How to fill an array column with zeros while not changing the initial array?

I'm trying to obtain 2 separate arrays from a 2 column array of 44100 values. One of the arrays should have 0s for the right column and the other values kept intact, and the other vice-versa for the left column. So far, I have the following code:

def tone(f, sf, duration):  # frequency, sampling frequency and duration

   sample_times = np.arange(0,duration,1/sf)
   max_amplitude = 2**15 - 1
   cos_wave = np.cos(2 * np.pi * f * sample_times) * max_amplitude
   return cos_wave.astype("int16") # returns an array of dtype =int16

sf = 44100
f = random.randrange (20, 18000)
duration = 1

cos_wave = tone(f, sf, duration)

print(cos_wave.dtype)

 # Generate stereo sound by putting it in 2 columns
both_channels = cos_wave.reshape((int(sf/2), 2))
both_channels = both_channels.astype ('int16')
print (both_channels)
`# Fill one column with zeros to get mono sound for each channel`
copy_both_channels = np.copy (both_channels)
print (copy_both_channels)
both_channels[:, 0] =  0 # 1st column filled with zeros

The problem is that I'm modifying the initial array (both_channels) when filling one of the columns with 0s. I tried using the np.copy(arr) function but that didn't help either as it doesn't fill the columns with 0s. Can anyone tell me of a way to achieve this?