Solution combustion synthesis of nano-catalysts with

Journal of Catalysis 364 (2018) 112–124

Contents lists available at ScienceDirect

Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat

Solution combustion synthesis of nano-catalysts with a hierarchical structure G. Xanthopoulou a,⇑, O. Thoda a,b, S. Roslyakov c, A. Steinman c, D. Kovalev d, E. Levashov c, G. Vekinis a, A. Sytschev d, A. Chroneos b a

Institute of Nanoscience and Nanotechnology, NCSR ‘‘Demokritos”, Agia Paraskevi Attikis 15310, Greece Faculty of Engineering, Environment and Computing, Coventry University, Priory Street, CV1 5FB Coventry, UK National University of Science and Technology ‘‘MISIS”, Moscow 119049, Russia d Institute of Structural Macrokinetics and Materials Science, Russian Academy of Sciences Chernogolovka, Moscow Region, Russia b c

a r t i c l e

i n f o

Article history: Received 19 January 2018 Revised 20 March 2018 Accepted 4 April 2018

Keywords: Solution combustion synthesis Nano catalysts Liquid-phase hydrogenation Nickel-based catalysts

a b s t r a c t The structure, composition, surface area and catalytic activity of Solution Combustion Synthesis (SCS) catalysts are all influenced by the conditions of preparation and in particular, the glycine concentration in the initial SCS solution. NMR was used to monitor the formation of glycine-nickel nitrate complexes in solution before SCS initiates, IR high speed temperature measurements have allowed to elucidate the mechanism of synthesis during SCS, dynamic X-ray analysis and thermogravimetric analysis have clarified the mechanisms of phase formation during SCS, BET analyses have shown the regularity of pore formation and SEM and TEM studies have indicated the regularities involved during microstructure formation. Regular three-dimensional (3D) flowerlike Ni-NiO hierarchical architectures were synthesized by SCS. The results have revealed a three-dimensional percolation network with hierarchical structure on the basis of nano-structured metal oxides and metals synthesized during SCS. Such hierarchically nanoporous catalysts have versatile structural properties such as increased surface area and large overall pore volume that can alleviate diffusional limitations of conventional nanocatalysts with solely microporous framework. This is important for liquid phase heterogeneous catalysis. These new insights provide a valuable capability for optimizing the selectivity and activity of SCS catalysis and will no doubt be of significant interest to a wide range of researchers working in catalysis and other fields. Ó 2018 Elsevier Inc. All rights reserved.

1. Introduction A combination of SHS and reactive solution approaches leads to the so-called solution (or aqueous) combustion synthesis (SCS) method [1]. This process typically involves a reaction in solution of metal nitrates and different fuels. SCS entails significant advantages over other combustion-based methods and the most critical are mentioned below: (1) The initial reactants mixing takes place in the liquid state, facilitating control over homogeneity and stoichiometry of the reaction products. (2) The possibility of impurity ions incorporation in the oxide hosts is a viable option, in order to prepare materials of industrial interest (pigments, phosphors, catalysts, etc.).

⇑ Corresponding author. E-mail address: [email protected] (G. Xanthopoulou). https://doi.org/10.1016/j.jcat.2018.04.003 0021-9517/Ó 2018 Elsevier Inc. All rights reserved.

(3) The process is fast and requires no special equipment, which makes it preferable over SHS methods. (4) The method is very rapid allowing the formation of metastable phases [2]. Solution Combustion Synthesis (SCS) is a simple, but important technique for the synthesis and processing of nano-structured metals, oxides, spinels, alloys, intermetallics etc. SCS is based on exothermic redox reactions between nitrates (oxidizers) and organic substances (reducers) to directly produce nanostructured powders [3,4]. More specifically, after preheating to moderate temperatures (approximately 150–200 °C), the reaction can be initiated and the combustion front propagates a rapid (typically 0.001–0.1 m s1) high-temperature (1000–3000 °C) wave in a self-sustained manner through the system, leaving behind fine solid product of tailored composition [1]. The mechanism ruling the SCS process and the key role played by the reducing agent is crucial, as well. It is interesting to mention that various fuels have been employed for specific classes of oxidic

113

G. Xanthopoulou et al. / Journal of Catalysis 364 (2018) 112–124

materials. Various water soluble organic substances are utilized as source of carbon and hydrogen that, on combustion, form carbon dioxide and water generating heat during the reaction [5,6]. Furthermore, certain fuels, such as glycine, act as complex formation agents with the metal cations, which enable high mixing homogeneity in solution and hence avoid segregations [7,8], Such compounds that contain NAN bonds are well-known to better assist the combustion reaction; however, there is the drawback of releasing NOx emissions during combustion. In some cases, the use of highly viscous fuels, such as glycerol, facilitates better distribution of the final desired product on the structured catalyst support [9]. The reaction between a metal nitrate as an oxidizer and glycine (C2H5NO2) as a fuel is typically presented as follows:

5 5 v u CH2 NH2 COOH þ v ðu 1ÞO2 9 4 10 25 5u þ 9 N2 ðgÞ CO2 ðgÞ þ ! M v OV ðsÞ þ v u u H2 OðgÞ þ v 2 9 18 18

M v ðNO3 Þv þ

where m is the metal valence and a tunable parameter, u is the fuelto-oxidizer ratio. In case of u = 1, all oxygen required for complete combustion of fuel is derived from the oxidizer, while u > 1 signifies fuel-rich (or lean) conditions [10]. According to the equation above, the solid products of SCS should be oxides, and thousands of different oxide-based nanopowders are indeed synthesized by this technique. However, an appropriate fuel/oxidizer ratio in fuel rich conditions results in the formation of metallic phases, instead or co-existing with oxides [7]. Kumar et al. investigated the reaction pathways of metal nanopowders using nickel and copper as examples [11,12]. They reported, concerning the nickel-glycine system, that nickel oxide was formed at the early stage in the reaction front, and it was subsequently reduced to nickel in the post combustion zone. It is well-known in the field of catalysis, that nano-sized catalysts exhibit impressively high activity and selectivity [13]. Nanocomposites containing metals and metal oxides can also be obtained by SCS and offer a strong promise for increasing both the catalytic selectivity and the activity of nano-catalysts [14,15]. The activity of catalysts of different structures and compositions was studied during hydrogenation in the liquid phase, since this process is extremely sensitive to the slightest structural changes in the catalyst [16]. Heterogeneous catalytic hydrogenation uses finely crushed metal catalysts such as platinum, palladium, ruthenium, rhodium, osmium, copper and nickel either in pure form as pellets or coated on inert carriers [17–24] which are insoluble in organic solvents. The most active among them are ruthenium and rhodium, but platinum and nickel are the most widespread, the latter mainly because it is much cheaper. Hydrogenation of alkenes with a catalyst is usually carried out at ambient pressure and at a temperature of 20–80° C in alcohol, acetic acid or water. SCS produces nano-powders in a single step by a simple and rapid process [3,4]. Their specific surface area is usually small (up to 10 m2/g), while in some cases it can be extremely high. For example, Cross et al. [15] report Ni nanoparticles with extremely high surface area (155 m2/g) supported on fumed silica (SiO2) as catalysts for the ethanol decomposition toward hydrogen at low temperatures (200 °C). In order to prevent the undesired oxidation process that follows the formation of the highly dispersed nickel nanoparticles, the synthesis was conducted in an inert atmosphere. Interestingly, the low oxygen concentration during combustion passivated the nickel nanoparticles through the formation of a thin amorphous oxide layer. In the present study, hydrogenation was carried out in the liquid phase at 80 °C on Ni/NiO SCS derived nano-catalysts. It was found that their physical properties and structure depend in a

complex way on the parameters of SCS processing, composition of the initial SCS solution [25] and even the amount of water used in the initial solution [26]. Various reaction mechanisms (mainly at the atomic-level) have been identified as being active during the synthesis of these materials which critically influence their catalytic properties. Controlled hydrogen adsorption studies have helped to clarify the main mechanisms involved. Understanding the interrelationships between the processing parameters and the ensuing structure has allowed a degree of optimization of the catalytic properties of the new catalysts. 2. Methodology 2.1. Materials

Material

Manufacturer

Assay

Nickel (II) nitrate hexahydrate for analysis [Ni(NO3)26H2O] Glycine for synthesis [CH2NH2COOH] Maleic acid [HOOCCH = CHCOOH]

Merck

99.0– 102.0% 99.0%

PanReac AppliChem Riedel-de Haën

99.0%

2.1.1. Solution combustion synthesis of Ni-based catalysts The initial solution contains nickel nitrate hexahydrate (Ni (NO3)26H2O) as the oxidizer with glycine as the reducer. In the series of experiments carried out, four types of samples were synthesized keeping the quantity of nickel nitrate constant at 9.34 g in the initial mixture with 75 ml distilled water, adding specific amounts of glycine (80, 60, 50 and 40 wt% of nickel nitrate) to achieve fuel to oxidizer molar ratio (u) of 2.78, 2.08, 1.74 and 1.4 for the production of Ni and NiO where u = 1.0 corresponds to the stoichiometric composition for the equation (1), below.

3NiðNO3 Þ2 6H2 O þ C2 H5 NO2 ! 2Ni þ NiO þ NO2 5 11 O2 þ 2NO þ N2 þ 2CO2 þ 18O þ 2 2

ð1Þ

Each solution was pre-heated in a borosilicate glass beaker on a hot plate with mild magnetic stirring until the temperature reached 70 °C and then placed in a pre-heated furnace at 500 °C to initiate SCS, yielding an extremely friable foam-like structure consisting of nano-sized powders lightly bound together. Once SCS is completed the beaker is removed from the furnace and allowed to cool at room temperature. 2.2. In-situ monitoring and characterisation of the SCS process Determination of the temperature and velocity of the combustion wave during SCS was carried out using three 100 µm diameter K-type (chromel–alumel) thermocouples placed in and above the solution. A PICO TC-08 conditioner was used to convert, filter and amplify the temperature signals which are recorded at a rate of 0.5 ms. No attempt was made to move the thermocouples since these measurements were used only to pinpoint the reaction time and correlate it to the appearance of combustion phenomena. In any case, this work is being extended in order to compare the measured reaction temperatures with calculated adiabatic reaction temperatures and will be published later. The temperature of the combustion wave was also measured in-situ using a high-speed infrared camera (FLIR Systems, model A655SC). The speed of the combustion wave was calculated by analysis of the data obtained from the high-speed infrared camera.

114


For in-situ observation of the combustion process and study of the dynamics of phase formation in the reaction front of the combustion, wave time-resolved (dynamic) X-ray diffraction (DXRD) was used. This method is a powerful tool for studying the mechanism of phase formation, since it can monitor a sample’s phase compositional changes in real time during the propagation of the reaction wave. For this purpose, the samples are placed in a small PTFE container of dimensions 20x10x2mm inside a furnace equipped with an X-ray source with a copper anode. A monochromator made of pyrolytic graphite was attached to the X-ray tube. The window of the chamber is made of a beryllium plate to ensure minimum possible X-ray absorption. Diffracted X-rays from the sample are recorded in the horizontal plane by using a 1dimensional position-sensitive detector which can receive data over a range of 2h of 30° to 75°. The detector control, synchronization of measurements and data capture and processing were all carried out by means of a computer. Thermogravimetric analysis (TGA) and differential thermal analysis (DTA) were used to monitor the change in sample weight and any endothermic and exothermic thermal transformations occurring during SCS using an STA 449F1 instrument from NETZSCH-Gerätebau GmbH, Germany. Heating rate for all samples (weight of 17–21 mg) was 30 °C/min between 40 °C and 500 °C and all experiments were conducted under Ar with a flow rate of 80 ml/ min in an alumina crucible.

An attempt was made to detect the formatting complexes in the precursor solution during preheating using 2D NMR spectra on a Bruker Avance 500 MHz spectrometer at 298 K. All data were acquired and processed with Topspin 1.3 software. Three different compositions were tested by NMR following 10–15 min sonication. The first one was a mixture of glycine and deuterium oxide (D2O) with concentration of 0.74 mg/1 ml D2O. The second test included a mixture of glycine, nickel nitrate and deuterium oxide [composition (glycine 0.74 mg + + nickel nitrate hexahydrate 1.3 mg)/1 ml D2O], while during the third test the quantity of nickel nitrate hexahydrate changed from 1.6 mg to 2.6 mg. The scanning electron microscope (SEM) that has been used for the study of specimens is Quanta Inspect from FEI Inspect system and the chosen points’ analysis has been yielded by EDS graphs from EDAX apparatus. Before SEM analysis, the specimens were spatter coated with gold (coating thickness 5–10 nm). The transmission electron microscope that has been used in this work is CM 20 FEI from Philips Company. 2.4. Catalytic hydrogenation of maleic acid The activity of the SCS catalysts was studied for aqueous hydrogenation of maleic acid producing succinic acid, as presented in the Eq. (2) below. The installation, as well as its operating principles, is extensively discussed in a previous work [25,26].

ð2Þ 2.3. Characterisation of the SCS products The atomic structure and microstructure of the synthesised catalysts were determined by several methods. Atomic structure was determined by X-ray diffraction measurements on a Siemens Spellman DF3 spectrometer with Cu-Ka radiation. For semiquantitative XRD analysis, 10% KCl was added in all samples as an internal standard. By normalising the XRD results against the 10% KCl internal standard and calculating the peak ratio of intensities of particular peaks (making sure that they are unique for each phase being analysed) it is possible to determine the relative content of each phase in the material. The hkl of the peaks selected for each phase were: 111 for nickel, 101 for nickel oxide and 100 for potassium chloride. B.E.T. specific surface area and pore distribution measurements were carried out on a GAPP V-Sorb 2800 Analyser using nitrogen (99.9%) with helium (99.999%) as carrier gas. All samples were subjected to degassing under vacuum (310-4 Torr) in two stages during pretreatment. During the first stage the samples were heated up to 80 °C for 45 min, while in the second stage they are heated up to 150 °C for another 45 min.

The catalytic reaction takes place in a shaker-reactor enclosed inside a water bath kept at 80 °C. An amount of 1.0 g of powdered catalyst is added in the catalytic reactor with 30 ml of distilled water. After this, the maleic acid is added in a quantity (0.26 g) calculated for reaction with 50 ml of hydrogen at atmospheric pressure. A calomel electrode is used in the reactor in order to monitor hydrogen adsorption and the hydrogen saturation of the catalysts in the solution is monitored over one hour. The Pt electrode is immersed in the solution containing the catalyst. Both catalyst saturation with hydrogen and hydrogenation are carried out under continuous horizontal shaking at a rate of about 500 cycles per minute. Measurements of the amount of reacted hydrogen are taken every minute at atmospheric pressure. To check reproducibility, each hydrogenation test was repeated at least three times. 3. Results and discussion Of particular current interest worldwide is the development of technological methods for obtaining hierarchical three-

1mm

1mm

Fig. 1. Thermal photos of the combustion wave during SCS in the system with u = 1.4 using IR camera (left) and high speed camera (right).


dimensional network structures permeated with nano-sized pores. Such three-dimensional percolation structures of nano-composites on the basis of metal oxides and metals obtained by combustion in solution provide the capability for increasing the selectivity and activity in catalysis. During the SCS combustion experiment with u = 1.4, two different cameras were used in order to observe combustion synthesis processes as well as to capture the combustion wave during its development. As shown in Fig. 1, the width of the combustion wave is approximately 1 mm and the temperature during combustion in the wave is in the range of 650–730 °C. The evolution of combustion temperature with time for the three different u values used is presented in Fig. 2a–c. In each case, combustion regime and maximum combustion temperature change. As indicated in Fig. 2a-c and taking into account additional temperature measurements with thermocouples in and above the solution, increasing glycine concentration in the initial mixture from u = 1.4 to u = 2.08 and to u = 2.78 leads to a decrease in the maximum combustion temperature from over 750 °C for u = 1.4 (in this case thermocouple readings showed temperatures as high as 1150 °C) to about 700 °C for u = 2.08 and about 680 °C for u = 2.78. Interestingly, in the case of the fuel-rich specimen (u = 2.78), after the end of the reaction carbon particles were observed coating the inside of the container which continued to burn in a flash mode one of which is indicated by the small sharp peak at about 200 sec in Fig. 2c. These results are in line with expectations since the maximum combustion temperature would be expected to decrease as we move away from the stoichiometric composition u = 1.0. In addition, these combustion observations also allowed a clarification of the changes in the combustion regime itself. While for u = 1.4 combustion occurred ‘‘in the volume”, where combustion initiates from many points in the whole volume of the gel, in the cases where u is higher combustion initiates and proceeds in the selfpropagating wave regime. For example, in the sample with u = 2.08 (Fig. 2b), two combustion waves were observed: one wave ran from top to bottom with a low speed (about 0.06 cm/s) and as soon it reached the centre a second wave started at the bottom of the specimen which propagated towards the first with a higher speed of about 0.3 cm/s. Therefore, in this sample the two-wave combustion resulted in a double heat treatment of the products, one from each wave. These differences in regime of combustion are reflected in the products’ composition and structure (Fig. 3). SCS products in general display a pronounced 3-dimensional dendritic foam-like structure as illustrated on Fig. 3 which shows how materials appear after SCS. The different colours that appear in Fig. 3 indicate variations in microstructure and crystal orientation. The dendritic structure is due to the formation of hydrates while still in solution in combination with extensive gas emission during SCS. Glycine reacts with nickel nitrate producing glycine-nickel nitrate tetrahydrate [C2H13N3NiO12] according to Fleck et al. [27]. An attempt was made to find this complex by NMR (Nuclear Magnetic Resonance) and the results are presented on Fig. 4. Three different compositions were tested: the first one was a mixture of glycine and deuterium oxide (D2O) (black line), the second test included a mixture of glycine, nickel nitrate and deuterium oxide (blue line), while during the third test the quantity of nickel nitrate hexahydrate was doubled (pink line). There are two peaks demonstrated on Fig. 4. The first peak represents the mixture of glycine and nickel nitrate in two different concentrations (1.3 and 2.6 mg/mL D2O), while glycine shows a second peak at 3.47 ppm. It appears that by increasing nickel nitrate’s concentration, the peaks of nickel nitrate and glycine become wider. This may be evidence of the synthesis of this complex, since no evidence has been found

115

in the literature of any paramagnetism in nickel nitrate, whether or not it is in solution with glycine. In any case, paramagnetism is indicated when a substance is weakly attracted to a magnetic field and generally occurs when there are unpaired electrons in the material. Indeed, paramagnetism is found in the salts of some of the first row transition metals, i.e. manganese through nickel since these metal ions have unpaired electrons in degenerate d orbitals, as predicted by Hund’s rule. However, salts of those ions that have strong-field ligands are

Fig. 2. Temperature evolution during SCS for u values a) u = 1.4b) u = 2.08c) u = 2.78 (SCS preheating temperature 300 °C).

116


diamagnetic [28] since the strong-field ligands can split the energy levels of the d orbitals so that they are no longer degenerate. At the same time, for the hydrated nickel nitrates Ni(NO3)22H2O, Ni(NO3) 24H2O and Ni(NO3)26H2O, zero-field splitting, anisotropy of magnetic susceptibility, proximal and far magnetic order interactions, as well as metamagnetism have all been reported [29]. Measurements of magnetic properties have been carried out over a wide range of temperatures and magnetic fields in a SQUID magnetometer and a vibration magnetometer. The Ni(NO3)24H2O, displays single-ion anisotropy and low-symmetry environment. In the Ni(NO3)26H2O, short-range order is observed, associated with the fact that the exchange interaction is much less than the zerofield splitting of Ni2+ ion [14]. Thus, on balance, these results tend to confirm that the widening of the spectral peaks observed in Fig. 4, are probably due to complex formation. Significantly, the SCS products’ final structure exhibited noticeable differences when glycine’s concentration was varied. A possible explanation of this phenomenon would be that increasing of fuel concentration leads to higher combustion temperatures thereby enhancing sintering. On the other hand, the observed aggregation process at high u could be connected with extended combustion due to fuel excess, soot origination and soot burningout process after combustion. Under these conditions, sintering of nanoparticles is possible even as low as 500 °C. Results obtained from the thermogravimetric (TG) and differential thermal analysis (DTA) analysis results seem to support such a conclusion. These tests were carried out for compositions with / = 1.4 and / = 2.8 and are shown in Fig. 5. It can be seen that the auto-ignition temperature for both cases is very low at about 200 °C. Decreasing nickel nitrate concentration in the initial SCS mixture results in decreasing formation of nitrogen oxides gases. As shown in Fig. 5, TG curve drops 50% (u = 2.78) and 70% (u = 1.4) in the range of temperatures 25–200 °C due to the decreased amount of released NO and NO2. Moreover, the step-like weight loss is connected to the loss of three lattice water molecules from nickel nitrate hexahydrate at 95 °C and all lattice water molecules at 110 °C. At approximately 150 °C glycine starts to burn and at 200° C there is a minimum of the DTA curve, which signifies maximum heat release from solution combustion synthesis. Even after SCS has completed, a further weight loss is recorded as glycine’s decomposition products have only reacted partially during combustion and soot burning-out takes place at this point. In the case of u = 2.78 (Fig. 5a) this process is more noticeable on the TG curve. This is also shown in the DTA curves where in both cases two exothermic peaks are evident corresponding to the SCS exothermic effect and the burning out soot process.

The development of the various phases as a function of amount of water in the SCS mixture is shown in the series of XRD spectra in Fig. 6. All catalysts tested (Fig. 6) had two main products: metal Ni and NiO. These products are the result of the following multi-branch reaction cascade: 54 C

85:4 C

NiðNO3 Þ2 6H2 O ! NiðNO3 Þ2 4H2 O ! NiðNO3 Þ2 2H2 O

ðiÞ

3NiðNO3 Þ2 6H2 O þ C2 H5 NO2 ! 2Ni þ NiO þ NO2 5 11 O2 þ 2NO þ N2 þ 2CO2 þ 18O þ 2 2 ðiiÞ 7 2NiðNO3 Þ2 6H2 O ! 2NiO þ NO2 þ NO þ N2 þ 12H2 O þ O2 2

ðiiiÞ

NiðNO3 Þ þ H2 O ! NiOHNO3 ; NiOHNO3 þ H2 O ! NiðOHÞ2 þ HO3

ðivÞ

CH2 NH2 COOH ! ½CH2 COO þ NH3

ðvÞ

HNO3 þ NH3 ! H2 O þ N2 þ H2

ðviÞ ðviiÞ

NiO þ H2 ! Ni þ H2 O

ðviiiÞ

1 Ni þ O2 ! NiO þ O2 2

ðixÞ

2CH2 NH2 COOH þ 6O2 ! 4CO2 þ 5H2 O þ NO þ NO2

ðxÞ

2CH2 NH2 COOH þ 4O2 ! 2C þ 2CO2 þ 5H2 O þ NO þ NO2

ðxiÞ

C þ O2 ! CO2

ðxiiÞ

The phase composition of the products, as determined by the comparative XRD analysis and crystal lattice spacing relative ratio, as a function of fuel to oxidizer ratio (u) in the SCS solution are presented in Fig. 7, where the variation between Ni/KCl and NiO/KCl are shown together for comparison in Fig. 7a where the catalytic activity is also shown. As shown in Fig. 7a, the maximum nickel content in the SCS products is found for u = 1.4, where the nickel oxide’s concentration is minimum. In addition, the catalytic activity shows a direct

Fig. 3. Influence of u (glycine/nitrate molar ratio) on the structure of SCS catalyst: a) u = 1.4b) u = 1.74c) u = 2.08 d) u = 2.78.


dependence on nickel concentration in the final product which is correlated with the finding that the activity curve has the same gradient as the nickel concentration curve. Comparing results in Figs. 7a and 2b, it is seen that during the SCS reaction for u = 2.08 two well-defined combustion waves (expanding from top to bottom) appeared during SCS. It is probable that during the first wave, there was hydrogen formation in the gas phase (reaction (vi)) followed by further reduction of NiO (reaction (viii) in liquid phase and immediate partial oxidation of nickel (reaction (ix)) dur-

117

ing the second wave as a result of the high temperatures achieved during the previous exothermic reactions. In the other cases, the oxidation process is less intense which explains the minimum observed in the Ni concentration curve and the maximum of NiO concentration for this particular sample. The reason why only for u=2.08 were two combustion waves observed is that this is the stoichiometric ratio for hydrogen production (reaction (vi)), which is a highly exothermic reaction and under these conditions reduction of nickel oxide is favoured.

Fig. 4. NMR spectra of Ni(NO3)2 6H2O and glycine in D2O with the two spectrum peaks magnified.


12

0

DTA

-50 -100 -150 -200 -250 -300 100

200 300 400 Temperature, oC

500

4 10 3.5 8

2.5 6 2 4 2

0.5 0 1.4

1.74

2.08

2.78

φ 50

(a)

0

DTA

-100 -150

DTA, μV

-50

-200 -250 -300 200 300 400 Temperature, oC

500

600

Ni crystal lace spacing (d), Å

TG

100

1

0

100 90 80 70 60 50 40 30 20 10 0 0

1.5

Ni (1,1,1)/KCl (1,0,0) NiO (1,0,1)/KCl (1,0,0) Acvity

600

(a)

Mass loss, %

3

2.04

4.5 d

4

acvity 2.035

3.5 3 2.5

2.03 2 1.5 2.025

Acvity, ml/gr*min

0

4.5

Acvity, ml/g*min

TG

Rao of XRD spectrum peaks

50

100 90 80 70 60 50 40 30 20 10 0

DTA, μV

Mass loss, %

118

1 0.5

2.02

(b)

0 1.4

1.74

2.08

2.78

φ

Fig. 5. TGA-DTA analysis of samples with / = 2.78 (a) and / = 1.4 (b).

(b) 1: KCl 2: Ni 3: NiO

2

4000 2000

2

φ = 1.4 1

2

1

0 0

20

2

1

40

60

80

2 100

2 4000 2000

2

φ = 1.74

2

Intensity

1 20

2

1

1

0 0

40

60

80

2 100

4000 3 2000

φ = 2.08

3 1 20

3

2

1

32

0 0

40

1 60

4000

32 3

2 3 2

80

100

2

φ = 2.78

2000

1

3

3

2 3

1

0 0

Fig. 7. Dependence of relative crystal lattice spacing (ratio of XRD peaks) and catalytic activity on fuel to oxidizer ratio in initial SCS mixture.

20

40

60

1

2 3 3 80

2

2 100

2 theta Fig. 6. XRD spectrum peaks of the SCS catalysts on the basis of various ratios (u) between nickel nitrate hexahydrate and glycine.

The results shown in Fig. 7b also reveal that the nickel crystal lattice spacing of 2,036 Å for h k l = 1 1 1 is (close to) optimum for maleic acid hydrogenation. In Fig. 8 a sequence of 50 XRD traces obtained by Time-resolved X-ray diffraction (TRXRD) studies (measured each second) describes the process of phase formation during SCS for composition with u = 2.08. The spectra are presented in both 2D of 2H vs time as well as in 3D of 2H vs intensity vs time. The initial solution

nitrate-glycine gives a diffraction spectrum with wide amorphous halo that indicates an absence of phases in the crystalline state. On SCS ignition nickel lines (1 1 1) and (2 0 0) appear first and within 6 s their intensity increases. A little later, a reduction of their amplitude with simultaneous appearance of lines (1 0 1), (0 1 2) and (1 1 0) of NiO is observed. This means that after formation Ni, its content starts to decrease with time while the concentration of NiO increases. It is therefore feasible to conclude that the second wave observable for SCS with u = 2.08, is connected with oxidation of Ni which has formed in the initial stage. The reason why only in this ratio nickel nitrate –fuel there are 2 waves is that at u = 2.08 there are stoichiometric conditions for hydrogen production (reaction (vi)), which is highly exothermic reaction and these conditions reduction of nickel oxide with further oxidation of nickel is easy. The temperature profile measured during actual SCS combustion is presented on Fig. 9. The exothermic peaks visible correspond to the formation of hydrogen (reaction vi), nickel (reaction viii) and soot combustion (reaction xii). Note that the speed of heating during the DTA tests and temperature profile analysis was different which explains the different number of detected exothermic peaks in each analysis. In Fig. 9, the measurements were taken within the gel (‘‘bottom thermocouple”) and within the gas phases (‘‘top thermocouple”). Fig. 9 (bottom thermocouple) indicates that the SCS reactions were mainly active within the gel where the concentration of components is higher. In the gas phase (top thermocouple) the exothermic effects detected were mainly due to reactions (vi) and (vii), which concern hydrogen formation. It is also possible that reaction (viii), where nickel is formed, influences the formation of

119

(101) NiO

(200) Ni

Intensity, imp.

(111) Ni


50

2Θ , deg. 30 amorphous galo 35

(101) NiO

40

(012) NiO (111) Ni

50

Θ

2 , deg.

45

(200) Ni 55

60

(110) NiO

65

70

time, s Fig. 8. Results of time-resolved X-ray diffraction analysis in an SCS gel with u = 2.08.

Combuson temperature, oC

800 700

φ=2.78

φ=1.4

600 500 400 300 φ=1.4 top φ=1.4 boom φ=2.78 top φ=2.78 boom

200 100 0 0

200

400

600

800

1000

Combuson me, sec Fig. 9. Temperature curves for samples with u = 0.4 and u = 0.8 in the initial mixture (preheating temperature 500 °C). These measurements were carried out during actual SCS and resulted in more detailed data concerning all the reactions yielded.

dendrites (when volume of product is significantly increased) which results in the exothermic effect detected by the top thermocouple. Comparison of Figs. 2 and 9 indicates that, at 500 °C, there are bigger differences in combustion temperature than at 300 °C, which may be attributed to the different conditions of the experiments (temperature and volume of initial mixture) and the different methods of measurement. The kinetic and conversion curves obtained from catalytic maleic acid hydrogenation to succinic acid on nickel SCS catalyst are presented on Fig. 10. As shown on Fig. 10a, maximum conversion of maleic acid was measured for the catalysts produced with u = 1.4 and u = 1.74. The most active catalyst (Fig. 10b) and the one with the highest conversion is the one with fuel/oxidizer ratio of 1.4. Activity is, in many cases, connected with active component’s concentration in the catalyst, its parameters of crystal lattice structure and specific surface area.


8

4.5

90

7

4

6

3.5

Speciﬁc surface area, m2/g

100

% Conversion

80 70 60 50 40

φ=1.4 φ=1.74 φ=2.08 φ=2.78

30 20 10

2.5 4 2 3

1.5

2 1

SSA

1

acvity

0.5 0

0

0

0

20

40 60 80 Reacon me, min

100

1.4

120

1.74

2.08

2.78

φ

(a)

(a)

4.5

4.5

4

3.5

Acvity, ml/gr*min

φ=1.74 φ=1.4 φ=2.08 φ=2.78

4 cm3 H2/g Ni catalyst

3

5

Acvity, ml/g*min

120

3 2.5 2 1.5

3.5

φ=1.4

3

φ=1.74

2.5 2

φ=2.78

1.5 1

φ=2.08

0.5

1

0 40.08

0.5 0 0

10

20 30 ΣVH2, cm3

40

50

(b)

41.72 47 Ni crystallite size, nm

49.15

(b) Fig. 11. (a) Influence of glycine concentration on the surface area and activity of the SCS catalysts and (b) influence of Ni crystallite size on the activity for various u values.

Fig. 10. Dependence of maleic acid hydrogenation (a) conversion and (b) velocity on fuel to oxidizer ratio u.

BET specific surface area measurements indicate strong dependence of surface area on the u ratio in the initial SCS mixture (Fig. 11a). The most active catalysts (with u = 1.4 and u = 1.74) display higher surface area (7m2/g) than the catalyst with u = 2.08, which would explain the enhanced activity. On the other hand, the catalyst with u = 2.78 also displays high surface area, but is less active in hydrogenation process which illustrates the fact that there are many parameters that influence specific surface area – in complex and often opposing ways - and subsequently catalytic activity. The measured decrease in the surface area of the catalysts (and associated decrease of their catalytic activity) with increasing fuel content is probably related to the increase in Ni crystallite size which results from increasing sintering of crystallites (Fig. 11b). The measured increase of surface area and activity at u = 2.78 (Fig. 11.a) is probably connected with changing of the SCS mechanism at high concentration of fuel as hydrogen production is favoured which also leads to reduction of NiO to Ni (Fig. 7a) which is a catalyst in this reaction. The catalyst’s crystallite size is an important factor for catalytic activity and, as can be seen in Fig. 11b, reduction of crystallite size leads to increased surface area. This is mainly attributed to the better dispersion of nickel on nickel oxide when the nickel crystallites are smaller. According to the obtained data, for u = 1.4, the minimum of nickel crystallite size corresponds to maximum activity in maleic acid’s hydrogenation. The hydrogen adsorption–desorption curves obtained by the BET method are shown in Fig. 12a–b, where the hysteresis curves obtained for SCS Ni catalysts produced here are shown. Based on the IUPAC classification of isotherms, the curves obtained are prob-

ably Type IV. Their main characteristic is their hysteresis loop, which is associated with capillary condensation taking place in mesopores and limited uptake over a range of high P/Po. The type of hysteresis loop (at u = 1.4) shown in Fig. 12 is Type H3, which is generally observed with aggregates of plate-like particles giving rise to slit-shaped pores. In the case of SCS composition with u = 2.78 (Fig. 12b) the hysteresis loop is much wider than in case of u = 1.4 (Fig. 12a) probably because of difficulty in desorption due to different pore structure, since in this case, pores are cylindrical with narrow necks. Such narrow ‘‘bottle neck” pores may be the result of soot oxidation on the surface of the catalyst, since the pores on the surface become narrower due to sintering. In such narrow neck pores, nitrogen desorption is slow, and so is for maleic acid as well. With such a pore shape, adsorption and desorption phenomena are quite laborious, and consequently reactions cannot proceed which leads to low catalytic activity for the catalyst (u = 2.78) which has relatively high specific surface area (Fig. 12a). The shape of the pores affects heat release during SCS, the cooling rate after SCS and hydrogen adsorption–desorption, all of which influence the structure of the catalysts and their activity. The pore shape of the catalysts made with u = 2.78 means that heat release is slow and cooling rate is low in comparison with the other compositions. The above contribute to more intense sintering conditions which results in a more compact catalyst structure (Fig. 3d). It also explains why in the catalyst with u = 2.78 there is less cumulative micropore volume (Fig. 13b) and more mesopores than in case of catalyst with u = 1.4 as shown in Fig. 13a. As a result, the structure of all four synthesized catalysts is highly complicated: they display nanopores, mesopores and macropores, but propor-

121

Desorpon

0.0035

0.001

0.003

1.5 1

0.0025

0.0008

0.002 0.0006 0.0015 0.0004 0.0002

0.5

0 1

0 0.2

0.4 0.6 0.8 Relave pressure, P/Po

10

100

dV/dW

0.001

Cumulave

0.0005

0 1,000

Width, nm

1

(a) 0.004

0.0003

(a) 1.6

0.0035

dV/dW, cm3/g*nm

0.00025

1.4 1.2 1 0.8

0.003 0.0002

0.0025 0.002

0.00015

0.0015

0.0001 dV/dW Cumulave

0.00005

0.6

0

0.4

1

Adsorpon 0.2

10

100

0 1,000

(b)

0 0.2

0.0005

Width, nm

Desorpon 0

0.001

Cumulave pore volume, cm3/g STP

2

0

Quanty of adsorbed N2, cm3/g STP

0.0012

Adsorpon dV/dW, cm3/g*nm

Quanty of adsorbed N2, cm3/g STP

2.5

Cumulave Pore Volume, cm3/g STP


0.4 0.6 0.8 Relave pressure, P/Po

1

Fig. 13. Pore size distribution curves and cumulative pore volume of SCS Ni catalyst (a) u = 1.4, (b) u = 2.78.

(b) 4.5

tion between them is different, as illustrated in the example of two of the catalysts in Fig. 13. Diffusion of the initial SCS mixture’s components into the catalysts’ pores plays a significant role due to the fact that the hydrogenation reaction of maleic acid is in the liquid phase and at relatively low temperature (80 °C). For the same reason, adsorption of maleic acid and hydrogen is the limiting stage in the liquid phase hydrogenation. Fig. 14 shows the dependence of the catalysts’ activity for hydrogen adsorption in the solvent (water) on their specific surface area. The above results show that the concentration of glycine with respect to the amount of nickel nitrate in the SCS mixture (i.e. ratio u) influences the catalyst’s composition, structure, surface and activity. The results of TEM and SEM examinations of the Ni nanopowder produced by SCS are shown in Figs. 15 and 16, together with EDX analyses of selected structures. The images in Figs. 15 and 16 indicate that the nano-structured SCS catalyst has hierarchical percolation structure. The TEM images (Fig. 15) showed that the NiO nanoparticles were highly aggregated. The particles were found to be very close to spherical with a homogeneous distribution. Fig. 16a depicts typical aggregates derived by SCS, displaying tight three-dimensional flakes tens of micrometers in size. The flakes possess a relatively smooth surface. With the addition of NaF into the precursor SCS solution, the three-dimensional flake structure was broken up and switched to highly fluffy and porous networks consisting of fine nanoparticles (Fig. 16b). The aggregates shown in Figs. 15 and 16 were very similar throughout, indicating uniform particle shape and size. Their microstructure exhibits uniformly distributed foam-like (mem-

% H2 surface coang

96

H2 coang

4

acvity

3.5

94

3

92

2.5

90

2 1.5

88

1 86

Acvity, ml/gr*min

98

Fig. 12. Hysteresis curve of a SCS Ni catalyst (a) u = 1.4, (b) u = 2.78.

0.5

84

0 1.4

1.74

2.08

2.78

φ Fig. 14. Influence of glycine concentration (u ratio) on the hydrogen adsorption on SCS nickel catalysts and their catalytic activity.

brane) structure with a ‘‘wall” of about 5–7 lm in thickness (Fig. 16b). Fig. 16a shows large pores in the sub-mm range, while Fig. 16b displays the higher magnification with sub-lm pores; the constituting NiO particles are not yet visible. We suggest that such high amount of porosity is the result of extensive gas release during combustion. According to the EDS data in Fig. 16, the porous ‘‘wall” consists mainly of fine porous nanoparticles based on Ni and O (Fig. 16e). Fig. 16e illustrate that the catalyst surface composition is different from bulk composition (Figs. 6 and 16a). Bulk composition for this catalyst is mainly Ni (Fig. 6) with traces of NiO which is also confirmed in the SEM/EDX examination in Fig. 16a. Surface analysis of pores by high resolution SEM (Fig. 16b and d) show that the surface of the pores is covered by NiO (estimated at 55.4–68.1%

122


This kind of structure and composition of the catalysts as shown in Fig. 16a and 16c strongly indicate some kind of percolation type of hierarchical structure, where the sponge like structures are nickel with NiO on the inner surfaces of pores. It is possible that such a distinctive difference in composition between the body of the pores and their surface may result in higher selectivity of these catalysts since the difference in microstructure of Ni and NiO plays different roles in the adsorption- desorption process of hydrogenation reaction. It therefore appears that SCS offers an efficient route for the production of high selectivity three-dimensional percolation network hierarchical structures of catalytic nano-composites of metals with metallic oxide coatings.

4. Conclusions

Fig. 15. TEM images of SCS Ni-based nano-powders (u = 1.74) with 75 ml distilled water. The particle size on the left is several hundred nanometers; while the particle size of the powder on the right is only 20–30 nm.

NiO). This means that during SCS the dendrites formed consist mainly of Ni but the surface of the pores is oxidized because of lower heat loss, lower rate of cooling after reaction which means there is enough time for oxidation inside the pores. The hierarchical order of a structure or a material may be defined as the number of levels of scale with recognized structure. Uniform three-dimensional (3D) flowerlike Ni-NiO hierarchical architectures assembled from nanosheet building blocks have been successfully fabricated via a simple and direct SCS method (Fig. 16b and d). Hierarchically nanoporous catalysts have versatile structural properties such as increased surface area and large pore volume that can alleviate diffusional limitations of conventional nanocatalysts with solely microporous framework (which is important for liquid phase heterogeneous catalysis. Hierarchical structure formation also confirm TEM analysis (Fig. 15).

The catalytic activity of SCS-derived nano-structured Ni-based hydrogenation catalysts depend on composition (nickel and nickel oxide), specific surface area, crystal lattice spacing and pore size and shape. Thermogravimetric analysis showed that for glycine concentration greater than the stoichiometric ratio (u = 2.08) between nickel nitrate and glycine, the decomposition product (soot) reacts partially during combustion but most of soot oxidation takes place after SCS is completed. This leads to increased sintering and aggregation processes, changing the open shape of the pores to ‘‘bottle neck”, changing the ratio between nanomicro and macro pores and consequently changing adsorption – desorption process and catalytic activity. Data obtained from IR high speed camera show that the combustion mechanism depends on the glycine concentration – at high glycine concentration there is a mechanism change from combustion in volume to combustion in self-propagating regime. It was found by temperature profile measurements during SCS that the reactions take place mainly within the gel where the concentration of components is higher. In the gas phase, above the SCS solution, the exothermic effects detected mainly concern hydrogen formation. Uniquely, in sample with u = 2.08 (stoichiometric ratio) two waves occurred during combustion: hydrogen formation takes place in first wave followed by reduction of nickel oxide in the second wave together with Ni oxidation. This is illustrated by the observed minimum of Ni and maximum of NiO in this case. When hydrogen is formed in the gas phase, a combustion wave propagates from the top to the bottom of the reactor while the second combustion wave is initiated at the bottom (within the gel) – first by reduction of Ni and then by partial Ni oxidation, because of the higher temperature developed. Widening of NMR spectra peaks observed may be connected with glycine-nickel nitrate complex dendritic formation. During SCS, the dendrites formed consist mainly of Ni, but surface of pores is oxidized because of lower heat loss, lower speed of cooling after reaction which gives sufficient time for oxidation of the inside surface of the pores. Understanding the interrelationships between the processing parameters and the ensuing structure has allowed a degree of optimization of the catalytic properties of the new catalysts. The three-dimensional percolation-like network and hierarchical structure of nano-composites on the basis of metal oxides and metals obtained by combustion in solutions provides a distinct possibility of increasing the selectivity and activity of such catalysts.

123


(c) (a)

(b)

(d)

Spectrum 1 2 3

55.5 31.9

Std. deviaon 31.9

(e) Fig. 16. SEM examination and EDX analysis of SCS catalyst (/ = 2.08) of ‘‘wall” of pores (a) and of inner surface of pores (b). The morphology of the synthesized NiO high porosity agglomerates is shown in c, d and e.

124


Acknowledgements The authors would like to express their gratitude to N. Boukos (Fig. 15) for TEM, A. Shchukin for SEM (Fig. 16d) and K. Yannacopoulou for the NMR measurement in Fig. 4. Part of this work was carried out with financial support from the Ministry of Education and Science of the Russian Federation in the framework of the Increase Competitiveness Program of MISiS (project no. K2-2016-002). References [1] P. Dinka, A.S. Mukasyan, In situ preparation of oxide-based supported catalysts by solution combustion synthesis, J. Phys. Chem. B 109 (46) (2005) 21627– 21633. [2] S. Specchia, E. Finocchio, G. Busca, V. Specchia, Combustion synthesis, in: M. Lackner, F. Winter, A.K. Agarwal (Eds.), Handbook of Combustion, Wiley-VCH, Weinheim, 2010, pp. 439–472. [3] A. Varma, A. Mukasyan, K. Rogachev, Manukyan, Solution combustion synthesis of nanoscale materials, Chem. Rev. 116 (23) (2016) 14493–14586. [4] S. Mukasyan, P. Dinka, Novel approaches to solution-combustion synthesis of nanomaterials, Int. J. Self Propag. High Temp. Synth. 16 (1) (2007) 23–35. [5] K.C. Patil, S.T. Aruna, T. Mimani, Combustion synthesis: an update, Curr. Opin. Solid State Mater. Sci. 6 (6) (2002) 507–512. [6] H. Birol, C.R. Rambo, M. Guiotoku, D. Hotza, Preparation of ceramic nanoparticles via cellulose-assisted glycine nitrate process: a review, RSC Adv. 3 (2013) 2873. [7] W. Wen, J.-M. Wu, Nanomaterials via solution combustion synthesis: A step nearer to controllability, RSC Adv. 4 (2014) 58090–58100. [8] M. Lackner (Ed.), Combustion Synthesis: Novel Routes to Novel Materials, Betham Science Publishers, 2010. [9] S. Specchia, G. Ercolino, S. Karimi, C. Italiano, A. Vita, Solution combustion synthesis for preparation of structured catalysts: a mini-review on process intensification for energy applications and pollution control, Int. J. Self Propag. High Temp. Synth. 26 (2017) 166–186. [10] K.V. Manukyan, A. Cross, S. Roslyakov, S. Rouvimov, A.S. Rogachev, E.E. Wolf, A. S. Mukasyan, Solution combustion synthesis of nano-crystalline metallic materials: mechanistic studies, J. Phys. Chem. C 117 (2013) 24417–24427. [11] A. Kumar, E.E. Wolf, A.S. Mukasyan, Solution combustion synthesis of metal nanopowders: nickel-reaction pathways, AlChe J. 57 (8) (2011) 2207–2214. [12] A. Kumar, E.E. Wolf, A.S. Mukasyan, Solution combustion synthesis of metal nanopowders: copper and copper/nickel alloys, AlChe J. 57 (12) (2011) 3473– 3479. [13] S. Specchia, C. Galletti, V. Specchia, Solution combustion synthesis as intriguing technique to quickly produce performing catalysts for specific applications, Stud. Surf. Sci. Catal. 175 (2010) 59–67.

[14] S.L. Gonzalez-Cortis, F.E. Imbert, Fundamentals, properties and applications of solid catalysts prepared by solution combustion synthesis (SCS), Appl. Catal. A Gen. 452 (2013) 117–131. [15] A. Cross, S. Roslyakov, K.V. Manukyan, S. Rouvimov, A.S. Rogachev, D. Kovalev, Eduardo E. Wolf, A.S. Mukasyan, In situ preparation of highly stable ni-based supported catalysts by solution combustion synthesis, J. Phys. Chem. C 118 (45) (2014) 26191–26198. [16] G.C. Bond, The mechanism of the hydrogenation of unsaturated hydrocarbons on transition metal catalysts, Adv. Catal. 15 (1965) 91–226. [17] U.K. Singh, M.A. Vannice, Kinetics of liquid-phase hydrogenation reactions over supported metal catalysts – a review, Appl. Catal. A Gen. 213 (1) (2001) 1–24. [18] A. Corma, P. Serna, Chemoselective hydrogenation of nitro compounds with supported gold catalysts, Science 313 (5785) (2006) 332–334. [19] H. Kobayashi, Y. Ito, T. Komanoya, Y. Hosaka, P.L. Dhepe, K. Kasai, K. Hara, A. Fukuoka, Synthesis of sugar alcohols by hydrolytic hydrogenation of cellulose over supported metal catalysts, Green Chem. 13 (2011) 326–333. [20] A. Borodzinski, G.C. Bond, Selective hydrogenation of ethyne in ethene‐rich streams on palladium catalysts Part 1: Effect of changes to the catalyst during reaction, Catal. Rev. Sci. Eng. 48 (2) (2006) 91–144. [21] W. Baratta, L. Fanfoni, S. Magnolia, K. Siega, P. Rigo, Benzo [h] quinoline pincer ruthenium and osmium catalysts for hydrogenation of ketones, 2010, 9, 1419– 1423. [22] P.A.Z. Suarez, J.E.L. Dullius, S. Einloft, R.F. De Souza, J. Dupont, The use of new ionic liquids in two-phase catalytic hydrogenation reaction by rhodium complexes, Polyhendron 15 (7) (1996) 1217–1219. [23] J. Xiong, J. Chen, J. Zhang, Liquid-phase hydrogenation of o-chloronitrobenzene over supported nickel catalysts, Catal. Commun. 8 (3) (2007) 345–350. [24] S. Toppinen, T.K. Rantakylä, T. Salmi, J. Aittamaa, Kinetics of the liquid-phase hydrogenation of benzene and some monosubstituted alkylbenzenes over a nickel catalyst, Ind. Eng. Chem. Res. 35 (6) (1996) 1824–1833. [25] O. Thoda, G. Xanthopoulou, G. Vekinis, A. Chroneos, Parametric optimisation of solution combustion synthesis catalysts and their application for the aqueous hydrogenation of maleic acid, Catal. Lett. (2017), https://doi.org/10.1007/ s10562-017-2279-y. [26] G. Xanthopoulou, O. Thoda, E.D. Metaxa, G. Vekinis, A. Chroneos, Influence of atomic structure on the nano-nickel-based catalyst activity produced by solution combustion synthesis in the hydrogenation of maleic acid, J. Catal. 348 (2017) 9–21. [27] M. Fleck, L. Bohaty´, catena-Poly[[[tetraaquanickel(II)]-l-glycine-j2O:O0 ] dinitrate], Acta Crystallogr. Sect. E 61 (9) (2005) 1890–1893. [28] G.A. Boudreaux, L.N. Mulay, Theory and Applications of Molecular Paramagnetism, John Willey & Sons, California, USA, 1976. [29] V.V. Shutov, O.S. Volkova, E.N. Lapsheva, I.V. Morozov. Magnetic properties of Rb3Ni2(NO3)7 and Ni(NO3)2, IV Euro-Asian Symposium ‘‘Trends in MAGnetism” EASTMAG-2010, June 28 – July 2, Ekaterinburg, 2010, p. 350.