Continuous-Flow Alkane Dehydrogenation by Supported Pincer-Ligated Iridium Catalysts at Elevated Temperatures

Article abstract of DOI:10.1021/acscatal.8b01497

Pincer-ligated iridium complexes of the form [Ir(^R4PCP)L] (^R4PCP = κ³-C₆H₃-2,6-(XPR₂)₂; X = CH₂, O; R = tBu, iPr) are efficient homogeneous alkane dehydrogenation catalysts that have been reported to be highly active at temperatures of 240 °C or below. In this work, silica-supported [Ir(C₂H₄)(p-^tBu₂PO-^tBu4POCOP)] (1/SiO₂) was used to study a model continuous-flow gas-phase acceptorless alkane dehydrogenation system. This particular supported framework is thermally stable at temperatures up to 340 °C, 100 °C above the highest temperature at which analogous homogeneous complexes have been reported to show stable activity, with observed butane dehydrogenation rates of ca. 80 mol_butenes mol_cat.^-1 h^-1. Solid-state ³¹P MAS NMR and ATR IR are used to demonstrate that the backbone pincer ligand remains intact and coordinated at 340 °C. The complex is fully converted to [Ir(CO)(p-^tBu₂PO-^tBu4POCOP)] (3/SiO₂) above 300 °C. 3/SiO₂ is observed to be catalytically active at the higher temperatures tested, and reaction rates are comparable to those of 1/SiO₂. 3/SiO₂ and 1/SiO₂ act as resting states for the active 14-electron fragment, through dissociation of the CO or olefin ligand, respectively. Given that 3/SiO₂ is air resistant at ambient temperature and is structurally stable and catalytically active at elevated temperatures, it is a suitable candidate as a catalyst for the highly endothermic acceptorless dehydrogenation of alkanes.

Full text of DOI:10.1021/acscatal.8b01497

Subscriber access provided by Kaohsiung Medical University
Article
Continuous Flow Alkane Dehydrogenation by Supported
Pincer-Ligated Iridium Catalysts at Elevated Temperatures
Boris Sheludko, Molly T Cunningham, Alan S. Goldman, and Fuat E. Celik
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.8b01497 • Publication Date (Web): 16 Jul 2018
Downloaded from http://pubs.acs.org on July 16, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
Continuous Flow Alkane Dehydrogenation by Supported PincerꢀLigated Iridium Catalysts
at Elevated Temperatures
1
,2
1
1
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Boris Sheludko, Molly T. Cunningham, Alan S. Goldman, Fuat E. Celik *
1
Department of Chemistry and Chemical Biology
2
Department of Chemical and Biochemical Engineering
Rutgers, The State University of New Jersey, Piscataway, NJ
9
8 Brett Road, Piscataway, New Jersey 08854, United States
*
Tel.: +1 848 445 5558. Eꢀmail: fuat.celik@rutgers.edu
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Abstract
Pincerꢀligated iridium complexes of the form [Ir( PCP)L] ( PCP = κ ꢀC H ꢀ2,6ꢀ
R
4
R
4
3
6
3
(XPR ) ; X = CH , O; R = tBu, iPr) are efficient homogeneous alkane dehydrogenation catalysts
2
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
that have been reported to be highly active at temperatures of 240 °C or below. In this work,
t
tBu4
silicaꢀsupported [Ir(C H )(pꢀ Bu POꢀ POCOP)] (1/SiO ) was used to study a model
2
4
2
2
continuousꢀflow gasꢀphase acceptorless alkane dehydrogenation system. This particular
supported framework is thermally stable at temperatures up to 340 °C, 100 °C above the highest
temperature at which analogous homogeneous complexes have been reported to show stable
ꢀ1
ꢀ1
activity, with observed butane dehydrogenation rates of ca. 80 molbutenes molcat h . Solidꢀstate
3
1
P MASꢀNMR and ATR IR are used to demonstrate that the backbone pincer ligand remains
t
intact and coordinated at 340 °C. The complex is fully converted to [Ir(CO)(pꢀ Bu POꢀ
2
tBu4
POCOP)] (3/SiO ) above 300 °C. 3/SiO is observed to be catalytically active at the higher
2
2
temperatures tested, and reaction rates are comparable to those of 1/SiO . 3/SiO and 1/SiO act
2
2
2
as resting states for the active 14ꢀelectron fragment, through dissociation of the CO or olefin
ligand respectively. Given that 3/SiO is air resistant at ambient temperature and is structurally
2
stable and catalytically active at elevated temperatures, it is a suitable candidate as a catalyst for
the highly endothermic acceptorless dehydrogenation of alkanes.
Key Words
alkane dehydrogenation, pincer ligand, silica, supported molecular catalyst, thermal stability
2
ACS Paragon Plus Environment

Page 3 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
. Introduction
Olefins are highly versatile chemical building blocks, useful as precursors for polymers,
fuels, and a wide range of chemicals. They are commercially obtained from dehydrogenation of
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
alkanes in highꢀenergy processes such as STAR, Catofin and Oleflex. However, such systems
have significant shortcomings, including the fact that catalysts used industrially deactivate over
time by coking due to the high temperatures used and need to be regenerated, leading to nonꢀ
trivial downtime and energy expenditure. In addition, products need to be separated due to the
numerous olefins produced from even shortꢀchain alkane cracking.
2
ꢀ4
With the advent of the "shale gas revolution,"
a new feedstock for such
dehydrogenation processes has emerged. Onꢀpurpose dehydrogenation plants for light alkanes
5
such as propane are being opened or planned at an increased rate. these processes also suffer
from the same issues as more traditional petroleum cracking including a need for highꢀenergy
1
input, a lack of selectivity and catalyst deactivation over time.
Selectivity generally presents a major challenge for highꢀtemperature alkane conversion
processes. Thermal cracking is likely at high temperatures, leading to nꢀalkyl radicals which can
then perform hydrogen atom abstraction, or radical addition to olefins to form longer chain
alkanes which can crack further. Shorter contact times can typically mitigate this, leading to a
higher selectivity for the desired dehydrogenation products with minimal side products formed.
6
R
4
R4
3
Pincerꢀligated iridium catalysts of the form [Ir( PCP)L] ( PCP = κ ꢀC H ꢀ2,6ꢀ(XPR )
2 2;
6
3
t
i
X = CH , O; R = Bu, Pr) have been studied extensively in the context of homogeneous
2
(solutionꢀphase) alkane dehydrogenation both for their high activity as well as their
regioselectivity for formation of terminal olefin. Dehydrogenation by such catalysts can occur at
or below 240 °C, allowing for reaction conditions much milder than current industry standards.
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The lower equilibrium conversion at these milder conditions do not allow these catalysts to reach
the higher olefin yields achieved in industrial processes at much higher temperatures, but they
may be incorporated into tandem processes that upgrade and consume the olefins as they form,
7
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
producing valueꢀadded products in a single reactor. Since their discovery, various frameworks
including [Ir(POCOP)], [Ir(PCOP)], [Ir(PSCOP)], [Ir(CCC)], and [Ir(NCN)], as well as
complexes of metals other than iridium have been developed and investigated for various
8
iPr4
reactions including dehydrogenation. Recently, [Ir( PCP)] complexes analogous to the parent
tBu4
9
[
Ir( PCP)] complex have been used in the dehydrogenation of gasꢀphase alkane, functioning
as discreet molecular catalysts in a solventless system, suggesting their potential for
dehydrogenation of shorterꢀchain olefins such as ethane and propane.
Acronym
Ir(PCP)]
Substituents
[
Z = P, E = E = CH
1
2
2
[
Ir(POCOP)]
[Ir(PCOP)]
Z = P, E = E = CH
1 2
2
Z = P, E = CH , E = O
1
2
2
[
Ir(PSCOP)]
Ir(CCC)]
Ir(NCN)]
Z = P, E = S, E = O
1 2
Z = C, E = E = CH
1 2 2
Z = N, E = E = CH
1 2 2
[
[
In order to develop industrially relevant processes based on alkane dehydrogenation by
pincerꢀiridium complexes, it would be desirable to implement such selective catalysts in a
continuousꢀflow system. Supporting catalyst (precursors) of the form [Ir(COD)L ] on metal
2
oxide surfaces has afforded dehydrogenation activity which remained stable with the initial
1
0ꢀ12
addition of excess phosphine.
Pincerꢀligated catalysts have also previously been supported to
4
ACS Paragon Plus Environment

Page 5 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
R4
R4
the same end: supported complexes of the form [Ir(Xꢀ PCP)] and [Ir(Xꢀ POCOP)] (X =
1
3
functional group) on alumina also result in active dehydrogenation catalysts.
Iridium complexes of triptyceneꢀbased pincer ligands have also been supported on
1
4
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
alumina, presumably via a Lewis acidꢀLewis base interaction anchoring the catalyst in place.
More recently, MOFꢀsupported pincerꢀiridium complexes have also been observed to be
catalytically active in hydrogenation which, while being the reverse of acceptorless
dehydrogenation, suggests promise in the development of systems incorporating MOFꢀsupported
15
dehydrogenation catalysts.
In all reports of supported pincerꢀligated catalysts, a backbone substituent, most often in
the paraꢀ position of a central aryl group, is required to maintain activity. In the case of alumina
as the supporting material, Lewis basic substituents such as ꢀOK, ꢀOMe, and ꢀNMe have been
2
1
3
incorporated into active, supported catalysts. In the absence of such a paraꢀsubstituent,
oxidative addition of support hydroxyl groups to the iridium center occurs, resulting in Ir(III)
1
5ꢀ17
species.
Such catalysts are effective in olefin hydrogenation, with reported turnover
ꢀ1
16,17
frequencies of up to 0.95 s and stable activity for several days in the case of ethylene.
However, these Ir(III) catalysts are not expected to be good alkane dehydrogenation catalysts
compared to pincerꢀligated Ir(I) complexes as they would not have access to the established Ir(I)ꢀ
Ir(III) couple responsible for dehydrogenation with molecular catalysts.
While more active for dehydrogenation than their unfunctionalized analogues, γꢀaluminaꢀ
tBu4
t
tBu4
supported [Ir(pꢀKOꢀ POCOP)] and [Ir(pꢀ Bu POꢀ POCOP)] catalysts still show a 60ꢀ80%
2
drop in activity upon recycling under transfer dehydrogenation conditions, while γꢀaluminaꢀ
tBu4
supported [Ir(pꢀMe Nꢀ PCP)] showed no added recyclability as compared with solutionꢀphase
2
systems. Deactivation can occur by a variety of mechanisms including inhibition by products,
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
thermal decomposition, or undesirable interaction with the support such as the generation of a
restingꢀstate Ir(III) species outside of the catalytic cycle of interest.
In this work, we report development and kinetic studies of a continuousꢀflow gasꢀphase
system for alkane dehydrogenation by pincerꢀiridium complexes. Such a system allows access to
reaction conditions unattainable in a batch system – fine control over pressure and temperature,
reagent partial pressures, and residence time. Unlike in batch or semiꢀbatch experiments where
they may accumulate, products are continuously removed from the reactor, limiting the extent of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
t
tBu4
secondary reaction or product inhibition. In this system, we use [Ir(C H )(pꢀ Bu POꢀ POCOP)]
2
4
2
on various supports as a model catalyst precursor, and we demonstrate catalyst stability under
acceptorless alkane dehydrogenation conditions. Through MASꢀNMR, XPS and ATR FTIR, we
describe catalyst speciation as a function of reaction conditions, providing insight into the
catalyst species involved in the reaction.
2
2
. Experimental Methods
.1 Catalyst Preparation
All chemical syntheses and material preparations were performed under an airꢀfree
t
tBu
4
atmosphere unless otherwise noted. The ethylene complexes [Ir(C H )(pꢀ Bu POꢀ POCOP)]
2
4
2
tBu4
(
1) and [Ir(C H )( POCOP)Ir] (2) were synthesized according to previously established
2 4
1
3
t
tBu4
procedures. The carbonyl complex [Ir(CO)(pꢀ Bu POꢀ POCOP)] (3) and dihydride complex
2
t
tBu4
[IrH (pꢀ Bu POꢀ POCOP)] (4) were obtained from 1 using procedures used for similar
2
2
18
complexes.
6
ACS Paragon Plus Environment

Page 7 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
t
tBu
4
(
pꢀ Bu POꢀ POCOP): To a Schlenk flask charged with a stir bar were added 100 mg (0.79
2
mmol) phloroglucinol and 76 mg (3.17 mmol) sodium hydride and dissolved in 15 mL THF with
the evolution of hydrogen. This mixture was stirred at reflux for two hours, at which point 570
mg (3.17 mmol) chloroꢀdiꢀtertꢀbutylphosphine was added and let stir an additional two hours.
Volatiles were removed in vacuo and the solid residue was extracted several times with pentane
and filtered through a kimwipe plug. The washes were combined and volatiles were pulled off in
vacuo at 55 °C to yield the product as 391 mg (0.70 mmol, 88.6%) of a white solid. Purity was
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
13
confirmed by comparison to literature.
t
tBu
4
[
IrClH(pꢀ Bu POꢀ POCOP)]: To a Schlenk flask charged with a stir bar were added 229 mg
2
t
tBu
4
(0.409 mmol) (pꢀ Bu POꢀ POCOP) and 137 mg (0.204 mmol) [Ir(COD)Cl] and dissolved in
2
2
1
0 mL toluene. This solution was stirred at reflux for 16 hours, by which point the solution was
deep red in color. All volatiles were removed in vacuo and the residue was extracted several
times with benzene and filtered through a kimwipe plug. Solvent was removed in vacuo once
3
1
again to leave the product as a bright red microcrystalline powder (256 mg, 80% yield). P and
1
13
H NMR spectra matched well with those in the literature.
t
tBu
4
[
Ir(C H )(pꢀ Bu POꢀ POCOP)] (1): To a Schlenk flask charged with a stir bar were added 100
2 4 2
t
tBu
4
mg (0.127 mol) [IrClH(pꢀ Bu POꢀ POCOP)] and 14.3 mg (0.127 mmol) potassium tertꢀ
2
butoxide and dissolved in 5 mL benzene. This solution was stirred under ethylene at room
temperature for 16 hours, with a slight change in color to redꢀbrown. All volatiles were removed
in vacuo and the residue was extracted several times with benzene and filtered through a
kimwipe plug. Solvent was removed in vacuo once again to leave the product as a brown
3
1
1
microcrystalline powder (90 mg, 91% yield). P and H NMR spectra matched well with those
1
3
in the literature.
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
tBu4
(
POCOP): To a Schlenk flask charged with a stir bar were added 202 mg (1.83 mol)
resorcinol and 92 mg (3.83 mmol) sodium hydride and dissolved in 10 mL THF (caution:
evolution of hydrogen). This mixture was stirred at reflux for one hour, at which point 0.73 mL
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(690 mg, 3.8 mmol) chloroꢀdiꢀtertꢀbutylphosphine was added and let stir an additional 16 hours.
Volatiles were removed in vacuo and the solid residue was extracted several times with pentane
and filtered through a Kimwipe plug. The washes were combined and volatiles were pulled off in
3
1
1
vacuo at 55 °C to yield the product as a translucent colorless oil (637 mg, 89% yield). P and H
19
NMR spectra matched well with those in the literature.
tBu4
[
IrClH( POCOP)]: To a Schlenk flask charged with a stir bar were added 195 mg (0.489
tBu4
mmol) ( POCOP) and 150 mg (0.223 mol) [Ir(COD)Cl] and dissolved in 5 mL toluene. This
2
solution was stirred at reflux for 24 hours, by which point the solution was deep red in color. All
volatiles were removed in vacuo and the residue was extracted several times with toluene and
filtered through a kimwipe plug. Solvent was removed in vacuo once again to leave the product
3
1
1
as a brown microcrystalline powder (246 mg, 88% yield). P and H NMR spectra matched well
19
with those in the literature.
tBu4
[
(
Ir(C H )( POCOP)] (2): To a Schlenk flask charged with a stir bar were added 50.7 mg
2 4
tBu4
0.0810 mmol) [IrClH( POCOP)] and 9.5 mg (0.085 mmol) potassium tertꢀbutoxide and
dissolved in 10 mL toluene. This solution was stirred under ethylene at room temperature for 16
hours, with a slight darkening in color. All volatiles were removed in vacuo and the residue was
extracted several times with pentane and filtered through a kimwipe plug. Solvent was removed
in vacuo once again to leave the product as a brown microcrystalline powder (43 mg, 86% yield).
3
1
1
13
P and H NMR spectra matched well with those in the literature.
8
ACS Paragon Plus Environment

Page 9 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
t
tBu4
[
[
Ir(CO)(pꢀ Bu POꢀ POCOP)] (3): To a J. Young NMR tube was added 30 mg (0.039 mmol)
2
t
tBu
4
Ir(C H )(pꢀ Bu POꢀ POCOP)] (1) dissolved in 400 ꢁL C D This solution was then degassed
2
4
2
6
6.
and charged with 1 atm of carbon monoxide. The NMR tube was spun overnight, by which point
the color of the solution was a dark orange and quantitative conversion to the carbonyl complex
was observed by NMR along with generation of free ethylene. Volatiles were removed in vacuo
and the isolated yellow solid was used without further purification. Elemental analysis calculated
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
for C H O P Ir (777.91 g/mol): C 47.86, H 7.26; found: C 42.35, H 8.79. H NMR (400 MHz,
3
1
56
4 3
3
3
C D ): δ = 1.09 (d, J = 11.6 Hz, 18 H, 2 x P(tBu) ), 1.31 (virtual t, J = 7.2 Hz, 36 H, 4 x
6
6
P,H
2
P,H
3
1
1
P(tBu) ), 7.00 (s, 2 H, ArꢀH). P{ H} NMR (162 MHz, C D ): δ = 154.3 (s, 1 x uncoordinated
2
6
6
P(tBu) ), 202.6 (s, 2 x coordinated P(tBu) ).
2
2
t
tBu4
[
[
IrH (pꢀ Bu POꢀ POCOP)] (4): To a J. Young NMR tube was added 10 mg (0.013 mmol)
4 2
tBu
4
Ir(C H )(pꢀtBu POꢀ POCOP)] (1) dissolved in 400 ꢁL D ꢀtoluene This solution was then
2
4
2
8
.
degassed and charged with 1 atm of H . The NMR tube was spun overnight and >99%
2
conversion to the H and H complexes was observed by NMR along with generation of ethane.
2
4
Volatiles were removed in vacuo and the isolated brown solid was used without further
purification. Elemental analysis calculated for C H O P Ir (751.92 g/mol): C 47.92, H 7.77;
30 58
3 3
1
found: C 44.05, H 9.80. H NMR (400 MHz, D ꢀtoluene): dihydride: δ = ꢀ8.42 (bs, 2H, IrꢀH),
8
3
3
1
.10 (d, J = 11.6 Hz, 18 H, 2 x P(tBu) ), 1.30 (virtual t, J = 7.13 Hz, 36 H, 4 x P(tBu)2),
P,H 2 P,H
3
7
.09 (s, 2 H, ArꢀH). tetrahydride: δ = ꢀ17.6 (bs, 4 H, IrꢀH), 0.95 (d, J = 14.7 Hz, 18 H, 2 x
P,H
P(tBu) ), 7.00 (s, 2 H, ArꢀH). Note: Signals from chelating tertꢀbutyl protons overlapped with
2
31
1
signals from dihydride and thus were not observed. P{ H} NMR (162 MHz, D ꢀtoluene):
8
dihydride: δ = 156.4 (s, 1 x uncoordinated P(tBu) ), 211.5 (s, 2 x coordinated P(tBu) ).
2
2
tetrahydride: δ = 189.3 (s, 2 x coordinated P(tBu) ). Note: Signal from chelating phosphorus
2
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
atoms overlapped with signals from the same in the dihydride complex and thus were not
observed.
Mg(Al)O was prepared by postꢀsynthesis modification of hydrotalcite. Aluminum nitrate
nonahydrate, magnesium nitrate hexahydrate, sodium carbonate and sodium hydroxide were
obtained from Acros and used as received. Hydrotalcite was synthesized by the coprecipitation
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
0,21
method.
A solution prepared from Al(NO ) ·9H O and Mg(NO ) ·6H O in deionized water
3 3 2 3 2 2
was added dropwise into a stirred solution of Na CO and NaOH in deionized water at a constant
2
3
temperature of 60 °C, and stirring continued for an hour after mixing. The rinsed and filtered
hydrotalcite precipitate was dried at 110 °C, then ground. The powder was calcined at 700 °C for
three hours in a tube furnace to yield the hydrotalciteꢀderived Mg(Al)O.
Mesoporous silica (SigmaꢀAldrich) and γꢀalumina (Strem) were obtained commercially.
SiO , Al O , and Mg(Al)O were calcined in a gasꢀtight reactor equipped with valves to seal the
2
2
3
ꢀ1
samples in dry air following calcination. Calcination was carried out under 100 mL min using a
ꢀ1
ꢀ1
2
°C min ramp to 550 °C, held for 3 hours, and cooled to room temperature at 2 °C min . The
reactor was then sealed using the valves and transported to a highꢀvacuum line, evacuated, and
brought into a glovebox for use as catalyst support.
In a typical experiment, 5 mg of catalyst (1, 2, or 3) was supported onto 150 mg of
support (SiO , Al O , or Mg(Al)O) using incipient wetness impregnation. Approximately 3.2
2
2
3
wt% catalyst/support was prepared by dissolving the catalyst in a volume of benzene or toluene
equivalent to the pore volume of the support material. The nominal catalyst coverage calculated
from the amount of metal catalyst impregnated in the support is approximately 0.8 metal wt%
based on iridium. The actual amount of iridium deposited was determined by ICPꢀMS analysis
(Robertson Microlit Laboratories) to be 0.71 wt% and by XPS to be 0.77 wt%, in good
10
ACS Paragon Plus Environment

Page 11 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
agreement with the expected value. After impregnation, the solid was dried for one hour in the
glovebox prior to packing in the gasꢀphase flow reactor. The appearance of 1/SiO is of a redꢀ
2
purple powder, 2/SiO is a brown powder, and 3/SiO is a yellow powder.
2
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
.2 Steadyꢀstate Gasꢀphase Continuousꢀflow Catalytic Data
Reactions were carried out in a 6.4 mm outer diameter (OD) quartz tube reactor with an
expanded section of 12.5 mm OD packed with quartz wool to hold the catalyst powder in place.
The reactor was packed with supported catalyst under argon atmosphere and sealed with valves
prior to connection to the gasꢀflow manifold, and connected under inert gas flow. The reactor
was placed inside a resistively heated ceramic furnace with external temperature control, and the
catalyst bed temperature was measured with a Kꢀtype thermocouple placed in direct contact with
the catalyst bed.
Butane (Airgas, 99.99%) was used as received. Helium (Praxair, 99.999%) was passed
through an onꢀstream oxygen and moisture trap. Reaction products were analyzed using an
Agilent 7890B GC equipped with a GSꢀGASPRO capillary column (0.32 mm x 60 m) fitted with
a splitter plate connected to a flame ionization detector and a 5977A mass spectrometer. Unless
stated otherwise, all experiments were carried out at 1.27 atm total pressure by throttling a needle
valve located downstream from the reactor.
Gas flow rates reported were measured at room temperature and pressure. For
experiments at elevated pressure, the gas flow rates were increased to maintain constant
volumetric flowrate at the reactor pressure.
2
.3 Batch Catalytic Data
In a typical batch experiment, 1.5 mL glass ampules were sealed with either 100 µL
solution of 1 or 2 (1 mM) in dodecane or the equivalent amount of catalyst supported on 6.7 mg
11
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
SiO (to match catalyst loading used in continuousꢀflow runs) in the presence of 100 µL
2
dodecane. These ampules were heated from room temperature to 340 °C and held for 30 minutes
or 24 hours, at which point the oven was cooled down and the vials were removed. The ampules
were opened in a glovebox and the liquid contents sampled via GC. In the case of recycling
experiments, all volatiles were removed by mild heating under vacuum. Fresh dodecane was
added and the ampules were resealed. All experiments were performed in duplicate.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
.4 Catalyst Characterization
Solution phase NMR spectra were recorded on a Varianꢀ400 or Varianꢀ500 spectrometer.
3
1
The samples for solidꢀstate P MAS NMR spectra were packed into 3.2 mm zirconia solidꢀstate
3
1
NMR rotors under an argon atmosphere and sealed with tightꢀfitting rotor caps. Solidꢀstate P
MAS NMR spectra were recorded on a Bruker Avance III spectrometer operating at 162 MHz
with magic angle spinning (MAS) of 10 KHz. The rotors were spun with air and no change to
sample spectra was observed on the time scale of spectrum collection. All spectra were recorded
at room temperature. Spinning side bands were observed at 10 KHz intervals upfield and
downfield of immobile species’ signals, corresponding to the frequency of the rotor spin.
Attenuated total reflectanceꢀFourier transform infrared (ATRꢀFTIR) spectra were
measured on a Bruker ALPHA spectrometer equipped with a germanium crystal plate.
XꢀRay Photoelectron Spectroscopy (XPS) was measured under UHV using a Thermo
Fisher KꢀAlpha XPS instrument equipped with a monochromatic Al Kα line with photon energy
2
2
of 1486.7 eV. The estimated depth sensitivity in this geometry is about 10 nm. Spectra were
23
deconvoluted using the method described by Nesbitt and Banerjee. All reported binding
energies are in reference to the C (1s) peak at 284.8 eV.
12
ACS Paragon Plus Environment

Page 13 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
Multiꢀpoint BET surface area and pore volume distribution were measured at ꢀ196 °C by
nitrogen physisorption using an Autosorb1 (Quantachrome). Pore diameter and volume were
measured via NLDFT equilibrium physisorption at ꢀ196 °C with nitrogen as the adsorbent and
silica as the oxygen model. Samples were degassed under nitrogen at 100 °C for six hours prior
to surface area measurement in a 6 mm tube.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
. Results and Discussion
O P
Ir
O P
t
tBu
4
t
tBu
4
t
tBu4
[Ir(C H )(pꢀ Bu POꢀ
[Ir(C H )( POCOP)] [Ir(CO)(pꢀ Bu POꢀ POCOP)] [IrH (pꢀ Bu POꢀ POCOP)]
2 4 2 2 2
2
4
2
tBu
4
POCOP)]
1
2
3
4
O
P
O
P
H
Ir
O
P
O
Si
O
O
O
1/SiO2
2/SiO2
3/SiO2
4/SiO2
Figure 1. Molecular catalyst species (1, 2, 3, and 4) prepared and supported on SiO in the
2
present study. Expected structures for covalently anchored supported molecular catalysts from
chemisorption (1/SiO , 2/SiO , 3/SiO and 4/SiO ). Silica surface represented by one SiO
2
2
2
2
4
tetrahedron.
3
.1 Asꢀprepared Supported Catalyst
13
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Upon supporting 1 on SiO , chemisorption is expected to lead to a condensation reaction
2
between the paraꢀsubstituent phosphinite group and a surface hydroxyl group to form a covalent
2
4
SiꢀOꢀC bond between the molecular catalyst and the surface of the silica. The resulting
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
supported complex, 1/SiO , is anchored to a single site on the surface of the support, as
2
illustrated in Figure 1. The redꢀpurple color (Figure S1) of the supported complex is consistent
with the color of the complex in solution prior to supporting on SiO . To confirm that supported
2
31
catalyst structure was consistent with 1/SiO and with previous reports, solidꢀstate P MASꢀ
2
NMR was conducted to examine the environment of the chelating phosphorus atoms of the
pincer ligands.
3
1
Figure 2a shows the solutionꢀphase P NMR spectrum of 1, with two signals for
13
phosphorus corresponding to the paraꢀsubstituent phosphinite group at 150.6 ppm and the
1
3
chelating phosphorus atoms at 181.0 ppm with an approximate peak area ratio of 1:2,
consistent with the structure of 1. Upon supporting 1 on SiO , the majority of the complex
2
undergoes condensation with surface silanol groups at room temperature to generate 1/SiO . The
2
spectrum of the supported complex (Figure 2b) shows the formation of two major new peaks at
1
74.7 ppm and 72.5 ppm. A small amount of complex 1 remained in physisorbed or molecular
1
3
form, evidenced by the minor peaks at 150.6 and 181.0 ppm. The new peak at 174.7 ppm is
assigned to the chelating phosphorus atoms of the supported complex 1/SiO , shifted upfield
2
from 181.0 ppm for the physisorbed species. The chemisorbed and covalently anchored species
is anchored strongly to the surface and is relatively immobile, as indicated by the presence of
spinning side bands which would be absent in the case of more mobile species. The new peak at
24
7
2.5 ppm is assigned to diꢀtertꢀbutylphosphine oxide , the leaving group from the condensation
reaction of the para substituent phosphinite group of 1. The peak area ratio between the 72.5
14
ACS Paragon Plus Environment

Page 15 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
ppm and 174.7 ppm peaks was approximately 1:2, consistent with the stoichiometry of the
condensation reaction between 1 and a silanol group to form 1/SiO and diꢀtertꢀbutylphosphine
2
oxide. This binding interaction preserves the iridium in the Ir(I) oxidation state.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The interaction of pincer complexes lacking paraꢀsubstituents, including 2, with SiO2
surfaces has been studied extensively and leads to interaction through the iridium center as
1
6,17,25,26
shown for 2/SiO in Figure 1.
These chemisorbed Ir(III) complexes have been shown to
2
be active for hydrogenation reactions in batch and continuous flow systems.
The impact of supporting molecular iridium complexes on surface area was investigated
using the carbonyl complex 3 instead of the olefin complex 1. Pincerꢀtype iridium carbonyl
complexes have been anecdotally reported to be more stable with respect to air exposure relative
to olefin complexes. For example, the time scale over which the color of the complex changes
upon air exposure is significantly longer for carbonyl complexes, which we have also observed.
This higher airꢀtolerance permitted the collection of accurate BET surface area measurements
with low risk of the catalyst changing oxidation state or structure over the timeꢀscale of the BET
2
ꢀ1
experiment. The surface area of the calcined SiO was 578 m g . Upon supporting 3, the surface
2
2
ꢀ1
3
ꢀ1
area of 3/SiO decreased to 510 m g and the pore volume decreased from 0.77 cm g to 0.62
2
3
ꢀ1
cm g indicating the presence of catalyst within the mesopores of the support. Also observed
,
were a small drop in larger diameter pores (6 – 10 nm pores) and a decrease in the modal pore
width from 5.4 to 5.1 nm, both consistent with a submonolayer surface coverage of iridium
2
7
complexes. Assuming 1.67 silanol groups per square nanometer for SiO calcined at 550 °C,
2
the expected ratio of SiOH groups to Ir atoms is approximately 3,000.
3
1
Comparison of the solutionꢀphase P NMR of 3 (Figure 2h) with the solidꢀstate NMR of
3/SiO (Figure 2g) confirmed that the carbonyl complex was supported in a manner analogous to
2
15
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
that of the olefin complex 1/SiO As observed with 1, the chemical shift of the paraꢀphosphinite
2
.
1
3
group in 3 at 151.9 ppm disappeared upon supporting on SiO , yielding a signal at 73.4 ppm
2
attributable to diꢀtertꢀbutylphosphine oxide. The substitution of the ethylene ligand by CO has a
strong effect on the chemical shift of the chelating phosphorus atoms, resulting in a downfield
shift from 181.0 ppm in 1 to 200.1 ppm in 3; the corresponding chemical shifts for the supported
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
species are 174.7 ppm and 202.8 ppm for 1/SiO and 3/SiO , respectively.
2
2
The dihydride analogue of 1, species 4, could be considered as a viable catalyst precursor
for supporting on SiO , and yet was found to interact with the surface in a very different way
2
than 1/SiO . 4 was prepared and supported on SiO using the same procedure as for 1 and 3. The
2
2
3
1
P NMR showed a significant shift in peak corresponding to the chelating phosphorus atoms
from 206.8 ppm in the solution phase spectrum of 4 to 175.3 ppm in the MAS spectrum of the
supported 4/SiO (Figure S2). The new peak is similar in chemical shift to the chelating
2
16
phosphorus atoms in 2/SiO at 170 ppm, where a surface hydroxyl group is oxidatively added
2
to the iridium center, with simultaneous loss of H . A paraꢀphosphinite peak appears in the
2
spectrum of 4/SiO at 153.4 ppm, similar to 1/SiO . However, the spectrum of 4/SiO shows
2
2
2
spinning side bands for this peak indicating an immobilized species, unlike in 1/SiO , where
2
spinning side bands for this peak are totally absent, indicating a mobile and physisorbed species.
The preservation of the paraꢀphosphinite peak shows that the complex has not anchored via
condensation reaction with a silanol group. As this binding mode leads to an Ir(III) complex,
activity towards dehydrogenation is not expected.
These results serve to shed light on future rational design of such supportable catalysts.
Even with an established supporting strategy, a highly reactive and unhindered iridium center
will preferentially react with the surface and form an undesired product. While the olefin and
16
ACS Paragon Plus Environment

Page 17 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
carbonyl ligands in 1 and 3 were sufficient to prevent interaction of the Ir center with surface
silanol groups, the analogous iridium dihydride complex behaves much more similarly to 2 when
supported on silica. It is important to note, however, that since the oxidative addition of surface
silanol groups is reversible while phosphinite generation is irreversible, given sufficient time
under reaction conditions (200 °C under alkane flow), this species may ultimately form the paraꢀ
supported surface species, presumably with bound olefin in the fourth coordination site.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a)
b)
*
*
*
*
*
*
*
≠
c)
d)
*
≠
*
*
*
*
≠
e)
*
≠
*
f)
*
g)
h)
*
*
350
300
250
200
150
100
50
0
δ
3
1
31
Figure 2. Solutionꢀphase P NMR of a) 1 and h) 3 in C D ; solidꢀstate P MASꢀNMR of
6
6
freshly supported b) 1/SiO and g) 3/SiO , and cꢀf) 1/SiO following continuous butane
2
2
2
dehydrogenation for 1.5 hours at 200 °C, 250 °C, 300 °C, and 340 °C respectively. Spinning
≠
ꢀ1
̇
sidebands denoted with *. Catalyst = 1/SiO , V = 20 mL min , P
= 0.32 atm, P = 0.95
2
total
butane
He
atm, Ptotal = 1.27 atm.
3
.2 SteadyꢀState Continuous Flow Butane Dehydrogenation
The primary advantage of a properlyꢀsupported heterogeneous catalyst is the trivial
separation of catalyst from reactants and products, and this is best exemplified in continuous
17
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
flow reaction conditions. To explore the effectiveness of supported pincerꢀtype iridium catalysts,
the acceptorless dehydrogenation of butane over catalyst 1/SiO was selected.
2
Prior to the present study, 240 °C was the highest temperature for which data was
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
available for catalytic alkane dehydrogenation by pincerꢀligated iridium complexes. Above this
temperature, rate data were found to be unreliable as the catalyst decomposed and yielded a
13
black solid that precipitated from solution or collected on reactor vessel walls. To test the
temperature stability of supported 1/SiO , steadyꢀstate butane dehydrogenation was measured at
2
temperatures between 200 °C and 340 °C, shown in Figure 3a. These experiments revealed an
unprecedented stable reactivity for pincerꢀligated iridium catalysts at 340 °C. Previously, 200ꢀ
2
8
240 °C was considered high temperature for molecular dehydrogenation catalysts.
Due to the unfavorable thermodynamics for alkane dehydrogenation at the temperatures
studied here, it is important to note that reaction can potentially reach equilibrium conversion of
the starting alkane within the catalyst bed of the reactor. Reaction rates reported in this study
primarily denote the net rate of forward reaction under these conditions, although reaction rates
far from equilibrium can be measured by manipulating the residence time in the reactor (see e.g.
Figure 6 discussed below).
ꢀ1
At 200 °C the net forward rate of reaction was low, on the order of 1 h , but the activity
31
was stable and did not decline with time on stream. The P MASꢀNMR spectrum of the catalyst
after 1.5 hours of exposure to reaction conditions at 200 °C (Figure 2c) was essentially
indistinguishable from freshly supported 1/SiO , except that no signal could be found for
2
physisorbed catalyst species, suggesting that the heating resulted in physisorbed species
covalently attaching to the surface. The similarities in the NMR spectra between 1/SiO freshlyꢀ
2
supported and after 1.5 hours of reaction at 200 °C indicate the similar chemical environments of
18
ACS Paragon Plus Environment

Page 19 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
the chelating phosphorus atoms, which did not change significantly with the identity of the
coordinating olefin (ethylene prior to reaction, and presumably butene(s) after reaction). The
catalyst remained unchanged in color, indicative of very little change in electronic properties of
the complex upon binding a different, albeit very similar, ligand (Figure S1).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
31
Upon heating the supported catalyst to 250 °C, the P MASꢀNMR spectrum showed a
new small peak at 201.2 ppm (Figure 2d). A reduction in the diꢀtertꢀbutylphosphine oxide peak
was also observed relative to the peak corresponding to the chelating phosphorus atoms, possibly
due to desorption of the molecule from the surface at higher temperature. At this temperature, the
reaction rate was still stable and so higher reaction temperatures were attempted. At 300 °C, a
ꢀ1
steady butene formation rate of 35 h was observed over 1.5 hours of time on stream. At this
31
temperature (Figure 2e), the peak in the P MASꢀNMR spectrum corresponding to the olefin
complex at 174.7 ppm had disappeared, and the peak at 201 ppm had grown. Thus, the resting
state of the catalyst was now a different species than the starting olefin complex, though it
showed similar reactivity. The diꢀtertꢀbutylphosphine oxide peak also disappeared, indicating
that desorption of the volatile molecule was complete.
ꢀ1
Upon continued heating of the catalyst, a reaction rate of 84 h was reached at a
temperature of 340 °C, and was stable for 1.5 hours. By this temperature, the color had changed
31
to brownꢀyellow, though the P MASꢀNMR (Figure 2f) showed no significant changes relative
to the spectrum at 300 °C. The peak at 201 ppm coincided well with the freshly supported
carbonyl complex 3/SiO (this is further discussed in Section 3.4). It is worth noting that there is
2
no evidence for deactivation of the catalyst via oxidation at the phosphine positions or
dimerization up to at least 340 °C, both of which would result in significant changes in the
chemical shift of the corresponding phosphorus signals.
19
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
00
360
350
a)
b)
4
4
3
40
00
60
300
250
200
150
320
7
5
2
5
0
5
0
280
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
240
200
100
1000
320
1750
0
250
500
750
1000
1250
1500
Time (min)
Time (min)
Figure 3. Catalyst activity as a function of temperature. a) 200 ꢀ 340 °C. b) 360 ꢀ 440°C.
Catalyst = 1/SiO , V = 20 mL min , P
ꢀ1
̇
= 0.32 atm, P = 0.95 atm, P = 1.27 atm.
2
total
butane
He total
Heating the catalyst above 340 °C under butane dehydrogenation conditions led to a color
change to gray and then black (Figure S1), the appearance of a redꢀorange band of color on the
quartz tube wall downstream of the reactor heated zone, and declining reaction rates as a
function of time on stream (Figure 3b). The color change is consistent with that observed with
pincerꢀiridium catalysts in solution when heated above reaction temperatures of ca. 200ꢀ240 °C.
2
9
The band of color downstream of the reactor likely arises from the sublimation of a small
amount of pincerꢀiridium complex at high temperature, and the leachate then condensing in a
31
1
colder spot downstream of the heated zone. This condensate was analyzed by P and H
solutionꢀphase NMR and showed several peaks consistent with symmetric, chelated pincer
species.
While the activity of the catalyst was not stable above 340 °C, it did continue to increase
as a function of temperature despite eight hours of exposure to temperatures where the activity
ꢀ1
was unstable. The highest initial reaction rate measured at any temperature was 315 h at 440
ꢀ1
°
C, which declined to 186 h after four hours. This suggests that some of the supported
molecular catalyst survived at 440 °C and was responsible for the reaction rates observed, while
the concentration of such species decreased as a function of time. Continued exposure to reaction
20
ACS Paragon Plus Environment

Page 21 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
ꢀ1
conditions at 440 °C for over three days led to a further decline in reaction rate to 149 h , at
which it stabilized. This new steady state activity can therefore be attributed to the thermal
decomposition product of the molecular catalyst, which retains some lower activity. The reaction
temperature was subsequently reduced to compare the steady state activity of the decomposed
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ1
catalyst to the anchored molecular catalyst. At 340°C, the reaction rate had dropped to 21 h ,
ꢀ1
ꢀ1
down from 77 h in the fresh catalyst, and at 200 °C, the reaction rate was 0.27 h , down from
ꢀ
1
1
.1 h .
31
P MASꢀNMR of the catalyst following reaction at 440 °C showed no strong signals,
indicating a complete loss of chelating pincer ligands. This means that both the oxidation of the
chelating phosphines and dimerization of the anchored complexes are unlikely deactivation
mechanisms. The P (2p) region of the XPS spectrum (Figure 4) also showed a complete loss of
phosphorus signal after heating to 440 °C, compared to the strong phosphorus signal in the asꢀ
prepared sample. Thus, relative to the liquid phase, covalently supporting the catalyst on SiO2
greatly increases the thermal stability of the pincerꢀIr catalysts and delays the onset temperature
for catalyst deactivation by ca. 100 °C.
Analysis of the Ir (4f) region of the asꢀprepared sample XPS spectrum (Figure 4) shows a
spinꢀorbit doublet consistent with Ir(I) species, while lacking evidence of metallic iridium or
30
Ir(0). Following reaction conditions for over three days at 440 °C, the peaks broadened slightly
and shifted to higher binding energy. The shift in the spectrum to higher binding energy indicates
an increase in the oxidation state upon decomposition of the catalyst, rather than a reduction to
iridium metal. The similar Ir (4f) peak intensity between the asꢀprepared and spent catalyst
indicates that the catalyst loss due to sublimation discussed above was small compared to the
total catalyst loading, and not a major source of deactivation. The loss of chelating phosphorus
21
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
from the pincer ligand and the oxidation of the iridium are responsible for the reduction in
activity following high temperature treatment.
4
40 °C
440 °C
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
asꢀprepared
asꢀprepared
138
136
134
132
130
128
70
68
66
64
62
60
58
Binding Energy (eV)
Figure 4. P (2p) and Ir (4f) regions of XPS spectra of asꢀprepared 1/SiO compared with those of
2
ꢀ1
̇
a sample after exposure to reaction conditions at 440 °C for three days. V
= 20 mL min ,
total
Pbutane = 0.32 atm, P = 0.95 atm, P = 1.27 atm.
He
total
While the reactivity of 1/SiO observed at 340 °C appeared to be stable over 1.5 hours of
2
reaction time, the fact that 360 °C was sufficient to cause deactivation necessitated a longer timeꢀ
onꢀstream experiment to determine if any deactivation occurs at 340 °C. The results, shown in
Figure 5, show that even after three days of continuous operation, there was no appreciable
decline in activity for butene formation, with a calculated turnover number over 3,500.
Covalently anchoring pincerꢀIr species on high surface area SiO support creates a true
2
heterogeneous catalyst with enhanced thermal stability relative to the molecular homogeneous
catalyst. This stability seems likely to arise from the reduced mobility of the Ir species, limiting
their ability to engage in intermolecular reactions.
22
ACS Paragon Plus Environment

Page 23 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
8
0
60
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
0
0
0
2
0
10 20 30 40 50 60 70 80
Time (hours)
ꢀ
1
̇
Figure 5. Catalyst activity as a function of time on stream. Catalyst = 1/SiO , V = 5 mL min ,
Pbutane = Ptotal = 1.27 atm. T = 340 °C.
2
total
These results compare favorably with prior reports of homogeneous catalysis by solutionꢀ
R4
phase [Ir( PCP)]ꢀtype pincer complexes. To the best of our knowledge, there are no prior
reports on batchꢀphase acceptorless linear alkane dehydrogenation making use of
R4
tBu
4
[
Ir( POCOP)]ꢀtype pincer complexes. [Ir( PCP)] complexes afforded no discernible products
at reflux in the acceptorless dehydrogenation of nꢀdecane (at the boiling point of nꢀdecane, 174
3
1
°C) in a sealed system. In the dehydrogenation of nꢀundecane (bp 196 °C) at reflux in an open
ꢀ1
system that allows the escape of hydrogen, an initial rate of 46 h has been observed with
tBu
4
32
[
IrH2( PCP)], which decreased by about a factor of two within the first two hours. In
contrast, the continuousꢀflow gasꢀphase system described here displays negligible activity loss
over several days.
The approach to chemical equilibrium between butane and the reaction products was
explored by manipulation of the space time in the reactor. At long space times, with sufficient
contact time between reactant (butane) and product gases (1ꢀbutene, cisꢀ2ꢀbutene, transꢀ2ꢀ
butene, hydrogen) and the catalyst, the gases should be fully equilibrated with each other. This
23
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
equilibrium conversion is the maximum achievable at a fixed set of reaction conditions. The
measured reaction rate at this point is the net forward reaction rate at equilibrium, and will not be
affected by further increasing the space time. Reducing the space time will lead to a monotonic
decrease in conversion, accompanied by an increase in the net forward reaction rate as the
concentration of product gases is reduced. This expected behavior was observed during a space
time (τ) test at 340 °C. Increasing the space time led to an increase in conversion, with a plateau
at 3.7% conversion (Figure 6). Conversely, the net reaction rate reduced with increasing space
time as reactants and products approached equilibrium concentration in the gas phase. However,
at the shortest space times, the reaction was sufficiently far from equilibrium, and much larger
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ1
net reaction rates were measured. The net reaction rate was only 13.3 h at sufficiently long
ꢀ1
space times, but increased to 200 h at very short contact times. The continuous reaction
conditions used in the present study thus have another advantage in the study of molecular
catalysts relative to batch operation, in that the true reactivity of the catalyst at sufficiently small
contact times can be studied. Extrapolating the measured data to zero space time and zero
conversion gives the intrinsic forward reaction rate, as the reverse reaction rate is zero at zero
ꢀ1
conversion. The value obtained from Figure 6 is approximately 235 h at 340 °C. Assuming that
the reaction rate is first order in butane when far from equilibrium, this gives a first order rate
3
ꢀ1 ꢀ1
constant of 3.6 × 10 atm h .
24
ACS Paragon Plus Environment

Page 25 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
5
4
3
2
1
2
1
1
00
60
20
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
8
0
0
0
4
0
0
.0000
0.0005
0.0010
0.0015
0.0020
ꢀ
1
τ (molcatalyst min L )
Figure 6. Effect of space time on catalyst activity and conversion. Catalyst = 1/SiO , T = 340 °C,
2
ꢀ1
̇
V
= 4 ꢀ 100 mL min , P
= 0.32 atm, P = 0.95 atm, P = 1.27 atm.
total
butane He total
While the above calculation assumed a first order dependence between the butane partial
pressure and the reaction rate when far from equilibrium, the expected relationship when close to
equilibrium is quite different. If the partial pressures of olefin (Pbutenes) and hydrogen (P ) from
H2
the dehydrogenation of butane are assumed to be equal to one another, then the equilibrium
2
constant (K ) can be rearranged to give: PbutenesꢂPH2 = Pbutenes = K ꢂPbutane. The partial pressure
eq
eq
of butenes, and therefore the formation rate of butenes, is therefore expected to have ½ꢀorder
dependence on Pbutane regardless of temperature. The measured dependence in rate of formation
of butenes at 200 °C and 340 °C on butane partial pressure at long space time was found to
follow a ½ꢀorder power law (Figure 7), confirming that at long space time the reaction reached
equilibrium regardless of temperature.
25
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
1
1
1
60
40
20
00
80
60
1
00×
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0.0
0.5
1.0
1.5
2.0
Pbutane (atm)
Figure 7. Effect of butane partial pressure at varying temperature and space time. Catalyst =
/SiO , at T = 200 °C ( ) (values ×100) and 340 °C ( ),V = 20 ꢀ 32 mL min , P
ꢀ1
̇
1
2
= 0.32 ꢀ
2
total
butane
atm, P = 0.95 ꢀ 0 atm, P = 1.27 ꢀ 2 atm. Curves fit to ½ power law functions.
He Tot
The selectivity towards each butene isomer as a function of temperature between 200 °C
and 340 °C is given in Figure 8. Selectivity was constant as a function of time on stream, a
further indication of the stability of the active site. Selectivity also did not vary with residence
time (Figure S3), regardless of the distance from equilibrium conversion. The product
distribution did vary with temperature up to 340 °C, but only slightly, and closely matched
changes expected from equilibrium calculations. As predicted by equilibrium, the selectivity
between butenes declined as transꢀ2ꢀbutene>cisꢀ2ꢀbutene>1ꢀbutene. Equilibrium calculations
predict a higher yield of transꢀ2ꢀbutene and a lower yield of 1ꢀbutene than that observed
experimentally. In contrast to the stable selectivity with time on stream at 340 °C over three
days, the selectivity changed continually at 440 °C over a similar time period (Figure S4). This
highlights the stable nature of the active site at a temperature insufficient to cause catalyst
decomposition, compared to significant changes in the nature of the catalyst active site at the
higher temperature.
26
ACS Paragon Plus Environment

Page 27 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
0
0
0
0
0
.8
.6
.4
.2
.0
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
00
250
300
350
Temperature (°C)
Figure 8. Effect of temperature on selectivity between 200 – 340 °C. Dashed lines indicate
expected product distribution from equilibrium calculations. 1ꢀbutene ( ), transꢀ2ꢀbutene ( ),
and cisꢀ2ꢀbutene ( ). Catalyst = 1/SiO , V = 20 mL min , P
Ptotal = 1.27 atm, T = 200 – 340 °C.
ꢀ1
̇
= 0.32 atm, P = 0.95 atm,
2
total
butane
He
3
.3 Effect of Catalyst Support
The influence of support acidity on the catalytic properties of supported 1 was
investigated by comparing SiO to γꢀAl O , a more acidic support, and hydrotalciteꢀderived
2
2
3
Mg(Al)O, a less acidic support. Mg(Al)O is a slightly basic support that is inert towards reaction
21
of olefins at elevated temperatures. Experimental results show that dehydrogenation activity of
1
supported on all three oxide supports was comparable and stable over several hours (Figure 9),
with nearly identical steady state activities between. The choice of support and its acidity did not
appear to influence the activity of the catalyst. The selectivity of the reaction was similarly
unaffected.
27
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
0
0
.3
.2
.1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0.0
0
50
100
150
200
250
time (min)
Figure 9. Comparison of 1ꢀbutene formation rates when supported on SiO ( ), Mg(Al)O ( )
and Al O ( ). Catalyst = 1/support, V = 20 mL min , P = Ptotal = 1.27 atm, T = 200 °C.
2
ꢀ1
̇
2
3
total
butane
3
.4 Supported Carbonyl Complex as an Active Catalyst
In section 3.2, butane dehydrogenation conditions at temperatures above 300 °C lead to
3
1
the formation of a P MASꢀNMR signal consistent with a carbonyl complex 3/SiO . This
2
assignment was further examined by ATR FTIR. The spectra of asꢀprepared samples of 1,
1/SiO , 3, and 3/SiO were examined for evidence of carbonyl stretching frequencies (Table 1).
2
2
As expected, the ethylene complexes 1 and 1/SiO showed no carbonyl stretches. The molecular
2
ꢀ1
complex 3 showed a single intense peak at 1932 cm . Supported 3/SiO showed two signals; a
2
ꢀ1
ꢀ1
weaker signal also at 1932 cm possibly from physisorbed 3, and a stronger signal at 1945 cm
1
3
in good agreement with the reported C=O stretch of an authentic 3/SiO structure.
2
Table 1. CꢀO stretching frequencies of complexes in solution and adsorbed on SiO .
2
ꢀ1
Sample
Condition
asꢀprepared
asꢀprepared
asꢀprepared
ṽ
CO (cm )
intensity
1
ꢀ
ꢀ
ꢀ
ꢀ
1
3
/SiO2
3
1932
932
1945
strong
1
weak
strong
ꢀ
/SiO2
asꢀprepared
a
a
1
1
/SiO2
/SiO2
postꢀreaction at 200 °C
postꢀreaction at 340 °C
ꢀ
1945
strong
a
̇
V
= 20 mL/min, PButane = PTot = 1.27 atm, Duration = 1.5 hours.
Tot
28
ACS Paragon Plus Environment

Page 29 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
3
1
As the 201.2 ppm peak in the P MASꢀNMR was not observed after reaction at 200 °C,
no carbonyl peak was expected in ATR FTIR either, and none was observed. Following reaction
ꢀ1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
at 340 °C, a strong band at 1945 cm corresponded well with the asꢀprepared and reference
13
spectra, providing further evidence that the catalyst was converted from ethylene/olefin
complex at lower temperatures to carbonyl complex at higher temperatures under reaction
conditions, while maintaining the chelated pincerꢀIr structure.
If the CO complex 3/SiO were formed in situ from olefin complex 1/SiO under reaction
2
2
conditions, at temperatures where the conversion of 1/SiO → 3/SiO was complete, the activity
2
2
of either precursor would be expected to be the same. Indeed, this can be seen in Figure 10. At
31
3
40 °C, where the P MASꢀNMR spectrum showed only carbonyl complex present, the activity
of both precursors were the same. Interestingly, under identical temperature profiles, the
decomposition profiles at temperatures above 340 °C, where activity was unstable, also fully
coincided. Again, this points to identical molecular species present, even at 440 °C, which
rapidly lose activity by the same deactivation mechanism with the same deactivation kinetics.
However, at temperatures below 340 °C, the activity of the carbonyl complex was significantly
lower than that of the olefin complex. The much higher expected binding affinity of carbon
monoxide for the iridium center relative to ethylene (or other light olefin) presumably explains
this. The active catalyst species in either case is the 14ꢀelectron complex with a coordination site
9
,33
opened by ligand dissociation. Both 1/SiO and 3/SiO can lose their respective small ligands
2
2
to generate this species (Figure 11), but ethylene is much more labile than carbon monoxide.
Importantly, the composition of the catalyst appears to be uniform even at the highest
temperatures at which stable activity is observed. This is in contrast to results reported with
29
ACS Paragon Plus Environment