19'76 703The Chemistry of Trioxodinitrates. Part 1. Decomposition of SodiumTrioxodinitrate (Angeli's Salt) in Aqueous SolutionBy Martin N. Hughes and Peter E. Wimbledon, Chemistry Department, Queen Elizabeth College, CampdenThe kinetics of decomposition of sodium trioxodinitrate, Na,[N,O,], have been measured over the range pH 1-10a t several temperatures. At pH > 4, the rate-determining step is breakdown of [HN,O,]- to [NO,]- and N,O.The pK for [HN,O,]- a t / = 0.25 rnol dm-3 and 25 "C i s 9.35, and A,,,,. = 237 nm. At lower pH values, the rateincreases with increasing acidity with production of NO, which is the predominant product at pH 2. At these pHvalues, kobe., the measured first-order rate constant, increases with increasing concentration of Na2[N,0,]. This isattributed to a reaction with trace amounts of HNO,, giving NO and reforming [NO,]-.rather than to an enhancedinstability of H2N203. Extrapolation of kobs. to zero [N,0,2-] at pH < 3 gives kobs.* which corresponds to decompo-sition via [ H N,0,] - ; kobs* decreases with decreasing pH, showing that H2N203 is stable compared to [ H N,0,] -in the absence of nitrite and allowing the estimation of pKlCa.3.0. Addition of [NO,]- to Na2[N,03] at pH 5 resultsin the production of NO, the use of [15N02]- showing that this i s not attributable to the disproportionation of HN02and also that both molecules of NO produced in the reaction are derived from the nitrogen atoms of [N,0,]2-.The acid-catalysed HN0,-catalysed reaction a t lower pH obeys the rate equation Rate cc [H+] [HNO,] [H2N,O9],but the value of the third-order rate constant is too high for a diffusion-controlled electrophilic nitrosation reaction.Added ethanol at these pH values has a very marked inhibitory effect on the rate, so it is suggested that formationof NO results from a free-radical chain reaction.Hill Road, London W8 7AHWHILE the structure of the anion of Angeli's salt,Na2[N,0& ;known as sodium hyponitrate, sodiumtrioxodinitrate(II), or sodium oxohyponitrite l] has nowbeen determined2 to be [ON=N0,]2-, there is stilluncertainty over the mechanism of the decompositionof the salt in aqueous solution.The stoicheiometry ofthe reaction is known3'* to be dependent on pH. Di-nitrogen oxide and nitrite ions are formed at high pH,while nitrogen mono-oxide is reported to be the onlyproduct in acidic solutions, although there is someuncertainty over the highest pH at which this occurs.The use of [l5N,O3I2- shows that the nitrite product isderived from the nitrogen atom in [N2Od2- that isbonded to two oxygen atoms, suggesting that one modeof decomposition involves N=N cleavage.Kineticstudies 496 on the decomposition are limited, but it hasbeen suggested that the inhibition of decomposition byadded hydroxide is indicative of the involvement of[HN,O,]- in the rate-determining step. Values forpK, and pK, have been determined' at several ionicstrengths (pK, 2.39, pK, 9.36 at 0.25 mol dm-3 ionicstrength), and it has been observed that decompositionin acidic solutions is fast.We have made a thoroughstudy of the kinetics of decomposition over a wide rangeof pH a i d temperature, together with lSN studies.Since the completion of this work, Bonner and Ravidshave reported a detailed study of the stoicheiometry ofthe reaction using [O15NNOJ2-, together with somekinetic results. In general, our observations are inagreement with those of Bonner and Ravid, but we havemade much more extensive kinetic studies, particularlyat pH (3, and as a result have a different interpretationof the kinetics in more acidic solutions.C. C. Addison, G. A. Gamlen, and R. Thompson, J . Chem.Sac,, 1962, 338.H. Hope and M. R. Sequeira, Inorg. Chem., 1973,12, 286.L. Cambi, Bey., 1936, B69, 2027. * J . Veprek-Siska, V.Pliska, F. Smirous, and F. Visely,Coll. Czech. Chem. Comm., 1969, 24, 687.ti D. K. Hendrickson and W. L. Jolly, Inovg. Chem., 1969, 8,693.EXPERIMENTALAngeli's salt was prepared froin liydroxylamine andethyl nitrate 1 and recrystallised several times from dilutealkali (rather than water, to cut down self decomposition).The purity was checked by analysis (Found: N, 23.0.Calc.: 22.96%) and U.V. spectroscopy (cmx. 7 900 dm3mol-1 cm-1 at 248 nm; lit.,I 8 300). Angeli's salt isoxidised to nitrite quite readily in air, and so was stored inan atmosphere of nitrogen, while all solutions werethoroughly deoxygenated before use. Reproducible kineticresults were obtained for four preparations, but a fifthpreparation gave ca. 30% lower rate constants althoughwith a similar pH dependence.All other reagents were ofthe best grade available.U.V. studies were made in silica cells in Unicam SP 500and SP 800 spectrophotometers. Kinetic runs werestarted by addition of the appropriate amount of solidNa,{N,O,] to thermostatted solutions of appropriate pHand ionic strength. The reaction was followed by with-drawing known volumes of the reaction solution andrunning them into known volumes of 1.0 mol dm-, sodiumhydroxide solution, the final solutions being alkaline.Optical densities were measured a t 248 nm. Initial con-centrations of [N,0,I2- were measured by extrapolation ofthe absorption at 248 nm to zero time. Some runs werefollowed directly in the thermostatted cell compartment ofthe Unicam SP 500 by noting the change in absorptionwith time at 248 nm and pH 10, and a t 237 nm (Amx.for[HN,O,]-) and pH 4-10. Good agreement was obtainedfor the two methods for reactions at pH 2 4 , but only thesampling technique could be used a t pH (4.The gaseous products of the reaction were studied bymass spectrometry (A.E.I. MS 20). Solid Na,[N,O,] wasplaced in one leg of a Y tube and all other components werein the second leg. After several cycles of freezing toliquid-nitrogen temperatures, evacuating, and thawing,for thorough degassing of the solution, the reaction wasstarted (at room temperature) by tipping the Y tube.6 P. A. S. Smith and G. E. Hein, J. Amer. Chem. Soc., 1960,7 P. E. Sturrock, J. D. Ray, and H. R. Hunt, Inorg.Chetu.,* F. T. Bonner and B. Ravid, Inorg. Chew, 1976, 14, 588.82, 6731.1963, 2, 649704 J.C.S. DaltonOn completion of the reaction the solutions were frozen in successive half-lives, and the independence of kobs., thesolid carbon dioxideacetone slush and the relative concen- measured first-order rate constant, of the concentration oftrations of the various products were measured. A similar trioxodinitrate. Many reactions were followed for fourtechnique was used for the 16N studies in which the nitrite- half-lives and the remainder for at least two half-lives.trioxodinitrate reaction was studied by [16N02]- (95% en- Values of hobs. (Table 2) were unaffected by the addition ofriched; British Oxygen Company, ' Prochem '). For ethanol (up to 4% v/v) and were independent of the natureTABLE 1Variation of nitrite production with pH a t 25 "CpH 5.06 * 4.83 * 4.68 4.50 4.46 * 4.03 * 4.14 3.66 3.20 2.70 2.33 1.76A[N0,-]/A[N20,2-J 0.84 0.83 0.73 0.83 0.80 0.76 0.72 0.62 0.44 0.34 0.19 0.10* Acetate buffer, other results in citrate buffer.studies of the gaseous products of the trioxodinitrate-nitrite reaction, thoroughly mixed Ka,[N,O,] and sodiumnitrite were placed in one leg of the tube.Buffers (borax, phosphate, acetate, and citrate) wereprepared according to literature specifications.Values ofpH were checked by the use of a Pye model 290 pH meter.RESULTSU.V. Studies on the Reactiovl Stoicheionzetry.-The reactionwas studied over the full pH range by repetitive scanning ofthe U.V.spectrum. In borate and phosphate buffers (pH7.5-10 and 5.4-7.5 respectively) there was a decrease inthe Na,[N,O,] absorption with time, either of the [N2OS]2-absorption (Amx. 248 nm) or of the [HN,O,]- absorption(Amax. 236-237 nm) as observed by Bonner and Ravid,*while there was an increase in absorption of a band withA,, 212-215 nm due to [NO,]-. A sharp isosbestic pointwas observed at 223-226 nm, together with one between190 and 200 nm. The quotient A[NO,-]/A[N,0,2-] wasca. 0.8 for pH >4.5, although it is unwise t o attach toomuch significance to the deviation of this ratio from unityin view of the experimental uncertainty. In acetate andcitrate buffers (pH 3.8-5.4 and 2.0-3.9 respectively) thesituation is more complicated.The absorption due t o[HN,O,]- decreased with time, but the absorption due t othe production of [NO,]- appeared at higher wavelength,overlapping with the [HN,O,]- absorption. The spectrumof [NO,]- in these buffers is similar and we attribute thisbehaviour to formation of nitrosyl acetate or citrate, whileat pH < 3 molecular nitrous acid is formed. It is clear,however, that the production of nitrite decreased at pH< 4 and that only trace amounts could be detected inperchloric acid solutions of pH < 2 (Table 1).Analysis of Gaseous Products.-This showed that onlydinitrogen oxide could be detected at pH >3, that N,O insmall amounts and nitrogen mono-oxide could be found inthe range pH 2-3, and that NO is the predominant gaseousprocluct a t lower pH.These results confirm the generalconclusions previously reported, except that, like Bonnerand Ravid,8 we did not observe production of NO atpH r 3 . However, as described later, when [NO,]- wasadded, NO was produced at pH <5, and this may explainthe discrepancy in the literature in view of the readyoxidation of [N,0,]2- t o [NO,]-, which will then interferewith subsequent reactions of [N,O,le-.Kinetic Studies, pH >,4.-Most runs were carried out at26 "C and 0.25 mol dm-, ionic strength (maintained withsodium perchlorate). Occasionally, erratic runs wereobserved and were disregarded. Otherwise the repro-ducibility was fairly good. Over the range pH 4-10.5the reaction is of first order in [S,0,2-], as evidenced byexcellent plots of log D against time, the constancy ofand concentration of the buffer used.It may be seen thatbobs. increased with decreasing pH until pH cu. 8, and wasTABLE 2Decomposition of trioxodinitrate in aqueous solution a tBuffers are pH >4, 25 O C , and I = 0.25 mol dm-,.detailed in the text10.059.909.899.599.409.379.129.079.058.848.758.738.668.638.578.137.647.3611.417.713.219.412.716.113.012.817.912.08.621.516.113.611.013.612.09.21.18 7.211.83 7.001.58 6.872.72 6.872.33 6.663.12 6.223.18 5.873.55 5.853.98 5.575.07 5.444.65 5.335.78 5.424.80 5.066.30 4.396.18 4.026.676.446.8515.315.613.89.78.612.68.668.808.909.513.114.319.010.78.06.506.926.627.126.366.786.306.556.646.796.956.426.606.787.02then independent of pH down to pH 4 (Figure). It isquite clear that in this pH range decomposition takes placethrough the anion [HN,O,]-, values of hobs.following closelyIIPHEHect of pH on the decomposition of [N,0J2- at 26 "Cthe variation of the quotient [HT\;,O,-]/{[Hn',O,-] $-[N,0,2-] 1 calculated from the literature value of pK, 9.36.Alternatively, a graphical treatnient of the data in the rang1976 705pH 8-10 allows the calculation of pK2 9.35, in excellentagreement with the literature value. Equations (1) and[HN203]- " ~ 0 3 1 ~ - + Hf (1)(2) hold. Cleavage of the N=N bond in @N203]- gives[HN,O,]- ----t +N20 + $H20 + [NO,]-[KO,]- and [TU'OJ-, one of which will be protonated depend-ing on the site of protonation of [HN203]-.The pK of[HNOI- is 4.7, so for most of the present pH range theproducts of N=N cleavage will be [NOJ- and [NO]-. It iscustomary to assume that [NO]- dimerises to give cis-hyponitrite which decomposes rapidly to N20 and water.10Kinetic Studies at pH <4.-These were made a t 25 "Cand 0.25 mol dm-3 ionic strength. There was a rapidincrease in the measured first-order rate constant at pH < 3.An immediate explanation is that this reflects the con-version of [HN,O,]- into H2N203 which decomposes more(2)that the sulphamic acid-nitrous acid reaction is diffusioncontrolled.Values of hoba. a t pH < 3 were dependent on the age andbatch of Na2[N,03] as these contain different amounts of[NO,]-.All the runs reported were made over a shortperiod of time on one batch of Na2[N20,]. A standard runwas repeated to ensure that the nitrite concentration wasnot varying enough to invalidate a comparison betweenthese runs. Increase in the concentration of [N203]2-(and hence of catalytic [NOJ) speeds up the NO-producingreaction and has no effect on [HN203]- decomposition. Webelieve that the dramatic change in stoicheiometry withpH is accounted for more satisfactorily by this new reactionthan by the suggestion that i t results from decompositionof H,N203 rather than [HN,O,]-. The different nature ofthe reaction giving NO is shown by the effect of addedalcohol.This had no effect a t pH 5.5, but a t pH 2.38 theaddition of 1% ethanol caused a 38% decrease in bobs,while 2% ethanol caused a 80% decrease.TABLE 3Decomposition of trioxodinitrate in aqueous solution a t pH <4, 25 "C, and I = 0.25 mol dni-,3.613.513.613.253.253.263.153.163.163.063.063.062.842.842.725.119.431.215.340.851.816.949.356.615.632.242.013.730.830.85.075.957.207.80 1.8010.9 4.9012.2 6.209.68 4.1817.7 12.218.8 13.316.8 13.318.8 15.313.1 10.626.4 23.930.8 29.89.58 6.08PH1 o4k0b8./S-'k,/dm3 mol-l s-lk,(A) c/dm3 inol-' s-'10-3k3 d/dms mol-? s-l10-3k3(A) '/dm6 molP1.181.201.202.472.472.343.904.133.647.737.769.703.25 3.156.00 5.601.19 2.431.86 3.633.31 4.986.13 8.362.72 31.8 29.4 28.42.72 43.8 44.2 43.22.64 16.6 20.6 18.62.64 34.6 46.4 43.42.64 38.7 44.1 42.12.64 60.5 73.9 71.72.30 55.7 148 1482.30 81.4 218 2182.19 20.6 92.7 92.72.19 46.1 174 1742.19 51.6 268 2682.19 63.4 175 1761.91 46.5 292 2921.91 48.7 346 3461.91 67.3 410 4101.91 86.7 584 6843.06 2.84 2.72 2.64 2.39 2.19 1.913.50 2.50 ca.1.0 ca. 2 ca. 0 ca. 0 ca. 03.89 7.74 9.63 12.2 26.6 41.4 65.65.32 9.45 11.1 13.9 28.2 43.4 67.26.11 6.64 5.83 6.07 5.63 6.72 6.469.47 8.71 7.30 7.34 6.17 7.22 5.679.009.9011.912.610.911.926.526.845.237.750.032.862.871.060.967.4Values of kobo.* were obtained by extrapolation of kobo. to zero [NzOs2-]; at pH 2.30, 2.19, and 1.91, kob,.NN 0.kO~,.*)/[Ka,N,OJ. d k , = K,(A)/[H+].for the varying amounts of H,N,03 with pH, mean value 7 400 dms mol-, s-l.a k , = (kobo. -* Includes a correction c Includes a correction for the varying amounts of HNO, with pH.rapidly with the production of NO. However, whileindividual runs in this pH region gave excellent first-orderplots, values of hob& increased with increasing Na2[N203](Table 3). Bonner and Ravid did not report this pheno-menon. We attribute the behaviour to a reaction betweena tioxodinitrate species and nitrous acid, which is catalyticin nitrous acid. The nitrite is present as an impurity inAngeli's salt but is difficult t o determine quantitatively.When working with Na,[N,O,] containing zero or very low[NO2]- a short induction period was observed, during whichnitrite was formed by [HN203]- decomposition, followed byan increase in the reaction rate.We attempted to eliminatethe nitrite reaction by the use of nitrite scavengers, butadded excesses of urea or sulphamic acid had no effect,presumably because hydrogen trioxodinitrate is more re-active with nitrous acid. This is strange in view of the fact31. Gratzel, S. Taniguchi, and A. Henglein, Ber., 1970, 74,1003.A plot of kob& against [Na,N,OJ was linear at each pH.Extrapolation t o zero concentration gives a value for kobs.corresponding to the rate constant for the [HN,O,]- de-composition (designated hob.*). These values, shown onthe Figure and in Table 3, indicate that self decompositionof Na,[N,OJ decreases a t pH >3, i.e.H2N20, is stablecompared to [HN,03]-, and pK, ca. 3.0 in reasonableaccord with the literature value. There is a similarity withhyponitrite l1 where H2N,02 and [N20,]2- are stable com-pared to [HN,OJ-. Values of (hob& - k&e.*) give first-order rate constants for the reaction producing NO. Valuesof (hob*. - kob,*)/[Na2N,03] = k , were constant a t each pH(Table 2), showing that the reaction is first order in theconcentration of nitrous acid (the dominant nitrite speciesa t these pH values). This results from the fact that thenitrite concentration will be directly proportional to thatlo M. N. Hughes, Quavt. Rev., 1968, 22, 1.l1 M. N. Hughes and G. Stedman, J . Chem. Soc., 1963, 1539706 J.C.S.Daltonprobably by dimerisation and decomposition of [NO]-to N,O. The site of protonation of the anion is notof [N20,l2-. Values of k , therefore represent a minimumvalue, true values being obtained by multiplying by[Na,N,O,]/[NaNO,]. Unfortunately this quotient is notaccurately known, but the sample used in these runscontained ca. 5% nitrite. Also in Table 2 are values ofk , = k2/[Hf] [equation (3)], corrected to allow for theconversion of [NO,]- into HNO, and [HN,O,]- intoH,N,O, over the pH range studied. The latter calculationwas based on a pK, value for H2N203 of 3.32, obtained byRate = k,[H+][HNO,][H,N,O,] (3)a method of successive approximations. This gives K , =7 400[Na,N,0,]/[NaN02]. The constancy of k , is only fairbut is better than that obtained for the rate constant forequation (a), which also requires a very low pKl value forH2N203 of ca. 1.Rate a [H+] [HNO,] [HN,O,-] (4)The Nitrous Acid-Hydrogen Trioxodinitrate Reactiow-We attempted to confirm the existence of this reaction byobserving the effect of added [NO,]- on the rate andproducts of the decomposition of Na,[N,O,] over the rangepH 2-4.At all acidities this resulted in an increasedrate of decomposition, while at pH 4, where NO is notnormally a product of the decomposition reaction, additionof [NO,]- in equimolar amounts resulted in extensiveproduction of NO. Tests on NO production via nitrousacid disproportionation in the range pH 2-5 showedthat this is small compared to NO production in theHN0,-H,N,O, reaction.This was confirmed by the use of[16NO,]-, which resulted in the production of NO with only15% of the labelled product present, showing that theNO is not produced via nitrous acid decomposition. Italso appears probable by comparison with HNO, experi-ments that this l6NO is derived from disproportionation ofnitrous acid, and that both nitrogen atoms in H,N,O, endup in NO, so confirming the catalytic role of HNO,.The HN0,-H2N,0, reaction at pH >4 followed adifferent course, the trioxodinitrate being stabilised by[NO,]-. This may reflect the absence of the molecularacids, and the existence of an equilibrium between [HN,O,]-and its decomposition product, addition of [NO,]- thereforefavouring the recombination of HNO with [NO,]- ratherthan its dimerisation and decomposition.The reactionbetween HNO, and H2N,03 is being studied further atpresent.Ternfieratwe Dependence at pH >,4.-The reaction hasbeen studied in detail at 35 "C and briefly at other tem-peratures (Table 4). The pK, value at 35 "C is 9.30,indicative of a small value for the heat of ionisation of[HN,O,]-. Values of the enthalpy and entropy of activ-ation were 100 kJ mol-1 and 33.8 J K-I mol-I.TABLE 4Variation of rate with temperaturep~ e,pc 104kOb./s-1 PH e,l0c 104kobs./~-19.87 36 6.67 8.06 36 26.29.81 36 7.77 7.66 36 29.19.66 36 9.45 7.66 40 58.39.60 36 10.1 9.60 40 23.29.13 36 16.3 7.66 30 16.28.83 35 20.9 9.60 30 6.35DISCUSSIONThe interpretation of the results at pH >4 seems clearcut, and involves homolytic fission of [HN,O,]-, followedH-' t [N,0J2- z+= [HN,O,]-[HN,O.J- -+ HOX -+ [NOJ +known so it is possible that [NO]- and HNO, areproduced.Alternatively, the enhanced instability ofthe [HN,OJ- ion may result from A'-protonation giving-0-NH-6 (0)-0 - or O=&-NH (O-) , and f acilit atiiigN-N bond cleavage.The results at pH (4 are less easy to interpret.Bonner and Ravid attribute the increase in rate withdecreasing pH to decomposition via H2N203, the reactionat pH ca. 3 giving N,O and [NO,]-. However, theysuggest that protonation of HNO occurs at lower pHvalues and that this species reacts with [NO,]- insteadof dimerising, so accounting for the change in stoicheio-metry.We prefer a scheme involving the reactionbetween HNO,, H,N,O,, and H+, but agree that (a) NOis not formed by a simple dehydration reaction ofH2N203, and ( b ) that HNO, is implicated. Thesuggestion that NO is not produced by dehydration ofH2N203 is in accord with pulse-radiolysis studieswhich show that NO is the anhydride of the speciesH,NO,.In theory, the two mechanisms suggested for NOproduction could be distinguished by the use of [15N0,J-.The reaction between equimolar amounts of Na,[N,OJand Na[15N02] should give NO containing 25% 15N0if Bonner and Ravid's scheme holds, and zero 16N0 inour scheme. We have found intermediate levels of16N0 to be produced (i.e. ca. lS"/b), but cannot distin-guish precisely enough between this and the amount ofl5NO produced by disproportionation to categoricallyrule out Bonner and Ravid's scheme.However, oui-kinetic results cast doubt on their mechanism, as itcannot accommodate the observed first-order depend-ence on the concentration of HNO,, while the results inthe presence of ethanol show the reaction at pH (4 tobe more complicated than that suggested by them.Rate equation (3) is a familiar one l2 in nitrous acidchemistry, and is usually interpreted in terms of theattack of [H,NO,]+ on a substrate, although attack by[NO]+ must also be considered. Ridd12 noted that[H,NO,] + does not distinguish markedly betweennucleophiles, but there appears to be a limit for negativelycharged and neutral nucleophiles resulting from theeffect of charge on the encounter rate between thereactants.The limiting rate for neutral nucleophiles (S)a t 25 "C is given l3 by k, = 5 000 dm6 mol" s-l inRate = K,[H+][HNOJ[S]. In the present work k, z7 400[Na,N,O,]/[NaNO,] and is probably ca. 100 000dm6 mol-2 s-1, much greater than the limiting value.This suggests that the reaction of HNO, with H,N,O,+N,O 3- +H,O + [NO,]- (6)l2 J. H. Ridd, Quart. R e g . , 1961, 15, 418.13 I<. Xl-1\lallah, P. Collings, and G. Stcdrnan, J.C.S. DaZto?t,1974, 34691976 707does not involve a simple electrophilic nitrosation.This is confirmed by the effect of added ethanol, whichindicates the presence of a chain reaction, and by thefailure of excess of sulphamate to inhibit the reaction.There is a further parallel with hyponitrous acid.This reacts with nitrous acid l4,l5 according to rateequation (7). The value of k, is ca. 400 times greaterRate cc [HNO,][H,N,OJthan the limit for reaction with negative nucleophiles.As with hyponitrous acid, it appears then that the mostprobable mechanism for the decomposition of Angeli’sl4 M. N. Hughes and G. Stedman, J . Chem. SOL, 1963, 4230.= k,[H+l [HNO21 CHN20,-I (7)J. R. Buccholz and R. E. Powell, J . Amer. Chem. Soc.,1965, 87, 2350.salt at pH (3 is through a chain reaction that isinitiated by nitrous acid attack, probably giving anoxygen-nitrosated species.H+ + HNO, [H2N0.J+ =+= wO]+ + H20 (8)(9) [NO]+ + H,N20, ---t ONON=NO(OH) + H+ONON=NO(OH) --). NO, + N,O + OH (10)(12)NO, + H2N203 + HNO, + OH + 2N0 (11)OH + HNO, ---t H,O + NO,OH + OH -+ H20, -+ H20 + &O, (13)We thank Dr. G. Stedman for discussion, and the S.R.C.for the award of a studentship (to P. E. W.).[6/1625 Received, 18th Awgztst, 1976