Iron Nanoparticles Embedded in Graphitic Carbon Matrix as Heterogeneous Catalysts for the Oxidative C?N Coupling of Aromatic N?H Compounds and Amides

Accepted Article

Title: Iron nanoparticles embedded in graphitic carbon matrix as

heterogeneous catalysts for the oxidative C-N coupling of

aromatic N H compounds and amides

Authors: Jinbao He, Amarajothi Dhakshinamoorthy, Ana Primo, and

Hermenegildo García

This manuscript has been accepted after peer review and appears as an

Accepted Article online prior to editing, proofing, and formal publication

of the final Version of Record (VoR). This work is currently citable by

using the Digital Object Identifier (DOI) given below. The VoR will be

published online in Early View as soon as possible and may be different

to this Accepted Article as a result of editing. Readers should obtain

the VoR from the journal website shown below when it is published

to ensure accuracy of information. The authors are responsible for the

content of this Accepted Article.

To be cited as: ChemCatChem 10.1002/cctc.201700429

Link to VoR: http://dx.doi.org/10.1002/cctc.201700429

A Journal of

www.chemcatchem.org

10.1002/cctc.201700429

ChemCatChem

Iron nanoparticles embedded in graphitic carbon matrix as

heterogeneous catalysts for the oxidative C-N coupling of aromatic N-H

compounds and amides

Jinbao He,^aAmarajothi Dhakshinamoorthy,*^bAna Primo^aand Hermenegildo Garcia*^a,c

^aInstituto Universitario de Tecnología Química CSIV-UPV, Universitat Politècnica de

València, Av. De los Naranjos s/n, 46022, Valencia, Spain.

^bSchool of Chemistry, Madurai Kamaraj University, Tamil Nadu, India 625 021.

^cCentre of Excellence for Advanced Materials Research, King Abdulaziz University,

Jeddah, Saudi Arabia. E-mail: hgargia@qim.upv.es

Abstract

Fe or Co nanoparticles (NPs) and two nanoparticulate Fe-Co alloys having

different Fe/Co atomic ratio with average particle size ranging from 10.9 to 26.5 nm

embedded in turbostratic graphitic carbon matrix have been prepared by pyrolysis at

o

900 C under inert atmosphere of chitosan powders containing Fe²⁺and Co²⁺ions in

various proportions. The resulting Fe/Co NP@C samples have been evaluated as

heterogeneous catalysts for the oxidative C-N coupling of amides and aromatic N-H

compounds. It was observed that sequential addition of two aliquots of tert-butyl

hydroperoxide (TBHP) in an excess of N,N-dimethylacetamide (DMA) as solvent affords

the corresponding coupling product in high yields, the most efficient catalyst being Fe

NP@C. Fe NP@C is reusable and exhibits a wide scope. The catalytic activity of Fe is

supported by the use of highly pure Fe salt and by the observation that purposely

addition of Cu²⁺impurities even plays a detrimental effect on the catalytic activity.

Mechanistic studies quenching with TEMPO have shown that the amide radical is the

key reaction intermediate and the role of Fe NP@C is to generate the first radicals by

TBHP decomposition.

- 1 -

10.1002/cctc.201700429

ChemCatChem

Introduction

Metal nanoparticles (MNPs) are catalysts for a wide range of organic reactions

including oxidations, reductions, homo- and cross couplings, and rearrangements

among other transformations.^[1-4]One general problem of MNPs that can limit its

application in catalysis is their low stability under reaction conditions due to the

tendency of small MNPs to increase their particle size.^[4]Large particle size is generally

associated to low catalytic activity.^{[2, 5, 6]}A general strategy to minimize particle growth

is incorporation of MNPs in porous polymeric or inorganic matrices that by interaction

with the particle or by confinement stabilizes its size.^[7-10]Besides stabilization of

particle size distribution, the matrix can assist the catalytic reaction by adsorption or

preactivation of the substrates, there being a common observation that the nature and

structure of the matrix exerts a strong influence on the intrinsic activity of these

MNPs.^[11-13]

In this context, it has been recently reported that MNPs and metal oxides obtained

by pyrolysis of metal complexes can exhibit a unique catalytic activity due to the

favorable influence of the carbonaceous matrix formed in the process surrounding the

MNPs.^{[14, 15]}Similar to the pyrolysis of molecular metal complexes, metal organic

frameworks have also been used as precursors of MNPs and metal oxides wrapped

with carbon.^[16-21]Although active carbons and other types of carbons are among the

preferred supports of MNPs,^[22-24]the difference in the present case is the pyrolytic

formation procedure of the carbon-wrapped MNPs, the type of carbon that surrounds

the MNPs and the unique activity and stability of MNPs that can be achieved.

Other aspect to be considered is the nature of the metal. Although NPs of noble

metals are widely used as catalysts, for the sake of sustainability and affordability,

there is an increasing interest in exploiting the potential of abundant, base metals as

catalysts.^{[25, 26]}In this context, Fe being the most abundant metal has been under the

- 2 -

10.1002/cctc.201700429

ChemCatChem

spotlight with regard to its ability to promote organic reactions and couplings in

particular.^{[26, 27]}

In this context, in the present article a general procedure to prepare MNPs

embedded in a graphitic carbon matrix (denoted as MNP@C) from natural

biopolymers is presented and used to obtain a series of carbon wrapped Fe, Co and Fe-

Co alloys. The catalytic activity of these MNP@C samples for the oxidative C-N cross

coupling of aromatic N-H compounds and amides has been determined and details of

the reaction mechanism involving radicals unveiled. Related precedents have reported

the oxidative C-N coupling of aromatic N-H compounds with amides using ditert-butyl

peroxide as oxidizing agent and homogeneous Fe²⁺catalysts^{[28, 29]}as well as a Cu metal

organic framework as heterogeneous catalyst.^[30]However the use of MNPs wrapped

on graphitic carbon that are currently under quite intense investigation due to the

unique catalytic activity frequently observed^[31-34]has not yet been reported as

catalysts for this reaction.

Results and discussion

Catalyst preparation

o

Preparation of MNPs@C was performed by submitting to pyrolysis at 900 C under

inert atmosphere powdered samples of chitosan containing adsorbed Fe²⁺or Co²⁺

metal ions or the corresponding Fe²⁺-Co²⁺mixture. These chitosan samples containing

Fe²⁺/Co²⁺metal ions were obtained by water evaporation of acid aqueous solutions of

chitosan to which the appropriate amounts of the corresponding transition metal ions

(5 wt% of metal ion with respect to dry chitosan weight) were added. Scheme 1

illustrates the preparation procedure. In a series of previous studies,^{[35, 36]}it has been

shown that pyrolysis of chitosan under inert atmosphere renders a N-doped

turbostratic graphitic carbon residue with some residual oxygen content (about 8

wt.%) that can be easily exfoliated by sonication. When these samples contain metals,

- 3 -

10.1002/cctc.201700429

ChemCatChem

they can spontaneously segregate in a different phase, while becoming simultaneously

reduced to their metallic state by the so-called carbochemical reduction.^[37]Thus, upon

heating at high temperature a mixture of metal oxide and carbon, the oxide becomes

reduced to metal, while carbon becomes oxidized to CO₂.^[38]The same behavior should

occur in the present case in the pyrolysis of Fe²⁺/Co²⁺-containing chitosan samples,

resulting in the formation of the expected MNP@C. Table 1 summarizes the series of

four MNP@C samples prepared in the present study and the elemental composition of

these materials according to chemical analysis. Theχvalue in Fe-Coχ NP@C samples

denotes the Fe/Co atomic ratio.

Scheme 1. Steps in the preparation of the MNP@C samples: (i) dissolution in acetic acid

aqueous solutions; (ii) water evaporation; (iii) pyrolysis at 900 ℃ under Ar flow.

Table 1. Chemical analysis of the MNP@C samples. The percentage of metals was determined

by ICP analysis and the percentage of N, C, H was determined by combustion analysis. All

values are in wt%, and it is assumed that the difference to 100 % is due to O content.

Sample

N

C

H

Fe

Co

Fe NP@C

0.98

89.76

85.23

86.02

87.50

0.14

0.16

0.17

1.02

0.37

0.62

-

Fe-Co1.42 NP@C 0.97

Fe-Co0.86 NP@C 1.04

0.26

0.72

0.61

Co NP@C

1.36

- 4 -

10.1002/cctc.201700429

ChemCatChem

Besides chemical analysis, the samples obtained according to Scheme 1 were

characterized by XRD, Raman spectroscopy and electron microscopy. In XRD, the

presence in significant proportion of the corresponding metal carbides or metal oxides

was ruled out due to the absence or very low intensity of the corresponding diffraction

peaks. Thus, although particularly in the case of Fe NP@C, the presence of some weak

[39]

diffraction peaks some of the compatible with the presence of cementite (Fe₃C)

appearing at 2 37 (112) and 39 (200) ^owere recorded, their intensity was very low. In

contrast, those peaks indicating the presence of Fe or Co metals were very intense. In

the case of samples containing mixtures of Fe and Co, the minor variation in the

position of the diffraction peaks at 2 about 44 ^osupports that the NPs are real Fe-Co

alloys, rather than independent Fe and Co NPs. Figure 1 shows XRD patterns of three

selected MNP@C samples. By using Scherrer equation, the average sizes of MNPs of

the four samples were estimated as 26.5, 16.3, 17.0, 10.9 nm for Fe NP@C, Fe-Co1.42

NP@C, Fe-Co0.86 NP@C, Co NP@C, respectively, which agree well with the statistical

average particle sizes obtained from TEM measurements. It should be commented that

although no changes was observed in the XRD patterns after sample preparation and

storage of the solids in vials at the ambient atmosphere, it is very likely that the

outermost external part of the metal surface can become partially oxidized.

Co(111)

(c)

Co(200)

Co(220)

Fe-Co(110)

Fe-Co(211)

Fe-Co(200)

(b)

Fe(110)

Fe(211)

Fe(200)

(a)

0

20

40

2θ(°)

60

80

- 5 -

10.1002/cctc.201700429

ChemCatChem

Figure 1. XRD patterns of the MNP@C prepared according to Scheme 1: (a) Fe NP@C; (b) Fe-

Co1.42 NP@C; (c) Co NP@C. The broad peak at 2θ value of 27^ocorresponds to the graphitic

carbon residue and some weak peaks in plot (a) could indicate the presence of some Fe₃C (see

text).

The nature of carbon residue derived from chitosan was determined by Raman

spectroscopy. Figure 2 shows the Raman spectrum for Fe NP@C. Very similar Raman

spectra to those previously reported for turbostratic graphitic carbons lacking any

metal were recorded for all the samples.^[36]Thus, the expected G and D bands

appearing at 1596 and 1354 cm^-1, respectively, were recorded together with two 2D

defined peaks at 2940 and 2686 cm^-1. This is in accordance with the presence in the

graphene sheet of defects due to carbon vacancies, holes and the residual N atoms

from chitosan.^[36]A quantitative indicator of the defect level of graphenic sheets can

be given by determining the relative intensity of the G vs. the D band (I_G/I_D) that in the

present case was 1.26 for the four MNP@C samples, corresponding to a defective,

doped G. Interestingly, no vibration peaks that could be attributed to the presence of

metal oxides that should appear at low wavenumber values were observed for any of

the MNP@C samples before their use as catalyst, in good agreement with XRD

patterns showing that no oxides are detectable. In contrast, the recovered Fe NP@C

catalyst after the C-N coupling between benzimidazole and N,N-dimethylacetamide

(DMA) showed the presence of iron oxide indicating that the oxidation by tert-butyl

hydroperoxide (TBHP) takes place under the experimental conditions.

- 6 -

10.1002/cctc.201700429

ChemCatChem

6000

5000

4000

3000

2000

1000

0

500

1000

1500

2000

2500

3000

3500

Raman shift (cm^-1)

Figure 2. Raman spectra of fresh (a) and the recovered (b) Fe NP@C catalyst after the N-C

coupling reaction.

To prove accessibility to Fe and Co NPs even if they are embedded within the

graphitic carbon matrix, the Fe NP@C and Co NP@C samples were treated with an

aqueous solution of diluted HCl, observing that between 80 and 90 % of the initial

metal content becomes dissolved under these conditions (see experimental section).

This proves that the surface of Fe and Co metals are accessible to interact with TBHP

and reagents in spite their embedment in a carbon matrix.

As expected, TEM imaging of the MNP@C samples detected the presence of MNPs

embedded in a carbonaceous matrix. Figure 3 shows a selected set of images to

illustrate the structure and morphology of the Fe NP@C samples and images of other

samples are presented in Figures S.1-S.3. MNPs were clearly detectable using dark field

- 7 -

10.1002/cctc.201700429

ChemCatChem

that also allows estimation of the particle size distribution. From these images, it was

clear that the MNP size was not homogeneous and that there was a broad dispersion

of sizes below 50 nm. However, the corresponding histogram of particle size counting

a statistically relevant number of particles shows that particles of sizes between 5 and

10 nm are the most abundant in terms of number in the MNP@C samples.

Elemental mapping of Fe and Co shows that there is coincidence in the location of

these two elements (see Fig. S5 in supporting information), thus confirming the XRD

information that the MNPs are real Fe-Co alloys.

Figure 3. TEM images of Fe NP@C at low (a,b) and higher (c) magnification that show the

distribution of Fe NPs embedded in the graphitic turbostratic carbon (c). Panel (d) shows an

- 8 -

10.1002/cctc.201700429

ChemCatChem

image taken under dark field and its inset shows the statistical particle size distribution of Fe

NPs.

The relatively small particle size is remarkable considering the high temperature

o

(900 C) to which the MNP@C samples are submitted during the preparation

procedure and the time taken in the pyrolysis (6 h). In one of the most common

procedures, Fe NPs of small particle size similar to those obtained in the present case

were obtained by chemical reduction using reverse phase emulsions employing

oleylamine-water emulsions.^[40]In this regard, although there are also particles of

much larger size, the present procedure allows obtaining in a single step Fe/Co NPs of

small diameters and the carbon matrix in which it is embedded. It is suggested that, as

claimed in previous cases, chitosan and the carbonaceous residues derived therefrom

during the different phases of the pyrolysis are thwarting and restricting the growth of

the Fe/Co NPs by interacting with them.^[41]In related precedents, it has been found

[41]

[45]

that Au,^{[42, 43]}Cu

and even TiO₂^[44]and CeO₂

NPs of relatively small size are

formed when the procedure starts with the corresponding metal ions embedded in

chitosan.

Figure 4. C 1s (left) and Fe 2p (right) peaks of the high resolution XP spectra of Fe NP@C.

- 9 -

10.1002/cctc.201700429

ChemCatChem

The composition of the uppermost part of the Fe NP@C catalyst was also

characterized by XPS analysis and the observed peaks as well as the best fitting to

individual components are given in Figure 4. The deconvoluted XPS of C 1s spectrum

(Figure 4, left) can be fit to three main peaks at 283.0, 284.5 and 288.5 eV that can be

assigned to C en defects, graphenic C atoms and sp³C atoms bonded to oxygen,

respectively. Furthermore, analysis of Fe 2p spectrum (Figure 4, right) reveals the

existence of Fe²⁺and Fe³⁺at the binding energy values of 709 and 711 eV, respectively.

The presence of Fe^2+/3+on the outermost part of Fe NPs is relevant from the catalytic

point of view and complements the information by XRD that indicates that the

predominant oxidation state of Fe is Fe⁰. It has to be noted that XPS analysis probes

the composition of the outermost surface of the NPs, while XRD probes the whole

sample.

Oxidative C-N coupling

As commented in the introduction the purpose of the present study was to

evaluate the catalytic activity of the Fe/Co NP@C samples as heterogeneous catalysts

for the oxidative C-N coupling. Oxidative C-N couplings have the advantages that do

not require halide derivatives as substrates. Preliminary catalytic screenings were

carried out using Fe NP@C as catalyst, optimizing the amount of TBHP, solvent and

temperature for the coupling of benzimidazole and DMA. The reaction is indicated in

Scheme 2. Initially, 1,2-dichloroethane (DCE) was used as solvent, since the available

literature data suggested that this could be one of the most suitable solvents for the

reaction.^[28]However, in contrast to the report in the literature,^[28]under our reaction

conditions the C-N coupling product (3a) was observed as a primary, but instable

product, undergoing a subsequent conversion to N-(2-chloroethyl)benzimidazole (3b)

as final and stable product. Figure 5 shows a time-conversion plot of the catalytic

- 10 -

10.1002/cctc.201700429

ChemCatChem

oxidative C-N coupling using DCE as solvent. 3b clearly derives from the reaction of

benzimidazole 3a with the solvent. Thus, in order to avoid the decomposition of 3a,

the reaction was attempted in other solvents. It was observed, however, that the

reaction completely fails using tert-butanol, toluene, acetonitrile or ethyl acetate as

solvents.

Scheme 2. Oxidative C-N coupling of benzimidazole (1a) and DMA (2a) promoted by TBHP and

Fe NP@C.

80

60

40

20

0

5

10

15

20

25

Time(h)

Figure 5. Time-yield plots determined by GC for the oxidative C-N coupling promoted by Fe

NP@C in DCE as solvent: (■) 3a; (●) 3b. Reaction conditions: 1a (0.1 mmol ), Fe NP@C (1.6

mol%), 2.5 equiv oxidant (5 M solution of TBHP in decane), DCE (2 ml), Ar (2.5 bar), 110 ^oC.

In contrast with the previous negative results in some solvents, C-N coupling

between benzimidazole and DMA could be also observed using an excess of DMA that

was acting under these conditions as reagent and solvent. In DMA, however, although

the selectivity to 3a was almost complete, conversion of 1a was unsatisfactorily low,

- 11 -

10.1002/cctc.201700429

ChemCatChem

about 35 %. Figure 6 shows the corresponding temporal evolution of the formation of

3a under these conditions.

70

60

50

40

30

20

10

adding an extra amount of TBHP

0

5

10

15

20

Time(h)

Figure 6. Time-yield plots determined by GC for two twin coupling reactions under the

same conditions, adding an extra amount of TBHP (1.5 equiv TBHP from a 5 M solution of

TBHP in decane) after 4 h reaction time in one of the experiments (■), but not in the

other (▼). The temporal profiles of control experiment without any catalyst (▲) and a

deactived catalyst obtained by boiling Fe NP@C in acetamide before use (●) are also

plotted. Reaction conditions: 1a (0.1 mmol ), Fe NP@C (1.6 mol%), DMA (2 ml), Ar (2.5

bar), 110 ^oC, 2.5 (or 4.0) eq. of TBHP from a 5 M TBHP solution in decane.

A likely reason for the low yield of 3a could be the spurious decomposition of

TBHP without promoting C-N coupling. This possibility was confirmed by performing an

additional experiment in which the oxidative C-N coupling reaction was initially started

under the normal conditions (see caption of Figure 6), but adding an additional 1.5 eq.

- 12 -

10.1002/cctc.201700429

ChemCatChem

amount of TBHP when the reaction rate decreased considerably that is at about 4 h

reaction time (see Figure 6). It was observed that upon addition of a second amount of

TBHP, a significant increase in the yield of 3a from about 30 to 50 % was quickly

achieved, reaching a final 3a yield of 65 % at 20 h. It is interesting to mention that

under the present experimental conditions no other product besides 3a was observed

and unreacted starting material 1a was also recovered at final reaction time. Thus, the

incomplete yield is probably due to the spurious decomposition of TBHP.

Considering the chemical structure of DMA and its reactivity through a N-methyl

substituent, a structurally related solvent that could also be suitable to overcome the

limitation of other solvents would be acetamide. Surprisingly, no benzimidazole

conversion was observed using acetamide. Titration of TBHP after 24 h reaction time

indicates that only a residual percentage of 11 % of the initial TBHP remains, even

though no 3a was formed. It could be due that acetamide could act as a poison of Fe

NP@C catalyst by strong hydrogen bonding of acetamide to the metal surface. This

possibility was checked by performing an experiment in which Fe NP@C was boiled

o

first in acetamide at 110 C for 4 h, then, recovered, washed with DMA and used as

catalyst in DMA as solvent. It was observed that the catalytic activity of acetamide-

treated Fe NP@C was much lower than that of the fresh untreated sample (see Figure

6, plot ●). This decrease in catalytic activity after contacting with acetamide suggests

that the material becomes deactivated by this solvent, thus, providing some hints to

rationalize the negative results observed in acetamide, in spite of its similarity with

DMA.

After having found that an excess of DMA is a suitable medium to perform the

oxidative coupling and optimized reaction conditions, the influence on the catalytic

performance of the percentage of Co on the Fe/Co NP@C catalyst was checked by

- 13 -

10.1002/cctc.201700429

ChemCatChem

performing a series of reactions under the same conditions, but using the other

MNP@C catalysts of the series. As it is shown in Figure 7, it was found that the

presence of Co in the catalyst is detrimental for the activity that undergoes a gradual

decrease as the percentage of Co increases.

70

60

50

40

30

20

10

0

5

10

15

20

25

Time(h)

Figure 7. Time-yield plots determined by GC for the oxidative C-N coupling of benzimidazole

and DMA in the presence of a series of catalysts: (■) Fe NP@C; (●) Fe-Co1.42 NP@C; (▲) Fe-

Co0.86 NP@C; (▼) Co NP@C. Reaction conditions: 1a (0.1 mmol), catalyst (1.6 mol%), two

consecutive additions of 2.5 and 1.5 equiv of TBHP (5 M solution in decane), DMA (2 ml), Ar

(2.5 bar), 110 ^oC.

To gain insight into the heterogeneity of the MNP@C catalyzed reaction, the solid

catalysts were removed by filtration after 24 h reaction, and the clear solutions

analyzed by ICP analysis. Irrespective of the MNP@C material used, the Fe amount

leached to the solution was 0.72 g, while, for instance, the initial Fe amount of the Fe

- 14 -

10.1002/cctc.201700429

ChemCatChem

NP@C (1.6 mol%) with a Fe loading of 1.02 wt % was 93 g. Thus, more than 99.2 % of

the Fe initially present was retained in the catalyst. It was, however, observed that

filtration of the Fe NP@C catalyst does not completely stops the reaction. Once

initiated the reaction, formation of C-N coupling product also occurs even after

filtration of the Fe NP@C catalyst at 4 h, although in about 40 % lower percentage than

in its presence. This formation of product 3a in the absence of catalyst is proposed to

be related to the reaction mechanism involving C-centered radicals as reaction

intermediates and the role of Fe NP@C as radical initiator as it would be discussed

below.

To learn more about the role of leached Fe species as homogeneous catalyst,

three control experiments using soluble precursors with Fe in different initial oxidation

states were carried out. Specifically the compounds used were Fe(CO)₅, FeCl₂or

Fe(OAc)₃, with Fe in 0, +II and +III oxidation state. The amount of these materials used

was such to provide a weight of Fe about twice (1.5 g) determined by ICP at the final

reaction time. It was, however, observed that although compound 3a was detected at

final reaction time, the percentage of C-N coupling was in all cases below 5 %,

indicating that the contribution of the leached Fe to the observed yield of 3a is

negligible.

Reusability and catalyst stability.

When working on heterogeneous catalysis, reusability and catalyst stability are

important issues that have to be addressed. Aimed at determining catalyst stability,

the same Fe NP@C sample was submitted to a series of consecutive uses under the

optimal reaction conditions. As shown in Figure 8, a gradual decrease in the catalytic

activity was observed. Careful perusal of the time-conversion plots indicate, however,

that the decay in the catalytic activity is mainly apparent due to the lower 3a yield

- 15 -

10.1002/cctc.201700429

ChemCatChem

achieved in the first TBHP addition (see data points at times shorter than 4 h in Figure

8), while the profile corresponding to the second addition is very similar with no decay

in the reaction rates for five reuses. Thus, the decay observed in the yield at final

reaction time seems to be due to the operation of a very fast spurious decomposition

of TBHP occurring for the used Fe NP@C sample at the initial stages of the recycling

and/or some unavoidable loss of catalyst mass during reuse. It seems, however, that

after this variation at initial times, the time-conversion plots are parallel upon reuse,

indicating that the reaction rates for the oxidative C-N coupling do not change upon

reuse and that the activity of Fe NP@C sample remains constant besides the spurious

TBHP decomposition. It could be that this fast TBHP decomposition is an artefact

caused by the recycling procedure.

70

60

50

40

30

20

10

0

5

10

15

20

25

Time(h)

Figure 8. Time-yield plots determined by GC for the reusability test of the catalyst: (■)

first use; (●) second use; (▲) third use; (★) fourth use; (◆) fifth use. Reaction

conditions: 1a (0.1 mmol), Fe NP@C (1.6 mol%), two consecutive additions of 2.5 and

1.5 equiv of TBHP (5 M solution in decane), DMA (2 ml), Ar (2.5 bar), 110 ^oC.

- 16 -

10.1002/cctc.201700429

ChemCatChem

Furthermore, the Fe NP@C sample resulting after three consecutive reuses was

again imaged by TEM trying to determine variations of Fe particle size. It was,

however, determined that the average Fe particle size does not undergo an obvious

variation upon use of the sample as catalyst. In addition, also the morphology of the C

matrix remained unaltered. Images of one of the used samples are presented in Figure

9, together with the corresponding particle size histogram.

Figure 9. TEM images of three times-used Fe NP@C under bright (a) and dark (b) field.

The inset shows the particle size distribution of this used Fe NP@C sample.

Role of Cu impurities

An additional point of concern using Fe as catalyst is the possible role of impurities

on the observed catalytic activity. It should be noted that the samples in the present

study were prepared using a Fe salt precursor with a nominal purity of 99.99 % and,

therefore, the maximum expected proportion of metal impurities should be below 100

ppm. Considering the precedents that have shown that Cu is the most common Fe

impurity that could contribute to the apparent catalytic activity of Fe in other cross

coupling reactions,^{[27, 46]}the influence of the presence of Cu on the catalytic activity of

- 17 -

10.1002/cctc.201700429

ChemCatChem

Fe NP@C was investigated by preparing two additional Fe NP@C samples purposely

containing 10 and 20 ppm of Cu. The results are presented in Figure 10. It was

observed that the presence of Cu in minute amounts even plays in this case a

detrimental influence, decreasing somewhat the initial reaction rate of Fe NP@C from

1.23 mmol/g·h to 0.98 mmol/g·h and the yield at final time from 66.4% to 64.3%.

Extrapolation at 0 ppm of Cu content of the plot of the initial reaction rate against the

level of Cu impurity strongly supports that Fe has intrinsic catalytic activity for the

oxidative C-N coupling.

70

60

50

40

30

20

10

0

5

10

15

20

25

Time(h)

Figure 10. Time-yield plots determined by GC for the oxidative C-N coupling with Fe

salt precursor containing two different proportions of Cu impurities: (■) Fe salt

precursor with a purity of 99.99 % ; (▲) Fe salt precursor with a purity of 99.99 %

contaminated with 10 ppm of Cu; (●) Fe salt precursor with a purity of 99.99 %

contaminated with 20 ppm of Cu. Reaction conditions: 1a (0.1 mmol), Fe NP@C (1.6

mol%), two consecutive additions of 2.5 and 1.5 equiv of TBHP (5 M solution in

decane), DMA (2 ml), Ar (2.5 bar), 110 ^oC.

- 18 -

10.1002/cctc.201700429

ChemCatChem

Reaction mechanism

The reaction mechanism of the oxidative C-N coupling, even in homogeneous

phase, has not been yet disclosed.^[28]In order to gain some insights on the reaction

mechanism, a series of control experiments were performed in the presence of TEMPO

as a quencher of possible C-centered radical intermediates. In this way, the reaction of

DMA with benzimidazole using Fe NP@C as catalyst in the presence of TEMPO did not

afford the expected product 3a. In contrast, adducts between DMA with TEMPO (A)

and the C-N coupling product with TEMPO (B) were observed (Scheme 3) from the

analysis of the reaction mixture by GC-MS. On the other hand, a similar control

experiment of DMA, TBHP and TEMPO in the absence of benzimidazole resulted

exclusively in the formation of adduct A between DMA and TEMPO. These experiments

strongly support that the reaction proceeds through the formation of the radical from

DMA.

- 19 -

10.1002/cctc.201700429

ChemCatChem

Scheme 3. Possible reaction mechanism, role of Fe NP@C catalyst and quenching of

intermediates by TEMPO.

Furthermore, if the adduct of DMA and TEMPO is first generated by TBHP in the

presence of Fe NP@C and, then, the Fe NP@C is removed, benzimidazole is added and

- 20 -

10.1002/cctc.201700429

ChemCatChem

o

the mixture heated at 140 C, formation of the C-N coupling product 3a was observed

concomitantly with the disappearance of the DMA-TEMPO adduct A. This experiment

is interpreted considering that the DMA-TEMPO adduct is acting as a dormant radical

of DMA, since upon heating it would reversibly release some DMA radical. This DMA

radical would couple with benzimidazole without the need of any catalyst. In this way,

the role of Fe NP@C catalyst would be merely to act as initiator decomposing TBHP,

generating the first radicals that would form C-centered DMA radicals. Since the

homogeneous catalyst reported for this reaction are Fe(OAc)₂and FeCl2 and XPS

reveals that the surface of Fe NPs has a +II and +III oxidation state, it is probably that

these surface species act similarly to the homogeneous counterparts coordinating to

TBHP and, then, splitting the reagent in the initial oxygen-centered radicals (Scheme 3).

This role of Fe generating radicals and the intermediacy of radicals is compatible with

the observation indicated above that filtration of Fe NP@C does not completely stops

the reaction. Since radical reactions occur through radical chain mechanism,

depending on the length of the propagation steps some product formation can be

observed even upon removal of the initiator, although the initiator can generate new

radicals at any reaction time.

Scope of the reaction

Finally the scope of the reaction was screened by using Fe NP@C as catalyst to

promote the oxidative C-N coupling of amides with other N-H substrates. The results

are presented in Table 2. As it can be seen there, high to moderate yields were

obtained in all cases, except in the case of 2-phenylbenzimidazole, where low yields

were attained. It seems that for 2-phenybenzimidazole steric encumbrance due to the

presence of a 2-phenyl substituent at the neighbor position should make C-N coupling

more difficult. In contrast, 2-methylbenzimidazole affords the expected C-N coupling

- 21 -

10.1002/cctc.201700429

ChemCatChem

product in even higher yields than the parent benzimidazole. Similar increase in the

yield of the coupling product is observed comparing the results obtained with pyrazole

and 3,5-dimethylpyrazole. Besides DMA, the scope of the oxidative C-N coupling was

also expanded using two other amides, namely, caprolactam and butyrolactam,

reaching good to moderate yields under the general reaction conditions.

Table 2. Products formed in the oxidative C-N coupling of DMA with N-H substrates

catalyzed by Fe NP@C. Reaction conditions: N compound (0.1 mmol), Fe NP@C (1.6

mol%), two consecutive additions of 2.5 and 1.5 equiv of TBHP (5 M solution in

decane), DMA (2 ml), Ar (2.5 bar), 110 ^oC, 24 h.

Product Yields^a

^athe numbers in brackets correspond to the values determined by GC and when

there is a second value, it corresponds to the isolated yield; b the reaction was

performed with 5 mol % of Fe NP@C.

- 22 -

10.1002/cctc.201700429

ChemCatChem

Conclusions.

In the present work, it has been shown that chitosan containing first row

transition metal ions, Fe²⁺and Co²⁺in the present case, are suitable precursors to form

MNPs, either of a single metal or NP alloys, embedded within graphitic carbon matrix.

The resulting carbonaceous MNP@C composites can be suitable materials to be used

as heterogeneous catalysts. Their catalytic activity has been evaluated in the present

case by the oxidative C-N coupling, finding a large influence of the nature of the metal

catalyst and the solvent in this process. In spite of the apparent decrease in catalytic

activity, Fe NP@C appears as a stable catalyst for the process with a broad scope. The

reaction mechanism involves the generation of amide radicals by decomposition of

TBHP promoted by the Fe NP@C catalyst.

Experimental section

Reagents

Commercially available reagents and solvents were purchased from Aldrich and

used without further purification, except chitosan that was purified by filtration of acid

aqueous solutions.

Synthesis of MNP@C

Briefly, Fe-Co1.42 NP@C, Fe-Co0.86 NP@C were prepared by mixing chitosan (400

mg) in a 20 ml aqueous solution of Fe(OAc)₂(20.5 mg) and Co(OAc)₂·H₂0 (20.0 mg or

28.3 mg, respectively). HOAc (250 μL) was added to dissolve chitosan completely. Fe

NP@C and Co NP@C were prepared following the same procedure, but only using

Fe(OAc)₂(20.5 mg) or Co(OAc)₂·H₂0 (28.3 mg), respectively. After 2 h under magnetic

string at room temperature, the solutions were dried by removing the water heating

- 23 -

10.1002/cctc.201700429

ChemCatChem

the solution at 70 ℃ overnight. The resulting samples were pyrolyzed under Ar flow

(50 ml min^-1), increasing the temperature at a rate of 10 ℃ min^-1up to 900 ℃ with a

holding time of 2 h.

Characterization techniques

Raman spectra were recorded at ambient temperature with 514 nm laser

excitation on a Renishaw In Via Raman spectrometer equipped with a CCD detector.

TEM images were recorded by using a Philips CM 300 FEG system with an operating

voltage of 100 kV. X-ray diffraction (XRD) patterns were obtained by using a Philips X’

Pert diffractometer and copper radiation (CuK_a= 0.1541178 nm). Quantitative ICP-OES

measurements were performed by using a 715-ES Varian apparatus. GC analysis was

performed by using a Bruker 430-GC instrument equipped with a flame-ionization

detector (FID) and a HP-5MS column (30 m×0.25 mm×0.25 mm) whose stationary

phase is constituted by crosslinked 5% phenylmethylsilicone. Product isolation was

carried out by column chromatography of the residue after removing DMA under

reduced pressure using silica gel as stationary phase and CH2Cl2 as eluent.

Catalytic tests

Typically, 1.6 mol% (total metal vs. substrate) of metal catalyst was added to a

reaction mixture of benzimidazole (0.1 mmol) in DMA (2 ml) inside a reinforced glass

reactor. Then, the suspension was magnetically stirred while purging with Ar and

o

heated at 110 C with a preheated silicone bath. Once the reaction temperature was

reached, the system was closed and pressurized with Ar (2.5 bar). Then, TBHP (5M

solution in decane) was introduced to the reaction mixture in two steps, 0.25 mmol at

the starting time, and 0.15 mmol at 4 h reaction time. The temporal evolution of the

reaction was followed by analyzing periodically known aliquots of solution taken with a

- 24 -

10.1002/cctc.201700429

ChemCatChem

syringe provided with a filter (0.2 m) to remove the solid catalyst. The aliquots were

immediately analyzed by GC. Indole was used as internal standard and yields were

based on GC analysis of the reaction mixture containing weight amounts of indole as

external standard. Experiments using Fe(CO)₅, FeCl₂or Fe(OAc)₃as homogeneous

catalysts were carried out in the same conditions as indicated above with 5.3, 3.4, 6.5

g for each of this catalyst, respectively, corresponding to 1.5 g of Fe. The amounts

were taken by weighing the amounts in mg dissolving in 1 ml of DMA and taking 1 L

of the solution with a syringe. Formation of compound 3a in yields below 5 % was

observed after 24 h.

Reuse experiment

Once the reaction was finished, the catalyst was recovered from the reaction

mixture by filtration (Nylon filter, 0.2 m), washed three times with DMA (3 ×10 ml)

and dried in a oven (70 ℃) overnight. Then, the catalyst was used again in the next C-N

coupling reaction under the same conditions.

Accessibility tests.

50 mg of the as-synthesized samples (Fe NP@C, Fe-Co0.86 NP@C or Co NP@C,

respectively) were treated in 15 ml 0.5 M HCl at room temperature under stirring for

14 h. Then, the samples were filtrated and the amounts of leached metals were

determined by ICP. Comparing with the total metal content, 87.3% of total iron

content in Fe NP@C, 85.5% of total iron content and 87.5% of total cobalt content in

Fe-Co NP0.86@C, 88.4% of total cobalt content in Co NP@C was dissolved under these

conditions. As it is well known, the acid leaching under relatively diluted HCl at

ambient temperature can only remove the accessible iron and cobalt nanoparticles

embedded in the graphite shells. From these results, It seems that the Fe and Co

nanoparticles in the as-synthesized catalysts have similar metal accessibility.

- 25 -

10.1002/cctc.201700429

ChemCatChem

Acknowledgements

Financial support by the Spanish Ministry of Economy and Competitiveness

(Severo Ochoa and CTQ2015-69153-CO2-1) and Generalitat Valenciana (Prometeo

2013-014) is gratefully acknowledged. J.H. thanks the Chinese Ministry of Higher

Education for a doctoral fellowship at Valencia. A.D.M. thanks University Grants

Commission, New Delhi, for the award of Assistant Professorship under its Faculty

Recharge Program. A.D.M. also thanks the Department of Science and Technology,

India, for the financial support through Fast Track project (SB/FT/CS-166/2013).

References

[1]

[2]

[3]

[4]

D. Astruc, F. Lu, J. R. Aranzaes, Angew. Chem. Int. Ed. 2005, 44, 7852-72.

M. Haruta, Cattech 2002, 6, 102-15.

R. Narayanan, M. A. El-Sayed, J. Phys. Chem. B 2005, 109, 12663-76.

Y. Mikami, A. Dhakshinamoorthy, M. Alvaro, H. Garcia, Catal. Sci. Technol. 2013, 3, 58-

69.

[5]

[6]

[7]

[8]

[9]

[10]

M. Haruta, Catal. Today 1997, 36, 153-66.

M. Haruta, Gold Bull. 2004, 37, 27-36.

R. Chinchilla, C. Najera, Chem. Soc. Rev. 2011, 40, 5084-121.

A. Dhakshinamoorthy, S. Navalon, M. Alvaro, H. Garcia, ChemSusChem 2012, 5, 46-64.

V. Farina, Adv. Synth. Catal. 2004, 346, 1553-82.

R. J. White, R. Luque, V. L. Budarin, J. H. Clark, D. J. Macquarrie, Chem. Soc. Rev. 2009,

38, 481-94.

[11]

[12]

M. Comotti, W. C. Li, B. Spliethoff, F. Schuth, J. Am. Chem. Soc. 2006, 128, 917-24.

N. Lopez, T. V. W. Janssens, B. S. Clausen, Y. Xu, M. Mavrikakis, T. Bligaard, J. K.

Norskov, J. Catal. 2004, 223, 232-5.

[13]

[14]

[15]

M. Okumura, S. Nakamura, S. Tsubota, T. Nakamura, M. Azuma, M. Haruta, Catal. Lett.

1998, 51, 53-8.

S. Pisiewicz, D. Formenti, A.-E. Surkus, M.-M. Pohl, J. Radnik, K. Junge, C. Topf, S.

Bachmann, M. Scalone, M. Beller, ChemCatChem 2016, 8, 129-34.

F. A. Westerhaus, R. V. Jagadeesh, G. Wienhoefer, M.-M. Pohl, J. Radnik, A.-E. Surkus,

J. Rabeah, K. Junge, H. Junge, M. Nielsen, A. Brueckner, M. Beller, Nature Chem. 2013,

5, 537-43.

- 26 -

10.1002/cctc.201700429

ChemCatChem

[16]

A. Banerjee, R. Gokhale, S. Bhatnagar, J. Jog, M. Bhardwaj, B. Lefez, B. Hannoyer, S.

Ogale, J. Mater. Chem. 2013, 22, 19694-9.

[17]

[18]

J. A. Hu, H. L. Wang, Q. M. Gao, H. L. Guo, Carbon 2010, 48, 3599-606.

T. A. Wezendonk, Q. S. E. Warringa, V. P. Santos, A. Chojecki, M. Ruitenbeek, G.

Meima, M. Makkee, F. Kapteijn, J. Gascon, Farad. Discuss 2016, 10.1039/C6FD00198J.

T. A. Wezendonk, V. P. Santos, M. A. Nasalevich, Q. S. E. Warringa, A. I. Dugulan, A.

Chojecki, A. Koeken, M. Ruitenbeek, G. Meima, H. U. Islam, G. Sankar, M. Makkee, F.

Kapteijn, J. Gascon, ACS Catal. 2016, 6, 3236-47.

[19]

[20]

[21]

B. An, K. Cheng, C. Wang, Y. Wang, W. Lin, ACS Catal. 2016, 6, 3610-8

V. P. Santos, T. A. Wezendonk, J. J. Delgado Jaén, A. I. Dugulan, M. A. Nasalevich, H. U.

Islam, A. Chojecki, S. Sartipi, X. Sun, A. A. Hakeem, A. C. J. Koeken, M. Ruitenbeek, T.

Davidian, G. R. Meima, G. Sankar, F. Kapteijn, M. Makkee, J. Gascon, Nature Commun.

2015, 6, 6451.

[22]

[23]

K. P. De Jong, J. W. Geus, Catal. Rev. Sci. Eng. 2000, 42, 481-510.

S. H. Joo, S. J. Choi, I. Oh, J. Kwak, Z. Liu, O. Terasaki, R. Ryoo, Nature 2001, 412, 169-

72.

[24]

[25]

[26]

[27]

[28]

[29]

[30]

P. Serp, M. Corrias, P. Kalck, Appl. Catal. A: Gen. 2003, 253, 337-58.

Z. Chen, D. Higgins, A. Yu, L. Zhang, J. Zhang, Energy Environ. Sci. 2011, 4, 3167-92.

B. D. Sherry, A. Fuerstner, Acc. Chem. Res. 2008, 41, 1500-11.

I. Thome, A. Nijs, C. Bolm, Chem. Soc. Rev. 2012, 41, 979-87.

G. Saidulu, R. A. Kumar, K. R. Reddy, Tetrahedron Lett. 2015, 56, 4200-3.

Q. Xia, W. Chen, J. Org. Chem. 2012, 77 9366-73.

T. Truong, K. D. Nguyen, S. H. Doan, N. T. S. Phan, Appl. Catal. A: Gen. 2016, 510, 27-

33.

[31]

[32]

[33]

[34]

[35]

F. Chen, C. Topf, J. Radnik, C. Kreyenschulte, H. Lund, M. Schneider, A. E. Surkus, L. He,

K. Junge, M. Beller, Journal of the American Chemical Society 2016, 138, 8781-8.

X. J. Cui, Y. H. Li, S. Bachmann, M. Scalone, A. E. Surkus, K. Junge, C. Topf, M. Beller,

Journal of the American Chemical Society 2016, 138, 457-.

L. He, F. Weniger, H. Neumann, M. Beller, Angewandte Chemie-International Edition

2016, 55, 12582-94.

I. Ziccarelli, H. Neumann, C. Kreyenschulte, B. Gabriele, M. Beller, Chemical

Communications 2016, 52, 12729-32.

A. Primo, P. Atienzar, E. Sanchez, J. Maria Delgado, H. Garcia, Chem. Commun. 2012,

48, 9254-6.

[36]

[37]

A. Primo, E. Sanchez, J. M. Delgado, H. Garcia, Carbon 2014, 68, 777-83.

G. Abellan, M. Latorre-Sanchez, V. Fornes, A. Ribera, H. Garcia, Chem. Commun. 2012,

48, 11416-8.

[38]

M. Latorre-Sanchez, P. Atienzar, G. Abellan, M. Puche, V. Fornes, A. Ribera, H. Garcia,

Carbon 2012, 50, 518-25.

- 27 -

10.1002/cctc.201700429

ChemCatChem

[39]

E. Park, J. Q. Zhang, S. Thomson, O. Ostrovski, R. Howe, Metallurgical and Materials

Transactions B-Process Metallurgy and Materials Processing Science 2001, 32, 839-45.

S. Peng, C. Wang, J. Xie, S. Sun, J. Am. Chem. Soc. 2006, 128, 10676-7.

A. Primo, I. Esteve-Adell, J. F. Blandez, A. Dhakshinamoorthy, M. Alvaro, N. Candu, S.

M. Coman, V. I. Parvulescu, H. Garcia, Nature Commun. 2015, 6.

A. Primo, I. Esteve-Adell, S. N. Coman, N. Candu, V. I. Parvulescu, H. Garcia, Angew.

Chem. Int. Ed. 2014, 55, 607-12.

[40]

[41]

[42]

[43]

D. Mateo, I. Esteve-Adell, J. Albero, J. F. Sanchez Royo, A. Primo, H. Garcia, Nature

Commun. 2016, 7.

[44]

[45]

[46]

M. Buaki-Sogo, M. Serra, A. Primo, M. Alvaro, H. Garcia, ChemCatChem 2013, 5, 513-8.

C. Lavorato, A. Primo, R. Molinari, H. Garcia, ACS Catal. 2014, 4, 497-504.

S. L. Buchwald, C. Bolm, Angew. Chem. Int. Ed. 2009, 48, 5586-7.

Graphical Abstract

- 28 -

Article Doi

Iron Nanoparticles Embedded in Graphitic Carbon Matrix as Heterogeneous Catalysts for the Oxidative C?N Coupling of Aromatic N?H Compounds and Amides

DOI: 10.1002/cctc.201700429

Source and publish data:

Authors:

Article abstract of DOI:10.1002/cctc.201700429

Full text of DOI:10.1002/cctc.201700429

Products guided by the article

R&D Labs maybe for 1398412-30-0

Relevant to this article

Hot Product