Page 1 of 1

NEB to Dimer

Posted: Thu Nov 08, 2012 4:00 pm
by eduard
Dear Graeme,

I did a calculation with cNEB which I converged it till the forces on the images were of:

0 0.045449 -844.235414 0.000000
1 0.029524 -842.154981 2.080433
2 0.041430 -841.204198 3.031216
3 0.024248 -841.924670 2.310744
4 0.246999 -844.234075 0.001339

As I want to calculate the prefactor for the reaction I need smaller forces. For computational reasons I thought that it might be efficient to relax the trnasition state from this level with the Dimer method.

I run first nebresults.pl which gave me:

==============================================
Unziping the OUTCARs ... done
Do nebbarrier.pl ; nebspline.pl
Do nebef.pl
Do nebmovie.pl
Do nebjmovie.pl
Do nebconverge.pl
Zipping the OUTCARs again ... done

Forces and Energy:
0 0.045449 -844.235414 0.000000
1 0.029524 -842.154981 2.080433
2 0.041430 -841.204198 3.031216
3 0.024248 -841.924670 2.310744
4 0.246999 -844.234075 0.001339

Extrema 1 found at image 0.004137 with energy: -0.000050
Extrema 2 found at image 1.754931 with energy: 3.047522
Extrema 3 found at image 2.019890 with energy: 3.030753
Extrema 4 found at image 2.412900 with energy: 3.088996
Extrema 5 found at image 3.998680 with energy: 0.001333

=====================================================

upon which I run neb2dim.pl with the outcome:

=====================================================

FORMING DIMER BETWEEN IMAGES 2 and 3

FOR DIMER, REMEMBER TO SET IN THE INCAR:

ICHAIN = 2
IOPT = 2
IBRION = 3
POTIM = 0.0
EDIFF = 1E-7

DdR (DEFAULT = 5E-3)
DRotMax (DEFAULT = 4)
DFNMax (DEFAULT = 1.0)
DFNMin (DEFAULT = 0.01)

========================================

I started a dimer calculation in the dim directory using the INCAR:

ISTART = 0
ICHARG = 2
ENCUT = 550

#! SPRING=-5
#! IMAGES=3
#! ICHAIN=0
#! LCLIMB=.TRUE.

PREC = Accurate
ISPIN = 1
ISIF = 2
ISYM = 0

EDIFF = 1e-7
NELM = 90
NELMIN = 5
ALGO = Normal


EDIFFG = -0.010000
NSW = 32
POTIM = 0.0
IBRION = 3
IOPT = 2

ICHAIN=2
DdR=0.005
DRotMax=4
DFNMin=0.01
DFNMax=1.0

ISMEAR = 0
SIGMA = 0.11


LCHARG = FALSE
LWAVE = FALSE
LREAL = Auto



LPLANE = TRUE
NPAR = 32

To my surprise the Dimer calculation sarted pretty bad.


Step Force Torque Energy Curvature Angle
1 1.50979 16.31105 -840.91766 13.44073 26.49398
1 1.50979 4.81521 -840.91766 5.28904 15.77359
1 1.50979 4.28587 -840.91766 4.07293 9.74088
1 1.50979 3.40436 -840.91766 3.39773 7.47284
2 1.78227 3.91929 -840.84649 3.51438 8.01444
2 1.78227 3.28916 -840.84649 3.00863 8.28268
2 1.78227 3.20031 -840.84649 2.57524 5.30350
2 1.78227 2.35889 -840.84649 2.37935 7.52230
3 2.23477 3.15808 -840.70430 2.09134 5.59410
3 2.23477 2.27771 -840.70430 1.81796 6.63904
3 2.23477 2.54722 -840.70430 1.53788 3.84544
3 2.23477 1.68002 -840.70430 1.38536 5.94463


As you can see the curvature is positive althought according to NEB I should be close to a TS. The first iterations also do not give negative curvature ( althought, indeed, I notice that the positive value do decrease).
I am quite worry about the bad start of the DIMER ? Was the Cl-NEB TS not a TS ? Was the force too large in NEB ? 0.04 eV/A looked reasonable to me. I am not sure if I did some mistake when going from NEB -> DIMER. I would really appreciate your advice.

Few more informations about the system ( particlarly about the NEB run ) : I used, as previously seen, 3 images to span the path. The distances between the images are around 2 A:

1 intermediate image: NEB: distance to prev, next image, angle between 2.387168 2.389744 140.820767
2 intermediate image: NEB: distance to prev, next image, angle between 2.389744 2.380375 78.934033
3 intermediate image: NEB: distance to prev, next image, angle between 2.380375 2.376425 149.137467

As you notice the ath is not liniar ( the angle at image 2 - where the saddle points reside - is 80 degree ). Ideealy more than 3 images should have been used, but due to computational reasons I had to limit myself to these. In spite of this the relaxations went well, and the final structures provided by Cl-NEB in 01/ 02/ 03 directory looked reasonable. I wonder now: could it be that neb2dim.pl interpolate between the images 2 and 3, and because of the large distance and sharp angle push the TS away ?

I would really appreacite any advice from you, to know if I should continue the Dimer calculation, and how.

Thank you very much for your time !
Eduard

Re: NEB to Dimer

Posted: Thu Nov 08, 2012 4:04 pm
by eduard
PS. I am sorry for the type spelling. I spent all the night doing calculations and I am bit tired.

Re: NEB to Dimer

Posted: Fri Nov 09, 2012 6:28 pm
by eduard
Follow up data:

I started just for test the followings:

(test calculation 1) : - use the CONTCAR from NEB ( according to energies, this would be intermediate image 2) and MODECAR generated between the initial and final states ( with modemake.pl ). This is different from the approach posted above when I have used the outputs ( POSCAR, MODECAR) from neb2dim.pl.
With the CONTCAR and MODECAR so build, the DIMCAR after the first iterations looks like:

Step Force Torque Energy Curvature Angle
1 0.26905 2.42332 -841.15555 1.28876 8.66569
1 0.26905 2.06991 -841.15555 1.03492 7.39347
1 0.26905 1.84444 -841.15555 0.85121 7.92913

As you notice the curvature is still positive, but the value is not as large as before ( before was 13). Why ? Have I done something wrong before ?

(test calculation 2) : - use the CONTCAR from NEB ( image 2, as before) and MODECAR generated between the intermediate states 1 and 3 ( the states around the CONTCAR)
The DIMCAR is after first iterations:

Step Force Torque Energy Curvature Angle
1 0.26905 12.53783 -841.15555 8.00840 19.02101
1 0.26905 4.53157 -841.15555 3.70677 18.95893
1 0.26905 4.70282 -841.15555 2.39826 9.32045

The curvature is positive, and larger then before. Given that the POSCAR is the same, the choice of MODECAR ( between which states is generated) seems to be quite important.

(test calculation 3) : - use he same CONTCAR as in the above 2 cases, and MODECAR generated between the TS and the initial state.
DIMCAR:
Step Force Torque Energy Curvature Angle
1 0.26905 5.18074 -841.15555 2.73314 10.44582
1 0.26905 2.71375 -841.15555 1.84362 8.41597
1 0.26905 2.28075 -841.15555 1.50052 6.47738

Having a look over these 3 tests, it looks like the way the MODECAR is generated ( between which states: initial-final, 1-3, TS-initial ) influence quite a lot the way the relaxation starts. I am not sure why is so, because all these states ( inital, 1, TS=2, 3, final) are on the same path, isn't it ? So, why would it matter which one you chose ? Anyway, the best choice - according to these tests - is to build the MODECAR between the initial and final states.

Why does nebmake.pl provide poorer starting files ? The value for the curvature obtained after running nebmake.pl outputs ( POSCAR, MODECAR) is the worst.

Re: NEB to Dimer

Posted: Sat Nov 10, 2012 12:57 am
by eduard
Comparative results :

-> POSCAR,MODECAR from neb2dim.pl:

Step Force Torque Energy Curvature Angle
1 1.50979 16.31105 -840.91766 13.44073 26.49398
1 1.50979 4.81521 -840.91766 5.28904 15.77359
1 1.50979 4.28587 -840.91766 4.07293 9.74088
1 1.50979 3.40436 -840.91766 3.39773 7.47284
2 1.78227 3.91929 -840.84649 3.51438 8.01444
2 1.78227 3.28916 -840.84649 3.00863 8.28268
2 1.78227 3.20031 -840.84649 2.57524 5.30350
2 1.78227 2.35889 -840.84649 2.37935 7.52230
3 2.23477 3.15808 -840.70430 2.09134 5.59410
3 2.23477 2.27771 -840.70430 1.81796 6.63904
3 2.23477 2.54722 -840.70430 1.53788 3.84544
3 2.23477 1.68002 -840.70430 1.38536 5.94463

-> POSCAR = CONTCAR from image 2, DIMCAR obtained by using modemake.pl between IS ( initial state ) and FS ( final state):

Step Force Torque Energy Curvature Angl
1 0.26905 2.42332 -841.15555 1.28876 8.66569
1 0.26905 2.06991 -841.15555 1.03492 7.39347
1 0.26905 1.84444 -841.15555 0.85121 7.92913
1 0.26905 1.73114 -841.15555 0.69061 5.23087
2 0.41726 1.71855 -841.13655 0.50573 5.27606
2 0.41726 1.43809 -841.13655 0.50876 3.95697
2 0.41726 1.30036 -841.13655 0.40300 2.66405
2 0.41726 0.97147 -841.13655 0.34132 2.04778
3 0.57024 1.24058 -841.10143 0.33686 4.91774
3 0.57024 1.63187 -841.10143 0.36442 3.35126
3 0.57024 1.05844 -841.10143 0.28250 2.39760
3 0.57024 0.88609 -841.10143 0.21678 1.59988

-> POSCAR = CONTCAR from image 2, DIMCAR obtained by using modemake.pl between image 1 CONTCAR and image 3 CONTCAR :

Step Force Torque Energy Curvature Angl
1 0.26905 12.53783 -841.15555 8.00840 19.02101
1 0.26905 4.53157 -841.15555 3.70677 18.95893
1 0.26905 4.70282 -841.15555 2.39826 9.32045
1 0.26905 2.95664 -841.15555 1.66892 6.70682
2 0.67270 2.65004 -841.12359 1.31949 4.53322
2 0.67270 1.98271 -841.12359 1.15331 6.18485
2 0.67270 2.37317 -841.12359 0.91439 3.32267
2 0.67270 1.43193 -841.12359 0.81229 5.85146
3 1.07037 2.25075 -841.06665 0.69715 3.05628
3 1.07037 1.21584 -841.06665 0.61112 5.47035
3 1.07037 1.95141 -841.06665 0.49753 2.39803
3 1.07037 0.97615 -841.06665 0.46093 3.82337

-> POSCAR = CONTCAR from image 2, DIMCAR obtained by using modemake.pl between POSCAR ( image 2 CONTCAR ) and initial state (IS):

Step Force Torque Energy Curvature Angl
1 0.26905 5.18074 -841.15555 2.73314 10.44582
1 0.26905 2.71375 -841.15555 1.84362 8.41597
1 0.26905 2.28075 -841.15555 1.50052 6.47738
1 0.26905 2.05682 -841.15555 1.29150 3.67976
2 0.42067 1.53036 -841.12356 0.54498 2.98685
2 0.42067 1.10487 -841.12356 0.51514 2.39828
2 0.42067 1.05590 -841.12356 0.46339 1.84711
2 0.42067 0.96909 -841.12356 0.42881 1.67254
3 0.54396 1.19259 -841.07396 -0.17811 2.83707
3 0.54396 1.05361 -841.07396 -0.17275 1.93560
3 0.54396 0.86346 -841.07396 -0.21189 1.58758
4 0.54912 1.08940 -841.07532 -0.00764 2.07438

Questions :

(1) why when starting with neb2dim.pl outputs the initial data seems to be the poorest ( largest curvature = 13, largest force=1.5, largest torque=16, etc) ?
(2) althought I use the same POSCAR in the test cases 2,3,4, the use of a different MODECAR leads to relatively different initial conditons ( again, in terms of curvature, force, etc). I was a slightly surprised because we speak about states found on the same MEP and between which MODECAR is generated.
(3)- more technical - why in all the cases are 4 steps per iteration: e.g. 4 steps for iteration 1 in MODECAR, 4 steps for iteration 2, 4 steps for iterations 3 - BUT in the last tested case there are ONLY 3 steps for iteration 3, and then the iteration 4 starts ? I use the same INCAR always. Why does the last test case passes faster to iteration 4 ? ( it is also the only case when I get a negative curvature).
(4) I would have liked to run further, more than 1 test case. But, unfortunatelly, computationally I do not have the time. I think the best would be to continue the calculation which has already reach the positive curvature ? Or ... ?

Dear Graeme, I would really appreciate your support.

Wishing you all a great evening,
Eduard

Re: NEB to Dimer

Posted: Mon Nov 12, 2012 2:45 am
by graeme
Goodness, this is a lot of information. Also, you have done some good tests, but somehow I don't think you have the source of the problem yet.

One thing that I would start with is trying to minimize image 2. Your NEB spline interpolation:

Extrema 1 found at image 0.004137 with energy: -0.000050
Extrema 2 found at image 1.754931 with energy: 3.047522
Extrema 3 found at image 2.019890 with energy: 3.030753
Extrema 4 found at image 2.412900 with energy: 3.088996
Extrema 5 found at image 3.998680 with energy: 0.001333

shows a minimum at image 2. This is consistent with the data from your NEB calculation:

1 intermediate image: NEB: distance to prev, next image, angle between 2.387168 2.389744 140.820767
2 intermediate image: NEB: distance to prev, next image, angle between 2.389744 2.380375 78.934033
3 intermediate image: NEB: distance to prev, next image, angle between 2.380375 2.376425 149.137467

showing that there is a sharp angle at image 2. My guess is that the path is going into and out-of the minimum at image 2 along the same vector, which is giving rise to the sharp angle.

Then, with the neb2dim.pl script, it is interpolating halfway between image 2 and 3 in the estimate of the saddle (a long way from image 2), and since the path curves so much, the tangent is ill-defined and the dimer guess is far from the saddle. This is also consistent with your high initial dimer energy (~1eV higher than image 2) and the initial positive mode. It also explains the positive mode when you start from image 2.

If you minimize image 2 and find that it is indeed a local intermediate minimum, then run a new NEB from image 0 to image 2 with about 4 images and use the climbing image to find a saddle. If you find one, I expect a smoother path and a good initial guess for the dimer. Equivalently, you could also calculate the entire path with 7 or 9 images, again with the climbing image to find a saddle (or more likely two, one on either side of image 2).

Also, look at the geometries along the path and see if the mechanism makes sense in terms of this explanation.

One last thing: make sure that ISYM=0 in your calculations.

Good luck. It would be interesting to know how this turns out. And if you don't get to the bottom of it, you can post your initial band and we'll take a look.

Re: NEB to Dimer

Posted: Wed Nov 14, 2012 4:16 pm
by eduard
Dear Graeme,

I relaxed the 02 image, using 2 approaches:

(a) use as POSCAR the original configuration of the image 2 ( the one with which the NEB started, generated by the interpolation script). Following carefull unconstrained relaxation, I found that the the 02 image returns to the initial state ( 00 ). This would suggest that 02 is not a minimum ( if another saddle point would be present between 02 and 00, then probabbly this would stop the image 02 to fall into 00)

(b) use as POSCAR the LAST configuration obtained by the NEB ( the last CONTCAR from NEB). In this case, following the unconstrained relaxation, nothing happens. Moore precisely, following 7-8 relaxation steps. the VASP code stops finding convergence. I am not sure what this suggests. That 02 IS a minimum ? A saddle point ? And how does this fit with what I obtained at point (a) ?

I am really a bit confused. If you have any time, please, have a look over my band. I have just submitted now an email on the address henkelman at mail dot utexas dot edu

Thank you very, very much for all your time and help !

Re: NEB to Dimer

Posted: Thu Nov 15, 2012 2:58 pm
by graeme
A 7-image NEB converges nicely; here is the MEP output:

0 0.021689 -842.010700 0.000000
1 0.013746 -841.449500 0.561200
2 0.018819 -840.310600 1.700100
3 0.023155 -839.024000 2.986700
4 0.020080 -838.603700 3.407000
5 0.021659 -838.965400 3.045300
6 0.014454 -840.205200 1.805500
7 0.017117 -841.474800 0.535900
8 0.021689 -842.010700 0.000000

Also, a dimer calculation starting from an unconverged NEB converges smoothly as well:

Step Force Torque Energy Curvature Angle
1 37.74639 15.97367 -823.30752 10.40136 13.05355
1 37.74639 16.74381 -823.30752 6.71989 19.14399
1 37.74639 10.81138 -823.30752 3.10640 68.07784
1 37.74639 31.20113 -823.30752 -10.33356 57.18000
2 17.84786 15.67256 -828.07430 -8.72449 30.06326
2 17.84786 13.68475 -828.07430 -17.94333 16.15749
2 17.84786 9.80441 -828.07430 -21.50891 14.26648
2 17.84786 5.39798 -828.07430 -23.11027 7.81125
3 13.09341 3.38040 -831.03216 -17.39297 3.16885
3 13.09341 1.60417 -831.03216 -17.68159 1.83065
3 13.09341 1.08600 -831.03216 -17.78907 1.62588
3 13.09341 0.94010 -831.03216 -17.67082 0.79265
4 9.60257 4.15207 -833.24545 -12.93592 5.60081
4 9.60257 2.36302 -833.24545 -13.42569 2.80119
4 9.60257 1.26125 -833.24545 -13.49154 1.65553
4 9.60257 1.15429 -833.24545 -13.59001 1.68224
5 6.26520 1.82224 -834.79061 -8.78900 2.45722
5 6.26520 1.25615 -834.79061 -8.86844 1.36438
5 6.26520 0.77367 -834.79061 -8.99951 0.95639
6 4.21097 1.51310 -835.78890 -7.48993 3.19291
6 4.21097 1.09033 -835.78890 -7.61752 1.27771
6 4.21097 0.89603 -835.78890 -7.63450 1.10449
7 3.12534 1.53407 -836.48423 -5.48002 2.61113
7 3.12534 1.13942 -836.48423 -5.57638 1.39086
7 3.12534 0.87045 -836.48423 -5.62590 1.29625
8 2.81701 1.47728 -836.97042 -4.56701 2.62227
8 2.81701 1.06860 -836.97042 -4.64908 1.33953
8 2.81701 0.81599 -836.97042 -4.73375 1.16882
9 3.50327 1.31960 -837.14455 -3.45497 2.09994
9 3.50327 1.08980 -837.14455 -3.61006 1.31661
9 3.50327 0.85103 -837.14455 -3.66484 1.38407
10 2.60734 1.28488 -837.48623 -2.85404 2.27032
10 2.60734 0.81736 -837.48623 -2.84772 1.16385
11 1.53042 1.01348 -837.56635 -2.96772 2.57516
11 1.53042 0.99752 -837.56635 -2.97004 1.27218
12 1.57250 1.40444 -837.77390 -2.95347 2.83990
12 1.57250 1.12711 -837.77390 -2.91984 1.76449
12 1.57250 1.06583 -837.77390 -2.94139 1.46351
12 1.57250 0.90739 -837.77390 -2.93929 1.20099
13 1.67827 1.32498 -837.85777 -2.48832 2.55044
13 1.67827 0.80228 -837.85777 -2.50544 1.49306
14 1.89896 1.15382 -837.97431 -2.13233 2.56784
14 1.89896 1.03763 -837.97431 -2.14133 1.68453
14 1.89896 1.08539 -837.97431 -2.15928 1.56471
14 1.89896 0.94886 -837.97431 -2.23656 1.33103
15 0.81261 1.13680 -838.02444 -1.34601 2.26373
15 0.81261 0.91891 -838.02444 -1.39389 1.35624
16 1.89715 0.87987 -838.07107 -2.09294 1.75369
17 1.58349 1.19630 -838.17059 -1.13440 2.49583
17 1.58349 1.06247 -838.17059 -1.06629 2.12163
17 1.58349 1.18093 -838.17059 -1.06298 1.60070
17 1.58349 0.86607 -838.17059 -1.17720 1.21462
18 0.71007 1.36873 -838.21318 -1.45493 3.05475
18 0.71007 1.32073 -838.21318 -1.44682 1.74049
18 0.71007 0.87205 -838.21318 -1.43371 1.96923
19 0.81375 1.55526 -838.22776 -1.55122 2.49385
19 0.81375 0.93606 -838.22776 -1.49669 1.62545
20 1.41026 1.16182 -838.27883 -2.07046 2.54608
20 1.41026 1.12548 -838.27883 -2.13382 1.53496
20 1.41026 0.92018 -838.27883 -2.13841 1.42641
21 1.11320 1.44679 -838.33931 -1.44278 4.85704
21 1.11320 1.90084 -838.33931 -1.37370 2.45360
21 1.11320 1.08488 -838.33931 -1.43988 2.23927
21 1.11320 1.11704 -838.33931 -1.45062 1.75270
22 1.02297 1.06883 -838.38982 -1.08059 2.44322
22 1.02297 1.23011 -838.38982 -1.10899 1.94452
22 1.02297 1.09108 -838.38982 -1.16895 1.59963
22 1.02297 0.71127 -838.38982 -1.16483 1.30200
23 0.93655 0.90527 -838.42545 -1.34984 1.86920
24 0.94267 1.31649 -838.45523 -1.10286 3.17461
24 0.94267 1.53709 -838.45523 -1.14828 2.18548
24 0.94267 1.11851 -838.45523 -1.14866 1.93160
24 0.94267 1.14986 -838.45523 -1.20641 1.54493
25 0.93409 1.58223 -838.48111 -0.59256 3.19999
25 0.93409 1.40712 -838.48111 -0.67189 2.31547
25 0.93409 1.31276 -838.48111 -0.73816 2.51870
25 0.93409 1.43619 -838.48111 -0.76191 1.91161
26 1.10963 1.94061 -838.51159 -0.48983 3.75088
26 1.10963 1.62943 -838.51159 -0.48898 2.73911
26 1.10963 1.71675 -838.51159 -0.59343 2.42157
26 1.10963 1.24918 -838.51159 -0.68570 2.03131
27 1.01239 1.21408 -838.54180 -0.22423 2.79964
27 1.01239 1.28707 -838.54180 -0.27037 2.21626
27 1.01239 1.16761 -838.54180 -0.33978 1.99470
27 1.01239 1.18798 -838.54180 -0.38052 2.30839
28 0.99904 2.00689 -838.57548 -0.42145 3.60868
28 0.99904 1.48478 -838.57548 -0.46590 3.12694
28 0.99904 1.77515 -838.57548 -0.46665 2.69566
28 0.99904 1.36915 -838.57548 -0.55089 2.42093
29 1.08747 2.07337 -838.60943 0.22412 3.86743
29 1.08747 1.53891 -838.60943 0.12418 3.13643
29 1.08747 1.44199 -838.60943 0.00407 2.52827
29 1.08747 1.22052 -838.60943 -0.04868 2.26804
30 1.05252 1.64026 -838.65741 -0.36078 3.81790
30 1.05252 1.68328 -838.65741 -0.31216 2.48576
30 1.05252 1.20912 -838.65741 -0.37307 2.28683
30 1.05252 1.08258 -838.65741 -0.39263 1.92800
31 1.18450 1.61427 -838.71793 -0.08531 3.23860
31 1.18450 1.33040 -838.71793 0.00088 2.42273
31 1.18450 1.54899 -838.71793 -0.02638 2.18883
31 1.18450 1.56451 -838.71793 -0.03932 2.13628
32 1.03933 1.41689 -838.77069 0.25821 2.53621
32 1.03933 1.14477 -838.77069 0.27770 1.78663
32 1.03933 1.09611 -838.77069 0.24461 1.73089
32 1.03933 0.89324 -838.77069 0.25327 1.57722
33 1.08842 1.62370 -838.76361 0.05063 4.88193
33 1.08842 2.02494 -838.76361 0.04621 3.05893
33 1.08842 1.49192 -838.76361 -0.05425 3.11762
33 1.08842 1.83727 -838.76361 -0.02824 2.78960
34 0.83841 1.85012 -838.78559 -0.31639 3.56925
34 0.83841 1.63883 -838.78559 -0.32807 2.51644
34 0.83841 1.33222 -838.78559 -0.35989 2.34383
34 0.83841 1.40058 -838.78559 -0.39645 1.98683
35 0.65145 1.85955 -838.80377 0.25735 4.07464
35 0.65145 1.76762 -838.80377 0.16255 2.99598
35 0.65145 1.50431 -838.80377 0.10433 2.65647
35 0.65145 1.56973 -838.80377 0.10903 2.39195
36 0.78107 2.51964 -838.79689 -0.55221 5.35372
36 0.78107 1.99916 -838.79689 -0.56807 3.35049
36 0.78107 1.54493 -838.79689 -0.69043 3.08214
36 0.78107 1.68139 -838.79689 -0.70742 2.35949
37 0.73846 2.21075 -838.79851 -0.10859 4.53305
37 0.73846 1.78659 -838.79851 -0.18550 3.02348
37 0.73846 1.45367 -838.79851 -0.27805 2.74456
37 0.73846 1.54402 -838.79851 -0.30536 2.25614
38 0.55894 1.76472 -838.80781 -0.01699 3.89963
38 0.55894 1.73754 -838.80781 -0.05852 2.74769
38 0.55894 1.36672 -838.80781 -0.11742 2.38937
38 0.55894 1.16194 -838.80781 -0.14926 2.12802
39 0.56293 2.00650 -838.81847 -0.11370 3.89611
39 0.56293 1.68871 -838.81847 -0.18422 2.65175
39 0.56293 1.52714 -838.81847 -0.22914 2.79891
39 0.56293 1.54872 -838.81847 -0.29206 2.30502
40 0.62907 2.00738 -838.83525 -0.23240 4.16088
40 0.62907 1.51817 -838.83525 -0.26239 2.65747
40 0.62907 1.28687 -838.83525 -0.32609 2.30205
40 0.62907 1.21875 -838.83525 -0.33451 1.92690
41 0.61161 1.72996 -838.84858 -0.55719 4.37148
41 0.61161 1.91415 -838.84858 -0.59769 3.13585
41 0.61161 1.57780 -838.84858 -0.62772 2.52989
41 0.61161 1.61699 -838.84858 -0.67357 2.26980
42 0.38275 2.19987 -838.85509 0.11463 4.38312
42 0.38275 1.71591 -838.85509 -0.04637 2.76931
42 0.38275 1.29611 -838.85509 -0.11094 2.22968
42 0.38275 1.31650 -838.85509 -0.13043 1.97657
43 0.40044 1.77780 -838.85974 -0.49646 3.72841
43 0.40044 1.54235 -838.85974 -0.53336 2.33566
43 0.40044 1.18227 -838.85974 -0.56173 1.99161
43 0.40044 1.00647 -838.85974 -0.58103 1.76761
44 0.30775 1.52625 -838.86395 -0.30742 3.22619
44 0.30775 1.48887 -838.86395 -0.31694 2.14870
44 0.30775 1.20427 -838.86395 -0.34916 2.24183
44 0.30775 1.06934 -838.86395 -0.37837 1.69984
45 0.24262 1.75586 -838.86672 -0.53306 4.92109
45 0.24262 1.75990 -838.86672 -0.71024 2.78999
45 0.24262 1.71203 -838.86672 -0.76623 3.80843
45 0.24262 1.65144 -838.86672 -0.76436 2.68579
46 0.22342 2.56304 -838.86832 -0.16968 4.87651
46 0.22342 1.65482 -838.86832 -0.26313 2.67263
46 0.22342 1.57634 -838.86832 -0.31132 2.47998
46 0.22342 1.41888 -838.86832 -0.31199 2.18132
47 0.19933 1.88144 -838.86980 -0.12589 4.88738
47 0.19933 1.63437 -838.86980 -0.23642 2.66231
47 0.19933 1.38115 -838.86980 -0.27349 2.56066
47 0.19933 1.55370 -838.86980 -0.31951 2.16579
48 0.16385 1.85817 -838.87096 -0.69169 3.73348
48 0.16385 1.62882 -838.87096 -0.73654 2.73172
48 0.16385 1.45496 -838.87096 -0.79678 2.38897
48 0.16385 1.17265 -838.87096 -0.81706 2.04203
49 0.15140 2.33208 -838.87168 -0.78244 5.61383
49 0.15140 1.86924 -838.87168 -0.76483 3.12873
49 0.15140 1.53267 -838.87168 -0.88390 2.83238
49 0.15140 1.82571 -838.87168 -0.88379 2.51969
50 0.15140 --- -838.87168 --- ---

I'll send the calculation files to you for comparison.