In situ formed "weakly ligated/labile ligand" iridium(0) nanoparticles and aggregates as catalysts for the complete hydrogenation of neat benzene at room temperature and mild pressures

Article abstract of DOI:10.1021/la101390e

"Weakly ligated/labile ligand" nanoparticles, that is nanoparticles where only weakly coordinated ligands plus the desired catalytic reactants are present, are of fundamental interest. Described herein is a catalyst system for benzene hydrogenation to cyclohexane consisting of "weakly ligated/labile ligand" Ir(0) nanoparticles and aggregates plus dry-HCl formed in situ from commercially available [(1,5-COD)IrCl]₂ plus 40 ± 1 psig (～2.7 atm) H₂ at 22 ± 0.1 °C. Multiple control and other experiments reveal the following points: (i) that this catalyst system is quite active with a TOF (turnover frequency) of 25 h^-1 and TTO (total turnovers) of 5250; (ii) that the BF ₄^- and PF₆^- iridium salt precursors, [(1,5-COD)Ir(CH₃CN)₂]BF₄ and [(1,5-COD)Ir(CH₃CN)₂]PF₆, yield inferior catalysts; (iii) that iridium black with or without added, preformed HCl cannot achieve the TOF of 25 h^-1 of the in situ formed Ir(0)/dry-HCl catalyst. However and importantly, CS₂ poisoning experiments yield the same activity per active iridium atom for both the Ir(0)/dry-HCl and Ir black/no-HCl catalysts (12.5 h^-1 Ir^1-), but reveal that the Ir(0)/dry-HCl system is 10-fold more dispersed compared to the Ir(0) black catalyst. The simple but important and key result is that "weakly ligated/labile ligand" Ir(0) nanoparticles and aggregates have been made in situ as demonstrated by the fact that they have identical, per exposed Ir(0) activity within experimental error to Ir(0) black and that they have no possible ligands other than those desired for the catalysis (benzene and H₂) plus the at best poor ligand HCl. As expected, the in situ catalyst is poorly stabilized, exhibiting only 60% of its initial activity in a second run of benzene hydrogenation and resulting in bulk metal precipitation. However, stabilization of the Ir(0) nanoparticles with a ca. 2-fold higher catalytic activity and somewhat longer lifetime for the complete hydrogenation of benzene was accomplished by supporting the Ir(0) nanoparticles onto zeolite-Y (TOF of 47h^-1 and 8600 TTOunder otherwise identical conditions). Also reported is the interesting result that Cl^- (present as Proton Sponge ·H⁺Cl^-) completely poisons benzene hydrogenation catalysis, but not the easier cyclohexene hydrogenation catalysis under otherwise the same conditions, results that suggest different active sites for these ostensibly related hydrogenation reaction. The results suggest that synthetic routes to "weakly ligated/labile ligand" nanoparticles exhibiting improved catalytic performance is an important goal worthy of additional effort. 2010 American Chemical Society.

Full text of DOI:10.1021/la101390e

pubs.acs.org/Langmuir
© 2010 American Chemical Society
In Situ Formed “Weakly Ligated/Labile Ligand” Iridium(0) Nanoparticles
and Aggregates as Catalysts for the Complete Hydrogenation of Neat
Benzene at Room Temperature and Mild Pressures
†
‡
‡
,†
€
Ercan Bayram, Mehmet Zahmakıran, Saim Ozkar, and Richard G. Finke*
†Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523, and
‡Department of Chemistry, Middle East Technical University, 06531 Ankara, Turkey
Received April 7, 2010. Revised Manuscript Received May 19, 2010
“Weakly ligated/labile ligand” nanoparticles, that is nanoparticles where only weakly coordinated ligands plus the
desired catalytic reactants are present, are of fundamental interest. Described herein is a catalyst system for benzene
hydrogenation to cyclohexane consisting of “weakly ligated/labile ligand” Ir(0) nanoparticles and aggregates plus dry-
HCl formed in situ from commercially available [(1,5-COD)IrCl]2 plus 40 ( 1 psig (∼2.7 atm) H2 at 22 ( 0.1 °C.
Multiple control and other experiments reveal the following points: (i) that this catalyst system is quite active with a TOF
(turnover frequency) of 25 h-1 and TTO (total turnovers) of 5250; (ii) that the BF4- and PF6- iridium salt precursors,
[(1,5-COD)Ir(CH3CN)2]BF4 and [(1,5-COD)Ir(CH3CN)2]PF6, yield inferior catalysts; (iii) that iridium black with or
without added, preformed HCl cannot achieve the TOF of 25 h-1 of the in situ formed Ir(0)/dry-HCl catalyst. However
and importantly, CS2 poisoning experiments yield the same activity per active iridium atom for both the Ir(0)/dry-HCl
and Ir black/no-HCl catalysts (12.5 h-1 Ir1-), but reveal that the Ir(0)/dry-HCl system is 10-fold more dispersed compared
to the Ir(0) black catalyst. The simple but important and key result is that “weakly ligated/labile ligand” Ir(0)
nanoparticles and aggregates have been made in situ as demonstrated by the fact that they have identical, per exposed
Ir(0) activity within experimental error to Ir(0) black and that they have no possible ligands other than those desired for
the catalysis (benzene and H2) plus the at best poor ligand HCl. As expected, the in situ catalyst is poorly stabilized,
exhibiting only 60% of its initial activity in a second run of benzene hydrogenation and resulting in bulk metal
precipitation. However, stabilization of the Ir(0) nanoparticles with a ca. 2-fold higher catalytic activity and somewhat
longer lifetime for the complete hydrogenation of benzene was accomplished by supporting the Ir(0) nanoparticles onto
zeolite-Y (TOF of 47 h-1 and 8600 TTO under otherwise identical conditions). Also reported is the interesting result that
Cl- (present as Proton Sponge HþCl-) completely poisons benzene hydrogenation catalysis, but not the easier
3
cyclohexene hydrogenation catalysis under otherwise the same conditions, results that suggest different active sites for
these ostensibly related hydrogenation reaction. The results suggest that synthetic routes to “weakly ligated/labile
ligand” nanoparticles exhibiting improved catalytic performance is an important goal worthy of additional effort.
Introduction
catalyst system for neat-acetone hydrogenation consisting of
iridium(0) plus dry-HCl formed in situ from the H2 reduction
of commercially available 1,5-cyclooctadienechloroiridium(I) di-
mer, [(1,5-COD)IrCl]2. This Ir(0)/dry-HCl catalyst system is a
superior catalyst for acetone hydrogenation at low temperature
and pressureinterms of its activity (22 ( 0.1 °C and 40( 1 psig H2
“Weakly ligated/labile ligand” nanoparticles1-8 are of interest
to the nanoparticle catalysis community since their surfaces
should be readily available for catalysis following only the
dissociation of weakly coordinated ligands9 or solvent.3,10,11 In
a previous publication,11 we reported an easily prepared, highly
active and selective “weakly ligated/labile ligand” nanoparticle
(5) More recently, Evanoff and co-workers5a reported “essentially naked” Ag
nanoparticles with
a large H2O solvation cage which encapsulates the Ag
nanoparticles. In a later paper,5b the stability of the Ag nanoparticles were
attributed to electrostatic repulsion afforded by the low ionic strength solution
(Agþ and OH-). (a) Chumanov, G.; Evanoff, D. D. J. Phys. Chem. B 2004, 108,
13948. (b) Evanoff, D. D.; Chumanov, G. Chem. Phys. Chem. 2005, 6, 1221.
(6) Raabe, I.; Krossing, I. Angew. Chem., Int. Ed. 2004, 43, 2066.
*Corresponding author. E-mail: rfinke@lamar.colostate.edu.
(1) (a) The nomenclature “nanoparticle” is used herein rather than “nanocluster”.
Although as Schmid and Fenske have noted there is at present no “sharp
discrimination”1b between these two terms, there is an evolving nomenclature
where nanocluster is preferred when the composition and the structure of the
species are precisely defined whereas nanoparticle is more appropriate when a size
distribution is present.1b However, reports exist where “nanoparticle” has been
used to designate very precisely defined species.1c,d (b) Schmid, G.; Fenske, D.
Philos. Trans. R. Soc. A 2010, 368, 1207. (c) Jadzinsky, P. D.; Calero, G.; Ackerson,
C. J.; Bushnell, D. A.; Kornberg, R. D. Science 2007, 318, 430. (d) Heaven, M. W.; Dass,
A.; White, P. S.; Holt, K. M.; Murray, R. W. J. Am. Chem. Soc. 2008, 130, 3754.
(2) Ott, L. S.; Finke, R. G. Coord. Chem. Rev. 2007, 251, 1075; see especially
sections 4.1 and 4.2 and references therein to the work of others on what we have termed
the “weakly ligated/labile ligand” nanoparticles problem.
€
(7) Mondloch, J. E.; Ozkar, S.; Finke, R. G. Manuscript in preparation. As part of
this mini-review, a fuller survey of the literature related to “naked nanoparticles”, “solvent-
only stabilized nanoparticles”,2,3 and “weakly ligated/labile ligand nanoparticles”3 is
planned.
(8) (a) Meisel and coworkers8b find that a small amount of Agþ is present on the
surface of Ag nanoparticles, which in turn are negatively charged overall via
adsorbed OH- anions. (b) Merga, G.; Wilson, R.; Lynn, G.; Milosavljevic, B. H.;
Meisel, D. J. Phys. Chem. C 2007, 111, 12220.
(9) Strauss, S. H. Chem. Rev. 1993, 93, 927. Of interest here is the parallel between
the development of the “weakly coordinating/labile ligand” nanocluster problem to the
earlier “weakly coordinating anion” problem reviewed by S. H. Strauss.
(10) Noteworthy here, however, is that even metal-alkane bond dissociation
energies are ∼10 ( 3 kcal/mol, so that such putatively “naked” nanoparticles in
even an alkane solvent would really be better labeled “weakly ligated/labile ligand”
nanoparticles.10a-c (a) Yang, G. K.; Peters, K. S.; Vaida, V. Chem. Phys. Lett.
1986, 125, 566. (b) Simon, J. D.; Xie, X. J. Phys. Chem. 1989, 93, 291. (c) Morse, J. M.,
Jr.; Parker, G. H.; Burkey, T. J. Organometallics 1989, 8, 2471.
(3) Ott, L. S.; Finke, R. G. Inorg. Chem. 2006, 45, 8382; see the references therein
to putatively “solvent-only stabilized nanoparticles”. Also of interest is footnote 32
therein, reproduced here: “The “naked nanoparticle problem”, perhaps more accurately
called the “ligand-labile nanoparticle problem”, is the need for efficient, high yield
syntheses of metastable nanoparticles that have relatively easily removed ligands so that
they can be used for low-temperature syntheses of novel heterogeneous catalysts or other
applications of clean, ideally naked-surface nanoparticles.”
€
€
(4) Korall, B.; Bonneman, H. Angew. Chem., Int. Ed. 1992, 31, 1490.
(11) Ozkar, S.; Finke, R. G. J. Am. Chem. Soc. 2005, 127, 4800.
Langmuir 2010, 26(14), 12455–12464
Published on Web 06/10/2010
DOI: 10.1021/la101390e 12455

Article
Bayram et al.
pressure with an estimated TOF of 1.9 s-1), selectivity (95%
2-propanol, 5% diisopropyl ether), and total catalyst lifetime
(16400 TTOs). These results suggested that it would be of interest
to employ this easily formed, highly active Ir(0)/dry-HCl catalyst
system for the solventless reduction of other challenging hydro-
genations such as benzene hydrogenation.
forms Ir(0)/dry-HCl as highly active catalyst for neat-benzene
hydrogenation with 100% conversion to cyclohexane. The in situ
coproduction of dry-HCl is relevant to the catalytic activity as
demonstrated by multiple control experiments such as (i) scaven-
ging in situ formed Hþ via Proton Sponge, and (ii) comparing
the catalytic activity of two other iridium precursors (i.e., [(1,5-
COD)Ir(CH3CN)2]BF4 and [(1,5-COD)Ir(CH3CN)2]PF6) under
otherwise identical conditions. A comparison to iridium black,
employed as a benzene hydrogenation catalyst under otherwise
identical conditions, reveals a 10-fold lower activity (TOF of
2.5 h-1 for iridium black22 vs 25 h-1 for Ir(0)/dry-HCl), but CS2
poisoning studies indicate that ca. 2% of total iridium atoms are
catalytically active when starting with [(1,5-COD)IrCl]2, but only
0.2% when starting with iridium black (both assuming a 1:1 CS2:
Ir poisoning stoichiometry). Hence, the catalytic activity in each
case is the same per exposed Ir(0)—results that also demonstrate
the value of such active-site determination studies via quantitative
poisoning experiments. The high activity and good lifetime, 25 e
TOF e 1250 h-1 and 5250 e TTO e 262 500, of the Ir(0)/dry-HCl
catalyst system is consistent with the expected “weakly ligated/
labile ligand” nature of the Ir(0) nanoparticles and aggregates
where only the possible ligands are benzene, H2 (i.e., and
hydrides), and the weak to nonligand HCl.
The complete hydrogenation of aromatics is of interest12,13 for
a number of reasons, including the growing demand for cleaner-
burning, low-aromatic-content diesel fuels (that thereby minimize
powerful carcinogens14 in diesel exhaust particles that contribute
to asthma or nasal allergies15). Benzene hydrogenation to cyclo-
hexane is also important, cyclohexane being a key intermediate in
the production of the nylon precursor adipic acid.16 However,
benzene hydrogenation is notably difficult compared to simple
olefin hydrogenation17 due to the loss of the resonance stabilization
energy during benzene reduction and prior to the rate-determining
transition state.18 Consequently, benzene hydrogenation has his-
torically required higher temperatures and H2 pressures (i.e.,
g100 °C and ∼50 atm H2).19 Confirming this, only 17 studies
that we have been able to find report the complete hydrogenation
of benzene to cyclohexane at temperatures e25 °C (see Table SI-1
in the Supporting Information).20 Among those 17 studies, only
6 articles20b,c,f,j,o,q report neat benzene hydrogenation, the condi-
tions that will be employed herein. The reported neat-benzene
hydrogenation catalyst systems also tend to involve multistep,
sometime laborious catalyst preparation procedures.
Experimental Section
Materials and General Considerations. All commercially
obtainedcompounds were usedas receivedunlessindicatedother-
wise. Benzene (anhydrous, 99.8%), diethyl ether (g99.9%),
CH2Cl2 (anhydrous, g99.8%), Proton Sponge (1,8-bis(dimethyl-
amino)naphthalene, 99%), and CS2 (anhydrous, g99%) were
purchased from Aldrich Chemicals and transferred into a Vacuum
Atmospheres nitrogen atmosphere drybox. CD2Cl2 (Cambridge
Isotope Laboratories) was purchased in 1 mL glass ampules which
were then transferred into the drybox where NMR sample prepa-
rations were performed. [(1,5-COD)IrCl]2 and iridium black
(99.9%, 30-60 m2/gr) were purchased from Strem Chemicals.
Sodium zeolite-Y (Na56Y, Si/Al = 2.5) was purchased from
Zeolyst Inc. and slurried with 0.1 M NaCl solution to remove
cation defect sites, washed until free of chloride and calcined in dry
oxygen at 500 °C for 12 h before use. Unless otherwise stated all
studies were performed under oxygen- and moisture-free condi-
tions using a VacuumAtmospheres N2 drybox (always <5 ppm of
O2, and typically <1 ppm of O2, as monitored by a Vacuum
Atmospheres O2 level monitor). 1H NMR spectra were taken on a
Hence, still of interest is the complete hydrogenation of neat
benzene (i.e., solventless, green conditions21) via a highly active,
long-lived, readily available catalyst that operates under mild
conditions, (e25 °C and e10 atm H2 pressure). Also desirable is
clear documentation of the catalyst turnover frequency and total
turnover number, as well as determination of the important, but
too infrequently measured, percentage of active catalyst sites via
catalyst poisoning studies (e.g., with CS2).18
Herein, we report that in situ reduction of [(1,5-COD)IrCl]2
under 40 ( 1 psig (∼2.7 atm) initial H2 pressure at 22 ( 0.1 °C
(12) Widegren, J. A.; Finke, R. G. Inorg. Chem. 2002, 41, 1558.
(13) Schulz, J.; Patin, H.; Roucoux, A. Chem. Rev. 2002, 102, 3757.
(14) Enya, T.; Suzuki, H.; Watanabe, T.; Hirayama, T.; Hisamatsu, Y. Environ.
Sci. Technol. 1997, 31, 2772.
(15) (a) Casillas, A. M.; Hiura, T.; Li, N.; Nel, A. E. Ann. Allergy, Asthma,
Immunol. 1999, 83, 624. (b) Nal, A.; Diaz-Sanchez, D.; Ng, D.; Hiura, T.; Saxon, A. J.
J. Allergy Clin. Immunol. 1998, 102, 539.
1
(16) Arpe, H. J.; Weissermel, K. Industrial Organic Chemistry, 4th Ed.; Wiley-
VCH: New York, 2003.
Varian INOVA-300 instrument with 300.115 MHz for H. Gas
Chromatography (GC) measurements were performed using a
Hewlett-Packard 5890 series II GC with an MSD 5970 B. The GC
was equipped with a 30 m (0.25 mm i.d., 0,25 μm film thickness)
Supelco SPB-1 column and with an ionizing voltage of 70 eV. The
GC parameters were as follows: initial temperature, 50 °C; initial
time, 3 min; solvent delay, 2 min; temperature ramp, 10 °C/min;
final temperature, 270 °C; final time, 5 min; injector port tempera-
ture, 280 °C; detector temperature, 290 °C; injection volume,
0.2 μL. The iridium content of the Ir(0)/zeolite-Y sample was
determined by ICP-OES analysis (Leeman-DRE) after each
(17) (a) March, J. Advanced Organic Chemistry: Reactions, Mechanisms, and
Structure, 4th ed.; Wiley-Interscience: New York, 1992;(b) Fessenden, R. J.; Fessenden,
J. S. Organic Chemistry, 5th ed.; Brooks/Cole Publishing Company: Pacific Grove, CA,
1993.
(18) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003, 191, 187.
(19) Augustine, R. L. Heterogeneous Catalysis for the Synthetic Chemistry;
Marcel Dekker: New York, 1996.
(20) (a) Januszklewicz, K. R.; Alper, H. Organometallics 1983, 2, 1055.
(b) Seeberger, M. H.; Jones, R. A. J. Chem. Soc., Chem. Commun. 1985, 6, 373.
(c) Duan, Z.; Sylwester, A. P.; Hampden-Smith, M. J. Chem. Mater. 1992, 4, 1146.
(d) Schulz, J.; Patin, H.; Roucoux, A. Chem. Commun. 1999, 535. (e) Schulz, J.; Patin,
H.; Roucoux, A. Chem. Eur. J. 2000, 6, 618. (f) Nicholas, J. P.; Ahn, H.; Marks, T. J.
J. Am. Chem. Soc. 2003, 125, 4325. (g) Schulz, J.; Patin, H.; Roucoux, A. Adv. Synth.
Catal. 2003, 345, 222. (h) Mevellec, V.; Ramirez, E.; Phillippot, K.; Chaudret, B.;
Roucoux, A. Adv. Synth. Cat. 2004, 346, 72. (i) Park, I. S.; Kwon, M. S.; Kim, N.; Lee,
J. S.; Kang, K. Y.; Park, J. Chem. Commun. 2005, 5667. (j) Zhang, J.; Xie, Z.; Liu, Z.;
Wu, W.; Han, B.; Huang, J.; Jiang, T. Catal. Lett. 2005, 103, 59. (k) Roucoux, A.;
Phillippot, K.; Payen, E.; Granger, P.; Dujardin, C.; Nowicki, A.; Mevellec, V. New
J. Chem. 2006, 30, 1214. (l) Nowicki, A.; Zhong, Y.; Leger, B.; Rolland, J. P.; Bricout,
H.; Monflier, E.; Roucoux, A. Chem. Commun. 2006, 296. (m) Nowicki, A.; Le
Boulaire, V.; Roucoux, A. Adv. Synt. Catal. 2007, 349, 2326. (n) Park, I. S.; Kwon,
M. S.; Lee, J. S.; Kang, K. Y.; Park, J. Adv. Synth. Catal. 2007, 349, 2039.
(22) Attempts to mimic the in situ Ir(0)/dry- HCl catalyst system, by employing
iridium black as the catalyst while adding preformed dry-HCl in benzene, failed.
Specifically, independently prepared HCl(g), generated by reacting NaCl with
H2SO4, was bubbled into 10 mL of fresh benzene (at ∼20 °C and ∼1 atm for
∼30 min) to obtain saturated benzene/dry-HCl stock solution (the solubility of
HCl in benzene is 0.039 mol HCl per mol of benzene at 20 °C22a). This stock
solution was then diluted with fresh benzene to obtain the same amount of dry-HCl
as is present when starting with [(1,5-COD)IrCl]2, namely 0.052 mmol HCl was
added to the Ir(0) black catalyst. However, due to the necessary experimental
protocol involving flushing the apparatus with H2 to begin the benzene hydro-
genation reaction, the HCl was partially removed from the apparatus, resulting in
irreproducible results with a TOF of 0 to ∼10 h-1. Hence, these experiments were
abandoned. (a) Hydrogen Chloride. Kirk-Othmer Encyclopedia of Chemical
Technology; Wiley & Sons: New York, 2004; Vol. 13, pp 808-837.
€
(o) Zahmakıran, M.; Ozkar, S. Langmuir 2008, 24, 7065. (p) Hubert, C.; Denicourt-Nowicki,
ꢀ
A.; Guegan, J.-P.; Roucoux, A. Dalton Trans. 2009, 7356. (q) Pan, H.-B.; Wai, C. M.
J. Phys. Chem. C 2009, 113, 19782.
(21) Poliakoff, M.; Fitzpatrick, J. M.; Farren, T. R.; Anastas, P. T. Science 2002,
297, 807.
12456 DOI: 10.1021/la101390e
Langmuir 2010, 26(14), 12455–12464

Bayram et al.
Article
sample was completely dissolved in the mixture of HNO3:HCl
with a 1:3 ratio (theoretical Ir: 1%, found Ir: 1%). Powder X-ray
diffraction (XRD) patterns of Ir(0)/zeolite-Y samples were re-
corded with a Rigaku X-ray diffractometer (Model, Miniflex)
with Cu KR (30 kV, 15 mA, λ = 1.54051 A) radiation at room
temperature.
immediately followedbyclosing and evacuatingthe antechamber.
The binding energies were compared to the literature values.26
Transmission Electron Microscopy (TEM) Analyses. i. Ir-
(0)/Dry-HCl System. The final form of the hydrogenation
reaction solution is not homogeneous; instead, a brown solution
with black precipitate forms, so that the TEM images obtained as
follows cannot be 100% representative of the whole medium.
However, to obtain an idea about the particles present in solution,
TEM samples were harvested after 4 h of a standard conditions
benzene hydrogenation, and then at the end of this same reaction
(i.e., after 8.7( 0.1 h). TEM samples were preparedas follows:the
F-P bottle was detached from the hydrogenation line after 4 h of
the reaction via its quick connects, vented, and brought back into
the drybox. The solution was transferred with a disposable
polyethylene pipet into a clean, 5 mL glass vial. TEM samples
were prepared by dipping the SiO-TEM grid into the solution for
3 s. The F-P bottlewasthenresealedafter transferringthesolution
back into the Pyrex culture tube in it, brought out of the drybox,
reconnected to the line and pressurized to 40 ( 1 psig with H2. At
this point, collection of pressure versus time data was continued
(ignoring the ∼1 h gap required for the procedure). The same
TEM preparation procedure was repeated at the end of the
reaction, that is, after 8.7 ( 0.1 h. The resultant TEM grids were
then placed separately in a screw-capped glass vials, were sealed
and then were shipped to Clemson University for TEM analysis
via the expert assistance of Dr. JoAn Hudson and her staff. TEM
images were obtained with a Hitachi H7000 instrument operating
at 120 keV.
Hydrogenation Apparatus. All the hydrogenation reactions
were carried out on the previously described, custom-built pres-
surized hydrogenation apparatus which allows monitoring the H2
pressure decrease accompanying hydrogenations ((0.01 psig)
via a computer interface (LabView ver. 8.2) in real-time.23-25
A
Fischer-Porter (F-P) bottle was connected via its Swagelock TFE-
sealed Quick Connects to a hydrogenation line and Omega D1512
10 V A/D converter with RS-232 connection to a computer. The
hydrogen (>99.5) was purchased from Airgas and scrubbed via a
Trigonmoisturetrap and a TrigonTechnologies oxygen/moisture
trap to remove O2 and H2O followed by a Trigon Technologies
high capacity indicating oxygen trap.
Standard Conditions Procedure for the Complete Benzene
HydrogenationExperimentsStartingwith[(1,5-COD)IrCl]2.
To start, 17.5 mg (0.052 mmol in iridium) of the [(1,5-COD)IrCl]2
precatalyst was weighed in a 2 dram glass vial and then dissolved
in 1.0 mL (11.2 mmol) of benzene added via a 5.0 mL gastight
syringe; a clear, orange solution 52 mM in iridium resulted. The
solution was then transferred via a disposable polyethylene pipet
into a new 22 ꢀ 175 mm Pyrex culture tube containing a new
5/16 ꢀ 5/8 in. Teflon-coated stir bar. The culture tube was then
sealed inside the F-P pressure bottle and brought outside the
drybox. The F-P bottle was placed into a constant temperature
circulating bath at 22 ( 0.1 °C, and attached via Swagelock TFE-
sealed Quick-Connects to the hydrogenation line (which had
already been evacuated for at least 30 min to remove any trace
oxygen and water present, then refilled with purified H2 at 40 ( 1
psig). Stirringwas started (at 600 rpm) and the F-P bottle was then
purged 15 times with hydrogen (15 s per purge) and t = 0 was
started. When the hydrogen uptake ceased, the F-P bottle was
disconnected from the hydrogenation line, remaining H2 pressure
was released, and transferred back into the drybox, where a 9 in.
glass Pasteur pipet was used to withdraw a ca. 0.05 mL aliquot
from the culture tube. This aliquot was then added to 1 mL of
CD2Cl2 in an individual glass ampule, mixed under agitation
using the Pasteurpipet, and thentransferred intoanNMR sample
tube which was subsequently brought out of the drybox after
ii. Ir(0)/Zeolite-Y System. When the otherwise standard
conditions benzene hydrogenation reaction was complete (now
after 4.7 ( 0.1 h for this faster catalyst), the F-P bottle was
detached from the hydrogenation line via its quick connects,
vented, and brought back into the drybox. The Pyrex culture tube
inside F-P bottle was transferred into a Schlenk tube and then
brought to dryness under vacuum, yielding a gray solid. A small
amount of the powdered sample was placed on a copper grid.
Samples were examined at magnification between 100 and 400 K
by using JEM-2010F microscope (JEOL) operating at 200 keV.
Control Experiments of Benzene Hydrogenation Starting
with [(1,5-COD)IrCl]2 Plus 1.0 or 0.02 equiv of Proton
Sponge per equiv of Ir. Two separate control experiments were
performed starting with 17.5 mg (0.052 mmol in iridium) [(1,5-
COD)IrCl]2 in 1.0 mL (11.2 mmol) of benzene, but now plus 1.0
or 0.02 equiv of Proton Sponge (0.052 or 1.04 ꢀ 10-4 mmol,
respectively) per equiv of iridium under the Standard Conditions
of 22 ( 0.1 °C and 40 ( 1 psig initial H2 pressure. The results are
described in the Results and Discussion section.
1
sealing for H NMR investigation. The NMR analysis showed
the complete reduction of benzene (7.26 ppm, m) to cyclohexane
(1.44 ppm, s). None of the partially hydrogenated benzene
reduction product, cyclohexene, was observed.
Control Experiment of Cyclohexene Hydrogenation un-
der Standard Conditions Starting with [(1,5-COD)IrCl]2 in
X-ray Photoelectron Spectroscopy (XPS) Analyses. At
the end of the benzene hydrogenation reaction, the F-P bottle was
vented and the solution was brought to dryness under vacuum to
yield a gray solid. Some of that solid was then coated onto a XPS
sample holder and subsequently sealed in a desiccator under the
N2 atmosphere of the drybox. The desiccator-enclosed sample
was then removed from the drybox for analysis via a Physical
Electronics (PHI) Model 5800 XPS system equipped with a mono-
chromator (Al KR source, hυ = 1486.8 eV; system pressure e5 ꢀ
10-9 Torr) and a hemispherical analyzer to detect the ejected
photons; XPS analysis wasaccomplished with the expert assistance
of Pat McCurdy at the Central Instrument Facility of Colorado
State University, Department of Chemistry. To minimize expo-
sure of the sample to air, the desiccator was opened next to the
instrument antechamber and the sample holder was mounted
the presence of 1 equiv of Proton Sponge per equiv of Ir.
A
cyclohexene control hydrogenation experiment was performed
starting with 17.5 mg (0.052 mmol in iridium) [(1,5-COD)IrCl]2 in
1.0 mL (11.2 mmol) of benzene and 0.5 mL (4.9 mmol) cyclohexene
plus 1 equiv of Proton Sponge (0.052 mmol) under the standard
conditions of 22 ( 0.1 °C and 40 ( 1 psig initial H2 pressure. The
results are described in the Results and Discussion section.
Synthesis and Characterization of [(1,5-COD)Ir(CH3-
CN)2][X] (X: BF4 and PF6). These complexes were synthesized
and characterized according to the literature procedure.27 2.02 g
of [(1,5-COD)IrCl]2 (3.01 mmol) was dissolved in 42 mL of
CH2Cl2. Upon the addition of 10 mL of CH3CN, the dark red
solution turned bright yellow. Then, 6.01 mmol of AgX (where
X represents BF4 for [(1,5-COD)Ir(CH3CN)2]BF4 and PF6 for
[(1,5-COD)Ir(CH3CN)2]PF6) was added to precipitate AgCl. The
obtained solution was filtered into 200 mL of diethyl ether; yellow
microcrystals resulted of [(1,5-COD)Ir(CH3CN)2][X], where X:
BF4 or PF6. The resultant solution was then filtered, washed with
(23) Lin, Y.; Finke, R. G. J. Am. Chem. Soc. 1994, 116, 8335.
(24) (a) Ozkar, S.; Finke, R. G. J. Am. Chem. Soc. 2002, 124, 5796. (b) Ozkar, S.;
Finke, R. G. Langmuir 2002, 18, 7653.
€
€
(25) Lin, Y.; Finke, R. G. Inorg. Chem. 1994, 33, 4891.
(26) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K. D. Handbook of
X-ray Photoelectron Spectroscopy: Physical Electronics Inc.: Eden Praire, MN, 1995.
(27) Day, V. W.; Klemperer, W. G.; Main, D. G. Inorg. Chem. 1990, 29, 2345.
Langmuir 2010, 26(14), 12455–12464
DOI: 10.1021/la101390e 12457

Article
Bayram et al.
2 ꢀ 10 mL diethyl ether, and dried under vacuum for 12 h (80%
yield of [(1,5-COD)Ir(CH3CN)2]BF4 as a yellow solid; 81% yield
of [(1,5-COD)Ir(CH3CN)2]PF6 as a yellow solid). 1H NMR data
were compared to the literature27 (in parentheses) to confirm
the identity and purity of the complexes: 4.3 ppm, s (4.27 ppm, s);
2.5 ppm, s (2.53 ppm, s); 2.3 ppm, m (2.29 ppm, m); 1.8 ppm, m
(1.78 ppm, m).
to dryness under vacuum to yield a gray solid for both the Ir(0)/
dry-HCl and Ir(0)/zeolite-Y catalysts. In two separate experi-
ments these gray residues were redispersed in 1.0 mL (11.2 mmol)
of benzene, repressurized to 40 ( 1 psig with H2, and purged
15 times (15 s per purge), and the collectionof pressure versus time
data was restarted. The results are described in the Results and
Discussion section.
Control Experiments for Standard Conditions Benzene
Hydrogenation Starting with [(1,5-COD)Ir(CH3CN)2]BF4
and [(1,5-COD)Ir(CH3CN)2]PF6. Two separate standard
conditions benzene hydrogenation experiments were performed,
but now using the two different iridium precursors [(1,5-COD)Ir-
(CH3CN)2]BF4 and [(1,5-COD)Ir(CH3CN)2]PF6 in place of[(1,5-
COD)IrCl]2. The results are described in the Results and Discus-
sion section.
Control Experiments with Commercial Iridium Black.
Control experiments were performed by following the Standard
Conditions, but now using commercial iridium black in place of
[(1,5-COD)IrCl]2. The results are described in the Results and
Discussion section.
Catalyst Lifetime Experiments. Catalyst lifetime experi-
ments were performed by following the Standard Conditions
for the Ir(0)/dry-HCl and for Ir(0)/zeolite-Y catalyst, but with
(i) 1.2 mg (3.6 μmol in iridium) of [(1,5-COD)IrCl]2 precursor in
3.0 mL (33.6 mmol) of benzene corresponding to a maximum
possible 9333 TTOs, and (ii) 1.75 mg (5.25 μmol in iridium) of [(1,5-
COD)IrCl]2 precursor and 48 mg zeolite-Y in 5.0 mL (56 mmol) of
benzene corresponding to a maximum possible 10667 TTOs. The
results are described in the Results and Discussion section.
CS2 Poisoning Experiments. These experiments were each
started as if they were a Standard Conditions benzene hydrogena-
tion experiment. For each poisoning trial below (i.e., for each
different equivs of CS2 added), a fresh standard conditions
benzene hydrogenation was started and allowed to proceed for
5 h while pressure versus time data were collected. The F-P bottle
was then vented, taken into the drybox, and opened, and the
desired amount of CS2 was added. After the addition of CS2, the
F-P bottle was resealed, brought out of the drybox, reconnected
to the line and pressurized to 40 ( 1 psig with H2. At this point,
collection of pressure versus time data was continued (ignoring
the ∼1 h gap required for the procedure). For example, for the
Ir(0)/dry-HCl system, 0.1 equiv CS2 with respect to the total
iridium present was added and found to poison the previous
catalytic activity completely. In the next experiment, the amount
ofCS2 added was loweredto0.01 equiv; this inturn did not poison
the catalyst completely. The third experiment employed 0.03
equiv of CS2, and that did completely cease the catalytic activity.
Hence, ca. 0.02 equiv of CS2 was deemed sufficient to poison
completely the catalytic activity.
Curve-Fitting Trials of the Hydrogen Uptake Data.
Curve-fitting trials for concentration vs time data to the pre-
viously established 2-step28 or 4-step29 nanoparticle and agglom-
erated nanoparticle formation mechanisms were performed using
nonlinear least-squares fitting in Origin ver. 7.0 or MacKinetics,
respectively.
Procedure for the Complete Hydrogenation of Neat Ben-
zene Starting with [(1,5-COD)IrCl]2 plus Zeolite-Y (Na56Y,
Si/Al: 2.5). In a 2 dram glass vial, 17.5 mg (0.052 mmol in
iridium) of the precatalyst [(1,5-COD)IrCl]2 was weighed and
then dissolved in 1.0 mL (11.2 mmol) of benzene added via a
5.0 mL gastight syringe to yield a clear, orange solution; this solution
was then transferred via a disposable polyethylene pipet into a new
22 ꢀ 175 mm Pyrex culture tube containing a new 5/16 ꢀ 5/8 in.
Teflon-coated stir bar. Next, 480 mg zeolite-Y (corresponding to
1.0% wt Ir on zeolite-Y) was added into this solution and mixed for
half an hour. Then, the hydrogenation experiment was performed in
the same way as described previously in the section Standard
Conditions Procedure for the Complete Benzene Hydrogenation
Experiments Starting with [(1,5-COD)IrCl]2.
Results and Discussion
Product Characterization and Balanced Stoichiometry
for the Complete Hydrogenation of Benzene Starting with
[(1,5-COD)IrCl]2 as Precatalyst under Standard Condi-
tions. Figure 1 shows a typical benzene loss vs time plot for
the complete hydrogenation of neat benzene (1.0 mL, 11.2 mmol)
starting with the [(1,5-COD)IrCl]2 (17.5 mg, 0.052 mmol Ir)
precatalyst at 22 ( 0.1 °C and 40 ( 1 psig (∼2.7 atm) initial H2,
which we will refer to as “standard conditions”. The initial orange
color of the solution due to the [(1,5-COD)IrCl]2 precatalyst dark-
ens within 30 min of the initiation of the reaction by the addition of
H2 pressure, and that dark color remains throughout the entire
reaction. Complete hydrogenation of benzene was achieved after
8.7 ( 0.1 h as indicated by the cessation of H2 uptake and the
production of cyclohexane (100%) as the sole product and con-
firmed by 1H NMR as detailed in the Experimental Section. This
particular kinetic curve given in Figure 1 is not well fit by either our
2- or 4-step mechanisms of nanoparticle and agglomerated nano-
particle formation28,29 details of which are given in the Supporting
Information for the interested reader (Figures SI-1 and SI-2).
The balanced stoichiometry of the formation of Ir(0)/dry-HCl
from [(1,5-COD)IrCl]2 and concomitant benzene hydrogenation
under standard conditions is given in Scheme 1: the reduction of
1 equiv of [(1,5-COD)IrCl]2 requires 5 equiv of H2 and produces
2/n equiv of Ir(0)n plus 2 equiv of HCl by mass balance.24b The
evolution of 2 equiv of cyclooctane was confirmed by GC. The
formation of Ir(0) was confirmed by XPS30 on the vacuum-dried
Control Experiment for Benzene Hydrogenation Showing
That Transferring to the Drybox, Opening the F-P Bottle,
and Then Restarting the Reaction Does Not Cause a Detec-
table Loss of Activity. This experiment was performed to ensure
thatthe lossof activity seeninthe CS2 poisoning experimentisdue
to the added CS2, and not some other aspect of the necessary
manipulations cited above in the CS2 poisoning experiments. This
control experiment was started as if it were a standard conditions
benzene hydrogenation experiment. Pressure versus time data
were collected for 5 h. Then the F-P bottle was vented, taken into
the drybox, and opened. The F-P bottle was then resealed,
brought out of the drybox, reconnected to the line and pressurized
to 40 ( 1 psig with H2. At this point, collection of pressure versus
time data was continued (again ignoring the ∼1 h gap required for
the procedure). No detectable loss of activity due to the above
transfer procedure was observed in this control experiment.
Catalyst Redispersibility Experiments. These experiments
were each started as if they were a standard conditions benzene
hydrogenation experiment. After complete hydrogenation of
benzene, the F-P bottle was vented and the solution was brought
(28) (a) Watzky, M. A.; Finke, R. G. J. Am. Chem. Soc. 1997, 119, 10382 and
references therein. (b) Watzky, M. A.; Finke, R. G. Chem. Mater. 1997, 9, 3083.
(c) Aiken, J. D., III; Finke, R. G. J. Am. Chem. Soc. 1998, 120, 9545 and references
€
therein. (d) Widegren, J. A.; Aiken, J. D., III; Ozkar, S.; Finke, R. G. Chem. Mater.
2001, 13, 312 and references therein.
(29) (a) Besson, C.; Finney, E. E.; Finke, R. G. J. Am. Chem. Soc. 2005, 127,
8179. (b) Besson, C.; Finney, E. E.; Finke, R. G. Chem. Mater. 2005, 17, 4925.
(c) Finney, E. E.; Finke, R. G. Chem. Mater. 2008, 20, 1956.
(30) The resultant XPS binding energies were compared to the literature26
(in parentheses). The results confirm that the solid is Ir(0): 577 eV (4p1/2, 578 eV),
494 eV (4p3/2, 495 eV), 311 eV (4d3/2, 312 eV), 298 eV (4d5/2, 297 eV), 64 eV (4f5/2
64 eV), 61 eV (4f7/2, 61 eV).
,
12458 DOI: 10.1021/la101390e
Langmuir 2010, 26(14), 12455–12464

Bayram et al.
Article
Scheme 1. Balanced Stoichiometry for the Formation of Ir(0)/Dry-
HCl from the Precursor [1,5-COD)IrCl]2 and Subsequent Benzene
Hydrogenation under Standard Conditions
amount of H2 needed (∼14 psig H2) to convert (only) the
cyclohexene into cyclohexane. No benzene hydrogenation was
observed (1H NMR was used to confirm that the initial amount
of benzene still remained while the complete conversion of
cyclohexene into cyclohexane had occurred). This cyclohexene
(only) hydrogenation curve is fit roughly (R2 = 0.997) to the
2-step mechanism of nanoparticle nucleation and autocatalytic
growth,28 Figure 3. TEM investigation of the resultant product
once the cyclohexene hydrogenation was complete reveals Ir(0)
nanoparticles with >10 nm in diameter, Figure 4. The finding
that Proton Sponge•HþCl- poisons room temperature benzene
reduction catalysis, but not cyclohexene hydrogenation catalysis,
is of some interest in its own right and is suggestive of either a
different requirement for HCl in these two ostensibly related
hydrogenation reactions or possible different active sites for these
two reactions.
Figure 1. Typical benzene loss vs time plot for the catalytic
hydrogenation of benzene starting with 17.5 mg of [(1,5-COD)-
IrCl]2 (0.052 mmolIr) in1.0 mL(11.2mmol) benzeneat22( 0.1°C
with an initial H2 pressure of 40 ( 1 psig (∼2.7 atm). After an
induction period of 0.7 ( 0.1 h, the hydrogenation starts with an
initial rate of 2.0 ( 0.3 M benzene/h. Complete hydrogenation of
benzene into cyclohexane was achieved in 8.7 ( 0.1 h as confirmed
by 1H NMR analysis.
solid gray catalyst residue collected after the hydrogenation of
benzene was complete (i.e., after 8.7( 0.1 h). The darkening of the
solution within 30 min following the addition of H2 pressure is
consistent with the formation of Ir(0) nanoparticles,31 and con-
firming evidence for those nanoparticles was obtained by ex situ
TEM, Figure 2. The TEM images were taken at two different, key
stages of the reaction: after 4 h of reaction, Figure 2a, and at the
end of the reaction (after 8.7 ( 0.1 h), Figure 2b. The close inspec-
tion of these images reveals smaller, average diameter <10 nm,
nanoparticles present after 4 h of reaction but then larger,
average diameter >20 nm, nanoparticles present at the end of the
reaction—that is, that agglomeration of the Ir(0) nanoparticles to
larger particles occurs as the reaction proceeds.
Experiments Demonstrating the Requirement for the in
situ Formation of Dry-HCl for the Observed Catalytic
Activity. As mass balance requires and the balanced stoichio-
metry in Scheme 1 indicated, one equiv of HCl is produced in situ
per one equiv of Ir(I) reduced by H2.24b Control experiments were
done (vide infra) and reveal that the in situ formation of dry-HCl
is a required component of the observed, high catalytic activity
reported herein. Specifically, 1.0 equiv of Proton Sponge (1,8-bis-
(dimethylamino)naphthalene)32 per equiv of iridium was added
to an otherwise standard conditions benzene hydrogenation to
learn the effects of removing the Hþ (Proton Sponge is a strongly
basic, conjugate acid aqueous pKa = 12.3, weakly coordinating,
scavenger of Hþ that has been shown to be valuable in nanopar-
ticle syntheses24b). No detectable hydrogen uptake was observed in
the presence of 1.0 equiv of Proton Sponge, even after more than
10 h. Immediate darkening of the initially orange reaction solu-
tionupon the application of hydrogen pressure indicates that Ir(0)
nanoparticles were still formed,31 albeit ones significantly less
catalytically active in the presence of the resultant 1.0 equiv of
Proton Sponge•HþCl- that can serve as an iridium nanoparticle
ligand, Scheme 2. As a further control to confirm that a (less
active, more stable) nanoparticle catalyst is formed when 1.0
equiv of Proton Sponge is added, the same experiment was
repeated, except now with just one change: the much more easily
reduced substrate, cyclohexene (0.5 mL), also added. H2 uptake
now occurred, but ceased right after the required stoichiometric
In an attempt to see if more stable, yet still active iridium
nanoparticles could be formed via the addition of <1.0 equiv
Proton Sponge, the addition of 0.02 equiv was tried (the 0.02
equiv being picked since it matches the number of active iridium
atoms found by CS2 poisoning experiments, vide infra). Interest-
ingly, even just 0.02 equiv of Proton Sponge poisons the activity,
Figure 5.
-
-
Very interestingly, however, replacing Cl- by PF6 or BF4
does not yield a better catalyst as one might have guessed. Instead,
these salts yield inferior, totally inactive systems. When the
iridium precursors [(1,5-COD)Ir(CH3CN)2]PF6 (that should
yield HPF6 under H2) and [(1,5-COD)Ir(CH3CN)2]BF4 (that
should yield HBF4 under H2) were employed separately in an
otherwise standard conditions benzene hydrogenation experi-
ments (i.e., 0.052 mmol iridium concentration in 1.0 mL benzene
at 22 ( 0.1 °C and initial H2 pressure of 40 ( 1 psig), neither of the
non-Cl- containing, non-HCl generating precatalysts yielded any
catalytic activity whatsoever, Figure 6. Both [(1,5-COD)Ir-
(CH3CN)2]PF6 and [(1,5-COD)Ir(CH3CN)2]BF4 do, however,
yield black precipitates of bulk iridium metal after ∼9 h along
with a clear (and thus largely Ir(0) nanoparticle-plus aggregate-
free) solution. The inactivity of catalysts formed from the PF6- or
BF4- salts of the iridium precursors is, however, not completely
unexpected since we have previously seen that even these tradi-
tionally weakly coordinating anions PF6- and BF4- can coordi-
nate well to at least Ir(0) nanoparticle surfaces.3 The implication is
that Cl- binds Hþ in preference to the Ir(0) surface, while the
opposite is true for BF4- and PF6-.33 Restated, the Cl- precursor is
superior in yielding weakly ligated/labile ligand nanoparticles plus
-
aggregates as benzene hydrogenation catalysts than are the PF6
and BF4- precatalysts.
Catalyst Lifetime Demonstration for Ir(0)/Dry-HCl System
in the Complete Hydrogenation of Benzene. A catalyst lifetime
experiment reveals 5250 TTOs of complete benzene hydrogena-
tion over 320 h. The initial clear-yellow solution darkens as the
(31) Widegren, J. A.; Finke, R. G. J. Mol. Cat. A: Chem. 2003, 198, 317.
(32) Brzezinski, B.; Schroeder, G.; Grech, E.; Malarski, Z.; Sobczyk, L. J. Mol.
Soc., Perkin Trans. 1991, 2, 1643.
Langmuir 2010, 26(14), 12455–12464
DOI: 10.1021/la101390e 12459

Article
Bayram et al.
Figure 2. TEM images of the solution fraction of the heterogeneous reaction at different key stages of the reaction confirming the
agglomeration of the Ir(0) nanoparticles in the solution fraction: (a) TEM image after 4 h of reaction exhibiting an average diameter <10 nm,
and (b) TEM image of the resultant, final solution once benzene hydrogenation is complete (after 8.7 ( 0.1 h) showing an increased average
diameter of >20 nm.
Scheme 2. Stoichiometry for the Experiment in Which the Hþ of HCl Is Scavenged via 1.0 equiv of Proton Sponge (PS) To Yield More Stable, but
Less Catalytically Active, Ir(0) Nanoparticles
Figure 3. Cyclohexene hydrogenation curve ()) and approxi-
mate fit to 2-step mechanism of nanoparticle formation28 (-)
obtained after the addition of 1.0 equiv of Proton Sponge plus
0.5 mL of cyclohexene under otherwise standard conditions.
The fit is only approximate in this case (R2 = 0.997), not
unexpectedly since agglomerated nanoparticles and bulk metal
are among the products. The resultant rate constants are k1 =
0.352 h-1 and k2corr = 249 M-1 h-1 (k2 being corrected as is
proper by the stoichiometric factor of ∼95; see elsewhere for
the reasons for, and details of, this mathematically required
Figure 4. TEM image of the resultant nanoparticles after cyclo-
hexene hydrogenation was completed. The presence of large Ir(0)
nanoparticles with >10 nm in diameter indicates agglomerated
nanoparticles and bulk metal are among the products.
correction factor when the evolution of the catalyst is being
followed by the cyclohexene hydrogenation reporter reaction
method28).
iridium(0) nanoparticles are first generated, followed by the for-
mation of a precipitate of bulk iridium(0) metal (via XPS30) which
becomes visible after a few hours, eventually yielding a clear, color-
less (i.e., Ir(0) nanoparticle free) solution. The average TOF during
12460 DOI: 10.1021/la101390e
Langmuir 2010, 26(14), 12455–12464

Bayram et al.
Article
Figure 5. Standard conditions complete benzene hydrogenation
(i) without ()), (ii) with 0.02equiv (0), and (iii) with1.0 equiv (Δ) of
Proton Sponge addition. Even just 0.02 equiv of Proton Sponge
(∼1.0 equiv per active iridium, vide infra) yields negligible activity
in comparison to the experiment without Proton Sponge.
Figure 6. Comparison of hydrogenation activities of three different
iridium precursors for complete benzene hydrogenation: [(1,5-
COD)IrCl]2 (0), [(1,5-COD)Ir(CH3CN)2]BF4 (O), and[(1,5-COD)-
Ir(CH3CN)2]PF6 ()). Each experiment was performed under other-
wise identical following conditions: 0.052 mmol iridium in 1.0 mL
benzene (11.2 mmol) at 22 ( 0.1 °C and 40 ( 1 psig initial H2
pressure. In the cases of the [(1,5-COD)Ir(CH3CN)2]BF4 and [(1,5-
COD)Ir(CH3CN)2]PF6 precursors, catalytically inactive, poisoned
bulk Ir(0) are observed.
the 320 h total catalyst lifetime, and 5250 TTOs, is 16.4 h-1, before
complete deactivation by aggregation into bulk metal occurs. The
bulk Ir(0) metal product is of course the thermodynamic sink of
the system, one readily formed here since there is little DLVO
(Derjaugin-Landau-Verwey-Overbeek)34 or other stabilization
present for the initially formed Ir(0) nanoparticles, especially with
Cl- largely tied up as weakly to noncoordinating HCl.
Table 1. Comparison of Iridium Catalysts’ Activity and Percentage of
Active Iridium Atoms Determined by CS2 Poisoning Experiments
under Otherwise Standard Conditions
Redispersibility of the Ir(0)/Dry-HCl Catalyst System
under Standard Conditions. Since aggregated Ir(0) is formed as
catalysis proceeds, one would expect less than 100% catalytic
activity in a second cycle/reuse of the in situ formed Ir(0)/dry-HCl
catalyst system. In addition, the volatile HCl should be lost in
isolating the catalyst. As expected, only some (60%) of the pre-
vious initial catalytic activity is retained in a second, Stardard
Conditions benzene hydrogenation.
Control Experiments with Iridium Black and CS2 Poison-
ing Experiments Allowing the Comparison of the Activity
per Exposed Ir(0) of the In Situ Ir(0)/Dry-HCl System.
Iridium black was examined as a neat-benzene hydrogenation
catalyst and in the absence of HCl;22 an average TOF of 2.5 h-1
under otherwise Standard Conditions was seen, that is, 10-fold
lower than the TOF of 25 h-1 for the in situ Ir(0)/dry-HCl system,
Table 1. CS2 poisoning experiments35 uncovered the primary
source of the difference, ca. 0.02 equiv of CS2 per iridium is
sufficient to completely poison the benzene hydrogenation for
the in situ Ir(0)/dry-HCl system whereas only ca. 0.002 equiv of
CS2 is necessary for the Ir black/no-HCl system. Hence, the in situ
formation of the Ir(0)/dry-HCl system produced a 10-fold more
dispersed catalyst, Table 1. As a simple control to check the CS2
poisoning work, a standard conditions benzene hydrogenation
with 10 times more iridium black (i.e., using 0.520 mmol iridium
average active iridium atom average activity
activity percentage (%) via per active iridium
catalyst
system
iridium
(mmol) (TOF, h-1
)
CS2 poisoning
atom (h-1 Ir-1)
Ir(0)/dry-HCl 0.052
iridium black 0.052
iridium black 0.520
25
2.5
25
2
12.5
12.5
12.5
0.2
0.2
vs 0.052 mmol iridium for Standard Conditions) was performed.
As the above poisoning results predict, that experiment exhibited
the same overall catalytic activity as the 0.052 mmol Ir(0)/dry-
HCl system, showing an average turnover frequency of 25 h-1
Table 1.
,
The CS2 poisoning experiments, along with those using Proton
Sponge and those with the HCl, HPF6 and HBF4 generating
precatalyst salts above, reveal that the Ir(0)/dry-HCl catalyst system
(i) provides a 10-fold more highly dispersed, weakly ligated/labile
ligand nanoparticle plus aggregates catalyst, one where Cl- is
largely unavailable as a ligand for Ir(0) since it is tied up as HCl
and where the only other ligands are benzene, H2 and hydrides
formed from H2.
Two important findings here, then, are that: (a) Cl- containing
precursor systems that generate HCl in situ are preferred, weakly
ligated/labile ligand nanoparticle and aggregate catalyst systems;
and (b) that the active catalyst that results is indistinguishable
from surface Ir(0) of commercial iridium black, except that it is
formed in situ in a 10-fold higher dispersion.
Catalyst Activity Comparison of Ir(0)/dry-HCl System
with the Prior Highest Activity Catalysts. The CS2 poisoning
experiments allow a comparison of the catalytic activity to
the prior best catalysts. Employing 1:1 CS2:Ir poisoning stoichio-
metry assumption yields an estimated per-active site TOF of 1250 h-1
and TTOs of 262 500 for the Ir(0)/dry-HCl catalyst system
(The use of a 1:1 CS2:Ir poisoning stoichiometry provides the
most conservative, least favorable estimate as CS2 is known
to poison 7 or more active sites in some cases31). The highest
TOF values previously reported for the complete hydrogena-
tion of neat benzene at e25 °C and e10 atm are intrazeolite
(33) To determine any effect of CH3CN (present in the other two iridium
precursors along with BF4 or PF6) a control experiment was performed. Specifi-
cally, the addition of even 5 equiv (per iridium) of CH3CN to a standard conditions
benzene hydrogenation with [(1,5-COD)IrCl]2 yielded the same catalytic activity
within the experimental error as standard conditions benzene hydrogenation wit-h
[(1,5-COD)IrCl]2. The implication is that it is not the CH3CN, but rather the PF6
or BF4-, that is the catalyst poison.
(34) DLVO theory34a was developed to describe the stabilization of colloids in
1940s and relies on the Coulombic repulsion between anions and coordinatively
unsaturated electrophilic surface of colloids. Further details of the theory, its
application to transition metal nanoparticles stability, and implications of DLVO
stabilization of nanoparticles are well reviewed for the interested reader.2,3,34b (a)
Verwey, E. J. W.; Overbeek, J. T. G. Theory of the Stability of Lyophilic Colloids,
2nd ed.; Dover Publications: Mineola, NY, 1999; (b) Ott, L. S.; Cline, M. L.; Finke,
R. G. J. Nanosci. Nanotech. 2007, 7, 2400.
(35) Hornstein, B. J.; Aiken, J. D.; Finke, R. G. Inorg. Chem. 2002, 41, 1625.
Langmuir 2010, 26(14), 12455–12464
DOI: 10.1021/la101390e 12461

Article
Bayram et al.
Table 2. Comparison of the Catalytic Activity, Lifetime, and Catalyst Preparation Steps for the Present Ir(0)/Dry-HCl Catalyst System with the
Prior Two Best Catalyst Systems Identified from an Extensive Literature Search of Benzene Hydrogenation at Room Temperature or Lower
Conditions (e25 °C and e10 atm H2 pressure)a
authors
catalyst system
activity (TOF, h-1
)
lifetime (TTO)
catalyst preparation steps
ref
€
Ozkar and
neat benzene at 22 °C
and 2.7 atm H2
1040
2420
ion exchange followed by borohydride
reduction of Ru(III) to Ru(0) within
the cages of zeolite
sonochemical, synthesis of Rh
nanoparticles on carbon nanotubes
one-pot use of [(1,5-COD)IrCl]2 as a
precatalyst
20o
co-workersb
Wai and
neat benzene at 20 °C
and 10 atm H2
neat benzene at 22 °C
and 40 psig (∼2.7 atm) H2
1038 e TOF e 2414c
25 e TOF e 1250d
not reported
20q
co-workers
Finke and
co-workers
5250 e TTO e
this study
262 500d
a The complete table of the 17 prior, most relevant literature studies is provided in the Supporting Information for the interested reader, Table SI-136
.
b The activity and lifetime values in this report is not corrected for the number of exposed surface atoms; that is, the values given are lower limits. c The
lower limit TOF (of 1038 h-1) was defined as the number of molecules reacted per unit weight of catalyst per unit time. The upper limit TOF of 2414 h-1
was calculated from dispersion values evaluated from the mean size of Rh nanoparticles via TEM images.37 d The TOF and TTO values reported herein
are the lowest limit (i.e., considering all the iridum atoms are active catalysts) and then the estimated upper limit calculated from the CS2 poisoning
experiments and assuming a 1:1 CS2:Ir poisoning stoichiometry; see the main text for details.
Figure 7. TEM image of zeolite-supported iridium(0) nanoparticles taken at the end of the hydrogenation of 1.0 mL neat benzene starting
with 17.5 mg of [(1,5-COD)IrCl]2 precatalyst (0.052 mmol Ir) plus 480 mg of zeolite-Y at 22 ( 0.1 °C with an initial H2 pressure of 40 ( 1 psig
(∼2.7 atm). Scale bar corresponds to 25 nm.
ruthenium(0) nanoparticles with a TOF (uncorrected for active
sites) of 1040 h-1 and TTOs (also uncorrected) of 2420.20o The
second most active catalyst system is that reported by Wai and co-
However, supporting the Ir(0) nanoparticles on microporous and
macroporous materials (e.g., zeolite, Al2O3, TiO2, etc.), then test-
ing their activity and lifetime for benzene hydrogenation under
the same mild temperature conditions, is expected to yield an
improved catalyst lifetime.
workers with 1038 e TOF e 2414 h-1 20q
Table 2. Hence and
,
overall, the Ir(0)/dry-HCl catalyst system has comparable TOF
and superior TTO values, without any need for laborious catalyst
preparation steps, in comparison to the prior best two catalyst
systems, Table 2.36
Supporting Ir(0) Nanoparticles on Zeolite-Y to Obtain
More Stable and More Active Catalyst. The observed agglo-
merated Ir(0) nanoparticles and resultant bulk metal after ex-
tensive catalytic cycles make apparent the relatively low level of
nanoparticle stabilization in these weakly ligated/labile ligand
nanoparticles and aggregates. Restated, the weakness of weakly
ligated/labile ligand nanoparticles in solution-based catalysis is
just this, the lack of stabilization in solution of the nanoparticles.
To test this hypothesis, [(1,5-COD)IrCl]2 plus zeolite-Y
(Na56Y, Si/Al: 2.5) was prepared in situ as described in the
Experimental Section. This precatalyst was then used for in situ
generation of the Ir(0)/HCl catalyst during neat-benzene hydro-
genation to cyclohexane at 22 ( 0.1 °C and 40 ( 1 psig (∼2.7 atm)
initial H2 pressure. The initial orange color of the suspension
([(1,5-COD)IrCl]2 plus zeolite-Y powders) darkens within 30 min
of the initiation of the reaction with H2 pressure; TEM analysis
confirms the implied and expected formation of zeolite-supported
Ir(0) nanoparticles, Figure 7. After an induction period of 0.3 (
0.1 h, hydrogenation starts with an initial rate of 2.6 ( 0.3 M
benzene/h and the complete hydrogenation of benzene into
cyclohexane (100%) is achieved after 4.7 ( 0.1 h as confirmed
36b
€
(36) (a) S.O. wishes to point out that in a recent publication some wording
and information in Table S-1 of that publication36b was adapted, inadvertently
without proper referencing36b to the present paper, from the literature search first
done in 2007 as part of the present work and as summarized in Table SI-1 herein
(i.e., and not the other way around, as it might appear since that J. Am. Chem. Soc.
1
by H NMR analysis. Significantly, a smooth sigmoidal curve
for the evolution of the catalyst and catalytic activity is now
observed, along with an excellent (R2 = 0.999) fit to the 2-step
mechanism of nanoparticle formation,28 Figure 8, with rate
paper36b appeared first). (b) Zahmakıran, M.; Tonbul, Y.; Ozkar, S. J. Am. Chem.
€
Soc. 2010, 132, 6541.
12462 DOI: 10.1021/la101390e
Langmuir 2010, 26(14), 12455–12464

Bayram et al.
Article
Conclusions
The primary findings of the present work are:
(i)
That the “weakly ligated/labile ligand” nanoparticle
plus aggregates concept has been explored in ben-
zene hydrogenation starting with a [(1,5-COD)IrCl]2
that evolves under H2 to a Ir(0)/dry-HCl system that
is quite active for the 100% reduction of neat ben-
zene to cyclohexane at 22 ( 0.1 °C and 40 ( 1 psig
(∼2.7 atm) of initial H2 pressure;
(ii) That the CS2-active site corrected, per Ir(0) benzene
hydrogenation catalytic activity (TOF) and lifetime
(TTOs) are atthe high end of whathas beenobserved
at room temperature and mild pressures in compari-
son to the prior literature, 25 e TOF e 1250 h-1 and
5250 e TTO e 262 500;
(iii) That the 10-fold higher activity of the [(1,5-COD)-
IrCl]2 plus H2 system compared to Ir black/no-HCl
catalyst is due to the 10-fold higher dispersion of the
in situ formed Ir(0)/HCl catalyst. The number of
active sites in this more highly dispersed catalyst is
still just 0.02 equiv (2%) out of the total Ir present
(and if one uses a 1:1 CS2:Ir stoichiometry for the
poisoning; the real number of active sites is possibly
g7 fold higher31);
Figure 8. Data (O) and fit (—) for the catalytic hydrogenation of
benzene starting with 17.5 mg [(1,5-COD)IrCl]2 (0.052 mmol Ir)
plus 480 mg zeolite-Y (Ir theoretical, 1%; Ir found, 1%) in 1.0 mL
benzene at 22 ( 0.1 °C and 40 ( 1 psig (∼2.7 atm) initial H2
pressure. Following 0.3 ( 0.1 h of induction period, the hydro-
genation rate increases in sigmoidal curve, which is well fit by the
slow, continuous nucleation Ir(I) f Ir(0) (rate constant, k1), then
autocatalytic surface growth, Ir(I) þ Ir(0) f 2 Ir(0) (rate constant,
k2). This observed sigmoidal curve fits well (R2 = 0.999) to a 2-step
(iv) That, significantly, what one can properly call
weakly ligated/labile ligand Ir(0) nanoparticles and
aggregates have been made in situ as demonstrated
by the fact that they have identical, per exposed Ir(0)
activity within experimental error to Ir(0) black and
that they have no possible ligands other than those
desired for the catalysis plus the at best poor ligand
HCl. Further consistent with the weakly ligated/
labile ligand nanoparticle concept is that the iridium
complexes [(1,5-COD)Ir(CH3CN)2]BF4 and [(1,5-
COD)Ir(CH3CN)2]PF6, employed as precatalysts
under otherwise identical conditions yield negli-
gible benzene hydrogenation activity compared
to the Cl- containing, [(1,5-COD)IrCl]2 precata-
lyst. The implication is that BF4- and PF6- prefer
the Ir(0) surface rather than Hþ, where as Cl-
prefers Hþ leading to a more active Ir(0) catalyst.
nanoparticle formation mechanism28 with rate constants k1
0.081 ( 0.002 h-1 and k2corr = 22.8 ( 0.5 M-1 h-1
=
.
Scheme 3. Minimalistic, 2-Step Nanoparticle Nucleation Then
Autocatalytic Surface Growth Mechanism and Its Implied More
Detailed Steps (Right) for [(1,5-COD)IrCl]2 Precatalyst plus
Zeolite-Y (i.e., A Below) System en Route to the Zeolite-Y
Supported Ir(0) Nanoparticles (i.e., B Below) under H2
constants k1 = 0.081 ( 0.002 h-1 and k2corr = 22.8 ( 0.5 M-1 h-1
The minimalistic nanoparticle formation kinetic scheme and the
correspondent iridium/zeolite species are given in Scheme 3. The
ability to follow in real time, even if indirectly, the formation of a
supported heterogeneous catalyst in contact with solution is not
trivial and of considerable interest by itself; hence such kinetics
and mechanism of the formation of supported heterogeneous
catalysts is being vigorously pursued in separate studies in our
laboratories.38
.
-
That is, Cl- is a preferred ligand over BF4 and
PF6- in the presence of Hþ for at least the present
weakly ligated/labile ligand nanoparticle and ag-
gregates catalysts.
(v) That even 0.02 equiv of Cl- (Cl- formed from the
HCl plus Proton Sponge to give Proton Sponge•
HþCl-) poisons the room temperature benzene hy-
drogenation catalysis by the weakly ligated/labile
ligand nanoparticles and their aggregates. Restated,
Hþ is a key component of the present, weakly ligated/
labile ligand nanoparticle plus aggregates system.
(vi) That the weakness of the weakly ligated/labile
ligand nanoparticles, however, is their expected
poor stabilization due to the lack of stabilizing
ligands. Hence, aggregation and the formation of
bulk metal yields only 60% of the initial activity in
a second cycle of benzene hydrogenation.
The catalytic activity of the zeolite supported iridium(0) nano-
particles is increased almost 2-fold to a TOF of 47 h-1 (vs 25 h-1
for the unsupported Ir(0)/dry-HCl catalyst system), while the
TTO also increased some to 8600 TTOs over 232 h before
deactivation (vs 5250 over 320 h for the unsupported Ir(0)/dry-
HCl catalyst).
In a separate experiment, it was also shown that the vacuum-
dried gray powder form of the resultant Ir(0)/zeolite-Y exhi-
bited 89% of its initial activity in a second run of benzene
hydrogenation.
(vii) That one can, however, generate the Ir(0)/dry-HCl
catalyst in situ from supported [(1,5-COD)IrCl]2 on
zeolite-Y. The resultant Ir(0) nanoparticles are
more stable and exhibit modest improvements
in the catalyst activity (2-fold increase) and
(37) Boudart, M. Chem. Rev. 1995, 95, 661.
(38) (a) Mondloch, J. E.; Yan, X.; Finke, R. G. J. Am. Chem. Soc. 2009, 131,
6389. (b) Mondloch, J. E.; Wang, Q.; Frenkel, A. I.; Finke, R. G. Development Plus
Kinetic and Mechanistic Studies of a Prototype Supported-Nanoparticle Heterogeneous
Catalyst Formation System in Contact with Solution: Ir(1,5-COD)Cl/γ-Al2O3 and Its
Reduction by H2 to Ir(0)n/γ-Al2O3. J. Am. Chem. Soc., 2010, in press.
Langmuir 2010, 26(14), 12455–12464
DOI: 10.1021/la101390e 12463

Article
Bayram et al.
lifetime (1.6 fold increase) under otherwise identi-
cal conditions.
of this hypothesis are in progress and will be reported in due
course.7,38
(viii) That both Ir(0)/dry-HCl and Ir(0)/zeolite-Y catalyst
systems are also relatively “green” in that they satisfy
9 out of 12 proposed principles of green chemistry.21
The Ir(0)/zeolite-Y heterogeneous catalyst is of
course isolable, bottleable and reusable as is the case
with other supported-nanoparticle heterogeneous
catalysts, and finally
Acknowledgment. This work was supported by the Depart-
ment of Energy, Office of Basic Sciences, via DOE Grant SE-
FG02-03ER15453. Partial support by the Turkish Academy of
€
Sciences (to S. Ozkar) is also gratefully acknowledged.
Supporting Information Available: Literature table of
17 studies reporting the complete hydrogenation of benzene
under room temperature conditions and mild pressures
(e25 °C and e10 atm H2 pressure) and attempted fits of
the kinetic data in Figure 1 to either a 2- or 4-step mechan-
ism of nanoparticle and agglomerated nanoparticle forma-
tion, plus brief discussion of the results. This material is
available free of charge via the Internet at http://pubs.acs.
org.
(ix) That Proton Sponge•HþCl- poisons room tem-
perature benzene, but not cyclohexene, catalysis is
of interest and implies either a different requirement
for Hþ in these two, otherwise ostensibly related,
types of hydrogenation reactions or, possibly, dif-
ferent active sites for these two hydrogenations.
This paper is our third exploring the weakly ligated/labile
ligand nanoparticle catalysts hypothesis.3,11 Additional studies
12464 DOI: 10.1021/la101390e
Langmuir 2010, 26(14), 12455–12464