Inner-sphere two-electron reduction leads to cleavageand functionalization of coordinated dinitrogenLiam P. Spencer, Bruce A. MacKay, Brian O. Patrick, and Michael D. Fryzuk*

Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, BC, Canada V6T 1Z1

Edited by Richard R. Schrock, Massachusetts Institute of Technology, Cambridge, MA, and approved May 23, 2006 (received for review March 22, 2006)

Activation of molecular nitrogen by transition metal complexes isan area of current interest as investigations using the inert N2

molecule to produce higher-value organonitrogen compoundsintensify. In an attempt to extend the addition of hydride reagentsE-H (where E � BR2, AlR2, and SiR3) to the dinitrogen complex([NPN]Ta)2(�-H)2(�-�1:�2-N2) [1; where NPN � (PhNSiMe2CH2)2PPh],the reaction with zirconocene chlorohydride, [Cp2Zr(Cl)H]x, wasexamined. The crystalline product formed in 35% yield was deter-mined to be ([NP(N)N]Ta)(�-H)2(�-N)(Ta[NPN])(ZrCp2) (2) in whichthe coordinated N2 has been cleaved to form a phosphinimidebridging between Ta and Zr and a triply bridging nitride. Themechanism of this reaction was examined to determine the fate ofthe chloride and hydride ligands attached to Zr in the startingzirconocene reagent. Using the zirconocene dihydride dimer([Cp2ZrH2]2), a higher yield of 2 was obtained (76%), and H2 wasalso observed by 1H NMR spectroscopy. To probe the origin of theeliminated H2, the dideuterated dinitrogen complex ([NPN]Ta)2(�-D)2(�-�1:�2-N2) (d2-1) was allowed to react with ([Cp2ZrH2]2), whichresulted in the formation of ([NP(N)N]Ta)(�-D)2(�-N)(Ta[NPN])(ZrCp2), (d2-2), with no evidence of hydrogen for deuterium scram-bling between the starting zirconocene dihydride and the ditan-talum dinitrogen complex. Studies into the use of preformed Zr(II)and Ti(II) reagents were also performed. The proposed mechanisminvolves initial adduct formation that facilitates inner-sphere elec-tron transfer to cleave the N-N bond to form a species with bridgingnitrides, one of which is transformed by nucleophilic attack of aphosphine donor to generate the observed phosphinimide.

dinitrogen coordination � nitrogen fixation � nitrogen–nitrogenbond cleavage

I t has been just slightly more than 40 years since the first metalcomplex with molecular nitrogen coordinated as a ligand was

reported. The isolation of [Ru(NH3)5N2]2� by Allen and Senoffin 1965 (1) is considered to be a milestone in inorganic coordi-nation chemistry because it changed the way people thoughtabout this simple, abundant, yet intrinsically inert diatomicmolecule. That it could coordinate to a metal complex wasconsidered to be the key first step in the eventual discovery ofhomogeneous catalysts that might use N2 as a feedstock toproduce higher-value organonitrogen products (2). Unfortu-nately, this problem is still unsolved. While much has beenlearned about how to synthesize dinitrogen complexes (hundredsare now known) (3–7) and how N2 binds to one or more metalcomplexes (7–9), there are only a few reports of homogeneoussystems that involve turnover of molecular nitrogen, and thesesystems are not efficient or broadly applicable (10, 11).

Part of the problem is that the reactivity patterns for coordi-nated N2 are rather meager (4, 9, 12, 13). By far the most wellstudied reaction for dinitrogen complexes is protonation (14), inan effort to try to model the nitrogenase enzymes (15–18), whichconvert N2 to ammonia (NH3) catalytically. Surprisingly, whathas been discovered is that the observed products, which caninclude NH3, hydrazine (H2NNH2), or even liberated dinitrogen,depend on a number of factors: the choice of the central metal,the choice of the ancillary ligands, and the choice of the acid (3,11, 14, 19). Another extremely common, but unproductive,

reactivity pattern for coordinated N2 is displacement by otherdonors or reactants, which is particularly prevalent for late metaldinitrogen complexes.

Recently, there have been a small number of new reactionsdiscovered for coordinated N2 that potentially could be used incatalytic cycles. Stoichiometric cleavage of the strong N–N bondhas been observed for groups 5 and 6 metal complexes (20–25),certain dinuclear zirconium dinitrogen complexes have beenfound to react with H2 to generate N–H bonds (26–28), andterminal alkynes have been found to react with coordinated N2in a side-on bound dizirconium complex to generate a N–C bond(29). These processes have expanded the toolbox of fundamentalreactions for coordinated dinitrogen.

In 1998, we (30) reported a ditantalum dinitrogen derivativethat has the N2 unit bound in the unusual side-on end-on mode;the complex ([NPN]Ta)2(�-H)2(�-�1:�2-N2) [1; where NPN �(PhNSiMe2CH2)2PPh] is remarkable not only for the way N2binds but also because of its facile formation by exposure of N2to a dinuclear tetrahydride complex avoiding the use of strongreducing agents (31). More importantly, this dinitrogen complexhas been shown to exhibit a wide range of reactivity patterns thatinclude reaction with electrophiles, adduct formation with Lewisacids, and displacement of N2 by terminal alkynes (31–33).

A particularly intriguing series of reactions is the addition ofsimple hydride reagents E-H (E � BR2; AlBu2

i ; SiH2Bu) to thedinitrogen complex 1 (34–37); these reactions, which result infunctionalization of coordinated N2, are summarized in Fig. 1.Common to all of these hydride addition processes is N-N bondcleavage. This cleavage is initiated by E-H addition across theexposed end of the coordinated dinitrogen to form in some casesan isolable intermediate (A); the next step involves eliminationof H2, followed by N–N bond cleavage and subsequent rear-rangement. This last step is very dependent on the hydridesource; for example, hydroboration leads to ancillary liganddegradation (35), and hydroalumination leads to ancillary ligandmigration (36), both of which hamper the overall usefulness ofthese stoichiometric transformations. In the hydrosilylationstudy, clean conversion to the bis(silylimide) was observed forthe addition of butylsilane BuSiH3 (37).

In this article, we report our attempt to extend this kind of E-Haddition�functionalization to the hydrozirconation of theside-on end-on dinitrogen complex 1 by reaction with chloro-bis(cyclopentadienyl)hydridozirconium(IV), [Cp2Zr(Cl)H]x.What results is an unexpected reaction that involves N-N bondcleavage without zirconium–hydride addition or H2 eliminationfrom the dinitrogen complex 1.

Results and DiscussionZirconocene chlorohydride or Schwartz’s reagent,[Cp2Zr(Cl)H]x, is known (38) to add across alkenes, alkynes,

Author contributions: M.D.F. designed research; L.P.S. and B.A.M. performed research;L.P.S., B.A.M., B.O.P., and M.D.F. analyzed data; and M.D.F. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS direct submission.

Abbreviations: Cp, cyclopentadienyl; THF, tetrahydrofuran.

*To whom correspondence should be addressed. E-mail: fryzuk@chem.ubc.ca.

© 2006 by The National Academy of Sciences of the USA

17094–17098 � PNAS � November 14, 2006 � vol. 103 � no. 46 www.pnas.org�cgi�doi�10.1073�pnas.0602132103

Dow

nloa

ded

by g

uest

on

Mar

ch 2

3, 2

020

ketones, aldehydes, and nitriles. Anticipating that it would addin a similar fashion to that described in Fig. 1, its reactivity with1 was examined. Stirring a brown tetrahydrofuran (THF) solu-tion of 1 with 1 equiv of sparingly soluble [Cp2Zr(Cl)H]x for 2weeks generates an intense purple solution from which purplecrystals of 2 were isolated in 35% yield. The 1H NMR spectrumreveals a Cs symmetric species in solution with four inequivalentsilyl methyl resonances, one single Cp resonance at � 5.38, andbridging hydrides at � 11.4. Surprisingly no terminal tantalumhydride resonances were observed, indicating that hydrozir-conation to form a species analogous to A in Fig. 1 had notoccurred. The 31P NMR spectrum displays two singlets at � 7.3and � 46.7 with the downfield resonance split into a doublet(1JPN � 34.9 Hz) when 15N-labeled 1 is used. This coupling wasalso observed in the 15N{1H} NMR spectroscopy with theresonance at � �185.1 split into a doublet. Another 15N reso-nance was observed at � 228.4; however, these two resonancesare not mutually coupled, implying N-N bond cleavage hasoccurred.

Single crystals of 2 were grown from benzene and subjected tox-ray analysis. The solid-state molecular structure is shown inFig. 2. What is clearly evident is that N-N bond cleavage hasoccurred and the zirconocene fragment has been inserted be-tween the two nitrogen atoms. In addition, evident in thesolid-state structure is the formation of a phosphinimide, pre-sumably by the intramolecular attack of one of the ancillaryphosphine donors at an intermediate nitride species. Such aprocess has been reported earlier in the formation of dinucleartitanium phosphinimides (39) by reduction of a titanium pre-cursor under N2 and in the thermolysis of a preformed niobiumdinitrogen complex (40). These intramolecular processes have

analogy to the intermolecular attack of phosphines on metalnitrides (41–43).

Although no bridging hydrides were found in the differencemap, they were successfully modeled with XHYDEX (44) andconfirmed by the 1H NMR data. The Ta(1)-Ta(2) distance is2.7007(6) Å similar to other Ta(IV)-Ta(IV) species (45); thePAN bond distance of 1.595(9) Å in this unit is typical of othergroups 4 and 5 phosphinimide complexes (46, 47). The zirco-nium nitride moiety is somewhat uncommon in the literature;however, comparison of Zr1-N2 at 2.040(9) Å shows it to beshorter that other bridging zirconium nitrides that range from2.21 to 2.35 Å (48–50). There is also some resemblance to otherTa-nitride complexes in the literature (51). Other selected bonddistances are given in Fig. 2.

What is clearly missing in the product is the chloro substituentfrom the initially added [Cp2Zr(Cl)H]x reagent; also not obviousis the fate of the zirconium hydride. To further probe this, weexamined the chloride-free zirconocene dihydride reagent,[Cp2ZrH2]2, with the side-on end-on dinitrogen complex 1. Inthis case, the reaction to form 2 proceeded in much higher yield(76% by 31P NMR spectroscopy) and an observed by-productwas H2 (peak at � � 4.47 in 1H NMR spectrum in C6D6),although this was not quantified. To confirm the origin of theliberated H2, the reaction of the deuterated dinitrogen complex,([NPN]Ta)2(�-D)2(�-�1:�2-N2) (d2-1), with the zirconocene di-hydride reagent, [Cp2ZrH2]2, led to the observation of free H2(no HD) and loss of the bridging tantalum-hydride signal at � �11.4 as evidenced by 1H NMR spectroscopy.

To rationalize the mechanism with the information obtainedso far, a likely scenario is the initial reaction of [Cp2Zr(Cl)H]xwith the starting dinitrogen complex 1 by chloride for hydride

Fig. 1. Sequence of reactions for the addition of simple main-group hydride reagents E-H to the dinitrogen complex 1.

Spencer et al. PNAS � November 14, 2006 � vol. 103 � no. 46 � 17095

CHEM

ISTR

YSP

ECIA

LFE

ATU

RE

Dow

nloa

ded

by g

uest

on

Mar

ch 2

3, 2

020

exchange, which then generates [Cp2ZrH2]2 along with someas-yet-unknown tantalum chloride species. This would accountfor the low yield of the final product when the zirconocenechlorohydride reagent is used (Fig. 3). To test this further, weadded Cp2ZrCl2 to the starting dinitrogen complex 1 andobserved the formation of final product 2 (�20% by NMRspectroscopy). The much lower yield here is consistent with thedouble chloride for hydride exchange that must occur to gener-ate the zirconocene dihydride from zirconocene dichlorideand 1.

As detailed above, the observation of free H2 in the formationof 2 from the reaction of 1 with [Cp2ZrH2]2 suggests that thiszirconocene dihydride reagent is simply a source of the Zr(II)

species, Cp2Zr. To test this, the Zr(II) precursor,Cp2Zr(py)(Me3SiC'CSiMe3) (52), was allowed to react with 1.In this case the formation of 2 is nearly quantitative, stronglysuggestive that the hydrides of [Cp2ZrH2]2 reductively eliminateearly in the reaction to form 2. This kind of H2 elimination fromZr(IV) is rare to our knowledge (53, 54).

As an extension, we also examined the reaction of 1 with therelated titanocene derivative, Cp2Ti(Me3SiC'CSiMe3), aknown Ti(II) source (55). This reaction proceeds smoothly togenerate the analogous complex 3, which was characterized byNMR spectroscopy and elemental analysis. Although no x-raystructure was obtained, the spectroscopic data match 2 quiteclosely; for example, the 31P NMR data show two singlets at � �8.6 and � � 46.4, with the latter downfield resonance assigned tothe phosphinimide donor; this was confirmed by reaction with15N-labeled 1, which results in the appearance of a doublet at � �46.4 (1JPN � 34 Hz). The 1H NMR spectrum shows the same Cssymmetric set of resonances, and the two equivalent bridginghydrides are observed at � � 12.4.

A possible mechanism for this process is shown in Fig. 4. AfterH2 elimination, the Lewis acid-base adduct B forms (32), whichcan subsequently provide two electrons to cleave the N-N bondto form C, in which the central metal of the metallocene has aM(IV) formal oxidation sate. Intramolecular attack of thephosphine unit on the bridging nitride, the TaAN-MCp2 unit ofC, generates the final product 2 (39).

ConclusionsOur attempts to extend the E-H addition to the side-on end-ondinitrogen complex 1 by reaction with [Cp2Zr(Cl)H]x resulted inan unexpected transformation. Although N-N bond cleavage andfunctionalization at nitrogen both occur, the process by whichthese happen is fundamentally different from that observed formain-group element hydrides such as boranes, silanes, and alanes(Fig. 1). Instead of zirconocene hydride addition across theTa2N2 unit of 1, coordination of Cp2Zr to the side-on end-onbound dinitrogen unit is proposed, followed by a two-electroninner-sphere reduction, whereby the zirconocene fragment in-serts into the N-N bond of 1. The transient species featuringbridging nitrides undergoes intramolecular attack by the phos-phine donor to generate the phosphinimide unit along with abridging trimetallic nitride moiety. Exactly how the zirconocenefragment Cp2Zr forms from [Cp2Zr(Cl)H]x or [Cp2ZrH2]2 isunknown; however, labeling studies indicate that elimination ofdihydrogen occurs either after hydride for halide exchange in the

Fig. 2. ORTEP drawing of the solid-state molecular structure of 2 (spheroidsat 50% probability; hydrogen and phenyl ring carbon atoms other than ipsoomitted for clarity); the hydride ligands were modeled by using XHYDEX.Selected bond distances (Å) and angles (°), respectively, are: Ta1-Ta2, 2.7007,6; N1� � � N2, 2.925, 9; P1-N1, 1.595, 9; N1-Ta1, 2.102, 9; Ta1-N2, 2.139, 8; Ta2-N2,1.935, 9; N1-Zr1, 2.146, 9; N2-Zr1, 2.040, 9; Ta1-N3, 2.054, 9; Ta1-N4, 2.102, 9;Ta2-N5, 2.112, 10; Ta2-N6, 2.119, 9; Ta2-P2, 2.624, 3; Ta1-Zr1, 3.0319, 11;N1-Ta1-N2, 87.2, 3; P1-N1-Ta1, 121.5, 5; N1-Zr1-N2, 88.6, 3; Ta1-N2-Ta2, 82.9,3; N3-Ta1-N4, 107.0, 4; N5-Ta2-N6, 105.6, 4; P1-N1-Ta1-N2, �176.3, 6; P2-Ta2-Ta1-Zr1, �13.25, 11; P2-Ta2-N2-Ta1, �175.8, 2.

Fig. 3. Outcome of the reaction of the dinitrogen complex 1 with [Cp2Zr(Cl)H]x.

17096 � www.pnas.org�cgi�doi�10.1073�pnas.0602132103 Spencer et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

3, 2

020

case of the chlorohydridozircononcene or directly from thezirconocene dihydride.

Experimental SectionGeneral Considerations. Unless otherwise stated, all manipulationswere performed under an inert atmosphere of dry, oxygen-freedinitrogen or argon by means of standard Schlenk or gloveboxtechniques. Anhydrous hexanes and toluene were purchased fromAldrich, sparged with dinitrogen, and passed through activatedalumina and Ridox catalyst columns under a positive pressure ofnitrogen before use. Anhydrous pentane, benzene, THF, anddiethyl ether were purchased from Aldrich, sparged with dinitro-gen, and passed through an Innovative Technologies (Newbury-port, MA) Pure-Solv 400 Solvent Purification System. Nitrogen gaswas dried and deoxygenated by passage through a column contain-ing activated molecular sieves and MnO. Deuterated benzene wasdried by heating at reflux with sodium�potassium alloy in a sealedvessel under partial pressure, then trap-to-trap distilled, and freeze-pump-thaw degassed three times. Unless otherwise stated, 1H, 31P,1H{31P}, 31P{1H} NMR spectra were recorded on a BrukerAMX-500 instrument with a 5-mm broadband inverse probe op-erating at 500.1 MHz for 1H. 31P NMR spectra were referencedto either external or internal P(OMe)3 (� 141.0 ppm with respectto 85% H3PO4 at � 0.0 ppm). Elemental analyses were performedby M. Lakha of the University of British Columbia.Cp2Zr(py)(Me3SiC'CSiMe3) (52) and Cp2Ti(Me3SiC'CSiMe3)(55) were synthesized according to literature procedures as was 1and its isotopomers (31).

Synthesis of 2 Using [Cp2Zr(Cl)H]x. To an intimate mixture ofSchwartz’s reagent (70.4 mg, 0.265 mmol) and 1 (334 mg, 0.265mmol, 1 equiv) in a 50-ml Erlenmeyer flask equipped with as stirbar was added 10 ml of toluene and 10 ml of THF in a glove box.The mixture was capped and stirred for 1 week at 15°C, afterwhich the red-brown color of 1 was converted to a dark purple.The crude purple solid 2 was recovered on a frit after evapo-ration of solvents and trituration with hexanes. X-ray qualitycrystals were obtained from a cooled solution of THF (137 mg,92.7 mmol, 35% yield). These were also used for NMR spec-

troscopy and elemental analysis. 1H{31P} NMR (C6D6, 300 K,400 MHz): � �0.32, �0.04, �0.02, 0.08 (s, 6H each, SiCH3), 1.17(AMX-500 2JHH � 11 Hz, 2JPH � 36 Hz, 4H, PCH2), 2.19(AMX-500, 2JHH � 15 Hz, 2JPH � 43 Hz, 4H, PCH2), 5.38 (s, 10H,�5-C5H5), 6.48, 6.55, 6.74, 6.95, 6.99, 7.62 (d, t, 20 H total,N-C6H5), 7.99, 8.26 (dd, 4H, o-PC6H5), 7.08, 7.12, 7.17, 7.68, (d,t, 6H total, PC6H5), 11.66 (d, 2JHP � 18 Hz, TaHTa). 13C{1H}NMR (C6D6, 300 K, 100.61 MHz): � 0.2, 0.9, 1.0, 1.8 (SiCH3),11.9, 20.4 (PCH2), 106.1 (�5-C5H5), 113.1, 113.9, 114.7, 117.8,124.7, 125.2 (NC6H5), 123.2, 129.0, 135.4, 137.8 (P-C6H5), 153.2,155.9 (o-PC6H5). 31P{1H} NMR (C6D6, 300 K, 161.9 MHz): � 7.3ppm (s), 46.7 ppm (s). Analysis: calculated forC58H74N6P2Si4Ta2Zr, C 46.99; H 5.03; N 5.67. Found: C 47.12;H 5.21; N 5.51.

Synthesis of 2 Using [Cp2ZrH2]2. To an intimate mixture of[Cp2ZrH2]2 (16 mg, 0.036 mmol) and 1 (92 mg, 0.072 mmol) in aJ-Young tube was added 2 ml of d8-THF in a glove box. A sealedcapillary tube containing a P(OMe)3 standard solution was added,and the mixture was sealed and rotated on a mechanical stirrer for2 weeks at room temperature. The 1H and 31P NMR confirmed theexclusive formation of 2 in addition to unreacted 1 and H2 forma-tion (4.54 in d8-THF) (76% yield by 31P NMR after 2 weeks). Asimilar experiment with d2-1 and [Cp2ZrH2]2 yielded 2 and nobridging hydride resonances in the 1H NMR.

Synthesis of 2 Using Cp2ZrCl2. To an intimate mixture of Cp2ZrCl2(24 mg, 0.081 mmol) and 1 (102 mg, 0.081 mmol) in a J-Youngtube was added 2 ml of C6D6 in a glove box. A sealed capillarytube containing a P(OMe)3 standard solution was added, and themixture was sealed and rotated on a mechanical stirrer for 1week at room temperature. The 1H and 31P NMR confirmed theformation of 2 (20% yield by 31P NMR after 2 weeks) in additionto H2 formation (4.47 in C6D6).

Synthesis of 2 Using Cp2Zr(py)(Me3SiC'CSiMe3). To a stirred solutionof 1 (350 mg, 0.28 mmol) in 20 ml of toluene was addedCp2Zr(py)(Me3SiC'CSiMe3) (135 mg, 0.28 mmol) dissolved in 5ml of toluene. The dark brown solution immediately darkened, and

Fig. 4. Proposed mechanism of the reaction of Cp2M with 1.

Spencer et al. PNAS � November 14, 2006 � vol. 103 � no. 46 � 17097

CHEM

ISTR

YSP

ECIA

LFE

ATU

RE

Dow

nloa

ded

by g

uest

on

Mar

ch 2

3, 2

020

the solution was stirred for 8 h. The solvent was removed undervacuum, leaving a purple residue, which was triturated with pen-tanes until a purple solid was retained. The resulting precipitate wasrecovered on a glass frit (370 mg, 0.25 mmol, 90% yield).

Synthesis of 15N2-2. One equivalent of Cp2Zr(py)(Me3SiC'C-SiMe3) was allowed to react with 1 equiv of 15N2-1 in a mannersimilar to that outlined above for the synthesis of 2. 15N NMR(C6D6, 300 K, 40 MHz): � 228.4 (d, JPN � 5 Hz), �185.1 (d,1JPN � 35).

Synthesis of 3. To a stirred solution of 1 (300 mg, 0.24 mmol) in20 ml of toluene was added Cp2Ti(Me3SiC'CSiMe3) (83 mg,0.24 mmol) dissolved in 5 ml of toluene. The dark brown solutionimmediately darkened, and the solution was stirred for 8 h. Thesolvent was removed under vacuum, leaving a purple residue,which was triturated with pentanes to yield a purple solid (0.19mmol, 80% yield). 1H{31P} NMR (C6D6, 300 K, 400 MHz): ��0.45, �0.01, 0.07, 0.11 (s, 6H each, SiCH3), 1.50 (AMX-500,2JHH � 10 Hz, 2JPH � 36 Hz, 4H, PCH2), 2.05 (AMX-500, 2JHH �15 Hz, 2JPH � 42 Hz, 4H, PCH2), 5.20 (s, 8H, �5-C5H5), 6.62–7.24(m, 26H total, -ArH), 9.32 (dd, 4H, o-PC6H5), 12.40 (d, 2JHP �17 Hz, TaHTa). 13C{1H} NMR (C6D6, 300 K, 100.61 MHz): �

0.2, 0.9, 1.1, 1.8 (SiCH3), 12.1, 20.6 (PCH2), 104.5 (�5-C5H5),112.5, 113.2, 113.5, 118.5, 123.6, 126.2 (NC6H5), 122.9, 129.6,136.9, 139.4 (P-C6H5), 155.6, 156.1 (o-PC6H5). 31P{1H} NMR(C6D6, 300 K, 161.9 MHz): � 8.6 ppm (s), 46.4 ppm (s). Analysis:calculated for C58H74N6P2Si4Ta2Ti, C 48.40; H 5.18; N 5.84.Found: C 48.23; H 5.10; N 5.49.

Synthesis of 15N2-3. One equivalent of Cp2Ti(Me3SiC'CSiMe3)was allowed to react with 1 equiv of 15N2-1 in a manner similarto that outlined above for the synthesis of 3. 15N NMR (C6D6, 300K, 40 MHz): � 227.5 (d, JPN � 5 Hz), �185.5 (d, 1JPN � 34 Hz).

Details of the X-Ray Crystal Structure Analysis of 1. CCDC 290043contains crystallographic data for this article. These data can beobtained free of charge from The Cambridge CrystallographicData Center (www.ccdc.cam.ac.uk�data�request�cif). See alsoSupporting Text, which is published as supporting information onthe PNAS web site.

This work by supported by a Natural Sciences and Engineering ResearchCouncil Discovery grant (to M.D.F.) and Natural Sciences and Engi-neering Research Council postgraduate scholarships (to L.P.S. andB.A.M.).

1. Allen AD, Senoff CV (1965) Chem Commun 621–622.2. Fryzuk MD (2003) Chem Rec 3:2–11.3. Chatt J, Dilworth JR, Richards RL (1978) Chem Rev 78:589–625.4. Hidai M (1999) Coord Chem Rev 185�186:99–108.5. Shaver MP, Fryzuk MD (2003) Adv Synth Catal 345:1061–1076.6. Gambarotta S, Scott J (2004) Angew Chem Int Ed 43:5298–5308.7. MacKay BA, Fryzuk MD (2004) Chem Rev 104:385–401.8. Fryzuk MD, Johnson SA (2000) Coord Chem Rev 200�202:379–409.9. MacLachlan EA, Fryzuk MD (2006) Organometallics 27:1530–1543.

10. Komori K, Oshita H, Mizobe Y, Hidai M (1989) J Am Chem Soc 111:1939–1940.

11. Yandulov DV, Schrock RR (2003) Science 301:76–78.12. Hidai M, Mizobe Y (1995) Chem Rev 95:1115–1133.13. Figueroa JS, Piro NA, Clough CR, Cummins CC (2006) J Am Chem Soc

128:940–950.14. Leigh GJ (1992) Acc Chem Res 25:177–181.15. Einsle O, Tezcan FA, Andrade Susana LA, Schmid B, Yoshida M, Howard JB,

Rees DC (2002) Science 297:1696–1700.16. Kim J, Rees DC (1992) Nature 360:553–560.17. Kim J, Rees DC (1992) Science 257:1677–1682.18. Lee SC, Holm RH (2003) Proc Natl Acad Sci USA 100:3595–3601.19. Yandulov DV, Schrock RR (2005) Can J Chem 83:341–357.20. Kawaguchi H, Matsuo T (2002) Angew Chem Int Ed 41:2792–2794.21. Laplaza CE, Cummins CC (1995) Science 268:861–863.22. Laplaza CE, Johnson MJA, Peters JC, Odom AL, Kim E, Cummins CC,

George GN, Pickering IJ (1996) J Am Chem Soc 118:8623–8638.23. Caselli A, Solari E, Scopelliti R, Floriani C, Re N, Rizzoli C, Chiesi-Villa A

(2000) J Am Chem Soc 122:3652–3670.24. Clentsmith GKB, Bates VME, Hitchcock PB, Cloke FGN (1999) J Am Chem

Soc 121:10444–10445.25. Mindiola DJ, Meyer K, Cherry J-PF, Baker TA, Cummins CC (2000) Orga-

nometallics 19:1622–1624.26. Fryzuk MD, Love JB, Rettig SJ, Young VG (1997) Science 275:1445–1447.27. Pool JA, Lobkovsky E, Chirik PJ (2004) Nature 427:527–530.28. Pool JA, Bernskoetter WH, Chirik PJ (2004) J Am Chem Soc 126:14326–14327.29. Morello L, Love JB, Patrick BO, Fryzuk MD (2004) J Am Chem Soc

126:9480–9481.30. Fryzuk MD, Johnson SA, Rettig SJ (1998) J Am Chem Soc 120:11024–11025.

31. Fryzuk MD, Johnson SA, Patrick BO, Albinati A, Mason SA, Koetzle TF(2001) J Am Chem Soc 123:3960–3973.

32. Studt F, Johnson SA, MacKay B, Patrick BO, Fryzuk MD, Tuczek F (2005)Chem Eur J 11:604–618.

33. Shaver MP, Johnson SA, Fryzuk MD (2005) Can J Chem 83:652–660.34. Fryzuk MD, MacKay BA, Johnson SA, Patrick BO (2002) Angew Chem Int Ed

41:3709–3712.35. MacKay B, Johnson SA, Patrick BO, Fryzuk MD (2005) Can J Chem

83:315–323.36. MacKay BA, Patrick BO, Fryzuk MD (2005) Organometallics 24:3836–3841.37. Fryzuk MD, MacKay BA, Patrick BO (2003) J Am Chem Soc 125:3234–3235.38. Schwartz J, Labinger JA (1976) Angew Chem 88:402–414.39. Morello L, Yu P, Carmichael CD, Patrick BO, Fryzuk MD (2005) J Am Chem

Soc 127:12796–12797.40. Fryzuk MD, Kozak CM, Bowdridge MR, Patrick BO, Rettig SJ (2002) J Am

Chem Soc 124:8389–8397.41. Bennett BK, Saganic E, Lovell S, Kaminsky W, Samuel A, Mayer JM (2003)

Inorg Chem 42:4127–4134.42. Betley TA, Peters JC (2004) J Am Chem Soc 126:6252–6254.43. Seymore SB, Brown SB (2002) Inorg Chem 41:462–469.44. Orpen AG (1980) J Chem Soc Dalton Trans 2509–2516.45. Shaver MP, Fryzuk MD (2005) Organometallics 24:1419–1427.46. Courtenay SS, Stephan DW (2001) Organometallics 20:1442–1450.47. Yue N, Hollink E, Guerin F, Stephan DW (2001) Organometallics 20:4424–

4433.48. Bai G, Muller P, Roesky HW, Uson I (2000) Organometallics 19:4675–4677.49. Abarca A, Martın A, Mena M, Yelamos C (2000) Angew Chem Int Ed

39:3460–3463.50. Banaszak Holl MM, Wolczanski PT (1992) J Am Chem Soc 114:3854–3858.51. Banaszak Holl MM, Wolczanski PT, van Duyne GD (1990) J Am Chem Soc

112:7989–7994.52. Nitschke JR, Zuercher S, Tilley TD (2000) J Am Chem Soc 122:10345–10352.53. Chirik PJ, Henling LM, Bercaw JE (2001) Organometallics 20:534–544.54. Edelbach BL, Fazlur AK, Lachiotte RJ, Jones WD (1999) Organometallics

18:3170–3177.55. Varga V, Mach K, Heyrovsky J, Schmid G, Thewalt U (1994) J Organomet

Chem 475:127–137.

17098 � www.pnas.org�cgi�doi�10.1073�pnas.0602132103 Spencer et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

3, 2

020