I get an error 'Conversion from Z-matrix to cartesian coordinates failed' in Gaussian09. Any suggestions?

I'm doing my calculations using G09W. My OS is Windows 7 (64-bit). This error occurs in the frequency calculation. What's even weirder is that it seems to happen right after everything is already calculated (all the energies and corrections). But the output file doesn't contain the fragment that should say that all 4 convergence criteria are satisfied and that a stationary point is located.

It reads 'Error on Z-matrix line number 3: angle Alpha is outside the valid range of 0 to 180. Conversion from Z-matrix to cartesian coordinates failed:'

Apparently, Gaussian doesn't deal well with linear fragments and this molecule is linear. That is generally known. But how do I fix this? I can't really use the results obtained untill the error is dealt with, can I?

My input:

%nprocshared=4
%mem=1400MB
%chk=D:\!HNGH\aug-cc-pwCVTZ\CCSDT XeH2 freq.chk
#p freq=noraman rccsd(t)/genecp maxdisk=15900MB
Title Card Required

0 1
Xe
H   1   R12
H   1   R13           2   A213
R12           =        1.92821600
R13           =        1.92934900
A213          =        180.00000000
Xe     0
S   11   1.00
6456.4000000              0.0002500
974.0610000              0.0016300
214.4290000              0.0047330
31.3308000             -0.0336990
19.6130000              0.2439310
8.8860900             -0.7646530
2.3250800              0.9007940
1.1842200              0.4415130
0.4268540              0.0214160
0.2016450             -0.0030880
0.0921780              0.0008510
S   11   1.00
6456.4000000             -0.0001300
974.0610000             -0.0008850
214.4290000             -0.0023580
31.3308000              0.0072910
19.6130000             -0.0926390
8.8860900              0.3537950
2.3250800             -0.6158300
1.1842200             -0.4130600
0.4268540              0.5793340
0.2016450              0.6254620
0.0921780              0.1423370
S   1   1.00
2.444900D+00           1.000000D+00
S   1   1.00
1.535200D+00           1.000000D+00
S   1   1.00
0.4268540              1.0000000
S   1   1.00
0.2016450              1.0000000
S   1   1.00
0.1143000              1.0000000
S   1   1.00
0.0476000              1.0000000
P   9   1.00
191.1700000              0.0007650
22.5178000              0.0369520
10.0604000             -0.2806880
5.5576200              0.1427170
2.6962500              0.6356540
1.3076400              0.3932570
0.5306750              0.0410560
0.2224610             -0.0022510
0.0868810              0.0007740
P   9   1.00
191.1700000             -0.0003290
22.5178000             -0.0127810
10.0604000              0.1082740
5.5576200             -0.0720510
2.6962500             -0.2845800
1.3076400             -0.1852050
0.5306750              0.3583360
0.2224610              0.5883940
0.0868810              0.2572450
P   1   1.00
5.323800D+00           1.000000D+00
P   1   1.00
1.625100D+00           1.000000D+00
P   1   1.00
0.6953000              1.0000000
P   1   1.00
0.1135000              1.0000000
P   1   1.00
0.0420000              1.0000000
D   9   1.00
142.9710000              0.0007360
41.3057000              0.0053720
10.9222000             -0.0288270
6.8240800              0.1199100
3.6669600              0.3324640
1.9828900              0.3892090
1.0539900              0.2483590
0.5299030              0.0776110
0.2192000              0.0073660
D   1   1.00
2.520100D+00           1.000000D+00
D   1   1.00
1.254800D+00           1.000000D+00
D   1   1.00
0.5299030              1.0000000
D   1   1.00
0.2192000              1.0000000
D   1   1.00
0.0917000              1.0000000
F   1   1.00
5.615800D+00           1.000000D+00
F   1   1.00
1.614800D+00           1.000000D+00
F   1   1.00
0.4779000              1.0000000
F   1   1.00
0.2287000              1.0000000
G   1   1.00
1.753000D+00           1.000000D+00
****
H 0
aug-cc-pVTZ
****
XE     0
XE-ECP     4     28
g potential
1
2      1.0000000              0.0000000
s-g potential
3
2     40.0051840             49.9979620
2     17.8122140            281.0133030
2      9.3041500             61.5382550
p-g potential
4
2     15.7017720             67.4391420
2     15.2586080            134.8747110
2      9.2921840             14.6633000
2      8.5590030             29.3547300
d-g potential
6
2     15.1856000             35.4369080
2     14.2845000             53.1957720
2      7.1218890              9.0462320
2      6.9919630             13.2236810
2      0.6239460              0.0848530
2      0.6472840              0.0441550
f-g potential
4
2     20.8815570            -23.0892950
2     20.7834430            -30.0744750
2      5.2533890             -0.2882270
2      5.3611880             -0.3869240

• +1. Does it work if you convert your ZMAT to XYZ using this then just use XYZ in your Gaussian input? – Nike Dattani Jul 27 '20 at 20:21
• Can you tell gaussian to not use a zmatrix? Alternatively you can add a ghost atom – Cody Aldaz Jul 27 '20 at 21:51
• If the molecule is linear (any part of it can cause this error), or even nearly linear, the cleanest way with a z-matrix is to specify it with dummy atoms. In Cartesian coordinates, you should take the time to specify the redundant coordinates manually using linear bent descriptions instead of dihedral angles. If it is a tiny molecule with an inexpensive method, you can force it to optimise in Cartesian coordinates directly. This is very, very brute force though. – Martin - マーチン Jul 27 '20 at 23:16
• If you are able to add the full input (and the corresponding version of Gaussian) I'm sure you'll find someone to fix your problem here. – Martin - マーチン Jul 27 '20 at 23:18
• @QuantumX Okay, then if you're using ZMAT, there's always strange oddities that require dummy atoms (this is not just in Gaussian, but also in CFOUR, for example). Since I use CFOUR (not GAUSSIAN), I know of this page that might help with learning about dummy atoms for Z-matrices (search for "dummy"): slater.chemie.uni-mainz.de/cfour/…. Gaussian might have something similar, but the information on this CFOUR help page, will probably be helpful for anyone wanting to make a ZMAT with dummy atoms (regardless of the software they're using). – Nike Dattani Jul 29 '20 at 16:09

You should change your input to avoid defining a linear angle. This can be done by introducing a dummy atom. Here is an example with your input:

0 1
H
Xe   1   R12
X    2   1.0     1   A1
H    2   R13     3   A1    1    D1

R12  =   1.92821600
R13  =   1.92934900
A1   =   90.0000000
D1   =   180.000000


By defining the hydrogen as perpendicular to the dummy atom rather than 180 degrees with respect to each other, you can still work with a linear molecule, but not have to worry about the angles you have defined going out of range for Gaussian.

Also, you may want to use R12 for both $$\ce{Xe-H}$$ bond lengths. This molecule should be symmetric and specifying them with the same variable will enforce that throughout the calculation.