<div dir="ltr">Hi Natalie,<div><br></div><div>if your system is metallic, and you are using smearing, there is not necessarily any spin polarization, even if there is an odd number of electrons in the system. Imagine running a system with 27 Na atoms in your supercell, for instance. Then RKS and UKS give the same results.</div><div><br></div><div>Do you expect any spin polarization in your case?</div><div><br></div><div>Matt<br><br>On Monday, October 17, 2016 at 9:29:35 PM UTC+1, Natalie Austin wrote:<blockquote class="gmail_quote" style="margin: 0;margin-left: 0.8ex;border-left: 1px #ccc solid;padding-left: 1ex;"><div dir="ltr">Hello, <div><br></div><div>I was running a calculation (calculation 1) with an odd number of electrons. The default multiplicity was set to doublet. This calculation ran without the UKS tag. </div><div><br></div><div>I compared this calculation to one in which I included the UKS keyword in the input (calculation 2). The electronic energies of each calc1 and cal 2 were comparable to 10E-7. </div><div><br></div><div>I found it strange that CP2K would run a system with an odd number of electrons as RKS, is this because of the smearing?</div><div><br></div><div>Input:</div><div><div>&FORCE_EVAL</div><div> METHOD QS</div><div> &DFT</div><div>  BASIS_SET_FILE_NAME BASIS_MOLOPT</div><div>  POTENTIAL_FILE_NAME GTH_POTENTIALS</div><div>  &MGRID</div><div>   CUTOFF 500</div><div>   REL_CUTOFF 60</div><div>   NGRIDS 4</div><div>  &END MGRID</div><div>  &QS</div><div>    EPS_DEFAULT 1.0E-14</div><div>    MAP_CONSISTENT</div><div>  &END QS</div><div> &SCF</div><div>   SCF_GUESS ATOMIC</div><div>   EPS_SCF 1.0E-8</div><div>   MAX_SCF 500</div><div>   ADDED_MOS  200</div><div>   CHOLESKY INVERSE</div><div>   &SMEAR  ON</div><div>     METHOD FERMI_DIRAC</div><div>     ELECTRONIC_TEMPERATURE [K] 300</div><div>   &END SMEAR</div><div>   &DIAGONALIZATION</div><div>     ALGORITHM STANDARD</div><div>   &END DIAGONALIZATION</div><div>   &MIXING</div><div>     METHOD BROYDEN_MIXING</div><div>     ALPHA  0.1</div><div>     BETA   1.5</div><div>     NBROYDEN  8</div><div>   &END MIXING</div><div> &END SCF</div><div>  &XC</div><div>   &XC_FUNCTIONAL PBE</div><div>   &END XC_FUNCTIONAL</div><div>  &END XC</div><div>CHARGE -1</div><div>  &POISSON</div><div>   POISSON_SOLVER WAVELET</div><div>   PERIODIC NONE</div><div>  &END POISSON</div><div># UKS .TRUE.     => this is uncommented in the UKS calculation</div><div> &END DFT</div></div><div><br></div><div><br></div><div>Calculation 1:</div><div><div> DFT| Spin restricted Kohn-Sham (RKS) calculation               RKS</div><div> DFT| Multiplicity                               2</div><div> DFT| Number of spin states                           1</div><div> DFT| Charge                                  -1</div></div><div><br></div><div><br></div><div>Calculation 2</div><div><div> DFT| Spin unrestricted (spin-polarized) Kohn-Sham calculation        UKS</div><div> DFT| Multiplicity                               2</div><div> DFT| Number of spin states                           2</div><div> DFT| Charge                                  -1</div></div><div><br></div></div></blockquote></div></div>