Potential biases arising in the use of cascade impactors to estimate respiratory tract deposition patterns of lead-acid battery plant aerosols