Profiling of free fatty acids in wort using an isolator column-assisted LC-MS/MS approach
Journal of Food Composition and Analysis, Volume 140, 2025, 107215: Fig. 3. Comparison of measurement results of sweet wort with GC-FID and LC-MS/MS, significant differences are marked with *.
The study aims to develop a reliable and efficient LC-MS/MS method for determining medium and long chain free fatty acids (FFA) in wort, a critical component in brewing that influences beer quality. Using solid phase extraction and an isolator column, the method significantly reduces background interference and offers comparable accuracy to conventional methods, with advantages in speed, simplicity, and environmental impact.
This method was applied to wort samples from 20 German breweries to explore links between FFA and flavor instability in lager beer. Although no clear correlation was found, the method provides a valuable tool for future research into the complex roles of FFA across brewing-relevant matrices.
The original article
Profiling of free fatty acids in wort using an isolator column-assisted LC-MS/MS approach
Florian Lehnhardt, Sarina Lindtner, Martina Gastl
Journal of Food Composition and Analysis, Volume 140, 2025, 107215
https://doi.org/10.1016/j.jfca.2025.107215
licensed under CC-BY 4.0
Selected sections from the article follow. Formats and hyperlinks were adapted from the original.
The impact of fatty acids on the spontaneous foaming, known as gushing, in beer has been demonstrated. Unsaturated fatty acids can suppress or even prevent gushing, whereas lipids containing saturated fatty acids tend to exacerbate it (Carrington et al., 1972). Furthermore, the chain length of fatty acids plays a role in gushing, with only fatty acids from capric acid (C10) onwards contributing to foaming (Christian et al., 2011). Conversely, short chain fatty acids exhibit a negative correlation with gushing (Müller et al., 2010).
Medium and long chain fatty acids from C10 to C18 are considered detrimental to foam stability. However, it is emphasized that their impact on foam stability primarily occurs in combination with other molecules like proteins and strongly depends on the concentration of the respective fatty acids (Cozzolino and Degner, 2016, Gordon et al., 2018). Removing lipids from beer has been shown to improve foam stability (Dickie et al., 2001).
Fatty acids also influence the flavor profile of beer, intricately shaping its complex aroma throughout various stages of brewing. Fatty acids mainly contribute to beer flavor as a precursor for various volatile components. Certain fatty acids such as C18:2 and C18:1 have an influence on the aroma by contributing to the formation of aroma compounds such as pentanal, hexanal, 1-octen-3-ol and others (Nykänen and Nykänen, 1977, Pires et al., 2014, Vanderhaegen et al., 2006). Short and medium chain fatty acids themselves, such as C10, can have a flavor themselves and lead to a rancid and greasy impression (Clapperton, 1978). Longer chain fatty acids present in beer, such as C18:2 and C18:3, appear to have a negative effect on the flavor instability of beer. These fatty acids can lead through oxidative and enzymatic degradation to typical aging components such as (E)-2-nonenal (Vanderhaegen et al., 2006, Lehnhardt et al., 2019, Dale et al., 1977, Gottlieb et al., 1978, Kuroda et al., 2003). The highest values in literature are given for total wort content, as FFA increase after lautering of the front wort due to sparging. Depending on the amount of sparging the fatty acids are washed out of the spent grains and transferred to the total wort. Contents of around 40 mg/L are indicated in literature. Further, FFA content decreases over the course of wort boiling. At the end of the boiling process the total fatty acid content is less than 5 mg/L (Kühbeck et al., 2006, Narziss and Back, 2012, Bravi et al., 2014, Engelmann, 2021). Table 1 gives an overview of the impact of FFA on beer quality.
Despite their low concentrations in beer and wort, FFA play a key role in quality of beer and wort. Knowledge of the FFA content and composition during the brewing process enables quality control of the product in regard of the various functions of FFA as described above. It is therefore essential to have strong analytical tools for the determination of FFA.
However, none of the established methods, especially determination by GC-FID after derivatization (silylation) fulfil analytical requirements for robust process monitoring and quality control in the matrix wort. These are high sample throughput, sustainability or environmental aspects, and robust chromatographic performance. Therefore, the aim of this study was the development of a new method that (i) has simple sample preparation and short run times, (ii) does not require the use of environmentally harmful chemicals and (iii) gives reliable and robust results.
2. Material and methods
2.4. LC-MS/MS method
The development of the method was based on the publication of Koch et al. (2021). The analysis was performed on a LC system (1290 Infinity, Agilent, Waldbronn, Germany) coupled to a triple quad tandem mass spectrometer (G6470A, Agilent, Waldbronn, Germany). The injection volume was 1 µL. The analytes were separated on a Kinetex C8 Core Shell column at 30 °C (100 mm×2.1 mm, 2.6 µm particle size, pore size 10 nm, Phenomenex, Aschaffenburg). In addition, a precolumn was installed (Security Guard Ultra C8 2.1 mm, Phenomenex, Aschaffenburg). Additionally, an isolator column (C8 Core Shell 30 mm×2.1 mm, 2.6 µm particle size, 10 nm pore size, Phenomenex, Aschaffenburg) was installed after the LC pump and prior to the injection valve to offset background contaminants from the LC pump, autosampler, degasser and mobile phases. Ultrapure water was used for the eluent A. Eluent B consisted of an 80/20/0.1 ACN/MeOH/HAc mixture. At the beginning of the chromatographic run, the composition of the mobile phases was 60 % A and 40 % B, maintained for the first minute. Subsequently, a linear gradient was applied, decreasing the proportion of A to 10 % and increasing B to 90 % over 7 minutes. At 11 minutes, the composition shifted entirely to mobile phase B (100 %) and remained constant until the end of the run at 16 minutes. Finally, to equilibrate the column, the mobile phase composition reverted to 60 % A and 40 % B for the last 4 minutes, ensuring stability for subsequent analyses. The total run time for the method was 20 minutes.
Ionization was performed by negative electron spray ionization (ESI-(-)). The following source parameters were set: 45 psi spray gas, with a temperature of 150 °C and a drying gas flow rate of 5 L/min. Nitrogen was used as sheath nebulizer and drying gas. The capillary voltage was set to 3500 V (negative). The following Multiple Reaction Monitoring (MRM) mass transitions were set on the mass spectrometer. Representative fragment ions were only found for a few analytes (see Table 2). Data evaluation and device control were performed with the MassHunter software (Version 10.1).
2.6. GC method
To wort samples (10 mL), the internal standard solution (0.5 mg/L, Pentadecanoic acid) was added. A mixture of MeOH/Chloroform (25 mL, 1/2, v/v) was added and stirred for 5 min. Upon addition of a NaCl solution (44 mL, 0,58 %), the mixture was stirred again for 5 min. The resulting lower phase was separated and evaporated until dryness was reached. It was resuspended in 1 mL MTBE and again fully evaporated. Finally, 100 µL MTBE and 50 µL TMSH were added, the sample was transferred to a GC vial and immediately measured on a GC-FID (Clarus 590, Perkin Elmer, Rodgau, Germany) equipped with a Zebron ZB-FAME column (60 m, 0.25 mm, 0.2 µm, Phenomenex, Aschaffenburg, Germany). Therefore, 2 µL were injected with a split flow of 10 mL/min. The carrier gas was hydrogen and a flow of 3 mL/min was used. The following heating program was applied: 50 °C, 3 min hold; 4.25 °C/min to 135 °C; 25 °C/min to 185 °C; 13 °C/min to 250 °C. The FID was heated to 270 °C.
2.7. Aging indicators OFD-HS-SPME-GC-MS/MS
The determination of aging indicators was performed as described in the literature (Dennenlöhr et al., 2020). A gas chromatograph (Trace GC 1310, Thermo Fisher Scientific, Waltham, MA USA) coupled to a triple quad mass spectrometer (TSQ 9610, Thermo Fisher Scientific, Waltham, MA USA) and solid phase-micro extraction arrow autosampler (Triplus RSH, Thermo Fisher Scientific, Waltham, MA USA) were used for the analysis.
3. Results and discussion
3.2. Optimization of chromatographic separation using an isolator column
Background signals pose a pervasive challenge in LC-MS/MS-based fatty acid analysis (Koch et al., 2021, Kamphorst et al., 2011, Volpato et al., 2017) and were also encountered in the method developed here. The TIC of a blank measurement depicts interference signals eluting concurrently with the target analytes, rendering their differentiation unfeasible. Notably, these interfering signals are particularly prominent for analytes C16, C18 and C18:1. Consequently, the detection and quantification limits for these compounds had to be elevated due to these circumstances. Furthermore, determining LOD and LOQ based on signal-to-noise ratio proved ineffective. This is attributed to the pronounced nature of the interfering signals, which yielded signal to noise ratios exceeding 1000 even at lowest calibration concentrations (0.01, 0.05 resp. 0.1 mg/L).
Efforts were made to mitigate the intensity of interference signals by inducing fragmentation during ionization of the substances responsible for interference in blank measurements through elevating the spray gas temperature. However, doubling the temperature failed to rectify the blank value. Additionally, endeavors were undertaken to enhance the ionization of fatty acid molecules by more strongly acidifying solvent B or by opting for FA over HAc, aiming for a clearer separation between analyte and interfering signal. Nonetheless, this resulted in higher baseline signals in blank measurements and was consequently discontinued. It is noteworthy that the substances generating the interfering signals persisted despite employing purer solvents and glassware or cleansing the injection system, column, and SPE cartridges with isopropanol and methanol. As indicated in numerous other studies, the origin of these interference signals appears to stem from ubiquitous plastic products and their abrasion (Koch et al., 2021, Volpato et al., 2017, Schiesel et al., 2010).
The installation of an isolator column proved to be a possible measure to minimize the problem of interference signals. By using this method, the interfering signals could be separated from the analyte signals (Fig. 2) and thus a better LOD and LOQ could be given for most analytes (Table 3).
Journal of Food Composition and Analysis, Volume 140, 2025, 107215: Fig. 2. Comparison of a measurement of C16 in wort with (left) and without (right) a built-in isolator column. Noise signal is separated from analyte signal when using an isolator column.
3.4. Comparison with GC-FID
Both HPLC and GC methods for the analysis of FFA in wort are described in the literature (Bravi et al., 2014, Horák et al., 2009).
To compare the two methods, the sweet worts of the breweries described above were also measured using GC-FID. The results for most analytes are comparable, with significant differences observed only for C18:1 and C20 (2-sample t-test; df=42; α=0.05). In FFA analysis, numerous factors must be considered, including sample preparation, extraction efficiency, chromatographic separation, and detection methods. It is possible that the two methods differ in their sensitivity to specific fatty acids, potentially due to interactions between the analytes and the analysis conditions. The expectation that all analytes would deviate similarly arose from the assumption that any deviations would be attributable to sample preparation or detection. If derivatization were inadequate due to the complexity of the wort, it should affect all analytes equally. However, since most analysis results largely agree (see Fig. 3), it can be inferred that the complexity of the wort and its constituents does not significantly influence the choice of sample preparation or analysis method. In general, LC-MS/MS measurement tends to yield higher results.
Journal of Food Composition and Analysis, Volume 140, 2025, 107215: Fig. 3. Comparison of measurement results of sweet wort with GC-FID and LC-MS/MS, significant differences are marked with *.
A significant aspect of the comparison was the effort required for sample processing and analysis. The GC method is laborious, especially concerning sample preparation. Processing 10 samples requires 2 h for GC preparation, whereas it can be accomplished in just 1 h using SPE for LC-MS/MS. Additionally, consideration must be given to the analysis time on the instrument. A GC run takes approximately 2 h, whereas an LC run lasts only about 20 min.
In evaluating the environmental and user-friendliness of the chemicals utilized, the LC-MS/MS method must be rated superior. The use of toxic solvents such as chloroform is particularly noteworthy in this regard. Conversely, the LC-MS/MS method facilitates simpler sample preparation and employs fewer hazardous chemicals. However, it is worth noting that the LC-MS/MS method requires the use of plastic consumables, especially for SPE, which also has environmental implications.
It has to be mentioned that the cost of analysis (especially acquisition costs) are higher for the LC-MS/MS method.
4. Conclusion
The described mass spectrometry-based method is a strong analytical tool for the determination of FFA in worts and possibly other matrices. It shows clear advantages in various aspects compared to established methods. These are among others fast sample preparation and run times, robust performance, and avoidance of environmentally harmful chemicals. Thus, application of this newly developed method will provide insights into the complexity of FFA in brewing in the future.
A similar approach for the determination of FFA in wines by liquid-chromatography high-resolution mass spectrometry showed comparable analytical performance for most analytes. This method could also be adapted to beer wort (Kokotou, 2024). The here-in proposed method shows similar analytical performance though.
