An Analytical Technique for Identifying Tribromoanisole

In the past several years, pharmaceutical companies have recalled some of their products due to the presence of a musty, moldy, earthy odor (1–5). When a consumer opens a product container and smells an unusual odor, it indicates that the product has been compromised. Subsequent analysis of the recalled products determined that 2,4,6-tribromanisole (TBA) caused the odor.

TBA has been found to cause odor problems in the past. Whitfield and coworkers identified TBA as a possible source for a musty odor in sultanas that were packaged in polyethylene bags (6). They demonstrated that when these food packages were stored in the presence of TBA-contaminated fiberboard, the food could acquire the off odor or off flavor after only one week.

Chatonnet and colleagues also identified TBA as the compound responsible for a musty off odor in wine (7). The compound was present in wines that had a musty character during tasting. The wine acquired this flavor from being stored in an atmosphere contaminated with TBA.

TBA belongs to a group of organic compounds called haloanisoles, which are extremely powerful and can add musty, moldy, or earthy odors to many materials, such as food, beverages, and packaging. Haloanisole compounds contain at least one halogen atom (i.e., fluorine, chlorine, bromine, or iodine) as part of their chemical composition.

Other haloanisole compounds, such as 2,4,6-trichloroanisole (TCA), 2,3,4,6-tetrachloroanisole, and pentachloroanisole have similar odor characteristics and have been associated with cork taint in wine (8). They are extremely aromatic and have low odor-threshold concentrations i.e., (the minimum amount necessary for the compound to be smelled by the majority of people).

The minimum concentration necessary to produce the characteristic off odor of TBA and TCA is approximately 30 parts per quadrillion (9). These haloanisole compounds are a byproduct of microbial metabolism in products that have been treated with halophenols. The biomethylation process is performed by various types of filamentous fungi primarily as a biochemical self-defense mechanism. Figure 1 shows the chemical reaction during which the hydrogen of the hydroxyl group is replaced with a methyl group. The fungi convert the usually toxic halophenol compounds into less toxic haloanisole compounds.

Figure 1: Biomethylation reaction where 2,4,6-tribromophenol is converted to 2,4,6-tribromoanisole. (ALL FIGURES ARE COURTESY OF THE AUTHORS)

Haloanisoles can be present in many manufacturing and shipping areas. Trichlorophenol can be formed when chlorine reacts with the natural phenolic compounds found in wood and water. Cleaning or sanitizing materials are a possible source of chlorine. Tribromophenol (TBP) can be used as a flame retardant in brominated plastics, papers, and textiles. The compound also is used as a fungicide, a wood preservative, and an antiseptic agent.

TBP is also a synthetic intermediate used for producing many types of brominated flame retardants (10). In some parts of the world, TBP continues to be used as a fungicide and wood preservative. The compound is not a registered pesticide with the Environmental Protection Agency, and its use as a fungicide has been banned in the United States, Canada, and Europe (11).

For contamination to occur, TBP and the filamentatous fungi must be present. As the fungi consume the TBP, the local concentration of TBA increases as a function of time. Only after the odor-threshold concentration for TBA is exceeded does the characteristic odor of TBA become apparent.

TBA contamination occurs upon direct contact with the source, including air, in an enclosed area. Because TBA is readily absorbed by solid materials, the contamination can be instantaneous and can quickly spread throughout a facility. Products can be contaminated simply by environmental exposure.

When a moldy, musty, off odor emanates from a product, it is crucial that the problem be identified as quickly as possible to minimize the financial consequences and damage to the brand's reputation. The source of the contamination must be determined, and the area isolated. An analysis of samples with the odor may identify the specific compounds that are responsible, which can suggest possible mechanisms for the contamination.

One caveat about this type of contamination is that the typical smell of various packaging materials, shipping containers, and warehouses may be musty or moldy. Thus, a musty or moldy off odor may not be a contamination problem.

The largest problem associated with the analysis of TBA is the extremely low odor-threshold concentration. A concentration of 30 parts per quadrillion is well below the detection limits of any analytical instrument. Consequently, an individual will be able to smell the odor of the compound long before an analytical instrument can identify and measure it.

As mentioned earlier, the wine industry encountered a similar odor problem that resulting from TCA contamination. Only after the instrumental sensitivity increased and the detection limits decreased were scientists able to identify the compounds responsible for cork taint. Much of the analytical methodology developed for the analysis of TCA can be applied directly to the analysis of TBA.

A method can be developed for routine testing of TBA in samples using gas chromatography–mass spectrometry (GC–MS) instrumentation. However, TBA may not be the only compound in the sample responsible for the offending odor. Other compounds present could also contribute to it.

Since many foul-smelling compounds have extremely low odor thresholds, GC with olfactory detection provides an excellent methodology to identify aroma notes from a sample. These concentrations are generally below the detection limits for a mass spectrometer. By identifying the retention times for crucial odor notes, the scientist can focus on these areas and use the mass spectrometer to confirm the compound identification.

Environmental samples can be collected using two methodologies. In the first method, air is drawn continuously through a thermal desorption cartridge packed with an appropriate adsorbant material for specific periods of time. In the second method, solid-phase microextraction (SPME) fibers can be located throughout the facility. These two types of samples can be directly inserted and desorbed in the GC inlet for separation and analysis.

Many other types of samples, such as corrugated containers, plastic containers, and pharmaceutical products, can be analyzed. The sample is enclosed in a sealed container, in the headspace of which the volatile and semivolatile compounds equilibrate, and SPME fibers are used to sample the headspace of the container. Depending on the strength of the off odor, the volatility, and the headspace concentration of the key odorants, the length of the collection time can vary.

One advantage of SPME fiber technology is its ease of operation. Following headspace collection, the fiber is inserted directly into the hot inlet of the GC system, and the adsorbed compounds are flash desorbed into the column for GC analysis.

A case study

Figure 2 presents the total ion chromatogram obtained from the MS detector operating in scan mode during a multidimensional GC analysis. Not only does this result demonstrate the complexity of the GC analysis, but it also shows a crucial aspect of odor testing. The numerous peaks in the chromatogram represent the volatile and semivolatile compounds collected from the headspace of the sample. Both odorless and odorous compounds are represented by these peaks, but identifying the compounds associated with each of the peaks is an enormous task. From the list of compounds generated, determining which compounds are responsible for the off odor from this sample is an even larger task. Scientists ultimately determined that this sample was contaminated with TBA. Under these experimental conditions, TBA elutes at approximately 30 min.

Figure 2: Total ion chromatogram from a TBA contaminated sample.

By combining olfactometry with the GC–MS analysis, the retention times at which odor-causing compounds elute can be identified quickly. Figure 3 is an aromagram, which is a graphical presentation of the sample's aroma and intensity versus retention time in the GC–MS chromatogram format. Each aroma detected at the sniff port by the scientist during the olfactory analysis is assigned descriptors and a relative intensity. The aromagram shows numerous aroma notes that indicate the presence of compounds whose concentrations have exceeded their odor-threshold concentrations. Although numerous aroma compounds were sensed during the analysis, a single, large, character-defining odorant peak occurred at a retention time of approximately 30 min. It should be noted that the mass spectrometer operating in scan mode detected nothing but baseline noise at this retention-time region.

Figure 3: Aromagram from a TBA contaminated sample.

In Figure 4, the gas chromatogram and the aromagram have been combined, and the retention-time axis has been expanded from 24 to 32 min. Both the mass spectrometer trace from the gas chromatogram and the aromagram show several compounds in this region. However, the retention times for these compounds do not coincide with each other. Thus, the compounds indicated in the GC trace are odorless or below their odor-threshold concentration. The aromas indicated in the aromagram trace have exceeded their odor-threshold concentration, but are below the detection limit of MS. The same situation obtains for TBA, where the aromagram shows an intense peak that cannot be identified on the MS scan.

Figure 4: An expanded combination plot of the gas chromatogram (red curve) with the aromoagram (black curve).

This retention-time area must be investigated closely to confirm the identification of the compound with the mass spectrometer. Two methods may provide compound verification. The first method involves sample preconcentration of the headspace before analysis. The second method is to operate the mass spectrometer in selective ion monitoring (SIM) mode for ions representative of a short list of compounds. Nonetheless, in this example, the olfactory analysis that is carried out simultaneously during the GC–MS analysis was crucial in identifying the retention time of the compound responsible for the major odor effect.

Once the aroma or odor compounds are identified, method development can occur based on the results. Early researchers in the wine industry focused on the analysis of TCA using SPME followed by GC–MS analysis (12). The MS analysis was performed in the selective-ion mode. Quantification is based on an internal standard with a quantitation limit of 5 ng/L.

The odorous mixed trihalogenated anisoles have also been evaluated in water samples using a headspace–SPME and GC–MS method (13). By optimizing the parameters affecting collection efficiency, the detection limit for TCA was lowered to 0.25 ng/L.

This same technology has been applied to the analysis of food packaging (14). TCA, TBA, and pentachloroanisole were simultaneously analyzed in packaging samples by SPME and GC. The detection limits ranged from 0.43 to 1.32 ng/g.

Conclusion

The determination of haloanisoles, specifically TBA, presents a difficult analytical challenge. Due to their extremely low odor thresholds, these compounds generate odors that are smelled long before their concentrations are large enough to be identified instrumentally. Combining olfactometry analysis with multidimensional GC–MS creates an extremely useful analytical methodology for identifying aroma or odor notes from a sample. The chemical compounds responsible for the aroma notes can be verified using the mass spectrometer detector. In many cases, identification of the compound causing the off odor is the ultimate goal, so that the manufacturing process can be scrutinized to determine the compound's origin. However, the origin of a compound within a manufacturing process can be determined knowing the retention time and the aroma character of a compound alone. Finally, analytical protocols can be developed that monitor these low-concentration compounds as part of a quality-control program.

Roger J. Bleiler* is a senior scientist, Fred Kuhrt is an operations manager, and Don Wright is a senior scientist and consultant, all at MOCON, 2011 A Lamar Blvd., Round Rock, TX 78664, tel. 512.218.9873, fax 512.218.9875, roger.bleiler@mdgc.com.

*To whom all correspondence should be addressed.

References

1. P. Kavilanz, "Tylenol Recall: FDA Slams Company," www.money.cnn.com/2010/01/15/news/companies/over_the_counter_medicine_recall/index.htm, accessed Oct. 3, 2011.

2. D.J. DeNoon, "Lipitor Recall Grows by 19,000 Bottles," www.webmd.com/cholesterol-management/news/20101221/lipitor-recall-grows-by-19000-bottles, accessed Oct. 3, 2011.

3. Reuters, "UPDATE 2-Depomed Recalls 52 Lots of Diabetes Drug," www.reuters.com/article/2010/06/17/depomed-idUSSGE65G0D020100617, accessed Oct. 3, 2011.

4. P. Van Arnum, ePT, May 19, 2011, www.pharmtech.findpharma.com/pharmtech/News/JampJ-to-Recall-Select-Lots-of-Prezista-Due-to-Pos/ArticleStandard/Article/detail/723176, accessed Oct. 3, 2011.

5. FiercePharma, "Depomed Recalls Diabetes Med on TBA Contamination," June 17, 2010, www.fiercepharma.com/story/depomed-recalls-diabetes-med-tba-contamination/2010-06-17, accessed Oct. 3, 2011.

6. F.B. Whitfield, J.L. Hill, and K.J. Shaw, J. Agric. Food Chem. 45 (3), 889–893 (1997).

7. P. Chatonnet et al., J. Agric. Food Chem. 52 (5), 1255–1262 (2004).

8. C. Pizarro, N. Perez-del-Notario, and J.M. Gonzales-Saiz, J. Chromatogr. A 1143 (1–2), 26–35 (2007).

9. P. Tomboulian et al., Water Sci. Technol. 49 (9), 219–226 (2004).

10. Weifang Huitao Chemical, "2,4,6-Tribromophenol (TBP)," (Weifang City, China), www.huitaotech.com/huitaotechcom/cpshow.asp?id=147&lxid=0005&gclid=CI3g8IWO8KsCFaJgTAodOxlSKA, accessed Oct. 17, 2011.

11. M. Foley and C. Brindley, Pallet Enterprise, Aug. 1, 2011, www.palletenterprise.com/articledatabase/view.asp?articleID=3457, accessed Oct. 17, 2011.

12. T.J. Evans, C.E.Butzke, and S.E. Ebeler, J. Chromatogr. A 786 (2), 293–298 (1977).

13. A. Diaz, F. Ventura, and M.T. Galceran, J. Chromatogr. A 1064 (1), 97–106 (2005).

14. E.Martendal et al., Mikrochim. Acta 159 (3–4), 229–234 (2007).