When you increase the surface roughness of metal film the amount of absorbed light will increase. The reason is that surface reflectance decreases and more light penetrates into the films, so more of it will be absorbed - A =1-R (I assume that T=0). This is the way they increase the efficiency of Si solar cell. Have in mind, however, that the extinction coefficient (k) is material propertie and it is not possible it to be changed only be surface roghnening.

In general I agree with the answer of Tsvetanka Babeva. However I would like to add that reflectance is decreased only when the surface roughness is of specific kind, that is the surface is textured so that incident light is reflected only in the forward direction by the slanted surface (due to surface roughness/texture). This is called light trapping that is used in Si solar cells to increase light coupling into the cell by effectively decreasing the back-reflection of light.  But this does not mean a change in extinction coefficient, which is a fundamental material property. More light is absorbed in cell because more light enters it.

I absolutely agree with Dinesh Pandya. Thank you for the amendments

Thank you so much for your valuable answer. I also agree with the answer of Tsvetanka Babeva and Dinesh Pandya. But I have found the decrease in refractive index and extinction coefficient. Is it possible?

If I can have a look at your data may I can advise further.

You should be very careful when calculating optical properties of films with rough surfaces. The obtained data depend strongly on the model used. Do you use ellipsometry for optical characterization?

I do agree. It is very difficult to calculate ' k ' and 'n' for rough films that have significant diffuse reflectance. The formula for T and R from which these two optical constants are calculate assume multiple internal reflections in the film that occur only for planar interfaces. Therefore one has to be very careful. Ellipsomertry and effective medium theory have to be applied for correct evaluation. Data needs to be carefully processed using appropriate software.

Dear Sir/madm, 

I have attached the extinction coefficient figure here.

I have used Spectroscopic ellipsometry (VASE) to calculate optical constants of Ti thin films for varying surface roughness.

 I will specify Tsvetankas  answer . Not only increases the absorption and decreases transmittance due to scattering on surface roughness .

I have both type of samples

1. same thickness but different surface roughness

2. different thickness and different surface roughness

Yeah I agree with your answer but what about extinction coefficient, which is material properties.

The increase in surface roughness led to an increase in absorbance due to the decrease in reflectance . but we must take into account that extinction coefficient does not affected because it is constant for any material