August, 1955 KINETICS DECOMPOSITION

OF THERMAL NITRIC
OF LIQUID ACID 683

THE KINETICS OF THE THERMAL DECOMPOSITION OF NITRIC ACID

IN THE LIQUID PHASE
BY GLENND. ROBERTSON,
JR., DAVID
M. MASONAND WILLIAM
H. CORCORAN
Chemical Engineering Laboraiory, California Institute of Technolog2/,Pasadena, California
Received December 87, 1064

The kinetics of the liquid phase decomposition of HNOs in the temperature range 54 to 88' suggests that the unimolecular
decomposition of NzO~ which is known to exist in fuming nitric acid solutions plays a significant role in the rate-determining
step. A rate mechanism experimentally indistinguishable from that of the N9O6 hypothesis involves the product (NOz+)-
(NOa-). The effect of several inorganic additives, namely, HzO,N204,NeOJKN03, KHSO?, NOHS04, H2SO4 and HC104,
on the rate of decomposition was in agreement with these postulated rate mechanisms. The inhibition of the decomposition
rate by varioua additives tested was found to be insufficient to prevent the eventual attainment of high equilibrium pres-
sures resulting from the decomposition of fuming nitric acid when stored in closed containers for perjods of about a month
in the temperature range 54 to 88'. Of all the additives tested on both a molal and a weight basis, H20 gave the most
pronounced inhibiting effect on HNOa decomposition.

Introduction both liquid and gas phases are present. Therefore

The thermal decomposition of fuming nitric a more detailed study of the kinetics of the liquid
acid' introduces many problems in storage and phase decomposition was undertaken, and measure-
handling of this important material which is used ments of pressure as a function of time under iso-
as an oxidant for rocket fuels and as a nitrating choric conditions were used to follow the rate of
agent in chemical syntheses. Some of the factors formation of O2 and correspondingly the rate of
involved in the kinetics o€ its thermal decomposi- decomposition of HN03.
tion are discussed here.2 Description of Methods and Equipment
The work of other investigator^^-^ has estab-
lished the over-all stoichiometry of the decomposi- Measurement for
Apparatus Determining Decomposition Rate by
of Pressure as a Function of Time.-The
tion to be usual method of following the rate of a chemical reaction is to
4HNOa 4N02 + 2Hz0 + 0 2 (1) observe the change with time of the concentration of one of
the products or reactants. It is frequently more convenient,
for both the liquid and gas phase reactions. Avail- although not always feasible, to observe some physical
able equilibrium data' show that with an increase property of the system which is related to the concentration
in temperature the products of decomposition are of one of the species. I n the present system, either elec-
trolytic conductance9 or optical absorbanceI0could have been
favored. These products have been observed over a observed since these properties have been shown to vary
temperature range from room temperature to 80" measurably with composition of the acid. Their use as
for the liquid phase reaction3,6 and to 475" for analytical tools, however, would have required lengthy
the gas phase r e a c t i ~ n . ~At
. ~ the temperatures calibration procedures, and for this reason they were not
utilized.
encountered in the present investigation, ie., 54 to The increase in pressure resulting from the formation of
88", the formation of NO by the decomposition of oxygen offers an easily interpreted measure of the rate a t
YO2 from the equation which HNOl is decomposing. A sample of HN03 was con-
2N02 C-,2 N 0 + 02 (2)
fined in a precision-bore, glass capillary-tube (3 mm. i.d.)
by Fluorolubell which has a greater density than the liquid
is negligible.* acid. The liquid and gas phase volumes of the sample were
investigation^^^^ have shown that the rate of de- determined to within 10.0004 cc. by locating the gas-
liquid and liquid-liquid interfaces with a cathetometer.
composition of "03 in the vapor phase is not Concentric with the capillary tube was another tube which
sufficientlyhigh to account for the rates of pressure was vacuum jacketed. Mineral oil, the temperature of
rise observed under actual storage conditions when which could be controlled to within &0.05" of any desired
value between 35 and 90°, flowed rapidly along the capillary
(1) The term fuming nitric acid represents the ternary system tube in the annular space between it and the surrounding
H N O ~ N O P - H I Ofor concentrations of HNOs greater than about 0.75 tube. To ensure constant hysical equilibrium between
weight fraction. For the sake of brevity, N0z is used throughout phases, an iron pellet encasefin glass was driven back and
this paper to designate equilibrium mixtures of NOn and NzOi unless forth through the gas and liquid phase of the acid sample
otherwise specified, Compositions are expressed in weight fractions once per second by means of a pair of intermittently actu-
on a formal basis. Also HNOB designates equilibrium mixtures with ated elect,romagnet,s.
associated species such as (HNOdm. See R. Dalinon and R. Frey- The pressure within the capillary tube was transmitted
mann, Compl. rend., 211, 472 (1940). from the confining fluid through the mercury U-tube to an
(2) The material reported here is based largely on a thesis pre- oil-filled pressure balance which was capable of measuring
sented a t the California Infititute of Technology in 1953 by Glenn D . pressure with a precision of 10.005 atmosphere up to a total
Robertson, Jr., in partial fulfillment of the requirements for the degree
pressure of 22 atmospheres. This balance had been cali-
of Chemical Engineer. The work was part of a program sponsored brated against, a standard balance which was periodically
by the Jet Propulsion Laboratory a t the California Institute under checked against the vapor pressure of COZat the ice point.
Contract No. Da-04-495 ORD-18 for the U. 8. Army Ordnance Corps. While rate measurements were being made, t,he system
(3) H.H. Franck and W. Schirmer, 2. Eleklrochem., 64, 254 (1950). containing about, 1 ml. of liquid sample was maintained
(4) H.S. Johnston, L. Foering, Y. Tao and G. H. Messerly, J . A m . constant a t approximately 25% ulIageI2 by making small
Chem. Sac., 73, 2319 (1951). adjustments on a plunger in the hydraulic system, and the
(5) H.S. Johnston, L. Foering and R. J. Thompson, THISJOURNAL,
57,390 (1953). (9) G. D. Robertson, Jr., D. 111. Mason and B. H. Sage, Ind. Eng.
(6) D. M . Mason, L. L. Taylor and H. F. Keller, Jet Propulsion Chem., 44,2928 (1952).
Laboratory Report No. 20-72, Pasadena, California, 1953. (10) 9. Lynn, D. M . Mason and B. H. Sage, ibid., 46, 1963 (1954).
(7) W. R. Forsythe and W. F . Giauque, J. A m . Chem. SOC.,64, (11) An inert, fluorinated hydrocarbon liquid manufactured by
48 (1942). Hooker Electrochemical Co., Niagara Falls, N. Y.
( 8 ) D . M. Yost and H. Russell, Jr., "Systematic Inorganic Chemis- (12) Ullage is a term used to describe the percentage of the gas
try," Prentice-Hall, Inc., New York, N . Y., 1944. phase volume relative to the total volume of the system.
684 JR.,DAVID
GLENND. ROBERTSON, M. MASONAND WILLIAMH. CORCORAN VOl. 58
pressure was recorded with time. The rate of increase of the condensable components, as obtained from Fig. 2,14-11
pressure with time can be related to the rate at which nitric were constant during the measurements. By use of thc
acid decomposes as described in the section which follows. st,ated assumptions the following ex ression relating the
Fluorolube S was observed to be quite inert with respect rat,e of decomposition of "03, -d [%N03]/dt, to the ob-
to "0s a t eleyated.temperatures, and HzO and HN03 served pressure rise, dP/dt, was obtained. For a given
are immiscible with this liquid. NO2 appears, however, to value of ullage U,and Henry's law coefficient ct
be soluble in all roportions. This solubility could vitiate
experiments ma& over extended periods of time. How-
ever, since only the init,ial rate of the forwardSeaction of
equation 1 was of primary interest in this investigation, Pressure-time data in the figures and initial-rate-of-
experiments did not exceed 1hour in duration. In this time pressure-rise data in the tables were converted from the
the loss of NOz from the system by diffusion into Fluorolube esperimental value of ullage to an ullage of 25% so that they
S was insignificant. The method used is capable of re- are on a comparable basis. It was assumed that the volume
producing rates to within *I%. of liquid was independent of ullage in the range of ullnges
For convenience and to prevent contamination by mois- studied from the experimental ullage to the 25% ullage.
ture in the air, liquids were transferred by means of glass
syringes fitted with long stainless steel hypodermic needles. 5.00
The presence of dissolved air in the samples was not pre-
cluded.
Calculation of Decomposition Rate from Pressure-Time
Measurements.-It was assumed that while the experi-
ments were being conducted the gas phase was in continuous 2.00
equilibrium with the liquid phase and contained 02,HN03,
NO2, H20 and traces of inert gas (primarily N2) from air
initially dissolved in the liquid phase. Under isochoric,
isothermal conditions the pressure was measured as a func- 1.oo
tion of time. Decomposition in the gas phase was assumed
to be negligible. It was assumed that the gas phase obeyed
the perfect gas law and Dalton's law. From the stoichiom-
etry of equation 1, i t is evident that, for each mole of 02 0.50
formed, 4 moles of HN03decomposed and that cd
4

-d (moles "Os) = d (moles 02) f

dt dt (3) 2
Some of the 0 2 formed passed into the gas phase, and the Ea 0.20
remainder was dissolved in the liquid phase. The solubility
of 0 2 in fuming nitric acid is summarized in Fig. l . 1 3 0 2 3
0.10
6.6

0.05
6.2 0 KLEMENC AND RUPP

4
42 9 PERRY AND DAVIS
22 5.8 0.02 1 I
I
I I

-8- 5.4
2 20 40 60 80
X Temp., "C.
u 5.0 Fig. 2.-The effect of temperature on the total pressure of
+ vapor above fuming nitric acid.
8
.d

Nomenclature
4x 4.6 d differential operator
E, experimental apparent activation energy in kcal./mole
3 obtained from the variation of experimental rate with
3
$ 4.2 temp.
4
k s ecific rate constant
r/;l
K tgermodynamic equilibrium constant
[Od: CONCENTRATION OF
In natural logarithm
3.8 OXYGEN IN ACID, - P pressure (atm.)
MOLES PER LITER R universal gas constant
p0 = TPARTIAL PRESSURE of t time (sec.)
E N , ATMOSPHERES T temperature ("K., unless otherwise specified)
I 1 I I I U ullage, defined as Va/(Va +
VL)X 100 (%)
Henry's law coefficient (moles/l. atm.)
40 50 60 70 80 O(

Temp., "C. Materials

Fig. 1.-The solubility of oxygen in fuming nitric acid. HN03 was repared by vacuum distillation a t 40' from
a mixture of &03 and H2S04.7 The product was condensed
obeys Henry's law under the conditions encountered here. a t -79' and subsequently stored under refrigeration a t
It was assumed that during the initial stages of decomposi-
tion of HMO8 the ullage and Henry's law coefficients were (14) A. Klemenc and J. Rupp, Z. anorg. &?em. Chem., 194, 51
constant. It was also assumed that the vapor pressures of (1930).
(15) G . B. Taylor, Ind. Eng. Chem., 17,663 (1925).
( 1 3 ) G. D. Robertson, Jr., D. M . Mason and W. H. Corcoran, ( 1 6 ) E. P . Egan, Jr., ibid., 37, 303 (1945).
to be published in I n d . Eng. Chem. (17) J. H,Perry and D. S, Davis, Chern. Met. Eng., 41, 188 (1934).
August, 1955 KINETICSOF THERMAL
DECOMPOSITION
OF LIQUIDNITRICACID 685

-30". The acid thus prepared was analyzed for HzO

by acidimetric titration with NaOH and was always found
to contain less than 0.001 weight fraction HzO. Previous
experience with optical absorbancelo has shown that quanti-
ties of the species NO, too minute to be determined easily
by chemical analysis were sufficient to impart a definite
yellow color to HNOa, which itself is water-white a t 0".
Therefore no specific test for NO, was made, but the acid
was considered to be free of this species if, before a test,
the sample lacked any trace of yellow. Commercial NO2
was fractionated and dried over P z O ~ . Distilled HzO was
obtained from the available laboratory supply.
Materials used in the studies of the effects of additives
on the liquid phase decomposition of HNOa were obtained
from various sources. Anhydrous ",NO,, KNOa and
KHS04 of reagent-grade quality were carefully dried under
vacuum to remove any traces of adsorbed moisture. NOH-
SO4 was prepared from NO2 and concentrated in
accordance with a method tested by Elliott, et aZ.,Is NzOs
containing about 0.05 weight fraction of a mixture of
HNOI and NO,, was obtained by distilling HNOa to which
an excess of P~05 had been added.8
Anhydrous HzS04 was prepared by mixing concentrat,ed
H2SOaand SOa until an acidimetric titration indicated that
an excess of 0.005 weight fraction SO3 existed. Commercial
concentrated HClO4 containing 0.30 weight fraction H,O
was mixed with an excess of HzSOc and distilled to yield a
1 oduct which contained less than 0.006 weight fraction
H' .
,O as indicated by titration with a standard solution of
NaOH
ALL DATA CORRECTED
25 PERCENT ULLAGE
T O CORRESPOND

Results I
I
I
I
I
400 800 1200
Experimental Verification of Phase Equilibrium. TIME SECONDS.

-Original experimental measurements of pressure Fig. 3.-The effect of various additives on the initial p:essure
as a function of time for 25 tests are available rise due to decomposition of nitric acid at 87.8 .
elsewhere.19 From these tests rates of pressure the experimental conditions supersaturation of 0 2
rise with time a t zero time were calculated from the occurred in that phase. The pressure-time data
slope of the pressure-time curves corrected to without agitation were irregular, and, in most in-
%youllage at 71.1". The results are summarized stances at a given time the pressure without agita-
in Table I. In Table I1 the initial rates of decom- tion had a value less than that with agitation, indi-
position in terms of moles HN03/liter sec. are cating that where agitation was not present super-
shown and were calculated from the corresponding saturation of 0 2 in the liquid phase occurred. Ki-
initial rates of pressure rise obtained from pressure- netic data were thus obtained with agitation, and the
time curves adjusted to 25% ullage. Examples of rate of agitation was assumed to be sufficient to
pressure-time curves at 87.8" from which initial maintain physical equilibrium between the gas and
rates of pressure rise were obtained are shown cor- liquid phases.
rected to 25% ullage in Fig. 3. The increased pres- Homogeneity of Liquid Phase Reaction.-To
sures obtained with NHdN03 were attributed to the show whether or not the liquid phase reaction was
formation of N2 and N 2 0 as a result of decomposi- heterogeneous , the liquid-glass interfacial area
tion of ",NO3 in "03. The linearit,y of the was increased tenfold by adding glass wool to the
initial portion of the curves in Fig. 3 indicates that liquid phase. The presence of the glass wool pre-
the reverse of the reaction of equation 1 was not vented agitation and thus permitted supersatura-
taking place to any appreciable extent during the tion of the liquid phase by 0 2 . This effect was
period in which the initial rates were being deter-
mined. As the reverse reaction becomes more I
TABLE
prominent, curvature is observable, as is evident SUMMARYOF THE EFFECTOF VARIOUS ADDITIVES
ON THE
in the curves of Fig. 3 a t high values of time. INITIAL
RATE OF PRESSURE RISE IN LIQUIDNITRICACID
Eventually as equilibrium with respect to equation MIXTURESAT 71.1'
1 is attained, the rate of change of pressure ap- Initial rate
Amount, of pressure
proaches zero. Test mole rise,a atm./
To determine the practicability and limits of the no. Additive frac. 24 hr. d a y
method chosen to follow the course of the reaction, 21 NH~NO~" 0.0982 2080
several preliminary experiments were performed, 14 N206 .lo3 1570
For example, since O2 bubbles visibly formed within 3 None ... 194
the liquid phase, it was necessary to know if under 12 Nz04 .111 90
(18) G. A. Elliott, L. L. Kleist, F. J. Wilkins and H. W. Webb, J .
15 KNOi .lo5 57
Chem. Soc., 1219 (1926). 16 KHSO4 .lo9 44
(19) Material supplementary t o this article has been deposited a s 1s HzSO4 .lo8 25
Document number 4550 with the AD1 Auxiliary Publications Project,
Pliotoduplication Service, Library of Congress, Washington 25, D. (3.
17 NOHSOa .loo 20
A copy may be Becured by citing tlle Docititlent nuinher a n d by
19 HClOJ ,090 16
reliiitting $1.25 for ~~liotoiirints,
or $1.25 for 35 iniii. niicrofiliii in ad- 7 HzO ,099 10
vance b.v chock or tiioney order ilayahle t o : Chief, Photoduplication Tho initial rrrte of risr is givcn per 24-hr. day for acid
Yeryicc, Library of Congrcm. stored at 25% ullage. Measurements made a t 87.8".
(380 GLENND. ROBERTSON,
JR., DAVIDM. MASONAND WILLIAMH. CORCORAN Vol. 59

TABLD I1 artificial light of the same intensity as that which

OF INITIAL
SUMMARY OF NITRIC existed in the studies using the high pressure capil-
RATDOF DECOMPOSITION
ACIDIN THE LIQUIDPHASE lary tube. The constancy of pressure with time
d[HNOa1 for this system at 80" indicated that no appreciable
dt decomposition occurred.
Comgn.a of Canon. of sdditive,b x 106, Data obtained by Johnston, et aZ.,4Jfor the gas
Test the system, additive, mole Temp., moles/l.
no. wt. frao. moles/l. frac. OC. aec. phase thermal reaction indicate that it is heteroge-
1 1.oo "03 0 0 54.4 1.49 neous in nature below 300" and slow compared with
2 1.00 "01 0 0 . 62.8 5.06 the liquid phase reaction at the temperatures inves-
3 1.00 "03 0 0 71.1 16.1 tigated in the present study. The lowest tempera-
4 1.00 "08 0 0 87.8 140 ture investigated by Johnston, et aZ., was about
5 0.0088 HzO 0.69 0.030 71.1 3.34 130". An extrapolation of the reported rate con-
6 ,0088 HzO 0.68 .030 87.8 33.4 stants t o the highest temperature of the present
7 ,0304 HzO 2.39 .099 71.1 0.85 investigation, i s . , 87.8", yields the value of 2 X
8 .0304 HzO 2.34 .099 87.8 10.5 mole/l. sec. for the rate of disappearance of
9 ,0600 HzO 4.59 .183 87.8 1.84 "0, caused by gas phase decomposition. Thus (1

10 .0509 Nor 0.80 .035 71.1 11.1 at 25y0 ullage and assuming gas phase decomposi-
11 .0971 NOz 1.56 .069 71.1 8.60 tion Johnston's data give an upper limit on the
rate of "03
12
13
14
.0866 NOz
.165 NrOs
2.50
'"')
1.38
2.5
::::)
,111
71.1
,103 71.1 131
71.1
2.51
6.30 decomposing in the gas phase in con-
tact with 1 liter of pure liquid HNOa as 0.7 X
mole/sec. In the present investigation a rate
of 140 X mole/sec. (cf. test 4 in Table 11) was
15 ,1586 KNOB 2.6 .lo5 71.1 4.78 observed under these conditions. Thus it may be
16 ,2085 KHSO4 2.7 ,109 71.1 3.85 inferred that liquid phase decomposition is control-
17 .1836NOHSOi 2 . 5 ,100 71.1 1.60 ling. In the heterogeneous gas phase reaction the
18 .1584HzS04 2.4 ,108 71.1 2.03 energy of activation is about 5 kcal./mole, whereas
19 .1363 HCIO4 2.0 ,090 71.1 1.31 in the liquid phase decomposition,as isdiscussed sub-
20 .1363HClOi 2.0 .090 87.8 13.3 sequently, the experimental energy of activation lies
a When not expressed, the remainder of the sample in the range 32 t o 37 kcal./mole. A consideration of
consists of "08. b These values of concentration and the latter energy shows the large temperature de-
mole fraction for the additives are given on a formal basis pendence of the liquid phase decompositionand indi-
for the molecular species indicated. cates again that the gas phase reaction is not con-
minimized, however, by conducting the experiment trolling. It has thus been assumed that the rate
at 87.8" where the lack of external agitation was of decomposition in the gas phase is negligible in
shown to have a small effect. The data obtained the results that are presented.
General Effect of Additives on the Rate of Ther-
when the glass wool was present agreed very closely mal Decomposition.-Table I1 gives a summary
with those which were obtained in the absence of of the initial rates of decomposition of " 0 3 in
glass wool. This agreement indicates that the re- the liquid phase at the temperature specified in
action which occurred in the liquid phase was homo- the range 54.4 to 87.8". In this table composi-
geneous with respect to a glass surface. tions are expressed both in mole fraction and in
Predominance of Liquid Phase Reaction.-The moles per liter. Density data for t.he "03-
visual evidence that bubbles of 0 2 originated NOZ-HzO system used to convert composition from
throughout the liquid phase as decomposition weight fraction to mole fraction were obtained from
proceeded supports qualitatively the fact that in the work of Klemenc and RuppI4 and that of Berl.21
the range of temperatures investigated, 54 to 88", The rate data were derived from the pressure-time
the liquid phase reaction was controlling. Also datal9 by the method previously described. In-
there is evidence that the initial rate of pressure cluded in the data are the effects of various addi-
rise in pure H N 0 3 increased with a decrease in tives (HzO, NOz, Nz06, KN03, KHSO4, NOHS04,
ullage. This fact further supports a liquid phase HzS04and HC1O4) on the rate of decomposition. It
controlling reaction because the concentration and is seen from Tables I and I1 that, whereas it was
consequently the rate of thermal decomposition of possible to inhibit the reaction with some of the r
HNOI in the vapor phas aree not a function of ul- additives, this inhibition is insufficient to prevent
lage. It is known, however, that HNOI vapor can extensive decomposition for prolonged storage
decompose thermally4J and under the influence of periods.
light.20 Reynolds and Taylor20 showed that pho- HzO is most effective as an inhibitor. Figure 4
tochemical action is effective only for the decompo- depicts the effects of HzO concentration on the ini-
sition of HNO, in the gas phase and that the corre- tial rate of decomposition of HN03at temperatures
sponding rate is of a much lower order of magni- of 71.1 and 87.8". As the mole fraction of HzO
h d e than that of the liquid phasg thermal decom- is changed from 0 to 0.183, corresponding to 0 to
position being studied. To demonstrate that the 4.59 molea/liter, there is an 18-fold decrease in the
rate of the photochemical gas phase reaction was initial rate of decomposition of the "03.
negligible under the conditions encountered in the
present work, HNO, vapor a t about 35 mm. of mer- Where solid salts were added, the partial molar
cury in a 250-cc. spherical flask wab subjected to volumes were assumed to be negligible, and additive
volumes were used where &So4 and HClOl were
(20) W, G. Reynolds and W. H. Taylor, J . Chew&. S a c . , 101, 131
(1912). (21) E. Berl, Chem. M e t . Eng., 46, 234 (1939).
August, 1955 KINETICSOF THERMAL OF LIQUID
DECOMPOSITION NITRICACID 687
200

100

'I 50 I
3
-8 20
LD-

2 10
X
5
L
a
1 2

0.5
ALL DATA CORRECTED TO CQYIESPOND
1 2 3 4 5 TO 25 PERCENT ULLAGE
Formal water concn., moles/l.
Fig, 4.-The effect of water on the initial rate of decomposi-
tion of nitric acid.
2 4 e
introduced. For the kinetics studies most of the TIME Xm'3 SECONDS.
data on initial rates were obtained at 71.1°, since at Fig. 5.-The effect of temperature on the initial pressure rise
87.8" the rates with pure HNOI were too great to due to decomposition of pure nitric acid.
allow accurate determination.
Effect of Temperature on Rate of Thermal De- different temperatures for pure HNO, (tests 1 to 3).
composition.-To determine the effect of tem- Curves are also shown for mixtures of HN03 with
perature upon the rate of thermal decomposition, H 2 0 (tests 7 and 8) and with HClOl (tests 19 and
rate measurements with pure HNO, were made at 20). The values of E , calculated from the slopes of
54.4' (test l), 62.8' (test 2), 71.1" (test 3) and the straight lines for these three systems are pre-
87.8" (test 4). The pressure-time curves for these sented in Table 111.
tests are shown in Fig. 5. The rate of decom- 20 I I J I I I
position of "03 at 87.8" was so great that the
resulting data are not considered reliable for obtain-
ing quantitative values of initial rates. By the
method described previously, the initial slopes of 10
the curves in Fig. 5 are expressed in terms of the u'
initial rate of decomposition of "03. If the varia- P
tion of specific rate constants with temperature
were knwm, the Arrhenius equation could be used
to determine the energy of activation of the rate-
determining step in the decomposition. In this .n-

investigation the actual concentration of the react- 2 2

ing species was not determinable because of the
unknown extent of various equilibria existing in
X
z/a 1
6
-
HN03 solutions, as is discussed in a subsequent
section. Consequently no specific rate constants a
x
could be calculated. An apparent energy of activa- I
tion E, was obtained, however, by employing a re-
lationship of the form of the Arrhenius equation22
-0 HNO,+ 0.030 WEIGHT FRACTION HO ,
0- HNO, + 0. I 3 6 WEIGHT FRACTl,ON HCIO,
(5)
where
rate = - d ["Oal
~

dt (6)
2.8 2.9 3.0 3.1
1000/T (OK.-').
in moles per liter per second. Fig. 6.-The variation with temperature of the rate of de-
Figure 6 shows the results of a plot of the log of composition of nitric acid mixtures.
initial rate of decomposition versus 1000/T at three Analysis of Results
(22) For the sake of brevity, the term rate is used throughout this
report t o designate the rate of decomposition of "01. Since the
Possible Role in the Reaction Mechanism of
liquid phase reaction is controlling, the rate is expressed in terms of Various Equilibria in HN03 Solutions.-The work
moles of "Os decomposed per liter of liquid "01. of Franck and Schirmer3 where they proposed
688 GLENND. ROBERTSON,
JR.,DAVIDM. MASON
AND WILLIAMH. CORCORAN Vol. 59

TABLE I11 ?ZHN03 HzO C-,(HN0a)nHnO (13) +

APPARENTENERGY NITRIC
OF ACTIVATIONOF FUMING The existence of the foregoing equilibria which
ACID affect the actual concentration of N20sexisting in
EP solution prevents Franck and Schirmer's3 first-order
Test no. System koal./mole
1 to 3 Pure HNOa 32 rate expression from being confirmed without know-
19 and 20 +
H N 0 3 0.136 wt. frac. HC104 34 ing the actual concentration of NzOS. Consistent
7 and 8 +
HN03 0.030 wt. frac. H20 37 with their data and assumptions is the possibility
that the rate-determining species may not be N206
that the decomposition of "03 proceeds by the but instead may be one or both of the products of
first-order decomposition of Nz06 was done before its ionic dissociation such as NO3- or NO2+. In en-
the recent evidence for the self-i0nization~3-~~ of suing discussions the equilibria in reactions (7)
pure "03 and other equilibria in HN03 solutions through (13) are assumed to bo attained very rap-
was available. The newer data should be examined idly as compared with the rate of decomposition of
to see what bearing they may have on the mech- "03.
anism of decomposition. Inhibiting Effect of HzO and Accelerating Effect
It has been shown that the following ionic equilib- of NzOson Reaction Rate.-It was experimentally
rium exists in HN03 containing less than about observed (cf. Fig. 4) that a marked decrease in
5% HzO rate resulted from the addition of a relatively small
IC,NOn+ Nos- + HzO+ (7) amount of HzO to "03, the effect decreasing with
increase in water added. By the reaction of
The fractional ionization of pure HN03 a t 0" has equation 7 an increase in HzO tends to increase the
been determinedz8 from optical absorbance meas- concentration of HN03 and correspondingly of
urements to be about 0.05. This ionization may HzN03+(cf. eq. 8) whereas by the reaction of equa-
occur in two steps.2s The first step is a proton shift tion 13 an increase in H20 reduces the "03 and
resulting in the formation of the nitricacidium ion HzNO3+ concentration. The marked change in
HzNO3+and Nos-, that is rate of decomposition of "03 for such a relatively
2"Oa HzN03+ + Nos- (8) small added amount of H,O appears t o eliminate
Subsequently H2N03+almost completely dissoci- HN03 in the rate-determining step since the con-
ates into the nitronium ion NOz+ and HzOas centration of "0, is little affected by such small
additions of water. Although the quantitative ef-
f,NO*+ + HzO (9) fect of HzO on the concentration of H2N03+de-
and the net result is given by equation 7. pends upon the interdependent equilibria of equa-
NOZz7ionizes in H N 0 3 solution in accordance tions 7 through 13, the net qualitative effect is that
with the equation of increasing the amount of H2N03+,as is expressed
2N02 Nn04 NO+ + NO,- (10)
by equation 9. Such behavior eliminates %NO3+
as the rate-determining species. The effect of
The fractional ionization of NO2 in HNOa solutions HaO in decreasing the rate is in keeping with the
is quite large, being about 0.7 according to optical corresponding reduction of NO2+by equation 7 or
absorbance measurements.28 Nz06 also ionizes Nz06 by the reaction
to a large extent in HNO, s o l ~ t i o n sin ~ ~the
, ~ fol-
~ 2HNOs N2Os + HzO (14)
lowing ma,nner
NzOb NOn' + NOa- (11) That either N206 or NO2+ may be the rate-de-
termining species is also qualitatively supported by
The equilibrium position of equation 11 has been the marked increase in decomposition rate observed
shown26 to lie far to the right, even a t N206 concen- with the addition of NaOs (test 14 in Table 11)
trations as high as 0.G molar. "03 also is known which ionizes t o give NOz+ and NO3- by equation

2IINO3 + Nos- __
to undergo a solvation reaction with Nos- from 11. Thus a mechanism involving either NzOn
compounds such as E(N0329~30
(HNOa)2N03-
It is likely that NOa- formed by the ionization of position
NOz (cf. eq.10) reacts with HN03 in accordance mechanism
(12)
or NOz+ as the rate-determining species is in keep-
ing with the experimentally observed effect of H?O
reducing and of NzOs increasing the rate of decom-
of HN03, as is shown in Table 11. The
2
with equation 12. involving the first-order decompoai tioi I
Ramaii spectral and thermal studies30 have of N206 is ambiguous since the expression
shown that the following solvation of HNOI and - ["031 dt = k l s [NOz+l[N03-l (15)
HzOoccurs in concentrated " 0 3 solutions
(23) C . K. Ingold and D. J. Millen, J . Chem. Soc.. 2612 (1950). is indistinguishable from the equation
(24) W . J. Dunning and C. W. Nutt, Trans. Faraday So&, 47, 15
(1951). - ["031 = kla [N206] (16)
(25) E. G. Taylor, L. M. Lyne and A. G. Follows, Can. J . Chem., dt
29, 439 (1951).
(26) R. ,J. Gillespie, E. D. Hughes and C. K. Ingold, J . Chem. Soc.,
because of the interdependence of NOz+, NO$- and
2652 (1950). NzOsin the rapid equilibrium of equation 11.
(27) J. D. S. Goulden and D. J. Millen, ibid., 2620 (1950). Inhibition of Reaction Rate by HzS04 and
(28) S. Lynn, D. M. Mason and W. H. Corcoran, THISJOURNAL,HC1O4.-0ne means of differentiating between the
69, 298 (1955).
(29) J . Ch6din and S. B'BnnBnt, C o m p t . rand.. 228, 242 (l!l49).
mechanism involving N206 or NOz+alone is to add
(30) D. M . hIason, L. L. Taylor and S. .'l Vanyo, A n d . Cheni., i n components to the system so that NOz+is increased
press. and Nz06 is decreased in concentration; Has04
August, 1955 KINETICSOF THERMAL
DECOMPOSITION
OF LIQUIDNITRICACID 689

and HClO4 are such additives. It has been shown volving association of NO3- with HN03. At very
from Raman spectra that acids which are stronger low concentrations of NO3- it is expected that the
than H N 0 3 can donate a proton t o "03 to form inhibition of the reaction due to the salt effect would
H2N03+which dissociates, as is shown in equation be reduced and that a net increase in rate of decom-
scnted by
HNOi

Nz06
+
+ &SO4
__
9. The effect of adding H2S04 to "03

NOz+
is repre-

+ HzO + HSO4-
~ i i dthis equation combined with equation 14 gives
NOz+ + HSO4- + HNOa
(17)

(18)
position of HN03would occur.
(6.
The marked decrease in rate caused by NOHSOd
Table 11) indicates not only the salt effect
shown by KHS04 but in addition a greater inhibit-
ing effect, probably because NO+ reacts with NO,-
to form the species NOz and Nz04 in accordance

__
with equation 10. This decrease in N01- requires
'l'tm effect of adding HC104 to "03 is represented a shift to the right in the equilibrium of equation 11
1 'Y which lowers the concentration of N206 and, at the
HNOI + HClO4 NOz' + HzO + clod- (19) same time, increases NO2+. Thus in agreement
with the hypothesis that N206and not NOz+is the
and equation 19 combined with equation 14 gives rate-determining species, NOHSOd should cause
NzOb + HC104 NO$" + clod- + HNOa (20) more of a reduction in rate than KHSO4.
Apparent Energy of Activation of Reaction.-
From the high concentration of NO2+ detected by If the actual concentration of the molecular species
Raman the position of the equilibria in NzOswere known, a specific rate constant kle could
equations 18 and 20 lies very far t o the right. be calculated from the observed rate of decom-
It is seen from Table I1 that both H2S04 and HClO4 position, and the energy of actjvation E could then
markedly decrease the rate of decomposition of be obtained from the Arrhenius equation. How-
HN03. This behavior is in keeping with the hy- ever, since the actual concentration of NzOs is not
pothesis that N206,and not NO2+ alone, is the rate- known, the apparent activation energy defined by
determining species since both H2S04 and HC104 re- equation 5 is used.
duce the amount of NzOsand increase the amount of The experimental apparent energies of activa-
NOz+. tion obtained from the slopes in Fig. 6 are shown in
Effect of Several Salts Including Nitrates on Table I11 to vary between 32 and 37 kcal./mole.
Reaction Rate.-Both neutral salts and nitrates These data were obtained for pure HNO,, for an
(to a lesser extent) are found to decrease rate. "03 solution containing 0.136 weight fraction
Since the equilibria in equations 17 and 19 lie HC104, and for a solution containing 0.030 weight
far to the right, ms04 and KC104 can be con- fraction H20. Within experimental precision the
sidered neutral salts in "03, and they would exert apparent energies of activation for these different
little mass action effect on ionic equilibria such as HN03 solutions are in close agreement, and it can
those represented by equations 7 through 13. therefore be implied that H N 0 3js decomposing in a
Of these two salts only KHS04 was sufficiently similar fashion in these different solutions. A value
soluble (0.1 mole fraction) to permit its use for for the energy of activation of about 24 kcal./mole
evaluating the salt effect in this investigation. The for the decomposition of "03 was reported by
reduction of the decomposition rate by the salt Franck and Schirmer.3
effect which is caused by the addition of KHSOA
to HN03 is shown in Table 11, where it can be Conclusions
compared with other additives. The normal re- This investigation presents the effects of several
sult of an increase in ionic strength is to decrease additives on the rate of the thermal decomposition
the rate of a reaction occurring in solution between of HN03in the liquid phase. It has been shown that
two ions of unlike charge. Thus the observed de- the presence of any of these additives in quantities
crease in rate caused by the inert salt KHS04 is to up to 0.1 mole fraction does not inhibit the rate of
be expected from a rate expression (cf. eq. 15) con- decomposition sufficiently t o prevent the eventual
taining unlike ions. However, theories valid only approach to equilibrium for storage periods of the
at low ionic strength are not strictly applicable to order of months a t elevated ambient temperatures.
HN03 solutions since a t 70" the self-ionization of On both a molal and weight basis HzO was the most
HN03 alohe is sufficient to give an ionic strength effective inhibitor tested.
probably greater than 0.3 m~le/liter.*~ The kinetic studies suggest that the rate-deter-
It is evident from Table I1 that, for amounts of mining step in the decomposition reaction involves
either NO2 or K N 0 3 which are equivalent to the the first-order decomposition of the molecular spe-
neutral salt KHSOI, less inhibiting effect occurs cies N?06or the interaction of its products of ioni-
with the NO2 and KN03. Thus a net catalytic ef- zation, NOz+ and Nos-. Which of these mech-
fect which over-shadows the neutral salt effect is anisms prevails is indeterminable by the present
evidenced by these compounds which give NO3-. methods of measurement. It might be speculated
Such behavior would be expected if NzOs and not that N2O6 decomposes t o NO2 and NO, molecules
NO2+were the rate-determining species since equa- and that NO3 in turn yields NO2 and 0. Extensive
tion 11 shows that NOI- increases N205 and de- investigation, however, is necessary to establish a
creases NOz+. The increase in N2Os by NO3- ac- definite mechanism. An experimental energy of ac-
cording to equation 11is not as pronounced as might tivation has been determined for the decomposition
be due to the competing reaction equation 12 in- of pure HN03 and for the HN03-HpO and "0,-
(31) C. K. Ingold, D. J. Millen and H. G . Poole, J . Chem. Soc., 2676 HC104 systems. To det,ermine the actual energy of
(1950). activation and to substantiate N205 as the rate-de-
690 WALTERE. NERVIK VOl. 59

termining species, i t is suggested that a physico- fraction. If further studies.in this field are under-
chemical technique such as measurement of optical taken, it is suggested that more dilute solutions be
absorbance in the ultraviolet range32might be em- investigated. Such a procedure permits the use of
ployed to determine the concentration of NzOsand additives such as KC104 which, by virtue of their
perhaps NOa- present in “Os. low solubility in HN03, could not be investigated
The effects of the various additives used in this in the present work. Moreover, in dilute solutions
investigation were emphasized by introducing them the neutral salt effect of nitrates may be diminished
at relatively high concentrations of about 0.1 mole t o the extent that the expected acceleration of the
(32) R. N. Jones and G . D. Thorn, Con. J . Res., 27B,580 (1949). rate of decomposition by NOs- could be observed.

AN IMPROVED METHOD FOR OPERATING ION-EXCHANGE RESIN

COLUMNS IN SEPARATING THE RARE-EARTH ELEMENTS1
BY WALTERE. NERVIK
University of California, Radiation Laboratory Livermore Site, Livermore, CaE.
Received January 14, 1066

A ‘ I radient elution” method has been adapted for operating an ion-exchange resin column using an eluting agent in which
the p f I is changed continuously. The method shortens the column operating time required for separation of the rare
earth elements without adversely affecting their purity.

I. Introduction the individual rare earths in relatively short column

Procedures for separating individual members of operating times.
the rareearth group of elements through the ap- Although the author was not aware of it a t the
plication of synthetic cation exchange resins have time, the elution technique which was finally devel-
been described by many investigators.2-6 The ex- oped was quite similar to the “gradient elution”
act experimental conditions in each case depended chromatographic method originally described by
on the number of elements to be separated and the Alm, Williams and Tiselius.6 In addition, the
degree of separation desired. I n general, the rare physical arrangement of the various reservoirs and
earth elements were first adsorbed on the top por- column equipment components was similar to that
tion of a long narrow column of cation-exchange used by Busch, Hurlbert and Potter.’
resin, then selectively eluted by passing a solution Experimental
containing a suitable complexing agent through the
resin column and collecting portions of the eluent. Materials.-Commercial 12% cross-linked Dowex-50
cation-exchange resin of “minus 400” mesh size was graded
The degree of separation for a given experimental to obtain that portion which settled between 1.0 and 1.5
arrangement depends on variables such as tempera- cm./min. in distilled water. The resin was washed with 6
ture, type of cation-exchange resin, size of the resin M ammonium thiocyanate until the red ferric thiocyanate
particles, dimensions of the column, eluting agent color was no longer visible, then washed in turn with dis-
water, 6 N hydrochloric acid and distilled water a ain.
used, concentration and p H of the eluting agent, tilled
Finally, the resin was converted to the ammonium form
flow rate of the eluting agent, etc., but for a given with 1 M ammonium lactate and stored in distilled water
eluting agent the relative position of the various until loaded on the column.
rare earths remains approximately constant.8 An Baker The eluting agents were prepared from conductivity water,
and Adamson reagent grade 85% lactic acid, and con-
experimental consequence of this phenomenon is centrated ammonium hydroxide. All solutions were 1 M
the fact that if operating conditions are adjusted in total lactate concentration with the p H adjusted with con-
to effect separation of the heavy rare earths, the centrated ammonium hydroxide and measured on a Beck-
light rare earths require excessively long column man Model G pH meter. The solutions were also made
0.01 M in phenol t o prevent deterioration
operating times before being eluted. If conditions approximately
of the lactate.
are adjusted to give separation of the light rare Rare earth activities were obtained by bombarding a
earths in reasonable operating times, the heavy rare target consisting of alternate layers of pure uranium and
tantalum metal foil with 340 MeV. protons in the 184-inch
earths are usually eluted too quickly to give high Berkeley synchrocyclotron.
purity. Ordinarily the chemist has to make some in a solution of concentratedThe target foils were dissolved
hydrofluoric acid and nitric
compromise between purity of individual rare acid to which 5 mg. of lanthanum carrier and 1 mg. each of
earths and column operating time. This series of strontium, zirconium, niobium, molybdenum, ruthenium,
experiments has been undertaken to determine rhodium, palladium, tin, antimony, tellurium and barium
carriers had been added. The mixed rare earth activities
whether introduction of an additional variable, were then purified as a group by fairly well-known chemical
Le., continuous change in pH of the eluting solution, steps which included a fluoride precipitation in the presence
could be used to effect complete separation of all crf of dichromate, a barium sulfate “scavenge” in dilute acid
solution, precipitation of the rare earth hydroxides with
(1) This work was performed under the auspices of the U. S. Atomic ammonium hydroxide, assage of the rare earth group
Energy Commission. through a small Dowex 1-1 anion-exchange resin column in
(2) B. H. Ketelle and G . E. Boyd, J . Am. Chem. Sac., 69, 2800 concentrated hydrochloric acid (to remove those elements
(1947).
(3) 9. W. Mayer and E. C . Freiling, ibid., 76, 5647 (1953). (6) R. 1.A h , R . J. P . Williams and A. Tiselius, Acta Chem. Scand.1
(4) F. H. Spedding, J. E. Powell and E. J. Wheelwright, ibid., 76, 6, 825 (1952).
012 (1954). ( 7 ) H. Busch, R. B. Hurlbert and V. R . Potter, J . Bid. Chem., 196,
(5) E. C. Freiling and L. R. Bunney, i b i d . , 76, 1021 (1954). 717 (1952).