By Emma Sage, Coffee Science Manager, SCAA
Coffee is what is known in the food science field as a shelf-stable product, which after roasting does not spoil due to enzymatic or microbial processes (Illy and Viani 2005; Nicoli and others 1993; Anese and others 2006). However, in the specialty coffee industry, we are aware of the importance of chemical reactions and physical changes that occur after roasting (Nicoli and others 2009). Some of these changes are responsible for staling, or a perceptible negative flavor that increases over time, and affects the quality of the brew.
Capturing the exact nature, quantity and rate of staling is inherently challenging due to both to the diversity of flavors possible in the bean itself, and of the ephemeral nature of roasted coffee. The Specialty Coffee Association of America (SCAA) water quality handbook states “coffee’s flavor potential is constantly changing. As a result, when conducting a chemical or sensory analysis, coffee must be considered a moving target” (Beeman and others 2011). This concept epitomizes the problem with defining the science behind staling. The chemical and physical changes that occur in coffee after roasting make experimental control, repeatability, and data analysis all but impossible. However, this has not stopped a large cohort of researchers from tackling the challenge.
The General Causes of Coffee Staling
Roasting is ultimately responsible for much of coffee staling reactions, as it forms volatiles and creates pressure in coffee beans via internal gas build-up (Nicoli et al. 2009). Many complex chemical reactions and physical changes occur during roasting, a few of which play key roles in staling. As the beans heat up in the roaster, sugars and amino acids are reduced and increasing amounts of carbon dioxide are formed via Strecker degradation, which later off-gas. At the same time, a physical change occurs in the beans as the mass of the bean decreases and porosity increases, leading to a higher potential volatile compound diffusion rate (Labuza and others 2001). Simultaneously, the Maillard reaction produces compounds that have an affinity for oxygen and later contribute to lipid oxidation (Nicoli et al. 1993).
The loss of carbon dioxide from coffee occurs due to diffusion forces, which move molecules because of differences in pressure and/or gradients of molecule concentrations. When coffee is ground, the porosity and surface-to-volume ratio increase, which accelerates degassing and staling. A group of studies have found that losses of a few specific volatile compounds are responsible for a majority of coffee aroma loss. The first research to propose this found that methanethiol and 2-methylpropanal gave the most intense aroma notes and dissipated two hours after roasting, and that after eight days of storage, methanethiol decreased to about 30% of its original amount (Holscher and Steinhart 1992). Czerny and Scieberle (2001) and Sanz and others (2001) have also reported these compounds as key molecules lost in staling. New work has also focused on the ratios of certain compounds, such as 2-methylfuran/2-butanone, 2-furfurylthiol/hexanal (Marin and others 2008). However, any of the above compounds or ratios are only indicators of a broader group of reactions responsible for staling that have not yet been characterized (Nicoli et al. 2009).
One of the characteristic flavors of staling is rancidity, which is created by lipid degradation, the chemical oxidation or pyrolysis of fats and related compounds (Smith and others 2004; Vila and others 2005). In roasted arabica, lipids only account for about 15% of the dry weight (Illy and Viani 2005), but they significantly impact the flavor of staling. This process is accelerated by moisture (Smith et al. 2004), oxygen (Vila et al. 2005) temperature (Nicoli et al. 1993; Huynh-Ba and others 2001), and correlates with the surface area of ground coffee (Vila et al. 2005). However, it is also known that even coffees stored under vacuum or low oxygen can show lipid oxidation due to the presence of free radicals in coffee produced during the roasting process.
The rate of all of these changes and therefore the overall shelf life of coffee is dependent on the state of the coffee (whole bean vs. grind) and environmental conditions, such as temperature, moisture, and most importantly, oxygen availability (Nicoli et al. 1993; Illy and Viani 2005; Radtke-Granzer and Piringer 1981).
The Main External Factors Affecting Staling
Temperature affects the rate of coffee staling both chemically and physically. Chemically, temperature is positively correlated with the kinetics of chemical reactions (i.e. the Arrhenius equation), therefore, with warmer temperatures all chemical reactions in coffee are accelerated (Nicoli et al. 2009). Physically, pressure and concentration gradients between coffee and the external environment are influenced by temperature, which affects the rate of degassing of volatile compounds. Labuza et al. (2001) reported that degassing whole bean roasted coffee had a Q10 value of 1.5, which means that for every 10 degrees of temperature increase, the rate of degassing increases by 1.5-fold. They also showed that ground coffee had an accelerated Q10 value about twice that of whole beans. Nicoli and others (1993) found that temperature had a positive correlation with the release of carbon dioxide and other volatile compounds, and that these losses were most dramatic within the first few days of coffee storage. There is a solid body of literature on coffee that confirms that the chemical compounds linked to freshness decrease at increasing rates of temperature (Cappuccio and others 2001; Cardelli and Labuza 2001; Huynh-Ba et al. 2001). Generally, this relationship is characteristic of the curve in figure 1 from Cappuccio and others (2001):
On the other end of temperature research, of course, is the effect of cold temperatures and freezing on coffee. For a long time, there was generally a consensus that freezing is an adequate deterrent of staling (Sivetz 1979). A majority of studies that included refrigerating or freezing coffee as a method of preservation found that it slowed the reactions known to be a part of staling (Cappuccio et al. 2001; Ross and others 2006; Nicoli et al. 1993). However, none of these included any measure of differences that might have resulted from freezing or the associated temperature fluctuation.
Moisture (and water activity) also has a general positive correlation with coffee staling. Commonly, studies have found that if coffee is stored in a high moisture environment, it will pick up water and increase it water activity values, accelerating its loss of volatile compounds and therefore shortening its possible shelf life (Anese et al. 2006; Cardelli and Labuza 2001; Prescott and others 1937). Relatedly, the wide practice of water-quenching has been linked to a higher water activity in roasted coffee, which therefore would also lead to faster degassing (Baggenstoss and others 2007). The absorption of water after roasting can also occur over time, increasing with the humidity of the storage environment (Illy and Viani 2005; Apostolopoulos and Gilbert 1988).
Oxygen availability is deemed by many the primary enemy of roasted coffee and affects staling reactions in a variety of ways. Of course, factors that influence the interaction between oxygen and coffee bean or grounds, such as packing density, ground size, or bean surface area, also influence these reactions (Ross et al. 2006; Illy and Viani 2005). Oxidation can not only responsible for loss of some aroma compounds, but also the formation of off-flavors, such as rancidity (Prescott et al. 1937; Illy and Viani 2005).
It has been found that most of the compounds responsible for the aroma of freshly roasted beans are very susceptible to oxidation and can be lost quickly after roasting. Some work has determined that degradation of freshness occurs as soon as coffee has contact with oxygen. Poisson and others (2006) found that hexanal, formed by oxidation reactions, was immediately generated in roasted coffee in an unprotected setting. They also found that low-weight sulfurous compounds dissipated rapidly with exposure to oxygen. Labuza et al. (2001) determined that oxygen was the most important factor controlling the shelf life of coffee, and showed that reducing oxygen to 0.5% in a coffee container could increase shelf life by 20-fold. One research group found that for each 1% oxygen increase there is an increase of the rate of degradation of 10% (Cardelli and Labuza 2001). Even at very low levels of oxygen in packaged coffee (<2%), this oxygen has been found to migrate into coffee and facilitate oxidation reactions (Harris and others 1974).
Lipid oxidation occurs as degassing and volatile compound loss goes on and is also affected by oxygen availability (Nicoli et al. 1993; Prescott et al. 1937). Huynh-Ba et al. (2001) found that the oxidation of lipids to volatiles occurred during the first 24 hours after roasting and grinding. Another study found that trained sensory assessors were able to detect rancidity in coffee packed in air after four months (Marin et al. 2008). These volatiles undergo subsequent reactions to influence volatile secondary oxidation products, which contribute to the “rancid” flavor that can be found in stale coffee.
The Effect on Taste
Not all studies that have investigated the staling of coffee have included how these chemical reactions influence taste. However, when taste is included it is apparent how immediate staling is. Just one week after roasting, tasters in one study preferred soluble coffee that had been stored in a can with 0% oxygen over coffee stored under 2% oxygen (Harris et al. 1974). Ross and others (2006) found that sensory panelists preferred fresh coffee compared with two-week stored coffee, finding the coffee bitter, but also preferred two-week stored coffee over one-week stored coffee. A different research group, Cardelli and Labuza (2001), found that sensory testers detected a loss of quality in coffee with increases in oxygen partial pressure, water activity, and temperature, confirming that these environmental influences made their way to the cup. They deemed that oxygen had the most critical role, with a close to twenty-fold staling acceleration difference between 0% oxygen and average sea level oxygen concentrations.
Many studies on staling include a sensory evaluation of coffee aroma, as opposed to taste. Aroma testers in Steinhart and Holscher’s (1991) study noted that coffee one-week off-roast were “distinctly less odor-intensive” and showed “less aroma freshness.” The researchers determined that this was due to quickly dissipating “low boiling” components such as sulfur compounds, Strecker-aldehydes, and alpha-dicarbonyls. Sanz and others (2001) found that eight identified volatile compounds were positively correlated with the sensory rating of aroma freshness and that the greatest rate of freshness loss occurred in the first month of coffee storage (see Figure 2, below).
The Effect of Packaging on Staling
Packaging and how it buffers external influences on coffee can influence staling a great deal. However, there are very few published studies investigating or comparing specific packaging types. A small experiment on flexible packages with degassing valves found that unless there was a leak in the seal at the top of the bag, they contained 0% oxygen and above 40% carbon dioxide. However, they only used six bags for this experiment and found that half of them presented with leaks, ultimately voiding any conclusions that could have been made (Walter and others 2008). More frequently, studies investigated gas flushing. Most found that coffee flushed with inert gases or vacuum packaged fared better in taste tests (Bezman and others 2008). In a study by Alves and others (2001), nitrogen-flushed coffee had a six-month shelf life based on sensory analysis, as opposed to coffee bagged with no flushing, which had a three-month shelf life. Historically, there was an understanding that coffee would be adequately protected by vacuum packing (Sivetz 1979). Vacuum packing has been shown to be very effective, one study demonstrating 0% residual oxygen in sealed coffee bags after two rounds of vacuum packing and gas flushing (Sortwell 2008). Nicoli and others (1993) found that coffee vacuum packaged had a staling rate five times lower than beans packed in air. Finally, some research admitted that despite coffees packaging, a “secondary shelf life” begins once that packaging has been opened by the consumer, which may occur at an accelerated rate compared to normal, in-package, staling (Cappuccio et al. 2001; Anese et al. 2006). Setting up an experiment to investigate this type of staling would prove difficult, as coffee no doubt goes through a variety of environments after it is bought and opened by consumers. Also, there has been no attempt to account for the chemical reactions that may occur during this time.
Issues with Existing Research
After studying the available resources detailing studies carried out to investigate the mechanisms behind staling coffee, a few major issues come to light in our ability to apply the results to the specialty coffee industry. First, most studies that included sensory evaluation did so with aroma proxies for staling, rather than the taste evaluation that is commonly performed in the specialty industry. Second, the quality of coffee used in these studies was often below specialty grade. Third, the methods used to determine these proxies for staling include practices that would never or rarely happen at a coffee shop or home brewing situation, limiting the applicability to the specialty coffee industry. Finally, a large portion of the research covered in this review was never submitted for a peer-reviewed process, meaning that the quality of research varied and experimental design, implementation, and results cannot be fully verified.
Fundamentally, the reason that this body of literature exists about coffee is that researchers hope that they will come up with a proxy for coffee taste freshness that large companies can use to easily determine shelf life of their coffee (Kallio and others 1990; Marin et al. 2008; Ross et al. 2006; Czerny and Schieberle 2001). In the specialty coffee industry, we would no doubt ideally measure staling in brewed coffee with the quantitative cupping method (Lingle 2011). In fact, most of the chemical compounds present in roasted coffee that end up suspended in the coffee beverage are virtually unknown (Nicoli et al. 2009), whereas the chemistry of coffee odor has been more extensively documented (Semmelroch and Grosch 1995; Czerny and others 1999; Mayer and others 2000). Due to methodological constraints and the opinion that aroma proxies are sufficient, compounds that contribute to the coffee aroma are measured for proxies of taste. From these identified volatile compounds, researchers have determined some ratios that they feel epitomize the reactions that occur after coffee is roasted and begins to stale. These ratios are correlated to simultaneous sensory evaluations in order to support the conclusion that these metrics are useful. However, the identified compounds in these studies may be but a few of many possible volatile compounds present in coffee.
One of the most important causes for concern in the coffee staling literature is that experiments are conducted on coffees of low or varied quality, few focusing on coffee that would be considered specialty, nor was any coffee specifically evaluated as specialty. Since so little is known about staling chemistry, we can certainly gain knowledge from these studies, but the quality of coffee should still be taken into account. Half of the studies reviewed here were conducted on a robusta or robusta/arabica coffee varietal blend, as detailed in table 1 (below). Additionally, seven studies did not include coffee variety information at all in their reports, leaving out critical data for those interested in repeating their experiments or building on their results.
Another important methodological issue in the research investigating coffee staling is the technique used to identify volatile compounds important in staling. Researchers have tried to find a repeatable way that “freshness” can be measured, and for the scientific community this has become analysis via gas chromatography of headspace volatiles (Kallio et al. 1990; Holscher and Steinhart 1992), where roasted and ground coffee is heated and/or agitated (at up to 90°C and/or for up to 19 hours) so that it releases all of the possible volatile compounds, which are captured and then identified by reference compounds in relative abundances (see table 3, below). This technique, albeit scientifically valid, may not be answering questions that the specialty coffee industry would be interested in, since it effectively oxygenates, re-roasts, or burns coffee to allow volatiles to escape.
Also in these studies on staling, methods of quantifying staling and shelf life were not consistent. Ultimately, this comes down to the definition of shelf life and who should be able to determine it in coffee. Food scientists have investigated shelf life extensively in other food products. Smith and others (2004) in their paper describing the staling of baked goods, stated that staling was “almost any change…making it less acceptable to the consumer.” However, this definition of shelf life varies. Statistically, the food science industry tends to define it as over 50% rejection. Or over 50% of tasters feel the product is not “acceptable” (Fu and Labuza 1993). This 50% originates in the experimental design developed by food scientists to test the shelf life of a product called the Weibull Hazard Analysis (Gacula 1975). This is a test that increases the number of sensory testers as the quality of a product decreases. This test finally ends when a consistent 50% of participants deem the product unacceptable. In the research on coffee, Labuza et al. (2001) defined the end of shelf life when more than 50% of consumers became “displeased” with the product, but only used “acceptable” and “unacceptable” in their consumer rating scale. Ultimately, it must be decided how to identify the “acceptability limit” of staling coffee. This is stressed in a recent publication by Nicoli and others (2009). They discuss the difficult task of doing so in coffee, as it has poses no safety hazard or biological toxicity with age after roasting, as other food products do. They suggest that sensory evaluation is the only method that is acceptable, and that this should be based on customer acceptance limits, as “foods do not have sensory shelf lives without the consumer” (Nicoli et al. 2009).
Finally, almost half of the original studies researching coffee staling included in this review (12 of the 28) were not available in a peer-reviewed article. Despite the fact that these findings were presented at a professional conference, primarily the Association for Science and Information on Coffee (ASIC) meetings, it would much benefit the community to have a higher value placed on peer-reviewed studies. The peer-review process has merit on two levels, both of which should be important to the coffee industry. The first is the assurance of proper experimental techniques, sample size, and overall appropriateness of topic-specific methods. These checks and balances assure full disclosure, data availability, and experimental repeatability, which are an important stamp of approval and give readers the assurance of accuracy and quality. The second merit of peer review is that it allows those interested in continuing research to build upon previous work. A solid base of reliable, peer-reviewed research is essential for new ideas and work to be established.
Despite all the research being carried out on the staling of coffee, very few studies reviewed here were able to meet the scientific needs of the specialty coffee community. This has been the motivation for the SCAA to develop a plan to investigate coffee staling and its role in the specialty industry. During the next year, the SCAA and the Roasters Guild will be implementing a series of experiments to test both the speed of and first detectable staling, and the customer acceptability limit of coffee staling in different packaging methods. We will gather data from highly trained and practiced coffee cuppers on staling in a variety of coffees in a small number of packaging methods. Next, we will move to a limited customer taste test of coffees varying in age and packaging method. From this customer data we will be able to both understand the average customers’ ability to taste staling, and also assess the individual shelf-lives of different packaging methods used in the specialty community. Results of these experiments will be compiled and analyzed in a detailed report to be published by the SCAA.
Alves RMV, Mori EE, Milanez CR & Padula M. 2001. Roasted and Ground Coffee in Nitrogen Gas Flushing Packages – II. Proc. 19th ASIC. Trieste.
Anese M, Manzocco L & Nicoli MC. 2006. Modeling the Secondary Shelf Life of Ground Roasted Coffee. Journal of Agricultural and Food Chemistry 54(15):5571-5576.
Apostolopoulos D & Gilbert SG. 1988. Frontal Inverse Gas Chromatography as Used in Studying Water Sorption of Coffee Solubles. Journal of Food Science 53(3):882-884.
Baggenstoss J, Poisson L, Luethi R, Perren R & Escher F. 2007. Influence of water quench cooling on degassing and aroma stability of roasted coffee. Journal of Agricultural and Food Chemistry 55(16):6685-6691.
Beeman D, Songer P & Lingle T. 2011. Water Quality Handbook, 2nd ed. Long Beach, CA: Specialty Coffee Assocation of America.
Bezman Y, Nini D, Cohen Z, Ofek A, Bensal D & Garcia AL. 2008. Analytical Method for Monitoring Freshness of Roast and Ground Coffee. Proc. 22nd ASIC Campinas.
Cappuccio R, Full G, Lonzarich V & Savonitti O. 2001. Staling of Roasted and Ground Coffee at Different Temperatures: Combining Sensory and GC Analysis. Proc. 19th ASIC. Trieste.
Cardelli C & Labuza TP. 2001. Application of Weibull Hazard Analysis to the Determination of the Shelf Life of Roasted and Ground Coffee. LWT – Food Science and Technology 34(5):273-278.
Czerny M, Mayer F & Grosch W. 1999. Sensory Study on the Character Impact Odorants of Roasted Arabica Coffee. Journal of Agricultural and Food Chemistry 47(2):695-699.
Czerny M & Schieberle P. 2001. Changes in Roasted Coffee Aroma during Storage – Influence of the Packaging. Proc. 19th ASIC. Trieste.
Fu B & Labuza TP. 1993. Shelf-Life prediction: theory and application. Food Control 4(3):125-133.
Gacula MC. 1975. The Design OF Experiments For Shelf Life Study. Journal of Food Science 40(2):399-403.
Harris NE, Bishov SJ, Rahman AR, Robertson MM & Mabrouk AF. 1974. Soluble Coffee: Shelf Life Studies. Journal of Food Science 39(1):192-195.
Holscher W & Steinhart H. 1992. Investigation of Roasted Coffee Freshness with an Improved Headspace Technique Zeitschrift Fur Lebensmittel-Untersuchung Und-Forschung 195(1):33-38.
Huynh-Ba T, Courtet-Compondu MC, Fumeauz R & Pollien P. 2001. Early Lipid Oxidation in Roasted and Ground Coffee. Proc. 19th ASIC. Trieste.
Illy A & Viani R. 2005. Espresso Coffee: The Science of Qualtiy, 2nd ed. San Diego: Elsvier Academic Press.
Kallio H, Leino M, Koullias K, Kallio S & Kaitaranta J. 1990. Headspace of roasted ground coffee as an indicator of storage time. Food Chemistry 36(2):135-148.
Labuza TP, Cardelli C, Anderson B & Shimoni E. 2001. Physical Chemistry of Roasted and Ground Coffee: Shelf Life Improvement for Flexible Packaging. Proc. 19th ASIC. Trieste.
Leino M, Lapvetelainen A, Menchero P, Malm H, Kaitaranta J & Kallio H. 1992. Characterization of Stored Arabica and Robusta Coffees by Headspace-GC and Sensory Analysis. Food Quality and Preference 3:115-125.
Lingle T. 2011. The Coffee Cupper’s Handbook, 4th ed. Long Beach, CA: Specialty Coffee Association of America.
Marin K, Pozrl T, Zlatic E & Plestenjak A. 2008. A New Aroma Index to Determine the Aroma Quality of Roasted and Ground Coffee During Storage. Food Technology and Biotechnology 46(4):442-447.
Mayer F, Czerny M & Grosch W. 2000. Sensory study of the character impact aroma compounds of a coffee beverage. European Food Research and Technology 211(4):272-276.
Nicoli M, Calligaris S & Manzocco L. 2009. Shelf-Life Testing of Coffee and Related Products: Uncertainties, Pitfalls, and Perspectives. Food Engineering Reviews 1(2):159-168.
Nicoli MC, Innocente N, Pittia P & Lerici CR. 1993. Staling of Roasted Coffee: Volatile Release and Oxidation Reactions During Storage. Proc. 15th ASIC. Montpellier.
Poisson L, Koch P & Kerler J. 2006. Coffee Freshness Alteration of Roasted Coffee Beans Ground Coffee in the Presence of Oxygen and under Protective Conditions. Proc. 21st ASIC. Montpellier.
Prescott SC, Emerson RL & Peakes LV. 1937. The Staling of Coffee. Journal of Food Science 2(1):1-20.
Radtke-Granzer R & Piringer OG. 1981. Problems in the quality evaluation of roasted coffee though quantitative trace anlysis of volatile flavor components. Deutsche Lebensmittel-Rundschau 77(6):203-210.
Ross CF, Pecka K & Weller K. 2006. Effect of storage conditions on the sensory quality of ground Arabica coffee. Journal of Food Quality 29(6):596-606.
Sanz C, Pascual L, Zapelena MJ & Cid MC. 2001. A new ‘aroma index’ to determine the aroma quality of a blend of roasted coffee beans. Proc. 19th ASIC. Trieste.
Semmelroch P & Grosch W. 1995. Analysis of roasted coffee powders and brews by gas chromatography-olfactometry of headspace samples. LWT – Food Science and Technology 28(3):310-313.
Sivetz M. 1979. Coffee Technology. Westport, CT: The Avi Publishing Company, Inc.
Smith JP, Daifas DP, El-Khoury W, Koukoutsis J & El-Khoury A. 2004. Shelf Life and Safety Concerns of Bakery Products: A Review. Critical Reviews in Food Science and Nutrition 44(1):19-55.
Sortwell DR. 2008. Purging of Roasted Coffee Bean Containers. Proc. 22nd ASIC. Campinas.
Steinhart H & Holscher W. 1991. Storage-Related Changes of Low-Boiling Volatiles in Whole Coffee Beans. Proc. 14th ASIC. San Francisco.
Vila MA, Andueza S, de Peña MP & Concepción C. 2005. Fatty Acid Evolution During the Storage of Ground, Roasted Coffees. JAOCS, Journal of the American Oil Chemists’ Society 82(9):639.
Walter EHM, Capacla NC & Faria JAF. 2008. Evaluation of Flexible Packages With Degassing Valves for Roasted and Ground Coffee. Proc. 22nd ASIC.