OR WAIT 15 SECS
An interview with Hans-Gerd Janssen of Unilever Research and Development Vlaardingen and the University of Amsterdam. Janssen is involved with method development for gas chromatography, liquid chromatography, and mass spectrometry; the development of instrumentation for multidimensional chromatography for food analysis and biomacromolecular characterization; and problem solving and efficient routine analysis of foods (particularly oils and fats). This interview discusses the challenges encountered in the analysis of fats and oils, how multidimensional chromatography can be applied to the analysis, 2D LC for food-related compound analysis, the analysis of natural antioxidants in edible oils, and future steps in his research.
An interview with Hans-Gerd Janssen of Unilever Research and Development Vlaardingen, in Vlaardingen, The Netherlands. He is also a professor of Biomacromolecular Separations in the Institute of Molecular Sciences at the University of Amsterdam, in Amsterdam, The Netherlands.
Janssen is involved with method development for gas chromatography (GC), liquid chromatography (LC), and mass spectrometry (MS); the development of instrumentation for multidimensional chromatography for food analysis and biomacromolecular characterization; and problem solving and efficient routine analysis in the area of food analysis (particularly oils and fats).
What special challenges are encountered in the chromatographic analysis of fats and oils?
Edible oils and fats are very complex mixtures of many different groups of compounds. Important groups are the triacylglycerols, the partial acylglycerols, the fatty acids, wax esters, sterols and sterol esters, alkanes, phospholipids, and lysophospholipids. Together these groups probably account for over 99.9% of the mass of the sample. In addition to these main compounds there are numerous other compounds and groups of compounds that, although minor in terms of mass, are crucial for the quality of the oil. These include flavor and taste compounds, micronutrients such as vitamins and minerals, and undesired compounds such as residual solvents, polyaromatic hydrocarbons, pesticides, and glycidyl esters. The analytical challenge lies in the fact that all these factors together determine the quality and characteristics of an oil. Methods are needed to analyze gross compositions and to focus on traces of compounds such as off-flavors. A complicating factor is that the questions change almost from one day to the next. Today the ratio of mono- to diacylglycerides has to be measured, tomorrow the level of alcohols or proteins needs to be determined, and the day after tomorrow it is the dissolved oxygen content that should be quantified. This huge variability of questions requires analytical chemists to have a broad range of expertise and to be versatile. The situation is further aggravated by the fact that for some of the above mentioned parameters analytical methods have just recently become available or are not yet available at all. Finally, very often it is not known exactly what needs to be measured. Consider for example the following situation: A margarine shows a very short shelf life and gives rapid development of an off-taste. What is going on here? What do we actually have to measure? This is clearly a challenge that is very different from “just” having to measure specific compounds or compound groups.
So far I have not even mentioned the practical difficulties in the analysis: Edible oils are highly viscous and fats are solids. They stick to everything and do not dissolve very well. Lipids from the skin can contaminate samples, the injector and columns of the chromatograph readily become dirty, many of the contaminants are thermally labile, flavors are easily lost by evaporation, volatiles from the surroundings can be trapped in the oil samples, and components of plastic bottles and caps are extracted by the oil. Many difficulties can occur. The analyst must deal with these problems. Challenges occur in the decision about what to measure, in the selection of the approach, in setting up the analysis, and in actually running it.
How can multidimensional chromatography be applied to the analysis of fats and oils?
Multidimensional chromatography comprises methods where the first dimension is actually a sample preparation step as well as methods where two “real” separation steps are applied.
Sample preparation in edible oil analysis usually means selecting the target group of compounds, while discarding the bulk of the other species present. This group isolation is very nicely done in a normal-phase LC fractionation, exploiting the extreme selectivity of normal-phase LC stationary phases. After the fraction of interest has been isolated, it is sent to the second dimension, usually GC, where the compounds are separated according to chain length, the degree of unsaturation, or both. In multidimensional heart-cut LC–GC this process of group selection with a subsequent detailed analysis can be fully automated. If multiple groups of compounds are of interest, comprehensive LC´GC can be used. With such systems detailed information can be obtained on all compound groups present in just one analysis. This information on which groups are present and, within in a group, which different isomers or chain lengths occur really provides great insight in the composition of an edible oil sample.
In addition to being used for coupled sample preparation and analysis, multidimensional techniques can also be applied to provide greatly enhanced peak capacities. Such approaches are usually not applied to the sample as a whole, but only to specific target groups isolated from the oil or fat. Comprehensive GC´GC, for example, has been widely used for fatty acid methyl ester analysis. The huge number of fatty acids present — especially in animal oils and fats or if cis–trans separation is needed — cannot be separated using single dimensional GC, and thus requires the use of techniques with a much higher peak capacity.
Have you used two-dimensional LC to analyze other food-related compounds?
Whereas reliable instruments for comprehensive GC´GC have been available for more than a decade, instrumentation for comprehensive LC´LC has just recently become commercially available. At the University of Amsterdam, however, we have built several fully automated instruments for comprehensive two-dimensional LC´LC. A very typical food application we have studied is the separation of peptides, specifically in protein hydrolysates from samples such as yogurts. Such peptide mixtures are extremely complex. There is an almost infinite number of peptides present, with large differences in size and polarity. We have applied setups using size-exclusion chromatography in the first dimension with fast reversed-phase LC in the second. In addition we have also applied systems where reversed-phase LC was used in both dimensions. Although one might expect that the use of the same LC mode in both dimensions might not offer much additional peak capacity, the opposite is true. The exact nature of the stationary phase — C18 versus phenyl, for example — as well as the organic modifier and the gradient applied, strongly influence retention of peptides in reversed-phase LC. It is for this reason that a good orthogonality can be obtained in systems using reversed-phase LC in both dimensions. We have also used combinations of reversed-phase LC and hydrophilic interaction liquid chromatography (HILIC) in comprehensive systems for peptide analysis. Whereas these peptide analyses were all done using fully automated comprehensive LC´LC setups, we have also applied at-line comprehensive LC´LC for the separation of polyphenols. In these experiments fractions from an LC separation were collected, either manually or using a fraction collector, and the samples were then put in the autosampler of a second LC system for injection onto another column. Here we usually diluted the sample two- or threefold with water to improve band focusing on the second column.
An article of yours published earlier this year in LCGC (1) described the analysis of natural antioxidants in edible oils — in this case, maize oil — by LC fractionation with off-line detection. GC was used in the study to test the lipid oxidation inhibition of the LC fractions. Have you used the system to test other types of edible oils? If not, could it be applied to other oil types? Where any changes to the original method necessary?
Many natural samples contain antioxidants. To be honest, edible oils are not the first type of samples to look at when you are interested in natural antioxidants. On the other hand, if you are looking for a natural antioxidant to protect edible oils from turning rancid, it is logical to look at how nature itself protects edible oils. That was the main reason why we developed the setup for the identification of the antioxidants in maize oil. We have later used the system extensively to study other sources of antioxidants; these were not oils but extracts of spices and herbs, including the well-known rosemary extract. Whereas the system could be used without any modification for other oils, application to extracts of the spices and herbs required some minor modifications. We installed a guard column to protect the expensive semipreparative normal-phase LC columns and modified the gradient a bit. Actually the system is rather versatile, at least as long as you are looking for nonpolar, fat-soluble antioxidants.
What are the next steps in your research?
I think we, the analytical community, are rather good at measuring target compounds. Tell us what we need to measure and we will be able to do it. But there is some work to do here. In particular we should be quicker. Not in terms in of sample throughput — that is, how many samples can be done in a day — but more in terms of the time it takes to set up the method, have it validated, and present the results of the first set of samples. I think we need more universal methods, methods that have a broader scope and cover more analytes and applications, or methods that are more rapidly “tunable” so that changing to another application is faster. LC´LC and LC´GC could be routes in this.
In addition to developing universal methods, we are faced by an even more difficult task: We need to become better at finding causes for observations — for example, why does a particular oil sample have a much longer shelf life than the other samples? The analytics should not be driven by a series of uncertain hypotheses, but analytics should generate the hypothesis. Fingerprinting and profiling approaches are what we need here. Highly stable, highly detailed fingerprints, combined with good chemometrics is what we have to develop.
(1) R. Poort, H. Steenbergen, and H.-G. Janssen, LCGC No. America31(3), 232–239 (2013).