**3.1 Silver nanoparticles formation on the microfiltration membranes**

The recovery of Ag (I) was carried out using Ag+ 100 mg L-1 and HNO3 0.25 mol L-1 as the feed solution, and a solution of ascorbic acid (HA) 1 mol L-1 as the stripping solution. The contact time of the membrane with the feed and stripping solutions was 30 minutes, with a stirring speed of 600 rpm at both compartments.

After 30 minutes, the membrane was removed from the cell. The side of the membrane in contact with the feed solution showed a deposit, while any deposits were observed on the face in contact with the stripping solution. Figure 2a shows a part of the microfiltration membrane that was in contact with the feed solution. Figure 2b corresponds to an image obtained with the optical microscope.

Fig. 2. Images of (a) microfiltration membrane after contact with the feed (Ag (I)) and stripping (HA) solutions; and (b) small part of the membrane through an optical microscope

This deposit may have been caused by the formation of metallic silver by the reduction of Ag (I) with ascorbic acid. The deposit seems uniform both to the naked eye and through the optical microscope. In addition, the formation of silver on the membrane is consistent with the decrease concentration of Ag+ ions in the feed solution (see Figure 3). No concentration of Ag+ was detected in the stripping solution.

Silver Recovery from Acidic Solutions by Formation of

Nanoparticles and Submicroparticles of Ag on Microfiltration Membranes 445

A deeper more detailed analysis on the morphology of the deposits was carried out using Scanning Electron Microscopy (SEM). The micrographs of different membranes are shown in Figure 5. The silver particles formed, are distributed over the membrane and in its pores. The shape acquired by the metallic silver depends on the synthesis conditions. We observe a non-homogeneous 3D growth, with a hexagonal shape resemblance and a broad size

distribution. A similar result was recently reported by Masaharu et al., 2010.

(a) (b)

(a) (b)

and 800 rpm in the stripping solution

Fig. 6. Micrograph of silver particles on the membrane, same solution that in figure 5. (a) stirring speed 300 rpm, at both compartments. (b) stirring speed 1200 in the feed solution

(b) are the same sample with different magnification

Fig. 5. Micrograph of the silver particles on the membrane. Feed solution: 100 mg L-1 of Ag (I). Stripping solution: [HA] = 1 mol L-1. Stirring speed: 600 rpm at both compartments. (a) and

It is also evident, that the stirring speed has an impact on the size and location of the particles (on the surface or in the pores of the membrane). Figure 6 shows the micrographs of silver particles deposited on the microfiltration membrane at different stirring speeds.

Fig. 3. Variation of Ag (I) concentration as a function of time. ()Feed solution: [Ag(I)] = 100 mg L-1; [HNO3] = 0.25 mol L-1. () Stripping solution: [HA] = 0.25 mol L-1

Comparing the weight of the membrane before (m1 = 0.1447 g) and after the experiment (m2=0.1739 g), we found that the difference in weigh was 29.2 mg. While the initial concentration of Ag+ ions in the feed solution was 122 mg L-1 and the volume of both solutions was 250 mL, for each trial, the amount of metallic silver that could be deposited on the membrane is of 30.5 mg. This value is very close to the mass in excess of the membrane and it corresponds to the metallic silver deposited on the membrane. The yield of silver recovery in these conditions is around 96%. To confirm the presence of metallic silver on the membrane, an EDAX analysis was performed. Figure 4 shows the spectrum which indicates the peaks that correspond to metallic silver.

Fig. 4. EDAX spectrum of the membrane after contac twith the feed (Ag (I)) and stripping (HA) solutions

Fig. 3. Variation of Ag (I) concentration as a function of time. ()Feed solution: [Ag(I)] = 100 mg L-1; [HNO3] = 0.25 mol L-1. () Stripping solution: [HA] = 0.25 mol L-1

the peaks that correspond to metallic silver.

(HA) solutions

Comparing the weight of the membrane before (m1 = 0.1447 g) and after the experiment (m2=0.1739 g), we found that the difference in weigh was 29.2 mg. While the initial concentration of Ag+ ions in the feed solution was 122 mg L-1 and the volume of both solutions was 250 mL, for each trial, the amount of metallic silver that could be deposited on the membrane is of 30.5 mg. This value is very close to the mass in excess of the membrane and it corresponds to the metallic silver deposited on the membrane. The yield of silver recovery in these conditions is around 96%. To confirm the presence of metallic silver on the membrane, an EDAX analysis was performed. Figure 4 shows the spectrum which indicates

 keV Fig. 4. EDAX spectrum of the membrane after contac twith the feed (Ag (I)) and stripping

A deeper more detailed analysis on the morphology of the deposits was carried out using Scanning Electron Microscopy (SEM). The micrographs of different membranes are shown in Figure 5. The silver particles formed, are distributed over the membrane and in its pores. The shape acquired by the metallic silver depends on the synthesis conditions. We observe a non-homogeneous 3D growth, with a hexagonal shape resemblance and a broad size distribution. A similar result was recently reported by Masaharu et al., 2010.

Fig. 5. Micrograph of the silver particles on the membrane. Feed solution: 100 mg L-1 of Ag (I). Stripping solution: [HA] = 1 mol L-1. Stirring speed: 600 rpm at both compartments. (a) and (b) are the same sample with different magnification

It is also evident, that the stirring speed has an impact on the size and location of the particles (on the surface or in the pores of the membrane). Figure 6 shows the micrographs of silver particles deposited on the microfiltration membrane at different stirring speeds.

Fig. 6. Micrograph of silver particles on the membrane, same solution that in figure 5. (a) stirring speed 300 rpm, at both compartments. (b) stirring speed 1200 in the feed solution and 800 rpm in the stripping solution

Silver Recovery from Acidic Solutions by Formation of

required than those proposed here.

Nanoparticles and Submicroparticles of Ag on Microfiltration Membranes 447

solution), the particles maintain a certain homogeneity. A decahedron shape is observed in figure 7a. On the other hand, when the stripping solution was not stirred and the stirring speed on the feed solution was reduced (1000, 600, 300 rpm) the shape of the silver crystals gets very different. In the case of a stirring speed of 1000 rpm, the metallic silver takes crystal morphology in the shapes of cubes and rods. If the stirring speed is decreased to 600 rpm, the silver particles appear as rounded shapes but with traces of nucleation that form a cubic shape (figure 7b, 7c). At 300 rpm on the feed solution, the silver particles take decahedron shapes with an average particle size greater than in other cases. Also in this case, the large silver particles are occluded in the pores of the microfiltration membrane. It is important to observe that the proposed methodology is very suitable for obtaining metallic silver particles of different shapes and sizes. In fact, in the literature, in order to obtain different shape and size silver nanoparticles, more controlled and drastic conditions are

Finally, in the absence of stirring at both phases, the silver particles obtained on the membrane surface, clearly show the formation of hexagonal plates (Figure 8). These hexagonal plates come from the formation of dendrites on the surface, which is the first

stage in the process of crystallization of silver on the microfiltration membrane.

(a) (b)

**3.2 Effect of Ag (I)'s concentration on the recovery efficiency** 

solutions. (b) Higher magnification of the same sample

silver particles.

is under mass transfer of Ag+ ions.

Fig. 8. Micrograph of the silver particles on the membrane when no stirring is applied to the

We can conclude that hydrodynamics play an important role in the morphology and size of

In the next section we will discuss the chemical aspects that affect the process of reducing Ag+ ions by ascorbic acid. Additionally, we will analyze the conditions for efficient recovery of silver so we will have a better understanding of the mechanism under which the process

In order to evaluate the influence of Ag(I)'s concentration on the efficiency recovery of the proposed separation system, two tests with different concentrations of Ag(I) (25 and 100 mg L-1), were performed. For the first test we used a 0.25 mol L-1 of HNO3 as the feed solution

As in the previous conditions, the silver particles are distributed over the membrane. For a stirring speed of 350 rpm at both phases, the silver particles shape is well defined, showing a hexagonal 3D growth (figure 6a and 6b). For a higher stirring speed (Figure 6c and 6d), the shape of the silver particles is less homogeneous. As we will show later in this paper, the crystalline shapes of the silver particles are highly dependent on the nucleation speed and are based on hydrodynamic and chemical aspects. The hexagonal crystal plate shape has been reported earlier in the literature (Jixiang et al., 2007; Masaharu et al. 2010).

Figure 7 shows the micrographs of silver particles deposited on the microfiltration membranes when a low stirring speed (350 rpm) or no agitation is applied to the stripping solution and different stirring speeds applied to the feed solution. From this, it is clear that every stirring speed causes significant changes in the shape and size distribution of the silver particles. For a 350 rpm stirring speed in the stripping solution (to 600 rpm in the feed

Fig. 7. Micrograph of silver particles on the membrane obtained with different stirring speeds in the two compartments. (a) Stirring speed 350 rpm in the stripping solution and 600 rpm in the feed solution. (b), (c) and (d) No agitation of the stripping solution. To the feed solution a stirring speed of 1000, 600 and 300 rpm, is applied respectively

As in the previous conditions, the silver particles are distributed over the membrane. For a stirring speed of 350 rpm at both phases, the silver particles shape is well defined, showing a hexagonal 3D growth (figure 6a and 6b). For a higher stirring speed (Figure 6c and 6d), the shape of the silver particles is less homogeneous. As we will show later in this paper, the crystalline shapes of the silver particles are highly dependent on the nucleation speed and are based on hydrodynamic and chemical aspects. The hexagonal crystal plate shape has

Figure 7 shows the micrographs of silver particles deposited on the microfiltration membranes when a low stirring speed (350 rpm) or no agitation is applied to the stripping solution and different stirring speeds applied to the feed solution. From this, it is clear that every stirring speed causes significant changes in the shape and size distribution of the silver particles. For a 350 rpm stirring speed in the stripping solution (to 600 rpm in the feed

been reported earlier in the literature (Jixiang et al., 2007; Masaharu et al. 2010).

(a) (b)

(c) (d)

Fig. 7. Micrograph of silver particles on the membrane obtained with different stirring speeds in the two compartments. (a) Stirring speed 350 rpm in the stripping solution and 600 rpm in the feed solution. (b), (c) and (d) No agitation of the stripping solution. To the

feed solution a stirring speed of 1000, 600 and 300 rpm, is applied respectively

solution), the particles maintain a certain homogeneity. A decahedron shape is observed in figure 7a. On the other hand, when the stripping solution was not stirred and the stirring speed on the feed solution was reduced (1000, 600, 300 rpm) the shape of the silver crystals gets very different. In the case of a stirring speed of 1000 rpm, the metallic silver takes crystal morphology in the shapes of cubes and rods. If the stirring speed is decreased to 600 rpm, the silver particles appear as rounded shapes but with traces of nucleation that form a cubic shape (figure 7b, 7c). At 300 rpm on the feed solution, the silver particles take decahedron shapes with an average particle size greater than in other cases. Also in this case, the large silver particles are occluded in the pores of the microfiltration membrane. It is important to observe that the proposed methodology is very suitable for obtaining metallic silver particles of different shapes and sizes. In fact, in the literature, in order to obtain different shape and size silver nanoparticles, more controlled and drastic conditions are required than those proposed here.

Finally, in the absence of stirring at both phases, the silver particles obtained on the membrane surface, clearly show the formation of hexagonal plates (Figure 8). These hexagonal plates come from the formation of dendrites on the surface, which is the first stage in the process of crystallization of silver on the microfiltration membrane.

Fig. 8. Micrograph of the silver particles on the membrane when no stirring is applied to the solutions. (b) Higher magnification of the same sample

We can conclude that hydrodynamics play an important role in the morphology and size of silver particles.

In the next section we will discuss the chemical aspects that affect the process of reducing Ag+ ions by ascorbic acid. Additionally, we will analyze the conditions for efficient recovery of silver so we will have a better understanding of the mechanism under which the process is under mass transfer of Ag+ ions.
