Preservation of roasted and ground coffee during storage . Part 2 : Bulk density and intergranular porosity

Conservação de café torrado e moído durante o armazenamento. Parte 2: Massa específica e porosidade R E S U M O A determinação das propriedades físicas é fator importante na formulação de projetos de maquinários e dimensionamento de operações pós-colheita. De forma a permitir o dimensionamento correto e o uso desses maquinários, objetivou-se avaliar e determinar, durante o armazenamento, as propriedades físicas: massa específica unitária e aparente e porosidade intergranular, além de avaliar a influência dos níveis de torrefação e moagem sobre essas propriedades. Grãos de café cru (Coffea canephora e Coffea arabica) foram utilizados, descascados e secados e só então torrados em dois níveis: média clara (SCAA#65) e moderadamente escura (SCAA#45). Os grãos foram moídos em três granulometrias: fina (0,59 mm), média (0,84 mm) e grossa (1,19 mm), além do lote de café inteiro. Realizada a moagem as amostras foram armazenadas em duas temperaturas (10 e 30 oC) e analisadas em cinco tempos distintos de armazenamento (0, 30, 60, 120 e 180 dias). A torra média clara permitiu maiores valores das propriedades físicas; já as amostras de granulometria fina apresentaram aumento das massas específica unitária e aparente e porosidade. Conclui-se que a granulometria, o nível de torrefação e o tempo de armazenamento, afetaram as propriedades físicas do café.


Introduction
The preservation of whole and ground roasted coffee is necessary for its commercialization.Different factors can influence the preservation of coffee and, consequently, its final quality, such as roasting, grinding and storage.
Roast level has been reported to determine the final quality of the beverage (Melo, 2004;Baggenstoss et al., 2008), and significant changes to coffee's physical properties have been observed to occur during roasting (Mwithiga & Jindal, 2003).Grinding results in powdery products that can have different particle sizes according to the market's needs (Schmidt et al., 2008) and affect the physical properties of coffee (Geldart et al., 2009;Langroudi et al., 2010).
Storage of roasted and ground coffee is not recommended because grinding promotes cell breakage and, therefore, enables the loss of molecular components, negatively affecting quality.However, studying the physical properties of roasted and ground coffee is important due to potential market conditions.Previously processed products may need to be stored due to a lack of transportation, prices that make immediate commercialization impossible and the need to formulate coffee blends.In addition, bulk storage can be a great advantage, because it enables the mechanization of coffee processing, which substantially decreases the amount of labor required compared to traditional storage methods (Oliveira et al., 2014).Storage time was also observed to influence the physical properties of other products such as wheat flour, tea, whey (Teunou & Fitzpatrick, 2000;Iqbal & Fitzpatrick, 2006), dairy powders (Fitzpatrick et al., 2007) and poultry feed (Nóbrega & Nascimento, 2005;Lopes Neto et al., 2009).
The aim of the present study was to evaluate physical properties (true density, bulk density and intergranular porosity) of coffee roasted to different levels, ground to different sizes and stored at different temperatures.

Material and Methods
The present study was performed at the Laboratory of Physical Properties and Quality of Agricultural Products of the National Center of Storage Training (CENTREINAR) located at the Federal University of Viçosa-MG, Brazil, and at the Laboratory of Ceramic Materials of the Department of Metallurgical and Materials of the Federal University of Minas Gerais, in Belo Horizonte, MG.
Raw coffee beans (Coffea canephora and Coffea arabica), hulled and dried, were acquired from the regional markets of Zona da Mata, Minas Gerais.The beans were sorted to eliminate deteriorated, damaged and bored beans and to obtain a homogeneous lot with minimal defects.The initial moisture content of the beans was determined through gravimetry using a forced-air circulation oven at 105 ± 1 °C for 24 h in triplicate.The average initial moisture content was 12.61% (dry basis).
Following sorting, coffee beans from both species were roasted using a direct-flame LPG gas roaster with a rotating cylinder at 45 rpm with a 350-g raw coffee roasting capacity.The different roast levels were identified through monitoring and comparison of sample colors with the Agtron color reference standards, with the assistance of a trained professional.Two roast levels were obtained: medium light (ML) and moderately dark (MD), corresponding to the Agtron standards SCAA#65 and SCAA#45, respectively (Figure 1).
To guarantee the uniformity of the roast, a portable infrared thermometer (Mult-Time) with a temperature range of -50 to 500 ºC, 1 s response time and 0.1 ºC resolution was used.ML and MD roasted beans lost an average of 15.85 and 18.74 g when the roaster reached 285 ºC, respectively (Vargas-Elías, 2011).When each roast level was reached, the beans were removed from the roaster and immediately cooled at ambient temperature.
Following roasting, the beans were ground with a mill (Mahlkönig) into three different grind sizes according to the ABIC (2013) (Table 1): fine (0.59 mm), medium (0.84 mm) and coarse (1.19 mm).An additional lot was kept whole.
The processed samples were then placed in polypropylene bags and stored in B.O.D. chambers at two different temperatures (10 and 30 ºC) for six months.Physical properties of the stored samples were analyzed at five different times during storage (0, 30, 60, 120 and 180 days) as described below.
Source: ABIC ( 2013 Table 1.Classification based on the percentage retention in sieve numbers 12, 16, 20, 30 and at the sieve bottom pan, with agitation for 10 min and the rheostat set at position 5 A. True density (ρ u ) was determined using a helium pycnometer, according to Eqs. 1 and 2, with five replicates.The helium pycnometer used was a Multipycnometer MVP-4DC (Quantachrome Corporation, USA), which operates according to the gas displacement principle.
Bulk density (ρ ap ) was determined using a Burrows hectoliter weight scale, with 1-L capacity, on a steel plate, with five replicates.Intergranular porosity (ε) was determined indirectly according to Eq. 3 (Mohsenin, 1986), traditionally used for agricultural products.
The experimental design was completely randomized, with the number of replicates varying depending on the physical property measured.Two experiments were performed: one using Coffea arabica L. (cv.'Catuai vermelho') and another one using Coffea canephora Pierre (cv.'Emcapa 8111').ANOVA, followed by Tukey's test, was performed for both experiments, with significance at p < 0.05.All statistical analyses were performed using SAEG® software.For the comparison of different storage times, a regression analysis using average values was performed.The best-fit models were selected based on the highest coefficient of determination (R 2 ) and the significance of the parameters.

Results and Discussion
Roast level, storage temperature, storage time and grind size affected the true density (ρ u ) of roasted coffee, regardless of the coffee species tested.Thus, only the results of Arabica coffee will be presented (Figure 2).
Higher grind sizes, which result in smaller particle sizes, resulted in higher ρ u (Figure 2).This result is based on the fact that, for a given constant volume, finer ground coffees were more tightly agglomerated, resulting in higher weights per volume, whereas larger particles undergo less agglomeration, A. resulting in lower weights and, consequently, lower ρ u values (Singh et al., 1997).Medium light roasted coffee exhibited higher ρ u , regardless the coffee species (Figure 2).This result is explained by the variations in volume and weight during roasting: lighter roast (1) (2) (3) levels result in smaller volume increases and less weight loss, resulting in higher ρ u .Singh et al. (1997) studied Mexican and Colombian coffee and also concluded that higher roast levels result in decreased ρ u .Vargas-Elías (2011) reported a ρ u decrease in coffee with higher roast levels due to greater weight loss.
The differences in ρ u at different roast levels may also be explained by changes at the cellular level.Licciardi et al. (2005) reported that the triglycerides (oils) of coffee beans are mainly unaffected by roasting, suffering only slight hydrolysis and decomposition, with the release of fatty acids and the formation of volatile products.However, in darker roasted coffee, many cells break and oils may migrate to the surface (França et al., 2001), resulting in greater weight loss and decreases in ρ u .This effect has also been observed in milk powder (Fitzpatrick et al., 2004).
Storage at 10 ºC resulted in the maintenance of the initial ρ u values (post-roasting).This result was mainly due to maintenance of volume.This finding, together with the small changes in weight due to moisture absorption, resulted in small variations of ρ u during storage.Samples stored at 30 ºC exhibited greater variations in ρ u .This result was due to more agglomeration at the higher storage temperature, which affected ρ u .
During storage, ρ u exhibited a general tendency to decrease compared to its initial values (Figure 2), because true density depends on moisture content and decreases with increased moisture content (Resende et al., 2008;Oliveira et al., 2015).This behavior was observed in the present study; a more pronounced decrease in ρ u was observed starting at 60 days of storage and was accompanied by increased moisture content, as previously discussed.
Roast level, storage temperature, storage time and grind size also had effects on the bulk density of roasted coffee, regardless of the species studied.Thus, only the results of Robusta coffees during storage for six months will be presented (Figure 3).
Bulk density was influenced by the interaction between roast level and grind size, varying between 300 and 410 kg m -3 .These values are within the range of 300 to 450 kg m -3 reported by Illy & Viani (1995) for roasted coffee.
Whole coffee beans showed higher ρ ap compared with ground coffee.This result is directly related to the content of volatile components, which are related to the total weight of the product.Whole beans release CO 2 slowly, with complete release occurring after approximately 30 days, whereas 70% of the CO 2 content is released immediately following grinding.This results in higher weights for whole beans (Illy & Viani, 1995).
For ground coffee, ρ ap decreased with increased particle size.Yan & Barbosa-Canovas (1997) studied the compression characteristics of food powders and also reported this relationship between grind size and ρ ap .The authors explained that higher particle sizes result in the product occupying less volume, i.e., there are more pores and, therefore, lower weights per volume, resulting in lower ρ ap .
As observed for ρ u , medium light roasted coffee exhibited higher ρ ap due to its lower weight loss and lower roasting time.This effect was previously observed in several studies on coffee (Illy & Viani, 1995;Singh et al., 1997;Mwithiga & Jindal, 2003;Mendonça et al., 2009;Vargas-Elías, 2011).Moderately dark A. roasting results in greater weight loss and increased coffee volume and, therefore, lower ρ ap .The change in volume is related to the increase in internal cell pressure that occurs during heat transfer and in pyrolysis reactions, which are more intense during longer roasts (Borges et al., 2004).
Storage at 30 ºC resulted in a more pronounced decrease in ρ ap in comparison to storage at 10 ºC (Figure 3).Adequate storage temperatures enable the preservation and/or decrease the loss of cellular components of agricultural products, resulting in better preservation during the shelf life of a product, which is consistent with the present results.This effect was observed more clearly in medium light roasted coffee (ML), indicating that the roast level may have more significant effects on the physical properties of coffee than on the storage temperature.
An effect between coffee roast level and grind size was observed with respect to intergranular porosity.The results of samples stored at 30 ºC are not presented because no significant differences were observed between the two storage temperatures tested, regardless of the species of coffee evaluated.The variation in intergranular porosity during storage at 10 ºC is presented in Figure 4.
Intergranular porosity varied with grind level, with higher intergranular porosity observed in finely ground coffee and decreasing with increasing particle size.The lowest intergranular porosity was observed in whole beans (Figure 4), due to agglomeration.Schubert (1987) reported that intergranular porosity increases with decreased particle size, because the adhesion between particles permits a loose structure.Pegg & Shahidi (2007) reported that products with lower grind sizes tend to agglomerate, forming aggregates larger than the bean itself and resulting in higher intergranular porosity, which is in agreement with the present study.

Conclusion
True density, bulk density and intergranular porosity of coffee are proportional to the increased grind level, decreased roast level and decreased storage time.

Figure 2 .
Figure 2. True density of roasted Arabica coffee during storage at 10 and 30 ºC.Coffee beans were whole (A) or ground, with fine (B), medium (C) or coarse (D) grind sizes

Figure 3 .
Figure 3. Bulk density of roasted Robusta coffee during storage at 10 and 30 ºC.Coffee beans were whole (A) or ground, with fine (B), medium (C) or coarse (D) grind size