The scale of the problem
In the blog 2 of this series, we looked at the rationale behind the approach to measure mineral oil contamination from foods and packaging. This involves firstly fractionating MOSH and MOAH components using normal phase HPLC and then analysing each pooled fraction using Gas Chromatography coupled with flame ionisation detection (GC-FID). The aim of this is approach is not to measure any specific component in each pooled fraction (as there are potentially 1000s of molecules in each), instead the aim is to reliably detect and quantitate fractions of these as ‘slices’ of MOSH and MOAH humps. These resultant slices represent certain oil classes which elute in specific retention time windows.
In the previous blog, we also explored the fact that the food matrices are complicated and heterogeneous in composition and so effective sample preparation is challenging. We also discussed that some samples are dry and also depleted of oil and fats - such as found in certain dried foods and also from food packaging. On the other hand, other samples may be anhydrous but are rich in fats and oils, or others may be 100% oil such as food grade oils. Finally, other matrices may be both aqueous and some have a high percentages of natural fats and oils.
The wide heterogeneity of sample matrices means that pre-sample preparation is often conducted to remove an excess polar interferences and natural oils, which otherwise may cause masking of the mineral oil humps. Below, we will explore sample preparation tools which help to deplete these unwanted interferences but preserve the mineral oil content.
Step 1 - Removing polar contaminants
For many fatty and oil based food substances, n-hexane is commonly used to extract oils from foods and packaging. However, when the food sample has more water content, then the water miscible solvent ethanol is also used. Under these conditions, it is common to conduct a liquid liquid extraction (LLE) with n-hexane where the non-polar analytes will partition into upper hexane layer for further processing . The lower ethanolic layer, containing polar analytes such as sugars peptides and amino acids, will then be discarded.
Step 2 - Removal of natural fats and oils content
As discussed previously, a normal phase HPLC step is often used in the prefractionation step prior to GC analysis. However, if the sample is has a high fat / oil content, then this needs to be reduced prior to fractionation, otherwise the column will be overloaded
preventing effective fractionation. As triglycerides are a major class of naturally occurring fats and oils, the carboxy ester group can be targeted for saponification using potassium hydroxide. This results in the formation of polar potassium salts of the resultant fatty acids. This also includes saponification of any related esters and free fatty acids that maybe have also been present before addition of the KOH. The sample is then cleaned up twice by LLE (ethanol-water / n-hexane). As a result, any mineral oil contaminants (as well as any naturally occurring saturated, unsaturated and aromatic hydrocarbons) will partition into the in the upper n-hexane layer ready for further purification. Finally the salts in the lower layer are then discarded.
Next steps
Dependent on the make-up of the sample matrix, the initial LLE (step 1) and then saponification followed by LLE (step 2) steps, may be sufficient to then conduct successful HPLC fractionation of the hexane layer. However the sample may also contain naturally occurring aliphatic and unsaturated compounds which then coelute with the MOSH and MOAH humps respectively. As a result, further sample cleanup is needed to selectively deplete samples from these naturally occurring interferences.
Step 3 - Removing MOSH interferences
When we think of naturally occurring fats and oils, we often think of triglycerides, due to their high abundance and ubiquity. However there are a range of other naturally occurring lipid molecules, such as for example, sphingolipids and phospholipids. Furthermore, there are also a number of biogenic alkanes found in foods and oils too. The issue is that these natural alkanes will coelute with the MOSH humps and so will interfere with the results if not selectively removed. Thankfully natural alkanes have two properties which help to identify that they are biogenic rather than from a mineral source. The first is that they are linear and the second is that they exist with odd numbered carbon chains. In contrast, mineral oil alkanes are almost exclusively composed of branched and ring components. It transpires that columns packed with activated alumina (AlOx) will selectively bind strongly to n-alkanes, compared to alkanes with branched and ring components. As a result, AlOx can be used to remove these biogenic alkane interferences.
Step 4 - Removing MOAH interferences
Like the MOSH fraction, there are also naturally occurring unsaturated analytes which can elute with the MOAH hump on the GC chromatogram. These compounds include, squalene, sterols, and carotenoids. If these are observed, then an epoxidation step of double bonds with meta-chloroperbenzoic acid (mCPBA) in ethanol or performic acid (followed by quenching with aqueous sodium bisulfate solution) is conducted. The epoxidation step increases the polarity of many interferences, allowing them to be trapped on the normal phase HPLC column prior to GC-FID. When carefully controlled,
the reaction favours epoxidation of unsaturated aliphatic compounds over aromatic compounds. It must be noted though that it has been reported that some MOAH components are also lost, however the gains in sample purity are often outweighed by the presence of these analytical interfaces when in high abundance1.
Automation
It can be seen that preparing and analysing food / food oil, as well as food packaging, can be a complicated labour intensive multistep process. As a result, it is desirable to automate as much of the process as possible. Using the Trajan CHRONECT Workstation - MOSH MOAH, powered by CHRONOS software, all of the steps covered in this blog (and the previous blog) are fully automated on the same instrument. These steps are summarised below.
Automated functions on the CHRONECT Workstation - MOSH MOAH
· A one step process where a water - ethanol hexane LLE to remove polar analytes in the sample and saponification to remove natural fats and oils such as triglycerides and free fatty acids followed by LLE
· Removal of MOSH interferences using an online aluminium oxide cleanup.
· Removal of MOAH interferences using online epoxidation with performic and quenching with sodium bisulfate.
· HPLC fractionation using a 250 x 2.1 3 µL normal phase silica column.
· Backflush of HPLC column to remove build-up of polar components prior to injection of the next sample.
· Efficient transfer of HPLC (600 µL!) to GC using a HPLC-GC coupling unit onto two GC columns (one for MOSH and the other for MOAH) in one oven.
· Simultaneous analysis of MOSH and MOAH fractions
· Dedicated MOSH MOAH analytical software (Chrolibri ) which carries out the following functions
o Integration of broad chromatographic ‘humps’, or unresolved complex mixtures (UCMs).
o Any sharp peaks sitting on these humps are subtracted.
o The humps can be divided into retention time ranges, based on chain length The individual ranges can be integrated separately.
o dentification and verification functions, such as calculation of the aromatic content
o Overlay of the MOSH and MOAH chromatograms.
o Manual post-integration can be performed at any time.
o The baseline can be adjusted flexibly and easily.
o A method blank value can be subtracted.
o A result report can be run sequentially and easily.
o Data export to a text format is easily configurable (e.g. LIMS export).
Summary
Mineral oil contamination of foods presents serious health risks and so reliable identification of these substances is critical to ensure foods are safe to consume. The physicochemical properties of these substances can sometimes be similar to naturally occurring alkanes and alkenes. Moreover, food matrices (and food packaging) can be varied in terms of fat, oil and water content. As result, purification and reliable quantification of these MOSH MOAH components, presents technical challenge requiring multiple steps and employing hazardous chemicals such as performic. Thankfully Trajan has developed fully automated sample preparation and analytical MOSH MOAH CHRONECT workstations, eliminating manual handling of these processes.
If you would like to know more about CHRONECT Workstation - MOSH MOAH, then please contact us and one of our subject matter experts will be able to assist you.
References:
1 https://www.chromatographyonline.com/view/a-review-of-mosh-and-moah-analysis-in-food