REDUCTION OF NITRITE
2131
BY ?dOLYBDENUM(V)
The Reduction of Nitrite by Molybdenuim(V)
by Jean A. Frank2 and Jack T. Spence Chemistry Department, Utah State University, Logan, Utah
(Received January 16, 1964)
The kinetics and mechanism of the reduction of nitrite by YIo(V) in aqueous solutioii have been investigated. It was found that NO was produced in stoichioiiietric amounts during the reduction. The reaction was determined to be first order with respect to nitrite and hydrogen ion and zero order with respect to Mo(V) in the pH range 5.0-7.5. A negative salt effect on the reaction rate was observed. The activation energy, free energy, and entropy for the reaction have been determined. A mechanism, involving the formation of S O + from HK02 as the rate-determining step, has been developed.
Molybdenum has been shown to participate in thch enzymatic reduction of nitrate to nitrite.3 The subsequent biological reduction of nitrite to hyponitrite is apparently mediated by an iron and copper containing e i i ~ y i n e but , ~ little is known concerning its properties It is perhaps significant that a nitrite-reducing system obtained from bacteria has been reported to coiivert nitrite to nitric oxide,6 but whether or not a metal i s involved is not yet known. Recently, considerable interest has been shown in the ability of niolybdenuiii to catalyze the reduction of nitrate by various reducing agents in strong acid.6-'' It has been reported that Mo(V) will reduce hydroxylainine to ammonia in acidloand that it acts as a catalyst for the reduction of hydroxylamine by stannous chloride, again under strongly acidic conditions. l1 No reports have appeared, however, concerning the reduction of nitrite by Mo(V). There have been a number of studies concerning the reduction of nitrite or nitrous acid. The reactiom with ferrous ion'* and with nietallic ~ o p p e r l have ~ ) ' ~both been investigated. In addition, the reductions by sulfite, oxalate, formate, sulfamate, iodide, and several other species have been reported.15 Of particular interest, because of the nature of the intermediate, is the study of Anbar and Taube of the rate of exchange of oxygen between nitrike and water.I6 Preliminary work in this laboratory, undertaken as: part of a study of possible niodels for nitrate reductase, indicated that IIo(V) reduces nitrite to nitric oxide at a considerably faster rate than it reduces nitrate to nitrite." Because of its relation to nitrate reduction
and its possible biochemical significance, the reaction was investigated in detail.
Experimental MateriaEs. S~Iolybdenum(V)stock solutions in HCl were prepared and standardized as previously reported. l 8 The chloride ion concentrations and the ionic strength of the solutions were maintained constant by the addition of the proper amount of a stock solution of NaCI, prepared from Baker and Adamson reagent grade NaCl. (1) Journal Paper No. 384, Utah State Agricultural Experiment Station. (2) Abstracted from the thesis submitted by J. 4.Frank in partial fulfillment of the requirements for the 1'h.D. degree, Utah State University, 1963. (3) D. J. D. Nicholas and A. Nnson, J . Biol. Chem., 211, 183 (1957). (4) D. J. D . Nicholas, Satztre, 179, 800 (1957). (5) T. Yamanaka, A. Otaand, and K. Okunuki, Bioehim. Riophys. Acta, 53, 294 (1961). (6) G. P. Haight, J r . , Acta Chem. Scand., 15, 2012 (1961). (7) G. P. Hnight, Jr., P. Mohliner, and A. Kat%;,ibid.,16, 221 (1962). (8) G. P. Hnight, Jr., and A. Kntz, ibid., 16, 650 (1962). (9) J. M. Kolthoff and J. Hodnrx, J . Electroanal. Chcm., 5 , 2 (1963). (10) G. P. Haight, Jr.. and A. C. SuTift, J . Phys. Chem.. 6 5 , 1921 (1961). (11) G. P. Haight, Jr., and C . V. Frankenberg. Acta Chem. Seand., 15, 2026 (1961). (12) E. Abel, H. Schmitl, and I. Pollak, Monatsh., 69, 125 (1936). (13) Ti. Komuro and R. Kato, Bttlt. Aichi Gakugei. Chic., 2 , 43 (1953). (14) T.Komuro and K . Kato, Chem. Abstr., 48, 1122 (1952). (15) T. A. Turney nnd G. A. Wright, Chem. Rei,., 59, 497 (1959). (16) M. Anbar and H . Tnube, J . Am. Chem. SOC..76, 6243 (1954). (17) J. T. Spence and J. A. Frank, ibid., 8 5 , 116 (1963). (18) J. T. 8pence and G. Kallos, Inorg. Chem., 2 , 710 (1963).
Volume 68, S u m b e r 8
Augtis t , 1064
2132
In order to obtain reproducible results, the stock solutions of I\lo(V) were aged for 12 days before use. Standard solutions of NaNOzwere prepared froin Baker and Adamson reagent grade material that had been dried in uucuo over Pz05. Before use these solutions were analyzed For nitrite concentration by the method of Shinn.lQ Phosphate buffers were prepared from reagent grade KI-I~POIand NaOH and adjusted to the proper pH and ionic strength. Helium, used for flushing the reaction vessels, was obtained from Jlatheson Co. Its purity (99.99%) was found to be sufficient and no deoxygenation treatment was required. Xitric oxide was obtained in lecture bottles from Matheson Co. All other chemicals used were reagent grade. Methods. A silica spectrophotometric cell, which could be evacuated, was joined to a specially constructed reaction vessel for measurement of Mo(V) concentrations. This was essential because AIo(Tr) is easily oxidized by atornospheric oxygen. The buffer and sodium nitrite solutions were added to the vessel and flushed with helium for 30 min. The proper amount of Mo(V) stock solution, which had also been deoxygenated, was then added and mixed with a helium stream. The moment of mixing was recorded as zero time. A stopcock a t the bottom of the vessel was opened to allow the solution to flow into the evacuated cell. The cell was then sealed with an oxygen torch and placed in a constant temperature bath. The concentration of Mo(V) was determined by removing the cell momentarily from the thermostat and reading its absorbance at 289 mp in a Beckman DU spectrophotometer against a blank containing buffer and sodium nitrite. For following nitrite concentrations with time, helium was bubbled continuously through the reaction vessel under a slight positive pressure and samples of the proper volunie were withdrawn through a rubber diaphragm using a hypodermic syringe. The sample was then analyzed for nitrite.IQ When a run had a large excess of nitrite, the sample was prepared as above; but the sealed cell was allowed to remain in the spectrophotometer compartment for the entire reaction, due to the increased rate. The temperature in the compartment was maintained constant by the use of thermospacers. Unless one reactant was in large excess, duplicate runs were made, one being analyzed for lIo(V) and the other for nitrite. Runs were made at five pH values, from 4.68 to 7.50, to determine the effect on the rate. Lower pH values were not used because of the rapidity of the reaction and because of the autodecomposition of nitrite, which becomes appreciable in acidic solutions. Higher values were not used because of the great decrease in reaction rate and because of the tendency of l\lo(V) to precipiThe Journal of Physical Chemistry
JEAN A. FRANK AND JACK T. SPENCE
tate as the hydroxide in basic solution. Kinetic measurements were aIso made a t four temperatures, 20, 35, 42, and 50°, to obtain the activation energy for the reaction. Runs which contained initially both the products, Mo(V1) and YO, were made in order to investigate possible effects on the rate. Runs were made containing 1 k? Sa2S04in addition to the reactants to investigate ionic strength effects. Gas measurements to determine the stoichiometry of the reaction were made with a Warburg constant volunie respirometer, using the procedure of Umbreit, Burris, and StauffernZ0The buffer containing the nitrite was measured into the single arm Warburg flask and the Rlo(V) stock solution was measured into the side arm. The flask was shaken under a flow of helium in the thermostat for 1 hr. and then closed off. Thirty minutes were allowed to establish equilibrium before the initial reading of the manometer was made. The Mo(V) and nitrite solutions were mixed immediately after this reading and the flask shaken in the bath for the desired time. The final reading on the manometer was bhen taken. Identical solutions were prepared and the concentrations of Mo(V) and nitrite were determined with time as described. All pH measurements were made with a Beckman expanded scale pH meter. Rate constants and the activation energy were obtained from the slope of the proper plots, using the method of least squares.
Results Stoichiometry. Previous work, using gas chromatography, had identified the gas produced in the reduction of nitrite by R4o(V) as nitric 0xide.l’ Furthermore, no appreciable amounts of other nitrogen oxides were detected. In the present work, the average value for the number of moles of YO produced for each mole of nitrite consumed, measured at various reaction times, was found to be 1.02 with a standard deviation of =k0.03. In all cases, the concentrations of Mo(V) and nitrite were found to be the same, within experimental error, a t equal reaction times. Kinetics. The over-all order of the reaction was determined from the runs containing equal amounts of Mo(V) and nitrite, a t different total concentrations. It was found that under these conditions the half-lives for each run were the same, indicating an over-all firstorder reaction. When the results were plotted as firstorder reactions, straight lines were obtained for initial (19) M . B. Shinn, I n d . Eng. Ckem., Anal. Ed., 13, 33 (1941). (20) W. W. Umbreit, R. H. Burris, and J. T. Stnuffer, “Manometric Techniques,” 3rd Ed., Burgess Publishing Co., Minneapolis, Minn., 1957.
2133
REDUCTION OF KITRITEBY A'IOLYBDENUM(V)
2.8
4.80
3.2
4.00
3-
2 3.20
2 3.6 3 s"
E;
X
d
3
h
5
H
I
2.40
1.60
4.4
200
400
600
Time, min.
4
8
12
10
20
24
Time, min.
Figure 1. First-order plots for equal concentration runs; -log C ( C = [iMo(V)] =: [NOz-]) is plotted us. time. Numbers refer to runs as listed in Table I. T = 35", pH 6.23, 0.05 M phosphate buffer, fi = 0.77.
Figure 2. Zero-order plots for runs with excess S O 2 - ; concentration of Mo(Y) is plotted us. time. Numbers refer to rune as listed in Table I. T = 35", pH 6.23, 0.05 A4 phosphate buffer, A, = 0.77.
concentrations of 1.80 X lopa to 2.25 X lop4 A4 (Fig. 1). Below this concentration, considerable deviation from the first-order plots was observed ; however, no other simple order equations would fit the data in this region of low concentration. The order with respect to each reactant was determined froin runs in which the concentration of one reactant was in large excess. When Mo(V) was in excess, the same half-life was obtained as iii the equal concentration runs. When nitrite was in excess, the rate was independent of Mo(V) concentration, indicating a reaction first order in nitrite and zero order in Mo(V), within the above conceiitratioii range. The first-order rate constant obtained from the run with excess Mo(V) agreed very well with the first-order rate constaut of the equal concentration run containing the same nitrite concentration. The rate constant obtained from the run with excess nitrite was found also to be in excellent agreement with the other results. (In this case, the zero-order rate constant, k', obtained froin the equation,
rate = k[NOz-] = k', was divided by [NO2-]in order to obtain the first-order coiistaiit, k . ) In the runs with excess nitrite, the plot of concentration of Mo(V) vs. time gave a straight line until the Mo(V) concentration had fallen to approxilnately lop4M , in agreeinelit with the results of the equal concentration runs (Fig. 2). It was observed that the rate mas strongly dependent on pH, increasing rapidly with an increase in acid concentration. The data gave good first-order plots in all cases, indicating that the order did not change with pH. The reaction was found to be first order with respect to hydrogen ion concentration. This was determined by plotting the data according to the equation, rate = kl[NOz-][H+]n, and observing that consistent values of k , mere obtained oiily for n = 1. It was found that the rate of the reaction was unaffected by the presence of the products, N O or Mo(V1). It was found that an increase in ionic strength produced a significant decrease in the rate but did not affect the order. Because of the high ionic strength necessary Volume 68,2Ytimber 8
August, 1964
2134
JEANA. FRAXK ASD JACK T. SPENCE
for the reactions ( p = 0.77) no further attempts were made to st,udy this effect quantitatively. The activation energy for the reaction was determined from the runs at four temperatures a t the same pH, concentration, and ionic strength. The Arrhenius plot for this data is shown in Fig. 3. From the slope of the line the activation energy was determined t,o be 10.6 kcal. From the frequency factor obtained froin the intercept, a value of AS* of - 16.4 e.u. was obtained and A F * was calculated from Eyring's rate equation to be 15.6 kcal.
4.00
3.80
3.60
E
Table I : Rate Constants
Runa
lb 26 3* 4' 5' 6* 7' 8' 9' 1o b 1l b 125 13b 14' 15b 16' 17 18 19 20 21 22 23" 24"
M
s kl x 10moles/l
[hIo(V)11 x 104,
[NO?-lo
rnoles/l
moies/l
PH
OC.
18 0 9 00 6 00 4 50 2 25 0 900 0 450 9 00 18 0 5 12 3 67 17 6 9 00 4.50 9 26 9 00 9 00 4 50 9 00 4 50 9 00 4 50 18 0 9 00
18 0 9 00 6 00 4 50 2 25 0 900 0 450 4 50 0 132 100 00 100 00 17 6 9 00 4 50 9 26 9 00 9 00 4 50 9 00 4 50 9 00
6.23 6 23 6 23 6 23 6 23 6 23 6 23 6 23 6 23 6 23 6 23 7 50 5 90 5 90 5 04 4 68 6 23 6 23 6 23 6 23 6 23 6 23 6 23 6 23
35 35 35 35 35 35 35 35 35 35 35 35 35 35 35 35
x
T,
104,
4 50 18 0 9 00
mln
3 3 3 3 3
24 29 29 31 38
3 3 3 3 3 3 2 3 3
41 23 13 26 10 00 95 21 03
-1
4 79 mean
42 50 35 35 \
1
7 00 mean 1 07 1 34
All runs made in 0.05 ;M phosphate buffer with ,u = 0.77. value of IC, for runs 1-16 with standard deviation. k , = 3.20 & 0.14. Runs 23 and 24 contained 1.00 mole/l. of Na2S04. a
Discussion The stoichiometric data indicate that the over-all reaction is re!atirely simple and can be represented as
+ 4H+ + [L10(1~)]2
----f
2LTo(VI)
I
I
3.1
3.2
i / x~ 103.
I
1
3.3
3.4
Figure 3. Arrhenius plot of temperature dependence of rate constant; log is plotted vs. 1 / T . pH 6.23, 0.05 M phosphate buffer, ,u = 0.77.
In the concentration range from 1.80 X 1WS to 2.25 X ilf the reaction is first order in nitrite, first order in H f , and zero order in lIo(17). The rate-determining step, therefore, must involve the decomposition of either nitrite ion, nitrous acid, or some species foriiied from H+ and KO2- into an intermediate spccies which reacts quickly with h!Io(V) to give the products. (In the pH range studied, the nitrite is present predominantly as nitrite ion since K , for nitrous acid is 4.0 x Since the nitrite analysis measures total nitrite, this is equivalent to measuring the nitrite ion concentration.) This gives the experimental rate law rate
+ 2H20 + 2x0
(The exact species of Mo(V) and Mo(V1) present in aqueous solution are unknown ; but it appears that Mo(1') exists as a dimer21 and that both species occur as oxyanions.) The Journal of Physical Chemiatry
3.20
1 52 mean
20
' Average
2N02-
3.40
=
Isl[H+][NOS-]
(1)
The reactive intermediate cannot be HXO2 siiice the rate of formation of HSOz from H + and NOS- would be considerably greater than the over-all rate of the reaction. Although there is no direct evidence to support (21) G . P. Haight, Jr., J . rnorg. .\Tz(cl. Chem., 24, 663 (1962).
2135
REDUCTION OF NITRITEBY MOLYBDENUM(V)
it, S O + is a reasonable possibility for the intermediate. This species has been proposed for various other reacThe following possiblie tions of nitrous acid. mechanism, based on NO+ as the reactive intermediate, is proposed. 13114,16
H+
+ N02- iz HX02 Ka = [H+][XOz-]/[HN02] (I) HNOz -% KO+
SO+
NO+
+ OH-
+ OH- -%
HYOz
+ nlIo(V) --% NO $- Mo(V1)
(11) (III) (IV)
Sitrous acid, in equilibrium with H+ and NOz-, decomposes in the rate-controlling step (reaction 11) into ?;O+ and OH-. I n a subsequent fast step (reaetion IV) the S O + reacts with Mo(V) to give the products. The rate of the reaction is given by rate = k3 [NO+][Mo(V) ]
(2)
Applying the steady-state condition to [NO+]
d-[N o l + - 0 = k1["02] dt
the dimer, [Mo(V)]~,the same rate expression is obtained, except [;\ilo(Tr) ] is replaced by [ [Mo(V) 1 2 1 . The cause of the deviation froin second-order kinetics when the total Mo(V) concentration drops below M is unknown. It may be that the first term in the denoniinator of eq. 5 becomes significant a t this concentration. At present, however, this cannot be determined since the exact species of Mo(V) involved in the reaction is unknown. Further work to elucidate the nature and reactions of Mo(V) species in dilute solution is planned. The experimental entropy of activation, AS* = -16.4 e.u., is in agreement with the proposed niechanisni since it iniplies a greater degree of order in the transition state than in the reactants. Such a situation would arise because of the separation of charge occurring in the transition state as the neutral molecule decomposes into NO+ and OH-, with a subsequent increase in ordering of solvent niolecules. It should be noted that the experimental AS" was calculated from the experimental rate constant. If the proposed mechanisin is correct, this rate constant ( k ) involves the actual rate constant (ICl) and the ionization constant for "02 (Ka)
-
k = kl/Ku AS*exptl
[NO+] =
ki [HNOZ1 kz [OH-] k3 [Mo(V)1
+
(4)
Replacing [HXO2J and [OH-] from their respective equilibria and substilxting for [ S O + ]in eq. 2
When (k2/k3Ku