Applicability of a catalytic reduction method using a palladium–copper catalyst and hydrazine for the denitration of a highly concentrated nitrate salt solution

Abstract

Denitration of a highly concentrated sodium nitrate (NaNO₃) aqueous solution via a catalytic reduction method using a palladium–copper catalyst supported on carbon powder (Pd–Cu/C) and hydrazine (N₂H₄) was investigated. It was demonstrated that nitrate ion (NO₃ ⁻) in a 5 mol L⁻¹ NaNO₃ solution was completely reduced through an intermediate nitrite ion (NO₂ ⁻) to nitrogen compounds such as nitrogen, nitrous oxide, and ammonia. By comparing the reaction rates of NO₃ ⁻ and NO₂ ⁻ obtained using catalysts with various Pd–Cu compositions and different reductants (hydrogen (H₂) or N₂H₄), it was determined that the catalyst with a molar ratio of Pd:Cu = 1:0.66 provides the maximum reaction rates for NO₃ ⁻ and NO₂ ⁻ using N₂H₄, and that not only the reactions of NO₃ ⁻ and NO₂ ⁻ but also that of N₂H₄ were affected by the Pd–Cu composition.

Keywords:

• denitration
• reduction
• nitrate
• hydrazine
• catalyst
• reaction rate

1. Introduction

Low-level radioactive liquid wastes containing highly concentrated nitrate salts are generated from the operation of nuclear-fuel reprocessing facilities [1]. When the liquid waste is solidified with cement without denitration treatment and then disposed; first the solidified substance comes in contact with groundwater, and then the nitrates in the substance dissolve in the groundwater and diffuse into surrounding environment. Because nitrate ion (NO₃ ⁻) is an environmental pollutant [2] and is included in the waste before solidification, its removal is effective for preventing its diffusion into the environment from the waste. Therefore, a nitrate removal method is required to efficiently remove NO₃ ⁻ of high concentration from a large amount of waste and reduce the amount of radioactive waste produced. It has been reported that NO₃ ⁻ is reduced via a catalytic reduction method using a bi-metal catalyst such as palladium–copper (Pd–Cu) and a reductant such as hydrogen gas (H₂) [3 –17], and this method has a possibility of meeting such requirements. However, because this method was mainly studied for the denitration of low-concentrated nitrate salt solutions, the applicability of the method to the denitration of highly concentrated nitrate salt solutions must be confirmed.

Since Vorlop and Tacke [4] reported a reduction method for low concentrations of NO₃ ⁻ using supported Pd–Cu as the catalyst and H₂ as the reductant, many studies have been conducted [3 –17]. The mechanism of NO₃ ⁻ reduction with a Pd–Cu catalyst and H₂ reductant has been studied, and it was deduced that NO₃ ⁻ is reduced through the nitrite ion (NO₂ ⁻) to nitrous oxide (N₂O), nitrogen (N₂), or ammonia (NH₃) via hydrogen adsorbed on the catalyst surface [6,11,14,16,17]. Ilinitch et al. [14] reported that the maximum reaction rate for NO₃ ⁻ was achieved using a catalyst with a molar ratio of Pd:Cu = 1:0.8, and that the maximum reaction rate of NO₂ ⁻ was observed with a Pd monometallic catalyst.

Nitrate ion at 200 ppm was completely reduced by a Pd–Cu catalyst and H₂ reductant [11]. However, when H₂ is used in the denitration of highly concentrated nitrate salt solutions, for example, 5 mol L⁻¹ sodium nitrate (NaNO₃), it is guessed that the reaction rate will be slow because the concentration of H₂ in the solution is low. To increase the concentration of the reductant, the use of a water-soluble reductant is effective. Formaldehyde (HCHO); formic acid (HCOOH); hydrazine (N₂H₄); and reductants including boron, sulfur, phosphorus, and other metal ions are available as water-soluble reductants. However, given that the liquid wastes will be solidified with cement and then disposed, the products of the denitration reaction must not be highly corrosive and should not include any environmentally harmful substances or metal salts. Therefore, reductants including boron, phosphorus, and other metals are not suitable. When HCHO or HCOOH is used, they are oxidized to carbonate and the carbonate salts are precipitated in the denitration of highly concentrated nitrate salt solutions. With a reductant including sulfur, there is a concern that the catalyst will be poisoned by sulfur compounds [18]. Hydrazine, which produces N₂, is thus a suitable reductant.

In the present study, NO₃ ⁻ in a 5 mol L⁻¹ NaNO₃ aqueous solution was reduced using a Pd–Cu catalyst and hydrazine monohydrate (N₂H₄ · H₂O) reductant at 333 K. The denitration performance of the N₂H₄ was compared with that of H₂. The reaction rates of NO₃ ⁻ and NO₂ ⁻ were determined with catalysts having various Pd–Cu compositions, and the effect of the Pd–Cu composition on the reactivities of not only NO₃ ⁻ and NO₂ ⁻ but also N₂H₄ was discussed.

2. Experimental

2.1. Preparation and characterization of the catalysts

Pd–Cu catalysts supported on carbon powder (Pd–Cu/C) were prepared by an electroless method as follows. Palladium dichloride (PdCl₂, Sigma-Aldrich Co., LLC, 99%), copper dichloride (CuCl₂, Kojundo Chemical Laboratory Co., Ltd., 4N), and carbon powder (Kojundo Chemical Laboratory Co., Ltd., 2N7) were used. The carbon powder was suspended in a mixed solution of PdCl₂ (0–0.016 mol L⁻¹), CuCl₂ (0–0.013 mol L⁻¹), and 0.01 mol L⁻¹ hydrochloric acid (Wako Pure Chemical Industries, Ltd.), and then disodium ethylenediaminetetraacetate (Nacalai Tesque, Inc., 99.5%) was added to achieve 0.1 mol L⁻¹ in the suspension with stirring at 333 K. Next, the pH of the suspension was adjusted to 10 using a HORIBA F53 pH meter by dropwise adding a 5 mol L⁻¹ sodium hydroxide (Koso Chemical Co., Ltd., 96%) solution. Then, a solution containing an equimolar amount (relative to the metals) of sodium borohydride (Wako Pure Chemical Industries, Ltd., 90%) was added to the suspension. The carbon powder in the suspension was recovered by suction filtration, washed, and then dried at 333 K. Catalysts with the composition 0.47 mmol g⁻¹ Pd–x mmol g⁻¹ Cu/C (x = 0–0.39) and y mmol g⁻¹ Pd–0.39 mmol g⁻¹ Cu/C (y = 0–0.47) were prepared. Distilled and deionized water was used in the preparation of the catalysts. The naming convention used in this paper for the catalysts with various Pd–Cu compositions is shown by the following example: 0.47 mmol g⁻¹ Pd–0.39 mmol g⁻¹ Cu/C is denoted as 0.47 Pd–0.39 Cu/C.

The specific surface area of the carbon powder was measured using a SHIMADZU Gemini V surface area analyzer. The total metal contents of the catalysts were evaluated in following way: Pd and Cu of the 0.47 Pd–0.39 Cu/C were dissolved in an aqua regia, and then the Pd and Cu concentrations in the solution were determined using a SHIMADZU ICPS-8000E inductively coupled plasma atomic emission spectrometer (ICP-AES). The X-ray diffraction (XRD) patterns of the prepared Pd–Cu/C were also measured using a Rigaku MiniFlexII X-ray diffractometer.

2.2. Procedures

2.2.1. Denitration

Aqueous solutions containing NO₃ ⁻ or NO₂ ⁻ were prepared using NaNO₃ (Wako Pure Chemical Industries, Ltd., 99.0%) or sodium nitrite (NaNO₂, Wako Pure Chemical Industries, Ltd., 98.5%), respectively. An apparatus for denitration is shown in Figure 1. The apparatus consisted of a four necked Pyrex glass 500 mL separable flask in a water bath, a magnetic stirrer, a measuring cylinder, a peristaltic pump, a tygon tube, a condenser tube, and a recirculating cooler. Two grams of 0.47 Pd–0.31 Cu/C and 100 mL of 5 mol L⁻¹ NaNO₃ solution were placed in the flask. The suspension was stirred and maintained at 333 K by heating with the water bath. Forty milliliters of N₂H₄ · H₂O solution (Sigma-Aldrich Co., LLC, 98%) was then added by drops into the suspension for 0.5 h.

Figure 1. An apparatus for denitration experiment. (a) A four necked separable flask, (b) a water bath, (c) a underwater magnetic stirrer, (d) a measuring cylinder, (e) a peristaltic pump, (f) a tygon tube, (g) a condenser tube, and (h) a recirculating cooler.

In a second experiment, 0.2 g of the 0.47 Pd–0.39 Cu/C and 100 mL of 5 × 10⁻³ mol L⁻¹ or 0.5 mol L⁻¹ NaNO₃ solution were placed in the flask, and then the suspension was stirred and maintained at 333 K. The flow of H₂ (TOMOE SHOKAI Co., Ltd., 99.99%) into the suspension was maintained at a constant rate of 30 mL min⁻¹. Additional experiments using 0.2 g of 0.47 Pd/C, 0.47 Pd–0.39 Cu/C, or 0.39 Cu/C and 5 × 10⁻³ mol L⁻¹ of NaNO₂ solution with H₂ were also conducted using the abovementioned procedure.

In a third experiment, 0.2 g of a Pd–Cu/C catalyst and 100 mL of 0.5 mol L⁻¹ NaNO₃ or NaNO₂ solution were placed in the flask. The suspension was stirred and maintained at 333 K. Four milliliters of N₂H₄ · H₂O was then rapidly added. The catalysts used in these experiments included 0.47 Pd–x Cu/C (x = 0−0.39) and y Pd–0.39 Cu/C (y = 0−0.47).

2.2.2. Analysis

To analyze the reactions, a portion of the suspension was sampled, followed by filtration and dilution. The concentrations of NO₃ ⁻, NO₂ ⁻, and N₂H₄ in the diluted solutions were determined using a DIONEX (now THERMO SCIENTIFIC) ICS-1000 ion chromatographic apparatus. The pH value of the suspension was measured using a pH meter. The generated gas was collected and the component of the gas was analyzed using a HITACHI263-50 gas chromatography. The concentrations of the Pd and Cu ions in the suspension after denitration were determined by ICP-AES.

3. Results and discussion

3.1. Characterization of the catalysts

The Pd–Cu/C was prepared using an electroless method. The specific surface area of the carbon powder support was determined to be 13.57 m² g⁻¹. The metals supported on the carbon powder were dissolved in an aqua regia, and the concentrations of the Pd and Cu in the solution were determined. For this determination, the 0.47 Pd–0.39 Cu/C catalyst with the highest supported ratio of the prepared catalysts was used. The results are summarized in Table 1 together with the calculated concentrations of the Pd–Cu/C based on the preparation conditions. The found ratios of Pd and Cu were 0.46 and 0.37 mmol g⁻¹, respectively, which are in good agreement with the calculated values. This result suggests that nearly all the metal ions used for the catalyst preparation were deposited on the carbon powder.

Table 1. Calculated and found ratios for the Pd and Cu supported on carbon powder.

	Pd (mmol g^−1)	Cu (mmol g^−1)
Calculated	0.47	0.39
Found	0.46	0.37

Figure 2 shows the XRD patterns of the carbon powder and the 0.47 Pd–x Cu/C (x = 0−0.39) and y Pd–0.39 Cu/C (y = 0–0.47) catalysts. Three peaks were observed for the Pd/C in Figure 2(b). The peak at 2θ = 40.02° was attributed to Pd(111) and those at 42.43° and 44.55° were attributed to graphite [19,20]. The Pd(111) peak shifted with an increase in the Cu fraction of the catalyst. The Pd(111) peak appeared at 2θ = 40.35°, 40.57°, 40.66°, 40.84°, and 40.88° for the catalysts with Cu fractions of 0.14, 0.25, 0.40, 0.46, and 0.51, respectively. These shifts of the Pd(111) peak positions indicate the formation of an alloy of the Pd with the Cu having different lattice parameters [21]. In addition, the Pd(111) peak intensity decreased with an increase in the Cu fraction in the catalyst when the Pd concentration remained constant. This decrease in the peak intensity is likely caused by a decrease in the crystallinity of the Pd–Cu alloy (Figure 1(b)–(e)). Other peaks were observed for the catalysts with a Cu fraction greater than 0.68 (Figure 1(h)–(j)). These peaks were attributed to Cu oxide [22,23]. No peaks for Cu metal were observed in the catalysts having a high-Cu fraction, because the Cu(111) peak appears at approximately 2θ = 43.3°, which overlaps with the C peaks [24].

Figure 2. XRD patterns for the carbon powders 0.47 Pd–x Cu/C (x = 0−0.39) and y Pd–0.39 Cu/C (y = 0−0.47). (a) Carbon powder, mole fraction of Cu/(Pd + Cu) = (b) 0, (c) 0.14, (d) 0.25, (e) 0.40, (f) 0.46, (g) 0.51, (h) 0.68, (i) 0.81, and (j) 1.

3.2. Denitration

3.2.1. Denitration of a highly concentrated NaNO₃ solution using the Pd–Cu catalysts and N₂H₄

Nitrate ion in a 5 mol L⁻¹ NaNO₃ solution (100 mL) was reduced using the 0.47 Pd–0.31 Cu/C catalyst (2 g) suspended in the NaNO₃ solution, and 40 mL of N₂H₄ · H₂O was added by drops for 0.5 h. The variation in the concentration of NO₃ ⁻, NO₂ ⁻, and N₂H₄ in the suspension as a function of time was determined, and the results are plotted in Figure 3. The concentration of NO₃ ⁻ decreased from 5 mol L⁻¹ to 0.33 mol L⁻¹ after 0.5 h and fell below 1 × 10⁻³ mol L⁻¹ after 4 h. Nitrite ion was newly detected in the suspension, suggesting that the denitration proceeds through NO₂ ⁻ as an intermediate. The concentration of NO₂ ⁻ increased to 1.8 mol L⁻¹ after 0.25 h and decreased to 6.6 × 10⁻³ mol L⁻¹ after 7 h. The concentration of N₂H₄ was below the detection limit until 0.25 h, reached 3.7 mol L⁻¹ at 0.5 h, and then decreased to 0.45 mol L⁻¹ after 7 h. From these changes in the concentration of NO₃ ⁻, NO₂ ⁻, and N₂H₄ with time, it was assumed that the reaction of NO₃ ⁻ with N₂H₄ proceeds more rapidly than the addition rate of N₂H₄ · H₂O, but that the reaction of NO₂ ⁻ with N₂H₄ was slower than that. The pH of the suspension rose to 14 for 7.5 min and 14.6 after 7 h. About 99.9% of NO₃ ⁻ was removed with this method, and thus it was determined that the use of a Pd–Cu catalyst and N₂H₄ is applicable for the denitration of highly concentrated nitrate salt solutions.

Figure 3. Concentrations of NO₃ ⁻, NO₂ ⁻, and N₂H₄ in suspension as a function of time. Reaction solution: 100 mL of 5 mol L⁻¹ NaNO₃, catalyst: 2 g of 0.47 Pd–0.31 Cu/C, reductant: 40 mL of N₂H₄ · H₂O, drop rate of N₂H₄ · H₂O: 80 mL h⁻¹, temperature: 333 K. •: NO₃ ⁻, ○: NO₂ ⁻, □: N₂H₄.

However, NO₃ ⁻ was not reduced by mixing the NaNO₃ solution and N₂H₄ · H₂O in the absence of the catalyst at 333 K. Therefore, it was confirmed that the denitration reaction proceeds via a catalytic reaction. In addition, after completion of the denitration reaction, no dissolved Pd or Cu ions from the catalyst were found in the suspension. This result indicated that the catalyst was not corroded during the reaction with N₂H₄ in a highly concentrated NO₃ ⁻ solution.

The mechanism of NO₃ ⁻ reduction with Pd–Cu catalysts and H₂ has been widely studied, and it has been deduced that NO₃ ⁻ is reduced through NO₂ ⁻ to N₂, N₂O, and NH₃ via reaction with hydrogen adsorbed on the catalyst surface [6,11,14,16,17]. In the present study, a Pd–Cu catalyst was also used, and N₂, N₂O, and NH₃ were produced. In addition, it has been reported that the adsorbed hydrogen is generated by decomposition of N₂H₄ on a Pd or Cu surface [25 –27]. Therefore, in the reduction using N₂H₄, NO₃ ⁻ is likely reduced through the same mechanism as that used for the reduction using H₂.

3.2.2. Comparison of H₂ with N₂H₄

Nitrate ion in a 5 × 10^–3 mol L⁻¹ NaNO₃ solution (100 mL) was reduced using a 30 mL min⁻¹ H₂ flow in the presence of 0.47 Pd–0.39 Cu/C catalyst (0.2 g). Eighty percent of NO₃ ⁻ in the suspension was reduced in 2 h. However, in the denitration of a 0.5 mol L⁻¹ NaNO₃ solution, the concentration of NO₃ ⁻ did not decrease. Based on these results, it was concluded that H₂ is unsuitable for the reduction of highly concentrated of NO₃ ⁻ solutions, although H₂ is applicable for the reduction of low-concentrated NO₃ ⁻ solutions. On the other hand, NO₃ ⁻ in a 0.5 mol L⁻¹ NaNO₃ solution (100 mL) was effectively reduced using N₂H₄ · H₂O (4 mL) in the presence of 0.47 Pd–0.39 Cu/C (0.2 g). In this denitration reaction, 80% of the NO₃ ⁻ was reduced within 1 h. This result demonstrated that N₂H₄ is more effective than H₂ for the denitration of highly concentrated NaNO₃ solutions.

Elliot et al. reported that the solubility of H₂ at 343 K was 7.7 × 10⁻⁴ mol L⁻¹ in water and 4.7 × 10⁻⁴ mol L⁻¹ in a 2.0 mol L⁻¹ lithium nitrate solution [28]. Thus, the solubility of H₂ was of the same order of magnitude regardless of whether the solution contained any salt. It is expected that the concentration of H₂ in the present study would also be similar to the reported values. It has also been reported that NO₃ ⁻ strongly adsorbs onto the Pd–Cu catalyst surface [16], and H₂ adsorption onto the catalyst surface may be blocked by the high concentration of adsorbed NO₃ ^–. Therefore, it is presumed that the concentration of adsorbed H₂ on the catalyst surface is too low owing to the combination of the low-solubilized H₂ concentration and the defeat of H₂ in adsorption competition with NO₃ ^–. These factors may also be responsible for the reduced reaction rate of NO₃ ⁻ in the reaction with H₂ and the lack of a decrease in the NO₃ ⁻ concentration in the above experiment. On the other hand, for the reaction with N₂H₄, the initial concentration of the reductant was 0.8 mol L⁻¹, which is about 10³ times higher than the concentration of H₂, and was sufficient for the denitration of the 0.5 mol L⁻¹ NaNO₃ solution.

3.3. Reaction rate as a function of the Pd–Cu composition

The NO₃ ⁻ present in a 0.5 mol L⁻¹ NaNO₃ solution (100 mL) was reduced using the 0.47 Pd–x Cu/C (x = 0−0.39) and y Pd–0.39 Cu/C (y = 0−0.47) catalysts (0.2 g) with N₂H₄ · H₂O (4 mL). Similarly, the NO₂ ⁻ in a 0.5 mol L⁻¹ NaNO₂ solution (100 mL) was also reduced. Then, in order to decrease the temperature change caused by the heat of reaction, 1/10th of the concentrations of both the NaNO₃ and N₂H₄ reactants and the amount of catalyst were used for the denitration of a 5 mol L⁻¹ NaNO₃ solution, and also, the catalyst metal of the same quantity was used in all examinations without depending on difference in composition of the catalyst. In these reactions, the concentration of NO₃ ⁻ and NO₂ ⁻ decreased with time, and the rate of the decrease in the NO₃ ⁻ concentration depended on the Pd–Cu composition of the catalyst. For example, the concentration of NO₃ ⁻ decreased most rapidly using the 0.47 Pd–0.31 Cu/C catalyst, with 87% of NO₃ ⁻ reduced in 1 h, while only 55% of NO₃ ⁻ reduced in 1 h using the 0.47 Pd–0.16 Cu/C catalyst. Here, the reaction rate of NO₃ ⁻ reduction was defined as the amount of NO₃ ⁻ reduced per unit time and unit amount of Pd–Cu, and was calculated using the following procedure. First, an average slope for the initial 0.25 h of the reaction was determined from a graph of the amount of reduced NO₃ ⁻ versus time, and the average slope was divided by the amount of Pd–Cu in the catalyst used for the reaction. The reaction rate of NO₂ ⁻ reduction was also calculated using a similar procedure.

The reaction rates of NO₃ ⁻ and NO₂ ⁻ are plotted in Figure 4 as a function of the mole fraction of Cu in the catalyst. This figure shows the results obtained for the catalysts with a Cu fraction ranging from 0 to 1, i.e., the catalysts with 0.47 Pd–x Cu/C (x = 0−0.39) and y Pd–0.39 Cu/C (y = 0−0.47). Nitrate ion was not reduced by the Pd monometallic catalyst, but NO₃ ⁻ was reduced by the Pd–Cu catalysts (Figure 4). The reaction rate of NO₃ ⁻ increased with the Cu fraction and reached a maximum at a Cu fraction of approximately 0.4, and then decreased with an increase in the Cu fraction in the range of 0.4–1. With the Cu monometallic catalyst, NO₃ ⁻ was reduced. These results clarified that Cu is indispensable for the reduction of NO₃ ⁻. It has been reported that the reduction of NO₃ ⁻ to NO₂ ⁻ is initiated by the Cu sites, which are then converted to Cu₂O during the deoxidation of NO₃ ⁻ [14]. However, the formed Cu₂O must then be reduced back to Cu by the spillover hydrogen generated mainly on the Pd sites, thus regenerating the Pd–Cu alloy. Therefore, the reaction rate of NO₃ ⁻ increases with an increase in the Cu content in the low-Cu fraction range, but decreases with a decrease in the number of Pd sites in the high-Cu fraction range. In this denitration system, the catalyst with 0.47 Pd–0.31 Cu/C, which has a Cu fraction of 0.40, gives the best results for the deoxidation of NO₃ ⁻ and the regeneration of the Cu, and thus showed the highest rate for the reduction of NO₃ ⁻. This Cu fraction is similar to the results obtained in some previous reports [14,17].

Figure 4. Reaction rates of NO₃ ⁻ and NO₂ ⁻ using several catalysts with different Cu fractions. Reaction solution: both 100 mL of 0.5 mol L⁻¹ NaNO₃ and NaNO₂, catalyst: 0.2 g of 0.47 Pd–x Cu/C (x = 0−0.39) or y Pd–0.39 Cu/C (y = 0−0.47), reductant: 4 mL of N₂H₄ · H₂O, temperature: 333 K. •: NO₃ ⁻, ○: NO₂ ⁻.

Reduction of NO₂ ⁻ as the intermediate in the NO₃ ⁻ reduction was also studied. It was confirmed that NO₂ ⁻ is not reduced when a NaNO₂ solution is mixed with N₂H₄ · H₂O in the absence of the catalyst, as well seen with the NO₃ ⁻ solution mentioned above. As shown in Figure 4, NO₂ ⁻ was reduced by the Pd monometallic catalyst and the Pd–Cu bi-metal catalyst, but was not reduced by the Cu monometallic catalyst. Therefore, it was confirmed that the NO₂ ⁻ reduction was promoted by a catalytic reaction, and that Pd is indispensable for the reduction of NO₂ ⁻. The reaction rate of NO₂ ⁻ reached a maximum when the catalyst with a Cu fraction of approximately 0.4 was used. This result is different from those described in earlier reports, in which the reaction rate of the NO₂ ⁻ reduction by H₂ reached a maximum when a catalyst with a low-Cu fraction was used [10,14,17]. The decrease in the reaction rate of NO₂ ⁻ with an increase in the Cu fraction (Figure 4) can be explained by the decrease in the Pd content if Pd serves as the NO₂ ⁻ active site. The increase in the reaction rate of NO₂ ⁻ (Figure 4) is discussed below.

The reaction rate of NO₂ ⁻ was determined using H₂ instead of N₂H₄. Here, 100 mL of 5 × 10⁻³ mol L⁻¹ NaNO₂ solution and 0.2 g of 0.47 Pd/C, 0.47 Pd–0.39 Cu/C, and 0.39 Cu/C were used. The relationship between the reaction rate of NO₂ ⁻ and the Pd–Cu composition using H₂ and N₂H₄ is summarized in Table 2. When Pd/C and Pd–Cu/C were used, the reaction rate of NO₂ ⁻ was nearly the same as in the reduction with H₂. Because NO₂ ⁻ is mainly reduced on Pd surface, same reaction rates were observed for the catalysts with same Pd content. On the other hand, the rate with Pd–Cu/C was higher than that with Pd/C in the reduction with N₂H₄. This result indicates that the alloy of Pd with Cu strongly contributes to the increase in the N₂H₄ reactivity. Thus, the behavior of the NO₂ ⁻ reaction rate (Figure 4) can be explained by the change in reactivity of the N₂H₄ with the change in the Cu fraction.

Table 2. Comparison of the reaction rate of NO₂ ⁻ using H₂ and N₂H₄.

Reductant	Reaction rates (h^−1)
	Pd/C	Pd–Cu/C	Cu/C
H2 ^a	1.8	1.7	–
N2H4 ^b	26	106	–
Notes: ^a100 mL of 5 × 10^−3 mol L^−1 NaNO2; ^b100 mL of 0.5 mol L^−1 NaNO2. Reductant: 30 mL min^−1 of H2, 4 mL of N2H4 · H2O, temperature: 333 K. “–” means that reaction was not observed.

We concluded the mechanism of the denitration reaction using the Pd–Cu catalyst and N₂H₄ as follows. At first, NO₃ ⁻ was deoxidized to NO₂ ⁻ on Cu and next NO₂ ⁻ was reduced to N₂O, N₂, or NH₃ on Pd. The adsorbed hydrogen needed for the NO₃ ⁻ and NO₂ ⁻ reduction was generated by decomposition of N₂H₄ on Pd and Cu surface. The alloying of Pd and Cu increased reactivity of not only NO₃ ⁻ and NO₂ ⁻ but also N₂H₄, and the reaction rates of NO₃ ⁻ and NO₂ ⁻ showed the maximum when the composition of Pd:Cu was 1:0.66, as a result. By using a water-soluble reductant like N₂H₄, enough amount of the reductant could adsorb on the catalyst surface in competitive adsorption with NO₃ ⁻ in the suspension of highly concentrated NaNO₃ solution and the Pd–Cu catalyst, and then the denitration proceeded rapidly. It was confirmed that the chemical reduction method using the Pd–Cu catalyst and the N₂H₄ reductant had a high applicability for the denitration of the highly concentrated NaNO₃ solution.

4. Conclusion

A catalytic reduction method that is applicable for highly concentrated NaNO₃ solutions was demonstrated using the combination of a Pd–Cu catalyst and N₂H₄. Highest reaction rates for NO₃ ⁻ and NO₂ ⁻ reduction were achieved using a catalyst with a metal composition of approximately Pd:Cu = 1:0.66, and not only the reaction rates of NO₃ ⁻ and NO₂ ⁻ but also the reactivity of N₂H₄ were affected by the formation of a Pd–Cu alloy. To apply this method for practical use, the durability of the catalyst should be further studied in order to determine the optimized conditions for enhancing the catalyst lifetime and thus decreasing the operating costs of the catalyst. In addition, the selectivity for N₂O, N₂, and NH₃ should be studied to determine how to suppress the production of N₂O and NH₃ as a means for reducing the required off-gas treatment processes.

Additional information

Notes on contributors

Haruhiko Kadowaki

†Present address: Japan Atomic Energy Agency, 3 Myojin-cho, Tsuruga-shi, Fukui 914-8510, Japan.