Improved Chemical Profiling of Brazilian Cachaça Using Matrix-Compatible SPME Fibers and an Ionic Liquid GC Column

Fernando C. Fontanivea, PhD; Erica A. Souza-Silvaa, PhD; Elina, B. Caramãoa, Professor; Claudia A. Zinia, Professor; Janusz Pawliszynb, Professor; Leonard M. Sidiskyc, R&D Manager; Michael D. Buchananc, Product Manager; and Daniel Vitkuskec, Market Segment Manager
aDepartment of Chemistry, Universidade Federal do Rio Grande do Sul, Porto Alegre, RS Brazil
bDepartment of Chemistry, University of Waterloo, Waterloo, Ontario, Canada
cSupelco, Bellefonte, Pennsylvania, USA

Brazilian cachaça is the earliest known distilled spirit produced in all of the Americas, first produced between 1516 and 1532. With an annual production of 1.3 billion liters (7 L per capita in Brazil), it is the second most produced alcoholic beverage in Brazil (only behind beer). It is the third most distilled spirit consumed in the world. In 2013, cachaça was recognized as an exclusive Brazilian drink by the United States Department of the Treasury Alcohol and Tobacco Tax and Trade Bureau (TTB).1

Prior to the TTB rule, “Brazilian rum” was used to classify cachaça. Unlike rum, which is distilled from molasses (a by-product from refining sugarcane), cachaça is distilled directly from the juice of unrefined sugarcane. It is more multilayered than rum, containing a complex mixture of hundreds of flavor compounds in addition to water and 38-48% ethanol (76-96 proof ). Previous published works have documented the aroma compounds present in the headspace over cachaça using headspace solid phase microextraction coupled to comprehensive gas chromatography with time-of-flight mass spectrometry detection (HS-SPME-GCxGC/TOFMS).2-5 The work presented here was devised to 1) develop methodology to identify the flavor compounds present in the actual sample, not just the aroma compounds in the headspace, and 2) refine the GCxGC column set to better utilize the two-dimensional space.

Overcoated SPME Fibers

SPME is a great sample preparation tool for the reason that it integrates sampling, extraction, concentration, and sample introduction in a single device. Benefits include minimal sample handling, minimal solvent consumption, suitable with flexible sample volumes, is easily automated to allow high throughput, applicable to on-site measurements, and may be able to improve sensitivity. The SPME technique can be used in two sampling modes; 1) in a headspace, and 2) by direct immersion in a liquid sample.

Limitations exist when using a fiber that contains divinylbenzene (DVB) and/or Carboxen® adsorbent while performing direct immersion into a liquid sample that contains macromolecules (such as sugars). These macromolecules foul the fiber (physically stick to DVB and/or Carboxen adsorbents), competing with analytes for adsorption and/or do not desorb off the fiber during the analysis/bake step. This results in:

  • Production of artifacts during GC analysis
  • Poor method reproducibility and extraction efficiency
  • Shortened fiber lifetime

An approach to overcome this limitation is the use of overcoated SPME fibers.6 These are standard SPME fibers with an adsorbent type coating (e.g., DVB and/or Carboxen) in which a controlled layer of polydimethylsiloxane (PDMS) is then added to encapsulate the adsorbent coating. Scanning electron microscope (SEM) images showing the fiber tips of both a standard SPME fiber and an overcoated SPME fiber are shown in Figure 1. With the standard SPME fiber, matrix macromolecules can interact directly with any adsorbent-type coating, and can also wick along the fiber core. With the overcoated SPME fiber, any adsorbent-type coatings are protected by the PDMS overcoating (matrix macromolecules do not stick to PDMS). Additionally, access to the fiber core is eliminated. Another location where matrix macromolecules can adhere is the interface where the fiber is attached to the inner needle. As shown in Figure 2, the PDMS overcoating denies access to this interface. A compilation of the features and benefits associated with overcoated SPME fibers is shown in Table 1.


Fiber Tip Comparison

Figure 1. Fiber Tip Comparison


Fiber Attachment Comparison

Figure 2. Fiber Attachment Comparison


Table 1. Features and Benefits of Overcoated SPME Fibers


  • PDMS overcoating serves as a barrier (matrix macromolecules do not stick to PDMS)
  • Prohibits matrix macromolecules from sticking to adsorbent-type coatings (e.g., DVD and Carboxen)
  • Analytes migrate through PDMS overcoating and onto the adsorbent surface or into the adsorbent pores
  • Overcoating application seals the end of the fiber (matrix does not wick along fiber core)
  • Overcoating application seals fiber/inner needle interface (matrix cannot access interface)


  • Less background in chromatograms (no artifacts from matrix macromolecules)
  • Reduces matrix competition with analytes
  • Extended fiber life when fibers are used in a direct immersion mode
  • More durable fibers

To illustrate the result of the overcoating process, an overcoated SPME fiber was cut in half then subjected to SEM imagery. The resulting image of the cross section is shown in Figure 3. The approximate 30 μm thick PDMS overcoat layer is clearly visible using 900x magnification.


Cross Section of Overcoated Fiber

Figure 3. Cross Section of Overcoated Fiber

SPME Method Development

To achieve the primary goal (develop methodology to identify the flavor compounds present in the actual sample, not just the aroma compounds in the headspace), SPME fibers consisting of a 50/30 μm DVB/Carboxen/PDMS coating on a 1 cm long StableFlex™ core were used. The presence of multiple sorbents maximizes the variety of compound types extracted, useful for chemical profiling. Standard SPME fibers were used for sampling in the headspace, whereas overcoated SPME fibers were used for direct immersion in liquid cachaça. The overcoated SPME fibers performed well, as the sugars and other matrix macromolecules present in the sample did not adhere to the fibers.

Figure 4 shows the real second-dimensional chromatograms obtained from headspace sampling (green trace) and direct immersion sampling (orange trace) using optimized extraction and analysis conditions. As expected, the chemical profile obtained from each sampling mode is different. Earlier eluting compounds (more volatile) are present in the headspace analysis, whereas later eluting compounds (more semi-volatile) dominate the direct immersion analysis. This highlights the drawback of solely using headspace sampling for chemical profiling, namely, incomplete characterization.

Figure 5 depicts the same data as in Figure 4 after software processing into contour plots, and then further manipulation into apex plots so that compound classifications could be displayed. A total of 330 and 423 compounds were identified in the headspace and direct immersion analyses, respectively. Multiple compounds were detected in both analyses.

Ionic Liquid GC Columns

The stationary phase is the most critical parameter of any GC column. This is because it determines the selectivity of the column, and that selectivity influences resolution. In fact, changing stationary phase may be an effective way to increase resolution.7 GC stationary phases employing dicationic ionic liquids were established in 2005.8 Extensive evaluations of these columns have shown their main strength is unique selectivity. They are able to perform many applications with slight elution order changes compared to columns made with polysiloxane polymer or polyethylene glycol phases. This often results in increased resolution and/or shorter run times. Recently, the solvation parameter model (SPM) was used to characterize ionic liquid columns to determine which analyte-stationary phase interactions are possible.9 This information could then be useful to explain why they exhibit unique selectivity. Conclusions were that only the ionic liquids are capable of simultaneously providing:

  • Intense H-acceptor interactions (a constant)
  • Intense H-donor interactions (b constant)
  • Dipolar interactions (s constant)
  • π-π interactions (e constant)
  • Limited dispersive interactions (l constant)
Chemical Profile by Sampling Mode (Real Second-Dimensional Chromatograms)

Figure 4. Chemical Profile by Sampling Mode (Real Second-Dimensional Chromatograms)

sample/matrix: 9 mL cachaça in a 20 mL vial, pre-heated at 30 °C for 15 min with agitation, no salt added
SPME fiber: (HS) 50/30 μm DVB/Carboxen/PDMS on 1 cm StableFlex (Product No. 57298-U)
SPME fiber: (DI) Overcoated 50/30 μm DVB/Carboxen/PDMS on 1 cm StableFlex (custom)
extraction: (HS) 30 °C for 60 min with agitation
extraction: (DI) 30 °C for 60 min with agitation, rapid rinse of SPME fiber in deionized water before desorption
desorption process: 250 °C for 25 min
column: (1D) SLB-IL59, 30 m × 0.25 mm I.D., 0.20 μm (Product No. 28891-U)
column: (2D) 5% phenyl polysiloxane, 2 m × 0.25 mm I.D., 0.25 μm
oven: (1D) 50 °C (3 min), 3 °C/min to 295 °C (5 min)
oven: (2D) 85 °C (3 min), 3 °C/min to 330 °C (5 min)
oven: (modulator) temp. offset 35 °C, modulation period 10 sec
inj. temp.: 250 °C
detector: TOFMS, 250 °C, m/z = 45-400
MSD interface: 200 °C
carrier gas: helium, 1.0 mL/min
injection: 30:1 split
liner: 0.75 mm I.D., direct (SPME) type, straight design

(HS = headspace, DI = direct immersion, 1D = 1st dimension column, 2D = 2nd dimension column)


Chemical Profile by Sampling Mode (GCxGC Apex Plots)

Figure 5. Chemical Profile by Sampling Mode (GCxGC Apex Plots)

Conditions: same as listed in Figure 4.

GCxGC Column Sets

The secondary goal (refine the GCxGC column set to better utilize the two-dimensional space) was also achieved. GCxGC is one of the fastest growing areas in analytical chemistry due to its ability to resolve a large number of compounds, even in the most complex samples. No other chromatographic technique can match the level of detail it provides. Common detectors, including MS, can be used. It employs two columns in series, separated by a modulator. The role of the modulator is to collect fractions from the first column (often called the primary column, first dimension column, or 1D column) and focus them onto the second column (often called the secondary column, second dimension column, or 2D column). First dimension columns tend to be 30 m × 0.25 mm I.D., whereas second dimension columns tend to be short, cut-down lengths of stock columns (0.10 or 0.18 mm I.D. for non-MS detectors, and 0.18 or 0.25 mm I.D. for MS detectors).

One key to the successful operation of GCxGC is the two columns must have orthogonal selectivity, that is, they must utilize different retention mechanisms. The more different (more orthogonal), the better the overall performance will be. A common GCxGC column set uses a column containing a poly (50% diphenyl/50% dimethylsiloxane) stationary phase along with column containing a poly(5% diphenyl/95% dimethylsiloxane) stationary phase. Figure 6 shows the GCxGC apex plots obtained using such a column set. It was noticed that the GCxGC space was poorly utilized, as evidenced by the absence of data points in many areas. As illustrated in the GCxGC apex plots shown in Figure 7, an improved utilization of the GCxGC space was achieved when the 50% phenyl column was exchanged for an SLB-IL59 column, a column that contains an ionic liquid stationary phase.

Both the 50% phenyl column and 5% phenyl column can undergo similar analyte-stationary phase interactions. The difference is solely in the ratios of these interactions. Conversely, the ionic liquid column can undergo different interactions, so is more orthogonal to the 5% phenyl column than the 50% phenyl column is. This results in an improved utilization of the GCxGC space, allowing for a more complete identification of individual compounds.

GCxGC Space Utilization using Column Set 1

Figure 6. GCxGC Space Utilization using Column Set 1 (50% Phenyl × 5% Phenyl)

column: (1D) 50% phenyl polysiloxane, 30 m × 0.25 mm I.D., 0.25 μm
oven: (modulator) temp. offset 35 °C, modulation period 8 sec
All other conditions the same as listed in Figure 4.


GCxGC Space Utilization using Column Set 2 (SLB-IL59 × 5% Phenyl)

Figure 7. GCxGC Space Utilization using Column Set 2 (SLB-IL59 × 5% Phenyl)

the same as listed in Figure 4.


A more complete chemical profile of cachaça was obtained, compared to data previously reported. Headspace sampling and direct immersion sampling proved to be complementary. The practical aspects of overcoated SPME fibers demonstrated here create new opportunities for the SPME technique when applied to complex matrices, such as cachaça and other spirits and foods. Possibilities may also exist for on-site, in vitro, and in vivo sampling. The proper selection of a GCxGC column set allowed better utilization of the GCxGC space, which also contributed to improving the chemical profiling.


Carboxen is a registered trademark of Sigma-Aldrich Co. LLC




  1. TTB Amends the Distilled Spirits Identity Regulations to Recognize “Cachaça” as a Type of Rum and Distinctive Product of Brazil, announcements/press-releasefy1305-cache-td.pdf (accessed October 9, 2015).
  2. de Souza, M. D.; Vásquez, P.; Del Mastro, N. L.; Acree, T. E.; Lavin, E. H. Characterization of cachaça and rum aroma, J Agric Food Chem 2006, 54, 485-488.
  3. . Cardeal, Z. L.; de Souza, P. P.; Gomes da Silva, M. D. R.; Marriott, P. J. Comprehensive two-dimensional gas chromatography for fingerprint pattern recognition in cachaça production, Talanta, 2008, 74, 793-799.
  4. Vidal, E. E.; de Billerbeck, G. M.; Simões, D. A.; Schuler, A.; François, J. M.; de Morais Jr, M. A. Influence of nitrogen supply on the production of higher alcohols/esters and expression of flavour-related genes in cachaça fermentation, Food Chem, 2013, 138, 701-708.
  5. Menezes, H. C.; Paulo, B. P.; Paiva, M. J. N.; de Barcelos, S. M. R.;Macedo, D. F. D.; Cardeal, Z. L. Determination of polycyclic aromatic hydrocarbons in artisanal cachaça by DI-CF-SPME–GC/MS, Microchemical Journal, 2015, 118, 272-277.
  6. Souza-Silva, E. A; Pawliszyn, J. Optimization of fiber coating structure enables direct immersion solid phase microextraction and high-throughput determination of complex samples, Anal Chem, 2012, 84, 6933-6938.
  7. Barry, E. F. Columns: Packed and Capillary; Column Selection in Gas Chromatography. Modern Practice of Gas Chromatography, Fourth Edition. Grob, R. L.; Barry, E. F. Eds.; John Wiley & Sons, Inc.: New Jersey, 2004; 65-191.
  8. Anderson, J. L.; Ding, R.; Ellern, A.; Armstrong, D. W. Structure and properties of high stability geminal dicationic ionic liquids, J Am Chem Soc. 2005, 127, 593-604.
  9. Rodríguez-Sánchez, S.; Galindo-Iranzo, P.; Soria, A. C.; Sanz, M. L.; Quintanilla-López, J. E.; Lebrón-Aguilar, R. Characterization by the solvation parameter model of the retention properties of commercial ionic liquid columns for gas chromatography, J Chromatogr A, 2014, 1326, 96-102.


Related Links