Beruflich Dokumente
Kultur Dokumente
Catalytic decomposition of nitric oxide a t a concentration of 0.404 and 0.432% by volume the central section of the reactor to a maxi-
i n nitrogen was studied in a tubular flow reactor. The packing consisted of alumina pellets mum depth of 15.5 in. A uniform tempera-
impregnated with 0.1% by weight of platinum oxide and 3.0% by weight of nickel oxide. Tests ture was maintained in the catalyst bed for
were conducted at pressures of 1 to 15 atm. and temperatures from 800' to l,OOO°F. A rate all lengths up to the maximum by means of
equation correlating the data as a function of temperature, pressure, and compositions was de- heaters mounted externally.
veloped. A reaction mechanism compatible with the rate equation is proposed.
Reoctor Accessories
Although the catalytic decomposition Their tests were mainly qualitative Figure 1 shows a schematic diagram of
of nitric oxide is of both practical and and were conducted for the purpose of the reactor system. In order to allow con-
theoretical interest, there is very little applying the results to the treatment ditioning of the feed gas before entrance to
information on the subject in the liter- of exhaust from automobile engines. the reactor a preheater section consisting of
ature. The bulk of the experimental Later Fraser and Daniels (6) in 1958 a 15 ft. length of %-in., stainless steel
studies on catalytic decomposition oc- tubing coiled to a length of about 9 in.
reported on the catalytic decomposi- was mounted just upstream from the re-
curred during the period between 1923 tion of diluted nitric oxide over metal actor. The outflow leg of the tube had a
and 1934. 'I'he et€orts were directed oxides with low surface areas. I n this heater mounted on its external surface
mainly toward decomposition of pure work the temperature was varied from which allowed a variable energy input up
nitric oxide at pressures of 1 atm. or 1,300" to 1,90O"F., and helium was to 700 w. A porcelain tube which had two
less and at temperatures from 1,100" used to dilute the nitric oxide to 10% 1,850-w. heaters mounted on its exterior
to 2,500"F. With the exception of by volume. A zero-order reaction rate was used as a shield over the outflow leg.
Muller and co-workers (11) and was obtained, and this was attributed The whole assembly of preheater and exit
Schwab and co-workers (17), who to the complete occupation of active tube was located inside a cylindrical shield
used a variety of metals and metal ox- made from a porcelain tube closed at the
sites because of the high concentration top. The external portion of this porcelain
ides for cataIysts, the investigators of nitric oxide. outer shield was wrapped with a 2,200-w.
mainly studied platinum catalysts. Some work has also been done on heater, and finally the whole assembly was
In general (8, 11, 17, 22, 23) the the high-pressure decomposition ( 3 , 4, insulated with microquartz and Fiberglas
data indicated that the reaction was 5, 7, 16) of pure nitric oxide. Studies and covered with aluminum foil.
first order with respect to nitric oxide were made at pressures up to 800 atm.
and was retarded by adsorbed oxygen. and temperatures up to 575°F. Prod- Meosurement of Pressure and Temperature
The rate equation to describe such be- ucts of the decomposition were found Pressures above 10 lb./sq. in. gauge were
havior could be written as follows: to be nitrogen, oxygen, and various measured with a gauge which had a maxi-
oxides of nitrogen. Mechanism and rate mum reading of 400 lb./sq. in. and was
studies were not made, but it was as- located at the exit of the reactor. For
certained qualitatively that the rate of pressures below 10 lb./sq. in. gauge a
decomposition increased with increases differential pressure gauge was used and
Another form of the rate equation was was connected by a manifold system to
presented as in temperature and pressure. six pressure taps equally spaced along the
The purpose of the present work length of the reactor. When used only for
was to develop further knowledge measurement of the exit pressure rather
about the kinetics and mechanism of than pressure differences along the reactor,
the catalytic decomposition of nitric the gauge was connected between the re-
Bachman ( 2 ) proposed from his stud- oxide, especially for the case of low actor exit and the manifold vent. Both
ies however that in the temperature concentrations of nitric oxide in nitro- gauges were stipulated by the manufacturer
range of 1,900" to 2,500"F. the reac- gen. Selection of the temperature range to be accurate to within 0.5% of the full
tion was second order with respect to of 800" to 1,000"F. was made on the scale reading, and this accuracy was con-
firmed by checks with a pressure balance
nitric oxide but was also retarded by basis of possible application of the re- and a mercury manometer.
oxygen to give a rate equation of the sults to the dissociation of nitric oxide Temperatures in the preheater, the line
form in the exhaust gases from internal com- connecting the preheater to the reactor,
bustion engines. and the reactor were measured with ten
chromel-alumel thermocouples. They were
EQUIPMENT checked over the range of 700" to 1,lM)"F.
In later work the A m o u r Research against a standard platinum-platinum-
Foundation (1) screened a large num- Reactor 10%-rhodium thermocouple which had
ber of catalysts to determine their ef- The reactor used in the flow studies of been calibrated by the National Bureau of
fectiveness in decomposing nitric oxide the decomposition of nitric oxide in nitro- Standards.
which was present in nitrogen at a gen was constructed from type 304 stainless Feed and Product Analysis
concentration of 0.2% by volume. steel tubing. It had an outside diameter of
~
% in., an inside diameter of 0.495 in., and Analysis of the feed and discharge gases
R. R. Sakaida is with Union Carbide Corpora-
tion, Linde Division, Tonawanda, New York. a length of 46 in. Catalyst was supported in was accomplished by means of a gas
g
samples was 20 to 35 min. and was changes in feed rate for a given weight
erally required for the system to reach of catalyst did not affect the basic
steady state. governed by both the analysis time
mechanism of the reaction, namely
About 1 hr. before the first sample that diffusion was not a rate-control-
was taken from the discharge stream ling factor, so that the curves were
the flow of the product stream was ad- graphically differentiated to obtain
justed to a predetermined setting on the rate in accord with the expression
the rotameter without changing the
pressure. As discussed this was accom-
plished by means of a regulator valve J
based upon Thiele's work (19) showed but reported as 1 atm. The catalyst pellets were weighed
that for fractional conversions greater For all other conditions of tempera- and counted before and after use, and
than 0.04 the fraction of surface avail- ture, pressure, and flow rate the maxi- these data are listed in Table 2. The
able was always greater than 0.93. mum pressure drop through the reac- ratio of the weight of catalyst charged
Values of r obtained as a function tor was less thau 0.01 atm., and this to the reactor to that after use was re-
of x were used to test a series of postu- change was neglected. producible to within ? 6%, and the
lated rate equations. The find test of Feed Rote weight after use was used to calculate
a selected rate equation was to com- The maximum errors in the measure- the w / F ratio noted in Equation (5).
pare the results of the integration of ments of feed rate varied from -+- 3% Loss in weight was because of loss of
Equation (5) with experimental data. at the lower rates to * 1% at the moisture and fines in the activation of
the catalyst.
Temperature Control
higher rates. These were largely due
to uncertainties in reading the rotame-
The main source of difficulty in con- ter scale. EXPERIMENTAL RESULTS
ducting the experiments was the con-
trol of temperature. Control was Feed and Product Composition Doto Correlotion
manual and subject to fluctuations in The composition of the feed gas Values of x were plotted as functions
line voltage which were not eliminated was either 0.404 or 0.432% nitric ox- of w/F for parameters of constant tem-
until the last part of the experimental ide in nitrogen to within & 1%.Analy- perature and pressure. Figures 2, 3,
program. When the reactor was oper- ses of the mixtures by gas chroma- and 4 are plots of data from studies
ated in the temperature range of 800" tography showed the absence of water with 9 or 18 g. of catalyst in the reac-
to 1,000"F. a difference in setting of and other impurities which could be tor. The solid curves in each figure are
approximately 5 v. was required to separated from nitrogen or nitric oxide. plots of a generalized equation which
change the temperature by 100°F. Impurities which were not separable is to be discussed. Tabular summaries
Once the operating temperature was from nitrogen or nitric oxide were of all the data collected are available
obtained for a given pressure, it was negligible. (14).
controllable to 2 2°F. at a point when The fractional conversion of nitric
no fluctuation in line voltage occurred. oxide in the reactor was obtained from
There were times when temperature the following relation between chro- TABLE2. DATAON PACKING OF G-43
CATALYST BEDSIN THE REACTOR
changes during intervals between sam- matogram areas:
Pack- No. Bed Weights,?
area % NO in feed stream - area % NO in product stream ing of height*, Input Output
X = - (7) no. pellets (in.) (6.)
area % NO in feed stream
1 250 4.0 10.14 8.99
ples were as great as 10°F. When Equation ( 7 ) states implicitly that the 2** 500 8.0 20.05 17.97
3 500 7.6 20.05 18.00
those variations were observed, Sam- calibration factor to convert the area
ples were not taken until the operat- per cent to volume per cent is the * After activation and use.
ing temperature was steady for at least same for both the feed and product t Weight difference between input and output
due to removal of moisture and dust in activation
10 min. within the limits of t 2°F. streams. Knowledge of the fractional of catalvst.
In accordance with thermocouple conversion of the nitric oxide allowed 0 9 This catalyst charge was used in studies at
1 and 8 atm. and the other 18% charge at 15
readings the temperature difference computation of the per cent of oxygen atm.
0.4
0.4
x
z
0
c
VI
>
z
8
0.2
0.2
0
0 5 10 I5 C I I I I
5 10 IS 20
W/F, IG. CAl.)IHR.~/IG. MOLE NO1 X
WIF. ( G . CAT.)lHR>IlG. MOLE N O ) X
Fig. 3. Conversion of nitric oxide vs. w / f a t 8 atm. Fig. 4. Conversion of nitric oxide vs. w/F a t 15 atm.
In the correlation of the data for x that the only rate equation to fit all of oxygen and the experimental observa-
as a function of w / F it was assumed the curves individually was of the form tion that nitrogen was adsorbed on the
that: catalyst, it was necessary to modify
1. Side reactions and the reverse Equation (13) to account for the com-
reaction in the equation 2 N 0 N, + + petitive adsorptivity of the nitrogen.
0, were negligible. This effect was incorporated in the
2. The gas system was a mixture of The coefficients A and B obtained constant A; thus A was taken to be a
perfect gases. from the computer in fitting individual function of temperature and pressure,
3. The adsorption and reaction con- curves as well as all the curves were while B was considered to be a func-
stants were functions only of tempera- varied and scattered. It was then tion only of temperature.
ture. necessary to use trial-and-error calcu- A correlation of A as a function of
4. Diffusion steps were always rapid lations to correlate the coefficients of temperature and pressure gave an
and at equilibrium. Equation (13) as orderly functions of equation of the form
5. Only one of the possible chemi- temperature and pressure. -3,080
cal steps was rate controlling, while ~
I n predicting interfacial area in liquid-liquid systems there are several published reports which
give different results. Differences in methods of measurement of interfacial area and impeller
design and location account for some of these discrepancies.
For one pair of liquids, 40 parts by volume of ethylhexanol and 60 parts water, i t was found
that over wide ranges of impeller size to tank size ratio (0.27 to 0.67) that equal power per unit
volume gave equal interfacial area per unit volume.
A familiar rule-of-thumb for scaling sisted of 40 parts by volume of 2-ethyl- terned after that of Trice and Rodger
u p a mixing operation is to hold con- hexanol and 60 parts of water. The ( 7 ) . Essentially a beam of light is led
stant the energy input per unit volume. alcohol was commercial grade ( interfacial through a glass probe into the suspension.
When mixing involves the creation and tension, 13.9 dynes/cm.; density 0.834 Another probe transfers light transmitted
g./cc.; refractive index, 1.431 at 20°C.). by the suspension to a photocell (Figure
maintenance of interfacial area be-
As in previous studies ( 1 to 5, 7, 8) 1). The signal from the photocell is
tween two immiscible liquids, the rule light transmission was used as a measure amplified, compared with the initial beam,
has been supplanted by dimensionless of particle size. The apparatus is pat- and recorded.
correlations of interfacial area, impeller For measurement of power input a
diameter, and Weber number (2, 3, Chemineer ELB kit was used. The basic
5, 6). In the model system reported kit ( 9 ) consists of a ball-bearing mounted
here energy input per unit volume ap- motor and variable speed reducer, various
pears to be a more valid criterion than impellers and baffles, an 8- and a 16-
the dimensionless grouping. Specifi- gal. tank. The maximum of three baffles,
cally it has been found that the aver- symmetrically located, was used in the
above tanks and also in a l-gal. tank. The
age interfacial area per unit volume is dynamometer is simply a two-pan
proportional to the 0.4 power of energy balance, one pan of which is connected by
input per unit volume for a number of a wire to an arm from the speed reducer
geometric variations. mounting.
EXPERIMENTAL
LIGHT T R A N S M I S S I O N AS A
The system used for the correlation of M E A S U R E OF I N T E R F A C I A L AREA
Fig. 1. The light beam transmitted by the glass
interfacial area with power input con- Three groups of workers have re-
probe is interrupted by a 0.9 cm long path
Ferdinand Rodriguez is at Comell University, through the continuous phase ( I , ) or by the ported the measurement of interfacial
Ithaca, New York; L. C. Grotz is at the Uni-
versity of Wisconsin, Milwaukee, Wisconsin. agitated suspension ( I ) . area of liquid-liquid dispersions ( 1 , 4,