**E**ff**ect of Initial Salt Composition on Physicochemical and Structural Characteristics of Zero-Valent Iron Nanopowders Obtained by Borohydride Reduction**

#### **Denis Leybo \*, Marat Tagirov \*, Dmitry Arkhipov, Elizaveta Permyakova, Evgeny Kolesnikov and Denis Kuznetsov**

National University of Science and Technology "MISiS", 119049 Moscow, Russia; arhipov.di@misis.ru (D.A.); permyakova.es@misis.ru (E.P.); kolesnikov.ea@misis.ru (E.K.); dk@misis.ru (D.K.)

**\*** Correspondence: leybo.dv@misis.ru (D.L.); m1701562@edu.misis.ru (M.T.); Tel.: +7-499-237-2226 (D.L. & M.T.)

Received: 4 September 2019; Accepted: 18 October 2019; Published: 21 October 2019

**Abstract:** The effect of initial salt composition on characteristics of zero-valent iron nanopowders produced via borohydride reduction was studied. The samples were characterized by X-ray diffraction, scanning and transmission electron microscopy, and low-temperature nitrogen adsorption. The efficiency of Pb2<sup>+</sup> ions removal from aqueous media was evaluated. The use of ferric salts led to enhanced reduction kinetics and, consequently, to a smaller size of iron particles in comparison with ferrous salts. A decrease in the ionic strength of the synthesis solutions resulted in a decrease in iron particles. The formation of small highly-reactive iron particles during synthesis led to their oxidation during washing and drying steps with the formation of a ferrihydrite phase. The lead ions removal efficiency was improved by simultaneous action of zero-valent iron and ferrihydrite phases of the sample produced from iron sulphate.

**Keywords:** nano zero-valent iron; borohydride reduction method; wastewater treatment; iron nanopowders; lead ions

#### **1. Introduction**

Wastewater is one of the main challenges of humanity in the 21st century. The fast growth of the population leads to the diminution of clean water reserves both because of increased global utilization and worsening of the ecological situation as a result of a constant increase in production. As mentioned in the United Nations (UN) World Water Development report [1], water demand is continuously increasing by 1% annually, and is expected to reach 6000 km<sup>3</sup> per year in 2050. About 80% of the water used by industry is released without any treatment [2], which puts the ecosystem and human health at risk.

Several chemical species have been identified as potentially harmful to human health. Among them are nitrates, fluorides, arsenic, and selenium, which constitute the main part of inorganic pollutants found in many locations around the world and are considered as the main priority chemicals for wastewater treatment technologies [3]. Other substances threatening human health include recalcitrant organic dyes [4], chlorinated organic compounds [5], phenolic compounds [6], and heavy metals (e.g., Cr, and Pb) [7–9]. Lead ions fall into wastewater mainly from the mining industry, smelters, hydroelectric power plants, among others [10]. The presence of lead in drinking water leads to irreversible health effects that can affect every organ of the human body [11]. The recommended limit concentration was set in the guidelines for drinking water as 0.01 mg/L for Pb(II) [12]; however, in industrial wastewaters, the lead-ion concentrations approach 200–500 mg/L [13]. Even though producers try to lower or exclude lead from manufacturing processes, lead is still used and exerts detrimental impact on the ecosystem [14].

So far different wastewater treatment technologies have been suggested, e.g., adsorption [15], biosorption [16,17], membrane technology [18,19], biodegradation [20,21], ion exchange [22,23], ionic liquid-based treatment [24]. Sorption methods have several advantages, such as high efficiency uptake for heavy metal ions at low concentrations, energy saving and easy sorbent synthesis [25]. Among these technologies are nano zero valent iron (nZVI)-based techniques that have gained much attention of some scientific groups [26]. nZVI particles showed superior efficiency in the removal of antibiotics [27], phenolic compounds [28], heavy metals [29,30], inorganic pollutants [31], and organic dyes [32,33]. Our previous studies showed that nZVI particles are very effective for the decomposition of azo dyes [34] and as a fenton-like zinc removal catalyst [35]. nZVI have a higher reaction rate because of their very high specific surface area and surface energy for the redox process [36]. The small particle size significantly increases the diffusion properties of the materials, reduces the required amount of sorbent [37], and reduces the required contact time between nZVI and water to prevent oxidation during the synthesis [38].

The primary method for the preparation of standalone and supported nZVI particles used by scientific groups is the borohydride reduction of iron ions [39]. Although this method has been used for several decades, it is still insufficiently studied in terms of different parameters that influence the properties of the final product. R. Eljamal et al. [40] reported the results of a comprehensive study of eight relevant experimental parameters, such as concentration, reaction temperature, pH, mixing speed, and reaction time on composition, particles' size and the surface area of the produced nZVI. Their findings resulted in some improvement of nitrate reduction efficiency by 27% and phosphorus adsorption by 9.5%. N. Goldstein et al. [41] studied the influence of the organophosphate stabilizers on nZVI particles' size and zeta potential. H. Woo et. al. [42] investigated the effect of a washing solution and drying condition on the reactivity of nanosized nZVI synthesized by borohydride reduction. However, up to date, there has been no information in the literature on the effect of initial salt composition on the properties of nZVI powders produced by the borohydride reduction method. Although the authors of [41] reported differences in the stabilization mechanisms using different organic additives for ferrous and ferric salts, it is still unclear how these salts affect the borohydride reduction of iron.

Another issue related to the method is the need to use a protective atmosphere to avoid oxidation of the as-prepared nZVI particles. This concern imposes restrictions on the widespread use of the method because of the increased complexity of the overall process. Thus, it is highly desirable to identify conditions that will allow producing nZVI in an ambient atmosphere.

Our research investigated the effect of ferric and ferrous salts choice on structural characteristics and the physicochemical properties of nZVI samples synthesized by borohydride reduction. The obtained samples were characterized by X-ray diffraction, scanning and transmission electron microscopy, and low-temperature nitrogen adsorption techniques. A tentative scheme of the initial salt effect was proposed based on the experimental results obtained in this work. The nZVI samples were evaluated in the process of Pb2<sup>+</sup> ion removal from aqueous solution.

#### **2. Materials and Methods**

#### *2.1. Materials*

All the reagents (FeSO4·7H2O, Fe(NO3)3·9H2O, FeCl2·4H2O, FeCl3·6H2O, sodium carboxymethylcellulose (C6H7O2(OH)2CH2COONa), and NaBH4) used in this work were of analytical grade and used without any additional purification. MilliQ water and absolute isopropanol were used for the synthesis and water purification experiments.

#### *2.2. nZVI Synthesis*

All the synthesis procedures were performed in an ambient atmosphere. The nZVI samples were obtained by the reduction of iron ions from different salts according to the following reactions [40,43]:

$$\rm FeSO\_4 + 2NaBH\_4 + 6H\_2O = Fe + Na\_2SO\_4 + 2B(OH)\_3 + 7H\_2 \tag{1}$$

$$\rm Fe(NO\_3)\_3 + 3NaBH\_4 + 9H\_2O = Fe + 3NaNO\_3 + 3B(OH)\_3 + 10.5H\_2 \tag{2}$$

$$\text{FeCl}\_2 + 2\text{NaBH}\_4 + 6\text{H}\_2\text{O} = \text{Fe} + 2\text{NaCl} + 2\text{B(OH)}\_3 + 7\text{H}\_2\tag{3}$$

$$\rm FeCl\_3 + 3NaBH\_4 + 9H\_2O = Fe + 3NaCl + 3B(OH)\_3 + 10.5H\_2 \tag{4}$$

In detail, 2.5 g of FeSO4·7H2O/4 g of Fe(NO3)3·9H2O/2 g of FeCl2·4H2O/2.7 g of FeCl3·6H2O were dissolved in 30 mL of deionized water. 30 mL of 10 wt. % NaBH4 aqueous solution was added drop-by-drop for 10 min. Sodium borohydride was taken excessively to provide a complete reduction. The suspension was mixed during the next 30 min.

The precipitate was filtered through a 0.22 μm membrane filter; afterwards, it was washed from by-products with 250 mL of water and 250 mL of isopropanol, sequentially. The obtained sample was dried in a drying oven under vacuum at room temperature. For the sample obtained from iron nitrate, a possibility of passivating the nanoparticles' surface under the flow of 1% O2/N2 was also studied. For this experiment, the washing of the sample was performed in a Drexel flask with further drying under the flow of O2/N2 overnight. The washing of the sample, in this case, was done by decantation using a magnet to increase the sedimentation speed.

The reduction process was also performed in the presence of sodium carboxymethylcellulose (Na-CMC) [44] with Fe:Na-CMC molar ratio equal to 1:0.002, 1:0.0055, and 1:0.0108 to passivate the surface of iron nanoparticles obtained from Fe(NO3)3·9H2O by the reaction (2). The precipitate was separated by vacuum filtration followed by washing with isopropanol (250 mL) and drying under vacuum at room temperature as in the previous experiments.

#### *2.3. Characterization*

The X-ray diffraction analysis (XRD) of the samples was done on Diffray 401 diffractometer (Scientific Instruments, Russia) operating with Cr-Kα radiation (λ = 2.2909 Å). The crystallites' size was estimated using the Scherrer equation, and the Si standard was used for instrumental broadening elimination. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) were used to study the morphology of the obtained particles. The SEM analysis was done on Tescan Vega 3 SB instrument, and the LEO 912 ab omega was used for the TEM analysis. The specific surface area analysis was performed using a low-temperature N2 adsorption method with the Nova 1200e analyzer (Quantachrome, Boynton Beach, FL, USA) at 77 K. Before the analysis, the samples were degassed in vacuum at 120 ◦C for 3 h. The specific surface area was calculated using the Brunauer–Emmett–Teller (BET) equation.

The pH zero charge point (pHpzc) was measured using the pH drift method [45,46]. Samples of 2.5 mg nZVI were introduced into a known volume (10 mL) of a solution of 0.1 M KNO3. KNO3 was selected as an inert electrolyte. The initial pH (pHInitial) of the KNO3 solutions was adjusted from 3 to 12 by adding 0.1 M HCl or NaOH solutions. Suspensions were left to balance for 24 h in closed vessels at room temperature. Then, the pH values (pHfinal) were measured again. Based on the data obtained, we plotted the pHfinal vs. pHinitial dependence. pHpzc was determined by the intersection with the pHinitial = pHfinal line.

#### *2.4. Pb2*<sup>+</sup> *Ions Removal*

The water purification experiments were performed by measuring the concentration of Pb2<sup>+</sup> ions in an aqueous solution in the presence of the nZVI samples reduced from different salts. In detail, 200 mL of 1.2 mM Pb(NO3)2 solution was mixed with 50 mg of nZVI. The suspension was processed in an ultrasonic bath for 10 min for homogenization. After the ultrasonic treatment, the suspension was left in a shaker for 230 min. The probes of the solutions were taken in 1-mL portions after 10 min of the ultrasonic treatment and after 10, 20, 50, 80, 110, 140, 170, 200, and 230 min of mechanical mixing for our further analysis of the Pb concentration. The concentration of Pb in the supernatant solutions was determined by inductively coupled plasma mass spectrometry on ICAP Q instrument (Thermoscientific) in triplicates.

The lead removal capacity q (mg·g<sup>−</sup>1) was calculated for the samples after treatment for 240 min by the following equation:

$$\mathbf{q} = \frac{(\mathbf{C}\_0 - \mathbf{C}\_\pi) \cdot \mathbf{V}}{\mathbf{m}} \tag{5}$$

where C0 is the initial concentration of Pb2+, C<sup>τ</sup> is the concentration of Pb2<sup>+</sup> after 240 min of the experiment, V is the volume of the solution, and m is the mass of nZVI.

#### **3. Results and Discussion**

#### *3.1. The E*ff*ects of Initial Salt on the Phase Composition of the nZVI Samples*

The X-ray diffraction patterns of the powders obtained after the reduction of iron with sodium borohydride from FeSO4, FeCl2, and FeCl3 had characteristic peaks at 2θ angles equal to 68.6◦ and 106.0◦, corresponding to metallic iron (Figure 1, PDF card no. 06-0696). The data confirmed that iron precipitated in a zero valence state. The sample obtained from FeSO4 also contained the 2-line ferrihydrite (Fh) phase, Fe10O14(OH)2 [47], which was confirmed by the presence of the broad peaks at 2θ angles equal to 52.6◦ and 99.5◦ [48]. nZVI could not be obtained by reducing Fe3<sup>+</sup> ions from Fe(NO3)3 with NaBH4. This sample consisted of only the Fh phase. The XRD patterns show that the iron particles reduced from FeCl2 had the highest crystallinity (Figure 1), which was confirmed by the lowest full width at half maximum (FWHM) value.

**Figure 1.** The X-ray diffraction analysis (XRD) patterns of powders obtained by reduction from FeCl2 (**a**); FeCl3 (**b**); FeSO4 (**c**); Fe(NO3)3 (**d**).

The XRD pattern in Figure 1 shows that the best results were obtained by reducing iron from its chlorides. The content of the oxidized products in the final samples obtained from different salts decreased in the series Fe(NO3)3 > FeSO4 > FeCl3 ≈ FeCl2.

The ferrihydrite phase is usually synthetically produced by hydrolyzing ferric or ferrous salts [48]. Our observations on suspension's color change during the reduction of the iron nitrate salt suggested

that the formation of the Fh phase occurred after the end of the reduction process, which means that this was a secondary product of the iron particles oxidation. H. Pullin et al. [49] observed the formation of the Fh phase as a corrosion product of the nZVI particles in aqueous media. The authors stated that the formation of 2-line ferrihydrite, being intermediate to more stable iron oxide and oxohydroxide phases, was detectable during the entire experiment lasting for more than four weeks. Similar results were obtained by A. Liu et al. [50] during their study of the nZVI composition evolution in static water after ten days of the experiment.

The corrosion rate of nZVI is inversely proportional to the size of the particles comprising the powder. The results obtained in our work indicate the dependence of the final product composition on the initial salt used for the production of nZVI, which could be attributed to the different sizes of the particles formed during the synthesis. Thus, the use of iron nitrate led to the formation of extremely reactive small iron particles stable only under reductive conditions. The iron sulphate-derived sample only partially consisted of the Fh phase, which suggested the formation of more stable and larger particles during the borohydride reduction; whereas the use of iron chlorides allowed producing the pure iron phase.

During the reduction of the iron ions from the nitrate salt, the color change of the suspension from black to rusty was observed during the drying step. Thus, an additional experiment on the passivation of the surface of the iron particles reduced from nitrate during drying using an O2/N2 mixture was conducted to protect against oxidation. The XRD pattern of this sample is shown in Figure 2a. It can be seen from the results that the only crystalline phase present in the sample was sodium borate. Sodium borate forms as a by-product of the sodium borohydride hydrolysis reaction [51] and are usually removed during the washing procedure. The decantation is less efficient for the sample purification. The rusty color of the sample suggested that it should contain iron-based compounds, probably, in a low crystallinity form. To confirm this, the energy dispersive X-ray (EDX) analysis of the sample was done. The results (Figure 2b) show that the sample contained iron, most probably, in the form of the amorphous Fh phase. Therefore, this route is inappropriate for the production of nZVI.

**Figure 2.** The XRD pattern (**a**) and the energy dispersive X-ray (EDX) analysis (**b**) of the powder obtained by reduction from Fe(NO3)3 with passivation in a current of 1% O2/N2 in a Drexel bottle.

The second route for stabilizing the surface of the iron particles reduced from the nitrate salt in this work was the usage of Na-CMC as an additive. Sodium CMC is known to be soluble in water and poorly soluble in alcohols. Thus, changing the solvent during the washing procedure could result in the formation of the Na-CMC protective coating on the surface of the iron particles. The XRD pattern in Figure 3 shows that all the samples obtained using the sodium CMC as a passivating agent consisted of ferrihydrite phase. The results obtained in this work show that it is impossible to produce nZVI with a sodium borohydride reduction of the nitrate salt using the abovementioned techniques.

**Figure 3.** The XRD patterns of the powders obtained by reduction from Fe(NO3)3 with a concentration of CMC 1:0.002 (**a**), 1:0.0055 (**b**), and 1:0.0108 (**c**).

#### *3.2. The E*ff*ects of Initial Salt on Structural Characteristics of Iron-Based Samples*

#### 3.2.1. SEM Results

The SEM images of nanoparticles are shown in Figure 4. The nZVI particles reduced from FeSO4 formed aggregates of complex shapes (Figure 4a). The SEM images of the powder reduced from Fe(NO3)3 showed aggregates of the particles with an extended surface (Figure 4b). The image revealed that these aggregates consist of flake-shaped particles. The particles of the iron sample produced from FeCl2 and FeCl3 were arranged in chains (Figure 4c,d). The particle size of the sample produced from FeSO4 was approximately 50–60 nm, the cross-section particle size of the sample produced from FeCl2 was approximately 70–90 nm and of FeCl3 30–45 nm. The aggregates of the particles obtained from Fe(NO3)3 had sizes in a range from 500 nm to 1 μm.

**Figure 4.** The scanning electron microscopy (SEM) images of nZVI produced from FeSO4 (**a**), Fe(NO3)3 (**b**), FeCl2 (**c**), FeCl3 (**d**).

#### 3.2.2. TEM Results

The TEM images of the nanoparticles are shown in Figure 5. The aggregates of the nZVI particles obtained from FeSO4 had a developed morphology (Figure 5a). The sample consisted of flake-like particles as well as particles of an irregular form with a size of around 20 nm. The TEM images of the powder produced from Fe(NO3)3 showed that two-dimensional structures with a thickness of about 2 nm were obtained (Figure 5b). The nZVI particles obtained from FeCl2 and FeCl3 were arranged in chains and had a core-shell structure with a shell thickness of about 3–6 nm (Figure 5c,d). The particles' size of the iron samples reduced from FeCl2 was approximately 100 nm, and from FeCl3 was 20–30 nm.

**Figure 5.** The transmission electron microscopy (TEM) images of nZVI produced from FeSO4 (**a**), Fe(NO3)3 (**b**), FeCl2 (**c**), FeCl3 (**d**).

The results of the morphology study of the nZVI samples obtained in this work are well correlated with their composition and the results obtained by other authors. The 2D structures produced using the iron nitrate salt are characteristic of ferrihydrite being a poorly crystalline iron compound with the crystallites' size of 2–6 nm [52–54]. The structures similar to the obtained ones via the reduction of iron sulphate in this work were also observed by different researchers [55,56]. Mainly, these structures were assigned to the presence of iron oxides [57] on the surface of the Fe<sup>0</sup> particles. Our results showed that these structures correspond to the formation of ferrihydrite via oxidation of small reactive iron particles. The core-shell structured chains of the nZVI aggregates are the most commonly encountered morphology of iron produced by a borohydride reduction process [30,58,59]. The shell composition according to the XPS analysis was shown to be Fe2O3 [59], the most stable iron oxide, comprising the passivation coating preventing their fast oxidation upon exposure to air/water environment.

#### 3.2.3. Specific Surface Area and Crystallites' Size

The specific surface area and the crystallites' sizes of the iron samples reduced from various salts are presented in Table 1.


**Table 1.** The characteristics of the recovered powders.

The powder produced from iron nitrate had the largest surface area among the samples. The smallest specific surface area was identified for iron particles reduced from FeCl2. The obtained results are in good agreement with the particles' size and the morphology observations visible on the SEM and TEM images. Thus, the 2D morphology of the particles produced from Fe(NO3)3 salt resulted in the highest surface area. The presence of such two-dimensional structures in the sample produced from sulphate was also responsible for the high surface area of this sample. Because both samples containing 2D structures consisted of Fh, we could conclude that the particles of such morphology corresponded to the Fh phase. In the case of the iron chloride salts, the largest specific surface area was obtained for the powder produced from FeCl3, which is well correlated to the smaller size of the particles produced from this salt.

The crystallites' size for the iron particles reduced from FeSO4 had the smallest value among the samples although it could be underestimated because of the presence of the amorphous phase in the sample. The powders obtained from FeCl2 had the largest crystallites' size, almost ten times larger than those reduced from the FeCl3 salt. The comparison of these values with a transverse size of the particles measured from the SEM and TEM images led us to the conclusion that the chain aggregates produced from the iron chloride salts consisted of single-crystal domains of iron.

#### *3.3. The Scheme of Initial Salt Composition E*ff*ect on Reduction of Fe Ions*

The sodium borohydride reduction of iron was performed using Fe3<sup>+</sup> and Fe2<sup>+</sup> containing salts in present work according to the reactions [55,58]:

$$\rm{Fe^{2+} + 2BH\_4^- + 6H\_2O = Fe^0 + 2B(OH)\_3 + 7H\_2} \tag{6}$$

$$\rm Fe^{3+} + 3BH\_4^- + 9H\_2O = Fe^0 + 3B(OH)\_3 + 10.5H\_2 \tag{7}$$

The standard reduction potentials of these ions are different, with Fe2<sup>+</sup> having a more negative value of <sup>−</sup>0.447 V in comparison with <sup>−</sup>0.037 V for Fe3<sup>+</sup> ion. Consequently, it is easier to reduce ferric ion [60] from the solution. Since we can expect that iron ions form identical transition complexes [43], the rate constant of the reaction (6) is higher, which leads to the formation of smaller particles using ferric ion as the source of iron.

Among the salts containing identical iron ions the difference in the properties of the produced iron powders originates from the corresponding anions. Although the anions do not participate in the reactions (5) and (6) directly, they do form the ionic atmosphere that influences the reaction rate. According to the Brönsted–Bjerrum equation [61], an increase in the ionic strength of the solution results in a decrease of the reaction rate constant when the reacting ions have different signs. Thus, the formation of smaller iron particles is more favorable in the solutions of lower ionic strength.

Taking into account the abovementioned qualitative relationships, a tentative scheme of the initial salt composition effect on the reduction of Fe ions is proposed in this work (Figure 6). According to the scheme, the use of ferric ion-containing FeCl3 and Fe(NO3)3 salts leads to the formation of smaller particles during the reduction process as compared to the FeCl2 and FeSO4. The ionic strength of the solutions used in this work was calculated to be equal (in mol/L) to 1.8, 2.0, 2.4, and 2.6 for FeSO4, FeCl2, Fe(NO3)3, and FeCl3 initial salts, respectively. The iron particle formation rate is higher for the FeSO4 and Fe(NO3)3 initial salts due to their lower ionic strengths. After the formation of nZVI during the washing and drying steps, small highly reactive particles are oxidized by water and dissolved oxygen with the formation of the Fh phase in the case of Fe(NO3)3 and, partially, FeSO4 salts. Larger nZVI particles produced from the chlorides agglomerate in the chain structures with the formation of a passivation oxide layer on their surface.

**Figure 6.** The scheme of the nZVI production via a borohydride reduction of different salts.

#### *3.4. The E*ff*ects of Initial Salt Composition to Pb2*<sup>+</sup> *Ions Removal E*ffi*ciency*

The results of the removal of Pb2<sup>+</sup> ions removal are illustrated in Figure 7. The horizontal axis shows the total processing time as the sum of the ultrasonic and mechanical treatment. The concentration of the Pb2<sup>+</sup> ions rapidly decreased after the ultrasonic treatment for the nZVI samples obtained from all the salts. Then there was a gradual decrease in the concentration of all the samples within the next 120 min. There was a relatively slight fluctuation after 120 min of the experiment. The lead removal capacity q was equal to 744, 646, 462, and 650 mg·g−<sup>1</sup> for the nZVI samples obtained from FeSO4, Fe(NO3)3, FeCl2, and FeCl3, respectively, after the treatment for 240 min. Interestingly, the nZVI sample obtained from FeSO4 showed maximum efficiency in the removal of the Pb2<sup>+</sup> ions as compared to other nZVI samples.

Table 2 shows the comparison of the lead removal capacity results obtained in the present study and by other authors. Zhang Dongsheng et al. [30] gave q calculations for the removal of lead from the obtained nZVI and nZVI powders after freeze-drying. Chuang Yu et al. [62] reported the results of the adsorption of lead ions by GO-B-nZVI composite. Humera Jabeen et al. [63] in their work compared the adsorption capacity q for the synthesized powders of nZVI and composites based on graphene and nZVI (G-nZVI).

**Figure 7.** Pb2<sup>+</sup> ions removal efficiency by the nZVI samples produced from different salts.


**Table 2.** Comparison of the lead removal capacity results.

The comparison of the results with the characteristics of the samples obtained in this work reveals that there was a synergy between Fh and zero-valent iron, which led to the increased efficiency of the lead removal. Thus, the removal of the lead ions by the iron chloride originated samples was lower, which can be attributed to lower suspension stability of the nZVI particles in the solution as well as the lower specific surface area of these samples. These samples consisted of iron in a metallic form. The sample obtained from the iron nitrate had moderate efficiency. Although this sample did not contain zero-valent iron, its high specific surface area resulted in high adsorption capacity. The composition of the nZVI sample produced from sulphate ensured its high efficiency as a result of both positive actions of zero-valent iron capable of reducing lead ions [30] and ferrihydrite phase providing high surface area and the suspension stability of this sample.

Three mechanisms leading to the removal of lead ions from an aqueous solution are considered to be the main ones: sorption, reduction, and coprecipitation. During Pb2<sup>+</sup> ions removal experiments, the initial pH was 4.53 for all samples. The final pH also did not differ significantly and varied in the series 5.54, 5.52, 5.51, 5.48 for FeSO4, Fe(NO3)3, FeCl2, and FeCl3, respectively. An increase in pH is associated with the oxidation of nZVI in an acid solution. A relatively small change in pH during the

removal of lead ions indicates a slight contribution of the coprecipitation mechanism on the overall Pb ions removal.

An important characteristic of nZVI is the point of zero charge (pHpzc). pHpzc is the pH at which the number of positively and negatively charged centers on the surface of the powder is the same. If pH < pHpzc, part of the protons from the solution is adsorbed on the surface with the formation of a positive charge, part of the counterions (OH−) in the surface layer of the solution will be released and the pH of the contacting solution will increase, approaching pHpzc; otherwise, some OH<sup>−</sup> groups will sorb, and the surface will acquire a negative charge.

The results of determining pHpzc for nZVI samples are presented in Figure 8. The dashed line shows the position where pHinitial = pHfinal. pHpzc was determined by the intersection of the graphs with a dashed line. pHpzc were 7.9, 6.6, 9.1, and 8.8 for the nZVI samples obtained from FeSO4, Fe(NO3)3, FeCl2, and FeCl3, respectively. The pHpzc of the obtained nZVI samples are not favorable for the electrostatic adsorption of Pb2<sup>+</sup> cations. Therefore, adsorption also does not make a priority contribution to the removal of the lead ions.

**Figure 8.** pHfinal versus pHinitial during equilibration of nZVI powders in 0.1 M KNO3 solution.

The obtained results of measuring the pH of the solution and pHpzc of nZVI samples show that the processes of coprecipitation and electrostatic adsorption were not the main ones during the removal of Pb2<sup>+</sup> ions from the aqueous environment. Apparently, the reduction of Pb2<sup>+</sup> ions by nZVI made a predominant contribution to the process of water purification from lead ions.

#### **4. Conclusions**

Iron-based samples with different phase compositions, particles' size, and morphology can be obtained by a borohydride reduction of various ferric and ferrous iron salts. The use of ferric salts leads to the formation of smaller iron particles during the synthesis because of a higher value of the reduction potential and the enhanced reaction kinetics. The formation of such highly reactive particles results in their further oxidation by water and dissolved oxygen with the formation of 2D ferrihydrite. The oxidation cannot be avoided by O2/N2 passivation and Na-CMC assisted synthesis. Among the salts containing iron ions with the same oxidation state the difference in the corresponding nZVI samples' characteristics originated from different solution's ionic strength, which influences the reduction rate.

The synergistic action of zero-valent iron and ferrihydrite leads to superior Pb2<sup>+</sup> ion removal from water. This effect is easily achieved by the borohydride reduction of iron from sulphate salt. Pure ferrihydrite shows moderate lead ion removal efficiency despite the highest specific surface area among the samples since the adsorption is the only action mechanism. The phase pure zero-valent iron samples have the lowest Pb2<sup>+</sup> ions removal efficiency because of the lowest suspension stability and the adsorption capacity.

**Author Contributions:** Conceptualization, D.K.; formal analysis, D.L.; funding acquisition, D.A.; investigation, D.L. and M.T.; methodology, D.L., M.T., E.P., and E.K.; project administration, D.A.; supervision, D.K.; visualization, M.T.; writing–original draft, D.L. and M.T.; writing–review and editing, D.L., M.T., and D.K.

**Funding:** RFBR funded the reported study according to the research project No. 18-29-25051.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
