yes, the secondary phase is CeO2. which is not showing in XRD pattern. But it is visible in SEM

Generally, if the secondary or impurity phase is more than 5% then it is reflected in XRD pattern otherwise small amount can not detected through XRD. In your case by making the pellet you have applied certain pressure and then heated the sample at 1400 C that results in unreacted amount of Ce to incorporate into the host lattice.But I do agree with Roberto that at higher doping concentration of Ce, you may get the seconadary phase in XRD pattern even after pelletizing and sintering.

Dear all,

I have doped CeO2 upto 15 mol %.  But I have not seen any XRD peaks of CeO2 phase.

Dear karthikeyan, I have analysed the XRD data of sintered pellet but all the peaks I found are close to peaks of SrTiO3 (JCPDS)  with a small shifting in peaks with CeO2 addition.  No other peaks were found. So there is no possibility of Ce2O3 formation.

The shifting could show formation of solid solution... I have a similar problem - the XRD shows undoubted solid solutions with the increase of the second component, but at the same time phase differences appear on the phase image of AFM. I started thinking of some sort of stress or tensions on the grains borders...

yes you are right that CeO2 is incorporating in SrTiO3 but my question is that why the excess CeO2 is not showing in the XRD pattern.

Is it possible that these alloys (I assume that they are polycrystalline) possess conductivity anistropy?..

What evidence from SEM do you have for the secondary phase (CeO2) being present? If you have absolute concrete proof from SEM that you absolutely definitely have CeO2 crystallites in your films then there are several reasons why you might not see characteristic reflections in the XRD. 

My 0th order comment when it comes to XRD is plot it on a log[intensity] scale. The number of people who publishe papers with XRD on a linear scale and then claim to see no evidence of secondary phases may well be factually accurate, but then they haven't looked very hard.  If your main peaks are 105-107 cps then any impurity phases with a few hundreds of counts in their reflections will just not be seen on a linear scale but will poke your eye out on a log scale.  So if you are not plotting on a log scale then you should do so.  If you already are, then move on.  

Firstly, if you are using a lab-based x-ray source the limit of detection for a crystallite is a lot higher than for synchrotron radiation, i.e. you need more of the impurity phase to see peaks appearing above the background counts.  

Secondly, if you do have CeO2 crystallites, then could they be more or less randomly oriented, or have multiple crystallographic orientations?  In extremis you might have a lot of the unwanted CeO2 phase, but if you have small crystallites all randomly oriented then each crystallite might not contribute enough scattered intensity to be seen above the background counts.  

Thirdly, if you have a lot of CeO2, and all of it is crystallographically oriented in one preferential direction, then it might well be a case of each crystallite being very small and contributing a very weak and very broad peak to the XRD diffractogram.  

I cannot understand why everybody thinks that there is a problem with the XRD pattern.  The SEM observation also has its disadvantages. As it was pointed by Roberto Koferstein, according to similar phase diagrams there is high probability of solid solution formation (full incorporation of CeO2 into the titanate) up to ~30 % CeO2, which appears on the XRD pattern as shifting of the SrTiO3 lines. No CeO2 lines appear in this case. There may be Ce-rich and Sr-rich regions in the sample, which may appear as different phases on the SEM microphotographs. Other option, which may cause the "color differences" on the SEM, may be anisotropy of the electrical conductivity of the crystallites (different conductivity on the different cristallographic directions). If there is a nano-phase a hump should appear on the XRD, same for a glassy phase. 15 % is not small quantity of second component. Have a nice day!

Dear Rampyar Sing,

Go for slow scan XRD (long scan about 3-4 hrs) with small step size.  If you still do not find CeO2 peaks, then think of the way Temenuga Hristova Vasileva proposed. Of course if your CeO2 vol% too much low that of XRD detection range it may not be possible to detect with XRD, but majority cases slow scan XRD helps to decide what to do?.

all the best

XRD can not show additional phases when such phase contents are either less than 5% or amorphous in nature after sintering. On the other hand, SEM gives a morphological image which can clearly figure out secondary phases and grain boundary segregations even for minor constituents.