1: \begin{titlepage}
2:
3: \vspace{-5cm}
4:
5: \title{
6: \hfill{\normalsize UPR-1017-T} \\[1em]
7: {\LARGE $SU(4)$ Instantons on Calabi-Yau Threefolds with ${\mathbb Z}_{2} \times {\mathbb Z}_{2}$ Fundamental Group}
8: \\
9: [1em] }
10: \author{
11: Ron Donagi$^2$, Burt A. Ovrut$^1$, Tony Pantev$^2$ and
12: Ren\'e Reinbacher$^1$ \\[0.5em]
13: {\normalsize $^1$Department of Physics, University of Pennsylvania} \\[-0.4em] {\normalsize Philadelphia, PA 19104--6396}\\
14: {\normalsize $^2$Department of Mathematics, University of Pennsylvania} \\[-0.4em]
15: {\normalsize Philadelphia, PA 19104--6395, USA}\\ }
16: \date{}
17: \maketitle
18: \begin{abstract}
19: \noindent
20: Structure group $SU(4)$ gauge vacua of both weakly and strongly
21: coupled heterotic superstring theory compactified on torus-fibered
22: Calabi-Yau threefolds $Z$ with ${\mathbb Z}_{2} \times {\mathbb
23: Z}_{2}$ fundamental group are presented. This is accomplished by
24: constructing invariant, stable, holomorphic rank four vector bundles
25: on the simply connected cover of $Z$. Such bundles can
26: descend either to Hermite-Yang-Mills instantons on $Z$ or to twisted gauge
27: fields satisfying the Hermite-Yang-Mills equation corrected by
28: a non-trivial flat $B$-field. It is shown that large families of such
29: instantons satisfy the constraints imposed by particle physics
30: phenomenology. The discrete parameter spaces of those families are
31: presented, as well as a lower bound on the dimension of the continuous
32: moduli of any such vacuum. In conjunction with ${\mathbb Z}_{2} \times
33: {\mathbb Z}_{2}$ Wilson lines, these $SU(4)$ gauge vacua can lead to
34: standard-like models at low energy with an additional $U(1)_{B-L}$
35: symmetry. This $U(1)_{B-L}$ symmetry is very helpful in naturally
36: suppressing nucleon decay.
37:
38:
39:
40: \end{abstract}
41:
42: \thispagestyle{empty}
43:
44: \end{titlepage}
45:
46:
47:
48:
49:
50: \section{Introduction}
51:
52: The compactification of both weakly coupled \cite{gsw} and strongly
53: coupled \cite{hw1,hw2,w1} heterotic string theory on smooth Calabi-Yau
54: threefolds can potentially lead to phenomenologically interesting
55: particle physics models in the four uncompactified dimensions. In both
56: cases, the physical properties of the low energy theory are governed
57: by the $E_{8}$ gauge field vacua on the internal space. These vacua
58: determine the low energy gauge group and the number of quark and
59: lepton families. In addition, they contribute moduli which play an
60: important role in the structure of non-perturbative superpotentials
61: \cite{bdo1,bdo2} and in vacuum stability \cite{bo}. It follows that
62: constructing gauge field vacua, that is, instantons with structure
63: group $G\subseteq E_8$ on Calabi-Yau threefolds, is of fundamental
64: importance in the attempt to derive the real world from superstring
65: theory.
66:
67: It is well known how to construct $G$-instantons on four dimensional
68: real manifolds and on complex surfaces. This is accomplished via the
69: so-called ADHM construction \cite{adhm}. This construction yields all
70: information about the gauge connections and the associated holomorphic
71: vector bundles. However, until relatively recently, little was known
72: about how to construct arbitrary $G$-instantons on Calabi-Yau
73: threefolds. This was due to several factors, not the least of which
74: was the almost complete lack of knowledge about the metrics on such
75: manifolds. This situation changed significantly with the work of
76: \cite{fmw1}, \cite{don} and \cite{fmw2} where stable, holomorphic vector
77: bundles with arbitrary structure group were constructed on simply
78: connected elliptically fibered Calabi-Yau threefolds. These papers
79: exploited the theorem of Uhlenbeck and Yau \cite{uyau} and Donaldson
80: \cite{d} that proves that any such bundle must admit a unique gauge
81: connection that satisfies the hermitian Yang-Mills equation. That is,
82: finding $G$-instantons on an elliptically fibered Calabi-Yau threefold
83: is equivalent to constructing stable holomorphic vector bundles with
84: structure group $G$. It remains unknown how to solve the hermitian
85: Yang-Mills equation, but \cite{fmw1}, \cite{don} and \cite{fmw2} open the door
86: to constructing, relatively easily, the associated vector bundles.
87:
88: These results were extended and applied to both weakly and strongly
89: coupled heterotic superstring vacua in a series of papers. In
90: \cite{low2,dlow,low1}, vector bundles were constructed on simply
91: connected Calabi-Yau threefolds to produce low energy grand unified
92: theories with gauge groups such as $E_6$, $SO(10)$ and $SU(5)$. This
93: work was carried out within the context of compactified strongly
94: coupled heterotic string theory, known as heterotic M-theory
95: \cite{losw1,losw2}. Physically relevant phenomena such as $N=1$
96: supersymmetry breaking \cite{low4,low3}, five-brane moduli
97: \cite{dow}, non-perturbative superpotentials \cite{lopr} and small
98: instanton phase transitions \cite{opp} were explored in these
99: theories. These concepts were also used to develop a new theory of the
100: early universe, termed Ekpyrotic cosmology \cite{kost}. Although
101: originally presented in heterotic M-theory, most of this work applies
102: to compactifications of the weakly coupled heterotic string as well.
103:
104: As constructed, these grand unified theories need non-zero vacuum
105: expectation values to develop in the low energy theory in order to
106: break to the standard model gauge symmetry. A more direct way to
107: produce standard-like models is to allow Wilson lines. Mathematically
108: this amounts to adding non-trivial flat bundles to the holomorphic vector
109: bundles. However, this requires two new objects, torus-fibered
110: Calabi-Yau threefolds with non-trivial fundamental group and stable,
111: holomorphic vector bundles on such manifolds, to
112: be constructed. The first steps in this direction were presented in a
113: series of papers \cite{dopw-i,dopw-ii,dopw-iii,dopw-iv}. In these
114: papers, torus-fibered Calabi-Yau threefolds $Z$ with non-trivial
115: ${\mathbb Z}_2$ fundamental group were constructed and relevant
116: geometrical properties, such as the structure of $H_4(Z,{\mathbb Z})$,
117: were discussed. Then, a general method for building stable,
118: holomorphic vector bundles with structure group $G=SU(5)$ on these
119: manifolds was presented. The result was standard-like models at low
120: energy, obtained by breaking the $SU(5)$ low energy gauge group to
121: $SU(3)_C\times SU(2)_L\times U(1)_Y$ with a ${\mathbb Z}_2$ Wilson
122: line. These models, although promising, suffer from a potential
123: deficiency. That is, there is no natural mechanism to suppress nucleon
124: decay in this context. Although such suppression might be achieved, by
125: discrete symmetries for example, it has to be implemented on a
126: restricted, case by case basis.
127:
128: This has prompted the authors to consider a major generalization of
129: the work in \cite{dopw-i,dopw-ii,dopw-iii,dopw-iv} to include a much
130: wider range of fundamental groups and structure groups. The question
131: of nucleon decay has led us to construct, as the next step,
132: torus-fibered Calabi-Yau threefolds with ${\mathbb Z}_2\times {\mathbb
133: Z}_2$ fundamental group supporting stable, holomorphic vector bundles
134: with $G=SU(4)$ structure group. This choice of
135: structure group was motivated by the results of \cite{w2,w3}, where it
136: was pointed out that such theories could lead to standard-like models
137: with an additional $U(1)_{B-L}$ gauge group. This $U(1)_{B-L}$
138: symmetry, although not a complete mechanism for nucleon decay
139: suppression, is very helpful in suppressing the most egregious
140: dimension four operators. Also see \cite{rt}. In \cite{opr-i}, we
141: constructed torus-fibered Calabi-Yau threefolds $Z$ with ${\mathbb
142: Z}_2\times {\mathbb Z}_2$ and ${\mathbb Z}_2\times {\mathbb Z}_2\times
143: {\mathbb Z}_2$ fundamental group over a base surface $d{\mathbb
144: P}_9$. A detailed discussion of their geometry was given and their
145: moduli spaces presented. An important ingredient in constructing
146: standard-like models is the homology group $H_4(Z,{\mathbb Z})$. The
147: explicit construction of this group in the present context is
148: non-trivial. For this reason, we devoted a second paper \cite{opr-ii}
149: entirely to a discussion of $H_4(Z,{\mathbb Z})$ and its
150: properties. However, there remains the fundamental issue of
151: constructing $G=SU(4)$ gauge instantons on these Calabi-Yau
152: threefolds. That is explicitly accomplished in this paper.
153:
154: Specifically, we do the following. In Section~\ref{Z}, we review the
155: results of \cite{opr-i} and \cite{opr-ii} that will be required in our construction.
156: We also use this opportunity to set our notation. Section~\ref{physics} is devoted
157: to a general discussion of rank four vector bundles in this
158: context. First, bundles are constructed on elliptically fibered
159: Calabi-Yau threefolds $X$ with trivial fundamental group which admit a
160: freely acting automorphism group ${\mathbb Z}_2\times {\mathbb
161: Z}_2$. Any such bundle $V$ that is, additionally, equivariant under
162: ${\mathbb Z}_2\times {\mathbb Z}_2$ will then descend to produce a
163: stable, holomorphic rank four vector bundle $V_Z$ on the torus-fibered
164: Calabi-Yau threefold $Z=X/({\mathbb Z}_2\times {\mathbb Z}_2)$ with
165: ${\mathbb Z}_2\times {\mathbb Z}_2$ fundamental group. The
166: Donaldson-Uhlenbeck-Yau theorem allows us to identify a unique
167: connection of the bundle $V_{Z}$ as a solution of the hermitian
168: Yang-Mills equations on $Z$. Equivariance is a very strong constraint
169: in this context. A weaker condition, which also leads to physical
170: vacua, is the requirement that $V$ only be invariant under the action
171: of ${\mathbb Z}_2\times {\mathbb Z}_2$. For a detailed discussion of
172: invariance versus equivariance see Appendix~\ref{appendix2}. Stable
173: holomorphic bundles $V$ on $X$ which are invariant, but not
174: equivariant, give rise to twisted vector bundles $V_{Z}$ on $Z$, where
175: the twisting is given by a flat $B$-field $\boldsymbol{B}$. See
176: \cite{c,cks,dp}, Section~\ref{physics} and Appendix~\ref{appendix2}.
177: Every such bundle comes equipped with a gauge field $A$ which solves
178: its string theory equation of motion. In the presence of a
179: flat $B$-field $\boldsymbol{B}$, this equation
180: is a version of the hermitian Yang-Mills equation in which the right
181: hand side is modified by $\boldsymbol{B}$. That is, we have $F_{A} =
182: \boldsymbol{B}\cdot \operatorname{id}$. See Section~\ref{physics} for
183: details. Since the $B$-field is flat, we still obtain an $N=1$
184: supersymmetric theory in the four dimensional low-energy limit. When
185: the bundle $V$ is actually equivariant, then the $B$-field is gauge
186: equivalent to zero and we get back the usual hermitian Yang-Mills
187: instantons. Mathematically, the invariant stable holomorphic bundles on
188: $X$ correspond to stable, holomorphic vector bundles on a suitably
189: defined Deligne-Mumford gerbe over $Z$ \cite{dp}, but we will not
190: discuss this interpretation here.
191:
192: We outline the method we will use to produce invariant bundles, namely, by
193: extensions from invariant rank two bundles $V_i,\;i=1,2$ on $X$. In
194: each particular case, when we want to decide whether we are working in
195: a trivial $B$-field background, we check the resulting bundles for
196: equivariance as well. The constraints on $V$ required by particle
197: physics phenomenology, such as three families of quarks and leptons,
198: are presented and discussed. It was shown, for example in
199: \cite{schoen}, that any threefold $X$ of the above structure must be a
200: fiber product $X=B\times_{\cp{1}}B^{'}$ of two rational elliptic
201: surfaces $B$ and $B^{'}$. Each invariant rank two bundle $V_i$ on $X$
202: is explicitly composed from a rank two invariant bundle $W_i$ on $B$
203: and an invariant line bundle $L_i$ on $B^{'}$. The bundles $W_i$ and
204: $L_i$ are constructed, their moduli spaces described, and their Chern
205: character calculated in Section~\ref{Wbundles} and
206: Section~\ref{lbundles} respectively. Using this data, in
207: Section~\ref{Vbundles} rank two invariant vector bundles $V_i$ are
208: produced on $X$. These bundles are then used to construct ${\mathbb
209: Z}_2\times {\mathbb Z}_2$ invariant rank four vector bundles $V$ on
210: $X$ by extension. In order for such vector bundles to admit a
211: connection satisfying the hermitian Yang-Mills condition, it is
212: necessary that they have the property of stability. The concept of
213: stable holomorphic vector bundles is recalled in Section~\ref{stable}
214: and a necessary condition for stability presented. A more complete,
215: technical description of stability is given in
216: Appendix~\ref{appendix1}. Section~\ref{numerics} is devoted to
217: explicitly demonstrating that there is a large class of rank four
218: vector bundles $V$ on $X$ that satisfy the standard model constraints
219: and the necessary condition for stability. The discrete parameter
220: space of these vector bundles is presented. In Section~\ref{Stable} we
221: prove that, subject to a mild restriction on this parameter space, all
222: of these ${\mathbb Z}_2\times {\mathbb Z}_2$ invariant holomorphic
223: vector bundles are stable. This section relies on the technical
224: discussion of stability given in Appendix~\ref{appendix1}. Finally, in
225: Section~\ref{conclusions} we draw all of these results together into a
226: conclusion. We present a large class of rank four, stable holomorphic
227: vector bundles $V$ on $X$ that are ${\mathbb Z}_2\times {\mathbb Z}_2$
228: invariant and explicitly give their discrete parameter space. We also
229: use results derived throughout the paper to put a lower bound on the
230: complex dimension of the continuous moduli space of any such bundle
231: $V$ and give a concrete example. As we have already explained, any such
232: bundle $V$ descends to a twisted hermitian Yang-Mills connection of
233: standard model type on the torus-fibered Calabi-Yau threefold
234: $Z=X/({\mathbb Z}_2\times {\mathbb Z}_2)$ with ${\mathbb Z}_2\times
235: {\mathbb Z}_2$ fundamental group, thus accomplishing our goal. In the
236: process, we also analyze the conditions for the standard
237: model vacuum to appear in a trivial $B$-field background. We find that,
238: in our setting, the triviality of the $B$-field is incompatible with the
239: three generation condition. Although it is likely that our method can
240: be modified to produce standard model bundles with trivial $B$-field
241: background, we will not pursue this in the present paper since the
242: phenomenologically relevant constraints coming from the standard
243: model are insensitive to this issue.
244:
245: Admittedly, this paper is rather technical. This arises from the fact
246: that one is, in principle, solving the very non-linear hermitian
247: Yang-Mills equation on complicated Calabi-Yau threefolds whose metrics
248: are unknown. From this point of view, it seems remarkable that this
249: can be accomplished at all, let alone with the relatively
250: straightforward mathematical techniques used in this paper. Be that as
251: it may, the results of this paper are immediately relevant to pure
252: particle physics, opening the door to the construction of
253: standard-like models with suppressed nucleon decay within the context
254: of heterotic superstring theory. In several upcoming papers, we extend
255: our results to more general fundamental groups and structure groups and
256: give a detailed analysis of the Wilson line symmetry breaking patterns
257: \cite{dlor}. We will also present standard-like models with $U(1)_{B-L}$
258: symmetry \cite{dopr-iii} within this context.
259:
260:
261:
262:
263:
264:
265:
266:
267:
268:
269:
270:
271:
272:
273:
274:
275:
276:
277:
278: \section{Calabi-Yau Threefolds $Z$ with ${\mathbb Z}_2 \times {\mathbb
279: Z}_2$ Fundamental Group}\label{Z}
280:
281: Calabi-Yau threefolds $Z$ with non-trivial first homotopy group
282: \be\label{2.1} \pi_1(Z)={\mathbb Z}_2 \times {\mathbb Z}_2 \ee can be
283: constructed as follows. Let $X$ be a Calabi-Yau threefold with trivial
284: fundamental group that admits a freely acting ${\mathbb Z}_2 \times
285: {\mathbb Z}_2$ group of automorphisms preserving the volume form. Then
286: the quotient \be Z=X/({\mathbb Z}_2 \times {\mathbb Z}_2) \ee is a
287: smooth Calabi-Yau threefold with first homotopy group
288: (\ref{2.1}). There are many types of Calabi-Yau threefolds $X$ with
289: this property. However, for the construction of standard-like model
290: vacua in heterotic superstring theory, it is expedient \cite{dlow} to
291: choose $X$ to be elliptically fibered. That is, $X$ is fibered over a
292: base $B^{'}$, \be \pi: X \to B^{'}, \ee with the generic fiber of
293: $\pi$ being isomorphic to a smooth torus $T^2$. Furthermore, this
294: fibration admits a global section \be \sigma: B^{'}\to X, \ee which
295: turns each generic fiber into an elliptic curve. The mapping $\sigma $
296: is referred to as the zero section of $X$. As discussed in
297: \cite{opr-i}, choosing $X$ to be elliptically fibered renders
298: $Z=X/({\mathbb Z}_2 \times {\mathbb Z}_2)$ to be torus fibered, since
299: $Z$ does not admit a global section.
300:
301: The base space $B^{'}$ of an elliptically fibered Calabi-Yau threefold is restricted to be either a del Pezzo surface, $d{\mathbb P}_r$ for $r=1,\dots,8$, a Hirzebruch surface $F_r$ for any non-negative integer $r$, an Enriques surface $E$, certain blow-ups of these surfaces or a rational elliptic surface $d{\mathbb P}_9$. However, for the purpose of constructing Calabi-Yau threefolds admitting freely acting automorphism groups with ${\mathbb Z}_2$ factors, it is particularly convenient \cite{opr-i} to choose
302: \be\label{2.5}
303: B^{'}\cong d{\mathbb P}_9.
304: \ee
305: The reason for this is the following. Recall that any rational elliptic surface is itself elliptically fibered over a $\cp{1}$ base with the projection map
306: \be
307: \beta^{'}: B^{'}\to {\mathbb P}^{1'}.
308: \ee
309: Using this fact, it was shown, for example in \cite{opr-i}, that any Calabi-Yau threefold $X$ elliptically fibered over base (\ref{2.5}) must be the fiber product over $\cp{1}$ of two rational elliptic surfaces $B$ and $B^{'}$. That is,
310: \be\label{2.7}
311: X=B\times_{\cp{1}}B^{'}.
312: \ee
313: The fiber product permits the natural projections
314: \be
315: \xymatrix{
316: & X \ar[dl]_-{\pi'} \ar[dr]^-{\pi} & & \\
317: B \ar[dr]_-{\beta} & & B' \ar[dl]^-{\beta'}\ar@{}[r]|-{.} & \\
318: & \cp{1} & &
319: }
320: \ee
321: For any rational elliptic surface $B$, it is relatively straightforward to construct all of the involutions
322: \be
323: \tau_B: B \to B,
324: \ee
325: where $\tau_B^{2}=id$. This was carried out in \cite{opr-i}. By appropriately restricting the moduli space of $B$, one can find surfaces that admit an involution and, hence, a ${\mathbb Z}_2 $ automorphism group. Further restrictions of the moduli space allow several commuting involutions, leading to automorphism groups ${\mathbb Z}_2 \times {\mathbb Z}_2$, ${\mathbb Z}_2 \times {\mathbb Z}_2\times {\mathbb Z}_2$ and so on. With this in mind, one chooses two rational elliptic surfaces $B$ and $B^{'}$ each, for example, admitting a ${\mathbb Z}_2 \times {\mathbb Z}_2$ automorphism group. Let $\tau_{Bi}$ and $\tau_{B^{'}i}$ for $i=1,2$ be the associated generators. Then, as shown in \cite{opr-i}, the Calabi-Yau threefold $X$ constructed as the fiber product (\ref{2.7}) of $B$ and $B^{'}$ inherits a freely acting ${\mathbb Z}_2 \times {\mathbb Z}_2$ automorphism group generated by
326: \be\label{invX}
327: \tau_{Xi}=\tau_{Bi}\times_{\cp{1}}\tau_{B^{'}i},\;i=1,2,
328: \ee
329: as desired. This explains our choice of $B^{'}\cong d{\mathbb P}_9$ as the base surface of $X$.
330:
331: The explicit choice of the rational elliptic surfaces $B$ and $B^{'}$ required to produce standard-like model vacua in heterotic superstring theory is rather subtle. In this paper, we are interested in producing realistic theories with ${\mathbb Z}_2 \times {\mathbb Z}_2$ Wilson lines. Therefore, we must choose $B$ and $B^{'}$ to each admit an automorphism group containing ${\mathbb Z}_2 \times {\mathbb Z}_2$. As discussed in \cite{opr-i}, there exists a three parameter family of rational elliptic surfaces $B$ that admit exactly ${\mathbb Z}_2 \times {\mathbb Z}_2$ automorphism groups. However, as will become clear later in this paper, such surfaces do not generically lead to standard-like model vacua. To achieve such vacua, it is essential to further restrict $B$ and $B^{'}$ to each be contained in a specific two-parameter sub-family of rational elliptic surfaces. This two parameter sub-family was described in detail in \cite{opr-i}. Here, we will simply recall the relevant properties. Let $B$ be a $d{\mathbb P}_9$ surface in this two parameter sub-family. Then $B$ necessarily has two properties shared by all rational elliptic surfaces. First, $B$ is an elliptic fibration
332: \be\label{2.11}
333: \beta: B \to \cp{1}.
334: \ee
335: Second, the second homology group is given by
336: \be
337: H_2(B,{\mathbb Z})\cong {\mathbb Z}l \oplus \bigoplus_{i=1}^{9}e_i,
338: \ee
339: where $l$ is the pull-back of a line in $\cp{2}$ and $e_i,\;i=1,\dots,9$ are a set of divisors of $B$ with self-intersection $-1$. These divisors each intersect every fiber of (\ref{2.11}) only once. Note, however, that for $B$ in the two parameter sub-family, not all of the $e_i$ are sections of this fibration. More precisely, $e_3, e_5$ and $e_8$ each consist of two components. Hence, they are not irreducible and smooth, properties required of any section.
340:
341:
342: Generic $B\cong d{\mathbb P}_9 $ have twelve singular $I_1$ fibers. However, as shown in \cite{opr-i}, the restriction of $B$ to the two parameter sub-family coalesces these fibers pairwise to produce six $I_2$ Kodaira fibers. These $I_2$ fibers are reducible, each being of the form
343: \be
344: n_i \cup o_i
345: \ee
346: for $i=1,\dots,6$. Another important consequence of restricting $B$ to the two parameter sub-family is that the maximal number of commuting involutions is increased, from two in the three parameter case, to three. That is, $B$ has a ${\mathbb Z}_2 \times {\mathbb Z}_2\times {\mathbb Z}_2$ automorphism group generated by $\tau_{Bi},\;i=1,2,3$. Be that as it may, only two generators,
347: \be\label{2.14}
348: \tau_{Bi},\;i=1,2,
349: \ee
350: lift to a freely acting involution on $X$. Hence, we consider only the ${\mathbb Z}_2 \times {\mathbb Z}_2$ sub-group of automorphisms generated by (\ref{2.14}). It was shown in \cite{opr-i} that these generators have the form
351: \be\label{ex042}
352: \tau_{B1}=t_{e_6}\circ \alpha_B,\;\;\;\tau_{B2}=t_{e_4}\circ \alpha_B
353: \ee
354: where $\alpha_B$ is an involution on $B$ that leaves the zero section $e_9\equiv e$ invariant and $e_4, e_6$ are sections of $B$ that intersect each fiber at a point of order two. Note that
355: \be
356: \tau_{B1}\circ\tau_{B2}=t_{e_4+e_6}
357: \ee
358: which acts as a pure translation on the fibers.
359: %Since only a ${\mathbb Z}_2\times {\mathbb Z}_2$ sub-group of automorphisms li%fts to a free action on $X$, what, then, is the purpose of restricting $B$ to %the two parameter sub-family? The reason is the following.
360:
361: A class in $H_2(B,\mathbb Z)$ is called invariant if it is not transformed under the action of the ${\mathbb Z}_2 \times {\mathbb Z}_2$ automorphism group. Denote by
362: \be
363: H_2(B,\mathbb Z)^{inv} \subset H_2(B,\mathbb Z)
364: \ee
365: the homology sub-space of invariant classes. It was shown in \cite{opr-ii} that for $B$ in the two parameter sub-family
366: \be
367: \rank H_2(B,\mathbb Z)^{inv}=4.
368: \ee
369: Furthermore, a set of generators of $H_2(B,\mathbb Z)^{inv}\otimes {\mathbb Q}$ is given by
370: \be
371: i,\;f,\;n_1+o_2,\;M,
372: \ee
373: where
374: \be
375: \begin{split}
376: i=&2e_6+2e_4-n_1+o_2,\\
377: M=&2e_2-2e_9-n_1-n_2,\\
378: \end{split}
379: \ee
380: $f$ is the fiber class of $\beta: B \to \cp{1}$ and $n_1, n_2$ and $o_2$ are components of some $I_2$ fibers.
381: This choice of generators is motivated by their intersection numbers displayed in Table~\ref{1.6}. Note, in particular, that $M$ is orthogonal to the three other classes.
382: \begin{table}[!ht]
383: \begin{center}
384: \begin{tabular}{|c||c|c|c|c|} \hline
385: & $i$ & $f$ & $n_1+o_2$& $M$ \\ \hline\hline
386: $i$ & $-4$ & $4$ & $4$ & $0$ \\ \hline
387: $f$ & $4$ & $0$ & $0$ & $0$ \\ \hline
388: $n_1+o_2$ & $4$ & $0$ & $-4$ & $0$ \\ \hline
389: $M$ & $0$ & $0$ & $0$ & $-4$ \\ \hline
390: \end{tabular}
391: \caption{The intersection numbers of the four generators of $H_2(B,{\mathbb Z})^{inv}\otimes {\mathbb Q}$.}
392: \label{1.6}
393: \end{center}
394: \end{table}
395: It can easily be seen that these four invariant generators do not generate $H_2(B,{\mathbb Z})^{inv}$. For example,
396: \be\label{ex08}
397: \frac{1}{2}(i+M+f)=e_6+e_4+e_2-e_9-n_1-n_2+f
398: \ee
399: is an integral class in $H_2(B,{\mathbb Z})^{inv}$. We would not have found this class if we had restricted the coefficients of $i,f, n_1+o_2$ and $M$ to be integers. We will have to deal with issues arising from this subtlety later in this paper.
400: As discussed above, choosing both $B$ and $B^{'}$ to be in the two parameter sub-family allows one to construct a Calabi-Yau threefold $X=B\times_{\cp{1}}B^{'}$ that admits a freely acting ${\mathbb Z}_2 \times {\mathbb Z}_2$ group of automorphisms. This is generated by the involutions
401: \be
402: \tau_{Xi}=\tau_{Bi}\times_{\cp{1}}\tau_{B^{'}i},\;i=1,2,
403: \ee
404: where $\tau_{Bi}$ and $\tau_{B^{'}i}$ for $i=1,2$ are involutions generating ${\mathbb Z}_2 \times {\mathbb Z}_2$ on $B$ and $B^{'}$ respectively. Furthermore, as shown in detail in \cite{opr-ii}, $X$ inherits a homology sub-group
405: \be
406: H_4(X,{\mathbb Z})^{inv} \subset H_4(X,{\mathbb Z}),
407: \ee
408: each of whose classes is invariant under ${\mathbb Z}_2 \times {\mathbb Z}_2$ and which has
409: \be
410: \rank H_4(X,{\mathbb Z})^{inv}=7.
411: \ee
412: As we will see, both the rank and the structure of $H_4(X,{\mathbb Z})^{inv}$, which derive from the properties of $H_2(B,{\mathbb Z})^{inv}$, are required to achieve standard-like model vacua.
413: %This explains our restriction of both $B$ and $B^{'}$ to the two parameter sub-family.
414: