One of the most remarkable achievements of modern chemistry is the capability to draw accurate structural formulas for complex molecules. Considering that molecules are much too small to be seen by even the strongest optical microscopes, how is it that the connectivity and relative location of the component atoms can be assigned with confidence? Prior to 1950, molecular structures were deduced largely by chemical degradation of large complex molecules into smaller known fragments. The terpene hydrocarbon myrcene provides a simple example, as shown in the following diagram.
Combustion analysis yielded the ratio of carbon to hydrogen by measuring the amount of carbon dioxide and water produced, and a rough molecular weight was obtained from boiling point studies. The double bond functional groups were assayed by catalytic hydrogenation, and cleavage by ozonolysis yielded the known small molecules formaldehyde and acetone. Further oxidation formed succinic acid with the loss of carbon dioxide. The logical assembly of these fragments into the structure of myrcene, essentially a chemical jigsaw puzzle, required perseverance and the insight of experience on the part of early chemists.
If this were the present status of structure determination, knowledge of the characteristic reactivity of the common functional groups, and the nuances of their interactions with each other would be a minimal requirement for achieving a basic understanding of this field. Over the past fifty years, however, a more easily applied and interpreted set of tools for this purpose has been developed and are now widely used by chemists for the elucidation of molecular structures. Most of what we know about the structure of atoms and molecules comes from studying their interaction with light (electromagnetic radiation), a subject called spectroscopy, and a full introduction to spectroscopy will be found elsewhere in this text. Fortunately, spectroscopic applications to molecular structure are sufficiently straightforward that they can be appreciated without delving into the exact nature of the methods. In the following discussion, information provided by three different spectroscopic tools will be described.
To begin, consider the three isomeric C5H12 alkanes, molecular weight = 72. These are all low boiling liquids lacking reactive functional groups. Despite their similarity, they are easily distinguished by the number of structurally distinct carbon atoms in each, as disclosed by 13C nmr. In the following diagram equivalent carbon atoms are identified by color. The numbers next to each different carbon are in a sense magnetic addresses, called chemical shifts, produced in the course of the spectroscopic measurement. Because these chemical shifts are influenced by the full electron and nuclear distribution in each structure, identical values are seldom observed, even for similar kinds of carbon atoms (e.g. 1º, 2º, 3º or 4º). The nmr signals that identify each set of structurally identical carbons have different intensities, which may vary with the experimental conditions. Although these intensities are roughly proportional to the number of atoms of a given kind, other factors are influential. Consequently, 13C nmr provides a very good qualitative count of different kinds of carbon atoms in a molecule, but a poor quantitative count of each.
Introduction of a functional group into a structure usually increases the number of possible isomers. This is illustrated below for C5H10 and C5H8 hydrocarbons, which may incorporate carbon double and triple bonds as well as rings. All the isomers for a given formula are not shown, and each structurally unique carbon is designated by color only once in each structure. Double and triple bond functional groups have characteristic infrared signatures that are often used for identification. For purposes of this discussion, however, these are not necessary and the presence of such functions is easily established by addition of hydrogen, see the myrcene example above. These functions may also be identified by their characteristic 13C nmr chemical shifts. The five C5H10 alkene isomers all display the same number of 13C signals; nevertheless, they can be distinguished in part by their catalytic hydrogenation to either pentane or 2-methyl butane (shown above). Further identification may then be achieved by the ozonolysis reaction used in the myrcene case. Although the stereoisomeric 2-pentenes (2nd and 3rd structures in the top row) cannot be distinguished in this way, it is worth noting that two or more of their 13C signals are clearly different. The chemical shift values for the sp2 carbons in alkenes and the sp carbons in alkynes are not only different from the sp3 carbons of alkanes, but the magnitude of the difference is unexpected. An explanation for this will be found in the nmr spectroscopy section, but is not important for our present application of this technique.
Application of 13C nmr spectroscopy to a wide variety of organic compounds has confirmed the value of this tool for structure determination. By clicking on the above diagram, additional examples will be displayed. The top row shows isomers in which an electronegative chlorine is attached to a butyl group. In each isomer the carbon bearing the chlorine has a significantly larger chemical shift than the others. The first two compounds cannot be distinguished by the number of carbon signals; however, a careful analysis of the chemical shifts permits an assignment to be made. The second row shows three isomeric xylenes (dimethylbenzenes), and for comparison benzene itself and toluene (in the gray shaded box). Structural assignment of the isomers is easily made from the number of distinct carbon signals displayed by each. The chemical shifts of the sp2 carbons are roughly the same as in the alkenes. Finally, some isomeric oxygen compounds are listed in the bottom row. These include both unsaturated and cyclic compounds, as well as ether, alcohol and carbonyl functional groups. Infrared spectroscopy provides an excellent detector for such functions, and some characteristic absorptions observed for these compounds are given (in reciprocal centimeters, cm-1) below each structure. Once again, however, the 13C data are almost sufficient for a complete structure assignment. If a simple chemical aldehyde test is used to distinguish the first and second compounds this ambiguity is removed.
The following table outlines some general conclusions concerning the relationship of chemical shifts to carbon bonding that have become apparent from numerous measurements.
The importance of facile spectroscopic structure determination in organic chemistry may be illustrated by the vapor phase chlorination of 2,4-dimethylpentane. This reaction, shown by the following equation, generates three isomeric monochloro compounds. If all the C–H bonds were equally susceptible to this free radical substitution reaction the 1-chloro isomer would predominate by the statistical 6:1:1 ratio over the others. However, when the reaction is actually carried out at 100 ºC, the other isomers are formed in greater than expected amounts.
Experiments of this kind were instrumental in demonstrating that 3º-C–H and 2º-C–H bonds had increasingly lower bond dissociation energies relative to their 1º-analogs. Once the isomers were separated by fractional distillation, it was necessary to determine which was which. They all have the same molecular weight, but differ in the number of 13C signals. Inspection of their formulas indicates that A will have six signals, B will have three and C five signals. Predicting the chemical shifts is possible but not necessary for the structure assignments.