doi.org/10.26434/chemrxiv.7215320.v1

Competitive Ligand Exchange and Dissociation in Ru Indenyl ComplexesRoman Belli, Yang Wu, Hyewon Ji, Anuj Joshi, Lars Yunker, Lisa Rosenberg, J Scott McIndoe

Submitted date: 16/10/2018 • Posted date: 17/10/2018Licence: CC BY-NC-ND 4.0Citation information: Belli, Roman; Wu, Yang; Ji, Hyewon; Joshi, Anuj; Yunker, Lars; Rosenberg, Lisa; et al.(2018): Competitive Ligand Exchange and Dissociation in Ru Indenyl Complexes. ChemRxiv. Preprint.

Kinetic profiles obtained from monitoring the solution phase substitution chemistry of[Ru(η5-indenyl)(NCPh)(PPh3)2]+ (1) by both ESI-MS and 31P{1H} NMR are essentially identical, despite anenormous difference in sample concentrations for these complementary techniques. These studiesdemonstrate dissociative substitution of the NCPh ligand in 1. Competition experiments using differentsecondary phosphine reagents provide a ranking of phosphine donor abilities at this relatively crowdedhalf-sandwich complex: PEt2H > PPh2H >> PCy2H. The impact of steric congestion at Ru is evident also inreactions of 1 with tertiary phosphines; initial substitution products [Ru(η5-indenyl)(PR3)(PPh3)2]+ rapidly losePPh3, enabling competitive recoordination of NCPh. Further solution experiments, relevant to the use of 1 incatalytic hydrophosphination, show that PPh2H out-competes PPh2CH2CH2CO2But (the product ofhydrophosphination of tert-butyl acrylate by PPh2H) for coordination to Ru, even in the presence of a ten-foldexcess of the tertiary phosphine. Additional information on relative phosphine binding strengths was obtainedfrom gas-phase MS/MS experiments, including collision-induced dissociation (CID) experiments on the mixedphosphine complexes [Ru(η5-indenyl)PP’P’’]+, which ultimately appear in solution during the secondaryphosphine competition experiments. Unexpectedly, unsaturated complexes [Ru(η5-indenyl)(PR2H)(PPh3)]+,generated in the gas-phase, undergo preferential loss of PR2H. We propose competing orthometallation ofPPh3 is responsible for the surprising stability of the [Ru(η5-indenyl)(PPh3)]+ fragment under these conditions.

File list (3)

download fileview on ChemRxivbelli-wu_et_al.pdf (879.37 KiB)

download fileview on ChemRxivbelli-wu_et_al.docx (0.92 MiB)

download fileview on ChemRxivbelli-wu_et_al_SI.pdf (1.46 MiB)

Competitive Ligand Exchange and Dissociation in Ru Indenyl Complex-es

Roman G. Belli‡, Yang Wu‡, Hyewon Ji, Anuj Joshi, Lars P. Yunker, Lisa Rosenberg*, J. Scott McIn-

doe*

‡These authors contributed equally.

Department of Chemistry, University of Victoria, P. O. Box 1700, STN CSC, Victoria, British Columbia, Canada V8W 2Y2

Supporting Information Placeholder

ABSTRACT: Kinetic profiles obtained from monitoring the solution phase substitution chemistry of [Ru(η5-

indenyl)(NCPh)(PPh3)2]+ (1) by both ESI-MS and 31P{1H} NMR are essentially identical, despite an enormous difference in sample

concentrations for these complementary techniques. These studies demonstrate dissociative substitution of the NCPh ligand in 1.

Competition experiments using different secondary phosphine reagents provide a ranking of phosphine donor abilities at this rela-

tively crowded half-sandwich complex: PEt2H > PPh2H >> PCy2H. The impact of steric congestion at Ru is evident also in reac-

tions of 1 with tertiary phosphines; initial substitution products [Ru(η5-indenyl)(PR3)(PPh3)2]+ rapidly lose PPh3, enabling competi-

tive recoordination of NCPh. Further solution experiments, relevant to the use of 1 in catalytic hydrophosphination, show that

PPh2H out-competes PPh2CH2CH2CO2But (the product of hydrophosphination of tert-butyl acrylate by PPh2H) for coordination to

Ru, even in the presence of a ten-fold excess of the tertiary phosphine. Additional information on relative phosphine binding

strengths was obtained from gas-phase MS/MS experiments, including collision-induced dissociation (CID) experiments on the

mixed phosphine complexes [Ru(η5-indenyl)PP’P’’]+, which ultimately appear in solution during the secondary phosphine competi-

tion experiments. Unexpectedly, unsaturated complexes [Ru(η5-indenyl)(PR2H)(PPh3)]+, generated in the gas-phase, undergo pref-

erential loss of PR2H. We propose competing orthometallation of PPh3 is responsible for the surprising stability of the [Ru(η5-

indenyl)(PPh3)]+ fragment under these conditions.

INTRODUCTION

We recently reported the activity of a series of Ru(η5-

indenyl) complexes in the catalytic hydrophosphination of

activated alkenes.1 Although participation of this half-

sandwich species in C-C bond-forming catalysis is well-

established,2 its activity for heteroatom addition chemistry

such as hydrophosphination remains relatively under-

explored.3 Our preliminary investigations of this catalysis

show that competing coordination of substrate secondary

phosphine PR2H, ancillary PPh3, and product tertiary phos-

phine PR2R’ at Ru in this system plays an important role in the

speciation of the catalyst during the reaction. For example, 31P{1H} NMR spectra of catalytic mixtures from our prelimi-

nary studies show signals for multiple Ru-P-containing prod-

ucts, attributable to species containing one, two and three dif-

ferent phosphine ligands.1 The possibility of product inhibition

is a concern in harnessing such late metal systems in catalytic

hydrophosphination, but its importance will be highly suscep-

tible to the complex balance of steric and electronic effects

governing reactivity at this relatively crowded Ru center.

The coordination behaviour of phosphine ligands is well-

documented and even quantified,4 but most data concerning

their steric and electronic properties has been gathered for

tertiary phosphines. It is possible to estimate Tolman cone

angles5 or other steric parameters6 for secondary phosphines,

but the impact of the P-H bond on overall donor ability is not

well established. Thus, the synthesis and reactivity of Ru(η5-

indenyl) mixed phosphine complexes are important to our

ongoing studies and optimization of hydrophosphination ca-

talysis, and we have a particular interest in evaluating the be-

havior of secondary phosphines within the Ru coordination

sphere.7

Scheme 1. Activity of cation 1 for hydrophosphination

Among catalyst precursors we assessed in the hydrophos-

phination reactions of diphenylphosphine was [Ru(η5-

indenyl)(NCPh)(PPh3)2]+ (1, Scheme 1). Cations such as 1 and

its phosphine-substituted derivatives are straightforward to

analyze by ESI-MS,8 since they are readily transferred from

solution phase to gas phase during the electrospray process

without any need for adventitious association with some other

charged species.9 Indeed, we previously exploited the elec-

trospray technique to determine Ru-P bond dissociation ener-

gies (BDEs) for [Ru(η5-indenyl)(PPh2H)3]+ using Fourier

transform ion cyclotron resonance mass spectrometry (FTICR-

MS).10 Those quantitative studies involved collision-induced

dissociation (CID) of mass-selected ions within the FTICR

RPPh2H Ph2P R+

Ru

Ph3P

Ru

NCPhPh3P

1R = CN, CO2But

cell, and provided a lower limit for the Ru-PPh2H bond disso-

ciation energy (BDE) of 16.6 kcal/mol. We noted the difficul-

ty in comparing this absolute, gas phase value for an η5-

indenyl system with BDE values for Ru-PR3 available from

solution calorimetry studies of tertiary phosphine substitutions

at {Ru(η5-Cp*)Cl}4, given the dependence of the calorimetry-

derived values on solvent and temperature, and the observed

strong correlation of BDEs for the η5-Cp* system with steric,

as opposed to electronic, properties of the tertiary phosphines

studied.11 Ideally, a clearer picture of the relative binding en-

ergies of phosphines could be obtained by performing these

painstaking CID studies on a broad range of analogous cation-

ic complexes for a given half-sandwich system.

A simple and useful alternative for ranking the affinity of

disparate phosphines at a single, cationic metal fragment is the

study of its substitution reactions in real time using the “pres-

surized sample infusion mass spectrometry” (PSI-ESI-MS)

technique.12 In these experiments a continuous flow of the

stirred reaction mixture in solution is transferred directly into

the spectrometer under a pressure of ≤ 5 psi of inert gas. Pro-

vided the incoming and outgoing ligands differ in mass (e.g. a

competition experiment between PtBu3 and PnBu3 would be

uninformative), time-resolved speciation of all cationic com-

plexes in the mixture will result. Here we report the use of

PSI-ESI-MS to measure the relative kinetics of competitive

phosphine substitution reactions at complex 1, and we confirm

the results using 31P{1H} NMR analysis. These experiments,

complemented by further assessment in the gas phase of the

relative binding strengths of different phosphines in the mixed

phosphine substitution products [Ru(η5-indenyl)PP’P”]+, have

allowed us to rank a group of secondary and tertiary phos-

phines in terms of their affinities for the cationic ruthenium

center in this half-sandwich system.

RESULTS AND DISCUSSION

1. Substitution kinetics from PSI-ESI-MS and 31P{1H} NMR

The benzonitrile (NCPh) ligand in cation 1 is easily dis-

placed by a variety of phosphines. Indeed, the lability of NCPh

in this complex is such that unless very gentle ionization con-

ditions are used in the ESI-MS experiment (cone voltage 10

V),13 a significant proportion of the observed ion intensity

involves the coordinatively unsaturated [Ru(η5-

indenyl)(PPh3)2]+, from which the NCPh has been lost in the

desolvation process. Examination by PSI-ESI-MS of the reac-

tivity of 1 toward a series of phosphines (PPh2H, PEt2H,

PCy2H, PBun3) shows that the rate of substitution is insensitive

to the identity and concentration of the incoming ligand (vide

infra), consistent with a dissociative substitution mechanism

(Scheme 2). This result might seem surprising in the context

of the “indenyl effect”, an observation that indenyl complexes

tend to have accelerated substitution reactions relative to their

Cp analogues. The rate enhancement is commonly attributed

to the coordinative flexibility of the indenyl ligand in swap-

ping between η5- and η3-modes, which can facilitate an associ-

ative substitution mechanism.14 However, explicit examples of

such variable indenyl ligand hapticity do not include rutheni-

um complexes.15 A study on the insertion of alkynes into the

Ru-H bond of RuH(η5-indenyl)(dppm) found evidence for an

associative mechanism, but the authors acknowledge that ei-

ther hapticity change of the indenyl ligand or ring-opening of

the strained bis(diphenylphosphino)methane ligand could al-

low binding of the alkyne substrate.16a Moreover, a kinetic

analysis of the accelerated rate of phosphine substitution at

RuCl(η5-indenyl)(PPh3)2 (the neutral precursor to complex 1)

relative to its η5-Cp analogue found clear evidence for a disso-

ciative mechanism in both cases.16b In that report, the en-

hanced rate of substitution at the η5-indenyl complex (i.e. the

indenyl effect) was attributed to the ability of the electron-rich

indenyl ligand to stabilize the 16-electron intermediate formed

via a dissociative substitution pathway.17

Despite the dissociative nature of these NCPh substitution

reactions, when we monitor by PSI-ESI-MS the addition of a

1:1 mixture of PEt2H and PPh2H (10 equivalents each) to 1,

the time-dependent distribution of cationic species reveals a

dependence of the product ratio on the nature of the incoming

ligand (Figure 1a). The relative amounts of the two products

formed, [Ru(η5-indenyl)(PPh3)2(PPh2H)]+ (2a) and [Ru(η5-

indenyl)(PPh3)2(PEt2H)]+ (2b) tell us about the relative rates of

the rapid second step in the dissociative substitution (Scheme

2): PEt2H reacts slightly faster than PPh2H with the unsaturat-

ed, 16e– intermediate [Ru(η5-indenyl)(PPh3)2]+, giving a higher

proportion of product 2b (Scheme 2). This is consistent with

the fact that PEt2H is smaller and more electron-rich than

PPh2H; it is a stronger donor ligand.

Figure 1. (a) Competitive reactions of 1 with a 10:10 mixture of

PPh2H:PEt2H at 45°C in PhF, as monitored by PSI-ESI-MS. Cir-

cles are normalized experimental data; lines are simulated using

parameter estimation with COPASI. (b) 145.85 MHz 31P{1H}

NMR data for the same experiment in 2:1 CH2Cl2/C6D6 at RT.

Scheme 2. Competitive substitution of the nitrile ligand in

1 with a mixture of PPh2H and PEt2H

The kinetic profile we observe for this competition experi-

ment by ESI-PSI-MS was confirmed by an analogous experi-

ment monitored by 31P{1H} NMR (Figure 1b). These two ex-

periments were carried out under strikingly different condi-

tions. Most obvious is the much higher sample concentration

required for NMR, relative to MS. However, the effective

sampling frequency also differs for the two techniques; the

ESI-MS experiment produces one spectrum per second, while

the relatively long delays used to render the 31P{1H} NMR

experiment approximately quantitative gave us one spectrum

(128 scans) per 20 minutes. To address this difference, we

conducted the NMR experiment at room temperature instead

of 45°C, to slow it down. Based on the rule of thumb that reac-

tion rates double with each increase in temperature of 10°C,18

the NMR experiment should take about 5× as long as the PSI-

MS experiment, for this 24°C difference. The consumption of

1 to give 2a and 2b as monitored by NMR actually took about

18× as long (180 min as opposed to 10 min at 45°C when

monitored by ESI-MS); however further slowing of NMR tube

reactions can occur due to poor mixing.19 Nevertheless, the

two kinetic profiles shown in Figure 1 are essentially identical.

Given the factor of three million between concentrations used

for the MS and NMR experiments, this is compelling evidence

for the cleanly dissociative nature of the substitution reaction

and for the coordination behaviors of these two secondary

phosphines.

To tease out electronic vs. steric effects we also compared

the reactions of PPh2H and PCy2H with complex 1, since

PCy2H has similar electronic properties to PEt2H but is signif-

icantly bulkier (its cone angle of 148° is actually slightly larg-

er than that of PPh3 = 145°).4d The MS and NMR results in

Figure 2 show clearly the importance of steric hindrance of the

incoming phosphine; [Ru(η5-indenyl)(PPh3)2(PCy2H)]+ (2c)

forms much more slowly than the PPh2H complex 2a.

Figure 2. (a) Competitive reactions of 1 with a 10:10 mixture of

PPh2H and PCy2H at 45°C in PhF, as monitored by PSI-ESI-MS.

Circles are normalized experimental data; the lines were simulat-

ed using parameter estimation in COPASI. (b) 145.85 MHz 31P{1H} NMR data for the same experiment in 2:1 CH2Cl2/C6D6

at RT.

Based on the results shown in Figures 1 and 2, the relative

reactivity of these three secondary phosphines in the substitu-

tion of NCPh in 1 is PEt2H > PPh2H >> PCy2H. We used

COPASI20 to model the traces provided by ESI-MS, to extract

kinetic parameters for the competitive substitution reactions

(solid lines in Figures 1a and 2a).21 Along with the time-

dependent speciation data, COPASI is given the starting con-

centrations and the predicted sequence of elementary steps

(e.g. Scheme 2). COPASI then fits the data and provides k1

(rate constant for the nitrile dissociation, and for the overall

reaction); it also simulates accurately the ratios of the two k2

values, i.e. how much faster one of these fast, subsequent

phosphine association steps is than the other. The modeling is

consistent with rates for these association reactions that are

very fast regardless of the incoming phosphine. Since the sim-

ulation involved a relatively large number of independent pa-

rameters, we tested the rate constants resulting from one set of

initial concentrations (e.g. 10 equivalents of the 1:1 phosphine

mixture, Figures 1 and 2) against a different set of initial con-

centrations (e.g. 100 equivalents of the phosphines, Figures

S3, S5). Under these conditions, the simulation changed in the

same way as the experimental data. These results not only

validate the modeling but also provide further support for the

dissociative nature of the reaction, since the overall rate of

reaction remains the same with the concentration change (fall-

ing in the range k1 = 0.006 ± 0.001 for all experiments).

slow

Ph3P

Ru

NCPhPh3P

1NCPh Ph3P

RuPh3P

PPh2H PEt2H

fast faster

Ph3P

Ru

PEt2HPh3P

2bPh3P

Ru

PPh2HPh3P

2a

The trialkylphosphine PBun3 showed more complex reac-

tivity with cation 1 than did the secondary phosphines (Figure

3). Since there were multiple products, we did not pursue

competition experiments and instead examined just the simple

substitution of NCPh in 1 with PBun3. The rate of disappear-

ance of 1 is the same as in all previous experiments, but the

subsequent reactivity is more complex. As shown in Scheme

3, the first species to occupy the vacant coordination site in the

unsaturated intermediate [Ru(η5-indenyl)(PPh3)2]+ is PBun

3,

which is present in large excess. However, the resulting com-

plex [Ru(η5-indenyl)(PPh3)2(PBun3)]

+ (2d) forms only transi-

ently (Figure 3b); it never reaches more than 10% of the initial

abundance of 1. Complex 2d decomposes to two new com-

plexes, [Ru(η5-indenyl)(PPh3)(PBun3)2]

+ (3d) and

[Ru(η5-indenyl)(PPh3)(PBun3)(NCPh)]+ (4d), which corre-

spond to the substitution of one PPh3 ligand in 2d by PBun3

and NCPh, respectively (Scheme 3). The ratio of 3d and 4d

formed depends on the amount of PBu3 added: smaller excess-

es of PBun3 produce less of the former (Figures S8, S9).

Figure 3. (a) Reaction of 1 with 100 equivalents of PBun3 at 45°C

in PhF, as monitored by PSI-ESI-MS. Circles are normalized

experimental data and solid lines were simulated using COPASI.

The different intensity scale in (b) highlights the low concentra-

tions of transiently formed 2d. In MS experiments using 50 and

10 equivalents PBun3, signal due to this complex falls much lower

in the baseline, and 2d was not detected in the corresponding 31P{1H} NMR experiment (10 equivalents PBun

3, Figures S6, S7).

Scheme 3. Reaction sequence used to model MS data for

addition of excess PBun3 to 1 from Figure 3

The results shown in Figure 3 indicate that the PPh3 ligands

in 2d are labilized by the initial replacement of NCPh in 1 by

PBun3, presumably through steric crowding; one PPh3 is then

lost quickly (relative to the initial loss of NCPh) to form the

unsaturated complex [Ru(η5-indenyl)(PPh3)(PBun3)]

+. The

solution now contains three possible candidates to occupy that

vacant coordination site: one equivalent of PPh3, a large ex-

cess of PBun3, and one equivalent of NCPh. While the PPh3 is

clearly uncompetitive (low relative concentration, cone angle

145°), a high relative concentration of the somewhat less

bulky PBun3 (cone angle 132°) allows the formation of 3d. The

apparent non-innocence of the nitrile ligand under these condi-

tions is somewhat surprising; despite its much lower relative

concentration (NCPh:PBun3 is 1:10, 1:50, or 1:100, vide infra)

and its facile loss from 1, the much smaller NCPh can reoccu-

py the vacant site to give 4d, at a rate that is comparable to

that for the formation of 3d. The numerical modelling agrees

with this sequence of events. Not only did the parameter esti-

mation provide reasonable matches for the experimental data

obtained using 100 equivalents of PBu3 (solid lines in Figure

3), but also the model responded in the same way as the exper-

iment when the number of equivalents was dropped to 50 and

then 10 (Figures S8, S9).

As described in the introduction, these studies of secondary

and tertiary phosphine substitution chemistry are relevant to

the potential activity of this half sandwich system in the cata-

lytic hydrophosphination of alkenes.1 The ability of secondary

phosphines to compete effectively with the product tertiary

phosphine for binding at Ru will be critical to this activity, and

the above results suggest that this should not be a problem,

based primarily on steric arguments. Further support for this

premise comes from additional experiments examining the

reactions of 1 with PPh2H and the product of hydrophosphina-

tion of tert-butyl acrylate by this secondary phosphine:

PPh2(CH2CH2CO2But). The reaction of 1 with just product

phosphine gives a relatively similar product distribution to that

observed for PBun3 (vide supra, Figures S10, S11); monitoring

by 31P{1H} shows the nitrile substitution product 2e forms

transiently but the mixed phosphine nitrile complex

PBun3 NCPh

fast fast

Ph3P

Ru

PBun3

Ph3P

2d

Bun3P

Ru

PBun3

Ph3P

3d

fastPBun3

slow

Ph3P

Ru

NCPhPh3P

1NCPh Ph3P

Ru

Ph3P

slow

PPh3

Ru

PBun3

Ph3P

PhCN

Ru

PBun3

Ph3P

4d

[Ru(η5-indenyl)(PPh3)(PPh2CH2CH2CO2But)(NCPh)]+, 4e,

ultimately dominates the product mixture. Monitoring this

reaction by PSI-ESI-MS shows the same final product distri-

bution, but instead of complex 2e, the apparently unsaturated

complex [Ru(η5-indenyl)(PPh3)(PPh2CH2CH2CO2But)]+ is

observed transiently. (We attribute formation of this species to

facile dissociation of PPh3 from sterically congested 2e during

the electrospray process; it may be stabilized in the gas-phase

by chelation through the pendant ester group in the tertiary

phosphine.) This substitution chemistry may become im-

portant towards the end of catalysis when the ratio of product

phosphine to substrate PPh2H is very high, but in the presence

of a 10:10 mixture of PPh2H and PPh2CH2CH2CO2But only the

PPh2H complex 2a is observed (Figures S13, S14). Even in a

reaction of 1 with a 1:10 mixture of PPh2H and

PPh2CH2CH2CO2But, complex 2a dominates and only trace

amounts of PPh2CH2CH2CO2But-containing products are ob-

served, by ESI-MS (e.g. 4e and

[Ru(η5-indenyl)(PPh3)(PPh2H)(PPh2CH2CH2CO2But)]+, 6)

(Scheme 4, Figure S15). These results indicate that the tertiary

product phosphine does not effectively compete with PPh2H

for coordination at this cationic Ru center.

Scheme 4. Competitive substitution of the nitrile ligand in

1 with a 1:10 mixture of PPh2H and PPh2CH2CH2CO2But;

only 2a was observed by NMR; 2a, 4e and 6 were observed

by MS.

2. MS/MS experiments: an alternative gauge of relative

phosphine binding affinities

When the competitive substitution reactions involving addi-

tion of two secondary phosphines to complex 1 are allowed to

reach equilibrium (e.g. t >10 min for the PSI-ESI-MS experi-

ments, after complex 1 has been completely consumed in the

initial substitution reactions), the product mixtures include

complexes that result from further substitution of the PPh3

ligands in the mixed phosphine complexes 2a-c (e.g. Figures

S2, S4). This is a straightforward way to generate complexes

of the type [Ru(η5-indenyl)(PP’P”)]+, which provides an op-

portunity to directly compare the relative binding strength of

the phosphine ligands through MS/MS experiments. We can

isolate in the gas phase a single [Ru(η5-indenyl)(PP’P”)]+

complex from a complicated mixture of other complexes of

this type (e.g. [Ru(η5-indenyl)PxP’yP”z]+ where x, y, z = 0-3

and x+y+z = 3), and accelerate it through an argon-filled colli-

sion cell to initiate energetic collisions and raise the internal

energy of the ion to the point that unimolecular decomposition

reactions occur, eliminating neutral molecules (most often

intact L-type ligands in the first instance) and new product

ions. The relative propensity of the complex to lose one or

another ligand during this collision-induced dissociation (CID)

is gauged by the lowest collision voltage at which the cationic

dissociation product is detected and by the relative amount of

dissociation product detected as the collision voltage is in-

creased (collectively referred to below as a “track”). Dissocia-

tion product tracks tell us about the relative binding strength of

the ligands in the complex of interest.22 This presents a con-

venient alternative to actually quantifying ligand binding

strength (as described in the introduction), which requires

energy-resolved threshold CID techniques that often necessi-

tate instrumentation that is not commercially available,23 and

relies on clean fragmentation to a single product.24

We selected the ion [Ru(η5-

indenyl)(PPh3)(PEt2H)(PPh2H)]+, 7, from the competitive lig-

and substitution mixture that initially produced complexes 2a-

b (vide supra), and carried out CID over a range of collision

voltages. Figure 4 shows the relative intensities (abundances)

of 7 and the resulting product ions as a function of collision

voltage; these are referred to as “breakdown curves”.25

Figure 4. Breakdown curves for the precursor ion [Ru(η5-

indenyl)(PPh3)(PEt2H)(PPh2H)]+ (7).

There are three possible simple ligand dissociations that can

occur for complex 7, and we observe all three to varying de-

grees starting at the minimum voltage of ~0.6 V. The least

abundant is the ion generated via loss of PEt2H, suggesting

this ligand is the most tenaciously bound. Ions resulting from

loss of PPh2H and PPh3 appear with almost identical tracks;

overall the three tracks suggest an order of binding strength

PEt2H > PPh2H ~ PPh3. As discussed further below, these

product ions themselves are susceptible to further fragmenta-

tion via loss of a second phosphine ligand, which is why their

relative intensities drop to zero at collision voltages higher

than 2 V. Finally, at the highest collision voltages, loss of the

slow

Ph3P

Ru

NCPhPh3P

1NCPh Ph3P

RuPh3P

PPh2H

PPh2CH2CH2CO2But

fastest

Ph3P

Ru

PPh2HPh3P

2a

Ph3P

Ru

PPh3P

2e

= P

not observed

PPh3

Ru

PPh3P

NCPh PPh2H

PhCN

Ru

PPh3P

4eHPh2P

Ru

PPh3P

6

major

trace trace

final phosphine from all three monophosphine species occurs,

giving the common, highly unsaturated product [Ru(η5-

indenyl)]+.

When the analogous CID experiment was conducted with

the ion [Ru(η5-indenyl)(PPh3)(PCy2H)(PPh2H)]+ (8) (Figure

S16), PPh3 was the ligand most easily lost, with slightly lower

and almost identical tracks for loss of PCy2H and PPh2H. This

experiment highlights the relative importance of the steric and

electronic characters of the phosphines in this half-sandwich

system; PPh3 and PCy2H have almost identical cone angles,

but PCy2H should be more Lewis basic, so in this very steri-

cally crowded complex PPh3 is the most weakly bound. Col-

lectively, the CID experiments for 7 and 8 show that the great-

er steric pressure exerted by PCy2H relative to PEt2H has the

effect of discriminating between PPh3 and PPh2H, as well as

making PCy2H significantly easier to dissociate than PEt2H in

a comparable coordination environment. Overall, a binding

strength order between these four ligands emerges as PEt2H >

PPh2H ~ PCy2H ~ PPh3, where the order between the last three

ligands depends on the steric environment at the metal com-

plex.

Figure 5. Breakdown curves observed in MS/MS experiments for (a) [Ru(η5-indenyl)(PPh3)2(PEt2H)]+ (2b) and (b) the in-source generat-

ed ion [Ru(η5-indenyl)(PPh3)(PEt2H)]+. In both experiments, [Ru(η5-indenyl)(PPh3)(PEt2H)]+ converts cleanly into [Ru(η5-

indenyl)(PPh3)]+ without appreciable formation of [Ru(η5-indenyl) (PEt2H)]+.

3. MS/MS experiments: unexpected evidence for gas-phase

orthometallation

We also used CID experiments to examine the gas phase,

competitive dissociation chemistry of [Ru(η5-

indenyl)(PPh3)2(PR2H)]+ for R = Ph (2a) and Et (2b), since

the reliably facile loss of PPh3 from these complexes to gen-

erate coordinatively unsaturated fragments [Ru(η5-

indenyl)(PPh3)(PR2H)]+ should allow a simpler, two-way

comparison of the binding energies of the remaining phos-

phines. We were surprised by the results of this experiment

for the PEt2H complex 2b (Figure 5a); although we did ob-

serve almost exclusive initial loss of PPh3 to form [Ru(η5-

indenyl)(PPh3)(PEt2H)]+,26 continued ramping of the colli-

sion voltage showed subsequent, exclusive loss of PEt2H to

give [Ru(η5-indenyl)(PPh3)]+, instead of the loss of a second

equivalent of PPh3 that we expected based on the stronger

donor ability of PEt2H (vide supra). This points to unusual

stability of the gaseous [Ru(η5-indenyl)(PPh3)]+ fragment

under the conditions of the CID experiment. We reproduced

this result using an alternative experiment in which the ion

[Ru(η5-indenyl)(PPh3)(PEt2H)]+ was generated “in-source”,

by increasing the cone voltage such that the initial CID oc-

curs during the desolvation process (Figure 5b). This al-

lowed examination of the unsaturated precursor ion by

MS/MS independently of the presence of potentially compli-

cating fragments.

One possible explanation for the surprisingly facile loss of

PEt2H and apparent stability of the [Ru(η5-indenyl)(PPh3)]+

fragment (m/z 479) in these experiments involves gas-phase

reactivity of the coordinatively unsaturated intermediate

[Ru(η5-indenyl)(PPh3)(PEt2H)]+, which is “hot” (vibrational-

ly excited) from collisions (Scheme 5). Although the pres-

sure in the collision cell is too low to allow intermolecular

ion-molecule reactivity, intramolecular reactions could oc-

cur. In particular, we propose that this 16-electron species

undergoes orthometallation of the coordinated PPh3, a well-

known intramolecular C-H activation reaction that occurs

readily for Ru(II) complexes.27 This would generate a

Ru(IV) species (complex A, Scheme 5), from which dissoci-

ation of the neutral secondary phosphine ligand PEt2H would

now be favoured, relative to loss of the new hydride and 2-

(o-C6H4PPh2) ligands (fragment also m/z 479).

Scheme 5. Proposed gas-phase orthometallation chemis-

try resulting in preferential loss of PEt2H over PPh3 from

[Ru(η5-indenyl)(PPh3)(PEt2H)]+

Similar results were obtained for CID experiments involv-

ing the PPh2H complex 2a (Figure S17), although in this

case PPh2H was more competitive with PPh3, both for initial

loss (about 2% PPh2H is lost from 2a to give the

bis(triphenylphosphine) cation, compared to 0.1 % PEt2H

dissociation from 2b) and for the second ligand dissociation

(about 5% of the PPh2H remains bound to Ru, compared to

0% for PEt2H). These observations are reasonable: PPh3 is

lost more easily than PPh2H from 2a because it is signifi-

cantly larger than PPh2H and there are two PPh3 ligands;

also, as shown above, the PPh2H is a slightly stronger donor.

In addition, PPh2H is able to participate in orthometallation,

unlike PEt2H but like PPh3. However, its propensity to do so

in this context is less for two reasons: PPh2H has just four

ortho protons while PPh3 has six available to participate in

the C-H activation at Ru, and significantly, the PPh3 ortho

hydrogens are closer to the metal due to the larger PPh3 cone

angle, relative to that for PPh2H. Figure S17 shows that at

very high collision energies, [Ru(η5-indenyl)(PPh2H)]+ be-

gins to form as a product of decomposition of [Ru(η5-

indenyl)(PPh3)]+, which corresponds to a loss of benzyne

that can be attributed to a second orthometallation event.

CONCLUSIONS

We have demonstrated the facile evaluation of ligand sub-

stitution chemistry and relative binding strengths at cationic

metal complexes, through straightforward solution and gas-

phase experiments involving PSI-ESI-MS. The kinetics of

competitive ligand substitution reactions at a ruthenium in-

denyl cation as monitored by PSI-ESI-MS are reproduced

comprehensively by those obtained for analogous experi-

ments monitored by 31P{1H} NMR, despite a difference in

solution concentrations of more than six orders of magni-

tude. These solution experiments examining mixtures of

secondary and tertiary phosphines highlight especially the

impact of steric crowding at the ruthenium indenyl fragment,

which has implications for its activity in catalytic hydro-

phosphination chemistry; substrate secondary phosphines

should compete well for binding in the presence of (at least)

equimolar amounts of product tertiary phosphine. The ease

of isolating gaseous cations containing two or three distinct

phosphine ligands in the mass spectrometer allowed us to

further parse relative Ru-P binding strengths in this system

via competitive CID experiments. These gas-phase studies

reinforce some of our conclusions from the solution studies

concerning the balance of electronic and steric influences of

secondary and tertiary phosphines in ruthenium coordination

chemistry. However rigorous one-on-one comparison of

phosphine binding energies is precluded for this half-

sandwich system by apparent competing C-H activation

chemistry in the collision cell. Although such phosphine

orthometallation is routinely observed in solution chemistry,

its importance in gas-phase organometallic chemistry was

not previously established.

EXPERIMENTAL SECTION

General details and instrumentation

Chemicals and reaction mixtures were all handled under inert gas

atmosphere using standard glovebox and Schlenk techniques. Di-

chloromethane was freshly distilled from CaH or P2O5 and fluoro-

benzene was freshly distilled from P2O5 before use. Deuterated

benzene (C6D6) was stored over sodium/benzophenone and was

degassed by three freeze−pump−thaw cycles and vacuum-

transferred before use. All phosphines were purchased from Strem

Chemicals Inc. and used without further purification. The ruthenium

complex [Ru(η5-indenyl)(NCPh)(PPh3)2]+[B(C6F5)4]–, 1[B(C6F5)4],

was synthesized by a literature method.1

Mass spectra were collected with a Micromass Q-Tof Micro

mass spectrometer in positive ion mode. Key parameters: capillary

voltage, 2800 V; sample cone voltage, 10.0 V; extraction cone volt-

age, 1.0 V; desolvation temperature, 160 °C; source temperature, 60

°C; cone gas flow, 0 L/h; desolvation gas flow, 40 L/h; MCP volt-

age, 2700 V; collision voltage (MS study), 2 V; collision voltage

(MS/MS study), programmed by Autohotkey (MassLynx) to in-

crease the voltage by 1 V every 10.12 seconds from 0 to 100 V.

Mass spectrometric data was normalized to total ion current before

conducting kinetic analysis (this normalization step accounts for

differences in spray conditions from one spectrum to the next).

NMR spectra were recorded at ambient temperature (294K) on a

Bruker AMX 360 spectrometer operating at 145.85 MHz for 31P.

Chemical shifts are reported in ppm. 31P{1H} chemical shifts are

reported relative to 85% H3PO4(aq).

Ligand substitution reactions monitored using pressurized sam-

ple infusion (PSI)

An argon-pressurized (3 psi) Schlenk flask containing a solution

of 1[B(C6F5)4] (0.2 mg, 0.1 μmol) in fluorobenzene (10.5 ml) heat-

ed to 45 °C (or 60 °C for experiments with Ph2PCH2CH2CO2But)

was connected to the mass spectrometer via PEEK tubing, and a

stable signal established for the cation 1. The reaction was initiated

by the injection of a fluorobenzene solution (0.5 ml) of 10, 50 or

100 equiv of phosphine ligand(s). Amounts used: PPh2H (100

equiv, 13 µmol, 36 µL; 10 equiv, 1.3 µmol, 3.6 µL); PEt2H (100

equiv, 13 µmol, 18 µL ; 10 equiv, 1.3 µmol, 1.8 µL); PCy2H (100

equiv, 13 µmol, 39 µL; 10 equiv, 1.3 µmol, 3.9 µL); PBun3 (100

equiv, 13 µmol, 3.3 µL ; 50 equiv, 6.5 µmol, 1.6 µL; 10 equiv, 1.3

µmol, 0.33 µL); Ph2PCH2CH2CO2But (10 equiv, 1.3 µmol, 7 mg).

Ligand substitution reactions monitored using 31P{1H} NMR

Complex 1[B(C6F5)4] (30 mg, 20 µmol) was dissolved in a mix-

ture of dichloromethane (0.4 mL) and C6D6 (0.2 mL) and 1,10, or

20 equiv of various phosphine ligands were added. Amounts used:

PPh2H (10 equiv, 0.20 mmol, 35 μL; 1 equiv, 20 μmol, 3.5 μL),

PEt2H (10 equiv, 0.20 mmol, 23 μL), PCy2H (10 equiv, 0.20 mmol,

44 μL), PBun3 (10 equiv, 0.20 mmol, 50 μL), Ph2PCH2CH2CO2But

(10 equiv, 0.20 mmol, 62 mg). The progress of the reactions was

monitored by 31P{1H} NMR using a relaxation delay (D1) of 10 s.

MS/MS studies examining competitive ligand dissociation from

[Ru(η5-indenyl)PP’P”]+ and [Ru(η5-indenyl)PP’]+

The [Ru(η5-indenyl)PP’P”]+ complexes 5 and 6 were prepared in

situ by adding 10 equiv each of two secondary phosphine ligands to

a solution of 1[B(C6F5)4] (2 mg, 1 µmol) in dichloromethane (5

mL). The [Ru(η5-indenyl)PP’]+ complexes were generated during

the MS/MS experiment from precursor complexes [Ru(η5-

indenyl)(PPh3)2(PR2H)]+ (R = Ph (2a), Et (2b)), which were pre-

pared in situ by the addition of 10 equiv of one secondary phosphine

ligand to a solution of 1[B(C6F5)4] (2 mg, 1 µmol) in dichloro-

methane (5 mL). Amounts used: diphenylphosphine (36 µL, 13

µmol); diethylphosphine (18 µL, 13 µmol); dicyclohexylphosphine

(39 µL, 13 µmol).

MS/MS data for these tris(phosphine) complexes were collected

by isolating in the collision cell the appropriate m/z value for the

precursor ion of interest, and increasing the collision voltage 1 V

every 10 seconds while observing the resulting product ions.

Additional, analogous MS/MS experiments were performed on

coordinatively unsaturated [Ru(η5-indenyl)PP’]+ complexes that

were generated from the solutions containing 2a-b by increasing the

sample cone voltage to 30 V, to dissociate the most weakly bound

phosphine ligand during the in-source desolvation process.

ASSOCIATED CONTENT

Supporting Information

Additional reaction monitoring data acquired using ESI-MS and 31P{1H} NMR, and additional MS-MS data (pdf).

The Supporting Information is available free of charge on the

ACS Publications website.

AUTHOR INFORMATION

Corresponding Authors

* Email: lisarose@uvic.ca, mcindoe@uvic.ca Notes The authors declare no competing financial interests.

ACKNOWLEDGMENT

JSM thanks NSERC (Discovery and Discovery Accelerator

Supplements) for operational funding and CFI, BCKDF and the

University of Victoria for infrastructural support. LR thanks

NSERC (Discovery) for funding. RGB thanks University of

Victoria (graduate fellowship) and NSERC (CGS-M and PGS-

D) for funding.

REFERENCES

(1) Belli, R. G.; Burton, K. M. E.; Rufh, S. A.; McDonald, R.;

Rosenberg, L. Inner- and outer-sphere roles of ruthenium phosphido

complexes in the hydrophosphination of alkenes. Organometallics

2015, 34, 5637-5646.

(2) Leading references for the use of the Ru(η5-indenyl) fragment

in catalysis: (a) Trost, B. M.; Ryan, M. C.; Maurer, D. Development

of a coordinatively unsaturated chiral indenylruthenium catalyst.

Org. Lett. 2016, 18, 3166-3169; (b) Trost, B. M.; Ryan, M. C.

Indenylmetal catalysis in organic synthesis. Angew. Chem. Int. Ed.

2016, 56, 2862-2879; (c) Manzini, S.; Fernández-Salas, J. A.;

Nolan, S. P. From a decomposition product to an efficient and

versatile catalyst: the [Ru(η5-indenyl)(PPh3)2Cl] story. Acc. Chem.

Res. 2014, 47, 3089-3101; (d) Thamapipol, S.; Kündig, E. P.

Intramolecular Diels–Alder reactions using chiral ruthenium Lewis

acids and application in the total synthesis of ent-ledol. Org.

Biomol. Chem. 2011, 9, 7564-7570; (e) Fung, W. K.; Huang, X.;

Man; ManNg, S. M.; Hung, M. Y.; Lin, Z.; Lau, C. P. Dihydrogen-

bond-promoted catalysis:  catalytic hydration of nitriles with the

indenylruthenium hydride complex (η5-C9H7)Ru(dppm)H (dppm =

bis(diphenylphosphino)methane). J. Am. Chem. Soc. 2003, 125,

11539-11544; (f) Alvarez, P.; Gimeno, J.; Lastra, E.; García-

Granda, S.; Van der Maelen, J. F.; Bassetti, M. Synthesis and

reactivity of indenyl ruthenium(II) complexes containing the labile

ligand 1,5-cyclooctadiene (COD):  catalytic activity of [Ru(η5-

C9H7)Cl(COD)]. Organometallics 2001, 20, 3762-3771; (g) Koh, J.

H.; Jung, H. M.; Kim, M.-J.; Park, J. Enzymatic resolution of

secondary alcohols coupled with ruthenium-catalyzed racemization

without hydrogen mediator. Tetrahedron Lett. 1999, 40, 6281-6284;

(h) Yamamoto, Y.; Kitahara, H.; Hattori, R.; Itoh, K. Ruthenium-

catalyzed tandem [2 + 2 + 2]/[4 + 2] cycloaddition of 1,6-

heptadiyne with norbornene. Organometallics 1998, 17, 1910-1912.

(3) The activity of a RuCp* complex in similar hydrophosphina-

tion chemistry has been reported. Sues, P. E.; Lough, A. J.; Morris,

R. H. Reactivity of ruthenium phosphido species generated through

the deprotonation of a tripodal phosphine ligand and implications

for hydrophosphination. J. Am. Chem. Soc. 2014, 136, 4746-4760.

(4) (a) Santiago, C. B.; Guo, J.-Y.; Sigman, M. S. Predictive and

mechanistic multivariate linear regression models for reaction

development. Chem. Sci. 2018, 9, 2398-2412; (b) Fey, N.; Orpen,

A. G.; Harvey, J. N. Building ligand knowledge bases for

organometallic chemistry: Computational description of

phosphorus(III)-donor ligands and the metal–phosphorus bond.

Coord. Chem. Rev. 2009, 253, 704-722; (c) Leyssens, T.; Peeters,

D.; Orpen, A. G.; Harvey, J. N. How important is metal−ligand

back-bonding toward YX3 ligands (Y = N, P, C, Si)? An NBO

analysis. Organometallics 2007, 26, 2637-2645; (d) Tolman, C. A.

Steric effects of phosphorus ligands in organometallic chemistry

and homogeneous catalysis. Chem. Rev. 1977, 77, 313-348.

(5) See reference 4d. These calculations may not accurately re-

flect the much tighter C-P-H angles (95-97°) in secondary phos-

phines relative to the C-P-C angles in tertiary phosphines (99 -

110°). Quin, L. D., A Guide to Organophosphorus Chemistry.

Wiley:Hoboken: NJ, 2000.

(6) Percent buried volume (%Vbur) is an important ligand steric

parameter that can complement the Tolman cone angle in describing

the steric influence of phosphines on catalytic reactions. Wu, K.;

Doyle, A. G. Parameterization of phosphine ligands demonstrates

enhancement of nickel catalysis via remote steric effects. Nature

Chem. 2017, 9, 779; and references therein.

(7) For recent review of the coordination chemistry of secondary

phosphines see: Nell, B. P.; Tyler, D. R. Synthesis, reactivity, and

coordination chemistry of secondary phosphines. Coord. Chem.

Rev. 2014, 279, 23-42.

(8) Chisholm, D. M.; Scott McIndoe, J. Charged ligands for

catalyst immobilisation and analysis. Dalton Trans. 2008, 3933-

3945.

(9) In ESI-MS, neutral molecules often appear as [M+H]+ ions,

for example, and their intensity in spectra is thus strongly dependent

on a combination of basicity and surface activity.

(10) Westmore, J. B.; Rosenberg, L.; Hooper, T. S.; Willett, G.

D.; Fisher, K. J. Determination of ruthenium−phosphorus bond

dissociation energies by ES-FTICR mass spectrometry.

Organometallics 2002, 21, 5688-5691.

(11) Luo, L.; Nolan, S. P. Relative binding energies of sterically

demanding tertiary phosphine ligands to the Cp*RuCl (Cp* = 5-

C5Me5) moiety. Thermochemical investigation of coordinatively

unsaturated organoruthenium complexes. Organometallics 1994,

13, 4781-4786.

(12) (a) Vikse, K. L.; Ahmadi, Z.; Luo, J.; van der Wal, N.; Daze,

K.; Taylor, N.; McIndoe, J. S. Pressurized sample infusion: An

easily calibrated, low volume pumping system for ESI-MS analysis

of reactions. Int. J. Mass Spectrom. 2012, 323-324, 8-13; (b) Vikse,

K. L.; Woods, M. P.; McIndoe, J. S. Pressurized sample infusion for

the continuous analysis of air- and moisture-sensitive reactions

using electrospray ionization mass spectrometry. Organometallics

2010, 29, 6615-6618.

(13) Chisholm, D. M.; Oliver, A. G.; McIndoe, J. S. Mono-

alkylated bisphosphines as dopants for ESI-MS analysis of catalytic

reactions. Dalton Trans. 2010, 39, 364-373.

(14) (a) Hartwig, J. F., Organotransition Metal Chemistry: From

Bonding to Catalysis. University Science Books: 2010, p.250; (b)

Crabtree, R. H., The Organometallic Chemistry of the Transition

Metals. 5th ed.; Wiley:Hoboken: NJ, 2009, p.110.

(15) The seminal examples of 5- to 3-hapticity changes include

complexes from group 9 and group 6. See reference 14a and refer-

ences therein.

(16) (a) Bassetti, M.; Casellato, P.; Gamasa, M. P.; Gimeno, J.;

González-Bernardo, C.; Martín-Vaca, B. Insertion reactions of

alkynes into the Ru−H bond of indenylruthenium(II) hydride

complexes. Mechanism of the reaction of phenylacetylene with

[RuH(η5-C9H7)(dppm)] (dppm = bis(diphenylphosphino)methane).

Organometallics 1997, 16, 5470-5477; (b) Gamasa, M. P.; Gimeno,

J.; Gonzalez-Bernardo, C.; Martín-Vaca, B. M.; Monti, D.; Bassetti,

M. Phosphine substitution in indenyl- and

cyclopentadienylruthenium complexes. Effect of the η5 ligand in a

dissociative pathway. Organometallics 1996, 15, 302-308.

(17) For discussion of ground state contributions to the indenyl

effect, arising from weaker binding of the 5-indenyl ligand relative

to 5-Cp at coordinatively saturated metals, see: (a) Calhorda, M. J.;

Romão, C. C.; Veiros, L. F. The nature of the indenyl effect. Chem.

Eur. J. 2002, 8, 868-875; (b) Kubas, G. J.; Kiss, G.; Hoff, C. D.

Solution calorimetric, equilibrium, and synthetic studies of

oxidative addition/reductive elimination of cyclopentadiene

derivatives, C5R5H (R = H, Me, indenyl), to/from the metal

complexes M(CO)3(RCN)3/(η5-C5R5)M(CO)3H (M = chromium,

molybdenum, tungsten). Organometallics 1991, 10, 2870-2876.

(18) Pauling, L., College Chemistry. 2nd ed.; WH. Freeman and

Company: San Francisco, CA, 1957, p.406.

(19) Foley, D. A.; Dunn, A. L.; Zell, M. T. Reaction monitoring

using online vs tube NMR spectroscopy: seriously different results.

Magn. Reson. Chem. 2015, 54, 451-456.

(20) Hoops, S.; Sahle, S.; Gauges, R.; Lee, C.; Pahle, J.; Simus,

N.; Singhal, M.; Xu, L.; Mendes, P.; Kummer, U. COPASI—a

COmplex PAthway SImulator. Bioinformatics 2006, 22, 3067-3074.

(21) For a previous example of the use of COPASI in conjunc-

tion with reaction monitoring data obtained from ESI-MS, see: Luo,

J.; Theron, R.; Sewell, L. J.; Hooper, T. N.; Weller, A. S.; Oliver, A.

G.; McIndoe, J. S. Rhodium-catalyzed selective partial

hydrogenation of alkynes. Organometallics 2015, 34, 3021-3028.

(22) See for examples: (a) Bamford, K. L.; Chitnis, S. S.;

Stoddard, R. L.; McIndoe, J. S.; Burford, N. Bond fission in

monocationic frameworks: diverse fragmentation pathways for

phosphinophosphonium cations. Chem. Sci. 2016, 7, 2544-2552; (b)

Pike, S. D.; Pernik, I.; Theron, R.; McIndoe, J. S.; Weller, A. S.

Relative binding affinities of fluorobenzene ligands in cationic

rhodium bisphosphine η6–fluorobenzene complexes probed using

collision-induced dissociation. J. Organomet. Chem. 2015, 784, 75-

83.

(23) See for examples reference10 and (a) Couzijn, E. P. A.;

Zocher, E.; Bach, A.; Chen, P. Gas-phase energetics of reductive

elimination from a palladium(II) N-heterocyclic carbene complex.

Chem. Eur. J. 2010, 16, 5408-5415; (b) Torker, S.; Merki, D.; Chen,

P. Gas-phase thermochemistry of ruthenium carbene metathesis

catalysts. J. Am. Chem. Soc. 2008, 130, 4808-4814.

(24) Kobylianskii, I. J.; Widner, F. J.; Kräutler, B.; Chen, P. Co–

C bond energies in adenosylcobinamide and methylcobinamide in

the gas phase and in silico. J. Am. Chem. Soc. 2013, 135, 13648-

13651.

(25) (a) Butcher, C. P. G.; Dyson, P. J.; Johnson, B. F. G.;

Langridge-Smith, P. R. R.; McIndoe, J. S.; Whyte, C. On the use of

breakdown graphs combined with energy-dependent mass

spectrometry to provide a complete picture of fragmentation

processes. Rapid Commun. Mass Spectrom. 2002, 16, 1595-1598;

(b) Weinmann, W.; Stoertzel, M.; Vogt, S.; Wendt, J. Tune

compounds for electrospray ionisation/in-source collision-induced

dissociation with mass spectral library searching. J. Chromatogr. A

2001, 926, 199-209; (c) Weinmann, W.; Stoertzel, M.; Vogt, S.;

Svoboda, M.; Schreiber, A. Tuning compounds for electrospray

ionization/in-source collision-induced dissociation and mass spectra

library searching. J. Mass Spectrom. 2001, 36, 1013-1023; (d)

Begala, M.; Delogu, G.; Maccioni, E.; Podda, G.; Tocco, G.;

Quezada, E.; Uriarte, E.; Fedrigo, M. A.; Favretto, D.; Traldi, P.

Electrospray ionisation tandem mass spectrometry in the

characterisation of isomeric benzofurocoumarins. Rapid Commun.

Mass Spectrom. 2001, 15, 1000-1010; (e) Harrison, A. G. Energy-

resolved mass spectrometry: a comparison of quadrupole cell and

cone-voltage collision-induced dissociation. Rapid Commun. Mass

Spectrom. 1999, 13, 1663-1670; (f) Dookeran, N. N.; Yalcin, T.;

Harrison, A. G. Fragmentation reactions of protonated α-amino

acids. J. Mass Spectrom. 1996, 31, 500-508.

(26) We did see trace quantities of [Ru(η5-indenyl)(PPh3)2]+ in

this MS/MS experiment, but at less than 1/1000th the (normalized)

intensity of signal due to [Ru(η5-indenyl)(PPh3)(PR2H)]+ (see Figure

5a).

(27) (a) Derrah, E. J.; Pantazis, D. A.; McDonald, R.; Rosenberg,

L. A highly reactive ruthenium phosphido complex exhibiting Ru−P

π-bonding. Organometallics 2007, 26, 1473-1482; (b) Fryzuk, M.

D.; Montgomery, C. D.; Rettig, S. J. Synthesis and reactivity of

ruthenium amide-phosphine complexes. Facile conversion of a

ruthenium amide to a ruthenium amine via dihydrogen activation

and orthometalation. X-ray structure of

RuCl(C6H4PPh2)[NH(SiMe2CH2PPh2)2]. Organometallics 1991, 10,

467-473; (c) Cole-Hamilton, D. J.; Wilkinson, G. The reactions of

dihydridotetrakis(triphenylphosphine)ruthenium(II),

tetrakishydridotris(triphenylphosphine)ruthenium(II) and hydrido-

(η3-2-

diphenylphosphinophenyl)bis(triphenylphosphine)ruthenium(II)

with alkenes, dienes, ketones, aldehydes and weak acids. Nouv. J.

Chim. 1977, 1, 141-55; (d) James, B. R.; Markham, L. D.; Wang, D.

K. W. Stoicheiometric hydrogenation of olefins using

HRuCl(PPh3)3 and formation of an ortho-metallated ruthenium(II)

complex. J. Chem. Soc., Chem. Commun. 1974, 439-440.

11

For Table of Contents Only

Competitive binding of secondary and tertiary phosphines at a ruthenium indenyl complex reveals a

textbook dissociative substitution reaction whether studied at nanomolar concentrations by ESI-MS or at

millimolar concentrations by NMR, and provides, along with gas phase dissociation experiments, an

order of binding strengths for different phosphines. Fragmentation of unsaturated species show that

these activated complexes undergo orthometallation reactions.

download fileview on ChemRxivbelli-wu_et_al.pdf (879.37 KiB)

Competitive Ligand Exchange and Dissociation in Ru Indenyl ComplexesRoman G. Belli‡, Yang Wu‡, Hyewon Ji, Anuj Joshi, Lars P. Yunker, Lisa Rosenberg*, J. Scott McIndoe*

‡These authors contributed equally.

Department of Chemistry, University of Victoria, P. O. Box 1700, STN CSC, Victoria, British Columbia, Canada V8W 2Y2

Supporting Information Placeholder

ABSTRACT: Kinetic profiles obtained from monitoring the solution phase substitution chemistry of [Ru(η 5-indenyl)(NCPh)(PPh3)2]+ (1) by both ESI-MS and 31P{1H} NMR are essentially identical, despite an enormous difference in sample concentrationsfor these complementary techniques. These studies demonstrate dissociative substitution of the NCPh ligand in 1. Competitionexperiments using different secondary phosphine reagents provide a ranking of phosphine donor abilities at this relatively crowdedhalf-sandwich complex: PEt2H > PPh2H >> PCy2H. The impact of steric congestion at Ru is evident also in reactions of 1 withtertiary phosphines; initial substitution products [Ru(η5-indenyl)(PR3)(PPh3)2]+ rapidly lose PPh3, enabling competitiverecoordination of NCPh. Further solution experiments, relevant to the use of 1 in catalytic hydrophosphination, show that PPh2Hout-competes PPh2CH2CH2CO2But (the product of hydrophosphination of tert-butyl acrylate by PPh2H) for coordination to Ru, evenin the presence of a ten-fold excess of the tertiary phosphine. Additional information on relative phosphine binding strengths wasobtained from gas-phase MS/MS experiments, including collision-induced dissociation (CID) experiments on the mixed phosphinecomplexes [Ru(η5-indenyl)PP’P’’]+, which ultimately appear in solution during the secondary phosphine competition experiments.Unexpectedly, unsaturated complexes [Ru(η5-indenyl)(PR2H)(PPh3)]+, generated in the gas-phase, undergo preferential loss ofPR2H. We propose competing orthometallation of PPh3 is responsible for the surprising stability of the [Ru(η5-indenyl)(PPh3)]+

fragment under these conditions.

INTRODUCTIONWe recently reported the activity of a series of Ru(η5-

indenyl) complexes in the catalytic hydrophosphination ofactivated alkenes.1 Although participation of this half-sandwich species in C-C bond-forming catalysis is well-established,2 its activity for heteroatom addition chemistrysuch as hydrophosphination remains relatively under-explored.3 Our preliminary investigations of this catalysisshow that competing coordination of substrate secondaryphosphine PR2H, ancillary PPh3, and product tertiaryphosphine PR2R’ at Ru in this system plays an important rolein the speciation of the catalyst during the reaction. Forexample, 31P{1H} NMR spectra of catalytic mixtures from ourpreliminary studies show signals for multiple Ru-P-containingproducts, attributable to species containing one, two and threedifferent phosphine ligands.1 The possibility of productinhibition is a concern in harnessing such late metal systems incatalytic hydrophosphination, but its importance will be highlysusceptible to the complex balance of steric and electroniceffects governing reactivity at this relatively crowded Rucenter.

The coordination behaviour of phosphine ligands is well-documented and even quantified,4 but most data concerningtheir steric and electronic properties has been gathered fortertiary phosphines. It is possible to estimate Tolman cone

angles5 or other steric parameters6 for secondary phosphines,but the impact of the P-H bond on overall donor ability is notwell established. Thus, the synthesis and reactivity of Ru(η5-indenyl) mixed phosphine complexes are important to ourongoing studies and optimization of hydrophosphinationcatalysis, and we have a particular interest in evaluating thebehavior of secondary phosphines within the Ru coordinationsphere.7

Scheme 1. Activity of cation 1 for hydrophosphination

RPPh2H Ph2P R+

Ru

Ph3P

Ru

NCPhPh3P

1R = CN, CO2But

Among catalyst precursors we assessed in thehydrophosphination reactions of diphenylphosphine was[Ru(η5-indenyl)(NCPh)(PPh3)2]+ (1, Scheme 1). Cations suchas 1 and its phosphine-substituted derivatives arestraightforward to analyze by ESI-MS,8 since they are readilytransferred from solution phase to gas phase during theelectrospray process without any need for adventitiousassociation with some other charged species.9 Indeed, wepreviously exploited the electrospray technique to determine

Ru-P bond dissociation energies (BDEs) for [Ru(η5-indenyl)(PPh2H)3]+ using Fourier transform ion cyclotron resonancemass spectrometry (FTICR-MS).10 Those quantitative studiesinvolved collision-induced dissociation (CID) of mass-selected ions within the FTICR cell, and provided a lowerlimit for the Ru-PPh2H bond dissociation energy (BDE) of16.6 kcal/mol. We noted the difficulty in comparing thisabsolute, gas phase value for an η5-indenyl system with BDEvalues for Ru-PR3 available from solution calorimetry studiesof tertiary phosphine substitutions at {Ru(η5-Cp*)Cl}4, giventhe dependence of the calorimetry-derived values on solventand temperature, and the observed strong correlation of BDEsfor the η5-Cp* system with steric, as opposed to electronic,properties of the tertiary phosphines studied.11 Ideally, aclearer picture of the relative binding energies of phosphinescould be obtained by performing these painstaking CIDstudies on a broad range of analogous cationic complexes for agiven half-sandwich system.

A simple and useful alternative for ranking the affinity ofdisparate phosphines at a single, cationic metal fragment is thestudy of its substitution reactions in real time using the“pressurized sample infusion mass spectrometry” (PSI-ESI-MS) technique.12 In these experiments a continuous flow ofthe stirred reaction mixture in solution is transferred directlyinto the spectrometer under a pressure of ≤ 5 psi of inert gas.Provided the incoming and outgoing ligands differ in mass(e.g. a competition experiment between PtBu3 and PnBu3

would be uninformative), time-resolved speciation of allcationic complexes in the mixture will result. Here we reportthe use of PSI-ESI-MS to measure the relative kinetics ofcompetitive phosphine substitution reactions at complex 1, andwe confirm the results using 31P{1H} NMR analysis. Theseexperiments, complemented by further assessment in the gasphase of the relative binding strengths of different phosphinesin the mixed phosphine substitution products [Ru(η5-indenyl)PP’P”]+, have allowed us to rank a group ofsecondary and tertiary phosphines in terms of their affinitiesfor the cationic ruthenium center in this half-sandwich system.

RESULTS AND DISCUSSION1. Substitution kinetics from PSI-ESI-MS and 31P{1H} NMR

The benzonitrile (NCPh) ligand in cation 1 is easilydisplaced by a variety of phosphines. Indeed, the lability ofNCPh in this complex is such that unless very gentleionization conditions are used in the ESI-MS experiment (conevoltage 10 V),13 a significant proportion of the observed ionintensity involves the coordinatively unsaturated [Ru(η5-indenyl)(PPh3)2]+, from which the NCPh has been lost in thedesolvation process. Examination by PSI-ESI-MS of thereactivity of 1 toward a series of phosphines (PPh2H, PEt2H,PCy2H, PBun

3) shows that the rate of substitution is insensitiveto the identity and concentration of the incoming ligand (videinfra), consistent with a dissociative substitution mechanism(Scheme 2). This result might seem surprising in the context ofthe “indenyl effect”, an observation that indenyl complexestend to have accelerated substitution reactions relative to theirCp analogues. The rate enhancement is commonly attributedto the coordinative flexibility of the indenyl ligand inswapping between η5- and η3-modes, which can facilitate anassociative substitution mechanism.14 However, explicitexamples of such variable indenyl ligand hapticity do notinclude ruthenium complexes.15 A study on the insertion ofalkynes into the Ru-H bond of RuH(η5-indenyl)(dppm) found

evidence for an associative mechanism, but the authorsacknowledge that either hapticity change of the indenyl ligandor ring-opening of the strainedbis(diphenylphosphino)methane ligand could allow binding ofthe alkyne substrate.16a Moreover, a kinetic analysis of theaccelerated rate of phosphine substitution at RuCl(η5-indenyl)(PPh3)2 (the neutral precursor to complex 1) relative to its η5-Cp analogue found clear evidence for a dissociativemechanism in both cases.16b In that report, the enhanced rate ofsubstitution at the η5-indenyl complex (i.e. the indenyl effect)was attributed to the ability of the electron-rich indenyl ligandto stabilize the 16-electron intermediate formed via adissociative substitution pathway.17

Despite the dissociative nature of these NCPh substitutionreactions, when we monitor by PSI-ESI-MS the addition of a1:1 mixture of PEt2H and PPh2H (10 equivalents each) to 1,the time-dependent distribution of cationic species reveals adependence of the product ratio on the nature of the incomingligand (Figure 1a). The relative amounts of the two productsformed, [Ru(η5-indenyl)(PPh3)2(PPh2H)]+ (2a) and [Ru(η5-indenyl)(PPh3)2(PEt2H)]+ (2b) tell us about the relative rates ofthe rapid second step in the dissociative substitution (Scheme2): PEt2H reacts slightly faster than PPh2H with theunsaturated, 16e– intermediate [Ru(η5-indenyl)(PPh3)2]+, givinga higher proportion of product 2b (Scheme 2). This isconsistent with the fact that PEt2H is smaller and moreelectron-rich than PPh2H; it is a stronger donor ligand.

Figure 1. (a) Competitive reactions of 1 with a 10:10 mixture ofPPh2H:PEt2H at 45°C in PhF, as monitored by PSI-ESI-MS.Circles are normalized experimental data; lines are simulatedusing parameter estimation with COPASI. (b) 145.85 MHz31P{1H} NMR data for the same experiment in 2:1 CH2Cl2/C6D6 atRT.

Scheme 2. Competitive substitution of the nitrile ligand in1 with a mixture of PPh2H and PEt2H

slow

Ph3P

Ru

NCPhPh3P

1NCPh Ph3P

RuPh3P

PPh2H PEt2Hfast faster

Ph3P

Ru

PEt2HPh3P

2bPh3P

Ru

PPh2HPh3P

2a

The kinetic profile we observe for this competitionexperiment by ESI-PSI-MS was confirmed by an analogousexperiment monitored by 31P{1H} NMR (Figure 1b). Thesetwo experiments were carried out under strikingly differentconditions. Most obvious is the much higher sampleconcentration required for NMR, relative to MS. However, theeffective sampling frequency also differs for the twotechniques; the ESI-MS experiment produces one spectrumper second, while the relatively long delays used to render the31P{1H} NMR experiment approximately quantitative gave usone spectrum (128 scans) per 20 minutes. To address thisdifference, we conducted the NMR experiment at roomtemperature instead of 45°C, to slow it down. Based on therule of thumb that reaction rates double with each increase intemperature of 10°C,18 the NMR experiment should take about5× as long as the PSI-MS experiment, for this 24°C difference.The consumption of 1 to give 2a and 2b as monitored byNMR actually took about 18× as long (180 min as opposed to10 min at 45°C when monitored by ESI-MS); however furtherslowing of NMR tube reactions can occur due to poormixing.19 Nevertheless, the two kinetic profiles shown inFigure 1 are essentially identical. Given the factor of threemillion between concentrations used for the MS and NMRexperiments, this is compelling evidence for the cleanlydissociative nature of the substitution reaction and for thecoordination behaviors of these two secondary phosphines.

To tease out electronic vs. steric effects we also comparedthe reactions of PPh2H and PCy2H with complex 1, sincePCy2H has similar electronic properties to PEt2H but issignificantly bulkier (its cone angle of 148° is actually slightlylarger than that of PPh3 = 145°).4d The MS and NMR resultsin Figure 2 show clearly the importance of steric hindrance ofthe incoming phosphine; [Ru(η5-indenyl)(PPh3)2(PCy2H)]+ (2c)forms much more slowly than the PPh2H complex 2a.

Figure 2. (a) Competitive reactions of 1 with a 10:10 mixture ofPPh2H and PCy2H at 45°C in PhF, as monitored by PSI-ESI-MS.Circles are normalized experimental data; the lines weresimulated using parameter estimation in COPASI. (b) 145.85MHz 31P{1H} NMR data for the same experiment in 2:1CH2Cl2/C6D6 at RT.

Based on the results shown in Figures 1 and 2, the relativereactivity of these three secondary phosphines in thesubstitution of NCPh in 1 is PEt2H > PPh2H >> PCy2H. Weused COPASI20 to model the traces provided by ESI-MS, toextract kinetic parameters for the competitive substitutionreactions (solid lines in Figures 1a and 2a).21 Along with thetime-dependent speciation data, COPASI is given the startingconcentrations and the predicted sequence of elementary steps(e.g. Scheme 2). COPASI then fits the data and provides k1

(rate constant for the nitrile dissociation, and for the overallreaction); it also simulates accurately the ratios of the two k2

values, i.e. how much faster one of these fast, subsequentphosphine association steps is than the other. The modeling isconsistent with rates for these association reactions that arevery fast regardless of the incoming phosphine. Since thesimulation involved a relatively large number of independentparameters, we tested the rate constants resulting from one setof initial concentrations (e.g. 10 equivalents of the 1:1phosphine mixture, Figures 1 and 2) against a different set ofinitial concentrations (e.g. 100 equivalents of the phosphines,Figures S3, S5). Under these conditions, the simulationchanged in the same way as the experimental data. Theseresults not only validate the modeling but also provide furthersupport for the dissociative nature of the reaction, since theoverall rate of reaction remains the same with theconcentration change (falling in the range k1 = 0.006 ± 0.001for all experiments).

The trialkylphosphine PBun3 showed more complex

reactivity with cation 1 than did the secondary phosphines(Figure 3). Since there were multiple products, we did notpursue competition experiments and instead examined just thesimple substitution of NCPh in 1 with PBun

3. The rate ofdisappearance of 1 is the same as in all previous experiments,but the subsequent reactivity is more complex. As shown inScheme 3, the first species to occupy the vacant coordinationsite in the unsaturated intermediate [Ru(η5-indenyl)(PPh3)2]+ isPBun

3, which is present in large excess. However, the resultingcomplex [Ru(η5-indenyl)(PPh3)2(PBun

3)]+ (2d) forms onlytransiently (Figure 3b); it never reaches more than 10% of theinitial abundance of 1. Complex 2d decomposes to two newcomplexes, [Ru(η5-indenyl)(PPh3)(PBun

3)2]+ (3d) and[Ru(η5-indenyl)(PPh3)(PBun

3)(NCPh)]+ (4d), which correspondto the substitution of one PPh3 ligand in 2d by PBun

3 andNCPh, respectively (Scheme 3). The ratio of 3d and 4dformed depends on the amount of PBu3 added: smallerexcesses of PBun

3 produce less of the former (Figures S8, S9).

Figure 3. (a) Reaction of 1 with 100 equivalents of PBun3 at 45°C

in PhF, as monitored by PSI-ESI-MS. Circles are normalizedexperimental data and solid lines were simulated using COPASI.The different intensity scale in (b) highlights the lowconcentrations of transiently formed 2d. In MS experiments using50 and 10 equivalents PBun

3, signal due to this complex fallsmuch lower in the baseline, and 2d was not detected in thecorresponding 31P{1H} NMR experiment (10 equivalents PBun

3,Figures S6, S7).

Scheme 3. Reaction sequence used to model MS data foraddition of excess PBun

3 to 1 from Figure 3

PBun3 NCPhfast fast

Ph3P

Ru

PBun3Ph3P

2d

Bun3P

Ru

PBun3Ph3P

3d

fastPBun3

slow

Ph3P

Ru

NCPhPh3P

1NCPh Ph3P

RuPh3P

slow

PPh3

Ru

PBun3Ph3P

PhCN

Ru

PBun3Ph3P

4d

The results shown in Figure 3 indicate that the PPh3 ligandsin 2d are labilized by the initial replacement of NCPh in 1 byPBun

3, presumably through steric crowding; one PPh3 is thenlost quickly (relative to the initial loss of NCPh) to form theunsaturated complex [Ru(η5-indenyl)(PPh3)(PBun

3)]+. Thesolution now contains three possible candidates to occupy thatvacant coordination site: one equivalent of PPh3, a large excessof PBun

3, and one equivalent of NCPh. While the PPh3 isclearly uncompetitive (low relative concentration, cone angle145°), a high relative concentration of the somewhat lessbulky PBun

3 (cone angle 132°) allows the formation of 3d. Theapparent non-innocence of the nitrile ligand under theseconditions is somewhat surprising; despite its much lowerrelative concentration (NCPh:PBun

3 is 1:10, 1:50, or 1:100,vide infra) and its facile loss from 1, the much smaller NCPhcan reoccupy the vacant site to give 4d, at a rate that iscomparable to that for the formation of 3d. The numericalmodelling agrees with this sequence of events. Not only didthe parameter estimation provide reasonable matches for theexperimental data obtained using 100 equivalents of PBu3

(solid lines in Figure 3), but also the model responded in thesame way as the experiment when the number of equivalentswas dropped to 50 and then 10 (Figures S8, S9).

As described in the introduction, these studies of secondaryand tertiary phosphine substitution chemistry are relevant tothe potential activity of this half sandwich system in thecatalytic hydrophosphination of alkenes.1 The ability ofsecondary phosphines to compete effectively with the producttertiary phosphine for binding at Ru will be critical to thisactivity, and the above results suggest that this should not be aproblem, based primarily on steric arguments. Further supportfor this premise comes from additional experiments examiningthe reactions of 1 with PPh2H and the product ofhydrophosphination of tert-butyl acrylate by this secondaryphosphine: PPh2(CH2CH2CO2But). The reaction of 1 with justproduct phosphine gives a relatively similar productdistribution to that observed for PBun

3 (vide supra, FiguresS10, S11); monitoring by 31P{1H} shows the nitrilesubstitution product 2e forms transiently but the mixedphosphine nitrile complex [Ru(η5-indenyl)(PPh3)

(PPh2CH2CH2CO2But)(NCPh)]+, 4e, ultimately dominates theproduct mixture. Monitoring this reaction by PSI-ESI-MSshows the same final product distribution, but instead ofcomplex 2e, the apparently unsaturated complex[Ru(η5-indenyl)(PPh3)(PPh2CH2CH2CO2But)]+ is observedtransiently. (We attribute formation of this species to faciledissociation of PPh3 from sterically congested 2e during theelectrospray process; it may be stabilized in the gas-phase bychelation through the pendant ester group in the tertiaryphosphine.) This substitution chemistry may becomeimportant towards the end of catalysis when the ratio ofproduct phosphine to substrate PPh2H is very high, but in thepresence of a 10:10 mixture of PPh2H andPPh2CH2CH2CO2But only the PPh2H complex 2a is observed(Figures S13, S14). Even in a reaction of 1 with a 1:10 mixtureof PPh2H and PPh2CH2CH2CO2But, complex 2a dominates andonly trace amounts of PPh2CH2CH2CO2But-containingproducts are observed, by ESI-MS (e.g. 4e and[Ru(η5-indenyl)(PPh3)(PPh2H)(PPh2CH2CH2CO2But)]+, 6)(Scheme 4, Figure S15). These results indicate that the tertiaryproduct phosphine does not effectively compete with PPh2Hfor coordination at this cationic Ru center.

Scheme 4. Competitive substitution of the nitrile ligand in1 with a 1:10 mixture of PPh2H and PPh2CH2CH2CO2But;only 2a was observed by NMR; 2a, 4e and 6 were observedby MS.

slow

Ph3P

Ru

NCPhPh3P

1NCPh Ph3P

RuPh3P

PPh2H

PPh2CH2CH2CO2But

fastest

Ph3P

Ru

PPh2HPh3P

2a

Ph3P

Ru

PPh3P

2e

= P

not observed

PPh3

Ru

PPh3P

NCPh PPh2H

PhCN

Ru

PPh3P

4eHPh2P

Ru

PPh3P

6

major

trace trace

2. MS/MS experiments: an alternative gauge of relativephosphine binding affinities

When the competitive substitution reactions involvingaddition of two secondary phosphines to complex 1 areallowed to reach equilibrium (e.g. t >10 min for the PSI-ESI-MS experiments, after complex 1 has been completelyconsumed in the initial substitution reactions), the productmixtures include complexes that result from furthersubstitution of the PPh3 ligands in the mixed phosphinecomplexes 2a-c (e.g. Figures S2, S4). This is a straightforwardway to generate complexes of the type [Ru(η5-indenyl)(PP’P”)]+, which provides an opportunity to directly comparethe relative binding strength of the phosphine ligands throughMS/MS experiments. We can isolate in the gas phase a single[Ru(η5-indenyl)(PP’P”)]+ complex from a complicated mixtureof other complexes of this type (e.g. [Ru(η5-indenyl)PxP’yP”z]+

where x, y, z = 0-3 and x+y+z = 3), and accelerate it throughan argon-filled collision cell to initiate energetic collisions andraise the internal energy of the ion to the point thatunimolecular decomposition reactions occur, eliminatingneutral molecules (most often intact L-type ligands in the firstinstance) and new product ions. The relative propensity of thecomplex to lose one or another ligand during this collision-induced dissociation (CID) is gauged by the lowest collisionvoltage at which the cationic dissociation product is detectedand by the relative amount of dissociation product detected asthe collision voltage is increased (collectively referred tobelow as a “track”). Dissociation product tracks tell us aboutthe relative binding strength of the ligands in the complex ofinterest.22 This presents a convenient alternative to actuallyquantifying ligand binding strength (as described in theintroduction), which requires energy-resolved threshold CIDtechniques that often necessitate instrumentation that is notcommercially available,23 and relies on clean fragmentation toa single product.24

We selected the ion [Ru(η5-indenyl)(PPh3)(PEt2H)(PPh2H)]+,7, from the competitive ligand substitution mixture thatinitially produced complexes 2a-b (vide supra), and carriedout CID over a range of collision voltages. Figure 4 shows therelative intensities (abundances) of 7 and the resulting productions as a function of collision voltage; these are referred to as“breakdown curves”.25

Figure 4. Breakdown curves for the precursor ion [Ru(η5-indenyl)(PPh3)(PEt2H)(PPh2H)]+ (7).

There are three possible simple ligand dissociations that canoccur for complex 7, and we observe all three to varyingdegrees starting at the minimum voltage of ~0.6 V. The leastabundant is the ion generated via loss of PEt2H, suggestingthis ligand is the most tenaciously bound. Ions resulting fromloss of PPh2H and PPh3 appear with almost identical tracks;overall the three tracks suggest an order of binding strengthPEt2H > PPh2H ~ PPh3. As discussed further below, theseproduct ions themselves are susceptible to furtherfragmentation via loss of a second phosphine ligand, which iswhy their relative intensities drop to zero at collision voltageshigher than 2 V. Finally, at the highest collision voltages, lossof the final phosphine from all three monophosphine species

occurs, giving the common, highly unsaturated product[Ru(η5-indenyl)]+.

When the analogous CID experiment was conducted withthe ion [Ru(η5-indenyl)(PPh3)(PCy2H)(PPh2H)]+ (8) (FigureS16), PPh3 was the ligand most easily lost, with slightly lowerand almost identical tracks for loss of PCy2H and PPh2H. Thisexperiment highlights the relative importance of the steric andelectronic characters of the phosphines in this half-sandwichsystem; PPh3 and PCy2H have almost identical cone angles,but PCy2H should be more Lewis basic, so in this verysterically crowded complex PPh3 is the most weakly bound.

Collectively, the CID experiments for 7 and 8 show that thegreater steric pressure exerted by PCy2H relative to PEt2H hasthe effect of discriminating between PPh3 and PPh2H, as wellas making PCy2H significantly easier to dissociate than PEt2Hin a comparable coordination environment. Overall, a bindingstrength order between these four ligands emerges as PEt2H >PPh2H ~ PCy2H ~ PPh3, where the order between the last threeligands depends on the steric environment at the metalcomplex.

Figure 5. Breakdown curves observed in MS/MS experiments for (a) [Ru(η5-indenyl)(PPh3)2(PEt2H)]+ (2b) and (b) the in-source generatedion [Ru(η5-indenyl)(PPh3)(PEt2H)]+. In both experiments, [Ru(η5-indenyl)(PPh3)(PEt2H)]+ converts cleanly into [Ru(η5-indenyl)(PPh3)]+

without appreciable formation of [Ru(η5-indenyl) (PEt2H)]+.

3. MS/MS experiments: unexpected evidence for gas-phase orthometallation

We also used CID experiments to examine the gas phase,competitive dissociation chemistry of [Ru(η5-indenyl)(PPh3)2(PR2H)]+ for R = Ph (2a) and Et (2b), since thereliably facile loss of PPh3 from these complexes to generatecoordinatively unsaturated fragments [Ru(η5-indenyl)(PPh3)(PR2H)]+ should allow a simpler, two-way comparison of thebinding energies of the remaining phosphines. We weresurprised by the results of this experiment for the PEt2Hcomplex 2b (Figure 5a); although we did observe almostexclusive initial loss of PPh3 to form [Ru(η5-indenyl)(PPh3)(PEt2H)]+,26 continued ramping of the collision voltageshowed subsequent, exclusive loss of PEt2H to give [Ru(η5-indenyl)(PPh3)]+, instead of the loss of a second equivalentof PPh3 that we expected based on the stronger donor abilityof PEt2H (vide supra). This points to unusual stability of thegaseous [Ru(η5-indenyl)(PPh3)]+ fragment under theconditions of the CID experiment. We reproduced this resultusing an alternative experiment in which the ion [Ru(η5-indenyl)(PPh3)(PEt2H)]+ was generated “in-source”, byincreasing the cone voltage such that the initial CID occursduring the desolvation process (Figure 5b). This allowedexamination of the unsaturated precursor ion by MS/MSindependently of the presence of potentially complicatingfragments.

One possible explanation for the surprisingly facile loss ofPEt2H and apparent stability of the [Ru(η5-indenyl)(PPh3)]+

fragment (m/z 479) in these experiments involves gas-phasereactivity of the coordinatively unsaturated intermediate[Ru(η5-indenyl)(PPh3)(PEt2H)]+, which is “hot”(vibrationally excited) from collisions (Scheme 5). Althoughthe pressure in the collision cell is too low to allowintermolecular ion-molecule reactivity, intramolecularreactions could occur. In particular, we propose that this 16-electron species undergoes orthometallation of thecoordinated PPh3, a well-known intramolecular C-Hactivation reaction that occurs readily for Ru(II)

complexes.27 This would generate a Ru(IV) species (complexA, Scheme 5), from which dissociation of the neutralsecondary phosphine ligand PEt2H would now be favoured,relative to loss of the new hydride and 2-(o-C6H4PPh2)ligands (fragment also m/z 479).

Scheme 5. Proposed gas-phase orthometallationchemistry resulting in preferential loss of PEt2H overPPh3 from [Ru(η5-indenyl)(PPh3)(PEt2H)]+

Similar results were obtained for CID experimentsinvolving the PPh2H complex 2a (Figure S17), although inthis case PPh2H was more competitive with PPh3, both forinitial loss (about 2% PPh2H is lost from 2a to give thebis(triphenylphosphine) cation, compared to 0.1 % PEt2Hdissociation from 2b) and for the second ligand dissociation(about 5% of the PPh2H remains bound to Ru, compared to0% for PEt2H). These observations are reasonable: PPh3 islost more easily than PPh2H from 2a because it issignificantly larger than PPh2H and there are two PPh3

ligands; also, as shown above, the PPh2H is a slightlystronger donor. In addition, PPh2H is able to participate inorthometallation, unlike PEt2H but like PPh3. However, itspropensity to do so in this context is less for two reasons:

PPh2H has just four ortho protons while PPh3 has sixavailable to participate in the C-H activation at Ru, andsignificantly, the PPh3 ortho hydrogens are closer to themetal due to the larger PPh3 cone angle, relative to that forPPh2H. Figure S17 shows that at very high collisionenergies, [Ru(η5-indenyl)(PPh2H)]+ begins to form as aproduct of decomposition of [Ru(η5-indenyl)(PPh3)]+, whichcorresponds to a loss of benzyne that can be attributed to asecond orthometallation event.

CONCLUSIONSWe have demonstrated the facile evaluation of ligand

substitution chemistry and relative binding strengths atcationic metal complexes, through straightforward solutionand gas-phase experiments involving PSI-ESI-MS. Thekinetics of competitive ligand substitution reactions at aruthenium indenyl cation as monitored by PSI-ESI-MS arereproduced comprehensively by those obtained foranalogous experiments monitored by 31P{1H} NMR, despitea difference in solution concentrations of more than sixorders of magnitude. These solution experiments examiningmixtures of secondary and tertiary phosphines highlightespecially the impact of steric crowding at the rutheniumindenyl fragment, which has implications for its activity incatalytic hydrophosphination chemistry; substrate secondaryphosphines should compete well for binding in the presenceof (at least) equimolar amounts of product tertiaryphosphine. The ease of isolating gaseous cations containingtwo or three distinct phosphine ligands in the massspectrometer allowed us to further parse relative Ru-Pbinding strengths in this system via competitive CIDexperiments. These gas-phase studies reinforce some of ourconclusions from the solution studies concerning the balanceof electronic and steric influences of secondary and tertiaryphosphines in ruthenium coordination chemistry. Howeverrigorous one-on-one comparison of phosphine bindingenergies is precluded for this half-sandwich system byapparent competing C-H activation chemistry in the collisioncell. Although such phosphine orthometallation is routinelyobserved in solution chemistry, its importance in gas-phaseorganometallic chemistry was not previously established.

EXPERIMENTAL SECTIONGeneral details and instrumentation Chemicals and reaction mixtures were all handled under inert gas

atmosphere using standard glovebox and Schlenk techniques.Dichloromethane was freshly distilled from CaH or P2O5 andfluorobenzene was freshly distilled from P2O5 before use.Deuterated benzene (C6D6) was stored over sodium/benzophenoneand was degassed by three freeze−pump−thaw cycles and vacuum-transferred before use. All phosphines were purchased from StremChemicals Inc. and used without further purification. The rutheniumcomplex [Ru(η5-indenyl)(NCPh)(PPh3)2]+[B(C6F5)4]–, 1[B(C6F5)4],was synthesized by a literature method.1

Mass spectra were collected with a Micromass Q-Tof Micro massspectrometer in positive ion mode. Key parameters: capillaryvoltage, 2800 V; sample cone voltage, 10.0 V; extraction conevoltage, 1.0 V; desolvation temperature, 160 °C; sourcetemperature, 60 °C; cone gas flow, 0 L/h; desolvation gas flow, 40L/h; MCP voltage, 2700 V; collision voltage (MS study), 2 V;collision voltage (MS/MS study), programmed by Autohotkey(MassLynx) to increase the voltage by 1 V every 10.12 secondsfrom 0 to 100 V. Mass spectrometric data was normalized to totalion current before conducting kinetic analysis (this normalizationstep accounts for differences in spray conditions from one spectrumto the next).

NMR spectra were recorded at ambient temperature (294K) on aBruker AMX 360 spectrometer operating at 145.85 MHz for 31P.Chemical shifts are reported in ppm. 31P{1H} chemical shifts arereported relative to 85% H3PO4(aq).

Ligand substitution reactions monitored using pressurizedsample infusion (PSI)

An argon-pressurized (3 psi) Schlenk flask containing a solutionof 1[B(C6F5)4] (0.2 mg, 0.1 μmol) in fluorobenzene (10.5 ml)heated to 45 °C (or 60 °C for experiments withPh2PCH2CH2CO2But) was connected to the mass spectrometer viaPEEK tubing, and a stable signal established for the cation 1. Thereaction was initiated by the injection of a fluorobenzene solution(0.5 ml) of 10, 50 or 100 equiv of phosphine ligand(s). Amountsused: PPh2H (100 equiv, 13 µmol, 36 µL; 10 equiv, 1.3 µmol, 3.6µL); PEt2H (100 equiv, 13 µmol, 18 µL ; 10 equiv, 1.3 µmol, 1.8µL); PCy2H (100 equiv, 13 µmol, 39 µL; 10 equiv, 1.3 µmol, 3.9µL); PBun

3 (100 equiv, 13 µmol, 3.3 µL ; 50 equiv, 6.5 µmol, 1.6µL; 10 equiv, 1.3 µmol, 0.33 µL); Ph2PCH2CH2CO2But (10 equiv,1.3 µmol, 7 mg).

Ligand substitution reactions monitored using 31P{1H} NMRComplex 1[B(C6F5)4] (30 mg, 20 µmol) was dissolved in a

mixture of dichloromethane (0.4 mL) and C6D6 (0.2 mL) and 1,10,or 20 equiv of various phosphine ligands were added. Amountsused: PPh2H (10 equiv, 0.20 mmol, 35 μL; 1 equiv, 20 μmol, 3.5μL), PEt2H (10 equiv, 0.20 mmol, 23 μL), PCy2H (10 equiv, 0.20mmol, 44 μL), PBun

3 (10 equiv, 0.20 mmol, 50 μL),Ph2PCH2CH2CO2But (10 equiv, 0.20 mmol, 62 mg). The progress ofthe reactions was monitored by 31P{1H} NMR using a relaxationdelay (D1) of 10 s.

MS/MS studies examining competitive ligand dissociation from[Ru(η5-indenyl)PP’P”]+ and [Ru(η5-indenyl)PP’]+

The [Ru(η5-indenyl)PP’P”]+ complexes 5 and 6 were prepared insitu by adding 10 equiv each of two secondary phosphine ligands toa solution of 1[B(C6F5)4] (2 mg, 1 µmol) in dichloromethane (5mL). The [Ru(η5-indenyl)PP’]+ complexes were generated duringthe MS/MS experiment from precursor complexes [Ru(η5-indenyl)(PPh3)2(PR2H)]+ (R = Ph (2a), Et (2b)), which were prepared in situby the addition of 10 equiv of one secondary phosphine ligand to asolution of 1[B(C6F5)4] (2 mg, 1 µmol) in dichloromethane (5 mL).Amounts used: diphenylphosphine (36 µL, 13 µmol);diethylphosphine (18 µL, 13 µmol); dicyclohexylphosphine (39 µL,13 µmol).

MS/MS data for these tris(phosphine) complexes were collectedby isolating in the collision cell the appropriate m/z value for theprecursor ion of interest, and increasing the collision voltage 1 Vevery 10 seconds while observing the resulting product ions.

Additional, analogous MS/MS experiments were performed oncoordinatively unsaturated [Ru(η5-indenyl)PP’]+ complexes thatwere generated from the solutions containing 2a-b by increasing thesample cone voltage to 30 V, to dissociate the most weakly boundphosphine ligand during the in-source desolvation process.

ASSOCIATED CONTENT

Supporting InformationAdditional reaction monitoring data acquired using ESI-MS and31P{1H} NMR, and additional MS-MS data (pdf).

The Supporting Information is available free of charge on theACS Publications website.

AUTHOR INFORMATION

Corresponding Authors* Email: lisarose@uvic.ca, mcindoe@uvic.ca NotesThe authors declare no competing financial interests.

ACKNOWLEDGMENT

JSM thanks NSERC (Discovery and Discovery AcceleratorSupplements) for operational funding and CFI, BCKDF and theUniversity of Victoria for infrastructural support. LR thanksNSERC (Discovery) for funding. RGB thanks University ofVictoria (graduate fellowship) and NSERC (CGS-M and PGS-D) for funding.

REFERENCES(1) Belli, R. G.; Burton, K. M. E.; Rufh, S. A.; McDonald, R.;

Rosenberg, L. Inner- and outer-sphere roles of ruthenium phosphidocomplexes in the hydrophosphination of alkenes. Organometallics2015, 34, 5637-5646.

(2) Leading references for the use of the Ru(η5-indenyl) fragmentin catalysis: (a) Trost, B. M.; Ryan, M. C.; Maurer, D. Developmentof a coordinatively unsaturated chiral indenylruthenium catalyst.Org. Lett. 2016, 18, 3166-3169; (b) Trost, B. M.; Ryan, M. C.Indenylmetal catalysis in organic synthesis. Angew. Chem. Int. Ed.2016, 56, 2862-2879; (c) Manzini, S.; Fernández-Salas, J. A.;Nolan, S. P. From a decomposition product to an efficient andversatile catalyst: the [Ru(η5-indenyl)(PPh3)2Cl] story. Acc. Chem.Res. 2014, 47, 3089-3101; (d) Thamapipol, S.; Kündig, E. P.Intramolecular Diels–Alder reactions using chiral ruthenium Lewisacids and application in the total synthesis of ent-ledol. Org.Biomol. Chem. 2011, 9, 7564-7570; (e) Fung, W. K.; Huang, X.;Man; ManNg, S. M.; Hung, M. Y.; Lin, Z.; Lau, C. P. Dihydrogen-bond-promoted catalysis: catalytic hydration of nitriles with theindenylruthenium hydride complex (η5-C9H7)Ru(dppm)H (dppm =bis(diphenylphosphino)methane). J. Am. Chem. Soc. 2003, 125,11539-11544; (f) Alvarez, P.; Gimeno, J.; Lastra, E.; García-Granda,S.; Van der Maelen, J. F.; Bassetti, M. Synthesis and reactivity ofindenyl ruthenium(II) complexes containing the labile ligand 1,5-cyclooctadiene (COD): catalytic activity of [Ru(η5-C9H7)Cl(COD)].Organometallics 2001, 20, 3762-3771; (g) Koh, J. H.; Jung, H. M.;Kim, M.-J.; Park, J. Enzymatic resolution of secondary alcoholscoupled with ruthenium-catalyzed racemization without hydrogenmediator. Tetrahedron Lett. 1999, 40, 6281-6284; (h) Yamamoto, Y.;Kitahara, H.; Hattori, R.; Itoh, K. Ruthenium-catalyzed tandem [2 +2 + 2]/[4 + 2] cycloaddition of 1,6-heptadiyne with norbornene.Organometallics 1998, 17, 1910-1912.

(3) The activity of a RuCp* complex in similarhydrophosphination chemistry has been reported. Sues, P. E.;Lough, A. J.; Morris, R. H. Reactivity of ruthenium phosphidospecies generated through the deprotonation of a tripodal phosphineligand and implications for hydrophosphination. J. Am. Chem. Soc.2014, 136, 4746-4760.

(4) (a) Santiago, C. B.; Guo, J.-Y.; Sigman, M. S. Predictive andmechanistic multivariate linear regression models for reactiondevelopment. Chem. Sci. 2018, 9, 2398-2412; (b) Fey, N.; Orpen, A.G.; Harvey, J. N. Building ligand knowledge bases fororganometallic chemistry: Computational description ofphosphorus(III)-donor ligands and the metal–phosphorus bond.Coord. Chem. Rev. 2009, 253, 704-722; (c) Leyssens, T.; Peeters,D.; Orpen, A. G.; Harvey, J. N. How important is metal−ligandback-bonding toward YX3 ligands (Y = N, P, C, Si)? An NBOanalysis. Organometallics 2007, 26, 2637-2645; (d) Tolman, C. A.Steric effects of phosphorus ligands in organometallic chemistryand homogeneous catalysis. Chem. Rev. 1977, 77, 313-348.

(5) See reference 4d. These calculations may not accuratelyreflect the much tighter C-P-H angles (95-97°) in secondaryphosphines relative to the C-P-C angles in tertiary phosphines (99-110°). Quin, L. D., A Guide to Organophosphorus Chemistry.Wiley:Hoboken: NJ, 2000.

(6) Percent buried volume (%Vbur) is an important ligand stericparameter that can complement the Tolman cone angle in describingthe steric influence of phosphines on catalytic reactions. Wu, K.;Doyle, A. G. Parameterization of phosphine ligands demonstratesenhancement of nickel catalysis via remote steric effects. NatureChem. 2017, 9, 779; and references therein.

(7) For recent review of the coordination chemistry of secondaryphosphines see: Nell, B. P.; Tyler, D. R. Synthesis, reactivity, and

coordination chemistry of secondary phosphines. Coord. Chem.Rev. 2014, 279, 23-42.

(8) Chisholm, D. M.; Scott McIndoe, J. Charged ligands forcatalyst immobilisation and analysis. Dalton Trans. 2008, 3933-3945.

(9) In ESI-MS, neutral molecules often appear as [M+H]+ ions,for example, and their intensity in spectra is thus strongly dependenton a combination of basicity and surface activity.

(10) Westmore, J. B.; Rosenberg, L.; Hooper, T. S.; Willett, G.D.; Fisher, K. J. Determination of ruthenium−phosphorus bonddissociation energies by ES-FTICR mass spectrometry.Organometallics 2002, 21, 5688-5691.

(11) Luo, L.; Nolan, S. P. Relative binding energies of stericallydemanding tertiary phosphine ligands to the Cp*RuCl (Cp* = 5-C5Me5) moiety. Thermochemical investigation of coordinativelyunsaturated organoruthenium complexes. Organometallics 1994, 13,4781-4786.

(12) (a) Vikse, K. L.; Ahmadi, Z.; Luo, J.; van der Wal, N.; Daze,K.; Taylor, N.; McIndoe, J. S. Pressurized sample infusion: Aneasily calibrated, low volume pumping system for ESI-MS analysisof reactions. Int. J. Mass Spectrom. 2012, 323-324, 8-13; (b) Vikse,K. L.; Woods, M. P.; McIndoe, J. S. Pressurized sample infusion forthe continuous analysis of air- and moisture-sensitive reactionsusing electrospray ionization mass spectrometry. Organometallics2010, 29, 6615-6618.

(13) Chisholm, D. M.; Oliver, A. G.; McIndoe, J. S. Mono-alkylated bisphosphines as dopants for ESI-MS analysis of catalyticreactions. Dalton Trans. 2010, 39, 364-373.

(14) (a) Hartwig, J. F., Organotransition Metal Chemistry: FromBonding to Catalysis. University Science Books: 2010, p.250; (b)Crabtree, R. H., The Organometallic Chemistry of the TransitionMetals. 5th ed.; Wiley:Hoboken: NJ, 2009, p.110.

(15) The seminal examples of 5- to 3-hapticity changes includecomplexes from group 9 and group 6. See reference 14a andreferences therein.

(16) (a) Bassetti, M.; Casellato, P.; Gamasa, M. P.; Gimeno, J.;González-Bernardo, C.; Martín-Vaca, B. Insertion reactions ofalkynes into the Ru−H bond of indenylruthenium(II) hydridecomplexes. Mechanism of the reaction of phenylacetylene with[RuH(η5-C9H7)(dppm)] (dppm = bis(diphenylphosphino)methane).Organometallics 1997, 16, 5470-5477; (b) Gamasa, M. P.; Gimeno,J.; Gonzalez-Bernardo, C.; Martín-Vaca, B. M.; Monti, D.; Bassetti,M. Phosphine substitution in indenyl- andcyclopentadienylruthenium complexes. Effect of the η5 ligand in adissociative pathway. Organometallics 1996, 15, 302-308.

(17) For discussion of ground state contributions to the indenyleffect, arising from weaker binding of the 5-indenyl ligand relativeto 5-Cp at coordinatively saturated metals, see: (a) Calhorda, M. J.;Romão, C. C.; Veiros, L. F. The nature of the indenyl effect. Chem.Eur. J. 2002, 8, 868-875; (b) Kubas, G. J.; Kiss, G.; Hoff, C. D.Solution calorimetric, equilibrium, and synthetic studies ofoxidative addition/reductive elimination of cyclopentadienederivatives, C5R5H (R = H, Me, indenyl), to/from the metalcomplexes M(CO)3(RCN)3/(η5-C5R5)M(CO)3H (M = chromium,molybdenum, tungsten). Organometallics 1991, 10, 2870-2876.

(18) Pauling, L., College Chemistry. 2nd ed.; WH. Freeman andCompany: San Francisco, CA, 1957, p.406.

(19) Foley, D. A.; Dunn, A. L.; Zell, M. T. Reaction monitoringusing online vs tube NMR spectroscopy: seriously different results.Magn. Reson. Chem. 2015, 54, 451-456.

(20) Hoops, S.; Sahle, S.; Gauges, R.; Lee, C.; Pahle, J.; Simus,N.; Singhal, M.; Xu, L.; Mendes, P.; Kummer, U. COPASI—aCOmplex PAthway SImulator. Bioinformatics 2006, 22, 3067-3074.

(21) For a previous example of the use of COPASI in conjunctionwith reaction monitoring data obtained from ESI-MS, see: Luo, J.;Theron, R.; Sewell, L. J.; Hooper, T. N.; Weller, A. S.; Oliver, A. G.;McIndoe, J. S. Rhodium-catalyzed selective partial hydrogenationof alkynes. Organometallics 2015, 34, 3021-3028.

(22) See for examples: (a) Bamford, K. L.; Chitnis, S. S.;Stoddard, R. L.; McIndoe, J. S.; Burford, N. Bond fission inmonocationic frameworks: diverse fragmentation pathways for

phosphinophosphonium cations. Chem. Sci. 2016, 7, 2544-2552; (b)Pike, S. D.; Pernik, I.; Theron, R.; McIndoe, J. S.; Weller, A. S.Relative binding affinities of fluorobenzene ligands in cationicrhodium bisphosphine η6–fluorobenzene complexes probed usingcollision-induced dissociation. J. Organomet. Chem. 2015, 784, 75-83.

(23) See for examples reference10 and (a) Couzijn, E. P. A.;Zocher, E.; Bach, A.; Chen, P. Gas-phase energetics of reductiveelimination from a palladium(II) N-heterocyclic carbene complex.Chem. Eur. J. 2010, 16, 5408-5415; (b) Torker, S.; Merki, D.; Chen,P. Gas-phase thermochemistry of ruthenium carbene metathesiscatalysts. J. Am. Chem. Soc. 2008, 130, 4808-4814.

(24) Kobylianskii, I. J.; Widner, F. J.; Kräutler, B.; Chen, P. Co–Cbond energies in adenosylcobinamide and methylcobinamide in thegas phase and in silico. J. Am. Chem. Soc. 2013, 135, 13648-13651.

(25) (a) Butcher, C. P. G.; Dyson, P. J.; Johnson, B. F. G.;Langridge-Smith, P. R. R.; McIndoe, J. S.; Whyte, C. On the use ofbreakdown graphs combined with energy-dependent massspectrometry to provide a complete picture of fragmentationprocesses. Rapid Commun. Mass Spectrom. 2002, 16, 1595-1598;(b) Weinmann, W.; Stoertzel, M.; Vogt, S.; Wendt, J. Tunecompounds for electrospray ionisation/in-source collision-induceddissociation with mass spectral library searching. J. Chromatogr. A2001, 926, 199-209; (c) Weinmann, W.; Stoertzel, M.; Vogt, S.;Svoboda, M.; Schreiber, A. Tuning compounds for electrosprayionization/in-source collision-induced dissociation and mass spectralibrary searching. J. Mass Spectrom. 2001, 36, 1013-1023; (d)Begala, M.; Delogu, G.; Maccioni, E.; Podda, G.; Tocco, G.;Quezada, E.; Uriarte, E.; Fedrigo, M. A.; Favretto, D.; Traldi, P.Electrospray ionisation tandem mass spectrometry in thecharacterisation of isomeric benzofurocoumarins. Rapid Commun.

Mass Spectrom. 2001, 15, 1000-1010; (e) Harrison, A. G. Energy-resolved mass spectrometry: a comparison of quadrupole cell andcone-voltage collision-induced dissociation. Rapid Commun. MassSpectrom. 1999, 13, 1663-1670; (f) Dookeran, N. N.; Yalcin, T.;Harrison, A. G. Fragmentation reactions of protonated α-aminoacids. J. Mass Spectrom. 1996, 31, 500-508.

(26) We did see trace quantities of [Ru(η5-indenyl)(PPh3)2]+ inthis MS/MS experiment, but at less than 1/1000th the (normalized)intensity of signal due to [Ru(η5-indenyl)(PPh3)(PR2H)]+ (see Figure5a).

(27) (a) Derrah, E. J.; Pantazis, D. A.; McDonald, R.; Rosenberg,L. A highly reactive ruthenium phosphido complex exhibiting Ru−Pπ-bonding. Organometallics 2007, 26, 1473-1482; (b) Fryzuk, M.D.; Montgomery, C. D.; Rettig, S. J. Synthesis and reactivity ofruthenium amide-phosphine complexes. Facile conversion of aruthenium amide to a ruthenium amine via dihydrogen activationand orthometalation. X-ray structure of RuCl(C6H4PPh2)[NH(SiMe2CH2PPh2)2]. Organometallics 1991, 10, 467-473; (c)Cole-Hamilton, D. J.; Wilkinson, G. The reactions ofdihydridotetrakis(triphenylphosphine)ruthenium(II),tetrakishydridotris(triphenylphosphine)ruthenium(II) and hydrido-(η3-2-diphenylphosphinophenyl)bis(triphenylphosphine)ruthenium(II)with alkenes, dienes, ketones, aldehydes and weak acids. Nouv. J.Chim. 1977, 1, 141-55; (d) James, B. R.; Markham, L. D.; Wang, D.K. W. Stoicheiometric hydrogenation of olefins using HRuCl(PPh3)3

and formation of an ortho-metallated ruthenium(II) complex. J.Chem. Soc., Chem. Commun. 1974, 439-440.

For Table of Contents Only

Competitive binding of secondary and tertiary phosphines at a ruthenium indenyl complex reveals atextbook dissociative substitution reaction whether studied at nanomolar concentrations by ESI-MS or atmillimolar concentrations by NMR, and provides, along with gas phase dissociation experiments, anorder of binding strengths for different phosphines. Fragmentation of unsaturated species show thatthese activated complexes undergo orthometallation reactions.

12

download fileview on ChemRxivbelli-wu_et_al.docx (0.92 MiB)

Supporting information for

Competitive Ligand Exchange and Dissociation in Ruthenium Indenyl Complexes

Roman G. Belli‡, Yang Wu‡, Hyewon Ji, Anuj Joshi, Lars P. E. Yunker, Lisa Rosenberg*, J.

Scott McIndoe*

Department of Chemistry, University of Victoria, P.O. Box 1700 STN CSC, Victoria, BC V8W

2Y2, Canada.

Fax: +1 (250) 721-7147; Tel: +1 (250) 721-7173; +1 (250) 721-7181

E-mail: lisarose@uvic.ca, mcindoe@uvic.ca

‡These authors contributed equally.

Contents:

1. Monitoring substitution kinetics by PSI-ESI-MS and 31P{1H} NMR S2

1.1 Competitive substitution of 1 with mixtures of PPh2H and PEt2H S3

1.2 Competitive substitution of 1 with mixtures of PPh2H and PCy2H S5

1.3 Substitution of 1 with PBun3 S7

1.4 Individual and competitive substitutions of 1 with PPh2CH2CH2CO2But and PPh2H S11

2. MS/MS experiments: an alternative gauge of relative phosphine binding affinities S17

3. MS/MS experiments: unexpected evidence for gas-phase orthometallation S18

S2

1. Monitoring substitution kinetics by PSI-ESI-MS and 31P{1H} NMR

Figure S1. Preliminary experiments showing consumption of 1 in its reaction with 10 equiv of

PPh2H at 30°C (red), 45°C (blue), and 60°C (green) in PhF, as monitored by PSI-ESI-MS. The

lines are normalized experimental data; slight induction periods observed for the 45°C and 30°C

runs are attributable to slow mixing on addition of phosphine solutions. Inset shows an Eyring

plot with calculated activation parameters; the large positive DS‡ provides further evidence for

the dissociative substitution mechanism.

S3

1.1 Competitive substitution of 1 with mixtures of PPh2H and PEt2H (see also Figure 1)

Figure S2. 145.85 MHz 31P{1H} NMR spectra of the reaction of 1 with 10 equiv each of PPh2H

and PEt2H in 2:1 CH2Cl2/C6D6, showing speciation during and after complete consumption of

1.

S4

Figure S3. Competitive substitution of [Ru(η5-indenyl)(PPh3)2(NCPh)]+ (1) with a 100:100

mixture of PPh2H:PEt2H at 45°C in PhF, as monitored by PSI-ESI-MS. Circles are normalized

experimental data; the lines are simulated using parameter estimation with COPASI.

S5

1.2 Competitive substitution of 1 with mixtures of PPh2H and PCy2H (see also Figure 2)

Figure S4. 145.85 MHz 31P{1H} NMR spectra of the reaction of 1 with 10 equiv each of HPPh2

and HPCy2 in 2:1 CH2Cl2/C6D6, showing speciation during and after complete consumption of

1.

S6

Figure S5. Competitive substitution of [Ru(η5-indenyl)(PPh3)2(NCPh)]+ (1) with a 100:100

mixture of PPh2H:PCy2H at 45°C in PhF, as monitored by PSI-ESI-MS. Circles are normalized

experimental data; the lines are simulated using parameter estimation with COPASI.

S7

1.3 Substitution of 1 with PBun3 (see also Figure 3)

Figure S6. 145.85 MHz 31P{1H} NMR spectra of the reaction of 1 with 10 equiv PBun3 in 2:1

CH2Cl2/C6D6, showing speciation during and after complete consumption of 1.(O=PBun3 and

an unidentified impurity (*) derive from the PBun3 used in this experiment.)

S8

Figure S7. Reaction of 1 with 10 equiv of PBun3 at RT in 2:1 CH2Cl2/CcD6, as monitored by 31P{1H} NMR.

S9

Figure S8. Reaction of 1 with 50 equiv PBun3 at 45°C in PhF, as monitored by PSI-ESI-MS.

Circles are normalized experimental data; the lines are simulated using parameter estimation

with COPASI.

S10

Figure S9. Reaction of 1 with 10 equiv of PBun3 at 45°C in PhF, as monitored by PSI-ESI-MS.

Circles are normalized experimental data; the lines are simulated using parameter estimation

with COPASI. Although the COPASI fit has started to deviate from experimental at these low

concentrations, this plot shows that it responds in the correct direction (relative to the two higher

concentration experiments) to the change in initial concentrations.

S11

1.4 Individual and competitive substitutions of 1 with PPh2CH2CH2CO2But and PPh2H

Figure S10. 145.85 MHz 31P{1H} NMR spectra of the reaction of 1 with 10 equiv

PPh2CH2CH2CO2But in 2:1 CH2Cl2/C6D6, showing speciation during and after complete

consumption of 1. (The * marks an unidentified impurity (neutral, unreactive) in this batch of 1,

and the ‡ marks another unidentified, neutral impurity.)

S12

Figure S11. Reaction of 1 with 10 equiv PPh2CH2CH2CO2But at 60°C in PhF, as monitored by

PSI-ESI-MS. The lines are normalized experimental data. Formation of complex 2a results from

the presence of a <5% impurity of PPh2H in the tertiary phosphine reagent

PPh2CH2CH2CO2But. We attribute formation of the "unsaturated" fragment

[Ru(η5-indenyl)(PPh3)(PPh2CH2CH2CO2But)]+ to facile dissociation of PPh3 from the congested

substitution product 2e during the electrospray process, and its stabilization in the gas phase

by chelation through the pendant ester functionality as shown above.

S13

Figure S12. 145.85 MHz 31P{1H} NMR spectra of the reaction of 1 with 10 equiv PPh2H in 2:1

CH2Cl2/C6D6, showing speciation during and after complete consumption of 1. (The * marks an

unidentified impurity (neutral, unreactive) in this batch of 1, and the ‡ marks another unidentified,

neutral impurity.)

S14

Figure S13. 145.85 MHz 31P{1H} NMR spectra of the reaction of 1 with 10 equiv each of PPh2H

and PPh2CH2CH2CO2But in 2:1 CH2Cl2/C6D6, showing speciation during and after complete

consumption of 1.

S15

Figure S14. Reaction of 1 with 10 equiv each of PPh2H and PPh2CH2CH2CO2But at 60°C in

PhF, as monitored by PSI-ESI-MS. The lines are normalized experimental data.

S16

Figure S15. Reaction of 1 with 1 equiv PPh2H and 10 equiv PPh2CH2CH2CO2But at 60°C in

PhF, as monitored by PSI-ESI-MS. The lines are normalized experimental data.

S17

2. MS/MS experiments: an alternative gauge of relative phosphine binding affinities (see

also Figure 4)

Figure S16. Breakdown curves for the precursor ion [Ru(η5-indenyl)(PPh3)(PCy2H)(PPh2H)]+

(8) at relatively low collision voltages.

S18

3. MS/MS experiments: unexpected evidence for gas-phase orthometallation (see also

Figure 5)

Figure S17. Breakdown curves for the precursor ion [Ru(η5-indenyl)(PPh3)2(PPh2H)]+, 2a.

download fileview on ChemRxivbelli-wu_et_al_SI.pdf (1.46 MiB)