I BACKGROUND OF PLUTONIUM PARTS A supercritical fluid is the comprcssed, dense gas Stephanie J. Hale above the critical temperature. Liquefaction of a gas can EG&G Rocky Flats, Inc occur upon compression below the critical temperature, but above the critical temperature the gas cannot liquefy regardleas of the applied pressure and a single gas phase is ABSTRACT maintained. The critical point for carbon dioxide is 31'C and 74 bar (1088 psi) which means that the supercritical Supercritical fluid carbon dioxide is under investigation in fluid can be aUnined under relatively mild conditions. this work for use as a cleaning for the final cleaning of plutonium parts. These parts must be free of The phase diagram of CO, (Figure 1) shows the organic residue to avoid cormsion in the stockpile. Initial supcrcritical region above 74 bar and the 32" isotherm. It studies on stainless steel and full-scale mock-up parts can be seen that very liquid-like densities can be achieved indicate that the oils of interest are easily and adequately and still remain in the gas phase. cleaned from the metal surfaces with supercritical fluid carbon dioxide. Results from compatibility studies show that undesirable oxidation or other surface mctions are not occurring during cxposureof plutonium to the supercritical fluid. Cleaning studies indicate that the oils of interest are Temperalure ' removed from the plutonium surface under relatively mild conditions. These studies indicate that supercritical fluid carbon dioxide is a very promising cleaning medium for this application.

INTRODUCTION

Our aim in this work is to develop a cleaning process that van u-o clean plutonium parts. 'Ihe plutonium parts iiv are repeatedly exposed to various organic substances during the fabrication and assembly of weapons components. The plutonium is machined and requires organic based coolants, GasLiquid lubricants, and oils. These organic residues must be Equilibrium removed from the parts to avoid corrosion in the stockpile. In light of the President's recent initiative toward limiting production of new weapons and, consequently, the long- .25 SODensity .75 (@I) 1.0 1.2: term storage of these, it has become essential that these pa- are free from any organic residue. :L Typically, these organic residues are removed in vapor Fig. 1: Phase diagram of CO, degreasers using halogenated hydrocarbons me l,l,l- trichloroethane and carbon tetrachloride. Large amounts of The constant density curves for CO, (Figure 2) as a mixed waste are generated and a significant quantity of function of temperature and pressure illustrate that small rhcse volatile organic compounds are released to the changes in pressure can result in significant changes in the atmosphere. With the current environmental issues and density of the CO,. And, since the solubility of a regulatory requirements, it has become prudent to avoid the substance in CO, is dependent upon density, one can he- Use of halogenated hydrocarbons. Additionally, it is tune the system using pressure changes to obtain the Suspected that thesc leave a residue on the cleaned required solubilities. parts that can lead to undesirable corrosion reactions. What makes SCF CO, such a promising cleaning medium An alternative cleaning medium is needed that is is that these liquid-like densities can be achieved resulting environmentally acccptable, nonhazardous, nontoxic, in very liquid-like solvent properties. It is a very good noncombustible, riadily recyclable, low cost. not regulated, solvent for non-polar to slightly polar organic substances. compatible, and effective. Supercritical Fluid Carbon There is the added benefit of its gas-like characteristics Dioxide (SCF COJ is, in fact, a cleaning solvent that which means improved mass transport characte&ics such meets thcse criteria. Work on this project is focussed on as improved diffusivities and lower than Liquids. the evaluation of the compatibility and effectiveness of the ,: SCF CO, cleaning process. .. ..*

EXPERIMENTAL RESULTS

Preliminary tests

The preliminary work was accomplished by Motyl‘ in the early 1980’s. Stainless steel coupons were contaminated with Texaco Regal R&O 32 Oil (a commonly used machining oil) and cleaned in SCF CO, under conditions of pressure ranging from 1260 to 4000 psig, temperature ranging from 33 to 5OoC, densities ranging from 0.65 to 0.92 g/cc, cleaning times of 10 to 30 minutes; and at OI I different flowrates. Auger elcctron spectroscopy (AES) 0 10 20 30 40 50 showed a measurable but acceptably small residue. It was TEMPERANRE PCI found that this oil removal showed little pressure, Fig. 2: Constant density curves for CO, temperature, density, or flowrate dependency. Further studies using steel wool to simulate plutonium machining The benefit of no surface tension is especially useful when turnings proved that the cleaning process was adequate: cleaning complex geometries. Since it acts like a gas, it although, a strong density and flowrate dependency was expands to ii the entire volume of a chamber, thus seen for the removal of the same oil. exposing all of the surface area of the complexgwmetry to the cleaning medium. The studies continued using uranium. The uranium coupons were cleaned at 1029 psig and 34°C. The oil was A SCF CO, cleaning process is especially attractive with removed easily and there were no observable changes in respect to waste minimization. Typically, the organic appearanceafter the cleaning process. AES suggested the residue is removed with a hazardous solvent. Although possibility of CO, chemisorbed to the UOx surface. most of the solvent escapes to the atmosphere, a significant quantity contributes to the overall volume of waste, A preliminary compatibility test was done on plutonium in forming a mixed waste. A mixed waste is onc containinr a static test at 1500 psig and 50°C. AES of the plutonium disposition of this waste is difficult. The proposed plan to in the oxide layer before and after cleaning which suggests use a nonhazardous solvent to remove the residue compatibility of plutonium in supercritical fluid carbon circumvents the formation of a mixed waste. The dioxide. supercritical fluid can be expanded to the gas phase and recycled leaving only the organic residue and avoids the Subsequent tests addition of thc solvent to the waste volume. In recent subsequent studies’ full-scale mock-up partr were The schematic given in Figure 3 depicts a simplistic used to evaluate the cleaning of an appropriate shape. A version of the proposed system. The CO, is pumped, rinse analysis was used to evaluate the cleaning efficacy of compressed and sent through the cleaning chamber which the process using hexane and gas chromatography. A resides in a glovebox. The glovebox is necessary, of cleaning criterion has been calculated and established at course, when workingwith plutonium. It then goes through about 5 - 10 pg/cmz. These stainless steel hemishells were an expansion process where the oil is dropped out and the contaminated with the R&O oil and cleaned with SCF CO, gas is recycled. This may be more energy intensive than varying the operating conditions of pressure. temperature. using a low boiling solvent in a vapor degreaser. but the density, flow rate, and process times. In all cues the cost difference should be negligible outside the capital residue levels were far below the lowest requued Limit of expense of the initial system. 5 pglcm’. In fact, this oil was removed so easily, it was

GI- 80, not passible to optimize on any process parameters.

Nye Watch Oil is an oil commonly used in the gauging and contouring process for the finished parts. It was anticipated that this oil might be slightly more difficult to remove owing to a polar ester component in the formulation of thc oil. Removing this oil from the mock- up part did require densities above 0.71 gicc (1400 psig t and 35‘ C) as shown in Figure 4. This information provided the lower limits for operating parameters since it IJ is reasonable to assume that this oil is a typical - YIM..WClmCI-. - contaminating oil. Fig. 3: Schematic of SCF cleaning system Cleaning studies were then undertaken to evaluate the process for removal of Nye Watch Oil. At this time an analytical method has not becn approved for the plutonium am within which we are working. Therefore, the evaluation of oil removal was done by weight differential. While this method is adequate to determine cleanliness to the mg level, it is not sufficient to meet. the pg requirements. Fourier Transform Infrared (FTIR) techniques are being dcvelaped at LANL for use as an analytical rinse method as well as a surface analysis method for determining cleanliness levels for plutonium and should be implemcnted in the near future. At present, however, hndty 01 cubon dloxlde mg level detection provides valuable cleaning information.

Fig. 4: Depxndence of oil removal on density of CO, Figure S represents the results of the fust plutonium cleaning tests. The tests were run at CO, densities ranging A test was performed in which a hemishell was fitted with from 0.7 to 0.9 glcc, pressures ranging from 1653 to 4069 small coupons on both the convex and concave surfaces. psig. and the temperature was held at 40'C. In all but one The part with the coupons was contaminated with Nye case the oil put onto the coupon was completely removed Watch Oil and cleaned in SCF CO,. The coupons were by the SCF CO, process, within the limits of detection. then removed and evaluated for residue by X-ray Additionally, there has been no observable deleterious Photoelectron Spcct~~scopy(XPS). Although all the effects to the surfaces of these coupons. There was somc coupons were mom than adequately cleaned, it was indication in these tests that the plutonium coupons with discovered that the cnupons residing at the pole of the observable oxide layers before contamination with oil and hemishell had a substantially higher carbon layer than all cleaning were consistently less clean afier the SCF CO, the other coupons. This indicates that in some process than freshly polished plutonium surfaces. Since, in configurations the potential exists for nonuniform cleaning actuality, it is oxided surfaces that will be cleaned, this and this muues further investieation.- observation merits further investigation. However, it is also noted that even the oxlded surfaces yielded virtually

' ' % Plutonium studies completely cleaned surfaces after SCF CO, cleaning. ;3. . ..

National Laboratory (LANL) and plutonium coupon tests 0.87 were initiated. 'i~ The compatibility of plutonium in supercritical fluid carbon 0.41- 1 dioxide remained questionable since the thermodynamics of 0.3 t the oxidation of plutonium in carbon dioxide indicate a i Iavorable reaction. It is anticipated that the kinetics of the ilx 1. [ IJ I process are such that the reactien would not occur under 0.1 I- 1 normal operating conditions; however, the kinetics at I. "}I supercritical conditions have not becn evaluated. Although 0 i- f' rate data are not available for oxidation of plutonium by -0.1 ' ' '- ' ' ' I ' ' ' ' ' ' ' ' ' ' ' ' ' ' '

.. CONCLUSIONS

The successful removal of contaminating oils from full- sized mock-up (stainless steel) parts and plutonium coupons in SCF CO, indicates a promising alternative cleaning technology for the fmal cleaning of machined plutonium parts. Observations such as the slightly less effective cleaning of oxide surfaces and the potential for nonuniform cleaning in cemin configurations provide areas for further evaluation. However, all results thus far indicate that the SCF C02 cleaning process is an excellent method for this purpose.

This technology is particularly attractive for this application for many reasons including waste minimization and hazardous solvent elimination. The cleaning of plutonium parts is a unique application. Most cleaning applications do not involve cleaning plutonium or other reactive metals. However, this cleaning technology can be applied to cleaning complex geometries or in processes where conventional methods such as aqueous cleaning are not feasible.

REFERENCES

[l] K. M.hlotyl, USDOE Report, RFP-4150, Rockwell International, Rocky Flab Plant, Golden, CO, 1988.

[Z] S. I. Hale, D. R. Horrell, G. S. Fenner, I. M. Hnschke, I. P. Baiardo, RFP Report PPC-91-017, EG&G Rocky Flats, Inc., Golden CO, 1991.

[3] I. T. Wabcr, "Corrosion and Oxidation" in Plutonium Handbook, 0. J. Wick, ed., Vol. 2, Chapter 6, American Nuclear Society, LaGrange Park, IL, 1980.

[4] I. M.Harshke and S. J. Hale, LANL Formal Repon, in publication. Los Alamos National Laboratory, Los Alamos, NM. 1991.

a

a!

u a I 4 u' k! E I Y X 0" . 0

sa U

u1 cna zn

c c I

_. -1 CONSTANT-DENSITY CURV4S FOR C02 AS FUNCTION OF PRESSURE ANQ TEMPERATURE

350 -

300 -

h 250 - mL .Q v W 150 - E 3 v) v) W 100 - IT n 50 -

50

EGCG ROCKY FLATS - . n s" c-l Frr c-l u4 i= u2 wP= PC 5 in I SUPERCRITICAL FLUID CARBON DIOXIDE CLEANING i OF PLUTONfUM PARTS I WASTE MINI$IZATION

1 II

' dispose + - -. -+ Organic residue

I Mixed Waste I

~

Organic residue Non-hazardous Solvent 1 dispose,

Organic residue waste Heated Cleaning Chamber

Heated ) Expansion Valve

1 J c I - Mechanical or Thermal Compression - 0 a

L w 0 Q) 5 00 dc, N 0 u

4

I

.f

I. r? 1S

zz00 II ti li LE+* /" I -. REMOVAL OF OIL F ?OM Pu WITH SCF C02 GO2 Densii ?s 0.7 to 0.9 g/cc Contamina t: Nye Watch Oil

NORMALIZED Mg. OIL (6.18 = 1) 0.6 I

0.5

0.4

0.3

0.2 x xx* 0.1

0

-0.1 112233445566 7 7 8 8 9 9 10 10 11 11 12 12 RUN #' (OIL AP LIED: OIL REMOVED)

Run Time-30 in. Flow=O.O2 Ipm Temp.=40"C Pr, ;sure-1653-4069 psig