**3. Results and discussion**

### **3.1 Factors affecting metal distribution: initial reactants concentration**

The simulation model was successfully validated by comparison with experimental results. Au/Pt nanoparticles were synthesized in a 75% Isooctane/20% Tergitol/5% water microemulsion [42] (which can be characterized as flexible microemulsion)using different precursor concentrations. (〈cAuCl4 −〉 = 〈cPtCl6 <sup>2</sup>−〉 = 〈c〉 = 0.01, 0.08, 0.16, and 0.40 M. The resulting Au/Pt particles were studied by HR-STEM and their structures were revealed by cross sections scanned with EDX analysis. The studied conditions were reproduced by simulation, using concentrations 〈cAuCl4 −〉 = 〈cPtCl6 <sup>2</sup>−〉 = 〈c〉 = 2, 16, 32, and 64 metal precursors in each micelle, which corresponds to 0.01, 0.08, 0.16, and 0.40 M, respectively, in a micelle with a radius of 4 nm. As mentioned above, Au/Pt pair was characterized as vAu/vPt = 10 reduction rate ratio, and the 75% Isooctane/20% Tergitol/5% water microemulsion was simulated as a flexible surfactant film (*f* = 30, *kex* = 5).

**65**

**Figure 1.**

*Microemulsions as Nanoreactors to Obtain Bimetallic Nanoparticles*

The left column in **Figure 1** shows the simulated nanostructures obtained at each concentration. In order to compare the experimental and simulated nanostructures, the quantity of each metal crossed by a beam of 2 Å (approximate EDX beam size) was computed from each simulated final nanoparticle, and the theoretical STEM profiles were calculated. The STEM profiles of the average particle for each concentration are shown in center (theoretical) and right (experimental) columns of **Figure 1**. For a better comparison, experimental x-axis was changed from nm to counts, and the two kind of profiles were normalized to 1. Both profiles show the expected behavior: the surface (outer layers) is enriched in Pt, because of its slower reduction rate, and Au, which is reduced faster, accumulates in the core (inner layers). As concentration increases (see **Figure 1** from the top to the bottom), deeper Pt profiles are obtained. This means that the final nanostructure shows an improved metal segregation as concentration is higher. It is clearly observed in the histograms, which evolve from Au core covered by a mixed shell obtained at a low concentration to a more mixed Au core covered by a pure Pt shell as concentration

*Left column: simulated histograms for different initial concentrations (Au:Pt = 1:1). The proportion of pure metal in the layer is higher as the color becomes lighter (red: 100% Au, blue: 100% Pt, gray: 50% Au-Pt). Centre column: calculated STEM profiles for the average nanoparticle. Right column: measured STEM profiles for Au/Pt nanoparticles synthesized in a water/tergitol/isooctane microemulsion. Simulation parameters: flexible film (kex = 5, f = 30); reduction rate ratio (vAu/vPt = 100/10); and reducing agent concentration* 〈*cR*〉 *= 10*〈*M+*

*Adapted with permission from Ref. [42]. Copyright (2015) American Chemical Society.*

〉*.* 

*DOI: http://dx.doi.org/10.5772/intechopen.80549*

#### *Microemulsions as Nanoreactors to Obtain Bimetallic Nanoparticles DOI: http://dx.doi.org/10.5772/intechopen.80549*

*Microemulsion - A Chemical Nanoreactor*

using the values *f* = 30, *kex* = 5 [42].

results are averaged over 1000 runs.

**3. Results and discussion**

concentrations 〈cAuCl4

A rigid film, such as AOT/n-heptane/water microemulsion, was successfully reproduced considering a channel size *f* = 5, associated to *kex* = 1 free atoms exchanged during a collision [26]. In case of flexible film, both factors rise together, because a more flexible film produces more stable dimer and larger channel size, allowing a quicker exchange of isolated species as well as an exchange of larger particles [41]. That is, a flexible film is associated to a faster material intermicellar exchange rate. A more flexible microemulsion, such as 75% Isooctane/20% Tergitol/5% water microemulsion, was successful compared to simulation data

**2.5 Description of the metal distribution in the bimetallic nanoparticle**

The composition of each nanoparticle is revised at each step and monitored as a function of time. When all metal precursors were reduced and the content of all micelles remains constant over time, nanoparticle synthesis is considered to be finished. At this stage, the sequence of metal deposition of each particle (which is stored as a function of time) is stabilized. One simulation run produces a set of micelles, each one of them can carry one particle with different composition or be empty. At the end of each run, the averaged nanoparticle is calculated. Finally,

The intrastructure of each particle is calculated by analyzing the sequence in which the two metals are deposited on the nanoparticle surface. So that, each sequence is arranged in 10 concentric layers, assuming that final nanoparticle is spherical. Then, the averaged percentage of each metal is calculated layer by layer. The final bimetallic distribution is represented by histograms, in which the layer composition is described by a color grading, as stated in the following pattern: Au, Pt, and Rh are represented by red, blue, and green, respectively. As the proportion of pure metal in the layer is higher, the color becomes lighter. In order to illustrate the heterogeneity of nanoparticle composition, the number of particles with a given percentage of the faster reduction metal (Au in Au/Pt and Pt in Pt/Rh nanoparticles) in each of 10 layers is also represented in the histograms. This analysis is reproduced layer by layer, from the beginning of the synthesis (inner layer or core) to the end (outer layer or surface). To simplify, the metal distribution is also shown by means of concentric spheres, whose thickness is proportional to the number of

layers with the same composition, keeping the same color pattern.

microemulsion)using different precursor concentrations. (〈cAuCl4

−〉 = 〈cPtCl6

**3.1 Factors affecting metal distribution: initial reactants concentration**

The simulation model was successfully validated by comparison with experimental results. Au/Pt nanoparticles were synthesized in a 75% Isooctane/20% Tergitol/5% water microemulsion [42] (which can be characterized as flexible

〈c〉 = 0.01, 0.08, 0.16, and 0.40 M. The resulting Au/Pt particles were studied by HR-STEM and their structures were revealed by cross sections scanned with EDX analysis. The studied conditions were reproduced by simulation, using

each micelle, which corresponds to 0.01, 0.08, 0.16, and 0.40 M, respectively, in a micelle with a radius of 4 nm. As mentioned above, Au/Pt pair was characterized as vAu/vPt = 10 reduction rate ratio, and the 75% Isooctane/20% Tergitol/5% water

microemulsion was simulated as a flexible surfactant film (*f* = 30, *kex* = 5).

−〉 = 〈cPtCl6

<sup>2</sup>−〉 = 〈c〉 = 2, 16, 32, and 64 metal precursors in

<sup>2</sup>−〉 =

**64**

The left column in **Figure 1** shows the simulated nanostructures obtained at each concentration. In order to compare the experimental and simulated nanostructures, the quantity of each metal crossed by a beam of 2 Å (approximate EDX beam size) was computed from each simulated final nanoparticle, and the theoretical STEM profiles were calculated. The STEM profiles of the average particle for each concentration are shown in center (theoretical) and right (experimental) columns of **Figure 1**. For a better comparison, experimental x-axis was changed from nm to counts, and the two kind of profiles were normalized to 1. Both profiles show the expected behavior: the surface (outer layers) is enriched in Pt, because of its slower reduction rate, and Au, which is reduced faster, accumulates in the core (inner layers). As concentration increases (see **Figure 1** from the top to the bottom), deeper Pt profiles are obtained. This means that the final nanostructure shows an improved metal segregation as concentration is higher. It is clearly observed in the histograms, which evolve from Au core covered by a mixed shell obtained at a low concentration to a more mixed Au core covered by a pure Pt shell as concentration

#### **Figure 1.**

*Left column: simulated histograms for different initial concentrations (Au:Pt = 1:1). The proportion of pure metal in the layer is higher as the color becomes lighter (red: 100% Au, blue: 100% Pt, gray: 50% Au-Pt). Centre column: calculated STEM profiles for the average nanoparticle. Right column: measured STEM profiles for Au/Pt nanoparticles synthesized in a water/tergitol/isooctane microemulsion. Simulation parameters: flexible film (kex = 5, f = 30); reduction rate ratio (vAu/vPt = 100/10); and reducing agent concentration* 〈*cR*〉 *= 10*〈*M+* 〉*. Adapted with permission from Ref. [42]. Copyright (2015) American Chemical Society.*

increases (see decreasing red bar on the left and increasing blue bar on the right in histograms of **Figure 1**). This means that the nanostructure can be fine-tuned with sub-nanometer resolution, just by changing concentration.

A good agreement between experimental and theoretical results was attained, upholding the validity of the simulation model to predict the atomic structure of bimetallic nanoparticles. On this basis, the model becomes a strong tool to further enhance our knowledge of the complex mechanisms governing reactions in microemulsions and the impact of compartmentalization on final nanostructures.

The better metal segregation obtained as concentration increases is also observed when the two reductions are slowing down, as shown in **Figure 2**. This figure shows the final nanostructures obtained for the pair Pt/Rh (vPt/ vRh = 10/1 = 10), under the same synthetic conditions as used in **Figure 1** to prepare Au/Pt nanoparticles. The better metal separation cannot be attributed to a larger reduction rate ratio, because in both cases the faster metal is 10 times faster than the slower one (vAu/vPt = 100/10 = 10). The better metal separation obtained for Pt/Rh pair is more evident at low concentration, where an alloy is obtained for Au/Pt and a core-shell structure for Pt/Rh (compare histograms when <c>= 2 in **Figures 1** and **2**). This means that, although the reduction rate ratio is similar, when the reduction rate of the faster metal slows down as in Pt, the other metal (Rh) is delayed even more. As a consequence, both reactions take place at different stages of the synthesis, resulting in better segregated structures.

#### **3.2 Factors affecting metal distribution: reduction rate ratio**

The difference between the standard potentials of the two metal precursors is believed to be the most relevant factor to determine the kinetics and the resulting bimetallic arrangement [43]. As established in Eq. (2), the higher the difference between the standard potentials of the two metals, the higher the ratio between both reduction rates is. It results in the earlier reduction of the faster reduction metal, which builds up the core and becomes the seed for the subsequent deposition of the slower metal, which forms the surrounding shell. On the contrary, when the two reduction rates are almost similar, a mixed nanoalloy is expected. In spite of this argumentation was initially proposed for reactions in homogeneous media, and it does not take into account the confinement of reactants within micelles, it is frequently applied to explain results in microemulsion. As a rule, it is observed a tendency from nanoalloy to core-shell structure as difference in reduction potential is increased (see Table 1 in Ref. [44]). Previous simulation studies allow to clearly observe a better separation of the two metals as reduction rate ratio is larger (for a deeper discussion, see Ref. [28]).

#### **Figure 2.**

*Histograms show the number of particles with a given percentage of the faster reduction metal (Pt) in each layer at different concentrations. In all cases,* 〈*cR*〉 *= 10*〈*cPtCl6 <sup>2</sup><sup>−</sup>*〉*, and cPtCl6 <sup>2</sup><sup>−</sup>:RhCl6 <sup>3</sup><sup>−</sup> is in 1:1 proportion. Reduction rates: vPt = 10%, vRh = 1% (vPt/vRh = 10); flexible film (f = 30, kex = 5). Scheme color: Pt and Rh are represented by blue and green bars, respectively. Lighter colors mean less mixture. The nanostructure is also shown by colored concentric spheres, keeping the same color pattern.*

**67**

**Figure 3.**

*10*〈*cPtCl6*

ing sections).

**3.4 Kinetic study**

number of metal precursors M+

*Microemulsions as Nanoreactors to Obtain Bimetallic Nanoparticles*

**3.3 Factors affecting metal distribution: microemulsion composition**

To isolate the effect of microemulsion composition on nanoparticle structure, a particular pair of metals must be chosen and analyze if a change in the microemulsion composition leads to a different metal segregation. For example, Pt/Ru nanoparticles were obtained as nanoalloy, both for rigid (water/Brij-30/n-heptane [45]) and flexible films (water/Berol 050/isooctane 80 [46] or water/NP5-NP9/ cyclohexane [47]). But in this couple, the small difference in reduction potentials leads to quite similar reduction rates, which hinder metal segregation, even with a slow intermicellar exchange rate. As a matter of fact, when couples with higher reduction rate ratio are studied (such as Au/Ag, Au/Pt, and Au/Pd), an increase in surfactant flexibility results in the expected transition from a core-shell to a nanoalloy. As an example, alloyed Au/Pt nanoparticles were prepared using a flexible film such as water/Tergitol 15-S-5/isooctane [17] or water/TritonX-100/cyclohexane [48]. On the contrary, rigid films (water/AOT/isooctane [39] and water/ Brij 30/n-heptane [49]) lead to segregated structures. The simulation model also predicts this result, as shown in **Figure 3**, in which different Au/Pt (vAu/vPt = 10) arrangements were obtained by employing different values of surfactant film flexibilities. The ability of the microemulsion to minimize the difference between the reduction rates is clearly reflected in the progressive mixture of Au and Pt as increasing the intermicellar exchange rate (for a deeper discussion, see the follow-

The results shown in previous figures were obtained under isolation conditions,

that is, the reducing agent concentration is much higher than stoichiometry, so the change in R concentration during the course of the reaction is negligible. As a result, the metal reduction, which is a bimolecular reaction, appears to be first order, when the reaction media is homogeneous. In order to study how the confinement of reactants inside micelles would affect chemical kinetics, the depletion of the

= AuCl4

shown in **Figure 4** using different initial reactant concentrations. Left and right columns show results for Au/Pt and Pt/Rh couples, respectively. Au, Pt, and Rh are represented by dashed, solid, and dashed-dotted lines, respectively. **Figure 4A** and **B** was obtained by simulating a flexible film and **C** and **D** a rigid one. At first

*Number of particles with a given percentage of the faster reduction metal (Au) in each layer using three* 

*<sup>2</sup><sup>−</sup>*〉 *reduction rates: vAu = 100%, vPt = 10%. Scheme color: Au and Pt are represented by red and blue* 

*different microemulsion compositions (different f and kex parameters).* 〈*cAuCl4*

*bars, respectively. Lighter colors mean less mixture. Adapted with permission from Ref. [44].*

<sup>−</sup>, PtCl6<sup>−</sup>

<sup>2</sup><sup>−</sup>, RhCl6

<sup>3</sup><sup>−</sup>) was monitored

concentration versus time is

*<sup>−</sup>*〉 *=* 〈*cPtCl6*

*<sup>2</sup><sup>−</sup>*〉 *= 4;* 〈*cR*〉 *=* 

(M+

as the synthesis advances. The logarithmic plot of M+

*DOI: http://dx.doi.org/10.5772/intechopen.80549*

### **3.3 Factors affecting metal distribution: microemulsion composition**

To isolate the effect of microemulsion composition on nanoparticle structure, a particular pair of metals must be chosen and analyze if a change in the microemulsion composition leads to a different metal segregation. For example, Pt/Ru nanoparticles were obtained as nanoalloy, both for rigid (water/Brij-30/n-heptane [45]) and flexible films (water/Berol 050/isooctane 80 [46] or water/NP5-NP9/ cyclohexane [47]). But in this couple, the small difference in reduction potentials leads to quite similar reduction rates, which hinder metal segregation, even with a slow intermicellar exchange rate. As a matter of fact, when couples with higher reduction rate ratio are studied (such as Au/Ag, Au/Pt, and Au/Pd), an increase in surfactant flexibility results in the expected transition from a core-shell to a nanoalloy. As an example, alloyed Au/Pt nanoparticles were prepared using a flexible film such as water/Tergitol 15-S-5/isooctane [17] or water/TritonX-100/cyclohexane [48]. On the contrary, rigid films (water/AOT/isooctane [39] and water/ Brij 30/n-heptane [49]) lead to segregated structures. The simulation model also predicts this result, as shown in **Figure 3**, in which different Au/Pt (vAu/vPt = 10) arrangements were obtained by employing different values of surfactant film flexibilities. The ability of the microemulsion to minimize the difference between the reduction rates is clearly reflected in the progressive mixture of Au and Pt as increasing the intermicellar exchange rate (for a deeper discussion, see the following sections).

#### **3.4 Kinetic study**

*Microemulsion - A Chemical Nanoreactor*

increases (see decreasing red bar on the left and increasing blue bar on the right in histograms of **Figure 1**). This means that the nanostructure can be fine-tuned with

A good agreement between experimental and theoretical results was attained, upholding the validity of the simulation model to predict the atomic structure of bimetallic nanoparticles. On this basis, the model becomes a strong tool to further enhance our knowledge of the complex mechanisms governing reactions in microemulsions and the impact of compartmentalization on final nanostructures. The better metal segregation obtained as concentration increases is also observed when the two reductions are slowing down, as shown in **Figure 2**. This figure shows the final nanostructures obtained for the pair Pt/Rh (vPt/ vRh = 10/1 = 10), under the same synthetic conditions as used in **Figure 1** to prepare Au/Pt nanoparticles. The better metal separation cannot be attributed to a larger reduction rate ratio, because in both cases the faster metal is 10 times faster than the slower one (vAu/vPt = 100/10 = 10). The better metal separation obtained for Pt/Rh pair is more evident at low concentration, where an alloy is obtained for Au/Pt and a core-shell structure for Pt/Rh (compare histograms when <c>= 2 in **Figures 1** and **2**). This means that, although the reduction rate ratio is similar, when the reduction rate of the faster metal slows down as in Pt, the other metal (Rh) is delayed even more. As a consequence, both reactions take place at differ-

sub-nanometer resolution, just by changing concentration.

ent stages of the synthesis, resulting in better segregated structures.

The difference between the standard potentials of the two metal precursors is believed to be the most relevant factor to determine the kinetics and the resulting bimetallic arrangement [43]. As established in Eq. (2), the higher the difference between the standard potentials of the two metals, the higher the ratio between both reduction rates is. It results in the earlier reduction of the faster reduction metal, which builds up the core and becomes the seed for the subsequent deposition of the slower metal, which forms the surrounding shell. On the contrary, when the two reduction rates are almost similar, a mixed nanoalloy is expected. In spite of this argumentation was initially proposed for reactions in homogeneous media, and it does not take into account the confinement of reactants within micelles, it is frequently applied to explain results in microemulsion. As a rule, it is observed a tendency from nanoalloy to core-shell structure as difference in reduction potential is increased (see Table 1 in Ref. [44]). Previous simulation studies allow to clearly observe a better separation of the two metals as reduction rate ratio is larger (for a

*Histograms show the number of particles with a given percentage of the faster reduction metal (Pt) in each* 

*Reduction rates: vPt = 10%, vRh = 1% (vPt/vRh = 10); flexible film (f = 30, kex = 5). Scheme color: Pt and Rh are represented by blue and green bars, respectively. Lighter colors mean less mixture. The nanostructure is also* 

*<sup>2</sup><sup>−</sup>*〉*, and cPtCl6*

*<sup>2</sup><sup>−</sup>:RhCl6*

*<sup>3</sup><sup>−</sup> is in 1:1 proportion.* 

**3.2 Factors affecting metal distribution: reduction rate ratio**

deeper discussion, see Ref. [28]).

*layer at different concentrations. In all cases,* 〈*cR*〉 *= 10*〈*cPtCl6*

*shown by colored concentric spheres, keeping the same color pattern.*

**66**

**Figure 2.**

The results shown in previous figures were obtained under isolation conditions, that is, the reducing agent concentration is much higher than stoichiometry, so the change in R concentration during the course of the reaction is negligible. As a result, the metal reduction, which is a bimolecular reaction, appears to be first order, when the reaction media is homogeneous. In order to study how the confinement of reactants inside micelles would affect chemical kinetics, the depletion of the number of metal precursors M+ (M+ = AuCl4 <sup>−</sup>, PtCl6<sup>−</sup> <sup>2</sup><sup>−</sup>, RhCl6 <sup>3</sup><sup>−</sup>) was monitored as the synthesis advances. The logarithmic plot of M+ concentration versus time is shown in **Figure 4** using different initial reactant concentrations. Left and right columns show results for Au/Pt and Pt/Rh couples, respectively. Au, Pt, and Rh are represented by dashed, solid, and dashed-dotted lines, respectively. **Figure 4A** and **B** was obtained by simulating a flexible film and **C** and **D** a rigid one. At first

#### **Figure 3.**

*Number of particles with a given percentage of the faster reduction metal (Au) in each layer using three different microemulsion compositions (different f and kex parameters).* 〈*cAuCl4 <sup>−</sup>*〉 *=* 〈*cPtCl6 <sup>2</sup><sup>−</sup>*〉 *= 4;* 〈*cR*〉 *= 10*〈*cPtCl6 <sup>2</sup><sup>−</sup>*〉 *reduction rates: vAu = 100%, vPt = 10%. Scheme color: Au and Pt are represented by red and blue bars, respectively. Lighter colors mean less mixture. Adapted with permission from Ref. [44].*

sight, metal reductions obey first-order kinetics in both Au/Pt and Pt/Rh synthesis, as expected. Nevertheless, it is important to note that, with the exception of Rh, a time lag is required to reach the linear regime. Two points must be highlighted: First, the higher the concentration, the longer the time lag between the beginning of the synthesis and the achievement of the linear behavior. On the second hand, the time lag strongly depends on the intermicellar exchange rate, being longer as the exchange rate is slower (rigid film). Both factors (concentration and film flexibility) suggest that the rate-determining step is the intermicellar exchange rate at earlier reaction times, as explained as follows. The synthesis starts when the microemulsions containing the reactants are mixed. In order to be able to react, reactants must be located inside the same micelle. The reactants redistribution between micelles is dictated by the rate with which reactants can go through the channels communicating colliding micelles, that is, the intermicellar exchange rate. So, a slow exchange rate only allows the exchange of few reactants in each collision, which implies that more collisions are required to redistribute reactants and allow the reactants encounter. Therefore, a rigid film requires much longer lag times than a flexible one (compare **Figure 4A** with **B** and **C** with **D**, for any value of concentration). Apart from that, reactants redistribution is also affected by concentration, because the number of reactants which can traverse the intermicellar channel during an effective collision is restricted. Therefore, if concentration within micelle is large, more collisions (i.e., more time) are needed to make possible reactants redistribution. Finally, it is interesting to point out that this delay in reaching linear behavior disappears when a very slow chemical reduction takes place, as shown by Rh kinetics in right column of **Figure 4** (see dashed-dotted lines), which is linear from the beginning, at any

#### **Figure 4.**

*Plot of ln M<sup>+</sup> (number of metal salt) against time (in Monte Carlo step, mcs) using different initial concentration c (metal salts/micelle). Solid, dashed, and dashed-dotted represent Au, Pt, and Rh, respectively. A and B shows results for a flexible surfactant film (kex = 5, f = 30) and C and D for a rigid one (kex = 1, f = 5). Au/Pt pair (vAu/vPt = 100/10 = 10) is represented in A and C and Pt/Rh pair (vPt/vRh = 10/1 = 10) in B and D. Stars indicate the half-life (red, blue, and green means Au, Pt and Rh, respectively).*

**69**

**Figure 5.**

*Microemulsions as Nanoreactors to Obtain Bimetallic Nanoparticles*

concentration value. This behavior can be taken as indication that the reduction rate is so slow (only a 1% of reactants give rise to products) that the limiting step is the

Summarizing, the time lag required to achieve linear behavior in **Figure 4** reflects the time it takes for reactants to encounter. This time lag can be determinant of final metal segregation, because the inner layers of nanoparticle are building up

Pseudo-first-order rate constants, kobs, can be calculated from linear regime of the logarithmic plot as shown in **Figure 4**. The values of the pseudo-first-order rate constants are represented in **Figure 5** as a function of concentration. One can observe that the slopes of Au reduction are always higher than the slopes of Pt, which in turn is faster that Rh, as expected. Classical chemical kinetics in a homogeneous reaction medium establishes that kobs in bimolecular reactions does not depend on precursors concentrations. This is the case for Pt and Rh, whose kobs values did not depend neither on the concentration nor on the intermicellar exchange rate. In contrast, kobs values of Au are strongly influenced by both factors. To explain this behavior, one have to take into account that the limiting step in Au chemical reduction is the intermicellar exchange [38, 50], because of the extremely fast Au reduction rate. One can observe that the dependence of kobs on concentration decreases as intermicellar exchange rate is faster, until reaching an almost constant value at very fast intermicellar exchange rate (see gray line in **Figure 5**), as expected. In comparison, Pt and Rh reductions are so slow that their rates are not

One can conclude from **Figure 5** that intermicellar exchange rate exerts a different degree of influence depending on the reduction rate of the metal in comparison to the intermicellar exchange rate. This means that the compartmentalization of reaction medium takes part in chemical kinetics more or less depending on the metal nature. This different interplay between exchange rate and reduction rate has to be reflected in the metal segregation of final nanoparticle. It was proposed that if the intermicellar exchange rate can only modify the rate of metals whose

*kobs (slopes of the linear parts from the plots in* **Figure** *3) as a function of concentration for different microemulsion compositions and different metals. vAu = 100, vPt = 10, vRh = 1. Lines are only a guide to the eye.*

*DOI: http://dx.doi.org/10.5772/intechopen.80549*

reduction itself.

during this stage.

limited by the exchange rate.

*Microemulsion - A Chemical Nanoreactor*

sight, metal reductions obey first-order kinetics in both Au/Pt and Pt/Rh synthesis, as expected. Nevertheless, it is important to note that, with the exception of Rh, a time lag is required to reach the linear regime. Two points must be highlighted: First, the higher the concentration, the longer the time lag between the beginning of the synthesis and the achievement of the linear behavior. On the second hand, the time lag strongly depends on the intermicellar exchange rate, being longer as the exchange rate is slower (rigid film). Both factors (concentration and film flexibility) suggest that the rate-determining step is the intermicellar exchange rate at earlier reaction times, as explained as follows. The synthesis starts when the microemulsions containing the reactants are mixed. In order to be able to react, reactants must be located inside the same micelle. The reactants redistribution between micelles is dictated by the rate with which reactants can go through the channels communicating colliding micelles, that is, the intermicellar exchange rate. So, a slow exchange rate only allows the exchange of few reactants in each collision, which implies that more collisions are required to redistribute reactants and allow the reactants encounter. Therefore, a rigid film requires much longer lag times than a flexible one (compare **Figure 4A** with **B** and **C** with **D**, for any value of concentration). Apart from that, reactants redistribution is also affected by concentration, because the number of reactants which can traverse the intermicellar channel during an effective collision is restricted. Therefore, if concentration within micelle is large, more collisions (i.e., more time) are needed to make possible reactants redistribution. Finally, it is interesting to point out that this delay in reaching linear behavior disappears when a very slow chemical reduction takes place, as shown by Rh kinetics in right column of **Figure 4** (see dashed-dotted lines), which is linear from the beginning, at any

 *(number of metal salt) against time (in Monte Carlo step, mcs) using different initial concentration c (metal salts/micelle). Solid, dashed, and dashed-dotted represent Au, Pt, and Rh, respectively. A and B shows results for a flexible surfactant film (kex = 5, f = 30) and C and D for a rigid one (kex = 1, f = 5). Au/Pt pair (vAu/vPt = 100/10 = 10) is represented in A and C and Pt/Rh pair (vPt/vRh = 10/1 = 10) in B and* 

*D. Stars indicate the half-life (red, blue, and green means Au, Pt and Rh, respectively).*

**68**

**Figure 4.** *Plot of ln M<sup>+</sup>* concentration value. This behavior can be taken as indication that the reduction rate is so slow (only a 1% of reactants give rise to products) that the limiting step is the reduction itself.

Summarizing, the time lag required to achieve linear behavior in **Figure 4** reflects the time it takes for reactants to encounter. This time lag can be determinant of final metal segregation, because the inner layers of nanoparticle are building up during this stage.

Pseudo-first-order rate constants, kobs, can be calculated from linear regime of the logarithmic plot as shown in **Figure 4**. The values of the pseudo-first-order rate constants are represented in **Figure 5** as a function of concentration. One can observe that the slopes of Au reduction are always higher than the slopes of Pt, which in turn is faster that Rh, as expected. Classical chemical kinetics in a homogeneous reaction medium establishes that kobs in bimolecular reactions does not depend on precursors concentrations. This is the case for Pt and Rh, whose kobs values did not depend neither on the concentration nor on the intermicellar exchange rate. In contrast, kobs values of Au are strongly influenced by both factors. To explain this behavior, one have to take into account that the limiting step in Au chemical reduction is the intermicellar exchange [38, 50], because of the extremely fast Au reduction rate. One can observe that the dependence of kobs on concentration decreases as intermicellar exchange rate is faster, until reaching an almost constant value at very fast intermicellar exchange rate (see gray line in **Figure 5**), as expected. In comparison, Pt and Rh reductions are so slow that their rates are not limited by the exchange rate.

One can conclude from **Figure 5** that intermicellar exchange rate exerts a different degree of influence depending on the reduction rate of the metal in comparison to the intermicellar exchange rate. This means that the compartmentalization of reaction medium takes part in chemical kinetics more or less depending on the metal nature. This different interplay between exchange rate and reduction rate has to be reflected in the metal segregation of final nanoparticle. It was proposed that if the intermicellar exchange rate can only modify the rate of metals whose

#### **Figure 5.**

*kobs (slopes of the linear parts from the plots in* **Figure** *3) as a function of concentration for different microemulsion compositions and different metals. vAu = 100, vPt = 10, vRh = 1. Lines are only a guide to the eye.*

reduction is very fast [38], such as Au, only bimetallic nanoparticle including Au could be prepared with different metal distributions as a function of microemulsion composition (intermicellar exchange rate) by a one-pot method (see Table 1 in Ref. [44]). To the best of our knowledge, only Au/Pt, Au/Ag, and Au/Pd have been synthesized in a different intrastructure by different authors. Thus, when a rigid film (such as provided by AOT) is used, Au-Pt nanoparticles are arranged in a coreshell distribution [39]. On the contrary, more flexible surfactants such as Brij30 [49], tergitol [17], or TritonX-100 [48] give rise to alloyed nanoparticles. In relation to Au/Ag, alloys were obtained with TritonX-100 [51] and C11E3 and C11E5 [52], but an enriched in Ag surface was observed when microemulsion contains AOT [53]. Finally, AOT was also used to obtain core-shell Au-Pd nanoparticles [54] and alloys with Pd-rich surface [55]. In contrast, true Au-Pd alloys [4] have been obtained with Brij30 and TritonX-100. With that in mind, it could be suggested that metal segregation in the nanoparticle can be modified by a change in the microemulsion composition only when one of the metals is Au or another very fast reduction rate metal. Nevertheless, in spite of the agreement between theoretical and experimental data, this assumption is based on the kinetic constants, which were calculated from the linear plot shown in **Figure 4**. It must be emphasized that the linear regime is not fulfilled at initial stages of the reaction, when the core is been building up. With the aim of studying the relevance of the non-linear behavior at the beginning of the synthesis, the half-life, defined as the time that it takes for the reactant concentration to decay to half of its initial value, was calculated for each case. Stars in **Figure 4** show half-life under different synthesis conditions (Au, Pt, and Rh are represented by red, blue, and green stars, respectively). With the exception of Rh and Pt at very low concentration, half-life is usually smaller than the time needed to achieve the linear regime. As observed in **Figure 4C** and **D**, Au and Pt reductions in a rigid microemulsion and at high concentration have a half-life much earlier than linear plot. This means that not only the initial layers but also the middle ones are formed under a nonlinear regime. So, chemical reductions are still not a first-order process during the formation of a large part of the particle (for a deeper discussion on life time, see Ref. [38]).

### **4. Conclusions**

The generalized belief according to which the difference in the reduction potentials determines the intrastructure in a bimetallic nanoparticle should be improved. We propose that there are three potentially limiting factors which restrict chemical kinetics of bimetallic nanoparticles prepared from microemulsions: chemical reduction rate itself, exchange rate of reactants between micelles, and reactants concentration. The specific combination of these three factors determines the reaction rate of each metal, which in turn determines the sequence of metals deposition and the resulting bimetallic arrangement.

The kinetic study of Pt/M nanoparticles prepared via microemulsions under isolation conditions shows that chemical reductions are pseudo-first-order reactions, but not from the initial stages. At the beginning of the synthesis, the reactants encounter is dictated by the redistribution of reactants between micelles, which is controlled by the intermicellar exchange rate. As a result, the limiting step of faster reduction metals, such as Au, is the intermicellar exchange. On the contrary, microemulsion dynamics has a little effect if reduction rates are very slow (i.e., Rh). This means that compartmentalization of the reaction media has a different impact depending on the reduction rate of the particular metal. We are not referring only to the reduction rate of a metal in relation to the another metal

**71**

*Microemulsions as Nanoreactors to Obtain Bimetallic Nanoparticles*

in the pair but also how fast the reduction takes place in relation to the intermicellar exchange rate. Specifically, for a given reduction rates ratio and keeping fixed microemulsion composition and concentration, the fact that the two reactions were slow (as in Pt/Rh) leads to a better metal segregation than if both reactions are faster (as in Au/Pt). Therefore, with the exception of very slow metal reduction as Rh, intermicellar exchange rate drastically impacts on chemical kinetics, particularly at the beginning of the synthesis. This is not a minor matter, because it will be reflected in the composition of the core and middle layers of the resulting nanoparticle. So, the ability of microemulsion to manipulate the sequence of metal deposition, when the metal reductions are quite fast in relation to the intermicellar exchange rate, can be used to design the experiments to synthesize bimetallic particles with ad hoc nanoarrangements. This ability disappears when the two chemical reductions are slow, because of chemically controlled kinetics.

In this paper, computer simulation has proved to be very useful tool to identify

This work was supported by MINECO, Spain (MAT2015-67458-P, co-financed with FERDER Funds), from the European Union's H2020 research and innovation programme under grant agreement No. 646155 (INSPIRED) and from Xunta de Galicia (Programa REDES ED431D-2017/18). D.B. thanks for the postdoc grant

This research is dedicated to Prof. Julio Casado Linarejos, who teached us the

suitable synthesis parameters, which control metal segregation in a bimetallic nanoparticle. Further insights into the interplay between metal nature, exchange

rate, and final bimetallic structure can be gained from kinetics studies.

from Xunta de Galicia, Spain (POS-A/2013/018).

The authors declare no conflict of interest.

**Notes/thanks/other declarations**

fundamentals of chemical kinetics.

*DOI: http://dx.doi.org/10.5772/intechopen.80549*

**Acknowledgements**

**Conflict of interest**

*Microemulsions as Nanoreactors to Obtain Bimetallic Nanoparticles DOI: http://dx.doi.org/10.5772/intechopen.80549*

in the pair but also how fast the reduction takes place in relation to the intermicellar exchange rate. Specifically, for a given reduction rates ratio and keeping fixed microemulsion composition and concentration, the fact that the two reactions were slow (as in Pt/Rh) leads to a better metal segregation than if both reactions are faster (as in Au/Pt). Therefore, with the exception of very slow metal reduction as Rh, intermicellar exchange rate drastically impacts on chemical kinetics, particularly at the beginning of the synthesis. This is not a minor matter, because it will be reflected in the composition of the core and middle layers of the resulting nanoparticle. So, the ability of microemulsion to manipulate the sequence of metal deposition, when the metal reductions are quite fast in relation to the intermicellar exchange rate, can be used to design the experiments to synthesize bimetallic particles with ad hoc nanoarrangements. This ability disappears when the two chemical reductions are slow, because of chemically controlled kinetics.

In this paper, computer simulation has proved to be very useful tool to identify suitable synthesis parameters, which control metal segregation in a bimetallic nanoparticle. Further insights into the interplay between metal nature, exchange rate, and final bimetallic structure can be gained from kinetics studies.

## **Acknowledgements**

*Microemulsion - A Chemical Nanoreactor*

on life time, see Ref. [38]).

and the resulting bimetallic arrangement.

**4. Conclusions**

reduction is very fast [38], such as Au, only bimetallic nanoparticle including Au could be prepared with different metal distributions as a function of microemulsion composition (intermicellar exchange rate) by a one-pot method (see Table 1 in Ref. [44]). To the best of our knowledge, only Au/Pt, Au/Ag, and Au/Pd have been synthesized in a different intrastructure by different authors. Thus, when a rigid film (such as provided by AOT) is used, Au-Pt nanoparticles are arranged in a coreshell distribution [39]. On the contrary, more flexible surfactants such as Brij30 [49], tergitol [17], or TritonX-100 [48] give rise to alloyed nanoparticles. In relation to Au/Ag, alloys were obtained with TritonX-100 [51] and C11E3 and C11E5 [52], but an enriched in Ag surface was observed when microemulsion contains AOT [53]. Finally, AOT was also used to obtain core-shell Au-Pd nanoparticles [54] and alloys with Pd-rich surface [55]. In contrast, true Au-Pd alloys [4] have been obtained with Brij30 and TritonX-100. With that in mind, it could be suggested that metal segregation in the nanoparticle can be modified by a change in the microemulsion composition only when one of the metals is Au or another very fast reduction rate metal. Nevertheless, in spite of the agreement between theoretical and experimental data, this assumption is based on the kinetic constants, which were calculated from the linear plot shown in **Figure 4**. It must be emphasized that the linear regime is not fulfilled at initial stages of the reaction, when the core is been building up. With the aim of studying the relevance of the non-linear behavior at the beginning of the synthesis, the half-life, defined as the time that it takes for the reactant concentration to decay to half of its initial value, was calculated for each case. Stars in **Figure 4** show half-life under different synthesis conditions (Au, Pt, and Rh are represented by red, blue, and green stars, respectively). With the exception of Rh and Pt at very low concentration, half-life is usually smaller than the time needed to achieve the linear regime. As observed in **Figure 4C** and **D**, Au and Pt reductions in a rigid microemulsion and at high concentration have a half-life much earlier than linear plot. This means that not only the initial layers but also the middle ones are formed under a nonlinear regime. So, chemical reductions are still not a first-order process during the formation of a large part of the particle (for a deeper discussion

The generalized belief according to which the difference in the reduction potentials determines the intrastructure in a bimetallic nanoparticle should be improved. We propose that there are three potentially limiting factors which restrict chemical kinetics of bimetallic nanoparticles prepared from microemulsions: chemical reduction rate itself, exchange rate of reactants between micelles, and reactants concentration. The specific combination of these three factors determines the reaction rate of each metal, which in turn determines the sequence of metals deposition

The kinetic study of Pt/M nanoparticles prepared via microemulsions under

isolation conditions shows that chemical reductions are pseudo-first-order reactions, but not from the initial stages. At the beginning of the synthesis, the reactants encounter is dictated by the redistribution of reactants between micelles, which is controlled by the intermicellar exchange rate. As a result, the limiting step of faster reduction metals, such as Au, is the intermicellar exchange. On the contrary, microemulsion dynamics has a little effect if reduction rates are very slow (i.e., Rh). This means that compartmentalization of the reaction media has a different impact depending on the reduction rate of the particular metal. We are not referring only to the reduction rate of a metal in relation to the another metal

**70**

This work was supported by MINECO, Spain (MAT2015-67458-P, co-financed with FERDER Funds), from the European Union's H2020 research and innovation programme under grant agreement No. 646155 (INSPIRED) and from Xunta de Galicia (Programa REDES ED431D-2017/18). D.B. thanks for the postdoc grant from Xunta de Galicia, Spain (POS-A/2013/018).
