Noble Metal-Catalyzed Ammonia Generation by Formic Acid

Environ. Sci. Technol. 1997, 31, 984-992

Noble Metal-Catalyzed Ammonia Generation by Formic Acid Reduction of Nitrate in Simulated Nuclear Waste Media R. B. KING* AND N. K. BHATTACHARYYA Department of Chemistry, University of Georgia, Athens, Georgia 30602 H. D. SMITH AND K. D. WIEMERS Pacific Northwest National Laboratory, Richland, Washington 99352

Simulants for the Hanford Waste Vitrification Plant (HWVP) feed containing the major non-radioactive components Al, Cd, Fe, Mn, Nd, Ni, Si, Zr, Na, CO32-, NO3-, and NO2- were used to study the formic acid reduction of nitrate and/or nitrite to ammonia at 90 °C catalyzed by the noble metals Ru, Rh, and/or Pd found as fission products in waste from the reprocessing of irradiated uranium. Reactions of this type were monitored using gas chromatography to analyze the CO2, H2, NO, and N2O in the gas phase and a microammonia electrode to analyze the NH4+/NH3 in the liquid phase as a function of time. The rhodium-catalyzed reduction of nitrogen-oxygen compounds to ammonia by formic acid was found to exhibit the following features: (1) Nitrate rather than nitrite is the principal source of ammonia. (2) Ammonia production occurs at the expense of hydrogen production. (3) Supported rhodium metal catalysts are more active than rhodium in any other form, suggesting that ammonia production involves heterogeneous rather than homogeneous catalysis.

Introduction The most promising method for the disposal of highly radioactive nuclear wastes is a vitrification process in which the wastes are incorporated into borosilicate glass, the molten glass is poured into large stainless steel canisters, and the sealed canisters are disposed in geologic repositories. The method of feed preparation for the vitrification process in the Hanford Waste Vitrification Plant (HWVP) uses formic acid for glass redox and melter feed rheology control. However, the reducing properties of formic acid can lead to a number of difficulties in the processing of nuclear wastes. One difficulty is the catalytic decomposition of formic acid to hydrogen by

HCO2H f H2v + CO2v

(1)

The catalytic role in hydrogen generation from formic acid of the noble metals Ru, Rh, and Pd present in the wastes as uranium fission products (1) is discussed in previous papers (2, 3). In addition, formic acid reduction of the nitrite and/or nitrate present in nuclear wastes catalyzed by noble metal fission products can lead to the production of ammonia, which is a problem of considerable concern because of the generation of a potential ammonium nitrate explosion hazard from the pretreatment plant operation. This paper describes work

984

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 4, 1997

on the noble metal-catalyzed formic acid reduction of nitrite and/or nitrate to ammonia in simulated nuclear waste media.

Experimental Methods Feed Simulant Experiments. The experimental methods were similar to those used in the previously reported work (3) with the necessary modifications to allow monitoring of ammonia production. Thus, the reactions were conducted in glass reaction vessels of ∼550 mL total capacity equipped with a threaded plug and O-ring adapter. A pressure gauge was attached to the system. The three side arms of the vessel near the top were capped with rubber septa through which needles could be inserted to flush the system or to sample the gas phase. A fourth opening also capped by a rubber septum was used for the periodic withdrawal of liquid samples from the reaction mixture for ammonia analyses. The reaction vessel had two thermocouple wells. A thermocouple attached to a cycle heater was placed in one well to maintain a constant temperature. The other well was used to check the temperature of the reaction mixture independently with a second thermocouple. The reaction mixture was stirred magnetically during the course of the experiment. In typical experiments, a reaction vessel of the above type was charged with 50 mL of the feed simulant or other reaction medium, 10 mL of water, and a solid noble metal catalyst. After flushing the mixture with argon, the system was closed and the gas phase analyzed using the system described below. The solution was warmed to the desired reaction temperature (typically 90 °C) and kept there using a temperature controller. Formic acid was then added below the surface of the feed simulant from a 10-mL plastic disposable syringe driven by a syringe pump (Sage Instruments Model 355) set to deliver 0.0194 mL/min (1.164 mL/h) corresponding to 0.448 mmol of HCO2H/min (26.9 mmol of HCO2H/h). The syringe pump delivery rate was checked weekly with pure water. The pressure gauge was used to monitor the pressure buildup from the gases produced during the experiment and to assure the absence of leaks. Pressures in excess of 30 psig were frequently found to cause leakage and thus were avoided by limiting the amount of feed simulant used. Gas samples were withdrawn periodically at the reaction temperature using a gas syringe inserted into one of the rubber septa; these samples were analyzed immediately using gas chromatography as indicated below (Gas Analyses). Liquid samples were also withdrawn periodically using a 1-mL syringe and analyzed for NH4+/NH3 using the ammonia microelectrode described below. In some experiments, the pH of the reaction mixture was monitored using a pH electrode inserted into a neck of the reaction vessel and immersed in the liquid reaction mixture. The methods for the preparation of the simulant for a typical Hanford Waste Vitrification Plant (HWVP) neutralized current acid waste (NCAW) feed used in this work are given in the previous paper of this series (3). The raw materials used for the preparation of the standard NCAW simulant designated as UGA-12M1 are listed in Table 1. Gas Analyses. Hydrogen analyses were performed using a gas chromatograph (Varian 90P). The hydrogen was eluted on a 20 cm × 6 mm column packed with a 40/60 mesh 13× molecular sieve material obtained from Varian Corporation. Argon was used as the carrier gas. The column temperature was maintained at 80 °C. Carbon monoxide, carbon dioxide, nitrous oxide, and nitric oxide were monitored using a Fisher Model 1200 gas partitioner. The gases were separated on the basis of their size and polarity by means of two columns, a 2-m 80/100 mesh Columnpak PQ and a 3.3-m 13× molecular sieve column mounted in series using helium as the carrier

S0013-936X(96)00335-5 CCC: $14.00

 1997 American Chemical Society

TABLE 1. Sources of Components of Feed Simulant UGA-12M1 componenta

molarity

source

amountb (g)

water solubilityc

Al Cd Fe Mn

0.226 0.03 0.45 0.0314

Nd Ni Si Zr Na+ CO32NO3NO2-

0.0264 0.0392 0.0854 0.156 0.801d 0.125 0.116 0.435

Al(NO3)3‚9H2O Cd(NO3)3‚4H2O Fe(NO3)3‚9H2O KMnO4 Mn(NO3)2‚6H2O Nd(NO3)3‚6H2O Ni(NO3)2‚6H2O SiO2 ZrO(NO3)2‚6H2O Na2CO3 + NaNO3 + NaNO2 Na2CO3 NaNO3 NaNO2

84.8 9.2 181.8 2.0 5.4 11.6 11.6 5.1 52.9 see below 13.2 9.9 30.0

63.7 g/100g at 25 °C 215 g/100 g at 25 °C soluble 6.38 g/100 g at 20 °C 426.4 g/100 g at 0 °C 152.9 g/100 g at 25 °C 238.5 g/100 g at 0 °C insoluble soluble see below 7.1 g./100 g at 0 °C 92.1 g/100g at 25 °C 81.5 g/100 g at 15 °C

a The insoluble components are present as Al(OH) , Cd(OH) , Fe(OH) , MnO , Nd O , Ni(OH) , SiO , and ZrO . b Amount of source chemical for 3 2 3 2 2 3 2 2 2 a 1-L batch. c The water solubility figures for these salts are taken from the CRC Handbook of Chemistry and Physics. More detailed information on the solubilities of sodium nitrate, sodium nitrite, and sodium aluminate in similar simulated nuclear waste media is given by D. A. Reynolds and D. L. Herting, Report RHO-RE-ST-14P prepared for the U.S. Department of Energy under Contract DE-AC06-77RL01030. d The nominal Na+ concentration in the simulant is 0.88 M because the NaOH precipitant was incompletely washed from the oxide/hydroxide precipitates.

TABLE 2. Ammonia Analyses on Samples from Selected Titrations of Feed Simulants with Formic Acid final sample identification (amt, mL) 12M1 (50) 12M1 (50) 12M1 (50) 13M+Rh1 (50) 13M+Rh1 (50) 12M1 (50) 12M1 (50) water (50) NaNO2 (1.5 g) water (50), NaNO2 (1.5 g) 12M1 (50) 12M1 (50) water (60), NaNO2 (1.5 g), HCO2Na (6.2 g) water (60), NaNO2 (0.5 g), HCO2Na (6.2 g) water (60), HCO2Na (6.2 g) 12M1 (50)

catalyst (amt, mg) RhCl3‚3H2O (14) 5% Rh/C (117) 5% Rh/Al2O3 (115) RhCl3‚3H2O (14) 0 RhCl3‚3H2O (14) RhCl3‚3H2O (14) CuCl2 (100 mg) RhCl3‚3H2O (14) CuCl2 (100 mg) RhCl3‚3H2O (14) RhCl3‚3H2O (14) PdCl2 (12) PdCl2 (14) PdCl2 (14) PdCl2 (14) RuCl3‚3H2O (15)

gas. The temperatures of both columns were maintained at 50 °C. Sensitivity factors were determined from known amounts of pure certified samples of the gases of interest and were rechecked every week. Nitrogen dioxide could not be determined by gas chromatography because of the NO2 a N2O4 equilibrium. The data reported in the figures are results from single experiments. pH Monitoring. The pH of selected reactions was monitored using a pH electrode (Ingold HA405-DXK-58/120 Combination Xerolyt) with a 2-14 pH range and a 0-110 °C temperature range and resistant to pressures up to 16 bar at 25 °C and 6 bar at 100 °C without pressure compensation. Temperature compensation for the pH measurements was effected with a PT 100-S8-120 automatic temperature compensator. The pH electrode and temperature compensator were connected to an Ingold Model 2300 pH transmitter that could read both pH and temperature. Before each experiment, a two-point calibration of the pH electrode was performed using standard pH 2.00 and pH 9.00 buffers. Ammonia Analyses. Reaction mixtures were analyzed for ammonium ion produced by nitrate or nitrite reduction. In this connection, a 0.5-mL sample of the reaction mixture was diluted with 0.5 mL of distilled water and then made strongly basic with 1-2 drops of 10 N sodium hydroxide in order to convert NH4+ to NH3. The ammonia concentration was then determined using a MI-740 microammonia electrode ob-

HCO2H mL (added over min)

temp (°C)

H2 (mmol)

pH

NH3 (mmol)

3.63 (190) 3.53 (190)

89 ( 1 90 ( 1

1.4 0.2

4.3 6.9

0.0 15.0

3.67 (190) 3.53 (190) 3.53 (190) 3.71 (190) 3.40 (185)

88-91 90 ( 2 91 ( 2 70-90 90 ( 1

0.3 4.8 0.4 1.4 0.3

6.9 4.1 3.9 3.5 4.3

5.4 0.1 0.3 0.2 0.1

3.71 (190)

90 ( 1

6.3

2.9

0.2

3.46 (180) 3.61 (190) 3.67 (190) 0 0 0 3.63 (190)

91 ( 1 91 ( 1 90 ( 2 92 ( 1 92 ( 2 90 ( 2 90 ( 1

10.0 1.9 0 0.1 1.4 1.4 0

2.6 4.1 5.4 10.6 10.0 9.6 3.9

0.3 0.2 0.4 0.1 0.8 0 0.1

tained from Microelectrodes, Inc. (Londonderry, NH). This electrode is designed to measure ammonia gas and to record the ammonia concentration as millivolts on a standard pH meter; an Orion 290A portable pH meter was used as a millivolt meter for these studies. The ammonia electrode was calibrated before each analysis using 0.1, 0.01, and 0.001 N NH4Cl solutions made basic with 10 N sodium hydroxide.

Results Ammonia Analyses of Previous Runs. The concentrations of ammonium ion produced by selected syringe pump titrations of various nitrite and/or nitrate-containing feed simulants with 88 wt % formic acid in the presence of various noble metals for the work described in the previous paper (3) were determined using the analytical methods outlined in Ammonia Analyses. The ammonia analyses were performed on the residues remaining from the experiments, which had been stored in sealed screw cap bottles at ambient temperature (24 ( 7 °C) since completion of the experiments. The results summarized in Table 2 indicate that the only experiments in which significant quantities of ammonia were produced (>1 mmol) were the runs highlighted in boxes in Table 2 in which the rhodium catalysts were supported on charcoal or alumina. These experiments are also distinguished by the production of far too much carbon dioxide to arise solely from reactions of formic acid leading to carbonate

VOL. 31, NO. 4, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

985

FIGURE 1. Treatment of 50 mL of the NCAW feed simulant UGA12M1, 10 mL of water, and 0.118 g of 5% Rh/C with 85 mmol of 88 wt % formic acid at 90 ( 1 °C. This experiment produced 1.6 mmol of NH3 and 0.15 mmol of H2 at the end.

FIGURE 2. Treatment of 50 mL of the NCAW feed simulant UGA12M1, 10 mL of water, and 0.118 g of 5% Rh/Al2O3 with 85 mmol of 88 wt % formic acid at 90 ( 1 °C. This experiment produced 1.2 mmol of NH3 and 0.3 mmol of H2 at the end.

decomposition, MnO2 reduction to Mn2+, nitrite reduction to N2O, and formic acid decomposition to H2. In addition, the final pH in both of these experiments is much higher (6.9) than the final pH in any of the other formic acid titrations (2.9-4.3). Production of each mole of ammonia by nitrate reduction can convert 4 mol of formic acid to carbon dioxide and neutralizes a fifth mole of formic acid to give formate, i.e.

wt % formic acid in the presence of 5% Rh/C (Figure 1) and 5% Rh/Al2O3 (Figure 2), respectively, at 90 °C. Both of these experiments using supported rhodium metal catalysts are seen to produce significant amounts of ammonia. In fact, the amounts of ammonia produced in these experiments are significantly greater than the amounts of hydrogen. The use of supported rhodium metal catalysts appears to be essential for the production of relatively large amounts of ammonia. Thus, treatment of the NCAW feed simulant with 88 wt % formic acid in the presence of rhodium trichloride at 90 °C (Figure 3), which are the standard conditions used for many of the hydrogen generation studies described in the previous paper (3), is seen to produce much more hydrogen and much less ammonia. Since ammonia production appears to occur at the expense of hydrogen production, we suspect that the supported rhodium metal catalysts are functioning as hydrogenation catalysts for nitrite, nitrate, and/or nitrogen oxides. This idea is also supported by the observed decrease in nitric oxide concentration to zero in the experiments using the supported rhodium catalysts (e.g., 5% Rh/C and 5% Rh/ Al2O3), which also produced significant amounts of ammonia. A question of considerable interest is whether nitrite or nitrate is the source of the ammonia nitrogen. In this connection, an NCAW feed simulant prepared according to the standard protocol but leaving out both nitrite and nitrate (UGA-10XNN′1) was treated with formic acid in the presence of sodium nitrate as well as 5% rhodium on charcoal (Figure 4). The amount of ammonia produced in this experiment (1.6 mmol) was essentially identical to the amount of ammonia produced in an experiment using the complete NCAW feed simulant, containing nitrite (Figure 1). No nitrogen oxides were observed in the experiment with the nitrite-free feed simulant indicating that rhodium cannot catalyze the formic acid reduction of nitrate to N2O or NO. Two other experiments indicate the role of supported rhodium metal catalysts in ammonia production from nitrogen-oxygen species and formic acid:

5HCO2H + NO3- f NH3 + HCO2- + 4CO2 + 3H2O (2) The observed final pH of 6.9 is outside the range of the formic acid/formate buffer (pK of HCO2H ) 3.7) but in the range of the carbonic acid/bicarbonate buffer (pK1 of H2CO3 ) 6.4). Thus, some of the carbon dioxide liberated by the formic acid forms the bicarbonate ion and thus cannot be measured in the gas phase. These preliminary observations suggest that ammonia production by the rhodium-catalyzed reduction of nitrite or nitrate ion by formic acid is highly dependent on the form of the rhodium. Thus, highly dispersed rhodium on alumina or charcoal is a very active catalyst whereas soluble rhodium compounds (such as the nitrorhodium complexes obtained from rhodium chloride and nitrite ion) are much less active catalysts. Furthermore, rhodium co-precipitated as Rh2O3‚ xH2O with the iron fraction in the feed simulant UGA-13+Rh1 is not an active catalyst for ammonia production even though the hydrous rhodium oxide is presumably dispersed in the other insoluble metal hydrous oxides. Ammonia Monitoring of Runs with Supported Rhodium Catalysts. The initial studies were performed using the standard NCAW feed simulant and 5% rhodium on charcoal or alumina since the ammonia analyses summarized in Table 2 suggested that these supported rhodium metal catalysts produced the largest amounts of ammonia. Figures 1 and 2 show the production of H2, NH3, N2O, NO, and CO2 from the full NCAW feed simulant UGA-12M1 upon treatment with 88

986

9


was added to the aluminum nitrate before precipitation of the aluminum hydroxide according to the standard procedure (3). The presence of rhodium(III) in the aluminum hydroxide precipitate was indicated by the tan color of the precipitate and a colorless supernatant liquid. Treatment of this feed simulant with 88 wt % formic acid at 90 °C in the presence of sodium nitrate gave negligible amounts of ammonia and 0.4 mmol of H2. This experiment suggests that supported rhodium must be the free metal (i.e., in the zero oxidation state) to exhibit high activity for the reduction of nitrate to ammonia by formic acid. (2) Negligible amounts of ammonia were observed when the same UGA-10XNN′1/NaNO3 mixture was treated with 88 wt % formic acid at 90 °C in the presence of soluble rhodium trichloride rather than the 5% rhodium on carbon used in the experiment depicted in Figure 4. Important conclusions suggested by the experiments outlined above concerning the rhodium-catalyzed reduction of nitrogen-oxygen compounds to ammonia by formic acid include the following: (1) Nitrate rather than nitrite is the principal source of ammonia. (2) Ammonia production appears to occur at the expense of hydrogen production.

FIGURE 3. Treatment of 50 mL of the NCAW feed simulant UGA12M1, 10 mL of water, and 14 mg of RhCl3‚3H2O with 85 mmol of 88 wt % formic acid at 90 ( 1 °C. This experiment produced 0.4 mmol of NH3 and 0.9 mmol of H2 at the end.

FIGURE 4. Treatment of 50 mL of the nitrate- and nitrite-free NCAW feed simulant UGA-10XNN′1, 10 mL of water, 0.5 g of NaNO3, and 0.118 g of 5% Rh/C with 85 mmol of 88 wt % formic acid at 90 ( 1 °C. This experiment produced 1.6 mmol of NH3 but no nitrogen oxides or H2 at the end. (1)The usual protocol for the preparation of the nitriteand nitrate-free NCAW feed simulant (UGA-10XNN′1) was followed except that 0.284 g of hydrated rhodium trichloride

(3) Supported rhodium metal catalysts are more active than rhodium in any other form, suggesting that ammonia production involves heterogeneous rather than homogeneous catalysis. Studies with a Nitrogen-Free Feed Simulant. The protocol for the preparation of UGA-12M1 and related feed simulants (3) uses nitrates as metal sources so that such feed simulants contain significant quantities of nitrate even if an alkali metal nitrate is not added later since residual nitrate remains even after thorough washing of the metal hydroxide/ oxide precipitates. The presence of nitrate in the feed simulant complicates nitrogen balance studies such as those relating to the origin of ammonia. Our standard protocol for the preparation of a nitrate-free (as well as nitrite-free) NCAW feed simulant has therefore been modified to use metal chlorides rather than metal nitrates in equivalent quantities to those given in Table 1 and leaving out the NaNO2 and NaNO3. In connection with this modified protocol, we were concerned initially whether the KMnO4/MnCl2 reaction would be a satisfactory method for the preparation of nitrate-free and organic-free MnO2 because of the possibility of permanganate oxidation of MnCl2 to elemental chlorine. However, no evidence for the formation of Cl2 was observed during the feed simulant preparation process, apparently because the pH remains far too high for the oxidation of Cl- to Cl2 by MnO4- or MnO2. One liter of a carbonate-free version of this nitrogen-free NCAW feed simulant (designated as UGA9M2XC) was prepared for nitrogen balance studies. The experiments with this nitrogen-free feed simulant are summarized in Table 3. In all of these experiments 50 mL of UGA-9M2XC and 10 mL of water were treated with 84 mmol of 88 wt % formic acid at 90 ( 1 °C in the presence of the indicated noble metal derivative and NaNO2 and/or NaNO3 where indicated. In addition, Figure 5 compares the formic acid treatment of this feed simulant in the presence of 5% rhodium on carbon and added NaNO2 or added NaNO3. Both of these experiments are seen to produce ammonia, but the experiment with added nitrate produces more ammonia (1.2 mmol of NH3) than the experiment with added nitrite (0.6 mmol of NH3). Furthermore, in the experiment with added nitrate, the formation of ammonia began almost immediately upon addition of the formic acid at the reaction temperature whereas in the experiment with added nitrite, formation of ammonia began only after reaction of formic


9

987

TABLE 3. Studies Using Nitrogen-Free and Carbonate-Free NCAW Feed Simulant UGA-9M2XC additive (g)

noble metal (mg)

H2 (mmol)

CO2 (mmol)

N2O (mmol)

NO (mmol)

NH3 (mmol)

final pH

none none none NaNO3 (1.9) NaNO3 (1.9) NaNO3 (1.9)a NaNO3 (1.9) NaNO3 (1.9)a NaNO3 (1.9) NaNO2 (1.5) NaNO2 (1.5) NaNO2 (1.5)

none RhCl3‚3H2O (14) 5% Rh/C (116) 5% Rh/C (118) RhCl3‚3H2O (14) RhCl3‚3H2O (14) 5% Pd/C (120) 5% Pd/C (120) PdCl2 (13) 5% Rh/C (118) RhCl3‚3H2O (14) 5% Pd/C (120)

0.0 0.0 0.0 0.1 0.4 0.4 0.4 0.8 0.3 0.0 1.5 0.0

3.1 4.3 12.0 16.7 6.5 5.5 33.1 33.6 18.7 19.4 15.3 15.2

0.0 0.0 0.0 0.0 0.0 0.0 3.7 3.5 1.2 6.9 5.8 6.1

0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 4.5 3.3

0.0 0.0 0.0 1.2 0.0 0.0 1.4 1.4 0.2 0.6 0.1 0.2

2.9 3.0 3.9 4.0 3.2 3.0 5.0 5.0 4.0 4.2 3.7 3.8

a In this experiment, the NaNO was added to the feed simulant at least 12 h before the experiment in order to allow time for adsorption of nitrate 3 ion onto the solid phase before initiation of the reaction.

TABLE 4. Carbon Balance in the Treatment of NCAW Feed Simulants with 88 wt % Formic Acid at 90 °C in the Presence of Rhodium Catalysts sources of CO2

5% Rh/C Figure 1

5% Rh/Al2O3 Figure 2

RhCl3‚3H2O Figure 3

carbonate (1 mmol of CO2/mmol of CO32-) MnO2 (1 mmol of CO2/mmol of MnO2) 2NO2- f N2O (2 mmol of CO2/mmol of N2O) HCO2H f H2 (1 mmol of CO2/mmol of H2) NO3- f NH3 (4 mmol of CO2/mmol of NH3) total mmol of CO2 expected mmol of CO2 observed excess CO2 produced

6.2 1.5 13.4 0.2 6.4 27.7 30.4 2.7

6.2 1.5 13.9 0.3 4.8 26.7 30.3 3.6

6.2 1.5 11.6 0.9 1.6 21.8 20.4 -1.4

acid with nitrite ion was complete (Figure 5). This suggests that the nitrate ion produced by disproportionation of nitrous acid according to eq 3 is the source of ammonia in the experiment in which nitrogen is only added in the form of nitrite:

3HNO2 f 2NO + HNO3 + H2O

(3)

Figure 6 compares the formic acid treatments of the UGA9M2XC feed simulant at 90 °C using 5% Pd on carbon as the catalyst and either NaNO2 or NaNO3 as the nitrogen source. The reaction with NaNO3 generated considerable amounts of NH3, H2, and N2O whereas the reaction with NaNO2 generated no hydrogen and a much smaller amount of ammonia. The following additional points are of interest concerning the experiments summarized in Table 3: (1)If a supported rhodium catalyst is used, nitrogen oxides are observed only in experiments with added nitrite. The N2O remains until the end of the experiment whereas the initially produced NO is consumed during the experiment. The maximum amount of NO observed during this experiment was 7.3 mmol. The generation of N2O from nitrate in the presence of formic acid does not appear to be catalyzed by rhodium. (2) Insignificant H2 is produced in the absence of nitrite even in the experiment with RhCl3‚3H2O in accord with the requirement of added nitrite for the rhodium-catalyzed decomposition of formic acid to H2 + CO2 (eq 1) indicated in previous work (2, 3). (3) Palladium catalysts, unlike rhodium catalysts, catalyze the reduction of nitrate to N2O as well as NH3. Palladium(II) chloride is a considerably less active catalyst precursor than 5% Pd/C for the production of H2, N2O, and NH3 consistent with the relative activity of Pd catalysts for formic acid transfer hydrogenolysis of aryl halides (4).

988

9


(4) Addition of the NaNO3 to the feed simulant 1 day in advance of beginning the reaction in order to allow more time for adsorption onto the solid phases had no significant effect on the production of H2 or NH3 in experiments using either RhCl3‚3H2O or 5% Pd on carbon (Table 3). Carbon Balance Studies. The carbon balance in the ammonia generation experiments is of interest. Table 4 compares the observed amounts of CO2 with the total amount expected from carbonate acidification, MnO2 reduction, H2 generation, reduction of nitrite to N2O, and reduction of nitrate to ammonia upon treatment of the metal nitrate-based NCAW feed simulant with formic acid in the presence of various rhodium catalysts. The experiments producing relatively large amounts of ammonia are seen to produce excess CO2 even assuming the production of 4 mmol of CO2 for each mmol of NH3 as required by eq 2. A possible source of this excess CO2 is partial reduction of the considerable amounts of iron(III) in the feed simulant to iron(II) under the conditions where nitrate is reduced to ammonia. This hypothesis is supported by the fact that the metal hydroxide precipitate obtained upon the addition of base to samples for ammonia analysis appeared more green and less brown in samples containing appreciable amounts of ammonia, suggesting that some reduction of iron(III) to iron(II) might accompany ammonia production. Further experiments were performed to demonstrate that Fe(III) reduction contributes to the CO2 production. Some CO2 balance observations were also made in the experiments using the NCAW nitrogen-free and carbonatefree feed simulant UGA-9M2XC summarized in Table 3. In these experiments, the CO2 produced (3.1 mmol) in the absence of any additives must come from formic acid reduction of some of the metals in the feed simulant since this feed simulant contains no carbonate. Reduction of the 1.55 mmol of MnO2 present in 50 mL of UGA-9M2XC according to

a

b

c

FIGURE 6. Comparison of the effects of added NaNO2 and NaNO3 in the treatment of 50 mL of the nitrogen- and carbon-free feed NCAW simulant UGA-9M2XC, 10 mL of water, and 120 mg of 5% Pd/C with 85 mmol of 88 wt % formic acid at 90 ( 1 °C. The experiment with added NaNO2 produced 0.2 mmol of NH3 and insignificant H2 at the end whereas the experiment with added NaNO3 produced 1.4 mmol of NH3 and 0.4 mmol of H2 at the end. acid reduction of iron(III) to iron(II) can be catalyzed by the addition of rhodium. The feed simulant is 0.45 M in iron(III), which means that the 50 mL of UGA-9M2XC contains 22.5 mmol of Fe(III), which would give 11.2 mmol of CO2 according to FIGURE 5. Comparison of the effects of added NaNO2 and NaNO3 in the treatment of 50 mL of the nitrogen- and carbon-free NCAW feed simulant UGA-9M2XC, 10 mL of water, and 118 mg of 5% Rh/C with 85 mmol of 88 wt % formic acid at 90 ( 1 °C. The experiments with added NaNO2 and NaNO3 produced 0.6 and 1.2 mmol of NH3, respectively, at the end.

MnO2 + 3HCO2H f CO2v + 2H2O + Mn2+ + 2HCO2(4) should give only 1.55 mmol of CO2, suggesting that the remainder of the observed 3.1 mmol of CO2 must come from the reduction of Fe(OH)3. However, the amount of CO2 increases markedly in the presence of added rhodium, and the absence of any carbon sources other than the elemental carbon support of the 5% Rh/C suggesting that the formic

2Fe3+ + HCO2H f 2Fe2+ + CO2 + 2H+

(5)

The 12.0 mmol of CO2 observed in the experiment with 5% Rh/C but no added nitrogen species is very close to the 12.7 mmol of CO2 expected for complete reduction of 1.55 mmol of MnO2 and 22.5 mmol of Fe(OH)3. The predicted amount of CO2 from the experiment in Table 3 with a 5% Rh/C catalyst and added nitrate from formic acid reduction of MnO2 to Mn2+, Fe3+ to Fe2+, and nitrate to ammonia is 17.6 mmol (Table 5) as compared with the 16.7 mmol of CO2 observed in this experiment. The predicted amount of CO2 from the experiment in Table 3 with the 5% Rh/C catalyst and added nitrite from formic acid reduction of MnO2 completely to Mn2+ and Fe3+ completely to Fe2+, the observed 7.3 mmol of NO completely to N2O, the production of the remaining 3.3 mmol of N2O by


9

989

TABLE 5. CO2 Balance in Titration of UGA-9M2XC with Formic Acid in the Presence of 5% Rh/C and Nitrate at 90 °C Listed in Table 3 CO2 from reduction of (50 mL) (0.0314 M) ) 1.55 mmol of MnO2 CO2 from reduction of (50 mL) (0.45 M) ) 22.5 mmol of Fe3+ to Fe2+ CO2 from reduction of nitrate to 1.2 mmol of ammonia

1.5 mmol 11.3 mmol 4.8 mmol

total expected CO2

17.6 mmol

TABLE 6. CO2 Balance in Titration of UGA-9M2XC with Formic Acid in the Presence of 5% Rh/C and Nitrite at 90 °C Listed in Table 3 CO2 from reduction of (50 mL) (0.0314 M) ) 1.55 mmol of MnO2 CO2 from reduction of (50 mL) (0.45 M) ) 22.5 mmol of Fe3+ to Fe2+ CO2 from reduction of 7.3 mmol of NO to 3.6 mmol of N2O CO2 from reduction of nitrite to (6.9-3.6) ) 3.3 mmol of N2O CO2 from reduction of nitrate to 0.6 mmol of ammonia

1.5 mmol 11.3 mmol 3.6 mmol 6.6 mmol 2.4 mmol

total expected CO2

25.4 mmol

TABLE 7. Formation of Ammonia and Hydrogen in Limited Component Experiments additive (g)

noble metal (mg)

H2 (mmol)

CO2 (mmol)

N2O (mmol)

NO (mmol)

NH3 (mmol)

final pH

none NaNO3 (0.5) Fe/NaNO3a NaNO3 (0.5) NaNO2 (1.5) NaNO2 (1.6) NaNO2 (1.5) NaNO2 (1.5) NaNO3 (1.9) NaNO3 (1.0) NaNO3 (0.5) NaNO3 (0.2) none NaNO2 (1.5) NaNO2 (1.5) NaNO3 (1.9) none

5% Rh/C (112) 5%Rh/C (115) 5%Rh/C (115) 5%Rh/Al2O3 (115) 5%Rh/Al2O3 (117) 5% Rh/C (118) 5% Pd/C (120) 5% Pd/Al2O3 (120) 5% Pd/C (120) 5% Pd/C (120) 5% Pd/C (120) 5% Pd/C (120) 5% Pd/C (120) 5% Ru/Al2O3 (110) 5% Ru/C (110) 5% Ru/C (110) 5% Ru/C (110)

1.0 0.2 0.0 0.1 1.4 0.2 0.0 1.0 0.0 0.6 1.1 4.4 3.8 0.0 0.0 0.0 0.2

1.1 1.5 13.8 0.6 15.1 17.1 10.3 12.2 36.1 7.3 8.0 4.9 3.5 6.7 6.2 0.6 0.4

0.0 0.0 0.0 0.0 5.1 5.8 6.7 6.0 1.6 0.0 0.0 0.0 0.0 4.5 4.3 0.0 0.0

0.0 0.0 0.0 0.0 2.3 0.0 2.4 2.5 0.0 0.0 0.0 0.0 0.0 6.5 6.1 0.0 0.0

0.0 0.2 0.3 0.0 1.9 1.5 0.1 0.4 2.9 0.8 0.7 0.1 0.0 0.0 0.0 0.0 0.0

1.8 1.8 3.7 1.9 3.2 3.2 3.2 3.1 7.0 2.2 2.3 1.9 1.8 3.1 3.1 1.8 1.8

a 50 mL of a 0.45 M aqueous slurry of Fe(OH) , 10 mL of water, 0.5 g of sodium nitrate, and 115 mg of 5% rhodium on carbon were used for this 3 experiment.

nitrite reduction, and the reduction of nitrate to the observed 0.6 mmol of ammonia is 25.4 mmol (Table 6). Comparison of this predicted 25.4 mmol with the observed 19.4 mmol suggests that reduction of Fe3+ to Fe2+ is incomplete in the presence of nitrite. This is scarcely surprising in view of the oxidizing properties of nitrous acid according to the following half-reaction (5):

HNO2 + H+ + e- ) NO + H2O

E° ) +1.00 (6)

Limited Component Experiments. A series of limited component experiments was performed in order to gain a greater insight into the factors affecting ammonia formation. In these experiments, solutions of the additives of interest in 60 mL of water in the presence of the noble metal derivative of interest were treated at 90 ( 1 °C with 84 mmol of 88 wt % formic acid using the syringe pump. Supported noble metal catalysts were used for these experiments since the results from previous experiments discussed above suggest that supported noble metal catalysts are more active than soluble noble metal catalysts for formic acid reduction of nitrate and/ or nitrite to ammonia. The final compositions of the reaction mixtures (typically after 400 minutes reaction time) are summarized in Table 7. Two of the reactions of nitrate with formic acid in the presence of Pd/C are depicted in Figure 7. The following points are of interest concerning these limited component experiments:

990

9


(1) The CO2 produced in the Rh/C experiment with Fe(OH)3 indicates essentially complete formic acid reduction of the 22.5 mmol of Fe3+ present to Fe2+ probably catalyzed by the supported rhodium catalyst. The final color of this reaction mixture was pale green, consistent with the high conversion of Fe3+ to Fe2+. (2) Substantial amounts of hydrogen are produced in the Rh/Al2O3 experiment with added NaNO2 but not in the experiments with NaNO3. This again indicates the role of nitrite as a promoter for the rhodium catalysts for hydrogen generation from formic acid as suggested from previous work (2, 3). (3) Although RhCl3‚3H2O does not produce NH3 from nitrite at least in feed simulant experiments, Rh/C produces a significant amount of NH3 (1.5 mmol) but very little H2 (0.2 mmol) from nitrite. The production of H2, N2O, and NH3 accounts for essentially all of the CO2 since 0.2 + (5.8 × 2) + (1.5 × 3) ) 16.3 mmol as compared with the observed 17.1 mmol. (4) Pd/Al2O3 is more efficient than Pd/C at catalyzing the formic acid reduction of nitrite to ammonia. In addition, Pd/Al2O3 but not Pd/C catalyzes formic acid decomposition to hydrogen in the presence of nitrite. (5) The experiment with 1.9 g of NaNO3 using Pd/C had to be stopped after 138 min after the addition of only 61.3 mmol of formic acid because of excessive pressure (e.g., 36.1 mmol of CO2) and produced ∼5.4 mmol of N2 in addition to the gases listed above. The production of relatively large

TABLE 8. Effect of Nitrate Concentration on the Pd/C Catalyzed Reduction of Nitrate with Formic Acida NaNO3 (mmol)

H2 (mmol)b

CO2 (mmol)

N 2O (mmol)

N2 (mmol)

NH3 (mmol)

final pH

0 2.4 5.9 11.7 22.3

3.8 (12.2) 4.4 (11.4) 1.1c 0.6 (17.7)d 0e

3.5 4.9 8.0 7.3 36.4

0 0 0 0 1.6

0 0 0 0 5.3

0 0.1 0.7 0.8 2.9

1.8 1.9 2.3 2.2 7.0

a All experiments were performed using 60 mL of water and 120 mg of 5% Pd/C. b The rate of H2 production is given in parentheses in mmol/ day. c Maximum H2 produced ) 2.1 mmol at 80 min. Too few points were obtained to give a reliable H2 production rate. Rate of NH3 increase ) 2.7 mmol/day. Rate of H2 decrease during NH3 production ) 4.3 mmol/day. d Maximum H2 produced ) 1.2 mmol at 69 min. Rate of NH3 increase ) 4.1 mmol/day. Rate of H2 decrease during NH3 production ) 2.7 mmol/day. e The reaction was stopped after 138 min due to excessive pressure. No H2 was observed during any stage of the process.

at least under the reaction conditions of interest in this work. Supported ruthenium catalysts are only active for formic acid decomposition to hydrogen in the absence of both nitrite and nitrate (Table 7). The effect of NaNO3 concentration on the Pd/C catalyzed reduction of nitrate with formic acid was investigated at a range of nitrate concentrations (Table 8). At intermediate nitrate concentrations (5.9 and 11.7 mmol NaNO3/60 mL), some of the initially produced H2 was subsequently consumed suggesting a metal-catalyzed hydrogenation of nitrate to ammonia according to

NO3- + 4H2 f NH3 + 2H2O + OH-

(7)

The final pH in such experiments was 2.0 ( 0.3, which is close to the reported range for the hydrogenation of nitrate ion in the presence of a 5% Pd/C catalyst in aqueous phosphoric acid (6-8). Under such conditions, nitrate is reported to be hydrogenated not only to ammonia but also to hydroxylamine according to

NO3- + 3H2 f NH2OH + H2O + OH-

(8)

This suggests that hydroxylamine could be a byproduct on the formic acid treatment of nuclear wastes containing nitrate or even nitrite ion. FIGURE 7. Effect of NaNO3 concentration in its reaction with formic acid in the presence of 5% Pd/C. The reaction with 1.9 g of NaNO3 was stopped at 138 min because of excessive pressure and produced 2.9 mmol of NH3 at the end. In the experiment with 0.5 g of NaNO3, the rate of NH3 production corresponded to 62% consumption of the H2 initially produced. amounts of NH3 (2.9 mmol) led to an unusually high final pH (7.0). Use of a smaller amount of NaNO3 (0.5 g) allowed the reaction to be carried out to completion and led to the generation of considerable amounts of H2. The maximum amount of H2 generated in this experiment with 0.5 g of NaNO3 was 2.1 mmol, but the amount of H2 present in this experiment decreased to 1.1 mmol as 0.7 mmol of NH3 was generated in the later stages of the reaction. This suggests that the NH3 arises from H2 reduction of a higher oxidation state nitrogen intermediate. Furthermore, a comparison of the otherwise identical experiments using 1.9 and 0.5 g of NaNO3 suggests that, if the nitrate concentration is high enough, all of the H2 generated by formic acid decomposition is used for the reduction of nitrate to NH3 rather than being evolved as H2 gas. (6) Supported ruthenium catalysts are not active for the formic acid reduction of either nitrite or nitrate to ammonia,

Discussion The work summarized in previous papers (2, 3) suggests that the predominant source of hydrogen from the formic acid treatment of nuclear waste simulants containing nitrite and nitrate is a nitrite-promoted homogeneous rhodium-catalyzed process. A mechanism has been proposed for this process based on nitrorhodium(III) intermediates (3). In the case of ammonia generation from the formic acid treatment of the same nuclear waste feed simulants, the predominant process appears to be heterogeneous catalysis involving supported rhodium or palladium metal catalysts. Thus, 5% rhodium metal supported on carbon or alumina is a considerably more active catalyst for ammonia production than soluble rhodium trichloride. Nitrate rather than nitrite appears to be the major source of ammonia as shown using experiments in which either nitrate or nitrite is added to a special nitrogen-free feed simulant. The following observations indicate the close similarity between the noble metal-catalyzed formic acid reduction of nitrate to ammonia on one hand and organic hydrogenation processes (9-13) on the other hand: (1) The supported noble metal catalysts, namely, supported rhodium and palladium catalysts, are the same as those that are most active for organic hydrogenations under the mildest conditions.


9

991

(2) Ammonia production from the noble metal-catalyzed formic acid reduction of nitrate appears to occur at the expense of hydrogen production. For example, 5% Rh/C and Rh/Al2O3 are good catalysts for ammonia production but poor catalysts for hydrogen production whereas RhCl3 is a good catalyst for hydrogen production but a poor catalyst for ammonia production. Furthermore, in some of the experiments with supported palladium catalysts, some of the initially produced H2 in the closed system is consumed as ammonia is produced. (3) Formic acid is known (14-16) to be a transfer hydrogenation reagent for conversion of nitro compounds to amines in the presence of noble metal catalysts according to (R ) alkyl or aryl group)

(3) (4) (5) (6) (7) (8) (9) (10)

RNO2 + 3HCO2H f RNH2 + 2H2O + 3CO2v

(9)

Palladium on charcoal is among the most active catalyst for this reaction (9).

(11) (12) (13)

Acknowledgments We are indebted to the U.S. Department of Energy for support of this work at the University of Georgia through the Westinghouse Hanford Corporation and the Pacific Northwest National Laboratory operated by Battelle Memorial Institute.

Literature Cited (1) Choppin, G. R.; Rydberg, J. Nuclear Chemistry: Theory and Applications; Pergamon: Oxford, 1980; p 505. (2) King, R. B.; King, A. D., Jr.; Bhattacharyya, N. K.; King, C. M.; Landon, L. F. In Chemical Pretreatment of Nuclear Waste for

992

9


(14) (15) (16)

Disposal; Schulz, W. W., Horwitz, E. P., Eds.; Plenum: New York, 1994; pp 101-113. King, R. B.; Bhattacharyya, N. K.; Wiemers, K. D. Environ. Sci. Technol. 1996, 30, 1292. Anwer, M. K.; Sherman, D. B.; Roney, J. G.; Spatola, A. F. J. Org. Chem. 1990, 54, 1283. Huheey, J. E. Inorganic Chemistry: Principles of Structure and Reactivity Harper and Row: New York, 1983; Appendix F. Lopatin, V. L.; Lezhneva, K. A.; Kul’kova, N. V.; Temkin, M. I. Kinet. Katal. 1979, 20, 373; Engl. Transl. 1979, 20, 302. Kinza, H. Kinet. Katal. 1982, 23, 1511; Engl. Transl. 1982, 23, 1290. Temkin, M. I.; Kul’kova, N. V.; Lopatin, V. L. Kinet. Katal. 1982, 23, 1514; Engl. Transl. 1982, 23, 1294. Augustine, R. L. Catalytic Hydrogenation; Dekker: New York, 1965. Rylander, P. N. Catalytic Hydrogenation over Platinum Metals; Academic Press: New York, 1967. Freifelder, M. Practical Catalytic Hydrogenation; Wiley-Interscience: New York, 1971. Rylander, P. N. Hydrogenation Methods; Academic Press: London, 1985. C ˇ erven ´ y, L., Ed. Catalytic Hydrogenation; Elsevier: Amsterdam, 1986. Entwistle, I. D.; Jackson, A. E.; Johnstone, R. A. W.; Telford, R. P. J. Chem. Soc. Perkin Trans. 1 1977, 443. Cortese, N. A.; Heck, R. F. J. Org. Chem. 1977, 42, 3491. Ram, S.; Ehrenkaufer, R. E. Tetrahedron Lett. 1984, 25, 3415.

Received for review April 12, 1996. Revised manuscript received October 25, 1996. Accepted November 19, 1996.X ES960335+ X

Abstract published in Advance ACS Abstracts, February 1, 1997.

Noble Metal-Catalyzed Ammonia Generation by Formic Acid

Recommend Documents