<div dir="ltr"><div>Dear all, I apologize for sending a former e-mail from another account other than the subscribed one.<br><br></div>So, here is my post again:<br><br>I’m having trouble to replicate the result for the raman spectra for some hybrid perovskite,
namely CH3NH3PbI3, a.k.a MAPbI3 (4 MAPbI3 molecules/unit cell). I'm using the relaxed cell and atomic coordinates as in <cite>J. Phys. Chem. Lett.</cite>, <span>2014</span>, <span>5</span> (2), pp 279–284.<br><br>So, I’d
really thank you any advice and words of correction regarding to the
following issues:
<p>1-The thing is that my phonon calculations never converge, even using
an energy cutoff in the scf previous calculations of 1D-12. (Using
Norm-conserving LDA-level PZ PPs). WIll be necessary to lower the tr2_ph
value in ph.x to something greater than 1D-12 since the system is
somewhat huge (24 atoms)?</p><p>2- In a latter ph.x run attemp, I set
the tr2_ph
value to 1D-8 and the the initial convergence was achieved, but at the
stage of calculation the second derivatives, roots never converged, so i
can not get the dynamical matrix anytime. <br></p>
<p>So, for as a huge system as MAPbI3 is, It is likely to exist special
constraints, adjustments or requirements i'm missing, in order to get
the dynamical matrix and, using dynmat.x and other plotting tools, to
simulate the raman spectrum with the quantum espresso suite?</p>
<p>Excuse me if i’m missing several important theoretical or QE-related
aspects, but i really need some advice about all of this stuff.</p>
<p>Best regards, </p><div>-------------------------------------------------------<br><b>Dr.Sc. JOSUÉ CLAVIJO, Chem.</b><br><b style="font-size:12.8px">Christian, born again</b></div><div><b><br></b>Assistant Professor</div>Universidad Nacional de Colombia<br>Science College<br>Chemistry Department<br><br><div><div><br><br></div></div></div>