# Grids¶

There are three possible grid topologies in gimic: ‘std (or base)’ which is 3-dimensional (although the extend in one direction may be zero), ‘bond’ which is 2-dimensional, and ‘file’ which is an arbitrary set of points read from a file. ‘std’ grids are especially useful for visualisations of an entire current vector field. ‘bond’ grid are used to integrate the current that flows through the grid.

For an ‘std’ or ‘bond’ grid one can specify the spacing (type) of points in each dimension. The choices are ‘even’ for an equidistant grid, ‘gauss’ for a Gauss grid and ‘lobatto’ for a Gauss-Lobatto grid. The recommended choices are ‘even’ for ‘std’ and ‘gauss’ for ‘bond’. When a quadrature grid is specified the order of the quadrature must also be specified with the ’gauss_order’ keyword. The number of grid points in each direction is specified either explicitly using either of the array keywords ’grid_points’ or ’spacing’. If the chosen grid is not a simple even spaced grid, the actual number of grid points will be adjusted upwards to fit the requirements of the chosen quadrature.

The shape of the grid can also be modified by the ’radius’ key, which specifies a cutoff radius. This can be useful for integration. Sometimes it’s practical to be able to specify a grid relative to a well know starting point. The ’rotation’ keyword specifies Euler angles for rotation according to the x->y->z convention. Note that the magnetic field is not rotated, unless it is specified with ’magnet_axis=i,j or k’.

GIMIC automatically outputs a number of .xyz files containing dummy points to show how the grids defined actually are laid out in space.

## Basic grids¶

The ’std’ grid is defined by giving an ’origin’ and two orthogonal basis vectors ’ivec’ and ’jvec’ which define a plane. The third axis is determined from \vec k=\vec i\times\vec j. The array ’lengths’ specifies the grid dimensions in each direction.

### grid(std)¶

type=even
origin=-8.0, -8.0, 0.0
Origin of grid

ivec=1.0, 0.0, 0.0
Basis vector i

jvec=0.0, 1.0, 0.0
Basis vector j ( k = i x j )

lengths=16.0, 16.0, 0.0
Lenthts of (i,j,k)

spacing=0.5, 0.5, 0.5
Spacing of points on grid (i,j,k)

grid\_points=50, 50, 0
Number of gridpoints on grid (i,j,k)

rotation=0.0, 0.0, 0.0
Rotation of (i,j,k) -> (i’,j’,k’) in degrees. Given as Euler angles
in the x->y->z convention.


## Bond grids¶

The ’bond’ type grids define a plane through a bond, or any other defined vector. The plane is orthogonal to the vector defining the bond. The bond can be specified either by giving two atom indices, ’bond=[1,2]’, or by specifying a pair of coordinates, ’coord1’ and ’coord2’. The position of the grid between two atoms is determined by the ’distance’ key, which specifies the distance from atom 1 towards atom 2. For analysing dia- and paramagnetic contributions, the positive direction of the bond is taken to be from atom 1 towards atom 2. Since one vector is not enough to uniquely defining the coordinate system (rotations around the bond are arbitrary), a fixpoint must be specified using either the ’fixpoint’ atom index or the ’fixcoord’ keyword. This triple of coordinates is also used to fix the direction of the magnetic field when the ’magnet_axis=T’ is used.

The shape and size of the bond grid is specified by the keywords ‘width’ and ‘height’. Each of the four components is relative to a point on the line connecting the two reference atoms/coordinates. Typically (but not necessarily) the first component of both ranges is negative and the second component is positive.

width=[-1.5, 5.0]
height=[-5.0, 5.0]


### grid(bond):¶

type=gauss|lobatto
Use uneven distribution of grid points for quadrature
bond=1,2
Atom indices for bond specification
fixpoint=5
Atom index to use for fixing the magnetic field and grid orientation
coord1=0.0, 0.0, 3.14
Coordinate of atom 1
coord2=0.0, 0.0, -3.14
Coordinate of atom 2
fixcoord=0.0, 0.0, 0.0
Fixation coordinate
distance=1.5
Place grid ’distance’ between atoms 1 and atom 2
gauss_order=9