From owner-nwchem-users@emsl.pnl.gov Wed Nov 1 13:35:40 2006 Received: from odyssey.emsl.pnl.gov (localhost [127.0.0.1]) by odyssey.emsl.pnl.gov (8.13.8/8.13.8) with ESMTP id kA1LZdWW023565 for ; Wed, 1 Nov 2006 13:35:40 -0800 (PST) Received: (from majordom@localhost) by odyssey.emsl.pnl.gov (8.13.8/8.13.8/Submit) id kA1LZder023564 for nwchem-users-outgoing-0915; Wed, 1 Nov 2006 13:35:39 -0800 (PST) X-Authentication-Warning: odyssey.emsl.pnl.gov: majordom set sender to owner-nwchem-users@emsl.pnl.gov using -f X-IronPort-AV: i="4.09,379,1157353200"; d="scan'208,217"; a="10038343:sNHT83616309" X-MimeOLE: Produced By Microsoft Exchange V6.5 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C6FDFD.AB098D9F" Subject: [NWCHEM] RE: Hybrid functionals and PW in NWChem Date: Wed, 1 Nov 2006 13:35:36 -0800 Message-ID: In-Reply-To: X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: Hybrid functionals and PW in NWChem Thread-Index: AcbnQJTLP6MjC0bmQTeBoWbYI0ql6AAESo5ABAzIhmABndtdQA== From: "Bylaska, Eric J" To: "Deskins, Nathaniel A" Cc: X-OriginalArrivalTime: 01 Nov 2006 21:35:36.0922 (UTC) FILETIME=[AB4E33A0:01C6FDFD] Sender: owner-nwchem-users@emsl.pnl.gov Precedence: bulk This is a multi-part message in MIME format. ------_=_NextPart_001_01C6FDFD.AB098D9F Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Aaron, The answer to your questions are: 1) The "set nwpw:excited_ne 4 0" tells the program to compute the 4 lowest up/restricted virtual orbitals. 2) The pspw code is a gamma point calculation. For accurate calculations of crystals one needs to formally include molecular orbitals that have phases beyond the periodicity of the system that approach infinity. Hence you want to include molecular orbitals that have periodicities larger than the unit cell, while at the same time producing densities that have a periodiciy of the unit cell. One strategy for doing this is to multiply the wavefunctions by a phase factor exp(i*k*r), (the BAND code contained in nwpw allows you to do this). Another, less elegent, strategy is to just make your system size bigger. =20 3) yes Eric Bylaska > _____________________________________________=20 > From: Deskins, Nathaniel A =20 > Sent: Tuesday, October 24, 2006 9:01 AM > To: Bylaska, Eric J > Subject: RE: Hybrid functionals and PW in NWChem >=20 > Eric, > I have a few more questions.=20 >=20 > 1) I couldn't find anything in the manual for the following line. What > exactly does this keyword do?=20 >=20 > set nwpw:excited_ne 4 0 >=20 > 2) Why exactly is a larger cell needed for accurate band-gap > calculations? >=20 > 3) One approach to charged periodic systems is to add an oppositely > charged uniform background charge to compensate for the excess charge > added. Is this the approach used with NWChem? >=20 > Thank you, > Aaron >=20 >=20 >=20 > _____________________________________________=20 > From: Bylaska, Eric J =20 > Sent: Tuesday, October 03, 2006 6:13 PM > To: Deskins, Nathaniel A > Cc: Dupuis, Michel > Subject: RE: Hybrid functionals and PW in NWChem >=20 > Nathan, >=20 > Here's a simple input for a pbe0 calculation. It's fairly > straightforward, however it's best to initialize the calculation with > a pbe96 calculation. Note that for an accurate band-gap calculation > you will need to use a bigger cell then what I've given here. Hope > that helps. >=20 > Eric >=20 >=20 > title "quartz-alpha minimization and energy calculation" >=20 > #permanent_dir /home/bylaska/Projects/Solids/Quartz/sio2-9/perm3 > #scratch_dir /home/bylaska/Projects/Solids/Quartz/sio2-9/perm3 > permanent_dir ./perm > scratch_dir ./perm >=20 > start alpha-quartz >=20 > memory 900 mb >=20 > geometry units angstroms nocenter noautosym noautoz print > system crystal > lat_a 4.716d0 > lat_b 4.716d0 > lat_c 5.216d0 > alpha 90.d0 > beta 90.d0 > gamma 120.d0 > end > Si 0.48078 0.48078 0.00000 > Si 0.51922 0.00000 0.66667 > Si 0.00000 0.51922 0.33333 > O 0.58541 0.73559 0.78317 > O 0.15018 0.41459 0.11650 > O 0.26441 0.84982 0.44983 > O 0.41459 0.15018 0.88350 > O 0.73559 0.58541 0.21683 > O 0.84982 0.26441 0.55017 > end > set nwpw:cif_filename quartz >=20 > nwpw > simulation_cell > ngrid 32 32 32 > end > energy_cutoff 45.0 > ewald_ncut 8 > xc pbe96 > end >=20 > set nwpw:excited_ne 4 0 >=20 > set nwpw:minimizer 2 > task pspw energy >=20 >=20 > nwpw > xc pbe0 > end > task pspw energy > task pspw energy >=20 > _____________________________________________=20 > From: Deskins, Nathaniel A =20 > Sent: Tuesday, October 03, 2006 4:07 PM > To: Bylaska, Eric J > Subject: Hybrid functionals and PW in NWChem >=20 > Eric, > Michel mentioned to me that he thought you had coded some hybrid > functionals into the PW portion of NWChem. If this is correct, I would > be very interested in using them to compare with the DFT+U method (as > mentioned in my previous email). >=20 > Thank you, >=20 > Aaron Deskins > Post-Doctoral Research Associate > Molecular Interactions and Transformations Group > Pacific Northwest National Laboratory > Richland, WA >=20 ------_=_NextPart_001_01C6FDFD.AB098D9F Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable RE: Hybrid functionals and PW in NWChem

Aaron,
        The answer to your questions are:

        1)  The "set nwpw:excited_ne 4 = 0"  tells the program to compute the 4 lowest up/restricted = virtual orbitals.

        2) The pspw code is a gamma point = calculation.  For accurate calculations of crystals one needs to = formally include

         &nbs= p;      molecular orbitals that have phases = beyond the periodicity of the system that approach infinity.  Hence = you want

            to include molecular orbitals = that have periodicities larger than the unit cell, while at the same = time producing

         &nbs= p;      densities that have a periodiciy of the = unit cell.  One strategy for doing this is to multiply the = wavefunctions by

         &nbs= p;      a phase factor exp(i*k*r), (the BAND = code contained in nwpw allows you to do this).  Another, less = elegent, strategy is to      =            just make your = system size bigger. 

        3) yes

        =         =         =         Eric Bylaska

_____________________________________________
From:   Deskins, Nathaniel A 
Sent:   Tuesday, October 24, 2006 9:01 AM
To:     Bylaska, Eric J
Subject:       = RE: Hybrid functionals and PW in = NWChem

Eric,
  I have a few = more questions.

1) I couldn't find = anything in the manual for the following line. What exactly does this = keyword do?

    = set nwpw:excited_ne 4 0

2) Why exactly is a = larger cell needed for accurate band-gap calculations?

3) One approach to = charged periodic systems is to add an oppositely charged uniform = background charge to compensate for the excess charge added. Is this the = approach used with NWChem?

Thank you,
Aaron



_____________________________________________
From:   Bylaska, Eric J 
Sent:   Tuesday, October 03, 2006 6:13 PM
To:     Deskins, Nathaniel A
Cc:     Dupuis, Michel
Subject:       = RE: Hybrid functionals and PW in = NWChem

Nathan,

        Here's a simple input for a pbe0 calculation. = It's fairly straightforward, however it's best to initialize the = calculation with a pbe96 calculation.  Note that for an accurate = band-gap calculation you will need to use a bigger cell then what I've = given here.  Hope that helps.

        =         Eric


title = "quartz-alpha minimization and energy calculation"

#permanent_dir = /home/bylaska/Projects/Solids/Quartz/sio2-9/perm3
#scratch_dir   = /home/bylaska/Projects/Solids/Quartz/sio2-9/perm3
permanent_dir = ./perm
scratch_dir   ./perm

start = alpha-quartz

memory 900 mb

geometry units = angstroms nocenter noautosym noautoz print
  system = crystal
    = lat_a 4.716d0
    = lat_b 4.716d0
    = lat_c 5.216d0
    = alpha 90.d0
    = beta  90.d0
    = gamma 120.d0
  end
Si     0.48078   = 0.48078   0.00000
Si     0.51922   = 0.00000   0.66667
Si     0.00000   = 0.51922   0.33333
O      0.58541   = 0.73559   0.78317
O      0.15018   = 0.41459   0.11650
O      0.26441   = 0.84982   0.44983
O      0.41459   = 0.15018   0.88350
O      0.73559   = 0.58541   0.21683
O      0.84982   = 0.26441   0.55017
end
set = nwpw:cif_filename quartz

nwpw
  = simulation_cell
    = ngrid 32 32 32
  end
  energy_cutoff = 45.0
  ewald_ncut = 8
  xc = pbe96
end

set nwpw:excited_ne 4 = 0

set nwpw:minimizer = 2
task pspw = energy


nwpw
   xc = pbe0
end
task pspw = energy
task pspw = energy

_____________________________________________
From:   Deskins, Nathaniel A 
Sent:   Tuesday, October 03, 2006 4:07 PM
To:     Bylaska, Eric J
Subject:       = Hybrid functionals and PW in = NWChem

Eric,
  Michel mentioned to me that he = thought you had coded some hybrid functionals into the PW portion of = NWChem. If this is correct, I would be very interested in using them to = compare with the DFT+U method (as mentioned in my previous = email).


Thank you,

Aaron Deskins
Post-Doctoral Research = Associate
Molecular Interactions and = Transformations Group
Pacific Northwest National = Laboratory
Richland, WA

------_=_NextPart_001_01C6FDFD.AB098D9F--