1: \documentclass[10pt,paper]{JHEP3}

2: \usepackage[dvips]{graphicx}

3: %\usepackage{epsfig}

4: \usepackage{amsmath}

5: 

6: 

7: \def\zetaslash{\zeta\hskip-5.5pt/\hskip0.5pt}

8: \def\kslash{k\hskip-5.5pt/\hskip0.5pt}

9: 

10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

11: \setlength\overfullrule{10pt}

12: 

13: \skip\footins = 1\bigskipamount plus 2pt minus 4pt

14: 

15: 

16: \newcommand{\tr}{\mathop{\rm tr}}

17: 

18: \title{5-field terms in the open superstring effective action}

19: 

20: 

21: 

22: \author{Luiz Antonio Barreiro$^{1,2}$ and  Ricardo Medina$^{1}$\\

23: $^{1}$Instituto de Ci{\^e}ncias Exatas, Universidade Federal de Itajub\'{a}\\

24: \hphantom{$^{1}$}Itajub\'a, Minas Gerais, Brazil\\

25: $^{2}$Instituto de Ci{\^e}ncias Exatas, Funda\c{c}\~{a}o de Ensino e Pesquisa de Itajub\'{a}\\

26: \hphantom{$^{1}$}Itajub\'a, Minas Gerais, Brazil\\

27: \hphantom{$^{1}$}E-mails: \email{barreiro@unifei.edu.br,

28: rmedina@unifei.edu.br}.}

29: 

30: 

31: 

32: 

33: \abstract{Some time ago the bosonic and fermionic 4-field terms of

34: the non-abelian low energy effective action of the open

35: superstring were obtained, to all order in $\alpha'$. This was

36: done at tree level by directly generalizing the abelian case,

37: treated some time before, and considering the known expressions of

38: all massless superstring 4-point amplitudes (at tree level). In

39: the present work we obtain the bosonic 5-field terms of this

40: effective action, to all order in $\alpha'$. This is done by

41: considering the simplified expression of the superstring 5-point

42: amplitude for massless bosons, obtained some time ago.}

43: 

44: %\preprint{DFPD 03/TH/38}

45: 

46: \begin{document}

47: 

48: \section{Introduction}

49: 

50: It has been known for a long time that String Theory modifies the

51: Yang-Mills and Einstein lagrangians  by adding $\alpha'$ terms to

52: them \cite{Scherk1}. Since the middle eighties many results of

53: this type were found, among which it was seen that the low energy

54: dynamics of abelian open strings (corresponding to photons) was

55: described by the Born-Infeld lagrangian \cite{Fradkin1, Tseytlin1,

56: Abouelsaood1, Bergshoeff1, Leigh1}, as long as the field strength

57: is kept constant. This is an all order in $\alpha'$ result. Some

58: non-abelian $\alpha'$ corrections of the effective lagrangian were

59: also determined by that time using the scattering amplitude

60: approach \cite{Tseytlin1, Gross1, Kitazawa1}.\\

61: \noindent After the discovery of D-branes as Ramond-Ramond charged

62: states \cite{Polchinski0} and that the low energy dynamics of

63: non-abelian open superstrings captured an equivalent description

64: in terms of them \cite{Witten1}, a supersymmetric generalization

65: of the abelian Born-Infeld lagrangian, in this context, was

66: constructed \cite{Cederwall0, Cederwall0p, Bergshoeff0,

67: Aganagic1}. Besides these results, abelian corrections were

68: obtained in

69: \cite{Wyllard2, Andreev2, DeRoo1}, to all order in $\alpha'$.\\

70: \noindent A non-abelian generalization of the Born-Infeld

71: lagrangian, in the context of Superstring Theory, has been

72: proposed in \cite{Tseytlin2} by means of a symmetrized trace

73: prescription. The complete ${\cal O}({\alpha'}^4)$ terms of

74: \cite{Koerber2, Sevrin1, Nagaoka1} have been written following

75: this prescription for the $F^6$ terms. It has been seen that, in

76: the open superstring effective lagrangian, the covariant

77: derivative terms are as important as the $F^n$ ones

78: \cite{Hashimoto1, Bilal1}. Also, an interesting proposal about the

79: general structure of the non-abelian Born-Infeld action and its

80: covariant derivative terms has been made in \cite{Cornalba1} by

81: means of the Seiberg-Witten map \cite{Seiberg1}, but this result

82: has unknown coefficients and explicit expressions in $D=10$ are

83: not given. So, in this sense, in the non-abelian case, the usually

84: known exact results for the lagrangian terms are strictly

85: perturbative in $\alpha'$. This has been done by different methods

86: which include the 1-loop effective action for $N=4$ SYM

87: \cite{Refolli1, Grasso1}; deformations of the Yang-Mills

88: lagrangian considering either BPS solutions to it \cite{Fosse1,

89: Koerber1, Koerber2} or supersymmetry requirements \cite{DeRoo2,

90: Cederwall1, Cederwall2, Drummond1}; and the scattering amplitude

91: approach \cite{Tseytlin1, Gross1, Kitazawa1,

92: Brandt1}. \\

93: \noindent It is generally believed that scattering

94: amplitudes are only used in String Theory to find the $first$

95: $\alpha'$ corrections in the effective lagrangians, but this is

96: not true: it was first seen in \cite{DeRoo1} that all the

97: $\alpha'$ information of the open superstring 4-point amplitudes,

98: of massless bosons and fermions, can be taken to the effective

99: lagrangian, at least in the abelian case. This was soon

100: generalized to the non-abelian case and to the 4-point effective

101: actions of the NS-NS sector of Closed Superstring Theory

102: \cite{Chandia1}. In the present work we go further and find the

103: $D^{2n}F^5$ terms of the open superstring non-abelian effective

104: lagrangian, to all order in $\alpha'$. This is done by using the

105: 5-point amplitude, first completely calculated in

106: \cite{Brandt1}\footnote{In \cite{Kitazawa1} a partial computation

107: of this 5-point amplitude was done.} and afterwards simplified in

108: \cite{Machado1}. In contrast to the 4-point case, due to the

109: presence of poles in the $\alpha'$ terms of the scattering

110: amplitude, this is a very

111: much more complicated problem to solve.\\

112: \noindent We have organized this paper in a main body and

113: appendices. These last ones contain, besides conventions,

114: identities and tensors, some lengthy formulas and derivations,

115: which are important, but which would otherwise have turned the

116: main body too long. The structure of the main body of the paper is

117: as follows. In section \ref{345-point} we give a very brief review

118: about scattering amplitudes and the low energy effective

119: lagrangian inferred from them. In section \ref{34-point} we review

120: the known structure of the effective lagrangian, as far as 3 and

121: 4-point amplitudes are concerned. In section \ref{5-point} we

122: review the derivation of the 5-point amplitude (at tree level),

123: since this is the starting point for the present paper,  and we

124: confirm that this formula satisfies the usual properties of

125: scattering amplitudes in Open Superstring Theory. We have placed

126: the main result of this work, namely, the explicit $D^{2n}F^5$

127: terms of the effective lagrangian (to all order in $\alpha'$) in

128: section \ref{lagrangian}. This section also contains the

129: scattering subamplitude that leads to that lagrangian and some

130: explicit examples of $\alpha'$ terms up to ${\cal O}({\alpha'}^4)$

131: order\footnote{In principle, with the $\alpha'$ expansions that we

132: give in appendix \ref{expansions} we could explicitly write the

133: $D^{2n}F^5$ terms up to ${\cal O}({\alpha'}^6)$ order.}. Finally,

134: section \ref{final} contains a summary and final remarks,

135: including future directions. Throughout this work we have treated

136: the very involved calculations that arise in scattering

137: amplitudes using $Mathematica$'s $FeynCalc \ 3.5$ package.\\

138: 

139: 

140: 

141: \section{Review of scattering amplitudes and low energy effective lagrangian}

142: \label{345-point}

143: 

144: The $M$-point tree level scattering amplitude for massless bosons,

145: in superstring theory, is given by \cite{Schwarz1}

146: \begin{eqnarray}

147: {\cal A}^{(M)} = i \ (2 \pi)^{10} \delta(k_1 + k_2 + \ldots + k_M)

148: \cdot {\sum_{j_1, j_2, \ldots , j_M}}' \tr(\lambda^{a_{j_1}}

149: \lambda^{a_{j_2}} \ldots \lambda^{a_{j_M}}) \ A(j_1, j_2, \ldots,

150: j_M) \ , \label{general-amplitude}

151: \end{eqnarray}

152: where the sum $\sum'$ in the indices $\{j_1, j_2, \ldots, j_M\}$

153: is done over $non-cyclic$ equivalent permutations of the group

154: $\{1, 2, \ldots, M\}$. The matrices $\lambda^{a_{j}}$ are in the

155: adjoint representation of the Lie group\footnote{At tree level,

156: the Lie groups $SO(N)$ and $USp(N)$ have been shown to be

157: consistent with the description of superstring

158: interactions\cite{Schwarz1}. It has also been seen that $U(N)$ is

159: the appropriate gauge group when describing the low energy

160: dynamics of $N$ coincident D-branes \cite{Witten1}. In the present

161: work it will not be necessary to make any especial reference to

162: any of these Lie groups.} and $A(j_1,

163: j_2, \ldots, j_M)$, called $subamplitude$, corresponds to the

164: $M$-point amplitude of open superstrings which do not carry color

165: indices and which are placed in the ordering $\{j_1, j_2, \ldots,

166: j_M\}$ (modulo cyclic permutations).

167: 

168: Since the present work is based in the determination of the

169: effective lagrangian by means of the known expressions of the

170: scattering amplitudes\footnote{To the knowledge of the authors,

171: this method was first introduced in \cite{Scherk1}.}, in the

172: following subsections we briefly review the 3, 4 and 5-point (tree

173: level) subamplitudes of massless bosons in Open Superstring

174: Theory. From them, an on-shell effective lagrangian of the form

175: \begin{eqnarray}

176: {\cal L}_{\rm ef} = {\cal L}_{YM}  + {\cal L}_{D^{2n}F^4} + {\cal

177: L}_{D^{2n}F^5} \label{effective1}

178: \end{eqnarray}

179: emerges, where

180: \begin{eqnarray}

181: \label{LYM}

182: {\cal L}_{YM} &=& -\frac{1}{4} \tr(F^2) \\

183: \label{LD2nF4} {\cal L}_{D^{2n}F^4} &=& {\alpha'}^2 g^2 \tr(F^4) +

184: {\alpha'}^3 g^2 \tr(D^2F^4) + {\alpha'}^4 g^2 \tr(D^4F^4) + \ldots \\

185: \label{LD2nF5} {\cal L}_{D^{2n}F^5} &=& {\alpha'}^3 g^3 \tr(F^5) +

186: {\alpha'}^4 g^3 \tr(D^2F^5) + {\alpha'}^5 g^3 \tr(D^4F^5) + \ldots

187: \ .

188: \end{eqnarray}

189: The lagrangian in (\ref{effective1}) agrees with the open

190: superstring effective lagrangian up to 5-field terms: any

191: difference between them is sensible only to 6 or higher-point

192: scattering amplitudes. Also, there is no unique way in writing its

193: terms since some freedom arises from the $[D,D]F=-ig[F,F]$

194: relation, the Bianchi identity and integration by parts (see

195: \cite{Tseytlin1},\cite{Koerber1} and \cite{Grasso1}, for example),

196: thus, allowing to interchange some terms from ${\cal

197: L}_{D^{2n}F^4}$ and ${\cal L}_{D^{2n}F^5}$. In the next section we

198: will specify the convention that we will use in writing the

199: $D^{2n}F^4$ terms.

200: 

201: 

202: \section{3 and 4-point subamplitudes and their effective

203: lagrangian in Open Superstring Theory}

204: \label{34-point}

205: 

206: The tree level 3 and 4-point subamplitudes for massless bosons, in

207: Open Superstring Theory, have been known for a long time. Their

208: on-shell expressions are the following \cite{Schwarz1},

209: respectively:

210: \begin{eqnarray}

211: \label{A123}

212: A(1,2,3) & = & 2 g \left[\frac{}{} (\zeta_1 \cdot

213: k_2)(\zeta_2 \cdot \zeta_3) + (\zeta_2 \cdot k_3)(\zeta_3 \cdot

214: \zeta_1) + (\zeta_3 \cdot k_1)(\zeta_1 \cdot \zeta_2) \frac{}{}

215: \right] \ ,

216: \\

217: \label{A1234} A(1,2,3,4) & = & 8 \ g^2 \ {\alpha'}^2 \frac{\Gamma(-

218: \alpha' s) \Gamma(- \alpha' t)}{\Gamma(1- \alpha' s - \alpha' t)}

219: K(\zeta_1, k_1; \zeta_2, k_2; \zeta_3, k_3; \zeta_4, k_4) \ ,

220: \end{eqnarray}

221: where

222: \begin{eqnarray}

223: K(\zeta_1, k_1; \zeta_2, k_2; \zeta_3, k_3; \zeta_4, k_4) =

224: t_{(8)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4}

225: \zeta_{\mu_1}^1 k_{\nu_1}^1 \zeta_{\mu_2}^2 k_{\nu_2}^2

226: \zeta_{\mu_3}^3 k_{\nu_3}^3 \zeta_{\mu_4}^4 k_{\nu_4}^4

227: \label{kinematicfactor}

228: \end{eqnarray}

229: is a kinematic factor, $t_{(8)}$ being a known tensor

230: \cite{Schwarz1}. The $s$ and $t$ variables in (\ref{A1234}) are part

231: of the three Mandelstam variables, which are defined as

232: \begin{eqnarray}

233: s = -(k_1 + k_2)^2, \ \ \, t = -(k_1 + k_4)^2, \ \ \, u = -(k_1 +

234: k_3)^2 \ \ \ .

235: \label{Mandelstam}

236: \end{eqnarray}

237: 

238: The Gamma factor in (\ref{A1234}) has a completely known $\alpha'$

239: expansion (see appendix \ref{expansion-Gamma}) which begins like

240: \begin{eqnarray}

241: {\alpha'}^2 \frac{\Gamma(- \alpha' s) \Gamma(- \alpha' t)}

242: {\Gamma(1- \alpha' s - \alpha' t)} = \frac{1}{s t} - \frac{\

243: \pi^2}{6}{\alpha'}^2 + {\cal O}({\alpha'}^3)\,. \label{gammas}

244: \end{eqnarray}

245: Since the leading term in the $\alpha'$ expansion of $A(1,2,3,4)$ in

246: (\ref{A1234}) is the Yang-Mills 4-point subamplitude,

247: $A_{YM}(1,2,3,4)$, then it may also be written as

248: \begin{eqnarray}

249: A(1,2,3,4) & = &  {\alpha'}^2 st \ \frac{\Gamma(- \alpha' s)

250:  \Gamma(- \alpha' t)}{\Gamma(1- \alpha' s - \alpha' t)}

251: A_{YM}(1,2,3,4) \ . \label{A1234II}

252: \end{eqnarray}

253: This last formula will be useful in the next section, when making

254: a comparison between the 4 and 5-point subamplitudes.\\

255: \noindent From the expression of $A(1,2,3)$ in (\ref{A123}) it is

256: immediate that the effective lagrangian may be written with no

257: $D^{2n}F^3$ terms (as it has already been done in

258: (\ref{effective1})) since this amplitude contains no $\alpha'$

259: corrections and agrees with

260: the one from the Yang-Mills lagrangian.\\

261: \noindent Using the expression of $A(1,2,3,4)$ in (\ref{A1234})

262: the lagrangian ${\cal L}_ {D^{2n}F^4}$ has been determined some

263: time ago\cite{Chandia1}\footnote{In fact, not only the $D^{2n}F^4$

264: terms were determined in \cite{Chandia1}, but also all

265: boson-fermion and fermion-fermion terms which are sensible to

266: 4-point amplitudes.}, to all order in $\alpha'$. This was done by

267: directly generalizing the procedure considered in \cite{DeRoo1}

268: for the non-abelian case. The final result is the

269: following\footnote{There are three differences between eq.

270: (\ref{Seff-nonabelian}) and the corresponding one in section 2.2

271: of \cite{Chandia1}:

272: \begin{enumerate}

273: \item We have now written the $F^4$ terms of the integrand in an

274: abbreviated manner, by means of the $t_{(8)}$ tensor and, as a

275: consequence of the symmetries of this tensor, it has been possible

276: to write ${\cal L}_{D^{2n}F^4}$ in terms of only one function

277: $f_{\rm sym}$, instead of the three ones contained in the ${\cal

278: G}^{\ a_1 a_2 a_3 a_4}_{\rm \alpha' corr.}(s, t, u)$ function of

279: \cite{Chandia1}. \item We have introduced a symmetrized

280: prescription in the $\alpha'$ expansion of function $f$, explained

281: in eq. (\ref{fsym}), and that is why we have now called it $f_{\rm

282: sym}$. \item We have included derivative terms which were not

283: considered in \cite{Chandia1}, namely, the ones with

284: $D^2=D^{\alpha}D_{\alpha}$ operating on a same field strength.

285: \end{enumerate}

286: Neither item 2 nor item 3 invalidate the result of

287: \cite{Chandia1}, since there were considered results sensible up

288: to 4-point amplitudes only.}:

289: \begin{multline}

290: {\cal L}_{D^{2n}F^4}  = - \frac{1}{8} g^2 {\alpha'}^2 \

291:  \int \int \int \int \

292: \left\{ \frac{}{}\prod_{j=1}^{4} d^{10} x_j \ \delta^{(10)}(x-x_j)

293: \frac{}{} \right\} \times \\

294: \begin{split}

295: & \times f_{\rm sym} \left( \frac{}{} \frac{ (D_{1} + D_{2})^2 +

296: (D_{3} + D_{4})^2}{2} ,

297: \frac{(D_{1} + D_{4})^2 + (D_{2} + D_{3})^2}{2} \frac{}{} \right) \times \\

298: & \times t_{(8)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4}

299: \tr \biggl( F_{\mu_1 \nu_1}(x_1)F_{\mu_2 \nu_2}(x_2)F_{\mu_3

300: \nu_3}(x_3)F_{\mu_4 \nu_4}(x_4) \biggr) \ ,

301: \label{Seff-nonabelian}

302: \end{split}

303: \end{multline}

304: where the function $f$ is given by

305: \begin{eqnarray}

306: f(s, t) & = & \frac{\Gamma(- \alpha' s)\Gamma(- \alpha' t)}

307: {\Gamma(1 - \alpha' s - \alpha' t)} - \frac{1}{{\alpha'}^2 st} \ .

308: \label{f}

309: \end{eqnarray}

310: In (\ref{Seff-nonabelian}) `$f_{\rm sym}$' denotes that in the

311: power series of it, which involves powers of $s$ and $t$, we are

312: using a symmetrized convention:

313: \begin{eqnarray}

314: (s^k t^l)_{\rm sym} = \frac{( \ \mbox{Sum of all different

315: permutations of k powers of $s$ and l powers of $t$} \

316: )}{\left(\begin{array}{c}

317:                 k+l\\

318:                 k

319:                \end{array}\right)} \ .

320: \label{fsym}

321: \end{eqnarray}

322: \noindent In the case of ordinary numbers the formula in

323: (\ref{fsym}) coincides with $s^k t^l$, but this does not happen in

324: the case of covariant derivative operators, as it is in

325: (\ref{Seff-nonabelian}).\\

326: \noindent The terms of ${\cal L}_{D^{2n}F^4}$ were written

327: explicitly up to ${\cal O}({\alpha'}^5)$ order in \cite{Chandia1} .\\

328: \noindent The fact that $A(1,2,3,4)$ has no poles at any (non zero)

329: $\alpha'$ order made the construction of ${\cal L}_{D^{2n}F^4}$, in

330: (\ref{Seff-nonabelian}), quite direct: besides some trace factors,

331: the lagrangian is constructed in terms of the 4-point Gamma factor

332: in which the momenta are substituted by covariant derivatives

333: (appropriately symmetrized). All this implies an enormous

334: simplification with respect to other methods of obtaining the

335: $\alpha'$ correction terms in the effective lagrangian, since all of

336: them require, order by order in $\alpha'$, of an explicit

337: construction of gauge invariant terms with unknown coefficients

338: which are afterwards determined by some matching with Superstring

339: Theory 4-point amplitudes (see \cite{Tseytlin1}, \cite{Bilal1},

340: \cite{Koerber1} and \cite{Grasso1}, for example).\\

341: 

342: 

343: 

344: 

345: \section{The 5-point amplitude of massless bosons in Open Superstring Theory}

346: \label{5-point}

347: 

348: \noindent A first calculation of the 5-point amplitude for

349: massless bosons in Open Superstring Theory was done a long time

350: ago in \cite{Kitazawa1}. This was a partial calculation with

351: enough information to calculate the $F^5$ and the $D^2F^4$ terms

352: that appear at ${\cal O}({\alpha'}^3)$ order in the effective

353: lagrangian (\ref{effective1}). Long after this result, in

354: \cite{Koerber1} it was seen that, of all these terms, only the

355: $D^2F^4$ ones were correctly determined. The corrected $F^5$ terms

356: of \cite{Koerber1} were also confirmed subsequently by other

357: methods \cite{DeRoo2,Brandt1,Grasso1, Drummond1}.\\

358: \noindent Since the main result of the present work, namely, the

359: determination of the $D^{2n}F^5$ terms of the effective

360: lagrangian, lies on the computation of the 5-point subamplitude,

361: in the next subsection we will review the main steps of it in some

362: detail. In the subsection after the next one we will check that

363: our 5-point formula satisfies the usual properties of massless

364: open string subamplitudes \cite{Schwarz1, Mangano1}: cyclicity,

365: (on-shell) gauge invariance, world-sheet parity and

366: factorizability.

367: 

368: \subsection{Review of the 5-point subamplitude derivation}

369: \label{review-5}

370: 

371: \noindent A complete expression for the 5-point amplitude was

372: first obtained in \cite{Brandt1}:

373: \begin{eqnarray}

374: \lefteqn{ A(1, 2, 3, 4, 5)  =} && \nonumber\\&& 2 g^3 (2

375: \alpha')^2 \biggl[

376:   (\zeta_1 \cdot \zeta_2)(\zeta_3 \cdot \zeta_4)

377:       \Bigl\{   (\zeta_5 \cdot k_1)(k_2 \cdot k_3) L_2

378:      - (\zeta_5 \cdot k_2)(k_1 \cdot k_3) L_3 +

379: \nonumber \\ & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[ (\zeta_1

380: \cdot \zeta_2)(\zeta_3

381:   \cdot \zeta_4) \Bigl\{}

382: + (\zeta_5 \cdot k_3) \left( (k_2 \cdot k_4){L_3}'+(k_2 \cdot

383: k_3)L_2 \right)

384:   \Bigr\}+

385: \nonumber \\

386: & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[} + (\zeta_1 \cdot

387: \zeta_3)(\zeta_2 \cdot \zeta_4) \Bigl\{

388:     -(\zeta_5 \cdot k_1)(k_2 \cdot k_3) L_7 +

389: \nonumber

390: \\

391: && \hphantom{2 g^3 (2 \alpha')^2 \Bigl[+ (\zeta_1 \cdot

392: \zeta_3)(\zeta_2

393:     \cdot \zeta_4) \Bigl\{}

394: +  (\zeta_5 \cdot k_2) \left( (k_3 \cdot k_4){L_1}'-(k_2\cdot

395: k_3)L_7

396:   \right) -

397: \nonumber \\

398: & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[+ (\zeta_1 \cdot

399: \zeta_3)(\zeta_2

400:     \cdot \zeta_4) \Bigl\{}

401: - (\zeta_5 \cdot k_3) (k_1 \cdot k_2)L_1   \Bigr\}+ \nonumber\\

402: & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[} + (\zeta_1 \cdot

403: \zeta_4)(\zeta_2 \cdot \zeta_3) \Bigl\{   (\zeta_5 \cdot k_1)

404: \left(   (k_3 \cdot k_4){K_4}'-(k_1 \cdot k_3)K_5 \right) -

405: \nonumber\\

406:  & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[+ (\zeta_1

407: \cdot \zeta_4)(\zeta_2

408:     \cdot \zeta_3) \Bigl\{ }

409: - (\zeta_5 \cdot k_2)(k_1 \cdot k_3) K_5 + (\zeta_5 \cdot k_3)(k_1

410: \cdot k_2) K_4   \Bigr\}+ \nonumber

411: \end{eqnarray}

412: \begin{eqnarray}

413: & & \hphantom{2 g^3 (2

414: \alpha')^2 \Bigl[} + (\zeta_2 \cdot \zeta_3) \Bigl\{   (\zeta_5

415: \cdot k_1)  \left( (\zeta_1 \cdot k_2)(\zeta_4 \cdot k_3) L_2 -

416: (\zeta_1 \cdot k_3)(\zeta_4 \cdot k_2) L_7   \right)+

417: \nonumber \\

418: & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[+ (\zeta_2 \cdot

419: \zeta_3)\Bigl\{ } + (\zeta_5 \cdot k_2) \Bigl(   (\zeta_1 \cdot

420: k_3)(\zeta_4 \cdot k_1) K_5 + (\zeta_1 \cdot k_3)(\zeta_4 \cdot

421: k_3) {L_4}' +

422: \nonumber\\

423: && \hphantom{2 g^3 (2 \alpha')^2 \Bigl[+ (\zeta_2 \cdot

424: \zeta_3)\Bigl\{ +

425:     (\zeta_5 \cdot k_2)\Bigl(} +(\zeta_1 \cdot k_2)(\zeta_4 \cdot k_3)

426:     L_2 + (\zeta_1 \cdot k_4)(\zeta_4 \cdot k_3) {K_4}' \Bigr)-

427: \nonumber \\

428: & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[+ (\zeta_2 \cdot

429: \zeta_3)\Bigl\{} - (\zeta_5 \cdot k_3) \Bigl(   (\zeta_1 \cdot

430: k_2)(\zeta_4 \cdot k_1) K_4 + (\zeta_1

431:  \cdot k_3)(\zeta_4 \cdot k_2) L_7 +

432: \nonumber\\

433:  && \hphantom{2 g^3 (2 \alpha')^2 \Bigl[+ (\zeta_2

434: \cdot \zeta_3)\Bigl\{-

435:     (\zeta_5 \cdot k_3)\Bigl(}

436: +(\zeta_1 \cdot k_2)(\zeta_4 \cdot

437:  k_2) L_4 +

438: (\zeta_1 \cdot k_4)(\zeta_4 \cdot k_2) {K_5}'

439:   \Bigr)     \Bigr\}+ \nonumber\\

440:  & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[} + (\zeta_1

441: \cdot \zeta_4) \Bigl\{   (\zeta_5 \cdot k_2)  \left( (\zeta_2

442: \cdot k_1)(\zeta_3 \cdot k_4) K_2 - (\zeta_2 \cdot k_4)(\zeta_3

443: \cdot k_1) K_3   \right)- \nonumber \\ & & \hphantom{2 g^3 (2

444: \alpha')^2 \Bigl[+ (\zeta_1 \cdot \zeta_4) \Bigl\{} - (\zeta_5

445: \cdot k_1) \Bigl(   (\zeta_2 \cdot k_3)(\zeta_3 \cdot k_4) {K_4}'

446:  - (\zeta_2 \cdot k_4)(\zeta_3 \cdot k_2) {K_5}' +

447: \nonumber

448: \\

449: && \hphantom{2 g^3 (2 \alpha')^2 \Bigl[+ (\zeta_1 \cdot \zeta_4)

450: \Bigl\{- (\zeta_5 \cdot k_1)\Bigl(} +(\zeta_2 \cdot

451:  k_4)(\zeta_3 \cdot k_4) {K_1}' -

452: (\zeta_2

453:  \cdot k_1)(\zeta_3 \cdot k_4) {K_2}

454:   \Bigr)+

455: \nonumber

456: \\ & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[} + (\zeta_5

457: \cdot k_4) \Bigl(   (\zeta_2 \cdot k_1)(\zeta_3 \cdot k_2) K_4   +

458: (\zeta_2 \cdot k_1)(\zeta_3 \cdot k_1) K_1- \nonumber\\ &&

459: \hphantom{2 g^3 (2 \alpha')^2 \Bigl[+ (\zeta_5 \cdot k_4) \Bigl(}

460: - (\zeta_2 \cdot k_3)(\zeta_3 \cdot k_1) K_5 - (\zeta_2 \cdot

461: k_4)(\zeta_3 \cdot k_1) {K_3}

462:   \Bigr)     \Bigr\} \biggr] +

463: \nonumber

464: \\ & & \hphantom{2 g^3 (2 \alpha')^2 \Bigl[} +  \left( \

465: \mbox{cyclic permutations of indexes (1,2,3,4,5)} \ \right) \ .

466: \label{amplitude}

467: \end{eqnarray}

468: In this formula, $K_i$, ${K_i}'$, $L_i$ and ${L_i}'$ are $\alpha'$

469: dependent factors\footnote{In \cite{Brandt1} they were called

470: `kinematic factors' but we will not use this terminology any longer

471: since the word `kinematic' is usually reserved for expressions which

472: depend on both, momenta $and$ polarizations. In the present work we

473: will use the terminology `$\alpha'$ dependent factor' to denote an

474: expression which depends on $\alpha'$ $and$ the momenta $k_i$.},

475: defined by a double integral of the form

476: \begin{eqnarray}

477: \int_0^1 d x_3 \int_0^{x_3} d x_2 \ x_3^{2 \alpha' \alpha_{13}}

478: (1-x_3)^{2 \alpha'

479:   \alpha_{34}} {x_2}^{2 \alpha' \alpha_{12}} (1-x_2)^{2 \alpha'

480:   \alpha_{24}} (x_3-x_2)^{2 \alpha' \alpha_{23}} \cdot

481: \varphi(x_2, x_3) \ , \label{double-int}

482: \end{eqnarray}

483: where the function $\varphi(x_2, x_3)$ has a specific expression

484: for each of them (see appendix A.1 of \cite{Brandt1} for further

485: details). They can all be calculated as a product of a Beta and a

486: Hypergeometric function \cite{Kitazawa1} and they have a well

487: defined Laurent expansion in $\alpha'$ (after regularizing in some

488: cases). For example, in the case of the factors $K_2$ and $K_3$ we

489: have that \cite{Brandt1}\footnote{In \cite{Brandt1}, the

490: expansions of $K_2$ and $K_3$ contained the variables $\rho$ and

491: $\alpha_{24}$ which we have already substituted using the

492: relations (\ref{rho}), (\ref{alpha13}) and (\ref{alpha24}) of

493: appendix \ref{expansion-K3K}.}:

494: \begin{eqnarray}

495: K_2 & = & \frac{1}{(2 \alpha')^2} \left\{ \frac{1}{\alpha_{12} \

496: \alpha_{34}} \right\} - \frac{ \ \pi^2}{6} \left\{\frac{

497: \alpha_{51} \ \alpha_{12} - \alpha_{12} \ \alpha_{34} +

498: \alpha_{34} \ \alpha_{45}}{\alpha_{12} \ \alpha_{34}} \right\} +

499: \nonumber

500: \\&+& \zeta(3) \ (2 \alpha') \left\{  \frac{\alpha_{12}^2 \ \alpha_{51}

501:  - \alpha_{34}^2 \ \alpha_{12} +

502: \alpha_{45}^2 \ \alpha_{34} + \alpha_{51}^2 \ \alpha_{12} -

503: \alpha_{12}^2 \ \alpha_{34} + \alpha_{34}^2 \ \alpha_{45} - 2

504: \alpha_{12} \ \alpha_{23} \ \alpha_{34} }{\alpha_{12} \

505: \alpha_{34}} \right\}

506: + \nonumber\\

507: &+& {\cal O}((2\alpha')^2) \,, \label{K2}

508: \\

509: K_3 & = & \frac{ \ \pi^2}{6} - \zeta(3) \ (2 \alpha') \left\{

510: \alpha_{12} + \alpha_{23} + \alpha_{34} + \alpha_{45} +

511: \alpha_{51} \right\} + {\cal O}((2\alpha')^2) \, .

512: \label{K3}

513: \end{eqnarray}

514: \noindent In (\ref{K2}) and (\ref{K3}) we are using the notation

515: \begin{eqnarray}

516: \alpha_{ij} = k_i \cdot k_j \ \ \ \ (i,j = 1,2,3,4,5; \ i \neq j) \

517: . \label{alphaij}

518: \end{eqnarray}

519: 

520: \noindent In appendix A.2 of \cite{Brandt1}, among some relations,

521: the following were found:

522: \begin{eqnarray}

523: \alpha_{34} K_2 & = & \alpha_{13} K_1 + \alpha_{23} K_4

524: \nonumber \\

525: \alpha_{24} K_3 & = & \alpha_{12} K_1 - \alpha_{23} K_5

526: \nonumber \\

527: \alpha_{13} L_1 & = & \alpha_{34} {L_3}' - \alpha_{23} {L_4}'

528: \nonumber \\

529: \alpha_{12} K_2 & = & \alpha_{24} {K_1}' + \alpha_{23} {K_4}'

530: \nonumber \\

531: \alpha_{13} K_3 & = & \alpha_{34} {K_1}' - \alpha_{23} {K_5}'

532: \nonumber \\

533: \alpha_{24} {L_1}' & = & \alpha_{12} L_3 - \alpha_{23} L_4

534: \nonumber \\

535: \alpha_{34}{K_4}'-\alpha_{13}K_5 & = &

536: \alpha_{12}K_4-\alpha_{24}{K_5}' \,.

537: \label{relations1}

538: \end{eqnarray}

539: \noindent The first three of them were explicitly written in eq.

540: (A.6) of \cite{Brandt1}. The following three relations can be

541: obtained from the first ones by the $duality$ operation mentioned in

542: that appendix\footnote{From the point of view of the string

543: world-sheet, this $duality$ operation is nothing else than a $twist$

544: of a disk with five insertions, with respect to the fifth vertex.

545: This will be seen in the third item of subsection

546: \ref{properties}.}. All of these expressions can be derived using

547: the definition of each $K_i$, ${K_i}'$, $L_i$ and ${L_i}'$, as a

548: double integral (see appendix A.1 of \cite{Brandt1}), and

549: integration by parts. The last relation in (\ref{relations1}) comes

550: from demanding invariance of $K_6$ under the mentioned $duality$

551: operation and using its expression given in appendix A.2 of the same

552: reference.

553: 

554: \noindent Now, in \cite{Machado1} there were found additional

555: relations, independent of the ones in (\ref{relations1}):

556: \begin{eqnarray}

557: K_2 + (K_1-L_3+{K_1}'-{L_3}')/2 = 0,\nonumber \\

558: K_1-K_4+K_5 = 0, \nonumber \\

559: {K_1}'-{K_4}'+{K_5}' = 0, \nonumber \\

560: K_5+{K_5}'+(K_1-L_4+{K_1}'-{L_4}')/2 = 0, \nonumber \\

561: L_2-L_3-{L_4}'=0, \nonumber \\

562: L_2-{L_3}'- L_4=0, \nonumber \\

563: K_1-{K_1}'-(L_4 - {L_4}') = 0.

564: \label{relations2}

565: \end{eqnarray}

566: These independent relations come from the very definition of each

567: factor, since the function $\varphi(x_2, x_3)$ in every case is a

568: fraction. For example, for the factors $K_1$, $K_4$ and $K_5$, the

569: corresponding $\varphi(x_2, x_3)$ function is given by

570: \begin{eqnarray}

571: \varphi_{K_1}(x_2,x_3) = \frac{1}{x_2 \ x_3} , \ \ \ \

572: \varphi_{K_4}(x_2,x_3) = \frac{1}{x_2 \ (x_3-x_2)} , \ \ \ \

573: \varphi_{K_5}(x_2,x_3) = \frac{1}{x_3 \ (x_3-x_2)} \ \ ,

574: \label{fractions}

575: \end{eqnarray}

576: which can easily be seen to satisfy

577: \begin{eqnarray}

578: \varphi_{K_1}(x_2,x_3) - \varphi_{K_4}(x_2,x_3) +

579: \varphi_{K_5}(x_2,x_3) = 0 \ .

580: \label{sumoffractions}

581: \end{eqnarray}

582: The integrated version of eq. (\ref{sumoffractions}) is precisely

583: the second of the equations in (\ref{relations2}).

584: 

585: \noindent Besides all these relations, we now introduce a new

586: factor, $T$, defined as

587: \begin{eqnarray}

588: T = (2 \alpha')^2 \biggl[ \ \alpha_{12} \ \alpha_{34} \ K_2 + (

589: \alpha_{51} \ \alpha_{12} - \alpha_{12} \ \alpha_{34} +

590: \alpha_{34} \ \alpha_{45} ) \ K_3 \ \biggr]. \label{K2-K}

591: \end{eqnarray}

592: The same as $K_3$, this factor remains invariant under cyclic

593: permutations of indexes $(1,2,3,4,5)$. We prove this in appendix

594: \ref{expansion-K3K}.

595: 

596: \noindent The factors $K_2$ and $K_3$ have been written in terms of

597: Beta and Hypergeometric functions in equations (\ref{K2hyperg}) and

598: (\ref{K3hyperg}).

599: 

600: 

601: \noindent So, summarizing, the 5-point amplitude in

602: (\ref{amplitude}) is given in terms of sixteen factors $K_i$,

603: ${K_i}'$, $L_i$ and ${L_i}'$, which are related with the

604: additional factor $T$ by fifteen independent linear relations

605: given in (\ref{relations1}), (\ref{relations2}) and (\ref{K2-K}).

606: This allows to write the 5-point amplitude in (\ref{amplitude}) in

607: terms of only two factors, which we choose to be $T$ and

608: $K_3$\footnote{Eq. (\ref{amp2}) is the same as the one in Table 2

609: of \cite{Machado1}, but with not exactly the same choice of

610: momentum dependent factors and kinematic expressions.}

611: \begin{eqnarray}

612: A(1,2,3,4,5) =  T \cdot A(\zeta,k) \ + (2  \alpha')^2 \ K_3 \cdot

613: B(\zeta,k) \ . \ \label{amp2}

614: \end{eqnarray}

615: Here, $A(\zeta,k)$ and $B(\zeta,k)$ are two kinematical expressions

616: which are known explicitly, after all the substitutions of the

617: factors have been done in (\ref{amplitude}). They can be identified

618: with subamplitudes of specific terms of the effective lagrangian

619: (\ref{effective1}), as we will see in the next lines.

620: 

621: \noindent In appendix \ref{expansion-K3K} we have that the $\alpha'$

622: expansion of $T$ begins as

623: \begin{eqnarray}

624: T & = & 1 + {\cal O}((2  \alpha')^3) \ . \label{K}

625: \end{eqnarray}

626: Substituting the leading terms of the $\alpha'$ expansions of $T$

627: and $(2  \alpha')^2  K_3$ in (\ref{amp2}), we have that $A(\zeta,k)$

628: should agree with the Yang-Mills 5-point subamplitude and that

629: $B(\zeta,k)$ should agree with the $F^4$ terms 5-point subamplitude.

630: We have checked (computationally) that this really happens on-shell,

631: after using momentum conservation and physical state conditions.

632: 

633: \noindent So our final formula for the 5-point subamplitude is

634: \begin{eqnarray}

635: A(1,2,3,4,5) = T \cdot A_{YM}(1,2,3,4,5) + (2  \alpha')^2 K_3

636: \cdot A_{F^4}(1,2,3,4,5) \ .

637: \label{A12345final}

638: \end{eqnarray}

639: In (\ref{AYM5}) and (\ref{AF412345}) we give the expressions for

640: $A_{YM}(1,2,3,4,5)$ and $A_{F^4}(1,2,3,4,5)$, respectively.\\

641: \noindent The formula in (\ref{A12345final}) has exactly the same

642: structure of the corresponding one for the 4-point subamplitude,

643: written in (\ref{A1234II}), but with two kinematic expressions and

644: two factors which contain the $\alpha'$ dependence.

645: 

646: \subsection{Properties satisfied by the 5-point amplitude}

647: \label{properties}

648: 

649: \noindent On this subsection we confirm that $A(1,2,3,4,5)$, given

650: in (\ref{A12345final}), satisfies the usual properties of massless

651: open string subamplitudes, namely, cyclic invariance, on-shell gauge

652: invariance, world-sheet parity and factorizability.

653: 

654: 

655: \begin{enumerate}

656: \item{\underline{Cyclic invariance}:}\\

657: \noindent It was mentioned in the previous subsection, and it is

658: proved in appendix \ref{expansion-K3K}, that the factors $T$ and

659: $K_3$ are invariant under cyclic permutations of indexes

660: $(1,2,3,4,5)$. Now, once $A_{YM}(1,2,3,4,5)$ and

661: $A_{F^4}(1,2,3,4,5)$ are the Yang-Mills and the $F^4$ terms 5-point

662: subamplitudes, which are non-abelian field theory amplitudes, by

663: construction they are cyclic invariant. So, in this sense, the

664: cyclic invariance of $A(1,2,3,4,5)$ is already manifestly written in

665: formula (\ref{A12345final}).

666: 

667: \item{\underline{On-shell gauge invariance}:}\\

668: \noindent This property consists in that the subamplitude should

669: become zero if any of the polarizations $\zeta_i$ is substituted by

670: the corresponding momentum $k_i$, after using physical state

671: ($\zeta_j \cdot k_j = 0$) and on-shell ($k_j^2=0$) conditions for

672: all external string states, together with momentum

673: conservation\cite{Schwarz1}. That this indeed happens in

674: (\ref{A12345final}) can be understood from the fact that

675: $A_{YM}(1,2,3,4,5)$ and $A_{F^4}(1,2,3,4,5)$ are 5-point

676: subamplitudes that come from gauge invariant terms, so both of them

677: should independently become zero when doing $\zeta_i \rightarrow

678: k_i$ for any $i=1,2,3,4,5$. So the on-shell gauge invariance of

679: $A(1,2,3,4,5)$ is also manifestly

680: written in (\ref{A12345final}).\\

681: \noindent In any case, the explicit check of the on-shell gauge

682: invariance of $A_{F^4}(1,2,3,4,5)$ can be seen as follows. This

683: subamplitude is given by

684: \begin{eqnarray}

685: A_{F^4}(1,2,3,4,5)  =  \ 2 \  g^3

686: \left\{ \frac{}{} \left[ \frac{}{} K(\zeta_1, \zeta_2; \zeta_3,

687: k_3; \zeta_4, k_4; \zeta_5, k_5) \ + \right. \right. \ \ \ \ \ \

688: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \

689: \ \ \ \ \ \nonumber \\

690: + \frac{1}{\alpha_{12}} \left( \frac{}{} (\zeta_1 \cdot \zeta_2)

691: K(k_1, k_2; \zeta_3, k_3; \zeta_4, k_4; \zeta_5, k_5) + (\zeta_1

692: \cdot k_2) K(\zeta_2, k_1 + k_2; \zeta_3, k_3; \zeta_4, k_4;

693: \zeta_5, k_5) \ - \right. \nonumber \\

694: \left. \left. \left. - (\zeta_2 \cdot k_1) K(\zeta_1, k_1 + k_2;

695: \zeta_3, k_3; \zeta_4, k_4; \zeta_5, k_5) \frac{}{} \right)

696: \frac{}{} \right]+ \mbox{(cyclic permutations)} \frac{}{} \right\}

697: \ . \ \ \ \ \

698:  \ \ \ \ \

699: \label{AF412345}

700: \end{eqnarray}

701: It has been obtained as the coefficient of $\pi^2/6 \ (2 \alpha')^2$

702: in the $\alpha'$ series of the subamplitude

703: $A_{D^{2n}F^4}(1,2,3,4,5)$, given in (\ref{AD2nF4}). As mentioned in

704: appendix \ref{D2nF4}, in this formula the expression $K(A, a; B, b;

705: C, c; D, d)$ denotes the same kinematic construction of

706: (\ref{kinematicfactor}), evaluated in the corresponding variables.

707: Due to the symmetries of the $t_{(8)}$ tensor (see appendix

708: \ref{t8tensor}), it is not difficult to see that this last

709: expression becomes zero whenever any of the $\zeta_i$ is substituted

710: by $k_i$.

711: 

712: 

713: 

714: 

715: \item{\underline{World-sheet parity}:}\\

716: \noindent The tree level interaction of $n$ open strings is

717: described by a conformal field theory on a disk with $n$

718: insertions on its boundary\cite{Polchinski1}. A world-sheet parity

719: transformation, $\sigma \rightarrow l - \sigma$ (where $\sigma \

720: \varepsilon \ [0, \ l]$ is the string internal coordinate),

721: corresponds to a twisting of this disk with respect to any of

722: those insertions (in figure \ref{twist5}, for example, it

723: is shown a twisting with respect to the fifth insertion).\\

724: \begin{figure}[h]

725: \centerline{\includegraphics*[scale=0.8,angle=0]{twist.eps}}

726: \caption{Twisting of a disk with five insertions, with respect to

727: the fifth one. Under this operation the indexes $(1,2,3,4,5)$

728: transform as indicated in equation (\ref{twisting5}). }

729: \label{twist5}

730: \end{figure}

731: \noindent When there is an interaction of an odd number of open

732: strings, the subamplitude changes sign under a world-sheet parity

733: transformation \cite{Schwarz1, Polchinski1} . So, in the case of

734: five interacting superstrings, a twisting with respect to, say the

735: fifth one,

736: \begin{eqnarray}

737: (1,2,3,4,5) \rightarrow (4,3,2,1,5) \ , \label{twisting5}

738: \end{eqnarray}

739: should imply that

740: \begin{eqnarray}

741: A(1,2,3,4,5) = - A(4,3,2,1,5) \ . \label{twisting}

742: \end{eqnarray}

743: To see that the amplitude in (\ref{A12345final}) indeed satisfies

744: this condition it can be argued as follows. First, in appendix

745: \ref{K2K3Ktwisting} it is proved that the factors $T$ and $K_3$

746: are invariant under a twisting transformation. Second, the

747: Yang-Mills subamplitude $A_{YM}(1,2,3,4,5)$ already satisfies

748: (\ref{twisting}) \cite{Mangano1}\footnote{This twisting

749: transformation is also called `inversion' in \cite{Mangano1} since

750: due to the cyclic symmetry (\ref{twisting}) it can also be written

751: as $A(1,2,3,4,5) = - A(5,4,3,2,1)$ \ .}. And third, using the

752: symmetries of the $t_{(8)}$ tensor it is not difficult to prove

753: that $A_{F^4}(1,2,3,4,5)$, in (\ref{AF412345}), does also satisfy

754: the twisting condition (\ref{twisting}).

755: 

756: 

757: \item{\underline{Factorizability}:}\\

758: \noindent There are two non trivial tests that confirm the right

759: structure of the poles of $A(1,2,3,4,5)$ in (\ref{A12345final}). The

760: factorization comes when a particular limit is considered. In that

761: limit the subamplitude diverges with simple poles and its residues

762: factorize in terms of the 4-point subamplitude.\\

763: \noindent Here we follow very closely section 6 of \cite{Mangano1}.

764: \begin{itemize}

765: \item{\underline{Soft boson factorization}:}\\

766: This corresponds to the case when one of the bosons, say the fifth

767: one, becomes soft ($k_5 \rightarrow 0$). It is very well known (see

768: \cite{Mangano1}, for example) that the $n$-point Yang-Mills

769: subamplitude satisfies the factorization mentioned here. In the case

770: $n=5$ it has the following form:

771: \begin{eqnarray}

772: A_{YM}(1,2,3,4,5)_{k_5 \rightarrow 0} \sim \ g \left(\frac{}{}

773: \frac{\zeta_5 \cdot k_1}{k_5 \cdot k_1} - \frac{\zeta_5 \cdot

774: k_4}{k_5 \cdot k_4} \frac{}{} \right) \ A_{YM}(1,2,3,4) \ .

775: \label{softYM5}

776: \end{eqnarray}

777: Also, taking $k_5 \rightarrow 0$ in (\ref{AF412345}) it is not

778: difficult to see that $A_{F^4}(1,2,3,4,5)$ satisfies exactly the

779: same factorization.

780: Therefore, taking the $k_5 \rightarrow 0$ limit in

781: (\ref{A12345final}) and using that

782: \begin{eqnarray}

783: A_{YM}(1,2,3,4) & = & 2 \ g^2 \frac{K(\zeta_1, k_1; \zeta_2, k_2;

784: \zeta_3, k_3; \zeta_4, k_4)}{\alpha_{12} \alpha_{14}} \ , \

785: \nonumber \\

786: A_{F^4}(1,2,3,4) & = & - \ 2 \ g^2 \ K(\zeta_1, k_1; \zeta_2, k_2;

787: \zeta_3, k_3; \zeta_4, k_4) \ , \label{YMF4K}

788: \end{eqnarray}

789: we have that

790: \begin{eqnarray}

791: A(1,2,3,4,5)_{k_5 \rightarrow 0} & \sim & 2 g^3  \left(\frac{}{}

792: \frac{\zeta_5 \cdot k_1}{k_5 \cdot k_1} - \frac{\zeta_5 \cdot

793: k_4}{k_5 \cdot k_4} \frac{}{} \right) \left(\frac{}{}

794: \frac{1}{\alpha_{12} \alpha_{14}} T \ - \ (2 \alpha')^2 K_3

795: \frac{}{}

796: \right)\biggl|_{k_5=0} \times \nonumber \\

797: && \times \ K(\zeta_1, k_1; \zeta_2, k_2; \zeta_3, k_3; \zeta_4,

798: k_4) \ .

799: \label{almostsoft}

800: \end{eqnarray}

801: Now, if $k_5=0$ this implies that $\alpha_{45}=\alpha_{51}=0$ and

802: $\alpha_{14}=\alpha_{23}$ (see eq. (\ref{alpha14})), so we have that

803: \begin{eqnarray}

804: A(1,2,3,4,5)_{k_5 \rightarrow 0} & \sim & 2 g^3  \left(\frac{}{}

805: \frac{\zeta_5 \cdot k_1}{k_5 \cdot k_1} - \frac{\zeta_5 \cdot

806: k_4}{k_5 \cdot k_4} \frac{}{} \right) \left(\frac{}{}

807: \frac{1}{\alpha_{12} \alpha_{23}} T \ - \ (2 \alpha')^2 K_3

808: \frac{}{}

809: \right)\biggl|_{\alpha_{45}=\alpha_{51}=0} \times \nonumber \\

810: && \times \ K(\zeta_1, k_1; \zeta_2, k_2; \zeta_3, k_3; \zeta_4,

811: k_4) \ .

812: \label{almostsoft2}

813: \end{eqnarray}

814: In this last relation, the expression evaluated in

815: $\alpha_{45}=\alpha_{51}=0$ is basically the Gamma factor of the

816: 4-point amplitude evaluated in $\alpha_{12}$ and $\alpha_{23}$ (see

817: eq. (\ref{relate3})), which finally allows us to write

818: \begin{eqnarray}

819: A(1,2,3,4,5)_{k_5 \rightarrow 0} \sim \ g \left(\frac{}{}

820: \frac{\zeta_5 \cdot k_1}{k_5 \cdot k_1} - \frac{\zeta_5 \cdot

821: k_4}{k_5 \cdot k_4} \frac{}{} \right) \ A(1,2,3,4) \ .

822: \label{softA12345}

823: \end{eqnarray}

824: \item{\underline{Factorization of collinear poles}:}\\

825: \noindent This corresponds to the case of two consecutive bosons, say the first

826: and the second one, that become parallel. The Yang-Mills subamplitude

827: factorization, in this case, has the following form \cite{Mangano1}:

828: \begin{multline}

829: A_{YM}(\zeta_1, k_1;\zeta_2, k_2;\zeta_3, k_3;\zeta_4, k_4;\zeta_5, k_5)_{k_1 || k_2} \sim \\

830: \begin{split}

831: & \frac{1}{2(k_1 \cdot k_2)} \ V^{\mu} \frac{\partial}{\partial \zeta^{\mu}}

832: A_{YM}(\zeta, k_1+k_2;\zeta_3, k_3;\zeta_4, k_4;\zeta_5, k_5) \ ,

833: \end{split}

834: \label{collinearYM}

835: \end{multline}

836: where

837: \begin{eqnarray}

838: V^{\mu} = - g \biggl[(\zeta_1 \cdot \zeta_2) (k_1-k_2)^{\mu}-

839: 2(\zeta_2 \cdot k_1) {\zeta_1}^{\mu}+2(\zeta_1 \cdot k_2)

840: {\zeta_2}^{\mu} \biggr] \ \label{Vmu}

841: \end{eqnarray}

842: comes from the Yang-Mills 3-point vertex given in

843: (\ref{YMvertex3}):

844: \begin{eqnarray}

845: V_{\mu}= -i \ g \ V^{(3)}_{YM \ \mu_1 \mu_2 \mu}(k_1, k_2, -k_1

846: -k_2) \ \zeta_1^{\mu_1}\zeta_2^{\mu_2} \ . \label{vertex3}

847: \end{eqnarray}

848: Notice that the subamplitude on the left handside of

849: (\ref{collinearYM}) is a 5-point subamplitude while the one on the

850: right handside

851: is a 4-point one.\\

852: \noindent Now, when considering the $k_1 || k_2$ limit in

853: (\ref{AF412345}) the dominant term is precisely the one which has

854: the denominator $\alpha_{12}$ (recall that $k_1 \cdot k_2 = 0$ in

855: this limit, due to the on-shell condition). The three terms in the

856: residue of $\alpha_{12}$ in (\ref{AF412345}) can be seen to match

857: with the corresponding ones of $V^{\mu}$, when substituted in

858: (\ref{Vmu}), so $A_{F^4}(1,2,3,4,5)$ also

859: satisfies the factorization in (\ref{collinearYM}). \\

860: \noindent So, in the same way as was done in the previous item, taking the

861: $k_1 || k_2$ limit in (\ref{A12345final}) leads to

862: \begin{multline}

863: A(\zeta_1, k_1;\zeta_2, k_2;\zeta_3, k_3;\zeta_4, k_4;\zeta_5, k_5)_{k_1 || k_2} \sim \\

864: \begin{split}

865: & \frac{1}{2(k_1 \cdot k_2)} \ V^{\mu} \frac{\partial}{\partial \zeta^{\mu}} \biggl[

866: 2 \ g^2 \ K(\zeta, k_1+k_2;\zeta_3, k_3;\zeta_4, k_4;\zeta_5, k_5) \biggr] \ \times \\

867: & \times \left(\frac{}{} \frac{T}{[(k_1+k_2) \cdot k_3] \

868: [(k_1+k_2) \cdot k_5]} \ - \ (2 \alpha')^2 K_3 \frac{}{}

869: \right)\biggl|_{\alpha_{12}=0} \ .

870: \end{split}

871: \label{collinearA1}

872: \end{multline}

873: Using relations (\ref{alpha13}) and (\ref{alpha25}) it can easily be seen that

874: $(k_1+k_2) \cdot k_3 = \alpha_{45}$ and $(k_1+k_2) \cdot k_5=\alpha_{34}$ when

875: $\alpha_{12}=0$, so the factor in the third line in (\ref{collinearA1}) becomes,

876: once more, the Gamma factor of the 4-point subamplitude (see (\ref{relate3})).

877: So we finally have that:

878: \begin{multline}

879: A(\zeta_1, k_1;\zeta_2, k_2;\zeta_3, k_3;\zeta_4, k_4;\zeta_5, k_5)_{k_1 || k_2} \sim \\

880: \begin{split}

881: & \frac{1}{2(k_1 \cdot k_2)} \ V^{\mu} \frac{\partial}{\partial \zeta^{\mu}}

882:  A(\zeta, k_1+k_2;\zeta_3, k_3;\zeta_4, k_4;\zeta_5, k_5) \ .

883: \end{split}

884: \label{collinearA3}

885: \end{multline}

886: \end{itemize}

887: \end{enumerate}

888: 

889: \section{5-field terms in the low energy effective lagrangian}

890: \label{lagrangian}

891: In the case of the $D^{2n}F^4$ terms

892: lagrangian, ${\cal L}_{D^{2n}F^4}$, given in

893: (\ref{Seff-nonabelian}), its determination was quite direct from

894: the corresponding scattering subamplitude (\ref{A1234}).

895: Unfortunately, the same procedure does not work in the case of

896: ${\cal L}_{D^{2n}F^5}$: for each of the two terms in the 5-point

897: subamplitude in (\ref{A12345final}), there does not exist a local

898: lagrangian, at each order in $\alpha'$, which reproduces the

899: corresponding subamplitude\footnote{Only for the first term in

900: (\ref{A12345final}), at order $0$ in $\alpha'$, it is possible to

901: find a local lagrangian, namely, the Yang-Mills lagrangian.}.

902: There only exists an effective lagrangian which reproduces the

903: \underline{sum} of the two terms in (\ref{A12345final}).  This can

904: be argued by noticing that the $\alpha'$ expansion for each of

905: those terms does not satisfy the factorizability property but, as

906:  it was seen in (\ref{softA12345}) and (\ref{collinearA3}), this

907:  property is indeed satisfied if \underline{both} terms are

908:  included.\\

909: \noindent In this section we deal with the main result of this

910: work, namely, the 5-field terms in the effective lagrangian. By

911: this we mean the explicit determination of the $D^{2n}F^5$

912: terms\footnote{The $D^{2n}F^4$ terms also contain 5-field terms,

913: but in the present context we refer to those which explicitly

914: contain 5 field strengths and covariant derivatives of them.} by

915: means of the 5-point subamplitude (\ref{A12345final}). These terms

916: do not contribute to the \underline{abelian} effective lagrangian:

917: in the abelian limit all of them become zero\footnote{In the case

918: of an odd number of strings, the abelian limit of the scattering

919: amplitude (\ref{general-amplitude}) gives a null result and,

920: therefore, no interaction term for them goes in the effective

921: lagrangian. This happens due to the world-sheet parity

922: antisymmetry, (\ref{twisting}),

923: of the string subamplitudes.}.\\

924: \noindent The determination of the $D^{2n}F^5$ terms, by means of

925: the complete 5-point subamplitude, is by far more complicated than

926: the corresponding one of the $D^{2n}F^4$ terms \cite{Chandia1}

927: given in (\ref{Seff-nonabelian}). This happens because, at every

928: order in $\alpha'$, say ${\alpha'}^{k+3}$ ($k=0, 1, 2, \ldots $),

929: the 5-point subamplitude receives contributions not only from the

930: $D^{2k}F^5$ terms, but also from the $D^{2k+2}F^4$ ones. The

931: contribution of the first type of terms has no poles, but the

932: contribution of the second type of

933: terms contains poles and also regular terms. \\

934: \noindent In the next subsection we first write the final

935: expression for the $D^{2n}F^5$ terms scattering subamplitude (and

936: leave the details of its derivation to appendix \ref{derivation}).

937: In the second subsection we write the desired lagrangian terms in

938: short notation (using tensors and $\alpha'$ dependent functions)

939: and in the last one we give explicit examples of the $D^{2n}F^5$

940: and the $D^{2n}F^4$ terms up to ${\cal O}({\alpha'}^4)$ order.

941: 

942: 

943: 

944: \subsection{The $D^{2n}F^5$ terms 5-point subamplitude}

945: \label{D2nF5}

946: 

947: From (\ref{effective1}) we have that the

948: contribution of the $D^{2n}F^5$ terms to the open superstring

949: 5-point subamplitude is obtained as

950: \begin{eqnarray}

951: A_{D^{2n}F^5}(1,2,3,4,5) = A(1,2,3,4,5) - A_{YM}(1,2,3,4,5) -

952: A_{D^{2n}F^4}(1,2,3,4,5) \ ,

953: \label{AD2nF5}

954: \end{eqnarray}

955: where the $D^{2n}F^4$ terms 5-point subamplitude,

956: $A_{D^{2n}F^4}(1,2,3,4,5)$, is given in (\ref{AD2nF4}). In appendix

957: \ref{derivation} we have worked with the expression in

958: (\ref{AD2nF5}), checking out that all poles cancel (as expected),

959: arriving to the following expression:

960: \begin{eqnarray}

961: A_{D^{2n}F^5}(1,2,3,4,5) &=& g^3 \ \biggl\{ \frac{}{} \biggl[

962: H^{(1)} \cdot h^{(1)}(\zeta,k) + P^{(1)}

963: \cdot p^{(1)}(\zeta,k)+\nonumber \\

964: &&\hphantom{\biggl\{ \frac{}{} \biggl[} + U^{(1)} \cdot

965: u^{(1)}(\zeta,k) + W^{(1)} \cdot w^{(1)}(\zeta,k)+  Z^{(1)} \cdot

966: z^{(1)}(\zeta,k) \biggr] +

967: \nonumber \\

968: &&\hphantom{\biggl\{ \frac{}{} \biggl[} + (\frac{}{} \mbox{cyclic

969: permutations} \frac{}{}) \frac{}{} \biggr\} \ + g^3 \ \Delta \cdot

970: \delta(\zeta,k) , \ \

971: \label{AD2nF5nopoles}

972: \end{eqnarray}

973: where $H^{(1)}$, $P^{(1)}$, $U^{(1)}$, $W^{(1)}$, $Z^{(1)}$ and

974: $\Delta$ are $\alpha'$ dependent factors (obtained from the Gamma,

975: the $T$ and the $K_3$ factors) which are given in appendix

976: \ref{otherfactors}, while $h^{(1)}(\zeta,k)$, $p^{(1)}(\zeta,k)$,

977: $u^{(1)}(\zeta,k)$, $w^{(1)}(\zeta,k)$, $z^{(1)}(\zeta,k)$ and

978: $\delta(\zeta,k)$ are kinematical expressions (with no poles) which

979: depend on the polarizations $\zeta_i$ and momenta $k_i$

980: ($i=1,2,3,4,5$). There is no summation over cyclic permutations of

981: the term `$g^3 \ \Delta \cdot \delta(\zeta,k)$' because $\Delta$ and

982: $\delta(\zeta,k)$ are

983: already cyclic invariant expressions.\\

984: \noindent The kinematical expressions in (\ref{AD2nF5nopoles}) are

985: given by

986: \begin{eqnarray}

987: \label{h23} h^{(1)}(\zeta,k) &=& t_{(10)}^{\mu_1 \nu_1 \mu_2 \nu_2

988: \mu_3 \nu_3 \mu_4 \nu_4 \mu_5

989: \nu_5}\zeta^1_{\mu_1}k^1_{\nu_1}\zeta^2_{\mu_2}k^2_{\nu_2}

990: \zeta^3_{\mu_3}k^3_{\nu_3}\zeta^4_{\mu_4}k^4_{\nu_4}

991: \zeta^5_{\mu_5}k^5_{\nu_5} \ ,

992: \\

993: \label{p34} p^{(1)}(\zeta,k) &=& 2 \ {(\eta \cdot

994: t_{(8)})_1}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5

995: \nu_5}\zeta^1_{\mu_1}k^1_{\nu_1}\zeta^2_{\mu_2}k^2_{\nu_2}

996: \zeta^3_{\mu_3}k^3_{\nu_3}\zeta^4_{\mu_4}k^4_{\nu_4}

997: \zeta^5_{\mu_5}k^5_{\nu_5} \ ,

998: \\

999: \label{u34} u^{(1)}(\zeta ,k) &=&\biggl[ \frac{1}{2} \

1000: t_{(10)}^{\mu _{1}\nu _{1}\mu _{2}\nu _{2}\mu _{3}\nu _{3}\mu

1001: _{4}\nu _{4}\mu _{5}\nu _{5}}\alpha _{34} + 2 \ t_{(8)}^{\mu

1002: _{4}\nu _{4}\mu _{5}\nu _{5}\mu _{1}\nu _{1}\mu _{2}\nu _{2}}

1003: k_{1}^{[\mu _{3}}k_{2}^{\nu _{3}]} + 2 \ t_{(8)}^{\mu _{3}\nu

1004: _{3}\mu _{4}\nu _{4}\mu _{5}\nu _{5}\mu

1005: _{1}\nu _{1}} k_{1}^{[\mu _{2}}k_{3}^{\nu _{2}]} - \notag \\

1006: &&- 2 \ t_{(8)}^{\mu _{1}\nu _{1}\mu _{2}\nu _{2}\mu _{3}\nu

1007: _{3}\mu _{4}\nu _{4}} k_{1}^{[\mu _{5}}k_{4}^{\nu _{5}]} - 2 \

1008: t_{(8)}^{\mu _{5}\nu _{5}\mu _{1}\nu _{1}\mu _{2}\nu _{2}\mu

1009: _{3}\nu _{3}}

1010: k_{1}^{[\mu _{4}}k_{5}^{\nu _{4}]} \biggr] \times \nonumber \\

1011: && \times \ \zeta _{\mu _{1}}^{1}k_{\nu _{1}}^{1}\zeta _{\mu

1012: _{2}}^{2}k_{\nu _{2}}^{2}\zeta _{\mu _{3}}^{3}k_{\nu _{3}}^{3}\zeta

1013: _{\mu _{4}}^{4}k_{\nu _{4}}^{4}\zeta _{\mu _{5}}^{5}k_{\nu _{5}}^{5}

1014: \ ,

1015: \\

1016: \label{w51} w^{(1)}(\zeta ,k) &=&4 \ t_{(8)}^{\mu _{2}\nu _{2}\mu

1017: _{3}\nu _{3}\mu _{4}\nu _{4}\mu _{5}\nu _{5}} \zeta _{\mu

1018: _{1}}^{1}k_{\nu _{1}}^{1} \zeta _{\mu _{2}}^{2}k_{\nu

1019: _{2}}^{2}\zeta _{\mu _{3}}^{3}k_{\nu _{3}}^{3}\zeta _{\mu

1020: _{4}}^{4}k_{\nu _{4}}^{4}\zeta _{\mu _{5}}^{5}k_{\nu _{5}}^{5}

1021:  k_{2}^{[\mu _{1}}k_{5}^{\nu

1022: _{1}]}  \ ,

1023: \\

1024: \label{z23} z^{(1)}(\zeta,k) &=& 4 \ t_{(8)}^{\mu_2 \nu_2 \mu_3

1025: \nu_3 \mu_4 \nu_4 \mu_5 \nu_5}\zeta^1_{\mu_1}k^1_{\nu_1}

1026: \zeta^2_{\mu_2}k^2_{\nu_2}\zeta^3_{\mu_3}k^3_{\nu_3}

1027: \zeta^4_{\mu_4}k^4_{\nu_4}\zeta^5_{\mu_5}k^5_{\nu_5} \biggl(

1028: k_{3}^{[\mu _{1}}k_{4}^{\nu _{1}]}-k_{2}^{[\mu_{1}}k_{5}^{\nu

1029: _{1}]} \biggr) \ ,

1030: \\

1031: \label{delta}

1032: \delta(\zeta,k) &=& \frac{1}{5}\biggl[ t_{(10)}^{\mu

1033: _{1}\nu _{1}\mu _{2}\nu _{2}\mu _{3}\nu _{3}\mu _{4}\nu _{4}\mu

1034: _{5}\nu _{5}}\alpha _{51}\alpha _{12}

1035:  + 4 {(\eta \cdot t_{(8)})_1}^{\mu _{1}\nu _{1}\mu _{2}\nu

1036: _{2}\mu _{3}\nu _{3}\mu _{4}\nu _{4}\mu _{5}\nu _{5}}\alpha

1037: _{23}\alpha _{45} \biggr] \times \nonumber \\

1038: && \times \ \zeta _{\mu _{1}}^{1}k_{\nu _{1}}^{1}\zeta _{\mu

1039: _{2}}^{2}k_{\nu _{2}}^{2}\zeta _{\mu _{3}}^{3}k_{\nu

1040: _{3}}^{3}\zeta _{\mu _{4}}^{4}k_{\nu _{4}}^{4}\zeta _{\mu

1041: _{5}}^{5}k_{\nu _{5}}^{5} + (\frac{}{} \mbox{cyclic permutations}

1042: \frac{}{}) \ .

1043: \end{eqnarray}

1044: The expression for the $(\eta \cdot t_{(8)})_1$ tensor of

1045: (\ref{p34}) and (\ref{delta}) is given in equation (\ref{etat8})

1046: and the expression for the $t_{(10)}$ tensor is

1047: given in appendix \ref{t10tensor} . \\

1048: \noindent The expression for $A_{D^{2n}F^5}(1,2,3,4,5)$ in

1049: (\ref{AD2nF5nopoles}) is such that, besides containing no poles and

1050: being cyclic invariant, satisfies (on-shell) gauge invariance and

1051: world-sheet parity\footnote{The upper index $(1)$ in all factors

1052: $H^{(1)}$, $P^{(1)}$, $\ldots$ , denotes that they are invariant

1053: under a twisting transformation with respect to index $1$.

1054: Similarly, the upper index $(1)$ in all kinematical expressions

1055: $h^{(1)}(\zeta,k)$, $p^{(1)}(\zeta,k)$, $\ldots$ , denotes that they

1056: change their sign under a twisting transformation with respect to

1057: index $1$.} on $each$ group of terms, that is, on $[ H^{(1)} \cdot

1058: h^{(1)}(\zeta,k) + (\mbox{cyclic permutations}) ]$, $\ldots$, $[

1059: Z^{(1)} \cdot z^{(1)}(\zeta,k) + (\mbox{cyclic permutations}) ]$ and

1060: $\Delta \cdot \delta(\zeta,k)$, separately. These conditions are

1061: enough to find a local lagrangian for each of those terms, at each

1062: order in $\alpha'$.

1063: 

1064: 

1065: 

1066: \subsection{The main formula}

1067: \label{main}

1068: 

1069: From formula (\ref{AD2nF5nopoles}), for

1070: $A_{D^{2n}F^5}(1,2,3,4,5)$, and the kinematical expressions

1071: (\ref{h23})-(\ref{delta}), the following effective lagrangian can

1072: be obtained for the $D^{2n}F^5$ terms, up to terms which are

1073: sensible to 6 or higher-point amplitudes\footnote{When considering

1074: the $\alpha'$ expansion of eq. (\ref{LD2nF5final}), whenever two

1075: covariant derivatives $D_{\alpha}$ and $D_{\beta}$ (with $\alpha

1076: \neq \beta$) operate on a $same$ field strength $F_{\mu \nu}(x)$,

1077: the order in which they operate does not matter since the

1078: difference between the two possibilities will be sensible only to

1079: 6 or higher-point amplitudes.}:

1080: \begin{multline*}

1081: {\cal L}_{D^{2n}F^5}  =  \ i \ g^3 \ \int \int \int \int \int

1082: \left\{ \frac{}{}\prod_{j=1}^{5} d^{10} x_j \ \delta^{(10)}(x-x_j)

1083: \frac{}{} \right\} \times \\

1084: \begin{split}

1085: & \times  \ \biggl[ \ \biggl\{ \frac{1}{32} H^{(1)} ( - D_{2} \cdot

1086: D_{3}, - D_{3} \cdot D_{4} \ , - D_{4} \cdot D_{5} ) \

1087: t_{(10)}^{\mu_1 \nu_1 \mu_2 \nu_2

1088: \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} + \\

1089: &+ \frac{1}{16} P^{(1)} ( - D_{1} \cdot D_{2}, - D_{2} \cdot D_{3},

1090: - D_{3} \cdot D_{4} \ , - D_{4} \cdot D_{5}, - D_{5} \cdot D_{1} ) \

1091: {(\eta \cdot t_{(8)})_1}^{\mu_1 \nu_1 \mu_2 \nu_2

1092: \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} \biggr\} \times \\

1093: &\times \tr\biggl( F_{\mu_1 \nu_1}(x_1)F_{\mu_2 \nu_2}(x_2)F_{\mu_3

1094: \nu_3}(x_3)F_{\mu_4 \nu_4}(x_4)F_{\mu_5 \nu_5}(x_5) \biggr) \ \ -

1095: \\

1096: & - \ \ U^{(1)} ( - D_{1} \cdot D_{2}, - D_{2} \cdot D_{3}, - D_{4}

1097: \cdot

1098: D_{5}, - D_{5} \cdot D_{1} ) \times \\

1099: & \times \biggl\{\frac{1}{64} t_{(10)}^{\mu_1 \nu_1 \mu_2 \nu_2

1100: \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} \tr\biggl( F_{\mu_1

1101: \nu_1}(x_1)F_{\mu_2 \nu_2}(x_2)D^{\alpha}F_{\mu_3

1102: \nu_3}(x_3)D_{\alpha} F_{\mu_4 \nu_4}(x_4)F_{\mu_5 \nu_5}(x_5)

1103: \biggr) + \\

1104: & + \frac{1}{16} t_{(8)}^{\mu_4 \nu_4 \mu_5 \nu_5 \mu_1 \nu_1 \mu_2

1105: \nu_2} \tr\biggl( D^{\mu_3}F_{\mu_1 \nu_1}(x_1)D^{\nu_3}F_{\mu_2

1106: \nu_2}(x_2)F_{\mu_3 \nu_3}(x_3)F_{\mu_4 \nu_4}(x_4)F_{\mu_5

1107: \nu_5}(x_5) \biggr)+\\

1108: & + \frac{1}{16} t_{(8)}^{\mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5 \mu_1

1109: \nu_1} \tr\biggl( D^{\mu_2}F_{\mu_1 \nu_1}(x_1)F_{\mu_2

1110: \nu_2}(x_2)D^{\nu_2}F_{\mu_3 \nu_3}(x_3)F_{\mu_4 \nu_4}(x_4)F_{\mu_5

1111: \nu_5}(x_5) \biggr)- \\

1112: & - \frac{1}{16} t_{(8)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4

1113: \nu_4} \tr\biggl( D^{\mu_5}F_{\mu_1 \nu_1}(x_1)F_{\mu_2

1114: \nu_2}(x_2)F_{\mu_3 \nu_3}(x_3)D^{\nu_5}F_{\mu_4 \nu_4}(x_4)F_{\mu_5

1115: \nu_5}(x_5) \biggr)- \\

1116: & - \frac{1}{16} t_{(8)}^{\mu_5 \nu_5 \mu_1 \nu_1 \mu_2 \nu_2

1117: \mu_3 \nu_3} \tr\biggl( D^{\mu_4}F_{\mu_1 \nu_1}(x_1)F_{\mu_2

1118: \nu_2}(x_2)F_{\mu_3 \nu_3}(x_3)F_{\mu_4

1119: \nu_4}(x_4)D^{\nu_4}F_{\mu_5 \nu_5}(x_5) \biggr) \ \biggr\} \ \

1120: -\\

1121:  & - \ \ \frac{1}{8} W^{(1)} ( - D_{1} \cdot D_{2}, - D_{2} \cdot

1122: D_{3}, - D_{3} \cdot D_{4} \ , - D_{4} \cdot D_{5}, - D_{5} \cdot

1123: D_{1} )

1124: \times \\

1125: & \times t_{(8)}^{\mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5}

1126: \tr\biggl( F_{\mu_1 \nu_1}(x_1)D^{\mu_1}F_{\mu_2

1127: \nu_2}(x_2)F_{\mu_3 \nu_3}(x_3)F_{\mu_4

1128: \nu_4}(x_4)D^{\nu_1}F_{\mu_5 \nu_5}(x_5) \biggr) \ \ -

1129: \end{split}

1130: \end{multline*}

1131: \begin{multline}

1132: \begin{split}

1133:  & - \ \

1134: \frac{1}{8} Z^{(1)} ( - D_{1} \cdot D_{2}, - D_{2} \cdot D_{3}, -

1135: D_{3} \cdot D_{4} \ , - D_{4} \cdot D_{5}, - D_{5} \cdot D_{1} )

1136: \times \\

1137: & \times t_{(8)}^{\mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5}

1138: \biggl\{ \tr\biggl( F_{\mu_1 \nu_1}(x_1)F_{\mu_2

1139: \nu_2}(x_2)D^{\mu_1}F_{\mu_3 \nu_3}(x_3)D^{\nu_1}F_{\mu_4

1140: \nu_4}(x_4)F_{\mu_5 \nu_5}(x_5) \biggr)

1141: - \\

1142: & \hphantom{\times t_{(8)}^{\mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4

1143: \mu_5 \nu_5} } - \tr\biggl( F_{\mu_1 \nu_1}(x_1)D^{\mu_1}F_{\mu_2

1144: \nu_2}(x_2)F_{\mu_3 \nu_3}(x_3)F_{\mu_4 \nu_4}(x_4)D^{\nu_1}F_{\mu_5

1145: \nu_5}(x_5) \biggr) \biggr\} \ \ + \\

1146: & + \ \ \frac{1}{160} \Delta( - D_{1} \cdot D_{2}, - D_{2} \cdot

1147: D_{3}, - D_{3}

1148: \cdot D_{4} \ , - D_{4} \cdot D_{5}, - D_{5} \cdot D_{1} ) \times \\

1149: & \times \biggl\{ t_{(10)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3

1150: \mu_4 \nu_4 \mu_5 \nu_5}  \tr\biggl( D^{\alpha} D^{\beta} F_{\mu_1

1151: \nu_1}(x_1)D_{\alpha}F_{\mu_2 \nu_2}(x_2)F_{\mu_3

1152: \nu_3}(x_3)F_{\mu_4

1153: \nu_4}(x_4)D_{\beta}F_{\mu_5 \nu_5}(x_5) \biggr) + \\

1154: & + 4 {(\eta \cdot t_{(8)})_1}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3

1155: \mu_4 \nu_4 \mu_5 \nu_5}  \times \\

1156: & \hphantom{ + 4 }  \times \tr\biggl( F_{\mu_1

1157: \nu_1}(x_1)D_{\alpha}F_{\mu_2 \nu_2}(x_2)D^{\alpha} F_{\mu_3

1158: \nu_3}(x_3)D_{\beta} F_{\mu_4 \nu_4}(x_4)D^{\beta}F_{\mu_5

1159: \nu_5}(x_5) \biggr) \ \biggr\} \ \ \ \biggr] \ .

1160: \end{split}

1161: \label{LD2nF5final}

1162: \end{multline}

1163: Notice that in this result we have used that the factors $P^{(1)}$,

1164: $W^{(1)}$, $Z^{(1)}$ and $\Delta$ depend in $all$ five (independent)

1165: $\alpha_{ij}$ variables, while $H^{(1)}$ and $U^{(1)}$ depend on a

1166: $lower$ number of them:

1167: \begin{eqnarray}

1168: \label{explicitH1} H^{(1)} & = & H^{(1)}(\alpha_{23}, \alpha_{34},

1169: \alpha_{45}) \ ,

1170: \\

1171: \label{explicitU1} U^{(1)} & = & U^{(1)}(\alpha_{12}, \alpha_{23},

1172: \alpha_{45}, \alpha_{51}) \ ,

1173: \end{eqnarray}

1174: as may be seen directly in the formulas of appendix

1175: \ref{otherfactors}.\\

1176: \noindent That formula (\ref{LD2nF5final}) is indeed correct may

1177: be confirmed by calculating the 1 PI 5-point function of it,

1178: $\Gamma_{\mu_1 \mu_2 \mu_3 \mu_4 \mu_5}^{(5) \ a_1 a_2 a_3 a_4

1179: a_5}(x_1, x_2, x_3, x_4, x_5)$, and then calculating its Fourier

1180: Transform\footnote{In the case of the 1 PI 4-point function and

1181: the corresponding subamplitude, for example, this was done in the

1182: appendices of \cite{DeRoo1} and \cite{Chandia1}.}. From this last

1183: one it may be checked that the corresponding subamplitude is the

1184: one given in (\ref{AD2nF5nopoles}).

1185: 

1186: \subsection{Some $\alpha'$ terms}

1187: \label{some}

1188: 

1189: In this subsection we will see two applications of formula

1190: (\ref{LD2nF5final}), in order to see how it works. Before doing

1191: so, notice that the $\alpha'$ dependent factors of

1192: (\ref{AD2nF5nopoles}) and (\ref{LD2nF5final}) do not begin their

1193: power expansions at the same order in $\alpha'$. As may be seen

1194: from the explicit expansions of them in appendix

1195: \ref{otherfactors}: $H^{(1)}$ and $P^{(1)}$ begin at ${\cal

1196: O}({\alpha'}^3)$ order; $U^{(1)}$, $W^{(1)}$ and $Z^{(1)}$ begin

1197: at ${\cal O}({\alpha'}^4)$ order and $\Delta$ begins at ${\cal

1198: O}({\alpha'}^5)$ order.\\

1199: \noindent As a first application in which the 5-point amplitude is

1200: important, let us see the case of the non-abelian effective

1201: lagrangian of the open superstring at order ${\cal

1202: O}({\alpha'}^3)$ (which was first completely and correctly

1203: calculated in \cite{Koerber1}). At this order, as may be seen in

1204: (\ref{LD2nF4}) and (\ref{LD2nF5}), the lagrangian contains $F^5$

1205: and $D^2F^4$ terms. The first ones can be taken from the

1206: lagrangian in (\ref{LD2nF5final}) and the second ones from the

1207: lagrangian in (\ref{Seff-nonabelian}), giving:

1208: \begin{eqnarray}

1209: {{\cal L}_{\rm ef}}^{(3)}& = &\zeta(3) \ {\alpha'}^3 \biggl[ \ - i

1210: \frac{ \ g^3}{4} \biggl \{ t_{(10)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3

1211: \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} + 2 {(\eta \cdot t_{(8)})_1}^{\mu_1

1212: \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5

1213: \nu_5} \biggr\} \times \nonumber \\

1214: && \hphantom{\zeta(3) \ {\alpha'}^3 [ \ } \times \tr\biggl( F_{\mu_1

1215: \nu_1}F_{\mu_2 \nu_2}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4}F_{\mu_5 \nu_5}

1216: \biggr) \ + \frac{ \ g^2}{2}

1217: t_{(8)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4} \times \nonumber \\

1218: && \hphantom{\zeta(3) \ {\alpha'}^3 [ \ } \times \biggl\{

1219: \tr\biggl( D^{\alpha}F_{\mu_1 \nu_1}D_{\alpha} F_{\mu_2

1220: \nu_2}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4} \biggr) + \tr\biggl(

1221: D^2F_{\mu_1 \nu_1}F_{\mu_2 \nu_2}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4}

1222: \biggr)

1223: \biggr\} \ \biggr] \ . \nonumber \\

1224: \label{L3}

1225: \end{eqnarray}

1226: In order to compare (\ref{L3}) with one of the expressions of the

1227: known result we use that

1228: \begin{eqnarray}

1229: D^2F_{\mu \nu} =  D^{\alpha}(D_{\alpha} F_{\mu \nu}) =

1230: D_{\mu}(D_{\alpha} {F^{\alpha}}_{\nu}) - D_{\nu}(D_{\alpha}

1231: {F^{\alpha}}_{\mu}) + 2 \ i \ g \ [ F_{\alpha \mu},

1232: {F_{\nu}}^{\alpha} ] \ , \label{D2}

1233: \end{eqnarray}

1234: which may be derived using the Bianchi identity and the $[D,D]F =

1235: -ig[F,F]$ relation (see (\ref{Bianchi}) and (\ref{group4}),

1236: respectively).\\

1237: Substituting (\ref{D2}) in (\ref{L3}), and dropping out on-shell

1238: terms (i.e. the ones which contain $D_{\alpha}

1239: {F^{\alpha}}_{\mu}$)\footnote{In \cite{Tseytlin1} it was seen, at

1240: least up to ${\cal O}({\alpha'}^2)$ order, that the on-shell terms

1241: in the effective lagrangian are not sensible to on-shell

1242: scattering amplitudes.} leads to

1243: \begin{eqnarray}

1244: {{\cal L}_{\rm ef}}^{(3)}& = &\zeta(3) \ {\alpha'}^3 \biggl[ \ - i

1245: \frac{ \ g^3}{4}  ( t_{(10)}+4 (\eta \cdot t_{(8)})_1)^{\mu_1

1246: \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} \tr\biggl(

1247: F_{\mu_1 \nu_1}F_{\mu_2 \nu_2}F_{\mu_3

1248: \nu_3}F_{\mu_4 \nu_4}F_{\mu_5 \nu_5} \biggr) \ +  \nonumber \\

1249: && \hphantom{\zeta(3) \ {\alpha'}^3 [ \ } + \frac{ \ g^2}{2}

1250: t_{(8)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4}

1251: \tr\biggl( D^{\alpha}F_{\mu_1 \nu_1}D_{\alpha} F_{\mu_2

1252: \nu_2}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4} \biggr) \ \biggr] \ .

1253: \label{newL3}

1254: \end{eqnarray}

1255: \noindent Using the explicit expressions of the $t_{(8)}$ and the

1256: $t_{(10)}$ tensors, and a basis of $F^5$ and $D^2F^4$ terms taken

1257: from \cite{Grasso2}\footnote{See appendix B of this reference.},

1258: it may be verified that the $D^2F^4$ terms of (\ref{newL3}) agree

1259: with the ones in eq. (4.17) of \cite{Grasso1} and that the $F^5$

1260: terms of (\ref{newL3}) agree, on-shell, with the ones of the

1261: mentioned equation after going to $D=4$\footnote{In $D=4$, due to

1262: the Cayley-Hamilton theorem, there is an identity involving the

1263: $F^5$ terms which reduces the number of them which are

1264: independent. That is the reason of why our $F^5$ terms in

1265: (\ref{newL3}) only agree with the ones which go in the effective

1266: action of $N=4$ SYM in \cite{Grasso1} after going to $4$

1267: dimensions. We thank D. Grasso for explaining this to us.}. In

1268: this last reference it was seen that the lagrangian in equation

1269: (4.17) of it agrees (in $D=4$) with the one in \cite{Koerber1},

1270: which was also confirmed in \cite{DeRoo2} and

1271: \cite{Brandt1}.\\

1272: \noindent The next interesting example consists in the ${\cal

1273: O}({\alpha'}^4)$ terms of the effective lagrangian which are

1274: sensible to 5-point amplitudes. We take the $D^2F^5$ terms from

1275: (\ref{LD2nF5final}) and the $D^4F^4$ ones from the lagrangian in

1276: (\ref{Seff-nonabelian}), giving:

1277: \begin{eqnarray}

1278: {\cal L'}^{(4)}& = & {\cal L}_{\rm D^2F^5} + {\cal L}_{\rm D^4F^4}

1279: \ , \label{L4}

1280: \end{eqnarray}

1281: where

1282: \begin{eqnarray}

1283: {\cal L}_{\rm D^2F^5} & = &  i \ \frac{\ \pi^4}{90} \ g^3 \

1284: {\alpha'}^4 \biggl[ \ \ - \frac{1}{8} t_{(10)}^{\mu_1 \nu_1 \mu_2

1285: \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} \biggl \{ 4 \tr\biggl(

1286: F_{\mu_1 \nu_1}F_{\mu_2 \nu_2}F_{\mu_3 \nu_3}D^{\alpha}F_{\mu_4

1287: \nu_4}D_{\alpha}F_{\mu_5 \nu_5} \biggr) +

1288: \nonumber \\

1289: && + 4 \tr\biggl( F_{\mu_1 \nu_1}D^{\alpha}F_{\mu_2

1290: \nu_2}D_{\alpha}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4}F_{\mu_5 \nu_5}

1291: \biggr) + \tr\biggl( F_{\mu_1 \nu_1}F_{\mu_2

1292: \nu_2}D^{\alpha}F_{\mu_3 \nu_3}D_{\alpha}F_{\mu_4 \nu_4}F_{\mu_5

1293: \nu_5} \biggr) \biggr\} -

1294: \nonumber \\

1295: && - \frac{1}{4} {(\eta \cdot t_{(8)})_1}^{\mu_1 \nu_1 \mu_2 \nu_2

1296: \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} \biggl\{ 4 \tr\biggl(

1297: F_{\mu_1 \nu_1}F_{\mu_2 \nu_2}D^{\alpha}F_{\mu_3

1298: \nu_3}D_{\alpha}F_{\mu_4 \nu_4}F_{\mu_5 \nu_5} \biggr) +  \nonumber \\

1299: && + \tr\biggl( F_{\mu_1 \nu_1}F_{\mu_2 \nu_2}F_{\mu_3

1300: \nu_3}D^{\alpha}F_{\mu_4 \nu_4}D_{\alpha}F_{\mu_5 \nu_5} \biggr) +

1301: \tr\biggl( F_{\mu_1 \nu_1}D^{\alpha}F_{\mu_2

1302: \nu_2}D_{\alpha}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4}F_{\mu_5 \nu_5}

1303: \biggr) + \nonumber \\

1304: && + 4 \tr\biggl( D^{\alpha}F_{\mu_1 \nu_1}F_{\mu_2 \nu_2}F_{\mu_3

1305: \nu_3}F_{\mu_4 \nu_4}D_{\alpha}F_{\mu_5 \nu_5} \biggr) + 4

1306: \tr\biggl( D^{\alpha}F_{\mu_1 \nu_1}D_{\alpha}F_{\mu_2

1307: \nu_2}F_{\mu_3

1308: \nu_3}F_{\mu_4 \nu_4}F_{\mu_5 \nu_5} \biggr) \biggr\} \ -   \nonumber \\

1309: && -  \frac{1}{4} t_{(10)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3

1310: \mu_4 \nu_4 \mu_5 \nu_5} \tr\biggl( F_{\mu_1 \nu_1}F_{\mu_2

1311: \nu_2}D^{\alpha}F_{\mu_3 \nu_3}D_{\alpha} F_{\mu_4 \nu_4}F_{\mu_5

1312: \nu_5}

1313: \biggr) - \nonumber \\

1314: && -  t_{(8)}^{\mu_4 \nu_4 \mu_5 \nu_5 \mu_1 \nu_1 \mu_2 \nu_2}

1315: \tr\biggl( D^{\mu_3}F_{\mu_1 \nu_1}D^{\nu_3}F_{\mu_2

1316: \nu_2}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4}F_{\mu_5

1317: \nu_5} \biggr)- \nonumber \\

1318: && -  t_{(8)}^{\mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5 \mu_1 \nu_1}

1319: \tr\biggl( D^{\mu_2}F_{\mu_1 \nu_1}F_{\mu_2

1320: \nu_2}D^{\nu_2}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4}F_{\mu_5

1321: \nu_5} \biggr)+ \nonumber \\

1322: && +  t_{(8)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4}

1323: \tr\biggl( D^{\mu_5}F_{\mu_1 \nu_1}F_{\mu_2 \nu_2}F_{\mu_3

1324: \nu_3}D^{\nu_5}F_{\mu_4 \nu_4}F_{\mu_5

1325: \nu_5} \biggr)+ \nonumber \\

1326: && +  t_{(8)}^{\mu_5 \nu_5 \mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3}

1327: \tr\biggl( D^{\mu_4}F_{\mu_1 \nu_1}F_{\mu_2 \nu_2}F_{\mu_3

1328: \nu_3}F_{\mu_4 \nu_4}D^{\nu_4}F_{\mu_5

1329: \nu_5} \biggr) \  -  \nonumber \\

1330: && -  \frac{1}{2} t_{(8)}^{\mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4

1331: \mu_5 \nu_5} \tr\biggl( F_{\mu_1 \nu_1}D^{\mu_1}F_{\mu_2

1332: \nu_2}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4}D^{\nu_1}F_{\mu_5 \nu_5}

1333: \biggr) +  \nonumber

1334: \end{eqnarray}

1335: \begin{eqnarray}

1336:  && +  \frac{1}{8} t_{(8)}^{\mu_2 \nu_2 \mu_3

1337: \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} \biggl\{ \tr\biggl( F_{\mu_1

1338: \nu_1}F_{\mu_2 \nu_2}D^{\mu_1}F_{\mu_3 \nu_3}D^{\nu_1}F_{\mu_4

1339: \nu_4}F_{\mu_5

1340: \nu_5} \biggr)- \nonumber\\

1341:  && \hphantom{  +  \frac{1}{8}

1342: t_{(8)}^{\mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} \biggl\{

1343: } - \tr\biggl( F_{\mu_1 \nu_1}D^{\mu_1}F_{\mu_2 \nu_2}F_{\mu_3

1344: \nu_3}F_{\mu_4 \nu_4}D^{\nu_1}F_{\mu_5 \nu_5} \biggr) \biggr\} \ \

1345: \ \biggr] \label{LD2F5}

1346: \end{eqnarray}

1347: and

1348: \begin{eqnarray}

1349: {\cal L}_{\rm D^4F^4} & = & \frac{\ \pi^4}{2880} \ g^2 \

1350: {\alpha'}^4 \ t_{(8)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4

1351: \nu_4} \ \biggl[ \ \ \ 9 \tr\biggl( D^2 D^2F_{\mu_1 \nu_1}F_{\mu_2

1352: \nu_2}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4} \biggr) + \nonumber \\

1353: &+&  16  \tr\biggl( D^{\alpha}F_{\mu_1 \nu_1}D_{\alpha}F_{\mu_2

1354: \nu_2}D^{\beta}F_{\mu_3 \nu_3}D_{\beta}F_{\mu_4 \nu_4} \biggr) + 4

1355: \tr\biggl( D^{\alpha}F_{\mu_1 \nu_1}D_{(\beta} D_{\alpha)}F_{\mu_2

1356: \nu_2}D^{\beta}F_{\mu_3 \nu_3}F_{\mu_4

1357: \nu_4} \biggr) + \nonumber \\

1358: &+&  18 \tr\biggl( D^{2}F_{\mu_1 \nu_1}D^{2}F_{\mu_2

1359: \nu_2}F_{\mu_3 \nu_3}F_{\mu_4 \nu_4} \biggr) + 9 \tr\biggl( D^{2}

1360: F_{\mu_1 \nu_1}F_{\mu_2\nu_2}D^{2}F_{\mu_3 \nu_3}

1361: F_{\mu_4 \nu_4} \biggr) + \nonumber \\

1362: &+& 16 \tr\biggl( D^{\alpha}D^{\beta}F_{\mu_1

1363: \nu_1}D_{\alpha}D_{\beta}F_{\mu_2 \nu_2}F_{\mu_3 \nu_3}F_{\mu_4

1364: \nu_4} \biggr) + 34 \tr\biggl( D^{(2} D^{\beta)}F_{\mu_1

1365: \nu_1}D_{\beta}F_{\mu_2 \nu_2}F_{\mu_3 \nu_3}F_{\mu_4

1366: \nu_4} \biggr) + \nonumber \\

1367: &+&  34 \tr\biggl( D^{2}F_{\mu_1 \nu_1}F_{\mu_2

1368: \nu_2}D^{\beta}F_{\mu_3 \nu_3}D_{\beta}F_{\mu_4 \nu_4} \biggr) \ \

1369: \ \biggr] \ . \label{LD4F4}

1370: \end{eqnarray}

1371: At this order the symmetrized prescription mentioned in

1372: (\ref{fsym}) begins to apply: that is why some terms in ${\cal

1373: L}_{\rm D^4F^4}$ contain symmetrized products. Using the explicit

1374: expressions of the $t_{(8)}$ and the $t_{(10)}$ tensors,

1375: integration by parts, the Bianchi and the $[D,D]F=-ig[F,F]$

1376: identities, and dropping out the on-shell terms, the expressions

1377: in (\ref{LD2F5}) and (\ref{LD4F4}) could be made explicit and

1378: reduced to a minimum length,  but we will leave that work to be

1379: done somewhere else and just content ourselves with those

1380: expressions as they are, in order to see how formulas

1381: (\ref{LD2nF5final}) and (\ref{Seff-nonabelian}) have operated at

1382: this order in $\alpha'$. In any case, we have checked that the

1383: abelian limit of (\ref{LD4F4}) agrees

1384: completely with the $\partial^4F^4$ terms of \cite{Andreev1}.\\

1385: \noindent The lagrangian in (\ref{L4}) (with the expressions in

1386: (\ref{LD2F5}) and (\ref{LD4F4})) is not the complete lagrangian at

1387: ${\cal O}({\alpha'}^4)$ because the $F^6$ terms should also be

1388: present (that is why we have called it ${\cal L'}^{(4)}$ instead

1389: of ${\cal L}^{(4)}$). A complete lagrangian at this order has been

1390: determined in \cite{Koerber2} by

1391: other method and confirmed in \cite{Sevrin1} and \cite{Nagaoka1}.\\

1392: The procedure to continue writing the $D^{2n}F^5$ terms from

1393: (\ref{LD2nF5final}) and the $D^{2n}F^4$ ones from

1394: (\ref{Seff-nonabelian}) is quite direct using the expansions of

1395: all the $\alpha'$ dependent factors. In the case of the Gamma

1396: factor, the coefficients of the $\alpha'$ expansion are known up

1397: to infinity, while in this work we have expanded the $T$ and

1398: $(2\alpha')^2 K_3$ factors (from which the $P^{(1)}$, $U^{(1)}$,

1399: $W^{(1)}$ and $\Delta$ factors have been obtained) only up to

1400: ${\cal O}({\alpha'}^6)$ order. So in this section we could, in

1401: principle, go on with the $D^{2n}F^5$ terms up

1402: to that order, but to save space we will not do it here.\\

1403: In spite of what was mentioned in the previous paragraph, the

1404: structure of the $\alpha'$ expansions of $T$ and $(2\alpha')^2 K_3$

1405: is \underline{completely} known since, as it is proved in detail in

1406: appendix \ref{expansion-K3K}, at each order in $\alpha'$ only cyclic

1407: polynomial invariants (in the $\alpha_{ij}$ variables)

1408: appear\footnote{Further more, the polynomials which appear at each

1409: order in $\alpha'$ should also be invariant under a twisting

1410: transformation, so this diminishes a little the number of them which

1411: appear in the expansion. In appendix \ref{polynomial} we give a

1412: complete list of them up to sixth degree \cite{Medina1}.}. So the

1413: only unknowns are the coefficients which go with each of those

1414: cyclic polynomial invariants: this is a considerably much reduced

1415: number of unknowns than the ones which in principle should be

1416: determined at every $\alpha'$ order (see appendix \ref{polynomial}

1417: for further details).

1418: 

1419: 

1420: \section{Summary and final remarks}

1421: \label{final} \noindent We have succeeded in finding the

1422: $D^{2n}F^5$ terms of the open superstring effective lagrangian, to

1423: all order in $\alpha'$. This lagrangian is of importance in the

1424: corresponding sector of the $SO(32)$ Type I theory and in the

1425: description of the low energy interaction of D-branes. To our

1426: knowledge, this is the second $non$-$abelian$ result, reported in

1427: the literature, which is complete (in the sense that all the terms

1428: in the lagrangian and its coefficients are given in a closed form)

1429: and which has been determined to all order in $\alpha'$; the first

1430: one being the determination of the

1431: $D^{2n}F^4$ in \cite{Chandia1}.\\

1432: \noindent This has been a very much more complicated problem than

1433: the determination of the $D^{2n}F^4$ terms, which consisted in a

1434: non-abelian generalization of a previous result in \cite{DeRoo1}.

1435: We have found the general formula and we have given expansions

1436: which allow us to compute explicitly the $D^{2n}F^5$ terms up to

1437: ${\cal O}({\alpha'}^6)$ order. We have checked that the $F^5$

1438: terms of our lagrangian, together with the $D^2F^4$ ones found

1439: previously (see eq. (\ref{Seff-nonabelian})), agree with the known

1440: result at ${\cal O}({\alpha'}^3)$ order. At ${\cal

1441: O}({\alpha'}^4)$ order, it would be interesting to see if our

1442: $D^2F^5$ and $D^4F^4$ terms agree with the ones which are sensible

1443: to 4 and 5-point

1444: amplitudes in eqs. (1.5) and (1.6) of \cite{Sevrin1}.\\

1445: \noindent The starting point of all this huge labor has been the

1446: determination of the 5-point subamplitude of massless bosons in

1447: Open Superstring Theory, which we have found in terms of two

1448: kinematical expressions and two $\alpha'$ dependent factors

1449: (see eq. (\ref{A12345final})).\\

1450: \noindent Another important step that we have done, in order to be

1451: able to find all the $D^{2n}F^5$ terms, has been the determination

1452: of the scattering subamplitude $A_{D^{2n}F^5}(1,2,3,4,5)$ as a sum

1453: of terms which have no poles and have manifest cyclic and

1454: (on-shell) gauge invariance, as well as world-sheet parity

1455: symmetry (see eq. (\ref{AD2nF5nopoles})). We have done this by

1456: using the known $t_{(8)}$ and a new $t_{(10)}$ tensor in all the

1457: kinematical

1458: expressions involved in that subamplitude.\\

1459: \noindent The explicit expression of ${\cal L}_{D^{2n}F^5}$

1460: depends directly on the previously found expression of ${\cal

1461: L}_{D^{2n}F^4}$, because the scattering subamplitude

1462: $A_{D^{2n}F^5}(1,2,3,4,5)$ is calculated in terms of

1463: $A_{D^{2n}F^4}(1,2,3,4,5)$ (see eq. (\ref{AD2nF5})). Now, given

1464: that there is not a unique way in choosing the terms of ${\cal

1465: L}_{D^{2n}F^4}$\footnote{Since the only requirement for this

1466: lagrangian is to reproduce superstring 4-point amplitudes, there

1467: is some freedom in the way its covariant derivatives are present

1468: on it.}, and the fact that the $\alpha'$ factors in

1469: (\ref{AD2nF5nopoles}) are not all independent (see, for example,

1470: eq. (\ref{almostthere}) ), it may happen that a final lagrangian

1471: ${\cal L}_{D^{2n}F^5}$ could eventually be found in terms of only

1472: the $t_{(10)}$ tensor, as mentioned in \cite{Machado1}. In the

1473: present work we have not look

1474: further in this direction.\\

1475: \noindent The calculations of this work have been extremely long

1476: and, based on the experience we have had in solving them out, we

1477: think it is impossible to have done them without any computer

1478: assistance. In spite of this fact, and contrary to what it is

1479: generally believed, the main result of our paper suggests that it

1480: is indeed possible to take to the level of the effective

1481: lagrangian, to all order in $\alpha'$, the information of (tree

1482: level) superstring scattering amplitudes: at least we have

1483: succeeded on this subject,

1484: up to 5-field terms, in the case of Open Superstring Theory.\\

1485: \noindent On a future paper \cite{Barreiro1} we will use the

1486: results of the present work to determine the tree level

1487: ${\alpha'}^4R^5$ terms in the effective lagrangians of the type II

1488: theories, by means of the KLT relations \cite{Kawai1}, and we will

1489: consider the possibility of finding and $all$ $\alpha'$ result for

1490: those actions, as it was done in \cite{Chandia1} in the case of

1491: the 4-Riemann tensor terms.

1492: 

1493: 

1494: 

1495: 

1496: \section*{Acknowledgements}

1497: 

1498: R. M. would like to acknowledge useful conversations with F.

1499: Machado, with whom the first steps of this work were given. He

1500: would also like to thank A. Sevrin, N. Berkovits, F. Brandt, P.

1501: Koerber, S. Stieberger and especially D. Grasso and A. Tseytlin,

1502: for useful conversations. This research has been partially

1503: supported by the Brazilian agencies CNPq and FAPEMIG.

1504: 

1505: 

1506: \appendix

1507: 

1508: \section{Conventions and identities}

1509: \label{conventions}

1510: 

1511: \begin{enumerate}

1512: \item Metric, symmetrization and antisymmetrization over spacetime

1513: indexes:

1514: 

1515: We use the following convention for the Minkowski metric:

1516: \begin{eqnarray}

1517: \label{metric} \eta_{\mu \nu} = \mbox{diag}(-, +, \ldots , +) \ .

1518: \end{eqnarray}

1519: The symmetrization and antisymmetrization convention that we use,

1520: on the spacetime indexes of a product of two  vectors $A$ and $B$,

1521: is the following:

1522: \begin{eqnarray}

1523: \label{symmetrization} A^{(\mu}B^{\nu)}=\frac{1}{2} (

1524: A^{\mu}B^{\nu}+A^{\nu}B^{\mu}) \ , \\

1525: \label{antisymmetrization} A^{[\mu}B^{\nu]}=\frac{1}{2} (

1526: A^{\mu}B^{\nu}-A^{\nu}B^{\mu} ) \ .

1527: \end{eqnarray}

1528: 

1529: 

1530: \item Gauge group generators, field strength and covariant derivative:\\

1531: 

1532: Gauge fields are matrices in the Lie group internal space, so that

1533: $A_{\mu} = A^{\mu}_{\ a} \lambda^a$, where the $\lambda^a$ are the

1534: generators (in a matrix representation) which satisfy the usual

1535: relation

1536: \begin{eqnarray}

1537: \mbox{tr}(\lambda^a \lambda^b) = \delta^{ab} \ .\label{group1}

1538: \end{eqnarray}

1539: The field strength and the covariant derivative are defined by

1540: \begin{eqnarray}

1541: \label{group2} F_{\mu \nu} & = & \partial_{\mu} A_{\nu} -

1542: \partial_{\nu} A_{\mu}

1543: - ig [ A_{\mu}, A_{\nu}] \ , \\

1544: \label{group3} D_{\mu} \phi & = & \partial_{\mu} \phi - ig [

1545: A_{\mu}, \phi ] \ ,

1546: \end{eqnarray}

1547: and they are related by the identity

1548: \begin{eqnarray}

1549: [D_{\mu}, D_{\nu}] \phi & = & -ig \ [ F_{\mu \nu}, \phi ] \ .

1550: \label{group4}

1551: \end{eqnarray}

1552: Covariant derivatives of field strengths satisfy the Bianchi

1553: identity:

1554: \begin{eqnarray}

1555: \label{Bianchi} D^{\mu} F^{\nu \rho} + D^{\rho} F^{\mu \nu} +

1556: D^{\nu} F^{\rho \mu} = 0 \ .

1557: \end{eqnarray}

1558: 

1559: \end{enumerate}

1560: 

1561: \section{Tensors}

1562: \label{tensors}

1563: 

1564: \subsection{$t_{(8)}$ and $(\eta \cdot t_{(8)})_1$ tensors}

1565: \label{t8tensor}

1566: 

1567: \noindent The $t_{(8)}$ tensor\footnote{An explicit expression for

1568: it may be found in equation (4.A.21) of \cite{Schwarz1}.},

1569: characteristic of the 4 boson scattering amplitude, is

1570: antisymmetric on each pair $(\mu_j ,\nu_j)$ ($j=1,2,3,4$) and is

1571: symmetric under any exchange of such of pairs.  It satisfies the

1572: identity\footnote{Formula (\ref{t8}) has been taken from appendix

1573: A of \cite{DeRoo1}.}:

1574: \begin{multline}

1575: t^{(8)}_{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4}

1576: A_1^{\mu_1 \nu_1} A_2^{\mu_2 \nu_2} A_3^{\mu_3 \nu_3} A_4^{\mu_4

1577: \nu_4}  = \\

1578: \begin{split}

1579: &  -2 \biggl( \mbox{Tr}(A_1 A_2)\mbox{Tr}(A_3 A_4) + \mbox{Tr}(A_1

1580: A_3)\mbox{Tr}(A_2 A_4) +

1581: \mbox{Tr}(A_1 A_4)\mbox{Tr}(A_2 A_3) \biggr)+ \\

1582: & + 8 \biggl( \mbox{Tr}(A_1 A_2 A_3 A_4) + \mbox{Tr}(A_1 A_3 A_2

1583: A_4) + \mbox{Tr}(A_1 A_3 A_4 A_2)  \biggr) \ ,

1584: \end{split}

1585: \label{t8}

1586: \end{multline}

1587: where the $A_j$ tensors are antisymmetric and where `Tr' means

1588: the trace over the spacetime indexes.\\

1589: \noindent A ten index tensor, which is also antisymmetric on each

1590: pair $(\mu_j ,\nu_j)$, can be constructed from the Minkowski

1591: metric tensor and the $t_{(8)}$ one, as follows:

1592: \begin{eqnarray}

1593: {(\eta \cdot t_{(8)})_1}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3

1594: \mu_4 \nu_4 \mu_5 \nu_5} & = & \eta^{\nu_3 \nu_4} t_{(8)}^{\mu_3

1595: \mu_4 \mu_5 \nu_5 \mu_1 \nu_1 \mu_2 \nu_2} + \eta^{\mu_3 \mu_4}

1596: t_{(8)}^{\nu_3 \nu_4 \mu_5 \nu_5 \mu_1 \nu_1 \mu_2 \nu_2} -

1597: \nonumber \\

1598: &&- \ \eta^{\mu_3 \nu_4} t_{(8)}^{\nu_3 \mu_4 \mu_5 \nu_5 \mu_1

1599: \nu_1 \mu_2 \nu_2} - \eta^{\nu_3 \mu_4} t_{(8)}^{\mu_3 \nu_4 \mu_5

1600: \nu_5 \mu_1 \nu_1 \mu_2 \nu_2} \ . \label{etat8}

1601: \end{eqnarray}

1602: This tensor appears in the 5-point amplitude of the open

1603: superstring. It also changes sign under a twisting

1604: transformation\footnote{See the third item of subsection

1605: \ref{properties} for further details about a twisting transformation

1606: on the disk.} with respect to index 1, that is,

1607: \begin{eqnarray}

1608: {(\eta \cdot t_{(8)})_1}^{\mu_1 \nu_1 \mu_5 \nu_5 \mu_4 \nu_4

1609: \mu_3 \nu_3 \mu_2 \nu_2} = - {(\eta \cdot t_{(8)})_1}^{\mu_1 \nu_1

1610: \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} \ .

1611: \label{etat8twisting}

1612: \end{eqnarray}

1613: We have used the subindex $1$ in the $(\eta \cdot t_{(8)})$ tensor

1614: as a reminder of this relation.

1615: 

1616: 

1617: \subsection{$t_{(10)}$ tensor}

1618: \label{t10tensor}

1619: 

1620: The $t_{(10)}$ tensor is another ten index tensor that appears in

1621: the 5-point amplitude of the open superstring. It is linearly

1622: independent to the $(\eta \cdot t_{(8)})_1$ one. It can be

1623: constructed by the following procedure:

1624: 

1625: \begin{enumerate}

1626: \item Using the Minkowski metric tensor, we make all possible ten

1627: index

1628: tensorial structures:%

1629: \begin{equation*}

1630: \eta ^{\mu _{i}\nu _{n}}\eta ^{\mu _{j}\nu _{o}}\eta ^{\mu _{k}\nu

1631: _{p}}\eta ^{\mu _{l}\nu _{q}}\eta ^{\mu _{m}\nu _{r}},\text{ }\eta

1632: ^{\mu _{i}\nu _{n}}\eta ^{\mu _{j}\nu _{o}}\eta ^{\mu _{k}\nu

1633: _{p}}\eta ^{\mu _{l}\mu _{m}}\eta ^{\nu _{r}\nu _{q}}\text{ and

1634: }\eta ^{\mu _{i}\nu _{n}}\eta ^{\mu _{j}\mu _{k}}\eta ^{\nu _{p}\nu

1635: _{o}}\eta ^{\mu _{l}\mu _{m}}\eta ^{\nu _{r}\nu _{q}},

1636: \end{equation*}%

1637: with $i\neq j\neq k\neq l\neq m$ and $n\neq o\neq p\neq q\neq r$.

1638: These

1639: indexes are in the range $1, \ldots, 5$. When applied to the general expression $%

1640: \zeta _{\mu _{1}}^{1}k_{\nu _{1}}^{1}\zeta _{\mu _{2}}^{2}k_{\nu

1641: _{2}}^{2}\zeta _{\mu _{3}}^{3}k_{\nu _{3}}^{3}\zeta _{\mu

1642: _{4}}^{4}k_{\nu _{4}}^{4}\zeta _{\mu _{5}}^{5}k_{\nu _{5}}^{5}$,

1643: many terms will be null due to the physical state condition

1644: $\left( \zeta ^{i}\cdot k^{i}=0\right) $. Despising these terms,

1645: 544 structures remain.

1646: 

1647: \item Making a linear combination of these 544 terms, we demand

1648: antisymmetry on each pair $(\mu _{j},\nu _{j})$ and also a change

1649: of sign under a twisting transformation with respect to index $1$

1650: (the same as in eq. (\ref{etat8twisting}) ). This procedure

1651: results in a ten index tensor with only $14$ free parameters,

1652: which will be our \textit{ansatz} for the $t_{(10)}$ tensor.\\

1653: Unlike the $t_{(8)}$ tensor, the $t_{(10)}$ one \underline{is not

1654: cyclic invariant} in the pair of indexes

1655: $(\mu_{j},\nu_{j})$\footnote{Due to the symmetries of the 4-point

1656: subamplitude, the $t_{(8)}$ tensor is expected to be cyclic

1657: invariant in the pair of indexes $(\mu_{j},\nu_{j})$, but in fact

1658: it happens to be $completely$ symmetric with respect to those

1659: indexes\cite{Schwarz1}.}.

1660: 

1661: \item Then, the last $14$ parameters are determined by

1662: substituting $t_{(10)}$ in the kinematical expression

1663: $h^{(1)}(\zeta ,k)$, in (\ref{h23}), and  comparing it with the

1664: corresponding kinematical structure in eq. (\ref{AD2nF5nopolesII})

1665: of appendix \ref{derivation}. This comparison is not immediate

1666: since both expressions only agree after using on-shell and

1667: physical state conditions, together with momentum conservation. At

1668: the end we obtain an expression which satisfies an identity

1669: similar to that in (\ref{t8}):

1670: \begin{multline}

1671: t^{(10)}_{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5

1672: \nu_5} A_1^{\mu_1 \nu_1} A_2^{\mu_2 \nu_2} A_3^{\mu_3 \nu_3}

1673: A_4^{\mu_4 \nu_4} A_5^{\mu_5 \nu_5} = \\

1674: \begin{split}

1675: & \hphantom{+ } - 8 \left[  \ \ \mbox{Tr}(A_1 A_2)\mbox{Tr}(A_3

1676: A_4 A_5) + \mbox{Tr}(A_1 A_3)\mbox{Tr}(A_2 A_4 A_5) +

1677: \mbox{Tr}(A_1 A_4)\mbox{Tr}(A_2 A_3 A_5) + \right. \\

1678: & \hphantom{+ \ \ \ 8} + \mbox{Tr}(A_1 A_5)\mbox{Tr}(A_2 A_3 A_4)

1679: + \mbox{Tr}(A_2 A_3)\mbox{Tr}(A_1 A_4 A_5) +

1680: \mbox{Tr}(A_2 A_4)\mbox{Tr}(A_1 A_3 A_5) + \\

1681: & \hphantom{+ \ \ \ 8} + \mbox{Tr}(A_2 A_5)\mbox{Tr}(A_1 A_3 A_4)

1682: + \mbox{Tr}(A_3 A_4)\mbox{Tr}(A_1 A_2 A_5) +

1683: \mbox{Tr}(A_3 A_5)\mbox{Tr}(A_1 A_2 A_4) + \\

1684: & \hphantom{+ \ \ \ 8} \left. + \mbox{Tr}(A_4 A_5)\mbox{Tr}(A_1

1685: A_2 A_3) \ \ \right] + \ 48 \ \mbox{Tr}(A_1 A_2 A_3 A_4 A_5) \ +

1686: \\

1687: &+ 16 \left[ \ \ \mbox{Tr}(A_1 A_2 A_3 A_5 A_4) + \mbox{Tr}(A_1

1688: A_2 A_4 A_3 A_5) + \mbox{Tr}(A_1 A_2 A_5 A_3 A_4) + \right. \\

1689: & \hphantom{+ \ 16}+ \mbox{Tr}(A_1 A_2 A_4 A_5 A_3) -

1690: \mbox{Tr}(A_1 A_2 A_5 A_4 A_3) + \mbox{Tr}(A_1 A_3 A_2 A_4 A_5) -

1691: \\

1692: & \hphantom{+ \ 16}- \mbox{Tr}(A_1 A_3 A_2 A_5 A_4) +

1693: \mbox{Tr}(A_1 A_4 A_2 A_3 A_5) + \mbox{Tr}(A_1 A_5 A_2 A_3 A_4) -

1694: \\

1695: & \hphantom{+ \ 16} \left. - \mbox{Tr}(A_1 A_4 A_2 A_5 A_3) -

1696: \mbox{Tr}(A_1 A_5 A_2 A_4 A_3) \ \ \right] \ ,

1697: \end{split}

1698: \label{t10}

1699: \end{multline}

1700: where the $A_j$ fields are antisymmetric. From (\ref{t10}) an

1701: explicit expression of the $t_{(10)}$ tensor may be obtained, once

1702: its symmetry properties are considered.

1703: \end{enumerate}

1704: 

1705: 

1706: \section{$\alpha'$ dependent factors}

1707: \label{expansions}

1708: 

1709: \subsection{Gamma factor}

1710: \label{expansion-Gamma} As remarked in \cite{DeRoo2}, using the

1711: Taylor expansion for $\mbox{ln} \ \Gamma(1+z)$ \footnote{See

1712: formula (10.44c) of \cite{Arfken}, for example.},

1713: \begin{eqnarray}

1714: \mbox{ln} \ \Gamma(1+z) = -\gamma z + \sum_{k=2}^{\infty} (-1)^k

1715: \frac{\zeta(k)}{k} z^k \ \ \ \ \ \ \ (-1 < z \leq 1) \ ,

1716: \label{Taylor}

1717: \end{eqnarray}

1718: it may be proved that the explicit $\alpha'$ expansion for the

1719: Gamma factor in eq. (\ref{A1234}) is given by

1720: \begin{eqnarray}

1721: {\alpha'}^2 \frac{\Gamma(- \alpha' s) \Gamma(- \alpha' t)}

1722: {\Gamma(1- \alpha' s - \alpha' t)} = \frac{1}{st} \cdot

1723: \mbox{exp} \left\{ \sum_{k=2}^{\infty} \frac{\zeta(k)}{k}

1724: {\alpha'}^k (s^k+t^k-(s+t)^k) \right \}   \ . \label{formula1}

1725: \end{eqnarray}

1726: \noindent Up to ${\cal O}({\alpha'}^6)$ terms this gives:

1727: \begin{eqnarray}

1728: {\alpha'}^2 \frac{\Gamma(- \alpha' s) \Gamma(- \alpha' t)}

1729: {\Gamma(1- \alpha' s - \alpha' t)} & = & \frac{1}{st} -\frac{\

1730: \pi^2}{6} {\alpha'}^2 - \zeta(3) (s+t) \ {\alpha'}^3  -

1731: \frac{ \ \pi^4}{360}(4 s^2 + st + 4 t^2) \ {\alpha'}^4 \nonumber \\

1732:  & & + \left[ \frac{\ \pi^2}{6} \zeta(3) \ st (s+t) - \zeta(5)

1733: (s^3 + 2 s^2 t + 2 s t^2+ t^3)  \right]  {\alpha'}^5 \nonumber \\

1734:  & & + \left[ \frac{}{} \frac{1}{2} \zeta(3)^2 st(s+t)^2 - \frac{\

1735:  \pi^6}{15120}(16 s^4 + 12 s^3 t + 23 s^2 t^2 + 12 s t^3 + 16 t^4)

1736:  \right] {\alpha'}^6

1737: \nonumber \\

1738:  & &   + \ {\cal O}({\alpha'}^7) \ .

1739: \label{expansionGamma}

1740: \end{eqnarray}

1741: The function $f(s,t)$, defined in (\ref{f}), consists in this

1742: Gamma factor (divided by $\alpha'^2$) with the pole subtracted, so

1743: it has a well defined power series:

1744: \begin{eqnarray}

1745: f(s,t) = \sum_{m=0}^{\infty} \sum_{n=0}^{\infty} a_{mn} \ s^m t^n

1746: \ {\alpha'}^{m+n} \ ,

1747: \label{fpower}

1748: \end{eqnarray}

1749: where $a_{mn}=a_{nm}$, with $a_{00}= -\pi^2/6$,

1750: $a_{01}=-\zeta(3)$, $a_{11}=-\pi^4/360$, $a_{02}=-\pi^4/90$, etc.

1751: 

1752: \subsection{$K_2$, $K_3$ and $T$ factors}

1753: \label{expansion-K3K}

1754: 

1755: In this section we briefly review the definitions of the factors

1756: $K_2$ and $K_3$, given in appendix A.1 of \cite{Brandt1} and we deal

1757: with the factor $T$ introduced in eq. (\ref{K2-K}) of this work. We

1758: study the twisting and the cyclic symmetry and see how they are

1759: present in the $\alpha'$ expansion of $K_3$ and $T$.

1760: 

1761: \subsubsection{The definitions and the $\alpha_{ij}$ variables}

1762: \label{K2K3Kdefinitions}

1763: 

1764: The factors $K_2$ and $K_3$ were defined in \cite{Brandt1} as

1765: \begin{eqnarray}

1766: \label{K2-int}

1767: K_2 & = & \int_0^1 d x_3 \int_0^{x_3} d x_2 \

1768: x_3^{2 \alpha' \alpha_{13}} (1-x_3)^{2 \alpha'

1769:   \alpha_{34}-1} {x_2}^{2 \alpha' \alpha_{12}-1} (1-x_2)^{2 \alpha'

1770:   \alpha_{24}} (x_3-x_2)^{2 \alpha' \alpha_{23}} \ , \ \ \ \ \ \ \ \ \\

1771: \label{K3-int}

1772: K_3 & = & \int_0^1 d x_3 \int_0^{x_3} d x_2 \

1773: x_3^{2 \alpha' \alpha_{13}-1} (1-x_3)^{2 \alpha'

1774:   \alpha_{34}} {x_2}^{2 \alpha' \alpha_{12}} (1-x_2)^{2 \alpha'

1775:   \alpha_{24}-1} (x_3-x_2)^{2 \alpha' \alpha_{23}} \ .

1776: \end{eqnarray}

1777: 

1778: 

1779: \noindent After making the substitution $x_2 = u \cdot x_3$ in the

1780: inner integral (keeping $x_3$ constant) they become

1781: \begin{eqnarray}

1782: \label{K2-int2} K_2 & = & \int_0^1 d x_3 \int_0^{1} d u \ x_3^{2

1783: \alpha' \rho} (1-x_3)^{2 \alpha'

1784:   \alpha_{34}-1} {u}^{2 \alpha' \alpha_{12}-1} (1-u x_3)^{2 \alpha'

1785:   \alpha_{24}} (1-u)^{2 \alpha' \alpha_{23}} \ , \\

1786: \label{K3-int2}

1787: K_3 & = & \int_0^1 d x_3 \int_0^{1} d u \ x_3^{2

1788: \alpha' \rho} (1-x_3)^{2 \alpha'

1789:   \alpha_{34}} {u}^{2 \alpha' \alpha_{12}} (1-u x_3)^{2 \alpha'

1790:   \alpha_{24}-1} (1-u)^{2 \alpha' \alpha_{23}} \ .

1791: \end{eqnarray}

1792: Here,

1793: \begin{eqnarray}

1794: \rho = \alpha_{12} + \alpha_{13} + \alpha_{23} \ .

1795: \label{rho}

1796: \end{eqnarray}

1797: As was seen in \cite{Kitazawa1}, these factors may be written in

1798: terms of Euler Beta functions and a generalized Hypergeometric

1799: function, as

1800: \begin{eqnarray}

1801: K_2 & = & B(2 \alpha' \alpha_{12}, 1 + 2 \alpha' \alpha_{23})

1802: \cdot B(2 \alpha' \alpha_{34}, 1 + 2 \alpha' \rho) \cdot \nonumber

1803: \\

1804: & & \cdot \ {}_3F_2(1 + 2 \alpha' \rho, 2 \alpha' \alpha_{12}, -2

1805: \alpha' \alpha_{24}; \ 1 + 2 \alpha' \rho + 2 \alpha' \alpha_{34},

1806: 1 + 2 \alpha' \alpha_{12} + 2 \alpha' \alpha_{23}; \ 1) \ ,

1807: \label{K2hyperg} \\

1808: K_3 & = & B(1+2 \alpha' \alpha_{12}, 1 + 2 \alpha' \alpha_{23})

1809: \cdot B(1 + 2 \alpha' \alpha_{34}, 1 + 2 \alpha' \rho) \cdot

1810: \nonumber

1811: \\ & & \cdot \ {}_3F_2(1 + 2 \alpha' \rho,1+ 2 \alpha' \alpha_{12},1

1812: -2 \alpha' \alpha_{24}; \ 2 + 2 \alpha' \rho + 2 \alpha'

1813: \alpha_{34}, 2 + 2 \alpha' \alpha_{12} + 2 \alpha' \alpha_{23}; \

1814: 1) \ . \ \ \ \ \ \ \label{K3hyperg}

1815: \end{eqnarray}

1816: \noindent We will not use (\ref{K2hyperg}) and (\ref{K3hyperg}) to

1817: find any $\alpha'$ expansion in the present work. In

1818: \cite{Brandt1}, using (\ref{K2-int2}) and (\ref{K3-int2}), the

1819: $\alpha'$ expansions of $K_2$ and $K_3$ were found up to ${\cal

1820: O}(\alpha')$ terms. The double integrals that appear at each power

1821: of $\alpha'$ were calculated using Harmonic Polylogarithms

1822: \cite{Remiddi1}.

1823: 

1824: \noindent It is important to note that the ten $\alpha_{ij}$

1825: variables defined in (\ref{alphaij}) are not all independent once

1826: the on-shell (${k_i}^2=0$) and the momentum conservation

1827: conditions are taken into account. In fact, only five of them are

1828: independent. When finding the $\alpha'$ expansions of the factors

1829: $K_2$ and $K_3$ we have chosen $\alpha_{12}$, $\alpha_{23}$,

1830: $\alpha_{34}$, $\alpha_{45}$ and $\alpha_{51}$, as independent

1831: variables. The rest of them are given by

1832: \begin{eqnarray}

1833: \label{alpha13}

1834: \alpha_{13} = \alpha_{45} - \alpha_{12} - \alpha_{23} \\

1835: \label{alpha14}

1836: \alpha_{14} = \alpha_{23} - \alpha_{51} - \alpha_{45} \\

1837: \label{alpha24}

1838: \alpha_{24} = \alpha_{51} - \alpha_{23} - \alpha_{34} \\

1839: \label{alpha25}

1840: \alpha_{25} = \alpha_{34} - \alpha_{12} - \alpha_{51} \\

1841: \label{alpha35}

1842: \alpha_{35} = \alpha_{12} - \alpha_{45}

1843: -\alpha_{34}

1844: \end{eqnarray}

1845: Notice that considering (\ref{alpha13}), the $\rho$ variable in

1846: (\ref{rho}) coincides with $\alpha_{45}$.\\

1847: 

1848:  The $T$ factor has

1849: already been defined in eq. (\ref{K2-K}) as

1850: \begin{eqnarray*}

1851: T = (2 \alpha')^2 \biggl[ \ \alpha_{12} \ \alpha_{34} \ K_2 + (

1852: \alpha_{51} \ \alpha_{12} - \alpha_{12} \ \alpha_{34} +

1853: \alpha_{34} \ \alpha_{45} ) \ K_3 \ \biggr].

1854: \end{eqnarray*}

1855: $T$ and $K_3$ are related to the Gamma factor of the 4-point

1856: amplitude by

1857: \begin{eqnarray}

1858: (2 \alpha')^2 \frac{\Gamma(2 \alpha' \alpha_{34}) \

1859:  \Gamma(2 \alpha' \alpha_{45})}{\Gamma(1+ 2 \alpha' \alpha_{34} + 2 \alpha'

1860:  \alpha_{45})} \ = \ \frac{1}{\alpha_{34}

1861:  \alpha_{45}} \

1862:  T\biggl|_{\alpha_{12}=0} \ - \ (2 \alpha')^2 K_3\biggl|_{\alpha_{12}=0} \ .

1863: \label{relate3}

1864: \end{eqnarray}

1865: To see this we notice that using the definition of $T$, already

1866: mentioned, the relation in (\ref{relate3}) is equivalent to

1867: \begin{eqnarray}

1868: \{ \alpha_{12} \ K_2 \}\biggl|_{\alpha_{12}=0} = \alpha_{45}

1869: \frac{\Gamma(2 \alpha' \alpha_{34})

1870:  \Gamma(2 \alpha' \alpha_{45})}{\Gamma(1+ 2 \alpha' \alpha_{34} + 2 \alpha'

1871:  \alpha_{45})} \ ,

1872: \label{relate32}

1873: \end{eqnarray}

1874: which can be proved using (\ref{K2hyperg}) and the fact that the

1875: Hypergeometric function appearing there becomes $1$ when

1876: $\alpha_{12}=0$.\\

1877: \noindent The following relations (and the analog ones, obtained

1878: by cyclic permutations of them) can also be proved:

1879: \begin{eqnarray}

1880: \label{K3Gamma}

1881: (2 \alpha')^2

1882: K_3\biggl|_{\alpha_{12}=\alpha_{23}=0} & = & - (2 \alpha')^2

1883: \frac{ \ \alpha_{34}f(-2 \alpha_{34}, -2 \alpha_{45})-

1884: \alpha_{51}f(-2 \alpha_{45}, -2 \alpha_{51}) \

1885: }{\alpha_{34}-\alpha_{51}} \ ,

1886: \\

1887: \label{TGamma}

1888: T\biggl|_{\alpha_{12}=\alpha_{23}=0} \ - \ 1& = & -

1889: (2 \alpha')^2 \alpha_{34}\alpha_{45}\alpha_{51} \frac{ f(-2

1890: \alpha_{34}, -2 \alpha_{45})-f(-2 \alpha_{45}, -2 \alpha_{51}) \

1891: }{\alpha_{34}-\alpha_{51}} \ ,

1892: \\

1893: \label{T0}

1894: T\biggl|_{\alpha_{12}=\alpha_{34}=0} \ - \ 1& = & 0 \ .

1895: \end{eqnarray}

1896: 

1897: 

1898: 

1899: \subsubsection{Twisting symmetry}

1900: \label{K2K3Ktwisting}

1901: 

1902: The twisting transformation was already commented (in the third item

1903: of subsection \ref{properties}) to be equivalent to a parity

1904: transformation in the string world-sheet. In this section we prove

1905: that $K_2$, $K_3$ and $T$ are invariant under a twisting

1906: transformation with respect to the fifth insertion on the disk, that

1907: is, they are invariant under the transformation of indexes given in

1908: (\ref{twisting5}). For this purpose we use that

1909: \begin{eqnarray}

1910: \int_0^1 d x_3 \int_0^{x_3} d x_2 \ g(x_2, x_3) = \int_0^1 d x_3

1911: \int_0^{x_3} d x_2 \ g(1-x_3, 1-x_2) \ . \label{int-twisting}

1912: \end{eqnarray}

1913: This can be easily proved by first changing the order of

1914: integration on the left integral and then making the substitution

1915: \begin{eqnarray}

1916: \left\{ \begin{array}{ccc}

1917:        x_2 & = & 1 - {x_3}'\\

1918:        x_3 & = & 1 - {x_2}'

1919:        \end{array}

1920:        \right.

1921: \label{substitution1}

1922: \end{eqnarray}

1923: Applying the result in (\ref{int-twisting}) to the integrals in

1924: (\ref{K2-int}) and (\ref{K3-int}) we have the desired relation

1925: \begin{eqnarray}

1926: K_i(\alpha_{12}; \alpha_{13}; \alpha_{23}; \alpha_{24};

1927: \alpha_{34}) = K_i(\alpha_{34}; \alpha_{24}; \alpha_{23};

1928: \alpha_{13}; \alpha_{12}) \ \ \ (i=2,3) \ . \label{Ki}

1929: \end{eqnarray}

1930: 

1931: \noindent Once $K_2$ and $K_3$ are invariant under the twisting

1932: transformation in (\ref{twisting5}), from its definition in

1933: (\ref{K2-K}), it is immediate that $T$ is also invariant under the

1934: same transformation.

1935: 

1936: \subsubsection{Cyclic symmetry}

1937: \label{K3Kcyclic}

1938: 

1939: \noindent It is easy to see that $K_2$ is not cyclic invariant.

1940: For this purpose it is enough to look at the first term in the

1941: $\alpha'$ expansion of it, in (\ref{K2}), which clearly does not

1942: respect the cyclic invariance. In this section we will prove that

1943: $K_3$ and $T$ remain invariant under the following cyclic

1944: permutation of indexes:

1945: \begin{eqnarray}

1946: (1,2,3,4,5) \rightarrow (2,3,4,5,1) \ .

1947: \label{permutation}

1948: \end{eqnarray}

1949: So we need to prove that

1950: \begin{eqnarray}

1951: \label{K3cyclic}

1952: K_3(\alpha_{12}; \alpha_{13}; \alpha_{23};

1953: \alpha_{24}; \alpha_{34}) & = & K_3(\alpha_{23}; \alpha_{24};

1954: \alpha_{34};

1955: \alpha_{35}; \alpha_{45}) \ , \\

1956: \label{Kcyclic} T(\alpha_{12}; \alpha_{13}; \alpha_{23};

1957: \alpha_{24}; \alpha_{34}) & = & T(\alpha_{23}; \alpha_{24};

1958: \alpha_{34}; \alpha_{35}; \alpha_{45}) \ .

1959: \end{eqnarray}

1960: 

1961: 

1962: \begin{itemize}

1963: \item \underline{Cyclic invariance of $K_3$}:\\

1964: 

1965: \noindent Making the substitution

1966: \begin{eqnarray}

1967: \left\{ \begin{array}{ccc}

1968:        x_2 & = & 1 - {x_3}'\\

1969:        x_3 & = & (1 - {x_3}')/(1-{x_2}')

1970:        \end{array}

1971:        \right.

1972: \label{substitution2}

1973: \end{eqnarray}

1974: in the integral expression of $K_3$, in (\ref{K3-int}), it is not

1975: difficult to arrive to

1976: \begin{eqnarray}

1977: \label{K3cyclicproof} K_3( \alpha_{12};  \alpha_{13};

1978: \alpha_{23};  \alpha_{24};  \alpha_{34}) & = & K_3( \alpha_{23};

1979: \alpha_{24};  \alpha_{34};  -\alpha_{13}-\alpha_{23}-\alpha_{34};

1980:  \alpha_{12}+\alpha_{13}+\alpha_{23}) \ . \ \ \ \ \ \

1981: \end{eqnarray}

1982: Now, after considering the expressions for $\alpha_{13}$ and

1983: $\alpha_{35}$, given in (\ref{alpha13}) and (\ref{alpha35}), this

1984: last relation becomes the one in (\ref{K3cyclic}).

1985: 

1986: \item \underline{Cyclic invariance of $T$}:\\

1987: 

1988: The proof of the cyclic invariance in this case is very much more

1989: involved. We begin noticing that the desired condition

1990: (\ref{Kcyclic}) is equivalent to

1991: \begin{eqnarray}

1992: \alpha_{23} \ \alpha_{45} \ K_2(\alpha_{23}; \alpha_{24};

1993: \alpha_{34};

1994: \alpha_{35}; \alpha_{45}) \ & = & \  \alpha_{12} \ \alpha_{34} \

1995: K_2(\alpha_{12}; \alpha_{13};

1996: \alpha_{23};

1997: \alpha_{24}; \alpha_{34}) \ - \nonumber \\

1998: & & - \ (\alpha_{13}+ \alpha_{23}) \ \alpha_{24} \ K_3(\alpha_{12}; \alpha_{13};

1999: \alpha_{23};

2000: \alpha_{24}; \alpha_{34}) \ , \ \ \ \ \ \ \

2001: \label{aux-relation}

2002: \end{eqnarray}

2003: once the definition of $T(\alpha_{12}; \alpha_{13}; \alpha_{23};

2004: \alpha_{24}; \alpha_{34})$, given in (\ref{K2-K}), together with

2005: the

2006: relations for the $\alpha_{ij}$, given in (\ref{alpha13})-(\ref{alpha35}),

2007: have been considered.\\

2008: 

2009: \noindent We will prove (\ref{aux-relation}) in five steps:

2010: \begin{enumerate}

2011: \item Using the definition (\ref{K2-int}) we have that

2012: \begin{eqnarray}

2013: K_2(\alpha_{23}; \alpha_{24};

2014: \alpha_{34};

2015: \alpha_{35}; \alpha_{45}) =

2016: \hspace{9.0cm}

2017: \nonumber \\

2018: \int_0^1 d {x_3}' \int_0^{{x_3}'} d {x_2}' \

2019: {x_3}'^{2 \alpha' \alpha_{24}} (1-{x_3}')^{2 \alpha'

2020:   \alpha_{45}-1} {x_2}'^{2 \alpha' \alpha_{23}-1} (1-{x_2}')^{2 \alpha'

2021:   \alpha_{35}} ({x_3}'-{x_2}')^{2 \alpha' \alpha_{34}} \ ,

2022: \nonumber \\

2023: \label{K2prime}

2024: \end{eqnarray}

2025: and making the substitution (\ref{substitution2}) for

2026: $K_2(\alpha_{12}; \alpha_{13}; \alpha_{23};

2027: \alpha_{24}; \alpha_{34})$ and \\

2028: $K_3(\alpha_{12}; \alpha_{13}; \alpha_{23};

2029: \alpha_{24}; \alpha_{34})$, given in (\ref{K2-int}) and (\ref{K3-int}),

2030: we have that

2031: \begin{eqnarray}

2032: \label{K2transformed}

2033: K_2(\alpha_{12}; \alpha_{13}; \alpha_{23};

2034: \alpha_{24}; \alpha_{34}) \ =

2035: \hspace{9.0cm}

2036: \nonumber \\

2037: \int_0^1 d {x_3}' \int_0^{{x_3}'} d {x_2}' \

2038: {x_3}'^{2 \alpha' \alpha_{24}} (1-{x_3}')^{2 \alpha'

2039:   \alpha_{45}} {{x_2}'}^{2 \alpha' \alpha_{23}} (1-{x_2}')^{2 \alpha'

2040:   \alpha_{35}-1} ({x_3}'-{x_2}')^{2 \alpha' \alpha_{34}-1} \ ,

2041: \nonumber \\

2042: \\

2043: K_3(\alpha_{12}; \alpha_{13}; \alpha_{23};

2044: \alpha_{24}; \alpha_{34}) \ =

2045: \hspace{9.0cm}

2046: \nonumber \\

2047: \int_0^1 d {x_3}' \int_0^{{x_3}'} d {x_2}' \

2048: {x_3}'^{2 \alpha' \alpha_{24}-1} (1-{x_3}')^{2 \alpha'

2049:   \alpha_{45}} {x_2}'^{2 \alpha' \alpha_{23}} (1-{x_2}')^{2 \alpha'

2050:   \alpha_{35}-1} ({x_3}'-{x_2}')^{2 \alpha' \alpha_{34}} \ .

2051: \nonumber \\

2052: \label{K3transformed}

2053: \end{eqnarray}

2054: 

2055: \item Doing integration by parts in (\ref{K2prime}), with respect

2056: to the ${x_2}'$ variable, it may be proved that

2057: \begin{multline}

2058: K_2(\alpha_{23}; \alpha_{24};

2059: \alpha_{34};

2060: \alpha_{35}; \alpha_{45})  = \\

2061: \begin{split}

2062: \frac{\alpha_{35}}{\alpha_{23}}

2063: M(\alpha_{23}; \alpha_{24};

2064: \alpha_{34};

2065: \alpha_{35}; \alpha_{45})

2066: +\frac{\alpha_{34}}{\alpha_{23}}

2067: N(\alpha_{23}; \alpha_{24};

2068: \alpha_{34};

2069: \alpha_{35}; \alpha_{45}) \ ,

2070: \end{split}

2071: \label{int-parts-1}

2072: \end{multline}

2073: where

2074: \begin{eqnarray}

2075: M(\alpha_{23}; \alpha_{24}; \alpha_{34};

2076: \alpha_{35}; \alpha_{45}) \ =

2077: \hspace{9.0cm}

2078: \nonumber \\

2079: \int_0^1 d {x_3}' \int_0^{{x_3}'} d {x_2}' \

2080: {x_3}'^{2 \alpha' \alpha_{24}} (1-{x_3}')^{2 \alpha'

2081:   \alpha_{45}-1} {{x_2}'}^{2 \alpha' \alpha_{23}} (1-{x_2}')^{2 \alpha'

2082:   \alpha_{35}-1} ({x_3}'-{x_2}')^{2 \alpha' \alpha_{34}} \ ,

2083: \nonumber \\

2084: \label{M}

2085: \\

2086: N(\alpha_{23}; \alpha_{24}; \alpha_{34};

2087: \alpha_{35}; \alpha_{45}) \ =

2088: \hspace{9.0cm}

2089: \nonumber \\

2090: \int_0^1 d {x_3}' \int_0^{{x_3}'} d {x_2}' \

2091: {x_3}'^{2 \alpha' \alpha_{24}} (1-{x_3}')^{2 \alpha'

2092:   \alpha_{45}-1} {x_2}'^{2 \alpha' \alpha_{23}} (1-{x_2}')^{2 \alpha'

2093:   \alpha_{35}} ({x_3}'-{x_2}')^{2 \alpha' \alpha_{34}-1} \ .

2094: \nonumber \\

2095: \label{N}

2096: \end{eqnarray}

2097: 

2098: \item Now, doing integration by parts in (\ref{K3transformed}), with

2099: respect to the ${x_3}'$ variable, it may also be proved that

2100: \begin{multline}

2101: K_3(\alpha_{12}; \alpha_{13};

2102: \alpha_{23};

2103: \alpha_{24}; \alpha_{34})  = \\

2104: \begin{split}

2105: \frac{\alpha_{45}}{\alpha_{24}}

2106: M(\alpha_{23}; \alpha_{24};

2107: \alpha_{34};

2108: \alpha_{35}; \alpha_{45})

2109: +\frac{\alpha_{34}}{\alpha_{24}}

2110: K_2(\alpha_{12}; \alpha_{13};

2111: \alpha_{23};

2112: \alpha_{24}; \alpha_{34}) \ .

2113: \end{split}

2114: \label{int-parts-2}

2115: \end{multline}

2116: 

2117: \item Noticing that

2118: \begin{eqnarray}

2119: \frac{1}{(1-{x_3}')({x_3}'-{x_2}')} =

2120: \frac{1}{(1-{x_2}')(1-{x_3}')} +

2121: \frac{1}{(1-{x_2}')({x_3}'-{x_2}')} \ ,

2122: \label{partial-fractions}

2123: \end{eqnarray}

2124: and using (\ref{K2transformed}), (\ref{M}) and (\ref{N}), we have that

2125: \begin{eqnarray}

2126: N(\alpha_{23}; \alpha_{24};

2127: \alpha_{34};

2128: \alpha_{35}; \alpha_{45}) =

2129: M(\alpha_{23}; \alpha_{24};

2130: \alpha_{34};

2131: \alpha_{35}; \alpha_{45}) +

2132: K_2(\alpha_{12}; \alpha_{13};

2133: \alpha_{23};

2134: \alpha_{24}; \alpha_{34}) \ .

2135: \nonumber \\

2136: \label{NMK2}

2137: \end{eqnarray}

2138: 

2139: \item Finally, eliminating $M(\alpha_{23}; \alpha_{24};

2140: \alpha_{34};

2141: \alpha_{35}; \alpha_{45})$ and $N(\alpha_{23}; \alpha_{24};

2142: \alpha_{34};

2143: \alpha_{35}; \alpha_{45})$ from equations (\ref{int-parts-1}),

2144: (\ref{int-parts-2}) and (\ref{NMK2}), leads precisely to (\ref{aux-relation}).

2145: \end{enumerate}

2146: \end{itemize}

2147: 

2148: \noindent Now that the proof of (\ref{K3cyclic}) and

2149: (\ref{Kcyclic}) has been done, then it may be repeated for the

2150: rest of the cyclic permutations, guaranteeing that $K_3$ and $T$

2151: remain invariant under them. We have also checked this.

2152: 

2153: 

2154: 

2155: \subsubsection{The $\alpha'$ expansions of $(2 \alpha')^2K_3$ and $T$}

2156: \label{K3Kexpansions}

2157: 

2158: A detailed calculation of the $\alpha'$ expansions of $K_2$ and

2159: $K_3$ was done in \cite{Medina1}, up to ${\cal O}((2 \alpha')^4)$

2160: terms in both cases. This was done using Harmonic Polylogarithms

2161: \cite{Remiddi1} and the $harmpol$ package of FORM \cite{Vermaseren1}

2162: (see appendix A.3. of \cite{Brandt1} for more details about the type

2163: of calculations involved). The $\alpha'$ expansion of $T$ was also

2164: obtained in \cite{Medina1}, up to ${\cal O}((2 \alpha')^6)$ terms,

2165: by directly using its definition (\ref{K2-K}) and the expansions of

2166: $K_2$ and $K_3$. In this section we will only write the $\alpha'$

2167: expansions of $(2 \alpha')^2 K_3$ and $T$.

2168: 

2169: \noindent It is not very difficult to prove that the $\alpha'$

2170: expansions of $(2 \alpha')^2 K_3$ and $T$ are power series, that

2171: is, they do not have poles in $\alpha'$ (as $K_2$

2172: does\footnote{See eq. (\ref{K2}).}). So, at every order in

2173: $\alpha'$, there only appear polynomial expressions which, after

2174: using the relations (\ref{alpha13})-(\ref{alpha35}), may be

2175: written in terms of the five $\alpha_{ij}$ variables mentioned in

2176: section \ref{K2K3Kdefinitions}. Each of these polynomial

2177: expressions respects the $twisting$ and the $cyclic$ symmetry,

2178: that were seen to be satisfied by $(2 \alpha')^2 K_3$ and $T$ in

2179: sections \ref{K2K3Ktwisting} and \ref{K3Kcyclic}. The result

2180: obtained, written in terms of polynomial cyclic invariants (which

2181: will be specified in the next subsection), is the following

2182: \cite{Medina1}:

2183: \begin{eqnarray}

2184: (2 \alpha')^2 K_3 & = & (2 \alpha')^2 \ \zeta(2) \ I_1^{(0)} - \

2185: (2 \alpha')^3 \ \zeta(3) \ I_1^{(1)} \ + \ (2 \alpha')^4 \

2186: \frac{2}{5} \zeta(2)^2 \ ( \ I_1^{(2)} +

2187: \frac{1}{4}I_2^{(2)} + I_3^{(2)} \ ) + \nonumber \\

2188:  & + &  (2 \alpha')^5 \ \biggl[ \frac{}{} -\zeta(5) \

2189:  ( \ I_1^{(3)}+I_3^{(3)}+I_4^{(3)} \ ) + ( \ -2

2190:  \zeta(5)+\zeta(2)\zeta(3) \ ) \ ( \ I_2^{(3)}+I_5^{(3)}+I_6^{(3)} \ )+

2191:  \nonumber \\

2192:  & &\hphantom{(2 \alpha')^5 [}  + ( \ \frac{7}{2}

2193:  \zeta(5)-2\zeta(2)\zeta(3) \ ) \ I_7^{(3)} \frac{}{} \biggr] +

2194:  \nonumber \\

2195:  & + &  (2 \alpha')^6 \ \biggl[ \frac{}{} \frac{8}{35}\zeta(2)^3 \

2196:  ( \ I_1^{(4)}+I_3^{(4)}+I_4^{(4)}+I_7^{(4)} \ )

2197:  +\nonumber \\

2198:   &  & \hphantom{(2 \alpha')^6 \ \biggl[}

2199:  ( \ \frac{6}{35}\zeta(2)^3-\frac{1}{2}\zeta(3)^2 \ ) \

2200:  ( \ I_2^{(4)}+I_5^{(4)}+I_8^{(4)}-I_9^{(4)}

2201:  -I_{12}^{(4)}+I_{13}^{(4)} \ )+

2202:  \nonumber \\

2203:  & &\hphantom{(2 \alpha')^5 [}  + ( \ \frac{23}{70}

2204:  \zeta(2)^3-\zeta(3)^2 \ )( \ I_6^{(4)}+ I_{10}^{(4)} \ ) + ( \ -\frac{26}{105}

2205:  \zeta(2)^3+\zeta(3)^2 \ ) \ I_{11}^{(4)} + \nonumber \\

2206:  & &\hphantom{(2 \alpha')^5 [}  + ( \ -\frac{67}{105}

2207:  \zeta(2)^3+2 \zeta(3)^2 \ ) \ I_{14}^{(4)} \frac{}{} \biggr] \ + \ {\cal O}( \ (2

2208:  \alpha')^7 \ ) \ ,

2209:  \label{K3final} \\

2210: T & = & I_1^{(0)} \  - \ (2 \alpha')^3 \ \zeta(3) \ I_6^{(3)} \ +

2211: \ (2 \alpha')^4 \ \frac{2}{5} \zeta(2)^2 \ ( \ I_8^{(4)} +

2212: \frac{1}{4}I_{10}^{(4)} +

2213: I_{13}^{(4)} + I_{14}^{(4)} \ ) + \nonumber \\

2214:  & + &  (2 \alpha')^5 \ \biggl[ \frac{}{} -\zeta(5) \ ( \ I_{10}^{(5)}+I_{15}^{(5)}

2215:  +I_{19}^{(5)}+I_{22}^{(5)}+I_{25}^{(5)} \ ) + \nonumber \\

2216:   & & \hphantom{(2 \alpha')^5 [} + ( \ -2

2217:  \zeta(5)+\zeta(2)\zeta(3) \ ) \ ( \ I_{12}^{(5)}+I_{16}^{(5)}+I_{18}^{(5)} \ )+

2218:  \nonumber \\

2219:  & &\hphantom{(2 \alpha')^5 [}  + ( \ \frac{7}{2}

2220:  \zeta(5)-2\zeta(2)\zeta(3) \ ) \ ( \ I_{23}^{(5)}+I_{24}^{(5)} \ ) \frac{}{} \biggr] +

2221:  \nonumber \\

2222:  & + &  (2 \alpha')^6 \ \biggl[ \frac{}{} \frac{8}{35} \zeta(2)^3 \

2223:  ( \ I_{10}^{(6)}+I_{15}^{(6)}

2224:  +I_{21}^{(6)}+I_{26}^{(6)}+I_{30}^{(6)}+I_{33}^{(6)}+I_{41}^{(6)} \ )

2225:  +\nonumber \\

2226:   &  & \hphantom{(2 \alpha')^6 \ \biggl[}

2227:  ( \ \frac{6}{35}\zeta(2)^3-\frac{1}{2}\zeta(3)^2 \ ) \

2228:  ( \ I_{12}^{(6)}+I_{18}^{(6)}

2229:  +I_{29}^{(6)}-I_{31}^{(6)}-I_{32}^{(6)}+I_{34}^{(6)}-I_{36}^{(6)}-I_{38}^{(6)}

2230:   \ )+

2231:  \nonumber \\

2232:  & &\hphantom{(2 \alpha')^5 [}  + ( \ \frac{23}{70}

2233:  \zeta(2)^3-\zeta(3)^2 \ ) \ (   \ I_{20}^{(6)}+I_{27}^{(6)})

2234:  + \nonumber \\

2235:   & &\hphantom{(2 \alpha')^5 [} + ( \ -\frac{26}{105}

2236:  \zeta(2)^3+ \zeta(3)^2 \ ) \ ( \ I_{39}^{(6)} + I_{40}^{(6)} \ ) +

2237:  ( \ -\frac{67}{105}

2238:  \zeta(2)^3+2 \zeta(3)^2 \ ) \ I_{37}^{(6)} \frac{}{} + \nonumber \\

2239:   & &\hphantom{(2 \alpha')^5 [} + ( \ -\frac{109}{210} \zeta(2)^3

2240: -\frac{3}{2} \zeta(3)^2 \ ) \ I_{42}^{(6)} \biggr] \ + \ {\cal O}(

2241: \ (2 \alpha')^7 \ ) \ .

2242:  \label{Kfinal}

2243: \end{eqnarray}

2244: Here $I_{j}^{(i)}$ denotes the $j$-th polynomial cyclic invariant of

2245: degree $i$.

2246: 

2247: 

2248: 

2249: \subsubsection{Polynomial cyclic invariants}

2250: \label{polynomial}

2251: 

2252: \noindent The polynomial cyclic invariants are uniquely determined

2253: up to a global factor which we have chosen to be $1$. In the next

2254: lines we list them up to sixth degree. At degrees $0$, $1$, $2$,

2255: $3$, $4$, $5$ and $6$ there are, respectively, one, one, three,

2256: seven, fourteen, twenty six and forty two linearly independent

2257: polynomial cyclic invariants.

2258: \begin{itemize}

2259: \item Degree 0:

2260: \begin{eqnarray}

2261: I_{1}^{(0)} = 1 \ .

2262: \label{I10}

2263: \end{eqnarray}

2264: \item Degree 1:

2265: \begin{eqnarray}

2266: I_{1}^{(1)} = \alpha_{12} + \alpha_{23} + \alpha_{34} +

2267: \alpha_{45} + \alpha_{51} \ .

2268: \label{I11}

2269: \end{eqnarray}

2270: \item Degree 2:

2271: \begin{eqnarray}

2272: I_{1}^{(2)} & = & \alpha_{12}^2 + \alpha_{23}^2 + \alpha_{34}^2 +

2273: \alpha_{45}^2 + \alpha_{51}^2 \ , \nonumber \\

2274: I_{2}^{(2)} & = & \alpha_{12}\alpha_{23} + \alpha_{23}\alpha_{34}

2275: + \alpha_{34}\alpha_{45}  + \alpha_{45}\alpha_{51}  +

2276: \alpha_{51}\alpha_{12}  \ , \label{Icyclic2} \\

2277: I_{3}^{(2)} & = &

2278: \alpha_{12}\alpha_{34} + \alpha_{23}\alpha_{45} +

2279: \alpha_{34}\alpha_{51}  + \alpha_{45}\alpha_{12}  +

2280: \alpha_{51}\alpha_{23}  \ . \nonumber

2281: \end{eqnarray}

2282: \item Degree 3:

2283: \begin{eqnarray}

2284: I_{1}^{(3)} & = & \alpha_{12}^3 + \alpha_{23}^3 + \alpha_{34}^3 +

2285: \alpha_{45}^3 + \alpha_{51}^3 \ , \nonumber \\

2286: I_{2}^{(3)} & = & \alpha_{12}^2 \alpha_{23} + \alpha_{23}^2

2287: \alpha_{34}  + \alpha_{34}^2 \alpha_{45}  + \alpha_{45}^2

2288: \alpha_{51}  + \alpha_{51}^2 \alpha_{12}  \ , \nonumber \\

2289: I_{3}^{(3)} & = & \alpha_{12}^2 \alpha_{34} + \alpha_{23}^2

2290: \alpha_{45}  + \alpha_{34}^2 \alpha_{51}  + \alpha_{45}^2

2291: \alpha_{12}  + \alpha_{51}^2\alpha_{23}  \ , \nonumber \\

2292: I_{4}^{(3)} & = & \alpha_{12}^2 \alpha_{45} + \alpha_{23}^2

2293: \alpha_{51}  + \alpha_{34}^2 \alpha_{12}  + \alpha_{45}^2

2294: \alpha_{23}  + \alpha_{51}^2\alpha_{34}  \ , \label{Icyclic3} \\

2295: I_{5}^{(3)} & = & \alpha_{12}^2 \alpha_{51} + \alpha_{23}^2

2296: \alpha_{12}  + \alpha_{34}^2 \alpha_{23}  + \alpha_{45}^2

2297: \alpha_{34}  + \alpha_{51}^2\alpha_{45}  \ , \nonumber \\

2298: I_{6}^{(3)} & = & \alpha_{12}\alpha_{23}\alpha_{34} +

2299: \alpha_{23}\alpha_{34}\alpha_{45}   +

2300: \alpha_{34}\alpha_{45}\alpha_{51}  +

2301: \alpha_{45}\alpha_{51}\alpha_{12}  +

2302: \alpha_{51}\alpha_{12}\alpha_{23}   \ , \nonumber \\

2303: I_{7}^{(3)} & = & \alpha_{12}\alpha_{34}\alpha_{45} +

2304: \alpha_{23}\alpha_{45}\alpha_{51}   +

2305: \alpha_{34}\alpha_{51}\alpha_{12}  +

2306: \alpha_{45}\alpha_{12}\alpha_{23}  +

2307: \alpha_{51}\alpha_{23}\alpha_{34}   \ . \nonumber

2308: \end{eqnarray}

2309: \item Degree 4:

2310: \begin{eqnarray}

2311: I_{1}^{(4)} & = & \alpha_{12}^4 + \alpha_{23}^4 + \alpha_{34}^4 +

2312: \alpha_{45}^4 + \alpha_{51}^4 \ , \nonumber \\

2313: I_{2}^{(4)} & = & \alpha_{12}^3 \alpha_{23} + \alpha_{23}^3

2314: \alpha_{34}  + \alpha_{34}^3 \alpha_{45}  + \alpha_{45}^3

2315: \alpha_{51}  + \alpha_{51}^3 \alpha_{12}  \ ,

2316: \nonumber \\

2317: I_{3}^{(4)} & = & \alpha_{12}^3 \alpha_{34} + \alpha_{23}^3

2318: \alpha_{45}  + \alpha_{34}^3 \alpha_{51}  + \alpha_{45}^3

2319: \alpha_{12}  + \alpha_{51}^3\alpha_{23}  \ .

2320: \nonumber \\

2321: I_{4}^{(4)} & = & \alpha_{12}^3 \alpha_{45} + \alpha_{23}^3

2322: \alpha_{51}  + \alpha_{34}^3 \alpha_{12}  + \alpha_{45}^3

2323: \alpha_{23}  + \alpha_{51}^3\alpha_{34}  \ .

2324: \nonumber \\

2325: I_{5}^{(4)} & = & \alpha_{12}^3 \alpha_{51} + \alpha_{23}^3

2326: \alpha_{12}  + \alpha_{34}^3 \alpha_{23}  + \alpha_{45}^3

2327: \alpha_{34}  + \alpha_{51}^3 \alpha_{45}  \ .

2328: \nonumber \\

2329: I_{6}^{(4)} & = & \alpha_{12}^2 \alpha_{23}^2 + \alpha_{23}^2

2330: \alpha_{34}^2  + \alpha_{34}^2 \alpha_{45}^2  + \alpha_{45}^2

2331: \alpha_{51}^2  + \alpha_{51}^2 \alpha_{12}^2  \ ,

2332: \nonumber \\

2333: I_{7}^{(4)} & = & \alpha_{12}^2 \alpha_{34}^2 + \alpha_{23}^2

2334: \alpha_{45}^2  + \alpha_{34}^2 \alpha_{51}^2  + \alpha_{45}^2

2335: \alpha_{12}^2  + \alpha_{51}^2 \alpha_{23}^2  \ ,

2336: \nonumber \\

2337: I_{8}^{(4)} & = & \alpha_{12}^2\alpha_{23}\alpha_{34} +

2338: \alpha_{23}^2\alpha_{34}\alpha_{45}   +

2339: \alpha_{34}^2\alpha_{45}\alpha_{51}  +

2340: \alpha_{45}^2\alpha_{51}\alpha_{12}  +

2341: \alpha_{51}^2\alpha_{12}\alpha_{23}   \ ,

2342: \label{Icyclic4} \\

2343: I_{9}^{(4)} & = & \alpha_{12}^2\alpha_{23}\alpha_{45} +

2344: \alpha_{23}^2\alpha_{34}\alpha_{51}   +

2345: \alpha_{34}^2\alpha_{45}\alpha_{12}  +

2346: \alpha_{45}^2\alpha_{51}\alpha_{23}  +

2347: \alpha_{51}^2\alpha_{12}\alpha_{34}   \ ,

2348: \nonumber \\

2349: I_{10}^{(4)} & = & \alpha_{12}^2\alpha_{23}\alpha_{51} +

2350: \alpha_{23}^2\alpha_{34}\alpha_{12}   +

2351: \alpha_{34}^2\alpha_{45}\alpha_{23}  +

2352: \alpha_{45}^2\alpha_{51}\alpha_{34}  +

2353: \alpha_{51}^2\alpha_{12}\alpha_{45}   \ ,

2354: \nonumber \\

2355: I_{11}^{(4)} & = & \alpha_{12}^2\alpha_{34}\alpha_{45} +

2356: \alpha_{23}^2\alpha_{45}\alpha_{51}   +

2357: \alpha_{34}^2\alpha_{51}\alpha_{12}  +

2358: \alpha_{45}^2\alpha_{12}\alpha_{23}  +

2359: \alpha_{51}^2\alpha_{23}\alpha_{34}   \ ,

2360: \nonumber \\

2361: I_{12}^{(4)} & = & \alpha_{12}^2\alpha_{34}\alpha_{51} +

2362: \alpha_{23}^2\alpha_{45}\alpha_{12}   +

2363: \alpha_{34}^2\alpha_{51}\alpha_{23}  +

2364: \alpha_{45}^2\alpha_{12}\alpha_{34}  +

2365: \alpha_{51}^2\alpha_{23}\alpha_{45}   \ ,

2366: \nonumber \\

2367: I_{13}^{(4)} & = & \alpha_{12}^2\alpha_{45}\alpha_{51} +

2368: \alpha_{23}^2\alpha_{51}\alpha_{12}   +

2369: \alpha_{34}^2\alpha_{12}\alpha_{23}  +

2370: \alpha_{45}^2\alpha_{23}\alpha_{34}  +

2371: \alpha_{51}^2\alpha_{34}\alpha_{45}   \ ,

2372: \nonumber \\

2373: I_{14}^{(4)} & = & \alpha_{12}\alpha_{23}\alpha_{34}\alpha_{45} +

2374: \alpha_{23}\alpha_{34}\alpha_{45}\alpha_{51} +

2375: \alpha_{34}\alpha_{45}\alpha_{51}\alpha_{12} +

2376: \alpha_{45}\alpha_{51}\alpha_{12}\alpha_{23} +

2377: \alpha_{51}\alpha_{12}\alpha_{23}\alpha_{34}  \ . \nonumber

2378: \end{eqnarray}

2379: \item Degree 5:

2380: \begin{eqnarray}

2381: I_{1}^{(5)} & = & \alpha_{12}^5 + \alpha_{23}^5 + \alpha_{34}^5 +

2382: \alpha_{45}^5 + \alpha_{51}^5 \ ,

2383: \nonumber \\

2384: I_{2}^{(5)} & = & \alpha_{12}^4 \alpha_{23} + \alpha_{23}^4

2385: \alpha_{34}  + \alpha_{34}^4 \alpha_{45}  + \alpha_{45}^4

2386: \alpha_{51}  + \alpha_{51}^4 \alpha_{12}  \ ,

2387: \nonumber \\

2388: I_{3}^{(5)} & = & \alpha_{12}^4 \alpha_{34} + \alpha_{23}^4

2389: \alpha_{45}  + \alpha_{34}^4 \alpha_{51}  + \alpha_{45}^4

2390: \alpha_{12}  + \alpha_{51}^4\alpha_{23}  \ ,

2391: \nonumber \\

2392: I_{4}^{(5)} & = & \alpha_{12}^4 \alpha_{45} + \alpha_{23}^4

2393: \alpha_{51}  + \alpha_{34}^4 \alpha_{12}  + \alpha_{45}^4

2394: \alpha_{23}  + \alpha_{51}^4\alpha_{34}  \ ,

2395: \nonumber \\

2396: I_{5}^{(5)} & = & \alpha_{12}^4 \alpha_{51} + \alpha_{23}^4

2397: \alpha_{12}  + \alpha_{34}^4 \alpha_{23}  + \alpha_{45}^4

2398: \alpha_{34}  + \alpha_{51}^4 \alpha_{45}  \ ,

2399: \nonumber \\

2400: I_{6}^{(5)} & = &  \alpha_{12}^3 \alpha_{23}^2 + \alpha_{23}^3

2401: \alpha_{34}^2  + \alpha_{34}^3 \alpha_{45}^2  + \alpha_{45}^3

2402: \alpha_{51}^2  + \alpha_{51}^3 \alpha_{12}^2  \ ,

2403: \nonumber \\

2404: I_{7}^{(5)} & = &  \alpha_{12}^3 \alpha_{34}^2 + \alpha_{23}^3

2405: \alpha_{45}^2  + \alpha_{34}^3 \alpha_{51}^2  + \alpha_{45}^3

2406: \alpha_{12}^2  + \alpha_{51}^3 \alpha_{23}^2  \ ,

2407: \nonumber \\

2408: I_{8}^{(5)} & = &  \alpha_{12}^3 \alpha_{45}^2 + \alpha_{23}^3

2409: \alpha_{51}^2  + \alpha_{34}^3 \alpha_{12}^2  + \alpha_{45}^3

2410: \alpha_{23}^2  + \alpha_{51}^3 \alpha_{34}^2  \ ,

2411: \nonumber \\

2412: I_{9}^{(5)} & = &  \alpha_{12}^3 \alpha_{51}^2 + \alpha_{23}^3

2413: \alpha_{12}^2  + \alpha_{34}^3 \alpha_{23}^2  + \alpha_{45}^3

2414: \alpha_{34}^2  + \alpha_{51}^3 \alpha_{45}^2  \ ,

2415: \nonumber \\

2416: I_{10}^{(5)} & = &  \alpha_{12}^3\alpha_{23}\alpha_{34} +

2417: \alpha_{23}^3\alpha_{34}\alpha_{45}   +

2418: \alpha_{34}^3\alpha_{45}\alpha_{51}  +

2419: \alpha_{45}^3\alpha_{51}\alpha_{12}  +

2420: \alpha_{51}^3\alpha_{12}\alpha_{23}   \ ,

2421: \nonumber \\

2422: I_{11}^{(5)} & = &  \alpha_{12}^3\alpha_{23}\alpha_{45} +

2423: \alpha_{23}^3\alpha_{34}\alpha_{51}   +

2424: \alpha_{34}^3\alpha_{45}\alpha_{12}  +

2425: \alpha_{45}^3\alpha_{51}\alpha_{23}  +

2426: \alpha_{51}^3\alpha_{12}\alpha_{34}   \ ,

2427: \nonumber \\

2428: I_{12}^{5)} & = &  \alpha_{12}^3\alpha_{23}\alpha_{51} +

2429: \alpha_{23}^3\alpha_{34}\alpha_{12}   +

2430: \alpha_{34}^3\alpha_{45}\alpha_{23}  +

2431: \alpha_{45}^3\alpha_{51}\alpha_{34}  +

2432: \alpha_{51}^3\alpha_{12}\alpha_{45}   \ ,

2433: \nonumber \\

2434: I_{13}^{(5)} & = &  \alpha_{12}^3\alpha_{34}\alpha_{45} +

2435: \alpha_{23}^3\alpha_{45}\alpha_{51}  +

2436: \alpha_{34}^3\alpha_{51}\alpha_{12}  +

2437: \alpha_{45}^3\alpha_{12}\alpha_{23}  +

2438: \alpha_{51}^3\alpha_{23}\alpha_{34}   \ ,

2439: \nonumber \\

2440: I_{14}^{(5)} & = &  \alpha_{12}^3\alpha_{34}\alpha_{51} +

2441: \alpha_{23}^3\alpha_{45}\alpha_{12}   +

2442: \alpha_{34}^3\alpha_{51}\alpha_{23}  +

2443: \alpha_{45}^3\alpha_{12}\alpha_{34}  +

2444: \alpha_{51}^3\alpha_{23}\alpha_{45}   \ ,

2445: \label{Icyclic5} \\

2446: I_{15}^{(5)} & = &  \alpha_{12}^3\alpha_{45}\alpha_{51} +

2447: \alpha_{23}^3\alpha_{51}\alpha_{12}   +

2448: \alpha_{34}^3\alpha_{12}\alpha_{23}  +

2449: \alpha_{45}^3\alpha_{23}\alpha_{34}  +

2450: \alpha_{51}^3\alpha_{34}\alpha_{45}   \ ,

2451: \nonumber \\

2452: I_{16}^{(5)} & = &  \alpha_{12}^2\alpha_{23}^2\alpha_{34} +

2453: \alpha_{23}^2\alpha_{34}^2\alpha_{45}  +

2454: \alpha_{34}^2\alpha_{45}^2\alpha_{51}  +

2455: \alpha_{45}^2\alpha_{51}^2\alpha_{12}  +

2456: \alpha_{51}^2\alpha_{12}^2\alpha_{23}   \ ,

2457: \nonumber \\

2458: I_{17}^{(5)} & = &  \alpha_{12}^2\alpha_{23}^2\alpha_{45} +

2459: \alpha_{23}^2\alpha_{34}^2\alpha_{51}  +

2460: \alpha_{34}^2\alpha_{45}^2\alpha_{12}  +

2461: \alpha_{45}^2\alpha_{51}^2\alpha_{23}  +

2462: \alpha_{51}^2\alpha_{12}^2\alpha_{34}   \ ,

2463: \nonumber \\

2464: I_{18}^{(5)} & = &  \alpha_{12}^2\alpha_{23}^2\alpha_{51} +

2465: \alpha_{23}^2\alpha_{34}^2\alpha_{12}  +

2466: \alpha_{34}^2\alpha_{45}^2\alpha_{23}  +

2467: \alpha_{45}^2\alpha_{51}^2\alpha_{34}  +

2468: \alpha_{51}^2\alpha_{12}^2\alpha_{45}   \ ,

2469: \nonumber \\

2470: I_{19}^{(5)} & = &  \alpha_{12}^2\alpha_{34}^2\alpha_{23} +

2471: \alpha_{23}^2\alpha_{45}^2\alpha_{34}  +

2472: \alpha_{34}^2\alpha_{51}^2\alpha_{45}  +

2473: \alpha_{45}^2\alpha_{12}^2\alpha_{51}  +

2474: \alpha_{51}^2\alpha_{23}^2\alpha_{12}   \ ,

2475: \nonumber \\

2476: I_{20}^{(5)} & = &  \alpha_{12}^2\alpha_{34}^2\alpha_{45} +

2477: \alpha_{23}^2\alpha_{45}^2\alpha_{51}  +

2478: \alpha_{34}^2\alpha_{51}^2\alpha_{12}  +

2479: \alpha_{45}^2\alpha_{12}^2\alpha_{23}  +

2480: \alpha_{51}^2\alpha_{23}^2\alpha_{34}   \ ,

2481: \nonumber \\

2482: I_{21}^{(5)} & = &  \alpha_{12}^2\alpha_{34}^2\alpha_{51} +

2483: \alpha_{23}^2\alpha_{45}^2\alpha_{12}  +

2484: \alpha_{34}^2\alpha_{51}^2\alpha_{23}  +

2485: \alpha_{45}^2\alpha_{12}^2\alpha_{34}  +

2486: \alpha_{51}^2\alpha_{23}^2\alpha_{45}   \ ,

2487: \nonumber \\

2488: I_{22}^{(5)} & = &  \alpha_{12}^2\alpha_{23}\alpha_{34}\alpha_{45}

2489: + \alpha_{23}^2\alpha_{34}\alpha_{45}\alpha_{51} +

2490: \alpha_{34}^2\alpha_{45}\alpha_{51}\alpha_{12} +

2491: \alpha_{45}^2\alpha_{51}\alpha_{12}\alpha_{23} +

2492: \alpha_{51}^2\alpha_{12}\alpha_{23}\alpha_{34}  \ ,

2493: \nonumber \\

2494: I_{23}^{(5)} & = &  \alpha_{12}^2\alpha_{23}\alpha_{34}\alpha_{51}

2495: + \alpha_{23}^2\alpha_{34}\alpha_{45}\alpha_{12} +

2496: \alpha_{34}^2\alpha_{45}\alpha_{51}\alpha_{23} +

2497: \alpha_{45}^2\alpha_{51}\alpha_{12}\alpha_{34} +

2498: \alpha_{51}^2\alpha_{12}\alpha_{23}\alpha_{45}  \ ,

2499: \nonumber \\

2500: I_{24}^{(5)} & = &  \alpha_{12}^2\alpha_{23}\alpha_{45}\alpha_{51}

2501: + \alpha_{23}^2\alpha_{34}\alpha_{51}\alpha_{12} +

2502: \alpha_{34}^2\alpha_{45}\alpha_{12}\alpha_{23} +

2503: \alpha_{45}^2\alpha_{51}\alpha_{23}\alpha_{34} +

2504: \alpha_{51}^2\alpha_{12}\alpha_{34}\alpha_{45}  \ ,

2505: \nonumber \\

2506: I_{25}^{(5)} & = &  \alpha_{12}^2\alpha_{34}\alpha_{45}\alpha_{51}

2507: + \alpha_{23}^2\alpha_{45}\alpha_{51}\alpha_{12} +

2508: \alpha_{34}^2\alpha_{51}\alpha_{12}\alpha_{23} +

2509: \alpha_{45}^2\alpha_{12}\alpha_{23}\alpha_{34} +

2510: \alpha_{51}^2\alpha_{23}\alpha_{34}\alpha_{45}  \ ,

2511: \nonumber \\

2512: I_{26}^{(5)} & = &

2513: \alpha_{12}\alpha_{23}\alpha_{34}\alpha_{45}\alpha_{51} \ .

2514: \nonumber

2515: \end{eqnarray}

2516: \item Degree 6:

2517: \begin{eqnarray}

2518: I_{1}^{(6)} & = &  \alpha_{12}^6 + \alpha_{23}^6 + \alpha_{34}^6 +

2519: \alpha_{45}^6 + \alpha_{51}^6 \ ,

2520: \nonumber \\

2521: I_{2}^{(6)} & = &  \alpha_{12}^5 \alpha_{23} + \alpha_{23}^5

2522: \alpha_{34}  + \alpha_{34}^5 \alpha_{45}  + \alpha_{45}^5

2523: \alpha_{51}  + \alpha_{51}^5 \alpha_{12}  \ ,

2524: \nonumber \\

2525: I_{3}^{(6)} & = &  \alpha_{12}^5 \alpha_{34} + \alpha_{23}^5

2526: \alpha_{45}  + \alpha_{34}^5 \alpha_{51}  + \alpha_{45}^5

2527: \alpha_{12}  + \alpha_{51}^5\alpha_{23}  \ ,

2528: \nonumber \\

2529: I_{4}^{(6)} & = &  \alpha_{12}^5 \alpha_{45} + \alpha_{23}^5

2530: \alpha_{51}  + \alpha_{34}^5 \alpha_{12}  + \alpha_{45}^5

2531: \alpha_{23}  + \alpha_{51}^5\alpha_{34}  \ ,

2532: \nonumber \\

2533: I_{5}^{(6)} & = &  \alpha_{12}^5 \alpha_{51} + \alpha_{23}^5

2534: \alpha_{12}  + \alpha_{34}^5 \alpha_{23}  + \alpha_{45}^5

2535: \alpha_{34}  + \alpha_{51}^5 \alpha_{45}  \ .

2536: \nonumber \\

2537: I_{6}^{(6)} & = &  \alpha_{12}^4 \alpha_{23}^2 + \alpha_{23}^4

2538: \alpha_{34}^2  + \alpha_{34}^4 \alpha_{45}^2  + \alpha_{45}^4

2539: \alpha_{51}^2  + \alpha_{51}^4 \alpha_{12}^2  \ ,

2540: \nonumber \\

2541: I_{7}^{(6)} & = &  \alpha_{12}^4 \alpha_{34}^2 + \alpha_{23}^4

2542: \alpha_{45}^2  + \alpha_{34}^4 \alpha_{51}^2  + \alpha_{45}^4

2543: \alpha_{12}^2  + \alpha_{51}^4 \alpha_{23}^2  \ ,

2544: \nonumber \\

2545: I_{8}^{(6)} & = &  \alpha_{12}^4 \alpha_{45}^2 + \alpha_{23}^4

2546: \alpha_{51}^2  + \alpha_{34}^4 \alpha_{12}^2  + \alpha_{45}^4

2547: \alpha_{23}^2  + \alpha_{51}^4 \alpha_{34}^2  \ ,

2548: \nonumber \\

2549: I_{9}^{(6)} & = &  \alpha_{12}^4 \alpha_{51}^2 + \alpha_{23}^4

2550: \alpha_{12}^2  + \alpha_{34}^4 \alpha_{23}^2  + \alpha_{45}^4

2551: \alpha_{34}^2  + \alpha_{51}^4 \alpha_{45}^2  \ ,

2552: \nonumber \\

2553: I_{10}^{(6)} & = &  \alpha_{12}^4\alpha_{23}\alpha_{34} +

2554: \alpha_{23}^4\alpha_{34}\alpha_{45}   +

2555: \alpha_{34}^4\alpha_{45}\alpha_{51}  +

2556: \alpha_{45}^4\alpha_{51}\alpha_{12}  +

2557: \alpha_{51}^4\alpha_{12}\alpha_{23}   \ , \nonumber

2558: \end{eqnarray}

2559: \begin{eqnarray}

2560: I_{11}^{(6)} & = &  \alpha_{12}^4\alpha_{23}\alpha_{45} +

2561: \alpha_{23}^4\alpha_{34}\alpha_{51}  +

2562: \alpha_{34}^4\alpha_{45}\alpha_{12}  +

2563: \alpha_{45}^4\alpha_{51}\alpha_{23}  +

2564: \alpha_{51}^4\alpha_{12}\alpha_{34}   \ ,

2565: \nonumber \\

2566: I_{12}^{(6)} & = &  \alpha_{12}^4\alpha_{23}\alpha_{51} +

2567: \alpha_{23}^4\alpha_{34}\alpha_{12}  +

2568: \alpha_{34}^4\alpha_{45}\alpha_{23}  +

2569: \alpha_{45}^4\alpha_{51}\alpha_{34}  +

2570: \alpha_{51}^4\alpha_{12}\alpha_{45}   \ ,

2571: \nonumber \\

2572: I_{13}^{(6)} & = &  \alpha_{12}^4\alpha_{34}\alpha_{45} +

2573: \alpha_{23}^4\alpha_{45}\alpha_{51}  +

2574: \alpha_{34}^4\alpha_{51}\alpha_{12}  +

2575: \alpha_{45}^4\alpha_{12}\alpha_{23}  +

2576: \alpha_{51}^4\alpha_{23}\alpha_{34}   \ ,

2577: \nonumber \\

2578: I_{14}^{(6)} & = &  \alpha_{12}^4\alpha_{34}\alpha_{51} +

2579: \alpha_{23}^4\alpha_{45}\alpha_{12}  +

2580: \alpha_{34}^4\alpha_{51}\alpha_{23}  +

2581: \alpha_{45}^4\alpha_{12}\alpha_{34}  +

2582: \alpha_{51}^4\alpha_{23}\alpha_{45}   \ ,

2583: \nonumber \\

2584: I_{15}^{(6)} & = &  \alpha_{12}^4\alpha_{45}\alpha_{51} +

2585: \alpha_{23}^4\alpha_{51}\alpha_{12}  +

2586: \alpha_{34}^4\alpha_{12}\alpha_{23}  +

2587: \alpha_{45}^4\alpha_{23}\alpha_{34}  +

2588: \alpha_{51}^4\alpha_{34}\alpha_{45}   \ ,

2589: \nonumber \\

2590: I_{16}^{(6)} & = &  \alpha_{12}^3 \alpha_{23}^3 + \alpha_{23}^3

2591: \alpha_{34}^3  + \alpha_{34}^3 \alpha_{45}^3  + \alpha_{45}^3

2592: \alpha_{51}^3  + \alpha_{51}^3 \alpha_{12}^3  \ ,

2593: \nonumber \\

2594: I_{17}^{(6)} & = &  \alpha_{12}^3 \alpha_{34}^3 + \alpha_{23}^3

2595: \alpha_{45}^3  + \alpha_{34}^3 \alpha_{51}^3  + \alpha_{45}^3

2596: \alpha_{12}^3  + \alpha_{51}^3 \alpha_{23}^3  \ ,

2597: \nonumber \\

2598: I_{18}^{(6)} & = &  \alpha_{12}^3\alpha_{23}^2\alpha_{34} +

2599: \alpha_{23}^3\alpha_{34}^2\alpha_{45} +

2600: \alpha_{34}^3\alpha_{45}^2\alpha_{51}  +

2601: \alpha_{45}^3\alpha_{51}^2\alpha_{12}  +

2602: \alpha_{51}^3\alpha_{12}^2\alpha_{23}   \ ,

2603: \nonumber \\

2604: I_{19}^{(6)} & = &  \alpha_{12}^3\alpha_{23}^2\alpha_{45} +

2605: \alpha_{23}^3\alpha_{34}^2\alpha_{51} +

2606: \alpha_{34}^3\alpha_{45}^2\alpha_{12}  +

2607: \alpha_{45}^3\alpha_{51}^2\alpha_{23} +

2608: \alpha_{51}^3\alpha_{12}^2\alpha_{34}   \ ,

2609: \nonumber \\

2610: I_{20}^{(6)} & = &  \alpha_{12}^3\alpha_{23}^2\alpha_{51} +

2611: \alpha_{23}^3\alpha_{34}^2\alpha_{12} +

2612: \alpha_{34}^3\alpha_{45}^2\alpha_{23}  +

2613: \alpha_{45}^3\alpha_{51}^2\alpha_{34} +

2614: \alpha_{51}^3\alpha_{12}^2\alpha_{45}   \ ,

2615: \nonumber \\

2616: I_{21}^{(6)} & = &  \alpha_{12}^3\alpha_{34}^2\alpha_{23} +

2617: \alpha_{23}^3\alpha_{45}^2\alpha_{34} +

2618: \alpha_{34}^3\alpha_{51}^2\alpha_{45}  +

2619: \alpha_{45}^3\alpha_{12}^2\alpha_{51} +

2620: \alpha_{51}^3\alpha_{23}^2\alpha_{12}   \ ,

2621: \nonumber \\

2622: I_{22}^{(6)} & = &  \alpha_{12}^3\alpha_{34}^2\alpha_{45} +

2623: \alpha_{23}^3\alpha_{45}^2\alpha_{51} +

2624: \alpha_{34}^3\alpha_{51}^2\alpha_{12}  +

2625: \alpha_{45}^3\alpha_{12}^2\alpha_{23} +

2626: \alpha_{51}^3\alpha_{23}^2\alpha_{34}   \ ,

2627: \label{Icyclic6} \\

2628: I_{23}^{(6)} & = &  \alpha_{12}^3\alpha_{34}^2\alpha_{51} +

2629: \alpha_{23}^3\alpha_{45}^2\alpha_{12} +

2630: \alpha_{34}^3\alpha_{51}^2\alpha_{23}  +

2631: \alpha_{45}^3\alpha_{12}^2\alpha_{34} +

2632: \alpha_{51}^3\alpha_{23}^2\alpha_{45}   \ ,

2633: \nonumber \\

2634: I_{24}^{(6)} & = &  \alpha_{12}^3\alpha_{45}^2\alpha_{23} +

2635: \alpha_{23}^3\alpha_{51}^2\alpha_{34} +

2636: \alpha_{34}^3\alpha_{12}^2\alpha_{45}  +

2637: \alpha_{45}^3\alpha_{23}^2\alpha_{51} +

2638: \alpha_{51}^3\alpha_{34}^2\alpha_{12}   \ ,

2639: \nonumber \\

2640: I_{25}^{(6)} & = &  \alpha_{12}^3\alpha_{45}^2\alpha_{34} +

2641: \alpha_{23}^3\alpha_{51}^2\alpha_{45} +

2642: \alpha_{34}^3\alpha_{12}^2\alpha_{51}  +

2643: \alpha_{45}^3\alpha_{23}^2\alpha_{12} +

2644: \alpha_{51}^3\alpha_{34}^2\alpha_{23}   \ ,

2645: \nonumber \\

2646: I_{26}^{(6)} & = &  \alpha_{12}^3\alpha_{45}^2\alpha_{51} +

2647: \alpha_{23}^3\alpha_{51}^2\alpha_{12} +

2648: \alpha_{34}^3\alpha_{12}^2\alpha_{23}  +

2649: \alpha_{45}^3\alpha_{23}^2\alpha_{34} +

2650: \alpha_{51}^3\alpha_{34}^2\alpha_{45}   \ ,

2651: \nonumber\\

2652: I_{27}^{(6)} & = &  \alpha_{12}^3\alpha_{51}^2\alpha_{23} +

2653: \alpha_{23}^3\alpha_{12}^2\alpha_{34} +

2654: \alpha_{34}^3\alpha_{23}^2\alpha_{45}  +

2655: \alpha_{45}^3\alpha_{34}^2\alpha_{51} +

2656: \alpha_{51}^3\alpha_{45}^2\alpha_{12}   \ ,

2657: \nonumber\\

2658: I_{28}^{(6)} & = & \alpha_{12}^3\alpha_{51}^2\alpha_{34} +

2659: \alpha_{23}^3\alpha_{12}^2\alpha_{45} +

2660: \alpha_{34}^3\alpha_{23}^2\alpha_{51}  +

2661: \alpha_{45}^3\alpha_{34}^2\alpha_{12} +

2662: \alpha_{51}^3\alpha_{45}^2\alpha_{23}    \ ,

2663: \nonumber\\

2664: I_{29}^{(6)} & = &  \alpha_{12}^3\alpha_{51}^2\alpha_{45} +

2665: \alpha_{23}^3\alpha_{12}^2\alpha_{51} +

2666: \alpha_{34}^3\alpha_{23}^2\alpha_{12}  +

2667: \alpha_{45}^3\alpha_{34}^2\alpha_{23} +

2668: \alpha_{51}^3\alpha_{45}^2\alpha_{34}   \ ,

2669: \nonumber\\

2670: I_{30}^{(6)} & = &  \alpha_{12}^3\alpha_{23}\alpha_{34}\alpha_{45}

2671: + \alpha_{23}^3\alpha_{34}\alpha_{45}\alpha_{51} +

2672: \alpha_{34}^3\alpha_{45}\alpha_{51}\alpha_{12} +

2673: \alpha_{45}^3\alpha_{51}\alpha_{12}\alpha_{23} +

2674: \alpha_{51}^3\alpha_{12}\alpha_{23}\alpha_{34}   \ ,

2675: \nonumber\\

2676: I_{31}^{(6)} & = & \alpha_{12}^3\alpha_{23}\alpha_{34}\alpha_{51}

2677: + \alpha_{23}^3\alpha_{34}\alpha_{45}\alpha_{12} +

2678: \alpha_{34}^3\alpha_{45}\alpha_{51}\alpha_{23} +

2679: \alpha_{45}^3\alpha_{51}\alpha_{12}\alpha_{34} +

2680: \alpha_{51}^3\alpha_{12}\alpha_{23}\alpha_{45}   \ ,

2681: \nonumber\\

2682: I_{32}^{(6)} & = & \alpha_{12}^3\alpha_{23}\alpha_{45}\alpha_{51}

2683: + \alpha_{23}^3\alpha_{34}\alpha_{51}\alpha_{12} +

2684: \alpha_{34}^3\alpha_{45}\alpha_{12}\alpha_{23} +

2685: \alpha_{45}^3\alpha_{51}\alpha_{23}\alpha_{34} +

2686: \alpha_{51}^3\alpha_{12}\alpha_{34}\alpha_{45}    \ ,

2687: \nonumber\\

2688: I_{33}^{(6)} & = & \alpha_{12}^3\alpha_{34}\alpha_{45}\alpha_{51}

2689: + \alpha_{23}^3\alpha_{45}\alpha_{51}\alpha_{12} +

2690: \alpha_{34}^3\alpha_{51}\alpha_{12}\alpha_{23} +

2691: \alpha_{45}^3\alpha_{12}\alpha_{23}\alpha_{34} +

2692: \alpha_{51}^3\alpha_{23}\alpha_{34}\alpha_{45}     \ ,

2693: \nonumber\\

2694: I_{34}^{(6)} & = &  \alpha_{12}^2\alpha_{23}^2\alpha_{34}^2 +

2695: \alpha_{23}^2\alpha_{34}^2\alpha_{45}^2 +

2696: \alpha_{34}^2\alpha_{45}^2\alpha_{51}^2  +

2697: \alpha_{45}^2\alpha_{51}^2\alpha_{12}^2 +

2698: \alpha_{51}^2\alpha_{12}^2\alpha_{23}^2   \ ,

2699: \nonumber\\

2700: I_{35}^{(6)} & = &  \alpha_{12}^2\alpha_{23}^2\alpha_{45}^2 +

2701: \alpha_{23}^2\alpha_{34}^2\alpha_{51}^2 +

2702: \alpha_{34}^2\alpha_{45}^2\alpha_{12}^2  +

2703: \alpha_{45}^2\alpha_{51}^2\alpha_{23}^2 +

2704: \alpha_{51}^2\alpha_{12}^2\alpha_{34}^2    \ ,

2705: \nonumber\\

2706: I_{36}^{(6)} & = &

2707: \alpha_{12}^2\alpha_{23}^2\alpha_{34}\alpha_{45} +

2708: \alpha_{23}^2\alpha_{34}^2\alpha_{45}\alpha_{51} +

2709: \alpha_{34}^2\alpha_{45}^2\alpha_{51}\alpha_{12} +

2710: \alpha_{45}^2\alpha_{51}^2\alpha_{12}\alpha_{23} +

2711: \alpha_{51}^2\alpha_{12}^2\alpha_{23}\alpha_{34}   \ ,

2712: \nonumber\\

2713: I_{37}^{(6)} & = &

2714: \alpha_{12}^2\alpha_{23}^2\alpha_{34}\alpha_{51} +

2715: \alpha_{23}^2\alpha_{34}^2\alpha_{45}\alpha_{12} +

2716: \alpha_{34}^2\alpha_{45}^2\alpha_{51}\alpha_{23} +

2717: \alpha_{45}^2\alpha_{51}^2\alpha_{12}\alpha_{34} +

2718: \alpha_{51}^2\alpha_{12}^2\alpha_{23}\alpha_{45}   \ ,

2719: \nonumber\\

2720: I_{38}^{(6)} & = &

2721: \alpha_{12}^2\alpha_{23}^2\alpha_{45}\alpha_{51} +

2722: \alpha_{23}^2\alpha_{34}^2\alpha_{51}\alpha_{12} +

2723: \alpha_{34}^2\alpha_{45}^2\alpha_{12}\alpha_{23} +

2724: \alpha_{45}^2\alpha_{51}^2\alpha_{23}\alpha_{34} +

2725: \alpha_{51}^2\alpha_{12}^2\alpha_{34}\alpha_{45}   \ ,

2726: \nonumber\\

2727: I_{39}^{(6)} & = &

2728: \alpha_{12}^2\alpha_{34}^2\alpha_{23}\alpha_{45} +

2729: \alpha_{23}^2\alpha_{45}^2\alpha_{34}\alpha_{51} +

2730: \alpha_{34}^2\alpha_{51}^2\alpha_{45}\alpha_{12} +

2731: \alpha_{45}^2\alpha_{12}^2\alpha_{51}\alpha_{23} +

2732: \alpha_{51}^2\alpha_{23}^2\alpha_{12}\alpha_{34}  \ ,

2733: \nonumber \\

2734: I_{40}^{(6)} & = &

2735: \alpha_{12}^2\alpha_{34}^2\alpha_{23}\alpha_{51} +

2736: \alpha_{23}^2\alpha_{45}^2\alpha_{34}\alpha_{12} +

2737: \alpha_{34}^2\alpha_{51}^2\alpha_{45}\alpha_{23} +

2738: \alpha_{45}^2\alpha_{12}^2\alpha_{51}\alpha_{34} +

2739: \alpha_{51}^2\alpha_{23}^2\alpha_{12}\alpha_{45}  \ ,

2740: \nonumber \\

2741: I_{41}^{(6)} & = &

2742: \alpha_{12}^2\alpha_{34}^2\alpha_{45}\alpha_{51} +

2743: \alpha_{23}^2\alpha_{45}^2\alpha_{51}\alpha_{12} +

2744: \alpha_{34}^2\alpha_{51}^2\alpha_{12}\alpha_{23} +

2745: \alpha_{45}^2\alpha_{12}^2\alpha_{23}\alpha_{34} +

2746: \alpha_{51}^2\alpha_{23}^2\alpha_{34}\alpha_{45}  \ ,

2747: \nonumber \\

2748: I_{42}^{(6)} & = &

2749: \alpha_{12}^2\alpha_{23}\alpha_{34}\alpha_{45}\alpha_{51} +

2750: \alpha_{23}^2\alpha_{34}\alpha_{45}\alpha_{51}\alpha_{12} +

2751: \alpha_{34}^2\alpha_{45}\alpha_{51}\alpha_{12}\alpha_{23} +

2752: \alpha_{45}^2\alpha_{51}\alpha_{12}\alpha_{23}\alpha_{34} +

2753: \nonumber \\

2754: & & + \alpha_{51}^2\alpha_{12}\alpha_{23}\alpha_{34}\alpha_{45} \

2755: .

2756: \nonumber

2757: \end{eqnarray}

2758: \end{itemize}

2759: 

2760: 

2761: \subsection{$H^{(1)}$, $P^{(1)}$, $U^{(1)}$, $W^{(1)}$, $Z^{(1)}$ and

2762: $\Delta$ factors}

2763: \label{otherfactors}

2764: 

2765: The $\alpha'$ dependent factors of formula (\ref{AD2nF5nopoles}),

2766: except for $Z^{(1)}$, are all given explicitly in terms of the

2767: function $f$ of (\ref{f}), the factor $K_3$ of (\ref{K3-int}) and

2768: the factor $T$ of (\ref{K2-K})\footnote{In this formula, the factors

2769: $H^{(j)}$ (with $j=2,3,4,5$) are constructed by means of cyclic

2770: permutations of the factor $H^{(1)}$. For example, to obtain the

2771: expression for $H^{(2)}$ we make the following change of indexes in

2772: the $\alpha_{ij}$ variables: $(1,2,3,4,5) \rightarrow

2773: (2,3,4,5,1)$.}:

2774: \begin{eqnarray}

2775: \label{H23} H^{(1)} &=& - (2 \alpha')^2 \frac{f(-2 \alpha_{23}, -2

2776: \alpha_{34})-f(-2 \alpha_{34}, -2

2777: \alpha_{45})}{\alpha_{23}-\alpha_{45}} \ ,

2778: \\

2779: \label{P34} P^{(1)} &=&  (2 \alpha')^2 \frac{ \ \ K_3 -

2780: K_3\biggl|_{\alpha_{34}=0}}{\alpha_{34}}  \ ,

2781: \\

2782: \label{U34} U^{(1)} &=& \frac{(2 \alpha')^2\biggl\{f(-2 \alpha_{51},

2783: -2 \alpha_{12}) + \ K_3\biggl|_{\alpha_{34}=0} \biggr\}  - \

2784: \alpha_{45}H^{(3)} \ - \ \alpha_{23}H^{(4)}}

2785: {\alpha_{23}\alpha_{45}} \ ,

2786: \\

2787: \label{W51} W^{(1)} &=& \frac{(2 \alpha')^2 \biggl[K_3 -

2788: K_3\biggl|_{\alpha_{51}=0} - K_3\biggl|_{\alpha_{12}=0}

2789: \biggr]+\alpha_{23}H^{(1)}- (2 \alpha')^2 f(-2 \alpha_{34}, -2

2790: \alpha_{45}) \ }{\alpha_{51}\alpha_{12}} \ ,

2791: \\

2792: \label{Z23} Z^{(1)} &=& (2 \alpha')^2  \times \nonumber

2793: \\

2794: && \times \frac{G(-2 \alpha_{34}, -2

2795: \alpha_{45},-\alpha_{25}-\alpha_{34},-\alpha_{23}-\alpha_{45})-G(-2

2796: \alpha_{34}, -2

2797: \alpha_{23},-\alpha_{25}-\alpha_{34},-\alpha_{23}-\alpha_{45})}{\alpha_{23}-\alpha_{45}}

2798: \ , \nonumber

2799: \\

2800: \\

2801: \label{Delta} \Delta &=&  \frac{ \ T -1 - \biggl[\frac{}{}\biggl\{

2802: T\biggl|_{\alpha_{23}=0}- \ 1 \ - \

2803: \alpha_{12}\alpha_{23}\alpha_{34}H^{(5)} \biggr\}+(\mbox{cyclic

2804: permutations})\frac{}{}\biggr]

2805: }{\alpha_{12}\alpha_{23}\alpha_{34}\alpha_{45}\alpha_{51}} \ ,

2806: \end{eqnarray}

2807: where the function $G(a,b,c,d)$ that appears in (\ref{Z23}) is

2808: defined by the $\alpha'$ series

2809: \begin{eqnarray}

2810: G(a,b,c,d) = \sum_{m=1}^{\infty} \sum_{n=0}^{\infty} \frac{ \

2811: a_{mn} \ {\alpha'}^{m+n}}{\left(\begin{array}{c}

2812:                           m+n \\

2813:                           m

2814:                           \end{array} \right)}

2815: \sum_{p=0}^{m-1} \sum_{q=0}^{n} \left(\begin{array}{c}

2816:                           m-1-p+n-q \\

2817:                           m-1-p

2818:                           \end{array} \right)

2819:                           \left(\begin{array}{c}

2820:                           p+q \\

2821:                           p

2822:                           \end{array} \right)

2823: a^{m-1-p} \ b^{n-q} \ c^p \ d^q \ . \nonumber \\

2824: \label{G}

2825: \end{eqnarray}

2826: Here, the (completely known) coefficients $a_{mn}$ are the ones that

2827: appear in the expansion of the function $f(s,t)$, in (\ref{fpower}).

2828: May be it is possible to write the function $G(a,b,c,d)$ in a closed

2829: form (in terms of Hypergeometric functions, for

2830: example), but we have not succeeded in doing so.\\

2831: \noindent $G(a,b,c,d)$ begins like

2832: \begin{eqnarray}

2833: G(a,b,c,d) & = & -\zeta(3) \ \alpha' - \frac{ \ \pi^4}{720}

2834: \biggl( \ 8a + 8c + b + d  \biggr) \

2835: {\alpha'}^2 + \nonumber \\

2836: && + \biggl( -\frac{1}{3} \zeta(5) \

2837: (3a^2+3ac+3c^2+2b^2+2bd+2d^2+4ab+2ad+2bc+4cd) + \nonumber \\

2838: && \hphantom{ + \biggl( }+ \frac{ \ \pi^2}{18}\zeta(3)

2839: (b^2+bd+d^2+2ab+ad+bc+2cd) \biggr) \ {\alpha'}^3 \ +{\cal

2840: O}({\alpha'}^4). \label{Gpower}

2841: \end{eqnarray}

2842: \noindent Using eqs. (\ref{K3hyperg}), (\ref{K2-K}) and

2843: (\ref{K2hyperg}), the $\alpha'$ factors defined in equations

2844: (\ref{H23})-(\ref{Z23}) can all be written in terms of the Gamma

2845: factor, Beta and Hypergeometric functions. It is also not

2846: difficult to see that these factors are all invariant under a

2847: twisting transformation with respect to index $1$. The factor

2848: $\Delta$ in (\ref{Delta}) is invariant under a twisting

2849: transformation with respect to any of the five

2850: indexes and it is also cyclic invariant.\\

2851: \noindent Formulas

2852: (\ref{H23}), (\ref{P34}) and (\ref{Z23}) can be understood to have

2853: no poles since the numerator is a power series of the $\alpha_{ij}$

2854: and, whenever the denominator is zero, so happens with the

2855: numerator. This means that, in each of those

2856: cases, the numerator is divisible by the corresponding denominator.\\

2857: \noindent In the case of formulas (\ref{U34}), (\ref{W51}) and

2858: (\ref{Delta}) it is not straight forward to see that those

2859: expressions contain no poles, but the argument is similar: in all

2860: cases the numerator involves a power series of the $\alpha_{ij}$

2861: and, whenever the denominator is zero, so happens with the

2862: numerator. For example, in the case of (\ref{U34}) and (\ref{W51}),

2863: using (\ref{K3Gamma}) (and one of the cyclic permutations of that

2864: relation) it can be proved that the numerator becomes zero if any of

2865: the $\alpha_{ij}$ of the denominator are zero. In the case of

2866: (\ref{Delta}), using (\ref{TGamma}) and (\ref{T0}) (and the cyclic

2867: permutations of them) it can be proved that if any of the five

2868: $\alpha_{ij}$ of the denominator is zero,

2869: so happens with the numerator.\\

2870: \noindent Up to ${\cal O}({\alpha'}^6)$ order, the expansions of the

2871: factors defined in (\ref{H23})-(\ref{Delta}) are given by

2872: \begin{eqnarray}

2873: \label{H23exp} H^{(1)} &=&  \ -\zeta(3)(2

2874: \alpha')^3+\frac{\pi^4}{360}\biggl[4\alpha_{45}+4\alpha_{23}+\alpha_{34}\biggr](2

2875: \alpha')^4+ \biggl[\frac{\

2876: \pi^2}{6}\zeta(3)\alpha_{34}\biggl(\alpha_{34}+\alpha_{45}+\alpha_{23}\biggr)-

2877: \nonumber

2878: \\

2879: && \ \ \ \

2880: -\zeta(5)\biggl(2\alpha_{34}\alpha_{23}+\alpha_{23}^2+2\alpha_{45}\alpha_{34}+2\alpha_{34}^2

2881: +\alpha_{45}^2+\alpha_{45}\alpha_{23}\biggr)\biggr](2 \alpha')^5 +

2882: \nonumber \\

2883: && + \biggl[ -\frac{1}{2}\zeta(3)^2

2884: \alpha_{34}\biggl(2\alpha_{23}\alpha_{34}+\alpha_{45}\alpha_{23}+2\alpha_{45}\alpha_{34}

2885: +\alpha_{45}^2+\alpha_{23}^2+\alpha_{34}^2\biggr) + \frac{ \

2886: \pi^6}{15120}\biggl(16\alpha_{45}\alpha_{23}^2+\nonumber

2887: \\

2888: &&\hphantom{+ \biggl[ } +12\alpha_{34}\alpha_{23}^2

2889: +12\alpha_{45}\alpha_{34}\alpha_{23}+16\alpha_{23}\alpha_{45}^2+23\alpha_{23}\alpha_{34}^2

2890: +23\alpha_{45}\alpha_{34}^2+16\alpha_{45}^3+12\alpha_{34}^3+

2891: \nonumber \\

2892: &&\hphantom{+ \biggl[ } +16\alpha_{23}^3

2893: +12\alpha_{34}\alpha_{45}^2\biggr) \biggr](2 \alpha')^6 + \ {\cal

2894: O}( \ (2 \alpha')^7 \ )  \ ,

2895: \\

2896: \label{P34exp} P^{(1)} &=& -\zeta(3)(2

2897: \alpha')^3+\frac{\pi^4}{360}\biggl[4\alpha_{34}+\alpha_{45}+\alpha_{23}

2898: +4\alpha_{51}+4\alpha_{12}\biggr] (2 \alpha')^4+  \biggl[\frac{\

2899: \pi^2}{6}\zeta(3)\biggl(\alpha_{23}\alpha_{45}-  \nonumber

2900: \\

2901: && \ \ \ \ -2\alpha_{12}\alpha_{45}+\alpha_{45}^2

2902: +\alpha_{23}^2+\alpha_{34}\alpha_{45}+\alpha_{12}\alpha_{23}-2\alpha_{51}\alpha_{23}

2903: +\alpha_{23}\alpha_{34}+\alpha_{45}\alpha_{51}-2\alpha_{51}\alpha_{12}\biggr)-

2904: \nonumber

2905: \\

2906: && \ \ \ \ -\frac{1}{2}\zeta(5)\biggl(2\alpha_{34}^2+

2907: 2\alpha_{12}^2+4\alpha_{23}^2+2\alpha_{12}\alpha_{34}+2\alpha_{51}\alpha_{34}

2908: +2\alpha_{51}^2 +4\alpha_{45}^2+4\alpha_{23}\alpha_{34}+\nonumber

2909: \\

2910: && \ \ \ \

2911: +4\alpha_{12}\alpha_{23}-7\alpha_{12}\alpha_{45}-7\alpha_{51}\alpha_{23}

2912: +4\alpha_{23}\alpha_{45}+4\alpha_{45}\alpha_{51}-7\alpha_{51}\alpha_{12}+\nonumber

2913: \\

2914: &&\ \ \ \ +4\alpha_{34}\alpha_{45}\biggr)\biggr](2 \alpha')^5

2915: +\biggl[

2916: \frac{1}{2}\zeta(3)^2\biggl(2\alpha_{51}^2\alpha_{23}+2\alpha_{12}^2\alpha_{45}

2917: -\alpha_{23}^2\alpha_{45}+\alpha_{23}^2\alpha_{51}-\alpha_{51}^2\alpha_{45}

2918: -\nonumber \\

2919: && \ \ \

2920: -\alpha_{34}^2\alpha_{45}-\alpha_{34}^2\alpha_{23}+\alpha_{12}^2\alpha_{51}

2921: -2\alpha_{12}\alpha_{23}^2-2\alpha_{45}^2\alpha_{51}-2\alpha_{23}^2\alpha_{34}

2922: -2\alpha_{45}^2\alpha_{34}+\alpha_{45}^2\alpha_{12}+\nonumber \\

2923: && \ \ \ +\alpha_{51}^2\alpha_{12}

2924: -\alpha_{45}^2\alpha_{23}-\alpha_{12}^2\alpha_{23}-\alpha_{45}^3-\alpha_{23}^3

2925: +2\alpha_{12}\alpha_{34}\alpha_{51}-2\alpha_{23}\alpha_{34}\alpha_{45}

2926: +\alpha_{12}\alpha_{45}\alpha_{34}+\nonumber

2927: \\

2928: && \ \ \ +4\alpha_{51}\alpha_{12}\alpha_{23}

2929: +\alpha_{51}\alpha_{23}\alpha_{34}+4\alpha_{12}\alpha_{23}\alpha_{45}

2930: +4\alpha_{45}\alpha_{51}\alpha_{23}-\alpha_{12}\alpha_{23}\alpha_{34}

2931: -\alpha_{34}\alpha_{45}\alpha_{51}+\nonumber

2932: \\

2933: && \ \ \ +4\alpha_{45}\alpha_{51}\alpha_{12}\biggr) + \frac{ \

2934: \pi^6}{45360}\biggl(-52\alpha_{51}^2\alpha_{23}-52\alpha_{12}^2\alpha_{45}

2935: +36\alpha_{23}^2\alpha_{45}-28\alpha_{23}^2\alpha_{51}+36\alpha_{51}^2\alpha_{45}

2936: +\nonumber \\

2937: && \ \ \

2938: +36\alpha_{34}^2\alpha_{45}+36\alpha_{34}^2\alpha_{23}-28\alpha_{12}^2\alpha_{51}

2939: +69\alpha_{12}\alpha_{23}^2+69\alpha_{45}^2\alpha_{51}+69\alpha_{23}^2\alpha_{34}

2940: +69\alpha_{45}^2\alpha_{34}-\nonumber

2941: \\

2942: && \ \ \ -28\alpha_{45}^2\alpha_{12}-28\alpha_{51}^2\alpha_{12}

2943: +36\alpha_{45}^2\alpha_{23}+36\alpha_{12}^2\alpha_{23}+48\alpha_{34}^3+36\alpha_{45}^3

2944: +36\alpha_{23}^3+48\alpha_{51}^3+\nonumber

2945: \\

2946: && \ \ \ +48\alpha_{12}^3+48\alpha_{34}\alpha_{51}^2

2947: -52\alpha_{12}\alpha_{34}\alpha_{51}+69\alpha_{23}\alpha_{34}\alpha_{45}

2948: -28\alpha_{12}\alpha_{45}\alpha_{34}-134\alpha_{51}\alpha_{12}\alpha_{23}

2949: -\nonumber \\

2950: && \ \ \

2951: -28\alpha_{51}\alpha_{23}\alpha_{34}-134\alpha_{12}\alpha_{23}\alpha_{45}

2952: -134\alpha_{45}\alpha_{51}\alpha_{23}+36\alpha_{12}\alpha_{23}\alpha_{34}

2953: +36\alpha_{34}\alpha_{45}\alpha_{51}-\nonumber

2954: \\

2955: && \ \ \ -134\alpha_{45}\alpha_{51}\alpha_{12}

2956: +48\alpha_{51}\alpha_{34}^2+48\alpha_{12}\alpha_{34}^2+48\alpha_{34}\alpha_{12}^2\biggr)\biggr](2

2957: \alpha')^6+{\cal O}(  \ (2 \alpha')^7 \ )  \ ,

2958: \end{eqnarray}

2959: \begin{eqnarray}

2960: \label{U34exp} U^{(1)} &=& \frac{\pi^4}{90}(2

2961: \alpha')^4+\biggl[-\frac{\ \pi^2}{3}

2962: \zeta(3)\biggl(\alpha_{51}+\alpha_{12}\biggr)

2963: +\frac{1}{2}\zeta(5)\biggl(7\alpha_{51}+7\alpha_{12}-2\alpha_{45}

2964: -2\alpha_{23}\biggr) \biggr](2 \alpha')^5 + \nonumber

2965: \\

2966: &&+ \biggl[

2967: \frac{1}{2}\zeta(3)^2\biggl(4\alpha_{51}\alpha_{12}+\alpha_{51}^2+\alpha_{12}^2

2968: +2\alpha_{51}\alpha_{23}+2\alpha_{45}\alpha_{12}+\alpha_{45}\alpha_{51}

2969: +\alpha_{12}\alpha_{23}\biggr) + \nonumber \\

2970: && \ \ \ + \frac{ \

2971: \pi^6}{22680}\biggl(-14\alpha_{45}\alpha_{51}+24\alpha_{23}^2-26\alpha_{45}\alpha_{12}

2972: -26\alpha_{51}\alpha_{23}+24\alpha_{45}^2-67\alpha_{51}\alpha_{12}

2973: -\nonumber \\

2974: && \ \ \ -14\alpha_{12}^2

2975: +24\alpha_{45}\alpha_{23}-14\alpha_{51}^2-14\alpha_{12}\alpha_{23}\biggr)\biggr](2

2976: \alpha')^6 + \ {\cal O}( \ (2 \alpha')^7 \ ) \ ,

2977: \\

2978: \label{W51exp} W^{(1)} &=& \frac{\pi^4}{360}(2

2979: \alpha')^4+\biggl[\frac{ \

2980: \pi^2}{6}\zeta(3)\biggl(\alpha_{45}+\alpha_{23}-2\alpha_{34}

2981: +\alpha_{51}+\alpha_{12}\biggr)

2982: +\frac{1}{2}\zeta(5)\biggl(7\alpha_{34}-4\alpha_{45}-\nonumber

2983: \\

2984: && \ \ \ -4\alpha_{23} -4\alpha_{51}-4\alpha_{12}\biggr)\biggr](2

2985: \alpha')^5 + \biggl[

2986: -\frac{1}{2}\zeta(3)^2\biggl(\alpha_{45}^2+2\alpha_{45}\alpha_{51}

2987: +\alpha_{51}^2+\alpha_{12}^2+\alpha_{23}^2-\nonumber\\

2988: && \ \ \

2989: -4\alpha_{23}\alpha_{45}+2\alpha_{12}\alpha_{23}+\alpha_{45}\alpha_{12}

2990: -2\alpha_{34}^2

2991: -\alpha_{34}\alpha_{51}+\alpha_{51}\alpha_{23}-4\alpha_{23}\alpha_{34}

2992: -\alpha_{12}\alpha_{34}-\nonumber \\

2993: && \ \ \ -4\alpha_{34}\alpha_{45}+2\alpha_{51}\alpha_{12}\biggr) +

2994: \frac{ \

2995: \pi^6}{45360}\biggl(36\alpha_{51}\alpha_{23}-28\alpha_{12}\alpha_{34}+69\alpha_{51}\alpha_{12}

2996: +36\alpha_{23}^2+36\alpha_{45}^2+\nonumber

2997: \\

2998: && \ \ \ +69\alpha_{12}\alpha_{23}+36\alpha_{51}^2+36\alpha_{12}^2

2999: -134\alpha_{23}\alpha_{34}-52\alpha_{34}^2+69\alpha_{45}\alpha_{51}

3000: -134\alpha_{34}\alpha_{45}+\nonumber \\

3001: && \ \ \

3002: +36\alpha_{45}\alpha_{12}-28\alpha_{34}\alpha_{51}-134\alpha_{23}\alpha_{45}\biggr)\biggr](2

3003: \alpha')^6 + \ {\cal O}( \ (2 \alpha')^7 \ )  \ ,

3004: \\

3005: \label{Z23exp} Z^{(1)} &=&  -\frac{ \ \pi^4}{1440}(2

3006: \alpha')^4+\biggl(-\frac{ \

3007: \pi^2}{72}\zeta(3)+\frac{1}{6}\zeta(5)\biggr)\biggl(-\alpha_{12}+3\alpha_{45}-\alpha_{51}

3008: +6\alpha_{34}+3\alpha_{23}\biggr)(2 \alpha')^5 + \nonumber

3009: \\

3010: && +

3011: \biggl[\frac{1}{192}\zeta(3)^2\biggl(21\alpha_{45}^2+6\alpha_{12}\alpha_{51}

3012: -16\alpha_{23}\alpha_{51}

3013: +3\alpha_{12}^2+3\alpha_{51}^2+21\alpha_{23}^2+96\alpha_{34}\alpha_{45}-

3014: \nonumber \\

3015: &&\hphantom{ +

3016: \biggl[}-16\alpha_{45}\alpha_{12}+72\alpha_{34}^2-16\alpha_{23}\alpha_{12}

3017: -24\alpha_{34}\alpha_{51}-24\alpha_{34}\alpha_{12}-16\alpha_{45}\alpha_{51}

3018: +96\alpha_{23}\alpha_{34}+\nonumber

3019: \\

3020: &&\hphantom{ + \biggl[}+30\alpha_{45}\alpha_{23}\biggr) -\frac{ \

3021: \pi^6}{362880}\biggl(276\alpha_{23}\alpha_{34}-72\alpha_{34}\alpha_{51}

3022: -46\alpha_{45}\alpha_{51}

3023: +90\alpha_{45}\alpha_{23}+63\alpha_{23}^2+\nonumber

3024: \\

3025: &&\hphantom{ + \biggl[}+216\alpha_{34}^2 -46\alpha_{23}\alpha_{51}

3026: +18\alpha_{12}\alpha_{51}+9\alpha_{12}^2+9\alpha_{51}^2

3027: +63\alpha_{45}^2-46\alpha_{23}\alpha_{12}

3028: +276\alpha_{34}\alpha_{45}-\nonumber \\

3029: &&\hphantom{ + \biggl[}-46\alpha_{45}\alpha_{12}

3030: -72\alpha_{34}\alpha_{12}\biggr)\biggr](2 \alpha')^6 + \ {\cal O}( \

3031: (2 \alpha')^7 \ ) \ ,

3032: \\

3033: \label{Deltaexp} \Delta &=& \biggl(\frac{3}{2}\pi^2

3034: \zeta(3)-\frac{35}{2}\zeta(5)\biggr)(2 \alpha')^5 + \nonumber

3035: \\

3036: && \ \ \  +

3037: \biggl(-\frac{3}{2}\zeta(3)^2+\frac{109}{45360}\pi^6\biggr)

3038: \biggl(\alpha_{12}+\alpha_{23}+\alpha_{34}+\alpha_{45}+

3039: \alpha_{51}\biggr)(2 \alpha')^6 + {\cal O}(  \ (2 \alpha')^7 \ ) \

3040: .

3041: \end{eqnarray}

3042: 

3043: 

3044: 

3045: 

3046: \section{The Yang-Mills 5-point subamplitude}

3047: \label{YM5} The calculation of the Yang-Mills 5-point subamplitude

3048: was already considered in the appendix D of

3049: \cite{Brandt1}\footnote{In \cite{Brandt1} it was called as `5-point

3050: amplitude'.}. Its (off-shell) expression is the following:

3051: \begin{multline}

3052: A_{YM}(1,2,3,4,5)= -i \ g^3 \times \\

3053: \begin{split}

3054: &\times \biggl\{ \frac{ {V_{YM}}_{\mu_1 \mu_2}^{ (3) \ \rho}(k_1,

3055: k_2, -k_1-k_2){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_3

3056: \mu_4}(-k_3-k_4,k_3,k_4) {V_{YM}}^{(3)}_{\ \mu_5 \rho

3057: \sigma}(k_5,k_1+k_2, k_3+k_4)}{(k_1+k_2)^2

3058: (k_3+k_4)^2} \ - \\

3059: &\hphantom{-i \ g^3 \biggl\{ }- \ \frac{{V_{YM}}_{\mu_1 \mu_2}^{

3060: (3) \ \rho}(k_1, k_2, -k_1-k_2){V_{YM}}^{(4)}_ {\rho \mu_3 \mu_4

3061: \mu_5}(k_1+k_2, k_3, k_4, k_5)}{(k_1+k_2)^2} \biggr\}

3062: \zeta_1^{\mu_1}\zeta_2^{\mu_2}\zeta_3^{\mu_3}\zeta_4^{\mu_4}

3063: \zeta_5^{\mu_5} \ + \\

3064: &+ \ \left( \frac{}{}\mbox{cyclic permutations} \frac{}{} \right)

3065: \ ,

3066: \end{split}

3067: \label{AYM5}

3068: \end{multline}

3069: where the Yang-Mills 3 and 4-point vertices (which do not carry

3070: color indices) are given, respectively, by

3071: \begin{eqnarray}

3072: \label{YMvertex3}

3073: V^{(3)}_{YM \ \mu_1 \mu_2 \mu_3}(k_1, k_2, k_3)

3074: &=& -i \ [\frac{}{} \eta_{\mu_1 \mu_2} (k_1-k_2)_{\mu_3} +

3075: \eta_{\mu_2 \mu_3}

3076: (k_2-k_3)_{\mu_1} + \eta_{\mu_3 \mu_1} (k_3-k_1)_{\mu_2} \frac{}{}] \ , \ \ \ \ \ \ \\

3077: \label{YMvertex4} V^{(4)}_{YM \ \mu_1 \mu_2 \mu_3 \mu_4}(k_1, k_2,

3078: k_3, k_4 ) &=& - \ [\frac{}{} \eta_{\mu_1 \mu_2} \eta_{\mu_3

3079: \mu_4} - 2 \eta_{\mu_1 \mu_3} \eta_{\mu_2 \mu_4} + \eta_{\mu_4

3080: \mu_1} \eta_{\mu_2 \mu_3} \frac{}{}] \ .

3081: \end{eqnarray}

3082: We have not worked any further trying to find a tensor notation

3083: which would shorten the expression of $A_{YM}(1,2,3,4,5)$ in

3084: (\ref{AYM5}).

3085: 

3086: \section{The $D^{2n}F^4$ terms 5-point subamplitude}

3087: \label{D2nF4}

3088: 

3089: The 5-point subamplitude the comes from the lagrangian ${\cal

3090: L}_{D^{2n}F^4}$, given in (\ref{Seff-nonabelian}), is given by:

3091: \begin{multline}

3092: A_{D^{2n}F^4}(1,2,3,4,5) = -2 \ (2 \alpha')^ 2 \ g^3 \ \biggl\{ \ \

3093: f(-2 \alpha_{34}, -2

3094: \alpha_{45}) \ \frac{T_{12}(\zeta,k)}{\alpha_{12}}  \ +  \\

3095: \begin{split}

3096: & \ \ + \ K(\zeta_1, k_1; \zeta_2, k_2; \zeta_3, k_3; \zeta_4,

3097: k_4) \ S_{5}(\zeta_5; k_1, k_2, k_3, k_4,k_5; \alpha') \ \

3098: \biggr\} \ + \ \left( \frac{}{}\mbox{cyclic permutations}

3099: \frac{}{} \right) \ ,

3100: \end{split}

3101: \label{AD2nF4}

3102: \end{multline}

3103: where

3104: \begin{eqnarray}

3105: T_{12}(\zeta,k) & = & \alpha_{12} \ K(\zeta_1, \zeta_2; \zeta_3,

3106: k_3; \zeta_4, k_4; \zeta_5, k_5)  + (\zeta_1 \cdot \zeta_2) K(k_1,

3107: k_2; \zeta_3,

3108: k_3; \zeta_4, k_4; \zeta_5, k_5) + \nonumber \\

3109: &+& (\zeta_1 \cdot k_2) K(\zeta_2, k_1 + k_2; \zeta_3, k_3; \zeta_4,

3110: k_4; \zeta_5, k_5)  -  (\zeta_2 \cdot k_1) K(\zeta_1, k_1 + k_2;

3111: \zeta_3, k_3; \zeta_4, k_4; \zeta_5, k_5)  \

3112: \nonumber \\

3113: \label{T12}

3114: \end{eqnarray}

3115: and

3116: \begin{multline}

3117: S_{5}(\zeta_5; k_1, k_2, k_3, k_4,k_5; \alpha') =\ \biggl[ \

3118: \left( \frac{}{} G(-2 \alpha_{34},-2 \alpha_{23},- \alpha_{12}-

3119: \alpha_{34},-\alpha_{14}- \alpha_{23}) + \right. \\

3120: \begin{split}

3121: % \ \  \hspace{4.2cm}

3122: \left.+ \ G(-2 \alpha_{12},-2 \alpha_{23}, - \alpha_{12}-

3123: \alpha_{34},-\alpha_{14}- \alpha_{23}) \frac{}{} \right) \ \zeta_5

3124: \cdot (k_1 + k_2) \ \ +

3125: \\

3126: % \hphantom{\ \ + \ K(\zeta_1, k_1; \zeta_2, k_2; \zeta_3, k_3;\zeta_4, k_4) \ }

3127: + \ \left( \frac{}{} G(-2 \alpha_{23}, -2 \alpha_{34},- \alpha_{14}-

3128: \alpha_{23},-\alpha_{12}- \alpha_{34}) - \right. \\

3129: % \ \ \left. \hspace{4.2cm}

3130: \left.- \ G(-2 \alpha_{23},-2 \alpha_{12}, - \alpha_{14}-

3131: \alpha_{23},-\alpha_{12}- \alpha_{34}) \frac{}{} \right) \ \zeta_5

3132: \cdot (k_1 + k_4) \ \biggr] \ .

3133: \end{split}

3134: \label{S12}

3135: \end{multline}

3136: \noindent In these formulas, by $K(A, a; B, b; C, c; D, d)$ we mean

3137: the same expression of the 4-point amplitude kinematic factor, in

3138: (\ref{kinematicfactor}), evaluated in the corresponding variables.

3139: The function $f(s,t)$ in the second line of (\ref{AD2nF4}) is the

3140: one defined in (\ref{f}) and the function $G(a,b,c,d)$ that goes in

3141: the second square bracket of (\ref{AD2nF4}) is the one defined in

3142: (\ref{G}).

3143: 

3144: \noindent The derivation of our formula in (\ref{AD2nF4}) is quite

3145: nontrivial and very lengthy. We will not give all the details

3146: here, but will comment some aspects about it. It was obtained by

3147: first finding the 4 and 5-point vertices, $V^{(4)}_{\mu_1 \mu_2

3148: \mu_3 \mu_4}(k_1, k_2, k_3, k_4)$ and $V^{(5)}_{\mu_1 \mu_2 \mu_3

3149: \mu_4 \mu_5}(k_1, k_2, k_3, k_4, k_5)$ (which do not carry color

3150: indices), of the lagrangian in (\ref{Seff-nonabelian}). This was

3151: done along the same lines of the Appendices of

3152: \cite{DeRoo1,Chandia1}, that is, by means of the 1-particle

3153: irreducible 4 and 5-point functions in momentum space.\\

3154: In the case of the 1-particle irreducible 4-point function,

3155: $\Gamma_{\mu_1 \mu_2 \mu_3 \mu_4}^{(4)} (k_1, k_2, k_3, k_4)$, the

3156: calculation was straight forward since only the abelian terms of

3157: ${\cal L}_{D^{2n}F^4}$ were required. \\

3158: In the case of the 1-particle irreducible 5-point function,

3159: $\Gamma_{\mu_1 \mu_2 \mu_3 \mu_4 \mu_5}^{(5)} (k_1, k_2, k_3, k_4,

3160: k_5)$, the calculation was much more involved than the previous

3161: one. On each $D^{2n}F^4$ term there are two types of 5-field terms

3162: present:

3163: \begin{enumerate}

3164: \item The ones that come from the non-abelian $F^4$ terms

3165: ($F^4 \rightarrow A^2 \cdot (\partial A)^3$), with the covariant

3166: derivatives acting as ordinary ones ($D^{2n} \rightarrow

3167: {\partial}^{2n})$.\\

3168: \noindent In the scattering amplitude (\ref{AD2nF4}) their

3169: contribution is the one that goes with the $f$ factor, with no

3170: poles.

3171: \item The ones that pick an $A^{a}_{\mu}$ term from the covariant

3172: derivatives ($D^{2n} \rightarrow A \cdot {\partial}^{2n-1}$) and

3173: the other four from the abelian part of the $F^4$ term

3174: ($F^4 \rightarrow (\partial A)^4$).\\

3175: \noindent These are the terms of the scattering amplitude

3176: (\ref{AD2nF4}) which are written in terms of the function

3177: $G(a,b,c,d)$. They begin to contribute only at order ${\alpha'}^3$

3178: (while the rest of the terms begin to contribute at ${\alpha'}^2$

3179: order) .

3180: \end{enumerate}

3181: Once the 4 and 5-point vertices have been found, to all order in

3182: $\alpha'$, the calculation of the 5-point subamplitude in

3183: (\ref{AD2nF4}) can be done using the corresponding Feynman rules.

3184: This subamplitude receives contributions from one particle

3185: irreducible (1 PI) diagrams and one particle reducible ones which

3186: contain only simple poles, as shown in

3187: figure \ref{twodiagrams}.\\

3188: \begin{figure}[h]

3189: \centerline{\includegraphics*[scale=0.7,angle=0]{diagrams.ps}}

3190: \caption{Type of 5-particle diagrams, with $\alpha'$ dependence,

3191: which come from the effective lagrangian. The Feynman diagram in

3192: (a) is one particle irreducible (1 PI), while the one in (b) is

3193: one particle reducible. } \label{twodiagrams}

3194: \end{figure}

3195: \noindent By construction, $A_{D^{2n}F^4}(1,2,3,4,5)$ remains

3196: invariant under cyclic permutations of indexes \\

3197: $(1,2,3,4,5)$. It is not difficult to see that it also remains

3198: invariant under a

3199: world-sheet parity (or twisting) transformation.\\

3200: Since the lagrangian ${\cal L}_{D^{2n}F^4}$ is gauge invariant,

3201: $A_{D^{2n}F^4}(1,2,3,4,5)$ should also satisfy (on-shell) gauge

3202: invariance. This is a very non trivial test of formula

3203: (\ref{AD2nF4}). In fact, demanding that it should become zero

3204: after doing $\zeta_5 \rightarrow k_5$, it leads to the following

3205: relation to be satisfied:

3206: \begin{eqnarray}

3207: f(-2 \alpha_{12}, -2 \alpha_{23}) - f(-2 \alpha_{23}, -2

3208: \alpha_{34}) & = &  \left[ \frac{}{} G(-2 \alpha_{34},-2

3209: \alpha_{23},- \alpha_{12}-

3210: \alpha_{34},-\alpha_{14}- \alpha_{23}) +  \right. \nonumber \\

3211: & + & \left.   \ G(-2 \alpha_{12},-2 \alpha_{23}, - \alpha_{12}-

3212: \alpha_{34},-\alpha_{14}- \alpha_{23}) \frac{}{} \right]

3213: (\alpha_{51}+ \alpha_{25})  + \nonumber

3214: \\

3215: &&  + \ \left[ \frac{}{} G(-2 \alpha_{23}, -2 \alpha_{34},-

3216: \alpha_{14}- \alpha_{23},-\alpha_{12}-

3217: \alpha_{34}) - \right. \nonumber \\

3218: & - &  \left.  \ G(-2 \alpha_{23},-2 \alpha_{12}, - \alpha_{14}-

3219: \alpha_{23},-\alpha_{12}- \alpha_{34}) \frac{}{} \right]

3220: (\alpha_{51} + \alpha_{45})  . \nonumber \\

3221: \label{gauge-inv-AD2nF4}

3222: \end{eqnarray}

3223: We have checked this relation using the expansions of $f(s,t)$ and

3224: $G(a,b,c,d)$, given in (\ref{fpower}) and (\ref{G}), respectively,

3225: up to ${\cal O}({\alpha'}^{15})$ terms.

3226: 

3227: \section{Derivation of the $D^{2n}F^5$ terms 5-point subamplitude}

3228: \label{derivation}

3229: 

3230: Together with the derivation of formulas (\ref{A12345final}) and

3231: (\ref{AD2nF4}), the calculations of the present appendix

3232: constitute an essential part of this work. There are three main

3233: achievements that we arrive to in the expression for

3234: $A_{D^{2n}F^5}(1,2,3,4,5)$: it explicitly has no poles; it is

3235: written as a sum of terms that have manifest cyclic, (on-shell)

3236: gauge invariance and world-sheet parity symmetry; it is written in

3237: terms of tensors in such a way that a local

3238: lagrangian can be found almost directly in terms of them.\\

3239: We have divided the calculations of this appendix in five steps:\\

3240: 

3241: \noindent {\bf \underline{Step 1}:} \ \ Treatment of the poles of $T

3242: \cdot A_{YM}(1,2,3,4,5)$ in

3243: (\ref{A12345final}).\\

3244: 

3245: \noindent We begin writing the mentioned term as

3246: \begin{multline}

3247: T \cdot A_{YM}(1,2,3,4,5)= A_{YM}(1,2,3,4,5) - i g^3 \biggl\{ \ \

3248: \biggl(T\biggl|_{\alpha_{12}=0}- \ 1 \ \biggr) \times \\

3249: \begin{split}

3250: &\times \biggl[ \frac{ {V_{YM}}_{\mu_1 \mu_2}^{ (3) \ \rho}(k_1,

3251: k_2, -k_1-k_2){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_3

3252: \mu_4}(-k_3-k_4,k_3,k_4) {V_{YM}}^{(3)}_{\ \mu_5 \rho

3253: \sigma}(k_5,k_1+k_2, k_3+k_4)}{4 \alpha_{12} \alpha_{34}} \ - \\

3254: &- \ \frac{{V_{YM}}_{\mu_1 \mu_2}^{ (3) \ \rho}(k_1, k_2,

3255: -k_1-k_2){V_{YM}}^{(4)}_ {\rho \mu_3 \mu_4 \mu_5}(k_1+k_2, k_3, k_4,

3256: k_5)}{2 \alpha_{12}} \biggr]

3257: \zeta_1^{\mu_1}\zeta_2^{\mu_2}\zeta_3^{\mu_3}\zeta_4^{\mu_4}

3258: \zeta_5^{\mu_5} \ \ + \\

3259: &+ \ \left( \frac{}{}\mbox{cyclic permutations} \frac{}{} \right) \

3260: \ \biggl\} \  \hspace{0.9cm} - \hspace{0.9cm} i g^3 \biggl\{ \ \

3261: \biggl(T \ - \ T\biggl|_{\alpha_{12}=0} \biggr) \times \\

3262: &\times \biggl[ \frac{ {V_{YM}}_{\mu_1 \mu_2}^{ (3) \ \rho}(k_1,

3263: k_2, -k_1-k_2){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_3

3264: \mu_4}(-k_3-k_4,k_3,k_4) {V_{YM}}^{(3)}_{\ \mu_5 \rho

3265: \sigma}(k_5,k_1+k_2, k_3+k_4)}{4 \alpha_{12} \alpha_{34}} \ - \\

3266: &- \ \frac{{V_{YM}}_{\mu_1 \mu_2}^{ (3) \ \rho}(k_1, k_2,

3267: -k_1-k_2){V_{YM}}^{(4)}_ {\rho \mu_3 \mu_4 \mu_5}(k_1+k_2, k_3, k_4,

3268: k_5)}{2 \alpha_{12}} \biggr]

3269: \zeta_1^{\mu_1}\zeta_2^{\mu_2}\zeta_3^{\mu_3}\zeta_4^{\mu_4}

3270: \zeta_5^{\mu_5} \ \ + \\

3271: &+ \ \left( \frac{}{}\mbox{cyclic permutations} \frac{}{} \right) \

3272: \ \biggl\} \ \ ,

3273: \end{split}

3274: \label{term1}

3275: \end{multline}

3276: Here we have split the left-hand member of the equality into three

3277: terms on the right-hand side:  $A_{YM}(1,2,3,4,5)$, the term which

3278: contains a factor $( \ T|_{\alpha_{12}=0} \ - \ 1 \ )$ (and cyclic

3279: permutations) and the term which contains a factor $( \ T \ - \

3280: T|_{\alpha_{12}=0} \ )$ (and cyclic permutations). We have used the

3281: on-shell expression of $A_{YM}(1,2,3,4,5)$ in these last two

3282: terms\footnote{This has been done by substituting

3283: $(k_i+k_j)^2=2\alpha_{ij}$ in the denominators of (\ref{AYM5}).} and

3284: we have also used the fact that $T$ is cyclic invariant.\\

3285: In principle, $T \cdot A_{YM}(1,2,3,4,5)$ should have double poles

3286: because $A_{YM}(1,2,3,4,5)$ does, but it happens that the mentioned

3287: expression has only simple ones. To see this we notice that $( \ T \

3288: - \ T|_{\alpha_{12}=0} \ )$ is factorable by $\alpha_{12}$ and also,

3289: using (\ref{K2-K}) and (\ref{relate32}), it may be proved that $( \

3290: T|_{\alpha_{12}=0} \ - \ 1 \ )$ is factorable by

3291: $\alpha_{34}\alpha_{45}$. Using this information in (\ref{term1}) we

3292: can arrive to the following result:

3293: \begin{multline}

3294: T \cdot A_{YM}(1,2,3,4,5)= A_{YM}(1,2,3,4,5) - i g^3 \biggl\{ \ \

3295: \frac{1}{\alpha_{12}} \ \ \biggl[ \ \ \

3296: \biggl(\frac{T|_{\alpha_{12}=0}- \ 1 \ }

3297: {\alpha_{34}\alpha_{45}}\biggr)

3298: \times \\

3299: \begin{split}

3300: & \times \biggl(\frac{\alpha_{45}}{4} {V_{YM}}_{\mu_1 \mu_2}^{ (3) \

3301: \rho}(k_1, k_2, -k_1-k_2){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_3

3302: \mu_4}(-k_3-k_4,k_3,k_4) {V_{YM}}^{(3)}_{\ \mu_5 \rho

3303: \sigma}(k_5,k_1+k_2, k_3+k_4) + \\

3304: &\hphantom{\times \biggl( \ } + \frac{\alpha_{34}\alpha_{45}}{2} \

3305: {V_{YM}}_{\mu_1 \mu_2}^{ (3) \ \rho}(k_1, k_2,

3306: -k_1-k_2){V_{YM}}^{(4)}_ {\rho \mu_3 \mu_4 \mu_5}(k_1+k_2, k_3, k_4,

3307: k_5) \biggr) \ \ + \\

3308: &\hphantom{T \cdot A_{YM}(1,2,3,4,5)= A_{YM}(1,2,3,4,5) - i g^3

3309: \biggl\{ \ \frac{1}{\alpha_{12}} \ \  }  + \ \biggl(\frac{T \ -

3310: \ T|_{\alpha_{45}=0} \ }{\alpha_{45}}\biggr) \times \\

3311: & \times \frac{1}{4} {V_{YM}}_{\mu_1 \mu_2}^{ (3) \ \rho}(k_1, k_2,

3312: -k_1-k_2){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_3

3313: \mu_4}(-k_3-k_4,k_3,k_4) {V_{YM}}^{(3)}_{\ \mu_5 \rho

3314: \sigma}(k_5,k_1+k_2, k_3+k_4) \ \ \ \biggr] \times \\

3315: & \times \

3316: \zeta_1^{\mu_1}\zeta_2^{\mu_2}\zeta_3^{\mu_3}\zeta_4^{\mu_4}

3317: \zeta_5^{\mu_5} \ \ \ \ + \ \ \ \ \left( \frac{}{}\mbox{cyclic

3318: permutations} \frac{}{} \right) \ \ \biggr\} \ \ \ \ - \\

3319: &- i g^3 \biggl\{ \ \ \biggl(\frac{T \ - \ T|_{\alpha_{12}=0} \

3320: }{\alpha_{12}}\biggr) \frac{1}{2} {V_{YM}}_{\mu_1 \mu_2}^{ (3) \

3321: \rho}(k_1, k_2, -k_1-k_2){V_{YM}}^{(4)}_ {\rho \mu_3 \mu_4

3322: \mu_5}(k_1+k_2, k_3, k_4, k_5) \ \times \\

3323: & \times \

3324: \zeta_1^{\mu_1}\zeta_2^{\mu_2}\zeta_3^{\mu_3}\zeta_4^{\mu_4}

3325: \zeta_5^{\mu_5} \ \ \ \ + \ \ \ \ \left( \frac{}{}\mbox{cyclic

3326: permutations} \frac{}{} \right) \ \ \biggr\} \ ,

3327: \end{split}

3328: \label{term2}

3329: \end{multline}

3330: in which is manifest that $T \cdot A_{YM}(1,2,3,4,5)$ contains only

3331: simple poles: they come in the first curly bracket of this

3332: equation.\\

3333: 

3334: \noindent {\bf \underline{Step 2}:} \ \ Treatment of the poles of

3335: $(2 \alpha')^2 K_3 \cdot A_{F^4}(1,2,3,4,5)$ in

3336: (\ref{A12345final}).\\

3337: 

3338: \noindent In a similar way as it was done in Step 1, using the

3339: expression in (\ref{AF412345}) for $A_{F^4}(1,2,3,4,5)$, we arrive

3340: to:

3341: \begin{eqnarray}

3342: (2 \alpha')^2 K_3 \cdot A_{F^4}(1,2,3,4,5) & = & -2 g^3 \biggl\{ (2

3343: \alpha')^2 f(-2 \alpha_{34}, -2 \alpha_{45}) \frac{T_{12}(\zeta,

3344: k)}{\alpha_{12}} + \left( \frac{}{}\mbox{cyclic permutations}

3345: \frac{}{} \right) \ \biggr\} + \nonumber \\

3346: &+&2 g^3 \biggl\{ \biggl(\frac{T|_{\alpha_{12}=0}- \ 1 \ }

3347: {\alpha_{34}\alpha_{45}}\biggr) \frac{T_{12}(\zeta, k)}{\alpha_{12}}

3348: + \left( \frac{}{}\mbox{cyclic permutations}

3349: \frac{}{} \right) \ \biggr\} + \nonumber \\

3350: &+& 2 g^3 \biggl\{ (2 \alpha')^2 \biggl(\frac{K_3 \ - \

3351: {K_3}|_{\alpha_{12}=0} \ }{\alpha_{12}}\biggr) T_{12}(\zeta, k)+

3352: \left( \frac{}{}\mbox{cyclic

3353: permutations} \frac{}{} \right) \ \biggr\} \ , \nonumber \\

3354: %\begin{split}

3355: %\end{split}

3356: \label{term3}

3357: \end{eqnarray}

3358: where $T_{12}(\zeta, k)$ is given in (\ref{T12}).\\

3359: The simple poles appear in the first and second lines of

3360: (\ref{term3}). In the second line of this equation we used that

3361: \begin{eqnarray}

3362: (2 \alpha')^2 K_3\biggl|_{\alpha_{12}=0} \ +  \ \ (2 \alpha')^2 f(-2

3363: \alpha_{34}, -2 \alpha_{45}) \ = \ \frac{T|_{\alpha_{12}=0}- \ 1 \ }

3364: {\alpha_{34}\alpha_{45}} \ , \label{identity}

3365: \end{eqnarray}

3366: which is equivalent to (\ref{relate3}), once the double pole of that

3367: equation has been subtracted in both sides of it.\\

3368: We notice that the simple poles in the first line of (\ref{term3})

3369: are \underline{exactly the same ones} of\\

3370: $A_{D^{2n}F^4}(1,2,3,4,5)$ (see eq. (\ref{AD2nF4}) ).\\

3371: 

3372: \noindent {\bf \underline{Step 3}:} \ \ Taking the poles to

3373: $A_{D^{2n}F^4}(1,2,3,4,5)$.\\

3374: 

3375: \noindent Using the expressions in (\ref{term2}) and (\ref{term3})

3376: we have that

3377: \begin{multline}

3378: T \cdot A_{YM}(1,2,3,4,5)+(2 \alpha')^2 K_3 \cdot A_{F^4}(1,2,3,4,5)

3379: = \\

3380: \begin{split}

3381: &=A_{YM}(1,2,3,4,5) - 2 g^3 \biggl\{ (2 \alpha')^2 f(-2 \alpha_{34},

3382: -2 \alpha_{45}) \frac{T_{12}(\zeta, k)}{\alpha_{12}}   + \left(

3383: \frac{}{}\mbox{cyclic

3384: permutations} \frac{}{} \right) \ \biggr\} \ +\\

3385: &+ g^3 \biggl\{ \ \ \frac{1}{\alpha_{12}} \ \ \biggl[ \ \ \

3386: \biggl(\frac{T|_{\alpha_{12}=0}- \ 1 \ }

3387: {\alpha_{34}\alpha_{45}}\biggr)

3388: \times \biggl( \ 2 \ T_{12}(\zeta, k) \ \ -  \\

3389: & - i \frac{\alpha_{45}}{4} {V_{YM}}_{\mu_1 \mu_2}^{ (3) \

3390: \rho}(k_1, k_2, -k_1-k_2){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_3

3391: \mu_4}(-k_3-k_4,k_3,k_4) {V_{YM}}^{(3)}_{\ \mu_5 \rho

3392: \sigma}(k_5,k_1+k_2, k_3+k_4) + \\

3393: &\hphantom{\times \biggl( \ } +i \frac{\alpha_{34}\alpha_{45}}{2} \

3394: {V_{YM}}_{\mu_1 \mu_2}^{ (3) \ \rho}(k_1, k_2,

3395: -k_1-k_2){V_{YM}}^{(4)}_ {\rho \mu_3 \mu_4 \mu_5}(k_1+k_2, k_3, k_4,

3396: k_5) \biggr) \ \ - \\

3397: &\hphantom{T \cdot A_{YM}(1,2,3,4,5)= A_{YM}(1,2,3,4,5) - i g^3

3398: \biggl\{ \ \frac{1}{\alpha_{12}} \ \  }  - \ \biggl(\frac{T \ -

3399: \ T|_{\alpha_{45}=0} \ }{\alpha_{45}}\biggr) \times \\

3400: & \times \frac{i}{4} {V_{YM}}_{\mu_4 \mu_5}^{ (3) \ \rho}(k_4, k_5,

3401: -k_4-k_5){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_1

3402: \mu_2}(-k_1-k_2,k_1,k_2) {V_{YM}}^{(3)}_{\ \mu_3 \rho

3403: \sigma}(k_3,k_4+k_5, k_1+k_2) \ \ \ \biggr] \times \\

3404: & \times \

3405: \zeta_1^{\mu_1}\zeta_2^{\mu_2}\zeta_3^{\mu_3}\zeta_4^{\mu_4}

3406: \zeta_5^{\mu_5} \ \ \ \ + \ \ \ \ \left( \frac{}{}\mbox{cyclic

3407: permutations} \frac{}{} \right) \ \ \biggr\} \ \ \ \ - \\

3408: &- i g^3 \biggl\{ \ \ \biggl(\frac{T \ - \ T|_{\alpha_{12}=0} \

3409: }{\alpha_{12}}\biggr) \frac{1}{2} {V_{YM}}_{\mu_1 \mu_2}^{ (3) \

3410: \rho}(k_1, k_2, -k_1-k_2){V_{YM}}^{(4)}_ {\rho \mu_3 \mu_4

3411: \mu_5}(k_1+k_2, k_3, k_4, k_5) \ \times \\

3412: & \times \

3413: \zeta_1^{\mu_1}\zeta_2^{\mu_2}\zeta_3^{\mu_3}\zeta_4^{\mu_4}

3414: \zeta_5^{\mu_5} \ \ \ \ + \ \ \ \ \left( \frac{}{}\mbox{cyclic

3415: permutations} \frac{}{} \right) \ \ \biggr\} \ + \\

3416: &+ 2 g^3 \biggl\{ (2 \alpha')^2 \biggl(\frac{K_3 \ - \

3417: {K_3}|_{\alpha_{12}=0} \ }{\alpha_{12}}\biggr) T_{12}(\zeta, k)+

3418: \left( \frac{}{}\mbox{cyclic

3419: permutations} \frac{}{} \right) \ \biggr\} \ .

3420: \end{split}

3421: \label{term4}

3422: \end{multline}

3423: Now, we first notice that

3424: \begin{eqnarray}

3425: \biggl(\frac{T \ - \ T|_{\alpha_{45}=0} \ }{\alpha_{45}}\biggr) -

3426: \alpha_{34}\biggl(\frac{T|_{\alpha_{12}=0}- \ 1 \ }

3427: {\alpha_{34}\alpha_{45}}\biggr) \label{M4512}

3428: \end{eqnarray}

3429: is factorable by $\alpha_{12}$\footnote{The expression in

3430: (\ref{M4512}) has a power series expansion in $\alpha_{12}$ and

3431: using (one of the cyclic permutations of) (\ref{T0}) it can be

3432: proved that it becomes $0$ when $\alpha_{12}=0$.}, so we will call

3433: it $\alpha_{12}M_{12}$, where $M_{12}$ has a well defined $\alpha'$

3434: series expansion with no poles. Eliminating $( \ T \ - \

3435: T|_{\alpha_{45}=0} \ )/\alpha_{45}$ from it and substituting it in

3436: (\ref{term4}), together with the expression of

3437: $A_{D^{2n}F^4}(1,2,3,4,5)$, given in (\ref{AD2nF4}), we arrive to

3438: \begin{multline}

3439: T \cdot A_{YM}(1,2,3,4,5)+(2 \alpha')^2 K_3 \cdot A_{F^4}(1,2,3,4,5)

3440: = \\

3441: \begin{split}

3442: &=A_{YM}(1,2,3,4,5)+A_{D^{2n}F^4}(1,2,3,4,5) + g^3 \biggl\{ \ \

3443: \biggl(\frac{T|_{\alpha_{12}=0}- \ 1 \ }

3444: {\alpha_{34}\alpha_{45}}\biggr) \times \ \biggl( \ \ \

3445: \frac{1}{\alpha_{12}} \

3446:  \biggl[ \ 2 \ T_{12}(\zeta, k) \ -  \\

3447: & - i \frac{\alpha_{45}}{4} {V_{YM}}_{\mu_1 \mu_2}^{ (3) \

3448: \rho}(k_1, k_2, -k_1-k_2){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_3

3449: \mu_4}(-k_3-k_4,k_3,k_4) {V_{YM}}^{(3)}_{\ \mu_5 \rho

3450: \sigma}(k_5,k_1+k_2, k_3+k_4) - \\

3451: & -i \frac{\alpha_{34}}{4} {V_{YM}}_{\mu_4 \mu_5}^{ (3) \ \rho}(k_4,

3452: k_5, -k_4-k_5){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_1

3453: \mu_2}(-k_1-k_2,k_1,k_2) {V_{YM}}^{(3)}_{\ \mu_3 \rho

3454: \sigma}(k_3,k_4+k_5, k_1+k_2) \ + \\

3455: &+i \frac{\alpha_{34}\alpha_{45}}{2} \ {V_{YM}}_{\mu_1 \mu_2}^{ (3)

3456: \ \rho}(k_1, k_2, -k_1-k_2){V_{YM}}^{(4)}_ {\rho \mu_3 \mu_4

3457: \mu_5}(k_1+k_2, k_3, k_4, k_5) \ \biggr] \ + \\

3458: &+i \frac{\alpha_{34}}{2} \ {V_{YM}}_{\mu_4 \mu_5}^{ (3) \

3459: \rho}(k_4, k_5, -k_4-k_5){V_{YM}}^{(4)}_ {\rho \mu_1 \mu_2

3460: \mu_3}(k_4+k_5, k_1, k_2, k_3) \ \ \ \biggr) \

3461: \zeta_1^{\mu_1}\zeta_2^{\mu_2}\zeta_3^{\mu_3}\zeta_4^{\mu_4}

3462: \zeta_5^{\mu_5} \ \ + \\

3463: &+ \ \left( \frac{}{}\mbox{cyclic permutations} \frac{}{} \right) \

3464: \ \ \biggr\} \ \ \ \ - \ \ i g^3 \biggl\{ \ \ M_{12} \ \times

3465: \\

3466: &\times \biggl( \frac{1}{4} {V_{YM}}_{\mu_4 \mu_5}^{ (3) \

3467: \rho}(k_4, k_5, -k_4-k_5){V_{YM}}^{(3) \sigma }_{\ \ \ \mu_1

3468: \mu_2}(-k_1-k_2,k_1,k_2) {V_{YM}}^{(3)}_{\ \mu_3 \rho

3469: \sigma}(k_3,k_4+k_5, k_1+k_2) \ - \\

3470: &- \frac{\alpha_{12}}{2} \ {V_{YM}}_{\mu_4 \mu_5}^{ (3) \ \rho}(k_4,

3471: k_5, -k_4-k_5){V_{YM}}^{(4)}_ {\rho \mu_1 \mu_2 \mu_3}(k_4+k_5, k_1,

3472: k_2, k_3) \ \biggr) \

3473: \zeta_1^{\mu_1}\zeta_2^{\mu_2}\zeta_3^{\mu_3}\zeta_4^{\mu_4}

3474: \zeta_5^{\mu_5} \ \  + \\

3475: & + \  \left( \frac{}{}\mbox{cyclic

3476: permutations} \frac{}{} \right) \ \ \biggr\} \ + \\

3477: &+ 2 g^3 \biggl\{ (2 \alpha')^2 \biggl(\frac{K_3 \ - \

3478: {K_3}|_{\alpha_{12}=0} \ }{\alpha_{12}}\biggr) T_{12}(\zeta, k)+

3479: \left( \frac{}{}\mbox{cyclic permutations} \frac{}{} \right) \

3480: \biggr\} \ + \\

3481: & +2 \ g^3 \ (2 \alpha')^ 2  \ \biggl\{ \ K(\zeta_1, k_1; \zeta_2,

3482: k_2; \zeta_3, k_3; \zeta_4, k_4) \ S_{5}(\zeta_5; k_1, k_2, k_3, k_4,k_5; \alpha')  \ + \\

3483: &\hphantom{+2 \ g^3 \ (2 \alpha')^ 2  \ \biggl\{ \ }  + \left(

3484: \frac{}{}\mbox{cyclic permutations} \frac{}{} \right) \ \biggr\} \ .

3485: \end{split}

3486: \label{term5}

3487: \end{multline}

3488: This is a huge expression. At this moment our main interest lies

3489: in the term in the square bracket (the one that goes multiplying

3490: the factor $1/\alpha_{12}$) and its cyclic permutations because,

3491: besides the poles of $A_{D^{2n}F^4}(1,2,3,4,5)$, those are the

3492: only places where poles could come from. We have verified

3493: computationally that, using on-shell and physical state

3494: conditions, together with momentum conservation, the term in the

3495: square bracket is factorable by $\alpha_{12}$, thus eliminating

3496: all poles which are not contained in $A_{D^{2n}F^4}(1,2,3,4,5)$.\\

3497: 

3498: \noindent {\bf \underline{Step 4}:} \ \ Derivation of an expression for

3499: $A_{D^{2n}F^5}(1,2,3,4,5)$ as a sum of terms which are gauge invariant

3500: and have no poles.\\

3501: 

3502: \noindent Since the left hand-side of (\ref{term5}) is precisely

3503: $A(1,2,3,4,5)$, using (\ref{AD2nF5}) we have that the sum of curly

3504: brackets in (\ref{term5}) corresponds to $A_{D^{2n}F^5}(1,2,3,4,5)$.

3505: This expression explicitly has no poles and is (on-shell) gauge

3506: invariant as a whole, but each of the terms in curly brackets is not

3507: individually gauge invariant. This means that, in order to write

3508: $A_{D^{2n}F^5}(1,2,3,4,5)$ as a sum of terms which have both

3509: properties, some redistribution of the terms in (\ref{term5}) should

3510: be done. Only

3511: after succeeding in doing this redistribution it will be possible to

3512: find a local lagrangian for each group of terms.\\

3513: \noindent This is a very non trivial step and in the following lines

3514: we will only summarize the operations which took us to the desired

3515: expression:

3516: \begin{enumerate}

3517: \item In (\ref{term5}) we substitute the following expression (and

3518: the cyclic permutations of it in the corresponding cases) for the

3519: kinematical expression $T_{12}(\zeta,k)$, defined in (\ref{T12}):

3520: \begin{eqnarray}

3521: T_{12}(\zeta,k) & = & {(\eta \cdot t_{(8)})_4}^{\mu_1 \nu_1 \mu_2

3522: \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5 \nu_5} \zeta _{\mu

3523: _{1}}^{1}k_{\nu _{1}}^{1}\zeta _{\mu _{2}}^{2}k_{\nu _{2}}^{2}\zeta

3524: _{\mu _{3}}^{3}k_{\nu _{3}}^{3}\zeta _{\mu _{4}}^{4}k_{\nu

3525: _{4}}^{4}\zeta _{\mu _{5}}^{5}k_{\nu _{5}}^{5} + \nonumber

3526: \\

3527: &&+ (\zeta^1 \cdot k^2) K( \zeta^2, k^2; \zeta^3, k^3; \zeta^4, k^4;

3528: \zeta^5, k^5) - (\zeta^2 \cdot k^1) K( \zeta^1, k^1; \zeta^3, k^3;

3529: \zeta^4, k^4; \zeta^5, k^5) \ . \nonumber \\

3530: \label{T12II}

3531: \end{eqnarray}

3532: The term with the $(\eta \cdot t_{(8)})_4$ tensor is the part of

3533: $T_{12}(\zeta,k)$ which is gauge invariant. That term changes sign

3534: under a twisting transformation with respect to index $4$.

3535: \item In the term $S_{5}(\zeta_5; k_1, k_2, k_3, k_4,k_5; \alpha')$

3536: given in (\ref{S12}), which also goes in (\ref{term5}), we

3537: substitute the following relation (and the corresponding cyclic

3538: permutations of it):

3539: \begin{multline}

3540: G(-2 \alpha_{34},-2 \alpha_{23},- \alpha_{12}-

3541: \alpha_{34},-\alpha_{14}- \alpha_{23}) + \ G(-2 \alpha_{12},-2

3542: \alpha_{23}, - \alpha_{12}-

3543: \alpha_{34},-\alpha_{14}- \alpha_{23}) = \\

3544: \begin{split}

3545: &\frac{f(-2 \alpha_{12}, -2 \alpha_{23}) - f(-2 \alpha_{23}, -2

3546: \alpha_{34})}{\alpha_{34}- \alpha_{12}} \ - \  \left[ \frac{}{} G(-2

3547: \alpha_{23}, -2 \alpha_{34},- \alpha_{14}- \alpha_{23},-\alpha_{12}

3548:  - \alpha_{34}) \ - \right. \\

3549: & \left. - \  G(-2 \alpha_{23},-2 \alpha_{12}, - \alpha_{14}-

3550: \alpha_{23},-\alpha_{12}- \alpha_{34}) \frac{}{} \right] \

3551: \left(\frac{\alpha_{51} + \alpha_{45}}{\alpha_{34}-

3552: \alpha_{12}}\right) .

3553: \end{split}

3554: \label{G1G2}

3555: \end{multline}

3556: This relation has been obtained from (\ref{gauge-inv-AD2nF4}) and in

3557: the denominators we have used, from (\ref{alpha25}), that

3558: $\alpha_{51}+\alpha_{25}=\alpha_{34}-\alpha_{12}$.

3559: \item We then substitute in (\ref{term5}) the following

3560: expressions for $M_{12}$ and $(T|_{\alpha_{12}=0}- \ 1 \ )/

3561: (\alpha_{34}\alpha_{45})$ (and the corresponding cyclic

3562: permutations of them):

3563: \begin{eqnarray}

3564: \label{M12} M_{12}&=& \alpha_{51} H^{(3)} +

3565: \alpha_{51}\alpha_{23}U^{(1)}+\alpha_{23}\alpha_{34}U^{(3)}

3566: +\alpha_{34}\alpha_{51}U^{(5)} +

3567: \alpha_{34}\alpha_{51}\alpha_{23}\Delta  \ , \ \ \ \ \ \ \ \ \\

3568: \label{G12} \frac{T|_{\alpha_{12}=0}- \ 1 \ }

3569: {\alpha_{34}\alpha_{45}}&=& \alpha_{23} H^{(1)} + \alpha_{51}

3570: H^{(2)}+\alpha_{23}\alpha_{51}U^{(4)} \ .

3571: \end{eqnarray}

3572: These relations have been obtained from the definition of

3573: $M_{12}$, given in (\ref{M4512}), and using the definitions of the

3574: $H^{(k)}$, $U^{(k)}$ and $\Delta$ factors of appendix

3575: \ref{otherfactors}, together with (\ref{identity}). \\

3576: \noindent As an outcome, the resulting expression of

3577: $A_{D^{2n}F^5}(1,2,3,4,5)$ will depend on kinematical expressions

3578: and on the factors $H^{(k)}$, $P^{(k)}$, $U^{(k)}$, $Z^{(k)}$

3579: ($k=1, \ldots , 5.$) and $\Delta$. It will no longer depend

3580: (explicitly) on the $T$ and $K_3$ factors.

3581: \item We next require on-shell gauge invariance in the resulting

3582: expression of $A_{D^{2n}F^5}(1,2,3,4,5)$ by doing, for example,

3583: $\zeta_1 \rightarrow k_1$ and demanding that the expression should

3584: vanish after using on-shell and physical state conditions, together

3585: with momentum conservation. This leads us to the following condition

3586: to be satisfied:

3587: \begin{eqnarray}

3588: \alpha_{51}P^{(3)}-\alpha_{12}P^{(4)}=

3589: \alpha_{51}H^{(2)}-\alpha_{12}H^{(5)}+\alpha_{23}\alpha_{51}U^{(4)}-

3590: \alpha_{12}\alpha_{45}U^{(3)} \ ,

3591: \label{almostthere}

3592: \end{eqnarray}

3593: which can be seen to be valid after using the definitions of the

3594: $H^{(k)}$, $P^{(k)}$ and $U^{(k)}$ factors, given in appendix

3595: \ref{otherfactors}.\\

3596: \noindent In order for the condition (\ref{almostthere}) to be

3597: automatically satisfied, without needing to make any substitutions,

3598: we introduce a new $\alpha'$ dependent factor, $W^{(1)}$, and its

3599: cyclic permutations $W^{(k)}$ ($k=2,3,4,5$). This factor has been

3600: defined in eq. (\ref{W51}) and it may be proved that it satisfies

3601: the following relations:

3602: \begin{eqnarray}

3603: P^{(4)} = H^{(5)}+\alpha_{45}U^{(3)}+\alpha_{51}W^{(1)} \ ,

3604: \\

3605: P^{(3)} = H^{(2)}+\alpha_{23}U^{(4)}+\alpha_{12}W^{(1)} \ ,

3606: \label{relateW}

3607: \end{eqnarray}

3608: which automatically fulfill (\ref{almostthere}).

3609: \end{enumerate}

3610: 

3611: \noindent {\bf \underline{Step 5}:} \ \ Final expression for

3612: $A_{D^{2n}F^5}(1,2,3,4,5)$.\\

3613: 

3614: \noindent From the previous step we find the following expression

3615: for $A_{D^{2n}F^5}(1,2,3,4,5)$:

3616: \begin{eqnarray}

3617: A_{D^{2n}F^5}(1,2,3,4,5) &=& g^3 \ \biggl\{ \frac{}{} \biggl[

3618: H^{(1)} \cdot h(\zeta,k) + P^{(1)}

3619: \cdot p^{(1)}(\zeta,k)+\nonumber \\

3620: &&\hphantom{\biggl\{ \frac{}{} \biggl[} + U^{(1)} \cdot u(\zeta,k) +

3621: W^{(1)} \cdot w(\zeta,k)+  Z^{(1)} \cdot z^{(1)}(\zeta,k) \biggr] +

3622: \nonumber \\

3623: &&\hphantom{\biggl\{ \frac{}{} \biggl[} + (\frac{}{} \mbox{cyclic

3624: permutations} \frac{}{}) \frac{}{} \biggr\} \ + g^3 \ \Delta \cdot

3625: \delta'(\zeta,k) \ ,

3626: \label{AD2nF5nopolesII}

3627: \end{eqnarray}

3628: where, by now, only $p^{(1)}(\zeta,k)$ and $z^{(1)}(\zeta,k)$ (and

3629: their cyclic permutations) are known to be the same kinematical

3630: expressions of equations (\ref{p34}) and (\ref{z23}), respectively.\\

3631: \noindent In (\ref{AD2nF5nopolesII}) we have an expression which, on

3632: $each$ group of terms, has no poles and is (on-shell) gauge

3633: invariant. We have also verified that the world-sheet parity

3634: condition (\ref{twisting}) is satisfied by $each$ group of terms in

3635: (\ref{AD2nF5nopolesII}). But unfortunately, for $h(\zeta,k)$,

3636: $u(\zeta,k)$, $w(\zeta,k)$ and $\delta'(\zeta,k)$ we only have long

3637: (computer saved) expressions, whose specific structure is not

3638: explicitly known.\\

3639: \noindent Our final labor has then been to determine the structure

3640: of $h(\zeta,k)$, $u(\zeta,k)$, $w(\zeta,k)$ and $\delta'(\zeta,k)$.

3641: In appendix \ref{t10tensor} we have explained with some detail how

3642: we have obtained an on-shell equivalent expression for $h(\zeta,k)$

3643: (which we have called $h^{(1)}(\zeta,k)$ in equations

3644: (\ref{AD2nF5nopoles}) and (\ref{h23})) in which the gauge symmetry

3645: is manifest. This has been done by introducing a ten index tensor,

3646: $t_{(10)}^{\mu_1 \nu_1 \mu_2 \nu_2 \mu_3 \nu_3 \mu_4 \nu_4 \mu_5

3647: \nu_5}$, which is antisymmetric on each pair $(\mu_j \ \nu_j)$ and

3648: which has the twisting symmetry mentioned in eq. (\ref{etat8twisting}).\\

3649: \noindent For each remaining kinematical expression we have also

3650: used an $ansatz$, in analogy to $h^{(1)}(\zeta,k)$, which consists

3651: in the most general expression that can be constructed and which is

3652: manifestly gauge invariant. It is quite remarkable that using

3653: on-shell, physical state and momentum conservation conditions, we

3654: have been able to determine their structure completely in terms of

3655: only two tensors: the known $t_{(8)}$ and the same $t_{(10)}$ that

3656: we have mentioned in the above lines. The resulting expressions are

3657: the ones that we have named $u^{(1)}(\zeta,k)$, $w^{(1)}(\zeta,k)$

3658: and $\delta(\zeta,k)$, respectively, in subsection \ref{D2nF5}.\\

3659: \noindent The final expression that we have for

3660: $A_{D^{2n}F^5}(1,2,3,4,5)$ is, then, the one in

3661: (\ref{AD2nF5nopoles}), which has no poles, is cyclic and

3662: (on-shell) gauge invariant on $each$ group of terms, as explained

3663: in subsection \ref{D2nF5}, and which also has the world-sheet

3664: parity symmetry manifest. All this final step has been

3665: fundamental, in order to go from the subamplitude in

3666: (\ref{AD2nF5nopoles}) to the effective lagrangian, ${\cal

3667: L}_{D^{2n}F^5}$, in (\ref{LD2nF5final}).

3668: 

3669: \begin{thebibliography}{10}

3670: 

3671: \bibitem{Scherk1} J.~Scherk, J.~H. Schwarz, \emph{Dual models

3672: for nan-hadrons}, \npb{81}{1974}{118}.

3673: 

3674: \bibitem{Fradkin1}

3675: E.S. Fradkin and A.A. Tseytlin, \emph{Nonlinear electrodynamics

3676: from quantized strings}, \plb{163}{1985}{123}.

3677: 

3678: \bibitem{Tseytlin1}

3679: A.A. Tseytlin, \emph{Vector field effective action in the open

3680:   superstring theory}, \npb{276}{1986}{391}; Erratum-ibid:

3681:   \npb{291}{1987}{876}.

3682: 

3683: \bibitem{Abouelsaood1} A. Abouelsaood, C. G. Callan, C. R. Nappi and

3684: S. A. Yost, \emph{Open strings in background gauge fields},

3685: \npb{280}{1987}{599}.

3686: 

3687: \bibitem{Bergshoeff1} E. Bergshoeff, E. Sezgin, C. N. Pope and P.

3688: K. Townsend, \emph{The Born-Infeld action from conformal

3689: invariance of the open superstring}, \plb{188}{1987}{70}.

3690: 

3691: \bibitem{Leigh1} R. G. Leigh, \emph{Dirac-Born-Infeld action from

3692: Dirichlet sigma model}, \mpla{4}{1989}{2767}.

3693: 

3694: \bibitem{Gross1}

3695: D.J. Gross and E.~Witten, \emph{Superstring modifications of

3696:   Einstein's equations}, \npb{277}{1986}{1}.

3697: 

3698: \bibitem{Kitazawa1}

3699: Y.~Kitazawa, \emph{Effective lagrangian for the open superstring

3700: from a 5-point function}, \npb{289}{1987}{599}.

3701: 

3702: \bibitem{Polchinski0}

3703: J.~Polchinski, \emph{Dirichlet-Branes and Ramond-Ramond Charges},

3704: \prl{75}{1995}{4724} [\hepth{9510017}].

3705: 

3706: \bibitem{Witten1}

3707: E. Witten, \emph{Bound states of strings and p-branes},

3708: \npb{460}{1996}{35}, [\hepth{9510135}].

3709: 

3710: \bibitem{Cederwall0}

3711: M.~Cederwall, A. von Gussich, B.E.W. Nilsson, P. Sundell and A.

3712: Westerberg, \emph{The Dirichlet Super Three-Brane in

3713: Ten-Dimensional Type IIB Supergravity}, \npb{490}{1997}{163},

3714: [\hepth{9610148}] .

3715: 

3716: \bibitem{Cederwall0p}

3717: M. Cederwall, A. von Gussich, B.E.W. Nilsson, P. Sundell and A.

3718: Westerberg, \emph{The Dirichlet Super-$p$-Branes in

3719: Ten-Dimensional Type IIA and IIB Supergravity},

3720: \npb{480}{1997}{179} [\hepth{9611159}].

3721: 

3722: \bibitem{Bergshoeff0}

3723: E. Bergshoeff and P.K. Townsend, \emph{Super-D-branes},

3724: \npb{490}{1997}{145}, [\hepth{9611173}].

3725: 

3726: \bibitem{Aganagic1}

3727: M. Aganagic, C. Popsescu and J. H. Schwarz, \emph{Gauge-invariant

3728: and gauge-fixed D-brane actions}, \npb{495}{1997}{99},

3729: [\hepth{9612080}].

3730: 

3731: \bibitem{Wyllard2}

3732: N.~Wyllard, \emph{Derivative corrections to D-brane actions with

3733: constant background fields}, \npb{598}{2001}{247}

3734: [\hepth{0008125}].

3735: 

3736: \bibitem{Andreev2}

3737: O. ~Andreev, \emph{More about partition fntion of open bosonic

3738: string in background fields and string theory effective action},

3739: \plb{513}{2001}{207}.

3740: 

3741: \bibitem{DeRoo1}

3742: M.~De~Roo and M.G.C. Eenink, \emph{The effective action for the

3743: 4-point functions in abelian open superstring theory},

3744: \jhep{0308}{2003}{036} [\hepth{0307211}].

3745: 

3746: \bibitem{Tseytlin2}

3747: A.A. Tseytlin, \emph{On non-abelian generalization of Born-Infeld

3748:   action in string theory}, \npb{501}{1997}{41} [\hepth{9701125}].

3749: 

3750: \bibitem{Koerber2}

3751: P.~Koerber and A.~Sevrin, \emph{The non-abelian D-brane effective

3752: action through order ${\alpha'}^4$}, \jhep{10}{2002}{046}

3753: [\hepth{0208044}].

3754: 

3755: \bibitem{Sevrin1} A. Sevrin and A. Wijns, \emph{Higher order terms

3756: in the non-abelian D-brane effective action and Magnetic

3757: background fields}, \jhep{08}{2003}{059} [\hepth{0306260}].

3758: 

3759: \bibitem{Nagaoka1}

3760: S. ~Nagaoka, \emph{Fluctuation Analysis of Non-abelian Born-Infeld

3761: Action in the Background Intersecting D-branes}, {\it Prog. Theor.

3762: Phys.} {\bf 110} (2004) 1219 [\hepth{0307232}].

3763: 

3764: \bibitem{Hashimoto1} A. Hashimoto and W. Taylor,

3765: \emph{Fluctuation spectra of tilted and intersecting D-branes from

3766: the Born-Infeld action}, \npb{503}{1997}{193} [\hepth{9703217}].

3767: 

3768: \bibitem{Bilal1}

3769: A.~Bilal, \emph{Higher-derivative corrections to the non-abelian

3770: Born-Infeld action}, \npb{618}{2001}{21} [\hepth{0106062}].

3771: 

3772: \bibitem{Cornalba1}

3773: L. Cornalba, \emph{On the general structure of the non-abelian

3774: Born-Infeld action}, {\it Adv. Theor. Math. Phys.} {\bf 4} (2002)

3775: 1259 [\hepth{0006018}].

3776: 

3777: \bibitem{Seiberg1}

3778: N. Seiberg and E. Witten, \emph{String Theory and noncommutative

3779: geometry}, \jhep{09}{1999}{032} [\hepth{9908142}].

3780: 

3781: \bibitem{Refolli1}

3782: A.~Refolli, A.~Santambrogio, N.~Terzi and D.~Zanon, \emph{$F^5$

3783:   contributions to the non-abelian Born-Infeld action from a

3784:   supersymmetric Yang-Mills five-point function}, \npb{613}{2001}{64}

3785:   [\hepth{0105277}]; Erratum-ibid: \npb{648}{2003}{453}.

3786: 

3787: \bibitem{Grasso1}

3788: D. T. Grasso, \emph{Higher order contributions to the effective

3789: action of N=4 super Yang-Mills}, \jhep{0211}{2002}{012}

3790: [\hepth{0210146}].

3791: 

3792: \bibitem{Fosse1}

3793: L. De Foss\'e, P. Koerber and A. Sevrin, \emph{The uniqueness of

3794: the nonabelian Born-Infeld action}, \npb{603}{2001}{413}

3795: [\hepth{0103015}].

3796: 

3797: \bibitem{Koerber1}

3798: P.~Koerber and A.~Sevrin, \emph{The non-abelian Born-Infeld action

3799:   through order ${\alpha'}^3$}, \jhep{10}{2001}{003}

3800:   [\hepth{0108169}]; \emph{Testing the ${\alpha'}^3$ term in the

3801:   non-abelian open superstring effective action}, \jhep{09}{2001}{009}

3802:   [\hepth{0109030}].

3803: 

3804: \bibitem{DeRoo2}

3805: A.~Collinucci, M.~De~Roo and M.G.C. Eenink, \emph{Supersymmetric

3806: Yang-Mills theory at order ${\alpha'}^3$}, \jhep{06}{2002}{024}

3807: [\hepth{0205150}].

3808: 

3809: \bibitem{Cederwall1} M. Cederwall, B.E.W. Nilsson and D. Tsimpis,

3810: \emph{The structure of maximally supersymmetric Yang-Mills theory:

3811: constraining higher-order corrections}, \jhep{06}{2001}{034}

3812: [\hepth{0102009}].

3813: 

3814: \bibitem{Cederwall2} M. Cederwall, B.E.W. Nilsson and D. Tsimpis,

3815: \emph{D=10 super Yang-Mills at ${\alpha'}^2$},

3816: \jhep{07}{2001}{042} [\hepth{0104236}].

3817: 

3818: \bibitem{Drummond1}

3819: J.~M. ~Drummond, P.J. Heslop, P. S. Howe and S.F. Kerstan,

3820: \emph{Integral invariants in N=4 SYM and the effective action for

3821: coincident D-branes}, \jhep{0308}{2003}{016}   [\hepth{0305202}].

3822: 

3823: \bibitem{Brandt1}

3824: F. Brandt, F. Machado and R. Medina, \emph{The open superstring

3825: 5-point amplitude revisited}, \jhep{07}{2002}{071}

3826: [\hepth{0208121}].

3827: 

3828: \bibitem{Chandia1}

3829: O. Chandia and R. Medina, \emph{4-point effective actions in open

3830: and closed superstring theory}, \jhep{0311}{2003}{003}

3831: [\hepth{0310015}].

3832: 

3833: \bibitem{Machado1}

3834: F. Machado and R. Medina, \emph{The open superstring and the

3835: non-abelian Born Infeld theory}, $Nucl. \ Phys.$ \ {\bf B} \

3836: (Proc. Suppl.) \ {\bf 127} \ (2004) \ 166.

3837: 

3838: \bibitem{Schwarz1}

3839: J. H. Schwarz, \emph{Superstring Theory},

3840: \newjournal{Phys.\ Rept.\ }{PR}{89}{1982}{233}.

3841: 

3842: \bibitem{Mangano1} M. L. Mangano, S. J. Parke, \emph{Multi-parton

3843: amplitudes in Gauge theories}, $Phys. \ Rept.$ {\bf 200}, No.6

3844: (1996) \ 301.

3845: 

3846: \bibitem{Polchinski1}

3847: J.~Polchinski, \emph{String theory}, vol.\ 1: \emph{String

3848: theory}, Cambridge, UK: Univ.\ Pr.\ 1998.

3849: 

3850: \bibitem{Grasso2} D. Grasso, \emph{Higher order contributions to the effective

3851: action of N=2 super Yang-Mills}, \jhep{0409}{2004}{054}

3852: [\hepth{0407264}].

3853: 

3854: \bibitem{Andreev1}

3855: O. D. ~Andreev and A. A. Tseytlin, \emph{Partition function

3856: representation for the open superstring effective action:

3857: cancellation of M\"obius infinities and derivative corrections to

3858: Born-Infeld lagrangian}, \npb{311}{1988}{205}.

3859: 

3860: \bibitem{Medina1}

3861: R. Medina and C.V. Paiva, \emph{Uso de Polilogaritmos Harm\^onicos

3862: na determina\c{c}\~ao de algumas integrais impr\'oprias

3863: m\'ultiplas}, Scientific Initiation work done at Universidade

3864: Federal de Itajub\'a (Minas Gerais, Brasil), portuguese version,

3865: 2004, http://www.ici.unifei.edu.br/mestrado{\underline{

3866: }}fisica/pdf/2003-2004{\underline{ }}relatorio{\underline{

3867: }}claudio{\underline{ }}paiva.pdf .

3868: 

3869: \bibitem{Barreiro1} L. A. Barreiro and R. Medina, work in progress.

3870: 

3871: \bibitem{Kawai1} H. Kawai, D. C. Lewellen and S. -H. H. Tye,

3872: \npb{269}{1986}{1}.

3873: 

3874: \bibitem{Arfken}

3875: G. ~Arfken, \emph{Mathematical Methods for Physicists}, third

3876: edition, San Diego, USA: Academic Press, 1985.

3877: 

3878: \bibitem{Remiddi1}

3879: E.~Remiddi and J.A.M. Vermaseren, \emph{Harmonic polylogarithms},

3880: \ijmpa{15}{2000}{725} [\hepph{9905237}].

3881: 

3882: \bibitem{Vermaseren1}

3883: J.A.M. Vermaseren, \emph{New features of form} [math-ph/0010025].

3884: 

3885: \end{thebibliography}

3886: 

3887: \end{document}

3888: