Ultrasound-assisted hydration of finger millet (Eleusine Coracana) and its effects on starch isolates and antinutrients

Highlights

• •Ultrasound-assisted hydration process was developed and optimized for finger millet.
• •Hydration was optimum at 65.73% amplitude, 26.13 min treatment and 1:3 grain-water ratio.
• •Phytate and tannin (antinutrients) were reduced by 66.98 and 62.83%, respectively.
• •Morphology, water binding capacity, and solubility of the isolated starch improved.
• •The ultrasound-assisted hydration is a better alternative for finger millet.

Abstract

Finger millet (Eleusine Coracana) is rich in nutrients and minerals. The iron and calcium contents are comparatively higher than other cereal crops. Finger millet also has some antinutrients such as tannins and phytates, that needs to be removed for maximum health benefits. Traditionally, these antinutrients are removed by the hydration process. The conventional hydration process is time cumbersome and often results in poor quality grains. Ultrasonication during hydration of finger millet could reduce the processing time and antinutrient content in finger millet. The ultrasound amplitude, treatment time, and grain to water ratio during hydration were optimized. An ultrasound amplitude of 66%, treatment time of 26 min, and a grain to water ratio of 1:3 resulted in best desirability parameters with a reduction in phytate and tannin contents of the finger millet by 66.98 and 62.83%, respectively. Ultrasonication during hydration increased the water binding capacity and solubility of the finger millet starch. XRD study of the starch isolates confirmed the increased crystallinity of the particles. FESEM of the starch isolates also confirmed that ultrasound-assisted hydration of finger millet resulted in the desired size reduction and homogeneous distribution of starch particles. The optimized ultrasound-assisted hydration could be adopted and scaled up for bulk processing of finger millets.

1. Introduction

Finger millet (Eleusine coracana) is a cereal crop, gaining considerable popularity recently due to its higher nutrient content and higher adaptability to stressed climatic conditions. Finger millet is native to Asia and Africa and is being used as a staple food. The crop is rich in dietary fiber, calcium, iron, β-glucan, amylose, antioxidant, and total sugars [1]. Finger millet is gluten-free, suitable for people having celiac disease. It also has antimicrobial, hypoglycemic, anti-ulcerative, and wound healing properties [2]. However, it also has some antinutrients, such as tannins and phytates [3]. These antinutrients are responsible for changing the nutritive quality of the cereal crop. Tannins reduce protein digestion, and phytates chelate the minerals, decreasing the availability for absorption in the human gut [4]. Hydration of the crop during processing helps in increasing the availability of nutrients and reducing the antinutrients. It also softens the texture of the cereal grain for smooth grinding and extraction [5]. The conventional hydration process is time cumbersome and results in fermentation, off-flavor development, and leaching of nutrients. The longer hydration time also increases the chances of microbial contamination. Thermal hydration could decrease the processing time. However, the high temperature in this method affects the nutritional and biochemical properties [5]. There is a need to develop a faster non-thermal hydration process for finger millet. Ultrasound-assisted hydration has been reported to be an effective non-thermal method for cereal and legumes [4], [5]. The application of ultrasound treatment is associated with lower processing time, lower energy consumption, and higher nutrient retention in food. It also alters the starch characteristics of food grains [8], [9]. The present work explores the possibility of an ultrasound-assisted hydration protocol for finger millet. The ultrasound amplitude, treatment time, and grain to water ratio during hydration have been optimized for faster hydration, reduced antinutrients, i.e., tannin and phytates, and better starch characteristics of the finger millet.

2. Materials and methods

The finger millet was procured from the locally at Rourkela, Odisha, India. The grains were appropriately cleaned and stored in the laboratory at ambient conditions with an initial moisture content of 12.3% (dry basis, db). These samples were used for further experiments.

2.1. Ultrasound-assisted hydration

The ultrasound-assisted hydration experiments were designed using a three-factor Box-Behnken design in Design-Expert software (Ver-11.0, Stat-Ease Inc., USA). The independent process parameters and their ranges were decided based on the results of preliminary experiments. Finger millet (50 g) was soaked in distilled water with varying grain to water ratio (1:3 to 1:6) in a glass beaker. A probe-type ultra-sonicator (Qsonica sonicators, Newtown, USA) with a maximum power input of 700 W and 20 kHz frequency was used for the experiments. The Schematic diagram of ultrasound-assisted hydration of finger millet is shown in Fig. 1. The ultrasound amplitude was varied between 30 and 70% with a pulse on/off time of five seconds. The ultrasound treatment/hydration time was varied in the range of 10 to 30 min. The energy reading (Joule, J) on the instrument was noted as the applied energy for each treatment. The calorimetric enery for each treatment was calculated as per Kikuchi et al. (2011) [10]. The corresponding energy values have been given in Table 1. The experiments were done in triplicate at seventeen combinations of process parameters with five center and twelve axial points (Table 1). The data were fitted and analyzed using a non-linear second order regression equation (Eq. (1)).

$$y_i=b_0+\sum _{i=1}^{\infty }{b}_i{x}_i+\sum _{i=1}^{\infty }{b}_{\mathit{ii}}{x}_i^2+\sum _{i=1}^{\infty }\sum _{i=1}^{\infty }{b}_{\mathit{ij}}{x}_i{x}_j$$ (1)

where x_i (i = 1, 2, 3) are independent variables (ultrasound amplitude, treatment time, and grain to water ratio), respectively, and b₀, b_i, b_ii, and b_ij are coefficients for intercept, linear, quadratic, and interactive effects respectively.

Fig. 1.

Schematic diagram of ultrasound assisted hydration of finger millet.

Table 1.

Experimental design of ultrasound-assisted hydration process according to Box-Behnken design.

Run	Ultrasound amplitude (%)	Treatment time (min)	Grain to Water ratio	Applied energy (J)*	Calorimetric energy (J)**
1	30	30	1:4.5	28,847	7524
2	70	20	1:6	46,285	16,302
3	50	30	1:6	42,701	13,794
4	50	20	1:4.5	36,220	11,286
5	50	20	1:4.5	36,220	12,226
6	50	20	1:4.5	36,220	11,756
7	50	20	1:4.5	36,220	12,226
8	30	20	1:3	17,201	5016
9	30	20	1:6	18,394	8778
10	50	20	1:4.5	36,220	12,226
11	50	30	1:3	52,126	9405
12	70	20	1:3	47,049	10,032
13	70	30	1:4.5	72,078	16,929
14	50	10	1:6	16,283	8778
15	30	10	1:4.5	8609	4702
16	70	10	1:4.5	25,546	7524
17	50	10	1:3	15,929	6897

*As per instrument reading; **Calculated value as per Kikuchi et al. (2011) [10].

Finger millet was also hydrated following the conventional method with a constant grain to water ratio (1:3), hydration temperature (70 °C), and hydration time of two hours [11], [12]. The results of the ultrasound-assisted hydration process were compared with that of conventionally hydrated finger millet samples.

2.2. Process optimization and statistical analysis

The ultrasound-assisted hydration process was optimized based on quality characteristics. Design Expert (Ver. 11) software was used for numeric optimization by response surface methodology. The validity of the developed model was perceived with the optimal condition predicted by the model. The average experimental values of quality characteristics were compared with the predicted values from the model. After model validation, the optimized sample was compared with that of conventionally hydrated and raw finger millets by using Duncan’s test in SPSS software (Ver. 22).

The experimental data were statistically analyzed by ANOVA in Design Expert (Ver. 11) software to determine the effect of independent variables on the dependent quality characteristics. The adequacy of the model was checked by a lack of fit and R² value.

2.3. Quality characterization of hydrated finger millet

Various quality characteristics such as moisture content, total phenolic content, tannin content, and phytate content of raw, conventionally hydrated, and ultrasound hydrated finger millet samples were determined. The starch was isolated from each of the samples. Water-binding capacity, solubility, and degree of gelatinization were determined by the following methods.

2.3.1. Moisture content (MC)

The moisture content of the samples was determined following the standard oven-dry method [7]. The samples were dried in a hot air oven at 105 °C for 24 h. The pre and post drying weights of the samples were used to calculate the moisture content.

2.3.2. Total phenolic content (TPC)

The total phenolic content of the sample was determined as per Siroha et al. [13] with little modification. The finger millet extract was prepared by incubating the ground sample in 80% methanol at 30 °C for 15 h instead of shaking. It was further centrifuged at 3000g for 10 min. An aliquot (0.5 ml) of the supernatant was taken in a test tube for total phenolic content determination. Sodium carbonate (6 g/L) followed by Folin-ciocalteu reagent (20 fold) were added for further reaction to get molybdenum blue color. The colored samples were incubated in the dark for 90 min, and the absorbance was taken at 725 nm using a UV–Vis Spectrophotometer (AU 2701, Systronics India Ltd.) with Gallic acid (R² = 0.995) as the standard. Total phenolic content was expressed as mg of gallic acid equivalent (GAE) per 100 g of sample.

2.3.3. Tannin content (TC)

The tannin content of the finger millet sample was determined by the modified Vanillin-HCL staining method [14]. Instead of shaking, the sample was incubated while extraction. One gram ground sample in 10 ml of acidified methanol (1 ml concentrated HCl in 100 ml methanol) was incubated at 30 °C for 15 h. The sample was then centrifuged at 10000 rpm for 15 min. An aliquot (1 ml) of the supernatant was taken in a test tube. Vanillin-HCl staining reagent was prepared by adding equal volumes of vanillin solution (4 g/100 ml methanol) and acidified methanol (8 ml concentrated HCl in 100 ml methanol). Tannic acid solution (1 g of tannic acid in 100 ml of acidified methanol) was used as the standard (R² = 0.9969). The Vanillin-HCl reagent (5 ml) was added to the extract and the standard. The samples were incubated in the dark for 20 min, and then the absorbance was measured using a UV–Vis Spectrophotometer (AU 2701, Systronics India Ltd.) at 500 nm against Vanillin-HCl reagent as blank. The tannin content is expressed as mg of tannic acid equivalent per 100 g of sample.

2.3.4. Phytate content

The phytate content of the finger millet sample was determined as previously described [14]. A 0.1 g of sample was taken in 10 ml of HCl and was shook for 1 h. The sample was then centrifuged at 5000 rpm for 15 min. An aliquot (0.5 ml) of supernatant was taken in the test tube, and 1 ml ammonium iron sulfate (0.2 g of ammonium iron sulfate in 100 ml HCl and made up to 1000 ml distilled water) was added. The contents were boiled for 30 min and then rapidly cooled to 25 °C in an ice water bath. Two milliliters of Bipyridine solution (1 g of bipyridine and 1 ml of thioglycolic acid in 100 ml of water) were added to the sample and allowed further reaction for one minute. The absorbance was measured at 519 nm against distilled water by UV–Vis Spectrophotometer (AU 2701, Systronics India Ltd.). The phytic acid sodium salt hydrate was used as the standard. The standard solution (R² = 0.9925) was prepared by taking 130 mg of phytic acid sodium salt hydrate in 100 ml of 0.1 N HCl. This stock solution was used for different dilutions to get different concentrations.

2.3.5. Degree of gelatinization

The degree of gelatinization (DG) of finger millet was determined following a previously described method [15]. The powdered sample (0.2 g) was dissolved in distilled water (10 ml) and stirred for 5 min. The stirred sample was then centrifuged at 1500 rpm for 25 min. The supernatant of 1 ml was diluted to 10 ml by adding distilled water. Further, 0.1 ml of iodine solution was added. The final solution was used to measure the absorbance of the water-soluble components of the sample.

Similarly, alkali solubilized solution was prepared using 10 ml of 10 M KOH. The absorbance of both water and alkali solubilized samples were taken at 600 nm by a UV–Vis Spectrophotometer (AU 2701, Systronics India Ltd.). The iodine reagent was taken as blank. The degree of gelatinization was calculated by using the following equation.

$$DG\left(\%\right)=\frac{\mathit{Absorbance}ofwatersolubalizedsample}{\mathit{Absorbance}ofalkalisolubalizedsample}\times 100$$ (3)

2.3.6. Starch isolation

Starch was isolated following the previously described alkali method [5] with minor changes. The finger millet powder was sieved and mixed with 10 ml of 0.2% NaOH solution in a vortex mixer (30 min at 1800 rpm), and the supernatant was removed. The precipitate was again added with 10 ml of 0.2% NaOH solution in a vortex mixer (30 min at 1800 rpm), and the supernatant was again removed. The procedure was repeated three times. The starch precipitate was resuspended in 100 ml of distilled water and neutralized to pH 7.0 by adding 0.1 N HCl solution. The neutralized starch dispersion was centrifuged at 1800g for 15 min. The supernatant was removed, and the remaining solid was washed with 95% ethanol. The final supernatant was removed, and the isolated starch was dried overnight at 40 °C. The dried starch was stored in a ziplock pouch in a desiccator for further analysis.

2.3.7. Water binding capacity of isolated starch

The water-binding capacity (WBC) of isolated starch from each experiment was determined by the previously described method [5] with minor modification. Isolated starch of 0.25 g was taken in 10 ml of distilled water and mixed in a vortex mixer (10 min at 1800 rpm). The dispersion was then centrifuged at 3000g for 10 min. The supernatant was removed, and wet starch was weighed and dried overnight at 110 °C. The weight of dried starch was taken, and water-binding capacity was calculated from Eq. (4).

$$\mathit{Water}bindingcapacity(\%)=\frac{(WS-{\mathit{Starch}}_{\mathit{dwb}})}{{\mathit{Starch}}_{\mathit{dbw}}}\times 100$$ (4)

where: WS = weight of wet starch sediment; Starch_dwb = weight of dried starch; Starch_dbw = dry basis starch weight

2.3.8. Water solubility of isolated starch (WS)

The solubility of isolated starch was determined by the previously described method [16]. Isolated starch of 0.1 g in 10 ml of distilled water was heated in a water bath at 80 °C for 30 min. The suspension was then centrifuged at 3000g for 15 min. An aliquot (5 ml) was carefully removed and was dried overnight at 110 °C. The weight of swollen starch was taken. The water solubility of isolated starch was calculated by Eq. (5).

$$\mathit{Water}solubility=\frac{\mathrm{D}\mathrm{S}}{{\mathit{Starch}}_{\mathit{dbw}}}\times 100$$ (5)

where: DS = Weight of dried supernatant (g); Starch_dbw = dry basis starch weight (g)

2.3.9. Surface morphology of isolated starch

The surface morphologies of the isolated starch from the raw, conventionally hydrated, and optimized ultrasound hydrated finger millet were analyzed by Field Emission Scanning Electron Microscopy (FESEM, Nova NanoSEM, FEI, USA) at 10 kV. The powder was placed on the plasma coated surface covered with carbon adhesive tape.

2.3.10. X-ray diffraction of isolated starch

The crystallinity of the isolated starch from the raw, conventionally hydrated, and optimized ultrasound hydrated finger millet was analyzed by an X-ray diffractometer (XRD, D8 Advance A25, Brucker, USA). The scattering angle and step sizes were 5–40° and 0.02, respectively. The scan rate was set as 1°/min. The relative crystallinity was calculated by the following Eq. (6). The areas under the peaks were analyzed by Origin (2016) software. The total area under the curve was the summation of all the areas under different peaks.

$$\mathit{Relative}crystallinity(\%)=\frac{\mathit{Area}underpeaks}{\mathit{Total}areaunderthecurve}\times 100$$ (6)

3. Results and discussion

3.1. Model fitting for ultrasound-assisted hydration

The experimental data (Table 2) on various quality characteristics such as moisture content, total phenolic content, tannin content, phytates, degree of gelatinization of ultrasound hydrated finger millet, and water binding capacity and water solubility of isolated starch were fitted to the model given in Eq. (1). The ANOVA results for the test of significance (p < 0.05) and lack of fits are shown in Table 3. The R², adjusted R², and coefficient of variation (CV) of the model for each of the quality characteristics are given in Table 3 and discussed in the subsequent sections. The influences of linear, quadratic, and interaction terms on each of the quality characteristics are also shown in Table 3. A second-order regression equation fitted well for the results with an insignificant lack of fit. The uncoded regression coefficients of the model are shown in Table 4. For each of the quality characteristics, the adequate precision values were more than four, suggesting adequacy of the model.

Table 2.

Quality parameters of ultrasound hydrated finger millet.

Run	Moisture content (%db)	Total phenolic content (mg GAE/100 g)	Tannin content (mg TAE/100 g)	Phytates (mg/100 g)	Degree of Gelatinisation (%)	Water binding capacity of starch (%)	Water solubility of starch (%)
1	64.88 ± 0.21	203.52 ± 0.11	96.22 ± 0.21	486.20 ± 0.04	164.16 ± 0.36	189.00 ± 0.41	09 ± 0.01
2	71.44 ± 0.03	190.98 ± 0.11	62.50 ± 0.09	237.55 ± 0.02	177.58 ± 0.26	279.60 ± 0.65	00 ± 00
3	51.77 ± 0.12	217.54 ± 0.36	72.31 ± 0.15	473.03 ± 0.02	163.24 ± 0.26	244.80 ± 0.33	04 ± 0.01
4	32.46 ± 0.03	195.35 ± 0.31	44.19 ± 0.23	417.49 ± 0.04	125.26 ± 0.60	178.40 ± 0.61	31 ± 0.03
5	42.43 ± 0.96	206.19 ± 0.33	51.23 ± 0.29	458.18 ± 0.04	144.48 ± 0.20	167.94 ± 0.05	33 ± 0.03
6	38.54 ± 0.02	215.96 ± 1.58	64.02 ± 0.20	443.65 ± 0.05	149.31 ± 1.09	158.23 ± 0.04	33 ± 0.02
7	42.40 ± 0.03	212.35 ± 1.05	59.12 ± 0.11	419.24 ± 0.05	138.06 ± 0.26	128.35 ± 0.05	33 ± 0.02
8	49.63 ± 0.04	223.38 ± 1.54	130.36 ± 0.44	440.42 ± 0.02	138.15 ± 0.47	141.20 ± 0.02	00 ± 00
9	26.60 ± 0.28	237.05 ± 1.23	154.45 ± 0.07	438.16 ± 0.07	133.79 ± 0.47	153.65 ± 0.33	21 ± 0.04
10	45.27 ± 0.03	242.92 ± 0.59	56.48 ± 0.12	479.43 ± 0.03	152.98 ± 2.02	178.33 ± 0.27	34 ± 0.02
11	46.86 ± 0.11	255.78 ± 0.82	42.58 ± 0.35	455.70 ± 0.02	169.63 ± 0.83	274.40 ± 0.41	18 ± 0.01
12	56.89 ± 0.03	233.08 ± 0.24	59.02 ± 0.07	185.32 ± 0.02	158.91 ± 1.29	287.80 ± 0.08	09 ± 0.02
13	76.65 ± 0.03	274.64 ± 1.39	35.83 ± 0.12	225.29 ± 0.04	194.04 ± 1.19	298.46 ± 0.04	04 ± 0.01
14	29.00 ± 0.01	291.94 ± 1.82	122.50 ± 0.12	425.28 ± 0.02	139.69 ± 0.41	218.40 ± 0.02	18 ± 0.02
15	25.39 ± 0.01	308.40 ± 2.26	187.83 ± 0.12	423.25 ± 0.07	127.43 ± 0.50	153.20 ± 0.06	03 ± 0.03
16	67.93 ± 0.05	222.10 ± 2.93	77.35 ± 0.14	295.12 ± 0.04	166.80 ± 0.88	243.20 ± 0.05	04 ± 0.02
17	40.98 ± 0.21	249.87 ± 1.11	135.21 ± 0.16	435.61 ± 0.03	133.31 ± 0.49	209.60 ± 0.03	01 ± 0.02

± Values represent standard deviations

Table 3.

ANOVA of linear, quadratic, and interaction terms on each response for ultrasound hydrated finger millet.

Variance source	MC (% db)		TPC (mg GAE/100 g)		TC (mg TAE/100 g)		Phytate (mg /100 g)		DG (%)		WBC of IS (%)		WS of IS (%)
	SS	P-value	SS	P-value	SS	P-value	SS	P-value	SS	P-value	SS	P-value	SS	P-value
Model	3780.22	0.0003	16256.62	< 0.0048	31588.45	< 0.0001	1.432E + 05	0.0001	5366.14	0.0055	47231.73	0.0009	2843.95	< 0.0001
A	1415.39	< 0.0001	332.19	0.2486	13957.86	< 0.0001	89200.32	< 0.0001	2237.81	0.0008	27849.18	< 0.0001	32.00	0.0185
B	738.43	0.0006	1824.90	< 0.0214	9518.55	< 0.0001	464.52	0.4111	1917.04	0.0013	4152.34	0.0109	10.13	0.1297
C	30.23	0.2691	75.65	0.5671	248.53	0.0373	405.70	0.4409	25.56	0.5709	34.24	0.7642	28.13	0.0243
AB	236.70	0.0121	6195.26	0.0010	627.25	0.0047	4407.63	0.0310	22.52	0.5943	94.67	0.6200	9.00	0.1496
AC	353.06	0.0046	777.60	0.0956	106.19	0.1375	742.29	0.3057	132.60	0.2179	106.61	0.5992	225.00	< 0.0001
BC	71.32	0.1078	1612.46	0.0276	450.29	0.0107	191.27	0.5924	40.77	0.4773	368.64	0.3401	240.25	< 0.0001
A2	794.90	0.0005	25.82	0.7360	2927.01	< 0.0001	45617.94	< 0.0001	497.02	0.0344	1478.49	0.0796	982.42	< 0.0001
B2	95.10	0.0708	5197.57	< 0.0016	1354.15	0.0005	1357.36	0.1788	440.19	0.0430	6728.84	0.0033	660.53	< 0.0001
C2	33.48	0.2469	70.61	0.5800	1719.51	0.0003	842.77	0.2775	2.53	0.8569	5033.03	0.0069	423.16	< 0.0001
Residual	146.86	–	1468.61	–	264.43	–	4255.81	–	506.39	–	2463.34	–	24.05	–
Lack of Fit	48.68	0.6178	218.70	0.8692	32.83	0.8988	1484.30	0.5928	30.50	0.96	746.21	0.6589	19.25	0.0695
Pure Error	98.18	–	1249.91	–	231.59	–	2771.51	–	475.90	–	1717.14	–	4.80	–
R2	0.96	–	0.92	–	0.99	–	0.97	–	0.91	–	0.95	–	0.99	–
Adjusted R2	0.92	–	0.81	–	0.98	–	0.93	–	0.80	–	0.89	–	0.98	–
Predicted R2	0.76	–	0.69	–	0.97	–	0.81	–	0.79	–	0.71	–	0.89	–
C.V. %	9.62	–	6.19	–	7.20	–	6.22	–	5.61	–	9.10	–	12.36	–
Adeq. Precision	13.54	–	10.39	–	32.36	–	16.12	–	9.87	–	11.37	–	23.42	–

*MC: Moisture content; TPC: Total phenolic content; TC: Tannin content; WBC: Water binding capacity of isolated starch; WS: Water solubility of isolated starch; DG: Degree of gelatinization

Table 4.

Regression coefficients for quality parameters of ultrasound hydrated finger millet.

Coefficients	Moisture content (%db)	Total phenolic content (mg GAE/100 g)	Tannin content (mg TAE/100 g)	Phytates(mg /100 g)	Degree of Gelatinisation (%)	Water binding capacity of starch (%)	Water solubility of starch (%)
b0	130.42	434.70	717.73	−57.40	177.87	491.58	−316.12
b1	−3.41**	−2.79	−9.16**	22.02**	−2.51**	−1.45**	4.99*
b2	−0.28*	−19.38*	−16.95**	−0.19	−0.99**	−12.05*	7.82
b3	−11.30	31.58*	−82.68*	29.41*	−1.05	−118.27	64.18
b4	−0.04*	0.20**	0.06*	−0.17	−0.01	0.02	−0.01
b5	0.31*	−0.46	−0.17	0.45	0.19	−0.17	−0.25
b6	0.28	−1.34*	0.71*	0.46	−0.21	−0.64	−0.52
b7	0.03*	0.01	0.07**	−0.26**	0.03*	0.05	−0.04
b8	0.05	0.35*	0.18*	0.18	0.10	0.40*	−0.13
b9	−1.25	+1.82	8.98*	−6.29	−0.35	15.37*	−4.46

b_{0 to} b₉ are the regression coefficients of developed model; ** Significant at P < 0.001; *Significant at P < 0.05

3.2. Effect of ultrasound-assisted hydration on quality characteristics of finger millet

As discussed earlier, various quality characteristics were determined for ultrasound hydrated finger millet and were compared with those of conventionally hydrated and raw finger millets. The effects of hydration and ultrasound-assisted hydration on these quality characteristics are discussed below.

3.2.1. Moisture content

The moisture content (MC) of the grain influences the textural properties and cooking time. It affects different quality characteristics of the product. The moisture content of the finger millet was significantly (P < 0.05) affected by the ultrasound amplitude and treatment time. The moisture content increased with an increase in amplitude and time (Table 2). Ultrasound exposure at higher amplitude for extended time disrupts and dislodge the cells creating microcavities and results in higher mass transfer by diffusion. This helps in improving hydration [6], [7], [15], [16], [17]. The higher values of R² (0.96), adj. R² (0.92) and lower value of CV (9.62) (Table 3) confirmed higher accuracy and lesser relative deviation of the data in the proposed model. Maximum hydration of the samples could be achieved with a moisture content of 76.65% (db) by ultrasound treatment for 30 min at 70% amplitude (Table 1, Table 2).

3.2.2. Total phenolic content

The total phenolic content (TPC) of the grain signifies antioxidant activity. The TPC value of the finger millet was significantly (P < 0.05) affected by the ultrasound amplitude and treatment time. It decreased with an increase in ultrasound amplitude and time (Table 2). At higher ultrasound amplitude and time with lower grain to water ratio, higher energy is generated, increasing cavitation in the cellular structure, thereby increasing solvent penetration [8]. The higher penetration of solvent may be responsible for the higher loss of phenolic compounds from the grain to the medium. The lower content of total phenolic compounds after ultrasound-assisted hydration may also be due to the reduction of galloyl moieties at lower grain to water ratio during the hydration process [14], [15], [19]. The effects of linear, quadratic, and interaction terms of the model are shown in Table 3. The higher values of R² (0.92) and adj. R² (0.81), and lower value of CV (6.19) (Table 3) confirmed higher accuracy and lesser relative deviation of the data in the proposed model. The lowest and highest total phenolic contents of the ultrasound hydrated finger millet were 190.98 mg GAE/100 g (70% amplitude, 20 min) and 308.89 mg GAE/100 g (30% amplitude, 10 min) (Table 1, Table 2).

3.2.3. Total tannin content

Tannin has a tendency of precipitating proteins in the food and, therefore, acts as an anti-nutritional factor [4]. It decreases the bioavailability of essential nutrients. The total tannin content in the millet sample was between 35.83 and 187.83 mg TAE/ 100 g (Table 2), which varied based on the moisture content. The higher amount of water in the sample significantly (P < 0.05) reduces the total tannin content. It may be due to the hydrolysis of ester linkage in the tannin during the homolysis of water into OH^– and H⁺ [20]. The tannin content could be decreased significantly (P < 0.001) with an increase in ultrasound amplitude and treatment time. Ultrasound helps convert hydrolyzable tannic acid into gallic acid, thereby reducing the total tannin content of the sample. It also induces leaching out of the condensed tannin from the sample, thereby decreasing total tannin content [20]. The effects of linear, quadratic, and interaction terms of the model are shown in Table 3. Higher values of R² (0.99) and adj. R² (0.98) and lower value of CV (7.20) (Table 3) confirmed higher accuracy and lesser relative deviation of the data in the proposed model.

3.2.4. Phytate content

The phytate content in the millet samples was in the range of 185.32 to 486.2 mg/ 100 g of the sample. Similar kinds of observations have been reported by Udeh et al. [21]. The phytates chelate the minerals available in the millets and thus act as antinutrients [4]. Ultrasound treatment of the samples could reduce phytate content significantly. The effects of linear, quadratic, and interaction terms on the phytate content of finger millet are shown in Table 3. The phytate contents decreased with an increase in ultrasound amplitude and soaking time. At higher ultrasound amplitude, heat is generated in the sample, thereby increasing the temperature. The heat induces the chemical degradation of phytate to lower inositol phosphate, thereby decreasing the phytate content of the sample [4], [20]. Higher values of R² (0.97) and adj. R² (0.93) with lower CV (6.22) confirms higher accuracy and lesser relative deviation.

3.2.5. Degree of gelatinization

Millets have a lower gelatinization temperature of 17 °C [21]. Therefore even at normal ambient conditions degree of gelatinization is higher. The degree of gelatinization of hydrated finger millet flour was in the range of 125.26 to 194.04% (Table 2). The degree of gelatinization increased significantly (P < 0.05) with increased ultrasound amplitude and treatment time. The effects of linear, quadratic, and interaction terms are shown in Table 3. The further increase in the degree of gelatinization may be due to the heat generation during ultrasound treatment. Ultrasound treatment increases the agglomeration of starch molecules and helps in higher gelatinization [8], [22], [23], [24], [25]. Higher R² (0.91) and adj. R² (0.80) with lower CV (5.61) confirmed higher accuracy and lesser relative deviation in the results.

3.2.6. Water-binding capacity of isolated starch

Water-binding capacity of the starch isolated from the ultrasound-treated hydrated finger millet was determined. The water-binding capacity of the isolated starch ranged from 128.35 to 298.46% (Table 2). A similar range of values has been reported for brown rice (Park and Han, 2016). The water-binding capacity increased significantly with the increase in ultrasound amplitude and treatment time. Higher water-binding capacity may be attributed to the formation of micro-jets and higher shearing forces under ultrasound that breaks the starch granules and facilitates the higher water penetration and binding inside the pores [8], [9], [22], [23]. The effects of linear, quadratic, and interaction terms of independent parameters on the water-binding capacity of isolated starch are shown in Table 3. Higher R² (0.95) and adj. R² (0.93) with lower CV (9.10) confirmed higher accuracy and lesser relative deviation in the results.

3.2.7. Water solubility of isolated starch

Water solubility of the isolated starch was in the range of 0–34%. With the increase of ultrasound amplitude from 30 to 50% while treatment, water solubility increased. The increased solubility may be due to the higher affinity of degraded starch molecules towards water molecules [23]. Beyond 50% ultrasound amplitude, water solubility decreased. It may be due to the higher mechanical degradation of starch at higher amplitude, consequently solubilizing and leaching out some amylose and amylopectin molecules from the millet [5]. The significant effect of linear, quadratic, and interaction terms in terms of a sum of squares and p-value are shown in table 3. Higher accuracy of data was observed with higher R² (0.99) and adj. R² (0.98) values. There was a lesser relative deviation of results with a lower CV (12.36) value.

3.3. Process optimization and validation

Ultrasound-assisted hydration parameters were optimized for removal of antinutrients, such as tannin and phytates, and retention of phenolic compounds. Also, improvement of isolated starch suitability for different value addition process was necessary. The conditions for process optimization was based on the desirability function. The optimum hydration could be achieved for finger millet at 66% of ultrasound amplitude, 26 min soaking time, and 1:3 grain to water ratio. The optimized values for the factors and responses after model prediction are shown in Table 5. The predicted optimized values were validated by experiments in the laboratory with replications. The actual experimental values were close to the predicted values for the quality characteristics, confirming the correctness of the developed regression model.

Table 5.

Optimized condition for ultrasound-assisted hydration of finger millet.

Variables	Target	Model optimized values	Optimum Experimental values**
Ultrasound amplitude (%)	In range	65.73	66
Treatment time (min)	In range	26.16	26
Grain: water	minimize	1:3	1:3
Applied Energy (J)	In range	–	58,756
Calorimetric energy (J)	In range	–	10,032
MC (% db)	In range	52.31	51
TPC (mg GAE/100 g)	maximum	264.64	255
TC (mg TAE/100 g)	minimum	42.24	56
Phytates (mg PA/100 g)	minimum	255.28	219
WBC (%)	maximum	298.46	273
WS (%)	maximum	15.93	16
DG (%)	minimum	160.08	159
Desirability	–	0.710	–

*MC: Moisture content; TPC: Total phenolic content; TC: Tannin content; WBC: Water binding capacity of isolated starch; WS: Water solubility of isolated starch; DG: Degree of gelatinization.

**Values given as nearest whole numbers to accommodate experimental errors during ultrasonication.

3.4. Comparison of raw, conventionally hydrated, and optimized ultrasound hydrated finger millet

The quality characteristics data (Table 6) for the raw, conventionally hydrated, and ultrasound hydrated finger millets were compared by the Duncan test in SPSS software (Ver. 22, USA). The moisture content of conventionally hydrated finger millet was higher than the raw and the optimized ultrasound hydrated samples. Higher temperature (70 °C) during conventional hydration increases the expansion and contraction of pores within the grain, thereby decreasing diffusion resistance. It increases the diffusion process and results in higher moisture content in the millet grains [18]. Though the moisture content of the optimized ultrasound hydrated finger millet was close to that of conventionally hydrated samples, ultrasound treatment alone is not sufficient to achieve maximum hydration [7]. Significant differences (P < 0.05) were observed for the various quality characteristics of optimized ultrasound hydrated and conventionally hydrated finger millets. Higher total phenolic content was observed in the ultrasound hydrated finger millet compared to conventionally hydrated and raw finger millet samples. It may be due to the conversion of tannic acid to gallic acid during the sonication process [20]. The tannin content of the conventionally hydrated finger millet was lower than that of the ultrasound hydrated finger millet. It may be due to higher leaching of condensed tannin (red color compound) into the water in conventional hydration. Also, the hydrolyzable tannin in the finger millet is converted into gallic acid at high temperature, thereby lowering the overall tannin content of the conventionally hydrated finger millet [26]. The phytate content of the ultrasound hydrated finger millet was lesser than that of the conventionally hydrated finger millet [26], [27], [28], [29]. During ultrasound treatment, the impact of bubbles in the microcavities increases significantly. The process results in the free radical generation and phytase activation. The whole process reduces the phytate content of the ultrasound hydrated samples [22]. No significant difference (P > 0.05) could be observed in the water-binding capacities of isolated starches from the ultrasound hydrated and conventionally hydrated finger millets. The water solubility of conventionally hydrated finger millet starch was lower than the raw finger millet starch. It may be because of the higher temperature during conventional hydration, strengthening the bonds between amylose and amylopectin molecules, thereby reducing solubility [30]. The water solubility of the isolated starch from the ultrasound hydrated finger millet was higher than that of the conventionally hydrated sample. It may be due to the increased binding of water molecules with the free hydroxyl groups of amylose and amylopectin generated by ultrasound [23].

Table 6.

Factors and quality of raw, conventionally hydrated, and ultrasound hydrated finger millets.

Factors	Raw finger millet	Conventionally hydrated finger millet	Ultrasound hydrated finger millet***
Grain: water	–	1:3	1:3
Hydration time (min)	–	120	26
Ultrasound amplitude (%)	–	–	66
MC (% db)	12.35 ± 0.02c	65.49 ± 0.04a	51b
TPC (mg GAE/100 g)	223.45 ± 0.08c	253.30 ± 0.44a	255b
TC (mg TAE/100 g)	151.87 ± 0.08c	49.15 ± 0.09a	56b
Phytates (mg PA/100 g)	664.55 ± 7.15c	391.02 ± 1.00a	219b
DG (%)	0.00c	137.24 ± 0.18a	159b
WBC (%)	193.56 ± 0.89b	272.01 ± 0.41a	273a
WS (%)	11.42 ± 0.14c	1.03 ± 0.05a	16b

*Means followed by different letters in the same row are significantly different by Duncan’s test (P < 0.05)

** MC: Moisture content; TPC: Total phenolic content; TC: Tannin content; WBC: Water binding capacity of isolated starch; WS: Water solubility of isolated starch; DG: Degree of gelatinization.

*** Values given as nearest whole numbers to accommodate experimental errors during ultrasonication.

3.5. Surface morphology of isolated starch

The surface morphologies of isolated starches were studied using FE-SEM (Fig. 2). It could be observed that hydration and ultrasound reduce the particle size and alter polymeric structures of finger millet starch. The average particle sizes of isolated starches from the raw, conventionally hydrated, and ultrasound hydrated finger millet samples were 10.58 ± 0.98, 6.45 ± 0.54, and 5.27 ± 0.51 µm, respectively. The hydration process reduces the particle size of isolated starch. However, the size distribution of starch granules was heterogeneous, as seen in the conventionally hydrated finger millet starch isolates (Fig. 2b). The conventional hydration process ruptures and damages the starch granules that may affect the overall quality of the isolated starch. During the hydration process, the ultrasound application further reduced the particle size and resulted in homogenous grain size distribution in the starch isolates (Fig. 2c). It also reduced the rupture and damage of starch granules. The starch granules had smaller sizes and polyhedron shapes with smooth surface morphology that improves the gelatinization, water binding capacity, and water solubility of isolated starch (Table 6). During ultrasound treatment, the bubbles inside the starch microcavities generate shear forces that break the starch granules mechanically and smoothen the granule surface. It also results in a high-pressure gradient inside the starch microcavities that degrades the starch polymeric chain [23].

Fig. 2.

FE-SEM images of finger millet starch (a) isolated from raw sample (b) isolated from conventionally hydrated sample, and (c) isolated from ultrasound hydrated finger millet sample.

3.6. XRD analysis

The X-ray diffraction pattern was analyzed for three different starch samples isolated from raw, conventionally hydrated, and ultrasound hydrated finger millet (Fig. 3). The higher intensity peaks were obtained at 14.3, 17.7, 20.3, 26.9, and 27.5° for all the starch samples. The peaks obtained for starch isolates from raw and conventionally hydrated finger millet overlapped, showing similar relative crystallinity (15.37% & 15.84%). The lower relative crystallinity of the starch isolates from conventionally hydrated samples may be due to the asymmetric molecular arrangement of the starch particle only in the amorphous region. The relative crystallinity of ultrasound hydrated finger millet starch was 24.7%. The systematically arranged linear polymeric chains of starch in the amorphous region are disrupted during ultrasound treatment, thereby increasing the relative crystallinity of starch isolates [9], [25]. The result is reconcilable with FE-SEM (Fig. 2c), showing homogenous starch granule structure after ultrasonication. The raw finger millet starch was A-type (diameter > 9.9 µm) [31], whereas the conventionally hydrated and ultrasound hydrated finger millet starches were B-type (diameter < 9.9 µm).

Fig. 3.

X-ray diffraction pattern of raw finger millet starch (RFMS), conventionally hydrated finger millet starch (CFMS), and ultrasound hydrated finger millet starch (USFMS).

4. Conclusions

The effects of ultrasound amplitude, treatment time, and grain to water ratio on various quality characteristics such as moisture content, total phenolic content, tannin content, phytates, degree of gelatinization, water binding capacity, and water solubility of isolated starches from all the finger millet samples were examined and compared. Better quality could be achieved at optimized ultrasound-assisted hydration conditions at 66% of ultrasound amplitude, 26 min of treatment time, and 1:3 of grain to water ratio. Application of ultrasound during hydration decreased hydration time and antinutrients of the sample compared to the raw and conventionally hydrated finger millets. The process also improved the structural morphology, water binding capacity, and water solubility of the isolated starch. The ultrasound-assisted hydration is a better alternative for finger millet processing with reduced antinutrients and improved starch quality. The process could be scaled up for industrial use.

CRediT authorship contribution statement

Shweta Yadav: Methodology, Investigation, Formal analysis, Writing - original draft. Sabyasachi Mishra: Conceptualization, Investigation, Writing - review & editing, Visualization, Supervision, Resources. Rama Chandra Pradhan: Conceptualization, Investigation, Writing - review & editing, Visualization, Supervision, Resources.

Declaration of Competing Interest

The Authors do not have any conflict of Interest.