[molpro-user] DF-LCCSD(T) for open-shell

Mon Oct 15 14:49:58 BST 2012

Dear Prof. Werner,

Many thanks for your reply. I have been running some RHF/LUCCSD(T) calculations using molpro2012, as suggested. Most calns appear to work well. However, one caln crashed with the following:

 Reference energy:                   -788.07526886

 ITER.     SQ.NORM      CORR.ENERGY   TOTAL ENERGY   ENERGY CHANGE        DEN1      VAR(S)    VAR(P)  DIIS     TIME
   1      1.63280147    -2.13546247  -790.21073133    -2.13546247    -0.02719991  0.68D-04  0.96D-02  1  1    22.25
   2      1.65445020    -2.16492616  -790.24019502    -0.02946369    -0.00116298  0.88D-05  0.46D-03  2  2    28.12
   3      1.65443861    -2.16399928  -790.23926814     0.00092688    -0.00012816  0.13D-05  0.54D-04  3  3    34.02
   4      1.65470311    -2.16406672  -790.23933558    -0.00006744    -0.00000267  0.32D-06  0.11D-05  4  4    40.28
   5      1.65478847    -2.16412335  -790.23939221    -0.00005663    -0.00000026  0.62D-07  0.97D-07  5  5    46.32
   6      1.65482105    -2.16414878  -790.23941764    -0.00002543    -0.00000003  0.13D-07  0.87D-08  6  6    52.44
   7      1.65482823    -2.16415234  -790.23942121    -0.00000357     0.00000000  0.23D-08  0.69D-09  6  1    58.53
   8      1.65483059    -2.16415382  -790.23942268    -0.00000148     0.00000000  0.35D-09  0.56D-10  6  2    64.66

 Spin contamination <S**2-Sz**2-Sz>     0.00074012

 Norm of t1 vector:      0.04643300      S-energy:    -0.00376389      T1 diagnostic:  0.00061213
 Norm of t2 vector:      0.80788277      P-energy:    -2.16038994
                                         Alpha-Beta:  -1.57162110
                                         Alpha-Alpha: -0.29977811
                                         Beta-Beta:   -0.28899073

 MP2 energy of close+weak+distant pairs    -0.09337685

 Reference energy                    -788.075268861925
 Correlation energy                    -2.164153821781
 !RHF-LRMP2 STATE 1.1 Energy         -790.239422683706

 Starting LUCCSD calculation

 Reference energy:                   -788.07526886

 ITER.     SQ.NORM      CORR.ENERGY   TOTAL ENERGY   ENERGY CHANGE        DEN1      VAR(S)    VAR(P)  DIIS     TIME
   1             NaN            NaN            NaN            NaN            NaN  0.19D-01       NaN  1  1    77.89
   2             NaN            NaN            NaN            NaN            NaN       NaN       NaN  2  2    90.78
   3             NaN            NaN            NaN            NaN            NaN       NaN       NaN  3  3   101.62

The above was run with basis sets of VDZ quality. The same job with AVDZ quality was completed with no problem. Is there any reason for the above failure ?

Many thanks for your helps.

Regards,
Edmond.

Ps. My input is as follows:

***, TS (34-1 + OH => adduct1b) lccsd(t)/VDZ//m062x/6-31G**
! default frozen core
memory,125,m
gdirect
symmetry,nosym;orient,noorient

geometry={ang;
 C,0,-2.945728331,-0.0696541128,0.5020110341
 C,0,-1.9177463076,-0.9139236919,0.7236359301
 C,0,-0.5826413726,-0.8250331928,0.0449986445
 H,0,-3.8973818072,-0.2241753134,0.9923019475
 H,0,-2.8151818823,0.8772978846,-0.0107774801
 H,0,-2.080190265,-1.8203951198,1.3004242415
 H,0,-0.5784631414,-1.5647657764,-0.766844108
 C,0,0.6049847744,-1.1657764952,0.941351034
 O,0,-0.3936537473,0.3809024256,-0.7118800023
 N,0,-0.3100837309,1.5717274779,0.0232076837
 O,0,-0.1416508604,2.5211882008,-0.6782771613
 O,0,-0.4242507922,1.5057126962,1.2180045182
 H,0,0.4005533709,-2.1115375928,1.4479884817
 H,0,0.8242263276,-0.3934850335,1.6754421803
 O,0,1.7575605727,-1.4578740089,0.1433350527
 N,0,2.5510356285,-0.3573086408,-0.1745563949
 O,0,3.4681736594,-0.6455501116,-0.8810559482
 O,0,2.2464236038,0.7120552195,0.2894756015
 O,0,-3.4608460902,-1.1148490026,-1.2450436959
 H,0,-2.8692426091,-0.588796812,-1.8129885592}

gprint,orbitals;

basis={
default,vdz
set,mp2fit
default,vdz/mp2fit
set,jkfit
default,vdz/jkfit }

rhf
DF-LUCCSD(T)

---

eof

________________________________________
From: Hans-Joachim Werner [werner at theochem.uni-stuttgart.de]
Sent: 03 October 2012 12:18
To: Lee E.P.F.
Cc: molpro-user at molpro.net
Subject: Re: [molpro-user] DF-LCCSD(T) for open-shell

Yes, DF-LUCCSD(T) can treat high-spin open-shell molecules using RHF reference functions. However, it is still
somewhat experimental and therefore not yet fully described in the manual. Nevertheless, you can try to use it in 2012.1,
just as the closed-shell program.
An extension to DF-LUCCSD(T)-F12 is already functional in the development version and will be added to 2012.1 soon.
Best regards
Joachim Werner

Am 03.10.2012 um 13:05 schrieb "Lee E.P.F." <E.P.Lee at soton.ac.uk>:

> Dear Authors/Users,
>
> Am I correct that new features of molpro2012 includes the following:
>
> - Density fitted local coupled-cluster methods: DF-LCCSD(T), DF-LUCCSD(T), DF-LRPA
>
> Can DF-LUCCSD(T) deal with open-shell molecules ? From the online molpro2012 manual, it has only the following in 31 LOCAL CORRELATION TREATMENTS:
>
> The local correlation program of MOLPRO can currently perform closed-shell LMP2, LMP3, LMP4(SDTQ), LCISD, LQCISD(T), and LCCSD(T) calculations.
>
> This suggests that it cannot deal with open-shell species. Am I correct ?
>
> Regards,
> Edmond Lee
>
>
> _______________________________________________
> Molpro-user mailing list
> Molpro-user at molpro.net
> http://www.molpro.net/mailman/listinfo/molpro-user