242: limit would be ignored because an off-shell propagator
4243: carries the large $P_+$ momentum. However,
4244: if we split this diagram into two pieces as in Fig.~D8,
4245: \begin{center}
4246: \epsfxsize=4.5in
4247: \epsffile{ehes92.ps}
4248:
4249: Fig.~D8 Splitting The Diagram of Fig.~D7
4250: \end{center}
4251: the first piece cancels with the first piece of Fig.~D2.
4252:
4253: If the vertical
4254: quark line were on shell so that the full numerator, and not just the asymptotic
4255: $\gamma_-$ piece, were present, the second piece of Fig.~D8 would be zero.
4256: In fact, if we use the mass-shell condition
4257: $P_+k_-' \sim (p'_{\perp} + k'_{\perp})^2$ we obtain
4258: $$
4259: \frac{(\pt'+\kt')^*}{P_+k_-}~\sim ~\frac{k'_-}{k_-}~
4260: \frac{(\pt'+\kt')^*}{(p'_{\perp} + k'_{\perp})^2}~= ~
4261: \frac{k'_-}{k_-}~
4262: \frac{1}{(\pt' + \kt')}
4263: \auto\label{d2}
4264: $$
4265: Since both $k_-'$ and $k_-$ are finite in the momentum region
4266: we are considering, (\ref{d2})
4267: is of the same form as the second term in Fig.~D2. In this case, therefore,
4268: a superficially non-leading
4269: asymptotic contribution, with bad large transverse momentum
4270: behavior, gives a contribution that can not be nelected.
4271:
4272: We now consider the additional tree
4273: diagrams that would be involved
4274: in a Ward identity for the right side internal
4275: vector meson line. From the above discussion it follows that, after we have added
4276: all such diagrams and carried out the analogous cancelations to those above,
4277: there will be one surviving
4278: contribution that will give an amplitude of the form of Fig.~D1. This
4279: will come from the tree diagram shown in Fig.~D9.
4280: \begin{center}
4281: \epsfxsize=3.5in
4282: \epsffile{ehes98.ps}
4283:
4284: Fig.~D9 The Tree Diagram Giving the Surviving Amplitude
4285: \end{center}
4286: The piece of this diagram that we have picked out
4287: would vanish if both the quark and antiquark
4288: vertical lines were on-shell. From Fig.~D8 and (\ref{d2}) it is clear that
4289: this piece gives a ``superficially non-leading'' amplitude of the form
4290: $$
4291: \frac{(\pt'+\kt')^*}{P_+k_-}~\kt~\frac{(\pt''-\kt')^*}{P_-k_+}~
4292: \auto\label{d3}
4293: $$
4294: which, after we use the mass-shell conditions for $P_+$ and $P_-$,
4295: gives the amplitude
4296: $$
4297: \bigl(~\frac{k'_+k'_-}{k_+ k_-}~\bigr)~~~
4298: \frac{\kt}{(\pt' + \kt')(\pt'' - \kt')}
4299: \auto\label{d4}
4300: $$
4301: Since both $k'_+k'_-$ and $k_+ k_-$ are finite, this is, indeed, an amplitude
4302: of the form of Fig.~D1.
4303:
4304: We conclude that the large transverse momentum behavior of the
4305: amplitude in Fig.~3, which combines with the loop
4306: amplitude in the top half of Fig.~1 to give the anomaly, does not cancel
4307: after the imposition of Ward identities. In this respect, therefore,
4308: nothing is gained by implementing the Ward identity cancelations.
4309: However, the lack of cancelation is
4310: entirely due to the asymptotic on-shell nature of the quark and antiquark lines.
4311: This raises a general issue of principle. Including the remaining amplitude
4312: that would put these lines exactly on-shell would apparently cancel the behavior
4313: (\ref{d4}). Yet this amplitude would normally be neglected as contributing only to
4314: non-leading high-energy behavior. It can contribute to the leading
4315: behavior only if there are large transverse momentum divergences.
4316: In fact, as we have emphasized, to carry out
4317: the Ward identity cancelations we have actually included several diagrams
4318: that would normally be considered non-leading.
4319:
4320: The normal procedure is to effectively assume in advance (and then
4321: justify a posteori) that transverse momenta will be
4322: sufficiently cut-off after the summation over all diagrams. The
4323: leading parts of diagrams can then be safely extracted without worrying
4324: about transverse momentum divergences. The occurrence of the anomaly
4325: enhancement phenomenon in the diagrams that we have discussed could imply that
4326: in many other diagrams large transverse
4327: momenta are also sufficiently important that the normal methods are inadequate.
4328: If this is the case, then it is likely that the general transverse momentum
4329: diagram formalism will fail. Since there would then be no reggeon diagram
4330: formalism, $t$-channel unitarity
4331: is also likely to fail.
4332: The conclusion, which is really the main conclusion of this paper
4333: (as we already stated in Section 8),
4334: is that to use the transverse momentum diagram formalism (and therefore
4335: to ensure $t$-channel unitarity) it is essential to initially
4336: employ a transverse momentum cut-off.
4337:
4338: \newpage
4339:
4340: \begin{thebibliography}{99}
4341:
4342: \bibitem{fkl} V.~S.~Fadin, E.~A.~Kuraev and L.~N.~Lipatov, {\it Sov. Phys.
4343: JETP} {\bf 45}, 199 (1977).
4344:
4345: \bibitem{bs} J.~B.~Bronzan and R.~L.~Sugar, {\it Phys. Rev.} {\bf D17},
4346: 585 (1978), this paper organizes into reggeon diagrams the results from
4347: H.~Cheng and C.~Y.~Lo, Phys. Rev. {\bf D13}, 1131 (1976),
4348: {\bf D15}, 2959 (1977).
4349:
4350: \bibitem{fs} V.~S.~Fadin and V.~E.~Sherman, Sov. Phys. JETP {\bf 45},
4351: 861 (1978).
4352:
4353: \bibitem{fl1} V.~S.~Fadin and L.~N.~Lipatov, {\it Nucl. Phys.} {\bf B406},
4354: 259 (1993).
4355:
4356: \bibitem{fl} V.~S.~Fadin and L.~N.~Lipatov, {\it Nucl. Phys.} {\bf B477},
4357: 767 (1996) and further references therein.
4358:
4359: \bibitem{jb} J.~Bartels, {\it Z. Phys.} {\bf C60}, 471 (1993) and further
4360: references therein.
4361:
4362: \bibitem{arw93} A.~R.~White, {\it Int. J. Mod. Phys.} {\bf A8}, 4755 (1993).
4363:
4364: \bibitem{arw02} A.~R.~White, {\it Phys. Rev.} {\bf D66}, 056007 (2002).
4365:
4366: \bibitem{arw021} A.~R.~White, {\it Phys. Rev.} {\bf D66}, 045009 (2002).
4367:
4368: \bibitem{arw84} A.~R.~White, {\it Phys. Rev.} {\bf D29}, 1435 (1984).
4369:
4370: \bibitem{rk} For a detailed review and further references, see R.~Kirschner,
4371: {\it Nucl. Phys. Proc. Suppl.} {\bf 51C}, 118 (1996).
4372:
4373: \bibitem{gth} G. 't Hooft, {\it Nucl. Phys.} {\bf B33}, 173 (1971).
4374:
4375: \bibitem{awb} S.~l.~Adler and W.~B.~Bardeen, {\it Phys. Rev.}
4376: {\bf 182} 1517-1536 (1969).
4377:
4378: \bibitem{arw98} A.~R.~White, {\it Phys. Rev.} {\bf D58}, 074008 (1998).
4379:
4380: \bibitem{cri} A.~A.~Migdal, A.~M.~Polyakov and K.~A.~Ter-Martirosyan,
4381: {\it Zh. Eksp. Teor. Fiz.} {\bf 67}, 84 (1974);
4382: H.~D.~I.~Abarbanel and J.~B.~Bronzan, {\it Phys. Rev.} {\bf D9}, 2397 (1974).
4383:
4384: \bibitem{kms} R.~Kirschner, L.~Mankiewicz, A.~Schafer and L.~Szymanowski,
4385: {\it Z. Phys.} {\bf C74}, 501 (1997).
4386:
4387:
4388: \end{thebibliography}
4389:
4390: \end{document}
4391:
4392: