More necessary information about the sample is needed to answer the question. If the film is thick enough, then x-rays do not penetrate the film and the substrate peaks do not show up. So, in the two samples which show peaks of the substrate, either the film became thinner (if the temperature is high enough for evaporation) or the film is destroyed in some regions of the sample, exposing the substrate to the beam. This possibility can be checked by examining the film morphology carefully. Also, the materials (if more than one) of the film could have reacted at higher temperatures to form another material which is more transparent to x-rays (although this possibility is not great in light of absence of information!). Any way more information is required to reach a convincing answer, and the answers I provided are only possibilities.    

First of all what is ur instrument settings: thin film configuration or Bragg Brentano?

Did u take the pattern of ur substrate? (attention substrate orientation). What is ur film method?

About "peak corresponding to substrate", cud u please provide the XRD pattern. Maybe can't be seen because it still amorphous or fine crystallite (SRO).

I suggest u try with powder sample (if possible). Anneal the sample as what u did and compare the pattern.

More necessary information about the sample is needed to answer the question. If the film is thick enough, then x-rays do not penetrate the film and the substrate peaks do not show up. So, in the two samples which show peaks of the substrate, either the film became thinner (if the temperature is high enough for evaporation) or the film is destroyed in some regions of the sample, exposing the substrate to the beam. This possibility can be checked by examining the film morphology carefully. Also, the materials (if more than one) of the film could have reacted at higher temperatures to form another material which is more transparent to x-rays (although this possibility is not great in light of absence of information!). Any way more information is required to reach a convincing answer, and the answers I provided are only possibilities.    

I agree withe Sami. It might be due to phase change during the temperature. But fr that you have to look that whether at that point any phase change can occur or not. Other thing is from you explanation it seems that coating is multi layer so  might be some layer has been destroyed or changed to some other composition due to some diffusional process between the layers. Some elements may be diffused to some upper layer and change the composition and making the new phase. So, need to look for all these possibilities. 

I agree with Sami Mahmoud. This can be attributed to the film thickness.Further information’s on the samples (composition and thickness) are needed to explain why the substrate peaks as well as the chromium peaks are not observed  in the XRD patterns .

The Bragg Brentano geometry in general is useful also to collect thin film pattern. It seems that at lower temperature the film is more stable and thick and able to cover the diffraction of the substrate. At high temperature the film could be more thick or can be completely lost/decomposed by the effect of high temperature. Are you sure that your sample is still present? Have you performed only XRD analysis on your film? What kind of sample you deposited? Inorganic or organic?

Dear Murali Krishnan: I forgot to tell you that in a diffraction experiment with no spectroscopy involved, we should not use the term "XRD spectrum" to refer to XRD data, but rather, the term "XRD pattern"

The strange behavior of your MgO-2.0.0 peak can be caused by the miscut angle of your MgO crystals. In single crystalline substrates, cut from a larger crystal, the sample surface is usually not exactly parallel to the aimed for crystal plane. This miscut angle can easily be a few degrees. When you perfectly align the sample surface on your diffractometer, the crystal will therefore be miss aligned. When the normal to the crystal plane is in the diffraction plane, the peak will be invisible if the miscut angle is larger then the divergence angle. If the normal to the crystal plane is outside the diffraction plane your substrate peak will be invisible when the miscut angle is larger than the axial divergence. As the axial divergence is usually much larger than the in-plane divergence it is possible to either see or not see the substrate peak depending on how your sample is put in the diffractometer.

You can find the miscut angle of your substrate by making a pole figure of the MgO-2.0.0 peak, or make a series of rocking curves of that peak while rotating your sample in between measurements.

You can use this effect to your benefit by misaligning your substrate crystal plane on purpose, using an omega offset. Thereby making the normally huge substrate peak invisible and increasing the visibility of your thin film crystal structures.

I strongly confirm Prof. Sami Mahmood 

My sincere thanks to all of you especially for Sami and Harry for their clear and useful explanation. Thank you so much for correcting me from terming it as XRD spectrum.

The absence of MgO in some cases can be due to the miscut angle. However, I can not find any reason for the absence of Cr peak. Please find the XRD data for the annealed sample attached. The structure of my sample is:

 Cr (10 nm)/ [Sm(3.2 nm)/Co (7.6 nm)] x 25 /Cr (20 nm)/MgO(100)

All the samples are of same structure but annealed at different temperatures given in the figure.

Thanks

The evolution of the peak profiles indicate progressive crystallization in the film. The peaks in the sample annealed at 500 C are broad, indicating poor crystallization of textured Sm2Co7 phase along the [110] crystallographic direction (following your assignment for the peaks). The crystallite size cal be evaluated using Stokes and Wilson method, and this should give crystallite size below 10 nm in the sample annealed at 500 C. From the breadth of the peak (for the (110) and (220) reflections) I could conclude that the crystallite size increased by probably more than double value for the sample annealed at  700 C. Also, increasing the annealing temperature seems to disrupt the texture where the (200) reflection develops. Also, improvement of the crystallization of the SmCo3 phase is evident from the sharpening of the corresponding peaks with increasing annealing temperature. These observations could indicate the occurrence of serious changes in the morphology of the film. The structural peak (110) corresponding to bcc Cr should be located in the angular range 44 - 45 degrees. Evidence of such peak is available at low annealing temperatures. The way I see the provided data is that obvious solid state reactions took place at higher temperatures, leading to better crystallization of some phases and changes in the film morphology and crystal orientations. The Cr could be incorporated in the new evolving phases, and the pattern of the sample annealed at 650 C should be investigated in more depth, especially concerning the appearance of a structural peak just above 30 degrees.

In fact the samples crystallize by increasing the temperature. Are you sure that the peak at about 43° that you see at high temperature pertains to MgO? The  SmCo5 phase presents a main peak at about 43°, overlapped with that of MgO. In addition at 650°C many other unexplained peaks that request attention are present.

What is your film thickness? If it is thin (>500-1000mircons) Bragg-Brentano XRD might not be suitable for your film. X-ray will penetrate through the film and reach the substrate resulting in substrate peaks. Grazing incidence XRD on the other hand will be effective for the surface planes. In that case you can miss some planes aswell. You need to set the angle carefully.

And regarding the substrate peaks on heating, the reason is the micro cracks produced at elevated temperatures. At higher temperature the substrate material (can) diffuse into the film. You need to go through the thermochemistry of the film and substrate material for that.

I don't know what you want but if you want to sort it out without effort you can go for mirco-raman or XPS.