Energy minimization

When LTRD is used, the gradient norm will be minimized using the full Newton algorithm. Additionally, a full Hessian is computed at each step making LTRD the most reliable (and expensive) method for gradient minimization. LTRD can also be used with OPTMAX=0 to build and diagonalize a Hessian matrix in a given subset of internal coordinates. Thermodynamic quantities may also be computed from the results of an LTRD calculation. See Chapter 4, Computational Procedures for a more complete discussion about the LTRD and NEWTON quadratic optimization methods.

Simulated annealing

The use of the regular AMPAC keyword LTRD along with GANNEAL, TSANNEAL, or MANNEAL causes the slower but more reliable LTRD algorithm to be used in the critical point searching procedure. See Chapter 13, Simulated Annealing for a more complete discussion about the effect of using LTRD or NEWTON in simulated annealing calculations.

Abbreviation:

none

Requires:

none

Input File (optimization/opt_ltrd.dat):

  am1 rhf singlet bonds t=auto ltrd grad inertial
Conrotary cyclization TS
LTRD on TS, INERTIAL
 C              0.000000  0    0.000000  0    0.000000  0    0    0    0
 C              1.368591  1    0.000000  0    0.000000  0    1    0    0
 C              1.467760  1  101.846287  1    0.000000  0    2    1    0
 C              1.467760  1  105.846287  1  -18.651781  1    1    2    3
 H              1.088208  1  133.192008  1  157.464541  1    1    2    3
 H              1.088208  1  125.192008  1  145.464541  1    2    1    4
 H              1.096007  1  117.078638  1  133.669157  1    4    1    2
 H              1.095904  1  124.139890  1  139.767746  1    3    2    1
 H              1.098257  1  124.343753  1  -63.184369  1    4    1    2
 H              1.098289  1  117.265822  1  -67.054326  1    3    2    1
 0              0.000000  0    0.000000  0    0.000000  0    0    0    0