First, your line "freq = 0:binHzConv:100;" is not correct. Frequency does not range between 0 and 100 Hz in your example, but rather -50 to 50. This is the so-called "two sided" spectrum. I have some more details on that here: http://www.mechanicalvibration.com/Making_matlab_s_fft_functio.html
The second thing, is that I suggest you plot both amplitude and phase together. I.e. using the "simpleFFT" script in the above web page:
[frq, amp1, phase1] = simpleFFT( signal, fs);
[frq, amp2, phase2] = simpleFFT( filtered_signal, fs);
figure;
subplot(2,1,1);
plot( frq, amp1, frq, amp2);
xlim([0 10])
subplot(2,1,2);
plot( frq, phase1, frq, phase2);
xlim([0 10])
You will see that your filtered signal only has a large amplitude near 6 Hz. For example, at 2 Hz, it's about 100 times smaller than at 6 Hz. This is of course exactly what you had intended by use of the filter. But the thing to keep in mind that the phase of a signal with zero amplitude is basically meaningless. It's like asking whether zero is a "positive zero" or a "negative zero". The phase at 6 Hz in your filtered signal does match exactly the phase at 6 Hz in your original signal. At other frequencies it does not match, just due to numerical roundoff, but you don't really care because the amplitude is zero and phase is meaningless there anyway.
It may help you to understand if you plot the real() and imag() components of X and Z instead of amplitude and phase.
