Microwave‐assisted extraction of bioactive compounds from Sarawak Liberica sp. coffee pulp: Statistical optimization and comparison with conventional methods

Abstract Coffea liberica, commonly known as Liberica coffee, is a kind of coffee that originated in Liberia, a West African country. It is considered a less‐known coffee bean variety, which accounts for less than 2% of commercially produced coffee worldwide. In this study, the influences of optimization of microwave‐assisted extraction (MAE) on the total phenolic content (TPC), total flavonoid content (TFC), and total carbohydrate content (TCC) of bioactive compounds extracted from Sarawak Liberica sp. coffee pulp were studied. Response surface methodology was adopted with a face‐centered central composite design to generate 34 responses by taking three microwave parameters into consideration, microwave power (watt), time of irradiation (second), and solvent‐to‐feed ratio as independent variables. As a result, the findings revealed that optimum extraction conditions were conducted as follows: microwave power of 700 W, time of irradiation of 180 s, and solvent‐to‐feed ratio of 86.644:1. While under optimal extraction conditions, MAE outperformed conventional maceration extraction in terms of extraction efficiency and had no significant difference (p < .05) with Soxhlet extraction on the extraction of TPC (12.94 ± 2.25 mg GAE/g), TFC (9.84 ± 0.38 mg QE/g), and TCC (876.50 ± 64.15 mg GE/g). Present work advances the usage of Sarawak Liberica sp. coffee for the development of functional products and aids in reducing environmental pollution by utilization of coffee pulp waste.

Coffee manufacturing generates a variety of by-products, including wasted coffee grounds, coffee wastewater, coffee pulp, coffee husks, coffee silver skin, and coffee parchment (Reichembach & Petkowicz, 2020).Coffee pulp makes up most of the solid waste produced during the wet processing of coffee.Studies on coffee pulp have confirmed coffee pulp to be abundant in health-promoting nutrients such as fiber, proteins, carbohydrates, minerals (especially potassium), tannins, polyphenols, and caffeine (Murthy & Madhava Naidu, 2012).In addition, Sarawak Liberica sp.coffee pulp has been found to have high levels of phenolic and flavonoid compounds, reducing sugar, and a higher level of radical scavenging activity (Nillian et al., 2020), indicating its potential as a natural source for obtaining bioactive compounds of interest and further developing into functional food.For example, antioxidant-rich food products such as energy bars, cereal bars, or snack bars can be developed from the extracted phenolic and flavonoid compounds.Besides, it can also be used as an ingredient in prebiotic supplements, such as capsules or powders, due to its high level of reducing sugars, which promote the growth of beneficial bacteria in the gut.However, coffee pulp is said to harm the ecosystem by contaminating lakes and rivers close to coffee processing facilities.If marketable by-products are extracted from the coffee pulp in large-scale commercial secondary processing installations using a technology that is beneficial to the environment, this issue might turn into an opportunity (Manasa et al., 2021).
Therefore, it may be better to exploit the coffee pulp as a source for extracting bioactive compounds.
To isolate bioactive ingredients from natural sources, extraction is the initial and most crucial stage.The knowledge of effective bioactive compound extraction from Liberica needs to be improved.It has been discovered that traditional extraction methods, including maceration, infusion, and superficial extraction take a long time, cost a lot to operate, and provide minimal output (Fang et al., 2011).
However, Ince et al. (2012) found that microwave-assisted extraction (MAE) may be used for the substitution of traditional extraction methods and increase bioactive compounds' yield while also speeding up the extraction process and using less solvent.One of the cutting-edge extraction methods, MAE, has been widely used to separate, evaluate, and quantify bioactive components.In MAE, solvents are heated when they encounter solid or liquid samples (or heated samples, such as fresh tissues), facilitating the partition of chemicals relevant to the sample into the solvent.According to published research, several variables, for example, the type of solvent, power, microwave irradiation time, sample-to-solvent ratio, and matrix type, affect MAE yields (Mandal et al., 2007).In light of the fact that coffee pulp may be used to obtain bioactive compounds, it will be useful to determine the experimental conditions for extracting bioactive compounds with the highest yield using MAE.
To achieve maximum extraction efficiency, the influence of MAE process factors and their interactions were assessed using a response surface methodology (RSM).With several experiment runs, this method predicts the most compelling circumstances while allowing simultaneous optimization of all factors.Therefore, this study aims to evaluate the optimization of MAE conditions to obtain the maximum yield of total phenolic, flavonoid, and carbohydrate content from Sarawak Liberica sp.coffee pulp and to compare the effectiveness of MAE with conventional maceration extraction (CME) and Soxhlet extraction (SE).

| Materials
The Sarawak Liberica sp.coffee pulp was used in this study.The coffee pulp samples were provided by RekaJaya Plantation Sdn.Bhd. in Kuching, Sarawak and were sent to the laboratory and refrigerated at 4°C prior to investigation.Unless otherwise mentioned, all chemicals and reagents involved in the study were of analytical grade.

| Sample preparation of Sarawak Liberica sp. coffee pulp fibrous powder
Sarawak Liberica sp.coffee pulp was made into powder form according to the guidelines recommended by Larrauri (1999).Firstly, coffee pulp samples were collected and cleaned with distilled water to get rid of the dirt.The weight of the pulp samples was determined.The pulp was then processed using a wet-milling method with a 1 g of pulp to 1 ml of water ratio.This process was performed using a food blender (Sharp blender EM131BK).A cotton cloth filter was used to separate the juice and pomace from the pulp.After that, the pomace was weighed and measured.The pomace was then washed in water for 5 min.Next, cotton cloth filter was used to filter the pomace, and the pomace was dried in an oven dryer (Drying oven FCD-3000) for 18 h at 50°C.After drying, the pomace was ground into powder form using the food blender (Sharp blender EM131BK).Then, the powder was sieved to pass through a standard 150μm sieve to obtain Sarawak Liberica sp.coffee pulp fibrous powder.The powder was kept at −20°C prior to analysis.

| Experimental design
The method used for extraction and optimization was MAE under different parameters and levels as shown in Table 1.For comparison, CME and SE were also conducted.In this study, RSM with face-centered central composite design (CCD) was applied.
CCD was selected to investigate the effects of multiple variables on the extraction of bioactive compounds and to determine the optimal conditions for maximizing the extraction of bioactive compounds from Sarawak Liberica sp.coffee pulp.CCD allows the investigation of both linear and quadratic effects of the input variables on the response and provides insight into the curvature of the response surface, allowing for the identification of optimum conditions for maximum response.Three microwave parameters: microwave power (watt, X 1 ), time of irradiation (seconds, X 2 ), and solvent-to-feed ratio (X 3 ) were examined for their impacts on three response variables: total phenolic content (TPC, mg GAE/g extract, Y 1 ), total flavonoid content (TFC, mg QE/g extract, Y 2 ), and total carbohydrate content (TCC, mg GE/g extract, Y 3 ).According to the minimum and maximum readings of the parameters selected based on literature data and early experimental findings, each parameter was examined at two levels: high (+1) and low (−1) levels.Methanol was used as the solvent for the extraction of bioactive compounds due to its high polarity, which allows it to effectively dissolve and extract a wide range of polar and semi-polar compounds from the sample matrix.
Design-Expert® Software (Version 13.0; Stat-Ease, Inc.) was used to carry out the experimental design.

| Microwave-assisted extraction
Microwave-assisted extraction was performed using a microwave oven (Elba EMO-2072SV) operating at 700 W output power, 2450 MHz frequency, and 20 L capacity.Sarawak Liberica sp.coffee pulp fibrous powder was mixed with methanol (CH 3 OH, 99% v/v) to obtain the ratio as indicated in Table 2.Then, extraction was carried out at the respective settings of microwave power and irradiation time (Table 2).Thereafter, the mixture was centrifuged twice at 5000 rpm, 4°C for 10 min by a centrifuge machine (Hanil Scientific, Inc.; Supra R22).Next, the supernatant was collected and added with 1 mL of trichloroacetic acid (C 2 HCl 3 O 2 , 10% w/v) for the precipitation of protein (Novák & Havlíček, 2016).The precipitate was removed and the supernatant was collected and dialyzed against distilled water for 12 h using a dialysis bag (molecular weight cut-off, MWCO 500 Da) to eliminate the unwanted low-molecular-weight compounds.Liquid extract that had not been dialyzed was collected and kept at 4°C prior to investigation.

| Conventional maceration extraction
Conventional maceration extraction was carried out at room temperature in the following manner: 5 g of Sarawak Liberica sp.coffee pulp fibrous powder was mixed with 350 mL of methanol (CH 3 OH, 99% v/v).Then, the mixture was kept at room temperature overnight on a shaker (Shaker JEIO TECH SK-71).
Thereafter, the mixture was centrifuged twice at 5000 rpm, 4°C for 10 min by a centrifuge machine (Hanil Scientific, Inc.; Supra R22).Next, the supernatant was collected and added with 1 mL of trichloroacetic acid (C 2 HCl 3 O 2 , 10% w/v) for the precipitation of protein.The precipitate was removed and the supernatant was collected and dialyzed against distilled water for 12 h using a dialysis bag (molecular weight cut-off, MWCO 500 Da) to eliminate the unwanted low-molecular-weight compounds.
Liquid extract that had not been dialyzed was collected and kept at 4°C prior to investigation.

| Soxhlet extraction
A measured amount of Sarawak Liberica sp.coffee pulp fibrous powder (5 g) was placed in a cellulose thimble of the Soxhlet extractor.An amount of 350-mL methanol (CH 3 OH, 99% v/v) was added into a round bottom flask of the Soxhlet extractor.Then, the Soxhlet extractor was set up and methanol was heated using the isomantle heater.The extraction time was 10 h and the temperature was set at 65°C.Thereafter, the mixture was centrifuged twice at 5000 rpm, 4°C for 10 min by a centrifuge machine (Hanil Scientific, Inc.; Supra R22).Next, the supernatant was collected and added with 1 mL of trichloroacetic acid (C 2 HCl 3 O 2 , 10% w/v) for the precipitation of protein.The precipitate was removed and the supernatant was collected and dialyzed against distilled water for 12 h using a dialysis bag (molecular weight cut-off, MWCO 500 Da) to eliminate the unwanted low-molecular-weight compounds.Liquid extract that had not been dialyzed was collected and kept at 4°C prior to investigation.

| Total phenolic content
Total phenolic content of the extract obtained from MAE, CME, and SE was investigated following the method developed by Velioglu et al. (1998) with some modifications.For sample preparation, liquid extract was dried in a fume hood (Köttermann GmbH fume hood) and resuspended with methanol (CH 3 OH, 99% v/v) to obtain a final concentration of 1 mg/mL.Briefly, 0.2 mL of the sample was mixed with 0.5 mL of stock Folin-Ciocalteu's reagent (10% w/v).Next, 0.8 mL of bicarbonate (NaHCO 3 , 7.5% w/v) solution was added to the mixture.The mixture was vortexed and allowed to stand for 30 min at room temperature with continual shaking.The absorbance of the mixture was obtained using a Ultraviolet-visible spectrophotometer (Shimadzu UV-1900i) at 765 nm.A calibration curve (absorbance vs. concentration) was constructed using gallic acid solution (0.1-0.5 mg/mL in methanol [CH 3 OH, 99% v/v]) as a standard.The results were expressed as milligram of gallic acid equivalent per gram of dry extract (mg GAE/g extract).
TA B L E 1 Microwave parameters and their levels used in CCD.Abbreviations: TCC, total carbohydrate content; TFC, total flavonoid content; TPC, total phenolic content.a Parameters: X 1 , microwave power (watt); X 2 , time of irradiation (second); X 3 , solvent-to-feed ratio.
for another 5 min.Thereafter, 2.0 mL of sodium hydroxide (NaOH, 1 M) and 4.0 mL of distilled water were added to the mixture, and the test tube was vortexed before the measurement of absorbance at 510 nm using Ultraviolet-visible spectrophotometer (Shimadzu UV-1900i).A calibration curve (absorbance vs. concentration) was made using quercetin (0.2-1.0 mg/mL in 99% [v/v] methanol).The results were presented as milligram of quercetin equivalent per gram of dry extract (mg QE/g extract).

| Total carbohydrate content
Total carbohydrate content of the extract obtained from MAE, CME, and SE was tested by phenol-sulfuric method as described by Dubois et al. (1956).At first, liquid extract was dried in a fume hood (Köttermann GmbH fume hood) and resuspended with methanol (CH 3 OH, 99% v/v) to obtain a final concentration of 1 mg/mL.

| Statistical analysis
Experiments were done in triplicate and data were presented as mean ± standard deviation.SPSS software (version 23; IBM Corporation) was used to carry out the statistical analysis.Analysis of variance (ANOVA)'s Tukey test, Kruskal-Wallis test, and independent t-test were applied to determine the significant difference among the means at 95% confidence interval (p < .05).

| Experimental outcome of design matrix
A total of 34 experimental runs were carried out at random using the levels described in Table 1 and the conditions are shown in Table 2, which included 14 various combinations of experiments conducted in duplicate and six center points, as well as the experimental results of response of the study.

| Model fitting
In the present study, a face-centered CCD was selected to investigate the influences and interactions of the MAE parameters (microwave power, time of irradiation, and solvent-to-feed ratio) on the TPC, .9614and .9114for TCC.The fact that the differences between adjusted R 2 and predicted R 2 of the three models were all less than .2implies that the predicted and actual experimental results were in reasonable agreement with each other.Other than that, the low coefficient of variation values (CV = 7.40%, 14.69%, and 8.99% for TPC, TFC, and TCC, respectively) indicates that the experimental results of the fitted models were reliable and reproducible.It was recommended that the models had a decent simulation of the extraction experiment based on the findings.
Predictions about the tested responses can be made using the following coded equations (Equations 1-3).By contrasting the factor coefficients, the coded equation may be used to determine the relative significance of the factors.A positive effect is indicated by a positive coefficient, and vice versa.
Equation (1) shows that TPC obtained increases as all the parameters (X 1 , X 2 , and X 3 ), interaction effect X 1 X 2 and X 1 X 3 , and quadratic terms X 2 2 increase, and it increases as the quadratic terms X 2 1 and X 2 3 decrease.TPC obtained has a positive relation with microwave power (X 1 ), time of irradiation (X 2 ), solvent-to-feed ratio (X 3 ), interaction effect (microwave power and time of irradiation, X 1 X 2 ), interaction effect (microwave power and solvent-to-feed ratio, X 1 X 3 ), and the quadratic effect X 2 2 , whereas it has a negative relation with the quadratic effects X 2 1 and X 2 3 .There was no significant interactive effect (p = .7030[>.05]) of X 2 X 3 on TPC.This indicates that the TPC of the bioactive compounds from Sarawak Liberica sp.coffee pulp was not affected by the interaction effect between time of irradiation and solvent-to-feed ratio.The factor coefficient of X 1 (2.78) has the highest number among all the parameters.Thus, microwave power had the most significant effect on TPC.
TA B L E 3 Analysis of variance (ANOVA) of quadratic model for TPC, TFC, and TCC.Similar to TPC, Equation (2) shows that TFC obtained increases as all the parameters (X 1 , X 2 , and X 3 ), interaction effect X 1 X 2 and X 1 X 3 , and quadratic terms X 2 2 increase, and it increases as the quadratic terms X 2 1 and X 2 3 decrease.TFC obtained has a positive relation with microwave power (X 1 ), time of irradiation (X 2 ), solventto-feed ratio (X 3 ), interaction effect (microwave power and time of irradiation, X 1 X 2 ), interaction effect (microwave power and solventto-feed ratio, X 1 X 3 ), and the quadratic effect X 2 2 , whereas it has a negative relation with the quadratic effects X 2 1 and X 2 3 .There was no significant interactive effect (p = .7030[>.05]) of X 2 X 3 on TFC.This indicates that the TFC of the bioactive compounds from Sarawak Liberica sp.coffee pulp was not affected by the interaction effect between time of irradiation and solvent-to-feed ratio.The factor coefficient of X 1 (2.58) has the highest number among all the parameters.Therefore, the most significant effect on TFC was caused by microwave power.
As indicated in Equation ( 3), the TCC obtained increases as all the parameters (X 1 , X 2 , and X 3 ) and the interaction effect X 1 X 2 and X 1 X 3 increase, and it increases as the interaction effect X 2 X 3 and the quadratic terms X 2 2 decrease.TCC obtained has a positive relation with microwave power (X 1 ), time of irradiation (X 2 ), solvent-to-feed ratio (X 3 ), interaction effect (microwave power and time of irradiation, X 1 X 2 ), and interaction effect (microwave power and solvent-to-feed ratio, X 1 X 3 ), whereas it has a negative relation with the interaction effect (time of irradiation and solvent-to-feed ratio, X 2 X 3 ) and quadratic effects X 2 2 .There were no significant quadratic effects of X 2 1 (p = .2393[>.05]) and X 2 3 (p = .2586[>.05]) on TCC.This indicates that the TCC of the bioactive compounds from Sarawak Liberica sp.coffee pulp was not affected by the quadratic effects of X 2 1 and X 2 3 .The factor coefficient of X 1 (176.97)has the highest number among all the parameters.Hence, microwave power was the most significant influencing factor on TCC.
The mathematical models in terms of coded factors for the tested responses can be represented as follows:

| Analysis of the significance of MAE parameters on TPC
The effects of MAE parameters and their mutual interactions on the TPC of the bioactive compounds from Sarawak Liberica sp.coffee pulp are presented in Table 3 and Figure 1a-c.
Figure 1a shows the response surface plot of TPC influenced by parameters X 1 (microwave power) and X 2 (time of irradiation) indicated as strongly significant with a p-value of <.0001.The yield of TPC increased with the increase in microwave power from 119 to 534 W and time of irradiation from 60 to 180 s, respectively, then decreased as microwave power increased from 534 to 700 W. Similar results were obtained for MAE of polyphenols from spent espresso coffee grounds (Ranic et al., 2014), from lotus (Nelumbo nucifera Gaertn.)seeds (Zhang et al., 2012), and from avocado (Persea americana Mill.) seeds (Weremfo et al., 2020), in which the TPC increased with the increase in microwave power to a certain level, beyond which a drop in TPC was found.This phenomenon could be explained by the increased solubility of polyphenols at higher microwave power.The temperature of the extraction process is influenced by microwave power; an increase in power level will result in a higher temperature.High temperatures could break down chemical bonds or reduce intermolecular interaction, which increases the solubility of specific compounds and increases the number of extractable compounds in the solution (Xu et al., 2017).
They can also soften plant tissue and increase solvent penetration into the sample, increasing the mass transfer rate (Dahmoune et al., 2013).Additionally, microwaves could break down the sample tissue and cell wall by creating a pressure differential between the interior and exterior of the plant cells or tissues.This results in the release of additional phenolic compounds.However, heat-sensitive compounds, particularly phenolic compounds, can degrade at high temperatures, hence the decrease in the amount of TPC of bioactive compounds extracted from Sarawak Liberica sp.coffee pulp (Ali et al., 2018).As for the time of irradiation, the result was similar to Figure 1b depicts the response surface plot of TPC influenced by parameters X 1 (microwave power) and X 3 (solvent-to-feed ratio) (1) (2) TA B L E 4 Fit statistics.
which is shown as significant (p = .0017).The extraction yield of TPC was likewise enhanced when the solvent-to-feed ratio increased from 10:1 to 90:1, but as the ratio continued to increase, the extraction yield of TPC gradually decreased.Higher solvent-to-feed ratio indicates that a greater amount of solvent was used.The results agree with the previous findings on traditional maceration extraction of phenolics from dried chokeberry reported by Ćujić et al. (2016); where phenolics increase when more solvent is used.
This may be due to the saturation of phenolics in the solvent.When the amount of sample increased, more solvent is required to fully submerge the sample, hence minimizing the interaction between sample and solvent (Milutinović et al., 2014).In general, materials expand excessively when an extraction solvent is used in a bigger volume, which makes it simpler to break cell walls and extract polyphenols.However, in the presence of too much solvent, more energy will be absorbed by the solvent, and this reduces the material's capacity to absorb microwaves, which results in less effective extraction and a reduced polyphenol concentration in the extract (Maran et al., 2013).This could be the reason TPC yield decreased slightly with the further increase in solvent-to-feed ratio from 90:1 to 100:1.
The response surface plot of TPC as being influenced by the parameters X 2 (time of irradiation) and X 3 (solvent-to-feed ratio) is shown in Figure 1c, which is shown as nonsignificant with a p-value of .7030.The p-value was larger than .05,showing that X 2 X 3 's interaction effect on TPC yield was insignificant.This suggests that the solvent-to-feed ratio and time of irradiation each had an individual impact on the TPC yield but not their interaction effect.

| Analysis of the significance of MAE parameters on TFC
Table 3 and Figure 2a-c illustrate the effects of microwave power, time of irradiation, and solvent-to-feed ratio on the TFC of the bioactive compounds from Sarawak Liberica sp.coffee pulp.
As presented in Table 3, the linear effects of microwave power and time of irradiation and the interaction effect of microwave power and time of irradiation pair showed a strongly significant (p < .001)positive effect on TFC.In comparison, the quadratic terms of microwave power exhibited a strongly significant (p < .001)negative effect on TFC yield.Furthermore, the linear effect of solvent-to-feed ratio, the interaction effect of microwave power and solvent-to-feed ratio pair, and the quadratic terms of the time of irradiation showed a significant (p = .0016,p = .0402,and p = .0111,respectively) positive effect on TFC.In contrast, the quadratic terms of solvent-tofeed ratio showed a significant (p = .0016)negative effect on TFC.
Nevertheless, the interaction effect of time of irradiation and solvent-to-feed ratio pair exhibited a nonsignificant (p = .0714)negative effect on TFC, which means that the pair does not have any effect on TFC.An increase in microwave power, time of irradiation, and solvent-to-feed ratio from 119 to 617 W, 30 to 180 s, and 10:1 to 65:1, respectively, induces a rise in TFC.Further increasing microwave power and solvent-to-feed ratio from 617 to 700 W and 65:1 to 100:1 caused a decrease in TFC.The linear, interactive, and quadratic effects affecting TFC yield were very similar to those of TPC, which shows that the same factors that impacted the extraction of TPC also affected the extraction of TFC.It is reasonable to anticipate the observed outcomes, as flavonoids are a type of phenolic compound.Similar findings were reported by Woumbo et al. (2021), who noted that the responses of TPC and TFC were both influenced by the same effects for the MAE of polyphenols from soybean meal using the parameters (solvent/dry matter ratio, power, and time).

| Analysis of the significance of MAE parameters on TCC
As depicted in Figure 3a, the response surface plot of TCC influenced by parameters X 1 (microwave power) and X 2 (time of irradiation) is indicated as strongly significant with a p-value of <.0001.
The TCC yield increased with the microwave power and time of irradiation, and then, a decline was found as time of irradiation increased from 145 to 180 s.The result was consistent with  et al., 2015).Along with increasing microwave power, dipole rotations would also rise, leading to power degeneration inside the reaction mixture.This causes the reaction mixture to heat up fast, increasing the polysaccharide yield.The TCC increased with time of irradiation, and reached its maximum yield at 145 s, then fell.
The assumption is that the thermal accumulation in the extraction solution caused by the microwave energy's absorption accelerated the polysaccharide dissolution process into the solution for 145 s (Zheng et al., 2011).The deterioration of polysaccharides may occur, nevertheless, if they are exposed to microwave radiation over an extended period of time.
Figure 3b,c present the response surface plot of TCC influenced by parameters X 1 X 3 and X 2 X 3 , respectively, indicated as significant with a p-value of .0045and .0020,respectively.From the figures, it is observed that the yield of TCC increased with increasing solventto-feed ratio.This can be explained by the fact that increasing solvent volume makes the compounds more mobile from plant matrix, which would account for the rise in the TCC at a certain solvent ratio.
Moreover, increasing the solvent-to-feed ratio encouraged a rise in the extraction yields of bioactive compounds since a smaller amount of samples in the same volume of solvent might result in less saturation and thus boost overall extraction yields (Pham et al., 2017).
Similar results were observed in the extraction of bioactive compounds from coffee pulp waste, where the extraction yields declined with the sample-to-solvent ratio increased (Tran et al., 2022a(Tran et al., , 2022b)).

| Optimization and model validation
Based on the RSM analysis, an optimization procedure was performed to simultaneously maximize the yield of TPC, TFC, and TCC of the bioactive compounds from Sarawak Liberica sp.coffee pulp.
The optimal conditions for the maximum yield of TPC, TFC, and TCC were determined to be a microwave power of 700 W, time of irradiation of 180 s, and solvent-to-feed ratio of 86.644:1 with desirability of 0.977.To demonstrate the models' validity in predicting the optimal conditions for the responses, experiments were performed under the optimized conditions, and the outcomes were compared with the predicted optimal yields.As shown in Table 5, the percentage difference between predicted and experimental values are less than 5%, and at 95% confidence level, the experimental values fell within the PI range generated by Design Expert Software, which confirms that the optimal conditions for the maximum yield of TPC, TFC, and TCC were validated successfully and can be used for further extraction of bioactive compounds from Sarawak Liberica sp.coffee pulp.

| Comparison of MAE with CME and SE
Based on the optimal extraction conditions (microwave power of 700 W, time of irradiation of 180 s, and solvent-to-feed ratio of 86.644:1) for maximizing the extraction yield, the best resulting MAE process was used to compare with CME and SE on the extraction efficiency of TPC, TFC, and TCC and the results are shown in Table 6.

| Comparison of bioactive compounds' yield with those reported in previous studies
A comparison of the TPC, TFC, and TCC yields of bioactive compounds extracted from different parts of various coffee species is presented in Table 8.
Although the extraction yield of TPC and TFC from Liberica sp.coffee pulp in the present study was found to be lower compared to the other coffee species and parts listed in  Note: Results are presented as mean ± standard deviation of three replicates and the values in the same column with same superscript letter are not significantly different from each other (Tukey's test, p > .05).Abbreviations: TCC, total carbohydrate content (mg GE/g); TFC, total flavonoid content (mg QE/g); TPC, total phenolic content (mg GAE/g).
TA B L E 9 Yield of total phenolic content, total flavonoid content, and total carbohydrate content in different parts of Liberica sp.coffee.Abbreviations: TCC, total carbohydrate content (mg GE/g); TFC, total flavonoid content (mg QE/g); TPC, total phenolic content (mg GAE/g).

Sample
was slightly higher than those reported by Insanu et al. (2021).The slight difference in yields could be due to variations in the extraction conditions used in each study, such as solvent type, temperature, and duration of extraction.These variables can greatly influence the efficiency of extraction and the amount of bioactive compounds that can be extracted.The differences in coffee by-products used in each study are important to consider because they can affect the composition and concentration of bioactive compounds in the extract.Different parts of the coffee plant may contain varying amounts of specific compounds (Pohl et al., 2013).The information on total carbohydrate content determination using phenol-sulfuric method is lacking.

| CON CLUS ION
In conclusion, the optimal conditions for MAE of total phenolic, flavonoid, and carbohydrate content from Sarawak Liberica sp.
Saifullah et al. (2020), who reported that longer MAE irradiation period resulted in larger yield of bioactive compounds extracted from lemon-scented tee tree leaves.Time of irradiation represents the length of time the sample is submerged in the extraction solvent while being microwave irradiated.Longer radiation times resulted in more diffusion of the bioactive chemicals into the solvent, which increased extraction yields.

F
I G U R E 1 3D plot for total phenolic content presenting (a) the interaction effect of microwave power and time of irradiation.(b) Interaction effect of microwave power and solvent-to-feed ratio.(c) Interaction effect of time of irradiation and solvent-to-feed ratio.F I G U R E 2 3D plot for total flavonoid content presenting (a) the interaction effect of microwave power and time of irradiation.(b) Interaction effect of microwave power and solvent-to-feed ratio.(c) Interaction effect of time of irradiation and solvent-to-feed ratio.F I G U R E 3 3D plot for total carbohydrate content presenting (a) the interaction effect of microwave power and time of irradiation.(b) Interaction effect of microwave power and solvent-to-feed ratio.(c) Interaction effect of time of irradiation and solvent-to-feed ratio.
Thirugnanasambandham et al. (2015) andChen et al. (2007), who claimed that the increased microwave power and time of irradiation could accelerate the MAE of polysaccharide and the negligible effect was observed when the irradiation time exceeded a certain period.Radiation from microwave energy speeds up the lysis of the cell wall.Rapid transmission of magnetic energy to biomolecules increases the amount of power wasted inside the solvent, causing molecular mobility in the plant materials, and enhancing polysaccharide extraction effectiveness simultaneously(Maran Results are presented as mean ± standard deviation of three replicates and the values in the same column with same superscript letter are not significantly different from each other (Tukey test, p > .05).
coffee pulp were successfully optimized by RSM.Results suggest that three parameters, microwave power, time of irradiation, and solvent-to-feed ratio significantly enhanced the extraction yield of total phenolic, flavonoid, and carbohydrate content.The optimal MAE conditions generated by RSM are microwave power of 700 W, time of irradiation of 180 s, and solvent-to-feed ratio of 86.644:1.Under these conditions, the experimental results agreed with the predicted values; thus, the RSM models were verified.Comparison with CME confirmed the efficiency of MAE as it revealed the advantages of higher extraction yield within a shorter time of extraction.No significant difference was observed (p < .05) between MAE and SE on the extraction efficiency of TPC, TFC, and TCC.Nonetheless, MAE has the financial benefit of short extraction time with less solvent as compared to SE.The findings of this work offer fundamental understanding of MAE optimization for extraction of phytochemicals and may help with future research on bioactive compound extraction from coffee pulp for industrial purposes.

Microwave parameters Unit Notation Levels Low (−1) High (+1)
Design matrix of central composite design (CCD) with TPC, TFC, and TCC as responses.
mL.Firstly, 0.4 mL of the sample was mixed with 0.3 mL of sodium nitrite (NaNO 2 , 5% w/v).The mixture was let to stand for 5 min.Then, 0.3 mL of aluminum chloride (AlCl 3 , 10% w/v) was added to the mixture and shaken well.The mixture was allowed to stand TA B L E 2 TFC, and TCC of the bioactive compounds from Sarawak Liberica sp.coffee pulp.The experimental results for TPC, TFC, and TCC are depicted in Table2.The data demonstrated a significant dependency on the extraction conditions, indicating that the extraction process should be optimized.
As indicated by fit statistics in Table4, the level of correspondence between experimental and predicted values of TPC, TFC, and TCC was stated by R 2 values (TPC-.9845,TFC-.9576, and TCC-.9614).It indicates that the variables in the design matrix explain the variation occurring in TPC, TFC, and TCC by 98.45%, 95.76%, and 96.14%, respectively.The adjusted R 2 and predicted R 2 values were .9782and .9656for TPC, .9402and .8985for TFC, and

Table 8
Guglielmetti et al. (2019)kin as reported byGuglielmetti et al. (2019).This indicates that Liberica sp.coffee pulp can be a potential source of carbohydrates that can be used in various food applications.The high carbohydrate content in Liberica sp.coffee pulp worthwhile to explore the potential of Liberica sp.coffee pulp as a source of bioactive compounds for functional food and nutraceutical industries.The extraction yield of TCC from Liberica sp.coffee pulp was significantly higher compared to the TCC yield of Arabica sp.coffee husks and

Table 9
Predicted and experimental values of the responses at optimum extraction conditions.Note: Results are presented as mean ± standard deviation of three replicates and the values in the same row with same superscript letter are not significantly different from each other (Tukey and Kruskal-Wallis tests, p > 0.05).Consumed sample, time, and solvent during the preparation by three different extraction techniques.Yield of total phenolic content, total flavonoid content, and total carbohydrate content in different parts of various coffee species.
shows the comparison of TPC, TFC, and TCC yields of bioactive compounds extracted from different parts of Liberica sp.coffee.Based on the results, the TPC yield from Liberica sp.coffee pulp in the present study was the lowest among all, and its TFC yield TA B L E 5 TA B L E 8