JUL 83 '83 82:58PM ND CHEM. ENG.

Final Report

P.5111

Supercritical Carbon Dioxide-Soluble Ligands for Extracting Actinide Metal Ions from Porous Solids (DOE DE-FG07-98ER14924)

Joan F. Brennecke Department of Chemical Engineering

University of Notre Dame Notre Dame, IN 46556

(574) 631-5847 FAX (574) 631-8366 Email: jfb@nd.edu

July 3,2003

This project was part of a collaborative effort with Argonne National Laboratories (EMSP Project Number 64965) and Loyola University. The investigators from the other institutions are listed below:

Dr. Mark L. Diem (Lead Principal Investigator), Chemistry Division, Argonne National Laboratory, Argonne, ZL 60439. Telephone: (630) 252-3647; e-mail: mdietz@anl.gov.

br. Richard E. Barrans, Jr. (Co-Investigator), Chemistry Division, Argonne National Laboratory, Argonne, 1L. 60439. (no longer at ANL)

Dr. Albert W. Herlinger (Co-Investigator), Depaitment of Chemistry, Loyola University Chicago, Chicago, IL 60626. Telephone: (773) 508-3 127; e-rnail: AherlinQ wpo.it.luc.edu.

Graduate studentlpostdoctorate involvement:

Two graduate students at the University of Notre Dame were involved with this project. One, Dr. Aaron M. Scurto, graduated with his Ph.D. in August, 2002. He is currencly pursuing a post-doctoral position at the hsititut fur Technische Chemie und Makromolekulare Chernie at RWTH-Aachen in Gcrmany. The other Christopher M. Lubbers, who graduated with his M. S. degree in August, 2000. He is currently employed by Goodyear in Akron, Ohio.

Research Objective:

Numerous types of actinide-bearing waste materials are found throughout the DOE complex. Most of these wastes consist of large volumes of non-hazardous marerials contaminated with relarivcly small quantities of actinide elemcnts. Separation of thesc wastes into their inert and radioactive components would dramatically reduce the costs of stabilization and disposal. For example, the DOE is responsible for decontaminating

1

JUL 83 '83 82:58PM ND CHEM. ENG. P. 6/11

concrete within 7000 surplus contaminated buildings. The best technology now available for removing surface contamination from concrete involves removing the surface layer by grit blasting, which produces a large volume of blasting residue containing a small amounc of radioactive material. Disposal of this residue is expensive because of its large volume and fine particulate nature. Considerable cost savings would result from sepwation of the radioactive constituents and stabilization of the concrele dust. Similarly, gas diffusion plants for uranium enrichment contain valuable high-purity nickel in the form OF diffusion barriers. Decontamination is complicated by the extremely fine pores in these barriers, which ixe not readily accessible by most cleming techniques. A cost-effective method for the removal of radioactive contaminants would release this valuablc material for salvage.

The objective of the overall (Argonne National LaboratorylNoue DamelLoyola) project was to develop novel, substituted diphosphonic acid ligands that can be used for supercritical carbon dioxide extraction (SCDE) of actinide ions from solid wastes. Specifically, selccted diphosphonic acids, which are known to form extremely stable complexes with actinides in aqueous and organic solution, were to be rendered carbon dioxide-soluble by the introduction of appropriate alkyl- or silicon-containing substituents. The metal complexation chemistry of these new ligands in SC-C02 were then to be investigated and techniques for their use in actinide extraction from porous solids developed. ANL and Loyola wcrc fully responsible for synthesis, Notre Dame was responsible for measuremen1 and modeling of the share behavior of the ligands and metal linarid comdexes with SC-CO,, and ANL was responsible for actinide extraction from porous solids with the new ligands.

Results:

The University of Notrc Dame goal was to measure the phase behavior of the various siloxane derivatized diphosphonic acid compounds synthesized by the groups at ANL and Loyola, in order to provide feedback on which compounds wouId be best as metal extractants in supercritical CO,. Unfortunately, the ANL and Loyola groups found that those compounds wefe much more difficult to synthesize than originally anticipated. Over the four years of the grant (3 years plus a one year no-cost extension), we received only a very limited number of ligand samples. As indicated below, we investigated the phase behavior of the nine viable liquid samples that we received with CO,. Therefore, we significantly expanded the scope of our part of the project by investigating the phase behavior with C02 of commercially available extractants and mecal chelate complexes.

Our accomplishments 10 datc fall into six categories: 1.) VLE measurements of oligo(dimethy1siloxane) substituted gem-diphosphonates, 2) SFE measurements of two Eu complexes, 3) VLE of commercially-available P-diketones, 4) thermodynamic modeling of the VLE of the P-diketanes, and 5) modeling of the solubility of metal p- di ketonate complexes in multicomponent supercritical mixtures.

2

JUL 83 '03 82:51PW ND CHEW. ENG. P.7/11

1 ) VLE measurcments of olig.o(dimethvisiloxanc) substituted nem-diphosphonates

We have measured the bubble points (i.e., the liquid phase compositions) of nine siloxane substituted gem-diphosphonates with CO;! at 40°C and pressures to about 95 bar. These compounds were oligo(dimethylsi1oxane) substituted gem-diphosphonate esters, Le., not actual diphosphonic acids chat could be used as metal ligands. These were measured with the static high pressure apparatus described in the original proposal. At low pressures, the incorporation of COz into the diphosphonatc-rich phases depends strongly upon the number and nature of the substituent groups present. The parent diphosphonic ester with no siloxane functionality, absorbs less C02 than one with a linear trisiloxane substituent. A compound bearing two such trisiloxane groups absorbs even more CO2. The specific structure of the siloxane substituent is also important. Siloxy- substitution at the phosphonate alkyl esters also enhances the uptake of C02. Siloxane substitution also enhances the CO? affinity of diphosphonates bearing larger alkyl ester groups. The ligand-rich phases of all these compounds consist primarily of CO2 even at 90 bar. Extrapolating the mass friction trends to higher pressure suggests that the mixture critical points of all chese compounds with COz are at pressures of about 100-130 bar at 40 "C. Above the critical pressure, thc COn/diphosphonate mixtures would be single phases aL all compositions. Although siloxane substituents increase the molecular mass and rcduce the volatility of the diphosphonates, they c l e ~ l y enhmce the affinity for co;. 2) SFE measurements of two Eu complexes

We attempted to measure the solubility of two Europium complexes of diphosphonic acid in COZ. These compounds were made wirh conventional diphosphonic acid ligands; Le., ones that have not been derivatized to increase C02 solubility. They were examined as a baseline - to determine how low the solubilities of metal complexes with conventional diphosphonic acids were. The compounds were Eu2(TMSP[MDP])3 and Eu~(EH~[MDP])~. We determined that we were able to detect both of these compounds by fluorescence spectroscopy and developed an analytical method to do that. The solubility measurements we made at 40 OC and 300 bar, using the dynamic extraction apparatus described in the original proposal. We found that there were absolutely no detectable amounts of either compound in the extract. Based on the estimated detection limits, we determincd that the solubilily of both of these compounds must be less than 5 x IO-' mole fxac ti on.

3) VLE of commerciallv-available &diketonet

We measured the bubble points of ten cominercially available P-diketones. They are 2,4-pentanedione (acac), 2,2,6,6-tetramethyl-3,5-heptancdione (thd), 2,2,7-trirnethyl-3,5- octanedione (tod), 1-phenyl-butane- 1,3-dione (Bzac), 1,1,1 -trifluoro-2,4-pentanedione (tfa), 1 , 1,1,5,5,5-hexafluoropentane-2,4-dione (HFA), 4,4,4-trifluoro-1-(2-f~ryl)-butane- 1,3-dione (FTFA), 4,4,4-irifluoro-1-(2-thienyl)-butane-1,3-dione (TTFA), 4,4,4-trifluoro- I-phenyl-buranc-lY3-dione (BTFA), and 6,6,7,7,8,8,8-heptafluoro-2,2-dimethyl-octane- 3,5-dione (fod). Using the static high pressure apparatus, we measured the liquid phase

3

JUL a3 'a3 B ~ : ~ I P M ND CHEM. ENG. P. 8/11

compositions, molar volumes and mixture critical points of these ten compounds at temperatures between 30 and 60°C and pressures to about 95 bar. All of these compounds have high CO;! solubilities and many have mixture critical points below 1,OO bar. An example of some of the data is shown in Figure 1, below. As expected, the C02 solubility decreases with increasing temperature- Mixture molar volumes decrease dramatically with increasing pressure but then level off, as is typical for high pressure liquid/C02 systems. Although higher molecular weight, [he fluorinated compounds do take up more CO;! than their non-fluorinated counterparts. For instance, the COZ solubility in fod is greater than in tfi- Moreover, the COZ solubility in both of these fluorinated compounds is greater than in the non-fluorinated chelating agents.

160

140

120 L : 100

; 8 0

I

m

0 In

P 2 60

4 0

2 0

0

60 C

0 .00 0.20 0.40 0 -60 0.80 1 .oo M o l e Fraction CO,

I I Figure 1 Phase behavior of various P-diketone ligands with CO,.

4) Theimodvnamic modeling of the VLE of the 6-dikctones

In addition, we conducted thermodynamic modeling of the bubble points and liquid rnolcu- volumes of the ten commercially available p -diketone complexes. We used the Peng Robinson equation of state with Van der Waals 1 mixing rules. The critical tcmperitures, pressures and acentric factors of these compounds were not available so they were estimated or exuapolated from available data using stiindard techniques. The only parameters fit to the bubble point and liquid molar volume data are a temperature- independent k, for each compound. The magnitude of these ki, values is quite small, indicating that very little correction is needed to t h e gcometrk mean mixing rule. The Peng-Robinson equation of state does a superb job of modcling the bubble point data. The liquid molar volumes are not as accurate but do show the correct trends.

4

JUL 83 ’83 82:52PM ND CHEM. ENG. P. 9/11

As far as we know, these are rhe first vapor liquid equilibrium measurements of any P-dikerone chelating agenrs with CO2. This informarion is important for determining what conditions are necessary to obtain a single phase COzlcheIating agent mixtures to be fed to the extractor. TL is also important in modeling the solubility OF metal chelate complexes when cxcess chelating agent is present (see below).

5 ) Modeling of the solubility of metal D-diketonate coniplexes in multicomponent suDercritica1 mixtures

The solubility of a variety of metal chelate complexes in supercritical CO2 are available in the literature. However, in a real metal extraction process. one would inevitably have excess ligand present in order to allow high removal of thc metal contaminant. Thus, what is really of interest is the solubility of the metal chelate complex in a COzlexcess ligand mixture. Using published metal chetate/C& solubility data and the chelaring agentKO2 VLE data measured above, we have been able to cstimate the solubility of meul chelates in CO?/excess ligand mixtures using the Peng-Robinson equation of state. For instance, we are able to predict the solubility of Cr(acach in a COz/acac mixture. The model predicts that there should be some significant solubility increase just by having the excess ligand prcsenr.

Summary :

In this project, we measured the phase behavior of C02 with the series of oligo(dimethylsi1oxane) substitutcd gem-diphosphonate esters. They showed high affinity for COz, thus validating the hyporhesis that the addition of silyl groups should significantly increase the solubility of diphosphonates in C02. We have also investigated the phase behavior of a series of ten cormnercidly available P-diketones, showed that their phase behavior with C02 could bc accuratcly modeled with the Peng-Robinson equation of state, and used that model to predicL increases in the solubility of metal chelate complexes in SC COz when excess ligand is present.

University of Notre Dame Publications resulting from this project:

Gang Xu, Aaron M. Scurto, Marcelo Castier, Joan F. Brennecke and Mark A. Stadtherr, “Reliable Compucation of High Pressure Solid-Fluid Equilibrium,” Proceedings of COBEQ 2000, Aquas de Siio Pedro, Brazil, Sept. 24-27,2OOO.

Erik J. Roggeman, Aaron M. Scurto and Joan F. Brennecke, “Spectroscopy, Solubility and Modeling of Cosolvent Effects on Metal Chelate Complexes in Supercritical Carbon Dioxide Solutions,” Ind. En% Chern. Res., 40,2001, p. 980-989.

Aaron M. Scurto, Christopher M, Lubbeis, and Joan F. Brennecke, “Thermodynamic Aspects of the in-situ Extraction OP Metals with Supercritical Fluids,” Proceedings of the 2”‘ International Meeting on High Pressure Chemicai Engineering, Hamburg, Germany, March 7-9,2001,

5

JUL 83 ’83 82:52PM ND CHEM. ENG. P. 18/11

Joan F. Brennecke, Aaron M. Scuno, Christopher M. Lubbers, Lynnertc A. Blanchard. Jennifcr L. Anthony and Edward J. Maginn, “Environmental Applications of Supercritical Fluids,” Proceedings of the 4Ih Brazilian Conference on Supercritical Fluids, Salvador, Bahia, Brazil, October 9-1 1,2001.

Aaron M. Scurto, Christopher M. Lubbers and Joan F. Brennecke, “Supercritical COz for the Extraction of Metals,” Proceedings of the 5Ih International Symposium of the ESIQIE-IPN, Mexico City, Mexico, May 29 - 3 I, 2002.

Aaron M. Scuno, Gang Xu, Joan F. Brennecke and Mark A. Stadthen, “Phase Behavior and Reliable Computation of High Pressure Solid-Fluid Equli b r i m with Cosolvents,” submitted to Tnd. Enp. Chem. Res., 2003.

Chris M. Lubbers, Aaron M. Scurto, and Joan F. Brennecke, “Experimental Measurement and Modeling of the Vapor-Liquid Equilibrium of P-Diketones with COz,” ACS Symposium Series, A. S. Gopalan, ed., in press, 2003.

6