**7. Experimental design to process optimization**

Response surface methodology (RSM) is a methodology used to improve process via very few essays, reducing cost and time. The RSM uses statistical and experimental design tools to obtain an optimal response, which is useful for making the right decision. The process performance is very complex due to numerous parameters that affect their behavior. RSM allows built process behavior maps based on mathematical models containing the significant parameters to achieve the maximum, target or minimum process performance.

The optimization of complex processes locates the best experimental conditions at which the process presents the minimum or maximum performance (yield, efficiency, etc.). The use of experimental design for optimizing processes has several advantages: less treatment time, low cost, and efficient use of resources, such as materials, equipment, and workforce. Besides, it uses tools of numerical regression to fit the data to mathematical models to predict values on the region of studied factor levels.

## **7.1 Use of experimental design on coffee waste**

The experimental design has been used to optimize the extraction conditions of coffee parchment waste (CP) [86], antioxidant phenolic compounds from coffee silverskin (CS) [8], total phenolic compound and caffeine from SCG [86], coffee oil from SCG [87], the removal conditions of free fatty acid of SCG [88], the conditions to reducing sugar from SCG [89], organic acids [90] and alcohol production from coffee waste [91], and the conditions for the quantification of heavy metals (Cd(II) and Pb(II)), where a carbon-paste electrode modified with SCG was used as a working electrode [92].

## *7.1.1 Type of experimental designs*

The experimental design tools most used are the central composite design, the Box-Behnken design, and the Plackett-Burman design.

Box-Behnken experimental design was used by Mirón-Mérida et al. [86] to maximize the extract yield, total phenolic content, antioxidant activity, and caffeine content on CP simultaneously. The effects of three parameters on the responses were studied: liquid/solid ratio (10, 30, and 50), extraction temperature (45, 60, and 75°C), and ethanol percentage (50, 75, and 100%). The maximum extract yield of 2.36% was achieved at 75°C with 66.76% ethanol as a solvent and with 50 of liquid/solid ratio. The maximum caffeine extracted was 1.513 g caffeine kg<sup>−</sup><sup>1</sup> CP at 74.35°C and 69.64% ethanol with 33.47 of liquid/solid ratio. The highest total phenolic content of 2.84037 g gallic acid kg<sup>−</sup><sup>1</sup> CP was obtained at 14.33 liquid/solid ratio, 70.74% ethanol, and 75°C. For the maximum extraction of 12.69 μmol Trolox g<sup>−</sup><sup>1</sup> , CP of antioxidant activity was attained at liquid/solid ratio of 50, temperature of 75°C, and ethanol of 59.47%. Finally, the optimal extraction conditions were established at 75°C with 41 liquid/solid ratio using 70% of aqueous ethanol as solvent.

**149**

*Revalorization of Coffee Waste*

*DOI: http://dx.doi.org/10.5772/intechopen.92303*

CS dry matter, during 30 min at 60–65°C [8].

0.8 g sample loading weight at 195°C.

Ballesteros et al. [8] used a 23 face-centered central composite design to maximize the extraction of antioxidant phenolic compounds and oxidant activity from CS. The effects of ethanol concentration (20 and 90%), solvent/solid ratio (10 and 40 ml/g), and extraction time (90 and 30 min) were studied on the two responses. The highest phenolic compounds of 13 mg gallic acid equivalents/g CS, with the maximum antioxidant activity of 18.24 μmol Trolox equivalents/g CS and 0.83 mmol Fe(II)/g CS, were achieved at 60% ethanol as solvent, a ratio of 35 ml/g

Shang et al. [87] developed a two-stage experimental statistical analysis to optimize extraction conditions for total phenolics (mg/g) and caffeine (mg/g) from SCG. First, the process parameter was screened through a Plackett-Burman experiment design to identify the significant parameters of the pressurized liquid extraction method that affect the extraction efficiency, using six parameters at two levels: temperature (80 and 160°C), the concentration of ethanol in water (25 and 75%), extraction time (5 and 20 min), pressure (500 and 2500 psi), sample loading weight (0.5 and 2.5 g), and flush (20 and 100%). The most critical parameters affecting total phenolics and caffeine extraction were temperature and sample loading weight, at 95°C and 0.8 g, respectively. In the second optimization stage, a second-order central composite experimental design, employing the two significant parameters, was used to maximize the total phenolics and caffeine. The highest total phenolic compounds of 22.91 mg/g and caffeine extraction of 9.66 mg/g were achieved with

Pichai and Krit [88] applied response surface methodology to optimize the effects on the coffee oil yield for the solvent extraction process of the ratio of DSCG-hexane (1:8–1:22 g/g) and extraction time (6–34 min). According to the optimal conditions of 1:22.5 g/g mass ratio of DSCG-to-hexane and 30.4 min of extraction time under the 30°C of room temperature, the highest coffee oil yield

Mueanmas et al. [89] used a central composite design to investigate the effect on the FFA removal percentage of the mole ratio (5–15) of MeOH-free fatty acid (FFA), the quantity of catalyst (5–15 wt%), the reaction temperature (50–70°C), and the reaction time (30–120 min). The maximum predicted (95.06%) and experimental (93.88%) of FFA removal was attained at 9.1:1 mol ratio of MeOH/FFA with 11.7

Ravindran et al. [90] proposed a central composite design to maximize the reducing sugar yield of SCG, after enzymatic saccharification of pretreated biomass and ultrasound-assisted potassium permanganate oxidation. The effects of five parameters on the responses were studied: 77.08 FPU/mL of cellulase (biomass loading 1–5 g/50 ml), 72.23 U/mL of hemicellulase (biomass loading 0.3–1.5 ml/50 ml), pH (4.8–6.6), and incubation time (24–120 h). A maximum reducing sugar yield of 35.64 mg/mL of reaction volume was estimated with a high biomass loading of 5 g/50 mL, 1.5 mL/50 mL of cellulase, 0.37 mL/50 mL of hemicellulase, pH 6.7, and a low incubation time of 24 h. The experimental values obtained using the optimized

parameters are in the range of total reducing sugar of 35.15 ± 0.2 mg/mL.

Montoya et al. [91] developed a Plackett-Burman design to evaluate the effect of the parameters on H2, organic acids, and alcohol production from coffee waste. The coffee waste was pretreated using a consortium of bacteria and fungi (indigenous from coffee waste) with hydrolytic and fermentation activity in a hydrothermal reactor. The parameters of pH (4.0–7.0), temperature (30–50°C), agitation (0–180 rpm), headspace (50–70%), percentage of bioaugmentation (without microbial consortium to 20%), the concentration of coffee pulp and husk (2–6 g/L), coffee processing wastewater (7–30 g COD/L), and yeast extract (0–2 g/L) were studied. Under the optimum conditions of 30°C, 180 rpm, 50% headspace, without

estimated (14.75 wt%) and experimental (14.68 wt%) was reached.

wt% of catalyst and 97.2 min of reaction time at 65°C.

#### *Revalorization of Coffee Waste DOI: http://dx.doi.org/10.5772/intechopen.92303*

*Coffee - Production and Research*

footprint (0.96 kg CO2-e/kg of BSG).

**7. Experimental design to process optimization**

maximum, target or minimum process performance.

**7.1 Use of experimental design on coffee waste**

Box-Behnken design, and the Plackett-Burman design.

highest total phenolic content of 2.84037 g gallic acid kg<sup>−</sup><sup>1</sup>

as a working electrode [92].

*7.1.1 Type of experimental designs*

compost, adsorbent, bioactive compounds, and pharmaceutical products [72]. The production of xylitol, activated carbon, phenolic acid, lactic acid, and heat using brewer's spent grain as the feedstock of a biorefinery [60, 85]. The process sustainability of biorefinery was demonstrated, thanks to the economic margin (62.25%), the potential environmental impact (0.012 PEI/kg products), and the carbon

Response surface methodology (RSM) is a methodology used to improve process via very few essays, reducing cost and time. The RSM uses statistical and experimental design tools to obtain an optimal response, which is useful for making the right decision. The process performance is very complex due to numerous parameters that affect their behavior. RSM allows built process behavior maps based on mathematical models containing the significant parameters to achieve the

The optimization of complex processes locates the best experimental conditions at which the process presents the minimum or maximum performance (yield, efficiency, etc.). The use of experimental design for optimizing processes has several advantages: less treatment time, low cost, and efficient use of resources, such as materials, equipment, and workforce. Besides, it uses tools of numerical regression to fit the data to mathematical models to predict values on the region of studied factor levels.

The experimental design has been used to optimize the extraction conditions of coffee parchment waste (CP) [86], antioxidant phenolic compounds from coffee silverskin (CS) [8], total phenolic compound and caffeine from SCG [86], coffee oil from SCG [87], the removal conditions of free fatty acid of SCG [88], the conditions to reducing sugar from SCG [89], organic acids [90] and alcohol production from coffee waste [91], and the conditions for the quantification of heavy metals (Cd(II) and Pb(II)), where a carbon-paste electrode modified with SCG was used

The experimental design tools most used are the central composite design, the

CP at 74.35°C and 69.64% ethanol with 33.47 of liquid/solid ratio. The

, CP of antioxidant activity was attained at liquid/solid ratio

14.33 liquid/solid ratio, 70.74% ethanol, and 75°C. For the maximum extraction of

of 50, temperature of 75°C, and ethanol of 59.47%. Finally, the optimal extraction conditions were established at 75°C with 41 liquid/solid ratio using 70% of aqueous

CP was obtained at

Box-Behnken experimental design was used by Mirón-Mérida et al. [86] to maximize the extract yield, total phenolic content, antioxidant activity, and caffeine content on CP simultaneously. The effects of three parameters on the responses were studied: liquid/solid ratio (10, 30, and 50), extraction temperature (45, 60, and 75°C), and ethanol percentage (50, 75, and 100%). The maximum extract yield of 2.36% was achieved at 75°C with 66.76% ethanol as a solvent and with 50 of liquid/solid ratio. The maximum caffeine extracted was 1.513 g caf-

**148**

feine kg<sup>−</sup><sup>1</sup>

12.69 μmol Trolox g<sup>−</sup><sup>1</sup>

ethanol as solvent.

Ballesteros et al. [8] used a 23 face-centered central composite design to maximize the extraction of antioxidant phenolic compounds and oxidant activity from CS. The effects of ethanol concentration (20 and 90%), solvent/solid ratio (10 and 40 ml/g), and extraction time (90 and 30 min) were studied on the two responses. The highest phenolic compounds of 13 mg gallic acid equivalents/g CS, with the maximum antioxidant activity of 18.24 μmol Trolox equivalents/g CS and 0.83 mmol Fe(II)/g CS, were achieved at 60% ethanol as solvent, a ratio of 35 ml/g CS dry matter, during 30 min at 60–65°C [8].

Shang et al. [87] developed a two-stage experimental statistical analysis to optimize extraction conditions for total phenolics (mg/g) and caffeine (mg/g) from SCG. First, the process parameter was screened through a Plackett-Burman experiment design to identify the significant parameters of the pressurized liquid extraction method that affect the extraction efficiency, using six parameters at two levels: temperature (80 and 160°C), the concentration of ethanol in water (25 and 75%), extraction time (5 and 20 min), pressure (500 and 2500 psi), sample loading weight (0.5 and 2.5 g), and flush (20 and 100%). The most critical parameters affecting total phenolics and caffeine extraction were temperature and sample loading weight, at 95°C and 0.8 g, respectively. In the second optimization stage, a second-order central composite experimental design, employing the two significant parameters, was used to maximize the total phenolics and caffeine. The highest total phenolic compounds of 22.91 mg/g and caffeine extraction of 9.66 mg/g were achieved with 0.8 g sample loading weight at 195°C.

Pichai and Krit [88] applied response surface methodology to optimize the effects on the coffee oil yield for the solvent extraction process of the ratio of DSCG-hexane (1:8–1:22 g/g) and extraction time (6–34 min). According to the optimal conditions of 1:22.5 g/g mass ratio of DSCG-to-hexane and 30.4 min of extraction time under the 30°C of room temperature, the highest coffee oil yield estimated (14.75 wt%) and experimental (14.68 wt%) was reached.

Mueanmas et al. [89] used a central composite design to investigate the effect on the FFA removal percentage of the mole ratio (5–15) of MeOH-free fatty acid (FFA), the quantity of catalyst (5–15 wt%), the reaction temperature (50–70°C), and the reaction time (30–120 min). The maximum predicted (95.06%) and experimental (93.88%) of FFA removal was attained at 9.1:1 mol ratio of MeOH/FFA with 11.7 wt% of catalyst and 97.2 min of reaction time at 65°C.

Ravindran et al. [90] proposed a central composite design to maximize the reducing sugar yield of SCG, after enzymatic saccharification of pretreated biomass and ultrasound-assisted potassium permanganate oxidation. The effects of five parameters on the responses were studied: 77.08 FPU/mL of cellulase (biomass loading 1–5 g/50 ml), 72.23 U/mL of hemicellulase (biomass loading 0.3–1.5 ml/50 ml), pH (4.8–6.6), and incubation time (24–120 h). A maximum reducing sugar yield of 35.64 mg/mL of reaction volume was estimated with a high biomass loading of 5 g/50 mL, 1.5 mL/50 mL of cellulase, 0.37 mL/50 mL of hemicellulase, pH 6.7, and a low incubation time of 24 h. The experimental values obtained using the optimized parameters are in the range of total reducing sugar of 35.15 ± 0.2 mg/mL.

Montoya et al. [91] developed a Plackett-Burman design to evaluate the effect of the parameters on H2, organic acids, and alcohol production from coffee waste. The coffee waste was pretreated using a consortium of bacteria and fungi (indigenous from coffee waste) with hydrolytic and fermentation activity in a hydrothermal reactor. The parameters of pH (4.0–7.0), temperature (30–50°C), agitation (0–180 rpm), headspace (50–70%), percentage of bioaugmentation (without microbial consortium to 20%), the concentration of coffee pulp and husk (2–6 g/L), coffee processing wastewater (7–30 g COD/L), and yeast extract (0–2 g/L) were studied. Under the optimum conditions of 30°C, 180 rpm, 50% headspace, without

bioaugmentation, 2 g/L pulp and husk coffee, 30 gCOD/L coffee processing wastewater, and 2 g/L yeast extract, estimated production of 82 ml H2 was achieved.

Finally, Estrada-Aldrete et al. [92] applied a central composite design to optimize the quantification of two heavy metals (Cd(II) and Pb(II)) at trace levels using a paste carbon electrode of spent coffee grounds, which was chemically modified by citric acid. The metal quantification was carried out by differential pulse anodic stripping voltammetry technique. The electrodeposition potential (−1200, −950, and −700 mV) and accumulation time (30, 75, and 120 s) were employed as design parameters. The optimal conditions to achieve the maximum Pb(II) anodic peak current of 2.09 × 10<sup>−</sup><sup>4</sup> A were − 1200 mV electrodeposition potential and 120 s accumulation time. The maximum Cd(II) anodic peak current of 1.385 × 10<sup>−</sup><sup>3</sup> A obtained at −1155 mV potential and 76 s time.
