You can do it by 3D Confucian Microscope or AFM through height different between your deposited layer and substrate. 

but as total substrate is covered by the film ,it is not an easy task. i have tried with parafin film and aluminium foil covering,and then sputtered.still unable to get an easy way.

It is easy to use Energy dispersive Xray analysis to estimate thickness of any multilayers as elemental peaks are easily identified. However, thickness standards are available or may be generated in the lab where micro-balance and UHV evaporation units are available. Contact US National Bureau of Standards in Washington Dc to supply you such Standard Reference Materials!

That is called a Step procedure.It is required when uses Diamond tip profiler.One gets distance thickness this way.Alternately one can get mass thickness using EDXRF method. Knowing density one can get the thickness. 

You can use few possibities:

1) Nobody said about calowear tester

2) You can prepare cross-section and observe coating thickness with SEM or TEM (in depends on your expecting thickness)

3) AFM

P.S. I dont know about your substrate but i advice you to deposite your coating also onto Si-Wafer. This wafer is easy to break and you can with help of SEM observe your thickness

For metal films the best technique is using liftoff and a step profiler, as you suggest. You need a sharp boundary and a calibrated profiler (Dektak, AFM, whatever). To get a sharp boundary I have had consistent success using a high-melting wax (not paraffin!) like Apiezon W. Dissolve in tricholorethane (TCE), paint a narrow stripe or small dot. Deposit your film. Then liftoff the wax with TCE--it comes off very clean but rinse with acetone and isopropanol before the TCE evaporates. This assumes (1) the deposition is at room temp (2) the film thickness is less than about 1 micron. Y can get data good to about 5% or 1 nm. Less than 1% error would require great care in calibration.

If you are depositing at elevated temperature then a much more complex process would be needed. Al foil and other shadowing approaches generally don't give good results because the boundary they create is tapered (due to penumbra effects) not abrupt.

Cross-section SEM works but calibrations are rarely reliable to better than 10% unless you put a lot of effort into calibration. Also it's not trivial to get a cross section that doesn't have misleading artifacts.

If you have very flat films you can use XRD to observe low angle reflections (or sidelobes) that are a precise measure of thickness.

Rutherford backscattering will give you the thickness but not with great accuracy.

A common method to measure thickness is with a quartz microbalance (crystal thickness monitor) but this assumes you know the density accurately--at the 5% error level this is rarely true sputtered films. (Procedure: measure the resonant freq of the crystal before and after sputtering; the difference is proportional to the mass deposited.) 

For oxides (not your case) ellipsometry works extremely well, as already suggested.

Prof. van Dover is mistaken about Rutherford backscattering spectrometry (RBS) when he says RBS "will give you the thickness but not with great accuracy".  On the contrary:  we have just demonstrated that RBS is a primary direct reference technique capable of the highest thickness accuracy available from any non-destructive model-free thin film method (Colaux et al, Analyst, 2015:  see link).

You have had a list of techniques already but the majority of them are integral methods depending on prior knowledge of the sample.  So XRF or SEM-EDS have to assume a simple layer structure,  AFM only measures the linear thickness at a point (without any compositional information),  ellipsometry depends very sensitively on both knowledge of the appropriate optical constants and a detailed model of the layer structure.

RBS on the other hand tells you not only how many atoms/cm2 are present but where they are in the film thickness - a differential analysis,  not an integral one.  And if you want to look for oxidation (highly likely in your case) you can easily slide into non-Rutherford (elastic) backscattering (EBS),  simply by using the 3038 keV resonance in the 16O(a,a)16O reaction (see details in the link).

Of course,  all the other techniques have their own power,  and most of them are widely available and convenient.  If you have them use them!  But if you really want to know what your films are,  then use RBS and the related techniques.  You have several labs in India able to supply this for you.

I should remark that I did mixed metal silicides on silicon by RBS long ago (Barradas et al, APL, 1997,  see link).  I should also mention that by using the ion beam analysis (IBA) methods synergistically we can unambiguously determine mixed metal boro-silicides (where the metal ratios are determined by particle-induced X-ray emission,  PIXE,  and the boron profile is determined by nuclear reaction analysis, NRA:  Barradas et al, 2009).

http://dx.doi.org/10.1039/c4an02316a

http://dx.doi.org/10.1063/1.119524

http://dx.doi.org/10.1016/j.nimb.2010.02.092

Agree with Chris comment. Suggestion re. PIXE is similar to my suggestion about using EDXRF technique as both these measure mass thickness!

What is the expected thickness?

200 nm- 900 nm

First you have to make step and further use AFM or surface profiler tool to measure the thickness of your film.

As it's a bit thick to be measured by optical method, I would suggest to use XRF method. Do calibration against SEM and it should be stable enough by time.

For sure there are other ways but I am suggesting what I have experienced. Just ned to choose the right detector type

Good luck.

sudhir, after making steps also in afm i am not getting the proper thickness.

I would like to suggest the following paper:

http://journals.aps.org/pra/abstract/10.1103/PhysRevA.91.042704

With this method is possible measure the absolute thickness and density (this method combine MEIS and REELS)

Using physical means such as Dektak requires sharp steps, preferably a repeating pattern to handle baseline slope issues. If the sample is not perfectly "horizontal" (it never is), than its base slope will mask the transition of film/substrate IF the step is not sharp. You mentioned nothing about what temperature you are working at, and what type of substrate you are using?. Polished c-Si substrates are flat enough to use Dektak, but I'm not sure if glass substrates are sufficiently flat.

If you are working at ambient temperatures, you could try placing narrow strips of tape on your substrate, then pealing them off after, to reveal the step pattern. Otherwise, you would need to put a suitable mask after deposition, and acid etch your film away (as long as the acid does not also etch your substrate).

While it is certainly true that the electrical resistivity of any metal depends sensitively on the deposition conditions, you could use an ohm meter to get rough idea. Just make your samples into thin strips and measure its resistance from end to end.

You mentioned Si. What is your substrate? Is it some sort of glass? Do you have access to a spectrophotometer? If so, looking at its interference pattern is a simple way. Of course, there are sophisticated ways to process optical data (see my papers in JVST) to do this. Also, in the regions of absorption, you could measure transmittance, and calculate the Absorbance. IF you know the absorption constant (you could always use some published [n,k] data to figure this out, even approximately), then it is easy to convert Absorbance, Abs = log10(To/T), into absorption constant (alpha), which involves the film thickness (d) and Log10(e), i.e., alpha = Abs / (d x log10(e)) = 2.3025 x Abs / d, so d = 2.3025 x Abs / alpha.

Of course, anything you do will depend on the details, the density in particular, where the density of thin films usually differs from bulk handbook values (think about columnar growth, with void structure, etc.). Si is a particularly interesting example, where there are tens of thousands of papers on Si, and a-Si showing all sorts of microstructure. One of the reasons I like measuring the interference pattern optically is that there is plenty of information to also yield refractive index [n], which relates to density. So these optical measurements gives both [n,d], and from [n], you can get density. Yes, the spacing between the interference peaks depend on the product of n x d, BUT, the values of the minimum/maximum depend uniquely on the refractive indices of the film (nf) and substrate (ns). You should be able to figure this out fairly simply using Excel calculations.

what type of substrate you are using?

Silicon and glass substrate, over which I am depositing Si Co simultaneously.

Jyoti

you can do SEM cross-section of the film on both glass and silicon substrate.

i did the same but i got reflection and brightness of the thin film over glass and over silicon irregularities seen, which didnt give me accurate data.

You can measure the thickness either by profilo meter or by Ellipsometry test.

i'm also interested how to measure deposited film thickness during magnetron sputtering. what is the dynamic method during deposition?