Hi Anagha,
There is no real reason why the two ends of the plot should look the same. Say your elements are numbered 0 to 23 from left to right, and theta = +- 90 degrees.
In one case you are observing from the left end and there is increasing time-delay phase shift from element 0 to 23. In the other case you are observing from the right end and there is decreasing time-delay phase shift from element 0 to 23. To this you have to add the increasing 10-degree-inspired phase shifts of the elements themselves, which are the same in both cases. It's not a symmetric situation.
It so happens that for d = (1/2)*lambda the amplitude is the same at both ends, but if you change to, say, d = .53 lambda you get the plot below.
I believe it's correct to use 20*log10 of abs(amplitude) which gives twice as many dB as in your plot.
