Sensory profile of conilon coffee brews from the state of Espírito Santo , Brazil

The objective of this work was to present a sensory description of conilon coffee (Coffea canephora) brews produced from genotypes developed for the state of Espírito Santo, Brazil. The genotypes 'Diamante ES8112', 'ES8122', and 'Centenária ES8132' were evaluated. Nine samples from three clones (genotypes) of each cultivar, from two different regions of the state, were analyzed. A panel of tasters for coffee sensory evaluation was trained to obtain a descriptive profile and to select terms. Beverages were described as showing a higher intensity of powder fragrance, characteristic aroma of coffee brew, body, characteristic flavor and aftertaste, and a lower intensity of bitterness, acidity, astringency, and defects. The terms bitterness, strong, body, astringency, and coffee aftertaste were the most relevant in the characterization of the beverages of the studied C. canephora cultivars. Acidity is the attribute with the greatest impact on the overall quality of C. canephora brew. The coffee brews of the clones of 'Diamante ES8112', 'ES8122', and 'Centenária ES8132' are classified as of traditional quality, but near the superior category.


Introduction
Brazil is the largest coffee country in the world, with an estimate of 44.77 million bags (60 kg) for the 2017 crop season.Currently, it is the second largest conilon coffee (Coffea canephora) producer, with the production concentrated in the states of Espírito Santo (58%) and Rondônia (18%) (Acompanhamento…, 2017).
Overall, C. canephora production grew more than 400% in the last 56 years, from 12.6 million bags in 1960, to 66.6 million bags in the 2015/2016 season (USDA, 2017).Coffea canephora has assisted Brazil to maintain its position as the largest coffee grower in the world.In the 2017 season, it represented approximately 24% of the total Brazilian production (Acompanhamento…, 2017).Another relevant aspect of this species is its adaptation to regions with low altitude and high temperatures, and its resistance to prolonged water deficits (Ferrão et al., 2017), which allows of the expansion of the planted area, as it shows fewer restrictions for growing regions than C. arabica.
The C. canephora breeding program of Instituto Capixaba de Pesquisa, Assistência Técnica e Extensão Rural (Incaper) has developed different genotypes of high yield, adaptation, and production stability associated to other important agronomic characteristics, such as cup-quality and maturation of fruit in different seasons.The use of cultivars with different fruit-ripening seasons (early, medium, and late maturation) is advantageous for the coffee grower, as they allow of harvesting in a staggered way and during a longer period, with a better management of the workforce and of the use of coffee yards and dryers (Ferrão et al., 2009(Ferrão et al., , 2017)).
Coffee quality varies along the production chain, which goes from the coffee-grower to the endconsumer.For growers, it varies by join productivity, price, and crop facility, while for consumers, coffee quality is associated with price, flavor, and aroma, health effects, geographical origin, and environmental and sociological aspects (Leroy et al., 2006).
The Associação Brasileira da Indústria de Café (ABIC) recommends a sensory evaluation of the overall quality of the beverages, in order to monitor the quality of roasted coffee commercialized in Brazil (ABIC, 2017).This requirement is already present in specific legislation of some states in the country that are traditional producers of C. arabica.The state of São Paulo stands out for having a regulation establishing the minimum quality standards for classification, such as the Resolutions SAA no. 30 and no. 31 of June 22, 2007, and SAA no. 19 of April 5, 2010(São Paulo, 2007b, 2007c, 2010).Other Brazilian states adopt similar criteria for purchasing roasted coffee products for consumption in the public administration sphere, as those described by the decree 2916 of 06/25/2008 from Paraná state (Paraná, 2008), decree 44.661 of 11/26/2007 from Minas Gerais state (Minas Gerais, 2007), and decree 1674-R of 05/25/2006 from Espírito Santo state (Espírito Santo, 2006).The cupquality evaluation of the coffee brewd was included as mandatory in the national legislation, according to the Normative Instruction 16/2010 (Brasil, 2010), revoked later by the Normative Instruction 7/2013 (Brasil, 2013), justified by the difficulty in maintaining trained sensory panels and certified laboratories to carry out the evaluation.
It is important to highlight that any roasted coffee sold in Brazil -regardless of the species -certified by the Coffee Quality Program (CQP), or subjected to inspection by state and federal agencies, is subjected to a coffee brew sensory quality evaluation.The evaluation is based on the ABIC quality regulation (ABIC, 2017), or on specific standards for each Brazilian state, whenever applicable.Based on the CQP, roasted coffee brews with General Recommendable Quality, either coffee beans or grounded coffee, are those constituted of C. arabica, or blended with C. canephora, that meet the overall quality requirements by ABIC (2017).However, the sensory characteristics evaluated by current regulations apply to characteristics of C. arabica brews, undervaluing the quality evaluation in the case of the inclusion of C. canephora in the blends.Except for Moura et al. (2007), who studied the effect of adding C. canephora to blends, the literature shows no studies related to the quality of beverages of this coffee species.
The objective of this work was to present a sensory description of C. canephora brews produced from genotypes of clonal cultivars developed for the state of Espírito Santo, Brazil.
Samples (500 g of each coffee genotype) were collected during 2014, from 36-month-old representative crops, in different regions of the state of Espírito Santo, as follows: an experimental farm of Marilândia, located in the municipality of Marilândia, in the northwestern region of the state (19º24'S, 40º31'W); and experimental farm of Bananal Norte, in the municipality of Cachoeiro do Itapemirim, in the southern region of the state (20º75'S, 41º29'W).Manual harvesting of early-maturing genotypes started in May and was followed by that of mediummaturing (June) and late-maturing (July) ones; only cherry-stage fruit were collected.Coffee was dried in the coffee yard, in the sun, and then it was processed.Nondefective beans of sieve size 16 (6.5 mm) were selected.Green coffees were kept in plastic bags at room temperature, until roasting in February 2016.
Each sample (genotype) was represented by a mixture of the green beans originated from each experimental farm.Samples were roasted in a gas pilot roaster (Rod Bel, São Paulo, SP, Brazil) for 15-17 min, at temperatures between 210 and 230°C.Roasting was standardized to obtain weight loss of 16.5±0.2g 100 g -1 , described by Mendes et al. (2001) as the optimum for C. canephora.Coffees showed a roast degree of 65.8±6.7,evaluated by the E10-CP process analyzer (Agtron Inc., Reno, NV, USA), corresponding to a medium to light roasting, according to the Agtron System.
Samples were ground at fine granulometry, according to the recommendation of ABIC (2017), using the Top laboratory mill (La Spaziale Brasil & America Latina, São Paulo, SP, Brazil).
The coffee brew was prepared, according to the ABIC recommendations (ABIC, 2017), through percolation with paper filter (Melitta, São Paulo, SP, Brazil), using 50 g of roasted and ground coffee to 500 mL of mineral water (Ouro Fino, Curitiba, PR, Brazil), at the 1:10 proportion of coffee:water, at 92°C.The samples were stored in thermal bottles and kept until analyses, up to 30 min.
Beverages were characterized for caffeine and total chlorogenic acid contents according to Corso et al. (2016), and total titratable acidity according to Scholz et al. (2013).Analyses were carried out in duplicate, and results were expressed as a range for each cultivar (considering the three genotypes, and the two growing sites).Caffeine and chlorogenic acids (µg mL -1 of coffee brew), and the total titratable acidity (mL of sodium hydroxide for 20 mL of the coffee brew) were reported.
Sensory analyses were carried out at the laboratory of physical and sensorial analysis of Instituto de Tecnologia de Alimentos de Campinas (Ital), and registered in the laboratory network for sensory evaluation of coffee brew quality (ABIC, 2017).The sensory panel, previously trained, was composed by ten assessors, between 28 and 56 years of age, including three men and seven women.This panel is retrained and validated annually, and participants' experience with coffee sensorial evaluation was between 5 to 10 years.
The analyses were performed under red light in individual booths, equipped with the Compusense Five system, version 5.6 (Compusense, Guelph, ON, Canada), for data collection.Data were analyzed by the XLStat 2016.01.26779 software (Addinsoft, Paris, France).
Assessors received 50 mL of the beverages (at a temperature between 66 and 69°C) in porcelain cups, black inside and white outside, identified with a randomized three-digit code.Samples were presented in a monadic way.During the session, the assessors first evaluated the descriptive profile of the coffee brew, and then they started the procedure of term selection.The order of presentation of the samples was randomized for each assessor.Three sessions were performed, and during each one, three different samples were evaluated, to obtain a genuine duplicate of each coffee brew.
The research was approved by the human research ethics committee of Universidade Estadual de Londrina (CAAE 21939713.3.0000.5231).
The descriptive profile analysis was carried out according to the Resolution SAA no. 30 of June 22, 2007, and SAA no.28 of June 1, 2007 (São Paulo, 2007a, 2007c), in a procedure equivalent to that advocated by ABIC (2017).The assessors filled out an individual score sheet for each sample, at each evaluation.Each parameter -powder fragrance, characteristic aroma of coffee brew, defects, acidity, bitterness, characteristic flavor of coffee brew, residual flavor (aftertaste), body, and astringency -, as well as the overall quality of the beverages, were evaluated individually using an unstructured 10 cm scale.One sample classified as superior was used as a reference, similarly to the described by Moura et al. (2007) and Mori et al. (2003).Data were evaluated by the analysis of variance, considering the samples as the source of variation, and Tukey's test, at 10% probability.Results for each parameter were expressed as mean values for the sensory panel (duplicate analysis and ten assessors).
For the overall quality parameter, the analysis was based on the Resolutions SAA no. 19, SAA no. 30, and SAA no.31, which define the minimum quality standards for roasted coffee; for beans and grounded coffee, classified as traditional, coffee shows an overall quality score from 4.5 to 6.0; for the rank superior, overall quality score is from 6.0 to 7.3; or gourmet, overall quality score from 7.3 to 10 (São Paulo, 2007b, 2007c, 2010).
For the procedure for selection of terms, a Focus Group session was previously performed, in order to choose the descriptive terms to be used.The 172 terms presented in the flavor wheel from Specialty Coffee Association of America (SCAA) -94 of which are associated to coffee characteristics, and 78 are associated with coffee defects (SCAA, 2015) -were used for discussion.The group was requested to highlight the term related to Brazilian roasted coffees, which were noticed during routine sensory evaluations.Therefore, 22 terms were selected for further sample analyses, as follows: sweet, acid, bitter, slightly sweet, slightly acid, slightly bitter, astringent, nutty/hazelnut, chocolate, watery, characteristic residual flavor of coffee, floral/fruity, smoky/tobacco, strong, weak, burnt, caramelized, herb-like/tea, fullbodied, noncharacteristic residual flavor, spice/black pepper/nutmeg, and chemical /resinous/medicinal.
The evaluation was done straight after the descriptive profile.Assessors were asked to select the terms among the 22 previously defined, which would apply to coffee brews.Terms presentation order was randomized for each sample and assessor, according to Ares et al. (2014).Among the 22 presented terms, the selected ones were those showing difference in the frequency of citation among coffee brews at 20% probability, and those cited by more than 20% of the assessors.Terms selection results were reported as frequency of citation and were subjected to the Cochran Q-test, followed by the Marascuilo multiple comparison test, at 5% probability.

Results and Discussion
In general, the scores of the coffee brews from the genotypes of three C. canephora cultivars were classified in the categories traditional and superior (Table 1).Except for acidity and overall quality, there was no difference among the mean scores of the attributes for the nine coffee brews evaluated, considering the three genotypes of each of the three cultivars.
Coffee brews showed a high intensity of the attributes powder fragrance, characteristic aroma of coffee brew, body, characteristic flavor of coffee brew, and residual flavor with mean scores from 5.6 to 6.1 (Table 1).The body of coffee brews of the three evaluated cultivars (mean score 6.0) was superior to that reported by Aguiar et al. (2001), who obtained values between 4.81 and 5.66 for beverages of three C. arabica cultivars (Ouro Verde, Tupi, and Obatã).The panel showed consensus, that is, coefficients of variation between 8 and 11%, in describing these attributes, even considering that they were evaluating C. canephora beverages.The coffee brews were also described as showing less bitterness (mean score 4.6), acidity (scores from 3.6 to 4.1), and astringency (mean score 3.8).Moura et al. ( 2007) reported a score of 3.4 for astringency of 100% C. canephora beverage, similarly to that observed in the present study.The authors also reported that the C. canephora brew showed acidity score of 1.9, which is inferior to the obtained one for a beverage of 100% C. arabica (score 3.6).It is important to emphasize that coffee brews in the present study obtained acidity scores close to those reported by Moura et al. (2007) for a C. arabica brew.
Acidity was the only specific attribute for which the difference between cultivars was observed.The beverage from 'Diamante ES8112', genotype 105P, was described as less acid than that from the 'ES8122', genotype 202I (Table 1).For the attributes acidity and astringency, a great variability (CV around 16%) was observed among the scores assigned by the panel.
The sensory parameters and ranges required by the recommendations are characteristic of beverages of C. arabica species.Therefore, it was expected that some specific sensory attributes of the C. canephora coffee brew, such as low acidity and bitterness, and marked astringency (Williams et al., 1989) could cause a disadvantage for the classification of C. canephora beverages.However, these beverage parameters (Table 1) were very near to the superior category, according to resolution SAA no.28 of June 1, 2007(ABIC, 2017).
Although the raw material used in the present study had no defects and coffee was collected and processed under adequate conditions, beverages from genotypes of the three cultivars were described as showing defects (score 3.6) (Table 1).It should be highlighted that the parameter "defect" showed the greater variability among the assessors (CV of 21%), indicating a greater difficulty and lesser consensus in the evaluation by the panel.Probably, according to the sensory evaluation protocol used, specific characteristics of C. canephora beverages may have been considered as "defects".However, the defect scores were inferior to those reported by Schmidt et al. (2012) for C. arabica 'Iapar 59', cultivated in the state of Paraná in different types of soil, and subjected to medium and dark-roasting processes (scores from 5.01 to 5.54).
Beverages from genotypes of the three cultivars had overall quality scores from 5.5 to 6.0 (Table 1), at the superior limit of the traditional category.These scores were higher than those reported by Moura et al. (2007) for a coffee blend with 70% of C. arabica (score 5.4) and by Schmidt et al. (2012) for C. arabica 'Iapar 59' (with 4.32 to 4.88).The beverage from cultivar 'ES8122' genotype 202I was distinguished by an overall quality score of 6.0 (superior quality coffee), which is higher than that assigned to the beverage from 'Diamante ES8112' genotype 105P (Table 1).This behavior stressed the importance of acidity to the overall quality evaluation of coffee brews.
When evaluated by the specific protocol for C. canephora of the Coffee Quality Institute, conilon coffee brews of 'Diamante ES8112', 'ES8122', and 'Centenária ES8132' stood out for showing a high-cup quality, from 77.50 to 79.01, and were classified as "Premium" with 70 to 80 points ('Centenária ES8132'..., 2015;'Diamante ES8112'..., 2015;'ES8122' Jequitibá..., 2015).The present research confirms this information, and shows that even when evaluated by a general protocol, the genotypes from the evaluated cultivars showed a good potential that was considered close to a superior category coffee brew.There is a great interest in cultivars that show distinct fruit-ripening seasons and greater production viability (staggered harvesting, helping labor management, and the use of coffee yards and dryers).It is also important that these cultivars have similar coffee brew sensory characteristics, to allow of the production of C. canephora with good quality in a longer period during the year; besides, the characteristics of these cultivar allow of the use of the studied genotypes in greater quantities and proportion by the industry.
Pesq. agropec.bras., Brasília, v.53, n.9, p.1061-1069, Sept. 2018 DOI: 10.1590/S0100-204X2018000900010 highlighted, and it was mentioned by assessors for all beverages, and frequency of citation varied from 50 to 77%.The term slightly bitter was mentioned, but in a much lower proportion.Astringency (22 a 44%), coffee characteristic flavor, and the terms strong and full-bodied were also considerably cited for all coffee brews.Gloess et al. ( 2013) compared sensory and composition characteristics of coffee brews prepared by different methods, and concluded that caffeine content is positively associated with bitterness, and the content of chlorogenic acids, with astringency.
Beverages from the genotypes of 'Diamante ES8112', 'ES8122', and 'Centenária ES8132' showed caffeine content from 1,176 to 1,660 µg mL -1 (Table 3).Caffeine content of 314 to 1,153 µg mL -1 was reported for C. canephora beverages (Rodrigues et al., 2007;Perrone et al., 2012), but the coffee brews were prepared with a lower proportion of coffee/water than that reported by the present study.Perrone et al. (2012) found total chlorogenic acids from 633.8 to 1,417.5 µg mL -1 , for C. canephora beverages (produced with coffee roasted with weight loss near to that of the present study); these values are similar to those of 493 to 1,130 µg mL -1 , observed in the present study, for coffee brews from the three cultivars (Table 3).
The term acid was mentioned, but there was a high perception that the beverages showed low acidity (Table 2).The term slightly acid showed a difference in the frequency of citation among the beverages, varying from 'Centenária ES8132' (genotype 303T), less acid, with 44% of the citations to 'ES8122' (genotype 203I), with 5% of citations. (1)Frequency of citation by the sensory panel (ten assessors), with two evaluations per sample. (2)Numbers refer to the genotypes of cultivars.Means unidentified with letters do not differ significantly by the Marascuilo test, at 5% probability, from others in the same row.Means followed by different lowercase letters, in the same row, show significant difference between genotypes, by the Marascuilo test, at 5% probability.

Table 3 .
Contents of caffeine, total chlorogenic acids, and total titratable acidity (TTA) of conilon coffee (Coffea canephora) brews of the three genotypes of each cultivar.