Wedge and Dash Stereochemistry Drawing Conventions

You drew the active enantiomer, exported a SMILES for the assay request, and the computational chemist came back asking why the @/@@ descriptor pointed at the wrong configuration. Nine times out of ten the structure was right in your head and wrong on the canvas — a wedge on the wrong bond, or two wedges on one center leaving the configuration undefined. The wedge and dash stereochemistry drawing conventions are simple, but they only do their job when the 2D drawing, the R/S label, and the SMILES stereo descriptor all agree. This is a step-by-step on drawing them so all three line up.

Wedge, dash, and plain line: what each bond means

Three bond styles encode the third dimension on a flat drawing:

• Bold wedge — the bond points toward the viewer, out of the plane of the screen.
• Dashed (hashed) bond — the bond points away from the viewer, behind the plane.
• Plain line — the bond lies in the plane of the screen.

At a tetrahedral stereocenter you typically draw two plain bonds in the plane, one wedge coming forward, and one dash going back. That arrangement fully specifies the spatial layout of the four substituents. The reader (and the software) reconstructs the 3D geometry from those four directions, so the four bonds need to be unambiguous about which way each substituent goes.

Where the narrow end of the wedge goes

This is the convention that gets drawn backwards most often. The narrow (pointed) end of the wedge or dash sits at the stereocenter; the wide end is at the substituent atom. The point is the apex at the chiral atom, fanning out to the group that is coming toward or going away from you.

One stereocenter needs exactly one out-of-plane bond to be defined relative to the in-plane bonds. In practice chemists draw one wedge and one dash for visual clarity, but a single wedge against three plain bonds is also a complete specification — the geometry of a tetrahedron means the remaining bond direction is implied. This is the basis of the published one-wedge convention for unambiguous stereo drawings.

Make the drawing agree with R/S

R/S is the absolute, viewer-independent label for a stereocenter, assigned by the Cahn–Ingold–Prelog (CIP) priority rules. To check that your wedge/dash drawing produces the configuration you intend:

1. Rank the four substituents by CIP priority: highest atomic number first at the first point of difference, then work outward along each branch when the first atoms tie. Multiple bonds count as duplicated atoms.
2. Orient the lowest priority away from you. If the lowest-priority group is already on a dash, you are looking down the correct axis. If it is on a wedge, you are looking from the wrong side — determine the rotation and then reverse your answer.
3. Trace highest → second → third. Clockwise is R; counterclockwise is S.

The drawing and the label have to match: if you meant to draw (S) but the CIP trace comes out clockwise with the lowest priority pointing back, the wedge is on the wrong bond. The full CIP procedure with worked rotations is laid out in the LibreTexts CIP R-S sequence rules. For double bonds the analogous label is E/Z, assigned by comparing the CIP priority of the groups on each end of the double bond — higher priorities on the same side is Z, opposite sides is E.

Make the drawing agree with the SMILES descriptor

A SMILES string encodes the same stereochemistry in text. Tetrahedral configuration is written with @ or @@ after the chiral atom, and double-bond geometry with the / and \ slash bonds. Per the Daylight SMILES specification, @ means the listed neighbors are arranged anticlockwise when viewed with the first neighbor toward you, and @@ means clockwise. Critically, @/@@ are defined by the order the atoms are written in the SMILES, not by CIP priority — so there is no fixed @-equals-R rule. The descriptor and the R/S label can both be correct while looking unrelated.

The thing that must hold is consistency: a 2D wedge/dash drawing and the SMILES stereo descriptors have to describe the same molecule. The reliable way to keep them in sync is to let the toolkit do the conversion rather than hand-translating — draw the wedge/dash on the canvas and read back the canonical SMILES, or paste a SMILES and confirm the wedges render the way you expect. RDKit, for instance, assigns CIP labels from either the molfile wedges or the SMILES parity and will tell you when they disagree; the mechanics are documented in the RDKit Book stereochemistry section. If you are starting from a SMILES and want to see whether its stereo matches your intent, turning a SMILES into a structure in the browser shows the wedges the descriptor actually implies. The distinction between a plain canonical SMILES and one that carries stereo is worth understanding directly — see canonical versus isomeric SMILES for how the stereo layer rides on top of the connectivity string.

Common mistakes (problem → fix)

Two wedges on one stereocenter

Problem: Drawing two bonds from the same chiral atom as wedges (or two as dashes) makes the configuration ambiguous — you have specified two groups coming forward and not enough constraint on the geometry, so a parser either rejects it or guesses. Fix: Use exactly one out-of-plane bond per stereocenter as the defining bond (one wedge, with the other three as plain lines), or the conventional one-wedge-one-dash pair. Never two of the same kind on one center.

Wide end of the wedge at the stereocenter

Problem: The wedge is drawn fat-end-first at the chiral atom, which can reassign the stereo to the wrong atom on import. Fix: Always place the narrow point at the stereocenter and the wide end at the substituent.

Drawing meant nothing — flat structure for a chiral molecule

Problem: All plain bonds at a known stereocenter. The molecule is chiral but the drawing is undefined, so the exported SMILES carries no @/@@ and downstream tools treat it as racemic or unspecified. Fix: Add the defining wedge or dash; if you genuinely mean a mixture, say so explicitly rather than leaving it implied by an undefined drawing.

Drawing and SMILES disagree after a hand edit

Problem: You edited the canvas after exporting a SMILES, and now the text string describes the old configuration. Fix: Recompute identifiers from the current structure after every edit — the canonical SMILES, InChI, and InChIKey should always be read back from the molecule as it stands on the canvas, not from a cached earlier copy.

The workflow that keeps all three representations honest is to draw the wedge/dash on a canvas, let the toolkit assign the CIP label, and read back the stereo-bearing SMILES — then verify the label matches your intended R/S. ChemStitch draws single, double, triple, wedge, and dash bonds in the browser and validates the structure with RDKit, so a defined stereocenter round-trips to a SMILES with the correct @/@@ rather than silently dropping the configuration.