Home

User's Guide

1. 2 EDIFF SYMPREC 1e 3 SYMPREC 1e 3 IBRION ISIF ISIF ISIF NSW NSW NSW ISMEAR 1 ISMEAR 1 ISMEAR 1 SIGMA 0 2 SIGMA 0 2 SIGMA 0 2 POTIM 0 050 POTIM 0 100 POTIM 0 05 PSTRESS 0 01 PSTRESS 0 01 PSTRESS 0 01 EDIFFG 4e 2 EDIFFG 4e 1 EDIFFG le 3 e POTCAR should be provided by the user The order of element of POTCAR must be the same with the setting tag of NameOfAtoms f submit sh is the job submission file for VASP calculations Here is an example of submit sh bin sh mpdboot mpiexec n 12 bin vasp 4 6 gt vasp log 2 gt dev null Once all the input files are ready you can simply type calypso x gt caly log E to execute the CALYPSO code in the sequential mode Or you can write this command into the pbs script and submit it CALYPSO will generate the results folder which contains all the output files To analyze the CALYPSO results just type following command cd results cak py cif 81 The index of the structures sorted by enthalpies in the ascending order is presented in Analysis_Output dat file and these newly predicted structure files with cif format are shown in dir_0 1 directory Please see section 2 6 3 for further information about the analysis of CALYPSO results 82 4 Special Topic 4 1 The Parallel mode The parallel mode of CALYPSO enables the geometrical optimization of structures through desirable number of CPUs The parallel
2. HAHAHAHA AA ARA AA AAA AA RARA END 2D Parameters A b We strongly suggest that you use the multi stage strategy for structure relaxation For VASP the number of INCAR_ INCAR_1 INCAR 2 is in accord with the setting tag of NumberOfLocalOptim In this example we use three INCAR files for structure relaxation INCAR_1 and INCAR_2 are used to perform very coarse structure relaxation keeping the volume fixed INCAR_3 is used to perform full structure relaxation with accurate precision 31 INCAR_1 SYSTEM B3N2 PREC LOW EDIFF IBRION ISIF INCAR_2 SYSTEM B3N2 PREC LOW EDIFF IBRION ISIF NCAR_3 SYSTEM PREC ENCUT EDIFF IBRION B3N2 Accurate 400 2e 4 2 NSW ISMEAR SIGMA 0 2 POTIM 0 50 LWAVE FALSE LCHARG FALSE PSTRESS 0 001 EDIFFG 4e 2 NSW ISIF ISMEAR NSW 6 SIGMA ISMEAR POTIM 0 200 SIGMA 0 2 LWAVE FALSE POTIM 0 10 LCHARG FALSE LWAVE FALSE PSTRESS 0 001 LCHARG FALSE EDIFFG le 2 PSTRESS 0 001 EDIFFG le 4 3 0 c POTCAR should be provided ATTENTION The order of elements in POTCAR must be identical to the element order in the setting tag of NameOfAtoms d submit sh is the job submission file for performing the VASP calculations Here is an example of submit sh bin sh mpdboot mpiexec n 12 bin vasp 4 6 gt vasp log 2 gt dev null Once all input files are ready you can just type calypso x gt caly l
3. SIESTA the number of sinput_ sinput_1 sinput_2 is in accord with the setting tag of NumberOfLocalOptim In this example we use three SIESTA input files sinput_1 sinput_2 sinput_3 and sinput_4 for structure relaxation sinput_1 and sinput_2 are used to perform very coarse structure relaxation while sinput_3 is used to perform full structure relaxation with accurate precision 47 sinput_1 SystemName siesta SystemLabel siesta NumberOfSpecies 2 NumberOfAtoms 10 block ChemicalSpeciesLabel 1 6 C 2 1H Sendblock ChemicalSpeciesLabel PAO BasisSize SZ kgrid_cutoff 8 0 Ang MeshCutoff 80 Ry PAO EnergyShift 0 02 Ry XC functional GGA XC authors PBE MaxSCFIterations 100 DM MixingWeight 0 150 DM Tolerance 1 d 4 DM NumberPulay 5 Relaxation smearing etc ElectronicTemperature 3000 K MD TypeOfRun cg Optim Broyden true MD VariableCell true MD ConstantVolume false MD MaxForceTol 1d 3 ev Ang MD NumCGsteps 5 Use Save CG true Use Save XV true MD Broyden Initial Inverse Jacobian 0 20 MD RemoveIntramolecularPressure true MD TargetPressure 100 0 GPa ZM ForceTolLen 0 04 eV Ang ZM ForceTolAng 0 0001 eV deg ZM MaxDisplLen 0 1 Ang ZM MaxDisplAng 20 0 deg AtomicCoordinatesFormat NotScaledCartesianAng Sinclude Zmatrix data block MM Potentials 1 1 C6 16 292 0 5 22 C6 0 735 0 5 1 2 C6 3 185 0 5 Sendblock MM Potentials sinput_2 SystemName siesta SystemLabel siesta NumberOfSpecies 2 Numbe
4. true Use Save XV true MD RemoveIntramolecularPressure true MD TargetPressure 100 000010 GPa ZM ForceTolLen 0 04 eV Ang ZM ForceTolAng 0 0001 eV deg ZM MaxDisplLen 0 1 Ang ZM MaxDisplAng 20 0 deg AtomicCoordinatesFormat NotScaledCartesianAng Sinclude Zmatrix data block MM Potentials 1 1 C6 16 292 1 392 2 2 C6 0 735 2 025 1 2 C6 3 185 1 649 Sendblock MM Potentials e The psedopotential psf C psf and H psf in this case should be provided by the user f submit sh is the job submission file for performing the SIESTA calculation Here is an example of submit sh bin sh siesta lt siesta fdf gt siesta out Note the SIESTA output file should be named as siesta out Once all the input files are ready you can just type calypso x gt caly log E to execute the CALYPSO code in the sequential mode Or you can write this command into the pbs script and submit it CALYPSO will generate the results folder which contains all the output files of CALYPSO To analyze the CALYPSO results just type following command cd results cak py cif The index of the structures sorted by enthalpies in the ascending order is presented in Analysis Output dat file and these newly predicted structure files with cif format are shown in dir_0 1 directory Please see section 2 6 3 for further information about the analysis of CALYPSO results 51 3 5 Variable Stoichiometry Structure Prediction T
5. 50 LWAVE FALSE LCHARG FALSE ISTART PSTRESS 0 001 EDIFFG 4e 2 INCAR_3 SYSTEM local optimization PREC Normal EDIFF 2e 4 IBRION 1 ISIF NSW 4 ISMEAR 0 SIGMA 0 05 POTIM 0 1 LWAVE FALSE LCHARG FALSE PSTRESS 0 001 EDIFFG 2e 3 3 0 INCAR_META SYSTEM META PREC Accurate ENCUT 400 EDIFF le 6 IBRION 1 ISMEAR 5 SIGMA 0 05 METAGGA MBJ LASPH TRUE LMINTAU TURE NELM 80 NEDOS 800 INCAR_2 SYSTEM local optimization PREC Normal EDIFF 2e 2 IBRION 2 ISIF NSW ISMEAR 0 SIGMA 0 05 POTIM 0 2 LWAVE FALSE LCHARG FALSE PSTRESS 0 001 EDIFFG 4e 1 INCAR_4 SYSTEM local optimization ENCUT 400 EDIFF le 5 IBRION 2 ISIF NSW 8 ISMEAR 0 SIGMA 0 05 POTIM 0 1 LWAVE FALSE LCHARG FALSE PSTRESS 0 001 EDIFFG 2e 3 3 0 76 d POTCAR should be provided by the user The order of element in POTCAR must be the same with the setting tag of NameOfAtoms Meta GGA calculations require POTCAR files that include information on the kinetic energy density of the core electrons To check whether a particular POTCAR contains this information type grep kinetic POTCAR This should yield at least the following lines for each element on the file kinetic energy density mkinetic energy density pseudized and for PAW datasets with partial core corrections kinetic energy density partial e submit s
6. INCAR files should be therefore provided The same multiple optimization procedure applies also to the use of other optimization codes More details can be found in examples Default value 4 PsoRatio real This variable defines what percentage of the structures per generation should be produced by PSO The rest of structures will then be randomly generated with symmetry constraints Default value 0 6 PopSize integer The population size 1 e the total number of structures per generation Normally a larger population size is needed for a larger system Default value 30 Kgrid reall real2 The precision of the K point sampling for local geometric optimization for VASP and SIESTA codes The Brillouin zone sampling uses a grid of spacing 27xKgrid 12 AT Reall controls the precision of the first two or three local optimizations and the real2 with denser K points controls the last optimization The smaller value gives finer optimization results Note that this setting of Kgrid is only applicable to the uses of VASP SIESTA and Quantum Esspresso codes For uses of other ab initio codes eg CASTEP CP2K and Gaussian etc please prepare appropriate Kgrid settings in their input files Default value 0 12 0 06 Command string The command to submit structure optimization jobs in your computer Default value submit sh MaxStep integer The maximum number of generations to be executed for the entire structure prediction simulati
7. LOW EDIFF 3e 2 IBRION 2 ISIF 2 NSW 40 ISMEAR 0 SIGMA 0 05 POTIM 0 50 LWAVE FALSE LCHARG FALSE ISTART 0 PSTRESS 3000 EDIFFG 4e 2 INCAR_3 SYSTEM local optimization PREC Normal EDIFF 2e 4 IBRION 1 ISIF NSW ISMEAR 0 SIGMA 0 05 POTIM 0 1 LWAVE FALSE LCHARG FALSE PSTRESS 3000 EDIFFG 2e 3 3 0 INCAR_2 SYSTEM local optimization PREC Normal EDIFF 2e 2 IBRION 2 ISIF NSW ISMEAR 0 SIGMA 0 05 POTIM 0 2 LWAVE FALSE LCHARG FALSE PSTRESS 3000 EDIFFG 4e 1 INCAR_4 SYSTEM local optimization ENCUT 600 EDIFF le 5 IBRION 2 ISIF NSW 8 ISMEAR 0 SIGMA 0 05 POTIM 0 1 LWAVE FALSE LCHARG FALSE PSTRESS 3000 EDIFFG 2e 3 3 0 d POTCAR should be provided ATTENTION The order of elements in POTCAR must be identical to the element order in the setting tag of NameOfAtoms e submit sh is the job submission file for the VASP calculations Here is an example of submit sh bin sh mpdboot mpiexec n 12 bin vasp 4 6 gt vasp log 2 gt dev null 28 Once all input files are ready you can simply type calypso x gt caly log amp to execute the CALYPSO in the sequential mode Or you can write this command into the pbs script and submit it in a queue system CALYPSO run will generate the results folder in the current directory you have specified The resul
8. cluster 3 To set up passwordless connection from local machine to remote cluster you need to copy the public key from you local machine directory ssh to the remote cluster Here is the list of commands you need to execute On the local machine user local ssh keygen user local scp ssh id_rsa pub test cluster calypso cn ssh local pub On the remote cluster test cluser cd ssh test cluser cat local pub gt gt authorized_keys test cluser rm local pub 4 Prepare the pbs script vasp pbs to calculate the vasp Here is an example cat vasp pbs bin bash PBS L nodes 1 ppn 6 PBS j oe PBS V cd PBS_O_WORKDIR 84 if f mpd conf then bin echo secretword dfadfs gt gt mpd conf bin chmod 600 mpd conf fi Intel MPI Home MPI_HOME 0pt intel impi 4 0 0 027 setup Nums of Processor NP cat PBS_NODEFILE lwc l echo Numbers of Processors NP Number of MPD N_MPD cat PBS_NODEFILE luniqlwc l echo started mpd Number AN MDDI setup mpi env em64t MPI_HOME bin64 mpdboot r ssh n N_MPD f PBS_NODEFILE running program for i in123 do cp INCAR_ 1 INCAR python writekp py 0 1 mpiexec n NP share apps vasp vasp 4 6 gt log 2 gt dev null cp CONTCAR POSCAR done cp INCAR_4 INCAR python writekp py 0 07 mpiexec n NP share apps vasp vasp 4 6 gt log 2 gt dev null Note the blue digits can be modified to control the precis
9. each atomic species 10 separated by space Taking MgB and MgSiO3 as examples one will write NameOfAtoms Mg B and NameOfAtoms Mg Si O respectively Default value There is no default You must define it NumberOfAtoms integerl integer2 integer3 Number of atoms for each chemical species per formula unit Taking MgB and MgSiO as examples one writes NumberOfAtoms 1 2 and NumberOfAtoms 1 1 3 respectively Default value There is no default NumberOfFormula integerl integer2 Defining the number of formula units in the simulation cell integer and integer2 are the minimal and maximal number of formula units used in the simulation cell respectively If integer equals to integer2 then only single choice of formula unit will be adopted For example NumberOfFormula 4 4 indicates one single run of 4 formula units in your simulation while NumberOfFormula 1 4 indicates that four separate structure prediction runs for 1 2 3 and 4 formula units respectively will be performed Default value 1 4 Volume real The volume in unit of angstrom 3 per formula unit If you cannot provide a good estimation on the volume please use the default value The program will automatically generate an estimated volume by using the ionic radii of given atoms Default value 0 DistanceOflon realy real realiz realy realy real End Minimal interatomic distances in unit of angst
10. for geometry optimization 4 INCAR and POTCAR The VASP input files Now you are ready to run the program calypso x gt caly log After a successful run of the program several VASP output files and a new directory named as results will be generated in your working directory 2 6 Basic inputs and outputs 2 6 1 Input dat The main input file named as input dat contains all necessary parameters for the structure prediction The file consists of input tags that can be given in any order or be omitted while the default values are used Below we offer a quick view of the syntax of the tags The general syntax is tag labels valuel value2 value3 For matrix input it starts with tag and ends with End Values that are not specified in the input dat file are assigned as default values Input values should be separated by space The labels are case sensitive All text following the character is taken as comment e Logical values can be given as T or True or F or False Below are brief descriptions on necessary input parameters SystemName string A string for description of the targeted system max 40 characters Defualt value CALYPSO NumberOfSpecies integer Number of different atomic or chemical species in the system For example it is 2 for MgB Default value There is no default You must provide the number NameOfAtoms stringl string2 string3 Elemental symbols of
11. in the search space based on efficient algorithms over the particle s position and velocity We quote from website of http www swarmintelligence org PSO has been successfully applied in many research and application areas It is demonstrated that PSO can get better results in a faster cheaper way compared with other methods Another reason that PSO is attractive is that there are few parameters to adjust One version with slight variations works well in a wide variety of applications PSO has been used for 5 approaches that can be used across a wide range of applications as well as for specific applications focused on a specific requirement 1 3 History of PSO on Structure Prediction Although PSO algorithm has been employed to various optimization problems the application of PSO in structure prediction started only recently It was attempted for isolated systems small clusters and molecules by Call Zubarev amp Boldyrev in 2007 However this effort did not lead to any practical application The CALYPSO team independently initialized the idea of applying PSO algorithm into structure prediction in 2006 Ma and Wang before Call et als work and made the first application of PSO algorithm into structure prediction of extended systems e g 3D crystals by Wang Lv Zhu amp Ma in 2010 2D layers by Luo et al in 2011 and Wang et al in 2012 2D surface reconstruction by Lu et al in 2014 Structure searching efficiencie
12. nocnnncnn no nn cra crac rn nccn nano 45 3 5 Variable Stoichiometry Structure Prediction conc cccnnnnnnno 52 3 6 Surface Structure PredictiOM ooococonncconnnononononaconeccnrncnononononononanonoccnnnnno noc non none EEN 56 3 6 1 Diamond 111 surface reconstruction prediction sese eee 56 3 6 2 Hydrogenated diamond 100 surface reconstruction ee eee eee 58 3 7 Design of Superhard Matenals sese eee 64 3 8 Design of 2D Material With Adsorption sss eee eee 68 3 9 Design of Optical Materials with Desirable Electronic Band OGoap sees 74 3 10 Crystal Structure Prediction with Fixed Cell Parameters or Atomic Positions G Sp cial MOP iii id 4 1 The Parallel mode iii as 4 2 Remote SUDMISSI cananea 4 3 The split e a is 5 Selected PuBliCations ias 6 Acknowledeements eee ee eee eee 1 Introduction CALYPSO is an efficient structure prediction method and its same name computer software The approach requires only chemical compositions for given compounds to predict stable or metastable structures at certain external conditions e g pressure The method can also be used to design multi functional materials e g superhard materials optical materials etc The CALYPSO package is protected by the Copyright Protection Center of China with the Registration No 2010SR028200 and Classification No 61000 7500 1 1 Meaning of CALYPSO CALYPSO is a short name of Crystal structure AnaLYsis by Particle Swarm Optimization It was
13. prediction This is a must process since it enables the reduction of noise of energy surfaces and the generation of physically justified structures CALYPSO routinely provides O Crystal structure prediction section 3 1 e 2 dimensional layer structure prediction section 3 2 e Clusters or nanoparticles structure prediction section 3 3 e Molecular crystal structure prediction section 3 4 Surface reconstruction structure prediction section 3 6 Structural design of superhard materials section 3 7 O Structural design of 2D material with adsorption section 3 8 Structural design of optical materials section 3 9 For more details on the methodologies and formalisms of CALYPSO please read the references cited below References CALYPSO Software Yanchao Wang Jian Lv Li Zhu and Yanming Ma CALYPSO A Method for Crystal Structure Prediction Comput Phys Commun 183 2063 2012 Crystal Structure Prediction Yanchao Wang Jian Lv Li Zhu and Yanming Ma Crystal structure prediction via particle swarm optimization Phys Rev B 82 094116 2010 Cluster Structure Prediction Jian Lv Yanchao Wang Li Zhu and Yanming Ma Particle Swarm Structure Prediction on Clusters J Chem Phys 137 084104 2012 Two Dimensional Layer Structure Prediction 1 Xinyu Luo Jihui Yang Hanyu Liu Xiaojun Wu Yanchao Wang Yanming Ma Su Huai Wei Xingao Gong and Hongjun Xiang Predicting Two Dimensional Boron Carbon Comp
14. structure relaxation 53 keeping the volume fixed INCAR_3 is used to perform full structure relaxation with medium precision while INCAR_4 is used to perform very accurate calculation INCAR_1 SYSTEM local optimization PREC LOW EDIFF 3e 2 IBRION 2 ISIF 2 NSW 40 ISMEAR 0 SIGMA 0 05 POTIM 0 50 LWAVE FALSE LCHARG FALSE ISTART 0 PSTRESS 3000 EDIFFG 4e 2 INCAR_3 SYSTEM local optimization PREC Normal EDIFF 2e 4 IBRION 1 ISIF NSW ISMEAR 0 SIGMA 0 05 POTIM 0 1 LWAVE FALSE LCHARG FALSE PSTRESS 3000 EDIFFG le 3 3 0 INCAR_2 SYSTEM local optimization PREC Normal EDIFF 2e 2 IBRION 2 ISIF NSW ISMEAR 0 SIGMA 0 05 POTIM 0 2 LWAVE FALSE LCHARG FALSE PSTRESS 3000 EDIFFG le 2 INCAR_4 SYSTEM local optimization ENCUT 600 EDIFF le 5 IBRION 2 ISIF NSW 8 ISMEAR 0 SIGMA 0 05 POTIM 0 1 LWAVE FALSE LCHARG FALSE PSTRESS 3000 EDIFFG le 4 3 0 d POTCAR should be provided by the user The order of element of POTCAR must be the same with the setting tag of NameOfAtoms e submit sh is the submission job file for performing the VASP calculations Here is an example of submit sh 54 bin sh mpdboot mpiexec n 12 bin vasp 4 6 gt vasp log 2 gt dev null Once all input files are ready you can just type calypso x gt caly log E to execute the CA
15. will generate the results folder in current directory which contains all most the outputs files of CALYPSO To analyze the results just type following command cd results cak py cif The index of the structures sorted by energies in the ascending order is presented in Analysis Output dat file and these newly predicted structure files with cif format are shown in dir_orign directory Please see section 2 6 3 for further information about the analysis of CALYPSO results 44 3 4 Molecular Structure Prediction This section is to show the example for the structure prediction with fixed rigid molecules using CALYPSO code Here SIESTA code was used for geometry optimization and enthalpy calculations The CALYPSO input file of input dat rigid molecular structure information file of MOL SIESTA input files of sinput_ and pseudopotential file of psf are needed The following files should be presented in the working directory Files Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO containing controllable key parameters sinput_ Input files for SIESTA MOL Rigid molecular structure information with Z_ Matrix format psf Pseudopotential for SIESTA submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre b The following parameters in input dat are shown for the CH system 45 Volume 60
16. 0 DistanceOfIon 1 0 1 0 1 0 1 0 End PsoRatio 0 6 Command sh submit sh DistOfMol c MOL file contains the rigid molecular structure information with Z_ Matrix format The Z_matrix is a way to represent a system built of atoms with internal coordinates It provides a description of each atom in a molecule in terms of its atomic type bond length bond angle 46 and dihedral angle The detailed descriptions about Z Matrix can be found at http en wikipedia org wiki Z matrix chemistry Here we take CH4 as an example D 0 0 1 1 1 0 0 0 0011 109 48 0 0 001 109 48 120 00 0 109 48 120 00 0 The first line gives the total number of the atoms in the rigid molecular structure information the following lines give the internal coordinates of atoms Nspecie IJK RAT ifr ifa ift Nspecie integer indicates the species number of the atom I J K integer atomic order in molecule is used to define the internal coordinates for current atomic coordinates R real bond length is made by current atom with respect to I atom A real bond angle is made by current atom with respect to I and J atoms T real dihedral angle is made by current atom with respect to I J and K atoms Ifr ifa ift integer flags that indicate whether r a and t should be varied during local structural relaxation 0 for fixed and 1 for varying d We strongly suggest that you use the multi stage strategy for structure relaxation For
17. 50000000 00000000 50000000 00000000 50000000 25000000 21796252 21796252 21796252 21796252 21796252 21796252 21796252 21796252 48632006 0 0 0 0 0 0 0 0 0 Les UDE Ba e e E e Be Ben Les B Ba a Be e e e Be Haase e e Be SY Haase se e Be le e belle e e e e e e We Besides CIF format is also supported However one needs to add _selective tag to specify which atoms are allowed to relax Please see the following example file 60 data_Surf _amcsd formula title H8C48 _symmetry_Int_Tables number Cell length a _cell length b _cell length e _cell angle alpha _cell angle beta _cell angle gamma Loop 5 052205 5 052205 19 824233 90 000000 90 000000 90 000000 _space group symop operation XYZ X Y Z loop _atom_site label _atom_ site fract_x _atom_ site fract_y _atom_site fract_z selective 0 00000000 0 00000000 0 50000000 0 50000000 0 25000000 0 25000000 0 75000000 0 75000000 0 25000000 0 75000000 0 25000000 Aa FO E e a A A 0 25000000 75000000 25000000 75000000 25000000 75000000 25000000 75000000 50000000 00000000 00000000 ke B Ba Ba e e e e e a Be 74778343 74778343 74778343 74778343 70273190 70273190 70273190 70273190 65768037 65768037 65768037 kr le WC d DFTB input files hsd formated input files are needed to relax structures dftb_in hsd_PRE_1 dftb_in hsd
18. ALYPSO PI PP User s Guide CALYPSO version 4 0 Aug 30 2014 Description The CALYPSO user s guide describes how to run and use various features of the structural prediction program CALYPSO This guide shows the capabilities of the program how to use these capabilities the necessary input files and formats and how to run the program both on uniprocessor machines and in parallel http www calypso cn Copyright O 2010 2014 CALYPSO Developers Group All Rights Reserved CALYPSO License Structural Prediction software CALYPSO is available free of charge for non commercial use by individuals academic or research institutions upon completion and submission of the online registration form available from the CALYPSO web site http www calypso cn getting calypso Commercial use of the CALYPSO software requires A CALYPSO COMMERCIAL LICENSE Commercial use includes 1 integration of all or part of the Software into a product for sale lease or license by or on behalf of Licensee to third parties or 2 distribution of the Software to third parties that need it to commercialize product sold or licensed by or on behalf of Licensee These requests can be directed to Prof Yanming Ma by email mym calypso cn Registration Requirements In completing the online registration form individuals may register in their own name or with their institutional or corporate affiliations Registration information must includ
19. G SCF xqc MaxCycle 565 conver 5 OPT Cartesian MaxCycle 1500 MaxStep 29 1l00se GEOM checkpoint NoSymm dr38 Geometry Optimization 0 1 there is an empty line d submit sh is the job submission file for the Gaussian calculation Here is an example of submit sh bin sh g09 lt gsinput gt gsoutput Once all the input files are ready you can just type calypso x gt caly log amp to execute the CALYPSO in the sequential mode Or you can write this command into the pbs script and submit it CALYPSO will generate the results folder which contains all most the output files To 36 analyze the results just type following command cd results cak py cif The index of the structures sorted by energies in the ascending order is presented in Analysis_Output dat file and these newly predicted structure files with cif format are shown in dir_origin directory Please see section 2 6 3 for further information about the analysis of CALYPSO results 37 3 3 2 Tutorial for Ti clusters In this tutorial a CALYPSO structure search for Ti4 clusters is presented The local structural optimizations are performed by VASP so you have to install the VASP code in your machine All the reference files can be found in Path to CALYPSO Example Ti4_VASP The following files should be presented in the working directory Files Description calypso x The executable file for ru
20. LYPSO code in the sequential mode Or you can write this command into the pbs script and submit it CALYPSO will generate the results folder which contains all the output files of CALYPSO To analyze the results just type following command cd results cak py cif The index of the structures sorted by enthalpies in the ascending order is presented in dir_Li H Analysis_Output dat file and these newly predicted structure files with cif format are shown in dir_0 1 directory 55 3 6 Surface Structure Prediction This section introduces the surface structure prediction module It includes a description of the required input files a typical run of the program and result analysis The most important input files are input dat control parameters SUBSTRATE surf the structure files of substrate both cif and VASP POSCAR format are supported and input files for VASP DFTB Most of the output files reside in results directory Note that this part of CALYPSO package is currently in the early stage of development there might be large changes on the control parameters and output files in the following versions For surface structure prediction DFTB and VASP package are supported for local optimization by now This section will give a general introduction to the files used to run the program in the first place then a relatively detailed description of each files followed by two example runs of the program via VASP and DFTB packa
21. Material With Adsorption is performed Adsorption T It determines which method of generating structures should be adopted AdsorptionStyle 69 The information of adatom The specific number Name of atomic element Numbers of adtoms Atomic style l atom 2 molecule or functional group Type of bond The 1st style is needed InformationOfAdatoms H 4 1 end The number of supercell SuperCell 2 2 The range of lenth of bond RangeOfZAxis 1 4 1 2 If true the prediction is performed on bothsides of 2D materials BothSide T The numbers of adatoms on bothsides AdatomsOnBothsides 2 2 end Minimal distance between adatoms and substrate Unit is in angstrom DistanceOfAdatom 1 0 1 0 1 0 1 5 end It determines which method of generating structures with symmetry should be adopted AdsorptionSymmetry 0 HAHAAARARAARR End of adsorption prediction HAHAHA ARA AR ARA ARA c SUB dat contains the 2D substrate structure information It is written in the format of Cartesian coordinate Here we take graphene as an example 70 C 1 0 2 4594000 0 0000000 0 0000000 1 2297000 2 1298750 0 0000000 0 0000000 0 0000000 15 0000000 2 Direct 0 66667 v 0 50000 0 33333 i 0 50000 points 2 0 66667 0 50000 0 33333 0 50000 The structure information is given in the format of POSCAR Note that the first line gives the element s name of substrate If 2 is chosen for AdsorptionStyl
22. N 2 ISIF NSW ISMEAR 0 SIGMA 0 05 POTIM 0 2 LWAVE FALSE LCHARG FALSE PSTRESS 0 001 EDIFFG le 2 INCAR_4 SYSTEM local optimization ENCUT 600 EDIFF le 5 IBRION 2 ISIF NSW 8 ISMEAR 0 SIGMA 0 05 POTIM 0 1 LWAVE FALSE LCHARG FALSE PSTRESS 0 001 EDIFFG le 4 3 0 e submit sh is the submission job file for performing the VASP calculations Here is an example of submit sh bin sh mpdboot mpiexec n 12 bin vasp 4 6 gt vasp log 2 gt dev null 66 Once all the input files are ready you can just type calypso x gt caly log amp to execute the CALYPSO code in the sequential mode Or you can write this command into the pbs script and submit it CALYPSO will generate the results folder which contains all outputs files of CALYPSO To analyze the results just type following command cd results cak py cif hard The index of the structures sorted by hardness in the descending order is presented in Analysis_Output dat file and these newly predicted structure files with cif format are shown in dir_0 1 directory Please see section 2 6 3 for further information about the analysis of CALYPSO results 67 3 8 Design of 2D Material With Adsorption This section is to show the example for the structure prediction of graphane Here DFTB code was used for geometry optimization and energy calculations The CALYPSO input file of inp
23. PEAT HEE EEE End PERE EEE EEE TEEPE EE EEE HH HHH The Parameters for structure prediction with rigid Molecules 4 AAT Examples are given in section 3 2 MAAT MOL ogical when this tag is set as True structure prediction with fixed rigid molecules will be performed This is a useful technique for structure design of molecular systems especially when rigid molecules are known constituent of certain structures Note that a file named as MOL is needed to define the rigid molecule Please refer to examples section 3 4 for settings of MOL Default value False NumberOfTypeMolecule integer The number of different types of molecules in the simulation cell 15 Default value There is no default You must supply this variable NumberOfMolecule integer integer2 The number of molecules for different molecular species Default value There is no default You must supply this variable DistOfMok real The minimum distance in unit of angstrom between two rigid molecules Default value 1 5 JUDO DO DO DO DO UU OOO End IU DO DO DO O UD HHHHHHHHHHHHHHHH The parameters for special constraints H HHHHHHHHH TAH Examples are given in section 3 10 AA SpeSpaceGroup integerl integer2 Defining the space groups ranging from integerl to integer2 for generation of structures If integerl equals to integer2 structure generations will be confined to this specified space group This option is particularly useful when one trie
24. _PRE_2 Control parameters line between will be filled automatically via the program and users shouldn t change these values Besides users should provide Slater Koster SK files for DFTB calculation Some SK files can be found in this link http www dftb org parameters 61 Geometry GenFormat Geometry GenFormat lt lt lt dftb gen lt lt lt dftb gen Driver ConjugateGradient Driver ConjugateGradient EHHH EHEH MovedAtoms MovedAtoms EHHH EHHH MaxForceComponent 1E 3 MaxForceComponent 1E 4 MaxSteps 1000 MaxSteps 1000 OutputPrefix geom out OutputPrefix geom out Hamiltonian DFTB Hamiltonian DFTB SCC Yes SCC Yes MaxSCCIterations 500 MaxSCCIterations 300 SCCTolerance 1E 4 SCCTolerance 1E 5 SlaterKosterFiles SlaterKosterFiles Type2FileNames Type2FileNames Prefix AN Prefix Separator Separator Suffix skf Suffix skf LowerCaseTypeName No LowerCaseTypeName No EEEH EEEH MaxAngularMomentum MaxAngularMomentum EEEH EEEH Mixer Broyden Mixer Broyden MixingParameter 0 2 MixingParameter 0 2 Filling Fermi Filling Fermi Temperature Kelvin Temperature Kelvin 300 0 300 0 EHHH EHHH KPointsAndWeights KPointsAndWeights EHHH EHEH Options Options ParserOptions ParserOptions ParserVersion ParserVersion e submit sh is the submission job file for perform th
25. c positions Here both POSCAR and cif file formats are supported For POSCAR file format users can use Selective dynamics tag to define the 17 relaxable atomic layers otherwise all the atoms in the substrate will remain un relaxed during surface reconstruction calculations For cif file format users need to insert selective tag beneath atom ste fract_z and then include T after atomic positions to define the relaxable atomic layers while F for the unrelaxable bulk Please refer to examples section 3 6 for more details By defining this variable to Automatic or Auto the substrate will be generated automatically from its bulk structure See below controlling parameters for generation of substrate based on its bulk structure Default value SUBSTRATE surf SurfaceAtoms string integer string integer End Defining the reconstructed surface region with desirable compositions in a format of mx2 matrix The row m rank of the matrix is determined by NumberOfSpecies For each row the matrix contains two columns The first column string is the elemental symbol of each chemical species followed by the number of atoms for such a chemical species integer Please refer to the Examples for more detailed setting of these parameters Default value There is no default You must supply this variable SpaceSaving logical If this parameter is set as True the output files for
26. confirmation of their results 16 Default value False MaxNumAtom integer The maximal number of atoms allowed in the simulation cell Default value 20 CtrlRange integer integer 12 integer integer End Defining the compositional range for each type of constituent atom in the binary A B system For atom A x varies from integer to integer while for atom B y varies from integer to integer Default value CtrlRange 1 6 1 6 End PEPE EEE ETE EE End HH ETE EEE HHHHHHHHHHE The parameters for surface reconstruction prediction 4H 44 JUDO DO AEE TP Examples are given in section 3 6 HHHRHRHHHHHHHHHH LSurface ogical The flag specifies whether surface reconstruction prediction will be performed Default value False CalSubstrate logical This tag determines whether the energy of the substrate is calculated Ideal substrate energy is usually used as reference energy to assess the stability of the reconstructed surface Default value False SurfaceThickness real This variable in unit of angstrom specifies the thickness of surface reconstruction and should be set as a value slightly larger than the double distance of two adjacent atomic layers in bulk materials Default value 3 0 Substrate string This variable allows the users to define their own substrates string is the name of the user defined substrate file which contains the structure data of the substrate i e lattice parameters and atomi
27. d POSCAR formatted are supported DFTB input files dftb_in hsd PRE 1 dftb_in hsd PRE 2 SK files submit sh A script to run VASP or DFTB a The calypso x can be found in Path to CALYPSO Sre b input dat contains all the parameters to run the surface structure predictions In this example we show how to control the program to adopt user specified substrate and use DFTB code to relax structures locally Here is an example input dat LSurface T Surface reconstruction predictions ICode 8 DFTBt Kgrid 0 2 PopSize 20 MaxStep 30 PsoRatio 0 6 SurfaceThickness 2 0 SurfaceAtoms atomic symbol count C 4 H 4 End Substrate SUBSTRATE surf c SUBSTRATE surf is the user defined substrate files The surface will be generated on top of this substrate Both VASP and CIF files can be recognized by CALYPSO Note that this file is optional One can set the Substrate control parameter in input dat to Auto or Automatic in order to generate substrate automatically via crystal info Please see previous example for more information For VASP format chemical symbols should be inserted just before the line of the numbers of atoms 59 Diamond 100 1 00000 5 054258 0 000000 0 000000 0 000000 5 054258 0 000000 0 000000 0 000000 19 467375 H C 8 24 Selective dynamics Direct 10783118 10783118 60783118 60783118 39216882 39216882 89216882 89216882 25000000 oO 00000000 50000000 00000000
28. di and di2 define the minimal C C and C H distances respectively while da and d define the minimal H C and H H distances respectively THEE End HAAL HHHH Parameters for design of optical materials with desirable band gap JUDO DO TATTA Examples are given in section 3 9 HHHRHRHHHHHHHHHH Band_edge logical When this flag is set as True design of optical materials with desirable band gap will be performed Default value False TarBandGap real This parameter is the desirable band gap for design of material THERE End Fee 22 The input files of VASP INCAR_ and pseudopotential file POTCAR are needed for local structure optimization using VASP The input files of SIESTA sinput_ and pseudopotential files psf are needed for local structure optimization using SIESTA The input files of GULP ginput_ are needed for local structure optimization using GULP The input files of CP2K cp2k inp_ are needed for local structure optimization using CP2K 2 6 2 CALYPSO Outputs All the major output files are listed in the folder of results CALYPSO_input dat Backup of the initial input file similar dat It contains the geometrical structure parameters of predicted structures pso_ini_ It includes the information of the initial structures of the th generation pso_opt_ It includes the enthalpy and structural parameters of the geometrically optimized structures of the th generation pso_sor_ The en
29. e DFTB calculation Here is an example of submit sh bin sh export OMP_NUM_THREADS 4 share apps dftb gt dftb log 2 gt amp 1 Once all the input files are ready you may run the structure prediction via the following 62 command or alternatively put this execution command into a job submit script calypso x gt caly log The result analysis process of surface structure prediction is similar to the crystal structure predictions Please see section 2 3 1 for more information You may run the following commands cd Results cak py cif You will see the surface formation enthalpy order of predicted structures in Analysis Output dat file and these newly predicted structure files with cif format can be found in dir_0 1 directory Please see section 2 6 3 for further information about the result analysis 63 3 7 Design of Superhard Materials This section is to show the examples for design of superhard materials using CALYPSO code Here VASP code is used for geometry optimization and enthalpy calculations The CALYPSO input file of input dat VASP input files of INCAR_ and pseudopotential file of POTCAR are needed The following files should be presented in the working directory File Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO containing controllable key parameters INCAR Input files for VASP POTCAR Pseudopot
30. e in input dat The adsorption site is needed The line below the word point gives numbers of sites in the cell The next few lines give the direct coordinates of adsorption sites d DFTB input files hsd formated input files are needed to relax structures dftb_in_1 hsd dftb_in_2 hsd Besides users should provide Slater Koster SK files for DFTB calculation Some SK files can be found in this link http www dftb org parameters 71 Geometry GenFormat Geometry GenFormat lt lt lt dftb gen lt lt lt dftb gen Driver ConjugateGradient Driver ConjugateGradient MaxForceComponent 1E 3 MaxForceComponent 1E 4 MaxSteps 1000 MaxSteps 1000 OutputPrefix geom out OutputPrefix geom out LatticeO0pt Yes LatticeO0pt Yes Hamiltonian DFTB Hamiltonian DFTB SCC Yes SCC Yes MaxSCCIterations 500 MaxSCCIterations 300 SCCTolerance 1E 4 SCCTolerance 1E 5 SlaterKosterFiles SlaterKosterFiles Type2FileNames Type2FileNames Prefix Prefix Separator Separator Suffix skf Suffix skf LowerCaseTypeName No LowerCaseTypeName No MaxAngularMomentum MaxAngularMomentum Mixer Broyden Mixer Broyden MixingParameter 0 2 MixingParameter 0 2 Filling Fermi Filling Fermi Temperature Kelvin Temperature Kelvin 300 0 300 0 EHHH HHH Options Options ParserOptions ParserOptions ParserVers
31. e name title e mail and mailing address of a person Please sign the CALYPSO license and send the scanned copy to calypso calypso cn The CALYPSO code will be sent Contact Information The best contact way is to send correspondence by email to calypso calypso cn or to Professor Yanming Ma mym calypso cn or Dr Yanchao Wang wyc calypso cn Mailing address Prof Yanming Ma State Key Lab of Superhard Materials Jilin University 130012 Changchun P R China Phone 86 431 85168276 FAX 86 43 1 85 168276 Personal webpage http mym calypso cn Contents T Introduction rat a a aia Ae aa 5 1 Meaning of CALYPSO onaren a diciendo 5 12 TEE E 5 1 3 History of PSO on Structure Prediction conc cc cnnnnnnnnnnns 6 2 CALYPSO Progra Mini a deed See 7 SL Program Features oe eee EE ee eea gege EENEG 7 2 2 Major N EE 7 KS Bug Repite nio Ea ar EEEE idas 9 Sue EE 9 2 5 Execution of CALYP SO aid 9 2 6 Basic inp ts and Outputs a a a a a aas 10 2 6 DMAP Ut dat a a ea aaa a aa E E A REA E EE E EE EARE 10 2 6 2 CAL ee RS 23 2 6 3 Analysis Of Results 23 EXIME esa 26 3 1 Crystal Structure Predicas dich ood dees teceebeas a aia at 26 3 2 Two Dimensional Structure Prediction sss sese 30 3 3 Cluster Structure PredictiO aii AA ee 34 3 3 1 Tutorial for Becl sterS neioii a a lacio 34 33 2 Futorial TO HEE CAUSAS A can dan Ee 38 3 3 3 Futorial Tor Bis clUSte Sia AEN 41 3 4 Molecular Structure Prediction nono nonn nono no
32. e number of relaxable layers in the substrate Default value 2 CapBondsWithH logical If this parameter is set as True the dangling bonds in the bottom side of the slab will be passivated via pseudo hydrogen atoms Default value True IDO ODO DO UU DOE End PIU ODO DUDO HHHHHHHHHHHH The parameters for design of superhard materials 7822924449902 IU DO OOO Example is given in section 3 7 AAA Hardness ogical when this tag is set as True structure design of superhard materials will be performed Here the fitness function is hardness instead of energy in a standard structure prediction After the CALYPSO design of superhard materials one can plot out hardness versus energy map to dig out the candidate structures with high hardness and low energies Please refer to for method description Default value False PEPE EH ETE ETE EH EEE End PEPER ETE ETE EEE The parameters for Adsorption of Two Dimensional Layer Materials FEEL DO ODO Examples are given in section 3 5 MA Adsorption logical When this flag is set as True the prediction of 2D materials with adsorption will be performed Default value False AdsorptionStyle integer It determines which method of generating structures should be adopted in the simulation 1 Generating structures freely 2 Generating structures with fixed position of adatom Default value 1 NumberOfTypeAtom integer The number of different types of adatoms in the s
33. ed by users request O It is written in Fortran 90 and memory is allocated dynamically 2 2 Major Techniques The success of CALYPSO method is on account of its efficient integration of several major structure dealing techniques l Structural evolution through PSO algorithm PSO is best known for its ability to overcome large barriers of energy landscapes by making use of the swarm intelligence and the smart self learning Both global and local PSO algorithms have been implemented The global PSO has the advantage of fast convergence while local PSO is good at avoiding structure premature ready for dealing with complex systems Symmetry constraints on generation of random structures to ensure the creation of physically feasible structure reduce searching space and enhance the structural diversity during evolution Two structural characterization techniques for eliminating similar structures and partitioning energy surfaces for local PSO structure searches 1 bond characterization matrix technique ii atom centered symmetrical function technique Introduction of random structures per generation with controllable percentage to enhance structural diversity during evolution 5 Interface to a number of local structural optimization codes varying from highly accurate DFT methods to fast semi empirical approaches that can deal with large systems Local structural optimization is the most time consuming part of CALYPSO structure
34. ed in the working directory File Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO containing controllable key parameters cell dat The input file contains fixed cell parameters and atomic positions INCAR Input files for VASP POTCAR Pseudopotential for VASP submit sh submission job file a The calypso x can be found in Path to CALYPSO Src b The parameters in input dat are shown as below c The cell dat file contains the fixed information of cell parameters and partial atomic positions In this example the atomic positions of C are fixed The first three lines indicate the lattice matrix The forth line is the number of atoms to be fixed followed with the fixed coordinates of atomic positions d We strongly suggest that you use the multi stage strategy for structure relaxation For VASP the number of INCAR_ INCAR_1 INCAR_2 is in accord with the setting tag of NumberOfLocalOptim In this example we use three INCAR files for structure 80 relaxation INCAR_1 and INCAR_2 are used to perform very crude structural relaxations with fixed volume and INCAR_3 is used to perform full structure relaxation with accurate precision INCAR_1 INCAR_2 INCAR_3 SYSTEM optimisation SYSTEM optimisation SYSTEM optimisation PREC LOW PREC Normal PREC accurate EDIFF 3e 2 EDIFF 2e 2 ENCUT 520 0 IBRION 2 IBRION
35. ed surface as determined by the Miller indices h k and for surface reconstruction Users might also use 4 numbered h k i D Miller index for hexagonal and rhombohedral lattice systems Default value There is no default value MatrixNotation interger intergery2 interger interger End This control parameter is used to define the reconstructed cell in unit of the ideal surface unit cell It is a 2X2 matrix Reconstructed cell can be obtained via multiplying this matrix by the unit lattice vectors For example reconstructed surface cell lattice vectors blue in the figure can be generated via the unit cell lattice vectors green in the figure multiplied by this matrix i e Ki _ E SE Ei b linteger21 integer22 la e 0 e e e e N e a Default value there is no default value for this variable SlabVacuumThick real This parameter in units of angstrom defines the vacuum region where the whole slabs is separated from its periodic images Default value 10 0 SlabDepth real This parameter defines which layer resides at the top of the slab By default the topmost layer is the one that contains the origin of the bulk unit cell Users can control the position of the topmost layer of the substrate via tuning this parameter Default value 0 0 19 SlabNumLayers integer This parameter controls the number of layers in the substrate Default value 6 NumRelaxedLayers integer This parameter controls th
36. ential for VASP submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre b The following parameters in input dat are shown for the carbon system SystemName Carbon NumberOfSpecies 1 NameOfAtoms NumberOfAtoms 1 64 Volume 14 0 al DistanceOofIon 1 30 QEnd Hardness c We strongly suggest that you use the multi stage strategy for structure relaxation For VASP the number of INCAR_ INCAR_1 INCAR 2 is in accord with the setting tag of NumberOfLocalOptim In this example we use four INCAR files for structure relaxation INCAR_1 and INCAR_2 are used to perform very coarse structure relaxation keeping the volume fixed INCAR_3 is used to perform full structure relaxation with medium precision and INCAR_4 is used to perform very accurate calculations 65 INCAR_1 SYSTEM local optimization PREC LOW EDIFF 3e 2 IBRION 2 ISIF 2 NSW 40 ISMEAR 0 SIGMA 0 05 POTIM 0 50 LWAVE FALSE LCHARG FALSE ISTART 0 PSTRESS 0 001 EDIFFG 4e 2 INCAR_3 SYSTEM local optimization PREC Normal EDIFF 2e 4 IBRION 1 ISIF NSW ISMEAR 0 SIGMA 0 05 POTIM 0 1 LWAVE FALSE LCHARG FALSE PSTRESS 0 001 EDIFFG le 3 3 0 d POTCAR is provided by the user The order of element must be the same with the setting tag of NameOfAtoms INCAR_2 SYSTEM local optimization PREC Normal EDIFF 2e 2 IBRIO
37. ges 3 6 1 Diamond 111 surface reconstruction prediction This subsection will introduce how to predict the 111 surface reconstruction of diamond using VASP as the local relaxation code In this example run the substrate will be generated automatically from crystal structures To perform surface structure prediction one needs to create a directory and prepare the following files File Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO specifying the controllable key parameters INCAR Input files for VASP POTCAR Pseudopotential for VASP submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre b input dat This file specifies all the control parameters that tell the program how to do the job of global surface structure search Most of the parameters have default values and do not need to be changed Note that the Substrate tag is specified to Automatic which means the substrate will be generated from the bulk crystal Here is an example input dat for diamond 111 surface reconstruction prediction 56 LSurface T Surface reconstruction predictions ICode 1 VASP Kgrid 0 2 PopSize 20 MaxStep 30 PsoRatio 0 6 SurfaceThickness 2 0 SurfaceAtoms atomic symbol count Following parameters are used to build the surface from bulk crystal info reconstruction symmetry Substrate Au
38. h is the job submission file for performing the VASP calculations and we must use the vasp 5 vasp 4 version is not support META GGA Here is an example of submit sh bin sh mpdboot mpiexec n 12 bin vasp 5 2 2 gt vasp log 2 gt dev null Once all the input files are ready you can simply type calypso x gt caly log to execute the CALYPSO in the sequential mode Or you can write this command into the pbs script and submit it CALYPSO will generate the results folder in current directory which contains all most the outputs files of CALYPSO To analyze the results just type following command cd results cak py cif 77 The index of the structures sorted by band gap in the ascending order is presented in Analysis_Output dat file and these newly predicted structure files with cif format are shown in dir_0 1 directory Please see section 2 6 3 for further information about the analysis of CALYPSO results 78 3 10 Crystal Structure Prediction with Fixed Cell Parameters or Atomic Positions This section is to show the examples for the three dimensional crystal structure prediction with fixed cell parameters or partial atomic positions using CALYPSO code Here VASP code is used for geometry optimization and enthalpy calculations The CALYPSO input file of input dat and VASP input files of INCAR_ and pseudopotential file of POTCAR are needed The following files should be present
39. he ionic radii of given atoms Default value 0 0 MultiLayer integer It defines the number of layers you would like to design Default value 1 DeltaZ real The distortion value in unit of angstrom along the c direction i e perpendicular to the a b plane of the layer If this parameter is set to be zero then a strict flat layer will be designed Otherwise a buckled layer is designed Default value 0 2 LayerGap real The gap between two layers i e the inter layer distance in unit of angstrom If this parameter is set as zero then one single layer will be designed Default value 5 0 VaccumGap real This variable defines the separations in unit of angstrom between the designed single layer or multi layers and its nearest neighboring periodic images This value should be large enough to ensure that interactions between the designed layer and its nearest neighboring periodic images are negligible Default value 10 0 LayerType integer integer integer 3 integer integer integers End Design of multi layers materials with desirable compositions in a format of mxn matrix The row m and column n ranks of the matrix are determined by the MultiLayer and NumberOfSpecies respectively For each row i e each layer the matrix column values are defined as the number of atoms for each chemical species as listed in the NameOfAtoms We take the design of double layers of B C N materials as an e
40. his section is to show the example for variable stoichiometry structure prediction using CALYPSO code Here VASP code was used for geometry optimization and enthalpy calculations The CALYPSO input file of input dat VASP input files of INCAR_ and pseudopotential file of POTCAR are needed The following files should be presented in the working directory File Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO containing controllable key parameters INCAR Input files for VASP POTCAR Pseudopotential for VASP submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre b The following parameters in input dat are shown for the LiH system 52 NumberOfAtoms NumberOfFormula Volume DistanceOfIon 1 0 0 8 0 8 0 8 End Kgrid 0 12 0 06 Command sh submit sh The MaxStep 30 If True Variational Stoichiometry structure prediction is performed VSC T vscEnergy 0 0 The Max Number of Atoms in unit cell MaxNumAtom 20 The Variation Range for each type atom CtrlRange 1117 c We strongly suggest that you use the multi stage strategy for structure relaxation For VASP the number of INCAR_ INCAR_1 INCAR 2 is in accord with the setting tag of NumberOfLocalOptim In this example we use four INCAR files for structure relaxation INCAR_1 and INCAR 2 are used to perform very crude
41. ile behind the marker line HERRERA amp FORCE EVAL METHOD QS amp DFT BASIS SET FILE NAME EMSL_ BASIS SETS POTENTIAL FILE NAME POTENTIAL amp MGRID CUTOFF 250 REL CUTOFF 50 amp END MGRID SOS EPS DEFAULT 1 0E 10 42 SEND QS amp POISSON PSOLVER MT amp END POISSON amp SCF SCF_GUESS ATOMIC amp OT ON MINIMIZER DIIS amp END OT MAX SCF 100 EPS SCF 1 0E 5 amp PRINT amp RESTART OFF amp END amp END amp END SCF amp XC amp XC_FUNCTIONAL PBE amp END XC_FUNCTIONAL amp END XC amp END DFT amp SUBSYS HEHEHE amp KIND B BASIS SET 6 311Gxx POTENTIAL GTH PBE q3 amp END KIND amp END SUBSYS SEND FORCE _EVAL amp GLOBAL PROJECT cp2k RUN_TYPE GEOMETRY OPTIMIZATION PRINT LEVEL LOW amp TIMINGS THRESHOLD 0 000001 SEND SEND GLOBAL MOTION amp GEO OPT OPTIMIZER BFGS MAX ITER 1000 MAX FORCE 0 00045 RMS_FORCE 0 0003 SEND GEO OPT amp END MOTION c The users should provide the EMSL_BASIS_SETS and POTENTIAL files which contain the basis set and pseudopotential for running CP2K code d submit sh is the job submission file for performing the CP2K calculations Here is an example of submit sh bin sh mpirun lsf cp2k popt i cp2k inp gt out cp2k 2 gt amp 1 Once all the input files are ready you can just type calypso x gt caly log amp to execute the code in the sequential mode Or you can write this command into the pbs script and submit it CALYPSO
42. imulation cell Default value There is no default You must supply this variable 20 InformationOfAdatoms string 1 integer integer 13 string integer integer End Defining the information of adatoms in a format of mx3 matrix The row m rank of the matrix is determined by NumberOfTypeAtom For each line the matrix contains three columns The first column string is the elemental symbol of each adatom followed by the number of adatoms in the simulation cell integer The third column integer is the type of the adatom Here 1 specifies atoms and 2 specifies molecules or functional groups We take graphane as an example The 7x3 matrix is defined as InformationOfAdatoms H 1 1 End Here H is the elemental symbol of adatom 1 H atom is adsorbed and indicates the adatom is atom instead of molecules or functional groups Default value There is no default You must supply this variable SuperCell integer integer2 Defining the number of creating a supercell Default value 1 1 RangeOfZAxis reall real2 Defining the region of length of the bond between the atoms of 2D material and adatoms reall specifies the maximum of the bond length and real2 specifies the minimum of the bond length Default value 1 7 1 2 AdsorptionSymmetry integer It determines which method of generating structures with symmetry should be adopted 0 Generating the structures with no symmetry constraint 1 Generating the structures w
43. ion 4 ParserVersion 4 HHH Ht HHH HH e submit sh is the submission job file for perform the DFTB calculation Here is an example of submit sh bin sh export OMP_NUM_THREADS 4 share apps dftb gt dftb log 2 gt amp 1 Once all the input files are ready you may run the structure prediction via the following command or alternatively put this execution command into a job submit script calypso x gt caly log 72 The result analysis process of surface structure prediction is similar to the crystal structure predictions Please see section 2 3 1 for more information You may run the following commands cd Results cak py cif You will see the surface formation enthalpy order of predicted structures in Analysis Output dat file and these newly predicted structure files with cif format can be found in dir_0 1 directory Please see section 2 6 3 for further information about the result analysis 73 3 9 Design of Optical Materials with Desirable Electronic Band Gap This section is to show the examples for design of optical materials with desirable electronic band gap using CALYPSO code Here VASP code was used for geometry optimization and enthalpy calculations The CALYPSO input file of input dat and VASP input files of INCAR and pseudopotential file of POTCAR are needed The following files should be presented in the working directory Files Description ca
44. ion of the k mesh sampling for local optimization 5 When you have prepared the necessary files such as POTCAR INCAR_ vasp pbs submitremote sh you can execute the calypso x on your local machine 85 4 3 The split mode The split mode of CALYPSO can provide a more flexible way to relax the structures The structural files with POSCAR format POSCAR_ are generated by calypso x You can relax these structures on any computer clusters or workstations Here is the workflow l 2 Enabling the split mode by setting Split T in the input dat Executing the calypso x to generate the structural files in POSCAR format of POSCAR_1 POSCAR 2 for one particular generation After optimizing all these structures one by one on all possible computer clusters please copy the OUTCAR and CONTCAR to OUTCAR 1 and CONTCAR_1 or QUTCAR 2 and CONTCAR 2 and then moves all these files to the directory where you executed calypso x Executing calypso x again to generate the POSCAR_ files for the next generation To repeat the steps 2 4 until the halting criterion is reached 86 5 Selected Publications 10 11 12 13 14 15 Reactions of xenon with iron and nickel are predicted in the Earth s inner core Nature Chem 6 644 2014 Caesium in high oxidation states and as a p block element Nature Chem 5 846 2013 Self assembled ultrathin nanotubes on diamond 100 surface Natu
45. ith usual symmetry constraint 2 Generating the structures with transferred symmetry constraints Default value 0 TypeOfSubstrate integerl If AdsorptionSymmetry equals 1 the parameter is needed Defining the type of the lattice of 2D substrate 1 Oblique 2 Rectangular 3 Square 4 Hexagonal Default value There is no default You must supply this variable BothSide logical when this tag is set as True the prediction will be performed on both sides 21 Default value False AdatomsOnBothsides integer integer integer integer end The number of adatoms is absorbed on both sides It is written in a format of mx2 matrix The row m rank of the matrix is determined by NumberOfTypeAtom The type of adatom of each line is determined by InformationOfAdatoms For each line the matrix contains three columns The first column is the number of adatoms on one side The second column is the number of adatoms on the other side Default value There is no default You must supply this variable DistanceOfAdatom real real realy realy end Minimal interatomic distances in unit of angstrom are in a format of nxn matrix The rank n of the matrix is determined by the sum of NumberOfTypeAtom and the element of substrate If we take graphane as an example with NumberOfTypeAtom 1 and one type of element in the substrate the 2x2 matrix is defined as DistanceOfAdatom di di2 dd vend Here
46. lowing files should be presented in the working directory Files Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO containing controllable key parameters gsinput_ Input files for Gaussian submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre b The parameters in input dat are shown below NumberOfSpecies 1 NameOfAtoms B NumberOfAtoms 6 ule NumberOfFormula 1 1 34 lDistanceO0flon T47 End t c Ialgo 2 PsoRatio my P ine Command MaxStep LMC T Cluster Vacancy c We strongly suggest that you use the multi stage strategy for structure relaxation For Gaussian the number of gsinput_ gsinput_1 and gsinput_2 is in accord with the setting tag NumberOfLocalOptim In this example we use two gsinput files for structure relaxation The gsinput_1 is used to perform very coarse structure relaxation and gsinput_2 is used to perform structure relaxation with medium precision During the structure relaxation gsinput_ files will be renamed as gsinput and atomic coordinates will be automatically attached by CALYPSO 35 chk calypso chk nprocs 12 sMem 12GB p RHF PBEPBE 3 21G SCF fermi MaxCycle 565 conver 4 OPT Redundant MaxCycle 50 MaxStep 29 loose NoSymm dr38 Geometry Optimization gchk calypso chk nprocs 12 sMem 12GB p RHF PBEPBE 3 21
47. lypso x The executable file for running CALYPSO program input dat The input file of CALYPSO containing controllable key parameters INCAR Input files for VASP POTCAR Pseudopotential for VASP submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre We take the structure prediction of carbon C as an example b The parameters in input dat are shown below SystemName C NumberOfSpecies NumberOfAtoms 1 prediction crystal strucute base on band gap Band_edge T the target band gap TarBandGap 2 5 4 ha 1e Volume Mi DistanceO0flon 1 0 1 0 End Kgrid 0 12 0 06 0 04 JL mha H rne Command sh submit sh MaxStep 30 c We strongly suggest that you use the multi stage strategy for structure relaxation For VASP the number of INCAR_ INCAR_1 INCAR_2 is one larger than the setting tag NumberOfLocalOptim In this example we use four INCAR files for structure relaxation and one INCAR files for calculating band gap INCAR_1 and INCAR 2 are used to perform very coarse structure relaxation keeping the volume fixed INCAR_3 is used to perform full structure relaxation with medium precision INCAR_4 is used to perform very accurate calculations and INCAR_ META is used to perform very accurate band gap calculations 75 INCAR_1 SYSTEM local optimization PREC LOW EDIFF 3e 2 IBRION 2 ISIF 2 NSW 40 ISMEAR 0 SIGMA 0 05 POTIM 0
48. mode of CALYPSO supports torque pbs system 1 The keywords Parallel and NumberOfParallel are needed for using the parallel mode For example you can add Parallel T and NumberOfParallel 4 to the input dat NumberOfParallel defines the number of structure relaxations you want to run in parallel 2 Modify the pbs script for generating the machinefile The machinefile is a file that contains a list of the possible machines on which you want your MPI program to run For example you should add cat PBS_NODEFILE gt machinefile before the line of executable calypso x usually calypso x gt caly 1log 3 Modify the script for vasp execution For example you can write the submit sh as below mpiexec machine snodefile n 12 vasp gt out vasp 2 gt dev null 83 4 2 Remote Submission The remote submission mode will allow the calypso running on a local machine while the VASP is running on the remote cluster This feature is yet under testing on Torque PBS system at present Here is the workflow 1 You should set the keywords RemoteParallel T to enable the remote submission mode And you can define the number of structures you want to relax at the same time by setting the keyword NumberOfParallel which just like in the parallel mode 2 Modify the submitremote sh file server test010 60 36 168 port 22 Please modify the server and the port to adapt your
49. nning CALYPSO program input dat The input file of CALYPSO containing controllable key parameters INCAR Input files for VASP POTCAR Pseudopotential for VASP submit sh Job submission file a The calypso x can be found in Path to CALYPSO Src b The parameters in input dat are shown below T ADistanceO0flon 2 6 End Command sh submit sh MaxStep 50 c In this example geometry optimization is performed by VASP code and only one input file for VASP named as INCAR_ 1 is needed SYSTEM Ti PREC Accurate ENCUT 300 EDIFF le 4 IBRION 2 ISIF 2 NSW 1000 ISMEAR 0 SIGMA 0 0001 POTIM 0 100 LCHARG FALSE LWAVE FALSE EDIFFG 0 01 ISPIN ISYM 39 d POTCAR should be provided by the user e submit sh is the job submission file for performing the VASP calculations Here is an example of submit sh bin sh mpdboot mpiexec n 12 bin vasp 4 6 gt vasp log 2 gt dev null Once all the input files are ready you can just type calypso x gt caly log E to execute the CALYPSO in the sequential mode Or you can write this command into the pbs script and submit it CALYPSO will generate the results folder which contains all most the output files To analyze the CALYPSO results just type following command cd results cak py cif The index of the structures sorted by energies in the ascending order is presented in Analysis_Output da
50. o your machine For Mac OS X please download the version of CALYPSO_ MAC 4 0 tar gz For 32 bit version of Linux please download the version of CALYPSO_x86 tar gz For 64 bit version of Linux please download the version of CALYPSO_x64 tar gz Installing the program is simple Please use the following commands tar zvxf Calypso _ tar gz cd CALYPSO_ The execution file calypso x can be found in the bin fold 2 5 Execution of CALYPSO A fast way to test your installation of CALYPSO code can be found in the directory of Tests Please simply run the commands cd CALYPSO_ Tests cp CALYPSO_ bin calypso x calypso x gt caly log Other examples are provided in the path to package Examples directory This directory contains basic input dat and shell scripts for running CALYPSO It is advisable to create independent directory for each job making a clean and neat environment We describe here in details an example of cubic boron nitride BN on how to run structure prediction by CALYPSO Here VASP code was used for geometric structure optimization and total energy calculations Please go to your working directory and do mkdir BN cd BN cp path to package Examples BN The following necessary files should appear in the working directory 1 input dat The CALYPSO input file 2 calypso x The executable file for structure prediction 3 submit sh Executable shell scripts to submit a job
51. og E to execute the CALYPSO in the sequential mode Or you can write this command into the pbs script and submit it CALYPSO will generate the results folder which contains all outputs files To analyze the results of CALYPSO just type following command cd results 32 cak py cif The index of the structures sorted by enthalpies in the ascending order is presented in Analysis_Output dat file and these newly predicted structure files with cif format are shown in dir_0 1 directory Please see section 2 6 3 for further information about the analysis of CALYPSO results 33 3 3 Cluster Structure Prediction This section is to show the examples for the cluster structure prediction using CALYPSO code In section 3 3 1 an example for boron clusters that composed of 6 atoms Bg is presented The Gaussian09 code was used for geometry optimization and energy calculations In section 3 3 2 an example for T clusters is given and the local optimization is performed by VASP Finally an example for Bj clusters is provided in section 3 3 3 The CP2K code was used for geometry optimization and energy calculations 3 3 1 Tutorial for Be clusters In this tutorial a CALYPSO structure search for Bg clusters is presented The local structure optimizations are performed by Gaussian09 code so you have to install the Gaussian09 in your machine All the reference files can be found in Path to CALYPSO Example B6_Gaussian The fol
52. on Typically a larger number of generations are needed for a larger system Default value 50 PickUp logical If this variable is set as True structure prediction will start from a specified generation see PickStep where a previous structure prediction job was unexpectedly interrupted Default value False PickStep integer At which generation the previously interrupted calculation should be re started Default value There is no default If PickUp True you must define this variable MaxTime integer The running wall time limit in unit of seconds of one single structure optimization If the optimization time goes beyond MaxTime the current structure optimization will be terminated Default value 7200 LMC ogical This variable determines whether the Metropolis criterion should be applied during the PSO structure evolution For cluster structure prediction it is highly recommended to set this variable as True Default value False HHHHH The parameters for structural prediction of two dimensional layers HHH JUDO TAPE EAE TP Examples are given in section 3 2 MA 2D logical when this tag is set as True structure prediction of 2 dimensional layer will be performed Default value False 13 Area real The area in unit of angstrom 2 per formula unit If you cannot provide a good estimation on the area please use the default value The program will automatically generate an estimated area by using t
53. originally designed to predict 3 dimensional 3D crystal structures Now CALYPSO has a more generalized meaning of structure prediction able to deal with structures ranging from 0D to 1D 2D and 3D CALYPSO with all capitalized letters is the only name in the field of structure prediction But the word Calypso has diverse meanings Calypso is the name of one of the Nereids sea nymphs in Greek mythology Calypso also refers to companies music places etc Have a look at Wikipedia http en wikipedia org wiki Calypso CALYPSO structure prediction software takes the advantage of structure evolution via PSO algorithm one of swarm intelligence schemes However many other efficient structure dealing techniques e g symmetry constraints bond characterization matrix etc were also implemented in CALYPSO We found that all these techniques implemented are equivalently important for the structure prediction It is therefore more appropriate to name the developed structure prediction method as a CALYPSO method 1 2 Why PSO As an unbiased global optimization method PSO is inspired by the choreography of a bird flock and can be viewed as a distributed behavior algorithm that performs multidimensional search see e g Kennedy amp Eberhart 1995 PSO is metaheuristic as it makes few or no assumptions about the solutions and can search very large spaces of candidate solutions dubbed as particles by moving them
54. ounds by the global optimization method J Am Chem Soc 133 16285 2011 2 Yanchao Wang Maosheng Miao Jian Lv Li Zhu Ketao Yin Hanyu Liu and Yanming Ma An effective Structure Prediction Method for Layered Materials Based on 2D Particle Swarm Optimization Algorithm J Chem Phys 137 224108 2012 Design of Superhard Materials Xinxin Zhang Yanchao Wang Jian Lv Chunye Zhu Qian Li Miao Zhang Quan Li and Yanming Ma First Principles Structural Design of Superhard Materials J Chem Phys 138 114101 2013 Surface Reconstruction Structure Prediction Shaohua Lu Yanchao Wang Hanyu Liu Maosheng Miao and Yanming Ma Self assembled ultrathin nanotubes on diamond 100 surface Nat Commun 5 3666 2014 2 3 Bug Report The CALYPSO package has been thoroughly tested for numerous systems by the CALYPSO team and other users and has been progressively improved by adding new features and eliminating bugs We would greatly appreciate comments suggestions and criticisms by the users of CALYPSO in case of bug report the users can contact the authors and send a copy of both input and output by E mail to the Ma group calypso calypso cn 2 4 Compilation For CALYPSO installation basic UNIX knowledge is required The users should be acquainted with the tar gzip awk sed and make commands of the UNIX environment Currently CALYPSO runs on Mac OS X and Linux Please download the different version of CALYPSO according t
55. rOfAtoms 10 block ChemicalSpeciesLabel 1 6 C 2 1H Sendblock ChemicalSpeciesLabel PAO BasisSize SZ Sendblock PAO BasisSizes kgrid_cutoff 8 0 Ang MeshCutoff 80 Ry PAO EnergyShift 0 01 Ry XC functional GGA XC authors PBE MaxSCFIterations 100 DM MixingWeight 0 150 DM Tolerance 1 d 5 DM NumberPulay 5 ElectronicTemperature 3000 K MD TypeOfRun cg Optim Broyden true MD VariableCell true MD ConstantVolume false MD MaxForceTol 2d 5 eV Ang MD NumCGsteps 40 Use Save CG true Use Save XV true MD Broyden Initial Inverse Jacobian 0 20 MD RemoveIntramolecularPressure true MD TargetPressure 100 000010 GPa ZM ForceTolLen 0 04 eV Ang ZM ForceTolAng 0 0001 eV deg ZM MaxDisplLen 0 1 Ang ZM MaxDisplAng 20 0 deg AtomicCoordinatesFormat NotScaledCartesianAng Sinclude Zmatrix data block MM Potentials 1 1 C6 16 292 1 392 2 2 C6 0 735 2 025 12 C6 3 185 1 649 Sendblock MM Potentials sinput_3 SystemName siesta SystemLabel siesta NumberOfSpecies 2 NumberOfAtoms 20 block ChemicalSpeciesLabel 1 6 C 2 1H Sendblock ChemicalSpeciesLabel PAO BasisSize DZ kgrid_cutoff 8 0 Ang MeshCutoff 80 Ry PAO EnergyShift 0 002 Ry XC functional GGA XC authors PBE MaxSCFIterations 100 DM MixingWeight 0 150 DM Tolerance 1 d 5 DM NumberPulay 5 ElectronicTemperature 3000 K MD TypeOfRun cg Optim Broyden false MD VariableCell true MD ConstantVolume false MD MaxForceTol 2d 5 eV Ang MD NumCGsteps 60 Use Save CG
56. re Commun 5 3666 2014 High Pressure Partially lonic Phase of Water Ice Nature Commun 2 563 2011 Pressure stabilized lithium caesides with caesium anions beyond the 1 state Nature Commun 5 4861 2014 Stacking Principle and Magic Sizes of Transition Metal Nanoclusters Based on Generalized Wulff Construction Phys Rev Lett 111 115501 2013 Global structural optimization of tungsten borides Phys Rev Lett 110 136403 2013 Towards direct gap silicon phases by the inverse band structure design approach Phys Rev Lett 110 118702 2013 Cagelike Diamondoid Nitrogen at High Pressures Phys Rev Lett 109 175502 2012 Predicted novel high pressure phases of lithium Phys Rev Lett 106 015503 2011 Substitutional Alloy of Bi and Te at High Pressure Phys Rev Lett 106 145501 2011 Novel Superhard Carbon C Centered Orthorhombic C8 Phys Rev Lett 107 215502 2011 Novel Superhard Carbon C Centered Orthorhombic C8 Phys Rev Lett 107 215502 2011 Spiral Chain O4 Form of Dense Oxygen Proc Natl Acad Sci USA 109 751 2012 Direct Band Gap Silicon Allotropes 87 16 17 18 19 J Am Chem Soc 136 9826 2014 Structural Evolution of Carbon Dioxide under High Pressure J Am Chem Soc 135 14167 2013 Tetragonal Allotrope of Group 14 Elements J Am Chem Soc 134 12362 2012 Predicted lithium boron compounds under high pressure J Am Chem Soc 134 18699 2012 Predic
57. rom in a format of nxn matrix The rank n of the matrix is determined by the NumberOfSpecies If we take MgB gt as an example with NumberOfSpecies 2 the 2x2 matrix is defined as DistanceOflon dy dy dad End Here du and di2 define the minimal Mg Mg and Mg B distances respectively while di and doa define the minimal B Mg and B B distances respectively Default value 0 7 A lalgo integer It defines which PSO algorithm should be adopted in the simulation 11 1 global PSO algorithm 2 local PSO algorithm 3 ABC algorithm with symmetry Default value 2 ICode integer It defines which code should be used for local structure optimization during the structure prediction 1 VASP SIESTA GULP PWSCF CASTEP CP2K Gaussian DFTB 9 LAMMPS Default value 1 oN DN MN A UU N NumberOfLocalOptim integer It defines how many times a structure should be optimized once it is generated during structure prediction The reason why we optimize the structure by several times is because the generated structures are often far from their local minima A single fine structure optimization typically leads to a non converged calculation Our tests indicated that three or four steps optimizations of one structure are the best solution During the course of these multiple optimizations the optimization degree should increase gradually coarse W medium fine If one uses VASP for structure optimization three or four input
58. s of isolated systems have been substantially improved by the CALYPSO team Lv Wang Zhu 8 Ma in 2012 where the success of this application has been backed up with the introduction of various efficient techniques e g bond characterization matrix for fingerprinting structures symmetry constraints on structure generation etc 2 CALYPSO Program 2 1 Program Features Predictions of the energetically stable metastable structures at given chemical compositions and external conditions e g pressure for 0D nanoparticles or clusters section 3 3 2D layers section 3 2 2D surface reconstructions section 3 6 and 3D crystals section 3 1 O Functionality driven design of novel functional materials e g superhard materials section 3 7 design of 2D material with adsorption see section 3 8 and optical materials with desirable electronic band gap section 3 9 respectively O Options for the structural evolutions using global or local PSO O Structure predictions with automatic variation of chemical compositions section 3 5 Incorporation of various structure constraints e g fixed rigid molecules section 3 4 fixed cell parameters fixed space group or fixed atomic positions section 3 10 CALYPSO is currently interfaced with VASP CASTEP Quantum Espresso GULP SIESTA FPLO Gaussian and CP2K codes for local geometrical optimization and total energy calculations Its interface with other codes can also be implement
59. s to perform a fixed space group calculation For a general structure prediction please use the default value All 230 space groups will be allowed for generation of structures Default value 1 230 FixCell ogical If this tag is set as True the structure prediction with fixed cell parameters will be performed Note that a file named as cell dat is needed to define the fixed cell parameters Please refer to examples section 3 8 for settings of cell dat Default value False FixAtom logical If this tag is set as True the structure prediction with fixed atomic positions will be performed Note that a file named as cell dat is needed to define the fixed atomic positions Please refer to examples section 3 8 for settings of cell dat Default value False PEPE EEE RRHH End HH HR The parameters for structure prediction with variational stoichiometry JUDO DO TALE TTP Examples are given in section 3 35 HHHH HTH VSC logical If this tag is set as True structure prediction of automatic variation of chemical compositions will be performed This technique is designed to explore all possible stoichiometries for given binary systems e g A B system at once However one has to take his her own risk for the use of this technique The search space has been significantly enlarged due to the existence of large number of possible stoichiometries We highly recommend separate simulations with fixed stoichiometries for
60. structure relaxation will be deleted Since the output files for structure relaxation are very large it is desirable to save the storage space in hard driver by deleting these redundant files However one might set this parameter as False in case of debugging the results Default value True ECR logical If this parameter is set as True only those initial structures that meet electron counting rule are accepted For more information on electron counting rule please refer to the paper by Lu et al Nat Commun 5 3666 2014 Note that ECR only works for semiconductors with large band gaps Default value False The controlling parameters below are used to generate substrate from bulk crystal structure ONLY when substrate tag is set as Automatic or Auto for short CifFilePath string This tag specifies the name of crystal structure in CIF format whose surface structure prediction is going to be performed One essential step in building the substrate from crystal structure is finding the surface unit cell of the defined slab orientation 18 And finding the surface unit cell needs the crystal lattice system information One of the simplest ways is providing the crystal symmetry information in the first place Therefore we require the crystal structure in CIF format MUST contain symmetry information Default vale There is no default value MillerIndex A k I This variable defines the target
61. t file and these newly predicted structure files with cif format are shown in dir_origin directory Please see section 2 6 3 for further information about the analysis of CALYPSO results 40 3 3 3 Tutorial for B12 clusters In this tutorial a CALYPSO structure search for B z clusters is presented The local structural optimizations are performed by CP2K so you have to install the CP2K code in your machine All the reference files can be found in Path to CALYPSO Example B12_cp2k The following files should be presented in the working directory Files Description calypso x The executable file for running CALYPSO program EMSL_BASIS_SETS Basis sets files for CP2K POTENTIAL Pseudopotential files for CP2K input dat The input file of CALYPSO containing controllable key parameters cp2k inp_ Input files for CP2K submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre b The following parameters in input dat are shown for the B cluster NumberOfFormula BDietance tton 1 7 End Ialgo Tr L LLC PsoRatio 41 Command sh submit sh Cluster Vacancy 12 12 12 c In this example geometry optimization is performed by CP2K code and only one input file for CP2K named as ep2k inp_ 1 is needed During the structure relaxation cp2k inp 1 file will be renamed as cp2k inp and atomic coordinates will be automatically inserted into this f
62. t files of INCAR_ and pseudopotential file of POTCAR are needed The following files should be presented in the working directory Files Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO containing controllable key parameters INCAR Input files for VASP POTCAR Pseudopotential file for VASP submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre We take the structure prediction of cubic boron nitride CBN as an example b The parameters in the input dat are shown below NumberOfSpecies NameOfAtoms B NumberOfAtoms 26 The Volume ADistanceO0flon 1 0 0 8 0 8 0 8 End 2 NumberOfLocalOptim Kgrid 0 12 0 06 Command sh submit sh MaxStep 30 c We strongly suggest that you use the multi optimization strategy for structure relaxation For VASP the number of INCAR_ INCAR_1 INCAR 2 files is in accordance with the setting tag NumberOfLocalOptim In this example four INCAR files are used for structure relaxation INCAR_1 and INCAR_2 files are used to perform very coarse structure relaxation with the fixed volume and INCAR_3 and INCAR 4 files are used to perform full structure relaxation ie variable lattice parameters variable volumes and variable atomic coordinates with medium and accurate precision respectively 27 INCAR_1 SYSTEM local optimization PREC
63. thalpy sorted in ascending order of the th generation struct dat Necessary information for all predicted structures the space group number the volume the number of atoms etc 2 6 3 Analysis of Results CALYPSO calculations typically generate a large number of structures It is necessary to devise a versatile tool for data analyses We here develop a CALYPSO_ANALYSIS KIT CAK allowing automatic structure analysis 2 6 3 1 Installation of CAK CAK should be installed on a Linux or a Mac OS X system Structure analysis through CAK tool needs the use of mathematic lib of numpy package in Python You might need to link the path of Python to the variable PYBIN in your Makefile Installation of CAK is simple S cd CALYPSO_ANALYSIS_KIT make Then add source somepath caly sh in bashrc 2 6 3 2 The CAK Commands Please go to the directory results and type cak py cd path to calculation results cak py 23 An output file named as Analysis Output dat will be generated By default Analysis_Output dat contains space group and enthalpy information of 50 energetically best structures Note that space groups of structures are determined with a tolerance of 0 1 by default Below we show several more advanced analyzing options cak py a All the structures generated by CALYPSO will be analyzed and output cak py n integer The specified number of best structures will be analyzed and ou
64. ting Two Dimensional Boron Carbon Compounds by the Global Optimization Method J Am Chem Soc 133 16285 2011 88 6 Acknowledgements We gratefully acknowledge the suggestions of Hongjun Xiang Yunlong Liao Sabino Maggi and Idoia G de Gurtubay and the support of all the CALYPSO ers 89
65. tomatic MatrixNotation 2 0 0 1 End UseCifFile T CifFilePath dia vasp cif MillerIndex BAN SlabDepth 0 0 SlabNumLayers NumRelaxedLayers CapBondsWithH c VASP input files For VASP users these files must present in the current work directory INCAR_1 INCAR 2 POTCAR and submit sh INCAR_1 and INCAR 2 are the central input files for VASP The structures will be relaxed by a set of relatively low precision control parameters in INCAR 1 first then the coarsely relaxed structures will be further relaxed by a set of relatively high precision parameters in INCAR_2 Users do NOT need to cat POTCAR of each atomic species together the program will do the job automatically However one has to copy POTCAR for each element to this directory and rename it to POTCAR format while the star stands for the name of the element Taking surface structures constructed by C and H atoms for example one needs to copy POTCAR of C and H to this directory and rename them to POTCAR C and POTCAR H respectively For pseudo hydrogens with fractional valence electrons such as Z 0 75 its POTCAR should be renamed to POTCAR H 75 57 INCAR_1 INCAR_2 PREC EDIFF EDIFFG IBRION ISIF LOW ENCUT 3e 2 EDIFF 4e 2 EDIFFG 2 IBRION ISIF NSW 0 H SIGMA 0 05 ISMEAR 0 050 POTIM FALSE LCHARG FALSE LWAVE 0 ISTART o Ln nou lH S L a Lo L Doo l we NSW ISMEAR POTIM LWAVE LCHARG ISTART Wa pes 0 05 oon Udy E
66. tput cak py t real One of the specified values ranging from 0 01 to 1 0 will be used as the tolerance for symmetry analyses of given structures The default value is 0 1 cak py m reall real2 realn Allowing multi tolerance values separated by space for symmetry analyses of given structures cak py cif m reall real2 realn The structure files with cif format will be generated in directories of dir_n n indicates the tolerance value for symmetry analysis cak py vasp m reall real2 realn The VASP POSCAR files containing structure information will be generated in directories dir ni n indicates the tolerance value for symmetry analysis cak py pri vasp cif The primitive cells of structures will output cak py p Plotting the figure of lowest enthalpy as a function of generations cak py hard The structures will output in the descending order of hardness 2 6 3 3 The Output Files 1 Analysis_Output dat 24 At least three columns are present in this file by default The first column is the index of the structures sorted by enthalpies in the ascending order the numbers in bracket indicate the structure number out of all structures generated in the CALYPSO calculation The second column shows the enthalpy data The corresponding space group numbers as obtained by the desirable tolerance value for the structures are listed in the third column 2 Conve
67. ts folder contains all the outputs files of CALYPSO To analyze the results please simply type the following commands cd results cak py cif The index of the structures sorted by enthalpies in the ascending order will be presented in the Analysis_Output dat file and all predicted structure files by CALYPSO with cif formats are shown in the dir_0 1 directory Please see section 2 6 3 for more information about the analyses of CALYPSO results 29 3 2 Two Dimensional Structure Prediction This section is to show the examples for the two dimensional structure prediction using CALYPSO code Here VASP code was used for geometry optimization and enthalpy calculations The CALYPSO input file of input dat and VASP input files of INCAR_ and pseudopotential file of POTCAR are needed The following files should be presented in the working directory Files Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO containing controllable key parameters INCAR Input files for VASP POTCAR Pseudopotential for VASP submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre b The following parameters in input dat are shown for the B N two dimensional system SystemName B3N2 NumberOfLocalOptim 3 30 Kgrid 0 1 0 07 Command sh submit sh LayerGap 5 VaccumGap 5 LayerType 1 1 2 1 LAtom Dis 1 4
68. ut dat 2D substrate structure information file of SUB dat DFTB input files of dftb_in_ hsd and Slater Koster SK files of skf are needed File Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO containing controllable key parameters SUB dat 2D substrate structure information Dftb_in_ hsd Input files for DFTB skf SK files for DFTB submit sh Job submission file a The calypso x can be found in Path to CALYPSO Sre b The following parameters in input dat are shown for the graphane system 68 HAHAHAHA The Basic Parameters of CALYV DSO AAA AAA AAA L NameOfAtoms C H It determines which algorithm should be adopted in the simulation Ialgo The proportion of the structures generated by PSO PsoRatio 0 6 It determines which local optimization method should be interfaced in the simulation ICode 8 The Number of local optimization for each structure NumberOfLocalOptim 2 The precision of the K point sampling for local optimization Kgrid 0 1 0 06 The command to perform local optimization calculation e g VASP SIESTA on your computer Command sh submit sh The population size Normally it has a larger number for larger systems PopSize 30 the max step of iteration MaxStep 30 HAHAHAHA The Parameters For 2D Material With Adsorption 4 If True a prediction of 2D
69. xample Here we have MultiLayer 2 NumberOfSpecies 3 and NameOfAtoms B C N If we purposely design the materials with the first layer having 6 C and 4 N atoms and the second layer having 3 B AC and 7 N atoms The 2x3 matrix can be written as follows LayerType 064 14 347 End Default value There is no default You must define these numbers LAtom_Dis real The minimal interatomic distance in unit of angstrom Default value 1 0 PEPER EH ETE EH EEE End HAHAHAHA HHHHHHHHHHHHHH The parameters for cluster structure prediction H HHHHHHHHHHH IU DO ODO Examples are given in section 3 3 MAAT Cluster ogical when this tag is set as True a cluster structure prediction is performed Default value False Vacancy reall real2 real3 The isolated cluster is placed into an orthorhombic box where the periodic boundary condition is applied This variable defines the separations in unit of angstrom between the studied cluster and its nearest neighboring periodic images It should be large enough to ensure that interactions between the studied cluster and its nearest neighboring images are negligible Default value 10 0 10 0 10 0 For cluster structure prediction we do not recommend the use of VASP for the structural optimization for large systems since computationally VASP calculations are very expensive We recommend using SIESTA CP2K and Gaussian codes when cluster sizes are larger than 10 atoms PERE EEE E
70. xhull dat This file contains two columns The first column shows various stoichiometries for a given binary system The second column presents the lowest enthalpy for each stoichiometry 3 plot dat This file contains two columns The first column shows the generation number while the second gives the lowest enthalpy for each generation 4 UCell_m_n vasp This file contains the structure data in conventional cell in the POSCAR format of VASP m and n indicate the enthalpy ranking number and the space group number of the structure respectively 5 PCell_m_n vasp This file contains the structure data in primitive cell in the POSCAR format of VASP m and n indicate the enthalpy ranking number and the space group number of the structure respectively 7 m_n cif This file contains the structure data in conventional cell in the cif format m and n indicate the enthalpy ranking number and the space group number of the structure respectively 8 m_n_p cif This file contains the structure data in primitive cell in the cif format m and n indicate the enthalpy ranking number and the space group number of the structure respectively 25 3 Examples 3 1 Crystal Structure Prediction This section is to show the examples for the three dimensional crystal structure prediction using CALYPSO code Here VASP code was used for geometry optimization and enthalpy calculations The CALYPSO input file of input dat and VASP inpu
71. y ge ir nology l d submit sh It is the script for local relaxation via VASP Here is an example of submit sh bin sh mpiexec n 12 share apps vasp vasp 5 2 gt vasp log 2 gt 81 Once all the input files are ready you may run the structure prediction via the following command under unix linux operating system or alternatively put this execution command into a job submision script calypso x gt caly log The result analysis process of surface structure prediction is similar to the crystal structure predictions Please see section 2 3 1 for more information You may run the following commands cd results cak py cif You will see the surface formation enthalpy order of predicted structures in Analysis Output dat file and these newly predicted structure files with cif format can be found in dir_0 1 directory Please see section 2 6 3 for further information about the result analysis 3 6 2 Hydrogenated diamond 100 surface reconstruction Example to perform hydrogenated diamond 100 surface reconstruction prediction using predefined substrate file and DFTB local relaxation will be demonstrated in this subsection Files needed to run this example are listed below 58 File Directory Description calypso x The executable file for running CALYPSO program input dat The input file of CALYPSO specifying the controllable key parameters SUBSTRATE surf Substrate file cif an

User's Guide

Contents

Download Pdf Manuals

Related Search

Related Contents