PHYSICAL AND MECHANICAL PROPERTIES CORRELATION OF COFFEE FRUIT (Coffea arabica) DURING ITS RIPENING

CORRELACIÓN DE PROPIEDADES FÍSICAS Y MECÁNICAS DEL FRUTO DE CAFÉ (Coffea arabica) DURANTE SU MADURACIÓN

 

IVÁN DARÍO ARISTIZÁBAL TORRES
D.Sc. Profesor Asociado, Universidad Nacional de Colombia Sede Medellín, idaristi@unal.edu.co

JOSÉ JAIME CARVAJAL HERRERA
Ingeniero Agrícola, Universidad Nacional de Colombia Sede Medellín, jjcarvaj@unal.edu.co

CARLOS EUGENIO OLIVEROS TASCÓN
Ph.D. Investigador Principal, Ingeniería Agrícola, CENICAFE, carlos.oliveros@cafedecolombia.com

 

Received for review July 15^th, 2011, accepted February 2^th, 2012, final version February, 12^th, 2012

 

---

ABSTRACT: There were correlated CIELAB color coordinates and reflectance in the visible spectrum of the exocarp of Coffea arabica coffee fruits of the Colombia variety in 9 different development stages, with physical properties such as fresh and dry mass, moisture content, and diameters; and mechanical properties such as detaching tensile force, fracture force, firmness, deformation, and fracture energy in the breaking point. A multivariate linear regression analysis was run to obtain models to accurately predict properties of the coffee fruit from the chromatic coordinates (a* and/or b*) or reflectance for discriminating wave lengths of different stages of development of the coffee fruit, with a confidence level of 99% and an R2 between 78.5 and 99.9%. Results could be used for the design of harvesting and classification systems, as well as for a rapid determination of coffee quality.

KEYWORDS: Coffea arabica, color, reflectance, physical properties, mechanical properties

RESUMEN: Se correlacionaron las coordenadas del espacio de color CIELAB y la reflectancia en el espectro visible del epicarpio del fruto de café Coffea arabica, variedad Colombia, de 9 diferentes estados de desarrollo, con propiedades físicas como masa fresca y seca, contenido de humedad y diámetros, y propiedades mecánicas como fuerza de desprendimiento del fruto, firmeza, fuerza de fractura, deformación y energía requerida para fracturar la pulpa. Utilizando análisis de regresión lineal multivariado se obtuvieron modelos que predicen significativamente propiedades del fruto de café a partir del conocimiento de las coordenadas cromáticas (a* y/o b*) o de la reflectancia para diferentes longitudes de onda discriminantes de estados de desarrollo, con un nivel de confianza del 99% y un R2 entre 78,5 y 99,9 %. Los resultados obtenidos podrían ser utilizados para el diseño de sistemas de recolección, clasificación y determinación rápida de la calidad del café.

PALABRAS CLAVE: Coffea arabica, color, reflectancia, propiedades físicas, propiedades mecánicas

---

 

1. INTRODUCTION

In order to improve the coffee fruit (Coffea arabica) processing, different physical [1-3] and mechanical properties [3-6] have been investigated. Some authors [7-11] have measured fruit color using different color models as well as subjective and quantitative measurement methods, depending on each study: RGB; Royal Horticultural Society Colour Chart; Pantone color chart; normalized reflected light peak; image processing to measure: RGB, HSI, YIQ, YCbCr, Ohta color spaces; and opponent color models. The CIELAB color spacing is a Cartesian coordinate system defined by three rectangular coordinates (L*, a*, b*); L* is the lightness (L*= 100, white; L*= 0, black), a* is from green (-a*) to red (+a*), and b* is from blue (-b*) to yellow (+b*).

However, color measurement using the CIELAb for coffee fruits of the variety Colombia, or reflectance in the visible spectrum have not been reported, and the latter is a conventional method for describing the color of any object. Neither has the color been associated with physical properties such as mass, moisture content, size, or mechanical properties such as detachment force, firmness, fracture force, deformation, or fracture energy, which are measurements obtained by destructive tests applying compressive stress. Studies exist in other fruits where the colorimetry has been related to other properties. For example, [12] non-destructive estimation of pigments with spectral reflectance in apple fruit was obtained; [13] the CIELAB color was correlated with texture and concentration of carotenoid in tomatoes; [14] color in CIELUV and CIELAB spaces were measured and related to pigments in olives; [15] spectral wavelengths were identified to determine skin defects in citrus fruit; and [16] the relation between physical, mechanical, and chemical properties in different stages of the ripeness of banana.

In order to satisfy the demands of coffee drinkers, it is important and interesting to know the relation between different physical and mechanical properties via color of fruit, which are closely associated with grain quality. The aim of this study was to correlate the color coordinate of CIELAB and reflectance (in the visible spectrum, 400 to 700 nm) of the epidermis of coffee fruits of the Colombia variety to different physical and mechanical properties for 9 stages of ripeness.

 

2. MATERIALS AND METHODS

This research was carriedout at the Experimental station El Rosario of Centro Nacional de Investigaciones de Café, CENICAFE, located in the municipality of Venecia, Antioquia-Colombia(05° 58' north latitude, 75° 43'west longitude,and 1,630 maltitude, average temperature of 20.1°C) and in the laboratories of the Agricultural Process and Quality Controlat the Universidad Nacional de Colombia, Sede Medellín.

A plot of Coffea arabica L. Colombia variety under full sun exposure was used, with adistance of 1 m x 1 m, planted in May of 1999 and renewed by stumpingin February 2006. One hundred two trees were randomly selected and the main flowering, which was held on April 18 of 2009, was marked. With the scale for measuring fruit ripeness used by [3], there were 9 different stages of fruit development defined as 182, 189, 196, 203, 210, 217, 224, 231, and 238 days after blossom DAA, Table 1) the physical and mechanical properties were measured for each stage of development as shown in Table 2, using the methodology and procedures evaluated by [4,17-19].

Table 1. Stages of development of coffee fruit Colombia variety

Table 2. Properties evaluated for each stage of development of coffee fruit Colombia variety

Figure 1. Diameters and symmetry planes of the coffee cherryfruit

With the average values for each one of the properties of coffee fruits cherries evaluated in 9 stages of fruit development, the correlation of the coordinates L*, a*, b*, and reflectance with mechanical and physical properties were performed, using multivariate linear model analysis with the Statgraphics®statistical software version 5.1.

Characteristic diameter was also correlated with physical and mechanical properties. Correlations were performed at a confidence level of 99%. For each equation, the significance level P of the model, the standard error of estimation,and the coefficients of determination were obtained.

 

3. RESULTS AND DISCUSSION

3.1. Correlation of physical and mechanical properties with the chromaticity coordinates color CIELAB.
Table 3 presents results for each of the regression models of the form Y = C0 ± C1·L* ± C2·a* ± C3·b*predicting properties of coffee fruit from the knowledge of chromaticity coordinates a* and b*.

Table 3. Correlation between physical and mechanical properties with CIELAB coordinates of coffee fruits, Colombia variety

Luminosity L* showed no significant effect in predicting different properties. Properties that showed highest fit (R2) of the linear model with the chromatic variables are shown as well. For all mechanical and physical properties studied, except for deformation with plates in plane E1, polar diameter (P) and moisture content of fruits (Mc), a significant linear relationship with the chromaticity coordinates a* or b*, or with both, was found. According to Table 3, some physical and mechanical properties (Y) can be explained with the monochrome knowledge a* or b* and other variables with a*and b* bichrome coordinates. For example, fruit size, expressed through the equatorial diameter (E1, E2) and the characteristic diameter (Dc) can be estimated knowing the chroma a* indicating whether the fruit is ripe (red, a* > 0) or unripe (green, a*< 0). Fracture force and fruit pulp firmness properties measured with sharp tip in the equatorial plane E1 are those that can be predicted with better explanation (R2 > 99%) from a* and b* color of the skin of coffee fruit between 182 and 238 days of development. For the rest of properties, the explanation for their color variation ranged between 78.5 and 98.6%. Fruit size (E2 and E1) can be predicted from a* with a correlation of 95.34 and 89.43%, respectively. The fresh and dry mass of coffee fruits can be predicted from a* and b* and with a correlation of 92.1 and 78.5%, respectively. The fracture force in the equatorial plane E2 can be estimated from the chroma a*, with a correlation of 97.9%, and the tensile force required to detach the fruit can be predicted from chroma a* and b*, with a correlation of 88.9%.

3.2. Physical and mechanical properties correlation with reflectance (%)
Multiple linear regression models of physical and mechanical properties with reflectance were evaluated for wavelengths (l) of 550, 590, 640, 670 and 700 nm in which the highest discrimination of ripening stages of coffee was obtained. Results for each model of the form Y = C0 ± C1·l550 + C2·l590 ± C3·l640 ± C4·l670 ± C5·l700 are shown in Table 4.

Table 4. Physical and mechanical properties correlation with reflectance for discriminating wavelengths (l)of coffee fruits, Colombia variety

A statistically significant linear correlation with different wavelengths (550 ≤ l ≤ 700 nm) was found for all mechanical and physical properties studied, except for the polar diameter (P), moisture content, and deformation with parallel plates in E1. Firmness and fracture force properties of fruit measured with a sharp tip on the equatorial plane E1, equatorial diameter (E1 and E2), deformation with sharp tip E1, fracture force with sharp tip in E1, fracture force with sharp tip in E2, and firmness between parallel plates in the plane E2, can be predicted with a major explanation (R2 > 99%) from the reflected light at different wavelengths for fruits from 182 to 238 days of development. For the remaining properties, the explanation of their variability with regard to reflectance ranged between 92.33 and 98.99%. The equatorial diameter E2 was the property that presented a higher correlation with the reflectance of the fruit. This physical property associated with the fruit size is significantly correlated to light reflected in the green-yellow, yellow-orange, orange-red, and red regions of the visible spectrum. Firmness using sharp tip in the equatorial plane E1, mechanical property indicator of fruit resistance, is also significantly correlated to light reflected in the green-yellow, orange-red, and red regions of the visible spectrum. The greater the reflected light from the fruit surface in the first regions of the spectrum, reflecting more in the red (700 nm), the lower the firmness will present itself, and therefore the more ripe the fruit will be.

Similar analysis can be performed for the model predicting the detachment force and fracture force according to the light reflected from the fruit surface in different regions of the visible spectrum.

3.3. Physical and mechanical properties correlation with the characteristic diameter (Dc)
Table 5 shows linear models of the form Y = Intercept ± Slope*Dc, that relate different physical and mechanical variables to the characteristic diameter (Dc) of coffee fruitsduring their stage of ripeness. It wasfound that Dc is not correlated with: the moisture content, the fracture energy with sharp tip in the E2 plane, the deformation with plates in E1, or the detaching force (Ft).The other properties were significantly correlated (P < 0.01) to Dc at a confidence level of 99% and with the R2 ranging between 78.2 and 99.6%. It was found that fresh mass (M) is the most closely related variable to Dc with a correlation coefficient of 99.57%. According to Table 5, the deformation between parallel plates in the three compressive planes decreases during ripening. The opposite case was found for fruit deformation with a sharp tip, which increases with the ripeness in both different compressive planes.

Table 5. Physical and mechanical properties correlation to the characteristic diameter (Dc) of coffee fruit, Colombia variety

Models indicate that firmness, independent of the attachment used (sharp tip or parallel plates), is inversely proportional to Dc. This corroborates with what was expected, the fruit, having developed, increases in size and becomes softer up to the overripe stage, when it is close to drying off in the tree, the Dc diminishes, and the firmness is less. A similar trend was found by [20] in pears: as the fruit grows its firmness diminishes. The other mechanical properties (fracture force and energy)also diminish when the fruit size is increased. The Dc variable is not correlated to moisture content or the fruitdetachment force. For the fracture energy variable with a sharp tip on the E2 plane, fracture deformation with parallel plates on the E1 plane, and deformation with a sharp tip on the E2 plane, the R2 changed between 58.9 and 68.9%. The rest of the variables were correlated to the diameter Dc at a confidence level of 99% (P < 0.01). Firmness and fresh mass evaluated at [3] did not demonstrate correlation with Dc. Those differences with this work could be explained by the methodology used in both studies.

 

4. CONCLUSIONS

Different linear models with a confidence level of 99% and a correlation that changed between 78.5 and 99.9%, allow for one to predict different physical and mechanical properties of coffee fruits Colombia variety, with stages of development between 26 and 34 weeks from the knowledge of the chromatic coordinates (a* and b*) or from skin reflectance at different wavelengths.

Mechanical and physical properties studied (except for the deformation with plates in the plane E1, the polar diameter and moisture content of fruits)showed a significant correlation with chromaticity coordinates a*and b*.

Physical and mechanical properties of the coffee fruit are significantly correlated with reflectance, except for the polar diameter, dry mass, moisture content, and the deformation by fruit compression between plates in the equatorial plane.

Physical and mechanical properties of coffee fruit are significantly correlated to their characteristic diameter, except for moisture content, the fracture energy with sharp tip on the E2 plane, the deformation with plates at the E1, and the detachment force of the fruit.

Correlations found in this research could be used for the design of collecting and sorting systems and quick and accurate determination of the quality of coffee fruit Colombia variety.

 

ACKNOWLEDGEMENTS

We thank the Research Division of the Universidad Nacional de Colombia, Medellín (DIME) for research funding and the Centro Nacional de Investigaciones del Café (CENICAFE) for their technical support.

 

REFERENCES

[1] Salazar, M. R., Chaves, B., Riaño, N., Arcila, J., y Jaramillo, A., Crecimiento del fruto de Coffea arabica var. Colombia. Cenicafé, 5(2), pp. 41-50, 1994.         [ Links ]
[2] Roa, G., Oliveros, C.E., Álvarez, J., Ramírez, C.A., Sanz, J.R., Álvarez, J.R., Dávila, M.T., Zambrano, D.A., Puerta, G.I. y Rodríguez, N., Beneficio ecológico del café. Cenicafé. Federación Nacional de Cafeteros de Colombia. Chinchiná (Colombia), 1999.         [ Links ]
[3] Marín, S.M., Arcila, J., Montoya, E.C. y Oliveros, C.E., Cambios físicos y químicos durante la maduración del fruto de café (Coffe arabica L. var. Colombia). Cenicafé. 54(3), pp. 208-225, 2003.         [ Links ]
[4] Álvarez, E., Álvarez, F., Oliveros, C.E. y Montoya, E.C., Propiedades físico-mecánicas del fruto y del sistema fruto-pedúnculo del café variedad Colombia. Rev. Fac. Nal. Agr. Medellín, 52(2), pp. 701-720, 1999.         [ Links ]
[5] Couto, S.M., Batista, C., Devilla, I.A., E Paim, V.T., Características de frutos de café sob compressao. Revista Brasileira de Engenharia Agrícola e Ambiental, 6(1), pp. 117-122, 2002.         [ Links ]
[6] Oliveros, C.E., Montoya, E.C. y Ayala, A., Efecto de la broca del café en la firmeza del grano en los estados de cereza, pergamino húmedo y pergamino seco. Cenicafé. 53(1), pp. 25-33, 2002.         [ Links ]
[7] Montes, N.L., Osorio, G., Prieto, F., y Angulo, F., La visión artificial aplicada al proceso de producción del café. Dyna Rev. Fac. Minas, No.133, pp. 41-49, 2001.         [ Links ]
[8] Ramos, P., Valdés, C., Gómez, E., Sanz, J.R. y Solarte, E., Características espectrales de la luz reflejada por frutos de café (Coffea arabica). Revista Colombiana de Física, 38(2), pp. 822-825, 2006.         [ Links ]
[9] Sandoval, Z.L. y Prieto, F.A., Caracterización de café cereza empleando técnicas de visión artificial. Rev. Fac. Nal. Agr. Medellín, 60 (2), pp. 4105-4127, 2007.         [ Links ]
[10] Cárdenas, S.I., Caracterización morfológica y agronómica de la colección núcleo de café (Coffea arabica L.) del CATIE. [Tesis de Magister Scientiae]. Turrialba, Costa Rica: CATIE, 2007.         [ Links ]
[11] Sanz, J.R., Ramos, P.J. and Oliveros, C.E., Optoelectronic Analyzer of Coffee Fruits. In: Proceedings Book, World Congress on Engineering and Computer Science 2008. San Francisco, CA, USA, pp. 631-636, 2008.         [ Links ]
[12] Merzlyak, M.N., Solovchenko, A. E. and Gitelson, A.A., Reflectance spectral features and non-destructive estimation of chlorophyll, carotenoid and anthocyanin content in apple fruit. Postharvest Biology and Technology, 27, pp. 197-211, 2003.         [ Links ]
[13] Zapata, L., Gerard, L., Davies, C., Oliva, L. y Schvab, M., Correlación matemática de índices de color del tomate con parámetros texturales y concentración de carotenoides. Ciencia, Docencia y Tecnología, 34, Año XVII, pp. 207-226, 2007.         [ Links ]
[14] Moyano, M. J., Meléndez, A. J., Alba, J. and Heredia, F.J., A comprehensive study on the colour of virgin olive oils and its relationship with their chlorophylls and carotenoids indexes (II): CIELUV and CIELAB uniform colour spaces. Food Research International, 41, pp. 513-521, 2008.         [ Links ]
[15] Balasundaram, D., Burks, T. F., Bulanon, D. M., Schubert, T. and Lee, W. S., Spectral reflectance characteristics of citrus canker and other peel conditions of grapefruit. Postharvest Biology and Technology, 51, pp. 220-226, 2009.         [ Links ]
[16] Soltani, M., Alimardani, R., and Omid, M., Comparison of Some Chromatic, Mechanical and Chemical Properties of Banana Fruit at Different Stages of Ripeness. Modern Applied Science, 4 (7), pp. 34-41, 2010.         [ Links ]
[17] Carvajal, J.J., Evaluación cuantitativa del color del fruto de café (Coffea arabica L. var. Colombia) y su relación con algunas propiedades físico-mecánicas durante el proceso de maduración. [Trabajo de Grado Ingeniero Agrícola]. Medellín, Colombia: Universidad Nacional de Colombia. Facultad de Ciencias Agropecuarias, 2010.         [ Links ]
[18] Ciro, H. J., Estudio dinámico de la rama de café para el desarrollo de la cosecha mecánica por vibración. [Trabajo de Grado Ingeniero Agrícola]. Medellín, Colombia: Universidad Nacional de Colombia. Facultad de Ciencias Agropecuarias, 1997.         [ Links ]
[19] AMERICAN SOCIETY OF AGRICULTURAL AND BIOLOGICAL ENGINEERING (ASABE). Standards engineering practices data. 53rd edition. ASABE, USA. 2006.         [ Links ]
[20] Parra, A., Hernández, J.E. y Camacho, J.H., Estudio de algunas propiedades físicas y fisiológicas precosecha de la pera variedad triunfo de viena. Rev. Bras. Frutic., Jaboticabal, 28(1), pp. 55-59, 2006.         [ Links ]