Dear Sarjana, SERS is a way to enhance the Raman signal based on a nice trick. When Raman and Krishnan discovered the phenomena they used solar light and then mercury lamps, but only the laser made it a spectroscopic technique due to the high intensity of the laser radiation... but even laser energy density is low for Raman, because it is a very inefficient phenomena (usually just 1 in 10000 photons excite a molecule). The trick of SERS was discovered by Fleishman, Hendra and McQuillan in 1973-74 using rough silver electrodes to measure the pyridine spectrum. They thought initially that it was due to the huge sufarce area of this electrode, but around 1977 Van Duyne et al and others understood that it was caused by a very local concentration of electromagnetic field (EM), plus some chemical contribution. These two contributions are the responsible of the SERS enhancement. Often the EM contribution is the bigger, and it is due to localized surface plasmon resonances (LSPR) or hot-spots.

Why do you need a metal?

Properly, more than a metal you need some material with free chages, better if they are electrons. When an EM field, like that of a light, for example a laser, hit a material with free electrons, they interact with the EM field, well mainly with the electric part of the EM field, so you have something like waves of electrons in the surface of your metal. These are Surface plasmons resonances

Why do you need it be rough??

Because we want it localized. Why? because we want a crazy high field concentration close to our molecules to excite them.

The roughnes you need is nanometric, and therefore you get SERS with nanoparticles, for example nanospheres. Imagine a nanosphere in an electric field (your laser) the charges inside oscillate with the field, now take a snapshot: the electrons are all in the north pole of the sphere, so the south pole is now positive. the next time you have a reverse situation, so you can generate a highly concentrate field there and if your molecule is chemically bonded, adsorbed or close to the sphere, it will experience the effect of such a high field and the probability to be excited will be increased by a million times or more.

The thing is even better if you have a dimer, two particles very very close each other, in the gap between them the field can concentrate even more and therefore if your molecule is there, its Raman signal will be even higher.

Let me draw it with an example at the macro level: imagine you have a sea, it is wavy, now it hits a wide beach, not so much effect, isn´t it?. Now imagine it reach a rocky rough coast, when it hits the water splash out, you are concentrating the energy. That is the function of the roughness of SERS substrates.

There are nice reviews and books about SERS, here you have a couple:

Article Surface-enhanced Raman spectroscopy: A brief retrospective

Hope it helps

Simply it produced resonance phenomenon in metal, increased EM signals and in this way EM factor play major contributions in enhancement of raman factor due to localized electron.

Just an additional point, the roughness should preferably be nanoscale roughness rather than microscale/centimeter-scale roughness. Nanoscale roughness on metallic surface leads to the confined oscillation of electrons upon excitation of a wavelength that is typically larger than the 'roughness' dimension.

---------------------------------

High quality Silver (Ag) nanocubes and Surface-enhanced Raman Scattering (SERS) substrates available at:

https://silverfactorytechnology.com/