Seeing Structure: Spectroscopy and Instrumental Analysis

Recall that an atom can only absorb or emit photons whose energy exactly matches a jump between its energy levels. In emission spectroscopy, atoms are excited (in a flame or plasma) and the wavelengths of light they give off as electrons drop back down identify the elements. In absorption spectroscopy, light passes through the sample and atoms in the ground state remove the wavelengths that match their jumps. Atomic absorption spectroscopy (AAS), developed by Alan Walsh in Australia in the 1950s, uses this to measure a single metal at parts-per-billion levels. To test drinking water for lead, the sample is sprayed into a flame (or heated in a graphite furnace) to break it into free atoms. A lamp whose cathode is made of lead emits lead's own spectral lines, such as 283.3 nm, and shines them through the atoms. Only lead atoms can absorb those exact wavelengths, so other metals do not interfere. The detector measures how much light is lost, and the absorbance is proportional to the concentration of lead. Standards of known concentration give a calibration line, from which the sample's concentration is read. Canada's limit for lead in drinking water is 5 µg/L.

Infrared photons have too little energy to move electrons, but just the right energy to make bonds vibrate. Think of a bond as a spring joining two balls: it stretches and bends at a natural frequency, and it absorbs IR light of exactly that frequency. Stronger (stiffer) bonds and lighter atoms vibrate faster, so they absorb at higher frequency. IR spectra use the wavenumber, the number of waves per centimetre ($\tilde{\nu }=1/\lambda$, in ${\text{cm}}^{-1}$), which is proportional to frequency and energy. That is why a triple bond absorbs at a higher wavenumber than a double bond, which absorbs higher than a single bond, and why bonds to light H atoms (O–H, N–H, C–H) appear above 2800 ${\text{cm}}^{-1}$. The region above about 1500 ${\text{cm}}^{-1}$ identifies functional groups. Below it, the fingerprint region is a complex pattern unique to each compound, matched against a database to confirm an identity. Molecules whose vibrations absorb IR are also why ${\text{CO}}_2$ and ${\text{H}}_2\text{O}$ are greenhouse gases.

A mass spectrometer vaporizes a tiny sample and bombards it with high-energy electrons, which knock an electron off each molecule to form a positive molecular ion, ${\text{M}}^{+}$. Many of these ions break into fragments. All the ions are accelerated by an electric field and separated by their mass-to-charge ratio (m/z), either by bending them in a magnetic field (lighter ions bend more) or by timing their flight down a tube. The spectrum shows relative abundance against m/z. The molecular ion peak, usually the highest m/z, gives the molar mass. The tallest peak, the base peak, is set to 100%. Fragments reveal pieces: a loss of 15 means a ${\text{CH}}_3$ group, a loss of 17 an OH. Isotopes leave fingerprints too. Chlorine is about 75% chlorine-35 and 25% chlorine-37, so a compound with one Cl atom shows peaks at M and M+2 in a 3 : 1 ratio; bromine (about 50% bromine-79 and 50% bromine-81) gives M and M+2 of equal height. Each carbon has a 1.1% chance of being carbon-13, so the small M+1 peak reveals how many carbons there are. Anti-doping labs use mass spectrometry to detect banned drugs at billionths of a gram.

Some nuclei, including hydrogen-1 and carbon-13, behave like tiny magnets. In nuclear magnetic resonance (NMR) spectroscopy, the sample sits in a strong magnetic field, where these nuclei line up with or against the field, and radio waves of just the right frequency flip them. The exact frequency depends on the electrons around each nucleus, so hydrogens in different chemical environments give separate signals, and the area of each signal is proportional to the number of H atoms producing it. Ethanol, ${\text{CH}}_3{\text{CH}}_2\text{OH}$, gives three signals with areas 3 : 2 : 1. Magnetic resonance imaging (MRI) applies the same physics to the hydrogen nuclei in the water and fat of the body, adding magnetic field gradients so each signal can be traced to a location. It uses radio waves, not ionizing radiation; Paul Lauterbur and Peter Mansfield shared the 2003 Nobel Prize in Physiology or Medicine for it.

X-ray crystallography finds the positions of every atom. X-rays have wavelengths of about 0.1 nm, similar to the spacing between atoms, so a crystal's regular array of molecules diffracts them into a pattern of spots. Mathematics converts the pattern into a map of electron density showing where each atom sits. Dorothy Crowfoot Hodgkin solved the structures of penicillin (1945), vitamin B12 (1956) and, after more than 30 years of work, insulin (1969), winning the 1964 Nobel Prize in Chemistry. Rosalind Franklin's X-ray images of DNA fibres, especially Photo 51 (1952), revealed its helical shape and dimensions. They were shown to James Watson without her knowledge and were essential to the double-helix model; she died in 1958, before the 1962 Nobel Prize was awarded to Watson, Crick and Wilkins.