In a large enough population we can assume that selections are independent. Then we can use binominal distribution. In this case, the probability of getting a defective part is 0.10. If we sat X is the number of defective parts in a sample of two then X can have three values: 0, 1, 2.
Then the function is: