plutonium and uranium hexafluoride hydrolysis kinetics

Downloaded by UNIV OF NEBRASKA-LINCOLN on September 8, 2015 | http://pubs.acs.org Publication Date: January 1, 1967 | doi: 10.1021/i260021a018

Design Considerations

d

A major difficulty to be anticipated in use of loop tuning for sampled-data systems is recognition of the existence of conditional stability. Thus, Figures 4a and 5 show that if the

G,(s)

quasi-steady-state response is aperiodic, it may require considerable effort to determine whether the controller gain is at, above, or below the ultimate gain. This is particularly true if a chromatographic or similar measurement is used, so that only the sampled output, shown on the middle channel, may be observed. Even when conditional stability is recognized, it is difficult to use the information. The effective test input shown by the manipulated variable in the third channel of Figures 4 a and 5 is a sequencie of aperiodic square pulses. Dynamic analysis will require taking the numerical Fourier transforms, instead of the simple sinusoidal result one obtains in the continuous-data system. If such calculations are to be used, it is probably simpler to make a pulse test (7) a t the outset, than to search for the ultimate gain. I n light of these considerations, it is not difficult to see why K / K , (or gain margin) is not a good design criterion, as the duthors observed previously ( 5 ) . I n the case where thc quasi-steady-state response is periodic, as in Figure 4 b , it should not be difficult to find K , experimentally. The period of the quasi-steady-state response is 2T, and thus is not dependent on the process time constant, 7. This is in contrast to the continuous-data case where the period corresponds to the frequency at which the process exhibits 180’ of phase lag. A test a t this crossover frequency is very important for controller design, because of its close relation to system stability. Hence, again the dynamic test is of limited value in the sampled-data case, because the input form is not well suited to gibing dynamic information, and because the observed frequency is more directly related to the sampling period, T , than to the process time constant, T . Nomenclature

delay time of process as a fraction of major time constant = e -T/T = output of system

=

a

b

0)

Gp(5)

H K

= = =

eO-(n+’)T/T

process error ratio of two time constants for second-order process = controller transfer function = process transfer function = transfer function of zero-order hold = loop gain of system =

dz

K7l

=

ultimate gain

m

= manipulated variable

i

T = sampling period T/+ = normalized sampling period ( T / T ) ~=, normalized ~ sampling period for which K , is maximized for T / T 2 a 2

a

6 7

e-transform variable constant constant = major time constant of process = = =

Ac knowledgment

T h e financial assistance of T h e Esso Research and Engineering Co., The Monsanto Co., and T h e Procter and Gamble Co., and the generous donation of computing time by the Purdue University Computer Science Center are gratefully acknowledged.

References

(1) Coughanowr, D. R., Koppel, L. B., “Process Systems Analysis and Control,” McGraw-Hill, New York, 1965. ( 2 ) Hartwigsen, C. C., Mortimer, K., Ruopp, D. E., Zoss, L. M.; “Analysis of Sampling Rates and Controller Settings for Direct Digital Control,” ISA Conference, Preprint 30.1-4-65 (October 1965). (3) Kuo, B. C., “Analysis and Synthesis of Sampled-Data Control Systems,” pp. 164-69, Prentice-Hall, Englewood Cliffs, N. J., 1963. (4) Lindorff, D. P., “Theory of Sampled-Data Control Systems,” ChaD. 4. Wilev. New York. 1965. (5) Mbsler, H. A , Koppel, L. B., Coughanowr, D. R., IND.END. CHEM.PROCESS DESIGN DEVELOP. 5,297 (1966). (6) Ziegler, J. G., Nichols, N. B., Trans. A S M E 6 4 , 759 (1942). RECEIVED for review March 28, 1966 ACCEPTEDSeptember 29, 1966

PLUTONIUM AND URANIUM HEXAFLUORIDE HYDROLYSIS KINETICS ROBERT W. KESSIE Chemical Engineering Division, Argonne National Laboratory, Argonne, Ill.

THE reaction of the highly volatile plutonium hexafluoride (PuFs) with atmo,,pheric moisture provides a basis for secondary containment of the highly toxic plutonium. Plutonium is removed from the air directly on the solid surfaces of reaction or in the case of aerosol formation is removed by high efficiency filters. T h e product of the hydrolysis reaction has been identified as plutclnyl fluoride (2, 4,7) (Pu02F2) and the over-all reaction is PuFs(g)

+ 2HzO(g)

+

+

PuOZF~(S) 4HF(g)

(1)

Uranium hexafluoride (UF6) gives homologous reaction prod-

ucts with similar reaction rates. Since uranium is much less toxic than plutonium, UF6 was used to check out techniques and procedures with much less stringent safety precautions than would be required with plutonium. The objective of this investigation was to measure the hydrolysis rates of PuFc. These data would be useful in evaluating the safety of containment facilities used for PuF8. T h e initial experiments were performed with UF6. I n preliminary experiments of the hydrolysis of UF6 on filters, the following observations were made principally from direct examination of the reaction product in an electron V O L 6 NO. 1

JANUARY 1967

105


The hydrolysis rates of gaseous plutonium hexafluoride and uranium hexafluoride were measured in a packed bed at room temperature. The reaction product i s solid plutonyl fluoride (PuOzF2) or uranyl fluoride (UO~FZ), respectively. At low concentrations of the hexafluorides in the gas phase, the reaction rate i s directly proportional to both the partial pressure of the hexafluoride and moisture and the area of the solid surfaces available for reaction. Adsorbed hexafluoride and HF on the reaction surface reduce the reaction rate at partial pressures greater than 0.001 mm. of Hg. Reaction in the gas phase was detected at moisture partial pressures greater than 1 .O mm. of Hg and results in part of the solid reaction product being produced as fine particles, less than 0.1 micron in diameter, in the gas phase.

microscope. The initial deposit of uranium was a t isolated spots about 100 A. apart. With more extensive reaction the deposit on the fibers appeared as a continuous thin layer, the thickness of which was below the resolution of the instrument (50% relative humidity for 24-hour periods, the deposit transformed to a coarsely textured, highly crystalline material, which was presumably hydrated uranyl fluoride. This material was not formed when a sample of the deposit was exposed to humid air for less than 2 hours. Hydrolysis of Uranium Hexafluoride in Packed Beds

Apparatus a n d Procedure. Surface reaction kinetics of the hydrolysis of dilute mixtures of UF6 were measured in a packed bed of 3.2-mm. glass spheres in the apparatus shown in Figure 1. Glass was used for the packing for several reasons. First, it is the major surface present in high efficiency filters used on existing facilities containing PuF6. Second, it does not react significantly either during the hydrolysis reaction or during the subsequent leaching reaction. Also, the smooth surface provides a reproducible, readily measured area for reaction. The bed was contained in a 1.5-inch i.d. acrylic (Plexiglas) tube and was supported on stainless steel screen separators. The two entering nitrogen gas streams, one containing UF6 and the other containing moisture, were rapidly mixed by a stirrer rotating at 500 r.p.m. Results of scouting tests indicated that the reaction rates were independent of stirrer speeds between 500 and 1500 r.p.m. An extremely dry source of gas was needed for this work, and nitrogen from a liquid nitrogen supply was found satisfactory as the carrier gas, since its dew point was very low (- 100’ F. corresponding to 3 p.p.m. moisture). The U F 6 stream was generated by passing a small flow of dry nitrogen over a bed of ‘/,-inch NaF pellets saturated with UF6. Equilibrium of the gas with the bed was attained, since this system was very similar to that used in the original measurement of the equilibrium pressure of over the “LTF6-3NaFcomplex” as a function of temperature ( 3 ) . The vapor pressure of UF6 was given as: 106

I&EC PROCESS DESIGN A N D DEVELOPMENT

logp (mm. Hg) = 10.88

5090 -T

where T is the absolute temperature, O K. The flow rate of nitrogen through the pellet bed was 83.7 ml. per minute in all runs and the temperature of the bed was adjusted from 134’ to 212’ C. to give the desired pressure of UFB in the feed stream to the hydrolysis apparatus. The UF6-containing stream from the NaF bed was mixed with additional dry nitrogen to give a total feed rate of 3.3 liters per minute. The humidified nitrogen stream was produced by mixing dry nitrogen with moist nitrogen that had been saturated in a water bubbler. By changing the proportion of wet and dry stream flow rates, the moisture content of the mixed stream could be adjusted. The moisture content was determined by passing a small part of this stream to an automatic dew point recorder while the bulk stream was fed to the packed-bed hydrolysis apparatus. The exit gas stream from the packed bed of glass spheres passed through three AEC filter media disks mounted in series. A clean bed of glass beads and clean filters were used for each experiment. All the flow rates and temperatures were adjusted and allowed to stabilize before starting the UF6 flow. The filters were inserted about 5 minutes after the UF6 flow was started and were removed shortly before the LTF6 flow was stopped, to eliminate the errors due to transient conditions. At the end of each run, the separate bed segments and filters were each immersed in 100 ml. of 10% nitric acid containing 0.01% Aerosol (wetting agent obtained from E. H. Sargent and Co.). The solution was analyzed for uranium by the fluorophotometric method. Results a n d Discussion. Hydrolysis rates were calculated from the uranium deposited on the surfaces of glass spheres in each packed-bed segment. The quantities of uranium found in the various packed-bed segments and also the quantity collected on each of the filter disks are reported in Table I. In all experiments, there was a large excess of water over that required for complete reaction. The hydrolysis rates increased with an increase in the partial pressure of either water or UF6. The rates ranged from 0.3 X 10-4 to 0.1 mg. of U/(hr.)(sq. cm.). The highest reaction rate was less than the mass transfer rate (5) calculated from known physical properties, indicating that the reaction is not mass transfer-limited. The ratios of the quantity of uranium collected on the first filter to that retained on the packed-bed segment immediately preceding the filter varied from about 0.1 to 25, the majority being less than 1. The higher ratios indicate the presence of gas-phase reaction (fume formation) and the low filtration efficiency of the packed bed for such a reaction product. Since the hydrolysis reaction produces a solid reaction product (L?O&’*),any surface at which the reaction takes place will be covered with the reaction product. This is true for all solid surfaces, Tvhether they be fixed solid parts of the equipment, or particles being carried in the gas as fume. The reaction rate measured on the surface of packed beds may be expected to be the same as at the surface of a particle suspended

& TO MOTOR

PLEXIGLAS

TUBE-


STAINLESS STEEL SCREEN SEPARATORS

3 - A E C FILTER MEDIA DISKS

SLIDING FILTER Dfi AW ER

--.SATURATED GAS

TO FURTHER REACTION AND CLEAN-UP FILTRATION

Figure 1.

Packed-bed apparatus for hydrolysis for P U R or UFO

in the gas phase under the same conditions of gas composition and pressure a t the .reaction surfaces. Thus reaction rates measured in the packed beds may be useful in predicting the particle growth rate arid size distribution of the fume produced in the gas-phase part of the reaction. T o define more clearly the effects of partial pressures of UF6, water, and HF, the experimental data were correlated using a rate equation proposed by Hougen and Watson (6) and Walas (9) for reaction catalyzed by solid surfaces. The selected rate equation is:

(3)

computer to estimate the constants in the rate equation by a least squares method. In this method the equations were solved and integrated to give the amount of uranium deposited on each bed segment for all runs. The constants were then varied until the total relative error was a minimum. For mass transfer: =

kD[PCFe(g)

- PUFe1

(4)

where

kD = mass transfer coefficient = partial pressure of UFs in the gas phase, mm. of Hg puFe= partial pressure of UFG at the reacting surfaces, mm. of Hg

,bUFe(g)

where r = reaction rate, :mg. U/(hr.) (sq. cm.) k = rate constant P = partial pressure of subscripted component a t the surface, mm. of H g K = adsorption constants

The rate equation and the following equations for describing mass transfer and ma.teria1 balance were programmed for a

Material balance was based on the reaction: UFe(g)

+ 2H20(g)

+

U O ~ F ~ ( -P S )4HF(g)

(5)

When the fraction of UF6 unreacted, x , is defined by the following equation PUFe(g)

= 'P'UFe

(6)

material balance on the other reactants gives VOL. 6

NO. 1 J A N U A R Y 1 9 6 7 107

Table 1.

Surface Reaction Data:for:UraniumlHexatluoride Column diameter. 1 . 5 inches

Linear velocity. Room temperature.

M m . Hg

UM-11 UM-12 UM-13 UM-14 UM-15 UM-16 UM-18 UM-19 UM-20 UM-22 UM-14e UM-23 UM-21

0.00310 0.00310

a

b c


P'H~o, Feed Rate,

d e

9.7 cm./sec. 23-25' C.

Bed,a, M g . U p e r Bed Segmentb M m . HE Mg. U/Hr. M i n . O-l/z in. 1 / ~ - 7 in. 7-2 in. 2-3 in. 0.050 15.9 55.2 1.7c 1 .o 0.9

POUFb,

Run

Hydrolysis Kinetics

0.15 0.27 0.42 0.72 1.3

t

15.9 15.9 15.9 15.9 15.9 15.9 15.9 1.62 7.96 15.9 77.3 158.5

56.0 55.0 58.4 52.5 56.8 61.7 58.7 91.2 90.0 58.4 26.9 25.0

4.2~ 1.6 1.3 0.6 0.3 0.15

1.6 1.3 0.7 0.10 0.04 0.12d

0.9 0.20 0.07 0.02d 0.03d 0.14d 0.lld 0.015d 0.025 0.07 3.5

t Filter," Min.

50.0 50.0 50.0 50.0

3.2 0.00310 4.3 0.00310 4.7 50.0 0.00310 5.2 50.0 0.00310 3.8 4.2 50.0 5.1 2.8 0.10 0.10d 0.00310 50.0 0.20 0.000314 0.43 0.017 0.014d 50.0 0.43 0.00155 4.2 1.0 0.26 70.0 0.42 4.3 1.3 0.7 50.0 0.00310 0.0150 0.43 5.4 3.4 4.9 i5.0 0.0308 0.42 7.5 5.9 8.2 ... 15.0 t bed and t Jiller represent lime during which reactant frow was maintained through bed and filter, respectively. Bed area = 332 sq. cm./in., wall area = 30.4 sq. cm.lin., area of screen between beds = 20.7 sq. cm. Beds in U M - 7 1 and 72 dioided into 7-in. segments only. 0.00310

7st

0.48 0.35 0.063 0.040 0.04i 0.070 1.0

2.8 0.002 0.004 0.040 . _

0.60 3.6

M g . U per Filter 2nd 3rd

0.40 0.099 0.0023

n.nm . ..-

0.38 0,035 0.0017 n nniz o.0011 0,0014 0.007

0.002

0,002

0.003 0.006 0.17 0.07

0.0004

n.nos

_. ..-

0.43 0.99

0.08 0,0008

n mi 0.13 0.38

2

Not used in computation of rate equation constants, since values were afected by fume collection. Run UM- 74 tabulated twice to aid in comparison of data.

PBzO

= poH?O

PHF

- 2poUF,(1

-

= ~ P " u F , (~ x)

r dS

-- -F

dx

(7)

(8) (9)

where

p"

partial pressure of subscripted component at the inlet, mm. of Hg S = surface area of reaction bed F = input rate of UF6 =

The result from the computer calculation is

The average error between the correlation and the measured (standard deviation) for individual bed segrates was =t26YG ments. Data from eight bed segments were not used in the computation because of collection of gas-phase reaction products (fume). This was apparent as a sudden large decrease in slope of a plot of r us. S. Inclusion of a term for moisture in the denominator did not improve the correlation. Tests of the rate equation with the denominator squared also gave a poorer correlation. Hydrolysis of Plutonium Hexafluoride in Packed Beds

Apparatus. The hydrolysis rate for PuFG was measured in an apparatus similar to that used for UF6 experiments. Air, dried to about 0.1 p.p.m. of moisture, was used as a carrier for PuFG. Air with a controlled moisture content between 0.1 and 10 mm. of Hg was mixed with the PuF6 carrier stream for studying the hydrolysis reaction rates. Another air stream saturated with moisture was combined with the mixed stream after passage through the packed bed of glass spheres and AEC filter disks. This final addition of moisture ensured the complete hydrolysis of unreacted hexafluoride and facilitated final cleanup of the gas before discharging to the glove box atmosphere. The equipment for supplying these three streams is shown in the top of Figure 2. The remaining equipment, shown in the lower half of Figure 2 , includes hexafluoride purification and metering systems, and the hydrolysis apparatus. This part of the equipment was enclosed in the glove box for PuF6 experiments. 108

I&EC PROCESS DESIGN AND DEVELOPMENT

Plutonium hexafluoride, in solid form, was purified by vacuum pumping on the material at -80" C. The solid PuF6 was then sublimed and recondensed to release any impurities trapped within the solid. The vapor pressure was measured at temperatures slightly below room temperature and compared with reported values for pure PuF6 (8) as a check of the purity of the hexafluoride. The purification process was repeated until the vapor pressure became constant a t the value reported for the pure PuF6. A weighed quantity of PuF6 was then charged to the ballast tank, which was then pressurized to 1900 mm. of Hg with dry air and the contents were mixed by maintaining a temperature gradient across the ballast tank, the bottom of the tank being 5' C. hotter than the top; this gradient was maintained for a t least 16 hours. During a run the PuFo-air mixture was metered with a rotameter at a constant rate between 50 and 100 ml. per minute. This mixture was then further diluted with dry air to produce the feed to the hydrolysis equipment. After each experimental run, the plutonium deposited on the bed segments, filters, mixer, etc., was dissolved by submerging the separated units in a measured volume of 10% nitric acid plus 0.01% wetting agent (Aerosol), for 24 hours, The solution was then mixed and sampled for determination of plutonium and/or uranium. Complete dissolution of plutonium was verified in several instances by counting of dried media. Results and Discussion. The PuF6 experimental results are reported in Table 11. The partial pressures in the initial mixed gas ranged from 0.06 to 5.6 mm. of Hg of water vapor and 0.00001 to 0.02 mm. of Hg of PuF6. The data for PuF6 were fitted by the same rate equation used for the UF6 hydrolysis data. T h e best constants in this form of the rate equation are:

r =

53.5 P P ~ F ~ P H ~ o 1 f 1 4 2 P ~ u ~ a 1.9PHF

+

(11)

The average error between the rates from the correlation and the measured rates was 81 % (standard deviation) for individual bed segments. This form of the rate equation does not correlate the PuF6 data as well as it did the UF6 data; but, considering the wide range of the partial pressures used in the experimental program, the equation is regarded as a reasonable empirical correlation. From the data reported here it appears that aerosol formation increases with increase in moisture concentration primarily

J

AIR

SUPPLY"

HUMIDIFIERS

" DRY

AIR

CHECK VALVE


1

n

DIFFUSION

PUMP REACTION AND

Tb

VACUUM PUMP

Figure 2.

Schematic drawing of packed-bed equipment for hydrolysis of PuFs or uF6

and is dependent to a lesser extent on concentration of hF6. A separate study (7) ai; high concentrations of PuF6 resulted in only aerosol being detected. These particulates had a geometric mean diameter of 0.017 micron with 99% of the particles less than 0.1 1 micron, determined by measurements with an electron microscope. The ratio of PUF6 to UF6 reaction rates a t similar concentrations of reactants, as indicated by the correlations, ranges from 0.28 to 2.4. The extent of departure of this ratio from unity is not highly significant when compared with the average error of the PuF6 results from the correlation. Hydrolysis of Mixed I'lutonium and Uranium Hexafluorides in Packed Beds

Three experiments were made with an approximately equimolar mixture of PuFs and UFOto determine more accurately the relative reaction mtes of PuFs and UF6 in the apparatus previously described. The hydrolysis rates of h F 6 were expected to be similar in mixed and unmixed experiments; a similar effect was also expected with UFs. Instead, an entirely new effect is indicated by the data in Table I:[. The mixed PuF6 hydrolysis rates are higher than the rates obtained in experiments in which PuFe alone was hydrolyzed in air with low moisture concentrations. The UF6 hydrolysis rates in the mixture are lower than the unmixed UFe hydrolysis rates. Further, in the mixed experiments PuF6 hydrolysis rates were uniformly higher than the

UF6 hydrolysis rates. These observations are substantiated by the following approach. The hydrolysis rate may be represented by the equation: = f PXFg

(12)

where f is a function of the reactant and product partial pressures and phtF6is the partial pressure of the hexafluoride. For the data reported in Tables I and 11, in any single experimental run, f changes more slowly than r or fiXF6. Assuming that f is not widely variant, an average value off between any two sections of the packed bed can be computed by integrating the rate equation. The fraction of h F 6 or UFO(MFF,) unreacted, x , is defined by: pXF6 = X p " M F 0

(13)

where poMF6is the initial value of occurs. Material balance gives:

pars before any reaction (14)

r d S = -Fdx

where S is surface area in the packed bed, sq. cm., and F is the feed rate of plutonium, mg. of Pu per hour. Solving the above equations for f and integrating gives: Fln

VOL. 6

XI -

NO. 1

JANUARY 1967

109

Table II. Surface Reaction Data for Plutonium-Uranium Hexafluorides Hydrolysis Kinetics

Column diameter. 1 .5 inches Linear velocity. 9 . 8 cm./sec. Temperature. 22-25' C. Run time. 60.0 min.

Run

M g . M per Bed Segment0 0-1/2 in. 1/2-7 in. 1-2 in. 2-3 in.

M g . M per Filter 1st 2nd

Plutonium Hexafluoride Experiments 1.067 0.0619 0.0863b 0.0604b 0.231 Pu 0.00098 5.06 0.506 0.139 0.0438 0.0515 2.31 Pu 0.00150 7.84 0.878 1.49 1.037 1.98 1.30 Pu 0.00217 11.29 PU 0.0221 114.4 16.6 12.2 22.0 14.9 13.8 0.316 0.0301 0.0317 0.0198 0.0531 Pu 0.000303 1.56 0.0209 0.0219 0.0133 0.0186 Pu 0.000181 0.938 0.192 Pu 0.0000115 0.059 0.00313 0.00102 0.0012 0.00212 0.00101 Mixed Hexafluoride Experiments 0.0550 0.0406 0.0339 0.229 0.875 Pu 0.00078 3.97 PM-11 5 . 3 0.945 0.086 0.124 0.114 Ir o.00108 5.52 0.694 . . . ~ . ~ 1.24 0.1215 0.184 0.0350 0.1004 4.63 PM-10 0.43 pu 0.00090 0.056 1.37 0.512 0.94 0.769 U 0.00113 5.80 0.699 0.1022 0.0368 0.0151 0.0362 2.51 PM-12 0.052 Pu 0.00049 0.499 0.736 0.748 0.481 0.358 U 0.00094 4.82

PM-4 PM-1 PM-6 PM-5 PM-8 PM-9 PM-7


Feed.

P'HnO, Mm. Hg

5

b

5.6 0.54 0.060 0.49 0.45 0.46 0.45

~~

3rd

Mg. M Mixer

Mg. M Cylinder and Screens

0.483 3.07 1.068 3.72 1.37 3.23 20.7 13.8 0.963 0.148 0.0933 0.000135 0.000246 0.578 0.000073 0.000073 0.0471 0.00317

0.00123 0.00152 0.00088 0.394

0.00215 0.00237 0.00059 0.0364

0.0001

0.000098 0.00022 0.00020 0.00022 0.00019 0.00019 0.000024 0.000110 0.355 0.325 0.0001

2.35 3.14 2.42 1.49 1.32 0.918

0.387 0.412 0.529 0.665 0.300 0.404

Bed area = 332 sq. cm./in., wall area = 30.4 sq. c m . / i n . , area of screen between beds = 20.7 sq. cm. Not used in computation of rate equation constants, since values were affected by fume collection.

E E

-1-

20

c I

i-

I2.0

t

0.02

0.05

0.1

0.2

0.5

I.o

2 .o

5.0

IO

Po H20,m m Hg Figure 3.

Pseudo-first-order rate constant for PuFe or UFdMFe) hydrolysis

Using Equations 12 and 13 gives

xz

rzfl

The plutonium or uranium concentrations deposited on the packed-bed surfaces give the average value of r in each bed section. S2 S1is the total surface area between the bed mid-

-

110

points where rl and r2 are determined. A first approximation to.7, calledJ', is obtained by setting f l = f 2 . I t can be proved that

I & E C PROCESS DESIGN AND DEVELOPMENT

As an example, i f f decreases by 10% and r decreases by a factor of 100,

E-

EXFERIMENTAL -EXTRAPOLATED RANGE RANGE

asymptotic values at low pressures are indicated as short dashed lines on the ordinate. Plutonium removal at high efficiency by the surface reaction is approximated best by selecting f for x = 0, since over most of the reaction area x is very near zero; this approximation underestimates the final concentration. Conclusions

pa MFs- m rn H g


Figure 4. Pseudo-first-order rate constant from rate equations

Thus the difference between 1and $‘ is only 2% in an interval when f is decreasing by 10%. T h e first approximation to computed from the data of Tables I and I1 is plotted as a function of PoHsoin Figure 3. For values o f j less than 40 orpoH20greater than 1.0 mm. of Hg, the first and second bed segments were used because the rapid decrease in concentration present in the first two beds was masked in the last bed by even small collection of product from the gas-phase reaction. T h e curves in the figure illustrate the statements, made earlier, about the relative hydrolysis rates. T h e dependence o f f on the pressure of hexafluoride becomes small at pressiires less than 0.001 mm. of Hg. The dependence at higher values is indicated in Figure 4. The values were computed from Equations 10, 11, and 12. T h e moisture values are conservative lower limits and were selected from the frequency distribution of atmospheric moisture a t Argonne to have probabilities of 0.001 and 5% of having lower values than 0.25 and 1.8 mm. of Hg, respectively. T h e

T h e measured reaction rates of plutonium and uranium hexafluoride provide a basis for understanding and estimating the performance of air cleanup systems used on process facilities containing PuFe or UFe. The remaining limitations to predicting performance of practical systems are probably restricted by the uncertainty of high efficiency air filtration performance with small particles in the size range of 0.02- to 0.2micron diameter. Acknowledgment

The author acknowledges the efforts of L. J. Marek and B. J. Misek in the construction of equipment and the collection of data, computer programming by A. J. Strecok, and the interest shown by W. J. Mecham, D. Ramaswami, J. Fischer, and A. A. Jonke. literature Cited

(1 ). Argonne National Laboratory, Chemical Engineering Division, Summary Report, April, May, June 1963, U. S. At. Energy Comm., ANL-6725,164 (1964). (2) Argonne National Laboratorv. Chemical Engineerine Division “Semiannual Report, Juli-December, 19u63, U. “S. At. Energy Comm., ANL-6800,247 (1964). (3) Cathers, G. I., Bennett, M. R., Jolley, R. L., Znd. Eng. Chem. 50,1709 (1958). (4) Florin, A. E., et al., J . Znorg. Nucl. Chem. 2, 379 (1956). (51 Guuta. A. S.. Thodos, G.. A.Z.CI2.E. J . 8.608 (1962). (6) HoLgen, A. O., Watson, K. M., “Chemical Process Principles,’’p. 922, IViley, New York, 1947. (7) Mandleburg, C. J.. et a/., J . Inorz. ’Vucl. Chem. 2, 365 (1965). ( 8 ) Steindler, M. J., “Properties of Plutonium Hexafluoride,” ‘c. S. At. Energy Comm., ANL-6753 (1963). ( 9 ) FYalas, S. M., “Reaction Kinetics for Chemical Engineers,” p. 160, McGraw-Hill, New York, 1959. 1

,

RECEIVED for review March 31, 1966 ACCEPTEDSeptember 22, 1966 Work performed under the auspices of the U. S. Atomic Energy Commission.

DIALYSIS FOR SEPARATING SOLUTES OF DIFFERENT MOLECULAR WEIGHTS Process Optimization C . C . OLDENBURG

Stauffer Chemical Go., Richmond, Calif.

IALYSIS

has been u!jed as a physical separation method on a

Dcommercial scale fix over 30 years to recover recycle caustic in rayon manufacture. Liquor fed to dialysis equipment is an aqueous solution containing about 17% sodium hydroxide and 27, hemicellulose (4, 8). About 90% of the sodium hydroxide passes through the membranes to result in a diffusate containing 9% sodiurn hydroxide. The hemicellulose mole1 Present address, Chevron Research Co., Richmond, Calif.

cules are so large that they do not pass through the membranes. The dialyzate contains all the hemicellulose and about 10% of the I i a O H present in the liquor. This particular application of dialysis recovers a valuable low molecular weight inorganic (NaOH) from a less valuable, larger molecular weight organic material (hemicellulose). The diffusate contains the valuable product, and the dialyzate is the waste stream. Other similar applications include sulfuric acid VOL. 6

NO. 1

JANUARY 1 9 6 7

111

plutonium and uranium hexafluoride hydrolysis kinetics

Recommend Documents