◼ Tutorial 8 — Ring-current maps and ACID: who carries the current?
Tutorial 7 gave the ring current as numbers: NICS, Icyc, bond currents. This tutorial draws the same current as pictures: arrows in a plane, streamlines, and ACID surfaces with the current drawn on them, as in Herges’ AICD program. And it adds one thing AICD cannot do: split the current, point by point, into the part carried by delocalized electrons (EDDB) and the part carried by localized ones (EDLB) — without choosing any orbitals.
What you will learn
Exporting the induced current density in 3D (--export-vtk jcur,acid,divj) and in a plane (--export-map jcur).
Drawing current maps, streamlines and ACID surfaces with the scripts shipped with runEDDB.
Reading the EDLB/EDDB partition of the current: benzene’s inner paratropic loop and pentalene’s paratropic current are carried by localized electrons.
Checking whether to trust the arrows (the div J table), and getting a qualitative map without any NMR job.
Before you start
The induced current density J(r) is the first-order response of the electrons to a magnetic field B along the ring normal n; per unit field it is measured in nA/T/Å2. Seen from the +n side, a diatropic current circulates clockwise (aromatic, shielding inside the ring) and is drawn red; a paratropic one circulates anticlockwise (antiaromatic) and is drawn blue. runEDDB splits the current exactly as it splits the shielding: Jtotal = JEDLB + JEDDB + Jother at every point (NMR & currents page of the Manual).
Two Gaussian 16 jobs, B3LYP/def2-TZVPP with NMR=CSGT IOp(10/33=2), are provided (the route sections are in the .gjf files). The .log file must lie next to the .fchk: it carries the exact NMR response; without it runEDDB falls back to the approximate lane and warns that NMR data are missing.
The zip (7.6 MB) holds _nmr.fchk, _nmr.log and _nmr.gjf for benzene and pentalene. The pictures are drawn by Python scripts installed with the program in ~/runEDDB/scripts/ (they need numpy and matplotlib; the 3D pictures also need POV-Ray).
--nics-rings 6 — find the six-membered ring and put the NICS probes on it; --rcs adds its ring current Icyc.
-a — the π-NOBDs. The tables then report the π-part: the first column is EDLB + EDDB of the π-NOBDs, and everything else (σ, core) goes to other. The 3D file gets an extra field, J_EDDB_sel.
--export-vtk jcur,acid,divj — the current vector field, the ACID scalar and div J on a 3D grid (benzene_nmr.JCUR.vtk, plus JCUR_atoms.vtk for the skeleton).
--export-map jcur,jcur-eddb — the same current in a plane 1 Å above the ring (.map.dat files).
benzene_run.out
⋯ banner and first stage lines omitted ⋯
▶ Computing NMR shielding partition (CTOCD-DZ)... 5s
WARNING: CTOCD-DZ translational sum rule 81.2 % < 90 %: the shieldings at the nuclei
are not converged in this basis (NICS away from the nuclei is fine).
⋯ files, div J table and shielding table omitted ⋯
│ NICS = -σ (ppm) AT RING CENTROIDS AND PROBE POINTS, SELECTED NOBDs │
└────────────────────────────────────────────────────────────────────────────────────┘
R1 1-2-4-6-5-3-1
Probe NICS EDLB_G EDDB_G other
──────── ═══════════ ─────────── ─────────── ───────────
R1 NICS(0) -19.27789 -2.22896 -17.04893 11.99041
R1 NICS(0)zz -35.54055 -4.11929 -31.42126 18.91675
R1 NICS(1) -9.61042 -1.11106 -8.49937 0.44280
R1 NICS(1)zz -28.76576 -3.33431 -25.43145 -0.93880
──────── ═══════════ ─────────── ─────────── ───────────
NICS = -σ; zz = along the ring normal; NICS(1) = mean of the +-1 A probes
┌────────────────────────────────────────────────────────────────────────────────────┐
│ RING CURRENT STRENGTHS (nA/T), SELECTED NOBDs │
└────────────────────────────────────────────────────────────────────────────────────┘
Ring Item total EDLB_G EDDB_G
──── ──────────────────────── ═══════════ ─────────── ───────────
R1 I_cyc 11.30130 1.30967 9.99162
──── ──────────────────────── ═══════════ ─────────── ───────────
I_cyc: min flux centre->vacuum
▶ Printing summary of EDDB calculations:
Total ED_G population: 5.98357 per atom: 0.49863
Total EDLB_G population: 0.69022 per atom: 0.05752
Total EDDB_G population: 5.29335 per atom: 0.44111
▶ Calculation completed (with 1 warning!) on 2026-10-29. Total time: 16s.
The warning concerns the shieldings at the nuclei (a basis-set limit of the CSGT response) and does not affect NICS or the current away from them. The π-ring current is 11.30 nA/T, of which 9.99 is carried by the delocalized π-electrons and 1.31 by the localized ones (without -a the table gives the full current, 11.97 nA/T: the σ-electrons add 0.67).
Figure 1. Benzene, plane 1 Å above the ring: Jtotal | JEDLB | JEDDB. Colour = diatropic component (red +, blue −), each panel on its own scale; arrows = in-plane current.
At 1 Å the picture is the familiar diatropic π-torus, and it is JEDDB (maximum about 2.5 nA/T/Å2, right). JEDLB is several times weaker and shows two opposite circulations: diatropic along the C–C and C–H bonds outside, paratropic inside the ring. Add --stream for streamlines:
Figure 2. Streamlines of Jtotal and JEDDB, 1 Å above benzene.
In the plane of the ring (--plane fit,0) the σ-electrons dominate. The inner paratropic loop is seen in every picture of the total current, including AICD:
Figure 3. Benzene in the molecular plane: Jtotal | JEDLB | JEDDB.
Part 2 — Benzene in 3D: ACID with the current on it
Each command draws the ACID isosurface of one channel with that channel’s current on it and renders it with POV-Ray. --colour sign colours the arrows red (diatropic) or blue (paratropic); without it they are all green, as in AICD, and the sense is read from the arrows themselves.
Figure 4. Benzene: ACIDtotal, ACIDEDLB and ACIDEDDB with J on the surface (red = diatropic, blue = paratropic).
Conclusion 1. The inner paratropic loop that every total-current picture of benzene shows (left; the only picture AICD gives) sits entirely in the localized channel: ACIDEDLB carries a diatropic current along the C–H periphery and a paratropic one along the inner edge (about 95 diatropic vs 41 paratropic arrows), while the ACIDEDDB torus is diatropic throughout (189 vs 3). The well-known σ/π picture of the benzene current comes out of the density partition, not out of a choice of orbitals.
Part 3 — Pentalene: a paratropic current carried by localized electrons
⋯ banner, stage lines and other tables omitted ⋯
│ RING CURRENT STRENGTHS (nA/T), SELECTED NOBDs │
└────────────────────────────────────────────────────────────────────────────────────┘
Ring Item total EDLB_G EDDB_G
──── ──────────────────────── ═══════════ ─────────── ───────────
R1 I_cyc -21.18167 -15.94730 -5.34748
R2 I_cyc -21.17947 -15.94935 -5.34393
──── ──────────────────────── ═══════════ ─────────── ───────────
I_cyc: min flux centre->vacuum
▶ Printing summary of EDDB calculations:
Total ED_G population: 7.96846 per atom: 0.56918
Total EDLB_G population: 5.83064 per atom: 0.41647
Total EDDB_G population: 2.13782 per atom: 0.15270
▶ Calculation completed (with 1 warning!) on 2026-10-29. Total time: 33s.
Figure 5. Pentalene (8π, antiaromatic), plane 1 Å above: Jtotal | JEDLB | JEDDB, each on its own scale.
Figure 6. Pentalene: ACIDtotal and ACIDEDLB at 0.05, ACIDEDDB at 0.015 a.u.
Table 1. Benzene and pentalene, B3LYP/def2-TZVPP CSGT, the π-part (-a): π = EDLB + EDDB. Icyc in nA/T, NICS in ppm, populations in e.
benzene: π
EDLB
EDDB
pentalene (R1 = R2): π
EDLB
EDDB
Icyc
11.30
1.31
9.99
−21.18
−15.95
−5.35
NICS(0)
−19.28
−2.23
−17.05
13.22
10.15
3.07
NICS(1)zz
−28.77
−3.33
−25.43
62.07
46.33
15.74
π-population
5.98
0.69
5.29
7.97
5.83
2.14
Conclusion 2. Pentalene’s strong paratropic π-current, −21.2 nA/T, is carried three quarters by localized π-electrons (−15.9); EDDB adds −5.3. Only 2.1 of its 8 π-electrons are delocalized (benzene: 5.3 of 6). The ACIDEDDB surface appears only at a low isovalue, as a thin band along the perimeter. A strong paratropic current is a response of localized electrons: “antiaromaticity” read from a current is not a measure of (missing) delocalization. It is the moral of Tutorial 7, shown as a picture instead of a number.
Part 4 — Should you trust the arrows? The div J table
A current induced by a static field has no sources: div J = 0 everywhere. A finite basis and a finite grid break this a little, and --export-vtk divj measures by how much. The table comes with every 3D export:
benzene_run.out
│ div J OF THE INDUCED CURRENT DENSITY, FIELD n = (0.000, 0.000, 1.000) │
└────────────────────────────────────────────────────────────────────────────────────┘
max|div J| sources valence max valence src
──────── ═══════════ ─────────── ─────────── ───────────
total 47.48944 17.44254 3.09824 11.15972
EDLB_G 33.20238 15.81658 4.34343 8.68754
EDDB_G 13.16691 8.93824 3.48246 5.14549
other 17.73000 14.59154 6.38912 9.46656
EDDB_G(sel) 13.17863 8.73758 3.47916 5.04320
──────── ═══════════ ─────────── ─────────── ───────────
max in nA/T/A^3; sources = sum of div J > 0 x dV = sinks, nA/T; valence: r > 0.5 A
pentalene_run.out
│ div J OF THE INDUCED CURRENT DENSITY, FIELD n = (0.000, 0.000, 1.000) │
└────────────────────────────────────────────────────────────────────────────────────┘
max|div J| sources valence max valence src
──────── ═══════════ ─────────── ─────────── ───────────
total 108.46038 31.92494 5.66675 21.55961
EDLB_G 63.66221 46.14135 14.46700 30.37891
EDDB_G 6.96959 13.15459 4.74317 9.41318
other 42.05529 45.27073 21.04966 33.30009
EDDB_G(sel) 6.84523 12.82912 4.66591 9.32900
──────── ═══════════ ─────────── ─────────── ───────────
max in nA/T/A^3; sources = sum of div J > 0 x dV = sinks, nA/T; valence: r > 0.5 A
Read the total row: it is the error budget. The largest values are always at the nuclei; the valence columns (r > 0.5 Å from every nucleus) are what the pictures show.
The valence sources are of the order of the ring current itself (11.2 nA/T for benzene, 21.6 for pentalene), but they are spread over the whole molecule: the ring current read as the flux through a plane agrees with the one integrated along a contour around the ring (11.3 and 11.3 nA/T for benzene). Trust the circulation pattern; do not over-read a single arrow near a nucleus.
The partial rows add up to more than the total row: the EDLB and EDDB channels exchange current with each other, so each of them alone has sources and sinks. That is a property of the partition, not an error.
Figure 7. div J of the three channels, 1 Å above benzene.
Part 5 — Without an NMR job
A plain geometry is enough for a qualitative picture. runEDDB then computes the current from its GFN2-xTB orbitals (the APPROXIMATE lane, marked in the stage lines):
Figure 8. Coronene from a bare geometry (GFN2-xTB, approximate): Jtotal and JEDDB 1 Å above the plane. Diatropic rim, paratropic hub.
The pattern — a diatropic current around the rim and a paratropic one around the hub — is the one known from DFT, and it finishes in half a minute. The magnitudes are not: the uncoupled xTB response is about 40 % too weak (rim Icyc 10.7 nA/T; see the GFN2-xTB page). Use this lane to look, and an NMR job to measure.
Conclusions
The ring current itself is always read from the flux, Icyc; the pictures tell who carries it.
In benzene the diatropic π-current is EDDB; the inner paratropic loop is EDLB.
In pentalene the paratropic current is mostly EDLB: a strong current, few delocalized electrons.
What to try next
--nmr-mos all,-homo: the current without the HOMO — which orbitals make pentalene paratropic?
--field-axis for a non-planar molecule or a macrocycle (the Manual describes the cost and the choices).
Open .JCUR.vtk in ParaView: Contour on ACID_total, Glyph on J_total.
acid2pov.py --style aicd for the gold-and-green look of the AICD program.
Where to go next. The NMR & currents page of the Manual lists every current-density option and file; Tutorial 7 covers the numbers behind these pictures.