UT theoretical chemistry code forum

I report the bader charge analysis results of metallic Li.
I changed the NGX(Y,Z) grids for the bader analysis charge calculation of metallic Li in the case 1-4.
The case 5 indicates the result using the potpaw(Li_sv) containing two 1s electrons and 2s electron.

It was found that the charges on Li atoms do not converge 1.0000 even if NGX(Y,Z) increase.
For example, the charge for Li1 atom is 1.7411 --> 1.1190 --> 1.7522 --> 1.7530.
In contrast, the bader analysis using potpaw(Li_sv) gives me the charge of 3.0000 though NGX(Y,Z)=24(default value)!
Is the method using the sum of core charge and valence charge appropriate?
Is the case of metallic Li peculiar?

[Case 1. potpaw(Li) NGX(Y,Z)F=25]
# X Y Z CHARGE MIN DIST
----------------------------------------------------------------
1 0.0000 0.0000 0.0000 1.7411 1.5984
2 1.6820 1.6820 1.6820 0.2589 0.9228
----------------------------------------------------------------
NUMBER OF ELECTRONS: 2.00000

# X Y Z CHARGE ATOM DISTANCE
-------------------------------------------------------------------------
1 0.0000 0.0000 0.0000 0.2589 1 0.0000
2 2.5230 0.0000 1.6820 0.1238 1 1.8806
3 0.0000 0.8410 1.6820 0.1332 1 1.8806
4 0.0000 1.6820 0.8410 0.1232 1 1.8806
5 0.0000 1.6820 2.5230 0.1232 1 1.8806
6 2.5230 1.6820 0.0000 0.1238 1 1.8806
7 0.0000 2.5230 1.6820 0.1140 1 1.8806
8 0.8410 0.0000 1.6820 0.1238 1 1.8806
9 0.8410 1.6820 0.0000 0.1238 1 1.8806
10 1.6820 1.6820 1.6820 0.2589 2 0.0000
11 1.6820 0.0000 0.8410 0.1232 1 1.8806
12 1.6820 0.0000 2.5230 0.1232 1 1.8806
13 1.6820 0.8410 0.0000 0.1140 1 1.8806
14 1.6820 2.5230 0.0000 0.1332 1 1.8806
-------------------------------------------------------------------------


[Case 2. potpaw(Li) NGX(Y,Z)F=50]
# X Y Z CHARGE MIN DIST
----------------------------------------------------------------
1 0.0000 0.0000 0.0000 1.1190 1.0002
2 1.6820 1.6820 1.6820 0.8810 1.0002
----------------------------------------------------------------
NUMBER OF ELECTRONS: 2.00000

# X Y Z CHARGE ATOM DISTANCE
-------------------------------------------------------------------------
1 0.0000 0.0000 0.0000 0.2512 1 0.0000
2 2.4894 0.0000 1.6820 0.1249 2 1.8658
3 0.0000 0.8747 1.6820 0.1303 2 1.8658
4 0.0000 1.6820 0.8074 0.1247 1 1.8658
5 0.0000 1.6820 2.5567 0.1247 1 1.8658
6 2.4894 1.6820 0.0000 0.1249 2 1.8658
7 0.0000 2.5567 1.6820 0.1194 1 1.8658
8 0.8747 0.0000 1.6820 0.1249 2 1.8658
9 0.8747 1.6820 0.0000 0.1249 2 1.8658
10 1.6820 1.6820 1.6820 0.2512 2 0.0000
11 1.6820 0.0000 0.8074 0.1247 1 1.8658
12 1.6820 0.0000 2.5567 0.1247 1 1.8658
13 1.6820 0.8074 0.0000 0.1194 1 1.8658
14 1.6820 2.5567 0.0000 0.1303 1 1.8658
-------------------------------------------------------------------------


[Case 3. potpaw(Li) NGX(Y,Z)F=100]
# X Y Z CHARGE MIN DIST
----------------------------------------------------------------
1 0.0000 0.0000 0.0000 1.7522 1.7480
2 1.6820 1.6820 1.6820 0.2478 1.0531
----------------------------------------------------------------
NUMBER OF ELECTRONS: 2.00000

# X Y Z CHARGE ATOM DISTANCE
-------------------------------------------------------------------------
1 0.0000 0.0000 0.0000 0.2478 1 0.0000
2 2.5230 0.0000 1.6820 0.1254 1 1.8806
3 0.0000 0.8410 1.6820 0.1281 1 1.8806
4 0.0000 1.6820 0.8410 0.1253 1 1.8806
5 0.0000 1.6820 2.5230 0.1253 1 1.8806
6 2.5230 1.6820 0.0000 0.1254 1 1.8806
7 0.0000 2.5230 1.6820 0.1226 1 1.8806
8 0.8410 0.0000 1.6820 0.1254 1 1.8806
9 0.8410 1.6820 0.0000 0.1254 1 1.8806
10 1.6820 1.6820 1.6820 0.2478 2 0.0000
11 1.6820 0.0000 0.8410 0.1253 1 1.8806
12 1.6820 0.0000 2.5230 0.1253 1 1.8806
13 1.6820 2.5230 0.0000 0.1281 1 1.8806
14 1.6820 0.8410 0.0000 0.1226 1 1.8806
-------------------------------------------------------------------------


[Case 4. potpaw(Li) NGX(Y,Z)F=200]
# X Y Z CHARGE MIN DIST
----------------------------------------------------------------
1 0.0000 0.0000 0.0000 1.7530 1.7771
2 1.6820 1.6820 1.6820 0.2470 1.0798
----------------------------------------------------------------
NUMBER OF ELECTRONS: 2.00000

# X Y Z CHARGE ATOM DISTANCE
-------------------------------------------------------------------------
1 0.0000 0.0000 0.0000 0.2470 1 0.0000
2 2.5230 0.0000 1.6820 0.1255 1 1.8806
3 0.0000 0.8410 1.6820 0.1269 1 1.8806
4 0.0000 1.6820 0.8410 0.1255 1 1.8806
5 0.0000 1.6820 2.5230 0.1255 1 1.8806
6 2.5230 1.6820 0.0000 0.1255 1 1.8806
7 0.0000 2.5230 1.6820 0.1241 1 1.8806
8 0.8410 0.0000 1.6820 0.1255 1 1.8806
9 0.8410 1.6820 0.0000 0.1255 1 1.8806
10 1.6820 1.6820 1.6820 0.2470 2 0.0000
11 1.6820 0.0000 0.8410 0.1255 1 1.8806
12 1.6820 0.0000 2.5230 0.1255 1 1.8806
13 1.6820 2.5230 0.0000 0.1269 1 1.8806
14 1.6820 0.8410 0.0000 0.1241 1 1.8806
-------------------------------------------------------------------------

[Case 5. potpaw(Li_sv) NGX(Y,Z)F=24]
# X Y Z CHARGE MIN DIST
----------------------------------------------------------------
1 0.0000 0.0000 0.0000 3.0000 0.8966
2 1.6803 1.6803 1.6803 3.0000 0.8966
----------------------------------------------------------------
NUMBER OF ELECTRONS: 6.00003

# X Y Z CHARGE ATOM DISTANCE
-------------------------------------------------------------------------
1 0.0000 0.0000 0.0000 2.2738 1 0.0000
2 0.0000 2.5205 1.6803 0.1213 2 1.8786
3 0.0000 0.8402 1.6803 0.1428 1 1.8786
4 0.0000 1.6803 0.8402 0.1309 1 1.8786
5 0.0000 1.6803 2.5205 0.1320 2 1.8786
6 0.8402 0.0000 1.6803 0.0996 1 1.8786
7 0.8402 1.6803 0.0000 0.0996 1 1.8786
8 1.6803 1.6803 1.6803 2.2738 2 0.0000
9 1.6803 0.0000 0.8402 0.1320 1 1.8786
10 1.6803 0.0000 2.5205 0.1309 2 1.8786
11 1.6803 0.8402 0.0000 0.1213 1 1.8786
12 1.6803 2.5205 0.0000 0.1428 2 1.8786
13 2.5205 0.0000 1.6803 0.0996 2 1.8786
14 2.5205 1.6803 0.0000 0.0996 2 1.8786
-------------------------------------------------------------------------

This sounds very similar to what was discussed in a previous thread ... viewtopic.php?t=368

Dear yohsi,
I also have the same situation. I tried Li in bcc and fcc structure, setting NGX(Y,Z) up to 500 (crazy? I know). It just wouldn't change a bit. The total charge is always correct, but the number of electrons of individual atoms are never correct. You have bcc case. In fcc case, I have one Li which is extraordinarily larger in volume than the rest three. I also have more than 6000 Bader maxima which is nothing but noise. I also plotted the charge density distribution, but it looks nothing strange. I am also wondering what is going on with Li.
Riche
PS: I used Li_sv with 1s and 2s.