22.5.1 Catalytic Hydrodechlorination of Trichloroethylene The reactive properties of the bimetallic Fe/Ni nanoparticles in PVDF membranes dip-coated with PAA were examined by reduction of TCE in water at room temperature. Trichloroethylene contamination of groundwater is widely reported in the literature [ 22 ]. Figure 22.11 shows the batch reaction (at pH 6) of TCE with Fe/Ni (Ni = 25 wt%) nanoparticles in PAA/PVDF membranes. The TCE transformation rate can be described by a simple pseudo-fi rst-order model. Complete dechlorination of TCE was achieved within
(a) (d)
2 nm
(b) (Fe map) (Pd map)
HRTEM
100 nm
EDS mapping
(c)
Figure 22.10 Characterization of Fe/Pd nanoparticles. (a) Scanning transmission electron microscopy image of Fe/Pd (Pd = 2.3 wt%) nanoparticles; (b) energy dispersive spectroscopy (EDS) mapping image of Fe; (c) EDS mapping image of Pd; (d) high- resolution transmission electron microscope image of Fe/Pd nanoparticles
Savage_Ch22.indd 323
Savage_Ch22.indd 323 11/11/2008 3:52:26 PM11/11/2008 3:52:26 PM
2 hours. Ethane and Cl - were formed as the only products in the headspace and aqueous phase, respectively. No chlorinated intermediates were detected and 92 percent carbon balance and 93 percent Cl balance were obtained in the whole system, indicating a direct reaction pathway from TCE to ethane for bimetallic Fe/Ni (Ni = 25 wt%) nanoparticles.
The complete conversion of TCE to ethane with bimetallic nanoparticles is totally diff erent from the sequential reductive dechlorination with monometallic Fe system. For the Fe system TCE is transformed to dichloroethylene (DCE) to vinyl chloride (VC) and fi nally to ethylene and ethane [ 40 ]. The presence of the secondary metal on nanosized Fe changes the reaction pathway
0.0 0.5 1.0 1.5 2.0 2.5
0.00 0.04 0.08 0.12 (a)
TCE
Concentration (mM)
Time (h)
Ethane Mass balance Cl-/3
0.0 0.2 0.4 0.6 0.8 0
2 4 6 8 10
-Ln(C/C0)/ρm
Time (h) kSA = 0.44 l h-1m-2
(b)
TCE
Cis-DCE
Trans-DCE
1,1-DCE e- e-
e-
e-
VC e-
Ethylene H2
Ethane Sequential reduction pathway of TCE by dissociative electron transfer with Fe0 particles
Direct reduction pathway of TCE by catalytic hydrodechlorination with Fe/Ni or Fe/Pd nanoparticles
TCE Ethane
Figure 22.11 (a) Batch reaction of trichloroethylene (TCE) dechlorination and products formation (ethane and chloride) with Fe/Ni (Ni = 25 wt%, post coat Ni) nanoparticles in polyvinylidene fl uoride (PVDF) membranes dip-coated with polyacrylic acid (PAA). ρm = 0.2 g L–1; (b): Schematic diagram for reductive dechlorination pathways of TCE with iron and Fe/Ni or Fe/Pd systems. From [22].
Savage_Ch22.indd 324
Savage_Ch22.indd 324 11/11/2008 3:52:28 PM11/11/2008 3:52:28 PM
dramatically [ 9 , 10 ]. In the monometallic Fe system, the dechlorination mechanism is preferably explained by dissociative electron transfer resulting in the formation of less chlorinated radicals as intermediates [ 41 ]. While in the bimetallic system, Fe is considered as the reductant for water to generate hydrogen and TCE is dechlorinated by catalytic hydrodechlorination [ 42 ] in the presence of Ni, resulting in the direct reduction to ethane. Another advantage of coating the secondary metal is to prevent the conversion of the Fe 0 to an oxide form (Fe x O y ) that can deactivate the nanoparticle surface [ 41 ].
It has been reported in earlier studies that dechlorination reaction can be described by a pseudo-fi rst-order kinetic model. The TCE degradation rate by Fe-based bimetallic nanoparticles is considered as fi rst order in terms of both TCE concentration and the concentration of metals available in the solution.
Therefore, the following Equation 22.2 has been used [ 43 ] to describe this pseudo-fi rst-order reaction model.
= − SA s m
dC k a p C
dt (22.2)
Regression of the kinetic data can be used to determine the surface area normalized reaction rate constant k SA . Since k SA is the characterized fi rst-order reaction rate, it should be independent of variance of reaction conditions such as initial TCE concentration, metal mass and the volume of reaction system.
22.5.2 Eff ect of Dopant Material and Nanoparticle Structure In order to further understand the catalytic dechlorination mechanism and bimetallic nanoparticle reactivity, TCE dechlorination experiments were conducted with bimetallic nanoparticles with diff erent type of dopant metal and diff erent structure. Figure 22.12 shows the normalized rate constant ( k SA ) of TCE dechlorination using PAA/PVDF membrane based Fe/Cu nanoparticles, alloy Fe/Ni nanoparticles (simultaneous reduction of Fe and Ni), and core- shell Fe/Ni nanoparticles (post-coating Ni). The second metal composition was kept the same (25 wt%) for all the three reaction systems. As expected, the core-shell Fe/Ni nanoparticles exhibit higher TCE degradation rate than the alloy Fe/Ni nanoparticles. k SA for core-shell nano Fe/Ni is about four times higher than that of alloy nano Fe/Ni. It has been demonstrated that most Ni atoms are located at the out side of iron surface for the core-shell structure.
The alloy structure has a homogenous distribution of iron and nickel atoms inside particle, which results in lesser amount of Ni atoms on the surface. Since the Ni is the active sites for the catalytic hydrodechlorination reaction and the reaction only takes place at the particle surface, the lower k SA for alloy Fe/Ni nanoparticles is due to the less active sites (Ni atoms) on the surface.
Compared to the Fe/Ni nanoparticles, Fe/Cu nanoparticles show a much slower reaction rate. k SA for core-shell Fe/Cu nanoparticles is about 30 times
Savage_Ch22.indd 325
Savage_Ch22.indd 325 11/11/2008 3:52:29 PM11/11/2008 3:52:29 PM
0 0.05
0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 (a)
0 0.5 1 1.5 2
Fe/Ni(25 wt% Ni) in PVDF/PAA, post coating Ni
Fe/Ni(25 wt% Ni) in PVDF/PAA, simultaneous reduction of Fe and Ni Fe/Cu(25 wt% Cu) in PVDF/PAA, post coating Cu
-Ln(C/C0) as × ρm
kSA= 0.014 L h-1 m-2 Fe/
F Ni
Time (h) (b)
0.0 0.2 0.4 0.6 0.8 1.0
0.0 0.2 0.4 0.6 0.8 1.0
C2H4/(C2H4+C2H6)
TCE Conversion Nano Fe/Cu (Cu=25wt%) Nano Fe/Ni (Ni=25wt%)
kSA= 0.12 L h-1 m-2 kSA= 0.45 L h-1 m-2
Figure 22.12 (a) Trichloroethylene (TCE) degradation with various bimetallic nanoparticles in polyacrylic acid (PAA)/polyvinylidene fl uoride (PVDF) membranes:
((o) Fe/Ni (Ni = 25 wt%, post-coating Ni); ((•) Fe/Ni (Ni = 25 wt%, simultaneous reduction of Fe and Ni); (Δ) Fe/Cu (Cu = 25 wt%, post-coating Cu). (b) Ethylene and ethane formation from TCE dechlorination with Fe/Ni (Ni = 25 wt%, post-coating Ni) nanoparticles and Fe/Cu nanoparticles (Cu = 25 wt%, post-coating Cu).
Savage_Ch22.indd 326
Savage_Ch22.indd 326 11/11/2008 3:52:29 PM11/11/2008 3:52:29 PM
lower than core-shell Fe/Ni nanoparticle at the same dopant composition. In order to understand the reactivity diff erence between Cu and Ni, the product formation was measured for TCE degradation with Fe/Cu and Fe/Ni nanoparticles. As shown in Fig. 22.12 , the dominating product for Fe/Cu system is ethylene, suggesting the main reduction pathway is by electron transfer. In contrast, ethane is the only product formed in the Fe/Ni system due to the catalytic hydrogenation mechanism. This observation is consistent with the result reported in the literature [ 44 ] for the cis -dichloroethylene ( cis - DCE) dechlorination by Fe/Ni and Fe/Cu particles. Ni and Pd are proven the most active dopant for the reductive dechlorination reaction due to the high hydrogenation activity.
22.5.3 Catalytic Hydrodechlorination of Selected Polychlorinated Biphenyls
To investigate the catalytic properties of Fe/Pd nanoparticles synthesized in PAA/PVDF membranes, we studied the reductive hydrodechlorination of selected PCBs [ 23 ] using the bimetallic nanoparticles. PCBs are among the most important chlorinated aromatic compounds that cause a stringent environmental problem due to their hydrophobic nature and excellent chemical stability. The dechlorination mechanism and kinetic rates were investigated using membrane supported Fe/Pd nanoparticles. In order to understand and quantify the role of second dopant metal, we studied the dechlorination rates as a function of Pd content on Fe as well as the reaction temperature.
Figure 22.13 shows the dechlorination of 15.6 mg/L PCB 77 in 65/35 vol.%
ethanol/water solution with Fe/Pd (Pd: 2.3 wt%) immobilized inside PAA/
PVDF membranes. High concentration of ethanol in the solution matrix was used because of the lower solubility of PCB 77 in water. As shown in the fi gure, the membrane supported Fe/Pd nanoparticles exhibit extremely fast dechlorination rate. Complete degradation of PCB 77 by Fe/Pd in PAA/PVDF membrane was achieved within 2 hours. Biphenyl was formed as the main dechlorination product. PCB77 was completely transformed to biphenyl after 2 hours. The degradation of PCB77 by Fe/Pd nanoparticles occurred in a sequential reduction pathway, which is indicated by the detected less chlorine intermediates. All the PCB intermediates were only identifi ed in the low concentration level within 1 hour. It has been proven in the literature [ 8 ] that non-ortho-chlorinated PCB congeners dechlorinate faster than the ortho- chlorinated isomers due to the eff ect of higher steric hindrance for ortho-position.
The reactivity of the chlorine substitutents decreases in the order para ≈ meta
> ortho [ 45 ]. Due to the increasing torsion angles with the increase of ortho substitution, non-ortho substituted congeners could adsorb in a closed planar position with nanoparticles, which is benefi cial for the reductive dechlorination.
The more positive reduction potentials measured in the literature [ 46 ] also support the increased reactivity of non-ortho substituted congeners.
Savage_Ch22.indd 327
Savage_Ch22.indd 327 11/11/2008 3:52:30 PM11/11/2008 3:52:30 PM
22.5.4 Dechlorination Effi ciency of Diff erent Polychlorinated Biphenyls
In theory, there are 209 diff erent PCB congeners. In order to understand and quantify diff erent PCBs dechlorination reactions, degradation of PCB4, PCB44, PCB77 were conducted with same Fe/Pd nanoparticles immobilized in PAA/PVDF membrane. Figure 22.14 plots the yield and selectivity for biphenyl as a function of reaction time. It can be seen that biphenyl yield is about the same for all three PCBs, whereas the biphenyl selectivity varies signifi cantly. Although the biphenyl selectivity for the three diff erent PCBs reaches about the same value at the end of reaction, PCB4 shows much higher selectivity than the other two PCBs within 1 hour. The higher selectivity is due to the negligible formation of intermediate 2-chlorobiphenyl as the only intermediate is detected in trace level. The degradation of PCB77 and PCB44 has nine and fi ve diff erent intermediates, respectively. Each individual intermediate is detected in trace amount, but total amount of intermediates are not negligible. This indicates that the degradation of PCB77 and PCB44 does not result in the correspondingly equal amount formation of biphenyl.
Biphenyl is formed by sequential reduction pathway. The lower biphenyl selectivity is due to the complex reduction pathway for higher number chlorine substituted PCB congeners.
0.0 0.5 1.0 1.5 2.0 2.5
0.000 0.015 0.030 0.045 0.060
Concentration (mM)
Time (h)
PCB77 (3,3',4,4') PCB37 (3,4,4') PCB35 (3,3',4) PCB15 (4,4')
PCB12and13 (3,4' and 3,4) PCB11 (3,3')
PCB3 (4) PCB2 (3) Biphenyl Carbon Balance
Figure 22.13 Batch reaction of PCB77 with Fe/Pd (Pd = 2.3 wt%) in pore-fi lled polyacrylic acid (PAA)/polyvinylidene fl uoride (PVDF) membrane. Metal loading: 0.8 g L–1. Initial PCB77 concentration: 15.6 mg L–1.
Savage_Ch22.indd 328
Savage_Ch22.indd 328 11/11/2008 3:52:30 PM11/11/2008 3:52:30 PM
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0 0.5 1 1.5 2
Time (h)
Biphenyl Yield
PCB77 (3,3',4,4'-tetrachlorobiphenyl) PCB4 (2,2'-dichlorobiphenyl)
PCB 44 (2,3,2’,5’- tetrachlorobiphenyl)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0 0.5 1 1.5 2
Time (h)
Biphenyl Selectivity PCB77 (3,3',4,4'-tetrachlorobiphenyl)
PCB4 (2,2'-dichlorobiphenyl)
PCB 44 (2,3,2’,5’- tetrachlorobiphenyl) (a)
(b)
Figure 22.14 Polychlorinated biphenyl 4 (PCB4), PCB44, and PCB77 dechlorination with Fe/Pd nanoparticles in polyacrylic acid (PAA)/polyvinylidene fl uoride (PVDF) membranes. Metal loading is kept same (0.8 g L–1). (a) Biphenyl yield; (b) biphenyl selectivity. (Biphenyl yield = biphenyl formed/initial amount of parent PCB; biphenyl selectivity = biphenyl formed/parent PCB consumed.)
Savage_Ch22.indd 329
Savage_Ch22.indd 329 11/11/2008 3:52:30 PM11/11/2008 3:52:30 PM
22.5.5 Catalytic Activity as a Function of Palladium Coating Content
In order to understand the role of Pd as the second dopant, the batch dechlorination rates of DiCB (2,2′-dichlorobiphenyl) were measured as a function of Pd coating content ( Fig. 22.15 ). The k SA is 0.017, 0.068, and 0.166 L h –1 m –2 for Fe/Pd nanoparticles with 0.6, 2.3, and 5.6 wt% Pd, respectively.
In the bimetallic system, the role of Fe is to generate hydrogen by corrosion reaction, whereas Pd serves as the catalyst and the chlorine atom in DiCB is mainly replaced by hydrogen on the Pd surface [ 8 ]. Therefore the Pd atoms are considered as the surface reactive sites for the dechlorination of DiCB. The variation of the k SA as a function of Pd content is due to the diff erence of reactive sites. By normalizing the k SA in terms of Pd content (reactive sites), we found the same reaction rate of Fe/Pd nanoparticles. The modifi ed reaction model shown in Equation 22.3, developed by Johnson, provided a better way to understand and quantify the eff ect of variation in reactivity of diff erent metal system [ 23 , 43 ].
2
SA m s s m
−dC = = Γ k p a C k a p C
dt (22.3)
where k 2 is the second order rate constant at a particular type of site (L h –1 mol –1 ) and Γ is the surface concentration of reactive sites (mol m –2 ). In this model, k SA is expressed as the product of k 2 and Γ, which is more reasonable when dechlorination reaction preferentially occurs at the reactive catalytic surface sites (bimetallic system).
Based on the 30 nm average diameter of nanoparticles, we calculated the Pd coverage and surface Pd atoms for diff erent Fe/Pd nanoparticles by using a Pd atom cross-sectional area of 0.0787 nm 2 [ 47 ]. Our calculations indicate that Fe/
Pd nanoparticles with 0.6, 2.3, and 5.6 wt% Pd have 0.1, 0.4, and 0.97 layers of Pd atoms, respectively. Since maximum Pd coverage is less than one layer, all the Pd atoms are considered as surface reactive sites. Γ for the three diff erent nanoparticles with 0.6, 2.3, and 5.6 wt% Pd is 2.20 × 10 –6 , 8.43 × 10 –6 , and 2.05 × 10 –5 mol m –2 , respectively. It should be noted that total surface area was used in all k SA calculations. By applying Γ in the Equation 22.3, k 2 was determined to be 7,727, 8,066, and 8,098 L h –1 mol –1 , respectively. The enhanced reaction rate ( k SA ) is only due to the increase of surface Pd atoms.
High-resolution STEM-EDS mapping images were also acquired in Fig. 22.15 to compare the Pd atoms distribution for diff erent Pd coating nanoparticles.
The STEM-EDS mapping technique presents us a 2D image of 3D sample in transmission [ 48 ]. All the Fe/Pd nanoparticles show a core/shell structure with Fe rich in the core region and Pd rich in the shell region. More Pd atoms were deposited on the iron surface and the Pd shell layer became denser with the increased Pd content. In spite of the limited spatial resolution in the EDS mapping, the distribution of Pd atoms is still in qualitative agreement with the result based on the calculation. This result implies that a uniform Pd coating
Savage_Ch22.indd 330
Savage_Ch22.indd 330 11/11/2008 3:52:31 PM11/11/2008 3:52:31 PM
0 1 2 3 4 5 0
1 2 3 4
0.6 wt% Pd 2.3 wt% Pd 5.6 wt% Pd
-Ln (C/C0)
Time (h) kSA = 0.068 Lh-1 m-2
kSA = 0.017 L h-1 m-2 kSA = 0.166 L h-1m-2
Fe
Pd
50 nm 0.6 wt% Pd
30 nm Fe
Pd
5.6 wt% Pd Fe
Pd
50 nm 2.3 wt% Pd
Figure 22.15 Best linear fi t of kSA for dechlorination of DiCB (2,2’- dichlorobiphenyl) with various Fe/Pd nanoparticles in polyacrylic acid (PAA)/polyvinylidene fl uoride (PVDF) membranes. Metal loading: 0.8 g L–1. Bottom images: high-resolution Scanning transmission electron microscopy (STEM)–energy dispersive spectroscopy (EDS) mapping of Fe and Pd in nanoparticles.
Savage_Ch22.indd 331
Savage_Ch22.indd 331 11/11/2008 3:52:31 PM11/11/2008 3:52:31 PM
with controllable thickness can be obtained by post-reduction of Pd 2+ with Fe nanoparticles immobilized in membrane phase.