Reduction of Copper Oxide by Low-Temperature

0 downloads 0 Views 2MB Size Report
CSIR-Institute of Minerals and Materials Technology, Bhubaneswar 751013, India ..... Sabat KC, Rajput P, Paramguru RK, Bhoi B, Mishra BK (2014) Plasma Chem Plasma ... Lee HG (1999) Chemical thermodynamics for metals and materials.
Plasma Chem Plasma Process DOI 10.1007/s11090-016-9710-9 ORIGINAL PAPER

Reduction of Copper Oxide by Low-Temperature Hydrogen Plasma K. C. Sabat1 • R. K. Paramguru2 • B. K. Mishra3

Received: 7 March 2016 / Accepted: 17 April 2016 Ó Springer Science+Business Media New York 2016

Abstract The paper presents experimental results of reduction of cupric oxide (CuO) by low-temperature hydrogen plasma in a microwave-assisted plasma set-up. The experiments were carried out at low microwave powers in the range of 600–750 W and low hydrogen flow rates in the range of 0.833 9 10-6 to 2.5 9 10-6 m3 s-1. In all the experiments for reduction of CuO with hydrogen plasma, an initial induction period was observed in the kinetic plots. The induction period decreases with increase in pressure or temperature. The induction period leads to the formation of active sites for adsorption of H2. After the induction period, fast autocatalytic reduction takes place followed by a sluggish period towards the end. The reduction process proceeds in sequential steps through the formation of sub-oxides. The kinetic data fits the Avrami-Erofeev equation with ‘n’ value close to 3. The resultant activation energy measured during hydrogen plasma processing is around 75.64 kJ mol-1. This is lower compared to activation energies measured by other methods of reduction indicating a clear advantage. Keywords Reduction of cupric oxide  Low-temperature hydrogen plasma  Induction period  Avrami-Erofeev equation

Introduction Reduction of metal oxides by hydrogen is a standard method to prepare active catalysts and electronic devices [1]. Cupric oxide, CuO, is an antiferroelectric semiconductor with a bandgap of about 1.4 eV [2]. It can be reduced in a sequence CuO ? Cu4O3 ? Cu2O ? Cu

& B. K. Mishra [email protected] 1

Visiontek Consultancy Services Pvt. Ltd, Bhubaneswar 751016, India

2

KIIT University, Bhubaneswar 751024, India

3

CSIR-Institute of Minerals and Materials Technology, Bhubaneswar 751013, India

123

Plasma Chem Plasma Process

with the formation and subsequent transformation of sub-oxides to lower oxides leading to the formation of metallic copper [3]. Thus, mixtures of CuOx and Cu are used in the fabrication of components in microelectronics [3]. Further, CuO is also used as a catalyst or catalyst precursor in many chemical processes involving hydrogen [3, 4]. In the past, some studies have been made to understand the reduction of CuO by hydrogen [3–7]. Though the reduction may proceed in a sequence as mentioned above, the reduction with hydrogen follows this sequence only when hydrogen partial pressure is low. When hydrogen is supplied in excess of the stoichiometric amount the reduction is direct CuO ? Cu [3]. The important point in CuO reduction by hydrogen is the existence of an induction period, during which active sites are probably formed that have a high affinity for the adsorption and dissociation of H2 [3, 4, 8]. The induction period is reduced at a higher temperature, hydrogen pressure, and concentration of defects in the oxide substrate [3, 4]. After the induction period, the reaction proceeds autocatalytically with an apparent activation energy of 60.61 kJ mol-1 for the reaction: CuO ? Cu, which is easier than Cu2O ? Cu. The activation energy for the later reaction is 114.53 kJ mol-1 [3]. Recently, some studies were made using low-temperature hydrogen plasma for reduction of iron oxide [9–11] and cobalt oxide [12]. In both the cases, successful reduction took place with a decrease in activation energy due to the presence of activated species in hydrogen plasma. Further, it is argued based on thermodynamic data that the equilibrium partial pressure of atomic hydrogen needed to reduce these oxides decreases with a decrease in temperature unlike that for the reduction with molecular hydrogen. Activated species present in hydrogen plasma are also known to remove kinetic barriers during the reaction [13]. The underlying question, whether hydrogen plasma having atomic H or other active hydrogen species can reduce the induction period in the CuO reduction, remains unresolved. Therefore, reduction of CuO with low-temperature hydrogen plasma was chosen as the topic for this study. The low-temperature hydrogen plasma used for reduction consists of various species such as H2 molecules, atomic hydrogen (H), ionic hydrogen (H?) and excited hydrogen molecules (H2*). Accordingly, the reduction may proceed through the following reactions steps: 4CuO þ H2 =H2  =2H=2ðHþ þ e Þ ¼ Cu4 O3 þ H2 O

ð1Þ

Cu4 O3 þ H2 =H2  =2H=2ðHþ þ e Þ ¼ 2Cu2 O þ H2 O

ð2Þ

2½Cu2 O þ H2 =H2  =2H=2ðHþ þ e Þ ¼ 2Cu þ H2 O

ð3Þ

The overall reaction is 4½CuO þ H2 =H2  =2H=2ðHþ þ e Þ ¼ Cu þ H2 O

ð4Þ

The Ellingham diagram (Fig. 1) and the Bauer–Glaessner diagram (Fig. 2) are drawn using the following reactions:

123

2Cu2 O þ O2 ¼ 4CuO

ð5Þ

4Cu þ O2 ¼ 2Cu2 O

ð6Þ

2H2 þ O2 ¼ 2H2 O

ð7Þ

4H þ O2 ¼ 2H2 O

ð8Þ

4ðHþ þ e Þ þ O2 ¼ 2H2 O

ð9Þ

Plasma Chem Plasma Process

Fig. 1 Ellingham diagram for copper–copper oxides and H2–H2O

Fig. 2 Bauer–Glaessner diagram of variation in hydrogen partial pressure with temperature

Ellingham diagram as in Fig. 1 shows the favorable positioning of copper oxides with respect to hydrogen for reduction. The remarkable fact is that the atomic hydrogen line is very much below (more than 1000 kJ at all temperatures) than the lines for both the copper oxides. This indicates that atomic hydrogen is a very strong reductant. The strength of ionic hydrogen as a reductant is still much more. The lines for the excited hydrogen molecules is expected to lie between the molecular and atomic hydrogen lines suggesting that it is a stronger reductant than molecular hydrogen. It may be mentioned that the Ellingham

123

Plasma Chem Plasma Process

diagram shows the standard Gibbs free energies of formation (DG°) of oxides as a function of temperature (T) and the associated equilibrium constant (K). It is mathematically written as DG° = - RTlnK. Considering K = pH2O/pH2 in the case of hydrogen reduction and putting this value in the DG° expression and after simplifying, the Bauer–Glaessner [B–G] equilibrium diagram can be worked out [12, 13]. The diagram shows the equilibrium relationship between the partial pressure of molecular hydrogen or atomic hydrogen, and the temperature for a given reaction. The thermodynamic data used for these diagrams have been taken from various sources [14–17]. The B–G diagram showing equilibrium of various copper oxides is redrawn for copper oxide system in Fig. 2. Figure 2 gives two other important inferences concerning the reduction of copper oxides: first, the equilibrium partial pressure of H2/H required to reduce cuprous oxide to metallic copper is lower than that to reduce cupric oxide to cuprous oxide at all temperatures, and second, as temperature goes down the equilibrium partial pressure needed for reduction also goes down, and this decrease is much sharper for atomic hydrogen than for molecular hydrogen. The excited molecular species are expected to come in between the molecular and atomic hydrogen. Thus, a lower temperature is more favorable for reduction by atomic hydrogen, which is highly significant for reduction of copper oxides in lowtemperature hydrogen plasma.

Experimental Details The reduction experiments were carried out in a reactor fabricated and supplied by IMAT Pvt. Ltd., India. The schematic of the reactor chamber is shown in Fig. 3. The reactor comprises two parts: (a) the reactor chamber, where plasma is created and reduction takes place, and (b) the operation rack, which controls the operating parameters of the experiment. The reactor chamber is made of aluminum alloy (Al7035) and it is cylindrical in shape with approximately 0.4 m diameter and 0.1 m height. The reactor chamber is double walled where chilled water flows between the walls so that the surface temperature of the chamber is maintained at room temperature. A sample holder made of molybdenum is located at the center of the reaction chamber. It is placed on a water-cooled copper plate, which cools the molybdenum holder. Thus, the molybdenum holder is not heated during reduction. As shown in Fig. 3, a quartz ring is provided surrounding the copper plate which basically prevents the plasma from hitting the inner walls of the reactor chamber. In any given experiment, the sample is placed on the molybdenum sample holder, which lies within the plasma region. The sample is essentially a pellet measuring approximately as 20 9 10-3 m in diameter and 2 9 10-3 m in height. The chamber is closed after placing Fig. 3 Schematic of the reactor chamber

123

Plasma Chem Plasma Process

the sample on the molybdenum sample holder. Then the chamber is evacuated and a base pressure of 0.1 Pa is maintained inside the chamber with the help of a rotary pump. The operation rack which lies below the reactor chamber incorporates a microwave generator having a power supply up to 6000 W. A magnetron is used to generate microwaves of a standard frequency of 2.45 9 109 s-1 which is guided by a waveguide to the chamber. In the chamber, the high-frequency microwaves interact with the incoming hydrogen gas (99.999 % pure) supplied to the reactor to produce the hydrogen plasma. The plasma plume covers the entire surface of the pellet where reduction takes place. Ocean Optics Spectrasuite USB Spectrometer (USB4000) was used for diagnosing the plasma composition. The optical emission spectra (OES) were recorded by the optical fiber from USB4000, placed in an optical view port located at the top of the chamber. The spectra were detected in the range from 300 9 10-9 to 350 9 10-9 m at regular intervals of 60 s. The CuO powder used in this study was sourced from J.T. Baker Ltd. The powder was subjected to phase analysis using an X’Pert PRO-PANalytical X-ray diffraction (XRD) system. The XRD of CuO is presented in Fig. 4. It was found by XRD analysis that the major constituent present in the sample is CuO with traces of Cu4O3 as per the JCPDS File No. 98-002-8583 and JCPDS File No. 98-006-2016 respectively. Several pellets were made out of CuO powder and then one by one they were subjected to plasma reduction in the chamber. The starting time of reduction is noted after the plasma is struck and all the parameters are stabilized. The parameters include hydrogen flow rate, chamber pressure, microwave power, and interface temperature. These were constantly monitored throughout the duration of the experiments. After each experiment, the weight loss during reduction was measured with a digital weighing balance with accuracy 0.1 9 10-6 kg. The reduced samples were analyzed by XRD to determine the phases.

Results and Discussion Effect of Hydrogen Flow Rate Figure 1, 2 indicate quite a favorable situation for reduction of CuO with hydrogen. So the experiments on the reduction of cupric oxide were conducted at relatively lower hydrogen

Fig. 4 XRD of raw material of cupric oxide

123

Plasma Chem Plasma Process Table 1 Experimental results of reduction of cupric oxide by hydrogen plasma R/10-6 m3 s-1, P/W, T/s

Pressure/ 103 Pa

Temp/K

Initial wt/ 10-3 kg

Final wt/ 10-3 kg

Wt. Loss/ 10-3 kg

0.833, 600, 600

0.666–0.853

743–843

2.272

2.222

0.050

10.9

0.833, 600, 1020

0.666–1.133

743–763

2.301

2.251

0.049

10.7

0.833, 600, 1920

0.666–0.933

743–983

2.261

2.034

0.227

50.0

0.833, 600, 2700

0.666–1.333

743–973

2.108

1.746

0.361

85.2

0.833, 600, 3720

0.666–1.333

743–973

2.362

1.923

0.438

92.2

1.000, 600, 300

0.666–1.533

743–818

2.452

2.425

0.026

5.3

1.000, 600, 600

0.666–2.213

743–823

2.442

2.371

0.071

14.4

1.000, 600, 1800

0.666–2.466

743

2.311

1.913

0.397

85.6

1.000, 600, 2700

0.666–2.533

743–1023

2.469

1.979

0.489

98.5

1.000, 600, 3600

0.666–1.893

743–1023

1.817

1.450

0.366

100.2

1.333, 600, 300

0.666–1.066

743–973

2.41

2.370

0.039

8.2

1.333, 600, 600

0.666–1.333

743–1073

2.524

2.397

0.127

25.0

1.333, 600, 900

0.666–1.226

743–1073

2.435

2.161

0.274

55.9

1.333, 600, 1200

0.666–1.466

743–1173

2.454

2.026

0.428

86.7

1.333, 600, 1800

0.666–1.533

743–1073

2.441

1.974

0.467

95.1

1.333, 600, 3600

0.666–1.533

743–1173

2.541

2.029

0.511

100.1

2.500, 750, 300

0.666–1.999

743–998

2.370

2.279

0.091

19.1

2.500, 750, 600

0.666–2.399

743–1098

2.430

2.153

0.276

56.6

2.500, 750, 900

0.666–2.533

743–1173

2.434

1.971

0.462

94.5

2.500, 750, 1200

0.666–2.399

743–1173

2.490

1.978

0.511

102.1

2.500, 750, 1800

0.666–2.533

743–1173

2.433

1.931

0.501

102.4

The first column shows H2 Flow rate (R) in 10

-6

3

Reduction, in %

-1

m s , Microwave Power (P) in W, Time (T) in s

flow rates of 0.833 9 10-6 to 2.5 9 10-6 m3 s-1 and at a microwave power of 600–750 W. The experimental results are shown in Table 1. The pressure and temperature build up as well as reduction level versus time plots are shown in Fig. 5a, b. The pressure built up was within 0.666 9 103–2.533 9 103 Pa and the interface temperature range was 743–1173 K. It may be noted that both the pressure as well as the interface temperature increase with an increase in hydrogen flow rate and microwave power. Commensurate to the indications obtained from the Ellingham diagram and Bauer Glaessner diagram, good reduction of cupric oxide was obtained. The rate plots (Fig. 6a–d) show ‘‘S’’ type plots with an induction period in the beginning followed by autocatalytic reduction—a similar observation reported by other authors for reduction of copper oxides by hydrogen gas [3, 4, 7]. However, unexpectedly, the initial induction period persisted even in the presence of excited hydrogen species as the reductant. It is observed from Fig. 6a–d, that the induction period apparently extends for few seconds. As evident from the figures, the induction period gets shortened at a higher pressure of hydrogen or higher temperature. The induction period is highest for lower pressure and lower interface temperature as shown in Fig. 6a and it is lowest in Fig. 6d due to higher hydrogen pressure and temperature. After this induction period, the reduction becomes autocatalytic and moves faster. The autocatalytic reduction rate increases with increase in flow rate. The extent of reduction increases with increase in hydrogen flow rate, microwave power, and time. At hydrogen flow rate of 0.833 9 10-6 m3 s-1 and microwave power of 600 W and around 3720 s of elapsed time, the reduction level was found to be

123

Plasma Chem Plasma Process

Fig. 5 Variation of pressure (a), interface temperature (b), at various hydrogen flow rate, microwave power, and time

Fig. 6 Plots of percentage reduction versus time at various hydrogen flow rates and microwave power (a) 0.833 9 10-6 m3 s-1 and 600 W, (b) 1 9 10-6 m3 s-1 and 600 W, (c) 1.333 9 10-6 m3 s-1 and 600 W, (d) 2.5 9 10-6 m3 s-1 and 750 W

92.2 % that reached up to 94.5 % at 2.5 9 10-6 m3 s-1 hydrogen flow rate, 750 W microwave power, and 900 s of elapsed time. Figure 7a–d presents the XRD analysis of the reaction products at various stages of reduction. In the case of 0.833 9 10-6 m3 s-1 hydrogen flow rate and 600 W microwave power (Fig. 7a), after the first 600 s in the induction period, the phases present are CuO (JCPDS File No. 98-002-8583), Cu2O (JCPDS File No. 98-002-1472); traces of Cu4O3

123

Plasma Chem Plasma Process

Fig. 7 XRD pictures of the reduction product at various flow rates of hydrogen and microwave power a 0.833 9 10-6 m3 s-1 and 600 W, b 1 9 10-6 m3 s-1 and 600 W, c 1.333 9 10-6 m3 s-1 and 600 W, d 2.5 9 10-6 m3 s-1 and 750 W

(JCPDS File No. 98-006-2016). The peaks representing metallic Cu (JCPDS File No. 98-006-2672) were not observed by then. It was about 1920 s after the start of the experiment, the peaks for Cu metal became prominent along with the presence of all the three other oxide phases. However, the plots for the reaction products obtained at 2700 and 3720 s show Cu peaks prominently with a couple of CuO peaks with very low intensity. In the case of 1 9 10-6 m3 s-1 hydrogen flow rate at the same microwave power of 600 W (Fig. 7b), the results are similar i.e., after 600 s from the start, the phases of CuO and Cu2O can be observed without any peak corresponding to metallic Cu. After 1800 s, the peaks corresponding to Cu became prominent along with the other three species. The plots for the products obtained at 2700 and 3600 s show mostly Cu peaks with a few CuO peaks of very low intensity. Also, at 1.333 9 10-6 m3 s-1 of hydrogen flow rate and 600 W power (Fig. 7c), only the CuO and Cu2O are present during the induction period. The metallic copper appears after 600 s. When hydrogen flow rate is 2.5 9 10-6 m3 s-1, and the microwave power is 750 W, Cu peaks appear even by 300 s. As evident from the XRD figures, the predominant reduction reaction occurring during the initial stage (i.e. induction period) is CuO to Cu2O. The autocatalytic reaction follows the induction period in which metallic copper forms from Cu2O. All these results show that the reduction is taking place in a sequential manner: CuO ? Cu4O3 ? Cu2O ? Cu. Earlier reports [3, 4, 7] indicate that the reduction reaction is direct CuO ? Cu with gaseous hydrogen at atmospheric pressure although the sub-oxides appear when the reaction takes place under vacuum or at

123

Plasma Chem Plasma Process

high temperatures. In the present study, both vacuum and high temperatures are prevalent and hence the sequential reduction. Since the reduction is a multistep process, this is not straight-forward. The percentage reduction was calculated from the ratio of the weight loss of the sample during reduction and the stoichiometric amount of oxygen present in the sample before experiment.

Kinetics and Activation Energy The rate plots shown in Fig. 6 and the XRD results presented in Fig. 7 indicate that (i) there is an induction period followed by fast autocatalytic reaction and (ii) CuO gets reduced in a sequential mode: CuO ? Cu4O3 ? Cu2O ? Cu. This type of ‘‘S’’ plots with induction period is better explained by Avrami-Erofeev universal kinetic model, which is mostly used for reactions involving nucleation and grain growth [18]. This rate equation has already been used for hydrogen reduction of magnetite [19] and hematite [20]. The equation is written as: ½lnð1  aÞ1=n ¼ kapp  t

 or; ln½lnð1aÞ ¼ n  ln kapp þ n  ln(tÞ

ð10Þ

where, a is fractional reduction degree, t is time, kapp is the apparent rate constant, n is the avrami parameter. This equation has been derived independently by Johnson and Mehl, Avrami, and Erofeev [18]. It is a special case for n = 1 representing first order kinetics. Figure 8 presents plots of ln [-ln(1 - a)] versus ln(t) for the kinetic data generated from the experimental results of reduction of copper oxide (Table 1). The slope, ‘n’ obtained is between 2 and 3. As shown in Fig. 8, a slope of *3 was observed for 2.5 9 10-6 m3 s-1 and 750 W. The slope decreased with a decrease in flow rate and power. An attempt was made to estimate the activation energy of the process using rate plots presented in Figs. 5 and 6. The time for 40 % reduction (t0.4) was determined from the rate plot (Fig. 6), and the inverse of this parameter was used as the reaction rate to generate the Arrhenius plot (Fig. 9) which resulted in an activation energy value of 75.64 kJ mol-1. This value should be taken as the resultant activation energy of reactions 1 and 3 since both the oxide species, CuO, and Cu2O as well as metallic Cu are present according to the XRD data (Fig. 7).

Fig. 8 Rate plots for reduction of CuO following Avrami-Erofeev equation (a) 0.833 9 10-6 m3 s-1 and 600 W, (b) 1 9 10-6 m3 s-1 and 600 W, (c) 1.333 9 10-6 m3 s-1 and 600 W, (d) 2.5 9 10-6 m3 s-1 and 750 W

123

Plasma Chem Plasma Process

Fig. 9 Arrhenius Plot for Copper oxide from (a) 20 % reduction, (b) 40 % reduction, (c) 70 % reduction, (d) induction time, (e) ln (kapp)

Encouraged by the above results, another attempt was made to use t0.2 and t0.7 as rate parameters and to determine if reactions 1 and 3 could be dealt separately. However, both these parameters resulted in activation energy values (72.19 and 71.82 kJ mol-1, respectively) closer to 75.64 kJ mol-1. The inverse of induction period (1/tind) was also taken as the rate and was fitted into the Arrhenius plot that resulted in the activation energy of 75.63 kJ mol-1, which is closer to 75.64. Kim et al. [3] have reported activation energy value of 60.61 kJ mol-1 for reduction of CuO to metallic Cu by H2 and 114.53 kJ mol-1 for reduction of Cu2O to Cu by hydrogen. Thus, the activation energy obtained from the present study may be considered as a good match for the reduction of CuO directly to Cu by H2. However, the involvement of Cu2O in the process during fast autocatalytic reduction is apparent from the XRD plots (Fig. 7). Hence, it seems there is a lowering of activation energy (from 114.53 to 75.64 kJ mol-1) due to active species present in hydrogen plasma. From Fig. 8, the Y-intercept of the plots, ln (kapp) values were generated and fitted into the Arrhenius equation in Fig. 9 resulting in an activation energy value of 73.32 kJ mol-1. All these activation energy values may be considered close to 75.64 kJ mol-1 obtained by using reduction at t0.4 in the same figure, which corresponds to the constant growth portion of the reduction plots, where Cu2O to Cu reduction is predominant. Thus lowering of activation energy is for reduction of Cu2O to Cu can only be due to interaction between oxygen and some other hydrogen species of the plasma in addition to H2.

Reaction Mechanism The reduction reactions are basically represented by two processes: the first process involves the conversion of hydrogen gas into a plasma state and the second process involves reduction of CuO in a sequential mode i.e., CuO ? Cu4O3 ? Cu2O ? Cu. The reduction plot shows ‘‘S’’ type curve, which consists of induction period followed by fast autocatalytic reduction. The induction period is a main feature of the process. During the induction period, active sites are being formed that have a high efficiency for the adsorption and dissociation of H2 [3, 8, 21]. During this period, the CuO mainly converts to Cu2O with very little decrease in molar volume (5 %). This small decrease in molar

123

Plasma Chem Plasma Process

volume gives less space for the dissociated H to reach to the reduction interface. So the embedding of H into the bulk CuO doesn’t take place. Kim et al. [3] concluded that the embedding of H is negligible during the induction period. Though the active sites are formed during this period, the reduction rate remains low. The induction period is followed by autocatalytic reduction. During the autocatalytic reduction, metallic copper forms from the Cu2O which in turn was formed during the induction period. The decrease in molar volume associated with the transformation from Cu2O to Cu is 40 %. This high decrease in molar volume gives more space for penetration of H. During the autocatalytic reduction, it has also been reported that the H penetrates into the bulk CuO phase and then the removal of O becomes faster [4]. This embedding of H into the bulk CuO phase changes the crystal dimensions [3]. To establish this aspect, the crystal dimensions of the CuO used here is estimated from the XRD data. The parameters are (in angstrom) as follows: a = 4.694, b = 3.422, c = 5.119 for the original CuO sample used which was changed slightly to a = 4.651, b = 3.446, c = 5.197 for the reduced products obtained after 300 s at 2.5 9 10-6 m3 s-1 of hydrogen flow. This increase in the dimensions, specifically in b and c parameters, suggests embedding of H as reported by Kim et al. [3]. A careful look at the temperature profile (Fig. 5) reveals that between 240–420 and 720–840 s period some wide peaks are visible which correspond to the linear part of the fast reaction time in the rate plot (Fig. 7). This may be due to reactions involving the formation of metallic Cu that are exothermic [7]. This observation explains the autocatalytic reaction. The fast reduction during the autocatalytic period has been explained by the high efficiency of adsorption and dissociation of H2 at the active sites formed during induction period [3, 8, 19–22]. These active sites support the rapid dissociation of H2 at the surface, during the autocatalytic reduction [3, 4, 8]. As evident from the XRD analysis, the predominant reduction during the autocatalytic period is Cu2O to Cu, which leads to a significant decrease in molar volume (40 %). This decrease in molar volume gives more space for the migration of H into the bulk of the oxide. The penetration of H to the reduction interface becomes significant. Once a significant amount of H is available in the oxide sample, the removal of oxygen as water becomes faster, as detected by mass spectroscopy. This gives rise to a faster reduction during autocatalytic period [3]. From their Rietveld analysis of the XRD data, the difference electron density maps were obtained which showed extra electron density at a position consistent with H in a hydroxo group. They concluded that the hydroxo-like species could be the precursors for the removal of O atoms as water. They proposed neutron diffraction to confirm the formation of hydroxo like group. However, the hydroxo group (OH band) has been observed in several other studies in the wavelength range of 306–322 nm [23]. So, the optical emission spectra were obtained and analyzed during the reduction study.

Optical Emission Spectra The OES spectra at different times within the induction period, for the experiments carried out at 0.833 9 10-6 m3 s-1 and 600 W was obtained. These plots are shown in Fig. 10. As evident from Fig. 10, there are two broad peaks at the wavelength of 307 and 309 nm, that are characteristics of the presence of the OH spectrum. This sort of spectrum is frequently observed in many kinds of flames and hot gases containing oxygen and hydrogen [24, 25]. The OH spectrum is also observed in the range of wavelength of 306–320 nm [26] and 306–322 nm [23]. These features are clearly seen in Fig. 10, which has been taken during the induction period. As reported earlier, there is formation of active sites during the induction and these active sites have a high efficiency for the adsorption

123

Plasma Chem Plasma Process

Fig. 10 The OES spectra of the reduction process

and dissociation of H2 [3, 8, 21]. Also, it has been reported that hydroxo-like species are probably formed during the induction period, which could be the precursors for removal of O as water [3]. All these findings point to the fact that formation of OH takes place during the induction period. As evident from our XRD results, the reduction occurs in sequence CuO ? Cu4O3 ? Cu2O ? Cu. Out of which the Cu4O3 is a metastable state. So, the presence of Cu4O3 is negligible in the XRD figures. During the induction period, there is no formation of metallic copper. During this period, there is only the formation of Cu2O from CuO. It may be noted that only around 10 % reduction occurs during the induction period, indicating that minor reduction progresses during this period. The formation of Cu2O from CuO gives rise to a decrease in molar volume by 5 %. This decrease in molar volume is not good enough for fast reduction occurred during the autocatalytic reduction period. Kim et al. [3] studied the reduction of CuO with molecular hydrogen. During the induction period, they proposed the formation of active sites, which have high efficiency for adsorption of H2. Our findings of OH formation during induction period supports the fact that this formation of OH, give rise to the active sites, which may be responsible for the fast autocatalytic reduction.

Microwave Power Density In order to get a perspective on the effect of operating parameters on conversion of hydrogen gas to its plasma state and its impact on the reduction process, a mention about average microwave power density (MWPD) is warranted [27, 28]. The average power density is regulated through careful monitoring of microwave power and pressure. In the present study, the pressure and temperature reported in Table 1 and Fig. 5 fall within a narrow range of 0.666 9 103–2.533 9 103 Pa and 743–1173 K, respectively. The estimated MWPD was within a narrow range of 4 9 106 to 10 9 106 W m-3. Thus, the position of hydrogen plasma is expected to be similar to that in the case of cobalt reduction [12]. In actual terms, the expected mole fraction of atomic hydrogen should be in the range

123

Plasma Chem Plasma Process

of 0.01–0.02 [27] and the dominant hydrogen species may be vibrationally excited hydrogen molecules. The H2 gas temperature (Tg) range may be between 1600 to 2000 K and the H2 vibration temperature (Tv) may vary between 1800 to 2200 K; the difference between these temperatures being only 200 K, the gas and vibration temperatures are practically in quasi-equilibrium. The plasma (gas) would transfer energy as well as the reductant species while interacting with the oxide substrate. The data presented in Fig. 1 is indicative enough that molecular hydrogen is sufficient to cause reduction of CuO and Cu2O in the chosen temperature range. This has been demonstrated in previous studies by various workers [3, 4, 7]. The vibrationally excited H2 molecules (m = 0–4), which are stronger reductants than molecular hydrogen are expected to play a significant role in the reduction process. The rate of emergence of vibrationally excited hydrogen species from the plasma and appearance at the copper oxide plasma interface may be significant.

Conclusions The following conclusions may be drawn from the present investigation on the reduction of copper oxide in hydrogen plasma: (i)

(ii)

(iii)

(iv)

(v)

The free energy values indicate that molecular hydrogen is a good reductant for reducing copper oxides. However, the H and H? species present in hydrogen plasma can be much stronger reductants. The equilibrium partial pressure of hydrogen required for reducing copper oxides decreases with decrease in temperature implying possibility of low-temperature reduction. The decrease is much sharper for H, in comparison to H2. Reduction of copper oxide encountered an initial induction period during which the OH band was observed. This supports the earlier findings of the formation of active sites for adsorption and dissociation of H2. This initial phase was followed by a fast autocatalytic reduction to metallic copper. The overall reduction reaction followed sequential steps through the formation of sub-oxides. The reduction reaction followed an ‘S’ shaped kinetic behavior which could be explained by applying Avrami-Erofeev equation resulting in a slope of nearly 3. The slope decreased with decrease in pressure, temperature, and microwave power. The activation energy of the process was found to be 75.64 kJ mol-1. It was observed that most of the reduction takes place mainly through the involvement of reduction of Cu2O to Cu, which has high activation energy. Thus the lowering of activation energy value as measured may be considered due to plasma processing.

References 1. Li J, Mayer JW, Tu KN (1992) Phys Rev B 45:5683 2. Henrich VE, Cox PA (1994) The surface science of metal oxides. Cambridge University Press, Cambridge 3. Kim JY, Rodriguez JA, Hanson JC, Frenkel AI, Lee PL (2003) J Am Chem Soc 125(35):10684–10692 4. Rodriguez JA, Kim JY, Hanson JC, Pe´rez M, Frenkel AI (2003) Catal Lett 85(3–4):247–254 5. Sawada Y, Tamaruz H, Kogomax M, Kawasek M, Hashimoto K (1996) J Phys D Appl Phys 29:2539–2544

123

Plasma Chem Plasma Process 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

Fleisch TH, Mains GJ (1982) Appl Surf Sci 10(1):51–62 Yamukyan MH, Manukyan KV, Kharatyan SL (2009) J Alloys Compd 473(1):546–549 Rodriguez JA, Hanson JC, Frenkel AI, Kim JY, Pe´rez M (2002) J Am Chem Soc 124(2):346–354 Rajput P, Bhoi B, Sahoo S, Paramguru RK, Mishra BK (2013) Ironmak Steelmak 40(1):61–68 Rajput P, Sabat KC, Paramguru RK, Bhoi B, Mishra BK (2014) Ironmak Steelmak 41(10):721–731 Bhoi B, Mishra BK, Paramguru RK, Pradhan SK, Mukherjee PS, Sahoo S, Das SK (2014) US Pat 8728195 Sabat KC, Paramguru RK, Pradhan S, Mishra BK (2015) Plasma Chem Plasma Process 35:387–399 Sabat KC, Rajput P, Paramguru RK, Bhoi B, Mishra BK (2014) Plasma Chem Plasma Process 34(1):1–23 Kubaschewski O, Alcock CB (1979) Metallurgical thermochemistry. Pergamon Press, Newyork Chase MW (1998) NIST-JANAF Thermochemical Tables, J Phys Chem Ref Data Monogr 9. American Institute of Physics, USA Turkdogan ET (1980) Physical chemistry of high temperature technology. Academic Press, New York Lee HG (1999) Chemical thermodynamics for metals and materials. Imperial College Press, London Piotrowski K, Mondal K, Wiltowski T, Dydo P, Rizeg G (2007) Chem Eng J 131(1):73–82 Wang H, Sohn HY (2013) Metall Mater Trans 44B(1):133–145 Chen F, Mohassab Y, Jiang T, Sohn HY (2015) Metall Mater Trans 46B(3):1133–1145 Chen JG, Fischer DA, Hardenbergh JH, Hall RB (1992) Surf Sci 279(1):13–22 Wang HP, Yeh CT (1983) J Chin Chem Soc 30(2):139–143 Parigger CG, Guan G, Hornkohl JO (2003) App Opt 42(30):5986–5991 Dieke GH, Crosswhite HM (1962) J Quant Spectrosc Radiat Transf 2(2):97–199 De Izarra C (2000) J Phys D Appl Phys 33(14):1697–1704 Wilson EW, Mackey JE, Keller BD, Goertzen EJ, Clements SA, Rivenbark CD, Cox C (2005) OH emission spectra of hybrid rocket motors using PMMA. AIAA Paper 3905. doi:10.2514/6.2005-3905 Hassouni K, Gicquel A, Capitelli M, Loureiro J (1999) Plasma Sour Sci Technol 8(3):494 Hassouni K, Grotjohn TA, Gicquel A (1999) J Appl Phys 86(1):134–151

123