SUPERLATTICE

SUPERLATTICE defines the relationship between the real-space unit vectors of the bulk and those of the surface.

Default: Detect automatically from bulk. See below for details.

Syntax examples:

SUPERLATTICE = c(2x2)
SUPERLATTICE = p(2x1)
SUPERLATTICE = (2x1)                    # same as p(2x1)
SUPERLATTICE = (sqrt(2) x sqrt(2))R45   # same as M = 1 1, -1 1
SUPERLATTICE = (2*sqrt(2) x sqrt(2))R45
SUPERLATTICE M = 1 1, -1 1              # same as (sqrt(2) x sqrt(2))R45

If SUPERLATTICE is not defined, ViPErLEED will perform a search for possible unit cell vectors for the bulk (as defined by LAYER_CUTS and N_BULK_LAYERS), and choose two vectors that minimize the unit cell area and circumference. The resulting SUPERLATTICE matrix will then be further optimized to bring the bulk unit cell to its highest symmetry form (preserving the area). Note that for many systems, multiple choices of the bulk unit cell are possible, but result in different SUPERLATTICE definitions; this choice will be made arbitrarily if SUPERLATTICE is not actively defined.

Even when SUPERLATTICE is defined, an automatic search for the best bulk unit cell will run. If a unit cell with a smaller area is found, the program will warn and stop. If required, halting can be suppressed using the HALTING parameter.

Note

For Wood notation, use only positive float values. The separator in the brackets should be the letter ‘x’, as in the examples.
In matrix notation use the convention 'M = m11 m12, m21 m22’, with a_surf = m₁₁·a_bulk + m₁₂·b_bulk, and b_surf = m₂₁·a_bulk + m₂₂·b_bulk, i.e., the bulk row vectors a_bulk and b_bulk get multiplied with M from the left to obtain the surface row vectors a_surf and b_surf respectively. In matrix notation all elements should be signed integers (floats will also be accepted, but this is not recommended).

Note

Make sure to get the relation between the structure in the POSCAR file and the superlattice right. For example, in c(4x8) “4” must refer to the a and “8” to the b lattice vector of the POSCAR, not reverse, unless SYMMETRY_CELL_TRANSFORM specifies that the superstructure cell differs from the cell given in the POSCAR file.

When using Wood notation, the regular expression interpreting the input is:

(?P<type>[PCpc]*)\s*\(\s*(?P<g1>.+)\s*[xX]\s*(?P<g2>.+)\s*\)\s*[rR]*\s*(?P<alpha>[\d.]*)

The expressions g1 and g2 in (g1 x g2) are interpreted afterwards; here, only sqrt(value) and multiplicative expressions are parsed, so '2*sqrt(2)’ will be read correctly, but e.g. '(1/3)*sqrt(3)’ will yield an error. For these cases, using a matrix is usually better.