In Breuer's book, he deduces quantum master equation using following steps:
$(1). \frac{d}{dt}\rho(t)=-i[H_{I},\rho(t)]$
then the solution for equ.(1) can be written as
$(2).{\rho(t)}=\rho(0)-i\int_{0}^{t}ds[H_{I},\rho(t)]$
By plugging equ(2) to equ(1), we will have
$(3).\frac{d}{dt}\rho_{s}(t)=-\int_{0}^{t}ds\ \ tr_{B}[H_{I}(t),[H_{I}(s),\rho(s)]]$
where $H_{I},\rho$ are interaction hamiltonian and density matrix for the whole system in interaction picture. $tr_{B}$ is partial trace over reservoir.
In the book, the authors say to obtain equ(3),
$(4).tr_{B}[H_{I},\rho(0)]=0$ is assumed.
So my question is why do we assume equ.(4), is there any physics behind it? Like we take Bohn approximation, the coupling between between the system and the reservoir is small?