How do I delete the spectral images caused due to sampling in frequency domain?

One of my module has a requirement to shift the frequency of a transmitter signal and view its spectrum.

For example, say, my transmitter signal is at center frequency 160MHz and the sample rate is 81.92MHz. I am shifting the signal step by step from 0MHz to 80MHz. What I expect to see on the spectrum is that the signal disappears at some point of time when the shifting frequency is more than the window size i.e. 81.92MHz but I see that the signal appears again. (Which I think is because of the spectral images caused due to sampling). Please find the spectrums attached.

I do not want to see the observed behavior. I designed an algorithm to make the fft bins zero on the right or left depending on the direction of shift but I feel it is a round about solution. Any suggestion would help!

For frequency shift, I used the formula,

SigOut = SigIn.*exp(I*2pi*t*(f_c-f_shift))

f_c = center frequency, f_shift = shift frequency

Answer 1

[Image Analyst]

"How do I delete the spectral images?" Use the delete() function and pass it the filename.

[Bhavanithya Thiraviaraja]

I'm sorry, they are not files. As you can see in the image attached, it is in the spectrum.

[Image Analyst]

If they're not files, but variables, you can use the clear() function to delete variables.