In very simple terms, you can see one transformer block as one layer of a (vanilla) neural network.

One attention head represents one neuron, one transformer block with multiple attention head represents one layer with multiple neurons.

If you see it this way, this question is same as: Why do we need a neural network with 5 hidden layers each having 10 neurons while we can create a neural network with a single hidden layer with 50 neurons ?

The early layers see lower level patterns while the deeper layers see more abstract and high-level patterns and hence stacking of layer is needed.