Many compounds -- and in particular, many compounds of biological interest -- contain more than one chiral atom. It can be very useful to be able to talk about the relative configuration of various asymmetric centers even in cases where the absolute configuration is not known. Consider the case of tartaric acid:
 |
 |
 |
 |
| (R,R)-tartaric acid A |
(S,S)-tartaric acid B |
meso-tartaric acid C |
tartaric acid D |
A known, specific, single configuration could be depicted using diagram A, B, or C. Diagram D should be used when nothing is known about the absolute configuration of the two chiral atoms; a substance represented by D might, on further analysis, prove in fact to be A, B, or C, but it might also prove to be any combination or mixture of the three..
Those diagrams can depict many common situations, but not all. For example, it is common to have a substance containing a racemate of the (R,R)-form A and its (S,S)-enantiomer B, but definitely not containing any of the meso-diastereoisomer C. In that case, the relative configuration of the two stereocenters is known (both "same" according to CIP rules, or both "different" according to the specific use of hashes and wedges above), but the racemic quality cannot easily be depicted without additional descriptive text.
That racemic mixture might be further characterized, and the two enantiomeric forms crystallized separately. At that point, each separate enantiomer has a known relative configuration, but it still might not be known which crystal actually contains the (R,R)-form and which contains the (S,S)-form. There is no easy way to depict this situation either.
Tartaric acid is a very simple example, with only two chiral atoms. The same problems are only multiplied in structures that contain more than two. As a still-simple example, a structure with four asymmetric centers might be known to have a specific relative configuration between the first two centers, a different relative configuration between the last two centers, but have nothing known about the relation between the first two and the last two.
It can be useful to discuss relative configurations even in a compound with a single chiral center; in that case, the relative stereochemistry would be in relation to some quantity that might contain many moles of individual molecules. Relative configurations of this sort play an important role in chemical syntheses, where a starting material and its reactive transformations are known. Reduction of an asymmetric carbonyl such as butan-2-one will produce a sample that is (probably) racemic. That racemate can be resolved, for example by chiral chromatography, into two fractions of known relative configuration, but unknown absolute configuration. One of those samples may be further analyzed, possibly via X-ray crystallography, to determine its absolute configuration. In such a sequence, the carbonyl moiety may itself be part of a larger structure that has other stereocenters, and each of those other stereocenters may be identified in racemic, relative, or absolute terms. This procedure can quickly become very complex.
There are a variety of obscure conventions for indicating relative configuration. The most common (but still obscure) of these conventions says that bold and hashed wedges should be used for absolute configuration, unwedged bold and unwedged hashed lines represent relative configuration and racemic character, while hollow wedges and dashed lines represent relative configuration only. [Maehr, H. A proposed new convention for graphic presentation of molecular geometry and topography. J. Chem. Educ., 1985, 62, 114-120.] [Maehr, H. Graphic Representation of Configuration in Two-Dimensional Space. Current Conventions, Clarifications, and Proposed Extensions. J. Chem. Inf. Comput. Sci., 2002, 42, 894-902.]. In 2002, Elsevier MDL added extensive support for relative tetrahedral configurations to their ISIS system. A white paper on this subject is currently available at http://www.mdl.com/solutions/white_papers/downloads/public/Enhanced_Stereochemical_Representation.pdf. Other conventions have been used by various other authors and by various chemical software applications. Other conventions have been used by various other authors and by various chemical software applications. Yet other conventions have been used to describe relative configuration in nomenclature [E. L. Eliel, S. H. Wilen. Stereochemistry of Organic Compounds. pp. 117-124. John Wiley & Sons, Inc., New York (1994).], although generally with little direct application to depiction styles. At this point, there does not seem to be any consensus for a preferred depiction style. This is an unresolved area. Given the lack of consensus among different systems and the poor general acceptance of any given one, the safest approach currently is to depict multiple structures explicitly, optionally accompanied by additional descriptive text as discussed below.
The goals of structure representation here are the same as elsewhere: it is important to indicate clearly what chemical substances are intended by the authors. In the case of relative configuration, it is critical not to conflate unknown, unspecified, and racemic in any way. For a given center there are at least the following values:
Unfortunately, it is extremely difficult to represent all of these different possibilities in a legible manner simply by using variations of bond types (the Maehr proposal) or atom-based indicators (the Elsevier MDL approach).
In the absence of any other sort of indicators, a structural diagram is expected to represent the specific substance drawn. If stereochemical hashed wedged and/or solid wedged bonds are present in the diagram, it will represent a single stereoisomer. If no stereochemical information is present in the diagram, it represents a substance with unknown configuration.
This recommendation is at odds with the historic (last 20-25 years) behavior of the MDL family of software (ChemBase, MACCS, ISIS/Draw, ISIS/Base, etc.). In MDL software, diagrams were assumed to represent a racemate of the absolute configuration drawn and its enantiomer unless an additional "chiral flag" was added explicitly. The present recommendations, however, correspond to prevailing usage within printed chemical communications such as journal articles, books, patents, and so on. More recent software from MDL has also moved in the same direction.
Other types of relative configuration should be represented with a more verbose description than a single structure. In some situations, it might be acceptable to list separate depictions for all stereoisomers present. In printed materials, it would usually be appropriate to have some additional text, such as the word "and" present, to indicate that the two structures were both intended. It is also permissible to depict a single stereoconfiguration with a longer piece of descriptive text, such as "and enantiomer", or "(±)-form".
| ACCEPTABLE | ACCEPTABLE | ACCEPTABLE |
 |
 |
 |
Non-alphabetic symbols should not be used to separate multiple depictions. In particular, a plus sign should not be used to replace the word "and" since the plus sign already has several other meanings (as a charge, or as a separator of multiple reactants or products in a reaction) that could cause further confusion..
| NOT ACCEPTABLE | NOT ACCEPTABLE | NOT ACCEPTABLE |
 |
 |
 |
The general recommendations for placement of annotations should be followed when positioning any such explanatory text. In particular, the explanatory text should be positioned no further from the structural diagram than a distance equal to the median bond length used in the diagram. In no case should it be positioned closer to some other diagram than the one to which it applies. Positioning of the explanatory text below or to the right of the structure is preferred, but such positioning may conflict with other symbols including reaction arrows and compound numbers in journal articles. If such positioning would possiblye create any ambiguity or confusion, any other positioning is also acceptable.
The recommendations presented here are primarily focused on the depiction of substances that can be represented with a single structure; a comprehensive discussion of arbitrary mixtures of substances is outside the scope of this document.
The use of the terms "racemate" and "rac" should be avoided because it has historically had several different meanings. The most common historical meanings appear to be "a 50:50 mixture of enantiomers" and, colloquially, "an equilibrium mixture of enantiomers"]. To ensure understanding, the clearest and most accurate description should always be used.
The terms "relative" and "rel" should similarly be avoided unless accompanied with a clear explanation of what the structure is relative to. It remains acceptable to use the term "relative" in a generic sense, as it has been used in these recommendations.
| PREFERRED | NOT ACCEPTABLE | PREFERRED | NOT ACCEPTABLE |
|
|
 |
 |
 |
