计算化学公社

标题: 关于优化晶面 [打印本页]

作者
Author: wufangfang 时间: 2016-6-6 15:00
标题: 关于优化晶面
我建立了一个8x8x1的氧化锆（110）晶面，切了3层，固定2层，结果优化失败了，这是因为我建立的晶面太大的吗？

Publications of specific relevance to this calculation:

Density functional:
P91 exchange: Perdew, Physica B 172, 1 (1991)

The Generation and Use of Delocalized Internal Coordinates in Geometry optimization;
Baker Kessi Delley: J. Chem. Phys., 105, 192 (1996)
Andzelm Fitzgerald King-Smith: Chem. Phys. Lett., 335, 321 (2001)

Fast Calculation of Electrostatics in Crystals and Large Molecules;
Delley: J. Phys. Chem. 100, 6107 (1996)

Warning: ecp          takes precedence over using  symmetrized basis
Calculation is Spin_restricted

Lattice:
translation vector [a0] 1 77.722167850 0.000000000 0.000000000
translation vector [a0] 2 0.000000000 54.957960030 0.000000000
translation vector [a0] 3 0.000000000 0.000000000 44.319746809
Cell volume    189309.662 a0^3
Warning: at least C1 SYM file should be there for optgeom
Oxygen    nbas=  1, z= 8, nrfn=  7, rcut=  8.69, e_ref= -0.056141 Ha
                              rcore=  0.00  zval= 8.00 8.00
n=1  L=0  occ= 2.00 e=    -18.912531Ha    -514.6364eV
n=2  L=0  occ= 2.00 e=    -0.879876Ha    -23.9426eV
n=2  L=1  occ= 4.00 e=    -0.334433Ha       -9.1004eV
n=2  L=0  occ= 0.00 e=    -2.144815Ha    -58.3634eV
n=2  L=1  occ= 0.00 e=    -1.593916Ha    -43.3727eV
n=3  L=2  occ= 0.00 e=    -2.722222Ha    -74.0755eV
n=3  L=2  occ= 0.00 e=    -1.388838Ha    -37.7922eV  eliminated
Zirconium nbas=  2, z=  40, nrfn=  7, rcut=  8.69, e_ref= -0.039312 Ha
  Zr ECP ECP28MWB : 28 3 32    rcore=  3.58  zval=  12.00  40.00
n=4  L=0  occ= 2.00 e=    -2.007769Ha    -54.6342eV
n=4  L=1  occ= 6.00 e=    -1.178718Ha    -32.0745eV
n=4  L=2  occ= 2.00 e=    -0.104892Ha       -2.8543eV
n=5  L=0  occ= 2.00 e=    -0.139354Ha       -3.7920eV
n=4  L=2  occ= 0.00 e=    -0.722500Ha    -19.6602eV
n=5  L=0  occ= 0.00 e=    -0.640828Ha    -17.4378eV
n=5  L=1  occ= 0.00 e=    -0.462659Ha    -12.5896eV

Symmetry orbitals C1
n  norb representation
1 12032       a
  total number of valence orbitals:  12032

cell charge=             0.000000 active electron number 7168.0
(without charge=          7168.0 )

Integration in 3d (fixed mesh): ipa= 6  f3= F
npri,iomax,iomin, thres,  rmaxp, sp
0 6 1 0.00010 10.00 1.00
Integration points, checksum  0 448040 3584256 7168.086841

matrix diagonalization has max memory requirements  4969.8 MB    277.4 273.5  4419.0

real array elements, matrices vectors etc: 3864.3 MB
integer arrays                         :    69.2 MB
min recommended for all-incl workspace : 3889.6 MB
Total memory allocated for arrays       : 4969.9 MB
Total disk space needed for TPVEC    1.10 GB

Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
HP-MPI licensed for Accelrys.
Host 0 -- ip 10.1.255.253 -- ranks 0 - 7

host | 0
======|======
0 : SHM

Prot -  All Intra-node communication is: SHM

Checked out license feature: MS_dmol_MP <v2010.090> [for msi] (1 copy)

Checked out license feature: MS_dsolid_MP <v2010.090> [for msi] (1 copy)

MPI Application rank 0 killed before MPI_Finalize() with signal 9

DMol3.pl message: DMol3 job finished in 0 hr 44 min 35 sec.

作者
Author: liu_tiao 时间: 2016-6-6 16:52
总共多少原子啊。真空层多大。

作者
Author: wufangfang 时间: 2016-6-6 19:37
180左右个原子，真空层18A，k点1x11,您说会不会k点太小了，一般有什么经验或规律之类的吗？

作者
Author: liu_tiao 时间: 2016-6-6 19:48

wufangfang 发表于 2016-6-6 19:37
180左右个原子，真空层18A，k点1x11,您说会不会k点太小了，一般有什么经验或规律之类的吗？

超胞的大小是多少呢。你这个体系太大，先用很低精度试试吧

作者
Author: wufangfang 时间: 2016-6-6 19:48
我建立了一个8x8x1的氧化锆（110）晶面，切了3层，固定2层，结果优化失败了，这是因为什么原因？

Publications of specific relevance to this calculation:

Density functional:
P91 exchange: Perdew, Physica B 172, 1 (1991)

The Generation and Use of Delocalized Internal Coordinates in Geometry optimization;
Baker Kessi Delley: J. Chem. Phys., 105, 192 (1996)
Andzelm Fitzgerald King-Smith: Chem. Phys. Lett., 335, 321 (2001)

Fast Calculation of Electrostatics in Crystals and Large Molecules;
Delley: J. Phys. Chem. 100, 6107 (1996)

Warning: ecp          takes precedence over using  symmetrized basis
Calculation is Spin_restricted

Lattice:
translation vector [a0] 1 77.722167850 0.000000000 0.000000000
translation vector [a0] 2 0.000000000 54.957960030 0.000000000
translation vector [a0] 3 0.000000000 0.000000000 44.319746809
Cell volume    189309.662 a0^3
Warning: at least C1 SYM file should be there for optgeom
Oxygen    nbas=  1, z= 8, nrfn=  7, rcut=  8.69, e_ref= -0.056141 Ha
                              rcore=  0.00  zval= 8.00 8.00
n=1  L=0  occ= 2.00 e=    -18.912531Ha    -514.6364eV
n=2  L=0  occ= 2.00 e=    -0.879876Ha    -23.9426eV
n=2  L=1  occ= 4.00 e=    -0.334433Ha       -9.1004eV
n=2  L=0  occ= 0.00 e=    -2.144815Ha    -58.3634eV
n=2  L=1  occ= 0.00 e=    -1.593916Ha    -43.3727eV
n=3  L=2  occ= 0.00 e=    -2.722222Ha    -74.0755eV
n=3  L=2  occ= 0.00 e=    -1.388838Ha    -37.7922eV  eliminated
Zirconium nbas=  2, z=  40, nrfn=  7, rcut=  8.69, e_ref= -0.039312 Ha
  Zr ECP ECP28MWB : 28 3 32    rcore=  3.58  zval=  12.00  40.00
n=4  L=0  occ= 2.00 e=    -2.007769Ha    -54.6342eV
n=4  L=1  occ= 6.00 e=    -1.178718Ha    -32.0745eV
n=4  L=2  occ= 2.00 e=    -0.104892Ha       -2.8543eV
n=5  L=0  occ= 2.00 e=    -0.139354Ha       -3.7920eV
n=4  L=2  occ= 0.00 e=    -0.722500Ha    -19.6602eV
n=5  L=0  occ= 0.00 e=    -0.640828Ha    -17.4378eV
n=5  L=1  occ= 0.00 e=    -0.462659Ha    -12.5896eV

Symmetry orbitals C1
n  norb representation
1 12032       a
  total number of valence orbitals:  12032

cell charge=             0.000000 active electron number 7168.0
(without charge=          7168.0 )

Integration in 3d (fixed mesh): ipa= 6  f3= F
npri,iomax,iomin, thres,  rmaxp, sp
0 6 1 0.00010 10.00 1.00
Integration points, checksum  0 448040 3584256 7168.086841

matrix diagonalization has max memory requirements  4969.8 MB    277.4 273.5  4419.0

real array elements, matrices vectors etc: 3864.3 MB
integer arrays                         :    69.2 MB
min recommended for all-incl workspace : 3889.6 MB
Total memory allocated for arrays       : 4969.9 MB
Total disk space needed for TPVEC    1.10 GB

Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
Signal Caught:          15
HP-MPI licensed for Accelrys.
Host 0 -- ip 10.1.255.253 -- ranks 0 - 7

host | 0
======|======
0 : SHM

Prot -  All Intra-node communication is: SHM

Checked out license feature: MS_dmol_MP <v2010.090> [for msi] (1 copy)

Checked out license feature: MS_dsolid_MP <v2010.090> [for msi] (1 copy)

MPI Application rank 0 killed before MPI_Finalize() with signal 9

DMol3.pl message: DMol3 job finished in 0 hr 44 min 35 sec.

作者
Author: wufangfang 时间: 2016-6-6 19:50
我选择了精度是Medium

作者
Author: liu_tiao 时间: 2016-6-6 20:18

wufangfang 发表于 2016-6-6 19:50
我选择了精度是Medium

真空层能否也调小一些。

作者
Author: wufangfang 时间: 2016-6-6 20:32
因为我还要在晶面上放吸附分子，如果真空层太小的话，放不进去。

作者
Author: liu_tiao 时间: 2016-6-6 20:44

wufangfang 发表于 2016-6-6 20:32
因为我还要在晶面上放吸附分子，如果真空层太小的话，放不进去。

什么分子。带苯环的吗

作者
Author: zyj19831206 时间: 2016-6-6 21:10

wufangfang 发表于 2016-6-6 19:48
我建立了一个8x8x1的氧化锆（110）晶面，切了3层，固定2层，结果优化失败了，这是因为什么原因？[/backcolo ...

加没加smearing?有没有应用slab correction?有没有降低density mixing和提高DIIS size?

作者
Author: wufangfang 时间: 2016-7-1 09:00
不好意思，最近忙晕了，可以打开了，结果我们计算平台故障了

欢迎光临计算化学公社 (http://bbs.keinsci.com/)