Evolution of Petroleum Assays During the Last 50 Years

Assays to characterize properties of petroleum precede the days of instrumental analysis. The fluorescence indicator assay (FIA) was one of the first. It was used to separate petroleum and process feeds by adsorption chromatography operating in the displacement mode. The sample containing saturates (alkanes and naphthenes), olefins, and aromatics was poured over a bone dry silica column in a quartz tube. The analytes separate according to their increasing affinity for activated silica. Saturates move the fastest, followed by olefin, aromatics, and compounds containing heteroatoms (asphaltenes). The position of the displacement bands was measured with UV light. The hydrocarbon composition was determined by measuring the percentage of the wetted bed length attributed to the hydrocarbon type. Although it was not very precise, it was the best available technology that provided useful numbers to production and refinery operators.

Fast forward 60 years: Assays are faster and there is higher resolution. For example, a team at Schlumberger DBR Technology Center (Edmonton, Alberta, Canada) has developed an automated microfluidic apparatus for the rapid assay of asphaltenes in petroleum samples.1 The absorbance of the total sample is measured at 600 nm. The sample is then diluted 40× with heptane, which precipitates the asphaltene fraction. Absorbance of the hexane-soluble fraction, maltene, is recorded and subtracted from the total to the concentration of asphaltenes. Typical cycle time for the automated system is 300 sec. Prior assay protocols required a few days. The microfluidic device can be cleaned and reused. Schlumberger has made several for internal use. The firm plans to make a commercialization decision in the fall of 2013.

At the 10th GC×GC Symposium held at the Renaissance Hotel in Palm Springs, CA (May 12–14, 2013), many of the applications showed analysis of petroleum feeds and process streams. A few papers also reported assays of biofuels. GC×GC is also known as Comprehensive GC and 2-D GC and its little brother LC×LC. Both are designed to increase the peak capacity of a separation by using two columns with orthogonal selectivity. This is analogous to 2-D electrophoresis, which has been used by the proteomics set for about 50 years. In 2-D electrophoresis, proteins are separated according to molecular weight and pI.

A typical experiment involves a GC fitted with a long capillary column. The selectivity is usually nonpolar, which provides a separation that correlates with boiling point. Indeed, this is the basis for “simulated distillation” assays. However, as the boiling point increases, the molecular complexity rapidly expands, creating peak overlap, and even coelution.

In practice, GC×GC adds peak capacity and hence improves resolution by connecting a short column to the long column with a modulator, which traps segments of the effluent from the long column and flashes these onto a second column for rapid separation. The result is a 2-D plot of peaks that has much greater peak capacity. Simplistic theoretical estimates are the peak capacity of the first multiplied by the capacity of the second. Typical numbers are 400 for the first dimension and 10 for the second. In practice, one only achieves about 30–50% of the expected improvement since much of the theoretical space is empty. For example, there are few low boiling polar analytes. Analytes with very high polarity do not elute from the second column.

Mass spectrometry adds still another dimension, hence the name 3-D GC. This also increases the effective peak capacity by a factor of 10. In numbers, we have a capacity of 400 for the first column times 10 for the second, and another 10× for the third, for a total possible capacity of 40,000 and a practical capacity of 10,000–20,000.

How useful is this? Dr. Tadeusz Gorecki’s lecture at the GC×GC meeting cited the criteria of Calvin Giddings:

1. “Peak resolution is severely compromised when the number of components present in a sample exceeds one-third of the peak capacity” and

2. “In order to resolve 98% of the components, the peak capacity must exceed the number of components by a factor of 100.”

Thus a useful GC separation of petroleum would be limited to about 200 components. This is low on the scale of things, so insufficient peak capacity is an issue.

The two-day symposium covered a wide range of experimentally important details and applications other than petroleum. However, the message is clear: GC×GC×MS is today a viable experimental approach for the assay of volatile analytes in lipids, aromas, petrochemicals, environmental applications, and fuels (both petro and bio).


  1. Schneider, M.H.; Sieben, V.J. et al. Measurement of asphaltenes using optical spectroscopy on a microfluidic platform. Anal. Chem.2013, 85, 5153–60.
  2. Giddings, J.C.; Davis, J.M. Statistical theory of component overlap in multicomponent chromatograms. Anal. Chem.1983, 55, 418.
  3. Giddings, J.C. Sample dimensionality: a predictor of order-disorder in component peak distribution in multidimensional separation. J. Chromatogr. A1995, 703(1–2), 3–15.

Also see http://www.americanlaboratory.com/913-Technical-Articles/140493-37th-International-Symposium-on-Capillary-Chromatography/.

For more on gas chromatography, please see http://www.americanlaboratory.com/913-Technical-Articles/136800-Advances-in-Gas-Chromatography-Instrumentation-at-Pittcon-2013/.

Robert L. Stevenson, Ph.D., is a Consultant and Editor of Separation Science for American Laboratory/Labcompare; e-mail: rlsteven@yahoo.com.