Determination of The Mineral Composition of Foam-Mat Dried Ngu Powder Using Response Surface Methodology

Foam-mat drying of Ngu using egg white as a foaming agent was investigated. Three concentrations of Ngu sample-60, 80, and 100% were used for the study. The Ngu samples were enriched with egg white at 20, 25, and 30% concentration respectively at room temperature. The mixture was beaten for 5, 9, and 12 mins. A central composite response surface design was employed to investigate the impact of whipping duration, egg white, and Ngu concentration on the mineral composition of dried Ngu powder. Regression models showed that the experimental variables had significant (p<0.05) effects on the dry Ngu powder's iron, zinc, calcium, sodium, phosphorus, potassium, and magnesium contents. The optimal values for all responses were achieved using a combination of 10 mins of whipping time, 20.8% egg white, and 100% Ngu concentration. The optimal values for the mineral content of Ngu powder were determined to be 17.46 mg/g for iron, 1.03mg/g for zinc, 0.58 mg/g for calcium, 12.72 mg/g for sodium, 2.26 mg/g for phosphorus, 1.36 mg/g for potassium and 0.26 mg/g for magnesium.


INTRODUCTION
Food waste is one of the most challenging issues humankind is currently facing worldwide (Garcia-Garcia et al., 2017).Currently, food systems are extremely inefficient: it is estimated that between onethird and one-half of the food produced is lost before reaching a human mouth (Griffin et al., 2009).Nevertheless, reduction of the current levels of food waste must be accompanied by better management of the waste: inevitably there will always be some food waste.Food waste management, recovery, and utilization involve activities where discarded food materials are collected, sorted, processed, and converted into edible or non-edible new products (Thyberg and Tonjes, 2015).Agricultural waste materials such as palm bunch waste, cocoa pod, plantain peels, banana peels, maize cob, sugar beet waste, etc. contain a good percentage of potash (Hung Mo et al., 2014).When these materials are burnt, the resulting ashes contain oxides of potassium and sodium which when dissolved in water yield alkalis -the corresponding hydroxides, a basis for the production of various emulsions (Onyekwere, 1996).The filtrate obtained from the filtration of the mixture of this ash and water normally has a brown color and can emulsify oil, thus producing an emulsion with it.
One of the largest agricultural waste is the palm inflorescence (Uzodinma et al., 2014).In the Eastern part of Nigeria, some communities use the filtrate to produce an edible emulsion base (a mild soap) called "Ncha" or "Ngu".Ngu is an African salad dressing (water-in-oil emulsion) commonly used in the preparation of African salad (Abacha, Ighu), bitter yam (Dioscorea dumetorums), and processed oil bean seed (Ugba).
Emulsions are inherently thermodynamically unstable systems as when mixed, the oil and the aqueous phases continually strive to minimize contact (Rousseau, 2000).Drying the "Ngu" emulsion and rendering it into powder is therefore very important to create a convenient form of this popular traditional emulsion salad dressing and preserve the integrity of the emulsion.
The foam-mat drying technique is a process in which liquid or semi-solid food products like fruit juices, vegetable puree, or pastes from cereals are converted into stable foams by whipping air or inert gas in the presence of a food-grade foaming agent and/or stabilizer, with subsequent drying to reduce product moisture to 2-2.5% (Hardy and Jideani, 2017).Foam-mat drying can process hard-to-dry materials to produce products of the desired and required properties thereby retaining volatiles that otherwise would be lost during drying of nonfoamed materials (Wilson et al., 2012).The process produces an end product with favorable rehydration/controlled density and retains volatiles that would be lost when using other forms of drying (Kudra and Ratti, 2006).Foam-mat dried products contain improved reconstitution characteristics due to their open structure.More interest has developed in foam-mat drying because of the simplicity, cost-effectiveness, high-speed drying, and improved product quality it provides (Kandasamy et al., 2012).
Response surface methodology (RSM) is a collection of mathematical and statistical techniques that are useful for designing experiments and analysis of problems in which the response of interest is influenced by several variables and objectives (Iwe, 2000;Kalil et al., 2000).The designs capable of generating a response surface include central composite and Box-Behnken designs (Lucas, 1994).The response surface methodology is being increasingly used in the industry, with even more recent emphasis and application in chemical and processing fields on finding regions where there is improvement in response rather than finding the optimum response (Myers et al., 1989).The multivariate approach in RSM reduces the number of experiments, improves statistical interpretation possibilities, and indicates interaction among different parameters.Combinatorial interactions of drying parameters with the quality of dehydrated products and the optimum processes may be developed using an effective experimental design procedure.This study investigated the impact of specific experimental factors, utilizing the response surface methodology, on the mineral content of foam-mat dried Ngu powder.

Materials
Prepared Ngu (produced from 5litres of distilled water, 500g of Ncha, 800g of melted red palm oil, and 100g of Ogiri) was purchased from a professional Ngu maker at Amuda Isuochi, Umunneochi Local Government Area, Abia State, Nigeria.Fresh eggs were purchased from a local market in Umuahia, Abia State, Nigeria.

Foaming agent preparation
The foaming agent used was egg white/albumin.Egg white was obtained from fresh whole eggs and prepared for use as a foaming agent according to the method described by Lomakina and Mikova (2005).Fresh whole eggs (24 in number) were washed with distilled water and allowed to dry.Then, the egg shell was cracked with a stainless fork to make a small opening for the egg white to flow out into a clean bowl.The egg white was then whipped for 1-2 mins and kept (1-3 mins) at room temperature until ready for use.The egg white was used at different concentrations of 20, 25 and 30% (w/w).

Ngu foam preparation and foam-mat drying
The method for the production of foammat dried fruit and vegetable powder described by Balasubramanian et al. (2012) was modified and adopted for the production of foam-mat dried Ngu powder.One hundred milliliters (100 ml) of Ngu was mixed separately with the foaming agent (egg white) at different concentrations of 20, 25, and 30% EW (weight-to-weight basis).The different mixtures were blended using a highspeed mixer (Kenwood, 900 w) at speed of 6 for 5, 9, and 12 mins respectively (according to the different whipping times in the experimental design) at room temperature.Thereafter, the foamed Ngu emulsion was spread on trays (wrapped with foil), placed in the hot air oven (UNISCOPE SM9023 Surgifield Medicals, England), and dried at 70 o C for about 6-9 h.The dried Ngu was scraped off from the trays, pulverized, and packaged in a transparent bowl plate.

Experimental design
A central composite rotatable response surface design (CCD) for K = 3 described by Cornell (1990) was used (Equation 1) . =  0 + ∑    + ∑    2 + ∑∑      + (1) Where Y = dependent variable   and   = independent variables K = number of independent variables  0 = intercept (constant and regression coefficient of the model)  = random error terms Twenty-five (25) experimental runs generated based on CCD (K = 3) were used to study the linear, interactive, and quadratic effects of the independent experimental variables.The runs include five (5) experiments at the start, center, and corner points respectively.The center point (0, 0, 0) was replicated six times.The corner and start points were not replicated (Iwe et al., 1998).

Determination of mineral composition of dried Ngu powder
The dry ash acid extraction method (James, 1995) was used for mineral extraction.The weighed sample was burnt to ash in a muffle furnace.The resulting ash was dissolved in 10 ml of 2M HCl solution and diluted to 100 ml in a volumetric flask.It was filtered and the filtrate was then used for the determination of the different mineral elements.

Determination of Calcium and Magnesium
These minerals were determined by complexiometric titrimetry in which the versanate EDTA titrimetric method was employed.Twenty ml of the extract was dispersed into a conical flask and treated with pinches of the masking agents (sodium, potassium ferro cyanide).The flask was shaken to dissolve the mixture and 20 ml of ammonia buffer was added to raise the pH to 10.00 (a point at which both calcium and magnesium form complexes with EDTA).The mixture was titrated against 0.02N EDTA solution using Erichrome Black T as an indicator.A reagent blank was also titrated in each case, it turned from deep red to a permanent blue endpoint.The titration value represents both Ca 2+ and Mg 2+ alone in the test samples.Titration of calcium alone was done in similarity with the above titration however, 10% NaOH was used in place of ammonia buffer and solochrome dark blue indicator in place of Erichrome Black T. The Ca 2+ and Mg 2+ content from the titer values obtained was calculated as shown in Equation 2

Determination of Phosphorus
The phosphorus content in the samples was determined by the Molydbo vandate colorimetric method (James, 1995).A measured volume of the dry ash (2mg) digest of the sample was dispersed into a 50-volume flask.The same volume of distilled water and standard phosphorus solution were measured into different flasks, at the same time, to serve as a reagent blank and standard, respectively.Two milliliters (2ml) of the phosphorus color reagent (Molybdo vandate solution) was added to each of the flasks and the mixture was allowed to stand at room temperature for 15 mins.The content of each flask was diluted to a 50ml mark with distilled water and its absorbance was measured using a spectrophotometer at a wavelength of 540nm with the reagent blank at zero.The phosphorus content was calculated using Equation 3 below; P (mg/g) = 100 × Au × C × Vt ( 3

Determination of Sodium and Potassium
These were determined using the flame photometer (AOAC, 1990).The instrument digital flame photometer was set up according to the manufacturer's instructions.It was switched on and allowed to stand for about 15 mins to equilibrate.On the other hand, standard sodium and potassium solutions were prepared separately and serially diluted to contain 10, 8, 6, 4, and 2pp of Sodium (Na) and Potassium (K) respectively.Furthermore, 1 ml of each standard was aspirated into the instrument after calibration and sprayed over the non-luminous flame.The optical density of the resulting emission from each standard solution was recorded.The appropriate element filter (Na or K) was put in place with the standards measured before flaming while the test sample extract was measured in time and plotted into a standard curve which was used to extrapolate the content of each test element and calculated as shown below; Na or K (mg/g) = X x Vt x D x 100 Va x W (4) Where:

Determination of Iron
Iron was determined using the phenylthroline method as described by (James, 1995).The sample (5ml) with 3ml concentrated HCL and 1ml hydroxylamine solution were mixed in a test tube.A few glass beads were added and the mixture was heated till it started boiling.It was then cooled to room temperature and transferred to a 50ml volumetric flask.Ten milliliters of ammonium acetate buffer solution and 2ml phenylthroline solution were added and distilled water to the mark.It was thoroughly mixed and allowed to stand for about 10-15 mins for maximum development.The absorbance was read at 510nm.A standard curve using 2.0, 4.0, 6.0, 8.0, and 10.0 ml/std iron solution was prepared and put into a 100volumetric flask from the stock solution.

Determination of Zinc
Zinc was determined using an atomic absorption spectrophotometer as described by James (1995).The standard series and suitable diluted sample and blank digests were aspirated into an atomic absorption spectrophotometer calibrated for zinc measurement at a wavelength of 213.9nm.A calibrated curve of the absorbance readings of the standard solution against the concentrations was plotted and used to determine the concentration of the unknown.The concentration of the Zinc in the dried sample was calculated as follows: Zn (mg/g) =

Optimization
The desirability method of optimization described by Nazni and Gracia (2014) was adopted in the optimization of these multiresponses simultaneously, formulation composition and setting the optimization goals.The application of the desirability function combines all the responses in one measurement and offers the possibility to predict optimum levels for the independent variables.Experimental analysis, literature information, research objectives, and panelists' preferences were considered in maximizing/minimizing the independent variables and the responses.Sodium is the only response that was minimized.Low sodium is necessary for the system due to its support for cell functionality without promoting heartbeats or high blood pressure (Agu et al. 2023).The general approach was to first convert each response (yi) into an individual desirability function (di) that varies for the range 0 ≤ di ≤ 1, whereas if the yi is at its goal or target then, di = 1, and if the response is outside acceptable region di = 0. Then the design variables are chosen to maximize the overall desirability, D = (d1 x d2 x ……….dn) 1/n (6) Where n = number of responses

Statistical analysis
The Statistical Package for Social Sciences (SPSS, version 20, IBM SPSS Statistics 20, IBM Corporation, Armonk, NY) was used to obtain the mean, standard deviation, and analysis of variance (ANOVA) was used to judge for significance (p≤0.05).Means were separated using Duncan's Multiple Range Test.

RESULTS AND DISCUSSION Mineral composition of dried Ngu powder
Results of the mineral composition and pH values of dried Ngu powder are shown in Table 1 below.

Iron (Fe)
Results in Table 1 show that there was a significant difference in the Fe content of the dried Ngu samples which varied from 17.00 to 17.71mg/g.Uzodinma et al. (2014) reported the Fe content of palm filtrate from palm bunch waste products to be 0.72 mg/g.Differences in the raw materials utilized, as well as environmental and experimental conditions, can explain the variation in Fe content.Dry Ngu powder produced from the formulation having 60% Ngu concentration, 30% egg white and whipped for 12 mins had the highest Fe content (17.71mg/g), while dry Ngu powder produced from the formulation having 60% Ngu concentration, 25% egg white and whipped for 9 mins and dry Ngu produced from the formulation having 80% Ngu concentration, 25% egg white and whipped for 5mins had the lowest Fe content (17.00mg/g).This variation in Fe content could be a result of the concentration of the egg white, as the Fe content was observed to vary directly with the quantity of egg white used.The same observation was made by Jyothsna et al. (2023) who reported that albumin has a statistically significant positive correlation with serum iron.On the other hand, Werner et al. (2022) noted that providing eggs daily for 6 months did not affect iron status or anemia prevalence among young Malawian children.
Response surface analysis showed (Table 2) that the iron content of the dried Ngu powder was significantly (p≤0.05)influenced by the quadratic effect of the egg white concentration.The non-significant effect of whipping time was not surprising as the whipping time is not expected to destroy or enhance the Fe content of the food.The model coefficients are presented in Table 2.An acceptable regression coefficient (R 2 ) value of 70.2% shows that 70% variability in the data is explained by the deduced model.Although the model was not significant, the non-significant (p>0.05)lack of fit in addition to the acceptable R 2 value ≥70%, shows the predictability of the response (iron) by the model (Gbenyi et al., 2015).
Concerning optimization, the optimum possible Fe value is 17.46mg/g.To obtain this, a whipping time of 10 mins, egg white concentration of 20.8%, and Ngu concentration of 100% should be used.The final model equation is: Fe = 17.20 + 0.21X2 2 -0.025X2X3 (7) Where, X2 = Egg white concentration and X3 = Ngu concentration

Magnesium (Mg)
There was a significant (p<0.05)difference in the Mg content of the dry Ngu powder, which varied from 0.180 to 0.535mg/g as shown in Table 1.It was observed that the Mg content increased with a decrease in the Ngu concentration.This is evident in the negative coefficient of estimate (-0.050) shown by the regression analysis.The magnesium (Mg) content of palm filtrate has been reported in various studies.Vijayakumari et al. (2017) found that the pulp of the palmyra palm, a type of palm, contains 10.2 mg/100g of magnesium.Similarly, Santana et al. (2020) noted a magnesium content of 0.70-1.78% in the forage palm, another type of palm.Furthermore, Uzodinma et al. (2014), reported the Mg content of palm filtrate from palm bunch waste products to be 0.028mg/g.The concentration of magnesium (Mg) in palm filtrate is influenced by various factors.Szydłowska-Czerniak et al. (2013) found a significant positive correlation between Mg and other metals in palm oils, suggesting that the presence of these metals may affect Mg concentration.Also, differences in the raw materials utilized may be another reason for the varied results.
Regression analysis showed (Table 3) that the linear effects of egg white concentration and Ngu concentration, interaction between whipping time and egg white concentration, and interaction between egg white and Ngu concentrations significantly (p<0.05)affected the magnesium content of dry Ngu powder.These variables accounted for a 74% variation in the Mg content of the product.The significant effect on the Mg content of the dried Ngu powder and the non-significant lack of fit of the model confers goodness of fit (Gbenyi et al., 2015).The optimum value of Mg is 0.26mg/g, obtainable by using 100% Ngu concentration, 20.8% egg white, and 10 mins whipping time.The final model equation is: Mg = 0.21 + 0.033X2 -0.050X3 -0.059X1X2 -0.046X2X3 ( 8)

Potassium (K)
The potassium content of Ngu powder as shown in Table 1 ranged from 0.86 to 1.56mg/g.Dry Ngu powder produced from 60% Ngu concentration, 20% egg white, whipped for 5 mins had the lowest K content (0.86mg/g) while dry Ngu powder produced from 100% Ngu concentration, 30% egg white, whipped for 5 mins and 80% Ngu concentration, 30% egg white, whipped for 9mins had the highest K content (1.56mg/g).It was observed that the K content of the product increased with the concentrations of Ngu and egg white.This could be attributed to the high K content of egg white which is 169mg/100g (USDA, 2015).
Regression results (Table 4) showed that the model was significant (p<0.05).The linear effect of whipping time, egg white, and Ngu concentrations significantly (p<0.05)affected the K content of dry Ngu powder.The quadratic term of the whipping time and interactions between whipping time and egg white concentration; egg white and Ngu concentrations were also significant (p<0.05).These accounted for 88% variation in the K content of the dry Ngu powder.The non-significant (p>0.05)lack of fit with the acceptable R 2 value of 88% shows the predictability of the response (K) by the model.Results further showed that all independent variables were significant (p<0.05).From optimization, the optimum K value is 1.36mg/g, obtainable from using 100% Ngu concentration, 20.8% egg white, and 10 mins whipping time.The final model equation: K = 1.43 + 0.054X1 + 0.11X2 + 0.13X3 -0.12X1 2 -0.070X1X2 ( 9)

Phosphorus (P)
The phosphorus content of the dry Ngu powder ranged from 1.34mg/g to 2.51mg/g as shown in Table 1.There was a significant (p<0.05)difference in the phosphorus content of the dry Ngu product but no particular trend was observed.Uzodinma et al. (2014) reported no trace of P in the filtrate from palm bunch waste.The concentration of phosphorus in palm filtrate is influenced by a variety of factors.Saleh (2012) found that phosphorus levels significantly affected the chemical composition of date palm fruit, suggesting a potential impact on palm filtrate.Oviasogie and Uzoekwe (2011) demonstrated that the use of palm oil mill effluent can increase the concentration of available phosphorus in soil, which may subsequently affect the phosphorus content of palm filtrate.Finally, Maranguit et al. (2017) discussed the impact of land-use change on phosphorus availability in tropical soils, indicating that changes in land use, such as the conversion of forests to palm plantations, could also affect the phosphorus concentration in palm filtrate.
Regression analysis showed (Table 5) that the linear effects of whipping time, egg white, and Ngu concentrations were significant (p<0.05) and they accounted for 94% of the total variation in the P content of the product.Results further showed that the quadratic effect of whipping time and interactions between whipping time and egg white; whipping time and Ngu concentration significantly (p<0.05)affected the P content of the product.The insignificant (p>0.05)lack of fit and the R 2 value of 94% show the predictability of the response (P) by the model.From optimization, the optimum P content in dry Ngu powder is 2.26mg/g and is obtainable from 100% Ngu concentration, 20.8% egg white, and 10 mins whipping time.The final model equation is:

Zinc (Zn)
Results in Table 1 show that there was a significant (p<0.05)difference in the zinc content of dry Ngu powder which ranged from 1.025mg/g to 1.045mg/g The Zinc content of dry Ngu powder was lower than that of palm bunch waste (10.75mg/g) and higher than the filtrate from palm bunch waste product (0.72mg/g) as reported by Uzodinma et al. (2014).Differences in the raw materials utilized may be the reason for the varied results.
Regression analysis showed (Table 6) that the quadratic effect of Ngu concentration significantly (p<0.05)affected the Zn content of dry Ngu powder and they accounted for 71% of the total variation in the Zn content of the product.An acceptable regression coefficient (R 2 ) value of 0.7078 shows that 71% variability in the data is explained by the model.Optimization shows that optimum Zinc is 1.03%, obtainable from 100% Ngu concentration, 20.8% egg white, and 10 mins whipping time.The final model equation is: Zn = 1.03 + 2.714 -E-003X3 2 (11)

Calcium (Ca)
Dry Ngu powder produced from 80% Ngu concentration, 20% egg white, and whipped for 9 mins had the highest calcium content (0.805mg/g) while dry Ngu powder produced from 60% Ngu concentration, 30% egg white, 12 mins whipping time had the lowest calcium content (0.295%).There was a significant (p<0.05)difference in the calcium content of dry Ngu powder as shown in Table 1.However, no definite trend was observed.The calcium content of the product compared well with the calcium content of filtrate from palm bunch waste product which is 0.96mg/g (Uzodinma et al., 2014).The interaction between calcium and fatty acids, such as palmitic acid, may affect the calcium concentration in the palm filtrate (Watras et al. 1984).Optimization shows the optimum calcium value is 0.58mg/g, obtainable from 100% Ngu concentration, 20.8% egg white, and 10 mins whipping time.
Regression analysis showed (Table 7) that the model was insignificant (p>0.05).The linear effect of whipping time significantly (p<0.05)affected the calcium content of the dry Ngu powder.The insignificant (p>0.05)model and significant lack of fit in addition to the low R 2 value of 42% shows the unpredictability of the response (Ca) by the model.The final model equation is: Ca = 0.05 -0.080X1 ( 12)

Sodium (Na)
The sodium content of the dry Ngu powder varied from 9.33 to 15.06mg/g (Table 1).
There was a significant (p<0.05)difference in the Na content of dry Ngu powder.It was observed that the Na content increased with egg white concentration and decreased with an increase in Ngu concentration.This may be attributed to the high Na content of egg white (154mg/100g) reported by Roe et al. (2013).According to Uzodinma et al. (2014), the sodium content of oil palm biogenic waste ranged from 0.00mg/g (filtrate from palm bunch waste) to 0.013mg/g (palm bunch waste).Regression analysis (Table 8) showed that the model was significant (p<0.05).The linear effect of egg white and Ngu concentrations significantly (p<0.05)affected the Na content of dry Ngu powder.The quadratic term of whipping time and egg concentration; interactions between whipping time and egg concentration and egg concentration and Ngu concentration significantly (p<0.05)affected the Na content of the product.These accounted for 81% of the total variation in the Na content of the product.There was a significant (p<0.05)lack of fit and a high coefficient of determination (R 2 ) value of 81% signifying the adequacy of the model.Regarding optimization, the optimum value of sodium content in the dry Ngu powder is 12.72mg/g, obtainable from 100% Ngu concentration, 20.8% egg white, and 10mins whipping time.

CONCLUSION
Dry Ngu powder was produced by adopting selected whipping time, egg white, and Ngu concentration.There was a significant (p<0.05)difference in the mineral composition of dry Ngu powder.Iron content ranged from 17.01 to 17.05mg/g and regression analysis showed that the quadratic effect of egg white concentration had significant effects on the Fe content of the samples.The magnesium content of dry Ngu powder varied from 0.11 to 0.54mg/g.Linear effects of Ngu concentration, the interaction between whipping time and egg white, and the interaction between egg white and Ngu concentration had significant (p<0.05)effects on the magnesium content of dry Ngu powder.Potassium content varied from 0.86 to 1.56mg/g.Regression analysis showed that the linear effect of whipping time, egg white, and Ngu concentration had significant effects on the K content of the product.The quadratic term of whipping time and interaction between whipping time and egg white concentration were also significant (p<0.05).The phosphorus content varied from 1.34 to 2.51mg/g.Regression analysis showed that the linear effect of the independent variables had significant (p<0.05)effects on the phosphorus content of the dry Ngu powder.The quadratic effect of whipping time also significantly affected the dry Ngu powder.Zinc content varied from 1.025 to 1.045 mg/g.Calcium content ranged from 0.30 to 0.81 mg/g.Sodium content varied from 9.33 to 15.06 mg/g.Optimization showed the optimum value obtainable was 17.4mg/g of Fe, 0.26mg/g of magnesium, 1.36 mg/g of potassium, 2.26 mg/g of phosphorus, 1.03 mg/g of Zinc,.12.72 mg/g of sodium, and the optimum value for calcium was 0.58 mg/g This research recommends that oil palm processing industries should incorporate into their processes the utilization of at least part of their palm lunch wastes for Ngu powder production as a way of recycling waste.
Food Scientech Journal 6(1) 2024, pp 67-82 below: Ca or Mg (mg/g) = 100 × T -B (N × Ca or Mg) of titrant (0.02N EDTA) Vt = Total volume of extract Va = Volume of extract titrated extract volume Va = Volume of extract analyzed W = Weight of sample Au = absorbance of sample

Food
Scientech Journal 6(1) 2024, pp 67-82 X = Concentration of the test element from the curve Vt = Total extracted volume Va = Volume of extract analyzed D = Dilution factor W = Weight of sample

(
−)× ××1000  (5) Where: a =concentration of Zn in the solution b = concentration of Zn in the mean values of blanks v = final volume of the digestion process w = weight of the sample taken f = the dilution factor

Table 1 .
Mineral composition of dry Ngu powder Means bearing different superscripts in the same column are significantly (p<0.05)different