| |
|
|
|
Jaguar needs certain types of files in order to run a job. An input file must be created, of course, but additional files specifying the basis functions, data for the initial guesses, dealiasing functions, grids, and cutoffs used during a run are generally necessary as well. For most calculations, unless other files are specified, Jaguar uses the files default.basis, default.atomig, default.daf, default.grid, and default.cutoff in the data directory. For many solvation calculations, Jaguar also uses the file default.lewis. All of these files are provided by Schrödinger.
If you want to use other data files than those described above, you have several options. First, you can create a new data directory and put files in it whose names and formats match those described above, then specify that data directory from the interface. Note that in order to be able to list that directory as a possible choice, you will have to edit the jaguar.hosts file you are using for the job so that it includes the directory in a `datapath' line. (See section 6.1 for information on jaguar.hosts files.) A second option is to edit the input file for the job from outside of the interface, adding or editing "BASISFILE," "ATOMIGFILE," "DAFFILE," "GRIDFILE," "CUTOFFFILE," and/or "LEWISFILE" lines with the paths and names of the files you want to use. See section 8.2 for more details. However, if you specify a.cutoff file called "accurate.cutoff," "quick.cutoff," or "solvent.cutoff," the program will assume you are using an outdated file and will reset the name to "default.cutoff," so be careful about using these names for new files.
The rest of this chapter contains descriptions of the basis, atomic initial guess, dealiasing function, grid, cutoff and Lewis files. Even if you do not plan on creating your own versions of these files, you may want to skim this chapter if you are curious about the methods used in Jaguar.
The basis sets available for use in Jaguar appear in a file called default.basis, in the standard data directories. Portions of this file are shown in this section; you may wish to refer to them as you read the description of the file.
The basis sets are described in turn. Basis sets at the top of the file do not contain effective core potentials, and will be described first here. The basis sets with effective core potentials, whose names begin with "LA," will be described later.
Each basis set description begins with a blank line. The next line (or lines) must begin with the word `BASIS,' followed by one space. That label is followed by one or more names of the basis set to be described: the name of the basis set as given in Table 3.8.1 or Table 3.8.2 in Section 3.8, and any other names which describe the same basis set (e.g. STO3G and STO3G). The basis set names are separated by commas, and must include `*' and/or `+' characters, if those are allowed for that basis set. (`**' or `++' character strings are sufficient to describe the `*' and `+' cases also, and the `*' characters can be listed either before or after the `+' characters.) The next notation in the line, `5D' or `6D,' sets the default number of functions for d shells when using that basis set, as described in Section 3.8.
"Backup" basis set name or names, which are each preceded by the word `BACKUP,' may follow on the same line. If any sets are listed after the word `BACKUP,' it indicates that if an atom is not found in the current basis set (whose name is listed before `BACKUP'), its basis function will be obtained from the list of backup basis sets. (If there is more than one backup name listed, the basis function for the atom will come from the first backup set listed that contains that atom.) Note that the numbers of d shells specified in the backup basis sets are ignored. Also, polarization or diffuse functions are chosen according to the basis set specified by the calculation; that is, *, **, +, or ++ options on backup basis sets are ignored if they do not agree with the options on the basis set chosen for the calculation.
The basis set description continues with a set of lines describing the basis functions on each atom. The information for each atom begins with a line containing the element's symbol (e.g. He). The atomic symbol must not be preceded by any spaces or characters. The next line begins with the type of function being described (S, P, or D, for instance). If this label is `SP,' the corresponding set of data describes an s and a p function whose Gaussians have the same exponents. The next number in that line is the polarization/diffuse function parameter. If it is a 1, it indicates a polarization function which is included in the basis set if the basis set name ends in a *, as described in Section 3.8. If the number is a 2, it indicates a `**' basis set function; if -1, a `+' basis set function; if -2, a `++' basis set function. Otherwise, the number should be 0.
The rest of the numbers on that line determine the way that Jaguar will contract some of the functions, and the "range" of each function. The numbers before the dash (-) describe how many of the functions are included in that contraction. For example, if there were two such numbers, 2 and 1, the line would indicate that Jaguar would contract the first two Gaussians provided immediately below into one contracted function, and would treat the third Gaussian as an uncontracted function.
If you want to add or change a basis set to a .basis file, you should probably contract together all Gaussians whose exponents are greater than 0.3. The default.basis information generally follows this rule, although there are some exceptions (see the Li s and p function information in the sample file below for an example).
The numbers after the dash describe the range of each such function. There should be one such number for each contraction number before the dash. A zero indicates that the contracted function will be treated as a long-range function, while a 1, 2, 3, or 4 indicate various types of short-range functions. These assignments help determine the symmetrization of the Fock matrix components by the "side choosing" method described in reference [13]. These range values are only used in pseudospectral calculations, so if your basis set will be used for non-pseudospectral calculations, just use a `0' as a place holder for each range value. (Pseudospectral calculations require that grids and dealiasing functions exist for the basis set. These will be defined in the default.grid and default.daf files, respectively; see below.)
The Gaussians in the contraction are listed next, with the first number in each of these lines describing the exponent for the Gaussian, and the second its coefficient in the contraction. The Gaussians should be listed in decreasing size of exponent. If both s and p functions are being described, the second number on the line corresponds to the coefficient for that Gaussian in the s function's contraction, and the third number corresponds to the p function's contraction coefficient. The data for that atom ends with a line containing 4 * characters, with no spaces or other characters preceding them.
When all of the atoms for that basis set have been listed, ending with the obligatory `****' line, the next basis set is listed, in the same manner described above.
The beginning of the default.basis file is shown below to illustrate most of these points.
Basis sets containing effective core potentials (ECPs) are described in a slightly more complicated fashion. First, the string `ECP' must appear between the `5D' or `6D' label and the `BACKUP' label. This string indicates that the basis set description contains information about the effective core potential associated with the basis set.
As for the basis sets without effective core potentials, each atom in the set is described in turn. The description begins with the basis function, which is in the same format as those described above. After a line with two asterisks (`**'), the effective core potential is described.
The first line in the effective core potential description contains the element symbol (e.g. Na) and two numbers. The first of these indicates the maximum angular component in the core, and the second indicates the number of electrons replaced by the effective core potential. Next, the information for various angular components is listed. The first set of such information contains the local components of the ECP and should begin with a line saying "D_AND_UP," "F_AND_UP," or "G_AND_UP," which indicates that the maximum angular components to be described are 2, 3, or 4, respectively.
Following that line, the different terms for this angular component are given. Each line describes a term of the form
, listing the parameters n, a, and C (from left to right) in a free format. Next, the lower angular components are listed in increasing order (S, P, D, etc.) in the same fashion.
A line with four `*' characters appears the end of the description of each atom's ECP. When all atoms have been described in turn, the next basis set is described.
The example below shows the beginning of the description in default.basis of the LAV2D and LAV2P basis sets. Note that these basis sets only differ in their choice of what basis set to use for atoms which are not described by the effective core potential.
If you want to set up your own .basis file, you can do so, if you use the format described above. Generally, you must also create an altered version of the .atomig file, which is described in Section 9.2, although if you are just adding polarization functions to the basis set, and these functions are identified by the polarization/diffuse function parameter described earlier in this section, you can continue to use the usual .atomig file. Make sure your new .basis file contains the 631G basis set, because the initial guess program needs this basis set. If you alter the basis functions in the default.basis file only slightly, you can use the same names for the basis sets. If you change them a great deal, you should use a new name, so that Jaguar will not attempt to use grids or dealiasing functions that do not match the new basis set. If you change a basis set name to something Jaguar does not recognize, runs using that basis set will use all-analytic methods (see the subsection Analytic Corrections in Section 3.9, or the information on the input file gen section keyword nops in the subsection Keywords for SCF Methods in section 8.6).
To use the file in a Jaguar calculation, you must add a line in the form
to the input file for the job. You can specify a file on another host, or under another account name on that host, by listing the file name in the format <host>:<file path and name on host>, or <user>@<host>:<file path and name on host>.
The file default.atomig contains the results of Hartree-Fock calculations on atoms for various basis sets. By default, the initial guess is constructed from wavefunctions in this file. When the basis set to be used for the calculation is 631G, MSV, LAV2P, LAV2D, LAV3P, LAV3D, LACVP, or LACVD (or any variant of these sets involving polarization or diffusion functions (e.g. 631G*)), the initial guess is formed from the wavefunctions obtained from individual calculations on the atoms in the molecule which were calculated using that same basis set (ignoring polarization and diffusion functions). Therefore, if you change the .basis file, you need to change the .atomig file correspondingly, and vice versa.
For other basis sets, the wavefunctions used to construct the initial guess are obtained by projecting either the appropriate atomic wavefunction in default.atomig onto the basis set actually being used for the molecular calculation. The 631G wavefunction is used whenever possible; when a 631G atomic wavefunction is not listed for a particular atom, the MSV one is used for that atom. For atoms beyond Xe in calculations using the LAV1S basis set, the LAV2P atomic results are used.
As in the default.basis file, the basis sets listed in the default.atomig are listed in turn, and for each basis, the information for each atom is listed. Each basis set section begins with a blank line, which is followed by one or more lines reading `BASIS,' followed by one space, and ending with the name or names (separated by a space and/or comma) of all basis sets for which the atomic calculations listed immediately after that line apply. The basis set names are listed in Table 3.8.1 and Table 3.8.2 in Section 3.8.
Next, the information for each atom follows. The first line lists the atomic symbol for the atom, followed by information which is simply a comment and is not read in. The next line lists two numbers. The first of these numbers gives the number of basis functions for that atom and basis set, as listed in the default.basis file, and the second gives the number of electrons for that atom included in an effective core (0 for the basis sets whose names do not start with "LA"). The line after that lists the orbital number (1 if it is the first orbital listed for that atom, 2 if it is the second, and so on), the orbital occupation (i.e, the number of electron pairs in that orbital), and the orbital energy in Hartrees. That orbital's coefficients for each basis function for the given atom and basis set(s) follow on the next line(s).
When all of the orbitals for that atom have been specified, a line with 4 `*' characters indicates the end of the information for that atom, and the data for the other atoms is listed. Similar information for each other basis set follows.
If you want to set up your own .atomig file, you can do so, if you use the format described above. To use the file in a Jaguar calculation, you must add a line saying
to the input file for the job. You can specify a file on another host, or under another account name on that host, by listing the file name in the format <host>:<file path and name on host>, or <user>@<host>:<file path and name on host>.
When Jaguar fits a function's grid point values to a basis set to find the applicable basis set coefficients for the function, it uses dealiasing functions to reduce errors. The dealiasing functions span the function space determined by the grid more completely than the basis functions, so a function on the grid can be better described using the dealiasing functions than by the basis functions alone. The basis set coefficients for the function can then be determined by using the overlap between the dealiasing functions and the basis set functions, which is determined analytically.
Some basis functions die off slowly and require long-range functions centered on each atom in the molecule, while others die off quickly over distance and can be described with short-range dealiasing functions centered on the nearby atoms. The latter type can employ different dealiasing functions, depending on the distance between the atom upon which the relevant basis function is centered and the atom upon which the short-range dealiasing functions are to be centered. If the atoms are the same, "home atom" dealiasing functions are used; otherwise, the distance between the two atoms determines whether the dealiasing functions used should be those for first-order or one of the other higher-order neighbors. (To see this connectivity information for a system, use the gen section keyword setting ip12 = 2.) If the two atoms are further away than the farthest neighbor range specified, no dealiasing functions on one atom are used in calculating the contribution of a short-range basis function centered on the other atom.
The dealiasing functions themselves are simple polynomials multiplied by Gaussian functions, and are s-type, p-type, and so on, depending on the polynomial. Uncontracted dealiasing functions are simply formed by specifying the exponent of the Gaussian function. Contracted dealiasing functions are defined as linear combinations of the appropriate type of functions; the coefficients and exponents for the linear combination are the same as those used in the basis set for the contracted basis functions for the relevant function types (1s, 2s, 2px, etc., depending on the molecule and the basis set). Thus, a dealiasing uncontracted function can be specified by dictating the type (s, p, d, etc.) and the exponent desired for the Gaussian, while a contracted Gaussian function can be specified by dictating the type and referencing which set of contraction coefficients and exponents are desired.
The File Format and Description subsection below describes the file that determines the dealiasing functions for a calculation. Sets of dealiasing functions must be provided for each grid used in the calculation. Comments about a sample file refer to the sample .daf file in the subsection Sample File.
The first line of a dealiasing function file contains a character string which includes the version number of Jaguar. This string should be "dafv" followed immediately by four digits giving the version number times 100. Lead zeros are added if necessary.
The next line is made up of two integers. The first integer dictates the number of dealiasing function sets provided for each atom type; each set is used for a particular grid during the calculation. The ordering of the sets used for each grid type is determined by the parameters named dcoarse, dmedium, and so on, which are specified in the gen section of the input file. (By default, the coarse grid is listed first, then the medium, fine, ultrafine, and gradient grids, in that order.)
The second number in the second line gives the number of ranges described in each of these dealiasing function sets. The ranges correspond to particular RwR blocks for the calculation. One of these ranges is the long range, basically covering the whole molecule; another is the home atom range, which actually only includes the relevant atom itself; and the rest are increasingly large neighbor ranges. The number of ranges should currently not exceed 10. The sample file's second line indicates that for each basis set, five dealiasing function sets are specified for each atom, and that each of these sets contains dealiasing functions for a total of six ranges: the long-range functions, the functions for the home atom, and the functions for four other neighbor ranges.
The distances defining the neighbor ranges are set in the next line of real values, in units of bohr. (Please note, however, that generally only the third neighbor range is actually used.) The first distance specifies that if the basis function whose coefficient is being evaluated is to be approximated by short-range dealiasing functions, then the dealiasing functions for first-order neighbors will be used for each atom within this distance of the atom upon which the basis function is centered (except for the basis function atom itself, for which the home atom dealiasing functions will be used). The second distance defines which atoms are considered second-order neighbors to each other, and so on. Since the number of neighbor ranges includes not only these ranges but also the long range over the entire molecule and the home atom range consisting of the relevant atom itself, the number of neighbor ranges actually specified in this line of the .daf file should be two less than the number of ranges listed in the previous line. Thus, in the sample file, the distances listed specify the neighbor ranges for first- through fourth-order neighbors.
The rest of the .daf file contains the dealiasing function sets for each atom type within each basis set. The data for each basis set should begin with a line listing the basis set name (as listed in Table 3.8.1 and Table 3.8.2 in Section 3.8), including the `*' characters indicating the polarization functions (e.g. 631G**). The first line for each atom type for that basis set should list three integers: the atomic number for that atom type, the number of uncontracted dealiasing functions about to be listed for each neighbor range in each set, and the corresponding number of contracted dealiasing functions. In the sample file, the first atom whose dealiasing functions are listed is hydrogen, since the atomic number listed is 1. The same line says that ten uncontracted functions and two contracted functions will be specified for each range in the five sets of dealiasing functions for hydrogen.
The second line for the same atom type should list real dealiasing exponents for each uncontracted function. The exponents specify what functions can be used. For instance, in the sample file, hydrogen's s-type uncontracted basis function from the first exponent would be
, while the p-type uncontracted basis function for the same exponent would be
. (The N's are normalization constants.)
Beneath those two lines, the dealiasing function sets for that atom type should be listed set by set. By default, the first set will be used for the coarse grid, the second for the medium grid, and so on, with the last set corresponding to the gradient. This ordering can be changed in the gen section of the input file. Each set should contain a line for each neighbor range; the long-range functions should be specified first, then the home atom functions, then the functions for each neighbor range, in increasing order. Within each line, there should be several integers, one for each uncontracted function, then one for each contracted function. These integers dictate how to construct the actual functions from the exponents (just given in the .daf file for uncontracted functions, and already established in the .basis file for contracted functions) and contraction coefficients for contracted functions (also established in the .basis file). If the value is 1, an s-type function will be constructed using the relevant exponent or exponents; if 2, a p-type function; if 4, a d-type function; if 8, an f-type function; and if 16, a g-type function. To construct more than one of these types of functions with the same exponent or exponents, the relevant numbers should be added together (for instance, 1 + 2 + 4 = 7 for s, p, and d).
The exponent or exponents for each of these functions will be determined by the position of the entry in the row. The uncontracted functions are described first, in the same order as their exponents were listed earlier, and the contracted functions corresponding to the contracted functions found in the .basis file are described next, in the same order as in the basis set file. (Uncontracted functions in the basis set file should be ignored.) Finally, the first derivatives of the basis set file contracted functions will be calculated, and the values listed for these "extra" functions will correspond to the functions generated this way, in order of the function they were generated from and, within that order, of increasing complexity (s before p, etc.). For instance, if the basis set contained contracted functions for 1s, 2s, and 2p orbitals, the derivatives would be listed in the following order: a p-type function resulting from the derivative of the 1s function, a p-type function resulting from the derivative of the 2s function, an s-type function resulting from the first term of the derivative of the 2p function, and a d-type function resulting from the second term of the derivative of the 2p function.
The last six lines of the sample .daf file correspond to the gradient dealiasing function set for He (note that the atomic number specified for those five dealiasing function sets was 2). The first line of this set describes this set's long-range dealiasing functions centered on the He atom, which will be used when coefficients for long-range basis functions are to be calculated, as explained above. The second value on this line, 3, dictates that uncontracted s-type and p-type (1 + 2 = 3) basis functions are to be constructed using the second exponent provided for this atom (0.145957). The second line of the set, which describes this set's He-centered dealiasing functions to be used when calculating the coefficients for He-centered short-range basis functions (the home atom line of the set), has a value of 1 entered in the eleventh column, meaning that an s-type contracted function will be calculated using the exponents provided for the first contracted function for He in the basis set. Since this basis set only provides one contracted function for He, the 1s function, whose derivative is a p-type function, the last number entered on that line (2) dictates that a p-type function be constructed, using the contraction coefficients and exponents that correspond to that derivative function, as explained in the previous paragraph.
The following sample .daf file lists the dealiasing sets for H and He for a 631G** basis set. Note that since a FORTRAN free-format read statement reads lines until all variables are assigned values, blank lines may be added for readability, and data may be spread over multiple lines.
The grid input file (.grid file) determines the grids used during the calculation. Each grid type, for example, "coarse" or "ultrafine," is constructed from grids assigned to each atom in the molecule. Therefore, for any basis set for which the pseudospectral method is used, the grid file must contain grids for each grid type used, where each of these grid types in turn requires atomic grids for each element in the molecule. The assignment of grids to grid types is performed within the input file, using the gen parameters gcoarse, gmedium, and so on.
The first line of a .grid file contains a character string which includes the version number of Jaguar. This string should be "gridv" followed immediately by four digits giving the version number times 100. Lead zeros are added if necessary.
The next line should consist of an integer which gives the number of grid types described in the file. For instance, this number would be six if the grids specified were of the types coarse, medium, fine, ultrafine, eldens (for electron density calculations), and gradient. By default, Jaguar uses the coarse grid for electron density calculations and the ultrafine grid for gradient calculations, and the "extreme" grid is included for testing purposes, so the number of grid types in the file default.grid is actually five. Jaguar uses the grids upon each atom in the molecule provided by the .grid file to generate molecular grids.
All grids for each basis set are then listed in turn. The basis set is identified with a BASIS line and containing its name, and is followed by a blank line.
Each molecular grid description starts with two comment lines, usually a blank line followed by a descriptive line. The next line contains an integer flag which determines which points from the atomic grids for the atoms in a molecule are included in the molecular grid. Jaguar generates a boundary plane between the two atoms and perpendicular to the vector between them, disposing of any points from one atom that are on the other atom's side of the boundary plane. The integer flag determines the location of this plane: if the flag is 0, the plane is located so that the ratio of the distances of the atoms to the plane is the same as the ratio of their covalent radii, while if it is -1, the boundary plane is set where the grid point density from each atom, on the vector between the atoms, is equal. The grid point density is determined as a spline fit of the density for each shell, where each shell's density is determined as the number of points for that shell divided by the shell volume, which is the volume between the spheres whose radii are the average of the current and previous shell radii, and the current and following shell radii.
After the flag for the grid, information for each atomic grid is provided. The first line of each atomic grid section contains two integers, one providing the atomic number for that atom and the other giving the number of shells to be described. Currently, this second number should be 30 or less. The next line contains that number of entries defining the radial shell spacing, listing the radius of each shell in bohr. Grid points for that shell will be placed at that radius, in a pattern determined by the integers given in the third line. This last line of integers represents the density of the angular grid for each shell. The values are explained below.
The default.grid file for Jaguar version 4.1 begins as follows:
Since a FORTRAN free-format read statement reads lines until all variables are assigned values, blank lines have been added between atomic grids for readability. Data may be spread over multiple lines.
As explained above, the beginning of the default.grid file indicates that five grid types are listed for each atom (corresponding to the coarse, medium, fine, ultrafine, and gradient grids. All coarse grids for 631G (with or without the polarization functions indicated by the `**') will set the boundary plane between atoms (described earlier) at the point where the grid point densities are the same for the two atoms, because of the `-1' flag. Next, seven shells apiece are specified for H (atomic number 1), He (atomic number 2), and Li (atomic number 3). The actual default.grid file continues with a list of coarse atomic grids for the other atoms in the basis set, followed by the medium, fine, and ultrafine atomic grids in the same format, before proceeding to define the grids for another basis set in the same manner.
The possible values of the numbers on the angular grid line are listed in Table 9.4.1,
along with the corresponding number of points per angular shell and the degree of the highest spherical harmonic which the grid integrates exactly, when relevant. (The full references are provided near the end of the User's Guide.)
The cutoff file specifies parameters to be used for the various iterations of an SCF calculation. The file to be used is determined by the `CUTOFFFILE' entry in the input file, as described in section 8.2; if the input file has no such line, Jaguar uses the file default.cutoff from the data directory. Note that if the `CUTOFFFILE' entry is "accurate.cutoff," "solvent.cutoff," or "quick.cutoff," the program will interpret the setting as "default.cutoff."
The first line of a cutoff file contains a character string which includes the version number of Jaguar. This should be "cutv" followed by four digits giving the version number times 100. Lead zeros are added if necessary. A comment on the same line can follow the version string.
The next five lines each have five numbers. Each line describes a particular level of accuracy to be used for the calculation. The first line provides the information necessary to run a calculation with all-ultrafine pseudospectral grids and with "tight" cutoffs, and corresponds to an accuracy level setting of ultrafine from the interface, as described in Section 3.9 under Accuracy Level, or to the keyword setting iacc = 1 in the gen section of the input file, as described in the Keywords for SCF Methods subsection in section 8.6. The second line gives the parameters for the accurate level (iacc = 2), while the third line provides information for the quick level (iacc = 3). The last two lines are filled with zeroes, since they need to be present, but are not yet used.
In each of these rows, the columns describe which cutoff sets are used for various SCF iterations. The cutoff sets themselves are provided later in the file, and dictate the level of analytic "corrections," the grid, and the non-default values of the gen section cutoff keywords (cut1, for example). The cutoff sets are described in more detail below. The columns reflect a scheme in which calculations are broken down into preliminary and final sets of iterations. The iterations from the beginning of the first SCF calculation in a run are considered to be part of the preliminary set, while the iterations from the end of the first SCF calculation, or from any subsequent set of SCF iterations, are considered to be part of the final set. For instance, for a solvation calculation, the SCF iterations for the analysis of the converged gas phase wavefunction are preliminary iterations followed by final iterations, while the SCF iterations for all subsequent SCF calculations, those including the solvent effects, are final iterations. Jaguar determines how many iterations are preliminary and how many are final for the initial SCF calculation.
The number in the first column in each of the five accuracy level lines dictates the cutoff set used for the first iteration in the preliminary sequence: if the number is a 1, the first cutoff set listed in the file is used; if it is a 5, the fifth is used, and so on. The number in the second column provides the cutoff set used for updates during the preliminary sequence of iterations. The third and fourth columns describe the cutoff sets used for the first and updating iterations in the final sequence, respectively. Finally, the last column dictates the cutoff sets used for non-SCF calculations, as for gradient calculations.
The first six lines of the default.cutoff file, which illustrate these points, are as follows:
The rest of the .cutoff file consists of the cutoff sets. Each set is specified by one line with four integers, sometimes followed by lines containing explicit cutoff keyword values, and ending with a blank line. The four integers represent the variables jcor and kcor (described below), the grid number, and the number of cutoff values to follow immediately below. The grid number should be 1 for the coarse grid, 2 for the medium grid, 3 for the fine grid, and 4 for the ultrafine grid, 5 for the charge grid, 6 for the gradient grid, 7 for the electron density cubic grid, 8 for the DFT medium grid, or 10 for the DFT gradient grid, where these grids are specified by the keywords gcoarse, gmedium, gfine, gufine, gcharge, ggrad, geldens, gdftmed, and gdftgrad. (The Grid and Dealiasing Function Keywords subsection of 8.6 contains more information on these keywords.)
The next lines specify each cutoff by number (e.g. 22 for the variable cut22) and value. Thus, the cutoff set:
means that jcor is 5, kcor is 2, the ultrafine grid is used, and that three cutoff values which differ from the defaults follow. The next three lines set the cutoff values cut21, cut22, and cut24. (If you need more information on cutoffs, please contact Schrödinger, Inc.)
The variables jcor and kcor determine what analytic corrections are calculated for a particular SCF iteration. The meanings of their possible values are shown in Table 9.5.1.
The variables a, b, and c in the table refer to distinct atoms.
To perform an all-analytic calculation, you can set the keyword nops in the gen section of the input file to 1. All-analytic calculations use the cutoff keyword values in the .cutoff file, but ignore the jcor, kcor, and pseudospectral grid information.
The Lewis file determines how van der Waals radii for calculations using the Jaguar solvation module are set according to chemical functional groups. By default, for neutral molecules in water, the program calculates a Lewis dot structure for the molecule or system, scans the Lewis file for radius information for each atom and sets radii for relevant atoms, then sets any radii not determined by the Lewis file according to the atomic section or the standard, default value. Settings for radii not included the Lewis file are described in Section 3.5 and section 8.9 and are listed in Table E near the back of the User's Guide. If you do not want the atomic radii that determine the dielectric continuum boundary to change according to the chemical environment of the atom, set the solvation keyword isurf to 0 in the gen section. Otherwise, Jaguar will alter some radii for neutral molecules by using the default.lewis file from the data directory, unless you specify your own .lewis file in a LEWISFILE line in the input file, as described in section 8.2.
If radii are being set according to a Lewis file, Jaguar first computes a Lewis dot structure for the input geometry to determine each atom's bonds and hybridization type. The element and chemical environment of each atom determine its atom type. When Jaguar reads the Lewis file, it sets the atom's van der Waals radius to the value dictated by the first atom type description in the Lewis file that matches that atom. For instance, if the atom were a methyl carbon and the first atom type description in the file was of a carbon bound to a hydrogen, the radius would be set to the radius matching that description, even if a later line in the Lewis file described a carbon bound to three hydrogens.
Atom types are determined by an atom's element and by any combination of the following other properties:
· bonding type, which is determined by the bond orders of the bond(s) the atom forms and the element(s) to which the atom is bonded
· hybridization type, which describes the hybridization and element of atoms to which the original atom is bonded
· aromaticity of the ring the atom is in, if any (benzene carbons, for instance, form an aromatic ring). An aromatic ring is defined here by the Huckel Rule: if the ring contains 4n + 2 pi electrons, where n is any non-negative integer, it is considered to be aromatic.
The Lewis file first determines the bonding types and hybridization types that will be recognized, then lists atomic radii for various atom types. The file contains different versions of this information for LMP2 calculations than it does for other wavefunction types. Therefore, the first non-blank line of the file should begin
(with any comment allowed after this string), indicating that the information following that line is for HF, DFT, or GVB wavefunctions. After all the information in the file for these calculations, the file should contain this line:
followed by information for LMP2 wavefunctions.
The bonding type information for HF, DFT, or GVB wavefunctions should follow the first line describing the calculation type. The first line of this information should begin
and the rest of the bonding type information should not contain any blank lines except the last line, which signals the end of bonding type information.
Bonding type information should be listed for each relevant element in turn. The information for the first atom should follow immediately after the "BONDING TYPE 01" label. The first character of the information for that atom should begin with the atom's atomic number. The following lines should describe up to five "groups" of bonds for that atom. Each group must begin with the word
(with no preceding spaces) and must contain information for bonds of bond orders 1, 2, and 3, with a comment line identifying each bond order. The group is simply a list of bonded atoms and bond orders for the element being described-for instance, Group 2 for carbon could describe C=C and C=O bonds by specifying that for bond order 2, Group 2 contains two elements with atom numbers 6 and 8. The first line under each bond order label must list the number of elements for that bond order and that group (2 for the C=C and C=O example); if this number is 0, the next line must list the atomic numbers for those elements (6 and 8 in the example).
Here is the beginning of a sample .lewis file illustrating a list of bonding type information for carbon, including some comments to further explain the file format:
The number of spaces at the beginning of the lines described above is irrelevant for all lines except the "Group" lines.
After all the groups have been specified for a particular atom, the file should contain a line containing three asterisks (`***') to indicate the next element's bonding types are about to be described (in the same format). After all desired bonding types are described for all appropriate elements, the bonding type information should end with a blank line.
The hybridization type information in the Lewis file includes up to five groups for each element described, where each group indicates a set of elements and hybridizations for those elements. The hybridization applies to the atom to which the original element is bonded. The information for hydrogen's first group, for instance, could list C (atomic number 6) with sp2 hybridization, allowing a later line in the Lewis file to set a particular radius for hydrogen atoms bonded to sp2 carbons.
The format of the hybridization type information is very similar to that of the bonding type information. The first line of this information (for HF, GVB, or DFT calculations) should begin
and the rest of the hybridization type information should not contain any blank lines except the last line, which signals the end of hybridization type information.
Hybridization type information should be listed for each relevant element in turn. The information for the first atom should follow immediately after the "HYBRIDIZATION TYPE 01" label. The first character of the information for that atom should begin with the atom's atomic number. The following lines should describe up to five hybridization "groups" for that atom. Each group must begin with the word
(with no preceding spaces). The group is simply a list of bonded atoms for all relevant hybridization types of those bonded atoms-for instance, Group 2 for hydrogen could describe hydrogens bonded to sp carbons by listing carbon's atomic number under an sp hybridization label. Because there is no default number of hybridizations described for each group (unlike for the bonding type information, where each group contained sets for three bond orders), the first line under each group label must begin with the number of hybridizations described for that group (after any number of spaces).
The next line dictates a hybridization for the bonded elements about to be described. Hybridization labels must start with five spaces, followed by one of the following character strings:
For each hybridization, the bonded elements with that hybridization are then listed in two lines, the first indicating the number of elements and the second indicating the elements themselves, as for the bonding type information.
Information for any following atoms should be preceded by a line with three asterisks, and a blank line indicates the end of the hybridization type information, as for the bonding type information.
The beginning of the hybridization information in a sample .lewis file, illustrating a list of hybridization type information for hydrogen and carbon, is shown below, with some comments to further explain the file format:
The number of spaces at the beginning of the lines described above is irrelevant for all lines except the "Group" lines and the hybridization labels.
After all desired hybridization types are described for all appropriate elements, the hybridization type information should end with a blank line.
The Lewis file can be used to make non-default choices for van der Waals radii of atoms in particular chemical environments, or even to reset the default radii for particular elements. After Jaguar's lewis program analyzes an input geometry's Lewis dot structure, it sets the atom's van der Waals radius to the value dictated by the first atom type description of element and chemical environment in the Lewis file that matches that atom with no contradiction. If no such matching description exists in the Lewis file, the atom is assigned the default radius for that element.
Atom type descriptions in the Lewis file should be preceded by a heading beginning
for information applying to HF, GVB, or DFT wavefunctions, or
for information for LMP2 wavefunctions. After that, each atom type description is listed. Blank lines are allowed in an atom type description list, and as long as some spacing exists between numbers and comments on each line, the number of spacing characters is irrelevant. However, keep in mind that the order of the atom type descriptions is important since the first matching description will always be used.
Each line describing an atom type has six integers, one real number, and an optional comment, in that order. The integers describe the atom type, while the real number sets the radius in Angstroms for that atom type. The six integers describe the following characteristics, in turn:
· aromaticity of that ring according to Huckel Rule (aromatic rings have 4n + 2 pi electrons, where n is a non-negative integer)
All six integer values and a corresponding radius value must always be listed in an atom type description line, and the atomic number must correspond to an actual element. However, any or all of the other five integer values can be set to -1, a wild card entry indicating that any value for that characteristic matches that atom type description. To reset a default radius for hydrogen, for instance, you could put the following line before any other descriptions of hydrogen atoms:
and the van der Waals radius for all hydrogen atoms would be set to 1.10 Å.
To describe the hybridization of the atom itself, the atom type description line's second integer should take on one of the values indicated in Table 9.6.1
The description of the atom's bonding type uses the groups listed in the bonding type information described in the subsection Describing Bonding Types in the Lewis File above (unless it is -1). Any positive integer for bonding type describes the number of bonds the atom has in each of the bonding type groups for its element and/or the number of all other bonds the atom has. A bonding type group describes elements of bonded atoms and orders of those bonds, as described in the subsection Describing Bonding Types in the Lewis File. The third integer in an atom type description line determines how many bonds the atom forms of each bonding type group g for an atom of a particular element, where g indicates the order of the bonding type groups listed for that element. The number of bonds from group g is indicated by the 10g digit in the integer.
For example, if g were 1 and the atom being described were carbon, g would correspond to the first bonding type group listed for carbon, and a bonding type integer value of 40 (4 x 101) would indicate that that carbon atom had four bonds from carbon's Group 1 bonding type information. If the Lewis file contained the bonding type information example provided in the subsection Describing Bonding Types in the Lewis File, which included the lines:
the integer value of 40 would describe a methane carbon. The same sample Lewis file information, whose key Group 2 information for carbon appears in these lines:
would mean that this radius information line
would describe a carbon atom (6) with one bond from carbon's Group 2 (a double bond to either C or O) and two bonds from carbon's Group 1 (single bonds to H), and would set such an atom's radius to 2.00 Å (unless another matching description preceded that line).
The rightmost digit in the integer describing bonding type specifies the number of bonds formed by the atom which are not of any of the forms described in the groups for that atom's bonding type information. A double or triple bond counts as one bond, not two or three, and lone pairs should not be included in the bond count.
The digits of the bonding type integer must describe all of an atom's bonding in order to match the atom information. For example, if the Lewis file described above contained no group for C-C bonds in the bonding type information, the integer "200" would only describe a carbon atom with one double bond to another C or O and no other bonds, while the integer "202" would adequately describe a carbon with one double bond to another carbon and two single bonds to other carbon atoms.
The fourth integer in an atom type description, which describes hybridization type, or the elements and hybridization of the atoms to which an atom is bound, works almost the same way as the integer describing bonding type. As it does for bonding types, the digit g places from the rightmost digit in the integer represents the gth group in the hybridization type information for that element (see the subsection Describing Hybridization Types in the Lewis File above for more information), while the rightmost digit specifies the number of bonds to elements and hybridization types that do not fit into any of the groups described for the element of the atom being evaluated. For example, suppose only one hybridization group were described for carbon in the sample Lewis file, as follows:
Then this atom type description line in a Lewis file would accurately match the middle carbon in methylethylene (H2C=CH-CH3):
as would the following line, which also contains the proper settings for the middle carbon's hybridization and bonding type:
As for the integer describing bonding type, the total of the digits in the fourth integer should be the same as the number of bonds (three for this example, remembering that the double bond counts as one bond)-that is, all bonds should be accounted for (unless, of course, the integer is -1).
The fifth and sixth integers describe the ring the atom is in, if any. If the fifth integer is a positive number n, it indicates that the atom description corresponds to an atom in a ring of size n. (For example, a benzene carbon is in a ring of size 6.) If the fifth number is a negative number -n, the description corresponds to an atom in a ring of size n or smaller, unless the fifth integer is -1, in which case the question of the atom's ring environment is ignored completely. The size n should not be more than 20.
The sixth integer indicates whether the description corresponds to an atom in an aromatic ring as defined by the Huckel Rule (4n + 2 electrons in ring, where n is a non-negative integer). If the sixth integer is 1, the description corresponds to an aromatic ring; if it is 0, the description corresponds to a non-aromatic ring; and if it is -1, the aromaticity of the ring is irrelevant. Note, however, that aromaticity is not evaluated if the fifth integer (describing ring size) is -1. To describe aromaticity without regard to ring size, you should generally set the fifth integer to -20 and the sixth to 1, corresponding to atoms in aromatic rings of size 20 or less.
The radius settings contained in Jaguar's default.lewis file are used for any relevant atoms in all default solvation calculations in water with Jaguar's solvation module, except for calculations on ions or on molecules containing atoms with atomic numbers greater than 18. By default, the program uses the first Lewis dot structure generated to evaluate the radii, and the solvation calculation also includes a correction term (the first shell correction factor) that depends on that Lewis dot structure. If the Lewis dot structure does not correspond to that desired for the molecule, the keyword lewstr should be changed to correspond to a better structure, as described in the gen section description in Chapter 8. To avoid using Lewis dot structures for either correction terms or radius settings, set the gen section keyword isurf to 0. To use Lewis dot structures to set radii but not for correction terms, isurf should be 0 but the keyword ivanset should be 1. All Lewis dot keywords are explained in section 8.6 under GVB and Lewis Dot Structure Keywords.
The radius settings in the file default.lewis, which appears in the standard data directory, were optimized for HF, GVB, and LMP2 solvation calculations in water with Jaguar's solvation module that included the default correction terms for the cavity and surface area. The molecules used for radius optimization were the molecules containing carbon, hydrogen, oxygen, nitrogen, and sulfur from reference [115]. All calculations used a 631G** basis set. Geometries were obtained from gas phase optimizations at the HF, GVB, and LMP2 levels. For both the geometry optimizations and the solvation energy calculations, the GVB and LMP2 treatment was restricted to heteroatom pairs.
|
Schrödinger, Inc. http://www.schrodinger.com Voice: (503) 299-1150 Fax: (503) 299-4532 help@schrodinger.com Last updated: Thu Oct 11 19:11:03 2001 |
| |
|
|
|