A novel strategy for producing nano-particles from date seeds and enhancing their phenolic content and antioxidant properties using ultrasound-assisted extraction: A multivariate based optimization study.

Date seeds from the date palm fruit are considered as a waste and they are known to contain several bioactive compounds. Producing nanoparticles from the date seeds can enhances their effectiveness and their utilization as novel functional food ingredients. In this study, date seed nanoparticles (DSNPs) synthesized using acid (HCl) hydrolysis method (HCl concentration of 38% and hydrolysis time of 4 days) was found to have particle size between 50 and 150 nm. The obtained DSNPs were characterized by measuring particle size and particle charge (Zetasizer), morphology using scanning electron microscope (SEM), and determination of the functional groups using fourier-transform infrared spectroscopy (FTIR). DSNPs were further treated with green extraction technology [ultrasound-assisted extraction (UAE)] using water-based and methanol-based solvent for optimizing the extraction of the bioactive compounds by implementing response surface methodology (RSM). The UAE of DSNPs were analysed for set of responses including total phenolic content (TPC), total flavonoid content (TFC), 1,1-diphenyl-2-picrlthydrazyl (DPPH) radical scavenging activity, ferric ion reducing antioxidant power (FRAP), and 2,2'-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) radical scavenging activity. Three-factor and four-factor Box-Behnken design (BBD) of three models (Synthesis of DSNPs, UAE with water, and UAE with methanol) was performed. The results showed that in UAE of DSNPs using water-based solvent, the key independent factors effecting the TPC and TFC and antioxidant activities were S:L ratio (40:1 mg/ml) and treatment time (9 min). Whereas the methanol-based UAE of DSNPs was mostly affected by US amplitude/power (90%) and methanol concentration (80%). All models were further optimized using response optimizer in Minitab and the generated predicted values were very comparable to the actual obtained results which confirm the significance and validity of all RSM models used. The phenolic compounds identified from DSNPs consisted mainly of 3,4-Dihydroxy benzoic acid, ferulic acid, and p-coumaric acid. The present study demonstrated a successful method for synthesising DSNPs as well as documented the optimum UAE conditions to maximize the extraction of polyphenolic compounds from DSNPs and enhancing their antioxidant activities to be used in food application.

Introduction

The date palm (Phoenix dactylifera L.) is mainly grown in arid and semi-arid countries and has a significant role and status especially in terms of economy and as a source of abundant food resource. The date palm fruit and its byproducts are known for their high nutritional and medicinal values such as antioxidant, anticancer, antimicrobial, and hypoglycemic activities [48], [14], [17]. Recently, the cultivation of date palm has remarkably increased due to its ability to grow at high temperatures and low irrigation requirements that suits the uncertain future regarding food security and availability in the region with such prevailing climatic conditions. Previous study has reported a noticeable expansion in the production of dates around 600,000 metric tons from 2017 to 2019 [46]. The above-mentioned exponential production of date fruit and its processing has resulted in the production of huge quantities of date wastes including seeds, leaves, and spikelets which if not managed or utilized properly will lead to environmental problems and increase processing costs [28]. Apart from being consumed directly, date fruit is also processed into different products, such as date syrup, paste, confectionery, sweets, and candies which resulted into production of huge quantity of date seeds which are considered as nutrient rich bioresources seeking efficient valorization. Hence, novel processing techniques that can be implemented to valorize these date palm wastes will open diverse possibilities of utilizing these nutrient and bioactive rich wastes in food and other relevant applications.

Adapting various extraction methods to extract phenolic compounds from the seeds of a food material have been investigated such as pulsed electric field [18], [42], [44], super critical fluid [25], [50], [41], microwave-assisted extraction (MAE) [6], [7], [51], and ultrasound-assisted extraction (UAE) [4], [24], [32]. However, previous studies have suggested that UAE is efficient and reliable in extracting phenolic compounds from the various seeds of different fruits [11], [9], [40], [43]. However, a new direction that aims to produce nanoparticles from date seeds should be investigated and further exploration of bioactive properties of the synthesized date seed nanoparticles (DSNPs) should be thoroughly studied. Applications of nanotechnology have gained wide application in food industry with increasing utilization of nanoparticle in fields like food processing, food packaging, functional food development, food safety, detection of foodborne pathogens, and shelf-life extension of food and/or food products [56]. The main improved properties exhibited by having particles in the nanoscale include enhanced solubility, bioavailability, and functional properties compared to regular sized and coarse bulk materials [35], [33]. The reduced particle size and enhanced physiochemical properties could facilitate its incorporation in food as a functional ingredient. Moreover, the production of new functional food ingredients in the form of nanoparticles rich in bioactive compounds from natural sources which are considered as waste is very desirable outcome that can grab the attention of food processor and have a good share in the food processing market [38], [52], [54]. Till date, date seeds have been mainly subjected to extraction of bioactive compounds using conventional solvents and techniques [3], [49], [5]. There is an immense scope for implementing some novel processing techniques and approaches to valorise the date seeds effectively.

Therefore, the aim of this novel study is to synthesize date seed nanoparticles (DSNPs) in a powder form that have enhanced solubility in water and exhibits high phenolic content and antioxidant properties. Further, the DSNPs were subjected to extraction by water-based solvent and methanol-based solvent using green extraction technique (ultrasound-assisted extraction (UAE)). In order to optimize the synthesis of DSNPs and the extraction of phenolic compounds from DSNPs, response surface methodology (RSM) was used as a statistical tool to propose a valid processing condition for date seed valorization.

Materials and methods

Materials

Date seeds (khalas variety) were collected at mature stage (Tamr) of the date fruit which were obtained from the Al FOAH Date company in Al Ain, Abu Dhabi. The chemicals including 1,1-diphenyl-2-picrlthydrazyl (DPPH), 2,4,6-Tris (2-pyridyl)-s-triazine (TPTZ), 2,2′-azino-dis [3-ethylbenzthiazoline sulfonate (ABTS), 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox), gallic acid, quercetin, and Folin–Ciocalteu phenol reagent were procured from Sigma–Aldrich Chemical Co. (St. Louis, MO). The remaining ACS grade chemicals and solvents were obtained from Fisher Scientific (Nepean, ON).

Sample preparation

Date seeds were first milled using an industrial seed grinder at Al FOAH experimental farm in Al Ain, UAE followed by further grinding using an electric grinder (Nikai, India) to attain a fine powder. Date seed powder (DSP) produced were sieved through a 250 µm mesh filter to obtain a uniform particle size. The sieved powder was stored at −20 °C until used for further analysis.

Date seed nanoparticles (DSNPs) preparation

For the synthesis of DSNPs, 2.5 g, 5 g, and 10 g of DSP were weighted and mixed with 100 ml 38%, 19%, and 9.5% HCL solution in a 250 ml Scott bottle. The bottle containing the mixture was placed on a magnetic hotplate stirrer adjusted to temperatures of 30 °C, 40 °C, and 50 °C at a constant speed of 1200 rpm for 3 days, 4 days, and 5 days. After hydrolysis, the mixtures were centrifuged at 10,528 xg for 20 min, then the supernatant was collected and dried on a hotplate at 50 °C in a laminar flow hood for 24 hr to obtain the DSNPs in powder form. The dried DSNPs were washed using Deionized (DI) water several times and pH was adjusted to 6 then dried again and stored in polyethylene Ziploc bags at −20 °C until used for further analysis.

Ultrasound-assisted extraction (UAE) of DSNPs

To extract phenolic compounds from DSNPs, 2 g, 3 g, and 4 g of DSNPs were mixed with 100 ml water-based solvent, while 4 g were dissolved in 100 ml methanol-based solvent. The mixture was subjected to ultrasound-assisted extraction using ultrasound system (model SFX, Branson, Mexico) and temperature was maintained between 37 and 40 °C by placing the samples in cold water bath during US treatment. UAE of DSNPs was performed at three different amplitudes including 70%, 80%, and 90% and treatment time of 3, 6, and 9 min. All ultrasonicated samples were dried at 50 °C, then samples were stored at −20 °C until further use. UAE of DSNPs was carried out in triplicates on three different batches of DSNPs.

Response surface methodology

Optimization of DSP hydrolysis

RSM with four factors, with each factor having three levels (43) was applied to an experimental design based on Box–Behnken design (BBD) and was used to optimize the hydrolysis conditions of DSP to produce DSNPs. In the BBD, 27 experiments with three central points were included in the design. The hydrolysis time (X1) (3 days, 4 days, and 5 days), hydrolysis temperature (X2) (30, 40, and 50 °C), HCl concentration (X3) (9.75%, 19.5%, and 38%), and solid: liquid (S:L) ratio (X4) (0.025 g/ml, 0.05 g/ml, and 0.10 g/ml) were selected as the independent factors, and the particle size and TPC as response factors. The randomized design table of DSP hydrolysis factors with their different levels and responses are presented in Table 1. The obtained results of particle size and TPC responses of triplicate measurements were analysed by multiple linear regression analysis to get the regression coefficients following a second-order polynomial model.

Table 1

Box–Behnken experimental design for optimization of acid hydrolysis process with particle size reduction, surface charge, and TPC of DSP as response factors.

Runs	Independent variables				Dependent Variables
	Hydrolysis time X1	Temperature X2	HCl conc. X3	S: L X4	Particle size (nm)	TPC (mg GAE/g DSNPs)	Surface charge (mV)
1	1	0	0	−1	642.31 ± 12.88de	13.49 ± 0.05 k	−2.76 ± 0.35f
2	0	1	−1	0	3267.62 ± 32.1c	12.27 ± 0.08m	−0.61 ± 0.12a
3	−1	0	1	0	450.88 ± 13.75ef	15.74 ± 0.12g	−4.15 ± 0.41k
4	−1	0	−1	0	4778.49 ± 41.6a	7.85 ± 0.13r	−0.53 ± 0.14a
5	0	1	0	1	763.37 ± 20.7de	18.41 ± 0.07d	−2.48 ± 0.36e
6	0	0	0	0	675.79 ± 20.3de	13.52 ± 0.11k	−2.57 ± 0.11e
7	0	0	0	0	683.55 ± 11.26de	13.73 ± 0.04jk	−2.72 ± 0.32f
8	0	−1	1	0	122.57 ± 7.82fg	15.93 ± 0.16g	−4.89 ± 0.45l
9	0	0	−1	1	4283.25 ± 29.4b	12.04 ± 0.04mn	−0.68 ± 0.08a,b
10	0	−1	0	−1	778.65 ± 23.1de	10.26 ± 0.09o	−2.51 ± 0.25e
11	0	1	1	0	85.42 ± 3.59g	20.01 ± 0.10b	−5.47 ± 0.44m,n
12	0	−1	0	1	831.77 ± 23.7d	14.38 ± 0.15hi	−3.84 ± 0.19i,j
13	1	0	1	0	76.92 ± 5.09g	19.44 ± 0.19c	−6.21 ± 0.23o
14	0	0	1	1	97.47 ± 5.34g	20.53 ± 0.08a	−5.68 ± 0.47n
15	0	0	1	−1	88.32 ± 6.16g	16.78 ± 0.13f	−5.32 ± 0.52m
16	−1	0	0	−1	873.44 ± 53.9d	9.51 ± 0.21p	−2.14 ± 0.34d
17	0	0	−1	−1	3589.21 ± 93.8c	8.32 ± 0.14q	−0.73 ± 0.10b
18	0	1	0	−1	658.66 ± 38.4de	14.68 ± 0.11h	−3.05 ± 0.54g
19	1	1	0	0	622.35 ± 44.6de	17.59 ± 0.06e	−3.41 ± 0.21h
20	0	−1	−1	0	4419.44 ± 320.1b	8.26 ± 0.17qr	−0.55 ± 0.09a
21	−1	0	0	1	898.75 ± 44.2d	13.76 ± 0.23jk	−2.77 ± 0.13f
22	1	0	−1	0	3442.55 ± 430.3c	11.62 ± 0.15n	−1.29 ± 0.06c
23	−1	−1	0	0	901.89 ± 67.3d	9.89 ± 0.08op	−2.30 ± 0.39d
24	1	−1	0	0	704.72 ± 41.4de	14.07 ± 0.16ij	−3.96 ± 0.33j
25	1	0	0	1	670.11 ± 16.74de	18.21 ± 0.05d	−3.87 ± 0.22j
26	0	0	0	0	680.88 ± 84.9de	13.95 ± 0.22j	−3.65 ± 0.49i
27	−1	1	0	0	858.48 ± 48.1d	12.72 ± 0.18l	−2.59 ± 0.27e

X1: hydrolysis time (days) (−1 = 3 days, 0 = 4 days, and 1 = 5 days); X2: hydrolysis temperature (°C) (−1 = 30 °C, 0 = 40 °C, and 1 = 50 °C); X3: HCl concentration (%) (−1 = 9.5%, 0 = 19%, and 1 = 38%); X4: solid: liquid ratio (mg/ml) (−1 = 20 mg/ml, 0 = 30 mg/ml, and 1 = 40 mg/ml).

Where Y is the predicted responses (particle size and TPC), Xi and Ji are the independent factors, (hydrolysis time, hydrolysis temperature, HCl concentration, and S:L ratio), and β0, βi, βii, and βij are model constant, linear, squared, and 2-way interaction effects.

Optimization of Ultrasound-assisted extraction of phenolic compounds from DSNPs

For UAE of bioactive compounds from DSNPs, RSM design with three factors, with each factor having three levels (33) was developed based on Box–Behnken design (BBD) and implemented to optimize the extraction of phenolic compounds from DSNPs using water-based and methanol-based solvents. In the first BBD, 15 experiments with three central points were carried out. The ultrasound Amplitude (X1) (70%, 80%, and 90%), treatment time (X2) (3, 6 and 9 min), and solid: liquid (S: L) (X3) (20 mg/ml, 30 mg/ml, and 40 mg/ml) with water as solvent were selected as the main independent factors, and the TPC, TFC, DPPH radical scavenging activity, FRAP, and ABTS radical scavenging activity as response factors. For the second BBD, 15 experiments with three central points were conducted. The ultrasound Amplitude (X1) (70%, 80%, and 90%), treatment time (X2) (3, 6 and 9 min), and methanol concentration (X3) (60%, 70%, and 80%) were chosen as the main independent factors, and the TPC, total flavonoids, DPPH radical scavenging activity, FRAP, and ABTS radical scavenging activity as response factors. The complete randomized design tables of both DSNPs extraction (water-based and methanol-based solvents) factors with their levels and response factors are presented in Table 2 and Table 3, respectively. The obtained results of TPC, TFC, DPPH, FRAP, ABTS responses of triplicate measurements for both designs were analysed by multiple linear regression analysis to get the regression coefficients following a second-order polynomial model:Where Y is the predicted responses (TPC, TFC, DPPH, FRAP, and ABTS), Xi and Ji are the independent factors, (US amplitude, treatment time, and S:L ratio/methanol concentration), and β0, βi, βii, and βij are model constant, linear, squared, and 2-way interaction effects.

Table 2

Box–Behnken experimental design for the optimization of UAE process of DSNPs using water-based solvent with TPC and TFC and antioxidant activities as response factors.

Runs	Amplitude (%)X1	Time (min)X2	S:L ratio (mg/ml)X3	TPC (mg GAE/g DSNPs)	TFC (mg QE/g DSNPs)	DPPH (mmol trolox/g DSNPs)	FRAP (mmol trolox/g DSNPs)	ABTS (mmol trolox/g DSNPs)
1	90	9	30:1	26.79 ± 0.11b	2.61 ± 0.09b	21.52 ± 0.09b	35.28 ± 0.16b	12.54 ± 0.04b
2	80	6	30:1	24.32 ± 0.08d	2.29 ± 0.12bcd	18.73 ± 0.13ef	32.55 ± 0.09d	8.97 ± 0.03f
3	90	3	30:1	20.81 ± 0.15g	1.81 ± 0.08efg	16.46 ± 0.06i	33.50 ± 0.12c	10.21 ± 0.08e
4	70	6	40:1	26.15 ± 0.11c	2.54 ± 0.19bc	20.39 ± 0.16d	33.42 ± 0.22c	10.57 ± 0.11d
5	70	6	20:1	19.95 ± 0.12h	1.72 ± 0.09fg	15.40 ± 0.08j	29.01 ± 0.08h	6.16 ± 0.05j
6	80	9	20:1	18.63 ± 0.09j	1.54 ± 0.11g	13.77 ± 0.12l	32.03 ± 0.11e	6.27 ± 0.12j
7	70	9	30:1	22.07 ± 0.01f	1.99 ± 0.18def	17.18 ± 0.14h	31.93 ± 0.08e	7.94 ± 0.08h
8	70	3	30:1	19.04 ± 0.17i	1.59 ± 0.02g	14.24 ± 0.21k	29.78 ± 0.13g	6.71 ± 0.05i
9	80	9	40:1	26.86 ± 0.09b	2.62 ± 0.03b	21.01 ± 0.07c	35.43 ± 0.16b	11.09 ± 0.08c
10	80	6	30:1	23.83 ± 0.18e	2.21 ± 19cd	18.51 ± 0.13fg	33.30 ± 0.09c	8.76 ± 0.06f
11	80	6	30:1	23.47 ± 0.15e	2.16 ± 0.11de	18.24 ± 0.12g	33.32 ± 0.10c	8.33 ± 0.14g
12	90	6	40:1	29.48 ± 0.10a	2.98 ± 0.12a	23.77 ± 0.13a	36.91 ± 0.14a	14.93 ± 0.10a
13	90	6	20:1	18.36 ± 0.16j	1.51 ± 0.01g	15.48 ± 0.07j	31.81 ± 0.11e	11.21 ± 0.04c
14	80	3	20:1	15.63 ± 0.13k	1.11 ± 0.10h	11.04 ± 0.14m	30.25 ± 0.14f	5.12 ± 0.02k
15	80	3	40:1	23.47 ± 0.12e	2.17 ± 0.14d	19.02 ± 0.08e	33.28 ± 0.24c	8.24 ± 0.06g

X1: ultrasound amplitude (%) (−1 = 70%, 0 = 80%, and 1 = 90%); X2: treatment time (min) (−1 = 3 min, 0 = 6 min, and 1 = 9 min); X3 solid: liquid ratio (mg/ml) (−1 = 20 mg/ml, 0 = 30 mg/ml, and 1 = 40 mg/ml).

Table 3

Box–Behnken experimental design for the optimization of UAE process for DSNPs using Methanol-based solvent with TPC and TFC and antioxidant activities as response factors.

Runs	Amplitude (%) X1	Time (min) X2	Methanol conc. (%) X3	TPC (mg GAE/g DSNPs)	TFC (mg QE/g DSNPs)	DPPH (mmol trolox/g DSNPs)	FRAP (mmol trolox/g DSNPs)	ABTS (mmol trolox/g DSNPs)
1	80	6	70	55.75 ± 0.06d	4.63 ± 0.03d	49.43 ± 0.12g	59.89 ± 0.22e	20.31 ± 0.15ef
2	80	9	60	51.52 ± 0.10g	3.56 ± 0.06f	47.26 ± 0.09h	57.64 ± 0.19g	18.47 ± 0.12h
3	80	9	80	57.22 ± 0.19c	5.47 ± 0.09c	54.75 ± 0.13c	65.25 ± 0.16c	24.68 ± 0.16b
4	70	6	60	46.51 ± 0.27i	2.31 ± 0.11i	33.21 ± 0.07m	42.73 ± 0.14l	15.04 ± 0.08k
5	70	6	80	52.71 ± 0.42ef	3.34 ± 0.07fg	41.64 ± 0.11j	50.84 ± 0.09i	20.56 ± 0.13e
6	70	3	70	46.95 ± 0.02hi	2.54 ± 0.04hi	34.91 ± 0.06l	44.35 ± 0.17k	16.24 ± 0.21j
7	80	3	80	52.98 ± 0.29e	3.42 ± 0.12fg	51.48 ± 0.08d	61.07 ± 0.14d	21.72 ± 0.19d
8	80	6	70	55.64 ± 0.47d	3.87 ± 0.05e	50.06 ± 0.13f	59.67 ± 0.24e	19.54 ± 0.14g
9	90	6	70	57.12 ± 0.19c	3.89 ± 0.09e	49.37 ± 0.16g	58.75 ± 0.16f	18.43 ± 0.25hi
10	70	9	70	51.92 ± 0.48fg	3.21 ± 0.13g	39.14 ± 0.12k	48.49 ± 0.21j	17.98 ± 0.15i
11	80	6	70	55.69 ± 0.26d	4.59 ± 0.08d	50.23 ± 0.07ef	59.74 ± 0.11e	19.87 ± 0.11fg
12	80	3	60	47.69 ± 0.49h	2.64 ± 0.03h	43.54 ± 0.09i	52.97 ± 0.18h	16.38 ± 0.18j
13	90	9	70	61.67 ± 0.09b	6.37 ± 0.11b	55.32 ± 0.13b	65.86 ± 0.13b	24.07 ± 0.10c
14	90	3	70	55.91 ± 0.66d	4.82 ± 0.06d	50.53 ± 0.11e	61.04 ± 0.25d	22.15 ± 0.13d
15	90	6	80	63.64 ± 0.22a	6.82 ± 0.040a	56.78 ± 0.04a	67.31 ± 0.07a	25.42 ± 0.06a

X1: ultrasound amplitude (%) (−1 = 70%, 0 = 80%, and 1 = 90%); X2: treatment time (min) (−1 = 3 min, 0 = 6 min, and 1 = 9 min); X3 methanol concentration (%) (−1 = 60%, 0 = 70%, and 1 = 80%).

Characterization of synthesized DSNPs

The particle size and surface charge of the synthesized DSNPs were analysed through Zetasizer, Nano series, HT Laser, ZEN3600 (Malvern Instrument, UK). Regarding the morphology of synthesized DSNPs was determined using analytical scanning electron microscope (SEM) JSM-6010PLUS/LA (JEOL, Japan). For determining the functional groups of DSP (unhydrolyzed), HCl hydrolysed DSP for 4 days, and 4 days HCl hydrolysed DSP treated for 9 min with ultrasound (90% amplitude) samples, FTIR spectrophotometer (PerkinElmer Spectrum Two, Serial no.103146, UK) in the range of 400 to 4000 cm−1 at a resolution of 8 cm−1 was used. The comparison of FTIR spectrum of these samples were done to provide insight of any changes in the structural and functional groups occurring in samples after hydrolysis and treatment using ultrasound.

Total phenolic content (TPC)

The TPC of hydrolysed DSP and DSNPs was determined following the Folin–Ciocalteu method according to Olatunde et al., [39] with few modifications. The TPC was expressed as mg gallic acid equivalent (GAE) per gram of DSNPs using the standard curve equation (y = 0.007x + 0.0217, R2 = 0.99). All samples were analysed in triplicate. For details on the methodology of TPC, please refer to supplementary material section S-1.1.

Total flavonoid content (TFC)

The TFC of DSNPs was determined following the method described by Barakat et al., [8]. The TFC was expressed as mg Quercetin equivalents per g of DSNPs. All samples were analysed in triplicate. For details on the methodology of TFC, please refer to supplementary material section S-1.2.

Antioxidant activity of DSNPs

DPPH radical scavenging activity assay

Determination of 1,1-diphenyl-2-picrlthydrazyl radical scavenging activity (DPPH) assay of DSNPs was done following the method described by Maqsood et al. [30] and Abdul-Hamid et al., [1] with few modifications. The obtained results were expressed as mg trolox equivalents per gram of DSNPs (mmol TE/g DSNPs). All samples were analysed in triplicate. For details on the methodology of DPPH radical scavenging activity assay, please refer to supplementary material section S-1.3.

FRAP assay

The ferric ion reducing antioxidant power (FRAP) of DSNPS was determined following the method described by Hinkaew et al., [20], with some modifications. The obtained values were expressed as mmol TE/g DSNPs. All samples were analysed in triplicate. For details on the methodology of FRAP assay, please refer to supplementary material section S-1.4.

ABTS radical scavenging activity assay

The antioxidant capacity against 2,2′-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) was determined as described by Maqsood and Benjakul [29] and as modified by Yang et al., [53]. Obtained ABTS results were expressed as mmol TE/g DSNPs and all samples were analysed in triplicate. For details on the methodology of ABTS radical scavenging activity assay, please refer to supplementary material section S-1.3.

Identification and quantification of DSNPs phenolic compounds

DSNPs samples with highest TPC in each of the 3 models were analysed for identification of phenolic compounds. The phenolic compounds present in DSNPs were analysed using a Thermo Scientific Dionex Ultimate 3000 UHPLC system equipped with a quaternary Series RS pump and a Thermo Scientific Dionex Ultimate 3000 Series TCC-3000RS column compartment with a Thermo Fisher Scientific Ultimate 3000 Series WPS-3000RS autosampler and a rapid separations PDA detector controlled by Chromeleon 7.2 Software (Thermo Fisher Scientific, Waltham, MA, USA and Dionex Softron GmbH subsidiary of Thermo Fisher Scientific, Bremen, Germany). DSNPs samples (1 ml) was filtered through a 0.45 μm micropore membrane filter (PTFE, Waters, Milford, MA, USA) and injecting them into the UHPLC-PDA system. Liquid chromatography was performed using an UHPLC C18 column (Suplecosil LC-18-DD 150 X 4.6 mm) The mobile phase constitutes 0.1% formic acid in water (eluent A) and 0.1% formic acid on methanol (eluent B). The gradient program was as follows: 0–0.1 min, B (15%); 0.1–7.1 min, B (65%); 7.1–7.9 min, B (95%); 7.9–8.0 min, B (95%); 8.0–10.0 min, B (15%). The flow rate in the mobile phase was constant at 1.0 ml/min, the column temperature was set at 30 °C and the effluents was measured at 280 nm. Pure phenolic standards were run to calibrate a standard curve with a R2 = 0.95.

Statistical analysis

All the analyses carried out in this study were performed in triplicate, and the average of all results were presented with the standard deviation. The design of all models and analysing the RSM data was performed using Minitab software. The analysis of variance (ANOVA) was done to show the linear, square, and 2-way interaction effects of the various independent factors on all response factors covered in this study. The validity, precision, and significance of the models were evaluated using several statistical parameters such as coefficient of determination (R2), adjusted R2, adequate precision, and coefficient of variation (CV). Significance of the response factors was determined at 95% (p < 0.05).

Results and discussion

The particle size and particle charge of synthesised DSNPs were analysed by Zetasizer, Nano series, HT Laser. The results obtained from this study suggest that the hydrolysis using HCl for 4–5 days was required to reduce the particle size to the targeted range between 50 and 150 nm (Table 1). The lowest particle size of 76.92 nm was found in DSNPs obtained by 5 days of acid hydrolysis (38% HCl) treatment. Moreover, Tables S1 and S2 showed that treating DSNPs with US had a significant effect in further reducing the particle size of DSNPs from 85.42 nm to 66.45 nm due to the cleavage phenomenon resulted from cavitation bubbles [13]. The particle charge obtained for HCl hydrolysed samples for 3, 4, and 5 days was in the range of −0.53 to −6.21 mV (Table 1). The particle charge of −6.21 was found in the DSP sample hydrolysed for 4 days at 40 °C using 38% HCl with 40:1 S:L ratio (Table 1). The particle charge of US treated samples using water-based solvent and methanol-based solvent was between −4.31 to −6.28 mV and −5.16 to −6.43 mV, respectively (Tables S1 and S2), which show no significant change in the charge compared to DSNPs produced by 4 days HCl hydrolysis (Table 1). The morphology of the synthesized DSNPs were determined using SEM and the images collected confirmed that the synthesized DSNPs did not have a uniform shape of particles (Fig. 1). SEM images also displayed that the DSNPs treated with US were overall having the more uniform shape and the lower size of the particles than the HCl treated samples (Fig. 1). The results suggested that US was able to disrupt the date seed particle in a more homogenous manner to produce the particles of homogenous shape and size.

Fig. 1

Scanning electron microscopy images of DSNPs produced using A) HCl hydrolysed DSP for 4 days (x1000) B) HCl hydrolysed DSP for 4 days (x1500) C) DSNPs treated for 6 min with ultrasound (90% amplitude) (x1000) D) DSNPs treated for 6 min with ultrasound (90% amplitude) (x1500). E) FTIR spectrum for unhydrolyzed date seed power (initial DSP), acid hydrolysed (38% HCl) DSP for 4 days, and DSNPs treated with US (90% amplitude) for 6 min.

The functional groups present in the samples were identified using FTIR spectrophotometer and peak obtained for each sample were compared with data reported by Coates [12]. The FTIR spectra obtained for all samples have confirmed that overall, there was no noticeable change between HCl hydrolysed DSP for 4 days, and DSNPs treated for 9 min with US (Fig. 1). However, the DSP (unhydrolyzed) showed two intense peaks between 2900 and 3000 cm−1 which were either absent or having low intensity in other samples. Moreover, only the peak obtained around 1366–1368 cm−1 was found in the DSP (unhydrolyzed) and US treated DSNPs sample, while all other peaks were common in all above mentioned samples (Table 4 and Fig. 1). The peaks obtained around 3290–3316 cm−1 in all analysed samples are usually interpreted as hydrogen bonding that is found in alcohols, water, and hydrates, further some previous studies have interpreted these peaks as amino compounds and ammonium compounds [12], [34]. Previous studies on DSP have never reported the presence of nitrogenous base materials hence the possibility of having an amino compound in the date seed is very slim. The presence of peak such as 1367, 511.91, 482.4, and 460 cm−1 which lay between 600 and 800 cm−1 and 1300–1600 cm−1, are due to the adsorption of a hydroxyl functional group (alcohol) by a hydrogen-bonded OH. The peaks around 2910–2934.53 cm−1 are linked to the methylene carbon-hydrogen bond (C–H) stretching vibration [22], [36]. The peaks obtained between 1718 and 1744 cm−1 are regularly linked with the presence of a carbonyl compound (ester, aldehyde, ketone, etc) since these compounds mainly fall in the absorption range of 1700–1750 cm−1 [36]. For the peaks that lay between 807.98 and 808.91 cm−1 which confirm the C–H out-of-plane bending vibrations and existence of aromatic structure. Finally, peaks found around 1022.98–1054 cm−1 are a result of CH–O–CH stretching and can be linked to many organic compounds present in date seeds [12], [36]. Overall, the FTIR spectra of DSP was slightly affected upon HCl and US treatments with some intense peaks (2910 and 2924 cm−1) being absent in the hydrolysed and US treated samples.

Table 4

FTIR spectra peaks and the functional groups for DSP and DSNPs produced by acid hydrolysis and those treated with US.

Functional group	Functional group wavelength	Peaks of unhydrolyzed DSP	Peaks of hydrolysed DSP	Peaks of DSNPs treated with US	Notes
C–H	670–900	808	814.98	808.71	Aromatic 1.3 Disubstitution
	800–860	460	482.4	511.91	Aromatic C_H out-of-plane bend
					1.4-Disubstitution (para)
					Aromatic C–H out of plane bend
C-O	1050–1150	1054	1030.01	1022.98	Alcohol C-O
C-O	1210–1320	1246	1216	1213.06	Acid C-O
-C–H	1350–1480	1376	N/A	1367.19	Alkane -C–H
		1456
C = C	1620–1680	1744	1720.94	1718.00	Alkenyl C = C stretch
=CH2	2915–2935	2910	2934.53	2927.37	Methylene C–H stretch
		2924
O–H	3200–3400	3292.18	3316	3290.67	Polymeric OH stretch

Modelling of the hydrolysis optimization experiment and determining the effect of acid hydrolysis conditions on particle size and TPC of DSNPs

The experimental conditions selected for each hydrolysis factor in the Box-Behnken design, and the obtained responses (Particle size and TPC) are presented in Table 1. The lowest particle size of 76.92 nm was achieved at 5 days hydrolysis (38% HCl) at 40 °C using 30 mg/ml S:L ratio, while the highest TPC values of 20.53 mg GAE/g DSNPs was achieved at 4 days hydrolysis at 40 °C using HCl 38% concentration and 40 mg/ml S:L ratio (Table 1). Overall, the particle size and TPC values of DSNPs were in the range of 76.92–4419.44 nm and 7.85–20.53 mg GAE/g, respectively. Previous studies that used water-based solvent to extract TPC from date seed powder have reported the extraction of 160 g FE/kg and 4430 mg GAE/100 g, which is significantly higher when compared to the reported TPC extracted from DSNPs [3], [30]. The difference in the amount of TPC extracted can be linked to the loss of some phenolic compounds during acid hydrolysis process required to produce the nanoparticles. Increasing the hydrolysis time, HCl concentration, and temperature significantly improved the production of DSNPs in the desired size of <150 nm while extracting the maximum TPC, while the solid: liquid ratio did not have significant effect on the reduction of particle size, however, it had significant effect on TPC (P < 0.05).Table 5 show the obtained results after multiple regression analysis of data for both particle size and TPC that include the coefficients of the models that determine the level of significance. The p-value of the proposed particle size model was 0.0000 which confirm the significance of the model. Hydrolysis time and concentration of HCl had significant effect on the particle size with p values of 0.004 and 0.000, respectively, while hydrolysis temperature was non-significant with p value 0.067 and solid to liquid ratio had no significant effect with p value 0.244. Furthermore, all the variables quadratic effect on particle size was non-significant with p-values > 0.05, except the HCl2 which had a significant effect with p-value 0.000 (Table 5). Similarly, almost all the 2-way interaction of variables had no significant effect on particle size with p-values > 0.05, apart from the interactions between hydrolysis time*HCl concentration and temperature*HCl concentration which had significant effect with p-values of 0.045 and 0.024, respectively. These findings reveal that HCl concentration had the most significant effect on the reduction of the particle size of DSP with 76.28% contribution, followed by hydrolysis time (0.99%). The squared effect of HCl concentration had also high contribution of 18.55%. The effectiveness of the synthesis of DSNPs was demonstrated using the following prediction equation:Where YPS is particle size (nm) and X1, X2, and X3 represent the hydrolysis time (min), hydrolysis temperature (°C), and HCl concentration (%), respectively.

Table 5

Regression coefficients values estimated for particle size and TPC of HCl hydrolysed DSP.

Term	Particle size				TPC
	Coefficient estimate	95% CI lower	95% CI upper	P-value	Coefficient estimate	95% CI lower	95% CI upper	P-value
Model
β0	680	409	951	0.000	13.73	13.26	14.21	0.000
Linear
β1	−216.9	−352.4	−81.4	0.004	2.08	1.84	2.32	0.000
β2	−125.3	−260	10.3	0.06	1.91	1.67	2.15	0.000
β3	−1904.9	−2040.4	−1769.4	0.000	4.01	3.77	4.25	0.000
β4	76.2	−59.4	211.7	0.24	2.02	1.79	2.26	0.000
Square
β11	112.3	−91.0	315.6	0.25	−0.37	−0.73	−0.01	0.044
β22	−0.9	−204.2	202.3	0.99	0.20	−0.16	0.56	0.243
β33	1335.0	1131.7	1538.2	0.000	0.24	−0.12	0.60	0.167
β44	19.1	−184.2	222.4	0.84	0.44	0.08	0.79	0.020
2-way interaction
β12	−10.0	−244.0	225.0	0.93	0.17	−0.24	0.59	0.382
β13	240.0	6.0	475	0.045	−0.02	−0.43	0.40	0.928
β14	1.0	−234.0	235.0	0.99	0.12	−0.29	0.53	0.548
β23	279.0	44.0	513.0	0.024	0.02	−0.39	0.43	0.928
β24	13.0	–222.0	248.0	0.91	−0.10	−0.51	0.32	0.617
β34	−171.0	−406.0	64.0	0.14	0.008	−0.41	−0.42	0.969
P-Value	0.0000				0.0000
F-value	86.96				168.49
R2	0.990				0.995
Adjusted R2	0.979				0.989
Predicted R2	0.944				0.972
Mean	1331.36				13.96
SD	21.548				0.3799
CV%	1.62				2.72
Adequate Precision	30.23				23.57
P-value (lack of fit)	0.079				0.240

For TPC as a response factor, hydrolysis time, temperature, HCl concentration, and solid to liquid ratio had significant effect on the TPC with p values 0.000, 0.000, 0.000, and 0.000, respectively. Furthermore, all the variables quadratic effect and 2-way interaction effect on TPC were non-significant with p-values > 0.05, except the hydrolysis time2 and S:L2 which had a significant effect with p-values 0.044 and 0.020, respectively. These findings reveal that HCl concentration had the most significant effect on TPC extraction with 56.25% contribution, followed by hydrolysis time (15.15%). The efficacy of TPC extraction from DSNPs was shown using the following prediction equation:Where YTPC is TPC (mg GAE/g DSNPs) and X1, X2, X3, and X4 represent the hydrolysis time (min), hydrolysis temperature (°C), HCl concentration (%), and S:L ratio (mg/ml), respectively.

The results of ANOVA, regression coefficients, and model adequacy of RSM are shown in Table 5. The coefficients of determination R2 of particle size and TPC were 0.990 and 0.995, respectively confirming that almost 99% of the total parameter’s variance was described by the quadratic polynomial models generated using Eq. 1. Adjusted R2 for particle size and TPC were 0.979 and 0.989, respectively, while the predicted R2 for both models were 0.944 and 0.972, respectively (Table 5). The comparability of adjusted R2 and predicted R2 values in both models and the fact that all the R2 values were close to 1 confirm that the models are highly significant and confirm the correlation between predicted and experimental values was high. The obtained two model p-values (0.000) and lack of fit (>0.05) for both models confirms that the models were highly significant and well fitted. The adequacy precision of the model for both particle size and TPC were 30.23 and 23.57, respectively which represent a good signal-to-noise ratio since both values are higher than 4 [37]. Moreover, low coefficient of variation (CV) values was obtained for both the responses (Particle size and TPC) (1.62 and 2.72), respectively, that were below 5% which signifies models’ precision and reproducibility [31]. All statistical findings presented in Table 5 verified that both models are valid and reliable to be used to optimize HCl hydrolysis conditions for synthesizing DSNPs and enhancing the extraction efficiency of the total phenolic compounds from DSNPs.

Furthermore, the effects of the independent factors on particle size were investigated by creating surface plots. The influence of the independent variables on the reduction of DSP is presented in Fig. 2. At constant HCl concentration and solid to liquid ratio, the particle size decreased as hydrolysis time and temperature were increased, reaching the maximum particle size reduction at 4.5–4.8 days and 50 °C (Fig. 2a). As the hydrolysis time and HCl concentration were increased at fixed hydrolysis temperature and S:L ratio, the particle size was significantly decreased (Fig. 2b). While fixing the temperature and HCl concentration at 40 °C and 19%, respectively, increasing the hydrolysis time was critical to reduce the particle size of DSP, but S:L ratio had no effect on particle size reduction (Fig. 2c). Finally, Fig. 2d illustrates the effect of hydrolysis temperature and HCl concentration on the particle size reduction at constant hydrolysis time and S: L ratio, a decrease in particle size was clearly noticed as the hydrolysis temperature and HCl concentration were increased.

Fig. 2

Response surface plot of particle size (nm) of DSNPs based on hydrolysis time vs temperature (a), hydrolysis time vs HCl concentration (b), temperature vs HCl concentration (c), and hydrolysis time vs S:L ratio (d), and response surface plot of TPC of DSNPs based on hydrolysis time vs temperature (e), hydrolysis time vs HCl concentration (f), hydrolysis time vs S:L ratio (g), and temperature vs HCl Concentration (h).

For TPC, at constant HCl concentration and solid to liquid ratio, the TPC increases with hydrolysis time and temperature, reaching the maximum at 4–4.5 days and 50 °C (Fig. 2e). Significant increase in TPC was observed when hydrolysis time and HCl concentration were increased at fixed hydrolysis temperature and S:L ratio, however the TPC tended to decrease when the hydrolysis time was >4.5 days (Fig. 2f). Meanwhile, at constant temperature and HCl concentration, increasing the hydrolysis time and S:L ratio enhanced the TPC (Fig. 2g). Finally, increasing the temperature and HCl concentration increased TPC extraction at fixed hydrolysis time and S:L ratio (Fig. 2h).

Modelling of the UAE optimization experiment using water-based solvent and determining the effect of extraction conditions on TPC, TFC, and antioxidant activities of DSNPs

The DSP extracted with water and treated with optimized UAE conditions (6 min, 90% amplitude) was analysed for TPC and DPPH radical scavenging activity to serve as a control and the results reported were 38.18 ± 0.86 mg GAE/g DSP and 141.37 ± 3.356 mmol TE/g DSP, respectively. The effect of various factors involved in UAE of DSNPs on the TPC and TFC and antioxidant activities using water as solvent was investigated, and the results are presented in Table 2. The highest TPC, TFC, DPPH radical scavenging activity, FRAP, and ABTS radical scavenging activity corresponding to 29.48 mg GAE/g DSNPs, 2.98 mg QE/g DSNPs, 23.77 mmol trolox/g DSNPs, 36.91 mmol trolox/g DSNPs, and 14.93 mmol trolox/g DSNPs, respectively were recorded at 90 % amplitude, 6 min treatment time, and 40 mg/ml S:L ratio. A positive trend in extraction efficiency of total phenolic compounds with higher antioxidant activity was seen with increasing the level of each factor of UAE. Overall, the TPC, TFC, DPPH, FRAP, and ABTS radical scavenging activities of DSNPs were in the range of 15.63–29.48 mg GAE/g, 1.11–2.98 mg QE/g DSNPs, 11.04–23.77 mmol trolox/g, 29.01–36.91 mmol trolox/g DSNPs, and 5.12–14.93 mmol trolox/g DSNPs, respectively (Table 2). Date seed powder (Khalas variety) extracted using 60 % ethanol that was analysed to determine the antioxidant activities, reported DPPH radical scavenging activity of 3200 µmol TE/g, ABTS radical scavenging activity of 4800 µmol TE/g, and FRAP activity of 2400 µmol TE/g [30]. The good antioxidant activities of the synthesized DSNPs are due to the improved physical characteristics and small particle size. The multiple regression analysis of data for both TPC and DPPH radical scavenging activity that include the coefficients of the models which determine the level of significance are shown in Table 6 and Table 7 include the results for TFC, FRAP, and ABTS radical scavenging activity. The p-values of the proposed TPC, TFC, DPPH radical scavenging activity and FRAP models were all 0.001, while for ABTS radical scavenging activity the model’s p-value was 0.000 which confirm the significance of all models. Ultrasound amplitude, treatment time, and S:L ratio had significant effect on the TPC and TFC with p values 0.025, 0.002, and 0.000, respectively. The effect of all the quadratic variables on TPC and TFC were non-significant with p-values > 0.05, except the treatment time2 which had a significant effect with p-value of 0.008. Regarding the 2-way interaction of variables, only the interaction between ultrasound power*S:L ratio had significant effect while all the other variables had no significant effect. These findings reveal that S:L ratio had the most significant effect on TPC and TFC from DSNPs with 67.28% and 66.74% contribution, respectively followed by treatment time with 14.29% and 14.74%, respectively. The effectiveness of TPC and TFC from DSNPs was represented using the following equations:Where YTPC is TPC (mg GAE/g DSNPs), YTFC is flavonoids (mg QE/g DSNPs), and X1, X2, and X3 represent the ultrasound amplitude (%), treatment time (min), and methanol concentration (%), respectively.

Table 6

Regression coefficients values estimated for TPC and DPPH radical scavenging activity of ultrasound-assisted extracted (UAE) DSNPs using water-based solvent.

Term	TPC				DPPH radical scavenging activity
	Coefficient estimate	95% CI lower	95% CI upper	P-value	Coefficient estimate	95% CI lower	95% CI upper	P-value
Model
β0	23.87	22.51	25.24	0.000	18.50	17.32	19.67	0.000
Linear
β1	1.03	0.20	1.87	0.025	1.25	0.53	1.97	0.007
β2	1.92	1.09	2.76	0.002	1.59	0.87	2.31	0.002
β3	4.17	3.34	5.01	0.000	3.56	2.84	4.28	0.000
Square
β11	0.32	−0.91	1.55	0.531	0.70	−0.36	1.76	0.149
β22	−0.202	−3.25	−0.79	0.008	−1.85	−2.91	−0.79	0.007
β33	−0.71	−1.94	0.52	0.197	−0.44	−1.50	0.62	0.338
2-way interaction
β12	0.74	−0.44	1.92	0.169	0.53	−0.49	1.55	0.238
β13	1.23	0.05	2.41	0.044	0.83	−0.19	1.84	0.092
β23	0.10	−1.08	1.28	0.840	−0.18	−1.20	0.84	0.665
P-Value	0.0010				0.0010
F-value	26.67				27.27
R2	0.979				0.980
Adjusted R2	0.943				0.944
Predicted R2	0.803				0.798
Mean	22.590				17.652
SD	0.9193				0.7923
CV%	4.07				4.49
Adequate Precision	26.87				25.36
P-value (lack of fit)	0.129				0.057

Table 7

Regression coefficients values estimated for TFC, FRAP and ABTS radical scavenging activity of ultrasound-assisted extracted (UAE) DSNPs using water-based solvent.

Term	TFC		FRAP		ABTS
	Coefficient estimate	P-Value	Coefficient estimate	P-Value	Coefficient estimate	P-Value
Model
β0	2.222	0.000	33.057	0.000	8.688	0.000
Linear
β1	0.132	0.026	1.670	0.000	2.189	0.000
β2	0.260	0.002	0.983	0.005	0.945	0.000
β3	0.5528	0.000	0.1992	0.000	2.009	0.000
Square
β11	0.055	0.420	−0.197	0.539	1.850	0.000
β22	−0.275	0.007	−0.237	0.464	−1.188	0.001
β33	−0.088	0.216	−0.072	0.819	0.181	0.301
2-way interaction
β12	0.099	0.161	−0.093	0.759	0.274	0.129
β13	0.163	0.042	0.173	0.573	−0.172	0.305
β23	0.007	0.914	0.092	0.762	0.425	0.037
P-Value	0.001		0.002		0.000
F-value	27.76		20.97		119.96
R2	0.980		0.974		0.995
Adjusted R2	0.945		0.928		0.987
Predicted R2	0.810		0.786		0.956
Mean	1.99		32.79		9.137
SD	0.0988		0.5747		0.3016
CV%	4.96		1.75		3.30
Adequate Precision	6.58		30.53		19.89
P-value (lack of fit)	0.182		0.329		0.615

The impact of all the independent factors on antioxidant activities was highly significant with p-values < 0.05 (Table 6 and Table 7). However, quadratic effect of all the factors and 2-way interactions on DPPH radical scavenging activity (Table 6) and FRAP (Table 7) were non-significant with p-values > 0.05, except the treatment time2 which had a significant effect on DPPH radical scavenging activity with p-value of 0.007. For the ABTS radical scavenging activity, almost all the quadratic effect of all the factors had significant effect with p-values < 0.05 and only the interaction between US amplitude*S:L ratio had significant effect with p-value 0.037, while other 2-way interactions were not significant. The factors that enhanced the DPPH, FRAP, and ABTS activities of treated sample were S:L ratio, US amplitude, and treatment time. The S:L ratio factor contributions in DPPH, FRAP, and ABTS models were 64.61%, 49.64%, and 32.73%, respectively. The effect contribution percentage of US amplitude on DPPH, FRAP, and ABTS models was 7.99%, 34.89, and 38.87, respectively. The contribution of the effect of treatment time on DPPH, FRAP, and ABTS was 12.86%, 12.08%, and 7.25%, respectively. The efficacy of enhancing the antioxidant activities of DSNPs samples was demonstrated using the following prediction equations:Where Y DPPH is DPPH (mmol trolox/g DSNPs), Y FRAP is FRAP (mmol trolox/g DSNPs), Y ABTS is ABTS (mmol trolox/g DSNPs), and X1, X2, and X3 represent the US amplitude (%), treatment time (min), and S:L ratio (mg/ml), respectively.

Moreover, the results of ANOVA, regression coefficients, and model adequacy for UAE of DSNPs using water-based solvent are presented in Table 6 and Table 7. The R2 of TPC, TFC, DPPH radical scavenging activity, FRAP, and ABTS radical scavenging activity models were 0.979, 0.980, 0.980, 0.974, and 0.995, respectively confirming that almost 97–99% of the total parameter’s variance was explained by the quadratic polynomial models generated using Eq. 2. The adjusted R2 for all response factors were 0.943, 0.945, 0.944, 0.928, and 0.987, respectively, while the predicted R2 were 0.803, 0.810, 0.798, 0.786, and 0.956, respectively confirming high significance and correlation of the models. The obtained model p-values were all < 0.05 and lack of fit (>0.05) which confirms that the models were highly significant and well fitted. The adequacy precision for all models was valid with 26.87, 6.58, 25.36, 30.53, and 19.89, respectively. Moreover, low coefficient of variation (CV) values was obtained for the models (4.07%, 4.96%, 4.49%, 1.75%, and 3.30%). All statistically calculated data presented in Table 6 and Table 7 confirmed that the models are valid and reliable to be used to maximize the ultrasound-assisted extraction of phenolic compounds from DSNPs with high antioxidant activities using water-based solvent.

The effect of the independent factors on TPC and DPPH activity of DSNPs was quite similar and is shown by the surface plots presented in Fig. 3. At constant solid to liquid ratio of 30 mg/ml, the TPC and TFC and antioxidant activity increased when the US amplitude and treatment time was set to the highest levels, reaching the maximum at 90% amplitude and 9 min treatment time (Fig. 3a, d, g, j and m). While the US amplitude and S:L ratio were increased to the highest levels at constant treatment time of 6 min, the response of all the factors significantly increased (Fig. 3b, e, h, k and n). Further, as the US amplitude was fixed at 80 %, the TPC and TFC and antioxidant activities were optimized by increasing the treatment time and S:L ratio (Fig. 3c, f, i, l and o).

Fig. 3

Response surface plot of TPC (mg GAE/g DSNPS) based on US amplitude and treatment time (a), US amplitude vs S:L ratio (b), and treatment time vs S:L ratio (c). Flavonoids (mg QE/g DSNPS) based on US amplitude vs treatment time (d), US amplitude vs S:L ratio (e), and treatment time vs S:L ratio (f); Response surface plot of DPPH radical scavenging activity (mmol trolox/g DSNPS) based on US amplitude vs treatment time (g), US amplitude vs S:L ratio (h), and treatment time vs S:L ratio (i); Response surface plot of FRAP (mmol trolox/g DSNPS) based on US amplitude vs treatment time (j), US amplitude vs S:L ratio (k), and treatment time vs S:L ratio (l); Response surface plot of ABTS radical scavenging activity (mmol trolox/g DSNPS) based on US amplitude vs treatment time (m), US amplitude vs S:L ratio (n), and treatment time vs S:L ratio (o).

Modelling of the UAE optimization experiment using methanol-based solvent and determining the effect of extraction conditions on TPC, TFC, and antioxidant activities of DSNPs

The DSP extracted with methanol (80%) and treated with optimized UAE conditions (6 min, 90% amplitude) was analysed for TPC and DPPH radical scavenging activity to serve as a control and the results reported were 68.92 ± 1.79 mg GAE/g DSP and 281.73 ± 2.383 mmol TE/g DSP, respectively. The effect of various factors involved in UAE on the TPC and TFC and the antioxidant activities of DSNPs using methanol-based solvent is shown in Table 3. The maximum values for TPC, TFC, DPPH radical scavenging activity, FRAP, and ABTS radical scavenging activity were recorded as 63.64 mg GAE/g DSNPs, 6.82 mg QE/g DSNPs, 56.78 mmol trolox/g DSNPs, 67.31 mmol trolox/g DSNPs, and 25.42 mmol trolox/g DSNPs, respectively when 90 % amplitude, 6 min treatment time, and 80 % methanol concentration was used. A study done by [5] reported the TPC extraction of 5.32 g GAE/100 g from date seed powder (Bousthammi variety) using 80 % methanol as solvent at 35 °C for 12 h using an orbital shaker-incubator, which is comparable with the values obtained in this study. Another study reported a TPC extraction efficiency of 430 g FE/kg and 510 FE/kg from unhydrolyzed date seed powder (Khalas variety) using 80 % ethanol and 80 % acetone [30]. An increase in the TPC and antioxidant activity was recorded with increasing the level of each factor [US amplitude (%), treatment time (min), and methanol concentration (%)]. Overall, the TPC, TFC, DPPH radical scavenging activity, FRAP, and ABTS radical scavenging activity of DSNPs based extracts were in the range of 46.51–63.64 mg GAE/g, 2.31–6.82 mg QE/g DSNPs, 33.21–56.78 mmol trolox/g, 42.73–67.31 mmol trolox/g DSNPs, and 15.04–25.42 mmol trolox/g DSNPs, respectively (Table 3). It was observed that TPC and TFC as well as antioxidant activities recorded in ultrasound-extracted samples using methanol-based solvent were significantly higher than those obtained with water-based solvent (P < 0.05) which comply with previous studies reporting that methanol is the more potent solvent in extraction of phenolic compounds [2], [15].

Table 8 present the results of the multiple regression analysis of the data for both TPC and DPPH radical scavenging activity that include the coefficients of the models that determine the level of significance and Table 9 include the multiple regression analysis of the data for TFC, FRAP, and ABTS radical scavenging activity. The model results showed that the p-values of < 0.05 were obtained for TPC, TFC, DPPH radical scavenging activity, FRAP, and ABTS radical scavenging activity which confirm the significance of the models. The independent variables (US amplitude, treatment time, and methanol concentration) had significant (P < 0.05) effect on TPC and TFC of methanol-based UAE of DSNPs. Moreover, the quadratic effect of UAE factors on TPC were significant with p-values < 0.05, except for US amplitude2 which had no significant effect on TPC with p-value 0.190. However, quadratic effect of the UAE factors on TFC were insignificant with p-value > 0.05. The 2-way interaction of variables showed non-significant effect with p-values > 0.05, except US amplitude*methanol concentration interaction effect on TFC found to be significant (p-value = 0.047). The results confirmed that US amplitude had the most significant effect on TPC and TFC in DSNPs with 59.15% and 54.13% contribution, followed by methanol concentration with contribution of 20.54% and 21.71%, respectively [16], [19]. The effectiveness of TPC and FLV extraction from DSNPs was presented using the following equations:Where YTPC is TPC (mg GAE/g DSNPs), YTFC is TFC (mg QE/g DSNPs), and X1, X2, and X3 represent the US amplitude (%), treatment time (min), and methanol concentration (%), respectively.

Table 8

Regression coefficients values estimated for TPC and DPPH radical scavenging activity of ultrasound-assisted extracted (UAE) DSNPs using methanol-based solvent.

Term	TPC				DPPH
	Coefficient estimate	95% CI lower	95% CI upper	P-value	Coefficient estimate	95% CI lower	95% CI upper	P-value
Model
β0	55.695	54.70	56.69	0.000	49.91	49.27	50.54	0.000
Linear
β1	5.03	4.42	5.64	0.000	7.89	7.50	8.28	0.000
β2	2.35	1.74	2.96	0.000	2.00	1.61	2.39	0.000
β3	2.97	2.35	3.58	0.000	3.91	3.52	4.30	0.000
Square
β11	0.53	−0.37	1.43	0.190	−4.47	−5.04	−3.90	0.000
β22	−2.11	−3.01	−1.21	0.002	−0.46	−1.04	0.11	0.094
β33	−1.23	−2.13	−0.33	0.017	−0.19	−0.76	0.39	0.441
2-way interaction
β12	0.20	−0.67	1.06	0.582	0.14	−0.41	0.69	0.543
β13	0.08	−0.78	0.94	0.822	−0.26	−0.81	0.30	0.288
β23	0.10	−0.76	0.97	0.776	−0.11	−0.67	0.44	0.623
P-Value	0.0000				0.0000
F-value	83.6				437.3
R2	0.993				0.999
Adjusted R2	0.982				0.996
Predicted R2	0.895				0.986
Mean	54.194				47.78
SD	0.6724				0.4296
CV%	1.24				0.89
Adequate Precision	34.92				33.85
P-value (lack of fit)	0.005				0.518

Table 9

Regression coefficients values estimated for TFC, FRAP and ABTS radical scavenging activity of ultrasound-assisted extracted (UAE) DSNPs using methanol-based solvent.

Term	TFC		FRAP		ABTS
	Coefficient estimate	P-Value	Coefficient estimate	P-Value	Coefficient estimate	P-Value
Model
β0	4.363	0.000	59.767	0.000	19.907	0.000
Linear
β1	1.313	0.000	8.319	0.000	2.531	0.000
β2	0.649	0.004	2.226	0.000	1.089	0.007
β3	0.831	0.001	4.047	0.000	3.007	0.000
Square
β11	0.095	0.637	−4.578	0.000	−0.123	0.747
β22	−0.223	0.290	−0.253	0.166	0.327	0.408
β33	−0.368	0.108	−0.281	0.132	0.079	0.835
2-way interaction
β12	0.220	0.279	0.170	0.309	0.045	0.902
β13	0.475	0.047	0.112	0.487	0.368	0.338
β23	0.282	0.179	−0.122	0.451	0.218	0.559
P-Value	0.002		0.000		0.001
F-value	21.01		989.91		30.93
R2	0.974		0.999		0.982
Adjusted R2	0.928		0.998		0.951
Predicted R2	0.785		0.992		0.748
Mean	4.10		57.04		20.06
SD	0.362		0.300		0.695
CV%	8.78		0.53		3.44
Adequate Precision	12.77		35.42		25.84
P-value (lack of fit)	0.706		0.083		0.180

Moreover, all the independent factors used in UAE experimental model had significant effect on the antioxidant activities with all factors having p-value < 0.05. Additionally, all the variables quadratic effect and 2-way interactions on antioxidant activities were non-significant with p-values > 0.05, except the US amplitude2 which had a significant effect with p-value of 0.000 for the model fitted for DPPH and FRAP radical scavenging activity. Therefore, the factors that contributed the most to enhance the antioxidant activities of the DSNPs based extract were US amplitude and methanol concentration. For DPPH radical scavenging activity, the main factors contributions were US amplitude (68.91%) and methanol concentration (16.91%). While for FRAP, US amplitude contributed 54.13% and methanol concentration contributed 21.71%. Further, for ABTS radical scavenging activity, US amplitude methanol concentration contributed 68.91% and 16.91%, respectively. The effectiveness of enhancing the antioxidant activities of DSNPs samples was demonstrated using the following prediction equations:Where YDPPH is DPPH (mmol trolox/g DSNPs), YFRAP is FRAP (mmol trolox/g DSNPs), ABTS is ABTS (mmol trolox/g DSNPs), and X1, X2, and X3 represent the US amplitude (%), treatment time (min), methanol concentration (%), respectively.

The influence of the independent factors on all response factors was comparable and surface plots presents in Fig. 4 confirm that. At fixed methanol concentration, the TPC and flavonoids extraction and antioxidant activities were increased as US amplitude and treatment time was set to the highest levels (Fig. 4a, d, g, j and m). When the US amplitude and methanol concentration were increased at constant treatment time, the extraction of TPC and FLV and antioxidant activities were significantly increased (Fig. 4b, e, h, k and n). As the US amplitude was fixed at 80% all the response factors were maximized by increasing the treatment time and methanol concentration (Fig. 4c, f, i, l and o).

Fig. 4

Response surface plot of TPC (mg GAE/g DSNPS) based on US amplitude vs treatment time (a), US amplitude vs methanol concentration (b), and treatment time vs methanol concentration (c); Response surface plot of TFC (mg QE/g DSNPS) based on US amplitude vs treatment time (d), US amplitude vs methanol concentration (e), and treatment time vs methanol concentration (f); Response surface plot of DPPH radical scavenging activity (mmol trolox/g DSNPS) based on US amplitude vs treatment time (g), US amplitude vs methanol concentration (h), and treatment time vs methanol concentration (i); Response surface plot of FRAP (mmol trolox/g DSNPS) based on US amplitude vs treatment time (j), US amplitude vs methanol concentration (k), and treatment time vs methanol concentration (l)’ Response surface plot of. ABTS radical scavenging activity (mmol trolox/g DSNPS) based on US amplitude vs treatment time (m), US amplitude vs methanol concentration (n), and treatment time vs methanol concentration (o).

The results of ANOVA, regression coefficients, and model adequacy for TPC, TFC and antioxidant activities of DSNPs treated with UAE and methanol-based solvent are presented in Table 8 and Table 9. The R2 of TPC, TFC, DPPH radical scavenging activity, FRAP, and ABTS radical scavenging activity models were 0.993, 0.974, 0.999, 0.999, and 0.982, respectively confirming that almost 97.5–99% of the total parameter’s variance was described by the quadratic polynomial models generated using Eq. 2. Adjusted R2 for the response factors were 0.982, 0.928, 0.996, 0.998, and 0.951, respectively, while the predicted R2 were 0.895, 0.785, 0.986, 0.992, and 0.748, respectively confirming high significance and correlation of these model parameters. The obtained models p-values (<0.05) and lack of fit (>0.05) confirmed that the models were highly significant and well fitted. The adequacy precision for all response factors models was valid with 34.92, 12.77, 33.85, 35.42, and 25.84, respectively. Moreover, low coefficient of variation (CV) values for almost all models was obtained. All the supporting results presented in Table 8 and Table 9 indicated that the models were valid and reliable to be used to optimize the UAE of phenolic compounds using methanol-based solvent and producing DSNPs based extracts with high antioxidant activities.

Optimization of acid hydrolysis process and UAE process of DSNPs using water and methanol-based solvents

Optimization of hydrolysis and UAE processes for DSNPs was done using response optimizer in Minitab software to attain the maximum yield of TPC, TFC and the antioxidant activities from DSNPs with particle size in the nanoscale (<150 nm). To further validate all the model designs, the desirability function (DF) obtained from the response optimization step of the models was evaluated. The values of the DF lie between 0 and 1, where 0 is attributed for undesirable response and 1 is for optimum response [21]. Table 10 presents the required optimum conditions for all models to achieve the maximum predicted values for all the response factors, except for particle size as the target is to reduce the particle size to the nanoscale (<150 nm). The hydrolysis conditions required to produce the lowest particle size (68.3 nm) were 5 days HCl, 30 °C-40 °C temperature, 38% HCl concentration, and 30:1 S:L ratio, while to enhance the TPC from DSNPs (24.46 mg GAE/g DSNPs) recommended conditions are 5 days HCl hydrolysis, 50 °C, 38% HCl concentration, and 40:1 mg/ml S:L ratio. The optimum conditions for UAE of DSNPs using water-based solvent to yield the highest TPC (30.86 mg GAE/g DSNPs), TFC (3.16 mg QE/g DSNPs),DPPH radical scavenging activity (24.91 mmol trolox/g DSNPs), FRAP (37.37 mmol trolox/g DSNPs), and ABTS (15.31 mmol trolox/g DSNPs)of treated samples were 90% US amplitude, 7.5–8 min treatment time except FRAP that needed 9 min treatment time to achieve the highest predicted value, and 40:1 mg/ml S:L ratio. Moreover, for the UAE process conditions using methanol-based solvent the highest TPC (63.90 GAE/g DSNPs), TFC (7.64 mg QE/g DSNPs), DPPH (58.43 mmol trolox/g DSNPs), FRAP (69.42 mmol trolox/g DSNPs), and ABTS (27.54 mmol trolox/g DSNPs) predicted values can be achieved at 88.4%-90% US amplitude, 7.9–9 min treatment time, and 80% methanol concentration. The desirability function for all the optimized models in UAE (water and methanol-based solvent) was 1 which was the desirable response. Moreover, the actual results reported in this study was very close to the predicted results generated using response optimization which validated the models used.

Table 10

Optimization parameters of HCl hydrolysis and UAE of DSNPs and validation of predicted and experimental values under optimum conditions.

Response factor		Optimum factors conditions		Highest values obtained
Acid hydrolysis model	Hydrolysis time (days)	Temperature (°C)	HCl conc. (%)	S:L ratio (mg/ml)	Response optimizer results	Actual results
Particle size (nm)	5 (5)	30–40 (40)	38 (38)	30:1 (30:1)	68.30	76.92
TPC (mg GAE/g DSNPs)	5 (4)	50 (40)	38 (38)	40:1 (40:1)	24.46	20.53
UAE model with water as solvent	US amplitude (%)	Treatment time (min)	S:L ratio (mg/ml)		Response optimizer results	Actual results
TPC (mg GAE/g DSNPs)	90 (90)	8 (6)	40:1 (40:1)		30.86	29.48
TFC (mg QE/g DSNPs)	90 (90)	7.97 (6)	40:1 (40:1)		3.16	2.98
DPPH (mmol trolox/g DSNPs)	90 (90)	7.55 (6)	40:1 (40:1)		24.91	23.77
FRAP (mmol trolox/g DSNPs)	90 (90)	9 (6)	40:1 (40:1)		37.37	36.91
ABTS (mmol trolox/g DSNPs)	90 (90)	8.1 (6)	40:1 (40:1)		15.31	14.93
UAE model with methanol as solvent	US amplitude (%)	Treatment time (min)	Methanol conc. (%)		Response optimizer results	Actual results
TPC (mg GAE/g DSNPs)	90 (90)	7.9 (6)	80 (80)		63.90	63.64
TFC (mg QE/g	90 (90)	9 (6)	80 (80)		7.64	6.82
DPPH (mmol trolox/g DSNPs)	88.8 (90)	9 (6)	80 (80)		58.43	56.78
FRAP (mmol trolox/g DSNPs)	89.4 (90)	9 (6)	80 (80)		69.42	67.31
ABTS (mmol trolox/g DSNPs)	90 (90)	9 (6)	80 (80)		27.45	25.42

Values presented in brackets denotes the actual process conditions used in the experimental assay.

Phenolic compounds present in the DSNPs

The identified and quantified major phenolic compounds present in the DSNPs samples that had the highest TPC content from each model are presented in Table 11. Samples extracted using water-based solvent reported 6 phenolic compounds, while samples extracted using methanol-based solvent showed 8 phenolic compounds. Phenolic compounds detected in DSNPs were categorized into hydroxybenzoic (3,4-Dihydroxy benzoic acid, gallic, 4 hydroxy benzoic acid, and vanillic acid) and hydroxycinnamic (p-coumaric, caffeic, ferulic acid, and cinnamic acid). For samples extracted using water-based solvent, ferulic acid was the main phenolic compound present, after that 3,4-Dihydroxy benzoic acid, while p-coumaric and vanillic acid were found in small amounts and caffeic acid and cinnamic acid were not detected. Regarding samples extracted using methanol-based solvent, 3,4-Dihydroxy benzoic acid was the main phenolic compound present, followed by p-coumaric, whereas caffeic acid and cinnamic acid were present in small quantities. 3,4-Dihydroxy benzoic acid is considered to have good antioxidant activities [27], [23], [47]. Further, benzoic acid, ferulic acid and p-coumaric acid are reported to possess strong antioxidant and antimicrobial activities [10], [26], [45], [55]. The finding of this study confirms that presence of several phenolic compounds in DSNPs that have high antioxidant and antimicrobial activities. Therefore, DSNPs can have wide and promising applications in the production of functional foods. Further studies are needed for fractionation and in-depth analysis of phenolic compounds.

Table 11

Major Phenolic compounds identified in DSNPs produced with acid hydrolysis and those treated with US using water and methanol-based solvents.

Phenolic compound (mg/100 g DSNPs)	Sample description
	Acid hydrolyzed DSP for 4 daysa	DSNPs extracted using water-based solventb	DSNPs extracted using methanol-based solventc
Gallic acid	6.33 ± 1.47	7.83 ± 0.08	8.91 ± 0.37
3,4-Dihydroxy-benzoic acid	5.13 ± 1.22	10.19 ± 1.04	27.05 ± 1.83
4-hydroxy-benzoic acid	2.71 ± 0.09	2.80 ± 0.13	11.17 ± 0.64
Vanillic acid	2.28 ± 1.16	4.31 ± 0.09	12.45 ± 0.72
Caffeic acid	n/a	n/a	3.15 ± 0.17
p-coumaric acid	0.58 ± 0.06	6.02 ± 0.08	17.62 ± 1.34
Ferulic acid	13.52 ± 1.86	12.42 ± 0.17	7.52 ± 0.28
Cinnamic acid	n/a	n/a	2.76 ± 0.07
Total	30.54 ± 2.72	43.58 ± 2.31	90.63 ± 3.47

a = HCL 4 days, 40 °C, 38% HCl, and 40:1 S:L ratio using water-based solvent; b = DSNPs extracted with 90% US amplitude, 6 min, and 40:1 S:L ratio using water-based solvent; c = DSNPs with 90% US amplitude, 6 min, and 80% methanol using methanol-based solvent.

Conclusion

This study reports a novel strategy to valorize date seeds by producing DSNPs under optimum conditions. Synthesis of date seed nanoparticles (DSNPs) with high bioactive properties using acid hydrolysis method depended on several factors such as hydrolysis time, temperature, concentration of acid, and solid: liquid ratio used. Optimization of the hydrolysis process have been performed using RSM and it was found that hydrolysis time (4–5 days) and concentration of acid (38%) were the main factors playing a role to produce DSNPs in the desirable size of <150 nm (76.92 nm) and high TPC (20.53 mg GAE/g DSNPs). Green extraction process (Ultrasound-assisted extraction: UAE) was utilized to enhance the TPC, TFC, and antioxidant activities of DSNPs using water-based solvent and methanol-based solvent. For water-based solvent, S:L ratio (40:1 mg/ml) and treatment time (6 min) were the independent and decisive factors for obtaining the highest TPC (29.48 mg GAE/g DSNPs), TFC (2.98 mg QE/g DSNPs) and antioxidant activity. Whereas for the UAE process coupled with methanol-based solvent, US amplitude (90%) and methanol concentration (80%) had the most significant influence in enhancing the TPC (63.63 mg GAE/g DSNPs) and TFC (6.82 mg QE/g DSNPs) as well as enhancing the antioxidant activities of DSNPs. The analysis of the extracted phenolic compounds from DSNPs have confirmed that 3,4-Dihydroxy benzoic acid, ferulic acid, and p-coumaric acid were the main phenolic compounds present in DSNPs. This study for the first time presents an efficient and novel method to synthesis nanoparticles from date seed which can be explored for diverse applications in food and other relevant industries.

CRediT authorship contribution statement

Hussein Mostafa: Methodology, Experimentation, Writing – original draft. Jennifer Osamede Airouyuwa: Assisted in Experimentation, and data analysis. Sajid Maqsood: Conceptualization, Funding acquisition, Methodology, Writing – review & editing, Resources, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.