[BUG] `.to_pandas` from booleans with missing values returns `None` instead of `np.nan`

**Steps/Code to reproduce bug**
xref https://github.com/rapidsai/cudf/pull/20899#discussion_r2628899171

```
In [1]: import cudf

In [2]: cudf.Series([True, float("nan"), False], nan_as_null=True).to_pandas()
Out[2]:
0 True
1 None
2 False
dtype: object

In [3]: import pandas as pd

In [4]: pd.Series([True, False]).reindex([0, 1, 2])
Out[4]:
0 True
1 False
2 NaN
dtype: object
```

**Expected behavior**
While pandas still does recognize `None` as a missing value for bool data, `np.nan` is the more canonical missing value.