Citation
The Design, Synthesis, and Application of Volatile Organometallic Precursors for Heterogeneous Catalysis, Photoactivated Chemical Vapor Deposition, and Electron Beam Induced Deposition

Material Information

Title:
The Design, Synthesis, and Application of Volatile Organometallic Precursors for Heterogeneous Catalysis, Photoactivated Chemical Vapor Deposition, and Electron Beam Induced Deposition
Creator:
Brannaka, Joseph A
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (140 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Chemistry
Committee Chair:
MCELWEE-WHITE,LISA ANN
Committee Co-Chair:
CASTELLANO,RONALD K
Committee Members:
WEAVER,JASON F
MILLER,STEPHEN ALBERT
JAMES,DELORES CORINNE SUZETTE
Graduation Date:
12/19/2014

Subjects

Subjects / Keywords:
Carbon ( jstor )
Catalysis ( jstor )
Catalysts ( jstor )
Chlorides ( jstor )
Electron beams ( jstor )
Irradiation ( jstor )
Ligands ( jstor )
Oxygen ( jstor )
Ruthenium ( jstor )
Vapor deposition ( jstor )
Chemistry -- Dissertations, Academic -- UF
beam -- catalysis -- catalyst -- chemical -- deposition -- design -- electron -- heterogeneous -- induced -- organometallic -- precursor -- synthesis -- vapor -- volatile
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Chemistry thesis, Ph.D.

Notes

Abstract:
The design and synthesis of organometallic precursors has proven very useful in the chemical vapor deposition (CVD) field. By taking a mechanistic approach the deposition temperatures have been lowered and cleaner deposits have been formed. A similar approach will be key in improving three additional fields: heterogeneous catalysis, photoactivated chemical vapor deposition (PACVD), and electron beam induced deposition (EBID). Precursors were designed and synthesized with hydroxyl containing ligands to direct the precursor to the undercoordinated sites of the oxide surface for a model study on heterogeneous catalysis. Unfortunately the hydroxyl containing precursors decomposed before sublimation and thus were not able to be delivered to the surface. A volatile precursor that did not contain the desired functional group was synthesized and deposited in a largely random manner. Designed precursors for PACVD on self-assembled monolayers yielded selective deposition of ruthenium; however the cyclopentadienyl and allyl ligands proved difficult to remove under these conditions. The acidity of the terminal groups of the substrate proved important to the deposition process. The use of designed ruthenium precursors has given insight into the mechanism of EBID. Allyl ligands contribute to carbon contamination in the deposit while carbonyl ligands dissociate cleanly under electron irradiation. Halogens were not removed during deposition but have been easily removed by post-deposition processing. These contributions will help the design of better precursors for all three applications in the future. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2014.
Local:
Adviser: MCELWEE-WHITE,LISA ANN.
Local:
Co-adviser: CASTELLANO,RONALD K.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2016-12-31
Statement of Responsibility:
by Joseph A Brannaka.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
12/31/2016
Resource Identifier:
974007318 ( OCLC )
Classification:
LD1780 2014 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

CationicMetallogermyleneandDicationicDimetallodigermenes: SynthesisbyChlorideAbstractionfrom N HeterocyclicCarbeneStabilizedChlorometallogermylenesKoyaInomata,TakahitoWatanabe,andHiromiTobita *DepartmentofChemistry,GraduateSchoolofScience,TohokuUniversity,Aoba-ku,Sendai980-8578,Japan*SSupportingInformationABSTRACT: ReactionofNHC-stabilizeddichlorogermylenes(NHC= N -heterocycliccarbene)withananionic tungstencomplexproducedNHC-stabilizedchlorometallogermylenes.Subsequentchlorideabstractionfromthe productswithNaBAr4(Ar=3,5-(CF3)2C6H3)gavea cationicmetallogermyleneordicationicdimetallodigermenes.Digermene,R2Ge GeR2,hasattractedgreatinterest becauseofitspeculiar trans -bentstructureandhigh reactivity.1Sincetheisolabledigermenewas rstreportedby Lappertetal.in1976,2digermeneswithavarietyof substituentshavebeenprepared.Thus,Lappertetal. synthesizedtetraalkyldigermeneDsi2Ge GeDsi2(Dsi=CH(SiMe3)2)bythereactionofastablegermylene[N(SiMe3)2]2Ge:withLiDsi.2Masamuneetal.reportedthe synthesisoftetraaryldigermeneAr 2Ge GeAr 2(Ar =2,6Me2C6H3)byphotochemicalredistributionofcyclotrigermane cyclo -(Ar 6Ge3).3Theyalsoreportedthedirectformationof tetraaryldigermeneDipp2Ge GeDipp2(Dipp=2,6-iPr2C6H3) byreductivecouplingofextremelycongesteddichlorogermane Dipp2GeCl2.4ThisreactionwasalsousedbyKiraetal.forthe synthesisofsilyl-substituteddigermenes(R3Si)2Ge Ge(SiR3)2(R3=iPr2Me,tBuMe2,iPr3).5Althoughvarioustypesof digermenehaveeverbeenreported,transitionmetalsubstituteddigermeneshaveneverbeensynthesized. NHC-stabilizedchlorogermyleneshaverecentlybeenrecognizedasusefulprecursorsforavarietyofuniquelow-valent germaniumspecies.6Inspiredbytheseresearches,webeganto applyNHC-stabilizedchlorogermylenesasprecursorsforthe synthesisoftransitionmetal-substitutedlow-valentgermanium species.Asaresult,wehavesucceededinthesynthesisofa metallogermyleneCp(CO)3WGe[GeCl(Mes)2](MeIiPr)(MeIiPr =1,3-diisopropyl-4,5-dimethylimidazol-2-ylidene),byintroductionofametalfragmenttoanNHC-stabilizedchlorogermylene GeCl[GeCl(Mes)2](MeIiPr),asapotentialprecursorforelusive digermavinylidenecomplexesM Ge GeR2.7Herewereport thesynthesisofNHC-stabilizedchlorometallogermylenesand theirconversiontothe rstexamplesofacationicmetallogermyleneordicationicdimetallodigermenesbychloride abstraction. ReactionsofGeCl2(NHC)(NHC=IPr,6d,eMeIiPr,6aMeIMe6g,h)(IPr=1,3-bis(2,6-diisopropylphenyl)imidazole-2-ylidene,MeIMe=1,3,4,5-tetramethylimidazol-2-ylidene)with [Li(thf)2][Cp * W(CO)3]intolueneatroomtemperaturefor 2hgaveNHC-stabilizedchlorometallogermylenesCp * (CO)3WGeCl(NHC)( 1a ,NHC=IPr95%; 1b ,NHC=MeIiPr71%; 1c ,MeIMe46%)(Scheme1).Thesecomplexes werefullycharacterizedby1HNMR,13CNMR,andIR spectroscopy,elementalanalysis,andX-raycrystalstructure analysis.Inthe13CNMRof 1a c ,threesignalsassignedtothe COligandswereobserved.Thisimpliesthatthegermanium centerof 1a c ischiralandthiscausestwoCOligandscisto thegermaniumligandmutuallyinequivalent. Themolecularstructureof 1c isshowninFigure1.TheW centeradoptsafour-leggedpiano-stoolgeometryandtheGe centerispyramidalized(thesumofbondanglesaroundGe: 315.1 ° ).Thisisattributabletotheexistenceofalonepairon theGeatom.TheW Gebondlength(2.7413(11)Å)is slightlyshorterthanthoseofotherbase-stabilizedmetallogermylenesCp(CO)3WGe(MeIiPr)GeCl(Mes)2(2.8127(8) Å)7andCp(CO)3WGe[N(SiMe3)C(Ph)C(SiMe3)(2C5H4N)](2.852(1)Å).8TheGe Cbondlength(2.083(10) Å)undergoesverylittlechangefromthatofthestarting dichlorogermyleneGeCl2(MeIMe)(2.082(3)).6gTheGe Cl bondlength(2.330(3)Å)iscomparabletothatof GeCl2(MeIMe)(2.3363(1)and2.3019(1)Å).6gTreatmentofCp * (CO)3WGeCl(IPr)( 1a )withNaBAr4(Ar =3,5-(CF3)2C6H3)in uorobenzeneatroomtemperature immediatelygaveacationicmetallogermylene[Cp * (CO)3WGe(IPr)](BAr4)( 2 )asbluecrystalsin59%yield (Scheme2).9AlthoughtheNMRandUV/visspectraof 2 in solutionwereunabletobemeasuredduetoitsinstabilityand poorsolubilityofthecrystalsin uorobenzene, o -dichlorReceived: June16,2014 Scheme1.SynthesisofNHC-Stabilized Chlorometallogermylenes1a c Communication pubs.acs.org/JACS ©XXXXAmericanChemicalSocietyAdx.doi.org/10.1021/ja506018f | J.Am.Chem.Soc. XXXX,XXX,XXX XXX

PAGE 2

obenzene,andacetonitrile,thesolid-statestructureof 2 was determinedbyX-raycrystalstructureanalysis(Figure2).There aretwoindependentmoleculeswithnearlyidenticalstructures inanasymmetricunitcell.TheGecenterof 2 istwocoordinateandtheW Ge Cbondisgreatlybent (112.4(2) ° ).10TheW Gebondlength(2.5787(10)Å)10is signi cantlyshorterthanthatofthechlorometallogermylenes 1a c (2.7413(11) 2.7894(8)Å).11Thisshorteningis attributabletotheincreaseof back-donationfroma lledd orbitaloftheWtotheemptyporbitaloftheGe,inother words,delocalizationofapositivechargeovertheW Ge C bond,whichisre ectedinthehigh-wavenumbershiftofthe CObandsfrom 1a (1950,1876,1839cm 1)to 2 (1996,1932, 1907cm 1). Clabstractionfromchlorometallogermyleneswithsterically lesshinderedNHCs, 1b and 1c ,ledtodi erentresults.Thus, reactionof 1b withNaBAr4in uorobenzenea ordeda dicationicdimetallodigermene 3 ,adimerofacationic metallogermylene,asorange/greendichroiccrystalsin57% yield(Scheme3).Inthecaseofthereactionof 1c withNaBAr4underthesamereactionconditions,adicationiccomplex 4 was obtainedasredcrystalsin61%yield.Thesecrystalswere almostinsolubleinorganicsolvents,butwerecharacterizedby IRspectroscopy,elementalanalysis,andX-raycrystallography. Thestructuresofthedicationicpartsof 3 and 4 andthe NewmanprojectionsalongtheGe Gebondareshownin Figures3and4,respectively.Theselectedbondlengthsand anglesof 3 and 4 areshowninTable1.Complex 3 takesa twisted Z -con gurationaccompaniedbya trans -bentgeometry ( trans -bentangle=34.6and35.3 ° ;twistangle=50.0 ° ).12The Ge Gebondlengthof 3 (2.4286(11)Å)isintherangeof previouslyreportedGe Gedoublebondlengthsofdigermenes (2.21 2.51Å).1eOntheotherhand,complex 4 adoptsan untwisted E -con gurationwitha trans -bentgeometry( trans Figure1. ORTEPdrawingof 1c withthermalellipsoidsat50% probabilitylevel.Hydrogenatomsareomittedforclarity.Selected bondlengths(Å)andangles(deg):W Ge2.7413(11),W C(1) 1.973(12),W C(2)1.950(11),W C(3)1.969(12),Ge Cl2.330(3), Ge C(14)2.083(10),C(1) O(1)1.143(14),C(2) O(2)1.168(14), C(3) O(3)1.159(14),W Ge Cl109.51(8),W Ge C(14) 110.8(2),Cl Ge C(14)94.8(3). Scheme2.ClAbstractionfromChlorometallogermylene1a Figure2. ORTEPdrawingofoneofthetwoindependentcationic partsof 2 withthermalellipsoidsat50%probabilitylevel.Allhydrogen atomsareomittedforclarity.Selectedbondlengths(Å)andangles (deg):W(1) Ge(1)2.5780(10),W(1) C(1)1.979(11),W(1) C(2)1.981(11),W(1) C(3)1.978(9),Ge(1) C(14)2.032(8), C(1) O(1)1.155(13),C(2) O(2)1.161(13),C(3) O(3) 1.173(11),W Ge C(14)112.6(2). Scheme3.ClAbstractionfromChlorometallogermylenes 1band1c Figure3. (a)ORTEPdrawingofthedicationicpartof 3 withthermal ellipsoidsat50%probabilitylevel.Allhydrogenatomsareomittedfor clarity.(b)ANewmanprojectionalongtheGe(1) Ge(2)bondof 3 . JournaloftheAmericanChemicalSociety Communicationdx.doi.org/10.1021/ja506018f | J.Am.Chem.Soc. XXXX,XXX,XXX XXXB

PAGE 3

bentangle=30.5 ° ;twistangle=0 ° )12andthemoleculehasan inversioncenteratthemidpointoftheGe Gebond.TheGe Gebondlengthof 4 (2.345(2)Å)isdistinctlyshorterthanthat of 3 .Thesedi erencesareattributabletothestericrepulsion betweenthesubstiuentsontheNatomsofNHCsand/orthe metalfragments.Namely,morestericallyhindered 3 hasamore distortedandelongatedGe Gebondincomparisonwiththat of 4 .Interestingly,thestructureofthegemylenemoiety Cp * (CO)3WGe(NHC)in 3 closelyresemblesthatin 4 (Table 1).TheW Gebondlengthof 3 and 4 aremuchlongerthan thatof 2 (2.5780(10)Å).Thesebondelongationssuggestthat the back-donationfromWtoGein 3 and 4 issigni cantly weakerthanthatin 2 .TheGe Cbondlengthsof 3 (2.033(8) and2.053(8)Å)and 4 (2.029(11)Å)areshorterthanthoseof 1b (2.087(7)Å)11and 1c (2.083(10)Å),respectively.This impliesthatthepositivechargeof 3 and 4 ispartially distributedattheNHCrings.13AsshowninFigure5,tworesonancestructuresarepossible for 4 .Structure A isregardedasadimetallodigermene-1,2diyliumioninwhichavacantporbitalontheGeiscoordinated byMeIMe.Structure B isregardedasadicationic dimetallodigermene.Toshedlightonthevalidityoftwo resonancestructures,atheoreticalcalculationforamodel dicationiccomplex[{Cp(CO)3WGe(IMe)}2]2+( 4 )(IMe= 1,3-dimethylimidazol-2-ylidene)wascarriedoutattheB3LYP level.11ThelengthsofW Ge(2.611Å),Ge C(2.006Å),and Ge Ge(2.324Å)bondsintheoptimizedstructureagreewell withthoseinthecrystalstructureof 4 .Wibergbondindex (WBI)analysisindicatesthattheGe Gebondhasa considerabledoublebondcharacter(Ge Ge:1.432).The naturalpopulationanalysis(NPA)showsthatboththeGe atomandtheIMemoietyarehighlypositivelycharged(+0.65 and+0.44,respectively),whilethetungstenmoietyisslightly negativelycharged( 0.09).Theseresultscorroboratethatthe contributionsofresonancestructures A and B areequally signi cantin 4 . Insummary,wesynthesizedNHC-stabilizedchlorometallogermylenes 1a c asprecursors.Chlorideabstractionfromthe bulkiestNHC-stabilizedchlorogermylene 1a gaveacationic metallogermylene 2 .Ontheotherhand,chlorideabstraction fromlesshindered 1b and 1c ledtodimerizationaccompanied byformationofaGe Gebondtoprovidedicationic dimetallodigermenes 3 and 4 with Z -and E -con gurations, respectively.ThebulkinessofsubstituentsontheNatomsof NHCsisanimportantfactortocontrolnotonlythereaction routesbutalsothestereochemistryoftheproducts.ASSOCIATEDCONTENT*SSupportingInformationText, gures,tables,andCIF lesgivingsyntheticprocedures andcharacterizationdatafor 1 4 ,detailsofthecrystal structurere nementfor 1 4 ,detailsofthecalculationofa modeldicationcomplexof 4 ,andX-raycrystallographicdata for 1 4 .ThismaterialisavailablefreeofchargeviatheInternet athttp://pubs.acs.org.AUTHORINFORMATIONCorrespondingAuthortobita@m.tohoku.ac.jpNotesTheauthorsdeclarenocompeting nancialinterest.ACKNOWLEDGMENTSThisworkwassupportedbytheMinistryofEducation, Culture,Sports,ScienceandTechnology,Japan(Grants-in-Aid forScienti cResearchNos.22350024and23750054). Figure4. (a)ORTEPdrawingofthedicationicpartof 4 withthermal ellipsoidsat50%probabilitylevel.Allhydrogenatomsareomittedfor clarity.(b)ANewmanprojectionalongtheGe Ge bondof 4 . Table1.SelectedBondLengths(Å)andAngles(deg)fortheDicationicPartsof3and434 W(1) Ge(1)2.6386(9)W Ge2.6368(11) W(2) Ge(2)2.6511(8) Ge(1) C(14)2.033(8)Ge C(14)2.029(11) Ge(2) C(38)2.053(8) Ge(1) Ge(2)2.4274(11)Ge Ge 2.345(2) W(1) Ge(1) C(14)110.4(2)W Ge C(14)109.4(3) W(2) Ge(2) C(38)110.3(2) W(1) Ge(1) Ge(2)135.20(4)W Ge Ge 134.22(7) W(2) Ge(2) Ge(1)135.47(4) C(14) Ge(1) Ge(2)97.5(2)C(14) Ge Ge 104.2(3) C(38) Ge(2) Ge(1)97.9(2) Figure5. Twopossibleresonancestructuresfor 4 . JournaloftheAmericanChemicalSociety Communicationdx.doi.org/10.1021/ja506018f | J.Am.Chem.Soc. XXXX,XXX,XXX XXXC

PAGE 4

REFERENCES(1)(a)Tsumuraya,T.;Batcheller,S.A.;Masamune,S. Angew.Chem., Int.Ed.Engl. 1991 , 30 ,902.(b)Driess,M.;Gru tzmacher,H. Angew. Chem.,Int.Ed.Engl. 1996 , 35 ,828.(c)Power,P.P. J.Chem.Soc., DaltonTrans. 1998 ,2939.(d)Weidenbruch,M. Eur.J.Inorg.Chem. 1999 ,373.(e)Fischer,R.C.;Power,P.P. Chem.Rev. 2010 , 110 ,3877. (2)Goldberg,D.E.;Harris,D.H.;Lappert,M.F.;Thomas,K.M. J. Chem.Soc.,Chem.Commun. 1976 ,261. (3)Masamune,S.;Hanzawa,Y.;Williams,D.J. J.Am.Chem.Soc. 1982 , 104 ,6136. (4)Park,J.;Batcheller,S.A.;Masamune,S. J.Organomet.Chem. 1989 , 367 ,39. (5)Kira,M.;Iwamoto,T.;Maruyama,T.;Kabuto,C.;Sakurai,H. Organometallics 1996 , 15 ,3767. (6)(a)Rupar,P.A.;Staroverov,V.N.;Ragogna,P.J.;Baines,K.M. J.Am.Chem.Soc. 2007 , 129 ,15138.(b)Rupar,P.A.;Jennings,M.C.; Baines,K.M. Organometallics 2008 , 27 ,5043.(c)Rupar,P.A.; Staroverov,V.N.;Baines,K.M. Science 2008 , 322 ,1360.(d)Thimer, K.C.;Al-Rafia,S.M.I.;Ferguson,M.J.;McDonald,R.;Rivard,E. Chem.Commun. 2009 ,7119.(e)Sidiropoulos,A.;Jones,C.;Stasch,A.; Klein,S.;Frenking,G. Angew.Chem.,Int.Ed. 2009 , 48 ,9701. (f)Rupar,P.A.;Staroverov,V.N.;Baines,K.M. Organometallics 2010 , 29 ,4871.(g)Filippou,A.C.;Chernov,O.;Blom,B.;Stumpf,K.W.; Schnakenburg,G. Chem. Eur.J. 2010 , 16 ,2866.(h)Ruddy,A.J.; Rupar,P.A.;Bladek,K.J.;Allan,C.J.;Avery,J.C.;Baines,K.M. Organometallics 2010 , 29 ,1362.(i)Katir,N.;Matioszek,D.;Ladeira, S.;Escudie ,J.;Castel,A.Angew.Chem.,Int.Ed. 2011 , 50 ,5352. (j)Jana,A.;Huch,V.;Scheschkewitz,D. Angew.Chem.,Int.Ed. 2013 , 52 ,12179. (7)Tashita,S.;Watanabe,T.;Tobita,H. Chem.Lett. 2013 , 42 ,43. (8)Leung,W.-P.;Chiu,W.-K.;Thomas,C.W.M. Organometallics 2012 , 31 ,6966. (9)Thecloselyrelatedsiliconanalogue,[Cp * (CO)3CrSi(SIdipp)] (BAr4)(SIdipp=1,3-bis(2,6-diisopropylphenyl)imidazolidin-2-ylidene),hasbeensynthesizedbyreactionofacationicsilylynecomplex [Cp * (CO)2Cr Si(SIdipp)](BAr4)withCO.Filippou,A.C.;Baars, B.;Chernov,O.;Lebedev,Y.N.;Schnakenburg,G. Angew.Chem.,Int. Ed. 2014 , 53 ,565. (10)Anaverageofthevaluesfortwoindependentmoleculesinthe asymmetricunitcell. (11)SeetheSupportingInformation. (12)Inadigermene(R1)2Ge Ge(R2)2,the trans -bentangleis de nedasthetiltangleoftheR-Ge RplanetowardtheGe Ge bond.Thetwistangleisde nedasadihedralanglebetweentheR1 Ge R1andR2 Ge R2meanplanes. (13)In 1b and 1c ,theGeissaturated(sp3hybridized),whilein 3 and 4 ,theGeissp2hybridized.Assuggestedbyareviewer,this di erenceofhybridizationcouldalsoaccountfortheshorteningofthe Ge Cbond. JournaloftheAmericanChemicalSociety Communicationdx.doi.org/10.1021/ja506018f | J.Am.Chem.Soc. XXXX,XXX,XXX XXXD



PAGE 1

THE DESIGN, SYNTHESIS, AND APPLICATION OF VOLATILE ORGANOMETALLIC PRECURSORS FOR HETEROGENEOUS CATALYSIS, PHOTOACTIVATED CHEMICAL VAPOR DEPOSITION, AND ELECTRON BEAM INDUCED DEPOSITION By JOSEPH BRANNAKA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014

PAGE 2

© 2014 Joseph Brannaka

PAGE 3

To my loving wife for supporting me all the way

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my research advisor, Professor Lisa McElwee White for her constant guidance and support. She has given me guidance and direction in my research, while still allowing me the freedom to pursue my research ideas. Our conversations in her office have been intellectually stimulating and intensely useful both in research and other endeavors. I would like to thank my collaborators. First I would like to thank Dr. Jason Weaver from University of Florida Chemical Engineering department and his student Juhee Choi. I was able to learn a great deal from both a s we worked together on a research project. I would also like to thank Dr. Amy Walker from University of Texas Dallas and her students Zhiwei Shi and Jing Yang. Although we have not met in person, their contributions to our research project have been signi ficant. Their expertise in deposition methods and SAMs has greatly contributed to the enclosed work. I would like to thank Dr. Howard Fairbrother and his student Julie Spencer for their contributions to the EBID project. They have greatly elevated the valu e of this body of work. I also have to thank Dr. Howard Fairbrother for making the connection with Dr. Oddur Ingólfsson and his student Rachel Thorman at the University of Iceland who have made great contributions to this body of work. I would like to than k the members of my committee, Dr. Stephen Miller, Dr. Ronald Castellano, Dr. Jason Weaver, Dr. Leslie Murray, and Dr. Delores James for their guidance, support and willingness to listen to me talk. I would like to thank the members of my groups for their contributions to my scientific understanding and the stimulating conversations we had. They were always ready to proofread, examine my slides, and provide feedback on my research. I would

PAGE 5

5 like to especially thank Sarah Goforth, Jennifer Johns, Ciera Gerack , Kelsea Johnson, and Arijit Koley, for their help, guidance and friendship. I would like to thank my parents for raising me and believing in me. I would like to especially thank my Mom for fostering and developing in me a love for science. They supported me and encouraged me to follow my love for science into the career I have chosen. I would like to thank my wife Maira for supporting me and loving me during this process. She has provided the ear to vent to, and the hand to massage the weary back on a num ber of occasions. I would also like to thank my daughter Ella for sharing her Daddy with his dissertation. She provided the much needed break and encouragement by just a simple smile.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURE S ................................ ................................ ................................ ........ 10 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 16 Chemical Vapor Deposition ................................ ................................ .................... 16 Ligands Affecting Volatility ................................ ................................ ................ 17 Ligands Affecting Decomposition Temperature and Pathway .......................... 21 Precursors for Heterogeneous Catalysis ................................ ................................ 25 Precursors for Photoactivated Chemical Vapor Deposition ................................ .... 27 Precursors for Electron Beam Induced Deposition ................................ ................. 29 Summary of Precursor Design Parameters ................................ ............................ 31 2 DESIGN, SYNTHESIS, AND APPLICATION OF ORGANOMETALLIC PRECURSORS FOR HETEROGENEOUS CATALYSTS ................................ ...... 33 Heterogeneous Catalysis ................................ ................................ ........................ 33 Hete rogeneous Catalysts Used in Industry ................................ ...................... 33 Process of Forming Heterogeneous Catalysts ................................ ................. 34 Precursors for Heterogeneous Catalysts ................................ .......................... 35 Drawbacks to Heterogeneous Catalysts and How to Deal with Them .............. 37 Design ................................ ................................ ................................ ..................... 40 Synthesis of Organometallic Precursors ................................ ................................ . 42 Application ................................ ................................ ................................ .............. 49 Summary of Results ................................ ................................ ................................ 58 3 DESIGN, SYNTHESIS, AND APPLICATION OF ORGANOMETALLIC PRECURSORS FOR PHOTOACTIVATED CHEMICAL VAPOR DEPOSITION ON SELF ASSEMBLED MONOLAYERS ................................ ............................... 60 Photoactivated Chemical Vapor Deposition ................................ ............................ 60 Organic Thin Films ................................ ................................ ........................... 6 0 Metallization Processes ................................ ................................ .................... 61 Phys ical Vapor Deposition ................................ ................................ ......... 62 Lift Off Method ................................ ................................ ........................... 65 Chemical Vapor Deposition ................................ ................................ ........ 65 Design ................................ ................................ ................................ ..................... 70

PAGE 7

7 Synthesis of Organomet allic Precursors for PACVD on SAMs ............................... 72 Application of Volatile Ruthenium Compounds for PACVD on SAMs ..................... 76 Summary of Results from PACVD on SAMs ................................ ........................... 87 4 DESIGN, SYNTHESIS, AND APPLICATION OF ORGANOMETALLIC PRECURSORS FOR ELECTRON BEAM INDUCED DEPOSITION ...................... 90 Electron Beam Induced Deposition ................................ ................................ ......... 90 Current Density ................................ ................................ ................................ 91 Electron Energy ................................ ................................ ................................ 92 Scan Pattern ................................ ................................ ................................ ..... 92 Writing Directio n ................................ ................................ ............................... 93 Post Deposition Processing ................................ ................................ ............. 93 Extra Irradiation ................................ ................................ .......................... 93 Extra Irradiation with Reactive Gas ................................ ............................ 94 Annealing ................................ ................................ ................................ ... 95 Reactive Gases ................................ ................................ ................................ 96 Precursor Identity ................................ ................................ ............................. 96 Design of Precursors for Electron Beam Induced Deposition ............................... 100 Synthesis of Precursors for Electron Beam Induced Deposition ........................... 108 Surface Science Studies of Precursors ................................ ................................ . 110 Summary of EBID Results ................................ ................................ .................... 118 5 EXPERIMENTAL ................................ ................................ ................................ .. 120 General Procedures ................................ ................................ .............................. 120 Synthesis ................................ ................................ ................................ .............. 120 Synthesis of 2,2' bipyridine N oxide ( 2 5 ) ................................ ....................... 120 Synthesis of 6 cyano 2,2' bipyridine ( 2 6 ) ................................ ...................... 121 Synthesis of 6 methoxycarbonyl 2,2' bipyridine ( 2 7 ) ................................ ..... 121 Synthesis of 6 hydroxymethyl 2,2' bipyridine ( 2 8 ) ................................ ......... 122 Synthesis of 2 9 ................................ ................................ .............................. 122 Synthesis of 2 10 ................................ ................................ ............................ 123 Synthesis of 2 11 ................................ ................................ ............................ 123 Synthesis of 2 14 ................................ ................................ ............................ 124 Synthesis of 2 15 ................................ ................................ ............................ 124 Synthesis of 2 18 ................................ ................................ ............................ 124 Synthesis of 2 1a ................................ ................................ ............................ 125 Synthesis of [RuCp(CO) 2 ] 2 ( 2 20 ) ................................ ................................ ... 125 Synthesis of CpRu(CO) 2 Me ( 2 2 ) ................................ ................................ ... 126 Synthesis of [Cp*RuCl 2 ] 2 ( 2 22 ) ................................ ................................ ...... 126 Synthesis of (C 5 Me 4 CH 2 Cl)Ru(CO) 2 Cl ( 2 23 ) ................................ ................. 127 Synthesis of (C 5 Me 4 CH 2 OH)Ru(CO) 2 Cl ( 2 3 ) ................................ ................. 127 Synthesis of 3 C 3 H 5 Ru(CO) 3 Br ( 4 1c ) ................................ ........................... 127 Synthesis of 3 C 3 H 5 Ru(CO) 3 Cl ( 4 1b ) ................................ ........................... 128 Synthesis of 3 C 3 H 5 Ru(CO) 3 I ( 4 1d ) ................................ ............................. 128 Synthesis of 3 C 4 H 7 Ru(CO) 3 Cl ( 4 1e ) ................................ ........................... 129

PAGE 8

8 Synthesis of 3 C 4 H 7 Ru(CO) 3 Br ( 4 1f ) ................................ ............................ 129 Synthesis of 3 C 4 H 7 Ru(CO) 3 Cl ( 4 1g ) ................................ ........................... 129 Synthesis of 3 C 4 H 7 Ru(CO) 3 Br ( 4 1h ) ................................ ........................... 130 LIST OF REFERENCES ................................ ................................ ............................. 131 BIOGRAPH ICAL SKETCH ................................ ................................ .......................... 140

PAGE 9

9 LIST OF TABLES Table page 1 1 Selected precursors and the comparison of materials deposited by EBID and CVD. ................................ ................................ ................................ ................... 30 1 2 The design parameters for the desired uses of the precursors. .......................... 32 2 1 Possible identities of the desorbed species from Figure 2 12. ........................... 52 3 1 m/z values for ruthenium containing ions from SIMS of 3 1 on MHA. ................. 82 4 1 Commonly used EBID precursors.. ................................ ................................ .... 97 4 2 Precursors synthesized for EBID and their volatility. ................................ ........ 108

PAGE 10

10 LIST OF FIGURES Figure page 1 1 Common ligands for CVD of metal oxides . ................................ ......................... 18 1 2 Common ruthenocene s used to test for volatility ................................ ................ 20 1 3 Sodium and zirconium fluoroalkoxides used to determine volatility in Reference 10. ................................ ................................ ................................ ..... 21 1 4 Thermolysis of zirconium containing compound 1 26 . ................................ ........ 23 1 5 Decomposit ion pathway of tungsten nitride precursor 1 33 . ............................... 24 1 6 Synthesis of CVD precursor 1 37 . ................................ ................................ ...... 25 1 7 Stepwise mechanism of PACVD on S AMs . ................................ ........................ 28 2 1 Complexes proposed for use in preparing heterogeneous catalysts. ................. 41 2 2 Synthesis and protection of the ligand. ................................ ............................... 43 2 3 Synthesis of the model complex. ................................ ................................ ........ 44 2 4 Attempted synthesis of 2 14 . ................................ ................................ .............. 45 2 5 Transmetallation investigations conducted using boron based reagents. ........... 46 2 6 The synthesis of protected complex 2 14 . ................................ .......................... 46 2 7 Alternate synthesis of unprotected complex 2 1a . ................................ .............. 47 2 8 The synthesis of volat ile complex 2 2 . ................................ ................................ 48 2 9 The synthesis of the hydroxyl derivative 2 3 . ................................ ...................... 49 2 10 The mechanism of the formation of complex 2 23 ................................ .............. 49 2 11 TGA of compound 2 1a . ................................ ................................ ..................... 50 2 12 The TPD measurements of compound 2 2 after dosing for 5 minutes. .............. 51 2 13 TPD of compound 2 2 after being exposed to the STO surface for 15 seconds to 5 minutes. ................................ ................................ ......................... 53 2 14 XPS of clean STO, dosed with precursor, and heated to 700 °C. ...................... 54 2 15 XPS of the Ru 3d and Sr 3p region for dosed and heated 2 2 on STO. ............. 54

PAGE 11

11 2 16 High resolution XPS of Ru 3p and Ti 2p region of 2 2 on STO. .......................... 55 2 17 ISS of compound 2 2 deposited on STO and heated to 700 ° C. ........................ 56 2 18 STM image (500Å x 500Å) of clean STO(12 2 1) annealed at 1275 K for 30 minutes. ................................ ................................ ................................ .............. 57 2 19 STM image (1000Å x 1000Å) of STO(12 2 1) dosed with 2 2 for 45 seconds at 1.1 x 10 9 Torr. ................................ ................................ ................................ 57 3 1 The process of forming SAMs and the structure and interactions once formed . ................................ ................................ ................................ ............... 61 3 2 Molecular electronic device structure. Overall structure (top) and detailed view of monola yer (bottom) ................................ ................................ ................ 62 3 3 The lift off process i llustrated ................................ ................................ .............. 66 3 4 TOF SIMS of patterned SAMs after PACVD. Detection of Al+, AlO , and Au 2 HDT recorded ................................ ................................ .............................. 70 3 5 Proposed precursors for PACVD on SAMs. ................................ ....................... 71 3 6 The mechanism of loss of Cp as ................... 72 3 7 Synthesis of volatile compound 3 1 . ................................ ................................ ... 72 3 8 The synthesis of compound 3 2 . ................................ ................................ ......... 73 3 9 Attempted replacement of bromide on 3 2 to form compound 3 5 . ..................... 74 3 10 Synthesis of compound 3 6 . ................................ ................................ ............... 74 3 11 Possible reductive elimination products from 3 6 . ................................ .............. 75 3 12 Attempted synthesis of compound 3 1 2 . ................................ ............................ 75 3 13 The experimental set up for the deposition of the ruthenium precursor. ............. 76 3 14 XPS of bare MHA SAM: a) gold 4f, b) carbon 1s and ruthenium 3d, c) oxygen 1s, and d) ruthenium 3p. ................................ ................................ ..................... 77 3 15 XPS of MHA after dosing with compound 3 1 : a) gold 4f, b) carbon 1s, c) ruthenium 3d, and d) ruthenium 3 p regions ................................ ........................ 78 3 16 XPS of MHA after dosing with 3 1 and rinsing with water and ethanol. .............. 78 3 17 XPS of clean MHL SAM on gold substrate: a) gold 4f, b) carbon 1s, c) oxygen 1s, and d) ruthenium 3p regions. ................................ ........................... 79

PAGE 12

12 3 18 XPS of MHL SAM after dosing with ruthenium precursor for 5 minutes: a) gold 4f, b) carbon 1s, c) ruthenium 3d, and d) ruthenium 3 p regions ................. 80 3 19 XPS of MHL SAM after dosing by ruthenium precursor and rinsing with ethanol and water: a) gold 4f, b) carbon 1s, c) ruthenium 3d, and d ) ruthenium 3p regions. ................................ ................................ ......................... 80 3 20 SIMS of ruthenium precursor dosed on MHA: a) ruthenium ion region and b) ruthenium cyclopentadienyl ion region. ................................ .............................. 81 3 21 SIMS of 3 1 dosed on MHL: a) ruthenium ions and b) ruthenium cyclopentadienyl ions. ................................ ................................ ........................ 82 3 22 XPS of 3 2 deposited on MHA (carboxyl t erminated) SAM. ............................... 83 3 23 TOF SIMS of 3 2 deposited on MHA: a) ruthenium ion region and b) ruthenium allyl ion region. ................................ ................................ ................... 85 3 24 XPS of 3 2 deposited on a) MHL and b) HDT. ................................ ................... 86 3 25 TOF SIMS of the bromide region of 3 2 deposited on a) MHA and b) MHL. ...... 87 3 26 Proposed mechanism for PACV D of 3 1 . ................................ ........................... 89 4 1 Treatment of ruthenium carbide deposition with elec tron beam and oxygen gas ................................ ................................ ................................ ...................... 95 4 2 The typical EBID expe riment. . ................................ ................................ .......... 101 4 3 The surface science approach to understanding the EBID decomposition pa thway ................................ ................................ ................................ ............ 102 4 4 The carbon:platinum ratio of MeCpPtMe 3 deposited by EBID as a function of electron radiation determined by XPS. ................................ ............................. 103 4 5 Mass spectrometry results from a) gas phase compound 4 2 , b) irradiated with electron beam, c) on the surface being radiated with electron beam, and d) reference spectrum of methane. ................................ ................................ ... 103 4 6 The mechanism of loss of Cp as ................. 104 4 7 Mass spectrum of volatile species from electron beam irradiation of tungsten hexacarbonyl. ................................ ................................ ................................ ... 105 4 8 Effect of electron beam on adsorbed tungsten hexacarbonyl. .......................... 106 4 9 The deposition and post deposition processing of platinum deposits from 4 3 . 106 4 10 Proposed EBID precursors. ................................ ................................ .............. 107

PAGE 13

13 4 11 Precursors synthesized for EBID. ................................ ................................ ..... 108 4 12 Synthesis of compounds 4 1b through 4 1e . ................................ .................... 108 4 13 Synthesis of 4 1a from allyl bromide and triruthenium dodecacarbonyl. ........... 109 4 14 Synthesis of 4 1f . ................................ ................................ .............................. 109 4 15 Synthesis of compound 4 1g . ................................ ................................ ........... 110 4 16 Literature synthesis of 4 1g . ................................ ................................ ............. 110 4 17 Synthesis of compound 4 1h . ................................ ................................ ........... 110 4 18 Mass spectrum of gas phase 4 1c . ................................ ................................ ... 112 4 19 The mass spectrum of volatilized species from electron beam irradiation of 4 1c on HOPG surface. ................................ ................................ ....................... 112 4 20 XPS data of 4 1c on HOP G with increasing electron irradiation dosage. ......... 113 4 21 Changes in the fractional coverage of adsorbed oxygen and bromine atoms a nd changes in Ru 3d 5/2 peak position for 1 2 nm 4 1c films plotted as a function of electron dose . ................................ ................................ ................. 114 4 22 Post deposition removal of bromine via extra irradiation as determined by XPS peak height. ................................ ................................ .............................. 115 4 23 Mass spectrum of irradiated 4 1b on HOPG. ................................ .................... 115 4 24 Ruthe nium peak position and oxygen peak area measured by XPS as a function of irradiation time. ................................ ................................ ............... 116 4 25 Chlorine peak area for XPS of 4 1b deposited on HOPG as a function of irradiation time. ................................ ................................ ................................ . 117 4 26 Chlorine area from XPS of 4 1b deposited on HOPG plotted as a function of electron dose, indicating loss of chlorine upon post deposition irradiation. ...... 117

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE DESIGN, SYNTHESIS, AND APPLICATION OF VOLATILE ORGANOMETALLIC PRECURSORS FOR HETEROGENEOUS CATALYSIS, PHOTOACTIVATED CHEMICAL VAPOR DEPOSITION, AND ELECTRON BEAM INDUCED DEPOSITION By Joseph Brannaka December 2014 Chair: Lisa McElwee Whit e Major: Chemistry The design and synthesis of organometallic precursors has proven very useful in the chemical vapor deposition (CVD) field. By taking a mechanistic approach the deposition temperatures have been lowered and cleaner deposits have been for med. A similar approach will be key in improving three additional fields: heterogeneous catalysis, photoactivated chemical vapor deposition (PACVD), and electron beam induced deposition (EBID). Precursors were designed and synthesized with hydroxyl contai ning ligands to di rect the precursor to the undercoordinated sites of the oxide surface for a model study on heterogeneous catalysis. Unfortunately the hydroxyl containing precursors decompo sed before sublimation and thus were not able to be delivered to the surface. A volatile precursor that did not contain the desired functional group was synthesized and deposited in a largely random manner. Designed precursors for PACVD on self assembled monolayers yielded selective deposition of ruthenium; however the cyclopentadienyl and allyl ligands proved difficult

PAGE 15

15 to remove under these conditions. The acidity of the terminal groups of the substrate proved important to the deposition process. The use of designed ruthenium precursors has given insight into the mecha nism of EBID. Allyl ligands contribute to carbon contamination in the deposit while carbonyl ligands dissociate cleanly under electron irr adiation. Halogens were not removed during deposition but have been easily removed by post deposition processing. Thes e contributions will help the des ign of better precursors for all three applications in the future.

PAGE 16

16 CHAPTER 1 INTRODUCTION The research in this dissertation involves the design of precursors for three applications: heterogeneous catalysis, photoactivated chemical vapor deposition, and electron beam induced deposition. Because v ery little work has been done in the design of pr ecursors for these applications, p recursor design for chemical vapor deposition ( CVD ) , which has seen extensive work, will be used as a starting point . First the CVD process will be introduced and then a brief overview of the precursor design will be discussed. These design rules will then be adapted to the precursors for heterogeneous catalysis, photoactivated chemical vapor deposition, and electron beam induced deposition. Chemical Vapor Deposition The CVD process has four steps: 1 1. The precursor is carried in the gas phase to the substrate, during which time it can react wi th one or more of the carrier gases 2. The precursor is adsorbed onto the substrate 3. Energy is directed to the precursor on the substrate (usually thermal, but plasma and light are known as well ) which facilitates the decomposition of the precursor and the for mation of the thin fi lm 4. The byp roducts are desorbed and removed in the gas phase CVD offers a significant advantage over physical deposition processes such as sputtering due to its conformal coverage. 2 Physical deposi tion methods generally involve line of sight deposition, which can cause poor conformality on high aspect ratio features in devices . Because of this , CVD has become an attractive alternative to physical methods for the electronic industry.

PAGE 17

17 CVD precursors need to be highly volatile to reach the substrate in significant quantities in the gas phase. 1 , 2 An alternative to this is liquid injection CVD and aerosol assisted CVD (AACVD) , in which the precursor solution is dispersed and carried to the substrate in droplets. For these methods , the precursor must be soluble in inert solvents and stable to the dispersion conditions. The other primary requirement for CVD precursors is that they should react in a clean and selective manner to deposit th e desired compound (sometimes the pure metal and sometimes an oxide or similar compound) on the substrate (and later , the already deposited film) and not the reactor walls. Predicting the decomposition pathway fo r specific precursors can be difficult due t o the difference between the deposition conditions (>300 °C) and the conditions generally studied for decomposition (near room temperature ) . 2 Ligands A ffecting Volatility The volatility of the precursor can be tuned by changing the ligands attached to the metal. Hubert Pf alzgraf and Guillon report common ligands used to increase the vapor pressure of precursors for CVD of metal oxides (Figure 1 1). 3 diketonate ligands ( 1 1 ) are very common due to the volatility of their complexes . 4 This ligand provides G roup 2 and 3 metals with nearly the only known volatile species. For alkali diketonates , the volatility generally decreases with increased mass and ionic radius. Fluorinated derivatives (such as hfac and tfac) have shown increased volatility, however , there is the possibility of including fluorine in the deposited film. The ketoesterate ligand ( 1 2 ) is a less commonly used de diketonate that can be used in precursors for metal oxides. 5 These are known to sublime a t low temperatures and are used to deposit hafnium and iron oxides. 6 ketoimin ate

PAGE 18

18 ligand ( 1 3 ) is another well known derivative that provides volatile compound s useful for CVD. 7 Figure 1 1 . Common ligands for CVD of metal oxides. Adapted from Scheme 1 on page 2 23 of R eference 3 . T he N O N and N O donors ( 1 4 through 1 7 ) have also provided relatively stable, volatile precursors (Figure 1 1) . 3 diketonates and derivatives can protect the metal from attack by sat urating the coordination sphere, which preve nts olig omer formation and raise s volatility . One strategy for increasing the volatility of a precursor is to break up the crystal packing by adding long chain alkyls and heteroleptic ligands. 2 , 3 The problem with adding long or bulky alkyl chains is this also increases the molecular weight, which increases

PAGE 19

19 the dispersion forces. Thus , sometimes adding increased bulk can result in less volatile species. Siddiqi et al. report ed volatility studies on common ruthenocenes that can be used for CVD (Figure 1 2 ) ; this study illustrates some design rules for improving volatility . 8 Isothermal thermogravimetry was co nducted on these complexes and all but compounds 1 1 5 , 1 1 7 , and 1 1 8 sublimed cleanly , leaving little to no residue. Compounds 1 15 , 1 17 , and 1 18 decomposed before sublimation , most likely due to their strong intermolecular forces from the carboxyl ic ac id for 1 17 and 1 18 and the asymmetric ketone for 1 15 . Compounds 1 1 1 and 1 1 3 sublimed cleanly at 95 °C with 1 1 3 taking more time , as would be expected with the stronger intermolecular for ces from the amine group. Compound 1 1 6 required a temperature o f 144 °C and took considerably longer than any of the previous compounds. This is due to its high symmetry and strong intermolecular forces from the ketones. Compounds 1 1 9 and 1 20 required 143 °C to sublime with 1 1 9 taking longer than 1 20 ; both compounds have broken the symmetry from 1 1 6 and added a bulky alkyl group to further break up the crystal packing. However , the added mass seemed to balance out the e ffect, resulting in very similar volatility to 1 1 6 . Compound 1 1 9 is less volatile than 1 20 due to the added mass. Compound 1 2 1 sublimed at 90 °C and at a faster rate than 1 19 or 1 20 .

PAGE 20

20 Figure 1 2 . Common ruthenocenes used to test for volatility . Adapted from Figure 1 on page 1173 of R eference 8 . Based on the high volatility of the silyl contain ing complex, more silylated rut henocenes have been synthesized and tested in CVD. 9 In general , more silyls increase the volatility and thus lower the sublimation temperature. Adding fluorine to the ligands is a nother method to increase the volatility of an organometallic complex due to the low polarizability of the fluorine . Samuels et al. report a study on the effect of structure on volatility of zirconium and sodium flu o roalkoxides (Figure 1 3 ) , which provide s a good case study for the effect of fluorine on volatility. 10 The volatility was determined by TGA and the sodium compounds were found to have relative volat ilities of 1 22 b < 1 22 c < 1 22 a < 1 22 d . This shows that with increased fluorination there is increased volatility. The zirconium comple xes have a slightly different trend; their order of volatility was 1 23 b < 1 23 c < 1 23 d < 1 23 a . This illustrates

PAGE 21

21 that generally, fluorinating ligands can lead to higher volatility, but there can be exception s to the rule. Figure 1 3 . Sodium and zirconium fluoroalkoxides used to determine volatility in R eference 10 . Li gands A ffecting Decomposition Temperature and Pathway For many years , the emphasis was simply on making more volatile precursors; however there is another very important factor: the decomposition temperature and pathway. If the temperature can be lowered , thermally sensitive substrate s can be employed . If the pathway can be understood better, the deposition composition can be tuned appropriately to deposit the desired compound. Some selected case studies will be examined to observe th e evolution of precursor design based on a mechanistic understanding. The precursor that was commonly used to deposit ZrO 2 was Zr(O R ) 4 (R = i Pr, t Bu) , compound 1 24 , due to its volatility from the sterically bulky alkyl groups forcing the complex to be monomeric. The drawback to this precursor is that it reacts with the trace oxygen and w ater present in the CVD reactor due to the unfilled coordination sphere . 11 This precursor can be stabilized to contact with water and air by adding one or more bidentate ligands , diketonate, to combine the stability to ambient water of diketonates with the lower thermal stab ility and high vapor pressure of the

PAGE 22

22 alkoxides. The complexes of Zr(OR) 2 (thd) 2 (thd = 2,2,6,6 tetra methylheptane 3,5 dionate) showed increased stability to water, but also an increased thermal stability and decreased volatility. 12 Another method to fill the coordination sphere is t o tether a weak electron donor to the alkoxide group. An example of this is Zr(O t Bu) 2 (mmp) 2 (mmp = 1 methoxy 2 methyl 2 propanolate) which has proven be more stable to moisture and have comparable volatility and decomposition temperature to the a l koxide. Thermolysis studies can give an increased understanding of the decomposition pathway of potential CVD precursors and thus help design better pr ecursors. Zechman et al. reports a potential zirconium CVD precursor and its thermolysis studies (Figure 1 4). 13 The decomposition likely followed a similar p athway to the pinacol rearrangement in organic chemistry. The organic form begins with proton transfer and then loss of the hydroxyl, but in this form , the two zirconium at oms bonded to the bridging oxygen take the place of proton transfer. Thus the carbon oxygen bond is broken, followed by a methyl shift to eventually form compound 1 27 . Compound 1 28 hydride elimination, with com pound 1 29 acting as the base . This precursor turned out to not be useful due to its low volatility, yet it illustrates how known organic reactions can be used to design precursors that selectively decompose to the desired compound, in this case ZrO 2 .

PAGE 23

23 Figure 1 4 . Thermolysis of zirconium containing compound 1 26 . 13 Adapted from Scheme 1 on page 5329 of R eference 2 . The CVD of metal nitrides provides some good examples of the mechanistic approach to precursor design. Tungsten nitride and tungsten carbonitride are attractive diffusion barriers in t he electronic industry and thus much work has been done to deposit these via CVD. 14 Bchir et al. report the deposition of Cl 4 (RCN)W(NR') (R = Me, Ph, R' = allyl, iPr, Ph) , compound 1 30 , to form tungsten nitride and tungsten carbonitride. 15 There was a direct rel ationship between the bond dis sociation energy of the C N bond and the acti vation energy of the deposition, which indicates that the rate limiting step of the deposition is b reaking the C N bond. This provides design rules for further precur sors to either weaken this bond or to remove it entirely.

PAGE 24

24 The McElwee White group has used this information to design CVD precursors that decompose at lower temperatures. Compound 1 3 1 was synthesized as a precursor for CVD due to the weaker N N bo nd replacing the C N bond of 1 30 . 16 Decomposition and computational studies revealed the d ecomposition pathway (Figure 1 5 ) . Figure 1 5 . Decomposition pathway o f tungsten nitride precursor 1 33 . The decomposition could take two pathways, either breaking the C N bond or the N N bond. The products from the C N bond breaking were not observed during in situ Raman spectroscopy experiments and DFT calculations showed this pathway to be higher in energy. This precursor was found to deposit films of WN x C y at 300 °C, while 1 30 deposited material at 450 °C. 14 The deposition temperature was lowered significantly by weakening the bond; the next step was to re move the bond entirely. Compound 1 37 , which has no N C (imido) or N N (hydrazido ) bond to break, was synthesized from WCl 6 (Figure 1 6 ). 17 This precursor was found to deposit films at 125 °C. 18

PAGE 25

25 Figure 1 6 . Synthesis of CVD precursor 1 37 . Based on these results , t he same type of mechanistic based precursor design c ould be applied to other deposition techniq u es . Much work has been done on precursor design for conventional CVD but very little has been done on similar techniques. First , the techniques will be introduced and then the design of precursors for that technique will be discussed. Precursors for Heterogeneous Catalysis The formation of an oxide supported heterogeneous catalyst has four steps: 19 1. The oxide surface is pretreated to remove unwanted species adsorbed on the surface 2. The precursor is exposed and anchored to the support 3. The unanchored precursor s are removed 4. The ligands still attached to the precursor are removed The precursor and the support must be chosen with particular care to make sure that upon exposure to the support the precursor can react w ith it. For example, if metal chl orides are used , the surface can have hydroxyls . 20 The hydroxyls react with the chlorides to form hydrogen chloride and the oxygen now binds with the metal in the

PAGE 26

26 empty coordin ation site left by the chloride. Metal alkoxides also require a protonated surface to remove the alkoxide and anchor the precursor to the support. 21 The remaining alkoxides can be rem oved by heat via dehydrogenation reactions. The research discussed in this dis sertation involves reversing the conventional roles of the support and the precursor. The standard route involves the attack of a hydr oxyl from the oxide support at the metal of the precursor, while the route discussed here involves a hydroxyl on the precursor attacking an exposed metal of the oxide support. The exposed metals of t he oxide support occur at undercoordinated sites such as step and kink sites. The step and kink site s are much more active for catalysis 22 and this method would theoretically make more well defined and active heterogeneous catalysts. Precursors for conventional synthesis of heterogeneous catalysts need to be vola tile and possess ligands that can be removed by heat and/or reaction with surface hydroxyls. 19 These precursors can be treated with various gases upon anchoring to help facilitate the loss of the ligand. In many wa ys , the requirements for these precursors are very similar to the CVD precursor requirements: both need to be volatile and the ligands can be removed in similar manners. The precursors for the research described in this dissertation (Chapter 2) have differ ent requirements than conventional synthesis of heterogeneous catalysts. These precursors still need to be volatile, but they also need to contain a hydroxyl functional group. The other ligands would ideally be relatively unreactive to facilitate the react ion of the hydroxyl with the surface.

PAGE 27

27 Precursors for Photoactivated Chemical Vapor Deposition Photoactivated chemical vapor deposition, as the name implies, is similar to conventional CVD except the precursor is activated by light rather than heat. Conven tional CVD involves the precursor reaching the substrate in the gas phase and then decomposing on the substrate via thermal means to form the desired deposit. For photoactivated chemical vapor deposition , the decomposition is initiated by photolysis in the gas phase rather than by heat. Photoactivated chemical vapor deposition (PACVD) f alls into two categories: pyroly t i c and photolytic deposition. 23 The method that is of interest for this research is the photolytic deposition of organometallic precursors on self assembled m onolayers. This is a four step process (Figure 1 7 ): 1. The precursor is volatilized and exposed to UV light which removes one ligand 2. The SAM interacts with the empty coordination site of the precursor 3. Possible removal of other ligands by the terminal groups 4. The remaining ligands are removed under mild thermal conditions Based on this process, the precu rsors have several requirements : volatility, one or more photo labile ligand, protically sensitive ligands, and thermally labile ligands. Therefore these precur sors have some properties in common with conventional CVD precursors, namely , volatility and thermally labile ligands. However, these ligands must be much more sensitive to heat because the temperatures attained will be far less than conventional CVD. 23 One o r more ligand s need to be photo labile to form an empty coordination site upon UV irradiation. For this to occur in any significant amount , the quantum yield for the loss must be high. Common ligands that fit this criteria are carbonyls and nitrogen donors. It could be useful, though not necessarily required , for some of the ligands to be removed by interaction with the protic SAM. Many ligands are

PAGE 28

28 known to do this, though the most likely ones are hydrides and alkyls. Halides are known to be removed by hydroxyls as discu ssed above , but usually require high temperatures or more acidic groups. Figure 1 7. Stepwise mechanism of PACVD on SAMs. 1) Ligand is removed in gas phase by UV light, 2) the SAM binds to the empty coordination site of the precursor, 3) the protic terminal groups remove one or more ligands, and 4 ) the remaining ligands are removed via mild thermal conditions. For designing these precursors , several things need to be considered. First the precursor needs to be volatile, but many of the ligands discussed above that facilitate volatility are not easily removed. These ligands can be removed under standard CVD con ditions which include high temperature and reactive gases, but those condit ions are not compatible with SAM s . The ligands need to help volatility while still being easily removed. The oxidation state of the metal is also important. The precursor should have the metal at the (0 ) oxidation state or a clear pathway to reduce the metal through reductive elimination of certain ligands or reaction with the SAM.

PAGE 29

29 Precursors for Electron Beam Induced Deposition Electron beam induced deposition (EBID) is the process of depositing metal b y electron bombardment of a gas phase precursor molecule. 24 The precursor is volati lized under vacuum and directed to the substrate via a gas nozzle. The electron beam is directed th rough the flux of precursor to deposit the metal on the substrate with size and shape control at the nanoscale . This process has extensive application in lithographic mask repair. 25 The precursors used for this process have very dif ferent requirements than CV D precursors, however, CVD precursors are currently being used for EBID. 26 The use of CVD precursors has led to significa nt contamination in the deposit because CVD precursors decompose through thermal processes rather than electr onic flux . Some selected precursors and their results in CVD and EBID show th e di fference in mechanism (Table 1 1 ). While EBID precursors do need to be volatile , like CVD precursors , their reactivity needs to be different. Au(acac)Me 2 is a well known gold CVD precursor th at gives nearly pure gold films. H owev er , when it is used for EBID , the atomic percent of metal in the deposit is less than 20%. Very similar results are found for MeCpPtMe 3 . Deposits are nearly pure metal for CVD and less than 20% Pt for EBID. Ruthenium and tungsten precursors, Ru(EtCp) 2 and W(CO) 6 , respectively, give even worse results for EBID at 10% metal in the deposits . This illustrates that using CVD precursors for EBID is not ideal. By examining precursors that have given promising results, design parameters for EBID precursors can be determined. Tungsten hexafluori de has given 8 0 100% metal deposits , 27 which indicates that flu oride, and per haps other halide s, can be

PAGE 30

30 acceptable ligands for EBID precursors. These can be removed by the EBID process or by post deposition irradiation. Table 1 1. Selected precursors and the comparison of materials deposited by EBID and CVD. Precursor Percentage of metal in deposit for CVD Percentage of metal in deposit for EBID References Au(acac)Me 2 100% <20% 26 MeCpPtMe 3 100% <20% 26 Ru(EtCp) 2 100% 10% 28 , 29 W(CO) 6 >95% 10% 24 , 30 Another ligand that has shown promising results is carb onyl. 31 Tungsten hexacarbonyl gives poor deposits, but on closer inspection , more than two of the carbonyls are volatilized upon irradiation , with the remaining carbonyl incorporated into the deposit. This shows that a sm all number of carb onyl ligands can be good , but too many can cause contamination. Alkyl groups can be useful in EBID precursors. 32 The MeCpPtMe 3 precursor loses one of the methyl groups prior to deposition. The other two methyl groups and the cyclopentadienyl group are incorporated in to the deposit. This illustrates that alkyl groups can be acceptable ligands in EBID precursors. Sev eral considerations must be taken into account when designing EBID precursors. First, the precursors need to be volatile. However, many of the groups that commonly cause volatility in CVD precursors are not suitable due to their incorporation into the deposit in EBID. Ligands must be chosen that will decompose into volatile species upon electron bombardment. In addition, these precursors would ideally be simple monodentate ligands. Many precursors are volatilized by adding bulky groups but this is not viable for these precursors. The ligands need to be given the optimal ability to leave and therefore

PAGE 31

31 s not important due to the high flux of electrons and their ability to reduce the metal. Summary of Precursor Design Parameters Common procedure in all three fields is to use precursors that are commercially available. In the few cases where they are not, the precursor is known from the CVD literature. This gives a good starting point, considering volatile compounds are needed. However, the decomposition mechanism in all three cases is not th e same as conventional CVD. To gain a better understanding of this mechanism, researchers must move away from the current mentality of commercially available and CVD precursors and move toward precursors designed specifically for these respective uses. Onc e a better understanding of the mechanism is achieved, even better precursors can be designed that will give much better results than the currently used ones. Each of these uses requires different properties of the organometallic complex (Table 1 2 ) . For t he heterogeneous catalyst precursor, the complex must be volatile. The ligands mu st be removable under mild thermal conditions or reaction with the surface oxides. And prefer ably one or more of these ligands would include a hydroxyl or amine functional gro up. Metals should also be employed that would yield useful catalysts. For PACVD , the complexes would need to be volatile as well. One or more of the ligands should be removable under UV light . For use in patterning, t he metal should ideally show a preferen ce for hydroxyl or carboxyl terminated SAMs and have little to no reactivity with methyl terminated SAMs , which could be achieved by having a ligand that reacts with the protic surface .

PAGE 32

32 For EBID , the complex needs to be volatile. The ligands need to be re m ovable by electron beam and form volatile components. These ligands are typically simple monodentate ligands such as alkyls and carbonyls. Table 1 2 . The design parameters for the desired uses of the precursors. Application Volatile Ligands Removable by: Other Features Heterogeneous Catalyst Yes Mild thermal conditions Hydroxyl or amine PACVD Yes UV light and protic surface followed by mild thermal conditions EBID Yes Electron beam

PAGE 33

33 CHAPTER 2 DESIGN , SY N THESIS , AND APPLICATION OF ORGANOMETALLIC PRECURSORS FOR HETEROGEN E OUS CATALYSTS 1 Heterogeneous Catalysis Catalysis is a very important process for the manufacturing of goods, due to its economic, environmental, and synthetic advantages over non catalyzed reactions. Each year, the value of goods produced that used a catalytic process during their formation is $2.4 trillion. 33 Approximately 20% of all the goods manufactured in the United States of America are produ ced using a catalytic process. In the year 2000 , it was estimated that the value of the heterogeneous catalysts on the market was $6.5 billion. Heterogeneous catalysis occurs when the cat alyst is in a different phase from the reaction mixture. This typical ly involves the reaction in the liquid or vapor phase with the catalyst in the solid phase. Homogeneous catalysis has the catalyst in the same phase as the reaction. Hetero geneous catalysts offer five advantages over homogeneous catalysts : 34 1. Separation and recovery of the catalyst are easy. 2. Heterogeneous catalysts are extremely durable. 3. Heterogene ous catalysts are generally more active than homogeneous catalysts. 4. The catalyst is stable to a wide range o f temperatures, which can be especially useful for high temperature industrial reactions. 5. They can be easily handled for large scale operations. Heterogeneous Catalysts Used in Industry Because of these advantages, h eterogeneous catalysis is used widely in industry. Some of the common industrial processes that use heterogeneous catalysis are: the Haber Bosch process, the Ostwald pro cess, and the Contact process. 1 Surface science work conducted by Juhee Choi under supervision of Jason Weaver in the University of Florida Chemical Engineering Department.

PAGE 34

34 The Haber Bosch process is incredibly important because i t transforms essentially inert dinitr ogen into chemically useful ammonia. There is an abundance of nitrogen available in the atmosphere, b ut very little that is reactive, because t he triple bond of dinitrogen is extremely stable. 35 Due to the large quantities of nitrogen needed, this process uses approximately 2% of the global energy supply. 36 , 37 This process occurs at approximately 400 ° C and 100 atm. The catalyst used is generally an iron based heterogeneous catalyst. 38 The Ostwald process is the manufacturing of nitric acid from ammonia. First , the NH 3 is oxidized to NO and NO 2 , which then react with water to form nitric acid. In 1988 , this process produc ed about 20 million metric ton s of nitric acid. This reaction generally occurs at 1100 K and atmospheric pressure over a catalyst of Pt 10% Rh gauze. 39 Originally , this was achieved with a pure platinum cata lyst, but the catalyst was found to have improved reactivity when rhodium was combined with it. At atmospheric pressure , the catalyst last s up to 1 year before be ing replaced. 40 The Contact process is the synthesis of sulfuric acid by oxidizing sulfur dioxide. T he annual production of sulfuric acid by this process was 145 million tons in 1992 of which more than half was used in the production of phosphoric acid for fertilizer. The process g enerally occurs at 970 1170 K. Originally , platinum was used as the cataly st but was discontinued in favor of vanadium(V) oxide. 33 , 40 Process of Forming Heterogeneous Catalysts Many heterogeneous catalysts are the result of grafting or anchoring a transition metal to an oxide surface. The purpose of this is to combine the advantages of homogeneous catalysis (versatility and selectivity) with the durability and ease of separation of the heterogeneous catalyst. 19 Anchoring generally occurs through a

PAGE 35

35 covalent bond from a functional group on the substrate to an inorganic or organometallic precursor. 41 The general process of preparing an oxide supported heterogeneous catalyst has four s teps. 19 First, the support is pretreated, in a pro cess generally involving heat , to remove any physisorbed water. This water, or other unwanted species, can cause side reactions to occur with the precursor. In some c ases , this step also involves the dehydroxylation of the surface. Second, the metal complex is anchored to the substrate. This o ccurs immediately upon contact ing the surface. The metal complex reacts with the hydroxyl on the surface via oxidative addition or nucleophilic substitution . This step typically takes place in anhydrous solvents or in the vapor phase. Complexes are known to form one, two, or three bonds with the surface. Third, not all of the precursor react s with the substrate and th e unanchored m etal complexes must be removed. The se are removed by inert gas flushing, in the case of vapor phase reactions, or by washing with pure solvent for liquid phase reactions. Finally , the ligands on the complex must be removed either by hydrolysis, with the wa ter in the liquid or vapor ph ase , or thermal decomposition. Precursors for Heterogeneous Catalysts The metal complexes used as precursors are of three general types: metal chlorides, metal alkoxides, and organometallics. 19 For metal chlorides , the chlorides must be removed for the catalyst to be active. The binding mode of the metal can generally be determined by the amount of HCl evolved during the reaction. 20 The binding mode can be affected by the pretreatment temperature as well as the reaction temperature. 42 , 43 One of the drawbacks to using me tal chlorides is the resulting structure has a low quantity of grafted metal. This could be the result of low reactivity

PAGE 36

36 with the substrate, chlorination of the substrate, or hydrolysis of the metal substrate bond. 19 Metal alkoxides are another class of precursor which gives very good dispersion of the metal on the surface. The most prevalent ligands used are the ethoxy and acet ylacetonate ligands. Inumaru et al. reported the deposition of VO(OEt) 3 on silica at 423 K. 21 The grafting and desorption of unreacted precursor occurred at 723 K. This was used to catalyze the oxidative dehydrogenation of ethanol to acetaldehyde. This grafted complex was t hree to four times more active than the analogue prepared by the impregnation technique, a process in which the complex is dissolved in a solvent and exposed to a porous surface. 19 Organo metallic complexes used to sy nthesize heterogeneous catalysts fall into two sub categories: metal allyls and metal carbonyls. 19 The metal allyls have been used frequently for many years. Iwasawa has compiled a table of common metal allyl precurs ors and their uses in catalysis , showing mostly early transition metal allyls being used for all major types of polymerization. 34 Typically , they are anchored by a metal allyl reacting with the hydroxy l of the substrate. The metals chromium, molybdenum, and tungste n typically form two bonds with the surface, while n ickel, palladium, and platinum typicall y form one bond . Both one and two bonds are known with titanium, zirconium, and rhodium. 19 Metal carbonyls have three main differences from other metal complexes, including metal allyls. First, the binding of metal carbonyls to t he surface is much more complex, which can lead to various bonding modes , including Van der Waals , ionic, and covalent. Second, the metal carbonyl is zerovalent, making the reaction with the surface

PAGE 37

37 generally an oxidative addition. Third, the complexes can be mononuclear, polynuclear, or heteronuclear. There are three primary reasons for the prevalence of the use of these complexes for heterogeneous catalysis. 19 First, metal carbonyls are frequently used in homogeneous catalysis , and by combining the well known reactions with the advantages of heterogen eous catalysis , the desired effect can be achieved . Second, grafted mononuclear and polynuclear complexes can be formed on the decomposition of the metal carbonyls. Third, small particles of the metal can be achieved from these complexes, giving a high sur face area and therefore , a more active catalyst. Drawbacks to Heterogeneous Catalysts and How to Deal with Them Although heterogeneous catalysts are very prevalent and have adva ntages, there are still four disadvantages to them : 34 1. They usually contain more than one type of active site 2. The surface is generally not well defined 3. Heterogeneous catalysts are usually less selective due t o the variation in active sites 4. Only the metal ato ms on the s urface can be used in catalysis Three of the four disadvantages can be linked back to a lack of uniformity in the surface structure. The metal has considerable interaction with the surface and if the local structure changes, that affects t he av ailable interactions. 44 A strong metal support interaction was found with anatase titania supported palladium when exposed to c arbon monoxide and hydrogen gas. However , when rutile titania was used instead, there was little to no interaction. Thus, the local structure surround ing the metal centers on the surface can have a significant effect on t he efficiency of the catalysis. 45 Copper on the surface of zirconia was used to catalyze the reaction of carbon dioxide or carbon monoxide with hydrogen

PAGE 38

38 gas to fo rm methanol. Two different phases of zirconia were u sed, monoclinic and tetragonal. Using the same percentage of copper and the same surface area ratio, the monoclinic phase was shown to be far more active toward bo th carbon monoxide and carbon dioxide. If a catalyst could be designed that contained the metal centers in a well defined, uniform distribution, the activity of the catalyst would be greatl y improved. The question remains whether there is a way to bind the metal centers to specific sites and if so, can they be bound to areas that are more active for catalysis . A brief survey of the li terature suggests that the hydroxyl or amine groups do provide selective binding at undercoordinated surface sites. One particular experiment involving ethanol is of considerable interest. 46 A relatively defect free surface of titanium oxide was prepared while o ther surfaces wer e prepared in a similar manner, but then exposed to electron bombardment for various lengths of time to remove oxygen atoms, leaving the undercoordinated titanium behind. Each of the surfac es was exposed to ethanol and t he binding and re activity of the eth anol was measured by temperature programme d reaction spectroscopy (TPRS). The reactivity of the adsorbed ethanol showed significant changes based on the number of defects . Not only did the reactivity change, the total a mount adsorbed changed as well; t he s urfaces with defects had more ethanol bound to the su rface than those that did not. This indicate s that ethanol was preferentially binding to the undercoordinated sites. This fits with a mechanistic understanding because the undercoordinated sites have exp osed metals while the terraces have terminal oxides. A hydroxyl functional group is unlikely to react with a n

PAGE 39

39 oxide or hydroxyl, but is far more likely to react with the exposed metal centers on the undercoordinated sites. Furthe r studies have been conduct ed to determine if amine functionalities have similar reactivity as predicted based on their similar acid base chemistry to hydroxyl functional groups . Ammonia and simple alkyl amines bind selectively at point defect sites; this , in turn , seems to block th e neighboring sites, thus lowering the total amount bound as compared to a defect free surface. 47 , 48 This binding appears to occur from the nitrogen atom directly to the Ti 4+ in the TiO 2 surface. Pyridine , on the other hand , was found to interact with the surface primarily by fa cial coordination. 49 One of the fundamental scientific questions regard ing heterogeneous catalysis is whether the reaction occurs on the f lat surface or on the defects . 50 C arbon carbon bond breaking occurs three to five times faste r on a kinked platinum surface. K inked palladium surface offers the fastest rate for methane decomposition. 22 The synthesis of ammonia on iron an d ruthenium occur s nine orders of magnitude faster at the steps than on the terrace. 51 Based on density functional theory (DFT) calculations, Li et al. concluded that defects are always favored for dissociative reactions , while associati ve reactions may or may not be favored. 22 Thus the active sites for many reactions occur at defect sites, such as the kink or step sites. If an organometallic complex could be tether ed vicinal to an undercoordinated site by the linking hydroxyl or amine g roup, the metal from the precursor could have favorab le interactions with the metal of the substrate . This would be achieved essentially by reversing the roles of the substrate and the precursor from standard

PAGE 40

40 heterogeneous catalyst synthesis. The precursor would have the hydroxyl functional group that reacts with the metal of the oxide support. Design As outlined above , the first criteria for heterogeneous catalyst precursors in this study is the incl usion of functional groups that direct the binding to low coordination step and kink sites . This is achieved by tethering a hydroxyl or an ami ne moiety to a ligand . As discussed above , these functional groups cause selective binding at the low coordination sites . Another feature of the desired complexes is low reactivity of the ligands so that reaction occurs only at the hydroxyl or amine groups. This will i ncrease the selectivity by having only one group directing the binding site . The ligands must also b e removable via mild thermal means . Some of the ligands that fit these criteria would be bipyridines, cyclopentadienyl systems, carbonyls, halogens, and alkyls. 52 The final feature required i s the volatility of the complex that allows it to reach the s urface. This presents a problem because t he design features outlined above, specifically the hyd roxyl or amine moiety, decrease the volatility of the complex. For small organic molecules, e.g . methanol, this is not a problem, however larger organometallic complexes containing hydroxyl functional groups may or may not prove volatile enough. This will need to be tested. Thus complexes (Figure 2 1 ) have been proposed that combine the needed functional groups with the unreactive and easily removed ligands . Com pound 2 1 has the hydroxyl functional group tethered to a bipyridine in a position where elimination to leave the hydroxyl attached to the surface, is impossible . It also has the easily removable methyl ligands or the less easily removed but less reactive chlorides

PAGE 41

41 attached to the palladium . The methyls would provide easier removal, but could affect the reactivity and direction of the precursor on the surface. The chlorides would be less reactive though still likely to be removable. The bipyridine ligand is a known compound and the position of the tethered hydroxyl will hold the palladium in close proximity to the undercoordinated sites; h owever, it could make the formation o f the complex more difficult due to the steric bulk surrounding the binding sites . T he volatility of both 2 1a and 2 1b is unknown. Figure 2 1. Complexes proposed for use in preparing heterogeneous catalysts. Compound 2 2 is a known volatile compound that contains the desired unreactive ligands and volatility. Two of the ligands, carbonyl and methyl, should be removable. The cyclopentadienyl (C p) ligand should be unreactive, however it usually forms strong bonds with the metal and might be difficult to remove. This complex does not , however , include the hydroxyl or amine functional groups . This complex provides a model compound to help understand the reactivity with the surface without the hydroxyl functional groups. Compound 2 3 provides an attractive compromise by combining the features of compound s 2 1 and 2 2 . It has a similar overall structure to compound 2 2 but includes the hydroxyl moiety from compound 2 1 . It has also replaced the methyl of 2 2 with a less reactive chloride . This compound will be less volatile than compound 2 2 ; however

PAGE 42

42 its volatility will need to be tested to determ ine if it is volatile enough to be dosed to the surface. Palladium is a well known catalyst in both homogeneous catalysis and heterogeneous catalysis and was thus chosen for this model system . It is known to catal yze many common reactions including the Heck reaction, 53 the Suzuki reaction, 54 and hyd rogenation reactions. 55 These and ot her reactions catalyzed by palladium are important for the pharmaceutical and fine chemical industries. Ruthenium was chosen for its versatile applicability in heterogeneous catalysis. Ruthenium on alumina is very active in alkane hydrogenolysis , 56 olefin hydrogenatio n, and isomerization. 57 , 58 The most important catalytic use for ruthenium is Fischer Tropsch chemistry, 59 the process of turning carbon monoxide and hydrogen gas into hydrocarbon chains. Ruthenium is the most active and produc es the highest weight chains of the commonly used heterogeneous catalysts. Synthesis of Organometallic Precursors The ligand 2 8 was synthesized as reported in the literature (Figure 2 2). 60 First , a nitrogen of 2 4 was oxidized using meta chloroperoxybenzoic acid to produce compound 2 5 . Next, TMSCN was used to add the nitrile group to the ring to form 2 6 . The cyanide was converted to the ester using acid and methanol to reach compound 2 7 , which was then reduced using sodium borohydride to obtain compound 2 8 . Next , the hydroxyl group was protected usin g t butyldimethylsilyl to form compound 2 9 . This group was chosen due to its reported stabi lity to column chromatography and i t s removal via fluoride ion treatment . 61

PAGE 43

43 Figure 2 2. Synthesis and protection of the ligand. Synthesis of the palladium complex is based on the synthesis reported in the literature for (bpy)PdMe 2 . 62 This same method was employed using model compound 2 13 to test the feasibility without wasting the more synthetically challenging compound 2 9 (Figure 2 3). First, N,N,N',N' tetramethylethylenediamine (TMEDA ) was added to p alladium (II) chloride to form compound 2 10 . Next the chlorides were replaced with methyl groups using methyl lithium to reach compound 2 11 , which proved less thermally stable than was reported in the literature 63 and was stored at 0 decomposition. Finally , compound 2 13 was added to obtain compound 2 12 . This reaction did not work as well as reported for 2,2' bipyridine , however the crude produ ct was confirmed via 1 H NMR.

PAGE 44

44 Figure 2 3. Synthesis of the model complex. The desired ligand, compound 2 9 , was used in a similar reaction based on these promising results (Figure 2 4); however the reaction did not prove successful. One possible reason for this is the added steric bulk raises the activation energy for the reaction, thus slowing down the process and allowing the thermally sensitive starting material to decompose . The reaction was attempted at both 78 C an d 0 C and the result was simply a mixture of the two starting materials. If they were subsequently allowed to warm to room temperature, decomposition occurred. An al ternate synthetic route involving reversing the steps of the reaction was attempted. Compo und 2 9 was added first and then the chlor ides were replaced by methyls. However , a new method of substitution for the chlorides was employed because the very basic methyl lithium would most li kely deprotonate the benzyl position af ter deprotonating the hy droxyl. Therefore , boron based transmetallation reagents were investigated.

PAGE 45

45 Figure 2 4. Attempted s ynthesis of 2 14 . Compound 2 1 5 was synthesized from 2 13 and palladium ( II) chloride and used to study the reactivity of boron based transmetallating agents (Figure 2 5). By using phenylboronic acid and potass ium carbonate as base , compound 2 16 was obtained ; however , a significant amount of 2 13 was produced as well . Very sim ilar results were obtained with the methyl and ethyl derivatives. The solvent was changed as well as the time of reaction, yet the results remained the same ; one of the chlorides had been replaced with an alkyl or aryl group, yet this product was only obta ined in a ~1:3 ratio with the ligand removed from the metal cent er. When base optimization was performed, it was discovered that cesium carbonate caused the formation of the symmetric product, compound 2 12 . However, the ratio of desired product, 2 12 , to undesired product, 2 13 , was not improved. Further optimization studies were performed for concentration and solvent, yet no significant improvement on the ratio was observed.

PAGE 46

46 Figure 2 5. Transmeta llation investigations c onducted using boron based reagents. The de sired ligand 2 9 was used with the plan to recover the unattached ligand and reuse it . Compound 2 9 was added to the palladium to form 2 18 in nearly quantitative yields (Figure 2 6). In a surprising result, the replacement of the chlorides was achieved successfully to form 2 14 with no sign of the ligand being removed from the metal center. This could possibly be due to a solubility issue as this compound was far more soluble than the model system. Figure 2 6. The synthesis of protected complex 2 14 . With the complex now synthesized, the only step remaining was the dep rotection of the hydroxyl group, which proved to be far more difficult than anticipa ted . The treatment with fluoride ions in the form of tetrabutylammonium fluoride decomposed the complex rather than yi elding the deprotected product. While attempting to isolate the

PAGE 47

47 ligand, it had been discovered that using silica column chromatography res ulted in removal of the protecting group . This was attempted with the complex yet it remained protected through the column. Because deprotection was proving difficult , the unprotected ligand was added to palladium (II) chloride (Figure 2 7). This reaction failed, f orm ing a wide mixture of products. A similar reaction was attempted by adding 2 8 to Pd(cod)Cl 2 , attempting to replace the cyc looctadiene with the bipyridine to form 2 1a . 1 H NMR indicated that the bipyridine had bound to the palladium, however , the cyclooctadien e appeared to be bound as well. The mass spectrum of the compound revealed the absence of any chlorine at oms in the compound, suggest ing the replacement of the chlorides rather than the cyclooctadiene to form compound 2 19 . Another method involved reacting 2 8 with PdCl 2 in refluxing methanol. 64 This produce d the desired product, 2 1a . Figure 2 7. Alternate synthesis of unprotected complex 2 1a . Next, the synthesis of the ruthenium based complexes was attempted (Figure 2 8) . First, triruthenium dodecacarbonyl was refluxed with freshly cracked cyclopentadiene in heptane to afford the dimer 2 20 . A sodium mercury amalgam was

PAGE 48

48 generated by adding mercury to molten sodium in refluxing toluene. The toluene wa s then removed and 2 20 was added in THF to stir overnight at room temperature . The reduced complex was then reacted with methyl iodide to form the desired complex 2 2 , which was then purified by sublimation. A similar synthesis was reported in the literat ure using K(BHEt 3 ) as a reductant to afford a 30% yield. 65 The method employed in this dissertation achieved a 43% yield while using a rea gent that is currently cheaper. Figure 2 8. The syn thesis of volatile complex 2 2 . Next , the hydroxyl con taining complex 2 3 was synthesized (Figure 2 9 ) . First ruthenium (III) chloride hydrate was refluxed with 1,2,3,4,5 pentamethylcyclopentadiene in ethanol to yield dimer 2 22 . This was exposed to air, which allowed one oxygen atom to be inserted between the ruthenium atoms. The complex was then dehydrated by losing a single hydrogen from a methyl group on each cyclopentadienyl ligand and the bridging oxygen. Carbon mon oxide was then bubbled through the solution of the compound in dichloromethane, which caused the transfer of one chloride to the methyl on the cyclopentadienyl ligand to form complex 2 23 (Figure 2 10 ). 66 This was then reacted with collidine in water and THF to form the desired complex 2 3 , which was purified by alumina column chromatography. 67

PAGE 49

49 Figure 2 9. The synthesis of the hydroxyl derivative 2 3 . Figure 2 10. The mechanism of the formation of complex 2 23 . Adapted from R eference 66 . Application If the compound is non volatile, no matter the other properties , it is useless for this method of synthesizing heterogeneous catalysts , as discussed previously, because it cannot be delivered to the surface. Compound 2 1a was first tested for volatility by taking the melting point and it was found to decompose at about 200 ° C. Sublimation of

PAGE 50

50 the compound was attempted under vacuum and decomposition still occurred at about the same temperature. T hermogravimetric analysis (TGA) showed decomposition beginning at about 200 ° C and continuing step wise to 600 ° C with no plateau (Figure 2 11 ). These data show that , although the complex contains many of the features desired in a precursor, it is not useful for the current purpose due to its lack of volatility. Figure 2 11. TGA of compound 2 1a . C ompound 2 2 was found to sublime at 59 ° C at atmospheric pressure , while a t reduce d pressure, 100 mTorr, the complex sublimed at room temperature. This is comparable to wha t is reported in the literature. Davison et al. report the sublimation of this complex at 100 mTorr and 40 °C. 68

PAGE 51

51 Figure 2 12. The TPD measurements of compound 2 2 after dosing for 5 minutes. C omplex 2 2 was then used as a precursor in the model study by depositing it on strontium titanate (STO) . The strontium titanate was cut across a high Miller index face to give a surface with many step and kink sites. Compound 2 2 was dosed on the prepared surface under ultra high vacuum (UHV) co nditions at 100 ° C for 5 minutes at a pressure of 2.0 nanoTorr. The surface was then slowly warmed and the desorbed species were measured by mass spe ctrometry in a process known as temperat ure programmed desorption (TPD) ( Figure 2 12 ) . The desorbed compounds show signs of decomposition, as well as desorption of the unreacted precursor (Table 2 1) . Water is a known contaminant, explaining its presence. The presence of carbonyl shows the decomposition of the precursor. Peaks

PAGE 52

52 are also present for methoxy, a possible product from the methyl reacting with a surface oxygen. The peaks at 43 and 44 could be an acetyl group, perhaps from ethanol contamination or reaction between carbonyl and methyl . The peak at 63 possibly corresponds to a methyl reacting with a surface titanium. The largest peak , at 167, most likely corresponds to cyclopentadienyl ruthenium after the loss of carbonyls and methyl. The peak at 182 likely corresponds to the precursor after loss of the carbonyls. Table 2 1. Possible identitie s of the desorbed species from Figure 2 12 . Mass (amu) Possible Identit y 18 H 2 O + 28 CO + 31 CH 3 O + 43 CH 3 CO + 44 CH 3 COH + 59 ? 63 CH 3 Ti + 72 ? 161 ? 167 CpRu + 182 CpRuCH 3 + 281 ? D esorption of the compound of m/z 167 was measured under different adsorption conditions ( Figure 2 13 ) . These were dosed at 2.0 nanoTorr with the surface cooled to 100 ° C. The time of exposure varied from 15 seconds to 5 minutes. Predictably , as the time of exposure increased, so did the amount of complex deso rbing from the surface. The tall peak corresponds to a multilayer while the wide low shoulder that reaches nearly to 350 K corresponds to a monolayer slowly desorbing , indicating positive interactions between the surface and the precursor.

PAGE 53

53 Figure 2 13. TPD of compound 2 2 after being exposed to the STO surface for 15 seconds to 5 minutes. A n X ray photoelectron spectrum (XPS) was taken of the surface under three con ditions: clean, exposed to precursor, exposed t o precursor and heated to 700 ° C ( Figure 2 14 ) . The surface was exposed to the precursor by repeated adsorption and desorption. Ruthenium remains even after heating the surface to 700 ° C. A high resolution XPS of the ruthenium 3d region (Figure 2 1 5) shows the peaks shift upon heating, indicating the reduction of the ruthenium(II) . This is more p ronounced in the 700 °C heating , but is noticeable in the lower temperatures as well . A high resolution XPS was taken of the ruthenium 3p and titanium 2p region (Figure 2 16) , which shows a similar e ffect, the peaks are shifted as the surface is heated. This indicates that the ruthenium complex is being affected by the heat, perhaps losing its ligands.

PAGE 54

54 Figure 2 14. XP S of clean STO, d osed with precursor, and heated to 700 ° C. Figure 2 15. XPS of the R u 3d and Sr 3p region for dosed and heated 2 2 on STO.

PAGE 55

55 Figure 2 16. High resolution XPS of Ru 3p and Ti 2p region of 2 2 on STO. Next, ion scattering spectroscopy (ISS) was conducted on the sample (Figure 2 17). There is no ruthenium present on the clean STO surface, while the dosed surface does contain ruthenium. The surface that had been dosed and then heated to 700 ° C lost the ruthenium peak which seems to contradict the XPS data. Howeve r, this technique only probes one or two monolayers. The lack of ruthenium peaks for the heated surface, combined with the presence of ruthenium peaks in XPS for the same surface, indicate that the surface is likely reformed on exposure to heat, causing the ruthenium to be encapsulated by the strontium and/or titanium.

PAGE 56

56 Figure 2 17. ISS of compound 2 2 deposited on STO and heated to 700 ° C. Compound 2 2 appears to decompose in small quantities to form some ruth enium containi ng compound. C ompound 2 2 was then dosed in smaller quantities at near room temperature to discover if this would decompose to form metallic ruthenium on the kink sites. Scanning tunneling microscopy was conducted on the clean, kinked substrate (Figure 2 1 8) , showing the steps and kinks on the surface. Compound 2 2 was dosed to the surface for 45 seconds at 1.1 x 10 9 Torr and STM was taken again (Figure 2 19). This shows significantly rougher features than the clean surface. It is unclear whether the ruthe nium was deposited in a random manner or if the surface structure was reformed in some manner due to exposure to the precursor.

PAGE 57

57 Figure 2 18. STM image (500 Ã… x 500 Ã… ) of clean STO(12 2 1) annealed at 1275 K for 30 minutes. Figure 2 19. STM image (1000 Ã… x 1000 Ã… ) of STO(12 2 1) dosed with 2 2 for 45 seconds at 1.1 x 10 9 Torr.

PAGE 58

58 Compound 2 3 was tested for v olatility by heating the sample and found to melt at approximately 190 ° C, but seemed to decompose just above that temperature as the liquid changed co lor. Sublimation was attempted under vacuum, yet it seemed to decompose rather than sublime under standard vacuum conditions. This was tried under UHV conditions to determine if the reduced pressure would allow this compound to v olatilize rather than decom pose; however , decomposition was still observed. Based on the non volatile nature of the compound, it was determined to not be useful for this study. It seems that hydroxyl containing organometallic complexes generally decompose rather than sublime. This i s due to the strong hydrogen bonding occurring between molecules which makes it more difficult to break these bonds than to break the covalent bonds in the complex. Summary of Results As has been outlined above, a plan was devised to synthesize more well d efined and active heterogeneo us catalysts. The plan required reaction of organometallic complexes containing hydroxyl or amine moieti es with an oxide surface, in this case strontium titanate (STO). When initially proposed, the volatility of the organometallic precursor was not known . Low volatility became apparent as the experiments were conducted. This le d to an interesting dilemma: complexes with the dir ecting groups were not volatile and volatile complexes did not contain the desired direc ting groups. Compounds 2 1 and 2 3 contain ed the desired directing group but were unable to be delivered to the surface due to their non volatile nature. Compound 2 2 , on the other hand , was volatile and was transported to the surface, however , it did not c ontain the desired directing groups and seemed to deposit in a largely random manner .

PAGE 59

59 Another unforeseen consequence of the reaction was the surface did not seem to be static. The proposal involved removing the ligands via non thermal plasma to when the su rface was heated to 700 ° C, it seemed to restructure , encapsulating the ruthenium under strontium and/or titanium. Although selective placement of ruthenium on the surface was not achieved, important information was obtained. It was discovered that carbonyl and methyl ligands would react with the surface without significant heating. Ruthenium could form fairly stable bonds with the surface, enoug h to stay present after heating to 700 ° C. However, the structure appeared to restructure at those temperatures encapsulating the ruthenium.

PAGE 60

60 CHAPTER 3 DESIGN , SY N THESIS , AND APPLICATION OF ORGA NOMETALLIC PRECURSORS FOR PHOTOACTIVATED CHEMICAL VAPOR DEPOS I TION ON SELF ASSEMBLED MONOLAYERS 2 Photoactivated Chemical Vapor Deposition Metallized organic thin films have received increased interest because of potential device applications. These can be used in flexible electronics, solar cells, and bio sensors, 69 and in miniaturization of devices. 70 Organic Thin Films Self assembled monolayers (SAMs) have been of great interest to the scientific community since the early 1980s. 71 The firs t reported instance of a SAM in 1946 garnered very little interest and for many years this field went largely uninvestigated. 72 Eventually , interest was aroused and many different syste ms have been investigated, of which t h e most prevalent are alkane thiolates on gold. 71 There has been an increas e in interest in SAMs for two r easons. First, self assembly is known in nature, which leads to supermolecular organizations that result in complex systems and there is an academic interest in understanding this mechanism. Second, there is a shift in focus from pure chemistry to interdis ciplinary sciences due to its use in technology . Much research interest has been generated on the interface of chemistry with physics, biology, and engineering. SAMs have been of particular interest due to their relevance to technology. The advantage of SA Ms over molecular beam epitaxy and chemical vapor deposition (CVD) is the highly ordered nature of the films generated by self assembly. SAMs can 2 PACVD conducted by Zhiwei Shi (compound 3 1) and Jing Yang (compound 3 2) under supervision of Amy Walker at University of Texas Dallas.

PAGE 61

61 incorporate a range of functional groups, giving control over a large variety of surfaces and interactions. 73 Figure 3 1. The process of forming SAMs and the struc ture and interactions once formed. Adapted from Figure 1 on page 1534 of R eference 71 . SAMs are formed by first immersing the substrate in a solution of surfactant (Figure 3 1) . The surface active material in the solution spontaneously bonds with the surface and has positive interactions with other surfactants. Because of the simplicity of this process, SAMs are generally low cost and synthetically flexible. 23 SAMs are generally bound to the su bstrate through a thiol functional group and are differentiated based on their functionality on the terminal end. Metallization Process es For SAMS to be used in devices, a robust metalli c contact must be added which will provide electrical contact and prevent the degradation of the monolayer. Several methods have been developed to add this metallic contact. Chen et al . report ed the fabrication of a molecular electronic device using 2' amino 4 ethynylphenyl 4' ethynyl phenyl 5' nitro 1 benzenethiolate as the SAM . 74 Gold was then slowly e vaporated onto the SAM (Figure 3 2 ).

PAGE 62

62 Figure 3 2 . Molecular electronic device structure. Overall structure (top) and detailed view of monolayer (bottom). Adapt ed from Figure 1 on page 1550 in R eference 7 4 . Physical Vapor Deposition When the metal is evaporated onto the SAM by physical vapor deposition (PVD), the metal is known to react with the surface, penetrate to the substrate, and degrade the monolayer. 75 In 2003 Walker et al. reporte d the deposition of copper, aluminum, and silver onto methoxy terminated SAMs on gold. In all three cases there was a competition between reaction with the surface and penetration to the substrate. 76 In 2004 the same group reported studies involving copper, aluminum, silver and gold deposited on methoxy te rminated SAMs. 77 Aluminum pe netrated to the substr ate until a 1:1 aluminum:gold ratio was reached, at which point no further penetration occurred. For silver and copper , both penetration and overlayer nucleation occurs, with t he ratio of silve r and copper to gold having no effect. Gold has no interaction with the

PAGE 63

63 terminal methoxy, it simply penetrates to deposit a new layer of gold on the existing gold substrate, leaving the intact SAM floating on the new layer of gold. Walker et al. reported the deposition of gold and titanium on a SAM composed of 4 (phenylethynyl) phenylethynyl] benzen e thiol in 2004. 78 The gold behaved the same as with the methoxy terminated SAMs, penetrating and forming a floating SAM. The titanium, a much more reactive metal, reacted aggre ssively with the SAM, transforming it into titanium carbide. It was concluded that, under these conditions, gold and titanium would not be useful for devices if PVD was used. In 2005 Walker et al. reported the PVD of calcium and titanium on methoxy terminated SAMs. 79 The titanium reacted with the terminal methoxy and the alkyl chains simultaneou sly to form carbides , hydrides, and oxides until enough had built up to form a metallic overlayer which stopped further degradation. The calcium immediately degraded the term inal groups and only when that was complet e did the alkyl begin degrading. Both metals degraded the SAM significantly, though titanium did so at a much faster rate. Reactivity of PVD on SAMs generally falls into two categories: unreactive/less reactive metals and reactive metals. For less reactive metals, like copper, silver, gold, or aluminum, there is a competition between nucleation of islands at the interface, reaction with the terminal groups, and penetration to the substrate. More reactive metals, such as titanium, degrade the SAM to form oxi des or carbides. 75 Another fa ctor that affects the metal organic interaction is the sticking probability, (Eq uation 3 1 ) ( 3 1)

PAGE 64

64 where N is the number of adsorbed metal atoms while N 0 is the number of incident metal atoms. Thran et al. reported the deposition of silver on various polymers, which resu lted in quite a variation of . 80 For polyimides, the sticking probability was close to unity, while for Teflon AF the sticking probability is 0.002. Zaporojtchenko et al. reported the sticking p robabilities of copper and gold in comparison to the already reported silver. 81 A ll three exhibited the same trend; w ith the polyimides all had near un ity (0.95) sticking probability, while f or Teflon AF all were very small (0.02 0.002). It would be expected that the more reactive metals would have a higher sticking probability in general; this is not always the case. Tighe et al. reported the PVD of titanium onto a methyl terminated SAM in which the stick ing probability started out very low, approximately 0.1, and slowly increased as the total number of titanium atoms increased. 82 Titanium is a more reactive metal yet its sticking probability is initially quite low. Collaud et al. investigated the sticking probability of magnesium on polypropylene and discovered that for untreated polypropylene the stic king probability was essentially zero. 83 By treating the surface with plasm a , the sticking probability was raised to 0.3. This is still small, showing that more reactive metals do not necessarily have a higher sticking probability. Haick and Cahen summarize the drawbacks of PVD for deposition of metals on SAMs as follows: 70 1. It is difficult to control the thick ness and growth of the deposit 2. The metal could degrade the monol ayer rather than deposit on it 3. It is likely to pene trate to the surf ace or form nucleation islands 4. If the metal leave s channels to the substrate, devic es made from them would short

PAGE 65

65 Lift Off Method Vilan and Cahen repo rt an alternate method (Figure 3 3 ) for depositing the metal on the SAM. 84 First the metal film is evaporated using standard PVD onto a solid support that binds weakly to the metal. Next the metal film is detached from the surface by wetting with one or more solvents and floated on a liquid. A monolayer can then be adsorbed onto the metal film, the target substrate, or both. Finally the metal film is While this is an attra ctive alternative, there are still significant drawbacks and difficulties. The thickness of the film needed to optimize detachment is not known and might not correspond to the thickness needed for the device. The film is also known to wrinkle if the solven t is not removed in a quick a nd uniform manner. The wetting of the film can also cause repulsion from the target substrate. Due to these factors this method is difficult to reproduce. Chemical Vapor Deposition Chemical vapor deposition (CVD) provides an at tractive alternative to PVD and the lift off method just described, due to its control over the thickness, its chemical selectivity, and the wide variety of metals that are available. In this process a precursor containing the desired metal is volatilized and comes in contact with the hot substrate. The precursor molecule decomp oses to, ideally, the desired deposit on the substrate. 85

PAGE 66

66 Figure 3 3 . The lift off process illustrated. Adapte d from Figure 1 on page 797 of R eference 84 .

PAGE 67

67 A drawback for CVD on SAMs is that it generally requires temperatures of 200 ° C or greater. 23 Bensebaa et al. repor t the thermal behavior of alkylthiolate S AMs on gold: at ~350 K (75 °C) the SAM undergoes an irreversible transformation to a liquid and at ~400 K (125 °C) the SAM begins significant desorption . 86 Therefore the temperature needed for conventional CVD would not allow the SAMs to remain intact during deposition. CVD on SAMs at lower de position temperatures has been reported , thoug h the results are not ideal. Wohlfart et al . were able to deposit gold from methyltrimethylphosphine gold(I) on thiol terminated SAMs. 87 The precursor was heated to 70 ° C under vacuum and exposed to the SAM on gold. T he gold formed nanoparticles on the surface which, after repeated exposure, coalesced into a rough deposition. T his occurred exclusively on the thiol terminated SAMs which could be used to pattern the gold onto a surface. Weckenman et al . reported the depo sition of palladium from ( cyclopentadienyl )( allyl ) palladium onto thiol and methyl terminated SAMs. 88 First a monolayer of palladium was deposited selectively on the thiol terminated SAM without any deposition on the methyl terminated SAM. This palladium remained Pd(II) bound to the SAM by a Pd S bond and still bound to the allyl . Upon further exposure Pd(0) was deposited unselectively across methyl terminated SAM and on top of the already deposited Pd(II). To deposit Pd(0) in a selective manner, another method was devised. Trimethylamine alane was deposited selectively on hydroxy l terminated SAMs. The palladium precursor was then exposed to this surface and P d(0) was immediately deposited on the aluminum activated surface.

PAGE 68

68 In 2008 Lu et al. reported the deposition of aluminum and alumina on hydroxyl, carboxyl, and methyl terminate d SAMs. 89 Under an inert atmosphere , trimethylaluminum was found to deposit alumina on all three SAMs. Under ultra high vacuum (UHV) conditions, the alumina was not found on the methyl terminated SAMs, indicating the reaction with residual water and oxygen in the reaction vessel as its source . By rinsing with organic solvents the alumina could be removed from the methyl terminated SAM. Und er vacuum the precursor deposited met allic aluminum on the hydroxyl and carboxyl terminated SAMs with no deposition on the methyl terminated SAM. These properties were then used to pattern aluminum onto a substra te. An alternative to the thermal activation is the photoactivation of the precur sor. As has already been discussed, thermal activation is generally unfavorable for deposition on SAMs due to the thermal instability of the SAMs. Photoactivation can occur by two different processes. The first method is laser pyrolytic deposition which in volves heating the substrate via laser irradiation for the precursor to deposit onto it. This method for deposition has been around since the 1970s, at which time CO 2 based lasers were used for deposition on SiO 2 . 90 Tsao and Ehlrich reported the deposition of aluminum from triisobutylaluminum by using a CO 2 laser to heat a quartz substrate. 91 The other process by which deposition can occur is photolytic deposition which is preferable for the metallization of SAMs because it offers a lower temperature that is less likely to disturb the SAM. Photolysis can occur in the gas pha se or after adsorption onto the surface. 23 Calloway et al . were able to photochemically deposit aluminum from trimethylaluminum vapor on quartz, silicon, and sapphire using ultraviolet light to initiate the decomposition. 92 A mask w as used to pattern the aluminum onto the substrate .

PAGE 69

6 9 In 2012 Shi et al . reported the application of this method for deposition of aluminum on SAMs. 23 First the gold substrate was soaked in a solution of the relevant alkanethiol to form the SAMs. The SAMs used were methyl, hydroxyl, and carboxyl terminated. After the SAMs were formed, trimethylaluminum vapor was introduced into the chamber with and with out UV light. It was discovered that the samples that had been exposed to UV light contained approximately twice the amount of aluminum on the hydroxyl and carboxyl terminated SAMs. No aluminum was deposited on the methyl terminated SAMs with or without ra diation. The UV light was introduced parallel to the surface to limit the interactions with the SAM. To determine if the SAM was disturbed during the deposition, a control experiment was run without the trimethylaluminum present. Based on time of flight se condary ion ma ss spectrometry (TOF SIMS) approximately 1% of the hydroxyl and carboxyl terminal groups were photooxidized, while 4% of the methyl terminal groups underwent photooxidation. Next Shi et al. implemented a process to pattern the aluminum deposi tion by combining methyl and carboxyl terminated SAMs on the surface. First the carboxyl terminated SAM was applied to the surface. Next, a mask was used to cover part of it before exposing it to direct UV light to photooxidize the uncovered SAM. The mask was removed and the substrate exposed to a solution of hexadecanethiol to form the methyl terminated SAM. Then PACVD was used to deposit aluminum on the SAMs. Deposition occurred selectively on the ca rboxyl terminated SAMs (Figure 3 4).

PAGE 70

70 Figure 3 4. TOF SIMS of patterned SAMs after PACVD. Detection of Al + , AlO , and Au 2 HDT recorded. 23 This is a promising new technique of placing metal contacts on SAMs that needs to be better understood. New precursors need to be studied to gain a better understanding of the mechanism of deposition as well as to expand the metals available to deposit in this manner. The properties required for these precursors are as follows: 1. Volatile 2. Ligands are labile to: i) Protic surfaces and/or ii) UV light Design Conventional CVD precursors have generally been used for the current research application . Shi et al ., using simi lar conditions as employed here for PACVD used trimethylaluminum (TMA) which is a simple, volatile, commercially available, organometallic precursor that is used for conventional CVD. 23 To gain a better understanding of the reactivity, and to help design new precursors, less common precursors must be tried to determine which ligands are appropriate and which are not. Compounds 3 1 and 3 2 have been proposed to have the desired properties (Figure 3 5 ) . C ompound 3 1 contains the methyl ligand that proved successful in the aluminum compound because it could be lost via reaction with the protic surface or UV

PAGE 71

71 light . It also contains carbonyl ligands that have a quantum yield of 0.4 for their photolability. 93 It also contains the cyclopentadienyl ligand (C p) , a common ligand in the CVD literature. This experiment was designed to determin e if C p could be labilized by a mild therma l secondary reaction. This complex, as discussed in C hapter 2, is volatile, a necessary feature for this technique . Compound 3 2 is similar, but contains some key differences. Rather than a C p, an allyl 3 organic ligands are appropriate for PACVD precursors . The cyclopentadienyl ligand is known to be lost upon ligand association 5 3 1 and then to labilize. An ex ample of this was reported by Casey and Connor , in which the loss of C p from a rhenium complex was achieved by adding excess phosphine ligand and heating (Figure 3 6). 94 The allyl ligand will therefore give a good model to help determine the process of C p loss in this system. Compound 3 2 conta ins carbonyl and halide ligands . The carbonyl has a reported quantum yield of 0.7 for its loss, somewhat higher than for 3 1 . 95 The halide has not been tested under these conditions and thus will provide new information regarding appropriate ligands and complexes. It could be removed by the protic surface, but it is unclear if the surface will be acidic enough to do that. There is no reporte d sublimation temperature in the literature on this complex; this will need to be tested and evaluated. Figure 3 5 . Proposed precursors for PACVD on SAMs.

PAGE 72

72 Figure 3 6 . The mechanism of loss of C p as reported by Casey and Connor. 94 Synthesis of Organometallic Precursors for PACVD on SAMs Compound 3 1 was synthesized as reported in C hapter 2 (referred to there as 2 2 ) and shown in Figure 3 7 . T his compound was then used as reported below. Figure 3 7 . Synthesis of volatile compound 3 1 .

PAGE 73

73 Compound 3 2 was synthesized using a modified literature procedure ( Figure 3 8 ) . 96 The reported procedure did not specify the precise amount of time requi red to form 3 2 ; this was discovered to be a crucial detail. Upon refluxing triruthenium dodecacarbonyl with allyl bromide in isooctane, severa l color changes were observed. The reaction mixture started as an orange solution with solid triruthenium dodecacarbonyl on the bottom. As the solution was heated , the color changed to a deep red/orange color as the triruthenium dodecacarb onyl dissolved an d then slowly changed to a bright yellow color. If the solution was left refluxing a dull yellow solid would form. This solid was found to be i nsoluble in nearly all solvents and it was never characterized. However, if the reflux was stopped at the bright yellow solution, the desired product was obtained. The average time for this to occur was 20 minutes. The solution was removed under vacuum and the resulting yellow solid was sublimed at reduced pressure to form a white s olid. Although the lite rature repor ted the color of the product as yellow, a ll other spectroscopic data fit with the data i n the literature. It was concluded that there was a minor impurity in the compound prepared by Sbrana et al. that led to the yellow color. 96 Figure 3 8 . The synthesis of compound 3 2 . Next, it was attempte d to replace the bromi de with a methyl group to form a closer analogue to 3 1 (Figure 3 9 ). Multiple reagents were tested including: methyl lithium, methyl magnesium bromide, methyl magnesium iodide, methyl boronic acid, and trimethyl aluminum. In all cases the desired product, 3 5 , was not obtained. The

PAGE 74

74 color of the solution generally changed to a dark brown/black color. This could be cause d by nucleophilic attack on the allyl group, or by reductive elimination. 97 , 98 Figure 3 9 . Attempted replacement of bromide on 3 2 to form compound 3 5 . To determine if this was the case, compound 3 6 , which has two important properties, was synthesize d (Fig ure 3 10 ) . First, the allyl group of 3 6 is larger, mak ing the product of attack or elimina tion a liquid rather than a gas which can be detected via 1 H NMR studies. Second, the allyl is asymmetric and if the reductive elimination favors primary over secondary a compound could be designed with only secondary positions available in which the bromide could be replaced without affecting the allyl group. If reductive elimina 1 allyl, then compound 3 7 is produced, while compound 3 8 is produced if reductive elimi nation occurs from the 1 a llyl (Figure 3 11 ). 1 H NMR studies of 3 6 reacting with methyl lithium revealed the presence of 3 8 without the presence of 3 7 . Figure 3 10 . Synthesis of compound 3 6 .

PAGE 75

75 Figure 3 11 . Possible reductive elimination products from 3 6 . Due to these results, it was hypothesized that a compound containing only secondar y allylic positions would create a more stable complex. The ligand was synthesized using a literature procedure ( Figure 3 12 ) . First, crotonaldehyde was reacted with methyl lithium to form the alcohol 3 10 . 99 The hydroxyl was replaced with a bromide by reaction with phosphorus tribromide to form 3 11 . 100 This was then reacted with triruthenium dodecacarbonyl, but rather than form 3 12 , the compound appeared to decompose , most likely due to the added steric bulk surrounding the ligand. Figure 3 12 . Attemp ted synthesis of compound 3 12 .

PAGE 76

76 Application of Volatile Ruthenium Compounds for PACVD on SAMs After testing t he volatility of compound 3 1 , as reported in C hapter 2, the complex was used to deposit ruthenium on SAMs (Figure 3 13 ) . The precursor was placed in a vial attached to the dosing chamber via a vacuum line . The SAMs were formed on a gold substrate and pl aced in the dosing chamber. The dosing chamber was exposed to UV light while the precursor was heated by a water bath to 40 °C. The SAMs used were 16 m ercaptohexadecanoic acid (MHA) and 16 mercaptohexadecanol (MHL) . Figure 3 13. The experimental set up for the deposition of the ruthenium precursor. T h e clean SAMs were c haracterized by XPS (Figure 3 14 ) before exposing the MHA to the ruthenium precur sor. This was done by heating the precursor to 40 °C and exposing to a reduced pressure of 10 4 Torr. The resulting gas was exposed to t he SAM on gold for five minutes before recharacterizing the surface using XPS (Figure 3 15 ). The gold showed no change, indicating the lack of interact ion with the gold substrate, while the carbon peaks shifted slightly, interaction with the

PAGE 77

77 ruthenium. The ruthen ium 3d and 3p peaks were present indicating that the rutheniu m is deposited on the SAMs and not penetrating through to the substrate. Figure 3 14 . XPS of bare MHA SAM: a) g old 4f, b) carbon 1s and ruthenium 3d, c) oxygen 1s, and d) ruthenium 3p. Based on the binding energies, it a ppears that ruthenium (0) (BE = 280.7eV) is not present. 101 The other issue to be determined is if the ruthenium is actually bound to the surface or if it is simply resting on it. The surface was then rinsed with water and ethanol consecutively and XPS was retaken (Figure 3 16 ) which showed ruthenium still present which indicates that the ruthenium is bound to the SAM.

PAGE 78

78 Figure 3 15. XPS of MHA after dosing with compound 3 1 : a) g old 4f, b) carbon 1s, c) rut henium 3d, and d) ruthenium 3p regions. Binding energy: Ru 3d 5/2 : 282.4 eV; Ru 3p 3/2 : 464.2 eV. Figure 3 16 . XPS of MHA after d osing with 3 1 and rinsing with water and ethanol.

PAGE 79

79 Next, the same procedure was carried out on a M HL SAM on gold substrate. T he clean SAM was characterized by XPS ( Figure 3 17 ) before the MHL was expo sed to the ruthenium precursor using the same conditions as for the MHA. The surface was t hen rechar acterized using XPS (Figure 3 18 ), showing similar results to the MHA. The ruthen ium deposited on the SAM without showing interaction with the gold substrate, indicating a lack of penetration. Figure 3 1 7 . XPS of clean MHL SAM on gold substrate: a) g old 4f, b) carbon 1s, c) oxygen 1s, and d) ruthenium 3p regions. Based on the bindin g energies, it appears that, just as with the carboxyl termi nated SAMs, it is not ruthenium (0) (BE = 280.7 eV) present. 101 The ruthenium on MHL has essentially the same binding energy of ruthenium on MHA SAMs. The surface was then rinsed with water and ethanol consecutively and XPS was taken ( Figure 3 19 ) to determine if the ruthenium was bound or simply resting on the SAM . There is still

PAGE 80

80 clearly rut henium present, the amount appears to have decreased , as was the case with MHA. Figure 3 18 . XPS of MHL SAM after dosing with ru thenium precursor for 5 minutes: a) g old 4f, b) carbon 1s, c) ruthenium 3d, and d) ruthenium 3p regions. Binding energy: Ru 3 d 5/2 : 282.4 eV; Ru 3p 3/2 : 464.1 eV . Figure 3 19 . XPS of MHL SAM after dosing by ruthenium precursor and rinsing wit h ethanol and water: a) g old 4f, b) carbon 1s, c) ruthenium 3d, and d) ruthenium 3p regions.

PAGE 81

81 Figure 3 20 . SIMS of ruthenium precurso r d osed on MHA: a) r uthenium ion region and b) rut henium cyclopentadienyl ion region . S econdary ion mass spectrometry (SIMS) was performed on clean, do sed, and rinsed MHA (Figure 3 20 ), showing evidence of ruthenium ions and organoruthenium ions in the rinsed SAMs (Table 3 1). The most prominent peaks corresponded to the ruthenium cyclopentadienyl fragment. This fragment, along with others containing this fragment, show that cyclopentadienyl ligand forms very strong bonds that are not easily brok en. These are similar r esults to what was obtained in C hapter 2. The hydroxyl terminated SAM, HML, showed v ery similar results (Figure 3 21 ).

PAGE 82

82 Table 3 1. m /z values for ruthenium containing ions from SIMS of 3 1 on MHA. Mass Peaks (amu) Possible Ruthenium Compounds 128, 130, 127, 126, 125, 122 C 2 H 2 Ru + 141, 143, 140, 139, 138, 135 C 3 H 3 Ru + 154, 156, 153, 152, 151, 148 C 4 H 4 Ru + 167, 169, 166, 165, 164, 161 C 5 H 5 Ru + 180, 182, 179, 178, 177, 176 C 6 H 6 Ru + 195, 197, 194, 193, 192, 189 C 5 H 5 CORu + 199, 201, 198, 197, 196, 193 C 5 H 5 O 2 Ru + 232, 234, 231, 230, 229, 226 C 10 H 10 Ru + Figure 3 21 . SIMS of 3 1 dosed on MHL: a) r uthenium ions and b) ruthenium c yclopentadienyl ions .

PAGE 83

83 Figure 3 22. XPS of 3 2 deposited on MHA (carboxyl terminated) SAM. High resolution XPS of a) carbon 1s and ruthenium 3d and b) ruthenium 3p regions. Next, compound 3 2 was dosed to SAMs using a method similar to the one used for 3 1 . The compound was exposed to the MHA SAM for 32 minutes by reducing the

PAGE 84

84 pressure to 4.5 x 10 4 Torr and heati ng the sample to 90 °C. The sample was irradiated with a 30 W D 2 a rc lamp. Compound 3 2 required higher temperatures and more time than 3 1 , most likely due to the lower vapor pressure. The surface was then c haracterized by XPS (Figure 3 22 ). The carbon 1s binding energies increased, showing interaction of the metal/precursor with the SAM. The ruthenium 3d peak was present, indicating the presence of a ruthenium species. Its binding energy was 281.5 eV, which indicates that the ruthenium is still oxidized. T he ruthenium 3p peak , however, could not be distinguished from noise. T ime of flight secondary ion mass spectrometry (TOF SIMS) of the surface clearly indicated the presence of ruthenium and ruthenium containing species (Figure 3 23 ). T he allyl is st ill present in the deposit, indicating incomplete decomposition. The same procedure was employed for the MHL and HDT SAMs; however little to no ruthenium containing species were deposited. The XPS showed no sign of ruthenium 3d or 3p (Figure 3 24 ). This was expected from the HDT SAMs; however the lack of deposition on MHL was a surprising result. This shows a clear distinction between the reactivity of 3 1 and 3 2 .

PAGE 85

85 Figure 3 23. TOF SIMS of 3 2 deposited on MHA: a) r uthenium ion region and b) r uthenium allyl ion region.

PAGE 86

86 Figure 3 24 . XPS of 3 2 deposited on a) MHL and b) HDT . In contrast, bromide was observed on all three SAMs, as in dicated by TOF SIMS (Figure 3 25 ). The peaks were larger and more intense for the carboxyl terminated SAM than f or the methyl or hydroxyl terminated SAMs.

PAGE 87

87 Figure 3 25 . TOF SIMS of the bromide region of 3 2 deposited on a) MHA and b) MHL . Summary of Results from PACVD on SAMs These results provide an interesting insight into the reactivity of precursors for PACVD on SAMs. Compound 3 1 de monstrates that the methyl group is easily removable, either due to the UV radiation or more likely the protic surface. This shows

PAGE 88

88 that methyl groups, and possibly other alkyls, are suitable li gands for PACVD o n SAMs. T he carbonyl ligands were easily removed, most likely due to the UV radiation, showing them to be ideal ligands for this purpose. Although the observation of Ru+ ions were evidence of some Cp loss , the C p ligand was largely u naffected by the reacti on, showing this ligand is not suitable for PACVD . 5 organic ligands are not appropriate for PACVD 1 organic ligands, such as methyl, have shown to be useful. T he associative ligand induced loss of C p involves ring slip to 3 and 1 , and finally loss of ligand. So the question 5 3 3 1 transition. For this reason compound 3 2 was tested to determine the reactivity of the allyl ligand. The results from 3 2 showed th at the allyl does not leave under these conditions. The results also showed a dif ferent reactivity with the SAMs than seen with 3 1 . Compound 3 2 only deposited material on carboxyl terminated SAMs while 3 1 deposited material on carboxyl and hydroxyl terminated SAMs. This change in reactivity is likely due to the bromide rather than the allyl (Figure 3 26 ) . In both cases the carbonyl is lost via irradiation, and the ruthenium binds to the SAM using the now open coordination si te. For 3 1 the methyl is removed presumably via the acidic hydroxyl or carboxyl groups . For 3 2 the bromide will be harder to remove via acid and likely only the carboxyl ic acid is strong enough to remove it. Synthesis of a methyl derivative of 3 2 would help determine i f it was in fact the bromide causing the change in reactivity.

PAGE 89

89 Figure 3 26 . Proposed mechanism for PACVD of 3 1 .

PAGE 90

90 CHAPTER 4 DESIGN , SY N THESIS , AND APPLICATION OF ORGANOMETALLIC PRECURSORS FOR ELECTRON BEAM INDUCED DEPOSITION 3 Electron Beam Induced Deposition Electron beam induced deposition (EBID) is a deposition technique in which an electron beam is passed through a precursor in the gas phase to deposit metal on a substrate. 24 First the precursor is adsorbed onto the surface of the substrate. When the molecules are irradiated by the electron beam, the precursor breaks apa rt into volatile and non volatile components. The volatile compounds are removed while the non volatile compounds adhere to the surface. Ideally this deposition only occurs in the immediate area surrounding the electron beam. It is generally performed in e lectron microscopes, making the observation of the patterned deposit uncomplicated. This lithographic technique was first reported in 1934 by Steward as an unwanted growth from contamination gases in the system. 102 Steward concluded that film could be depos ited by electron bombardment in the presence of silicon oil vapor. 103 Shortly afterward Baker and Morris reported the deposition of metallic films via electron impact in the presence of organometallic complexes containing tin and lead. 104 This field has been growing steadily ever since, with a surge of developments in the late 1980s . 3 Surface science performed by Julie Spencer under supervision of Howard Fairbrother at Johns Hopkins University. Gas phase studies conducted by Rachel Thorman under supervision of Oddur Ingólfsson at Iceland University.

PAGE 91

91 This technique is very useful for two industrial applications : mask repair 25 and circuit edit. 105 Photomasks are used extensively in the microelectronics industry to transfer a pattern to the substrate. These masks can be damaged in use or in production and EBID has proven to be a useful technique in repairing the damages. Circuits are fabricated in large quantities and sometimes errors can be included in the fabrication. EBID allows the individual edit of circuits to repair or change them as needed. There has also been academic in terest in its use for the following: diodes, 106 photonic crystals, 107 probes, 108 electron sources, 109 conducting wires, 110 micro superconducting quantum interference devices, 111 nanotweezers, 112 seeds for nanotube growth, 113 and conducting or nonconducting joining technique . 114 , 115 One of the major problems still associated with EBID is the composition of the deposit. Ideally the metal would be 100% of the deposit, yet in many cases it is less than 20%. The composition can be affected by current density, electron energy, scan pattern, writing direction, extra irradiation, anneali ng, gas pressure, reactive gases, and precursor identity. 24 Each of these factors will be examined in turn, in the context of the precursor. Current Density The metal content of the deposit of has been demonstrated to increase with increasing current density. Koops has shown that using Me 2 Au(tfac), Mo(CO) 6 , and CpPtMe 3 with an increased beam current increases metal content. 116 However, the gold compound was the only one that was found to be very effective. The percent age of molybdenum increased yet still stayed below 10%. The platinum content increased as well, yet stayed below 15%.

PAGE 92

92 It has been hypothesized that this effect is due to two pa rallel processes, one due to charge and the other due to heating of the substrate. 24 With a higher current density, the time during which the deposit is exposed to charge decreases. This leads to a shift from a n electron limited regime to a precursor limited regime. As the higher charge flux is hitting the precursor over a shorter time, more of the volatile components are desorbed from the surface leaving a higher concentration of the non volatile component s (in cluding the metal). The other process involves electron beam induced heating. A higher temperature can also help the desorption of the volatile components. This raises the percentage of the non volatile components , which include the metal. Electron Energy The literature contains conflicting reports on the effect of electron energy on composition of the deposits. According to Weber , the metal content increases with decreased electron energy. 117 This result was found for Me 2 Au(tfac), Me 2 Au(hfac), Me 2 Au(acac), and Mo(CO) 6 , most likely as a result of longer exposure to accumulated charge. 24 The effect was postulated to result from a larger number of secondary electrons due to the lower energy of the primary electrons. Not all researchers h ave found the same correlation. Hoyle fabricated 50 nm thick film depos its from W(CO) 6 that showed no difference in composition with different electron energies. 118 Cicoria reported the deposition of rectangles from (RhCl( PF 3 ) 2 ) 2 to have no correlation with the electron energy. 119 Scan Pattern In some cases the scan pattern can have an effect on the composition of the deposit. Utke et al. patterned (hfac)Cu(VMTS) onto a copper surface in vacuum. 120 S elf standing horizontal rods were formed with all of the precursor elements present. On the

PAGE 93

93 end of one of the self standing rods, a vertical tip was grown. The t ip and surrounding area was composed mostly of Cu , with F , O, and Si not detected and the C content greatly reduced. T his was due to longer time under the beam for the tip than the thin rods. This causes a greater heat transfer from the beam, thus reaching higher temperatures which are more likely to decompose t he precursor into volatile and non volatile components. However, Cicoria 119 found no correlation between scan pattern and composition for the previously discussed rhodium rectangle deposits. Writing Direction The position of the gas nozzle can ha ve an effect on the composition of the deposit. Utke et al. found that when depositing a tip from Co 2 (CO) 8 , a cobalt free area was found directly in front of the gas nozzle. 121 T his effe ct was due to the cooling by the precursor gas. As the deposit cools it is less likely to desorb the volatile components, leaving a lower percentage of metal in that area. Post Deposition Processing Another class of methods to change the composition of the deposit is to treat the material after it has been deposited. These post deposition processes are of three general categories: annealing, extra irradiation, and extra irradiation while exposed to a reactive gas. Extra Irradiation For Fe 3 (CO) 12 deposits, it was found that upon extra irradiation , the conductivity increases greatly. 122 With W(CO) 6 , the effe ct was similar: there was a 20 f old decrease in resistivity. 118 A free standing wire of 10 nm width was grown from tetraethyl orthosilicate, 123 and then exposed to extra irradiation. The silicon content increased, while the wire narrowed to 1 nm thickness an d finally broke. This is an unwanted side

PAGE 94

94 effect that would have to be compensated for by forming larger deposits prior to the post deposition process. Extra irradiation improves the composition for the same reason that a higher current density gives impr oved composition. 24 The extra irradiation can cause further decomposition of the precursor as well as the formation of new volatile components from residual gas in the background. An additional effect is due to the electron beam induced heating, as discussed above. Extra Irradiation with Reactive Gas Another post deposition process that can improve the composition of the deposit involves further electron beam irradiation while exposi ng the de posit to a reactive gas. Noh et al. report the deposition of ruthenium from bis(ethylcyclopentadienyl) ruthenium(II). 29 The initial deposit was found to have 9 carbon atoms per ruthe nium atom. This was then exposed to an electron beam and oxygen gas to convert the deposited carbon into carbon dioxide (Figure 4 1 ). The carbon was removed, however the remaining ruthenium was found to be ruthenium(II) oxide rather than pure ruthenium met al. The deposit was also found by scanning electron micrographs to be porous if the processing occurred at a higher temperature.

PAGE 95

95 Figure 4 1 . Treatment of ruthenium carbide deposition with electron beam and oxygen gas. Ada pted from Figure 2 of R eference 29 . Annealing Annealing, the process of heating the deposit after deposition, can lead to improved composition. Fe(CO) 5 was used to deposit amorphous freest anding rods, which were annealed at 600 °C, after which the carbon and oxygen were barely detected while the shape remained largely unchanged. 124 In some cases , the deposit is known to incorporate some of the substrate, which increases levels of contaminants. 125 It has also been known to completel y destroy some nanostructures. When a pattern of platinum dots of a fe w nanometers in diameter was annealed at 800 ° C , the pattern was deformed. 24

PAGE 96

96 Reactive Gases Addition of another gas to the precursor during deposition can convert non volatile contam inants into volatile components and thus improve the percentage of the desired component (usually the metal) in the deposit. Matsui and Mori were able to improve the tungsten percentage from 95% to 100% by including H 2 along with the WCl 6 precursor. 126 Folch et al. reported deposition of gold from Me 2 Au(hfac), yielding a deposit of 2 3% gold. 127 When 130 mTorr of Me 2 Au(hfac) was combined with 3 Torr of water, gold content increased to 20%. When 2 Torr of oxygen and 8 Torr of argon were added , the per centage of Au went up to 50%. It was hypothesized that the water and oxygen are ionized by the beam and react with the carbon to form CO and CO 2 . The inclusion of oxygen or other reactive gases sometimes has no effect. Wang et al. deposited platinum from P t(PF 3 ) 4 with and without oxygen present. 128 T he oxygen had a negligible effect, increasing the platinum from 17% to 22%. Tseng found that mixing w ater with Me 3 PtCp had no effect on the composition of the deposit. 129 Precursor Identity The precursor identi ty is the most important factor determining the composition of the deposit. Nearly all of the precursors used for EBID are commercially available (Table 4 1) . van Dorp and Hagen reported the ten most popular precursors used in EBID up t o 2008. 24 The most common metal containing precursors are W(CO) 6 , Fe(CO) 5 , and CpPtMe 3 . The only purely inorganic precursor in the top ten is WF 6 . All of these are known from the CVD literature and many are commercially available.

PAGE 97

97 Table 4 1. Commonly used EBID precursors. Adapted from Table 1 on pages 30 31 of Reference 130 . Deposit Precursor Formula Vapor Pressure at Room Temperature Commercially Available Reference Al AlCl 3 Yes 130 Al(CH 3 ) 3 Yes 130 Au AuCl 3 Yes 130 Me 2 Au(hfac) 700 mTorr Yes 130 Me 2 Au(acac) 8 mTorr Yes 130 Me 2 Au(tfac) 40 mTorr Yes 130 PF 3 AuCl No 130 (CO)AuCl Yes 131 C C 8 H 8 10 25 Torr Yes 130 C 12 H 26 to C 18 H 38 Yes 130 C 16 H 10 Yes 130 C 3 H 4 O 2 Yes 130 C 3 H 6 O 2 Yes 130 C 2 H 4 O 2 Yes 130 CH 2 O 2 Yes 130 C 8 H 8 Yes 130 C 2 H 4 Yes 130 Co Co 2 (CO) 8 Yes 130 Co(CO) 3 NO Yes 132 Cr Cr(CO) 6 10 Torr Yes 130 Cu Cu(hfac) 2 0.004 mbar Yes 130 Cu(hfac)(VTMS) 420 Torr Yes 130 Cu(hfac)(MHY) 0.2 mbar No 130 Cu(hfac)(DMB) 1.3 mbar No 130 Fe Fe(CO) 5 3 Torr Yes 130 Fe(C 5 H 5 ) 2 Yes 130 Ga Ga(CH 3 )/AsH 3 Yes 130 D 2 GaN 3 No 130 Mo Mo(CO) 6 78 mTorr Yes 130 Ni Ni(CO) 4 10 Torr Yes 130 Ni(C 5 H 5 ) 2 17 mTorr Yes 130 Os Os 3 (CO) 12 Yes 130 Pd Pd(OOCCH 3 ) 2 Yes 130 Pd(C 3 H 5 )(C 5 H 5 ) No 130 Pt CpPtMe 3 54 mTorr Yes 130 ( MeCp )Pt Me 3 54 mTorr Yes 130 Pt(PF 3 ) 4 No 130 Re Re 2 (CO) 10 Yes 130 Rh [RhCl(PF 3 ) 2 ] 2 55 mTorr No 130 [RhCl(CO) 2 ] 2 0.25 Pa Yes 130

PAGE 98

98 Table 4 1. Continued Deposit Precursor Formula Vapor Pressure at Room Temperature Commercially Available Reference Ru Ru 3 (CO) 12 Yes 130 Ru(CH 3 CH 2 C 5 H 4 ) 2 Yes 29 Si SiH 2 Cl 2 Yes 130 SiO 2 Si(C 2 H 5 O) 4 1.5 Torr Yes 130 SiO x Si(OCH 3 ) 4 420 Torr Yes 130 W W(CO) 6 17 mTorr Yes 130 WF 6 Yes 130 WCl 6 Yes 130 A precursor for EBID needs the following properties: 24 , 133 1. Volatile at room temperature 2. Evaporate s completely without leaving residue 3. D ecompose s in a fast, clean, and selective manner 4. Stable to vacuum 5. Stable during storage and supply 6. Stable to small concentrations of water and air 7. Non toxic 8. Inexpensive No currently used precursor fits all of these req uirements. W(CO) 6 and PtCpMe 3 give deposits that are frequently around 10% metal. 24 The commonly used inorganic precursor , WF 6 , is stable and deposits 80% 100% metal, but is highly c orrosive. 134 The precursors AuCl(PF 3 ) and D 2 GaN 3 yield pure deposits and do not corrode the equipment , yet are highly unstable a nd can be hazardous . 24 Both AuCl 3 and AlCl 3 are known to form rods that contain only the carbon and silicon substrate. 135 The vast majority of the precursors use d for EBID are commercially available. Synthesis of novel complexes could expand the range of precursors and potentially develop precursors that are ideal for EBID. But, very little work has been done to determine what type of precursor is most favorable. A few studies regarding the

PAGE 99

99 mechanism of decomposition have been done that can help guide the design of future precursors. Luiser et al . repor ted a comparative study on four copper precursors: Cu(hfac) 2 , (hfac)Cu(MHY), (hfac)Cu(VTMS), and (hfac)Cu(DMB). 136 Cu(hfac) 2 and (hfac)Cu(MHY) both formed deposits of 13% 14% copper and 75% 80% carbon. For (hfac)Cu(VTMS) and (hfac)Cu(DMB) , the copper content was 15% 20% , while the carbon content was 60% 70%. None of the deposits showed the presence of fluorine. Thi s shows that the deposit composition is not entirely dependent on the stoichiometry of the precursor and that fluoride is not generally incorporated into deposits. The current models, rel ying largely on CVD, are inadequate to u nderstand and predict the decomposition under EBID. Density functional theory calculations showed the most favorable pathway for the thermal decomposition of (RhCl(PF 3 ) 2 ) 2 involves the loss of the PF 3 ligands. 137 However, when the precursor was deposited it was found to have 60% rhodium, 12% 25% phosphorus, 2% 13% chlorine, 3% 20% oxygen and nitrogen, and no fluorin e. If all PF 3 is lost as the intact ligand, there should be either no phosphorus or both phosphorus and fluorine present in the deposit if the ligand remained . The e beam induced decomposition must break the P F bond prior to breaking at least some of the Rh P bonds. Similar results were obtained with (RhCl(CO) 2 ) 2 with carbon being deposited but not oxygen. 119 This indicate s the breaking of the ca rbon oxygen bond during decomposition. Fairbrother et al. have done more extensive research on the de composition of carbonyl containing complexes using Co(CO) 3 NO as a model. 132 T he first step involved the ejection of one or more carbonyls and the cleavage of the nitrogen oxygen triple

PAGE 100

100 bond. This left the partially decarbonylated cobalt intermediate that contained some nitride. If the temperature is low , further decomposition leads to a carbon containing deposit. If the temperature is high, thermal desorption of carbonyl happens prior to the decomposition , lowering the carbon content and thus raising the cobalt percentage. Another study compared the composition of deposits from Me 2 Au(acac), Me 2 Au(tfac), and Me 2 Au(hfac). 117 The composition of the deposit for all precursors was 10% Au, 20% O, and 70% C. No fluorine was found in any of the deposits. This again shows that the deposit composition is not de pendent on the stoichiometric composition of the precursor. 24 The decomposition in the presence of the electron beam does not necessarily follow the thermal decomposition pathway. Mo re work is needed to understand the decomposition under these conditions. According to van Dorp and Hagen 24 beam induced deposition in devices is the lack of control over the composition of the adjusting the beam parameters and post proc precursor is mostly determined by the fact whether it is used often and whether it is readily available. Nearly all of the precursors used for deposition stem from the CVD world and there has been hardly any search for pr Design of Precursors for Electron Beam Induced Deposition T he field of EBID has proven very useful, but still has some fa ctors that are limiting its use; t he primary factor is th e amount of organic contamination in the deposit ion. One commonly used precursor is Me 2 Au(acac), which generally yield s <20% gold in the deposit . 138 The other >80% usually is composed of carbon and oxygen. The problem with this high level of contamination is the resulting deposition has a n

PAGE 101

101 increased resistivity. Platinum wires deposited from MeCpPtMe 3 by EBID generally or nanowires. 26 , 139 The primary reason for the residual organic contamination is the use of organometallic precursors designed for CVD , which is not a good model because the deposition pathway is likely much different. 26 There has been a limited amount of work on designing precursors specifically for EBID . In one experiment, chloro(trifluorophosphine)gold(I) was designed and used to deposit gold on carbon nanotubes , which afforded pure gold deposition . 140 However this precursor was unstable with a shelf life of only a few days, making this an unsuitable precursor for industrial application. Nevert heless this result shows that precursors designed for the purp ose of EBID can be effective and that more work is required in this field . In a typical EBID experiment a gas nozzle directs the precursor to the surface, where a focused electron beam induces d eposition (Figure 4 2 ) . Fairbrother is using a surface science approach to gain a better understanding of the reaction mechanism, by allowing more characterization techniques to be used (Figure 4 3). Figure 4 2 . The typical EBID experiment. Graphic prepared by Fairbrother and used with permission.

PAGE 102

102 a thin film of precursor is deposited on the cooled substrate in UHV. The precursor is then irradiated with a broad electron beam to induce deposition. The reaction and deposit can then be analyzed by XPS, RAIRS (reflection absorption infrared spectroscopy), and mass spectrometry. This technique allows an in depth look at the process as it is occurring. Figure 4 3 . The surface science approach to understanding the EBID decomposition pathway. Graphic prepared by Fairbrother and used with permission. This method has been used to determine the decomposition pathway of the very common precursor , tr imethyl(methylcyclopentadienyl)platinum (IV) ( 4 2 ) , used for making nanost ructures and nanowires . 141 Compound 4 2 is used frequently in CVD where pure platin um metal is deposited. When 4 2 is used for EBID , less than 20 % of the deposition is platinum , with the remainder largely being carbon. 26 Compound 4 2 was adsorbed onto a gold surface in a study conducted by Fairbrother et al.. 32 XPS was conducted before and after the electron beam irradiation. Upon irradiation, the carbon cont ent was lowered to 89%, with platinum making up the other 11% . The change corresponds to a loss of exactly one carbon atom for every precursor molecule (Figure 4 4 ) . This correlates well with the results from mass

PAGE 103

103 spectrometry that show the evol ution of methane upon electr on ir radiation (Figure 4 5 ). Essentially the same results were observed f or CpPtMe 3 , showing that the carbon lost is most likely one attached directly to the platinum. Figure 4 4 . The carbon:platinum ratio of MeCpPtMe 3 deposited by EBID as a function of electron radiation determined by XPS. Figure 4 5 . Mass spectrometry results from a) gas phase compound 4 2 , b) irradiated with electron beam, c) on the surface being radiated with electron beam, and d) reference spect rum of methane.

PAGE 104

104 These data indicate that one methyl group is lost, while the cyclopentadienyl ligand and the other methyl groups are trapped and the carbon is deposited along with the platinum. This shows that, at least in some cases, the methyl ligand ca n be an acceptable ligand for EBID. The cyclopentadienyl ligand , however , is not useful for EBID. Under standard laboratory conditions in solution, t he cyclopentadienyl ligand is 5 3 1 and then to labilize upon addi tion of ligand . An example of this was r eported by Casey and Connor in which Cp was removed from a rhenium complex by adding excess phosphine ligand and heating (Figure 4 6 ). 94 Figure 4 6 . The mechanism of loss of C p as reported by Casey and Connor. 94 The question remains which step is limit 5 3 3 1 transi tion that stops the reaction from occurring under these

PAGE 105

105 conditions? 3 form of C 3 allyl ligand. Allyl has been chosen as the ligand of choice t o further understand the decomposition mechanism of EBID. For this experiment to be e ffective, the other ligands would ideally be ligands with known reactivity in EBID. In 2013 , Fairbrother used the surface science approach to determine the reactivity of t ungsten hexacarbonyl in EBID. 31 It was determ ined by mass spectrometry that upon electron beam irradiation carbonyls were lost (Figure 4 7 ) . However, upon completion, the deposit was found to contain carbon and oxygen. This shows that some carbonyl s are ejected, but some still remain on the surface . Upon analysis , it was determined that 2.0 2.6 equivalents of carbonyl were eject ed per tungsten atom (Figure 4 8 ). This shows that, at least for this complex, more than two or three carbonyls can prove problematic. Therefore , the designed precursor should have three or less carbonyls present. Figure 4 7 . Mass spectrum of volatile species from ele ctron beam irradiation of tungsten hexacarbonyl.

PAGE 106

106 Figure 4 8 . Effect of electron beam on adsorbed tungsten hexacarbonyl. Some ligands remain but can be removed in post deposition processing. In 2011 , Fairbrother conducted an investigation on the mechanism of deposition from tetrakis(trifluorophosphine)platinum ( 4 3 ) . 142 Upon electron beam irradiation , one phosphine per platinum was lost. However, upon further radiation , the remaining phosphines decomposed and the fluoride ions were lost. The final compositio n of the deposit is platinum, phosphorus , and oxygen incorporated from water in the system (Figure 4 9 ) . However, when an actual EBID experiment was conducted, it was found that no post depos ition processing was needed to remove the fluoride ions, and there was no oxygen . This shows that fluoride, and probably other halides, are acceptable ligands for the EBID precursors. Figure 4 9 . The deposition and post deposition processing of platinum deposits from 4 3 . To summarize, an investigation of the decomposition pathway of a n allyl metal complex could provide useful information on the scope of useful ligands for EBID . Other acceptable ligands for this complex include carbonyls and halides. The complex is required to be volatile to be delivered to the surface . For this study, compound 4 1 ( Figure 4 10 ) was proposed.

PAGE 107

107 Figure 4 10 . Propo sed EBID precursors . This com plex fits the design parameters: its ligands are allyl, carbonyl and halide. There are three carbonyls present, which will need to be tested to determine whether there is still carbonyl contamination in the deposit . There is a hal ide ligand present which should be easily removed during deposition or post deposition processing. The volatility of this compound is not reported in the literature and must be tested. The halogen can be var ied to determine if this causes any e ffect in the deposition pathway. Ruthenium has not been used extensively in EBID experimen ts, however it does have some key applications. A thin layer of ruthenium has been known to shield a molybdenum silicon extreme ultraviolet reflective mirror. This can be use ful for extreme ultravio let lithography, which is used for forming >70 nm features on silicon wafers. 143 This mask would be essentially transparent to extreme ultraviolet, yet would provide good mechanical and chemical protection. EBID pro vides an attractive method of repairing this mask. Triruthenium dodecacarbonyl has been used for EBID, though no quantitative compositional analysis has been conducted. 144 In 2 014 , Noh et al. conducted an EBID experiment using bis(ethylcyclopentadienyl)ruthenium(II) as a precursor. 29 As expected, the resulting deposit had large amounts of carbon contamination, with an average composition of RuC 9 . After extensive treatment with oxygen gas and further electron irradiation, the carbon content was removed; however the deposition thickness was greater than it should be for pure ruthenium. The XPS thickness was consistent with a

PAGE 108

108 RuO 2 layer, not the desired layer of pure ruthenium metal. There is clearly considerable room for improvement of ruthenium EBID precursors . Synthesis of Precursors for Electron Beam Induced Deposition Compounds 4 1a through 4 1h (Fig ure 4 11 and Table 4 2 ) were synthesized using a modified prep from Sbrana et al. as reported in Chapter 3. 96 Compounds 4 1 b th rough 4 1e were synthesized as reported from the allyl halide and triruthenium dodecacarbonyl except sublimation resulted in higher purity (Figure 4 12 ) . Figure 4 11 . Precursors synthesized for EBID. Table 4 2 . Precursors synthesized for EBID and their volatility. R 1 R 2 X Compound Sublimation Temperature ( ° C) Sublimation Pressure (mTorr) H H F 4 1 a ND 4 ND H H Cl 4 1 b 27 30 90 H H Br 4 1 c 30 32 80 H H I 4 1 d 28 30 85 Me H Cl 4 1 e 32 35 85 Me H Br 4 1 f 32 35 95 H Me Cl 4 1 g 24 110 H Me Br 4 1 h 25 30 100 Figure 4 12 . Synthesis of compounds 4 1b through 4 1 e . Compound 4 1 a was syn thesized by first synthesizing t he allyl fluoride from allyl bromide and potassium fluoride as reported in the literature (Figure 4 13 ) . 145 The allyl 4 ND = not determined

PAGE 109

109 fluoride was collected in a pressure flask containing triruthenium dodecacarbonyl under liquid nitrogen , before capping and slowly warming to 60 ° C to react for 6 hours. The reaction afforded a small amount of the desired product. The reason for the lower yield is likely due to the gaseous n ature of allyl fluoride as opposed to the liquid allyl chloride, bromide, and iodide. This limited the interaction with the ruthenium and thus thermal decomposition occurred faster. The product was clearly identifiable by 1 H and 13 C NMR, but the elemental analysis showed significantly higher carbon content than sh ould be present. Figure 4 13 . Synthesis of 4 1 a from allyl bromide and triruthenium dodecacarbonyl. Figure 4 14 . Synthesis of 4 1 f . Compound 4 1 f was synth esized in the same manner as 4 1 a e (Figure 4 14 ). Compound 4 1 g was sy nthesized in a similar manner to form the asymmetric allyl (Figure 4 15 ) . Complex 4 1g has been rep orted in the literature where it was synthesized from a butadiene complex and hydrogen chloride ( Figure 4 16 ) . 146 The synthesis reported in this dissertation is superior because it is a one step synthesis from commercially availab le starting materials.

PAGE 110

110 Figure 4 15 . Synthesis of compound 4 1 g . Figure 4 16 . Literature synthesis of 4 1 g . Compound 4 1h was synthesized from the 1 halo 2 b utene as opposed to the 3 halo 1 butene used in the synthesis of 4 1 g (Figure 4 17 ) . This shows that either starting material c an form the same product. C omplex 4 1h showed similar properties to 4 1 g . Figure 4 17 . Synthesis of compound 4 1 h . Co mpounds 4 1 b and 4 1 c were chosen for the EBID mechanistic studies. These complexes sufficiently volatile to be useful while contain ing fewer carbon s than the methyl allyl derivative s and are thus likely to produce less carbon contamination in the deposit . These complexes are representative of the precursors containing other halogens and should provide enough information to determine if the identity of the halogen is important. If the identity of the halogen does prove important other compounds will be teste d. Surface Science Studies of Precursors Compound 4 1 c was placed in the dosing vial and exposed to UHV conditions (10 9 Torr) while heating slowly to 40 ° C. This was then exposed to the surface of highly

PAGE 111

111 ordered pyrolytic graphite (HOPG) that had been cooled by liquid nitrogen to 168 °C , allowing the precursor to adsorb onto the carbon surface. The mass spectrum was then taken of the compound desorbing from the surface (Figure 4 18) , which shows some bromide, allyl, carbo nyl, and smaller fragments. The surface was then irradiated with an electron beam (9.36 x 10 16 e /cm 2 ) and the mass spectrum was taken of the species vo latilized during this process (Figure 4 19 ). The major fragment observed is carbonyl. Some of the other fragments (CO 2 and H 2 ) are attributed to the system. The precursor adsorbed onto the surface was analyzed by XPS, showing the unreacted precursor . The surface was then irradiated with increasing number of electrons (Figure 4 20) . The non irradiated surfac e showed at least two different forms of carbon (carbonyl and HOPG/allyl). The carbonyl peak disappears quickly upon irradiation while the other peak remains unchanged. This indicate s that the carbonyl is ejected upon irradiation while the ally l remains. T his agrees with the mass spectrometry , which showed carbonyl upon electron beam irradiation but no allyl fragments. However, this is not conclusive due to the large presence of carbon in the substrate, making detection of carbon deposits difficult.

PAGE 112

112 Figure 4 18 . Mass spectrum of gas phase 4 1 c . Figure 4 19 . The mass spectrum of volatilized species from electron beam irradiation of 4 1 c on HOPG surface.

PAGE 113

113 Figure 4 20 . XPS data of 4 1 c on HOP G with increasing electron irradiation dosage. Close ups of carbon 1s and ruthenium 3d (left), oxygen 1s (middle), and bromine 3d (right) regions. The ruthenium very clearly shifts toward lower binding energies upon electron irradiation , ending at 280.9 eV compared to metallic ruthenium at 280.1 eV . The oxygen p eak from carbonyl shrinks upon irradiation and nearly disappears upon higher dosage , all while shifting its binding energy . The bromine peak , however , seems to remain largely unchanged upon further irradiation. This leads to the conclusion that the carbony l leaves upon irradiation. This data indicate s that nearly all of it has disassociated as oppose d to only 2.0 2.6 equivalents previously reported for tungsten hexacarbonyl. It is possible that the bromide ruthenium bonds are breaking upon irradiation, but it is clear that the bromine is remaining on the surface.

PAGE 114

114 Figure 4 21 shows the changes in area of the O(1s) and Br(3d) regions as a function of electron dose, as well as the peak position from the Ru(3d) peak . The oxygen peak decreases at the same rate the ruthenium peak shifts, indicating the same process that removes the oxygen is making the ruthenium more metallic. However , it also s hows that the bromine is not removed at this point in the irradiation . These results are consistent with th e electron do se fully decomposing the precursor to a decarbonylated matrix containing ruthenium, carbon, and bromine. Figure 4 21 . Changes in the fractional coverage of adsorbed oxygen and bromine atoms and changes in Ru 3d 5/2 peak position for 1 2 nm 4 1 c films plo tted as a function of electron dose. Each relative concentration (O/O t=0 , Br/Br t=0 ) and peak position was determined by XPS. The initial O t= 0 and Br t=0 values were measured prior to electron irradiation. Fil ms were adsorbed onto HOPG at 168 ° C and ex posed to 500 eV electrons. Next, extra irradiation was applied in a post deposit ion process. With enough electron radiation , the bromine was removed. This was observed by XPS as the peak area decreased. T he peak position for ru thenium 3d shifted at the same rate as the

PAGE 115

115 bromine peak decreased , showing the same process that removes the bromide is converting the ruthenium to be more metallic (Figure 4 22) . Figure 4 22. Post deposition removal of bromine via extra irradiation as determined by XPS peak hei ght. Figure 4 23. Mass spectrum of irradiated 4 1 b on HOPG.

PAGE 116

116 The chloride complex , 4 1 b , was used fo r the same experiments. T he precursor was first adsorbed onto the surface under the same conditions as 4 1 c and then irradiated with an electron beam (7.4 9 x 10 16 e /cm 2 ) . Mass spectra of the v olatilized species were then obtained (Figure 4 23 ). Figure 4 24. Ruthenium peak po sition and oxygen peak area measured by XPS as a function of irradiation time. The XPS was taken of the deposit on the surface afte r increased irradiation times. The peak position of ruthenium was observed to move toward a more metallic binding energy upon further irradiation. The carbon and the oxygen from the carbonyl were observed to disappear . The chlorine peak decreased in size at firs t, but then remained unchanged, demonstrating that the oxygen loss is happening at the same rate as the ruthenium changes. This indicates the same process is involved in both carbonyl loss and conversion of ruthenium. Figure 4 25 shows the plot of chlorine peak area as a

PAGE 117

117 fu nction of irradiation time, showing initial loss of some chlorine, but not at the same ra te as the loss of carbonyl. These results indicate that the process involving loss of chlorine is different than the process involving loss o f carbonyl and conversion of ruthenium. Upon further irradiation the chlorin e was loss i n a similar manner to the bromin e from 4 1b (Figure 4 26). Fi gure 4 25 . Chlorine peak area for XPS of 4 1b deposited on HOPG as a function of irradiation time. Figure 4 26. Chlorine area from XPS of 4 1b deposited on HOPG plotted as a function of elec tron dose, i ndicating loss of chlor ine upon post deposition irradiation.

PAGE 118

118 To summarize, the d ata indicate that the electron stimulated reactions for both precursors are essentially the sa me. The first stage is electron stimulated carbonyl desorption and allyl dec omposition. The XPS data indicate that approximately 80% of the oxygen is removed; the remaining oxygen could originate from the precursor or could be from residual water in the ch amber. The mass spectrum showed the evolution of CO during irradiation, again showing desorption of the carbonyl. In contrast, there was no sign of allyl desorption during the irradiation. The ha logen content in both cases remains largely unchanged during this stage. The ruthenium at this time is reduced nearly to metallic ruthenium , based o n the binding energy in the XPS . The next stag e of the experiment is electron stimulated desorption of halog ens. Upon post deposi tion processing, both the bromin e and th e chlorin e were removed upon increased electron flux. This was evident from the decreased peak area in the XPS of the halogen peak. The ruthenium binding energy was shifted during loss of halogen fr om 280.9 eV to 280.5 eV after 80% of the halogen was lost. This is nearly the binding energy of metallic ruthenium (280.1 eV). Summary of EBID Results The results from this study can be used in the design of future EBID precursors. First, several previ ously reported results have been confirmed. The carbonyl has been shown to volatilize upon irradiation. The compounds tested have three carbonyls and all, or nearly all , of the carbonyls were lost upon irradiation. The halogens, in this case bromine and ch lorine, were found to remain on the surface upon initial irradiation. However, upon further post deposition irradiation, the halogens were remov ed and the ruthenium was reduced. The decomposition could be step wise, the carbonyl must leave before the halog en, or the loss of halogen could simply be a slower process. More

PAGE 119

119 studies will need to be performed to determine which of these pathways is appropriate. These results do confirm that carbonyls and halogens can be good ligands for EBID precursors. Pr evious results have shown that C p ligands are inappropriate for use in EBID . These results indicate that 3 C p, is also not appropriate. There is no evidence of its loss upon i rradiation and the deposit appears to be a matrix containi ng carbon and ruthenium. However , this precursor shows significant improvement over the most recently reported ruthenium precursor , (EtCp) 2 Ru, which formed a layer of RuC 9 ; 29 if all of the allyl were retained the product would be a layer of RuC 3 . These results illustrate the advantage of designing precursors for EBID. The next logical step would be to synthesize a precursor containing only the proven ligands, carbonyl and halides. This prec ursor should give pure ruthenium deposition.

PAGE 120

120 CHAPTER 5 EXPERIMENTAL General Procedures Unless specified otherwise, all manipulations were performed under an inert atmosphere (N 2 ) using standard Schlenk or glovebox techniques. All reaction solve nts were p urified using an MBrau n MB SP solvent pu rification system prior to use. 1 H and 13 C NMR spectra were recorded on Gemini 300, Mercury 300, or VXR 300 spectrometers using residual protons of deu terated solvents for reference. All chemicals were used as receiv ed without further purification. Synthesis Synthesis of 2,2' b ipyridine N oxide ( 2 5) A sol ution of 2,2' bipyridine (7.00 g, 44.8 mmol) in CH 2 Cl 2 (50 mL) was prepared in a 250 mL round bottom flask. In a 100 mL round bottom flask , a solution of m chloroperb enzoic acid (11.02 g, 77%, 49.18 mmol) in CH 2 Cl 2 (80 mL) was prepared . This was transferred to an addition funnel and ad ded dropwise to the bipyridine solution at 0 °C . Upon completion of the addition the solution was slightly yellow . The solutio n was left to warm to room temperature and to stir overnight , at which time the solution had become a milky white color . The solution was c ooled in an ice bath and 100 mL of 2M NaOH was added . The organics were separated , extracted twice more , and dried wi th anhydrous Na 2 CO 3 . The solvent was removed by rotary evaporation . Off white solid was obtained and recrystallized by dissolving in hot ether and adding hexanes . The s olution was cooled overnight in the freezer and filtered to obtain 4.21 g (55%) of white product. The product was characterized by comparison to literature data . 60

PAGE 121

121 Synthesis of 6 cyano 2,2' bipyridine ( 2 6) A solution of 2,2' b ipyridine N oxide (1.0366 g, 6.02 04 mmol) in CH 2 Cl 2 (12 mL) was stirred at room temperature . A syringe of Me 3 SiCN (0.83 mL, 6.6 mmol) was added very carefully to the so lution . After 5 10 minutes , d ime thyl carba m o yl chloride (0.61 mL, 6.6 mmol) was added via syringe . The orange solutio n was left to stir under N 2 in the hood for 7 days . After 2 3 days the solution had become darker . At the end of the 7 days , ether (80 mL) was added . The solution bubbled and a white solid precipitate d . The so lution was transferred to a sepa ratory funnel and washed with 5% NaHCO 3 (2 x 50 mL) and brine (2 x 50 mL) . The solution was dried using anhydrous Na 2 CO 3 and solvent was evaporated using the rotary evaporator . The crude solid was recrystallized using methylene chloride /hexanes to obtain 0 .9805 g (90%) of pure product. The product was characterized by comparison to literature data . 60 Synthesis of 6 methoxycarbonyl 2,2' bipyridine ( 2 7) A soluti on of 6 cyano 2,2' bipyridine (0.5990 g, 3.306 mmol) in methanol (50 mL) was stirred in a 250 mL round bottom flask . Hydrochloric acid (10 mL, 12 M) was added and solution became warm . The s olution was refluxed for 24 hours under N 2 . After 24 hours , the solution was cooled to room temperature , and the solution was concentrate d to 10 mL . The s olution was cooled in an ice bath and 30 mL of H 2 O was added . Sodium carbonate was added until the pH reached 7 . The solution was then extracted with methylene chloride (4 x 50 mL) and evaporated to yield an off white solid (0.6713 g, 94 % crude yield) , w hich was purified by s ublimation orr pressure and cooled using a dry ice/acetone bath . Yield: 61% of a pure white solid. The product was characterized by com parison to literature data. 60

PAGE 122

122 Synthesis of 6 hydrox ymethyl 2,2' bipyridine ( 2 8) A solution of 6 methoxycarbonyl 2,2' bipyridine (0.4578 g, 2.13 7 mmol) in ethanol (15 mL) was stirred at room temperature while NaBH 4 (0.2565 g, 6.78 0 mmol) was added . This solution was then refluxed for 14 hours . The solution was allowed to cool, then water (1.4 mL) and sulfuric acid (0.4 mL) were added . The solution was then stirred at room temperature for 12 hours , and 2 M NaOH solution was added until the pH reached 9 10 . Addition of NaOH caused a white solid to precipitate . This solution was stirred for 2 more hours and then filtered , and concentrate d to a minimum on the rotatory evaporator . The remaining solution was extracted using 1% MeOH in CH 2 Cl 2 (3 x 10 mL ) . The organics were combined and evaporated to afford a dark oil (0.3310 g, 83% y ield) behind . The product was characterized by comparison to literature data . 60 Synthesis of 2 9 6 hydroxymethyl 2,2' bipyridine ( 2 8 ) (0.3310 g, 1.7 7 8 mmol), t butyldimethylsilyl chloride (0.2963 g, 1.966 mmol) and a catalytic amount of DMAP were dissolved in CH 2 Cl 2 (40 mL) . T riethyla mine (0.30 mL, 2.1 mmol) was added . The solution was then stirred and the solution was stirred at room temperature overnight . The solutio n was then transferred to a sepa ratory funnel and washed with DI water (20 mL) . The aqueous layer was extracted with methylene chloride (2 x 20 mL) . The organics were combined and washed with saturated ammonium chloride solution (15 mL) and dried with MgSO 4 . The solvent was then removed on the rotary evaporator to afford a light brow n oil, which was chromatographed on alumina with methylene chloride as the eluent. The product was obtained as a white/clear oil (0.2256 g, 44 %). 1 H NMR (300 MHz , CDCl 3 ) 8.69 (ddd, J = 0.9, 1.7, 4.8 Hz, 1 H), 8.37 (dt, J = 1.1, 7.9 Hz, 1 H), 8.23 (d, J = 7.8 Hz, 1 H), 7.84 (t, J = 7.8 Hz, 1 H), 7.81 (dd, J = 2.0, 4.1 Hz, 1 H), 7.54 (dd, J = 1.1, 7.7 Hz, 1 H),

PAGE 123

123 7.30 (ddd, J = 1.2, 4.8, 7.5 Hz, 1 H), 4.93 (s, 2 H), 0.99 (s, 9H), 0.16 (s, 6 H) . 13 C NMR (75 MHz, CDCl 3 ) 160.8, 156.2, 155.0, 149.0, 137.3, 136.6, 123. 4, 121.0, 119.9, 119.1, 66.2, 25.9, 18.3, 5.4 . Synthesis of 2 10 To a dry, flushed S chlenk flask , 10 mL of dry acetonitrile was added. Palla dium(II) c hloride (0.3020 g, 1.70 3 mmol) was added while stirring before refluxing for 1 hour . Much of the solid dissolved and what remained was an orange color as opposed to the dark brown color of the starting material . The solution was cooled t o room temperature and N,N,N',N' tetrame thylethanediamine (0.39 mL, 2.6 mmol) was added . Immediately , a y ellow green precipitate formed. The solution was filtered and the solid washed with ether and air dried . The yellow green solid (0.4 593 g, 92 %) was insoluble in most organic solvents . A melting point of the product was obtained, then the compound was use d without purification . 62 Synthesis of 2 11 Compound 2 10 (0.222 g, 0.756 m mol) was placed in a dried and flushed S chlenk flask and dry ether (5 mL) was added . The mixture was cooled to 40 an acetonitrile/dry ice bath . At this point , a 1.5 M solution of methyl lithium in ether (1.09 mL, 1.6 mmol) was added . Th e solution was stirred for approximately 5 minutes and then the acetonitrile/dry ice bath was replaced by an ice water bath . The reaction mixture slow ly changed from a yellow to a gray . A small amount of ice cold water was added after t he mixture had stirred for approximately 1 hour . The solutio n was then transferred to a sepa ratory funnel . The aqueous layer was extracted with ether (2 x 10 mL) and the organics were combined and washed with brine (15 mL) . The organics

PAGE 124

124 were evaporated using the rotary evaporator to obtain a wh ite solid (0.141 g, 7 4% ) . The product was characterized by comparison to literature data . 62 Synthesis of 2 14 A Schlenk flask was prepared containi ng compound 2 18 (0.0611 g, 0.128 mmol), methylboronic acid (0.0303 g, 0.506 mmol) and cesium carb onate (0.1729 g, 0.5307 mmol). Dry THF (12 mL) was added and the solution was stirred for 2 hours. The s olven t was removed under vacuum. The o range solid was then redissolved in methylene chloride and filtered. Solvent was removed under vacuum to obtain the product as an orange solid (0.0412 g, 74 %). 1 H NMR (300 MHz, CDCl 3 ) 9.16 (d, J = 3.9 Hz, 1H), 7.82 8.07 (m, 5H), 7.57 (ddd, J = 1.1, 5.9, 7.3 Hz, 1H), 5.50 (s, 2H), 0.98 (s, 9H), 0.37 ( s, 3H), 0.33 (s, 3H), 0.20 (s, 6 H) . 13 C NMR (75 MHz, CDCl 3 156.7, 155.7, 151.0, 140.1, 140.4, 125.7, 122.6, 121.4, 66.6, 25.9, 18.2, 5.3, 5.8, 6.2. Synthesis of 2 15 Dry acetonitrile (30 mL) was adde d to a dried Schlenk flask under nitrogen. Palladium (II) chloride (0.3168 g, 1.787 mmol) was added while stirring. The dark brown solution was refluxed until it became a lighter orange color. The solution was then co oled to room temperature and 2 13 (0. 492 3 g, 2.672 mmol) was added. After stirring for approximately ten minutes , the solution was filtered to obtain a light yellow powder (0.6240 g, 97 %). The product was characterized by comparison to literature data. 147 Synthesis of 2 18 Dry acetonitrile (10 mL) was added to a dried Schlenk flask under nitrogen. Palladium(II) chloride (0.0284 g, 0.160 m mol ) was added while stirring. The dark brown solution was refluxed until it became a lighter orange color. The solution was then cooled to room temperature and a solution of 2 9 in CH 2 Cl 2 (1.02 mL, 0.16 M, 0.16

PAGE 125

125 mmol) was added . After stirring for approximate ly 10 minutes , the solven t was removed under vacuum. Chloroform was added to the o range solid and the solution was then filtered. The solvent was removed by rotary evaporation to obtain the product (0.0750 g, 98 % yield) as an orange solid. 1 H NMR (300 MHz, CDCl 3 ) 9.28 (d, J = 5.7 Hz, 1H), 8.07 8.17 (m, 4H), 7.87 (dd, J = 2.2, 7.2 Hz, 1H), 7.45 (ddd, J = 1.5, 5.7, 7.2 Hz, 1H), 5.44 (s, 2H), 0.95 (s, 9H), 0.16 (s, 6H) . 13 C NMR (75 MHz, CDCl 3 ) 169.4, 156.9, 155.5, 151.2, 140.3, 140.0, 125.8, 122.9, 121. 0, 66.7, 25.9, 18.2, 5.3 . Synthesis of 2 1a Palladium (II) chloride (0.1080 g, 0.6091 mm ol) was combined with 2 8 (0.0519 g, 0. 279 mmol) in 7 mL of methanol. The s olution was refluxed for 15 hours and the n allowed to cool to room temperature. The s olvent was removed via reduced pressure to afford an orange solid . 1 H NMR (300 MHz, DMSO d 6 ) 8.88 (br s , 1H), 8.52 (d, J = 8.2 Hz, 1H), 8.45 (d, J = 6.5 Hz, 1H), 8.29 8.37 (m, 2H), 7.84 (d, J = 8.2 Hz, 1H), 7.74 (dd, J = 6.5 Hz, 1H), 5.20 (br s , 2H) . 13 C NMR (75 MHz, DMSO d 6 151.5, 140.1, 140.0, 125.9, 123.0, 121.1, 66.7. Anal. Calcd for C 11 H 10 Cl 2 N 2 O Pd : C, 36.34 ; H, 2. 77 ; N, 7.7 1 . Found: C, 34.36 ; H, 2.86 ; N, 7.01 . Synthesis of [RuCp(CO) 2 ] 2 ( 2 20) Ru 3 (CO) 12 (1.5122 g, 2.3652 mmol) was added to 50 mL of dry heptane in the glovebox . Freshly cracked cyclopentadiene (2.4 mL, 28 mmol) was added to the stirred suspension. The suspension was refluxed for 1 h ou r. The reaction mixture was then opened to the atmosphere and the solution was concentrated heptane (50 mL) was added, the solution was refluxed for 2 h ou rs . The solution was allowed to cool, then filtered . The precipitate was washed with hexanes to obtain the

PAGE 126

126 product as a brown powder (1.2907 g, 82 % yield). The product was characterized by comparison to literature data . 148 Synthesis of CpRu(CO) 2 Me ( 2 2) Sodium (0.3914 g, 17.02 mmol) was added to a three neck ed round bottom flask and dry toluene (10 mL) was added to the flask. Toluene was refluxed to obtain molten sodium. Mercury (0.96 mL) was very slowly and carefully added to the molten sodium. The toluene was removed by heating and transferring to a Dean St ark trap. The sodium mercury amalgam was then allowed to cool to room temperature. At room temperature it became solid. A THF solution of [CpRu(CO) 2 ] 2 (1.2184 g, 2.7419 mmol) was then transferred from a dry Schlenk flask to the amalgam. The solution was st irred overnight. The o rganic layer was then transferred to another dry flask and methyl iodide (0.85 mL, 14 mmol) was added. The s olution was left to stir for 1 h ou r. The s olvent was then carefully removed via trap to trap distillation to obtain a d ark bro wn solid. The solid was sublimed beginning at 100 mT orr and room temperature and warming up to 40 ° C to yield pure 2 2 as a white solid (0.5665 g, 43% yield) . The product was characterized by comparison to literature data . 65 Synthesis of [Cp*RuCl 2 ] 2 ( 2 22) Ruthenium (III) chloride hydrate (0.4916 g, 2.370 mmol) was added to ethanol (10 mL). Pentamethylcyclopentadiene (0.85 mL, 5.4 mmol) was then added to the solution and refluxed for 3 h ou rs. The solution was allowed to cool to room temperature and sti rred overnight . The dark powder was filtered and washed by ethanol and diethyl ether to obtain pure 2 22 (0.4304 g, 59 % yield). The product was characterized by comparison to literature data . 149 , 150

PAGE 127

127 Synthesis of (C 5 Me 4 CH 2 Cl)Ru(CO) 2 Cl ( 2 23) [(C 5 Me 5 )RuCl 2 ] 2 (0.8154 g, 1.327 mmol ) was dissolved in methylene chloride while open to the atmosphere. Carbon monoxide gas was bubbled through the solution for 3 0 minutes before the solution was stirred for 2 h ou rs. The solvent was then removed under vacuum to obtain a dark solid. This was then dissolved in diethyl ether, filtered and the filtrate was concentrated under reduced pressure. Product was then purified via column chromatography using Florisil and diethyl ether as an eluent. The solvent was removed to obta in the product as a yellow solid (0.1284 g, 13 % yield) . The product was characterized by comparison to literature data . 66 Synthesis of (C 5 Me 4 CH 2 OH )Ru(CO) 2 Cl ( 2 3) (C 5 Me 4 CH 2 Cl)Ru(CO) 2 Cl ( 0.3092 g, 0.8536 mmol) was dissolved in THF (11 mL ) and d eionized water (1.4 mL ). Collidine (0.12 mL , 0.91 mmol) was added and the solution was refluxed for 2 hours. The solution was concentrated to a minimum under vacuum. The solution was extracted with ether (3 x 10 mL ) and the solvent was removed under vacuum. The p roduct was chromatographed on silica with ether as the eluent. The product dried under vacuum to obtain 2 3 as a yellow solid (0.1570 g, 53 %). The product was characterized by comparison to literatur e data. 151 Synthesis of 3 C 3 H 5 Ru(CO) 3 Br ( 4 1c) The precursor was synthesized using a modified literature procedure. 96 Ru 3 (CO) 12 (1.0 073 g, 1.5756 mmol ) was added to 20 mL of 2,2,4 trimethylpentane under nitrogen. Allyl bromide (5.0 mL , 58 mmol) was added and the mixture was immediately refluxed. Upon heating , the Ru 3 (CO) 12 dissolved turning the solution deep r ed. After 20 minutes of refluxing , the solution turned yellow and the solvent was re moved in vacuo. The yellow crude product was sublimed at 30 ° C and 80 mTorr to

PAGE 128

128 obtain 4 1c as a white solid (1.2907 g, 89 %). 1 H NMR (CDCl 3 J = 13 .1, 1.0 Hz), 4.11 (dt, 2H, J = 7.8, 1.1 Hz), 5.17 (tt, 1H, J = 13.1, 7.8 Hz). The compound was identified by comparison to literature data. 96 Synthesis of 3 C 3 H 5 Ru(CO) 3 Cl ( 4 1b) The precursor was synthesized using a modified literature procedure. 9 6 Ru 3 (CO) 12 (0.9 799 g, 1.533 mmol ) was added to 20 mL of 2,2,4 trimethylpentane under nitrogen. Ally l chloride (4.6 mL , 56 mmol ) was added and the mixture was immediately refluxed. Upon heating , the Ru 3 (CO) 12 dissolved turning the solution deep red. Aft er 50 minutes of refluxing , the solution turned yellow and the solvent was removed in vacuo. The yellow crude product was sublimed at 30 ° C and 90 mTorr to obtain 4 1b as a white solid (0.7967 g, 66%). 1 H NMR (CDCl 3 = 13.3, 1.5 Hz), 4.19 (dd, 2H, J = 7.9, 1.5 Hz), 5.30 (ttd, 1H, J = 13.3, 7.9, 1.5 Hz). The compound was identified by comparison to literature data. 96 Synthesis of 3 C 3 H 5 Ru(CO) 3 I ( 4 1d) The precursor was synthesized using a modified literature procedure. 9 6 Ru 3 (CO) 12 (0.1137 g, 0.1178 mmol) was added to 20 mL of 2,2,4 trimethylpentane under nitrogen. Allyl iodide ( 0.60 mL , 3.0 mmol) was added and the mixture was immediately refluxed. Upon heating , the Ru 3 (CO) 12 dissolved turning the solution deep red. Af ter 20 minutes of refl uxing , the solution turned orange and the solvent was removed in vacuo. The orange crude product was sublimed at 28 ° C and 85 mTorr to obtain 4 1d as a pale orange solid ( 0.1337 g, 71 %). 1 H NMR (CDCl 3 3.48 (d, 2H, J = 13.0 Hz), 3.96 (d, 2H, J = 7.9 Hz), 4.93 (tt, 1H, J = 13.0, 7.9 Hz) . The compound was identified by comparison to literature data. 96

PAGE 129

129 Synthesis of 3 C 4 H 7 Ru(CO) 3 Cl ( 4 1e) The precursor was synthesized using a modified literature procedure. 96 Ru 3 (C O) 12 ( 0.1024 g, 0.1602 mmol) was added to 20 mL of 2,2,4 trimethylpentane under nitrogen. 3 C hloro 2 methyl 1 propene (0.58 mL , 5.9 mmol) was added and the mixture was immediately refluxed. Upon heating , the Ru 3 (CO) 12 dissolved turning the solution deep red. After 5 0 minutes of refluxing , the solution turned yellow and the solvent was re moved in vacuo. The yellow crude product was sublimed at 32 ° C and 85 mTorr to obtain 4 1e as a white solid ( 0.0901 g, 68 %). 1 H NMR (C DCl 3 2.14 (s, 3H), 2.92 (s, 2H), 4.06 (s, 2H). The compound was identified by comparison to literature data. 96 Synthesis of 3 C 4 H 7 Ru(CO) 3 Br ( 4 1f) Ru 3 (CO) 12 ( 0.0864 g, 0.135 mmol) was added to 20 mL of 2,2,4 trimethylpentane under nitrogen. 3 B romo 2 methyl 1 propene (0.50 mL , 5.0 mmol) was added and the mixture was immediately refluxed. Upon heating , the Ru 3 (CO) 12 dissolved turning the solution deep red. After 20 minutes of refluxing , the solution turned yellow and the solvent wa s removed in vacuo. The yellow crude product was sublimed at 32 35 ° C and 95 mTorr to obtain 4 1f as a white solid ( 0.0612 g, 47 %). 1 H N MR (CDCl 3 2.15 (s, 3H), 3.13 (s, 2H), 4.00 (s, 2H) . 13 C NMR (75 MHz, CDCl 3 ) 25.6, 59.2, 188.9 . Anal. Calcd for C 7 H 7 BrO 3 Ru: C, 26.26 ; H, 2.20; N, 0.00. Found: C, 26.83; H, 2.20; N, 0.01 . Synthesis of 3 C 4 H 7 Ru(CO) 3 Cl ( 4 1g) Ru 3 (CO) 12 ( 0.1579 g, 0.2470 mmol) was added to 20 mL of 2,2,4 trimethylpentane under nitrogen. 3 C hl oro 1 butene (0.92 mL , 9.1 mmol) was added and the mixture was immediately refluxed. Upon heating , the Ru 3 (CO) 12 dissolved

PAGE 130

130 turning the solution deep red. After 5 0 minutes of refluxing , the solution turned yellow and the solvent was removed in vacuo. The yellow crude product was sublimed at 25 °C and 11 0 mTorr to obtain 4 1g as a white solid ( 0.0687 g, 34 %). 1 H NMR (CDCl 3 , 300 2.05 (d, 3H, J = 6.2 Hz), 2.63 (d, 1H, J = 12.5 Hz), 3.90 (d, 1H, J = 7.9 Hz), 4.00 (dq, 1H, J = 12.5, 6.1 Hz), 5.22 (td, 1H, J = 12.5, 7.9 Hz) ppm . 13 C NMR (75 MHz, CDCl 3 ) 20.5, 55.8, 83.7, 108.3 ppm. Anal. Calcd for C 7 H 7 ClO 3 Ru: C, 26.26; H, 2.20; N, 0.00. Found: C, 26.47; H, 1.81; N, 0.13. Synt hesis of 3 C 4 H 7 Ru(CO) 3 Br ( 4 1h) Ru 3 (CO) 12 ( 0.1312 g, 0.2052 mmol) was added to 20 mL of 2,2,4 trimethylpentane under nitrogen. Crotyl bromide (0.79 mL , 7.6 mmol) was added and the mixture was immediately refluxed. Upon heating , the Ru 3 (CO) 12 dissolved turning the solution deep red. After 3 0 minutes of refluxing , the solution turned yellow and the solvent was removed in vacuo. The yellow crude product was sublimed at 25 °C and 10 0 mTorr to obtain 4 1h as a white solid ( 0.1263 g, 64 %). 1 H NMR ( CDCl 3 , 300 MHz): 2.01 (d, 3H, J = 6.2 Hz), 2.83 (d, 1H, J = 12.4 Hz), 3.83 (d, 1H, J = 7.9 Hz), 4.19 (dq, 1H, J = 12.4, 6.2 Hz), 5.09 (td, 1H, J = 12.4, 7.9 Hz) ppm. 13 C NMR (75 MHz, CDCl 3 ) 20.5, 52.9, 80.9, 107.8 ppm. Anal. Calcd for C 7 H 7 BrO 3 Ru: C, 30.50; H, 2.56; N, 0.00. Found: C, 31.88; H, 2.41; N, 0.02 .

PAGE 131

131 LIST OF REFERENCES (1) Gaur, R.; Mishra, L.; Siddiqi, M. A.; Atakan, B. RSC Adv. 2014 , 4 , 33785. (2) McElwee White, L. Dalton Trans. 2006 , 5327. (3) Hubert Pfalzgraf, L. G.; Guillon, H. Appl. Organomet. Chem. 1998 , 12 , 221. (4) Tiitta, M.; Niinisto, L. Chem. Vap. Deposition 1997 , 3 , 167. (5) Dhar, S.; Dharmaprakash, M. S.; Shivashankar, S. A. Bull. Mat. Sci. 2008 , 31 , 67. (6) Dhar, S.; Shalini, K.; Shivashankar, S. Bull. Mat. Sci. 2008 , 31 , 723. (7) Liu, Y. H.; Cheng, Y. C.; Tung, Y. L.; Chi, Y.; Chen, Y. L.; Liu, C. S.; Peng, S. M.; Lee, G. H. J. Mater. Chem. 2003 , 13 , 135. (8) Siddiqi, M. A.; Siddiqui, R. A.; Atakan, B.; Roth, N.; Lang, H. Materials 2010 , 3 , 1172. (9) Tuchscherer, A.; Georgi, C.; Roth, N.; Schaarschmidt, D.; Rueffer, T.; Waechtler, T.; Schulz, S. E.; Oswald, S.; Gessner, T.; Lang, H. Eur. J. Inorg. Chem. 2012 , 4867. (10) Samuels, J. A.; Folting, K.; Huffman, J. C.; Caulton , K. G. Chem. Mater. 1995 , 7 , 929. (11) Jones, A. C. J. Mater. Chem. 2002 , 12 , 2576. (12) Jones, A. C. Chem. Vap. Deposition 1998 , 4 , 169. (13) Zechmann, C. A.; Folting, K.; Caulton, K. G. Chem. Mater. 1998 , 10 , 2348. (14) McElwee White, L.; Koller, J. ; Kim, D.; Anderson, T. J. ECS Trans. 2009 , 25 , 161. (15) Bchir, O. J.; Green, K. M.; Ajmera, H. M.; Zapp, E. A.; Anderson, T. J.; Brooks, B. C.; Reitfort, L. L.; Powell, D. H.; Abboud, K. A.; McElwee White, L. J. Am. Chem. Soc. 2005 , 127 , 7825. (16) Lee , J.; Kim, D.; Kim, O. H.; Anderson, T.; Koller, J.; Denomme, D. R.; Habibi, S. Z.; Ehsan, M.; Eyler, J. R.; McElwee White, L. J. Electrochem. Soc. 2012 , 159 , H545. (17) McClain, K. R.; O'Donohue, C.; Shi, Z. W.; Walker, A. V.; Abboud, K. A.; Anderson, T. ; McElwee White, L. Eur. J. Inorg. Chem. 2012 , 4579.

PAGE 132

132 (18) McClain, K. R.; O'Donohue, C.; Koley, A.; Bonsu, R. O.; Abboud, K. A.; Revelli, J. C.; Anderson, T. J.; McElwee White, L. J. Am. Chem. Soc. 2014 , 136 , 1650. (19) Averseng, F.; Vennat, M.; Che, M. In Handbook of Heterogeneous Catalysis ; Wiley: 2008. (20) Chien, J. C. W. J. Am. Chem. Soc. 1971 , 93 , 4675. (21) Inumaru, K.; Okuhara, T.; Misono, M. J. Phys. Chem. 1991 , 95 , 4826. (22) Liu, Z. P.; Hu, P. J. Am. Chem. S oc. 2003 , 125 , 1958. (23) Shi, Z. W.; Lu, P.; Walker, A. V. Langmuir 2012 , 28 , 16909. (24) van Dorp, W. F.; Hagen, C. W. J. Appl. Phys. 2008 , 104 . (25) Edinger, K.; Becht, H.; Bihr, J.; Boegli, V.; Budach, M.; Hofmann, T.; Koops, H. W. P.; Kuschnerus, P .; Oster, J.; Spies, P.; Weyrauch, B. J. Vac. Sci. Technol., B 2004 , 22 , 2902. (26) Spencer, J.; Rosenberg, S.; Barclay, M.; Wu, Y. C.; McElwee White, L.; Howard Fairbrother, D. Appl. Phys. A 2014 , 1. (27) Bell, D. A.; Falconer, J. L.; Lu, Z. M.; McConic a, C. M. J. Vac. Sci. Technol., B 1994 , 12 , 2976. (28) Matsui, Y.; Hiratani, M.; Nabatame, T.; Shimamoto, Y.; Kimura, S. Electrochem. Solid State Lett. 2002 , 5 , C18. (29) Noh, J. H.; Stanford, M. G.; Lewis, B. B.; Fowlkes, J. D.; Plank, H.; Rack, P. D. A ppl. Phys. A 2014 , 1. (30) Lai, K. K.; Lamb, H. H. Thin Solid Films 2000 , 370 , 114. (31) Rosenberg, S. G.; Barclay, M.; Fairbrother, D. H. Phys. Chem. Chem. Phys. 2013 , 15 , 4002. (32) Wnuk, J. D.; Gorham, J. M.; Rosenberg, S. G.; van Dorp, W. F.; Madey, T. E.; Hagen, C. W.; Fairbrother, D. H. J. Phys. Chem. C 2009 , 113 , 2487. (33) Ertl, G., Knozinger, H., and Weitkamp, J., Handbook of Heterogeneous Catalysis ; Wiley VCH,Weinheim, 1997; Vol. 1. (34) Iwasawa, Y. Tailored metal catalysts ; D. Reidel Pub. Co.: Dordrecht ; Boston : Hingham, MA, U.S.A, 1986.

PAGE 133

133 (35) Galloway, J. N.; Dentener, F. J.; Capone, D. G.; Boyer, E. W.; Howarth, R. W. Biogeochemistry 2004 , 70 , 153. (36) Service, R. F. S cience 2014 , 345 , 610. (37) Smith, B. E. Science 2002 , 297 , 1654. (38) Ertl, G. J. Vac. Sci. Technol., A 1983 , 1 , 1247. (39) Hickman, D. A.; Huff, M.; Schmidt, L. D. Ind. Eng. Chem. Res. 1993 , 32 , 809. (40) Schlögl, R.; Wagner, E.; Fetzer, T.; Wagner, J.; Nehb, W.; Adlkofer, J.; Pachaly, B.; Keim, W. In Handbook of Heterogeneous Catalysis ; Wiley: 2008, p 1697. (41) Haber, J. Pure Appl. Chem. 1991 , 63 , 1227. (42) Fricke, R.; Jerschkewitz, H. G.; Kazanskii, V. B.; Shvets, V. A.; Ohlmann, G. React. Kinet. Catal. Lett. 1981 , 18 , 465. (43) Haukka, S.; Lakomaa, E. L.; Root, A. J. Phys. Chem. 1993 , 97 , 5085. (44) Li, Y. Z.; Fan, Y. N.; Yang, H. P.; Xu, B. L.; Feng, L. Y.; Yang, M. F.; Chen, Y. Chem. Phys. Lett. 2003 , 372 , 160. (45) Jung, K. T.; Bell, A. T. Catal. Lett. 2002 , 80 , 63. (46) Farfan Arribas, E.; Madix, R. J. J. Phys. Chem. B 2002 , 106 , 10680. (47) Farfan Arribas, E.; Madix, R. J. J. Phys. Chem. B 2003 , 107 , 3225. (48) Markovits, A.; Ahdjoudj, J.; Minot, C. Surf. Sci. 1996 , 365 , 649. (49 ) Suzuki, S.; Yamaguchi, Y.; Onishi, H.; Fukui, K.; Sasaki, T.; Iwasawa, Y. Catal. Lett. 1998 , 50 , 117. (50) Somorjai, G. A. J. Mol. Struc. Theochem 1998 , 424 , 101. (51) Dahl, S.; Logadottir, A.; Egeberg, R. C.; Larsen, J. H.; Chorkendorff, I.; Tornqvist , E.; Norskov, J. K. Phys. Rev. Lett. 1999 , 83 , 1814. (52) Weaver, J. F., Personal Communication. (53) Prockl, S. S.; Kleist, W.; Kohler, K. Tetrahedron 2005 , 61 , 9855. (54) Kohler, K.; Heidenreich, R. G.; Soomro, S. S.; Prockl, S. S. Adv. Synth. Catal. 2008 , 350 , 2930.

PAGE 134

134 (55) Molnar, A.; Sarkany, A.; Varga, M. J. Mol. Catal. A: Chem. 2001 , 173 , 185. (56) Bond, G. C.; Slaa, J. C. J. Mol. Catal. A: Chem. 1995 , 101 , 243. (57) Kohler, J. U.; Krauss, H. L. J. Mol. Catal. A: Chem. 1997 , 123 , 49. (58) Sierra , M. C. S.; Ruiz, J. G.; Proietti, M. G.; Blasco, J. J. Mol. Catal. A: Chem. 1995 , 96 , 65. (59) Schulz, H. Appl. Catal., A 1999 , 186 , 3. (60) Norrby, T.; Borje, A.; Zhang, L.; Akermark, B. Acta Chem. Scand. 1998 , 52 , 77. (61) Becker, M. H.; Chua, P.; Downham, R.; Douglas, C. J.; Garg, N. K.; Hiebert, S.; Jaroch, S.; Matsuoka, R. T.; Middleton, J. A.; Ng, F. W.; Overman, L. E. J. Am. Chem. Soc. 2007 , 129 , 11987. (62) Degraaf, W.; Boersma, J.; Smeets, W. J. J.; Spek, A. L.; Vank oten, G. Organometallics 1989 , 8 , 2907. (63) Graaf, W. d.; Boersma, J.; Grove, D.; Spek, A. L.; Koten, G. v. Recl. Trav. Chim. Pays Bas 1988 , 107 , 299 (64) Butschke, B.; Schwarz, H. Organometallics 2010 , 29 , 6002. (65) Joseph, M. F.; Page, J. A.; Baird, M. C. Organometallics 1984 , 3 , 1749. (66) Fan, L.; Turner, M. L.; Hursthouse, M. B.; Malik, K. M. A.; Gusev, O. V.; Maitlis, P. M. J. Am. Chem. Soc. 1994 , 116 , 385 (67) Knowles, D. R. T.; Adams, H.; Maitlis, P. M. Orga nometallics 1998 , 17 , 1741. (68) Davison, A.; McCleverty, J. A.; Wilkinson, G. J. Chem. Soc. 1963 , 1133. (69) Vilan, A.; Yaffe, O.; Biller, A.; Salomon, A.; Kahn, A.; Cahen, D. Adv. Mater. 2010 , 22 , 140. (70) Haick, H.; Cahen, D. Prog. Surf. Sci. 2008 , 83 , 217. (71) Ulman, A. Chem. Rev. 1996 , 96 , 1533. (72) Bigelow, W. C.; Pickett, D. L.; Zisman, W. A. J. Colloid Interface Sci. 1946 , 1 , 513. (73) Ulman, A.; Evans, S. D.; Shnidman, Y.; Sharma, R.; Eilers, J. E.; Chang, J. C. J. Am. Chem. Soc. 1991 , 113 , 1499.

PAGE 135

135 (74) Chen, J.; Reed, M. A.; Rawlett, A. M.; Tour, J. M. Science 1999 , 286 , 1550. (75) Nagy, G.; Walker, A. V. J. Phys. Chem. C 2007 , 111 , 8543. (76) Walker, A. V.; Tighe, T. B.; Reinard, M. D.; Haynie, B. C.; Allara, D. L.; Winograd, N. Chem. Ph ys. Lett. 2003 , 369 , 615. (77) Walker, A. V.; Tighe, T. B.; Cabarcos, O. M.; Reinard, M. D.; Haynie, B. C.; Uppili, S.; Winograd, N.; Allara, D. L. J. Am. Chem. Soc. 2004 , 126 , 3954. (78) Walker, A. V.; Tighe, T. B.; Stapleton, J.; Haynie, B. C.; Upilli, S.; Allara, D. L.; Winograd, N. Appl. Phys. Lett. 2004 , 84 , 4008. (79) Walker, A. V.; Tighe, T. B.; Haynie, B. C.; Uppili, S.; Winograd, N.; Allara, D. L. J. Phys. Chem. B 2005 , 109 , 11263. (80) Thran, A.; Kiene, M.; Zaporojtchenko, V.; Faupel, F. Phys. Rev. Lett. 1999 , 82 , 1903. (81) Zaporojtchenko, V.; Behnke, K.; Thran, A.; Strunskus, T.; Faupel, F. Appl. Surf. Sci. 1999 , 144 45 , 355. (82) Tighe, T. B.; Daniel, T. A.; Zhu, Z. H.; Uppili, S.; Winograd, N.; Allara, D. L. J. Phys. Chem. B 2005 , 109 , 21 006. (83) Collaud, M.; Nowak, S.; Kuttel, O. M.; Schlapbach, L. J. Adhes. Sci. Technol. 1994 , 8 , 435. (84) Vilan, A.; Cahen, D. Adv. Funct. Mater. 2002 , 12 , 795. (85) Walker, A. V. J. Vac. Sci. Technol., A 2013 , 31 . (86) Bensebaa, F.; Ellis, T. H.; Badia, A.; Lennox, R. B. J. Vac. Sci. Technol., A 1995 , 13 , 1331. (87) Wohlfart, P.; Weiss, J.; Kashammer, J.; Winter, C.; Scheumann, V.; Fischer, R. A.; Mittler Neher, S. Thin Solid Films 1999 , 340 , 274. (88) Weckenmann, U.; Mittler, S.; Kramer, S.; Aliganga, A. K. A.; Fischer, R. A. Chem. Mater. 2004 , 16 , 621. (89) Lu, P.; Demirkan, K.; Opila, R. L.; Walker, A. V. J. Phys. Chem. C 2008 , 112 , 2091. (90) Ehrlich, D. J.; Tsao, J. Y. J. Vac. Sci. Technol., B 1983 , 1 , 969. (91) Tsao, J. Y.; Ehrlich, D. J. Appl. Phys. Lett. 1984 , 45 , 617.

PAGE 136

136 (92) Calloway, A. R.; Galantowicz, T. A.; Fenner, W. R. J. Vac. Sci. Technol., A 1983 , 1 , 534. (93) Kazlauskas, R. J.; Wrighton, M. S. Organometallics 1982 , 1 , 602. (94) Casey, C. P.; O'Conno r, J. M. Organometallics 1985 , 4 , 384. (95) Wuu, Y. M.; Wrighton, M. S. Organometallics 1988 , 7 , 1839. (96) Sbrana, G.; Braca, G.; Piacenti, F.; Pino, P. J. Organomet. Chem. 1968 , 13 , 240 (97) Lloyd Jones, G. C.; Krska, S. W.; Hughes, D. L.; Gouriou, L.; Bonnet, V. D.; Jack, K.; Sun, Y.; Reamer, R. A. J. Am. Chem. Soc. 2003 , 126 , 702. (98) Jackstell, R.; Gómez Andreu, M.; Frisch, A.; Selvakumar, K.; Zapf, A.; Klein, H.; Spannenberg, A.; Röttger , D.; Briel, O.; Karch, R.; Beller, M. Angew. Chem., Int. Ed. Engl. 2002 , 41 , 986. (99) Langlois, J. B.; Alexakis, A. Angew. Chem., Int. Ed. 2011 , 50 , 1877 (100) Davies, S. G.; Fox, J. F.; Jones, S.; Price, A. J.; Sanz, M. A.; Sellers, T. G. R.; Smith, A. D.; Teixeira, F. C. J. Chem. Soc., Perkin Trans. 1 2002 , 1757 (101) Wagner, C. D.; Muilenberg, G. E. Handbook of X ray photoelectron spectroscopy: a reference book of standard data for use in X ray photoelectron spectroscopy ; Physical Electronics Divi sion, Perkin Elmer Corp., 1979. (102) Stewart, R. L. Physical Review 1934 , 45 , 0488. (103) Christy, R. W. J. Appl. Phys. 1960 , 31 , 1680. (104) Baker, A. G.; Morris, W. C. Rev. Sci. Instrum. 1961 , 32 , 458. (105) Harriott, L. R.; Wagner, A.; Fritz, F. J. Vac. Sci. Technol., B 1986 , 4 , 181. (106) Miura, N.; Ishii, H.; Shirakashi, J.; Yamada, A.; Konagai, M. Appl. Surf. Sci. 1997 , 113 , 269. (107) Koops, H. W. P.; Hoinkis, O. E.; Honsberg, M. E. W.; Schmidt, R.; Blum, R.; Bottger, G.; Kuligk, A.; Liguda, C.; Eich, M. Microelectron. Eng. 2001 , 57 8 , 995. (108) Edinger, K.; Gotszalk, T.; Rangelow, I. W. J. Vac. Sci. Technol., B 2001 , 19 , 2856. (109) Edgcombe, C. J.; Valdre, U. Philos. Mag. B 2002 , 82 , 987. (110) Li, P. G.; Jin, A. Z.; Tang, W. H. Phys. Status Solidi A 2006 , 203 , 282.

PAGE 137

137 (111) Umbach, C. P.; Washburn, S.; Webb, R. A.; Koch, R.; Bucci, M.; Broers, A. N.; Laibowitz, R. B. J. Vac. Sci. Te chnol., B 1986 , 4 , 383. (112) Boggild, P.; Hansen, T. M.; Tanasa, C.; Grey, F. Nanotechnology 2001 , 12 , 331. (113) Mukawa, T.; Okada, S.; Kobayashi, R.; Fujita, J.; Ishida, M.; Ichihashi, T.; Ochiai, Y.; Kaito, T.; Matsui, S. Jpn. J. Appl. Phys., Part 1 2005 , 44 , 5639. (114) Ding, W.; Dikin, D. A.; Chen, X.; Piner, R. D.; Ruoff, R. S.; Zussman, E.; Wang, X.; Li, X. J. Appl. Phys. 2005 , 98 . (115) Molhave, K.; Madsen, D. N.; Rasmussen, A. M.; Carlsson, A.; Appel, C. C.; Brorson, M.; Jacobsen, C. J. H.; Bo ggild, P. Nano Lett. 2003 , 3 , 1499. (116) Koops, H. W. P.; Kretz, J.; Rudolph, M.; Weber, M.; Dahm, G.; Lee, K. L. Jpn. J. Appl. Phys., Part 1 1994 , 33 , 7099. (117) Weber, M.; Rudolph, M.; Kretz, J.; Koops, H. W. P. J. Vac. Sci. Technol., B 1995 , 13 , 461. (118) Hoyle, P. C.; Cleaver, J. R. A.; Ahmed, H. J. Vac. Sci. Technol., B 1996 , 14 , 662. (119) Cicoira, F.; Hoffmann, P.; Olsson, C. O. A.; Xanthopoulos, N.; Mathieu, H. J.; Doppelt, P. Appl. Surf. Sci. 2005 , 242 , 107. (120) Utke, I.; Luisier, A.; Hoffmann, P.; Laub, D.; Buffat, P. A. Appl. Phys. Lett. 2002 , 81 , 3245. (121) Utke, I.; Michler, J.; Gasser, P.; Santschi, C.; Laub, D.; Cantoni, M.; Buffat, P. A.; Jiao, C.; Hoffmann, P. Adv. Eng. Mater. 2005 , 7 , 323. (122) Bruk, M. A.; Zhikharev, E. N.; Grigor'ev, E. I.; Spirin, A. V.; Kal'nov, V. A.; Kardash, I. E. High Energy Chem. 2005 , 39 , 65. (123) Frabboni, S.; Gazzadi, G. C.; Spessot, A. Appl. Phys. Lett. 2006 , 89 . (124) Takeguchi, M.; Shimojo, M.; Furuya, K. Nanotechnolo gy 2005 , 16 , 1321. (125) Guise, O.; Marbach, H.; Levy, J.; Ahner, J.; Yates, J. T. Surf. Sci. 2004 , 571 , 128. (126) Matsui, S.; Mori, K. J. Vac. Sci. Technol., B 1986 , 4 , 299. (127) Folch, A.; Tejada, J.; Peters, C. H.; Wrighton, M. S. Appl. Phys. Lett. 1995 , 66 , 2080. (128) Wang, S.; Sun, Y. M.; White, J. M. Appl. Surf. Sci. 2005 , 249 , 110.

PAGE 138

138 (129) Tseng, A. A. Nanofabrication: Fundamentals and Applications ; World Scientific: Singapore, 2007. (130) Silvis Cividjian, N.; Hagen, C. W. In Advances in Imaging and Electron Physics, Vol 143 2006; Vol. 143, p 1. (131) Mulders, J. J. L.; Veerhoek, J. M.; Bosch, E. G. T.; Trompenaars, P. H. F. J. Phys. D: Appl. Phys. 2012 , 45 . (132) Rosenberg, S. G.; Barclay, M.; Fairbrother, D. H. J. Phys. Chem. C 2013 , 1 17 , 16053. (133) Hoffmann, P. In Presentation at the First Internation EBID Workshop Delft, 2006. (134) Bell, D. A.; McConica, C. M.; Falconer, J. L.; Lu, Z. M. J. Electrochem. Soc. 1994 , 141 , 2884. (135) Shimojo, M.; Bysakh, S.; Mitsuishi, K.; Tanaka, M.; Song, M.; Furuya, K. Appl. Surf. Sci. 2005 , 241 , 56. (136) Luisier, A.; Utke, I.; Bret, T.; Cicoira, F.; Hauert, R.; Rhee, S. W.; Doppelt, P.; Hoffmann, P. J. Electrochem. Soc. 2004 , 151 , C535. (137) Seuret, P.; Cicoira, F.; Ohta, T.; Doppelt, P.; Ho ffmann, P.; Weber, J.; Wesolowski, T. A. Phys. Chem. Chem. Phys. 2003 , 5 , 268. (138) Utke, I.; Hoffmann, P.; Melngailis, J. J. Vac. Sci. Technol., B 2008 , 26 , 1197. (139) Botman, A.; Hesselberth, M.; Mulders, J. J. L. J. Vac. Sci. Technol., B 2008 , 26 , 2464. (140) Brintlinger, T.; Fuhrer, M. S.; Melngailis, J.; Utke, I.; Bret, T.; Perentes, A.; Hoffmann, P.; Abourida, M.; Doppelt, P. J. Vac. Sci. Technol., B 2005 , 23 , 3174. (141) Koops, H. W. P.; Kaya, A.; Weber, M. J. Vac. Sci. Technol., B 1995 , 13 , 2 400. (142) Landheer, K.; Rosenberg, S. G.; Bernau, L.; Swiderek, P.; Utke, I.; Hagen, C. W.; Fairbrother, D. H. J. Phys. Chem. C 2011 , 115 , 17452. (143) Singh, M.; Braat, J. J. M. Opt. Lett. 2001 , 26 , 259. (144) Scheuer, V.; Koops, H.; Tschudi, T. Micro electron. Eng. 1986 , 5 , 423. (145) J. Payne, M.; J. Cole Hamilton, D. J. Chem. Soc., Dalton Trans. 1997 , 3167.

PAGE 139

139 (146) Ihrman, K. G., Coffeld, Thomas H. Butadiene iron subgroup tricarbonyl hydrohalide compounds and their preparation; United States Patent; 1964. (147) Foley, S. R.; Stockland, R. A.; Shen, H.; Jordan, R. F. J. Am. Chem. Soc. 2003 , 125 , 4350. (148) Nataro, C.; Angelici, R. J. Inorg Chem 1998 , 37 , 2975. (149) Kölle, U.; Kossakowski, J.; Klaff, N.; Wesemann, L.; Englert, U.; Herberich, G. E. Angew. Chem. 1991 , 103 , 732. (150) Navarro Clemente, M. E.; Juárez Saavedra, P.; Cervantes Vásquez, M.; Paz Sandoval, M. A.; Arif, A. M.; Ernst, R. D. Organometallics 2002 , 21 , 592. (151) Knowles, D. R. T.; Adams, H.; Maitlis, P. M. Organometallics 1998 , 17 , 1741

PAGE 140

140 BIOGRAPHICAL SKETCH Joseph Brannaka is the second son to Charles and Nancy Brannaka. He was homeschooled all the way through high school by his mother a schoolteacher who instilled a love for the sc iences in him. He completed a Bachelor of Science in Chemistry in 2009 from Southern Adventist University located in Chattanooga, Tennessee. In 2009, he joined the doctoral program in Chemistry at the University of Florida. While pursuing his degree, Dr. B rannaka worked as a research associat e under the guidance of Lisa McE lwee White and has worked with numerous collaborators across the nation and in Europe. Dr. Brannaka has had his research presented at local, regional, national, and international conferen ces. During this time, he also earned the Chemistry Teaching Assistant award for his outstanding contribution as an assistance instructor. Throughout his undergraduate and graduate education, he has participated in student supportive clubs as well as provi ded tutoring/mentoring for fellow students. Aside from his academic achievements, Dr. Brannaka enjoys outdoor activities such as canoeing and hiking and is a devoted husband to his wife, Maira and a proud father to their daughter, Isabella.