Radiation Chemistry of the Aqueous Nitrate System. y Radiolysis of


Radiation Chemistry of the Aqueous Nitrate System. y Radiolysis of...

18 downloads 79 Views 1MB Size

MALCOLM DAKIELS AND ERICE. WIGG

1024

Radiation Chemistry of the Aqueous Nitrate System. I. y

Radiolysis of Dilute Solutions

by Malcolm Daniels and Eric E. Wigg Radiation Center and Department of Chemistry, Oregon State University, CorvalEis, Oregon (Received May 26, 1966)

The y radiolysis of the aqueous nitrate system has been investigated as a function of dose, intensity, temperature, concentration, and pH, with oxygen, nitrite, hydrogen, and hydrogen peroxide as scavengers. I n neutral solution a simple seven-stage mechanism is deduced which quantitatively accounts for the results. The rate-constant ratios k(e02)/k(eNOa-) and k(OH Hz)/k(OH NOz-) are evaluated as 2.5 f 0.2 and 0.8 X respectively, and only two rate constants of this mechanism remain unknown. Yields of primary products are deduced or measured, and it is shown that stoichiometry is obtained by including g(0J = 0.1. The mechanism in alkaline solution changes owing to the ionization of OH and the subsequent reactions of the 0- species. It is shown that 0- reacts rapidly with HzOz,k ( 0 H202) = 1.3 X 1Olo M-’ sec-l. Oxygen is found to compete with H202 for 0- and also to react with NOa2-. The complete mechanism in alkaline solution involves six reactions; for these only three rate constants are unknown and the ratio of two of these is determined, k(IV03202)/l~(N03~- H20) = 2 x 105.

+

+

+

+

+

+

+

Introduction

Experimental Section

Despite the considerable volume of work which has been carried out on the radiation chemistry of the aqueous nitrate system,’ no detailed mechanism in quantitative agreement with experimental observations has yet been proven; indeed the experimental observations reported by different workers are often discordant.2 Accordingly, we have undertaken to investigate the radiation-induced decomposition of the nitrate ion in aqueous solution over a wide range of experimental variables including dose, intensity, temperature, concentration, pH, and selected scavengers. I n this paper we discuss in detail the results for dilute solutions, both neutral and alkaline, for which we are able to propose simple self-consistent mechanisms and to substantiate them quantitatively. Reasons are suggested to account for the disagreement found among other workers. Lastly, it is shown that stoichiometry in the neutral solution yields of primary products (from water radiolysis) can only be achieved by including a G(OJ = 0.1, which we derive from the experimental results of Mahlman.

The water used throughout this work was triply distilled, water from a Barnstead still being distilled first from basic permanganate solution and then from acid dichromate solution. Baker Analyzed reagent grade sodium nitrate was further purified by the filtration of concentrated solutions followed by recrystallization from triply distilled water. It was found that filtering and recrystallization lowered the G(N02-) in 6 M NaN03 by about 0.4, while very little effect could be observed below 1 A4 KaN03. Further recrystallization was not found to be necessary. Reagent grade sodium hydroxide was purified according to the method of d’Ans and IMatt~ier.~The use of unpurified NaOH gave higher Not- yields a t pH values above 13

The Journul of Physical Chemistry

(1) See, for example, A. K. Pikaev, Russ. Chem. Rev., 29, 235 (1 960). (2) & L. I. Hyder, J . Phys. Chem., 69, 1858 (1965). (3) 1%.A. Mahlmsn, J . Chem. Phys., 35, 936 (1961). (4) J. d’Ans and J. Mattner, Angew. Chem., 64, 448 (1952).

RADIATION CHEMISTRY

O F THE

AQUEOUS NITRATESYSTEM

compared to identical runs with purified NaOH. Reagent grade sodium nitrite, unstabilized hydrogen peroxide, and nitric acid were used without further purification. Purification of sweeping gas was accomplished, in the case of hydrogen and helium, by passage through an activated charcoal-filled U tube maintained a t liquid nitrogen temperature. Oxygen and air were bubbled, first through concentrated sulfuric acid and then through triply distilled water. Gas saturation was carried out by passing the gas through the solution contained in a cylindrical bubbler of 50-cm length and 3-cm i.d. equipped with a fritted-glass gas dispersion tube. This arrangement gave essentially complete sweeping in 15 min with a gas flow rate of about 50 cc/min. The irradiation cells were constructed from 25-mm Pyrex glass tubing having a volume of about 25 cc. A 3-cm length of l-mm bore capillary tubing a t each end of the cell, one end being equipped with a 5/20 -$- joint, permitted filling to be done directly from the bubbler, the cell first being filled with the sweeping gas. The cells, when ready for irradiation, contained no gas volume and the capillary tubing served essentially to eliminate exchange of sweeping gas with the atmosphere. Two Co60sources were employed in this work. Most of the results were obtained using a 2500-curie unit which provided a homogenous field with a maximum dose rate of 4.0 X lozoev 1.-* min-', as determined by the ferrous sulfate dosimeter taking G(Fe3+) = 15.5 ions/ 100 ev. Repeated dosimetry runs were reproducible to within I.%. A few of the results reported here were obtained through the use of a 2700-curie source whose maximuin dose rate was 6.2 X lozoev 1.-l min-I. The reproducibility of repeated dosimetry runs in this case was better than 2%. Nitrite was determined using the method outlined by Shinn5 with a slight modification. The sulfanilamide solution was added prior to the addition of HC1. This order of addition was found to eliminate the acidcatalyzed thermal reaction between HzOz and NOzwhich caused serious errors when the acid was added to the solution first. A value for the molar extinction coefficient of the diazoamino compound was found to be 5.23 X lo4 at 540 mp. Hydrogen peroxide was determined by the iodide method, as modified by Schwartz and Salzman.6 The molar extinction coefficient for the Is- was determined to be 2.37 X lo4a t 352 mp. G values were in all cases determined from yield-dose plots having a t least three points each and usually more. These points fell on a straight line, in most cases with a precision better than 0.5%. In high-pH solutions, the scatter was somewhat greater. A few

1025

12

*

10

1.8

.3 2

6

0

4

8

2

0 0

1

2

3 4 Dose, ev/l. X

5

6

7

8

Figure 1. Yield-dose curves for NOZ- and HZOZformation a t an average intensity of 3 X 1020ev 1.-l min-l: A, NOZformation from 4 x 10-2 M He-swept pH 12 NaN03 solution; B, NOz- formation from 8 X M NO8- in neutral, air-equilibrated solution; C, Hz02formation from 1.0 X 10-1 M neutral aerated NaN03 solution; D, same as B with IN02-10= 4 p M ; E, HzOzformation a t pH 12, for He-swept 4 X 10-2 M N08- solution.

representative plots are shown in Figure 1. A repeat of any particular run gave reproducibility within 1%. G(H202) values, determined a t very low conversions, are good to f 3 % , close to the uncertainty (*2%) reported by Hochanadel and Casey.' The G(NOZ-) values we believe good to f2%. pH measurements were made with a Corning expanded-scale pH meter in conjunction with a Corning triple-purpose glass electrode. Considerable drift at pH values of the unbuffered nitrate solutions between 5 and 9 resulted in an uncertainty of a few tenths of a pH unit in this range. The G values given in this paper represent yields calculated on the basis of total energy absorbed in the solution. Energy deposition in concentrated solutions was calculated by assuming the absorption of energy to be proportional to electron density, a valid assumption for Co60 y rays and for the low atomic weight elements which were involved.

Results and Discussion A . Neutral Solution. (i) Yield-Dose Curves. Although the yield-dose curves for NOz- are quite linear, in most cases they extrapolate to a small intercept, usually from 0.5 to 1 p M . The situation is similar to ( 5 ) M. B. Shinn, Ind. Eng. Chem. AWE. Ed., 13, 33 (1941). (6) H. A. Schwartz and A. J. Salzman, Radiation Res., 9,502 (1958). (7) C. J. Hochanadel and R. Casey, ibid., 25, 198 (1965).

Volume 71,Number 4 March 1967

MALCOLM DANIELSAND ERICE. WIGG

1026

Table I: Biniolecular Rate Constants for Reactions of eaq-, H, and OH with NOa- and Radiolytic Products NOa-

N01-

1.1 x 10'0"

ea,-

H OH

'

'

1.2

x

10'0"

9.0 4.5 2.26

x x x

107 1072

sec-l)

H2

HzOz

4.6 x 109" 3 . 5 x 10Qb 6 X 1Vu 2.5 X log

8.2 X logb 2.4 x 107'

(k-l

0 2

4 0 7

4.5 3.5

1 0 7 ~

x x

1.9

x

10'OE

2.1

x

10'0

107~ 107~

Ref loa. Ref 11. Ref 13a. J. H. Baxendale, E. M. Fielden, C. Capellos, J. M. Francis, J. V. Davies, M. Ebert, C. W. Gilbert, A. Appleby, G. Scholes, and 31. Simic, J . Am. J. P. Keene, E. J. Land, and A. J. Swallow, Nature, 201, 468 (1964). e Ref 10b. Chem. Soc., 85, 3891 (1963). J. Rabani and G. Stein, J. Chem. Phys., 37, 1865 (1962). ' S. Gordon, E. J. Hart, and J. K. Thomas, J. Phys. Chem., 68, 1262 (1964). Ref 13b. H. A. Schwarz, J. Phus. Chem., 67, 2827 (1963). Ref 14.

'

'

I

1.1

IO-'

10-8

10-2

10-1

I

1

10

1

10

[NaNOsl, M .

h

B

X

1.5

9 1.0

,

$ 0.5 E; 0

10-4

2.0 1.5

10-8

10-2 10 -1 [NaNOa], M .

I

0

I

that found by Allen and Holroydsa for the hydrogen peroxide formation in aerated water and by Johnson and Weisssb for Ce(1V) reduction. A true initial yield thus cannot be determined with any confidence for the irradiation of nitrate without added OH scavenger and the G values reported here are "linear" G values obtained as in Figure 1. As such, G(X02-) is then independent of nitrite concentration. We suggest that the higher slope at (NOz-) < 3 p M is due to incomplete scavenging of OH by XO2- and this is supported by the fact that addition of 4 pM X02- prior to irradiation eliminates the intercept without significantly changing the slope (Figure 1, curve D). Accordingly, the G(n'02-) reported here (Figure 2a) must be considered to be determined under conditions of complete OH scavenging by K'Op-, which is not surprising in view of the high rate constant for this reaction (Table I) and the low reactivity of nitrate to OH. I n previous work2J' an attempt was made to determine "true" initial G values. Such values, higher than those found here, correspond to varying degrees of OH scavenging, depending on the concentration of KO*- allowed to build up, and are not susceptible to quantitative interpretation. From the rate constant, gOH, and intensity, it can be shown that concentrations of -lo4 M NOZ- would have to be determined to obtain a good estimate of the initial slope pertaining to conditions of no back reaction. Current techniques do not allow this. I n addition, traces of organic impurities will have an unaccountable effect on such initial G values. Accordingly, all G values reported here have been determined with consideration for the presence of radical scavengers, e.g., KO2- for OH, O2 and NOa- for e- and H, and H2 for OH. (8) (a) A. 0. Allen and R. A. Holroyd, J . Am. Chem. Soe., 77, 5852 (1955); (b) G. R. A. Johnson and J. Weiss, Proc. Roy. SOC.(London), A240, 189 (1957). (9) A. K. Pikaev, P. Y. Glazunov, and A. A. Yukubovick. Kinetika i Kataliz, 4, 835 (1963).

The Jourlacal of PhysaCal C h m k t r y

RADIATION CHEMISTRY OF

THE

AQUEOUS NITRATESYSTEM

(ii) Oxygen Scavenging i n Neutral Solution. a. Nitrite Formation. G values of nitrite formation for helium-swept, oxygen-saturated, and air-equilibrated solutions are shown as a function of nitrate concentration in Figure 2a. It can be seen that oxygen has a pronounced effect particularly a t lower nitrate concentrations. I n view of the known high reactivities of O2 and NO3- toward the solvated electron, it may be anticipated that they will compete on about equal terms for the electron. At this stage an adequate qualitative picture can be presented in which capture of the electron by S O 3 - leads to nitrite formation and capture by the oxygen leads to hydrogen peroxide, the OH being scavenged by NOZ-. The H atom will act similarly to the electron though with different rate constants and the reduction mechanism can be written as

+ Nos- +N0z2H + N032- + H + + e-

n'o3- -+ "03-

HT\;03- -+ OH-

(1) (2)

"03-

(3)

+ NO2

(4)

Competing reactions in the presence of oxygen are known to be

+ H+

e-

(5)

0 2 -+ 0 2 0 2

+HOz

(6)

and the sequence is completed by 2N02

+ € 1 2 0 + + 2H+ + NO%OH + NOz- +OH- + NOz so3-

(7) (8)

Clearly such a mechanism, in conjunction with known rate constants, predicts simple competition kinetics between oxygen and nitrate for the solvated electron and this is also suggested by the shapes of the curves of Figure 2a. However, before this can be tested quantitatively, the observed yields must be corrected for scavenging of electrons and hydrogen atoms in the spur by increasing concentrations of nitrate ions. This has been using the data Of and the adjusted yields are shown in Figure 2b. Scavenging M in the oxyplateaus are now clearly seen a t gen-containing solution and a t lo-' M in the He-swept solution, the difference being ascribed to the higher concentration of nitrate required to scavenge H atoms in the latter case. The data of Figure 2b now follow the competition kinetics of eq A over a certain range of nitrate concentration (Figure 3) __-1

G(N02-)

1 +-kA -Go

(02)

Go (Not-)

(A)

1027

A

250pU

[OZ]

1

21

200

400

800

600

1000

1200

[NaNOal-', M-1.

Figure 3. Competition kinetics of 0 2 and NO,- for the solvated electron; data of Figure 2b.

and JCA is evaluated as 2.5 f 0.2. This is, within the limits of experimental error, identical with the ratios of lcs/lcl in the literature (1.8 f O.3loaand 2.3 f 0.310br11) and we thus feel the competition can be ascribed to reactions 1and 5. The intercepts of Figure 3 give the limiting values of G(N02-) when all electrons are scavenged by nitrate. These values, 0.29 for air-equilibrated solutions and 0.27 for oxygen-saturated solutions, are related to the primary radical yields by eq I G(N02-)

'/&e-

- figOH)

(1) where 1 - fl is the fraction of OH radicals scavenged by molecular yield H202. Experimental evaluation of the extent of this scavenging (see below) leads to the same limiting nitrite yield a t both oxygen concentrations. =

1/2(ge- - gOH)

=

0.20

f

0.03

(11)

that scavIt is apparent from the ratio kz/k1 enging of H will only occur at higher concentrations. The plateau corresponding to this can be seen in the He-swept series of Figure 2b, curve A, a t lo-' M , and from this we estimate '/2(ge-

+ gH - gOH) = 0.44 * 0.02

(111)

b. Hydrogen Peroxide Formation. G(H202) as a function of nitrate concentration is presented in Figure 2c for oxygen-saturated, air-equilibrated, and heliumswept solutions. It can be seen that the magnitude of G(H202), in general, corresponds to the molecular yield with certain qualifications. Below significant increases are found in the presence of 02, such as would be expected from the competition be(10) (a) J. K. Thomas, S. Gordon, and E. J. Hart, J . Phys. C h e n . , 68, 1262 (1964); (b) M. S. Matheson, Radiation Res. Suppl., 4 , 1 (1964). (11) J. H. Baxendale, E. M. Fielden, and J. P. Keene, Proc. Roy. SOC.(London), A286, 320 (1965).

Volume 71,Number .G March 1967

MALCOLM DANIELS AND ERICE. WIGG

1028

tween O2 and NO3- for the electron, presented above. I n each case, curves B and C meet curve A a t the concentration where the corresponding G(iY02-) curves of Figure 2b reach their plateau values. However, it can also be seen that in the absence of oxygen some molecular yield H202 is consumed, complete protection only being attained a t -lo-' M NO3-. A consideration of the various species involved, along with the known rate constants, indicates that the attacking species is most likely the OH radical. The extent of scavenging is dependent on nitrate concentration since G(S02-) increases with increasing nitrate concentration, while gHz02 does not. Such a conclusion does not seriously invalidate the assumption of complete OH radical scavenging by NO2- below lo-' M KO3-, since we calculate that only -2.0% of the OH radicals are scavenged by H202 and experimentaIly we measure -5%. (iii) Scavenging by Hydrogen. It is clear from the previous results and discussion that a key role in the mechanism is played by the OH and NOz- reaction. To confirm that this does, in fact, account for the OH radicals under present conditions and to determine gOH, a radical scavenger was sought which would compete with NO2- for OH and a t the same time not introduce new uncertainties into the system. Organic scavengers used in previous work, notably by Allan,12 have not been free of this fault and gOH could not be determined. Although the reactivity of molecular hydrogen to OH is low, it is possible to adjust conditions to allow scavenging by Hz and the evaluation of G (NO%-) a t limiting scavenging conditions. At 6.4 X M NO3- we find G(N02-) = 3.13 in Hz-saturated solution. The dependence of G(N02-) on concentration of hydrogen and nitrite clearly shows competition kinetics (Figure 4). From the above mechanism, with the addition of reaction 9

eq B can be derived, where Go(N02-) is the limiting value attained when scavenging of OH by H2 is complete 1

Go(N02-) -- G(NO2-)

-~ - 1

1 +-kBQOH QOH

032)

(NO*-)

(B) From the slope we obtain ~(oH+H~)/~(oH+No~-) = 0.8 X 10-2, which can be compared with literature values of 1.8 X 138 and 1.4 X 10-2.13b,14After correcting for scavenging of electrons in the spur by NO3-, this mechanism, in which the oxidizing OH The Journal of Physical Chemistry

1000 1500 [Hzl/[NOz-l.

500

2000

2500

Figure 4. Competition kinetics of Hz and Not- for OH in M NOa-; intensity, 7.5 X 10l8ev 1.-l mine'. 6.4 X

radicals are converted into the reducing H atoms, allows us to write the reduction yield, G(NOZ-), as l/z(ge-

+ gH + gOH) = 2.97

f

0.08

(IV)

(iv) Primary Product Yields and Material Balance. Expressions 11, 111, and IV allow the evaluation of the primary radical yields independent of any assumptions of material balance in these primary species and hence allow such an assumption to be tested. Thus from relations I11 and IV, we obtain gOH and the sum of the reducing species (gegH).

+

gOH = 2.53 g(e-

f

+ H) = 3.41

0.10 f

0.10

Use of relation I1 then allows the evaluation of geas 2.93 f 0.16 and by difference we have gH = 0.48 f 0.26. I n view of the interest attached to these values, it was desirable to obtain other independent relations allowing the determination of the primary yields. This we have done using I- as a scavenger and analyzing for NO2- and 212 H202 by the methods already described. It was anticipated that all the OH radicals could easily be scavenged by I- ( ~ ( o H + I - ) 3 X lo9 M-' sec-I). If NOz is unreactive toward I-, G(N02-) would be expected to be -1.8; if however it is reduced

+

-

KO2

+ I- --+302- + I

(12) J. T. Allan, J. Phys. Chem., 68, 2697 (1964). (13) (a) M. S. Matheson and J. Rabani, ibid., 69, 1324 (1965); (b) H. A. Schwarz and A. 0. Allen, J . Am. Chem. Soe., 77, 1324 (1955). (14) J. K. Thomas, Trans. Faraday SOC.,61, 702 (1965).

RADIATION CHEMISTRY OF THE AQUEOUS NITRATE SYSTEM

+

then G(NO2-) should be -3.6 and the yield of ZIz HzOz should be correspondingly higher. Experimentally, the irradiation of oxygen-free nitrate (6.2 X M ) containing 1 X M I- resulted in measured values of G(NO2-) = 3.76 and G(ZIIz HzOz) = 3.93. The second mechanism is thus pertinent and we have

+

G(N02-) = ge-

+ gH + 2[gH2 - G(Hz)J = 3.76

and

from which we evaluat,e ge-

+ gH = 3.44

and

1029

Schwarz, discussing this question recently,15 concluded that a material balance deficit does exist in neutral solution to the extent of 0.6 f 0.2. The uncertainty in our present work may be reduced by considering, not the derived g values, but the experimental G values. When this is done we have, for complete scavenging conditions, ZG(NOz-) G(HJ = 0.87 whereas G (HzOz) = 0.75, clearly showing a deficit of -0.25 in oxidized products. It is our opinion then that a deficit exists, although it cannot be determined with precision. We are further strengthened in our opinion by the existence of another set of experimental data on this system. JIahlman3 has measured oxygen formation from nitrate solutions in the range 1.0-7.0 M . We find his results can be treated quite simply in terms of energy deposition in nitrate and water using

+

G(Oz)

gOH = 2.58 Clearly these results are in excellent agreement with those derived earlier and have the advantage that gH) is :tlmost a direct measurement. (geThe situation concerning radical yields in neutral solution has been reviewed recently by S c h u ~ a r zwho ,~~ concluded that it has “progressed to the point of confusion.” Comparison of our results with others in the recent literature is therefore of interest, taking as our starting point the “standard” values quoted by Allen,I6 gOH = 2.2, ge- = 2.80, and gH = 0.65. The gOH for the nitrate system is significantly larger, though similar to the value of 2.59 reported by Hochanadel for the CO H, system’ and somewhat less than Sutton’s value” of 2.9 for the KO system. The present ge- is perhaps larger, but the total yield of reducing species gH = 3.41 is close to Allen’s value of 3.45 and gesomewhat lower than gH for acid solution. Our lower value of gH, if significant, can perhaps be ascribed to H + reaction. scavenging by KO3- of the eI n addition to the radical yields we also determine gHrOz = 0.75 4: 0.04 so that together with gHz = 0.45 reported by Mahlman3 all the values necessary for a complete accounting of this system in terms of the commonly considered species are available. When this gH 2gHz = 4.3, is done we find a(-HzO) = Zgeequal to acid solution, but 4.0 calculated as ZgOH 2gHzOZ. Thus this system appears to be yet another example of the material balance deficiency discussed by AJlen,16 the deficiency in this case amounting to -0.3. Uncertainties in the determinations of the separate g values become cumulative15when stoichiometry is considered in this way and it is a valid question whether or not a lack of stoichiometry actually exists.

+

+

+

+

+ +

+

=

+

G ( ~ Z ) H , O ~ H , OG ( ~ Z ) N O ~ - ~ N O ~ -

where f is the fractional energy deposition and G is the appropriate coefficient. Calculating the f values from the electron densities leads to the results shown in Figure 5. The existence of an intercept clearly indicates that oxygen is formed from water in the radiolysis of dilute nitrate solution with a yield G(02) 0.1. This amount is close to that required to give material balance in the primary yields. Homever, the radical nature of this conclusion requires that other interpretations of this data be considered. Thus there may be a systematic error in the oxygen analysis not found in the hydrogen analysis, but this error must also be linearly related to fNOa-/fHzO,which we consider to be somewhat unlikely. Moreover, extrapolation of the linear relation to -0.1 M may not be valid, the relationship changing. However, there is further evidenceI8 from the radiolysis of nitrate solutions containing HzO’* that 0 2 does originate from the water. The extrapolation of G(O’5) seems reasonable and the variation of the yield of oxygen from the water with r\To3- concentration suggests that the oxygen may originate in the atomic form.Ig The sequence of results-(a) the material balance deficit in conventional species, (b) formation of oxygen with a G value filling the deficit, and (e) isotopic evidence that the oxygen does come from the waterseems to be the first experimental evidence supporting

-

(15) H. A. Schwarz, Ann. Rea. Phiis. Chem., 16, 347 (1965).

(16)A. 0.Allen, & d & h Res. S U P P t . , 4, 54 (1964). (17) W. A. Seddon and H. C. Sutton, Trans. Faraday Sac., 59, 2333 (1963). (18) H. A. Mahlman, J . phys. Chem., 67, 1466 (1963). (19) M. Daniels and E. E. wigg, Science, 1 5 3 , 1533 (1966).

Volume 71,.\-umber

4 March 1967

1030

nfALCOLM

p'

0.2

00

i I

' 0.1

0.2

0.3

0.4

0.5

0.6

fNOB-/fHfO.

Figure 5. G ( 0 2 ) / f ~ ~us.o ~ N O ~ - / ~ H results ~O; of Mahlman.ls

leading to nitrite formation, whereas in neutral solution, it is essentially unaffected. The following modification of the mechanism for neutral solution, involving ionization of OH and HzOz, quantitatively accounts for this behavior. The OH radical is known to ionize in alkaline solution, with a pK = 11.9*Oto give the 0-species

The Journal of Physical Chemistry

+

+

OH OH- If 0H~O (10) and we propose that this species can react with HzOz just as does OH, but with greater rate constant, so that H202is consumed. 0-

Allen's p r ~ p o s a l 'that ~ 0 atoms may be formed as a primary radiolysis product. It will be of interest to see if other systems can be investigated in sufficient detail to provide further evidence in this respect. (v) Intensity Variation. No effect of intensity was found for G[h'Oe-) or G(HZO2)between 6.2 X lozoand 8.5 X 10" ev I.-' min-'. The increase observed in dilute solution by Pikaev, et U Z . , ~ between 3 X lozo and 6 X 10" ev 1.-' min-' indicates that the recombination of OH in the bulk of the solution becomes significant a t these very high dose rates. However, his results should probably be reevaluated in the light of the mechanism presented here. (vi) Temperature Variation. An increase in solution temperature from 25 to 60" was found to produce M a negligible change in G(N02-) for 1.6 X NaN03. (vii) Comparison with Other Work. As mentioned earlier, the method of determining G(N02-) by other workers precludes the possibility of direct comparison with our work. G(H202) could be compared; however the data are scanty and diverse, making comparison pointless. Such behavior would be expected if trace organic impurities were present, as the work of Allan12 shows. The values obtained by AllanlZ for ge- and gH of 2.80 and 0.45 in the methanol nitrate system are in reasonakle agreement with those presented here, although his mechanism is incomplete and gOH could not be determined. B. Alkaline Solutions. (i) E$ect of p H in OxygenFree Solutiorz. There are many reports in the literature of increased nitrite formation in alkaline solution, but no satisfactory explanation has yet been offered in any of them. Accordingly, we have investigated the formation of nitrite and hydrogen peroxide in oxygen-free solution as :i function of pH, with the results shown in Figure 6. The previously reported increase in G(NOz-) is found; however there is a concurrent and equivalent consumption of hydrogen peroxide. Clearly, hydrogen peroxide is involved in the reaction sequence

DANIELS AND ERICE. W I G G

+ HzOz +HzO + 02-

(11)

The reaction sequence is completed by 02- reducing HN03- (an equivalent ionic form of NO2) to nitrite. 02-

+ H?r'Os-

+02

+ KO,- + OH-

(12)

The evidence for this reaction comes from the scavenging limit (see below). The kinetic expression for the consumption of hydrogen peroxide derived from this mechanism (supposing nitrite to be relatively unreactive to 0-) is given by the equation G(H202)

=

gHzOz -

where KOHis the equilibrium constant of reaction 10. From the experimental results, k11 has been calculated NOz-) using the values pK(0H) = 11.9 and k(OH = 2.5 X lo9M-' sec-' for a range of OH- concentrations (column 2 of Table 11); a definite trend with pH is seen. However, ionization of H202,pK = 11.75,21 occurs to the same extent as OH

+

HzOz

+ OH- IfH2O + HOz-

(13)

and taking account of this leads to the values in column 3 of Table 11, showing now no trend with pH. sec-' Iwas-' An average value for k n of 1.3 X 1Olo & then used to calculate the curve shown in Figure 6 (broken curve).22 (20) J. Rabani and M. S. Matheson, J . Am. Chem. Soc., 86, 3175 (1964). (21) .M. Evans and N. Uri, Trans. Faraday Soc., 45, 244 (1949). (22) A referee has pointed out that in view of the closeness of p K (OH) and pX(HzOz), reaction 11 cannot be distinguished from OH HOz- if equilibria 10 and 13 are maintained. However, it is doubtful if equilibrium can be regarded as being maintained in the presence of such rapid disturbing reactions as ( l l ) , a point emphasized by G . E. Adams, J. W. Boag, and B. D. Michael, Trans. Faraday SOC.,61, 492 (1965). This is also indicated by Figure 6 in which the curves are clearly not symmetrical about either of the pK values. However, some ambiguity remains which cannot be resolved by the present type of experiments and must await the results of appropriately designed pulse radiolysis experiments. For the present, we show that a satisfactory interpretation may be given by reaction 11.

+

RADIATION CHEMISTRY OF

THE

AQUEOUSNITRATESYSTEM

1031

+ ~(o-+H&oH(OH-) + k(o-+H~o~)KoH(OH-) 1

~(oH+H~)

A4

= 21*(H2)[

~(oH+H~o~)

(D) The variation of A$ with (OH-) calculated from this expression using known rate constants and our value of k11 gives excellent agreement with the experimental results as shown in Figure 8. Accordingly, we feel the mechanism proposed above and the deduced value of k11 to be reliable. The high value for kll is rather unexpected, but not im2

3

4

5

6

8

7

9

1 0 1 1 1 2 1 3 1 4

PH.

Figure 6. Variation of G(N02-) and G(H202) with pH in helium-swept, 3 x 10-2 M NaN03 solutions; intensity, 1020ev L-1 min-1.

, ) Expression C Table 11: Calculated Values for k ( 0 - + ~ ~ 0from k(O-+HzO%)v

PH

M -1 8ec -1, assuming no effect due to ionization of Hi02

k (O-+H10z)t M -1 8eo -1, asauming HOndoed not react with 0 -

10.20 11.05 12.25

1.1 x 1010 0.69 X 1Olo 0.49 X 1O1O

1 . 2 x 1010 0.83 X 1O1O 1.8 x 1010

20

40

60

80

100

200

[HzOzlo, p M .

Further support for this mechanism comes from two sources. First, addition of hydrogen peroxide prior to irradiation causes a further increase in G(N02-) as predicted by the mechanism, up to a limit of 2.80 (see Figure 7, curve A). This limit is attained a t quite low concentrations of H202 (50 p M ) , which is not unreasonable in view of the high value of k11, and can ‘/z(ge- - gOH). Second, be identified with gOH by use of kll determined here, we can account for the results obtained by Hochanadel in the photolysis of hydrogen peroxide in the presence of hydrogen a t alkaline pH.23 The lower rate of consumption of H201 observed in alkaline solution was ascribed to a “reaction of OH- with OH to give a species, o-, which does not react with Hz.” However, it has recently been shown that13ak ( 0 H2)/k(OH H2) 2, implying that the alternative explanation must hold; Le., k(0H202) > k(OH -t H202). Using competing reactions 11and 14

Figure 7. Dependence of G(N02-) on added HZOZat various oxygen concentrations, 3 X low2M NaNOa, pH 12. [OZ]:0,zero; A, 100 p M ; 0, 150 p M ; 0, 250 p M ; V, 1250 p M .

+

+

+

-

+

4 x

1.5

I

v

B

1.0

0.5 0

2

4

6

8

10

(14)

Figure 8. +( -HzOZ) us. pH. Curve calculated using expression D; data from ref 23.

we have derived eq D for the change in 4( -HzOz) in the presence of H,.

(23) C. J. Hochanadel, Radiation Rea., 17, 286 (1962).

O--

+ H2 +OH- + H

12

PH.

Volume 71, Number

4 March 1967

MALCOLM DANIELS AND ERICE. WIGG

1032

possibly high for a diff usion-controlled reaction. Using the approximate Smoluchowski equation for the rate constant of a diff usion-controlled reaction

we calculate k = 2.0 X 10'O M-I sec-' from the following parameters: ?(Ow) = 1.76 A (crystal and r(H2Oz) = 1.20 A (estimated from bond lengths), cm2 D ( 0 - ) assumed equal to D(0H-) = 6.5 X see-' 25a and D(H202) = 2.5 X cm2sec-'. 25b Adjustment of these values is not justified, although no doubt a closer value could be calculated. Such a calculation does serve to show, however, that the deduced value of kll is not unrealistic. Last, it should be pointed out that the present mechanism eliminates the necessity of postulating reaction of 0- with Nos-, suggested by Hyder2 to account for the increase of G(N02-) in alkaline solution. (ii) EfTect of Oxygen at p H 12. I n the absence of added HzOzvery little change is observed with increase in oxygen concentration up to 150 p M . Higher concentrations, however, decrease G(N02-) to 0.42 in oxygen-saturated solution, a value essentially identical with that obtained in neutral oxygen-saturated solution at the same nitrate concentration (3 X M). Superfically this mould appear to indicate that radical yields are unchanged from neutral pH; but, as will be shown, the reaction mechanism in the presence of oxygen does not allow this value to be identified immediately with '/Z(ge- - gOH). At the Same time, hydrogen peroxide yields increase with oxygen concentration from 0.12 (at 0%= 0) to 0.87 in oxygen-saturated solution. It is Clem from these results that oxygen is competing for the species attacking hydrogen peroxidel thus 1Owering the nitrite yield and protecting the molecular yield H202. I n view of the previously deduced mechanism t,his reaction is most probably

0-

+

0 2

+0 3 -

(15)

followed by

limiting value obtained by adding HZOZto 02-free solutions. Experimentally this is not found. Although adding H202 to 02-containing solutions does increase G(NO2-), complete reversal is not observed; rather, in each case, a limiting value is found which depends on the oxygen concentration (Figure 7). Thus the simple competition of reactions 15 and 11 is not adequate and the results indicate that O2 is also reacting with another precursor of NO,-. To account for this effect, we suggest that O2can react with N032-

Nos2-

+

0 2

+ so2 + + NOS0 2

(16)

Csapski and Dorfman26 have reported a transient absorption which they attributed to the 0 3 - species and Adams, Boag, and Michael2' have measured the rate constant of reaction 15. Such a mechanism can easily be verified. By adding HzOz prior to irradiation to solutions containing oxygen, it should be possible to regain the high values found in the absence of oxygen and indeed achieve the The Journal of Physical Chemistry

(17)

02-

in competition with

Nos2-

+ H2O +

"03-

+ OH-

(18) Thus for the dependence of G(N02-) on oxygen concentration in the presence of sufficient H202 for plateau conditions, we have

where Ga(N02-)is the value of 02-freesolution. Figure 9 shows the data, t,o fit this result and we evaluate k17/ kls as 2 X lo5. Such behavior will not be observed in neutral solution if k3 > k17. (iii) Primary Yields and Material Balance at p H 12. From the mechanism deduced here, G(N02-) in oxygenfree solution (Figure 6) can be identified with

+

gH202- G(H202) 1/2(ge- - gOH) = 1.15 (V) I n the presence of excess 02,the hydrogen peroxide and nitrite yields are related by the expression G(H202) - gH202 = I/*(ge- - gOH) - G(N02-)

(VI) Consistency of interpretation is shown by the fact that these expressions are in close agreement, giving from our best values

+

'/2(9e- - gOH) gH202= 1.22 (VII) With excess hydrogen peroxide in oxygen-free solution, we also have G(N02-) given by 1/2(ge-

03-

+n'o3- f

+ gOH) = 2.60

(vm)

Clearly if the molecular yield hydrogen peroxide could be measured separately, the primary radical yields (24) "Handbook of Chemistry and Physics," 44thied, The Chemical Rubber Publishing Co., Cleveland, Ohio, 1963. (25) (a) M.Eigen and L. De Maeyer, Proc. R o y . Soc. (London), A247, 505 (1958); (b) H. A. Schwara, Radiation Res. Suppl., 4, 89 (1964). (26) G. Caapski and L. M. Dorfmnn, J . Phys. Chem., 68, 1169 (1964). (27) G. E. Adams, J. W. Boag, and B. D. Michael, Nature, 205, 898 (1965).

RADIATION CHEMISTRY OF

THE

AQUEOUSNITRATESYSTEMS

1033

species, although scavenging conditions are very similar to neutral solution, implying k(0- H) > k(OH H), which is hardly likely, or the H species is not formed, or g(eH)alk < g(eHIneUtral. Taking gHz = 0.4033material balance can be considered and we can distinguish two likely situations. If gH202 is unchanged from neutral solution at 0.75 then we have the same deficit in material balance as in neutral solution, but gH2Ozneed only be somewhat higher a t 0.80 for material balance to obtain. Further discussion is not warranted until gH202is unambiguously determined. The difficulty encountered in the present work of unambiguously determining gHz02 must cause reconsideration of the methods of determination of other reported values, particularly in view of the very rapid reaction of 0- with HzOZ. It must be demonstrated experimentally that H202 is protected against 0-, otherwise apparently low values may be obtained, and it cannot be assumed that 0- behaves as OH, for example k(0- H2O2)>> k(OH HzOz)whereas k(0- Br-)