1. Introduction
Silica nanoparticles are one of the most known stable, biocompatible, inert, and non-toxic materials [
1,
2]. Furthermore, its silanol-rich surface makes it a perfect support for a wide range of applications via anchoring different functional groups [
2,
3,
4]. The synthesis of uniform silica nanoparticles with a particle size in the range 100–1000 nm started by hydrolysis condensation of TEOS from aqueous solution catalyzed with ammonium in the early of 1960 [
5]. Later on, much smaller silica nanoparticles were successfully prepared via microemulsion technique through the dispersion of water phase within a continuous oil phase with the aid of a suitable surfactant at the interface as stabilizing agent [
5]. The micro water droplets encapsulated within the surfactant core act as a microreactor for the hydrolysis condensation of TEOS into nano-silica with excellent control over the particle size distribution via controlling the size of the droplets. Further improvement was achieved by adding a co-surfactant to increase the stability of the produced micro emulsion [
5].
Next, researchers focused on the ability to control and improve the surface properties of the nano-silica such as electrical and optical [
6], mechanical strength [
7], hydrophobicity, and hydrophilicity [
8]. It was reported that increasing hydrophilicity of the surface enhances the water flux and accordingly facilitates reaching the active sites on the surface by the target components in the aqueous phase [
8]. One effective way of improving hydrophilicity is to anchor the surface with amino groups (SiO
2–NH
2). In addition to this, it is reported that amino groups attached to the surface improve the affinity of the surface to some target molecules such as heavy metal ions, CO
2, and some biological components [
2]. Pandey et al. (2021) anchored 3-aminopropyltrimthoxysilane (APTMS) to the surface of nano-silica to enhance its affinity toward capturing CO
2 [
9], while Zhang et al. (2021) anchored 3-aminopropyl) triethoxysilane (APTES) to prepare an efficient adsorbent to extract heparin [
10]. Devlin et al. (2021) prepared enzyme-modified silica nanoparticles to attract staphylococcus aureus and disperse biofilms [
11]. Thus, silica nanoparticles have become one of the most attractive materials for different applications such as catalysis, adsorption, chromatography, ceramics, biomedical devices, electronic substances, stabilizers, coating, sensors, and many others [
1,
2,
5].
Accordingly, the demand for specific designs and synthesis of silica nanoparticles has emerged to meet the requirements. All the reported surface modifications of the nano-silica were achieved by additional steps after the synthesis of the bare nano-silica, which consumes more effort, time, and chemicals. Meddeb et al. (2021) grafted β-cyclodextrin on the surface of silica nanoparticles [
6]. Wang et al. (2021) followed a post-synthesis procedure to introduce the multi-hydroxyl-containing gemini surfactant to the surface of silica nanoparticles at 70 °C and applied it for the adsorption of methyl orange [
12]. Liu et al. (2021) anchored the thymol functionality to the silica nanoparticles by two steps via a simple impregnation method and studied its antimicrobial efficiency [
13]. Bamane and Jagtap (2021) followed a two-step technique through grafting of glycidyloxypropyl trimethoxysilane on the surface of nano-silica followed by modifying it with dimethyl propionic acid for self-cleaning coatings applications [
8].
High levels of heavy metals in water are toxic to humans, animals, and aquatic systems [
14]. Thus, it is vital to remove these ions from wastewater using a variety of methods such as chemical precipitation, membrane filtration, electrochemical technique, ion exchange, and adsorption [
15]. Above all, adsorption is very attractive due to its low cost, simple operation and maintenance, and also can be very selective, efficient, and fast if a suitable adsorbent is used for the target pollutant [
15]. Zinc will be used in this work due to its extensive usage in metallurgy, power plants, transportation, and construction. Zinc is a major nutrient for the human body at the micro-level of concentration. However, a high intake of zinc may result in stomach cramps, skin irritations, anemia, vomiting, and damage of the pancreas, negatively affecting the immune system and the protein metabolism [
14]. Accordingly, the World Health Organization (WHO) has set its limit in drinking water to 3 ppm [
16].
The main challenge is to control the morphology, size, and functionality of the materials to improve their performances in the targeted applications [
2]. This work is aiming to synthesize amine-modified nano-silica via a simple one-step method under very mild conditions. In addition, this route of introducing amine functionality to the silica nanoparticles is expected to allow achieving high loading of amine that is evenly distributed within the structure, in addition to saving time, effort, and chemicals if the correct ratios of materials were used for the synthesis.
3. Results and Discussion
The crystal structure was studied through the powder X-ray diffractometer (λ = 1.54056 nm) at a scanning rate of 0.02 °/s in the 2θ range of 5° to 80°.
Figure 1 shows that the XRD pattern of all samples shows a major characteristic broad peak at 2θ = 23° indicating an amorphous nature of the obtained nanocomposites.
Figure 2 shows the FTIR for the four nanocomposites and the DA used in the synthesis. SNP and DA spectra show a broad peak around 3400 cm
−1, which indicates a hydrogen bond due to the existence of hydroxyl (-OH) and (H
2O) adsorbed on the surface of the nanocomposite [
17,
18]. The other common peak between SNP and DA spectra is located around 1630 cm
−1 and is related to the hydroxyl group (-OH) [
17]. The disappearance of the broad peak at 3400 cm
−1 in the SNP-0.5DA, SNP-1.0DA, and SNP-1.5DA indicated that the hydroxyl groups (-OH) have been involved in the reaction with DA during the synthesis process. Similarly, the band at 980 cm
−1 in the SNP and DA spectra corresponds to the stretching vibration of Si-OH, this band decreased by increasing the amount of DA added to the synthesis mixture of SNP-0.5DA, SNP-1.0DA, and SNP-1.5DA due to the consumption of the silanol groups in the reaction [
10]. The peak that appears in all samples at 1050 cm
−1 is related to the C-O stretch [
17]. The peaks around 580 cm
−1 and 765 cm
−1 in the four nanocomposites are attributed to the Si-O- Si stretching vibration [
7] and bending vibrations [
9]. The possible reason for the undetected band at 3400 cm
−1 of the N-H stretching vibrations is the strong interaction between the terminal NH
2 of the aminopropyl groups and the unreacted surface hydroxyls on the silica [
19]. A similar observation was reported in the literature with no detection of N-H stretching vibrations of amine-modified silica [
20,
21,
22]. The weak band centered at around 1450 cm
−1 appeared and was assigned to the NH
2 deformation mode of the amine group [
20].
The thermal stability of the nanocomposites was investigated via thermogravimetric analysis (TGA).
Figure 3 shows the mass loss of solid for each nanocomposite as a function of temperature. It shows that the four nanocomposites are very stable until 250 °C with a maximum loss that did not exceed 5% for SNP and SNP-0.5DA and 7% for SNP-1.0DA and SNP-1.5DA. The mass loss during this stage is mainly related to the loss of weakly adsorbed molecules on the surface such as N
2, CO
2, and H
2O. As the temperature increased above 250 °C, a clear difference between the four nanocomposites can be noticed. For SNP, the loss in this stage is mainly related to the loss of water strongly attached to the surface or within the pores in addition to the dissociation of any surfactant or co-surfactant residuals within the sample with a total mass loss of 18%. In addition, SNP-0.5DA, SNP-1.0DA, and SNP-1.5DA lost more mass (25%, 30%, and 37%, respectively) due to the decomposition of the amino group, which extends to 600 °C. The higher the amount of DA added to the synthesis mixture, the higher the DA content of the prepared nanocomposite, and accordingly more mass loss was obtained by the TGA.
Figure 4 shows the results of the dynamic light scattering analysis (DLS). It is obvious that increasing the amount of DA added to the synthesis mixture resulted in larger hydrodynamic particle size with a broader distribution. The center of particle size distribution has been shifted from ≈34 nm for SNP to ≈168, 235, and 295 nm for SNP-0.5DA, SNP-1.0DA, and SNP-1.5DA, respectively, in agreement with the results obtained by SEM to be discussed later. This may be explained as adding the DA to the synthesis mixture increases the condensation of silica resulting in a higher growth rate and thus larger particle size. It is also possible that the higher DA content increases the tendency of the particles to aggregate, as will be discussed later with SEM and BET analysis.
The N
2 adsorption profile shown in
Figure 5 indicated that the porous structure of the nanocomposites is also significantly affected by the amount of DA used for the preparation (
Table 1). The SNP isotherm reveals the presence of mesoporous (2 nm < pores < 50 nm) with the largest surface area (26.95 m
2g
−1) and pore volume (0.1823 cm
3g
−1), which is related to the amorphous nature and low degree of aggregation for this sample [
23]. Adding DA to the synthesis mixture resulted in nonporous structure for both SNP-0.5DA and SNP-1.0DA with extremely small specific surface area and pore volume, as shown in
Table 2. On the other hand, nanocomposite prepared with 1.5 mL DA exhibits isotherms with open porous structures (
Figure 5) and a specific surface area of 9.836 m
2g
−1 and pore volume of 0.0768 cm
3g
−1. The drastic decrease in the specific surface area and pore volume of both SNP-0.5DA and SNP-1.0DA is related to the higher degree of aggregation in these two samples, which resulted in the loss of the majority of the pores. The effect of adding DA to the synthesis mixture on the porous structure of the nanocomposites is in agreement with the effect on the adsorption efficiency, to be discussed later since the availability of high surface area with accessible active sites within the pores is one of the most important factors affecting the efficiency of adsorption [
24].
The surface charge of SNP-DA is a result of the protonation or deprotonation of surface OH and NH
2 groups. Accordingly, the pH
0, which is the pH value of the solution corresponding to zero surface charge, is the dependence of adsorption of Zn
+2 on the solution pH [
25].
Figure 6 shows the ∆pH (pH
i-pH
e) versus pHi equilibrium curve. The x-intercepts of these curves represent the pH
0 of SNP, SNP-0.5DA, SNP-1.0DA, SNP-1.5DA which are 4.9, 6.3, 6.6, and 7.0, respectively, as recorded in
Table 2. The pH
0 of SNP is a higher-than-expected value for SiO
2 (around 2.5), which may be related to the presence of some surfactant residual within the silica structure since mild extraction was used for the removal of the surfactant rather than calcination. This is supported by the 18% mass loss of SNP at temperature higher than 250 °C (
Figure 3). Adding DA to the synthesis mixture resulted in shifting the pH
0 to a higher value, and also this pH
0 increased by increasing the amount of DA added. A similar observation about the zero-point charge of the silica and the effect of introducing amine was reported in the literature [
26]. The obtained pH
0 values imply that in aqueous solution with a pH less than pH
0, but higher than pK
b of the DA (around 4), the surface will be rich with ≡SiOH
2+ species, while at any pH lower than the pK
b, the SNP-DA surface will bear a more positive charge as a result of the protonation of both OH and NH
2 groups. Thus, the material will not be an effective Zn
+2 adsorbent as a result of repulsive forces. On the other hand, the SNP-DA will bear a negative charge if placed in a solution with a pH higher than 7.0 due to the deprotonation of the OH groups. Accordingly, and since the target of this work was to adsorb the Zn
+2 cations, all the adsorption tests were performed after adjusting the pH at 7.2 to ensure a negatively charged surface and at the same time to prevent the precipitation of Zn(OH)
2. Ali et al. (2020) prepared thiosemicarbazide-modified nanosilica and reported a pH
ZCP of 6.5 [
27]. They applied the sample for the adsorption of copper from a solution and reported maximum removal efficiency of 94% at a pH of 7.0.
Figure 7 shows the SEM images for the prepared samples. SNP sample shows almost spherical particles with narrow size distribution and some tendency to agglomeration. After incorporating the DA within the nanocomposite structure, a higher degree of aggregation resulted in irregular particle shapes and wider size distribution, in agreement with the DLS analysis discussed earlier. This higher degree of aggregation obtained after adding DA may be related to the enhanced hydrophilicity of the surface in addition to the tendency to form hydrogen bonds between surface hydroxyl groups and the amine groups attached to the surface. This growth of particle size and aggregation resulted in lower surface areas and pore volume, as obtained by nitrogen adsorption. Accordingly, less surface area will be available for adsorption, as will be next discussed.
Figure 7.
The SEM images of the nanocomposites were prepared with different amounts of DA at two different magnifications.
Figure 7.
The SEM images of the nanocomposites were prepared with different amounts of DA at two different magnifications.
Table 2.
Characteristics of the prepared nanocomposites.
Table 2.
Characteristics of the prepared nanocomposites.
Sample | pH0 | BET (m2g−1) | Pore Size PJH (nm) | Pore Volume PJH (cm3g−1) | TGA Mass Loss (%) | Morphology |
---|
SNP | 4.9 | 26.95 | 25.09 | 0.182 | 18.2 | Spherical |
SNP-0.5DA | 6.3 | 0.848 | 56.58 | 0.023 | 25.7 | Irregular |
SNP-1.0DA | 6.6 | 3.982 | 23.14 | 0.032 | 29.7 | Irregular |
SNP-1.5DA | 7.0 | 9.836 | 15.65 | 0.077 | 36.9 | Irregular |
Figure 8a shows the effect of initial Zn
2+ concentration on the adsorption capacity that was achieved by each sample. As the initial concentration increase, the driving force for diffusion of ions from the bulk solution to the adsorbent surface increase, and accordingly higher adsorption can be reached till equilibrium is established and all the available adsorption sites are being occupied.
Figure 8a shows that the equilibrium adsorption capacities follow the following order SNP-1.5DA > SNP > SNP-0.5DA > SNP-1.0DA. The maximum adsorption capacity was achieved by SNP-1.5DA, which could be attributed to the high surface area and pore size (
Table 2), and a surface rich with DA groups that are accessible for Zn
+2. Samples SNP-0.5DA and SNP-1.0DA have lower adsorption capacity than SNP despite the existence of the DA. This is mainly related to the aggregation that occurred in these samples which resulted in very low accessibility of the majority of the DA groups within the blocked pores. On the other hand, SNP sample showed a better adsorption capacity since it has the highest surface area, and its surface is rich with OH groups as shown by FTIR in
Figure 2. These OH groups were deprotonated since the pH of the solution was higher than the pH
0, and, accordingly, the surface will be negatively charged and attract the positive Zn
+2 cations.
Figure 8b represents the effect of increasing the initial concentration on the removal efficiency. As the concentration increases the removal efficiency decreases as a result of the saturation of the surface with Zn
+2. Sample SNP-1.5DA was able to 100% remove Zn
+2 from a 20-ppm solution, 98% from a 40-ppm solution, and 95% from a 60 ppm solution. While SNP, SNP-0.5DA, and SNP-1.0DA removed 89%, 74%, and 73% from the 20-ppm solution, respectively.
Figure 8c,d show the change of concentration and removal efficiency as a function of time. For samples SNP-0.5DA, and SNP-1.0DA there was a very fast drop in C
t and an increase in R% with the equilibrium reached in about 5 min., which is mainly because these samples have a high degree of aggregation, low surface area, and, accordingly, all the accessible adsorption sites (DA groups) are located on the external surface area, so there was no need for diffusion within the porous structure. For the other two samples, three stages are shown, a fast drop in C
t and an increase in R% during the first 5 min, which is related to the adsorption achieved by the active sites available on the external surface area of the samples. The second stage (5–90 min) was slower since adsorption is taking place on the active sites within the pores, and thus more time is needed for diffusion. The third stage is the equilibrium stage (after 90 min) where both C
t and R% remained constant as a result of the occupation of all active sites. Even though SNP has a higher surface area than SNP-1.5DA, the latter one showed faster kinetic than the SNP. This may be attributed to the effect of DA, which is expected to increase the affinity toward Zn
+2.
Equilibrium data were fitted with the Langmuir model while the kinetic data were fitted with the pseudo-second-order model following the nonlinear regression with the aid of Origin 2020 software, and the fitting results are shown in
Figure 8e,f and
Table 3.
Figure 8e,f show a good agreement between the experimental data (points) and the models (dashed lines) with the regression parameters shown in
Table 3. For the two models and all the samples, the obtained R
2 was greater than 90, indicating the feasibility of the models used. Langmuir’s model usually suggests monolayer adsorption with a higher chance of chemical rather than physical adsorption. On the other hand, the pseudo-second-order model suggests that each Zn
+2 occupies two DA groups. This suggests another reason for the low adsorption capacity obtained by SNP-0.5DA and SNP-1.0DA. Since the amount of DA in these samples is lower than in SNP-1.5DA, as suggested by the TGA analysis (
Figure 3), the distribution of the DA groups on the surface may not be crowded enough to give two DA groups close enough to hold one Zn
+2. The Langmuir constant (K
L) indicates the extent of interaction between the Zn
+2 and the adsorbents. The K
L value of SNP-1.5DA is the highest. This is due to the high content of NH
2 groups that are accessible to the Zn
+2, leading to the highest q
m value of 136 mgg
−1. On the other hand, SNP-0.5DA and SNP-1.0DA have lower K
L and q
m values than those of SNP despite their content of NH
2. This is mainly due to the restricted accessibility of these functional groups and the loss of surface area as a result of aggregation and pore blockage. The rate constant K
P2 shows that the fastest kinetics was obtained by SNP-1.0DA followed by SNP-0.5DA. This is mainly because these two samples have no or very limited porosity, and thus all the available adsorption sites are located on the external surface of the samples. This implies that for the Zn
+2 ions to reach these active sites they have to diffuse in the bulk solution only, while for SNP-1.5DA most of the active sites are located within the porous structure. Accordingly, the Zn
+2 ions need to diffuse first in the bulk solution, which is usually the fastest step. Then, it will diffuse within the pores to reach the internal NH
2 groups, and this is the slowest and most controlling step for the adsorption process.
Figure 8.
The dependence of equilibrium adsorption capacity on Zn+2 initial concentration (a), removal efficiency on Zn+2 initial concentration (b), Zn+2 concentration on time (c), equilibrium adsorption capacity on Zn+2 equilibrium concentration (d), adsorption capacity on time (e) and removal efficiency on time (f).
Figure 8.
The dependence of equilibrium adsorption capacity on Zn+2 initial concentration (a), removal efficiency on Zn+2 initial concentration (b), Zn+2 concentration on time (c), equilibrium adsorption capacity on Zn+2 equilibrium concentration (d), adsorption capacity on time (e) and removal efficiency on time (f).