Are you asking for a statistical modeling approach to estimating the relationship between stress (measured how?) and doping concentration (of what?)? If so, you need to provide more information about your data, and clarify what exactly you hope to achieve.
for particular length of material like 0-6 cm . initially concentration is high upto 2 cm but as increase the length concentration reduce but not zero.
Nice graphic display of the data! I don't understand what you mean by "i need stress/pressure vs that length /thickness...from that curve?" Are these two additional variables in your model? Or are you simply interested in modeling out the relationship illustrated in the graph? If the latter, then I would suggest you use regression splines, since you clearly see the non-linear relationship, but it looks more like a hockey stick than a polynomial.
I have analyzed your curve. Your graph shows that the nonlinear change of Boron concentration is negatively proportional to the depth into the surface. Both the Boron concentration and the depth are nonlinear numbers. In other words, the nonlinear change of nonlinear numbers Y is negatively proportional to the nonlinear change of nonlinear numbers X.
In summary
The nonlinear change of Boron concentration is negatively proportional to the depth into the surface.
The differential equation for the above phenomenon is d(q(Yu – Y)) = Kd(q(Xu – X))
The integral equation of the above phenomenon is q(Yu – Y) = Kq(Xu – X) + qC
The asymptote of the depth Xu is 1.5 μm.
From Fig. 1C-2, the equation for the first curve is y = 2.5845x-0.302 , i.e., (Yu – Y) = C(Xu – X)-K with K = -0.302, C = 2.5845. We can also write the equation as q(Yu – Y) = -Kq(Xu – X) + qC
The graphical analyses and the tables are given in the attached PDF file. I also attached two more files for explaining the nonlinear concept.