There are many unknowns (substrate, cement, coating materials).  Please check the following:

• Make sure that the design is really considering the prism material as incident and ambient media, and make sure that you know well the refractive index of the prism and cement;
• Did you apply the coating on one prism only, and then cement it with another prism?  or did you make the coating on both prism, and then cement?
• Maybe the cement material gets into the coating:  you could try (if possible) to end the coating with a thick layer of material (SiO2 maybe) that will match the prism and cement (away from the main part of the coating)
• You could try to "optical contact" the prisms (i.e. use very flat prism surfaces before coating, and just contact them together after (if they are flat enough, they will stay together without using any cement.

These are just some ideas.  But really, I would first double-check the design to make sure that there are no errors there.

If your designed is perfect then check optical properties of the cement.

Dear sir,

i have done coating on one one hexagonal prism and cemented it with another.

before cementing each other the optical performance of the hexagonal prism is showing the good characteristics which exactly matches the design spectra.

but once the right angle hexagonal prisms cemented with each other the peak shift is happening in a shift of 150-200 nm.

cementing material we are using the company Norland optics adhesive  UV cement  part number: P/N 6116 ( also attached the adhesive cements link )

the substrate material we are using is SK2

2) we are not doing coating on both prisms, on one prism coating is done and cemented with other.

Dear Vemuri,

I checked the data about the substrate and the cement:  I have no explanation for the short wavelength shift that you see.

A part from the shift, do you see a large deformation of the spectrum?  Is it something that you could use to characterize the coating (curve fitting playing with some parameters of the design)?    Could it be related to the measurement?  How are you measuring the properties?  Does the light entering the device is monochromatic, or white light?  Do you always get the same shift from the same device when you are repeating the measurement?

Thank you sir,

The shift in peak wavelength is happening all ways when cementing the prisms with each other.

regarding the light source: we are using the broadband halogen source range 400 - 800 nm, while measuring it will give monochromatic source.

we tried with different designs also in the prisms, but the results is same shift in wavelength is there, also amplitude of the peak is decreases and FWHM  is increasing sir.  

 sir please send any material regrading the Dichroic beam splitter designs.

my email id is: [email protected] 

Dear Vemuri,

I still don't know why you see that shift, but I suspect it comes from the measurement.

You wrote:  "regarding the light source: we are using the broadband halogen source range 400 - 800nm, while measuring it will give monochromatic source".

Does it mean that a beam of white light is incident on the device, and after it passes through it, the transmitted beam is sent to a spectrometer in order to get the transmitted spectrum?

The reason I am asking for precision is that if the device, for some reason, affect the alignment of the beam that is sent to the spectrometer, it might shift (somehow) the image that is coming from the gratings, and thus shift the whole spectrum.

One way to check if the shift is coming from the measurement would be to introduce between the prisms some material with an absorption that does not depend on the angle of incidence (band-gap, dye,...)  For example, if you have a colored plastic sheet, or some colored ink that you could sandwich between two glass substrates, you could (i) measured the transmittance at normal incidence between two flat slides of glass, and (ii) repeat the measurement with the material between the prisms, in the same configuration that you used for your previous measurements.  Because the absorption in the material does not depend on the angle of incidence, you should not see a shift between the measurements done with the slides and with the prisms.  If you observe a shift, then it confirms that there is something wrong with the measurements.

Since you confirmed that the design is okay, I don't know what else could cause the shift.

sir we have measured the hexagonal beam splitter with  with two spectrometers, they gave same result. after cementing the both prisms each other. shift is observed. 

sir, similar experiment we have also done in the plane substrate, we have cemented each other but the shift is not observed sir.

also,amplitude of the cemented plane substrate and with out cemented one, no change observed.

but the experiment when we performed in the hexagonal beamsplitter.

cemented prisms----we observed shift of 150-200 nm, and also amplitude also decreased by ~half    

with out cemented prism--- we observed same performance as seen in plane substrate.

please give e-mail id sir, i will attach some documents for you sir.

Vemuri (sorry for the delay, busy with other stuff...)

My email is: [email protected].  You can send me more detailed figures about the measurement setup, that would help me with our troubleshooting.  thanks

Based on your last reply, you really narrowed the problem to the cement; "with out cemented prism" means that you measured the performance with the prisms in contact with each other (attached somehow), but without the cement to glue them.  right?

dear sir, (sorry for the delay sir, due to holiday vacation i am unable to send you immediate reply sir)

I mention in the last reply, there are two separate substrates. one is plane 70X70 mm substrate and other is hexagonal prism.

Without Cementing Measurement

It means we measured both the plane substrate and hexagonal prism individually (without contact to any 2nd plane substrate or prism) and achieved correct answer (as per the modeled answer).

With Cementing Measurement

It means we cemented both i.e. the plane substrate with another uncoated plane substrate and prism with another prism. then we measured the results We observed varying results in case of prism while the plane substrates after cementing also gave the same results. 

Regarding  this query i am sending a mail from ( [email protected]) for you sir. 

Dear sir,

i Have send you email sir,  please check the details sir

Vemuri, I just sent an email to you;  I think I found the reason for your observed shift.

The reason lies in the design and the incident medium that was considered for the beam incident at 45° on the coating.  You have to do the design with the prism material as incident and exit media if you want to get the right quarterwave thicknesses for the coating materials; if you mistakenly take AIR as the incident medium, you will get thickness values that are too small and the resulting peak will be around 450nm instead of 650nm.  I did some calculations that confirm this hypothesis.

I hope that it is clear and that it will help you find a solution to your problem

Best Regards!!

Thank you sir. for clarifying the problem in beam splitter design.