amarillo national resource center for plutonium · 2012. 11. 18. · chemical company, inc. the...
TRANSCRIPT
ANRCP-1998-14September 1998
Amarillo NationalResource Center for PlutoniumA Higher Education Consortium of The Texas A&M University System,Texas Tech University, and The University of Texas System
Collector for Recovering Galliumfrom Weapons Plutonium
C. V. Philip, R. G. Anthony, and S. ChokkaramDepartment of Chemical Engineering
Texas A&M University
Edited by
Angela L. WoodsTechnical Editor
600 South Tyler • Suite 800 • Amarillo, TX 79101(806) 376-5533 • Fax: (806) 376-5561
http://www.pu.org
This report wasprepared with thesupport of the U.S.Department of Energy(DOE) CooperativeAgreement No. DE-FC04-95AL85832.However, any opinions,findings, conclusions,or recommendationsexpressed herein arethose of the author(s)and do not necessarilyreflect the views ofDOE. This work wasconducted through theAmarillo NationalResource Center forPlutonium.
ANRCP-1998-14
AMARILLO NATIONAL RESOURCE CENTER FOR PLUTONIUM/A HIGHER EDUCATION CONSORTIUM
A Report on
Collector for Recovering Gallium from Weapons Plutonium
C. V. Philip, R. G. Anthony, and S. ChokkaramDepartment of Chemical Engineering
Texas A&M UniversityCollege Station, TX 77843
Submitted for publication to
Amarillo National Resource Center for Plutonium
September 1998
ii
Collector for Recovering Gallium from Weapons Plutonium
C. V. Philip, R. G. Anthony, and S. ChokkaramDepartment of Chemical Engineering
Texas A&M University, College Station, TX 77843
Abstract
Currently, the separation of galliumfrom weapons plutonium involves the use ofaqueous processing using either solventextraction or ion exchange. However, thisprocess generates significant quantities ofliquid radioactive wastes.
A Thermally Induced GalliumRemoval process, or TIGR, developed byresearchers at Los Alamos NationalLaboratories, is a simpler alternative toaqueous processing. This research examinedthis process, and the behavior of galliumsuboxide, a vapor that is swept away bypassing hydrogen/argon over galliumtrioxide/plutonium oxide heated at 1100°Cduring the TIGR process. Through
experimental procedures, efforts were made toprevent the deposition of corrosive galliumonto furnace and vent surfaces.
Experimental procedures includedthree options for gallium removal andcollection: (1) collection of gallium suboxidethrough use of a “cold finger”; (2) collectionby in situ air oxidation; and (3) collection ofgallium on copper.
Results conclude all three collectionmechanisms are feasible. In addition, galliumtrioxide exists in three crystalline forms, andeach form was encountered during eachexperiment, and that each form will have adifferent reactivity.
iii
TABLE OF CONTENTS
1. INTRODUCTION.......................................................................................................................1
2. EXPERIMENTAL PROCEDURE .............................................................................................3
2.1 THREE OPTIONS FOR GALLIUM REMOVAL AND COLLECTIONS............................4 2.1.1 Collection of Gallium Suboxide by “Cold Finger”...........................................................4 2.1.2 Collection by In Situ Air Oxidation of Gallium Suboxide Vapors...................................5 2.1.3 Collection of Gallium on Copper......................................................................................6
2.2 REACTIONS OF METALS AND ALLOYS IN GALLIUM SUBOXIDE VAPOR.............7 2.2.1 Effect of Reaction Temperature on Gallium Removal .....................................................7 2.2.2 Effect of Copper to Ga2O3 Bed Distance on Gallium Collection .....................................7 2.2.3 BET Surface Area Measurements.....................................................................................8
2.3 VOLATILIZATION OF GALLIUM SUBOXIDE .................................................................8
3. CONCLUSIONS.........................................................................................................................9
iv
LIST OF TABLES
Table 1: XRD Peak Position, d-Spacing Values of White and Gray Products ................................15
Table 2: XRD Peak Positions, d-Spacings, and Relative Intensities of Gallium Oxides.................18
Table 3: XRD Peak Positions and d-Spacing Values for the Deposited Products onStainless Steel and Zirconium Wires ..................................................................................23
Table 4: Summary Table for the Gallium Experiments Conducted with Different Metals .............28
Table 5: Effect of Pretreatment on the BET Surface Area of Gallium Trioxide..............................29
Table 6: Effect of H2 Flow Rate on the % Conversion of TIGR Process Product ..........................30
Table 7: Experimental Results for Thermally Induced Gallium Removal (TIGR) ProcessUsing “Cold Finger” ...........................................................................................................31
v
LIST OF FIGURES
Figure 1: Schematic Diagram of the Experimental Setup Used for Gallium Removal ...............11
Figure 2: Experimental Setup for Gallium Removal ...................................................................12
Figure 3: Experimental Setup for Gallium Removal by “Cold Finger” Deposition of Product ..13
Figure 4a: Mid-IR Spectra of Deposited White Product and Gallium Trioxide..........................14
Figure 4b: Far-IR Spectra of Deposited White Product and Gallium Trioxide...........................14
Figure 5: X-Ray Diffraction Pattern of White Product and Gray Product ...................................15
Figure 6a: Mid-IR Spectra of Deposited Gray Product ...............................................................16
Figure 6b: Far-IR Spectra of Deposited Gray Product.................................................................16
Figure 7a: Mid-IR Spectra of Gallium Trioxide after the Reaction.............................................17
Figure 7b: Far-IR Spectra of Gallium Trioxide after the Reaction..............................................17
Figure 8: X-Ray Diffraction Pattern of Gallium Trioxide Before Reaction and GalliumTrioxide after Reaction from Run N-23 (bottom) .........................................................18
Figure 9: Experimental Setup for Gallium Removal by In Situ Oxidation of Ga2O ...................19
Figure 10: Far-IR Spectra of Re-Oxidized Product and White Product ......................................20
Figure 11: Experimental Setup for Gallium Removal using Copper Collector ...........................21
Figure 12: Experimental Setup for Gallium Removal using Stainless Steel Collector ...............22
Figure 13: X-Ray Diffraction Pattern of the Scratch from Stainless Steel Wire from Run N-4and Crumbled Zirconium Wire from Run 42 ..............................................................23
Figure 14: % Ga2O3 Conversion as a Function of Reaction Temperature in H2/Ar Flow ...........24
Figure 15: % Ga2O3 Conversion as a Function of Reaction Temperature in H2 Flow ................24
Figure 16: % Ga2O3 Conversion as a Function of Copper Distance............................................25
Figure 17: % Ga2O3 Conversion as a Function of Copper Distance............................................25
vi
Figure 18: Effect of Hydrogen Pretreatment on BET Surface Area of Gallium Trioxide ...........26
Figure 19: % Conversion of TIGR Process Product as a Function of Reaction Time.................26
Figure 20: % Conversion of TIGR Process Product as a Function of Flow Rate .........................27
1
1. INTRODUCTIONCurrently, separation of gallium from
weapons plutonium involves the use ofaqueous processing using either solventextraction or ion exchange, but this processgenerates significant quantities of liquidradioactive waste. At Los Alamos NationalLaboratory, researchers have been developinga simpler alternative referred to as theThermally Induced Gallium Removal (TIGR)process. In this process, vaporized galliumsuboxide is swept away by passinghydrogen/argon over galliumtrioxide/plutonium oxide heated at 1100°C orhigher. During the TIGR process some of thegallium suboxide prematurely decomposes togallium metal and gallium trioxide anddeposits on furnace and vent surfaces. Thedeposition of this corrosive gallium must beprevented. The efforts at Texas A&MUniversity are directed towards learning moreabout the behavior of gallium suboxide anddetermining how the reduction of vaporizedgallium suboxide on a copper surface can beused to develop an effective “galliumcollector,” which could be incorporated intothe TIGR process. It was discovered thatcopper catalyzes the reduction of vaporizedgallium suboxide to gallium and the dropletsof gallium collect on the copper forming aGa/Cu alloy2. When a copper wire wasplaced in a stream of gallium suboxidevapors, gallium metal was collected on the tipof the wire forming droplets that dropped offas they gained weight. Gallium collectedmostly on the tip of the wire, suggesting thatby placing a copper collector above the oxidebed, all the gallium could be collected as analloy with copper. Deposition of galliumsuboxide on a “cold finger,” and in situ vaporphase oxidation of gallium suboxide tononvolatile gallium trioxide are the other twooptions under consideration. Experimentswere conducted to test and evaluate theseoptions.
The copper-collector is believed to bethe best choice because of its clean galliumcollection without the formation of anysuboxide or trioxide dust, which could depositon parts of the furnace and vent. Ga/Cu alloyis chemically stable as well as easier to handlethan the oxides. Since the reduction ofgallium suboxide takes place on the tip of thecopper wire, the reaction is very fast and doesnot need a large copper surface area. A largecopper rod could be used as the collector andthe drippings from the rod will have high Gacontent.
Incorporation of a copper-collector inthe TIGR process is an attractive option.Several experiments were conducted toidentify the benefits of a copper-collector aswell as the reactions of other materials in gasstreams with vapors of gallium suboxide.Elemental analysis by atomic absorption (AA)spectrometry, FT-IR (both mid and far IR)spectrometric analysis, BET surface area andpore size distribution measurements, andpowder X-ray diffraction scanning were usedfor the characterization of materials.
3
2. EXPERIMENTAL PROCEDURECerium Oxide powder (< 5 micron),
Gallium trioxide and copper wire (AAstandard grade) were obtained from AldrichChemical Company, Inc. The experimentalsetup is shown in Figure 1. Two LindbergBlue/M furnaces were used so that twoexperiments could be run simultaneously.The Lindberg Blue/M furnace has a 10 X 10X 10 cm uniformly heated zone. The furnacevessel, a quartz “U” tube, was made bybending 61 cm long, 6 mm O.D. X 4 mm IDquartz tubing (Heraeus Amersil Inc., Austin,TX, Tubinger, HSQ300, 4 X 6 X 1220); thesides were 2 cm apart. About 10 cm of thequartz “U” tube was inside the heated zone, 9cm of the reactor was outside the heated zonebut inside the insulation and the outer metalcabinet of the furnace. About 10 cm of thequartz 'U' tube was outside in air. Figure 2shows temperature profile of the furnace andthe quartz “U” tube. When the quartz “U”tube was inside furnace, which wasmaintained generally at 9000C, the ends werecool enough to touch. The use of 1/8” Teflontubing for the gas lines to and from the quartz“U” tube, made handling easier. Swagelok1/4 to 1/8 reducing unions andGraphite/Vespel ferrules (Altech Associates)were used to connect the Teflon tubes to thequartz “U” tube. The outlet of the quartz “U”tube was connected to a gas flow meter (J&Wscientific, ADM 1000). The flow meterdisplayed the gas flow and the digital signalfrom the flow meter was connected to a PCusing RS232 ports. Soft Wedge software wasused to collect the flow meter data into anExcel spreadsheet. The loaded quartz “U”tube was easily placed in and taken out of thefurnace as needed. Either pure hydrogen or10% hydrogen in Argon was used as the feedgas. Hydrogen was added to a cylinder withArgon to obtain the latter. The feed gas wasalways purified with a High Capacity GasPurifier from Supelco that uses a heated
proprietary material for removing O2 and H20from the feed gas. Typically 200mg ofGa2O3/CeO2 or Ga2O3 was positioned on oneside of the quartz “U” tube using two quartzwool plugs. Copper wire was placed on theother side. The system was pressurized withhelium and tested for leaks using a Gow MacHelium Leak Detector (Gow Mac Model 21-250). The success of each experimentdepends on a leak free system, as well asoxygen free feed gas.
The reducing gas (H2 or H2/Ar) wasallowed to flow first through the oxide bed,and then around the copper wire. Generally aflow rate of 10ml/minute was used forhydrogen and a flow rate of 35ml/min wasused for H2/Ar. The feed gas and the flowrates were changed as needed. For eachexperiment the loaded quartz “U” tube withreducing gas-flow was heated in the furnacefor several hours at a constant temperature.Experiments were performed for thetemperature range of 700-1000°C. Ananalytical balance (Sartorius model BP11OS)was interfaced with a PC using a special cablefrom Sartorius for connecting the RS232ports. The software, Soft Wedge, allowedbalance readings to be directly recorded in anExcel spreadsheet.
First, the quartz “U” tube wasweighed. A quartz wool plug was placed atthe bottom of the tube and weighed. Theoxide mixture (200mg) was placed above theplug and the tube was weighed again.Another quartz wool plug was placed abovethe oxide and then the tube was weighedagain. A copper wire (about 2g) was weighedand positioned in the other side of the quartztube. The quartz tubing with the contentswere weighed before and after eachexperiment. If any oxide mixture was left atthe end of an experiment, it was referred to asgallium trioxide treated with reducing gas andwas carefully removed from the tube andweighed. Copper wire and Ga/Cu alloy
4
droplets were weighed. The weight gain bythe copper wire and the Ga/Cu alloyrepresents the gallium removed from theoxide bed. The weight gain was used forcalculating the TIGR conversion. Inexperiments where copper was not used,gallium suboxide was collected at the cooler(> 700°C) parts of the quartz tube; this wasreferred to as TIGR condensed product. Itwas carefully removed, weighed and used forcalculating the conversion. The chemicalcomposition of TIGR condensed product isunknown. Each experiment was associatedwith a number of recorded weights, whichwere used to calculate the weight of oxidemixture before and after the reaction, theweight of gallium collected on copper wire,and the weight of gallium in Ga/Cu alloydroplets. A Varian AA30 atomic absorptionspectrometer was used for the determinationof copper in the Ga/Cu alloy. It is assumedthat Ga was the only other element in thealloy; the weight of Ga in the alloy wasdetermined by subtracting the weight of Cufrom the weight of the alloy. The sampleswere also sent to Texas Tech University forICP analysis. We have also sent samples toGalbraith Laboratory for elemental analysis byICP; however, they could not determine thecomposition of the alloy due to matrixinterference. They were getting a higher valuefor copper (78%) and lower value for gallium(35%) with a total of more than 100%.
The FT-IR spectrometer (NicoletMagna 560), was equipped with two beamsplitters (KBr for mid IR and Solid Substratefor far IR) and three detectors (DTGS withKBr window for Mid IR, DTGS withpolyethylene window for far IR and MCTAfor Diffuse Reflectance and mid IR studies).The FT-IR setup allows the analysis ofsamples in the form of pellets, films, andpowders. Powder samples were filled in microcups and scanned for the Mid IR spectrumusing the MCTA detector and the Diffuse
Reflectance (DRIFT) accessory(SpectraTech). Almost all the samples wereanalyzed as cesium chloride pellets using thetransmission spectrometry. Each pellet (13-mm dia.) was prepared by pressing 100 mg of1 to 10% well-mixed sample in cesiumchloride (99.999%, Aldrich Chemical Co.).The pellets were placed in the sample holderand scanned for both mid and far IR spectrawhile changing the beam splitter and detectoras needed. BET surface area and pore sizedistributions of each sample was measuredusing a Micrometric ASAP 2000 analyzer. X-ray diffraction spectrometer (Scintag, XDS2000) was used for scanning the powder XRDpattern of each sample.
2.1 THREE OPTIONS FOR GALLIUMREMOVAL AND COLLECTION
During the TIGR process some of thegallium suboxide prematurely decomposes togallium metal and gallium trioxide anddeposits on furnace and vent surfaces. Toprevent such deposition of this corrosivegallium three options were suggested. Theyare “cold finger collection” of galliumsuboxide, vapor phase oxidation of galliumsuboxide to nonvolatile gallium trioxide, anduse of a “copper collector.” The followingexperiments were conducted to test andevaluate these options.
2.1.1 Collection of Gallium Suboxide by“Cold Finger”
The experiments were conducted tostudy the TIGR condensed white powder onthe cooler areas of the quartz “U” tube asshown in Figure 3. Mid and far IR spectra ofthe white powder and gallium trioxide isshown in Figure 4. The white powder has adifferent IR spectrum compared to galliumtrioxide as received. Of the three oxides ofgallium (Ga2O3, GaO, Ga2O), GaO is veryunstable and not expected to form. Ga2O isvolatile and deposits on the cooler areas of the
5
quartz tube. Later, it decomposes or getsoxidized to more stable products. Most of thewhite power could be gallium trioxide, butwith a crystal structure different from that ofgallium trioxide as received. Sometimes thedeposit was gray-colored or turned gray. Thepowder XRD patterns of the white and grayproducts are shown in Figure 5 and listed inTable 1. Chemical reactions responsible forthe color change are not well understood.Mid and far IR spectra of the gray product areshown in Figures 6 a and b, respectively. Thespectrum is almost identical with that of whitepowder (Figure 4). It is reasonable to assumethat most of the gray material is galliumtrioxide. There may be a small amount ofgallium present, which could cause the graycolor. Gallium suboxide is known todecompose into gallium and gallium trioxide.Oxygen, even at extremely low levels, iscapable of oxidizing gallium suboxide togallium trioxide. Most of the time the TIGRcondensed product was pure white andcrystalline. Long reaction times converted allthe gallium trioxide to vapors of galliumsuboxide and deposited it on the cooler partsas gallium suboxide, which is later oxidizedto gallium trioxide. The reaction was alsostopped when half of the oxide was convertedand the samples of gallium trioxide remainingin the tube were analyzed using FT-IR andpowder XRD scanning. Mid and far IRspectra of gallium trioxide left in the quartzreactor are shown in Figures 7 and b,respectively. The IR spectrum of galliumtrioxide left after some reaction is differentfrom that of TIGR condensed product, as wellas from that of gallium trioxide as received(Figure 4). Heating gallium trioxide in areducing atmosphere causes the volatilizationof gallium suboxide, while changing thestructure gallium trioxide left behind. It isunknown whether gallium trioxide is merelydehydrated or has undergone phasetransformation. It appears that a material with
new crystal structure is produced. Thepowder XRD pattern of gallium trioxidebefore and after heating in a reducingatmosphere are shown in Figure 8 and listedin Table 2. Sometimes gallium trioxide alsoturned gray when heated in a reducingatmosphere. When H2 was used instead ofH2/Ar, shiny flakes or droplets of galliumwere formed. So the gray color could be dueto tiny amounts of elemental gallium. It couldbe concluded that under certain conditions,gallium suboxide could further be reduced tometallic gallium by hydrogen. Since all theoxides of gallium are white, the gray colorshould be due to elemental gallium.Hydrogen can convert some of the galliumsuboxide deposited to elemental gallium. Theconditions under which gallium suboxide isreduced to metallic gallium are not clearlyunderstood.
2.1.2 Collection by In Situ Air Oxidation ofGallium Suboxide Vapors
The experimental setup is shown inFigure 9. An airflow of 5 ml/minute wasintroduced into the stream of feed gas andgallium suboxide vapors. A white powderdeposited near the air-feed gas (H2/AR)mixing zone. No white powder deposited atthe cooler parts of the tube as in the case of“cold finger” experiments. The IR spectrumand powder XRD pattern of the white powderwere obtained. The IR spectrum (Figure 10)is similar to that of TIGR condensed whiteproduct (Figure 4). Neither IR spectra nor theXRD patterns (Figure 8) indicate the presenceof any gallium trioxide in the as received formor the form obtained after heating in areducing gas. The gallium trioxide could befurther reduced by hydrogen back to galliumsuboxide and finally brought into the coolerparts of the tube and deposited as galliumsuboxide. Apparently this was not happening.Even the reduction of gallium trioxide togallium suboxide is inhibited by traces of
6
oxygen. That is why there was no depositionof gallium suboxide on the cooler areas of thequartz tube. If heavier grains fell below theair-mixing zone, the hydrogen rich zone couldreduce and produce gallium suboxide vapors.It is not known whether the latter ishappening.
2.1.3 Collection of Gallium on CopperThe experimental setup is shown in
Figure 11. A coiled copper wire is kept in thepath of gallium suboxide vapors. Coppercatalyzes the reduction of gallium suboxide toelemental gallium, which was collected at thecopper wire tip as droplets that melt at 650OC(50% Ga based on phase diagram). It is quiteprobable that during the melting pointexperiments, the droplet melting observed issurface melting while the interior could stillbe solid. There may be more copper insidethe droplet. The copper in the droplet wasestimated at about 60% by elemental analysisusing atomic absorption (AA) spectrometry.It is concluded that the droplets containedabout 40% gallium. The inner part of thedroplet is rich in Cu and the outer part is richin Ga. When the droplet breaks off it mayhave an inner core rich in Cu. The use of alarger rod of copper may result in Ga/Cu alloy
rich in Ga, which would drop off rather thanbreaking off parts of the thinner wire. Whencopper is used as a collector, no white depositwas collected in the cooler areas of the quartz“U” tube. Several experiments wereconducted to study various aspects of galliumcollection on copper; they are discussed indetail later.
The three current options forimproving the TIGR process, which could beused for Ga removal from weaponsplutonium, are tested by the three types ofexperiments discussed above. Coppercollector is believed to be the best choicebecause of the clean collection of Ga withoutgallium suboxide or trioxide that can becarried away into other parts of the furnaceand vent. Ga/Cu alloy is also easier to handlethan the oxides. Since the reduction ofgallium suboxide takes place at the tip of thecopper wire, the reaction is very fast and doesnot need a high copper surface area. A heavycopper rod could be use as the collector, anddrippings from the rod will have higher Gacontent. Experiments should be performed todetermine the relationship between thegeometry of the copper collector and theGa/Cu alloy composition.
The Oxidation and Reduction Reactions of Gallium Species
Ga2O3 + 2H2 → Ga2O + 2H2O (1)
Ga2 O + 2Cu → Cu2 O + 2Ga (2)
Cu2O + H2 → 2Cu + H2O (3)
Ga2O3 + 3H2 → 2Ga + 3H2O (4)
3Ga2O → 4Ga + Ga2O3 (5)
Ga2O3 + H2 → 2GaO + H2O (6)
2Ga + O2 → 2GaO (7)
Ga2O + O2 → Ga2O3 (8)
7
4Ga + 6O2 → 3Ga2O3 (9)
2Ga + N2 → 2GaN (10)
Heat of Formation of Gallium Oxides
2Ga + 3/2O2 → Ga2O3 + 257 61 kcal/mole (11)
Ga2O + O2 → Ga2O3 + 175 62 kcal/mole (12)
2Ga + 1/2O2 → Ga2Osolid + 82 62 kcal/mole (13)
3Ga2O → 4 Ga + Ga2O3 + 11 6 6 kcal/mole (14)
Equations 1-9 list all the possible reactions ofgallium species in the presence of hydrogen,nitrogen, oxygen and copper. Even a smallamount of oxygen can suppress the Garemoval process. The system must be leakproof and oxygen free. Nitrogen is known toreact with Ga to produce Gallium nitride,which is easily oxidized to Ga2O3. Morestudy is needed in this area.
2.2 REACTIONS OF METALS ANDALLOYS IN GALLIUM SUBOXIDEVAPOR
Experiments were conducted byreplacing the copper wire in Figure 12 withwires of other metals and alloys. Theexperimental conditions were the same as the“copper collector” experiments using H2/Ar asthe feed gas and a temperature of 9000 C.Several metals and alloys were tested todetermine their activity towards galliumsuboxide vapors. The XRD pattern of powderproduct from the two experiments are shownin Figure 13 and listed in Table 3. All ofthese experiments with correspondingobservations or comments are listed in Table4. Platinum and rhodium/ platinum (10%)alloy has some Ga collecting ability but not asgood as a copper collector. A gray powderdeposition was observed on most materials
including stainless steal. Zirconium crumbledin gallium vapor forming a gray granularmaterial. Tungsten and molybdenum werefound to be very inert. It could be concludedthat copper is the best material for galliumcollection. Tungsten and molybdenumpossibly could be used for building thefurnace vessel for the TIGR process since theyare the mostly inert.
2.2.1 Effect of Reaction Temperature onGallium Removal
Figure 14 shows the effect oftemperature on gallium removal using H2/Arand copper collector. Figure 15 shows theeffect of temperature on gallium removalusing H2 and copper collector. The lowesttemperature at which gallium could becollected on copper is 8000C. Highesttemperature tested was 10000C.
2.2.2 Effect of Copper to Ga2O3 BedDistance on Gallium Collection
A series of experiments wereconducted using the setup in Figure 11 todetermine if the distance between the copperand the gallium oxide bed has any effect ongallium collection on copper. Figure 16shows the result of the experiments usingH2/Ar. Figure 17 shows the result of the
8
experiments using H2. The distanceapparently has no effect on the conversion aslong as the copper is within the hot zone.
2.2.3 BET Surface Area MeasurementsThe change in the surface area of
gallium trioxide with respect to reactionconditions shown are in Figure 18 and alsolisted in Table 5.
2.3 VOLATILIZATION OF GALLIUMSUBOXIDE
Beta-gallium oxide was prepared byreducing gallium trioxide in a quartz reactorusing either hydrogen or hydrogen/argonusing the “cold finger” experiment. Galliumsuboxide first condenses on the cooler parts ofthe quartz tube and then undergoes chemicalchanges to produce a white powder that issometimes mixed with gray color. Galliumsuboxide is very unstable and could easily beoxidized to gallium trioxide. Both the IRspectrum and the XRD pattern of TIGRcondensed product is very different from thatof gallium trioxide as received. The Ga2O3
bed in the “cold finger” (Figure 4) wasreplaced by TIGR condensed product. In one
series of experiments, pure hydrogen waspassed. Another series of experiments usedthe conditions of copper collector experiments(Figure 12). The disappearance of TIGRcondensed product and the weight gain due tothe formation of Ga/Cu alloy on copper tipwere monitored. The results of these twoexperiments are listed in Table 6. Galliumsuboxide is initially volatilized from Ga2O3
bed by passing hydrogen or H2/Ar. The vaporcondenses on cooler parts of the furnace. Thevolatilization of gallium suboxide, oncecondensed, is a slow process probably due tothe chemical changes it is undergoing.Gallium suboxide is very unstable andexposure to air probably oxidizes it to galliumtrioxide but with a different crystal structureand IR absorption spectrum. The chemicalnature of the product is unknown and it isreferred to as TIGR condensed product on“cold finger.” The volatilization of suboxidefrom TIGR condensed product is favored by aflow of a reducing gas such as H2 and H2/Ar.Figure 19 shows the effect of reaction time onconversion of TIGR condensed product in H2
and Figure 20 shows the effect of flow rate onconversion of TIGR condensed product in H2.
9
3. CONCLUSIONSGallium trioxide exists in α, β, and γ
crystalline forms. The alpha form is stableand commercially available. It appears that thethree forms are encountered during theexperiments. It is quite likely that all of theseforms will have different reactivity. Forexample β-gallium trioxide could easily bereduced to gallium suboxide. If the mostreactive form of gallium trioxide is producingduring the HYDOX process, the galliumremoval could be accomplished in short andless rigorous conditions.
The distance between copper collectorand the oxide bed does not have an effect ongallium removal or conversion. In addition,hydrogen gives a better conversion thanH2/Ar. The rate of conversion increases withincrease in temperature and gas flow rates.
When gallium trioxide is heated in H2
or H2/Ar, gallium suboxide vaporized anddeposited cooler parts of the quartz “U” tubeand a white powder is produced. However,sometimes the powder is gray colored. Whenheating at 950°C in a reducing gas (H2 orH2/Ar), both white and gray powderssublimed in about thirty minutes. This is asignificant decrease in reaction rate comparedto the as-received powder.
Air oxidation of gallium suboxidevapors produces a nonvolatile white powderwhich deposits on the quartz “U” tube insidethe furnace. The far IR spectrum of theoxidized product is very similar to that of thewhite deposit collected on the cooler parts ofthe quartz during the heating of Ga2O3 in H2
or H2/Ar. It should be noted that traceamounts of oxygen and perhaps nitrogencould be affecting the TIGR process. Anoxygen sensor capable of monitoring very lowlevels of oxygen is needed to better quantifythis effort.
11
Figure 1: Schematic Diagram of the Experimental Setup used for Gallium Removal
High Capacity Gas Purifier
Reactor Furnace
H2 / Ar
Metering Valve
Regulator Digital Flowmeter
Data Acquisition
12
Figure 2: Experimental Setup for Gallium Removal
H2 / Ar 35 ml/min
10.1cm
832°C755°C
433°C476°C
538°C
667°C597°C
355°C403°C
858°C878°C888°C894°C896°C896°C
896°C896°C898°C898°C899°C
Ga2O3
Quartz Packing
Temperature profile in the furnace
330°C358°C389°C
451°C479°C552°C620°C690°C753°C
Heating Zone10.1cm
Top
Bottom
894°C
900°C898°C897°C
896°C895°C
890°C
882°C869°C
813°C843°C
Insulation Zone
Temperature profile in the Reactor( H2 /Ar flow = 35ml/min )
93°C125°C
164°C210°C245°C268°C306°C
Outlet
65°C
13
Figure 3: Experimental Setup for Gallium Removal by “Cold Finger” Deposition of Product
355°C
899°C
Ga2O3
Quartz Packing
858°C
896°C
896°C
898°C
888°C
476°C
755°C
597°C
433°C
Temperature profile in the furnace
Inlet H2/Ar
Heating Zone
Insulation Zone
330°C
10.1cm
Top
Bottom10.1cm900°C
389°C
451°C
813°C
897°C
895°C
890°C
869°C
690°C
552°C
Ga2O
Temperature profile in the Reactor( H2 /Ar flow = 35ml/min )
210°C
268°C
65°C
125°C
Outlet
14
Figure 4a: Mid-IR Spectra of Deposited White Product (top) and Gallium Trioxide (bottom),Respectively
Figure 4b: Far-IR Spectra of Deposited White Product (top) and Gallium Trioxide (bottom),Respectively
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
Abs
orb
anc
e
500 1000 1500 2000 2500 3000 3500 4000
cm-1
-1.6
-1.4
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
Abs
orb
anc
e
250 300 350 400 450 500 550 600 650 700
cm-1
15
Figure 5: X-Ray Diffraction Pattern of White Product (top) and Gray Product (bottom),Respectively. [White Product is from Run 50 and the Gray Product is from Run 69.]
Table 1: XRD Peak Positions, d-Spacing Values of White and Gray Products
White Product Grey ProductTwo Theta d Spacing, A Intensity I/Imax Two
Thetad Spacing, A Intensity I/Imax
30.14 2.96 208752 100.0 30.10 2.97 33202 66.630.51 2.93 149150 71.4 30.48 2.93 26943 54.131.75 2.82 133579 64.0 31.74 2.82 47417 95.133.51 2.67 29677 14.2 33.48 2.67 14092 28.335.23 2.55 125696 60.2 35.22 2.55 49835 100.037.49 2.40 94724 45.4 37.45 2.40 19847 39.838.43 2.34 81441 39.0 38.42 2.34 23662 47.545.85 1.98 44435 21.3 43.05 2.10 6004 12.048.66 1.87 29448 14.1 45.82 1.98 11703 23.557.62 1.60 29087 13.9 48.65 1.87 8000 16.159.93 1.54 25628 12.3 57.62 1.60 10198 20.560.58 1.53 29334 14.1 59.91 1.54 7416 14.962.71 1.48 23566 11.3 60.96 1.52 8163 16.464.14 1.45 23713 11.4 64.13 1.45 7968 16.064.71 1.44 77935 37.3 64.68 1.44 19632 39.4
16
Figure 6a: Mid-IR Spectrum of Deposited Gray Product. Conditions: Reaction Temperature =950oC, Treatment Gas = H2/Ar (35 ml/min) and Reaction Time = 2.0 hr.
Figure 6b: Far-IR Spectrum of Deposited Gray Product. Conditions: Reaction Temperature =950oC, Treatment Gas = H2/Ar (35 ml/min) and Reaction Time = 2.0 hr.
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
Abs
orb
anc
e
500 1000 1500 2000 2500 3000 3500 4000
cm-1
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Abs
orb
anc
e
250 300 350 400 450 500 550 600 650 700
cm-1
17
Figure 7a: Mid-IR Spectrum of Gallium Trioxide after the Reaction. Conditions: ReactionTemperature = 950oC, Treatment Gas = H2/Ar (35 ml/min) and Reaction Time = 2.0 hr.
Figure 7b: Far-IR Spectrum of Gallium Trioxide after the Reaction. Conditions: ReactionTemperature = 950oC, Treatment Gas = H2/Ar (35 ml/min) and Reaction Time = 2.0 hr.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
Abs
orb
anc
e
250 300 350 400 450 500 550 600 650 700
cm-1
*
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Abs
orb
anc
e
500 1000 1500 2000 2500 3000 3500 4000
cm-1
18
Figure 8: X-Ray Diffraction Pattern of Gallium Trioxide Before Reaction and Gallium Trioxideafter Reaction from Run N-23 (bottom)
Table 2: XRD Peak Positions (20), d-Spacings and Relative Intensities of Gallium Oxides
Gallium Trioxide Before Reaction Gallium Trioxide After Reaction (Run N-23)Two Theta D Spacing, A Intensity I/Io Two
ThetaD Spacing, A Intensity I/Io
24.88 3.58 11796 19.6 30.42 2.94 63618 52.330.57 2.92 4832 8.0 30.78 2.90 62273 51.232.10 2.79 8640 14.3 32.03 2.79 121612 100.034.14 2.62 60314 100.0 33.78 2.65 30610 25.235.62 2.52 8356 13.9 35.51 2.53 119589 98.336.37 2.47 45803 75.9 37.76 2.38 34828 28.641.80 2.16 10655 17.7 38.71 2.32 69851 57.450.54 1.80 23921 39.7 43.29 2.09 16533 13.655.38 1.66 35527 58.9 46.14 1.97 35898 29.559.29 1.56 8825 14.6 48.94 1.86 31500 25.963.59 1.46 17265 28.6 57.87 1.59 34639 28.565.04 1.43 27827 46.1 60.18 1.54 26661 21.974.04 1.28 13652 22.6 61.22 1.51 34378 28.376.60 1.24 4262 7.1 64.38 1.45 23815 19.682.57 1.17 4262 7.1 64.97 1.43 82125 67.5
19
Figure 9: Experimental Setup for Gallium Removal by In-Situ Oxidation of Ga2O
10.1cm
Ga2O3
Quartz Packing
896°C
898°C
899°C
896°C
858°C
888°C
755°C
Temperature profile in the furnace
597°C
476°C
433°C
355°C
Inlet H2/Ar
Heating Zone10.1cm
Top
Bottom900°C
897°C
895°C
690°C
890°C
869°C
813°C
Ga2O3
Insulation Zone
Temperature profile in the Reactor( H2 /Ar flow = 35ml/min )
65°C
389°C
330°C
451°C
552°C
125°C
210°C
268°C
Outlet
Air
SS tubing
20
Figure 10: Far-IR Spectra of Reoxidized (In-Situ) Product (top) and White Product (bottom).Conditions for Reoxidized Product: Reaction temperature (RT) = 1000oC, Treatment Gas = H2/Ar
(35 ml/min), Reoxidation Gas = Air (5 ml/min) and Reaction Time = 2 hr. Conditions for DepositedWhite Product: Reaction Temperature = 950oC, H2/Ar (35 ml/min) for 2 hr.
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
Abs
orb
anc
e
250 300 350 400 450 500 550 600 650 700
cm-1
21
Figure 11: Experimental Setup for Gallium Removal using Copper Collector
10.1cm
755°C
Quartz Packing
Ga2O3898°C
899°C
896°C
858°C
888°C
896°C
Temperature profile in the furnace
355°C
476°C
597°C
433°C
Inlet H2/Ar
Heating Zone
Insulation Zone
690°C
Gallium/Copper Alloy
10.1cm
Top
Bottom900°C
897°C
895°C
890°C
869°C
813°C
Temperature profile in the Reactor
H2/Ar flow=35ml/min330°C
389°C
451°C
552°C
125°C
210°C
268°C
Outlet
65°C
Copper Wire
22
Figure 12: Experimental Setup for Gallium Removal using Stainless SteelCollector
10.1cm
755°C
Quartz Packing
Ga2O3898°C
899°C
896°C
858°C
888°C
896°C
Temperature profile in the furnace
355°C
476°C
597°C
433°C
Inlet H2/Ar
690°C
Heating Zone10.1cm
Top
Bottom900°C
897°C
895°C
890°C
869°C
813°C
Insulation Zone
Temperature profile in the Reactor( H2 /Ar flow = 35ml/min )
330°C
389°C
451°C
552°C
125°C
210°C
268°C
65°C
Outlet
Steel Wire
23
Figure 13: X-Ray Diffraction Pattern of the Scratch from Stainless Steel Wire from Run N-4 (top)and Crumbled Zirconium Wire from Run 42 (bottom), Respectively
Table 3: XRD Peak Positions and d Spacing Values for the Deposited Products onStainless Steel and Zirconium Wires
Scratch from Stainless Steel Wire Crumbled Zirconium WireTwo Theta d Spacing, A Intensity I/Imax Two Theta d Spacing, A Intensity I/Imax
30.11 2.97 7871 70.9 28.29 3.15 17727 34.330.49 2.93 3903 35.1 31.58 2.83 13550 26.231.72 2.82 11106 100.0 32.53 2.75 51648 100.033.48 2.67 3142 28.3 36.08 2.49 7057 13.735.22 2.55 10852 97.7 36.19 2.48 7421 14.437.47 2.40 2753 24.8 36.58 2.45 11243 21.838.39 2.34 4885 44.0 37.81 2.38 6145 11.943.02 2.10 1910 17.2 38.23 2.35 17802 34.544.09 2.05 7036 63.4 45.84 1.98 8529 16.545.83 1.98 3058 27.5 54.20 1.69 5398 10.548.66 1.87 2154 19.4 54.35 1.69 5112 9.957.61 1.60 3334 30.0 55.35 1.66 4529 8.860.92 1.52 2627 23.7 63.47 1.46 4379 8.564.07 1.45 2095 18.9 67.96 1.38 4360 8.464.69 1.44 5848 52.7 76.40 1.25 7659 14.8
24
Figure 14: % Ga2O3 Conversion as a Function of Reaction Temperature in H2/Ar Flow.Conditions: Flow rate =35 ml/min, sample wt=200 mg and reaction time=2 hr.
Figure 15: % Ga2O3 Conversion as a Function of Reaction Temperature in H2 Flow.Conditions: Flow rate=35 ml/min, sample wt =200 mg and reaction duration = 2 hr.
0
10
20
30
40
850 900 950 1000 1050
Temperature, oC
% C
onv
ers
ion
0
10
20
30
40
50
60
70
80
90
750 800 850 900 950 1000 1050
T emperature, oC
%
Co
nve
rsio
n
25
0
10
20
30
40
50
0 2 4 6 8
Cu Distance, cm
% C
onv
ers
ion
Figure 16: % Ga2O3 Conversion as a Function of Copper Distance.Conditions: Pretreatment gas=H2 /Ar (35 ml/min) reaction duration=2 hr.
Figure 17: % Ga2O3 Conversion as a Function of Copper Distance.Conditions: Reaction Temperature=900oC, Pretreatment
Gas=H2 (35 ml/min), and Reaction Duration=2 hr.
0
20
40
60
80
100
0 2 4 6 8 10 12
Cu Distance, cm
% C
onve
rsio
n
26
Figure 18: Effect of Hydrogen Pretreatment on BET Surface Area of Gallium Trioxide.Condition: Pretreatment Gas=H2 (35 ml/min) Reaction Duration=2 hr.
Figure 19: %Conversion of TIGR Process Product as a function of Reaction TimeConditions: Reaction Temperature=1000oC, H2 flow=35 ml/min, Cu distance=1.2 to 1.3 cm and
sample size =20 mg.
0
20
40
60
80
100
0 5 10 15 20
Run T ime, Minutes
% C
onve
rsio
n
0
10
20
30
0 200 400 600 800 1000
T emperature, oC
BET
Surfa
ce
are
a, m
2 /g
Fresh Ga2O3
27
Figure 20: % Conversion of TIGR Process Product as a function of Flow Rate. Conditions: Reactiontemperature = 1000oC, Reaction Time = 10 minutes, Sample Size = 20 mg and Pretreatment Gas = H2.
0.0
20.0
40.0
60.0
80.0
100.0
0 10 20 30 40 50 60
Flow Rate, ml/min
% C
onv
ers
ion
28
Tab
le 4
: S
umm
ary
Tab
le fo
r th
e G
alliu
m E
xper
imen
ts C
ondu
cted
with
Diff
eren
t Met
als
Run
#T
ype
of M
etal
Sub
stra
teR
eact
ion
Tem
pera
ture
,o C
Typ
e of
Gas
& (
Flo
wR
ate,
ml/m
in)
Dur
atio
n of
Pre
trea
tmen
tC
om
men
ts
6930
4 S
S w
ire90
0H 2/
Ar
(35)
24S
S w
ire a
sta
ble
and
smal
l am
ount
of d
ark
prod
uct
was
dep
osite
d on
the
ss w
ire.
7031
6 S
S w
ire90
0H 2/
Ar
(35)
24S
S w
ire a
sta
ble
and
smal
l am
ount
of d
ark
prod
uct w
as d
epos
ited
on th
e ss
wire
.26
Zirc
oniu
mw
ire90
0H 2
/Ar
(15)
24Z
ircon
ium
wire
cru
mbl
ed
27P
t-R
h90
0H 2
(35)
24S
mal
l dro
plet
s of
gal
lium
form
ed.
28P
t-R
h90
0H 2/
Ar
(35)
24V
ery
smal
l Gal
lium
dro
plet
s fo
rmed
.37
Tita
nium
900
H 2/A
r (3
5)24
Tita
nium
wire
was
sta
ble
but t
urne
d in
to d
ark.
38N
b fo
il90
0H 2
/Ar
(35)
24N
b fo
il w
as s
tabl
e bu
t tur
ned
into
dar
k.41
Ste
rling
silv
er90
0H 2/
Ar
(35)
25S
terli
ng s
ilver
mel
ted
and
also
gai
ned
som
e w
eigh
tdu
ring
the
reac
tion
and
appe
ars
to s
omew
hat
reac
tive
with
gal
lium
sub
oxid
e.12
7T
ungs
ten
filam
ent
950
H 2 (
20)
5T
ungs
ten
filam
ent w
as s
tabl
e an
d th
ere
is n
ow
eigh
t gai
n on
the
filam
ent.
N-2
7M
olyb
denu
mw
ire10
00H 2
(20
)5
Mol
ybde
num
was
sta
ble
and
a w
hite
dep
osits
of g
alliu
m s
ubox
ide
was
dep
osite
d on
the
cool
er p
arts
of M
olyb
denu
m w
ire.
29
Tab
le 5
: E
ffect
of P
retr
eatm
ent o
n th
e B
ET
Sur
face
Are
a* o
f Gal
lium
Trio
xide
Sam
ple
Des
crip
tion
Tem
pera
ture
, o C
Pre
trea
tmen
t Gas
BE
T S
urfa
ce A
rea,
m2 /g
Run
#
Ga 2
O3
Bef
ore
Rea
ctio
n20
0N
one
17.6
N/A
Ga 2
O3
Afte
r R
eact
ion
500
H 2 (
2 hr
)14
.0R
un N
-37
Ga 2
O3
Afte
r R
eact
ion
800
H 2 (
2 hr
)11
.5R
un 1
46
Ga 2
O3
Afte
r R
eact
ion
900
900
H2
(4 h
r)H
2 (2
hr)
6.7
6.2
Run
74
Run
79
Ga 2
O3
Afte
r R
eact
ion
950
950
H2
(5.5
hr)
H2
(2 h
r)6.
74.
0R
un 9
2R
un 8
7
Ga 2
O3
Afte
r R
eact
ion
1000
Air
(5 h
r)10
.9R
un 1
39
* P
rior
to B
ET
mea
sure
men
ts th
e sa
mpl
es w
ere
dega
ssed
und
er v
acuu
m a
t 200
° C fo
r 10
hr.
30
Tab
le 6
: E
ffect
of H
2 F
low
Rat
e on
the
% C
onve
rsio
n of
TIG
R P
roce
ss P
rodu
ct
(Con
ditio
ns: T
empe
ratu
re =
100
0o C
, Pre
trea
tmen
t Gas
= H 2 an
d S
ampl
e S
ize
= 2
0 m
g.)
Run
##
Flo
w R
ate,
ml/m
inR
un T
ime,
min
utes
% C
onve
rsio
n ba
sed
on
Cu
wt g
ain
S
ampl
e w
t los
s
111
355
27.6
30.4
110
3515
92.4
98.1
113
1030
59.2
64.8
109
3510
55.9
49.5
112
5010
65.5
77.2
31
Tab
le 7
: E
xper
imen
tal R
esul
ts fo
r T
herm
ally
Indu
ced
Gal
lium
Rem
oval
(T
IGR
) P
roce
ss U
sing
“C
old
Fin
ger”
Run
Rea
ctio
nT
empe
ratu
re, o C
Red
ucin
gG
asG
as F
low
Rat
e, m
l/min
Rea
ctio
nD
urat
ion,
hr
% C
onve
rsio
n ba
sed
on
Dep
osite
d
S
ampl
e W
t los
s
9990
0H 2
102
6.3
6.4
7690
0H 2
352
16.6
18.1
6495
0H 2
/Ar
302
6.3
5.5
5795
0H 2
/Ar
352
11.8
10.3
6395
0H 2
/Ar
352
10.6
9.3
8795
0H 2
252
24.4
26.6
8495
0H 2
352
41.4
39.1
9195
0H 2
352
39.9
39.6
8995
0H 2
354
55.5
52.8
6610
00H 2
/Ar
352
29.3
24.2