It is a particle of a scalar field Φ; a scalar field that always comes with gravity. In standard general relativity, Newton's constant, or equivalently, the Planck mass is always constant. If we "promote" this constant to a dynamical field, what we would get is the dilaton.
So, in Kaluza-Klein theories, after dimensional reduction, the effective Planck mass varies as some power of the volume of compactified space. This is why volume can turn out as a dilaton in the lower dimensional effective theory.
Although string theory naturally incorporates Kaluza–Klein theory (which first introduced the dilaton), perturbative string theories, such as type I string theory, type II string theory and heterotic string theory, already contain the dilaton in the maximal number of 10 dimensions. However, on the other hand, M-theory in 11 dimensions does not include the dilaton in its spectrum unless it is compactified. In fact, the dilaton in type IIA string theory is actually the radion of M-theory compactified over a circle, while the dilaton in E8 × E8 string theory is the radion for the Hořava–Witten model. (For more on the M-theory origin of the dilaton, see .)
In string theory, there is also a dilaton in the worldsheet CFT. The exponential of its vacuum expectation value determines the coupling constant g, as ∫R = 2πχ for compact worldsheets by the Gauss-Bonnet theorem and the Euler characteristic χ = 2 − 2g, where g is the genus that counts the number of handles and thus the number of loops or string interactions described by a specific worldsheet.
Therefore the coupling constant is a dynamical variable in string theory, unlike the case of quantum field theory where it is constant. As long as supersymmetry is unbroken, such scalar fields can take arbitrary values (they are moduli). However, supersymmetry breaking usually creates a potential energy for the scalar fields and the scalar fields localize near a minimum whose position should in principle be calculable in string theory.
Dilaton action 
The dilaton-gravity action is
This is more general than Brans–Dicke in that we have a dilaton potential.
See also 
- Fujii, Y. (2003). "Mass of the dilaton and the cosmological constant". Prog. Theor. Phys. 110 (3): 433–439. arXiv:gr-qc/0212030. Bibcode:2003PThPh.110..433F. doi:10.1143/PTP.110.433.
- Hayashi, M.; Watanabe, T.; Aizawa, I. & Aketo, K. (2003). "Dilatonic Inflation and SUSY Breaking in String-inspired Supergravity". Modern Physics Letters A 18 (39): 2785–2793. arXiv:hep-ph/0303029. Bibcode:2003MPLA...18.2785H. doi:10.1142/S0217732303012465.
- Alvarenge, F.; Batista, A. & Fabris, J. (2005). "Does Quantum Cosmology Predict a Constant Dilatonic Field". International Journal of Modern Physics D 14 (2): 291–307. arXiv:gr-qc/0404034. Bibcode:2005IJMPD..14..291A. doi:10.1142/S0218271805005955.
- Lu, H.; Huang, Z.; Fang, W.; Zhang, K. (2004). "Dark Energy and Dilaton Cosmology". arXiv:hep-th/0409309 [hep-th].
- Wesson, Paul S. (1999). Space-Time-Matter, Modern Kaluza-Klein Theory. Singapore: World Scientific. p. 31. ISBN 981-02-3588-7.