4775
RADIATION CHEMISTRY OF FORMIC ACID
Sept. 20, 1960
[CONTRIBUTION FROM
ARCONNENATIONAL LABORATORY, ARGONNE,
ILLINOIS
1
Radiation Chemistry of Aqueous Formic Acid Solutions. Effect of Concentration’ BY DONALD SMITHIES~ AND EDWIN J. HART RECEIVED JANUARY 11, 1960 The radiolysis of formic acid solutions by C060 y-rays at concentrations in the range 1.0 to 26.6 AB, pure formic acid is reported Carbon dioxide, carbon monoxide and hydrogen are the sole gaseous products and formaldehyde is a reactive intermediate. G(CO2) reaches values as high as 12, increases as (HCOOH):/z at high dose rates, and as (dose rate)-’/a a t constant concentration. These results indicate a chain reaction. The reaction sequence COOH HCOOH = COz HCO H20 HCO HCOOH = HCHO COOH explains the chain foimation of carbon dioxide. G(H2) decreases from 3.2 in dilute solutions t o 2.4 a t 5-10 ’$1formic acid and then remains unchanged with increasing concentration. G(C0.) rises from zero with increase of formic acid concentration, rapidly at first and then more gradually t o 1.25 in pure formic acid. Direct excitation of formic acid by water subexcitation electrons is suggested as the mechanism explaiging the carbon monoxide yields at low concentrations. Some supporting photochemical studies at 1860, 2537 and 2669 A. are also reported.
++
++ +
Introduction Irradiation of dilute aqueous solutions of formic acid has been used to establish free radical and molecular product yields. Even though formic acid is not a particularly efficient free radical ~ c a v e n g e r the , ~ formic acid-water system has particular merit in the study of radiation effects in concentrated solutions since the components are miscible in all proportions. In aqueous solutions of formic acid of coccentration to lo-* 12.I and a t PH less than 4.0, carbon dioxide and hydrogen, the sole products of yradiolysis, are formed with equal yields. These yields are substantially independent of concentration and dose rate.3.5 The mechanism consistent with these results is Hz0 --+ H , OH, Hz, HzOz OH f HCOOH ---+HzO COOH H HCOOH +H P COOH Hz02 --+ COZ HzO OH COOH COOH + COz HCOOII COOH
+ + + + +
+ + +
The over-all reaction is HCOOH
--+ Hz
(1) (2) (3) (4) (5)
+ COn
The “molecular” hydrogen peroxide formed in (1) disappears via a chain mechanism involving steps (4) and (2).3 The molecular hydrogen and hydrogen peroxide yields formed by pairwise recombination of hydrogen and hydroxyl radicals within the “spur” are represented by g(H2) and g(H202), respectively. These yields are expected to decrease as the formic acid concentration is increased because of reactions 2 and 3. The above mechanism predicts that the carbon dioxide yield increases as g(H2) decreases but is independent of g (Hz02). In this paper G(CO2) is used as a measure of the efficiency of free radical formation in waterby excitation as well as ionization since it is known (1) Based on work performed under t h e auspices of t h e U. S. Atomic Energy Commission. (2) D e p a r t m e n t of Chemistry, The University, Leeds, England. ( 3 ) E. J . H a r t , (a) THIS JOURNAL,73, 68 (1951); ( h ) ibid.. 7 6 , 4198 (1954); ( c ) i b i d . , 7 6 , 4312 (1954); (d) J . Phys. Chem., 56, 584 (1952); (e) Radiation Research, 1, 53 (1954). (4) (a) E. J. H a r t , T H I S J O U R N A L , 74, 4174 (1952); (b) J . L. Weeks a n d M. S. Matheson, i b i d . , 78, 1273 (1956); (c) J . H. Baxendale a n d D. Smithies, 2 . Phys. Chem., 7 , 242 (1956). (5) H. Fricke, E. J. H a r t a n d H. P. S m i t h , J. Chem. P h y s . , 6, 239 (1938)
that water activated of dissociated by light of wave length 1849 or 1980 A. decomposes formic acids6 Results are reported for formic acid radiolysis in the concentration range from 1.0 to 26.6 A I , pure formic acid in order to determine primary yields for water decomposition and to investigate the effect of superimposing formic acid ionization and dissociation processes on water radiolysis. Experimental Formic acid (Matheson 98-100%) was distilled at 20 cm. pressure through a 30 plate fractionation column at 59’ to give a product having T Z ~ O D of 1.3715. The previously described techniques of irradiation, dosimetry, solution deaeration, cell filling and Van Slyke gas analysis were used.3 The gaseous irradiation products were separated from the solution in the Van Slyke apparatus, collected in a storage tube and the contents analyzed by conventional techniques viz. absorption by 1 M KOH for COZand mass spectrometer analysis for H Zand CO. Formaldehyde was determined by a modification of the method of Laska.’ This consisted of adding 10 ml. of a 45 p M solution of phenylhydrazine hydrochloride t o 5 ml. of the solution being estimated and standing 5 minutes. Two ml. of 1.234Yc potassium ferricyanide were then added, stood 5 minutes, 5 ml. of 12 N HzSO4added and t)e whole made up to 25 ml. The optical density at 5200 A. was measured. Formaldehyde concentration was calculated from the equation C = 41Dd where C = formaldehyde concentration in p Af D = optical density in 1 cm. cell d = dilution Carbon dioxide has appreciably greater solubility in formic acid solutions than in water. The Bunsen solubility coefficients (vol. COz/vol. sol.) measured by Adams and Anderson8 in 5.0, 10.0, 15.0, 20.0 and 26.6 Mforniic acid are 1.03, 1.30, 1.67, 2.37 and 4.94, respectively. These coefficients were used in calculating the concentration of carbon dioxide in the irradiated solutions from equilibrium pressures existing above the solutioxi in the Van Slyke pipet.
Results Hydrogen, carbon dioxide and carbon monoxide plots of product formed ‘c‘erszL.7 absorbed dose are linear for a particular set of initial conditions. Even a t the lowest dose rate used where carbon dioxide is formed mainly by a chain mechanism its formation is linear with dose. In contrast, the formaldehyde concentration rapidly reaches a low steady (6) J . L. Weeks a n d M. S . Matheson, unpublished results. (7) F. Laska, C h e m . L i s l y , 33, 375 (1939). (8) G. E. Adanis and A. R. Anderson, U. S. Atomic Energy Commission, Report .4FI,-59t)l, Aiiril, 1959.
I)OXALD SMITHIES AND
4776
EDWIN J. HART
Vol. 82
9.0
8.0
i
7.0
6.0
5.0
0
5.0
IO
0
15.0
(HCOOH)
20 0
25.0
30.0
0
(MI.
Fig. 1 -Effect of formic acid concentration and dose rate on G(C02), G ( H d and G(C0): 0, G(C02) a t 35 X l O l Q ev./l. min.; 0,G(CO2) a t 2 2 X l O I 9 ev./l. Inin.; A, G(CO2) at 0.33 X 1 O I y ev./l. min.; e, G(H2); 8 , G(C0).
state level. Each of these four compounds is a primary product. Effect of Concentration.-G(C02) increases from 4.08, the value in 0.1 M solutions (the first point in Fig. l), with increasing forniic acid concentration, rapidly a t first and finally a t each dose rate reaches concentration independent yields which increase as the dose rate decreases (see Fig. 1). G(C0) abruptly rises from zero with increase of formic acid concentration and thrn more gradually. G(H2) however, decreases from 3.2 in dilute solution to about 2.4 a t 5-10 M and then remains unchanged up to pure formic acid. G(C0) and G(H2) are independent of dose rate and the yield of these shown in Fig. 1 are averages of the values obtained a t dose rates of (33, 2.2 and 0.33) X 1019 ev./l.min. When the dose rate is 33 X l O I 9 ev./l.min., G(C02) increases as (HCOZH)l/* up to 15 151 acid. At a dose rate of 2.2 X l o L 9ev./l.min. this square root relation holds up t o 10 144 formic acid. (See Fig. 2.) Effect of Dose Rate.-At constant formic acid concentration in the range from 5 to 20 M , G(C0)2 increases from 5.76 a t 120 X 1019 ev./l.niin. to 21.3 a t 0.05 X l O I 9 ev./l.min. In pure formic acid, G(C02) is practically independent of dose rate, as also are G(C0) and G(H2) over the whole range of concentration. Effect of Temperature.-G(COz), G(C0) and G(H2) gradually increase in the temperature range from 6 to 78" for 5.0 M solutions (see Table I). G(CO2) increases more rapidly than G(C0) or G(H2) both of which increase a t about the same rate. This work was carried out using a high dose rate, 107 X 1019 ev./l.min., thereby minimizing the chain reactions. Plots of log A G(COZ),log G(CO), and log G(HJ us. 1/1' give activation energies for the formation of these products falling in the range from 3.0 to 7.6 kcal./mole [AG(CO,) = G(CO2)(obsd.) - G(C02)(infinite intensity)]. G(CO2) at infinite intensity isobtained from the extrapolated points for plots of G(C02) see eq. 9 or 9'. vs. (Ia)''~;
1.0
2.0 ( ti C O O ti )"Z
4.0
3.0 (
M )'I2
5.0
I
Fig. 2.-G(COa) as a function of square root of formic acid concentration: 0, 35 X 1019 cv./l. Inin.; 0 , 2 . 2 X ev./l. min.
Effect of Ferric Sulfate, Benzoquinone and Formaldehyde.-Ferric sulfate and benzoquinone, free radical scavengers, reduce G(C02), G(H2) and G(C0) in 5 ild formic acid (see Table 11). In agreement with its greater eficiency as a hydrogen atom scavenger, benzoquinone reduces G(H2) more than ferric sulfate does. Relatively, G(C0) is reduced more than G(Il2). At low dose rates where G(C02) becomes large, the efiectiveness of ferric sulfate in reducing the yield increases (see Table 11). Initially added 0.001 1 11 formaldehyde in 5.0 M formic acid causes little significant change in G(COS) or G(H2) although G(C0) is reduced. G(-HCHO) is 6.15 showing that formaldehyde disappears with about the established radical yield. TAULB I EFFECT OF TEMPERATURE ON THE RADIOLYSIS" OF 5 A I FORMIC ACID Temp., ' C .
G(COd
6 24 35 52-55 75-80 78 Dose rate = 107
5.53 5.81 ,5,94 7.15 7.42 6.93 X loi9 ev./l.
G(Hd
G(C0)
1.93 2.26 2.26 2.67 2 83 3.60 min.
0.34 0.50 0.61 1.27 1.39 1.46
TABLE 11 EFFECTO F ADDITIVES O N G(C02), G(H2) AND G ( C 0 ) IRRADIATED 5 X FORMIC ACID I x 10-19 ev./l. min.
G(C0z)
G(Hd
G(C0)
0.52 5.93 2.37 Sone 5.67 2 . 1 0 .16 0 . 0 1 AI I?,+++ .71 Sone 12.48 2.62 0.003 M F e + + + 7 20 ( 2 . 7 1 CO iHI) ,001M Q4 4.59 1.55 0.10 ,001 hf RCHOb 6.42 2.18 .22 .1 h' H2S04 6 63 2 . 7 8 .C,2 , 1 ArH2S04+ 0 . 0 1 5 . 8 7 2 26 .IO M Fe+++ Q = benzoquitione. b G( -1ICHO) = 6.15.
34.9 34.9 0.33 0 33 34.9 34.9 34.9 34 9 5
Additive
IN
Sept. 20. 1960
PHOTOCHEMICAL
0.0002 0.012
5.0 5 0 5.0 5.0 26.6 26.6
RADIATION CHElLrISTRY O F FORMIC ACID
ent but the intensity exponent has not been established. Possible propagation steps in the chain reaction are
TABLE I11 DECOMPOSITION O F FORMIC .4CID
1860 0 . 6 8 1800 0 . 3 7
0 0.16 ( .20) 2669 1 . 9 8 .26 2537 1 . 6 6 .14 2537 2 . 2 8 .14 1860 1.03 .58 2669 0 . 8 9 .13 .80 1860 0 . 9 6
0.60 .2T ( .36) .15
.04 .04 .Oi6 .0066
4777
+
+ +
COOH HCOOH .--f COz f HzO HCO HCO HCOOH -3 H C H O f COOH
0.0041 .0042
(6) (7)
A trimolecular termination step consistent with the observed dependence on formic acid concentration and similar t o the one proposed for peroxide radiolysis is
.0036 .0268 ,0032 17.0 0.0345 100.0
2COOH
+ HCOOH + COI 4-2HCOOH
(8)
Using a mechanism consisting of reactions 1, 2, 3, 4, 6, 7 and 8 and assuming steady state concentrations of intermediate free radicals, expression 9 for Photochemical Decomposition of Formic Acid.Aqueous formic acid solutions are decomposed by G(C02)can be derived, This is in accord with the observed concentration and dose rate dependencies light of wave lengths 2669, 2537 and 1860 Preliminary results are reported in Table 111. shown in Figs. 2 and 3. Carbon dioxide is the major product, lormed with a quantum yield, ~ ( C O Z )of , about 1.0 a t the shortest wave lengthoand a t high intensity. At g(H) g ( 0 H ) Vi (HCOOH)’/l X = 2537 - 2669 A. and a t lower intensities, ks 2ks I . 1 / 1 (9) +(COz) becomes greater than 2.0. Carbon mon- whereg(I1202),g(H) and g(0H) are the molecular oxide yields are dependent on wave length and product and free radical yields a t formic acid concenformic acid concentration, the shorter the wave tration (HCOOH); kg and ks are the rate constants length and the higher the formic acid concentra- of reactions 6 and 8, respectively; I , is the rate of tion the higher are the yields. Hydrogen is a energy input in evJ6.02 X l o z 5 l.sec. I n this minor product in all cases reported in Table 111, derivation the carboxyl radical is assumed to be although its relative importance increases with much less reactive than the formyl radical. Reacdecreasing formic acid concentration. tion 6 is then the rate determining step and (8) becomes the effective termination step. The acDiscussion The radiation chemistry of dilute aqueous solu- tivation energy of 3.0 to 7.6 kcal./mole estimated tions of formic acid is adequately explained by above is ascribed to reaction 6 if one assumes zero reactions 1-5, but a t concentration above 0.01 activation energy for the radical--radical terminaM more complex reactions are involved and both tion reaction 8. While the (HCOOH)’/a relation seems definite ionization and excitation effects in water and formic in Fig. 2, the results shown in Fig. 3 indicate a acid must be considered. Another complicating zero order dependence on formic acid concentrafactor is the changing ionic and molecular species tion. Note that the curves are nearly parallel and of formic acid. In order to hold the discussion displaced to higher G(C02) as the formic acid conwithin reasonable limits, it will be confined to centration increases. If the square root dependaccounting for the products and the dependence of ence held over the intensity range covered in Fi . their yields on formic acid concentration. Carbon Dioxide.-In concentrated formic acid 3, then G(C02) should increase as (HC0OH)’f. solutions a new feature, the inception of a carbon Zero order kinetics can be derived by assuming dioxide producing chain reaction, appears; hence that complex formation of the chain carrying COOH G(C02) no longer provides a measure of the radical is followed by bimolecular termination COOH HCOOH +COOH.HCOOH molecular product and free radical yields as it does in dilute solutions. This reaction increases COOH.HCOOH + COz HCO H20 (6’) G(CO1) but has little or no effect on G(C0) and 2COOH.HCOOH + COz 3HCOOH (8’) G(Hz). Significant features relating to the production Now one obtains of carbon dioxide are the high values for G(C0)2, G(CO2) = g(H) ‘(OH) + g(HzOz) + 2 the proportionality of G(C02) to (dose rate)-’/’ in the concentration range 1 to 20 M (see Fig. 3) and the inhibitory effects of ferric sulfate and benzoquinone. I n addition, G(CO2) increases as which is consistent with the data presented in Fig. 3. [HCOOH]’/z a t least a t 35 X 1019 ev./l. min. Further refinement of the data is necessary in order (see Fig. 2). These results not only support to explain the apparent anomaly between the the conclusion that a chain reaction producing (HCOOH)’/I dependence of Fig. 2 and the zero orcarbon dioxide occurs but suggest that the mech- der dependence of Fig. 3. It may also be possible anism parallels that for radiolysis of hydrogen that the first two terms of equations 9 and 9’ dep e r o ~ i d e . ~The quantum yield, d(COz) for photol- pend on (HCOOH)’/%. ysis of 5.0 ill formic acid is also intensity dependEquations 9 or 9’ predict that G(CO2) extrapolated to infinite dose rate, G(C02) provides a meas(9) E. J. Hart and M. s. Matheson, Discussions Faraday SOC, 12, 169 (1952). ure of g(HzO?) 4- ’/dg(H) g ( 0 H ) ) . Figure 4 .048
a.
+
+
1
+
+
+
+
DONALD SMITHIES A N D EDWIN J. HART
4778
y
5.0
4.0 0
‘
1.0
I
I
2.0
3.0
(I
) J ~ I0 ’ 0 ( P u / e m i
I 4.0
I
i J
5.0
n,y’/z,
Fig. 3.-G(COs) as a function of inverse square root of dose rate: 0, 1 N HCOOH; 3 , 5 M HCOOII; A, 20 .li HCOOH.
shows that G(COz), rises rapidly a t formic acid concentrations below 5 AT to a constant value of about 6.5 a t the intermediate concentrations and then rises again above 20 111. Since the chain reaction is eliminated at infinite dose rate, G(C02), originates from primary nonchain processes occurring in the water and formic acid. At formic acid concentrations above 1 M the fraction of the energy absorbed by the formic acid is no longer negligible. Assuming ( 1 ) that the direct effect is independent of formic acid concentration, ( 2 ) that G(CO,) = 8.0 for pure formic acid (later work a t higher dose rates established that C;(C02), for pure formic acid is nearer 8.0 than the value of about 9.0 indicated by the results of Fig. l), (3j that the energy absorbed by a component of a solution is proportional to the electron fraction of that component, G(COa) froin reactions originating in water radiolysis can be calculated. (Lower curve in Fig. 4.) This curve shows, as expected, that the yield of carbon dioxide from primary processes occurring in water increases with formic acid conceritration. This increase is expected not only because of scavenging of the precursors of molecular hydrogen €1
+ H --+ H ?
(101
but also because of interference with the radical recombination reaction €T
+ OH +H20
(11)
Scavenging of the precursors of molecular hydrogen peroxide produces no increase in G(C02) because of reaction 4. If reaction l l is assumed to have twice the probability of reaction 10, then the expected increase in G(C02) is given by 3g(H2) = 3 X 0.43 = 1.29. G(C02) increases from 4.1 in 0.1 M to a maximum of 5.3 in 10 A f formic acid. This result indicates that scavenging of radicals is complete in 10 3 2 formic acid arid that 5.3 water molecules are dissociated by y-rays per 100 ev. absorbed by the water.
0
Vol. 82
5
15
IO
(HCOOH)
20
25
(M).
Fig. 4.--Non-chain G(C02), and G ( C 0 2 ) ~as s ~a function of formic acid concentration: 0, G(COl),; 0 , G(CO?)H?O
Hydrogen.-G(Hz) decreases as the formic acid concentration is increased up to 10 AT and then remains substantially unchanged to 26.6 Icf, pure formic acid. I t is pertinent to ask why a change in the slope of the G(H2) vs. concentration curve occurs. The initial decrease in G(H2) can be explained by reactions 12 and 13 competing with (3) as suggested by Garrison and co-workers’O or by electrons captured by formic acid.3c
+ H20 f HCOOH
HCO
(13)
Hydrogen is formed in reaction 3, whereas (12) and (13) yield a formyl radical which is normally removed via reaction 7 . According to this scheme G(Ha) tends to zero a t high formic acid concentrations and is replaced by formaldehyde. This conclusion is supported by the photochemical data showing that d(H2) is substantially zero (Table 111). However, in 26.6 M formic acid G(H2) = 2.2 which indicates that the ionization processes operating are more complex. The following reasonable processes producing H atoms are revealed by electron impact studies on formic acid and its deuterated
+ H + ?e+ I1 HCOOH- ---+ HCOO- + 1%
HCOOH
+ e-
HCOOH+
--f
CI-I02+
+ (M)+
CH02+
(1-1) (1> +(Hz) or d(C0) a t waye lengths of 2537 and 2669 k . but not a t 1860 A. (see Table 111). In view of the (13) E. Gorin and H. S . Taylor, THISJOURNAL, 66, 2042 (1934). (14) W. N. Herr and W. A . Noyes, Jr., ibid., 50, 2345 (1958). (15) H. Thiele, Ber., 40, 4914 (1907). 116) A . J . Allmand and L. Reeve, J . Chem. Soc., 129, 2852 (1926).
would explain the decrease in G(C0). Acknowledgments.-The authors appreciate the help given by Messrs. Ben Holt, Chester Plucinslti and Kenneth Jensen in the analysis of hydrogen, carbon monoxide and formaldehyde. (17) E . J. Hart, THIS JOURNAL, 81, 608.5 (1959).
