| Abstract | In this study physical changes occurring during coffee drying and roasting were investigated in two stages. The first stage focused on drying of parchment coffee in a re-circulating rotary conduction-type heating unit aided by a screw conveyor and feeder. The second stage dealt with the roasting of coffee beans in a separate small batch rotary unit. In both heating units, the temperature of moving beans or heating surface could be maintained at specified levels with small variations during coffee drying and roasting experiments. Experimental variables in coffee drying tests included exposure time, bean temperature, heating surface temperature, batch load, and rotational speed ranging from 0 to 420 minutes, 50 to 70°C, 100 to 140°C, 15 to 40 kg, and 5 to 20 RPM, respectively. The coffee beans were dried from an initial moisture content of about 90% to 13% dry basis approximately, and associated changes in breakage susceptibility, specific gravity and surface color were evaluated. Results of drying experiments indicated exponential relationships between moisture content and time which could be represented by simple and modified logarithmic drying models. The drying constants in the fitted models were found to be sensitive to heating surface temperature and product loading without being greatly affected by the rotation speed. However, another drying model, which was developed based on product temperature-time history under different operating conditions, estimated the changes in moisture content of beans more accurately than either of the two logarithmic models. The color and specific gravity of coffee beans exhibited minimum changes as a result of drying operations. The susceptibility of coffee beans breakage decreased with the lowering of moisture and attained minimum values at moisture contents ranging from 20 to 30%. However, the breakage susceptibility increased sharply with a further reduction in moisture content. Roasting of coffee beans was carried out in the batch-type rotary heating unit using about 2 kg of test sample under constant heating surface of product temperature ranging from 200 to 240°C and 170 to 210°C, respectively. Comparable changes occurred in other physical properties such as bulk density, breakage susceptibility and transmittance of coffee brew. The roasting duration was varied from 0 to 70 minutes through intermittent sampling of coffee beans at selected time intervals. There were marked changes in physical conditions of coffee beans during roasting. The specific gravity, color lightness index L, hue angle and pH of brew changed from 1.2 to 0.5, 50 to 12, 75 to 24 and 5 to 6.5, respectively. Comparable changes occurred in other physical properties such as bulk density, breakage susceptibility and transmittance of coffee brew. It was possible to relate the changes in color indices, specific gravity, bulk density, susceptibility to breakage and light transmittance through brew prepared from roasted coffee beans with product time-temperature history and develop predictive models with R2 >0.94. susceptibility of coffee beans to breakage in increased throughout the roasting period. Although light transmittance in brew depended upon the wavelengths, its values increased at the start of roasting and then declined for all wavelengths in the range of 400 to 900 nm. Roasting of coffee samples which had been dried by condition heating at different temperatures and under sun, did not reveal any significant differences in terms of physical changes taking place in roasted coffee beans. The predictive models accounting for physical changes during roasting of coffee could lead to the development of computer based control systems to produce product of consistent quality according to consumer preference. Also it should be possible to automate the roasting process because the process variables are not equipment-specific and can be monitored on-line. |