Selective Catalytic Oxidation of Ammonia to Nitrogen on CuO-CeO 2 Bimetallic Oxide Catalysts

This study addresses the performance of the selective catalytic oxidation (SCO) of ammonia to N2 over a CuO-CeO2 bimetallic oxide catalyst in a tubular fixed-bed reactor (TFBR) at temperatures from 423 to 673 K in the presence of oxygen. CuO-CeO2 bimetallic oxide catalyst was prepared by co-precipitation with Cu(NO3)2 and Ce(NO3)3 at various molar concentrations. This study tested operational stability and investigated how the influent NH3 concentration (C0 = 500-1000 ppm) influences the capacity to remove NH3. The catalysts were characterized using XRD, FTIR, PSA, SEM and EDX. Ammonia was removed by oxidation in the absence of CuOCeO2 bimetallic oxide catalyst, and the formation of copper (II) and cerium (IV) oxide active sites was confirmed. Additionally, the effects of the NH3 content of the carrier gas on the catalyst’s reaction rate (r) were observed. The results revealed that the extent of catalytic oxidation of ammonia in the presence of a CuO-CeO2 bimetallic oxide catalyst was a function of the molar ratio Cu:Ce in the bimetallic catalyst. The kinetics of catalyzed NH3 oxidation are described using the rate expression of the Eley-Rideal kinetic model. Also, experimental results indicate a reasonable mechanism for the catalytic oxidation of ammonia.


INTRODUCTION
Extensively adopted into industrial processes, ammonia (NH 3 ) can be eliminated in numerous ways.NH 3 is used in the ammonium nitrate and nitric acid production industry, livestock feedlots, urea manufacturing plants, the nitrogen fertilizer application industry, fossil-fuel combustion and petroleum refineries, as well as the refrigeration industry.Ammonia is a toxic inorganic gas with a pungent odor under ambient conditions, and is potentially harmful to public health (Mojtahedi and Abbasian, 1995;Giroux et al., 1997;Kurvits and Matra, 1998;Amblard et al., 2000;Hung et al., 2003;Sotiropoulou et al., 2004).Conventional biological, physical and chemical treatments, including biofilters (Kim et al., 2000;Chung et al., 2001), stripping (Huang et al., 2000), scrubbing with water (Burch and Southward, 2001), post-combustion control (Hasegawa and Sato, 1998), microwave-plasma discharge (Wójtowicz et al., 2000), electrochemical oxidation (de Vooys et al., 2001) and the use of activated carbon fibers (ACFs) for soot adsorption (Mangun et al., 1999;Muenter and Koehler, 2000), only trigger a phase transformation and may yield a contaminated sludge and/or an adsorbent, both of which require further treatment.The maintenance and operating costs associated with these physical and/or chemical methods are high.
Therefore, the removal and the control and prevention, of the emission of ammonia emission from air and waste streams are important.Discharges present challenges because environmental laws and regulations on safe discharge are becoming increasingly strict.
More recently, catalytic oxidation has been established to increase the effectiveness of AOP (Advanced Oxidation Processes) technology using dedicated catalysts, which potentially shorten the reaction times of oxidation, and allow it to proceed under milder operating conditions.The selective catalytic oxidation (SCO) of ammonia in a stream to molecular nitrogen and water is one method for solving problems of ammonia pollution (Wöllner et al., 1993;Schmidt-Szałowski et al., 1998;Wang et al., 1999;Curtin et al., 2000;Lietti et al., 2000;Amblard et al., 2000;Escandón et al., 2002;Dravell et al., 2003).The catalytic oxidation of ammonia has been reported to proceed as follows; The SCO process that involves ammonia should be selective for nitrogen (reaction 1) and prevent further oxidation of nitrogen (reactions 2 and 3).Earlier work on ammonia oxidation was reviewed by Il'chenko (1975aIl'chenko ( , 1975bIl'chenko ( , 1976)), who focused on the reaction mechanism of ammonia oxidation, and compared catalytic activities.Few catalysts have been used in oxidizing ammonia in the gaseous phase.For instance, Amblard et al. (2000) demonstrated excellent selective conversion of ammonia to nitrogen (＞90%) by γ-Al 2 O 3 -supported Ni by selective catalytic oxidation.Moreover, Wang et al. (1999), who developed Ni-based catalysts for oxidizing fuel gas generated by gasifying biomass, found that fresh Ni-based catalysts were more active at lower temperatures in decomposing ammonia, and the partial pressure of hydrogen in the flue gas is a critical factor that governed ammonia oxidation.Liang (2000) studied the oxidation of ammonia in a fixed-bed microreactor in the temperature range 873-1023 K at GHSV=1800-3600 hr -1 .They found that the conversion of ammonia reached 98.7% and 99.8% Nitrogen compounds have been similarly examined and have been demonstrated reversibly to inhibit the catalyst (Richardson, 1989).However, the kinetics of the catalytic oxidation of NH 3 on metal composite catalysts have not been comprehensively examined.
Copper oxide is a highly active transition metal, and has been considered to be a potential substitute for noble metal-based emission control catalysts (Gang et al., 1999).Copper oxide on cerium oxide is known to constitute an efficient catalyst for various reactions, such as the combustion of CO and the water-gas-shift (WGS) reaction (Sedmak et al., 2004;Fu et al., 2003).
Furthermore, the most important washcoat component of a catalyst is cerium, which is added to the washcoat as a stabilizer and an oxygen-storage component.Cerium stabilizes the washcoat layer improves the thermal resistance enhances the catalytic activity of precious metals, and provides oxygen storage/release capacity (OSC) (Golunski et al., 1995;Petryk and Kolakowska, 2000;Wey et al., 2002;González-Velasco et al., 2003).The interaction between copper oxide and the cerium is complex, because various copper-cerium interactions can result in synergistic effects, enhancing catalytic characteristics (Skårman et al., 2002).However, little work has been performed on the use of CuO-CeO 2 bimetallic oxide catalyst to evaluate the reactive characteristics of these active metals in selective catalytic oxidation.Therefore, given the low cost of copper, this study considers the effect of the CuO-CeO 2 bimetallic oxide catalyst system on the oxidation of ammonia-containing gas streams with various parameters, and the kinetics of the removal of ammonia from the gas-phase in an SCO process.X-ray powder diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), SEM, energy-dispersive X-ray spectrometry (EDX) and a particle size analyzer (PSA), were all employed to characterize the CuO-CeO 2 bimetallic oxide catalyst.

Preparing CuO-CeO 2 Bimetallic Oxide Catalyst
CuO-CeO 2 bimetallic oxide catalysts were prepared by co-precipitating copper (II) nitrate (GR grade, Merck, Darmstadt, Germany) with cerium (III) nitrate (GR grade, Merck, Darmstadt, Germany) using K 2 CO 3 (0.2 M) at four molar ratios: 6:4, 7:3, 8:2 and 9:1.The pH of the coprecipitating aqueous solution was 11.5 ± 0.2.After filtration, the co-precipitating aqueous solution was washed five times with deionized water, and dried at 393 K.These compounds were then calcined at 773 K in an air stream for 4 hours.The powder thus produced was formed into tablets using acetic acid as a binder.The tablets were later reheated at 573 K to burn the binder out of the CuO-CeO 2 bimetallic oxide catalyst.They were then crushed and sieved into particles of various sizes from 0.15 and 0.25 mm, for later use.

Characterizing the Solid Phase
X-ray diffractograms were obtained using a Diano-8536 diffractometer with CuKα radiation as the source.During analysis, the sample was scanned from 20 to 80° at a rate of 0.4°/min.Diffuse reflectance FTIR spectra of species adsorbed on the catalyst were measured at room temperature using a Bruker Vector 22 FTIR spectrometer, equipped with a diffuse reflectance attachment with a resolution of 4 cm -1 (Bruker, Germany).The changes in the sizes of the catalytic particles were measured using a laser light-scattering particle size analyzer (PSA, Coulter LS100, USA).
Scanning electron microscopy, using an energy-dispersive X-ray spectrometer (SEM/EDX, JEOL, JSM-6400, Kevex, DeltaII), yielded the morphology of the catalysts and provided information on the distribution of copper and cerium on the surfaces of the catalysts.

Reaction System
Experiments were conducted on a tubular fixed-bed flow quartz reactor (TFBR).Two flowing gases, NH 3 and O 2 , were used to prepare the feed mixture in the diluting gas, helium, which flowed into the inlet of the reactor.A mass flow regulator was used to control independently the flows of ammonia and oxygen.Extremely pure helium was used as a carrier gas at a flow rate from 8 to 13 L/min, controlled using a mass flow meter (830 Series Side-Trak TM , Sierra, Monterey, CA, USA).The mass of each catalyst was 1 g and the empty bed volume was approximately 1.2 cm 3 .An inert material formed from (hydrophilic and inert) γ-Al 2 O 3 spheres was used to increase the interfacial area between the solid and the gas phase to increase the mass transfer of ammonia from gaseous streams.This approach resembled that of Huang (2000), who conducted experiments on the catalytic oxidation of ammonia.A reaction tube with a length of 300 mm and an inner diameter of 28 mm was placed inside a split tube furnace.The tube that contained the catalyst was placed in the same furnace.The temperature was measured using two type-K thermocouples (KT-110, Kirter, Kaohsiung, Taiwan), each with a diameter of 0.5 mm, these were located in front of and behind the catalytic bed.The thermocouples were also connected to a PID controller (FP21, Shimaaen, Tokyo, Japan) to maintain the temperature in the tube within ±0.5%.The concentration of the feed gas (GHSV, 92,000 ml/h-g) was maintained at 1,000 ppm NH 3 and the O 2 concentration was 4%.The catalyst was not deactivated during testing.

Analyses
Before and after the reaction, samples were automatically injected through a sampling valve into a gas chromatograph (Shimadzu GC-14A), equipped with a thermal conductivity detector.A stainless-steel column (Porapak Q 80/100mesh) was used to separate and determine the concentrations of N 2 O isothermally at 100 ℃ .The areas associated with the signals were electronically measured using a data integrator (CR-6A, Shimadzu, Kyoto, Japan).Dilute sulfuric acid was used to scrub the residual ammonia in the vapor gas and the amount present was measured using a Merck kit (Merck, Spectroquant Vega 400, Darmstadt, Germany).The concentrations of NO, NO 2 ,, and O 2 in the gas samples were monitored continuously during combustion at a particular location, using a portable flue gas analyzer (IMR-3000, Neckarsulm, Germany).Data were collected when the SCO reaction was in a steady state, typically after 20 min at each temperature.Each temperature was maintained for 90 min to allow the system to enter a steady state.Most experiments were repeated once to ensure reproducibility, and similar results were always obtained.

RESULTS AND DISCUSSION
Characterizing CuO-CeO 2 Bimetallic Oxide Catalyst Fig. 2 presents the X-ray diffraction (XRD) patterns obtained at various copper-cerium molar ratios, verifying that CuO and CeO 2 are the dominant phases of the copper-cerium composite catalyst.Dominant CuO diffraction peaks were observed at around 2θ=35. 5, 38.8, 48.7, 53.4, 61.4, 66.3 and 68.0 o when a copper-cerium composite catalyst was used.This result is similar to that obtained by Ning (2001).Moreover, CeO 2 diffraction peaks were present at approximately 2θ=28.5, 33.1, 47.5 and 56.7 o when a copper-cerium composite catalyst was used.This result is similar to that reported by Hashimoto (2000).An earlier investigation demonstrated that CeO 2 is the most active phase in a catalytic reaction, because it has been demonstrated to be a strong promoter of an oxygen-storage medium, whenever noble metals are used as major catalysts (Golunski et al., 1995;Petryk and Kolakowska, 2000).Cerium dioxide in a copper catalyst can be assumed to promote the formation of the active phase of CuO under the conditions of ammonia oxidation.Additionally, CeO 2 -based materials can act as oxygen buffers by storing/releasing O 2 , because the automotive three-way catalytic converter contains the cerium (III)-cerium (IV) redox couple (Kašpar et al., 1999).Thus, X-ray powder diffraction confirmed the formation of copper (II) and cerium (IV) oxide active sites, synergistically affecting the Cu/Ce cooperation in the redox mechanism.Fig. 3 compares the FTIR spectra at various copper-cerium molar ratios, and also confirms the presence of the CuO-like phase and the CeO 2 phase on the surface of the copper-cerium composite catalyst.Fig. 3 reveals that the peaks associated with the CuO-like phase on the framework are associated with a peak at around 1,384 cm -1 (Sadykov et al., 2001;Hadjiivanov, 2000).Fig. 3 also shows that the CeO 2 phase on the framework is associated with peaks at around 1,630 cm -1 , 1,524 cm -1 and 1,380 cm -1 (Centi and Perathoner, 1998).CuO and CeO 2 are normally accepted to exhibit great synergy effect when they are both prepared as a composite CuO/CeO 2 catalyst (Sedmak et al., 2004).Therefore, we suggest here that the catalytic activity of the CuO-CeO 2 bimetallic oxide catalyst system in oxidizing ammonia may be explained by the reversible redox behavior of CuO/CeO 2 couples in promoting the bifunctional mechanism.
The change in the sizes of particles of the catalyst was determined using the laser lightscattering method, as depicted in Fig. 4. The mean particle size converged to approximately 14.3, 14.6, 10.3, and 7.7 μm, for fresh copper-cerium composite catalysts with the four molar ratios: 6:4, 7:3, 8:2 and 9:1.However, the diameters of the catalyst decreased as the copper content declined, indicating that co-precipitation aggregates the low metal molar ratio.

Oxidizing Ammonia
Catalysts with various copper and cerium contents were prepared, characterized and tested for their effectiveness in the SCO process.Fig. 5 plots the effects of composite catalysts with various copper-cerium molar ratios on the extents of conversion of NH 3 , in terms of the extent of removal during SCO.The copper-cerium molar ratio of the copper-cerium composite catalyst governs the extent of catalytic oxidation of NH 3 over it.The extents of conversion of NH 3 were 99.0% and 92.0% when the catalytic oxidation proceeded under particular operating conditions over coppercerium composite catalysts with molar ratios of 6:4 and 9:1, respectively.Notably, a higher cerium content was associated with a higher NH 3 conversion.The overall maximum selectivity of N 2 production varied from 27% to 86%, and that of NO production varied from 0% to 13% over the range 23-99% NH 3 conversion at NH 3 concentrations of 1,000 ppm over the copper-cerium (6:4) composite catalyst (Table 1).The only intermediate species was NO.The decrease in the nitrogen selectivity is primarily related to the formation of NO.This result is consistent with the findings of Gang (1999).Nitrogen gas is believed to be generated mostly by the dissociation of NO produced by the oxidation of adsorbed NH 3 (Bradely et al., 1995).Accordingly, this work considers the hypothesis that NH 3 and oxygen may be adsorbed to specific sites on the coppercerium (6:4) composite catalyst, accelerating the conversion of NH 3 to nitrogen.Based on these results, nitrogen was the dominant gas and a small amount of NO was detected in the resultant stream.This result is similar to those of Li (1997) and Curtin (2000).Fig. 7 indicates that disaggregated and dispersed phases were formed when the surface of the catalyst was aged or when poisoning was caused by plugging, indicating that the porosity of the particles had changed.These crystal phases may be responsible for the high activity of the catalysts.The findings also confirm that the dispersion phenomena of the catalyst increased the efficiency of the removal of NH 3 .The elemental distribution proportional to these particles was homogeneous, as shown by the electron backscattering in SEM micrographs (Figs. 6 and 7).
Furthermore, elemental dot mapping photographs reveal that, after the elemental composition of the surfaces of the test catalysts had been activated, the metal content of the copper-cerium composite catalyst was slightly changed.Copper was present at an elemental percentage of 62% to 58% and cerium was present at an elemental percentage of 41% to 36% of the catalyst.(c) Ce.Test conditions : 1,000 ppm NH 3 in He, O 2 =4%, R.H.=12%, GHSV=92,000 ml/h-g.

Reaction Kinetics
The effect of the NH 3 content of the carrier gas on the reaction rates (r) associated with the copper-cerium composite catalyst was considered.The reaction rate (r) of NH 3 is defined as follows.
( ) ( ) where ν represents the volume flow rate (ml/s) of NH 3 ; V is the reaction volume (ml), and C i0 and C i are the concentrations of NH 3 at the inlet and the outlet, respectively.The ratio F/V specifies the space velocity.X i and X i0 represent the conversion of NH 3 at the inlet and the outlet, respectively.Fig. 8 presents the effects of NH 3 concentration and reaction temperature on the reaction rate over the CuO-CeO 2 (6:4) bimetallic oxide catalyst.The reaction rate (r) increases almost linearly with the NH 3 content at various temperatures.The reaction rates also increases with temperature above 423 K.The results also reveal that the CuO-CeO 2 (6:4) bimetallic oxide catalyst is very active in ammonia decomposition.Therefore, the extent removal of ammonia increased markedly with the temperature of the influent stream because the retention time during the catalytic process is reduced.The catalytic effect on the reduction of ammonia reduction was insignificant in the influent stream with ammonia concentrations from 500 to 1,000 ppm.
Increasing the ammonia concentration to 500 ppm dramatically reduced the extent of the removal of ammonia.Fig. 8. Effect of various ammonia concentrations and different temperatures with reaction rates on the CuO-CeO 2 (6:4) bimetallic oxide catalyst for the conversion of NH 3 .Test conditions: 500-1,000 ppm NH 3 in He, O 2 =4%, R.H.=12%, GHSV=92,000 ml/h-g.Fig. 9 plots ln (1-X) versus time at various reaction temperatures, using a CuO-CeO 2 (6:4) bimetallic oxide catalyst.A higher space time resulted in lower catalytic efficiency at a particular temperature.The plot is of a linear relationship between ln (1-X) and space time, and supports the fact that the reaction is of first order.Previous studies by Huang (2000) and Long (2001) have found a first order dependency of the rate on NH 3 concentration.The reaction rate constants obtained from various kinetic models normally vary with temperature.The Arrhenius Equation can be determined using a typical method for calculating the activation energy of the reaction, where t is defined as follows.
( ) where A is the frequency factor (sec -1 ); Ea is the activation energy (kcal/mol); R is the ideal gas constant, and T represents the absolute temperature (K).Therefore, the reaction order can be determined from the slopes of the lines in Fig. 9.The linearity of the plots reveals that the kinetics are approximately first-order of the NH 3 oxidation varies from 0.55 to 0.94.Moreover, the reaction constant k is determined from the intercept of each line.Also, the plot ln (1-X) versus space time in Fig. 10 is a straight line.The intercept and gradient and the Arrhenius relationships give the pre-exponential factor [A = 2.02×10 8 sec -1 ] and the apparent activation energy of the oxidation of NH 3 (E a = 14.8 kcal/mol).The model has been suggested to be appropriate for the oxidation of ammonia over CuO-CeO 2 bimetallic oxide catalyst, since the R 2 value (R 2 = 0.9285) for the fitting of the rate constants is relatively high.
The consistency of the results and first-order reaction kinetics supports the assertion that the catalytic reaction of NH 3 proceeds as follows.
In this mechanism, the above steps are assumed to be first-order in the respective gaseous species.
Reaction (6) describes the adsorption of oxygen molecules onto the surface active site of the catalyst.
The adsorbed oxygen molecule is decomposed and considered to be the active oxygen atom.In Eq. ( 7), the active oxygen atom reacts with the gaseous NH 3 molecule and yields the products.The reaction rates, 1 r and 2 r , are defined as follows.
( )  Therefore, a pseudo steady state condition is specified, where α is the number of oxygen molecules involved in the NH 3 oxidation.Classical kinetics and Eq. ( 10) establish the following rate expression.
Eq. ( 11) can be simplified as a linear function: The rate expression Eq. ( 12) refers to the oxidation NH 3 of with Cu/Ce bimetallic composite catalysts.It is determined by the Eley-Rideal (ER) kinetic expressions (Satterfield, 1991), which involve the invitation of oxygen to copper or cerium metal and the reaction of NH 3 in the gas phase with the adsorbed oxygen.When the adsorption reaction between the metal and the oxygen molecule proceeds, 1 k 12), as is discussed below. .The reaction mechanism presented herein was considered the theory to elucidate the reaction kinetics.
Hence, the kinetic model was derived by assuming that a reaction proceeds between the adsorbed oxygen and the gaseous ammonia (Eley-Rideal type mechanism), by the adsorption and reduction of the ammonia onto an oxidized copper or cerium surface.Such results are consistent with those of Zawadzki (2003) and Lee (2003).

CONCLUSIONS
A composite catalyst of copper with cerium promoted the removal of ammonia by oxidation.
The overall maximum selectivity of N 2 production varied from 27% to 86% and that of NO production varied from 0% to 13% over an NH 3 conversion range of 23% to 99% over a CuO-CeO 2 (6:4) bimetallic oxide catalyst.A kinetic rate expression was proposed to describe the data under the various conditions investigated.Moreover, correlation equations, in the form of Arrhenius's Law, were obtained from experimental data to specify the efficiencies of the removal of ammonia at various temperatures and residence periods in a catalytic process.The removal efficiency increases with the temperature and as the residence period is reduced.An Eley-Ridealtype kinetic model of the reaction between adsorbed oxygen and a gas-phase reactant molecule appears to represent NH 3 oxidation adequately as a first-order reaction.It is highly consistent with the experimental data.The apparent activation energy is 14.8 kcal/mol.This kinetic model is suited to reactions that involve high concentrations of NH 3 in the presence of oxygen.In this work, catalytic oxidation was found have potential for treating highly concentrated ammonia streams, helping industrial plants satisfy discharge regulations.
on nitrided MoNx/α-Al 2 O 3 and NiMoNy/α-Al 2 O 3 catalysts, respectively.Schmidt-Szałowski (1998) also developed a hypothetical model of the effect of these catalysts and their activity and selectivity in oxidizing ammonia.Lou and Chen (1995a) employed a catalyst that comprised a foam Pt, Ni and Cr alloy to elucidate the kinetics of the catalytic incineration of butanone and toluene.He determined that the Mars and van Krevelen model was suitable for describing the catalytic incineration of the VOCs.Lou and Lee (1997) used a Pt/Al 2 O 3 alloy catalyst to clarify the kinetics of the catalytic incineration of trichloromethane.He applied power-rate law kinetics and found that the reaction was first-order in the trichloromethane concentration and that the activation energy was 16.2 kcal/mol.Lou and Lee (1995b) also utilized a 0.05% Pt/Ni/Cr alloy catalyst to investigate the kinetics of the catalytic incineration of trichloromethane.According to their results, the Mars and van Krevelen Model was appropriate for describing the catalytic incineration of these VOCs.Gangwal (1988) used a 0.1% Pt, 3% Ni/ γ-Al 2 O 3 catalyst to elucidate the kinetics of deep catalytic oxidation with n-hexane and benzene.According to their results, the Mars and van Krevelen model was favorable for explaining the catalytic combustion of a binary mixture, at temperatures between 433 and 633 K. Recently,Auer and Thyrion (2002)  employed an integral fixed-bed reactor over an La 0.9 Ce 0.1 CoO 3 perovskite catalyst to examine the kinetics of total oxidation with methane.They favored the SED (Sequential Experimental Design) model to explain the catalytic combustion of water in feed, at temperatures from 633 to 773 K. Spivey (1987) excellently reviewed recent work on the modeling of the catalytic oxidation of VOCs.

Fig. 2 .
Fig. 2. XRD pattern of the various metal content on the CuO-CeO 2 bimetallic oxide catalyst for the conversion of NH 3 .

Fig. 3 .
Fig. 3. FTIR pattern of the various metal content on the CuO-CeO 2 bimetallic oxide catalyst for the conversion of NH 3 .

Fig. 4 .Fig. 5 .
Fig. 4. Changes in particle sizes distribution of the various metal content on the CuO-CeO 2 bimetallic oxide catalyst for the conversion of NH 3 .
Figs. 6 and 7 depict the surface morphological changes of copper-cerium (6:4) composite catalyst, identified by scanning electron microscopy (SEM) with elemental dot mapping photographs, for copper-cerium with various metal contents.These figures thus provide information on the surface structures of fresh and aged catalysts.The observations of agglomerates by SEM reveal that the composite catalyst particles are shaped into fine sheet-rodlike.Moreover, Fig.6shows a catalyst whose surface is more aggregated and crystalline than that in Fig.7.
a first-order reaction with NH 3 .This result is consistent with ( )

Fig. 11 Fig. 11 .
Fig.11compares the predicted and experimental conversions, obtained using an isothermal differential fixed-bed reactor at temperatures from 423-673 K.The figure indicates that the model accurately predicts the extent of conversion of NH 3 , which increased greatly from 19.0% (423 K) to 97.0% (673 K).The removal efficiency increases with temperature and as the residence period decreases.The deviation is increased, particularly in the temperature range region 573-673 K at a high concentration of NH 3 (1000 ppm).In such a case, increasing the temperature promotes the