The ABINIT Users Mailing List ( CLOSED )

["Zhenhua Zeng" <myid520@163.com>]

Dear abinit users, 
    Now, I want to predict using oxygen. So I test some well-known situations, such as bond length and bond energy (atomisation energy) of oxygen. The error of bond length is less than 1%. But the error of bond energy is very large. The calculated result using every pseudo potential downloaded from the web cite is between 8.5ev and 9.7ev. But The experimental data is 5.12ev. The result calculated using VASP and CASTEP are between 5.5ev and 6.5ev. The input file is given at below which based on Tutorial\t21.in. In the input file, the acell and ecut have been tested for the convergence of calculation. I want to know why it differs so much from experimental data and other calculations Does there have something wrong in my input file, the method used in tutorial to calculate bond energy (atomisation energy) can’t be used here, Or there has something wrong else? In addition, Can I use oxygen for prediction only depends on the good tested result of bond length? Your help will be greatly appreciated!
Best regards.
Zhenhua Zeng
Hunan University, PR. china

####  input files   ##############
#calculate oxygen molecular and atom
#common parameters
ndtset 2
ixc 1 or ixc 3 or ixc 7 or ixc 11 for different pseudo potential
acell 3*12  Bohr 
ecut  30.0  Hatree
nstep 100
ngkpt 2 2 2
kptopt 1
ntypat 1
znucl  8
diemac 1.0   
diemix 0.5   

#the parameters for oxygen atoms
nsppol1 2             ##SPin POLarization calculation of oxygen atom
spinat1 0.0 0.0 1.0
natom1 1           
typat1 1
xcart1 3*0.0        
toldfe1 1.0d-6

#The parameters for oxygen molecular
natom2 2   
typat2 1 1        
xcart2   -1.145 0.0 0.0    #the bond length of oxygen molecular is 
          1.145 0.0 0.0    #about 2.29 Bohr
ionmov2  3  
ntime2  100  
tolmxf2  5.0d-5 
toldff2  5.0d-6

・1.5G 全国最大空间――网易126免费邮箱 http://www.126.com 火热抢注中>>>

  体验30兆超大附件发送，享受100％卡巴斯基专业杀毒……更多超G感受等您来体验！