Synthesis of tungsten complexes that contain hexaisopropylterphenyl- substituted triamidoamine ligands, and reactions relevant to the reduction of dinitrogen to ammonia

341

Synthesis of tungsten complexes that contain

hexaisopropylterphenyl-substituted triamidoamine

ligands, and reactions relevant to the reduction of

dinitrogen to ammonia

Dmitry V. Yandulov and Richard R. Schrock

Abstract: [HIPTN₃N]WCl (WCl) can be synthesized readily by adding H₃[HIPTN₃N] to WCl₄(DME) followed by

LiN(SiMe₃)₂([HIPTN₃N]^3–= [(HIPTNCH₂CH₂)₃N]^3–where HIPT = 3,5-(2,4,6-i-Pr₃C₆H₂)₂C₆H₃= HexaIsoPropylTer-

phenyl). Reduction of WCl with KC₈in benzene under N₂yields WN=NK. WN=NK is readily oxidized in THF by

ZnCl₂to yield zinc metal and WN₂. Reduction of WN₂to [WN₂]^–is reversible at –2.27 V vs. FeCp₂in 0.1 mol/L

+/0

+

′

[Bu₄N][BAr₄]/PhF electrolyte (Ar′ = 3,5-(CF₃)₂C₆H₃), while oxidation of WN₂to [WN₂] is also reversible at –0.66 V.

Protonation of WN=NK by [Et₃NH][OTf] in benzene yields WN=NH essentially quantitatively. Protonation of WN=NH

′

at N_βwith [H(OEt)₂][BAr₄] in ether affords [W=NNH₂][BAr₄] quantitatively. Electrochemical reduction of

′

[W=NNH₂][BAr₄] in 0.1 mol/L [Bu₄N][BAr₄]/PhF is irreversible at scan rates of up to 1 V/s. Addition of NaBAr₄and

NH₃to WCl in PhF yields [W(NH₃)][BAr₄]. Electrochemical reduction of [W(NH₃)][BAr₄] in 0.1 mol/L [Bu₄N]-

[BAr₄]/PhF is irreversible at –2.06 V vs. FeCp₂at a scan rate of 0.5 V/s. Treatment of [W(NH₃)][BAr₄] with

triethylamine and [FeCp₂][PF₆] in C₆D₆, followed by LiN(SiMe₃)₂, yielded WϵN. Treatment of [W(NH₃)][BAr₄] with

′

+/0

′

LiBHEt₃(1 mol/L in THF) results in formation of WH, which is converted to WH₃upon exposure to an atmosphere

′

of H₂. Attempts to prepare WN=NH by treating WN₂with [2,6-LutH][BAr₄] and CoCp₂yielded only [W=NNH₂][BAr₄].

*

′

[W=NNH₂][BAr₄] is reduced to W=NNH₂by CoCp₂, but this species disproportionates to yield WN=NH, WϵN, and

ammonia. Reduction of [W(NH₃)][BAr₄] with CoCp₂does not yield any observable W(NH₃). Attempted catalytic re-

*

′

duction of dinitrogen using WN₂as the catalyst under conditions identical or similar to those employed for catalytic

reduction of dinitrogen by MoN₂and related Mo complexes failed. Single crystal X-ray studies were carried out on

′

W-N=NK, WN₂, W-N=NH, [W=NNH₂][BAr₄], and [W(NH₃)][BAr₄].

Key words: dinitrogen, reduction, tungsten, ammonia.

Résumé : Le complexe [HIPTN₃N]WCl (WCl) peut fa c 3 i l 5 e 7 ment être synthétisé par l’addition de H₃[HIPTN₃N] à du

WCl₄(DME), suivie d’une additon de LiN(SiMe₃)₂([HIPTN₃N]^3–= [(HIPTNCH₂CH₂)₃N]^3–dans lequel HIPT = 3,5-

(2,4,6-i-Pr₃C₆H₂)₂C₆H₃= HexaIsoPropylTerphényle). La réduction de WCl par du KC₈, dans du benzène, sous atmos-

phère d’azote, conduit à la formation de WN=NK. Sous l’action du ZnCl₂dans du THF, le WN=NK s’oxyde faci-

′

lement pour conduire à la formation de zinc métallique et de WN₂. Dans un électrolyte à 0,1 mol/L de [Bu₄N][BAr₄]/PhF

(Ar′ = 3,5-(CF₃)₂C₆H₃), la réduction du WN₂en [WN₂]^–est réversible à –2,27 V vs. FeCp₂alors que l’oxydation du

+/0

[WN₂] en [WN₂]⁺est aussi réversible à –0,66 V. La protonation du WN=NK par le [Et₃NH][OTf], dans le benzène,

conduit à la formation de WN=NH, avec un rendement pratiquement quantitatif. La protonation du WN=NH au niveau

′

du N_βavec du [H(OEt)₂][BAr₄], dans l’éther, conduit à la formation de [W=NNH₂][BAr₄] avec un rendement quantita-

′

tif. La réduction électrochimique du [W=NNH₂][BAr₄] dans du [Bu₄N][BAr₄]/PhF à 0,1 mol/L est irréversible à des

vitesses de balayage allant jusqu’à 1 V/s. L’addition de NaBAr₄et de NH₃au WCl dans du PhF conduit à la forma-

tion de [W(NH₃)][BAr₄]. La réduction électrochimique du [W(NH₃)][BAr₄] dans du [Bu₄N][BAr₄]/PhF à 0,1 mol/L est

irréversible à –2,06 V vs. FeCp₂, à une vitesse de balayage de 0,5 V/s. Le traitement du [W(NH₃)][BAr₄] par de la

′

+/0

′

triéthylamine et du [FeCp₂][PF₆] dans du C₆D₆, suivi de LiN(SiMe₃)₂conduit à la formation de WϵN. Le traitement

′

de [W(NH₃)][BAr₄] par du LiBHEt₃(1 mol/L dans du THF) conduit à la formation de WH qui peut être transformé

en WH₃par exposition à un atmosphère de H₂. Des essais en vue de préparer le WN=NH par traitement du WN₂avec

*

′

du [2,6-LutH][BAr₄] et du CoCp₂ne fournissent que du [W=NNH₂][BAr₄]. Sous l’influence du CoCp₂, le [W=NNH₂]-

′

[BAr₄] est réduit en W=NNH₂, mais cette espèce subit une réaction de dismutation conduisant à la formation de

WN=NH, de WϵN et d’ammoniac. La réduction du [W(NH₃)][BAr₄] par le CoCp₂ne fournit aucune trace de

*

′

W(NH₃). Des essais en vue d’effectuer la réduction du diazote en utilisant du WN₂comme catalyseur, dans des

Received 11 August 2004. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 30 March 2005.

D.V. Yandulov¹and R.R. Schrock.²Department of Chemistry, Massachusetts Institute of Technology, Cambridge, MA 02139,

USA.

¹Present address: Department of Chemistry, Stanford University, Stanford, CA 94305-5080, USA.

²Corresponding author (e-mail: rrs@mit.edu).

Can. J. Chem. 83: 341–357 (2005)

doi: 10.1139/V05-013

342

Can. J. Chem. Vol. 83, 2005

conditions identiques ou semblables à celles utilisées pour la réduction catalytique du diazote par le MoN₂et les com-

plexes apparentés du molybdène n’ont pas donné les résultats escomptés. Des études de diffraction des rayons X sur un

′

cristal unique ont été effectuées sur le W-N=NK, le WN₂, le W-N=NH, le [W=NNH₂][BAr₄] et le [W(NH₃)][BAr₄].

Mots clés : diazote, réduction, tungstène, ammoniac.

[Traduit par la Rédaction] Yandulov and Schrock

Introduction

Fig. 1. A drawing of [HIPTN₃N]Mo(N₂) = MoN₂.

Discovery of the first dinitrogen complex of a transition

metal, [Ru(NH₃)₅(N₂)]²⁺, in 1965 (1) provided support for

the proposal that dinitrogen might be reduced to ammonia at

a molybdenum center in nitrogenase (only the FeMo nitro-

genase was known at the time (2–4)). As more and more

types of transition-metal-catalyzed reactions were discovered

in the following decades, scientists became optimistic that

catalysts would be discovered for the reduction of dinitrogen

to ammonia, or for the combination of dinitrogen with ele-

ments other than hydrogen (5–16). Unfortunately, although

hundreds of man years have been directed toward this goal

in the last 40 years, and although hundreds of dinitrogen

complexes of all transition metals in groups 4 through 10 (if

we include those detected at low temperature in matrices)

are now known, progress has been relatively slow.

Chatt, a pioneer, along with Hidai, in dinitrogen chemistry

involving W(0) and Mo(0) phosphine complexes (5–7, 12,

14) believed that dinitrogen could be reduced to ammonia at

a single metal center through end-on binding of dinitrogen,

and was the first to show that up to 2 equiv. of ammonia per

metal could be formed from M(0) dinitrogen complexes

(M = Mo or W) upon addition of protons, the six electrons

being provided by the metal (5). If dinitrogen could be re-

duced to ammonia catalytically through the addition of six

external protons and six external electrons, and if a distinct

intermediate were formed upon addition of each proton and

electron, then 12 intermediates could be imagined, or a total

of 14 if one includes a dinitrogen complex and a complex

that contains no nitrogen or nitrogen-derived ligand. Al-

though examples of almost all of the proposed intermediates

for reduction of dinitrogen at a single Mo or W center in a

“Chatt cycle” have been isolated, the ligand environment

(L_x) of the metal has varied from compound to compound,

and no catalytic reaction to give ammonia was ever achieved

(see, however, the synthesis of silylamines catalytically

(17)). Significant progress has been made toward utilizing

atmospheric dinitrogen to make organic molecules (6, 7, 18–

26). One catalytic reduction is known in which dinitrogen is

reduced to a mixture of hydrazine (largely) and ammonia (9,

27, 28). This system requires molybdenum, is catalytic with

respect to molybdenum, produces >90% hydrazine, and

takes place in methanol in the presence of Mg(OH)₂and a

strong reducing agent (e.g., Na amalgam). Unfortunately,

few details concerning the mechanism of this intriguing re-

action have been established. In this context, it should be

noted that ammonia can be formed from hydrazine via

disproportionation (to dinitrogen and ammonia) in the pres-

ence of transition metals. In fact, [Ru(NH₃)₅(N₂)]²⁺was first

prepared by treating Ru salts with hydrazine (1).

plore the chemistry of Mo complexes that contain a triami-

doamine ([(ArNCH₂CH₂)₃N]^3–= [ArN₃N]^3–, Ar = aryl)

ligand (30, 31). To prevent formation of what we believe to

be relatively stable and unreactive bimetallic [ArN₃N]Mo-

N=N-Mo[ArN₃N] complexes, maximize steric protection of

a metal coordination site in a monometallic species, and pro-

vide increased solubility in nonpolar solvents, we synthe-

sized species that contain a [HIPTN₃N]^3–ligand, where

HIPT = 3,5-(2,4,6-i-Pr₃C₆H₂)₂C₆H₃(HexaIsoPropylTerphenyl,

see Fig. 1) (32, 33). Starting with MoCl (Mo = [HIPTN₃N]Mo),

we showed that we could isolate and characterize many in-

termediates in

a hypothetical Chatt-like reduction of

dinitrogen, all of which contain the same [HIPTN₃N]^3–sup-

porting ligand (Fig. 2). These intermediates include para-

magnetic Mo(N₂) (1), diamagnetic [Mo(N₂)]^–(2), diamagnetic

′

Mo-N=N-H (3), diamagnetic {Mo=N-NH₂}{BAr₄} (4; Ar′ =

3,5-(CF₃)₂C₆H₃), diamagnetic MoϵN (7), diamagnetic

′

{Mo=NH}{BAr₄} (8), paramagnetic {Mo(NH₃)}{BAr₄} (12),

and paramagnetic Mo(NH₃) (13). Extensive ¹⁵N labeling

studies, NMR studies, and X-ray studies (of 1, 2 (as two dif-

ferent Mg derivatives), 7, and 12 (33), and 3, 4, 8, and 13

(34)) all reveal a trigonal pocket in which N₂or its reduced

products are protected to a dramatic degree by three 3,5-

(2,4,6-i-Pr₃C₆H₂)₂C₆H₃(HIPT) rings clustered around it.

Compounds 5, 6, 10, and 14 (Fig. 2) have not yet been ob-

served. The oxidation state of the metal varies between

Mo(III) and Mo(VI) in the species as they are drawn in

Fig. 2. No other unambiguous example of one of these com-

plexes (3) is known, presumably either because in other sys-

tems bimolecular processes that involve H radical or proton

transfer result in facile decomposition, or the synthetic meth-

ods are incompatible with the ligands that are present. Steric

protection of the nitrogen entities within the pocket provided

by the bulky HIPT groups are believed to be an important

reason why several of the most unusual species can be pre-

pared.

We have been interested in dinitrogen chemistry of early

metals in relatively high oxidation states for more than two

decades (29). Approximately 4 years ago, we began to ex-

Yandulov and Schrock

343

Fig. 2. Proposed intermediates in the reduction of dinitrogen at a

[HIPTN₃N]Mo (Mo) center through the step-wise addition of

protons and electrons.

It has been known for some time that certain bacteria

grown in the presence of tungsten will produce a “FeW

nitrogenase”, but that enzyme will not reduce dinitrogen to

ammonia (2, 3, 38). Therefore, we became interested in pre-

paring some tungsten analogs of the [HIPTN₃N]Mo species

shown in Fig. 2, exploring conversion of one to another, and

testing at least one of them for catalytic dinitrogen reduction

activity. We report this work here.

Results

Synthesis of WCl and diazenido species

The starting material for this chemistry ([HIPTN₃N]WCl

(WCl)) can be readily synthesized by a two-step method

analogous to that used to prepare [HIPTN₃N]MoCl (33) as

1

shown in eq. [1]. The H NMR spectrum of high-spin WCl

(d²) contains sharper resonances than does the spectrum of

MoCl; this difference, and the pattern of relative shifts of the

corresponding resonances between Mo and W complexes is

very similar to that observed previously for a series of

related [ArN₃N]MCl (M = Mo, W) (30, 39, 40) and

[Me₃SiN₃N]MCl (41–43) complexes. Like MoCl, WCl is

exceedingly air- and moisture sensitive, so the utmost care is

required to ensure that solvents are pure, dry, and free of ox-

ygen.

Reduction of MoCl with magnesium in THF under an at-

mosphere of dinitrogen led to salts that contain the [Mo(N₂)]^–

anion (33). However, attempts to prepare [W(N₂)]^–salts in

this manner were not successful; WCl was slowly converted

into a complex mixture of unidentified products. Reduction

of WCl is markedly accelerated when potassium graphite

(KC₈) is employed as the reductant in THF or DME, but a

complex mixture of products is still formed. However, when

WCl is treated with KC₈in benzene under dinitrogen, clean

and nearly quantitative conversion of WCl to diamagnetic

W-N=NK takes place over a period of 1 to 2 days at room

temperature (eq. [2]). No intermediates could be detected by

¹H NMR when this reaction was performed in C₆D₆. Deep

red W-N=NK was isolated in good yield and was fully char-

acterized. Since W-N=NK can be dissolved in THF without

decomposition (vide infra), failure of the reductions in THF

and DME must be attributed to a some side reaction or reac-

tions prior to binding of dinitrogen to the metal, presumably

at the W(III) stage. WN=NK is readily soluble in aliphatic

We then showed that it is possible in fact to reduce dini-

trogen catalytically in heptane using {2,6-lutidinium}{BAr₄}

as the proton source and decamethylchromocene as the elec-

tron source (35). The results of 16 runs using four different

Mo derivatives (1, 3, 7, or 12) revealed that ~8 equiv. of am-

monia are formed with an efficiency of 63%–66% based

′

on reducing equivalents available.

A run employing

Mo(¹⁵N¹⁵NH) under ¹⁵N₂yielded entirely ¹⁵N-labeled am-

monia. We assume at this stage that any equivalents of re-

ducing agent that are not consumed in making ammonia are

consumed to form hydrogen, and we now know that no

hydrazine is formed.³The fact that compounds that contain

hexamethylterphenyl or hexa-tert-butylterphenyl variations

of the [HIPTN₃N]^3–ligand are relatively inefficient catalysts

for catalytic reduction of dinitrogen to ammonia (37) sug-

gests that the catalytic reaction is extremely finely balanced

and that perhaps most or all of the intermediates shown in

Fig. 2 are actually involved. All are likely to be equilibria

with exception of the actual N—N bond cleavage reaction

+

(Mo=N-NH₃+ e → MoϵN + NH₃).

³W. Weare. Unpublished results.

344

Can. J. Chem. Vol. 83, 2005

Fig. 3. Thermal ellipsoid drawing of WN=NK (50% probability level, hydrogen atoms omitted).

solvents and can be recrystallized without change from

pentane.

different, with one HIPT group in contact with K being ro-

tated by only 10° out of the plane of the amide nitrogen

(N1), while the other two HIPT groups are rotated by an av-

erage of 28.3° outside of the plane of the amide nitrogen to

which they are attached. “Encapsulation” of the K ion be-

tween two aryl rings accounts for the high solubility of W-

N=NK in aliphatic solvents (binding of sodium or potassium

ions to aryl rings in high oxidation state amido complexes

has considerable precedent in the literature (44–47)). The

N—N (1.220(8) Å) and W—N (1.849(7) Å) bond distances

are consistent with a considerably greater degree of reduc-

tion of the N₂unit in W-N=NK relative to that in Mo-

N=NMgBr(THF)₃(1.156(8) and 1.863(7) Å, respectively)

(33). Greater backbonding from the tungsten center in W-

N=NK (vs. the molybdenum center in Mo-N=NMgCl(THF)₃

(33)) is also evident from the relative values for ν_NNin the

two species (1745 cm^–1(W) and 1812 cm^–1(Mo), respec-

tively, in benzene). The W-N_α-N_βangle is essentially linear

(177.3(6)°), as expected.

The structure of W-N=NK, as determined through a single

crystal X-ray study, is shown in Fig. 3 (see also Table 1 for

crystal data and Table 2 for selected M—N and N—N bond

distances and angles in W and Mo analogs, along with se-

lected IR and NMR data. A full set of X-ray data is available

as supporting information.⁴The potassium ion is 2.536(7) Å

from N6, the β nitrogen in the unsubstituted diazenido ligand,

and is supported through six interactions with aryl carbon at-

oms in two Trip (2,4,6-i-Pr₃C₆H₂) aromatic rings of two

HIPT substituents. The potassium ion resides 3.229 and

3.301 Å from the ring centroids, and all K···H contacts are

>2.92 Å. Interaction between the potassium and the aromatic

rings is not symmetric, with K···C distances spanning a

range of 0.52 Å for each ring. The presence of the potassium

ion has no noticeable effect on the C—C bond distances in

the two aryl rings with which it interacts. Understandably,

the orientations of the three HIPT groups are significantly

⁴Supplementary data may be purchased from the Directory of Unpublished Data, Document Delivery, CISTI, National Research Council

Canada, Ottawa, ON K1A 0S2, Canada (http://cisti-icist.nrc-cnrc.gc.ca/irm/unpub_e.shtml for information on ordering electronically).

CCDC 261864–261868 contain the crystallographic data for this manuscript. These data can be obtained, free of charge, via

ww.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ,

U.K.; fax +44 1223 336033; or deposit@ccdc.cam.ac.uk).

Yandulov and Schrock

345

Table 1. Crystal data and structure refinement for [HIPTN₃N]W-N=NK, [HIPTN₃N]WN₂, [HIPTN₃N]W-N=NH, {[HIPTN₃N]W=N-

a

′

NH₂}[BAr₄], and {[HIPTN₃N]W(NH₃)}[BAr₄].

′ ′

{W-N=NH₂}[BAr₄] {W(NH₃)}[BAr₄]

W-N=NK

WN₂

W-N=NH

Empirical formula

C₁₁₄H₁₅₉KN₆W

C₁₁₄H₁₅₉N₆W

C₁₁₄H₁₆₀N₆W

C₁₄₆H₁₇₃BF₂₄N₆W

C₁₄₆H₁₇₄BF₂₄N₅W

Formula weight

Crystal system

Space group

a (Å)

b (Å)

c (Å)

1 836.42

Monoclinic

P2(1)/n

18.297(4)

40.166(9)

19.528(4)

90

117.600(5)

90

12 718(5)

4

1 797.32

Monoclinic

Cc

16.0231(9)

40.125(2)

17.9933(10)

90

93.8110(10)

90

11 542.7(11)

4

1 798.33

Triclinic

P1

2 662.56

Triclinic

P1

2649.56

Triclinic

P1

18.2377(14)

20.1004(15)

37.763(3)

86.923(2)

85.346(2)

75.098(2)

13 326.0(17)

4

15.7795(10)

20.6647(14)

26.2549(17)

85.2740(10)

83.6730(10)

81.9220(10)

8405.1(10)

2

15.8411(9)

20.5894(11)

26.3012(14)

84.9740(10)

83.8560(10)

81.8820(10)

8420.7(8)

2

α (°)

β (°)

γ (°)

Volume (Å³)

Z

Density (calcd., Mg/m³)

Absorption coeff. (mm^–1)

F(000)

0.959

0.981

3912

1.034

1.044

3836

0.896

0.904

3840

1.052

0.755

2772

1.045

0.754

2760

Crystal size (mm³)

θ range (°)

Index ranges

0.20 × 0.18 × 0.07 0.27 × 0.15 × 0.14 0.20 × 0.10 × 0.03 0.21 × 0.14 × 0.12

0.18 × 0.16 × 0.13

0.78–27.00

–20 ≤ h ≤ 20

–15 ≤ k ≤ 26

–33 ≤ l ≤ 33

57 597

36 583 (0.0817)

99.5

None

1.01–25.00

–14 ≤ h ≤ 21

–47 ≤ k ≤ 47

–21 ≤ l ≤ 23

62 953

22 389 (0.1916)

99.9

None

1.37–27.00

–20 ≤ h ≤ 20

–50 ≤ k ≤ 51

–10 ≤ l ≤ 22

35 587

16 642 (0.0509)

99.9

None

1.05–25.00

–21 ≤ h ≤ 18

–23 ≤ k ≤ 15

–44 ≤ l ≤ 44

69 284

46 258 (0.1000)

98.6

Empirical^b

0.78–27.00

–18 ≤ h ≤ 20

–26 ≤ k ≤ 25

–33 ≤ l ≤ 33

57 366

36 498 (0.0387)

99.4

Empirical^b

Reflections collected

Indep. reflections (R(int))

Compl. to θ max (%)

Absorption correction

Data/restraints/param.

Goodness-of-fit on F²

22 389/0/1 135

0.938

16 642/2/1 126

0.965

46 258/21/2 251

0.928

36 498/0/1 639

0.961

36 583/0/1 631

0.776

Final R indices (I > 2σ(I)) R1 = 0.0733

wR2 = 0.1517

R1 = 0.0434

wR2 = 0.0984

R1 = 0.0504

wR2 = 0.1004

1.765 and –1.530

R1 = 0.1026

wR2 = 0.2264

R1 = 0.2094

wR2 = 0.2632

1.295 and –1.619

R1 = 0.0611

wR2 = 0.1489

R1 = 0.0805

wR2 = 0.1577

2.172 and –1.566

R1 = 0.0612

wR2 = 0.1147

R1 = 0.1048

wR2 = 0.1235

1.244 and –1.090

R indices (all data)

R1 = 0.1362

wR2 = 0.1732

2.872 and –1.777

Peak and hole (e Å^–3)

^aIn all cases the temperature = 193(2) K, the wavelength = 0.710 73 Å, and the refinement method full-matrix least-squares on F².

^bEmpirical absorption correction was applied using the SADABS v2.10 program (36).

Table 2. A comparison of structural and spectroscopic properties of selected [HIPTN₃N]W and [HIPTN₃N]Mo complexes.^a

+b

Compound property

MoN₂

WN₂

WN=NK

MoN=NH

WN=NH

Mo=NNH₂

W=NNH₂

M-N_α(Å)

1.963(5)

1.061(7)

1990

1924

—

1.945(6)

1.132(8)

1888

1826

—

1.849(7)

1.220(8)

1745

1.810(7)

1.25(1)

1587

1.732(9)

1.29(1)

1542

1.737(3)

1.310(5)

—

1.745(3)

1.315(5)

—

N_α-N_β(Å)

ν(¹⁴N_α¹⁴N_β) (cm^–1)

ν(¹⁵N_α¹⁵N_β) (cm^–1)

δ(¹⁵N_α¹⁵N_β) (ppm)

¹J(¹⁵N_β¹⁵N_α, H) (Hz)

1688

1523

1499

—

357, 350

409, 233

394, 229

358, 140

344, 131

—

11.2

15.9, 53.6

15.4, 53.1

11.2, 90.5

10.1, 83.6

^aIR and ¹⁵N NMR data in C₆D₆at 22 °C, unless stated otherwise. The Mo compounds can be found in refs. 33 and 34.

^bAnion = BAr₄′^–.

Doubly labeled W-¹⁵N=¹⁵NK reveals sharp doublet reso-

nances for ¹⁵N_α(at 357 ppm) and ¹⁵N_β(at 350 ppm) in the

¹⁵N NMR spectrum in C₆D₆at room temperature. The ¹⁵N_α

and ¹⁵N_βresonances in W-¹⁵N=¹⁵NK are relatively close to-

gether, yet fall well within the range typical for compounds

of this general type (48, 49). The ¹H NMR spectrum of

W-N=NK in C₆D₆at room temperature features slightly

broadened resonances, yet is otherwise consistent with a

C_3v-symmetric structure. Apparently the interactions be-

tween the potassium ion and the aryl rings are weak enough

so that intramolecular ligand fluxionality in W-N=NK, in-

cluding rotation of the HIPT rings around the N—C_ipso

bonds, is not hindered significantly at room temperature.

Treatment of a C₆D₆solution of W-N=NK with Bu₄NCl

results in a rapid color change from maroon-red to dark

1

green. Both IR and H NMR spectra of the green solution

show complete conversion of W-N=NK to a single diamag-

netic product with a ν_NNfrequency shifted to higher energy

346

Can. J. Chem. Vol. 83, 2005

Fig. 4. Thermal ellipsoid drawing of WN₂(50% probability level, hydrogen atoms omitted).

by 56 cm^–1from what it is in W-N=NK. These observations

are analogous to those made for related Mo diazenides (33)

and indicate formation of a species in which there is essen-

tially no interaction between the cation and the β nitrogen,

i.e., formation of [W-N=N]^–. However, all attempts to crys-

tallize [W-N=N][Bu₄N] (e.g., by precipitation with pentane)

have failed. This compound appears to decompose to yield

two major pentane-soluble diamagnetic tungsten complexes

and Bu₃N. We speculate that part of the decomposition

consists of a reaction between [WN₂]^–and Bu₄N⁺to yield

W-N=NBu and Bu₃N. In contrast, [Mo-N=N][Bu₄N] can be

isolated and is stable at room temperature (33). A greater

nucleophilicity of [WN₂]^–compared to [MoN₂]^–at N_β

would be consistent with the observed greater degree of

backbonding from tungsten to the N₂unit and would not be

surprising.

Fig. 5. Electrochemical behavior of MN₂(M = Mo or W) in

′

0.1 mol/L [Bu₄N][BAr₄]/PhF, recorded at a glassy carbon elec-

trode (3 mm diam.) at a scan rate of 500 mV/s. The potential

+/0

scale is referenced to FeCp₂

.

Synthesis of a terminal dinitrogen complex of W(III)

Like various {Mo-N=N}^–salts (33), WN=NK is readily

oxidized in THF by ZnCl₂to yield zinc metal and the neu-

tral dinitrogen complex, WN₂, essentially quantitatively. The

WN₂complex was isolated as an orange solid in 53% yield

1

starting from WCl and has been fully characterized. The H

although one might expect the degree of this type of lib-

ration of a bound dinitrogen to decrease as backbonding to

dinitrogen increases. There is clearly a greater degree of

backbonding to N₂in WN₂than in MoN₂also according to

ν(N-N) frequencies, i.e., the value for ν(N-N) is nearly

100 cm^–1lower in WN₂(1888 cm^–1) than in MoN₂(1990 cm^–1).

All other structural details are normal for [HIPTN₃N]^3–com-

plexes (see supporting information).⁴

NMR spectrum of WN₂in C₆D₆resembles that of the low-

spin d³MoN₂analog (33), although the ligand backbone res-

onances of the W complex are broader and shifted with re-

spect to those for MoN₂(at 35, –11, and –23 ppm for WN₂

and at 23 (NCH₂), –5 (4,6-H), and –34 (NCH₂) ppm in

MoN₂). WN₂is the only known terminal dinitrogen complex

of tungsten in which a triamido/amine ligand is present and

the only one of W(III) to our knowledge.

A single crystal X-ray study of WN₂(Fig. 4) shows it

to have a structure analogous to that of MoN₂(33) with an

N—N bond distance of 1.132(8) Å (cf. 1.098 Å in free N₂

and 1.061(7) Å in MoN₂). The observed “shorter” N—N

bond length in MoN₂than in free dinitrogen is believed to

be a consequence of a slight “libration” of the bound dinitro-

gen about the center of the N—N bond, which would make

the N—N bond appear slightly shorter than it actually is

(34). The actual N—N bond distance in WN₂therefore actu-

ally also could be somewhat longer than what is observed,

Electrochemical reduction of WN₂to [WN₂]^–is reversible

+/0

and takes place at –2.27 V vs. FeCp₂

in 0.1 mol/L

′

[Bu₄N][BAr₄]/PhF electrolyte (Fig. 5) (the advantages of

[Bu₄N]⁺salts of relatively unreactive and weakly coordinat-

ing anions as electrolytes in polar, but relatively unreactive

solvents such as fluorobenzene, were demonstrated more

than 10 years ago (50), and have been explored more sys-

tematically by Geiger and co-workers in the last several

0/–

years (51–55)). The WN₂potential is 260 mV more nega-

tive than the corresponding E°′(MoN₂^0/–) value (34), consis-

tent with tungsten being more difficult to reduce than

Yandulov and Schrock

347

molybdenum, as expected. The tungsten dinitrogen complex

also undergoes a reversible electrochemical oxidation, but at

the same potential as the molybdenum dinitrogen complex

N=NH group. A greater degree of backbonding from W is

also evident from a lower value for ν(N-N) in WN=NH

(1542 cm^–1) than in MoN=NH (1587 cm^–1). Each molecule

of WN=NH is noticeably asymmetric within the

[HIPTN₃N]^3–ligand. One W—N(amide) distance is slightly

shorter than the other two in each molecule, and the HIPT

group attached to that amido nitrogen is rotated out of its

plane by an average of 68° in the direction opposite to the

average turn of 28° found for the other two HIPT groups in

each molecule. As a result, the three HIPT groups in each

molecule do not form a pseudo three-fold symmetric propel-

ler arrangement around the apical pocket similar to what is

found in 13 other [HIPTN₃N]Mo and [HIPTN₃N]W deriva-

tives. The hydrazide ligand is bent at both W (N5-W-N4 =

173.1(4)°) and N5 (W-N5-N6 = 173.6(9)°) in the same di-

rection, primarily toward the unique N(amide) (N3A in the

molecule shown in Fig. 6). The diazenide hydrogen was not

located, but was added and refined in the position shown. It

should be noted that N_α-N_β-R angles in diazenido compounds

vary, e.g., from 132° in [(Me₃SiNCH₂CH₂)₂NCH₂CH₂NMe₂)]-

Mo(CH₃)(N₂TMS) to 170° in {[(Me₃SiNCH₂CH₂)₂N-

CH₂CH₂NMe₂]Mo-N=NSiMe₃}OTf (56). Therefore, we

cannot say with any degree of certainty exactly where H_βis

located. To our knowledge, MoN=NH (33, 34) and WN=NH

are the only structurally characterized examples of parent

diazenide complexes of any metal.

+/0

(–0.66 V, Fig. 5). The similarity in the MN₂potentials for

Mo and W can be ascribed to an effective “leveling” of the

HOMO energies in the two MN₂complexes as a conse-

quence of increased π donation from the metal to N₂in WN₂

(∆ν(N-N) = 102 cm^–1, vide supra). However, the HOMO en-

ergies in [MoN₂]^–and [WN₂]^–are not approximately the

same, and [WN₂]^–therefore remains a stronger one-electron

reductant than [MoN₂]^–by 260 mV under the conditions de-

scribed.

′

Synthesis of WN=NH and [W=NNH₂][BAr₄]

Protonation of WN=NK by [Et₃NH][OTf] in benzene is

virtually instantaneous and yields WN=NH essentially quan-

titatively (eq. [3]). Diamagnetic WN=NH was isolated in

55% yield starting from WCl and has been fully character-

1

ized. The N=NH resonance appears at 8.43 ppm in the H

NMR spectrum in C₆D₆, flanked by ¹⁸³W satellites with

³J(¹⁸³W-H) = 17.4 Hz. The IR spectrum in C₆D₆exhibits a

ν(N-N) absorption at 1542 cm^–1, a frequency 45 cm^–1lower

than the corresponding value in MoN=NH (34). The ¹H

NMR spectrum of doubly labeled W-¹⁵N=¹⁵NH features a

doublet of doublets for the ¹⁵N=¹⁵NH resonance (¹J(¹⁵N_β-

2

H) = 53.1 Hz, J(¹⁵N_α-H) = 6.7 Hz) and two doublets of

doublets at 394 (¹⁵N_α) and 229 (¹⁵N_β) ppm (¹J(¹⁵N-¹⁵N) =

15.4 Hz) in the ¹⁵N NMR spectrum. These NMR data are

similar to those observed for MoN=NH (Table 2). Unlike

MoN=NH, however, WN=NH is quite stable thermally, suf-

fering only minor decomposition after a solution in C₆D₆

had been heated to 65 °C for 11 days. In contrast, MoN=NH

decays in C₆D₆with a half-life of ~7 h at 61 °C, forming

predominantly MoH, when prepared via a method analogous

to that shown in eq. [3]. This decomposition was shown to

be promoted by traces of acid (33) that are believed to re-

main after the Mo reaction analogous to that shown in eq. [3].

The X-ray crystal structure of WN=NH is shown in

Fig. 6. Owing to the presence of two independent molecules

in the unit cell and the difficulty of obtaining adequately

sized crystals of this compound, the quality of this structure

is mediocre (R1 = 10%). However, it was possible to refine

all non-hydrogen atoms anisotropically. The critical geomet-

rical parameters of the two crystallographically independent

molecules are statistically indistinguishable (within 3σ). The

average W—N_α(1.732(9) Å) and N_α—N_β(1.286(11) Å)

distances are shorter and longer, respectively, than those in

MoN=NH (1.810(7) and 1.25(1) Å, Table 2), consistent with

a greater degree of backbonding from tungsten into the

Protonation of WN=NH at N_βwith [H(OEt)₂][BAr₄] in

′

ether affords the hydrazido(2–) species [W=NNH₂][BAr₄]

essentially quantitatively (eq. [4]). This W(VI) species can

be isolated in good yield as a pale-yellow solid and has been

+

fully characterized. The NNH₂resonance appears at

1

5.96 ppm in the H NMR spectrum in C₆D₆, flanked by

barely resolved ¹⁸³W satellites (¹J(¹⁸³W-H) = 6 Hz). Labeled

15 15

′

[W= N NH₂][BAr₄] features a doublet of doublets for the

+

¹⁵N¹⁵NH₂resonance in the ¹H NMR spectrum (¹J(¹⁵N_β-

H) = 83.6 Hz, J(¹⁵N_α-H) = 1.8 Hz) and a doublet at 344

2

+

(¹⁵N_α-¹⁵N_βH₂) and a triplet of doublets at 131 ppm (¹⁵N_α-

+

¹⁵N_βH₂, J(¹⁵N-¹⁵N) = 10.1 Hz) in the ¹⁵N NMR spectrum.

′

While the spectra of [W=NNH₂][BAr₄] are generally similar

to those of [Mo=NNH₂][BAr₄], ¹⁵N NMR data for

′

[W=NNH₂][BAr₄] reveal a reduction of ¹J(¹⁵N_β-H) to

′

83.6 Hz in [W=NNH₂][BAr₄] from 90.5 Hz in [Mo=NNH₂]-

′

[BAr₄] (34), consistent with an increased p character in the

bonding at N_βas a consequence of a greater degree of

′

backbonding to the NNH₂unit in [W=NNH₂][BAr₄].

The structure of [W=NNH₂][BAr₄] (Fig. 7) is similar to

other {[HIPTN₃N]M}[BAr₄] (M = Mo, W) derivatives that

have been characterized in an X-ray study. The N_α—N_β

′

distance (1.315(5) Å) is essentially identical to that in

348

Can. J. Chem. Vol. 83, 2005

Fig. 6. Thermal ellipsoid drawing of W-N=NH (50% probability level, selected hydrogen atoms omitted).

′

difference therefore better reflects the difference in the

HOMO energy that results from substitution of W for Mo

alone. Consistent with this hypothesis, a large difference be-

tween Mo and W is also found for the M(NH₃)^+/0potentials

(vide infra).

[Mo=NNH₂][BAr₄] (1.310(5) Å). Since the N_α—N_βdis-

tance in [W=NNH₂][BAr₄] is similar to that in WN=NH

′

(1.29(1) Å), the amount of N_α—N_βdouble bond character

must be significant, as one would expect in a M-N-N π sys-

tem of this type. The NNH₂hydrogens were not located,

+

but were refined in calculated positions. Regardless of the

particular orientation chosen for their representation, their

presence in the model had a slight, but constant, effect on

the refined N—N distance.

′

Synthesis of [W(NH₃)][BAr₄], its structure, and

reactivity

′

Electrochemical reduction of [W=NNH₂][BAr₄] in 0.1 mol/L

When WCl is treated with NaBAr₄and NH₃in PhF

[W(NH₃)][BAr₄] is formed nearly quantitatively (eq. [5]).

This reactivity mirrors that established for MoCl (33), ex-

cept that fluorobenzene had to be used as a solvent because

WCl was found to be incompatible with dichloromethane.

The ammonia adduct was isolated in good yield and fully

′

[Bu₄N][BAr₄]/PhF at a glassy carbon electrode is irrevers-

ible at scan rates of up to 1 V/s and takes place at I_pc

=

+/0

–2.2 V vs. FeCp₂(100 mV/s scan rate). The Mo analog,

[Mo=NNH₂][BAr₄], undergoes a quasi-reversible reduction

at –1.56 V under the same conditions (34), which allows us

to estimate the difference between the Mo=NNH₂

W=NNH₂couples as being 400–500 mV. The magnitude

of this difference is considerably greater than that found for

[MN₂]^0/–(260 mV) and [MN₂]^+/0(0 mV) potentials (M =

Mo or W, vide supra). Apparently π donation of electron

density from M into the [NNH₂]^2–system does not differ

′

+/0

1

and

characterized. Its H NMR spectrum in C₆D₆features a pat-

+/0

tern of paramagnetically shifted and broadened resonances

characteristic of a high-spin d²W environment in complexes

of this general type. The resonances in W(NH₃)⁺are

paramagnetically shifted to a greater extent than those of

WCl, a phenomenon that was also observed for Mo(NH₃)⁺

vs. MoCl.

+/0

substantially for Mo and W, and the M=NNH₂potential

Yandulov and Schrock

Fig. 7. Thermal ellipsoid drawing of [W=NNH₂][BAr₄] (50% probability level, selected hydrogen atoms omitted).

349

′

The X-ray crystal structure of [W(NH₃)][BAr₄] (Fig. 8) is

essentially identical to that of [Mo(NH₃)][BAr₄] (33). The

of [Mo(NH₃)][BAr₄] is fully reversible both in PhF

(0.1 mol/L [Bu₄N][BAr₄]) at a glassy carbon electrode and

′

M—N(ammonia) distances are 2.210(4) Å in [W(NH₃)][BAr₄]

vs. 2.24(1) Å in [Mo(NH₃)][BAr₄] (Table 2). The ammonia

protons were not located, but are shown in calculated posi-

tions.

Electrochemical reduction of [W(NH₃)][BAr₄] in 0.1 mol/L

THF (0.4 mol/L [Bu₄N][PF₆]) at a platinum disk at all scan

rates and takes place at E°′(Mo(NH₃)^+/0) = –1.63 V vs.

′

+/0

′

FeCp₂in PhF (0.1 mol/L [Bu₄N][BAr₄]). The difference

in M(NH₃)^+/0potentials (M = Mo or W) is thus ~430 mV,

+/0

similar to the difference between the Mo=NNH₂

W=NNH₂potentials (vide supra). Chemical reductions of

both [W=NNH₂][BAr₄] and [W(NH₃)][BAr₄] are described

and

′

+/0

′

[Bu₄N][BAr₄]/PhF at a glassy carbon electrode reveals an ir-

′

reversible wave at a scan rate of 0.5 V/s (Fig. 9). The couple

appears to become more reversible at faster scan rates (1 and

2 V/s) at –2.06 V vs. FeCp₂^+/0, although it cannot be called

reversible even at 2 V/s. In contrast, one-electron reduction

below.

Deprotonation of [W(NH₃)]⁺with LiN(SiMe₃)₂/4TMEDA

in C₆D₆yields a single paramagnetic product, which we pre-

350

Can. J. Chem. Vol. 83, 2005

′

Fig. 8. Thermal ellipsoid drawing of [W(NH₃)][BAr₄] (50% probability level, selected hydrogen atoms omitted).

sume to be (at least initially) W(NH₂). However, attempts to

isolate the product in analytically pure crystalline form so

′

far have failed. Treatment of [W(NH₃)][BAr₄] with triethy-

lamine and [FeCp₂][PF₆] in C₆D₆, followed by LiN(SiMe₃)₂

yielded WϵN essentially quantitatively (eq. [6]). The nitride can

be prepared more conveniently via salt metathesis of WCl

with 3 equiv. of NaN₃in a mixture of C₆H₆and MeCN over

a period of 16 h at room temperature (eq. [7]), the method

utilized originally in the synthesis of [Me₃SiN₃N]WϵN (42).

The IR spectra of WϵN and MoϵN in C₆D₆are virtually

identical in the 1200–900 cm^–1region, with the exception

of an absorption at 1012 cm^–1in the MoϵN spectrum

(ν(Mo-N)) and at 1024 cm^–1in the WϵN spectrum. There-

fore, we assign the 1024 cm^–1absorption to the ν(W-N)

stretch. As in [Me₃SiN₃N]MϵN complexes, ν(M-N) is

slightly higher for WϵN than it is in MoϵN. The ¹⁵N NMR

spectrum of Wϵ¹⁵N shows a resonance for the nitride at

827 ppm (vs. 898 ppm in Moϵ¹⁵N).

′

Treatment of a C₆D₆solution of [W(NH₃)][BAr₄] with

LiBHEt₃(1 mol/L in THF) results in a color change to

brown and formation of a largely one paramagnetic species,

which is tentatively assigned as the monohydride complex,

WH, on the basis of reactivity similar to that exhibited by

MoH (33). However, a small amount of a diamagnetic prod-

uct is present initially, which becomes the sole observable

species when the reaction mixture is exposed to an atmo-

sphere of H₂for ~10 min (eq. [8]); this diamagnetic species

(WH₃) can be isolated in 42% yield. This compound was

identified on the basis of the similarity of its NMR and IR

spectra with those of the crystallographically characterized

Yandulov and Schrock

351

′

Fig. 9. Electrochemical behavior of [W(NH₃)][BAr₄] in

sults in a subtle color change, formation of a yellow precipi-

tate of [CoCp₂][BAr₄], complete consumption of the acid,

′

0.1 mol/L [Bu₄N][BAr₄]/PhF, recorded at a glassy carbon elec-

trode (3 mm diam.) at the scan rates shown in V/s. The potential

and conversion of approximately half of the starting WN₂to

+/0

′

scale is referenced to FeCp₂

.

[W=NNH₂][BAr₄] in 3 h at 22 °C (eq. [9]). This mixture re-

mains unchanged over a period of 24 h. WN=NH is con-

′

verted quantitatively into [W=NNH₂][BAr₄] upon reaction

with 1 equiv. of [2,6-LutH][BAr₄] in C₆D₆. Therefore, we

′

propose that WN=NH is formed in the reduction of WN₂in

the presence of [2,6-LutH][BAr₄] and that WN=NH then

consumes proton relatively irreversibly to give

[W=NNH₂][BAr₄], thereby limiting consumption of WN₂

when 1 equiv. of acid is present to ~50%. In contrast, the

analogous reaction between MoN₂and [2,6-LutH][BAr₄]

and CoCp₂rapidly yields MoN=NH quantitatively (34).

[Mo=NNH₂][BAr₄] is considerably more acidic than

′

a

′

[W=NNH₂][BAr₄], as judged from the fact that treatment of

′

MoN=NH with 1 equiv. of [2,6-LutH][BAr₄] in C₆D₆gives

′

an equilibrium mixture of MoN=NH and [Mo=NNH₂][BAr₄]

(44:56); therefore, conversion of MoN₂to MoN=NH can be

complete when MoN₂is treated with only 1 equiv. of [2,6-

LutH][BAr₄] and CoCp₂. The most important point is that a

′

one-electron reduction–protonation of WN₂to yield

WN=NH is feasible, although this process appears to be

slower than conversion of MoN₂to MoN=NH under analo-

gous conditions.

complex, [Me₃SiN₃N]WH₃(57). Like [Me₃SiN₃N]WH₃,

WH₃features a low-field resonance of intensity three in its

′

Isolated [W=NNH₂][BAr₄] reacts rapidly with 2 equiv. of

1

¹H NMR spectrum in C₆D₆at 11.99 ppm with J(¹⁸³W-H) =

CoCp₂(E°′ = –2.01 V vs. FeCp₂^+/0in PhF) in C₆D₆, as indi-

*

24 Hz (cf. 10.01 ppm and 23 Hz in [Me₃SiN₃N]WH₃), in ad-

dition to characteristic ligand resonances. The IR spectrum

of WH₃in C₆D₆shows two broad absorptions at 1958 and

1966 cm^–1consistent with ν(W-H) stretching modes (cf.

1903 and 1887 cm^–1in [Me₃SiN₃N]WH₃in Nujol). How-

ever, unlike [Me₃SiN₃N]WH₃, WH₃can be induced to lose

H₂, by exposure to a high vacuum at 70 °C to yield a mix-

ture that contains a considerable amount of WH after a few

hours. This difference in behavior can be attributed to less

electron donation by the [HIPTN₃N]^3–ligand compared to

the [Me₃SiN₃N]^3–ligand. It should be noted that what is pro-

posed to be [HIPTN₃N]MoH₃cannot be isolated since it

readily loses dihydrogen to form [HIPTN₃N]MoH at room

temperature (34).

cated by the formation of

*

a yellow precipitate of

′

[CoCp₂][BAr₄]. However, instead of the expected one-

electron reduced (W(V)) product (W=NNH₂), a proton

NMR spectrum of the reaction mixture recorded 20 min af-

*

′

ter mixing [W=NNH₂][BAr₄] and CoCp₂revealed that only

WN=NH and WϵN were present in a 52:48 ratio. After

17 h, the ratio of WN=NH to WϵN had not changed, but a

well-resolved triplet for free ammonia could be observed.

This ammonia was collected by vacuum transfer and shown

(using the indophenol analytical method) to amount to 0.43

′

equiv. of the original [W=NNH₂][BAr₄]. Therefore, one-

′

electron reduction of [W=NNH₂][BAr₄] predominantly re-

sults in the net reaction shown in eq. [10]. In fact, this reac-

15

15 15

′

tion was used to prepare Wϵ N from [W= N NH₂][BAr₄].

This stoichiometry is consistent with the disproportionation

of the initial reduction product, W=NNH₂, a W(V) species,

into a W(IV) and a W(VI) species, WN=NH and WϵN, re-

spectively. A plausible sequence of elementary electron- and

Reduction of dinitrogen on W with protons and electrons

′

Treatment of WN₂with 1 equiv. of [2,6-LutH][BAr₄]

(2,6-Lut = 2,6-lutidine) and 2 equiv. of CoCp₂in C₆D₆re-

352

Can. J. Chem. Vol. 83, 2005

Scheme 1. A proposed sequence of reactions that yields the net

eq. [10] and (it is proposed) Scheme 1 (34). One of the rea-

sons to investigate the chemistry of [W=NNH₂][BAr₄] was

process shown in eq. [10]. Unobservable intermediates are shown

′

in boxes. The counterion (BAr₄^–) is excluded for clarity. PT =

′

the possibility that W=NNH₂would be more stable than

Mo=NNH₂and that W=NNH₂, therefore, could at least be

observed. Instead, however, we find that W=NNH₂is even

less stable than Mo=NNH₂. In each case, the high propen-

sity of M=NNH₂to disproportionate can be attributed to a

combination of a high Brønsted basicity of N_βin combina-

tion with strong reducing properties. We attribute the strong

reducing properties to the presence of one of the five elec-

trons in the M-N_α-N_βπ system in a π orbital that has a node

at N_α; four electrons are in the two perpendicular lowest en-

ergy bonding molecular orbitals. This combination of prop-

erties makes M=NNH₂susceptible to both protonation by its

own oxidized form and oxidation by its own protonated

form, at least in a thermodynamic sense. If both steps are

kinetically feasible, then disproportionation proceeds readily,

perhaps in part because of the fact that the step in which am-

monia is formed (N—N bond cleavage) is almost certainly

irreversible. It is important to note that disproportionation of

M=NNH₂intermediates has no adverse consequences for

overall reduction of dinitrogen to ammonia, since it merely

leads to species that are themselves susceptible to further re-

actions that lead to the formation of ammonia. Dispropor-

tionations of d¹species and related intermolecular proton or

electron transfer reactions are likely to complicate the mech-

anism of ammonia formation depicted in Fig. 2, when all

species are in solution.

proton transfer, ET = electron transfer.

proton-transfer steps that comprises the overall process is

proposed in Scheme 1.

*

′

[10] [W=N-NH₂]BAr ₄+ 1.0CoCp₂

C₆D₆

⎯⎯⎯→ 0.5W-N=NH + 0.5W≡N

Attempted conversion of W(NH₃)⁺to WN₂, and

catalytic reduction of N₂

*

′

+ 0.5 NH₃+ [CoCp₂]BAr₄

The results discussed in the previous section show that it

is possible to reduce WN₂to yield ammonia and WϵN in

the presence of acid and reductant. Since reduction of

MoϵN to Mo(NH₃)⁺is a relatively easy transformation in

the Mo dinitrogen reduction cycle (35), while reduction of

Mo(NH₃)⁺to Mo(NH₃) and transformation of Mo(NH₃) into

MoN₂can be observed (34), we examined the feasibility of

+

Initial one-electron reduction of W=NNH₂yields as yet

unobserved W=NNH₂, possibly in a rapid equilibrium. A

small amount of W=NNH₂then protonates W=NNH₂to

+

yield the proposed hydrazidium(2–) species (W=NNH₃),

+

while W=NNH₂is transformed into WN=NH. The

+

hydrazidium(2–) species (W=NNH₃) then oxidizes

W=NNH₂to regenerate W=NNH₂, and in the process is

reduced to form WϵN and ammonia. Essentially, the dispro-

portionation of W=NNH₂is catalyzed by W=NNH₂.

It is possible that the disproportionation shown in

Scheme 1 requires some time to proceed to completion and

that the initial “W(V) = NNH₂” exists in a fast electron-

transfer equilibrium with W=NNH₂and (or) CoCp₂

which causes extensive broadening and (or) complete disap-

pearance of the resonances for W=NNH₂and CoCp₂in

the H NMR spectrum. A similar situation was encountered

in the reversible reduction of{[HTBTN₃N]Mo(NH₃)}[BAr₄]

with CrCp₂in C₆D₆(37). However, irreversibility of the

converting W(NH₃)⁺into WN₂.

+

*

′

Reduction of [W(NH₃)][BAr₄] with 2 equiv. of CoCp₂in

+

C₆D₆proceeds to some extent, as judged by the rapid forma-

*

′

tion of a yellow precipitate of [CoCp₂][BAr₄]. However, the

¹H NMR spectrum of the resulting solution is unusually fea-

tureless, i.e., few resonances beyond isopropyl methyl reso-

nances and some aryl resonances can be observed. Since the

¹H NMR spectrum of Mo(NH₃) is similar to that of other d³

Mo complexes in this general family, we expected W(NH₃)

to be observable and to have a very similar spectrum to that

of Mo(NH₃).

+

*

,

+

*

1

′

*

′

The mixture of [W(NH₃)][BAr₄] and W(NH₃) remained

′

electrochemical reduction of [W=NNH₂][BAr₄] suggests

unchanged for days and also did not change after being ex-

posed to 800 psi (1 psi = 6.894 757 kPa) of N₂for 48 h.

Thermolysis at 60 °C for a day resulted only in minor de-

composition and formation of WϵN and WH₃. Treatment of

WN₂with ~4 equiv. of NH₃in C₆D₆under argon also re-

sulted in no change after 15 days, again in sharp contrast to

the behavior in the analogous molybdenum system where an

equilibrium between Mo(NH₃) into MoN₂is readily estab-

lished in a closed system over a period of several hours. Al-

though it has not been possible to transform W(NH₃) into

WN₂, W(NH₃) appears to be relatively stable, at least as

that the sequence of reactions following the initial electron

transfer takes place rapidly at 22 °C. We are puzzled by the

apparently fast rate of net proton transfer in the absence of

conspicuous proton-transfer mediators. However, we note

that small amounts of the free ligand, H₃[HIPTN₃N], are

typically formed in a decomposition reaction, and could eas-

ily mediate the proton-transfer step. The actual mechanism

of the proton-transfer step remains to be clarified.

′

Reduction of [Mo=NNH₂][BAr₄] at room temperature

proceeds in a manner closely analogous to that shown in

Yandulov and Schrock

Table 3. Results of the attempted catalytic reduction of dinitrogen with WN₂.^a

353

d

Reductant

E°′(THF)^b

E°′(PhF)^c

Acid

[2,6-LutH][BAr₄]

pK_a(H₂O)

Equiv. of NH₃

*

′

CrCp₂

–1.47

–1.84

–1.63

–2.01

6.68

10.67

1.31

1.51

0.62

*

′

[2,6-LutH][BAr₄]

′

[Et₃NH][BAr₄]

CoCp₂

*

CoCp₂

^aSingle runs of all experiments were carried out with 5.84 µmol of WN₂, 36 equiv. of the reductant, and 48 equiv. of the acid,

under conditions otherwise identical to those previously described (35, 37).

+/0

^bV vs. FeCp_2+/0in 0.4 mol/L [Bu₄N][PF₆].

^cV vs. FeCp₂in 0.1 mol/L [Bu₄N][BAr₄′].

^dThe maximum possible is 12 equiv. of ammonia (36 electrons available).

+

*

′

generated by reduction of [W(NH₃)][BAr₄] with CoCp₂.

dinitrogen at various stages of reduction (WN=NH₂, WϵN,

and W(NH₃)⁺). As expected, tungsten variants of known

molybdenum complexes are characterized by greater nucleo-

philicities (at N_β) and lower reduction potentials. Interest-

ingly, however, the differences between Mo and W potentials

are considerably less pronounced for complexes that contain

However, this conclusion is difficult to reconcile with a lim-

ited degree of reversibility of the electrochemical reduction

′

of [W(NH₃)][BAr₄] ([Bu₄N][BAr₄]/PhF) (Fig. 9). There-

fore, the exact nature of the putative “W(NH₃)” remains

obscure, and we cannot exclude the possibility that some re-

action involving oxidation of W(III) (e.g., intramolecular

NH addition) is facile. In any case, it seems likely that

W(NH₃) is not strictly analogous to well-behaved Mo(NH₃)

(34).

0/–

a dinitrogen ligand, e.g., 260 mV for MN₂and 0 mV for

MN₂^+/0. Structural and spectroscopic data (Table 2) signify

the more electron-rich nature of all other tungsten deriva-

tives without exception. We had hoped to observe and possi-

bly isolate W variants of several proposed Mo species that

are either exceedingly difficult to isolate, and in some cases

even observe in situ. However, many of the W complexes,

e.g., [WN₂][Bu₄N] and W(NH₃), were in fact less stable

than the analogous Mo complexes. Also W=NNH₂dispro-

portionates into a mixture of WN=NH, WϵN, and ammonia

much faster than does Mo=NNH₂. Although dinitrogen in

WN₂could be reduced to ammonia with acid and reductant

combinations similar to those used in the catalytic reduction

of dinitrogen with [HIPTN₃N]Mo species, catalytic turnover

could not be achieved as a consequence of (inter alia) major

difficulties in converting W(NH₃)⁺into WN₂. Complications

of this magnitude in a system where Mo analogs are inter-

mediates in the catalytic reduction of dinitrogen to ammonia

would appear to doom tungsten as a plausible site for dini-

trogen reduction. If MoFe and WFe nitrogenases are struc-

turally analogous, and if dinitrogen is reduced at Mo in the

MoFe nitrogenase, then one can imagine why reduction of

dinitrogen by the WFe nitrogenase might fail.

Treatment of the mixture containing partly reduced

W(NH₃)⁺in C₆D₆with [Et₃NH][BAr₄] produced a new

′

paramagnetic species that gave rise to considerable quanti-

ties of both W(NH₃)⁺and WH₃after several days, but no

WN₂or products of its reduction. The net reaction of

W(NH₃)⁺with a reductant and an acid thus results primarily

in reduction of protons (presumably) and formation (ulti-

mately) of WH₃(eq. [11]), but dinitrogen is not bound or re-

duced.

*

′

[11] [W(NH₃)][BAr₄] + 4CoCp₂+ 3[Et₃NH][BAr₄]

*

′

⎯⎯→ WH₃+ NH₃+ 4[CoCp₂][BAr₄] + 3Et₃N

In view of the previous observations, it is perhaps not sur-

prising that attempted catalytic reduction of dinitrogen using

WN₂as the catalyst under conditions identical or similar to

those employed for catalytic reduction of dinitrogen by Mo

complexes (35) failed (Table 3). This can be attributed

chiefly to the inability to convert W(NH₃)⁺into WN₂to any

significant extent on the timescale of a typical catalytic ex-

periment (6 h), as was independently established in the ex-

periments previously described. Ammonia is formed, but

less than 2 equiv., i.e., at most 1.51 equiv. with CoCp₂as the

reductant and [2,6-LutH][BAr₄] as the acid. Therefore,

while it is possible to cleave the N—N bond and form the

first equivalent of ammonia, and to form some WNH_x^y+spe-

cies that yield additional ammonia upon workup, dinitrogen

is not bound and reduced further under the conditions em-

ployed.

Experimental section

*

General

All air- and moisture-sensitive compounds were manipu-

lated using standard Schlenk and glovebox techniques under

an atmosphere of nitrogen in flame- and oven-dried glass-

ware (including NMR tubes). Ether, pentane, and toluene

were purged with nitrogen, passed through activated alumina

columns (58), and freeze–pump–thaw degassed three times.

THF, DME, 1,4-dioxane, and benzene were distilled from

dark purple Na–benzophenone ketyl solutions. Dichloro-

methane and acetonitrile were distilled from CaH₂, while

fluorobenzene was distilled from P₂O₅under N₂. All dried

and deoxygenated solvents were stored over molecular

sieves in a nitrogen-filled glovebox. Small quantities of hy-

drocarbon and ether solvents, as well as all such deuterated

NMR solvents, were purified a second time by vacuum

transfer from dark-purple solutions and (or) suspensions of

Na–benzophenone ketyl (or CaH₂for CH₂Cl₂, PhF, and

′

Conclusions

We have been able to synthesize a number of [HIPTN₃N]W

complexes analogous to the [HIPTN₃N]Mo species that are

proposed intermediates in the catalytic reduction of

dinitrogen to ammonia (35), among them the first terminal

dinitrogen complex of W(III) (WN₂), the second structurally

characterized example of a parent diazenide of any metal

(WN=NH), and three other tungsten derivatives containing

354

Can. J. Chem. Vol. 83, 2005

CD₂Cl₂), degassed, and stored in gas-tight solvent bulbs inside

a glovebox. Molecular sieves (3 Å) and Celite^®were acti-

vated in vacuo at 230 °C for several days. Cobaltocene (sub-

the residue with pentane (60 mL) afforded a dark orange

microcrystalline solid, which was collected by filtration on a

glass frit after standing the mixture at –35 °C overnight,

washed with cold pentane, and dried in vacuo at 85 °C for

*

(sublimed), [CoCp₂][PF₆] (Strem),

limed), CrCp₂

1

(Me₃Si)₂NLi (sublimed), anhydr. ZnCl₂, LiBHEt₃(1 mol/L

THF), and NaN₃(Aldrich) were used as received, unless in-

dicated otherwise. WCl₄(DME) (59), H₃[HIPTN₃N] (33),

several hours. Yield: 11.28 g (6.25 mmol, 67%). H NMR

(C₆D₆, 20 °C) δ: 18.81 (s, 6H, 4′,6′-H), 7.49 (s, 12H,

3,5,3′′,5′′-H), 3.47 (br s, 12H, 2,6,2′′,6′′-CHMe₂), 3.18 (sept,

J_HH= 6.6 Hz, 6H, 4,4′′-CHMe₂), 2.98 (s, 3H, 2′-H), 1.58 (d,

J_HH= 6.6 Hz, 36H, 4,4′′-CH(CH₃)₂), 1.34 (s, 36H,

2,6,2′′,6′′-CH(CH₃)₂), 0.83 (br s, 36H, 2,6,2′′,6′′-CH(CH₃)₂),

–27.06 (s, 6H, NCH₂), –50.1 (br s, 6H, NCH₂). Anal. calcd.

for C₁₁₄H₁₅₉ClN₄W (%): C 75.87, H 8.88, N 3.10, Cl 1.96;

found: C 75.97, H 8.78, N 2.97, Cl 2.10.

′

NaBAr₄(60), [H(OEt₂)₂][BAr₄] (60), and [2,6-LutH][BAr₄]

(35) were prepared according to the published procedures or

slight modifications thereof. Potassium graphite (KC₈) was

prepared by stirring potassium metal, freshly cut in a nitro-

gen-filled glovebox, and graphite in 1:8 stoichiometry with a

glass-coated stirbar under 700 torr (1 torr = 133.322 Pa) of

*

Ar at 140 °C for an hour. CoCp₂was prepared from

*

[CoCp₂][PF₆] and a slight excess of KC₈in THF over 4 h

[HIPTN₃N]WN=NK

′

and sublimed. [Et₃NH][BAr₄] was prepared from Et₃NHCl

A mixture of WCl (500 mg, 0.277 mmol) and KC₈

(93.6 mg, 0.692 mmol) was stirred in benzene (6 mL) under

1 atm of N₂(1 atm = 101.325 kPa) with a glass-coated mag-

netic stirbar for 42 h. The resulting dark maroon-red mixture

was taken to dryness in vacuo and the solid residue was ex-

tracted with pentane. The pentane extracts were filtered

through Celite^®and concentrated to 5 mL; a black-red crys-

talline solid formed within in minutes. The mixture was kept

at –35 °C for several days and the product was then col-

lected on a frit, washed with cold pentane, and dried in

vacuo at 35 °C for 2 h. Yield: 310 mg (0.169 mmol, 61%).

′

and 1.05 equiv. of NaBAr₄in CH₂Cl₂. All metal complexes

were stored in the dark, under N₂at –35 °C. ¹H, ¹⁹F, and ¹⁵

N

NMR spectra were recorded on a Varian Mercury 300 (¹H

300 MHz, ¹⁹F 282 MHz), Varian Unity 300 (¹H 300 MHz)

or a Varian Inova 500 (¹H 500 MHz, ¹⁵N 50.7 MHz) spec-

trometers and referenced to the residual protio solvent peaks

(¹H) or external neat PhF (¹⁹F, –113.15 ppm relative to

CFCl₃) and neat MeCN (¹⁵N, +245.5 ppm relative to neat

NH₃at 303 K (61)). Attempted catalytic dinitrogen reduc-

tion runs and analyses of ammonia by the indophenol method

(62–65) utilized procedures, equipment, and reagents de-

scribed previously (35, 37). IR spectra were recorded on a

Nicolet Avatar 360 FT-IR spectrometer in 0.2 mm KBr solu-

tion cells; frequencies are given in cm^–1. Elemental analyses

were performed by H. Kolbe Mikroanalytisches Laborator-

ium, Mülheim an der Ruhr, Germany.

1

IR (C₆D₆): 1745 (ν_N-N). IR (pentane): 1747 (ν_N-N). H NMR

(C₆D₆, 20 °C) δ: 7.54 (br s, 6H, 4′,6′-H), 7.16 (s, 12H,

3,5,3′′,5′′-H), 6.61 (s, 3H, 2′-H), 3.75 (br s, 6H, NCH₂),

3.42 (sept, J_HH= 6.8 Hz, 12H, 2,6,2′′,6′′-CHMe₂), 2.84 (br

sept, J_HH= 6 Hz, 6H, 4,4′′-CHMe₂), 1.93 (br s, 6H, NCH₂),

1.26 (d, J_HH= 6.5 Hz, 36H, 4,4′′-CH(CH₃)₂), 1.24 (d, J_HH

=

6.0 Hz, 36H, 2,6,2′′,6′′-CH(CH₃)₂), 1.2 (br sh, 36H,

2,6,2′′,6′′-CH(CH₃)₂). Anal. calcd. for C₁₁₄H₁₅₉KN₆W (%):

C 74.56, H 8.73, N 4.58; found: C 74.43, H 8.59, N 4.45.

X-ray quality crystals were obtained from a supersatu-

rated heptane solution at room temperature.

Electrochemistry

Electrochemical measurements were carried out using a

′

BAS CV-50W potentiostat, 0.1 mol/L [Bu₄N][BAr₄] in PhF

electrolyte, and a standard three-electrode cell assembly with

glassy carbon (3.0 mm diam.) disk working electrode, plati-

num wire auxiliary electrode, and a AgCl-coated silver wire

reference electrode. All measurements were referenced inter-

nally with FeCp₂, CoCp₂, or MoN₂(35), as appropriate.

Compensation of internal solution resistance was used

to lower the peak separation of reversible couples to 100–

200 mV using the standard routines available with the BAS

CV-50W potentiostat and accompanying software. An expla-

nation of the technique can be found on the web at

http://www.epsilon-web.net/Ec/manual/Setup/ir_comp.html.

[HIPTN₃N]W¹⁵N=¹⁵NK

A mixture of WCl (500 mg, 0.277 mmol) and KC₈

(93.6 mg, 0.692 mmol) in benzene (6 mL), prepared under

regular N₂in a 25 mL flask fitted with a Teflon^®valve, was

freeze–pump–thaw degassed five times, pressurized to 1 atm

(1 atm = 101.325 kPa) with ¹⁵N₂purified with solid Na–

Ph₂CO and stirred with a glass-coated magnetic stirbar for

50 h. A work-up procedure analogous to that used in the

preparation of the unlabeled complex yielded 285 mg

(0.155 mmol, 56%) of a black-red solid, identified by spec-

troscopic data as the labeled compound enriched with ¹⁵N at

the N_αand N_βdiazenido sites to the extent of only 92% as a

consequence of some initial reaction under ¹⁴N₂to yield

[HIPTN₃N]W¹⁴N=¹⁴NK. IR (C₆D₆): 1745 (ν_14N–14N),

1688 (ν_15N–15N). ¹⁵N NMR (C₆D₆, 20 °C) δ: 356.6 (d,

J(¹⁵N_α-¹⁵N_β) = 11.2 Hz, ¹⁵N_α), 349.5 (d, J(¹⁵N_α-¹⁵N_β) =

11.2 Hz, ¹⁵N_β).

[HIPTN₃N]WCl

A mixture of WCl₄(DME) (3.897 g, 9.37 mmol) and

H₃[HIPTN₃N] (14.93 g, 9.40 mmol) was stirred vigorously

in benzene (100 mL) for 2 h to give a dark red-brown homo-

geneous solution. Solid (Me₃Si)₂NLi (4.868 g, 29.1 mmol)

was added in portions over a period of 10 min. The solution

was stirred for an additional 2 h and taken to dryness in

vacuo. The residue was dried at 70 °C overnight and ex-

tracted with benzene. The extracts were filtered through

Celite^®, and the filtrates were concentrated and centrifuged

in portions to remove residual LiCl. The resulting homoge-

neous solution was then taken to dryness in vacuo and the

residue was heated at 85 °C for several hours. Trituration of

[HIPTN₃N]WN₂

A mixture of WCl (500 mg, 0.277 mmol) and KC₈

(93.6 mg, 0.692 mmol) was stirred in benzene (6 mL) under

1 atm of N₂(1 atm = 101.325 kPa) with a glass-coated mag-

Yandulov and Schrock

355

[HIPTN₃N]W¹⁵N=¹⁵NH

netic stirbar for 42 h. The resulting mixture was filtered

through Celite^®and brought to dryness in vacuo. Solid

ZnCl₂(30.2 mg, 0.222 mmol) and THF (6 mL) were added

to the residue. After stirring the solution for a few minutes, a

dark orange-red solution and a precipitate of Zn metal

formed. The mixture was taken to dryness in vacuo and

heated to 75 °C for 2 h. The solid residue was extracted with

pentane and the pentane extracts were filtered through

Celite^®. The extracts were taken to dryness in vacuo. Addi-

tion of Me₄Si (7 mL) to the solid residue led to the forma-

tion of a microcrystalline orange solid after 1 h. The mixture

was kept at –35 °C for several days and the solid product

was collected on a frit, washed with cold Me₄Si, and dried

in vacuo at 75 °C for 2 h. Yield: 265 mg (0.147 mmol,

W¹⁵N=¹⁵NK (40 mg) was treated with [Et₃NH][OTf]

(1.05 equiv.) in C₆D₆in an NMR tube to yield the title com-

pound, 92% ¹⁵N-enriched, essentially quantitatively. The so-

lution was taken to dryness in vacuo and the residue was

dissolved in C₆D₆for spectroscopic analysis. IR (C₆D₆):

1

1499 (ν₁₅_Ν−₁₅_ΝΗ). H NMR (C₆D₆, 20 °C) δ: 8.42 (s, 0.08H,

1

2

¹⁴N_α=¹⁴N_βH), 8.41 (dd, J(¹⁵N_β-H) = 53.1 Hz, J(¹⁵N_α-H) =

6.7 Hz, J(¹⁸³W-H) = 17.1 Hz, 0.92H, ¹⁵N_α=¹⁵N_βH) (other

resonances are identical to those measured for the unlabeled

sample). ¹⁵N NMR (C₆D₆, 20 °C) δ: 393.5 (dd, J(¹⁵N_α-

2

¹⁵N_β) = 15.8 Hz, J(¹⁵N_α-H) = 6.5 Hz, ¹⁵N_α=¹⁵N_βH), 229.3

(dd, ¹J(¹⁵N_β-H) = 53.9 Hz, J(¹⁵N_α-¹⁵N_β) = 14.9 Hz,

¹⁵N_α=¹⁵N_βH).

1

53%). IR (C₆D₆): 1888 (ν_N-N). H NMR (C₆D₆, 20 °C) δ:

35.2 (br s, 6H, NCH₂), 6.79 (s, 12H, 3,5,3′′,5′′-H), 2.83 (s,

6H, 4,4′′-CHMe₂), 2.58 (br s, 36H, 2,6,2′′,6′′-CH(CH₃)₂),

1.29 (s, 36H, 4,4′′-CH(CH₃)₂), 0.66 (s, 36H, 2,6,2′′,6′′-

CH(CH₃)₂), –0.25 (s, 2H, 2′-H), –0.95 (s, 12H, 2,6,2′′,6′′-

CHMe₂), –11.1 (br s, 6H, 4′,6′-H), –22.5 (br s, 6H, NCH₂).

Anal. calcd. for C₁₁₄H₁₅₉N₆W (%): C 76.18, H 8.92, N 4.68;

found: C 75.93, H 9.08, N 4.57.

′

{[HIPTN₃N]W=NNH₂}[BAr₄]

An ether (4 mL) solution of WN=NH (100 mg,

55.6 µmol) was stirred and treated with solid

′

[H(OEt₂)₂][BAr₄] (59.1 mg, 58.4 µmol) at –35 °C. The solu-

tion was stirred for 1 h while being allowed to warm up to

room temperature, and then taken to dryness in vacuo. The

resulting solid was extracted with pentane and the pentane

extracts were filtered through Celite^®and concentrated to

~3 mL. Seeding this solution with a minute amount of

X-ray quality crystals were obtained from a supersatu-

rated heptane solution at room temperature.

′

{[HIPTN₃N]Mo=NNH₂}[BAr₄] (33) led to a gradual forma-

tion of pale yellow microcrystals over a period of ~1 h. The

mixture stood at –35 °C for several days and the product

was isolated by filtration, washed with cold pentane, and

dried in vacuo at 60 °C. Yield: 106 mg (39.8 µmol, 72%).

¹H NMR (C₆D₆, 20 °C) δ: 8.36 (br s, 8H, C₆H₃-3,5-(CF₃)₂-

2,6-H), 7.66 (br s, 4H, C₆H₃-3,5-(CF₃)₂-4-H), 7.13 (s, 12H,

3,5,3′′,5′′-H), 6.79 (br d, 6H, 4′,6′-H), 6.62 (br t, 3H, 2′-H),

5.96 (s, 2H, J(¹⁸³W-H) ~ 6 Hz, N-NH₂), 3.78 (br t, 6H,

NCH₂), 2.88 (sept, J_HH= 6.9 Hz, 6H, 4,4′′-CHMe₂), 2.67

(sept, J_HH= 6.8 Hz, 12H, 2,6,2′′,6′′-CHMe₂), 2.32 (br t, 6H,

NCH₂), 1.33 (d, J_HH= 6.6 Hz, 36H, 4,4′′-CH(CH₃)₂), 1.11

[HIPTN₃N]W(¹⁵N₂)

A mixture of W¹⁵N=¹⁵NK (10 mg) and ZnCl₂(0.6 mg,

0.8 equiv.) in THF (0.6 mL) was stirred for 10 min. The

mixture was filtered through Celite^®and the filtrate was

taken to dryness in vacuo. The solid residue was dissolved in

C₆D₆for spectroscopic analysis. IR (C₆D₆): 1888 (ν_14N–14N),

1826 (ν_15N–15N).

[HIPTN₃N]WN=NH

A mixture of WCl (500 mg, 0.277 mmol) and KC₈

(93.6 mg, 0.692 mmol) was stirred in benzene (6 mL) with a

glass-coated magnetic stirbar for 42 h under 1 atm of N₂

(1 atm = 101.325 kPa). The resulting dark maroon-red mix-

ture was treated with solid [Et₃NH][OTf] (69.6 mg,

0.277 mmol), stirred for 30 min, and taken to dryness in

vacuo. The solid residue was extracted with pentane and the

pentane extracts were filtered through Celite^®. Upon concen-

tration of this solution to ~5 mL, the product precipitated as

a cotton-like solid. Repeated warming and cooling of the

mixture to –35 °C induced crystallization of the product in

the form of a microcrystalline solid sufficiently dense to al-

low isolation by filtration. The resulting beige solid was

washed with cold pentane and dried in vacuo at 60 °C.

Yield: 273 mg (0.152 mmol, 55%). IR (C₆D₆): 1542 (ν_N-NH).

¹H NMR (C₆D₆, 20 °C) δ: 8.43 (s, 1H, J(¹⁸³W-H) = 17.4 Hz,

N=NH), 7.30 (br d, 6H, 4′,6′-H), 7.20 (s, 12H, 3,5,3′′,5′′-H),

6.64 (br t, 3H, 2′-H), 3.68 (br t, 6H, NCH₂), 3.14 (sept,

(d, J_HH= 6.8 Hz, 36H, 2,6,2′′,6′′-CH(CH₃)₂), 0.98 (d, J_HH

=

6.8 Hz, 36H, 2,6,2′′,6′′-CH(CH₃)₂). ¹⁹F NMR (C₆D₆, 20 °C)

δ: –62.4 (s, CF₃). Anal. calcd. for C₁₄₆H₁₇₃BF₂₄N₆W (%): C

65.86, H 6.55, N 3.16; found: C 65.74, H 6.46, N 3.08.

X-ray quality crystals were obtained from a supersatu-

rated heptane solution at room temperature.

15 15

′

{[HIPTN₃N]W= N NH₂}[BAr₄]

W¹⁵N=¹⁵NH, generated in situ as described above, was

′

treated with 1 equiv. of [H(OEt₂)₂][BAr₄] in C₆D₆in an

NMR tube to yield the title compound, 92% ¹⁵N-enriched,

1

essentially quantitatively. H NMR (C₆D₆, 20 °C) δ: 5.99 (s,

0.08 H, ¹⁴N_α-¹⁴N_βH₂), 5.99 (dd, ¹J(¹⁵N_β-H) = 83.6 Hz,

²J(¹⁵N_α-H) = 1.8 Hz, 0.92 H, ¹⁵N_α-¹⁵N_βH₂) (other reso-

nances are identical to those measured for the unlabeled

sample). ¹⁵N NMR (C₆D₆, 20 °C) δ: 343.8 (d, J(¹⁵N_α-¹⁵N_β) =

10.2 Hz, 1N, ¹⁵N_α-¹⁵N_βH₂), 131.1 (td, J(¹⁵N_β-H) = 83.8 Hz,

1

J(¹⁵N_α-¹⁵N_β) = 9.9 Hz, 1N, ¹⁵N_α-¹⁵N_βH₂).

J_HH= 6.6 Hz, 12H, 2,6,2′′,6′′-CHMe₂), 2.93 (sept, J_HH

=

6.9 Hz, 6H, 4,4′′-CHMe₂), 1.88 (br t, 6H, NCH₂), 1.37 (d,

J_HH= 6.6 Hz, 36H, 4,4′′-CH(CH₃)₂), 1.24 (d, J_HH= 6.6 Hz,

36H, 2,6,2′′,6′′-CH(CH₃)₂), 1.14 (d, J_HH= 6.6 Hz, 36H,

2,6,2′′,6′′-CH(CH₃)₂). Anal. calcd. for C₁₁₄H₁₆₀N₆W (%): C

76.14, H 8.97, N 4.67; found: C 76.00, H 8.96, N 4.55.

X-ray quality crystals were obtained from a supersatu-

rated heptane solution at room temperature.

[HIPTN₃N]WϵN

A mixture of WCl (500 mg, 0.277 mmol) and NaN₃

(54.0 mg, 0.831 mmol) in a mixture of benzene and

acetonitrile (4:6 mL) was stirred vigorously for 16 h. The

mixture became lighter in color and was taken to dryness in

vacuo. The solid residue was dried at 80 °C in vacuo and ex-

356

Can. J. Chem. Vol. 83, 2005

tracted with benzene. The benzene extracts were filtered

through Celite^®and the filtrates were taken to dryness in

vacuo. Addition of pentane (3 mL) led to the precipitation of

a pale beige solid after 1 h at room temperature. The mixture

was allowed to stand at –35 °C for several days, and the

product was then isolated by filtration, washed with cold

pentane, and dried at 70 °C in vacuo. Yield: 286 mg

(0.160 mmol, 58%). IR (C₆D₆): 1024 (ν_Mo-N). ¹H NMR

(C₆D₆, 20 °C) δ: 8.01 (br s, 6H, 4′,6′-H), 7.21 (s, 12H,

3,5,3′′,5′′-H), 6.59 (s, 3H, 2′-H), 3.66 (br t, 6H, NCH₂), 3.10

(sept, J_HH= 6.6 Hz, 12H, 2,6,2′′,6′′-CHMe₂), 2.93 (sept,

J_HH= 6.8 Hz, 6H, 4,4′′-CHMe₂), 1.66 (br t, 6H, NCH₂),

brown-red mixture was stirred for 30 min, freeze–pump–

thaw degassed, and exposed to 1 atm of H₂(1 atm =

101.325 kPa), which caused some lightening in color. The

mixture was stirred for 1 h and brought to dryness in vacuo.

The resulting solid was dried at 80 °C and extracted with

benzene. The benzene extracts were filtered through Celite^®

and the filtrate was taken to dryness. Addition of pentane

(4 mL) to the residue led to a gradual formation of a light

brown solid. The mixture was left standing at –35 °C for

several days and the product was then isolated by filtration,

washed with cold pentane, and dried in vacuo at 70 °C. As

was explained in the text, WH₃loses hydrogen in the solid

state at elevated temperatures under vacuum, forming an ob-

servable quantity of a paramagnetic compound tentatively

assigned as the monohydride WH. Thus, solid WH₃isolated

as described previously was stored under an atmosphere of

H₂overnight, which nearly eliminated the WH impurity.

Yield: 208 mg (0.117 mmol, 44%). IR (C₆D₆): 1966, 1958

1.37 (d, J_HH= 6.9 Hz, 36H, 4,4′′-CH(CH₃)₂), 1.24 (d, J_HH

=

6.6 Hz, 36H, 2,6,2′′,6′′-CH(CH₃)₂), 1.10 (br d, J_HH= 6.3 Hz,

36H, 2,6,2′′,6′′-CH(CH₃)₂). Anal. calcd. for C₁₁₄H₁₅₉N₅W

(%): C 76.78, H 8.99, N 3.93; found: C 76.57, H 9.01, N

3.94.

1

[HIPTN₃N]Wϵ¹⁵N

(both broad, ν_W-H). H NMR (C₆D₆, 20 °C) δ: 11.99 (s, 3H,

J(¹⁸³W-H) = 24 Hz, W(H)₃), 7.66 (s, 6H, 4′,6′-H), 7.21 (s,

12H, 3,5,3′′,5′′-H), 6.54 (s, 3H, 2′-H), 3.65 (s, 6H, NCH₂),

3.10 (sept, J_HH= 6.8 Hz, 12H, 2,6,2′′,6′′-CHMe₂), 2.93

(sept, J_HH= 6.9 Hz, 6H, 4,4′′-CHMe₂), 2.01 (s, 6H, NCH₂),

15 15

′

[W= N NH₂][BAr₄], generated in situ as described

*

above, was treated with 2 equiv. of CoCp₂in C₆D₆in an

NMR tube to yield a mixture consisting predominantly of

W¹⁵N=¹⁵NH and the title compound. ¹⁵N NMR (C₆D₆,

20 °C) δ: 827.0 (s, Wϵ¹⁵N), 393.5 (dd, W¹⁵N_α=¹⁵N_βH), 229.3

(dd, W¹⁵N_α=¹⁵N_βH).

1.37 (d, J_HH= 6.9 Hz, 36H, 4,4′′-CH(CH₃)₂), 1.24 (d, J_HH

=

6.9 Hz, 36H, 2,6,2′′,6′′-CH(CH₃)₂), 1.12 (d, J_HH= 6.6 Hz,

36H, 2,6,2′′,6′′-CH(CH₃)₂). Anal. calcd. for C₁₁₄H₁₆₂N₄W

(%): C 77.25, H 9.21, N 3.16; found: C 77.08, H 9.28, N

3.06.

′

{[HIPTN₃N]W(NH₃)}[BAr₄]

Ammonia (57 mL, 360 torr (1 torr = 133.322 Pa),

~1.11 mmol) was vacuum transferred from a bronze-colored

Na solution to a frozen mixture of WCl (500 mg,

Acknowledgements

′

0.277 mmol) and NaBAr₄(540.2 mg, 0.610 mmol) in

fluorobenzene (10 mL). The mixture was thawed and stirred

at room temperature for 2 h, at which time ¹H NMR analysis

of an aliquot taken from the reaction mixture showed com-

plete consumption of WCl. The mixture was taken to dry-

ness in vacuo, and the solid residue was dried at 60 °C and

extracted with pentane. The pentane extracts were filtered

through Celite^®and the filtrate was concentrated to 3 mL.

This solution was seeded with a minute quantity of crystal-

R.R.S. is grateful to the National Institutes of Health (GM

31978) for research support and both R.R.S. and D.V.Y. are

grateful to William E. Geiger for advice on electrochemistry

in weakly polar media.

References

1. A.D. Allen and C.V. Senoff. J. Chem. Soc. Chem. Commun.

621 (1965).

′

line {[HIPTN₃N]Mo(NH₃)}[BAr₄], which led to rapid for-

mation of a bright yellow microcrystalline solid. The

mixture was left standing at –35 °C for several days and the

product was then isolated by filtration, washed with cold

pentane, and dried in vacuo at 60 °C, turning light orange.

2. B.K. Burgess and D.J. Lowe. Chem. Rev. 96, 2983 (1996).

3. B.K. Burgess. ACS Symp. Ser. 535, 144 (1993).

4. B.K. Burgess. Chem. Rev. 90, 1377 (1990).

5. J. Chatt, J.R. Dilworth, and R.L. Richards. Chem. Rev. 78, 589

(1978).

6. M. Hidai and Y. Mizobe. Chem. Rev. 95, 1115 (1995).

7. M. Hidai. Coord. Chem. Rev. 185–186, 99 (1999).

8. M.D. Fryzuk and S.A. Johnson. Coord. Chem. Rev. 200–202,

379 (2000).

1

Yield: 642 mg (242 mmol, 87%). H NMR (C₆D₆, 20 °C) δ:

18.66 (s, 6H, 8.27, 4′,6′-H), 7.41 (s, 12H, 3,5,3′′,5′′-H), 7.34

(s, 8H, C₆H₃-3,5-(CF₃)₂-2,6-H), 7.21 (s, 4H, C₆H₃-3,5-

(CF₃)₂-4-H), 3.15 (sept, J_HH= 6.9 Hz, 6H, 4,4′′-CHMe₂),

2.7 (v br s, 12H, 2,6,2′′,6′′-CHMe₂), 1.54 (d, J_HH= 6.6 Hz,

36H, 4,4′′-CH(CH₃)₂), 1.01 (br s, 36H, 2,6,2′′,6′′-CH(CH₃)₂),

0.6 (v br s, 36H, 2,6,2′′,6′′-CH(CH₃)₂), –0.20 (s, 3H, 2′-H),

–28.79 (s, 6H, NCH₂), –63.7 (br s, 6H, NCH₂). Anal. calcd.

9. T.A. Bazhenova and A.E. Shilov. Coord. Chem. Rev. 144, 69

(1995).

10. C.J. Pickett. J. Biol. Inorg. Chem. 1, 601 (1996).

11. F. Barriere. Coord. Chem. Rev. 236, 71 (2003).

12. R.L. Richards. Coord. Chem. Rev. 154, 83 (1996).

13. R.A. Henderson, G.J. Leigh, and C.J. Pickett. Adv. Inorg.

Chem. Radiochem. 27, 197 (1983).

14. J. Chatt and G.J. Leigh. Chem. Soc. Rev. 1, 121 (1972).

15. A.D. Allen, R.O. Harris, B.R. Loescher, J.R. Stevens, and R.N.

Whiteley. Chem. Rev. 73, 11 (1973).

C

₁₄₆H₁₇₄BF₂₄N₅W (%): C 66.16, H 6.62, N 2.64; found: C

65.97, H 6.68, N 2.52.

X-ray quality crystals were obtained from a supersatu-

rated heptane solution at room temperature.

[HIPTN₃N]WH₃

′

A benzene (5 mL) solution of [W(NH₃)][BAr₄] (700 mg,

16. B.A. MacKay and M.D. Fryzuk. Chem. Rev. 104, 385 (2004).

17. K. Komori, H. Oshita, Y. Mizobe, and M. Hidai. J. Am. Chem.

Soc. 111, 1939 (1989).

0.264 mmol) was treated with LiBHEt₃(1 mol/L THF,

264 µL, 0.264 mmol) at room temperature. The resulting

Yandulov and Schrock

357

18. M. Mori, K. Hori, M. Akashi, M. Hori, Y. Sato, and M. Nishida.

Angew. Chem. Int. Ed. 37, 636 (1998).

44. M.G. Fickes, A.L. Odom, and C.C. Cummins. Chem. Commun.

1993 (1997).

19. K. Hori and M. Mori. J. Am. Chem. Soc. 120, 7651 (1998).

20. M. Mori, M. Hori, and Y. Sato. J. Org. Chem. 63, 4832 (1998).

21. M. Akashi and M. Mori. Heterocycles, 59, 661 (2003).

22. K. Ueda and M. Mori. Tetrahedron Lett. 45, 2907 (2004).

23. M. Mori. J. Heterocycl. Chem. 37, 623 (2000).

24. M. Akashi, Y. Sato, and M. Mori. J. Org. Chem. 66, 7873

(2001).

25. M. Akashi, M. Nishida, and M. Mori. Chem. Lett. 465 (1999).

26. M. Mori. J. Organometal. Chem. 689, 4210 (2004).

27. A.E. Shilov. J. Mol. Catal. 41, 221 (1987).

28. A.E. Shilov. Russ. Chem. Bull. Int. Ed. 52, 2555 (2003).

29. R.R. Schrock. Pure Appl. Chem. 69, 2197 (1997).

30. G.E. Greco and R.R. Schrock. Inorg. Chem. 40, 3850 (2001).

31. G.E. Greco and R.R. Schrock. Inorg. Chem. 40, 3860 (2001).

32. D.V. Yandulov and R.R. Schrock. J. Am. Chem. Soc. 124,

6252 (2002).

45. J.C. Peters, A.L. Odom, and C.C. Cummins. Chem. Commun.

1995 (1997).

46. J.B. Greco, J.C. Peters, T.A. Baker, W.M. Davis, C.C. Cummins,

and G. Wu. J. Am. Chem. Soc. 123, 5003 (2001).

47. J.S. Figueroa and C.C. Cummins. Angew. Chem. Int. Ed. 43,

984 (2004).

48. J. Mason. Chem. Rev. 81, 205 (1981).

49. J. Mason. In Multinuclear NMR. Edited by J. Mason. Plenum

Press, New York. 1987.

50. M.G. Hill, W.M. Lamanna, and K.R. Mann. Inorg. Chem. 30,

4687 (1991).

51. N.G. Connelly and W.E. Geiger. Chem. Rev. 96, 877 (1996).

52. R.J. LeSuer and W.E. Geiger. Angew. Chem. Int. Ed. 39, 248

(2000).

53. N. Camire, U.T. Mueller-Westerhoff, and W.E. Geiger. J.

Organomet. Chem. 637–639, 823 (2001).

33. D.V. Yandulov, R.R. Schrock, A.L. Rheingold, C. Ceccarelli,

54. N. Camire, A. Nafady, and W.E. Geiger. J. Am. Chem. Soc.

124, 7260 (2002).

and W.M. Davis. Inorg. Chem. 42, 796 (2003).

34. D. Yandulov and R.R. Schrock. Inorg. Chem. 44, 1103 (2005).

35. D.V. Yandulov and R.R. Schrock. Science, 301, 76 (2003).

36. G.M. Sheldrick. SADABS: A program for absorption correc-

tion of crystallographic data. Version 2.10 [computer pro-

gram]. University of Goettingen, Goettingen, Germany. 1996.

37. V. Ritleng, D.V. Yandulov, W.W. Weare, R.R. Schrock, A.R.

Hock, and W.M. Davis. J. Am. Chem. Soc. 126, 6150 (2004).

38. S. Siemann, K. Schneider, M. Oley, and A. Mueller. Biochem-

istry, 42, 3846 (2003).

39. G.E. Greco, A.I. Popa, and R.R. Schrock. Organometallics, 17,

5591 (1998).

40. S.W. Seidel, R.R. Schrock, and W.M. Davis. Organometallics,

17, 1058 (1998).

55. S. Trupia, A. Nafady, and W.E. Geiger. Inorg. Chem. 42, 5480

(2003).

56. M.B. O’Donoghue, W.M. Davis, and R.R. Schrock. Inorg. Chem.

37, 5149 (1998).

57. D.A. Dobbs, R.R. Schrock, and W.M. Davis. Inorg. Chim.

Acta, 263, 171 (1997).

58. A.B. Pangborn, M.A. Giardello, R.H. Grubbs, R.K. Rosen, and

F.J. Timmers. Organometallics, 15, 1518 (1996).

59. C. Persson and C. Andersson. Inorg. Chim. Acta, 203, 235

(1993).

60. M. Brookhart, B. Grant, and A.F. Volpe, Jr. Organometallics,

11, 3920 (1992).

61. M. Witanowski, L. Stefaniak, S. Szymanski, and H. Januszewski.

41. R.R. Schrock, S.W. Seidel, N.C. Mösch-Zanetti, K.-Y. Shih,

M.B. O’Donoghue, W.M. Davis, and W.M. Reiff. J. Am. Chem.

Soc. 119, 11 876 (1997).

42. N.C. Mösch-Zanetti, R.R. Schrock, W.M. Davis, K.

Wanninger, S.W. Seidel, and M.B. O’Donoghue. J. Am. Chem.

Soc. 119, 11 037 (1997).

J. Magn. Reson. 28, 217 (1977).

62. A.L. Chaney and E.P. Marbach. Clinical Chem. 8, 130 (1962).

63. M.W. Weatherburn. Anal. Chem. 39, 971 (1967).

64. T.T. Ngo, A.P.H. Phan, C.F. Yam, and H.M. Lenhoff. Anal.

Chem. 54 (1982).

65. M.J. Dilworth and K. Fisher. Anal. Biochem. 256, 242 (1998).

43. R.R. Schrock, K.-Y. Shih, D. Dobbs, and W.M. Davis. J. Am.

Chem. Soc. 117, 6609 (1995).

This article has been cited by:

1. Jeremy M. Smith, Deepak Subedi. 2012. The structure and reactivity of iron nitride complexes. Dalton Transactions 41:5,

1423. [CrossRef]

2. Richard R. SchrockCatalytic Reduction of Dinitrogen to Ammonia by Molybdenum 25-50. [CrossRef]

3. Michael R. Reithofer, Richard R. Schrock, Peter Mu#ller. 2010. Synthesis of [(DPPNCH 2 CH 2 ) 3 N] 3- Molybdenum

Complexes (DPP = 3,5-(2,5-Diisopropylpyrrolyl) 2 C 6 H 3 ) and Studies Relevant to Catalytic Reduction of Dinitrogen.

Journal of the American Chemical Society 132:24, 8349-8358. [CrossRef]

4. Richard R. Schrock. 2008. Catalytic Reduction of Dinitrogen to Ammonia by Molybdenum: Theory versus Experiment.

Angewandte Chemie International Edition 47:30, 5512-5522. [CrossRef]

5. Richard R. Schrock. 2008. Die katalytische Reduktion von Distickstoff zu Ammoniak mit Molybdän: Theorie und

Experiment. Angewandte Chemie 120:30, 5594-5605. [CrossRef]

6. Christopher J. PickettMolecular Nitrogen and Dinitrogen Complexes . [CrossRef]

7. Markus Hölscher, Walter Leitner. 2006. DFT Investigation of the Potential of [H-M{(NHCH2CH2)3X}] Catalysts (M = Mo,

Ru, Os; X = N, P) for the Reduction of N2 to NH3 by H2. European Journal of Inorganic Chemistry 2006:21, 4407-4417.

[CrossRef]

8. Alastair J. Nielson, Chaohong Shen, Joyce M. Waters. 2006. Molecular engineering of coordination pockets in chloro-tris-

phenoxo complexes of titanium(IV). Polyhedron 25:10, 2039-2054. [CrossRef]

Article Doi

Synthesis of tungsten complexes that contain hexaisopropylterphenyl- substituted triamidoamine ligands, and reactions relevant to the reduction of dinitrogen to ammonia

DOI: 10.1139/v05-013

Source and publish data:

Authors:

Article abstract of DOI:10.1139/v05-013

Full text of DOI:10.1139/v05-013

Products guided by the article

R&D Labs maybe for 860612-80-2

Relevant to this article

Hot Product