Scholarlink

Purpose <p>The kinetics of the Agtron index (a standardised quantification of roast degree, R<sub>d</sub>) in Arabica and Robusta coffee, and their relationship to physical and spectral parameters, have not been previously examined. This study aimed to: (1) examine the kinetics of characteristic changes in two coffee species (Arabica and Robusta) during roasting, focusing on modelling the R<sub>d</sub> in the form of Agtron index value kinetics; (2) investigate the correlation between the Agtron index and coffee physical parameters; and (3) explore its relationship with spectral characteristics across the visible to short-wave near-infrared (VIS–SWNIR) range (410–940&#xa0;nm).</p> Methods <p>Coffee beans (500&#xa0;g per batch) were roasted using a drum roaster at different duration to produce different R<sub>d</sub> values within a time range of 3–11&#xa0;min with the same initial temperature (190&#xa0;°C), followed by analysis of the physical, chemical, and spectral properties. Data analysis included kinetic modelling for the Agtron index parameter, correlation analysis, and hierarchical clustering to assess interrelationships.</p> Results <p>The Agtron index decreased consistently with roasting time (stage 2) for both coffee species, following both zero- and first-order (type II) kinetics with high R<sup>2</sup> and low RMSE values. The combined data show high positive correlations between Agtron and L*, b*, and C* (R between 0.91 and 0.95), indicating that these physical parameters can serve as reliable alternatives to the Agtron index. Furthermore, the Agtron values exhibit very high correlations (R = 0.99) with the spectral data at 810&#xa0;nm, 860&#xa0;nm, and 900&#xa0;nm.</p> Conclusions <p>These findings provide a robust foundation for the development of an alternative roasting level measurement instrument (Agtron) based on the VIS–SWNIR approach, particularly using readily available low-cost multispectral sensor modules.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Kinetic Modelling of Agtron-Based Roast Degrees and Their Association with Physical and Spectral Changes in Arabica and Robusta Coffee

  • 期刊论文

  • 发表日期：

    2026年1月5日

查看全文
  • Diang Sagita,
  • Sutrisno Suro Mardjan,
  • Slamet Widodo,
  • Suparlan Suparlan,
  • Pradeka Brilyan Purwandoko,
  • Ari Rahayuningtyas,
  • Hari Hariadi,
  • Sandi Darniadi,
  • Lista Eka Yulianti,
  • Dadang Dayat Hidayat,
  • Doddy Andy Darmajana

摘要

Purpose

The kinetics of the Agtron index (a standardised quantification of roast degree, Rd) in Arabica and Robusta coffee, and their relationship to physical and spectral parameters, have not been previously examined. This study aimed to: (1) examine the kinetics of characteristic changes in two coffee species (Arabica and Robusta) during roasting, focusing on modelling the Rd in the form of Agtron index value kinetics; (2) investigate the correlation between the Agtron index and coffee physical parameters; and (3) explore its relationship with spectral characteristics across the visible to short-wave near-infrared (VIS–SWNIR) range (410–940 nm).

Methods

Coffee beans (500 g per batch) were roasted using a drum roaster at different duration to produce different Rd values within a time range of 3–11 min with the same initial temperature (190 °C), followed by analysis of the physical, chemical, and spectral properties. Data analysis included kinetic modelling for the Agtron index parameter, correlation analysis, and hierarchical clustering to assess interrelationships.

Results

The Agtron index decreased consistently with roasting time (stage 2) for both coffee species, following both zero- and first-order (type II) kinetics with high R2 and low RMSE values. The combined data show high positive correlations between Agtron and L*, b*, and C* (R between 0.91 and 0.95), indicating that these physical parameters can serve as reliable alternatives to the Agtron index. Furthermore, the Agtron values exhibit very high correlations (R = 0.99) with the spectral data at 810 nm, 860 nm, and 900 nm.

Conclusions

These findings provide a robust foundation for the development of an alternative roasting level measurement instrument (Agtron) based on the VIS–SWNIR approach, particularly using readily available low-cost multispectral sensor modules.