Convergence issues with Ibrion=1 and vasp5.2.12 for multiple systems

Message

akulkarni34 · #1 Post by **akulkarni34** » Wed Oct 10, 2012 4:15 am

Hello,
I have been a VASP user for a while but have now started noticing issues with VASP convergence recently for multiple systems.
Specifically, for my system I carry out geometry optimization with the CG method (with or without spin, POTIM = 0.1) and end up with the ZBRENT warning (curvature = 0). Since, I can't find an optimum POTIM value for the ZBRENT ionic step, I 'choose' a reasonable POTIM and shift to Quasi Newton and keep a track of the maximum force on the atoms and the maximum displacement.
--
Ionic Steps
Step 1: max disp= 38.30604158 f= 0.13431398 for atom O (104)
Step 2: max disp= 0.00129534 f= 0.13439252 for atom O (104)
Step 3: max disp= 0.00003000 f= 0.09108829 for atom O (151)
Step 4: max disp= 0.00107634 f= 0.09241037 for atom O (104)
Step 5: max disp= 0.00049285 f= 0.13178783 for atom O (104)
Step 6: max disp= 0.00088798 f= 0.11513393 for atom O (104)
Step 7: max disp= 0.00092136 f= 0.12207442 for atom O (104)
Step 8: max disp= 0.00139334 f= 0.12624884 for atom O (104)
Step 9: max disp= 0.00200125 f= 0.17448056 for atom O (151)
--
In each case the ionic displacements are small (one hundredth of an angstrom) but the forces seem to oscillate or diverge. I cannot seem to get them to be lower thatn 0.03 eV/ang.

I have tried the following
1. Try CG with potim= 0.1, 0.01, 0.001 to try and find 'optimum'. It always give the ZBRENT error.
2. Try QN with potim = 0.1, 0.05 .... Even though I am so close to the minimum the force does not decrease
3. Increase the SCF accuracy by changing EDIFF and the minimum number of SCF cycles
4. Change NFREE = 0, 1, 5 10 .. none of them work
5. I can't change ENCUT = 400 as I am comparing different structures at the same encut value
6. Check the compiler version (intel 11.1.059 and 12.0.0.089) and vasp version (12 and 13).
7. I am still having convergence issue both with spin on and spin off.

If I am close to the minima (as the curvature is zero), shouldn't the QN method converge immediately? This is happening for multiple systems and I can't figure out what the issue could be. I have been trying for more than a couple of months without any success.

I'd really appreciate some feedback.
Thank you,

Regards,
Ambarish

POTCAR
----
TITEL = PAW_PBE H 15Jun2001
TITEL = PAW_PBE C 08Apr2002
TITEL = PAW_PBE N 08Apr2002
TITEL = PAW_PBE O 08Apr2002
TITEL = PAW_PBE Cu 05Jan2001

INCAR
----
SYSTEM=Charge_calculation

# Files to read and write
LCHARG= .TRUE.
LWAVE = .TRUE.
LAECHG= .FALSE.
LVTOT = .FALSE.

# spin polarized
ISPIN=2
ICHARG=1
MAGMOM = 16*0 72*0 8*0 64*0 6*1 6*-1 #16 72 8 64 12
LORBIT = 11

# Electronic steps
PREC = Accurate
ENCUT = 400.000
ENAUG =650.000
NELMDL= -4
NELMIN= 4
EDIFF = 1E-04
ALGO = Fast
NSIM = 4
LREAL = Auto

# Ionic steps
NSW = 1000
IBRION= 1
ADDGRID = .TRUE.
#NFREE = 10
EDIFFG= -3E-02
ISIF = 2
POTIM = 0.1
ISMEAR= 0
NPAR = 4
LPLANE= .TRUE.

# Grimmes method
LVDW = .TRUE.
VDW_RADIUS = 30.0
VDW_SCALING = 0.75 s6 value for PBE
VDW_D = 20.0 parameter d in the dampling function
---

POSCAR
---
test
0.990000000000000
18.7136999999999993 0.0000000000000000 0.0000000000000000
9.3568499999999997 16.2065395999999993 0.0000000000000000
9.3568499999999997 5.4021799000000001 15.2796721000000009
H C N O Cu
16 72 8 64 12
Selective dynamics
Direct
0.0314546404051822 0.7328471105721247 0.5014210140064779 T T T
0.0297079775603472 0.5017603893969818 0.7347034973128451 T T T
0.2667527858351689 0.4961938025785905 0.2638207152459398 T T T
0.2674088685279770 0.2644375198557865 0.4964735844306852 T T T
0.7322329007340624 0.5041262269399471 0.0268993629906784 T T T
0.7318979230394315 0.7356378115211368 0.0286195767091692 T T T
0.5022229941855245 0.0286606096771823 0.7357693887204564 T T T
0.7349372714090950 0.0268293258726596 0.5038754407313335 T T T
0.2655199095730504 0.9732324930681678 0.2651977114929445 T T T
0.4977794428439334 0.9720969757581450 0.2664512953404236 T T T
0.2625215332374135 0.4974785512533964 0.9720280624581271 T T T
0.4955602663596327 0.2661217051180355 0.9706865047450199 T T T
0.9686618267112393 0.4976792549760107 0.2667913260820983 T T T
0.9710395179639552 0.2639822071878066 0.4975844943321546 T T T
0.5025170034831945 0.7336955721574689 0.7321543529298061 T T T
0.7352561252204430 0.5020960062383529 0.7319764177380066 T T T
0.0266930254135202 0.5676871455943954 0.5683119618795689 T T T
0.0323120242999455 0.6119503030115542 0.6125776503590579 T T T
0.0326439676895392 0.6985013381652361 0.5688717791324317 T T T
0.0315010070187796 0.5690039387810518 0.6996143806917675 T T T
0.0319648049435121 0.7430330203947391 0.6116970093542794 T T T
0.0305395673599559 0.6123304751360424 0.7438123156363912 T T T
0.0266795880782160 0.8371179168446513 0.5637443402379853 T T T
0.0233474548165889 0.5653932287939553 0.8384923374029011 T T T
0.0308698527415197 0.6996309579405573 0.6993875596877110 T T T
0.1642137569207709 0.4302449084836067 0.4293299644893470 T T T
0.2582987959693865 0.3858549693511201 0.3855515148089448 T T T
0.3013021475659238 0.4292103473934300 0.2987235135240902 T T T
0.3016489324218647 0.2993595095281715 0.4292805645667302 T T T
0.3882518039525445 0.3867659487425664 0.2546420413991544 T T T
0.3888222598647282 0.2556643401389718 0.3867360491552819 T T T
0.4283387358133155 0.4341873254417804 0.1599305511845300 T T T
0.4303620681589307 0.1616477288683499 0.4344844028668105 T T T
0.4320904537806973 0.2999375609174625 0.2993766248404284 T T T
0.5704424587150616 0.5663446581423447 0.0226689148414055 T T T
0.6107000249276175 0.6135291098387239 0.0300292388433967 T T T
0.5669046422873255 0.7003281938903977 0.0319144657440145 T T T
0.6977285234698456 0.5710961793630637 0.0294387553999455 T T T
0.6102322222436578 0.7445457116402555 0.0317026817790197 T T T
0.5687411228864871 0.8385637774676993 0.0269130632376131 T T T
0.6974560775899811 0.7008231229686606 0.0307299760947297 T T T
0.7407725152571834 0.6144350744311355 0.0299829586781568 T T T
0.8349590514158605 0.5701766233657143 0.0238469615110206 T T T
0.5690691261216742 0.0232706014607518 0.5707034817022043 T T T
0.6132455197534110 0.0294218327746241 0.6145676249227153 T T T
0.5696957890155152 0.0298197346438362 0.7011400498102468 T T T
0.6128487952254718 0.0290289022729068 0.7454145799250146 T T T
0.7001752251260775 0.0288475860019731 0.5710557719501045 T T T
0.5654550314646565 0.0232782556492474 0.8397593727121317 T T T
0.7444476294236664 0.0285370093089614 0.6138265838712237 T T T
0.7003450676522520 0.0286669956211352 0.7012457070357556 T T T
0.8388989373247273 0.0221618931560139 0.5663691331969118 T T T
0.1622163842522782 0.9756531876827486 0.4278442276399677 T T T
0.2565563154124961 0.9695251091853125 0.3870602761827414 T T T
0.3004097972643590 0.9705874608622913 0.2999124717651044 T T T
0.4311467058464787 0.9773807861466337 0.1626328814962017 T T T
0.3006450881062386 0.9689408132411280 0.4308274183784889 T T T
0.3870993663713686 0.9707381401130746 0.2569769444590434 T T T
0.4305662530439303 0.9701955208053852 0.3005857779106677 T T T
0.3877036764432140 0.9697837475621301 0.3876664828293550 T T T
0.4347570757641749 0.9757580706924323 0.4290711234682729 T T T
0.1589013926990945 0.4347229361570106 0.9768472314185984 T T T
0.2536288207471321 0.3875914091208595 0.9696943037652010 T T T
0.2985207535859735 0.3006871832268335 0.9681204228539392 T T T
0.2975939476973446 0.4304088786843907 0.9697706607768244 T T T
0.4333683641839866 0.1622713896243650 0.9745585115813518 T T T
0.3857431969541562 0.2565905876086940 0.9686145566139374 T T T
0.4283176299030663 0.3007208859902952 0.9688735183172613 T T T
0.3846170727130697 0.3873597969859182 0.9689172905136517 T T T
0.4288064302343300 0.4320504471537673 0.9738699030230715 T T T
0.9687541883684809 0.3000154130046259 0.2999634094134248 T T T
0.9716252690221433 0.4363270831798626 0.1628616809055520 T T T
0.9747944915781781 0.1610963236738497 0.4345761784937485 T T T
0.9675359679808226 0.3877764168596466 0.2566129089061943 T T T
0.9689345875239441 0.2554923894780764 0.3873273066072590 T T T
0.9677026700356592 0.4301571506368507 0.3010018073815382 T T T
0.9689731850178710 0.2992600687017351 0.4304308959964563 T T T
0.9690214596627157 0.3860361357558801 0.3872087730337914 T T T
0.9763574187244417 0.4297572032073753 0.4307600270791518 T T T
0.5699797385950707 0.6992134188601870 0.6979759687597690 T T T
0.5692399296917015 0.5680155398389437 0.8349010857023158 T T T
0.5660407508877472 0.8382109542751981 0.5710977974284699 T T T
0.6134637436811465 0.6124516680413126 0.7408943896197402 T T T
0.6132526595459008 0.7437806624018281 0.6116290033840182 T T T
0.7007964769662830 0.7000662719597421 0.5683934251286075 T T T
0.7004972157037964 0.5693969295398345 0.6978118887645884 T T T
0.7450234759971872 0.6126097599875676 0.6113565842041800 T T T
0.8395911862256692 0.5653958751250332 0.5706350741170336 T T T
0.0276600726137025 0.7472208554582975 0.7463380916687482 T T T
0.5247406330190387 0.2540058885325480 0.2532943358968928 T T T
0.4741669657729822 0.7462150468837000 0.0327589194569062 T T T
0.7471027976625557 0.0261334549793207 0.7485963874966202 T T T
0.2545505424215165 0.9680212062263306 0.5236102284785358 T T T
0.9721979962926756 0.2531433683004055 0.2521540439623274 T T T
0.7480109715761756 0.7478479882981637 0.4773149800128820 T T T
0.2531824722757034 0.2545345329517040 0.9667240877101591 T T T
0.0264119074522771 0.6117108632556785 0.4889716125283541 T T T
0.0202977698980699 0.4906968064274217 0.6150258600898931 T T T
0.0132377595699680 0.8726786082021999 0.4902934023603134 T T T
0.0251718440403447 0.4857488732272404 0.8709370126676179 T T T
0.0340844297515667 0.8735580605922225 0.6013506914448769 T T T
0.0135262823848723 0.6105066505118817 0.8786759555231586 T T T
0.9536013054901714 0.7917240910869022 0.7829252589590054 T T T
0.0993932167855757 0.7375570524067203 0.7443382769692786 T T T
0.1280961179348254 0.3856043922356401 0.5078906663530185 T T T
0.1283314346772920 0.5080505270490205 0.3829711374037283 T T T
0.3807438284572169 0.5119969242147773 0.1264269285319793 T T T
0.3795955722046745 0.1243726294049130 0.5070215836303367 T T T
0.5059494544228960 0.3909693123889380 0.1208511479647651 T T T
0.5124290935130800 0.1268957679410119 0.3970792942110131 T T T
0.5822818902608078 0.2655888324152492 0.2575681651429148 T T T
0.5355919582145668 0.2088844464497902 0.2151061044502367 T T T
0.4927673835800621 0.6096265895009644 0.0180971207267199 T T T
0.6179454323542188 0.4886071683177725 0.0191043781244153 T T T
0.4867118255791860 0.8735051071667725 0.0365269021401143 T T T
0.6195453856101558 0.8756493779133490 0.0116356652328914 T T T
0.8725419133995033 0.4962068138406518 0.0107526799757304 T T T
0.8693508098641677 0.6108273292516090 0.0300961666054597 T T T
0.4167405284816941 0.7345232993835062 0.1062763316417818 T T T
0.4630373099092234 0.7912938284028104 0.9604399692308236 T T T
0.4871338168783322 0.0318421978793000 0.6110420959611774 T T T
0.6183223279735872 0.0075388891956658 0.4979795409028779 T T T
0.4938300871568695 0.0060580027887788 0.8771061178536824 T T T
0.6017160607539214 0.0343865847722593 0.8747473431803767 T T T
0.8712449453363360 0.0259698541777330 0.4861460000670701 T T T
0.8789342235268216 0.0110165484979829 0.6124032144965708 T T T
0.7825236917344318 0.9525565233031974 0.7945688421781760 T T T
0.7456500399245993 0.0981512043020172 0.7377004997134674 T T T
0.1308557435233137 0.9667154363468673 0.3865145917400608 T T T
0.1212167896552429 0.9922864845876184 0.5000519424299149 T T T
0.3819126514225625 0.9932443205478224 0.1233539408445861 T T T
0.5130774075933016 0.9687357166323131 0.1297385484927316 T T T
0.3984446294888719 0.9645959917236533 0.5114464523737411 T T T
0.5061324561139913 0.9934867509156174 0.3775721816459950 T T T
0.1190512369078054 0.3897535403692099 0.9858214289369904 T T T
0.1261591239953944 0.5140862974391538 0.9759048806354182 T T T
0.3976144140723614 0.1271965660840621 0.9629268608473056 T T T
0.5043249313754875 0.1250524274324135 0.9928992387774517 T T T
0.3815907350534134 0.5089570545411101 0.9804882531914327 T T T
0.5084264533052523 0.3887030285280743 0.9731692722385120 T T T
0.9644683900898213 0.3985541591709347 0.1264207950223285 T T T
0.9882522192160714 0.1202548641245594 0.3881816769975774 T T T
0.9838367529206068 0.5104304975431357 0.1274188408973625 T T T
0.9683612981052371 0.1295552445455221 0.5156112514281124 T T T
0.9808341733394270 0.5075851405290215 0.3850093018322613 T T T
0.9799021079833354 0.3841309459782932 0.5088339152011248 T T T
0.4898404459989499 0.6116378103380076 0.8705419368101214 T T T
0.4948689069170510 0.8752915549578417 0.6230577843428566 T T T
0.6164968699576272 0.4911824464104594 0.8712672644245005 T T T
0.6024099944300378 0.8732503929543490 0.4887851756622008 T T T
0.8788575245027097 0.6088756059448722 0.4932515361078514 T T T
0.8727765416262988 0.4870307767296108 0.6175803831888662 T T T
0.0463018513887616 0.2173804945261090 0.2070360158397929 T T T
0.7850730030460199 0.7916357666033514 0.4709015279019327 T T T
0.2570574679807853 0.2677203717050666 0.8928321432827793 T T T
0.2572358995217792 0.8945375650496826 0.5807575538440394 T T T
0.2182840750993724 0.0399990421014366 0.5351114891643741 T T T
0.7458564205849308 0.7393340173250682 0.4166662082289213 T T T
0.2164550584695072 0.2076891687241083 0.0384636913044805 T T T
0.9006380347001319 0.2544653370327630 0.2620740212010935 T T T
0.4314368363956531 0.9997889009357193 0.5707614839429125 T T T
0.4296773359767223 0.5681566383150276 0.9980771348826756 T T T
0.0032782707338975 0.5661380330263268 0.4308188163047912 T T T
0.4314141680114737 0.0003541614762466 0.0004877694303379 T T T
0.0008759054007812 0.9972975252663382 0.4293990655374074 T T T
0.9958610034229408 0.4343176637857518 0.9993795752510620 T T T
0.5682028715408406 0.0000889960317890 0.4325640171039207 T T T
0.5687441899690014 0.4325016689756594 0.9963010249533223 T T T
0.0009103163049948 0.4299707219072807 0.5687054732325382 T T T
0.5681524635029344 0.0004361116244331 0.0018732431360659 T T T
0.0003017602909264 0.0008320452448580 0.5657829623101353 T T T
0.9974066110771946 0.5697036247663809 0.0024294224745891 T T T

[ Edited ]

akulkarni34 · #2 Post by **akulkarni34** » Wed Oct 10, 2012 4:31 am

I should also mention that the energies are well converged (changes of 1E-3 to 1E-4) but are sometimes positive. This is very surprising as the QN method should move in the direction of lower forces and thus also reduce the energy of the system.
Thank you,

Regards,
Ambarish

tlchan · #3 Post by **tlchan** » Thu Oct 11, 2012 4:14 am

A minimization algorithm decreases the step size when the trial step made previously increases the total energy. The step size keeps on decreasing till the total energy decreases for the chosen direction. If this occurs, it usually means that there are inconsistencies between the total energy and the force.

akulkarni34 · #4 Post by **akulkarni34** » Thu Oct 11, 2012 4:25 am

Hmm. That seems to make sense. If I understand the CG and QN correctly, CG uses the 'energy' to get an approximate hessian and thus the energy should necessarily decrease. The QN method works based on forces to get the direction (not energy) and does not have an explicit constraint on reducing the energy of the system, though ideally that should be the case. Please correct me if I am wrong.

Is there something that I should try to check for this inconsistency that you mentioned? Would this be fixed by using smaller and smaller value of POTIM for QN?
Thank you for you help!

Regards,
Ambarish

akulkarni34 · #5 Post by **akulkarni34** » Sun Oct 14, 2012 4:01 am

Any help from the admin? Can someone try running these calculations to figure out the problem?