<html>
  <head>
    <meta content="text/html; charset=ISO-8859-1"
      http-equiv="Content-Type">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <div class="moz-cite-prefix"><br>
      What about the evolution of the constrained magnetization?<br>
      (Please also make sure that the values specified in
      starting_magnetization make sense, as suggested by L. Paulatto
      Sir).<br>
      <br>
      My suggestion was to vary lambda in small steps (say 0.5). I'm
      surprised that you managed to converge the calculation with such
      high lambda values.<br>
      <br>
      Anyway, the constrain energy looks way too large, your system is
      probably still far from the target.<br>
      <br>
      <br>
      GS<br>
      <br>
      On 03/04/2014 01:57 PM, paresh rout wrote:<br>
    </div>
    <blockquote
cite="mid:CANRJWo8DqdwiauyLYUJC51uHMTJV8F3E_KDYSqSHaMVzUzsctA@mail.gmail.com"
      type="cite">
      <meta http-equiv="Content-Type" content="text/html;
        charset=ISO-8859-1">
      <div dir="ltr">
        <div>
          <div>Respected Sclauzero   sir,<br>
          </div>
            Thanks for your reply. According to your suggestion, I
          varied my Lambda value from 0,5,.......150 ry. Although 
          calculated constrained energy are decreasing but upto 150 ry
          the constrained energy and the estimated scf accuracy are not
          the same order. Here I am providing my constrained energy with
          various lambda value.<br>
          <br>
        </div>
        Lambda      Constraint_Energy<br>
        0           0.00000000<br>
        5           36.93411685<br>
        10          69.54815816                 <br>
        15          6.65653915   estimated scf accuracy    <         
        7.6E-13 Ry<br>
        20          7.88546052<br>
        25          8.88385707<br>
        30          9.71513061        <br>
        35          10.42697250<br>
        40          11.05006563<br>
        45          11.60072229<br>
        50          12.08887057<br>
        70          13.54966033<br>
        80          14.05546257<br>
        90          14.45159513<br>
        100         14.75974550<br>
        110         14.99680383<br>
        120         15.17624003<br>
        130         15.30876396 <br>
        140         15.40310437         estimated scf accuracy   
        <          9.9E-13 Ry                        <br>
        150         15.46632278          estimated scf accuracy   
        <          9.9E-13 Ry<br>
        <br>
        <div>
          <div>
            <div>
              <div class="">
                <div class="">
                  <table class="" cellpadding="0">
                    <tbody>
                      <tr class="">
                        <td class="">
                          <table class="" cellpadding="0">
                            <tbody>
                              <tr>
                                <td>
                                  <div class="">
                                    <div class="">
                                      <table class="" cellpadding="0">
                                        <tbody>
                                          <tr class="">
                                            <td class="">
                                              <table class=""
                                                cellpadding="0">
                                                <tbody>
                                                  <tr>
                                                    <td><br>
                                                    </td>
                                                  </tr>
                                                </tbody>
                                              </table>
                                            </td>
                                          </tr>
                                        </tbody>
                                      </table>
                                    </div>
                                  </div>
                                </td>
                              </tr>
                            </tbody>
                          </table>
                        </td>
                      </tr>
                    </tbody>
                  </table>
                </div>
              </div>
            </div>
          </div>
        </div>
      </div>
      <div class="gmail_extra"><br>
        <br>
        <div class="gmail_quote">On Tue, Mar 4, 2014 at 3:06 PM,
          Sclauzero Gabriele <span dir="ltr"><<a
              moz-do-not-send="true"
              href="mailto:gabriele.sclauzero@mat.ethz.ch"
              target="_blank">gabriele.sclauzero@mat.ethz.ch</a>></span>
          wrote:<br>
          <blockquote class="gmail_quote" style="margin:0 0 0
            .8ex;border-left:1px #ccc solid;padding-left:1ex">Dear
            Paresh,<br>
            <br>
                 in my understanding you should start with a very small
            lambda value (e.g. 0.1), make sure the calculation has
            converged (not always trivial), then restart with a larger
            value.<br>
            It is important to tune the steps by which you increase
            lambda. Increasing it by steps of 5 seems too much to me, I
            would suggest you to try much smaller steps, say between 0.1
            and 0.5.<br>
            <br>
            There are two reasons why the energy increases: the first is
            because you are constraining your system out of its ground
            state, but that's exactly what one would expect. The other
            is the contribution from the penalty energy (E_constrain, it
            should be printed after each scf step), which is used to
            impose the constraint.<br>
            An important thing is that this energy term is not physical
            and becomes negligible once your system reaches the target
            state.Therefore one should monitor this constraint energy,
            together with the constrained quantity, and make sure it
            goes to zero at some point.<br>
            <br>
            Once lambda is large enough and you reached the targeted
            state, E_constrain should be negligible w.r.t. the total
            energy and of the same order of the estimated scf accuracy.
            From that point on, the energy should not change if you
            further increase lambda, because your system fulfills
            (almost) exactly the constraint, so that E_constrain should
            stay to a very low value.<br>
            <br>
            HTH<br>
            <br>
            <br>
            GS<br>
            <div class=""><br>
              Respected Lorenzo sir,<br>
                              Thanks for your replay but my question was
              how to fix the proper LAMBDA value for any constrained
              system as the energy is increasing with increase in lambda
              value.Please suggest me something regarding lambda.<br>
              <br>
              <br>
              <br>
            </div>
            <div>
              <div class="h5">On Mon, Mar 3, 2014 at 5:54 PM, Lorenzo
                Paulatto <<a moz-do-not-send="true"
                  href="mailto:lorenzo.paulatto@impmc.upmc.fr">lorenzo.paulatto@impmc.upmc.fr</a><mailto:<a
                  moz-do-not-send="true"
                  href="mailto:lorenzo.paulatto@impmc.upmc.fr">lorenzo.paulatto@impmc.upmc.fr</a>>>
                wrote:<br>
                On 03/03/2014 12:39 PM, paresh rout wrote:<br>
                Dear all,<br>
                         I am doing  spin polarized  calculations on a
                multiferroic compound . In some cases I am doing
                constrained magnetic calculations to get the Low-spin
                and High-spin state . For this I am using<br>
                    constrained_magnetization='atomic'<br>
                    starting_magnetization(1) = 5.0<br>
                    starting_magnetization(2) = -3.0<br>
                    starting_magnetization(3) = 0.0<br>
                    starting_magnetization(4) = 0.0<br>
                         lambda =0,5,10,20,25,........etc<br>
                <br>
                <br>
                <br>
                 
                 +--------------------------------------------------------------------<br>
                   Variable:       starting_magnetization(i), i=1,ntyp<br>
                <br>
                   Type:           REAL<br>
                   Description:    starting spin polarization on atomic
                type 'i' in a spin<br>
                                   polarized calculation. Values range
                between -1 (all spins<br>
                                   down for the valence electrons of
                atom type 'i') to 1<br>
                                   (all spins up). Breaks the symmetry
                and provides a starting<br>
                                   point for self-consistency. The
                default value is zero, BUT a<br>
                                   value MUST be specified for AT LEAST
                one atomic type in spin<br>
                                   polarized calculations, unless you
                constrain the magnetization<br>
                                   (see "tot_magnetization" and
                "constrained_magnetization").<br>
                                   Note that if you start from zero
                initial magnetization, you<br>
                                   will invariably end up in a
                nonmagnetic (zero magnetization)<br>
                                   state. If you want to start from an
                antiferromagnetic state,<br>
                                   you may need to define two different
                atomic species<br>
                                   corresponding to sublattices of the
                same atomic type.<br>
                                   starting_magnetization is ignored if
                you are performing a<br>
                                   non-scf calculation, if you are
                restarting from a previous<br>
                                   run, or restarting from an
                interrupted run.<br>
                                   If you fix the magnetization with
                "tot_magnetization",<br>
                                   you should not specify
                starting_magnetization.<br>
                 
                 +--------------------------------------------------------------------<br>
                <br>
                <br>
                kind regards<br>
                <br>
                <br>
                --<br>
                Dr. Lorenzo Paulatto<br>
                IdR @ IMPMC -- CNRS & Université Paris 6<br>
                phone:+33 (0)1 44275 084 / skype: paulatz<br>
                www:  <a moz-do-not-send="true"
                  href="http://www-int.impmc.upmc.fr/%7Epaulatto/"
                  target="_blank">http://www-int.impmc.upmc.fr/~paulatto/</a><br>
                mail: 23-24/4é16 Boîte courrier 115, 4 place Jussieu
                75252 Paris Cédex 5<br>
              </div>
            </div>
          </blockquote>
        </div>
      </div>
    </blockquote>
    <br>
  </body>
</html>