Abstract
In the context of the current, flourishing specialty coffee industry, the processing of the coffee bean at the origin is increasingly becoming the main differentiation tool for specialty coffee producers around the world. The processing method of coffee at the origin may be one of three main types, depending on how the fleshy parts of the coffee fruit are removed and how the coffee bean is dried. The three main processing methods are often referred to as the washed process, pulped natural process and natural process.
The natural process consists of drying the whole coffee fruit like a prune prior to removing the pulp. It is the oldest processing method known and is currently gaining interest from the specialty coffee industry. The reasons for this are threefold – the ‘characteristic’ flavour of natural coffee is sought as a means for differentiation; natural coffee is a key ingredient of high quality espresso blends, and it is generally considered as more sustainable than other methods as it does not require water. The specialty coffee industry often describes the characteristic flavour of naturals as ‘fruity’ or ‘winey’, which is a profile sometimes called ‘Mocha’. The flavour of natural coffee is thus currently in the focus of the specialty coffee industry. However, little is known of how the flavour is formed, the compounds responsible for it and how the process might be controlled so as to allow a reproducible expression of that flavour. Thus, the overall aim of this research was to understand the flavour of Arabica natural coffee as a product class and to understand the main formation pathway during the post-harvest drying process. This aim was achieved through a series of studies, to address 5 specific objectives.
The first objective was to develop a rapid flavour-profiling method based on the analysis of coffee cupping (coffee tasting) data. The second objective was to determine the effect of postharvest processing methods (washed, natural and pulped natural) on the sensory profile of coffee. Both objectives were addressed in Chapter 3, where the effect of the main post-harvest processing methods on the sensory profile of coffee was investigated. Coffee tasting data from previous, unpublished research was analysed using a novel technique called Descriptive Cupping. The cupping data pertained to coffee cherries from 22 farms in the state of Guerrero, Mexico, which were processed using four post-harvest treatments: natural process, pulped natural process and two variations of the washed process including fermentation under water and dry fermentation. The samples were then hulled and assessed by licensed Q Graders (certified coffee tasters) using the Specialty Coffee Association of America (SCAA) cupping (coffee tasting) protocol. The qualitative descriptors provided by the cuppers were used to investigate differences in the sensory profile of the coffees produced from the four methods. Descriptor categories used by the cuppers (descriptor subgroups) were consolidated and integrated into a frequency table. Global chi-square (χ2), χ2 per-cell and non-symmetric correspondence analysis (NSCA) were performed on the frequency table.
The third objective was to characterise the sensory profiles of coffee processed using natural post-harvest methods. This objective was addressed in Chapter 4. Samples (22) of natural coffee, received from different origins, were assessed by a cupping panel trained at the University of Otago, and flavour profiles were generated using Descriptive Cupping. Chapter 4 also addressed the second objective by assessing washed coffee samples (9, coming from the same farms as 9 of the natural coffees) as points of reference.
The fourth objective was to identify flavour compounds responsible for key flavours of coffee processed using natural post-harvest methods. It was addressed in Chapter 5, where the same samples used for Chapter 4 were characterised using analytical methods. The headspace of the roasted coffee beans was analysed using Solid-Phase Extraction-Gas Chromatography coupled to Olfactometry and Mass Spectrometry (SPE-GC-MS/O) and Static Headspace- Gas Chromatography coupled to Olfactometry and Mass Spectrometry (SH-GC-O/MS). The headspace of the green coffee beans was analysed as supplementary data using Proton-Transfer-Reaction Mass Spectrometry (PTR-MS). The analytical results and the Descriptive Cupping results were analysed together using Multiple Factor Analysis (MFA).
The fifth objective was to investigate the effect of different natural post-harvest processing treatments (by favouring different fermentative populations) on flavour profiles and specific flavour compounds. This objective was addressed in Chapters 6 and 7. In Chapter 6, different natural coffee processing treatments (7) were compared, using the same raw material. The processing treatments were designed to vary the drying rate of different sections of the drying curve to favour fermentation by different kinds of microbial populations. A washed coffee sample was also produced as a witness from the same raw material. The resulting samples were characterised using Descriptive Cupping by a cupping panel from New Zealand and another cupping panel from Mexico. The results from both panels were compared using MFA. In Chapter 7, these samples were characterised by the analytical methods (GC-MS/O and PTR-MS) used for the previous studies and both analytical and Descriptive Cupping data were integrated in an MFA.
An effect of the processing method on the flavour profile was found in Chapter 3, whereby the natural samples were characterised by tropical-fruity, dried-fruity, red-fruity or fermented flavour notes. Washed samples showed floral, spicy, or nutty flavour notes, while pulped naturals tended to have a flavour profile that could range from natural to washed coffee. In the case of Chapter 4, the effect of the country of origin on coffee flavour was found to be more important than the effect of processing. This was linked to different geographically concentrated factors, including the drying technology. Only when natural coffees showed a ‘Mocha’ character (fruitiness or wineyness) did they become clearly different from washed coffees. Moreover, the absence or presence of the ‘Mocha’ character was also found to be one of the main sources of variation within the natural coffees as a group.
In Chapter 5, thermal degradation products of sugars and Maillard reaction products were found to have an important role in the differentiation of samples. However, the main factors explaining the ‘Mocha’ character were associated with both amino acid catabolism (of valine, isoleucine and leucine) and ethanol fermentation. The esters produced by the ethyl esterification of 2- and 3-methylbutanoic acid were linked to a fresh, red-fruity character (strawberry, blueberry), which is the most common description of specialty natural coffees.
In Chapter 6 similar product spaces were obtained by the New Zealand and Mexican panels and coherence in the use of descriptors such as fermented was achieved. Overall, Descriptive Cupping was shown to be a rapid method for profiling coffee flavour with satisfactory levels of discriminating power and satisfactory panel-to-panel agreement.
In Chapters 6 and 7, the most distinct natural drying treatment was found to be the one termed ‘honeying’ treatment for the purpose of this thesis, which involved suspending the drying for the first two days. This resulted in an intense fermentation with a high production of ethanol, which confirmed ethanol plays a key role in the formation of fruity esters.
This research achieved its objectives by characterising the flavour of natural coffee from the sensory and analytical points of view. Findings from this research contribute to the field of coffee science by providing the industry and future researchers with an explanation for the formation of characteristic fruity and winey flavours in natural coffee. Based on the findings of this research, coffee producers will be able to better control the flavour outcome of natural coffee. Further research is needed to characterise the role of non-volatile constituents of natural coffee, as well as to understand the role of specific fermentation types in the development of natural coffee flavour. Generally speaking, further research is needed to understand how to better control the coffee flavour outcome.