ABSTRACT

Flour blends of African breadfruit (Treculla africana), corn (Zea mays) and partially defatted soybean (G'Iycine max) were extruded in a single screw Brabender Laboratory extruder (Model DCE 330, New Jersey) at varying feed composition (fc), feed moisture, (fm) and screw speeds (ss). A central composite rotatable response surface design was employed to study the effects of independent process variables on product responses at five levels. The five levels embrace the following factors: fc = 40, 55, 70, 85 and 100% (breadfruit); fm = 15, 18, 21, 24 and 27%; and ss = 100, 120, 140, 160 and 180 rpm. Based on the design, a total of 25 runs were generated. Response surface data on each run were statistically regressed, using the stat graphic computer software. Three dimensional response surface contour figures were plotted. Results show that optimization of extrusion process involving these three raw materials was achieved. Extradite acceptability was achieved at 25% defatted soy substitution for breadfruit (fc 70% breadfruit), feed moisture of 18% and screw speed of 140rpm. Amino acid retention of 95.01, 89.79, 88.57% for fc values of 100, 70 and 40% (breadfruit), respectively showed adequacy of process conditions. The polynomial equations developed from the models showed that all the process variables had significant (p ~ 0.05) linear and quadratic effects on bulk density, but showed no such significant effects on water solubility index, residence time distribution, and overall acceptability. Individually, feed composition showed significant (p ~ 0.05) linear and quadratic effects also on extrudate temperature, specific mechanical energy torque, expansion ratio, crude protein and trypsin inhibitor activity (TIA). They showed significant cross-product (fc* fm) effects on carbohydrate and crude protein and cross-product (fc ss) effects on bulk density and appearance. Feed moisture had significant (p ~ 0.05) linear and quadratic effects also on, residence time, bulk density, nitrogen solubility index, crude protein and moisture content. On the other hand, screw speed showed significant (p ~ 0.05) linear and quadratic effects on water absorption index, throughput, nitrogen solubility index, TIA, moisture and energy content. Values of 32.40 - 35.50% were obtained for crude protein corresponding to optimum process variables of 40, 18 - 24% and 140rpm for fc, fm and ss, respectively. Minimum residence time of 20s was obtained at process variable values of fc = 55°h, fm = 27% and ss = 180rpm. Minimum (1.08) and maximum (2.16) values of TIA were observed with process variables (fc:fm:ss) of 100: 18: 120 and 40:18:160rpm, respectively.