This may be the one of the reason, up to the certain concentration of Cu loading there would be mono-layer formation of Cu on TiO2 , hence uniform distribution of Cu on the surface of TiO2 support. Once the surface is saturated with Cu. Above this concentration of Cu, there would be multi-layer formation of Cu on TiO2 support. Hence, the excess Cu would be physically adsorbed. Which might desorb  in subsequent processing steps, leaving less amount of Cu loading than the theoretically expected.  This can be interpreted by XRD and SEM.

Dear K.Y. Nandiwale,

Nice to hear about your explaination. Please refer me, If any supportive paper with this concept.

Analogy could be i have two hands and i can hold two coconuts on my palm one on each. so if i have extra coconuts i cannot hold them as effective surface area to carry them is not there.

Similarly when effective surface are of support is occupied no more loading can occur.

@Kakasaheb and @Mithil 

Question arise that if the catalyst has reached the saturated state then further loading should not add anything on it but the original add should remain constant or not???? it should show plateau at rather than decrease it concentration???? 

What you guys say

When the Cu loading on TiO2 increases certain loading reached to maximum loading. At the loading Cu is well dispersed on the support, When further increase in Cu loading Cu is agglomeration on the TiO2 support . The active sites were blocked  by Cu, these can be clearly seen in XRD and TEM and BET surface area.

How you are estimating the decrease in Cu level of loading on to the support?

`Through EDX Analysis

This result seems indeed unlikely : one would expect a plateau except if, for some reason (e. g., very high or very low pH)  the surface of the support decreases a lot under the conditions used for the wet impregnation at the highest loadings. I believe that you should first question the accuracy of the method used for quantifying copper and try complementary methods.

Initially the at lower loadings the Cu is well dispersed on the support.  Once monolayer is formed then the metal starts to agglomerate and forms bigger crystallites.  In impregnation whatever you load on to support it will sit there, it will not slip off unless you used for any reaction where there is a chance of leaching.  Since bigger crystallites are formed the accessible sites of Cu is only on the surface. The sites inside the surface are not accessible.  EDAX is only a surface technique depending on the size of crytal or surface thickness you try to measure.  Over certain loading EDAX is not able to penetrate enough to the bul the measure the actual amount of Cu present.  Hope this will help.

In my opinion you must go for ICP and XRD analysis to get a robust conclusion.

I agree with Dr Srivasta.  It is a matter of access.  Access may be limited due to the size of the crystallites.  It also depends on the solution chemistry and you actual ion sizes.  Depending on the concentration you may have a complex ion formed which may result in limited access through the pores.