Introduction

Microwave-assisted nanoparticle synthesis is a powerful tool, providing precise control over nanomaterial growth and morphology, with reasonable low energy consumption. In addition, quasi-uniform MW heating has been shown to significantly reduce not only reaction times but also to minimize side effects that improve both the rate and reproducibility of the synthesis. During NPs synthesis, real-time thermal control, and relatively low thermal gradients in the reaction mixture are needed to produce high-quality materials. Among MW heating parameters the reaction temperature, MW power applied, and uniform MW field distribution are critical for producing the highest structural quality and the narrow particle size distribution1,2,3.

In addition, dielectric heating in contrast to convective heating takes advantage of the selective energy transfer to materials that exhibit high dielectric losses, thus increasing the local microscopic temperature of the reaction. Asakuma et al.4 demonstrated that during MW-assisted nanoparticle synthesis MW energy is distributed to each particle causing overheating due to the heat generation by NPs and the thermal energy concentration at the particle surface. The overheating effect can be manifested by formation of gaseous microbubbles on the particle surface even at relatively low MW power level of tens of Watts5.

Although the fast synthesis of porous materials by MW heating is relatively well established, the mechanism and engineering causing the enhancement of the reaction rates are still unknown6,7. Several effects have been proposed to explain why synthesis time drastically decreased under MW irradiation including faster and more uniform heating of the reaction mixture, more effective interaction between species within the mixture, local superheating of the reaction components and creation of hot spots2. Additionally, there is no quantitative report to suggest which stage of nanoparticles formation: nucleation, particle growth or both are accelerated by MW irradiation8.

Mesoporous nanocomposites are preferentially synthesized based on mesoporous templates such as mesoporous silica6,7,8, which can be impregnated with a suitable precursor9. In order to fabricate hollow nanofibers with very high pore diameter a more advanced method such as coaxial electrospinning is suitable10.

For the synthesis of mSeNPs in aqueous medium, we adopted the hydrothermal method reported by Sun et al.11, where ZnNPs were used as hard templates, ascorbic acid (AA) was the reductant, and cetyltrimethylammonium bromide (CTAB) acted as the dispersing and capping agent of the NPs. Originally, the synthesis was running at room temperature12,13, while CTAB and AA concentration ranged 5 to 10 mM and 0.5 to 10 mM, respectively. Such nano-sized heterostructures are commonly formed through seeding-growth of one material on seed particles of another, yielding core/shell or composite nanoparticles. However, the detailed mechanism of formation of such hybrids has not been addressed. After removal of zinc template, the resulting SeNPs have a mesoporous structure.

Our overall strategy in this study was to clarify some of the characteristics of the MW methodology in the mSeNPs synthesis. The synthesis of mSeNPs via zinc as a hard template method has been performed for the first time under MW heating at 5.8 or 2.45 GHz. In particular, the thermal features of the 5.8 GHz MW irradiation used in the synthesis of mSeNPs are compared to 2.45 GHz and conventional convective heating (CH), with a special emphasis on the temperature effects and the characteristics of different MW reactors14,15.

Despite earlier observations that MW frequencies have no effect on the size and shape of the gold NPs synthesized in polar media16, we proceeded to examine the frequency effects of the 5.8 and 2.45 GHz MW irradiation on the synthesis of mSeNPs in aqueous solution of CTAB at concentration exceeding critical micelle concentration (cmc) level of 0.92 mM17,18. Additionally, the beneficial role of CTAB on the MW-assisted synthesis of mesoporous nanomaterials is attributed to promoting the self-organization of the micelles that enhances a “supramolecular templating” mechanism of particle growth.

Up to now, the number of reports concerning MW-assisted nanoparticle synthesis with the use of 5.8 GHz operated reactors is very limited16. Superheating of nonpolar solvents was achieved most efficiently by the 5.8 GHz MW irradiation, compared to 2.45 GHz frequency. It was shown that the merits of the 5.8 GHz frequency were prompt heating and superheating caused by the shallow penetration depth of these microwaves.

Mesoporous nanoparticles also reveal great features like high volume for drug loading in drug delivery systems. In this study, we further synthesized mSeNPs with high loading efficiency to combine them with the antifungal carbamates for enhancing pesticide bioavailability while reducing its dosage13,19.

Materials and methods

Ultra-pure water (15 MΩ cm) obtained from a Milli-Q Elix 3 water purification system Millipore (France) was used throughout. Sodium selenite (99%), cetyltrimethylammonium bromide (CTAB, ≥ 99%), ascorbic acid (≥ 99%), thiol-dPEG®4-acid (> 95%, HPLC grade) and zinc nanopowder with average particle size in the range of 40–60 nm (≥ 99% trace metals basis) were purchased from Sigma Aldrich and used for mSeNPs synthesis. Anhydrous ethanol (99.8%) and hydrochloric acid (36.5–38%, p.a. grade) were purchased from Avantor, Poland. Sodium diethyldithiocarbamate trihydrate (NaDDTC, ACS reagent) from Merck, Switzerland was dissolved in water and incorporated in mSeNPs. Selenium dioxide (99.8%, trace metal basis) from Sigma Aldrich, Germany and orange juice obtained from fresh oranges purchased from local market were applied to synthesize non-porous SeNPs, which were used as a reference for mSeNPs.

The 5.8 GHz microwave apparatus

The 5.8 GHz microwave apparatus is built of three main components, namely a solid-state MW Source, a Control Subsystem, and an MW applicator20. The design of the 5.8 GHz frequency MW generator (max. output power 40 Watts) follows a typical structure of a semiconductor-based MW power source, including a PLL frequency synthesizer, digitally controllable attenuator, a cascade of microwave amplifiers and an output forward and reflected power measurement circuitry. The cylindrical glass reactor (10 mm i.d., maximal volume of 10 mL) containing the sample solution is positioned in the MW cavity, which is a resonant structure optimized to uniformly deliver MW energy to the volume of reaction medium. A tunable electrical short is used to tune the field pattern inside the cavity.

A row of five infrared sensors enables non-invasive monitoring of temperature distribution along the height of the reaction flask and, together with an automatic temperature-control system, allows for continuous monitoring and control (± 0.1 °C) of the reaction temperature. The preset time-dependent temperature profile is followed automatically by continuously adjusting the applied MW power allowing the power to fluctuate to maintain the reaction temperature. A magnetic stirring is provided to enhance both temperature and NPs dispersion uniformity in the reaction flask and to enable reproducible reaction conditions.

The Control Subsystem is based on custom control software running on the Raspberry Pi 4 microcomputer.

The 2.45 GHz microwave apparatus

A single mode MW reactor MAGNUM II 2.45 GHz (Ertec, Wrocław, Poland) designed for chemical syntheses in 100-mL PTFE closed vessel was used to synthesize the mSeNPs. Continuous MW irradiation was provided by a magnetron (maximum power, 300 W), coupled to a heating block through a coaxial cable. The incident and reflected MW power levels were measured by the power monitor. A temperature profile was programmed as a constant temperature mode where the MW power is set to the highest constant power value that can reach and maintain the temperature of about 100–120 °C for a fixed duration of synthesis time. In all experiments, the vessel was filled with 20 mL of sample, and the MW power and the MW irradiation time were 40 W and 30 min, respectively. Once the reaction is completed, the water-cooling system is turned on to terminate the reaction.

CH was performed using a magnetic stirrer equipped with temperature-controlled heating plate IKA® C-MAG HS 7 (IKA, China). Incubation of mSeNPs suspension with sodium diethyldithiocarbamate was performed in a Multi-Therm shaker with heating and cooling (Benchmark Scientific, China). The suspensions were centrifuged by MPW-352RH centrifuge (MPW Med. Instruments, Poland).

Synthesis of nanoparticles

Initially, 73 mg CTAB was dispersed in 16 mL of water, and, after 30 min, 10 mg Zn, 5 mg PEG-SH was added, and then the mixture was sonicated within 5 min. Then, 17 mg Na2SeO3 was added and stirred for 2 h. Next, 8 mL of the mixture was heated in 5.8 GHz MW reactor after adding 2 mL of 1 mg mL−1 AA solution for 30 min, while stirring. Initially, the ramp period was set to 60 °C min−1 to reach the reaction temperature of 80 °C in the least amount of time. Next, the synthesis was running at constant temperature within 30 min, and then cooled. After centrifugation of the reaction mixture, the pellet was dispersed in 20 mL of ethanol:HCl mixture (39:1) and refluxed at 50 °C for 12 h to remove CTAB and Zn, then the dispersion was centrifuged at 10,000 rpm for 20 min. Finally, after supernatant removal, mSeNPs were dispersed in water.

Also, the reaction was performed in a 2.45 GHz MW reactor, and 20 mL of the reaction mixture was placed in 100-mL PTFE vessel sealed with a PTFE cap. Alternatively, the hydrothermal synthesis (20 mL of the reaction mixture) was carried out in a glass flask placed on a heating plate at 80 °C, while stirring.

Synthesis of mSeNPs/DDTC

As obtained mSeNPs were resuspended in 5 mL of distilled water and 2 mL of 20 mg L−1 NaDDTC aqueous solution was added to the mSeNPs dispersion and stirred at room temperature for 18 h, then allowed to stay for 5 days.

Nanoparticle characterization techniques

For optical characterization of Se/ZnNPs, mSeNPs and mSeNPs/DDTC, a V-560 UV–Vis spectrophotometer (JASCO, Germany) was used. All the samples were analyzed in 0.5 cm path length cuvettes from 250 to 700 nm. Derivative spectra calculations of the as-produced mSe/ZnNPs and mSeNPs were carried out on custom software based on the Savitzky-Golay algorithm.

Freeze-dried mSe/ZnNPs and mSeNPs were characterized by the method developed by Borowska et al.21 using single-particle microwave plasma optical emission spectrometry (SP-MWP-OES). The SP-MWP-OES experimental set-up consisted of six-phase rotating field 2.45 GHz MWP source (Ertec-Poland, Wrocław, Poland) as an excitation source and miniature optical spectrometer AvaSpec-2048XL with optical fiber FC-UV800-2-ME-SR (Avantes, Netherland) operating in time resolved analysis mode as an OES detection system (see also Table S1). The sample was introduced into the plasma, one particle at a time, through a vertically positioned glass tube and a conical dilution chamber. Raw data was acquired using vendor software AvaSoft 8.7 and transferred to Microsoft Excel. Microscopic images were taken using a Hitachi SU8230 ultra-high-resolution field-emission scanning electron microscope (Hitachi High-Technologies Corporation, Japan) at 10.0-kV accelerating voltage. Magnifications used were in the range from × 5 k to × 250 k.

Evaluation of the synthesis yield

ICP-OES (Integra XL, GBC Scientific Equipment) was used to quantify the conversion of selenite ions to mSeNPs. The measurements were carried out on the bulk mixture obtained after the microwave reaction, but after acid digestion, to assess the total amount of selenium (mSeNPs plus unreacted selenite) in the sample. Next, 10 mL of the mSeNPs dispersion was centrifuged at 10,000 rpm for 20 min. The ICP-OES analysis for this supernatant determined the amount of unreacted selenite present in the final mixture, thereby giving us the values to calculate the yield of the reaction. The measurements were carried out in triplicate for duplicate synthesis reactions and with the detection of Se both at a wavelength of 196 and 204 nm, independently, to double the experimental results (see also Table S2). Limits of detection for Se were 10 ng mL−1.

Results and discussion

Hard-templating synthesis of spherical mSeNPs with the use of ZnNPs has been developed by Zhao et al.12, and further improved by Sun et al.11, for use as a nano-carrier for anti-tumor drug delivery. The synthesis was running in aqueous 5 mM CTAB solution at room temperature for 2–6 h, without any heating, and Zn template was not removed after mSeNPs preparation. A mechanism of mSeNPs formation has not been studied. Next, the method has been modified by Liu et al.13, who added the Zn template removal procedure.

In this paper, we are focused on the MW-assisted synthesis of mSeNPs in an aqueous micellar solution of CTAB running at elevated temperature of about 100 °C for 30 min after reductant addition. The surfactant forms hydrophobic domains of self-assembled micelles acting as micro-reactors to synthesize well-distributed stable Se nuclei of nano-size dimensions but also as a micellar template for obtaining a mesoporous structure. Moreover, ZnNPs in the dispersion medium may play a role in achieving stable formation of Se nuclei on ZnNPs surface which exhibits catalytic activity towards the deposition of multiple Se nuclei, similarly to the typical seeding process of metallic NPs22. In turn, by providing enough time (2 h) for the CTAB micelles to adhere to the ZnNPs before synthesis the reduction rate of the Se precursor is decreased. Also, it can induce interfacial phenomena to inhibit subsequent growth of the mesoporous Se shells and to direct the formation of branched particles.

Microwave heating effect

The MW-assisted synthesis can be divided into three reaction stages: temperature ramping to initiate nucleation; a growth regime manipulated by reaction time and temperature; and a rapid thermal quenching step to control reaction termination. In the 5.8 GHz microwave reactor, the nucleation process is achieved by rapidly increasing the temperature from room temperature to 80 °C at full MW power (40W). Once the preset temperature is reached, the output power level is automatically reduced to hold the temperature constant, allowing for a controlled growth stage within a specified period of time. After that, the microwave source is turned off to quench the reaction that minimizes particle size distribution resulting from Ostwald ripening. All three stages in the MW-assisted synthesis are critical in the formation of a narrow size distribution of particles.

In contrast to conventional electric heating, the MW-assisted nanoparticle synthesis is characterized by rapid and homogeneous heating of the NPs dispersion, thus reaction rates proceed faster. The heating performances of microwaves, including MW frequency used, differ from one MW apparatus to another. In the case of 5.8 GHz system, the rapid MW heating was controlled based on averaged value of five temperatures measured along the height of the reaction vessel, and the preset profile (desired time and temperature) was followed automatically by continuously adjusting the applied MW power. Both the applied power and temperature were closely monitored so as to extract the characteristic thermal features of the microwaves. In the case of 2.45 GHz system, rise in temperature of the reaction mixture was intentionally matched to the set temperature and time value using microwave power level of 40 Watts.

Power and temperature reaction profiles during mSe/ZnNPs synthesis for two different MW heating modes are demonstrated in Fig. 1. A remarkable difference in heating efficiency was observed between the 5.8 and 2.45 GHz microwaves. The ramping period ranged from 1 to 5 min and initial ramping rates of temperature were 60 °C min−1 and 15 °C min−1 for the 5.8 GHz and 2.45 GHz, respectively, under a continuous applied MW power of 40 Watts. Thus, the rate of increase of temperature was nearly fourfold faster at 5.8 GHz than that for 2.45 GHz.

Fig. 1
figure 1

Power-temperature reaction profiles during microwave synthesis: (A) at 5.8 GHz, (B) at 2.45 GHz.

Full size image

In the 5.8 GHz reactor utilizing a solid-state generator, it is feasible to couple the generator to a set of temperature sensors via a feedback loop and to smoothly adjust the MW power to keep the desired temporary temperature value as shown in Fig. 1A. Exploratory experiments ascertained that fluctuations in temperature measured during the growth stage were less than 1 °C. In the same way, it is possible to enforce a predefined temperature course over time. In consequence, overheating of the reacting compounds can be avoided, and repeatability of the reaction conditions can be guaranteed. In contrast, in the 2.45 GHz reactor, temperature stabilization in the second stage of the synthesis is much more difficult, due to the lower controllability of the system (Fig. 1B).

Despite the heating performance of the system used, several MW induced effects should be considered, that may affect the temperature of the synthesis or may cause the reaction rates to proceed faster than the measured temperature would suggest.

It is known that MW irradiation causes selective heating of specific reaction components due to different electric permittivity. First, the presence of particles and CTAB micelles cause non-equilibrium MW energy distribution leading to interfacial instabilities23. During micelles formation, the MW field may create a more favorable alignment between molecules leading to increased losses (chaperone effect). Alternatively, absorption of the impingent MW electric field may cause raising the internal energy of the polar CTAB molecules, which is known as configurational or fictive temperature. In the syntheses of CdSe and CdTe nanoparticles, the configurational energy from selective MW heating (non-thermal effect) was found to support template activation, thus improving the reaction rates24. In fact, a configurational temperature cannot be measured by a thermometer. Thus, reaction temperature is measured at the vessel wall with infrared sensors, but the configurational temperature remains unknown. On the other hand, surfactant-induced capping effect phenomena may suppress the particle growth as shown by Takai et al.25.

In addition, the introduction to the reaction medium a high dissipation agent, such as zinc metal particle, would give a more rapid, non-equilibrium local heating of the dielectric particles within the reaction mixture. These ZnNPs embedded in micelles act as MW-activated templates within the reaction mixture and specifically adsorb MW energy (for Zn, the penetration depth is 2500 nm at 2.45 GHz26), leading to overheating nanoparticle surface and the thermal energy concentration at the solid–liquid interface, thus consistently influencing both the Se nucleation and particle growth kinetics. Such template activation appears to be a specific non-thermal phenomenon of microwaves. In addition, the presence of a heat dissipation agent plays a beneficial role in the synthesis of porous materials, namely more uniform and smaller spherical particles are formed.

In our synthesis conditions, MW irradiation penetrated the ZnNP-loaded CTAB micelles causing the temperature to rise by dielectric loss and to some extent by conduction loss heating, due to MW absorption by water molecules.

Finally, one more MW-induced effect should be mentioned. For NPs dispersion media, Asakuma et al.4 observed the formation of relatively stable gaseous microbubbles around the particles as a rapid thermal response on the specific MW absorption by the particle using even tens of Watts power level. Hence, it seems that the effect exhibits remarkable non-thermal component. Consequently, the special effects caused by MW-metal interaction cannot be ignored in this synthesis mechanism and may play an important role in nanoparticle nucleation and growth.

For a better understanding of the MW heating mechanism, we next examined the dielectric characteristics of the CTAB aqueous solution (10 mM), the CTAB/Zn micellar solution, and the Se/ZnNPs in CTAB solution (see Fig. S1 and Table S3). The corresponding measured dielectric loss factors (ε′′) were 9.86, 10.00 and 10.41, respectively. Variations of dielectric characteristics, summarized in Table 1 revealed mild and gradual increase of dielectric loss tangent of NPs dispersion that reflects a change in MW-particle interaction caused by particle growth.

Table 1 Summary of dielectric characteristics for various media throughout MW-assisted synthesis of mSe/ZnNPs.
Full size table

Microwave frequency effect

MW-assisted syntheses are commonly running using 2.45 GHz operated systems. The number of studies on the MW frequency effect on NPs formation is limited and the conclusions are ambiguous7,16. Since the dielectric constant of any material is frequency dependent, the MW frequency is also expected to play a role in NPs synthesis. Caponetti et al.27 observed that size and surface morphology of CdSNPs were frequency dependent, with 12 GHz producing the largest particles, and frequencies above and below that producing smaller particles. For barium titanate NPs synthesis, Nyutu et al.28 reported that MW frequencies higher than 2.45 GHz (up to 5.5 GHz) lead to more narrow particle size distributions and more spherical particles. The results for NiNPs MW-assisted synthesis presented by Ashley et al.15 clearly show that the kinetics of nucleation and growth are enhanced by increasing MW frequency up to 18 GHz. The enhanced nucleation rate is due to the increased configurational energy of the NPs dispersion. In turn, Horikoshi et al.16 observed that MW frequency (2.45 vs. 5.8 GHz) had no effect on the size and shape of the AuNPs synthesized in polar media.

The influence of frequency on a MW reaction is explored for the growth of mSeNPs using AA as a mild reducing agent. The growth of the NPs follows a catalytic growth mechanism allowing the effect of MW power and frequency on the kinetics of the reaction to be systematically evaluated. The growth kinetics was followed at two different frequencies (5.8 and 2.45 GHz) and at the same power level chosen to keep the heating rate constant for each reaction.

The magnitude of configurational energy change is dependent on the absorption of the MW energy by a single nanoparticle, which will be dependent on the frequency and applied power. Most materials exhibit much higher losses at 5.8 GHz than at 2.45 GHz, which is equivalent to more efficient conversion of MW energy into heat and results in faster temperature increase. Moreover, the shorter wavelength enables operation with lower power levels, as well as more effective control of the reaction.

Since each material will have a unique electromagnetic absorption cross section, NPs dispersion can experience selective heating, which impacts the rate of heating and, thus, the overall kinetics of the reaction. For metallic NPs the dielectric dispersion generally increases with increasing MW frequency, and therefore an increase in heating rate is anticipated at higher MW frequencies. Also, the high-frequency MW fields can induce focusing effect between neighboring particles that promotes NPs agglomeration during MW irradiation. In addition, the formation of microbubbles around the particles affects aggregate size and shape, as postulated by Asakuma et al.29.

The formation and time-dependent evolution of nanosized selenium particles have also been monitored using UV–vis spectrophotometry. In addition, SEM was used to assess the particle shape and particle size characteristics.

Morphology observations of mesoporous Se/ZnNPs

According to Sun et al.11, no 1D nanoparticles were grown by running the reaction at room temperature and without MW irradiation. SEM measurements indicated spherical-shaped nanoparticles with a size distribution in the range 55–80 nm.

In our study, at an early stage of both the MW-assisted and CH-synthesis the distorted spherical morphology was observed for mSe/Zn nanograins (Fig. 2B), with the size ranging from 80 to 100 nm, which was regulated by the dimension of inverse CTAB micelles formed on the single zinc NP. However, a closer look revealed a fine grain surface structure, which is made of Se nuclei.

Fig. 2
figure 2

SEM images for selenium nanograins (B) and final mesoporous Se anisotropic structures: MW heating at 5.8 GHz (A and C); MW heating at 2.45 GHz (D), conventional convective heating (E and F).

Full size image

As shown on SEM images (Fig. 2C), uniform nearly-spherical Se nuclei are obtained from MW heating at both microwave frequencies (diameter from 8 to 10 nm, with some up to 12 nm) under temperature conditions otherwise identical to those of CH, which produced slightly larger nuclei with a broader size distribution between 10 and 16 nm (Fig. 2F). MW irradiation promotes rapid and homogenous nucleation in producing finer particles at shorter time.

Moreover, when the synthesis was running at elevated temperature, 1D structures have been predominantly produced, namely rod-like particles in the case of CH (Fig. 2E,F) or multi-branched nanostructures (Fig. 2A) for the 5.8 GHz MW synthesis. The obtained mSeNPs are uniform in size and morphology. Interestingly, under 2.45 GHz reaction conditions, the final product exhibits mixed morphologies i.e. a number of rod-like particles together with multi-branched structures and also spherical nanoparticles (Fig. 2D). Clearly, the MW heating and frequency effect contribute to the formation of the mSe/Zn NPs in the CTAB micelles aqueous media.

At 2.45 GHz and by lowering the temperature ramping rate to 15 °C min−1, the population of multi-branched particles is greatly reduced, and primarily rod-like nanoparticles are formed (length = 4–5 µm) from the overpopulated spherical nanograins in comparison to the 5.8 GHz reaction, which produces multi-branched structures (arm length = 0.72 ± 0.23 µm).

This indicates that the use of microwaves not only enhances the reaction rate of reduction of selenite anions to Se atoms and the formation of Se spherical nuclei but also increases the growth rate and modified the final mSeNPs shape. This suggests that nucleation is temperature-dependent, and more nucleation occurs at a slower ramp rate. Thus, mechanistic variations in MW heating versus CH heating cause differences in heat exchange at the microscopic level which leads ultimately to variations in shape of hydrothermally fabricated mSeNPs. In addition, the morphology of single nanoparticles was examined by STEM and the images are presented in Fig. S2.

The influence of CTAB should be considered as an additional factor accompanying the MW-dependent growth, as it is known as the ligand capable of enhancing anisotropic shapes. Several studies30 indicated that adsorption of cationic CTAB on the surface of SeNPs is facet-selective. Thus, preferred growth of mSeNPs at ⟨111⟩ direction takes the form of nanorods or nanowires. In turn, CTAB induced capping effect prevents the particle growth.

Evaluation of formation and stability of mesoporous SeNPs by UV–Vis spectrophotometry

The formation of mSe/Zn core–shell NPs in CTAB micellar media in a MW cavity and time-dependent changes in their polydispersity, size, and shape have been examined by using UV–Vis spectrophotometry to follow the increase in the absorbance for mSe/ZnNPs specific wavelengths during MW irradiation.

UV–Vis spectra of both mSe/ZnNPs in CTAB micelles media and mSeNPs in water after template removal, registered at different stages of formation at reaction times of 1, 5, 15, and 30 min (ramp time + hold time) are displayed in Fig. 3, respectively. In the case of mSe/ZnNPs spectra, wavelengths below 300 nm cannot be measured due to the strong absorption of the reaction medium in this region. In general, the spectra of mSe/ZnNPs are rather featureless (Fig. 3A), except for the broad and weak absorption band at 260–270 nm and a rising broad absorption band above 610 nm. The former establishes the formation of spherical mSe/Zn nanograins where Se shell contribution corresponds to a particle of a substitute diameter of about 40 nm, following from the Mie formalism31. The second absorption band is associated with the formation of anisotropic shaped nanoparticles between 5 and 30 min of the synthesis.

Fig. 3
figure 3

Time-dependent UV–Vis absorption spectra of as-synthesized selenium nanoparticles on nanozinc templates (A) and their purified products after zinc removal (B) recorded at t = 1 (orange line), 5 (blue line), 15 (grey line), and 30 min (yellow line) for 40 W irradiation at 5.8 GHz. The red arrows indicated changes of absorbance at characteristic peaks in the course of reaction.

Full size image

The spectra of mSeNPs aqueous dispersion obtained after zinc and excess CTAB removal are more readable (Fig. 3B) and show that the band at 260–270 nm disappears within 15 min of the synthesis, while the absorbance of the second band gradually increases and the position of λmax redshifts from about 650 nm (5 min) to 690 nm (15 min) and moves back to 660 nm (30 min), indicating first an increase then a decrease of the aspect ratio of anisotropic particles.

Comparing absorption spectra of mSe/ZnNPs (Fig. 3A) with those of mSeNPs (Fig. 3B), it has been observed that the bimetallic particles show a different spectral profile from the mSeNPs with almost diminished Se-related absorption bands at 260 and 660 nm. It is clear that Se and Zn interact with each other in the case of mSe/ZnNPs, likely by the existence of so-called plasmonic “antenna effect”, well known for hybrid metal–semiconductor NPs32.

To get more qualitative and quantitative information from obtained UV–Vis spectra, derivative spectrophotometry was applied that uses first or higher derivatives of absorbance with respect to wavelength to resolve weak overlapping bands. A characteristic of the derivative spectrum is that it has a sharper peak than the original absorption spectrum and consequently weak bands and shoulders can be identified.

For all spectra shown in Fig. 3, after optimization of derivatization parameters, the first derivative absorption spectra were calculated based on Savitzky-Golay algorithm (with 35 spectrum points and a second degree polynomial) and displayed in Figs. 4 and 5.

Fig. 4
figure 4

Time-dependent first derivative UV–Vis absorption spectra collected over synthesis time to examine the formation of SeNPs on nanozinc templates: (A–D) at 5.8 GHz, t = 1, 5, 15, and 30 min; (E) at 2.45 GHz, t = 30 min; (F) non-porous SeNPs. Wavelengths < 300 nm cannot be measured due to the strong absorption of the reaction medium in this region.

Full size image
Fig. 5
figure 5

Time-dependent first derivative UV–Vis absorption spectra collected over synthesis time to examine the formation of template-free mesoporous SeNPs: (a–d) at 5.8 GHz, t = 1, 5, 15, and 30 min; (e) at 2.45 GHz, t = 30 min.

Full size image

In the case of mSe/ZnNPs (Fig. 4), where the absorption bands are so weak, it is much easier to observe them as peaks using the first-order derivative spectra. As can be seen from Fig. 4A–D, time-dependent 1st derivative spectra of as-synthesized at 5.8 GHz mSe/ZnNPs show a similar trendline but differ in both the number and amplitude of characteristic peaks, whereas the trendline for mSe/ZnNPs spectrum at 2.45 GHz is somewhat different (Fig. 4E). To further confirm the gradual formation of anisotropic shaped particles, one can observe negative peaks at the same wavelength ranges 370–400 nm and 660–680 nm, representing transverse and axial resonances, respectively. For comparison, a quite different 1st derivative spectrum registered for non-porous SeNPs at 2.45 GHz is presented in Fig. 4F.

Consequently, time-dependent 1st derivative spectra of mSeNPs at 5.8 GHz (Fig. 5A–D) and mSeNPs at 2.45 GHz (Fig. 5E) have been calculated. Here, positive peaks at the same wavelength ranges 260–300 nm and 550–690 nm can be regarded as resonances of spherical and anisotropic shaped particles, respectively. It is noteworthy that the peak representing spherical particles disappears after 5 min of the synthesis, and a gradual increase of the second peak (anisotropic particles) is observed between 5 and 30 min.

Selected data from Figs. 3 and 5 allowing the rate of growth of mSe/ZnNPs to be directly monitored by plotting the change in the 1st derivative value at 265 nm as a function of time, as absorbance will scale linearly with mSe/ZnNPs particle number concentration. The appearance of the band at 265 nm reflects the formation of spherical nanograins. A high number of spherical particles is obtained after 1 min reaction with size of around 40 nm (as shown in Fig. 3A).

Interestingly, the UV–Vis data (Fig. 6) indicates that spherical mSe/ZnNPs have been consumed to form rod-like structures, as can be seen in the gradual increase of the absorption peak at 650 nm. Also, the A650/A265 ratio increases with the synthesis time up to 15 and 30 min for the mSe/ZnNPs prepared by 5.8 GHz MW heating, respectively, indicating the successive change of particle morphology. The change in A650/A265 ratio is fitted to the sigmoidal growth curve behavior typical for the autocatalytic mechanism.

Fig. 6
figure 6

Time-dependent change in absorbance of zinc-template free mesoporous selenium nanoparticles dispersion at 265 nm ( quasi-spherical grains) and at 650 nm ( nanorods) based on first derivative spectra registered at t = 1, 5, 15, and 30 min (see Fig. 3) for 40 W irradiation at 5.8 GHz for monitoring the rate of growth of selenium nanostructures.

Full size image

Single particle measurements of elemental composition and size of mSe/ZnNPs and mSeNPs

For further chemical and physical characterization of mSeNPs, plasma-based optical emission spectrometric techniques have been used to provide qualitative and quantitative information about elemental composition and size of the nanomaterials studied. SP-MWP-OES is able to provide information on a particle-by-particle basis by using very high data acquisition frequencies for measuring individual nanoparticles21. At first, ZnNPs powder purchased from commercial source (at 40 and 60 nm nominal size) was analyzed to derive the particle size for later reference. The peak value is close to the manufacturer supplied data (Fig. 7A).

Fig. 7
figure 7

Analytical results of nanopowders measurements by SP-MWP-OES: (A) histogram of particle size distribution of pristine ZnNPs; (B) correlation between Zn and Se events for as-synthesized selenium nanoparticles on nanozinc templates; (C) correlation between Zn and Se events for mesoporous selenium nanoparticles after zinc-template removal.

Full size image

Diameters of ZnNPs measured in this study were consistent with those originally measured by TEM. Using the peak value of 50 nm for ZnNPs and assuming the spherical shape of mSe/Zn nanograins, the width of selenium shell (solid, non-porous) can be calculated based on UV–Vis measurements at 265 nm (Fig. 3A). The estimated value of Se shell width at about 4 nm is reasonable. It gives the average diameter of Se/Zn nanograin at about 60 nm, but in the case of non-porous selenium shell, versus 80–100 nm measured by SEM for mesoporous nanograins (Fig. 2B).

Next, the multi-element capability of SP-MWP-OES has been exploited to examine the elemental composition of mSe/ZnNPs and target mSeNPs as the technique can simultaneously detect Se and Zn at a single particle level. Analysis by SP-MWP-OES reveals the presence of zinc and selenium in individual particles from both nanomaterials studied (see Table S4). A plot of the time correlation between Se and Zn signals for all particles in mSe/ZnNPs sample (Fig. 7B) shows that the signals are well correlated and size-dependent for all particles. On the plot, each experimental point represents a single particle as a set of two values (signal intensity for Se and Zn, respectively). The closer a given point is located to the origin of coordinate system, the smaller the nanoparticle is. Obtained results indicate that most zinc nanoparticles are covered by a detectable amount of selenium. If mesoporous selenium is uniformly coated onto spherical ZnNPs, the correlation of the coating component and base material component should be linear. The linear trendline express a constant value of Zn/Se mass ratio for the measured particle population. Actually, it is related to the relationship between zinc mass and a particle mass for each particle in the sample. The linear regression coefficient value of 0.9745 is relatively high that confirms reproducibility of particle formation process. A group of experimental points placed on y axis can be regarded as excess ZnNPs insufficiently coated with selenium, and de facto represents impurity of as-obtained mSe/ZnNPs. In turn, three circled in red points represent statistically mSe/ZnNPs with selenium overgrowth. The single particle analysis of mSeNPs after zinc template removal confirmed effectiveness of the purification procedure (Fig. 7C). Most mSeNPs do not contain measurable amounts of zinc, and in the case of about 5% of all mSeNPs, zinc is removed with at least 80% efficiency.

Finally, the yield of MW-assisted synthesis has been assessed based on the Se mass balance for two consecutive runnings. Selenium has been determined by ICP-OES at its two atomic lines independently using external calibration and the yield of synthesis was found to be 80 ± 2%.

Mechanism of the nucleation and growth of mSe/ZnNPs

Based on the experimental results, we suggest the following mechanism for the MW-assisted formation of the mSe/ZnNPs using ZnNPs as a hard template and CTAB as a micellar template. At least three factors contribute to the balance between nucleation and growth kinetics of mSe/ZnNPs i.e. the formation of micellar template, the reaction temperature and MW heating effects.

In the aqueous medium ZnNPs are spatially separated by amphiphilic CTAB micelles, giving rise to assemblies that promotes the formation of selenium nuclei followed by a growth process via selenium-decorated ZnNPs toward zinc-selenium core–shell structure with a porous selenium shell. In addition, CTAB molecules stabilize the nanoparticles and may also impart some control in the growth process. At low temperature, a relatively fast nucleation step is followed by a slow growth process, leading to overproduction of quasi-spherical selenium nuclei and spherical mSe/Zn nanograins.

CTAB micelles encapsulate hydrophobic ZnNPs within micelle cores based on the interfacial instability route23 with essentially every micelle containing encapsulated metal particles18. In addition, it causes suppression of the micelle/water interfacial tension that improves mass transport between bulk water and micelle core.

The growth of mSeNPs follows a catalytic two-step Finke-Watzky mechanism, where a slow reduction of selenite in solution by ascorbic acid is followed by a second step of fast reduction of the anionic precursor at the surface of the ZnNP which exhibits catalytic activity towards the deposition of multiple Se nuclei. The attachment of Se nuclei preferentially occurs on the ZnNPs surfaces with lower energy barriers, as a result of weak intermolecular force. The diffusion of selenite anions and Se atoms across the micelle regulates the formation of mesoporous structure of the self-assembling mSe/Zn nanograins. As in the initial stage of Se shell growth, when selenium precursor is reduced into the ZnNP-loaded CTAB micelles, the UV–Vis spectrum is identical to that expected for Se spheres dispersion, with a well-defined absorption band at 265 nm.

During the conventional heating assisted synthesis, the presence of CTAB promotes the surfactant directed anisotropic growth of mSe/ZnNPs, probably driven by the oriented attachment mechanism. The mSe/Zn nanograins attach to each other and grow into a rod due to the surface modification effect of CTAB which has a tendency to form elongated rod-like micellar structures6,23 and assists in mSe/Zn rods formation as supramolecular template. This behavior is consistent with the hypothesis that microwaves interact both with water molecules and ZnNPs thereby increasing the temperature responsible for the faster SeNPs growth.

The use of MW irradiation brings additional effects to the nucleation and growth of mSe/ZnNPs. The dimensions of ZnNPs are sufficiently small comparable to the penetration depth, hence it can effectively absorb MW energy due to their electrical property and can be heated very quickly in the MW field. This leads to a condition where the configurational temperature localized around the micelles is significantly higher in temperature than that measured for the bulk solution. These hot surfaces on metallic ZnNPs would speed up mass transfer and enhance their catalytic activity, thus the energetic barrier for heterogeneous nucleation is much lower than that for homogeneous nucleation.

Another effect of MW irradiation is that it may induce localized ionic currents on the ZnNPs hot surface in an alternating MW field providing additional driving force (so called lightning rod effect) for directional nanograins aggregation into nanorods. These unusual heating effects cannot be achieved by conventional means. As the mSe/Zn nanograins aggregates gradually into a rod, their resonance band at about 650 nm becomes slightly red shifted.

The reaction rate for mSe/ZnNPs growth is a function of MW absorption effectiveness and thus will be dependent on the applied power, frequency, irradiation time, and concentration of the reactants. It is reasonable to assume the higher temperature ramping rates will lead to higher configurational temperature, achieved for MW irradiation at 5.8 GHz, the burst of nucleation proceeds faster than continuous nucleation leading to a lag growth rate of spherical mSe/Zn nanograins under assistance of facet-selective CTAB adsorption followed by the formation of anisotropic shapes of both rod-like and multi-branched structures. Finally, due to the faster directional growth of mSe/ZnNPs and limited extent of lightning rod effect, the formation of multi-branched structures seems to be preferred at 5.8 GHz. Also, the role of microbubbles in formation of multi-branched NPs may be considered, according to Asakuma et al.29.

Fabrication of mSeNPs-carbamate nanopesticide

To demonstrate a potential application of mSeNPs we have obtained a nanocomposite by loading DDTC into mesoporous structure of SeNPs. Carbamates are widely used as pesticides and also selenium particles have been used previously as a nanopesticide13,33. The UV–Vis spectra of the reaction mixture and supernatants after incubation indicate that DDTC interacts with mSeNPs to form DDTC-loaded mesostructured SeNPs. DDTC shows three characteristic absorption bands at 207 nm, 262 nm and 283 nm in the UV–Vis spectrum, one of them coinciding with the absorption band of mSeNPs at 265 nm (Fig. 8A). Therefore, the absorption band at 207 nm was selected as a reference for DDTC and its corresponding absorbance for quantitative examination. Also, the disappearance of the 283 nm band provides evidence for the formation of the nanocomposite (see also Fig. S3).

Fig. 8
figure 8

Time-dependent UV–Vis absorption spectra of mesoporous SeNPs-DDTC nanocomposites recorded at 0 (blue line), 5 (orange line) and 18 h (grey line) of incubation (A); Time-dependent change of absorbance at 207 nm for non-(♦) and mesoporous SeNPs-DDTC nanocomposites synthesized at 2.45 GHz (●) and at 5.8 GHz (▲) recorded at t = 0, 5, 18, and 120 h of incubation (B).

Full size image

Time-dependent UV–Vis absorbance spectra were collected over time (t = 0, 5 and 18 h) to monitor the DDTC-loading process into mSeNPs produced at 5.8 GHz. By measuring the UV–Vis spectra for different concentrations of DDTC, the standard curve was drawn, and the loading rate of mSeNPs in 10 mM DDTC solution was calculated to be 59% (Fig. 8B). For comparison the procedure was carried out with both mSeNPs at 2.45 GHz and non-porous SeNPs. It is clear from Fig. 8B that the loading process is most effective and mesopores of mSeNPs at 5.8 GHz are relatively easily accessible.

Conclusions

The MW-stimulated CTAB-directed hydrothermal synthesis to controllably fabricate mesoporous multi-branched SeNPs by hard templating in self-assembled ZnNPs-loaded CTAB micelles has been developed. The essence of this new synthetic route is to exploit the self-assembly of CTAB molecules into micelles at the ZnNP–water interface and subject the sample to MW irradiation. Water has a medium loss tangent, so the ZnNP-containing CTAB micelles are heated directly and rapidly.

A number of distinct and beneficial effects of MW irradiation as an alternative to convective heating have been demonstrated. Microwave heating effects has proven to be a key factor in the synthesis of multi-branched NPs with uniform shape and size. The study revealed that by varying the heating mode and MW frequency, mSe/ZnNPs morphological control may be significantly improved. This provides the opportunity of controlling the Se/Zn nanostructures from spherical nuclei to spherical grains without use of any heating, to nanorods or multi-branched assemblies using conventional or MW heating, respectively, by varying the heating conditions (temperature ramping rate as well as target temperature and time of the reaction). It is clear, therefore, that some special thermal and non-thermal effects of microwaves must be involved in the growth of these metal–semiconductor nanoparticles. Moreover, particle material having dielectric property was also found to be an important factor for mSeNPs size during MW irradiation. A solid-state microwave generator operating at 5.8 GHz has been applied for the first time for NPs synthesis. The advantage of the generator over a magnetron-based system (2.45 GHz) is its tenability—both frequency and power of a solid-state generator can be precisely tuned, providing excellent controllability of reaction conditions including temperature ramp rate and temperature holding during the growth stage.

A series of electron microscopy characterization results suggest that the growth of mesoporous selenium shell on the zinc hard template is governed by a nucleation and seeding growth mechanism followed by temperature-dependent assembly of spherical nanograins and MW-controlled agglomeration. Two stages of irradiation may be useful for well-balanced nucleation and growth process during the MW irradiation: rapid and more nucleation at higher temperature ramping rate and shorter MW irradiation time in the first stage and particle growth at lower MW power in the second stage.

The nucleation mechanism is strongly affected on the non-thermal way by selective MW heating of ZnNP-containing CTAB micelles and could be controlled microscopically providing control of template activation, selenium nucleation and growth. The mSe nanograins could grow along the ⟨111⟩ direction through a “lightning rod” effect until all sphere-like particles had been completely consumed, eventually growing into mSe/Zn nanorods. The growth kinetics of the multi-branched mSe/ZnNPs can be controlled mainly by controlling the reaction temperature and MW-heating conditions including MW frequency.

Determination of chemical composition at a single particle level reveals new insights on the formation of nanoscale hybrids containing metal templates using the microwave reduction method. Using SP-MWP-OES, it has been shown that the chemical composition of mSe/ZnNPs is stable and particle size-independent, and the zinc template removal procedure is highly effective.