Abstract
In this experiment, benzene was treated with hydrogen in the presence of Pt/SiO2 as an example that can be applied on industrial operations to show that the most recently modernized methods of neutron total scattering that cover a wide Q-range allow the use of appropriate catalytic chemistry that can be measured on time domain to probe directly on site the hydrogenation of benzene within the catalyst pore. Based on this method, it is possible to study the rates of reaction on length scales that can provide measurements equivalent to the size of an atom or a nanometre. Furthermore, the method simultaneously provides insight into the mechanism chemical reactions from the atomistic structural angle.
Introduction
Catalysis, particularly heterogeneous catalysis plays an important role in the production of various chemicals all over the world. Indeed, the manufacturing process of approximately 80% of manmade materials involves catalyzed chemical reactions. Heterogeneous catalysts are used in catalyzing most of the reactions involved in the industrial production of most of these chemicals. The importance of catalysis is evidenced by the fact that nearly 35% of the worldwide GDP depends on it. Most reactions that are driven by catalysts are carried out in the liquid phase. Even though these materials and processes are important, people have few insights into the multiple interphases and their impact on the rate of the overall process of reactions. In most cases, the sampling of liquid or gas uptake are used to monitor liquid phase reactions that rely on heterogeneous catalysts. This is the case because obtaining on site spatially and chemically-determined information on the microscopic scale is difficult. Techniques that use nuclear magnetic resonance spectroscopy are capable of obtaining information concerning diffusion and spatially-resolved activity within packed-bed system. However, the resolutions of these techniques do no go beyond micron units. Neutron scattering is a highly effective technique in probing condensed matter with an atomic resolution. However, total scattering techniques have not been recognized for use in measuring the kinetics of real systems because they exhibit long data acquisition times. Several time-resolved studies, such as those focusing on micelle aggregation and evolution, polymerization processes, and rheochaos and flow instability, have applied neutron scattering and x-ray techniques in which the elastic scattering of the rays, x-rays, is recorded at low angles. However, due to the nature of these techniques, these studies have not been able to provide information on the chemical and atomistic regime. In this experiment, we show that wide Q-range and total neutron scattering techniques used at the modern pulsed neutron facilities can obtain information concerning the kinetic reaction processes. In this case, the information can be obtained over a wide scale covering small and wide angle regimes simultaneously.
The treatment of benzene with hydrogen to form cyclohexane under platinum catalyst has been used in this experiment to that neutron methods can probe the reaction within the catalysts pore. Many studies have been conducted to gain insight into the mechanism of the reaction involved during the treatment of benzene with hydrogen. Most of these studies have focused on the reactions that take place in liquid and solid phases in which a variety of catalysts that include supported Ru, Ni, Pd, Pt, and Fe. The hydrogenation of benzene is important for commercial purposes because its product, cyclohexane, is widely used in the production of nylon and caprolactam. In both cases, cyclohexane serves as a precursor. Caprolactam is used in the production of nylon and in the removal of benzene from gasoline. The production of caprolactam promotes the use of cleaner fuels.
Experimental Details
Neutron Diffraction
In a total scattering experiment, the interference differential interferences, F(Q), is the structural quantity that is measured. In this case, Q represents the scattering angle (2θ) and the momentum transfer vector magnitude in the scattering process for neutrons of incident wavelength (λ). Q is defined as shown in the equation below:
Q=4πλsinθ Equation 1
A weighted sum of the factors of partial structure, Sij(Q), forms F(Q). Sij(Q) reflects the pair correlations between atoms of type i and j as shown below:
FQ=i,j2-δijcicjbibjSijQ-1 Equation 2
Based on equation 2 shown above, δij represents the Kronecker delta intended for avoiding possible double counting of the air terms. Ci,cj, and bi,bj,on the other hand, represent the atomic fractions and coherent scattering lengths of the types of atoms. Fourier transform affecting the corresponding radial distribution function, G(r), which is weighted by the system’s atomic density, r, relates the interference differential cross-section as shown in the equation below:
Gr=(12π3ρ0∞4πQ2FQsinQrQrdQ Equation 3
While F(Q) in equation 2 is formed from weighted sum of factors of partial structure, Sij(Q),
G(r) is formed from a weighted sum of the partial pair distribution functions, gij(r). However, both functions are formed in a similar manner. It is recommended that the differential correlation function, D(r), should be used when in experiments that involve the investigation of nanostructured materials. Since the present experiment involves the investigation of a nanostructured material, this functions is recommended. The differential correlation, D(r), is derived from the radial distribution, G(r), as shown in the equation below:
Dr=4πρrG(r) Equation 4
The alternate tendency of F(Q) and G(r) clearly define the structural features at longer distance.
In the present experiment, all the measurements of neutron diffraction were made on the Near and InterMediate Range Order Diffractometer (NIMROD). The exercise was conducted at the ISIS Second Target Station Facility of the Rutherford Appleton Laboratory in Oxford, United Kingdom. NIMROD provides a wide Q-range that can access correlations lying between <0.1 nm and ˷30 nm. Furthermore, NIMROD can provide high flux. The most important aspect of the instrument is that it exhibits high neutron count rate at intermediate and molecular order length scales. This property makes it possible for short acquisition times to be employed. As a result of applying short acquisition times, extraction of kinetic information becomes relatively easier. In the present experiment, Pt/SiO2 was used in the instrument at a temperature level of 296 K and pressure of 0.25 bar D2 (g) to conduct benzene hydrogenation. In other words, the measurements, in this case, probed the protium replacement by deuterium in benzene benzene-d6, Scheme1. Benzene benzene-d6, Scheme1 is a heavier isotope. Consequently, its neutron scattering characteristics are more favorable. This made it possible for its probing by the measurements. The catalyst, Pt/SiO2, was exposed to D2 in benzene-d6 at room temperature to prereduce it before performing neutron diffraction. The mixture of the catalyst and D2 in benzene-d6 was then pumped out of the system up to the point where the catalyst diffraction pattern did remain constant. The catalyst pore was then refilled with benzene-d6 vapor and then followed with D2.
Synthesis
All the chemicals used in this experiment were supplied by Qmx or Aldrich. Besides, the chemicals were used without any further purification. In case any chemical was not bought from Aldrich or Qmx, a statement to that effect is given. A procedure previously reported by Grün and colleagues was modified and then used to synthesize the mesoporous SiO2. In this case, 5g of 5 g of n-hexadecyltrimethylammonium bromide (C16TMABr, 0.014 mol) was dissolved in 100 g of ultrapure water (distilled, deionized >18 MU) by heating at 318 K to form a surfactant solution. Next, 120.0 g of absolute ethanol (EtOH, 2.6 mol) and 26.4 g of aqueous ammonia (32 wt%, 0.5 mol) were added to the surfactant solution. After being stirred for fifteen minutes, 9.4 g of triethyl orthosilicate (TEOS) (0.044 mol) was added to the solution. As a result, a gel with the molar composition (TEOS)(C16TMABr)0.3(NH3)11(H2O)144(EtOH)58 was formed. The gel was then stirred for two hours before being transferred to an autoclave lined with Teflon where it was aged at a temperature of 378K for forty-eight hours. A white precipitate was obtained.
The next step involved filtering and washing the precipitate with 200cm3 of ultra-pure water and 200cm3 of ethanol. The precipitate was then dried overnight at a temperature of 363K forming a white powder. The powder was then calcined at 823 K for 5 h with a heating rate of 2K per minute in air. Next, using platinum nitrate (Johnson-Matthey) as the platinum precursor, 5 wt% Pt/SiO2 was prepared through incipient wetness impregnation technique. Once the impregnation process was complete, the material was dried at the temperature of 393K for twelve hours before being calcinated at 773K for 4 hours.
Batch Liquid Phase Hydrogenation of Benzene
The experiments involving the hydrogenation of benzene were conducted in an Autoclave Engineers' high-pressure reactor with a capacity of 100cm3. Under normal circumstances, 0.25 g of 5 wt% Pt/MCM-41 catalyst and 20 cm3 of benzene were used to charge the reactor. The purging of the reactor was also conducted three times using nitrogen gas (N2). The mixture of Pt/MCM-41 catalyst and benzene was then heated and agitated at 323 K and 1500 rpm respectively. Next, the reactor was purged with hydrogen gas (Hs) three times before being pressurized to 5 bar (the 5 bar corresponded to t=0). In order to monitor the reaction, sampling was carried out at regular intervals. In this case, the samples were analyzed using a Perkin Elmer GC that is equipped with a DB-1 capillary column and a FID detector. These conditions enabled the conversion of benzene by 98% after a period of twenty-four hours of the hydrogenation reaction.
Catalyst Characterisation
The prepared catalyst is expected to have a regular array of uniform parallel channels with a hexagonal pattern as shown in the images of transmission electron microscope (TEM) with a high resolution shown in figure 1. The TEM image also reveals dark spots that are less than 1nm in size and are dispersed within the channels. These spots are the Pt nanoparticles. It is important to conduct a textural characterization of the catalyst to be able to interpret the data of the whole system. The conventional methods for measuring the porosity of the materials produced in the experiment involve the use of adsorption isotherms such as BET. Nevertheless, these techniques do not provide detailed characterization. Since neutron methods provide more detailed characterization with one probe than the methods that use adsorption isotherms, it is advisable to use them instead of using the methods that apply adsorption isotherms.
The F(Q) obtained from the evacuated Pt/SiO2 is shown in Figure. 2. A direct description of the porous framework of the material can be derived from the evacuated Pt/SiO2. If a Lorentzian profile is fitted to the first diffraction peak, a center of 0.190Å-1 with FWHM of 0.022 Å-1, which corresponds to a pore spacing of 33.1Å (s.d. 1.48 Å), is obtained. Figure 3 shows a Direct Fourier transform of the measured empty Pt/SiO2 sample. The Direct Fourier transform shows the differential correlation functions, D(r). The Pt/SiO2 data clearly reveals the spacing of the pores in the material. In this case, the spacing, d, is 33 Å. The data also shows that the oscillation of this period extends to beyond 100 Å. If measurements of x-ray diffraction of the same material are obtained, d spacing of 34.6 Å is revealed. On comparing this to the d spacing obtained in neutron diffraction measurements (33 Å), one can deduced that the two measurements are in good agreement.
Results and Discussion
Reactant/Product Characterisation
Benzene-d6 was absorbed into the pores of the catalyst through the capillary condensation of the vapor due to the measurement of the evacuated catalyst. The pores were saturated after five minutes. Besides, the structure factor did not show any further changes as shown in figure 2. These observations show that the first diffraction peak increased in terms of the intensity. Besides, the data show that the peak of signal complexity increased beyond 1 Å-1. A decrease in the intensity of the first diffraction peak is also observed in cases where the pores become filled with a material that possesses a scattering length density (SLD) (benzene-d6,) that offers less contrast to the silica than the pore that is evacuated. The change in scattering length density (∆SLD) for the deuteriated benzene is equal to 1.93 * 10-6 Å-2. On the other hand, the change in scattering length density for empty pore is equivalent to 3.47 * 10-6 Å-2. Therefore, the deuteriation of benzene reduces the effect of the interaction of the features of the pores on the F(Q). The increase of signal complexity above 1Å-1 commonly occurs in cases where additional correlations of atomic scale that correspond to the absorbed benzene-d6 are involved. The “washing out” of the periodic correlation of the pore matches the reduction in the intensity of the first diffraction peak. The periodic pore correlation is said to be “washed out” because the oscillations at long r appear to be almost totally damped due to contrast matching as shown in figure 3.
Figure 2 shows a rise in the first diffraction peak obtained after exposure to D2(g) for 9 h. The increase in the first diffraction peak is consistent with cyclohexane-d12 formation from benzene hydrogenation. The SLD of cyclohexane-d12gives a contrast with bulk SiO2 that is comparable to the contrast of the empty pore (3.24 * 10-6 cf. 3.47 * 10-6 Å-2). Therefore, it promotes the reappearance of the first diffraction peak. This explains why the long r oscillations are restored in the D (r) to a moderate extent as seen in figure 3. Additionally, a strong feature identified at Q = 1.2 Å-1 implies that a different species from benzene that is not found at t = 0 is present.
The direct Fourier transform of the initial system loaded with benzene is shown in figure 4. The figure also shows the mean of the last six datasets collected that correspond to the final 0.5 h of the study where further reaction occurs. It also shows the last six datasets obtained from the pure liquids. Both the systems with pure benzene and benzene-loaded systems display C-C correlations at 1.397 Å. However, the systems do not show these C-C correlations after reaction. In the same way, there is consistency between the new feature at ̰1.55Å and the cyclohexane C-C distance. The cyclohexane C-C distance overlaps with the Si-O peak at 1.59 Å. The Si-O peak occurring at1.59 Å is characteristic of the porous catalyst. Figure 4 also shows that there is a remarkable increase in C-D peak intensity after the hydrogenation reaction has taken place. This shows that the conversion of benzene to cyclohexane during the hydrogenation reaction was complete. To confirm this claim, the discharged catalyst was extracted with n-hexane and analysed by GC-MS. The GC-MS analysis showed that only cyclohexane-d12 was present. These observations are similar to those made from the batch liquid phase reactions in which the same catalyst was used. Furthermore, the data is consistent with the literature data. In this case, both the literature data and the data obtained from the batch reactions that take place in liquid phase show that not much of cyclohexene and cyclohexadiene are formed. Theoretical studies also suggest that cyclohexene and cyclohexadiene constitute the major pathway of the reaction. Therefore, the theoretical studies are consistent with the present studies in this sense.
Time-Resolved Data
The structural changes across the molecular, atomic, and mesoscopic length scales is simultaneously illustrated by the progression of F(Q) as a function of time as shown in figure 5a. The kinetics that corresponds to various scales of length can be obtained by means of taking portions at specific Q-values with reference to time in the system. Figure 5b illustrates this as well as the matching exponential fits. The first diffraction peak, which has a Q value of 0.19 Å-1, is characterized by a rate constant, k0, of 2.146 h-1. In addition, the peak has a component, k1, of 0.138 h-1. Based on these values of the rate constant and component, the first diffraction peak is said to have both a fast rate constant and a slow component. On the other hand, the peaks with the following Q values exhibit an intermediate third rate: 1.2 Å-1, 3.1 Å-1, and 4.3 Å-1. It is important to note that the variation in contrast between the mesoporous substrate and the absorbed material is not the only factor contributing to the rise in the first diffraction peak’s intensity. Rather, the first diffraction peak increase is also attributed to a second process on another timescale.
When the data (1 < Q < 2π Å-1) was inspected, it was found that no other changes occurred at the same rate as the fast component (k0). Consequently, it is unlikely that there is a relationship between this first process and benzene. Besides, this observation shows that the first process is attributed to the dissociative adsorption of D2. It was also observed that upon the introduction of D2, the temperature increased rapidly and by a relatively large magnitude over the first thirty minutes as shown in figure 5c. This observation is consistent with the first process caused by the dissociative adsorption of D2. The slow component, k1, is also not likely to be related to benzene since it does not imply that a chemical change took place. Rather, it is most likely associated with the mass product transport within the pore (pore diffusion). The change observed at Q = 1.22 Å-1 is an indication of the type of molecular interaction involving molecules that are nearest to one another. On the other hand, the change observed at Q = 3.1 Å-1 as well as that observed at Q = 4.3 Å-1 are indicative of the chemical changes related to atoms occurring within the system. The time constants with which these three features change are similar. In addition, there is a correlation between all the three features and the formation of the cyclohexane product, and the reduction process. Based on these time constants, it can be deduced that the diffusion of liquid (k1) is the rate-determining step of the whole process. This is the case because the rate of liquid diffusion is the slowest of all the steps in the process. However, the reaction is not regulated by the dissociation of D2 (k0). Instead, it is governed by the process of hydrogenation (k2,3,4). These results help in clarifying the following process outline for the reaction, at room temperature conditions and 250 mBar D2:
D2g →2Dads,k0=2.146 h-1
C6D6l+6Dads → C6D12l, k2,3,4≈0.35 h-1
C6D12lpore diffusion, k1=0.136 h-1
Discussion
For many years, there has been an awareness of the need for more comprehensive information concerning the evaluation of kinetics of catalysts. For instance, the widely known “turnover frequency” (TOF) concept should be replaced with a concept that provides measure that is more quantitative. In this case, measures that provide information about the system’s rate equation are preferred. In the past, it has been difficult to attain such characterization, particularly for processes that are conducted in liquid phase and catalysed heterogeneously. This has been the case because the rate equation commonly contains processes that include both mass transport and surface reaction. Each of these processes has distinct rate constants. However, so far, none of the techniques that have been used has been able to provide detailed information. This study has shown that simultaneous kinetic information about chemical reaction and mass transport within the catalyst pore can be obtained using the total neutron scattering technique. In the present experiment, in order to understand how the design of the catalyst may be developed, one should consider the important role that liquid diffusion through the pores plays in determining the reaction rate. Indeed, while the addition of solvent with the aim of increasing the rate of diffusion can promote the performance of the catalyst, improving the surface reaction processes (increasing the reactivity or the number of the active sites) does not have a significant effect on catalyst performance. Based on this concept, the reaction kinetics and the temporal/spatial changes can be correlated in future. Consequently, it will be possible to obtain important information about the catalytic process.
Conclusions
This experiment has directly probed the kinetics occurring within the heterogeneous catalyst pores in the process of reducing benzene in liquid phase with the help of neutron diffraction technique. Therefore, the experiment provides a method that can be used to comprehensively examine the correlations between the structure and reactivity for these complex systems, thus making it possible to decouple mass transport effects in the catalyst reactions and the reactions in which not less than one step of the mechanisms involved require the adsorption of at least one reactant. Furthermore, the technique makes it possible for the spatial and structural data to be associated with the kinetic features of the underlying processes.
Acknowledgements
We are indebted to JohnsonMatthey and EPSRC for their support in providing financial support under the CasTech project. We also thank P. Hawkins and M. Kibble for their contribution during the experiments involving neutron diffraction. We also acknowledge and appreciate the support given by the beamtime allocation from the Science and Technology Facilities Council (experiment RB1220486, DOI: 10.5286/ISIS.E.24089729) in the experiments at the ISIS Pulsed Neutron and Muon Source.
