SCF convergence issue in NEO calculations

Questions
#1

I am running SP NEO-DFT calculations with Q-chem 6.2. I am facing SCF convergence issue. Following are my input and (the end of) output.
Input:
$molecule
1 1
C 1.22127 0.70510 0.00000
C 1.22127 -0.70510 0.00000
C 0.00000 1.41021 0.00000
C 2.44340 1.41070 0.00000
C -0.00000 -1.41021 -0.00000
C -1.22127 0.70510 -0.00000
C 2.44340 -1.41070 0.00000
C 0.00000 2.82140 0.00000
C 3.65444 0.70022 0.00000
C 2.43363 2.81472 0.00000
C 3.65444 -0.70022 0.00000
C -1.22127 -0.70510 -0.00000
C 1.22081 3.51495 0.00000
C -0.00000 -2.82140 -0.00000
C -2.44340 1.41070 -0.00000
C 2.43363 -2.81472 0.00000
C -1.22081 3.51495 -0.00000
C -2.44340 -1.41070 -0.00000
C 1.22081 -3.51495 0.00000
C -2.43363 2.81472 -0.00000
C -1.22081 -3.51495 -0.00000
C -3.65444 0.70022 -0.00000
C -2.43363 -2.81472 -0.00000
C -3.65444 -0.70022 -0.00000
H 4.59989 1.22945 0.00000
H 3.36468 3.36890 0.00000
H 4.59989 -1.22945 0.00000
H 1.23521 4.59835 0.00000
H 3.36468 -3.36890 0.00000
H -1.23521 4.59835 -0.00000
H 1.23521 -4.59835 0.00000
H -3.36468 3.36890 -0.00000
H -1.23521 -4.59835 -0.00000
H -4.59989 1.22945 -0.00000
H -3.36468 -3.36890 -0.00000
H -4.59989 -1.22945 -0.00000
H 0 0 -0.2
$end

$rem
BASIS = def2-TZVP
GUI = 2
JOBTYPE = SP
MAX_SCF_CYCLES = 100000
METHOD = B3LYP
NEO = true
NEO_EPC = epc172
NEO_N_SCF_CONVERGENCE = 5
NEO_VPP = 0
SCF_ALGORITHM = GDM
SCF_CONVERGENCE = 5
SYM_IGNORE = 1
XC_GRID = 000099000302
$end

$neo_basis
H 37
S 1 1.000000
4.0 1.0
S 1 1.000000
8.0 1.0
S 1 1.000000
16.0 1.0
S 1 1.000000
32.0 1.0
P 1 1.000000
4.0 1.0
P 1 1.000000
8.0 1.0
P 1 1.000000
16.0 1.0
P 1 1.000000
32.0 1.0
D 1 1.000000
4.0 1.0
D 1 1.000000
8.0 1.0
D 1 1.000000
16.0 1.0
D 1 1.000000
32.0 1.0

$end

Output:
47021 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47007 failed line searches. Resetting
47022 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47008 failed line searches. Resetting
47023 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47009 failed line searches. Resetting
47024 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47010 failed line searches. Resetting
47025 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47011 failed line searches. Resetting
47026 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47012 failed line searches. Resetting
47027 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47013 failed line searches. Resetting
47028 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47014 failed line searches. Resetting
47029 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47015 failed line searches. Resetting
47030 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47016 failed line searches. Resetting
47031 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47017 failed line searches. Resetting
47032 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47018 failed line searches. Resetting
47033 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47019 failed line searches. Resetting
47034 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47020 failed line searches. Resetting
47035 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47021 failed line searches. Resetting
47036 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47022 failed line searches. Resetting
47037 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47023 failed line searches. Resetting
47038 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47024 failed line searches. Resetting
47039 -922.3728114280 1.57e-04 Line search: overstep
GDM:: 47025 failed line searches. Resetting

#2

Does the output you posted correspond to the initial restricted SCF? I think setting THRESH to 14 in your REM should fix this. You do not need to set MAX_SCF_CYCLES to 100000 in your input. MAX_SCF_CYCLES = 100 should suffice.

#3

Thanks, setting Thresh value already work for me.

#4

For medium-sized molecules (and definitely for large ones), I recommend THRESH=12 or tighter. See here:
https://doi.org/10.1021/acs.jpca.4c00283