All the reagents were of analytical grade, for the determination of ascorbic acid (AA) we used 2,4-dinitrophenylhydrazine (Lobachemie, India), glacial acetic acid (Scharlau, Spain), bromine water, thiourea, sodium acetate, metaphosphoric acid, sulfuric acid (98%, m/v), the reagents were acquired from Movilab (Lima, Peru). For iron determination, 1, 10-phenanthroline (Biolab, Argentina), ammonium iron (II) sulfate (Fe(SO₄)(NH₄)₂(SO₄)6H₂O) (Movilab), hydroxylamine hydrochloride (Scharlau, Spain), hydrochloric acid (37% m/v) was purchased from (J.T.Baker, Mexico), nitric acid (65% m/v) was purchased from (Merck, Germany). AA C₆H₈O₆ and ferrous sulfate (FeSO₄) were purchased from Movilab.

Yellow sweet potatoes (Ipomoea batatas) were acquired from a local supermarket in Sullana (Peru) and kept in storage at 27 ± 3°C for 2 to 3 days prior to utilization. The sample had a soluble solids content of 10 ± 1 °Brix and a moisture content of 79.8 ± 1.5%. The sweet potato samples were sliced into pieces measuring11 mm x 11 mm x 5 mm for further processing.

Iron and AA were incorporated following the methodologies reported by Carvalho et al. (2021) and Rojas et al. (2019). First, the solution was prepared containing ferrous sulfate at 0.1% w/v and AA at 1% w/v (FA). Next, sweet potato slices were added to a beaker containing FA solution in a 1:5 w/v ratio. The above samples were placed in an the ultrasonic bath (Elmasonic P 30H, Germany) and treated at different frequencies (37 and 80 kHz) for 10, 20, 30, 40, 50, and 60 min., and ultrasound power of 100 watts and temperature inside the ultrasonic bath was 30°C. A control sample (without ultrasonic treatment) was also prepared in parallel. Subsequently, both the ultrasound-treated and control samples were drained, and dehydrated in a tray dryer (Dehydrator ST-04, 30–90°C, 0 − 15 h), at 45 ± 2°C for 8 h at air circulation speed (2.5 m/s). The Schematic diagram of the ultrasound-assisted incorporation of iron and AA in sweet potatoes is shown in Figure 1.

Total AA (ascorbic acid + dehydroascorbic acid) was quantified using the UV spectrophotometric method described by Rahman et al. (2007). Ascorbic acid is oxidized to dehydroascorbic acid by the action of bromine water in an acidic medium, generated by the presence of acetic acid. The AA was extracted from the samples; for this, 1 g of the sample was mixed with 20 mL of 5% w/v metaphosphoric acid solution and 10% v/v acetic acid centrifuged at 2500 rpm for 4 min. Next, the obtained mixture was filtered with Whatman No. 1 filter. Next, 3 mL of filtered solution (SF) was reacted with 0.15 mL of bromine water, and then gently stirred. Additionally, 0.80 mL of 10% m/v thiourea was added to remove excess bromine, then 1 mL of 2,4- dinitrophenylhydrazine solution was added and gently stirred. To finish the reaction, the samples were maintained in an incubator (Usamed, DNP-9052A) at 37°C temperature for 3 h; subsequently, the samples solutions were immediately cooled in an ice bath, added 1 mL of 85% m/v H₂SO₄ with constant stirring and made up to 5 mL with distilled water. The absorbance of the colored solution obtained was measured at 521 nm in a UV–visible spectrophotometer (Thermo Scientific, model Genesys 150). The calibration curves were obtained with standard solutions of AA ranging from 2.5 to 20 mg/L, following the same procedure used for the samples, and the concentrations were expressed in mg of AA (AAE) /100 g. All determinations were performed in triplicate and all steps were repeated for the blank sample.

The determination of total iron was performed according to the AOAC (1994) method (944.02) with adaptations. In this method, Fe²⁺ reacts with 1,10-phenanthroline to form a characteristic red-orange complex that absorbs strongly in the region around 510 nm of the visible spectrum. Hydroxylamine hydrochloride solution was applied as a reducing agent to reduce Fe ³⁺ to Fe ²⁺.

To perform the analysis, a crucible was used to weigh 1 g of the sample and which was subsequently placed in a muffle furnace (Sel-Horn “R- 8 L) at 550 ± 15°C for a period of 5 h. Next, 0.1 mL of nitric acid was added to the ashes obtained and evaporated on a hot plate. Next, 1 mL of hydrochloric acid was added and again evaporated on a hot plate. Finally, 0.2 mL of hydrochloric acid was added, and the volume was filled up to 10 mL with distilled water. For quantification, an aliquot of 2 mL of the above solution was reacted with 0.6 mL of hydroxylamine hydrochloride, then allowed to stand in a dark place for 5 min, followed by the addition of 1.5 mL of acetate buffer and 0.6 mL of 1,10-phenanthroline, stirred gently and distilled water was added to make up the 5 mL. Finally, the solution was agitated in a vortex (Cole-Parmer, model SA8) at 1000 rpm × 10s and kept at rest for 10 min. The absorbance was measured at 510 nm in a UV–visible spectrophotometer. The calibration curves were obtained with standard solutions of Fe²⁺ ranging from 0.5 to 5.0 mg/L, following the same procedure applied to the samples, and the concentrations were expressed in mg /100 g. All determinations were performed in triplicate and all steps were repeated for the blank sample.

For screening purposes, a 2² full factorial design, with two levels (−1 and + 1) was applied to the two independent factors frequency (37–80 kHz) and ultrasound exposure time (10 and 60 min) in the responses of iron (mg/100 g) content and AA (mg AAE/100 g), in a total of 4 runs (Table 1). All the experiments were performed in triplicate.

Apart from the factorial design, aliquots were evaluated every 10 min (20, 30, 40, and 50 min) in 37 and 80 kHz to process control, in a total of 8 runs. For the analysis of results, the Analysis of Variance method (ANOVA) and Tukey’s test were used to determine differences between different treatments. The process was evaluated using factorial design with Statgraphics Centurion XVI software and the analysis was performed with IBM-SPSS software.

Table 2 presents the iron and AA content in the dehydrated sweet potato after the different treatments with and without ultrasound. The control dehydrated sweet potato (CT) presented an iron content of 2.15 ± 0.01 mg/100 g and an AA content of 63.61 ± 0.50 mg/100 g, values higher than those found by Amagloh et al. (2022), who reported an AA content in fresh sweet potato ranging from 3.040 to 16.698 mg AAE/100 g. Peruvian food composition tables indicated an even lower AA content of 0.80 mg/100 g was reported for dehydrated sweet potato; however, the iron values were 2.90 mg/100 g (María et al., 2017); which is in agreement with the findings of the present study. Both ultrasonic (US) and non-ultrasonic (NUS) treatments were effective in incorporating iron and AA into the sweet potato. Longer treatments resulted increased the content of both iron and AA content; however, ultrasonic frequency, as well as the interaction (frequency versus time), did not influence AA content; on the other hand, in the case of iron content, time and interaction were significant, as observed in Table 2.

According to Mashkour et al. (2018), the use of a 37 kHz frequency generated higher micronutrient incorporation compared to an 80 kHz frequency, due to the cavitation phenomenon induced by the US, the cavitational collapse produces the cellular rupture in the plant tissue, which allows the increased of cell permeability, resulting in better mass transfer. Lower US frequencies generate longer wavelengths, and longer compression cycles and produce a violent cavitational implosion, resulting in greater cell disruption. In contrast, Yu et al. (2016) testing three different frequencies (25, 45, and 100 kHz) to enrich peanuts with resveratrol, used US as a pretreatment, the optimal frequency was found to be 100 kHz. The difference in these results may be attributed in the different in the food matrix, pretreatment methods, and frequencies utilized.

The iron content was incorporated in greater quantity in the treatments with longer exposure times (Figure 2B), for instance (Table 2) the treatment at 37 kHz for 50 and 60 min, showed an iron content of 72.96 and 105.91 mg/100 g. This represents an increase of 3,293 and 4,826% with respect to the control, and at 80 kHz the values were 60.52 and 77.98 mg/100 g in both times showing an increase in iron content of up to 2,715 and 3,427%, respectively. These results were higher than those observed at the same times in NUS. Incorporating iron in sweet potato with US at 30 min allows to obtain values equivalent to those demonstrated in the twice the duration in NUS, this demonstrates the effectiveness of the application of US. The iron content shown in each treatment studied, supplements the recommended daily intake of iron for pregnant women, which recommendation requires the consumption of 23 to 27 mg of iron/day. Additionally, the dietary iron requirement for adolescents is also covered, which ranges varies from 8 to 15 mg/day, depending on age and sex (Ministerio de Salud [MINSA], 2016).

Currently, there are no precedents for incorporating iron into sweet potatoes; however, in other foods, it has been possible to incorporate iron and other micronutrients into other foods using the same technology. Carvalho et al. (2021) propose the iron fortification of pineapple chips by using microencapsulation, ethanol, ultrasound, and convective drying. The authors mentioned that the longer time (30 min) of pre-treatment in ethanol and US increased the iron content up to 1157.5%; however, compared to the pre-treatment in ethanol, the US treatment did not significantly improve the iron incorporation. Rojas et al. (2019) showed a similar result for the incorporation of iron and microencapsulated carotenoids in pumpkin and apple. The authors found that by applying US treatment for 30 min, an increase of more than 1,000% of iron incorporation was found and its application helped to improve the dispersion of ethanol microparticles, despite this, the US did not significantly affect the results compared to the treatment without US. This is consistent with Miano and Augusto (2018) found that the iron content in beans increases significantly at longer exposure times, they obtained an increase of 1418.6 mg/kg of iron during 510 min of processing.

The amount of AA incorporation in sweet potato depends significantly on the time factor, as shown in Table 2, which ranges from 93.32 to 152.73 mg AAE/100 g in NUS; US at 37 kHz from 103.71 to 392.65 mg AAE/100 g and US at 80 kHz from 91.82 to 235.70 mg AAE/100 g is observed; despite the frequency, factor did not evidence a significant statistical influence, at 37 kHz the highest AA ranges were observed (Figure 2A). Showing effective incorporation of AA from 63 to 517%, with respect to the control. In fact, 20 min of treatment was sufficient to exceed 60 min in NUS, achieving 50% higher AA. In similar studies, under the vitamin fortification approach in other food matrices, similar behavior was observed; for instance, Tiozon et al. (2021) incorporated folic acid (vitamin B) into rice using ultrasound as a pretreatment for 1, 3, and 5 min, found that longer sonication times resulted in greater incorporation and increased folic acid levels by 1,982% in brown rice and by 4,054 times in milled rice.; in addition, US improved the micronutrient retention capacity after cooking, by 93.53% (brown) and 86.52% (milled); however, the sonication time to be used should be evaluated according to the nature of the micronutrient, food matrix, among other factors. The incorporation of vitamin B5 in rice was study for different durations of 5, 15, 25, and 35 min; 25 min of sonication produced the greatest mass transfer by the rupturing of the cell walls (changes in cell microstructure), which provided an absorption of up to 140% more pantothenic acid than in the untreated sample; because a long time produced the gelatinization due to temperature variation (42–45°C) (Bonto et al., 2018).

The aim of the optimization of the process was to obtain the parameters that allow a greater increase in the iron and AA content, as shown in the surface response plot in Figure 3. As observed, the lowest frequency (37 kHz) and the longest exposure time (60 min) were the optimal parameters for the incorporation of iron and AA, obtaining an iron content of 105.91 ± 0.03, an increase of 4928.99% and an AA content of 392.65 ± 4.84, an increase of 610.65% compared to its initial content. The basis for these results, as mentioned above, is due to the increased cell permeability of the feed tissue as a result of the cavitation collapse caused by the ultrasonic treatment. Similar results were obtained by Oladejo and Ma (2016), in their study optimized the US-assisted osmotic dehydration process in sweet potato (Ipomoea batatas) using methodology response surface (MRS); the optimized values were the following, frequency of 33.93 kHz, sucrose concentration 35.69% and the exposure time was 30 min, this time was the maximum value, concluding that the use of US allowed more mass transfer and that the lower frequency favors the results.

Similarly, Azarpazhooh et al. (2020) investigated the impregnation of grape pomace phenols in aloe vera using ultrasound-assisted osmotic dehydration (25 kHz, 200–400 W), and optimized the process using a MRS. The optimum values were 50% sucrose, 50°C, 59% amplitude, 20% grape pomace extract and 173 min exposure, with the maximum time value being 210 min. They study concluded that that a low frequency (25 kHz) with an exposure time of 173 min achieves a higher gain of solids, leading to increase the content and retention of phenolic compounds.

Sucheta et al. (2020), optimized the process of spice impregnation in black carrots assisted with US (37 kHz), they used a MRS design, whose optimum values were 8.18% NaCl, 4.30% spice mixture, and 14.47 min of ultrasound exposure. It should be noted that the maximum exposure time was 15 min, and the US improved mass transfer.

Mashkour et al. (2018) obtained different results, they optimized the process of iron fortification of potatoes by vacuum impregnation and US pre-treatment, the optimized values were; 37 kHz frequency and 24 min exposure time. The authors conclude that longer exposure time affect the nutritional content as they observed leakage of compounds from the potato; however, this is a side effect of the vacuum impregnation technique which was enhanced by the US. Similar results were reported by Maleki et al. (2020), implying that prolonged ultrasound exposure time damages carrot 303 microtissues by reducing mass transfer, the optimum values were 25 kHz frequency and 10 min.