Introduction

The Ullmann reaction, after being introduced in 1901, has been used widely for C–C bonding between two aromatic nuclei using copper as a catalyst (Handa et al. 2015; Altenhoff et al. 2003). Ni and Pd were used as catalysts in later modifications of this reaction for aryl bromides or chlorides (Snelders et al. 2009; Pratap et al. 2009; Martin and Buchwald 2008). Different catalysts for the Ulmann reaction have been reported using palladium complexes (Hennings et al. 1999) which usually use several reducing agents such as zinc (Qafisheh et al. 2002), sodium formate (Arcadi et al. 1990) and trimethylamine (Han et al. 2000). Despite good efficiency of these catalysts in the Ulmann reaction, purification of products is reported to be challenging (Jiang and Cai 2007) due to the homogeneous reaction conditions. To solve this problem, heterogeneous catalysts have been developed to simplify their recovery and reusability (Li et al. 2017). Some polymer-supported transition metal catalysts are silica-supported metallocene/MAO (Grasa et al. 2002), silica‐supported zirconocene (Bourissou et al. 2000), silica-supported metals (Baran et al. 2017), layered double hydroxide supported nano palladium and clay-supported catalyst (Tehrani and Basiryan 2015). The main heterogeneous catalysts are expensive, not biodegradable, and unstable thermodynamically, and tend to aggregate in bulk metal form.

Cellulose nanocrystals (CNC) are an abundant, accessible, highly crystalline biopolymer that can be extracted from cellulosic fibers like wood, cotton, non-wood plants, agricultural residues, bacteria and algae (Tehrani and Basiryan 2015; Fraschini et al. 2014; Brito et al. 2012; Lopez et al. 2010; Favier et al. 1995; Le Normand et al. 2014; Silvério et al. 2013). Production, modification and application of cellulose nanocrystals have been considered by researchers and companies in the last decade.

Cellulose nanocrystals have been reported as a substrate for metal containing nano catalysts due to their high surface area, active surface functional groups, water suspension ability, high crystallinity, chirality, high mechanical strength, renewability, biodegradability, and non-toxicity (Yan et al. 2012; Roman 2015; Yanamala et al. 2014; Kümmerer et al. 2011; Kaushik and Moores 2016). Cellulose nanocrystals indicate un-conventional colloidal behavior in liquid form which can be useful for different applications (Kovacs et al. 2010; Moon et al. 2011; Habibi et al. 2010; Lin and Dufresne 2014). Using nano-cellulose as a heterogeneous support has been reported for different catalysts like palladium, platinum, copper, nickel and silver nano-particles (Keshipour and Adak 2016; Keshipour and Khalteh 2016; Kaushik and Moores 2016; Alesi et al. 2008; Cirtiu et al. 2011; Huang et al. 2014; Reddy et al. 2006). Cellulose fibers were reported as a substrate for nano palladium (0) for the Suzuki reaction (Fu et al. 2015; Jamwal et al. 2011; Hu et al. 2016), Heck reaction (Keshipour et al. 2013; Li et al. 2017; Cirtiu et al. 2011), the Ullmann cross-coupling reaction (Zhou et al. 2012), and Au-coupling reaction (Huang et al. 2013).

We recently developed a nano catalyst for the formation of C–O bonds between activated aryl halides and phenol derivatives (Khalilzadeh et al. 2011, 2014; Keipour et al. 2016; Salmanpour et al. 2013). In our continued efforts towards the development of heterogeneous catalysts for selective organic transformations, our hypothesis is that cellulose nanocrystals not only can be used as a substrate for our catalyst but that it can enhance catalytic reactions due to its special colloidal behavior. Therefore, in this research we developed nano-cellulose 2-(1H-Benzo[d]imidazol-2-yl) aniline (CNC-BIA-Pd) as a strong and novel nano-catalyst for organic chemistry. As a part of this research, activated and inactivated phenols and aryl halides were reacted using the developed catalyst in DMSO at 80 °C to synthesize diaryl ether with short reaction times (15–60 min). The design and procedure of the catalyst preparation is illustrated in Scheme 1.

Scheme 1
scheme 1

Procedure for the synthesis of CNC-BIA-Pd

Full size image

Experimental

Chemicals and instruments

All chemicals were purchased from Merck and Fluka and used as received. A Vector 22-Bruker FT-IR was used for Fourier transform infrared (FT-IR) in the range of 400–4000 cm−1 at room temperature. The X-ray diffraction (XRD) of the Nano-cellulose and its modified catalyst prepared as powder and were measured with a Philips PW 1830 X-ray diffractometer with Cu Kα source (λ = 1.5418 Å). The data sets were collected in reflection geometry in the range of 10° ≤ 2θ ≤ 80° at room temperature. Thermo gravimetric analysis (TGA) was carried out on a Stanton Red Craft STA-780 (London, UK) using N2 and O2 as carrier gas with a temperature ramp of 10 °C/min from room temperature to 650 °C. A CHNS/O analyzer (Vario Micro cube, Elemental Analysis system (GmbH, Hanau, Germany) was employed for the determination of elemental analysis with helium as the carrier gas. The morphology of the prepared nano catalysts was investigated using scanning electron microscopy (SEM; EM-3200, KYKY) followed by determination of the elemental composition by energy dispersive X‐ray spectrometry (EDX). Morphology of nano catalysts was carried out with transmission electron microscopy (TEM; EM-10C, ZEISS). The deposited amount of palladium nanoparticles was determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) (Perkin Elmer Optima 2000 DV ICP-AES). X-ray photoelectron spectroscopy (XPS) measurements were recorded on an electron spectrometer XPS/UVS—SPECS System with PHOIBOS 150 analyzer equipped with Al/Mg Kα radiation. NMR spectra of products were recorded with a Bruker DRX-400 AVANCE instrument (400.1 MHz for 1H, 100.6 MHz for 13C) in DMSO-d6 as solvent.

Preparations of cellulose nanocrystals (CNC) (1)

Cellulose nanocrystals were prepared by acidic hydrolysis of Whatman filter paper (#1) as reported in the literature (Sadeghifar et al. 2011). An amount of 2 g of cellulose fiber was heated in 100 mL of 2.5 M HBr for 3 h at 100 °C under ultra-sonication. The hydrolyzed fibers were diluted with deionized water followed by centrifugation. The washing/centrifugation cycles were repeated five times to remove excess acid and water-soluble fragments. The fine cellulose nanoparticles started to disperse in the aqueous supernatant after reaching a pH around 5 and then were collected using centrifugation at 12,000 RPM for 60 min. The product was kept in a refrigerator without drying.

Tosylation of cellulose nanocrystals yielding CNC-Tos (2)

CNC-Tos was synthesized using a reported method (Sadeghifar et al. 2011; Feese et al. 2011). An amount of 0.5 g of non-dried cellulose nanocrystals was washed with pyridine five times to replace water with pyridine. The mixture in pyridine was stirred for 2 days at room temperature after the addition of tosyl chloride (0.9 g, 5 mmol). An amount of 100 mL of ethanol was added to the reaction mixture and precipitated material was collected by filtration. The product was washed with ethanol (50 mL) five times and kept in a refrigerator without drying (Scheme 1, CNC-Tos).

Preparation of CNC-BIA (4)

An amount of 400 mg of CNC-Tos in methanol was washed and centrifuged three times with DMF to exchange methanol with DMF. An amount of 400 mg of 2-(1H-Benzo[d]imidazol-2-yl) aniline (BIA) was added to the mixture and stirred for 24 h at 100 °C. An amount of 50 mL of deionized water was added to the reaction mixture to recover precipitated product using centrifugation. The product was washed with ethanol and water several times to purify the final product. Purified product was then washed with DMF three times to replace the water and the washed product was kept in a refrigerator (Scheme 1, CNC-BIA).

Preparation of CNC-BIA-Pd catalyst

An amount of 0.50 g of CNC-BIA in DMF was added to a solution of PdCl2 (0.10 g, 0.45 mmol) in 10 mL of DMF under N2 atmosphere and the mixture was stirred for 24 h at 60 °C. After completion of the reaction, the mixture was cooled to room temperature and the resulting product was collected by filtration. The obtained solid black colored product was washed carefully with ethanol (2 × 25 mL), then with diethyl ether (2 × 25 mL) and then three times with distilled water (3 × 25 mL), and finally dried in a vacuum oven at room temperature (Scheme 1, CNC-BIA-Pd).

General procedure of Ullman reaction of aryl halides with phenols by nano catalyst

In a mixture of aryl halide (1 mmol) and phenol derivatives (1.2 mmol) in DMSO (2.5 mL) specific amounts of base and 3.8 mg of CNC-BIA-Pd (0.2 mol%) as nano catalyst were added. The mixture was stirred at 80 °C for different times. Finally, the solid catalyst was filtered and washed carefully with distilled water and absolute ether. To recover the Ullman reaction product, the solution was vaporized and the residue solid was purified by plate chromatography on silica gel and characterized with FT-IR, 13C NMR and 1H NMR techniques (Scheme 2).

Scheme 2
scheme 2

Ullman reaction of aryl halides and phenols in the presence of CNC-BIA-Pd

Full size image

Results and discussion

Evaluation of catalyst preparation

The focus of the present report is the preparation of cellulose nanocrystals and its surface modification with a catalyst to be used in the Ullmann coupling reaction. Evidence of cellulose nanocrystals preparation was indicated by TEM images (Fig. 1).

Fig. 1
figure 1

TEM image of prepared cellulose nanocrystals

Full size image

Due to the use of hydrobromic acid for nano-cellulose preparation, there are no sulfonated groups on the cellulose surface which are usually created when using sulfuric acid. The absence of sulfonate groups on the surface of the cellulose nanocrystals increases the thermal and chemical stability of the material.

The new catalyst was prepared using surface modification of cellulose nanocrystals through chemical bonding with 2-(1H-Benzo[d]imidazol-2-yl) aniline (BIA). Chemical bonding of catalyst on the cellulose nanocrystal surface should provide longer lasting catalyst with less leaching. It also should prevent catalyst aggregation. The covalent attachment of the ligand on the cellulose nanocrystals surface was achieved via surface tosylation of CNC (1) with tosyl chloride in DMF to prepare CNC-Tos (2) followed by reaction with 2-(1H-Benzo[d]imidazol-2-yl) aniline (BIA) yielding CNC-BIA (3). In the final step, PdCl2 solution was reacted with CNC-BIA in DMF solution to prepare the final CNC-BIA-Pd (4) catalyst.

Figure 2 indicates FT-IR spectra of CNC, CNC-Tos and CNC-BIA-Pd. The absorption peaks at 1060 cm−1 and 1373 cm−1 reveal vibration of C–O–C in pyranose ring of glucose in cellulose nanocrystals structure and C–H vibrations, respectively. The absorption bands at 3446 cm−1 and 2998 cm−1 were assigned to the O–H and C–H stretching vibrations, respectively (Baran et al. 2017). Sulfonyl symmetrical and unsymmetrical stretching vibration in tosylated cellulose nanocrystals (CNC-Tos) are visible at 1163 cm−1 and 1350 cm−1 respectively. Absorption bands at 1440 cm−1 and 1543 cm−1 confirm the aromatic structure of the tosyl group (Tehrani and Basiryan 2015) in the sample. After the amination process by 2-(1H-Benzo[d]imidazol-2-yl) aniline, hydroxyl groups on the CNC surface were converted to amino groups. Therefore, the absorption bands of sulfonyl groups at 1163 cm−1 and 1350 cm−1 were not apparent in the FT-IR spectrum of the CNC-BIA-Pd (Yan et al. 2012). New absorption peaks at 1653 and 1260 cm−1 are related to the N (sp2) and N (sp3)–C bonds of the stretching vibration and the vibration in CNC-BIA-Pd. In addition, the intensity of the absorption band at 650 cm−1 that is assigned to the bending vibration of C–H in the pyridine heterocyclic ring is reduced after formation of the CNC-BIA complex with Pd (Hu et al. 2016). The FT-IR spectra confirmed the grafting of 2-(1H-Benzo[d]imidazol-2-yl) aniline (BIA) as ligand connected to Pd metal on cellulose nanocrystals.

Fig. 2
figure 2

FTIR spectra of CNC, CNC-Tos and CNC-BIA-Pd

Full size image

Figure 3 indicates a comparison of crystal structures of pure cellulose nanocrystals and its surface modified form with catalyst (CNC-BIA-Pd). The pattern for prepared CNC is very much for a mixture of cellulose I and II, with strong peaks at 12° and 20° [the (020) cellulose II peak at 22° is merged with the (200) cellulose I peak at 22.5°] (French 2014). Relative crystallinity was calculated from the intensity measurements using the Segal method (Segal et al. 1959) (Eq. 1).

$${\text{X}}_{\text{C}} \% \, = \, \left( {\left( {{\text{I}}_{200} - {\text{I}}_{\text{AM}} } \right)/{\text{I}}_{200} } \right) \, 100\%$$
(1)

where I200 represents the maximum intensity of (200) lattice diffraction peak at a diffraction angle around 2θ = 22.5°, IAM represents the intensity scattered by the amorphous component in the sample, evaluated as the lowest intensity at 18°. The crystallinity of the unmodified cellulose nano particles was around 76% which is in agreement with a previously reported result (Sadeghifar et al. 2011). However, the calculated crystallinity number is only an estimate due to the strong cellulose II (110) peak at around 20°. The CNC maintained its morphology and crystal structure after modification, which is important to maintain the surface area, thermal and physical properties, and colloidal properties of the material. The crystalline structure of palladium is also clearly visible in the modified cellulose nanocrystals as evidenced by the presence of catalyst on the surface. The index peaks at 2θ = 40.0°, 46° and 68° are ascribed to diffractions from various lattice planes of (111), (200) and (220) present in the cubic Palladium (Hu et al. 2016).

Fig. 3
figure 3

XRD spectra of CNC and CNC-BIA-Pd

Full size image

To have an estimation of catalyst loading on the surface of cellulose nanocrystals, nitrogen content determination, EDX and thermal analysis were carried out. Nitrogen content in the product was 2.1% or 0.021 g of nitrogen per gram of final product. Considering the presence of three nitrogen atoms in the prepared catalyst structure, the amount of catalyst should be around 4.4 × 10−4 mol ((0.021/3)/15)) per each gram of the product.

Thermo-gravimetric analysis (TGA) on the cellulose nanocrystals and the final product in nitrogen and oxygen atmospheres are indicated in Fig. 4a, b, respectively. Under both gasses, modified cellulose nanocrystals indicated stability up to 220 °C, which makes the material possible to be used for high temperature reactions. After the degradation of samples at temperatures up to 550 °C in nitrogen atmosphere, the modified cellulose nanocrystals indicated more char than the pure cellulose nanocrystals. The remaining char in cellulose nanocrystals was 24% whereas in the modified sample it was around 31% (Fig. 4a). The modified samples contain catalyst with high levels of carbon and Pd, which leads to higher char mass after thermal degradation.

Fig. 4
figure 4

Thermogravimetric analysis of CNC and CNC-BIA-Pd in nitrogen (a) and oxygen (b)

Full size image

Figure 4b illustrates the sample behavior after heating using oxygen gas, which combusts all organic materials and leaves only minerals such as Pd for this study. The remaining sample weight after burning in oxygen for pure cellulose nanocrystals and the modified sample were zero and 4.7% respectively. An amount of 4.7% Pd in the product indicates the presence of 0.047 g Pd/g product. Due to presence of one atom of Pd in the catalyst, the mole of Pd in each gram of the prepared catalyst was calculated to be around 4.4 × 10−4 (0.047/106.4) which is very close to the number calculated with nitrogen content.

SEM images (Fig. 5) indicate precipitation of nano catalyst on the cellulose nanocrystals surface. After functionalization of CNC with 2-(1H-Benzo[d]imidazol-2-yl) aniline (BIA) connected to Pd, particles with nano size were observed on the CNC-BIA-Pd (final product).

Fig. 5
figure 5

SEM images of CNC (left) and CNC-BIA-Pd (right)

Full size image

To obtain additional insight into the shapes and particle sizes of the prepared nano catalyst, transmission electron microscopy (TEM) was used. As clearly indicated in Fig. 6, dark areas in the images revealed Pd grafted to 2-(1H-Benzo[d]imidazol-2-yl) aniline (BIA) on the surface of the cellulose nanocrystals.

Fig. 6
figure 6

TEM images of CNC-BIA-Pd

Full size image

EDX was used to determine the elemental compositions of the CNC-BIA-Pd (Fig. 7). The obtained results indicate the presence of C, O and Pd in the prepared catalyst. It can be concluded that Pd nano particles were embedded into 2-(1H-Benzo[d]imidazol-2-yl) aniline (BIA) inserted on the surface of cellulose nanocrystals. These observations confirm that chemical modification of the CNC was achieved.

Fig. 7
figure 7

EDX spectrum of CNC-BIA-Pd

Full size image

The ICP-AES analysis was performed to determine the amount of Pd (56,140 mg/Kg) in CNC-BIA-Pd as catalyst. The inductively coupled plasma–atomic emission spectroscopy (ICP-AES) analysis showed the weight percentage of the Pd to be 5.2%, in agreement with results obtained from thermal analysis and CHN.

The XPS spectrum of the Pd nanoparticles, dispersed on CNC-BIA-Pd for the Pd 3d region is presented in Fig. 8. The results show that for the binding energies of Pd 3d5/2, two peaks are observed at about 334.5 and 337.8 eV, and for the Pd 3d3/2 there are also two peaks at about 341.4 and 343.1 eV, respectively. This indicates that the Pd are in both forms, metallic state Pd(0) and Pd(II), simultaneously on the CNC-BIA-Pd catalyst (Gniewek et al. 2005; Narayana et al. 1985).

Fig. 8
figure 8

XPS spectrum of Pd 3d region of CNC-BIA-Pd

Full size image

A possible mechanism of reaction between cellulose and catalyst is proposed in Scheme 3 (Wu et al. 2013). It is believed that the electron-rich feature of the hydroxyl groups in the cellulose structure reduce PdCl2 to Pd NPs. In the next step the functional groups on the CNC surface containing oxygen by electrostatic interaction act as anchor points to immobilize Pd NPs. The BIA in this catalyst acts as homogeneous distributer of Pd(II) on the CNC surface after reduction.

Scheme 3
scheme 3

The proposed reduction mechanism of Pd NPs using CNC as reducing agent

Full size image

Ullmann coupling reaction using CNC-BIA-Pd catalyst

To evaluate the performance of the developed catalyst in the Ullmann reaction, the effects of catalyst loading, solvents, temperature, base and reaction times were investigated on a Carbon–Oxygen reaction. A reaction optimization was carried out using phenol and 4-nitro iodo benzene as reactants and CNC-BIA-Pd as catalysts. Table 1 indicates the effects of catalyst loading, temperature and solvents on the reaction efficiency. The best reaction efficiency (96%) occurred when using 0.2 mol% of catalyst, using 3.75 mmol of K2CO3 as the base at 80 °C for 45 min reaction time in the presence of DMSO as a solvent (Table 1).

Table 1 Effect of base and solvent on the Ullmann coupling reaction
Full size table

Generally, the Ullmann reaction indicates weak performance when using active phenols with a withdrawing group (EWG) and inactive aryl halides (with electron donor group). However, CNC-BIA-Pd catalyst indicated better results compared with the classic and improved Ullmann catalysts when using inactive aryl halides and inactive phenols (Banwell et al. 2011).

Moreover, the effect of the substitute groups of the substrate on the yields of the Ullmann coupling reactions using the developed catalyst was examined, and the yields were in the order of para > ortho > meta. Additionally, the order of the catalytic performance of the substrates (containing aryl bromide, aryl chloride and aryl iodide) was determined as follows: I > Br > Cl. All results are in agreement with the previous reported results for Ullmann reaction (Baran et al. 2017).

Turn over number (TON) and turn over frequency (TOF) values were calculated for all Carbon–Oxygen Ullmann reactions, and the results are listed in Table 2. The results indicated that CNC-BIA-Pd catalyst gave remarkable TON and TOF values with small loading of the catalyst in a short time. These values indicated that the catalyst can be used efficiently for different Ullmann coupling reactions.

Table 2 Effect of the Pd catalyst on C–O Ullmann coupling reaction for aryl halides (I, F, Cl, Br)
Full size table

In addition, the catalytic efficiency of CNC-BIA-Pd catalyst was evaluated against different commercial palladium salts with the model reaction under optimum conditions (Table 2). These tests indicated that the CNC-BIA-Pd catalyst had higher catalytic activity as well as TON and TOF values than the commercial palladium catalysts. All of the synthesized compounds were further identified using the GC/MS and 1H NMR techniques and their spectra are presented in supplementary data.

Another important issue concerning the application of a heterogeneous catalyst is its reusability and stability under reaction conditions. To gain insight into this issue, catalyst recycling experiments were carried out using the Ullmann reaction of 4-nitro iodobenzene and phenol over CNC-BIA-Pd. The results are tabulated in Table 3.

Table 3 Reusability of CNC-BIA-Pd in Ullmann C–O reaction
Full size table

After each cycle, the catalyst was filtered off, washed with water, diethyl ether and acetone. Then it was dried in an oven at 60 °C and reused in the Ullmann reaction. The results indicated that CNC-BIA-Pd could be reused without losing its effectiveness eight times. It should be mentioned that there was low Pd leaching (about 5%) during the reaction and the catalyst exhibited high stability even after eight cycles (Table 3).

Conclusion

A new Pd catalyst bonded to cellulose nanocrystals was developed (CNC-BIA-Pd) and characterized by different spectroscopic methods. The amount of bonded catalyst was calculated as 0.031 mol per each mole of glucose unit in the cellulose nanocrystals. The modified cellulose nanocrystal based catalyst showed thermal stability up to 220 °C which makes it possible to be used for high temperature reactions. The catalytic performance of the catalyst was tested for the Ullmann reaction for aryl fluoride, chloride, bromide and iodides at 80 °C. The highest conversion yield (96%) was obtained at 80 °C with DMSO as solvent using 3.75 mmol K2CO3 as base. The best conversion yield was obtained with 0.2 mol% catalyst for each mole of substrate. The effect of the substitute groups on the substrate on the Ullmann coupling reactions yield was determined and displayed efficiency in the order of para > ortho > meta. The order of the catalytic performance on the substrates (containing aryl bromide, aryl chloride and aryl iodide) was shown to be in the order of I > Br > Cl. The developed catalyst indicated high TON (490) and TOF (1960) values. The reusability of the catalyst activity was tested and the catalytic performance remained high after eight cycles of use. In conclusion, the developed CNC-BIA-Pd catalyst indicated high thermal stability, reusability, and high product yields in Ullmann reactions using small catalyst loadings.