What is the correct input for the GEOM_OPT_MODE

Questions
#1

I am trying to calculate a TS using a guess from fsm. The exact hessian is calculated with freq and scratch file save. Then I examined the normal modes which seemed the second mode is more relevant to the TS I am trying to compute. This is my input sections:
$rem
JOBTYPE sp
xc_grid = 3
geom_opt_driver optimize
SYM_IGNORE TRUE
SYMmetry FALSE
unrestricted true
BASIS = def2-svp
METHOD = WB97X-D
SCF_GUESS_MIX = 5
SCF_CONVERGENCE = 7
MEM_TOTAL = 50000
MAX_SCF_CYCLES 250
SCF_ALGORITHM diis_gdm
$end

@@@
$molecule
read
$end

$rem
JOBTYPE ts
xc_grid = 3
geom_opt_driver optimize
SYM_IGNORE TRUE
SYMmetry FALSE
unrestricted true
BASIS = def2-svp
METHOD = WB97X-D
SCF_GUESS = read
SCF_CONVERGENCE = 7
MEM_TOTAL = 50000
GEOM_OPT_MAX_CYCLES = 1500
GEOM_OPT_DMAX = 30
MAX_SCF_CYCLES 250
SCF_ALGORITHM diis
GEOM_opt_hessian = read
GEOM_OPT_MODE = 2
$end

$geom_opt
STEP_SEARCH_ALGORITHM = EFA
$end

But this is the error message:

Transition state search - taking P-RFO step

Q-Chem fatal error occurred in module progman/main.C, line 179:

Problem running Q-Chem: mode != 0

Please submit a crash report at Q-Chem Crash Reporter

Is it not allowing me to do a mode selection for hessian with imaginary frequency or is there something you guys know I should do?

#2

First of all, if you’re going to read a Hessian in the 2nd job then you need to compute one in the first job, so set JOBTYPE=FREQ in the first job.

Second, based on my experiments it looks like this feature is only available in the new optimizer (which is the default, where available) so delete the GEOM_OPT_DRIVER line in $rem. Most (perhaps all) of the $geom_opt variables are only available in the new driver.

With those changes, I did get a similar job to run (with a small molecule and basis set).