next up previous
Next: 2.4 Effective Bulk and Up: 2 Classical Dynamics Previous: 2.2 Center of Mass

2.3 Induced Bulk and Boundary Actions

In this subsection we wish to discuss those features of the brane classical dynamics which are instrumental for the subsequent evaluation of the quantum path-integral.
In agreement with the restricted reparametrization invariance of the action (10), as discussed in the previous subsection, we first set up a ``canonical formulation'' which preserves that same symmetry through all computational steps. This means that all the world indices $m\ , n\ , \dots$ are raised, lowered and contracted by means of the center of mass metric $\overline{g}_{\, mn}$. From the brane action (10) we extract the brane relative momentum $P^m{}_\mu$ and the corresponding Hamiltonian $H$:

$\displaystyle P^m{}_\mu$ $\textstyle \equiv$ $\displaystyle {\partial L\over \partial \partial_m\, Y^\mu(\, \sigma\, )}= -
\overline{g}^{\, mn}\, \partial_n\, Y_\mu(\, \sigma\, )\,$  
$\displaystyle H$ $\textstyle \equiv$ $\displaystyle P^{ m}{}_\mu\, \partial_m\, Y^\mu -L
= -{1\over 2}\, \left[ \, \overline{g}_{ \, m n}\, P^{
m}{}_\mu\, P^{n}{}^\mu + m_{p+1} \, p\, \right]\ .$ (11)

Thus, we can write the action in the following canonical form
$\displaystyle S$ $\textstyle =$ $\displaystyle \int_{\Sigma_{p+1}} d^{ p+1}\sigma\, \sqrt{\, \overline g\,} \,
\left(\, P^{ m}{}_\mu \, \partial_m \, Y^{\mu} - H\, \right)$  
  $\textstyle =$ $\displaystyle \int_{\Sigma_{p+1}} d^{ p+1}\sigma\, \sqrt{\, \overline g\, }\,
\...
...e{g}_{\, ab}\, P^{ a}{}_\mu \, P^{ b}{}^\mu +
p\, m_{p+1}\, \right) \, \right].$ (12)

The first term in Eq.(12) can be rewritten as follows
\begin{displaymath}
\int_{\Sigma_{p+1}} d^{ p+1}\sigma\, \sqrt{\, \overline g\, ...
...,\sqrt{\,\overline
g\,}
\, P^{ m}{}_\mu\, \right)\, \right]\ .
\end{displaymath} (13)

According to Eq. (13) we can write the total action as the sum of a boundary term plus a bulk term
$\displaystyle S$ $\textstyle =$ $\displaystyle S_B[\, \partial {\Sigma_{p+1}} \, ] + S_J[\, {\Sigma_{p+1}}
\, ]$  
  $\textstyle =$ $\displaystyle \int_{\Sigma_{p}} d^{p} s \, \sqrt{h}\, N _{n}\, p^n{}_\mu\,
y^\m...
... m}{}_\mu - \int_{\Sigma_{p+1}} d^{p+1}\sigma\,
\sqrt{\, \overline g\, }\, H\ ,$ (14)

where $ p^n{}_\mu$ and $y^\mu$ are the momentum and coordinate of the boundary, $ d^{p} s \sqrt{h} N _{n}$ represents the oriented surface element of the boundary, and $\nabla_m$ stands for the covariant derivative with respect to the metric $\overline{g}_{\, mn}$. The distinctive feature of this rearrangement is that the bulk coordinates $Y
^\mu(\sigma)$ enter the action as Lagrange multipliers enforcing the classical equation of motion:
\begin{displaymath}
{\delta S\over\delta \, Y^\mu(\, \sigma\, ) }=0
\quad \Longr...
...\left(\,
\sqrt{\, \overline g\, }\, P^{m}{}_\mu\, \right)=0\ .
\end{displaymath} (15)

The general solution of Eq. (15) may be expressed as follows

$\displaystyle P^{ m}{}_\mu$ $\textstyle =$ $\displaystyle \overline{g}^{ \, mn}\,\partial_n\, \phi_\mu +
{1\over p!}\, \ove...
...,
Y^{ \mu_{p+1}} +
\partial_{[\, m_2}\, A_{\mu m_3 \dots m_{p+1}\, ] }\,\right)$  
  $\textstyle =$ $\displaystyle \overline{g}^{ \, mn}\, \partial_n \, \phi_\mu +
{1\over p!}\, \o...
...\, P^{0)}{}_{\mu m_2\dots m_{p+1}}+
F(\, A\, )_{\mu m_2\dots m_{p+1}}\, \right)$ (16)

with
$\displaystyle \Box_{\overline g} \, \phi^\mu(\, \sigma\, )$ $\textstyle =$ $\displaystyle 0\,$ (17)
$\displaystyle \partial_m \, P^{0)}{}_{\mu\mu_2\dots \mu_{p+1}}$ $\textstyle =$ $\displaystyle 0\ .$ (18)

Here, the components $\phi_\mu$ represent local harmonic modes on the bulk, and $\overline{\epsilon}^{\, mm_2\dots m_{p+1}}\equiv (\overline
g)^{-1/2}\, \delta^{[\, m m_2\dots m_{p+1}\, ]}$ stands for the totally antisymmetric tensor. Moreover, the constant antisymmetric tensor $ P^{0)}{}_{\mu\mu_2\dots \mu_{p+1}}$ represents the volume momentum zero-mode, or collective-mode, that describes the global volume variation of the brane.

In order to be able to treat $\phi$, $P^{0)}$ and $A$ as independent oscillation modes, we demand that the following orthogonality relations are satisfied:

$\displaystyle {\mathcal{C}} _{\mathrm{I}}$ $\textstyle \equiv$ $\displaystyle \int_{\Sigma_{p+1}} d^{p+1}\sigma\, \sqrt{\, \overline g\, }\,
\o...
...hi^\mu\,
\partial_{m_2}\, Y^{\mu_2}\dots \partial_{m_{p+1}}\, Y^{\mu_{p+1}} =0,$ (19)
$\displaystyle {\mathcal{C}} _{\mathrm{II}}$ $\textstyle \equiv$ $\displaystyle \int_{\Sigma_{p+1}} d^{p+1}\sigma \,\sqrt{\,\overline g\,}
\overl...
..._m\, \phi^\mu\,
\partial_{[\, m_2}\, P ^{0)} _{\mu m_2 \dots m_{p+1}\, ]} =0\ ,$ (20)
$\displaystyle {\mathcal{C}} _{\mathrm{III}}$ $\textstyle \equiv$ $\displaystyle \int_{\Sigma_{p+1}} d^{p+1}\sigma\, \sqrt{\,\overline g\,}\,
\ove...
...\, \partial_m\, \phi^\mu\,
\partial_{[\, m_2}\, A_{m_3 \dots m_{p+1}\, ]} =0\ .$ (21)

The three orthogonality constraints on the bulk determine the field behavior on the boundary through Stokes' theorem. In particular ${\mathcal{C}} _{\mathrm{I}}$ gives
\begin{displaymath}
{\mathcal{C}} _{\mathrm{I}}=\int_{\Sigma _p}\, \phi^\mu \,
d...
...\mu_{p+1}}=0\quad\Longrightarrow\quad \phi^\mu(\,\vec s\, )=0,
\end{displaymath} (22)

which is a Dirichlet boundary condition, whereas in ${\mathcal{C}} _{\mathrm{III}}$ two integrations remain:
$\displaystyle {\mathcal{C}} _{\mathrm{III}}$ $\textstyle =$ $\displaystyle \int_{\Sigma _{p}} d^{p}s \,
F_\mu{}^{m_2\dots m_{p+1}}\, y^\mu\partial_{m_2}\, y^{\mu_2}\dots
\partial_{m_{p+1}}\, y^{\mu_{p+1}}$  
    $\displaystyle -\int_{\Sigma_{p+1}} d^{p+1}\sigma\, \sqrt{\, \overline g\, }
\le...
...\partial_{ m_3}\, Y^{\mu_3}\dots \partial_{ m_{p+1}}\, Y^{\mu_{p+1}}
\right)\ .$ (23)

Since the two integrations are carried over the boundary and the bulk respectively, the orthogonality condition can be satisfied only if each integral is identically vanishing,

$\displaystyle {\mathcal{C}} _{\mathrm{III}} =0 \quad$ $\textstyle \Longrightarrow$ $\displaystyle A^\mu{}_{m_3\dots m_{p+1}}(\, \vec s\, )=\partial_{[\, m_3 }\,
\Lambda^\mu\, {}_{m_4\dots m_{p+1}\, ] }$  
    $\displaystyle \mathrm{and} \quad \partial_{m_2}\, F_\mu{}^{m_2 m_3\dots
m_{p+1}}=0
\ .$ (24)

Thus, $A$ must solve free Maxwell-type equations on the bulk and reduce to a pure gauge configuration on the boundary. Note that under these conditions for $\phi$ and $A$, ${\mathcal{C}} _{\mathrm{II}}$ is satisfied as well.
In summary, the classical solution for the brane momentum reduces the original field content of the model to:


next up previous
Next: 2.4 Effective Bulk and Up: 2 Classical Dynamics Previous: 2.2 Center of Mass

Stefano Ansoldi