Gas Chromatography
Another method for separation of organic compounds is gas chromatography (also called gas-liquid
chromatography, gc, and glc). This is an instrumental method that requires that the compounds to be separated are
volatile and do not decompose with heating. Just as in extraction, this methodology requires the organic compound to
distribute itself between two “layers” or phases. One phase is a stationary liquid and the other is a moving vapor
phase.
Stepwise, the procedure goes like this. The compounds to be separated are injected, by means of a syringe with
a very small bore, through a septum in the injection port and into a moving stream of carrier gas. The compounds are
then carried into the column, which is the heart of the separation. The column is either walled coated with the
stationary phase or packed with what inert beads of diatomaceous earth which have been coated in the liquid phase. As
the compounds to be separated reach the column they will distribute themselves between being in the gas phase and
being dissolved in this liquid phase. Of course, that suggests that the nature of liquid is very important to the
separation, as “like dissolves like.” That which does not dissolve immediately moves on farther down the column
where it encounters additional liquid phase. That which does dissolve will soon be re-distributed between “clean”
carrier gas the liquid phase. This repetitive distribution creates a separation. Things that are the most soluble in the
liquid phase will cling to the stationary phase longer and take longer to move down the column. Things that are not as
soluble in the liquid phase will spend more time in the mobile gas phase and move down the column faster. In either
case, all compounds will eventually exit the column. How do we know when they exit? Since most organic compounds
do not have color, we need something other than visible monitoring. One of the most common methods is a detector
called a thermal conductivity detector. The detector “senses” the presence of the now separated compounds in the
carrier gas and sends an electric signal to some type of recording device as long as there is anything other than carrier
gas in the detector.
The sample size is very small, even for most packed columns. For the 1/8 inch diameter packed columns in the
instrument you will be using, the suggested volume of sample is 0.5 microliter. In other words, one drop of liquid from
an ordinary pipette or burette would contain enough for about 50 injections. The injection port must be heated to
ensure that the entire sample goes into the column in the gas phase. If there is any material in the injection that will not
vaporize, the instrument may be fitted with a glass sleeve inside the injection port which will eventually plug-up with the
residue but can be replaced.
The carrier gas must be pure, inert, and preferably have a low molecular weight. The carrier gas for this lab is
helium which is chromatographic grade (pure – dry) which has a molar mass of 4.0026m/mole. The only possible carrier
gas of lower molecular weight is hydrogen (molar mass = 2.0159gm/mole), and it is not completely inert as it is known to
react with pi bonds in the presence of metal catalysts (most columns are made of copper metal). The importance of
molar mass will become clearer when we discuss the detector.
The column, as described above, has two working components – the liquid phase and the inert beads. The
beads should be inert, mechanically strong, and be of similar sizes. While diatomaceous earth is neither inert nor pure,
it can be processed to meet our needs. If it is acid washed, metallic impurities have been removed but there are still
some reactive –OH groups present. These –OH groups are chemically converted to more inert groups which also create
a greater physical strength in the beads. This treated and washed diatomaceous earth is then sieved to ensure a fairly
narrow range of diameters to ensure even packing. The liquid phase, if we were to observe it at room temperature,
might appear more like a wax or very viscous axle grease. Remember, however, that the column is heated and as
temperature increases, viscosity decreases. To prepare the packing for a column, one decides what percentage by
weight of the packing should be the liquid phase. Since the compounds to be separated are dissolved in the liquid for
part of their time in the column, the greater the percent by weight of liquid, the longer the compounds spend in the
column. The amount of liquid is sometimes referred to as the percent liquid loading. What are our criteria for these
liquids? First, they should have a fairly high boiling point so that they will not boil off the column while you are
running the separation. Therefore, most columns come with instructions that they are not to be heated over a certain
temperature for that very reason. Second, the liquid should be such that the compounds you wish to separate have a
different solubility in that liquid. If we use a non-polar stationary liquid for two polar compounds, they will probably
spend very little dissolved time and, therefore, will exit the column (elute) at very similar times. However, if we use a
stationary liquid that is moderately polar (such that one compound is more soluble than the other is) the compound
which is the more soluble will spend the most time dissolved in the stationary phase and elute last. Therefore, the
liquid components of the column and the liquid loading are generally written on a tag attached to the column so that
we can select what we need.
There are two basic types of simple detectors. The universal detector is the thermal conductivity detector
which responds to how efficiently the carrier gas cools a heated filament. The sample is not destroyed and can be
collected at the column exit (if you are very patient). Another class of simple detectors is called ionizing detectors.
They convert a portion of the compounds eluting into ions and measure the flow of ions between charged plates at
the exit of the column. Some ionization detectors create ions by burning the sample in a flame and others by passing
the carrier gas over a radioactive source, which provides electrons to be “stuck” onto the sample molecules. These two
types of detectors simply create a signal when any compound other than the carrier gas passes through the detector.
They do not give any information about the chemical nature of these compounds. More modern equipment can include
detectors such as a mass spectrometer, which not only create a signal when the compound passes but also give
structural information about the compound.
How does a thermal conductivity (or TC) detector work? A TC detector is a Wheatstone bridge composed of 4
resistors in a circuit. Two resistors have carrier gas flowing over them at all times. The other two have the gas coming
off the column passing over them. Initially, the circuit is balanced so that the resistance on both sides of the detector (
carrier gas only and carrier gas + eluants) is equal and the amount of current passing through each side is equal. Physics
tells us that resistance changes as temperature changes; therefore, as long as both sides have only carrier gas moving
over the filaments, the resistance remains equal. However, the smaller a molecule is, the more efficiently it cools the
filament. Therefore, when a compound other than He enters one side of the detector, the filament heats up, and it’s
resistance changes. This change in resistance causes a different amount of electricity to pass through one side as
opposed to the other. This difference in electric current can be measured and an electric signal sent to a recording
device. Clearly not all compounds cool the filament equally well. Therefore, equal amounts of compounds A and B may
give a different size signal from the detector. In order to quantify the amount of material represented by a signal, one
simply runs the same kind of calibration curve ( mass vs. magnitude of signal) you have run for the Beer-Lambert Law.
For this lab, the electronic signal is being sent to an integrator which will not only give us a plot of millivolts vs
time, but will also do the integration of the peaks to give the area under the peak. It is the area under the peak which is
correlated with the amount of compound, not peak height. Each component will be identified by its retention time –
the amount of time between the injection of the sample and the maximum detector response (top of the peak).
Retention time is usually given in minutes or seconds, but when it is measured from a chart record, it can also be
given in units of distance i.e. cm or mm. That is because the chart paper is moving at a constant rate. Therefore, a
retention time of 1 minute could also be expressed as 1 cm if the chart paper is moving at 1 cm per minute.
Since we could use several different stationary liquid phases, we sometimes need a standard measure of how
well two compounds are separated on a given column. The measure of the efficiency of separation is called resolution.
Resolution is measured for two peaks of comparable size as follows:
Resolution =
(the distance between the peak maximum in mm or cm)
½ [ width of peak 1at the base + width of peak 2 at the base]
If the distance between the peak maxima is equal to the average of the peak width, we say that there is baseline
separation and the resolution is 1. Resolutions of less than 1 may require changing the column. Resolutions of more
than 1 are certainly acceptable, but you may choose to change the column so as to reduce the resoltion number and
spend less time per injection.
As we use columns, they change. Sometimes very small amounts of the liquid phase are slowly baking off the packing.
Sometimes compounds we inject stick and don’t come off. When things like this happen, the efficiency of a column changes; it’s
ability to separate compounds changes. When this change reaches some specified value (pre-determined by the user and procedure),
the column must be replaced. In order to measure the “health” of the column, chromatographers frequently measure the number of
theoretical plates. For now let’s define a theoretical plate as one equilibrium between being in the liquid phase and being in the gas
phase. That’s not exactly right, but it will suffice for now. If we inject a known compound at a given carrier gas rate and temperature,
we can measure the Theoretical Plates available from the column as follows:
Theo. Plates = 16 * [ retention time / width of peak at the base]2
While a distillation such as you ran would probably represent less than 100 plates, gas chromatography frequently will
exhibit several hundred to several thousand plates. Another related technique called high performance liquid chromatography
(HPLC) which uses a second liquid phase rather than the inert carrier gas, frequently has hundreds of thousands of plates.