Shrinkage Evaluation of Five Different Varieties of Coffee Berries during the Drying Process

ARTICLE IN PRESS Available online at www.sciencedirect.com

Biosystems Engineering (2003) 86 (4), 481–485 doi:10.1016/j.biosystemseng.2003.08.012 PH}Postharvest Technology

Shrinkage Evaluation of Five Different Varieties of Coffee Berries during the Drying Process P.C. Afonso, Jr1; P.C. Corr#ea2; F.A.C. Pinto2; C.P. Sampaio2 1

Embrapa Caf!e, Edif!ıcio Sede da Embrapa, Bras!ılia, DF 70770-901, Brazil; e-mail of corresponding author: [email protected] 2 Department of Agricultural Engineering, Universidade Federal de Vi-cosa, Vi-cosa, MG 36571-000, Brazil; e-mail: [email protected] (Received 29 October 2002; accepted in revised form 13 August 2003)

The objective of this work was to evaluate the effect of moisture content variation on reduction of superficial area, volume and equivalent sphere diameter of coffee berries. Four varieties of Coffea arabica (cv Catua!ı Vermelho, Catua!ı Amarelo, Mundo Novo and Catimor) and one variety of Coffea canephora (cv Conilon) were used. From the results obtained, it was concluded that moisture content in the coffee berries affects its physical properties causing significant decrease of the superficial area, volume and diameter of the equivalent sphere during a drying process. The varieties of coffee had different shrinkage behaviour. The Conilon coffee had the highest level of berry shrinkage, the volume decreased 35% during the drying process from 1
38 to 0
12 dry basis (d.b.) berry moisture content. The shrinkage behaviour during the drying process was well explained by a polynomial model with coefficient of determination greater than 90%. # 2003 Silsoe Research Institute. All rights reserved Published by Elsevier Ltd

1. Introduction The physical behaviour of coffee berries is the main factor when developing equipment, processes and simulation models for this crop. Information about volume, superficial area, shape and dimensions of a product are important for studies regarding mass and heat transfer as well as airflow within a granular mass of the product. These parameters and the moisture content are used for predicting storage and drying conditions of grain and cereals. The information can also be used for quality losses prediction during the postharvest period. Changes in the original dimensions and shape occur simultaneously and water diffusion affects the rate of the moisture loss during drying. Researchers have pointed out that the volume change process is one of the main sources of error for drying simulation models of biological products (Lang & Sokhansanj, 1993). Most of the mathematical models used to simulate the drying process of agricultural products have neglected the volume changing during dehydration (Brooker et al., 1992). However, those models have been improved incorporating the volume change phenomenon (Lang et al., 1994). 1537-5110/$30.00

There is a significant volume shrinkage during the drying process of high moisture content products such as vegetables (Hatamipour & Mowla, 2002; Lozano et al., 1993). Hatamipour and Mowla (2002) reported a 90% of volume shrinkage when carrots were dehydrated from 9
00 to 0 d.b. moisture content. On the other hand, the volume shrinkage is less significant during drying process of lower moisture content products such as grain (Sun et al., 2002; Afonso Jr et al., 2000). Afonso Jr et al. (2000) reported a 20% of volume shrinkage when grains of millet were dried from 0
27 to 0
09 d.b. moisture content. Moisture content of coffee berries at harvest can range from 4
00 to 0
18 d.b. (Pinto et al., 2002). Physical property changes during the drying process have been studied for different products (Vilela, 1977; Bala & Woods, 1984; Weber, 1995; Mcminn & Magee, 1997; Couto et al., 1999; Ruffato et al., 1999). These changes tend to be more evident for high moisture content biological products such as coffee. Vilela (1977) studied coffee berries shrinkage, and reported a decrease in volume of 34% when drying the product from 2
03 to 0
23 d.b. moisture content. The agricultural product shrinkage depends, not only on the moisture content variation, but also on the drying 481

# 2003 Silsoe Research Institute. All rights reserved Published by Elsevier Ltd

ARTICLE IN PRESS 482

P.C. AFONSO Jr ET AL.

process condition and product shape (Ratti, 1994; Zogzas et al., 1994). Most fruits and grains do not present a defined geometric shape. Therefore, a known shape is usually assumed for solving problems where the product geometry is important. Ellipsoidal geometric form with three characteristic dimensions}major, middle and minor axes}has been used to represent most of the agricultural products (Agrawal et al., 1972). Another approach is to use a sphere with the same volume as the product for representing it (Almeida, 1979; Soares, 1986). To find the diameter of such volume equivalent sphere, Mohsenin (1986) suggest the following equation: D ¼ ða b cÞ1=3

2. Materials and methods Four varieties of Coffea arabica (cv Catua!ı Vermelho, Catua!ı Amarelo, Mundo Novo and Catimor) and one variety of Coffea canephora (cv Conilon) were used in this study. The berries were harvested with approximately 1
50 d.b. moisture content, and dried to approximately 0
12 d.b. moisture content. An oven with circulating air was used to dry the berries at 3538C drying air temperature. Following the recommendation of Pimenta and Vilela (2001), the moisture content was measured by the oven method, until a constant weight was reached by using oven temperatures of 10538C. Three samples were used for each test. During the berry drying process, three orthogonal characteristic dimensions of the berry were measured by using a digital calliper. Assuming the berry had an oblate ellipsoid shape (Fig. 1), ten randomly sampled coffee fruits were measured for each test. As proposed by Mohsenin (1986), the volume of each fruit was estimated by pabc 6

b a

Fig. 1. Fruit characteristic dimensions a, b and c for assumed oblate ellipsoid shape

ð1Þ

where: D is the equivalent sphere diameter in mm; and a, b, c are orthogonal characteristic dimensions in mm. There have been some studies on coffee berry shrinkage during the drying process (Vilela, 1976; Couto et al., 1999). However, these studies were based on only one variety of the plant. Thus, the objective of this work was to study the effect of moisture content reduction on superficial area, volume and diameter of equivalent sphere for different varieties of coffee berries.

V¼

c

ð2Þ

where V is the berry volume in mm3. The unitary shrinkage volume of each variety was evaluated by the ratio between the initial volume of the berry and the volume at each moisture content.

The superficial area of the berry was estimated by the equation proposed by Mohsenin (1986):
 ½ðb þ cÞ=22 1þe 2 ln ð3Þ A ¼ 2pa þ p 1 e e where: A is the superficial area in mm2; and the eccentricity e is given by (   )1=2 ððb þ cÞ=2Þ 2 e¼ 1 ð4Þ a The diameter of a sphere having an equivalent was estimated using Eqn (1). The software STATISTICA 5
0 (StaSoft, Inc., OK) was used to fit the experimental data into the following polynomial model: Y ¼ b0 þ b1 U þ b2 U 2 þ E

ð5Þ

where: Y is the experimental physical characteristics of coffee berries, such as superficial area in mm2, volume in mm3, diameter of volume equivalent sphere in mm3 or volume shrinkage; b0 is the model constant; b1 and b2 are the model coefficients; U is the moisture content in dry basis; and E is the random error.

3. Results and discussion The experimental and estimated data for the superficial area, volume, diameter of volume equivalent sphere, and volume shrinkage are shown in Figs 2–5. The physical characteristics decreased during the drying process. The product structural modifications associated with cellular variation, due to water losses, are the causes for the berry shrinkage (Mcminn & Magee, 1997). The berry dehydration reduces the inter-cellular spaces. Thus, the berry dry matter takes the place where there was water before. Although the Coffea canephora (cv Conilon) had smaller orthogonal characteristic dimensions than those of Coffea arabica (cv Catua!ı Vermelho, Catua!ı Amarelo,

ARTICLE IN PRESS 483

SHRINKAGE EVALUATION OF COFFEE BERRIES

650.00

1.05 1.00 0.95 0.90

550.00

Volume ratio

Area, mm2

600.00

500.00 450.00 400.00 350.00 0.00

0.25

0.50

0.75 1.00 1.25 1.50 Moisture content (d. b.)

1.75

2.00

Fig. 2. Experimental and model estimated values of superficial area for coffee berries during the drying process; }, Catua!ı Vermelho; n, Mundo Novo; &, Catua!ı Amarelo; , Catimor; *, Conilon; }, estimated value 1200.00 1100.00 Volume, mm3

1000.00 900.00 800.00 700.00 600.00 500.00 400.00 0.00

0.25

0.50

0.75 1.00 1.25 1.50 Moisture content (d. b.)

1.75 2.00

Fig. 3. Experimental and model estimated values of volume for coffee berries during the drying process; }, Catua!ı Vermelho; n, Mundo Novo; &, Catua!ı Amarelo; , Catimor; *, Conilon; }, estimated value

Equivalent sphere diameter, mm

13.50 13.00 12.50 12.00 11.50 11.00 10.50 10.00 9.50 0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75 2.00

Moisture content (d. b.)

Fig. 4. Experimental and model estimated values of equivalent sphere diameter for coffee berries during the drying process; }, Catua!ı Vermelho; n, Mundo Novo; &, Catua!ı Amarelo; , Catimor; *, Conilon; }, estimated value

0.85 0.80 0.75 0.70 0.65 0.60 0.00

0.25

0.50

0.75 1.00 1.25 1.50 Moisture content (d. b.)

1.75

2.00

Fig. 5. Experimental and model estimated values of unitary volume shrinkage ratio V/V0 for coffee berries during the drying process; V, coffee berry volume at a specific moisture content;V0, initial coffee berry volume; }, Catua!ı Vermelho; n, Mundo Novo; &, Catua!ı Amarelo; , Catimor; *, Conilon; }, estimated value

Mundo Novo and Catimor), it had a higher level of shrinkage. Among the arabica coffee varieties, the Catua!ı Vermelho and Catua!ı Amarelo, the superficial area, volume and diameter of volume equivalent sphere were similar at the beginning and end of the drying process (Figs 2–4). The physical changes in Catimor and Mundo Novo varieties were also similar. Thus, the shrinkage behaviour of coffee berries was different, not only between species, but also among varieties. Analysing Fig. 5, it is noted that although Conilon started with the lowest moisture content, it ended up with the highest volume shrinkage, about 35%. Catimor berries also had about 35% of volume shrinkage. However, its berries were at a higher level of moisture content at the harvesting, leading to a greater initial volume than those of the other berries. The differences among the coffee berries also appeared in the polynomial model constants and coefficients (Table 1). The shrinkage behaviour during the drying process was well explained by polynomial model since most of the coefficient of determinations were greater than 90%. Shrinkage behaviour differences among the coffee species and varieties could be because of the differences in their external structures. For instance, the robusta coffee had thinner skin than that of arabica coffee (Matiello, 1998). Although the main focus of this research was in shrinkage behaviour of coffee berries during drying, the results showed that coffee berries shrinkage could not be negligible when modelling the coffee drying process. Using products that presented less shrinkage than coffee (Lang & Sokhansanj, 1993), researches have been improving drying simulation by incorporating the

ARTICLE IN PRESS 484

P.C. AFONSO Jr ET AL.

Table 1 The polynomial model constants and coefficients for superficial area A, volume V, equivalent sphere diameter D and unitary volume shrinkage ratio V/V0 of different coffee berries varieties; V0, initial berry volume; R2, coefficient of determination Physical characteristics

R2

Model constant and coefficients b0

Catua!ı Vermelho A V D V/V0

434
327 636
091 10
653 0
698

Catua!ı Amarelo A V D V/V0

445
342 656
207 10
796 0
757

MundoNovo A V D V/V0

b1 48
498* 16
320n.s 0
195n.s 0
018n.s

b2 } 92
068* 0
392** 0
101*

0
960 0
978 0
976 0
978

44
397* 131
505* 0
656* 0
152*

} } } }

0
797 0
939 0
934 0
939

523
001 747
023 11
282 0
711

34
856* 179
029* 0
797* 0
170*

} } } }

0
802 0
965 0
961 0
965

Catimor A V D V/V0

525
047 769
038 11
336 0
670

50
639** 127
494** 0
460** 0
111**

54
734* 175
203* 0
720* 0
153*

0
945 0
976 0
979 0
976

Conilon A V D V/V0

398
354 566
097 10
221 0
700

0
380n.s 0
155** 0
070** 0
191**

57
889** 228
063* 1
154* 0
282*

0
850 0
966 0
973 0
966

*

Significant at the level of 1% of probability for the test T. Significant at the level of 5% of probability for the test T. n.s Not significant. **

shrinkage into the model (Spencer, 1972; Lang et al., 1994).

(3) the Conilon coffee berries resulted in the highest level of shrinkage; and (4) the shrinkage behaviour during the drying process was well explained by a polynomial model.

4. Conclusions From the investigation of shrinkage behaviour of five different varieties of coffee berries, the following conclusions were drawn: (1) the shrinkage behaviour of coffee berries were different not only between species (Coffea canephora and Coffea arabica) but also among varieties (Conilon, Catua!ı Vermelho, Catua!ı Amarelo, Mundo Novo and Catimor); (2) the berry physical characteristics}superficial area, volume, equivalent sphere diameter and unitary volume shrinkage}decreased during the drying process;

Acknowledgements This research was sponsored by EMBRAPA ! ). The authors have been sponsored by (PNP&CAFE the Brazilian Agencies CNPq and FAPEMIG. All the mentioned supports are gratefully acknowledged.

References Afonso Jr P C; Corr#ea P C; Andrade E T (2000). An!alise da varia-ca* o das propriedades f!ısicas e contra-ca* o volum!etrica dos gr*aos de milheto (Pennisetum glaucum) durante o

ARTICLE IN PRESS SHRINKAGE EVALUATION OF COFFEE BERRIES

processo de dessor-ca* o [Analysis of physical properties variation and volume shrinkage of millet}Pennisetum glaucum}grains during the drying process]. Revista Brasileira de Armazenamento, 25(1), 15–21 Agrawal K K; Clary B L; Schroeder E W (1972). Mathematical models of peanut pod geometry. ASAE Paper No 72–315 Almeida B V (1979). Determina-ca* o das propriedades f!ısicas de am#endoas de cacau [Evaluation of the properties of cocoa]. Master Thesis, Universidade Federal de Vi-cosa, Vi-cosa, Brazil Bala B K; Woods J L (1984). Simulation of deep bed malt drying. Journal of Agricultural Engineering Research, 30(3), 235–244. Brooker D B; Bakker-Arkema F W; Hall C W (1992). Drying and Storage of Grains and Oilseeds. AVI, Westport, USA Couto S M; Magalh*aes A C; Queiroz D M; Bastos I T (1999). Massa espec!ıfica aparente e real e porosidade de gr*aos de caf!e em fun-ca* o do teor de umidade [Densities and porosity of coffee grain for different moisture content]. Revista Brasileira de Engenharia Agr!ıcola e Ambiental, 3(1), 61–68 Hatamipour M S; Mowla D (2002). Shrinkage of carrots during drying in an inert medium fluidized bed. Journal of Food Engineering, 55, 247–252 Lang W; Sokhansanj S (1993). Bulk volume shrinkage during drying of wheat and canola. Journal of Food Process Engineering, 16(4), 305–314 Lang W; Sokhansanj S; Rohani S (1994). Dynamic shrinkage and variable parameters in Bakker–Arkema’s mathematical simulation of wheat and canola drying. Drying Technology, 12(7), 1687–1708 Lozano J E; Rotstein E; Urbicain M J (1993). Shrinkage, porosity and bulk density of foodstuffs at changing moisture content. Journal of Food Science, 48(5), 1497–1502 Matiello J B (1998). Caf!e conillon: como plantar, tratar, colher, preparar e vender [Conillon coffee: how to plant, !/ manage, harvest, hand and sell]. MAA/SDR/PROCAFE PNPC, Rio de Janeiro Mcminn W A M; Magee T R A (1997). Physical characteristics of dehydrated potatoes}Part I. Journal of Food Engineering, 33(1–2), 37–48

485

Mohsenin N N (1986). Physical Properties of Plant and Animal Materials. Gordon and Breach Publishers, New York Pimenta C J; Vilela E R (2001). Qualidade do caf!e (Coffea arabica L.), lavado e submetido a" diferentes tempos de amontoa no terreiro [Washed coffee}Coffea arabica L.}quality for different heaping time]. Revista Brasileira de Armazenamento, v. especial(2), 03–10 Pinto F A C; Corr#ea P C; Queiroz D M; Oliveira A S C (2002). Propriedade espectral para estimar a umidade dos frutos de caf!e [Spectral properties for moisture content estimation of coffee fruits]. XXXI Congresso Brasileiro de Engenharia Agr!ıcola, Salvador, BA, Brazil Ratti C (1994). Shrinkage during drying of foodstuffs. Journal of Food Engineering, 23(1), 91–105 Ruffato S; Corr#ea P C; Martins J H; Mantovani B H M; Silva J N (1999). Influ#encia do processo de secagem sobre a massa espec!ıfica aparente, massa espec!ıfica unit!aria e porosidade de milho-pipoca [Influence of drying process over the densities and porosity of popcorn]. Revista Brasileira de Engenharia Agr!ıcola e Ambiental, 3(1), 45–48 Soares J B (1986). Curvas de secagem em camada fina e propriedades f!ısicas de soja (Glicine max L.) [Thin-layer drying curves and physical properties of soybean]. Master Thesis, Universidade Federal de Vi-cosa, Vi-cosa, Brazil Spencer H B (1972). A revised model of the wheat drying process. Journal of Agricultural Engineering Research, 17(2), 189–194 Sun Z; Yang W; Siebenmorgen T J; Stelwagen A M; Jia C; Cnossen A G (2002). Using dynamic shrinkage tests to study fissure propagation in rice kernels. Transaction of ASAE, 45(5), 1501–1504 Vilela E R (1977). Secagem de caf!e em terreiro e silo com energia solar [Coffee drying using terrace and solar bin drier]. Master Thesis, UNICAMP, Campinas, Brazil Weber E A (1995). Armazenagem Agr!ıcola [Agricultural Storage]. Gr!afica e Editora La Salle, Porto Alegre, Brazil Zogzas N P; Maroulis Z B; Kouris D M (1994). Densities, shrinkage and porosity of some vegetables during air drying. Drying Technology, 12(7), 1653–1666